Single-Locus-Sequence-Based Typing of the mgpB Gene Reveals Transmission Dynamics in Mycoplasma genitalium

Sexually transmitted infections (STIs) by Mycoplasma genitalium are a major problem worldwide, especially given their marked and rapid propensity for developing antimicrobial resistance. Since very few treatment options exist, clinicians face an important challenge in the management of the infection. In this scenario, little is known regarding the transmission dynamics of M. genitalium and the epidemiology of antimicrobial resistance. This mgpB-based molecular typing study, conducted among 54 asymptomatically infected individuals prospectively recruited from an STI screening service, reveals two distinct epidemiological clusters that significantly correlate with sexual conduct in heterosexuals and men who have sex with men (MSM), respectively.

ABSTRACT

Sexually transmitted infections (STIs) by Mycoplasma genitalium are a major problem worldwide, especially given their marked and rapid propensity for developing antimicrobial resistance. Since very few treatment options exist, clinicians face an important challenge in the management of the infection. In this scenario, little is known regarding the transmission dynamics of M. genitalium and the epidemiology of antimicrobial resistance. This mgpB-based molecular typing study, conducted among 54 asymptomatically infected individuals prospectively recruited from an STI screening service, reveals two distinct epidemiological clusters that significantly correlate with sexual conduct in heterosexuals and men who have sex with men (MSM), respectively. This well-defined structuration suggests the presence of two independent sexual networks with little connectivity between them. On the other hand, the study demonstrates the multiclonal feature of the emergence of antibiotic resistance in M. genitalium to both macrolides and fluoroquinolones. The high prevalence of macrolide resistance in M. genitalium among MSM, influenced by dense network connectivity and strong antibiotic selective pressure, may correspond to allodemics affecting other STIs such as gonorrhea, syphilis and enteric pathogens. Collaterally, the structural and functional impact of mutations in the mgpB gene, encoding the major adhesin P140 (MgpB), may require further investigation.

INTRODUCTION

Mycoplasma genitalium was isolated for the first time in 1981 from two men with nongonococcal urethritis (1). However, the first polymerase chain reactions (PCRs) for its detection were not developed until the early 1990s (2). Nowadays, M. genitalium is well recognized as a major cause of urethritis and further urogenital syndromes in both men and women (3). Because of the limited structural complexity of the bacterium, M. genitalium antimicrobial susceptibility is restricted to a very small range of antibiotics. Current international guidelines recommend the macrolide azithromycin, given as an extended dose, for treatment of uncomplicated M. genitalium infection (4). Nevertheless, the wide use of this antibiotic for the syndromic management of sexually transmitted infections (STIs) has likely enhanced the spread of macrolide resistance worldwide (5, 6). Single nucleotide mutations in domain V of the 23S rRNA gene are consistently linked with this resistance (7). In this scenario, the broad spectrum fluoroquinolone moxifloxacin is currently the only efficient therapeutic alternative in most parts of the world (4). However, moxifloxacin treatment failures are emerging in alarming numbers in many countries and are associated with polymorphisms in the quinolone-resistance determining region (QRDR) of the parC gene, mainly affecting amino acids S83 and D87 (M. genitalium numbering) (8–10).

Sexual transmission in M. genitalium was first suggested by looking at concordance of infections between sexual partners (11), and, shortly after, it was evidenced with molecular epidemiology using a single-locus-sequence-based typing (SLSBT) system in positive samples from infected couples (12). This SLSBT system was originally developed from a diagnostic PCR assay (2). After revealing high efficiency and excellent discriminatory power, the mgpB (MG191) gene region targeted by the MgPa-1/MgPa-3 primer set was proposed for molecular typing purposes (13). Subsequently, this efficient molecular typing method, based on the sequence of a 231 bp nonrepeated region of the MgPa operon (14), has been robustly used in general epidemiological studies, where it demonstrates reproducibility and stability (12, 15, 16). Nevertheless, little is known regarding the structure of sexual networks and the transmission and spread of antimicrobial resistance in M. genitalium.

Here, we present a wide, region-specific and cross-sectional combined analysis of molecular epidemiology, antibiotic resistance, and epidemiological data from M. genitalium asymptomatic infections. With this approach, our study aimed to explore the transmission dynamics of M. genitalium infection and the spread of antimicrobial resistance.

MATERIALS AND METHODS

Sample selection.

