Applied Microbiology and Biotechnology (2019) 103:685–694

https://doi.org/10.1007/s00253-018-9527-y

MINI-REVIEW

Bacteriophage production processes

Katja Jurač 1 & Dominik Nabergoj 2 & Aleš Podgornik 1,2

Received: 13 November 2018 / Accepted: 16 November 2018 / Published online: 24 November 2018
# Springer-Verlag GmbH Germany, part of Springer Nature 2018

Abstract
High quantities of bacteriophages are currently used in the food industry and agriculture. However, growing antibiotic resistance
of bacteria has recently awakened the interest to use bacteriophages for the treatment of bacterial infections in humans indicating
that even higher quantities will be required in the future. High demand combined with a wide range of applications requires also
efficient bacteriophage production processes operating at low production costs and with high productivity. To achieve this goal,
different approaches were introduced and extensive studies of various parameters affecting bacteriophage formation were
investigated. In this mini-review, we provide a short overview about different operation modes of bacteriophage production
such as batch, semi-continuous and especially continuous with the pros and cons of each. We present factors affecting bacterial
physiological state, its effect on phage formation and provide a description of methods for determination of bacteriophage growth
parameters, through which bacteriophage formation is obtained. Understanding of described phenomena and inclusion of
potential occurrence of mutations and selection in continuous systems enables evaluation of continuous process productivity
and its optimization.

Keywords Bacteriophage production . Continuous bioprocess . Bacteriophage growth parameters . Bacteria physiological state

Introduction antibiotics. Since the resistance of bacteria to antibiotics has

nowadays become an alarming problem (Wittebole et al.
The ability of ultra-microscopic viruses to induce an effect on 2014; Lin et al. 2017), the interest for the use of phages has
bacteria was first mentioned in a paper by Frederick Twort increased and thereby also a need for the extensive production
(1915), which were later identified by Felix d’ Herelle of phages and their products like phage cocktails
(1917) as bacteriophages, viruses that infect only bacterial (Merabishvili et al. 2009). That the potential for phage treat-
cells. Later on, significant importance of bacteriophages ment of human infections is genuine, is reflected in the fact
(shorter phages) was recognized for a treatment of bacterial that many companies (AmpliPhi Biosciences, Pherecydes
infections in humans, called phage therapy. But the discovery Pharma, Technophage, Novolytics) are engaged in clinical
of antibiotics caused a decline in the development and use of studies using phages (Kingwell 2015; Sharma et al. 2017;
phages for medical purposes (Abedon et al. 2011; Ahmed Abedon 2018; AmpliPhi Biosciences Corporation 2018;
et al. 2012; Summers 2012; Kiros et al. 2016; Kolesnikova Pherecydes Pharma 2018). For example, Pherecydes Pharma
et al. 2017). Phage most important characteristics are that they focused on the use of their product named Phagoburn for the
can be highly specific to a particular bacterial strain, meaning treatment of infected wounds with Escherichia coli or
that phages have the ability to selectively kill only one patho- Pseudomonas aeruginosa. AmpliPhi Biosciences has been
gen bacterial strain while other symbiotic bacteria remain in- carrying out clinical studies on phage preparations for the
tact (Ross et al. 2016). In this way phages also differ from treatment of skin infections with Staphylococcus aureus
(Kingwell 2015; Sharma et al. 2017). While in the countries
of Poland, Russia, and Georgia phage therapy is routinely
* Aleš Podgornik
used for certain medical treatments already for decades, it is
[email protected]
expected, due to positive results of studies using phage cock-
1
Faculty of Chemistry and Chemical Technology, University of tails (Merabishvili et al. 2009; Wright et al. 2009; Fong et al.
Ljubljana, Večna pot 113, Ljubljana, Slovenia 2017; Schooley et al. 2017), that phage therapy will cover a
2
Center of Excellence for Biosensors, Instrumentation and Process wide range of treatments within the next 10 years also in other
Control – COBIK, Tovarniška cesta 26, Ajdovščina, Slovenia countries (Chan et al. 2013; Roach and Debarbieux 2017).
686 Appl Microbiol Biotechnol (2019) 103:685–694

Besides the use of phages for the treatment of bacterial Phage cultivation
infections in humans or for the treatment of bacterial infec-
tions in animals (Barrow 2001), phages can be implemented Cultivation of phages is traditionally performed in shaker
also in many other fields. The use of phages has emerged in flasks or in stirred tank bioreactors (Sargeant 1969).
agriculture against plant infections (Balogh et al. 2010), where Especially in a stirred tank bioreactor, phages can be produced
a product like AgriPhage™ is implemented (Monk et al. in different operation modes: batch, semi-continuous, and
2010). Phages have an important role in the food industry in continuous. A description of phage production using different
increasing safety and preventing contamination of food. For modes of operation and their advantages and disadvantages
example, phage products like LISTEX™ P100 (Chibeu et al. were recently presented by Agboluaje and Sauvageau (2018).
2013) and ListShield™ (Monk et al. 2010) are used to prevent In batch mode, high phage concentrations can be achieved,
the growth of Listeria monocytogenes while product but changes in the conditions during the process and especial-
EcoShield™ is used to prevent contamination with ly long preparation time, including cleaning, decrease its over-
Escherichia coli O157:H7 (Monk et al. 2010). Besides treat- all productivity (Agboluaje and Sauvageau 2018). This can be
ment of food products, phages are also used for control and overcome by the implementation of continuous production.
detection of different bacterial strains such as Campylobacter However, the process of switching to continuous biotechno-
jejuni (Hagens and Loessner 2007) and Salmonella through logical processes, especially in biopharmaceuticals produc-
products like SALMONELEX™ and BacWash™ (Ly- tion, is demanding due to an increased complexity
Chatain 2014). A similar application is extended also in clin- (Jungbauer 2013; Rathore 2015). But once implemented this
ical diagnostics for detection and typing of bacterial pathogens difficulty is compensated by higher productivity and reduced
(Schofield et al. 2012), where a system like Vidas UP® was costs. In addition, better quality of a product and its consis-
developed (Hagens and Loessner 2014). Phages are also in- tency is obtained, due to easier control (Rathore 2015).
teresting as potential carriers of vaccines (Haq et al. 2012). Following this objective, one of the approaches for obtaining
Detailed description of potential use and different phage ap- high final phage concentrations is a semi-continuous two-
plications is beyond the scope of this review, since extensive stage self-cycling process (Sauvageau and Cooper 2010). In
information can be found in many excellent reviews published this setup, bacteria are grown in the first bioreactor, operating
elsewhere (Marks and Sharp 2000; Matsuzaki et al. 2005; however in a batch mode. Before bacteria reach stationary
Clark and March 2006; Jones et al. 2007; García et al. 2008; phase, half of the bioreactor media containing bacteria is trans-
Letchumanan et al. 2016; Criscuolo et al. 2017). ferred into the second bioreactor. Their infection is initiated
Due to a widespread use of lytic phages and increasing with the portion of phages produced in the second bioreactor
industrial interest in the use of lysogenic phages (Yosef et al. and intentionally left from an earlier production cycle
2014; Harrison and Brockhurst 2017) further optimization of (Sauvageau and Cooper 2010; Agboluaje and Sauvageau
phage production processes is needed to increase productivity 2018). Significant increase in productivity is demonstrated,
and consequently decrease production costs. Significant diver- which could be further improved when bacterial cells are
sity of phage and bacteria systems (Clokie et al. 2011) dictates immobilized in alginate gel due to easier separation from the
a need for a specific in-depth research about production of each phages (Roy et al. 2018). However, the self-cycling system
individual system such as e.g. lytic phage T4 production requires a high level of control and monitoring (Sauvageau
(Smrekar et al. 2008; Sauvageau and Cooper 2010; Nabergoj and Cooper 2010; Agboluaje and Sauvageau 2018).
et al. 2018a), lysogenic phage production, such as lambda Operational complexity is significantly reduced
phage, where stress like the temperature shift, can cause a implementing real continuous phage production process omit-
transition between lysogenic to a lytic cycle (De Czekala ting cycles. The simplest approach consists of a single biore-
et al. 1972; Clark et al. 1986) and M13 filamentous phage, actor, also called chemostat. In such setup, there is a continu-
used in industrial applications due to its long-term viability ous flow of media through the bioreactor. Bacteria are intro-
and importance in phage display (Branston et al. 2011; duced to the system and allowed that steady state is achieved,
Warner et al. 2014). To optimize large-scale phage production, followed by an addition of phages. Ideally, robust steady state
it is important to understand the interactions between bacteria is achieved, continuously producing phages. However, long-
and phage and its reproduction process, which can be lytic, term continuous production of phages in a single-stage biore-
lysogenic, or in some cases non-lytic for filamentous phages. actor proved to be unfeasible, due to the long-term simulta-
In the present review, we try to give an overview about neous presence of phages and bacteria, resulting in spontane-
possible operation modes for phage production with a special ous mutations of both (Novick and Szilard 1950; Chao et al.
emphasis on continuous processes. Taking into account spec- 1977; Varon 1979; Lenski and Levin 1985; Spanakis and
ificity of bacteria–lytic phage interactions and especially the Horne 1987). The resistance of bacteria is caused by the loss
effect of bacteria physiological state on phage formation, the or modification of the specific phage receptors. Phages adapt
productivity of continuous systems is discussed. to bacterial mutations avoiding resistance. This dynamics is
Appl Microbiol Biotechnol (2019) 103:685–694 687

