Characterization of outer membrane vesicles released by clinical isolates of Neisseria gonorrhoeae

The sexually transmitted pathogen Neisseria gonorrhoeae releases membrane vesicles including outer membrane vesicles (OMVs) during infections. OMVs traffic outer membrane molecules, such as the porin PorB and lipo‐oligosaccharide (LOS), into host innate immune cells, eliciting programmed cell death pathways, and inflammation. Little is known, however, about the proteome and LOS content of OMVs released by clinical strains isolated from different infection sites, and whether these vesicles similarly activate immune responses. Here, we characterized OMVs from four N. gonorrhoeae isolates and determined their size, abundance, proteome, LOS content, and activation of inflammatory responses in macrophages. The overall proteome of the OMVs was conserved between the four different isolates, which included major outer membrane and periplasm proteins. Despite this, we observed differences in the rate of OMV biogenesis and the relative abundance of membrane proteins and LOS. Consequently, OMVs from clinical isolates induced varying rates of macrophage cell death and the secretion of interleukin‐1 family members, such as IL‐1α and IL‐1β. Overall, these findings demonstrate that clinical isolates of N. gonorrhoeae utilize membrane vesicles to release proteins and lipids, which affects innate immune responses.


K E Y W O R D S
gonorrhoea, inflammasome, inflammation, macrophage, OMV, proteomics

INTRODUCTION
Neisseria gonorrhoeae is a diplococci bacterium that colonizes genital, rectal and oral mucosa surfaces resulting in gonorrhoea in 86.9 million people of 15-49 years of age globally in 2016 [1].Gonorrhoea is the second most common sexually transmitted disease with 583,405 reported cases in the USA in 2018, an increase of 63 percent compared to 2014 [2,3].In addition to increased infection rates, N. gonorrhoeae isolates that have developed resistance to front-line antibiotics have now been detected in several regions around the globe [4,5].In the absence of new antibiotics, alternative treatment options that control bacteria numbers and tissue-damaging inflammatory responses associated with gonorrhoea are needed.The development of new interventions depends on a better understanding how N. gonorrhoeae causes infections and inflammation.
N. gonorrhoeae can colonize mucosal surfaces without symptoms and thus remain undetected and underdiagnosed [6].N. gonorrhoeae infections persist due to immune evasion depending on phase and antigen variation of surface molecules, including Opa, adherence factor pili and lipo-oligosaccharide (LOS) [7].In addition, LOS can undergo molecular mimicry through addition of sialic acid on the terminal sugar, thereby resembling human glycosphingolipids [8].Despite these mechanisms to evade immunity, symptomatic gonorrhoea is associated with excessive innate immune responses resulting in tissue damaging inflammation without efficient elimination of N. gonorrhoeae [6].This notion is best understood in urethral and cervical infections due to the access to human male volunteers and female reproductive organs, respectively [9].Ascending N. gonorrhoeae infections into the upper genital tract such as the fallopian tubes can cause pelvic inflammatory disease (PID), which is associated with infertility, ectopic pregnancy, and chronic pelvic pain [10].PID severity correlates with the levels of pro-inflammatory cytokines and influx of inflammatory immune cells, such as neutrophils, to the fallopian tubes [11,12].Besides reproductive organs, N. gonorrhoea can infect other mucosal surfaces and trigger inflammation, as observed with oropharyngeal gonorrhoea [13].Molecules secreted by N. gonorrhoeae and responses by host cells are thought to drive inflammation [14].
Membrane vesicles have emerged as a novel mechanism by which bacteria interact with other cells, including host cells [15].Amongst the different bacterial membrane vesicles, outer membrane vesicles (OMVs) primarily originate from the outer membrane of live bacteria [16].N. gonorrhoeae OMVs range from 50 to 250 nm in size and are enriched in outer membrane proteins relative to inner membrane and cytoplasmic proteins [17][18][19].The factors that regulate OMV biogenesis in N. gonorrhoeae, which releases OMVs spontaneously at high rates compared to other Gram-negative bacteria like Escherichia coli, remain ill-defined [19].The lack of Braun's lipoprotein in N. gonorrhoeae which anchors the outer membrane to the peptidoglycan layer, likely promotes OMV biogenesis [20,21].In addition, the downregulation of the phospholipid transporter MlaA due to iron starvation results in increased vesiculation in Neisseria sp.[22].Thus, the biogenesis of OMVs is affected by environmental factors, such as those encountered at the mucosal surface [19].
Thus, to understand the role of OMVs in N. gonorrhoeae-associated pathogenesis, we have determined the size, quantities, and proteome of OMVs released by four clinical isolates originating from urethra, pharynx, cervix, and vagina.We show that these OMVs trigger innate immune responses in macrophage to varying degrees.

