Structural Elements of the Signal Propagation Pathway in Squid

ARTICLE pubs.acs.org/JPCB

Structural Elements of the Signal Propagation Pathway in Squid Rhodopsin and Bovine Rhodopsin Minoru Sugihara,* Wataru Fujibuchi, and Makiko Suwa* Computational Biology Research Center (CBRC), National Institute of Advanced Industrial Science and Technology (AIST), AIST Tokyo Waterfront BIO-IT Research Building, 2-4-7 Aomi, Koto-ku, Tokyo 135-0064, Japan ABSTRACT: Squid and bovine rhodopsins are G-protein coupled receptors (GPCRs) that activate Gq- and Gt-type G-proteins, respectively. To understand the structural elements of the signal propagation pathway, we performed molecular dynamics (MD) simulations of squid and bovine rhodopsins plus a detailed sequence analysis of class A GPCRs. The computations indicate that although the geometry of the retinal is similar in bovine and squid rhodopsins, the important interhelical hydrogen bond networks are different. In squid rhodopsin, an extended hydrogen bond network that spans ∼13 Å to Tyr315 on the cytoplasmic site is present regardless of the protonation state of Asp80. In contrast, the extended hydrogen bond network is interrupted at Tyr306 in bovine rhodopsin. Those differences in the hydrogen bond network may play significant functional roles in the signal propagation from the retinal binding site to the cytoplasmic site, including transmembrane helix (TM) 6 to which the G-protein binds. The MD calculations demonstrate that the elongated conformation of TM6 in squid rhodopsin is stabilized by salt bridges formed with helix (H) 9. Together with the interhelical hydrogen bonds, the salt bridges between TM6 and H9 stabilize the protein conformation of squid rhodopsin and may hinder the occurrence of large conformational changes that are observed upon activation of bovine rhodopsin.

’ INTRODUCTION G-protein coupled receptors (GPCRs) are the target of ∼30% of total marketed drugs. Consequently, much effort has been put into solving their structures and understanding their mechanisms of action.1,2 The first breakthrough in the structural biology of GPCRs was the determination of the crystal structure of bovine rhodopsin.3 7 Remarkable advances have taken place in the past 2 years. The structures of squid rhodopsin8,9 and bovine opsin,10,11 as well as the crystal structures of ligand-mediated GPCRs (adrenergic, adenosine, dopamine, and chemokine receptors), were solved and deposited in the Protein Data Bank.12 16 Those crystal structures enable the detailed analysis of the structural determinants of GPCR function. Squid and bovine rhodopsins contain the same ligand molecule, retinal, but activate different types of G-proteins: squid rhodopsin activates Gq-type G-proteins, whereas bovine rhodopsin activates Gt-type G-proteins. Although the conformation of retinal in squid rhodopsin is 6-s-cis, 11-cis, 12-s-trans (i.e., the same as that found in bovine rhodopsin), Murakami and Kouyama have shown that the retinal polyene chain has a less distorted configuration in squid rhodopsin than in bovine rhodopsin.8 The hydrogen-bonding partner of the retinal is either the carbonyl group of Asn87, or the hydroxyl group of Tyr111. In contrast, in bovine rhodopsin the counterion of the retinal is the negatively charged Glu113. The absorption of a photon causes the isomerization of retinal to the all-trans configuration. This geometrical change triggers the propagation of a signal to the cytoplasmic surface where the coupling with G-proteins and the activation occur.17 The retinal r 2011 American Chemical Society

binding pockets of squid and bovine rhodopsins contain a highly conserved tryptophan residue in transmembrane helix (TM) 6 near the β-ionone ring facing an interhelical cavity filled with water molecules. Analysis of crystallized GPCRs has revealed the presence of internal water molecules and their interaction with conserved motifs is associated with the activation of GPCRs.17 21 Water molecules in the cavity are surrounded by the conserved residues, including an aspartate residue (Asp83 in bovine rhodopsin or Asp80 in squid rhodopsin). Fourier transform infrared (FTIR) experiments on bovine rhodopsin have shown that Asp83 is protonated and the hydrogen bond network around Asp83 is important for the selective stability in the meta-II state.22 Meanwhile in squid rhodopsin, hydrogen bonds are formed between Asp80 and water molecules present in the cytoplasmic cavity but the protonation state of Asp80 remains an important open question. One remarkable difference between the crystal structures of squid and bovine rhodopsins is that TM5 and TM 6 of squid rhodopsin protrude into the cytoplasmic medium and are longer than those of bovine rhodopsin in the inactive dark state. Squid rhodopsin contains 12 more residues in cytoplasmic loop (CL) 3 than bovine rhodopsin and the crystal structure of squid rhodopsin reveals a helical structure in this region (TM5 and TM6). Furthermore, squid rhodopsin possesses an additional helix (H) 9 located after H8 (See Figure 1). Received: October 24, 2010 Revised: April 1, 2011 Published: April 21, 2011 6172

dx.doi.org/10.1021/jp1101785 | J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B

Figure 1. Structures of squid (on the left side in green) and bovine (on the right side in blue) rhodopsins with key residues and water molecules (in blue). In squid rhodopsin (on the left side), residues are colored according to their conservation rates in GPCRs in red (more than 80%), in yellow (more than 60%), or in magenta (the rest). In bovine rhodopsin (on the right side), aromatic residues appear in orange. Superscripts in the residues of bovine rhodopsin indicate Ballesteros Weinstein numbering for conserved GPCR residues.40 The most conserved residue in a transmembrane helix (number x) is denoted by x.50 and all other residues on the same transmembrane helix are numbered relative to the x.50 reference residue.

