RESEARCH

Evidence for
Elizabethkingia anophelis
Transmission from Mother
to Infant, Hong Kong
Susanna K.P. Lau,1 Alan K.L. Wu,1 Jade L.L. Teng,1 Herman Tse,1 Shirly O.T. Curreem,
Stephen K.W. Tsui, Yi Huang, Jonathan H.K. Chen, Rodney A. Lee, Kwok-Yung Yuen, Patrick C.Y. Woo

Elizabethkingia anophelis, recently discovered from mos- cholerae, and mycobacteria (3–14). However, the routine
quito gut, is an emerging bacterium associated with neo- application of this method in diagnostic microbiology and
natal meningitis and nosocomial outbreaks. However, infection control, especially for less well-defined, emerging
its transmission route remains unknown. We use rapid pathogens, is yet to be explored.
genome sequencing to investigate 3 cases of E. anoph- Elizabethkingia anophelis is a recently discovered bac-
elis sepsis involving 2 neonates who had meningitis and 1
terium isolated from the midgut of the Anopheles gambiae
neonate’s mother who had chorioamnionitis. Comparative
genomics revealed evidence for perinatal vertical transmis-
mosquito in 2011 (15). The genus Elizabethkingia also in-
sion from a mother to her neonate; the 2 isolates from these cludes E. meningoseptica (previously named Chryseobac-
patients, HKU37 and HKU38, shared essentially identical terium/Flavobacterium meningosepticum) and E. miricola
genome sequences. In contrast, the strain from another (16). E. meningoseptica causes neonatal sepsis and infec-
neonate (HKU36) was genetically divergent, showing only tions in immunocompromised persons. E. anophelis has
78.6% genome sequence identity to HKU37 and HKU38, also recently been reported to cause neonatal meningitis in
thus excluding a clonal outbreak. Comparison to genomes the Central African Republic, and a nosocomial outbreak
from mosquito strains revealed potential metabolic adapta- was reported in an intensive care unit in Singapore (17–19).
tions in E. anophelis under different environments. Mater- However, the role of mosquitoes or other sources in the
nal infection, not mosquitoes, is most likely the source of transmission of E. anophelis remains unclear.
neonatal E. anophelis infections. Our findings highlight the
In 2012, we encountered 3 cases of Elizabethkingia
power of genome sequencing in gaining rapid insights on
transmission and pathogenesis of emerging pathogens.
sepsis associated with meningitis in 2 neonates and cho-
rioamnionitis in a neonate’s mother in a hospital in Hong
Kong. Three strains of Elizabethkingia-like, gram-neg-

M icrobial genome sequencing can enhance diagnosis

and control of infectious diseases (1,2). Its ultimate
molecular resolution is superior to other phenotypic and
ative bacilli sharing similar phenotypic characteristics
were isolated from the 3 patients, but confident identifi-
cation results were not obtained by matrix-assisted laser
genotypic tests and enables not only rapid microbial iden- desorption ionization/time-of-flight (MALDI-TOF) mass
tification but also characterization of transmission events. spectrometry and 16S rRNA gene sequencing. Moreover,
The technique has been applied in large-scale infectious clinical and microbiological data did not provide adequate
disease outbreaks such as those caused by Escherichia coli clues about the possible transmission route. We therefore
O104:H4, Staphylococcus aureus, Streptococcus pyogenes, attempted to use draft genome sequencing to rapidly dis-
Enterococcus faecium, Pseudomonas aeruginosa, Vibrio sect transmission pathways and confirm the identity of
the species.
Author affiliations: The University of Hong Kong, Hong Kong
(S.K.P. Lau, J.L.L. Teng, H. Tse, S.O.T. Curreem, Y. Huang, Materials and Methods
J.H.K. Chen, K.-Y. Yuen, P.C.Y. Woo); State Key Laboratory of
Emerging Infectious Diseases, Research Centre of Infection and Setting and Patients
Immunology, Carol Yu Centre for Infection, Hong Kong (S.K.P. The 3 patients were hospitalized in an acute regional hos-
Lau, H. Tse, K.Y. Yuen, P.C.Y. Woo); Pamela Youde Nethersole pital, Pamela Youde Nethersole Eastern Hospital, which
Eastern Hospital, Hong Kong (A.K.L. Wu, R.A. Lee); School of is situated in the eastern area of Hong Kong Island. This
Biomedical Sciences, The Chinese University of Hong Kong, study was approved by the Institute Review Board, Hos-
Hong Kong (S.K.W. Tsui) pital Authority, Hong Kong (reference HKEC-2013-051).
DOI: http://dx.doi.org/10.3201/eid2102.140623 1
These authors contributed equally to this article.

232 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015
 E. anophelis Transmission from Mother to Infant

