PmrC (EptA) and CptA Negatively Affect Outer Membrane Vesicle Production in Citrobacter rodentium

Although OMVs secreted by Gram-negative bacteria fulfill multiple functions, the molecular mechanism of OMV biogenesis remains ill defined. Our group has previously shown that PmrC (also known as EptA) and CptA maintain OM integrity and provide resistance to iron toxicity and antibiotics in the murine pathogen Citrobacter rodentium. In several enteric bacteria, these proteins modify the lipid A and core regions of lipopolysaccharide with phosphoethanolamine moieties. Here, we show that these proteins also repress OMV production in response to environmental iron in C. rodentium. These data support the emerging understanding that lipopolysaccharide modifications are important regulators of OMV biogenesis in Gram-negative bacteria.

arn operon that is responsible for L-Ara4N addition to lipid A (31). We previously showed that the PmrAB-regulated genes pmrC (eptA) and cptA are important in maintaining OM integrity. Strains lacking pmrAB, pmrC (eptA), or cptA exhibited increased susceptibility to iron toxicity and antibiotics and increased OM permeability (31).
Environmental iron is an important contributor to bacterial oxidative stress. In the Fenton reaction, hydrogen peroxide is converted to a more reactive hydroxyl radical through the oxidation of Fe 2ϩ into Fe 3ϩ (35). These reactive hydroxyl radicals are highly damaging to lipid membranes, proteins, and nucleic acids (35). In this study, we explored the hypothesis that environmental iron and the iron-induced genes pmrC (eptA) and cptA affect OMV biogenesis in C. rodentium. We showed that production of OMVs is greatly increased in C. rodentium strains lacking the PmrAB TCS or the pmrC (eptA) and cptA genes. These results may further relate LPS modifications to OMV biogenesis and reveal a new role for PmrC (EptA) and CptA in C. rodentium.

RESULTS
The PmrAB TCS negatively affects OMV production. C. rodentium and related Enterobacteriaceae use the PmrAB TCS to sense and respond to environmental iron and modify the OM (Fig. 1A) (36). The role of PmrAB in OMV production was investigated by measuring the amounts of OMVs produced by the C. rodentium ΔpmrAB strain in the presence of FeCl 2 (PmrAB-activating conditions). OMVs were isolated as described in the Materials and Methods. OMV production was measured using the detergentcompatible (DC) protein assay, which measures the total protein content of solubilized OMVs, and by the fluorescent probe FM1-43, which measures the lipid content of biological material with intact membranes. In a comparison of OMV production between the wild-type and ΔpmrAB strains grown at 50 M FeCl 2, the ΔpmrAB strain exhibited a modest but significant increase in vesicle production, as measured by Under conditions of high Fe 3ϩ , the cytoplasmic PmrA response regulator becomes phosphorylated and active, resulting in differential expression of several genes, including pmrC (eptA) and cptA (31). (B) Schema showing PmrC (EptA) and CptA-mediated modifications of LPS. In E. coli, PmrC (EptA) and CptA add pEtN groups to the lipid A and core regions of LPS, respectively (36). Percentages refer to the similarity of the PmrA, PmrC (EptA), and CptA proteins between E. coli and C. rodentium. protein and lipid content (Fig. 2). Complementation-mediated overexpression of pmrAB resulted in drastic decreases of OMV production below wild-type levels (Fig. 2). Similar trends were observed with C. rodentium grown at various concentrations of FeCl 2 (see Fig. S1 in the supplemental material). Collectively, these results indicate that the PmrAB TCS negatively affects OMV production in the presence of environmental iron.
PmrC (EptA) and CptA negatively affect OMV production. In C. rodentium and related species, the PmrAB TCS directly regulates transcription of pmrC and cptA, which are important in maintaining OM integrity and providing resistance to iron and antibiotics ( Fig. 1A) (31,37). OMV production was assessed in C. rodentium ΔpmrC (ΔeptA) and ΔcptA strains grown at various concentrations of FeCl 2 . At each concentration of FeCl 2 , the ΔpmrC (ΔeptA) ΔcptA strain displayed an approximate 2-fold increase in protein and lipid concentrations of OMV preparations, indicating that these genes negatively affect OMV formation ( Fig. 3 and Fig. S2). The ΔpmrC (ΔeptA) and ΔcptA strains each exhibited a 1.5-fold increase in OMV production relative to that of the wild type, indicating that these genes both similarly impact vesicle formation. Overexpression of either pmrC (eptA) or cptA in both wild-type and knockout backgrounds reduced vesicle production below the levels of wild-type and empty vector controls ( Fig. 3 and Fig. S2). Together, these results show that OMV production is negatively influenced by the PmrAB-regulated proteins PmrC (EptA) and CptA in response to environmental iron in C. rodentium.
