Conformation of a double-membrane-spanning fragment of a G protein-coupled receptor: effects of hydrophobic environment and pH.

Overcoming the problems associated with the expression, purification and in vitro handling of membrane proteins requires an understanding of the factors governing the folding and stability of such proteins in detergent solutions. As a sequel to our earlier report (Biochim. Biophys. Acta 1747(2005), 133-140), we describe an improved purification procedure and a detailed structural analysis of a fragment of the mu-opioid receptor ('TM2-3') that comprises the second and third transmembrane segments and the extracellular loop that connects them. Circular dichroism (CD) spectroscopy of TM2-3 in 2,2,2-trifluoroethanol gave a helical content similar to that predicted by published homology models, while spectra acquired in several detergents showed significantly lower helical contents. This indicates that this part of the mu-opioid receptor has an intrinsic propensity to be highly helical in membrane-like environments, but that in detergent solutions, this helical structure is not fully formed. Proteolysis of TM2-3 with trypsin showed that the helical portions of TM2 and TM3 are both shorter than their predicted lengths, indicating that helix-helix interactions in the full-length receptor are apparently important for stabilizing their conformation. Lengthening the alkyl chain of the detergent led to a small but significant increase in the helicity of TM2-3, suggesting that hydrophobic mismatch could play an important role in the stabilization of transmembrane helices by detergents. Protonation of aspartic acid residues in detergent-solubilized TM2-3 also caused a significant increase in helicity. Our results thus suggest that detergent alkyl chain-length and pH may influence membrane protein stability by modulating the stability of individual transmembrane segments.

A method for expression and purification of soluble, active Hsp47, a collagen-specific molecular chaperone.

Hsp47 is regarded as a collagen-specific chaperone with several suggested roles in collagen biosynthesis under normal and disease conditions. We describe here a procedure for the expression and purification of Hsp47 in Escherichia coli using the IMPACT expression system (New England Biolabs) where the guest gene is fused to the adduct, intein, with a chitin-binding domain. Use of this system resulted in relatively high levels of soluble Hsp47 compared to other available protocols, especially when the bacterial cells were induced at 14 degrees C instead of 37 degrees C. The cell lysate was passed through a chitin-Sepharose affinity column and Hsp47 was cleaved from intein using beta-mercaptoethanol. Minor degradation products were subsequently removed using a hydroxylapatite column to yield milligram amounts of pure and active protein suitable for structural studies. Gel electrophoretic analysis of the purified protein indicated the presence of a small proportion of trimeric species when non-reducing conditions were used. The ability to form a trimer may be important for its role as a chaperone. The IMPACT system allows for radiolabelling of purified Hsp47 with (35)S for use in binding experiments. Illustrative data on collagen binding by (35)S-Hsp47 are shown.

Homology model of dihydropyridine receptor: implications for L-type Ca(2+) channel modulation by agonists and antagonists.

