Article pubs.acs.org/biochemistry
Mutational Replacements at the “Glycine Hinge” of the Escherichia coli Chemoreceptor Tsr Support a Signaling Role for the C‑Helix Residue Andrea Pedetta,† Diego Ariel Massazza,‡ María Karina Herrera Seitz,† and Claudia Alicia Studdert*,§ †
Instituto de Investigaciones Biológicas, Universidad Nacional de Mar del Plata-CONICET, Mar del Plata, Buenos Aires, Argentina Instituto Nacional de Tecnología en Materiales, Universidad Nacional de Mar del Plata-CONICET, Mar del Plata, Buenos Aires, Argentina § Instituto de Agrobiotecnología del Litoral, CONICET-Universidad Nacional del Litoral, Santa Fe, Santa Fe, Argentina ‡
S Supporting Information *
ABSTRACT: Bacterial chemoreceptors are dimeric membrane proteins that transmit signals from a periplasmic ligandbinding domain to the interior of the cells. The highly conserved cytoplasmic domain consists of a long hairpin that in the dimer forms a four-helix coiled-coil bundle. The central region of the bundle couples changes in helix packing that occur in the membrane proximal region to the signaling tip, controlling the activity of an associated histidine kinase. This subdomain contains certain glycine residues that are postulated to form a hinge in chemoreceptors from enteric bacteria and have been largely postulated to play a role in the coupling mechanism, and/or in the formation of higher-order chemoreceptor assemblies. In this work, we directly assessed the importance of the “glycine hinge” by obtaining nonfunctional replacements at each of its positions in the Escherichia coli serine receptor Tsr and characterizing them. Our results indicate that, rather than being essential for proper receptor−receptor interaction, the “glycine hinge” residues are involved in the ability of the receptor to switch between different signaling states. Mainly, the C-helix residue G439 has a key role in shifting the equilibrium toward a kinase-activating conformation. However, we found second-site mutations that restore the chemotactic proficiency of some of the “glycine hinge” mutants, suggesting that a complete hinge is not strictly essential. Rather, glycine residues seem to favor the coupling activity that relies on some other structural features of the central subdomain.
C
helical bundle of four antiparallel helices. This length is variable in MCPs from different organisms, according to seven-residue long (two turns of helix) insertions and/or deletions that appeared symmetrically in both arms of the hairpin over the course of evolution.2 The overall coiled-coil character of the hairpin is thus maintained by the periodicity of hydrophobic residues occupying positions a and d of each amino acid heptad. Chemoreceptors are classified into families or classes named after the number of heptads present in the hairpin. Three distinct subdomains can be recognized in the cytoplasmic domain of different length-defined families. The membrane distal tip mediates interactions between dimers to form trimers of dimers of the same or different specificity. On the outside of the trimers, this region also interacts with the dimeric histidine kinase CheA and the essential coupling protein CheW. These interactions interconnect different trimers and result in the large, hexagonally arranged arrays
hemotaxis allows bacteria to detect chemical gradients with great sensitivity and modulate their movement accordingly to find optimal conditions under which to thrive. Amazingly conserved among Bacteria and Archaea, the machinery responsible for that behavior consists of a few proteins that group together in large arrays and direct a cell’s motility by means of controlling the phosphorylation level of a response regulator protein. In spite of the wealth of molecular details that are available today (see ref 1 for a recent review), there are still open questions about how exactly the signal is transmitted from the exterior of the cell to the central histidine kinase. Chemoreceptors, or MCPs (for methyl-accepting chemotaxis proteins), which are transmembrane homodimeric rod-shaped proteins of α-helical structure (see the scheme in Figure 1A), perform signal transmission. In canonical MCPs, the periplasmic ligand-binding domain is located between two transmembrane segments. A HAMP domain, consisting of a four-helix bundle of parallel helices, connects the second transmembrane segment to the highly conserved cytoplasmic domain that is the hallmark of chemoreceptors. It consists of a hairpin that is ∼20 nm long and forms another coiled-coil © 2017 American Chemical Society
Received: May 11, 2017 Revised: June 28, 2017 Published: June 30, 2017 3850
DOI: 10.1021/acs.biochem.7b00455 Biochemistry 2017, 56, 3850−3862
Article
Biochemistry
Figure 1. Architecture and conservation of the “glycine hinge” of the chemoreceptors. (A) Cartoon representation of a Tsr homodimer showing important signaling features. Cylindrical bars represent α-helical regions. Methylation residues are shown as dark gray circles. White circles represent the glycine residues that form the postulated hinge within the coupling subdomain. The inset shows a molecular view of the “glycine hinge” region obtained with PyMol. G340, G341, and G439 residues of the dark gray subunit are colored pink; the other subunit is colored light gray. (B) Conservation analysis of the coupling subdomain in the 36H chemoreceptor class. After retrieving 2428 receptors belonging to 36H class, we aligned their cytoplasmic domain using the MAFFT engine, and that alignment was used to analyze the amino acid conservation of the coupling subdomain with WebLogo3. This subdomain encompasses N-helix residues 321−362 and C-helix residues 420−461 (E. coli serine receptor Tsr numbering); “glycine hinge” residues are denoted with light blue stars (G340 and G341 in the N-helix and G439 in the C-helix); knob positions a and d are denoted with red arrowheads.