Outer membrane vesicle purification
OMVs were isolated from overnight 500 mL broth cultures, supplemented with 1% DMIV and 0.01 M NaHCO

Electron microscopy
OMVs in PBS (10 µL, 500 µg/mL protein concentration) were applied on carbon-coated 300-mesh copper grids (3.05 mm diameter: 0.4 × 2 mm single slot) for 10 s and excess fluid was soaked in filter paper.Girds were then stained with 1% (w/v) uranyl acetate negative stain for 30 s and washed twice with PBS.The excess fluid was again soaked in filter paper.The grids were viewed with a Jeol-1400 transmission electron microscope operated at 60 KV.

Nanoparticle tracking analysis
OMV concentration and size was determined using the nanoparticle tracking analyser ZetaView basic PMX-120 NTA (Particle Metrix, Germany) as previously described [39,40].OMVs were diluted in 1 mL of DPBS to an appropriate concentration between 1 and 5 × 10 7 particles per mL.Instrument calibration was performed using 102 nm polystyrene beads (ThermoFisher, USA) according to manufacturer's guidelines.Measurements were taken using a 405 nm 68 mW laser and CMOS camera, at 11 cell positions and captured 60 frames per

Significance Statement
In this study, we used four clinical strains of Neisseria gonorrhoeae that have been isolated from the cervix, vagina, pharynx and urethra, and characterized their outer membrane vesicles (OMVs).We identified that the different OMVs shared a conserved set of proteins, including major outer membrane proteins that have been previously been identified to important during infections.Importantly, our analysis showed that the relative abundance of proteins can varying between the isolates.Similarly, the major outer membrane lipid, LOS, showed strain-specific modifications and packaging into membrane vesicles.

Coomassie staining
An aliquot of the filtered OMVs (20 µL) collected as described above was used to measure protein concentration using bicinchoninic acid assay, according to manufacturer's instructions.The remaining OMV samples were normalized to 150 µg/mL using PBS.Then, five parts of normalized OMV samples were added to one part of 6 X SDS loading buffer.The equal volume of the prepared OMVs and bacterial lysates were run at 125 V for 2 h in 15% resolving SDS-PAGE.The gel with separated proteins was then stained with Coomassie blue overnight at room temperature with gentle shaking.The gel was washed with destain (40% ethanol, 7.5% acetic acid) to reduce background staining.

Immunoblot analysis
Purified OMVs and macrophages were analysed by immunoblot analysis as described previously [33].BMDMs were seeded at concentration of 2.5 × 10 5 per well in a 24 well plate and treated with OMVs (50 µg/mL based on total protein) the next day for 24 h at 37˚C.The supernatant was treated with 6 x SDS loading buffer and the cells were scraped of the plates in 1 x SDS loading buffer.Both the supernatant and cell lysates were heated at 95˚C for 10 min.Then, the equal volume of supernatant and cell lysates were run at 300 V for 1.

