CAFASP-1: Critical assessment of fully automated structure prediction methods

Complementarity of the residue-level protein function and structure predictions in human proteins

Sequence-based predictors of the residue-level protein function and structure cover a broad spectrum of characteristics including intrinsic disorder, secondary structure, solvent accessibility and binding to nucleic acids. They were catalogued and evaluated in numerous surveys and assessments. However, methods focusing on a given characteristic are studied separately from predictors of other characteristics, while they are typically used on the same proteins. We fill this void by studying complementarity of a representative collection of methods that target different predictions using a large, taxonomically consistent, and low similarity dataset of human proteins. First, we bridge the gap between the communities that develop structure-trained vs. disorder-trained predictors of binding residues. Motivated by a recent study of the protein-binding residue predictions, we empirically find that combining the structure-trained and disorder-trained predictors of the DNA-binding and RNA-binding residues leads to substantial improvements in predictive quality. Second, we investigate whether diverse predictors generate results that accurately reproduce relations between secondary structure, solvent accessibility, interaction sites, and intrinsic disorder that are present in the experimental data. Our empirical analysis concludes that predictions accurately reflect all combinations of these relations. Altogether, this study provides unique insights that support combining results produced by diverse residue-level predictors of protein function and structure.

Computational Prediction of Secondary and Supersecondary Structures from Protein Sequences

Many new methods for the sequence-based prediction of the secondary and supersecondary structures have been developed over the last several years. These and older sequence-based predictors are widely applied for the characterization and prediction of protein structure and function. These efforts have produced countless accurate predictors, many of which rely on state-of-the-art machine learning models and evolutionary information generated from multiple sequence alignments. We describe and motivate both types of predictions. We introduce concepts related to the annotation and computational prediction of the three-state and eight-state secondary structure as well as several types of supersecondary structures, such as β hairpins, coiled coils, and α-turn-α motifs. We review 34 predictors focusing on recent tools and provide detailed information for a selected set of 14 secondary structure and 3 supersecondary structure predictors. We conclude with several practical notes for the end users of these predictive methods.

Determining protein similarity by comparing hydrophobic core structure

Formal assessment of structural similarity is - next to protein structure prediction - arguably the most important unsolved problem in proteomics. In this paper we propose a similarity criterion based on commonalities between the proteins' hydrophobic cores. The hydrophobic core emerges as a result of conformational changes through which each residue reaches its intended position in the protein body. A quantitative criterion based on this phenomenon has been proposed in the framework of the CASP challenge. The structure of the hydrophobic core - including the placement and scope of any deviations from the idealized model - may indirectly point to areas of importance from the point of view of the protein's biological function. Our analysis focuses on an arbitrarily selected target from the CASP11 challenge. The proposed measure, while compliant with CASP criteria (70-80% correlation), involves certain adjustments which acknowledge the presence of factors other than simple spatial arrangement of solids.

Protein inter-domain linker prediction using Random Forest and amino acid physiochemical properties

BACKGROUND

Protein chains are generally long and consist of multiple domains. Domains are distinct structural units of a protein that can evolve and function independently. The accurate prediction of protein domain linkers and boundaries is often regarded as the initial step of protein tertiary structure and function predictions. Such information not only enhances protein-targeted drug development but also reduces the experimental cost of protein analysis by allowing researchers to work on a set of smaller and independent units. In this study, we propose a novel and accurate domain-linker prediction approach based on protein primary structure information only. We utilize a nature-inspired machine-learning model called Random Forest along with a novel domain-linker profile that contains physiochemical and domain-linker information of amino acid sequences.

RESULTS

The proposed approach was tested on two well-known benchmark protein datasets and achieved 68% sensitivity and 99% precision, which is better than any existing protein domain-linker predictor. Without applying any data balancing technique such as class weighting and data re-sampling, the proposed approach is able to accurately classify inter-domain linkers from highly imbalanced datasets.

CONCLUSION

Our experimental results prove that the proposed approach is useful for domain-linker identification in highly imbalanced single- and multi-domain proteins.

A dynamic model of the proteins that form the initial iron-sulfur cluster biogenesis machinery in yeast mitochondria

The assembly of iron-sulfur clusters (ISCs) in eukaryotes involves the protein Frataxin. Deficits in this protein have been associated with iron inside the mitochondria and impair ISC biogenesis as it is postulated to act as the iron donor for ISCs assembly in this organelle. A pronounced lack of Frataxin causes Friedreich's Ataxia, which is a human neurodegenerative and hereditary disease mainly affecting the equilibrium, coordination, muscles and heart. Moreover, it is the most common autosomal recessive ataxia. High similarities between the human and yeast molecular mechanisms that involve Frataxin have been suggested making yeast a good model to study that process. In yeast, the protein complex that forms the central assembly platform for the initial step of ISC biogenesis is composed by yeast frataxin homolog, Nfs1-Isd11 and Isu. In general, it is commonly accepted that protein function involves interaction with other protein partners, but in this case not enough is known about the structure of the protein complex and, therefore, how it exactly functions. The objective of this work is to model the protein complex in order to gain insight into structural details that end up with its biological function. To achieve this goal several bioinformatics tools, modeling techniques and protein docking programs have been used. As a result, the structure of the protein complex and the dynamic behavior of its components, along with that of the iron and sulfur atoms required for the ISC assembly, have been modeled. This hypothesis will help to better understand the function and molecular properties of Frataxin as well as those of its ISC assembly protein partners.

Computational prediction of secondary and supersecondary structures

The sequence-based prediction of the secondary and supersecondary structures enjoys strong interest and finds applications in numerous areas related to the characterization and prediction of protein structure and function. Substantial efforts in these areas over the last three decades resulted in the development of accurate predictors, which take advantage of modern machine learning models and availability of evolutionary information extracted from multiple sequence alignment. In this chapter, we first introduce and motivate both prediction areas and introduce basic concepts related to the annotation and prediction of the secondary and supersecondary structures, focusing on the β hairpin, coiled coil, and α-turn-α motifs. Next, we overview state-of-the-art prediction methods, and we provide details for 12 modern secondary structure predictors and 4 representative supersecondary structure predictors. Finally, we provide several practical notes for the users of these prediction tools.

