JOURNAL OF CLINICAL MICROBIOLOGY, Oct. 2010, p. 3681–3689 Vol. 48, No. 10 0095-1137/10/$12.00 doi:10.1128/JCM.00565-10 Copyright ? 2010, American Society for Microbiology. All Rights Reserved.
Molecular Surveillance of Clinical Neisseria gonorrhoeae
Isolates in Russia †
Elena N. Ilina,* Nina Y. Oparina, Egor A. Shitikov, Alexandra D. Borovskaya, and Vadim M. Govorun
Research Institute for Physico-Chemical Medicine of Ministry of Public Health of Russian Federation, Moscow, Russia
Received 16 March 2010/Returned for modi，cation 28 April 2010/Accepted 12 July 2010
The choice of adequate methods for epidemiological purposes remains a challenging problem in Neisseria gonorrhoeae molecular monitoring. In this study, the collection of geographically unrelated gonococci (n 103) isolated in Russian clinics was comparably tested by (i) a traditional serotyping scheme, (ii) por typing, (iii) Neisseria gonorrhoeae multiantigen sequence typing (NG-MAST), and (iv) multilocus sequence typing (MLST). It is shown that, according to sequencing data, a third of the strains carried new porB1 alleles, as well as tbpB ones, and more than half of the samples had new sequence types (STs) as determined by NG-MAST or MLST. The discriminatory power for each typing method was calculated by using the Hunter-Gaston discriminatory index, D. Commonly, modern nucleic acid-based typing methods (por typing, NG-MAST, and MLST) appeared to be more ef；cient than the classical serotyping scheme. While the traditional serotyping gave a D value of 0.82, the por typing, NG-MAST, and MLST approaches yielded D values of 0.97, 0.98, and 0.91, respectively. Each typing technique revealed the distribution of gonococci slightly correlated with their geographical sources. However, only the MLST method STs were highly associated with certain phenotypes. Although ST1594, ST1892, and ST6720 were typical for susceptible gonococci, ST1901 and ST6716 were undoubtedly associated with a multidrug-resistant phenotype. We conclude that every tested nucleic acid-based typing method is suitable for N. gonorrhoeae molecular surveillance. However, the MLST method seems to serve large-scale epidemiological purposes, whereas the NG-MAST and por typing approaches are more appropriate for the investigation of local outbreaks.
The choice of an optimal epidemiological approach for Neis- sequences. It possesses rather high discriminatory power (39) seria gonorrhoeae typing remains a problem for public health and presents data in a format comparable with serotyping data. control strategies. The worldwide spread of drug-resistant It should be mentioned that PorB1 protein plays a key role in strains requires typing methods to be introduced into a na- the adhesion of gonococci to epitheliocytes and re！ects the tional surveillance programs that are already realized within pathogenicity of isolates. Moreover, in earlier studies the mul- the Australian Gonococcal Surveillance Programme (33). In tidrug-resistant phenotype of N. gonorrhoeae was bound to Russia, strains that are resistant to penicillin, tetracycline, and certain serotypes determined in accordance with the antigenic ！uoroquinolones are common, reaching up to 60% prevalence properties of PorB1 proteins (7, 21, 29). in some regions (14). As is generally known, most resistance Today, the N. gonorrhoeae multiantigen sequence typing mechanisms in N. gonorrhoeae are linked to mutations in (NG-MAST) (19) appears to be a leading method for N. gon- genomic DNA, and a wide dissemination of such mutations has orrhoeae typing. It is based on the analysis of internal frag- been demonstrated in subjects with gonorrhea in the Russian ments of two hypervariable genes, porB and tbpB, encoding population (11). In fact, chromosomally mediated drug resis- super，cial gonococcus antigens and therefore under positive tance expands clonally in a bacterial population. Thus, tracing selection (19). An online database (www.ng-mast.net) collects the spread of gonorrhea by using strain differentiation methods current information on the sequence types of gonococcus iso- is of great importance. In addition, population genetics are lates in different regions of the world, and it can be an ef，cient important in understanding the evolutionary history, epidemi- instrument for objective estimation of the genetic variability of ology, and population dynamics of pathogens. the microbial population and tracing the spread of infection Widespread adoption of the molecular diagnosis of gonor- (2). Moreover, an association of the NG-MAST sequence type rhea could compromise traditional bacteriological cultivation with the antibiotic resistance pro，le has been demonstrated by in routine practice and highlights the need for the development some investigators (20, 26). of molecular tools for N. gonorrhoeae typing. Several promising In addition, NG-MAST, which is usually carried out on approaches exist in this ，eld. One of them is por typing (7, 8, DNA extracted from a pure bacterial culture, can be per- 27), based on comparative analysis of porB gene nucleotide formed directly from noncultured samples such as a piece of clothing (18) or clinical specimens (40). Although some re- searchers have reported a successful application of NG-MAST * Corresponding author. Mailing address: Research Institute of to urogenital specimens (urine specimens and swabs from the Physicochemical Medicine, Malaya Pirogovskaya St., 1a, 119435, Moscow, Russia. Phone: 7-499-2450471. Fax: 7-499-2467721. E-mail: cervix, urethra, and vagina) and to rectal swabs, this typing firstname.lastname@example.org. method was found to be less suitable for throat swabs due to † Supplemental material for this article may be found at http://jcm cross-reaction with commensal Neisseria species. It seems that .asm.org/. direct typing schemes utilizing probe hybridization methods Published ahead of print on 21 July 2010.