Samples were collected at Drassanes Exprés, as a part of a resistance-guided M. genitalium screening study (17), between October 2017 and January 2018. Briefly, Drassanes Exprés is a point-of-care service for rapid STI screening belonging to the Vall d´Hebron University Hospital (HUVH) in Barcelona, Spain. Here, testing for human immunodeficiency virus (HIV), syphilis, Chlamydia trachomatis, and Neisseria gonorrhoeae is routinely offered to asymptomatic men and women. Individuals reporting symptoms and sexual contacts with infected partners are excluded from this screening program. Thus, during the study period, participants were also tested for M. genitalium and genotypic markers of antibiotic resistance.

All participants provided their consent and ethical approval for the study was obtained from the HUVH Ethics Committee (209/17).

Laboratory procedures.

For the parent study (17), first-void urine samples, vaginal swabs, and rectal swabs (if receptive anal intercourse was reported) from asymptomatic participants were tested for M. genitalium and macrolide resistance using the ResistancePlus MG kit (SpeeDx, Sydney, Australia) at the Microbiology Department of the HUVH (18). Mutants were then confirmed by sequencing of the 23S rRNA gene using a previously described methodology (7). Fluoroquinolone resistance-associated parC mutations were also retrospectively studied as previously described (19). Positive primary samples were stored at –20°C and sent to the Statens Serum Institut in Copenhagen, Denmark, for subsequent analyses.

For the SLSBT procedure, specimens were reextracted using 20% Chelex 100. PCR amplification and sequencing of the 231 bp region of the mgpB gene was performed as previously described using the MgPa-1/MgPa-3 primer set (2, 12). M. genitalium G37 published genome sequence was used as a reference (accession number NC_000908.2) (20).

Data analysis.

The epidemiological characteristics of individuals were collected from a routine questionnaire that every patient at the Drassanes Exprés must complete. Sexual behavior was defined regarding the gender of sexual partners reported by the participant. Statistical analyses were performed using Stata (StataCorp., College Station, TX, USA).

Sanger sequences were analyzed using MEGA (version 6.0) software. Furthermore, to visualize the relationships between genotype profiles, a maximum likelihood phylogeny was constructed with IQ-TREE (version 1.6.10) software using a K3Pu+F+I model. Branch support values were generated from 1,000 bootstrap replicates. The phylogenetic tree was then complemented using iTOL (version 4.4.2).

RESULTS AND DISCUSSION

A total of 70 M. genitalium-positive specimens from 66 individuals were collected during the recruiting period of the parent study (17). Of these, 65 samples from 61 participants were suitable for the subsequent SLSBT analysis and sequencing was successful in 57 samples (87.7%) from 54 individuals (88.5%) (Fig. S1 in the supplemental material).

An important aspect of a typing method is reproducibility, and this characteristic has already been established for the present typing system (12). In our study, three participants (two men and one woman, identified as MGSP_118-9, MGSP_144-5, and MGSP_124-5, respectively) had M. genitalium infections in both genital and rectal locations at the same time. These concurrent infections found in different anatomical sites shared the same sequence profile in each individual, confirming the high reproducibility of this molecular typing method.

Overall, 32 different sequence types (STs) were described among the 54 infected individuals included in the study. None of the STs was identical with the reference ST_G37. Despite the remarkable genetic heterogeneity, some strains from different participants revealed an identical ST (Table 1). This is especially the case for ST_5, found in 15 individuals, of whom 14 (93.3%) were men who have sex with men (MSM). This cluster may suggest either the spread of a certain clone in the MSM population or, alternatively, a common mgpB ST. In addition, similar and identical STs among participants appeared at different time points during the study period.

TABLE 1.

Sequence types present in more than one individual^e

Sequence type	Participant (MGSP)	Sexual behavior	Infection site	HIV status	MR genotypea	QR genotypeb	Study week
ST_1	8	MSW	Urethra	N	WT	WT	2
	133	MSW	Urethra	N	WT	WT	6
	165	MSW	Urethra	N	WT	WT	12
ST_2	110	MSM	Rectum	N	A2071G	G248T (S83I)	3
	150	Women	Vagina	N	WT	WT	9
ST_3	117	MSM	Rectum	N	A2071G	WT	4
	126	MSM	Rectum	P	A2071G	WT	6
	146	MSM	Rectum	P	A2071G	WT	9
	152	MSM	Rectum	N	A2071G	WT	11
ST_4	118-9	MSMBI	U&R	N	A2072G	WT	4
	158	MSM	Rectum	N	A2071G	WT	12
ST_5	80	MSM	Rectum	N	A2072G	WT	2
	154	MSM	Rectum	N	A2072G	WT	11
	149	MSM	Rectum	N	ND	WT	9
	62	MSM	Rectum	P	A2072G	WT	2
	144-5	MSM	U&R	N	A2072G	WT	8
	129	MSM	Rectum	P	MXDc	WT	5
	160	MSM	Rectum	N	A2071G	WT	12
	173	MSM	Rectum	N	MXDd	WT	13
	147	MSW	Urethra	N	WT	WT	9
	172	MSM	Urethra	P	A2072G	WT	13
	171	MSM	Rectum	N	MXDd	WT	13
	163	MSM	Rectum	N	WT	WT	12
	116	MSM	Rectum	N	ND	WT	4
	161	MSMBI	Urethra	N	A2071G	WT	12
	132	MSM	Urethra	N	A2071G	WT	7

