Zap the clap with DNA: a novel microbicide for preventing Neisseria gonorrhoeae infection

ABSTRACT

Each year, Neisseria gonorrhoeae (Ngo) causes over 1.5 million new infections in the United States, and >87 million worldwide. The absence of a vaccine for preventing gonorrhea, the rapid emergence of multidrug-resistant and extremely drug-resistant Ngo strains, and the limited number of antibiotics available for treating gonorrhea underscore the importance of developing new modalities for addressing Ngo infection. Here, we describe DNA-based microbicides that kill Ngo but not commensals. Previously, we showed that Ngo is killed when it takes up differentially methylated DNA with homology to its genome. We exploited this Achilles heel to develop a new class of microbicides for preventing Ngo infection. These microbicides consist of DNA molecules with specific sequences and a methylation pattern different from Ngo DNA. These DNAs kill low-passage and antibiotic-resistant clinical isolates with high efficiency but leave commensals unharmed. Equally important, the DNAs are equally effective against Ngo whether they are in buffered media or personal lubricants. These findings illustrate the potential of this new class of practical, low-cost, self-administered DNA-based microbicides for preventing Ngo transmission during sexual intercourse.

INTRODUCTION

Sexually transmitted infections (STIs) are a serious health concern worldwide. The World Health Organization (WHO) estimates that 1 million new STIs are acquired every day around the globe, and 374 million people are infected with at least one major STI pathogen (N. gonorrhoeae, C. trachomatis, T. pallidum and T. vaginalis) (1).

Gonorrhea is the second most reported bacterial STI. Each year, Neisseria gonorrhoeae (Ngo) causes over 1.5 million new infections in the United States, and >87 million worldwide (2). The most common sites of infections are the genitals, rectum, and throat (3). In men, untreated gonococcal infection can cause epididymitis; in women, infection can result in pelvic inflammatory disease, infertility, abortion, and ectopic pregnancy as sequelae (2). In a large proportion of men and women (10%–20%), gonorrhea often escapes diagnosis because symptoms are mild or lacking (4–7). In both genders, Ngo can cause disseminated infection by entering the bloodstream (8, 9), and on rare occasions, Ngo can cause endocarditis and meningitis (10–12).

A significant challenge in treating gonorrhea is the ability of Ngo to develop antibiotic resistance (13). Sulfonamides were the first antibiotics prescribed to treat Ngo infections. Four years after its introduction, the pathogen developed resistance to these compounds (14). Over time, Ngo has acquired and maintained resistance to all antibiotics used for its treatment (13). Currently, cephalosporins are the last line of defense against this pathogen (15). The emergence of multidrug-resistant and extremely drug-resistant Ngo strains (16–19), coupled with the limited number of antibiotics available for treating Ngo infections, led the World Health Organization to include this pathogen in its Priority 2 list for research and development of new antibiotics (20).

Currently, three antibiotics are in phase 3 clinical trials: zoliflodacin, solithromycin, and gepotidacin (21–25). Among these, zoliflodacin shows the most promise for treating uncomplicated gonorrhea (26). Zoliflodacin functions by inhibiting bacterial type II topoisomerases by binding to gyrase at sites different from fluoroquinolones (27). Point mutations in DNA GyrA (S91F and D95G/D95A) and ParC (E91G) are directly correlated with high resistance levels to fluoroquinolones (28, 29); however, strains carrying these mutations are sensitive to zoliflodacin (27). Due to the propensity of Ngo to rapidly develop resistance to antibiotics, there is concern that once this antibiotic is in use, resistance will inevitably emerge. Indeed, a multidrug-resistant Ngo isolates with resistance to zoliflodacin (GyrB D428V) was identified in Japan in 2000 (30). These observations emphasize the need to develop new modalities for treating and preventing Ngo infections.

In the last 50 years, multiple efforts have been made to develop a vaccine against Ngo (31, 32). These efforts, which originally focused on antigenically variable surface proteins, were unsuccessful (33–37). Only two vaccine candidates, a heat-killed whole-cell lysate and a purified pilin have reached human clinical trials; however, these vaccines failed to protect against gonococcal urethritis (38–40). More recently, the incidence of gonorrhea declined in persons receiving a vaccine against N. meningitidis serogroup B; subsequently, this vaccine was shown to elicit cross-protective antibodies against Ngo (41–44). This observation, together with the rise of multidrug-resistant Ngo strains, refueled the interest in developing Ngo vaccines (45). However, until there is a vaccine, the search for new strategies for preventing gonorrhea infections continues.

