Antimicrobial Resistance Profiles of Human Commensal Neisseria Species

Pathogenic Neisseria gonorrhoeae causes the sexually transmitted infection gonorrhea. N. gonorrhoeae has evolved high levels of antimicrobial resistance (AR) leading to therapeutic failures even in dual-therapy treatment with azithromycin and ceftriaxone. AR mechanisms can be acquired by genetic transfer from closely related species, such as naturally competent commensal Neisseria species. At present, little is known about the antimicrobial resistance profiles of commensal Neisseria. Here, we characterized the phenotypic resistance profile of four commensal Neisseria species (N. lactamica, N. cinerea, N. mucosa, and N. elongata) against 10 commonly used antibiotics, and compared their profiles to 4 N. gonorrhoeae strains, using disk diffusion and minimal inhibitory concentration assays. Overall, we observed that 3 of the 4 commensals were more resistant to several antibiotics than pathogenic N. gonorrhoeae strains. Next, we compared publicly available protein sequences of known AR genes, including penicillin-binding-protein 2 (PBP2) from commensals and N. gonorrhoeae strains. We found mutations in PBP2 known to confer resistance in N. gonorrhoeae also present in commensal Neisseria sequences. Our results suggest that commensal Neisseria have unexplored antibiotic resistance gene pools that may be exchanged with pathogenic N. gonorrhoeae, possibly impairing drug development and clinical treatment.


Introduction
Neisseria gonorrhoeae, the etiologic agent of gonorrhea, is the second most commonly reported bacterial sexually transmitted infection in the US [1], and a worldwide public health concern. The World Health Organization (WHO) estimates there were 87 million new cases globally of N. gonorrhoeae in 2016 [2]; this incidence is increasing in many countries including the USA [3]. The Centers for Disease Control and Prevention (CDC) classifies N. gonorrhoeae as an "urgent threat" due to the emergence of antimicrobial resistance (AR) and multidrug resistance (MDR) [4][5][6][7][8][9]. The spread of AR has led to increasing rates of untreatable gonorrhea [10], including reports of untreatable and harder to treat pharyngeal N. gonorrhoeae, acquired through oral sex [11][12][13], where N. gonorrhoeae shares the environment with closely related commensal Neisseria.
Pathogenic and commensal Neisseria species exchange and transfer genes via natural competence and transformation [14]. Neisseria will exchange and transfer genes at high rates, as long as they share an identical or similar DNA uptake Sequences (DUS) and the corresponding DNA import complex [15][16][17][18]. Hence, in this study, we characterize the antimicrobial resistance profiles of commensal Neisseria species and explore their potential role as antibiotic resistance gene reservoirs for pathogenic Neisseria species. We selected and characterized commensal Neisseria species across a spectrum of genetic relatedness, including N. lactamica, N. elongata, N. cinerea, and N. mucosa. Of these, N. lactamica is the closest relative to the pathogens N. gonorrhoeae and N. meningitidis, while N. elongata appears to be the most distant relative to the pathogens [19][20][21]. N. cinerea

Commensal Neisseria Display Increased Resistance Levels to Several Antibiotics, Detected by Disk Diffusion Assays
In order to assess the levels of antibiotic resistance of commensal Neisseria, we performed disk diffusion assays (DDA) with 10 commonly used antibiotics, including 3 beta-lactams and 7 protein synthesis inhibitors as described in the Methods. Our DDA results revealed that some commensal Neisseria species were particularly resistant to azithromycin and ceftriaxone ( Figure 1 and Supplementary Table S1), the first-line treatment drugs against N. gonorrhoeae, as well as to erythromycin, the antibiotic applied on newborns' eyes to prevent conjunctivitis neonatorum. All commensal species displayed