reflected initially in population oscillations that can later result connected bioreactors, again to prolong contact time and in-
in a stable state (Bohannan and Lenski 1997; Mizoguchi et al. crease the degree of infection (Mancuso et al. 2018).
2003). Therefore, chemostat is an effective tool to study co-
evolution of bacteria and phage system or coevolution of mul-
tiple phages and single bacteria. Coevolution, which enables Effect of bacteria growth conditions on phage
increased species diversity, is directly affected by cultivation production
conditions, for example, resource availability, which is in
chemostat controlled by dilution rate (Forde et al. 2008; To obtain optimal continuous production of phages, besides
Koskella and Brockhurst 2014). It was theoretically and ex- operation mode, also operating conditions (such as tempera-
perimentally demonstrated that in the chemostat, even under ture, pH, feed regime, bioreactor size, and configuration) af-
single substrate limiting conditions, multiple quasispecies of fecting bacterial growth environment and by this indirectly
bacteria and phages can coexist (Bohannan and Lenski 2000; phages formation, should be considered. Differences in phage
Weitz et al. 2005). Alternativelly, complex dynamics between concentration found under different growth conditions clearly
bacteria and phages in chemostat can result in bacterial resis- indicate that bacteria physiological state affects phage forma-
tance and extinction of phage as reviewed by Dennehy (2012). tion rate (Hadas et al. 1997). For this reason, the studies under
Since chemostat is a useful system to investigate resistance non-optimal bacterial growth conditions were performed
development in bacteria and phage, it can also serve for eval- (Hadas et al. 1997; You et al. 2002; Golec et al. 2014;
uation of phage therapy efficacy, where the appearance of Nabergoj et al. 2018b). Obtained results are not only impor-
resistant bacteria represents main drawback (Kunisaki and tant for the design of the production process, but also for
Tanji 2010). different phage applications. For example, in the food industry
Clearly, for a long-term robust phage production, mutations where phages are used for suppression of bacteria, changes in
have to be minimized. This was addressed by developing physiological states of bacteria due to environment conditions
multi-stage connected bioreactors, introduced already in (e.g., changes in temperature, substrate limitation) may lead to
1966 (Jacobson and Jacobson 1966). Setup consisted of a a reduced infection efficiency and consequently poor elimina-
bioreactor where bacteria are continuously growing. In the tion of contamination (Fister et al. 2016). Furthermore, it is
second bioreactor, the flow of bacteria from the first bioreactor essential to investigate the effect of media composition since
is mixed with a continuous flow of phages needed for infec- certain components such as salt concentration and type of
tion. The second bioreactor is connected to several consecu- ions, as well as presence of organic cofactors such as amino
tive bioreactors to ensure long enough residence time, what acid L-tryptophan affect phage adsorption (Ellis and Delbrück
allows all bacteria cells to lyse and produce phages, before 1939; Anderson 1949; Adams 1959; Tucker 1961; Storms
they are washed out of the system. Although the possibility et al. 2010; Mojica and Brussaard 2014). Media composition
of continuous phage production was successfully also significantly affects bacteria physiological state, includ-
demonstrated, the occurrence of resistant bacterial mutants ing cellular composition and metabolism and thereby also
was not completely eliminated. Husimi et al. (1982) presented phage formation (Adams 1959; Kutter et al. 1994; You et al.
a similar configuration with some important modifications 2002; Guttman et al. 2005; Choi et al. 2010; Taj et al. 2013;
resulting in the elimination of bacterial mutations. They de- Egli 2015). In the case of extreme substrate limitation, phage
signed a continuous system consisting of only two connected formation can be completely stopped, despite infection, as is
bioreactors called cellstat. In the first one, bacteria are contin- the case in hibernation mode (Bryan et al. 2016).
uously growing and pumped into the second bioreactor, where Due to a plethora of factors influencing bacterial physio-
phages are located to interact with them. Their system differs logical state in many different ways, making any general con-
from the previous configuration as there is no continuous flow clusion almost impossible, several studies focused on the in-
of phages into a cellstat. More importantly, the occurrence of fluence of bacteria physiological state on phage production,
bacterial mutations is overcome by the use of high dilution keeping all parameters but limiting component constant
rates in the second bioreactor, being above the maximal (Gáspár et al. 1981; Hadas et al. 1997; You et al. 2002;
growth rate of bacteria cultivated in the first bioreactor. This Nabergoj et al. 2018b). Gáspár et al. (1981) studied its influ-
can be achieved with lower volume in the second bioreactor in ence by determining the number of formed phages from bac-
comparison to the volume of the first bioreactor. In this way, terial cells withdrawn at different stages of the batch process,
bacteria that are generated in the second bioreactor and could especially during exponential and limitation phase of bacteria
mutate due to a phage presence, are washed out of the system growth. After withdrawal, bacterial samples were immediate-
and can not therefore accumulate. The phages achieve a ly infected with phages and the number of newly formed
steady-state, as they can compensate their outflow by much phages was determined. It was shown that samples withdrawn
faster formation (Wang 2006; Shao and Wang 2008; Lindberg during the exponential growth phase exhibited a similar num-
et al. 2014). The system can be further upgraded into three ber of newly formed phages, while their number was
688 Appl Microbiol Biotechnol (2019) 103:685–694