We have previously demonstrated that commensal Neisseria elongata (Nel) kills Ngo in vitro, and the presence of Nel accelerates Ngo clearance in a mouse model of lower genital tract infection (46). Neisseria are naturally competent organisms (47). Ngo is killed when it takes up commensal chromosomal DNA, and when this DNA attempts to undergo homologous recombination with the pathogen chromosome. The methylation signatures of Nel and Ngo DNA are different: while adenines are methylated in Nel DNA, cytosines are methylated in Ngo (46). The evidence strongly suggests that the restriction methylation systems in Ngo recognize Nel DNA in the heteroduplexes as foreign and cause nicks in these structures (46, 48, 49). DNA repair enzymes cannot keep pace with the large number of nicks along the chromosome, leading to cell death. Importantly, Nel DNA does not kill the commensal Neisseria species tested. These findings led us to hypothesize that DNA can be developed as a novel class of microbicides that act specifically on Ngo (50). Here we describe the engineering of DNA molecules containing Ngo sequences that kill Ngo but not commensal Neisseria species or Lactobacillus iners or Gardnerella vaginalis, two commensals often residing in the female genital tract. Finally, we describe the mechanisms by which a small proportion of Ngo cells resist killing by DNA molecules and discuss the potential of these resistant cells to cause infection.

RESULTS

Nel DNA kills a variety of Ngo isolates

As a first step toward testing the hypothesis that differentially methylated DNA can kill Ngo, we determined whether commensal Neisseria elongata (Nel) DNA can kill a diverse group of Ngo isolates (Table S1) from an STD clinic in Durham, North Carolina, one of which, UNC020, has been shown to lose viability when cocultured with Nel (46). And an additional group of strains from the Antibiotic Resistance Isolate Bank (CDC-FDA Antibiotic Resistant Isolate Bank. Atlanta (GA): CDC) (51). The killing assay was performed as described by Kim et al. (46), using 10 µg/mL of purified Nel DNA (see Materials and Methods). Nel DNA killed four of the five tested isolates as efficiently as it killed laboratory strain MS11, the positive control (90%–95% cells killed). It did not kill low-passage isolate UNC003 (Fig. 1).

Fig 1.

Low-passage and antibiotic-resistant Neisseria gonorrhoeae isolates are killed by N. elongata DNA. A killing assay was performed to test the ability of Nel DNA to kill low-passage Ngo clinical isolates from the STD clinic, Durham, North Carolina, and antibiotic-resistant isolates from the CDC AR bank. For each experiment, 5 × 10⁵ CFUs were incubated with 10 µg/mL of DNA in 250 µL of GC broth containing Kellogg’s supplements. The bacteria were incubated for 5 h at 37°C/5% CO₂. Bacteria were harvested, serially diluted, and plated to enumerate the survivors. Bars represent the average of at least three biological replicates. Statistical analysis was done using one-way ANOVA with Tukey’s post-test using GraphPad Prism 9 software. ns, not significant compared to Nel DNA. Abbreviations: *P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001; ns, not significant; AR, antibiotic resistant.

Microbicide design

We sought to design DNA molecules that will specifically target Ngo and not commensals. To this end, we constructed DNA molecules whose sequences were thought to be present in Ngo but rare in, or absent from, human-adapted commensal Neisseria (52). Sequences in this group include tdfF (NGFG_RS00105), tdfH (NGFG_RS05315), and iga (NGFG_RS01485) (Data S1 and S2). Our in silico analysis showed that tdfF was found in some commensals; tdfH was present in all closed Neisseria genomes except Nel, but not in draft Neisseria genomes (Table S2); iga was not found in any human commensal species. We also constructed a second set of DNA molecules whose sequences have homology to Ngo genes essential for DNA replication and maintenance, reasoning that these DNAs, when attempting to recombine with the essential Ngo genes, will lead to cell death. DNA sequences in this category are cytosine DNA methyltransferase genes known to be active (48, 53), namely, ngoAXIV (NGFG_RS02014), ngoAI (NGFG_RS11080), and ngoAXV (NGFG_RS03570) (Data S1). Finally, we constructed a third group of DNAs to test the idea that any sequence encoding an essential Neisseria gene can act as a microbicide. These sequences, from genes involved in DNA replication and maintenance, correspond to parC (NGFG_02066), gyrA (NGFG_01171), and gyrB (NGFG_01684) (Data S1).

Promoter-less sequences within the open reading frame of each locus, designed to omit start codon and termination signals, and containing DNA Uptake Sequences (DUS) preferred by Ngo, DUS-12 (51) flanking both ends, were PCR-amplified using primers listed in Table S3. The amplicons were cloned into pCR-Blunt and transformed into E. coli DH5 alpha. The resulting plasmids were used for killing assays. The names of the constructs begin with the letters MB, for MicroBicide, followed by a letter or number that identifies the cloned sequence; for instance, MB-I indicates the DNA molecule contains iga sequences (Table S4). Individual DNA molecules were also concatenated: constructs MB-FHI (tdfF, tdfH, and iga); MB-14–1 (ngoAXVI and ngoAI), and MB-14-1-15 (ngoAXIV, ngoAI, and ngoAXV) (see Materials and Methods).

