Biotecnologia Bacteriana
Biotecnologia Bacteriana
Biotecnologia Bacteriana
Facultad de Medicina
Departamento de Biología Molecular
Estudio de la red de
regulación global de R388
dirección el trabajo que lleva por título “Estudio de la red de regulación global de
R388”.
Durante estos años ha pasado mucha gente por el laboratorio y cada uno de
ellos ha aportado su granito de arena a este trabajo. Por ello sería muy largo dar
las gracias uno a uno y, con lo despistada que soy yo seguro que me dejo a
alguien importante, así que a todos vosotros os digo: ¡muchas gracias!
Eso sí, no puedo pasar sin destacar a las personas que con las que he
compartido trabajo y que son responsables directas de que esta tesis haya salido
adelante. Raúl, mi jefecillo para la familia, que además de un gran compañero
ha sido casi como un codirector de tesis. Ana, que sin ella nunca hubiera
acabado este trabajo, por su ayuda, por sus ánimos, por todo. Carlos y Lillo,
compartir trabajo (y penas) con ellos ha sido un placer. Fernando (el jefe
supremo para la familia) por su ayuda y paciencia, que mira que hace años que
empezamos con esto. Yera por las jornadas maratonianas de laboratorio
compartidas, por todas las conversaciones y cervezas. Y, por supuesto, a Mapi
que siempre está ahí para todo lo que haga falta, desde darte de comer hasta
describirte todos los plásmidos del mundo. He tenido la suerte de compartir el
trabajo con ellos, de tenerles siempre dispuestos para ayudarme y compartir
buenos momentos tanto profesionales como personales. Gracias, de verdad.
Venga va, antes de que me riñan, a mis chicas del café que aunque ya deberían
saber ellas que tengo mil cosas por las que darles las gracias, sé que Sheila lo
primero que va a mirar de la tesis es esta página. Y a mi familia, que no tengo ni
que decírselo, pero si no fuese por ellos esto ni siquiera habría empezado.
A mi familia
INTRODUCCIÓN 3
1. CONJUGACIÓN BACTERIANA Y PLÁSMIDOS 3
1.1. Transferencia genética horizontal 3
1.2. La conjugación bacteriana 5
1.3. Plásmidos y grupos de incompatibilidad 8
1.4. El esqueleto IncW 13
1.5. El plásmido R388 como prototipo de la familia IncW 15
2. REGULACIÓN TRANSCRIPCIONAL 18
2.1. Motivos de red (Alon, 2007) 18
2.2. Redes de regulación en plásmidos 23
2.3. Estudio de la regulación transcripcional de R388 25
3. MEDIDA DE LA EFICIENCIA DE CONJUGACIÓN 29
OBJETIVOS 35
MATERIALES Y MÉTODOS 39
A. MATERIALES 39
1. CEPAS 39
2. MEDIOS DE CULTIVO, SELECCIÓN Y CONSERVACIÓN 40
3. OLIGONUCLEÓTIDOS 41
4. PLÁSMIDOS 43
B. MÉTODOS 47
1. MÉTODOS GENERALES DE BIOLOGÍA MOLECULAR 47
1.1. Extracción y purificación de DNA 47
1.2. Transformación 47
1.3. Amplificación de DNA: PCR 48
1.4. Digestiones enzimáticas 49
1.5. Ligaciones 49
1.6. Secuenciación de DNA 50
1.7. Electroforesis de DNA 50
1.8. Fijación de células 51
2. CONSTRUCCIÓN DE PLÁSMIDOS 51
3. MÉTODO DE WANNER Y DATSENKO 53
4. CONJUGACIONES 55
4.1. Conjugaciones en filtro 55
4.2. Conjugaciones en placa 55
5. ENSAYOS CON FLUORESCENCIA 56
5.1. Determinación de la actividad transcripcional 56
5.2. Citometría de flujo 58
6. ENSAYOS DE ESTABILIDAD 59
7. MEDIDA DEL NÚMERO RELATIVO DE COPIAS DE PLÁSMIDO 60
8. TEORÍA: ESTIMACIÓN DE LA TASA DE CONJUACIÓN A TIEMPO FINAL (γ) 61
9. OBTENCIÓN DE CURVAS DE CRECIMEINTO Y CÁLCULO DEL TIEMPO DE
GENERACIÓN 63
10. PURIFICACIÓN DE ResP DE R388 64
11. ENSAYOS DE RETARDO EN GEL CON RESP (EMSA) 65
RESULTADOS 69
1. ESTUDIO DE LA REGULACIÓN TRANSCRIPCIONAL DEL PLÁSMIDO R388 69
1.1. Determinación de la actividad transcripcional de los promotores de R388 69
1.2. Actividad de los promotores en presencia de R388 73
1.3. Obtención de un sistema de expresión controlada para los reguladores 75
1.4. Búsqueda combinatoria de reguladores transcripcionales en R388 77
1.5. Deleciones de R388 82
1.6. Aproximación a la caracterización de resP 85
1.7. Efecto de perturbaciones ambientales en la red 91
1.8. Actividad de los promotores en presencia de otros plásmidos genéticamente
parecidos a R388 97
2. APROXIMACIÓN AL ESTUDIO DEL FENOTIPO DE LOS IncW 100
2.1. Funcionamiento de la replicación: medida del número de copias por
cromosoma. 100
2.2. Ensayos de estabilidad del plásmido R388, plásmidos IncW y derivados
R388 102
2.3. Ensayos de conjugación 106
3. SISTEMATIZACIÓN DE LA MEDIDA DE LA CONJUGACIÓN 112
3.1. Puesta a punto de un ensayo de conjugación en el citómetro 112
3.2. Estimación de la cinética de conjugación del oriT de R388 117
3.3. Medida de la conjugación de plásmidos modelo por citometría 121
3.4. Comparación de la capacidad de propagación de los plásmidos modelo 124
DISCUSIÓN 131
1. LA RED DE REGULACIÓN TRANSCRIPCIONAL DE R388 131
2. EFECTO DE LA VARIACIÓN EN EL GENOMA DE LOS IncW EN LA
REGULACIÓN Y EL FENOTIPO 138
3. MEDIDA DE LA CONJUGACIÓN EN EL CITÓMETRO 140
CONCLUSIONES 145
BIBLIOGRAFÍA 149
ANEXO PUBLICACIONES 159
INTRODUCCIÓN
INTRODUCCIÓN
La HGT puede tener lugar entre distintos reinos en una u otra dirección,
habiéndose descrito este proceso entre bacterias (Jain, et al., 1999), desde bacterias a
arqueas y viceversa (Rest & Mindell, 2003) (Gophna, et al., 2004), desde arqueas a
eucariotas (Andersson, et al., 2005), desde bacterias a eucariotas (Watkins & Gray,
2006), desde eucariotas a bacterias (Guljamow, et al., 2007) e incluso entre eucariotas
(Nedelcu, et al., 2008).
3
Desde un punto de vista aplicado, hoy en día la HGT cobra especial importancia
debido a que estos mecanismos son responsables de la diseminación de los genes de
resistencia a antibióticos algo que actualmente se ha convertido en un importante
problema de salud a nivel mundial (de la Cruz & Davies, 2000).
Existen tres mecanismos principales por los que tiene lugar la HGT en el mundo
bacteriano que se resumen en la figura I1. Son los siguientes:
4
Generalmente en los fenómenos de HGT, la transferencia de información
genética está mediada por lo que se conoce como “elementos genéticos móviles”
(MGEs). La primera referencia a los vehículos de transferencia genética horizontal la
encontramos en los años 40, cuando Barbara McClintock descubrió la existencia de
genes capaces de saltar de un cromosoma a otro, algo sorprendente porque rompía con
el dogma de la genética que predicaba que los cromosomas se heredan como unidades
discretas y estables. Por este descubrimiento Barbara McClintock, recibió el premio
Nobel en 1983. Distinguimos diferentes tipos de elementos:
c) Fagos. Los fagos son los virus que infectan a las bacterias. Fueron descubiertos en
1913 por Frederick Twort y son los responsables de los procesos de transducción. A
pesar de ser ubicuos, se considera que la aportación que hacen a los fenómenos de
transferencia en su globalidad es la menos significativa.
5
más allá del mundo procariota, pudiendo ocurrir incluso entre especies muy alejadas
filogenéticamente. Por ejemplo, en condiciones de laboratorio se ha observado
transferencia de DNA conjugativo de bacterias a levaduras o a células de plantas
(Heinemann & Sprague, 1989). En la naturaleza un buen ejemplo de la versatilidad de
este proceso es el sistema de transferencia de DNA del género bacteriano
Agrobacterium al núcleo de células vegetales, mediante una maquinaria de
transferencia muy similar al aparato conjugativo de las bacterias Gram-negativas (Lessl
& Lanka, 1994).
Las bases moleculares del proceso conjugativo han sido ampliamente estudiadas
debido a su importancia biológica. Por un lado este fenómeno presenta importantes
implicaciones desde el punto de vista evolutivo constituyendo un mecanismo que genera
variabilidad genética; por otro lado contribuye de manera determinante a la diseminación
de factores de virulencia (Ziebuhr, et al., 1999) y la propagación de genes de resistencia
a antibióticos (Mazel & Davies, 1999). Además, debido al peculiar procesamiento del
DNA que tiene lugar en la conjugación, se especula sobre la posibilidad de utilizar los
elementos que participan en este mecanismo como herramienta biotecnológica (Llosa
& de la Cruz, 2005).
6
que se descubrió, y se considera el prototipo de los estudios de conjugación. Pasaron
muchos años hasta que se descubrió que la conjugación es un mecanismo extendido
en la naturaleza, y que los plásmidos conjugativos están presentes en prácticamente
todos los microorganismos conocidos. Actualmente el número de plásmidos
conjugativos o movilizables que se descubren sigue creciendo día a día.
1 2 3
4 5 6
7
Figura I2. Conjugación bacteriana (modelo “shoot and pump”).
(1) La conjugación comienza cuando se ponen en contacto dos células bacterianas y una
de ellas lleva un plásmido conjugativo. (2) Se estabiliza el par conjugativo. (3) La relaxasa
corta una cadena del plásmido y se pone en contacto con el sistema de transporte. (4) El
complejo covalente relaxasa-ssDNA es transportado a la célula receptora. (5) Se
recirculariza el plásmido y se sintetiza la cadena complementaria. También en la célula
donadora se sintetiza la hebra desplazada por RCR. (6) Como resultado, cada una de las
células se queda con una copia del plásmido original.
8
copias suelen codificar sistemas de estabilidad destinados a asegurar una correcta
segregación entre las células hijas, mientras que los plásmidos con alto número de
copias muchas veces confían en una distribución estocástica. Los sistemas de
mantenimiento o herencia estable constituyen un conjunto de herramientas moleculares
que los plásmidos utilizan para asegurarse de que la probabilidad de que se generen
bacterias libres de plásmido en la división celular sea muy baja o de que las pocas que
se generen sean eliminadas de la población. Existen tres grupos de mecanismos de
herencia estable: los sistemas de resolución de multímeros, los sistemas de adicción
molecular y los sistemas de partición activa (Zielenkiewicz & Ceglowski, 2001).
9
El primer criterio aplicado en la clasificación de plásmidos estaba relacionado con
la transferencia conjugativa. En sus investigaciones sobre los factores de resistencia a
antibióticos Watanabe descubrió que ciertos plásmidos eran capaces de inhibir la
transferencia de F cuando corresidían en la misma célula donadora. Suponiendo que
ésta era una propiedad fundamental, dividió a los plásmidos en dos grupos (fi+ y fi-) en
función de si inhibían o no la fertilidad del plásmido F (Watanabe, 1963). Más adelante
Meynell y Datta demostraron que esta característica estaba relacionada con el tipo de
pilus sexual que producía el plásmido (Datta, et al., 1966), designando los dos primeros
grupos de incompatibilidad: la clase F (fi+) y la clase I (fi-). La aparición de nuevos tipos
de pili sexual y el descubrimiento de plásmidos no conjugativos hicieron que la
clasificación basada en la inhibición de la fertilidad de F quedase obsoleta, por lo que
se buscó un método alternativo.
Que dos plásmidos incompatibles tienden a ser más cercanos es una regla que
generalmente se cumple, no obstante, el fenómeno de la incompatibilidad es complejo:
los mecanismos por los que se produce son variados y en ocasiones mutaciones
puntuales en los orígenes de replicación hacen a dos plásmidos compatibles, aunque
compartan una extensa identidad de secuencia (Wilson & Figurski, 2002). Los
problemas asociados con la definición de plásmidos incompatibles (distintos
mecanismos de incompatibilidad, exclusión de superficie, plásmidos con varios
replicones, etc…) y sus dificultades técnicas, han hecho que el método tradicional de
clasificación esté actualmente en desuso, habiendo sido sustituido por el tipado del
replicón o la comparación directa entre las secuencias de DNA (Couturier, et al., 1988).
No obstante, la terminología de grupos de incompatibilidad sigue siendo usada
habitualmente en la literatura, especialmente en aquellos grupos en los que existe una
buena correlación entre la identidad de secuencia y el grupo de incompatibilidad (grupos
IncF, IncP, IncN, IncW, etc...).
10
El complejo IncF
11
resistencia que portan así como en algunas funciones de la región de herencia estable
(Schluter, et al., 2007).
El grupo IncW
Los plásmidos IncW reunían algunas características que los hacían un interesante
modelo de estudio: fueron durante mucho tiempo los plásmidos conjugativos conocidos
de menor tamaño, tienen un amplio rango de hospedador y en el caso de pSa se
demostró que posee la habilidad inusual de inhibir la oncogenicidad del plásmido pTi de
Agrobacterium tumefaciens (Close & Kado, 1991).
Estudios tempranos en pSa, R388 y R7K (Bradley & Cohen, 1976) determinaron
que los pili IncW son filamentos rígidos, de 10–12nm de ancho, con una longitud media
12
de unos 450 nm. En los géneros Escherichia, Salmonella, Shigella y Pseudomonas, el
pilus W está presente a una frecuencia media de hasta 3 pili por célula. Junto a los
plásmidos IncP e IncN se los considera a los IncW prototipo de los plásmidos
conjugativos que conjugan preferentemente sobre medio sólido, presentando baja
eficiencia de conjugación en líquido (a diferencia de las clases IncF, IncH e IncI, que son
capaces de conjugar con alta eficiencia en medio líquido).
De los tres plásmidos IncW caracterizados originalmente tan solo pSa y R388
fueron estudiados posteriormente. Los estudios sobre el mecanismo de conjugación de
esta familia se llevaron a cabo principalmente en R388, mientras que el plásmidos pSa
fue utilizado fundamentalmente en estudios sobre el mecanismo de inhibición de la
oncogenicidad del pTi y del replicón IncW. Actualmente R388 es el miembro del grupo
estudiado en más detalle y su secuencia completa se ha empleado para inferir muchas
de las características de la familia (Fernandez-Lopez, et al., 2006).
13
et al., 2000)), R7K (GeneBank AM901564) y R388 (GeneBank BR000038) puso de
manifiesto que los plásmidos IncW mantienen una identidad a nivel de secuencia del
DNA superior al 95%.
14
Figura I3. Estructura del armazón y diferentes inserciones de MGEs en los cinco
IncW analizados (R388, pSa, pIE522, pIE321 y R7K).
Los distintos elementos que componen el armazón IncW se muestran en blanco y negro,
mientras que los genes específicos de cada uno de los plásmidos aparecen destacados
en colores. Tomado de (Revilla, et al., 2008).
Como se muestra en la figura I3, todas estas inserciones de MGEs en los distintos
plásmidos se encuentran en la zona correspondiente a las funciones identificadas como
de mantenimiento estable del plásmido, no afectando en ningún momento a los genes
de la región tra. Además, estas inserciones no producen grandes reorganizaciones en
el armazón, conservándose en todos los casos la sintenia. En el caso de los plásmidos
con el integrón In1 (R388/pSa/pIE522) éste se encuentra insertado en el CDS36. En el
caso de los otros dos plásmidos, en R7K las inserciones afectan a los genes nuc, osa y
klcB, mientras que en pIE321 las inserciones afectan al extremo terminal de ardK.
15
korA orf34 korB orf5 qacEdelta1
trwN sulI
trwL orfA
dhfr
trwM int
intI1
trwK tnpM
resP
trwJ repA
eex
OriV Integrón, resistencia a antibióticos
trwI kfrA
Replicación
trwH nuc1
trwG R388 Estabilidad y mantenimiento
nuc2
trwF 33926 pb osa Orígenes de transferencia y replicación
Ldr1 Procesamiento conjugativo del DNA (Dtr)
trwE
ardC Formación del pilus conjugativo (Mpf)
trwD orf7
orf8
orf9
ssb
trwC
ardK
trwB Ldr2
trwA stbA orf14 orf12
klcb Región replicación
Región conjugación oriT stbB
stbC y mantenimiento
En base a las funciones asignadas a los ORFs, los genes de R388 se dividen en
5 grupos funcionales que se localizan en regiones discretas del genoma que
denominamos módulos. Estos módulos se organizan en dos sectores mayores: un
sector contiene las funciones de mantenimiento general del plásmido (fondo azul en la
figura I4) y el otro contiene los genes dedicados a la conjugación (fondo rojo en la figura
I4); en gris se acota la región correspondiente al integrón.
16
fuera de los IncW. Cada gen presenta historias evolutivas distintas y se habrían
incorporado al genoma a tiempos diferentes, procedentes de fuentes diferentes
(Fernandez-Lopez, et al., 2006).
Dado que R388 tiene solo 2-3 copias por célula, parece poco probable que el
plásmido carezca de sistemas que garanticen su correcta segregación, sin embargo, en
el análisis de la secuencia no se identificaron ATPasas características de ninguno de
los principales sistemas de estabilidad, ni tampoco ninguna evidencia de locus toxina-
antitoxina, encontrándose únicamente un sistema de resolución de dímeros codificado
por la resolvasa ResP. Se han identificado algunos genes sueltos que codifican para
homólogos de proteínas conocidas por jugar un papel secundario en distintos sistemas
de herencia estable (kfrA, nuc1, nuc2 y osa) y que se encuentran agrupados en un
mismo operón (Chen & Kado, 1994) y sería posible que estos genes junto con algunos
situados en su vecindad constituyan algún tipo de sistema de herencia estable aun no
conocido. En los últimos años, el trabajo llevado a cabo por Guynet et al (Guynet, et al.,
2011) sugiere que el operón stbABC de R388 tiene un papel importante en la estabilidad
del plásmido, constituyendo un sistema atípico de estabilidad ligado a la transferencia
conjugativa. StbA es una proteína de unión al DNA (ParB-like) que es estrictamente
necesaria para la estabilidad de R388. Se une a una serie de iterones presentes en el
plásmido (IT-B) y se cree que podría actuar emparejando el plásmido al cromosoma
para utilizar de forma indirecta el sistema de partición bacteriano. StbB tiene un motivo
ATPasa de tipo Walker (ParA-like) pero no parece ser necesaria para la estabilidad del
plásmido, sin embargo, se ha visto que favorece la conjugación. Se ha propuesto que el
sistema StbA-StbB pueda actuar como un interruptor entre los mecanismos de
transferencia vertical y horizontal, constituyendo la primera evidencia de interrelación
entre los procesos de segregación y conjugación plasmídica (Guynet, et al., 2011).
17
2. REGULACIÓN TRANSCRIPCIONAL
Las entradas (inputs) a la red son señales que llevan información de condiciones
del medio externo o de la concentración de los propios componentes de la red, mientras
que las salidas (outputs) son respuestas, producidas por variaciones en la tasa de
transcripción de un conjunto de genes dado. En definitiva, la red es un sistema de
procesamiento de la información, tanto interna como externa, que se encuentra
continuamente ajustando la tasa de síntesis de los componentes de la maquinaria
celular en base a las señales que recibe (Albert, 2005).
a) Regulación simple.
18
interacciones adicionales. Cuando comienza la transcripción, la concentración del gen
X aumenta hasta alcanzar un estado estacionario cuyo nivel será igual a la proporción
entre las tasas de producción y degradación del sistema. El tiempo de respuesta del
sistema, que es definido como el tiempo que lleva alcanzar la mitad entre los niveles
inicial y final, es igual a la vida media del producto génico de manera que cuanto mayor
sea la tasa de degradación menor será el tiempo de respuesta. En el caso de proteínas
que no sean activamente degradadas, el tiempo de respuesta será el correspondiente
a un tiempo de generación.
Dentro de los sistemas de regulación simple existen otros motivos en los que el
factor de transcripción actúa controlando su propia producción estableciéndose una
autorregulación que podrá ser positiva o negativa según se trate de un activador o un
represor.
19
* Autorregulación positiva (PAR): tiene lugar cuando un factor de transcripción activa
su propia producción (Figura I5). Los efectos en este caso son los contrarios a los del
NAR: los tiempos de respuesta se ralentizan y se aumenta la variabilidad, pudiendo
producirse distribuciones bimodales.
20
filtro de desactivación. La presencia de un filtro de activación producirá un retraso en el
desencadenamiento de la respuesta tras la aparición de la señal S (adición de S; señal
ON) pero no habrá ningún retraso cuando ésta se elimina (eliminación de S; señal OFF)
lo que permite evitar que se produzca la respuesta ante pequeños pulsos espurios de
señal. Un filtro de desactivación, por el contrario, no genera retraso tras la aparición de
la señal sino al desaparecer ésta, evitando en este caso que el sistema se desactive en
respuesta a pérdidas momentáneas de la señal.
* I1-FFL: este motivo presenta las propiedades de actuar como un generador de pulsos
y acelerador de respuesta. En este caso los dos brazos del FFL actúan en oposición de
manera que X activa Z pero al mismo tiempo lo reprime a través de Y.
21
función de este motivo de red es permitir la expresión coordinada de un grupo de genes
con función compartida, pudiendo generar programas de expresión temporal de manera
similar al FFL de salida múltiple.
22
de E. coli que conocemos se organiza en torno a unos cuantos DOR fundamentales,
que engloban centenares de genes distintos (Shen-Orr, et al., 2002).
Los plásmidos parecen estar a medio camino entre los fagos y los cromosomas
en su organización genética y la existencia de una regulación coordinada podría ser una
consecuencia de la existencia de esqueletos plasmídicos estables. El estudio de las
redes de regulación de los distintos plásmidos conjugativos vendría a confirmar la
existencia de una organización global estable en sus genomas frente a la mera
yuxtaposición de módulos independientes.
23
* Regulación CUP de pKM101:
Figura I9. Modelo para la regulación de los genes ccg controlados por
elementos CUP en pKM101.
Las proteínas regulatorias ArdK and ArdR (en los círculos) se unirían a las regiones
CUP de los promotores inhibiendo la expresión de los genes ccg que llevan estas
secuencias. Los dos reguladores (ArdK y ArdR) actúan como represores de su
propia síntesis. Tomado de (Delver & Belogurov, 1997).
24
funciones de mantenimiento y en negro se representa un transposón. Las flechas
horizontales indican los promotores. Tomado de (Bingle & Thomas, 2001).
a) Señales de regulación
25
Figura I11. Señales de regulación en R388.
En la figura se muestran las señales de regulación identificadas en R388 mediante el
análisis informático de la secuencia y su localización en el plásmido. En el alineamiento
de las secuencias, los residuos invariables se muestran en rojo sobre fondo amarillo
mientras que los residuos conservados se muestran en negro sobre fondo azul. Tomado
de (Fernandez-Lopez, et al., 2006).
- IT-B2: formado por tres repeticiones, presente en la región intergénica entre ardC
y ORF7.
- IT-B3: formado por cinco repeticiones, situado entre la segunda LDR y ORF12.
26
- .IT-B4: formado por dos repeticiones, situado en la región intergénica entre klcB
y ORF14.
27
Figura I12. Esquema de la red de regulación propuesta para ArdK y alineamiento de
las secuencias de la familia SDR-A.
Las flechas coloreadas en verde indican los ORFs detectados en esta región. Las flechas
marcadas en rayas representan las dos repeticiones directas largas presentes en el
genoma de R388. Se indica la posición de ArdK (flecha roja) así como su circuito de
represión. En la parte inferior se muestra un alineamiento de las secuencias SDR-A.
Tomado de (Fernandez-Lopez, 2007).
28
azul corresponden a genes regulados por StbA, pero que no tienen una función conocida.
Los genes de la región en los que no se detectó la presencia de iterones IT-B y que por
tanto no estarían regulados por StbA, se muestran en color anaranjado Las flechas
transparentes con fondo de cuadros indican las dos repeticiones largas (LDRs) presentes
en el genoma de R388. Tomado de (Fernandez-Lopez, 2007).
29
frecuencia de conjugación, esto es, la tasa entre el número de transconjugantes (T)
respecto al número de donadores (D) o receptores (R).
Este método ha sido ampliamente utilizado debido a que permite una estimación
robusta de la tasa de conjugación, que es insensible a factores como la densidad celular
inicial, la proporción inicial D/R y el tiempo al que se obtienen las muestras de
conjugación, al menos en poblaciones bien mezcladas como las conjugaciones en
líquido (Simonsen, et al., 1990); (Zhong, et al., 2012). Según este modelo la estimación
de punto final de la tasa de conjugación (γep) viene dada por la siguiente ecuación:
𝑇𝑇. 𝑁𝑁
𝛾𝛾𝑒𝑒𝑒𝑒 = 𝜓𝜓𝑚𝑚𝑚𝑚𝑚𝑚 (𝑁𝑁 − 𝑁𝑁0 )−1 ln (1 + )
𝐷𝐷. 𝑅𝑅
donde γep es la tasa de transferencia (ml cel-1 h-1), ψmax es la tasa de crecimiento
bacteriano (h-1), y D, R, T y N son las densidades en un tiempo concreto (cel ml-1) de
donadores, receptores, transconjugantes y el total de las células de cultivo
respectivamente.
30
Recientemente, Zhong et al. (Zhong, et al., 2012) analizaron la posibilidad teórica
de aplicar este método a la conjugación en superficie, usando datos simulados de un
modelo basado en un sistema interactivo de partículas. Concluyeron que las tasas de
transferencia (γ) eran medidas más robustas que las frecuencias de transferencia (T/D
o T/(R+T)), tanto en conjugación en líquido como en conjugación asociada a superficies,
pero solo si las poblaciones de donadores y receptores constituían una capa confluyente
bien formada. Se demostró que, a diferencia de los ratios T/D, T/R o T/N la estimación
de punto final de γ es insensible a la proporción D/R y, hasta cierto punto, al tiempo de
muestreo. (Zhong, et al., 2012).
31
32
OBJETIVOS
OBJETIVOS
A partir del estudio de la regulación del plásmido nos planteamos además otros
objetivos:
35
36
MATERIALES Y
MÉTODOS
MATERIALES Y MÉTODOS
A. MATERIALES
1. CEPAS
Escherichia coli
Cepa Características/Genotipo Referencia
F- endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG
DH5α Φ80dlacZΔM15 Δ(lacZYA-argF)U169, hsdR17(rK- (Grant, et al., 1990)
mK+), λ–
lacIq rrnB3 ΔlacZ4787 hsdR514 DE(araBAD) 567
BW27783 (Khlebnikov, et al., 2001)
DE(rhaBAD) 568 DE(araFGH) Φ(ΔaraEp PCP8-araE)
Otras bacterias
39
Vibrio cholerae
N16961 RifR. Biovar Eltor, serovar O:1 CIP106855
CIP106855
Vibrio cholerae
N16961 SmR. Biovar Eltor, serovar O:1 CIP106851
CIP106851
Alfaproteobacteria
Agrobacterium
Derivada de C58 curada de sus plásmidos naturales (Rosenberg and
tumefaciens
(pTi-, pAT-) Huguet, 1984)
GMI9023
Para los ensayos de fluorescencia el medio rico LB fue sustituido por medio
mínimo M9 preparado utilizando sales M9 comerciales de Sigma (6,8 g/L Na2HPO4; 3
g/L KH2PO4; 1 g/L NH4Cl; 0,5 g/L NaCl), esterilizadas mediante autoclave 121 ºC durante
20 min, a las que posteriormente se añadió 2 mM MgSO4 estéril. El medio se suplementó
con casaminoácidos (0,2% p/v) y glicerol (0,5% p/v).
40
Tetraciclina Clorhidrato de
10 µg/ml En etanol 50%
(Tc10) tetraciclina (Apollo)
Estreptomicina Sulfato de
300 µg/ml En agua
(Sm300) estreptomicina (Apollo)
Rifampicina Rifampicina 99%
50 µg/ml En DMSO
(Rif50) (Apollo)
Dihidrocloruro de
Espectinomicina
20 µg/ml espectinomicina En agua
(Sp100)
pentahidratada (Apollo)
3. OLIGONUCLEÓTIDOS
41
orf14dir GTCGTCCTCGAGACAAACCATCTACCAAGAAG Construcción pGP11
orf14ev GTCGTCGGATCCTCATGTGTGACTCCTGCTTT Construcción pGP11
trwAdir GTCGTCCTCGAGCGTCCGTTTCATTCACTTGT Construcción pGP12
trwArev GTCGTCGGATCCTACTTGGATGGGGTCGCCTA Construcción pGP12
42
CGCGACCGCGCGAAGGGCGGCAAGGGCGCTTTTCA
Construcción de
KfrAKamiK2 TTGTTTGCCTCCTGT
R388ΔkfrA-osa
GGTACCAACCCCAGAGTCCCGCTCAG
AACCGCCCGGCCAGATCAATGCAGGAGTCAAAGCTA
Construcción de
Orf12KamiK1 TGAAGCGCAAATTTGGTACCATCAAGAGACAGGATG
R388Δorf12
AGGATCGT
TGGTGCTTTCTGCTCACGGCGTCGGCTCCGGCTCAA
Construcción de
Orf12KamiK2 TACTCGTGGGGCAGGGTACCAACCCCAGAGTCCCG
R388Δorf12
CTCAG
GGAAAGGGCGGGTTCCCCCGCCCTCCCCTCGGTCA Construcción de
Orf14KamiK2 AATGTGCGCGGCGGTGGTACCAACCCCAGAGTCCC R388Δorf14 y
GCTCAG R388ΔkfrA-orf14
CATCCGCAATTCTAGTGTAGAAGGAGAGTAGAAAAG
Construcción de
KorBKamiK1 TGAATGACCTCAAGGGTACCATCAAGAGACAGGATG
R388ΔkorB
AGGATCGT
GTTCCTCCTGGCCTTACGGCCCGATCTTCAATTCATG
Construcción de
KorBKamiK2 ATACTTCCAGTTCGGTACCAACCCCAGAGTCCCGCT
R388ΔkorB
CAG
Oligo directo para clonar
ResPNdeI GGCTGCATATGCCAAAATACTATGCCTA
resP_R388 en pET29C
Oligo reverso para clonar
ResPXhoI GCTATCTCGAGGGCATGTTGTTCCCTCCCCG
resP_R388 en pET29C
Secuenciación del MCS
pZE05 CCAGCTGGCAATTCCGA
en pUA66
Secuenciación del MCS
GFPseq GGGACAACACCAGTG
en pUA66
Secuenciación del MCS
pBAD1 CTGTTTCTCCATACCCGTT
en pBAD33
Secuenciación del MCS
pBAD2 CTCATCCGCCAAAACAG
en pBAD33
Secuenciación clones
pRODseq1 CTGGAAAGCGGGCAGTGAGCGCAACG
pAC
4. PLÁSMIDOS
Nº Acceso o
Plásmido Resistencia Replicón Descripción
referencia
Plásmido reportero GFP, para
pUA66 Km pSC101 (Zaslaver, et al., 2006) clonar una secuencia controlando la
transcripción de gfpmut2
Expresión controlada por ParaBAD,
pBAD33 Cm p15A (Guzman, et al., 1995)
inducible por arabinosa
Sobreexpresión de proteínas con
pET29c Km pBR322 Novagen cola de histidinas, controlada por el
promotor de T7.