FIG. 1. Phylogenetic tree based on analysis of the entire porB1 gene nucleotide sequences. Serovars P1A and P1B are indicated by vertical lines.
Clinical N. gonorrhoeae strains (n 103) are named according to Table S2 in the supplemental material. Isolate origins: Irkutsk (n 22), black square; Samara (n 10), gray square; Murmansk (n 14), large black circle; St. Petersburg (n 29), small black circle; Arkhangelsk (n 28), large gray circle. Multiresistant isolates (i.e., isolates resistant to at least two antibiotics) are indicated by thick lines.
VOL. 48, 2010 N. GONORRHOEAE MOLECULAR TYPING 3683
TABLE 1. Variability of entire porB gene nucleotide sequences for 103 N. gonorrhoeae clinical isolates
aNo. of: Z-test result Nucleotide Locus Length (bp) Polymorphic variability ( ) Isolates Alleles dN dS dN dS dN dS sites 257 porB 933–1020 0.062 103 58 0.0013 0.0032 1.0 1.0 0.0076 0.0038 porB1a 933 11 2 18 0.004 porB1b 990–1020 92 56 135 0.030 0.21 0.44 1.0 a A Z-test was performed to check the null hypothesis as to whether the porB allele is under purifying selection (dN dS), diversifying positive selection (dN dS), or neutral selection (dN dS). dN, nonsynonymous substitutions; dS, synonymous substitutions.
The DNA sequences of housekeeping genes were uploaded on the http://pubmlst (16, 17) for a broad spectrum of clinical samples are more .org website. convenient than sequencing. NG-MAST or MLST. Characterization of all isolates by NG-MAST and MLST The opposite approach—analysis of conservative, presum- was performed as originally described (1, 19). The corresponding sequences were ably selective neutral housekeeping genes—is taken by mul- submitted to the NG-MAST (http://www.ng-mast.net) or MLST (http://pubmlst tilocus sequence typing (MLST) (1, 38). Possessing a suf，cient .org/neisseria) websites for comparison with the existing alleles for determination of the allele types and sequence types (STs) of the isolates. number of allelic variants and characterized by the slow accu- Data analysis. Sequences were aligned by using CLUSTAL X 1.8 software mulation of mutations, these genes re！ect the natural evolu- (http://www.clustal.org). At the next step, aligned nucleotide sequences were tion of the microbial population and common trends in the converted to MEGA version 4.0.2 software (http://www.megasoftware.net), and spread of gonorrhea. distance matrices were estimated (using the nucleotide Tajima-Nei model). Fur- ther classi，cation was constructed by using the neighbor-joining method based Often, the effectiveness of N. gonorrhoeae surveillance de- on a bootstrap tree stability test (500 iterations). pends on the methods used for species identi，cation and the Visual mapping of multidrug resistance and place of strain isolation were typing of clinical isolates. However, considering the variety of carried out by using the Dendroscope tree editor (http://www-ab.informatik.uni typing systems but the lack of world experience in their appli- -tuebingen.de/software/dendroscope/welcome.html) (10). cation to monitoring gonococci, it has not yet been determined The discriminatory power for each typing method was de，ned by calculation of a single numerical index of discrimination (i.e., the Hunter-Gaston discrimi- whether there is a unique objective method for revealing the natory index [D]) (9), based on the probability that two unrelated isolates would relationships between clinical isolates. The goal of the present be placed into different typing groups: study was to evaluate a number of approaches, including por typing, NG-MAST, MLST, and serotyping for typing geo- S 1 graphically unrelated gonococcal isolates in Russia. D 1 nn1 jjN N 1 j 1 (This study was partially presented during the 16th Interna- tional Pathogenic Neisseria Conference in 2008 in the Nether- where N is the number of unrelated strains tested, S is the number of different lands.) types, and nis the number of strains belonging to the jth type. j The calculated data was interpreted with commonly accepted decision that a D of 0.90 is desirable for a con，dent typing method (9). MATERIALS AND METHODS A recombination test to determine the index of association and the stan- SBacterial isolates. N. gonorrhoeae clinical isolates collected in different regions dardized index of association (Iand IA, respectively) was performed using Aof Russia from 2004 to 2005 were obtained from the Central Research Institute the START2 program (12). Allele pro，le data were analyzed in eBURST of Dermatology and Venereology. All isolates were identi，ed as N. gonorrhoeae (version 3; http://eburst.mlst.net/v3/enter_data/single/) to de，ne clonal com- and tested for serotype and for susceptibility to penicillin G (PEN), tetracycline plexes or groups (6). (TET), cipro！