Molecular epidemiology of Neisseria gonorrhoeae strains circulating in Indonesia using multi-locus variable number tandem repeat analysis (MLVA) and Neisseria gonorrhoeae multi-antigen sequence typing (NG-MAST) techniques

This study was part of a larger study on gonorrhea epidemiology in Indonesia, which was partly reported elsewhere [2].

In 2014, clients of eight STI service facilities in three major cities in Indonesia (Jakarta, Yogyakarta, and Denpasar) were invited to participate in the study. A written informed consent was obtained from all participants. All participants were 16 years old or older at the day of recruitment. The participants completed a self-administered questionnaire regarding demographics, behavioral and clinical information. Based on the questionnaire data, three risk groups were defined: heterosexual males, heterosexual females, and a combined group of men who have sex with men and transwomen (MSM).

The clinicians collected urethral samples from males and transwomen, and endocervical samples from females. Primary culture and simple identification were performed at the laboratory of the Faculty of Biology, Universitas Gadjah Mada in Yogyakarta, Indonesia. Isolates were transferred on dry ice to the reference laboratory of the Public Health Service of Amsterdam, the Netherlands, where re-culture and complete identification were performed, as previously described [2]. The pure cultured Ng isolates were stored at −80 °C in beads (Microbank®, BioTrading, Mijdrecht, the Netherlands).

MLVA and NG-MAST have been used to type Ng strains at the reference laboratory in the past decade [5, 9]. A database containing molecular typing data of the strains, as well as characteristics of the patients from whom the strains were collected, was built in BioNumerics® version 7.5 (Applied-Maths, Austin, Texas, USA). To understand how Indonesian strains relate to circulating Ng strains in another setting, we used Dutch strains collected at the STI clinic of Amsterdam, the Netherlands, from two molecular epidemiology studies on azithromycin resistance [11], and ceftriaxone resistance [12]., All Dutch strains were fully susceptible for tested antibiotics as they acted as control strains in the two studies. The Dutch strains were collected between January 2012 and December 2015.

DNA lysates were prepared by boiling one stored bead seeded with cultured Ng isolates in 100 μL of phosphate buffered saline at 95 °C for 15 min.

MLVA was performed as described [9]. In short, DNA lysates were used for two different multiplex PCRs in a Bio-Rad C1000 PCR System (Bio-Rad, Hercules, California, USA) to amplify five variable number of tandem repeat sequence (VNTR) loci, i.e. VNTR04–03, VNTR04–10, VNTR07–02, VNTR30–01 and VNTR16–01. The PCR products were diluted 1:20 in water and 2 μl of each diluted sample was mixed with 18 μl of a 1:450 in water diluted GeneScan LIZ 500 size standard (Applied Biosystems, Foster City, California, USA). After heat denaturation for 5 min at 95 °C, the fragments were separated on an ABI 3130 Genetic Analyzer (Applied Biosystems, Foster City, California, USA) using the fragment analysis module. Sizing and calculation of the number of repeats of each VNTR were performed with GeneMarker® version 1.8 (SoftGenetics, LLC, State College, Pennsylvania, USA). An MLVA profile was a string consisting of integers that corresponds to the number of repeats found in the five VNTR loci [9]. This string was used for cluster analysis.

NG-MAST was also performed, as described [10]. In short, DNA lysates were used for two different PCR targets to amplify parts of the porB and tbpB genes in the Bio-Rad C1000 PCR System. The size of the amplimers in each PCR product was estimated using a QIAxcel® Advanced System (Qiagen, Hilden, Germany). PCR products with sufficient amplimers were transferred to the sequencing facility of the Academic Medical Center, University of Amsterdam, the Netherlands. An online international database of NG-MAST (www.​ng-mast.​net) was used as reference to determine allele numbers of porB and tbpB gene fragments, and to assign sequence types (STs). In addition, we checked the susceptibility profile of Indonesian strains against cephalosporins (ceftriaxone and cefixime), as previously reported [2], and examined its distribution by NG-MAST types.

We built a BioNumerics® database integrating participants’ characteristics and molecular typing data of Indonesian and Dutch reference strains.

Minimum spanning trees (MSTs) were constructed using MLVA profiles [9]. Two MSTs were constructed, i.e. an MST using data of Indonesian strains only and another MST using combined data of both Indonesian and Dutch strains. A cluster was defined as a group of ≥5 strains that differed in maximally one VNTR locus [9]. Clusters were assigned capital letters.

Based on NG-MAST STs, a genogroup was defined among strains that had one identical allele (porB or tbpB), and had another allele differing in ≤4 base pairs (bp) (tbpB) or ≤5 bp (porB), as described [13]. Genogroups were named after the most frequent STs within that genogroup. A genogroup consisting of ≥5 strains was considered as major genogroup.

Various participants’ characteristics, i.e. geographical location (city of the recruiting clinic), risk group, and age group, and NG-MAST genogroups were indicated in the MSTs.

