- Email: [email protected]
β-domain
Biochemical and Biophysical Research Communications 497 (2018) 368e373
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Crystal structure of a substrate-binding protein from Rhodothermus marinus reveals a single a/b-domain Ji-Eun Bae a, b, In Jung Kim c, Kyung-Jin Kim a, b, *, Ki Hyun Nam c, d, ** a
School of Life Sciences, KNU Creative BioResearch Group, Kyungpook National University, Daegu 41566, Republic of Korea KNU Institute for Microorganisms, Kyungpook National University, Daegu 41566, Republic of Korea c Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 02841, Republic of Korea d Institute of Life Science and Natural Resources, Korea University, Seoul 02841, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 February 2018 Accepted 8 February 2018 Available online 9 February 2018
Substrate-binding proteins (SBPs) bind to specific ligands and are associated with membrane protein complexes for transport or signal transduction. Most SBPs recognize substrates by the hinge motion between two distinct a/b domains. However, short SBP motifs are often observed in protein databases, which are located around methyl-accepting chemotaxis protein genes, but structural and functional studies have yet to be performed. Here, we report the crystal structure of an unusually small SBP from Rhodothermus marinus (named as RmSBP) at 1.9 Å. This protein is composed of a single a/b-domain, unlike general SBPs that have two distinct domains. RmSBP exhibits a high structural similarity to the Cterminal domain of the previously reported amino acid bound SBPs, while it does not contain an Nterminal domain for substrate recognition. As a result of the structural comparison analysis, RmSBP has a putative SBP that is different from the previously reported SBP. Our results provide insight into a new class of substrate recognition mechanism by the mini SBP protein. © 2018 Elsevier Inc. All rights reserved.
Keywords: Substrate-binding protein SBP ABC transport a/b-domain
1. Introduction The ATP-binding cassette (ABC) transporters are found in all kingdoms, which use the hydrolysis of ATP to translocate solutes across biological membranes [1e3]. In this system, the substratebinding protein (SBP) or covalently bonded substrate-binding domain (SBD) capture the solute and deliver the molecule for transfer by the transmembrane domain of the ABC transporter [4,5]. Among them, the SBP is a key determinant of substrate specificity and has a high affinity towards ABC uptake systems [6]. Many SBP-dependent ABC transporters recognize a broad range of ligands such as metal ions, amino acids, sugars, and peptides [6]. In general, substrate-bound SBPs are involved in the activation of the ABC transporter through interactions with the ABC transporter [7,8].
* Corresponding author. School of Life Sciences, KNU Creative BioResearch Group, Kyungpook National University, Daegu 41566, Republic of Korea. ** Corresponding author. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 02841, Republic of Korea. E-mail addresses: [email protected] (K.-J. Kim), [email protected] (K.H. Nam). https://doi.org/10.1016/j.bbrc.2018.02.086 0006-291X/© 2018 Elsevier Inc. All rights reserved.
SBPs have low or no amino acid sequence similarities but have highly conserved three-dimensional structural folds [7]. These SBPs commonly consist of two a/b domains and are connected by a hinge region, which comprises one to three interconnecting strands [5]. The two domains have an open conformation and a closed conformation when a specific substrate is bound between the two domains [8]. Based on the available structures from the Protein Data Bank (PDB), SBP shows four distinct structural states of (i) open-unliganded (ii) open-liganded (iii) closed-unliganded, and (iv) closed-liganded [7]. In the absence of substrates, most SBPs have open-unliganded states and some have closed-unliganded states [9,10], whereas the equilibrium between the open and closed conformation shifts towards the closed-liganded state in the presence of a substrate [7]. Based on the features of the threedimensional structures, SBPs were divided into seven structural classes (cluster AeG) [5]. In these clusters, the SBPs have various sizes, ranging between 25 and 70 kDa and commonly have two a/b domains [7]. During a protein domain analysis using Pfam and InterPro, we found a 17.96 kDa (164 amino acids excluding the signal peptide) SBP domain from Rhodothermus marinus. This protein is smaller than previously known SBP family members. Although the analysis
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
of the primary structure predicted that this protein is part of the SBP family, there was a lack of information for understanding the functional relationship with general SBP proteins. To better understand the unusually small SBP, we have determined the crystal structure of the small SBP from Rhodobacter marinus (named as RmSBP) with a resolution of 1.9 Å. This protein has a single a/b-domain, which was structurally completely different compared with previously reported SBPs with two domains. Our structural data provided here for mini SBP now reveals more structural and functional diversity of the SBP family.
2. Materials and methods 2.1. Protein preparation The gene encoding RmSBP (UniProt entry: D0MDR1, residues 22e185) was cloned between the NdeI and XhoI sites of the pET28a vector (Novagen). The recombinant DNA was transformed into the Escherichia coli BL21 (DE3) strain. Expression of the recombinant protein was induced by 0.5 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) at 18 C for overnight. The cells were harvested by centrifugation and then resuspended in a buffer containing 50 mM Tris-HCl, pH 8.0, 200 mM NaCl, and 20 mM imidazole. After sonication, cell debris was removed by centrifugation. The lysate supernatant was loaded onto a Ni-NTA column (QIAGEN) and washed by buffer A. The proteins were eluted by a buffer containing 50 mM Tris-HCl, pH 8.0, 200 mM NaCl, and 300 mM imidazole. The protein was incubated overnight at 4 C with thrombin protease to remove the N-terminal hexahistidine-tag. Furthermore, this protein was concentrated and purified by size-exclusion chromatography on a HiLoad Sephacryl 100 16/60 column (GE Healthcare) via AKTA start (GE Healthcare) with buffer containing 10 mM Tris-HCl, pH 8.0 and 200 mM NaCl. The selenomethionine (Se-Met)-substituted RmSBP protein was expressed in E. coli BL21 (DE3) (Novagen) using the SeMet expression kit. The Se-Met protein was purified using a protocol as described for native protein purification.