oxacin (CIP), spectinomycin, and ceftriaxone as a part of a previ- Associations between sequence type and susceptibility level were examined by 2ous investigation (11). using chi-square () and Cramer (V) tests. All statistical calculations were Genetic analysis. Genomic DNA from N. gonorrhoeae was extracted according two-tailed and were conducted with a signi，cance level of at least P 0.05. to the method of Boom et al. (4). When necessary, the prepared DNA samples Statistical analysis was performed with Statistica 7.0 software. were stored at 20?C. Ampli，cation reactions were carried out in 10 l of 66 mM Tris-HCl (pH 9.0), )SO16.6 mM (NH424, 2.5 mM MgCl2, 0.2 mM concentrations of each de- RESULTS oxynucleoside triphosphate (dNTP), 5 pmol of each primers (see Table S1 in the In all, 103 clinical N. gonorrhoeae isolates from ，ve remote supplemental material), and 1 U of Taq polymerase (Fermentas, Lithuania) under the following conditions: 35 cycles of 94?C for 20 s, 60?C for 20 s, and 72?C regions of Russia—Irkutsk (n 22), Samara (n 10), for 15 s. A programmed thermocycler Tetrad DNA Engine (MJ Research, Inc.) Arkhangelsk (n 28), Murmansk (n 14), and St. Petersburg was used. (n 29)—were investigated (see Table S2 in the supplemental Dephosphorylation of the 5 -end phosphate groups of dNTPs and cleavage of material). primers in the postampli，cation reaction mixture was done by incubation with 0.5 U of shrimp alkaline phosphatase and 0.1 U of exonuclease I (both enzymes Protein-based typing method. The traditional serotyping from Fermentas, Lithuania) for 20 min at 37?C, followed by the inactivation by procedure divided all typing isolates into 11 different groups, heating for 10 min at 85?C. with a prevalence of the PIB serovar (92/103 [89.3%]). Within Sequencing procedure was performed by the modi，ed Sanger method using a this serovar the PIB2, PIB3, PIB5, and PIB22 serotypes were ABI Prism BigDye terminator cycle sequencing ready reaction kit and an ABI Prism 3100 genetic analyzer (Applied Biosystems, Hitachi, Japan) according to detected in 30 (29.1%), 27 (26.2%), 10 (9.7%), and 10 (9.7%) the manufacturer’s instructions. cases, respectively. The remaining PIB serovar isolates (n Analysis of the nucleotide sequences was carried out by using Vector NTI 15) belonged to ，ve different serotypes: PIB3/6 (n 4), PIB4 Advance v.9.0 software (Infomarks, Inc.). The DNA sequences of porB and tbpB (n 4), PIB8 (n 2), PIB9 (n 3), and PIB26 (n 2). genes collected during the present study were submitted to GenBank (accession Among the PIA serovar isolates, the PIA6 serotype was found numbers EU530732 to EU530817 and EU532618 to EU532759, respectively).
VOL. 48, 2010 N. GONORRHOEAE MOLECULAR TYPING 3685
in 10 (9.7%) isolates, and one sample belonged to the PIA10 here for the ，rst time. Generally, there was no clear geo-
graphic clustering of speci，c NG-MAST types found within the serotype. The calculated D value was found to be 0.82, which
country. In some cases, certain STs were found for gonococci shows that the discriminatory power of serotyping is not suf，-
isolated in the same place, for example, ST205 in Samara, cient to make it an effective typing method.
ST285 in Murmansk, and ST972 in Irkutsk. However, this Nucleic acid-based typing methods. Nucleic acid-based tech-
probably does not re！ect local outbreaks but rather the global niques, such as por typing, NG-MAST, and MLST, divided all
spread of infection. The calculated index of discrimination for of the isolates into 58, 61, and 30 groups, respectively.
this typing method was 0.98, which re！ected the strong dis- por typing. Alignment of the entire porB gene nucleotide
sequences revealed the suf，cient heterogeneity of N. gonor- criminatory power of NG-MAST.
