Genetic Diversity of Toxigenic and Nontoxigenic …Comparative Genomic Hybridization Bo Pang, 1†...
13
JOURNAL OF BACTERIOLOGY, July 2007, p. 4837–4849 Vol. 189, No. 13 0021-9193/07/$08.000 doi:10.1128/JB.01959-06 Copyright © 2007, American Society for Microbiology. All Rights Reserved. Genetic Diversity of Toxigenic and Nontoxigenic Vibrio cholerae Serogroups O1 and O139 Revealed by Array-Based Comparative Genomic Hybridization Bo Pang, 1 † Meiying Yan, 1 † Zhigang Cui, 1 Xiaofen Ye, 2 Baowei Diao, 1 Yonghong Ren, 2 Shouyi Gao, 1 Liang Zhang, 2 and Biao Kan 1 * State Key Laboratory for Infectious Disease Prevention and Control, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, P.O. Box 5, Changping, Beijing 102206, People’s Republic of China, 1 and National Engineering Research Center for Beijing Biochip Technology, 18 Life Science Parkway, Beijing 102206, People’s Republic of China 2 Received 25 December 2006/Accepted 21 April 2007 Toxigenic serogroups O1 and O139 of Vibrio cholerae may cause cholera epidemics or pandemics. Nontoxigenic strains within these serogroups also exist in the environment, and also some may cause sporadic cases of disease. Herein, we investigate the genomic diversity among toxigenic and nontoxigenic O1 and O139 strains by comparative genomic microarray hybridization with the genome of El Tor strain N16961 as a base. Conservation of the toxigenic O1 El Tor and O139 strains is found as previously reported, whereas accumulation of genome changes was documented in toxigenic El Tor strains isolated within the 40 years of the seventh pandemic. High phylogenetic diversity in nontoxigenic O1 and O139 strains is observed, and most of the genes absent from nontoxigenic strains are clustered together in the N16961 genome. By comparing these toxigenic and nontoxigenic strains, we observed that the small chromosome of V. cholerae is quite conservative and stable, outside of the superintegron region. In contrast to the general stability of the genome, the superintegron demonstrates pronounced divergence among toxigenic and nontoxigenic strains. Additionally, sequence variation in virulence-related genes is found in nontoxi- genic El Tor strains, and we speculate that these intermediate strains may have pathogenic potential should they acquire CTX prophage alleles and other gene clusters. This genome-wide comparison of toxigenic and nontoxigenic V. cholerae strains may promote understanding of clonal differentiation of V. cholerae and contribute to an under- standing of the origins and clonal selection of epidemic strains. Vibrio cholerae is a natural inhabitant in warm aquatic eco- systems (13). No firm understanding is available regarding how many pandemics have been caused by V. cholerae in the course of human events; however, it is generally accepted that at least seven distinct cholera pandemics have occurred since 1817 (2). Of the more than 200 O-antigen serogroups of V. cholerae identified, only O1 and O139 are known to cause epidemic and pandemic cholera. The fifth and sixth pandemics (caused by a classical biotype of the O1 serogroup) were confirmed etiolog- ically, and in 1961 the seventh and present pandemic began, caused by the O1 El Tor biotype (2). In 1992, a novel sero- group, O139, emerged and caused epidemic cholera in Bangla- desh and India (1, 5); the O139 serogroup remains confined to Southeast Asia (49). Cholera toxin is the major virulence factor of toxigenic V. cholerae (31), and colonization of the intestine is the first major step in Vibrio pathogenesis. The lysogenic infection of nontoxi- genic V. cholerae strains by CTX, which carries ctxAB, is an important mechanism for the emergence of new toxigenic strains (54). The toxin-coregulated pilus (TCP) is the receptor for CTX and plays an important role in intestinal coloniza- tion by V. cholerae. The TCP is located in the V. cholerae pathogenicity island (VPI). Other genomic islands, including Vibrio seventh-pandemic island I (VSP-I) and VSP-II, are present in toxigenic O139 and El Tor strains of the seventh pandemic (16). V. cholerae can coexist with a variety of zooplankton and phytoplankton (29), and such survival in aquatic systems is regarded as a prerequisite for cholera pandemics. A report showed that strains with virulence potential (from among the majority of genetically diverse nonpathogenic strains) can be enriched by passage through mammalian host intestines (20). Analysis of diverse V. cholerae strains isolated from different sources has revealed that the strains that can cause diarrhea are highly clonal and that they have evolved independently from different lineages of nontoxigenic environmental isolates. Additionally, it is known that the various nontoxigenic envi- ronmental lineages are composed of heterogeneous popula- tions (20, 32, 50). Several studies have indicated that virulence genes disperse among different serogroups of nontoxigenic strains that are present in aquatic ecosystems and that these aquatic strains can be virulence gene reservoirs for toxigenic strains (20, 21, 42). In order to understand the origins of toxigenic strains, the investigation of genomic differences among V. cholerae strains (especially nontoxigenic strains) will provide insight. The ge- nome contents of toxigenic and nontoxigenic strains have been studied by genome sequencing or comparative genome hybrid- * Corresponding author. Mailing address: State Key Laboratory for Infectious Disease Prevention and Control, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, P.O. Box 5, Changping, Beijing 102206, People’s Republic of China. Phone: 8610 61739458. Fax: 8610 61739156. E-mail: [email protected]. † B.P. and M.Y. contributed equally to this work. Published ahead of print on 27 April 2007. 4837 on May 24, 2020 by guest http://jb.asm.org/ Downloaded from
Transcript of Genetic Diversity of Toxigenic and Nontoxigenic …Comparative Genomic Hybridization Bo Pang, 1†...
JOURNAL OF BACTERIOLOGY, July 2007, p. 4837–4849 Vol. 189, No. 130021-9193/07/$08.00�0 doi:10.1128/JB.01959-06Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Genetic Diversity of Toxigenic and Nontoxigenic Vibrio choleraeSerogroups O1 and O139 Revealed by Array-Based
Comparative Genomic Hybridization�
Bo Pang,1† Meiying Yan,1† Zhigang Cui,1 Xiaofen Ye,2 Baowei Diao,1 Yonghong Ren,2Shouyi Gao,1 Liang Zhang,2 and Biao Kan1*
State Key Laboratory for Infectious Disease Prevention and Control, National Institute for Communicable Disease Controland Prevention, Chinese Center for Disease Control and Prevention, P.O. Box 5, Changping, Beijing 102206,
People’s Republic of China,1 and National Engineering Research Center for Beijing Biochip Technology,18 Life Science Parkway, Beijing 102206, People’s Republic of China2
Received 25 December 2006/Accepted 21 April 2007
Toxigenic serogroups O1 and O139 of Vibrio cholerae may cause cholera epidemics or pandemics. Nontoxigenicstrains within these serogroups also exist in the environment, and also some may cause sporadic cases of disease.Herein, we investigate the genomic diversity among toxigenic and nontoxigenic O1 and O139 strains by comparativegenomic microarray hybridization with the genome of El Tor strain N16961 as a base. Conservation of the toxigenicO1 El Tor and O139 strains is found as previously reported, whereas accumulation of genome changes wasdocumented in toxigenic El Tor strains isolated within the 40 years of the seventh pandemic. High phylogeneticdiversity in nontoxigenic O1 and O139 strains is observed, and most of the genes absent from nontoxigenic strainsare clustered together in the N16961 genome. By comparing these toxigenic and nontoxigenic strains, we observedthat the small chromosome of V. cholerae is quite conservative and stable, outside of the superintegron region. Incontrast to the general stability of the genome, the superintegron demonstrates pronounced divergence amongtoxigenic and nontoxigenic strains. Additionally, sequence variation in virulence-related genes is found in nontoxi-genic El Tor strains, and we speculate that these intermediate strains may have pathogenic potential should theyacquire CTX prophage alleles and other gene clusters. This genome-wide comparison of toxigenic and nontoxigenicV. cholerae strains may promote understanding of clonal differentiation of V. cholerae and contribute to an under-standing of the origins and clonal selection of epidemic strains.
Vibrio cholerae is a natural inhabitant in warm aquatic eco-systems (13). No firm understanding is available regarding howmany pandemics have been caused by V. cholerae in the courseof human events; however, it is generally accepted that at leastseven distinct cholera pandemics have occurred since 1817 (2).
Of the more than 200 O-antigen serogroups of V. choleraeidentified, only O1 and O139 are known to cause epidemic andpandemic cholera. The fifth and sixth pandemics (caused by aclassical biotype of the O1 serogroup) were confirmed etiolog-ically, and in 1961 the seventh and present pandemic began,caused by the O1 El Tor biotype (2). In 1992, a novel sero-group, O139, emerged and caused epidemic cholera in Bangla-desh and India (1, 5); the O139 serogroup remains confined toSoutheast Asia (49).