The DNA molecules were tested for their ability to kill Ngo laboratory strain MS11. The killing efficiency of each DNA was calculated as the number of viable bacteria in the DNA-treated sample divided by the number of viable bacteria in the buffer-treated sample (10 mM Tris pH 8), expressed as a percentage. Nel DNA served as the positive control and Tris buffer alone as the negative control. The killing efficiency of the DNAs varied (Fig. 2). MB-F (tdfF), MB-H (tdfH), and MB-I (iga) killed MS11 at moderate efficiency (22%–38%), while the concatenate MB-FHI killed at a higher efficiency (56%). Notably, MB-1, MB-14, MB-15, and the concatenates MB-14–1 and MB-14-1-15 were as effective in killing MS11 as Nel DNA (>90% efficiency). High killing efficiency was also observed for MB-A (gyrA), MB-B (gyrB), and MB-C (parC). The two concatenates together, MB-FHI and MB-14-1-15, killed MS11 less efficiently than when they were tested alone. We focused on MB-14-1-15 for the subsequent studies.

Fig 2.

Killing of Neisseria gonorrhoeae (Ngo) strain MS11 by DNA molecules. (A) Description of the constructs and loci selected for testing. (B) Killing efficiency of DNA constructs in (A). For each experiment, 5 × 10⁵ CFUs were incubated with 10 µg/mL of DNA in 250 µL of GC broth containing Kellogg’s supplements. The bacteria were incubated for 5 h at 37°C/5% CO₂. Bacteria were harvested and serially diluted to enumerate the survivors. Bars represent the average of at least three biological replicates. Statistical analysis was done using one-way ANOVA with Tukey’s post-test using GraphPad Prism 9 software. Abbreviations; ****P ≤ 0.0001; ns, not significant compared to Nel DNA.

Testing the effect of MB-14-1-15 on commensal bacteria

The effect of MB-14-1-15 was tested on human-adapted commensal Neisseria spp. and two common bacterial inhabitants of the female lower genital tract, Lactobacillus crispatus and Gardnerella vaginalis (Table 1). MB-14-1-15 did not kill any of the commensal Neisseria isolates. By contrast, Nel DNA exhibited a modest killing effect on N. lactamica and N. subflava. The reason for this is not clear. One plausible explanation is that the methylation patterns of N. lactamica and N. subflava DNA are more similar to that of Ngo than to Nel DNA. N. lactamica encodes between 14 and 17 restriction-modification systems (Table S5); at least 6 of those are predicted to methylate cytosines (http://rebase.neb.com/rebase/rebase.html). This may explain why Nel DNA has a moderate to high killing effect on N. lactamica. N. subflava encodes between 3 and 9 putative RM systems; as their nucleotide target(s) are unknown, we cannot explain the moderate to low killing efficiency of Nel DNA on this organism. We can only speculate that some of the RM systems are present in N. subflava methylated cytosines. The extent and type of methylation in commensal Neisseria requires further investigation.

TABLE 1.

MB-14-1-15 DNA does not harm commensal bacteria^b

Target	Nel DNA % killing (SD, P-value)	MB 14-1-15 % killing (SD, P-value)
Neisseria elongata ATCC 9315	16.67 (± 10.02, ns)	11.27 (± 11.49, ns)
Neisseria subflava NJ9703	34.83 (± 8.37, 0.03)	0.06 (± 1.4, ns)
Neisseria polysaccharea ATCC 43768	6.9 (± 7.9, ns)	−0.5a (± 7.4, ns)
Neisseria lactamica ATCC 23970	70.81 (± 20.49, 0.04)	−1.05a (1.86, ns)
Lactobacillus crispatus JV-V01	5.93 (± 12.95, ns)	8.6 (± 3.68, ns)
Gardnerella vaginalis 8151B	6.73 (± 6.9, ns)	8.45 (± 8, ns)

^a A negative value indicates bacteria in the test sample grew more than bacteria in the negative control sample.

^b SD, standard deviation; ns, not significant. P-value, Tris vs DNA treatment.

As expected, MB-14-1-15 DNA did not kill Lactobacillus or Gardnerella. These bacteria do not appear to have a Neisseria like DNA uptake system, to our knowledge, and their DNA recombination systems are little understood. They would therefore not take up MB-14-1-15.

Personal lubricants as a vehicle for delivering MB-14-1-15

Our findings so far suggest that DNA can potentially be used as an antimicrobial agent specifically to prevent Ngo infection. One method for introducing the DNA to the potential site of infection, the genital tract, is via a commonly used gel. To explore this idea, we suspended MB-14-1-15 in two commercially available personal lubricants, the water-based KY Jelly and the silicone-based Astroglide, and tested these formulations for their ability to kill Ngo laboratory strain MS11. The negative and positive controls are lubricants containing Tris buffer and Nel DNA, respectively (Fig. 3A). MB-14-1-15 in both water- and silicone-based personal lubricants killed MS11 as efficiently as Nel DNA incorporated into the same lubricants. KY jelly and Astroglide alone killed a small percentage of MS11; this may be due to the sensitivity of MS11 to the antiseptics in their formulation. These findings suggest that personal lubricants have high potential as delivery vehicles for the DNA molecules.

