Evaluating protein structures determined by structural genomics consortia

Structural genomics projects are providing large quantities of new 3D structural data for proteins. To monitor the quality of these data, we have developed the protein structure validation software suite (PSVS), for assessment of protein structures generated by NMR or X-ray crystallographic methods. PSVS is broadly applicable for structure quality assessment in structural biology projects. The software integrates under a single interface analyses from several widely-used structure quality evaluation tools, including PROCHECK (Laskowski et al., J Appl Crystallog 1993;26:283-291), MolProbity (Lovell et al., Proteins 2003;50:437-450), Verify3D (Luthy et al., Nature 1992;356:83-85), ProsaII (Sippl, Proteins 1993;17: 355-362), the PDB validation software, and various structure-validation tools developed in our own laboratory. PSVS provides standard constraint analyses, statistics on goodness-of-fit between structures and experimental data, and knowledge-based structure quality scores in standardized format suitable for database integration. The analysis provides both global and site-specific measures of protein structure quality. Global quality measures are reported as Z scores, based on calibration with a set of high-resolution X-ray crystal structures. PSVS is particularly useful in assessing protein structures determined by NMR methods, but is also valuable for assessing X-ray crystal structures or homology models. Using these tools, we assessed protein structures generated by the Northeast Structural Genomics Consortium and other international structural genomics projects, over a 5-year period. Protein structures produced from structural genomics projects exhibit quality score distributions similar to those of structures produced in traditional structural biology projects during the same time period. However, while some NMR structures have structure quality scores similar to those seen in higher-resolution X-ray crystal structures, the majority of NMR structures have lower scores. Potential reasons for this "structure quality score gap" between NMR and X-ray crystal structures are discussed.     

1.67-A X-ray structure of the B2 immunoglobulin-binding domain of streptococcal protein G and comparison to the NMR structure of the B1 domain.

The structure of the B2 immunoglobulin-binding domain of streptococcal protein G has been determined at 1.67-A resolution using a combination of single isomorphous replacement (SIR) phasing and manual fitting of the coordinates of the NMR structure of B1 domain of streptococcal protein G [Gronenborn, A. M., et al. (1991) Science 253, 657-661]. The final R value was 0.191 for data between 8.0 and 1.67 A. The structure described here has 13 residues preceding the 57-residue Ig-binding domain and 13 additional residues following it, for a total of 83 residues. The 57-residue binding domain is well-determined in the structure, having an average B factor of 18.0. Only residues 8-77 could be located in the electron density maps, with the ends of the structure fading into disorder. Like the B1 domain, the B2 domain consists of four beta-strands and a single helix lying diagonally across the beta-sheet, with a -1, +3 chi, -1 topology. This small structure is extensively hydrogen-bonded and has a relatively large hydrophobic core. These structural observations may account for the exceptional stability of protein G. A comparison of the B2 domain X-ray structure and the B1 domain NMR structure showed minor differences in the turn between strands and two and a slight displacement of the helix relative to the sheet. Hydrogen bonds between crystallographically related molecules account for most of these differences.     

Modeling the structure of Pyrococcus furiosus rubredoxin by homology to other X-ray structures.

The three-dimensional structure of rubredoxin from the hyperthermophilic archaebacterium, Pyrococcus furiosus, has been modeled from the X-ray crystal structures of three homologous proteins from Clostridium pasteurianum, Desulfovibrio gigas, and Desulfovibrio vulgaris. All three homology models are similar. When comparing the positions of all heavy atoms and essential hydrogen atoms to the recently solved crystal structure (Day, M. W., et al., 1992, Protein Sci. 1, 1494-1507) of the same protein, the homology model differ from the X-ray structure by 2.09 A root mean square (RMS). The X-ray and the zinc-substituted NMR structures (Blake, P. R., et al., 1992b, Protein Sci. 1, 1508-1521) show a similar level of difference (2.05 A RMS). On average, the homology models are closer to the X-ray structure than to the NMR structures (2.09 vs. 2.42 A RMS).     