Cholera toxin is the major virulence factor of toxigenic V.cholerae (31), and colonization of the intestine is the first majorstep in Vibrio pathogenesis. The lysogenic infection of nontoxi-genic V. cholerae strains by CTX�, which carries ctxAB, is animportant mechanism for the emergence of new toxigenicstrains (54). The toxin-coregulated pilus (TCP) is the receptor
for CTX� and plays an important role in intestinal coloniza-tion by V. cholerae. The TCP is located in the V. choleraepathogenicity island (VPI). Other genomic islands, includingVibrio seventh-pandemic island I (VSP-I) and VSP-II, arepresent in toxigenic O139 and El Tor strains of the seventhpandemic (16).
V. cholerae can coexist with a variety of zooplankton andphytoplankton (29), and such survival in aquatic systems isregarded as a prerequisite for cholera pandemics. A reportshowed that strains with virulence potential (from among themajority of genetically diverse nonpathogenic strains) can beenriched by passage through mammalian host intestines (20).Analysis of diverse V. cholerae strains isolated from differentsources has revealed that the strains that can cause diarrheaare highly clonal and that they have evolved independentlyfrom different lineages of nontoxigenic environmental isolates.Additionally, it is known that the various nontoxigenic envi-ronmental lineages are composed of heterogeneous popula-tions (20, 32, 50). Several studies have indicated that virulencegenes disperse among different serogroups of nontoxigenicstrains that are present in aquatic ecosystems and that theseaquatic strains can be virulence gene reservoirs for toxigenicstrains (20, 21, 42).
In order to understand the origins of toxigenic strains, theinvestigation of genomic differences among V. cholerae strains(especially nontoxigenic strains) will provide insight. The ge-nome contents of toxigenic and nontoxigenic strains have beenstudied by genome sequencing or comparative genome hybrid-
* Corresponding author. Mailing address: State Key Laboratory forInfectious Disease Prevention and Control, National Institutefor Communicable Disease Control and Prevention, Chinese Centerfor Disease Control and Prevention, P.O. Box 5, Changping, Beijing102206, People’s Republic of China. Phone: 8610 61739458. Fax: 861061739156. E-mail: [email protected].
† B.P. and M.Y. contributed equally to this work.� Published ahead of print on 27 April 2007.
4837
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
ization (CGH); one nontoxigenic O1 strain and four nontoxi-genic non-O1/non-O139 strains were included in these studies(16, 17). These investigations revealed that genomic contentsare quite conservative in O1 and O139 toxigenic strains (in-cluding a nontoxigenic TCP� strain) (16), whereas four patho-genic non-O1/non-O139 strains diverge from O1 and O139strains (17).
Serogroup transformation between O1 and non-O1 may nothappen frequently in the environment (20); hence, genomicanalysis among nontoxigenic serogroup O1 and O139 strainsmay provide further insight into the origins of the seventh-pandemic strains. Until now, only one nontoxigenic O1 strainwas analyzed by CGH (16). In the present study, we con-structed microarrays based on the genome sequence of El Torstrain N16961 to investigate the genomic difference of V. chol-erae and we used seven nontoxigenic O1 strains and threeO139 strains. We found that the nontoxigenic O1 and O139strains are demonstrably different from toxigenic strains. Wespeculate that the nontoxigenic strains carrying different tcpAalleles may represent different intermediate strains in the orig-ination of toxigenic strains which would be different from ex-tant toxigenic clones.
MATERIALS AND METHODS
Strains. The strains used in this study are listed in Table 1. Except for thestrains isolated in Bangladesh and India, toxigenic O1 El Tor strains wereisolated from three cholera epidemics in China during the seventh cholerapandemic (from 1961 to 2005). Serogroup O139 strains were isolated in differentChinese regions and in different years since their emergence in 1992. MO45 is anon-China exception that was isolated in India. For nontoxigenic O1 and O139
strains, strains with extensive differences in source and year of isolation wereselected. Four nontoxigenic El Tor strains containing TCP clusters were alsoincluded in this study.
Chromosomal DNA preparation. All strains were grown in 5 ml Luria-Bertanibroth to an optical density at 600 nm of 0.8. Chromosomal DNA was preparedwith NucleoSpin Tissue kits (Macherey-Nagel, Germany) according to the man-ufacturer’s protocol.
Construction of microarrays. DNA microarrays were constructed on the basisof the genomic content of N16961 (GenBank accession no. NC002505 andNC002506). An in-house high-throughput computer algorithm was used to de-sign PCR primers for all open reading frames (ORFs). The fundamental rules ofour computer algorithm dictates that all primers have a 61°C annealing temper-ature, PCR product lengths should be 200 to 1,000 bp, and amplicons shouldhave minimum sequence similarity. PCR amplifications were performed in100-�l reaction mixture volumes. A total of 3,687 ORFs were amplified success-fully for N16961. Products were confirmed by agarose gel electrophoresis. PCRproducts were precipitated with isopropanol and redissolved in 50% dimethylsulfoxide. For the 174 genes that were not successfully amplified by PCR, 70-mersense oligonucleotides were synthesized. PCR products and long oligonucleo-tides (in 50% dimethyl sulfoxide, 20 �M) were spotted with a SmartArrayMicroarrayer (CapitalBio Corp., Beijing, China). Each gene was printed intriplicate to facilitate subsequent statistical analyses. After printing, the slideswere baked at 80°C for 1 h and stored dry at room temperature until used. In all,the microarray contained 3,861 predicted ORFs representing 99.4% (3,861/3,885) of the genome of N16961, of which 2,758 ORFs are located on chromo-some 1 and 1,103 are on chromosome 2.
Prior to hybridization, the slides were rehydrated over a 65°C water bath for10 s and UV cross-linked at 250 mJ/cm2. The unimmobilized DNA was washedoff with 0.5% sodium dodecyl sulfate (SDS) for 15 min at room temperature, andthe SDS was removed by dipping the slides in anhydrous ethanol for 30 s. Theslides were spin dried at 1,000 rpm for 2 min.
Hybridization. Genomic DNA was fragmented by Dpn restriction endonucle-ase digestion and purified with the PCR Clean-up NucleoSpin Extract kit (Ma-cherey-Nagel, Germany). For each labeling reaction, 2 �g of digested DNA and4 �g of a random nonamer were heated to 95°C for 3 min and snap cooled on ice.
TABLE 1. Strains used for comparative analysis
Strain Serogroup, biotype Serotype Yr isolated Where isolated Origin ctxAB No. of genesabsent
No. of genes absentin superintegron
V0518 O1, El Tor Ogawa 2005 China Shrimp � 199 51SD95001 O1, El Tor Inaba 1995 China Water � 203 6093097 O1, El Tor Ogawa 1993 China Patient � 199 52V0550 O1, El Tor Inaba 2005 China Patient � 309 6619-22 O1, El Tor Ogawa 1977 China Patient � 244 17JS32 O1, El Tor Inaba 1990 China Patient � 309 1117743 O1, El Tor Ogawa 1977 China Water � 112 28FJ147 O1, El Tor Inaba 2005 China Patient � 34 16ZJ0596 O1, El Tor Inaba 2005 China Patient � 34 16V011506 O1, El Tor Inaba 2001 China Patient � 16 16200106 O1, El Tor Inaba 2001 China Water � 16 1693284 O1, El Tor Ogawa 1993 China Patient � 16 16FJ80004 O1, El Tor Inaba 1980 China Patient � 0 0D118 O1, El Tor Ogawa 1961 China Patient � 27 06312 O1, El Tor Ogawa 1961 Indonesia Patient � 0 0N16961 O1, El Tor Inaba 1975 Bangladesh Patient � 0 079005 O1, El Tor Ogawa 1979 China Patient � 0 0Wujiang-2 O1, El Tor Inaba 1980 China Patient � 0 1FJ0299357 O139 1999 China Water � 71 2193010 O139 1993 China Patient � 72 2293209 O139 1993 China Carrier � 74 22MO45 O139 1993 India Patient � 71 2198106 O139 1998 China Patient � 72 23200306 O139 2003 China Patient � 73 23SCH04082 O139 2004 China Water � 592 50B4 O139 1994 China Patient � 599 2394001 O139 1994 China Water � 405 105569B O1, classical Inaba 1948 India Patient � 94 35O395 O1, classical Ogawa 1964 India Patient � 77 29ZHE66 O1, classical Ogawa 1963 China Carrier � 143 27
4838 PANG ET AL. J. BACTERIOL.
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
FIG. 1. ORFs absent from or present in the test strains compared with N16961. Final absent and present calls for each gene were translatedto a binary code and analyzed with Cluster/TreeView. For each strain, a black line indicates the presence of a gene whereas a white line indicatesits absence. The GC content curve is shown on the right.
VOL. 189, 2007 ARRAY-BASED CGH ANALYSIS OF VIBRIO CHOLERAE 4839
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
TA
BL
E2.
