The Sec1/Munc18 Protein Groove Plays a Conserved Role in Interaction with Sec9p/SNAP-25.

The Sec1/Munc18 (SM) proteins constitute a conserved family with essential functions in SNARE-mediated membrane fusion. Recently, a new protein-protein interaction site in Sec1p, designated the groove, was proposed. Here, we show that a sec1 groove mutant yeast strain, sec1(w24), displays temperature-sensitive growth and secretion defects. The yeast Sec1p and mammalian Munc18-1 grooves were shown to play an important role in the interaction with the SNAREs Sec9p and SNAP-25b, respectively. Incubation of SNAP-25b with the Munc18-1 groove mutant resulted in a lag in the kinetics of SNARE complex assembly in vitro when compared with wild-type Munc18-1. The SNARE regulator SRO7 was identified as a multicopy suppressor of sec1(w24) groove mutant and an intact Sec1p groove was required for the plasma membrane targeting of Sro7p-SNARE complexes. Simultaneous inactivation of Sec1p groove and SRO7 resulted in reduced levels of exocytic SNARE complexes. Our results identify the groove as a conserved interaction surface in SM proteins. The results indicate that this structural element is important for interactions with Sec9p/SNAP-25 and participates, in concert with Sro7p, in the initial steps of SNARE complex assembly.

Discovery of O-(3-carbamimidoylphenyl)-l-serine amides as matriptase inhibitors using a fragment-linking approach.

Matriptase is a cell-surface trypsin-like serine protease of epithelial origin, which cleaves and activates proteins including hepatocyte growth factor/scatter factor and proteases such as uPA, which are involved in the progression of various cancers. Here we report a fragment-linking approach, which led to the discovery of O-(3-carbamimidoylphenyl)-l-serine amides as potent matriptase inhibitors. The co-crystal structure of one of the potent inhibitors, 6 in complex with matriptase catalytic domain validated the working hypothesis guiding the development of this congeneric series and revealed the structural basis for matriptase inhibition. Replacement of a naphthyl group in 6 with 2,4,6-tri-isopropyl phenyl resulted in 10 with improved matriptase inhibition, which exhibited significant primary tumor growth inhibition in a mouse model of prostate cancer. Compounds such as 10, identified using a fragment-linking approach, can be explored further to understand the role of matriptase as a drug target in cancer and inflammation.

Structure-guided discovery of 2-aryl/pyridin-2-yl-1H-indole derivatives as potent and selective hepsin inhibitors.

Hepsin, a type II transmembrane serine protease, is upregulated in prostate cancer and known to be involved in the progression of metastasis. Here we report a structure-guided approach, which resulted in the discovery of 2-aryl/pyridin-2-yl-1H-indole derivatives as potent and selective inhibitors of hepsin. Potent and selective inhibition of hepsin by compound 8 is likely due to interactions of the amidine group at the S1 site with the cyclohexyl ring from the 2-aryl group projecting towards the S1' site and the tert-hydroxyl group interacting with His57 side-chain as revealed by X-ray crystallography. Compounds 8 and 10, showed Ki of 0.1 µM for hepsin, and exhibited inhibition of invasion and migration of hepsin-overexpressing cell line. Compounds described here could serve as useful tool reagents to investigate the role of hepsin as a potential therapeutic target in cancer.

A vertebrate model for the study of lipid binding/transfer protein function: conservation of OSBP-related proteins between zebrafish and human.