Environmental iron influences OMV production. Since iron activates the PmrAB TCS and PmrAB-controlled genes, including pmrC (eptA) and cptA (31), we investigated the effects of various FeCl 2 concentrations on OMV production in C. rodentium. OMV formation was assessed in various C. rodentium strains grown at different concentrations of FeCl 2 . Concentrations up to 75 M FeCl 2 were used, since the presence of iron at concentrations of 100 M or higher inhibited bacterial growth due to iron toxicity (31). Unexpectedly, protein and lipid concentrations of OMV samples increased in each C. rodentium strain when grown in the presence of 25 and 50 M FeCl 2 ( Fig. 4A and B). Increasing the concentration of FeCl 2 to 75 M did not greatly increase OMV production (Fig. 4B). Substitution of FeCl 2 by FeSO 4 or FeCl 3 resulted in similar increases in OMV production in wild-type C. rodentium (Fig. S3). These results indicate that environmental iron increases OMV production in C. rodentium. Iron is known to cause oxidative stress by facilitating the production of damaging hydroxyl radicals (35). OMVs isolated from the C. rodentium wild-type, ΔpmrAB, and ΔpmrAB pWSKpmrAB strains grown in the presence of 50 M FeCl 2 were quantified by measuring protein and lipid contents (A) Protein content of OMVs was measured using the DC protein assay. Protein concentration was calculated using BSA as a standard. (B) The lipid content of OMV samples was determined by using the lipophilic dye FM 1-43. Cultures were grown to mid-log phase and normalized to an OD 595 of 0.5 prior to OMV isolation. Values shown are the means Ϯ standard error from three independent experiments. Significance was assessed relative to wild-type using one-way ANOVA; ****, P Յ 0.0001; ***, P Յ 0.001. Furthermore, increased OMV production has been associated with oxidative stress (18). In order to determine whether iron-induced oxidative stress was facilitating OMV formation, we added glutathione to N-minimal medium containing various concentrations of iron and measured OMV production. Glutathione is an antioxidant that can neutralize reactive oxygen species. We therefore grew C. rodentium in the presence of glutathione to neutralize reactive oxygen species to determine if the influence of iron on OMV production is dependent on reactive oxygen species or not.
We found that at each concentration of iron, the presence of 2 mM glutathione reduced OMV formation in various strains ( Fig. 4C and D; Fig. S4). This suggests that the iron-mediated increase in OMV formation was due to an increase in reactive oxygen species. In addition, the differences in OMV production between wild-type and ΔpmrAB and ΔpmrC (ΔeptA) ΔcptA strains still remained in the presence of glutathione, suggesting that the iron-mediated and the PmrC (EptA) and CptA-mediated responses to OMV formation are likely distinct from one another.
OMV quantification and size distribution by transmission electron microscopy. To determine whether the increased OMV production by C. rodentium in the absence of pmrAB, pmrC (eptA), and cptA strains is due to either increased number of OMVs or increased vesicle size, OMV samples isolated from cells grown in the presence of 50 M FeCl 2 were analyzed by transmission electron microscopy (TEM) (Fig. S5). Consistent with the above data, TEM micrographs revealed an increase in the total number of OMVs produced by strains lacking pmrAB, pmrC (eptA), or cptA in comparison to that produced by the wild type (Fig. 5). Complementation-mediated overexpression of pmrAB and of both pmrC (eptA) and cptA resulted in a decrease in terms of the numbers of vesicles produced in comparison to that produced by their respective genetic backgrounds (Fig. 5). Furthermore, images of whole cells showed a greater number of vesicles budding from the membranes of the ΔpmrAB and ΔpmrC (ΔeptA) ΔcptA strains in comparison to those of the wild-type and complemented strains (Fig. 6). The size distribution of OMVs isolated from each C. rodentium was consistent, as most OMVs were found to be between 20 to 100 nm (Fig. S6). Collectively, these data confirm that the number of OMVs produced by C. rodentium is negatively regulated by the PmrAB TCS and the PmrC (EptA) and CptA proteins. cptA, wild-type, and complemented strains were grown in the presence of 50 M FeCl 2 to mid-log phase and normalized to an optical density at 595 nm (OD 595 ) of 0.5. A "ϩ" indicates that a strain was transformed with pWSK129 or pACYC184. OMVs were isolated and quantified by measuring protein and lipid content. (A) Protein content of OMVs was measured using the DC assay. Protein concentration was calculated using BSA as a standard. (B) Lipid content of OMV samples was determined by using the lipophilic dye FM 1-43. Values shown are the means Ϯ standard error from three independent experiments. Significance was assessed relative to the wild type using a one-way analysis of variance (ANOVA); ****, P Յ 0.0001.

