The ubiquity of the insulin superfamily across the eukaryotes detected using a bioinformatics approach

The insulin superfamily is composed of a diverse group of proteins that share a common structural design whose most notable feature is a set of disulfide bonds. There is now sufficient experimental and bioinformatics evidence that it is represented in at least a number of well-investigated invertebrates, where they have been found to intervene mainly in complex processes such as mitosis, cell growth, castes differentiation, and fertility. In this article we automated a methodology first proposed elsewhere-that combines sequence similarity with assessing membership to the superfamily by conservation of structuraly key residues-to identify putative insulin-like peptides (ILPs) in completely sequenced genomes, and applied it as a pipeline to a group of 46 organisms both vertebrates and invertebrates. As a result, we were able to identify 1,653 putative members of the insulin superfamily, from 17 putative members in C. savigny to 58 in X. tropicalis. Moreover, we found that structural distinctions-such as peptides length-between functionally diverse members of the superfamily found in vertebrates, that is, insulins, IGFs, and relaxins, are not equally represented in invertebrates genomes, suggesting that such divergence has occurred only recently in the evolutionary history of vertebrates.     

Identification of a new subfamily of HNH nucleases and experimental characterization of a representative member, HphI restriction endonuclease

The restriction endonuclease (REase) R. HphI is a Type IIS enzyme that recognizes the asymmetric target DNA sequence 5'-GGTGA-3' and in the presence of Mg(2+) hydrolyzes phosphodiester bonds in both strands of the DNA at a distance of 8 nucleotides towards the 3' side of the target, producing a 1 nucleotide 3'-staggered cut in an unspecified sequence at this position. REases are typically ORFans that exhibit little similarity to each other and to any proteins in the database. However, bioinformatics analyses revealed that R.HphI is a member of a relatively big sequence family with a conserved C-terminal domain and a variable N-terminal domain. We predict that the C-terminal domains of proteins from this family correspond to the nuclease domain of the HNH superfamily rather than to the most common PD-(D/E)XK superfamily of nucleases. We constructed a three-dimensional model of the R.HphI catalytic domain and validated our predictions by site-directed mutagenesis and studies of DNA-binding and catalytic activities of the mutant proteins. We also analyzed the genomic neighborhood of R.HphI homologs and found that putative nucleases accompanied by a DNA methyltransferase (i.e. predicted REases) do not form a single group on a phylogenetic tree, but are dispersed among free-standing putative nucleases. This suggests that nucleases from the HNH superfamily were independently recruited to become REases in the context of RM systems multiple times in the evolution and that members of the HNH superfamily may be much more frequent among the so far unassigned REase sequences than previously thought.     

Expression, purification, and functional characterization of a stable helicase domain from a tomato mosaic virus replication protein

Tomato mosaic virus (genus, Tobamovirus) is a member of the alphavirus-like superfamily of positive-strand RNA viruses, which include many plant and animal viruses of agronomical and clinical importance. The RNA of alphavirus-like superfamily members encodes replication-associated proteins that contain a putative superfamily 1 helicase domain. To date, a viral three-dimensional superfamily 1 helicase structure has not been solved. For the study reported herein, we expressed tomato mosaic virus replication proteins that contain the putative helicase domain and additional upstream N-terminal residues in Escherichia coli. We found that an additional 155 residues upstream of the N-terminus of the helicase domain were necessary for stability. We developed an efficient procedure for the expression and purification of this fragment and have examined factors that affect its stability. Finally, we also showed that the stable fragment has nucleoside 5'-triphosphatase activity.     

The drug/metabolite transporter superfamily.

Previous work defined several families of secondary active transporters, including the prokaryotic small multidrug resistance (SMR) and rhamnose transporter (RhaT) families as well as the eukaryotic organellar triose phosphate transporter (TPT) and nucleotide-sugar transporter (NST) families. We show that these families as well as several other previously unrecognized families of established or putative secondary active transporters comprise a large ubiquitous superfamily found in bacteria, archaea and eukaryotes. We have designated it the drug/metabolite transporter (DMT) superfamily (transporter classification number 2.A.7) and have shown that it consists of 14 phylogenetic families, five of which include no functionally well-characterized members. The largest family in the DMT superfamily, the drug/metabolite exporter (DME) family, consists of over 100 sequenced members, several of which have been implicated in metabolite export. Each DMT family consists of proteins with a distinctive topology: four, five, nine or 10 putative transmembrane alpha helical spanners (TMSs) per polypeptide chain. The five TMS proteins include an N-terminal TMS lacking the four TMS proteins. The full-length proteins of 10 putative TMSs apparently arose by intragenic duplication of an element encoding a primordial five-TMS polypeptide. Sequenced members of the 14 families are tabulated and phylogenetic trees for all the families are presented. Sequence and topological analyses allow structural and functional predictions.     