Oxysterol-binding protein (OSBP) and OSBP-related (ORP) or OSBP-like (OSBPL) proteins constitute a family of lipid-binding/transfer proteins (LTPs) present in eukaryotes from yeast to man. The mechanisms of ORP function have remained incompletely understood. However, several ORPs are present at membrane contact sites and act as either lipid transporters or sensors that control lipid metabolism, cell signaling, and vesicle transport. Zebrafish, Danio rerio, has gained increasing popularity as a model organism in developmental biology, human disease, toxicology, and drug discovery. However, LTPs in the fish are thus far unexplored. In this article we report a series of bioinformatic analyses showing that the OSBPL gene family is highly conserved between the fish and human. The OSBPL subfamily structure is markedly similar between the two organisms, and all 12 human genes have orthologs, designated osbpl and located on 11 chromosomes in D. rerio. Interestingly, osbpl2 and osbpl3 are present as two closely related homologs (a and b), due to gene duplication events in the teleost lineage. Moreover, the domain structures of the distinct ORP proteins are almost identical between zebrafish and man, and molecular modeling in the present study suggests that ORD liganding by phosphatidylinositol-4-phosphate (PI4P) is a feature conserved between yeast Osh3p, human ORP3, and zebrafish Osbpl3. The present analysis identifies D. rerio as an attractive model to study the functions of ORPs in vertebrate development and metabolism.

Structure-guided discovery of 1,3,5 tri-substituted benzenes as potent and selective matriptase inhibitors exhibiting in vivo antitumor efficacy.

Matriptase is a serine protease implicated in cancer invasion and metastasis. Expression of matriptase is frequently dysregulated in human cancers and matriptase has been reported to activate latent growth factors such as hepatocyte growth factor/scatter factor, and proteases such as urokinase plasminogen activator suggesting that matriptase inhibitors could have therapeutic potential in treatment of cancer. Here we report a structure-based approach which led to the discovery of selective and potent matriptase inhibitors with benzene as central core having 1,3,5 tri-substitution pattern. X-ray crystallography of one of the potent analogs in complex with matriptase revealed strong hydrogen bonding and salt-bridge interactions in the S1 pocket, as well as strong CH-π contacts between the P2/P4 cyclohexyl and Trp215 side-chain. An additional interaction of the pendant amine at cyclohexyl with Gln175 side-chain results in substantial improvement in matriptase inhibition and selectivity against other related serine proteases. Compounds 15 and 26 showed tumor growth inhibition in a subcutaneous DU-145 prostate cancer mouse model. These compounds could be useful as tools to further explore the biology of matriptase as a drug target.

Visually interpretable models of kinase selectivity related features derived from field-based proteochemometrics.

Achieving selectivity for small organic molecules toward biological targets is a main focus of drug discovery but has been proven difficult, for example, for kinases because of the high similarity of their ATP binding pockets. To support the design of more selective inhibitors with fewer side effects or with altered target profiles for improved efficacy, we developed a method combining ligand- and receptor-based information. Conventional QSAR models enable one to study the interactions of multiple ligands toward a single protein target, but in order to understand the interactions between multiple ligands and multiple proteins, we have used proteochemometrics, a multivariate statistics method that aims to combine and correlate both ligand and protein descriptions with affinity to receptors. The superimposed binding sites of 50 unique kinases were described by molecular interaction fields derived from knowledge-based potentials and Schrödinger's WaterMap software. Eighty ligands were described by Mold(2), Open Babel, and Volsurf descriptors. Partial least-squares regression including cross-terms, which describe the selectivity, was used for model building. This combination of methods allows interpretation and easy visualization of the models within the context of ligand binding pockets, which can be translated readily into the design of novel inhibitors.

Discovery and Characterization of BAY-805, a Potent and Selective Inhibitor of Ubiquitin-Specific Protease USP21.

USP21 belongs to the ubiquitin-specific protease (USP) subfamily of deubiquitinating enzymes (DUBs). Due to its relevance in tumor development and growth, USP21 has been reported as a promising novel therapeutic target for cancer treatment. Herein, we present the discovery of the first highly potent and selective USP21 inhibitor. Following high-throughput screening and subsequent structure-based optimization, we identified BAY-805 to be a non-covalent inhibitor with low nanomolar affinity for USP21 and high selectivity over other DUB targets as well as kinases, proteases, and other common off-targets. Furthermore, surface plasmon resonance (SPR) and cellular thermal shift assays (CETSA) demonstrated high-affinity target engagement of BAY-805, resulting in strong NF-κB activation in a cell-based reporter assay. To the best of our knowledge, BAY-805 is the first potent and selective USP21 inhibitor and represents a valuable high-quality in vitro chemical probe to further explore the complex biology of USP21.

