Structural analysis of CACHE domain of the McpA chemoreceptor from Leptospira interrogans

Abstract

Leptospira is a genus of spirochaete bacteria highly motile that includes pathogenic species responsible to cause leptospirosis disease. Chemotaxis and motility are required for Leptospira infectivity, pathogenesis, and invasion of bacteria into the host. In prokaryotes, the most common chemoreceptors are methyl-accepting chemotaxis proteins that have a role play to detect the chemical signals and move to a favorable environment for its survival. Here, we report the first crystal structure of CACHE domain of the methyl-accepting chemotaxis protein (McpA) of L. interrogans. The structural analysis showed that McpA adopts similar α/β architecture of several other bacteria chemoreceptors. We also found a typical dimerization interface that appears to be functionally crucial for signal transmission and chemotaxis. In addition to McpA structural analyses, we have identified homologous proteins and conservative functional regions using bioinformatics techniques. These results improve our understanding the relationship between chemoreceptor structures and functions of Leptospira species.

Graphical Abstract

1. Introduction

Most bacteria possess a diversity of signal transduction pathways able to respond to stimuli and move for favorable (attractants) or deviate from unfavorable (repellents) environments. A variety of signals can act on bacteria, these signals are integrated and translated into specific responses to temperature, osmotic pressure, nutrients, pH change, and light. When this movement is directed to chemicals, is called chemotaxis. Bacterial chemotaxis helps the bacteria to escape from challenges such as scarcity of nutrients and the accumulation of toxic substances. Chemotaxis is essential for many important biological processes in pathogenic bacteria, for example, virulence, adhesion, host colonization, and the establishment of symbiotic relationships [1,2].

For bacteria to survive in an always-changing environment, a complex signal transduction system is necessary, e.g. the chemoreceptors. The chemoreceptors are localized in bacterial membrane and form homodimers that are stable in the presence or absence of ligands [3]. Methyl-accepting chemotaxis proteins (MCPs) are the most common chemoreceptors in bacteria. There are different classes of MCPs, which are classified according to the ligand-binding domain (LBD), membrane topology, sequence length, and conservation of their cytoplasmic domains [4,5]. Three domains form a general MCP chemoreceptor: (i) a LBD, (ii) a transmembrane-helix domain, and (iii) a cytoplasmic signaling domain. The LBDs can be located in the cytoplasm or periplasmic region, and participate in the cellular recognition of various chemical and physical signals [5]. There are different families of LBDs that have been classified based on their structure, such as CACHE (calcium channels and chemotaxis) and PAS (Per-Arnt-Sim) domains [6].

Leptospira are spirochete bacteria comprising various species that cause the zoonotic leptospirosis, an important re-emerging infectious disease. It is estimated that more than 1.0 million Leptospira spp. infections with approximately 60,000 deaths occur annually worldwide [7]. The progression and clinical aspects of infection are well documented in the human and animal host, yet the molecular and cellular mechanisms of pathogenesis remain poorly understood [8]. In humans, a broad symptoms are observed in patients, ranging from asymptomatic to fever and headache for the subclinical and mild cases, to hemorrhages, jaundice, and multi-organ failure in the severe cases (Weil’s syndrome) [9]. Leptospira species are highly motile, the chemotaxis and motility are essential for infection and dissemination within the host. Mutant strains with reduced motility have lost their infectious activity. Leptospira motility depend on the presence of two periplasmic flagella localized between peptidoglycan and the outer membrane, each one extending from one extremity of the cell [10].

Chemotaxis has been poorly explored in leptospires, being mostly restricted to Borrelia species. The leptospiral chemoreceptors and chemotaxis behavior have started to be addressed only recently [11,12]. Like many bacteria, Leptospira use chemoreceptors to sense and respond to chemical signals in their environment [12,13]. An important example of the chemical response of Leptospira occurs in relation to hemoglobin. A study of Yuri and colleagues showed that pathogenic species are attracted towards hemoglobin, indicating that chemotaxis is an important factor in leptospiral infection [14]. Other compounds such as sucrose, glucose, pyruvate, and palmitate also displayed chemotactic attraction for Leptospira, but this behavior was not identical among pathogenic and saprophytic species [10,15].

Here, we present the first structure of the methyl-accepting chemotaxis protein (McpA) CACHE domain from the pathogenic bacterium L. interrogans. The structure revealed similarity to other bacterial chemoreceptors with an important role in motility and chemotaxis. In addition to McpA structural analysis, we also identified homologous proteins and conserved functional regions using bioinformatics techniques.