369
2.2. Crystallization The recombinant RmSBP was crystallized by the hanging drop vapor diffusion method at 22 C. The protein solution (20 mg/ml) was mixed with an equal volume of a reservoir solution containing 0.1 M sodium acetate, pH 4.5, 0.2 M MgCl2, and 8e13% (w/v) PEG3350. Crystals were grown to 0.2 0.2 0.4 mm within a month. 2.3. Data collection Crystals were soaked in reservoir solution containing an additional 20% (v/v) glycerol and flash cooled in liquid nitrogen. X-ray diffraction datasets for the crystals were collected at 100 K on the beamline 7A at PLS-II (Pohang, Republic of Korea) using an ADSC Quantum Q270 CCD detector at a wavelength of 0.9794 Å. The diffraction datasets were processed and scaled using the HKL2000 program [11]. 2.4. Structure determination Initial phases were obtained with a Se-Met crystal diffracting to 1.9 Å by the Se single-wavelength anomalous dispersion (SAD) method using PHENIX AutoSol [12]. The structure model was automatically built using PHENIX AutoBuild [13], followed by manual model building using COOT [14]. The built model was refined by the program PHENIX [12]. The structural quality was verified by MolProbity [15]. The statistics of the refined model are summarized in Table 1. 2.5. Protein structure analysis Short length SBP protein sequences were retrieved from the Pfam [16] and InterPro [17] databases. The signal peptide and cellular location were identified by SignalP 4.1 [18]. For the in silico analysis, the protein-protein interaction was analyzed by STRING [19]. Sequence searches for close homologs were performed with
Table 1 Data collection and refinement statistics for RmSBP. Data collection Space group Cell dimensions a, b, c (Å) a, b , g ( ) Resolution (Å) Completeness Redundancy I/s(I) Rmergea Refinement statistics Resolution (Å) Rwork/Rfree (%)b B-factor (Averaged) Protein Water R.m.s deviations Bond lengths (Å) Bond angles ( ) Ramachandran plot (%) favored allowed
C2 63.90, 63.32, 34.67 90.00, 97.22, 90.00 50.0e1.90 (1.93e1.90) 98.0 (97.8) 6.6 (7.2) 48.38 (31.85) 13.3 (38.1) 44.80e1.90 15.37/19.84 12.74 23.80 0.023 1.139 98.7 1.3
Highest resolution shell is shown in parentheses. a Rmerge ¼ ShSijIi(hkl)_j/ShSiIi(hkl), where Ii(hkl) is the intensity of the ‘ith’ measurement of reflection hkl and is the weighted mean of all measurements of hkl. b Rwork ¼ SjjFobsj-jFcalcjj/SjFobsj, where Fobs and Fcalc are the observed and calculated structure-factor amplitudes respectively. Rfree was calculated as Rwork using a randomly selected subset (5%) of unique reflections not used for structure refinement.
370
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
BLAST [20] against the nonredundant sequence database obtained from Uniprot (http://www.uniprot.org/). The structure-based sequence alignment was carried out using Clustal Omega [21] and ESPRIPT [22]. Surface conservation was calculated and illustrated using the Consurf server [23]. Protein structure comparison was performed using the DALI server [24]. Figures of the structure were prepared using PyMOL (https://pymol.org/). The coordinates and structure factors of the RmSBP have been deposited in the PDB with accession code 5Z6V. 3. Result and discussion 3.1. Primary structure analysis In the gram-negative bacteria Rhodothermus marinus, an unusually small substrate-binding protein (gene name: Rmar_2176) was found by a search through the Pfam and InterPro domain databases. This protein consists of 185 amino acids and was annotated as an ABC-type uncharacterized transport system periplasmic component-like protein on UniProt (http://www.uniprot.org/ uniprot/D0MDR1). The SBPs of gram-negative bacteria are located in the periplasm [1], indicating that RmSBP contains the signal peptide. Analysis of the primary structure using the SignalIP4.1 [18] suggests that this protein possesses a signal peptide that cleaves between the Phe21 and Pro22 residues (Fig. S1). To analyze the protein interactions of RmSBP in silico, we used ‘standard’ for the Search Tool in the Retrieval of Interacting Genes (STRING) database. This result predicts that RmSBP interacts with the neighbor gene (Rmar_2177), which was annotated as a methyl-accepting chemotaxis protein (named as RmMPC) (Fig. S2A). This MPC protein is a family of receptors in bacteria that mediates chemotaxis to diverse signals, and contains a HAMP domain, transmembrane regions, and methyl-accepting transducer domain [25]. Moreover, 9 other ORFs
were predicted to have protein-protein interactions with RmSBP (Table S1). Next, we searched for homologs of RmSBP using BLAST. This search revealed 4 total homologous full-length SBP proteins, with a sequence identify ranging from 38.6% to 82.2% and an e-value below 7e-33 (Fig. S2B). Most of the homologous sequences were annotated as putative ABC transport substrate-binding protein or hypothetical proteins. The closest sequence to that of RmSBP was that of a putative ABC transport system substrate-binding protein (gene name: SAMN04488087_0329, named as RpSBP) from Rhodothermus profundi with a sequence identity of 82.2%. This RpSBP also has a neighboring gene encoding a methyl-accepting chemotaxis protein (SAMN04488087_0330, named as RpMPC), similar to that encoded by the RmSBP gene. The sequence identity between RmMPC and RpMPC was 88.4%. Although other homologous proteins with RmSBP had a sequence identify below 38.7%, these SBPs also had MCP in a neighboring gene cluster, with sequence identities ranging from 45.1% to 51.8% (Fig. S2A). Taken together, short sequence SBPs commonly have a neighboring MPC gene. 3.2. Overall structure The RmSBP protein lacking its signal peptide was overexpressed in E. coli and was eluted as a monomer (18 kDa) via size exclusion chromatography. The crystal belongs to the monoclinic C2 space group with unit-cell dimensions of a ¼ 86.594 Å, b ¼ 87.431 Å, and c ¼ 110.574 Å, with one molecule occupying the asymmetric unit. The initial electron density map was obtained by single-wavelength anomalous diffraction (SAD) with anomalous signals of three selenomethionines in a single domain. The electron density of the RmSBP structure was well defined for 158 amino acid residues from Glu27eMet184, excluding the N- and C-terminus. The final model was refined to a resolution of 1.9 Å, resulting in an Rwork of 15.6 and
Fig. 1. Overall structure of RmSBP. (A) Cartoon representation (left) and the backbone temperature factors (right) of RmSBP. (B) A topological diagram of RmSBP. The a-helices and b-strands are indicated by cylinders and arrows, respectively.