Microbiological Methods fever of 1 day’s duration. He was born at the same hospital
Bacterial cultures and phenotypic identification were per- 21 days earlier at 41 weeks’ gestation by vaginal delivery
formed according to standard protocols by using the Vitek and was discharged on day 3. Physical examination did not
II system (bioMérieux, Marcy l’Etoile, France). Antimi- show obvious infective focus. Serum C-reactive protein
crobial drug susceptibility testing was performed by E-test (CRP) was elevated to 109 mg/L. Lumbar puncture was
method for vancomycin and Kirby-Bauer disk diffusion for performed; analysis of cerebrospinal fluid (CSF) showed
other drugs; because interpretative criteria for Elizabeth- polymorph pleocytosis, elevated protein levels, and low
kingia were lacking, results were interpreted according to glucose levels (Table). Treatment was initiated for bacte-
Clinical and Laboratory Standards Institute for Pseudomo- rial meningitis with empirical intravenous ampicillin and
nas aeruginosa (20). MALDI-TOF mass spectrometry was cefotaxime. Blood and CSF cultures recovered a gram-
performed by the direct transfer method as described previ- negative bacillus, designated HKU36. Antimicrobial drugs
ously (21), with modifications by using the Bruker Dalton- were changed to vancomycin, piperacillin, and rifampin on
ics microflex LT system with Reference Library Biotyper day 3. The patient was discharged after 3 weeks of intrave-
version 3.1 (Bruker Daltonik GmbH, Leipzig, Germany). nous drug treatment, without neurologic sequelae (Figure
Full 16S rRNA gene amplification and squencing were 1). The neonate’s mother was admitted to the same hospi-
performed according to previously published protocols tal 1 day after the infant’s admission for postpartum fever,
with modifications (22,23). Pulsed-field gel electrophoresis chills, rigor, and abdominal pain. Transvaginal ultrasound
(PFGE) was performed by using the CHEF Mapper XA showed no retained gestational products. Serum CRP level
system (Bio-Rad, Hercules, CA, USA) and restriction en- was elevated to 109 mg/L; however, blood cultures were
donuclease XbaI as described previously (8,22). negative. She was treated with intravenous cefuroxime and
metronidazole and discharged on day 6 with oral cefurox-
Draft Genome Sequencing and Analysis ime and metronidazole.
The draft genome sequences of the 3 E. anophelis strains In November 2012, a 33-year-old woman in week 30
were determined by high-throughput sequencing with the of pregnancy (patient 2) was admitted to the same hospi-
Illumina HiSeq 2500 system (Illumina, San Diego, CA, tal because of prolonged premature rupture of membranes.
USA). Samples of 50 ng of genomic DNA were extract- She stayed at the same antenatal ward and in the same cu-
ed by using a genomic DNA purification kit (QIAGEN, bicle as the mother of patient 1 (Figure 1). Fever devel-
Hilden, Germany) from cultures grown overnight on blood oped in the patient 3 days after admission, and clinical tests
agar at 37°C, as described previously (24,25). Each sample showed peripheral leukocytosis with neutrophilia (Table).
was sequenced by 151-bp paired-end reads with mean li- Serum CRP was elevated to 108 mg/L. Treatment with in-
brary size of 350 bp. Sequencing errors were corrected by travenous penicillin G was commenced, and an emergency
k-mer frequency spectrum analysis using SOAPec (http:// lower segment cesarean section was performed. Placental
soap.genomics.org.cn/about.html). De novo assembly was and uterine swab cultures recovered a gram-negative bacil-
performed in SOAPdenovo2 (http://soap.genomics.org.cn/ lus, designated HKU37. Blood cultures were negative. An-
soapdenovo.html). Prediction of protein coding regions timicrobial drug treatment was changed to cefuroxime and
and automatic functional annotation was performed by metronidazole, followed by oral ciprofloxacin for 1 week.
using Glimmer3 (26) and the RAST (Rapid Annotations Her fever subsided, and she was discharged on day 8.
using Subsystem Technology) server (27). Antibiotic re- The baby girl (patient 3) of patient 2 was pale and
sistomes were identified by using the Antibiotic Resistance flaccid at birth; apnea of prematurity developed, requir-
Genes Database (28). BLASTn comparisons were run in ing cardiopulmonary resuscitation. Peripheral leukopenia
BLAST+ (http://blast.ncbi.nlm.nih.gov/Blast.cgi) with an and metabolic acidosis were also detected, and serum CRP
E-value cutoff of 10.0. In addition, manual annotation was level was elevated to 70.6 mg/L. Chest radiograph showed
performed on putative virulence and antibiotic resistance bilateral ground-glass appearance. Lumbar puncture was
genes by protein domain predictions and multiple sequence performed, and CSF showed lymphocytic pleocytosis with
alignments with orthologous genes. Intergenomic distance elevated protein levels and low glucose levels (Table). Ul-
was calculated by using Genome-to-Genome Distance Cal- trasound of the brain showed grade I to II intraventricular
culator 2.0 (http://ggdc.dsmz.de/distcalc2.php) (29). hemorrhages. Empirical intravenous ampicillin and ce-
fotaxime at meningitic dose was started. Blood and CSF
Results cultures recovered a gram-negative bacillus, designated
HKU38. Antimicrobial drug therapy was changed to intra-
Patients venous vancomycin, piperacillin/tazobactam, and rifampin
In July 2012, a 21-day-old male neonate (patient 1) was on day 3, continuing for 3 weeks. Necrotizing enterocolitis
admitted to Pamela Youde Nethersole Eastern Hospital for and neonatal jaundice developed, but both resolved with