2. Materials and methods

2.1. Recombinant protein production

L. interrogans serovar Copenhageni strain FIOCRUZ L1–130 McpA (LIC12921; accession number AAS71473.1) was cloned, expressed, and purified by the Seattle Structural Genomics Center for Infectious Disease following standard protocols previously described [16]. Briefly, the gene encoding amino acids 43 – 304 of the McpA was PCR-amplified using a genomic DNA template. The amplified products were cloned into ligation-independent cloning expression vector pBG1861 [17]. The expression vector provides a non-cleavable N-terminal 6xHis-tag (target ID LpinA.18975.a.B2.GE42898). The pBG1861-McpA vector was transformed into chemically competent E. coli BL21-Gold (DE3)pLysS AG cells. Purification was carried out using immobilized metal affinity chromatography (IMAC) and size-exclusion chromatography (SEC). The purified protein was concentrated to 19.51 mg/mL using an Amicon purification system (Millipore, Burlington, MA) in buffer A (20 mM HEPES pH 7.0, 300 mM NaCl, 5% glycerol, 1 mM Tris(2-carboxyethyl)phosphine). The protein was flash frozen in liquid nitrogen and stored at −80 °C.

2.2. Crystallization

The purified McpA was crystallized using the set up as sitting-drop vapor-diffusion trials using XJR Junior crystallization trays (Rigaku Reagents) with drop sizes of 0.4 μL protein solution plus 0.4 μL well solution equilibrated against 80 μL reservoir. Crystallization conditions were searched for using the commercial screens MCSG-1 (Microyitic), JCSG+ (Rigaku Reagents) and Morpheus (Molecular Dimensions). All plates were stored at 14°C. Crystals grew from condition JCSG+ screen A7 (20% [w/v] PEG 8000, 100mM CHES pH 9.5). The crystals were cryoprotected by passing them into the respective reservoir solutions supplemented with 20% (v/v) ethylene glycol before flash freezing in liquid nitrogen. For phasing, crystals from condition Microlytic MCSG1 screen condition H3 (20% (w/v) PEG 3350, 200 mM Lithium acetate) which were soaked for 15 sec a mix of 90% reservoir and 10% 2.5M Nal in ethylene glycol, and for another 15 s in a mix of 80% reservoir and 20% 2.5M Nal in ethylene glycol and flash frozen for in-house data collection.

2.3. Data collection and processing

The data set for phasing was collected in-house on a Rigaku FR-E+ 007 SuperBright rotating anode equipped with Rikaku VariMax optics and a Saturn 944+ detector, using CuKα X-rays. The data set for refinement was collected at the APS LS-CAT beamline 21-ID-F equipped with a C(111) monochromator, and a Rayonix MX-300 detector. The data were indexed, integrated and scaled with the XDS package [18].

2.4. Structure determination, refinement and analysis

The McpA structure was determined by the single wavelength anomalous dispersion (SAD) method using iodide ions as the heavy atoms [19]. Iodide sites were found using HySS [20]. The initial Phases were improved using PARROT [21]. The initial model was built with ARP/wARP [22]. The final model was refined using Phenix [23] and model building in Coot [24]. The structure solution and refinement statistics are provided in Table 1. The atomic coordinates and native and anomalous amplitudes of McpA have been deposited in the Protein Data Bank (PDB entry 6PZJ). Structure figures were analyzed and prepared using Dali server [25], Espript [26], PISA [27] and PyMOL [28].

Table 1.

Data collection and refinement statistics.

Data Collection	native	iodide
Wavelength	0.97872 Å	1.5418 Å
Resolution rangea (Å)	38.416– 1.75 (1.80 – 1.75)	50–1.90 (1.95–1.90)_
Space group	P141212	P141212
Unit cell (a, b, c (Å); α, β, γ (°))	72.09 72.09 116.89; 90, 90, 90	72.90, 72,90, 116.94; 90 90 90
Unique reflections	31892 (2319)	45775(3091)*ano_
Redundancy	11.48 (7.45)	12.67 (4.78) *ano
Completeness (%)	100.00 (100.00)	99.9 (89.8) *ano
Mean I/sigma (I)	30.66 (3.18)	28.03 (2.99) *ano
Wilson B-factor	32.859	22.028
Rmerge	0.046 (0.599)	0.061 (0.508) *ano
CC1/2	1 (0.875)	100 (81.4) *ano
Structure Refinement Statistics
R-work/R-free	0.1652/0.2053
protein	2168
Ligands/ions	9
solvent	223
Protein residues	264
RMS (bonds)	0.009
RMS (angles) Ramachandran plot	1.24
Favored/allowed/outliers (%)	98.47/1.53/0.0
Rotamer outliers (%)	0.88
Clashscore	3.77
Average B-factor	33.02
PDB code	6PZJ

^aStatistics for the highest-resolution shell are shown in parentheses.