43
Plásmido que expresa mCherry
pROD17 Ap, Km ColE1 (Shaner, et al., 2004) controlada por Plac
BR000038 Plásmido silvestre. Representante
R388 Tp, Su IncW
Datta 1972 de los IncW.
Martinez & de la Cruz,
pSU2007 Tp, Su, Km IncW Derivado KmR de R388
1988
Cm, Gm,
EU419764
pSa Km, Sm, Sp, IncW Plásmido silvestre. Familia IncW.
Watanabe 1968
Su
AM901564
R7K Ap, Sm, Sp IncW Plásmido silvestre. Familia IncW.
Coetzee 1972
EF633507
pIE321 Sm, Tc IncW Plásmido silvestre. Familia IncW.
(Gotz, et al., 1996)
EU247928
Gm, Km,
pIE522 IncW (Gotz, et al., 1996) Plásmido silvestre. Familia IncW.
Su, Tob
Derivado no movilizable de R388 al
pAP711∆ que se le ha eliminado el oriT y
Tc IncW (Demarre, et al., 2005)
OriT::tet sustituído por una resistencia a
tetracicilina (tetA)
pOX38Km marcado con un cassette
pAR109 Cm IncFI (Reisner, 2002)
cat-pLac-yfp* en el locus aph
pLG221 marcado con un cassette
pAR111 Cm IncI1 (Reisner, 2002)
cat-pLac-yfp* en el locus aph
pLG272 marcado con un cassette
pAR113 Cm IncI1 (Reisner, 2002)
cat-pLac-yfp* en el locus aph
RP4 marcado con un cassette cat-
pAR115 Cm IncP (Reisner, 2002)
pLac-yfp* en el locus aph
R1 marcado con un cassette tetRA-
pAR118 Cm IncFII (Reisner, 2002)
pLac-yfp* en el locus cat
R1drd19 marcado con un cassette
pAR120 Tc IncFII (Reisner, 2002)
tetRA-pLac-yfp* en el locus cat
pSU2007 marcado con un cassette
pAR145 Tc IncW (Reisner, 2002)
cat-pLac-yfp* en el locus aph
Vector
Plásmido Resistencia (sitios de Inserto Descripción
clonaje)
Clon de la región promotora del
pUA66 234 pb (del 4214
pGP1 Km operón resP-repA controlando la
(XhoI-BamHI) al 4447 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 302 pb (del 6502
pGP2 Km operón kfrA-osa controlando la
(XhoI-BamHI) al 6803 de R388)
expresión de gfpmut2
Clon de la región intergénica entre
pUA66 203 pb (del 7957
pGP3 Km parB y nuc controlando la expresión
(XhoI-BamHI) al 8159 de R388)
de gfpmut2
Clon de la región promotora del
pUA66 155 pb (del 9911
pGP4 Km operón ardC controlando la
(XhoI-BamHI) al 10065 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 169 pb (del 10959
pGP5 Km operón orf7-orf9 controlando la
(XhoI-BamHI) al 11127 de R388)
expresión de gfpmut2
Clon de la región intergénica entre
pUA66 86 pb (del 11287
pGP6 Km orf7 y orf8 controlando la expresión
(XhoI-BamHI) al 11367 de R388)
de gfpmut2
44
Clon de la región intergénica entre
pUA66 52 pb (del 11548
pGP7 Km orf8 y orf9 controlando la expresión
(XhoI-BamHI) al 11599 de R388)
de gfpmut2
Clon de la región promotora del
pUA66 258 pb (del 12353
pGP8 Km operón ssb-ardK controlando la
(XhoI-BamHI) al 12610 de R388)
expresión de gfpmut2
Clon de la región intergénica entre
pUA66 106 pb (del 12903
pGP9 Km ssb y ardK controlando la expresión
(XhoI-BamHI) al 13008 de R388)
de gfpmut2
Clon de la región promotora del
pUA66 181 pb (del 13911
pGP10 Km operón orf12 y KlcB controlando la
(XhoI-BamHI) al 14092 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 200 pb (del 14973
pGP11 Km operón orf14 controlando la
(XhoI-BamHI) al 15172 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 496 pb (del 17076
pGP12 Km operón stbA-stbC controlando la
(XhoI-BamHI) al 17571 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 496 pb (del 17571
pGP13 Km operón trwA-trwC controlando la
(XhoI-BamHI) al 17076 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 240 pb (del 26252
pGP14 Km operón trwH-trwD controlando la
(XhoI-BamHI) al 26491 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 190 pb (del 31741
pGP15 Km operón korA-trwI controlando la
(XhoI-BamHI) al 31930 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 171 pb (del 33121
pGP16 Km operón kikA-trwN controlando la
(XhoI-BamHI) al 33291 de R388)
expresión de gfpmut2
Clon de la región promotora del
pUA66 171 pn (del 33291
pGP17 Km operón korB controlando la
(XhoI-BamHI) al 33121 de R388)
expresión de gfpmut2
pUA66 389 pb (del 2013 Clon del promotor Pint del integrón
pGP18 Km
(XhoI-BamHI) al 3001 de R388) controlando la expresión de gfpmut2
pUA66 355 pb (del 2613 Clon del promotor Pant del integrón
pGP19 Km
(XhoI-BamHI) al 2967 de R388) controlando la expresión de gfpmut2
pUA66
pGP20 Km
(XhoI-BamHI)
pBAD33 Clon del ORF resP de R388 bajo el
pAR1 Cm resP_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF repA de R388 bajo el
pAR2 Cm repA_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF kfrA de R388 bajo el
pAR3 Cm kfrA_R388
(XbaI-HindIII) control del promotor ParaBAD.
pBAD33 Clon del ORF ardK de R388 bajo el
pAR4 Cm ardK_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF orf7 de R388 bajo el
pAR5 Cm orf7_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF orf8 de R388 bajo el
pAR6 Cm orf8_R388
(XbaI-HindIII) control del promotor ParaBAD.
pBAD33 Clon del ORF orf9 de R388 bajo el
pAR7 Cm orf9_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF orf12 de R388 bajo el
pAR8 Cm orf12_R88
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF klcB de R388 bajo el
pAR9 Cm klcB_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF orf14 de R388 bajo el
pAR10 Cm orf14_R388
(XbaI-HindIII) control del promotor ParaBAD
45
pBAD33 Clon del ORF stbA de R388 bajo el
pAR11 Cm stbA_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF trwA de R388 bajo el
pAR12 Cm trwA_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF korA de R388 bajo el
pAR13 Cm korA_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon del ORF korB de R388 bajo el
pAR14 Cm korB_R388
(XbaI-HindIII) control del promotor ParaBAD
pBAD33 Clon de gfpmut2 bajo el control del
pAR15 Cm gfpmut2 (pUA66)
(XbaI) promotor ParaBAD
pIC1 Tp R388 trwA- R388ΔtrwA
trwA- (Km R388ΔtrwA (Km insertada en el sitio
pIC2 Tp Km R388
insertion) de trwA)
pIC3 Tp R388 trwB- R388ΔtrwB
trwB- (Km R388ΔtrwB (Km insertada en el sitio
pIC4 Tp Km R388
insertion) de trwB)
pIC5 Tp R388 trwC- R388ΔtrwC
trwC- (Km R388ΔtrwA (Km insertada en el sitio
pIC6 Tp Km R388
insertion) de trwC)
pIC7 Tp R388 korB- R388ΔkorB
R388 korB- (Km R388ΔkorB (Km insertada en el sitio
pIC8 Tp Km
insertion) de korB)
R388
pIC9 Tp kfrA-orf14- R388ΔkfrA-orf14
R388 kfrA-orf14- (Km
pIC10 Tp Km R388ΔkfrA-orf14 (Km insertada)
insertion)
R388 R388Δorf12 que también ha perdido
pIC11 Tp orf12– LDRs-
la región entre las LDRs
orf12- LDRs- (Km R388Δorf12 (Km insertada en el
pIC12 Tp Km R388
sitio de orf12)
insertion)
Derivado KmS de pROD17.Contiene
pAC1 Ap pROD17 -
mCherry controlado por Plac
Derivado KmS de pROD17.Contiene
pAC2 Ap pROD17 dsred2
dsred2 controlado por Plac
Derivado KmS de pROD17.Contiene
pAC3 Ap pROD17 gfpmut2
gfpmut2 controlado por Plac
Derivado KmS de pROD17.Contiene
pAC4 Ap pROD17 mKate2
mkate2 controlado por Plac
Derivado KmS de pROD17.Contiene
pAC5 Ap pROD17 cfp
cfp controlado por Plac
Derivado KmS de pROD17.Contiene
pAC6 Ap pROD17 citrus
citrus controlado por Plac
46
B. MÉTODOS
1.2. Transformación
La transformación de DNA por medio de TSB (LB pH 6,1, con 10% de PEG, 5%
DMSO, 10 mM de MgCl2 y 10 mM de MgSO4) (Chung & Miller, 1988), es una técnica
sencilla y barata que permite conseguir frecuencias de hasta 2x108 transformantes por
µg de DNA. Además, las células que han adquirido competencia por este método
pueden ser congeladas para ser transformadas en otro momento sin una pérdida
significativa de eficiencia.
47
y se ponen a crecer las células a 37 ºC con agitación durante 1h para permitir la
expresión del marcador de selección. Posteriormente se siembran los cultivos en placas
de LB-agar con el antibiótico adecuado para seleccionar los transformantes.
48
en el caso de la comprobación de colonias, se utilizó la polimerasa Biotaq (Bioline).
Generalmente las reacciones de PCR se hicieron en un volumen final de 50 µl, y las
concentraciones de cada reactivo se eligieron según la recomendación de cada
fabricante.
1.5. Ligaciones
Los vectores de clonación y los fragmentos de PCR a ligar se digirieron con los
enzimas de restricción seleccionados según el método que se ha indicado
anteriormente. Para aumentar la eficiencia de la ligación y disminuir el número de falsos
positivos, se introduce un paso de desfosforilación en el vector utilizando fosfatasa
alcalina (Roche) que elimina los fosfatos terminales. Siguiendo las instrucciones del
fabricante se utilizan 10 U por reacción y se mantiene 1 hora a 37 ˚C. La reacción se
inactiva mediante incubación a 65˚C 10 min.
49
Para las reacciones de ligación se utilizó la ligasa del fago T4 (Fermentas). Las
reacciones se llevaron a cabo poniendo una relación vector/inserto 1:5 en 20 µl de
volumen final, implicando típicamente 50 - 100 ng de DNA y 4 U de enzima. Las
reacciones de ligación se llevaron a cabo a 22 ˚C durante toda la noche. Como control
negativo de las ligaciones se utilizó la misma reacción sin inserto, añadiendo el volumen
equivalente de agua en este caso hasta completar los 20 µl.
Para cargar las muestras en el gel se mezclan con 1/10 de su volumen de tampón
de carga (azul de bromofenol 0,25% (p/v), glicerol 30% (v/v)). Como marcador de peso
molecular se utilizó el patrón HyperLadder (Biolabs) o 100 bp para bajos pesos
moleculares (Amersham), en función del tamaño del DNA que se estaba analizando.
50
1.8. Fijación de células
Se prepara una disolución de PFA al 7,4% en agua (22,2 g PFA; 200 µl NaOH
10N. Disolver en agua milliQ caliente a 60 ºC, enrasar a 300 ml y filtrar con filtros de
0,22 µm que se hacen alícuotas y se congela. Cuando vamos a trabajar con el fijador,
las alícuotas se descongelan en un baño seco a unos 60-70 ºC.
Para fijar los cultivos, se tomaron 400 µl del cultivo o de la conjugación y se lavaron
con 1 ml de PBS 1x. Posteriormente se centrifugó 3 min a 8000 r.p.m. para precipitar
las células y el pellet se resuspendió en 200 µl de PBS 2x. A esta suspensión se añadió
200 µl del PFA 7,4% y se mezcló suavemente. La mezcla se incubó a temperatura
ambiente durante 15 min. en agitación para evitar que las células se agregasen en el
fondo del tubo. Pasado el tiempo de fijación, la mezcla se centrifugó 3 min a 8000 rpm
para eliminar el sobrenadante, se lavó con 1 ml de PBS 1x y finalmente las células se
resuspendieron en 500 µl de PBS 1x y se conservaron a 4 ºC.
2. CONSTRUCCIÓN DE PLÁSMIDOS
a) Librería pGP
Para analizar la capacidad de las regiones intergénicas del plásmido R388 de actuar
como promotores se construyó una librería clonando estas regiones en el vector
reportero pUA66. El vector pUA66 es un vector de bajo número de copias que
contiene el ORF de la proteína GFPmut2, con su propio sitio de unión al ribosoma,
precedida de dos dianas de restricción XhoI y BamHI. Al clonar secuencias de DNA
entre estas dianas puede ensayarse su actividad transcripcional valorando la
capacidad de promover la transcripción de GFP.
51
clonados en las dianas de restricción XhoI - BamHI del plásmido. Los clones
obtenidos fueron analizados por secuenciación, utilizando para ello el
oligonucleótido pZE05.
b) Librería pAR
Para realizar la librería, los ORFs de los posibles reguladores fueron obtenidos por
digestión usando las enzimas XbaI – HindIII de una librería preexistente de clones
de los ORFs de R388 en el vector pET3a. Los fragmentos obtenidos fueron
purificados y clonados en las dianas del plásmido. Los clones obtenidos fueron
analizados por secuenciación usando los oligonucleótidos pBAD1 y pBAD2. Así
mismo se comprobó la correcta sobreexpresión de las proteínas tras la inducción
con arabinosa mediante geles de tricina.
c) Plásmidos pIC
Se trata de una serie de deleciones de distintos ORFs del plásmido R388 realizadas
mediante recombinación homóloga. Para ello se utilizó un protocolo derivado del
método descrito por Wanner y Datsenko, que se describe en el apartado B3. Los
genes se sustituyeron por un gen de resistencia Km amplificado del vector pUA66,
que posteriormente fue eliminado mediante digestión con KpnI.
Los oligos y los plásmidos generados se encuentran recogidos en las tablas MM3 y
MM4 respectivamente.
d) Plásmidos pAC
52
Posteriormente pAC2, pAC3, pAC4 y pAC5 se obtuvieron por digestion EcoRI - SalI
de pAC1, lo que elimina el gen correspondiente a la mcherry, que fue sustituido por
un amplicón obtenido por PCR de los genes correspondientes a otras proteínas
fluorescentes: dsred2 (Bevis & Glick, 2002), mkate2 (Shcherbo, et al., 2009), venus
(Nagai, et al., 2002) y cfp (Rizzo, et al., 2004). Los clones se comprobaron por
secuenciación con el oligo pRODseq1.
Gen diana
R388
Recombinación
KmR
Cassette de PCR
KpnI KpnI
KmR
R388Δgen_diana::Km
KpnI KpnI
Figura MM1. Esquema de los pasos seguidos para la eliminación de ORFs en R388
siguiendo una modificación del método desarrollado por Wanner y Datsenko.
53
Para preparar las células electrocompetentes y generar recombinantes se
procedió según el método descrito por Lee et al. (Lee, et al., 2001). Se utilizó la cepa de
E.coli DY380 que contiene los genes implicados en la recombinación RED clonados bajo
el control de un represor sensible a temperatura, de manera que cuando la bacteria se
expone a una temperatura de 42 ºC, se induce la expresión de los genes recombinativos,
junto con un inhibidor de las nucleasas RecBCD que impide el ataque al DNA lineal
electroporado. Todo el trabajo rutinario con esta cepa se realizó a 30 ºC para evitar la
expresión innecesaria del sistema de recombinación.
Se diseñaron oligos (tabla MM3) para amplificar por PCR el gen de la resistencia
a Km de pAU66, añadiendo a ambos lados regiones de 30-50 pb de homología con el
gen que se quería eliminar. Una vez obtenida la banda correspondiente a la
amplificación del gen de resistencia, ésta se purificó.
Por otro lado, se introdujo el plásmido R388 por conjugación en la cepa DY380 y
se creció un cultivo del transconjugante a 30 ºC con agitación hasta DO600 0,5-0,7. A
continuación, se incubó el cultivo durante 15 minutos a 42 ºC para inducir la expresión
del sistema de recombinación y tras ese tiempo se procesaron las células para hacerlas
electrocompetentes (apartado B1.2 de materiales y métodos).
54
4. CONJUGACIONES
55
Los cultivos de donadores y receptores se crecieron a partir de colonia puntual en
medio LB con antibiótico y 0,5 mM IPTG (cuando quisimos activar la expresión del
promotor Plac), durante toda la noche a 37 ºC con agitación. Al día siguiente los cultivos
saturados se lavaron para eliminar los restos de antibiótico, centrifugando y
resuspendiendo en LB fresco, se midió la OD600 y se mezclaron donadores y recipientes
en las proporciones adecuadas en función de la OD obtenida, para conseguir la relación
donador:recipiente deseada en cada caso. La mezcla se centrifugó 5 min a 4000 g y se
resuspendió en 1/10 del volumen inicial de LB con 0,5 mM IPTG.
56
ello se siguió el método descrito en (Zaslaver, et al., 2006) con algunas modificaciones.
Las medidas se llevaron a cabo usando un lector de placas Victor3 de Perkin Elmer con
dispensador, equipado con filtros de excitación y emisión adecuados para leer
fluorescencia de GFP (485 ± 15; 535 ± 25) y absorbancia a 600 nm.
57
Figura MM2. Representación del sistema utilizado para determinar la actividad
transcripcional de promotores.
Los fragmentos que contienen los posibles promotores se clonan en el vector pUA66,
controlando la expresión de gfpmut2, de manera que si existe actividad promotora se
transcribirá el gen produciendo proteína fluorescente GFP. Los cultivos de células
conteniendo estas construcciones se miden utilizando un lector de placas Victor3, con
filtros para medir OD y fluorescencia. Con los datos obtenidos se obtiene el perfil de
expresión del promotor representando la fluorescencia por célula (GFP/OD) frente al
crecimiento de los cultivos (OD600), se identifica el estado estacionario y se calcula la
actividad del promotor promediando los valores.
58
En los experimentos de conjugación, los datos recogidos en P1 se analizaron
utilizando gráficos GFP vs. RFP o bien YFP vs. CFP, donde se establecen cuadrantes
para delimitar las regiones correspondientes a los cuatro tipos de poblaciones (figura
MM3)
6. ENSAYOS DE ESTABILIDAD
59
y 5 µl se tranfieron a un matraz con 5 ml de LB fresco. Una muestra de este cultivo
recién inoculado se diluyó y se sembró en placas de LB-agar para obtener el valor a
tiempo 0, el resto del cultivo se puso a crecer a 30 ºC o 37 ºC con agitación durante 12h.
Una vez que el cultivo ha vuelto a saturarse se transfieren de nuevo 5 µl a un matraz
con 5 ml de LB fresco, que se pone a crecer a 30 ºC o 37 ºC con agitación otras 12h.
Este mismo esquema se repitió cada 12h hasta obtener unas 100 generaciones.
Se consideró que cada paso de dilución 1/1000 daba lugar a 10 generaciones antes de
que el cultivo volviese a alcanzar la saturación (12h). Cada 20 generaciones los cultivos
fueron diluidos y plaqueados en LB-agar. La determinación de la fracción de células libre
de plásmido se hizo replicando 100 colonias de cada placa de LB-agar en placas de LB-
agar con el antibiótico de selección del plásmido y de LB-agar sin antibiótico. Las células
que han perdido en plásmido no crecen en la placa con el antibiótico de selección.
El ensayo de PCR tiempo real se realizó utilizando cuatro parejas de oligos que
amplifican regiones diana conservadas en los tres plásmidos y dos parejas de oligos
para amplificar dos genes de referencia en el cromosoma bacteriano, rho y dxs (ver
tabla MM5). La amplificación y análisis por qPCR se llevó a cabo usando un equipo iQ5
de Bio-Rad, determinándose el ciclo umbra (Ct) utilizando el software del equipo.
60
Previamente a los experimentos se comprobó que la eficiencia de todos los oligos era
equivalente utilizando cuatro diluciones 1/10 de la muestra y aplicando la fórmula
(E=10^(-1/pendiente)-1).
Para la obtención del número de copias relativo se calculó ΔCt (Ct cromosoma -
Ct plásmido) para cada uno de los plásmidos, obteniéndose el número de copias como
2ΔCt.
61
Con el fin de obtener un indicador robusto de la transferencia plasmídica,
seguimos el modelo propuesto por Levin (Levin, et al., 1979) según las adaptaciones de
Simonsen (Simonsen, et al., 1990) y Zhong et al. (Zhong, et al., 2012) para estimar la
tasa de conjugación γ.
donde ψ denota la tasa de crecimiento (h-1), γ la tasa de conjugación (ml cel-1 h-1) y e
(µg) es la cantidad de recurso requerido para producir una nueva célula.
donde γmax es el valor estimado por el método de punto final (desde ahora γep) y K es la
constante de media saturación (µg mL-1). De las asunciones sobre dependencia de
nutrientes de las constantes de crecimiento y conjugación, se puede inferir que
γ(C)/ψ(C) = γep/ψmax es una constante. Esta apreciación es crucial en la derivación de
62
las ecuaciones para calcular γ, como se explica en Levin et al. (1979) y Simonsen et al.
(1990). Simonsen demostró mediante simulaciones complejas que estas estimaciones
son bastante robustas y que las diferencias en las tasas de crecimiento de D, R y T
tienen poco efecto en el resultado (excepto cuando las células que llevan el plásmido
tienen ventaja en el crecimiento).
Del desarrollo posterior de las ecuaciones puede ser deducido que la estimación
de la tasa de conjugación de punto final cuando los transconjugantes no son capaces
de volver a transferir el plásmido vendría dada por:
𝑇𝑇
𝛾𝛾𝑒𝑒𝑒𝑒 = 𝜓𝜓𝑚𝑚𝑚𝑚𝑚𝑚 (𝐷𝐷 − 𝐷𝐷0 )−1 ln( + 1)
𝑅𝑅
donde D sería la densidad a tiempo final de los donadores y D0 la densidad inicial (cel
mL-1). Dado que en nuestro modelo experimental la frecuencia de conjugación viene
dada Y = T/(R + T), entonces la ecuación (2) se puede convertir en:
1
𝛾𝛾𝑒𝑒𝑒𝑒 = 𝜓𝜓𝑚𝑚𝑚𝑚𝑚𝑚 (𝐷𝐷 − 𝐷𝐷0 )−1 ln( )
1 − 𝑌𝑌
63
Los valores de absorbancia contenidos en el equipo se transformaron en
equivalentes de OD600 usando una curva calibrada con medidas realizadas en cubetas
de 1cm de lado en un espectrofotómetro convencional. Cada experimento consta de 8
réplicas de cada cultivo, que se promediaron obteniéndose también la desviación
estándar de las medidas, que se representa en las gráficas como barras de error de
cada punto de la curva.
El gen resP de R388 se amplificó utilizando los oligos resPNdeI y resPXhoI (tabla
MM3 de materiales y métodos) que añaden sitios de corte para las enzimas NdeI (en el
extremo 3’ del fragmento) y XhoI (en el 5’) y se clonó en el vector de expresión pET29c
(Novagen) utilizando los sitios de corte del vector. La construcción se introdujo por
electroporación en la cepa E. coli C41 (Miroux and Walker 1996) para su sobreexpresión
y purificación.
64
El sobrenadante obtenido se cargó en una columna His trap HP (GE healthcare)
previamente equilibrada con buffer A (50 mM Tris pH 7.5, 1M NaCl). Tras pasar la
muestra, las proteínas unidas a la columna fueron eluídas en un gradiente lineal de
imidazol (0–500 mM) utilizando buffer B (50 mM Tris pH 7.5, 1 M NaCl, 500 mM
imidazol).
Las fracciones en las que eluyó ResP se cargaron en una columna de gel filtración
Superdex 75 (GE healthcare) equilibrada con buffer C (Tris 50 mM pH 7.5, 150 mM
NaCl, 1 mM EDTA) y eluída con el mismo buffer. Se identificaron las fracciones con
ResP-His purificada y se guardaron a -80 ºC añadiendo 20% glicerol.
65
66
RESULTADOS
RESULTADOS
Numerosos estudios a nivel bioquímico han hecho del plásmido R388 uno de los
modelos fundamentales en el estudio de la transferencia conjugativa en enterobacterias.
Sin embargo, a pesar del conocimiento detallado de la bioquímica de las proteínas
implicadas en la conjugación, al comienzo de esta tesis existía un desconocimiento
general sobre otros aspectos fundamentales de la biología del plásmido. Estudios de
secuenciación revelaron la presencia de un gran número de secuencias repetidas en el
genoma del plásmido (Fernandez-Lopez, et al., 2006). Estos mismos estudios y otros
datos obtenidos en nuestro laboratorio (Fernandez-Lopez, 2007) indicaban la presencia
de al menos 5 reguladores transcripcionales y apuntaban la posible existencia de una
red global de control transcripcional en el plásmido.
69
Los clones de los posibles promotores se introdujeron en la cepa de E. coli
BW27783, que es la cepa utilizada preferentemente en prácticamente todos los
experimentos de este trabajo. Se trata de un derivado de E. coli K12, con la mutación
lacIq, con el sistema de transporte de arabinosa modificado y capaz de crecer en medio
mínimo, por lo que se ajusta muy bien a las necesidades de nuestros experimentos.
Tabla R1. Actividad transcripcional obtenida para los distintos clones pGP en unidades
arbitrarias de fluorescencia.
Actividad transcripcional SD
Plásmido Promotor
2
(u.a.f) (x10 ) (x102)
pGP1 432 43 PresP
pGP2 230 16 PkfrA
pGP3 <1 2 -
pGP4 1136 425 PardC
70
pGP5 808 96 PORF7
pGP6 <1 3 -
pGP7 <1 3 -
pGP8 758 59 Pssb
pGP9 <1 3 -
pGP10 465 53 PORF12
pGP11 197 32 PORF14
pGP12 859 77 PstbA
pGP13 161 29 PtrwA
pGP14 628 63 PtrwH
pGP15 578 54 PkorA
pGP16 672 80 PkikA
pGP17 9 5 PkorB
pGP18 4 9 Pint
pGP19 945 94 Pant
pGP20 <1 2 -
Figura R1. Actividad transcripcional calculada para cada uno de los posibles
promotores de R388 contenidos en la librería pGP.
Las medidas se realizaron en un lector de placas Victor3 según el procedimiento indicado
en materiales y métodos (B5.1), realizándose al menos 10 medidas independientes de
cada dato. Las columnas del gráfico representan el promedio de la actividad obtenida para
cada uno de los posibles promotores y las barras de error corresponden a la desviación
estándar. La actividad transcripcional se expresa como cantidad de fluorescencia por
célula (GFP/OD600).
71
De todas las regiones clonadas en la librería pGP, se consideró que contenían un
promotor aquellas que manifestaron una expresión de fluorescencia de, al menos, dos
veces la medida para el vector pUA66 vacío. Así, de las 20 regiones seleccionadas
inicialmente se determinó que 15 de ellas contendrían un promotor, mientras que las
otras no parecían manifestar actividad transcripcional significativa.
72
1.2. Actividad de los promotores en presencia de R388
Nos encontramos con que seis de los promotores de R388 (PardC, Porf7, Porf14,
Porf12, PkorB, PtrwH) bajaron sus niveles de expresión hasta el nivel del ruido de fondo,
73
mostrando valores por debajo de los obtenidos en las medidas de PlacZ reprimido (1*103
GFP/OD) que habíamos tomado como referencia de un promotor apagado. Otros siete
promotores (PresP, PkfrA, PssB, PstbA, PtrwA, PkorA, PkikA) mostraron valores
alrededor del mismo orden de magnitud de PlacZ reprimido (entre 1 y 7*103 GFP/OD).
Pint Pant
120000
3000
100000
2500
Promotor solo
80000
GFP/OD600
+ R388
GFP/OD600
40000
1000
20000
500
0 0
0,05 0,2 0,35 0,5 0,65 0,8 0,95 0,05 0,2 0,35 0,5 0,65 0,8 0,95
OD600 OD600
74
ii) Que la presencia de la red de regulación del plásmido mantiene estos
promotores altamente reprimidos, de lo que se deduce que la regulación en
el plásmido R388 esta mediada fundamentalmente por represores
transcripcionales.
iii) Que las diferencias entre el estado activado y el estado reprimido de los
promotores del plásmido son, de forma general, muy altas. Decimos por tanto
que los promotores de R388 tienen una alta ganancia.
Para poder estudiar el efecto de los posibles reguladores sobre los promotores,
era necesaria la utilización de un sistema que permitiese inducir la expresión del
regulador de manera controlada y medible.
Para superar este problema, Keasling et al., construyeron la cepa E.coli Bw27784,
que contiene el gen araE bajo un promotor de expresión constitutiva (Khlebnikov, et al.,
2001). De este modo se consigue una inducción de la expresión progresiva y controlable
de ParaBAD al añadir al medio de cultivo cantidades crecientes de arabinosa. Para
poner a prueba este funcionamiento utilizamos una construcción ParaBAD::gfpmut2 en la
cepa E.coli BW27783. Indujimos su expresión mediante cantidades crecientes de
arabinosa y medimos la actividad transcripcional mediante citometría de flujo (Figura
R4). Como se muestra en la figura, nuestros resultados no mostraron rastro de bi-
estabilidad, sino que toda la población se desplaza homogéneamente a lo largo del eje
X a medida que aumenta la concentración de arabinosa. Tras comprobar el
75
comportamiento homogéneo de la población, se midió también la respuesta poblacional
al incremento de las concentraciones de arabinosa mediante la realización de perfiles
de expresión del promotor a concentraciones crecientes de arabinosa, siguiendo el
procedimiento descrito en el apartado B5.1 de materiales y métodos.
[Arabinosa]
Eventos
Intensidad de fluorescencia
FL1
% (p/v)
76
Figura R5. Ajuste de la curva de respuesta a la inducción por arabinosa de
pBAD33::gfpmut2 en BW27783.
En el gráfico de la izquierda se representan los valores medios de fluorescencia obtenidos
para los cultivos de BW27783 (pBAD33::gfpmut2) inducidos, frente al rango de
concentraciones de arabinosa utilizado. En la parte superior del gráfico se muestran los
parámetros del ajuste (en el cuadro interior se muestra lo mismo utilizando los valores de
arabinosa en escala logarítmica).
arabinosa arabinosa
PBAD PBAD
Promotor gfpmut2
Promotor gfpmut2
Figura R6. Esquema del diseño experimental utilizado para identificar los posibles
reguladores.
Por un lado se mide la actividad transcripcional de los distintos clones de la librería pGP
en BW27783 con pBAD33 vacío para obtener la actividad del promotor sin regulador (a la
derecha de la figura). Por otro, se mide la actividad de los mismos clones en BW27783
con cada uno de los plásmidos de la librería pAR que contienen los posibles reguladores
de R388 (a la izquierda de la figura). Los perfiles de expresión se obtuvieron según el
procedimiento descrito en el aparatado B5.1 de MM, añadiendo al medio de cultivo
arabinosa para inducir la expresión de los reguladores clonados en pBAD33.
77
Se determinó la actividad transcipcional de cada uno de los promotores mediante
la obtención del perfil de expresión de fluorescencia en presencia del vector pBAD33
vacío y se comparó este resultado con el obtenido al medirlos en presencia de cada uno
de los posibles reguladores de R388. Los resultados obtenidos se encuentran recogidos
en la tabla R2.
78
presencia del vector pBAD33 vacío mientras que ningún otro regulador transcripcional
mostró un efecto significativo en PresP.
El promotor PkfrA fue reprimido por la proteína KfrA, como mostró la comparación
entre los perfiles de expresión de fluorescencia en presencia (+pAR3) y en ausencia
(+pBAD33) de KfrA. Al igual que para PresP, ningún otro regulador transcripcional alteró
la expresión de PkfrA, ni KfrA afectó a ningún otro promotor de R388.
Como era de esperar, se confirmó el rol regulador de TrwA sobre el operón trwABC
(Moncalian, et al., 1997), observándose una disminución de 150 veces en la expresión
del promotor PtrwA (pGP12) cuando se indujo la expresión de trwA con el plásmido
pAR12. No se detectó ningún efecto por parte de ningún otro regulador sobre este
promotor.