^aMacrolide resistance-associated mutations in the 23S rRNA gene (M. genitalium numbering).

^bFluoroquinolone resistance-associated mutations in the parC gene (M. genitalium numbering).

^cMGSP_129 presented both genotypes A2072G and wild-type in the 23S rRNA gene.

^dBoth MGSP_173 and MGSP_171 strains harbored genotypes A2072G, A2071G and wild-type in the 23S rRNA gene.

^eAbbreviations: ST, sequence type; HIV, human immunodeficiency virus; MR, macrolide resistance; QR, quinolone resistance; MSW, men who have sex with women; MSM, men who have sex with men; MSMBI, bisexual men; N, negative; P, positive; WT, wild-type (antibiotic resistance-associated mutations not detected); U&R, urethra and rectum; MXD, mixed genotypes; ND, not determined (i.e., infection reported as macrolide resistant by the ResistancePlus MG kit but not confirmed with sequencing).

The dendrogram in Fig. 1, based on the genotype profiles from the mgpB-based SLSBT method, was complemented with additional information: gender/sexual conduct, infection site, HIV serostatus, study week of specimen collection, and the presence of macrolide and fluoroquinolone resistance-associated genotypic markers in the 23S rRNA and the parC genes, respectively. Overall, individuals with M. genitalium infection can be separated into two distinct clusters that also significantly correlate with sexual conduct. The major genotypic cluster 1, which includes reference ST_G37, mainly comprises infections occurring in women and heterosexual men (MSW) (11/17; 64.7%), while major cluster 2 mostly comprises infections in MSM and bisexual men (MSMBI) (31/37; 83.8%); P < 0.001. This characteristic clustering structure reveals transmission dynamics of M. genitalium infections among a general asymptomatic population. The correlation between sexual behavior and genotype was translated into the structuration of well-defined epidemiological clusters, and may suggest the presence of two independent sexual networks with little connectivity between them. Consequently, the spread of genetic diversity and antimicrobial resistance may be limited to each sexual transmission group. In this hypothetical scenario, the MSMBI group acts as a bridge between the disjoint populations of women/MSW (major cluster 1) and MSM (major cluster 2). Despite that vaginal sex may be the major route of M. genitalium transmission in terms of risk (17, 21, 22), anal intercourse in MSM is likely the most common cause of spread because of dense sexual networks in this subpopulation (23). In this sense, several studies have reported higher prevalence of urethral infection among men encountering vaginal sex compared to strict MSM (16, 21, 22). The frequent subclinical course of infections in the vagina and, especially, in the rectum makes both locations a common but often unnoticed reservoir for M. genitalium infection and the selection of antibiotic resistance.

FIG 1.

Clustering dendrogram using the mgpB-based single-locus-sequence-based typing from M. genitalium strains of asymptomatically infected individuals. The sequence of the M. genitalium G37 strain is used as a reference (accession number NC_000908.2). Macrolide resistance-associated mutations in the 23S rRNA gene and fluoroquinolone resistance-associated mutations in the parC gene are also shown (M. genitalium numbering). The last column (w) accounts for the study week of specimen collection. The red dashed lines constitute clusters mostly represented by MSM. Bootstrapping values are presented in Fig. S2. Abbreviations: MSM, men who have sex with men; MSMBI, bisexual men; MSW, men who have sex with women; HIV, human immunodeficiency virus; ND, not determined (i.e., infection reported as macrolide resistant by the ResistancePlus MG kit, but not confirmed with sequencing); MXD, mixed genotypes (i.e., MGSP_129 presented genotypes A2072G and wild type, while both MGSP_173 and MGSP_171 strains harbored genotypes A2072G, A2071G, and wild type in the 23S rRNA gene).