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015 233
RESEARCH

Table. Clinical characteristics and results of testing for 3 patients infected with Elizabethkingia anophelis, Hong Kong, 2012*
Characteristics Patient 1 Patient 2† Patient 3
Patient age/sex 21 d/M 33 y/F 0 d/F
Signs/symptoms Fever Fever, PPROM Apnea at birth
Blood test results
Total leukocytes, × 109 cells/L 16.0 (5.0–19.5) 15.2 (3.7–9.3) 5.1 (10.0–27.0)
Neutrophils, × 109 cells/L 6.8 (2.0–9.5) 12.5 (1.8–6.2) 1.2 (5.0–17.0)
Lymphocytes, × 109 cells/L 6.8 (2.5–11.0) 1.7 (1.0–3.2) 3.4 (3.0–10.0)
Monocytes, × 10 cells/L
9
2.3 (0.2–1.2) 0.8 (0.2–0.7) 0 (0.5–2.0)
Hemoglobin, g/dL 14.0 (11.0–19.0) 10.7 (11.5–15.4) 16.1 (13.5–19.5)
Platelets, × 109/L 180 (180–460) 241 (160–420) 186 (100–300)
C-reactive protein, mg/L 109 (<8.0) 108 (<5.0) 70.6 (<8.0)
CSF test results
Total leukocytes, × 106 cells/L 1,445 NA 5,850
Polymorphs, % 67 NA 1
Lymphocytes, % 33 NA 99
Protein, g/L 1.33 (0.15–0.45) NA 2.69 (0.15–0.45)
Glucose, mmol/L 2.2 (2.8–4.4) NA 1.5 (2.8–4.4)
CSF/serum glucose, % 38 NA 24
Positive culture sites for E. anophelis Blood, CSF Placental swab, uterine swab Blood, CSF
Phenotypic characteristics of isolates
Colony pigment Pale yellow None None
Citrate utilization Negative Delayed positive Delayed positive
Antimicrobial drug susceptibilities of isolates
Ampicillin Resistant Resistant Resistant
Pipercillin Susceptible Susceptible Susceptible
Cefoperazone/sulbactam Susceptible Susceptible Susceptible
Cefotaxime Intermediate Resistant Resistant
Ceftazidime Resistant Resistant Resistant
Imipenem Resistant Resistant Resistant
Amikacin Resistant Resistant Resistant
Gentamicin Resistant Resistant Resistant
Kanamycin Resistant Resistant Resistant
Streptomycin Resistant Resistant Resistant
Tobramycin Resistant Resistant Resistant
Ciprofloxacin Susceptible Susceptible Susceptible
Moxifloxacin Susceptible Susceptible Susceptible
Tetracycline Resistant Resistant Resistant
Trimethoprim/sulfamethoxazole Susceptible Susceptible Susceptible
Rifampin Susceptible Susceptible Susceptible
Chloramphenicol Resistant Resistant Resistant
Vancomycin MIC, μg/mL 16 4 4
Antimicrobial drug regimen Ampicillin + cefotaxime; Penicillin G; cefuroxime + Ampicillin + cefotaxime;
vancomycin + piperacillin + metronidazole; ciprofloxacin vancomycin +
rifampin pipercillin/tazobactam +
rifampin
Complications None None Respiratory distress,
intraventricular
hemorrhage
*Reference ranges are shown in parentheses. PPROM, prolonged premature rupture of membranes; CSF, cerebrospinal fluid.
†Mother of patient 3.

treatment (Figure 1). The infant was discharged on day 54 E. meningoseptica (best match to E. meningoseptica strain
without neurologic sequelae. 002_NEB14 NFI, with scores of 2.106 and 2.007, respec-
tively), whereas strain HKU36 was only identified to the
Clinical and Microbiological Investigations genus level as Elizabethkingia species (best match to E. me-
The 3 isolates from these patients, HKU36–38, were non- ningoseptica strain 002_NEB14 NFI, with score of 1.853)
motile, oxidase-positive, non–glucose-fermenting, gram- (online Technical Appendix Figure 1). The isolates’ 16S
negative bacilli. Their phenotypic characteristics are rRNA gene sequences exhibited 99.1%–99.9% nucleotide
summarized in the Table and online Technical Appendix identities to those of E. anophelis type strain R26T (Gen-
Table 1 (http://wwwnc.cdc.gov/EID/article/21/2/14-0623- Bank accession no. EF426425) and 97.4%–99.9% nucleo-
Techapp1.pdf). The isolates were identified as E. menin- tide identities to those of E. meningoseptica strains depos-
goseptica by using the Vitek II identification system (bio- ited in GenBank (GenBank accession nos. HM056770.1,
Mérieux, Marcy L’Étoile, France). However, MALDI-TOF GU180602.1, JQ673498.1, FJ816020, AVCQ01000012,
mass spectrometry identified strains HKU37 and HKU 38 as FJ839441.1, JN201943.1, and AJ704540).

234 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015
 E. anophelis Transmission from Mother to Infant

Figure 1. Clinical course of illness in 3 patients infected with Elizabethkingia anophelis in whom sepsis developed and the mother of
patient 1, who had culture-negative postpartum fever, Hong Kong, 2012. Locations where patients were treated at the hospital and times
when they were home are noted. CSF, cerebrospinal fluid; leaking, leaking of amniotic fluid (membrane rupture).

The high sequence identities to both E. anophelis and CBYE010000001–CBYE010000032, CBYF010000001–

E. meningoseptica made the species identity of the 3 strains CBYF010000038; online Technical Appendix Table 2).
uncertain, despite 16S rRNA gene sequencing. Moreover, These contigs contained 3,654–3,667 predicted protein-
the strains exhibited minor differences in phenotypes and coding genes (Figure 2, panel A). Using Genome-to-
antibiogram (Table). Further, because the mothers stayed Genome Distance Calculator for intergenomic distance
in the same ward before delivery (although 4 months estimation, which enabled genome-based species delin-
apart), concerns of a possible nosocomial outbreak were eation analogous to traditional DNA–DNA hybridization
raised. However, environmental and water samples from method, we found that these genomes shared 78.3%–
the hospital and patients’ homes were culture-negative for 85.4% nucleotide identities to the draft genome sequence
E. anophelis. A program of enhanced infection control of E. anophelis type strain R26T, the initial isolate from
measures was enforced in the hospital, and no further cases an Anopheles gambiae mosquito (GenBank accession no.
were identified. NZ_ANIW00000000.1). However, the genomes shared
only 23.6%–23.7% nucleotide identities to the draft ge-
Genome Sequencing and Comparative Analysis nome sequence of E. meningoseptica type strain ATCC
of E. anophelis Genomes 13253T (GenBank accession no. BARD00000000.1) (Fig-
We sequenced the draft genomes of strains HKU36–38 to ure 2, panel B). Phylogenetic analysis using the draft ge-
investigate their genetic relatedness and confirm their spe- nomes and concatenated sequences of 69 housekeeping
cies identity. Sequencing generated 11–15 million paired- genes also supported the identification of the 3 strains
end reads per strain (estimated 410–540-fold coverage). as E. anophelis (Figure 3; online Technical Appendix
After de novo assembly, the 3 draft genomes ranged from Figure 2).
3.92–3.99 Mb in length (G + C content 35.4%–35.8%) The sequences from 52 contigs of strain HKU37 dem-
and were distributed in 42–52 large (>500 bp) contigs onstrated 99.4% nucleotide identity to those from 46 con-
(EMBLaccessionnos.CBYD010000001–CBYD010000042, tigs of strain HKU38, indicating that these draft genomes