Role of proline residues in the structure and function of a membrane transport protein

By use of site-directed mutagenesis, each prolyl residue in the lac permease of Escherichia coli at positions 28 (putative helix I), 31 (helix I), 61 (helix II), 89 (helix III), 97 (helix III), 123 (helix IV), 192 (putative hydrophilic region 7), 220 (helix VII), 280 (helix VIII), and 327 [helix X; Lolkema, J. S., et al. (1988) Biochemistry 27, 8307] was systematically replaced with Gly, Ala, or Leu or deleted by truncation of the C-terminus [i.e., Pro403 and Pro405; Roepe, P.D., et al. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 3992]. Replacements were chosen on the basis of side-chain helical propensity: Gly, like Pro, is thought to be a "helix breaker", while Ala and Leu are "helix makers". With the exception of Pro28, each prolyl residue can be replaced with Gly or Ala, and Pro403 and -405 can be deleted with the C-terminal tail, and significant lac permease activity is retained. In contrast, when Pro28 is replaced with Gly, Ala, or Ser, lactose transport is abolished, but permease with Ser28 binds p-nitrophenyl alpha-D-galactopyranoside and catalyzes active transport of beta-galactopyranosyl-1-thio-beta-D- galactopyranoside. Replacement of Pro28, -31, -123, -280, or -327 with Leu abolishes lactose transport, while replacement of Pro61, -89, -97, or -220 with Leu has relatively minor effects. None of the alterations in permease activity is due to inability of the mutant proteins to insert into the membrane or to diminished lifetimes after insertion, since the concentration of each mutant permease in the membrane is comparable to that of wild-type permease as judged by immunological analyses.(ABSTRACT TRUNCATED AT 250 WORDS)     

1H NMR studies of DNA recognition by the glucocorticoid receptor: complex of the DNA binding domain with a half-site response element.

The complex of the rat glucocorticoid receptor (GR) DNA binding domain (DBD) and half-site sequence of the consensus glucocorticoid response element (GRE) has been studied by two-dimensional 1H NMR spectroscopy. The DNA fragment is a 10 base-pair oligonucleotide, 5'd(GCTGTTCTGC)3'.5'd-(GCAGAACAGC)3', containing the stronger binding GRE half-site hexamer, with GC base pairs at each end. The 93-residue GR-DBD contains an 86-residue segment corresponding to residues 440-525 of the rat GR. Eleven NOE cross peaks between the protein and DNA have been identified, and changes in the chemical shift of the DNA protons upon complex formation have been analyzed. Using these protein-DNA contact points, it can be concluded that (i) the "recognition helix" formed by residues C460-E469 lies in the major groove of the DNA; (ii) the GR-DBD is oriented on the GRE half-site such that residues A477-D481, forming the so-called D-loop, are available for protein-protein interaction in the GR-DBD dimer on the intact consensus GRE; and (iii) the 5-methyl of the second thymine in the half-site and valine 462 interact, confirming indirect evidence [Truss et al. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 7180-7184; Mader et al. (1989) Nature 338, 271-274] that both play an important role in GR-DBD DNA binding. These findings are consistent with the model proposed by H?rd et al. [(1990) Science 249, 157-160] and the X-ray crystallographic complex structure determined by Luisi et al. [(1991) Nature 352, 497-505].     

Validation of helical tilt angles in the solution NMR structure of the Z domain of Staphylococcal protein A by combined analysis of residual dipolar coupling and NOE data

Staphylococcal protein A (SpA) is a virulence factor from Staphylococcus aureus that is able to bind to immunoglobulins. The 3D structures of its immunoglobulin (Ig) binding domains have been extensively studied by NMR and X-ray crystallography, and are often used as model structures in developing de novo or ab initio strategies for predicting protein structure. These small three-helix-bundle structures, reported in free proteins or Ig-bound complexes, have been determined previously using medium- to high-resolution data. Although the location and relative orientation of the three helices in most of these published 3D domain structures are consistent, there are significant differences among the reported structures regarding the tilt angle of the first helix (helix 1). We have applied residual dipolar coupling data, together with nuclear Overhauser enhancement and scalar coupling data, in refining the NMR solution structure of an engineered IgG-binding domain (Z domain) of SpA. Our results demonstrate that the three helices are almost perfectly antiparallel in orientation, with the first helix tilting slightly away from the other two helices. We propose that this high-accuracy structure of the Z domain of SpA is a more suitable target for theoretical predictions of the free domain structure than previously published lower-accuracy structures of protein A domains.     