Sum
mar
yof
abse
ntge
nes
acco
rdin
gto
cate
gory
inte
stst
rain
sco
mpa
red
tost
rain
N16
961
Stra
in
No.
ofge
nes
info
llow
ing
cate
gory
mis
sing
:
Tot
alC
ell
enve
lope
Cel
lula
rpr
oces
ses
Cen
tral
inte
r-m
edia
rym
etab
-ol
ism
Dis
rupt
edre
adin
gfr
ame
DN
Am
etab
-ol
ism
Ene
rgy
met
ab-
olis
m
Fat
tyac
idan
dph
osph
o-lip
idm
etab
-ol
ism
Hyp
o-th
etic
alpr
otei
ns
Mob
ilean
dex
trac
hro-
mos
omal
elem
ent
func
tions
Prot
ein
fate
Prot
ein
synt
hesi
s
Puri
nes,
pyri
mid
ines
,nu
cleo
side
s,an
dnu
cleo
tides
Reg
ulat
ory
func
tions
Tra
n-sc
ript
ion
Tra
nspo
rtan
dbi
ndin
gpr
otei
ns
Unk
now
nfu
nctio
n
Am
ino
acid
bios
yn-
thes
is
Bio
synt
hesi
sof
cofa
ctor
s,pr
osth
etic
grou
ps,a
ndca
rrie
rs
ElT
orC
TX
�
V05
181
116
05
30
268
10
011
33
121
00
199
SD95
001
212
60
53
027
80
00
103
312
40
020
393
097
113
80
53
026
80
00
103
212
00
019
9V
0550
1136
52
67
136
152
00
142
1315
90
030
919
-22
1132
32
46
028
122
00
112
1211
81
024
4JS
328
339
35
31
4315
00
17
33
175
00
309
7743
18
20
24
114
72
00
61
955
00
112
ElT
orC
TX
�
D11
81
20
00
00
51
00
02
10
150
027
6312
00
00
00
00
00
00
00
00
00
079
005
00
00
00
00
00
00
00
00
00
0F
J800
040
00
00
00
00
00
00
00
00
00
Wuj
iang
-20
00
00
00
00
00
00
00
00
00
9328
40
01
00
00
30
00
00
00
120
016
V01
1506
00
10
00
03
00
00
00
012
00
1620
0106
00
10
00
03
00
00
00
012
00
16F
J147
11
10
00
05
00
00
11
024
00
34Z
J059
61
11
00
00
50
00
01
10
240
034
O13
9C
TX
�
9301
06
04
03
52
150
00
01
03
330
072
9320
96
04
03
52
150
00
01
03
350
074
9810
66
04
03
52
150
00
01
03
330
072
2003
066
04
03
52
150
00
01
03
340
073
FJ0
2993
576
04
03
52
150
00
01
03
320
071
MO
456
04
03
52
140
00
01
03
330
071
O13
9C
TX
�
SCH
0408
224
6015
110
285
7115
101
243
552
249
01
592
B4
2567
121
1130
572
1410
12
425
5224
90
159
994
001
1243
143
411
446
132
00
244
1820
41
240
5
Cla
ssic
al56
9B1
53
01
11
122
00
03
20
630
094
O39
51
52
01
10
112
30
00
10
500
077
ZH
E66
15
20
49
221
20
01
121
571
34
143
4840 PANG ET AL. J. BACTERIOL.
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
Reaction buffer (10�), deoxynucleoside triphosphates, and Cy5-dCTP or Cy3-dCTP (Amersham Pharmacia Biotech, Piscataway, NJ) at final concentrations of120 �M each dATP, dGTP, and dTTP; 60 �M dCTP; and 40 �M Cy dye,respectively, were added. Klenow enzyme (1 �l; TaKaRa) was added, and themixture was incubated at 37°C for 60 min. The labeled DNA was purified withthe PCR Clean-up NucleoSpin Extract II kit (Macherey-Nagel) and resuspendedin elution buffer. The optical density at 260 nm was read. Labeled control andtest samples were quantitatively adjusted on the basis of the efficiency of Cy dyeincorporation and mixed into 30 �l hybridization solution (3� SSC [1� SSC is0.15 M NaCl plus 0.015 M sodium citrate], 0.2% SDS, 25% formamide, 5�Denhardt’s solution). DNA in the hybridization solution was denatured at 95°Cfor 3 min prior to loading onto a microarray. The microarray was hybridized at42°C overnight and washed at room temperature with two consecutive washingsolutions (0.2% SDS–2� SSC at 42°C for 5 min and 0.2% SSC for 5 min).
Imaging and data analysis. Microarrays were scanned with a LuxScan scan-ner (CapitalBio Corp.), and images were analyzed with SpotData software(CapitalBio Corp.). Analysis of CGH results was also accomplished by de-termination of fluorescence intensity ratios via a space- and intensity-depen-dent normalization procedure based on a LOWESS program. For each ex-periment, each ORF was spotted in triplicate onto each slide.
Classification of ORFs as absent or present in test strains. For each strain,ORFs with ratios lower than 0.2 were classified as absent whereas those with aratio greater than 0.6 were regarded as present. The presence or absence of anORF with ratios between 0.2 and 0.6 was determined by PCR confirmationassays to validate cutoff values as follows. About 80% of such ORFs in eachstrain were selected for a PCR assay. The positive cutoff was the highest ratio inall PCR-negative ORFs, while the negative cutoff was the lowest ratio of thePCR-positive ORF. Between the positive and negative cutoff values, some ORFswere PCR positive and others were negative; the PCR results did not correspondexactly to the highness and lowness of ratios. Finally, every ORF with a ratiobetween the positive and negative cutoff values was assayed by PCR and classi-fied according to its PCR result. The cutoff values in the various test strainsdiffered from each other because of sequence divergence between strains. Foroligonucleotide probe ORFs with ratios between positive and negative cutoffvalues, the absence or presence of the ORFs was decided according to the PCRresults obtained with primers derived from flanking ORFs.
Comparative phylogenomic analysis of microarray data. Final absent orpresent calls for each gene were translated into a binary code (gene present in agiven strain � 2, absent � 0.5) and analyzed by Cluster/TreeView. Microarraydata were processed with a practical extraction and reporting language proce-dure, and trees based on the unweighted-pair group method using averagelinkages (UPGMA) were constructed with BioNumerics (version 3.0; AppliedMaths BVBA, Belgium). The correlations between these trees were calculatedwith BioNumerics.
RESULTS
Overview of the array-based CGH results. Of the 3,773ORFs surveyed, 2,876 were common to all strains, representingthe functional core of the V. cholerae genome. However, 897(23.8%) ORFs were variably present in the strains but thestrains varied greatly from each other with respect to the num-ber of missing ORFs. The majority of the absent ORFs wereclustered together in regions of the N16961 genome that havea low GC content, suggesting possible gene transfer (Fig. 1).
The number of variant or missing genes in toxigenic El Torstrains is much lower (missing 0 to 34 genes) than in the otherstrains. Thus, the El Tor strains exhibit the highest genomicconservation and gene numbers, consistent with previous re-search (16), and the same conservation also exists in toxigenicO139 strains.
The large number of ORFs (112 to 599) missing from non-toxigenic strains is associated with a wide variety of diversecellular functions. The variation in ORFs among the differentV. cholerae strains that were classified with respect to func-tional categories is shown in Table 2. Of the 2,876 commongenes, the conservation categories are protein and amino acidbiosynthesis (100 and 99.9%, respectively) in toxigenic El Torstrains (Fig. 2). Genes in the category of mobile and extra-chromosomal element function exhibited high diversity, fol-lowed by those for cellular processes and central intermediarymetabolism function, especially in nontoxigenic strains. In themobile and extrachromosomal element category, subcategoriesthat contain phage- and transposon-related genes appeared tobe the most variable, suggesting a role in the diversification ofV. cholerae. Of the 33 ORFs in the cellular process categorythat exhibit differences between toxigenic and nontoxigenicstrains, 29 are pathogenesis-related genes.
The fluorescence intensity ratios of each gene of the toxi-genic strains to N16961 were distributed around 1.0 and wereclustered more narrowly than those of nontoxigenic strains.
FIG. 2. Variability of ORFs of different functional categories in different groups of V. cholerae strains. Each column in each category representsthe ratio of the average number of absent genes within each V. cholerae group to the total number of genes in this category of N16961.
VOL. 189, 2007 ARRAY-BASED CGH ANALYSIS OF VIBRIO CHOLERAE 4841
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
Among nontoxigenic strains, the overall percentages of ratiosof ORFs higher than 1.4 or lower than 0.4 in strains are greaterthan those of toxigenic strains, revealing that more sequenceheterogeneity exists in nontoxigenic strains (Fig. 3).
A phylogenetic tree of the test strains was also generated onthe basis of the microarray data (Fig. 4). All of the test strainscould be divided into seven groups in accordance with sero-groups, biotypes, TCP gene cluster possession, and tcpA allelesof El Tor derived or not, i.e., toxigenic El Tor strains, toxigenicO139 strains, classical strains, nontoxigenic TCP� El Torstrains with an El Tor-type tcpA sequence (designated TCPET),nontoxigenic TCP� El Tor strains with tcpA different fromthe El Tor type (designated TCPETvar), nontoxigenic TCP�
El Tor strains, and nontoxigenic O139 strains. In the groupof TCPETvar El Tor strains, the tcpA DNA sequences aredifferent from those of the currently epidemic El Tor andO139 strains (GenBank accession no. AF512425, AF512423,and AY052831), and these strains were negative for thepresence of the tcpA gene in microarray analyses.
