Kabat Database and its applications: future directions

The Kabat Database was initially started in 1970 to determine the combining site of antibodies based on the available amino acid sequences. The precise delineation of complementarity determining regions (CDR) of both light and heavy chains provides the first example of how properly aligned sequences can be used to derive structural and functional information of biological macromolecules. This knowledge has subsequently been applied to the construction of artificial antibodies with prescribed specificities, and to many other studies. The Kabat database now includes nucleotide sequences, sequences of T cell receptors for antigens (TCR), major histocompatibility complex (MHC) class I and II molecules, and other proteins of immunological interest. While new sequences are continually added into this database, we have undertaken the task of developing more analytical methods to study the information content of this collection of aligned sequences. New examples of analysis will be illustrated on a yearly basis. The Kabat Database and its applications are freely available at http://immuno.bme.nwu.edu.     

Amino acid sequence of rat liver cathepsin L

The complete amino acid sequences of the heavy and light chains of rat liver cathepsin L (EC 3.4.22.15) were determined at the protein level. The heavy and light chains consisted of 175 and 44 amino acid residues, respectively, and their Mr values without glycosyl groups calculated from these sequences were 18941 and 5056, respectively. The amino acid sequence was also determined from the N-terminal sequences of the heavy and light chains, and the sequences of cleavage fragments of the heavy chain with lysylendopeptidase and cyanogen bromide. The fragments were aligned by comparison with the amino acid sequence deduced from the sequence of cDNA of rat preprocathepsin L. The sequence of rat liver cathepsin L determined at the protein level was identical with that deduced from the cDNA sequence except that in the heavy chain, residues 176-177 (Asp-Ser) were not present at the C-terminus and alanine was replaced by proline at residue 125. Asn-108 in the heavy chain is modified with carbohydrate.     

Factor X activator from Vipera lebetina venom is synthesized from different genes

Vipera lebetina venom contains specific coagulant Factor X activator (VLFXA) that cleaves the Arg52-Ile53 bond in the heavy chain of human factor X. VLFXA is a glycoprotein that is composed of a heavy chain (HC) and two light chains (LC) linked by disulfide bonds. The complete amino acid sequences of the three chains of the factor X activator from V. lebetina snake venom are deduced from the nucleotide sequences of cDNAs encoding these chains. The full-length cDNA (2347 bp) sequence of the HC encodes an open reading frame (ORF) of 612 amino acids that includes signal peptide, propeptide and mature metalloproteinase with disintegrin-like and cysteine-rich domains. The light chain LC1 contains 123 and LC2 135 amino acid residues. Both light chains belong to the class of C-type lectin-like proteins. The N-termini of VLFXA chains and inner sequences of peptide fragments detected by liquid chromatography-electrospray ionization tandem mass spectrometry (LC MS/MS) from protein sequence are 100% identical to the sequences deduced from the cDNA. The molecular masses of tryptic fragments of VLFXA chains analyzed by matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS) also confirm the protein sequences deduced from the cDNAs. These are the first cloned factor X activator heavy and light chains. We demonstrate that the heavy and light chains are synthesized from different genes.     

Prediction of zinc-binding sites in proteins from sequence.

MOTIVATION: Motivated by the abundance, importance and unique functionality of zinc, both biologically and physiologically, we have developed an improved method for the prediction of zinc-binding sites in proteins from their amino acid sequences. RESULTS: By combining support vector machine (SVM) and homology-based predictions, our method predicts zinc-binding Cys, His, Asp and Glu with 75% precision (86% for Cys and His only) at 50% recall according to a 5-fold cross-validation on a non-redundant set of protein chains from the Protein Data Bank (PDB) (2727 chains, 235 of which bind zinc). Consequently, our method predicts zinc-binding Cys and His with 10% higher precision at different recall levels compared to a recently published method when tested on the same dataset. AVAILABILITY: The program is available for download at www.fos.su.se/~nanjiang/zincpred/download/     

Amino acid sequence of chitinase from Streptomyces erythraeus

The amino acid sequence of chitinase from Streptomyces erythraeus was determined by the conventional method. The amino acid sequences of tryptic peptides of the reduced and S-carboxymethylated protein were determined. The tryptic peptides were aligned by overlapping the amino acid sequences of chymotryptic peptides, lysyl endopeptidase peptides and cyanogen bromide fragments. S. erythraeus chitinase consists of 290 amino acid residues with the molecular weight of 30,400 and has two disulfide bridges at Cys(45)-Cys(89) and Cys(265)-Cys(272). The enzyme has no significant homology with other chitinases, lysozymes, and other proteins.     