A threonine to isoleucine missense mutation in the pericentriolar material 1 gene is strongly associated with schizophrenia

Markers at the pericentriolar material 1 gene (PCM1) have shown genetic association with schizophrenia in both a University College London (UCL) and a USA-based case-control sample. In this paper we report a statistically significant replication of the PCM1 association in a large Scottish case-control sample from Aberdeen. Resequencing of the genomic DNA from research volunteers who had inherited haplotypes associated with schizophrenia showed a threonine to isoleucine missense mutation in exon 24 which was likely to change the structure and function of PCM1 (rs370429). This mutation was found only as a heterozygote in 98 schizophrenic research subjects and controls out of 2246 case and control research subjects. Among the 98 carriers of rs370429, 67 were affected with schizophrenia. The same alleles and haplotypes were associated with schizophrenia in both the London and Aberdeen samples. Another potential aetiological base pair change in PCM1 was rs445422, which altered a splice site signal. A further mutation, rs208747, was shown by electrophoretic mobility shift assays to create or destroy a promoter transcription factor site. Five further non-synonymous changes in exons were also found. Genotyping of the new variants discovered in the UCL case-control sample strengthened the evidence for allelic and haplotypic association (P=0.02-0.0002). Given the number and identity of the haplotypes associated with schizophrenia, further aetiological base pair changes must exist within and around the PCM1 gene. PCM1 protein has been shown to interact directly with the disrupted-in-schizophrenia 1 (DISC1) protein, Bardet-Biedl syndrome 4, and Huntingtin-associated protein 1, and is important in neuronal cell growth. In a separate study we found that clozapine but not haloperidol downregulated PCM1 expression in the mouse brain. We hypothesize that mutant PCM1 may be responsible for causing a subtype of schizophrenia through abnormal cell division and abnormal regeneration in dividing cells in the central nervous system. This is supported by our previous finding of orbitofrontal volumetric deficits in PCM1-associated schizophrenia patients as opposed to temporal pole deficits in non-PCM1-associated schizophrenia patients. Caution needs to be exercised in interpreting the actual biological effects of the mutations we have found without further cell biology. However, the DNA changes we have found deserve widespread genotyping in multiple case-control populations.

DescFold: a web server for protein fold recognition

BACKGROUND

Machine learning-based methods have been proven to be powerful in developing new fold recognition tools. In our previous work [Zhang, Kochhar and Grigorov (2005) Protein Science, 14: 431-444], a machine learning-based method called DescFold was established by using Support Vector Machines (SVMs) to combine the following four descriptors: a profile-sequence-alignment-based descriptor using Psi-blast e-values and bit scores, a sequence-profile-alignment-based descriptor using Rps-blast e-values and bit scores, a descriptor based on secondary structure element alignment (SSEA), and a descriptor based on the occurrence of PROSITE functional motifs. In this work, we focus on the improvement of DescFold by incorporating more powerful descriptors and setting up a user-friendly web server.

RESULTS

In seeking more powerful descriptors, the profile-profile alignment score generated from the COMPASS algorithm was first considered as a new descriptor (i.e., PPA). When considering a profile-profile alignment between two proteins in the context of fold recognition, one protein is regarded as a template (i.e., its 3D structure is known). Instead of a sequence profile derived from a Psi-blast search, a structure-seeded profile for the template protein was generated by searching its structural neighbors with the assistance of the TM-align structural alignment algorithm. Moreover, the COMPASS algorithm was used again to derive a profile-structural-profile-alignment-based descriptor (i.e., PSPA). We trained and tested the new DescFold in a total of 1,835 highly diverse proteins extracted from the SCOP 1.73 version. When the PPA and PSPA descriptors were introduced, the new DescFold boosts the performance of fold recognition substantially. Using the SCOP_1.73_40% dataset as the fold library, the DescFold web server based on the trained SVM models was further constructed. To provide a large-scale test for the new DescFold, a stringent test set of 1,866 proteins were selected from the SCOP 1.75 version. At a less than 5% false positive rate control, the new DescFold is able to correctly recognize structural homologs at the fold level for nearly 46% test proteins. Additionally, we also benchmarked the DescFold method against several well-established fold recognition algorithms through the LiveBench targets and Lindahl dataset.

CONCLUSIONS

The new DescFold method was intensively benchmarked to have very competitive performance compared with some well-established fold recognition methods, suggesting that it can serve as a useful tool to assist in template-based protein structure prediction. The DescFold server is freely accessible at http://202.112.170.199/DescFold/index.html.

Insights into the evolutionary origins of clostridial neurotoxins from analysis of the Clostridium botulinum strain A neurotoxin gene cluster

BACKGROUND

Clostridial neurotoxins (CNTs) are the most deadly toxins known and causal agents of botulism and tetanus neuroparalytic diseases. Despite considerable progress in understanding CNT structure and function, the evolutionary origins of CNTs remain a mystery as they are unique to Clostridium and possess a sequence and structural architecture distinct from other protein families. Uncovering the origins of CNTs would be a significant contribution to our understanding of how pathogens evolve and generate novel toxin families.

RESULTS

The C. botulinum strain A genome was examined for potential homologues of CNTs. A key link was identified between the neurotoxin and the flagellin gene (CBO0798) located immediately upstream of the BoNT/A neurotoxin gene cluster. This flagellin sequence displayed the strongest sequence similarity to the neurotoxin and NTNH homologue out of all proteins encoded within C. botulinum strain A. The CBO0798 gene contains a unique hypervariable region, which in closely related flagellins encodes a collagenase-like domain. Remarkably, these collagenase-containing flagellins were found to possess the characteristic HEXXH zinc-protease motif responsible for the neurotoxin's endopeptidase activity. Additional links to collagenase-related sequences and functions were detected by further analysis of CNTs and surrounding genes, including sequence similarities to collagen-adhesion domains and collagenases. Furthermore, the neurotoxin's HCRn domain was found to exhibit both structural and sequence similarity to eukaryotic collagen jelly-roll domains.

CONCLUSION

Multiple lines of evidence suggest that the neurotoxin and adjacent genes evolved from an ancestral collagenase-like gene cluster, linking CNTs to another major family of clostridial proteolytic toxins. Duplication, reshuffling and assembly of neighboring genes within the BoNT/A neurotoxin gene cluster may have lead to the neurotoxin's unique architecture. This work provides new insights into the evolution of C. botulinum neurotoxins and the evolutionary mechanisms underlying the origins of virulent genes.

