Crystal structure of chemically synthesized vMIP-II.
Ying Li,Dongxiang Liu,Rong Cao,Santosh Kumar,Changzhi Dong,Jing An,Scott R Wilson,Yi-Gui Gao,Ziwei Huang
DOI: https://doi.org/10.1002/prot.21172
2007-01-01
Abstract:The viral inflammatory protein II (vMIP-II) is a mammalian chemokine homolog encoded by Kaposi's sarcoma-associated herpesvirus (KSHV).1, 2 Chemokines constitute the largest family of cytokines that play important roles in immune responses. The main biological function of chemokines is to mediate the recruitment of leukocytes to the site of inflammation and their activation.3 The activity of chemokines is effected by binding to the G-protein coupled chemokine receptors present on various subpopulations of leukocytes such as monocytes, macrophages, basophils, eosinophils, and lymphocytes.4 Based on the positions of the first two conservative cysteine residues, chemokines are classified into four subfamilies: CXC, CC, CX3C, and C. As a viral chemokine, vMIP-II shows a broad spectrum of binding to CXC, CC, and CX3C chemokine receptors. In particular, vMIP-II binds to both CXCR4 and CCR5, which are the only two major coreceptors of human immunodeficiency virus Type 1 (HIV-1). Although most mammalian chemokines are agonists to their corresponding receptors, vMIP-II is in most cases an antagonist to its receptors including CXCR4 and CCR5.2 All ligands of CXCR4 and CCR5 show anti-HIV activity in vitro by competing with HIV-1 for the binding of the coreceptors though some are more potent than others.5 The structural basis of the interaction between chemokines and their receptors was subjected to intense investigation in recent years. However, many of the details are still not known. Here, we report the crystal structure of vMIP-II in two crystal forms different from previously reported and discuss the structural features that are possibly related to the binding affinity and specificity. vMIP-II, viral macrophage inflammatory protein-II; MCP-2, monocyte chemoattractant protein-2; CXCR4, CXC chemokine receptor 4; CCR5, CC chemokine receptor 5; RMSD, root mean square deviations. vMIP-II was chemically synthesized in our lab according a previously published protocol.6 The chemically synthesized vMIP-II was crystallized by hanging drop vapor diffusion using the condition reported by Fernandez et al.7 Protein (10 mg/mL) was mixed at 1:1 ratio with reservoir solution, which is 100 mM sodium citrate buffer (PH = 5.6), 11% (w/v) PEG 4000, and 11% (w/v) 2-propanol. The data sets were collected at the 19-BM beamline of the Advanced Photon Source at Argonne National Laboratory (Argonne, IL) and the X12C beamline of Light Source (LS) at Brookhaven National Laboratory (Upton, NY). The crystals were first soaked in cryoprotectant (35% sucrose) briefly and frozen in liquid nitrogen. The data was processed with HKL2000 software package.8 The tetragonal crystal data set is 98.6% complete, consisting of 8305 unique reflections in the resolution range of 30–1.7 Å. The monoclinic crystal data set is 89.1% complete, consisting of 10,419 unique reflections in the resolution range of 30–2.3 Å. The crystal parameters and data collection conditions are summarized in Table I. The crystal structure of vMIP-II was solved by molecular replacement using the program AmoRe.9 For tetragonal crystal form, the A chain of the vMIP-II dimer (PDB code 1CM9) was used as the search model. The rigid body refinement was applied to the initial model. Then the structure was refined by simulated annealing using X-PLOR.10 The missing sidechains of the initial model were added to the structure with the program O11 whenever the corresponding electron density could clearly be seen in the 2Fo–Fc map. Solvent molecules were finally added and verified by electron density map. During the process of further refinement, solvent molecules with B-factors higher than 70 were removed. The vMIP-II structures in tetragonal and monoclinic crystal forms were determined to 1.7 and 2.3 Å, respectively, by molecular replacement method. Although the crystallization condition is the same as reported by Fernandez et al.,7 we did not obtain the monoclinic crystals belonging to the same space group as previously reported.7 Instead, we obtained both monoclinic crystals belonging to a different space group and tetragonal crystals. Similar phenomenon has been observed on the CC chemokine MCP-1.12 In the previously reported monoclinic crystal, a noncrystallographic dimer was observed in the asymmetric unit, whereas our tetragonal crystal contains a monomer and our monoclinic crystal contains two dimers in the asymmetric unit. The overall monomer structure is shown in Figure 1(A). The structure refinement statistics is summarized in Table II. An example of electron density map is shown in Figure 1(B). The tertiary structure is consistent with many previously reported CC chemokine structures. The flexible N-terminus and N-loop are followed by three antiparallel β-strands (residues 25–30, 39–44, 49–53) arranged in a Greek key motif with a C-terminal α-helix (residues 57–65) laid on top of them. Interestingly, a water molecule buried inside the protein was observed in both crystal forms, which