increased resistance to azithromycin as evidenced by the smaller zone of inhibition (ZoI) diameters (ZoI in the range 15.5-24.6 mm) than any of the N. gonorrhoeae strains (ZoI in the range 31.5-35.9 mm), including the highly resistant strains N. gonorrhoeae F89 and H041. For ceftriaxone, only N. lactamica displayed susceptibility levels similar to N. gonorrhoeae FA19 strain, with ZoI of 41.2 mm and 50.2 mm, respectively. Commensals N. cinerea and N. elongata were as resistant to ceftriaxone (ZoI 30.1 mm and 32.4 mm, respectively) as the resistant N. gonorrhoeae F89 and H041 (ZoI 31.5 mm and 30.8 mm, respectively), identified as models of therapeutic failure for ceftriaxone treatments. According to the CLSI guidelines, ZoI > 35 mm suggest that N. gonorrhoeae were susceptible to ceftriaxone [23]; hence, only N. lactamica, N. mucosa and N. gonorrhoeae FA19 can be considered susceptible to ceftriaxone; all other strains tested displayed resistance to it. The resistance of commensals to erythromycin resembled azithromycin's pattern where all commensals displayed increased erythromycin resistance levels (ZoI in the range 14.9-21.4 mm) compared to N. gonorrhoeae FA19 or the other N. gonorrhoeae strains tested (ZoI in the range 26.0-33.1 mm). Commensal Neisseria displayed a wide dispersion in erythromycin resistance levels.
Commensal Neisseria species displayed increased resistance to penicillin and ampicillin (ZoI in the range 29.7-34.0 mm and 28.3-31.6 mm, respectively) compared to the susceptible N. gonorrhoeae FA19 strain (ZoI 46.7 mm and 42.5 mm, respectively). According to the interpretive standards published by the CLSI [23], N. gonorrhoeae were considered resistant to penicillin when ZoI < 26 mm, such as N. gonorrhoeae MS11 and H041. Similarly, N. cinerea, N. mucosa, and N. elongata, but not N. lactamica, displayed higher resistance to chloramphenicol (ZoI in the range 25.6-27.3 mm) than the susceptible N. gonorrhoeae FA19 (ZoI of 42.2 mm). The other N. gonorrhoeae strains (MS11, F89, H041) displayed ZoI to chloramphenicol similar to the commensal species. N. cinerea, N. mucosa, and N. elongata, but not N. lactamica, were more resistant to tetracycline (ZoI in the range 24.9-25.6 mm) than N. gonorrhoeae FA19 (ZoI of 37.8 mm). Streptomycin was the only antimicrobial tested for which we did not observe a statistically significant (adjusted p-value > 0.0001, One-way ANOVA) difference between the ZoI diameters of commensals and N. gonorrhoeae FA19 (Figure 2, Tables 1 and 2). Only two commensals, N. cinerea and N. elongata displayed higher kanamycin resistance levels (ZoI of 21.5 and 23.1 mm, respectively) than N. gonorrhoeae FA19 (ZoI of 33.8 mm). It is important to note that the ZoI diameters of N. cinerea, N. mucosa, and N. elongata were found to be at least 33% smaller than those of N. gonorrhoeae FA19, for ampicillin, ceftriaxone, azithromycin, erythromycin, and chloramphenicol, and tetracycline, reflecting a wide gap in resistance levels between these 3 commensals and the susceptible N. gonorrhoeae FA19 strain (Supplementary Table S1).  Interestingly, both groups of commensal and pathogenic strains displayed similar levels of resistance to gentamicin (ZoI in the range 17.8-21.8 mm and 17.5-20.2 mm, respectively), a possible alternative to treat N. gonorrhoeae ( Figure 1 and Supplementary Table S1). These ZoI to gentamicin were slightly larger than the ZoI ≥ 16 mm categorized as the limit for susceptibility to gentamicin [24]. This suggests that the Neisseria species tested do not currently display antibiotic resistance against gentamicin, nor should they be considered a reservoir of gentamicin resistance. However, the small difference between the threshold of susceptibility (16 mm) and the range of ZoI observed (in the range 17.5-21.8 mm) suggests that a mutation, even if slightly decreasing the susceptibility level to gentamicin, could render these species resistant, and a gentamicin treatment ineffective.