L-type calcium channels (LCCs) are transmembrane (TM) proteins that respond to membrane depolarization by selectively permeating Ca(2+) ions. Dihydropyridine (DHP) agonists and antagonist modulate Ca(2+) permeation by stabilizing, respectively, the open and closed states of the channel. The mechanism of action of these drugs remains unclear. Using, as a template, the crystal structure of the KcsA K(+) channel (Doyle et al. (1998) Science 280, 69-77), we have built several homology models of LCC with alternative alignments of TM segments between the proteins. In each model, nifedipine was docked in the pore region and in the interface between repeats III and IV. Several starting structures were generated by constraining the ligand to residues whose mutations reportedly affect DHP binding (DHP-sensing residues). These structures were Monte Carlo-minimized with and without constraints. In the complex with the maximum number of contacts between the ligand and DHP-sensing residues and the lowest ligand-receptor energy, the drug fits snugly in the "water-lake" cavity between segments S6s, which were aligned with M2 segment of KcsA as proposed for Na(+) channel (Lipkind and Fozzard (2000) Biochemistry 39, 8161-8170). In the flattened-boat conformation of DHP ring, the NH group at the stern approaches the DHP-sensing tyrosines in segments IIIS6 and IVS6. Stacking interactions of IVS6 Tyr with the bowsprit aromatic ring stabilize the ligand's orientation in which the starboard COOMe group coordinates Ca(2+) ion chelated by two conserved glutamates in the selectivity filter. In the inverted teepee structure of LCC, the portside COOMe group approaches a bracelet of conserved hydrophobic residues at the helical-bundle crossing, which may function as the activation gate. The dimensions of the gate may readily change upon small rotation of the pore-forming TM segments. The end of the portside group is hydrophobic in nifedipine, (R)-Bay K 8644, and other antagonists. Favorable interactions of this group with the hydrophobic bracelet would stabilize its closed conformation. In contrast, (S)-Bay K 8644 and several other agonists have hydrophilic groups at the portside. Unfavorable interactions of the hydrophilic group with the hydrophobic bracelet would destabilize its closed conformation thereby stabilizing the open conformation. In the agonist-bound channel, Ca(2+) ions would permeate between the hydrophilic face of the ligand and conserved hydrophilic residues in segments IS6 and IIS6. Our model suggests mutational experiments that could further our understanding of the pharmacological modulation of voltage-gated ion channels.

Mutation of any site of N-linked glycosylation accelerates the in vivo clearance of recombinant rabbit antithrombin.

Antithrombin (AT) is a plasma protein with four sites of N-linked glycosylation. Asn 135 is incompletely glycosylated, and the resulting 3-glycan AT is cleared more rapidly in vivo than the 4-glycan form. The Asn codons in each of the four sites of glycosylation were altered in turn, to create four mutant rabbit AT cDNAs. Permanently transfected CHO cell lines were generated following transfection of the resulting constructs, encoding either the wild-type rabbit AT (AT-WT) or one of the four underglycosylated variants (AT-N96Q, AT-N135Q, AT-N155Q, and AT-N155Q). Comparison of the five resulting recombinant AT proteins revealed that the major AT species of each variant co-migrated on SDS gels, and migrated more rapidly than the major form of AT-WT. The shift in mobility, from 60 to 57 kDa, was consistent with the loss of one fully sialylated complex N-linked glycan. Neither the amount of AT secreted (range: 1.25 to 4.2 microg/10(6) cells/day) nor the kinetics of secretion differed significantly between cell lines expressing AT-WT or any of the AT variants. All forms of recombinant rabbit AT were capable of forming denaturation-resistant complexes with thrombin. Purification and radioiodination of each of the five recombinant AT proteins permitted pharmacokinetic analysis of their individual clearance in rabbits. While neither the equilibration half-life (t(0.5)alpha) nor the terminal catabolic half-life (t(0. 5)beta) differed significantly between plasma-derived rabbit AT and AT-WT, the t(0.5)beta of all the underglycosylated variants was decreased relative to that of AT-WT (maximum reduction in mean: from 70.1+/-3.2 h to 52.4+/-2.5 h). These results suggest that the overall extent of glycosylation, rather than the location within AT of the glycan chains, is a primary determinant of AT clearance.

Structure-function studies on hsp47: pH-dependent inhibition of collagen fibril formation in vitro.

