Antimicrobial Resistance Mechanisms, Multilocus Sequence Typing, and NG-STAR Sequence Types of Diverse Neisseria gonorrhoeae Isolates in KwaZulu-Natal, South Africa

ABSTRACT

Antimicrobial resistance (AMR) is a major challenge to managing infectious diseases. Africa has the highest incidence of gonorrhoea, but there is a lack of comprehensive data from sparse surveillance programs. This study investigated the molecular epidemiology and AMR profiles of Neisseria gonorrhoeae isolates in KwaZulu-Natal province (KZN), South Africa. Repository isolates from patients attending public health care clinics for sexually transmitted infection (STI) care were used for phenotypic and genotypic analysis. An Etest was performed to determine antimicrobial susceptibility. Whole-genome sequencing (WGS) was used to determine epidemiology and to predict susceptibility by detecting resistance-associated genes and mutations. Among the 61 isolates, multiple sequence types were identified. Six isolates were novel, as determined by multilocus sequence typing. N. gonorrhoeae sequence typing for antimicrobial resistance (NG-STAR) determined 48 sequence types, of which 35 isolates had novel antimicrobial profiles. Two novel penA alleles and eight novel mtrR alleles were identified. Point mutations were detected in gyrA, parC, mtrR, penA, ponA, and porB1. This study revealed a high prevalence of AMR (penicillin 67%, tetracycline 89%, and ciprofloxacin 52%). However, spectinomycin, cefixime, ceftriaxone, and azithromycin remained 100% effective. This study is one of the first to comprehensively describe the epidemiology and AMR of N. gonorrhoeae in KZN, South Africa and Africa, using WGS. KZN has a wide strain diversity and most of these sequence types have been detected in multiple countries; however, more than half of our isolates have novel antimicrobial profiles. Continued surveillance is crucial to monitor the emergence of resistance to cefixime, ceftriaxone, and azithromycin.

INTRODUCTION

Neisseria gonorrhoeae, an obligate human pathogen, causes the sexually transmitted infection (STI) gonorrhoea (1). Among men, gonorrhoea infection is often symptomatic, causing urethritis, whereas among women it can be asymptomatic until complications occur (2). Untreated gonorrhoea can lead to adverse events, including acute urethritis, cervicitis, pelvic inflammatory disease, infertility, abortion, ectopic pregnancy, maternal death, and, among neonates, can lead to blindness (3–6). The World Health Organization (WHO) reported a global prevalence of 87 million new N. gonorrhoeae infections per year (7). WHO statistics show that Africa has the highest prevalence of N. gonorrhoeae, which has increased from 1.7% to 1.9% in women and from 0.5% to 1.6% in men (8, 9).

Across Africa, STIs are mainly treated according to syndromic management guidelines (10, 11), which recommend treatment of signs and symptoms of a group of diseases rather than identifying and treating a specific disease (12). The disadvantages of this approach are that asymptomatic infections are not treated, antibiotics can be overused in symptomatic patients, there is a lack of partner notification and treatment, and there is no antibiotic susceptibility testing (AST), which limits the opportunity for molecular epidemiological surveillance (13). Surveillance would provide a comprehensive understanding of the extent of antimicrobial resistance (AMR) in South Africa.

Vaccine development for gonorrhoea is still under way, but antibiotics are currently the only treatment option (14). In a global context, N. gonorrhoeae has developed resistance to all recommended antibiotic classes used for its treatment, inclusive of penicillin, tetracyclines, fluoroquinolones, aminoglycosides, and, more recently, macrolides and cephems (ceftriaxone and cefixime) (15). Molecular resistance determinants identified for recommended antibiotics include sulfonamides (folP), penicillin (bla_TEM, mtrR, penA, ponA, and porB1b), spectinomycin (16S rRNA, rpsE), tetracycline (mtrR, rpsJ, and tetM), ciprofloxacin (gyrA, parC, and parE), cefixime (mtrR, penA, rpoB, and rpoD), ceftriaxone (penA, rpoB, and rpoD), and azithromycin (23S rRNA, ermABCF, ereAB, mefA, macAB, mtrR, mtrD, mtrC, rplV, and rplD) (16, 17).