For each antibiotic, we performed a One-way ANOVA statistical test, and observed that when comparing the commensals as a group to N. gonorrhoeae FA19, commensals were more resistant to 8 out of 10 antibiotics tested (except for streptomycin and gentamicin, adjusted p-value > 0.001). The difference was statistically significant (adjusted p-value < 0.0001) ( Figure 2, Tables 1 and 2). In addition, for antibiotics azithromycin and erythromycin, the difference in mean for commensals to pathogenic is a third or more than the mean ZoI diameter of the pathogenic strains, suggesting also that the commensals are more resistant than the resistant strains for azithromycin and erythromycin. It is important to note that as seen in Figure 1, there is a range of antimicrobial resistance level among commensals for azithromycin and erythromycin, with N. lactamica being more susceptible compared to the other commensals. Similarly, it is important to note that while N. gonorrhoeae MS11, F89, and H041 are labelled "resistant" strains, they are not resistant to all antibiotics.
Overall, N. cinerea, N. mucosa, and N. elongata appeared more resistant to antibiotics than N. lactamica, which is more closely related to N. gonorrhoeae and N. meningitidis. As shown in Figure 1, N. lactamica was the most susceptible of the four commensals to seven of the 10 antibiotics tested and the second-most susceptible in the three remaining antibiotics (see Figure 1). Therefore, N. lactamica was most similar in its antibiotic susceptibility profile to the susceptible pathogenic Neisseria gonorrhoeae FA19. This suggests that the other commensal strains-N. cinerea, N. mucosa, and N. elongata-may have distinct antimicrobial resistance mechanisms than N. gonorrhoeae FA19. Indeed, the profile of resistance for these 3 commensal species was very different from N. lactamica and the 4 N. gonorrhoeae strains, possibly suggesting that these commensals carry mutations and/or genes that confer increased antimicrobial resistance, in ways that have not been observed in N. gonorrhoeae. This requires further exploration of genomic data from commensal Neisseria species, both in known antimicrobial resistance genes and in regions that are not known to confer antimicrobial resistance.

Commensal Neisseria Display Increased Resistance Levels to Several Antibiotics, Detected by Minimal Inhibitory Concentrations
Next, we performed minimal inhibitory concentration (MIC) assays using a 2-fold serial dilution on GCB plates of azithromycin, ceftriaxone, penicillin, erythromycin, chloramphenicol, and gentamicin (see materials and methods for concentration ranges). We compared our results to susceptible and resistant strains of N. gonorrhoeae with values reported in the literature. Figure 3 shows differences in MIC to these 6 antibiotics for the commensal Neisseria species tested in the lab. Figure 4, Tables 3 and 4 report the results of the One-way ANOVA tests comparing the group of commensals to the reference. Similarly to our DDA analysis, commensal Neisseria species displayed a wide range of MIC levels. N. lactamica appeared more susceptible compared to the other commensal Neisseria species ( Figure 3 and Table S2). Indeed, N. cinerea, N. mucosa, and N. elongata showed increased ceftriaxone MIC values (0.128 µg/mL, 0.064 µg/mL, 0.064 µg/mL, respectively), reflecting increased resistance, compared to N. lactamica (0.008 µg/mL) or a susceptible N. gonorrhoeae strain (≤0.015 µg/mL). N. cinerea displayed ceftriaxone MIC values 16 times higher than N. lactamica; N. mucosa and N. elongata displayed ceftriaxone MIC values 8 times higher than N. lactamica. However, these values of MIC for commensal Neisseria were still below the MIC breakpoint for ceftriaxone described by Kirkcaldy et al. [25]. The MIC values for azithromycin in commensal Neisseria were lower (0.25-0.5 µg/mL) than the MIC breakpoint for resistant N. gonorrhoeae (≥2 µg/mL) described by Kirkcaldy et al. [25]; hence, commensals appeared sensitive to azithromycin, detected by MIC assays, still more resistant than N. gonorrhoeae FA19 [26]. All commensals were overall more resistant to penicillin than the susceptible strain of N. gonorrhoeae. N. elongata and N. mucosa appeared more resistant than N. cinerea and N. lactamica towards chloramphenicol and erythromycin; we did not find MIC values to these antibiotics for N. gonorrhoeae susceptible and resistant strains, we only found the erythromycin MIC value for N. gonorrhoeae FA19 [26]. These observations reinforce the need to analyze genomic sequences of commensal Neisseria to identify possible antimicrobial resistance genes and mutations.    