De esta manera, encontramos que las proteínas ResP, KfrA y TrwA regulan sus
propios operones y no muestran efecto sobre otros promotores, estableciéndose en
cada caso un circuito de feedback loop negativo simple. En la figura R7 se muestran los
perfiles de expresión se estos promotores en presencia y ausencia de regulación.
10000
GFP/OD600
GFP/OD600
79
determinamos que una cantidad de arabinosa de 10-3% (p/v) añadida al medio de cultivo
producía la inducción máxima de ardK sin llegar a alterar la velocidad de crecimiento de
las células. Al inducir ArdK con esta concentración de arabinosa encontramos que las
tasas de transcripción de PardC, Porf7, Pssb y Porf14 caen hasta los niveles de fondo
(1*102 unidades de GFP/OD). En el caso de Porf12, que también es reprimido por ArdK,
el índice de represión es menor, disminuyendo la actividad unas 2 veces cuando la
expresión de ardK no está inducida y unas 10 cuando está inducido con 10-3 % de
arabinosa (tabla R2). No se detectaron efectos de ArdK en los promotores que no
contenían los sitios de unión SDR-A.
50000
150000 80000
125000 40000
60000
100000 30000
75000 40000
20000
50000
20000
25000 10000
0 0 0
0,05 0,2 0,35 0,5 0,65 0,8 0,05 0,2 0,35 0,5 0,65 0,8 0,05 0,2 0,35 0,5 0,65 0,8
OD600 OD600 OD600
12000 120000
20000
10000 100000
GFP/OD600
80
En la figura se recogen los perfiles de expresión de los promotores PardC, Porf7, Pssb,
Porf12, Porf14 y PstbA obtenidos según se describe en el apartado B5.1 de materiales y
métodos. En verde se representa el perfil de actividad transcripcional de los promotores
aislados y en negro la actividad medida en presencia del plásmido R388. En tonos rojos
los perfiles obtenidos en presencia del regulador ArdK a mínima y máxima inducción, y en
azules los obtenidos en presencia de StbA a mínima y máxima inducción. La actividad
transcripcional se representa como cantidad de fluorescencia por célula (GFP/DO600)
frente al crecimiento del cultivo (OD600).
Por último, en la región del plásmido que codifica para la formación del pilus
conjugativo existen 4 promotores: PtrwH, PkorA, PkikA y PkorB (pGP14, pGP15, pGP16
y pGP17, respectivamente). Está descrito que estos promotores contienen repeticiones
de la familia SDR-B, identificadas como sitios de unión del regulador KorA (Fernandez-
Lopez, 2007). Así, la actividad transcripcional de estos promotores medida en presencia
de pAR13, que codifica para KorA, muestra un descenso significativo respecto a la
actividad obtenida en ausencia del regulador (con pBAD33 vacío), con índices de
represión que van desde 30 en el caso de PkikA a 90 en el caso de PtrwH.(Figura R9,
tabla R2).
81
PtrwH PkorA
80000 70000
70000
60000
60000
GFP/OD600 50000
GFP/OD600
50000
40000
40000
30000
30000
20000
20000
10000 10000
0 0
0,05 0,2 0,35 0,5 0,65 0,8 0,05 0,2 0,35 0,5 0,65 0,8
OD600 OD600
PkikA PkorB
100000
3000
Promotor solo
80000 2500 + R388
+ pAR13 (ara-)
+ pAR13 (ara+)
GFP/OD600
GFP/OD600
2000
60000 + pAR11 (ara-)
+ pAR11 (ara+)
1500
40000
1000
20000 500
0 0
0,05 0,2 0,35 0,5 0,65 0,8 0,05 0,2 0,35 0,5 0,65 0,8
OD600 OD600
82
Sin embargo, este planteamiento experimental resultó ser más complicado de lo
esperado inicialmente. En muchos casos no fue posible aislar derivados del plásmido
que fuesen knock-out puros, lo que sugiere que la deleción de ciertas partes del genoma
del plásmido resulta en plásmidos tóxicos o inestables. Por ejemplo, las deleciones de
stbA resultaron en un plásmido altamente inestable, que fuimos incapaces de aislar
adecuadamente. Una forma de evitar estos efectos deletéreos fue realizar deleciones
de grandes regiones u operones completos, obteniendo de esta manera, por ejemplo,
la deleción de kfrA con todo el operón (R388Δkfra-osa).
83
presencia del knock-out de trwA,(pIC1). La deleción de trwB, no mostró ningún efecto
sobre la actividad transcripcional de los promotores, obteniéndose el mismo resultado
que con R388 silvestre. Sin embargo, el mutante sin trwC sorprendentemente, produjo
una desregulación de PtrwA, lo que sugeriría un papel regulador de trwC hasta ahora
desconocido.
84
1.6. Aproximación a la caracterización de resP
ResP es una proteína de 218 aminoácidos, que presenta homología con las serín-
recombinasas sitio específicas de tipo 2. Su homólogo más cercano caracterizado es la
resolvasa del Tn21 con un 24% de identidad (Fernandez-Lopez, et al., 2006). El gen
resP está acoplado transcripcionalmente a repA formando el operón replicación de R388
y parece ser responsable de la regulación transcripcional de este operón (figura 10).
-35 SDR-C1
-10
'orf46
SDR-C2
+1
RBS1 RBS2
Integrón
Figura R10. Replicón de R388. En la figura se representan los componentes del operón
de replicación (resP-repA) junto con el oriV de R388.
a) Purificación de ResP
85
HisTrap HP (GE healthcare) y un segundo paso de cromatografía de gel filtración con
Superdex S75 (GE healthcare). El protocolo de purificación se encuentra recogido en
materiales y métodos en el apartado B9.
R388 PresP
backbone +1
-10 resP
'orf46 -35
S3
S2
S5
S6
86
distintos sustratos y los geles se tiñeron con SYBR Gold, escaneándose con un láser de
473 nm.
87
Encontramos que ResP retarda de forma específica al fragmento S2
correspondiente a la región promotora del operón resP-repA, no así la región
inmediatamente anterior correspondiente al sustrato S3 que se comporta igual que el
sustrato inespecífico S4. El sustrato S3 corresponde a la zona inmediatamente anterior
a PresP incluyendo el extremo 3’ del integrón In1. Este sustrato se generó para
descartar que ResP pudiese tener un sitio de unión solapando con el extremo del
integrón, como está descrito para la resolvasa upv1 del plásmido R46 (Tosini, et al.,
1998).
-35 -10 +1
88
En base a los datos de homología de secuencia, es presumible que ResP sea una
resolvasa y por tanto desempeñe un papel en la estabilidad plasmídica evitando la
formación de dímeros de plásmido (Fernandez-Lopez, et al., 2006). Para investigar la
habilidad de ResP para resolver cointegrados se construyeron dos plásmidos con el
replicón de R388 (figura R14 A y B). Uno de ellos (A) contiene el replicón W sin resP
mientras que el otro contiene el replicón completo (B). Además, estos plásmidos llevan
un cassette de resistencia cloranfenicol flanqueado por repeticiones directas (DR) que
contenían la región promotora de resP.
89
pBAD::resP (pAR1). Se crecieron cultivos de DH5α (plásmido C), DH5α (plásmido C +
pBAD33 vacío) y DH5α (plásmido C + pAR1) en LB con Km en el primer caso o LB con
Km/Cm/glucosa para los que contenía un pBAD. Los cultivos crecidos se diluyeron en
LB, sin antibiótico para la que solo llevan el plásmido C, y con Cm y arabinosa para las
que llevan el pAR1 (o el pBAD33). A continuación, se crecieron a 37 ºC durante unas 5
generaciones.
90
Derecha: gráficas de citometría de flujo correspondientes al resultado obtenido para los
cultivos se representa en los dibujos. En rojo de representas las dobles negativas; en azul
las CFP+GFP-, que corresponde a aquellas células en las que se ha eliminado el cassette;
en morado los eventos dobles positivos, que corresponden a las células con el plásmido
C intacto.
En conjunto, nuestros datos indican que ResP constituye una “moonlight protein”
(Jeffery, 1999) capaz de ejecutar dos acciones diferentes. Por un lado ResP actúa como
regulador transcripcional de los genes implicados en la replicación del plásmido. Por
otra parte, ResP mantiene su actividad resolvasa, siendo capaz de resolver
cointegrados.
El hecho de que los promotores del plásmido sean intrínsecamente fuertes, pero
se encuentren completamente silenciados por la acción de represores transcripcionales,
sugería que podrían existir situaciones en las cuales esta represión fuese levantada. Por
esta razón, decidimos analizar si el plásmido era capaz de responder a perturbaciones
ambientales. Para ello, se ensayó tanto la librería de promotores pGP solos como en
presencia de R388 (pGP+R388) en distintas condiciones, midiendo los perfiles de
expresión transcripcional en el lector de placas según el protocolo descrito en materiales
y métodos (apartado B5.1).
91
a) Temperatura
10000
1000 1000 1000
1000
100 100 100
100
10 10 10 10
1 1 1 1
10 10 10 10
1 1 1 1
10 10 10 10
1 1 1 1
100000
1000 1000
GFP/OD600
10000
Temperatura 30ºC
100 100 1000
Temperatura 37ºC
100
10 10 Temperatura 42ºC
10
1 1 1
92
barras de error muestran la desviación estándar de las medidas. El eje Y se muestra en
escala logarítmica para facilitar la representación de los datos.
b) Respuesta SOS
Otra condición que se decidió probar fue si alguno de los promotores presentaba
respuesta SOS. La repuesta SOS se produce en condiciones de estrés celular, por
ejemplo cuando tiene lugar daño en el DNA. Su activación da lugar al
desencadenamiento de cambios en la expresión de genes mediado por LexA
produciendo la activación de mecanismos de reparación del DNA. Se sabe que la
entrada de cadena sencilla en la célula puede desencadenar la respuesta SOS
(Baharoglu, et al., 2010), por lo que parece interesante comprobar si en un plásmido
conjugativo alguno de sus promotores puede verse afectado por esta respuesta de
estrés.
93
PresP PkfrA PardC PORF7
100000 100000 1000000 100000
100000
GFP/OD600
Figura R17. Medida del efecto de la respuesta SOS en los promotores de R388
En la figura se muestran los valores de actividad transcripcional en estado estacionario
para los promotores de R388 tanto solos (en gris) como en presencia de R388 (en negro),
tras inducir la respuesta SOS en las células. Las medidas se realizaron según el protocolo
B5.1 de materiales y métodos. Para desencadenar la respuesta SOS, tras 2h de
crecimiento de los culivos, las células se expusieron a luz UV 5 seg., UV 10 seg. o bien se
añadió 5 µg/ml de mitomicina C al medio y se pusieron a medir en el lector de placas. El
eje Y se muestra en escala logarítmica.
94
artefacto experimental y para confirmarlo haría falta hacer más medidas. En el caso de
Pint podría observarse una respuesta SOS tanto por mitomicina C como por radiación
UV, pero al tratarse de valores de muy bajos están demasiado influenciados por el ruido
de fondo como para poder considerarlos significativos. Para obtener unos datos más
robustos sería necesario repetir los experimentos para tener más réplicas y poder
analizar los datos en profundidad.
80000 60000
60000 40000
40000
20000
20000
0 0
95
Ninguna de las concentraciones de antibiótico parece desencadenar una
respuesta específica en ninguno de los promotores. En el caso del cloranfenicol no se
observa ningún efecto sobre el perfil de expresión de los promotores de R388, ni solos
ni en presencia de R388. En el caso de la rifampicina, su presencia en el medio produce
una reducción general de los niveles GFP/OD de los promotores, que se observa tanto
en ausencia como en presencia de R388, pero no parece desencadenarse ningún efecto
específico llamativo sobre alguno de los promotores en concreto.
Otro factor que pensamos que podría ser importante para el plásmido es la
presencia en el medio de posibles receptores para la transferencia por conjugación. Por
ello, se obtuvieron los perfiles de expresión de los promotores de R388 en presencia de
potenciales receptores conjugativos, mezclando el cultivo de las células con la librería
pGP con células BW27783 vacías. Para ello se crecieron cultivos de BW27783 con los
promotores solos y con los promotores + R388, y se mezclaron en un ratio 1:1 con un
cultivo de células BW27783 justo antes de comenzar las medidas de los perfiles de
expresión. Para mantener la misma cantidad de células productoras de GFP en los
cultivos control y en los mezclados con receptores, se optó por doblar el volumen a
medir en cada pocillo (en vez de 150 µl se pusieron 300 µl) en el caso de los cultivos
con receptores. Los perfiles de expresión obtenidos para cada uno de los promotores
se muestran en la figura R19.
100000 90000
40000
90000
80000
30000
60000
40000 20000
40000 40000
20000 10000
GFP/OD600
0 0 0 0
0.05 0.25 0.45 0.65 0.05 0.25 0.45 0.65 0.05 0.25 0.45 0.65 0.05 0.25 0.45 0.65
OD600
96
Figura R19. Perfiles de expresión de los promotores de R388 en presencia de
posibles receptores.
Se muestran los perfiles de expresión de los promotores R388 obtenidos según el
protocolo de medida de la actividad transcripcional descrito en materiales y métodos
(apartado B5.1). El cultivo con la librería de promotores se mezcló 1:1 con un cultivo de
BW27783 vacío (recipientes potenciales). Para cada promotor se muestra la curva en
ausencia y en presencia de R388, para un experimento control y para el experimento en
presencia de posibles receptores conjugativos.
97
Actividad promotor (x10 2 )
R388 R7K pIE321
P resP 64 ± 9 58 ± 7 38 ± 5
P kfrA 32 ± 4 26 ± 6 24 ± 2
P ardC 5 ±1 13 ± 4 3 ±2
P ORF7 1 ±1 <1 7 ±2
P ssb 35 ± 2 39 ± 3 5 ±2
P ORF12 2 ±2 4 ±3 19 ± 4
P ORF14 4 ±3 1 ±2 17 ± 2
P stbA 48 ± 9 117 ± 19 86 ± 9
P trwA 16 ± 3 185 ± 16 37 ± 10
P trwH <1 <1 6 ±2
P korA 36 ± 3 44 ± 2 87 ± 6
P kikA 59 ± 7 61 ± 3 77 ± 4
P korB 2 ±1 2 ±1 2 ±2
P int 1 ±1 2 ±2 1 ±3
P ant 1045 ± 66 917 ± 37 819 ± 67
98
A pesar de que el esqueleto del plásmido es un 97.5% idéntico entre R388 y R7K,
y 97% entre R388 y pIE321, encontramos bastantes variaciones en la actividad de los
promotores de R388 en presencia de la red de transcripcional de estos plásmidos.
Cuando la librería de promotores se probó en células que contenían R7K, los
promotores PtrwA, PstbA y PardC mostraron un aumento de actividad transcripcional
respecto a la medida en presencia de R388 mientras que el resto de promotores no
mostraron cambios significativos. El promotor que mostró una mayor variación fue PtrwA
con un incremento de unas 10 veces, obteniéndose un nivel de actividad similar al
medido para el promotor solo (ver tabla R1, apartado 1.1). Esto indicaría que, aunque
R7K tiene un gen trwA, su producto no es capaz de reprimir PtrwA de R388. En el caso
de PstbA y PardC, los niveles obtenidos no fueron tan altos como en el plásmido
reportero solo, por lo que parece que los reguladores del plásmido R7K aún son capaces
de reprimir estos promotores aunque solo parcialmente.
99
2. APROXIMACIÓN AL ESTUDIO DEL FENOTIPO DE LOS IncW
Para determinar si las diferencias genéticas entre los plásmidos IncW tenían un
impacto sobre la capacidad replicativa del plásmido, se determinó el número de copias
de los plásmidos R388, R7K y pIE321. Para ello empleamos PCR cuantitativa,
normalizando los valores obtenidos frente al cromosoma para obtener así el número de
equivalentes plasmídicos por copia cromosómica. El protocolo utilizado se describe en
el apartado B7 de materiales y métodos. Se utilizaron cuatro parejas de oligos que
hibridan en secuencias 100% conservadas en los tres plásmidos, utilizándose como
referencia cromosómica los genes rho y dxs del cromosoma bacteriano. El número de
copias relativo se determinó comparando el ciclo umbral obtenido para el plásmido y el
cromosoma (nº copias = 2ΔCt; ΔCt = Ct cromosoma - Ct plásmido). En la tabla R6 se
recogen los valores de Ct medidos y en la figura R21 se han representado los resultados
obtenidos al calcular el número de copias de plásmido por cromosoma de cada una de
las réplicas.
100
Dianas plásmido Cromosoma
orf7left-orf8right klcbleft-orf14right trwCleft-trwDright int14right-resPleft rho dxs
R388 13,36 13,4 13,1 13,53 15,52 15,33
R388 (dil. 1/2) 14,07 14,55 14,32 14,02 15,99 16,21
R388 (dil. 1/4) 15,12 14,69 14,32 14,39 16,76 16,5
R7K 14,25 13,81 13,27 13,46 15,72 15,51
R7k (dil. 1/2) 15,47 15,58 14,85 14,61 16,92 17,16
R7K (dil 1/4) 16,94 16,55 15,89 16,21 18,1 18,08
pIE321 18,48 18,68 17,36 17,74 17,41 17,48
pIE321 (dil 1/2) 21,16 20,73 20,64 20,04 19,42 19,56
pIE321 (dil 1/4) 21,59 21,99 20,93 21,16 20,18 20,38
6
Copias de plásmido por cromosoma
0
pIE321
R388
R7K
Tabla R6 y figura R21. Cálculo del número de copias de plásmido por cromosoma.
En la tabla se recogen los valores Ct obtenidos en los experimentos de PCR tiempo real
para las cuatro dianas plasmídicas y las dos cromosómicas, en las muestras de DNA total
extraído para R388, R7K y pIE321. En la figura se representa el número de copias por
cromosoma obtenido para las distintas réplicas de cada uno de los plásmidos (R388 en
blanco, R7K en azul y pIE321 en rojo).
101
AGGGTACGTGAAATCGCTAATC pIE321
Figura R22. Secuencias IT-A en oriV de los plásmidos IncW. En la figura se muestra
un alineamiento de la secuencia oriV de R388, R7K y pIE321, donde se observa como
pIE321 presenta una repetición más en el iterón de unión de repA (IT-A) que sus
compañeros. En la parte de arriba de la figura se destaca la secuencia extra encontrada
en pIE321 de manera más clara.
102
siguiendo el protocolo descrito en el apartado B6 de MM adaptado de (De Gelder, et al.,
2007). Se añadió también en los experimentos el pSU5000 que es una construcción que
contiene únicamente el replicón de R388 con una resistencia a Cm (Fernandez-Lopez,
2007).
Número de generaciones
0 20 40 60 80 100
R388 100 100 100 100 99 100
R7K 94 83 76 75 68 57
pIE522 98 97 96 96 96 96
pSa 100 99 99 98 95 96
pIE321 100 100 100 100 100 100
Replicón R388 91 78 65 54 44 29
100
% Células con plásmido
80
60
R388
40
pSa
pIE522
20 pIE321
R7K
Replicon R388 (pSU5000)
0
0 20 40 60 80 100
Generaciones
103
Como se puede observar, con excepción de R7K todos los plásmidos IncW
ensayados se mantienen completamente estables en E. coli durante las 100
generaciones, obteniéndose valores del 100% de la población en el caso de R388 y
pIE321 o de alrededor del 95% en pSa y pIE522. Sin embargo, en el caso de R7K el
plásmido parece que no es capaz de mantenerse estable y se va perdiendo, aunque
más lentamente que el pSU5000 que contiene solo el replicón y que se pierde de forma
constante.
Gammaproteobacteria Enterobacteriales
Klebsiella pneumoniae K6 ATCC700603
Salmonella typhimurium LT2
Gammaproteobacteria Pseudomonadales
Acinetobacter baumannii ATCC19606
Pseudomonas putida KT2440
Gammaproteobacteria Vibrionales
Vibrio cholerae CIP106855
Vibrio cholerae CIP106851
Alfaproteobacteria
Agrobacterium tumefaciens GMI9023
90
% células con plásmido
80 80 80
70 70 70
60 60 60
50 50 50
40 40 40
30 30 30
20 20 20
10 10 10
0 0 0
0 50 100 0 50 100 0 50 100
Generaciones Generaciones Generaciones
104
K.pneumoniae V.cholerae
100
S. typhimurium
100 100
90
% células con plásmido
90
Las dos especies en las que R388 resultó más inestable, V. cholerae y P. putida,
se seleccionaron para ensayar en ella la estabilidad del resto de plásmidos IncW.
Añadimos también al experimento los dos mutantes de R388 con deleciones en la región
de establecimiento y mantenimiento estable del plásmido, construcciones pIC10 y
pIC12, para ver si el cambio en esta región producía algún efecto sobre en la estabilidad
del plásmido. Se decidió meter en el ensayo estos dos mutantes por presentar
modificaciones en la región de establecimiento y mantenimiento del plásmido y por ser
esta zona es donde se concentran las principales diferencias entre los IncW. Los
plásmidos se introdujeron en las cepas de V. cholerae y P. putida por conjugación desde
E. coli BW27783 y se siguió la estabilidad durante 100 generaciones, de acuerdo al
105
protocolo descrito en el apartado B6 de MM. En el caso de P. putida se dejó fuera del
experimento R7K debido a la falta de un marcador de selección eficaz para llevar a cabo
el ensayo. En cada caso se realizaron 3 experimentos con 3 réplicas cada uno, los
resultados obtenidos se encuentran representados en la figura R25.
pSU2007
40 40 R388DkfrA-orf14
pSa
pIE522
20 20
pIE321
R388DOrf12DLDRs
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Generaciones Generaciones
106
En primer lugar, llevamos a cabo experimentos de conjugación con R388 desde
E. coli a las distintas especies de proteobacterias ensayadas en los experimentos de
estabilidad (apartado 2.2) así como de éstas a E. coli como cepa receptora. Las
conjugaciones se llevaron a cabo siguiendo el protocolo de conjugación descrito en el
apartado B4.2 de MM, incubando a 30 ºC (P. putida y A. tumefaciens) o 37 ºC (el resto),
con tiempos de conjugación de 1h excepto en V. cholerae y A. baumanii que fue de 4h.
Cada dato es el promedio de al menos 9 experimentos (la columna N recoge el número
de datos promediados en cada dato), calculado a partir de la media de los logaritmos.
En la tabla R8 se recoge las frecuencias de conjugación obtenida para el plásmido R388
entre E.coli y otras especies bacterianas.
3,0E-01
1E-01
3,0E-01
6,7E-02
6,2E-02
9,1E-02
4,9E-02
1E-02
3,3E-05
4,4E-05
6,4E-03
1,0E-02
1E-03
2,5E-03
8,3E-04
1E-04
1E-05
1E-06
107
Tabla R8, Figura R26. Frecuencias de conjugación de R388.
En negro se representan las frecuencias de conjugación obtenidas para R388 desde E.
coli hacia otras bacterias (ver eje) y en gris se representan las frecuencias obtenidas en la
conjugación inversa (desde las bacterias hacia E. coli). Los cálculos para obtener las
medias y desviaciones se han hecho utilizando los logaritmos de las frecuencias.
108
R388
Fc de E.coli a otras bacterias
R388DkfrA-orf14
Frecuencia de conjugación 1E+00
6,9E-01
3,6E-01
4,5E-01
1E-01
1,2E-03
1,0E-01
8,7E-02
1,7E-04
2,6E-04
2,2E-04
1E-02
1E-03
8,1E-06
6,3E-03
6,3E-06
2,8E-06
2,1E-03
1E-04
1E-05
1E-06
Tabla R9, Figura R27. Frecuencias de conjugación de R388 y pICG12 desde E. coli a
otras bacterias. En negro se representan las frecuencias de conjugación obtenidas para
R388 desde E. coli hacia otras bacterias (ver eje) y en gris las obtenidas para el mutante
que carece de los genes de la zona de establecimiento (desde kfrA a orf14, ambos
incluidos). Las columnas corresponden a la media de las frecuencias de conjuagación
obtenidas en los distintos experimentos y las barras de error a la desviación estandar de
los datos. Los cálculos para obtener las medias y desviaciones se han hecho utilizando
los logaritmos de las frecuencias y calculando después el antilogaritmo de la media.
Los valores obtenidos son similares en todos los casos excepto para P. putida, en
que la frecuencia de conjugación de R388 es considerablemente mayor que la del
mutante sin los genes de la zona de establecimiento. Curiosamente este mutante
también mostró un comportamiento diferente al de R388 en los ensayos de estabilidad
en P. putida.
109
F. conjugación por F. conjugación por
donador recipiente
Donador Receptor (media ± desvest) (media ± desvest) N
BW-Nx
(pSU2007) BW-Rif 6,2E-01 2,0E-01 5,0E-01 1,8E-01 20
BW-Nx (pSa) BW-Rif 1,5E-02 1,1E-04 6,3E-03 7,4E-05 15
BW-Nx (R7K) BW-Rif 2,9E-02 1,1E-02 3,7E-02 9,0E-03 9
BW-Nx (pIE321) BW-Rif 2,5E-01 4,5E-02 1,6E-01 3,4E-02 9
BW-Nx (pIE522) BW-Rif 6,1E-05 3,3E-06 5,5E-05 2,9E-06 18
BW-Nx (pIC11) BW-Rif 8,2E-01 5,8E-01 9,0E-01 4,3E-01 9
BW-Nx
(pSU2007) V. cholerae Rif 9,9E-04 6,8E-05 1,8E-02 3,4E-04 15
BW-Nx (pSa) V. cholerae Rif 4,3E-04 1,8E-05 6,6E-03 3,0E-04 13
BW-Nx (R7K) V. cholerae Rif 7,1E-04 3,0E-05 2,3E-03 2,8E-04 13
BW-Nx (pIE321) V. cholerae Rif 4,1E-04 2,0E-04 3,3E-03 2,5E-04 13
BW-Nx (pIE522) V. cholerae Rif 9,5E-05 6,2E-06 4,4E-04 3,1E-05 13
BW-Nx (pIC11) V. cholerae Rif 4,1E-04 2,0E-04 2,7E-03 1,6E-04 12
BW-Nx
(pSU2007) P. putida 4,1E-02 9,6E-04 3,8E-01 1,2E-02 18
BW-Nx (pSa) P. putida 9,9E-06 3,7E-07 1,0E-04 5,0E-06 9
BW-Nx (R7K) P. putida - - - - -
BW-Nx (pIE321) P. putida 1,9E+00 8,3E-05 3,5E+00 1,4E-03 9
BW-Nx (pIE522) P. putida 3,3E-06 6,1E-07 5,3E-05 2,5E-05 9
BW-Nx (pIC11) P. putida 3,2E-05 2,1E-06 2,4E-04 1,6E-05 18
E. coli x P. putida
Fc de los plásmidos IncW E. coli x V. cholerae
E. coli x E. coli
1E+00
Frecuencia de conjugación
1E-01
1E-02
1E-03
1E-04
6,7E-02
2,5E-03
2,3E-05
1,4E-03
8,1E-04
7,0E-02
9,0E-04
3,6E-05
1,2E-04
6,2E-05
6,6E-04
3,0E-01
6,8E-04
1,8E-02
7,3E-02
6,2E-01
1,3E-05
1E-05
1E-06
110
los datos. Los cálculos para obtener las medias y desviaciones se han hecho utilizando
los logaritmos de las frecuencias y calculando después el antilogaritmo de la media.
Los resultados globales de esta sección indican que, pese a su alto nivel de
identidad genética, los plásmidos del grupo IncW muestran diferencias fenotípicas en
las funciones de replicación, estabilidad y conjugación.
111
3. SISTEMATIZACIÓN DE LA MEDIDA DE LA CONJUGACIÓN
Este método es el mismo que el utilizado hace 20 años, es un medio válido pero
tedioso, lo que dificulta la realización de experimentos con muchas muestras o la
obtención de datos estadísticamente robustos. Por otra parte, la conjugación es un
proceso que transcurre de manera geométrica (los transconjugantes actúan como
donadores), por lo que las medidas de conjugación están sometidas a una alta
variabilidad. Por estas razones, decidimos poner a punto un ensayo sistemático que no
requiriese del plaqueo para medir las frecuencias conjugativas, para lo que elegimos la
citometría de flujo.
112
diferente (probablemente mayor) a la de los donadores primarios (Jacob & Wollman,
1956).
Las células receptoras utilizadas son BW-NxR que contienen un plásmido que
codifica para una proteína fluorescente roja, cuya expresión es controlada por Plac. Se
probaron tres construcciones similares pAC1, pAC2 y pAC4, cada uno de ellos
codificando una proteína roja distinta (tabla MM4), obteniéndose en los tres casos
resultados similares (datos no mostrados). La expresión de fluorescencia roja se utilizó
como marcador de células receptoras manteniéndose de forma permanente la
expresión desde Plac mediante la adición de 0,5 mM IPTG al medio de cultivo. Aunque
el pico máximo de excitación para estas proteínas difiere bastante de la longitud máxima
de emisión del láser (488 nm), comprobamos que son suficientemente excitadas dando
resultados reproducibles mediante la detección en los canales FL2 para DsRed2 o en
FL3 para mCherry y mKate2.
113
Figura R29. Esquema del diseño experimental para la medida de la conjugación en
el citómetro de flujo.
En la figura se muestran esquemáticamente los distintos componentes del ensayo de
conjugación por citometría. Representados en gris, los donadores contienen el plásmido
pAP711ΔoriT::Tet (que codifica para todos los genes de R388 pero carece de oriT) junto
con el pAR12 (PstbA-gfpmut2) y se recogen como eventos Q3 en el citómetro (dobles
negativos); los receptores, representados en rojo, contienen pAC1,2 o 4 (Plac-rfp) y se
detectan como eventos Q1 (RFP+GFP-); por último los transconjugantes, representados
en naranja, contendrán el plásmido pAC1,2 o 4 (Plac-rfp) junto con pAR12 (PstbA-
gfpmut2) por lo que se detectan como eventos Q2 (dobles positivos RFP+GFP+).
114
LB y una muestra se fijó con PFA 4% para el análisis por citometría y otra se plaqueó
en placas con los antibióticos de selección para donadores, receptores y
transconjugantes.
0.5 h 1h 2h 4h
RFP
6h 8h 10 h 24 h
RFP
115
experimentos de plaqueo obtenemos la proporción T/(T+R) al dividir el número de
colonias con fenotipo (NxApKm)R entre las colonias obtenidas con fenotipo (NxAp)R
También quisimos comprobar si la fijación con PFA4% puede tener algún efecto
sobre los resultados obtenidos en el análisis por citometría respecto al uso de muestras
frescas sin fijar. Para ello, obtuvimos las frecuencias de conjugación en el citómetro a
distintos tiempos de conjugación de las mismas muestras frescas y fijadas con PFA 4%.
La figura R32 confirma que la fijación de las células con PFA4% y conservación a 4 ºC
antes de la citometría no tiene efectos medibles en el ensayo. Dada la similitud de los
resultados cuando se miden células frescas o fijadas, el uso de la fijación permite
simplificar el ensayo, al poder conservar las muestras y procesarlas todas juntas en el
citómetro.
116
Figura R32. Efecto de la fijación de las
muestras en los ensayos de
citometría de flujo.
Comparación de las frecuencias de
conjugación obtenidas por citometría de
flujo para la misma muestra de células
medidas directamente o tras la fijación
con 4% PFA y mantenimiento a 4 ºC
durante 24h. La muestras se tomaron a
0, 0.5, 1.0, 1.5, 2.0 y 4.0 h de
conjugación. Cada punto corresponde a
la media de 4 experimentos individuales.
El resultado del experimento aparece recogido en la figura R33 que muestra las
frecuencias de conjugación por receptor (T/(R+T)) obtenidas a cada tiempo para cada
mezcla de conjugación.