Superintegron diversity and chromosome 2. The superinte-gron present in chromosome 2 of V. cholerae is an active geneacquisition structure with frequent lateral gene transfers thatare independent of the entire genome (25). Only five toxigenicEl Tor strains, D118, 6312, 79005, FJ80004, and Wujiang-2,exhibited the same superintegron content as N16961, whereasall of the other strains showed diversity in this region.
Considering the specific characteristics of the superintegron
and its diversity, we analyzed variation in the superintegronand its impact on chromosome 2, as well as on the whole V.cholerae genome, by establishing phylogenetic trees with andwithout the superintegron. For this purpose, we constructedphylogenetic trees of the test strains based on the comparativehybridization data of the whole genome without the superinte-gron (I�II�SI) (Fig. 5B), of chromosome 2 alone (Fig. 5C), ofchromosome 2 without the superintegron (II�SI) (Fig. 5A),and of the superintegron (SI) alone (Fig. 5D). The indexes ofsimilarity between SI and I�II�SI and between SI and II�SIwere found to be only 27.9 and 0%, respectively, indicating theinstability of the superintegron and demonstrating its differentevolutionary rate compared to the rest of genome. The simi-larity index also revealed frequent gene losses and gains in thesuperintegron, as well as recent evolution within this group ofbacteria.
The index of similarity between the cluster analyses of chro-mosome 2 (II) and chromosome 1 (I) is 91.6%, and the indexof similarity between chromosome 2 excluding the superinte-gron (II�SI) and chromosome 1 (I) is 93.7% (Fig. 5E). Theseresults indicate that chromosome 2 without the superintegronhas almost the same evolutionary rate as chromosome 1 if thesuperintegron region is excluded from consideration. The in-dexes of similarity between clusters (SI) and the whole genomeand between SI and II are 45.5 and 59.3%, respectively. In thephylogenetic trees of the genomes without the superintegron,all of the toxigenic strains cluster in the same branching point,
FIG. 3. Distribution of ORF hybridization signal ratios of test V. cholerae strains. The divergence of different strains from N16961 can be seen.Panels: A, ratio distribution; B, proportions of genes in different ratio ranges compared to the total number of genes in the microarray. At thebottom of panel B, the name of each strain is listed. All ratios higher than 1.6 were treated as 1.6.
4842 PANG ET AL. J. BACTERIOL.
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
similar to the tree based on sequences of multiple housekeep-ing genes (Fig. 6). This finding indicates that chromosome 2without the superintegron is conserved and stable, as is chro-mosome 1.
Comparison of serogroup O1 strains. (i) Toxigenic El Torstrains. We observed conservation of genome content amongthe 11 toxigenic strains examined though they originated fromdiverse geographic locations and in different years (1961 to2005). However, gene changes were observed among thesestrains according to the year of isolation. Thus, toxigenic ElTor from China can be divided into four subgroups that gen-erally correspond to the times in which they were isolated (Fig.4). The first subgroup contains a single strain, D118, isolatedfrom a patient in 1961 when the seventh pandemic started. Theonly difference between D118 and other toxigenic El Torstrains was the absence of intact VSP-II in D118, a DNAregion considered specific to seventh-pandemic strains. Thesecond subgroup consists of five strains isolated between 1961and 1980; within this group, strains 6312, 79005, Wujiang-2,and FJ80004 have the same genome content as N16961. Thethird subgroup consists of strains 93284, V011506, and 200106,which were isolated between 1993 and 2001; these strains aresimilar to N16961 but lack 16 ORFs. The fourth subgroupencompasses FJ147 and ZJ0596, which were isolated duringthe epidemic in 2005; 35 ORFs are absent from these twostrains. Of the 35 absent ORFs, 18 (VC0495 to VC0512) are inVSP-II.
To further examine this phenomenon of missing ORFs, wesequenced the genome region corresponding to this missingfragment in FJ147 and ZJ0596. A new fragment was insertedbetween VC0495 and VC0513 (Fig. 7). BLAST analysis of thisregion revealed that it contained an IS element carrying atransposase with 100% identity to transposases in the N16961
genome; additionally, the flanking sequences suggest that thisinsertion was a random event.
(ii) Nontoxigenic El Tor strains. Nontoxigenic strains exhib-ited much higher diversity than toxigenic strains. The numbersof ORFs absent ranged from 112 (3.0% of the 3,773 ORFsdetected in the array) for strain 7743 to 309 (8.2%) for strainJS32. These strains were divided into three clusters, (i) non-toxigenic TCPET strains (cluster I, strain 7743), (ii) nontoxi-genic TCPETvar strains (cluster II, strains 93097, V0518, andSD95001), and (iii) nontoxigenic TCP� strains (cluster III,strains JS32, 19-22, and V0550, which do not harbor VPI).These three clusters comprise different branches in the phylo-genetic tree (Fig. 4). The number of ORFs absent from theseclusters compared to N16961 increased from cluster I to clus-ter III (Table 2).
Cluster I includes one environmental strain, 7743, that has aVPI region with the same tcpA gene as toxigenic El Tor strains.In several biochemical tests, including sensitivity to the V.cholerae typing phages, sorbitol and mannitol fermentation,and hemolysis activity, strain 7743 shows the same character-istics as toxigenic El Tor strains but differs from the nontoxi-genic strains (data not shown). A total of 112 ORFs wereabsent from strain 7743. In comparison, on average, only 12ORFs were absent from toxigenic El Tor strains and 273 ORFswere absent from clusterII and III nontoxigenic El Tor strains.Moreover, in the phylogenetic tree, 7743 is clustered togetherwith toxigenic strains rather than with other nontoxigenic ones,and the cluster III strains diverge strongly from the epidemicstrains (Fig. 4).
Aside from VPI, we also found sequence variations betweencluster II and cluster III nontoxigenic strains in other genomeregions. One gene cluster, VC0919 to VC0925, containsexopolysaccharide (EPS) biosynthesis genes. This region was
FIG. 4. Phylogenetic structure (UPGMA tree) according to CGH data of the 30 strains of V. cholerae. Final absent and present calls for eachgene of each strain were translated to 0 and 1, respectively. The UPGMA tree shown was calculated from difference matrices with BioNumerics.The seven groups of strains are marked with different colors. Partial magnifications of the toxigenic El Tor strain and toxigenic O139 strain groupsare shown at the right.
VOL. 189, 2007 ARRAY-BASED CGH ANALYSIS OF VIBRIO CHOLERAE 4843
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
FIG. 5. Phylogenetic structures (UPGMA trees) of the strains according to microarray data of (A) chromosome 2 excluding the superintegron(II�SI), (B) the whole genome excluding the superintegron (I�II�SI), (C) chromosome 2 (II), (D) the superintegron (SI), and (E) chromosome 1 (I).The indexes of similarity between different phylogenetic trees were calculated by Dice tests with BioNumerics and are listed in panel F. The strain codesrepresent toxigenic El Tor strains, group A, including 6312 (A1), 79005 (A2), FJ80004 (A3), N16961 (A4), D118 (A5), Wujiang-2 (A6), 93284 (A7),V011506 (A8), 200106 (A9), FJ147 (A10), and ZJ0596 (A11); toxigenic O139 strains, group B, including MO45 (B1), 93010 (B2), 93209 (B3), 98106 (B4),FJ0299357 (B5), and 200306 (B6); classical strains, group C, including 569B (C1), O395 (C2), and ZHE66 (C3); nontoxigenic TCPET El Tor strains, groupD, including 7743 (D1); nontoxigenic TCPETvar strains, group E, including 93097 (E1), SD95001 (E2), and V0518 (E3); nontoxigenic TCP� El Tor strains,group F, including 19-22 (F1), JS32 (F2), and V0550 (F3); and nontoxigenic O139 strains, group G, including 94001 (G1), B4 (G2), and SCH04082 (G3).
4844
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
absent from cluster III nontoxigenic El Tor strains but presentin cluster II nontoxigenic El Tor strains by hybridization. How-ever, amplicons with the same size as that of N16961 in clusterIII nontoxigenic El Tor strains were obtained by using primersfor the flanking ORFs of VC0919 to VC0925. Restriction en-zyme fragment lengths of PCR products of this region revealeddifferent patterns for cluster III nontoxigenic El Tor strainsand N16961. Additionally, cluster III strains also varied among
themselves. These data indicate that multiple new sequencealleles exist in this region. Since the EPS are involved in biofilmformation and motility, as well as in the secretion of hemolysinand CT, in V. cholerae (34), variance of these genes mightsuggest differences in EPS roles and environmental survivalamong nontoxigenic TCP�, TCP�, and toxigenic El Torstrains.