Hsp47, a 47 kDa heat shock protein whose expression level parallels that of collagen, has been regarded as a collagen-specific molecular chaperone. Studies from other laboratories have established the association of Hsp47 with the nascent as well as the triple-helical procollagen molecule in the endoplasmic reticulum and its dissociation from procollagen in the Golgi. One of several roles suggested for Hsp47 in collagen biosynthesis is the prevention of aggregation of procollagen in the endoplasmic reticulum. However, no experimental evidence has been available to verify this suggestion. In the present study we have followed the aggregation of mature triple-helical collagen molecules into fibrils by using turbidimetric measurements in the absence and presence of Hsp47. In the pH range 6-7, fibril formation of type I collagen, as monitored by turbidimetry, proceeds with a lag of approx. 10 min and levels off by approx. 60 min. The addition of Hsp47 at pH 7 effectively inhibits fibril formation at and above a 1:1 molar ratio of Hsp47 to triple-helical collagen. This inhibition is markedly pH-dependent, being significantly diminished at pH 6. CD and fluorescence spectral data of Hsp47 in the pH range 4.2-7.4 reveal a significant alteration in its structure at pH values below 6.2, with a decrease in alpha-helix and an increase in beta-structure. This conformational change is likely to be the basis of the decreased binding of Hsp47 to collagen in vitro at pH 6.3 as well as its inability to inhibit collagen fibril formation at this pH. Our results also provide a functional assay for Hsp47 that can be used in studies on collagen and Hsp47 interactions.

CD, 1H NMR and molecular modeling studies of the interaction of Ca2+ with substance P and Ala7-substance P in a non-polar solvent.

The biologically relevant conformation of substance P is likely to be dictated by the lipid milieu wherein the hormone would interact with its receptor. Assuming that specific constraints to the hormone structure may be imparted by its interaction with Ca2+ ions in the low dielectric lipid medium, the interaction of substance P and its inactive analog, Ala7-substance P, has been characterized in a lipid-mimetic solvent. Circular dichroism (CD) and NMR spectral methods were employed to study the conformation of the free and Ca2+-bound forms of the peptides and the conformational changes that occur on Ca2+ binding. The results show that both peptides assume a helical structure in the non-polar solvent used, a mixture of acetonitrile and trifluoroethanol. The N-terminal region is, however, less ordered in the analog peptide compared with the native hormone. Ca2+ addition causes significant conformational changes in both the peptides. However, while substance P binds two Ca2+ ions in a cooperative manner, Ala7-substance P binds only one Ca2+ ion with a relatively weaker affinity. Computations of the minimum-energy conformations of the free and Ca2+-bound peptides were performed using interproton distances derived from nuclear Overhauser enhancement spectra of the two peptides, as well as the information provided by changes in proton chemical shifts caused by Ca2+ addition. Taken together, the results of this study suggest that differences in the interaction of substance P and Ala7-substance P with Ca2+ in the non-polar milieu, which in turn leads to differences in their Ca2+-bound conformations, may be the basis for the differences in their biological potencies.

Homology models of mu-opioid receptor with organic and inorganic cations at conserved aspartates in the second and third transmembrane domains.

Metal ions affect ligand binding to G-protein-coupled receptors by as yet unknown mechanisms. In particular, Na(+) increases the affinity for antagonists but decreases it for agonists. We had modeled the mu-opioid receptor (muR) based on the low-resolution structure of rhodopsin by G. F. X. Schertler, C. Villa, and R. Henderson (1993, Nature 362, 770-772) and proposed that metal ions may be directly involved in the binding of ligands and receptor activation (B. S. Zhorov and V. S. Ananthanarayanan, 1998, J. Biomol. Struct. Dyn. 15, 631-637). Developing this concept further, we present here homology models of muR using as templates the structure of rhodopsin elaborated by I. D. Pogozheva, A. L. Lomize, and H. I. Mosberg (1997, Biophys. J. 70, 1963-1985) and J. M. Baldwin, G. F. X. Schertler, and V. M. Unger (1997, J. Mol. Biol., 272, 144-164). Using the Monte Carlo minimization (MCM) method, we docked the Na(+)-bound forms of muR ligands: naloxone, bremazocine, and carfentanyl. The resultant low-energy complexes showed that the two positive charges in the protonated metal-bound ligands interact with the two negative charges at Asp(3.32) and Asp(2.50) (for notations, see J. A. Ballesteros and H. Weinstein, 1995, Methods Neurosci. 25, 366-426). MCM computation on morphine docked inside the model of muR by I. D. Pogozheva, A. L. Lomize, and H. I. Mosberg (1998, Biophys. J. 75, 612-634) yielded two binding modes with the ligand's ammonium group salt-bridged either to Asp(3.32) (generally regarded as the ligand recognition site) or to Asp(2.50). The latter is the presumed site for Na(+) ion, which is known to modulate ligand binding. Assuming that in the low-dielectric transmembrane region of muR, organic and inorganic cations would compete for Asp(3.32) and Asp(2.50), we propose that ligand binding, as visualized in the above models, would first displace Na(+) from Asp(3.32). A subsequent progress of the ligand toward Asp(2.50) would result in either the retention of Na(+) at Asp(2.50) in the case of antagonists or the displacement of Na(+) from Asp(2.50) in the case of agonists. The displaced Na(+) would move toward the salt-bridged Asp(3.49)-Arg(3.50) and disengage the salt bridge. This, in turn, would result in conformational changes at the cytoplasmic face of the receptor that facilitate the interaction with the G-protein.

