Commonality of Virulence-Promoting Function in Rhodococcus equi Virulence Associated Proteins (Vaps)

Rhodococcus equi is a Gram-positive facultative intracellular pathogen associated with life-threatening bronchopneumonial disease in foals. Key to R. equi ’ s intracellular survival in host macrophages is the production of virulence associated proteins (Vaps). Numerous vap genes are found on virulence plasmids isolated from di ﬀ erent species, and the Vaps share a high degree of sequence identity. VapA has been extensively studied, and although vapK and vapN genes from other R. equi virulence plasmids have been shown to be essential for R. equi intracellular survival, their mode of action is less characterised. We, therefore, examined whether VapK and VapN worked mechanistically in the same way as VapA. Indeed, like VapA, VapK and VapN neutralised lysosomal pH and reduced lysosomal hydrolase activity. A loss of VapA and R. equi virulence could be regained by the presence of either VapK or VapN. The acid-neutralisation activity was also observed to a lesser extent with VapB. There was a di ﬀ erential activity across these virulence-promoting Vaps with the most “ acid-neutralising ” activity found with VapN, then VapA and K, and ﬁ nally VapB. These data suggest that VapA production, which is often found in equine infections, can be substituted by VapK and B (produced by plasmids often found in porcine species) or VapN (produced by plasmids often isolated in bovine and human samples). These data imply that the molecular mechanism(s) that VapA uses to neutralise lysosomal acidity should also be seen in VapN and K which will help guide researchers in identifying their precise mode of action and aid the future development of targeted therapeutics.