117
A 1,0
0,9
0,6
50%
0,5 10%
0,4 5%
2,5%
0,3
0,2
0,1
0,0
0 2 4 6 8 10 12 14 16 18 20 22 24
Tiempo (h)
B C
11 11
10,5 10,5
log10 densidad recipientes
log10 densidad donadores
10 10
9,5 9,5
9 9
8,5 8,5
8 50% 8 50%
10% 90%
7,5 7,5
95%
5%
7 7 97,5%
2,5%
6,5 6,5
0 4 8 12 16 20 24 0 4 8 12 16 20 24
Con los datos obtenidos calculamos los valores de tasa de conjugación (γ)
mediante el método “End point” descrito por Simonsen (Simonsen, et al., 1990),
118
modificado como se describe en el aparatado B7 de materiales y métodos para poder
aplicarlo a conjugaciones llevadas a cabo en una superficie sólida. Al representar los
valores γ en vez de las frecuencias de conjugación, obtenemos el gráfico representado
en la figura R34 (A). En los cuatro ratios D/R ensayados la tasa de conjugación (γ)
disminuye ligeramente con el tiempo. Este descenso se muestra más claramente en la
figura R34 (B) en la que se ha representado una sola curva con la media de los valores
de los distintos ratios D/R.
A B -8,5
-8,5
50%
-9 -9
10%
5% -9,5
-9,5
2,5%
log10 γep
log10 γep
-10 -10
-10,5 -10,5
-11 -11
-11,5 -11,5
0 2 4 6 8 0 2 4 6 8
-10,5
para los distintos ratios D/R. (C) Valor de la
-11 tasa de conjugación para cada ratio D/R
-11,5
obtenido promediando las tasas a
obtenidas para los distintos tiempos. Cada
-12 color corresponde a un ratio D/R: 50% en
50% 10% 5% 2,5%
azul; 10% en verde; 5% en rojo y 2,5% en
D/R ratio
morado.
En la tabla R11 se encuentran recogidos los datos representados en las figuras. Con
ellos, obtenemos un valor medio de γ para oriT_R388 de 8*10-11 (-10,10 ± 0,45 (tabla))
que no cambia prácticamente con el ratio D/R y que presenta poca variación con el
tiempo de conjugación. La media varía en una desviación standard según el punto
temporal en el que se tomen los datos. Este indicador es claramente más robusto que
119
las proporciones de la frecuencia de conjugación que se ven muy influenciadas por las
condiciones experimentales.
Tabla R11. Valores de tasa de conjugación en punto final (γ) obtenidos para las distintas
proporciones de donadores (izquierda) y para los distintos tiempos (derecha).
1,0
no mixed Figura R35. Efecto de la mezcla en
0,9 mixed at 6h la conjugación.
Frecuencia conjugación
0,8
Se representan las frecuencias de
0,7
conjugación (Y = T/(R + T)) a las 24h,
0,6
de experiementos con diferentes ratios
0,5
D/R. En negro se representan las
0,4 frecuencias obtenidas a las 24h (no
0,3 mezcladas) y en gris se representa Y
0,2 medido a las 24h pero tras un paso
0,1 intermedio de resuspensión y mezcla
0,0 en LB fresco a las 6h de conjugación
50 10 5 2,5 (mezcladas a las 6h).
% donadores
120
3.3. Medida de la conjugación de plásmidos modelo por citometría
D R T
R388::yfp Plac::cfp Plac::cfp
Plásmido Plásmido
conjugativo conjugativo
121
(CFP-YFP+); los receptores, representados en azul, contienen pAC5 (Plac-cfp) y se
detectan como eventos Q1 (CFP+YFP-); por último los transconjugantes, representados
en verde, contendrán el plásmido pAC5 (Plac-cfp) junto con el plásmido conjugativo
marcado con yfp por lo que se detectan como eventos Q2 (dobles positivos CFP+YFP+).
En todos los casos se analizaron 20.000 eventos en P1 mediante gráficos YFP vs.
CFP donde establecemos las distintas poblaciones: las células donadoras se
identificaron como eventos YFP+CFP- (Q3); las receptoras son células YFP-CFP+ y son
detectadas en Q1; los transconjugantes son dobles positivos YFP+CFP+ que son
detectados en Q2. En la figura R37 se muestra un ejemplo de un experimento con varios
tiempos en el que se puede ver cómo a medida que aumenta el tiempo de conjugación
la población de dobles positivos (P3), correspondiente a los transconjugantes, va
aumentando al mismo tiempo que disminuye la población de receptores (P2).
Tiempo
CFP
122
correspondiente a R388 se ha colocado la obtenida en el apartado 2.1 para oriT_R388
(derecha).
0,9
Frecuencia conjugación
0,8 0,8 oriT_R388 5%
0,7 0,7
(T/(R+T))
(T/(R+T))
0,6 0,6
0,5 0,5
0,4 0,4
0,3 0,3
pSU2007 50%
0,2 0,2
0,1 pSU2007 5%
0,1
0,0 0,0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
Tiempo (h) Tiempo (h)
Como se puede observar las cinéticas de las curvas son ligeramente diferentes.
El plásmido conjugativo (panel izquierdo) conjuga hasta frecuencias un poco más altas
(0.8 frente a 0.7 T/(R+T)) y de forma más rápida: su cinética es de tipo sigmoideo,
mientras que el plásmido movilizable muestra una cinética de tipo parabólico. Esta
diferencia se aprecia de manera mucho más clara cuando reducimos la concentración
inicial de donadores del 50% al 5%. En este caso, las frecuencias alcanzadas con el
plásmido conjugativo son de 0,8-0,9 (T/(R+T)) frente a 0,3 (T/(R+T)) en el movilizable.
Usando la misma estrategia descrita para R388, aplicamos citometría de flujo para
monitorizar la cinética de conjugación de representantes de los principales grupos de
incompatibilidad, que incluían los plásmidos RP4 (IncP), pOX38 (IncFI), R1 (IncFI) y
ColIbp-9 (IncI). Además, para estos dos últimos plásmidos contábamos con dos
mutantes des-reprimidos para la expresión de las funciones de transferencia: R1drd19
(IncFII) y ColIbp-9dr (IncI). Los distintos grupos de incompatibilidad plásmidicos
muestran diferentes estructuras genómicas, preferencias de hospedador y prevalencia
en poblaciones naturales, lo que plantea la pregunta de si su potencial conjugativo
también será también diferente.
123
1,0 RP4-yfp 1,0 pOX38Km-yfp
Frecuencia conjugación 0,9 0,9
Frecuencia conjugación
0,8 0,8
0,7 0,7
(T/(R+T))
(T/(R+T))
0,6 0,6
0,5 0,5
0,4 0,4
0,3 0,3
RP4 50% pOX38 50%
0,2 0,2
0,1 RP4 5% 0,1 pOX38 5%
0,0 0,0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
Tiempo (h) Tiempo (h)
Frecuencia conjugación
0,8 R1 50%
0,8
0,7 0,7 R1 5%
(T/(R+T))
(T/(R+T))
0,6 0,6
0,5 0,5
0,4 0,4
0,3 0,3
R1drd19 50%
0,2 0,2
0,1 R1drd19 5% 0,1
0,0 0,0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
Tiempo (h) Tiempo (h)
0,9
Frecuencia conjugación
0,9
0,8 0,8 pLG272 50%
0,7 0,7
pLG272 5%
(T/(R+T))
(T/(R+T))
0,6 0,6
0,5 0,5
0,4 0,4
0,3 0,3
pLG221 50% 0,2
0,2
0,1 pLG221 5% 0,1
0,0 0,0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
Tiempo (h) Tiempo (h)
124
La conjugación permite a los plásmidos invadir nuevos receptores, por lo que
resulta un mecanismo fundamental en la propagación de estos elementos genéticos. Al
objeto de comparar la capacidad de propagación de los distintos plásmidos modelo
analizados, comparamos su cinética prestando atención a dos variables importantes: la
proporción de células que llevaban el plásmido (P+ = (T + D)/ (D+T+R)) y la proporción
de células donadoras (D/(D+T+R)). Con los datos obtenidos en el apartado anterior,
representamos la cinética de estas variables para cada uno de los plásmidos analizados.
125
cada muestreo temporal. Cada gráfico representa los resultados obtenidos para uno de
los plásmidos (RP4, R388, F, etc.) en conjugaciones en las que se utilizó un 50% de
donadores (D) iniciales.
Los resultados obtenidos para los plásmidos R388 (IncW), RP4 (IncP) y F (IncFI)
(figura R40, paneles A) indican que estos plásmidos lograron ocupaciones de la
población bacteriana cercanas al 100%. Sin embargo, los plásmidos ColIbp-9 y R1,
representantes respectivamente de los grupos de incompatibilidad IncI e IncFII,
incrementaron su prevalencia de manera más discreta (Figura R40, B y C, paneles
izquierdos). La comparación entre P+ y D nos permite, además, cuantificar el efecto que
el crecimiento de los donadores tiene sobre la prevalencia del plásmido en la población.
Como se aprecia en la figura, en los casos de ColIbp-9dr y R1 el incremento en la
prevalencia se debe, fundamentalmente, al crecimiento de los donadores. En el caso de
las variantes des-reprimidas de estos plásmidos (Figura R40 B y C, paneles derechos),
la prevalencia del plásmido llegó a niveles cercanos al 100%, siendo la contribución de
los donadores proporcionalmente mucho menor.
126
Figura R41. Proporción de células P+ y D en los ensayos de conjugación (1:20) para
los distintos plásmidos conjugativos.
Utilizando los datos obtenidos en las cinéticas de conjugación analizadas por citometría
de flujo, calculamos para cada instante de tiempo (t), la proporción de células que
contenían el plásmido (en naranja) y la proporción de donadoras (en azul). Los valores de
ambas proporciones, así como sus desviaciones estándar se muestran en la figura para
cada muestreo temporal. Cada panel representa los resultados obtenidos para
conjugaciones en las que se utilizó un 5% de D iniciales. Los recuadros verdes indican
aquellos valores para los que P+ > 50%.
En conjunto, estos datos nos permitieron establecer que RP4 mostró las mayores
tasas de conjugación, seguido de R1drd19 y R388. Los plásmidos F y ColIb-p9
mostraron valores intermedios, mientras que R1 y ColIbp-9dr incrementaron su
prevalencia debido, fundamentalmente al crecimiento de los donadores, sin que la
conjugación mostrase contribución aparente alguna.
127
128
DISCUSIÓN
DISCUSIÓN
Tal y como habían establecido los experimentos llevados a cabo en nuestro grupo,
el plásmido R388 presenta distintas familias de motivos de repetición de secuencia en
las zonas intergénicas (Fernandez-Lopez, et al., 2006). Estudios preliminares sugerían
estos sitios podrían corresponder a operadores para los distintos reguladores
transcripcionales plasmídicos (Fernandez-Lopez, 2007), mostrando que, en cuatro de
estas repeticiones era posible asociar un regulador transcripcional dado que:
131
(figura D1), con la única excepción de StbA cuyo circuito englobaría genes de los
módulos de mantenimiento y transferencia. Esta situación podría hacer pensar que cada
módulo funcional constituiría un “bloque de construcción”, tratándose de una unidad
funcional, regulatoria y filogenética. Sin embargo, la estructura filogenética del plásmido
pone de manifiesto una situación donde reguladores y dianas presentan orígenes
filogenéticos completamente distintos (Fernandez-Lopez, et al., 2006). Por tanto, la red
de regulación transcripcional del plásmido es capaz de integrar este mosaico genético,
dando lugar al establecimiento de módulos funcionales con una regulación coordinada.
132
presentar algunas propiedades intrínsecas que favorecen la supervivencia y
propagación para un plásmido BHR.
133
Esto se ve claramente en el caso de los promotores de la zona de establecimiento
en el hospedador, que se encuentran regulados tanto por StbA como por ArdK
constituyéndose un motivo de regulación de tipo DOR. Éste contiene dos reguladores
que actúan sobre los mismos promotores pero no interfieren entre sí. Como se ha visto
experimentalmente la presencia de uno de los dos reguladores produce una fuerte
represión en los promotores, de manera que para que los promotores se activen es
necesario que ambos reguladores se despeguen. La presencia de este motivo de red
mantiene PardC, Porf7, Porf12 and Porf14 completamente reprimidos, como se observa
en los perfiles de expresión en presencia del plásmido R388, y sugiere que el plásmido
será capaz de mantener reprimidos estos genes frente a desuniones transitorias de uno
de los dos represores. La medida de los perfiles de expresión en presencia de los R388
knock-out para ardK (pIC9 y pIC11) pone de manifiesto que los promotores permanecen
reprimidos en ausencia de ArdK pero manifestándose un aumento en la actividad
transcripcional cuando únicamente está regulados por StbA.
134
Previamente se habían hecho algunos experimentos coexpresando KorB y KorA desde
distintos vectores (Fernandez-Lopez, 2007), tratando de observar un efecto de KorB
sobre los promotores de R388 pero no se había conseguido ningún resultado que
indicase una función reguladora de KorB, obteniéndose los valores esperados por la
presencia de KorA. Este resultado señala a korB como un posible punto de interés para
un trabajo posterior, al tiempo que pone de manifiesto el interés de conseguir nuevos
knockout de R388, especialmente de ORFs sin función asignada o con homología con
reguladores, para tratar de ampliar el conocimiento sobre la regulación de las funciones
plasmídicas.
135
en cualquier caso preliminar y, aunque se ha comprobado que el gen trwA se encuentra
en el plásmido, habría que descartar que se haya visto afectado de alguna manera.
136
autorrepresión, favorecería la generación de respuestas transitorias rápidas y fuertes en
situaciones específicas, permitiendo al mismo tiempo que el plásmido se mantenga
fuertemente reprimido el resto del tiempo. De esta manera, la red transcripcional de
R388 parece estar diseñada para permanecer apagada la mayoría del tiempo pero
teniendo la capacidad de producir respuestas fuertes transitorias cuando es necesario.
Estas características podrían resultar útiles para la dispersión del plásmido dado que la
selección impone demandas opuestas a la fisiología plasmídica. Así, el plásmido
requiere de un delicado equilibrio entre la expresión de los genes necesarios para su
adecuado mantenimiento y propagación en la población y la carga metabólica causada
al hospedador (Paulsson, 2002).
Por otro lado, como muestran los modelos teóricos y las simulaciones, es
esperable que tras un evento de conjugación los circuitos de feedback negativo
experimenten overshooting transitorio, que será mayor cuanto más fuerte sea el
promotor. El plásmido conjugativo al entrar en un nuevo hospedador se encuentra en
una situación en donde está presente la maquinaria transcripcional pero la red de
regulación se encuentra temporalmente ausente, lo que provocaría la expresión de los
genes plasmídicos hasta que se alcancen las concentraciones de reguladores
necesarias para volver a reprimirlos.
137
similares ha evolucionado redes basadas en feedback loops negativos, está claro que
este diseño debe presentar alguna propiedad adaptativa interesante para el estilo de
vida plasmídico.
Las diferencias detectadas pueden deberse a dos razones principales: por un lado
a que la concentración intracelular de los reguladores transcripcionales es distinta o que
hay una variación en las secuencias de los represores o sus sitios de unión. Variaciones
en la concentración intracelular suelen a su vez ser producidas por distintos
mecanismos, incluyendo modificaciones en el número de copias del plásmido,
variaciones en la fuerza del promotor, en la estabilidad de los RNA o proteínas, etc.
138
De todas las diferencias observadas en las actividades de los promotores, solo en
el caso del pTrwA somos capaces de identificar su causa inmediata. Las variaciones de
secuencias en la región PtrwA-trwA entre R7K y R388 es considerablemente más alta
que en el resto del armazón (12.5% respecto a 2.5%). Un análisis más detallado revela
que los residuos de TrwA implicados en el reconocimiento del sitio del DNA, así como
los sitios de unión son diferentes (Tena, 2013) (Revilla, et al., 2008), lo que explicaría lo
observado en las actividades transcripcionales, donde TrwA_R7K resulta incapaz de
reprimir PtrwA_R388. El resto de variaciones en las actividades promotoras podrían
deberse a distintos mecanismos actualmente aún sin caracterizar. Es llamativo que
pIE321 modifica los niveles de expresión de 9 de los 13 promotores del armazón
mientras que R7K modifica la actividad de tres. pIE321 muestra solo un 0,5% más de
variación respecto a R388 lo que sugiere que una relación no-lineal entre el grado de
homología del genoma y su impacto en la red de regulación.
139
Los plásmidos R388, R7K y pIE321 fueron aislados de diferentes especies
bacterianas, representando el establecimiento exitoso del mismo core plasmídico en
tres entornos genómicos diferentes (Revilla, 2008). La conservación de un core
genómico común con distintas variantes, aisladas de distintas especies, es también una
característica común de otros plásmidos BHR como es el caso de los plásmidos IncP
(Schluter, et al., 2007). Basándonos en los datos obtenidos para R388 y sus variantes,
sugerimos que tras este aparentemente bajo grado de variación en la secuencia, los
plásmidos esconden una sorprendente habilidad para ajustar sus redes regulatorias sin
hacer grandes cambios en su estructura genómica. Esto contrasta con los estudios
hechos recientemente en E. coli que muestran que cambios en la conectividad o en la
expresión de los reguladores globales tienen poco efecto en el perfil transcripcional de
la red de regulación (Isalan, et al., 2008). Esta susceptibilidad al cambio genético
conferiría a los plásmidos la habilidad de adaptarse rápido a nuevos ambientes
genéticos facilitándoles la colonización de distintos tipos de células. Esta sería una
característica favorable para los plásmidos BHR que necesitan sobrevivir en un amplio
abanico de huéspedes.
140
discriminar las distintas poblaciones presentes en la mezcla conjugativa. Gracias a la
capacidad de procesado de esta técnica (Shapiro, 2005), el protocolo permite la
obtención de manera rápida de gran cantidad de datos obteniéndose estimaciones de
las frecuencias de conjugación equivalentes a las obtenidas mediante los protocolos de
plaqueo tradicionales, pero menos ruidosas. La obtención de datos más robustos y el
aumento de la velocidad de procesado de las muestras hacen de esta técnica una
herramienta interesante para el estudio de la conjugación.
141
142
CONCLUSIONES
CONCLUSIONES
ii) Un DOR controlado por ArdK y StbA regula los operones de la zona de
establecimiento en el receptor mientras que un SIM controlado por KorA regula
los promotores del Mpf.
iii) StbA establece un FFL sobre los promotores regulados por KorA,
colocando a StbA como regulador central del plásmido.
C.- La eliminación de korB en el plásmido R388 tiene un efecto sobre los promotores
del Mpf. KorB no actúa directamente sobre los promotores por lo que debe estar
actuando a través de otro factor. Esto pone de manifiesto la existencia de otro nivel de
regulación aún por caracterizar.
D- Los experimentos con ResP indican que constituye una “moonlight protein” capaz
de ejecutar dos acciones diferentes: actúa como regulador transcripcional de los genes
implicados en la replicación del plásmido y mantiene su actividad resolvasa, siendo
capaz de resolver cointegrados.
145
E.- A pesar de que los plásmidos IncW están caracterizados por una alta identidad a
nivel de secuencia, los distintos plásmidos muestran efectos diferentes sobre la red de
regulación y presentan importantes diferencias fenotípicas. Esto indica que pequeñas
variaciones a nivel de secuencia pueden tener importantes consecuencias sobre la
fisiología y rango de hospedador del plásmido.
146
BIBLIOGRAFÍA
BIBLIOGRAFÍA
Albert R (2005) Scale-free networks in cell biology. J Cell Sci 118: 4947-4957.
Alon U (2007) Network motifs: theory and experimental approaches. Nat Rev Genet 8:
450-461.
Andersson JO, Sarchfield SW & Roger AJ (2005) Gene transfers from nanoarchaeota to
an ancestor of diplomonads and parabasalids. Mol Biol Evol 22: 85-90.
Baharoglu Z, Bikard D & Mazel D (2010) Conjugative DNA transfer induces the bacterial
SOS response and promotes antibiotic resistance development through integron
activation. PLoS Genet 6: e1001165.
Belogurov AA, Delver EP & Rodzevich OV (1992) IncN plasmid pKM101 and IncI1
plasmid ColIb-P9 encode homologous antirestriction proteins in their leading regions. J
Bacteriol 174: 5079-5085.
Belogurov AA, Delver EP, Agafonova OV, Belogurova NG, Lee LY & Kado CI (2000)
Antirestriction protein Ard (Type C) encoded by IncW plasmid pSa has a high similarity
to the "protein transport" domain of TraC1 primase of promiscuous plasmid RP4. J Mol
Biol 296: 969-977.
Bevis BJ & Glick BS (2002) Rapidly maturing variants of the Discosoma red fluorescent
protein (DsRed). Nat Biotechnol 20: 83-87.
Bingle LE & Thomas CM (2001) Regulatory circuits for plasmid survival. Curr Opin
Microbiol 4: 194-200.
Boucher HW, Talbot GH, Bradley JS, et al. (2009) Bad bugs, no drugs: no ESKAPE! An
update from the Infectious Diseases Society of America. Clin Infect Dis 48: 1-12.
Boyer LA, Lee TI, Cole MF, et al. (2005) Core transcriptional regulatory circuitry in human
embryonic stem cells. Cell 122: 947-956.
149
Bradley DE (1980) Morphological and serological relationships of conjugative pili.
Plasmid 4: 155-169.
Bradley DE & Cohen DR (1976) Basic characterization of W-pili. J Gen Microbiol 97: 91-
103.
Camas FM, Blazquez J & Poyatos JF (2006) Autogenous and nonautogenous control of
response in a genetic network. Proc Natl Acad Sci U S A 103: 12718-12723.
Cavalli LL, Lederberg J & Lederberg EM (1953) An infective factor controlling sex
compatibility in Bacterium coli. J Gen Microbiol 8: 89-103.
Close SM & Kado CI (1991) The osa gene of pSa encodes a 21.1-kilodalton protein that
suppresses Agrobacterium tumefaciens oncogenicity. J Bacteriol 173: 5449-5456.
Close SM & Kado CI (1992) A gene near the plasmid pSa origin of replication encodes
a nuclease. Mol Microbiol 6: 521-527.
Coetzee JN, Datta N & Hedges RW (1972) R factors from Proteus rettgeri. J Gen
Microbiol 72: 543-552.
Cormack BP, Valdivia RH & Falkow S (1996) FACS-optimized mutants of the green
fluorescent protein (GFP). Gene 173: 33-38.
Chung CT & Miller RH (1988) A rapid and convenient method for the preparation and
storage of competent bacterial cells. Nucleic Acids Res 16: 3580.
Datta N & Hedges RW (1971) Compatibility groups among fi - R factors. Nature 234:
222-223.
Datta N & Hedges RW (1972) R factors identified in Paris, some conferring gentamicin
resistance, constitute a new compatibility group. Ann Inst Pasteur (Paris) 123: 849-852.
Datta N, Lawn AM & Meynell E (1966) The relationship of F type piliation and F phage
sensitivity to drug resistance transfer in R+F- Escherichia coli K 12. J Gen Microbiol 45:
365-376.
Davey HM & Kell DB (1996) Flow cytometry and cell sorting of heterogeneous microbial
populations: the importance of single-cell analyses. Microbiol Rev 60: 641-696.
De Gelder L, Ponciano JM, Joyce P & Top EM (2007) Stability of a promiscuous plasmid
in different hosts: no guarantee for a long-term relationship. Microbiology 153: 452-463.
150
de la Cruz F & Davies J (2000) Horizontal gene transfer and the origin of species: lessons
from bacteria. Trends Microbiol 8: 128-133.
del Campo I, Ruiz R, Cuevas A, Revilla C, Vielva L & de la Cruz F (2012) Determination
of conjugation rates on solid surfaces. Plasmid 67: 174-182.
Delver EP & Belogurov AA (1997) Organization of the leading region of IncN plasmid
pKM101 (R46): a regulation controlled by CUP sequence elements. J Mol Biol 271: 13-
30.
Dostal L, Khare D, Bok J, Heinemann U, Lanka E & Welfle H (2003) RP4 repressor
protein KorB binds to the major groove of the operator DNA: a Raman study.
Biochemistry 42: 14476-14482.
Dower WJ, Miller JF & Ragsdale CW (1988) High efficiency transformation of E. coli by
high voltage electroporation. Nucleic Acids Res 16: 6127-6145.
Eichenberger P, Fujita M, Jensen ST, et al. (2004) The program of gene transcription for
a single differentiating cell type during sporulation in Bacillus subtilis. PLoS Biol 2: e328.
Gogarten JP & Townsend JP (2005) Horizontal gene transfer, genome innovation and
evolution. Nat Rev Microbiol 3: 679-687.
Goldenfeld N & Woese C (2007) Biology's next revolution. Nature 445: 369.
151
Gophna U, Charlebois RL & Doolittle WF (2004) Have archaeal genes contributed to
bacterial virulence? Trends Microbiol 12: 213-219.
Gorai AP, Heffron F, Falkow S, Hedges RW & Datta N (1979) Electron microscope
heteroduplex studies of sequence relationships among plasmids of the W incompatibility
group. Plasmid 2: 485-492.
Grandoso G, Avila P, Cayon A, Hernando MA, Llosa M & de la Cruz F (2000) Two active-
site tyrosyl residues of protein TrwC act sequentially at the origin of transfer during
plasmid R388 conjugation. J Mol Biol 295: 1163-1172.
Grant SG, Jessee J, Bloom FR & Hanahan D (1990) Differential plasmid rescue from
transgenic mouse DNAs into Escherichia coli methylation-restriction mutants. Proc Natl
Acad Sci U S A 87: 4645-4649.
Guynet C, Cuevas A, Moncalian G & de la Cruz F (2011) The stb operon balances the
requirements for vegetative stability and conjugative transfer of plasmid R388. PLoS
Genet 7: e1002073.
Guzman LM, Belin D, Carson MJ & Beckwith J (1995) Tight regulation, modulation, and
high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol
177: 4121-4130.
Halary S, Leigh JW, Cheaib B, Lopez P & Bapteste E (2010) Network analyses structure
genetic diversity in independent genetic worlds. Proc Natl Acad Sci U S A 107: 127-132.
Heinemann JA & Sprague GF, Jr. (1989) Bacterial conjugative plasmids mobilize DNA
transfer between bacteria and yeast. Nature 340: 205-209.
Jain R, Rivera MC & Lake JA (1999) Horizontal gene transfer among genomes: the
complexity hypothesis. Proc Natl Acad Sci U S A 96: 3801-3806.
152
Kalir S & Alon U (2004) Using a quantitative blueprint to reprogram the dynamics of the
flagella gene network. Cell 117: 713-720.
Kalir S, Mangan S & Alon U (2005) A coherent feed-forward loop with a SUM input
function prolongs flagella expression in Escherichia coli. Mol Syst Biol 1: 2005 0006.
Kornacki JA, Burlage RS & Figurski DH (1990) The kil-kor regulon of broad-host-range
plasmid RK2: nucleotide sequence, polypeptide product, and expression of regulatory
gene korC. J Bacteriol 172: 3040-3050.
Kostelidou K & Thomas CM (2002) DNA recognition by the KorA proteins of IncP-1
plasmids RK2 and R751. Biochim Biophys Acta 1576: 110-118.
Lanza VF, de Toro M, Garcillan-Barcia MP, Mora A, Blanco J, Coque TM & de la Cruz F
(2014) Plasmid flux in Escherichia coli ST131 sublineages, analyzed by plasmid
constellation network (PLACNET), a new method for plasmid reconstruction from whole
genome sequences. PLoS Genet 10: e1004766.
Lawrence JG & Ochman H (2002) Reconciling the many faces of lateral gene transfer.
Trends Microbiol 10: 1-4.
Lederberg J & Tatum EL (1946) Gene recombination in Escherichia coli. Nature 158:
558.
Lee C, Kim J, Shin SG & Hwang S (2006) Absolute and relative QPCR quantification of
plasmid copy number in Escherichia coli. Journal of Biotechnology 123: 273-280.
Lee EC, Yu D, Martinez de Velasco J, et al. (2001) A highly efficient Escherichia coli-
based chromosome engineering system adapted for recombinogenic targeting and
subcloning of BAC DNA. Genomics 73: 56-65.
Lee TI, Rinaldi NJ, Robert F, et al. (2002) Transcriptional regulatory networks in
Saccharomyces cerevisiae. Science 298: 799-804.
Lessl M & Lanka E (1994) Common mechanisms in bacterial conjugation and Ti-
mediated T-DNA transfer to plant cells. Cell 77: 321-324.
Levin BR, Stewart FM & Rice VA (1979) The kinetics of conjugative plasmid
transmission: fit of a simple mass action model. Plasmid 2: 247-260.
Levin BR, Lipsitch M, Perrot V, et al. (1997) The population genetics of antibiotic
resistance. Clin Infect Dis 24 Suppl 1: S9-16.
153
Licht TR, Christensen BB, Krogfelt KA & Molin S (1999) Plasmid transfer in the animal
intestine and other dynamic bacterial populations: the role of community structure and
environment. Microbiology 145 ( Pt 9): 2615-2622.
Lilley AK & Bailey MJ (2002) The transfer dynamics of Pseudomonas sp. plasmid
pQBR11 in biofilms. FEMS Microbiol Ecol 42: 243-250.
Loper JE & Kado CI (1979) Host range conferred by the virulence-specifying plasmid of
Agrobacterium tumefaciens. J Bacteriol 139: 591-596.
Llosa M & de la Cruz F (2005) Bacterial conjugation: a potential tool for genomic
engineering. Res Microbiol 156: 1-6.
Llosa M, Bolland S & de la Cruz F (1991) Structural and functional analysis of the origin
of conjugal transfer of the broad-host-range IncW plasmid R388 and comparison with
the related IncN plasmid R46. Mol Gen Genet 226: 473-483.
Llosa M, Gomis-Ruth FX, Coll M & de la Cruz Fd F (2002) Bacterial conjugation: a two-
step mechanism for DNA transport. Mol Microbiol 45: 1-8.
Macartney DP, Williams DR, Stafford T & Thomas CM (1997) Divergence and
conservation of the partitioning and global regulation functions in the central control
region of the IncP plasmids RK2 and R751. Microbiology 143 ( Pt 7): 2167-2177.
Mangan S & Alon U (2003) Structure and function of the feed-forward loop network motif.
Proc Natl Acad Sci U S A 100: 11980-11985.
Mangan S, Itzkovitz S, Zaslaver A & Alon U (2006) The incoherent feed-forward loop
accelerates the response-time of the gal system of Escherichia coli. J Mol Biol 356: 1073-
1081.
Mazel D & Davies J (1999) Antibiotic resistance in microbes. Cell Mol Life Sci 56: 742-
754.
Monod J (1949) The growth of bacterial cultures. Ann. Rev. Microbiol. 3: 371–394.
Nagai T, Ibata K, Park ES, Kubota M, Mikoshiba K & Miyawaki A (2002) A variant of
yellow fluorescent protein with fast and efficient maturation for cell-biological
applications. Nat Biotechnol 20: 87-90.
Nedelcu AM, Miles IH, Fagir AM & Karol K (2008) Adaptive eukaryote-to-eukaryote
lateral gene transfer: stress-related genes of algal origin in the closest unicellular
relatives of animals. J Evol Biol 21: 1852-1860.
154
Norberg P, Bergstrom M & Hermansson M (2014) Complete nucleotide sequence and
analysis of two conjugative broad host range plasmids from a marine microbial biofilm.
PLoS One 9: e92321.
Normander B, Christensen BB, Molin S & Kroer N (1998) Effect of bacterial distribution
and activity on conjugal gene transfer on the phylloplane of the bush bean (Phaseolus
vulgaris). Appl Environ Microbiol 64: 1902-1909.
Okumura MS & Kado CI (1992) The region essential for efficient autonomous replication
of pSa in Escherichia coli. Mol Gen Genet 235: 55-63.
Paterson ES, More MI, Pillay G, et al. (1999) Genetic analysis of the mobilization and
leading regions of the IncN plasmids pKM101 and pCU1. J Bacteriol 181: 2572-2583.