Ca(2+) and Zn(2+) binding properties of peptide substrates of vertebrate collagenase, MMP-1.

To understand the role of Ca(2+) in vertebrate in the structure and action of collagenase, we have examined peptides that interact with recombinant human fibroblast collagenase for their affinities towards Ca(2+) and Zn(2+) in a non-polar solvent. Two of the peptides, GPQGIAGQ and GNVGLAGA, had sequences in collagen which are, respectively, cleaved and not cleaved by collagenase. A third peptide, PSYFLNAG, had a collagenase-cleaved sequence in ovostatin, a globular protein substrate. Peptides TVGCEECTV and CLPREPGL were derived from TIMP-1; the former competitively inhibits collagenase while the latter does not. The relative rates of hydrolysis of the peptides by collagenase had the order GPQGIAGQ>PSYFLNAG>GNVGLAGA. Circular dichroism spectral data in trifluoroethanol showed that while the TIMP control peptide, CLPREPGL, bound only Zn(2+), the other four peptides bound both Ca(2+) and Zn(2+) with definite stoichiometries. Ca(2+) could displace Zn(2+) in the substrate peptides while Zn(2+) displaced Ca(2+) in the TIMP peptide. GPQGIAGQ, PSYFLNAG and TVGCEECTV formed peptide:Ca(2+):Zn(2+) ternary complexes. Our results suggest that both collagen and globular protein substrates of collagenase may bind Ca(2+) and Zn(2+) in the enzyme's active site. This, in turn, may account for the known importance of the non-catalytic Ca(2+) and Zn(2+) in collagenase activity.

Interaction of gonadotropin-releasing hormone and its agonist analogs with Ca2+ in a nonpolar milieu. Correlation with biopotencies.

Extracellular Ca2+ is necessary for the action of gonadotropin-releasing hormone (GnRH). Assuming that this partly because of the interaction of the hormone with the relatively abundant extracellular Ca2+ in the low dielectric milieu of the bilayer plasma membrane, we studied the interaction of GnRH and five of its agonist analogs with Ca2+ under membrane-mimetic conditions. The peptides used, in increasing order of their reported gonadotropin-releasing activities, were: des-amide GnRH (or GnRH-OH); [Ala6]GnRH; [D-Ala6]GnRH; des-Gly10[D-Ala6,Pro9-NHEt]GnRH and, des-Gly10[D-Trp6,Pro9-NHEt]GnRH. Changes in the far-UV CD and fluorescence spectra of these peptides in trifluoroethanol were used to monitor conformational changes and obtain the Ca2+-binding isotherms. The data show that GnRH and its active analogs contain two Ca2+ binding sites, whereas the inactive analogs have only one. The extent of Ca2+ binding by the agonist peptides paralleled their biological potency ranking. The superactive analog des-Gly10[D-Trp6,Pro9-NHEt]GnRH exhibited the ability to transport Ca2+ ions across large unilamellar vesicles of dimyristoylphosphatidylcholine. Our study shows that significant differences among the GnRH and its analog peptides, suggestive of differences in their conformations, are manifested only in the presence of Ca2+. This observation would provide a basis for understanding GnRH action in terms of the hormone's interaction with Ca2+ in the lipid milieu.