rhoeae clinical isolates. The dendrogram constructed by the MLST. According to MLST data, the 30 different groups neighbor-joining algorithm is shown in Fig. 1. When a cutoff (STs) were discovered. Among them, 13 ST clusters of between value of 0.1 (genetic distance) was used, 58 clusters could be 2 and 25 isolates were identi，ed. ST6716 (n 25) were found
delineated. Expectedly, all gonococcal isolates were separated to be the most prevalent, the other largest groups were ST1901 into two large groups corresponding to the porB1a (n 11) (n 11), ST1905 (n 11), ST1892 (n 8), ST1584 (n 8),
and porB1b (n 92) alleles of the porB gene. Comparative and ST6720 (n 5). The 17 STs were represented by single analysis of the nucleotide polymorphism showed a much isolates. Concatenated sequences of seven housekeeping gene higher variability for the porB1b allele than for porB1a. Vari- fragments were analyzed by using START 2.0 software. A ability for the porB1b allele became apparent in a number of phylogenetic tree constructed by using the neighbor-joining exposed allelic variants and polymorphic sites within each method is shown in Fig. 3. The calculated D value was deter-
group. Moreover, 3-, 6-, or 12-nucleotide insertions, as well as mined to be 0.91 by MLST, which demonstrates this method’s
6-nucleotide deletions, were found for several isolates possess- applicability as an effective typing method. ing the porB1b allele. Analysis of the nonsynonymous/synony- Analysis by eBURST showed that four nonoverlapping mous substitution rate ratio revealed different trends in the groups (clonal complexes) contained STs, which matched at natural selection of porB alleles (Table 1). While the in！uence least one other ST at ，ve or more loci (Fig. 4). Four STs—
of positive diversifying selection was shown for the porB1b ST1901, ST1927, ST6716, and ST6715—were recognized as
allele, the porB1a allele seemed to be under the in！uence of hyperinvasive genotypes. Among them, ST1901 and ST6716 purifying selection. Commonly, the discriminatory index for were evolutionary related, sharing six of seven alleles. this typing method was 0.97, which is suf，cient for epidemio- Based on MLST data, the examined N. gonorrhoeae isolates
logical purposes. were compared to gonococcal isolates collected in the United Comparative analysis of internal fragments of porB genes Kingdom (1) in terms of allelic diversity (Table 2) and recom- (490 bp) via the public database on the NG-MAST website bination parameters (Table 3). The most changeable loci were (http://www.ng-mast.net) showed that a third of tested group found to be similar to these collections, and signi，cant linkage
carried new por alleles unrepresented in the nucleotide library. disequilibrium was detected for both populations. The 49 isolates formed six clusters carrying por-37 (n 10), Genotyping data and antimicrobial resistance. N. gonor-
por-90 (n 10), por-91 (n 8), por-164 (n 6), por-232 (n rhoeae clinical isolate distribution according to different nu- 7), and por-685 (n 8) alleles. Among these, the por-90 cluster cleic acid-based typing schemes (por typing, NG-MAST, and
was clearly associated with the PIA6 serotype, and the por-91, MLST) was compared to their susceptibility pro，les. Since
por-164, and por-232 clusters contained gonococci isolated in many STs were represented by occasional isolates, only clusters Irkutsk, Samara, and Murmansk, respectively. that contained ，ve or more members were considered. Among
them, NG-MAST ST205 (n 6), ST285 (n 7), and ST282 NG-MAST. Based on analysis of porB and tbpB fragments,
(n 5) were unambiguously associated with multidrug resis- all isolates were characterized by NG-MAST. We found 51 and
27 variants of porB and tbpB alleles, respectively, which re- tance, i.e., intermediate susceptibility or resistance to penicil- sulted in the assignment of 61 different sequence types (STs). lin, tetracycline and cipro！oxacin. NG MAST ST1043 (n 8)
A phylogenetic tree based on the concatenated sequences of included gonococci not susceptible to PEN, resistant to TET, two loci is shown in Fig. 2. The 57 (55%) isolates belonged to and susceptible to CIP. Appropriately, the same isolates also 15 different STs containing from two to eight members. The formed corresponding clusters in accordance with the por typ-
largest of them were ST1043 (n 8), ST285 (n 7), ST206 ing results.
(n 6), and ST972 (n 5). The other 46 STs were represented A relationship between the N. gonorrhoeae phenotype and
by individual isolates, including four penicillinase-producing N. certain STs was observed by MLST (Table 4). Although gonorrhoeae (PPNG) isolates collected in St. Petersburg (n ST1594, ST1892, and ST6720 were typical for susceptible 3) and Murmansk (n 1). Meanwhile, approximately half of gonococci, ST1901 and ST6716 were undoubtedly associated the isolates (n 50) belonged to the 39 new STs described with multidrug resistance. ST1905 comprised isolates display-
FIG. 2. Phylogenetic tree based on analysis of concatenated sequences of porB1 and tbpB gene fragments (NG-MAST scheme). Newly described STs are indicated in red; serovars P1A and P1B are indicated too. Clinical N. gonorrhoeae strains (n 103) are named according to Table S2 in the supplemental material. Isolate origins: Irkutsk (n 22), black square; Samara (n 10), gray square; Murmansk (n 14), large black circle; St. Petersburg (n 29), small black circle; Arkhangelsk (n 28), large gray circle. Multiresistant isolates (i.e., isolates resistant to at least two antibiotics) are indicated by thick lines.
FIG. 3. Phylogenetic tree based on analysis of concatenated sequences of seven housekeeping gene fragments (MLST scheme). Newly
described STs are indicated in red. Clinical N. gonorrhoeae strains (n 103) are named according to Table S2 in the supplemental material. Isolate origins: Irkutsk (n 22), black square; Samara (n 10), gray square; Murmansk (n 14), large black circle; St. Petersburg (n 29), small black circle; Arkhangelsk (n 28), large gray circle. Multiresistant isolates (i.e., isolates resistant to at least two antibiotics) are indicated by thick lines.