The rate and structural consequences of proline cis-trans isomerization in calbindin D9k: NMR studies of the minor (cis-Pro43) isoform and the Pro43Gly mutant

The EF-hand calcium-binding protein, calbindin D9k, exists in solution in the calcium-loaded state, as a 1:3 equilibrium mixture of two isoforms, the result of cis-trans isomerism at the Gly42-Pro43 peptide bond [Chazin et al. (1989) Proc. Natl. Acad. Sci. U.S.A. 86, 2195-2198]. Nuclear magnetic resonance (NMR) studies of the minor (cis-Pro43) isoform and the Pro43----Gly mutant are reported here. The rate of cis----trans isomerization at the Pro43 peptide bond in the wild-type protein was determined by line-shape analysis at elevated temperatures, using a sample in which all amino acids, except Ser and Val, were deuterated. The cis----trans rate is calculated to be 0.2 s-1 at 25 degrees C, corresponding to a free energy of activation, delta G, of 77 kJ/mol. The complete sequence-specific 1H NMR assignments of the cis-Pro43 isoform and the Pro43----Gly mutant in the calcium-loaded state have been obtained by using standard methods combined with comparisons to the previously assigned major (trans-Pro43) isoform. This has permitted detailed comparative analysis of 1H NMR chemical shifts, backbone scalar coupling constants, and nuclear Overhauser effects. The minor isoform has a global fold that is identical with that of the major isoform. Structural changes imposed by cis-trans isomerization at Pro43 are highly localized to the linker loop (containing Pro43) that joins the two EF hands. The Pro43----Gly mutant has a global fold that is identical with the wild-type protein, but does not exhibit conformational heterogeneity. Only very limited structural differences are observed between mutant and wild-type protein, and these are also highly localized to the linker loop. The ion-binding properties of the mutant, as determined by 43Ca and 113Cd NMR, are found to be very similar to the wild-type protein. These results provide crucial evidence that justifies the calculation of high-resolution three-dimensional structures of the Pro43Gly mutant, rather than of the conformationally heterogeneous wild-type protein.     

Chiral interaction in peptide molecules: effects of chiral peptide species on helix-sense induction in an N-terminal-free achiral peptide

Noncovalent chiral domino effect (NCDE) has been proposed as terminal-specific interaction upon a 3(10)-helical peptide chain, of which the helix sense is manipulated by an external chiral stimulus (mainly amino acid derivatives) operating on the N-terminus (Inai, Y., et al. J Am Chem Soc 2000, 122, 11731-11732; ibid., 2002, 124, 2466-2473; ibid., 2003, 125, 8151-8162). We have investigated here a helix-sense induction in an optically inactive N-terminal-free nonapeptide (1) through the screening of several peptide species that differ in chiral sequence, chain length, and C-terminal group. Helix-sense induction in peptide 1 depends largely on both the C-terminal chirality and carboxyl group in the external peptide, in which N-carbonyl-blocked amino acids, "monopeptide acids," should be the minimum requirement for effective induction. N-Protected mono- to tetrapeptides of L-Leu residue commonly induce a right-handed helix. NMR study and theoretical computation reveal that the N-terminal segment of peptide 1 binds the N-protected dipeptide molecule through multipoint coordination to induce a right-handed helix preferentially. The present findings not only will improve our understanding of the chiral roles in peptide or protein helical termini, but also might demonstrate potential applications to chirality-responsive materials based on peptide helical fragments.     

