- Email: [email protected]
The Crystal Structure of the Extracellular Domain of the Inhibitor Receptor Expressed on Myeloid Cells IREM-1
J. Mol. Biol. (2007) 367, 310–318
doi:10.1016/j.jmb.2007.01.011
COMMUNICATION
The Crystal Structure of the Extracellular Domain of the Inhibitor Receptor Expressed on Myeloid Cells IREM-1 José Antonio Márquez 1 , Elena Galfré 2 , Florine Dupeux 1 , David Flot 1 Oscar Moran 3 and Nazzareno Dimasi 2 ⁎ 1
European Molecular Biology Laboratory, Grenoble Outstation Polygon Scientifique, 6 Rue Jules Horowitz, 38000 Grenoble, France 2
Laboratorio di Medicina Molecolare, Istituto Giannina Gaslini, Largo Gerolamo Gaslini 5, 16147 Genova, Italy 3
Istituto di Biofisica, Consiglio Nazionale delle Ricerche, 16149 Genova, Italy
The immune receptors expressed on myeloid cells (IREM) are type I transmembrane proteins encoded on human chromosome 17 (17q25.1), whose function is believed to be important in controlling inflammation. To date, three IREM receptors have been identified. IREM-1 functions as an inhibitory receptor, whereas IREM-2 and IREM-3 serve an activating function. Here, we report the crystal structure of IREM-1 extracellular domain at 2.6 Å resolution. The overall fold of IREM-1 resembles that of a V-type immunoglobulin domain, and reveals overall close homology with immunoglobulin domains from other immunoreceptors such as CLM-1, TREM-1, TLT-1 and NKp44. Comparing the surface electrostatic potential and hydrophobicity of IREM-1 with its murine homologous CLM-1, we observed unique structural properties for the complementary determining region of IREM-1, which suggests that they may be involved in recognition of the IREM-1 ligand. Particularly interesting is the structural conformation and physical properties of the antibody's equivalent CDR3 loop, which we show to be a structurally variable region of the molecule and therefore could be the main structural determinant for ligand discrimination and binding. In addition, the analysis of the IREM-1 structure revealed the presence of four structurally different cavities. Three of these cavities form a continuous hydrophobic groove on the IREM-1 surface, which point to a region of the molecule capable of accommodating potential ligands. © 2007 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: myeloid cells; inhibitory receptors; activating receptors; immune receptors; X-ray crystallography
A precise balance between positive and negative signals controls the function of the immune system. These signals are mediated by activating or inhibitory receptors present on the surface of leukocytes, and are essential for their function in order to discriminate between normal cells or invading pathogens such as bacteria, viruses and parasites. In structural terms, inhibitory or activating receptors belong to two different structural folds: the immunoglobulin-like and the C-type lectin-like.1 Abbreviations used: IREM, immune receptors expressed on myeloid cells; ITIM, immunoreceptor tyrosine-based inhibition motif; CDReq, antibody's equivalent complementary determining region; rmsd, root-mean-square deviation; Ig, immunoglobulin-like domain; Ig-V, immunoglobulin-like variable domain. E-mail address of the corresponding author: [email protected]
Both families contain inhibitory and activating receptors. A marked difference between inhibitory and activating receptors is the fact that inhibitory receptors are able to deliver their intracellular signals directly.2 On the contrary, activating receptors need to associate with specialized signal transduction polypeptides because they usually lack any functional intracellular signaling domain. Inhibitory receptors possess a cytoplasmic immune receptor tyrosine-based inhibitory motif (ITIM) that, upon ligand engagement, becomes phosphorylated and thereafter recruits Src homology 2 (SH2) bearing phosphatase. This, in turn, will start a cascade of intracellular events promoting inhibitory signals.2 Ligand engagement of activating receptors results in the phosphorylation of tyrosine residues within the immune receptor tyrosine-based activating motif (ITAM) consensus of the associated adaptor proteins.3 This phosphorylation is needed for the successive recruitment of either protein tyrosine
0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
Structure of the IREM-1
kinases containing SH2 domains, or PI3-kinase. This will stimulate a series of intracellular events that will induce inflammatory responses. During the past several years, a number of families of activating and inhibitory receptors of the immunoglobulin fold have been identified in myeloid cells, which are a specialized class of hematopoietic cells whose function is important in the regulation of the innate immune responses.4–6 Among these, the immunoglobulin-like transcript 1,7 SIRP,8 CMRF-35/ IRp60,9–12 immune receptors expressed on myeloid cells (IREMs),13–15 and TREM family of receptors 16,17 have a major role in the regulation of the myeloid cell function. The newly identified IREM immunoreceptors are of special interest, because they belong to a family of molecules with a pivotal role in the functional regulation of myeloid cells, and because they have important immunoregulatory functions during inflammation.13–15 IREM receptors belong to a multigene family encoding both inhibitory and activating receptors that map on the human chromosome 17 (17q25.1), which encodes other receptors, such as CMRF-35 (CD300c) and CMRF-35H/IRp60 (CD300a). Three distinct IREM receptors have been identified: IREM1 or CD300f,13 IREM-2 or CD300e14 and IREM-3 or CD300b.15 IREM-1 functions as an inhibitory receptor, whereas IREM-2 and IREM-3 serve as activating receptors on myeloid cells. Topologically, they are type I transmembrane proteins with a single extracellular immunoglobulinlike domain. Human IREMs consist of an extracellular domain of about 130-150 amino acids residues, a membrane-spanning domain of about 20 residues, and a cytoplasmic tail of ten residues for IREM-2, 30 residues IREM-3, and about 100 residues for IREM-1. Their extracellular immunoglobulin domains share about 30% sequence homology, but have relatively low levels of sequence homology to classic immunoglobulin members. With the goals of obtaining structural information on the IREM receptors, which is important in order to understand the molecular and structural principles that could determine recognition of the still elusive IREM ligands, we present the crystal structure at 2.6 Å resolution of the extracellular immunoglobulin-like domain of the inhibitory receptor IREM-1. The extracellular immunoglobulin-like domain of IREM-1, corresponding to amino acid residues 21– 140, was assembled in vitro from a mixture of 21 chemically synthesized oligonucleotides with codon optimized for maximum protein expression in prokaryotes.18 Protein expression was carried out in Escherichia coli and the protein was expressed as inclusion bodies. The protein was refolded in vitro by dilution, and purified using metal-affinity, gelfiltration and ion-exchange chromatography. The protein was concentrated and used for crystallization experiments, which were carried out using high-throughput nanodrop protocols.18 Needle-like crystals were obtained at room temperature after a period of five days with approxi-
311 mately dimension of 300 μm × 10 μm × 10 μm.18 Because of the small size of the crystals, data collection was carried out at the microfocus beamline ID23-2 of the European Synchrotron Radiation Facility (ESRF) in Grenoble, which is optimized for data collection from very small crystals. IREM-1 crystallized in space group P3121 with cell dimension a = b = 54.23 Å, c = 72.01 Å and α = β = 90°, γ = 120°. 18 The crystal contains one molecule of IREM-1 per asymmetric unit. The structure was determined by molecular replacement using the program Phaser,19 and the atomic coordinates of the mouse CMRF-35-like myeloid receptor extracellular domain 1 (CLM-1)9 (PDB code 1ZOX) as a starting model. IREM-1 and CLM-1 share 58% sequence identity over 98 overlapping residues forming the core of the immunoglobulin-like domain. The initial 2Fo–Fc model map and the 2Fo–Fc composite omit map, were continuous and of excellent quality, facilitating the unambiguous modelling of residues 23–126. After several iterations of model building and refinement, the final model has an Rcryst and an Rfree of 21.8% and 25.5%, respectively (Table 1). The N-terminal residues glycine 18, isoleucine 19 and proline 20 that were part of the thrombin cleavage tag, were visible in the electron density map and, therefore, were included in the final model. The model stereochemistry is typical for this resolution (2.6 Å) (Table 1), with 86.3% of the residues in the most favoured regions, 11.6% in the additionally allowed regions, and 2.1% in the generously allowed regions of the Ramachandran plot. The rmsd of the model from the ideal is 0.008 Å for bond lengths and 1.7° for bond angles (Table 1). The overall structure of IREM-1 is similar to a typical Ig-V domain, and consists of ten β-strands forming two antiparallel Table 1. Refinement statistics A. Refinement Resolution range (Å) Number of reflections in working set Number of reflections in test set Rcrysta (%) Rfreeb (%) B. Statistics of the final model Number of residues Number of water molecules rmsd from ideality Bond lengths (Å) Bond angles (deg.) Average B-values (Å2) Protein atoms Water molecules Ramachandran plot statisticsc Most favored (%) Allowed (%) Generously allowed (%) Disallowed (%)