Primary structure of potato kunitz-type serine proteinase inhibitor

The serine proteinase inhibitor (PSPI-21) isolated from potato tubers (Solanum tuberosum L.) comprises two protein species with pI 5.2 and 6.3, denoted as PSPI-21-5.2 and PSPI-21-6.3, respectively. They were separated by anion exchange chromatography on a Mono Q FPLC column. Both species tightly inhibit human leukocyte elastase, whereas their interaction with trypsin and chymotrypsin is substantially weaker. The sequences of both PSPI-21-5.2 and PSPI-21-6.3 were determined by analysis of overlapping peptides obtained from the oxidized or reduced and S-pyridylethylated proteins after digestion with trypsin or pepsin. Both species of PSPI-21 are composed of two chains, named chains A and B, which are linked by a disulfide bridge between Cys(146) and Cys(157). The other disulfide bridge is located within the A chains between Cys(48) and Cys(97). The amino acid sequences of the large A chains of the two forms, consisting of 150 amino acids residues each, differ in a single residue at position 52. The small chains B, containing 37 and 36 residues in PSPI-21-6.3 and PSPI-21-5.2, respectively, have nine different residues. The entire amino acid sequences of the two inhibitors show a high degree of homology to the other Kunitz-type proteinase inhibitors from plants.     

VL-VH expression by monoclonal antibodies recognizing avian lysozyme

Seven BALB/c hybridoma antibodies directed against the protein antigen, hen egg-white lysozyme c (HEL), were characterized on the basis of their ability to bind lysozymes from 10 species of birds, and their ability to bind HEL competitively. The hybridomas were separable into three complementation groups based upon competitive interactions. The fine specificities of all antibodies were distinct, but two, HyHEL-8 and HyHEL-10, had very similar and overlapping reactivity patterns. To test the hypothesis that VL-VH pairing correlates with binding specificity, the N-terminal amino acid sequences were determined to identify the VL and VH isotopes (subgroups) of the anti-HEL antibodies. HyHEL-8 and -10 shared the VK23 light chain isotype and nearly identical heavy chains in Kabat subgroup I, whereas the heavy and light chain isotypes of all other antibodies differed from HyHEL-8 and -10 and from each other. The heavy and light chain isotypes expressed by HyHEL-8 and -10 are also expressed by XRPC-25, a DNP-binding myeloma protein that does not bind lysozyme. These results are discussed with respect to the contributions of various genetic sources of structural diversity to antibody functional diversity.     

Molecular cloning and sequence analysis of aggretin, a collagen-like platelet aggregation inducer.

A cDNA library derived from the Malayan-pit-viper (Calloselasma rhodostoma) venom gland was constructed in the phagemid vector. Using the information of the N-terminal amino acid sequences of two subunits of aggretin, synthetic mixed-base oligonucleotides were employed as a screening probe for colony hybridization. Separate cDNA clones encoding for the alpha and beta chains of aggretin were isolated and sequenced. The results revealed that mature alpha and beta chains contain 136 and 123 amino acid residues, respectively. Aggretin subunits show high degrees of identity with respective subunits (50-60% for alpha, 49-58% for beta) of C-type lectin-like snake venoms. The identity to rattlesnake lectin is relatively lower (i.e., 39 and 30%). All cysteine residues in each chain of aggretin are well conserved and located at the positions corresponding to those of C-type lectins. Thus, three intracatenary disulfide bridges and an interchain disulfide bond between Cys83(alpha) and Cys75(beta) may be allocated. This is the first report regarding the entire sequence of venom GPIa/IIa agonist. According to the alignment of amino acid sequences, hypervariable regions among these C-type lectin-like proteins were revealed. These hypervariable regions are proposed to be the counterparts directly interacting with different receptors or different domains of a receptor on the surface of platelet.     

Amino acid sequences of the human kidney cathepsins H and L

The complete amino acid sequences of human kidney cathepsin H (EC 3.4.22.16) and human kidney cathepsin L (EC 3.4.22.15) were determined. Cathepsin H contains 230 residues and has an Mr of 25116. The sequence was obtained by sequencing the light, heavy and mini chain and the peptides produced by cyanogen bromide cleavage of the single-chain form of the enzyme. The glycosylated mini chain is a proteolytic fragment of the propeptide of cathepsin H. Human cathepsin L has 217 amino acid residues and an Mr of 23720. Its amino acid sequence was deduced from N-terminal sequences of the heavy and light chains and from the sequences of cyanogen bromide fragments of the heavy chain. The fragments were aligned by comparison with known sequences of cathepsins H and L from other species. Cathepsins H and L exhibit a high degree of sequence homology to cathepsin B (EC 3.4.22.1) and other cysteine proteinases of the papain superfamily.     

