Prevalence Estimates of Predicted Pathogenic COL4A3-COL4A5 Variants in a Population Sequencing Database and Their Implications for Alport Syndrome.

BACKGROUND: The reported prevalence of Alport syndrome varies from one in 5000 to one in 53,000 individuals. This study estimated the frequencies of predicted pathogenic COL4A3-COL4A5 variants in sequencing databases of populations without known kidney disease. METHODS: Predicted pathogenic variants were identified using filtering steps based on the ACMG/AMP criteria, which considered collagen IV α3-α5 position 1 Gly to be critical domains. The population frequencies of predicted pathogenic COL4A3-COL4A5 variants were then determined per mean number of sequenced alleles. Population frequencies for compound heterozygous and digenic combinations were calculated from the results for heterozygous variants. RESULTS: COL4A3-COL4A5 variants resulting in position 1 Gly substitutions were confirmed to be associated with hematuria (for each, P<0.001). Predicted pathogenic COL4A5 variants were found in at least one in 2320 individuals. p.(Gly624Asp) represented nearly half (16 of 33, 48%) of the variants in Europeans. Most COL4A5 variants (54 of 59, 92%) had a biochemical feature that potentially mitigated the clinical effect. The predicted pathogenic heterozygous COL4A3 and COL4A4 variants affected one in 106 of the population, consistent with the finding of thin basement membrane nephropathy in normal donor kidney biopsy specimens. Predicted pathogenic compound heterozygous variants occurred in one in 88,866 individuals, and digenic variants in at least one in 44,793. CONCLUSIONS: The population frequencies for Alport syndrome are suggested by the frequencies of predicted pathogenic COL4A3-COL4A5 variants, but must be adjusted for the disease penetrance of individual variants and for the likelihood of already diagnosed disease and non-Gly substitutions. Disease penetrance may depend on other genetic and environmental factors.

Pervasive variation of transcription factor orthologs contributes to regulatory network evolution.

Differences in transcriptional regulatory networks underlie much of the phenotypic variation observed across organisms. Changes to cis-regulatory elements are widely believed to be the predominant means by which regulatory networks evolve, yet examples of regulatory network divergence due to transcription factor (TF) variation have also been observed. To systematically ascertain the extent to which TFs contribute to regulatory divergence, we analyzed the evolution of the largest class of metazoan TFs, Cys2-His2 zinc finger (C2H2-ZF) TFs, across 12 Drosophila species spanning ~45 million years of evolution. Remarkably, we uncovered that a significant fraction of all C2H2-ZF 1-to-1 orthologs in flies exhibit variations that can affect their DNA-binding specificities. In addition to loss and recruitment of C2H2-ZF domains, we found diverging DNA-contacting residues in ~44% of domains shared between D. melanogaster and the other fly species. These diverging DNA-contacting residues, found in ~70% of the D. melanogaster C2H2-ZF genes in our analysis and corresponding to ~26% of all annotated D. melanogaster TFs, show evidence of functional constraint: they tend to be conserved across phylogenetic clades and evolve slower than other diverging residues. These same variations were rarely found as polymorphisms within a population of D. melanogaster flies, indicating their rapid fixation. The predicted specificities of these dynamic domains gradually change across phylogenetic distances, suggesting stepwise evolutionary trajectories for TF divergence. Further, whereas proteins with conserved C2H2-ZF domains are enriched in developmental functions, those with varying domains exhibit no functional enrichments. Our work suggests that a subset of highly dynamic and largely unstudied TFs are a likely source of regulatory variation in Drosophila and other metazoans.

A systematic survey of the Cys2His2 zinc finger DNA-binding landscape.