Structures of the two 3D domain-swapped RNase A trimers

When concentrated in mildly acidic solutions, bovine pancreatic ribonuclease (RNase A) forms long-lived oligomers including two types of dimer, two types of trimer, and higher oligomers. In previous crystallographic work, we found that the major dimeric component forms by a swapping of the C-terminal beta-strands between the monomers, and that the minor dimeric component forms by swapping the N-terminal alpha-helices of the monomers. On the basis of these structures, we proposed that a linear RNase A trimer can form from a central molecule that simultaneously swaps its N-terminal helix with a second RNase A molecule and its C-terminal strand with a third molecule. Studies by dissociation are consistent with this model for the major trimeric component: the major trimer dissociates into both the major and the minor dimers, as well as monomers. In contrast, the minor trimer component dissociates into the monomer and the major dimer. This suggests that the minor trimer is cyclic, formed from three monomers that swap their C-terminal beta-strands into identical molecules. These conclusions are supported by cross-linking of lysyl residues, showing that the major trimer swaps its N-terminal helix, and the minor trimer does not. We verified by X-ray crystallography the proposed cyclic structure for the minor trimer, with swapping of the C-terminal beta-strands. This study thus expands the variety of domain-swapped oligomers by revealing the first example of a protein that can form both a linear and a cyclic domain-swapped oligomer. These structures permit interpretation of the enzymatic activities of the RNase A oligomers on double-stranded RNA.     

Primary structure of pregnancy zone protein. Molecular cloning of a full-length PZP cDNA clone by the polymerase chain reaction

A full-length cDNA clone of the human pregnancy zone protein (PZP) was cloned from the hepatocellular carcinoma cell line Hep3B. Based on the exon sequences of the PZP gene (Devriendt et al. (1989) Gene 81, 325-334; Marynen et al., unpublished data), primer pairs were designed to amplify six overlapping fragments of the PZP cDNA. The obtained cDNA is 4609 bp long and contains an open reading frame coding for 1482 amino acids, including a signal peptide of 25 amino acid residues. Comparison with the published partial PZP amino acid sequence (Sottrup-Jensen et al. (1984) Proc. Natl. Acad. Sci. USA 81, 7353-7357) and the PZP genomic sequences confirmed the identity as a PZP cDNA. 71% of the corresponding amino acid residues in PZP and human alpha 2-macroglobulin (alpha 2M) are identical and all cysteine residues are conserved. A typical internal thiol ester site and a bait domain were identified. A Pro/Thr polymorphism was identified at amino acid position 1180, and an A/G nucleotide polymorphism at bp 4097.     

Detergent binding to unmyristylated protein kinase A—Structural implications for the role of Myristate

Myristylation often governs the targeting of protein kinases to the plasma membrane. It is now known that a key member of the src family of protein tyrosine kinases, pp60^v-src, binds to the lipid bilayer of the plasma membrane via a myristylated amino terminal sequence. The mechanism of this interaction is not known; however, myristic acid (Myristic acid may also be referred to as Myristate) and residues 2 through 14 are also absolutely required (Resh and Ling, 1990). This review presents an analysis of crystal structures of detergent-modified recombinant and myristylated mammalian catalytic subunit of protein kinase A. Crystals of unmyristylated recombinant catalytic subunit of protein kinase A are grown in the presence of Mega 8, a glucamide-type of detergent, and only this detergent binds, which results in a resolution extension (Knightonet al., 1991a). Comparisons of these two structures reveal that the detergent association with the recombinant enzyme binds in exactly the same hydrophobic pocket of the protein occupied by myristic acid in the mammalian protein (Karlssonet al., 1993; Zhenget al., 1993a). Removal of the detergent through soaking results in the local unwinding of the first helix, helix A, and disorder of the canonical recognition sequence of the phosphorylation site, Ser 10 (Zhenget al., 1993b). These results suggest that anchoring the myristic acid inside the protein results in formation of a stable structural template, which includes the myristylated amino terminal sequence important for the recognition by protein kinases. This inside out motif might provide a structural paradigm for the recognition of myristylated proteins, including pp60^v-src.     

An NMR study of the helix V-loop E region of the 5S RNA from Escherichia coli

Experiments are described that complete the assignment of the imino proton NMR spectrum of the fragment 1 domain from the 5S RNA of Escherichia coli. Most of the new assignments fall in the helix V-loop E portion of the molecule (bases 70-78 and 98-106), the region most sensitive to the binding of ribosomal protein L25. The spectroscopic data are incompatible with the standard, phylogenetically derived model for 5S RNA, which makes all the base pairs possible in loop E with the sequences aligned in parallel (C70-G106, C71-G105, etc.) [see Delihas et al. (1984) Prog. Nucleic Acid Res. Mol. Biol. 31, 161-190]. Furthermore, the alternative loop E model proposed for spinach chloroplast 5S RNA by Romby et al. [(1988) Biochemistry 27, 4721-4730] does not apply to the closely homologous 5S RNA from E. coli. The 5S RNAs from E. coli and spinach chloroplasts do not have the same secondary structures in solution despite their strong sequence homologies, and neither appears to conform to the standard model for 5S RNA in the loop E region.     