Proteomics
Purified OMVs from ten independent bacterial cultures were pooled into two samples and solubilized in 1% (w/v) sodium deoxycholate, 100 mM HEPES, pH 8.1.The OMVs lysate was then boiled at 95˚C for 5 min and sonicated three times for 10 s with a cooling down step after every sonication.Then, the protein concentration of the lysate was measured using the BCA assay.100 µg of protein was denatured partner repository with the dataset identifier PXD042787 [43].The raw data files were analysed using MaxQuant to obtain protein identifications and their respective label-free quantification (LFQ) values [44].Statistical analysis was performed in Perseus and LFQ-Analyst [45,46].

Proteome analysis
To determine subcellular distribution of identified proteins,

Live cell imaging
Analysis of macrophage survival and the quantification of cell death rates were performed as described previously [47].Macrophages were labelled with the fluorescent dyes Cell Tracker Green (1 µM) (LifeTechnologies) and DRAQ 7 (0.6 µM) (Abcam 109202) to count total number of cells and dead cells, respectively.Macrophages were imaged using fully automated fluorescence microscope (Leica DMi8 epi-fluorescence microscope) that contained motorized stage and controlled environment (5% CO 2 and 37˚C) using a 10x objective (NA: 0.8).
Bright field and fluorescent images were taken every hour over 72 h.
The obtained images were analysed using MetaMorph customized journals, Microsoft Excel and GraphPad Prism to determine number of total dead cells (Draq7 positive cells).

Enzyme Linked Immunosorbent Assay (ELISA)
The supernatants from BMDMs were used to determine the release of cytokines (IL-1β, IL-α) using ELISA following the manufacturers protocol (R&D Systems, Catalog # DY401 and DY400).Briefly, a flat transparent 96 well plate was coated with 100 µL of capture antibody solution overnight at room temperature.The plate was washed with 0.05% Tween 20 in PBS and 100 µL of samples (diluted 1:20 in BSA) as well as reconstituted standards were added in duplicates for 2 h at room temperature.Samples were washed with wash buffer and 100 µL of detection antibody was added.Samples were incubated for 2 h at room temperature and washed with wash buffer.Next, 100 µL of streptavidin HRP was added.Samples were incubated for 20 mins at room temperature in dark, then washed with wash buffer.Further, 100 µL of substrate solution was added, incubated for 20 min at room temperature in dark.Finally, the development of colour was stopped using 50 µL stop solution and the optical density was measured using a microplate reader at 450 nm wavelength.

Statistical analysis
Data analysis was performed using GraphPrism (v9) and represented as mean or median +/standard error of the mean (SEM) from three biological replicates (or as indicated in the figure legends).Statistical analysis was performed using Student's t-test, one-way or two-way analysis of variance (ANOVA) with a Tuckey's multiple-comparisons test as indicated.

Clinical strains of N. gonorrhoeae release outer membrane vesicles
Previous studies on N. gonorrhoeae OMVs have focused on MS11-A and FA1090, which are commonly used strains in the field [19].Here, we focused on N. gonorrhoeae strains isolated from individuals that presented at local health clinics.The clinical isolates with the reference number R72.04, R72.14, R72.17, and R72.20 are referred to as N. gon-1, N. gon-2, N. gon-3, and N. gon-4, in this study, and were isolated from the pharynx, cervix, vagina, and urethra, respectively (Table 1).OMVs from these isolates, referred to NOMV-1, −2, −3, and −4, were overall similar in shape and size, as visualized by transmission electron microscopy (TEM) (Figure 1A).To quantify the size and concentrations of OMVs, we next used ZetaView Nanoparticle Tracking Analysis (NTA).The OMVs had a similar size distribution ranging from 50 to 300 nm in diameter, whereby NOMV-2 was marginally bigger in size (Figure 1B).The median diameter of the four OMVs, however, was not significantly different (Figure 1C).OMVs normalized to bacterial numbers, based on OD 600 , ranged from 2.4 × 10 10 to 1.3 × 10 13 particles/mL (∼500-fold difference) (Figure 1D).While NOMV-1, 2, and −3 reached similar concentrations, there were statistically fewer NOMV-4 particles per parent bacteria (p < 0.05) (Figure 1D).Consistent with reduced rates of OMV generation, N. gon-4 secreted lower levels of PorB, previously identified as the major outer membrane protein in OMVs, compared to the other strains, despite similar expression levels in the parent bacteria (Figure 1E).N. gon-4 also secreted lower levels of other proteins, as evidenced by total protein staining, compared to the other isolates (Figure 1E).NOMV-1, −2, and −3, contained similar levels of several major proteins, although some proteins appeared differentially enriched or reduced (Figure 1E).This data shows that clinical N. gonorrhoeae isolates produce OMVs that share similarities but can differ in their relative concentration which reflects PorB secretion.