Sterol binding by OSBP-related protein 1L regulates late endosome motility and function.

ORP1L is an oxysterol binding homologue that regulates late endosome (LE) positioning. We show that ORP1L binds several oxysterols and cholesterol, and characterize a mutant, ORP1L Δ560-563, defective in oxysterol binding. While wild-type ORP1L clusters LE, ORP1L Δ560-563 induces LE scattering, which is reversed by disruption of the endoplasmic reticulum (ER) targeting FFAT motif, suggesting that it is due to enhanced LE-ER interactions. Endosome motility is reduced upon overexpression of ORP1L. Both wild-type ORP1L and the Δ560-563 mutant induce the recruitment of both dynactin and kinesin-2 on LE. Most of the LE decorated by overexpressed ORP1L fail to accept endocytosed dextran or EGF, and the transfected cells display defective degradation of internalized EGF. ORP1L silencing in macrophage foam cells enhances endosome motility and results in inhibition of [(3)H]cholesterol efflux to apolipoprotein A-I. These data demonstrate that LE motility and functions in both protein and lipid transport are regulated by ORP1L.

First-in-Class Small Molecule to Inhibit CYP11A1 and Steroid Hormone Biosynthesis.

Binding of steroid hormones to their cognate receptors regulates the growth of most prostate and breast cancers. We hypothesized that CYP11A inhibition might halt the synthesis of all steroid hormones, because CYP11A is the only enzyme that catalyses the first step of steroid hormone biosynthesis. We speculated that a CYP11A inhibitor could be administered safely provided that the steroids essential for life are replaced. Virtual screening and systematic structure-activity relationship optimization were used to develop ODM-208, the first-in-class, selective, nonsteroidal, oral CYP11A1 inhibitor. Safety of ODM-208 was assessed in rats and Beagle dogs, and efficacy in a VCaP castration-resistant prostate cancer (CRPC) xenograft mouse model, in mice and dogs, and in six patients with metastatic CRPC. Blood steroid hormone concentrations were measured using liquid chromatography-mass spectrometry. ODM-208 binds to CYP11A1 and inhibited its enzymatic activity. ODM-208 administration led to rapid, complete, durable, and reversible inhibition of the steroid hormone biosynthesis in an adrenocortical carcinoma cell model in vitro, in adult noncastrated male mice and dogs, and in patients with CRPC. All measured serum steroid hormone concentrations reached undetectable levels within a few weeks from the start of ODM-208 administration. ODM-208 was well tolerated with steroid hormone replacement. The toxicity findings were considered related to CYP11A1 inhibition and were reversed after stopping of the compound administration. Steroid hormone biosynthesis can be effectively inhibited with a small-molecule inhibitor of CYP11A1. The findings suggest that administration of ODM-208 is feasible with concomitant corticosteroid replacement therapy.

Discovery and characterization of ORM-11372, a novel inhibitor of the sodium-calcium exchanger with positive inotropic activity.