Fig 3.

MB-14-1-15 efficiently kills all tested Neisseria gonorrhoeae isolates, whether dissolved in buffer (10 mM Tris, pH 8) or personal lubricants. (A) Laboratory-adapted isolates. (B) Low-passage clinical isolates from the STD clinic, Durham North Carolina. (C) Antibiotic-resistant isolates from the CDC AR bank. MB-14-1-15 was incorporated into commercially available personal lubricants KY Jelly and Astroglide. For each experiment, 5 × 10⁵ CFUs were incubated with 1 µg/mL of DNA in 250 µL of GC broth containing Kellogg’s supplements. Prior to the experiment, 10 µL of personal lubricants or 10 mM Tris, pH 8 was mixed with the DNA. The bacteria were incubated for 5 h at 37°C/5% CO₂. Bacteria were harvested and serially diluted to enumerate the survivors. Bars represent the average of at least three biological replicates. Statistical analysis was done using one-way ANOVA with Tukey’s post-test using GraphPad Prism 9 software. Abbreviations: MB, MB 14-1-5; Nel, Nel DNA, KY, KY Jelly, AG, Astroglide. *P ≤ 0.01; **P ≤ 0.1; ***P ≤ 0.001; ****P ≤ 0.0001; ns, not significant.

Testing MB-14-1-15 against low-passage and drug resistant Ngo isolates

MB-14-1-15 incorporated into KY Jelly, Astroglide, or Tris buffer were compared for their ability to kill two additional laboratory-adapted Ngo isolates commonly used in the field (FA1090 and FA19), four low-passage clinical isolates (UNC012, UNC014, UNC016, and UNC020) from the Durham County, NC, STD clinic, and four antibiotic-resistant isolates (AR173, AR174, AR200, and AR202) from the CDC and FDA Antibiotic Resistance Isolate Bank (Fig. 3; Table S1). The positive control was Nel DNA in KY Jelly, Astroglide, or Tris buffer, and the negative controls were KY Jelly, Astroglide, and Tris buffer without DNA.

MB-14-1-15 DNA suspended in KY Jelly and Astroglide killed all tested isolates as well as, or better than, Nel DNA in the same medium, with an overall efficiency above 95%. KY Jelly alone did not kill UNC014 and had only a mildly negative effect on laboratory-adapted FA1090 and FA19 and low-passage isolates UNC012, UNC016, and UNC020 (killing efficiencies ranging from 25% to 35%). It had a greater negative effect on AR174 (~80% killing efficiency). Astroglide alone had a similarly mild negative effect on UNC014, UNC020, and AR202; however, it had a much lower negative effect on the other strains (killing efficiencies ranging from 0% to 15%); it did not kill AR174. Thus, the ability of MB-14-1-15 to kill Ngo is retained in the presence of commercially available lubricants.

Investigating the mechanism(s) of Ngo resistance to killing by Nel DNA

In our assays, a small percentage of Ngo cells (<7%) from all except one Ngo strain survive DNA killing. We investigated the mechanism(s) for this resistance. To kill Ngo, the DNA must be taken up by the Type IV Pilus (Tfp). Ngo can switch Tfp expression ON and OFF at high frequency, through a mechanism known as phase variation (33, 54, 55). To determine whether resistant cells expressed Tfp, we measured the number of non-piliated CFUs recovered after the killing assays by plating survivor cells and examining their colony phenotype after overnight growth. The vast majority of UNC020 and AR173 survivors produced colonies with a non-piliated phenotype (10.94% in Tris control vs 99.5% in MB-14-1-15; 1.16% in Tris control vs 76.9% in MB-14-1-15; P ≤ 0.0001 and P ≤ 0.002 respectively), suggesting that resistance of UNC020 and AR173 cells to DNA killing was likely due to the absence of Tfp. By contrast, the number of MS11 survivor cells producing non-piliated colonies was only slightly higher in the DNA-treated sample than in the Tris control (1.9% vs 0.3%, respectively; P = 0.18), suggesting these resistant cells still expressed Tfp. More work is needed to determine the mechanism behind the resistance of MS11 cells to DNA killing.

One low-passage Ngo (UNC003) was poorly killed by Nel DNA (5.5% killing efficiency) and MB-14-1-15 (16.2% killing efficiency) (Fig. 1). On GCB agar, UNC003 cells produced small, rounded shapes typical of piliated colonies (Fig. 4A). UNC003 cells are motile and crawled together to form tight microcolonies, a typical behavior of motile cells (Fig. 4B; Video S1). However, UNC003 cells could not be transformed to Rif resistance in the presence of MS11 Rif^R DNA (Fig. 4C). These results suggest that although UNC003 was able to retract its Tfp, it was unable to take up DNA.

Fig 4.