Cys2His2 zinc fingers (C2H2-ZFs) comprise the largest class of metazoan DNA-binding domains. Despite this domain's well-defined DNA-recognition interface, and its successful use in the design of chimeric proteins capable of targeting genomic regions of interest, much remains unknown about its DNA-binding landscape. To help bridge this gap in fundamental knowledge and to provide a resource for design-oriented applications, we screened large synthetic protein libraries to select binding C2H2-ZF domains for each possible three base pair target. The resulting data consist of >160 000 unique domain-DNA interactions and comprise the most comprehensive investigation of C2H2-ZF DNA-binding interactions to date. An integrated analysis of these independent screens yielded DNA-binding profiles for tens of thousands of domains and led to the successful design and prediction of C2H2-ZF DNA-binding specificities. Computational analyses uncovered important aspects of C2H2-ZF domain-DNA interactions, including the roles of within-finger context and domain position on base recognition. We observed the existence of numerous distinct binding strategies for each possible three base pair target and an apparent balance between affinity and specificity of binding. In sum, our comprehensive data help elucidate the complex binding landscape of C2H2-ZF domains and provide a foundation for efforts to determine, predict and engineer their DNA-binding specificities.

De novo prediction of DNA-binding specificities for Cys2His2 zinc finger proteins.

Proteins with sequence-specific DNA binding function are important for a wide range of biological activities. De novo prediction of their DNA-binding specificities from sequence alone would be a great aid in inferring cellular networks. Here we introduce a method for predicting DNA-binding specificities for Cys2His2 zinc fingers (C2H2-ZFs), the largest family of DNA-binding proteins in metazoans. We develop a general approach, based on empirical calculations of pairwise amino acid-nucleotide interaction energies, for predicting position weight matrices (PWMs) representing DNA-binding specificities for C2H2-ZF proteins. We predict DNA-binding specificities on a per-finger basis and merge predictions for C2H2-ZF domains that are arrayed within sequences. We test our approach on a diverse set of natural C2H2-ZF proteins with known binding specificities and demonstrate that for >85% of the proteins, their predicted PWMs are accurate in 50% of their nucleotide positions. For proteins with several zinc finger isoforms, we show via case studies that this level of accuracy enables us to match isoforms with their known DNA-binding specificities. A web server for predicting a PWM given a protein containing C2H2-ZF domains is available online at http://zf.princeton.edu and can be used to aid in protein engineering applications and in genome-wide searches for transcription factor targets.

Deep sequencing of large library selections allows computational discovery of diverse sets of zinc fingers that bind common targets.

The Cys2His2 zinc finger (ZF) is the most frequently found sequence-specific DNA-binding domain in eukaryotic proteins. The ZF's modular protein-DNA interface has also served as a platform for genome engineering applications. Despite decades of intense study, a predictive understanding of the DNA-binding specificities of either natural or engineered ZF domains remains elusive. To help fill this gap, we developed an integrated experimental-computational approach to enrich and recover distinct groups of ZFs that bind common targets. To showcase the power of our approach, we built several large ZF libraries and demonstrated their excellent diversity. As proof of principle, we used one of these ZF libraries to select and recover thousands of ZFs that bind several 3-nt targets of interest. We were then able to computationally cluster these recovered ZFs to reveal several distinct classes of proteins, all recovered from a single selection, to bind the same target. Finally, for each target studied, we confirmed that one or more representative ZFs yield the desired specificity. In sum, the described approach enables comprehensive large-scale selection and characterization of ZF specificities and should be a great aid in furthering our understanding of the ZF domain.

Ensemble Detection of DNA Engineering Signatures.

Synthetic biology is creating genetically engineered organisms at an increasing rate for many potentially valuable applications, but this potential comes with the risk of misuse or accidental release. To begin to address this issue, we have developed a system called GUARDIAN that can automatically detect signatures of engineering in DNA sequencing data, and we have conducted a blinded test of this system using a curated Test and Evaluation (T&E) data set. GUARDIAN uses an ensemble approach based on the guiding principle that no single approach is likely to be able to detect engineering with perfect accuracy. Critically, ensembling enables GUARDIAN to detect sequence inserts in 13 target organisms with a high degree of specificity that requires no subject matter expert (SME) review.

