Acyl carrier protein structural classification and normal mode analysis

All acyl carrier protein primary and tertiary structures were gathered into the ThYme database. They are classified into 16 families by amino acid sequence similarity, with members of the different families having sequences with statistically highly significant differences. These classifications are supported by tertiary structure superposition analysis. Tertiary structures from a number of families are very similar, suggesting that these families may come from a single distant ancestor. Normal vibrational mode analysis was conducted on experimentally determined freestanding structures, showing greater fluctuations at chain termini and loops than in most helices. Their modes overlap more so within families than between different families. The tertiary structures of three acyl carrier protein families that lacked any known structures were predicted as well.     

Structural classification and properties of ketoacyl synthases

Ketoacyl synthases (KSs) catalyze condensing reactions combining acyl-CoA or acyl-acyl carrier protein (acyl-ACP) with malonyl-CoA to form 3-ketoacyl-CoA or with malonyl-ACP to form 3-ketoacyl-ACP. In each case, the resulting acyl chain is two carbon atoms longer than before, and CO2 and either CoA or ACP are formed. KSs also join other activated molecules in the polyketide synthesis cycle. Our classification of KSs by their primary and tertiary structures instead of by their substrates and the reactions that they catalyze enhances insights into this enzyme group. KSs fall into five families separated by their characteristic primary structures, each having members with the same catalytic residues, mechanisms, and tertiary structures. KS1 members, overwhelmingly named 3-ketoacyl-ACP synthase III or its variants, are produced predominantly by bacteria. Members of KS2 are mainly produced by plants, and they are usually long-chain fatty acid elongases/condensing enzymes and 3-ketoacyl-CoA synthases. KS3, a very large family, is composed of bacterial and eukaryotic 3-ketoacyl-ACP synthases I and II, often found in multidomain fatty acid and polyketide synthases. Most of the chalcone synthases, stilbene synthases, and naringenin-chalcone synthases in KS4 are from eukaryota. KS5 members are all from eukaryota, most are produced by animals, and they are mainly fatty acid elongases. All families except KS3 are split into subfamilies whose members have statistically significant differences in their primary structures. KS1 through KS4 appear to be part of the same clan. KS sequences, tertiary structures, and family classifications are available on the continuously updated ThYme (Thioester-active enzYme) database.     

Evolutionary origin and functional diversification of aminotransferases

Aminotransferases (ATs) are pyridoxal 5′-phosphate–dependent enzymes that catalyze the transamination reactions between amino acid donor and keto acid acceptor substrates. Modern AT enzymes constitute ∼2% of all classified enzymatic activities, play central roles in nitrogen metabolism, and generate multitude of primary and secondary metabolites. ATs likely diverged into four distinct AT classes before the appearance of the last universal common ancestor and further expanded to a large and diverse enzyme family. Although the AT family underwent an extensive functional specialization, many AT enzymes retained considerable substrate promiscuity and multifunctionality because of their inherent mechanistic, structural, and functional constraints. This review summarizes the evolutionary history, diverse metabolic roles, reaction mechanisms, and structure–function relationships of the AT family enzymes, with a special emphasis on their substrate promiscuity and multifunctionality. Comprehensive characterization of AT substrate specificity is still needed to reveal their true metabolic functions in interconnecting various branches of the nitrogen metabolic network in different organisms.     

Thioesterases: A new perspective based on their primary and tertiary structures

Thioesterases (TEs) are classified into EC 3.1.2.1 through EC 3.1.2.27 based on their activities on different substrates, with many remaining unclassified (EC 3.1.2.–). Analysis of primary and tertiary structures of known TEs casts a new light on this enzyme group. We used strong primary sequence conservation based on experimentally proved proteins as the main criterion, followed by verification with tertiary structure superpositions, mechanisms, and catalytic residue positions, to accurately define TE families. At present, TEs fall into 23 families almost completely unrelated to each other by primary structure. It is assumed that all members of the same family have essentially the same tertiary structure; however, TEs in different families can have markedly different folds and mechanisms. Conversely, the latter sometimes have very similar tertiary structures and catalytic mechanisms despite being only slightly or not at all related by primary structure, indicating that they have common distant ancestors and can be grouped into clans. At present, four clans encompass 12 TE families. The new constantly updated ThYme (Thioester‐active enzYmes) database contains TE primary and tertiary structures, classified into families and clans that are different from those currently found in the literature or in other databases. We review all types of TEs, including those cleaving CoA, ACP, glutathione, and other protein molecules, and we discuss their structures, functions, and mechanisms.     