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
an Rfree of 19.8. The RmSBP consisted of six a-helices, one 310-helix, and seven b-stands, and an a/b fold was adopted with an approximate dimension of 25 Å 25 Å 30 Å (Fig. 1A and Fig. 2). The strands of the b-sheet, with the exception of b6 and b7, had a parallel arrangement, forming a central core with all five strands surrounded by the four a-helices (Fig. 1B). The interior between the a-helices and the b-sheet was stabilized by numerous hydrophobic interactions (Fig. S3) and showed structural rigidity, whereas the surface helices exhibited a relatively high temperature B-factor (Fig. 1A). Taken together, RmSBP was composed of a/b domains similar to those of the existing SBPs but instead consisting of only one domain. 3.3. Structural comparison with other SBPs The DALI server was used to search for structurally homologous proteins. This search revealed 857 candidates that show a z score above 2.0 (41 for z > 10.0, 5 for z > 15). In particular, the RmSBP protein revealed a high Z-score to the ABC domain from Sreptococcus pneumonia (PDB code: 3LFT, Z score ¼ 19.0, sequence identity ¼ 21%, named as SpSBP) and an uncharacterized protein from Vibrio cholerae serotype O1 (PDB code: 3LKV, Z score ¼ 18.9, sequence identity ¼ 30%, named as VcSBP). However, since SpSBP (Uniprot entry A0A0H2UPV6, 303 amino acids excluding 41 amino acids from the signal peptide) and VcSBP (Unprot entry Q9KT04,
371
296 amino acid excluding 25 amino acids from the signal peptide) were composed of two a/b domains as in general SBPs, they were distinguished from the RmSBP having a single a/b domain (Fig. 3A). Nevertheless, in order to compare the similarity between these proteins, we performed a superimposition and confirmed that RmSBP had a highly similar C-terminal domain with that of SpSBP (r.m.s. devation of 2.3 Å for 149 Ca-atoms) and VcSBP (2.5 Å for 149 Ca-atoms) (Fig. 3B). In order to obtain insight into the putative substrate recognition of RmSBP, we further analyzed the substrate-binding site of the SpSBP and VcSBP structures complexed with L-tryptophan and Lphenylalanine, respectively, showing the closed conformation. In SpSBP, the L-tryptophan substrate interacted with seven amino acids (Gln284, His61, Ser63, Leu64, Ala138, Thr140, and Asp163) in the N-terminal domain and two amino acids (Ile242 and Asn244) in the C-terminal domain (Fig. 3C and Fig. S4A). In VcSBP, the phenylalanine substrate interacted with six amino acids (His38, Ala40, Thr97, Ala116, Thr118, and Asp141) in the N-terminal domain and two amino acids (Ile220 and Asn222) in the C-terminal domain (Fig. 3C and Fig. S4B). In the C-terminal domain of SpSBP and VcSBP, the substrate interacted with the Ile and Asn residues, respectively, via hydrophobic interactions and the carbonyl-group of the substrate, respectively. Each amino acid was in the same structural position, and the substrate was on a hydrophobic surface. In contrast, RmSBP have Glu107 and Ser109 at the same positions,
Fig. 2. Conservation and electrostatic surface analysis of RmSBP. (A) Surface conservation map of RmSBP. The surface is colored according to sequence conservation from most conserved (purple) to the most variable (cyan) based on an alignment of short sequence SBP from 5 other homologous proteins in Fig. S2B. A putative substrate-binding site is indicated by a dotted red arrow. (B) Electrostatic analysis of RmSBP revealed a highly charged region on the surface. The putative substrate-binding site is indicated by a dotted yellow arrow. All figures are displayed in the same orientation in (AeB). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
372
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
Fig. 3. Comparison of the overall structure of RmSBP with SpSBP (PDB code: 3LFT) and VsSBP (3LKV). (A) RmSBP has a single a/b domain, whereas other SPBs have two a/b domains. (B) Superimposition of the RmSBP (green) with the C-terminal domain of SpSBP (blue) and VcSBP (orange). (C) Detailed view of the putative substrate-binding surface of RmSBP. In SpSBP and VcSBP, the Ile and Asn residues interact with the substrate on a hydrophobic surface, whereas RmSBP has Glu and Ser residues in the same position with a negatively charged surface. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
which have a negatively charged surface (Fig. 3C). As a result, the RmSBP and closest structures provided by the DALI server have different characteristics, not only in the number of domains, but also in the potential substrate recognition surfaces. 4. Discussion We confirmed RmSBP to be a short ABC transport SBP and that it was located near the MCP gene. This protein is composed of one a/ b-domain, unlike other SBP proteins, which are conventionally composed of two domains. Therefore, it was expected that RmSBP would have a different mechanism of substrate binding corresponding to the conformational change caused by the hinge motion of the two a/b domains. We believed that there would be two possible substrate recognition mechanisms. First, RmSBP alone
would be capable of substrate recognition. Secondly, RmSBP and its unknown partner protein recognize substrates such as general SBP consisting of two domains. In this case, RmSBP and a partner protein will not use the hinge motion in recognizing the substrate. In order to understand the exact molecular function, further studies with RmSBP are required for identifying its substrates and the mechanism by which the substrates are recognized. Moreover, it will also be necessary to investigate whether RmSBP is involved in the general ABC transport system or related to neighbor MCP genes. Nevertheless, our structural studies of RmSBP will provide insights into understanding the new type of SBP family. Acknowledgments We thank the beamline staff at the MX beamlines at PLS-II at
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