They suggested that the “glycine hinge” might be crucial for the ability of receptors to alternate between ON and OFF conformations. Additionally, given the structure of the hinge, they proposed that it might permit some bending necessary for signaling and/or for efficient trimer formation. In the available crystal structures for chemoreceptors,5−8 the coupling subdomain, also called the flexible bundle, displays a canonical knobs-into-holes conformation, in which hydrophobic residues (knobs) fit into holes consisting of other residues in a neighboring helix within the four-helix bundle. Along the bundle, knob residues are arranged in square a−d layers. When small knob residues are paired with large ones in the opposite helix, the layers acquire a diamond shape. In E. coli’s receptors, the upper part of the flexible region consists of small knobs in the N-helix, and large knobs in the C-helix, which determines a stack of diamond-shaped layers in the same orientation. Below the hinge, large knobs predominate in the N-helix and small ones in the C-helix, constituting a second stack of diamond-shaped layers in cross-orientation with respect to the upper one. Alexander and Zhulin2 proposed that this organization lends instability to the structure. They observed this pattern in the coupling subdomain of different receptor classes, suggesting that it might have a signaling role.2 To improve our understanding of the role of the “glycine hinge” of chemoreceptors in higher-order organization and signaling, in this work we obtained and characterized a set of nonfunctional mutations of Tsr, the serine chemoreceptor of E. coli, at each of the positions that constitute it. Our results indicate that, rather than being involved in the formation of trimers of dimers, the “glycine hinge” has a key role in the ability of the receptor to alternate between active and inactive conformations. In particular, the C-helix glycine residue is
that are usually located at the cell poles and were imaged by cryoelectron microscopy in a variety of organisms.3 On the other hand, the membrane proximal region of the hairpin contains three to five methylatable residues (glutamates or deamidable glutamines), responsible for adaptation to stimuli. Upon detection of a gradient, chemoreceptors change their conformation, and the kinase CheA responds accordingly with an immediate change in activity. However, chemoreceptor conformation is also detected by the methyltransferase CheR and the methylesterase CheB (which is in turn activated by CheA-mediated phosphorylation). The opposing activities of the methylation enzymes result in a change of the methylation level of the receptor that ultimately restores prestimulus kinase activity, allowing the cell to respond to new stimuli. Between the signaling tip and the adaptation region, the coupling domain seems to play a central role in signal transmission. An alignment of MCPs from different organisms shows that this subdomain is notably less conserved than the other two,2 yet certain common features suggest possible requirements to fulfill its role. The coupling subdomain is split into two by glycine residues that in Escherichia coli receptors are located two in the N-helix and one in the C-helix of the hairpin, forming a ring termed the “glycine hinge”.4 Except for the 24H (24 heptads) class that lacks the entire coupling region, all the MCP classes have at least one glycine residue conserved in this location, which would suggest the functional importance of glycine residue(s) in the middle of the coupling subdomain. Coleman et al.4 found that alanine or cysteine replacements at each of the three “glycine hinge” residues in Tar, the aspartate receptor of Salmonella typhimurium, abrogated chemotactic function, rendering mutant receptors locked in their ON (kinase-activating) or OFF (kinase-inhibiting) conformations. 3851