C. rodentium
OMVs contain the periplasmic ␤-lactamase. To assess the biological activity of C. rodentium OMVs, the presence of the chromosomally encoded periplasmic ␤-lactamase (ROD_12321) in OMV preparations was investigated. ␤-Lactamase activity from OMVs isolated from the C. rodentium wild type, ΔpmrAB, ΔpmrC (ΔeptA) ΔcptA, ΔpmrC (ΔeptA), and ΔcptA strains grown in the presence of 50 M FeCl 2 was measured using the chromogenic substrate nitrocefin (31). Incubation of equal volumes of solubilized OMVs with nitrocefin resulted in substrate cleavage, indicating the presence of ␤-lactamase associated with C. rodentium OMVs. In comparison to the wild type, solubilized OMV preparations from the ΔpmrAB strain and strains lacking either pmrC (eptA) or cptA resulted in increased ␤-lactamase activity over 20 min of incubation with nitrocefin (Fig. 7). Given that our sample preparation involved resuspension of OMV with equal volumes, these differences support our earlier findings that OMV production is increased in the absence of pmrAB, pmrC (eptA), and cptA. To ensure that differences in activity were not due to changes in ␤-lactamase expression, periplasm was isolated from each strain, and ␤-lactamase activity was measured. Using equal volumes of periplasm, no differences in ␤-lactamase activity were observed between strains, suggesting that there are no differences in cellular expression of ␤-lactamase between C. rodentium strains (Fig. S7A). Furthermore, to determine whether there was selective enrichment of ␤-lactamase in OMVs between strains, vesicles isolated from each strain were normalized to the same concentration of total protein. There was no difference in Wild-type C. rodentium strains were grown in various concentrations of FeCl 2 , and OMVs were isolated and quantified. (C and D) OMVs were isolated from various C. rodentium strains grown in 50 M FeCl 2 with or without 2 mM glutathione. A "ϩ" indicates that a strain was grown with glutathione. Significance was compared to each strain grown without glutathione (a) and between the wild type grown with glutathione and the mutant strains grown with glutathione (b). Significance between wild-type and mutant strains grown without glutathione is not shown in these graphs. (A and C) Total protein content of OMVs was measured using the DC assay. Protein concentration was calculated using BSA as a standard. (B and D) Lipid content of OMV samples was determined by using the lipophilic dye FM 1-43. Values shown are the means Ϯ standard error from three independent experiments. *, P Յ 0.05; **, P Յ 0.01; ***, P Յ 0.001; ns, not significant.
␤-lactamase between normalized OMVs isolated from each strain, indicating that the same relative amounts of ␤-lactamase are loaded into OMVs in each C. rodentium strain (Fig. S7B).
C. rodentium OMVs contain the outer membrane protease CroP. To further assess the biological activity of C. rodentium OMVs, the presence of CroP associated with OMVs was measured. The relative amounts of CroP associated with equal volumes of OMVs from C. rodentium strains grown in the presence of 50 M FeCl 2 was determined by Western blotting. As estimated by measuring the intensity of the bands, the amount of CroP was approximately 2-to 3-fold higher in equal volumes of OMVs isolated from strains lacking pmrAB, pmrC (eptA), and cptA in comparison to that of the wild type, reflecting the higher number of OMVs produced by these mutated strains (Fig. 8A). Activity of CroP in OMV preparations was measured by incubating equal volumes of OMVs isolated from the C. rodentium strains with the second complement component (C2) fluorescence resonance energy transfer (FRET) substrate (38). In comparison to that of the wild type, equal volumes of OMVs isolated from the ΔpmrAB, ΔpmrC (ΔeptA) ΔcptA, ΔpmrC (ΔeptA), and ΔcptA strains resulted in a faster increase in fluorescence over 20 min of incubation, indicating increased CroP activity (Fig. 8B). OMVs isolated from the C. rodentium ΔcroP strain produced minimal fluorescence, indicating that most of the substrate cleavage is mediated by the CroP protease (Fig. 8B). Furthermore, we confirmed that these differences were due to increased OMV production and not changes in CroP expression or selective enrichment of OMVs, since CroP activity in whole cells and in OMVs normalized to the same protein concentration yielded similar activity curves across different C. rodentium strains (Fig. S8). Together, these results show that OMV preparations from C. rodentium contain biologically active CroP. Consistent with the TEM results, these data confirm that pmrAB, pmrC (eptA), and cptA negatively influence OMV production in C. rodentium.