Use of restrained molecular dynamics in water to determine three-dimensional protein structure: prediction of the three-dimensional structure of Ecballium elaterium trypsin inhibitor II

Refinement of distance geometry (DG) structures of EETI-II (Heitz et al.: Biochemistry 28:2392-2398, 1989), a member of the squash family trypsin inhibitor, have been carried out by restrained molecular dynamics (RMD) in water. The resulting models show better side chain apolar/polar surface ratio and estimated solvation free energy than structures refined "in vacuo." The consistent lower values of residual NMR constraint violations, apolar/polar surface ratio, and solvation free energy for one of these refined structures allowed prediction of the 3D folding and disulfide connectivity of EETI-II. Except for the few first residues for which no NMR constraints were available, this computer model fully agreed with X-ray structures of CMTI-I (Bode et al.: FEBS Lett. 242:285-292, 1989) and EETI-II complexed with trypsin that appeared after the RMD simulation was completed. Restrained molecular dynamics in water is thus proved to be highly valuable for refinement of DG structures. Also, the successful use of apolar/polar surface ratio and of solvation free energy reinforce the analysis of Novotny et al. (Proteins 4:19-30, 1988) and shows that these criteria are useful indicators of correct versus misfolded models.     

Comparison of simulated and experimentally determined dynamics for a variant of the Lacl DNA-binding domain, Nlac-P.

Recent advances in the experimentally determined structures and dynamics of the domains within LacI provide a rare context for evaluating dynamics calculations. A 1500-ps trajectory was simulated for a variant of the LacI DNA-binding domain, which consists of the first three helices in LacI and the hinge helix of the homologous PurR. Order parameters derived from dynamics simulations are compared to those obtained for the LacI DNA-binding domain with 15N relaxation NMR spectroscopy (Slijper et al., 1997. Biochemistry. 36:249-254). The MD simulations suggest that the unstructured loop between helices II and III does not exist in a discrete state under the conditions of no salt and neutral pH, but occupies a continuum of states between the DNA-bound and free structures. Simulations also indicate that the unstructured region between helix III and the hinge helix is very mobile, rendering motions of the hinge helix essentially independent of the rest of the protein. Finally, the alpha-helical hydrogen bonds in the hinge helix are broken after 1250 ps, perhaps as a prelude to helix unfolding.     

Spectroscopy-Based Modelling of the 3D Structure of the β Subunit of the High Affinity IgE Receptor

Abstract

The high affinity IgE receptor, possesses a tetrameric structure. The 243 residue β subunit is a polytopic protein with four hydrophobic membrane-spanning segments, whereas the individual α and γ subunits are bitopic proteins each containing one transmembrane domain in their monomeric form. In the proposed topographical model (Blank et al., 1989), the four trans-membrane α helices of the β subunit are connected by three loop sequences.

To study the individual subunits and intact receptor, this membrane protein was divided into domains such as its loop peptides, cytoplasmic peptides and transmembrane helices according to Blank et al., 1989. The 3D structure of the synthesized loop peptides and cytoplasmic peptides were calculated; CD and/or NMR data were used as appropriate to generate the resultant structures which were then used as data basis for the higher level calculations.

The four individual transmembrane helices of the β subunit were characterised, first of all, by mapping the relative lipophilicity of their surfaces using lipophilic probes. A second procedure, docking of the individual helices in pairs, was used to predict helix–helix interactions.

The data on the relative lipophilicity of the surfaces as well as the surfaces that favoured helix–helix interactions were used in combination with the spectroscopy-based structures of the loops and cytoplasmic domains to calculate via molecular dynamics, the helix arrangement and 3D structure of the β subunit of the high affinity IgE receptor. In the final analysis, the molecular simulations yielded two structures of the β subunit, which should form a basis for the modelling of the whole high affinity IgE receptor.     