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015 235
RESEARCH

Figure 2. Comparison of draft

genome sequence data of the
3 Elizabethkingia anophelis
strains from patients in Hong
Kong (HKU36–38), E anophelis
type strain R26T, and E.
meningoseptica type strain
ATCC 13253T. A) Distributions
of predicted coding sequence
function in genomes of E.
anophelis strains HKU36–38, E.
anophelis type strain R26T, and
E. meningoseptica type strain
ATCC 13253T according to SEED
Subsystems are shown. The
columns indicate the number of
proteins in different subsystems.
B) Circular representation of
sequence comparison between
the draft genome of strain HKU37
and other draft genomes as
labeled. Comparison generated
in Rapid Annotations using
Subsystem Technology (27).
Intensity of color indicates degree
of protein identity.

are essentially identical (Figure 2, panel B, and Figure 3). Figure 3), consistent with PFGE patterns (Figure 4). More-
The small intergenomic distance can be explained by slight over, a potential genetic island consisting of conjugative
differences in coverage or contig assembly; sequences of transposable elements was found in strains HKU37 and
2,000 high-coverage protein-coding genes were identical HKU38 but not in HKU36. Our results exclude a clonal
between HKU37 and HKU38. In contrast, these sequenc- outbreak, but the extremely close genetic relatedness be-
es demonstrated only 78.6% nucleotide identity to those tween strains HKU37 and HKU38 provides evidence for
from the 42 contigs of strain HKU36, indicating that strain vertical transmission from patient 2 to patient 3 (mother
HKU36 is genetically divergent (Figure 2, panel B, and to infant).

236 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015
 E. anophelis Transmission from Mother to Infant

Figure 3. Phylogenetic trees constructed by using draft genome sequences and concatenated sequences of 69 housekeeping genes of
3 Elizabethkingia anophelis strains from patients in Hong Kong (HKU36–38). A) Neighbor-joining tree constructed on the basis of draft
genome sequences using by using Genome-to-Genome Distance Calculator 2.0 (http://ggdc.dsmz.de/distcalc2.php; formula 1) and
Chryseobacterium gleum ATCC 35910 as the root. Arrow indicates route of mother-to-neonate transmission. B) Maximum-likelihood
tree constructed on the basis of 69 housekeeping genes, showing the relationship of E. anophelis strains HKU36–38 to related bacterial
species, using RAxML version 7.2.8 (http://sco.h-its.org/exelixis/software.html) and Weeksella virosa DSM 16922 as the root. A total
of 78,520 nt positions were included in the analysis. Bootstrap values were calculated from 1,000 replicates. Scale bars indicate mean
number of nucleotide substitutions per site on the respective branches. Gene names and accession numbers are given as cited in
GenBank (online Technical Appendix Table 2, http://wwwnc.cdc.gov/EID/article/21/2/14-0623-Techapp1.pdf). ‘E. meningoseptica’ strain
502 is a misidentified isolate that actually belongs to E. anophelis on the basis of draft genome sequencing.

Potential Virulence Factors and Resistance Further studies may investigate the possible role of AgDI
Genes in E. anophelis and potential adherence factors for vaginal colonization in
The association of E. anophelis with neonatal meningitis E. anophelis.
in this and previous reports (17,18) suggests that the bac- Similar to E. meningoseptica, the 3 E. anophelis iso-
terium may possess virulence factors that enable it to in- lates we identified are resistant to multiple antimicrobial
vade the central nervous system. The 3 draft genomes we drugs. We found various antimicrobial resistance genes
identified contain homologs of several virulence genes consistent with their resistance phenotypes, including
found in Listeria monocytogenes, which also causes neo- metallo-β-lactamase (blaGOB-1 and blaB14 in strain HKU36
natal meningitis. These genes include cell wall hydrolase and a novel blaGOB and blaB1 in strains HKU37 and HKU38)
A, which enables host cell invasion; phosphatidylino- and extended-spectrum β-lactamase (blaACME-1 in strains
sitol-specific phospholipase (PlcA) and listeriolysin O HKU37 and HKU38 and a potential novel blaACME-1 vari-
(LLO), which enable escape from the primary vacuole ant in strain HKU36). A comparison of these β-lactamases
of macrophages, and genes that enable survival in the to their corresponding orthologs in E. meningoseptica ge-
secondary vacuole of macrophages; and virulence clus- nomes revealed only 74%–85% amino acid identities, in-
ter protein B (VclB). Phosphatidylinositol-specific phos- dicating that E. anophelis and related bacteria are potential
pholipase, listeriolysin O, and virulence cluster protein reservoirs of novel β-lactamase genes (19,34,35). Other
B are located in the Listeria pathogenicity island LIPI-1 antimicrobial resistance genes found included multidrug-
(30,31). Moreover, the 3 genomes we identified contain resistance efflux pumps (ATP binding cassette superfam-
homologs of arylsulfatase and genes that enable invasion ily, major facilitator superfamily, resistance-nodulation-
of brain endothelial cells, which contribute to the abil- division families, multidrug and toxic-compound extrusion
ity of Escherichia coli to cross the blood–brain barrier in family) that potentially carry resistance to a variety of
neonatal meningitis (32). compounds; chloramphenicol acetyltransferase; amino-
Vertical transmission of E. anophelis from mother to glycoside 6-adenyltransferase; and tetracycline resistant
infant also suggests that the bacterium may be able to colo- gene. Moreover, a putative tetX gene was also identified;
nize the vagina before causing ascending chorioamnionitis this gene encodes a predicted flavin-dependent monooxy-
in the mother and neonatal infection through transplacental genase with tetracycline/tigecycline-degrading activity,
spread. A homologof the gene encoding agmatine deimi- although the 3 strains we identified are only resistant to tet-
nase, AgDI, which mediates acid tolerance in L. mono- racycline but remained susceptible to other related drugs,
cytogenes (33), was found in the E. anophelis genomes. including tigecycline.