Designed coiled coils promote folding of a recombinant bacterial collagen.

Collagen triple helices fold slowly and inefficiently, often requiring adjacent globular domains to assist this process. In the Streptococcus pyogenes collagen-like protein Scl2, a V domain predicted to be largely α-helical, occurs N-terminal to the collagen triple helix (CL). Here, we replace this natural trimerization domain with a de novo designed, hyperstable, parallel, three-stranded, α-helical coiled coil (CC), either at the N terminus (CC-CL) or the C terminus (CL-CC) of the collagen domain. CD spectra of the constructs are consistent with additivity of independently and fully folded CC and CL domains, and the proteins retain their distinctive thermal stabilities, CL at ∼37 °C and CC at >90 °C. Heating the hybrid proteins to 50 °C unfolds CL, leaving CC intact, and upon cooling, the rate of CL refolding is somewhat faster for CL-CC than for CC-CL. A construct with coiled coils on both ends, CC-CL-CC, retains the ∼37 °C thermal stability for CL but shows less triple helix at low temperature and less denaturation at 50 °C. Most strikingly however, in CC-CL-CC, the CL refolds slower than in either CC-CL or CL-CC by almost two orders of magnitude. We propose that a single CC promotes folding of the CL domain via nucleation and in-register growth from one end, whereas initiation and growth from both ends in CC-CL-CC results in mismatched registers that frustrate folding. Bioinformatics analysis of natural collagens lends support to this because, where present, there is generally only one coiled-coil domain close to the triple helix, and it is nearly always N-terminal to the collagen repeat.

An expanded binding model for Cys2His2 zinc finger protein-DNA interfaces.

Cys(2)His(2) zinc finger (C2H2-ZF) proteins comprise the largest class of eukaryotic transcription factors. The 'canonical model' for C2H2-ZF protein-DNA interaction consists of only four amino acid-nucleotide contacts per zinc finger domain, and this model has been the basis for several efforts for computationally predicting and experimentally designing protein-DNA interfaces. Here, we perform a systematic analysis of structural and experimental binding data and find that, in addition to the canonical contacts, several other amino acid and base pair combinations frequently play a role in C2H2-ZF protein-DNA binding. We suggest an expansion of the canonical C2H2-ZF model to include one to three additional contacts, and show that computational approaches including these additional contacts improve predictions of DNA targets of zinc finger proteins.

Predicting DNA recognition by Cys2His2 zinc finger proteins.

MOTIVATION: Cys(2)His(2) zinc finger (ZF) proteins represent the largest class of eukaryotic transcription factors. Their modular structure and well-conserved protein-DNA interface allow the development of computational approaches for predicting their DNA-binding preferences even when no binding sites are known for a particular protein. The 'canonical model' for ZF protein-DNA interaction consists of only four amino acid nucleotide contacts per zinc finger domain. RESULTS: We present an approach for predicting ZF binding based on support vector machines (SVMs). While most previous computational approaches have been based solely on examples of known ZF protein-DNA interactions, ours additionally incorporates information about protein-DNA pairs known to bind weakly or not at all. Moreover, SVMs with a linear kernel can naturally incorporate constraints about the relative binding affinities of protein-DNA pairs; this type of information has not been used previously in predicting ZF protein-DNA binding. Here, we build a high-quality literature-derived experimental database of ZF-DNA binding examples and utilize it to test both linear and polynomial kernels for predicting ZF protein-DNA binding on the basis of the canonical binding model. The polynomial SVM outperforms previously published prediction procedures as well as the linear SVM. This may indicate the presence of dependencies between contacts in the canonical binding model and suggests that modification of the underlying structural model may result in further improved performance in predicting ZF protein-DNA binding. Overall, this work demonstrates that methods incorporating information about non-binding and relative binding of protein-DNA pairs have great potential for effective prediction of protein-DNA interactions. AVAILABILITY: An online tool for predicting ZF DNA binding is available at http://compbio.cs.princeton.edu/zf/.