An information measure of the quality of protein secondary structure prediction

We describe an information-theory-based measure of the quality of secondary structure prediction (RELINFO). RELINFO has a simple yet intuitive interpretation: it represents the factor by which secondary structure choice at a residue has been restricted by a prediction scheme. As an alternative interpretation of secondary structure prediction, RELINFO complements currently used methods by providing an information-based view as to why a prediction succeeds and fails. To demonstrate this score's capabilities, we applied RELINFO to an analysis of a large set of secondary structure predictions obtained from the first five rounds of the Critical Assessment of Structure Prediction (CASP) experiment. RELINFO is compared with two other common measures: percent correct (Q3) and secondary structure overlap (SOV). While the correlation between Q3 and RELINFO is approximately 0.85, RELINFO avoids certain disadvantages of Q3, including overestimating the quality of a prediction. The correlation between SOV and RELINFO is approximately 0.75. The valuable SOV measure unfortunately suffers from a saturation problem, and perhaps has unfairly given the general impression that secondary structure prediction has reached its limit since SOV hasn't improved much over the recent rounds of CASP. Although not a replacement for SOV, RELINFO has greater dispersion. Over the five rounds of CASP assessed here, RELINFO shows that predictions targets have been more difficult in successive CASP experiments, yet the predictions quality has continued to improve measurably over each round. In terms of information, the secondary structure prediction quality has almost doubled from CASP1 to CASP5. Therefore, as a different perspective of accuracy, RELINFO can help to improve prediction of protein secondary structure by providing a measure of difficulty as well as final quality of a prediction.

Automated server predictions in CASP7

With each round of CASP (Critical Assessment of Techniques for Protein Structure Prediction), automated prediction servers have played an increasingly important role. Today, most protein structure prediction approaches in some way depend on automated methods for fold recognition or model building. The accuracy of server predictions has significantly increased over the last years, and, in CASP7, we observed a continuation of this trend. In the template-based modeling category, the best prediction server was ranked third overall, i.e. it outperformed all but two of the human participating groups. This server also ranked among the very best predictors in the free modeling category as well, being clearly beaten by only one human group. In the high accuracy (HA) subset of TBM, two of the top five groups were servers. This article summarizes the contribution of automated structure prediction servers in the CASP7 experiment, with emphasis on 3D structure prediction, as well as information on their prediction scope and public availability.

Aligning sequences to structures

Most newly sequenced proteins are likely to adopt a similar structure to one which has already been experimentally determined. For this reason, the most successful approaches to protein structure prediction have been template-based methods. Such prediction methods attempt to identify and model the folds of unknown structures by aligning the target sequences to a set of representative template structures within a fold library. In this chapter, I discuss the development of template-based approaches to fold prediction, from the traditional techniques to the recent state-of-the-art methods. I also discuss the recent development of structural annotation databases, which contain models built by aligning the sequences from entire proteomes against known structures. Finally, I run through a practical step-by-step guide for aligning target sequences to known structures and contemplate the future direction of template-based structure prediction.

A historical perspective of template-based protein structure prediction

This chapter presents a broad and a historical overview of the problem of protein structure prediction. Different structure prediction methods, including homology modeling, fold recognition (FR)/protein threading, ab initio/de novo approaches, and hybrid techniques involving multiple types of approaches, are introduced in a historical context. The progress of the field as a whole, especially in the threading/FR area, as reflected by the CASP/CAFASP contests, is reviewed. At the end of the chapter, we discuss the challenging issues ahead in the field of protein structure prediction.

Structure determinants and substrate recognition of serine carboxypeptidase-like acyltransferases from plant secondary metabolism

Structures of the serine carboxypeptidase-like enzymes 1-O-sinapoyl-beta-glucose:L-malate sinapoyltransferase (SMT) and 1-O-sinapoyl-beta-glucose:choline sinapoyltransferase (SCT) were modeled to gain insight into determinants of specificity and substrate recognition. The structures reveal the alpha/beta-hydrolase fold as scaffold for the catalytic triad Ser-His-Asp. The recombinant mutants of SMT Ser173Ala and His411Ala were inactive, whereas Asp358Ala displayed residual activity of 20%. 1-O-sinapoyl-beta-glucose recognition is mediated by a network of hydrogen bonds. The glucose moiety is recognized by a hydrogen bond network including Trp71, Asn73, Glu87 and Asp172. The conserved Asp172 at the sequence position preceding the catalytic serine meets sterical requirements for the glucose moiety. The mutant Asn73Ala with a residual activity of 13% underscores the importance of the intact hydrogen bond network. Arg322 is of key importance by hydrogen bonding of 1-O-sinapoyl-beta-glucose and L-malate. By conformational change, Arg322 transfers L-malate to a position favoring its activation by His411. Accordingly, the mutant Arg322Glu showed 1% residual activity. Glu215 and Arg219 establish hydrogen bonds with the sinapoyl moiety. The backbone amide hydrogens of Gly75 and Tyr174 were shown to form the oxyanion hole, stabilizing the transition state. SCT reveals also the catalytic triad and a hydrogen bond network for 1-O-sinapoyl-beta-glucose recognition, but Glu274, Glu447, Thr445 and Cys281 are crucial for positioning of choline.

Global Effects of Mistranslation from an Editing Defect in Mammalian Cells

Aminoacyl-tRNA synthetases prevent mistranslation, or genetic code ambiguity, through specialized editing reactions. Mutations that disrupt editing in bacteria adversely affect cell growth and viability, and recent work in the mouse supports the idea that translational errors caused by an editing defect lead to a neurological disease-like phenotype. To further investigate the connection of mistranslation to cell pathology, we introduced an inducible transgene expressing an editing-deficient valyl-tRNA synthetase into mammalian cells. Introducing mistranslation precipitated a disruption of cell morphology and membrane blebbing, accompanied by activation of caspase-3, consistent with an apoptotic response. Addition of a noncanonical amino acid that is misactivated, but not cleared, by the editing-defective enzyme exacerbated these effects. A special ambiguity-detecting sensor provided direct readout of mistranslation in vivo, supporting the possibility that decreased translational fidelity could be associated with disease.

Evidence for a protein-protein interaction motif on an acyl carrier protein domain from a modular polyketide synthase

During biosynthesis on modular polyketide synthases (PKSs), chain extension intermediates are tethered to acyl carrier protein (ACP) domains through phosphopantetheinyl prosthetic groups. Each ACP must therefore interact with every other domain within the module, and also with a downstream acceptor domain. The nature of these interactions is key to our understanding of the topology and operation of these multienzymes. Sequence analysis and homology modeling implicates a potential helical region (helix II) on the ACPs as a protein-protein interaction motif. Using site-directed mutagenesis, we show that residues along this putative helix lie at the interface between the ACP and the phosphopantetheinyl transferase that catalyzes its activation. Our results accord with previous studies of discrete ACP proteins from fatty acid and aromatic polyketide biosynthesis, suggesting that helix II may also serve as a universal interaction motif in modular PKSs.