Although X-ray crystallography yields invaluable information on proteins, because of the limited resolution, it is difficult to solve the geometries of retinal and water molecules. In particular, water molecules are difficult to characterize from X-ray crystallographic data. GPCRs are also highly flexible molecules23 and computations have contributed to the establishment of reliable protein models and the analysis of dynamical information. However, only a few theoretical investigations have been performed for squid rhodopsin.24 26 We address the question of how specific structural elements of squid and bovine rhodopsins contribute to the signal propagation pathway by performing combined quantum mechanical/ molecular mechanical (QM/MM) MD studies of squid and bovine rhodopsins, plus the sequence analysis of GPCRs. We focus on three structural features of bovine and squid rhodopsins. First we examine the geometry of retinal in the binding pocket and the retinal interactions that could participate in relaying structural perturbation upon retinal photoisomerization. Second, we assess the conservation and pattern of the hydrogen bond network that extends from the retinal binding pocket to the cytoplasmic site. Third, we determine the interactions that stabilize the elongated geometry of TM5 and TM6 to which G-proteins bind.

’ METHODS QM/MM MD Simulations. The simulation was performed with CHARMM package. Retinal (60 atoms including the hydrogen link atom) was calculated quantum mechanically by employing the SCC-DFTB method as implemented in CHARMM.27 29 The SCC-DFTB method has been successfully applied to QM/MM MD simulations of retinal proteins: the refinement of the geometry of retinal in the binding pocket of bovine rhodopsin6,30 or the mechanism of proton pumping of bacteriorhodopsin.31 A recent

ARTICLE

work on MD simulations of squid rhodopsin demonstrated that the force-field parameters used to describe retinal and its interactions have a significant effect not only on the structure and dynamics of retinal but also on the dynamics of the surrounding protein groups and water molecules.25 A reliable quantum mechanical (QM) description of retinal is necessary to assess the hydrogen bond network that extends from the retinal binding site to the cytoplasmic side of the protein. As the starting coordinates, we used chain A from the crystal structure of squid rhodopsin obtained by Murakami et al. (PDB: 2Z73).8 Water molecules were described with the TIP3P model.32 All titratable residues were modeled in their standard protonation states except Asp80, whose likely protonation state was investigated in a separate set of computations. Cys108 and Cys186 are linked via a disulfide bond. To conduct a direct comparison with the previous MD study on bovine rhodopsin,6 we used a model protein that contained no lipids and applied a harmonic constraint to the peptide backbone atoms inside the lipid region to maintain the shape of the protein. These constrained residues were 31 61 in TM1, 68 98 in TM2, 106 138 in TM3, 148 172 in TM4, 195 224 in TM5, 258 286 in TM6, and 294 317 in TM7. A harmonic constraint was also applied to the C- and N-termini and the sulfur atom of Cys337 to which a palmitoyl group was attached. The time step was 1 fs and the simulation was prolonged to 10 ns at 300 K (Nos
e canonical ensemble).33 Although, ideally, MD studies of GPCRs would be performed with a lipid membrane,24,25,34,35 it is difficult to combine the QM treatment of retinal with the embedding of the lipid environment due to the high computational costs. In addition, recent MD simulations of squid rhodopsin with a lipid membrane have indicated that the force-field parameters used to describe retinal and its interactions have a significant effect not only on the structure and dynamics of retinal but also on the dynamics of the sorrowing protein groups and water molecules.25 In consequence, we perform MD simulations by using a reliable QM method with a harmonic constraint. Statistical Analysis. We retrieved sequences of class A GPCRs (26 vertebrate species) from the SEVENS database (http://sevens.cbrc.jp).36,37 To identify the sequences, we compared the sequences in SEVENS and in SwissProt38 on the basis of two selection criteria: sequence identity and coverage. If two sequences in SEVENS and SwissProt have more than 80% sequence identity and more than 90% coverage of the whole sequence, we assumed that those two sequences are identical and used them statistical analysis. By using those criteria, 1314 sequences were identified. Sequences containing unknown residues (31 sequences) and/or stop codons (11 sequences) were excluded and finally 1272 sequences of 48 subfamilies except olfactory and gustatory receptors were selected (26 species). Eighty-one of the 1272 sequences were vertebrate opsins. Five invertebrate sequences (Japanese flying squid, northern European squid, squid, common cuttlefish, and giant octopus) were chosen from SwissProt.38 The transmembrane segments of the sequences were determined based on the available crystal structures of bovine rhodopsin,3 7 two adrenergic receptors,12,13 and an adenosine receptor14 to which multiple sequence alignments were performed by mafft39 and assigned with the avoidance of gaps in the transmembrane domains.

