Network Analysis of Sequence-Function Relationships and Exploration of Sequence Space of TEM ß-Lactamases.

The Lactamase Engineering Database (www.LacED.uni-stuttgart.de) was developed to facilitate the classification and analysis of TEM ß-lactamases. The current version contains 474 TEM variants. Two hundred fifty-nine variants form a large scale-free network of highly connected point mutants. The network was divided into three subnetworks which were enriched by single phenotypes: one network with predominantly 2be and two networks with 2br phenotypes. Fifteen positions were found to be highly variable, contributing to the majority of the observed variants. Since it is expected that a considerable fraction of the theoretical sequence space is functional, the currently sequenced 474 variants represent only the tip of the iceberg of functional TEM ß-lactamase variants which form a huge natural reservoir of highly interconnected variants. Almost 50% of the variants are part of a quartet. Thus, two single mutations that result in functional enzymes can be combined into a functional protein. Most of these quartets consist of the same phenotype, or the mutations are additive with respect to the phenotype. By predicting quartets from triplets, 3,916 unknown variants were constructed. Eighty-seven variants complement multiple quartets and therefore have a high probability of being functional. The construction of a TEM ß-lactamase network and subsequent analyses by clustering and quartet prediction are valuable tools to gain new insights into the viable sequence space of TEM ß-lactamases and to predict their phenotype. The highly connected sequence space of TEM ß-lactamases is ideally suited to network analysis and demonstrates the strengths of network analysis over tree reconstruction methods.

Systematic analysis of metallo-ß-lactamases using an automated database.

Metallo-ß-lactamases (MBLs) are enzymes that hydrolyze ß-lactam antibiotics, resulting in bacterial resistance to these drugs. These proteins have caused concerns due to their facile transference, broad substrate spectra, and the absence of clinically useful inhibitors. To facilitate the classification, nomenclature, and analysis of MBLs, an automated database system was developed, the Metallo-ß-Lactamase Engineering Database (MBLED) (http://www.mbled.uni-stuttgart.de). It contains information on MBLs retrieved from the NCBI peptide database while strictly following the nomenclature by Jacoby and Bush (http://www.lahey.org/Studies/) and the generally accepted class B ß-lactamase (BBL) standard numbering scheme for MBLs. The database comprises 597 MBL protein sequences and enables systematic analyses of these sequences. A systematic analysis employing the database resulted in the generation of mutation profiles of assigned IMP- and VIM-type MBLs, the identification of five MBL protein entries from the NCBI peptide database that were inconsistent with the Jacoby and Bush nomenclature, and the identification of 15 new IMP candidates and 9 new VIM candidates. Furthermore, the database was used to identify residues with high mutation frequencies and variability (mutation hot spots) that were unexpectedly distant from the active site located in the ßß sandwich: positions 208 and 266 in the IMP family and positions 215 and 258 in the VIM family. We expect that the MBLED will be a valuable tool for systematically cataloguing and analyzing the increasing number of MBLs being reported.

A standard numbering scheme for thiamine diphosphate-dependent decarboxylases.

BACKGROUND: Standard numbering schemes for families of homologous proteins allow for the unambiguous identification of functionally and structurally relevant residues, to communicate results on mutations, and to systematically analyse sequence-function relationships in protein families. Standard numbering schemes have been successfully implemented for several protein families, including lactamases and antibodies, whereas a numbering scheme for the structural family of thiamine-diphosphate (ThDP) -dependent decarboxylases, a large subfamily of the class of ThDP-dependent enzymes encompassing pyruvate-, benzoylformate-, 2-oxo acid-, indolpyruvate- and phenylpyruvate decarboxylases, benzaldehyde lyase, acetohydroxyacid synthases and 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexadiene-1-carboxylate synthase (MenD) is still missing.Despite a high structural similarity between the members of the ThDP-dependent decarboxylases, their sequences are diverse and make a pairwise sequence comparison of protein family members difficult. RESULTS: We developed and validated a standard numbering scheme for the family of ThDP-dependent decarboxylases. A profile hidden Markov model (HMM) was created using a set of representative sequences from the family of ThDP-dependent decarboxylases. The pyruvate decarboxylase from S. cerevisiae (PDB: 2VK8) was chosen as a reference because it is a well characterized enzyme. The crystal structure with the PDB identifier 2VK8 encompasses the structure of the ScPDC mutant E477Q, the cofactors ThDP and Mg(2+) as well as the substrate analogue (2S)-2-hydroxypropanoic acid. The absolute numbering of this reference sequence was transferred to all members of the ThDP-dependent decarboxylase protein family. Subsequently, the numbering scheme was integrated into the already established Thiamine-diphosphate dependent Enzyme Engineering Database (TEED) and was used to systematically analyze functionally and structurally relevant positions in the superfamily of ThDP-dependent decarboxylases. CONCLUSIONS: The numbering scheme serves as a tool for the reliable sequence alignment of ThDP-dependent decarboxylases and the unambiguous identification and communication of corresponding positions. Thus, it is the basis for the systematic and automated analysis of sequence-encoded properties such as structural and functional relevance of amino acid positions, because the analysis of conserved positions, the identification of correlated mutations and the determination of subfamily specific amino acid distributions depend on reliable multisequence alignments and the unambiguous identification of the alignment columns. The method is reliable and robust and can easily be adapted to further protein families.