significantly lower when cells were in limitation growth obtained by fitting experimental data with a mathematical
phase. Furthermore, the study of phage formation with syn- model describing the change of concentration in continuous
chronized bacteria culture in exponential growth phase dem- stirred tank bioreactor after pulse injection (Levenspiel 1999).
onstrated differences even during the cell growth cycle. The Such procedure allows determination of phage growth param-
newly formed cells produced a lower number of phages than eters without affecting bacterial physiological state; therefore,
cells closer to a cell division (Storms et al. 2013). phage growth parameter values accurately reflect the effect of
Despite the fact that chemostat is not suitable for the robust bacteria physiological state and can be used for modeling and
production of phages due to mutations, it represents, besides optimization of continuous phage production.
for coevolutionary studies, also an excellent tool for investi-
gating the effect of physiological state of bacteria on phage
formation, as growth conditions can be precisely regulated
and kept constant over time (Kick et al. 2017). Different but Design of continuous phage production
constant bacterial physiological states in chemostat can be process
achieved by varying a dilution rate (You et al. 2002; Golec
et al. 2014; Nabergoj et al. 2018b), therefore its effect on To design continuous phage production process a mathemat-
phage formation can be studied in details. Formation of lytic ical description is required. The mathematical model enables
phages is most commonly described through three phage prediction of a system dynamics and influence of each specific
growth parameters: latent period, burst size, and adsorption or combination of process parameters. As such, a single model
constant (Adams 1959). Adsorption constant (δ) is a constant can be used for designing the process for different bioreactor
describing rate of adsorption of phages on bacterial cells; burst configurations, scale up and optimization. Accurate process
size (b) represents a number of phages formed from a single description requires a specific description of bacteria–phage
infected bacterial cell, while latent period (L) represents the dynamics. Because lytic phages are the most promising can-
time between infection and cell lysis. Their values are deter- didates for phage therapy (Abedon et al. 2011), many re-
mined through measurement of phage concentration as report- searchers have focused on mathematical modeling of their
ed elsewhere (Hyman and Abedon 2009). Briefly, the bacte- dynamics (Campbell 1961; Grossi et al. 1977; Lenski 1988;
rial sample is taken from a bioreactor and transferred into an Abedon et al. 2001; Santos et al. 2014; Storms and Sauvageau
Eppendorf tube, where cells are infected with phages. 2015; Krysiak-Baltyn et al. 2016) that includes adsorption of
Determination of adsorption constant is carried out by taking phages onto bacteria, followed by the injection of phage ge-
small samples from the Eppendorf tube in short time intervals. nome into bacteria cell. Phage genome begins to control bac-
The bacterial cells are removed and free phage concentration teria metabolism that leads to a synthesis of progeny phages.
is determined via plaque assay to determine its decrease over The final stage of infection is bacteria cells lysis, when new
time (Kropinski et al. 2009). For determination of latent period phages are released (Adams 1959). Commonly, mathematical
and burst size, so-called one-step growth experiment is per- models about predators and prey (Lotka 1925; Volterra 1926)
formed (Hyman and Abedon 2009). After infection, bacterial serve as a basis of such description. Campbell (1961) and
cells and phages are left mixed for a few minutes until adsorp- Levin et al. (1977) extended that model by inclusion of latent
tion is mostly completed. Free phages are removed from in- period and variations in bacteria concentration due to a limi-
fected bacterial cells by centrifugation, while infected cells are tation of resources. The mass balance equation of lytic phage
resuspended in fresh medium. Concentration of newly formed growth based on phage growth parameters, proposed original-
phages over time is determined, again via plaque assay. In ly for the continuous system by Levin et al. (1977), was mod-
chemostat experiments, infected cells are returned into a bio- ified by Bull (2006) to describe the change in phage concen-
reactor, after phage removal, in order to minimize the effect of tration (P) over time:
sampling on bacterial cells (Golec et al. 2014). Nevertheless, dP
since the determination of phage growth parameters requires ¼ −Pðm þ δX Þ þ bδX ∙PL e−dL ð1Þ
dt
the withdrawal of bacteria from the bioreactor, their physio-
logical state can be affected, and consequently, phage growth Equation 1 requires values of three phage growth parame-
parameter values might not reflect the real influence of bacte- ters (adsorption constant δ, burst size b and latent period L),
ria physiological state. An alternative experimental approach the intrinsic death rate of cells (d), the death rate of free phages
for determination of those parameters in a continuous process (m) and bacteria concentration (X). The first term of Eq. 1
was introduced by Gáspár et al. (1979). Once steady-state of B−P(m + δX)^ represents a disappearance of phages due to
bacteria is achieved, an injection of phages directly into the adsorption and cell death rate, while the second term BbδX ·
bioreactor is performed. Sampling from the bioreactor is car- PLe−dL^ represents the formation of new phages through cell
ried out at a certain time interval and free phage concentration lysis diminished for a portion of cells, which die during t-L
is determined. Values of phage growth parameters are time interval without phage release. Although phage
Appl Microbiol Biotechnol (2019) 103:685–694 689