Pohang Acceleratory Laboratory for assistance in data collection. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MOE) (NRF2017R1D1A1B03033087 and 2017M3A9F6029736). Transparency document Transparency document related to this article can be found online at https://doi.org/10.1016/j.bbrc.2018.02.086. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.bbrc.2018.02.086. References [1] T. van der Heide, B. Poolman, ABC transporters: one, two or four extracytoplasmic substrate-binding sites? EMBO Rep. 3 (2002) 938e943. [2] C.F. Higgins, ABC transporters: from microorganisms to man, Annu. Rev. Cell Biol. 8 (1992) 67e113. [3] K.P. Locher, Mechanistic diversity in ATP-binding cassette (ABC) transporters, Nat. Struct. Mol. Biol. 23 (2016) 487e493. [4] F. Fulyani, G.K. Schuurman-Wolters, A.V. Zagar, A. Guskov, D.J. Slotboom, B. Poolman, Functional diversity of tandem substrate-binding domains in ABC transporters from pathogenic bacteria, Structure 21 (2013) 1879e1888. [5] G.H. Scheepers, A.N.J.A. Lycklama, B. Poolman, An updated structural classification of substrate-binding proteins, FEBS Lett. 590 (2016) 4393e4401. [6] A. Maqbool, R.S. Horler, A. Muller, A.J. Wilkinson, K.S. Wilson, G.H. Thomas, The substrate-binding protein in bacterial ABC transporters: dissecting roles in the evolution of substrate specificity, Biochem. Soc. Trans. 43 (2015) 1011e1017. [7] R.P. Berntsson, S.H. Smits, L. Schmitt, D.J. Slotboom, B. Poolman, A structural classification of substrate-binding proteins, FEBS Lett. 584 (2010) 2606e2617. [8] E. Biemans-Oldehinkel, M.K. Doeven, B. Poolman, ABC transporter architecture and regulatory roles of accessory domains, FEBS Lett. 580 (2006) 1023e1035. [9] G.A. Bermejo, M.P. Strub, C. Ho, N. Tjandra, Ligand-free open-closed transitions of periplasmic binding proteins: the case of glutamine-binding protein, Biochemistry 49 (2010) 1893e1902. [10] C. Tang, C.D. Schwieters, G.M. Clore, Open-to-closed transition in apo maltosebinding protein observed by paramagnetic NMR, Nature 449 (2007) 1078e1082. [11] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Meth. Enzymol. 276 (1997) 307e326.
373
[12] P.D. Adams, P.V. Afonine, G. Bunkoczi, V.B. Chen, I.W. Davis, N. Echols, J.J. Headd, L.W. Hung, G.J. Kapral, R.W. Grosse-Kunstleve, A.J. McCoy, N.W. Moriarty, R. Oeffner, R.J. Read, D.C. Richardson, J.S. Richardson, T.C. Terwilliger, P.H. Zwart, PHENIX: a comprehensive Python-based system for macromolecular structure solution, Acta Crystallogr D Biol Crystallogr 66 (2010) 213e221. [13] T.C. Terwilliger, R.W. Grosse-Kunstleve, P.V. Afonine, N.W. Moriarty, P.H. Zwart, L.W. Hung, R.J. Read, P.D. Adams, Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard, Acta Crystallogr. Sect. D Biol. Crystallogr. 64 (2008) 61e69. [14] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr D Biol Crystallogr 60 (2004) 2126e2132. [15] V.B. Chen, W.B. Arendall 3rd, J.J. Headd, D.A. Keedy, R.M. Immormino, G.J. Kapral, L.W. Murray, J.S. Richardson, D.C. Richardson, MolProbity: all-atom structure validation for macromolecular crystallography, Acta Crystallogr D Biol Crystallogr 66 (2010) 12e21. [16] R.D. Finn, P. Coggill, R.Y. Eberhardt, S.R. Eddy, J. Mistry, A.L. Mitchell, S.C. Potter, M. Punta, M. Qureshi, A. Sangrador-Vegas, G.A. Salazar, J. Tate, A. Bateman, The Pfam protein families database: towards a more sustainable future, Nucleic Acids Res. 44 (2016) D279eD285. [17] A. Mitchell, H.Y. Chang, L. Daugherty, M. Fraser, S. Hunter, R. Lopez, C. McAnulla, C. McMenamin, G. Nuka, S. Pesseat, A. Sangrador-Vegas, M. Scheremetjew, C. Rato, S.Y. Yong, A. Bateman, M. Punta, T.K. Attwood, C.J. Sigrist, N. Redaschi, C. Rivoire, I. Xenarios, D. Kahn, D. Guyot, P. Bork, I. Letunic, J. Gough, M. Oates, D. Haft, H. Huang, D.A. Natale, C.H. Wu, C. Orengo, I. Sillitoe, H. Mi, P.D. Thomas, R.D. Finn, The InterPro protein families database: the classification resource after 15 years, Nucleic Acids Res. 43 (2015) D213eD221. [18] H. Nielsen, Predicting secretory proteins with SignalP, Meth. Mol. Biol. 1611 (2017) 59e73. [19] D. Szklarczyk, A. Franceschini, S. Wyder, K. Forslund, D. Heller, J. HuertaCepas, M. Simonovic, A. Roth, A. Santos, K.P. Tsafou, M. Kuhn, P. Bork, L.J. Jensen, C. von Mering, STRING v10: protein-protein interaction networks, integrated over the tree of life, Nucleic Acids Res. 43 (2015) D447eD452. [20] R. Lavigne, D. Seto, P. Mahadevan, H.W. Ackermann, A.M. Kropinski, Unifying classical and molecular taxonomic classification: analysis of the Podoviridae using BLASTP-based tools, Res. Microbiol. 159 (2008) 406e414. [21] F. Sievers, D.G. Higgins, Clustal omega, Curr. Protoc. Bioinformatics 48 (2014) 11e16, 3 13. [22] P. Gouet, E. Courcelle, D.I. Stuart, F. Metoz, ESPript: analysis of multiple sequence alignments in PostScript, Bioinformatics 15 (1999) 305e308. [23] H. Ashkenazy, E. Erez, E. Martz, T. Pupko, N. Ben-Tal, ConSurf 2010: calculating evolutionary conservation in sequence and structure of proteins and nucleic acids, Nucleic Acids Res. 38 (2010) W529eW533. [24] L. Holm, P. Rosenstrom, Dali server: conservation mapping in 3D, Nucleic Acids Res. 38 (2010) W545eW549. [25] P. Derr, E. Boder, M. Goulian, Changing the specificity of a bacterial chemoreceptor, J. Mol. Biol. 355 (2006) 923e932.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Crystal structure of a substrate-binding protein from Rhodothermus marinus reveals a single a/b-domain Ji-Eun Bae a, b, In Jung Kim c, Kyung-Jin Kim a, b, *, Ki Hyun Nam c, d, ** a
School of Life Sciences, KNU Creative BioResearch Group, Kyungpook National University, Daegu 41566, Republic of Korea KNU Institute for Microorganisms, Kyungpook National University, Daegu 41566, Republic of Korea c Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 02841, Republic of Korea d Institute of Life Science and Natural Resources, Korea University, Seoul 02841, Republic of Korea b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 February 2018 Accepted 8 February 2018 Available online 9 February 2018
Substrate-binding proteins (SBPs) bind to specific ligands and are associated with membrane protein complexes for transport or signal transduction. Most SBPs recognize substrates by the hinge motion between two distinct a/b domains. However, short SBP motifs are often observed in protein databases, which are located around methyl-accepting chemotaxis protein genes, but structural and functional studies have yet to be performed. Here, we report the crystal structure of an unusually small SBP from Rhodothermus marinus (named as RmSBP) at 1.9 Å. This protein is composed of a single a/b-domain, unlike general SBPs that have two distinct domains. RmSBP exhibits a high structural similarity to the Cterminal domain of the previously reported amino acid bound SBPs, while it does not contain an Nterminal domain for substrate recognition. As a result of the structural comparison analysis, RmSBP has a putative SBP that is different from the previously reported SBP. Our results provide insight into a new class of substrate recognition mechanism by the mini SBP protein. © 2018 Elsevier Inc. All rights reserved.