2.5. Bioinformatics

All the bioinformatics predictions and analyses were performed with the sequence available in the National Center for Biotechnology Information (NCBI) as a reference (accession number AAS71473.1). The presence of conserved domains and motifs were predicted in the webservers Myhits, SMART [29], and SignalP 5.0 [30]. The protein topology and the presence of transmembrane sections were predicted with the tools Octopus [31], TMHMM [32], and TOPCONS [33]. The subcellular localization of the McpA protein was predicted in PSORTb 3.0.2 [34].

To identify homologs, the sequence of McpA (LIC12921; AAS71473.1) was used as a seed in iterative profile searches with the Position-Specific Iterated BLAST (PSI-BLAST) program, running against the non-redundant (NR) protein database of NCBI. The search was performed either including or excluding Leptospiraceae (taxid: 170). The homologs above the threshold found in the searches including or excluding Leptospiraceae (494 and 473 proteins, respectively), were compared by multiple sequence alignments in the webserver PROMALS3D [35], which provides the consensus amino acid residues and secondary structure of the aligned protein sequences.

The PSI-BLAST search (excluding Leptospiraceae) resulting sequences were used as seed for multiple sequence alignment and Neighbour-joining Phylogenetic Tree construction without distance corrections in the MUSCLE Webserver [36]. The MUSCLE phylogenetic tree data was implemented in the Interactive Tree Of Life (iTOL) program [37], to construct a circular tree without branch lengths. The tree was manually curated to collapse branches and assign taxonomic groups based in the NCBI Taxonomy database.

3. Results and discussion

3.1. Bioinformatics analysis

MCP sequences typically consist of a sensory domain, a HAMP linker domain, and a signaling domain that interacts with the cytoplasmic signaling proteins. All domains were predicted in L. interrogans serovar Copenhageni McpA (LIC12921; accession number AAS71473.1): (i) prokaryotic membrane lipoprotein lipid attachment site (LPAM_1) (region 1 – 27), (ii) CACHE sensory domain (region 32 – 295), which bind senses and bind the ligands, (iii) HAMP domain (region 334 – 386), and (iv) methyl-accepting chemotaxis protein (MCP) signaling domain (MCPsignal) (region 474 – 692), which transduces the signal to downstream signaling proteins in the cytoplasm (Fig. 1A and Supplementary Table 1). The region corresponding to the three-dimensional structure of LIC12921 (region 43 – 304) comprises the majority of the sensory domain (Fig. 1A and Supplementary Table 1).

Figure 1. Bioinformatics prediction of L. interrogans serovar Copenhageni McpA conserved domains, topology and phylogenetic tree.

(A) Predicted McpA topology. TM: transmembrane section; CC: coiled coil. Schematic representation of McpA insertion into leptospiral inner membrane. In blue, the ligand-binding region (Cache domain). In orange, the signal transduction region, comprising the MCPsignal domain and two methylation regions, which form antiparallel alpha-helices embedded in the cytoplasm. The transmembrane sections (TM1 and TM2) flank the ligand-binding domain (LBD), which loops to the periplasmic space. (B). L. interrogans serovar Copenhageni McpA (LIC12921) homologs were searched through PSI-BLAST excluding the Leptospiraceae. The search resulted in 473 proteins from different microorganisms. A Neighbour-joining Phylogenetic Tree without distance corrections was constructed in the MUSCLE webserver. The phylogenetic tree data was edited in the iTOL software. The grey circles correspond to collapsed branches formed by more than one seed protein.

While the HAMP and signaling domains are always cytoplasmic in MCPs, the membrane topology of the sensory domain varies, allowing the assembling of MCPs into groups [38]. LIC12921 can be classified as a Class I MCP, characterized by a periplasmic sensory domain anchored by an N-terminal transmembrane (TM) helix and connected by an internal TM helix to the HAMP linker and signaling domains (Fig. 1A). Most MCPs have this sensory topology. The signaling-transduction region is flanked on both sides by methylation regions (α-helices) that contain potential sites of methylation and demethylation (Fig. 1A). The level of methylation controls the attractant or repellent signal to the motility apparatus.