BACKGROUND AND PURPOSE: The lack of selective sodium-calcium exchanger (NCX) inhibitors has hampered the exploration of physiological and pathophysiological roles of cardiac NCX 1.1. We aimed to discover more potent and selective drug like NCX 1.1 inhibitor. EXPERIMENTAL APPROACH: A flavan series-based pharmacophore model was constructed. Virtual screening helped us identify a novel scaffold for NCX inhibition. A distinctively different NCX 1.1 inhibitor, ORM-11372, was discovered after lead optimization. Its potency against human and rat NCX 1.1 and selectivity against other ion channels was assessed. The cardiovascular effects of ORM-11372 were studied in normal and infarcted rats and rabbits. Human cardiac safety was studied ex vivo using human ventricular trabeculae. KEY RESULTS: ORM-11372 inhibited human NCX 1.1 reverse and forward currents; IC50 values were 5 and 6 nM respectively. ORM-11372 inhibited human cardiac sodium 1.5 (INa ) and hERG KV 11.1 currents (IhERG ) in a concentration-dependent manner; IC50 values were 23.2 and 10.0 µM. ORM-11372 caused no changes in action potential duration; short-term variability and triangulation were observed for concentrations of up to 10 µM. ORM-11372 induced positive inotropic effects of 18 ± 6% and 35 ± 8% in anaesthetized rats with myocardial infarctions and in healthy rabbits respectively; no other haemodynamic effects were observed, except improved relaxation at the lowest dose. CONCLUSION AND IMPLICATIONS: ORM-11372, a unique, novel, and potent inhibitor of human and rat NCX 1.1, is a positive inotropic compound. NCX inhibition can induce clinically relevant improvements in left ventricular contractions without affecting relaxation, heart rate, or BP, without pro-arrhythmic risk.

Discovery and development of ODM-204: A Novel nonsteroidal compound for the treatment of castration-resistant prostate cancer by blocking the androgen receptor and inhibiting CYP17A1.

We report the discovery of a novel nonsteroidal dual-action compound, ODM-204, that holds promise for treating patients with castration-resistant prostate cancer (CRPC), an advanced form of prostate cancer characterised by high androgen receptor (AR) expression and persistent activation of the AR signaling axis by residual tissue androgens. For ODM-204, has a dual mechanism of action. The compound is anticipated to efficiently dampen androgenic stimuli in the body by inhibiting CYP17A1, the prerequisite enzyme for the formation of dihydrotestosterone (DHT) and testosterone (T), and by blocking AR with high affinity and specificity. In our study, ODM-204 inhibited the proliferation of androgen-dependent VCaP and LNCaP cells in vitro and reduced significantly tumour growth in a murine VCaP xenograft model in vivo. Intriguingly, after a single oral dose of 10-30 mg/kg, ODM-204 dose-dependently inhibited adrenal and testicular steroid production in sexually mature male cynomolgus monkeys. Similar results were obtained in human chorionic gonadotropin-treated male rats. In rats, leuprolide acetate-mediated (LHRH agonist) suppression of the circulating testosterone levels and decrease in weights of androgen-sensitive organs was significantly and dose-dependently potentiated by the co-administration of ODM-204. ODM-204 was well tolerated in both rodents and primates. Based on our data, ODM-204 could provide an effective therapeutic option for men with CRPC.

ODM-203, a Selective Inhibitor of FGFR and VEGFR, Shows Strong Antitumor Activity, and Induces Antitumor Immunity.

Alterations in the gene encoding for the FGFR and upregulation of the VEGFR are found often in cancer, which correlate with disease progression and unfavorable survival. In addition, FGFR and VEGFR signaling synergistically promote tumor angiogenesis, and activation of FGFR signaling has been described as functional compensatory angiogenic signal following development of resistance to VEGFR inhibition. Several selective small-molecule FGFR kinase inhibitors are currently in clinical development. ODM-203 is a novel, selective, and equipotent inhibitor of the FGFR and VEGFR families. In this report we show that ODM-203 inhibits FGFR and VEGFR family kinases selectively and with equal potency in the low nanomolar range (IC50 6-35 nmol/L) in biochemical assays. In cellular assays, ODM-203 inhibits VEGFR-induced tube formation (IC50 33 nmol/L) with similar potency as it inhibits proliferation in FGFR-dependent cell lines (IC50 50-150 nmol/L). In vivo, ODM-203 shows strong antitumor activity in both FGFR-dependent xenograft models and in an angiogenic xenograft model at similar well-tolerated doses. In addition, ODM-203 inhibits metastatic tumor growth in a highly angiogenesis-dependent kidney capsule syngenic model. Interestingly, potent antitumor activity in the subcutaneous syngenic model correlated well with immune modulation in the tumor microenvironment as indicated by marked decrease in the expression of immune check points PD-1 and PD-L1 on CD8 T cells and NK cells, and increased activation of CD8 T cells. In summary, ODM-203 shows equipotent activity for both FGFR and VEGFR kinase families and antitumor activity in both FGFR and angigogenesis models.