Introduction
Rhodococcus equi (R. equi) is a soil-dwelling pathogenic actinomycete that causes pulmonary and extrapulmonary pyogranulomatous infections in a variety of animal species [1].R. equi infects various animal hosts, including pigs, sheep, cattle, goats [2], and camelids [3] but is most frequently associated with life-threatening bronchopneumonial disease in foals [4].R. equi is prevalent on horse-breeding farms worldwide and has a significant economic impact on the horse-breeding industry [5].Treatments for R. equi infections involve combi-nation drug therapies with rifampin and macrolides such as clarithromycin [6].These treatments can be protracted and expensive; moreover, they are not always successful, and there is increasing antibiotic resistance [7][8][9][10][11].Additionally, R. equi is increasingly isolated from immunocompromised humans [12][13][14][15][16], and antibiotic-resistant strains of R. equi represent a further threat to human health because of the risk of zoonotic transmission.
Alveolar macrophages phagocytose R. equi that then become enclosed in an intracellular compartment termed the R.equi-containing vacuole (RCV) in which the bacterium survives and begins to multiply eventually resulting in necrosis of the host cell [17].Essential for R. equi survival is a virulence plasmid whose prevalence varies according to host species [18][19][20][21][22][23][24].pVAPA, the first of these plasmids to be discovered, is generally associated with equine R. equi isolates.pVAPB is mostly closely linked to porcine R. equi infections while pVAPN is often found in R. equi in bovine and human isolates.Plasmid swapping (pVAPA for pVAPB) does not affect R. equi virulence in swine or equine macrophages [25].
The Vaps possess N-terminal secretion signals consistent with their export from the bacterial cell giving rise to mature proteins of 130-160 amino acid residues.Sequence alignment of the different Vap proteins shows that they share a high degree of homology in the last 110 or so amino acids with little homology in the preceding 25-40 residues at the N-terminus.The crystal structures of the C-terminal regions, which we refer to here as the core domain, of VapB, VapD, and VapG have been determined by X-ray crystallography [31][32][33] revealing a tightly packed eight-stranded β-barrel.The structures of these three Vaps are closely superimposable.Alphafold [34] indicates that the structures of the other Vaps will be very similar to the crystal structures of VapB, D, and G. Conspicuously absent from the crystal structures of the Vaps are obvious cavities or grooves that might represent a ligand binding site that would provide clues to their function.
What is apparent is that VapA disrupts endolysosomes [35] so that lysosomes are not reformed, and this disruptive activity can be accounted for by VapA's capacity to reduce the acidity of the endolysosomal compartment [36][37][38].Given the high degree of sequence identity across Vaps and the predicted similarities in their structures, we wanted to identify what features VapA possesses that are absent from the Vaps that do not promote R. equi virulence.Further, we wished to examine whether VapK and VapN (the reported equivalents of VapA) work in the same way as VapA, given that properties of VapA, such as the binding to phosphatidic acid [39], were not seen with VapK.
Our results indicate that VapA, K, and N work in a similar manner and that late endocytic compartment acidity is reduced, leading to a loss of lysosomal hydrolytic activity.This activity is also seen to a certain extent with VapB, which has the highest degree of homology to VapA.These data imply that the mechanism(s) that Vaps use to neutralise lysosomal acidity should be seen in VapA, K, and N which will guide researchers in identifying their precise mode of action and aid in the future development of targeted therapeutics.
2.5.Immunofluorescence.NRK or J774A.1 cells were seeded onto coverslips (no.1.5, Scientific Laboratory Supplies) and incubated overnight.If cells were fed purified proteins, then NRK cells were fed protein at 40 μg/ml (unless otherwise stated) and J774A.1 at 5 μg/mL (unless otherwise stated) for 20 h.If J774A.1 were infected with R. equi, then cells were treated as per the infection assay but fixed instead of lysed at appropriate times.Coverslips were rinsed with PBS, fixed in 4% (w/v) formaldehyde (04018-  cells were recorded for all samples.For DQ-BSA experiments, cells were treated as above, but instead of Lyso-Tracker, cells were incubated with 10 μg/ml DQ Red-BSA for 1.5 h; then, medium was changed for DQ-BSA-free medium at 37 °C for 1 h before analysis on the CytoFLEX S (as for LysoTracker).
2.9.Isolation of Lysosomes.Macrophages were incubated ±5 μg/ml VapA for 20 h before isolation of lysosomes as previously described [40].Briefly, macrophages are pulsed for 1 h with colloidal iron-dextran before being chased for 2 h by media without dextran.Cells are then gently broken by nitrogen cavitation, and a postnuclear supernatant (PNS) is generated.The PNS is then passed through an iron-filings column in the presence of a strong magnetic field, retaining the iron-filled lysosomes, and then the lysosomes are eluted from the column by removal of the magnetic field.
2.11.Holotomography.For holotomography, 1 h before the start of imaging, cells were put into a heated CO 2 chamber on the microscope to equilibrate.Three-dimensional (3D) quantitative phase imaging (QPI) and its correlative fluorescence images of live LGP120-GFP NRK cells were obtained using a commercial holotomography instrument (HT-2H, Tomocube), which is based on the Mach-Zehnder interferometry equipped with a digital micromirror device.A coherent monochromatic laser (λ = 532 nm) was divided into two paths, a reference and a sample beam, respectively, using a 2 × 2 single-mode fibre coupler.A 3D RI tomogram was reconstructed from multiple 2D holographic images acquired from 49 illumination conditions, a normal incidence, and 48 azimuthally symmetric directions with a polar angle (64.5 °).The digital micromirror device was used to control the angle of an illumination beam impinging onto the sample.The diffracted beams from the sample were collected using a high-numerical-aperture (NA) objective lens (NA = 1 2, UPLSAP 60XW, Olympus).The off-axis hologram was recorded using a CMOS image sensor (FL3-U3-13Y3MC, FLIR Systems).For 3D epifluorescence imaging, LGP120-GFP was excited using a LED light source (488 nm).Deconvolution of reconstructed 3D fluorescence images was performed using commercial software (AutoQuant X3, Media Cybernetics).The visualization of 3D RI maps and its correlative 3D fluorescence signal with green pseudocolour was performed using commercial software (TomoStudio, Tomocube).
2.12.Surface-Labelling of Bacteria.R. equi strains were grown overnight in BHI medium at 30 °C.Cultures were then diluted 20-fold into fresh BHI medium containing 4 mM D-Alanine azide (Flurochem) and allowed to grow for a further 3 hours at 30 °C.Bacteria were then washed twice with PBS before being resuspended in PBS containing 20 μM FITC-Alkyne (DBCO-PEG 4-5/6-FAM; Jena Bioscience).Samples were rotated at 37 °C, in the dark, for 30 min, and then washed twice with PBS before being used in infection assays.