System for accepting server predictions in CASP6

We describe the new CASP system for collecting and verifying predictions generated by servers. The system was developed to ensure reliable execution of the server assessment part of CASP, with particular emphasis on data consistency. Following the principle that predictions should not be modified by anyone but their authors and to allow a later meaningful assessment, submissions are now verified for correctness of format and contents within the strict 48 hour CASP deadlines for this type of submission. This article also provides an overview of the rules governing server participation in CASP6 and some statistics pertaining to servers in CASP6.

Identification and characterization of collagen-binding activity in Streptococcus mutans wall-associated protein: A possible implication in dental root caries and endocarditis

Streptococcus mutans is implicated in coronal and dental root decay, and in endocarditis. Comparative study of the amino acid sequence of S. mutans 47 kDa wall-associated protein A (WapA) revealed a collagen-binding domain (CBD) at the N-terminal region. Recombinant AgA (WapA truncated at the carboxyterminal end) was isolated, biotin-labeled, and analyzed by Solid Phase Binding Assay. The results showed that biotin-labeled AgA bound significantly and in a dose-dependent manner to immobilized collagen type I, and to a lesser extent to fibronectin, but not to collagen type IV or laminin. Binding of biotin-labeled S. mutans cells to collagen-coated surfaces was significantly inhibited by antibody to WapA or AgA (P<0.001). The results obtained confirmed the collagen-binding activity of CBD in AgA and WapA, and suggested that WapA may be used, not only as a vaccine against coronal and dental root caries, but also against S. mutans-mediated endocarditis.

The yeast cyclin-dependent kinase inhibitor Sic1 and mammalian p27Kip1 are functional homologues with a structurally conserved inhibitory domain

In Saccharomyces cerevisiae, Sic1, an inhibitor of Cdk (cyclin-dependent kinase), blocks the activity of S-Cdk1 (Cdk1/Clb5,6) kinase that is required for DNA replication. Deletion of Sic1 causes premature DNA replication from fewer origins, extension of the S phase and inefficient separation of sister chromatids during anaphase. Despite the well-documented relevance of Sic1 inhibition of S-Cdk1 for cell cycle control and genome instability, the molecular mechanism by which Sic1 inhibits S-Cdk1 activity remains obscure. In this paper, we show that Sic1 is functionally and structurally related to the mammalian Cki (Cdk inhibitor) p27Kip1 of the Kip/Cip family. A molecular model of the inhibitory domain of Sic1 bound to the Cdk2-cyclin A complex suggested that the yeast inhibitor might productively interface with the mammalian Cdk2-cyclin A complex. Consistent with this, Sic1 is able to bind to, and strongly inhibit the kinase activity of, the Cdk2-cyclin A complex. In addition, comparison of the different inhibitory patterns obtained using histone H1 or GST (glutathione S-transferase)-pRb (retinoblastoma protein) fusion protein as substrate (the latter of which recognizes both the docking site and the catalytic site of Cdk2-cyclin A) offers interesting suggestions for the inhibitory mechanism of Sic1. Finally, overexpression of the KIP1 gene in vivo in Saccharomyces cerevisiae, like overexpression of the related SIC1 gene, rescues the cell cycle-related phenotype of a sic1Delta strain. Taken together, these findings strongly indicate that budding yeast Sic1 and mammalian p27(Kip1) are functional homologues with a structurally conserved inhibitory domain.

Servers for protein structure prediction

The 1990s cultivated a generation of protein structure human predictors. As a result of structural genomics and genome sequencing projects, and significant improvements in the performance of protein structure prediction methods, a generation of automated servers has evolved in the past few years. Servers for close and distant homology modeling are now routinely used by many biologists, and have already been applied to the experimental structure determination process itself, and to the interpretation and annotation of genome sequences. Because dozens of servers are currently available, it is hard for a biologist to know which server(s) to use; however, the state of the art of these methods is now assessed through the LiveBench and CAFASP experiments. Meta-servers--servers that use the results of other autonomous servers to produce a consensus prediction--have proven to be the best performers, and are already challenging all but a handful of expert human predictors. The difference in performance of the top ten autonomous (non-meta) servers is small and hard to assess using relatively small test sets. Recent experiments suggest that servers will soon free humans from most of the burden of protein structure prediction.

Homology modeling of membrane proteins: A critical assessment

Evaluation and validation of homology modeling protocols are indispensable for membrane proteins as experimental determination of their three-dimensional structure is an arduous task. The prediction ability of Modeller, MOE, InsightII-Homology and Swiss-PdbViewer (SPV) with different sequence alignments CLUSTALW, BLAST and 3D-JIGSAW have been assessed. The sequence identity of the target and template was chosen to be in the range of 25-35%. Validation protocols to assess the structure, fold and stereochemical quality, are employed by comparing with experimental structures. Two different ranking schemes are suggested to evaluate the performance of each methodology based on the validation scores. While unambiguous preference for any given procedure did not surface, statistically Modeller and the sequence alignment technique, 3D-JIGSAW, gave best results amongst the chosen protocols. The present study helps in selecting the right protocols when modeling membrane proteins, which form a major class of drug targets.

The thermal adaptation of the nitrogenase Fe protein from thermophilic Methanobacter thermoautotrophicus

The nitrogenase Fe protein is a key component of the biochemical machinery responsible for the process of biological nitrogen fixation. The Fe protein is a member of a class of nucleotide-binding proteins that couple the binding and hydrolysis of nucleoside triphosphates to conformational changes. The nucleotide-dependent conformational changes modulate the formation of a macromolecular complex, and some members of the class include Galpha, EF-Tu, and myosin. The members of this class are highly interesting model systems for the analysis of aspects of thermal adaptability, since their mechanisms involve protein conformational change and protein-protein interactions. In this study, we have used our extensive knowledge of the structure of the Azotobacter vinelandii nitrogenase Fe protein in multiple structural conformations, and standard homology modeling approaches have been used to generate reliable models of the Fe protein from thermophilic Methanobacter thermoautotrophicus in the analogous structural conformations. The resulting structural comparison reveals that thermal adaptation of the M. thermoautotrophicus Fe protein is conferred by a number of factors, including increased structural rigidity that results from various structural changes within the protein interior. The analysis of hypothetical docking models and nitrogenase complex structures provides insights into the thermal adaptation of the protein-protein interactions that support macromolecular complex formation and catalysis at higher temperatures.