19.67–2.60 3814 201 21.8 25.5 112 50 0.008 1.7 25.7 32.3 86.3 11.6 2.1 0.0
a Rcryst = ∑| |Fo| – |Fc||/∑|Fo|, where Fc is the calculated structure factor. b Rfree is as for Rcryst but calculated for a randomly selected 5.0% of reflections not included in the refinement. c As calculated by PROCHECK30 and by MOLPROBITY.31
312 β-sheets (Figure 1(a)). The first β-sheet is formed by β-strands G-F-C-C′, and the second by β-strands BE-D-CVV. A 310 α-helix is formed between β-strands E and F (Figure 1(a)). There are two disulphide bonds in IREM-1 (Figure 1(a)). The disulphide bond Cys40–Cys108 is conserved in the immunoglobulin-like domains and is found in other IREM-1 homologues such as murine and human TREM-1,20–22 human TLT-1,23 murine CLM-19 (PDB entry code: 1ZOX), NKp4424 and pIgR.25 This disulphide bond plays a major structural role in the Ig-V type immunoglobulins because it locks the two β-sheets together. Cys54 and Cys62 form the second disulfide bond in IREM-1. This disulfide bond is important for stabilizing the βhairpin formed by β-strands C-C′. This disulphide bond is found in CLM-1, NKp44, TLT-1 and pIgR, but is not present in murine and human TREM-1. Two hydrophobic cores are present in IREM-1. The first hydrophobic core is located within the two β-sheets towards the carboxy-terminal end of the protein chain and is composed of residues valine 28, 38, 121, 123, leucine 36, tryptophan 53, methionine 95, and tyrosine 106. The second hydrophobic core is located in the protrusion formed by the bottom part of β-strands F-C-C′ and the upper part of β-strands F-G. This core is formed by residues tryptophan 52, 59, and 107 (Figure 1(b)). A feature of the IREM-1 crystal structure is the formation of a prominent protrusion from the main immunoglobulin β-sandwich body (Figure 1(b)). This protrusion is about 18 Å long, 15 Å wide and 11 Å deep. It is noteworthy that this protrusion is present also in other Ig-V type receptors, such as NKp44, mouse and human TREM1, mouse CLM-1 and the pIgR. Therefore, this region could have some function in “accommodating” the potential ligand(s) recognized by this class of receptors (see later). The top of this protrusion is blocked, in part, by the side-chain of tryptophan 59 (Figure 1(b)). The other amino acid residues that are filling up this protrusion with their side-chains are isoleucine 64, aspartate 116, glutamate 11, lysine 67, tryptophan 52 and tryptophan 107 (Figure 1(b)). Calculation of the surface electrostatic potential for this protrusion shows that a central negatively charged surface is surrounded by an extensive positively charged surface. The hydrophobicity analysis shows that the central part of this protrusion is moderately hydrophobic and is surrounded by polar residues. By inspecting the IREM packing in the crystal, we observed that this protrusion is solvent-exposed and it is not occupied by symmetry-related molecules. This could point to a rather accessible surface for potential binding of hydrophilic and positively charged ligands. The three-dimensional structure of IREM-1 was used as a probe for searching the RCSB Protein Data Bank using the Vector Alignment Search Tool from NCBI.26 This search identified 60 non-redundant protein structures similar to IREM-1. Amongst these were the crystal structures of murine and human TREM-1,20–22 human TLT-1,23 murine CLM-1 (PDB
Structure of the IREM-1
entry code: 1ZOX), NKp4424 and pIgR.25 In fact, structure-based sequence alignment of the immunoglobulin-like domains between these representative structural homologues, showed good overall sequence identities (Figure 2). Not surprisingly, because it was the only model that produced a molecular replacement solution, the closest structural homologue of IREM-1 is CLM-1. The rmsd between the structurally equivalent Cα atoms forming the core of the immunoglobulin-like domain in IREM-1 and CLM-1 is 1.06 Å (Figure 1(c)), whereas the calculated rmsd values for IREM-1 superimposed on TLT-1 is 2.61 Å, the rmsd between IREM-1 and human TREM-1 is 4.47 Å, and the rmsd between IREM-1 and NKp44 is 2.58 Å. CDReq3 is the most variable loop in IREM-1 when compared with its murine homologous CLM-1 (Figure 1(c)). As seen in Figure 1(c), when IREM-1 is superimposed on CLM-1, the major structural changes are observed in the CDReq3 conformation and in the conformation of the β-hairpin formed by β-strands C-C′. In fact, the CDReq3 in IREM-1 is shifted about 4 Å and is rotated about 35° anticlockwise with respect to the CDReq3 of CLM-1 (Figure 3(c), circle), with almost no structural change in the immunoglobulin body. This variation could suggest an area of variability that may move in response to ligation. The electrostatic potential surface and hydrophobicity that describe the surface of the CDReq regions are different in IREM-1 and CLM-1 (Figure 3). The CDReq1 net surface charge in IREM-1 is −1.4 e− (Figure 3(a)), and represents a surface potential of 364 kJ mol−1. The CDReq1 in CLM-1 is more negative, with a surface charge of −2.8 e− (Figure 3(d)), corresponding to a surface potential of 656 kJ mol−1. The CDReq1 hydrophobicity in IREM-1 and CLM-1 is relatively similar (Figure 3(c) and (f)). Conversely, the CDReq2 net surface charge in IREM-1 is −2.5 e–, which is more negative (Figure 3(a)) than CLM-1, which is −0.6 e− (Figure 3(d)). This corresponds to an electrostatic surface potential of 643 kJ mol−1 and 198 kJ mol−1 for IREM-1 and CLM-1, respectively. The CDReq2 hydrophobicity in IREM-1 and CLM-1 is similar (Figure 3(c) and (f)). Finally, the net surface charge for the CDReq3 is nearly neutral in both polypeptides, being −0.6 e− and −0.3 e− for IREM-1 (Figure 3(a)) and CLM-1 (Figure 3(d)), respectively, that represent an electrostatic surface potential of 121 kJ mol−1 for IREM-1 and 87 kJ mol−1 for CLM1. The CDReq3 hydrophobicity is higher in IREM-1 (Figure 3(c)) than that of CLM-1 (Figure 3(f)). These physical differences could be the structural determinants for ligand discrimination and binding of these two homologous molecules if they bind their ligands in a fashion similar to that of antibodies and T-cell receptors. In an attempt to identify the region of IREM-1 that could accommodate ligands, a detailed analysis of surface cavities using a probe radius of 1.4 Å was carried out. This analysis revealed that four distinct solvent-exposed cavities, termed 1, 2, 3 and 4, are present in IREM-1 (Figure 4). None of these cavities,
Structure of the IREM-1
313
Figure 1. Overall structure of IREM-1, features of the extended protrusion and comparison with CLM-1. (a) A ribbon representation of the IREM-1 structure. IREM-1 has the expected V-set Ig domain fold present in the immunoglobulins. The secondary structure elements (β-strands and the 3 10 α-helix) were defined using the program DSSP.41 The antibody's equivalent CDR loops (CDReq) are shown in red and are labelled. Disulphide bonds are shown and are labelled with their respective cysteine residues. In this ribbon representation and in all the subsequent diagrams, N and C denote the amino terminus and the carboxy terminus of the protein chain. The location of the potential Nglycosylation site, which involves asparagine 88, is shown with a red diamond. (b) An important feature of the IREM-1 structure is the formation of a prominent protrusion that extends from the main immunoglobulin body. This protrusion is about 18 Å long, 15 Å wide and 11 Å deep. As seen in this Figure, the top of this protrusion is blocked, in part, by the side-chains of tryptophan 59, aspartate 116 and glutamate 111. The other amino acid residues present at this protrusion are isoleucine 64, lysine 67, tryptophan 52 and tryptophan 107. (c) Superimposition of IREM-1 with its murine homologous CLM-1 gives an rmsd of 1.06 for all the structurally equivalent Cα atoms forming the core of the immunoglobulin-like domain (102 residues). The orientation of IREM-1 and CLM-1 here is equivalent to that in (a). In this superimposition diagram, IREM-1 is in blue and CLM-1 is in red. The β-hairpin formed by β-strands C-C′ is labelled. The location of the CDReq loops is shown schematically. The circle highlights the prominent structural difference of the CDReq3 between IREM-1 and CLM-1.