Keywords: Substrate-binding protein SBP ABC transport a/b-domain
1. Introduction The ATP-binding cassette (ABC) transporters are found in all kingdoms, which use the hydrolysis of ATP to translocate solutes across biological membranes [1e3]. In this system, the substratebinding protein (SBP) or covalently bonded substrate-binding domain (SBD) capture the solute and deliver the molecule for transfer by the transmembrane domain of the ABC transporter [4,5]. Among them, the SBP is a key determinant of substrate specificity and has a high affinity towards ABC uptake systems [6]. Many SBP-dependent ABC transporters recognize a broad range of ligands such as metal ions, amino acids, sugars, and peptides [6]. In general, substrate-bound SBPs are involved in the activation of the ABC transporter through interactions with the ABC transporter [7,8].
* Corresponding author. School of Life Sciences, KNU Creative BioResearch Group, Kyungpook National University, Daegu 41566, Republic of Korea. ** Corresponding author. Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 02841, Republic of Korea. E-mail addresses: [email protected] (K.-J. Kim), [email protected] (K.H. Nam). https://doi.org/10.1016/j.bbrc.2018.02.086 0006-291X/© 2018 Elsevier Inc. All rights reserved.
SBPs have low or no amino acid sequence similarities but have highly conserved three-dimensional structural folds [7]. These SBPs commonly consist of two a/b domains and are connected by a hinge region, which comprises one to three interconnecting strands [5]. The two domains have an open conformation and a closed conformation when a specific substrate is bound between the two domains [8]. Based on the available structures from the Protein Data Bank (PDB), SBP shows four distinct structural states of (i) open-unliganded (ii) open-liganded (iii) closed-unliganded, and (iv) closed-liganded [7]. In the absence of substrates, most SBPs have open-unliganded states and some have closed-unliganded states [9,10], whereas the equilibrium between the open and closed conformation shifts towards the closed-liganded state in the presence of a substrate [7]. Based on the features of the threedimensional structures, SBPs were divided into seven structural classes (cluster AeG) [5]. In these clusters, the SBPs have various sizes, ranging between 25 and 70 kDa and commonly have two a/b domains [7]. During a protein domain analysis using Pfam and InterPro, we found a 17.96 kDa (164 amino acids excluding the signal peptide) SBP domain from Rhodothermus marinus. This protein is smaller than previously known SBP family members. Although the analysis
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
of the primary structure predicted that this protein is part of the SBP family, there was a lack of information for understanding the functional relationship with general SBP proteins. To better understand the unusually small SBP, we have determined the crystal structure of the small SBP from Rhodobacter marinus (named as RmSBP) with a resolution of 1.9 Å. This protein has a single a/b-domain, which was structurally completely different compared with previously reported SBPs with two domains. Our structural data provided here for mini SBP now reveals more structural and functional diversity of the SBP family.
2. Materials and methods 2.1. Protein preparation The gene encoding RmSBP (UniProt entry: D0MDR1, residues 22e185) was cloned between the NdeI and XhoI sites of the pET28a vector (Novagen). The recombinant DNA was transformed into the Escherichia coli BL21 (DE3) strain. Expression of the recombinant protein was induced by 0.5 mM isopropyl b-D-1-thiogalactopyranoside (IPTG) at 18 C for overnight. The cells were harvested by centrifugation and then resuspended in a buffer containing 50 mM Tris-HCl, pH 8.0, 200 mM NaCl, and 20 mM imidazole. After sonication, cell debris was removed by centrifugation. The lysate supernatant was loaded onto a Ni-NTA column (QIAGEN) and washed by buffer A. The proteins were eluted by a buffer containing 50 mM Tris-HCl, pH 8.0, 200 mM NaCl, and 300 mM imidazole. The protein was incubated overnight at 4 C with thrombin protease to remove the N-terminal hexahistidine-tag. Furthermore, this protein was concentrated and purified by size-exclusion chromatography on a HiLoad Sephacryl 100 16/60 column (GE Healthcare) via AKTA start (GE Healthcare) with buffer containing 10 mM Tris-HCl, pH 8.0 and 200 mM NaCl. The selenomethionine (Se-Met)-substituted RmSBP protein was expressed in E. coli BL21 (DE3) (Novagen) using the SeMet expression kit. The Se-Met protein was purified using a protocol as described for native protein purification.