Pattishall KH, Acar J, Burchall JJ, Goldstein FW & Harvey RJ (1977) Two distinct types
of trimethoprim-resistant dihydrofolate reductase specified by R-plasmids of different
compatibility groups. J Biol Chem 252: 2319-2323.
Porter J, Deere D, Hardman M, Edwards C & Pickup R (1997) Go with the flow – use of
flow cytometry in environmental microbiology. FEMS Microbiology Ecology 24: 93-101.
Rest JS & Mindell DP (2003) Retroids in archaea: phylogeny and lateral origins. Mol Biol
Evol 20: 1134-1142.
Rizzo MA, Springer GH, Granada B & Piston DW (2004) An improved cyan fluorescent
protein variant useful for FRET. Nat Biotechnol 22: 445-449.
Ronen M, Rosenberg R, Shraiman BI & Alon U (2002) Assigning numbers to the arrows:
parameterizing a gene regulation network by using accurate expression kinetics. Proc
Natl Acad Sci U S A 99: 10555-10560.
Santillan M, Mackey MC & Zeron ES (2007) Origin of bistability in the lac Operon.
Biophys J 92: 3830-3842.
Sen D, Brown CJ, Top EM & Sullivan J (2013) Inferring the evolutionary history of IncP-
1 plasmids despite incongruence among backbone gene trees. Mol Biol Evol 30: 154-
166.
155
Setty Y, Mayo AE, Surette MG & Alon U (2003) Detailed map of a cis-regulatory input
function. Proc Natl Acad Sci U S A 100: 7702-7707.
Shaner NC, Campbell RE, Steinbach PA, Giepmans BN, Palmer AE & Tsien RY (2004)
Improved monomeric red, orange and yellow fluorescent proteins derived from
Discosoma sp. red fluorescent protein. Nat Biotechnol 22: 1567-1572.
Shcherbo D, Souslova EA, Goedhart J, et al. (2009) Practical and reliable FRET/FLIM
pair of fluorescent proteins. BMC Biotechnol 9: 24.
Shen-Orr SS, Milo R, Mangan S & Alon U (2002) Network motifs in the transcriptional
regulation network of Escherichia coli. Nat Genet 31: 64-68.
Simonsen L, Gordon DM, Stewart FM & Levin BR (1990) Estimating the rate of plasmid
transfer: an end-point method. J Gen Microbiol 136: 2319-2325.
Smillie C, Garcillan-Barcia MP, Francia MV, Rocha EP & de la Cruz F (2010) Mobility of
plasmids. Microbiol Mol Biol Rev 74: 434-452.
Suzuki H, Yano H, Brown CJ & Top EM (2010) Predicting plasmid promiscuity based on
genomic signature. J Bacteriol 192: 6045-6055.
Tait RC, Kado CI & Rodriguez RL (1983) A comparison of the origin of replication of pSa
with R6K. Mol Gen Genet 192: 32-38.
Tettelin H, Riley D, Cattuto C & Medini D (2008) Comparative genomics: the bacterial
pan-genome. Curr Opin Microbiol 11: 472-477.
Tosini F, Venanzi S, Boschi A & Battaglia PA (1998) The uvp1 gene on the R46 plasmid
encodes a resolvase that catalyzes site-specific resolution involving the 5'-conserved
segment of the adjacent integron In1. Mol Gen Genet 258: 404-411.
156
Watanabe T, Furuse C & Sakaizumi S (1968) Transduction of various R factors by phage
P1 in Escherichia coli and by phage P22 in Salmonella typhimurium. J Bacteriol 96:
1791-1795.
Yano H, Rogers LM, Knox MG, Heuer H, Smalla K, Brown CJ & Top EM (2013) Host
range diversification within the IncP-1 plasmid group. Microbiology 159: 2303-2315.
Yuh CH, Bolouri H & Davidson EH (1998) Genomic cis-regulatory logic: experimental
and computational analysis of a sea urchin gene. Science 279: 1896-1902.
Zhong X, Droesch J, Fox R, Top EM & Krone SM (2012) On the meaning and estimation
of plasmid transfer rates for surface-associated and well-mixed bacterial populations. J
Theor Biol 294: 144-152.
Zoueva OP, Iyer VN, Matula TI & Kozlowski M (2003) Analysis of pCU1 replication
origins: dependence of oriS on the plasmid-encoded replication initiation protein RepA.
Plasmid 49: 152-159.
157
158
PUBLICACIONES
Problems and paradigms DOI 10.1002/bies.200900164
The gene regulation function (GRF) provides an opera- encodes proteins involved in catalytic or scaffolding activities;
tional description of a promoter behavior as a function of these were called structural genes. Genes of the second kind
the concentration of one of its transcriptional regulators. are involved in controlling in what amounts and under what
Behind this apparently trivial definition lies a central
concept in biological control: the GRF provides the circumstances structural proteins are produced; they were
input/output relationship of each edge in a transcrip- therefore called regulatory genes. Transcriptional regulation
tional network, independently from the molecular inter- was recognized at that time as a core process of cell
actions involved. Here we discuss how existing methods physiology,(2) and has been a central topic of biological
allow direct measurement of the GRF, and how several research ever since. Our perspective on how transcriptional
trade-offs between scalability and accuracy have hin-
dered its application to relatively large networks. We activity is regulated has changed substantially over the past
discuss the theoretical and technical requirements for 60 years.(3) Today there are two, apparently paradoxical,
obtaining the GRF. Based on these requirements, we research trends on transcriptional control. On the one hand,
introduce a simplified and easily scalable method that analysis of transcriptional regulation has become ‘‘global’’.
is able to capture the significant parameters of the GRF. Originally, transcriptional regulators were thought to control
The GRF is able to predict the behavior of a simple
genetic circuit, illustrating how addressing the quantita- the production of a certain set of proteins in response to a
tive nature of gene regulation substantially increases our specific stimulus.(4) Now we know that regulators are
comprehension on the mechanisms of gene control. interconnected in intricate transcriptional networks,(5–9) and
that the outcomes of these networks to changes in the
Keywords: gene regulation function; network parameteri- environment are often complicated to assess. On the other
zation; plasmid regulatory network; transcriptional regulator hand, transcriptional control has become ‘‘individual.’’ The
original description of the lacþ/ phenotype consisted of a
Abbreviations: a, growth rate; b, production rate for graph showing how a bacterial population smoothly switched
repressor-free promoter; b0, production rate for repressor- its lac phenotype in response to a diauxic shift.(10) This
bound promoter; 3(-UTR, 3( untranslated region; AFU, behavior is still referred in textbooks as the canonical example
arbitrary fluorescent units; ara, arabinose; aTc, anhydrotetra- of deterministic control: a culture of E. coli grown with glucose
cycline; CFP, cyan fluorescent protein; GFP, green fluorescent and lactose as carbon sources will always exhibit diauxie.
protein; GRF, gene regulation function; K, dissociation However, at the single cell level, a culture of genetically
constant; OD, optical density; RBS, ribosome binding site; X, identical cells grown in the same culture medium often exhibits a
concentration of transcriptional repressor; Xss, concentration mixture of lacþ and lac phenotypes, and the transition between
of transcriptional repressor at the steady state; Y, product of them is purely stochastic.(11) Transcriptional regulation is
the target promoter; YFP, yellow fluorescent protein. therefore a probabilistic process, and this fact has many
implications for cell adaptation and differentiation.(12–14)
Introduction Although this situation might seem contradictory, actually it
is not: regulatory networks are holistic, but cells within a
In their seminal work on adaptation, Jacob and Monod population show strong individualism, based on the probabil-
proposed that there are two kinds of genes.(1) The first kind istic nature of control processes. There is a common
dimension in both aspects of transcriptional regulation. The
*Correspondence to: Dr. F. de la Cruz, C. Herrera Oria s/n, Santander,
Cantabria, 39011 Spain final outcome of a regulatory network or stochastic switch
E-mail: [email protected] within the cell depends not only on which factors are produced
but also on the amounts of that factor. To understand how cells large networks it is usually complicated to measure every
self-regulate, we need to describe quantitatively the nature of relevant biochemical parameter, like dissociation constants or
cell responses.(15) While it remains technically challenging to cooperativity indexes. Moreover, it is even more complicated
assess how many mRNAs and proteins a cell produces as a to establish how parameters measured in vitro reflect the
response to a certain input,(16,17) it is feasible to determine in vivo behavior.(20) A systematic method for measuring the
how much a promoter activity changes as a consequence of a GRF simplifies the task of assessing the relevant parameters
change in the concentration of its regulators. of a given network, and also provides an in vivo framework to
The gene regulation function (GRF) indicates the relation- interpret in vitro results. One of the ultimate goals of synthetic
ship between the intracellular concentration of a given biology is to generate a repertoire of standard biological parts
regulator and the transcriptional activity of its target that can be used in the design of higher order devices.(21) One
promoter.(18) Ideally, it is a continuous bijective function that important set of such parts should be a series of well-
renders a certain promoter activity for any given regulator characterized transcriptionally regulated promoters.(22) With-
concentration. This function is characteristic for a regulator/ out a precise description of the response dynamics of each
promoter pair and each promoter has as many GRFs as the component, the integration of complex devices becomes a
number of transcriptional regulators acting on it. The GRF is trial-and-error instead of a design-based process.
an input/output relationship that depends on a number of
intermediate biochemical events (regulator oligomerization,
binding site searching kinetics, DNA binding, interactions with Precise measurement of the GRF:
RNApol, etc. . .). Therefore, it can be considered a metafunc- The l-cascade method
tion: a mathematical abstraction that has no real process
counterpart, but represents the convolution of many. This is in The gold standard for measuring the GRF is the method
fact its main value, since it provides a complete description of developed by Rosenfeld et al.(18) In this work the authors
the regulator/promoter response space regardless of the measured the GRF for the PR promoter of bacteriophage
underlying mechanism. Disentangling how each intervening lambda in response to variations in the concentration of
molecular process contributes to the final observable GRF is protein cI, its transcriptional repressor. Regulator concentra-
a quintessential problem of transcriptional regulation,(19) and tion (input) was monitored by translational fusion of gene cI
has attracted extensive research efforts. A detailed analysis to the YFP (yellow fluorescent protein) gene. The transcriptional
of the molecular mechanisms of transcriptional regulation is activity of PR (output) was determined using a transcriptional
beyond the scope of the present work. We do not discuss how fusion with the CFP (cyan fluorescent protein) gene (Fig. 1).
to predict the GRF from our knowledge of each individual The regulator concentration was modulated by making its
molecular step, but instead on how to measure it experimen- expression dependent upon an upstream Tet promoter in a
tally. Specifically, we describe how existing methods for tetRþ background strain. Cells were grown in the presence of
accurate expression profiling and controlled protein expres- anhydrotetracycline to achieve full induction of repressor cI.
sion allow a direct inference of the GRF. Then the inducer was washed away, so that Tet promoter
The GRF is a powerful tool for systems and synthetic activity decreased over time to a final off state. Consequently,
biology. The behavior of a given network depends not only on the levels of cI-YFP were shown to decrease analogically due
its topology but also on its parameterization. For relatively to protein degradation and dilution in growing cells. The
Figure 1. The l-cascade method developed by Rosenfeld and Alon. A: Ideal representation of the GRF. Repressor concentration is decreased
analogically (x-axis), and the GFP production rate (y-axis) is measured as an indicator of target promoter transcriptional activity. B: Scheme of the
l-cascade regulatory circuit. The transcriptional repressor cI is tagged by translational fusion with YFP and expressed from the Ptet promoter.
TetR (constitutively produced) blocks Ptet expression. Repression can be lifted by the action of aTc. The target promoter (PR) transcriptional
activity is tracked by following the CFP fluorescence. C: Plot showing cI-YFP (red line, regulator) and CFP (green line, target promoter) dynamics
in a single cell. The inducer aTc is kept at high concentration for 2 h (gray area) and then washed away (white area). Concentration of cI starts to decrease
as judged by the decrease in YFP fluorescence. Simultaneously, CFP fluorescence increases. Figure modified from Rosenfeld et al. (18)
decrease in cI-YFP levels and the subsequent variation of PR- precisely measure promoter activities over time and to control
driven CFP expression were tracked over time for each the induction levels of a given regulator. In this work we
individual cell by real-time fluorescence microscopy, thus illustrate how existing methods can lead to a simplified
obtaining the GRF for the PR/cI promoter/regulator pair. The procedure to measure the GRF. We will show that important
biological meaning of the GRF and its predictive power were trade-offs exist between the exactness of the measurement
demonstrated later,(23) when the behavior of a simple circuit and the scalability of the procedure. In this respect, the
was predicted upon the measurement of the component’s method that we propose here must be considered optimized
GRF. The main advantage of the l-cascade method was its for scalability, as the l-cascade is optimized for accuracy. In
ability to make single-cell measurements in real time. This terms of scalability, the l-cascade paradigm was complemen-
allowed an accurate evaluation of the contribution of ted by efforts to determine quantitative kinetic parameters using
stochasticity in the dynamics of the system and prevented bulk culture measurements in model regulatory networks.(24)
underestimation of the cooperativity index, a problem that We propose a variation of the Rosenfeld method, based on
bulk population measurements might cause.(18) As pointed a two-plasmid system (one for expressing the regulator and
out by the authors, the only caveat was that its applicability to one for measuring the target promoter response) (Fig. 2). The
other regulatory systems might be complicated by the system consists of an arabinose inducible expression vector
refinement of the required measurements.(23) Nevertheless, (the input plasmid) and a GFP reporter plasmid to precisely
it served as a powerful proof of concept that it is possible to obtain determine the response in the target promoter (the output plasmid).
quantitative models of biological systems with predictive power. This method presents two main differences with respect to the
original parameterization of the GRF in the l-cascade system:
A scalable method for obtaining the GRF: (i) Single-cell measurements are replaced by bulk population
The two-plasmid method averages: This allows the use of simple equipment such
as fluorescence plate readers and flow cytometers as
The main obstacle for the applicability of the l-cascade measuring instruments, instead of time-lapse microscopy
method is that it requires both extensive genetic manipulation facilities, which are of limited availability. On the other
and technically demanding measurements. Regulators need hand, it decreases the precision of the measurement,
to be translationally fused with fluorescent reporters, which since the population GRF might show differences to
often results in non functional proteins. Quantifying each single-cell GRF due to underestimation of cooperativity
interaction, even in small networks with a discrete number of indexes (see Trade-offs and limitations in experimental
arrows, would be challenging. Nevertheless, the l-cascade measurements of the GRF below).
method demonstrated that the fundamental bases for (ii) Regulator concentrations are estimated, not directly mea-
resolving the GRF are apparently simple, i.e., the ability to sured: the major handicap for extending the l-cascade
Figure 2. Scheme of the two-plasmid measuring system. The input plasmid consists of an expression vector (pBAD33) in which the desired
regulator is cloned under the control of ParaBAD promoter (and thus is arabinose inducible). The output plasmid is a pSC101 derivate (pUA66)
engineered for analyzing the activity of a given promoter. The target promoter is cloned so that it drives expression of the GFPmut2 gene.
method to the analysis of other regulatory systems is the protocol described in Refs.(24–26) Promoters are cloned in the
need for a fluorescent protein/transcriptional regulator pUA66 reporter vector, which contains the GFP-mut2 gene
translational fusion. Translational fusions often result in downstream of a strong ribosomal binding site. Fluorescence
non-functional proteins. To overcome this limitation we production and cell growth (OD600) are simultaneously
avoided the direct measurement of the regulator concen- tracked so it is possible to determine the GFP/OD change
tration. For this purpose, the regulator gene was cloned rate. Since GFP-mut2 is a stable protein, the death rate can
under the inducible promoter of the arabinose operon be assumed to be caused exclusively by dilution in the
PARABAD. Expression from this promoter is inducible, growing cells. Under these conditions, promoter activity
so increasing concentrations of arabinose in the culture becomes equivalent to the maximum of the time derivative of
medium produce increasing levels of PARABAD expres- GFP levels per OD unit [(dGFP/dt)/OD].(24,25,27) To ensure
sion. Regulators were cloned so they always shared a that promoter activities were measured during steady-state
strong consensus ribosomal binding site and a common conditions, several modifications were applied to the original
30 -UTR. This was intended to minimize differential trans- procedure (Box 1). Cells were grown for longer times, starting
lational efficiencies when different regulators were cloned. from low OD levels (0.01). This ensures that the culture
Since it is still possible that post-translational modifica- undergoes a number of replication events that dilute the GFP
tions and protein degradation rates have an important levels originating from previous growth. Under our conditions,
effect on regulator concentrations, measurements were GFP/OD levels reach a plateau in which d(GFP/OD)/dt ¼ 0,
made during steady state. Under steady-state conditions indicative of the steady-state regime (Fig. 3A). GFP levels at
an invariant average concentration of regulator can be that point equal, by definition, the production rate divided by
assumed. Steady-state levels depend on production and
degradation rates. These rates are assumed to be inde-
pendent from each other, at least under a range of non-
extreme conditions (e.g., extremely high protein produc-
tion could saturate the protein degradation machinery).
Although degradation rates are unknown, they can be
considered constant for a given regulator on a given
genetic background. Variation in the steady-state concen-
tration for two different ara induction levels therefore
depend only on differences in production rates. Since it
is possible to measure the response of PARABAD to
increasing ara concentrations, changes in the regulator
concentration can be inferred. Thus, the two-plasmid
method measures relative changes: the regulator concen-
tration is changed in a known proportion and the relative
response of the target promoter is then measured.
the degradation/dilution rate. Since the protein is stable and Measuring the GRF II: Inducing the
the dilution rate can be obtained from the OD600 growth regulator
curves, promoter activities can thus be directly estimated:
The second technical challenge was to induce expression of
d GFP
OD GFP
¼ba (1.1) the transcriptional regulator in a controlled and measurable
dt OD
way. This was achieved by cloning the regulator under the
GFP control of an inducible promoter. However, not every inducible
d OD GFP
¼ba ¼0 (1.2) promoter is suitable for this purpose. Population measure-
dt SS OD SS
ments of dose/response curves for inducible promoters yield
GFP
b¼a an apparently continuous curve, indicative of an analogical
OD SS
increase in the output as the inducer concentration increases.
where GFP are the fluorescence levels, b the production rate However, in many cases these curves do not reflect the single-
per OD unit, and a the growth rate. This method yields promoter cell dynamics, and the apparent analogical behavior is the
activities equivalent to those described in Refs.(24,25,27)(Fig. 3B). product of the averaging of different proportions of cells being
The advantage is that, since the steady-state phase can be moni- either in on or in off state.(28) This is a relevant issue for
tored precisely, many data points can be averaged to estimate promoters that respond to a certain chemical modulator (an
steady-state levels more precisely. The reproducibility of these inducer or a corepressor). On many occasions the intracel-
measurements in different experiments was better than SD < 15%. lular concentration of the modulator is dependent on active
transport and the transporter is often activated by the
transported metabolite. Transporter activation creates a
Box 1:
positive feedback loop that drives system bi-stability.(29) To
Experimental protocol for expression profiling
overcome this issue, we took advantage of the engineered
Step 1: Cell growth PARABAD system developed by Keasling and coworkers,
Reporter strains were inoculated in M9 Medium plus which uses an E. coli strain (BW27783) that constitutively
kanamycin (25 mg/mL), casaminoacids (0.2%) and gly- expresses araE, the arabinose transporter.(30) The PARABAD
cerol (0.5%). Cells were grown for 16 h, at 37 8C with promoter can be induced by adding different concentrations of
vigorous aeration. arabinose to the growing medium. Since the transporter is no
longer under the control of the transported metabolite, the
Step 2: Measurement positive feedback is broken. The quasi-analogical dynamics of
Cultures were diluted 1:10 000 in the same medium in this system was confirmed by measuring the expression
96-well plates and incubated in a Victor3 fluorimeter at levels of E. coli BW27783 bearing a PARABAD::GFP
37 8C with orbital shaking for about 6 h. Fluorescence and transcriptional fusion by flow cytometry (Fig. 4A). As shown
absorbance were determined for each well every 5 min. To in Fig. 4, the whole population displaces along the X-axis as
counterbalance evaporation, water was injected after each the ara concentration is increased. The population response
three steps of fluorescence and absorbance measure- to increasing ara concentrations was measured by expression
ment. Water evaporated in the Victor3 multiplate reader at profiling as indicated in the previous section. Results are
37 8C was 0.28 mL/min/well. shown in Fig. 3B. PARABAD::GFP fluorescence values were
plotted against the corresponding ara concentrations (Fig.
Step 3: Data processing 4B). Data were fitted to a Michaelis–Menten function
Absorbance data points had background (data obtained (R2 ¼ 0.99), thus obtaining an expression that infers the
from culture medium with no cells) subtracted. Absor- PARABAD expression level as a function of the arabinose
bance values were transformed into OD600 equivalents concentration. When a transcriptional regulator is cloned in
using a calibrated curve between Victor 3 readouts and a the pBAD33 vector under expression of PARABAD, relative
regular spectrophotometer (width 1 cm). Fluorescence changes in steady-state concentrations can be inferred
background was obtained from cells containing the according to the standard curve obtained for the GFP.
promoter-less plasmid pUA66. Fluorescence background
was subtracted from the values obtained for the reporter
strains at the same OD600 (not necessarily at the same
time points). Growth rate (a) was calculated from Mathematical framework for determining
the OD600 data. Fluorescence/OD600 was plotted the GRF
against OD600 and the steady-state level was obtained
by averaging values during the steady-state. Since the GRF is a metafunction that does not represent a
single process, different levels of abstraction can be applied to
dY K X
¼b þ b0 aY (2.1)
dt K þX
|fflfflffl{zfflffl
ffl} K þX
|fflfflffl{zfflffl
ffl}
X free X bound
promoter promoter
fraction fraction
Figure 4. Response of the input plasmid to arabinose induction. where b and b0 correspond respectively to the activity rates for
A: Fluorescence distributions (x-axis indicates fluorescence, y-axis
repressor-free and repressor-bound promoters, K stands for
indicates cell number) obtained by flow cytometry analysis of a culture
of E. coli BW27783 containing pBAD33::GFPmut2. Each color indi- the dissociation constant and a indicates the growth rate of
cates a different arabinose concentration. B: Median fluorescence the culture. In steady state (dY/dt ¼ 0):
values obtained from the distributions shown above plotted against
the Ara concentration used (in % w/v). Inner chart: arabinose con-
centrations indicated in log scale (% w/v). bK þ b0 XSS
YSS ¼ (2.2)
aðK þ XSS Þ
We define R as the regulatory rank of a given promoter Y general constant K to the power of n (Kn). Therefore, the
for a given repressor X: generalized expression follows:
This parameter is promoter/repressor dependent, and It should be noticed that a value of n > 1 in the GRF does
indicates the expression space that a certain repressor can not imply that the number of binding events is precisely n. It
produce when acting upon a target promoter. Therefore, a only indicates, as in the Hill function, the apparent
given promoter with two repressors will show different R cooperativity grade in the final observable dynamics.
values depending on the ability of each repressor to interfere
with transcription initiation.
Introducing Ymax ¼ b/a and Ymin ¼ b0/a into Eq. (2.2) we
obtain
A case study: Determining the GRF for
the transcriptional repressor KorA
Ymax K þ Ymin XSS
YSS ¼ (2.6) To demonstrate the validity of the two-plasmid method, we
K þ XSS
obtained the GRF for KorA, a transcriptional repressor from
the broad host range plasmid R388.(31) KorA is the major
This expression indicates how the target promoter Y
transcriptional modulator for the conjugative mating system, a
responds to changes in the repressor X, and therefore can be
multiprotein secretion system that mediates DNA transfer
considered directly proportional to the GRF. However, it is
during plasmid conjugation.(32) KorA represses four promo-
probably easier to interpret the relative change of Y from its
ters in the plasmid, binding to a conserved sequence called
Ymax levels for each X concentration; this expression also
the KorA box. KorA controls its own expression, therefore
renders a dimensionless value. The transformation follows:
establishing a negative feedback loop.(32) We analyzed the
GRF of KorA on its own promoter (PkorA) using the procedures
YSS K þ Ymin=Ymax XSS
¼ (2.7) described above. Briefly, korA gene was PCR-amplified and
Ymax K þ XSS
cloned in expression vector pBAD33, under the control of
PARABAD promoter. To monitor PkorA expression levels, the
Ymax K þ XSS
¼ (2.8) promoter was PCR-amplified and cloned in the reporter
YSS K þ XSS R X vector pUA66. Maximal PkorA promoter activity (Ymax) was
Y
measured by expression profiling of E. coli BW27783
containing pUA66::PkorA, thus measuring the promoter
This expression is fairly intuitive: given an X concentration
activity in absence of its cognate regulator. Minimal PkorA
of repressor, the target promoter is Ymax/Y times repressed.
promoter activity (Ymin) was measured by introducing
Parameters Ymax and Ymin (and therefore R) can be
pBAD33::KorA and inducing KorA expression at saturating
experimentally measured. For measuring Ymax the promoter
concentrations of arabinose. Saturating concentrations were
activity is determined in a regulator-defective genetic back-
attained at 3.5E-5% w/v. Higher ara concentrations impaired
ground. Ymin can in turn be measured by inducing the
cell growth (data not shown). PkorA expression levels (Y) were
regulator expression to maximal levels [so X/(X þ K) ! 1]. For
measured as indicated in Box 1 under different concentrations
this model to work, Ymin must be attained within the
of arabinose. KorA levels were estimated by running in parallel
experimental rank of concentrations of X. Although the former
a set of E. coli BW27783 cultures containing pBAD33::GFP-
expression was derived for the case of a transcriptional
mut2 subjected to the same ara concentrations. Ymax/Y levels
repressor, the derivation for a transcriptional activator is trivial
were plotted against estimated concentrations of KorA (X)
taking into consideration that:
expressed in GFP fluorescence equivalents (Fig. 6A), and
bK
Ymin ðx Þ ¼ Yðx¼0Þ ¼ aK ¼ b=a data were adjusted to Eq. (2.9) using the Levenberg–
b
Ymax ðx Þ ¼ Yðx¼1Þ ¼ 0=a Marquardt algorithm for non-linear regression. Regression
yielded K ¼ 22 10 arbitrary fluorescent units (AFU). The
The former expressions assume a unique binding event of regulatory rank was R ¼ 33, indicating that KorA was able to
the regulator to the promoter. For the more general case in repress PkorA 33-fold from its native promoter activity. This
which n regulator molecules bind the promoter, we can use value was in accordance to the experimental determination of
the Hill approximation to say that the product of n different R (Ymax/Ymin). The cooperativity index was n ¼ 1.3 0.4,
dissociation constants (K1K2K3. . .Kn) can be made equal to a indicating a slight cooperative effect of KorA in PkorA
Figure 6. Measurement and performance of the GRF for KorA regulator acting on PkorA promoter. A: Repression ratio (Ymax/Y) of PkorA
promoter (y-axis) as a function of apparent KorA repressor concentration (x-axis). KorA concentrations is modified by inducing pBAD33::korA
with different levels of arabinose, and apparent KorA concentrations are calculated from calibrating curves (pBAD33::GFP) obtained in parallel.
Maximal steady-state fluorescence level achieved by PkorA promoter (Ymax, [KorA] ¼ 0) is divided by steady-state fluorescence levels achieved by
PkorA (Y ), and plotted in ordinates. Error bars indicate the standard deviation of eight independent experiments. Data were fitted to Eq. (2.9) with
R2 ¼ 0.98. B: A comparison between the predicted behavior of PkorA-KorA feedback loop (black crosses) with an experimental measurement of
PkorA-KorA dynamics (red dots). Experimental measurements were done in E. coli BW27783 cultures containing plasmid R388 (that contributes
the PkorA-KorA feedback loop) and pUA66: PkorA-GFP (reporter plasmid that infers PkorA expression levels). Predicted dynamics were obtained
from Eq. (2.1) using GFP/OD measured at OD600 ¼ 0.2 as initial X value. All parameters in the simulation were determined experimentally [K and
n obtained from data from (A), growth rate (a) obtained from growth curve and production rates b and b0 obtained from expression profiles without
KorA and at maximal KorA concentration, respectively].
transcriptional activity. The R2 value of the adjusted curve was Trade-offs and limitations in experimental
R2 ¼ 0.980. measurements of the GRF
To check the predictive power of the KorA/PkorA GRF
we simulated the time behavior of the negative feedback As we have shown, the two-plasmid method is able to extract
loop. We measured the behavior of this feedback loop sufficient information to determine the GRF using simple
experimentally by introducing plasmid R388 (that carries KorA measurement techniques and a basic mathematical treat-
under PkorA expression) into an E. coli BW27783 strain ment. It is based on two previously reported experimental
containing also pUA66::PkorA and profiling the expression procedures(25,30) and might be useful for systematic para-
levels along time. We fed the computational simulation with meterization of relatively large networks. However, as stated
the initial fluorescence value experimentally obtained above, several caveats should be taken into consideration
at OD600 ¼ 0.2. We then compared how the theoretical when applying this method. The fundamental one refers to
prediction and the experimental measurement behave as the bulk averaging in contrast with single-cell measurements. Cell
culture grew (Fig. 6B). The OD600 start point was chosen heterogeneity is a possible source of conflict: population
because fluorescence values for OD600 < 0.2 were found to measurements tend to underestimate cooperativity indexes
be extremely noisy. As shown in the figure, the computational due to convolution of the regulator concentration distribution
prediction based on the measured GRF follows the dynamics and the promoter response.(33) The two-plasmid method
of the experimental data until the cell culture enters stationary alleviates this problem using an induction system that exhibits
phase (OD600 ¼ 1). At this point both curves separate sharply, a unimodal distribution (Fig. 4A). However, since both
indicating that the GRF performs well as a dynamical regulator concentration and promoter activity distributions
predictor for cells in constant growth but not in stationary are averaged, the method will still underestimate the
cultures. This was to be expected since one of the steepness of the curve. To overcome this problem, a
assumptions of our model was that cells were in exponential simultaneous measurement at the single-cell level of both
growth. regulator concentration and promoter activity must be carried
out. Flow cytometry can determine both distributions if the display mutual interactions, so the function of regulators A and
transcriptional regulator and target promoter are tagged with B is not equal to the sum of their independent GRFs. Since the
fluorescent proteins with enough spectral separation and GRF is a function, and function orthogonality is strictly
similar maturation timescales. However, due to the small size defined, it can be used to test for this principle. It has been
and bacillary shape of E. coli, this technique is unsuitable for demonstrated that the multiple input functions for sugar
assessing bacterial size. As a consequence flow cytometry metabolism in E. coli can sometimes be separated,(36) so it
retrieves a convoluted distribution of ill-defined cell sizes and seems that, at least in some cases, biological control systems
fluorescence values. If a higher level of accuracy is desired, are orthogonal. The circumstances and the kinds of promoter
measurements can be made by fluorescence microscopy, architectures with which orthogonality in the GRF is allowed
which would render an equivalent procedure to the l-cascade still need to be elucidated.
method. Scalability and precision in the measurement are Parameterizing global regulatory networks is a formidable
therefore inversely related. According to the size of the task, and massive parallel efforts are needed. In that sense,
network and the level of refinement needed, the researcher the development of a standard unit system for promoter
will have to decide which method to use. In that respect, activities is urgently needed so that results obtained by
further research is needed to assess the comparative different methods and in different laboratories can be
performances of both methods. compared. Efforts are being made to obtain absolute scales
Common to any GRF determination method is the need for based on the effective counting of mRNAs and proteins,(17)
a regulator-defective genetic background. In the l-cascade, and to establish a set of reference promoters.(37) Developing
the authors used a phage protein so the host bacteria were standard scales would be useful for Systems Biology and is
naturally cI .(18) Similarly, we have used a plasmid- probably essential for the development of standardized parts
borne regulator, so the host strain is naturally KorA defective. in synthetic biology. It is difficult to envisage how a new
If one wants to determine the GRF for a chromosomally engineering discipline can be founded without an operational
encoded regulator, a knockout background is always needed, metric system.
since the full possible range of regulator concentrations Another crucial question is to determine what forces shape
must be sampled. If the two-plasmid method described the GRF. Cost-benefit theory for the GRF identifies three
here were to be used, this mutation must be introduced distinguishable levels of selection: efficiency, economy, and
into the E. coli BW27783 strain, since this genetic noise adaptation.(38) Efficiency indicates the ability of the GRF
background is needed for appropriate induction of to sense and respond in an appropriate timescale to changes
the ParaBAD promoter.(30) For that purpose, extensive in the input. Economy refers to the ability to produce
knockout libraries that allow P1 transduction, like the Keio responses only when the benefits for the bacterial fitness
collection,(34) can be used. exceed the cost of producing the response. Noise adaptation
indicates the capacity of the GRF to suppress or enhance cell-
to-cell variation to achieve efficient and economic responses.