DOI: 10.1021/acs.biochem.7b00455 Biochemistry 2017, 56, 3850−3862
Article
Biochemistry
derivatives were inoculated on tryptone semisolid agar plates (1% tryptone, 0.5% NaCl, and 0.25% agar) containing 12.5 μg mL−1 chloramphenicol and 0.45 μM sodium salicylate. The plates were incubated for 7−8 h at 30 °C. For Tar function, UU1624 cells (carrying a chromosomal copy of the tar gene as the only chemoreceptor) transformed with the pRR31 vector control or pCS12 derivatives were inoculated into minimal H1 semisolid agar plates22 supplemented with 0.15 mM aspartate and 2 mM glycerol and containing chloramphenicol (12.5 μg mL−1) and sodium salicylate at different concentrations. The plates were incubated for 20 h at 30 °C. TMEA Cross-Linking Competition Assay. UU1613 cells (Tar-S364C, cheAcheW, cheRcheB) transformed with pRR31 vector control or pCS12 derivatives were grown at 30 °C to mid log phase in tryptone broth supplemented with 0.45 μM sodium salicylate, harvested by centrifugation, and resuspended at an OD600 of 2 in 10 mM potassium phosphate (pH 7.0) and 0.1 mM EDTA. Cell suspensions (0.5 mL) were incubated for 5 min at 30 °C and then treated with 50 μM tris(2maleimidoethyl)amine (TMEA, Pierce) for 20 s at 30 °C. Reactions were quenched by the addition of 10 mM Nethylmaleimide. Cells were pelleted and then lysed by being boiled in 50 μL of sample buffer.19 Lysate proteins were analyzed by electrophoresis in 10% acrylamide, 0.05% bis(acrylamide) gels and visualized by immunoblotting with an anti-Tsr antiserum. As secondary antibodies, we used either Cy5-labeled (Amersham) or alkaline phosphatase-conjugated (Sigma) goat anti-rabbit immunoglobulin. Cy5-labeled antibodies were detected with a Storm 840 fluorimager (Amersham); alkaline phosphatase-conjugated antibodies were developed with nitro blue tetrazolium and 5-bromo-4chloro-3-indolyl phosphate (both from Sigma). Receptor Clustering Assay. Receptor clusters were visualized by fluorescence light microscopy in cells expressing a YFP−CheZ fusion protein as a reporter. Cells containing pFG1 (YFP−CheZ under IPTG control) and the vector control plasmid pRR31 or different pCS12 derivatives were grown at 30 °C in tryptone broth to mid log phase in the presence of 100 μM IPTG and 0.45 μM sodium salicylate. Cells were collected and examined essentially as described previously.23 Cell fields were photographed, and at least 100 cells were inspected by eye to determine the proportion of individual cells with one or more distinct bright spots of fluorescence, which are indicative of a receptor cluster. Tethered-Cell Assay. Strains containing pRR31 vector control or different pCS12 derivatives were tethered to microscope slides with anti-flagellin antiserum as described previously,24 examined under a phase-contrast microscope, and subjected to 10 mM L-serine or 2 M glycerol stimuli. For each strain, at least 100 rotating cells were observed for 15 s each and classified into one of five categories according to their pattern of rotation. Cells were observed within 3 min of the addition of the attractant. The overall percent of time spent in clockwise (CW) rotation was computed as a weighted sum:14 the percent of cells that rotated exclusively CW, plus 0.75 times the percent of cells rotating predominantly CW, plus 0.5 times the percent of cells reversing frequently, plus 0.25 times the percent of cells rotating predominantly, but not exclusively, counterclockwise (CCW). Tsr Methylation Assay. UU1626 or U2612 cells transformed with pCS12 derivatives were grown at 30 °C to mid log phase in tryptone broth with 0.45 μM sodium salicylate,
important for shifting the equilibrium toward the ON conformation. Nevertheless, an incomplete “glycine hinge” can be skipped by the generation of compensatory second-site mutations, in line with the conservation of only one or two of these residues in other receptor families.