’ RESULTS AND DISCUSSION Overview of Conserved Residues. The statistical analysis of amino acid of the protein family indicates what residues have 6173

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B

ARTICLE

Table 1. Conservation Rates of Residues in 1272 Sequences of GPCRs and 81 Sequences of Opsinsa

a The most conserved residues appear in the first entry, although the list is not complete in the order of decreasing size. In this study, we are not interested in the residues in parentheses. Residues that are conserved particularly in opsins appear in gray. Sequences were retrieved from the SEVENS database (http://sevens.cbrc.jp).36,37

critical roles in the family members or subgroups. The conservation rates of key residues in the retinal binding pocket and near the internal water molecules are listed in Table 1. The first and second columns are the residue positions in squid and bovine rhodopsins, respectively, and the third and fourth columns are the conservation rates of the residues in 1272 sequences of GPCRs and in 81 sequences of vertebrate opsins, respectively. The superscript indicates the Ballesteros Weinstein numbering for GPCR residues40 (see legend in Figure 1). The residues with the Ballesteros Weinstein numbering represents that of bovine rhodopsin. The retinal molecule is covalently attached to Lys at 7.43 and the conservation rate is only 6% in GPCRs (the largest population is Tyr: 36%). In contrast, this lysine residue is completely conserved (100%) in vertebrate and invertebrate opsins (Lys305 in squid rhodopsin or Lys2967.43 in bovine rhodopsin). In opsins, the counterion position (3.28) is occupied by a charged or polar residue (Glu 67%, Tyr 19%, and Asp 7%) that stabilizes the conjugated bond and the absorption maximum is shifted in the visible wavelength region to around 500 nm. Most GPCRs do not have a charged or polar residue (Glu 4% and Tyr 3%) at this position. The retinal binding pocket of squid rhodopsin contains many hydrophobic residues,8,9 most of which are also found in bovine rhodopsin.3 7 Seven positions near the β-ionone ring are occupied by aromatic residues in both squid and bovine rhodopsins: Phe205 (Phe2085.43), Phe209 (Phe2125.47), Phe270 (Phe2616.44), Trp274 (Trp2656.48), and Tyr277 (Tyr2686.51) in TMs, and Tyr177 (Tyr178EL2) and Tyr190 (Tyr191EL2) in extracellular loop (EL) 2. The sum of the conservation rates of aromatic residues (Tyr, Trp or Phe) in GPCRs at these positions are 27%, 80%, 91%, 94%, 70%, 15%, and 46% in the order of listed above. The first phenylalanine residue (at the 5.43 position) and the last two tyrosine residues (in EL2) are opsin-specific as these positions are highly occupied by one of the aromatic residues in opsin groups.

Squid and bovine rhodopsins share 30% sequence identity on average in the TM regions. Most of the conserved residues are localized in the helical domains, including the D(E)R3.50Y motif in TM3, the CWxP6.50Y motif in TM6, and the NP7.50xxY motif in TM7. The most conserved residue is Pro in the CWxP6.50Y motif (100% in GPCRs). The residues with the second and third highest conservation rates are Asn1.50 in TM1 (99%) and Asp2.50 in TM2 (97%), respectively. The side chains of these residues project into the interhelical cavity. This cavity is filled with water molecules and surrounded by highly conserved residues: Trp274 (Trp2656.48), Asn311 (Asn3027.49), and Tyr315 (Tyr3067.53). The calculated conservation rates in GPCRs are 78%, 80%, and 92% in the order of sequence numbering (See Table 1). Concerning the loop regions, significant differences between squid and bovine rhodopsins are found in the regions of CL3. The amino acid sequence in CL3 is completely conserved in five invertebrate rhodopsins that have ten charged residues (seven acidic and three basic residues) in this region; bovine rhodopsin has only three (two acidic residues and one basic residue). The analysis of vertebrate GPCRs shows that the length of CL3 varies considerably (average length: 34), ranging from 1 (C5a anaphylatoxin chemotactic receptor C5L2) to 230 (M3 acetylcholine muscalinic receptor). On the other hand, the other loops have nearly the same lengths and the average lengths are 7 (EL1), 24 (EL2), 13 (EL3), 7 (CL1), and 12 (CL2). After H8, squid rhodopsin possesses H9, which is lacking in bovine rhodopsin, and this region is rich in charged residues (six or seven acidic residues and one basic residue). In contrast, the sequence of bovine rhodopsin contains only three acidic residues in the corresponding region after H8. Retinal Geometry and Retinal Binding Pocket. An accurate description of retinal geometry is essential to understanding the activation mechanism of rhodopsin, because retinal geometry is the key structural determinant of the ultrafast isomerization reaction41 and the maximum absorbance,42 and the isomerization 6174

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B

ARTICLE

Figure 3. Overlay of residues in the retinal binding pockets from MD studies. Retinal and residues in squid rhodopsin (bovine rhodopsin) are in gray (red) and helixes of squid rhodopsin (bovine rhodopsin) are in green (blue). The conservation rates of the residues are indicated. Phe188 in squid rhodopsin (in cyanine) that occupies the Tyr191 position in bovine rhodopsin is shown in magenta.