The current WHO-recommended dual therapy of ceftriaxone (250 mg as a single intramuscular dose) and azithromycin (1 g as a single oral dose) is under threat. Extended-spectrum cephalosporin (ESC) treatment failures have occurred in many countries (18, 19) and the WHO Global Gonococcal Antimicrobial Surveillance Program (GASP) survey reported that >5% of isolates were resistant to ESC in Africa, Europe, Southeast Asia, and Western Pacific regions (20). All WHO regions, except for the Eastern Mediterranean, reported that >5% of isolates were resistant to azithromycin (20). South Africa is a participating country in GASP and the latest data show no resistance to cefixime, ceftriaxone, and azithromycin (20–22); however, the data only include resistance surveys from the city of Johannesburg in the Gauteng province (23). As per WHO recommendation, local and regional AMR surveillance should guide treatment (22).

Whole-genome sequencing (WGS) technology can be used to mine for antimicrobial resistance determinants from N. gonorrhoeae isolates (24). This can result in a genotypically inferred profile that is ideal for predicting AMR (25–27) and thus may be used to optimize treatment to slow the spread of drug resistance (28). WGS has a higher and more accurate resolution of N. gonorrhoeae strains compared to traditional genotyping for epidemiological and public health purposes (29). It can elucidate the emergence, spread, and evolution of AMR at local and international levels.

N. gonorrhoeae epidemiological diversity occurs due to variability of the geographical distribution and prevalence in different populations (30). A high index of discrimination is important to differentiate between related and unrelated strains and to identify circulating strains (31). Molecular epidemiology typing methods, including multilocus sequence typing (MLST) (32), multiantigen sequence typing (NG-MAST) (33), and N. gonorrhoeae sequence typing for antimicrobial resistance (NG-STAR) (34), are used to study the association between genotypes and AMR phenotypes. WGS-based typing methods are a highly discriminatory epidemiological surveillance tool. Core genome single nucleotide polymorphism (cgSNP) analysis of relatedness is calculated based on the variant sites, whereas in core genome MLST (cgMLST), hundreds of loci are compared gene-by-gene to determine isolate relatedness.

South Africa has nine provinces with 59.62 million people; of this, KwaZulu-Natal (KZN) has a population of 11.5 million people (http://www.statssa.gov.za/?p=13453). This is the first study from KZN, South Africa that describes the AMR profiles and molecular epidemiology of N. gonorrhoeae using WGS.

RESULTS

Patient data.

The 61 N. gonorrhoeae genomes described in this study were extracted from repository isolates collected from patients attending for STI care in KZN clinics. Samples included were from 31 males and 30 females. The specimen sources were urethral and vaginal swabs for male and female patients, respectively. A third of isolates (33%, n = 20) were from the 15- to 24-year-old age group and 67% (n = 41) from the 25- to 44-year-old age group.

Phenotypic antibiotic susceptibility.

Phenotypic characterization identified a high level of AMR among the 61 isolates. MIC values for penicillin, tetracycline, ciprofloxacin, spectinomycin, cefixime, ceftriaxone, and azithromycin are listed by year of isolation (Table 1). The prevalence of N. gonorrhoeae isolates with reduced susceptibility and resistance to penicillin, tetracycline, and ciprofloxacin was 67% (n = 41), 89% (n = 54), and 52% (n = 32), respectively. Isolates with reduced susceptibility and resistance to penicillin increased from 20% in 2013 to 100% in 2016; tetracycline resistance increased from 73% in 2013 to 100% in 2016, and ciprofloxacin resistance increased from 20% in 2013 to 56% in 2016 (Fig. 1). All isolates were susceptible to spectinomycin (MIC range 2 to 32 μg/ml), azithromycin (MIC range 0.016 to 0.38 μg/ml), cefixime (MIC of <0.016 μg/ml), and ceftriaxone (MIC of <0.002 to 0.003) (Table 1).

TABLE 1.

MIC range values by year of isolation^a

Yr of isolation	PEN (μg/ml)	CIP (μg/ml)	SPT (μg/ml)	TET (μg/ml)	AZM (μg/ml)	CFM (μg/ml)	CRO (μg/ml)
2013 (n = 15)
MIC range	<0.016–>256	<0.002–2	2–32	0.032–32	0.023–0.38	<0.016	<0.002–0.002
MIC50	0.047	0.002	16	12	0.094	<0.016	<0.002
MIC90	16	0.064	24	24	0.38	<0.016	0.002
2014 (n = 24)
MIC range	0.023–>256	<0.002–4	2–24	0.038–48	0.016–0.25	<0.016	<0.002–0.003
MIC50	0.125	0.032	12	16	0.094	<0.016	<0.002
MIC90	>256	3	16	32	0.25	<0.016	0.002
2015 (n = 13)
MIC range	0.016–>256	0.19–2	2–12	6–24	0.016–0.38	<0.016	<0.002
MIC50	32	0.75	6	12	0.064	<0.016	<0.002
MIC90	>256	2	12	24	0.38	<0.016	<0.002
2016 (n = 9)
MIC range	0.064–>256	<0.002–2	3–32	0.75–32	0.023–0.25	<0.016	<0.002–0.003
MIC50	0.125	0.75	12	12	0.064	<0.016	<0.002
MIC90	>256	1.5	32	16	0.25	<0.016	<0.002