■
MATERIALS AND METHODS Bacterial Strains. Strains were derivatives of E. coli K12 strain RP4379 and are listed in Table 1. Table 1. Bacterial Strains strain
relevant genotype
ref
RP9352
zea::Tn10 (tsr)DE7028 (tar-tap)DE5201 zdb::Tn5 (trg) DE100 (cheZ)DEm67−25 (aer)DE1 ygjG::Gm (tsr)DE7028 (tar-tap)DE5201 zdb::Tn5 (trg)DE100 ( flhD-f lhB)DEtr4 (tsr)DE7028 (trg)DE100 Tar-S364C (tsr)DE7028 zdb::Tn5 (trg)DE100 (tap-cheB) DE2241 Tar-S364C (tsr)DE7028 (trgDE100) (tap-cheB)DE2241 (cheA-cheW)DE2167 (aer)DE1 ygjG::Gm (tsr)DE7028 (tap)DEm365−4 zdb::Tn5 (trg)DE100 (aer)DE1 ygjG::Gm (cheA-tap)DE2260 (tsr)DE7028 zdb::Tn5 (trg)DE100 mutD5 zae::Tn10−13 (flhD-f lhA)DEtr4 (tsr)DE7028 zdb::Tn5 (trg)DE100 (tar-cheB)DE4346 (tsr)DE5547 (aer)DE1 ygjG::Gm (trg) DE4543 (tar-tap)DE4530 (tsr)DE5547 (aer)DE1 (trg)DE4543
13
UU1250 UU1581 UU1598 UU1613 UU1624 UU1626 UU1935 UU2610 UU2612
14 11 11 11 15 16 17 18 18
Plasmids. Plasmids derived from pACYC184,10 which confers chloramphenicol resistance, were pRR31 (salicylateinducible expression vector)11 and pCS12 (salicylate-inducible wild-type tsr).11 pFG1 (IPTG-inducible expression vector) (A. F. Gasperotti, unpublished data), encoding a functional fusion between the phosphatase CheZ and the yellow fluorescent protein (YFP), derives from pBR32212 and confers ampicillin resistance. Site-Directed Mutagenesis. Mutations were introduced into plasmids with the QuikChange Site-Directed Mutagenesis Kit (Stratagene), using pCS12 as the template. Candidate mutants were verified by sequencing the entire protein-coding region. Protein Expression. UU1581 cells (devoid of all chemotaxis proteins) complemented with pCS12 or its derivatives were grown at 30 °C to mid log phase in tryptone broth (1% tryptone and 0.5% NaCl) supplemented with 0.45 μM sodium salicylate, harvested by centrifugation (6000g), and resuspended at an OD600 of 2 in 10 mM potassium phosphate (pH 7.0) and 0.1 mM ethylenediaminetetraacetic acid (EDTA). Cells from 0.5 mL of the suspension were pelleted and lysed by boiling in 50 μL of sample buffer.19 Proteins released from the lysed cells were analyzed by electrophoresis in sodium dodecyl sulfate-containing polyacrylamide gels (SDS−PAGE) and visualized by immunoblotting with an antiserum directed against the highly conserved portion of the Tsr signaling domain20 in 10% acrylamide, 0.05% bis(acrylamide) gels.21 The Cy5-labeled (Amersham) antibody was used as the secondary antibody and detected with a Storm 840 fluorimager (Amersham). All gel images were analyzed with ImageQuant (Amersham). Chemotaxis Assay in Semisolid Agar Plates. For Tsr function, cells carrying the pRR31 vector control or pCS12 3852
DOI: 10.1021/acs.biochem.7b00455 Biochemistry 2017, 56, 3850−3862
Article
Biochemistry Table 2. Summary of Glycine Mutant Receptor Functional Propertiesa response to glycerolg Tsr
expressionb
chemotaxis in soft agar platesc
Tar functiond
CheA activatione
response to f L-serine
cheRcheB
CheR+ CheB+
methylationh buffer/Ser
wild type (−) G340I G340P G340T G341D G341R G341T G439A G439E G439H G439M G439P G439R G439S G439Y G439H/G469S G439Y/G431S G439Y/G469S G439Y/D471N
1.0 − 0.70 0.70 0.90 0.75 1.7 1.1 0.85 1.0 0.85 0.65 0.85 0.85 0.85 0.80 0.80 1.0 1.1 0.65
1.0 0.10 0.15 0.15 0.15 0.15 0.25 0.20 0.10 0.10 0.15 0.20 0.10 0.15 0.10 0.30 0.60 1.1 1.0 0.40
1.3 1.0 0.50 1.2 0.50 0.40 0.75 0.35 0.25 0.25 0.50 0.30 0.35 0.45 0.40 0.70 nd nd nd nd