Figure 2. Internal coordinates of retinals in squid rhodopsin and bovine rhodopsin: (a) bond lengths, (b) bond angles, and (c) dihedral angles. Retinals in crystal and calculated structures are shown by broken and solid lines, respectively (squid rhodopsin in green and bovine rhodopsin in blue).

of retinal initiates the downstream signal propagation. We used QM/MM computations to analyze in detail the geometry of retinal in squid rhodopsin and compared our findings with published data on bovine rhodopsin. Figure 2 shows the internal and optimized coordinates of the retinal chromophores in squid and bovine rhodopsins. The calculated bond length alternation (Figure 2a) is smaller than the experimentally determined one that was identified in a previous work of bovine rhodopsin.6 Meanwhile, the experimental and calculated bond angles along the conjugated chain of the retinal chromophore (Figure 2b) show rather good agreement except for the protonated Schiff base region. A comparison of the dihedral angles in the experimental structures (broken lines in Figure 2c) reveals major differences in the dihedral angles of the two cis-bonds, C6 C7 and C11dC12 bonds. However, after the MD calculations, the dihedral angles of these two cis-bonds converge to essentially the same angles (C6 C7, 44° in squid rhodopsin vs 41° in bovine rhodopsin, and C11dC12, 18° in squid rhodopsin vs 17° in bovine rhodopsin). Finally, the chirality of the retinal polyene chain is the same in the MD geometries of squid and bovine rhodopsins. One quantitative

difference is the dihedral angle of one of the isomerizing bonds, C12 C13: 157° (in squid rhodopsin) vs 169° (in bovine rhodopsin). It is suggested that the additional twist on this bond in squid rhodopsin accelerates the isomerization.41 The retinal polyene chain is a highly correlated π-electron system and the description of the ground and excited properties depends on the applied method.43,44 The SCC-DFTB method can reproduce the results from the B3LYP level of calculations.43 As a consequence SCC-DFTB tends to overestimate the π-electron delocalization thereby leading a more planar retinal geometry. The MM crystallographic refinement program45 does not have a wellestablished set of parameters for a π-electron system and it is wellknown that different configurations of retinal are found in available crystal structures. Previous work6,30,31 has shown that the retinal configurations are significantly improved by QM calculations. Although X-ray crystallography suggests that the amino acids in van der Waals contact with retinal in squid rhodopsin are altered from those seen in bovine rhodopsin,6 statistical analysis and MD calculations show that conserved aromatic amino acid residues close to retinal play an important role in maintaining the twisted retinal configuration in both squid and bovine rhodopsins. Figure 3 is an overlay of the minimized retinal binding pockets in squid rhodopsin (in gray) and bovine rhodopsin (in red). Seven positions near the β-ionone ring are occupied by highly conserved aromatic residues in opsins (See Table 1). The coordinates of the aromatic residues show only small changes in positions in squid and bovine rhodopsins, except for Tyr190 (Tyr191EL2) in EL2. The distance between any atoms in two corresponding residues is less than 3 Å, whereas the largest movement (about 6 Å) is found in the hydroxyl group of Tyr190 (in squid) and Tyr191EL2 (in bovine). The movement of Tyr190 in squid rhodopsin is induced by the interaction with Phe188, which occupies the Tyr191EL2 position in bovine rhodopsin (See Figure.3). The distance between the Cζ atoms in Phe188 and Tyr191EL2 is ∼1 Å. In contrast, an isoleucine residue (Ile189EL2), which is not in contact with Tyr191EL2, is found at this position in bovine rhodopsin. The conservation rate of Phe at this position is small (9% in 1272 GPCRs and 0% in 81 vertebrate opsins, but 100% in five invertebrate opsins). Previous theoretical calculation has suggested that Tyr191EL2 in bovine rhodopsin hinders the rotational motion of the retinal C9 methyl group in the cis trans isomerization, stabilizing the peculiar geometry of 6175