Immunofluorescence Lysosomal Size Quantification.
ImageJ (v 2.1.0/1.53c) was used to quantify lysosomal sizes.A minimum of 10 images was used per experiment, with n > 3. Images were acquired with a minimum of a 63X objective and a 2.5x optical zoom.Image thresholds were matched to match the immunofluorescence pattern before converting the image into an 8-bit binary image.The watershed function was applied, and the scale was set using the metadata from the image.The analyse particles function was used with size = 0 05 − 10 μm 2 , excluding edges and including holes.

Results
3.1.VapA, VapB, VapK, and VapN Core Regions Have a High Degree of Identity.Previously, we demonstrated that the core region of VapA alone (consisting of the β-barrel domain of VapA) was sufficient to disrupt the lysosomes [35].When full-length Vap sequences are aligned, VapA exhibits the highest degree of identity to VapB (77%) with more modest identity of 61%, to VapK1/2 (VapK1 and K2 differ only by one amino acid) and VapN, and less than 53% for all other Vap proteins.For the predicted core regions, VapA maintains the highest identity to Vap B (86%), but the identity to VapK and VapN increases to 78% and 72%, respectively.A heat map depiction of these analyses suggested that when considering just the core regions of Vaps, VapA, VapB, VapK, and VapN may have a commonality of function (Figure 1), which would be consistent with vap gene ancestry [23].).These data suggest that there is a threshold level of identity to VapA in which proteins perturb lysosomes.

LC3 Degradation in Autolysosomes Is
Inhibited by VapA, VapB, VapK, and VapN.VapA disrupts endolysosomes, and we have previously observed an increase in microtubuleassociated protein light chain 3 (LC3) protein in VapAtreated cells [35].LC3 is commonly used as an autophagosome marker, and its accumulation is an indicator of autolysosomal dysfunction.When NRK cells were treated with the Vaps (A, K, N, and B) which induce lysosomal swelling (as in Figure 2), LC3 levels were found to increase when assessed by both immunofluorescence (Figure 3(a)) and western blotting (Figure 3(b)).No LC3 accumulation was observed in untreated cells or in cells that were treated with VapD (Figure 3).Relative to the control and VapD, the immunofluorescence is clear for VapA, VapB, VapK, and VapN and especially striking for A and N in this assay.Altogether, these data indicate that VapA, B, K, and N disrupt endolysosomes.

3.4.
Vaps A, B, K, and N Added Exogenously Can Promote R. equi Intracellular Survival in Bacteria Lacking vapA.To further test whether Vaps B, K, and N are functionally the equivalent of VapA, mouse macrophages were infected with R. equi harbouring the virulence plasmid pVapA (R. equi strain 103+), without the virulence plasmid (R. equi strain 103-) or with R. equi 103+ lacking vapA (R. equi 103+ ΔvapA) [42].Without pVapA (R. equi 103-), R. equi fails to proliferate intracellularly, and a lack of proliferation is also observed when just vapA is deleted from the virulence plasmid (R. equi 103+ ΔvapA) (Figure 4).Macrophages were then infected with these R. equi strains in the presence of recombinant Vaps in the culture medium.Assessing colony forming units 24 h postinfection, R. equi 103+ bacterial survival was enhanced by the addition of proteins VapA (2.5-fold), B (1.25-fold), and K (1.5-fold) (Figure 4).While 103+ ΔvapA bacteria did not persist under normal conditions, the addition of recombinant VapA protein not only restored virulence but enhanced virulence (1.75-fold) above wild-type bacteria (R. equi 103+), as did the addition of VapB (1.3-fold) and VapK (1.25-fold) (Figure 4).Initially, VapN did not appear to restore virulence, but when macrophage monolayers were observed post-bacterial infection, the majority of the macrophages had detached from the culture wells.Reducing the concentrations of VapN in the experiment from 5 μg/ml to 1 μg/ml showed that VapN can aid bacterial survival (1.5-fold greater than 103+) when vapA is deleted (Figure 4).In addition to vapA, virR and virS (from the virulence plasmid) are required to confer virulence [28], and in all conditions tested, none of the recombinant Vaps could restore virulence in R. equi 103-strains (Figure 4).With all three R. equi strains, recombinant Vaps D, E, G, and L did not aid R. equi intracellular survival (Figure 4).In experiments similar to ours, recombinant VapK protein has previously been shown to restore virulence in a ΔvapA strain, but VapB did not have any effect [39].This latter study was using VapB at 100 nM, and in our study, 5 μg/ml equates to 400 nM.When we used VapB protein at 1 μg/ml (equivalent of 82 nM), we did not see VapB being able to aid R. equi intracellular survival.These data suggest that VapB, VapK, and VapN can substitute for VapA in assisting R. equi survival, but data with VapB and VapN indicated that the ability of a Vap protein to mimic VapA in aiding R. equi to survive intracellularly is dosage dependent.