369
2.2. Crystallization The recombinant RmSBP was crystallized by the hanging drop vapor diffusion method at 22 C. The protein solution (20 mg/ml) was mixed with an equal volume of a reservoir solution containing 0.1 M sodium acetate, pH 4.5, 0.2 M MgCl2, and 8e13% (w/v) PEG3350. Crystals were grown to 0.2 0.2 0.4 mm within a month. 2.3. Data collection Crystals were soaked in reservoir solution containing an additional 20% (v/v) glycerol and flash cooled in liquid nitrogen. X-ray diffraction datasets for the crystals were collected at 100 K on the beamline 7A at PLS-II (Pohang, Republic of Korea) using an ADSC Quantum Q270 CCD detector at a wavelength of 0.9794 Å. The diffraction datasets were processed and scaled using the HKL2000 program [11]. 2.4. Structure determination Initial phases were obtained with a Se-Met crystal diffracting to 1.9 Å by the Se single-wavelength anomalous dispersion (SAD) method using PHENIX AutoSol [12]. The structure model was automatically built using PHENIX AutoBuild [13], followed by manual model building using COOT [14]. The built model was refined by the program PHENIX [12]. The structural quality was verified by MolProbity [15]. The statistics of the refined model are summarized in Table 1. 2.5. Protein structure analysis Short length SBP protein sequences were retrieved from the Pfam [16] and InterPro [17] databases. The signal peptide and cellular location were identified by SignalP 4.1 [18]. For the in silico analysis, the protein-protein interaction was analyzed by STRING [19]. Sequence searches for close homologs were performed with
Table 1 Data collection and refinement statistics for RmSBP. Data collection Space group Cell dimensions a, b, c (Å) a, b , g ( ) Resolution (Å) Completeness Redundancy I/s(I) Rmergea Refinement statistics Resolution (Å) Rwork/Rfree (%)b B-factor (Averaged) Protein Water R.m.s deviations Bond lengths (Å) Bond angles ( ) Ramachandran plot (%) favored allowed
C2 63.90, 63.32, 34.67 90.00, 97.22, 90.00 50.0e1.90 (1.93e1.90) 98.0 (97.8) 6.6 (7.2) 48.38 (31.85) 13.3 (38.1) 44.80e1.90 15.37/19.84 12.74 23.80 0.023 1.139 98.7 1.3
Highest resolution shell is shown in parentheses. a Rmerge ¼ ShSijIi(hkl)_
370
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
BLAST [20] against the nonredundant sequence database obtained from Uniprot (http://www.uniprot.org/). The structure-based sequence alignment was carried out using Clustal Omega [21] and ESPRIPT [22]. Surface conservation was calculated and illustrated using the Consurf server [23]. Protein structure comparison was performed using the DALI server [24]. Figures of the structure were prepared using PyMOL (https://pymol.org/). The coordinates and structure factors of the RmSBP have been deposited in the PDB with accession code 5Z6V. 3. Result and discussion 3.1. Primary structure analysis In the gram-negative bacteria Rhodothermus marinus, an unusually small substrate-binding protein (gene name: Rmar_2176) was found by a search through the Pfam and InterPro domain databases. This protein consists of 185 amino acids and was annotated as an ABC-type uncharacterized transport system periplasmic component-like protein on UniProt (http://www.uniprot.org/ uniprot/D0MDR1). The SBPs of gram-negative bacteria are located in the periplasm [1], indicating that RmSBP contains the signal peptide. Analysis of the primary structure using the SignalIP4.1 [18] suggests that this protein possesses a signal peptide that cleaves between the Phe21 and Pro22 residues (Fig. S1). To analyze the protein interactions of RmSBP in silico, we used ‘standard’ for the Search Tool in the Retrieval of Interacting Genes (STRING) database. This result predicts that RmSBP interacts with the neighbor gene (Rmar_2177), which was annotated as a methyl-accepting chemotaxis protein (named as RmMPC) (Fig. S2A). This MPC protein is a family of receptors in bacteria that mediates chemotaxis to diverse signals, and contains a HAMP domain, transmembrane regions, and methyl-accepting transducer domain [25]. Moreover, 9 other ORFs
were predicted to have protein-protein interactions with RmSBP (Table S1). Next, we searched for homologs of RmSBP using BLAST. This search revealed 4 total homologous full-length SBP proteins, with a sequence identify ranging from 38.6% to 82.2% and an e-value below 7e-33 (Fig. S2B). Most of the homologous sequences were annotated as putative ABC transport substrate-binding protein or hypothetical proteins. The closest sequence to that of RmSBP was that of a putative ABC transport system substrate-binding protein (gene name: SAMN04488087_0329, named as RpSBP) from Rhodothermus profundi with a sequence identity of 82.2%. This RpSBP also has a neighboring gene encoding a methyl-accepting chemotaxis protein (SAMN04488087_0330, named as RpMPC), similar to that encoded by the RmSBP gene. The sequence identity between RmMPC and RpMPC was 88.4%. Although other homologous proteins with RmSBP had a sequence identify below 38.7%, these SBPs also had MCP in a neighboring gene cluster, with sequence identities ranging from 45.1% to 51.8% (Fig. S2A). Taken together, short sequence SBPs commonly have a neighboring MPC gene. 3.2. Overall structure The RmSBP protein lacking its signal peptide was overexpressed in E. coli and was eluted as a monomer (18 kDa) via size exclusion chromatography. The crystal belongs to the monoclinic C2 space group with unit-cell dimensions of a ¼ 86.594 Å, b ¼ 87.431 Å, and c ¼ 110.574 Å, with one molecule occupying the asymmetric unit. The initial electron density map was obtained by single-wavelength anomalous diffraction (SAD) with anomalous signals of three selenomethionines in a single domain. The electron density of the RmSBP structure was well defined for 158 amino acid residues from Glu27eMet184, excluding the N- and C-terminus. The final model was refined to a resolution of 1.9 Å, resulting in an Rwork of 15.6 and
Fig. 1. Overall structure of RmSBP. (A) Cartoon representation (left) and the backbone temperature factors (right) of RmSBP. (B) A topological diagram of RmSBP. The a-helices and b-strands are indicated by cylinders and arrows, respectively.