Signal transduction within G-protein coupled receptors via an ion tunnel: a hypothesis.

Based on molecular modeling of the complexes between the mu-opioid receptor and its ligands, we present a hypothesis that accounts for several of the experimental data including the importance of conserved polar residues in rhodopsin-like G-protein-coupled receptors and the effect of Na+ on the binding of ligands to these receptors. We propose that agonists, but not antagonists, would displace Na+ from its initial binding site at the conserved D2.50 residue in the second transmembrane alpha-helical segment, H2. The displaced Na+ would pass through a "gate" of conserved hydrophobic residues and move along a tunnel-like interface (formed of H2, H3 and H7) enriched with several conserved hydrophilic residues including D3.49. Interaction of Na+ with D3.49 would result in the breaking of a salt-bridge between D3.49 and the conserved R3.50 residue thus exposing the latter for interaction with the G-protein.

Docking of verapamil in a synthetic Ca2+ channel: formation of a ternary complex involving Ca2+ ions.

The mechanism by which diverse drugs modulate voltage-dependent Ca2+ channels is ill-understood. We have approached this problem by examining the interaction of verapamil with a 97-residue synthetic channel peptide (SCP) that exhibits functional similarities to authentic L-type Ca2+ channels in terms of cation selectivity and permeation as well as interaction with channel-activating and blocking drugs (Grove et al. (1991) Proc. Natl. Acad. Sci. USA 88, 6418). Different possibilities of binding of verapamil inside the Ca(2+)-bound SCP were simulated using the Monte Carlo-with-energy-minimization method. In the optimal mode of the binding, verapamil adopted a folded conformation and fit snugly in the pore. The dimethoxyphenyl groups of the drug interacted with two Ca2+ ions coordinated to the acidic residues of SCP, thus forming a ternary complex of the drug, Ca2+, and channel. The isopropyl group of verapamil abetted a ring of four Ile residues constituting the putative SCP gate. The occlusion of this gate by verapamil in this manner was strikingly similar to that accomplished by the methyl group of dihydropyridine drugs. In conjunction with an earlier study on SCP bound to dihydropyridine drugs (Zhorov and Ananthanarayanan (1996) Biophys. J. 70, 22), our data suggest that, in general, drug modulation of SCP would involve the interaction of the ligands with the pore-bound Ca2+ and with the hydrophobic gate. In light of the functional similarity between SCP and L-type Ca2+ channel, it is likely that the latter would also interact with drugs in a similar fashion.

Conformational and electrostatic similarity between polyprotonated and Ca(2+)-bound mu-opioid peptides.

In a previous paper (Zhorov and Ananthanarayanan, J. Biomol, Struct. Dynam. 1995, 13:1-13) we had calculated the minimum-energy conformations of monoprotonated and zwitterionic mu-opioid peptides and demonstrated the remarkable similarity between Ca(2+)-bound morphine on the one hand and the Ca(2+)-bound forms of these peptides on the other. We postulated that the Ca(2+)-bound forms of mu-opioids would activate the mu-receptor. To assess further the involvement of multiple positive charges on some of the mu-opioid ligands in their interaction with the receptor, we have, in this work, studied the geometry of five mu-opioid peptides containing two to four protonated groups and having chemical structures essentially different from the endogenous mu-opioid peptide Met-enkephalin (EK). Conformational space was searched using the Monte Carlo-with-energy-minimization method. Ca(2+)-bound forms of the selected peptides were found to be energetically unfavourable implying that one of the protonated groups plays a role similar to that Ca2+ plays in EK-Ca2+ complex. Bioactive conformations of the polyprotonated peptides were then selected using the criteria formulated earlier for Ca(2+)-bound ligands as well as additional criteria requiring ligands to have an elongated conical overall shape complementary to the interface between the transmembrane segments of mu-receptor. Low-energy conformations meeting these criteria were found in all the peptides considered, the protonated groups being separated from each other by about 8 and 16 A. The possible role of the ligands' cationic groups in mu-receptor activation is discussed.