^aPEN, penicillin; CIP, ciprofloxacin; SPT, spectinomycin; TET, tetracycline; AZM, azithromycin; CFM, cefixime; CRO, ceftriaxone.

FIG 1.

(a) AMR profiles of N. gonorrhoeae isolates collected from KwaZulu-Natal, South Africa, between 2013 and 2016. (b) Predominant AMR determinants for penicillin, tetracycline, and ciprofloxacin present per year.

Genome-based susceptibility predictions.

Prior to performing genome-based mutation predictions, WGS based contamination checks and species identification using BactInspector and Kraken2 showed that the 61 isolates in this study were 100% N. gonorrhoeae. Point mutations and genes associated with AMR, as determined by Pathogenwatch, are presented in Table 2 and Table S2 in the supplemental material. The following genetic mechanisms of AMR were associated with each of the antimicrobials to which the KZN N. gonorrhoeae isolates were nonsusceptible (i.e., reduced susceptibility or resistance). Penicillin resistance determinants identified included bla_TEM (n = 27/41, 66%), penA_ins346D (n = 41/41, 100%), ponA_L421P (n = 31/41, 76%), mtrR_G45D (n = 2/41, 5%), porB1b_G120K (n = 1/41, 2%), porB1b_A121N (n = 1/41, 2%), and mtrR_disrupted (n = 1/41, 2%). Tetracycline resistance determinants identified were tetM (n = 51/54, 94%), rpsJ_V57M (n = 49/54, 91%), and mtrR_disrupted (n = 1/54, 2%). Ciprofloxacin resistance determinants identified were gyrA_S91F (n = 32/32, 100%), gyrA_D95G (n = 20/32, 63%), gyrA_D95A (n = 13/32, 41%), parC_D86N (n = 4/32, 13%), S87N (n = 12/32, 38%), and S87I (n = 1/32, 3%). blaTEM and tetM were present in 100% of isolates resistant to penicillin (n = 25/25) and tetracycline (n = 50/50), respectively. Although N. gonorrhoeae susceptibility to sulphonamides is not routinely tested for, Pathogenwatch detects for R228S, a nonsynonymous mutation found in folP. This resistance determinant was detected in 98% of isolates.

TABLE 2.

Genetic mechanisms of AMR for penicillin, tetracycline and ciprofloxacin, determined by Pathogenwatch, for 61 N. gonorrhoeae isolates collected from KwaZulu-Natal, South Africa, between 2013 and 2016