Vaps Have Variable Activities in Inducing Lysosomal
Dysfunction.To examine whether Vaps A, B, K, and N have 5 Cellular Microbiology different biological activities, we revisited the lysosomal swelling assay, as depicted in Figure 2, using Vaps at a variety of concentrations (Figure 5).At 32 μg/ml, Vaps A, B, K, and N all cause lysosomal swelling to the same extent (Figure 5), as has been seen when proteins were used at 40 μg/ml (Figure 2).However, at lower protein concentrations, a hierarchy of the ability of Vaps A, B, K, and N to cause lysosomal swelling is observed, with VapN biological activity > VapA > VapK > VapB (Figure 5).The activity seen in the lysosomal swelling assay is slightly different to the infection assay, but the latter assay is dependent upon eukaryotic cell survival.Macrophage viability was assessed in the presence of varying concentrations of Vaps.At 4 μg/ml, macrophage viability was reduced to 88%, 78%, 50%, and 60% by Vaps B, A, K, and N, respectively (Figure S2), indicating that VapK and N are detrimental to cells at quite    Cellular Microbiology modest levels.Macrophage viability, therefore, affects the infection assay and explains why VapA and VapB, which do not have a major effect on macrophage viability, are seen to have more significant effects in promoting R. equi survival and why a lower concentration of VapN protein was required in the infection assay to observe R. equi proliferation.
3.6.VapA Reduces the Acidity of the Phagolysosome.We previously stated that VapA could swell lysosomal compartments and that these swollen compartments were acidic, as assessed by using LysoTracker [35].These data differed from observations with virulent R. equi, which were determined to be in nonacidic compartments [36].Lysosomes are made acidic by the action of an ATPase-dependent proton pump, the vacuolar-type ATPase (V-ATPase), which consists of two multi-subunit complexes, the membrane-associated V 0 domain and the peripherally associated V 1 domain.Recently, it has been suggested that VapA prevents incorporation of the V-ATPase (at least the E-subunit of the V 1 domain when overexpressed) into the phagolysosomal compartment and that proton leakage (by membrane permeabilisation) increases phagolysosomal pH, and this aids R. equi survival [37].Our earlier LysoTracker data may be explained by the use of different cell types and the observations that the pH of lysosomes    Cellular Microbiology can vary depending upon the location of the lysosome within the cell [43].Moreover, in some cell types, terminal storage lysosomes may not be acidic [44].While LysoTracker accumulates in compartments where the pH < 6 5, it is not sensitive to small changes in pH and, therefore, does not report the precise lysosomal pH [45].Furthermore, measurements of intracellular pH using fluid-phase reporter probes may run the risk of the probes localising to a compartment other than that occupied by the phagocytosed bacterium, particularly if the late endocytic compartment has been perturbed, as is the case when using VapA.Hence, to directly measure the pH of the R. equi-containing phagosomal compartment, we utilised metabolic labelling and bioorthogonal clickchemistry and grew R. equi strains in the presence of Dalanine azide and then surface labelled the cell wall using dibenzocyclooctyne (DBCO)-fluorescein.Fluorescein is pH sensitive, and using dual-wavelength ratiometric fluorescence measurements [41], we were able to measure the pH of the phagosomal compartment surrounding R. equi bacteria.R. equi strains grown in the presence of D-alanine azide have the same growth kinetics as strains grown in normal growth medium (data not shown) and exhibit unaltered replicative characteristics in mouse macrophages (Figure 6(a)).The pH of the phagosomal compartment was measured in the R. equi strains 103-, 103+, and 103+ ΔvapA.Nonreplicative R. equi (103-) were found in a phagosomal compartment of pH 5.7 (Figure 6(b)).This is very similar to the pH of macrophage lysosomes measured with a fluorescein-dextran probe (pH 5.6; Figure 6(b)).By contrast, R. equi (103+) harbouring the intact virulence plasmid were found in a less acidic environment (pH 6.6; Figure 6(b)).R. equi 103+ ΔvapA were found in a compartment with a pH of 6.1, which was significantly different from that of 103+ strains but not that of 103-strains.When macrophages were fed VapA-pHluorin-2 (a pHsensitive probe which, like fluorescein, can be excited at two wavelengths) [46], lysosomal pH was determined to be 7.1.These data validate previous studies indicating that R. equisecreted VapA can neutralise the pH of the endolysosomal compartment [37].
To examine whether the pH of lysosomal compartments can be neutralised by paralogous Vap proteins, we analysed LysoTracker accumulation and fluorescence in cells incubated with other Vaps, using flow cytometry.Using mouse macrophages and Vaps at 5 μg/ml, cells were gated for live (DAPI exclusion) and single cells (Figure 6(c)), and LysoTracker fluorescence was measured (Figure 6(d)).There was a 185-fold increase in the fluorescence signal compared to control cells that had not been incubated with LysoTracker (Figure 6(e)).The V-ATPase inhibitor bafilomycin-A 1 reduced the fluorescence signal to 3-fold above control cells (Figure 6(e)).Statistically, there was no difference in the LysoTracker fluorescence between the bafilomycin-A 1 treated cells and cells incubated with Vaps A, K, and N (Figures 6(d) and 6(e)).VapB did reduce the accumulation of LysoTracker compared to control LysoTracker cells (2.3 fold reduction), but not to the same extent as VapA, K, and N. VapN showed the greatest reduction in LysoTracker accumulation in cells, then VapK, then VapA, and then VapB.The loss of acidity is not due to cells being incubated with extracellular protein, since VapL (which has no observed effects on lysosomes) did not have any effects on the LysoTracker assay (Figures 6(d) and 6(e)).Together, these data indicate that VapA directly reduces the acidity of phagolysosomes, and this biological activity is also conferred by VapN, VapK, and, to a lesser extent, VapB.
To address whether the reduction in acidity is due to the loss of V-ATPase subunits from the VapA-positive compartment, macrophages were fed VapA or infected with R. equi strains (103-or 103+), and immunofluorescence was used to examine the location of V-ATPase subunits.Neither of the V-ATPase subunits ATP6V 0 d or ATP6V 1 D could be observed to be excluded from lysosomal membranes or R. equi containing phagosomes (Figure 7(a)).We then used a colloidal-iron dextran to purify lysosomes [40] from macrophages that had been incubated in the presence or absence of VapA.When analysing the purified lysosomes, from both conditions, there was an enrichment of LAMP1 and various subunits of the V-ATPase V 0 and V 1 complexes compared to the postnuclear supernatants (PNS), including ATP6V 1 E which had been reported to be excluded by VapA [37] (Figure 7(b)).Additionally, we indirectly measured the V-ATPase activity in intact lysosomes (-/+ VapA) using a general ATPase assay and attributing V-ATPase specific activity to the proportion of the ATPase activity which was inhibited by the V-ATPase inhibitor concanamycin-A.In the presence of VapA, there was no change in the proportion of ATPase activity which could be inhibited by concanamycin-A (Figure 7(c)).Together, our data indicate that VapA does not have a direct effect on the recruitment of V-ATPase subunits (of the antibodies that we could test) to the lysosome or V-ATPase activity.