VOL. 48, 2010 N. GONORRHOEAE MOLECULAR TYPING 3687
FIG. 4. Population snapshot of clinical N. gonorrhoeae isolates (n 103) typed in the present study. The snapshot was created by the eBURST algorithm applied for the analysis of MLST data. The circles represent STs differing in only one housekeeping gene sequence from the founder genotype (in the center). Lines connect the other evolutionarily related STs, i.e., those sharing six of seven alleles. Each centered group was considered a clonal complex. ST1594 was identi，ed as a singleton.
ing intermediate susceptibility to PEN, resistance to TET, and have been adopted and used to study gonococcal strains cir- (especially) susceptibility to CIP. culating in the Russian Federation. The traditional serotyping
procedure is based on an analysis of protein variability. The
other typing schemes—por typing, NG-MAST, and MLST— DISCUSSION are nucleic acid-based techniques. In recent years, many molecular approaches for N. gonor- Each tested technique typed all of the isolates, but their rhoeae typing have been suggested: opa typing (23, 25), por discriminatory powers were different. According to our data, typing (7, 15, 34, 39), ampli，ed fragment length polymorphism the discriminatory power of the traditional serotyping ap- (24, 32), restriction fragment length polymorphism (3, 15), and proach was 0.82 and did not satisfy the value of 0.90 desirable some others, but no single typing scheme has been generally for effective typing systems (9). Although this disadvantage of adopted, and the lack of such a standard typing method has serotyping was discovered 10 years ago (37), this method can impeded the sharing of epidemiological data between labora- still be useful when a comparison of modern and old data is tories. The recently developed NG-MAST (19) and MLST (1, required.
38) approaches look the most promising in this ，eld. The discriminatory power of each of the nucleic acid-based In our research, four different schemes of molecular typing
TABLE 3. Multilocus linkage disequilibrium analysis of N. gonorrhoeae isolates collected in the Russian TABLE 2. Allele frequencies within N. gonorrhoeae isolates aFederation and the United Kingdom collected in the Russian Federation and athe United Kingdom bIsolate type SVe Vo P IAIA (no. of isolates) bNo. of unique variants Isolate type (no. of isolates) RUS (103) 1.5081 3.0065 0.9935 0.000 0.1656 abcZ adk aroE fumC gdh pdhC pgm UK (108) 1.4803 1.3165 0.1244 0.008 0.0207 5 7 5 RUS (103) 2 2 4 3 a Values for the expected variance (Ve), the observed variance (Vo), and the 7 6 9 UK (108) 3 2 3 3 index of association (IA), with the associated P value and the standardized index Sa The most variable alleles are indicated in boldface. A), are given. of association (Ib RUS, isolates collected in the Russian Federation; UK, isolates collected in b RUS, isolates collected in the Russian Federation; UK, isolates collected in the United Kingdom. the United Kingdom.
3688 ILINA ET AL. J. CLIN. MICROBIOL.
TABLE 4. Distribution of 103 N. gonorrhoeae isolates by susceptibility level against different sequence types according to MLST
Most prevalent phenotype (% isolates with the a indicated phenotype) ST according to MLST No. of isolates (% total) P
TET PEN CIP
ST1594 0.0573 8 (7.8) S/I S (87.5) S (100.0)
ST1892 0.0079 8 (7.8) S (87.5) S (87.5) S (100.0) ST1901 0.0002 11 (10.7) I (100.0) R (90.1) R (100.0) ST1905 S (100.0) 0.0032 11 (10.7) I (81.8) R (81.8) ST6716 R (65.0) R (100.0) 0.0000 25 (24.3) I (84.0) ST6720 5 (4.8) S/I S (100.0) S (100.0) 0.1525
a S, susceptible; I, intermediate resistant; R, resistant (in accordance with the Clinical and Laboratory Standards Institute recommendations for penicillin PEN , tetracycline TET , and cipro！oxacin CIP ). STs strongly associated with multidrug resistance are indicated in boldface. S/I, Almost equal numbers of isolates with each phenotypes.
The similar variabilities of alleles in both populations re！ect techniques was quite high. Both por typing and NG-MAST
satis，ed the value of 0.95 recommended for new typing the universality of this typing scheme and its applicability for
worldwide studies. methods by the European Society of Clinical Microbiology and
Despite the signi，cant linkage disequilibrium detected for Infectious Diseases (ESCMID) (36), in contrast to the MLST
both groups, the N. gonorrhoeae population in Russia seems scheme.
Generally, the N. gonorrhoeae typing data obtained by por more homogeneous, more clonal than in the United Kingdom. typing and NG-MAST appeared to be very similar. Both meth- For Russian isolates, the observed variance was much greater ods divided gonococcal strains into two large nonoverlapping than the maximum variance obtained in 1,000 trials (P Sgroups in accordance with porB alleles (or serovars in terms of 0.000). In combination with the Iand IA values, this suggests Aproteins). Both methods had comparable discriminatory power that the population structure of the examined N. gonorrhoeae
(0.97 and 0.98 for por typing and NG-MAST, respectively) isolates is highly clonal. It was con，rmed by the eBURST
suf，cient for identifying transmission clusters of N. gonor- analysis data shown that the majority of gonococci can be rhoeae. In addition, there was no clear geographic clustering of combined into four clonal complexes with only a singleton speci，c por types or STs according to NG-MAST found within formed by ST1594.