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015 237
RESEARCH

Figure 4. Pulsed-field gel

electrophoresis (PFGE) analysis of
samples from patients in Hong Kong
showing 3 Elizabethkingia anophelis
strains compared with reference
Elizabethkingia isolates. A) PFGE
performed by using CHEF Mapper XA
system (Bio-Rad, Hercules, CA, USA)
and restriction endonuclease XbaI
shows that isolates from patient 2 and
patient 3 are indistinguishable, wheras
isolates from patient 1 possess distinct
PFGE patterns. Lane 1, E. anophelis
strain HKU37 from uterine swab
specimen of patient 2; lane 2, placental
swab specimen from patient 2; lane 3,
E. anophelis strain HKU38 from blood
of patient 3; lane 4, cerebrospinal fluid
from patient 3; lane 5, E. anophelis
strain HKU36 from blood of patient 1;
lane 6, cerebrospinal fluid from patient
1; lane 7, E. anophelis type strain
R26T; lane 8, E. meningoseptica type
strain ATCC 13253T; lane 9,
E. miricola type strain LMG22470T.
B) Dendrogram constructed with
PFGE data by similarity and clustering
analysis using the Dice coefficient
(1% tolerance and 0.5% optimization)
and the unweighted pair-group
method using average linkages with
GelCompar II (Applied Maths,
Sint-Martens-Latem, Belgium).

Comparison of Genomes from Human and also absent in the genome of E. meningoseptica type strain
Mosquito E. anophelis Strains ATCC 13253T, which suggests that mosquito strains of
E. anophelis strains R26T and Ag1 were isolated from mos- E. anophelis may be evolutionarily distinct from clinical
quitoes (35). Compared with those strains, the genomes of strains of E. anophelis and E. meningoseptica. More ge-
the 3 strains we identified possessed 33 unique hypothetical nome sequence data from other clinical and environmental
proteins. Moreover, the genetic island consisting of conju- strains of E. anophelis may shed light on the ecology, biol-
gative transposable elements found in strains HKU37 and ogy, and pathogenesis of E. anophelis.
HKU38 was also absent in the mosquito strains. In contrast
to the mosquito strains, which possessed genes encoding Discussion
for xylose isomerase (XylA) and xylulose kinase (XylB), This study demonstrates the power of draft genome se-
these 2 genes were absent in the 3 strains we identified. quencing to rapidly dissect transmission pathways for
This finding may reflect different requirements for sugar emerging bacterial infections. Our results showed that ver-
metabolism in E. anophelis under different environments. tical perinatal transmission had occurred from patient 2, a
Notably, despite the presence of XylA and XylB, E. anoph- pregnant woman who had chorioamnionitis, to patient 3, a
elis mosquito strain R26T did not produce acid from xylose neonate who had early onset neonatal meningitis. The in-
(15). However, this finding does not exclude the strain’s fective source for patients 2 and 3 was unlikely to have
ability to metabolize xylose, as D-xylulose 5-phosphate, been patient 1 or his mother. However, we speculate that
the product of XylA and XylB, can be used as a substrate the mother of patient 1 might also have had E. anophelis
for the pentose-phosphate pathway. XylA and XylB were chorioamnionitis, as evidenced by postpartum fever and

238 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015
 E. anophelis Transmission from Mother to Infant