TASSER-based refinement of NMR structures

The TASSER structure prediction algorithm is employed to investigate whether NMR structures can be moved closer to their corresponding X-ray counterparts by automatic refinement procedures. The benchmark protein dataset includes 61 nonhomologous proteins whose structures have been determined by both NMR and X-ray experiments. Interestingly, by starting from NMR structures, the majority (79%) of TASSER refined models show a structural shift toward their X-ray structures. On average, the TASSER refined models have a root-mean-square-deviation (RMSD) from the X-ray structure of 1.785 A (1.556 A) over the entire chain (aligned region), while the average RMSD between NMR and X-ray structures (RMSD(NMR_X-ray)) is 2.080 A (1.731 A). For all proteins having a RMSD(NMR_X-ray) >2 A, the TASSER refined structures show consistent improvement. However, for the 34 proteins with a RMSD(NMR_X-ray) <2 A, there are only 21 cases (60%) where the TASSER model is closer to the X-ray structure than NMR, which may be due to the inherent resolution of TASSER. We also compare the TASSER models with 12 NMR models in the RECOORD database that have been recalculated recently by Nederveen et al. from original NMR restraints using the newest molecular dynamics tools. In 8 of 12 cases, TASSER models show a smaller RMSD to X-ray structures; in 3 of 12 cases, where RMSD(NMR_X-ray) <1 A, RECOORD does better than TASSER. These results suggest that TASSER can be a useful tool to improve the quality of NMR structures.     

Three-dimensional structure of acyl carrier protein in solution determined by nuclear magnetic resonance and the combined use of dynamical simulated annealing and distance geometry

The solution conformation of acyl carrier protein from Escherichia coli (77 residues) has been determined on the basis of 423 interproton-distance restraints and 32 hydrogen-bonding restraints derived from NMR measurements. A total of nine structures were computed using a hybrid approach combining metric matrix distance geometry and dynamic simulated annealing. The polypeptide fold is well defined with an average backbone atomic root-mean-square difference of 0.20 +/- 0.03 nm between the final nine converged structures and the mean structure obtained by averaging their coordinates. The principal structural motif is composed of three helices: 1 (residues 3-12), 2 (residues 37-47) and 4 (residues 65-75) which line a hydrophobic cavity. Helices 2 and 4 are approximately parallel to each other and anti-parallel at an angle of approximately equal to 150 degrees to helix 1. The smaller helix 3 (residues 56-63) is at an angle of approximately equal to 100 degrees to helix 4.     

Structure of H/ACA RNP protein Nhp2p reveals cis/trans isomerization of a conserved proline at the RNA and Nop10 binding interface

H/ACA small nucleolar and Cajal body ribonucleoproteins (RNPs) function in site-specific pseudouridylation of eukaryotic rRNA and snRNA, rRNA processing, and vertebrate telomerase biogenesis. Nhp2, one of four essential protein components of eukaryotic H/ACA RNPs, forms a core trimer with the pseudouridylase Cbf5 and Nop10 that binds to H/ACA RNAs specifically. Crystal structures of archaeal H/ACA RNPs have revealed how the protein components interact with each other and with the H/ACA RNA. However, in place of Nhp2p, archaeal H/ACA RNPs contain L7Ae, which binds specifically to an RNA K-loop motif absent from eukaryotic H/ACA RNPs, while Nhp2 binds a broader range of RNA structures. We report solution NMR studies of Saccharomyces cerevisiae Nhp2 (Nhp2p), which reveal that Nhp2p exhibits two major conformations in solution due to cis/trans isomerization of the evolutionarily conserved Pro83. The equivalent proline is in the cis conformation in all reported structures of L7Ae and other homologous proteins. Nhp2p has the expected α-β-α fold, but the solution structures of the major conformation of Nhp2p with trans Pro83 and of Nhp2p-S82W with cis Pro83 reveal that Pro83 cis/trans isomerization affects the positions of numerous residues at the Nop10 and RNA binding interface. An S82W substitution, which stabilizes the cis conformation, also stabilizes the association of Nhp2p with H/ACA snoRNPs expressed in vivo. We propose that Pro83 plays a key role in the assembly of the eukaryotic H/ACA RNP, with the cis conformation locking in a stable Cbf5-Nop10-Nhp2 ternary complex and positioning the protein backbone to interact with the H/ACA RNA.