The victor/FRST function for model quality estimation

Scoring functions are widely used in the final step of model selection in protein structure prediction. This is of interest both for comparative modeling targets, where it is important to select the best model among a set of many good, "correct" ones, as well as for other (fold recognition or novel fold) targets, where the set may contain many incorrect models. A novel combination of four knowledge-based potentials recognizing different features of native protein structures is introduced and tested. The pairwise, solvation, hydrogen bond, and torsion angle potentials contain largely orthogonal information. Of these, the torsion angle potential is found to show the strongest correlation with model quality. Combining these features with a linear weighting function, it was possible to construct a robust energy function capable of discriminating native-like structures on several benchmarking sets. In a recent blind test (CAFASP-4 MQAP), the scoring function ranked consistently well and was able to reliably distinguish the correct template from an ensemble of high quality decoys in 52 of 70 cases (33 of 34 for comparative modeling). An executable version of the Victor/FRST function for Linux PCs is available for download from the URL http://protein.cribi.unipd.it/frst/.

Genomic organization, characterization, and molecular 3D model of GDE1, a novel mammalian glycerophosphoinositol phosphodiesterase

Glycerophosphodiester phosphodiesterase (GDPD) catalyzes the hydrolysis of deacylated glycerophospholipids to glycerol phosphate and alcohol. A mammalian glycerophosphoinositol phosphodiesterase, GDE1/MIR16, was recently identified as an interacting protein of the regulator of G protein signaling 16 (RGS16) providing a link between phosphoinositide metabolism and G protein signal transduction. To further understand the function and properties of human GDE1, we determined its genomic organization and its biochemical and structural characteristics. GDE1 encodes a 331-residue protein with two hydrophobic domains and contains a GDE domain that shares strong homologies with GDE1-related proteins as well as bacterial GDPDs. The human GDE1 gene is located on chromosome 16p12-p11.2 and contains six exons and five introns. A molecular 3D model, which was built based on the crystal structure of Escherichia coli GDPD (1YDY), provides the first structural information of human GDE1 and suggests a TIM barrel core as typically found in bacterial GDPDs. Furthermore, a model of the putative catalytic motif within the GDE domain was nearly identical to the corresponding domain of GDPD and highlights the individual core residues Glu97, Asp99, and His112, which are crucial to maintaining GDE1 catalytic activity. These studies provide important new insights into understanding the function of GDE1 and GDE1-related proteins.

Iodotyrosine Deiodinase Is the First Mammalian Member of the NADH Oxidase/Flavin Reductase Superfamily

The enzyme responsible for iodide salvage in the thyroid, iodotyrosine deiodinase, was solubilized from porcine thyroid microsomes by limited proteolysis with trypsin. The resulting protein retained deiodinase activity and was purified using anion exchange, dye, and hydrophobic chromatography successively. Peptide sequencing of the final isolate identified the gene responsible for the deiodinase. The amino acid sequence of the porcine enzyme is highly homologous to corresponding genes in a variety of mammals including humans, and the mouse gene was expressed in human embryonic kidney 293 cells to confirm its identity. The amino acid sequence of the deiodinase suggests the presence of three domains. The N-terminal domain provides a membrane anchor. The intermediate domain contains the highest sequence variability and lacks homology to structural motifs available in the common databases. The C-terminal domain is highly conserved and resembles bacterial enzymes of the NADH oxidase/flavin reductase superfamily. A three-dimensional model of the deiodinase based on the coordinates of the minor nitroreductase of Escherichia coli indicates that a Cys common to all of the mammal sequences is located adjacent to bound FMN. However, the deiodinase is not structurally related to other known flavoproteins containing redox-active cysteines or the iodothyronine deiodinases containing an active site selenocysteine.

Cloning and structural characterization of the 6-phosphogluconate dehydrogenase locus of the medfly Ceratitis capitata and the olive fruit fly Bactrocera oleae

The pentose phosphate cycle is considered as a major source of NADPH and pentose needed for nucleic acid biosynthesis. 6-Phosphogluconate dehydrogenase (6PGD), an enzyme participating in this cycle, catalyzes the oxidative decarboxylation of 6PGD to ribulose 5-phosphate with the subsequent release of CO(2) and the reduction of NADP. We have determined the genomic sequences of 6PGD of two species of Tephritidae, the medfly Ceratitis capitata and olive fruit fly Bactrocera oleae, and constructed a three-dimensional model of 6PGD of C. capitata based on the homologous known sheep structure. In a comparative study of 6PGD sequences from seven species, all the conserved and variable sites of the enzyme were analyzed and the regions of functional importance were localized, an attempt promoted also by the direct involvement of the enzyme in various human diseases. The enzymes between the two species of Tephritidae have a very high homology and further examination of the variable positions with respect to the highly conserved binding site residues enabled their grouping in three distinct categories, with possible association to dimer formation, functional specificity, and antigenicity. Moreover, placement of sequence differences on the 3-D model suggests probable sites accommodating variations appearing at the allozymic variants of both species.

A fold-recognition approach to loop modeling

A novel approach is proposed for modeling loop regions in proteins. In this approach, a prerequisite sequence-structure alignment is examined for regions where the target sequence is not covered by the structural template. These regions, extended with a number of residues from adjacent stem regions, are submitted to fold recognition. The alignments produced by fold recognition are integrated into the initial alignment to create an alignment between the target sequence and several structures, where gaps in the main structural template are covered by local structural templates. This one-to-many (1:N) alignment is used to create a protein model by existing protein-modeling techniques. Several alternative approaches were evaluated using a set of ten proteins. One approach was selected and evaluated using another set of 31 proteins. The most promising result was for gap regions not located at the C-terminus or N-terminus of a protein, where the method produced an average RMSD 12% lower than the loop modeling provided with the program MODELLER. This improvement is shown to be statistically significant.

Extended sequence and functional analysis of the butirosin biosynthetic gene cluster in Bacillus circulans SANK 72073

Butirosin produced by Bacillus circulans is among the clinically important 2-deoxystreptamine containing aminoglycoside antibiotics and its unique structure is found in (S)-4-amino-2-hydroxyburyric acid substituted at C-1 of 2-deoxystreptamine. Recently, the key part of the butirosin biosynthetic gene cluster has been identified from Bacillus circulans SANK 72073, however the whole gene for the biosynthesis awaited for identification. In the present study, we undertook extended analysis of the butirosin biosynthetic gene cluster and found nine additional open reading flames (ORFs), btrQ, btrR1, btrR2, btrT, btrU, btrV, btrW, btrX and orf1 in the cluster. In addition, we constructed disruption mutants of btrR1 and btrP-V, and found that the btr genes (ca. 24Kb) between btrR1 and btrP-V are at least required for the butirosin biosynthesis.