abdominal pain, which resulted in late-onset meningitis more closely related to E. anophelis than to E. meningo-
in her son owing to fastidious bacterial growth. Although septica in their 16S rRNA sequences (Figure 3; online
strain HKU36 did not belong to the same clone as strains Technical Appendix Figure 2) (37). These ambiguous, po-
HKU37/38, a polyclonal outbreak of E. anophelis sepsis in tentially misidentified strains may cause incorrect interpre-
the labor ward, in which case an environmental source is tations in suspected E. anophelis infections. For example,
likely, could not be excluded. the sequence of strain HKU36 possessed 99.8% nucleotide
The discovery of E. anophelis in mosquito gut has identity to that of E. meningingoseptica strain G3-1-08 but
raised suspicion that mosquitoes are the source of neonatal only 99.1% nucleotide identity to that of E. anophelis strain
meningitis cases in Africa (17). Although Anopheles mos- R26T. Furthermore, phenotypic tests such as acid produc-
quitoes are not found in Hong Kong, the role of local mos- tion from cellobiose and citrate utilization, previously pro-
quitoes as reservoirs for E. anophelis remains unknown. posed to be useful for identification of E. anophelis (15),
Nonetheless, the vertical transmission demonstrated in 1 are probably unreliable in differentiating among Elizabeth-
neonate makes mosquitoes unlikely as vehicles of trans- kingia species. For example, E. anophelis strain R26T pro-
mission in our cases. duces acid from cellobiose, but the 3 strains we identified
Our report provides genomic evidence for vertical do not; in addition, E. anophelis strains R26T, HKU37, and
transmission in neonatal meningitis. Whereas we cannot HKU38, but not strain HKU36, utilize citrate (online Tech-
ascertain how the mother(s) acquired the infection, our re- nical Appendix Table 1). Strain HKU36 displayed higher
sults prompt further work to assess the importance of ma- MIC of vancomycin than did strains HKU37 and HKU38
ternal source in neonatal meningitis caused by E. anoph- and type strains of E. anophelis, E. meningoseptica, and
elis and other bacterial agents. Maternal colonization with E. miricola, which correlates with previous reports on
Lancefield group B streptococcus (GBS) during pregnancy variable vancomycin susceptibilities in Elizabethkingia
is the primary risk factor for early onset neonatal disease. (38,39). The species identity of the 3 strains we identified
However, direct microbiological evidence for vertical was only resolved by intergenomic comparison. Inclusion
transmission is seldom available, especially for bacterial of E. anophelis in MALDI-TOF MS databases and recti-
agents other than GBS. Further genomic studies may help fication of 16S rRNA gene sequences of Elizabethkingia
investigate the role of vertical transmission in neonatal strains deposited in databases will enable accurate diagno-
meningitis caused by other bacteria. Current indications sis of more E. anophelis infections.
for intrapartum antimicrobial drugs prophylaxis have been The draft genome sequences we identified have en-
determined on the basis of risk factors for early onset GBS abled rapid exploration of novel β-lactamase and other
disease; therefore, intravenous penicillin G or ampicillin antimicrobial drug resistance genes and possible viru-
is often the standard empirical regimen used. However, if lence genes in E. anophelis, highlighting the potential of
further research determines that the mother may also be a genome sequencing in identifying novel drug-resistance
source of transmission for other bacterial agents, broader- mechanisms and guiding treatment regimens for emerging,
spectrum antimicrobial drugs may need to be considered multidrug-resistant bacteria (25,34,40). Because previous
as treatment for intrapartum fever or prolonged rupture cases of E. anophelis neonatal meningitis have been associ-
of membranes. ated with poor outcomes (17,18), further work to elucidate
E. anophelis is likely an underreported bacterium be- the pathogenesis and antimicrobial drug resistance patterns
cause it can be easily misidentified as E. meningoseptica, of this emerging pathogen may help improve clinical man-
which shares a similar phenotypic profile (17,19). The E. agement of illness. The findings of potential genes related
anophelis isolates from the recent outbreak reported in to neuroinvasion and acid tolerance and the unique genetic
Singapore were initially mistakenly identified as E. menin- characteristics in clinical strains of E. anophelis compared
goseptica (19,36). Of the 3 strains we identified, 2 were with mosquito strains may also provide insights on the abil-
misidentified as E. meningoseptica with MALDI-TOF ity of E. anophelis to adapt to different ecologic niches and
mass spectrometry, the state-of-the-art technology, which cause neonatal infection through vertical transmission.
is replacing conventional phenotypic identification in diag- In conclusion, the genome data we obtained for these
nostic laboratories. The reason for failure of MALDI-TOF cases offered superior discriminatory power that supported
mass spectrometry to identify these strains was that refer- appropriate infection control measures. The ability to dis-
ence E. anophelis strains are lacking in existing diagnostic tinguish different bacterial isolates often has critical im-
spectrum databases, as is the case with other less common- plications on practical infection-control management, but
ly encountered organisms (21). different strains of the same bacterial species may not be
Although 16S rRNA gene sequencing should provide distinguishable by their phenotypes because they reflect
sufficient resolution, some strains indexed as E. meningo- a tiny portion of the microbial genome. With better auto-
septica, such as strains G3-1-08 and 502, were actually mation and lower costs, draft genome sequencing, which