Vap Disruption of Lysosomal Acidity Correlates with
Disruption to Lysosomal Hydrolase Activity.With a loss of acidic compartments, we expected that there would be a concomitant loss of lysosomal hydrolase activity.Indeed, using a similar flow cytometry gating strategy as shown in Figure 6, macrophages were incubated with Vaps and then DQ-BSA (dye-quenched bovine serum albumin) before being analysed by flow cytometry.DQ-BSA is highly fluorescently labelled, and the fluorescence is quenched.Upon endocytosis and delivery to endolysosomes, digestion of the BSA occurs, and there is an increase in fluorescence as dequenching occurs.Thus, fluorescence is an indicator of proteolytic activity.When macrophages were incubated with DQ-BSA, there was a 23-fold increase in fluorescence compared to cells which had not been incubated with DQ-BSA cells (Figured 8(a) and 8(b)).The fluorescence in control cells was reduced by 20-fold when cells were incubated with bafilomycin-A 1 , indicating that hydrolase activity is pH dependent (Figures 8(a) and 8(b)).Similarly, when macrophages were fed Vaps A, B, K, or N, there was a loss of the fluorescent signal indicating a lack of DQ-BSA digestion (Figures 8(a) and 8(b)).The DQ-BSA data very much matched the LysoTracker data, with the effect of VapB being more modest compared to Vap A, K, and N and no effect with VapL.These data show that Vaps secreted from R. equi enable R. equi to survive longer in the phagolysosome by disrupting acidity and, thus, the activity of hydrolases.