Y-position cysteine substitution in type I collagen (alpha1(I) R888C/p.R1066C) is associated with osteogenesis imperfecta/Ehlers-Danlos syndrome phenotype.

The most common mutations in type I collagen causing types II-IV osteogenesis imperfecta (OI) result in substitution for glycine in a Gly-Xaa-Yaa triplet by another amino acid. We delineated a Y-position substitution in a small pedigree with a combined OI/Ehlers-Danlos Syndrome (EDS) phenotype, characterized by moderately decreased DEXA z-score (-1.3 to -2.6), long bone fractures, and large-joint hyperextensibility. Affected individuals have an alpha1(I)R888C (p.R1066C) substitution in one COL1A1 allele. Polyacrylamide gel electrophoresis (PAGE) of [(3)H]-proline labeled steady-state collagen reveals slight overmodification of the alpha1(I) monomer band, much less than expected for a substitution of a neighboring glycine residue, and a faint alpha1(I) dimer. Dimers form in about 10% of proband type I collagen. Dimer formation is inefficient compared to a possible 25%, probably because the SH-side chains have less proximity in this Y-position than when substituting for a glycine. Theoretical stability calculations, differential scanning calorimetry (DSC) thermograms, and thermal denaturation curves showed only weak local destabilization from the Y-position substitution in one or two chains of a collagen helix, but greater destabilization is seen in collagen containing dimers. Y-position collagen dimers cause kinking of the helix, resulting in a register shift that is propagated the full length of the helix and causes resistance to procollagen processing by N-proteinase. Collagen containing the Y-position substitution is incorporated into matrix deposited in culture, including immaturely and maturely cross-linked fractions. In vivo, proband dermal fibrils have decreased density and increased diameter compared to controls, with occasional aggregate formation. This report on Y-position substitutions in type I collagen extends the range of phenotypes caused by nonglycine substitutions and shows that, similar to X- and Y-position substitutions in types II and III collagen, the phenotypes resulting from nonglycine substitutions in type I collagen are distinct from those caused by glycine substitutions.

The collαgen III fibril has a "flexi-rod" structure of flexible sequences interspersed with rigid bioactive domains including two with hemostatic roles.

Collagen III is critical to the integrity of blood vessels and distensible organs, and in hemostasis. Examination of the human collagen III interactome reveals a nearly identical structural arrangement and charge distribution pattern as for collagen I, with cell interaction domains, fibrillogenesis and enzyme cleavage domains, several major ligand-binding regions, and intermolecular crosslink sites at the same sites. These similarities allow heterotypic fibril formation with, and substitution by, collagen I in embryonic development and wound healing. The collagen III fibril assumes a "flexi-rod" structure with flexible zones interspersed with rod-like domains, which is consistent with the molecule's prominence in young, pliable tissues and distensible organs. Collagen III has two major hemostasis domains, with binding motifs for von Willebrand factor, α2ß1 integrin, platelet binding octapeptide and glycoprotein VI, consistent with the bleeding tendency observed with COL3A1 disease-causing sequence variants.

Molecular structure of the collagen triple helix.

The molecular conformation of the collagen triple helix confers strict amino acid sequence constraints, requiring a (Gly-X-Y)(n) repeating pattern and a high content of imino acids. The increasing family of collagens and proteins with collagenous domains shows the collagen triple helix to be a basic motif adaptable to a range of proteins and functions. Its rodlike domain has the potential for various modes of self-association and the capacity to bind receptors, other proteins, GAGs, and nucleic acids. High-resolution crystal structures obtained for collagen model peptides confirm the supercoiled triple helix conformation, and provide new information on hydrogen bonding patterns, hydration, sidechain interactions, and ligand binding. For several peptides, the helix twist was found to be sequence dependent, and such variation in helix twist may serve as recognition features or to orient the triple helix for binding. Mutations in the collagen triple-helix domain lead to a variety of human disorders. The most common mutations are single-base substitutions that lead to the replacement of one Gly residue, breaking the Gly-X-Y repeating pattern. A single Gly substitution destabilizes the triple helix through a local disruption in hydrogen bonding and produces a discontinuity in the register of the helix. Molecular information about the collagen triple helix and the effect of mutations will lead to a better understanding of function and pathology.