Identification of putative zinc hydrolase genes of the metallo-beta-lactamase superfamily from Campylobacter jejuni

DNA fragments encoding two putative zinc-dependent hydrolases, designated GLX2-1 and GLX2-2, from a clinical isolate of Campylobacter jejuni, strain 012, were cloned and sequenced. GLX2-1 was encoded by a sequence of 798 bp and GLX2-2 by a sequence of 597 bp. The amino acid sequences deduced from C. jejuni DNA showed 99% and 100% identity, respectively, to putative zinc hydrolases reported from C. jejuni ATCC strain 11168, and also shared identity (28-43%) with several hypothetical conserved proteins and known zinc-dependent hydrolases and metallo-beta-lactamase superfamily proteins. A strictly conserved motif, -H-X-H-X-D-, characteristic of the metallo-beta-lactamase superfamily of proteins, including class B metallo-beta-lactamases, was identified in both proteins. Other conserved metal-binding ligands, characteristic of the metallo-beta-lactamase superfamily of proteins, were also identified. Functional beta-lactamase could not be expressed in either Escherichia coli or Campylobacter coli transformed with C. jejuni hydrolase-containing plasmids, suggesting that they do not function as metallo-beta-lactamases, although structurally they are consistent with the zinc metallo-hydrolase family of the beta-lactamase fold.     

Development of a high-throughput cloning strategy for characterization of Acinetobacter baumannii drug transporter proteins

Heterologous expression of membrane proteins in Escherichia coli often requires optimization to overcome problems with toxicity of the recombinant protein to the host cell. A number of Gateway-based destination vectors were constructed to investigate expression of membrane proteins using a high-throughput approach. These vectors were tested using putative drug transporter proteins from the multidrug and toxic compound extrusion (MATE) family and the resistance-nodulation-cell division superfamily encoded by the human pathogen Acinetobacter baumannii. Active transport of antibiotics and antiseptics mediated by efflux proteins contributes to the high level of multidrug resistance observed in A. baumannii. Substrates for 4 of the 5 putative efflux proteins investigated were identified using the expression vectors designed in this study. Additionally, a Gateway-based suicide vector was designed for construction of specific A. baumannii insertion disruption mutants. This knockout cloning strategy was tested and shown to be successful in inactivating AbeM4, a putative MATE family protein. Therefore, we have shown that the Gateway-based vectors constructed in this study are versatile tools that can be used for manipulation and characterization of membrane proteins.     

Novel Predicted Peptidases with a Potential Role in the Ubiquitin Signaling Pathway

A multi-pronged strategy including extensive sequence searches, structuralmodeling, and analysis of contextual information extracted from domainarchitectures, genetic screens, and large-scale protein-protein interaction analyseswas employed to predict previously undetected components of the eukaryoticubiquitin signaling system. Two novel groups of proteins that are likely to function asde-ubiquitinating and de-SUMOylating peptidases (DUBs) were identified. The firstgroup of putative DUBs, designated PPPDE superfamily (after Permuted Papain foldPeptidases of DsRNA viruses and Eukaryotes), consists of predicted thiol peptidaseswith a circularly permuted papain-like fold. The inference of the likely DUB functionof the PPPDE superfamily proteins is based on the fusions of the catalytic domain toUb-binding PUG (PUB)/UBA domains and a novel alpha-helical Ub-associated domain(the PLAP, Ufd3p and Lub1p or PUL domain) amongst different members of thePPPDE supefamily. The presence of the PPPDE superfamily proteins in mosteukaryotic lineages, including basal ones, such as Giardia, suggest a role indeubiquitination of highly conserved proteins involved in key cellular functions, suchas cell cycle control. In addition to eukaryotic proteins, the PPPDE superfamilyincludes predicted proteases from several groups of double-stranded RNA virusesand one single-stranded DNA virus. The apparent recruitment of DUBs for viralpolyprotein processing seems to represent a common theme in evolution of viruses.The second group of putative DUBs identified in this study is the WLM (Wss1p-likemetalloproteases) family of Zincin-like superfamily of Zn-dependent peptidases,which are linked to the Ub -system by virtue of fusions with the UB-binding PUG(PUB), ubiquitin-like and Little Finger domains. More specifically on the basis ofgenetic evidence the WLM family is implicated in de-SUMOylation. If validatedexperimentally, the WLM family proteins will represent the first case of a Zincin-likemetalloprotease involvement in Ub-signaling.     