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
an Rfree of 19.8. The RmSBP consisted of six a-helices, one 310-helix, and seven b-stands, and an a/b fold was adopted with an approximate dimension of 25 Å 25 Å 30 Å (Fig. 1A and Fig. 2). The strands of the b-sheet, with the exception of b6 and b7, had a parallel arrangement, forming a central core with all five strands surrounded by the four a-helices (Fig. 1B). The interior between the a-helices and the b-sheet was stabilized by numerous hydrophobic interactions (Fig. S3) and showed structural rigidity, whereas the surface helices exhibited a relatively high temperature B-factor (Fig. 1A). Taken together, RmSBP was composed of a/b domains similar to those of the existing SBPs but instead consisting of only one domain. 3.3. Structural comparison with other SBPs The DALI server was used to search for structurally homologous proteins. This search revealed 857 candidates that show a z score above 2.0 (41 for z > 10.0, 5 for z > 15). In particular, the RmSBP protein revealed a high Z-score to the ABC domain from Sreptococcus pneumonia (PDB code: 3LFT, Z score ¼ 19.0, sequence identity ¼ 21%, named as SpSBP) and an uncharacterized protein from Vibrio cholerae serotype O1 (PDB code: 3LKV, Z score ¼ 18.9, sequence identity ¼ 30%, named as VcSBP). However, since SpSBP (Uniprot entry A0A0H2UPV6, 303 amino acids excluding 41 amino acids from the signal peptide) and VcSBP (Unprot entry Q9KT04,
371
296 amino acid excluding 25 amino acids from the signal peptide) were composed of two a/b domains as in general SBPs, they were distinguished from the RmSBP having a single a/b domain (Fig. 3A). Nevertheless, in order to compare the similarity between these proteins, we performed a superimposition and confirmed that RmSBP had a highly similar C-terminal domain with that of SpSBP (r.m.s. devation of 2.3 Å for 149 Ca-atoms) and VcSBP (2.5 Å for 149 Ca-atoms) (Fig. 3B). In order to obtain insight into the putative substrate recognition of RmSBP, we further analyzed the substrate-binding site of the SpSBP and VcSBP structures complexed with L-tryptophan and Lphenylalanine, respectively, showing the closed conformation. In SpSBP, the L-tryptophan substrate interacted with seven amino acids (Gln284, His61, Ser63, Leu64, Ala138, Thr140, and Asp163) in the N-terminal domain and two amino acids (Ile242 and Asn244) in the C-terminal domain (Fig. 3C and Fig. S4A). In VcSBP, the phenylalanine substrate interacted with six amino acids (His38, Ala40, Thr97, Ala116, Thr118, and Asp141) in the N-terminal domain and two amino acids (Ile220 and Asn222) in the C-terminal domain (Fig. 3C and Fig. S4B). In the C-terminal domain of SpSBP and VcSBP, the substrate interacted with the Ile and Asn residues, respectively, via hydrophobic interactions and the carbonyl-group of the substrate, respectively. Each amino acid was in the same structural position, and the substrate was on a hydrophobic surface. In contrast, RmSBP have Glu107 and Ser109 at the same positions,
Fig. 2. Conservation and electrostatic surface analysis of RmSBP. (A) Surface conservation map of RmSBP. The surface is colored according to sequence conservation from most conserved (purple) to the most variable (cyan) based on an alignment of short sequence SBP from 5 other homologous proteins in Fig. S2B. A putative substrate-binding site is indicated by a dotted red arrow. (B) Electrostatic analysis of RmSBP revealed a highly charged region on the surface. The putative substrate-binding site is indicated by a dotted yellow arrow. All figures are displayed in the same orientation in (AeB). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
372
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
Fig. 3. Comparison of the overall structure of RmSBP with SpSBP (PDB code: 3LFT) and VsSBP (3LKV). (A) RmSBP has a single a/b domain, whereas other SPBs have two a/b domains. (B) Superimposition of the RmSBP (green) with the C-terminal domain of SpSBP (blue) and VcSBP (orange). (C) Detailed view of the putative substrate-binding surface of RmSBP. In SpSBP and VcSBP, the Ile and Asn residues interact with the substrate on a hydrophobic surface, whereas RmSBP has Glu and Ser residues in the same position with a negatively charged surface. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
which have a negatively charged surface (Fig. 3C). As a result, the RmSBP and closest structures provided by the DALI server have different characteristics, not only in the number of domains, but also in the potential substrate recognition surfaces. 4. Discussion We confirmed RmSBP to be a short ABC transport SBP and that it was located near the MCP gene. This protein is composed of one a/ b-domain, unlike other SBP proteins, which are conventionally composed of two domains. Therefore, it was expected that RmSBP would have a different mechanism of substrate binding corresponding to the conformational change caused by the hinge motion of the two a/b domains. We believed that there would be two possible substrate recognition mechanisms. First, RmSBP alone
would be capable of substrate recognition. Secondly, RmSBP and its unknown partner protein recognize substrates such as general SBP consisting of two domains. In this case, RmSBP and a partner protein will not use the hinge motion in recognizing the substrate. In order to understand the exact molecular function, further studies with RmSBP are required for identifying its substrates and the mechanism by which the substrates are recognized. Moreover, it will also be necessary to investigate whether RmSBP is involved in the general ABC transport system or related to neighbor MCP genes. Nevertheless, our structural studies of RmSBP will provide insights into understanding the new type of SBP family. Acknowledgments We thank the beamline staff at the MX beamlines at PLS-II at