DISCUSSION
The OM of Gram-negative bacteria is constantly remodeled in response to changing environments sensed by TCS. Covalent modifications of LPS are viewed as an adaptive response to acidic pH, presence of AMPs, and changing concentrations of divalent cations (23). Several enteric bacteria modify the lipid A and core regions of LPS with pEtN modifications. These modifications are primarily catalyzed by the pEtN transferases PmrC (EptA) and CptA, the expression of which is regulated by the PmrAB TCS. Previously, we have shown that PmrC (EptA) and CptA confer protection against iron toxicity and contribute to the maintenance of OM integrity in C. rodentium (31). In the present study, we provide evidence that these proteins negatively affect OMV production in C. rodentium. In the absence of pmrAB, pmrC (eptA), or cptA, the production of OMVs is markedly enhanced during the adaptive response to environmental iron  ( Fig. 2, 3, 5 and 8). These data support the emerging hypothesis that LPS remodeling impacts OMV biogenesis by either enhancing or reducing OMV production in response to changing environmental cues (21,32).
Iron-induced oxidative stress is one form of envelope stress. Iron induces the formation of reactive oxygen species that damage the envelope of Gram-negative bacteria (35). In the C. rodentium wild type, we found that OMV production is increased when cells are grown in the presence of up to 50 M FeCl 2 (Fig. 4). Our data also suggest that this iron-mediated induction of OMV formation is due to increased production of reactive oxygen species. By adding exogenous glutathione, which is an antioxidant, we found that there was reduced OMV formation at 50 M FeCl 2 (Fig. 4). Along the murine digestive tract, the concentration of free soluble iron has been estimated to be approximately 50 M (39). Thus, during colonization of the murine gastrointestinal tract, C. rodentium is likely to encounter concentrations of iron that may induce the formation of OMVs in vivo. Increasing the concentration of FeCl 2 to 75 M did not yield a subsequent increase in OMV production for wild-type C. rodentium (Fig.  4). Further increasing the concentration of FeCl 2 was toxic to C. rodentium (data not shown). To our knowledge, the present study is the first that shows the iron-mediated increase in OMV production. These findings are counterintuitive, considering that pmrC (eptA) and cptA, which negatively affect OMV production, are transcriptionally activated in the presence of iron. However, our data suggest that the PmrC (EptA) and CptAmediated influence on OMV production is independent from the iron-mediated effects of oxidative stress, since differences in OMV production between wild-type and ΔpmrC (ΔeptA) ΔcptA strains were still observed when grown in the presence of the antioxidant glutathione ( Fig. 4C and D).
In this study, we demonstrate that OMV production is enhanced in C. rodentium in the absence of pmrAB, pmrC (eptA), and cptA (Fig. 2, 3, 5 and 8). Complementationmediated overexpression with plasmids containing pmrAB, pmrC (eptA), and cptA drastically reduced OMV production below wild-type levels ( Fig. 2 and 3). We had previously showed that complementation of the ΔpmrC (ΔeptA) ΔcptA strain with either pWSKpmrC/eptA or pWSKcptA increases resistance to iron toxicity and antibiotics (31). Collectively, these data indicate that pmrC (eptA) and cptA are responsible for maintaining OM integrity and negatively affecting OMV production, suggesting that both processes are required for C. rodentium resistance to iron toxicity. Given that the PmrAB TCS directly regulates pmrC (eptA) and cptA expression (31), our data also support the notion that the increased OMV production observed in the C. rodentium ΔpmrAB strain is due to the downregulation of the pmrC (eptA) and cptA genes. Although it remains possible that other PmrAB-regulated genes contribute to the increase in OMV production, most of the observed effects are likely mediated by pmrC (eptA) and cptA, since no significant difference in OMV production was observed when comparing the ΔpmrC (ΔeptA) ΔcptA and ΔpmrAB strains.
