|Home | About | Journals | Submit | Contact Us | Français|
The histidine sensor kinase (HK) QseC senses autoinducer 3 (AI-3) and the adrenergic hormones epinephrine and norepinephrine. Upon sensing these signals, QseC acts through three response regulators (RRs) to regulate the expression of virulence genes in enterohemorrhagic Escherichia coli (EHEC). The QseB, QseF, and KdpE RRs that are phosphorylated by QseC constitute a tripartite signaling cascade having different and overlapping targets, including flagella and motility, the type three secretion system encoded by the locus of enterocyte effacement (LEE), and Shiga toxin. We modeled the tertiary structure of QseC's periplasmic sensing domain and aligned the sequences from 12 different species to identify the most conserved amino acids. We selected eight amino acids conserved in all of these QseC homologues. The corresponding QseC site-directed mutants were expressed and still able to autophosphorylate; however, four mutants demonstrated an increased basal level of phosphorylation. These mutants have differential flagellar, motility, LEE, and Shiga toxin expression phenotypes. We selected four mutants for more in-depth analyses and found that they differed in their ability to phosphorylate QseB, KdpE, and QseF. This suggests that these mutations in the periplasmic sensing domain affected the region downstream of the QseC signaling cascade and therefore can influence which pathway QseC regulates.
IMPORTANCE In the foodborne pathogen EHEC, QseC senses AI-3, epinephrine, and norepinephrine, increases its autophosphorylation, and then transfers its phosphate to three RRs: QseB, QseF, and KdpE. QseB controls expression of flagella and motility, KdpE controls expression of the LEE region, and QseF controls the expression of Shiga toxin. This tripartite signaling pathway must be tightly controlled, given that flagella and the type three secretion system (T3SS) are energetically expensive appendages and Shiga toxin expression leads to bacterial cell lysis. Our data suggest that mutations in the periplasmic sensing loop of QseC differentially affect the expression of the three arms of this signaling cascade. This suggests that these point mutations may change QseC's phosphotransfer preferences for its RRs.
Enterohemorrhagic Escherichia coli (EHEC) O157:H7 causes outbreaks of bloody diarrhea and hemolytic-uremic syndrome (HUS) worldwide. EHEC colonizes the colon, where it forms attaching and effacing (A/E) lesions on enterocytes. The hallmark of these lesions is the effacement of the microvilli and the rearrangement of the cytoskeleton to form a pedestal-like structure that cups the bacterium (1,–3). A/E lesion formation requires the expression of a type three secretion system (T3SS) encoded within the locus of enterocyte effacement (LEE) (4, 5). Several LEE-encoded proteins are also necessary for A/E lesion formation, such as the adhesin intimin (6) and its receptor Tir (7). The LEE-encoded T3SS translocates effector proteins encoded both within and outside the LEE region. One such effector is EspFu/TccP, which is essential for A/E lesion formation (8,–15). Additionally, EHEC also produces Shiga toxins (Stx) that are responsible for HUS (16). This toxin is encoded by an integrated lambdoid phage in the EHEC chromosome. The genes encoding Stx are located within the late-phage genes and expressed along with the phage's late genes when the phage enters the lytic cycle due to activation of the cell's SOS response (17,–19).
Among many signals, there are three chemical signals sensed by EHEC that have been extensively reported to activate transcription of its virulence genes: a bacterial autoinducer (AI-3) produced by the normal gastrointestinal flora and two hormones (epinephrine/norepinephrine) produced by the host (20). These signals are detected by two membrane-bound histidine sensor kinases, QseC and QseE, which subsequently relay this information to a complex regulatory cascade to activate the transcription of key virulence genes (21, 22). When in the presence of these signals, QseC first autophosphorylates and then transfers the phosphate to QseB, its cognate response regulator (RR). The qseBC genes are cotranscribed in an operon that is also autoregulated (23). QseB is involved in regulation of the flagellar and motility genes (24, 25) (Fig. 1). This RR directly binds to the promoters of qseBC and flhDC (the master transcription regulator of the flagellar regulon), thereby allowing expression of these genes to be influenced by exogenous signals from the host or host flora. In the case of flhDC, the operon is activated by phosphorylated QseB but is repressed by unphosphorylated QseB (24, 25).