Structural classification by the Lipase Engineering Database: a case study of Candida antarctica lipase A.

BACKGROUND: The Lipase Engineering Database (LED) integrates information on sequence, structure and function of lipases, esterases and related proteins with the alpha/beta hydrolase fold. A new superfamily for Candida antarctica lipase A (CALA) was introduced including the recently published crystal structure of CALA. Since CALA has a highly divergent sequence in comparison to other alpha/beta hydrolases, the Lipase Engineering Database was used to classify CALA in the frame of the already established classification system. This involved the comparison of CALA to similar structures as well as sequence-based comparisons against the content of the LED. RESULTS: The new release 3.0 (December 2009) of the Lipase Engineering Database contains 24783 sequence entries for 18585 proteins as well as 656 experimentally determined protein structures, including the structure of CALA. In comparison to the previous release 1 with 4322 protein and 167 structure entries this update represents a significant increase in data volume. By comparing CALA to representative structures from all superfamilies, a structure from the deacetylase superfamily was found to be most similar to the structure of CALA. While the alpha/beta hydrolase fold is conserved in both proteins, the major difference is found in the cap region. Sequence alignments between both proteins show a sequence similarity of only 15%. A multisequence alignment of both protein families was used to create hidden Markov models for the cap region of CALA and showed that the cap region of CALA is unique among all other proteins of the alpha/beta hydrolase fold. By specifically comparing the substrate binding pocket of CALA to other binding pockets of alpha/beta hydrolases, the binding pocket of Candida rugosa lipase was identified as being highly similar. This similarity also applied to the lid of Candida rugosa lipase in comparison to the potential lid of CALA. CONCLUSION: The LED serves as a valuable tool for the systematic analysis of single proteins or protein families. The updated release 3.0 was used for the evaluation of alpha/beta hydrolases. The HTML version of the database with new features is available at http://www.led.uni-stuttgart.de and provides sequences, structures and a set of analysis tools including phylogenetic trees and HMM profiles.

Prediction and analysis of the modular structure of cytochrome P450 monooxygenases.

BACKGROUND: Cytochrome P450 monooxygenases (CYPs) form a vast and diverse family of highly variable sequences. They catalyze a wide variety of oxidative reactions and are therefore of great relevance in drug development and biotechnological applications. Despite their differences in sequence and substrate specificity, the structures of CYPs are highly similar. Although being in research focus for years, factors mediating selectivity and activity remain vague. DESCRIPTION: This systematic comparison of CYPs based on the Cytochrome P450 Engineering Database (CYPED) involved sequence and structure analysis of more than 8000 sequences. 31 structures have been applied to generate a reliable structure-based HMM profile in order to predict structurally conserved regions. Therefore, it was possible to automatically transfer these modules on CYP sequences without any secondary structure information, to analyze substrate interacting residues and to compare interaction sites with redox partners. CONCLUSIONS: Functionally relevant structural sites of CYPs were predicted. Regions involved in substrate binding were analyzed in all sequences among the CYPED. For all CYPs that require a reductase, two reductase interaction sites were identified and classified according to their length. The newly gained insights promise an improvement of engineered enzyme properties for potential biotechnological application. The annotated sequences are accessible on the current version of the CYPED. The prediction tool can be applied to any CYP sequence via the web interface at http://www.cyped.uni-stuttgart.de/cgi-bin/strpred/dosecpred.pl.

The Thiamine diphosphate dependent Enzyme Engineering Database: a tool for the systematic analysis of sequence and structure relations.

BACKGROUND: Thiamine diphosphate (ThDP)-dependent enzymes form a vast and diverse class of proteins, catalyzing a wide variety of enzymatic reactions including the formation or cleavage of carbon-sulfur, carbon-oxygen, carbon-nitrogen, and especially carbon-carbon bonds. Although very diverse in sequence and domain organisation, they share two common protein domains, the pyrophosphate (PP) and the pyrimidine (PYR) domain. For the comprehensive and systematic comparison of protein sequences and structures the Thiamine diphosphate (ThDP)-dependent Enzyme Engineering Database (TEED) was established. DESCRIPTION: The TEED http://www.teed.uni-stuttgart.de contains 12048 sequence entries which were assigned to 9443 different proteins and 379 structure entries. Proteins were assigned to 8 different superfamilies and 63 homologous protein families. For each family, the TEED offers multisequence alignments, phylogenetic trees, and family-specific HMM profiles. The conserved pyrophosphate (PP) and pyrimidine (PYR) domains have been annotated, which allows the analysis of sequence similarities for a broad variety of proteins. Human ThDP-dependent enzymes are known to be involved in many diseases. 20 different proteins and over 40 single nucleotide polymorphisms (SNPs) of human ThDP-dependent enzymes were identified in the TEED. CONCLUSIONS: The online accessible version of the TEED has been designed to serve as a navigation and analysis tool for the large and diverse family of ThDP-dependent enzymes.