Cellular Microbiology
We then applied holotomography to cells which have been incubated with VapA.Holotomography measures the refractive indices of samples and offers a means to probe, noninvasively and quantitatively, the structural and chemical information of cells, including morphology, membrane dynamics, and dry mass of the cell [47].The refractive index can be correlated to density.Over 24 h, LAMP-1-GFP positive NRK cells were imaged in the presence of VapA (Figure 8(c)).At 0 h, the GFP signal did not correlate with any particular density compared to the surrounding cell, but 24 h post-VapA addition to the cells, the swollen LAMP1-GFP compartments correlated with areas of increased density (Figure 8(c)).As lysosomal density increased, there was an overall increase in cellular volume and dry mass (Figure 8(d)).At 48 h, the cell begins to lose mass as the cell begins to apoptose (Figure 8(d)).Livecell imaging of LAMP1-GFP cells shows that VapA-swollen LAMP1 compartments remain fusogenic at least 9 h post-VapA addition (Figure S4, Movie S1).Together, these data, along with the accumulation of LC3 (Figure 3), indicate that VapA reduces the acidity of late endocytic compartments, and there is a reduction in proteolytic activity, thereby causing an accumulation of undigested material.The endolysosomes are still fusogenic (Movie S1), and in NRK cells, material from endocytic and autophagy pathways continues to be delivered to the lysosomal compartment up to 36 h after the addition of VapA, as evidenced by the increase in cell dry mass (Figure 8(c)).The loss of endolysosomal acidity and proteolytic activity is also observed with Vaps N and K, and to a lesser extent Vap B.