In addition to phosphorylating its cognate RR, QseC also phosphorylates the noncognate RRs QseF and KdpE (25). Through these RRs, QseC regulates expression of the LEE and Stx genes at the transcriptional and posttranscriptional levels (25,–27). There are 41 genes within the LEE, and the majority of them are organized within five major operons (LEE1 to -5) (28,–32). The first gene within LEE1 is ler, which encodes the master transcriptional activator of all LEE genes (29, 32, 33). KdpE in conjunction with Cra (a member of the LacI family that uses fluctuations in sugar concentrations to activate or inhibit the expression of its target genes ) directly binds to the regulatory region of ler (LEE1) to activate expression of all the LEE genes in a cascade fashion (25, 26, 35). Cra is active during gluconeogenesis (34); consequently, the growth of EHEC under glycolytic conditions inhibits LEE expression while growth under gluconeogenic conditions activates expression of these genes, and this sugar-dependent regulation is achieved through KdpE and Cra (26).
The QseF RR is phosphorylated by both QseC and QseE (25). While QseC senses AI-3, epinephrine, and norepinephrine, QseE senses epinephrine/norepinephrine, phosphates, and sulfates (21, 22). QseF is involved in the regulation of the stxAB genes encoding Shiga toxin (25), and QseF and QseB directly activate expression of the glmY small RNA (sRNA) that is encoded upstream of the qseEGF operon (36). The glmY sRNA posttranscriptionally controls expression of the LEE and espFu genes, which are necessary for A/E lesion formation (37). The GlmY sRNA is known to stabilize the GlmZ sRNA (constitutively encoded elsewhere in the E. coli chromosome ), and GlmZ is the sRNA that base pairs with its target mRNAs in an Hfq-dependent manner (38). GlmY/GlmZ destabilizes the LEE4 and LEE5 mRNA and promotes translation of EspFu, modulating the levels of A/E lesions produced by EHEC (38).
Although the phosphorylation cascade and resulting genetic regulation controlled by QseC have been studied extensively, the structural requirements for kinase activity are still unclear. The purpose of this study was to identify amino acids in the periplasmic domain of QseC that affect its ability to function as a kinase. Despite the fact that the periplasmic domain is predicted to be the site of signal sensing, the amino acids described here appear to play a role in controlling the interaction between QseC and the three RRs, QseB, KdpE, and QseF.
QseC is present in several species of bacteria. We aligned the periplasmic domain of several QseC proteins to identify the most conserved amino acids in this domain to target for mutagenesis (39). When the periplasmic domain (amino acid [aa] residues 37 through 155) of EHEC QseC was aligned with the same region of homologous proteins, at least eight strictly conserved residues were identified (Fig. 2A). To better understand the position of these residues within the periplasmic domain, a model of this domain was predicted using the Modeller software. When the positions of the residues were visualized on the predicted model (Fig. 2B), the conserved residues appeared to be located in one of two structural regions. Residues D101, F117, R130, V145, Q147, and W129 were predicted to be located in the vicinity of a series of β-sheets, forming a potential binding pocket. In both CitA and DcuS, two histidine kinases (HKs) that resemble the topology and secondary structure of QseC, this binding pocket is the site of ligand binding (40, 41). For this reason, these residues were chosen for site-directed mutagenesis. Three of these residues (R130, V145, and Q147), plus two more residues (D45 and R152), were also strictly conserved. D45 and R152 were both located within or near α-helices proximal to the potential binding pocket. Due to their location and conservation within the predicted structure, all eight of these residues were chosen for site-directed mutagenesis.