Several other virulence-related genes, including ompU andpilA, also exhibited sequence variation within cluster I nontoxi-genic strains compared to seventh-pandemic strain N16961. Inthe array analysis, these genes in cluster III strains were shownto be absent whereas adjacent genes were present. We se-quenced the corresponding regions and found that the ompUsequences in cluster III strains diverged at the nucleotide level.Thus, the sequence identity was 66.4% (strains 19-22 andV0550, 73.7% amino acid sequence identity) and 74.1% (strainJS32, 79.5% amino acid sequence identity) with respect toN16961. OmpU is a pore-forming protein of the outer mem-brane of V. cholerae. Since V. cholerae lacking ompU has re-duced tolerance to bile salts, it is possible that strains withvariant ompU alleles differ in environmental survival ability(33). Similarly, pilA of cluster III strains was only 59.8% iden-tical, at the nucleotide level (11 to 49% amino acid sequenceidentity only), to pilA of N16961. Because pilA contributes tothe adherence and colonization abilities of bacteria (51, 18),such allelic variation may have significance for colonization.The DNA sequences of ompU and pilA of the cluster II strainsand strain 7743 were identical to the sequence of N16961.
Another region where sequence variation was observed be-tween cluster II and III nontoxigenic El Tor strains was thesuperintegron. Other differences were found. For instance,part of VPI-II (VC1773 to VC1788) is present in cluster III ofnontoxigenic El Tor strains but absent from cluster II strains.Likewise, fragments of VC1748 to VC1753 and VCA1042 toVCA1044 are present in cluster II strains but absent fromcluster III strains. VC1748 to VC1753 encode pqiA and fivehypothetical proteins, while VCA1042 to VCA1044 encode theCcm2-related protein and the TagE protein.
On the basis of the array constructed with the genome con-tent of N16961, in all tested toxigenic and nontoxigenic El Torstrains, a total of 455 ORFs were missing from various El Torstrains. Of these, 31 are commonly found only in toxigenic ElTor strains. Thirty of these common genes are genomic island
FIG. 6. Phylogenetic structure (UPGMA tree) of the strains ac-cording to the sequence analysis of four housekeeping genes, i.e., recA,mdh, gyrB, and dnaE.
FIG. 7. Schematic of insert fragment in VSPII of predominant toxigenic El Tor strains isolated in cholera epidemic regions in 2005. ORFs ofVC0496 to VC0512 were deleted and replaced by sequence encoding transposase subunits A and B; a 118-bp fragment of VC0495 was left. In theupstream and downstream regions of the inserted transposase sequence lie new 8- and 38-bp insertion sequences, respectively. An 8-bp invertedrepeat sequence is present in italic capital letters. The ORF numbers are the same as those in N16961. The insertion sequence is indicated as agray bold vertical line. VC0495� represents the residual 118 bp of VC0495. A palindromic sequence (in italics and capitalized) lies at both endsof the insertion sequence. An 857-bp fragment of the 1,455-bp intergenic sequence between VC0512 and VC0513 was left, and a new 51-bpfragment (AAAAATACAGAATATAAATACTGAAATAATATCTGAGATTAAAATAAAAAA) with 7-bp AT-rich reverse repeats (under-lined) at both ends is inserted at bp 347 upstream of the VC0513 starting position.
VOL. 189, 2007 ARRAY-BASED CGH ANALYSIS OF VIBRIO CHOLERAE 4845
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
components, including VSP-I, CTX�, and the superintegron.The remaining ORF, VC2346, is annotated as encoding theSmp protein, but its function is unknown.
Comparison of genome contents of serogroup O139 strains.(i) Differences between the genome contents of toxigenic ElTor and O139 strains. It was likely that the serogroup O139 V.cholerae strains originated from an El Tor-like strain (53), andthe results of our whole-genome content analysis support thisidea (Fig. 4). Compared to toxigenic El Tor strains, 77 ORFswere absent or highly divergent in all toxigenic O139 strainsand most of them are clustered in three genomic islands and inthe gene cluster encoding the O antigen. The first island rep-resents 83% of the gene cluster encoding enzymes involved inlipopolysaccharide antigen synthesis (Fig. 8). The second is-land spans the region from VC1761 to VC1787 and consists oftwo parts of VPI-II, including an ORF that encodes neuramin-idase. Neuraminidase may increase host cell susceptibility tocholera toxin (23). Possible explanations for the reemergenceof El Tor cholera in Bangladesh since 1994 (19) are the pres-ence of VPI-II in the El Tor biotype and the deletion of mostof VPI-II from O139 strains (30). The third island is the su-perintegron in which 19 ORFs are missing from toxigenic O139strains. Herein we show that the genes missing from the su-perintegron of toxigenic O139 strains are almost the same asthose of toxigenic El Tor strains isolated after 1993 (93284,200106, V011506, ZJ0596, and FJ147). Moreover, the firstthree toxigenic El Tor strains are clustered more closely withtoxigenic O139 strains in the phylogenetic tree based on thesuperintegron hybridization profile (Fig. 5D), suggesting aclose relationship between toxigenic O139 and El Tor strainsprevailing at that time.
On the basis of the N16961 genome content array, the toxi-genic O139 strains isolated from 1993 to 2003 showed similarcontents and clustered together (Fig. 4). Although O139 chol-era has existed for more than 10 years, it remains limited toSoutheast Asia and the strains have no obvious divergences intheir genomes.
(ii) Divergence of nontoxigenic O139 strains. In contrast tothe conservation of toxigenic O139 strains, the nontoxigenicO139 strains showed more diversity in both the number andcategories of absent ORFs (Table 2). These absent ORFs aredispersed on the two chromosomes. The toxigenic and non-toxigenic O139 strains have the same microarray hybridizationpatterns in the gene cluster involved in lipopolysaccharide an-tigen synthesis, confirming the genetic basis of the serogroupantigen of these strains (Fig. 8). Nonetheless, in the phylo-genetic tree, the nontoxigenic O139 strains are scatteredthroughout the tree, indicative of the large genetic diversity ofgenomic content among these strains.
DISCUSSION
In this study, array-based CGH was used to investigate thegenome diversity of serogroup O1 and O139 V. cholerae, in-cluding toxigenic and nontoxigenic strains. In the comparison,we used seven nontoxigenic O1 and three nontoxigenic O139strains. Analysis of the difference in genome components ofnontoxigenic strains compared to toxigenic strains may facili-tate tracing of the origin and evolution of seventh cholerapandemic strains.
Conservation of chromosome 2 and impact of the superinte-gron on strain diversification. It is unclear why V. cholerae hastwo chromosomes. Sequence comparison of chromosomes 1and 2 (strain N16961) indicates that chromosome 2 was orig-inally a megaplasmid captured by an ancestral Vibrio species;perhaps chromosome 2 provides certain evolutionarily selec-tive advantages for the bacteria (25). Our examination of mul-tiple strains (including nontoxigenic O1 and O139 strains) in-dicates that the proportion of absent genes in chromosome 2 ishigher than is the case in chromosome 1. Our analysis alsoshows that the majority of missing genes are located within thesuperintegron, a region of more than 200 ORFs. Phylogenetictrees of these strains, based on chromosome 2 but excludingthe superintegron, reveal that the gene content of chromosome2 is as conservative as that of chromosome 1. Therefore, ourCGH data suggest that chromosome 2 (aside from the super-integron) is conservative in V. cholerae. The presumed reasonthat the phylogenetic trees for chromosomes 1 and 2 do notcorrespond strictly is that the superintegron provides variabil-ity in genome content.
Phylogenetic trees based on microarray hybridization resultsof chromosome 2 and the superintegron alone revealed evo-lutionary trends (shared or otherwise) between toxigenic O1and O139 strains and between toxigenic O1 El Tor strains. Thefrequent gene gains and losses in the superintegron, charac-teristic of such integrons, represent recent genetic change andactive gene transfer. It has been reported that microevolutionis an ancient and perpetual event in Vibrionaceae resulting inthe loss, gain, and duplication of genes (47), and our resultsgenerally support that observation. The structural diversifica-tion of the genomes of toxigenic El Tor strains spanning 45years in this study favors this viewpoint.
The superintegron cassettes examined to date seemingly en-code adaptive rather than indispensable functions (48). In thepresent study, the nontoxigenic strains exhibited greater diver-sity than the toxigenic strains in the superintegron region. Theimpact of integrons on bacterial evolution remains to be eval-
FIG. 8. CGH results for gene clusters encoding enzymes involvedin O lipopolysaccharide antigen synthesis in test strains, from VC0241to VC0270 in N16961. For each strain, a black line indicates thepresence of a gene and a white line indicates its absence.
4846 PANG ET AL. J. BACTERIOL.
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
uated beyond their obvious role in the dissemination of anti-biotic resistance genes. The variation of the superintegron andits impact on evolution of V. cholerae can be explained moreaccurately only after elucidation of the function of the genesinside the structure.
Microevolution among the seventh-pandemic toxigenic ElTor strains. The seventh cholera pandemic was caused bytoxigenic El Tor strains, which were reported to be a conservedclone (16). In spite of the conservation observed in the 11toxigenic El Tor strains, minor differences among these strainscan be observed and have accumulated over time from 1961 tonow. In the phylogenetic tree for chromosome 2 (Fig. 5), toxi-genic strains can be clustered into different groups according tothe year of isolation, reflecting ongoing but exiguous evolution.