Conclusions and prospects Interestingly, there are severe trade-offs between these three
levels.(38) As an example, a strict GRF with a high
Despite its abstract nature, the GRF has proven predictive cooperativity index would be efficient (the cell would respond
power.(23) Although we have focused on methods used according to a certain threshold) and economic (no
for estimating the GRF in prokaryotes, similar approaches suboptimal responses under the threshold), but it would be
have been successfully employed in eukaryotes, where extremely vulnerable to input noise, since noise transmission
chromatin remodeling was found to be an essential increases linearly with cooperativity indexes.(38) Network
player determining the GRF.(35) These results illustrate the topology comes to play at this point, since network motifs such
possibility of calculating the output of a certain regulatory as negative feedback and feed-forward loops can decrease
network if we know the transfer functions of its individual noise.(38–40) In other cases, in which input fluctuations are fast
components. or undetectable (so the cell has to rely on ‘‘blind’’ decisions),
Besides being a powerful tool for network analysis, the noise in gene expression can not be deleterious but
GRF opens the possibility to analyze a burning question in beneficial.(41) Selective landscapes will be the ultimate force
synthetic biology: How do promoters integrate the action of shaping the GRF,(38) a function in which topological proper-
two or more regulators? It might happen that the signals ties, noise in gene expression, and response time require-
deployed from two different transcription factors to the same ments concur. Therefore, the GRF exceeds any utilitarian view
target promoter act independently from each other, so the final for design purposes: it is the keystone of gene regulation. It
function can be calculated as the sum of the individual establishes how inputs and outputs are computed by the cell,
contributions. This is usually referred to as orthogonality and how these responses are modified and evolve by
between two inputs. It is also possible that both regulators environmental constraints.
Acknowledgments: Work in the FdlC laboratory was sup- 19. Segal E, Widom J. 2009. From DNA sequence to transcriptional beha-
ported by grants BFU2008-00995/BMC (Spanish Ministry of viour: a quantitative approach. Nat Rev Genet 10: 443–56.
20. Cases I, de Lorenzo V. 2005. Promoters in the environment: transcrip-
Education), RD06/0008/1012 (RETICS research network, tional regulation in its natural context. Nat Rev Microbiol 3: 105–18.
Instituto de Salud Carlos III, Spanish Ministry of Health) 21. Endy D. 2005. Foundations for engineering biology. Nature 438: 449–53.
and LSHM-CT-2005_019023 (European VI Framework Pro- 22. Hasty J, McMillen D, Collins JJ. 2002. Engineered gene circuits. Nature
420: 224–30.
gram). 23. Rosenfeld N, Young JW, Alon U, et al. 2007. Accurate prediction of gene
feedback circuit behavior from component properties. Mol Syst Biol 3:
143.
References 24. Ronen M, Rosenberg R, Shraiman BI, et al. 2002. Assigning numbers to
the arrows: parameterizing a gene regulation network by using accurate
expression kinetics. Proc Natl Acad Sci U S A 99: 10555–60.
1. Jacob F, Monod J. 1959. Genes of structure and genes of regulation in 25. Zaslaver A, Bren A, Ronen M, et al. 2006. A comprehensive library of
the biosynthesis of proteins. C R Hebd Seances Acad Sci 249: 1282–4. fluorescent transcriptional reporters for Escherichia coli. Nat Methods 3:
2. Umbarger HE. 1964. Intracellular regulatory mechanisms: regulation in 623–8.
multicellular forms may be an elaboration upon the pattern evolved in 26. Kalir S, McClure J, Pabbaraju K, et al. 2001. Ordering genes in a flagella
microorganisms. Science 145: 674–9. pathway by analysis of expression kinetics from living bacteria. Science
3. Balleza E, Lopez-Bojorquez LN, Martinez-Antonio A, et al. 2009. 292: 2080–3.
Regulation by transcription factors in bacteria: beyond description. FEMS 27. Setty Y, Mayo AE, Surette MG, et al. 2003. Detailed map of a cis-
Microbiol Rev 33: 133–51. regulatory input function. Proc Natl Acad Sci U S A 100: 7702–7.
4. Stent GS. 1964. The Operon: on its third anniversary. Modulation of 28. Ozbudak EM, Thattai M, Lim HN, et al. 2004. Multistability in the lactose
transfer RNA species can provide a workable model of an operator-less utilization network of Escherichia coli. Nature 427: 737–40.
operon. Science 144: 816–20. 29. Santillan M, Mackey MC, Zeron ES. 2007. Origin of bistability in the lac
5. Shen-Orr SS, Milo R, Mangan S, et al. 2002. Network motifs in the Operon. Biophys J 92: 3830–42.
transcriptional regulation network of Escherichia coli. Nat Genet 31: 64–8. 30. Khlebnikov A, Datsenko KA, Skaug T, et al. 2001. Homogeneous
6. Ihmels J, Friedlander G, Bergmann S, et al. 2002. Revealing modular expression of the P(BAD) promoter in Escherichia coli by constitutive
organization in the yeast transcriptional network. Nat Genet 31: 370–7. expression of the low-affinity high-capacity AraE transporter. Microbiology
7. Kashtan N, Itzkovitz S, Milo R, et al. 2004. Topological generalizations of 147: 3241–7.
network motifs. Phys Rev E Stat Nonlin Soft Matter Phys 70: 031909. 31. Fernandez-Lopez R, Garcillan-Barcia MP, Revilla C, et al. 2006.
8. Madan Babu M, Teichmann SA. 2003. Evolution of transcription factors Dynamics of the IncW genetic backbone imply general trends in con-
and the gene regulatory network in Escherichia coli. Nucleic Acids Res 31: jugative plasmid evolution. FEMS Microbiol Rev 30: 942–66.
1234–44. 32. Seubert A, Hiestand R, de la Cruz F, et al. 2003. A bacterial conjugation
9. Gutierrez-Rios RM, Rosenblueth DA, Loza JA, et al. 2003. Regulatory machinery recruited for pathogenesis. Mol Microbiol 49: 1253–66.
network of Escherichia coli: consistency between literature knowledge 33. Cluzel P, Surette M, Leibler S. 2000. An ultrasensitive bacterial motor
and microarray profiles. Genome Res 13: 2435–43. revealed by monitoring signaling proteins in single cells. Science 287:
10. Jacob F, Monod J. 1961. Genetic regulatory mechanisms in the synthesis 1652–5.
of proteins. J Mol Biol 3: 318–56. 34. Baba T, Ara T, Hasegawa M, et al. 2006. Construction of Escherichia coli
11. Choi PJ, Cai L, Frieda K, et al. 2008. A stochastic single-molecule event K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst
triggers phenotype switching of a bacterial cell. Science 322: 442–6. Biol 2: 2006 0008.
12. Paulsson J. 2004. Summing up the noise in gene networks. Nature 427: 35. Kim HD, O’Shea EK. 2008. A quantitative model of transcription factor-
415–8. activated gene expression. Nat Struct Mol Biol 15: 1192–8.
13. Elowitz MB, Levine AJ, Siggia ED, et al. 2002. Stochastic gene expres- 36. Kaplan S, Bren A, Zaslaver A, et al. 2008. Diverse two-dimensional input
sion in a single cell. Science 297: 1183–6. functions control bacterial sugar genes. Mol Cell 29: 786–92.
14. Suel GM, Garcia-Ojalvo J, Liberman LM, et al. 2006. An excitable gene 37. Kelly JR, Rubin AJ, Davis JH, et al. 2009. Measuring the activity of
regulatory circuit induces transient cellular differentiation. Nature 440: BioBrick promoters using an in vivo reference standard. J Biol Eng 3: 4.
545–50. 38. Kalisky T, Dekel E, Alon U. 2007. Cost-benefit theory and optimal design
15. Kim HD, Shay T, O’Shea EK, et al. 2009. Transcriptional regulatory of gene regulation functions. Phys Biol 4: 229–45.
circuits: predicting numbers from alphabets. Science 325: 429–32. 39. Becskei A, Serrano L. 2000. Engineering stability in gene networks by
16. Raj A, van Oudenaarden A. 2009. Single-molecule approaches to autoregulation. Nature 405: 590–3.
stochastic gene expression. Annu Rev Biophys 38: 255–70. 40. Paulsson J, Ehrenberg M. 2001. Noise in a minimal regulatory network:
17. Golding I, Paulsson J, Zawilski SM, et al. 2005. Real-time kinetics of plasmid copy number control. Q Rev Biophys 34: 1–59.
gene activity in individual bacteria. Cell 123: 1025–36. 41. Wolf DM, Vazirani VV, Arkin AP. 2005. Diversity in times of adversity:
18. Rosenfeld N, Young JW, Alon U, et al. 2005. Gene regulation at the probabilistic strategies in microbial survival games. J Theor Biol 234: 227–
single-cell level. Science 307: 1962–5. 53.
Plasmid
journal homepage: www.elsevier.com/locate/yplas
a r t i c l e i n f o a b s t r a c t
Article history: A cytometric method for the estimation of end-point conjugation rates is developed and
Received 19 December 2011 adapted to surface conjugation. This method improves the through-put of conjugation
Accepted 5 January 2012 assays based on replica-plating and results in less noisy experimental data. Although con-
Available online 24 January 2012
jugation on solid surfaces deviates from ideal conditions in which cells are continuously
Communicated by Dr. Dhruba K. Chattoraj
mixed, results show that, within the limits of high initial population densities and short
mating times, end-point estimates of the conjugation rates are robust measurements. They
Keywords:
are independent of the donor/recipient ratios and, to some extent, of the sampling time.
Bacterial conjugation
Surface mating
Remixing the mating population in the course of a conjugation experiment results in a
Flow-cytometer boost in the frequency of transconjugants.
GFP Ó 2012 Elsevier Inc. All rights reserved.
Conjugation rate
0147-619X/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.plasmid.2012.01.008
I. del Campo et al. / Plasmid 67 (2012) 174–182 175
Table 1
Plasmids.
where cep is the transfer rate (mL cell1 h1), wmax is the bac- hsdR514 DE(araBAD)567 DE(rhaBAD)568 DE(araFGH)
terial growth rate (h1), and D, R, T and N are the densities u(DaraEp PCP8-araE) (Khlebnikov et al., 2001). Two
(cells mL1) of donors, recipients, transconjugants and total BW27783 spontaneous mutants, one resistant to nalidixic
culture cells, respectively, at any given time. Although the acid (BW-NxR) and another one resistant to rifampicin
end point estimate was originally developed for liquid cul- (BW-RfR) were obtained for this work by plating out
tures, due to its independence from changes in some exper- BW27783 in LB-agar containing the relevant antibiotic.
imental conditions and to its simple implementation, the Spontaneous mutants arose at a frequency of roughly
method has been used to measure transfer rates even in 108. The mutants were not further characterized. Bacteria
experimental set-ups involving surface-associated conjuga- were grown in Luria–Bertani broth (LB) or LB-agar at 37 °C
tion (Licht et al., 1999; Lilley and Bailey, 2002; Normander supplemented with the appropriate antibiotics. Plac
et al., 1998). More recently, Zhong et al. (2011) analyzed expression was induced by adding 0.5 mM IPTG to culture
the theoretical feasibility of application of the method to media. Antibiotics were used at the following concentra-
surface conjugation, using simulated data from an interac- tions: tetracycline (Tc) = 10 mg/L; ampicillin
tive particle system model. They concluded that transfer (Ap) = 100 mg/L; kanamycin (Km) = 25 mg/L; nalidixic acid
rates (c) were more robust measurements than transfer fre- (Nx) = 20 mg/L and rifampicin (Rf) = 50 mg/L.
quencies (T/D or T/(R + T)) both in liquid conjugation and in
surface-associated conjugation but only if the donor and re- 2.2. Plasmid construction
cipient populations are forming a well-mixed confluent
layer. It was shown that, unlike the T/D, T/R or T/N ratios, Plasmids used and details of their construction are
the end-point estimate of c is insensitive to the D/R ratio shown in Table 1. They were constructed by using standard
and, to some extent, to the sampling time. recombinant DNA technology (Sambrook and Russell,
In this work we develop a simplified method of measur- 2001). DNA inserts were obtained by endonuclease diges-
ing conjugation rates under laboratory conditions (on the tion or PCR amplification with specific oligonucleotides
surface of agar plates) by analyzing conjugation of simple and subsequent ligation to the appropriate plasmid vector.
oriTs. In this way, transconjugants are not converted to The integrity of all constructions was confirmed by DNA
new donors and we measure the transfer rates of the donor sequencing.
population sensu stricto. By using a cytometric analysis of
the mating populations, we can estimate c directly from 2.3. Conjugation assays
the cytometer readings. We provide experimental evidence
that supports the statement of Zhong et al. (2011) that c is Plate-mating experiments were carried out in 24-well
a robust parameter, resistant to changes in the initial plates containing 1.0 mL LB-agar per well. Donor and reci-
experimental conditions. We further demonstrate that pient strains were grown o/n from single colonies in LB
conjugation frequencies can be boosted by remixing the medium at 37 °C with the appropriate antibiotics. After
mating populations. washing, fresh suspensions from stationary phase cultures
of donor and recipient strains were mixed in appropriate
volumes according to their OD600 to get the desired pro-
2. Materials and methods portion of donor and recipient. The mixture was centri-
fuged for 5 min at 4000g and resuspended in 1/10
2.1. Bacterial strains and culture conditions volume of LB medium. Fifteeen microliters of this cell mix-
ture (corresponding to a total OD600 of 0.6 units (roughly
All experiments were carried out using the Escherichia 6 108 cells)) were placed directly onto an agar surface
coli K12 derivative BW27783 (lacIq rrnB3 DlacZ4787 within a 24 well plate (previously filled with 1.0 mL
176 I. del Campo et al. / Plasmid 67 (2012) 174–182
detected simultaneously in individual bacterial cells by (Demarre et al., 2005)) and the oriT-containing plasmid
using a standard flow cytometer (Maksimow et al., 2002). pGP12. As pAP711DoriT::tet lacks a functional oriT but con-
In the present work, we monitored plasmid transfer by tains the full conjugative system of R388, it is able to mobi-
using the green fluorescent protein GFPmut2 (Cormack lize pGP12 oriT but it is not transferred itself to the
et al., 1996) and three different red fluorescent proteins: recipients. Once pGP12 reaches the recipient, PstbA be-
DsRed2, mCherry and mKate2 (Table 1). We designed an comes de-repressed (by lack of StbA) and gfpmut2 expres-
E. coli-based system in which a plasmid containing the oriT sion takes place. Recipients were BW-NxR cells containing
region of plasmid R388 is mobilized to a recipient cell by a pAC1, pAC2 or pAC4 (Table 1), plasmid constructions in
non-mobilizable derivative of plasmid R388 (Fig. 1). We which rfp expression is driven by the lac promoter (Plac).
reasoned that a system in which the transferred plasmid The three different red fluorescent proteins were used with
is not able to conjugate among recipients would be a sim- similar results (data not shown). Although the major exci-
pler system to analyze. This is due to the fact that newly tation peak for these proteins differs from the laser maxi-
acquired plasmids in transconjugants might conjugate at mum emission wavelength (488 nm), they were excited
different, probably significantly higher rates, a phenome- efficiently, giving reproducible results by detection with
non called zygotic induction (Jacob and Wollman, 1956). the FL2 filter for DsRed2 or with the FL3 filter for mCherry
Plasmid pGP12 (Table 1) contains the oriT of plasmid and mKate2. RFP proteins were used as markers for the re-
R388 cloned upstream of the gfpmut2 reporter gene, so that cipient cells by maintaining high levels of Plac expression
GFP expression is controlled by the stbABC operon pro- by the addition of 0.5 mM IPTG. In such a system, recipient
moter. In the presence of R388, PstbA transcription is re- cells could be identified as RFP+ events while transconju-
pressed by the presence of the StbA repressor protein gants show a GFP+ RFP+ phenotype.
coded by plasmid R388DoriT. Donors were BW-RfR cells To analyze the kinetics of R388_oriT transfer to recipi-
containing the non-mobilizable R388 (pAP711DoriT::tet; ent cells, conjugation was measured using both flow
Fig. 1. Scheme of the cytometric conjugation assay. Donor cells (D) were derivatives of strain BW-RfR carrying the non-mobilizable plasmid R388DoriT and
pGP12 (carrying an oriT:: construction); they appear in the cytometer as the grey population in Q3. They were mated with recipient BW-NxR cells (R)
containing a Plac::RFP construction, appearing as the red population in Q1. In donor cells, GFP expression is repressed by the presence of R388DoriT. When
pGP12 is transfer to the recipient cell, the PstbA::GFP promoter is induced, therefore producing GFP in the transconjugants (T), identified as the RFP+ GFP+
orange population in Q2.
178 I. del Campo et al. / Plasmid 67 (2012) 174–182
Fig. 2. Time series of a representative conjugation experiment as measured by flow cytometry. FL1 vs. FL3 plots showing gated populations. FL3+
population correspond to recipient cells (Q1). The FL3+FL1+ population (Q2), corresponding to transconjugants, increases with time. The Q3 population
corresponds to non-fluorescent events and contains donor cells.
cytometry and the traditional replica-plating method. A points were obtained by averaging at least nine indepen-
1:1 (D/R) conjugative mixture was prepared, spread on dent experiments. Error bars represent the standard devia-
individual wells of 24 well plates and incubated at 37 °C tion for each data point. The same result was obtained at
(Section 2). At different time points, mating mixtures were all four tested D/R ratios (data not shown). Interestingly,
resuspended in LB and a sample either fixed using parafor- as shown in Fig. 3, the error bars of the cytometer experi-
maldehyde for cytometry or plated on antibiotic-contain- ments were consistently smaller than those of the replica-
ing plates for selection for donors, recipients or plating assays. This is to be expected, since we are sam-
transconjugants. In flow cytometric measurements, the pling a much larger number of cells by cytometry (about
bacterial population was gated in a SSC vs. FSC plot and 20,000) than be replica-plating (a few hundred). Based on
set as P1. Twenty-thousand P1 events were collected and these results, we conclude that flow cytometry measures
FL1 vs. FL2 (or FL3) plot was used to study the different
populations. Donor cells were collected as non-fluorescent 1.0
events while recipient cells expressing RFP protein were
0.9
detected in gate Q1 corresponding to FL3+ events. As con-
Conjugation frequency
0.6 A 1.0
0.9
0.5
0.8
0.4 0.7
(fixed cells)
0.6
0.3 0.5
0.4
0.2
0.3
0.1 0.2
0.1
0
0.0
0 0.1 0.2 0.3 0.4 0.5 0.6
0 4 8 12 16 20 24
Conjugation frequency (fresh cells) Time (h)
Fig. 4. Effect of cell fixation on the estimation of cytometric-assayed B 11
conjugation frequencies. Comparison of the conjugation frequencies
obtained by flow cytometry (as described in Section 2) for the same
10.5
A -8.5 1.0
not mixed
50% 0.9 mixed at 6h
Conjugation frequency
-9 10%
5% 0.8
-9.5 2,5% 0.7
log10 γ ep
-10 0.6
0.5
-10.5
0.4
-11 0.3
-11.5 0.2
0 2 4 6 8 0.1
Time (h) 0.0
B -8.5
50 10 5 2.5
-9 Donor %
-9.5 Fig. 7. Effect of population mixing. The figure represents frequencies
log10 γ ep
-11.5
0 2 4 6 8 for the control of the dissemination of AbR (Andersson
Time (h) and Hughes, 2011; Baquero et al., 2011; Bonten et al.,
C -9 2001). Second, plasmids are the workhorses of genetic
engineering and, nowadays, systems and synthetic biology.
-9.5 In fact, plasmids are biological substrates that can be in-
-10 volved in biological processes to encode, store and manip-
log10 γ ep
conjugative transfer). Rather, the populations of donors the repressed version of R1 gave a c = 2 1012 (Levin
and recipients are mixed well enough, so they effectively et al., 1979). These values are higher but consistent to
behave as if they were in solution. Interestingly, the fact those obtained in our surface mating experiments. The
that our transconjugants continue to act as recipients does possibility that liquid-mating is more efficient than sur-
not seem to affect the transfer rates either (it is included in face-conjugation as a general rule should be rigorously
the ‘‘structure effect’’, which becomes significant only after tested. The 3-log rate difference between R1 and its de-re-
long mating times). In fact, if any donor bacterium is going pressed version underlines the potential ‘‘interfering’’ ef-
to engage in conjugation less than one time per experi- fect of zygotic induction on the calculation of c, as
ment on average, the ‘‘structure effect’’ should not be rele- discussed in this work. The values of c obtained in other
vant. This is what appears to happen in the experiments, as publications analyzing surface conjugation cannot be com-
shown in Fig. 6. Cells multiply and conjugate for about 8 h. pared directly with ours since they use surface areas in-
Little else occurs afterwards. stead of volumes. When the c of oriT_R388 is converted
Each well in a 24-well plate has a surface of 177 mm2 to these units, it gives 4.76 109 cm2 cell1 h1. Then, it
and was filled with about 1.0 mL of LB-agar. Therefore, a can be compared to that of plasmid pQBR11, which was
confluent E. coli monolayer (if each bacterium occupies 1 1011 in surface conjugation experiments between
1 lm2 – a cylinder 2.0 lm long and 0.5 lm base diameter) two Pseudomonas putida strains, or 4 1014 from P. putida
will consist of approximately 2 108 bacteria. According to Pseudomonas fluorescens (Lilley and Bailey, 2002). Signif-
to our estimates, we place roughly 3–6 108 bacteria/ icantly higher transfer rates were observed by Normander
plate well at the start of each conjugation experiment. This et al. (1998) in surface conjugation of the TOL plasmid
number was calculated, to get a maximal conjugation fre- (ranging from 7 1011 to 1 107). It is difficult to con-
quency (T/D) in a set of preliminary experiments that used clude anything from this scanty data. Certainly, a more
1:1 D/R ratio and 1 h mating periods (C. Revilla, data not extensive analysis with different plasmids and conditions
shown). Therefore, the bacterial cells are already in close is needed before we can extract general conclusions about
contact at the beginning of the mating experiment, form- plasmid conjugation rates.
ing at least a one-cell layer on the agar surface. Assuming In summary, we propose that Eqs. (2) and (3) could be
a height of 0.1 mm for the water column above the agar useful as a rule to estimate c and thus compare the transfer
surface, the resulting volume will be roughly 0.02 mL/well dynamics of different plasmids (or mutants of conjugation
(that is, 5 1010 cells/mL). These high cell concentrations genes). In this way significant differences can be discov-
seem to be essential for the close contacts needed for con- ered in the kinetic parameters that are related to different
jugation to happen at optimal rates. According to the mea- plasmid designs. As proof of principle, we analyzed the
surements in Fig. 5, the number of bacteria in the mating mobilization of a plasmid containing just the oriT region
mixtures remains constant for about 1 h and then in- of plasmid R388, instead of the native plasmid itself. Since
creases by a factor of 10 to 100-fold in the next 4–6 h (cor- the oriT cannot transfer from primary transconjugants to
responding to 4–7 doublings), growing little afterwards. new recipients, we were sure to deal with a single conjuga-
The lag phase is to be expected since experiments start tion rate. Experiments such as those detailed in this work
from stationary phase cultures, grown the previous day. should be extended to different systems and conditions
Besides, the mating plates have to adjust to the mating to adequately describe and quantify the conjugation kinet-
temperature. In summary, it seems that the surface of ics of the most common transmissible plasmids.
the plates can accumulate several layers of growing bacte-
ria in close contact. This crowded environment seems to be
6. Conclusions
optimal for surface conjugation.
The purpose of this work was not to analyze the theo-
We have developed a cytometric protocol for the esti-
retical aspects of conjugation kinetics on solid surfaces
mation of end-point conjugation rates. Our method is fas-
(as rigorously examined, for instance, by Zhong et al.
ter and less noisy than the classical replica-plating method.
(2011)). We rather wanted to provide a simple protocol
Using the cytometric method for the analysis of conjuga-
to estimate conjugation rates (that can be used in models
tion of the oriT of plasmid R388 we demonstrate that the
of conjugation kinetics) for the type of experiments that
end-point method is insensitive of D/R ratios and, to a cer-
are usually performed in a bacterial genetics laboratory.
tain extent, of the time of data collection. Conjugation fre-
We ended up with two equations, one that fits datasets ob-
quencies are significantly increased by re-mixing the
tained by cytometric analysis (Eq. (2)) and a second one
mating populations, until most recipients are converted
that is adapted to replica-plating datasets (Eq. (3)). Due
to transconjugants.
to its speed and simplicity, and the fact that the cytometric
estimates are less noisy (see Fig. 3), we favor the use of the
cytometric method. Acknowledgments
Under these experimental conditions, our estimate for c
of oriT_R388 was 5.46 1011 mL cell1 h1. It should be This work was supported by Grants BFU2008-00995/
emphasized that this value considers a volume of 20 lL BMC from Ministerio de Ciencia e Innovación (MCINN,
for the mating mixture, as discussed above. This value of Spain), RD06/0008/1012 from Instituto de Salud Carlos
c can be compared to estimates in liquid mating experi- III, and 248919/FP7-ICT-2009-4 from the European VII
ments of E. coli plasmid F and of the derepressed plasmid Framework Program. Ms. María Aramburu is also acknowl-
R1-drd-19 (both giving c = 2 109 mL cell1 h1), while edged by her expert assistance with the flow cytometer.
182 I. del Campo et al. / Plasmid 67 (2012) 174–182
References Levin, B.R. et al., 1979. The kinetics of conjugative plasmid transmission:
fit of a simple mass action model. Plasmid 2, 247–260.
Licht, T.R. et al., 1999. Plasmid transfer in the animal intestine and other
Andersson, D.I., Hughes, D., 2011. Persistence of antibiotic resistance in
dynamic bacterial populations: the role of community structure and
bacterial populations. FEMS Microbiol. Rev. 35, 901–911.
environment. Microbiology 145 (Pt 9), 2615–2622.
Baquero, F. et al., 2011. Ecology and evolution as targets: the need for
Lilley, A.K., Bailey, M.J., 2002. The transfer dynamics of Pseudomonas sp.
novel eco-evo drugs and strategies to fight antibiotic resistance.
plasmid pQBR11 in biofilms. FEMS Microbiol. Ecol. 42, 243–250.
Antimicrob. Agents Chemother. 55, 3649–3660.
Llosa, M. et al., 1991. Structural and functional analysis of the origin of
Bevis, B.J., Glick, B.S., 2002. Rapidly maturing variants of the Discosoma
conjugal transfer of the broad-host-range IncW plasmid R388 and
red fluorescent protein (DsRed). Nat. Biotechnol. 20, 83–87.
comparison with the related IncN plasmid R46. Mol. Gen. Genet. 226,
Bonten, M.J. et al., 2001. Understanding the spread of antibiotic resistant
473–483.
pathogens in hospitals: mathematical models as tools for control.
Maksimow, M. et al., 2002. Simultaneous detection of bacteria expressing
Clin. Infect. Dis. 33, 1739–1746.
GFP and DsRed genes with a flow cytometer. Cytometry 47, 243–247.
Boucher, H.W. et al., 2009. Bad bugs, no drugs: no ESKAPE! An update
Monod, J., 1949. The growth of bacterial cultures. Ann. Rev. Microbiol. 3,
from the Infectious Diseases Society of America. Clin. Infect. Dis. 48,
371–394.
1–12.
Normander, B. et al., 1998. Effect of bacterial distribution and activity on
Cormack, B.P. et al., 1996. FACS-optimized mutants of the green
conjugal gene transfer on the phylloplane of the bush bean (Phaseolus
fluorescent protein (GFP). Gene 173, 33–38.
vulgaris). Appl. Environ. Microbiol. 64, 1902–1909.
de la Cruz, F., Davies, J., 2000. Horizontal gene transfer and the origin of
Sambrook, J., Russell, D.W., 2001. Molecular cloning: a laboratory manual.
species: lessons from bacteria. Trends Microbiol. 8, 128–133.
Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
de la Cruz, F. et al., 2010. Conjugative DNA metabolism in Gram-negative
Shaner, N.C. et al., 2004. Improved monomeric red, orange and yellow
bacteria. FEMS Microbiol. Rev. 34, 18–40.
fluorescent proteins derived from Discosoma sp. red fluorescent
Demarre, G. et al., 2005. A new family of mobilizable suicide plasmids
protein. Nat. Biotechnol. 22, 1567–1572.
based on broad host range R388 plasmid (IncW) and RP4 plasmid
Shcherbo, D. et al., 2009. Far-red fluorescent tags for protein imaging in
(IncPalpha) conjugative machineries and their cognate Escherichia coli
living tissues. Biochem. J. 418, 567–574.
host strains. Res. Microbiol. 156, 245–255.
Simonsen, L., 1990. Dynamics of plasmid transfer on surfaces. J. Gen.
Garcillan-Barcia, M.P. et al., 2011. Identification of bacterial plasmids
Microbiol. 136, 1001–1007.
based on mobility and plasmid population biology. FEMS Microbiol.
Simonsen, L. et al., 1990. Estimating the rate of plasmid transfer: an end-
Rev. 35, 936–956.
point method. J. Gen. Microbiol. 136, 2319–2325.
Jacob, F., Wollman, E.L., 1956. Processes of conjugation and recombination
Smillie, C. et al., 2010. Mobility of plasmids. Microbiol. Mol. Biol. Rev. 74,
in Escherichia coli. I. Induction by conjugation or zygotic induction.
434–452.
Ann. Inst. Pasteur (Paris) 91, 486–510.
Zaslaver, A. et al., 2006. A comprehensive library of fluorescent
Khlebnikov, A. et al., 2001. Homogeneous expression of the P(BAD)
transcriptional reporters for Escherichia coli. Nat. Methods 3, 623–628.
promoter in Escherichia coli by constitutive expression of the low-
Zhong, X. et al., 2011. On the meaning and estimation of plasmid transfer
affinity high-capacity AraE transporter. Microbiology 147, 3241–3247.
rates for surface-associated and well-mixed bacterial populations. J.
Levin, B.R. et al., 1997. The population genetics of antibiotic resistance.
Theor. Biol.
Clin. Infect. Dis. 24 (Suppl. 1), S9–S16.
Negative Feedback and Transcriptional Overshooting in
a Regulatory Network for Horizontal Gene Transfer
Raul Fernandez-Lopez, Irene del Campo, Carlos Revilla, Ana Cuevas, Fernando de la Cruz*
Instituto de Biomedicina y Biotecnologia de Cantabria (IBBTEC), Santander, Spain
Abstract
Horizontal gene transfer (HGT) is a major force driving bacterial evolution. Because of their ability to cross inter-species
barriers, bacterial plasmids are essential agents for HGT. This ability, however, poses specific requisites on plasmid
physiology, in particular the need to overcome a multilevel selection process with opposing demands. We analyzed the
transcriptional network of plasmid R388, one of the most promiscuous plasmids in Proteobacteria. Transcriptional analysis
by fluorescence expression profiling and quantitative PCR revealed a regulatory network controlled by six transcriptional
repressors. The regulatory network relied on strong promoters, which were tightly repressed in negative feedback loops.