The OMV proteome from N. gonorrhoeae clinical isolates
We have previously shown that N. gonorrhoeae package outer membrane and other proteins into OMVs for release [19].To determine the proteome of OMVs derived from clinical isolates of N. gonorrhoeae, we next performed quantitative mass spectrometry on NOMV-1, −2, −3, and −4.A total of 688 proteins were identified across all four NOMVs based on a predefined false discovery rate (FDR) threshold of 1%, a p-value < 0.05 and at least three different peptides per protein.
Similarly, outer membrane assembly machinery subunits BamA, BamC, BamD, and BamE, and opa proteins, such as OpA54 were detected in all NOMVs.The relative abundance of proteins varied between NOMVs.For instance, OpA54 was abundant in NOMV1 and NOMV4, but relatively less so in NOMV2 and NOMV3 (Figure S1).The NOMVs from clinical isolates also contained putative peptidyl prolyl isomerase, membrane lysozyme inhibitor of c-type lysozyme family protein and MtrE, the outer membrane protein of the multidrug efflux transporter, which have not previously been associated with NOMVs (Figure S1) [19].To further characterise the OMV proteome, we analysed the subcellular localization of the identified proteins.Of the 179 proteins identified in NOMV-1, 27% are outer membrane (45) and 24% periplasmic proteins (43), consistent with the close link of OMVs with these compartments.In addition, these OMVs contain 14% inner (cytoplasmic) membrane proteins and 31% cytoplasmic proteins (Figure 2B).NOMV-2, −3, and 4 contained proteins with similar distribution of subcellular localization with some variations (Figure 2B).Gene Ontology (GO) terms enrichment analysis identified 14 clusters of proteins whereby the most enriched cluster comprised cell outer membrane proteins and outer membrane assembly proteins (p-value-2.90E-06)(Figure 2C).NOMVs also contained a number of cytoplasmic proteins involved in metabolism, including central carbon pathways of glycolysis and TCA-cycle, which may reflect targeting to the inner membrane as is the case for succinate dehydrogenase (Figure 2C).Similar results were obtained for all NOMVs, suggesting that clinical isolates release OMVs that have broadly similar proteome and functional roles.