1.0 − 1.1 − 0.95 − 0.80 0.75 − − − − − − − − − 1.0 1.2 −
yes nd yes na no na no yes na na na na na na na na na no yes na
nd nd nd yes nd no nd nd no no no no no yes no yes yes nd nd yes
yes no nd nd nd nd nd nd no no no no no yes no yes yes nd nd yes
+/+++ nd +/+ +/+ +/+ +++/na +/++ +/+++ +++/na ++/na +++/na +++/na +++/na ++/+++ +++/na +++/na nd nd nd nd
a
E. coli cells carrying pRR31 (empty vector), plasmid pCS12 (wild-type Tsr), or its derivatives with nonfunctional point mutations at the glycine hinge were characterized as indicated below. All parameters were normalized to those of wild-type Tsr. Values of 1.0 were rounded to the nearest 0.1. A dash indicates a nondetectable response. nd means nondetermined, and na means does not apply. bExpression was measured in strain UU1581, which lacks all chemotaxis proteins. The total receptor protein content was analyzed by SDS−PAGE followed by immunoblotting with an antibody against the conserved region of Tsr. cChemotaxis in soft agar plates was assessed in strain UU2612 (ΔMCPs). After 7−8 h at 30 °C, ring diameters were measured. dInterference with Tar function was determined in minimal medium soft agar plates containing 0.45 μM sodium salicylate in strain UU1624 that expresses Tar as the only MCP. After 20 h at 30 °C, ring diameters were measured. eCheA activation was assessed by tethered-cell assays in strain UU2610 (ΔMCPs cheRcheB). fThe response to the attractant L-serine in excess was assessed in strain UU2610. It was assayed in only those derivatives that could activate the kinase. gThe response to the repellent glycerol in excess was assessed both in cheRcheB strain UU2610 and in CheR+ CheB+ strain UU2612. It was assayed in only those derivatives that could not activate the kinase. hMethylation was analyzed in strain UU1626 (cheAcheW) upon treatment with buffer or 100 mM L-serine. Qualitative categories were assigned by comparison with the band pattern of wild-type Tsr.
using the parental plasmid and wild-type Tsr-encoding pCS12 as negative and positive controls, respectively. Construction of WebLogo Patterns. MCP protein sequences were obtained from the 5830 bacterial reference proteomes (to avoid potential over-representation of species) available in the Uniprot database and applying the advanced search tool provided by the server. A total of 22504 matching sequences were identified as MCPs. Using the HMM profile26 kindly provided by R. Alexander, sequences were assigned to the different receptor length classes defined by Alexander and Zhulin.2 A sequence was assigned to a specific class if it was the top-scoring class and if the E value exponent was smaller than −100. Following this criteria, 2564 sequences were assigned to the 36H class. An alignment was done using the MAFFT online engine, and it was manually edited using AliView software. After the identification of the highly conserved MCP cytoplasmic motif, the position corresponding to the tip residue of the hairpin was identified, and all the sequences in the alignment were cut, leaving 126 residues to each side of the tip, which corresponds to the 18 heptads per helix that constitute the cytoplasmic domain of the 36H class receptor. Sequences were realigned, and the alignment was again manually cured. The final alignment of 2428 36H class MCPs was used to generate WebLogos using the online application WebLogo327 (http://weblogo.threeplusone.com).