We performed a One-way ANOVA test (alpha level 1 × 10 −4 ) on groups, for MIC data. Using MIC, we show that commensals tend to be more resistant to 4 antibiotics (except chloramphenicol and gentamicin) of the 6 tested compared to a reference N. gonorrhoeae susceptible strain. However, while erythromycin MIC showed statistically significant differences, the difference was not statistically significant for azithromycin, penicillin and ceftriaxone. One explanation is the great phenotypic variability within the group of commensals. The MIC results show that all commensals were more resistant to erythromycin than N. gonorrhoeae FA19 [26] (p-value < 1 × 10 −5 ). For other antibiotics, MIC results do not display statistically significant differences between commensals and the reference, but show a trend where the commensals are more resistant than the reference (Figure 4, Tables 3 and 4), in agreement with DDA results. In regard to chloramphenicol, we could only do a within-commensal group comparison. We observed that N. mucosa is more resistant to chloramphenicol than the other commensals (p-value < 1 × 10 −5 ).

Mutations in Penicillin-Binding Protein 2 (PBP2) and other Protein Sequences Can Partially Explain Resistance in Commensal Species
In order to explore the potential mechanistic basis for variation in resistance among Neisseria species, we compared protein sequence variation amongst commensal and pathogenic strains, focusing on previously identified genes associated with antibiotic resistance. One of these genes, penA, encodes the penicillin-binding-protein-2 (PBP2), known to modify penicillin and other beta-lactam drugs [27]. In N. gonorrhoeae, known mutations in penA lead to increased resistance [28]. N. gonorrhoeae resistant strains often contain nucleotide point mutations leading to amino acid changes, and/or contain a mosaic sequence of penA, which is a recombination of penA genes from N. perflava and N. cinerea [29][30][31][32], both commensal species regularly carried in the human oro-and nasopharynx, where other commensal Neisseria also reside, and where N. gonorrhoeae may be found in cases of pharyngeal gonococcal infections.
The penA sequences of commensal Neisseria species reveal genetic polymorphism. We aligned amino acid sequences of PBP2 from the different commensal Neisseria, N. gonorrhoeae susceptible strain LM306 (wild-type sequence, Ngo_WTSu) and resistant strain NG-3 (mosaic sequence, Ngo_mosaic), and 2 outgroup sequences from Eikenella corrodens (Eik) and Kingella oralis (Kor) (from the family Neisseriaceae), using BLASTp [33] and Clustal Omega [34] algorithms. We observed that mutations in PBP2 known to increase antibiotic resistance in N. gonorrhoeae, were present in N. mucosa (Nmu), N. elongata (Nel), and N. cinerea (Nci) ( Figure 5). On the other hand, N. lactamica (Nla) conserved the amino acids present in the susceptible N. gonorrhoeae strain LM306 (highlighted in yellow). Other described mutations were not present in the commensal Neisseria species (hence, are not shown in Figure 5), namely the insertion of an aspartate after position 345 of the wild-type susceptible sequence [30,35], and the mutations A501V/P [36,37], and G545S [38]. The differences observed among PBP2 sequences of commensal Neisseria species may partially explain the higher resistance levels observed in these species for beta-lactam antibiotics, penicillin, ampicillin, and ceftriaxone. Other genes are likely involved for betalactam and other antibiotic resistance. However, the causal relationship between genes, mutations, and antimicrobial resistance are less clear than for penA. Supplementary Table S3 summarizes mutations observed in AR genes of commensal Neisseria, Eikenella corrodens and Kingella oralis. We used rplD and rplV as examples for macrolide resistance, mtrR, ponA, and porB for cephalosporin resistance, rpsJ and plasmid-carried