Resistance genes/mutations identified by Pathogenwatcha
Antibiotic	Gene	Mutation	Total isolates with this gene/ mutation	2013 (n = 15)			2014 (n = 24)			2015 (n = 13)			2016 (n = 9)
				No. S (%)	No. I (%)	No. R (%)	No. S (%)	No. I (%)	No. R (%)	No. S (%)	No. I (%)	No. R (%)	No. S (%)	No. I (%)	No. R (%)
PEN				n = 12/15 (80)	n = 1/15 (7)	n = 2/15 (13)	n = 5/24 (21)	n = 10/24 (42)	n = 9/24 (37)	n = 3/13 (23)	n = 1/13 (8)	n = 9/13 (69)	0	n = 4/9 (44)	n = 5/9 (56)
	blaTEM		28 (46)	1/12 (8)	1/1 (100)	2/2 (100)			9/9 (100)		1/1 (100)	9/9 (100)			5/5 (100)
	penA	ins346D	61 (100)	12/12 (100)	1/1 (100)	2/2 (100)	5/5 (100)	10/10 (100)	9/9 (100)	3/3 (100)	1/1 (100)	9/9 (100)		4/4 (100)	5/5 (100)
	ponA	L421P	37 (61)	4/12 (33)			1/5 (20)	8/10 (80)	9/9 (100)	1/3 (33)	1/1 (100)	9/9 (100)		2 (50)	2/5 (40)
	mtrR	G45D	2 (3)		1/1 (100)			1/10 (10)
	porB1	G120K	1 (2)						1/9 (11)
	porB1	A121N	1 (2)						1/9 (11)
TET				n= 4 (27)	0	n= 11 (73)	n= 3 (12,5)	n= 3 (12,5)	n= 18 (75)	0	0	n= 13 (100)	0	n= 1 (11)	n = 8 (89)
	rpsJ	V57M	53 (87)	2/4 (50)		8/11 (73)	2/3 (67)	3/3 (100)	17/18 (94)			13/13 (100)		1/1 (100)	7/8 (88)
	tetM		51 (84)			11/11 (100)		1/3 (33)	18/18 (100)			13/13 (100)			8/8 (100)
CIP				n= 12 (80)	0	n= 3 (20)	n= 13 (54)	0	n= 11 (46)	0	0	n= 13 (100)	n = 4 (44)	0	n = 5 (56)
	gyrA	S91F	34 (56)	1/12 (8)		3/3 (100)	1/13 (8)		11/11 (100)			13/13 (100)			5/5 (100)
		D95G	22 (36)	1/12 (8)		3/3 (100	1/13 (8)		6/11 (55)			7/13 (54)			4/5 (80)
		D95A	12 (20)				1/13 (8)		5/11 (45)			6/13 (46)			1/5 (20)
	parC	D86N	5 (8)			1/3 (33)	1/13 (8)		1/11 (9)			1/13 (8)			1/5 (20)
		S87N	12 (20)						5/11 (45)			6/13 (46)			1/5 (20)
		S87I	1 (2)						1/11 (9)

^aEUCAST MIC breakpoints (mg/liter): PEN (S ≤ 0.06, R > 1), TET (S ≤ 0.5, R > 1), CIP (S ≤ 0.03, R > 0.06).

Molecular typing of isolates.

WGS data, determined MLSTs and NG-STAR STs for 61 (100%) of isolates. Molecular typing differentiated the 61 KZN N. gonorrhoeae isolates into 28 MLST STs (Fig. 2), of which 5 were new profiles (15652, 15653, 15654, 15655, and 15657) with novel gdh alleles (loci 1162, 1163, and 1164). The most common MLST ST was 1588 (18%), followed by STs 14601 (11%), 1582 (8%), 1599 (8%), 13942 (7%), 8136 (7%), 12970 (3%), 13781 (3%), 15655 (3%), and 1931 (3%). Seventeen MLST STs were each associated with a single isolate. NG-STAR differentiated the 61 isolates into 45 STs, with 30 novel STs (35 isolates). Two novel penA alleles (related to penicillin binding-protein 2) were found in three isolates, and eight novel mtrR alleles (related to overexpression of the MtrCDE efflux pump) were found in 10 isolates. The most common NG-STAR ST was 357 (7%) and 1760 (7%), followed by 1627 (5%), 1632 (5%), 1628 (3%), 3404 (3%), 3413 (3%), 3415 (3%), 3431 (3%), and 3434 (3%). Thirty-five NG-STAR STs were each associated with a single isolate. The phylogenetic analysis of the 61 isolates showed a high diversity among these KZN isolates, with a Simpsons index of diversity of 0.94 using MLST STs and 0.99 as determined using NG-STAR STs.

FIG 2.

Maximum likelihood phylogenetic tree using core genome SNP analysis for 61 KZN, South Africa N. gonorrhoeae isolates between 2013 and 2016, 14 WHO strains, and the reference genome FA 1090. Phenotypic susceptible (green)/intermediate (yellow)/resistant (purple) data using the EUCAST breakpoints shown for penicillin (PEN), ciprofloxacin (CIP), tetracycline (TET), spectinomycin (SPT), cefixime (CFM), ceftriaxone (CRO), and azithromycin (AZM). All KZN isolates were susceptible to SPT, AZM, CFM, and CRO. Resistance-associated genes/mutations are included as present (blue) or absent (gray).

Core genome SNP and cgMLST analysis.