The mammalian oxysterol-binding protein-related proteins (ORPs) bind 25-hydroxycholesterol in an evolutionarily conserved pocket.

OSBP (oxysterol-binding protein) homologues, ORPs (OSBP-related proteins), constitute a 12-member family in mammals. We employed an in vitro [3H]25OH (25-hydroxycholesterol)-binding assay with purified recombinant proteins as well as live cell photo-cross-linking with [3H]photo-25OH and [3H]photoCH (photo-cholesterol), to investigate sterol binding by the mammalian ORPs. ORP1 and ORP2 [a short ORP consisting of an ORD (OSBP-related ligand-binding domain) only] were in vitro shown to bind 25OH. GST (glutathione S-transferase) fusions of the ORP1L [long variant with an N-terminal extension that carries ankyrin repeats and a PH domain (pleckstrin homology domain)] and ORP1S (short variant consisting of an ORD only) variants bound 25OH with similar affinity (ORP1L, K(d)=9.7x10(-8) M; ORP1S, K(d)=8.4 x10(-8) M), while the affinity of GST-ORP2 for 25OH was lower (K(d)=3.9x10(-6) M). Molecular modelling suggested that ORP2 has a sterol-binding pocket similar to that of Saccharomyces cerevisiae Osh4p. This was confirmed by site-directed mutagenesis of residues in proximity of the bound sterol in the structural model. Substitution of Ile249 by tryptophan or Lys150 by alanine markedly inhibited 25OH binding by ORP2. In agreement with the in vitro data, ORP1L, ORP1S, and ORP2 were cross-linked with photo-25OH in live COS7 cells. Furthermore, in experiments with either truncated cDNAs encoding the OSBP-related ligand-binding domains of the ORPs or the full-length proteins, photo-25OH was bound to OSBP, ORP3, ORP4, ORP5, ORP6, ORP7, ORP8, ORP10 and ORP11. In addition, the ORP1L variant and ORP3, ORP5, and ORP8 were cross-linked with photoCH. The present study identifies ORP1 and ORP2 as OSBPs and suggests that most of the mammalian ORPs are able to bind sterols.

3D proteochemometrics: using three-dimensional information of proteins and ligands to address aspects of the selectivity of serine proteases.

The high similarity between certain sub-pockets of serine proteases may lead to low selectivity of protease inhibitors. Therefore the application of proteochemometrics (PCM), which quantifies the relationship between protein/ligand descriptors and affinity for multiple ligands and targets simultaneously, is useful to understand and improve the selectivity profiles of potential inhibitors. In this study, protein field-based PCM that uses knowledge-based and WaterMap derived fields to describe proteins in combination with 2D (RDKit and MOE fingerprints) and 3D (4 point pharmacophoric fingerprints and GRIND) ligand descriptors was used to model the bioactivities of 24 homologous serine proteases and 5863 inhibitors in an integrated fashion. Of the multiple field-based PCM models generated based on different ligand descriptors, RDKit fingerprints showed the best performance in terms of external prediction with Rtest2 of 0.72 and RMSEP of 0.81. Further, visual interpretation of the models highlights sub-pocket specific regions that influence affinity and selectivity of serine protease inhibitors.

Classification and comparison of ligand-binding sites derived from grid-mapped knowledge-based potentials.