Relative enrichment of proteins and lipids in OMVs
While the proteins secreted via OMVs are largely shared between the four different N. gonorrhoeae clinical isolates, we did observe differences in their relative amounts.For instance, several proteins were enriched in NOMV-3 compared to NOMV-4 but underrepresented compared to NOMV-2 (Figure 3A).Further analysis showed that the levels of a number of proteins varied across the four NOMVs (Figure 3B).This included uncharacterized proteins but also membrane-associated enzymes involved in dihydrodipicolinate synthesis (DapA and DapB, enriched in NOMV-3), a glycosyltransferase involved in LOS biosynthesis (LtgG enriched in NOMV-4), the glycerol-3-phosphate acyltransferase PlsY, serine/threonine transporter SstT (enriched in NOMV-3 and −4) and the pilus assembly proteins PilW (enriched in NOMV-2) and PilE (enriched in NOMV-3).While the major outer membrane proteins were conserved between the different NOMVs, the analysis suggested reduced levels of BamA in NOMV-2 compared to NOMV-3 (Figure 3A).We thus probed for specific proteins by immunoblot analysis using available antibodies against BamA and PorB.PorB was present at similar levels in NOMV-1, −2, and −3, and comparable to OMVs derived from the previously studied MS11-A strain (Figure 3C).In contrast, we detected decreased levels of PorB in NOMV-4, despite similar amounts of total protein (Figure 3C).BamA levels were similar between NOMV-4 and MS11-A OMVs (Figure 3C).BamA, however, was enriched in NOMV-3 and NOMV-1 compared to NOMV-2 (Figure 3C).Besides proteins, N. gonorrhoeae derived OMVs contain membrane lipids such as LOS.All N. gonorrhoeae strains produced LOS, albeit with different molecular weights which reflects LOS modifications (Figure 3C).The LOS patterns were similar between OMVs and parent bacteria, whereby NOMV-2 contained the least modifications (Figure 3C).NOMV-1 and 3 had increased levels of LOS compared to MS11-A, whereas detectable LOS was markedly reduced in NOMV-4 (Figure 3C).Taken together, OMVs from clinical isolates of N. gonorrhoeae contain varying levels of major outer membrane proteins and lipids.

Activation of macrophage immune responses by N. gonorrhoeae derived OMVs
We have recently identified that E. coli-derived OMVs (EOMVs) activate the cytosolic LPS sensor caspase-11, which resulted in pyroptotic cell death and caspase-1-dependent secretion of IL-1β from bone marrow-derived macrophages (BMDMs) [33].As shown previously, NOMVs from MS11-A failed to markedly activate pyroptosis compared to EOMVs, but nevertheless triggered delayed macrophage cell death (> 40 h post treatment), as determined by live-cell imaging to quantify loss of plasma membrane integrity over time (Figure 4A).The delayed macrophage cell death due to NOMVs has previously been shown to depend on mitochondrial apoptosis, rather than pyroptosis [33].Control (PBS) treated BMDMs remained viable over the 72 h (Figure 4A).This is consistent with the notion that NOMVs activate two distinct cell death pathways including LOS-dependent pyroptosis during the early interaction (phase 1 up to 48 h post treatment) and PorB-dependent mitochondrial apoptosis at later time points (phase 2).
Given that NOMV from clinical isolates contained similar proteins and LOS, but to varying levels, we next investigated immune responses in macrophages.NOMVs isolated from the four clinical strains triggered cell death in the majority of BMDMs over time, although with varying kinetics (Figure 4A).For instance, NOMV-1 and 2 caused higher rates of cell death in phase 1, whereas NOMV-4 was indistinguishable to NOMVs derived from MS11-A (Figure 4A).All NOMVs induced rapid cell death during phase 2, although NOMV-3 showed a delay of up to 10 h (Figure 4A).Activation of pyroptosis, but also apoptosis, leads to the processing and secretion of interleukin 1 family members.
Consistent with this notion, EOMVs triggered IL-1β and IL-1α secretion from BMDMs (Figure 4B, C, and D).In contrast, NOMVs derived from MS11-A triggered around five to ten times less IL-1α and IL-1β secretion compared to EOMVs, including at 24 and 48 h post treatment (Figure 4C and D).The levels of secretion of both cytokines were increased over time (48 h) in BMDMs treated with MS11-A NOMVs, likely reflecting increased levels of macrophage death at later time points (Figure 4C and D).Macrophages treated with NOMV-1 to −4 similarly induced the secretion of IL-1β and IL-1α (Figure 4C and D).
NOMV-2 and −3 triggered significantly higher concentrations of IL-1β compared to NOMV derived from MS11-A (Figure 4C).Taken together, OMVs released by clinical isolates of N. gonorrhoeae activate programmed cell death pathways and inflammatory responses in macrophages to varying degree.