X-Linked and Autosomal Recessive Alport Syndrome: Pathogenic Variant Features and Further Genotype-Phenotype Correlations.

Alport syndrome results from mutations in the COL4A5 (X-linked) or COL4A3/COL4A4 (recessive) genes. This study examined 754 previously- unpublished variants in these genes from individuals referred for genetic testing in 12 accredited diagnostic laboratories worldwide, in addition to all published COL4A5, COL4A3 and COL4A4 variants in the LOVD databases. It also determined genotype-phenotype correlations for variants where clinical data were available. Individuals were referred for genetic testing where Alport syndrome was suspected clinically or on biopsy (renal failure, hearing loss, retinopathy, lamellated glomerular basement membrane), variant pathogenicity was assessed using currently-accepted criteria, and variants were examined for gene location, and age at renal failure onset. Results were compared using Fisher's exact test (DNA Stata). Altogether 754 new DNA variants were identified, an increase of 25%, predominantly in people of European background. Of the 1168 COL4A5 variants, 504 (43%) were missense mutations, 273 (23%) splicing variants, 73 (6%) nonsense mutations, 169 (14%) short deletions and 76 (7%) complex or large deletions. Only 135 of the 432 Gly residues in the collagenous sequence were substituted (31%), which means that fewer than 10% of all possible variants have been identified. Both missense and nonsense mutations in COL4A5 were not randomly distributed but more common at the 70 CpG sequences (p<10-41 and p<0.001 respectively). Gly>Ala substitutions were underrepresented in all three genes (p< 0.0001) probably because of an association with a milder phenotype. The average age at end-stage renal failure was the same for all mutations in COL4A5 (24.4 ±7.8 years), COL4A3 (23.3 ± 9.3) and COL4A4 (25.4 ± 10.3) (COL4A5 and COL4A3, p = 0.45; COL4A5 and COL4A4, p = 0.55; COL4A3 and COL4A4, p = 0.41). For COL4A5, renal failure occurred sooner with non-missense than missense variants (p<0.01). For the COL4A3 and COL4A4 genes, age at renal failure occurred sooner with two non-missense variants (p = 0.08, and p = 0.01 respectively). Thus DNA variant characteristics that predict age at renal failure appeared to be the same for all three Alport genes. Founder mutations (with the pathogenic variant in at least 5 apparently- unrelated individuals) were not necessarily associated with a milder phenotype. This study illustrates the benefits when routine diagnostic laboratories share and analyse their data.

Peptide investigations of pairwise interactions in the collagen triple-helix.

Pairwise interactions have been studied for the major secondary structures in proteins. The present work extends the characterization of interactions between side-chains to the context of a collagen triple-helix. In this study, the most frequent Gly-X-Y tripeptide sequences in collagen are characterized in terms of interchain interactions between non-imino acid X and Y residues, through the use of host-guest peptides and statistical frequency analysis. Stabilities predicted on the basis of additivity show good agreement with experimental values for almost half of the peptides, indicating a lack of interaction. A small number of peptides have a stability lower than predicted, while a larger number are more stable than expected. Of all triplets containing residues of opposite charge, only Gly-Lys-Asp and Gly-Arg-Asp exhibit stabilizing electrostatic interactions, and these pairs are found together preferentially in collagens. Repulsion of like charges is observed in Gly-Arg-Lys, Gly-Lys-Arg, and Gly-Glu-Asp sequences, and a small degree of hydrophobic stabilization was observed for the Gly-Leu-Leu guest triplet. The data reported here help clarify basic principles of triple-helix stability. In addition, the experimentally determined stabilities of the tripeptide units found most frequently in collagens constitute a database useful for predicting triple-helix stability in peptides, collagens and other triple-helix-containing proteins.