Construction of TNF-binding proteins by grafting hypervariable regions of F10 antibody on human fibronectin domain scaffold

Tumor necrosis factor (TNF) plays a key role in the pathogenesis of various diseases. To study the possibility of constructing TNF-binding proteins by grafting hypervariable regions of immunoglobulins (CDR), we have replaced amino acid sequences of loops from the tenth type III domain of human fibronectin (¹⁰Fn3) by amino acid sequences of CDR from the light and heavy chains of the anti-TNF antibody F10. The assessment of TNF-binding properties of the resulting proteins by ELISA has revealed the highest activity of Hd3 containing sequences CDR-H1 and CDR-H2 of the antibody F10 and of Hd2 containing sequences CDR-H1 and CDR-H3. The proteins constructed by us on the fibronectin domain scaffold specifically bound TNF during Western blotting and also weakened its cytotoxic effect on L929 line cells. The highest neutralizing activity was demonstrated by the proteins Hd2 and Hd3, which induced, respectively, 10- and 50-fold increase in the EC₅₀ of TNF.     

Botulinum neurotoxin type B (strain 657): partial sequence and similarity with tetanus toxin

The type B neurotoxin (NT) isolated from Clostridium botulinum (strain 657) behaved as a mixture of single (unnicked) and dichain (nicked) proteins, both of Mr approximately 150 kDa. When the dichain NT was reduced by mercaptoethanol, the two chains migrated in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as separate polypeptides of Mr approximately 100 and 50 kDa that appeared similar to the heavy and light chains of other serotypes of botulinum NT. The N-terminal amino acid sequences of the two chains were determined. They were as follows: light chain: Pro-Val-Thr-Ile-Asn-Asn-Phe-Asn-Tyr-Asn-Asp-Pro-Ile-Asp-Asn-Asn-Asn-Ile- Ile-Met - Met-Glu-Pro-Pro-Phe-Ala-Arg-Gly-Met-Gly-Arg-Tyr-Tyr-Lys-Ala-Phe-Lys-Ile- Thr-Asp - Arg-Ile-Trp-Ile-; and heavy chain: Ala-Pro-Gly-Ile-X-Ile-Asp-Val-Asp-Asn-Glu-Asp-Leu-Phe-Phe-Ile-Ala-Asp-Ly s-Asn- Ser-Phe-Arg-Asp-Asp-Leu-. These two sequences matched exactly with those of the light and heavy chains of type B NT (strain Okra) of which only 16 and 18 residues were known (J. Biol. Chem. (1985) 260, 10461). The above sequences were different from those of type A NT. Immunoprecipitation reactions of type B NT isolated from strains 657 and Okra were indistinguishable against polyclonal anti-type B NT serum. These two preparations did not produce precipitin reactions with polyclonal anti-type A NT serum.(ABSTRACT TRUNCATED AT 250 WORDS)     

Separation, purification, partial characterization and comparison of the heavy and light chains of botulinum neurotoxin types A, B, and E

Clostridium botulinum produces botulinum neurotoxin (NT) in antigenically distinct forms. When isolated from bacterial cultures type E is a single chain, type B is a mixture of single and two-chain molecules, and type A is essentially a two-chain molecule (Mr approximately 150,000). Protease(s) in the cultures or trypsin nick single-chain NT to the two-chain form. The heavy (Mr approximately 100,000) and light (Mr approximately 50,000) chains of the two-chain molecule remain held together by -S-S-bond(s). The two chains are presumed to have different functions. NT binds to nerve cells via the heavy chain and then light chain enters the cell and blocks release of acetylcholine (Simpson, L. L. (1981) Pharmacol. Rev. 33, 155-188). We nicked single-chain NT to form the two-chain form with trypsin, minimizing secondary cleavages, then separated and purified the heavy and light chains using ion-exchange chromatography. The technique, with minor modifications, is a generalized method for types A, B, and E. These subunit chains (each a single band in sodium dodecyl sulfatepolyacrylamide gel electrophoresis) were analyzed for their complete amino acid compositions. The amino acid contents of the heavy and light chains agreed well with the parent two-chain molecule. This affirms that NT is composed of two chains. The two subunit chains are now usable for amino acid sequence and other studies. Comparison of the amino acid contents indicates more similarity among the light chains than the heavy chains of the three NT types, a similarity that agrees with our published partial amino acid sequences (first 13-18 residues) of these chains. Several (up to 9) different amino acid residues of the heavy chain (which is twice the size of the light chain) are present in double the number of corresponding residues in the light chain.     