the examined group. Only 6 STs discovered here by MLST were described else- As far as por typing is concerned, different trends in the where, and 24 STs were newly elaborated. It should be men- evolution of porB alleles were found. Earlier, such a phenom- tioned that ST1905 was previously identi，ed only in gonococci
enon was described by Smith et al. in 1995 (28) and Posada et isolated in Uzbekistan and Russia (1). In accordance with al. in 2000 (31). This fact can be explained by different epide- MLST ，ndings, there was no evidence for geographical struc- miologies in the P1A and P1B homology groups. It was shown turing among the isolates examined, and a strong relationship that blood isolates obtained during disseminated gonococcal between the susceptibility pro，le of N. gonorrhoeae and certain
infection generally belong to the P1A serovar (5, 29), whereas STs was discovered. Moreover, two STs clearly associated with isolates from mucosal surfaces more often belong to the P1B multidrug resistance were determined by eBURST to be evo- serovar (21). It may be that the invasive process requires cer- lutionarily related (Fig. 4). If we take into account the geo- tain antigens exposed on the microbial cell surface that eluci- graphic heterogeneity of each cluster (Fig. 4; see also Table S2 date the in！uence of purifying selection on PIA gonococci. in the supplemental material), it seems that the multiresistant In contrast to por typing, NG-MAST allows international strains in Russia spread as two large clonal groups correspond- comparisons of genotyping data. Unfortunately, most of the ing to MLST ST6716 and ST1901.
STs identi，ed in our research were newly described STs. Of 23 In conclusion, we assert that the methods based on nucleic known STs, ST1527, ST1534, ST1544, and ST1548 had previ- acid variability are more suitable for N. gonorrhoeae molecular
ously been detected in N. gonorrhoeae samples obtained in typing than the ones based on changes in proteins. Among Arkhangelsk (35). It should be noted that the most common these, the NG-MAST displayed the highest resolution that is
useful in monitoring short-term transmission patterns of the ST in a European population, ST225 (26), was found in one
isolate from Irkutsk. organism and is more appropriate for local outbreak investi- Although the discriminatory power of MLST was slightly gations. Meanwhile, MLST, in spite of its lower discriminating lower than that of por typing and NG-MAST, the obtained potential, seems to be the best choice for global epidemiolog- genotyping data looked very promising. First, the tested gono- ical studies.
cocci were divided into 30 different clusters aside from their ACKNOWLEDGMENTS serotypes. In spite of the fact that N. gonorrhoeae has earlier
been proposed to have a nonclonal population structure (22, This study was supported by the Russian Ministry of Health (devel-
opment contract 06/311 [30 May 2008]). 30), four clonal complexes were identi，ed in the analyzed We thank V. A. Karpov for oligonucleotide primer synthesis and group. M. M. Chukin and T. A. Akopian for assistance with the sequence Moreover, using data from public MLST site (http://pubmlst instrumentation. .org/neisseria) allowed us to compare the N. gonorrhoeae pop- We used the NG-MAST database (http://www.ng-mast.net), Depart- ulations in the Russian Federation and the United Kingdom. ment of Infectious Disease Epidemiology, Imperial College London,
VOL. 48, 2010 N. GONORRHOEAE MOLECULAR TYPING 3689
and the Neisseria MLST website (http://pubmlst.org/neisseria/) devel- of a strain cluster with reduced susceptibility to ceftriaxone. J. Clin. Micro- biol. 47:3540–3545. oped by Keith Jolley and Man-Suen Chan (13) and located at the 21. Morse, S. A., P. G. Lysko, L. McFarland, J. S. Knapp, E. Sandstrom, C. University of Oxford. Critchlow, and K. K. Holmes. 1982. Gonococcal strains from homosexual men have outer membranes with reduced permeability to hydrophobic mol- REFERENCES ecules. Infect. Immun. 37:432–438. 1. Bennett, J. S., K. A. Jolley, P. F. Sparling, N. J. Saunders, C. A. Hart, I. M. 22. O’Rourke, M., and E. Stevens. 1993. Genetic structure of Neisseria gonor- Feavers, and M. C. Maiden. 2007. Species status of Neisseria gonorrhoeae: rhoeae populations: a non-clonal pathogen. J. Gen. Microbiol. 