Using liposomes, we have previously reported that QseC increases its autophosphorylation in response to epinephrine, norepinephrine, and AI-3 (42, 43). Here, we again confirmed that QseC increases its autophosphorylation in liposomes in the presence of epinephrine. In liposomes, QseC's response to epinephrine can be observed between 2 and 10 min (Fig. 3A). Because it has been previously reported that QseC autophosphorylation could be monitored using membrane vesicles (44), we also assessed the conditions in these assays under which QseC would still respond to signals, given that membrane preparation for several site-directed mutants would be a more straightforward protocol than that for liposomes. We observed that in the membrane autophosphorylation of QseC was faster and that high levels of NaCl enhanced QseC autophosphorylation, decreasing its response to epinephrine. These results suggest that the conditions previously used to assess QseC autophosphorylation using Tris-buffered saline (TBS) buffer (containing 150 mM NaCl) (44) are not conducive to probing epinephrine sensing. QseC already autophosphorylated to high levels at 1 min, and increased concentrations of NaCl abolished its response to epinephrine, with 100 mM NaCl completely bypassing the response to this signal (Fig. 3B).
We changed the eight conserved residues in QseC's periplasmic domain (D101, F117, W129 D45, R130, V145, Q147, and R152) into alanines. These mutations were introduced into a cloned copy of qseC in a pBADMycHisA plasmid under the control of an arabinose-inducible promoter. The altered plasmids, as well as the plasmid carrying the wild-type (WT) gene, were used to complement strain VS138, a qseC-null mutant of EHEC. All of the mutant proteins were expressed (Fig. 4A) and were capable of autophosphorylation in membrane vesicles (Fig. 4B), suggesting that all of them maintained their kinase activity. The D101A, F117A, W129A, and V145A mutants had basal autophosphorylation activity comparable to that of WT QseC (Fig. 4B and andC).C). Notably, the D45A, R130A, Q147A, and R152A mutants demonstrated increased basal levels of autophosphorylation compared to WT QseC (Fig. 4B and andCC).
One of the QseC-regulated phenotypes in EHEC is motility. The qseC-null mutant (VS138) has reduced motility compared to the WT, and this phenotype can be complemented with qseC cloned in a plasmid (pVS155) (24, 45). We assayed VS138 (qseC-null mutant) complemented with WT QseC (pVS155) or the eight amino acid mutants for motility. When these strains were assayed for motility (Fig. 5A), only three strains were significantly hindered in motility: the D45A, R130A, and R152A mutants. The remaining strains were able to swim through soft agar at a rate similar to that of the strain complemented with WT qseC. Quantification of transcription of fliC, the gene encoding the major flagellin protein, was also used to determine the effect of these mutations on the flagellar biosynthesis pathway. Quantification of the fliC transcript by reverse transcription-quantitative PCR (qRT-PCR) identified two additional mutations, V145A and Q147A, that adversely affected the complemented strains' ability to produce flagellin (Fig. 5B). A potential explanation for the fact that these two strains were still motile is that even with lowered levels of fliC transcript, these strains still produced enough flagellin to assemble functional flagella. Interestingly, the D45A mutant appears to have fliC transcripts at levels equivalent to the WT, despite being less motile. This suggests that this mutation might be affecting some other aspect of flagellar biosynthesis or motility, such as motor assembly/rotation or flagellin secretion.
Next, we tested LEE and Stx expression in these mutants compared to that with WT QseC by qRT-PCR. Expression of ler (the master regulator of the LEE genes ) was decreased in six of the mutants but not in the Q147A mutant, where it was unchanged, and the R152A mutant, where it was increased (Fig. 6A). Expression of tir, which is required for A/E lesion formation (7), is also decreased in all mutants except for the Q147A mutant, in which its expression is highly increased. Expression of stx2 is significantly decreased in the W129A mutant and increased in the R152A mutant, being largely unchanged in the others (Fig. 6A). To delve deeper into QseC-regulated phenotypes, we streamlined further assays to four mutations that belonged to different phenotypic categories: D45A (increased basal level of autophosphorylation, decreased motility but unchanged fliC expression, decreased ler and tir expression, and unchanged stx2 expression), V145A (normal levels of basal autophosphorylation, unchanged motility but decreased fliC expression, decreased ler and tir expression, and unchanged stx2 expression), Q147A (increased autophosphorylation, unchanged motility but decreased fliC expression, unchanged ler but increased tir expression, and unchanged stx2 expression), and R152A (increased autophosphorylation, decreased motility and fliC expression, increased ler and decreased tir expression, and increased stx2 expression). Western blots for the LEE-secreted effector EspA are mostly congruent with the LEE transcriptional profiling, with decreased levels of secreted EspA in the D45A and V145A mutants and increased levels of secreted EspA in the Q147A and R152A mutants (Fig. 6B). It is worth noting that both the LEE5 (which contains tir) and LEE4 (which contains espA) operons are highly posttranscriptionally regulated (27, 46,–48), which can account for the differences in ler and tir transcription and EspA expression in the Q147A and R152A mutants. We also quantified A/E lesion formation in these mutants, and the D45A, VS145A, and R152A mutants formed A/E lesions at levels similar to those with WT QseC, but the Q147A mutant demonstrated enhanced A/E lesion formation (Fig. 6C and andDD).