Structure of the IREM-1
314
Figure 2. Structure-based sequence alignment of the extracellular immunoglobulin-like domains of IREM homologues. In this sequence, alignment of the secondary structural information was derived from the IREM-1 structure and with the program DSSP.40 The sequences shown are for the only immunoglobulin-like extracellular domain of human IREM-1 (NCBI AAP57942), human IREM-2 (NCBI NP_852114), human IREM-3 (NCBI AAV69612), human IRp60 (NCBI CAB66145), human NKp44 (NCBI AJ225109), human TREM-1 (NCBI NM_018643), human TLT-1 (NCBI BC100945), human CMRF-35 (NCBI X66171), mouse CLM-1 (NCBI AY457047), and a representative immunoglobulin V-domain of human TCR (NCBI NG_001332.2). The sequences are ordered from top to bottom on the basis of their sequence similarity as defined by the program CLUSTAL W.39 The numbering above the sequence corresponds to the unprocessed IREM-1 protein. Complete sequence conservation is labelled white with a red background, and red letters correspond to >60% sequence homology. The secondary structure elements are denoted by arrows (β-strands) and by squiggles (310 α-helix). The paired numbers at the bottom of the sequences indicate the bonded cysteine residue involved in the formation of the disulphide bridges. The sequences of the CDR-equivalent loops (CDReq) are labelled and boxed in black. The consensus sequence Asp-X-Gly/Asp-X-Tyr-X-Cys, which correspond to the signature of a V-type Ig domain is labelled with a star symbol above the sequence. The arrow denotes the potential N-glycosylation site in IREM-1 that involves asparagine 88.
as expected, involves the CDReq loops. Cavity 1 is located at the bottom of the protrusion that extends from the main immunoglobulin-like body (Figure 4). Cavities 2 and 3 are close to each other and are located towards the carboxy-terminal part of the molecule (Figure 4). Cavity 4 is located on the opposite site of the cavity 1 in a region that is beneath the CDReq2 loop (Figure 4). Cavity 1 is hydrophobic (Figure 4(a)) and possesses a negative electrostatic potential (Figure 4(c)). Cavity 2 is hydrophobic (Figure 4(a)) and has a positive electrostatic potential on one side and a negative electrostatic potential on the opposite side (Figure 4(c)). Cavity 3 is mainly hydrophilic (Figure 4(a)) and has an electrostatic surface potential like cavity 2 (Figure 4(c)). Cavity 4 is slightly hydrophobic (Figure 4(b)) with a neutral electrostatic surface potential (Figure 4(d)). Notably, cavities 1, 2 and 3 form a continuous groove on the IREM-1 surface (Figure 4(a) and (c)), which may point to an extended surface capable of accommodating hydrophilic and positive charged ligands. In summary, our structural results show that the immunoglobulin-like fold is conserved in IREM-1, and therefore is presumably conserved in the other immunoreceptors encoded on human chromosome 17 locus 17q25.1, but unique hydrophobicity and
electrostatic surface potentials could be used by these classes of immunoreceptors to achieve exquisite ligand binding specificity using the antibodyequivalent CDR loops, in a fashion analogous to that of T-cell receptors, or through unique structural cavities. Methods Cloning, protein expression, purification, crystallization and X-ray data collection is reported elsewhere.18 The structure was solved using the molecular replacement program Phaser,19 and the coordinates of the mouse CLM1 receptor extracellular domain (PDB code: 1ZOX) as the starting model. IREM-1 and CLM-1 immunoglobulin-like extracellular domains share 58.3% sequence identity. The molecular replacement search was carried out using data from 8–3.5 Å, and using space group P3121. Rigidbody refinement of the rotated and translated model using a resolution from 20 Å to3.0 Å gave an R-factor of 0.41 and an Rfree of 0.45. Several rounds of simulated annealing, conjugate gradient minimization and B-factor refinement followed by several iterations through individual steps of manual model building, maximum likelihood phase refinement and solvent molecule addition with XtalView,27 CNS28 and REFMAC29 have resulted in the current structure, which has a R-factor/Rfree of 0.21/0.25 for all data from 19.67–2.60 Å resolution. The final model
Figure 3. Electrostatic and hydrophobic features of the CDReq loops in IREM-1 and comparison with the murine homologous CLM-1. The view of these two molecules is from the top. The location of the three CDReq loops is shown schematically for both molecules. In (b) and (e), the CDReq1 surface is coloured pink, the CDReq2 surface is coloured orange and the CDReq3 surface is coloured green. This Figure highlights that fact that IREM-1 has unique antibody's equivalent complementary determinant region surface shape, electrostatic potentials and hydrophobicity (see the text for details). The electrostatic potential is coloured from −5 kJ mol−1 Å−2 (red, negative potential) to 5 kJ mol−1 Å−2 (blue, positive potential). Neutral potential (0 kJ mol−1 Å−2) is shown in white. The hydrophobicity is coloured over a scale of green (most hydrophobic amino acids) to white (least hydrophobic amino acids).
Structure of the IREM-1
315
316
Structure of the IREM-1
Figure 4. Cavities on IREM-1. The cavities on IREM-1 are shown as yellow dots and are labelled from 1 to 4. The locations of the CDReq loops are shown schematically and are labelled only in the ribbon diagram. In (b) and (d) the continuous groove formed by cavities 1, 2 and 3 is highlighted. This groove, which possesses a negative potential and is highly hydrophobic, may accommodate potential ligands. The electrostatic potential surfaces and hydrophobicity are coloured as in Figure 3. The ribbon orientation is equivalent to that of Figure 1(a). contains 112 amino acid residues and 50 water molecules. The first three amino acid residues are derived from the vector N-terminal tag, while the remaining 109 correspond to residues 21–129 of the IREM-1 sequence.
Model geometry was examined using PROCHECK30 and MOLPROBITY.31 Structure superimpositions were performed with Superpose,32 and visually inspected using the program XtalView.27 Structural similarities were
Structure of the IREM-1 analyzed using the program VAST.26 Solvent-accessible surface areas were calculated with CNS,28 using a probe radius of 1.4 Å. Electrostatic potentials were calculated in an aqueous solution at pH 7.0 and with 150 mM NaCl using the program APBS.33 The hydrophobicity surfaces were calculated by using the optimized matching hydrophobicity program PROTSCALE.34 Surface cavities were calculated with the programs CASTp,35 CAVER,36 and LIGSITEcsc,37 using a probe radius of 5 Å. The Figures were prepared with the program Pymol (DeLano Scientific, San Carlos, CA)†, and rendered with the GIMP image manipulation program‡. The N-glycosylation sites were analyzed with the program GlyPro.38 Sequence alignment was carried out using the program CLUSTAL W,39 and was rendered using the program Espript.40
317
7.
8.
9.
Protein Data Bank accession code Atomic coordinates and stricture factors have been deposited in the RCSB Protein Data Bank§ under the accession code 2NMS.
10.
11.
Acknowledgements We thank Professor Roy Mariuzza (Center for Advanced Research in Biotechnology, Rockville, Maryland, USA) for critical reading of the manuscript. This work was supported, in part, by the Fondazione Gerolamo Galini (Genova, Italy). We thank Alessandro Faravelli (Laboratorio di Medicina Molecolare, Istituto Giannina Gaslini, Genova, Italy) for help during protein expression. We thank the constant scientific support from Professor Lorenzo Moretta (Istituto Giannina Gaslini, Genova, Italy).
12.
References
15.
1. Natarajan, K., Dimasi, N., Wang, J., Mariuzza, R. A. & Margulies, D. H. (2002). Structure and function of natural killer cell receptors: multiple molecular solutions to self, nonself discrimination. Annu. Rev. Immunol. 20, 853–885. 2. Ravetch, J. V. & Lanier, L. L. (2000). Immune inhibitory receptors. Science, 290, 84–89. 3. Humphrey, M. B., Lanier, L. L. & Nakamura, M. C. (2005). Role of ITAM-containing adapter proteins and their receptors in the immune system and bone. Immunol. Rev. 208, 50–65. 4. van den Berg, T. K., Yoder, J. A. & Litman, G. W. (2004). On the origins of adaptive immunity: innate immune receptors join the tale. Trends Immunol. 1, 11–16. 5. Pandey, S. & Agrawal, D. K. (2006). Immunobiology of Toll-like receptors: emerging trends. Immunol. Cell Biol. 84, 333–341. 6. Dietrich, J., Nakajima, H. & Colonna, M. (2000).