formation is a discrete process, it can be approximated accu- introducing free phage concentration in terms of phage growth
rately over longer period of time by an exponential increase parameters (Bull et al. 2006):
that can be expressed according to Eq. 2 (Bull 2006): 
DP
P ¼ P0 ∙e λt
ð2Þ X∙ðbe−LDP −1Þ−
D2 δ
Pr ¼ P ∙   ð6Þ
δ D P
Equation 2 assumes constant concentration of bacteria, X∙ð1−e−LDP Þ þ
thus only the effect of phage multiplication without cell δ
growth is considered. Although phages have no metabolism,
where Dp is dilution rate in the second bioreactor of the
they are able to catalyze their own formation through the
cellstat. Dilution rate, which represents the bacterial physio-
growing bacteria, therefore their increase can be considered
logical state, affects phage growth parameters and conse-
as an autocatalytic reaction (Northrop 1938). According to
quently also cellstat productivity. The effect of dilution rate
this conclusion Eq. 2 represents an analogy to the growth of
in either first or second bioreactor was shown for the minia-
microorganisms, being a real autocatalytic process that can be
turized system (Nabergoj et al. 2018a) and for a system of
written as a reaction of a first-order kinetics (Eq. 3):
larger scale (Mancuso et al. 2018). Nabergoj et al. (2018a)
dX conducted experiments for two configurations of cellstat. In
¼ X ∙μ or X ¼ X 0 ∙eμt ð3Þ the first configuration, both bioreactor volumes were kept
dt
constant and dilution rate was changed by changing the flow
where μ [h−1] is a specific growth rate of microorganisms. rate through both bioreactors among experiments, therefore
Phage population growth rate λ therefore represents an analog the physiological state of bacteria in the first bioreactor was
to bacteria specific growth rate (Eq. 3). variable between experiments (Nabergoj et al. 2018a). In the
Combining Eqs. 1 and 2, one can correlate phage popula- second configuration, the flow rate through the system was
tion growth rate λ with phage growth parameters to obtain Eq. constant, while the volume of the second bioreactor was
4 (Bull 2006): changed between experiments, affecting only dilution rate in
  the second bioreactor, therefore having constant physiological
λ ¼ −m þ δX be−LðλþdÞ −1 ð4Þ state of bacteria in the first bioreactor (Nabergoj et al. 2018a).
Results showed that for each type of experiments an optimal
According to Eq. 4 phage population growth rate λ is dilution rate to obtain the highest productivity exists and that
completely determined by cell concentration, cell death rate, in configuration with constant bacteria physiological state
and four phage growth parameters (latent period, burst size, higher productivity is achieved (Fig. 1). Furthermore,
adsorption constant, and phage death rate). Assuming that in a Mancuso et al. (2018) investigated the effect of dilution rate
chemostat (especially at high dilution rates) intrinsic death rate in the first bioreactor, therefore changing physiological state
of free phages (m) and intrinsic death rate of bacterial cells (d) of bacteria, wherein the dilution rate in the second reactor was
are negligible, the Eq. 4 can be further simplified by setting kept constant. Similar trend of productivity and phage concen-
(m) and (d) to 0 (Podgornik et al. 2014). By developing cor- tration in dependence of dilution rate was demonstrated for
relations between phage growth parameters and dilution rate both miniaturized and larger system (Nabergoj et al. 2018a;
(D) in a chemostat (Nabergoj et al. 2018b), phage growth Mancuso et al. 2018). Configuration developed by Mancuso
parameters in Eq. 4 can be expressed in terms of dilution rate. and colleagues contained also the third bioreactor connected
Due to a rather complex structure of derived equation, original to the second one, what resulted in the higher phage concen-
expression was simplified by Eq. 5, to explicitly express the trations (Mancuso et al. 2018).
effect of dilution rate (therefore of bacteria physiological state) Although operation under optimal dilution rate would pro-
on phage population growth rate λ (Nabergoj et al. 2018b). vide maximal phage production, used mathematical models
λ ¼ λmax ∙D=ðK λ þ DÞ ð5Þ provide no information about the phage consistency since they
did not take into account potential occurrence of mutations.
Similarly, correlations connecting phage growth parame- While bacteria mutations can not accumulate in a cellstat as
ters and normalized bacteria specific growth rate (μ/μmax), dilution rate in the second bioreactor exceeds maximal specif-
were recently derived for a batch process (Krysiak-Baltyn ic bacteria growth rate (Husimi et al. 1982; Schwienhorst et al.
et al. 2018). 1996; Lindemann et al. 2002), this is not the case for phages.
Derived correlations can be used for determination of the Due to spontaneous mutations, certain heterogeneity might
system productivity. Productivity Pr of a continuous system is already be present in the initial phage population (Labedan
generally defined as a product of free phage concentration and and Legault-Demare 1974; Domingo et al. 1978; Storms and
dilution rate (Dunn et al. 1992). The productivity of cellstat Sauvageau 2014). Such subpopulations can differ in their af-
can be calculated from Eq. 6 (Podgornik et al. 2014), when finity toward bacterial host cells that can in extreme case even
690 Appl Microbiol Biotechnol (2019) 103:685–694

mathematically described by parameter termed adsorption ef-

12.00 ficiency (Storms et al. 2010). Initial phage population hetero-
geneity can be the reason for the experimentally observed
Phage concentration (108 PFU mL-1) and

occurrence of mutations in cellstat (Husimi et al. 1982;

10.00
Husimi 1989; Kong and Yin 1995), besides mutations that
productivity (108 PFU mL-1 h-1)

can develop during continuous bioprocess run. In both

8.00 cases, due to a selection pressure, selection among phage
strains occurs during bioprocess run, resulting in the domi-
6.00
nance of one specific strain (Husimi 1989; Husimi et al.
1993; Aita and Husimi 1994; Bull et al. 2006). Several,
mainly theoretical studies were performed to investigate
4.00
the possible outcomes of selection between strains. Aita
and Husimi (1994) presented the selection dynamics of
2.00 two non-lytic filamentous phage strains. Bacteria growth as
well as phage growth was assumed to be dependent on
temperature, being however different for each phage strain.
0.00
0 1 2 3 4 5 The temperature was changing periodically over time until
Dilution rate in bioreactor "P" (h-1)
steady-state was achieved. Simulation results presented in a
phase diagram (Fig. 2) demonstrated that pure single phage
Fig. 1 Change of cellstat productivity of phages (black dots) in
comparison with theoretical data (black solid line) and change of
population or their combination can occur, depending on the
concentration of phages (black triangles) in comparison with theoretical temperature oscillation frequency (τ) and dilution rate (Dc)
data (black dashed line) as a function of dilution rate for constant bacterial (Aita and Husimi 1994). Only phage P1 is present in the
physiological state. Change of cellstat productivity of phages (gray cellstat at higher dilution rates and short time of oscillation,
square) in comparison with theoretical data (black dashed-dotted line)
and change of concentration of phages (gray diamond) in comparison
while phage P2 remains in the cellstat for longer oscillation
with theoretical data (black dotted line) as a function of dilution rate for periods and lower dilution rates. At the intermediate condi-
variable bacterial physiological state. Adapted with permission from tions, coexistence of phages P1 and P2 is formed. As dem-
Springer Nature Customer Service Centre GmbH: [Springer onstrated also in Fig. 2, the combination of phage popula-
Nature][Applied Microbiology and Biotechnology][Effect of dilution
rate on productivity of continuous bacteriophage production in cellstat,
tion, which is achieved, depends on experimental conditions
Nabergoj D, Kuzmić N, Drakslar B, Podgornik A], [Copyright] (2018a) and properties of populations. These numerical results dem-
onstrate that phage selection can be driven by experimental
conditions, which can therefore be a powerful tool for phage
be absent (resulting in bacteria resistance toward specific production robustness but also Bcreation^ of phages of de-
phage subpopulation). Effect of different phage affinity was sired properties.

Fig. 2 Phase diagram of the steady-state phage population consisting of boundaries of the steady-state population to washout state in
two phage strains (P1 and P2) in cellstat. Abscissa: time of temperature dependence of the volume (Vc): (a) Vc = 20 ml and (b) Vc = 6 × 1018 ml.
oscillation τ and ordinate: dilution rate Dc. Boundaries between phases of Reprinted from [Journal of Theoretical Biology 168, Aita T, Husimi Y,
pure single phage population are divided with solid lines. Dashed-dotted Period dependent selection in continuous culture of viruses in a periodic
line presents the coexistence phase. Dashed lines (a) and (b) show environment, 281–289, Copyright (1994) with permission from Elsevier
Appl Microbiol Biotechnol (2019) 103:685–694 691