QseC phosphorylates three RRs: QseB, KdpE, and QseF. These RRs regulate different and overlapping targets of this signaling cascade, including phenotypes such as flagella, motility, and LEE and Stx expression (25) (Fig. 1). Because expression of flagella, LEE, and Stx probably occurs at different times and under diverse environmental conditions, we investigated the conditions under which these RRs are expressed. QseB is highly expressed in mid- and late-logarithmic growth by EHEC in LB medium (Fig. 7), which is the environmental condition conducive to flagellar expression and motility (25). KdpE is expressed only in late-mid-logarithmic growth in low-glucose Dulbecco's modified Eagle's medium (DMEM) (gluconeogenic conditions) (Fig. 7), which is the environmental condition under which the LEE genes are expressed at optimal levels (26). QseF is expressed throughout growth under all conditions tested, LB, low-glucose DMEM (gluconeogenic), and high-glucose DMEM (glycolytic) (Fig. 7), which could reflect the fact that QseF's regulation of the LEE and Stx occurs indirectly and, in the case of the LEE, posttranscriptionally (25, 27).
Because these RRs regulate different arms of the QseC signaling cascade, we reconstituted QseC in liposomes and assessed phosphotransfer of the QseC mutants to these three RRs. Phosphotransfer from QseC to QseB and KdpE could be quantified because these proteins differ in size, but phosphotransfer from QseC to QseF could not be quantified because they have the same molecular weight (Fig. 8B and andC).C). The D45A and R152A mutants demonstrated decreased phosphotransfer to both QseB and KdpE. The V145A mutant had unaltered phosphotransfer to QseB but decreased phosphotransfer to KdpE. The Q147A mutant had increased phosphotransfer to QseB and decreased phosphotransfer to KdpE (Fig. 8B and andCC).
To further probe the different interactions between QseC and these RRs, we assessed ler and stx2 transcription in the ΔqseC ΔkdpE mutant and the ΔqseBC strain complemented with WT QseC or the four site-directed mutant genes. In the ΔqseC ΔkdpE mutant, complementation with WT QseC had a discrete impact in increasing ler and stx2 transcription (Fig. 8D), which suggests that the QseC-KdpE arm of this signaling cascade plays an important role in the expression of these genes. Transcription of ler was highly increased (15-fold) by complementation with WT QseC in the ΔqseBC mutant (Fig. 8E), again suggesting an important role for KdpE but not QseB in QseC-dependent activation of ler (25). The levels of ler transcription in the D45A and VS145A mutants were comparable to that in the ΔqseC ΔkdpE mutant with WT QseC (Fig. 8D), suggesting that these point mutants act through KdpE in regard to ler transcriptional activation. ler expression was decreased in the ΔqseBC mutant with the D45A and V145A mutations (Fig. 8E), which may be reflective of the dependence on KdpE for this activation in these mutants, given that they demonstrate lower levels of KdpE phosphorylation than strains with WT QseC (Fig. 8B and andC).C). Transcription of ler was unchanged for the R152 mutant compared to that with WT QseC in the ΔqseC ΔkdpE and ΔqseBC strains (Fig. 8D and andE),E), again suggesting that posttranscriptional regulation (27, 46,–48) may play a role in the regulation of the LEE by this mutant. Transcription of ler was increased in the ΔqseC ΔkdpE mutant with Q147A, while it was decreased in the ΔqseBC mutant compared to the strain with WT QseC (Fig. 8D and andE).E). These data suggest that there is an unknown QseB-dependent regulation of ler transcription triggered by this mutation, given that QseB is overphosphorylated by the Q147A mutant (Fig. 8B) and ler transcription increases in the absence of KdpE, which in theory could increase the pools of phospho-QseB within the cell, and decreases in the absence of QseB.