Mechanism of resistance in UNC003 to DNA-mediated killing. (A) UNC003 produces piliated (P+) and non-piliated (P-) colonies on GCB agar for 16 h. (B) Microcolonies formed by UNC003 after 4 h of growth in GC broth containing Kellogg’s supplements. (C) UNC003 WT does not take up DNA unless complemented with ComP from Ngo MS11. DNA transformation assay was performed as described (see methods) using 10 µg of MS11 Rif^R DNA and a 4 h recovery. (D) Killing assay using isolate UNC003 wt and UNC003+comP_MS11. Results represent the average of at least three biological replicates. Statistical analysis was done using GraphPad Prism 9 software; one-way ANOVA with Tukey’s post-test for the killing assay and a Student’s t-test for the DNA transformation assay. Abbreviations: wt, wild type; ::erm, UNC003::erm; ::comP_MS11, UNC003::comP_MS11; MB, MB-14-1-5; Nel, Nel DNA; LOD, limit of detection. **P ≤ 0.05; ***P ≤ 0.001; ns, not significant.

Transformation in Neisseria requires the binding of DNA by the pilus-associated ComP protein (56); retraction of the Tfp then brings the DNA-ComP complex in close proximity to the cell for further processing for DNA entry (57). The absence of ComP does not have an impact on piliation and/or retraction (58). These observations led us to hypothesize that the inability of UNC003 to take up DNA was due to a lesion in comP. Attempts to amplify the comP locus with multiple flanking primers (Table S3) were unsuccessful. Next, we sequenced the genome of UNC003 (NCBI accession number CP154352). Analysis of the genome showed that comP is absent in UNC003. In Neisseria, comP is adjacent to comE, but in UNC003, comE (locus tag AAIB25_09815) is adjacent to a putative peptidase (locus tag AAIB25_09810). To confirm that the absence of a comP is responsible for the inability of UNC003 to take up DNA, a wild-type copy of MS11 comP was inserted into the UNC003 chromosome between trpB and iga (comP_MS11) using the plasmid pMR68-comP_MS11 (46). With this step, the transformation frequency of UNC003::comP_MS11 increased by 3 logs over the negative control, UNC003::erm, which contains an erythromycin marker between trpB and iga (Fig. 4C). Moreover, UNC003::comP_MS11 was killed by Nel DNA and MB-14-1-15 at high efficiency (98% and 97%, respectively) (Fig. 4D). These results strongly suggest that the absence of comP in UNC003 genome resulted in resistance to killing by the microbicide.

Finally, we searched the PUBMLST database (59) for the presence of comP in Neisseria isolates. We found that 5% and 2.6% of the published Ngo and Nme genomes (respectively), and ~2.5% of commensal Neisseria genomes, lack comP (Data S2). This suggests that resistance to DNA killing via this mechanism may be relatively rare.

Investigating the fate of the incoming DNA

To determine whether survivors of DNA killing may have incorporated the incoming DNA into their genome, we performed a killing assay on two isolates, AR173 and AR174 (Fig. S1). These isolates are resistant to the fluoroquinolone ciprofloxacin (Table S6). MB-A and MB-C carry gyrA and parC sequences, respectively. The sequences present in these antimicrobials confer ciprofloxacin sensitivity (28, 29). A disk diffusion antimicrobial susceptibility test was then performed on random survivor cells. All survivors from the MB-A killing assay were resistant to ciprofloxacin as no zone of clearance was visible around the disc. A small number of survivors from the MB-C killing assay had recovered some sensitivity to ciprofloxacin, a zone of clearance for 3 of 10 AR173 and 3 of 12 AR174 survivor colonies was visible and ranged from 22 to 24 mm (Fig. S2). Sequencing of parC showed that survivor cells had acquired the sensitivity allele from the input DNA molecule, suggesting that few bacteria can take up MB-C without being harmed.

Testing the effect of MB14-1-15 on Neisseria meningitidis

Neisseria meningitidis (Nme), a close relative of Ngo, occasionally causes genital infections that can lead to septic shock in women, and anogenital infections in men (60, 61). As Nel DNA also kills Nme (46), we tested the ability of MB-14-1-15 to kill Nme serogroup C strain 8013. Nel DNA and Tris buffer without DNA served as the positive and negative controls, respectively. As reported, Nel DNA killed Nme efficiently (82% ± 4.5, P = 0.009) (46); MB-14-1-15 killed Nme even more efficiently (92.9% 2.6, P < 0.0001). This result suggests that MB-14-1-15 can potentially prevent genital infections caused by Nme.

DISCUSSION

We reported previously that DNA can kill Ngo (46). Here, we described our efforts to leverage this finding by determining whether DNA molecules with defined sequences can be developed into a new class of microbicides for preventing Neisseria gonorrhoeae (Ngo) infection.