J.-E. Bae et al. / Biochemical and Biophysical Research Communications 497 (2018) 368e373
Pohang Acceleratory Laboratory for assistance in data collection. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MOE) (NRF2017R1D1A1B03033087 and 2017M3A9F6029736). Transparency document Transparency document related to this article can be found online at https://doi.org/10.1016/j.bbrc.2018.02.086. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.bbrc.2018.02.086. References [1] T. van der Heide, B. Poolman, ABC transporters: one, two or four extracytoplasmic substrate-binding sites? EMBO Rep. 3 (2002) 938e943. [2] C.F. Higgins, ABC transporters: from microorganisms to man, Annu. Rev. Cell Biol. 8 (1992) 67e113. [3] K.P. Locher, Mechanistic diversity in ATP-binding cassette (ABC) transporters, Nat. Struct. Mol. Biol. 23 (2016) 487e493. [4] F. Fulyani, G.K. Schuurman-Wolters, A.V. Zagar, A. Guskov, D.J. Slotboom, B. Poolman, Functional diversity of tandem substrate-binding domains in ABC transporters from pathogenic bacteria, Structure 21 (2013) 1879e1888. [5] G.H. Scheepers, A.N.J.A. Lycklama, B. Poolman, An updated structural classification of substrate-binding proteins, FEBS Lett. 590 (2016) 4393e4401. [6] A. Maqbool, R.S. Horler, A. Muller, A.J. Wilkinson, K.S. Wilson, G.H. Thomas, The substrate-binding protein in bacterial ABC transporters: dissecting roles in the evolution of substrate specificity, Biochem. Soc. Trans. 43 (2015) 1011e1017. [7] R.P. Berntsson, S.H. Smits, L. Schmitt, D.J. Slotboom, B. Poolman, A structural classification of substrate-binding proteins, FEBS Lett. 584 (2010) 2606e2617. [8] E. Biemans-Oldehinkel, M.K. Doeven, B. Poolman, ABC transporter architecture and regulatory roles of accessory domains, FEBS Lett. 580 (2006) 1023e1035. [9] G.A. Bermejo, M.P. Strub, C. Ho, N. Tjandra, Ligand-free open-closed transitions of periplasmic binding proteins: the case of glutamine-binding protein, Biochemistry 49 (2010) 1893e1902. [10] C. Tang, C.D. Schwieters, G.M. Clore, Open-to-closed transition in apo maltosebinding protein observed by paramagnetic NMR, Nature 449 (2007) 1078e1082. [11] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Meth. Enzymol. 276 (1997) 307e326.
373
[12] P.D. Adams, P.V. Afonine, G. Bunkoczi, V.B. Chen, I.W. Davis, N. Echols, J.J. Headd, L.W. Hung, G.J. Kapral, R.W. Grosse-Kunstleve, A.J. McCoy, N.W. Moriarty, R. Oeffner, R.J. Read, D.C. Richardson, J.S. Richardson, T.C. Terwilliger, P.H. Zwart, PHENIX: a comprehensive Python-based system for macromolecular structure solution, Acta Crystallogr D Biol Crystallogr 66 (2010) 213e221. [13] T.C. Terwilliger, R.W. Grosse-Kunstleve, P.V. Afonine, N.W. Moriarty, P.H. Zwart, L.W. Hung, R.J. Read, P.D. Adams, Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard, Acta Crystallogr. Sect. D Biol. Crystallogr. 64 (2008) 61e69. [14] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr D Biol Crystallogr 60 (2004) 2126e2132. [15] V.B. Chen, W.B. Arendall 3rd, J.J. Headd, D.A. Keedy, R.M. Immormino, G.J. Kapral, L.W. Murray, J.S. Richardson, D.C. Richardson, MolProbity: all-atom structure validation for macromolecular crystallography, Acta Crystallogr D Biol Crystallogr 66 (2010) 12e21. [16] R.D. Finn, P. Coggill, R.Y. Eberhardt, S.R. Eddy, J. Mistry, A.L. Mitchell, S.C. Potter, M. Punta, M. Qureshi, A. Sangrador-Vegas, G.A. Salazar, J. Tate, A. Bateman, The Pfam protein families database: towards a more sustainable future, Nucleic Acids Res. 44 (2016) D279eD285. [17] A. Mitchell, H.Y. Chang, L. Daugherty, M. Fraser, S. Hunter, R. Lopez, C. McAnulla, C. McMenamin, G. Nuka, S. Pesseat, A. Sangrador-Vegas, M. Scheremetjew, C. Rato, S.Y. Yong, A. Bateman, M. Punta, T.K. Attwood, C.J. Sigrist, N. Redaschi, C. Rivoire, I. Xenarios, D. Kahn, D. Guyot, P. Bork, I. Letunic, J. Gough, M. Oates, D. Haft, H. Huang, D.A. Natale, C.H. Wu, C. Orengo, I. Sillitoe, H. Mi, P.D. Thomas, R.D. Finn, The InterPro protein families database: the classification resource after 15 years, Nucleic Acids Res. 43 (2015) D213eD221. [18] H. Nielsen, Predicting secretory proteins with SignalP, Meth. Mol. Biol. 1611 (2017) 59e73. [19] D. Szklarczyk, A. Franceschini, S. Wyder, K. Forslund, D. Heller, J. HuertaCepas, M. Simonovic, A. Roth, A. Santos, K.P. Tsafou, M. Kuhn, P. Bork, L.J. Jensen, C. von Mering, STRING v10: protein-protein interaction networks, integrated over the tree of life, Nucleic Acids Res. 43 (2015) D447eD452. [20] R. Lavigne, D. Seto, P. Mahadevan, H.W. Ackermann, A.M. Kropinski, Unifying classical and molecular taxonomic classification: analysis of the Podoviridae using BLASTP-based tools, Res. Microbiol. 159 (2008) 406e414. [21] F. Sievers, D.G. Higgins, Clustal omega, Curr. Protoc. Bioinformatics 48 (2014) 11e16, 3 13. [22] P. Gouet, E. Courcelle, D.I. Stuart, F. Metoz, ESPript: analysis of multiple sequence alignments in PostScript, Bioinformatics 15 (1999) 305e308. [23] H. Ashkenazy, E. Erez, E. Martz, T. Pupko, N. Ben-Tal, ConSurf 2010: calculating evolutionary conservation in sequence and structure of proteins and nucleic acids, Nucleic Acids Res. 38 (2010) W529eW533. [24] L. Holm, P. Rosenstrom, Dali server: conservation mapping in 3D, Nucleic Acids Res. 38 (2010) W545eW549. [25] P. Derr, E. Boder, M. Goulian, Changing the specificity of a bacterial chemoreceptor, J. Mol. Biol. 355 (2006) 923e932.