Equilibrium thermal transitions of collagen model peptides.

The folding of collagen in vitro is very slow and presents difficulties in reaching equilibrium, a feature that may have implications for in vivo collagen function. Peptides serve as good model systems for examining equilibrium thermal transitions in the collagen triple helix. Investigations were carried out to ascertain whether a range of synthetic triple-helical peptides of varying sequences can reach equilibrium, and whether the triple helix to unfolded monomer transition approximates a two-state model. The thermal transitions for all peptides studied are fully reversible given sufficient time. Isothermal experiments were carried out to obtain relaxation times at different temperatures. The slowest relaxation times, on the order of 10-15 h, were observed at the beginning of transitions, and were shown to result from self-association limited by the low concentration of free monomers, rather than cis-trans isomerization. Although the fit of the CD equilibrium transition curves and the concentration dependence of T(m) values support a two-state model, the more rigorous comparison of the calorimetric enthalpy to the van't Hoff enthalpy indicates the two-state approximation is not ideal. Previous reports of melting curves of triple-helical host-guest peptides are shown to be a two-state kinetic transition, rather than an equilibrium transition.

Stability related bias in residues replacing glycines within the collagen triple helix (Gly-Xaa-Yaa) in inherited connective tissue disorders.

A missense mutation leading to the replacement of one Gly in the (Gly-Xaa-Yaa)n repeat of the collagen triple helix can cause a range of heritable connective tissue disorders that depend on the gene in which the mutation occurs. Osteogenesis imperfecta results from mutations in type I collagen, Ehlers-Danlos syndrome type IV from mutations in type III collagen, Alport syndrome from mutations in type IV collagen, and dystrophic epidermolysis bullosa from mutations in type VII collagen. The predicted rates of substitutions by different amino acids for glycine in the alpha1(I), alpha2(I), alpha1(III), alpha5(IV), and alpha1(VII) chains (encoded by COL1A1, COL1A2, COL3A1, COL4A5, and COL7A1, respectively) were compared with missense mutations in those chains that have been observed to cause disease. The spectrum of amino acids replacing Gly was not significantly different from that expected for the alpha1(VII) chains, suggesting that any Gly replacement will cause disease. The distribution of residues replacing Gly was significantly different from that expected for all other collagen chains studied, with a particularly strong bias seen for alpha1(I) and alpha1(III) collagen chains. The bias did not correlate with the degree of chemical dissimilarity between Gly and the replacement residues, but in some cases a relationship was observed with the predicted extent of destabilization of the triple helix. For alpha1(III) collagen chains, the more destabilizing mutations were identified more often than expected. For alpha1(I), the most destabilizing residues, Val, Glu, and Asp, and the least destabilizing residue, Ala, were underrepresented. This bias supports the hypothesis that the level of triple-helix destabilization determines clinical outcome.

Variation in the helical structure of native collagen.

The structure of collagen has been a matter of curiosity, investigation, and debate for the better part of a century. There has been a particularly productive period recently, during which much progress has been made in better describing all aspects of collagen structure. However, there remain some questions regarding its helical symmetry and its persistence within the triple-helix. Previous considerations of this symmetry have sometimes confused the picture by not fully recognizing that collagen structure is a highly complex and large hierarchical entity, and this affects and is effected by the super-coiled molecules that make it. Nevertheless, the symmetry question is not trite, but of some significance as it relates to extracellular matrix organization and cellular integration. The correlation between helical structure in the context of the molecular packing arrangement determines which parts of the amino acid sequence of the collagen fibril are buried or accessible to the extracellular matrix or the cell. In this study, we concentrate primarily on the triple-helical structure of fibrillar collagens I and II, the two most predominant types. By comparing X-ray diffraction data collected from type I and type II containing tissues, we point to evidence for a range of triple-helical symmetries being extant in the molecules native environment. The possible significance of helical instability, local helix dissociation and molecular packing of the triple-helices is discussed in the context of collagen's supramolecular organization, all of which must affect the symmetry of the collagen triple-helix.