Structural and sequence analysis of imelysin-like proteins implicated in bacterial iron uptake

Imelysin-like proteins define a superfamily of bacterial proteins that are likely involved in iron uptake. Members of this superfamily were previously thought to be peptidases and were included in the MEROPS family M75. We determined the first crystal structures of two remotely related, imelysin-like proteins. The Psychrobacter arcticus structure was determined at 2.15 Å resolution and contains the canonical imelysin fold, while higher resolution structures from the gut bacteria Bacteroides ovatus, in two crystal forms (at 1.25 Å and 1.44 Å resolution), have a circularly permuted topology. Both structures are highly similar to each other despite low sequence similarity and circular permutation. The all-helical structure can be divided into two similar four-helix bundle domains. The overall structure and the GxHxxE motif region differ from known HxxE metallopeptidases, suggesting that imelysin-like proteins are not peptidases. A putative functional site is located at the domain interface. We have now organized the known homologous proteins into a superfamily, which can be separated into four families. These families share a similar functional site, but each has family-specific structural and sequence features. These results indicate that imelysin-like proteins have evolved from a common ancestor, and likely have a conserved function.     

Complete inventory of ABC proteins in human pathogenic yeast, Candida albicans

The recent completion of the sequencing project of the opportunistic human pathogenic yeast, Candida albicans (http://www.ncbi.nlm.nih.gov/), led us to analyze and classify its ATP-binding cassette (ABC) proteins, which constitute one of the largest superfamilies of proteins. Some of its members are multidrug transporters responsible for the commonly encountered problem of antifungal resistance. TBLASTN searches together with domain analysis identified 81 nucleotide-binding domains, which belong to 51 different putative open reading frames. Considering that each allelic pair represents a single ABC protein of the Candida genome, the total number of putative members of this superfamily is 28. Domain organization, sequence-based analysis and self-organizing map-based clustering led to the classification of Candida ABC proteins into 6 distinct subfamilies. Each subfamily from C. albicans has an equivalent in Saccharomyces cerevisiae suggesting a close evolutionary relationship between the two yeasts. Our searches also led to the identification of a new motif to each subfamily in Candida that could be used to identify sequences from the corresponding subfamily in other organisms. It is hoped that the inventory of Candida ABC transporters thus created will provide new insights into the role of ABC proteins in antifungal resistance as well as help in the functional characterization of the superfamily of these proteins.     

Characterization of amalgam: a member of the immunoglobulin superfamily from Drosophila

The immunoglobulin superfamily is a diverse group of proteins that are involved in various aspects of cell surface recognition. Here, we report the characterization of amalgam (ama), a gene in the Antennapedia complex (ANT-C) of D. melanogaster that exhibits amino acid similarity to vertebrate neural cell adhesion molecules and other members of the immunoglobulin superfamily. The putative 333 amino acid ama protein consists of a signal sequence, three immunoglobulin-like domains, and a short slightly hydrophobic carboxy-terminal region. Antibodies against the ama protein reveal that it accumulates on the surface of various mesodermal and neural cells during embryogenesis. The function of this protein remains elusive, as no mutations have been recovered for ama during saturation EMS mutagenesis of this chromosomal region.     

The peroxidase-cyclooxygenase superfamily: Reconstructed evolution of critical enzymes of the innate immune system