As in the ΔqseC ΔkdpE mutant, transcription levels of stx2 were also discretely impacted in the ΔqseBC mutant (Fig. 8D and andE),E), suggesting that neither KdpE nor QseB plays an important role in QseC-dependent stx2 activation, which has been previously shown to occur mostly through QseF (25). However, while levels of transcription of stx2 are not different between the strain with WT QseC and any of the ΔqseBC mutants (Fig. 8E), its transcription is increased in the D45A, Q147A and R152A ΔqseC ΔkdpE mutants (Fig. 8D). It is also possible that several of these phenotypes are affected by dysregulation of QseF, but unfortunately, we could not measure QseF's phosphorylation by QseC.
To investigate whether any of these mutations impacted the ability of QseC to sense and respond to epinephrine, we assessed the expression of a qseBC-lacZ transcriptional fusion (qseBC autoregulates its own expression in response to epinephrine ) in the absence and presence of epinephrine in the qseC strain complemented with WT QseC or the point mutants. Expression of qseBC was enhanced by epinephrine in the strain with WT QseC and all four mutants, suggesting that all mutants can still sense this signal (Fig. 8A). However, the effect of epinephrine on qseBC transcription was enhanced in the V145A, Q147A, and R152A mutants, suggesting that they are hyperresponders to this signal (Fig. 8A).
From a structural standpoint, QseC is typical of most sensor kinases; however, it is one of the only kinases that cannot autophosphorylate constitutively when only its cytoplasmic domain is expressed in vitro, requiring proper insertion of the full protein into the lipid membrane for autophosphorylation to occur (49). This suggests that the structure of the periplasmic and transmembrane regions is integral to its ability to autophosphorylate and to transfer that phosphoryl group to an RR. The point mutations described in this study are all located in the periplasmic domain (Fig. 2), but the mutations that displayed the largest effect on QseC function (D45A, V145A, Q147A, and R152A) were located near the periplasmic face of the inner membrane. In addition, three of these mutations (V145A, Q147A, and R152A) are located adjacent to a predicted HAMP domain (residues 164 to 234) (Fig. 2) that is theorized to play a role in kinase multimerization (50). These mutations had a strong effect on QseC's ability to phosphorylate its RRs and may have differentially affected the arms of this tripartite signaling cascade (Fig. 5 to to8).8). These data suggest that modifications in the periplasmic sensing domain of a sensor kinase alter its abilities for phosphotransfer to RRs.
Interestingly, the effect seen on RR phosphorylation varied among regulators. This suggests that QseC interacts with the RRs in different ways. One explanation is that QseC has a different affinity for each of the three RRs in this study (QseB, KdpE, and QseF). The fact that KdpE was phosphorylated poorly after any mutation in the periplasmic domain might indicate that QseC has a lower affinity for this RR than it does for the other two (Fig. 8B). Any alteration to the periplasmic domain, particularly in the vicinity of the HAMP domain, may have an adverse effect on QseC's ability to form multimers and efficiently phosphorylate RRs. If this is the case, we would expect the RR with the lowest affinity with QseC to be affected the most.
It the case of QseB, the different mutations in QseC had a range of effects on RR phosphorylation. Some mutations (D45A and R152A) resulted in lower QseB phosphorylation, while the Q147A mutation resulted in higher QseB phosphorylation, and the V145A mutation did not appear to have any effect on phosphorylation. Motility is decreased in both D45A and R152A mutants, congruent with the lower level of phosphorylation of QseB by these mutants (Fig. 5 and and8).8). However, flagellin expression, although decreased in R152A, is not altered in V45A, suggesting that the motility defect in V45A does not occur at the level of flagellin expression and may be due to effects on flagellar assembly and/or motor assembly/rotation. The V145A mutant has decreased ability to phosphorylate QseB, congruent with the lower expression of fliC; however, it does not demonstrate any motility defects (Fig. 5 and and8),8), suggesting that other aspects of flagellar assembly and rotation can compensate for lower flagellin expression in this mutant. The Q147A mutant overphosphorylates QseB but has decreased fliC expression and nonaltered motility (Fig. 5 and and8),8), suggesting again that other aspects of flagellar assembly and rotation are altered by the mutation.