DISCUSSION
N. gonorrhoeae releases proteins and lipids via OMVs, which are sensed by innate immune cells, such as macrophages.Whether recent clinical isolates derived from various infection sites similarly release OMVs has not been addressed.Here, we characterized OMVs from recent clinical isolates of N. gonorrhoeae and determined vesicle size, quantity, and proteome profile.Our data suggests that N. gonorrhoeae isolates produce OMVs with a shared proteome, and that the OMV biogenesis and the relative amount of proteins and glycolipids can differ between the strains.Consequently, OMVs from different clinical isolates induce innate immune responses to varying degree.
The N. gonorrhoeae strains used here were initially isolated from different infection sites, including the pharynx, cervix, vagina, and urethra.
To analyse their OMVs, the isolates were used after low passage number and cultured under the same conditions.We detected over 700 proteins associated with each OMV, and stringent analysis identified Most strikingly, the N. gonorrhoeae isolate from the urethra produced relatively low amounts of OMVs (NOMV-4) compared to the other isolates.The mechanism for this remains to be identified.It is possible that regulatory and synthetic pathways that control OMV biogenesis were affected by genetic differences between the isolates.This may include iron uptake, known to be phase variable in N. gonorrhoeae [48].Low iron levels control the expression of retrograde phospholipid shuttling channel MlaA resulting in hypervesiculating phenotype [22].It is also tempting to speculate that niche related factors like oxygen concentrations, presence of antimicrobials, reactive oxygen species and the microbiome affect OMV biogenesis, which is thought to enable adaptation to different environmental conditions [6,[49][50][51][52].
The current model suggests that OMVs counteract cellular stress by enabling secretion of toxic products such as misfolded proteins in the periplasm [21].At this point, however, it remains unclear whether the proteome of the different OMVs analysed in this study reflected increased concentrations of unfolded proteins.Intriguingly, cervical gonococcal isolates can contain loss of function mutations in major efflux pumps, likely an adaptation to acidic conditions [49].Under these conditions, bacteria must employ alternative strategies, such as increased OMV biogenesis, to counteract the otherwise increased susceptibility to antibiotics and host-derived antimicrobial in the absence of efflux pumps.Any of these programs that affect OMV biogenesis, however, operated in bacteria that were grown under the same growth conditions, suggesting more stable alterations.Future studies will be based on larger sample sizes to establish whether rates of OMV production are consistently associated to infections sites.
Differences in the rate of production and the proteome of OMVs likely affect other host-pathogen interactions [53,54].For instance, clinical isolates of Acinetobacter baumannii that released increased levels of the virulence factor Omp38 and OmpA via OMVs induced heightened immune responses [55].Similar to OmpA, PorB is targeted to mitochondria of host cells in an OMV-dependent manner, resulting in delayed cell death by activating apoptosis [33,35,56].
Rather than activating apoptosis directly, recent findings have demonstrating that N. gonorrhoeae-derived OMVs primarily cause mitochondrial dysfunction, which is sensed by macrophages to induce the activity of the host pore-forming proteins BAX and BAK [33].
Similar to the MS11-A strain studied previously, the OMVs from the clinical isolates used here caused comparable rates of macrophage apoptosis, although apoptotic cell death was somewhat delayed in macrophages treated with NOMV-3.Whether this reflects differences in targeting mitochondria, protein delivery to mitochondria or the ability to interfere with the apoptotic machinery awaits clarification.
Caspase-11 activation is associated with the release of IL-1α and IL-1β, whereby the latter requires the NLRP3 inflammasome [30].Similar to LPS containing OMVs, we have recently shown that OMVs from N. gonorrhoeae MS11-A contain LOS and activate caspase-11 mediated pyroptosis in mouse macrophages, and this precedes apoptosis [33].
Here, we identified that rates of pyroptosis varied in macrophage after treatment with OMVs from clinical isolates.OMVs from MS11-A and NOMV-4 triggered low levels of pyroptosis compared to E. coli derived OMVs [33].In contrast, NOMV-1 and NOMV-2 activated significant higher rates of pyroptosis, and this was associated with relatively high levels of LOS.NOMV-3 also caused increased rates of pyroptosis compared to MS11-A derived OMVs, despite similar LOS levels.This may reflect differences in LOS species capable of activating caspase-11.
However, the lipid A moiety of bacterial LOS, which is thought to be recognized by caspase-11 [31], was similar between the isolates (Figure S2), suggesting additional mechanisms such as sugar modifications.
Indeed, immunoblot analysis demonstrated that LOS associated with OMVs and parent bacteria was differently modified amongst the isolates, which may include GlcNAc-dependent extensions of the α-chain as corresponding glycosyltransferases are phase variable in N. gonorrhoeae [57].OMV-associated LOS levels and its modifications likely also affected the rates of IL-1α and IL-1β release from macrophages.Consistent with this notion, macrophages exposed to NOMV-2 and NOMV-3 released higher levels of IL-1β compared to MS11-A OMV treatment, although well below compared to E. coli-derived OMV treatment.
Intriguingly, macrophage secreted relatively high levels of IL-1α after exposure to NOMV-1 and −2 at later time points (48 h post treatment).
Recent findings have demonstrated that the apoptotic machinery can engage inflammatory pathways [33,58].Future studies are aimed at deciphering how different OMVs activate inflammatory pathways and whether this correlates with clinical outcomes in gonorrhoea patients and in infection models.
In conclusion, our data demonstrated that clinical isolates of N. gonorrhoeae produced OMVs that contained a conserved core proteome.