Discussion
All these data indicate that there is a good correlation between the ability of certain Vaps to disrupt late endocytic pathway acidity, reduce lysosomal proteolytic activity, and then accumulate undigested material resulting in large endolysosomes which fail to condense and reform terminal lysosomes.While R. equi may be adapted to survive short term in a phagolysosomal environment [48], bacterial replication will be dependent upon the reduction of hydrolase activity.Our data show that in an environment with VapA-induced alkalynisation of the compartment, the compartment remains fusogenic, thereby maintaining the delivery of nutrients to the phagolysosomal compartment that can be utilised by R. equi.
The ability of any individual Vap to promote R. equi survival will be dependent upon its expression levels from the various virulence plasmids, and while VapB may not be as "potent" as VapA or VapN when compared side by side, pVAPB harbours vapB, vapK1, and vapK2 all of which can produce Vaps to disrupt endolysosomal function.Indeed, vapK1 and vapK2 are required for optimal growth of R. equi strains carrying a pVAPB plasmid and can complement growth-attenuated ΔvapA strains [30].Together, our data indicate that vapN in pVAPN is the equivalent of vapA in pVAPA, and vapB, vapK1, and vapK2 in pVAPB can all contribute to R. equi survival.In R. equi pVAPB strains, vapK1 and vapK2 are likely to be more critical than vapB in promoting intracellular survival.While distinct virulence plasmids may be associated with particular animal sources, the plasmid type (pVAPA, pVAPB, and pVAPN) does not determine the ability of R. equi to replicate inside a particular species host cell [25], and thus, any plasmid which expresses VapA or an equivalent protein such as VapK or VapN is of importance.
Importantly, VapA, K and N, and, to a certain extent, VapB all appear to have the same mechanism of action, in that they reduce late endocytic acidity and lysosomal hydrolase activity.This appears to contradict recent data reporting that VapN did not alter lysosomal hydrolase activity [38].This might be explained by differing flow cytometry analyses.In the earlier study, scatter alone was used to assess viability in the presence of 50 μg/ml Vap proteins.In our hands, this protocol led to cell death in approximately 60% of the mouse macrophages.Additionally, while we had previously shown that the N-terminal region of VapA is not critical for VapA function [35], we have not evaluated whether the N-terminal regions in other Vap-proteins provide a 12 Cellular Microbiology regulatory function.The more recent study by Hansen et al. [38] used full-length proteins (without the signal sequence), whereas our study here has only used the β-barrel core region.It has been demonstrated that VapA binds to phosphatidic acid [39] and can permeabilise artificial membranes allowing for proton leakage [37].However, VapB bound phosphatidic acid weakly, and VapK not at all [39].Since we find VapA has similar lysosomal disruptive properties to VapK, and VapB has limited lysosomal disruptive properties, the binding of VapA to phosphatidic acid may be an artefactual property of VapA.The modest permeabilisation of liposomes with VapA [48] is not recapitulated to the same extent with VapB or VapN, suggesting that this observation may also be an effect which is unrelated to the true biological activity of VapA.Furthermore, VapA is found primarily localising to intracellular membranes [37] and not the phagolysosomal limiting membrane (which we also see in Figure S5), and so it is unclear how membrane permeabilisation would occur if VapA is bound to lumenal-vesicular membranes.
The lysosomal pH gradient is generated by the action of a V-type ATPase (V-ATPase) which uses ATP to pump protons into the lysosomal lumen.The movement of protons would create a membrane potential which would restrain the activity of the V-ATPase if left uncompensated [49].Thus, there is a counterion conductance pathway which neutralises the charge imbalance.The counterion conductance pathway may be satisfied by the influx of cytosolic anions such as Cl -into the lysosome [50,51], Cl -/H + exchangers such as ClC-7 [52], or efflux of cations [53,54].The role of Cl -in maintaining lysosomal pH by complementing the V-ATPase is controversial [51][52][53], and more recent research has demonstrated that ClC-7 does not contribute to lysosomal pH, but instead, Cl -is used to activate lumenal hydrolases [55].It is interesting to speculate that if VapA was to disrupt Cl -entry, then potentially, this could either disrupt the activity of the V-ATPase (if Cl -is acting as a counterion) or lead to a loss of hydrolase activity, causing undigested material to accumulate and lysosomal size to increase.A doubling of lysosomal diameter equates to an 8-fold increase in volume which could then account for changes seen in lumenal pH if the number of H + ions remains the same.
In addition to Cl -/H + exchangers, lysosomes continually leak H + equivalents.Proton leakage can occur in many ways including the cotransport with organic solutes such as amino acids [56,57] or through a proton-activated proton channel such as TMEM175 [58,59].If the V-ATPase is not directly inhibited by VapA (our study here and [37]), then there has to be a change in the concentration of protons in the lysosomal lumen to account for the observed change in pH.VapA may permeabilise the membrane, but as mentioned, this is seen to a much lesser extent with VapB or VapN [38], whereas in our study, VapN has the highest lysosomal alkalinisation activity.VapA could accelerate protoncoupled solute export, though the removal of useable solutes to a phagosomal-membrane contained R. equi might be disadvantageous to the bacterium.VapA may accelerate proton leak through a channel such as TMEM175.Alternatively, VapA may change one of the numerous ion or solute channels in such a way as to alter the osmolarity of the lysosome, creating an increase in volume, which as mentioned would reduce the proton concentration and reduce acidity.It, therefore, remains to be seen as to how, mechanistically, VapA disrupts the proton gradient in phagolysosomes, but our study demonstrates that any biological activity of VapA should also be observed in VapN and VapK, which will help eliminate potential spurious observations.
To date, while most research has concentrated on the role of VapA in promoting R. equi virulence, our study here demonstrates that VapK and VapN have the same lysosomal disrupting ability.R. equi only needs to produce one of these proteins to aid in its intracellular survival, and thus, VapA should not be the sole focus in understanding how lysosomes are disrupted.Given the emergence of antibioticresistant R. equi, a biologic that can disrupt Vap A, K, and N functions is likely to prove useful in treating R. equi infections.Since Vap A, K, and N appear to work mechanistically in the same manner, then a single therapeutic to VapA, K, or N should be beneficial in treating an R. equi infection independent of the virulence plasmid harboured.