Subunit structures of three human myeloperoxidases

The heavy and the light subunits of human myeloperoxidase (donor: H2O2 oxidoreductase [EC 1.11.1.7]) I, II, and III were isolated from the reduced and S-carboxymethylated enzymes. These three enzymes have the same terminal amino acid sequences and similar chemical compositions in both subunits. The NH2-terminal sequences of the heavy and light subunits were determined to be Val-Asn-Cys-Glu-Thr- and Thr-Cys-Pro-Glu-Gln-, respectively; a heterogeneity was observed in the NH2-termini of the latter subunits for the three enzymes. As for COOH-termini, the sequences -(Asn, 2 Leu, Ala, Ser, Trp)-Arg-Glu-Ala and -Ala-Arg were obtained for the heavy and the light subunits, respectively. The heavy subunits contained 8-10 mol/mol of glucosamine. On the basis of these results and the amino acid sequence deduced from cDNA clones, the heavy subunits probably correspond to amino acids 279-744 and the light subunits to amino acids (164-167)-272. For the heavy subunits, Ser-745, which was predicted as the COOH-terminal amino acid from the nucleotide sequence, was removed. The light subunits were also processed at their COOH-termini by 6 residues. Four or five high mannose type carbohydrate chains were attached to the heavy subunits.     

Identification and location of a cysteinyl posttranslational modification in an amyloidogenic kappa1 light chain protein by electrospray ionization and matrix-assisted laser desorption/ionization mass spectrometry

Amyloid-deposited light chain (AL) amyloidosis is correlated with the overproduction of a monoclonal immunoglobulin light chain protein by a B-lymphocyte clone. Since the amyloid fibril deposits in AL amyloidosis most often consist of the N-terminal fragments of the light chain, the majority of studies have focused on the determination of the primary structure of the protein, and reducing agents have been used routinely in the initial purification process. In this study, two light chain proteins were isolated and purified, without reduction, from the urine of a patient diagnosed with kappa 1 (kappa1) AL amyloidosis. One protein had a relative molecular mass of 12,000 and the other 24,000. Electrospray ionization and matrix-assisted laser desorption/ionization mass spectrometry, in combination with enzymatic digestions, were used to verify the amino acid sequences and identify and locate posttranslational modifications in these proteins. The 12-kDa protein was confirmed to be the N-terminal kappa1 light chain fragment (variable region) consisting of residues 1-108 or 1-109 and having one disulfide bond. The 24-kDa protein was determined to be the intact kappa1 light chain containing a cysteinyl posttranslational modification at Cys214 and disulfide bonds located at Cys23-Cys88, Cys134-Cys194, and Cys214-Cys. The methods used in this report enable high-sensitivity determination of amino acid sequence and variation in intact and truncated light chains as well as posttranslational modifications. This approach facilitates consideration of the effect of cysteinylation on the native protein structure and the potential involvement of this modification in AL amyloidosis.     

Partial amino acid sequence of porcine elastase II. Active site and the activation peptide regions

The partial amino acid sequence of porcine elastase II, isolated from crude trypsin Type II, was determined. The enzyme consists of two polypeptide chains, a light chain composed of 11 residues, and a heavy chain (Mr = 23 500) with four intrachain disulfide bridges; the two chains are held together by one interchain disulfide bond. Elastase II was fragmented into several peptides by chemical cleavages with CNBr at the two methionine residues, 99 and 180 (chymotrypsinogen numbering), and with hydroxylamine at the peptide bond following DIP-Ser195. About 50% of the sequence was determined and the positions of 120 amino acids were located, including the light chain residues and most of the active site residues. The partial sequence shows 64% difference between porcine elastase II and elastase I and only 26% difference between porcine elastase II and bovine chymotrypsin B.     