We also examined the concordance between MLVA and NG-MAST by calculating the adjusted Wallace coefficient (AW) and its 95% confidence interval (CI) [14], using an online tool (www.​comparingpartiti​ons.​info). For each country of strains collection, the AWs were calculated if the number of strains that belonged to one of MLVA clusters and one of the major NG-MAST genogroups was large >(20 strains) [14]. The AW of MLVA to NG-MAST showed the probability of strains within an MLVA cluster to belong to the same major NG-MAST genogroup. The AW of NG-MAST to MLVA showed the probability of strains within the same NG-MAST genogroup to belong to the same MLVA cluster. The AWs range from 0.00 (no association between the two techniques) to 1.00 (a complete association).

The statistical analyses were performed in STATA version 13 (Stata Corp., College Station, Texas, USA).

From the 992 participants recruited in Indonesia, a total of 78 strains were successfully typed at the reference laboratory. The strains were from Yogyakarta (n = 44, 56.4%), Jakarta (n = 25, 32.1%), and Denpasar (n = 9, 11.5%) (Table 1). Strains were from heterosexual males (n = 5, 6.4%), heterosexual females (n = 30, 38.5%), and MSM (n = 43, 55.1%). Most of the participants (n = 46, 59.0%) were sex workers. Strains were mostly from participants aged 25–34 years (n = 34, 43.6%).

Regarding the Indonesian strains, seven clusters (cluster A-G) and 16 non-clustered strains were identified (Fig. 1 and Additional file 1: Table S1). Cluster A (n = 16) was the largest cluster. Cluster C was relatively close to cluster D, differing in only two integers of their MLVA profiles. Other clusters differed in three or more integers from each other (Fig. 1).

Clusters B, C, E and G predominantly consisted of strains from Yogyakarta, whereas clusters A and F predominantly consisted of strains from Jakarta. In clusters A, B, and F, we observed strains from all three cities and no cluster was fully specific to one geographical area (Fig. 1a). Clusters A, B, and G predominantly consisted of strains from MSM and cluster F was predominantly from heterosexual females (Fig. 1b). In clusters C, D, and E, we observed strains from all risk groups, however (Fig. 1b). Clusters B and D predominantly consisted of strains from participants aged 25–34 years (which constituted the largest age group), but in clusters E and G, we observed strains from participants of all four age groups (Fig. 1c).

The 16 non-clustered strains were mostly isolated from Yogyakarta (n = 10, 62.5%) and from MSM (n = 11, 68.8%). The non-clustered strains were: three small groups each consisting of three strains with an identical MLVA profile, one group of two strains with an identical MLVA profile, and five singletons.

The most frequently occurring NG-MAST genogroup was G1407 (51.3%, Table 1) which was found in all Indonesian cities. We found no correlation between NG-MAST genogroups and geographical area, risk group, or age group (data not shown).

We included 119 Dutch contemporary strains as reference strains (Table 1). Most were from MSM (n = 103, 86.6%), and the remainder were from heterosexual males (n = 6, 5.0%) and heterosexual females (n = 10, 8.4%). Strains from participants who aged 25–34 years (n = 39, 32.8%) composed the largest group, as was the case for the Indonesian participants.

Regarding both Indonesian and Dutch strains, we identified eight clusters, including the seven clusters that had been identified in the MST using Indonesian data only (Fig. 2, Table 2). Cluster C/D was a consolidation of cluster C and D found in MST using Indonesian data only, and one Dutch singleton. Two clusters (B and C/D) consisted of both Indonesian and Dutch strains. Two clusters (H and I) consisted of Dutch strains only. We also identified 118 non-clustered strains (16 Indonesian and 102 Dutch strains).

Indonesian strains (n = 64, 79.5%) were more likely to be included in a cluster than Dutch strains (n = 17, 14.3%, p < 0.001). An identical MLVA profile of all strains in a cluster was observed in six out of seven Indonesian clusters, whereas the two Dutch clusters showed VNTR differences within the cluster, suggesting more diversity in the Dutch clusters.

The three most frequent NG-MAST genogroups among Indonesian strains were G1407 (n = 40, 51.3%), G2992 (n = 13, 16.7%), and G9276 (n = 10, 12.8%). Four Indonesian strains with por allele 908 and tbpB allele 2413 had an unidentified ST, yet these strains could be included in genogroup G1407. Indonesian strains that belong to ST1407 did not show decreased susceptibility against cephalosporins (ceftriaxone and cefixime), as we previously reported [2]. NG-MAST data were available for 112 out of 119 Dutch strains (NG-MAST was unsuccessful in 7 strains). The three most frequent genogroups among Dutch strains were G2992 (n = 23, 19.3%), G21 (n = 15, 12.6%), and G5624 (n = 9, 7.6%) (Table 1).

In 55 Indonesian strains that could be included in an MLVA cluster and a major NG-MAST genogroup, the AW [95% CI] for MLVA to NG-MAST was 0.07 [0.00–0.27] and for NG-MAST to MLVA was 0.03 [0.00–0.12]. The AWs in Dutch strains were not calculated because of a small sample size (n = 13 strains).