Because QseC and QseF have similar molecular weights, making it difficult to quantify QseF autophosphorylation, we refrain from discussing the potential effects of these QseC mutations on QseC-QseF interactions.
Another structural difference is that the D45A mutation is located near the first transmembrane helix while the other three mutations are adjacent to the second transmembrane helix and the HAMP domain. Since the second transmembrane region and the HAMP domain are theoretically involved in multimerization, the presence of V145A, Q147A, and R152A mutations near this location might result in an alteration in structure or multimerization different from that caused by the D45A mutation and therefore have different effects on the phosphorylation of QseB and QseF. Moreover, one has to take into account that the RRs are differentially expressed under different environmental conditions (Fig. 7), which may change the pool of available RRs within the bacterial cell to be phosphorylated by QseC.
The notion that HKs are exquisitely faithful to their cognate RRs has been proposed using truncated HKs where only the cytoplasmic domain was utilized (51). Here, we show that mutations in the periplasmic sensing loop of QseC change phenotypes associated with the different RRs phosphorylated by this HK. Taken as a whole, this study indicates that the interactions between an HK and various RRs may be unique to each kinase/regulator pair and that it is possible to influence these interactions through specific mutations within the periplasmic domain of the kinase. By better understanding these effects, we may be able to better understand how kinases have evolved to regulate multiple two-component systems and how we can better target these kinases to selectively disrupt one kinase/regulator pair while leaving another kinase/regulator pair unaffected.
The plasmids and strains used in this study are listed in Table 1. Unless otherwise noted, all strains were grown in either Luria-Bertani (LB) medium or low-glucose Dulbecco's modified Eagle's medium (DMEM) at 37°C and 250 rpm.
Known QseC homologues were identified by BLAST search using EHEC QseC as the target sequence. Several pathogenic or commensal homologue examples (total of 12) were chosen, and the periplasmic domains (equivalent to residues 37 through 155 of the EHEC protein) of these homologues were aligned using CLC Sequence Viewer 6 (CLC bio). The structure of the periplasmic domain of EHEC QseC was predicted with Modeler version 9.10 (52, 53), using the known structures of E. coli CitA (40) and DcuS (41) as the template.
Nonpolar mutants were constructed using the λ red protocol to all be in the same genetic background (54). The ΔqseC ΔkdpE mutant was constructed by knocking out kdpE on VS138 using kdpEλRed-F and kdpEλRed-R primers (55). The ΔqseBC mutant was constructed by knocking out qseB on VS138 using qseBλRed-F and qseBλRed-R primers (25).
The QuikChange II site-directed mutagenesis kit (Agilent Technologies) was used to introduce alanine substitution mutations in a plasmid carrying a C-terminally His-tagged copy of the qseC gene under the control of the araBAD promoter (pVS155) (45). The primers (Table 2) were used to amplify pVS155, and the resulting PCR product was treated with DpnI and transformed into XL1-Blue cells. The mutations were confirmed by sequencing plasmid DNA.
Either pVS155 or derivatives containing point mutations were introduced into VS138, a qseC-null mutant of WT EHEC 86-24 (45). These strains were subsequently used in motility assays performed using the soft agar method, as previously described (24). Briefly, strains were stabbed in triplicate into soft agar motility plates (1% tryptone, 0.25% NaCl, 0.3% agar) containing 0.2% arabinose and incubated at 37°C. The diameter of the resulting halos was measured after 8 h.
Cultures were grown in low-glucose DMEM to an optical density at 600 nm (OD600) of 1.0. RNA from 3 biological replicates was extracted using the RiboPure bacterial isolation kit, according to the manufacturer's protocols (Ambion). qRT-PCR was performed as described previously (25). Briefly, diluted extracted RNA was mixed with validated primers (Table 2), RNase inhibitor, and reverse transcriptase (Applied Biosystems). The mixture was used in a one-step reaction utilizing an ABI 7500 sequence detection system. Data were collected using ABI Sequence Detection 1.2 software, normalized to endogenous rpoA levels, and analyzed using the comparative critical threshold cycle (CT) method. Analyzed data were presented as fold changes over WT levels. Student's unpaired t test was used to determine statistical significance. A P value of ≤0.05 was considered significant.