Besides temperature, other limiting resources can also quantities, as well as to drive phage mutations to evolve
serve to induce selective pressure. Effect of cultivating condi- phages being able to kill specific bacterial strain.
tions on the selection process between two lytic phages was
investigated by Bull et al. (2006). In cellstat, the change in the Funding The study was funded by the European Regional Development
Fund and Slovenian Ministry of Education, Science and Sport (project
dilution rate results in different phage concentrations and con-
BioPharm.Si).
sequently, the different ratio between bacteria and phages (dif-
ferent multiplicity of infection—MOI). At high phage concen-
Compliance with ethical standardsThis article does not con-
tration, bacteria become a limiting factor, which results in a tain any studies with human participants or animals performed by any of
competitive advantage of faster adsorbing phage, even in the the authors.
case of superinfection. As a selection criterion, a product of
free phage concentration and adsorption constant, which rep- Conflict of interest The authors declare that they have no conflict of
resents the rate of infection (ROI) of bacteria with bacterio- interest.
phage, was used. Because free phage concentration depends
on phage growth parameters, changing these parameters af-
fects ROI and thus also selection process. Simulations dem- References
onstrated that there exist many values of phage parameters, for
Abedon ST (2018) Phage Companies. http://companies.phage.org/.
which both phages can coexist, as well as conditions for which Accessed 14 June 2018
only target phage will prevail in the population (Bull et al. Abedon ST, Herschler TD, Stopar D (2001) Bacteriophage latent-period
2006; Podgornik et al. 2014). As phage growth parameters evolution as a response to resource availability. Appl Environ
are significantly affected by growing conditions, precise con- Microbiol 67:4233–4241
trol and understanding of phage properties enable to establish Abedon ST, Kuhl SJ, Blasdel BG, Kutter EM (2011) Phage treatment of
human infections. Bacteriophage 1:66–85. https://doi.org/10.4161/
conditions optimal for target phage formation. Despite that, as bact.1.2.15845
it is not possible to predict all types of mutations, it is neces- Adams MH (1959) Bacteriophages. Interscience publishers, Inc., New
sary to constantly control any changes in the infection effi- York
ciency of newly produced phages on fresh bacterial culture Agboluaje M, Sauvageau D (2018) Bacteriophage production in bioreac-
grown overnight (Jensen et al. 2015; Krysiak-Baltyn et al. tors. In: Azeredo J, Sillankorva S (eds) Bacteriophage therapy.
Methods in molecular biology, vol 1693. Humana Press, New
2018). In this way, the process will be stopped, if there occur York, pp 173–193. doi: https://doi.org/10.1007/978-1-4939-7395-
any unexpected changes in infection efficiency, which indi- 8_15
cate changes in the phage population. Before starting the pro- Ahmed K, Kaderbhai NN, Kaderbhai MA (2012) Bacteriophage therapy
cess again, the bioreactor must be emptied and thoroughly revisited. African J Microbiol Res 6:3366–3379. https://doi.org/10.
5897/AJMR11.968
cleaned to avoid any cross-contamination (Krysiak-Baltyn
Aita T, Husimi Y (1994) Period dependent selection in continuous culture
et al. 2018). of viruses in a periodic environment. J Theor Biol 168:281–289.
Cellstat, therefore, represents a powerfull tool for the pro- https://doi.org/10.1006/jtbi.1994.1109
duction of phages. While under certain growth conditions AmpliPhi Biosciences Corporation (2018) http://www.ampliphibio.com/
phage mutations can be minimized, other conditions favorite about-us/. Accessed 28 June 2018
prevalence of novel phage strains. As such cellstat can repre- Anderson TF (1949) The reactions of bacterial viruses with their host
cells. Bot Rev 15:464–505. https://doi.org/10.1007/BF02861677
sent a system for Bcreation^ of phages with required proper- Balogh B, Jones JB, Iriarte FB, Momol MT (2010) Phage therapy for
ties. By combining in-depth knowledge about the dynamics of plant disease control. Curr Pharm Biotechnol 11:48–57
specific bacteria–phage system with engineering parameters Barrow PA (2001) The use of bacteriophages for treatment and preven-
providing a proper controlled environment in a bioreactor, a tion of bacterial disease in animals and animal models of human
infection. J Chem Technol Biotechnol 76:677–682. https://doi.org/
long-term continuous production of phages seems to be
10.1002/jctb.436
feasible. Bohannan BJM, Lenski RE (1997) Effect of resource enrichment on a
chemostat community of bacteria and bacteriophage. Ecology 78:
2303–2315
Conclusion Bohannan BJM, Lenski RE (2000) The relative importance of competi-
tion and predation varies with productivity in a model community.
Am Nat 156:329–340. https://doi.org/10.1086/303393
The high amount of various phages requires a shift from tra- Branston S, Stanley E, Ward J, Keshavarz-Moore E (2011) Study of
ditional preparation processes toward optimized, preferably robustness of filamentous bacteriophages for industrial applications.
continuous operations. Understanding the interaction of spe- Biotechnol Bioeng 108:1468–1472. https://doi.org/10.1002/bit.
cific bacteria–phage system enables to design continuous pro- 23066
Bryan D, El-Shibiny A, Hobbs Z, Porter J, Kutter EM (2016)
cesses for phage production operating at maximal productiv- Bacteriophage T4 infection of stationary phase E. coli: life after
ity. One can envision that such processes can become a pro- log from a phage perspective. Front Microbiol 7:1391. https://doi.
cess of choice for robust production of target phages in large org/10.3389/fmicb.2016.01391
692 Appl Microbiol Biotechnol (2019) 103:685–694