The Intracellular Trafficking of the G Protein-coupled Receptor TPβ Depends on a Direct Interaction with Rab11

Intracellular trafficking pathways of cell surface receptors following their internalization are the subject of intense research efforts. However, the mechanisms by which they recycle back to the cell surface are still poorly defined. We have recently demonstrated that the small Rab11 GTPase protein is a determinant factor in controlling the recycling to the cell surface of the beta-isoform of the thromboxane A2 receptor (TPbeta) following its internalization. Here, we demonstrate with co-immunoprecipitation studies in HEK293 cells that there is a Rab11-TPbeta association occurring in the absence of agonist, which is not modulated by stimulation of TPbeta. We show with purified TPbeta intracellular domains fused to GST and HIS-Rab11 proteins that Rab11 interacts directly with the first intracellular loop and the C-tail of TPbeta. Amino acids 335-344 of the TPbeta C-tail were determined to be essential for the interaction of Rab11 with this receptor domain. This identified sequence appears to be important in directing the intracellular trafficking of the receptor from the Rab5-positive intracellular compartment to the perinuclear recycling endosome. Interestingly, our data indicate that TPbeta interacts with the GDP-bound form, and not the GTP-bound form, of Rab11 which is necessary for recycling of the receptor back to the cell surface. To our knowledge, this is the first demonstration of a direct interaction between Rab11 and a transmembrane receptor.

The complete sequence of the bovine torovirus genome

Viruses in the family Coronaviridae have elicited new interest, with the outbreaks caused by SARS-HCoV in 2003 and the recent discovery of a new human coronavirus, HCoV-NL63. The genus Torovirus, within the family Coronaviridae, is less well characterized, in part because toroviruses cannot yet be grown in cell culture (except for the Berne virus). In this study, we determined the sequence of the complete genome of Breda-1 (BoTV-1), a bovine torovirus. This is the first complete torovirus genome sequence to be reported. BoTV-1 RNA was amplified using long RT-PCR and the amplicons sequenced. The genome has a length of 28.475 kb and consisted mainly of the replicase gene (∼20.2 kb) which contains two large overlapping ORFs, ORF1a and ORF1b, encoding polyproteins pp1a and pp1b, respectively. Sequence analysis identified conserved domains within the predicted sequences of pp1a and pp1b. Sequence alignments and protein secondary structure prediction data suggest the presence of a 3C-like serine protease domain with similarity to the arterivirus 3C-like serine protease and a single papain-like cysteine protease domain with similarity to the picornavirus leader protease. The ADRP (APPR-1″) domain – unique to the Coronaviridae – was also located in BoTV pp1a. In addition, several hydrophobic domains were identified that are typical of a nidovirus replicase. Within the pp1b sequence the polymerase and helicase domains were identified, as well as sequences predicted to be involved in ribosomal frameshifting, including the conserved slippery sequence UUUAAAC and two potential pseudoknot structures.

Beyond the brush border: NHERF4 blazes new NHERF turf

The Na exchanger regulatory factor (NHERF) family of epithelial-enriched PDZ domain scaffolding proteins plays important roles in maintaining and regulating epithelial cell function. The NHERFs exhibit some overlap in tissue distribution and binding partners, suggesting redundant functions. Yet, it is clear that each NHERF protein exhibits distinct properties, translating into unique cellular functions. The work summarized in this review suggests the most recently identified family member, NHERF4, is the most divergent. Additional investigation is needed, however, to understand more completely the role of NHERF4 in the context of the NHERF family.

California serogroup Gc (G1) glycoprotein is the principal determinant of pH-dependent cell fusion and entry

Members of the California serogroup of orthobunyaviruses, particularly La Crosse (LAC) and Tahyna (TAH) viruses, are significant human pathogens in areas where their mosquito vectors are endemic. Previous studies using wild-type LAC and TAH181/57, a highly neurovirulent strain with low neuroinvasiveness (Janssen, R., Gonzalez-Scarano, F., Nathanson, N., 1984. Mechanisms of bunyavirus virulence. Comparative pathogenesis of a virulent strain of La Crosse and an avirulent strain of Tahyna virus. Lab. Invest. 50 (4), 447-455), have demonstrated that the neuroinvasive phenotype maps to the M segment, the segment that encodes the two viral glycoproteins Gn (G2) and Gc (G1), as well as a non-structural protein NSm. To further define the role of Gn and Gc in fusion and entry, we prepared a panel of recombinant M segment constructs using LAC, TAH181/57, and V22F, a monoclonal-resistant variant of LAC with deficient fusion function. These M segment constructs were then tested in two surrogate assays for virus entry: a cell-to-cell fusion assay based on T7-luciferase expression, and a pseudotype transduction assay based on the incorporation of the bunyavirus glycoproteins on an MLV backbone. Both assays demonstrated that Gc is the principal determinant of virus fusion and cell entry, and furthermore that the region delineated by amino acids 860-1442, corresponding to the membrane proximal two-thirds of Gc, is key to these processes. These results, coupled with structural modeling suggesting homologies between the carboxy region of Gc and Sindbis virus E1, suggest that the LAC Gc functions as a type II fusion protein.

Sphingomyelinase D from venoms of Loxosceles spiders: evolutionary insights from cDNA sequences and gene structure

Loxosceles spider venoms cause dermonecrosis in mammalian tissues. The toxin sphingomyelinase D (SMaseD) is a sufficient causative agent in lesion formation and is only known in these spiders and a few pathogenic bacteria. Similarities between spider and bacterial SMaseD in molecular weights, pIs and N-terminal amino acid sequence suggest an evolutionary relationship between these molecules. We report three cDNA sequences from venom-expressed mRNAs, analyses of amino acid sequences, and partial characterization of gene structure of SMaseD homologs from Loxosceles arizonica with the goal of better understanding the evolution of this toxin. Sequence analyses indicate SMaseD is a single domain protein and a divergent member of the ubitiquous, broadly conserved glycerophosphoryl diester phosphodiesterase family (GDPD). Bacterial SMaseDs are not identifiable as homologs of spider SMaseD or GDPD family members. Amino acid sequence similarities do not afford clear distinction between independent origin of toxic SMaseD activity in spiders and bacteria and origin in one lineage by ancient horizontal transfer from the other. The SMaseD genes span at least 6500bp and contain at least 5 introns. Together, these data indicate L. arizonica SMaseD has been evolving within a eukaryotic genome for a long time ruling out origin by recent transfer from bacteria.