† http://pymol.sourceforge.net/ ‡ http://www.gimp.org/ § http://www.rcsb.org/
13.
14.
16.
17.
18.
19. 20.
Human inhibitory and activating Ig-like receptors which modulate the function of myeloid cells. Microbes Infect. 3, 323–329. Colonna, M., Navarro, F., Bellon, T., Llano, M., Garcia, P., Samaridis, J. et al. (1997). A common inhibitory receptor for major histocompatibility complex class I molecules on human lymphoid and myelomonocytic cells. J. Expt. Med. 186, 1809–1818. Fujioka, Y., Matozaki, T., Noguchi, T., Iwamatsu, A., Yamao, T., Takahashi, N. et al. (1996). A novel membrane glycoprotein, SHPS-1, that binds the SH2domain-containing protein tyrosine phosphatase SHP-2 in response to mitogens and cell adhesion. Mol. Cell Biol. 12, 6887–6899. Chung, D. H., Humphrey, M. B., Nakamura, M. C., Ginzinger, D. G., Seaman, W. E. & Daws, M. R. (2003). CMRF-35-like molecule-1, a novel mouse myeloid receptor, can inhibit osteoclast formation. J. Immunol. 171, 6541–6658. Clark, G. J., Green, B. J. & Hart, D. N. (2000). The CMRF-35H gene structure predicts for an independently expressed member of an ITIM/ITAM pair of molecules localized to human chromosome 17. Tissue Antigens, 55, 101–109. Green, B. J., Clark, G. J. & Hart, D. N. (1998). The CMRF-35 mAb recognizes a second leukocyte membrane molecule with a domain similar to the poly Ig receptor. Int. Immunol. 10, 891–899. Cantoni, C., Bottino, C., Augugliaro, R., Morelli, L., Marcenaro, E., Castriconi, R. et al. (1999). Molecular and functional characterization of IRp60, a member of the immunoglobulin superfamily that functions as an inhibitory receptor in human NK cells. Eur. J. Immunol. 10, 3148–3159. Alvarez-Errico, D., Aguilar, H., Kitzig, F., Brckalo, T., Sayos, J. & Lopez-Botet, M. (2004). IREM-1 is a novel inhibitory receptor expressed by myeloid cells. Eur. J. Immunol. 34, 3690–3701. Aguilar, H., Alvarez-Errico, D., Garcia-Montero, A. C., Orfao, A., Sayos, J. & Lopez-Botet, M. (2004). Molecular characterization of a novel immune receptor restricted to the monocytic lineage. J. Immunol. 173, 6703–6711. Martinez-Barriocanal, A. & Sayos, J. (2006). Molecular and functional characterization of CD300b, a new activating immunoglobulin receptor able to transduce signals through two different pathways. J. Immunol. 177, 2819–2830. Allcock, R. J., Barrow, A. D., Forbes, S., Beck, S. & Trowsdale, J. (2003). The human TREM gene cluster at 6p21.1 encodes both activating and inhibitory single IgV domain receptors and includes NKp44. Eur. J. Immunol. 33, 567–577. Bouchon, A., Dietrich, J. & Colonna, M. (2000). Cutting edge: inflammatory responses can be triggered by TREM-1, a novel receptor expressed on neutrophils and monocytes. J. Immunol. 164, 4991–4995. Dimasi, N., Flot, D., Dupex, F. & Marquez, J. A. (2007). Expression, crystallization and X-ray data collection from microcrystals of the extracellular domain of the human inhibitory receptor expressed on myeloid cells IREM-1. Acta Crystallog. sect. F. (In the press). Storioni, L. C., McCoy, A. J. & Read, J. R. (2004). Likelihood-enhanced fast rotation functions. Acta Crystallog. sect. D, 60, 432–438. Kelker, M. S., Debler, E. W. & Wilson, I. A. (2004). Crystal structure of mouse triggering receptor expressed on myeloid cells 1 (TREM-1) at 1.76 Å. J. Mol. Biol. 344, 1175–1181.
Structure of the IREM-1
318 21. Kelker, M. S., Foss, T. R., Peti, W., Teyton, L., Kelly, J. W., Wuthrich, K. & Wilson, I. A. (2004). Crystal structure of human triggering receptor expressed on myeloid cells 1 (TREM-1) at 1.47 Å. J. Mol. Biol. 342, 1237–1248. 22. Radaev, S., Kattah, M., Rostro, B., Colonna, M. & Sun, P. D. (2003). Crystal structure of the human myeloid cell activating receptor TREM-1. Structure, 11, 1527–3155. 23. Gattis, J. L., Washington, A. V., Chisholm, M. M., Quigley, L., Szyk, A., McVicar, D. W. & Lubkowski, J. (2006). The structure of the extracellular domain of triggering receptor expressed on myeloid cells like transcript-1 and evidence for a naturally occurring soluble fragment. J. Biol. Chem. 281, 13396–13403. 24. Cantoni, C., Ponassi, M., Biassoni, R., Conte, R., Spallarossa, A., Moretta, A. et al. (2003). The threedimensional structure of the human NK cell receptor NKp44, a triggering partner in natural cytotoxicity. Structure, 11, 725–734. 25. Hamburger, A. E., West, A. P., Jr & Bjorkman, P. J. (2004). Crystal structure of a polymeric immunoglobulin binding fragment of the human polymeric immunoglobulin receptor. Structure, 11, 1925–1935. 26. Gibrat, J. F., Madej, T. & Bryant, S. H. (1996). Surprising similarities in structure comparison. Curr. Opin. Struct. Biol. 3, 377–385. 27. McRee, D. E. (1999). XtalView/Xfit. A versatile program for manipulating atomic coordinates and electron density. J. Struct. Biol. 125, 156–165. 28. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921. 29. Collaborative Computational Project, Number 4. (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760–763. 30. Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1993). PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallog. 26, 283–290. 31. Davis, I. W., Murray, L. W., Richardson, J. S. & Richardson, D. C. (2004). MOLPROBITY: structure
32.
33.
34.
35.
36.
37.
38. 39.
40. 41.
validation and all-atom contact analysis for nucleic acids and their complexes. Nucl. Acids Res. 32, W615–W619. Maiti, R., Van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). SuperPose: a simple server for sophisticated structural superposition. Nucl. Acids Res. 32, W590–W594. Baker, N. A., Sept, D., Joseph, S., Holst, M. J. & McCammon, J. A. (2001). Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl Acad. Sci. USA, 18, 10037–10041. Sweet, R. M. & Eisenberg, D. (1983). Correlation of sequence hydrophobicities measures similarity in three-dimensional protein structure. J. Mol. Biol. 171, 479–488. Dundas, J., Ouyang, Z., Tseng, J., Binkowski, A., Turpaz, Y. & Liang, J. (2006). CASTp: computed atlas of surface topography of proteins with structural and topographical mapping of functionally annotated residues. Nucl. Acids Res. 34, W116–W118. Petřek, M., Otyepka, M., Banáš, P., Košinová, P., Koča, J. & Damborský, J. (2006). CAVER: a new tool to explore routes from protein clefts, pockets and cavities. BMC Bioinformatics, 7, 316. Huang, B. & Schroeder, M. (2006). LIGSITEcsc: predicting ligand binding sites using the Connolly surface and degree of conservation. BMC Struct. Biol. X, 6–19. Bohne-Lang, A. & von der Lieth, C. W. (2005). GlyProt: in silico glycosylation of proteins. Nucl. Acids Res. 33, W214–W219. Higgins, D., Thompson, J., Gibson, T., Thompson, J., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl. Acids Res. 22, 4673–4680. Gouet, P., Courcelle, E., Stuart, D. I. & Metoz, F. (1999). ESPript: multiple sequence alignments in PostScript. Bioinformatics, 15, 305–308. Kabsch, W. & Sander, C. (1983). Dictionary of protein secondary structure: pattern recognition of hydrogenbonded and geometrical features. Biopolymers, 22, 2577–2637.