As the specificity of the chemotaxis receptor is defined by the sensory domain, the MCP extracytoplasmic sensing domain sequences tend to be highly variable, whereas the cytoplasmic signaling domains are highly conserved. We performed a PSI-BLAST search and all the hits above the threshold belong to the genus Leptospira (494 proteins). Multiple sequence alignment of these proteins shows the high degree of conservation in the C-terminal region of the LIC12921 homologs, corresponding to the α-helix-rich MCPsignal domain (Supplementary Fig. 1). The N-terminal region, comprising signal peptides, lipidation regions, and sensory domains (3D structure fragment) presents a low level of amino acid conservation, while the secondary structure tends to be, as expected, similar (Supplementary Fig. 1).

A PSI-BLAST search excluding Leptospiraceae was also performed, to find McpA (LIC12921) homologs in other taxons. The 473 sequences above the threshold retrieved after 1 iteration search were aligned through PROMALS3D, and the result revealed a highly variable N-terminal region, a moderately conserved sensory domain, a large gap region without similarities, and a C-terminal region with a high identity fragment (consensus sequence pIs.lh.lIp.IA.QTNlLALNAhlEAARAGEpG+GFA....VVAsEVRpLAppot...Ah.-Ipph [where conserved amino acid residues are bold and uppercase letters, ‘l’ is an aliphatic residue, ‘h’ is a hydrophobic residue, ‘o’ is an alcohol residue, ‘p’ is a polar residue, ‘s’ is a small residue, - and + are negatively and positively-charged residues, respectively]) (Supplementary Fig. 2) [35].

The sequences of the 473 homologs (plus the L. interrogans McpA) were used as a seed to construct a Neighbour-joining Phylogenetic Tree. The resulting tree (Fig. 1B) was manually curated in order to collapse branches and assign taxonomy groups. The grey circles represent collapsed clades formed by more than one seed protein. The Terrabacteria group (comprising mostly Firmicutes) tended to form a cluster (Fig. 1B, yellow circles). The Proteobacteria (Alpha, Beta, Gamma, and Delta – yellow, green, red, and blue stars, respectively) also tended to cluster in a uniform branch. Noteworthy, the L. interrogans representative McpA (LIC12921) clustered in a heterogeneous group formed by all the McpA homologs seed proteins from the Spirochaetes, Thermotogae, Synergistaceae, and FCB groups (Bacteroidetes), in addition to some proteins from microorganisms belonging to Proteobacteria and Terrabacteria groups. The Spirochaete proteins which resulted from the PSI-BLAST search and were included in the phylogenetic analysis belong to several Treponema species and also unidentified bacteria. No hits of the genus Borrelia, another relevant spirochete pathogen, resulted from the search.

3.2. Overall structure of CACHE domain of the McpA

L. interrogans full-McpA is a signal transduction protein with three different domains (Fig. 1A): The N-terminal region represents the LBD, which corresponds to the periplasmatic ligand sensing region. The structure of the N-terminal McpA_43–300 of L. interrogans serovar Copenhageni strain FIOCRUZ L1–130 was determined to 1.75 Å resolution. The protein was solved in the tetragonal space group P4₁2₁2 with one molecule in the asymmetric unit. The final model contains residues 43–300, 221 water molecules, one chloride ion, and two ethylene glycol molecules from the cryoprotectant (Table 1). No electron density could be observed for the last four C-terminal residues. The McpA_43–300 structure revealed a dual CAlcium channels and CHEmotaxis receptor (dCACHE_1) domain with a mixed α/β-structure containing a long N-terminal α-helix (residues 43–72) followed by two subdomains of similar folds, although not identical (Fig. 2A). The upper subdomain is termed membrane-distal (α2-β5 - residues 73–206) and the lower is termed membrane-proximal (α6-β12 - residues 207–300) relative to the inner membrane. The distal subdomain consists of a seven-stranded antiparallel β-sheet and three α-helices. The β-sheet is curved and creates a pocket that corresponds to a possible region for ligand binding. The pocket has a loop between β3 and α4, which acts as the entrance door of the active site. The core of the proximal subdomain contains a five-stranded antiparallel β-sheet, flanked by three α-helices, as shown by the secondary-structure topology (Fig. 2A).