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B retinal in the bathorhodopsin intermediate.30 Our analysis suggests that Phe188 interacts with the retinal C9 methyl group in the bathorhodopsin intermediate of squid rhodopsin. Of the seven aromatic residues in the retinal binding pocket, three are highly conserved in opsin groups, but not in GPCRs in general (see Table 1). These amino acid residues are found on the extracellular side of retinal and stabilize the inactive conformation of the covalently attached ligand (retinal). Unlike rhodopsin to which retinal is covalently bound, in most of GPCRs, various ligands enter from the extracellular side; the amino acid residues in EL2 may differ among the various GPCRs according to the ligands. Indeed, the extracellular surface of GPCRs is remarkably diverse and the recent crystal structure of CXCR4 chemokine GPCR indicates that EL2 is displaced from the ligand binding pocket. 16 Common residues appearing in TM6 undergo major conformational changes during receptor activation.46,47 In particular, Trp274 (Trp2656.48) in the vicinity of retinal participates in the hydrogen bond network mediated by water molecules and is known as the “trigger” that initiates the activation process after the cis trans isomerization. It also has an important function to transfer the signal via water molecules in the interhelical cavity to the cytoplasmic site.48,49 Water-Mediated Hydrogen Bond Network in the Interhelical Cavity. Water molecules in the interhelical cavity are present in the crystal structures of GPCRs3 8,10 14 and likely to be important for the proper function of activation.17 21 X-ray crystallography of squid rhodopsin revealed nine water molecules in the interhelical cavity:8 of these, three occupy similar locations in the crystal structure of bovine rhodopsin.5 7 Starting from the indole HN group of Trp274 (Trp2656.48), a “ladder” of the hydrogen bond network mediated by the water molecules is observed in the interhelical cavity of both squid (Figure 4, parts a, c, and e) and bovine rhodopsins (Figure 4, parts b and d). In the case of squid rhodopsin, the indole NH group of Trp274 forms a direct hydrogen bond with water that contacts with Ser307, Asp80, and Asn311, whereas in the case of bovine rhodopsin, a hydrogen bond is formed between the NH group of Trp2656.48 and the hydroxyl group of Ser2987.47. The different orientation of the hydrogen bond network first occurs at the serine residue (Ser307 or Ser2987.43). Asp80 (Asp832.50) is the sole acidic amino acid residue (conservation rate 97%) that is centered in the ladder of the hydrogen bond network (Figure 4). FTIR measurements of bovine rhodopsin have shown that Asp832.50 is protonated (neutral),22 whereas the protonation state of Asp80 in squid rhodopsin is unknown. To uncover the likely protonation state of Asp80 in the dark state of squid rhodopsin, we performed QM/MM MD simulations with Asp80 in the negatively charged state, and independent QM/MM MD simulations with Asp80 in the protonated state; for the protonated Asp80 state, we considered the proton on either the Oδ1 or Oδ2 atom (see Figure 4a) and the protonation at the symmetric or antisymmetric position. When the proton is on Asp80-Oδ2 (symmetric or antisymmetric position), Asp80 rotates around its Cβ-Cγ bond after the MD simulation and the protonated Oδ2 oxygen is located in the space initially occupied by Oδ1 oxygen; the new geometry of the Asp80 side chain remains stable during the remaining of the trajectory (see Figure 4, parts a and c). In contrast, when the proton is on Asp80-Oδ1 (symmetric or antisymmetric position), Asp80 does not rotate during MD runs: the geometry of the side

ARTICLE

Figure 4. Hydrogen bond network near the asparatate residue (Asp80 in squid rhodopsin or Asp832.50 in bovine rhodopsin) in helix 2. The conservation rates of residues are included. Asp is protonated at the antisymmetric (squid rhodopsin (a) and bovine rhodopsin (b)) or symmetric position (squid rhodopsin (c) and bovine rhodopsin (d)). (e) Deprotonated Asp80 in squid rhodopsin.

chain remains close to that in the starting crystal structure. The deviations of the Cγ atom of Asp80 from the initial crystal structure are 0.6 ( 0.2 and 1.2 ( 0.2 Å for the protonated models at the antisymmetric and symmetric positions, respectively. For the deprotonated model, the deviation is 0.9 ( 0.2 Å. When the MD simulation is performed with Asp80 in the negatively charged state (Figure 4e), a water molecule bridges Asp80-Oδ1 and the carboxyl oxygen atom of Asn311. The hydrogen bond network around Asp80 is also stable during the MD simulation. Regardless of the protonation state, three water molecules interacting with Asp80, Asn52, and Tyr315 maintain the stable hydrogen bond network, as shown in Figure 4. From the computations discussed above, we conclude that both the 6176