harvested by centrifugation, washed three times with 10 mM potassium phosphate (pH 7.0) and 0.1 mM EDTA, and finally resuspended at an OD600 of 1 in the same buffer with 10 mM lactate, 1 mM methionine, and 200 mg mL−1 chloramphenicol. After being incubated at 30 °C for 15 min, cell suspensions (0.5 mL) were challenged with 100 mM L-serine, 2 M glycerol, or an equivalent volume of buffer for 30 min. Cells were then pelleted and lysed by being boiled in 50 μL of sample buffer. Proteins released from the lysed cells were analyzed by SDS−PAGE in 11% acrylamide, 0.075% bis(acrylamide) gels and visualized by immunoblotting with the Tsr antiserum. Isolation of Second-Site Suppressor Mutations. To identify second-site mutations that suppressed the chemotaxis defects exhibited by the Tsr-G439 variants, plasmids encoding those variants were mutagenized by being passed through UU1935, a proofreading-deficient polymerase mutant.25 A pool of mutagenized plasmids was obtained from this strain and used to electroporate UU1250 cells. Transformants were inoculated in a streak on tryptone semisolid agar plates containing chloramphenicol (12.5 μg mL−1) and sodium salicylate (0.45 μM). After being incubated at 30 °C for 20−24 h, cells from outgrowing chemotaxis rings were picked, and colonies were purified on selective antibiotic medium. Mutagenized plasmids were purified, used to retransform UU1250 cells, and retested for suppressor properties in tryptone semisolid agar plates, 3853
DOI: 10.1021/acs.biochem.7b00455 Biochemistry 2017, 56, 3850−3862
Article
Biochemistry
Figure 2. Trimer and cluster formation capabilities of nonfunctional mutant Tsr proteins. (A) TMEA cross-linking competition assay. Tar-C trimers of dimers (Cys indicated by black dots) are engaged by the trifunctional cross-linker TMEA through their inner subunits. In this experiment, Tar-C dimers (light gray) are co-expressed with an excess of unmarked Tsr dimers. Trimer-proficient Tsr dimers (dark gray) compete for trimer formation, so they should form mixed trimers with Tar-C and eliminate or clearly reduce TMEA cross-linking products. Trimer-deficient Tsr molecules (white) should not interfere with the formation of Tar-C cross-linking products. (B) Representative results of the competition assay. UU1613 cells (TarS364C as the only receptor, cheAcheW, cheRcheB) carrying pCS12-I377P (NC, noncompetitor trimer-deficient Tsr mutant), pCS12 (C, competitor trimer-proficient wild-type Tsr), or pCS12 derivatives with different mutations were induced with 0 (−) or 1.2 μM (+) sodium salicylate, treated with TMEA, and analyzed by SDS−PAGE as described in Materials and Methods. (C) Representative results of a clustering assay. UU9352 cells (ΔMCPs, cheZ) co-transformed with pFG1 (YFP−CheZ) and empty vector pRR31 (tsr), pCS12 (Tsr), or its derivatives were grown to mid log phase. YFP−CheZ was induced at 100 μM IPTG, and receptors were induced at 0.45 μM sodium salicylate. Fluorescence images were obtained with a Nikon Eclipse Ti inverted microscope.
■
RESULTS A New Glance at the Conservation of the Coupling Subdomain in 36H Class Chemoreceptors. E. coli receptors have 126 residues (18 heptads) in each arm of the cytoplasmic hairpin and thus belong to the 36H class. Numbered from the tip to the membrane, the pairs of heptads (one from each helix) 1−4 constitute the signaling region, 5− 10 the coupling subdomain, and 11−18 the adaptation subdomain. Considering the large number of new genomes sequenced in the past several years, and taking into account the fact that the coupling region is the less conserved of the three subdomains but presents class-specific characteristics, we decided to reexamine the residue conservation of this subdomain within the 36H class. After collecting more than 22500 MCPs from 5830 bacterial reference proteomes available in the UniProt database, we applied the domain family boundaries defined by Alexander and Zhulin2 and retrieved receptors grouped according to their families (see Materials and Methods for details). We then built an alignment with 2428 36H class receptors and analyzed the coupling region. The alignment shows a notably high level of conservation of every residue that lies at position a or d, that is, residues that function primarily as knobs in the coiled coil, although they can also form parts of holes in adjacent packing layers (Figure 1B). One of the glycine residues that constitute the hinge belongs to this group (G340 in Tsr, position d, 76% frequency). The glycine adjacent to it (G341, position e in the coiled coil) shows a similar frequency (74%), whereas the C-helix glycine (G439
in Tsr, position e) is the most frequent one (87%). The general pattern is strikingly similar to what had been described by Alexander and Zhulin 2 upon an alignment of