Self-association of collagen triple helic peptides into higher order structures.

Interest in self-association of peptides and proteins is motivated by an interest in the mechanism of physiologically higher order assembly of proteins such as collagen as well as the mechanism of pathological aggregation such as beta-amyloid formation. The triple helical form of (Pro-Hyp-Gly)(10), a peptide that has proved a useful model for molecular features of collagen, was found to self-associate, and its association properties are reported here. Turbidity experiments indicate that the triple helical peptide self-assembles at neutral pH via a nucleation-growth mechanism, with a critical concentration near 1 mM. The associated form is more stable than individual molecules by about 25 degrees C, and the association is reversible. The rate of self-association increases with temperature, supporting an entropically favored process. After self-association, (Pro-Hyp-Gly)(10) forms branched filamentous structures, in contrast with the highly ordered axially periodic structure of collagen fibrils. Yet a number of characteristics of triple helix assembly for the peptide resemble those of collagen fibril formation. These include promotion of fibril formation by neutral pH and increasing temperature; inhibition by sugars; and a requirement for hydroxyproline. It is suggested that these similar features for peptide and collagen self-association are based on common lateral underlying interactions between triple helical molecules mediated by hydrogen-bonded hydration networks involving hydroxyproline.

Prediction of collagen stability from amino acid sequence.

An algorithm was derived to relate the amino acid sequence of a collagen triple helix to its thermal stability. This calculation is based on the triple helical stabilization propensities of individual residues and their intermolecular and intramolecular interactions, as quantitated by melting temperature values of host-guest peptides. Experimental melting temperature values of a number of triple helical peptides of varying length and sequence were successfully predicted by this algorithm. However, predicted T(m) values are significantly higher than experimental values when there are strings of oppositely charged residues or concentrations of like charges near the terminus. Application of the algorithm to collagen sequences highlights regions of unusually high or low stability, and these regions often correlate with biologically significant features. The prediction of stability from sequence indicates an understanding of the major forces maintaining this protein motif. The use of highly favorable KGE and KGD sequences is seen to complement the stabilizing effects of imino acids in modulating stability and may become dominant in the collagenous domains of bacterial proteins that lack hydroxyproline. The effect of single amino acid mutations in the X and Y positions can be evaluated with this algorithm. An interactive collagen stability calculator based on this algorithm is available online.

Electrostatic interactions involving lysine make major contributions to collagen triple-helix stability.

Important stabilizing features for the collagen triple helix include the presence of Gly as every third residue, a high content of imino acids, and interchain hydrogen bonds. Host-guest peptides have been used previously to characterize triple-helix propensities of individual residues and Gly-X-Y triplets. Here, comparison of the thermal stabilities of host-guest peptides of the form (Gly-Pro-Hyp)3-Gly-X-Y-Gly-X'-Y'-(Gly-Pro-Hyp)3 extends the study to adjacent tripeptide sequences, to encompass the major classes of potential direct intramolecular interactions. Favorable hydrophobic interactions were observed, as well as stabilizing intrachain interactions between residues of opposite charge in the i and i + 3 positions. However, the greatest gain in triple-helix stability was achieved in the presence of Gly-Pro-Lys-Gly-Asp/Glu-Hyp sequences, leading to a T(m) value equal to that seen for a Gly-Pro-Hyp-Gly-Pro-Hyp sequence. This stabilization is seen for Lys but not for Arg and can be assigned to interchain ion pairs, as shown by molecular modeling. Computational analysis shows that Lys-Gly-Asp/Glu sequences are present at a frequency much greater than expected in collagen, suggesting this interaction is biologically important. These results add significantly to the understanding of which surface ion pairs can contribute to protein stability.