VSP-I and VSP-II are the characteristic genome islandsidentified in the seventh-pandemic El Tor strains and toxigenicO139 strains (16, 44, 45). In this study, a toxigenic strain, D118,was found to lack VSP-II entirely. It is reported that twotoxigenic El Tor strains isolated in the 1970s lacked both VSP-Iand VSP-II or only VSP-II, and one toxigenic El Tor strainfrom 1937 lacked VSP-I, VSP-II, and VPI-2 (45). Such strainsresemble the presumed progenitors of the seventh-pandemicEl Tor strains (16, 45). Strain D118, used in this study, wasisolated in 1961, the year the seventh pandemic began. In thephylogenetic tree based on the sequences of four housekeepinggenes, D118 is in a lineage separate from that of other toxi-genic El Tor strains (Fig. 6). Therefore, D118 may represent astrain or group of strains that were converted into the predom-inant clone by the acquisition of VSP-II.
However, two epidemic strains isolated in 2005, FJ147 andZJ0596, were found to lack most of the ORFs in VSP-II, andthese two strains caused epidemics spanning thousands of ki-lometers and lasting several months in southeastern China.Therefore, the function of VSP-II in bacterial pathogenesis,environmental survival, and epidemic instigation remained tobe defined. Likely the seventh-pandemic strains are in contin-ual evolution, and it is expected that they may evolve into newpathogenic clones through the acquisition of new virulenceand/or adaptive factors.
Diversity of nontoxigenic strains and clonality of toxigenicstrains in serogroups O1 and O139. Previous studies revealedhigh diversity in nontoxigenic strains, especially the non-O1and non-O139 isolates, through molecular subtyping tech-niques including ribotyping, pulsed-field gel electrophoresis,housekeeping gene sequencing, and amplified fragment lengthpolymorphism (20, 32, 45, 50). Herein we used microarraymethods to investigate the genome contents of toxigenic andnontoxigenic O1 and O139 strains. Extensive genomic diversityamong these strains was again found.
Comparative analyses of nearly 300 microbial species hasdemonstrated that multiple forces have shaped prokaryoticgenomes during their dynamic evolution (22). The genomicdiversity of toxigenic and nontoxigenic O1 and O139 V.cholerae strains reflects the same dynamism of these otherprokaryote genomes.
Toxigenic O1 and O139 V. cholerae strains exhibit greaterclonality than environmental non-O1/non-O139 strains (32,45). Our study supports this observation by demonstrating thatthe genomic content of nontoxigenic V. cholerae strains is quitedifferent from that of closely related toxigenic strains. It has
been suggested that populations of bacteria typically consist ofabundant and heterogeneous isolates that rarely cause diseasebut that a small number of groups of closely related strains(clones or lineages) are particularly associated with epidemicsof disease (40). Serogroup O37 caused an outbreak in 1968(58). IS1004 fingerprinting, housekeeping gene sequencing,and multilocus enzyme electrophoresis analysis revealed thatepidemic O37 strains are phylogenetically clustered with O1and O139 isolates and that they are especially close to O1classical strains (3, 6, 45). It seems that, to date, choleraepidemics have been caused only by strains possessing an O1serogroup core genome, despite the fact that O37 and O139are non-O1 serogroups (36, 45).
A study in an area where cholera is endemic provided amodel for understanding the selection of toxigenic strains fromthe genetically diverse nonpathogenic strains in the environ-ment; such selection has been suggested to occur through en-richment in the mammalian host intestine (20). In areas wherecholera is endemic, large numbers of toxigenic V. cholerae cellsare present in the stool of patients which, when excreted intoenvironmental waters, contribute in part to the expansion ofcholera epidemics. A study indicated that by passing throughthe mammalian intestine, V. cholerae cells acquire a hyperin-fectious state that may also contribute to the epidemic spreadof cholera (41). We suppose that the rapid spread of toxigenicV. cholerae strains in areas where cholera is endemic makes theisolates in epidemics more conservative in clonality. In con-trast, nontoxigenic strains of diverse lineages do not undergosuch enrichment and exhibit extensive genetic diversity. Itshould also be considered whether the epidemic strains haveadvantages in maintenance and multiplication over nonpatho-genic strains in aquatic environments.
Our CGH analysis revealed that the O139 and El Tor strainsisolated in the coepidemic time period (10 more years since theemergence of O139 V. cholerae) exhibit superintegron missing-gene profiles essentially identical to those of El Tor strainsisolated before the emergence of O139 strains. It has beenproposed that serogroup O139 originated from an El Torstrain that obtained the O139 antigen biosynthesis and someother gene clusters (4, 52, 55). However, some portion ofVPI-II is absent from all of the test toxigenic O139 strainsisolated through the past years, and an additional five commonmissing genes in the superintegron in the toxigenic O139strains have been documented (in comparison with all toxi-genic El Tor strains). We speculate that these missing geneswere lost after the replacement of the O-antigen-syntheticgene cluster.
Genomic analysis of toxigenic and environmental nontoxi-genic O1, O139, and non-O1/non-O139 V. cholerae strains canprovide clues to V. cholerae evolution and the origin of patho-genic strains. Strains that can trigger epidemics may acquirevirulence genes successively (9, 10, 20, 45), rather than in asingle horizontal gene transfer event. In this study, we classi-fied the nontoxigenic O1 El Tor V. cholerae strains into threeclusters, named TCP�, TCPET, and TCPETvar strains. We spec-ulate that the nontoxigenic TCPET and TCPETvar strains rep-resent intermediate strains in different lineages evolving fromnontoxigenic to toxigenic strains. The acquisition of ctxABthrough the lysogenic conversion of CTX� may confer theability to initiate epidemics upon those strains (54). The gene
VOL. 189, 2007 ARRAY-BASED CGH ANALYSIS OF VIBRIO CHOLERAE 4847
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
of pIIICTX (previously known as OrfU) of CTX� encodes aprotein that can recognize TcpA as the receptor of phageinfection (26); however, tcpA has different sequence variants(12, 24, 28, 42, 43, 46). Variable gene sequences of pIIICTX
have also been documented (9, 11). The tcpA sequence ofnontoxigenic strain 7743 (also known as IEM101) used in thisstudy is identical to the sequence of seventh-pandemic El Torstrains, and the attB site for CTX� genome integration is alsoidentical (57). Strain 7743 can be infected by CTXET� andthereby converted into a toxigenic strain (37, 38, 57). More-over, rabbits immunized with 7743 and the live-vaccine candi-dates developed from 7743 can be protected against a chal-lenge with classical or El Tor strains (37–39, 57). Thisprotection demonstrates that strain 7743 has the same immu-nogenicity as toxigenic El Tor strains. Intact VPI, VPI-II, andTLC clusters are present in strain 7743, whereas VSP-I, VSP-II, and CTX� are absent. Taken together, these data suggestthat it is possible that strain 7743 represents an intermediate inthe origination of the seventh-pandemic clone.
It has been reported that the different TcpA alleles can serveas CTXET� and CTXcalc� receptors. However, a difference inthe infection frequency of CTX� among these strains has beendocumented (11); perhaps there are barriers to the cross-in-fection with some CTX�s with different pIIICTX alleles andnontoxigenic strains with different tcpA alleles. Perhaps, incomparison with strain 7743, nontoxigenic TCPETvar strainsmay have a reduced ability to acquire CTXET�. Such strainsmay provide mechanistic clues about evolutionary events in thedevelopment new toxigenic strains. In other words, perhapsthese strains (or similar strains) have the potential to becomenew dominant clones in future epidemics through the acquisi-tion of CTX� with the corresponding pIIICTX allele.
Obvious divergence in virulence-related genes such as tcpA(12, 21, 24, 28, 42, 43, 46), the gene for pIIICTX (9, 11), and rstR(9, 14, 15) in CTX� has been found in different strains. Sim-ilarly, sequence variance in the ompU, pilA, and EPS clusterwas also observed in nontoxigenic O1 V. cholerae in our study.It is inconvincible that this variation represents an accumula-tion of random mutations because the ORFs flanking thosegenes show little or no nucleotide variation.
V. cholerae can survive in aquatic environments and in hostintestines. Different gene expression mechanisms are used (7,27, 35, 41, 56) in these two niches. Besides the modulation ofgene expression, in different environments, gene sequencevariation may confer selective advantages on the strains invarious environments. The hypervariability at the tcpA locus isthought to be generated through multiple horizontal transferand recombination events, and tcpA diversity is presumed toreflect diversifying selection resulting from the host immuneresponse, susceptibility to CTX� (11). The sequence variabil-ity in type IV pili of several other bacterial species has alsobeen reported (8, 11). Therefore, we speculate that variation inthe ompU, pilA, and EPS gene cluster may contribute to func-tional differences (phenotypes) in V. cholerae as the strainsrespond to selective pressure.