Edited by I. Wilson (Received 15 November 2006; received in revised form 21 December 2006; accepted 3 January 2007) Available online 10 January 2007
doi:10.1016/j.jmb.2007.01.011
COMMUNICATION
The Crystal Structure of the Extracellular Domain of the Inhibitor Receptor Expressed on Myeloid Cells IREM-1 José Antonio Márquez 1 , Elena Galfré 2 , Florine Dupeux 1 , David Flot 1 Oscar Moran 3 and Nazzareno Dimasi 2 ⁎ 1
European Molecular Biology Laboratory, Grenoble Outstation Polygon Scientifique, 6 Rue Jules Horowitz, 38000 Grenoble, France 2
Laboratorio di Medicina Molecolare, Istituto Giannina Gaslini, Largo Gerolamo Gaslini 5, 16147 Genova, Italy 3
Istituto di Biofisica, Consiglio Nazionale delle Ricerche, 16149 Genova, Italy
The immune receptors expressed on myeloid cells (IREM) are type I transmembrane proteins encoded on human chromosome 17 (17q25.1), whose function is believed to be important in controlling inflammation. To date, three IREM receptors have been identified. IREM-1 functions as an inhibitory receptor, whereas IREM-2 and IREM-3 serve an activating function. Here, we report the crystal structure of IREM-1 extracellular domain at 2.6 Å resolution. The overall fold of IREM-1 resembles that of a V-type immunoglobulin domain, and reveals overall close homology with immunoglobulin domains from other immunoreceptors such as CLM-1, TREM-1, TLT-1 and NKp44. Comparing the surface electrostatic potential and hydrophobicity of IREM-1 with its murine homologous CLM-1, we observed unique structural properties for the complementary determining region of IREM-1, which suggests that they may be involved in recognition of the IREM-1 ligand. Particularly interesting is the structural conformation and physical properties of the antibody's equivalent CDR3 loop, which we show to be a structurally variable region of the molecule and therefore could be the main structural determinant for ligand discrimination and binding. In addition, the analysis of the IREM-1 structure revealed the presence of four structurally different cavities. Three of these cavities form a continuous hydrophobic groove on the IREM-1 surface, which point to a region of the molecule capable of accommodating potential ligands. © 2007 Elsevier Ltd. All rights reserved.
*Corresponding author
Keywords: myeloid cells; inhibitory receptors; activating receptors; immune receptors; X-ray crystallography
A precise balance between positive and negative signals controls the function of the immune system. These signals are mediated by activating or inhibitory receptors present on the surface of leukocytes, and are essential for their function in order to discriminate between normal cells or invading pathogens such as bacteria, viruses and parasites. In structural terms, inhibitory or activating receptors belong to two different structural folds: the immunoglobulin-like and the C-type lectin-like.1 Abbreviations used: IREM, immune receptors expressed on myeloid cells; ITIM, immunoreceptor tyrosine-based inhibition motif; CDReq, antibody's equivalent complementary determining region; rmsd, root-mean-square deviation; Ig, immunoglobulin-like domain; Ig-V, immunoglobulin-like variable domain. E-mail address of the corresponding author: [email protected]
Both families contain inhibitory and activating receptors. A marked difference between inhibitory and activating receptors is the fact that inhibitory receptors are able to deliver their intracellular signals directly.2 On the contrary, activating receptors need to associate with specialized signal transduction polypeptides because they usually lack any functional intracellular signaling domain. Inhibitory receptors possess a cytoplasmic immune receptor tyrosine-based inhibitory motif (ITIM) that, upon ligand engagement, becomes phosphorylated and thereafter recruits Src homology 2 (SH2) bearing phosphatase. This, in turn, will start a cascade of intracellular events promoting inhibitory signals.2 Ligand engagement of activating receptors results in the phosphorylation of tyrosine residues within the immune receptor tyrosine-based activating motif (ITAM) consensus of the associated adaptor proteins.3 This phosphorylation is needed for the successive recruitment of either protein tyrosine
0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.
Structure of the IREM-1
kinases containing SH2 domains, or PI3-kinase. This will stimulate a series of intracellular events that will induce inflammatory responses. During the past several years, a number of families of activating and inhibitory receptors of the immunoglobulin fold have been identified in myeloid cells, which are a specialized class of hematopoietic cells whose function is important in the regulation of the innate immune responses.4–6 Among these, the immunoglobulin-like transcript 1,7 SIRP,8 CMRF-35/ IRp60,9–12 immune receptors expressed on myeloid cells (IREMs),13–15 and TREM family of receptors 16,17 have a major role in the regulation of the myeloid cell function. The newly identified IREM immunoreceptors are of special interest, because they belong to a family of molecules with a pivotal role in the functional regulation of myeloid cells, and because they have important immunoregulatory functions during inflammation.13–15 IREM receptors belong to a multigene family encoding both inhibitory and activating receptors that map on the human chromosome 17 (17q25.1), which encodes other receptors, such as CMRF-35 (CD300c) and CMRF-35H/IRp60 (CD300a). Three distinct IREM receptors have been identified: IREM1 or CD300f,13 IREM-2 or CD300e14 and IREM-3 or CD300b.15 IREM-1 functions as an inhibitory receptor, whereas IREM-2 and IREM-3 serve as activating receptors on myeloid cells. Topologically, they are type I transmembrane proteins with a single extracellular immunoglobulinlike domain. Human IREMs consist of an extracellular domain of about 130-150 amino acids residues, a membrane-spanning domain of about 20 residues, and a cytoplasmic tail of ten residues for IREM-2, 30 residues IREM-3, and about 100 residues for IREM-1. Their extracellular immunoglobulin domains share about 30% sequence homology, but have relatively low levels of sequence homology to classic immunoglobulin members. With the goals of obtaining structural information on the IREM receptors, which is important in order to understand the molecular and structural principles that could determine recognition of the still elusive IREM ligands, we present the crystal structure at 2.6 Å resolution of the extracellular immunoglobulin-like domain of the inhibitory receptor IREM-1. The extracellular immunoglobulin-like domain of IREM-1, corresponding to amino acid residues 21– 140, was assembled in vitro from a mixture of 21 chemically synthesized oligonucleotides with codon optimized for maximum protein expression in prokaryotes.18 Protein expression was carried out in Escherichia coli and the protein was expressed as inclusion bodies. The protein was refolded in vitro by dilution, and purified using metal-affinity, gelfiltration and ion-exchange chromatography. The protein was concentrated and used for crystallization experiments, which were carried out using high-throughput nanodrop protocols.18 Needle-like crystals were obtained at room temperature after a period of five days with approxi-
311 mately dimension of 300 μm × 10 μm × 10 μm.18 Because of the small size of the crystals, data collection was carried out at the microfocus beamline ID23-2 of the European Synchrotron Radiation Facility (ESRF) in Grenoble, which is optimized for data collection from very small crystals. IREM-1 crystallized in space group P3121 with cell dimension a = b = 54.23 Å, c = 72.01 Å and α = β = 90°, γ = 120°. 18 The crystal contains one molecule of IREM-1 per asymmetric unit. The structure was determined by molecular replacement using the program Phaser,19 and the atomic coordinates of the mouse CMRF-35-like myeloid receptor extracellular domain 1 (CLM-1)9 (PDB code 1ZOX) as a starting model. IREM-1 and CLM-1 share 58% sequence identity over 98 overlapping residues forming the core of the immunoglobulin-like domain. The initial 2Fo–Fc model map and the 2Fo–Fc composite omit map, were continuous and of excellent quality, facilitating the unambiguous modelling of residues 23–126. After several iterations of model building and refinement, the final model has an Rcryst and an Rfree of 21.8% and 25.5%, respectively (Table 1). The N-terminal residues glycine 18, isoleucine 19 and proline 20 that were part of the thrombin cleavage tag, were visible in the electron density map and, therefore, were included in the final model. The model stereochemistry is typical for this resolution (2.6 Å) (Table 1), with 86.3% of the residues in the most favoured regions, 11.6% in the additionally allowed regions, and 2.1% in the generously allowed regions of the Ramachandran plot. The rmsd of the model from the ideal is 0.008 Å for bond lengths and 1.7° for bond angles (Table 1). The overall structure of IREM-1 is similar to a typical Ig-V domain, and consists of ten β-strands forming two antiparallel Table 1. Refinement statistics A. Refinement Resolution range (Å) Number of reflections in working set Number of reflections in test set Rcrysta (%) Rfreeb (%) B. Statistics of the final model Number of residues Number of water molecules rmsd from ideality Bond lengths (Å) Bond angles (deg.) Average B-values (Å2) Protein atoms Water molecules Ramachandran plot statisticsc Most favored (%) Allowed (%) Generously allowed (%) Disallowed (%)