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015 239
RESEARCH

offers a short turnaround time, may replace existing typing emm59 from Montana to Wyoming, USA. Emerg Infect Dis.
methods such as PFGE or multilocus sequence typing for 2014;20:679–81. http://dx.doi.org/10.3201/eid2004.130564
10. Köser CU, Holden MT, Ellington MJ, Cartwright EJ,
outbreak investigations. Brown NM, Ogilvy-Stuart AL, et al. Rapid whole-genome
sequencing for investigation of a neonatal MRSA outbreak.
Acknowledgments N Engl J Med. 2012;366:2267–75. http://dx.doi.org/10.1056/
We thank Cheung-Hing Foo for technical support in bacterial NEJMoa1109910
11. Loman NJ, Constantinidou C, Christner M, Rohde H, Chan JZ,
identification and members of the Centre for Genomic Sciences, Quick J, et al. A culture-independent sequence-based metagenomics
The University of Hong Kong, for their technical support in ge- approach to the investigation of an outbreak of Shiga-toxigenic
nome sequencing. Escherichia coli O104:H4. JAMA. 2013;309:1502–10.
http://dx.doi.org/10.1001/jama.2013.3231
This work was supported by Health and Medical Research Fund, 12. Price JR, Golubchik T, Cole K, Wilson DJ, Crook DW,
Food and Health Bureau, The Government of the Hong Kong Thwaites GE, et al. Whole-genome sequencing shows that
patient-to-patient transmission rarely accounts for acquisition of
Special Administrative Region; Strategic Research Theme Fund,
Staphylococcus aureus in an intensive care unit. Clin Infect Dis.
Committee for Research and Conference Grant, and University 2014;58:609–18. http://dx.doi.org/10.1093/cid/cit807
Development Fund, The University of Hong Kong; the Shaw 13. Pérez-Lago L, Comas I, Navarro Y, González-Candelas F,
Foundation; donation from Ms. Eunice Lam; and Consultancy Herranz M, Bouza E, et al. Whole genome sequencing analysis of
intrapatient microevolution in Mycobacterium tuberculosis:
Service for Enhancing Laboratory Surveillance of Emerging In-
potential impact on the inference of tuberculosis transmission.
fectious Disease for the HKSAR Department of Health. J Infect Dis. 2014;209:98–108. http://dx.doi.org/10.1093/infdis/jit439
14. Johnson PD, Ballard SA, Grabsch EA, Stinear TP, Seemann T,
Dr. Lau is a clinical professor in the Department of Microbiology Young HL, et al. A sustained hospital outbreak of vancomycin-
at The University of Hong Kong. Her research focuses on micro- resistant Enterococcus faecium bacteremia due to emergence of
bial genomics for studying emerging infectious diseases. vanB E. faecium sequence type 203. J Infect Dis. 2010;202:
1278–86. http://dx.doi.org/10.1086/656319
15. Kämpfer P, Matthews H, Glaeser SP, Martin K, Lodders N, Faye I.
References Elizabethkingia anophelis sp. nov., isolated from the midgut of
1. Loman NJ, Constantinidou C, Chan JZ, Halachev M, Sergeant M, the mosquito Anopheles gambiae. Int J Syst Evol Microbiol.
Penn CW, et al. High-throughput bacterial genome sequencing: 2011;61:2670–5. http://dx.doi.org/10.1099/ijs.0.026393-0
an embarrassment of choice, a world of opportunity. Nat Rev 16. Kim KK, Kim MK, Lim JH, Park HY, Lee ST. Transfer of
Microbiol. 2012;10:599–606. http://dx.doi.org/10.1038/nrmicro2850 Chryseobacterium meningosepticum and Chryseobacterium miricola
2. Fournier PE, Drancourt M, Raoult D. Bacterial genome sequencing to Elizabethkingia gen. nov. as Elizabethkingia meningoseptica
and its use in infectious diseases. Lancet Infect Dis. 2007; comb. nov. and Elizabethkingia miricola comb. nov.
7:711–23. http://dx.doi.org/10.1016/S1473-3099(07)70260-8 Int J Syst Evol Microbiol. 2005;55:1287–93. http://dx.doi.
3. Shah MA, Mutreja A, Thomson N, Baker S, Parkhill J, Dougan G, org/10.1099/ijs.0.63541-0
et al. Genomic epidemiology of Vibrio cholerae O1 associated with 17. Frank T, Gody JC, Nguyen LB, Berthet N, Le Fleche-Mateos A,
floods, Pakistan, 2010. Emerg Infect Dis. 2014;20:13–20. Bata P, et al. First case of Elizabethkingia anophelis meningitis in
http://dx.doi.org/10.3201/.eid2001.130428 the Central African Republic. Lancet. 2013;381:1876.
4. Snyder LA, Loman NJ, Faraj LA, Levi K, Weinstock G, http://dx.doi.org/10.1016/S0140-6736(13)60318-9
Boswell TC, et al. Epidemiological investigation of Pseudomonas 18. Bobossi-Serengbe G, Gody JC, Beyam NE, Bercion R. First
aeruginosa isolates from a six-year-long hospital outbreak using documented case of Chryseobacterium meningosepticum meningitis
high-throughput whole genome sequencing. Euro Surveill. in Central African Republic. Med Trop (Mars). 2006;66:182–4.
2013;18:20611. 19. Teo J, Tan SY, Tay M, Ding Y, Kjelleberg S, Givskov M, et al.
5. Harris SR, Cartwright EJ, Török ME, Holden MT, Brown NM, First case of E anophelis outbreak in an intensive-care unit.
Ogilvy-Stuart AL, et al. Whole-genome sequencing for analysis Lancet. 2013;382:855–6. http://dx.doi.org/10.1016/
of an outbreak of methicillin-resistant Staphylococcus aureus: a S0140-6736(13)61858-9
descriptive study. Lancet Infect Dis. 2013;13:130–6. 20. Clinical and Laboratory Standards Institute. Performance standards
http://dx.doi.org/10.1016/S1473-3099(12)70268-2 for antimicrobial disk susceptibility tests. Approved standard,
6. Walker TM, Ip CL, Harrell RH, Evans JT, Kapatai G, Dedicoat MJ, 11th ed. M02–A11. Wayne (PA): The Institute; 2012.
et al. Whole-genome sequencing to delineate Mycobacterium 21. Lau SK, Tang BS, Curreem SO, Chan TM, Martelli P, Tse CW,
tuberculosis outbreaks: a retrospective observational study. et al. Matrix-assisted laser desorption ionization–time of flight
Lancet Infect Dis. 2013;13:137–46. http://dx.doi.org/10.1016/ mass spectrometry for rapid identification of Burkholderia
S1473-3099(12)70277-3 pseudomallei: importance of expanding databases with pathogens
7. Bryant JM, Grogono DM, Greaves D, Foweraker J, Roddick I, endemic to different localities. J Clin Microbiol. 2012;50:3142–3.
Inns T, et al. Whole-genome sequencing to identify transmission of http://dx.doi.org/10.1128/JCM.01349-12
Mycobacterium abscessus between patients with cystic fibrosis: a 22. Woo PC, Lau SK, Teng JL, Que TL, Yung RW, Luk WK, et al.
retrospective cohort study. Lancet. 2013;381:1551–60. L Hongkongensis study group. Association of Laribacter
http://dx.doi.org/10.1016/S0140-6736(13)60632-7 hongkongensis in community-acquired gastroenteritis with travel
8. Tse H, Bao JY, Davies MR, Maamary P, Tsoi HW, Tong AH, et al. and eating fish: a multicentre case-control study.
Molecular characterization of the 2011 Hong Kong scarlet fever Lancet. 2004;363:1941–7. http://dx.doi.org/10.1016/
outbreak. J Infect Dis. 2012;206:341–51. http://dx.doi.org/10.1093/ S0140-6736(04)16407-6
infdis/jis362 23. Lau SK, Curreem SO, Lin CC, Fung AM, Yuen KY, Woo PC.
9. Brown CC, Olsen RJ, Fittipaldi N, Morman ML, Fort PL, Streptococcus hongkongensis sp. nov., isolated from a patient with
Neuwirth R, et al. Spread of virulent group A Streptococcus type an infected puncture wound and from a marine flatfish.