Ca2+-dependent antifreeze proteins. Modulation of conformation and activity by divalent metal ions.

The antifreeze proteins (AFPs) are structurally diverse molecules that share an ability to bind to ice crystals and inhibit their growth. The type II fish AFPs of Atlantic herring and smelt are unique among known AFPs in their requirement of a cofactor for antifreeze activity. These AFPs are homologous with the carbohydrate-recognition domains of Ca2+-dependent (C-type) lectins and require Ca2+ for their activity. To investigate the role of metal ions in the structure and function of type II AFPs, the binding of Ca2+ and other divalent cations to herring AFP was investigated. Binding studies using 45Ca2+ demonstrated that the AFP has a single Ca2+-binding site with a Kd of 9 microM. Proteolysis protection studies and measurement of antifreeze activity revealed a conformational change from a protease-sensitive and inactive apoAFP to a protease-resistant active AFP upon Ca2+ binding. Other divalent metal ions including Mn2+, Ba2+, and Zn2+ bind at the Ca2+-binding site and induce a similar change. A saturatable increase in tryptophan emission intensity at 340 nm also occurred upon Ca2+ addition. Whereas antifreeze activity appeared normal when Ca2+ or Mn2+ were bound, it was much lower in the presence of other metal ions. When Ba2+ was bound to the AFP, ice crystals showed a distinct difference in morphology. These studies demonstrate that herring AFP specifically binds Ca2+ and, consequently, adopts a conformation that is essential for its ice-binding activity.

Ca2+ binding and translocating properties of diflunisal.

We have characterized the binding of Ca2+ and Mg2+ to the anti-inflammatory drug diflunisal in acetonitrile and demonstrated the drug-mediated transport of Ca2+ across the lipid bilayer in unilamellar vesicles made of 1,2-dimyristoyl-sn-glycero-3-phosphocholine. Fluorescence and difference absorption spectral data show that diflunisal undergoes a significant conformational change on binding Ca2+ and Mg2+ forming, respectively, 1:2 and 1:1 cation:drug complexes with Kd in the low microM range. The kinetics of transport showed that Ca2+ was transported into the vesicle as 1:2 Ca2+:diflunisal sandwich complex. This suggests that the interaction of the drug with the cation in the lipid-mimetic solvent are similar. The biological relevance of the Ca(2+)-binding and translocating abilities of diflunisal is examined in light of the reported ionophoretic properties of several phospholipids as well as cyclooxygenase and lipoxygenase products.

Structural model of a synthetic Ca2+ channel with bound Ca2+ ions and dihydropyridine ligand.

Grove et al. have demonstrated L-type Ca2+ channel activity of a synthetic channel peptide (SCP) composed of four helices (sequence: DPWNVFDFLI10VIGSIIDVIL20SE) tethered by their C-termini to a nanopeptide template. We sought to obtain the optimal conformations of SCP and locate the binding sites for Ca2+ and for the dihydropyridine ligand nifedipine. Eight Ca2+ ions were added to neutralize the 16 acidic residues in the helices. Eight patterns of the salt bridges between Ca2+ ions and pairs of the acidic residues were calculated by the Monte Carlo-with-energy-minimization (MCM) protocol. In the energetically optimal conformation, two Ca2+ ions were bound to Asp-1 residues at the intracellular side of SCP, and six Ca2+ ions were arrayed in two files at the diametrically opposite sides of the pore, implying a Ca2+ relay mechanism. Nine modes of nifedipine binding to SCP were simulated by the MCM calculations. In the energetically optimal mode, the ligand fits snugly in the pore. The complex is stabilized by Ca2+ bound between two Asp-17 residues and hydrophilic groups of the ligand. The latter substitute water molecules adjacent to Ca2+ in the ligand-free pore and thus do not obstruct Ca2+ relay. The ligand-binding site is proximal to a hydrophobic bracelet of Ile-10 residues whose rotation is sterically hindered. In some conformations, the bracelet is narrow enough to block the permeation of the hydrated Ca2+ ions. The bracelet may thus act as a "gate" in SCP. Nifedipine and (R)-Bay K 8644, which act as blockers of the SCP, extend a side-chain hydrophobic moiety toward the Ile-10 residues. This would stabilize the pore-closing conformation of the gate. In contrast, the channel activator (S)-Bay K 8644 exposes a hydrophilic moiety toward the Ile-10 residues, thus destabilizing the pore-closing conformation of the gate.