We describe the application of knowledge-based potentials implemented in the MOE program to compare the ligand-binding sites of several proteins. The binding probabilities for a polar and a hydrophobic probe are calculated on a grid to allow easy comparison of binding sites of superimposed related proteins. The method is fast and simple enough to simultaneously use structural information of multiple proteins of a target family. The method can be used to rapidly cluster proteins into subfamilies according to the similarity of hydrophobic and polar fields of their ligand-binding sites. Regions of the binding site which are common within a protein family can be identified and analysed for the design of family-targeted libraries or those which differ for improvement of ligand selectivity. The field-based hierarchical clustering is demonstrated for three protein families: the ligand-binding domains of nuclear receptors, the ATP-binding sites of protein kinases and the substrate binding sites of proteases. More detailed comparisons are presented for serine proteases of the chymotrypsin family, for the peroxisome proliferator-activated receptor subfamily of nuclear receptors and for progesterone and androgen receptor. The results are in good accordance with structure-based analysis and highlight important differences of the binding sites, which have been also described in the literature.

Analysis of pH-dependent elements in proteins: geometry and properties of pairs of hydrogen-bonded carboxylic acid side-chains.

A rather frequent but so far little discussed observation is that pairs of carboxylic acid side-chains in proteins can share a proton in a hydrogen bond. In the present article, quantum chemical calculations of simple model systems for carboxyl-carboxylate interactions are compared with structural observations from proteins. A detailed structural analysis of the proteins deposited in the PDB revealed that, in a subset of proteins sharing less than 90% sequence identity, 19% (314) contain at least one pair of carboxylic acids with their side-chain oxygen atoms within hydrogen-bonding distance. As the distance between those interacting oxygen atoms is frequently very short ( approximately 2.55 A), many of these carboxylic acids are suggested to share a proton in a strong hydrogen bond. When situated in an appropriate structural environment (low dielectric constant), some might even form a low barrier hydrogen bond. The quantum chemical studies show that the most frequent geometric features of carboxyl-carboxylate pairs found in proteins, and no or symmetric ligation, are also the most stable arrangements at low dielectric constants, and they also suggest at medium and low pH a higher stability than for isosteric amide-carboxylate pairs. The presence of these pairs in 119 different enzymes found in the BRENDA database is set in relation to their properties and functions. This analysis shows that pH optima of enzymes with carboxyl-carboxylate pairs are shifted to lower than average values, whereas temperature optima seem to be increased. The described structural principles can be used as guidelines for rational protein design (e.g., in order to improve pH or temperature stability).

Nuclear Receptor-Like Structure and Interaction of Congenital Heart Disease-Associated Factors GATA4 and NKX2-5.

AIMS: Transcription factor GATA4 is a dosage sensitive regulator of heart development and alterations in its level or activity lead to congenital heart disease (CHD). GATA4 has also been implicated in cardiac regeneration and repair. GATA4 action involves combinatorial interaction with other cofactors such as NKX2-5, another critical cardiac regulator whose mutations also cause CHD. Despite its critical importance to the heart and its evolutionary conservation across species, the structural basis of the GATA4-NKX2-5 interaction remains incompletely understood. METHODS AND RESULTS: A homology model was constructed and used to identify surface amino acids important for the interaction of GATA4 and NKX2-5. These residues were subjected to site-directed mutagenesis, and the mutant proteins were characterized for their ability to bind DNA and to physically and functionally interact with NKX2-5. The studies identify 5 highly conserved amino acids in the second zinc finger (N272, R283, Q274, K299) and its C-terminal extension (R319) that are critical for physical and functional interaction with the third alpha helix of NKX2-5 homeodomain. Integration of the experimental data with computational modeling suggests that the structural arrangement of the zinc finger-homeodomain resembles the architecture of the conserved DNA binding domain of nuclear receptors. CONCLUSIONS: The results provide novel insight into the structural basis for protein-protein interactions between two important classes of transcription factors. The model proposed will help to elucidate the molecular basis for disease causing mutations in GATA4 and NKX2-5 and may be relevant to other members of the GATA and NK classes of transcription factors.

Discovery of ODM-201, a new-generation androgen receptor inhibitor targeting resistance mechanisms to androgen signaling-directed prostate cancer therapies.