Our group has previously established the role of PmrC (EptA) and CptA in C. rodentium (31). In other related enteric bacteria, homologs of these proteins function as pEtN transferases. In C. rodentium, a BLAST search revealed that PmrC (EptA) is 92% similar to its homolog in E. coli K-12 MG1655 and 91% similar to its homolog in Salmonella enterica serovar Typhimurium LT2 based on amino acid similarity scores. CptA in C. rodentium is 94% similar to its homolog in E. coli K-12 MG1655 and is 93% similar to its homolog in S. Typhimurium LT2 (40)(41)(42). The large amount of amino acid similarity shared between these bacteria suggests that PmrC (EptA) and CptA function similarly as pEtN transferases. However, structural analysis has yet to be performed to confirm that PmrC (EptA) and CptA definitively modify LPS with pEtN moieties.
To assess putative biological functions of C. rodentium OMVs, we tested for the presence of the periplasmic enzyme ␤-lactamase and the OM protease CroP, associated with OMV preparations. Activity assays showed that both enzymes were associated with OMV preparations isolated from the wild-type, ΔpmrAB and ΔpmrC (ΔeptA) ΔcptA strains ( Fig. 7 and 8). Consistent with the higher numbers of OMVs produced by the deletion mutants ( Fig. 4 and 5), OMVs resuspended with the same volumes isolated from the ΔpmrAB and ΔpmrC (ΔeptA) ΔcptA strains exhibited higher ␤-lactamase and CroP enzymatic activities than an equal volume of OMVs isolated from the wild-type strain.
In a recent study, Elhenawy et al. showed that the PhoPQ-regulated PagL lipid A deacylase enhances OMV formation in S. enterica (32). The C. rodentium genome does not contain the pagL gene; however, it is possible that other PhoPQ-regulated LPSmodifying enzymes fulfill a similar role in enhancing OMV production. In the current study, we show that PmrC (EptA) and CptA repress OMV formation in C. rodentium. Together, these studies might indicate that different LPS modifications have the ability to either enhance or repress OMV production.
Confirming the functional role of CptA and PmrC (EptA) in C. rodentium will help us understand how these proteins affect OMV production and/or membrane integrity in the presence of iron induced-stress. If these proteins are indeed pEtN transferases, as is the case in E. coli and Salmonella enterica, the decrease in OMV production may be a response to maintain pEtN-modified LPS in the OM, and, in turn, strengthen OM integrity (21,29,30,40,43). Furthermore, the role of CptA and PmrC (EptA) as pEtN transferases that repress OMV production is consistent with the previously proposed model for regulation of membrane lipid (21,30,32). As pEtN is added to LPS, this may minimize membrane curvature and subsequent OMV formation. Together, our data are the first to show that PmrC (EptA) and CptA effect OMV production induced by iron-dependent environmental stress.

MATERIALS AND METHODS
Bacterial strains, growth conditions, and reagents. All bacterial strains and plasmids used in this study are listed in Table 1. Bacteria were grown overnight in Luria-Bertani (LB) broth at 37°C with aeration. Bacteria were diluted 1:100 in N-minimal medium [50 mM Bis-Tris (pH 7.5), 7.5 mM (NH 4 )SO 4 , 5 mM KCl, 0.5 mM K 2 SO 4 , 0.5 mM KH 2 PO 4 , 38 mM glycerol, and 0.1% (wt/vol) Casamino Acids] supplemented with 0.2% glucose, 20 M MgCl 2 , and various concentrations of FeCl 2 and glutathione, as indicated. Cultures were grown statically for 48 h at 28°C, leading to an optical density at 595 nm (OD 595 ) of approximately 0.5. When appropriate for plasmid selection, kanamycin (Kan; 50 g/ml) and chloramphenicol (Cm; 35 g/ml) were added to the medium.
