Study isolates

Study isolates consisted of 35 N. gonorrhoeae isolates obtained from both male and female patients seeking treatment in selected clinics from four different regions in Kenya: Nairobi; Coastal Kenya; Nyanza (including the Kisumu and Kombewa study sites); and Rift Valley between 2013 and 2018. The isolates were collected and antimicrobial susceptibility testing done as part of an ongoing STI surveillance study titled “A surveillance study of antimicrobial susceptibility profiles of N. gonorrhoeae isolates from patients seeking treatment in selected clinics in Kenya” (WRAIR#1743, KEMRI#1908). Minimum inhibitory concentrations for: ceftriaxone; cefixime; penicillin; tetracycline; azithromycin; ciprofloxacin and spectinomycin; (S1 Table) [11] were determined using E-test® (Biomerieux) method according to manufacturer’s instructions [12,13] and the breakpoints interpreted with reference to European Committee on Antimicrobial Susceptibility Testing (EUCAST) version 8.0, 2018 standards (S2 Table). DNA isolation and whole genome sequencing was done as part of a retrospective laboratory based molecular sub-study titled “Molecular characterization of antimicrobial resistance genes in Neisseria gonorrhoeae isolates from Kenya through whole genome sequencing” which was nested in the STI surveillance programme. Genomic DNA was extracted using QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) and the quality and quantity of DNA determined by Qubit dsDNA HS Assay using Qubit 3.0 fluorometer, (Thermo Fisher Scientific Inc. Wilmington, Delaware USA) according to the manufacturer’s instructions.

Whole genome sequence data, assembly and annotation

Illumina Nextera XT kit (Illumina Inc. San Diego, CA, USA) was used to prepare libraries as per manufacturer’s instructions. Sequence reads were generated on Illumina MiSeq platform (Illumina, San Diego, CA, USA) using a paired-end 2 × 300 bp protocol [14]. Sequence reads were assembled with CLC Genomics Workbench (v12.0, Qiagen). Genome annotation and analysis were conducted with BIGSdb tools accessed through the PubMLST platform (www.pubmlst.org/neisseria). The sequence data generated in this study are publicly available. Raw sequence reads have been deposited in the NCBI BioProjects PRJNA481622 and PRJNA590515, accessible at https://www.ncbi.nlm.nih.gov/bioproject/PRJNA481622 and https://www.ncbi.nlm.nih.gov/bioproject/PRJNA590515 respectively. Genome assemblies are accessible via PubMLST at https://pubmlst.org/. Specific accession numbers for genome assemblies within PubMLST are provided in the S3 Table.

Multi-locus Sequence Typing, NG-MAST and NG STAR

Identification of multi-locus sequence types (MLST) was performed on assembled genome sequences using MLST version 1.8 (previously hosted at https://cge.cbs.dtu.dk/services/MLST/) available online at the Centre for Genome Epidemiology (CGE) [15]. At the time of analysis, this platform provided a streamlined workflow for batch processing and generated allele definitions directly comparable to PubMLST, ensuring consistency and interoperability of sequence type assignments. MLST typing is now supported through the updated CGE platform (https://cge.food.dtu.dk/services/MLST/) or directly via PubMLST.

A local BLAST search against NG-MAST porB and tbpB databases was used to identify porB and tbpB genes from the assembled genomes using BioEdit sequence alignment editor version 7.0.5 [16]. Determination of porB and tbpB alleles and NG-MAST sequence types (STs) was originally performed using the NG-MAST website (www.ng-mast.net), which at the time was the standard resource for allele assignment. NG-MAST typing is now maintained within PubMLST, and future analyses should use the PubMLST-hosted NG-MAST scheme to ensure sustainability and consistency of results.

Determination of N. gonorrhoeae Sequence Typing for Antimicrobial Resistance (NG-STAR) STs was carried out using the NG-STAR scheme implemented in PubMLST [7]. While PubMLST hosts a version of NG-STAR, the primary host for the database is https://ngstar.canada.ca/, which should be used for future analyses to access the most up-to-date allele definitions and sequence type assignments.