Complemented VS138 strains with pVS155 (WT QseC) or the site-directed mutants were grown in LB supplemented with 0.2% arabinose to an OD600 of 0.6, and cells were pelleted and lysed by five passages through an EmulsiFlex C3 emulsifier. The resulting lysate was cleared by centrifugation at 26,000 × g. Membranes were isolated from these lysates by ultracentrifugation at 179,000 × g. The pelleted membranes were solubilized in SDS-PAGE sample buffer, and proteins were separated on a 10% acrylamide gel. These proteins were transferred by Western blotting, probed with an anti-His antiserum primary antibody, and then incubated with a secondary antibody conjugated to streptavidin-horseradish peroxidase (HRP). ECL reagent (GE) was added and the membranes were exposed to film.
Secreted proteins were extracted and detected as previously described (4). Complemented VS138 strains with WT QseC (pVS155) or the point mutants were grown in low-glucose DMEM to an OD600 of 1.0. Cells were pelleted by centrifugation and suspended in phosphate-buffered saline. Supernatants were passed through a 0.22-μm-pore-size filter and treated with EDTA (5 mM final concentration), phenylmethylsulfonyl fluoride (PMSF) (50 μg/ml final concentration), and aprotinin (0.5 μg/ml final concentration). One hundred micrograms of bovine serum albumin (BSA) was added to each sample as a control. The volume of the supernatants was reduced to 50 μl using centrifugal filters with a molecular weight cutoff (MWCO) of 10,000. Ten micrograms of protein from each sample was separated on 4 to 15% SDS gradient gel and transferred to a polyvinylidene difluoride (PVDF) membrane. The blot was stained with 0.1% Ponceau S in 5% acetic acid to visualize the BSA loading control. The presence of secreted EspA was detected with polyclonal anti-EspA antibodies.
Assays were performed as described by Knutton et al. (56). Briefly, HeLa cells were grown on coverslips in wells containing DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin-glutamine (PSG) antibiotic mix at 37°C and 5% CO2 overnight to about 80% confluence. The wells were then thoroughly washed with phosphate-buffered saline (PBS) and replaced with fresh medium supplemented with arabinose (0.2% final concentration) and lacking antibiotics. Overnight static cultures of bacteria were then used to infect at a dilution of 100:1 (bacteria to DMEM). After a 6-h infection at 37°C and 5% CO2, the coverslips were washed, fixed, and permeabilized. The samples were then treated with fluorescein isothiocyanate (FITC)-labeled phalloidin and propidium iodide (PI) to visualize actin accumulation and bacteria, respectively. PI also stained HeLa nuclei red. The coverslips were then mounted on slides and visualized with a Zeiss Axiovert microscope. Pedestal formation was quantified as the percentage of pedestals formed per attached bacterium. Replicate coverslips from multiple experiments were quantified, and statistical analysis was performed using the Student t test. Serially diluted samples of the original bacterial cultures were also plated to confirm that similar CFU ratios were used for infection.
His-tagged QseC proteins and response regulators (QseB, KdpE, and QseF) were isolated from cultures grown to mid-log phase in LB and induced with either 0.2% arabinose (kinases) or 400 mM isopropyl-β-d-thiogalactopyranoside (IPTG) (response regulators) at 30°C overnight. The cells were pelleted, resuspended in 50 mM phosphate buffer containing 150 mM NaCl, and lysed by passage through an EmulsiFlex C3 emulsifier. The resulting lysate was cleared by centrifugation, and proteins were isolated using the standard nickel-nitrilotriacetic acid (Ni-NTA) technique, as described by the manufacturer (Qiagen).