tetM for tetracycline resistance, plasmid-carried chloramphenicol acetyltransferase (cat) for chloramphenicol resistance, and plasmid-carried str for streptomycin resistance. We were not able to identify genes or specific mutations described for gentamicin resistance. We observed that the majority (11/11) of the commensal sequences analyzed through BLASTp and the MSA viewer (NCBI) contained the mutations L6Q, D94N, R99Q, V17A, V45Q, T72V, S101C in rplV. Mutations K74S, D91N, V120A, V121I, K123A/S/D/E, T173Q/H, A190K/R in rplD were observed in 26/26 commensal sequences, associated with macrolide resistance [39]. Kanamycin and streptomycin resistance genes are carried by transposons that were not observed in the commensal sequences studied. The genes tetM and cat are known to be carried by plasmids, which have not been described for the commensal Neisseria strains studied. Overall, these observations are not enough to explain the vast intrinsic antimicrobial resistance by commensal Neisseria species described in this study. Hence, further analyses of known antimicrobial resistance genes and genomic data followed by mutagenesis and phenotypic analysis are necessary to understand and explain antibiotic resistance for non-beta-lactam antibiotics, such as azithromycin, chloramphenicol, and erythromycin.

Discussion
In this study, we characterized the antimicrobial resistance profiles of 4 commensal Neisseria species, closely related to pathogenic Neisseria, and clinically relevant given their natural niche, the human oral and nasal pharynx (ONP) [40]. We compared the resistance levels of N. lactamica, N. cinerea, N. mucosa, and N. elongata to 4 N. gonorrhoeae strains (FA19, MS11, F89, H041). We observed that N. cinerea, N. mucosa, and N. elongata generally displayed higher resistance levels than N. gonorrhoeae FA19 or N. lactamica. Given the high antimicrobial resistance (AR) profiles observed for the commensals, it is possible that commensals express AR genes, mutations, and/or mechanisms not yet identified in N. gonorrhoeae. Hence, commensals are likely antimicrobial resistance gene reservoirs for N. gonorrhoeae, particularly for azithromycin and erythromycin.
Among the ten antibiotics tested, ceftriaxone and azithromycin are particularly relevant as they are the current line of treatment against N. gonorrhoeae. Indeed, N. gonorrhoeae is treated with a dual therapy of oral azithromycin (1 g, single dose) and injectable intramuscular ceftriaxone (0.25 g, single dose) (or oral cefixime) [41], to which certain gonococcal strains display resistance, such as N. gonorrhoeae F89 and H041 from France and Japan, respectively [10,36]. While antibiotic resistance to this dual treatment is increasing, the rates of resistance are still below the 5% threshold needed to recommend new guidelines by the CDC. However, if this treatment fails, clinicians require a time-and resource-consuming antibiogram to assess the antibiotic susceptibility panel of the infectious strain. Gentamicin was the only antibiotic for which all strains displayed a similar level of susceptibility and were categorized as susceptible according to EUCAST and Bala et al. [24]. Gentamicin could be a possible drug of choice to treat extreme drug resistant N. gonorrhoeae, as all strains tested in our study show ZoI diameters larger than 16 mm, breakpoint for susceptibility to gentamicin [24]. Additionally, new data associated with gentamicin shows that N. gonorrhoeae WHO reference strains present an MIC of 4 µg/mL (dispersion 2-8 µg/mL) [24], which is also consistent with our MIC results for commensal Neisseria. However, gentamicin is an aminoglycoside used to treat Gram-negative infections causing bone, urinary tract and respiratory infections, endocarditis, meningitis, and pelvic inflammatory disease. Often, these infections have limited treatment alternatives. Knowing how easily N. gonorrhoeae acquires resistance whether through mutation or natural competence, it is absolutely indispensable that scientists and clinicians reflect on and model the outcome of prescribing gentamicin as a future drug of choice for N. gonorrhoeae. In addition, a recent report demonstrate that gentamicin as monotherapy is not a good alternative to treat pharyngeal gonorrhea [42].