We tested three groups of DNA molecules. Group I (MB-F, MB-H, MB-I, and MB-FHI) contains sequences that are present in almost all Ngo strains but rare in commensal Neisseria: tdfF, tdfH, and iga (62–65). Group 2 (MB-14, MB-1, MB-15, MB-14-1, and MB-14-1-15) contains sequences in DNA methyl transferases genes (ngoAI, ngoAXIV, and ngoAXV) that are essential for Ngo viability (48, 53). Group 3 (MB-A, MB-B, and MB-C) contains sequences for DNA maintenance and replication enzymes (gyrA, gyrB, and parC). DNA sequences not common in commensal Neisseria, namely MB-F, MB-H, MB-I, and the concatenate MB-FHI, killed Ngo at a lower efficiency than those containing essential sequences (Fig. 2). Why the concatenate MB-FHI did not perform as well as the single essential sequences is unclear. The genes targeted by Group 2 and Group 3 DNAs are essential for Ngo viability during in vitro growth (53), while those targeted by Group 1 are essential only for intracellular survival (62–64). Additional work will be needed to determine the reason(s) for this variability, and how these findings relate to our current model for DNA killing.

Concatenate MB-14-1-15 was very effective in killing laboratory-adapted Ngo isolates, low-passage clinical isolates, and antibiotic-resistant isolates (Fig. 3). Equally important, its killing efficiency was undiminished when it was suspended in two commercially available personal lubricants. These findings strongly suggest that personal lubricants containing DNA molecules such as MB-14-1-15 can be a practical and low-cost means of blocking Ngo transmission during sexual intercourse.

For reasons we do not understand, MB-14-1-15 was slightly less efficient at killing laboratory strains than low-passage and antibiotic-resistant isolates. One possible explanation for this might be that during continuous passage, the laboratory strains have lost their RM genes or downregulated their expression. MB-14-1-15 was inefficient in killing the commensal Neisseria species tested and did not harm Lactobacillus and Gardnerella vaginalis, two members of the female genital tract flora that are correlated with health and disease, respectively (66, 67). The latter finding is not surprising; to our knowledge, these bacteria do not have a Neisseria-like DNA uptake system and their recombination systems are undescribed.

Nel DNA varied in its ability to kill commensal Neisseria, with killing efficiencies ranging from low (6.9% for Npo) to moderate (34.8% for Nsu) to relatively high (70.8% for Nla). One possible explanation for the variability of these values may be that the methylation pattern of Nel DNA is more similar to that of Npo DNA than Nsu DNA and Nla DNA; hence, the greater ability of Nel DNA to kill Nla and Nsu. Supporting this speculation is the presence in Nla of 14 to 16 RM systems, of which at least 6 are predicted to methylate cytosines (http://rebase.neb.com/rebase/rebase.html). Npo and Nsu have 3–9 and 11 RM systems, respectively, but the nucleotide(s) targets are unknown. Another possible explanation is that RM loci in the Npo strain tested may be downregulated. Additional experimentation is needed to determine the mechanism(s) underlying Npo resistance to killing.

Most of the UNC020 and AR173 cells surviving DNA killing did not express Tfp, an attribute required for the uptake of the DNA molecule. However, most MS11 survivor cells express Tfp; why they are resistant to DNA killing remains to be determined. UNC003, the isolate most resistant to DNA killing, is missing comP, whose gene product binds the Neisseria-specific DNA uptake sequence required for transformation (56) and is dispensable for piliation and retraction (58). These findings demonstrate that multiple mechanisms are used by Ngo cells to acquire resistance to DNA killing.

Does the ability of Ngo to develop resistance to DNA killing lessen the value of DNA molecules as a microbicide? We believe not. In vitro infection studies have shown that Tfp mediates the initial attachment of Ngo to epithelial cells and neutrophils (68–72) and that it is required for biofilm formation and host cell signaling, two activities that are predicted to enhance Ngo survival in human host (73–77). Non-piliated cells cannot perform these functions. In human volunteer studies, it was found that 100% of colonies recovered from infected individuals are piliated; non-piliated bacteria are not infectious (78, 79). In practice, non-piliated cells surviving DNA killing would be less likely to be able to initiate infection. As soon as these cells revert to the piliated state, they will be susceptible to DNA killing. In this context, it is worth noting that the ID₅₀ of Ngo is unknown. This has not been evaluated in human volunteer studies, and current animal models are inappropriate for examining this issue. The number of cells surviving DNA killing may therefore be too low for initiating infection.

Of interest, a small number of cells surviving killing by MB-C, which target parC, have acquired the sequences of this DNA molecule and become sensitive to ciprofloxacin. This approach of using DNA molecules to kill Ngo may have an unintended benefit: in helping to curb antibiotic resistance. Point mutations in promoter regions or inside the ORFs of mtrR, penA, 16S and 23S are responsible for resistance to macrolides, spectinomycin, and cephalosporins (16, 17, 80). A microbicide containing sequences that confer sensitivity to these antimicrobials will test this hypothesis.