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B protonated Asp80 at Oδ1 and the deprotonated Asp80 are theoretically possible in squid rhodopsin. Parts b and c of Figure 4 show two possible protonation states of Asp832.50 (Oδ1 at symmetric or antisymmetric position). In the case of the antisymmetric protonated state, the highly conserved residue, Asn551.50, is involved in the interhelical hydrogen bond network. In squid rhodopsin, three water molecules are found on the cytoplasmic side of Asn52 and Asn311, and the hydrogen bond network extends to the highly conserved tyrosine residue (in the NP7.50xxY7.53 motif). The tyrosine residue has a different orientation in squid and bovine rhodopsins. The hydroxyl group in Tyr315 in squid rhodopsin forms a hydrogen bond with an internal water molecule in the interhelical cavity and behaves like a plug of the water cluster. In contrast, Tyr3067.53 in bovine rhodopsin points to the cytoplasmic side and forms hydrogen bonds with Asn732.40 and Thr621.57 via a water molecule that does not form a hydrogen bond with other internal water molecules in the interhelical cavity. Hence, the successively extended hydrogen bond network formed from Trp2656.48 is interrupted at the tyrosine residue. The number of water molecules may be related to the configuration of the highly conserved tyrosine residue. The geometrical differences of the hydrogen bond network determine the sequent signal propagation pathway. In squid rhodopsin, the movement of Trp274 leads to coupling with water. This result well coincides with the report of Ota et al., suggesting that internal water molecules alter their hydrogen bonds upon photoisomerization in squid rhodopsin but not in bovine rhodopsin.50 In bovine rhodopsin, the movement of Trp2656.48 as a pivot point causes the rearrangement of helixes to activate G-protein.48,49 The crystal structures of ligand-mediated GPCRs12 15 have revealed that the ligand molecules appear within 4 Å distance from Trp at 6.48. Cytoplasmic Loop Regions. It has been suggested that CL3 plays distinct roles in the interaction with G-protein.48 The crystal structure of bovine opsin (PDB: 3DQB) has revealed that TM5 and TM6 protrude into the cytoplasmic medium interacting with the R-subunit of G-protein,11 although it is plausible that these helices form a loop conformation (CL3) in the dark state. In contrast, TM5 and TM6 of squid rhodopsin in the dark state already extend into the cytoplasmic medium.8,9 Compared to bovine rhodopsin, squid rhodopsin has an additional helix, H9 (from Asp341 to Glu351), which contains a significant number of charged residues. To clarify the influence of H9 on the stability of TM5 and TM6, we performed a set of MD runs with and without H9. In the protein model without H9, the amino acids from Phe339 to Glu358 in the sequence are omitted. The root-mean-square deviation values of the CR positions are plotted in Figure 5(b). The deviations are measured from the average CR positions and are observed mainly in the loop regions connected by two TMs, in the C- and N-terminal regions, and in the region between H8 and H9. Remarkable differences of the MD runs with (in green) and without (in magenta) H9 are seen in the intercellular sides of TM5 and TM6 (refer to captions on the abscissa in Figure 5. The protruding regions of TM5 and TM6 are shown in gray). The protein model of squid rhodopsin with H9 is stable against the MD runs at 300 K, while the fluctuation in loop regions is enhanced without H9. Interestingly, peculiar fluctuation is noted near the end of H8 (Trp332), which is larger in the protein model with H9 than without H9. This fluctuation is caused by the mechanical contact of Trp332 with H9.

ARTICLE

Figure 5. (a) Interhelical salt bridges on cytoplasmic side. (b) Root mean square deviations of CR positions from average positions. Data from squid rhodopsin with and without H9 appear in green and magenta, respectively. Residues forming interhelical salt bridges are connected with broken lines. Helical regions are shown as boxes on the x-axis and the enlarged regions in TM 5 and TM 6 are in gray.

Figure 5a shows the salt bridges at the cytoplasmic side. Residues forming salt bridges in the MD runs are connected with broken lines in Figure 5b. Some of the salt bridges are not present in the X-ray structure but can be formed during the simulation. Three of seven acidic residues in H9 interact with TM6 through the salt bridges (Glu343-Arg258, Glu345-Arg247, and Asp349-Arg247). In addition, two more salt bridges between the residues in TM6 and those in the C-terminal region are formed: Glu358-Lys244 and Glu353Lys248. The importance of salt bridges in protein stability has been extensively investigated. In particular, the energetic contribution of complex salt bridges, in which one charged residue forms salt bridges with two or more residues simultaneously, is more than the sum of the energies of individual pairs.51 During the MD runs one stable complex salt bridge appears between TM6 and TM9: Arg247-Glu345 and Asp349. The interaction between the charged residues in TM6 and TM9 is essential to stabilize the elongated conformation of TM5 and TM6. Besides, H9 interacts with CL1 and CL2: Glu351-Lys63 (CL1), Glu351-Lys146 (CL2), and Glu353Lys145 (CL2), and these salt bridges act to stabilize the conformation of the cytoplasmic surface. Salt bridges on the cytoplasmic surface, together with the interhelical hydrogen bond network, impose a restraint on the protein conformation and therefore the large conformational changes found in the activation state of bovine rhodopsin may not occur in squid rhodopsin. As suggested based on 6177

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B the crystal structure, squid rhodopsin in the dark state already has the ability to form a complex with G-protein.8