Phylogeny analysis

To identify phylogenetic relationships among the study sequences, genome comparator tool hosted on pubmlst.org/neisseria was used to compare whole genome sequence and core genome-based phylogeny inferred using cgMLST N. gonorrhoeae v.1.0 scheme. The generated data was visualized using Inkscape. ITOL tool hosted on pubmlst.org/neisseria was used to create a neighbor-joining tree from concatenated seven NG-STAR loci nucleotide sequences. The core genome comparison data was annotated by NG-STAR, NG-MAST and MLST schemes. To analyze clustering of sequence types in global context, ITOL was used to generate a neighbor-joining tree based on NG-STAR, NG-MAST and MLST schemes. To provide regional and global context, 55 additional N. gonorrhoeae genome sequences were retrieved from PubMLST. These included 35 isolates from other African countries (Uganda, Malawi, South Africa, Ethiopia, Eritrea, Morocco, Ghana, Zambia, Lesotho, Madagascar, Rwanda, Burkina Faso and Cameroon) to capture regional genetic diversity, and a randomized sample of 20 non-African isolates representing major global lineages (such as ST-1901 and ST-1407). Selection was prioritized for available isolates with collection dates between 2010 and 2020 to align with the study period.

Statistical analysis

All proportions are reported with 95% confidence intervals (CIs) using the binomial exact method. Statistical comparisons were performed using Mann–Whitney U tests for continuous variables and Fisher’s exact tests for categorical variables. Given the small sample size and exploratory nature of this study, p-values were not corrected for multiple testing and should be interpreted with caution. We abstracted sex, age, marital status, and self-reported partner count from clinic records. For partner count, categories were collapsed to >2 vs ≤ 2/none due to sparse upper cells. Genomic novelty was indicated by the presence of a novel sequence type in any scheme (MLST, NG-MAST, NG-STAR).

Ethical consideration

Permission to carry out the study for both the STI surveillance study and the nested molecular sub-study were obtained from Kenya Medical Research Institute (KEMRI) Scientific and Ethics Review Unit) and Walter Reed Army Institute of Research Institutional Review Board as (WRAIR#1743, KEMRI#1908) and (WRAIR#1743A, KEMRI#3385) “respectively”. Consent to participate was not applicable for this study because it was retrospective laboratory based, used archived samples, and there was no interaction with human subjects. All data were fully anonymized before we accessed them and the research was qualified as not involving human subjects. The isolates were acquired in two batches on 23/1/2017 and 26/2/2018.

N. gonorrhoeae multi-locus sequence typing

A total of 22 MLST STs representing: 3 abcZ; 2 adk; 3 aroE; 6 fumC; 5 gdh; 3 pdhC; and 2 pgm different alleles were identified. Of the 35 sequences, 24 belonged to 14 known MLST STs; ST-1588; ST-1599; ST-1893; ST-1921; ST-1928; ST-1932; ST-8111; ST-8133; ST-11242; ST-11365; ST-11366; ST-11367; ST-11750; and ST-11976; while the remaining 11 sequences belonged to 8 new STs; ST-13613; ST-13614; ST-13763; ST-13764; ST-13766; ST-13779; ST-13780; and ST-13782 (S4 Table). The predominant ST was ST-1932 (n = 5, 14.3%) followed by ST-8133 (n = 4, 11.4%) and a novel ST-13782 (n = 3, 8.6%). In 2015, 44.4% (4/9) of the isolates displayed novel MLST types. Twenty isolates formed seven groups (two or more same ST) while 15 isolates formed singular MLST STs. Unlike ST-1932 and ST-13782, ST-8133 consisted of isolates recovered from one region; Nyanza. Region-based occurrence was also observed in ST-11365 and ST-13780 which consisted of two isolates each from Nyanza.