Computational simulations and theoretical analysis indicated that this architecture would show a transcriptional burst after
plasmid conjugation, linking the magnitude of the feedback gain with the intensity of the transcriptional burst.
Experimental analysis showed that transcriptional overshooting occurred when the plasmid invaded a new population of
susceptible cells. We propose that transcriptional overshooting allows genome rebooting after horizontal gene transfer, and
might have an adaptive role in overcoming the opposing demands of multilevel selection.
Citation: Fernandez-Lopez R, del Campo I, Revilla C, Cuevas A, de la Cruz F (2014) Negative Feedback and Transcriptional Overshooting in a Regulatory Network
for Horizontal Gene Transfer. PLoS Genet 10(2): e1004171. doi:10.1371/journal.pgen.1004171
Editor: Ivan Matic, Université Paris Descartes, INSERM U1001, France
Received May 13, 2013; Accepted December 26, 2013; Published February 27, 2014
Copyright: ß 2014 Fernandez-Lopez et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Work was financed by the Spanish Ministry of Economy and Competitivity (BFU2011–26608) and the European Seventh Framework Program (289326/
FP7-KBBE-2011-5 and 282004/FP7–HEALTH-2011-2.3.1–2) to FdlC. The funders had no role in study design, data collection and analysis, decision to publish, or
preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: [email protected]
Introduction copies within a given cell favors plasmids with higher copy
number, superior partition systems and higher transfer rates [9].
Horizontal gene transfer (HGT) is ubiquitous in bacteria. However, these processes come to a cost, since plasmid
Because its important role in bacterial adaptation, HGT has been consumption of resources imposes a metabolic burden that
traditionally compared to sexual reproduction in higher eukary- hampers host fitness. Competition between plasmid-containing
otes. In bacteria, however, HGT is not mediated by specific cells, on the contrary, selects for plasmids that minimize the
intracellular mechanisms, but it is the byproduct of the pervasive burden imposed onto the host. Both selection levels are thus
movement of a myriad of mobile genetic elements. These include intrinsically in conflict, and an adequate regulation of gene
transposons, phages, ICEs and, most notably, plasmids [1]. expression becomes essential to ensure plasmid survival [10].
Among them, broad host range (BHR) plasmids of Proteobacteria Transcriptional regulation is common in plasmids, and virtually
stand out because of their ability to colonize a wide range of all functions in plasmid physiology have been found to be under
bacterial species. This ability makes BHR plasmids efficient transcriptional control [11]. In some cases, this control is exerted
shuttles for HGT, clearly illustrated in the last decades by their in sophisticated and apparently redundant layers. For example,
leading role in the spread of antibiotic resistance genes among plasmid R1 replication is controlled simultaneously by an antisense
microbial populations [2]. RNA and a transcriptional repressor [12] [13]. In other cases, like
Bacterial plasmids are agents for HGT, but they themselves are in the broad host range plasmid RP4, transcriptional regulation is
genetic replicons with their own, idiosyncratic, evolutionary organized under a global network that coordinates all functions in
history [3,4]. Plasmid fitness depends on two basic physiological the plasmid’s physiology [14]. Unfortunately, despite our knowl-
functions: maintenance within the bacterial host and transfer into edge of the molecular biology of transcriptional regulation, our
new hosts; functions that are encoded in the plasmid genome. understanding of the signals that plasmid regulatory circuits
However, since plasmids can only exist inside a bacterial cell, their respond to is scarce. Plasmids from Gram+ bacteria regulate
fitness is also host dependent. Plasmids impose a burden on host conjugation according to external cues about the abundance of
fitness [5,6,7,8], which is dependent on the collective effect of the possible receptors. These cues are communicated in the form of
plasmid population within a given cell. The overall fitness of a specific pheromones [15]. Such systems are generally not found in
plasmid replicon therefore depends not only on its own phenotype, plasmids from Proteobacteria, with the remarkable exception of Ti-
but also on the phenotype of other co-residing plasmid copies. like plasmids from Alpha-Proteobacteria, which monitor external
This dependency on the group phenotype puts plasmids under conditions via a quorum sensing mechanism[16] [17]. Apart from
multilevel selection [9]. Multilevel selection forces plasmids to these and a few other cases, the input information that plasmid
confront a paradoxical situation. Competition between plasmid regulatory circuits monitor remains elusive.
Author Summary to background, and another six (PresP, PkfrA, PssB, PstbA, PkorA,
PkikA) showed values similar to those of repressed PlacZ (1*103
In the environment, bacteria often evolve by the acquisi- GFP/OD). The only promoters that showed no changes in the
tion of new genes from different species. Plasmids are presence of plasmid R388 were Pint and Pant. Interestingly, these
small DNA molecules that mediate horizontal gene transfer promoters do not belong to the plasmid backbone: they are part of
in bacteria, thus they are fundamental agents for the the In-3 integron platform, which incorporated recently, in
spread of antibiotic resistances. Plasmids replicate inside evolutionary terms, into the plasmid genetic structure [20].
the bacterial cytoplasm, and propagate infectiously by Therefore, when the full regulatory network was present, all
contact. Plasmids control these two ways of multiplication, promoters from the plasmid backbone were repressed, and kept at
but like many other symbionts they suffer from a tradeoff. levels lower or similar to LacI-repressed PlacZ.
If plasmids become very infective, they can spread fast and
To determine the transcriptional units of the plasmid, mRNA
successfully, but this damages the bacterial hosts they
levels during exponential growth were analysed by RT-qPCR.
depend upon. If, on the contrary, plasmids become very
mild, the host is able to grow better but the ability of Relative mRNA abundance was measured using a set of 66 primer
plasmids to infect new hosts is hampered. We have pairs, designed to cover the entire plasmid genome. From these 66
studied the regulatory mechanisms plasmids use to primer pairs, 60 showed efficiencies in the interval 0.9,E,1.2
overcome this paradox. We discovered that negative (Figure 2A, left upper panel) and were considered suitable for
feedback, a regulatory motif ubiquitous in the plasmid quantification. mRNA was extracted from cells growing in rich
network, allows transient activation of plasmid functions media at mid-exponential phase, and retrotranscribed into cDNA
immediately after plasmids invade a new host. This might as described in Materials and Methods. Using 300 ng total RNA,
be an adaptive strategy for plasmids to be highly infective we obtained an average threshold cycle (Ct) of 19.9 with cv = 0.12.
without damaging their hosts, and it illustrates a natural Results for each primer pair were normalized measuring the Ct
mechanism for DNA transplantation that could be imple- corresponding to 5 ng of purified plasmid DNA. Results
mented in synthetic genomic transplants. (Figure 2A, left lower panel), showed a tight distribution with an
average Ct of 14.2 and cv = 0.034. Relative abundances of
mRNAs were expressed as DCt = CtcDNA-CtDNA [21] and a
Trying to understand the fundamental principles of plasmid representation of the average DCt obtained for each primer pair in
transcriptional control, we analyzed the regulatory network of three independent experiments is shown in figure 2B. Known
plasmid R388. Plasmid R388 is the smallest BHR conjugative untranscribed regions, like the plasmid origin of transfer (between
plasmid found in Proteobacteria [3]. It shows an extensive host range, PstbA and PtrwA), yielded DCt = 28, while the most actively
which overlaps with that of plasmid RP4, another model BHR transcribed region corresponded to the integron cassettes, with
plasmid [18]. Remarkably, plasmid RP4 is phylogenetically DCt = 2. Overall, the transcriptional profile clearly delimitated the
unrelated to plasmid R388 [3]. This situation allowed us to boundaries between transcriptional units (Figure 2.B). A compar-
compare two plasmid networks that evolved independently, but ison between promoter strengths, determined by fluorescence
under similar selective constraints. Using fluorescence expression profiling, and mRNA abundances, measured by RT-qPCR
profiling and quantitative PCR, we found a global regulatory (figure 2A, left lower panel), showed that both measurements
network that controlled plasmid R388 transcription. Unlike the were not linearly correlated (r2 = 0.49), indicating that mRNA
archetypical plasmids from Gram+ bacteria or Ti-like plasmids, the processing and degradation also played significant roles in
network seemed to be unresponsive to environmental changes, or determining plasmid levels of expression.
quorum signals. The network was based on a basic regulatory To determine the topology of the plasmid regulatory network,
motif: a strong promoter controlled by a tight negative feedback we tried to ascribe each plasmid promoter to its cognate
loop (NFL). We show, computationally and experimentally, that regulators. ORFs from the plasmid genome that were either
this architecture induces transcriptional overshooting after hori- orphan, or showed homology to known transcriptional regulators,
zontal transfer of the plasmid. were considered potential candidates to encode a plasmid
regulator. These ORFs were cloned in expression vector pBAD33,
Results and the transcriptional activity of plasmid promoters was
measured in the presence of all putative regulators. Expression
Intergenic regions in plasmid R388 DNA were PCR-amplified profiles are shown in supporting figures S1 and S2, and steady-
and cloned in the low copy number reporter vector pUA66 [19] to state levels are indicated in Supporting Table S1. Results allowed
drive transcription of gfpmut2 after a strong ribosomal binding site. the identification of six plasmid proteins (ResP, KfrA, ArdK, StbA,
Out of the 19 intergenic regions cloned, 15 showed transcriptional TrwA and KorA) able to repress at least one of the plasmid
activities at least two-fold higher than the promoter-less vector, promoters. (Supporting Table S1 and Supporting figure S1).
and were considered to contain a promoter (Figure 1). To Among the regulators identified, we did not find any transcrip-
compare the transcriptional strength of these promoters against a tional activator. All regulators were repressors involved in negative
known standard, the activity of ParaBAD was measured at different feedback loops (Figure 3A). Three of them controlled only their
arabinose concentrations. ParaBAD promoter reached 105 GFP/ own promoter, thus constituting simple negative feedback loops
OD units at maximal induction, and 13 out of 15 R388 promoters (ResP, KfrA and TrwA). Another three (ArdK, StbA and KorA)
showed levels similar to this value (Figure 1, Supporting Table S1). controlled more complicated circuits. Protein ArdK repressed
Therefore R388 promoters, when assayed in the absence of the expression from PardC, Porf7, Porf12, Porf14 and Pssb, its own
plasmid network, have strong transcriptional activities. promoter (Supporting fig. S1). All these promoters direct the
Transcriptional activity decreased sharply when the promoters transcription of genes involved in the stable maintenance of the
were assayed in cells that also contained plasmid R388 (Figure 1, plasmid [3]. Therefore, ArdK seems to regulate the vegetative
Supporting Table S1). The repression fold exerted by the plasmid maintenance of plasmid R388. Similarly, protein KorA was found
ranged from 5 (PresP) to more than 500 fold (Porf7). Six promoters to regulate PtrwH, PkorA, PkikA, PkorB and its own promoter,
(PardC, Porf14, Porf12, PkorB, PtrwH, Porf7) dropped to levels close PkorA (Supporting Fig. S2). All these promoters are responsible for
Figure 1. Promoters in plasmid R388. The figure shows the location and transcriptional activity of promoters detected in plasmid R388 genome.
The location of each promoter is indicated by an arrow on the red circle. Each promoter receives the name of the first gene encoded in the
transcriptional unit. Bar charts indicate the expression levels when the promoter activity was measured alone (black columns) or in cells that also
contained plasmid R388 (grey columns). The expression levels (,GFP/OD.) represent the average GFP/OD (6102) level achieved by cells growing in
exponential phase. Each column represents the average and standard deviation of at least five independent experiments.
doi:10.1371/journal.pgen.1004171.g001
Figure 2. Transcriptional units in plasmid R388. (A) Statistics of the 66 primer pairs used to measure transcriptional levels in plasmid R388
(Upper right) Histogram showing the efficiency (calculated as indicated in materials and methods) of the primer pairs. (Upper right) Histogram
showing the Ct obtained in qPCR amplifications from plasmid cDNA. (Lower left) Histogram showing the Ct obtained in qPCR amplifications from
purified plasmid DNA. (Lower right) Scatter plot showing the relationship between the promoter activity (obtained from figure 1, in GFP/OD unit on
the y axis) and the mRNA levels measured by RT-qPCR (DCt, x axis). (B) Transcriptional landscape of plasmid R388. The graph shows the relative
abundance of mRNA, indicated as DCt = CtcDNA-CtDNA) along the plasmid genome. Each unit in the y axis corresponds to a 2 fold increase in mRNA.
Peaks correspond to highly transcribed regions and valleys correspond to non-transcribed regions. The highlighted blue lines indicate the
overlapping of the transcriptional units and the plasmid promoters identified in figure 1.
doi:10.1371/journal.pgen.1004171.g002
expression of the Type IV secretion system, involved in plasmid involved in a complex regulatory circuit was StbA. Gene stbA is part
conjugation. Therefore, KorA acted as the main transcriptional of an operon responsible for plasmid segregation [22] and was found
regulator for expression of the conjugative pilus. The third protein to decrease PstbA transcription 50-fold (Figure 3A, Supporting
Table S1). StbA also repressed transcription from promoters PardC, environmental changes. We modified growth medium (LB,
Porf7, Porf12 and Porf14; indicating that its target repertoire minimal M9), temperature (30, 37 and 42uC) and tested the
overlaps with that of ArdK (Supporting Fig. S1). Similarly, StbA presence of stressing agents, like sub-inhibitory concentrations of
repressed the promoters targeted by KorA, although the level of antibiotics and common triggers of the SOS response (Materials
repression exerted was significantly lower (2 to 10-fold decrease and Methods). As judged from fluorescent expression profiling,
compared to the 90-fold decrease produced by KorA on PtrwH) none of these signals was found to specifically activate any
(Supporting Fig. S2). Interestingly, previous work on StbABC promoter in the network (Supporting figures S3, S4 and S5). The
showed that this operon balances plasmid partition and conjugation possible effect of Escherichia coli recipient cells was also tested by co-
[23]. Results presented here indicate that StbA acts as a common culture in liquid media with plasmid free cells (Supporting figure
regulator for genes involved in the vegetative and conjugative S6). Since plasmid R388 can only conjugate on solid surfaces [24],
functions of the plasmid. these conditions prevented horizontal transfer of the plasmid,
These results allowed us to determine the topology of plasmid while allowing the donors to sense any potential signal from the
R388 transcriptional network, which is depicted in Figure 3B. The recipient cells. Again, the regulatory network was unresponsive
network is completely dominated by negative repression, and (Supporting figure S6), indicating that, in the conditions tested, the
promoter activation will depend on signals levering the action of network did not respond to any diffusible signal from the recipient
plasmid repressors. In order to identify the signals that the network cells. Altogether these results indicated that plasmid R388 does
responded to, we challenged the plasmid with a plethora of neither respond to pheromones (unlike many plasmids from Gram
Figure 3. Negative feedbacks and topology of plasmid R388 transcriptional network. (A). Each panel shows the transcriptional activity
(GFP/OD) (6102) of a given promoter, either alone or in the presence of each of the six transcriptional repressors (ResP, KfrA, ArdK, StbA, TrwA and
KorA). Repressors were produced from a co-residing expression vector pBAD33, and the negative control indicates the GFP/OD (6102) values
obtained in the presence of the empty vector. The upper diagrams show the location of each regulator with respect to its cognate promoter in
plasmid R388 (B) A graphical representation of expression profiling data (shown in Supporting figures S3 and S4) unveils the topology of the
regulatory circuitry. Coloured arrows indicate the position of transcriptional regulators within plasmid R388 genome (ResP in orange, KfrA in blue,
ArdK in purple, StbA in green, TrwA in red and KorA in navy blue). The regulatory links are coloured according to the same code. Red lines shown
over the ORF map correspond to the transcriptional units identified in Figure 2.
doi:10.1371/journal.pgen.1004171.g003
+ bacteria [15]), nor quorum sensing signals (unlike Ti plasmids the value of X at steady state (Xss), simulations showed that
from Agrobacterium [16,17]) nor SOS inducing agents (like many increasing the promoter strength (l) or decreasing K (increasing
phages and ICEs [25]) the strength of the repression, i.e. the affinity of the regulator for its
The absence of responses against environmental challenges, cognate site) increased correspondingly the size of the transcrip-
DNA damage or quorum signals suggested that plasmid regulation tional overshoot (Figure 4A). This dependency strongly suggested
is disconnected from the main sensory circuitry of the host cell. that there should be a correlation between the overshoot and the
However, sensing is not the only function that transcriptional relative strength of the repression exerted by the NFL. The relative
regulation can undertake; generating temporal programs, or strength of the feedback can be expressed in terms of feedback
guarding the cell homeostasis are also adaptive functions that gain (G) (Figure 4B). We define G as the ratio between the steady
arise purely from the internal dynamics of regulatory systems. In states shown by the open loop (without the repressor) and the
order to study the internal dynamics of the plasmid network, we closed loop (with the repressor)
used a simple quantitative model. Since all transcriptional
regulators in the plasmid were self-repressors (Figure 3), we used n
l1 l2 nz1
a simple ordinary differential equations (ODE) model of a negative G~ ð2Þ
feedback loop. Let X denote the mRNA and Y the protein b1 b2 k
concentrations for a given feedback loop, the system of differential
equations that describe the system follows:
This expression indicates that the feedback gain is directly
dependent on the transcriptional/translational strength (l1 l2) and
dx kn inversely correlated to the feedback constant k. Similarly, the
~ l1 n { b1 x
dt
|{z} k zyn |{z} overshoot (O) can be expressed as the ratio between the maximum
|fflfflfflfflfflffl{zfflfflfflfflfflffl} degradation
mRNA production value on X divided by its steady state. Then, by linearizing X
turnover
ð1Þ before the onset of the repression loop we can approximate O as:
dy
~ l2 x { b2 y
dt |{z} |{z}
|{z} production degradation Ox % 1{e{b1 tx~xmax G ð3Þ
protein
turnover
This approximation indicates that the stronger the gain (G) the
This equation is based on the assumption that, in the absence of higher the overshoot will be. This approximation holds for highly
repressor binding, mRNA is transcribed at rate l1, and translated non-linear systems, with high values of n (Figure 4C, left panel).
at rate l2. Repressor binding is modelled following simple mass- However, if we introduce a dimerization step where two monomers
action kinetics. This binding is characterized by a half maximal of repressor Y need to interact to form an active dimer, the
binding constant k, which is the ratio between the dissociation and approximation holds for all n (Figure 4C, right panel). The fact that
binding constants (k = koff/kon). The model allows cooperative nearly all transcription factors from Prokaryotes act as multimers
binding, with cooperativity index n (n = 1 for non cooperative indicates that this is a conservative assumption [28]. Equation 3
binding). Parameters b1 and b2 correspond to the degradation indicates that O is proportional to the gain G, and to the time to
rates of mRNA and the regulator, respectively. This simple ODE reach the maximal value of X (in the limit t = = .‘, e2bt<0 and
model has been extensively used in the literature, and was shown O<G). This means that O increases with higher delays, and the
to confer different properties, such as decrease the response time higher the feedback gain, the more prominent the transcriptional
and increase the stability of transcriptional sensory systems[26,27]. overshoot will be. Previous computational analysis of other feedback
These properties are characteristic of negative feedback loops loops showed similar dependencies between the intensity of the
whose components (mRNAs and proteins) are in steady state. overshoot and the strength of the feedback gain [29]. Therefore,
However, apart from these and other steady-state properties, simulations and theory predicted that a network architecture based
NFLs are known to exhibit transient behaviours while adjusting to on strong promoters, tightly repressed in negative feedback loops,
the steady-state. In electrical engineering it is well known that would exhibit significant transient overshooting after HGT. For
NFLs can overshoot before reaching steady-state when they start more complex circuits of the plasmid network that are under the
from initial zero conditions (x = 0 , y = 0 at t = 0). In biological control of two transcriptional regulators, transient overshooting is
contexts, this property has received little attention, the reason also expected (Supporting figure S7). Due to the OR logic that rules
being that daughter cells inherit not only the chromosome but also these circuits, the overshooting was dependent on the transcrip-
a proportional part of its regulatory elements. Thus, in the normal tional regulator that first achieved its effective values. This, in turn,
life of bacteria, transcriptional NFLs do not usually experience will depend of its affinity for the target promoter (k) and its own
situations with zero concentration of its constituents. However, transcriptional/translational strength, as in simple NFLs.
conjugative plasmids have a specific mechanism of invasion, To test whether this transcriptional overshoot could be detected
entering a cell in the form of ssDNA, without accompanying experimentally, we determined mRNA expression levels during
mRNAs or transcriptional regulators. Simulations of Eq 1. conjugative transfer of the plasmid. Conjugative mixes with 1:1
mimicking these conditions produced an overshoot, showing that donor/recipient cell ratio were allowed to mate for 0, 30, 60, 90
both the mRNA and the protein experienced a transitory burst and 120 min. Total RNA was extracted from time samples, and
and then relaxed to their steady state values (Figure 4A). While expression levels from the main plasmid operons were measured
mathematical analysis indicated that a temporal lag between the by RT-qPCR. Expression levels were normalized by the results
mRNA and the protein was enough to produce overshooting obtained from a constitutive chromosomal gene (dxs). In order to
(Supporting Text S1), computational analysis indicated that the check for possible mRNA increases due to conjugative replication
magnitude of this overshoot is heavily dependent on the of the plasmid, we measured the relative increase in non-
parameters of the system. Defining the magnitude of the overshoot transcribed regions (oriT), and also in promoters that were not
as the ratio between the maximal levels reached by X (Xmax) and negatively regulated (dhfR gene, controlled by Pant). Experimental
Figure 4. Transcriptional overshooting and its relationship with feedback gain. (A) Numerical simulations showing the effect of increasing
feedback gain on the magnitude of the transcriptional overshoot. Left panels show the relative abundance of mRNA (X) normalized by its steady-
state value (Xss) along time. Right panels show the phase-plane portrait of the system, where the x axis corresponds to the normalized mRNA values
(variable X in Ec.1) and the y axis corresponds to the normalized regulator levels (variable Y in Ec.1). Values were normalized by their respective
steady-state levels. Upper panels show the effect of increasing the feedback gain by decreasing the feedback constant K. Simulations were performed
with l1 = 10, b2 = 0.2, l2 = 10, b1 = 1 and n = 1. Lower panels show the effect of increasing the feedback gain by increasing the intrinsic transcription
rate l1. Simulations were performed with l1 = 0.01, b2 = 0.2, l2 = 10, b2 = 1 and n = 1. The figure shows that the maximal values of X and Y grow as
the feedback gain is increased, either by decreasing K or increasing l1 (B) Scheme showing the theoretical time evolution of an open loop (blue) and
a closed negative feedback loop (red). The feedback gain (G) corresponds to the ratio between the steady states of both systems, being all
parameters equal (blue dashed line). The overshoot (O) corresponds to the transient production above the steady-state levels experienced by the
closed loop when starting from initial conditions t = 0, x = 0, y = 0 (red dashed line) (C) Performance of the theoretical approximation described in Ec.
4 , compared to numerical simulations. Both panels show results obtained by numerical integration of Ec.1 (white dots) and predicted overshoots
obtained from Ec. 4 (black dots). All simulations and calculations were done in a system with parameters k = 0.01, l1 = 10, b2 = 0.2, l2 = 10, b1 = 1 and
changing the cooperativity of the repression (n, x axis). The left panel corresponds to a system where regulator Y is allowed to repress its own
synthesis immediately after translation, while the right panel corresponds to the same system but including the requirement of Y dimerization before
binding to DNA. Dimer formation is simulated by a simple ODE with Ka = 0.1 and Kd = 0.01. The inner graphs on the right chart show the phase-
portrait of the system, with mRNA on the x axis and regulator concentration on the y axis. As shown in the figure, when the number of binding sites is
higher than n = 10 the system becomes cyclostationary, opening the possibility of periodical bursts of transcription.
doi:10.1371/journal.pgen.1004171.g004
procedures are detailed in materials and methods. Results, shown (trwF, ardC, ssb and klcB) corresponded to promoters with the
in figure 5, indicated that the expression levels of oriT and dhfR higher gains (PtrwH, PardC, Pssb and Porf12). On the other
showed limited changes, while genes controlled by negative hand, those promoters with lower gains (PtrwA and PresP) also
regulation (resP, ardC, ssb, klcB, trwA and trwF) increased their yielded lower overshoots (trwA, repA in fig. 5), as predicted by
relative abundance immediately after conjugation. Experimental theory. Since conjugation is inherently asynchronous (newly
results showed that those genes that showed the highest induction formed transconjugants become donors and infect new receptors),
Figure 5. Promoter induction after horizontal transfer of the plasmid. RT-qPCR was used to measure mRNA levels. Bars indicate the relative
ratio of mRNA at each time point compared to the values obtained in the absence of conjugation (time 0). Asterisks indicate the statistical
significance of the differences observed * = p,0.1, ** = p,0.05 Experimental procedures and calculations are detailed in Materials and Methods and
expanded results are shown in Supporting Figure S4. Measurements represent the average of three independent mRNA extractions.
doi:10.1371/journal.pgen.1004171.g005
our population measurements resolved poorly the actual kinetics of selection imposes opposing demands on plasmid physiology that
the overshoot. Also, the kinetics of the overshoot for individual require a delicate equilibrium between the expression of plasmid
NFLs will depend critically on the mRNA half-life (Eq.3), which is functions and the burden imposed on the host cell [9].
also likely to be variable from gene to gene. As a consequence, the Understanding the regulatory mechanisms of plasmid transcrip-
decrease in the overshoot is only observable in some of the genes tional control might shed light on the way plasmids conciliate these
tested (ardC, ssb,trwH trwA). However, steady-state measurements requirements.
(equivalent to time 0 in figure 5) indicated that all promoters would In this work we describe the topology and dynamics of the
eventually return to basal levels. For the klcB gene, controlled by transcriptional network of plasmid R388, the smallest BHR
Porf12 promoter, which yielded no observable overshoot, it is not plasmid from Proteobacteria. The network consisted exclusively of
possible to state at this point whether the overshoot was obscured transcriptional repressors. This preference for transcriptional
by population effects, or the parameters of this promoter did not repression is in contrast with the situation described for the
yield any significant overshoot. regulatory networks of bacterial chromosomes. For example, in
This transient induction could have phenotypic effects on the host E.coli the number of transcriptional activators roughly equals the
cell. Plasmids impose a burden on the host, meaning that, in the number of repressors [33]. However, other transcriptional
absence of positive selection for plasmid-encoded traits like networks from BHR plasmids, like plasmid RP4, were also found
antibiotic resistances, plasmids must survive as parasitic entities to depend solely on transcriptional repressors [11]. In plasmid
[30,31,32]. It is conceivable that a transient increase in plasmid gene R388, transcriptional repression was exerted mainly in the form of
expression will translate into a higher burden to the host cell. To test negative feedback loops. These feedback loops showed high gains
whether any effect on host fitness could be observed, we measured (defined as the ratio between the expression levels of the open and
the growth rates of donor, recipient and transconjugant cells, the closed feedback loops). Although we are not aware of any
immediately after conjugation. We used two spontaneous Rifr and systematic, quantitative study of a plasmid regulatory network,
Nxr mutants of E.coli strain Bw27783, which showed no observable several independent studies have reported that the regulatory
differences in growth rate (figure 6A). Cells from both strains that network of plasmid RP4 contains strong promoters that are kept
had carried the plasmid for at least 10 generations exhibited a 17% tightly repressed by the plasmid regulators [14,34,35,36,37,38].
increase in the generation time when compared to plasmid free cells Remarkably, plasmids R388 and RP4 show similar broad host
(Figure 6A). This indicated that the plasmid exerted a measurable ranges, but they are not phylogenetically related [3,4,39]. This
burden on the host Plasmid conjugation assays were performed on indicates that both plasmids, which presumably suffer from
LB agar surfaces in a 1:1 donor/recipient ratio, and cells were analogous selective constrains, have independently evolved tran-
allowed to mate for 30 min. Conjugation was stopped by scriptional networks with analogous topologies.
resuspending cells from the solid surface, cells were diluted to Simulations and theory indicate that whenever a negative
OD600<0.01 in fresh LB, placed in agitation at 37 C and allowed to feedback loop has a gain higher than 1 and a certain time delay
grow for 5 h (Figure 6B). Growth rates were measured by plating on between the mRNA and the regulatory protein, the system would
selective antibiotics (materials and methods). Plasmid R388 does not show transient overshooting (Eq 3 and figure 4A). The actual
conjugate in liquid media, thus any variation in the proportion of production of the overshoot requires the system to begin with zero
donors, recipients and transconjugants must be due to relative initial concentration of transcriptional repressors (t = 0, x = 0,
differences in growth rates. Results, shown in figure 6C, indicated y = 0, in Eq 1), allowing the separation of timescales to produce a
that transconjugant cells suffered a remarkable decrease in growth period of repressor-free transcription. For conjugative plasmids,
rate immediately after conjugation, showing a first generation time this situation is met every time the plasmid enters into a new cell
of about 2.56times that of donor cells. However, after this long first by conjugation. In fact, any negative feedback loop that undergoes
generation, transconjugant cells recovered, achieving the same conjugation is likely to experience transient overshooting. It has
number of divisions as donor cells for the total duration of the been known for a long time that a lysogenic phage transferred by
experiment (7 generations). Similar results were obtained when Hfr conjugation (an artificial system that allows the horizontal
donor and recipient strains were reversed (Supporting Figure S8). transfer of the entire chromosome), can become lytic when
The observed growth deficit in the transconjugants could be a by- entering into a new host [40]. A similar behavior was also observed
product of the conjugation mechanism, which requires the piercing when an RFP-TetR autogenously regulated cassette was inserted
of the recipient cell by the transfer apparatus. To test whether this in the E.coli chromosome and transferred by Hfr conjugation into a
was the case, we carried out a similar experiment with a mobilizable new cell [41]. Transient overshooting is therefore an epiphenom-
plasmid. In this case, a small cloning vector carrying just the origin enon associated to negative feedback loops that experience some
of transfer (oriT) of plasmid R388 was mobilized into recipient cells sort of ‘‘genome rebooting’’, a condition where the transcription-
by means of an oriT2 mutant of plasmid R388. Under these al/translational machinery is present, but the regulatory network is
conditions, plasmid R388 does not move itself, but is still able to transitory absent.
produce a conjugative pilus and thus mobilize the small vector into Simulations and theory also indicated that the overshoot is
the recipient cells. Results show that vector mobilization did not expected to be higher whenever the feedback loop has a high gain.
produce a significant decrease in the growth rate of transconjugant Plasmid promoters were shown to contain feedback loops with
cells (Supporting Figure S8). This indicated that the transitory characteristic high gains. RT-qPCR analysis showed a transcrip-
deficit in growth rate was not due to cell injuries produced by the tional burst in 5 out of 6 plasmid promoters subjected to NFLs,
mechanism of conjugation, but was a consequence of the entry of when the plasmid transferred horizontally into a new population
the conjugative plasmid inside the recipient cell. (Figure 5). Untranscribed regions (oriT), or plasmid genes that were
not under the control of a negative regulator (dhfR), did not show
Discussion similar increases (Figure 5). This indicates that the observed effect
is not due to conjugation increasing the abundance of plasmid
The intensity of HGT in the microbial world, and the molecules within the population. Moreover, given that the
prevalence of plasmids in nature indicate that plasmids are conjugative mix contained a 50% ratio of donor/recipient cells,
successful in colonizing microbial populations. Yet multilevel the maximal increase that conjugation could cause is 2-fold. The
Figure 6. Transconjugants experience a growth deficit immediately after conjugation. (A) Generation times (in minutes) of E.coli Bw27783
with or without plasmid R388. Results represent the average and standard deviation of 12 experiments. (B)Scheme showing the experimental design
to test the effect of plasmid conjugation in early transconjugants. Donor and recipient cells were grown in LB broth in the presence of selective
antibiotics and mixed in 1:1 conjugations on LB-Agar. Conjugation was allowed to take place for 30 minutes and cells were resuspended in fresh LB.