In order to explore a potential genetic mechanism for variation in antibiotic resistance profiles, we analyzed protein sequences of PBP2, encoded by the penA gene, and other AR genes, in the four commensal strains along with the focal N. gonorrhoeae strain. We observed that mutations known to increase antibiotic resistance to beta-lactams were found in N. cinerea, N. mucosa, and N. elongata. However, these penA mutations were not found in N. lactamica which conserves the amino acids present in susceptible strains of N. gonorrhoeae strain LM306. These differences may partially explain the reduced susceptibility to betalactams observed in the commensals studied. In addition, the analysis of other possible AR gene mutations were also observed in commensal Neisseria species. This suggests that genomic analyses of commensal Neisseria species for antimicrobial resistance mechanisms could reveal known and new candidate genes, and mutations involved in antimicrobial resistance, in particular to azithromycin and ceftriaxone, first-line treatment antibiotics to treat N. gonorrhoeae infections. Further work is needed to explore the potential influence of plasmid-mediated penicillin resistance in commensal Neisseria species, which is highly prevalent in N. gonorrhoeae strains.
As with PBP2, several other genes are involved in antimicrobial resistance, such as mtrR, penA (PBP2), penC, ponA (PBP1), tetM, pilTQ, folP, mtrC, 23S rRNA, rpsJ, 16S rRNA, gyrB, gyrA, parC, prld, porB (PIB). Genetic analysis of these sequences for commensal Neisseria both at the nucleic acid and protein levels could help inform the antimicrobial resistance profile of the pathogen. Similarly, the analysis of gene expression of these genes in commensals under sub-MIC conditions could offer additional insight in the evolution of AR mechanisms and their regulation. Antibiotic resistance in commensals has been observed in several genera, often through transformation with plasmids carrying antibiotic resistance genes [43,44]. Antibiotic resistance is widely spread in N. gonorrhoeae, and researchers have focused their studies on genomic analyses of AR genes in pathogenic Neisseria [11]. However, given the potential for transfer of AR genes between Neisseria species [45][46][47][48][49], our results and those of Fiore et al. [39] emphasize the need to further investigate AR gene pools in commensal Neisseria species. Fiore et al. [39] initiated this work by performing a genomic and phenotypic study of the CDC AR panel of Neisseria species, using 6 antibiotics (penicillin, cefixime, ceftriaxone, tetracycline, azithromycin, and ciprofloxacin). Our study further complements that work by analyzing the antibiotic resistance profile of 8 Neisseria strains to 10 antibiotics. Together, these studies provide a gateway to understanding commensal bacteria mechanisms of resistance and help identify putative genes involved in the expression and regulation of these mechanisms.
Moving forward, given observations that antibiotic resistance levels were higher in N. cinerea, N. mucosa, and N. elongata, it would be critical to further compare genomic sequences of these species, to identify candidate genes involved in resistance mechanisms not yet observed in N. gonorrhoeae, and to continue the analysis of phenotypic antibiotic resistance level characterization of other commensal Neisseria species.