19.67–2.60 3814 201 21.8 25.5 112 50 0.008 1.7 25.7 32.3 86.3 11.6 2.1 0.0
a Rcryst = ∑| |Fo| – |Fc||/∑|Fo|, where Fc is the calculated structure factor. b Rfree is as for Rcryst but calculated for a randomly selected 5.0% of reflections not included in the refinement. c As calculated by PROCHECK30 and by MOLPROBITY.31
312 β-sheets (Figure 1(a)). The first β-sheet is formed by β-strands G-F-C-C′, and the second by β-strands BE-D-CVV. A 310 α-helix is formed between β-strands E and F (Figure 1(a)). There are two disulphide bonds in IREM-1 (Figure 1(a)). The disulphide bond Cys40–Cys108 is conserved in the immunoglobulin-like domains and is found in other IREM-1 homologues such as murine and human TREM-1,20–22 human TLT-1,23 murine CLM-19 (PDB entry code: 1ZOX), NKp4424 and pIgR.25 This disulphide bond plays a major structural role in the Ig-V type immunoglobulins because it locks the two β-sheets together. Cys54 and Cys62 form the second disulfide bond in IREM-1. This disulfide bond is important for stabilizing the βhairpin formed by β-strands C-C′. This disulphide bond is found in CLM-1, NKp44, TLT-1 and pIgR, but is not present in murine and human TREM-1. Two hydrophobic cores are present in IREM-1. The first hydrophobic core is located within the two β-sheets towards the carboxy-terminal end of the protein chain and is composed of residues valine 28, 38, 121, 123, leucine 36, tryptophan 53, methionine 95, and tyrosine 106. The second hydrophobic core is located in the protrusion formed by the bottom part of β-strands F-C-C′ and the upper part of β-strands F-G. This core is formed by residues tryptophan 52, 59, and 107 (Figure 1(b)). A feature of the IREM-1 crystal structure is the formation of a prominent protrusion from the main immunoglobulin β-sandwich body (Figure 1(b)). This protrusion is about 18 Å long, 15 Å wide and 11 Å deep. It is noteworthy that this protrusion is present also in other Ig-V type receptors, such as NKp44, mouse and human TREM1, mouse CLM-1 and the pIgR. Therefore, this region could have some function in “accommodating” the potential ligand(s) recognized by this class of receptors (see later). The top of this protrusion is blocked, in part, by the side-chain of tryptophan 59 (Figure 1(b)). The other amino acid residues that are filling up this protrusion with their side-chains are isoleucine 64, aspartate 116, glutamate 11, lysine 67, tryptophan 52 and tryptophan 107 (Figure 1(b)). Calculation of the surface electrostatic potential for this protrusion shows that a central negatively charged surface is surrounded by an extensive positively charged surface. The hydrophobicity analysis shows that the central part of this protrusion is moderately hydrophobic and is surrounded by polar residues. By inspecting the IREM packing in the crystal, we observed that this protrusion is solvent-exposed and it is not occupied by symmetry-related molecules. This could point to a rather accessible surface for potential binding of hydrophilic and positively charged ligands. The three-dimensional structure of IREM-1 was used as a probe for searching the RCSB Protein Data Bank using the Vector Alignment Search Tool from NCBI.26 This search identified 60 non-redundant protein structures similar to IREM-1. Amongst these were the crystal structures of murine and human TREM-1,20–22 human TLT-1,23 murine CLM-1 (PDB
Structure of the IREM-1
entry code: 1ZOX), NKp4424 and pIgR.25 In fact, structure-based sequence alignment of the immunoglobulin-like domains between these representative structural homologues, showed good overall sequence identities (Figure 2). Not surprisingly, because it was the only model that produced a molecular replacement solution, the closest structural homologue of IREM-1 is CLM-1. The rmsd between the structurally equivalent Cα atoms forming the core of the immunoglobulin-like domain in IREM-1 and CLM-1 is 1.06 Å (Figure 1(c)), whereas the calculated rmsd values for IREM-1 superimposed on TLT-1 is 2.61 Å, the rmsd between IREM-1 and human TREM-1 is 4.47 Å, and the rmsd between IREM-1 and NKp44 is 2.58 Å. CDReq3 is the most variable loop in IREM-1 when compared with its murine homologous CLM-1 (Figure 1(c)). As seen in Figure 1(c), when IREM-1 is superimposed on CLM-1, the major structural changes are observed in the CDReq3 conformation and in the conformation of the β-hairpin formed by β-strands C-C′. In fact, the CDReq3 in IREM-1 is shifted about 4 Å and is rotated about 35° anticlockwise with respect to the CDReq3 of CLM-1 (Figure 3(c), circle), with almost no structural change in the immunoglobulin body. This variation could suggest an area of variability that may move in response to ligation. The electrostatic potential surface and hydrophobicity that describe the surface of the CDReq regions are different in IREM-1 and CLM-1 (Figure 3). The CDReq1 net surface charge in IREM-1 is −1.4 e− (Figure 3(a)), and represents a surface potential of 364 kJ mol−1. The CDReq1 in CLM-1 is more negative, with a surface charge of −2.8 e− (Figure 3(d)), corresponding to a surface potential of 656 kJ mol−1. The CDReq1 hydrophobicity in IREM-1 and CLM-1 is relatively similar (Figure 3(c) and (f)). Conversely, the CDReq2 net surface charge in IREM-1 is −2.5 e–, which is more negative (Figure 3(a)) than CLM-1, which is −0.6 e− (Figure 3(d)). This corresponds to an electrostatic surface potential of 643 kJ mol−1 and 198 kJ mol−1 for IREM-1 and CLM-1, respectively. The CDReq2 hydrophobicity in IREM-1 and CLM-1 is similar (Figure 3(c) and (f)). Finally, the net surface charge for the CDReq3 is nearly neutral in both polypeptides, being −0.6 e− and −0.3 e− for IREM-1 (Figure 3(a)) and CLM-1 (Figure 3(d)), respectively, that represent an electrostatic surface potential of 121 kJ mol−1 for IREM-1 and 87 kJ mol−1 for CLM1. The CDReq3 hydrophobicity is higher in IREM-1 (Figure 3(c)) than that of CLM-1 (Figure 3(f)). These physical differences could be the structural determinants for ligand discrimination and binding of these two homologous molecules if they bind their ligands in a fashion similar to that of antibodies and T-cell receptors. In an attempt to identify the region of IREM-1 that could accommodate ligands, a detailed analysis of surface cavities using a probe radius of 1.4 Å was carried out. This analysis revealed that four distinct solvent-exposed cavities, termed 1, 2, 3 and 4, are present in IREM-1 (Figure 4). None of these cavities,
Structure of the IREM-1
313
Figure 1. Overall structure of IREM-1, features of the extended protrusion and comparison with CLM-1. (a) A ribbon representation of the IREM-1 structure. IREM-1 has the expected V-set Ig domain fold present in the immunoglobulins. The secondary structure elements (β-strands and the 3 10 α-helix) were defined using the program DSSP.41 The antibody's equivalent CDR loops (CDReq) are shown in red and are labelled. Disulphide bonds are shown and are labelled with their respective cysteine residues. In this ribbon representation and in all the subsequent diagrams, N and C denote the amino terminus and the carboxy terminus of the protein chain. The location of the potential Nglycosylation site, which involves asparagine 88, is shown with a red diamond. (b) An important feature of the IREM-1 structure is the formation of a prominent protrusion that extends from the main immunoglobulin body. This protrusion is about 18 Å long, 15 Å wide and 11 Å deep. As seen in this Figure, the top of this protrusion is blocked, in part, by the side-chains of tryptophan 59, aspartate 116 and glutamate 111. The other amino acid residues present at this protrusion are isoleucine 64, lysine 67, tryptophan 52 and tryptophan 107. (c) Superimposition of IREM-1 with its murine homologous CLM-1 gives an rmsd of 1.06 for all the structurally equivalent Cα atoms forming the core of the immunoglobulin-like domain (102 residues). The orientation of IREM-1 and CLM-1 here is equivalent to that in (a). In this superimposition diagram, IREM-1 is in blue and CLM-1 is in red. The β-hairpin formed by β-strands C-C′ is labelled. The location of the CDReq loops is shown schematically. The circle highlights the prominent structural difference of the CDReq3 between IREM-1 and CLM-1.