3. 2 .
VapA, VapB, VapK, and VapN All Induce Enlargement of Lysosomes.A good indicator of Vap function is a simple assay where purified recombinant Vaps are exogenously fed to eukaryotic cells and immunofluorescence is used to immunolabel the lysosome.The size of the lysosome can then be measured.We have shown that the core region of VapA protein can disrupt endolysosomes giving rise to swollen LGP120-positive (the rat equivalent of human LAMP1) 4 Cellular Microbiology endolysosomes [35].Based on the sequence identity of the core regions of Vaps, we predicted that Vaps B, K, and N would cause similar disruption to endocytic compartments.Indeed, swollen LGP120 positive structures, on average twice the diameter of normal lysosomes, were observed when the core region of Vaps B, K, and N were fed to normal rat kidney (NRK) cells (Figures 2(a) and 2(b)).In contrast, no change in the size of lysosomes was observed in cells fed with Vaps D, E, G, and L (Figures 2(a) and 2(b)

Figure 1 :
Figure 1: Vaps A, B, K, and N have a high degree of identity.Heat map representation of pairwise protein identities between full-length Vap proteins (a) and the β-barrel core region (b).Sequences used are in Figure S1.Red box highlights the increase in identity between Vaps A, B, K, and N compared to full-length proteins.

Figure 2 :
Figure 2: Vap-induced lysosomal swelling is not unique to VapA.(a).NRK cells were incubated without (Ctrl) or with 40 μg/ml recombinant Vaps (denoted by alphabetic character) for 20 h.Cells were fixed and the lysosomes immunolabelled with an anti-LGP120 monoclonal antibody, followed by an Alexa Fluor 555-labelled secondary antibody (red).Cell nuclei were stained with DAPI (blue).Scale bars 10 μm.(b).The size of lysosomes in the images was measured using automatic thresholding and quantification with ImageJ.Data are from 30 images in total from three separate experiments (10 images per experiment).Data are mean ± SEM (n = 3).* * * * p < 0 001 vs. control.

Figure 3 :
Figure 3: VapA, VapB, VapK, and VapN cause accumulation of LC3 in NRK cells.(a) NRK cells were incubated without (Ctrl) or with 40 μg/ml recombinant Vaps (denoted by alphabetic character) for 20 h.Cells were fixed and immunolabelled for LGP110 (green) and LC3 (red) before confocal imaging.Scale bars 10 μm.(b) NRK cells were incubated with protein as in (a) before lysing, and equal amounts of protein were immunoblotted for ɑ-tubulin or total LC3 (I and II).

Figure 5 :
Figure5: Vaps A, B, K, and N vary in their ability to promote lysosome dysfunction.NRK cells were incubated with or without varying concentrations of Vaps A, B, K, and N for 24 h.Cells were fixed and the lysosomes immunolabelled with an anti-LGP120 monoclonal antibody, followed by an Alexa Fluor 555-labelled secondary antibody.The lysosome sizes of images were quantified using automatic thresholding and quantification with ImageJ.Data are from the analyses of 10 images per condition (n = 1).Control lysosomal size is indicated by the dashed horizontal line.