N. gonorrhoeae multi-antigen sequence genotyping

In total, 29 NG-MAST sequence types were detected, encompassing 22 distinct porB alleles and 21 tbpB alleles. Nine of both the 22 porB and the 21 tbpB alleles were novel. Thirty‑one isolates (88.6%) belonged to 26 novel NG‑MAST STs, including ST‑18599, ST‑19087, ST‑19166 through ST‑19170, and ST‑19254 through ST‑19272. Four isolates (11.4%) belonged to three previously known STs: ST‑355 (n = 1), ST‑10134 (n = 2), and ST‑11752 (n = 1). The most common ST overall was ST‑19168, identified in four isolates (11.4%) from Nyanza. All four of these isolates also belonged to MLST ST‑8133, which was the second most frequent MLST type observed. Sequence types ST‑19255, ST‑19262, and ST‑10134 were each identified in two isolates, making them the second most common after ST‑19168. In total, four ST groups (ST‑19168, ST‑19255, ST‑19262, and ST‑10134) accounted for 10 isolates, while the remaining isolates each represented unique NG‑MAST STs (S5 Table). Regional clustering was observed only in two novel STs: ST‑19168 and ST‑19255, both of which comprised isolates recovered exclusively from Nyanza between 2014 and 2018.

N. gonorrhoeae Sequence Typing for Antimicrobial Resistance (NG STAR)

Twenty-three NG-STAR STs representing 9 mtrR, 5 penA, 1 23S rRNA, 3 gyrA, 6 parC, 2 ponA, and 6 porB different alleles were identified. Seven (20%) isolates belonged to 6 novel NG-STAR STs: ST-3179; ST-3182; ST-3183; ST-3184; ST-3185 and ST-3186 while the rest of the isolates belonged to 17 known NG STAR STs. The most common STs were ST-1586 and ST-2668 which both comprised 4 (11.4%) isolates each. One of the two predominant NG-STAR ST, ST-2668 comprised isolates from Nyanza which also belonged to the most common NG-MAST, ST-19168 and to the second common MLST, ST-8133. Six isolates from Nyanza shared identical NG-STAR, NG-MAST, and MLST sequence types. Nineteen isolates formed 7 NG STAR ST groups while the rest formed singular NG-STAR STs (S6 Table)

Temporal distribution of novel sequence types

Across 35 isolates analyzed between 2013 and 2018, the emergence of novel sequence types was observed across all three typing schemes. The yearly distribution of isolates was: 2013 (n = 2), 2014 (n = 4), 2015 (n = 9), 2016 (n = 9), 2017 (n = 7), and 2018 (n = 4). Due to the small sample sizes in 2013, 2014, and 2018, temporal trends in the proportion of novel sequence types should be interpreted with caution. Novel MLST STs were detected between 2014 and 2017, with the highest proportions observed in 2015–2016 (44.4% each). NG-MAST showed high rates of novelty throughout the study period, with novel STs identified in every year. While the proportion of novel NG-STAR STs appeared higher in 2017 (57.1%) and 2018 (75.0%), these figures represent a small number of isolates (4/7 and 3/4, respectively) and may not reflect broader population shifts (Fig 1). Collectively, these findings underscore the contrasting capacities of the three typing schemes: MLST captured moderate levels of novelty, NG‑MAST identified extensive novel diversity, and NG‑STAR revealed fewer but temporally clustered novel STs.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Temporal trends for novel MLST, NG-MAST, and NG-STAR sequence types (2013–2018).