In summary, DNA molecules that target essential Ngo genes, typified by MB-14-1-15, are excellent candidates for developing into microbicides that specifically kill pathogenic Neisseria. We had reported that DNA will kill Ngo provided it has homology to the Ngo genome, can be taken up by Ngo, and is methylated differently than Ngo DNA. Our present study extends these features to include DNAs that target genes essential for viability.

MATERIALS AND METHODS

Bacterial strains, tissue culture cells, and growth conditions

Bacteria used in this study are listed in Table S1. Neisseria strains were routinely grown in GCB agar or GC broth with Kellogg’s supplements at 37°C with 5% CO₂. E. coli strains were grown in LB agar or broth and incubated at 37°C. Lactobacillus iners was grown in MRS agar. Gardnerella vaginalis was grown on BHI with 5% defibrinated sheep blood. Lactobacillus iners and Gardnerella vaginalis were incubated at 37°C under anaerobiosis using GasPak (BD). When needed, kanamycin was added to a final concentration of 50 mg/L and erythromycin at 10 mg/L.

Chromosomal DNA extraction

Neisseria elongata (Nel) chromosomal DNA was extracted in a modified method according to the protocol described in references (46, 46). Briefly, Nel was grown on GCB agar for 16 h at 37°C/5% CO₂. Bacteria were harvested and lysed in GC lysis buffer, followed by sequential phenol, phenol-chloroform, and chloroform extractions. The DNA was precipitated using cold ethanol and 7.5 M ammonium acetate, then washed two times with ice-cold 70% ethanol. The DNA pellet was resuspended in 10 mM Tris pH 8. RNA was removed by treatment with RNAse (Thermo Scientific), followed by a second round of phenol-chloroform extraction. The purity of the DNA was assessed by determination of the 260/280 and 260/230 ratios using a Nanodrop instrument. DNA was considered pure and suitable for killing assay when the ration 260/80 was between 1.8 and 1.9 and the ratio 260/230 was >2.

Killing assay

Neisseria killing assays were performed as described in references (46). Briefly, piliated colonies were lawned onto GCB agar and incubated at 37°C/5% CO₂ for 14 h. 5 × 10⁵ CFUs were suspended in 250 µL of GC broth containing Kellogg’s supplements and added to one well of a 12-well plate. DNA was added to a final concentration of 10 µg/mL. Nel DNA was used as a positive control. 10 mM Tris, pH 8 was used as a negative control. Bacteria were incubated for 5 h at 37°C/5% CO₂. Bacteria were recovered, and dilutions were plated on GCB agar to enumerate the surviving bacteria. Bacteria growing in the presence of 10 mM, pH 8 Tris were considered to have a 100% survival rate.

PCR and cloning

Internal fragments of tdfF, tdfH, iga, ngoAXIV, ngoAI, ngoAXV, parC, gyrA, and gyrB lacking promoter, regulatory sequences, start and stop codons were PCR amplified with Phusion-HF (New England Biolabs), primers are listed in Table S2. Each amplicon was purified and cloned into pCR-Blunt (Life Technologies). Products were transformed into E. coli DH5-alpha. The insertion of the fragment was confirmed by digestion with EcoRI, followed by sequencing (Eurofin genomics).

Primers used to concatenate tdfF, tdfH, and iga, and ngoAXIV, ngoAI, and ngoAXV were designed with NEB builder assembly tools (New England Biolabs). Individual tdfF, tdfH, and iga, or ngoAXIV, ngoAI, and ngoAXIV were obtained by PCR with Phusion-HF using primers listed in Table S2. Sequences were concatenated by Gibson assembly (New England Biolabs) following the manufacturer’s instructions. Assembly products were cloned pCR-Blunt (Life Technologies) and transformed into E. coli DH5-alpha. Transformants were selected on LB with kanamycin 50 mg/L. The fragment insertion was confirmed by digestion with EcoRI, followed by sequencing (Eurofin genomics).

Plasmid extraction

Small-scale (5 mL) plasmid preps were conducted using a Gene Jet Plasmid kit (Thermo Scientific) following the manufacturer’s instructions. Large-scale plasmid extractions were performed by alkaline lysis (81). Briefly, E. coli was harvested from a 100 mL LB culture grown at 37°C with 250 rpm shaking for ~18 h. The bacterial pellet was resuspended in 10 mL of buffer 1 (50 mM Tris-HCl, pH 8.0; 10 mM EDTA; 100 mg/mL RNase A), then lysed by adding 20 mL of buffer 2 [0.2 M NaOH; 1% (wt/vol) SDS]. Lysates were neutralized with 15 mL of buffer 3 (3 M potassium acetate, pH 5.5). Cell debris was removed by centrifugation. The plasmid was precipitated in ethanol, washed 3 times with 70% ethanol, and then dried at room temperature. Dried plasmid was resuspended in 10 mM tris pH 8. RNA was removed by treatment with RNAse (Thermo Scientific), followed by a final phenol-chloroform extraction. The purity of the plasmid was assessed by determination of the 260/280 and 260/230 ratios using a Nanodrop. Plasmid was considered pure and suitable for killing assay when the ration 260/80 was between 1.8 and 1.88 and the ratio 260/230 was >2.