The total SNP sites represented the variation in the data. Phylogenetic analysis based on cgSNP analysis differentiated the 61 N. gonorrhoeae isolates into 3 clusters (Fig. 2). Cluster 1 grouped 27 isolates; cluster 2 grouped 32 isolates, and cluster 3 grouped two isolates. Although some of the clusters and subclusters were small (based on the small size of the overall data set), the clusters grouped based on AMR profiles and similarities in the MLST loci. In cluster 1, resistance to penicillin, ciprofloxacin, and tetracycline was mostly due to bla_TEM, penA_ins346D, gyrA_S91F, gyrA_D95G, tetM, rpsJ_V57M, and mtrR_A39T, while a subcluster (R74-R12) also had the rplD_G70D mutation. In cluster 2, all isolates were resistant to tetracycline, and a subcluster grouping isolates 311 to R36 was nonsusceptible to penicillin, ciprofloxacin, and tetracycline. The predominant resistance determinants in cluster 2 included bla_TEM, penA_ins346D, ponA_L421P, gyrA_S91F (in combination with either gyrA_D95G, gyrA_D95A, parC_D86N, or parC_S87N) and tetM (in combination with rpsJ_V57M or mtrR_A39T). Cluster 3 lacked the bla_TEM gene, and tetM was absent in all but 1 isolate. Predominant resistance determinants in cluster 3 included penA_ins346D, ponA_L421P, porB1b_G120K, gyrA_S91F, and rpsJ_V57M. The clusters also grouped together based on similarities in the MLST loci. Cluster 1 had the same abcZ, adk, and aroE loci; cluster 2 had the same adk, gdh, and pgm loci; and cluster 3 had a double locus variation in loci fumC and gdh. BIGSdb output showed that 2,092 of the core genome loci were included in this cgMLST analysis. As per Fig. S1, the clusters obtained from cgSNP analysis were confirmed by cgMLST analysis.

DISCUSSION

Our study is the first from KZN, South Africa and one of a few from Africa to provide a comprehensive WGS analysis of circulating N. gonorrhoeae genomes. These isolates showed a highly diverse strain population with a variety of AMR determinants. All isolates remained susceptible to the current treatment regimen, as well as the discontinued drug spectinomycin, and no resistance determinants were identified for these.

In contrast, AST using Etest methodology revealed that the prevalence of resistance to penicillin, tetracycline, and ciprofloxacin was high in N. gonorrhoeae isolated from our population. Our results concur with other findings from South Africa, including the GASP focal point, NICD (23, 35), and further support the implementation of the WHO recommendation to discontinue the use of these drugs for the treatment of N. gonorrhoeae in most countries (36). In our study, we saw an increase in resistance to penicillin from 20% to 100% over a 4-year period. The bla_TEM gene, which confers high-level resistance, was present in 66% (n = 27/41) of nonsusceptible and 100% (n = 25/25) of resistant isolates, and the remaining resistance was due to a combination of mutations. The determinant ponA1_L421P is involved in chromosomally mediated resistance to penicillin and was dominant in 76% (n = 31/41) of nonsusceptible isolates. Of these isolates with the ponA_L421P mutation, blaTEM was present in 68% (n = 21/31). N. gonorrhoeae resistance is multifactorial, and a combination of resistance mechanisms results in increased MICs. Resistance to tetracycline increased from 73% to 100% over a 4-year period. In our study, the rpsJ resistance-associated mutation was detected in tetracycline-susceptible isolates, which confirms that a combination of the tetM plasmid and rpsJ_V57M mutation is required to result in high-level resistance to tetracycline. Our study shows an increase in ciprofloxacin resistance from 20% in 2013 to 56% in 2016. AMR surveillance reported that, in 2015, 67% of N. gonorrhoeae isolates in Johannesburg had high-level ciprofloxacin-resistance (23), and this was also the case for high-risk men in Johannesburg (37).