Activation of androgen receptor (AR) is crucial for prostate cancer growth. Remarkably, also castration-resistant prostate cancer (CRPC) is dependent on functional AR, and several mechanisms have been proposed to explain the addiction. Known causes of CRPC include gene amplification and overexpression as well as point mutations of AR. We report here the pharmacological profile of ODM-201, a novel AR inhibitor that showed significant antitumor activity and a favorable safety profile in phase 1/2 studies in men with CRPC. ODM-201 is a full and high-affinity AR antagonist that, similar to second-generation antiandrogens enzalutamide and ARN-509, inhibits testosterone-induced nuclear translocation of AR. Importantly, ODM-201 also blocks the activity of the tested mutant ARs arising in response to antiandrogen therapies, including the F876L mutation that confers resistance to enzalutamide and ARN-509. In addition, ODM-201 reduces the growth of AR-overexpressing VCaP prostate cancer cells both in vitro and in a castration-resistant VCaP xenograft model. In contrast to other antiandrogens, ODM-201 shows negligible brain penetrance and does not increase serum testosterone levels in mice. In conclusion, ODM-201 is a potent AR inhibitor that overcomes resistance to AR-targeted therapies by antagonizing both overexpressed and mutated ARs. ODM-201 is currently in a phase 3 trial in CRPC.

Efficient methods for filtering and ranking fragments for the prediction of structurally variable regions in proteins.

The prediction of protein 3D structures close to insertions and deletions or, more generally, loop prediction, is still one of the major challenges in homology modeling projects. In this article, we developed ranking criteria and selection filters to improve knowledge-based loop predictions. These criteria were developed and optimized for a test data set containing 678 insertions and deletions. The examples are, in principle, predictable from the used loop database with an RMSD < 1 A and represent realistic modeling situations. Four noncorrelated criteria for the selection of fragments are evaluated. A fast prefilter compares the distance between the anchor groups in the template protein with the stems of the fragments. The RMSD of the anchor groups is used for fitting and ranking of the selected loop candidates. After fitting, repulsive close contacts of loop candidates with the template protein are used for filtering, and fragments with backbone torsion angles, which are unfavorable according to a knowledge-based potential, are eliminated. By the combined application of these filter criteria to the test set, it was possible to increase the percentage of predictions with a global RMSD < 1 A to over 50% among the first five ranks, with average global RMSD values for the first rank candidate that are between 1.3 and 2.2 A for different loop lengths. Compared to other examples described in the literature, our large numbers of test cases are not self-predictions, where loops are placed in a protein after a peptide loop has been cut out, but are attempts to predict structural changes that occur in evolution when a protein is affected by insertions and deletions.

Positioning of anchor groups in protein loop prediction: the importance of solvent accessibility and secondary structure elements.

The prediction of loop regions in the process of protein structure prediction by homology is still an unsolved problem. In an earlier publication, we could show that the correct placement of the amino acids serving as an anchor group to be connected by a loop fragment with a predicted geometry is a highly important step and an essential requirement within the process (Lessel and Schomburg, Proteins 1999; 37:56-64). In this article, we present an analysis of the quality of possible loop predictions with respect to gap length, fragment length, amino acid type, secondary structure, and solvent accessibility. For 550 insertions and 544 deletions, we test all possible positions for anchor groups with an inserted loop of a length between 3 and 12 amino acids. We could show that approximately 80% of the indel regions could be predicted within 1.5 A RMSD from a knowledge-based loop data base if criteria for the correct localization of anchor groups could be found and the loops can be sorted correctly. From our analysis, several conclusions regarding the optimal placement of anchor groups become obvious: (1) The correct placement of anchor groups is even more important for longer gap lengths, (2) medium length fragments (length 5-8) perform better than short or long ones, (3) the placement of anchor groups at hydrophobic amino acids gives a higher chance to include the best possible loop, (4) anchor groups within secondary structure elements, in particular beta-sheets are suitable, (5) amino acids with lower solvent accessibility are better anchor group. A preliminary test using a combination of the anchor group positioning criteria deduced from our analysis shows very promising results.