The authors have reconstructed the phylogenetic relationships of the main evolutionary lines of mammalian heme containing peroxidases. The sequences of intensively investigated human myeloperoxidase, eosinophil peroxidase, and lactoperoxidase, which participate in host defence against infections, were aligned together with newly found open reading frames coding for highly similar putative peroxidase domains in all kingdoms of life. The evolutionary relationships were reconstructed using neighbor-joining, maximum parsimony, and maximum likelihood methods. It is demonstrated that this enzyme superfamily obeys the rules of birth-and-death model of multigene family evolution and contains proteins with a variety of function that could be grouped in seven subfamilies. On the basis of occurrence and the fact that two main enzymatic activities are related with these metalloproteins, they propose the name peroxidase-cyclooxygenase superfamily for this widely spread group of heme-containing oxidoreductases. Well known structure-function relationships in mammalian peroxidases formed the basis for the critical inspection of all subfamilies. The presented data unequivocally suggest that predecessor genes of mammalian heme peroxidases have segregated very early in evolution. Before organisms developed an acquired immunity, their antimicrobial defence depended on enzymes that were recruited upon pathogen invasion and could produce antimicrobial reaction products. Thus, these peroxidatic heme proteins evolved to important components in the innate immune defence system. This work shows that even in certain prokaryotic organisms, genes encoding putative antimicrobial enzymes are found providing a group of bacteria with an evolutionary advantage over the others.     

The diversity of algal phospholipase D homologs revealed by biocomputational analysis

Phospholipase D (PLD) participates in the formation of phosphatidic acid, a precursor in glycerolipid biosynthesis and a second messenger. PLDs are part of a superfamily of proteins that hydrolyze phosphodiesters and share a catalytic motif, HxKxxxxD, and hence a mechanism of action. Although HKD‐PLDs have been thoroughly characterized in plants, animals and bacteria, very little is known about these enzymes in algae. To fill this gap in knowledge, we performed a biocomputational analysis by means of HMMER iterative profiling, using most eukaryotic algae genomes available. Phylogenetic analysis revealed that algae exhibit very few eukaryotic‐type PLDs but possess, instead, many bacteria‐like PLDs. Among algae eukaryotic‐type PLDs, we identified C2‐PLDs and PXPH‐like PLDs. In addition, the dinoflagellate Alexandrium tamarense features several proteins phylogenetically related to oomycete PLDs. Our phylogenetic analysis also showed that algae bacteria‐like PLDs (proteins with putative PLD activity) fall into five clades, three of which are novel lineages in eukaryotes, composed almost entirely of algae. Specifically, Clade II is almost exclusive to diatoms, whereas Clade I and IV are mainly represented by proteins from prasinophytes. The other two clades are composed of mitochondrial PLDs (Clade V or Mito‐PLDs), previously found in mammals, and a subfamily of potentially secreted proteins (Clade III or SP‐PLDs), which includes a homolog formerly characterized in rice. In addition, our phylogenetic analysis shows that algae have non‐PLD members within the bacteria‐like HKD superfamily with putative cardiolipin synthase and phosphatidylserine/phosphatidylglycerophosphate synthase activities. Altogether, our results show that eukaryotic algae possess a moderate number of PLDs that belong to very diverse phylogenetic groups.     

The Arabidopsis thaliana ABC protein superfamily, a complete inventory

We describe the first complete inventory of ATP-binding cassette (ABC) proteins from a multicellular organism, the model plant Arabidopsis thaliana. By the application of several search criteria, Arabidopsis was found to contain a total of 129 open reading frames (ORFs) capable of encoding ABC proteins, of which 103 possessed contiguous transmembrane spans and were identified as putative intrinsic membrane proteins. Fifty-two of the putative intrinsic membrane proteins contained at least two transmembrane domains (TMDs) and two nucleotide-binding folds (NBFs) and could be classified as belonging to one of five subfamilies of full-molecule transporters. The other 51 putative membrane proteins, all of which were half-molecule transporters, fell into five subfamilies. Of the remaining ORFs identified, all of which encoded proteins lacking TMDs, 11 could be classified into three subfamilies. There were no obvious homologs in other organisms for 15 of the ORFs which encoded a heterogeneous group of non-intrinsic ABC proteins (NAPs). Unrooted phylogenetic analyses substantiated the subfamily designations. Notable features of the Arabidopsis ABC superfamily was the presence of a large yeast-like PDR subfamily, and the absence of genes encoding bona fide cystic fibrosis transmembrane conductance regulator (CFTR), sulfonylurea receptor (SUR), and heavy metal tolerance factor 1 (HMT1) homologs. Arabidopsis was unusual in its large allocation of ORFs (a minimum of 0.5%) to members of the ABC protein superfamily.     