Percentage of isolates assigned to novel sequence types across MLST, NG‑MAST, and NG‑STAR typing schemes between 2013 and 2018. Bars represent the proportion of isolates with novel STs per year for each scheme (MLST in blue, NG‑MAST in orange, NG‑STAR in gray). MLST demonstrated moderate novelty, peaking in 2015–2016 (44.4%) but absent in 2018. NG‑MAST revealed extensive novelty, with several years (2014, 2017, 2018) reaching 100%. NG‑STAR exhibited the lowest overall novelty (20%), with novel STs concentrated in later years, particularly 2017 (42.9%) and 2018 (75%).

https://doi.org/10.1371/journal.pone.0335831.g001

Associations between epidemiological covariates and genomic novelty

Of the 35 cases (male 85.7%; median 20–29 years, IQR 22–31), “any novel sequence type” was common across demographic strata. No significant associations were detected between novelty and age (median 20–29 years vs. 30–39 years; U = 18.0, p = 0.322) or sex (male vs female: OR=7.25, p = 0.269). The high prevalence of novelty across strata suggests broad dissemination of novel lineages in the catchment population rather than concentration in specific demographic groups

Core genome-based phylogeny (cgMLST N. gonorrhoeae v.1.0 scheme)

Genome comparator tool identified 1485 loci as core genomes to N. gonorrhoeae. cgMLST N. gonorrhoeae v.1.0 scheme identified four clusters (group of more than 3 isolates) among the study isolates: Cluster 1 (n = 15); Cluster 2 (n = 6); Cluster 3 (n = 5) and Cluster 4 (n = 4). Cluster 1 was further sub-divided into two smaller but distinct sub-clusters; sub-cluster 1a (n = 9) and sub-cluster 1b (n = 6). Isolates in sub cluster 1b were from 3 regions and belonged to 3 different MLST, NG-MAST, and NG-STAR STs. Isolates in sub-cluster 1a belonged to: 5 MLST ST (3 novel and 2 existing STs), 8 NG_MAST STs (6 novel and 2 existing STs) and 4 NG-STAR STs. Eight of the nine isolates in this sub-cluster were all from Nyanza. Cluster 2 was formed by isolates from Nairobi and Nyanza belonging to: 2 existing MLSTs STs; 5 NG-MAST STs and 6 NG-STAR STs. Cluster 3 was formed by four isolates from Nyanza and one from Rift Valley. They belonged to; four different MLST and NG- MAST STs (all novel) and 2 NG-STAR STs. Isolates in cluster 4 do not form a true cluster but rather represent a set of genetically diverse strains, belonging to three existing MLST STs, four novel NG-MAST STs, and four NG-STAR STs, and were not as closely clustered as in the other three groups (Fig 2).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Core genome-based phylogeny inferred from 1,485 loci identified as core in 35 Kenyan Neisseria gonorrhoeae isolates using the cgMLST N. gonorrhoeae v.

1.0 scheme. Three distinct clusters (1-3) were formed; isolates designated as Cluster 4 do not represent a true cluster but instead comprise genetically diverse strains belonging to multiple sequence types. Shown in the leaf terminals are PubMLST IDs for the isolates, along with their MLST, NG-MAST, and NG-STAR sequence type assignments. For clarity, ‌‌all isolate identifiers and sequence types are summarized in S3 Table.

https://doi.org/10.1371/journal.pone.0335831.g002

NG-STAR phylogeny

Neighbor-joining tree inferred by ITOL identified 3 distinct clusters: Cluster A (n = 12); Cluster B (n = 9); and Cluster C (n = 6) (Fig 3). Cluster A was formed by 12 isolates from all the four sampled regions. These isolates had varied antibiotic susceptibility profiles. Ten of these 12 isolates belonged to 6 already known MLST STs while the remaining 2 isolates belonged to a novel MLST ST. On the contrary more diversity was observed in this cluster with regard to NG-MAST and NG STAR STs as these isolates belonged to ten NG-MAST and ten NG-STAR STs. Nine of these ten NG-MAST STs formed by 10 isolates were novel while one ST formed by two isolates was an existing one (ST-10134).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Neighbor-joining tree generated by ITOL from concatenated sequences of seven NG-STAR loci (mtrR, penA, 23S rRNA, gyrA, parC, ponA, and porB) for 35 Kenyan Neisseria gonorrhoeae isolates.