139:2601– evolutionary and epidemiological inferences from MLST. BMC Biol. 5:35. 2603. 2. Bilek, N., I. M. Martin, G. Bell, G. R. Kinghorn, C. A. Ison, and B. G. Spratt. 23. O’Rourke, M., C. A. Ison, A. M. Renton, and B. G. Spratt. 1995. Opa-typing: 2007. Concordance between Neisseria gonorrhoeae genotypes recovered from a high-resolution tool for studying the epidemiology of gonorrhea. Mol. known sexual contacts. J. Clin. Microbiol. 45:3564–3567. Microbiol. 17:865–875. 3. Bindu, M. P., P. Rajendran, S. P. Thyagarajan, N. P. Pramod, U. A. Rao, M. 24. Palmer, H. M., and C. Arnold. 2001. Genotyping Neisseria gonorrhoeae using Ganapathy, T. Kurien, and the STD Study Group. 2002. Molecular typing of ！uorescent ampli，ed fragment length polymorphism analysis. J. Clin. Mi- Neisseria gonorrhoeae from hospital and community isolates by restriction crobiol. 39:2325–2329. fragment length polymorphism. Int. J. STD AIDS 13:124–130. 25. Palmer, H. M., J. P. Leeming, and A. Turner. 2001. Investigation of an 4. Boom, R., C. J. A. Sol, M. M. M. Salimans, C. H. Jansen, P. M. E. Wertheim- outbreak of cipro！oxacin-resistant Neisseria gonorrhoeae using a simpli，ed van Dillen, and J. van der Noordaa. 1990. Rapid and simple method for opa-typing method. Epidemiol. Infect. 126:219–224. puri，cation of nucleic acids. J. Clin. Microbiol. 28:495–503. 26. Palmer, H. M., H. Young, C. Graham, and J. Dave. 2008. Prediction of 5. Cannon, J. G., T. M. Buchanan, and P. F. Sparlingi. 1983. Con，rmation of antibiotic resistance using Neisseria gonorrhoeae multi-antigen sequence typ- association of protein I serotype of Neisseria gonorrhoeae with ability to cause ing. Sex. Transm. Infect. 84:280–284. disseminated infection. Infect. Immun. 40:816–819. 27. Poh, C. L., Q. C. Lau, and V. T. Chow. 1995. Differentiation of Neisseria 6. Francisco, A. P., M. Bugalho, M. Ramirez, and J. A. Carric?o. 2009. Global gonorrhoeae IB-3 and IB-7 serovars by direct sequencing of protein IB gene optimal eBURST analysis of multilocus typing data using a graphic matroid and pulsed-，eld gel electrophoresis. J. Med. Microbiol. 43:201–207. approach. BMC Bioinform. 10:152. 28. Posada, G. D., K. A. Crandall, M. Nguyen, J. C. Demma, and R. P. Viscidi. 7. Fudyk, T. C., I. W. Maclean, J. N. Simonsen, E. N. Njagi, J. Kimani, R. C. 2000. Population genetics of the porB gene of Neisseria gonorrhoeae: different Brunham, and F. A. Plummer. 1999. Genetic diversity and mosaicism at the dynamics in different homology. Mol. Biol. Evol. 17:423–436. por locus of Neisseria gonorrhoeae. J. Bacteriol. 181:5591–5599. 29. Sandstrom, E. G., J. S. Knapp, L. B. Reller, S. E. Thompson, E. W. I. Hook, 8. Hobbs, M. M., T. M. Alcorn, R. H. Davis, W. Fischer, J. C. Thomas, I. and K. K. Holmes. 1984. Serogrouping of Neisseria gonorrhoeae: correlation Martin, C. Ison, P. F. Sparling, and M. S. Cohen. 1999. Molecular typing of of serogroup with disseminated gonococcal infection. Sex. Transm. Dis. Neisseria gonorrhoeae causing repeated infections: evolution of porin during 11:77–80. passage within a community. J. Infect. Dis. 179:371–381. 30. Smith, J. M., N. H. Smith, M. O’Rourke, and B. G. Spratt. 1993. How clonal 9. Hunter, P. R., and M. A. Gaston. 1988. Numerical index of the discrimina- are bacteria? Proc. Natl. Acad. Sci. U. S. A. 90:4384–4388. tory ability of typing systems: an application of Simpson’s index of diversity. 31. Smith, N. H., J. M. Smith, and B. G. Spratt. 1995. Sequence evolution of the J. Clin. Microbiol. 26:2465–2466. porB gene of Neisseria gonorrhoeae and Neisseria meningitidis: evidence of 10. Huson, D. H., D. C. Richter, C. Rausch, T. Dezulian, M. Franz, and R. Rupp. positive Darwinian selection. Mol. Biol. Evol. 12:363–370. 2007. Dendroscope: an interactive viewer for large phylogenetic trees. BMC 32. Spaargaren, J., J. Stoof, H. Fennema, R. Coutinho, and P. Savelkoul. 2001. Bioinform. 8:460. Ampli，ed fragment length polymorphism ，ngerprinting for identi，cation of 11. Ilina, E. N., V. A. Vereshchagin, A. D. Borovskaya, M. V. Malakhova, S. V. a core group of Neisseria gonorrhoeae transmitters in the population attend- Sidorenko, N. C. Al-Khafaji, A. A. Kubanova, and V. M. Govorun. 2008. ing a clinic for treatment of sexually transmitted diseases in Amsterdam, The Relation between genetic markers of drug resistance and susceptibility pro- Netherlands. J. Clin. Microbiol. 