Structure of the IREM-1
314
Figure 2. Structure-based sequence alignment of the extracellular immunoglobulin-like domains of IREM homologues. In this sequence, alignment of the secondary structural information was derived from the IREM-1 structure and with the program DSSP.40 The sequences shown are for the only immunoglobulin-like extracellular domain of human IREM-1 (NCBI AAP57942), human IREM-2 (NCBI NP_852114), human IREM-3 (NCBI AAV69612), human IRp60 (NCBI CAB66145), human NKp44 (NCBI AJ225109), human TREM-1 (NCBI NM_018643), human TLT-1 (NCBI BC100945), human CMRF-35 (NCBI X66171), mouse CLM-1 (NCBI AY457047), and a representative immunoglobulin V-domain of human TCR (NCBI NG_001332.2). The sequences are ordered from top to bottom on the basis of their sequence similarity as defined by the program CLUSTAL W.39 The numbering above the sequence corresponds to the unprocessed IREM-1 protein. Complete sequence conservation is labelled white with a red background, and red letters correspond to >60% sequence homology. The secondary structure elements are denoted by arrows (β-strands) and by squiggles (310 α-helix). The paired numbers at the bottom of the sequences indicate the bonded cysteine residue involved in the formation of the disulphide bridges. The sequences of the CDR-equivalent loops (CDReq) are labelled and boxed in black. The consensus sequence Asp-X-Gly/Asp-X-Tyr-X-Cys, which correspond to the signature of a V-type Ig domain is labelled with a star symbol above the sequence. The arrow denotes the potential N-glycosylation site in IREM-1 that involves asparagine 88.
as expected, involves the CDReq loops. Cavity 1 is located at the bottom of the protrusion that extends from the main immunoglobulin-like body (Figure 4). Cavities 2 and 3 are close to each other and are located towards the carboxy-terminal part of the molecule (Figure 4). Cavity 4 is located on the opposite site of the cavity 1 in a region that is beneath the CDReq2 loop (Figure 4). Cavity 1 is hydrophobic (Figure 4(a)) and possesses a negative electrostatic potential (Figure 4(c)). Cavity 2 is hydrophobic (Figure 4(a)) and has a positive electrostatic potential on one side and a negative electrostatic potential on the opposite side (Figure 4(c)). Cavity 3 is mainly hydrophilic (Figure 4(a)) and has an electrostatic surface potential like cavity 2 (Figure 4(c)). Cavity 4 is slightly hydrophobic (Figure 4(b)) with a neutral electrostatic surface potential (Figure 4(d)). Notably, cavities 1, 2 and 3 form a continuous groove on the IREM-1 surface (Figure 4(a) and (c)), which may point to an extended surface capable of accommodating hydrophilic and positive charged ligands. In summary, our structural results show that the immunoglobulin-like fold is conserved in IREM-1, and therefore is presumably conserved in the other immunoreceptors encoded on human chromosome 17 locus 17q25.1, but unique hydrophobicity and
electrostatic surface potentials could be used by these classes of immunoreceptors to achieve exquisite ligand binding specificity using the antibodyequivalent CDR loops, in a fashion analogous to that of T-cell receptors, or through unique structural cavities. Methods Cloning, protein expression, purification, crystallization and X-ray data collection is reported elsewhere.18 The structure was solved using the molecular replacement program Phaser,19 and the coordinates of the mouse CLM1 receptor extracellular domain (PDB code: 1ZOX) as the starting model. IREM-1 and CLM-1 immunoglobulin-like extracellular domains share 58.3% sequence identity. The molecular replacement search was carried out using data from 8–3.5 Å, and using space group P3121. Rigidbody refinement of the rotated and translated model using a resolution from 20 Å to3.0 Å gave an R-factor of 0.41 and an Rfree of 0.45. Several rounds of simulated annealing, conjugate gradient minimization and B-factor refinement followed by several iterations through individual steps of manual model building, maximum likelihood phase refinement and solvent molecule addition with XtalView,27 CNS28 and REFMAC29 have resulted in the current structure, which has a R-factor/Rfree of 0.21/0.25 for all data from 19.67–2.60 Å resolution. The final model
Figure 3. Electrostatic and hydrophobic features of the CDReq loops in IREM-1 and comparison with the murine homologous CLM-1. The view of these two molecules is from the top. The location of the three CDReq loops is shown schematically for both molecules. In (b) and (e), the CDReq1 surface is coloured pink, the CDReq2 surface is coloured orange and the CDReq3 surface is coloured green. This Figure highlights that fact that IREM-1 has unique antibody's equivalent complementary determinant region surface shape, electrostatic potentials and hydrophobicity (see the text for details). The electrostatic potential is coloured from −5 kJ mol−1 Å−2 (red, negative potential) to 5 kJ mol−1 Å−2 (blue, positive potential). Neutral potential (0 kJ mol−1 Å−2) is shown in white. The hydrophobicity is coloured over a scale of green (most hydrophobic amino acids) to white (least hydrophobic amino acids).
Structure of the IREM-1
315
316
Structure of the IREM-1
Figure 4. Cavities on IREM-1. The cavities on IREM-1 are shown as yellow dots and are labelled from 1 to 4. The locations of the CDReq loops are shown schematically and are labelled only in the ribbon diagram. In (b) and (d) the continuous groove formed by cavities 1, 2 and 3 is highlighted. This groove, which possesses a negative potential and is highly hydrophobic, may accommodate potential ligands. The electrostatic potential surfaces and hydrophobicity are coloured as in Figure 3. The ribbon orientation is equivalent to that of Figure 1(a). contains 112 amino acid residues and 50 water molecules. The first three amino acid residues are derived from the vector N-terminal tag, while the remaining 109 correspond to residues 21–129 of the IREM-1 sequence.
Model geometry was examined using PROCHECK30 and MOLPROBITY.31 Structure superimpositions were performed with Superpose,32 and visually inspected using the program XtalView.27 Structural similarities were
Structure of the IREM-1 analyzed using the program VAST.26 Solvent-accessible surface areas were calculated with CNS,28 using a probe radius of 1.4 Å. Electrostatic potentials were calculated in an aqueous solution at pH 7.0 and with 150 mM NaCl using the program APBS.33 The hydrophobicity surfaces were calculated by using the optimized matching hydrophobicity program PROTSCALE.34 Surface cavities were calculated with the programs CASTp,35 CAVER,36 and LIGSITEcsc,37 using a probe radius of 5 Å. The Figures were prepared with the program Pymol (DeLano Scientific, San Carlos, CA)†, and rendered with the GIMP image manipulation program‡. The N-glycosylation sites were analyzed with the program GlyPro.38 Sequence alignment was carried out using the program CLUSTAL W,39 and was rendered using the program Espript.40
317
7.
8.
9.
Protein Data Bank accession code Atomic coordinates and stricture factors have been deposited in the RCSB Protein Data Bank§ under the accession code 2NMS.
10.
11.
Acknowledgements We thank Professor Roy Mariuzza (Center for Advanced Research in Biotechnology, Rockville, Maryland, USA) for critical reading of the manuscript. This work was supported, in part, by the Fondazione Gerolamo Galini (Genova, Italy). We thank Alessandro Faravelli (Laboratorio di Medicina Molecolare, Istituto Giannina Gaslini, Genova, Italy) for help during protein expression. We thank the constant scientific support from Professor Lorenzo Moretta (Istituto Giannina Gaslini, Genova, Italy).
12.
References
15.
1. Natarajan, K., Dimasi, N., Wang, J., Mariuzza, R. A. & Margulies, D. H. (2002). Structure and function of natural killer cell receptors: multiple molecular solutions to self, nonself discrimination. Annu. Rev. Immunol. 20, 853–885. 2. Ravetch, J. V. & Lanier, L. L. (2000). Immune inhibitory receptors. Science, 290, 84–89. 3. Humphrey, M. B., Lanier, L. L. & Nakamura, M. C. (2005). Role of ITAM-containing adapter proteins and their receptors in the immune system and bone. Immunol. Rev. 208, 50–65. 4. van den Berg, T. K., Yoder, J. A. & Litman, G. W. (2004). On the origins of adaptive immunity: innate immune receptors join the tale. Trends Immunol. 1, 11–16. 5. Pandey, S. & Agrawal, D. K. (2006). Immunobiology of Toll-like receptors: emerging trends. Immunol. Cell Biol. 84, 333–341. 6. Dietrich, J., Nakajima, H. & Colonna, M. (2000).