Amino-terminal amino acid sequence and heterogeneity in glycosylation of rat renal renin

We isolated 7.4 mg of pure renin from 2 kg of rat kidneys using affinity chromatography on pepstatin-aminohexyl-Sepharose and an octapeptide renin inhibitor, H-77-Sepharose. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed that renin consists of two polypeptide chains linked by a disulfide bond, one of Mr = 36,000 (heavy chain) and the other of Mr = 3,000 (light chain). The amino-terminal 10-amino acid sequences of the heavy and the light chains were identical to the sequences beginning at Ser72 and Asp355, respectively, of the amino acid sequence of preprorenin deduced from the renin cDNA sequence. Amino acid sequencing of the carboxyl-terminal peptide of the heavy chain, generated by digestion with lysyl endopeptidase, showed that the carboxyl-terminal residue of the heavy chain is Phe. Thus, the propeptide of prorenin is cleaved after Thr71, followed by removal of two amino acids, Arg353 and Asn354, the result being formation of the heavy and light chains. Thus, the site of cleavage of rat prorenin is after a nonbasic amino acid, in contrast to the cleavage of the propeptide after a pair of basic amino acids in mouse submaxillary renin, human renal renin, and many secretory proteins. Treatment of renin with neuraminidase or glycopeptidase F had no apparent effect on the charge heterogeneity of renin. Glycosylation probably does not contribute to charge heterogeneity.     

IMGT standardized criteria for statistical analysis of immunoglobulin V-REGION amino acid properties

IMGT, the international ImMunoGeneTics information system(R) (http://imgt.cines.fr) is a high-quality integrated information system specializing in immunoglobulins (IG), T cell receptors (TR) and major histocompatibility complex (MHC) of human and other vertebrates. IMGT comprises IMGT/LIGM-DB, the comprehensive database of IG and TR sequences from human and other vertebrates (76 846 sequences in September 2003). In order to define the IMGT criteria necessary for standardized statistical analyses, the sequences of the IG variable regions (V-REGIONs) from productively rearranged human IG heavy (IGH) and IG light kappa (IGK) and lambda (IGL) chains were extracted from IMGT/LIGM-DB. The framework amino acid positions of 2474 V-REGIONs (1360 IGHV, 585 IGKV, 529 IGLV) were numbered according to the IMGT unique numbering. Two statistical methods (correspondence analysis and hierarchic classification) were used to analyze the 237 framework positions (80 for IGHV, 79 for IGKV, 78 for IGLV), for three properties (hydropathy, volume and chemical characteristics) of the 20 common amino acids. Results of the analyses are shown as standardized two-dimensional representations, designated as IMGT Colliers de Perles statistical profiles. They provide a characterization of the amino acid properties at each framework position of the expressed IG V-REGIONs, and a visualization of the resemblances and differences between heavy and light, and between kappa and lambda sequences. The standardized criteria defined in this paper, amino acid positions and property classes, will be useful to study the mutations and allele polymorphisms, to establish correlations between amino acids in the IG and TR protein three-dimensional structures and to extract new knowledge from V-like domains of chains, other than IG and TR, belonging to the immunoglobulin superfamily.     

The primary structure of human gamma-glutamyl transpeptidase

A cDNA hybridizable to that of rat gamma-glutamyl transpeptidase (GGT) was cloned from a cDNA library of human fetal liver. The insert of the cDNA clone contained 1866 bp consisting of an open reading frame (ORF) of 1709 bp (569 amino acids (aa), N-terminal portion truncated) and a 135-bp 3'-untranslated region followed by a polyadenylated tail. In parallel, amino acid sequences of N-terminal portions of heavy and light chains of a purified human GGT were determined. Two stretches of amino acid sequences identical to the N-terminal sequences of heavy and light chains were found in the ORF. We therefore concluded that the clone is a cDNA for human GGT. From the amino acid sequence deduced from cDNA, the heavy and the light chains of the purified enzyme are estimated to be composed of 351 aa (Mr 38,336) and of 189 aa (Mr 20,000), respectively. The heavy chain is preceded by a signal peptide of at least 29 aa presumed to be cleaved by bromelain treatment. Six putative N-glycosylation sites are present in the heavy subunit region and one in the light subunit region. Primary structure and hydrophobicity profile are closely similar to those of rat GGT.     

Evolution of RNA-dependent RNA polymerases of positive riboviruses

A comparative analysis is presented of 24 known amino acid sequences of RNA-dependent RNA polymerases of positive strand RNA viruses infecting animals, plants and bacteria. Using a newly proposed methodology of group alignment for weakly similar sequences, evolutionary conserved fragments of all these proteins were unambiguously aligned. A unique pattern (consensus) of 7 invariant amino acid residues was revealed which is absent from the sequences of other RNA and DNA polymerases and is thought to unequivocally identify the RNA-dependent RNA polymerases of positive strand RNA viruses. Based on the obtained alignment a tentative phylogenetic tree of viral RNA polymerases was constructed for the first time. The RNA-dependent RNA polymerases of positive strand RNA viruses are concluded to comprise a distinct family of evolutionary related proteins.