Figure 2. Overall fold and structural comparison of the McpA with other bacteria chemoreceptors.

(A) Ribbon representation of the McpA-LBD. α-helices and β-strands are represented as coils and arrows, respectively. The chloride ion is represented as a yellow sphere and 1,2-ethanediol is shown with the carbon atoms colored yellow, and oxygen atoms red. On top is represented membrane-distal domain and at the bottom membrane-proximal domain. The α1 in orange, belongs to both domains. The N- and C- terminal are labeled. The topology representation of the secondary structure elements. The a-helices are represented by rods and β-strands by arrows. (B) Cartoon representation of the CACHE_1 domain homodimer. The dimerization interface is formed mainly by hydrophilic residues of both distal and proximal domains.

Chemoreceptor proteins were described as homodimers [39]. To analyze the dimerization of leptospiral McpA, we performed a size-exclusion chromatography, in which McpA_43–300 migrated with an apparent molecular weight of 46 kDa (data not shown). However, this weight is smaller than expected for a dimer. Despite this difference, it is known that the dependence on the molecular weight on gel filtration size varies as a function of the shape of the macromolecule. The McpA structure revealed a significant dimeric interface similar to the observed for other structurally homologous bacterial chemoreceptors with the CACHE_1 domain [39]. This interface crystallographic contacts are formed by α1, α2, the loops between α1-α2 and β11-β−12. The interactions involve mainly the hydrophilic residues E47, R48, R65, S74, Q80, N92, and K286 as well as side chain of hydrophobic residues V91 (Fig. 2B). The dimer buried a total surface area of the 4385.1 Å and surface area of the complex 21150 Å. For many chemoreceptors, it has been proposed that dimerization is required to perform its biological function, as demonstrated for the family of histidine kinase sensor domains [39].

3.3. Structural comparisons

Based on a Dali search of the PDB, 23 nonredundant bacterial chemoreceptors with a global structural similarity (Z-score >10) were identified, despite a low sequence conservation between those receptors to McpA. Most of the matches are chemoreceptors with CACHE_1 domain suggesting that McpA adopts a conserved CACHE_1 overall topology (Fig. 2C and Fig. 3). Many of these structural homologs had identified ligands that include histamine, L-proline, L-arginine, cytosine, and proline betaine. These ligands occupied the same region of the distal domain (Supplementary Fig. 3). The Dali server revealed that McpA structure is highly similar to TlpQ chemoreceptor of P. aeruginosa with a Z-score of 30.7 and RMSD 1.9 Å (PDB 6FU4). TlpQ binds to histamine neurotransmitter in the membrane-distal subdomain. The residues E170, Y208, D210, and D239 play a central role in the recognition of histamine as a ligand to TlpQ chemoreceptor. In McpA structure, a similar ligand binding pocket was observed, with three out of four residues involved in TlpQ-histamine binding. The missing residue is D210, which was replaced by Y174 in McpA (Fig. 2D and Fig. 3). This substitution causes a steric impediment in the imidazole ring of the histamine, although the pocket maintains space for the binding of different ligands (Supplementary Fig. 4).

Figure 3. Structure-based sequence alignment of LBD-McpA with other bacteria chemoreceptors.

Secondary elements of LBD-McpA are noted above the alignment. Residues in red are higher conserved. Numbers correspond to the LBD-McpA sequence. The green triangles correspond to putative residues in active site from McpA based on comparison with TIpQ chemoreceptor.

Another identified chemoreceptor structure similar to McpA was the histidine kinase sensor of Methanosarcina mazei with Z-score of 28.8 and RMSD 1.9 Å (PDB 3LIB). The McpX of Sinorhizobium meliloti with Z-score of 26.5 and RMSD 2.6 Å have the proline betaine bound in the same region of histamine binding site from TlpQ, which suggests that this region is the possible binding site of the McpA. More distant structural homologues include Tlp1 chemoreceptor of Campylobacter jejuni - Z-score of 21.7 and RMSD 3.5 Å (PDB 4WY9) and histidine kinase sensor of Shewanella oneidensis - Z-score of 20.5 and RMSD 3.3 Å (PDB 3LIC). The best-ranking structures identified are available in the Supplementary Table 2.

Highlights.

• The first chemoreceptor structure from L. interrogans was determined to 1.75 Å

• McpA structure revealed a CACHE_1 domain with dimeric interface

• McpA structure showed some structural characteristics with bacterial chemoreceptors

• McpA protein showed evolution similarity with several microorganisms.