Bull JJ (2006) Optimality models of phage life history and parallels in in a bacteria-bacteriophage interaction. Evolution 62:1830–1839.
disease evolution. J Theor Biol 241:928–938. https://doi.org/10. https://doi.org/10.1111/j.1558-5646.2008.00411.x
1016/j.jtbi.2006.01.027 García P, Martínez B, Obeso JM, Rodríguez A (2008) Bacteriophages
Bull JJ, Millstein J, Orcutt J, Wichman HA (2006) Evolutionary feedback and their application in food safety. Lett Appl Microbiol 47:479–
mediated through population density, illustrated with viruses in 485. https://doi.org/10.1111/j.1472-765X.2008.02458.x
chemostats. Am Nat 167:E39–E51. https://doi.org/10.1086/499374 Gáspár S, Rontó G, Müller G (1979) Determination of the biological
Campbell A (1961) Conditions for the existence of bacteriophage. parameters of bacterium-phage complexes. Z Allg Mikrobiol 19:
Evolution 15:153–165. https://doi.org/10.2307/2406076 163–169
Chan BK, Abedon ST, Loc-Carrillo C (2013) Phage cocktails and the Gáspár S, Módos K, Gy R (1981) Complex method for the determinaton
future of phage therapy. Future Microbiol 8:769–783. https://doi. of the physiological parameters of bacterium-phage systems. In:
org/10.2217/fmb.13.47 Fedina L, Kanyár B, Kocsis B, Kollai M (eds) Mathematical and
Chao L, Levin BR, Stewart FM (1977) A complex community in a simple Computational methods in Physiology. Pergamon Press, Akadémiai
habitat: an experimental study with bacteria and phage. Ecology 58: Kiadó, Hungary, pp 141–146
369–378. https://doi.org/10.2307/1935611 Golec P, Karczewska-Golec J, Łoś M, Węgrzyn G (2014) Bacteriophage
Chibeu A, Agius L, Gao A, Sabour PM, Kropinski AM, Balamurugan S T4 can produce progeny virions in extremely slowly growing
(2013) Efficacy of bacteriophage LISTEX™P100 combined with Escherichia coli host: comparison of a mathematical model with
chemical antimicrobials in reducing Listeria monocytogenes in the experimental data. FEMS Microbiol Lett 351:156–161. https://
cooked turkey and roast beef. Int J Food Microbiol 167:208–214. doi.org/10.1111/1574-6968.12372
https://doi.org/10.1016/j.ijfoodmicro.2013.08.018 Grossi GF, Cesareni G, Liello F (1977) Development of phage popula-
Choi C, Kuatsjah E, Wu E, Yuan S (2010) The effect of cell size on the tions in a bacterial culture: a mathematical model. Z Naturforsch C
burst size of T4 bacteriophage infections of Escherichia coli B23. J 32:844–849
Exp Microbiol Immunol 14:85–91 Guttman B, Raya R, Kutter E (2005) Basic phage biology. In: Kutter E,
Clark JR, March JB (2006) Bacteriophages and biotechnology: vaccines, Sulakvelidze A (eds) Bacteriophages: biology and applications.
gene therapy and antibacterials. Trends Biotechnol 24:212–218. CRC Press, Boca Raton, pp 29–66
https://doi.org/10.1016/j.tibtech.2006.03.003 Hadas H, Einav M, Fishov I, Zaritsky A (1997) Bacteriophage T4 devel-
Clark DW, Meyer H-P, Leist C, Fiechter A (1986) Effects of growth opment depends on the physiology of its host Escherichia coli.
medium on phage production and induction in Escherichia coli K- Microbiology 143:179–185. https://doi.org/10.1099/00221287-
12 lambda lysogens. J Biotechnol 3:271–280. https://doi.org/10. 143-1-179
1016/0168-1656(86)90009-X Hagens S, Loessner MJ (2007) Application of bacteriophages for detec-
tion and control of foodborne pathogens. Appl Microbiol Biotechnol
Clokie MRJ, Millard AD, Letarov AV, Heaphy S (2011) Phages in nature.
76:513–519. https://doi.org/10.1007/s00253-007-1031-8
Bacteriophage 1:31–45. https://doi.org/10.4161/bact.1.1.14942
Hagens S, Loessner MJ (2014) Phages of Listeria offer novel tools for
Criscuolo E, Spadini S, Lamanna J, Ferro M, Burioni R (2017)
diagnostics and biocontrol. Front Microbiol 5:159. https://doi.org/
Bacteriophages and their immunological applications against infec-
10.3389/fmicb.2014.00159
tious threats. J Immunol Res 2017:1–13. https://doi.org/10.1155/
Haq IU, Chaudhry WN, Akhtar MN, Andleeb S, Qadri I (2012)
2017/3780697
Bacteriophages and their implications on future biotechnology: a
De Czekala A, Luk D, Bartl P (1972) Large-scale production of lambda
review. Virol J 9:9. https://doi.org/10.1186/1743-422X-9-9
bacteriophage and purified lambda deoxyribonucleic acid. Appl
Harrison E, Brockhurst MA (2017) Ecological and evolutionary benefits
Microbiol 23:791–795
of temperate phage: what does or doesn’t kill you makes you stron-
Dennehy JJ (2012) What can phages tell us about host-pathogen coevo- ger. BioEssays 39. https://doi.org/10.1002/bies.201700112
lution? Int J Evol Biol 2012:1–12. https://doi.org/10.1155/2012/ Husimi Y (1989) Selection and evolution of bacteriophages in cellstat.
396165 Adv Biophys 25:1–43
d'Herelle F (1917) An invisible microbe that is antagonistic to the dysen- Husimi Y, Nishigaki K, Kinoshita Y, Tanaka T (1982) Cellstat-a contin-
tery bacillus. C R Acad Sci 165:373–375 uous culture system of a bacteriophage for the study of the mutation
Domingo E, Sabo D, Taniguchi T, Weissmann C (1978) Nucleotide se- rate and the selection process of the DNA level. Rev Sci Instrum 53:
quence heterogeneity of an RNA phage population. Cell 13:735– 517–522
744 Husimi Y, Mori T, Ushijima S, Aita T (1993) Modes of population
Dunn IJ, Heinzle E, Přenosil JE, Ingham J (1992) Biological reaction changeovers in continuous culture of bacterial and viral mutants
engineering : principles, applications and modelling with PC simu- spectrum. J Theor Biol 165:577–591. https://doi.org/10.1006/jtbi.
lation. Wiley-VCH, Weinheim 1993.1208
Egli T (2015) Microbial growth and physiology: a call for better crafts- Hyman P, Abedon ST (2009) Practical methods for determining phage
manship. Front Microbiol 6:287. https://doi.org/10.3389/fmicb. growth parameters. In: Clokie MRJ, Kropinski AM (eds)
2015.00287 Bacteriophages: methods and protocols, Vol 1: isolation, character-
Ellis EL, Delbrück M (1939) The growth of bacteriophage. J Gen Physiol ization, and interactions. Humana Press, pp 175–202
22:365–384 Jacobson H, Jacobson LS (1966) Virustat, a device for continuous pro-
Fister S, Robben C, Witte AK, Schoder D, Wagner M, Rossmanith P duction of viruses. Appl Microbiol 14:940–952
(2016) Influence of environmental factors on phage–bacteria inter- Jensen KC, Hair BB, Wienclaw TM, Murdock MH, Hatch JB, Trent AT,
action and on the efficacy and infectivity of phage P100. Front White TD, Haskell KJ, Berges BK (2015) Isolation and host range
Microbiol 7:1152. https://doi.org/10.3389/fmicb.2016.01152 of bacteriophage with lytic activity against methicillin-resistant
Fong SA, Drilling A, Morales S, Cornet ME, Woodworth BA, Fokkens Staphylococcus aureus and potential use as a fomite decontaminant.
WJ, Psaltis AJ, Vreugde S, Wormald P-J (2017) Activity of bacte- PLoS One 10:e0131714. https://doi.org/10.1371/journal.pone.
riophages in removing biofilms of Pseudomonas aeruginosa iso- 0131714
lates from chronic rhinosinusitis patients. Front Cell Infect Jones JB, Jackson LE, Balogh B, Obradovic A, Iriarte FB, Momol MT
Microbiol 7:418. https://doi.org/10.3389/fcimb.2017.00418 (2007) Bacteriophages for plant disease control. Annu Rev
Forde SE, Thompson JN, Holt RD, Bohannan BJM (2008) Coevolution Phytopathol 45:245–262. https://doi.org/10.1146/annurev.phyto.
drives temporal changes in fitness and diversity across environments 45.062806.094411
Appl Microbiol Biotechnol (2019) 103:685–694 693