Whole genome sequencing

Genomic DNA of strain UNC003 was purified, sequenced, and annotated by the Plasmidsaurus service certified by Oxford Nanopore Technologies (Eugene, Or, USA). Libraries were created according to Plasmidsaurus standard workflow for library preparation. A total of 126,405 reads (660,775,844 bp total, N50 13,382 bp) were generated, raw coverage was 288×.

DNA transformation assay

DNA transformation assays were done as described by Dillard (82). Briefly, 5 × 10⁵ CFUs were incubated with 1 µg of Ngo MS11 Rif^R for 30 min in GC broth with 5 mM MgSO4. Bacteria was allowed to recover for 5 h in GC broth containing Kellogg’s supplements in six-well plates, a reaction without DNA was used as control. UNC003 transformed with an empty pMR68 was used as a negative control. Plates were incubated for 5 h. The contents of the plate were harvested, diluted, and plated on GCB Rif 40 µg/mL, incubated for 48 h at this time CFUs were counted. Transformation frequencies were calculated as (#Rif^R CFUs/Total CFUs)/1 µg DNA.

Motility assay

Two million cells from strain UNC003 were grown in 2 mL of GC broth containing Kellogg’s supplements and NaHCO₃ in 35-mm glass bottom dishes (MatTek). Bacteria were incubated for 1 h at 37°C/5% CO₂ to allow bacteria to settle before imaging. Dishes were placed in an environmental chamber (company) equilibrated at 37°C and 5% CO₂. Images were captured at 30-second intervals for 6 h, using a Nikon Eclipse Ti microscope with a 40× oil objective. The 6 h of video was compressed into 1 min 48 seconds using NIS elements imaging software.

Neisseria gonorrhoeae UNC003 comP complementation

Complementation of comP in UNC003 was done using plasmid pMR-68 carrying Ngo MS11 comP as previously described (46). Electroporation was carried out as described by Dillard (82). Briefly, bacteria from a 16-h lawn were collected in 1 mL of 0.3% sucrose. Bacteria were pelleted by centrifugation. The supernatant was removed and washed two more times with 0.3% sucrose. After the last wash, bacteria were suspended in 1 mL of 0.3% sucrose. 1 µg of DNA was added to the bacteria. Cells were transferred into an electroporation cuvette and electroporated at 2.5 kV, 200 Ω, and 25 µF. Bacteria were resuspended in 1 mL of GC broth containing Kellogg’s supplements and transferred to a GCB plate. After 5 h of incubation, bacteria were harvested, diluted, and plated onto GCB with erythromycin at 10 µg/mL. Transformants appeared after 24 h of incubation. Empty pMR68 was transformed separately into UNC003, for use as a negative control.

Antibiotic resistance determination

The strip test was used to verify the antibiotic resistance of the trains used. Bacteria were prepared by suspending half of a plate from a 16-h lawn of Ngo grown on GCB agar. Bacteria were adjusted to an OD₆₀₀= 0.6 equivalent to a 0.5 McFarland standard. Using a sterile swab, bacteria were spread on the surface of a GCB agar plate. Plates were allowed to dry for 10 min prior the addition of a ciprofloxacin strip on the plate. Plates were incubated at 37°C in 5% CO₂ for 20 h.

A sensi disc assay was used to determine the ciprofloxacin resistance of survivors after the killing assay. A survivor colony was picked with a sterile calcium alginate swab and spread on a one-quarter GCB agar plate. A sensidisc (BD, catalog number 231657) was placed in the center. Plates were incubated at 37°C in 5% CO₂ for 20 h. The diameter of zones of clearance (no-growth zones) was measured in mm.

Statistical analysis

Statistical differences in the killing assay were evaluated using one-way ANOVA with Turkey’s post-test correction, DNA transformation assay was evaluated using a Student’s t-test using GraphPad Prism 9 software.

DATA AVAILABILITY

The annotated genome sequence of UNC003 is available in NCBI under accession number CP154352.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00794-24.

Data Set S1. aac.00794-24-s0001.xlsx.

Microbicide sequences.

Data Set S2. aac.00794-24-s0002.xlsx.

Presence of comP, tdfF, tdfH, and iga in Neisseria strains.

Figure S1. aac.00794-24-s0003.pdf.

MB-B and MB-C efficiently kill antibiotic-resistant Neisseria gonorrhoeae isolates.

Figure S2. aac.00794-24-s0004.pdf.

Ciprofloxacin resistance of four surviving Ngo AR173 colonies after a killing assay.

Supplemental material. aac.00794-24-s0005.docx.

Supplemental legends; Tables S1 to S6.

Video S1. aac.00794-24-s0006.mov.

UNC003 crawls and forms tight microcolonies.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.