’ CONCLUSIONS On the basis of the squid and bovine rhodopsin crystal structures, we have investigated the structural features of the hydrogen bond network by using MD simulation and sequence analysis. The hydrogen bond network is extended from the retinal binding pocket to the cytoplasmic site, which is likely responsible for the signal propagation pathway. Compared to the crystal structure of bovine rhodopsin, that of squid rhodopsin indicates a highly twisted β-ionone ring, and a less distorted retinal polyene chain.8 However, MD simulation demonstrated that retinal converged to practically the same conformation as that found in the previous theoretical work of bovine rhodopsin.6,30 The retinal binding pocket, which contains highly conserved aromatic residues, stabilizes the inactive retinal conformation for the interaction with a photon in the two rhodopsins. The signal induced by a photon propagates from the retinal binding pocket; the highly conserved tryptophan residue (Trp274 in squid rhodopsin and Trp2656.48 in bovine rhodopsin) near the β-ionone ring is the first residue to interact with the isomerized retinal and through the tryptophan residue the signal is transferred to the protein environment.48,49 In squid rhodopsin, Trp274 interacts with water that plays a key role in the signal propagation, whereas bovine rhodopsin, Trp2656.48 directly interacts with Ser2987.47. This causes the first perturbation of the signal propagation. In squid rhodopsin, the hydrogen bond network extends approximately 13 Å to Tyr315 at the cytoplasmic side, and is stable regardless of the protonation state of Asp80. In bovine rhodopsin, the hydrogen bond network is interrupted at the coincident tyrosine residue, Tyr3067.54 (approximately 8 Å from Trp2656.51 to Asn3027.49), and the interhelical hydrogen bond network is rather loose compared with that of squid rhodopsin. Moreover, in squid rhodopsin, the salt bridges on the cytoplasmic surface stabilize the protein conformation, and consequently, rather small changes are likely to occur on the cytoplasmic side in the later intermediates. ’ ACKNOWLEDGMENT This work was supported by a Grant-in-Aid for Scientific Research (B) (Kakenhi 20300104) from the Japanese Ministry of Education, Culture, Sports, Science, and Technology. The authors thank Dr. Tsutomu Kouyama, Dr. Midori Murakami, and Mr. Yukiteru Ono for insightful discussions, and Dr. AnaNicoleta Bondar and Dr. Douglas J. Tobias for providing the calculated coordinates. Computation was performed by the Supercomputer System, Human Genome Center, University of Tokyo. ’ REFERENCES (1) Vauquelin, G.; Mentzer, B. G. Protein-coupled Receptors; John Wiley & Sons, Ltd.: West Sussex, England, 2007. (2) Rosenbaum, D. M.; Rasmaussen, S. G. F.; Kobilka, B. K. Nature Insight 2009, 459, 356–363. (3) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima, H.; Fox, B. A.; LeTrong, I.; Teller, D. C.; Okada, T.; Stenkamp, R. E.; Yamamoto, M.; Miyano, M. Science 2000, 289, 739–745. (4) Teller, D. C.; Okada, T.; Behnke, C. A.; Palczewski, K.; Stenkamp, R. E. Biochemistry 2002, 40, 7761–7777. (5) Okada, T.; Fujiyoshi, Y.; Silow, M.; Navarro, J.; Landau, E. M.; Shichida, Y. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5982–5987.

ARTICLE

(6) Okada, T.; Sugihara, M.; Bondar, A.-N.; Elstner, M.; Entel, P.; Buss, V. J. Mol. Biol. 2004, 342, 571–581. (7) Li, J.; Edwards, P. C.; Burghammer, M.; Villa, C.; Sehertler, G. F. X. J. Mol. Biol. 2004, 343, 1409–1438. (8) Murakami, M.; Kouyama, T. Nature 2008, 453, 363–368. (9) Shimamura, T.; Hiraki, K.; Takahashi, N.; Hori, T.; Ago, H.; Masuda, K.; Takao, K.; Ishiguro, M.; Miyano, M. J. Biol. Chem. 2008, 283, 17753–17756. (10) Park, J. H.; Scheerer, P.; Hofmann, K. P.; Choe, H.-W.; Ernst, O. P. Nature 2008, 454, 185–187. (11) Scheerer, P.; Park, J. H.; Hildebrand, P. W.; Kim, Y. J.; Kraub, N.; Choe, H.-W.; Hofmann, K. P.; Ernst, O. P. Nature 2008, 455, 497– 502. (12) Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmusen, S. G. F.; Thian, F. S.; Kobilka, T. S.; Choi, H.-J.; Kuhn, P.; Weis, W. K.; Kobilka, B. K.; Stevens, R. C. Science 2007, 318, 1258–1265. (13) Warne, T.; Serrano-Vega, J.; Baker, J. G.; Moukhametzianov, R.; Edwards, P. C.; Henderson, R.; Leslie, A. G. W.; Tate, C. G.; Schertler, G. F. X. Nature 2008, 454, 486–491. (14) Jaakola, V.-P.; Griffith, M. T.; Hason, M. A.; Cherezov, V.; Chien, E. V. T.; Lane, J. R.; IJzerman, A. P.; Stevens, R. C. Science 2008, 322, 1211–1217. (15) Chien, E. Y. T.; Liu, W.; Zhao, Q.; Katritch, V.; Han, G. W.; Hanson, M. A.; Shi, L.; Newman, A. H.; Javitch, J. A.; Cherezov, V.; Stevens, R. C. Science 2010, 330, 1091–1094. (16) Wu, B.; Chien, Y. T.; Mol, C. D.; Fenalti, G.; Liu, W.; Katritch, V.; Abagyan, R.; Brooun, A.; Wells, P.; Bi, F. C.; Hamel, D. J.; Kuhn, P.; Handel, T. M.; Cherezov, V.; Stevens, R. C. Science 2010, 330, 1066– 1071. (17) Weis, W. I.; Kobilka, B. K. Curr. Opin. Struct. Biol. 2008, 18, 734–740. (18) Angel, T. E.; Chance, M. R.; Palczewski, M. R. Proc. Natl. Acad. Sci. U.S.A. 2009, 21, 8555–8560. (19) Angel, T. E.; Gupta, S.; Jastrzebska, B.; Palczewski, K.; Chance, M. R. Proc. Natl. Acad. Sci. U.S.A. 2009, 34, 14367–14372. (20) Urizar, E.; Claeysen, S.; Dupí, X.; Govaerts, C.; Costagliola, S.; Vassart, G.; Pardo, L. J. Biol. Chem. 2005, 280, 17135–17141. (21) Pardo, L.; Deupi, X.; D€olker, N.; L
opez-Rodríguez, M. L.; Campillo, M. ChemBioChem 2007, 8, 19–24. (22) Nagata, T.; Terakita, A.; Kandori, H.; Schichida, Y.; Maeda, A. Biochemistry 1998, 37, 17216–17222. (23) Kobilka, B.; Schertler, G. F. X. Trends Pharmacol. Sci. 2008, 29, 79–83. (24) Jardon-Valadez, E.; Bondar, A.-N.; Tobias, D. J. Biophys. J. 2009, 96, 2572–2576. (25) Jardon-Valadez, E.; Bondar, A.-N.; Tobias, D. J. Biophys. J. 2010, 99, 2200–2207. (26) Sivakuma, S.; Altun, A.; Morokuma, K. Chem.—Eur. J. 2010, 16, 1744–1749. (27) MacKerell, A. D.; Bashford, D.; Bellott, M.; Dunbrack, R. D.; Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.; JosephMcCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, G.; Reiher, W. E.; Roux, B.; Shlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. J. Phys. Chem. B 1998, 102, 3586–3616. (28) Field, M. J.; Bash, P. A.; Karplus, M. J. Comput. Chem. 1990, 11, 700–733. (29) Cui, Q.; Elstner, M.; Kaxiras, E.; Frauenheim, T.; Karplus, M. J. Phys. Chem. B 2001, 105, 569–585. (30) Sugihara, M.; Buss, V. Biochemistry 2008, 47, 13733–13735. (31) Bondar, A.-N.; Smith, J. C.; Elstner, M. Theor. Chem. Acc. 2010, 125, 353–365. (32) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. J. Chem. Phys. 1983, 79, 926–735. (33) Nos
e, S. J. Chem. Phys. 1984, 81, 511–519. (34) Dror, R. D.; Arlow, D. H.; Borhani, D. W.; Jensen, M. Ø.; Piana, S.; Shaw, D. E. Proc. Nalt. Acad. Sci. U.S.A. 2009, 106, 4689–4694. 6178