While cefixime and ceftriaxone treatment failure has been reported in Japan, Europe, Canada, South America, and Australia, and extensively drug-resistant (XDR) strains were isolated from sex workers and MSM in Japan, France, and Spain (18), all isolates in our study were susceptible to spectinomycin, azithromycin, cefixime, and ceftriaxone. Although this finding is reassuring, cefixime-resistant N. gonorrhoeae was first reported among two cases in South Africa in 2012 (38). Increasing azithromycin resistance is being reported globally and threatens the effectiveness of dual therapy (39). In Johannesburg, a study from 2017 which included isolates collected from 2008 to 2015 showed the prevalence of resistance to extended spectrum cephalosporins was <1% and to azithromycin <5% (23). A study from KZN in 2019 reported 68% of isolates, collected from 2013 to 2014, were resistant to azithromycin, as determined by agar dilution (40). Another study from Johannesburg, which collected specimens between 2018 and 2019, showed that 15% of isolates from high-risk men were resistant to azithromycin (37) as determined by Etest. Our study shows that the antibiotics used in the syndromic management approach, ceftriaxone, and azithromycin, remained effective in 2016, as all isolates collected between 2013 and 2016 were susceptible at the lowest drug concentrations tested. Since all isolates tested susceptible to spectinomycin, it remains a drug that could be used in first-line treatment. Spectinomycin resistance is rare and, although it is speculated that if reintroduced as first-line treatment than resistance will be acquired rapidly, a combination of ceftriaxone and spectinomycin is currently used in Japan (18). South Korea also effectively treats gonorrhoea with spectinomycin and resistance has not been reported since 1993 (15). It is also recommended by WHO that local AMR data should guide the choice of therapy and spectinomycin (2 g IM single as a single dose) can be used as monotherapy for the treatment of genital and anorectal gonorrhoea infections (36).

WGS has a higher discriminatory power than conventional typing methods (29) and is important for analysing the emergence of AMR. We observed an overall correlation between cgSNP and cgMLST. The phylogenetic tree highlights grouped isolates with distinct AMR profiles. The majority of the MLST STs have been reported previously in many countries (Table S3). Limitations of this study include that we had a small sample size and the data obtained was from isolates collected during 2013 to 2016, such that more recent isolates are required for an updated antibiogram. The data represents patients attending primary health clinics for STI care, which excludes data from high-risk groups and obstetrics and gynecology patients.

In conclusion, antibiotic susceptibility testing of N. gonorrhoeae is crucial for etiological surveillance. We characterized N. gonorrhoeae isolates from KZN, South Africa using WGS. The isolates were highly diverse; six were novel MLST STs and the majority were novel to the NG-STAR database. No resistance to the current treatment regimen (azithromycin, cefixime, and ceftriaxone) was observed in these isolates, and MIC levels were well below the EUCAST threshold of susceptibility. However, resistance has been identified in other parts of South Africa and other African countries. Also, most of the MLST STs in this study were previously detected in many countries worldwide. This indicates a need for continued surveillance to monitor emerging resistance mechanisms and guide treatment for patient care. Until the new antibiotics for N. gonorrhoeae treatment (solithromycin, zoliflodacin, and gepotidacin) receive U.S. Food and Drug Administration (FDA) approval, the monitoring of the current treatment regimen is essential.

MATERIALS AND METHODS

Source of isolates.

The 61 N. gonorrhoeae isolates investigated in this study were from the University of KwaZulu-Natal Medical Microbiology repository. The specimens were collected from male and female patients attending KZN public health care clinics for STI care, during ethics-approved studies (2013 to 2016). Briefly, nine isolates were from 2016 from the CAPRISA 083 cohort of female patients aged 18 to 40 years (BFC410/15) (35), 13 isolates were from male patients (2015) aged 19 to 60 years (BE371/13); and 39 isolates were from male and female patients (2013 to 2014) aged 18 years and older (BE220/13). The 2016 WHO gonococcal reference strains (F, G, K, L, M, N, O, P, U, V, W, X, Y, Z) (41), ATCC 49226, and the reference genome FA 1090 were used in this study (Table S1 in the supplemental material).

Ethical approval.

Ethical approval for this study was granted by the Biomedical Research Ethics Committee of the University of KwaZulu-Natal BREC/00000097/2019.

Identification of Neisseria gonorrhoeae.

Stored N. gonorrhoeae isolates (vaginal and urethral specimens) were revived on nonselective Thayer Martin medium (supplemented with 1% Vitox) for 18 to 24 h in a 37°C 5% CO₂ incubator. Identification was confirmed using bright field microscopy, the Bactident Oxidase rapid test (Merck, Germany), and Phadebact Monoclonal GC test (Pharmacia, Sweden). WGS data confirmed the identification of N. gonorrhoeae using Kraken2 (42) and Pathogenwatch (17).

Phenotypic antibiotic susceptibility testing.