240 Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015
 E. anophelis Transmission from Mother to Infant

Int J Syst Evol Microbiol. 2013;63:2570–6. http://dx.doi. agmatine deiminase: biochemical characterization of its
org/10.1099/ijs.0.045120-0 catalytic properties and roles in acid tolerance. J Biol Chem.
24. Tse H, Tsoi HW, Leung SP, Lau SK, Woo PC, Yuen KY. 2013;288:26606–15. http://dx.doi.org/10.1074/jbc.M113.477380
Complete genome sequence of Staphylococcus lugdunensis strain 34. Matyi SA, Hoyt PR, Hosoyama A, Yamazoe A, Fujita N,
HKU09–01. J Bacteriol. 2010;192:1471–2. Gustafson JE. Draft genome sequences of Elizabethkingia
http://dx.doi.org/10.1128/JB.01627-09 meningoseptica. Genome Announc. 2013;1:e00444–13.
25. Woo PC, Lau SK, Tse H, Teng JL, Curreem SO, Tsang AK, et al. http://dx.doi.org/10.1128/genomeA.00444-13
The complete genome and proteome of Laribacter hongkongensis 35. Kukutla P, Lindberg BG, Pei D, Rayl M, Yu W, Steritz M, et al.
reveal potential mechanisms for adaptations to different Draft genome sequences of Elizabethkingia anophelis strains R26T
temperatures and habitats. PLoS Genet. 2009;5:e1000416. and Ag1 from the midgut of the malaria mosquito Anopheles
http://dx.doi.org/10.1371/journal.pgen.1000416 gambiae. Genome Announc. 2013;1:e01030–13.
26. Delcher AL, Bratke KA, Powers EC, Salzberg SL. Identifying http://dx.doi.org/10.1128/genomeA.01030-13
bacterial genes and endosymbiont DNA with Glimmer. 36. Balm MN, Salmon S, Jureen R, Teo C, Mahdi R, Seetoh T, et al. Bad
Bioinformatics. 2007;23:673–9. http://dx.doi.org/10.1093/ design, bad practices, bad bugs: frustrations in controlling an outbreak
bioinformatics/btm009 of Elizabethkingia meningoseptica in intensive care units. J Hosp
27. Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, Infect. 2013;85:134–40. http://dx.doi.org/10.1016/j.jhin.2013.05.012
et al. The RAST server: rapid annotations using subsystems 37. Quick J, Constantinidou C, Pallen MJ, Oppenheim B, Loman NJ.
technology. BMC Genomics. 2008;9:75. http://dx.doi.org/ Draft genome sequence of Elizabethkingia meningoseptica isolated
10.1186/1471-2164-9-75 from a traumatic wound. Genome Announc. 2014;2:e00355–14.
28. Liu B, Pop M. ARDB—antibiotic resistance genes database. http://dx.doi.org/10.1128/genomeA.00355-14
Nucleic Acids Res. 2009;37:D443–7.http://dx.doi.org/10.1093/ 38. Sarma S, Kumar N, Jha A, Baveja U, Sharma S. Elizabethkingia
nar/gkn656 meningosepticum: an emerging cause of septicemia in critically
29. Auch AF, Klenk HP, Göker M. Standard operating procedure for ill patients. J Lab Physicians. 2011;3:62–3.
calculating genome-to-genome distances based on high-scoring http://dx.doi.org/10.4103/0974-2727.78575
segment pairs. Stand Genomic Sci. 2010;2:142–8. http://dx.doi. 39. Teo J, Tan SY, Liu Y, Tay M, Ding Y, Li Y, et al.
org/10.4056/sigs.541628 Comparative genomic analysis of malaria mosquito vector-
30. Dussurget O. New insights into determinants of Listeria associated novel pathogen Elizabethkingia anophelis. Genome
monocytogenes virulence. Int Rev Cell Mol Biol. 2008;270:1–38. Biol Evol. 2014;6:1158–65. http://dx.doi.org/10.1093/gbe/evu094
http://dx.doi.org/10.1016/S1937-6448(08)01401-9 40. Holden MT, Feil EJ, Lindsay JA, Peacock SJ, Day NP,
31. Kreft J, Vázquez-Boland JA, Altrock S, Dominguez-Bernal G, Enright MC, et al. Complete genomes of two clinical
Goebel W. Pathogenicity islands and other virulence elements in Staphylococcus aureus strains: evidence for the rapid evolution
Listeria. Curr Top Microbiol Immunol. 2002;264:109–25. of virulence and drug resistance. Proc Natl Acad Sci U S A.
32. Hoffman JA, Badger JL, Zhang Y, Huang SH, Kim KS. 2004;101:9786–91. http://dx.doi.org/10.1073/pnas.0402521101
Escherichia coli K1 aslA contributes to invasion of brain
microvascular endothelial cells in vitro and in vivo.
Address for correspondence: Patrick C.Y. Woo, Department of
Infect Immun. 2000;68:5062–7. http://dx.doi.org/10.1128/
IAI.68.9.5062-5067.2000 Microbiology, The University of Hong Kong, Room 423, University
33. Cheng C, Chen J, Fang C, Xia Y, Shan Y, Liu Y, et al. Listeria Pathology Building, Queen Mary Hospital Compound, Pokfulam Road,
monocytogenes aguA1, but not aguA2, encodes a functional Hong Kong, China; email: [email protected]

Online subscription: wwwnc.cdc.gov/eid/subscribe.htm

Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 2, February 2015 241
This content is in the Public Domain.