Conclusion. Our analysis shows that nontoxigenic strainsfrom serogroups O1 and O139 are a collection of heteroge-neous clones that are part of the natural microflora of aquaticenvironments. Some strains carry virulence-related genes thatdiffer in sequence from those of the seventh-pandemic El Tor
strains; although these strains have no El Tor-type TCP, theymay nonetheless have suitable receptors that allow infection byother types of CTX�. For example, the CTX� with pIIICTX
differed from CTXET� and may infect the nontoxigenic strainswith tcpA sequences distinct from that of the El Tor type. Thepossibility is proposed that these strains can serve as raw ma-terial for the development of (future) epidemic clones throughacquisition of novel alleles of CTX�. These nontoxigenic O1and O139 strains may serve as models to understand howenvironmental nontoxigenic strains become toxigenic strainsvia gene loss and gain. Array-based CGH provides the abilityto rapidly examine the genomic contents of numerous bacterialisolates, thereby allowing determination of which genes arepresent and which are absent in nontoxigenic strains comparedto toxigenic ones, a powerful tool. We proposed that morerobust conclusions regarding the pathogenicity, environmentaladaptation, and evolution of V. cholerae can be achieved bywhole-genome sequencing and biological research on thesenontoxigenic strains.
ACKNOWLEDGMENTS
This research was supported by the National Basic ResearchPriorities Program (grant G1999054102 from Ministry of Scienceand Technology) and the Key Technologies R&D Program (grant2003BA712A05-04 from the Ministry of Science and Technology,People’s Republic of China).
REFERENCES
1. Anonymous. 1993. Large epidemic of cholera-like disease in Bangladeshcaused by Vibrio cholerae O139 synonym Bengal. Lancet 342:382–383.
2. Barua., D. 1992. History of cholera. Plenum, New York, NY.3. Beltran, P., G. Delgado, A. Navarro, F. Trujillo, R. K. Selander, and A.
Cravioto. 1999. Genetic diversity and population structure of Vibrio cholerae.J. Clin. Microbiol. 37:581–590.
4. Berche, P., C. Poyart, E. Abachin, H. Lelievre, J. Vandepitte, A. Dodin, andJ. M. Fournier. 1994. The novel epidemic strain O139 is closely related to thepandemic strain O1 of Vibrio cholerae. J. Infect. Dis. 170:701–704.
5. Bhattacharya, M. K., S. K. Bhattacharya, S. Garg, P. K. Saha, D. Dutta,G. B. Nair, B. C. Deb, and K. P. Das. 1993. Outbreak of Vibrio choleraenon-O1 in India and Bangladesh. Lancet 341:1346–1347.
6. Bik, E. M., R. D. Gouw, and F. R. Mooi. 1996. DNA fingerprinting of Vibriocholerae strains with a novel insertion sequence element: a tool to identifyepidemic strains. J. Clin. Microbiol. 34:1453–1461.
7. Bina, J., J. Zhu, M. Dziejman, S. Faruque, S. Calderwood, and J. Mekalanos.2003. ToxR regulon of Vibrio cholerae and its expression in vibrios shed bycholera patients. Proc. Natl. Acad. Sci. USA 100:2801–2806.
8. Blank, T. E., H. Zhong, A. L. Bell, T. S. Whittam, and M. S. Donnenberg.2000. Molecular variation among type IV pilin (bfpA) genes from diverseenteropathogenic Escherichia coli strains. Infect. Immun. 68:7028–7038.
9. Boyd, E. F., A. J. Heilpern, and M. K. Waldor. 2000. Molecular analyses ofa putative CTX� precursor and evidence for independent acquisition ofdistinct CTX�s by toxigenic Vibrio cholerae. J. Bacteriol. 182:5530–5538.
10. Boyd, E. F., K. E. Moyer, L. Shi, and M. K. Waldor. 2000. Infectious CTX�and the Vibrio pathogenicity island prophage in Vibrio mimicus: evidence forrecent horizontal transfer between V. mimicus and V. cholerae. Infect. Im-mun. 68:1507–1513.
11. Boyd, E. F., and M. K. Waldor. 2002. Evolutionary and functional analysesof variants of the toxin-coregulated pilus protein TcpA from toxigenic Vibriocholerae non-O1/non-O139 serogroup isolates. Microbiology 148:1655–1666.
12. Chakraborty, S., A. K. Mukhopadhyay, R. K. Bhadra, A. N. Ghosh, R. Mitra,T. Shimada, S. Yamasaki, S. M. Faruque, Y. Takeda, R. R. Colwell, and G. B.Nair. 2000. Virulence genes in environmental strains of Vibrio cholerae.Appl. Environ. Microbiol. 66:4022–4028.
13. Colwell, R. R., and W. M. Spira. 1992. The ecology of Vibrio cholerae, p.107–127. In D. Barua and W. B. Greenough III (ed.), Cholera. PlenumMedical Book Co., New York, NY.
14. Davis, B. M., H. H. Kimsey, W. Chang, and M. K. Waldor. 1999. The Vibriocholerae O139 Calcutta bacteriophage CTX� is infectious and encodes anovel repressor. J. Bacteriol. 181:6779–6787.
15. Davis, B. M., K. E. Moyer, E. F. Boyd, and M. K. Waldor. 2000. CTXprophages in classical biotype Vibrio cholerae: functional phage genes butdysfunctional phage genomes. J. Bacteriol. 182:6992–6998.
16. Dziejman, M., E. Balon, D. Boyd, C. M. Fraser, J. F. Heidelberg, and J. J.
4848 PANG ET AL. J. BACTERIOL.
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
Mekalanos. 2002. Comparative genomic analysis of Vibrio cholerae: genesthat correlate with cholera endemic and pandemic disease. Proc. Natl. Acad.Sci. USA 99:1556–1561.
17. Dziejman, M., D. Serruto, V. C. Tam, D. Sturtevant, P. Diraphat, S. M.Faruque, M. H. Rahman, J. F. Heidelberg, J. Decker, L. Li, K. T. Montgomery,G. Grills, R. Kucherlapati, and J. J. Mekalanos. 2005. Genomic characteriza-tion of non-O1, non-O139 Vibrio cholerae reveals genes for a type III secretionsystem. Proc. Natl. Acad. Sci. USA 102:3465–3470.
18. Essex-Lopresti, A. E., J. A. Boddey, R. Thomas, M. P. Smith, M. G. Hartley,T. Atkins, N. F. Brown, C. H. Tsang, I. R. Peak, J. Hill, I. R. Beacham, andR. W. Titball. 2005. A type IV pilin, PilA, contributes to adherence ofBurkholderia pseudomallei and virulence in vivo. Infect. Immun. 73:1260–1264.
19. Faruque, S. M., K. M. Ahmed, A. R. Abdul Alim, F. Qadri, A. K. Siddique,and M. J. Albert. 1997. Emergence of a new clone of toxigenic Vibriocholerae O1 biotype El Tor displacing V. cholerae O139 Bengal in Bangla-desh. J. Clin. Microbiol. 35:624–630.
20. Faruque, S. M., N. Chowdhury, M. Kamruzzaman, M. Dziejman, M. H.Rahman, D. A. Sack, G. B. Nair, and J. J. Mekalanos. 2004. Genetic diversityand virulence potential of environmental Vibrio cholerae population in acholera-endemic area. Proc. Natl. Acad. Sci. USA 101:2123–2128.
21. Faruque, S. M., M. Kamruzzaman, I. M. Meraj, N. Chowdhury, G. B. Nair,R. B. Sack, R. R. Colwell, and D. A. Sack. 2003. Pathogenic potential ofenvironmental Vibrio cholerae strains carrying genetic variants of the toxin-coregulated pilus pathogenicity island. Infect. Immun. 71:1020–1025.
22. Fraser-Liggett, C. M. 2005. Insights on biology and evolution from microbialgenome sequencing. Genome Res. 15:1603–1610.
23. Galen, J. E., J. M. Ketley, A. Fasano, S. H. Richardson, S. S. Wasserman,and J. B. Kaper. 1992. Role of Vibrio cholerae neuraminidase in the functionof cholera toxin. Infect. Immun. 60:406–415.
24. Ghosh, C., R. K. Nandy, S. K. Dasgupta, G. B. Nair, R. H. Hall, and A. C.Ghose. 1997. A search for cholera toxin (CT), toxin coregulated pilus (TCP),the regulatory element ToxR and other virulence factors in non-01/non-0139Vibrio cholerae. Microb. Pathog. 22:199–208.
25. Heidelberg, J. F., J. A. Eisen, W. C. Nelson, R. A. Clayton, M. L. Gwinn, R. J.Dodson, D. H. Haft, E. K. Hickey, J. D. Peterson, L. Umayam, S. R. Gill,K. E. Nelson, T. D. Read, H. Tettelin, D. Richardson, M. D. Ermolaeva, J.Vamathevan, S. Bass, H. Qin, I. Dragoi, P. Sellers, L. McDonald, T.Utterback, R. D. Fleishmann, W. C. Nierman, O. White, S. L. Salzberg,H. O. Smith, R. R. Colwell, J. J. Mekalanos, J. C. Venter, and C. M.Fraser. 2000. DNA sequence of both chromosomes of the cholera patho-gen Vibrio cholerae. Nature 406:477–483.