Jungbauer A (2013) Continuous downstream processing of Lindberg HM, Mckean KA, Wang I-N (2014) Phage fitness may help
biopharmaceuticals. Trends Biotechnol 31:479–492. https://doi. predict phage therapy efficacy. Bacteriophage 4:e964081. https://
org/10.1016/j.tibtech.2013.05.011 doi.org/10.4161/21597073.2014.964081
Kick B, Behler KL, Severin TS, Weuster-Botz D (2017) Chemostat stud- Lindemann BF, Klug C, Schwienhorst A (2002) Evolution of bacterio-
ies of bacteriophage M13 infected Escherichia coli JM109 for con- phage in continuous culture: a model system to test antiviral gene
tinuous ssDNA production. J Biotechnol 258:92–100. https://doi. therapies for the emergence of phage escape mutants. J Virol 76:
org/10.1016/j.jbiotec.2017.06.409 5784–5792. https://doi.org/10.1128/JVI.76.11.5784-5792.2002
Kingwell K (2015) Bacteriophage therapies re-enter clinical trials. Nat Lotka AJ (1925) Elements of physical biology. Williams and Wilkins,
Rev Drug Discov 14:515–516. https://doi.org/10.1038/nrd4695 Baltimore
Kiros A, Gashaw T, Teshale A (2016) Phage therapy; a review on the Ly-Chatain MH (2014) The factors affecting effectiveness of treatment in
biology and therapeutic application of bacteriophage. ARC J Anim phages therapy. Front Microbiol 5. https://doi.org/10.3389/fmicb.
Vet Sci 2:15–25. https://doi.org/10.20431/2455-2518.0204003 2014.00051
Kolesnikova SG, Tulyakova EN, Moiseeva IY (2017) Phage therapy: Mancuso F, Shi J, Malik D (2018) High throughput manufacturing of
present and future. J Phys Conf Ser 784:012027. https://doi.org/ bacteriophages using continuous stirred tank bioreactors connected
10.1088/1742-6596/784/1/012027 in series to ensure optimum host bacteria physiology for phage pro-
Kong D, Yin J (1995) Whole-virus vaccine development by continuous duction. Viruses 10:537. https://doi.org/10.3390/v10100537
culture on a complementing host. Biotechnology (NY) 13:583–558 Marks T, Sharp R (2000) Bacteriophages and biotechnology: a review. J
Koskella B, Brockhurst MA (2014) Bacteria-phage coevolution as a driv- Chem Technol Biotechnol 75:6–17. https://doi.org/10.1002/(SICI)
er of ecological and evolutionary processes in microbial communi- 1097-4660(200001)75:1<6::AID-JCTB157>3.0.CO;2-A
ties. FEMS Microbiol Rev 38:916–931. https://doi.org/10.1111/ Matsuzaki S, Rashel M, Uchiyama J, Sakurai S, Ujihara T, Kuroda M,
1574-6976.12072 Imai S, Ikeuchi M, Tani T, Fujieda M, Wakiguchi H (2005)
Kropinski AM, Mazzocco A, Waddell TE, Lingohr E, Johnson RP (2009) Bacteriophage therapy: a revitalized therapy against bacterial infec-
Enumeration of bacteriophages by double agar overlay plaque assay. tious diseases. J Infect Chemother 11:211–219. https://doi.org/10.
In: Clokie MRJ, Kropinski AM (eds) Bacteriophages: methods and 1007/s10156-005-0408-9
protocols, vol 1: isolation, characterization, and interactions. Merabishvili M, Pirnay JP, Verbeken G, Chanishvili N, Tediashvili M,
Humana Press, pp 69–76 Lashkhi N, Glonti T, Krylov V, Mast J, Van Parys L, Lavigne R,
Volckaert G, Mattheus W, Verween G, De Corte P, Rose T, Jennes S,
Krysiak-Baltyn K, Martin GJO, Stickland AD, Scales PJ, Gras SL (2016)
Zizi M, De Vos D, Vaneechoutte M (2009) Quality-controlled small-
Computational models of populations of bacteria and lytic phage.
scale production of a well-defined bacteriophage cocktail for use in
Crit Rev Microbiol 42:942–968. https://doi.org/10.3109/1040841X.
human clinical trials. PLoS One 4:e4944. https://doi.org/10.1371/
2015.1114466
journal.pone.0004944
Krysiak-Baltyn K, Martin GJO, Gras SL (2018) Computational model-
Mizoguchi K, Morita M, Fischer CR, Yoichi M, Tanji Y, Unno H (2003)
ling of large scale phage production using a two-stage batch process.
Coevolution of bacteriophage PP01 and Escherichia coli O157:H7
Pharmaceuticals (Basel) 11. https://doi.org/10.3390/ph11020031
in continuous culture. Appl Environ Microbiol 69:170–176. https://
Kunisaki H, Tanji Y (2010) Intercrossing of phage genomes in a phage
doi.org/10.1128/AEM.69.1.170-176.2003
cocktail and stable coexistence with Escherichia coli O157:H7 in
Mojica KDA, Brussaard CPD (2014) Factors affecting virus dynamics
anaerobic continuous culture. Appl Microbiol Biotechnol 85:1533–
and microbial host - virus interactions in marine environments.
1540. https://doi.org/10.1007/s00253-009-2230-2
FEMS Microbiol Ecol 89:495–515. https://doi.org/10.1111/1574-
Kutter E, Kellenberger E, Carlson K, Eddy S, Neitzel J, Messinger L, 6941.12343
North J, Guttman B (1994) Effects of bacterial growth conditions Monk AB, Rees CD, Barrow P, Hagens S, Harper DR (2010)
and physiology on T4 infection. In: Karam JD (ed) Molecular biol- Bacteriophage applications: where are we now? Lett Appl
ogy of bacteriophage T4. American Society for Microbiology, Microbiol 51:363–369. https://doi.org/10.1111/j.1472-765X.2010.
Washington, DC, pp 406–418 02916.x
Labedan B, Legault-Demare J (1974) Evidence for heterogeneity in pop- Nabergoj D, Kuzmić N, Drakslar B, Podgornik A (2018a) Effect of dilu-
ulations of T5 bacteriophage. J Virol 13:1093–1100 tion rate on productivity of continuous bacteriophage production in
Lenski RE (1988) Dynamics of interactions between bacteria and virulent cellstat. Appl Microbiol Biotechnol 102:3649–3661. https://doi.org/
bacteriophage. In: Marshall KC (ed) Advances in microbial ecology. 10.1007/s00253-018-8893-9
Plenum Press, New York, pp 1–44. https://doi.org/10.1007/978-1- Nabergoj D, Modic P, Podgornik A (2018b) Effect of bacterial growth
4684-5409-3_1 rate on bacteriophage population growth rate. Microbiologyopen 7:
Lenski RE, Levin BR (1985) Constraints on the coevolution of bacteria e00558. https://doi.org/10.1002/mbo3.558
and virulent phage: a model, some experiments, and predictions for Northrop JH (1938) Concentration and purification of bacteriophage. J
natural communities. Am Nat 125:585–602. https://doi.org/10. Gen Physiol 21:335–366
2307/2461275 Novick A, Szilard L (1950) Experiments with the chemostat on sponta-
Letchumanan V, Chan KG, Pusparajah P, Saokaew S, Duangjai A, Goh neous mutations of bacteria. Proc Natl Acad Sci U S A 36:708–719
BH, Ab Mutalib NS, Lee LH (2016) Insights into bacteriophage Pherecydes Pharma (2018) https://www.pherecydes-pharma.com/.
application in controlling Vibrio species. Front Microbiol 7:1114. Accessed 28 June 2018
https://doi.org/10.3389/fmicb.2016.01114 Podgornik A, Janež N, Smrekar F, Peterka M (2014) Continuous produc-
Levenspiel O (1999) Chemical reaction engineering, 3rd edn. Wiley, New tion of bacteriophages. In: Subramanian G (ed) Continuous process-
York ing in pharmaceutical manufacturing. Wiley-VCH Verlag GmbH &
Levin BR, Stewart FM, Chao L (1977) Resource-limited growth, com- Co. KGaA, Weinheim, pp 297–338
petition, and predation: a model and experimental studies with bac- Rathore AS (2015) Continuous processing for production of biotech ther-
teria and bacteriophage. Am Nat 111:3–24 apeutics. PDA J Pharm Sci Technol 69:333. https://doi.org/10.5731/
Lin DM, Koskella B, Lin HC (2017) Phage therapy: an alternative to pdajpst.2015.01072
antibiotics in the age of multi-drug resistence. World J Gastrointest Roach DR, Debarbieux L (2017) Phage therapy: awakening a sleeping
Pharmacol Ther 8:162–173. https://doi.org/10.4292/wjgpt.v8.i3. giant. Emerg Top Life Sci 1:93–103. https://doi.org/10.1042/
162 ETLS20170002
694 Appl Microbiol Biotechnol (2019) 103:685–694