Analysis of the distribution of functionally relevant rare codons.

BACKGROUND: The substitution of rare codons with more frequent codons is a commonly applied method in heterologous gene expression to increase protein yields. However, in some cases these substitutions lead to a decrease of protein solubility or activity. To predict these functionally relevant rare codons, a method was developed which is based on an analysis of multisequence alignments of homologous protein families. RESULTS: The method successfully predicts functionally relevant codons in fatty acid binding protein and chloramphenicol acetyltransferase which had been experimentally determined. However, the analysis of 16 homologous protein families belonging to the alpha/beta hydrolase fold showed that functionally rare codons share no common location in respect to the tertiary and secondary structure. CONCLUSION: A systematic analysis of multisequence alignments of homologous protein families can be used to predict rare codons with a potential impact on protein expression. Our analysis showed that most genes contain at least one putative rare codon rich region. Rare codons located near to those regions should be excluded in an approach of improving protein expression by an exchange of rare codons by more frequent codons.

Protein variants form a system of networks: microdiversity of IMP metallo-beta-lactamases.

Genome and metagenome sequencing projects support the view that only a tiny portion of the total protein microdiversity in the biosphere has been sequenced yet, while the vast majority of existing protein variants is still unknown. By using a network approach, the microdiversity of 42 metallo-ß-lactamases of the IMP family was investigated. In the networks, the nodes are formed by the variants, while the edges correspond to single mutations between pairs of variants. The 42 variants were assigned to 7 separate networks. By analyzing the networks and their relationships, the structure of sequence space was studied and existing, but still unknown, functional variants were predicted. The largest network consists of 10 variants with IMP-1 in its center and includes two ubiquitous mutations, V67F and S262G. By relating the corresponding pairs of variants, the networks were integrated into a single system of networks. The largest network also included a quartet of variants: IMP-1, two single mutants, and the respective double mutant. The existence of quartets indicates that if two mutations resulted in functional enzymes, the double mutant may also be active and stable. Therefore, quartet construction from triplets was applied to predict 15 functional variants. Further functional mutants were predicted by applying the two ubiquitous mutations in all networks. In addition, since the networks are separated from each other by 10-15 mutations on average, it is expected that a subset of the theoretical intermediates are functional, and therefore are supposed to exist in the biosphere. Finally, the network analysis helps to distinguish between epistatic and additive effects of mutations; while the presence of correlated mutations indicates a strong interdependency between the respective positions, the mutations V67F and S262G are ubiquitous and therefore background independent.

(S)-Selective MenD variants from Escherichia coli provide access to new functionalized chiral α-hydroxy ketones.

We report the first rationally designed (S)-selective MenD from E. coli for the synthesis of functionalized α-hydroxy ketones. By mutation of two amino acids in the active site stereoselectivity of the (R)-selective EcMenD (ee > 93%) was inverted giving access to (S)-5-hydroxy-4-oxo-5-phenylpentanoate derivatives with stereoselectivities up to 97% ee.

Computational tools for rational protein engineering of aldolases.

In this mini-review we describe the different strategies for rational protein engineering and summarize the computational tools available. Computational tools can either be used to design focused libraries, to predict sequence-function relationships or for structure-based molecular modelling. This also includes de novo design of enzymes. Examples for protein engineering of aldolases and transaldolases are given in the second part of the mini-review.

The isoelectric region of proteins: a systematic analysis.

BACKGROUND: Binding of proteins in ion exchange chromatography is dominated by electrostatic interactions and can be tuned by adjusting pH and ionic strength of the solvent. Therefore, the isoelectric region (IER), the pH region of almost zero charge near the pI, has been used to predict the binding properties of proteins. PRINCIPAL FINDINGS: Usually the IER is small and binding and elution is carried out at pH values near to the pI. However, some proteins with an extended IER have been shown to bind and elute far away from its pI. To analyze factors that mediate the size of the IER and to identify proteins with an extended IER, two protein families consisting of more than 7000 proteins were systematically investigated. Most proteins were found to have a small IER and thus are expected to bind or elute near to their pI, while only a small fraction of less than 2% had a large IER. CONCLUSIONS: Only four factors, the number of histidines, the pI, the number of titratable amino acids and the ratio of acidic to basic residues, are sufficient to reliably classify proteins by their IER based on their sequence only, and thus to predict their binding and elution behaviour in ion exchange chromatography.