Two regions in the N-terminal domain of ionotropic glutamate receptor 3 form the subunit oligomerization interfaces that control subtype-specific receptor assembly

The N-terminal domain (NTD) of alpha-amino-3-hydroxy-5-methylisoxazolepropionate (AMPA) and kainate glutamate receptors plays an important role in controlling subtype specific receptor assembly. To identify NTD subdomains involved in this process we generated AMPA glutamate receptor 3 (GluR3) mutants having intra-NTD substitutions with the corresponding regions of the kainate receptor GluR6 and tested their ability to form functional heteromers with wild-type subunits. The chimeric design was based on the homology of the NTD to the NTD of the metabotropic GluR1, shown to form two globular lobes and to assemble in dimers. Accordingly, the NTD was divided into four regions, termed here N1-N4, of which N1 and N3 correspond to the regions forming lobe-1 and N2 and N4 to those forming lobe-2. Substituting N1 or N3 impaired functional heteromerization but allowed protein-protein interactions. Conversely, exchanging N2 or N4 preserved functional heteromerization, although it significantly decreased homomeric activity, indicating a role in subunit folding. Moreover, a deletion in GluR3 corresponding to the hotfoot mouse mutation of the glutamate receptor delta2, covering part of N2, N3, and N4, impaired both homomeric and heteromeric oligomerization, thus explaining the null-like mouse phenotype. Finally, computer modeling suggested that the dimer interface, largely formed by N1, is highly hydrophobic in GluR3, whereas in GluR6 it contains electrostatic interactions, hence offering an explanation for the subtype assembly specificity conferred by this region. N3, however, is positioned perpendicular to the dimer interface and therefore may be involved in secondary interactions between dimers in the assembled tetrameric receptor.

Nuclear and Mitochondrial Localization Signals Overlap within Bovine Herpesvirus 1 Tegument Protein VP22

VP22, a tegument protein of bovine herpesvirus 1, accumulates in the nucleus of infected and transiently transfected cells. Previous studies indicated a possible regulatory function of VP22 within nuclei, but how VP22 enters nuclei is unknown. Despite the abundance of basic residues within this protein, no classic nuclear localization signal (NLS) motif has been identified. To identify the signal directing nuclear accumulation, a series of truncations, internal deletions, and point mutations were constructed. Fluorescence microscopy of cells transfected with VP22 constructs indicated that a sequence of 103 residues is necessary and sufficient for nuclear localization. This NLS sequence is conformation-sensitive in contrast to a classical sequential NLS. Energy depletion assays and co-immunoprecipitation suggested that this NLS sequence also binds histone H4, resulting in nuclear retention of VP22. In addition, a mitochondrial targeting sequence was identified at the C-terminal 49 amino acids, which overlapped the sequence required for nuclear targeting. Our findings demonstrate the diversity of VP22 protein to localize within the cell and provide the opportunity for VP22 to direct cargo specifically to different subcellular compartments.

Functional constraints of 6-phosphogluconate dehydrogenase (6-PGD) based on sequence and structural information

The pentose phosphate cycle is considered as a major source of NADPH and pentose needed for nucleic acid biosynthesis. 6-Phosphogluconate dehydrogenase (6PGD), an enzyme participating in this cycle, catalyzes the oxidative decarboxylation of 6PGD to ribulose 5-phosphate with the subsequent release of CO2 and the reduction of NADP. We have determined the amino acid sequence of 6PGD of Bactrocera oleae and constructed a three-dimensional model based on the homologous known sheep structure. In a comparative study of 6PGD sequences from numerous species, all the conserved and variable regions of the enzyme were analyzed and the regions of functional importance were localized, in an attempt promoted also by the direct involvement of the enzyme in various human diseases. Thus, analysis of amino acid variability of 37 6PGD sequences revealed that all regions important for the catalytic activity, such as those forming the substrate and coenzyme binding sites, are highly conserved in all species examined. Moreover, several amino acid residues responsible for substrate and coenzyme specificity were also found to be identical in all species examined. The higher percentage of protein divergence is observed at two regions that accumulate mutations, located at the distant parts of the two domains of the enzyme with respect to their interface. These peripheral regions of non-functional importance are highly variable and are predicted as antigenic, thus reflecting possible regions for antibody recognition. Furthermore, locating the differences between diptera 6PGD sequences on the three-dimensional model suggests probable positions of different amino acid residues appearing at B. oleae fast, intermediate, and slow allozymic variants.

Site-directed mutagenesis, proteolytic cleavage, and activation of human proheparanase

Heparanase is an endo-beta-D-glucuronidase that degrades heparan sulfate in the extracellular matrix and cell surfaces. Human proheparanase is produced as a latent 65-kDa polypeptide undergoing processing at two potential proteolytic cleavage sites, located at Glu109-Ser110 (site 1) and Gln157-Lys158 (site 2). Cleavage of proheparanase yields 8- and 50-kDa subunits that heterodimerize to form the active enzyme. The fate of the linker segment (Ser110-Gln157) residing between the two subunits, the mode of processing, and the protease(s) engaged in proheparanase processing are currently unknown. We applied multiple site-directed mutagenesis and deletions to study the nature of the potential cleavage sites and amino acids essential for processing of proheparanase in transfected human choriocarcinoma cells devoid of endogenous heparanase but possessing the enzymatic machinery for proper processing and activation of the proenzyme. Although mutagenesis at site 1 and its flanking sequences failed to identify critical residues for proteolytic cleavage, processing at site 2 required a bulky hydrophobic amino acid at position 156 (i.e. P2 of the cleavage site). Substitution of Tyr156 by Ala or Glu, but not Val, resulted in cleavage at an upstream site in the linker segment, yielding an improperly processed inactive enzyme. Processing of the latent 65-kDa proheparanase in transfected Jar cells was inhibited by a cell-permeable inhibitor of cathepsin L. Moreover, recombinant 65-kDa proheparanase was processed and activated by cathepsin L in a cell-free system. Altogether, these results suggest that proheparanase processing at site 2 is brought about by cathepsin L-like proteases. The involvement of other members of the cathepsin family with specificity to bulky hydrophobic residues cannot be excluded. Our results and a three-dimensional model of the enzyme are expected to accelerate the design of inhibitory molecules capable of suppressing heparanase-mediated enhancement of tumor angiogenesis and metastasis.