Estimates on the prevalence of antibiotic resistance in this asymptomatic cohort can be found in previous investigations (17, 19). In the current context of the exponential rise of antimicrobial resistance in M. genitalium worldwide, some studies have suggested selection of certain clones as a possible cause of this resistance spread (24, 25). However, in accordance with other investigations (16, 26–28), our study does not support the idea of a clonal phenomenon but demonstrates the multiclonal feature of the emergence of antibiotic resistance in M. genitalium to both macrolides and fluoroquinolones. Thus, regarding macrolide resistance, mutations A2071G and A2072G in the 23S rRNA gene (M. genitalium numbering) were randomly distributed among STs in our study. Despite the limited report of fluoroquinolone resistance-associated parC polymorphisms in our cohort, mutations G248A (S83N) and G259T (D87Y) appeared among distinct ST clones.

The challenge of antimicrobial resistance in MSM, particularly to macrolides, requires further discussion. As previously pointed out by several authors, macrolide resistance in M. genitalium is not homogeneously distributed among the general population but the prevalence in MSM is usually overwhelming (17, 29–31). Furthermore, our study suggests that a significant transmission of this resistance to heterosexual men and women is unlikely. This specific phenomenon in MSM relies on different structural and environmental characteristics such as a dense network connectivity and an antibiotic selective pressure enhanced by the wide use of azithromycin in STIs. The term “allodemics” was first introduced by Baquero et al. to describe the polyclonal spread of extended-spectrum-beta-lactamase (ESBL) producing bacteria in a clinic in Spain, as a consequence of the selective pressure exerted by the use of cephalosporins in that hospital (32). Thus, as already pointed out by Kenyon et al., the phenotype of macrolide resistance in MSM may also be an allodemic process, triggered by the previously described challenges, affecting a wide range of STIs such as gonorrhea, syphilis, enteric pathogens, and M. genitalium (33). Consequently, to avoid undesirable selection and spread of macrolide resistance in M. genitalium, azithromycin should be gradually replaced in the syndromic management of STIs, particularly in vulnerable populations such as MSM.

In contrast to repeat regions of the MgPa operon, which play a major role in antigenic variation, the mgpB nonrepeated region targeted in this typing procedure undergoes much less variation at both nucleotide and deduced amino-acid levels (14, 34). Figure S3 shows the alignment of the different putative amino acid sequence variants of the targeted mgpB region found among individuals. Remarkably, missense mutations concentrate in four amino acid positions: D96, S101, S107, and A117. The hypervariable nature of these residues was documented in a previous study analyzing clinical specimens collected in San Antonio, Texas (35). In addition, some amino acid variants, such as S101A, T104Q, T106Q, and S108G, seem to be associated since they concur simultaneously. Because the mgpB gene encodes the major adhesin P140 (MgpB), key for bacterial virulence and the human immune response (36, 37), the impact of the mutations detected in this study on structural or functional properties of this protein may require further investigation. Indeed, the mgpB nonrepeated region used for this typing method is predicted, using the I-TASSER server (38), to be surface exposed and antibody accessible and, therefore, it may play a role in immune system evasion (34). However, the underlying mechanisms and forces driving variation within this conserved region of P140 are currently unknown.

Some limitations to our study must be addressed. First, despite the appropriate discriminatory power of the mgpB-based SLSBT system, the use of additional genotypic characterizations may have been of interest for these sexual network analyses. Thus, the discriminatory power of the typing method could be enhanced with the use of a dual-locus-sequence-based typing (DLSBT) system combining the mgpB characterization and the analysis of a variable number of short-tandem-repeats (STRs) in the putative lipoprotein gene MG309 (15, 39). Nevertheless, this complementary procedure would have also affected the feasibility and performance of the typing method, since infections were presumably harboring low bacterial loads. Also, the reported conclusions are based on asymptomatic individuals. Since the majority of infections, particularly in the vagina and rectum, are silent and usually unnoticed, asymptomatic people are likely contributing most to the global spread of M. genitalium and antibiotic resistance. Consequently, results may be biased and might be different in symptomatic populations. Finally, infections reported as macrolide susceptible with the ResistancePlus MG kit were not confirmed with sequencing of the 23S rRNA gene, so false-negative results may have occurred (although they are likely negligible).

The research provides further evidence regarding the structure of transmission dynamics and sexual networks in M. genitalium infection. Furthermore, the investigation demonstrates the multiclonal selection and spread of antibiotic resistance for the bacterium, and assesses the specific allodemic situation of macrolide resistance in STIs among MSM. In conclusion, the findings deepen the natural history of M. genitalium and may have an impact on the management and control of the infection.