† http://pymol.sourceforge.net/ ‡ http://www.gimp.org/ § http://www.rcsb.org/
13.
14.
16.
17.
18.
19. 20.
Human inhibitory and activating Ig-like receptors which modulate the function of myeloid cells. Microbes Infect. 3, 323–329. Colonna, M., Navarro, F., Bellon, T., Llano, M., Garcia, P., Samaridis, J. et al. (1997). A common inhibitory receptor for major histocompatibility complex class I molecules on human lymphoid and myelomonocytic cells. J. Expt. Med. 186, 1809–1818. Fujioka, Y., Matozaki, T., Noguchi, T., Iwamatsu, A., Yamao, T., Takahashi, N. et al. (1996). A novel membrane glycoprotein, SHPS-1, that binds the SH2domain-containing protein tyrosine phosphatase SHP-2 in response to mitogens and cell adhesion. Mol. Cell Biol. 12, 6887–6899. Chung, D. H., Humphrey, M. B., Nakamura, M. C., Ginzinger, D. G., Seaman, W. E. & Daws, M. R. (2003). CMRF-35-like molecule-1, a novel mouse myeloid receptor, can inhibit osteoclast formation. J. Immunol. 171, 6541–6658. Clark, G. J., Green, B. J. & Hart, D. N. (2000). The CMRF-35H gene structure predicts for an independently expressed member of an ITIM/ITAM pair of molecules localized to human chromosome 17. Tissue Antigens, 55, 101–109. Green, B. J., Clark, G. J. & Hart, D. N. (1998). The CMRF-35 mAb recognizes a second leukocyte membrane molecule with a domain similar to the poly Ig receptor. Int. Immunol. 10, 891–899. Cantoni, C., Bottino, C., Augugliaro, R., Morelli, L., Marcenaro, E., Castriconi, R. et al. (1999). Molecular and functional characterization of IRp60, a member of the immunoglobulin superfamily that functions as an inhibitory receptor in human NK cells. Eur. J. Immunol. 10, 3148–3159. Alvarez-Errico, D., Aguilar, H., Kitzig, F., Brckalo, T., Sayos, J. & Lopez-Botet, M. (2004). IREM-1 is a novel inhibitory receptor expressed by myeloid cells. Eur. J. Immunol. 34, 3690–3701. Aguilar, H., Alvarez-Errico, D., Garcia-Montero, A. C., Orfao, A., Sayos, J. & Lopez-Botet, M. (2004). Molecular characterization of a novel immune receptor restricted to the monocytic lineage. J. Immunol. 173, 6703–6711. Martinez-Barriocanal, A. & Sayos, J. (2006). Molecular and functional characterization of CD300b, a new activating immunoglobulin receptor able to transduce signals through two different pathways. J. Immunol. 177, 2819–2830. Allcock, R. J., Barrow, A. D., Forbes, S., Beck, S. & Trowsdale, J. (2003). The human TREM gene cluster at 6p21.1 encodes both activating and inhibitory single IgV domain receptors and includes NKp44. Eur. J. Immunol. 33, 567–577. Bouchon, A., Dietrich, J. & Colonna, M. (2000). Cutting edge: inflammatory responses can be triggered by TREM-1, a novel receptor expressed on neutrophils and monocytes. J. Immunol. 164, 4991–4995. Dimasi, N., Flot, D., Dupex, F. & Marquez, J. A. (2007). Expression, crystallization and X-ray data collection from microcrystals of the extracellular domain of the human inhibitory receptor expressed on myeloid cells IREM-1. Acta Crystallog. sect. F. (In the press). Storioni, L. C., McCoy, A. J. & Read, J. R. (2004). Likelihood-enhanced fast rotation functions. Acta Crystallog. sect. D, 60, 432–438. Kelker, M. S., Debler, E. W. & Wilson, I. A. (2004). Crystal structure of mouse triggering receptor expressed on myeloid cells 1 (TREM-1) at 1.76 Å. J. Mol. Biol. 344, 1175–1181.
Structure of the IREM-1
318 21. Kelker, M. S., Foss, T. R., Peti, W., Teyton, L., Kelly, J. W., Wuthrich, K. & Wilson, I. A. (2004). Crystal structure of human triggering receptor expressed on myeloid cells 1 (TREM-1) at 1.47 Å. J. Mol. Biol. 342, 1237–1248. 22. Radaev, S., Kattah, M., Rostro, B., Colonna, M. & Sun, P. D. (2003). Crystal structure of the human myeloid cell activating receptor TREM-1. Structure, 11, 1527–3155. 23. Gattis, J. L., Washington, A. V., Chisholm, M. M., Quigley, L., Szyk, A., McVicar, D. W. & Lubkowski, J. (2006). The structure of the extracellular domain of triggering receptor expressed on myeloid cells like transcript-1 and evidence for a naturally occurring soluble fragment. J. Biol. Chem. 281, 13396–13403. 24. Cantoni, C., Ponassi, M., Biassoni, R., Conte, R., Spallarossa, A., Moretta, A. et al. (2003). The threedimensional structure of the human NK cell receptor NKp44, a triggering partner in natural cytotoxicity. Structure, 11, 725–734. 25. Hamburger, A. E., West, A. P., Jr & Bjorkman, P. J. (2004). Crystal structure of a polymeric immunoglobulin binding fragment of the human polymeric immunoglobulin receptor. Structure, 11, 1925–1935. 26. Gibrat, J. F., Madej, T. & Bryant, S. H. (1996). Surprising similarities in structure comparison. Curr. Opin. Struct. Biol. 3, 377–385. 27. McRee, D. E. (1999). XtalView/Xfit. A versatile program for manipulating atomic coordinates and electron density. J. Struct. Biol. 125, 156–165. 28. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998). Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallog. sect. D, 54, 905–921. 29. Collaborative Computational Project, Number 4. (1994). The CCP4 suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760–763. 30. Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Thornton, J. M. (1993). PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallog. 26, 283–290. 31. Davis, I. W., Murray, L. W., Richardson, J. S. & Richardson, D. C. (2004). MOLPROBITY: structure
32.
33.
34.
35.
36.
37.
38. 39.
40. 41.
validation and all-atom contact analysis for nucleic acids and their complexes. Nucl. Acids Res. 32, W615–W619. Maiti, R., Van Domselaar, G. H., Zhang, H. & Wishart, D. S. (2004). SuperPose: a simple server for sophisticated structural superposition. Nucl. Acids Res. 32, W590–W594. Baker, N. A., Sept, D., Joseph, S., Holst, M. J. & McCammon, J. A. (2001). Electrostatics of nanosystems: application to microtubules and the ribosome. Proc. Natl Acad. Sci. USA, 18, 10037–10041. Sweet, R. M. & Eisenberg, D. (1983). Correlation of sequence hydrophobicities measures similarity in three-dimensional protein structure. J. Mol. Biol. 171, 479–488. Dundas, J., Ouyang, Z., Tseng, J., Binkowski, A., Turpaz, Y. & Liang, J. (2006). CASTp: computed atlas of surface topography of proteins with structural and topographical mapping of functionally annotated residues. Nucl. Acids Res. 34, W116–W118. Petřek, M., Otyepka, M., Banáš, P., Košinová, P., Koča, J. & Damborský, J. (2006). CAVER: a new tool to explore routes from protein clefts, pockets and cavities. BMC Bioinformatics, 7, 316. Huang, B. & Schroeder, M. (2006). LIGSITEcsc: predicting ligand binding sites using the Connolly surface and degree of conservation. BMC Struct. Biol. X, 6–19. Bohne-Lang, A. & von der Lieth, C. W. (2005). GlyProt: in silico glycosylation of proteins. Nucl. Acids Res. 33, W214–W219. Higgins, D., Thompson, J., Gibson, T., Thompson, J., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl. Acids Res. 22, 4673–4680. Gouet, P., Courcelle, E., Stuart, D. I. & Metoz, F. (1999). ESPript: multiple sequence alignments in PostScript. Bioinformatics, 15, 305–308. Kabsch, W. & Sander, C. (1983). Dictionary of protein secondary structure: pattern recognition of hydrogenbonded and geometrical features. Biopolymers, 22, 2577–2637.
Edited by I. Wilson (Received 15 November 2006; received in revised form 21 December 2006; accepted 3 January 2007) Available online 10 January 2007