OMV protein quantification. Protein concentrations of OMV preparations were determined by using the detergent-compatible (DC) protein assay (Bio-Rad) (44). OMVs were solubilized with 0.1% Triton X-100 (4°C, 15 min), and the DC protein assay was performed according to the manufacturer's instructions. The OD 750 was measured using a Hitachi U-2010 spectrophotometer. Protein concentrations were determined using bovine serum albumin (BSA) as a standard. In experiments in which OMVs were normalized to the same protein concentrations, all purified OMV samples were normalized to a protein concentration of 1.5 mg/ml. OMV lipid quantification. OMV samples were quantified using the lipophilic dye FM 1-43 [N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino) styryl) pyridinium dibromide; Molecular Probes, Life Technologies], which provides an indirect measurement of lipid content, as shown previously (45). OMV samples were diluted 1:55 in PBS and transferred to a quartz cuvette. The FM 1-43 dye was added at a final concentration of 4.5 M. Following excitation at 479 nm, fluorescence emission at 600 nm was measured on a Varian Cary Eclipse fluorescence spectrophotometer, using 5-nm slit widths for both excitation and emission.
TEM. For bacterial cell imaging, culture aliquots were transferred onto carbon-coated copper grids and grown for an additional 24 h. Grids were fixed with 2.5% glutaraldehyde diluted in 50 mM sodium cacodylate (pH 7.2) and washed 3 times with 3% saccharose diluted in the same buffer. For OMV imaging, purified OMV samples were diluted 1:20 with Dulbecco's modified Eagle medium (DMEM), and 70 l aliquots were laid onto carbon-coated copper grids by ultracentrifugation (Airfuge; 20,000 lb/in 2 , 5 min). Excess liquid was discarded from the grids and samples were negatively stained with phosphotungstic acid (3% [wt/vol]) for 5 min and dried with filter paper. Specimens were examined blinded under a microscope (H-7100; Hitachi) operated at an accelerating voltage of 75 kV and at magnifications of ϫ40,000 for cells and ϫ30,000 for OMVs. The size and number of OMVs per micrograph were determined blinded using the Macnification software (Orbicular).
Nitrocefin assays. The activity of ␤-lactamase in OMVs and periplasm was determined by measuring cleavage of the chromogenic ␤-lactam nitrocefin (31). OMVs were solubilized with 0.1% Triton X-100 for 15 min at 4°C. Enzymatic assays were performed at 22°C in the presence of 43 M nitrocefin, and ␤-lactamase activity was monitored by measuring the OD 486 for 45 min, using a Powerwave X340 microplate reader (Bio-Tek instruments). Data were fitted to a Michaelis-Menten nonlinear regression curve using GraphPad Prism 7 software. Periplasmic protein extraction. C. rodentium cells were grown to an OD 595 of 0.5 in N-minimal medium and centrifuged (12,000 ϫ g, 4°C, 20 min). Cell pellets were resuspended with periplasting lysis buffer (1 M Tris HCl, 20% sucrose, 1 mM EDTA, 40 U/l lysozyme) and incubated at 22°C for 5 min. Purified water was added to the lysed cells, which were then incubated for 10 min at 4°C. Lysed cells were then centrifuged (4,000 ϫ g, 22°C, 20 min) to extract the supernatant containing periplasmic proteins.
Western blotting. Purified OMVs (10 l) were added to Laemmli sample buffer, boiled for 5 min, and resolved on a 10% SDS-PAGE gel. Proteins were transferred to a polyvinylidene difluoride membrane and Western blotting was performed using a rabbit polyclonal antibody raised against CroP and a peroxidaseconjugated anti-rabbit secondary antibody, as previously described (38). Quantity One software (Bio-Rad) was used to image and quantify protein bands.
CroP activity assays. The activity of the outer membrane protease CroP in OMVs or whole C. rodentium cells was determined by measuring the cleavage of the C2 FRET substrate [2Abz-SLGRKIQ-IK(Dnp)-NH 2 ; AnaSpec], as previously described (38). OMVs and C. rodentium cells grown to an OD 595 of 0.5 were diluted in PBS, transferred into a quartz cuvette, and incubated with the FRET substrate (6 M) at 22°C. Cleavage of the FRET substrate was monitored over 60 min by measuring fluorescence emission at 430 nm at an excitation wavelength of 325 nm using a Varian Cary Eclipse fluorescence spectrophotometer. Excitation and emission slit widths were set at 5 nm. Data were fitted to a Michaelis-Menten nonlinear regression curve using GraphPad Prism 7 software.

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at https://doi.org/10.1128/JB .00454-18.