is hydrogen-bonded to the mainchain carbonyl of Lys37 and the sidechain hydroxyl group of Thr31. It is conservative in different crystals as it is also present in the previously reported crystal structure.7 This water molecule may help to stabilize the conformation of the 30's loop between the first and the second β-strands. The RMSDs of Cα atoms between our structure and the previously reported crystal structure (PDB ID 1CM9) have been calculated. The RMSDs for chain A and chain B of the tetragonal crystal form are 0.56 and 0.52 Å, respectively. For the monoclinic crystal form, the RMSDs are 0.55, 0.42, 0.44, and 0.46 Å for A, B, C, and D chains, respectively. (A) The structural diagram of vMIP-II monomer containing residues 4–71 (tetragonal crystal form). (B) The residues 11–16 of vMIP-II monomer in tetragonal crystal form with 2Fo–Fc electron density map contoured at 1.0σ. The figure was produced by program O.11 (C) The solvent accessible surface of vMIP-II (this work), MIP-1α (PDB code 1B53), MIP-1β (PDB code 1HUM), RANTES (PDB code 1B3A), and MCP-2 (PDB code 1ESR). The surfaces are colored according to surface electrostatic potential calculated by DelPhi.13 Red color represents negative electrical potential and blue color represents positive electrical potential. The N-loop regions of each structure were circled. (D) Sequence alignment of vMIP-II, MIP-1α, MIP-1β, RANTES, and MCP-2. The N-loop loop region is represented in bold letters. By examining the molecules in neighboring asymmetric unit, it was found that vMIP-II forms crystallographic dimers in our tetragonal crystal through interactions between N-termini, which is the interaction pattern also observed in the asymmetric unit of previously reported crystal form. The dimer is stabilized by six pairs of H-bonds, among which the H-bond between sidechain carbonyl group of Gln16 and mainchain amide of Ser4 was not observed previously. There are also weak interactions between two dimers mainly through the contact of the N-terminus of one molecule with the C-terminus of a neighboring molecule. The C-terminal carboxyl group is hydrogen-bonded to the mainchain oxygen atom of Arg7 of the neighboring molecule. The interactions are also reflected on the relatively low B-factors of C-terminus, suggesting the stabilization of C-terminus by the neighboring molecule. The buried surface area is ∼300 Å2 for each dimer due to interactions on the dimer–dimer interface. Similar weak interactions are also observed in our monoclinic crystals but the relative orientation of two dimers is slightly different. In contrast, dimers do not interact with each other in the previously reported crystal structure. Therefore, vMIP-II seems to have a strong tendency to form dimers in a crystal environment but the interactions between dimers are weak and nonspecific. One of the motivations to study chemokine structures is to understand the structural basis for chemokines to interact with their receptors. According to the two-site binding model,14-16 the N-loop after the first pair of conservative cysteines participates in the initial ligand-receptor recognition. Among multiple mammalian chemokine ligands of CCR5, including MIP-1α, MIP-1β, RANTES, and MCP-2,17, 18 vMIP-II shares the highest sequence identity with MIP-1α and MIP-1β (52 and 47%, respectively). The sequence alignment of the five CCR5 ligands is shown in Figure 1(D). In the N-loop region, vMIP-II shares the most similar surface topology and electrostatic potential with RANTES among all the CCR5 ligands [Fig. 1(C)]. For these two ligands, the N-loop and surrounding regions are, to a large extent, positively charged, while MIP-1α and MIP-1β are almost apolar in the corresponding regions. Arg18 is one of the conserved residues in the N-loop region. The importance of this Arg18 was confirmed in MIP-1β by mutagenesis studies.19 However, it is not necessarily essential for all the ligands with RANTES as an example.20 For MIP-1α and MIP-1β, the sidechain of this Arg18 protrudes far out of the protein surface, whereas in all other ligands it assumes a conformation bending toward the surface. The extended sidechain conformations in MIP-1α and MIP-1β possibly facilitate the charge–charge interaction between ligands and receptors considering the aforementioned fact that the N-loop regions are relatively apolar for these two CCR5 ligands. Overall, the structural comparison of the N-loop region suggests that high sequence identity and the presence of conserved residues do not necessary lead to similar local geometry and electrostatic potential in the functional sites, which is probably one of the reasons why the structure scaffolds of all chemokines are very similar but the binding specificity and affinity are distinct. We thank Dr. Randy Alkire at SBC 19-BM beamline, Argonne National Laboratory, for his help on the data collection and Dr. Howard Robinson at Brookhaven National Laboratory for collecting the monoclinic crystal data set. The use of the Argonne National Laboratory Structural Biology Center beamlines at the Advanced Photon Source is supported by the U. S. Department of Energy, Office of Energy Research, under Contract No. W-31-109-ENG-38.