QseC autophosphorylation experiments were performed as described by Clarke et al. (21). Briefly, as described previously (24), the E. coli strain containing pVS155 (qseC::Myc-His) was grown at 37°C in LB to an OD600 of 0.7, at which point arabinose was added to a final volume of 0.2% and allowed to induce for 3 h (21). Nickel columns were utilized, according to the manufacturer's instructions (Qiagen). Autophosphorylation experiments were performed with QseC embedded in liposomes. Liposomes were reconstituted as described by Janausch et al. (57). Briefly, 50 mg of E. coli phospholipids (20 mg/ml in chloroform; Avanti Polar Lipids) was evaporated and then dissolved into 5 ml of potassium phosphate buffer containing 80 mg of N-octyl-β-d-glucopyranoside. The solution was dialyzed overnight against potassium phosphate buffer. The resulting liposome suspension was subjected to freeze/thawing in liquid N2. Liposomes were then destabilized by the addition of 26.1 mg of dodecyl-maltoside, and 2.5 mg of QseC-Myc-His was added, followed by stirring at room temperature for 10 min. Two hundred sixty-one milligrams of Bio-Beads was then added to remove the detergent, and the resulting solution was allowed to incubate at 4°C overnight. The supernatant was then incubated with fresh Bio-Beads for 1 h in the morning. The resulting liposomes containing reconstituted QseC-Myc-His were frozen in liquid N2 and stored at −80°C until used. Orientation of HKs in the liposome system has been established by previous groups (58) and can be concluded from the accessibility of ATP to the kinase site and anti-Myc antisera to the C-terminal QseC Myc tag without disruption of the liposomes (21). Twenty microliters of the liposomes containing QseC-Myc-His was adjusted to 10 mM MgCl2 and 1 mM dithiothreitol (DTT), no signal, or 50 μM epinephrine, frozen and thawed rapidly in liquid N2, and kept at room temperature for 1 h. A volume of 0.625 μl of [γ32P]dATP (110 terabecquerels [TBq]/mmol) was added to each reaction mixture. At the 1-, 2-, 5-, and 10-min time points, 20 μl of SDS loading buffer was added (21). The samples were run on SDS-PAGE gels without boiling, according to standard procedures (59), and visualized via a phosphorimager.
Phosphotransfer assays were performed by first adding DTT (1 mM final concentration), MgCl2 (0.5 mM final concentration), and epinephrine-bitartrate (50 μM final concentration) to an aliquot of liposome. These were then subjected to three freeze/thaw cycles using liquid nitrogen and held at room temperature for 1 h. In a 20-μl reaction mixture, response regulator was added to 10 μl of loaded liposome at a 1:1 molar ratio to the kinase, and the reaction was started by adding 0.3 μl of [γ32P]dATP (3 μCi). Reactions were stopped by adding 5 μl of SDS-PAGE sample buffer supplemented with additional SDS to a concentration of 18% (wt/vol). Proteins were resolved, without boiling, on 10% SDS-PAGE gels. Radiolabeled proteins were visualized by exposing the gel to a phosphorimaging screen. The bands were quantitated using the ImageQuant version 5.0 software (Amersham).
Membranes were prepared from VS138 (qseC mutant) with or without pVS155 (QseC in pBADMycHisA) grown in LB and induced with 0.02% arabinose. Total membranes were isolated by ultracentrifugation, suspended in 50 mM Tris buffer (pH 7.5) with 100 mM NaCl, and stored at −80°C. For the autophosphorylation assays, the membranes were washed with Tris-EDTA (TE) and assayed in 10 mM Tris (pH 7.5)–2 mM MgCl2–0 to 100 mM NaCl and 2.5 μM [γ32P]dATP (110 TBq/mmol) in the absence or presence of 100 μM epinephrine on ice for 0 to 3 min. Assays were stopped with sample buffer, and assay mixtures were loaded on a 10% SDS gel. Radiolabeled proteins were visualized by exposing the gel to a phosphorimaging screen. We also performed Western blotting with anti-His antisera on blots from these genes to ensure the identity of QseC.
We thank members of the Sperandio lab for collegial discussions of this work.
This work was supported by National Institutes of Health (NIH) grants AI053067, AI05135, AI077613, and AI114511. C.T.P. was supported through NIH training grant 5 T32 AI7520-14.
The contents of this article are solely the responsibility of the authors and do not represent the official views of the NIH NIAID.