Antibiotic susceptibility testing (AST) was performed using Etest (bioMérieux, Marcy l’Etoile, France) for all isolates, with GC agar-base medium supplemented with 1% Vitox (Oxoid) (43–45). The MIC was determined as the lowest concentration of the drug to inhibit the growth of the organism. The drugs and concentration ranges were as follows; penicillin (0.016 to 256 μg/ml), ciprofloxacin (0.002 to 32 μg/ml), ceftriaxone (0.002 to 32 μg/ml), cefixime (0.016 to 256 μg/ml), spectinomycin (0.064 to 1024 μg/ml), tetracycline (0.016 to 256 μg/ml), and azithromycin (0.016 to 256 μg/ml). Susceptibility was interpreted as per the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines (46), i.e., penicillin (susceptible [S] at ≤0.06 mg/liter; resistant [R] at >1 mg/liter), ciprofloxacin (S ≤ 0.03; R > 0.06 mg/liter), ceftriaxone and cefixime (S ≤ 0.125; R > 0.125 mg/liter), spectinomycin (S ≤ 64; R > 64 mg/liter), tetracycline (S ≤ 0.5; R > 1 mg/liter), and for azithromycin the epidemiological cutoff value is 1 mg/liter. Nitrocefin was used to detect β-lactamase production (47).

Whole-genome sequencing and assembly.

DNA was extracted using the PureLink Microbiome DNA purification kit (Thermo Fisher Scientific) as per the manufacturer’s instructions. Paired-end libraries were prepared using the Nextera DNA Prep kit, followed by sequencing (2 × 75 bp) on a NextSeq platform (Illumina, Inc., USA). Raw paired-end (PE) reads were initially run through the Jekesa pipeline v1.0 (48), the –s option was used to specify the MLST schema “neisseria” for N. gonorrhoeae WGS typing. Briefly, Trim Galore v0.6.2 (49) was used to filter the PE reads (Q > 30 and length > 50 bp). De novo assembly and polishing of assemblies were performed using SPAdes v.3.13 (50) and Shovill v1.1.0 (51), respectively. Assembly metrics were calculated using QUAST v5.0.2 (52).

Molecular typing and genome-based susceptibility predictions.

Assembled contigs were used for in silico typing via pubMLST (53) to obtain abcZ, adk, aroE, fumC, gdh, pdhC, and pgm alleles and the sequence type, and NG-STAR (34) to obtain the penA, mtrR, porB1b, ponA, gyrA, parC, and 23S rRNA alleles and the sequence type, these were confirmed using Pathogenwatch (17). Novel alleles and unique allele combinations were submitted to the respective database curators for allocation of a novel ST number.

Core genome SNP analysis.

Phylogenetic trees were constructed using cgSNP alignments. Scapper (54) was used for cgSNP alignment of 61 isolate genomes and 14 international reference strains (WHO-F [LT591897], WHO-G [LT591898], WHO-K [LT591908], WHO-L [LT591901], WHO-M [LT591904], WHO-N [LT591910], WHO-O [LT592146], WHO-P [LT592157], WHO-U [LT592159], WHO-V [LT592150], WHO-W [LT592163], WHO-X [LT592155], WHO-Y [LT592161], and WHO-Z [LT592153]). The global alignment with gaps and conserved columns removed was used to generate a phylogenetic tree using iQtree (55) and viewed on Microreact (56).

Core genome mulitilocus sequence typing analysis.

Core genome multilocus sequence typing (cgMLST) was performed with The BIGSdb Genome Comparator (53) using default parameters and the N. gonorrhoeae cgMLST v1.0 scheme on our collection of 61 gonococcal isolates and 14 international reference strains (WHO-F [LT591897], WHO-G [LT591898], WHO-K [LT591908], WHO-L [LT591901], WHO-M [LT591904], WHO-N [LT591910], WHO-O [LT592146], WHO-P [LT592157], WHO-U [LT592159], WHO-V [LT592150], WHO-W [LT592163], WHO-X [LT592155], WHO-Y [LT592161], and WHO-Z [LT592153]). A distance matrix based on the number of variable alleles resolved isolates into networks using the Neighbor-Net algorithm (57) and was generated by SplitsTree (58).

Statistical analysis.

Descriptive data are provided as a number with proportion. Antimicrobial susceptibility trends were analyzed using GraphPad Prism v5.0 (GraphPad Software Inc. CA, USA). For antimicrobials currently recommended for use, MIC, MIC₅₀ (minimum concentration needed to inhibit 50% of isolates), and MIC₉₀ (minimum concentration needed to inhibit 90% of isolates) values were determined for each year. Simpsons index of diversity was used to measure strain diversity (59).

Data availability.

The whole-genome sequence project has been registered in DDBJ/ENA/GenBank with the BioProject number PRJNA681740.