Large-scale, lineage-specific expansion of a bric-a-brac/tramtrack/broad complex ubiquitin-ligase gene family in rice

Selective ubiquitination of proteins is directed by diverse families of ubiquitin-protein ligases (or E3s) in plants. One important type uses Cullin-3 as a scaffold to assemble multisubunit E3 complexes containing one of a multitude of bric-a-brac/tramtrack/broad complex (BTB) proteins that function as substrate recognition factors. We previously described the 80-member BTB gene superfamily in Arabidopsis thaliana. Here, we describe the complete BTB superfamily in rice (Oryza sativa spp japonica cv Nipponbare) that contains 149 BTB domain-encoding genes and 43 putative pseudogenes. Amino acid sequence comparisons of the rice and Arabidopsis superfamilies revealed a near equal repertoire of putative substrate recognition module types. However, phylogenetic comparisons detected numerous gene duplication and/or loss events since the rice and Arabidopsis BTB lineages split, suggesting possible functional specialization within individual BTB families. In particular, a major expansion and diversification of a subset of BTB proteins containing Meprin and TRAF homology (MATH) substrate recognition sites was evident in rice and other monocots that likely occurred following the monocot/dicot split. The MATH domain of a subset appears to have evolved significantly faster than those in a smaller core subset that predates flowering plants, suggesting that the substrate recognition module in many monocot MATH-BTB E3s are diversifying to ubiquitinate a set of substrates that are themselves rapidly changing. Intriguing possibilities include pathogen proteins attempting to avoid inactivation by the monocot host.     

ST7 is a novel low-density lipoprotein receptor-related protein (LRP) with a cytoplasmic tail that interacts with proteins related to signal transduction pathways

In 1997, McCormick and co-workers identified a novel putative tumor suppressor gene, designated ST7, encoding a unique protein with transmembrane receptor characteristics [Qing et al. (1999) Oncogene 18, 335-342]. Using degenerate primers corresponding to the highly conserved region of the ligand-binding domains of members of the low-density lipoprotein receptor (LDLR) superfamily, Ishii et al. [Genomics (1998) 51, 132-135] discovered a low-density lipoprotein receptor-related protein (LRP) that closely resembles ST7. Later, another LRP closely resembling ST7 and LRP3 was found (murine LRP9) [Sugiyama et al. (2000) Biochemistry 39, 15817-15825]. These results strongly suggested that ST7 was also a novel member of the low-density lipoprotein receptor superfamily. Proteins of this superfamily have been shown to function in endocytosis and/or signal transduction. To evaluate the relationship of ST7 to the LDLR superfamily proteins and to determine whether ST7 may function in endocytosis and/or signal transduction, we used proteomic tools to analyze the functional motifs present in the protein. Our results indicate that ST7 is a member of a subfamily of the LDLR superfamily and that its cytoplasmic domain contains several motifs implicated in endocytosis and signal transduction. Use of the yeast two-hybrid system to identify proteins that associate with ST7's cytoplasmic domain revealed that this domain interacts with three proteins involved in signal transduction and/or endocytosis, viz., receptor for activated protein C kinase 1 (RACK1), muscle integrin binding protein (MIBP), and SMAD anchor for receptor activation (SARA), suggesting that ST7, like other proteins in the LDLR superfamily, functions in these two pathways. Clearly, ST7 is an LRP, and therefore, it should now be referred to as LRP12.     

Probing the substrate specificities of human PHOSPHO1 and PHOSPHO2

PHOSPHO1, a phosphoethanolamine/phosphocholine phosphatase, is upregulated in mineralising cells and is thought to be involved in the generation of inorganic phosphate for bone mineralisation. PHOSPHO2 is a putative phosphatase sharing 42% sequence identity with PHOSPHO1. Both proteins contain three catalytic motifs, conserved within the haloacid dehalogenase superfamily. Mutation of Asp32 and Asp203, key residues within two motifs, abolish PHOSPHO1 activity and confirm it as a member of this superfamily. We also show that Asp43 and Asp123, residues that line the substrate-binding site in our PHOSPHO1 model, are important for substrate hydrolysis. Further comparative modelling reveals that the active sites of PHOSPHO1 and PHOSPHO2 are very similar, but surprisingly, recombinant PHOSPHO2 hydrolyses phosphoethanolamine and phosphocholine relatively poorly. Instead, PHOSPHO2 shows high specific activity toward pyridoxal-5-phosphate (V(max) of 633 nmol min(-1) mg(-1) and K(m) of 45.5 microM). Models of PHOSPHO2 and PHOSPHO1 suggest subtle differences in the charge distributions around the putative substrate entry site and in the location of potential H-bond donors.