and alkylated by adding 1
µL of freshly made 1 M TCEP (Tris(2carboxyethyl) phosphine hydrochloride) and 8 µL of freshly made 0.5 M CAA (2-Chloroacetamide) at a final concentration of 10 mM and 40 mM, respectively.The mixture was boiled for 5 min.The samples were allowed to cool down to room temperature, pH adjusted to ∼8 and treated with trypsin (1:100) overnight whilst shaking at 37˚C.Next day, the samples were cleaned up using ziptips-a cartridge column packed with C18-to remove detergent and salt contents of the sam-ples.In brief, the ziptip was sequentially equilibrated three times with buffer B (50% acetonitrile, and 0.1% formic acid) and three times with buffer A (0.1% formic acid).The ziptip was then loaded with the protein sample (8 to 10 times), subsequently washed six times with buffer A and the bound proteins were eluted with buffer B. The acetonitrile was removed in a vacuum concentrator, and the sample was reconstituted in 10 µL buffer A. The samples were sonicated for 10 min in sonicator water bath, centrifuged for 5 min full speed and then transferred into a mass-spectrometry vials.Samples were analysed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) using a QExactive Plus Orbitrap mass spectrometer (ThermoScientific) attached to a nano-high performance liquid chromatography (HPLC) (Dionex Ultimate 3000, Thermo Scientific) using an Acclaim PepMap RSLC analytical column (75 µm × 50 cm, nanoViper, C18, 2 µm, 100 Å; ThermoScientific).The instrument was operated in data-dependent acquisition (DDA) mode to automatically switch between full scan MS and MS/MS acquisition using in-house parameters.The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE

F I G U R E 2
Proteomic analysis of OMVs derived from clinical N. gonorrhoeae isolates.(A) NOMV-1, −2, −3, and −4 were analysed by mass spectrometry and the identified proteins compared using a Venn diagram.(B) NOMV-1, −2, −3, and −4 associated proteins were analysed based on their predicted subcellular localization.(C) NOMV-1 associated proteins were analysed for functional annotation in DAVID.The top ten clusters are presented in the bar diagram.The enrichment score (E.score) as well as GO term are shown on the left of the bar diagrams.The bar diagrams represent the number of proteins present in the cluster as shown the bottom of the bar diagram.
105 core proteins with high confidence that were shared among the four OMVs.Proteins associated with these OMVs were derived primarily from the outer membrane as well as the periplasm and were functionally linked to membrane assembly and function (Enrichment score-5.2).Based on peptide counts, the OMVs from the four clinical isolates were enriched in outer membrane proteins, whereby PorB was the most abundant protein in all four OMVs.This observation is con-F I G U RE 3 Differential expression of OMV-associated proteins.(A) Volcano blots of differentially abundant proteins between NOMV-1, −2, −3, and −4 showing log2 fold change and log 10 adjusted p < 1.5 significant differences (student t-test).(B) Protein identification and abundance heat map of NOMV-1, −2, −3, and −4.Log2 centred intensities and p < 0.05 (One-way ANOVA).(C) Whole bacterial lysates of N. gonorrhoeae N. gon-1, −2, −3, and −4, and total protein of their OMVs (NOMV-1, NOMV-2, NOMV-3, and NOMV-4) were analysed by immunoblotting using anit-BamA, -PorB, and -LOS antibodies.Samples were normalized to total protein concentrations.Molecular weight markers are indicated on the left.F I G U R E 4 N. gonorrhoeae-derived OMVs activate macrophage cell death pathways.(A) Wild type bone marrow-derived macrophages (BMDMs) were treated with vehicle control (PBS) and 50 µg/mL OMVs derived from E. coli OMVs (EOMVs), N. gonorrhoeae MS11-a (NOMVs) and clinical isolates (NOMV-1, −2, −3, and −4).Rates of cell death (DRAQ7 positive) were determined using live-cell imaging.Mean and SEM from three independent experiments.(B) Secretion of mature IL-1β (p17) from BMDMs treated with OMVs was detected in culture supernatant by immunoblotting.Ponceau staining of whole bacterial lysates is included as loading control.(C and D).The secretion of cytokine IL-1β (C) and IL-1α (D) from BMDMs exposed to EOMVs, N. gonorrhoeae NOMVs and NOMV-1, −2, −3, and −4 at 24 and 48 h was determined using ELISA.Mean and SEM from three independent experiments *** indicate p-value < 0.001 and ns non-significant differences.sistentwith previous studies using OMVs from the commonly used N. gonorrhoeae isolates MS11-A and FA1090[19].Immune blot analysis using serum from meningococcal PorB immunized animals, however, indicated relatively low levels of PorB in OMVs derived from the urethral isolate (NOMV-4), despite normalized total protein amounts.It is possible that NOMV-4 contained a variant of PorB which was not readily recognized by the antibodies used.However, the antibodies recognized bacteria-associated PorB in N. gon-4, suggesting alternative mechanisms.Additional outer membrane proteins varied in their relative abundance.This included BamA which showed varying levels as detected by immune blot analysis and at least 25 other proteins based on peptide counts that were statistically significant between the NOMVs.This suggests that protein packaging into OMVs can vary between isolates.It is also possible that the identified proteins undergo natural variation, which may further affect our analysis.While the work here was performed on isolates undergoing minimal culturing to avoid natural variation, further studies are required to address how N. gonorrhoeae regulates selective protein packaging into OMVs and whether this is associated with the infection site.

Furthermore, the relative
abundance of proteins and lipids, as well as the concentration of OMVs can differ among the clinical isolates.This affected macrophage immune responses, including the programmed cell death pathways of pyroptosis and apoptosis.ACKNOWLEDGMENTS The authors thank Drs Rhys Dunstan and Simon Crawford for expert assistance with the gradient fractionator and transmission electron microscope, respectively.This study used BPA-enabled (Bioplatforms Australia) / NCRIS-enabled (National Collaborative Research Infrastructure Strategy) infrastructure located at the Monash Proteomics and Metabolomics Platform.M.K.-L. was supported by a Veski Inspiring Women Fellowship and by the Australian Research Council (DP190101655).The study was supported by NHMRC Ideas Grant (1163556) and TN by an ARC Future Fellowship (FT170100313).Open access publishing facilitated by Monash University, as part of the Wiley -Monash University agreement via the Council of Australian University Librarians.