Ross A, Ward S, Hyman P (2016) More is better: selecting for broad host Storms ZJ, Sauvageau D (2015) Modeling tailed bacteriophage adsorp-
range bacteriophages. Front Microbiol 7:1352. https://doi.org/10. tion: insight into mechanisms. Virology 485:355–362. https://doi.
3389/fmicb.2016.01352 org/10.1016/j.virol.2015.08.007
Roy B, Philippe C, Loessner M, Goulet J, Moineau S (2018) Production Storms ZJ, Arsenault E, Sauvageau D, Cooper DG (2010) Bacteriophage
of bacteriophages by Listeria cells entrapped in organic polymers. adsorption efficiency and its effect on amplification. Bioprocess
Viruses 10:324. https://doi.org/10.3390/v10060324 Biosyst Eng 33:823–831. https://doi.org/10.1007/s00449-009-
Santos SB, Carvalho C, Azeredo J, Ferreira EC (2014) Population dy- 0405-y
namics of a Salmonella lytic phage and its host: implications of the Storms ZJ, Brown T, Cooper DG, Sauvageau D, Leask RL (2013) Impact
host bacterial growth rate in modelling. PLoS One 9:e102507. of the cell life-cycle on bacteriophage T4 infection. FEMS
https://doi.org/10.1371/journal.pone.0102507 Microbiol Lett 353:63–68. https://doi.org/10.1111/1574-6968.
Sargeant K (1969) The deep culture of bacteriophage. In: Norris JR, 12402
Ribbons DW (eds) Methods in microbiology, vol 1. Academic Summers WC (2012) The strange history of phage therapy.
Press, London, pp 505–520 Bacteriophage 2:130–133. https://doi.org/10.4161/bact.20757
Sauvageau D, Cooper DG (2010) Two-stage, self-cycling process for the Taj MK, Ji X, Taj I, Hussani TM, Samreen Z, Wei Y (2013) Different
production of bacteriophages. Microb Cell Factories 9:81. https:// phase Escherichia coli effect on T4 bacteriophage lysis and produc-
doi.org/10.1186/1475-2859-9-81 tion. Int J Appl Nat Sci 2:73–76
Schofield DA, Sharp NJ, Westwater C (2012) Phage-based platforms for
Tucker RG (1961) The role of magnesium ions in the growth of
the clinical detection of human bacterial pathogens. Bacteriophage
Salmonella phage anti-R. J Gen Microbiol 26:313–323. https://doi.
2:105–283. https://doi.org/10.4161/bact.19274
org/10.1099/00221287-26-2-313
Schooley RT, Biswas B, Gill JJ, Hernandez-Morales A, Lancaster J,
Lessor L, Barr JJ, Reed SL, Rohwer F, Benler S, Segall AM, Twort FW (1915) An investigation on the nature of ultra-microscopic
Taplitz R, Smith DM, Kerr K, Kumaraswamy M, Nizet V, Lin L, viruses. Lancet 186:1241–1243. https://doi.org/10.1016/S0140-
McCauley MD, Strathdee SA, Benson CA, Pope RK, Leroux BM, 6736(01)20383-3
Picel AC, Mateczun AJ, Cilwa KE, Regeimbal JM, Estrella LA, Varon M (1979) Selection of predation-resistant bacteria in continuous
Wolfe DM, Henry MS, Quinones J, Salka S, Bishop-Lilly KA, culture. Nature 277:386–388. https://doi.org/10.1038/277386a0
Young R, Hamilton T (2017) Development and use of personalized Volterra V (1926) Fluctuations in the abundance of a species considered
bacteriophage-based therapeutic cocktails to treat a patient with a mathematically. Nature 118:558–560
disseminated resistant Acinetobacter baumannii infection. Wang IN (2006) Lysis timing and bacteriophage fitness. Genetics 172:
Antimicrob Agents Chemother 61:e00954–e00917. https://doi. 17–26. https://doi.org/10.1534/genetics.105.045922
org/10.1128/AAC.00954-17 Warner CM, Barker N, Lee S-W, Perkins EJ (2014) M13 bacteriophage
Schwienhorst A, Lindemann BF, Eigen M (1996) Growth kinetics of a production for large-scale applications. Bioprocess Biosyst Eng 37:
bacteriophage in continuous culture. Biotechnol Bioeng 50:217– 2067–2072. https://doi.org/10.1007/s00449-014-1184-7
221. https://doi.org/10.1002/bit.260500204 Weitz JS, Hartman H, Levin SA (2005) Coevolutionary arms races be-
Shao Y, Wang I-N (2008) Bacteriophage adsorption rate and optimal lysis tween bacteria and bacteriophage. PNAS 102:9535–9540. https://
time. Genetics 180:471–482. https://doi.org/10.1534/genetics.108. doi.org/10.1073/pnas.0504062102
090100 Wittebole X, De Roock S, Opal SM (2014) A historical overview of
Sharma S, Chatterjee S, Datta S, Prasad R, Dubey D, Prasad RK, Vairale bacteriophage therapy as an alternative to antibiotics for the treat-
MG (2017) Bacteriophages and its applications:an overview. Folia ment of bacterial pathogens. Virulence 5:226–235. https://doi.org/
Microbiol (Praha) 62:17–55. https://doi.org/10.1007/s12223-016- 10.4161/viru.25991
0471-x Wright A, Hawkins CH, Änggård EE, Harper DR (2009) A controlled
Smrekar F, Ciringer M, Peterka M, Podgornik A, Strancar A (2008) clinical trial of a therapeutic bacteriophage preparation in chronic
Purification and concentration of bacteriophage T4 using monolithic otitis due to antibiotic-resistant Pseudomonas aeruginosa ; a prelim-
chromatographic supports. J Chromatogr B Analyt Technol Biomed inary report of efficacy. Clin Otolaryngol 34:349–357. https://doi.
Life Sci 861:177–180. https://doi.org/10.1016/j.jchromb.2007.05. org/10.1111/j.1749-4486.2009.01973.x
048 Yosef I, Kiro R, Molshanski-Mor S, Edgar R, Qimron U (2014) Different
Spanakis E, Horne MT (1987) Co-adaptation of Escherichia coli and approaches for using bacteriophages against antibiotic-resistant bac-
Coliphage λvir in continuous culture. J Gen Microbiol 133:353– teria. Bacteriophage 4:e28491. https://doi.org/10.4161/bact.28491
360. https://doi.org/10.1099/00221287-133-2-353
You L, Suthers PF, Yin J (2002) Effects of Escherichia coli physiology on
Storms ZJ, Sauvageau D (2014) Evidence that the heterogeneity of a T4
growth of phage T7 in vivo and in silico. J Bacteriol 184:1888–
population is the result of heritable traits. PLoS One 9:e116235.
1894. https://doi.org/10.1128/JB.184.7.1888-1894.2002
https://doi.org/10.1371/journal.pone.0116235