26. Heilpern, A. J., and M. K. Waldor. 2003. pIIICTX, a predicted CTX� minorcoat protein, can expand the host range of coliphage fd to include Vibriocholerae. J. Bacteriol. 185:1037–1044.
27. Hsiao, A., Z. Liu, A. Joelsson, and J. Zhu. 2006. Vibrio cholerae virulenceregulator-coordinated evasion of host immunity. Proc. Natl. Acad. Sci. USA103:14542–14547.
28. Iredell, J. R., and P. A. Manning. 1997. Translocation failure in a type-4 pilinoperon: rfb and tcpT mutants in Vibrio cholerae. Gene 192:71–77.
29. Islam, M. S., B. S. Drasar, and R. B. Sack. 1994. The aquatic flora and faunaas reservoirs of Vibrio cholerae: a review. J. Diarrhoeal Dis. Res. 12:87–96.
30. Jermyn, W. S., and E. F. Boyd. 2002. Characterization of a novel Vibriopathogenicity island (VPI-2) encoding neuraminidase (nanH) among toxi-genic Vibrio cholerae isolates. Microbiology 148:3681–3693.
31. Kaper, J. B., J. G. Morris, Jr., and M. M. Levine. 1995. Cholera. Clin.Microbiol. Rev. 8:48–86.
32. Karaolis, D. K., R. Lan, and P. R. Reeves. 1995. The sixth and seventhcholera pandemics are due to independent clones separately derived fromenvironmental, nontoxigenic, non-O1 Vibrio cholerae. J. Bacteriol. 177:3191–3198.
33. Karunasagar, I., I. Rivera, B. Joseph, B. Kennedy, V. R. Shetty, A. Huq, I.Karunasagar, and R. R. Colwell. 2003. ompU genes in nontoxigenic Vibriocholerae associated with aquaculture. J. Appl. Microbiol. 95:338–343.
34. Kierek, K., and P. I. Watnick. 2003. Environmental determinants of Vibriocholerae biofilm development. Appl. Environ. Microbiol. 69:5079–5088.
35. Larocque, R. C., J. B. Harris, M. Dziejman, X. Li, A. I. Khan, A. S. Faruque,S. M. Faruque, G. B. Nair, E. T. Ryan, F. Qadri, J. J. Mekalanos, and S. B.Calderwood. 2005. Transcriptional profiling of Vibrio cholerae recovereddirectly from patient specimens during early and late stages of human infec-tion. Infect. Immun. 73:4488–4493.
36. Li, M., T. Shimada, J. G. Morris, Jr., A. Sulakvelidze, and S. Sozhamannan.2002. Evidence for the emergence of non-O1 and non-O139 Vibrio cholerae
strains with pathogenic potential by exchange of O-antigen biosynthesisregions. Infect. Immun. 70:2441–2453.
37. Liang, W., S. Wang, F. Yu, L. Zhang, G. Qi, Y. Liu, S. Gao, and B. Kan. 2003.Construction and evaluation of a safe, live, oral Vibrio cholerae vaccinecandidate, IEM108. Infect. Immun. 71:5498–5504.
38. Liu, G., M. Yan, W. Liang, G. Qi, Y. Liu, S. Gao, and B. Kan. 2006.Resistance of the cholera vaccine candidate IEM108 against CTX� infec-tion. Vaccine 24:1749–1755.
39. Liu, Y. Q., G. M. Qi, S. X. Wang, Y. M. Yu, G. C. Duan, L. J. Zhang, and S. Y.Gao. 1995. A natural vaccine candidate strain against cholera. Biomed.Environ. Sci. 8:350–358.
40. Maiden, M. C., J. A. Bygraves, E. Feil, G. Morelli, J. E. Russell, R. Urwin,Q. Zhang, J. Zhou, K. Zurth, D. A. Caugant, I. M. Feavers, M. Achtman, andB. G. Spratt. 1998. Multilocus sequence typing: a portable approach to theidentification of clones within populations of pathogenic microorganisms.Proc. Natl. Acad. Sci. USA 95:3140–3145.
41. Merrell, D. S., S. M. Butler, F. Qadri, N. A. Dolganov, A. Alam, M. B. Cohen,S. B. Calderwood, G. K. Schoolnik, and A. Camilli. 2002. Host-inducedepidemic spread of the cholera bacterium. Nature 417:642–645.
42. Mukhopadhyay, A. K., S. Chakraborty, Y. Takeda, G. B. Nair, and D. E.Berg. 2001. Characterization of VPI pathogenicity island and CTX� pro-phage in environmental strains of Vibrio cholerae. J. Bacteriol. 183:4737–4746.
43. Nandi, B., R. K. Nandy, A. C. Vicente, and A. C. Ghose. 2000. Molecularcharacterization of a new variant of toxin-coregulated pilus protein (TcpA)in a toxigenic non-O1/Non-O139 strain of Vibrio cholerae. Infect. Immun.68:948–952.
44. O’Shea, Y. A., S. Finnan, F. J. Reen, J. P. Morrissey, F. O’Gara, and E. F.Boyd. 2004. The Vibrio seventh pandemic island-II is a 26.9 kb genomicisland present in Vibrio cholerae El Tor and O139 serogroup isolates thatshows homology to a 43.4 kb genomic island in V. vulnificus. Microbiology150:4053–4063.
45. O’Shea, Y. A., F. J. Reen, A. M. Quirke, and E. F. Boyd. 2004. Evolutionarygenetic analysis of the emergence of epidemic Vibrio cholerae isolates on thebasis of comparative nucleotide sequence analysis and multilocus virulencegene profiles. J. Clin. Microbiol. 42:4657–4671.
46. Rhine, J. A., and R. K. Taylor. 1994. TcpA pilin sequences and colonizationrequirements for O1 and O139 Vibrio cholerae. Mol. Microbiol. 13:1013–1020.
47. Rowe-Magnus, D. A., A. M. Guerout, L. Biskri, P. Bouige, and D. Mazel.2003. Comparative analysis of superintegrons: engineering extensive geneticdiversity in the Vibrionaceae. Genome Res. 13:428–442.
48. Rowe-Magnus, D. A., A. M. Guerout, P. Ploncard, B. Dychinco, J. Davies,and D. Mazel. 2001. The evolutionary history of chromosomal super-integrons provides an ancestry for multiresistant integrons. Proc. Natl.Acad. Sci. USA 98:652–657.
49. Sack, D. A., R. B. Sack, G. B. Nair, and A. K. Siddique. 2004. Cholera.Lancet 363:223–233.
50. Singh, D. V., M. H. Matte, G. R. Matte, S. Jiang, F. Sabeena, B. N. Shukla,S. C. Sanyal, A. Huq, and R. R. Colwell. 2001. Molecular analysis of Vibriocholerae O1, O139, non-O1, and non-O139 strains: clonal relationshipsbetween clinical and environmental isolates. Appl. Environ. Microbiol. 67:910–921.
51. Stabb, E. V., and E. G. Ruby. 2003. Contribution of pilA to competitivecolonization of the squid Euprymna scolopes by Vibrio fischeri. Appl. EnvironMicrobiol. 69:820–826.
52. Stroeher, U. H., G. Parasivam, B. K. Dredge, and P. A. Manning. 1997.Novel Vibrio cholerae O139 genes involved in lipopolysaccharide biosyn-thesis. J. Bacteriol. 179:2740–2747.
53. Waldor, M. K., R. Colwell, and J. J. Mekalanos. 1994. The Vibrio choleraeO139 serogroup antigen includes an O-antigen capsule and lipopolysaccha-ride virulence determinants. Proc. Natl. Acad. Sci. USA 91:11388–11392.
54. Waldor, M. K., and J. J. Mekalanos. 1996. Lysogenic conversion by a fila-mentous phage encoding cholera toxin. Science 272:1910–1914.
55. Waldor, M. K., and J. J. Mekalanos. 1994. Vibrio cholerae O139 specific genesequences. Lancet 343:1366.
56. Xu, Q., M. Dziejman, and J. J. Mekalanos. 2003. Determination of thetranscriptome of Vibrio cholerae during intraintestinal growth and midexpo-nential phase in vitro. Proc. Natl. Acad. Sci. USA 100:1286–1291.
57. Yan, M., G. Liu, B. Diao, H. Qiu, L. Zhang, W. Liang, S. Gao, and B. Kan.2007. A Vibrio cholerae serogroup O1 vaccine candidate against CTXET�infection. Vaccine 25:4046–4055.
58. Zinnaka, Y., and C. C. Carpenter, Jr. 1972. An enterotoxin produced bynoncholera vibrios. Johns Hopkins Med. J. 131:403–411.
VOL. 189, 2007 ARRAY-BASED CGH ANALYSIS OF VIBRIO CHOLERAE 4849
on May 24, 2020 by guest
http://jb.asm.org/
Dow
nloaded from