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179

The Journal of Physical Chemistry B

ARTICLE

(35) Bhattacharya, S.; Hall, S. E.; Vaidehi, N. J. Mol. Biol. 2008, 382, 539–555. (36) Suwa., M.; Ono, Y.; Koga., H. Methods Mol. Biol. 2009, 577, 41–54. (37) Suwa, M.; Ono, Y. Synthesiology 2010, 3, 1–12. (38) Apweiler, R.; Bairoch, A.; Wu, C. H.; Barker, W. C.; Boeckmann, B.; Ferro, S.; Gasteiger, E.; Huang, H.; Lopez, R.; Magarane, M.; Martin, M. J.; Anatales, D. A.; O’Donovan, C.; Bedaschi, N.; Yeh, L.-S. L. Nucleic Acids Res. 2010, 32, 115–119. (39) Katoh, K.; Toh, H. Briefings Bioinf. 2008, 9, 286–298. (40) Ballesteros, J. A.; Weinstein, H. Method Neurosci. 1995, 25, 366–428. (41) Buss, V.; Oliver, W.; Sugihara, M. Angew. Chem., Int. Ed. 2002, 39, 2784–2786. (42) Schreiber, M.; Buss, V.; Sugihara, M. J. Chem. Phys. 2003, 119, 12045–12048. (43) Zhou, H.; Tajkhorshid, E.; Frauenheim, T.; Suhai, S.; Elstner, M. Chem. Phys. 2002, 277, 91–103. (44) Terstegen, F.; Buss, V. J. Mol. Struct. (THEOCHEM) 1998, 430, 209–218. (45) Br€unger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros, P.; Grosse-Kunstleve, R. W.; Jiang, J.-S.; Kuszewski, J.; Nilges, M.; Pannu, N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L. Acta. Crystallog. Sect. D 1998, 54, 905–921. (46) Filipek, S.; Teller, D. C.; Palczewski, K.; Stenkamp, R. Annu. Rev. Biophys. Biomol. Struct. 2003, 32, 375–97. (47) Park, P. S.-H.; Lodowski, D. T.; Palczewski, K. Annu. Rev. Pharmacol. Toxicol. 2008, 48, 107–141. (48) Crocker, E.; Eilers, M.; Ahuja, S.; Hornak, V.; Hirshfeld, A.; Sheves, M.; Smith, S. O. J. Mol. Biol. 2006, 357, 163–172. (49) Patel, A. B.; Crocker, E.; Reeves, P. J.; Getmanova, E. V.; Eilers, M.; Khorana, H. G.; Smith, S. O. J. Mol. Biol. 2005, 347, 803–812. (50) Ota, T.; Furutani, Y.; Terakita, A.; Shichida, Y.; Kandori, H. Biochemistry 2006, 45, 2841–2851. (51) Gvritishvili, A. G.; Gribenko, A. V.; Makhatadze, G. I. Protein Sci. 2008, 17, 1285–1290.

6179

dx.doi.org/10.1021/jp1101785 |J. Phys. Chem. B 2011, 115, 6172–6179