Plate counting was used to determine de number of donors (D) recipients (R) and transconjugants (Tc). (C) Results of the competition experiments
between D, R and Tc cells. (Left panel) Absolute numbers (cells/ml) of each species along the time course of the experiment. Each data point was
measured by triplicate. (Right panel) Proportion of plasmid containing cells that are transconjugants along the time course of the experiment. Since
plasmid R388 does not conjugate in liquid media, all variations in the relative proportions of Tc cells to D and R cells must be due to differences in
growth. Results show a decrease in the relative abundance of Tc cells compared to D cells, that recovers after t = 90 minutes. The lower bars indicate
the apparent generation times for each cell type calculated from data shown on the left panel.
doi:10.1371/journal.pgen.1004171.g006
PLOS Genetics | www.plosgenetics.org 10 February 2014 | Volume 10 | Issue 2 | e1004171
Negative Feedback in Horizontal Gene Transfer
increase of mRNA abundance was not due to cell growth either, compete with its neighbours [9], and thus the plasmid experiences
since results were normalized by the increase experienced by a lower vertical transmission rates. Although both selection processes
constitutively expressed chromosomal gene (dxs). Gene dxs showed are intrinsically in conflict, the timescales involved in each of them
a maximal increase of 2-fold, indicating that cell growth is a minor are different. The decrease on host fitness imposed by the plasmid
contributor to the observed bursts in mRNA levels. These results metabolic burden is usually low (% in the case of plasmid R388),
cannot be considered absolute quantifications of the transcrip- meaning that intercellular selection acts by the accumulation of
tional overshooting, because our measurements involved entire small fitness deficits over long periods of time [5,6,8]. On the other
populations (which contained donor and recipient cells), and hand, intracellular selection is more pronounced in the initial
bacterial conjugation is an asynchronous process. However, stages of infection, since a cell that has received the plasmid is still
although our quantitative results might be blurred by population susceptible to superinfection until the surface exclusion systems
effects, the general trend predicted by theory and simulations was have been deployed [9] [47]. Therefore, it is in the interest of the
sound: those promoted that showed higher gains also showed the first plasmid that enters into a cell to block the entrance of other
higher overshoots. plasmid copies, and to reach the steady-state copy number as soon
Plasmids are known to produce fitness deficits on their hosts. as possible. Transcriptional overshooting after HGT would allow
This effect has been usually ascribed to the metabolic burden the plasmid to produce a vigorous transcriptional response when
imposed by expression of plasmid genes. Therefore, any increase intracellular selection is more acute. The transient nature of this
in expression levels caused by transient overshooting might have response would guarantee that the long-term effects on intercel-
its counterpart in the growth rate of the host cell. We measured the lular selection are minimized. Indeed competition experiments
growth rates of newly formed transconjugants and found that the showed that, despite the severe initial effect on the host growth
plasmid induced a deficit that was transitorily high (250% increase rate, transconjugants recovered quickly and were able to achieve
in generation time), relaxing later to a 17% increase compared to the same number of cell divisions as the original donors. Note also
plasmid free cells. This was not caused by any physical damage that since transconjugants are able to act as donors, conjugation
produced by the mechanism of conjugation, and correlates in time results in an infectious process that proceeds geometrically in the
with the induction of plasmid genes after transfer. Altogether, population. If overexpression of conjugative functions results in
these results strongly suggest that overshooting after HGT has a increased transfer efficiency, a transient overshoot would provide
measurable impact on the host growth rate. Although this kind of the invading plasmids with higher infectivity. This property will be
effect has been traditionally ascribed to metabolic burden, it is also maintained as long as new cells are infected. If the availability of
possible that the toxic effects of specific plasmid proteins could possible receptors decreases, the overshoot transient nature
contribute. Since the growth deficit roughly corresponds to the guarantees that the plasmid population relaxes to a ‘‘silent’’ state,
time of overshoot decay (figure 5 and figure 6C) the most plausible minimizing the burden on the host and improving vertical
explanation is that growth deficit be caused by the transcription/ transmission. Such a mechanism would provide the plasmid
translation of the plasmid genes. This would also explain why, population with a mechanism to switch from horizontal to vertical
when the recipient cells recover, they grow as fast as recipient cells reproduction modes depending on the availability of susceptible
for a few generations. receptors. Other lines of evidence also point to this possibility. The
One intriguing question then is why has the plasmid evolved a stbABC operon of plasmid R388 has been shown to balance the
network based exclusively on NFLs, when this motif is likely to requirements for vegetative stability and conjugative transfer [23]
overshoot after conjugative transfer, temporarily hampering the The fact that transient overshooting is linked to genome
host growth rate? Other broad host range plasmids show rebooting is also interesting from a synthetic biology perspective.
convergent architectures, suggesting that despite this temporary Plasmids are nature counterparts of genomic transplantations. In
fitness deficit, negative feedback might have some adaptive fact, they can be considered as genetic devices for the unidirec-
property for the plasmid lifestyle. Indeed, negative feedback has tional injection of genomes into suitable recipient cells. So far,
been shown to exhibit a number of adaptive properties, speeding efforts to transplant whole chromosomes have been restricted to
up the response time of sensory regulatory networks [27], reducing species that share a high degree of genomic identity [48]. A close
transcriptional noise [42,43], driving noise to higher frequencies phylogenetic relationship implies that the regulatory networks of
and allowing easier filtering [44]. Speeding up the response is a both species might show some cross-reactivity, which could be
property associated to sensory systems, and so far the plasmid necessary to control the transplanted chromosome until it has built
network has not shown responses to any specific signals. Noise up its own regulatory system. Distantly related species, however,
control might be more interesting for plasmids, given that plasmid might show no cross-reactivity between their regulatory networks.
replication is extremely sensitive to fluctuations [45,46]. However Broad host conjugative plasmids are able to invade a wide variety
this problem is restricted mainly to replication, and does not of distantly related species. If we want to expand the range of
explain why the same regulatory strategy is widespread in the possible transplants, we need to deal with problems identical to
entire plasmid backbone. those faced by conjugative plasmids. In particular, how can a
It is also possible that transient overshooting provides an genome start up from just DNA and the transcriptional/
adaptive benefit for the particular lifestyle of conjugative plasmids. translational machinery? Negative regulation, with high feedback
Plasmids spread horizontally, by invading new cells, and vertically, gains and transient overshooting might be the solution evolved by
as the host cell reproduces. Like many other parasites that share natural plasmids.
this double reproductive strategy, plasmids suffer from opposing
selective forces, summarized in the observation that increased Materials and Methods
infectivity usually results in increased virulence. This inverse
relationship is well known in plasmids and phages [30,31,32], and Promoter library construction
if a given plasmid increases the expression levels of its own plasmid Strains used were Escherichia coli C41 (ompT hsdSB (rB2 mB2) gal
products (especially those that are cis-acting), it would also increase dcm (DE3)), E. coli Bw27783 (lacIqrrnB3 DlacZ4787 hsdR514
its intracellular fitness, at the cost of penalizing the host [9]. D(araBAD)567 D(rhaBAD)568 D(araFGH) W(DaraEp PCP8-araE))
Penalizing the host, in turn, decreases the ability of the host cell to [49] and E. coli JM109 (recA1, endA1, gyrA96, thi, hsdR17, supE44,
relA1, D(lac-proAB)/F9 [traD36, proAB+, lacIq, lacZDM15]) Primer harvested at OD600 = 0.5 and RNA was extracted using
oligonucleotides (Supporting table S2) were designed to flank each RNAEasy kit (Quiagen). Total RNA concentration was quantified
R388 intergenic region longer than 30 bp, according to R388 using Experion RNA StdSens Analysis kit (Biorad). cDNA was
genomic sequence (Genbank accession number BR000038) and obtained by reverse transcriptase (Omniscript, Qiagen) and the
purchased from Sigma. R388 plasmid DNA was extracted using relative concentration of the target genes was determined by
Sigma GenElute Miniprep kit and used as template for PCR qPCR (ICycler, Biorad) using the IQ SYBR Green Supermix kit
amplification. PCR amplification was carried out with Vent DNA (Biorad). To determine the cDNA abundance, the threshold cycle
polymerase (Biolabs) and consisted of 95uC for 10 min, then 28 (Ct) was determined using the ICycler software. A total of 66
cycles of 95uC for 30 s, 55uC for 30 s, 72uC for 30 s and a final primer pairs were manually designed to cover the entire genome of
step of 72uC for 5 min. PCR products were digested with XhoI and the plasmid, and the efficiency of each primer pair was determined
BamHI at 37uC for 2 h and the products purified using QIAquick measuring the Ct obtained from 3 different DNA concentrations
Gel Extraction Kit (Qiagen). Digested and purified fragments were (2.5, 5 and 10 ng). The sequence and efficiencies of each primer is
ligated into pUA66 plasmid DNA using T4 ligase (Roche) with shown in Supporting table S2. cDNA reactions were performed
overnight incubation at 16uC. Transformation was accomplished from 300 ng total RNA concentration and results were normalized
by electroporation into Bw27783 strain. Transformants were by the Ct obtained from 5 ng of plasmid DNA purified by alkaline
selected in LB-agar plates containing 25 mg/ml kanamycin. lysis.
Positive colonies were detected by PCR using pZE0.5 and
pZE0.6 primers. DNA from positive colonies was extracted and RT-qPCR measurements of gene expression during
insertions sequenced using the same primers as above. The set of conjugation
reporter plasmids obtained is indicated in Supporting table S2. To determine the relative expression of plasmid genes during
Plasmid R388 was transferred to the reporter strains by conjugation, 1 ml samples of donor (E.coli Bw27783 +R388
conjugation. Donor bacteria were E. coli JM109 containing plasmid) and recipient (E.coli Bw27783) cultures were mixed in a
plasmid R388 and recipient bacteria were E. coli BW27783 1:1 ratio, pelleted and resuspended in 100 l of fresh LB. Cells were
containing the corresponding reporter plasmid. Conjugations were then spread on LB-Agar and incubated at 37C for 30, 60 or
carried out as previously described [50]. 90 min. After each conjugation period, cells were resuspended in
2 ml of fresh LB and total RNA was extracted as described in the
Fluorescent expression profiling previous paragraph. Primer pairs used are shown in supporting
Transcriptional activity was determined by GFP expression table. S2 For each independent experiment measurements were
profiling as described in [51]. Reporter strains were inoculated performed in duplicate, and a total of 4 independent experiments
into enriched M9 Medium (M9 + casaminoacids 0.2%+ glycerol were performed for each time point. In order to avoid artefacts
0.5%) to which kanamycin (25 mg/ml) was added. To test the introduced by cell manipulation, the RNA concentration at time 0
effect of subinhibitory concentrations of antibiotics, we used was determined following the same manipulation procedure, but
rifampicin and chloramphenicol following the protocol described cells were immediately resuspended after being plated in LB-Agar.
in [52]. Results represent the average of at least 4 independent The relative concentration of target RNAs (r) was determined by
measurements. The S.O.S. response was induced exposing the r = E(DCt) where E is the efficiency of the primer pair, calculated as
cells to 5 or 10 seconds of UV light (254 nm, 15W). Mitomycin C in [21], and DCt = CtT = t2CtT = 0. Results were normalized to the
was used at a final concentration of 5 mg/ml. increase experienced by a chromosomal gene (dxs) using the DCt
method, where DCt = (CtT = t2CtT = 0)probe [21]. Relative error
Effects of regulatory proteins was propagated using the standard propagated error formula. For
Selected R388 ORFs containing potential transcriptional the relative amount of a test mRNA (Ct) with respect to a given
regulators were PCR amplified with primers indicated in table standard (Ctdxs) the aforementioned formula yields sx/,x.
S1. The resulting DNA segments were cloned in plasmid pET3a = ln(2) (var(Ct)+var(Ctdxs)22cov(Ct,Ctdxs))22 , where var stands for
(Addgene). The genetic manipulation techniques were the same as the variance and cov for the covariance. The statistical significance
those described above, except that NdeI and BamHI restriction was ascertained using a one handed t test.
endonucleases were used for cloning PCR products in pET3a
expression vector. After transforming to Escherichia coli C41 strain, Competition after conjugation experiments
protein expression was induced by adding 0.1 mM IPTG to Two spontaneous mutants resistant to rifampicin and nalidixic
exponentially growing cultures and visualized by denaturing acid from E.coli Bw27783 were obtained by plating in selective
polyacrylamide gel electrophoresis (data not shown). Then each antibiotics. The stability of the mutation was tested and the strains
gene was subcloned in plasmid pBAD33 using XbaI-Hind III were used as donor/recipients in conjugation experiments.
endonucleases. Plasmids obtained (table S2) were transformed to Growth rates were determined from cells growing in LB broth
E. coli Bw27783 containing the corresponding reporter plasmids at 37 C with agitation, and results showed that both strains had
for further analysis. To determine the effect of potential regulatory indistinguishable division times in such conditions (n = 12). In
proteins, pAR expression plasmids (Supporting table S2) were order to measure the growth rate of donor, recipient transconju-
transformed into E. coli Bw27783 containing the corresponding gant cells, we performed conjugations for t = 30 minutes. Saturat-
reporter plasmid. Protein expression was induced by adding ed cultures, grown overnight in LB at 37C, were and diluted 1/
appropriate concentrations of arabinose [51] to M9-broth and 1000 in fresh LB until cells reached an OD600 of 0.1. Donor and
fluorescence per OD unit (GFP/OD) was determined and recipient cells were mixed in a 1:1 ratio and concentrated 1000
compared to that produced by the same reporter strain when times. A total of 15 microliters were deposited onto a LB agar plate
containing the empty expression vector pBAD33. and let at 37C for 30 minutes to allow plasmid transfer. Cells were
then resuspended in 3 ml of LB broth and allowed to grow at 37C
Quantification of mRNA levels with agitation. Time samples were obtained every 30 min, and
Total RNA was extracted from E.coli Bw27783 containing cells were diluted in 16PBS to count the number of donor,
plasmid R388 and grown in LB media at 37C. Cells were recipients and transconjugants by plating in selective antibiotics.
Plating of early time points was performed 30 minutes after PBS expression profiles obtained with a co residing pAR13 vector
resuspension, to allow the antibiotic markers to express to (pBAD33::korA). Although PkorB fluorescence levels decreased in
adequate levels. We checked that this treatment did not artificially response to KorA the profile is not shown since the difference was
increased the number of cells due to growth in the PBS dilution. not statistically significant. C) Expression profiles of cultures
We plated dilutions from 1021 to 1026. The error introduced by containing Pint and Pant reporter vectors alone (black lines) and
the dilution was measured obtaining values typically around in the presence of R388 (red lines). Data shown represents the
cv = 0.1–0.2. This error was propagated to the actual number of average of at least four independent experiments.
cells and accounts for most of the variability observed in the (DOCX)
results.
Figure S3 Effects of sub-inhibitory concentrations of rifampicin
on plasmid promoters. (A) Expression profiles of plasmid R388
Computational analysis promoters, measured as described in Materials and Methods, in
Numerical integration of the ODE system was performed in the presence of rifampicin 3 mg/ml. Rifampicin produced a
Matlab (Mathworks) using the ODE23 algorithm. ODE23 is a general decrease in GFP/OD levels, either in the presence or the
Runge-Kutta algorithm with automatic step size. The absolute absence of plasmid R388. (B) Effect of rifampicin 3 mg/ml on
and relative tolerances were set at 10210 , tspan = 1000. bacterial growth rate. Growth curves were determined measuring
OD600 at different time points. The upper panel shows the
Supporting Information complete growth curve in a linear scale. The lower panel shows
the exponential growth phase in a semi-logarithmic scale. As
Figure S1 Expression profiles of the replication and mainte-
shown by the figure, rifampicin 3 mg/ml produced no detectable
nance promoters in the presence or absence of their transcriptional
effect on the growth rate, while the presence of plasmid R388
regulators. Each panel shows the expression profile of the reporter
decreased it significantly.
plasmid indicated above the panel (cloned promoter indicated in
(DOCX)
brackets). Expression profiles correspond to promoter alone (black
lines), in the presence of plasmid R388 (red lines), or when Figure S4 Effects of SOS response on plasmid promoters.
different regulators are expressed from a co residing pBAD33 Charts show the GFP/OD values achieved in steady-state by the
expression vector (blue and green lines). The effect of the promoters indicated in the figure. Expression profiling was
transcriptional regulators was tested without arabinose (darker performed as described in Materials and Methods. Cells were
lines, ara 2) and with maximum arabinose induction (lighter lines, treated with uv irradiation (254 nm, 15W) for 5 or 10 seconds.
ara +). Some transcriptional regulators were found to decrease the Mitomycin C was used at a concentration of 5 mg/ml. Those
growth rate when induced above a certain threshold. To discard promoters that were induced by SOS response were marked with
effects produced by impaired growth rate we measured, for each an asterisk (*). Pint showed a clear response to S.O.S induction
regulator, the rank of arabinose concentration that did not impair either by Mitomycin C or by UV irradiation. PtrwA showed a
bacterial growth (data not shown). Therefore maximum arabinose discrete 5 fold increase when the promoter was assayed alone, but
induction stands for the maximum concentration that did not that response could not be reproduced with co-residing plasmid
produce a measurable effect on growth rate, and it is variable for R388.
each regulator (ranging from 1023 to 1024% (w/v)). A) Expression (DOCX)
profiles from PresP and PkfrA promoters and response to ResP and
Figure S5 Temperature effects on plasmid promoters. Charts
KfrA respectively. B) Expression profiles from PardC, Porf7, Pssb,
show the GFP/OD values achieved in steady-state by the
Porf12, Porf14 and PstbA promoters and response to ArdK and
promoters indicated in the figure. Expression profiling was
StbA. Data shown represents the average of at least four
performed as described in Materials and Methods. Cells were
independent experiments.
grown at 37 C overnight, then diluted 1:10000 in fresh media, and
(DOCX)
then grown at the indicated temperatures.
Figure S2 Expression profiles of conjugation region and (DOCX)
response to their transcriptional regulators. Panels show the
Figure S6 Presence of potential receptors. Expression profiles of
expression profiles (obtained as in Materials and Methods) from
plasmid R388 promoters, obtained as described in Materials and
cultures containing the reporter plasmids indicated above each
Methods. The effect of potential recipients for horizontal transfer
panel (corresponding promoter indicated in brackets). Profiles
was tested by co-culture with empty E.coli Bw27783. Cells were
obtained with the reporter plasmid alone are indicated by black
mixed at 1:1 ratio before the measurement started. To obtain the
lines and by red lines when plasmid R388 was also present. Green
same amount of GFP-producing cells, the volume of recipient-
and blue lines indicate profiles obtained in the presence of a given
containing cultures was doubled. The only effect observed was a
regulator expressed from a co residing pBAD33 expression vector.
general decrease in fluorescence signal in those cultures that
The effect of the regulators was determined both with arabinose
contained recipients. Cell quenching probably caused this
induction (lighter lines, ara+) and without (darker lines, ara2). A)
unspecific effect.
Expression profiles of PtrwA containing reporter vector and
(DOCX)
response to R388 (red line) and TrwA (blue lines). B) Expression
profiles from reporter plasmids containing PtrwH, PkorA, PkikA Figure S7 Transient overshooting in StbA/KorA Incoherent
and PkorB and response to KorA and StbA transcriptional Feed Forward Loop (IFFL). In order to test whether the transient
regulators. Black lines represent expression profiles obtained from overshooting would also happen in more complex architectures
cultures containing the corresponding reporter vectors (indicated apart from simple NFLs, we simulated the behavior of the KorA-
above each panel) and red lines indicate the profiles of the same StbA IFFL loop present in the conjugation region (Upper panel).
reporter vector in the presence of a co residing R388. Green lines The parameters were introduced according to the results depicted
show the profile obtained when expression vector pAR12 in Table S1, which indicate the order of promoter strengths
(pBAD33::stbA) was present with (light green, ara+) and without (PstbA.PtrwH.PkorA) and indicated also the relative strengths
arabinose induction (dark green, ara2). Blue lines indicate the of repression exerted by the two regulators (KStbA_PkorA ..
KKorA_PkorA and KStbA_PtrwH..KKorA_PtrwH). Results shown in ratio and allowed to conjugate for 30 min at 37 C, on LB agar
the lower panel indicate that this IFFL architecture will also plates. Donors were E.coli Bw27783 Rifr containing plasmid
exhibit a transient overshoot. R388Dnic, and the mobilizable vector pSU4910 (Cmr). R388Dnic
(DOCX) encodes for the entire transfer system, but lacks the nic site needed
in cis for a DNA to be transferred by conjugation. Thus this strain
Figure S8 Growth rate deficit after horizontal gene transfer A.
is able to mobilize pSU4910 without transferring plasmid R388.
Growth rate after conjugation. (Left panel) Growth rate of
Recipients were E.coli Bw27783 Nxr. Experiments were performed
Recipients (R), Donors (D) and Transconjugants (T). Cells were
as in conjugation assays.
mixed at a 1:1 ratio and allowed to conjugate for 30 min. at 37 C,
(DOCX)
on LB agar plates. Donors were E.coli Bw27783 Rifr containing
plasmid R388, and recipients were E.coli Bw27783 Nxr. Cells Table S1 Promoter activities in the presence of plasmid R388
were then resuspended in liquid LB and allowed to grow. Cell and plasmid transcriptional regulators.
numbers were obtained by plating on appropriate antibiotic (DOCX)
combinations, as indicated in materials and methods. (Right
Table S2 Oligonucleotides and plasmids used in this study.
panel) Proportion of plasmid-containing cells that are transcon-
(XLSX)
jugants along time. The x axis indicates the timespan since cells
were taken out from conjugation mixtures. The y axis indicates the Text S1 Calculations on the Gain-Overshoot relationship.
proportion of transconjugants over plasmid-. containing cells (PDF)
(donors + transconjugants). Plasmid R388 does not conjugate in
liquid, thus any change in this proportion was due to growth Author Contributions
differences. Lower bars indicate the apparent generation times for Conceived and designed the experiments: RFL IdC CR AC FdlC.
each species, calculated from the data shown in the left panel. B. Performed the experiments: RFL IdC CR AC. Analyzed the data: RFL
Growth rate after mobilization. Growth rate of Recipients (R), IdC CR FdlC. Contributed reagents/materials/analysis tools: RFL IdC
Donors (D) and Transconjugants (T). Cells were mixed at a 1:1 CR FdlC. Wrote the paper: RFL FdlC.
References
1. Frost LS, Leplae R, Summers AO, Toussaint A (2005) Mobile genetic elements: 20. Revilla C, Garcillan-Barcia MP, Fernandez-Lopez R, Thomson NR, Sanders
the agents of open source evolution. Nat Rev Microbiol 3: 722–732. M, et al. (2008) Different pathways to acquiring resistance genes illustrated by
2. de la Cruz F, Davies J (2000) Horizontal gene transfer and the origin of species: the recent evolution of IncW plasmids. Antimicrob Agents Chemother 52:
lessons from bacteria. Trends Microbiol 8: 128–133. 1472–1480.
3. Fernandez-Lopez R, Garcillan-Barcia MP, Revilla C, Lazaro M, Vielva L, et al. 21. Pfaffl MW (2001) A new mathematical model for relative quantification in real-
(2006) Dynamics of the IncW genetic backbone imply general trends in time RT-PCR. Nucleic Acids Res 29: e45.
conjugative plasmid evolution. FEMS Microbiol Rev 30: 942–966. 22. Guynet C, de la Cruz F (2011) Plasmid segregation without partition. Mobile
4. Sen D, Brown CJ, Top EM, Sullivan J (2013) Inferring the Evolutionary History genetic elements 1: 236–241.
of IncP-1 Plasmids Despite Incongruence among Backbone Gene Trees. 23. Guynet C, Cuevas A, Moncalian G, de la Cruz F (2011) The stb operon
Molecular biology and evolution 30: 154–166. balances the requirements for vegetative stability and conjugative transfer of
5. Bouma JE, Lenski RE (1988) Evolution of a bacteria/plasmid association. plasmid R388. PLoS genetics 7: e1002073.
Nature 335: 351–352. 24. del Campo I, Ruiz R, Cuevas A, Revilla C, Vielva L, et al. (2012) Determination
6. Dionisio F, Conceicao IC, Marques AC, Fernandes L, Gordo I (2005) The of conjugation rates on solid surfaces. Plasmid 67: 174–182.
evolution of a conjugative plasmid and its ability to increase bacterial fitness. Biol 25. Waldor MK, Friedman DI (2005) Phage regulatory circuits and virulence gene
Lett 1: 250–252. expression. Current opinion in microbiology 8: 459–465.
7. De Gelder L, Ponciano JM, Joyce P, Top EM (2007) Stability of a promiscuous 26. Camas FM, Blazquez J, Poyatos JF (2006) Autogenous and nonautogenous control
plasmid in different hosts: no guarantee for a long-term relationship. of response in a genetic network. Proc Natl Acad Sci U S A 103: 12718–12723.
Microbiology 153: 452–463. 27. Rosenfeld N, Elowitz MB, Alon U (2002) Negative autoregulation speeds the
8. Haft RJ, Mittler JE, Traxler B (2009) Competition favours reduced cost of response times of transcription networks. J Mol Biol 323: 785–793.
plasmids to host bacteria. ISME J 3(7):761–9. 28. Schleif R (1988) DNA binding by proteins. Science 241: 1182–1187.
9. Paulsson J (2002) Multileveled selection on plasmid replication. Genetics 161: 29. Sarkar RR, Maithreye R, Sinha S (2011) Design of regulation and dynamics in
simple biochemical pathways. Journal of mathematical biology 63: 283–307.
1373–1384.
30. Levin BR, Stewart FM (1980) The population biology of bacterial plasmids: a
10. Thomas CM, Nielsen KM (2005) Mechanisms of, and barriers to, horizontal
priori conditions for the existence of mobilizable nonconjugative factors.
gene transfer between bacteria. Nat Rev Microbiol 3: 711–721.
Genetics 94: 425–443.
11. Bingle LE, Thomas CM (2001) Regulatory circuits for plasmid survival. Curr
31. Lundquist PD, Levin BR (1986) Transitory derepression and the maintenance of
Opin Microbiol 4: 194–200.
conjugative plasmids. Genetics 113: 483–497.
12. Polzleitner E, Zechner EL, Renner W, Fratte R, Jauk B, et al. (1997) TraM of 32. Stewart FM, Levin BR (1977) The Population Biology of Bacterial Plasmids: A
plasmid R1 controls transfer gene expression as an integrated control element in PRIORI Conditions for the Existence of Conjugationally Transmitted Factors.
a complex regulatory network. Mol Microbiol 25: 495–507. Genetics 87: 209–228.
13. Koraimann G, Teferle K, Markolin G, Woger W, Hogenauer G (1996) The 33. Shen-Orr SS, Milo R, Mangan S, Alon U (2002) Network motifs in the
FinOP repressor system of plasmid R1: analysis of the antisense RNA control of transcriptional regulation network of Escherichia coli. Nat Genet 31: 64–68.
traJ expression and conjugative DNA transfer. Mol Microbiol 21: 811–821. 34. Jagura-Burdzy G, Thomas CM (1992) kfrA gene of broad host range plasmid
14. Thomas CM (2006) Transcription regulatory circuits in bacterial plasmids. RK2 encodes a novel DNA-binding protein. J Mol Biol 225: 651–660.
Biochem Soc Trans 34: 1072–1074. 35. Jagura-Burdzy G, Thomas CM (1995) Purification of KorA protein from broad
15. Dunny GM, Johnson CM Regulatory circuits controlling enterococcal host range plasmid RK2: definition of a hierarchy of KorA operators. J Mol Biol
conjugation: lessons for functional genomics. Curr Opin Microbiol 14: 174–180. 253: 39–50.
16. Oger P, Kim KS, Sackett RL, Piper KR, Farrand SK (1998) Octopine-type Ti 36. Kostelidou K, Jones AC, Thomas CM (1999) Conserved C-terminal region of
plasmids code for a mannopine-inducible dominant-negative allele of traR, the global repressor KorA of broad-host-range plasmid RK2 is required for co-
quorum-sensing activator that regulates Ti plasmid conjugal transfer. Mol operativity between KorA and a second RK2 global regulator, KorB. J Mol Biol
Microbiol 27: 277–288. 289: 211–221.
17. McAnulla C, Edwards A, Sanchez-Contreras M, Sawers RG, Downie JA (2007) 37. Adamczyk M, Dolowy P, Jonczyk M, Thomas CM, Jagura-Burdzy G (2006)
Quorum-sensing-regulated transcriptional initiation of plasmid transfer and The kfrA gene is the first in a tricistronic operon required for survival of IncP-1
replication genes in Rhizobium leguminosarum biovar viciae. Microbiology 153: plasmid R751. Microbiology 152: 1621–1637.
2074–2082. 38. Bingle LE, Rajasekar KV, Muntaha S, Nadella V, Hyde EI, et al. (2008) A single
18. Suzuki H, Yano H, Brown CJ, Top EM (2010) Predicting plasmid promiscuity aromatic residue in transcriptional repressor protein KorA is critical for
based on genomic signature. J Bacteriol 192: 6045–6055. cooperativity with its co-regulator KorB. Mol Microbiol 70: 1502–1514.
19. Zaslaver A, Bren A, Ronen M, Itzkovitz S, Kikoin I, et al. (2006) A 39. Pansegrau W, Lanka E, Barth PT, Figurski DH, Guiney DG, et al. (1994)
comprehensive library of fluorescent transcriptional reporters for Escherichia Complete nucleotide sequence of Birmingham IncP alpha plasmids. Compila-
coli. Nat Methods 3: 623–628. tion and comparative analysis. J Mol Biol 239: 623–663.
40. Jacob F, Wollman EL (1956) [Processes of conjugation and recombination in 48. Lartigue C, Glass JI, Alperovich N, Pieper R, Parmar PP, et al. (2007) Genome
Escherichia coli. I. Induction by conjugation or zygotic induction]. Annales de transplantation in bacteria: changing one species to another. Science 317: 632–
l’Institut Pasteur 91: 486–510. 638.
41. Babic A, Lindner AB, Vulic M, Stewart EJ, Radman M (2008) Direct 49. Khlebnikov A, Datsenko KA, Skaug T, Wanner BL, Keasling JD (2001)
visualization of horizontal gene transfer. Science 319: 1533–1536. Homogeneous expression of the P(BAD) promoter in Escherichia coli by
42. Becskei A, Serrano L (2000) Engineering stability in gene networks by constitutive expression of the low-affinity high-capacity AraE transporter.
autoregulation. Nature 405: 590–593. Microbiology 147: 3241–3247.
43. Savageau MA (1974) Comparison of classical and autogenous systems of 50. Bolland S, Llosa M, Avila P, de la Cruz F (1990) General organization of the
regulation in inducible operons. Nature 252: 546–549. conjugal transfer genes of the IncW plasmid R388 and interactions between
44. Austin DW, Allen MS, McCollum JM, Dar RD, Wilgus JR, et al. (2006) Gene R388 and IncN and IncP plasmids. J Bacteriol 172: 5795–5802.
network shaping of inherent noise spectra. Nature 439: 608–611. 51. Fernandez-Lopez R, Del Campo I, Ruiz R, Lanza V, Vielva L, et al. Numbers
45. Das N, Valjavec-Gratian M, Basuray AN, Fekete RA, Papp PP, et al. (2005) on the edges: a simplified and scalable method for quantifying the gene
Multiple homeostatic mechanisms in the control of P1 plasmid replication. Proc regulation function. Bioessays 32: 346–355.
Natl Acad Sci U S A 102: 2856–2861.
52. Goh EB, Yim G, Tsui W, McClure J, Surette MG, et al. (2002) Transcriptional
46. Park K, Han E, Paulsson J, Chattoraj DK (2001) Origin pairing (‘handcuffing’)
modulation of bacterial gene expression by subinhibitory concentrations of
as a mode of negative control of P1 plasmid copy number. EMBO J 20: 7323–
antibiotics. Proceedings of the National Academy of Sciences of the United
7332.
States of America 99: 17025–17030.
47. Garcillan-Barcia MP, de la Cruz F (2008) Why is entry exclusion an essential
feature of conjugative plasmids? Plasmid 60: 1–18.