Interaction of oxytocin with Ca2+: I. CD and fluorescence spectral characterization and comparison with vasopressin.

Extracellular Ca2+ is required for the action of oxytocin and both the hormone and its receptor have binding sites for divalent metal cations. To characterize the cation-bound form of oxytocin, we monitored the binding of Ca2+ and Mg2+ to oxytocin as well as peptides representing its ring and tail regions in trifluoroethanol, a lipid-mimetic solvent, using CD and fluorescence spectroscopy. Binding Ca2+ (Kd approximately 50 microM) caused drastic CD and fluorescence changes leading to a helical conformation. Mg2+ caused CD changes smaller than and opposite to Ca2+. However, the helical structure was enhanced when both Ca2+ and Mg2+ were present together. CD changes in the tail peptide of oxytocin showed its ability to bind Ca2+ and Mg2+ whereas the vasopressin tail peptide did not bind either cation. CD spectral changes on Ca2+ and Mg2+ binding to tocinoic acid (the ring moiety of oxytocin) were much smaller than those of oxytocin. These data suggest that the tail segment of oxytocin potentiates Ca2+ binding by the ring. While vasopressin displayed a CD spectrum similar to that of oxytocin, CD spectra of its cation-bound forms were markedly different from those of oxytocin; the Ca(2+)-induced CD changes in vasopressin were very much smaller and of opposite sign, and Mg(2+)-induced ones significantly larger than in oxytocin. Taken together, our observations bring out the structural differences between oxytocin and vasopressin in the context of their interaction with Ca2+ and Mg2+. This may be relevant to understanding the differences in the bioactive conformations and receptor interactions of the two hormones.

Interaction of oxytocin with Ca2+: II. Proton magnetic resonance and molecular modeling studies of conformations of the hormone and its Ca2+ complex.

Drastic changes in the CD and fluorescence spectra of oxytocin [cyclo(Cys1-Tyr2-Ile3-Gln4-Asn5-Cys6)-Pro7-Leu8-Gly 9-NH2] occur on binding Ca2+ in trifluoroethanol (Ananthanarayanan and Brimble, preceding paper). To further characterize the conformation of the Ca(2+)-bound hormone, we carried out 1H-nmr measurements in deuterated trifluorethanol of oxytocin and its 1:1 Ca2+ complex. The one-dimensional nmr data identified residues involved in Ca2+ binding and the extent of their perturbation on Ca2+ addition. The 3JNH-CH coupling constants and two-dimensional nuclear Overhauser effect (NOE) spectral cross peaks confirmed the helical nature of the Ca2+ complex deduced from CD data. Interproton distances in the free hormone and its Ca2+ complex were estimated from the respective NOE data. Apparent global minimum-energy conformations of free and Ca2+ bound oxytocin were computed using the Monte Carlo with energy minimization protocol, with and without incorporating the NOE-derived distance constraints. Taken together, our results show Ca2+ binding to oxytocin to be a two-step process. The binding of the first Ca2+ brings the otherwise extended tail segment of oxytocin closer to the ring moiety so that it wraps around the cation. This causes the maximal extent of change in all the spectral parameters. The subsequent formation of the 2:1 Ca-oxytocin complex results in the tail detaching itself away from the ring so as to bind the second Ca2+ ion. This leads to further spectral changes in the hormone molecule. The tail segment plays a major role in both steps. These observations may be useful in understanding the structural basis of oxytocin action.