39:2335–2337. ，le of clinical Neisseria gonorrhoeae strains. Antimicrob. Agents Chemother. 33. Tapsall, J. W., E. A. Limnios, and D. Murphy on behalf of the Australian 52:2175–2182. Gonococcal Surveillance Programme. 2008. Analysis of trends in antimicro- 12. Jolley, K. A., E. J. Feil, M. S. Chan, and M. C. Maiden. 2001. Sequence type bial resistance in Neisseria gonorrhoeae isolated in Australia, 1997–2006. J. analysis and recombinational tests (START). Bioinformatics 17:1230–1231. Antimicrob. Chemother. 61:150–155. 13. Jolley, K. A., M. S. Chan, and M. C. Maiden. 2004. MlstdbNet: distributed 34. Thompson, D. K., C. D. Deal, C. A. Ison, J. M. Zenilman, and M. C. Bash. multilocus sequence typing (MLST) databases. BMC Bioinform. 5:86. 2000. A typing system for Neisseria gonorrhoeae based on biotinylated oligo- 14. Kubanova, A., N. Frigo, A. Kubanov, S. Sidorenko, T. Priputnevich, T. nucleotide probes to PIB gene variable regions. J. Infect. Dis. 181:1652– Vachnina, N. Al-Khafaji, S. Polevshikova, V. Solomka, M. Domeika, and M. 1660. Unemo. 2008. National surveillance of antimicrobial susceptibility in Neisse- 35. Unemo, M., V. Vorobieva, N. Firsova, T. Ababkova, I. Leniv, B. C. Haldo- ria gonorrhoeae in 2005–2006 and recommendations of ，rst-line antimicro- rsen, H. Fredlund, and V. Skogen. 2007. Neisseria gonorrhoeae population in bial drugs for gonorrhea treatment in Russia. Sex. Transm. Infect. 84:285– Arkhangelsk, Russia: phenotypic and genotypic heterogeneity. Clin. Micro- 289. biol. Infect. 13:873–878. 36. van Belkum, A., P. T. Tassios, L. Dijkshoorn, S. Haeggman, B. Cookson, 15. Lau, Q. C., V. T. Chow, and C. L. Poh. 1995. Differentiation of Neisseria N. K. Fry, V. Fussing, J. Green, E. Feil, P. Gerner-Smidt, S. Brisse, M. gonorrhoeae strains by polymerase chain reaction and restriction fragment Struelens, and The European Society of Clinical Microbiology and In- length polymorphism of outer membrane protein IB genes. Genitourin. fectious Diseases (ESCMID) Study Group on Epidemiological Markers Med. 71:363–366. 16. Ling, A. E., M. C. Bash, F. Lynn, N. A. Lister, P. Zhu, S. M. Garland, C. K. (ESGEM). 2007. Guidelines for the validation and application of typing Fairley, and S. N. Tabrizi. 2007. Evaluation of PorB variable region typing of methods for use in bacterial epidemiology. Clin. Microbiol. Infect. Neisseria gonorrhoeae using PCR-ELISA in samples collected from men who 2007(Suppl. 3):1–46. 37. have sex with men. J. Clin. Lab. Anal. 21:237–243. van Looveren, M., C. A. Ison, M. Ieven, P. Vandamme, I. M. Martin, K. 17. Lynn, F., M. M. Hobbs, J. M. Zenilman, F. M. Behets, K. Van Damme, A. Vermeulen, A. Renton, and H. Goossens. 1999. Evaluation of the discrimi- Rasamindrakotroka, and M. C. Bash. 2005. Genetic typing of the porin natory power of typing methods for Neisseria gonorrhoeae. J. Clin. Microbiol. protein of Neisseria gonorrhoeae from clinical noncultured samples for strain 37:2183–2188. characterization and identi，cation of mixed gonococcal infections. J. Clin. 38. Viscidi, R. P., and J. C. Demma. 2003. Genetic diversity of Neisseria gonor- Microbiol. 43:368–375. rhoeae housekeeping genes. J. Clin. Microbiol. 41:197–204. 39. 18. Martin, I. M. C., E. Foreman, V. Hall, A. Nesbitt, G. Forster, and C. A. Ison. Viscidi, R. P., J. C. Demma, J. Gu, and J. Zenilman. 2000. Comparison of 2007. Non-cultural detection and molecular genotyping of Neisseria gonor- sequencing of the por gene and typing of the opa gene for discrimination of rhoeae from a piece of clothing. J. Med. Microbiol. 56:487–490. Neisseria gonorrhoeae strains from sexual contacts. J. Clin. Microbiol. 38: 19. Martin, I. M. C., C. A. Ison, D. M. Aanensen, K. A. Fenton, and B. G. Spratt. 4430–4438. 2004. Rapid sequence-based identi，cation of gonococcal transmission clus- 40. Whiley, D. M., N. Goire, E. S. Ray, A. Limnios, S. B. Lambert, M. D. Nissen, ters in a large metropolitan area. J. Infect. Dis. 189:1497–1505. T. P. Sloots, and J. W. Tapsall. 2010. Neisseria gonorrhoeae multi-antigen 20. Monfort, L., V. Caro, Z. Devaux, A.-S. Delannoy, S. Brisse, and P. Sednaoui. sequence typing using non-cultured clinical specimens. Sex. Transm. Infect. 2009. First Neisseria gonorrhoeae genotyping analysis in France: identi，cation 86:51–55.