Disk Diffusion Assays
Neisseria strains plated on GCB agar plates, overnight, at 37 • C in 5% CO 2 atmosphere, were harvested with a sterile loop and suspended in supplemented GCB broth at OD 600nm 0.2 UA. Bacteria were spread on the plate using CLSI guidelines [52]. Briefly, a sterile cotton swab was dipped in the suspension and spread in one direction on the plate. The procedure (dip and spread) was repeated two additional times, every time rotating the plate by 120 degrees, to obtain a homogeneous lawn of bacterial growth throughout the plate. Plates were then allowed to dry for 10 min, and antibiotic disks were applied. To prevent overlay of antibiotics or zones of inhibitions (ZoI), we applied 3 antibiotics per plate. Zones of inhibition were measured using AntibiogramJ [53] and ImageJ [54]. Averages and standard deviation (SD) values were obtained from 3 independent experiments, each containing 3 biological replicates. All strains (4 commensal strains and 4 N. gonorrhoeae strains (FA19, MS11, F89 and H041), were tested against a panel of 10 antibiotics (Penicillin 10 UI, Kanamycin 30 µg, Streptomycin 10 µg, Azithromycin 15 µg, Ceftriaxone 30 µg, Erythromycin 15 µg, Tetracycline 30 µg, Ampicillin 10 µg, Chloramphenicol 30 µg, Gentamicin 10 µg), on pre-loaded disks (6 mm diameter) purchased from Hardy Diagnostics (Santa Maria, CA, USA), stored at −20 • C when not in use.

Minimal Inhibitory Concentrations
The minimal inhibitory concentration (MIC) was determined as the concentration of antimicrobial inhibiting 99.99% (or a 4-log 10 decrease) of bacterial growth. We used the plate dilution technique to quantify inhibition, following CLSI guidelines [55]. Briefly, three biological replicates of each bacterial species grown overnight on supplemented GCB plates at 37 • C, in 5% CO 2 , were harvested with sterile plastic loops, and suspended in supplemented GCB broth, at OD 600nm 0.2 UA. Five µL of each suspension were plated in triplicates (technical replicates), on supplemented GCB agar plates containing a range of antibiotics, in 2-fold serial dilutions. Plates contained 0.006 µg/mL to 2 µg/mL of Penicillin G, or 4 ng/mL to 512 ng/mL of Ceftriaxone, or 0.031 µg/mL to 4 µg/mL of Azithromycin, or 2 µg/mL to 256 µg/mL of Erythromycin, or 0.125 µg/mL to 16 µg/mL of Chloramphenicol, or 1 µg/mL to 64 µg/mL of Gentamicin. Plates were stored at 4 • C and used at most 5 days after plating. Three independent experiments were performed. The concentrations of antibiotics on plates were considered the minimal inhibitory concentrations where less than 4 colonies were observed per spot, as it suggests 99.99% growth inhibition.

Penicillin-Binding Protein 2 (PBP2) Sequence Alignment
Sequences from penicillin-binding protein 2 (PBP2) of several Neisseria and non-Neisseria Neisseriaceae sequences were obtained from NCBI. Amino acid sequences were aligned using ClustalOmega and BLASTp platforms, using the default parameters.

Data Analysis
Data were analyzed using R and RStudio [56], and plots were made using ggplot2 [57]. All experiments were performed at least 3 times independently. Each experiment contained 3 biological replicates of each species. Means and standard error of the means (SEM) were reported where applicable on the charts and tables. One-way ANOVA tests were used for DDA and MIC tests as described in Section 2. We used a conservative alpha threshold of 1 × 10 −4 given the number of pairwise comparisons performed per antibiotic.

Conclusions
Here, we explored the antimicrobial resistance profiles of commensal Neisseria species. We demonstrated that several commensal Neisseria species express high levels of resistance to antimicrobial compounds, in particular azithromycin and ceftriaxone, which form the current dual-therapy treatment against N. gonorrhoeae. We propose that commensal Neisseria species found in the human oral and naso-pharynx can be reservoirs of antimicrobial resistance genes for pharyngeal N. gonorrhoeae, which could increase the spread of antimicrobial resistance among already hard-to-treat N. gonorrhoeae. In particular, we observed that N. cinerea and N. elongata were particularly resistant to several antimicrobials used to treat N. gonorrhoeae, to levels higher than known N. gonorrhoeae resistant strains such as MS11, F89, and H041. Further genomics and transcriptomics analysis are needed to pinpoint the genetic and gene regulation bases for these antimicrobial resistance profiles.