Interaction of peptide substrates of fibroblast collagenase with divalent cations: Ca2+ binding by substrate as a suggested recognition signal for collagenase action.

To correlate structural data on substrates of human fibroblast collagenase with their interaction with the enzyme, we have studied: Ac-PLG-s-LLG-O-ethyl ester (I), Dnp-PLGLWA(d-Arg)-NH2 (II), AcGPEGLRVG-O-ethyl ester (III) and Succ-GPLGP-O-amidomethylcoumaryl ester (IV). Peptides I and II represent collagenase cleavage sequences in collagen, peptide III is a mimic for the cleavage site in alpha 2-macroglobulin and peptide IV represents a non-substrate model. Kinetic data showed that peptides I, II and III were substrates of the enzyme. In contrast, peptide IV was not acted upon by the enzyme. Circular dichroism data on the peptides showed that the peptides assume ordered structures in water and trifluoroethanol. In the latter solvent, peptides I and III bound Ca2+ and Zn2+ while peptide II bound Ca2+ but not Zn2+. Peptide IV did not bind either cation in this solvent. Together with the kinetic data, the results suggest that the collagenase cleavage segments in collagen and non-collagen substrates of collagenase could interact with Ca2+ and the enzyme to form a ternary complex. This, in turn, would imply a cofactor role for Ca2+ in collagenase action in addition to the solely structural role ascribed so far to this cation.

Conformational analysis of the Ca(2+)-bound opioid peptides: implications for ligand-receptor interaction.

Based on our earlier proposal on the role of Ca2+ in ligand-receptor recognition and the demonstration of the similarity of the Ca(2+)-bound forms of Met-enkephalin and morphine (Zhorov, B.S. and Ananthanarayanan, V.S., FEBS Lett. 354, 131-134 (1994)) we have undertaken the conformational analysis of a series of the Ca(2+)-bound opioid peptides aiming to find their conformations matching Ca(2+)-bound morphine. A Monte Carlo-with-energy-minimization method was used to calculate 14 opioid peptides in the presence of Ca2+. Low-energy conformations of the Ca2+ complexes of peptides with high mu-affinity were found to resemble closely morphine-Ca2+ complex. In contrast, the Ca2+ complexes of peptides with low mu-affinity did not. The results are relevant for understanding the structure-activity relations of opioid receptor ligands.

Interaction of calcium channel antagonists with calcium: structural studies on nicardipine and its Ca2+ complex.

Conformational features of nicardipine in acetonitrile, in the absence and presence of Ca2+, were investigated by one-dimensional NMR and difference absorption spectroscopy techniques. The data show that in acetonitrile solution the antiperiplanar form of nicardipine is dominant. The addition of Ca2+ to the drug solution caused marked changes in the difference absorbance spectra in the 200-400 nm region and in many of its 1H and 13C NMR resonances. The changes were most significant up to a ratio of 0.5 Ca2+:drug. Analysis of the binding data showed the predominant species to be a 2:1 drug:Ca2+ "sandwich" complex with an estimated dissociation constant of 100 microM at 25 degrees C. One-dimensional nuclear Overhauser effect (NOE) experiments revealed through-space connectivities in the drug before and after Ca2+ binding. These changes in conjunction with the changes in 1H and 13C chemical shifts suggest a structure in which the 4-aryl ring substitute of the pyridine moiety moves closer to the C3-side chain in the presence of Ca2+. This attraction is achieved via the chelation of the Ca2+ ion by the oxygen atoms in the m-NO2 of the aryl group and the COOCH2 group in the side chain of the dihydropyridine ring, and gives rise to a stable synperiplanar conformation. A preference for this conformation was also observed in the Ca2+ complex of nifedipine in acetonitrile as inferred from the rather limited NOE data obtained. Our study provides a detailed solution structure for nicardipine and also leads to a suggestion of a role for Ca2+ in the action of this and possibly other dihydropyridines.