Antimicrobial susceptibility profiles for isolates with reduced susceptibility to ceftriaxone (CRO; MICs > 0.002 mg/L) and those resistant to penicillin (PEN), tetracycline (TET), ciprofloxacin (CIP), and azithromycin (AZM) are indicated. NG-STAR sequence type labels are included to illustrate how specific STs cluster in relation to antimicrobial resistance profiles, complementing the phylogeny inferred from concatenated loci. Three distinct clusters (A-C) comprised isolates with varied MLST, NG-MAST and NG-STAR STs. Branch lengths represent genetic distances between sequences.

https://doi.org/10.1371/journal.pone.0335831.g003

The nine isolates forming cluster B were also from all the sampled regions and belonged to six existing MLST STs. The isolates belonged to six novel NG-MAST STs and six NG-STAR STs. They were all tetracycline resistant, while 8 of the 9 were ciprofloxacin resistant. Additionally, three isolates with a low-level resistance to azithromycin all belonged to this cluster. Cluster C was formed by six isolates; five from Nyanza and one from Rift Valley. Unlike cluster A and B which mostly consisted of existing MLST STs, cluster C isolates belonged to four novel MLST STs and one existing MLST ST. The isolates were all tetracycline and ciprofloxacin resistant. NG-STAR, NG-MAST and MLST region-based clustering was not observed in either of the two phylogenies as the identified clusters comprised isolates with varied STs. Resistance to ciprofloxacin, penicillin, and tetracycline was widespread across the three typing schemes (MLST, NG-MAST, NG-STAR).

Clustering of Kenyan isolates in a global and regional context

To assess the genetic relationship of Kenyan N. gonorrhoeae isolates within a broader context, a neighbor-joining tree was constructed using MLST, NG-MAST, and NG-STAR schemes, incorporating 50 additional sequences from PubMLST (Fig 4). The updated analysis reveals that Kenyan isolates are distributed across multiple distinct clusters, showing both regional affinity and evidence of localized evolution. A significant proportion of Kenyan isolates clustered closely with sequences from other African countries, including Uganda, Malawi, South Africa, Cameroon, and Burkina Faso. Notably, Kenyan isolates with novel NG-MAST and NG-STAR types often branched within these African-centric clades, indicating that while these specific sequence types are newly described, they have emerged from established regional genetic backgrounds.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Global and Regional Phylogenetic Context of Kenyan N. gonorrhoeae Isolates.

Neighbor-joining tree generated by ITOL using MLST, NG-MAST, and NG-STAR schemes (n = 90). The analysis includes 35 Kenyan isolates from the present study (2013−2018) and 55 additional genome sequences from PubMLST, comprising 35 regional African and 20 randomized global reference strains representing major multidrug-resistant (MDR) lineages. Kenyan isolates are distributed across multiple distinct clusters, showing strong regional affinity with other African sequences, particularly those from neighboring East African countries. In contrast, the Kenyan isolates remained phylogenetically distinct from the major global MDR lineages (ST-1901 and ST-1407),‌‌ which predominantly clustered within non-African clades. PubMLST IDs, MLST, NG-MAST, and NG-STAR sequence types are shown in the leaf terminals.

https://doi.org/10.1371/journal.pone.0335831.g004

In contrast, the Kenyan isolates remained phylogenetically distinct from the major global multidrug-resistant (MDR) lineages. A randomized sample of non-African isolates, including those from Europe (UK, Spain, France), the Americas (USA, Brazil, Argentina), and Asia (Japan, China, Vietnam), predominantly clustered within well-known high-risk lineages such as MLST ST-1901 and NG-MAST ST-1407 [17,18]. None of the Kenyan isolates from this study period (2013–2018) clustered within these pandemic MDR clades, reinforcing the observation that the Kenyan gonococcal population during this time was characterized by locally and regionally evolving lineages rather than the introduction of global MDR clones.