Isolation and Characterization of Active LINE and SINEs from the Eel

Long interspersed elements (LINEs) and short interspersed elements (SINEs) are retrotransposons. These elements can mobilize by the "copy-and-paste" mechanism, in which their own RNA is reverse-transcribed into complementary DNA (cDNA). LINEs and SINEs not only are components of eukaryotic genomes but also drivers of genomic evolution. Thus, studies of the amplification mechanism of LINEs and SINEs are important for understanding eukaryotic genome evolution. Here we report the characterization of one LINE family (UnaL2) and two SINE families (UnaSINE1 and UnaSINE2) from the eel (Anguilla japonica) genome. UnaL2 is approximately 3.6 kilobases (kb) and encodes only one open reading frame (ORF). UnaL2 belongs to the stringent type--thought to be a major group of LINEs--and can mobilize in HeLa cells. We also show that UnaL2 and the two UnaSINEs have similar 3' tails, and that both UnaSINE1 and UnaSINE2 can be mobilized by UnaL2 in HeLa cells. These elements are thus useful for delineating the amplification mechanism of stringent type LINEs as well as that of SINEs.

Tissue Distribution and Evolution of Fructosamine 3-Kinase and Fructosamine 3-Kinase-related Protein

Fructosamine 3-kinase (FN3K) and FN3K-related protein (FN3K-RP) catalyze the phosphorylation of the Amadori products ribulosamines, psicosamines, and, in the case of FN3K, fructosamines. BLAST searches in chordate genomes revealed two genes encoding proteins homologous to FN3K or FN3K-RP in various mammals and in chicken but only one gene, encoding a protein more similar to FN3K-RP than to FN3K, in fishes and the sea squirt Ciona intestinalis. This suggests that a gene duplication event occurred after the fish radiation and that the FN3K gene evolved more rapidly than the FN3K-RP gene. In agreement with this distribution, only one enzyme, phosphorylating ribulosamines and psicosamines but not fructosamines, was found in the tissues from a fish (Clarias gariepinus), whereas two enzymes with specificities similar to either FN3K or FN3K-RP were found in mouse, rat, and chicken tissues. FN3K is particularly active in brain, heart, kidney, and skeletal muscle. Its activity is also relatively elevated in erythrocytes from man, rat, and mouse but barely detectable in erythrocytes from chicken and pig, which correlates well with the low intracellular concentration of glucose in erythrocytes from these species. This is in keeping with the specific role of FN3K to repair protein damage caused by glucose. FN3K-RP was more evenly distributed in tissues, except for skeletal muscle where its activity was particularly low. This may be related to low activity of the pentose phosphate pathway in this tissue, as suggested by assays of glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase. This finding, together with the high affinity of FN3K-RP for ribulosamines, suggests that this enzyme may serve to repair damage caused by the powerful glycating agent, ribose 5-phosphate.

Exploiting evolutionary relationships for predicting protein structures

In the last few years there have been many developments in computational biology, particularly with regard to novel, imaginative exploitation of genomic data. Disappointingly, there has been a lack of progress in the methodology for prediction of protein structures. In the last several years, however, promising new methods have finally begun to emerge. These methods are increasing the power and scope of the methodology, but, most importantly, they are generating new areas of investigation that we believe will accelerate progress in the field. In this review we describe recent developments and highlight the implications of their success as well as areas where efforts should be focused.

Site-directed mutagenesis and protein 3D-homology modelling suggest a catalytic mechanism for UDP-glucose-dependent betanidin 5-O-glucosyltransferase from Dorotheanthus bellidiformis

In livingstone daisy (Dorotheanthus bellidiformis), betanidin 5-O-glucosyltransferase (UGT73A5) is involved in the regiospecific glucosylation of betanidin and various flavonols. Based on sequence alignments several amino acid candidates which might be essential for catalysis were identified. The selected amino acids of the functionally expressed protein, suggested to be involved in substrate binding and turnover, were substituted via site-directed mutagenesis. The substitution of two highly conserved amino acids, Glu378, located in the proposed UDP-glucose binding site, and His22, located close to the N-terminus, led to the complete loss of enzyme activity. A 3D model of this regiospecific betanidin and flavonoid glucosyltransferase was constructed and the active site modelled. This model was based on the crystallographic structure of a bacterial UDP-glucose-dependent glucosyltransferase from Amycolatopsis orientalis used as a template and the generated null mutations. To explain the observed inversion in the configuration of the bound sugar, semiempirical calculations favour an SN-1 reaction, as one plausible alternative to the generally proposed SN-2 mechanism discussed for plant natural product glucosyltransferases. The calculated structural data do not only explain the abstraction of a proton from the acceptor betanidin, but further imply that the reaction mechanism might also involve a catalytic triad, with similarities described for the serine protease family.

Structural components of SCAN-domain dimerizations

The SCAN or leucine-rich domain has been characterized as a highly conserved sequence in zinc finger transcription factors that mediates selective dimer formation between SCAN-domain-containing proteins. In order to accommodate various SCAN-domain sequence features, a minimal functional folding unit was defined on the premise of proper structural folding and biochemical binding. The 58-amino acid minimal functional units derived from each of four SCAN-domain protein families were subjected to a three-dimensional position-specific scoring matrix (3D-PSSM) and ungapped threading analysis. The resulting fold prediction represented the SCAN-domain's minimal functional unit as a bundle of three alpha helices folded to a core structure. In addition, the minimal functional folding unit biochemically retained the selective dimerization properties of the native proteins. In order to elucidate the structural components within the SCAN-domain that engage in binding interactions, we attempted to correlate the physicochemical helix properties, as represented by a hydropathy profile, with the experimental dimerization selectivities. The amino-terminal helix revealed the highest diversity measure among the three helices of the minimal functional unit and is therefore likely to offer critical surface-exposed binding residues. Indeed, by interchanging the amino-terminal helix between SCAN-domains without alteration of their structural frames consisting of conserved hydrophobic residues, a modulation of binding preferences was demonstrated. The minimal functional folding unit of SCAN-domains may therefore contain within the amino-terminal alpha helix structural components that determine selective dimerization patterns and combinatorial control of transcription factors.