VIDA: a virus database system for the organization of animal virus genome open reading frames

VIDA is a new virus database that organizes open reading frames (ORFs) from partial and complete genomic sequences from animal viruses. Currently VIDA includes all sequences from GenBank for Herpesviridae, Coronaviridae and Arteriviridae. The ORFs are organized into homologous protein families, which are identified on the basis of sequence similarity relationships. Conserved sequence regions of potential functional importance are identified and can be retrieved as sequence alignments. We use a controlled taxonomical and functional classification for all the proteins and protein families in the database. When available, protein structures that are related to the families have also been included. The database is available for online search and sequence information retrieval at http://www.biochem.ucl.ac.uk/bsm/virus_database/ VIDA.html.     

How well is enzyme function conserved as a function of pairwise sequence identity?

Enzyme function conservation has been used to derive the threshold of sequence identity necessary to transfer function from a protein of known function to an unknown protein. Using pairwise sequence comparison, several studies suggested that when the sequence identity is above 40%, enzyme function is well conserved. In contrast, Rost argued that because of database bias, the results from such simple pairwise comparisons might be misleading. Thus, by grouping enzyme sequences into families based on sequence similarity and selecting representative sequences for comparison, he showed that enzyme function starts to diverge quickly when the sequence identity is below 70%. Here, we employ a strategy similar to Rost's to reduce the database bias; however, we classify enzyme families based not only on sequence similarity, but also on functional similarity, i.e. sequences in each family must have the same four digits or the same first three digits of the enzyme commission (EC) number. Furthermore, instead of selecting representative sequences for comparison, we calculate the function conservation of each enzyme family and then average the degree of enzyme function conservation across all enzyme families. Our analysis suggests that for functional transferability, 40% sequence identity can still be used as a confident threshold to transfer the first three digits of an EC number; however, to transfer all four digits of an EC number, above 60% sequence identity is needed to have at least 90% accuracy. Moreover, when PSI-BLAST is used, the magnitude of the E-value is found to be weakly correlated with the extent of enzyme function conservation in the third iteration of PSI-BLAST. As a result, functional annotation based on the E-values from PSI-BLAST should be used with caution. We also show that by employing an enzyme family-specific sequence identity threshold above which 100% functional conservation is required, functional inference of unknown sequences can be accurately accomplished. However, this comes at a cost: those true positive sequences below this threshold cannot be uniquely identified.     

Revisit of aminotransferase in the genomic era and its application to biocatalysis

Aminotransferases (ATs) have useful applications in the chemical industry because of their capability of introducing amino group into ketones or keto acids as well as their high enantioselectivity and regioselectivity and broad substrate specificity. Abundant protein sequence databases and new powerful tools such as advanced computational structure modeling, multiple sequence analysis, and in vitro evolution have made it possible to understand the detailed reaction mechanisms of various ATs and to isolate and design novel enzymes for unnatural substrates. This, in turn, suggests that developing new integrated approaches to screen ATs are possible, but at the same time poses formidable technical challenges. Here, this paper reviews the use of family profile analysis to find the correlation between the type of ATs and their substrate specificities, the relation between the 3-D structures of ATs and their substrate specificities, and enzyme engineering for the synthesis of unnatural substrates.     

Classification and phylogeny for the annotation of novel eukaryotic GNAT acetyltransferases

The enzymes of the GCN5-related N-acetyltransferase (GNAT) superfamily count more than 870 000 members through all kingdoms of life and share the same structural fold. GNAT enzymes transfer an acyl moiety from acyl coenzyme A to a wide range of substrates including aminoglycosides, serotonin, glucosamine-6-phosphate, protein N-termini and lysine residues of histones and other proteins. The GNAT subtype of protein N-terminal acetyltransferases (NATs) alone targets a majority of all eukaryotic proteins stressing the omnipresence of the GNAT enzymes. Despite the highly conserved GNAT fold, sequence similarity is quite low between members of this superfamily even when substrates are similar. Furthermore, this superfamily is phylogenetically not well characterized. Thus functional annotation based on sequence similarity is unreliable and strongly hampered for thousands of GNAT members that remain biochemically uncharacterized. Here we used sequence similarity networks to map the sequence space and propose a new classification for eukaryotic GNAT acetyltransferases. Using the new classification, we built a phylogenetic tree, representing the entire GNAT acetyltransferase superfamily. Our results show that protein NATs have evolved more than once on the GNAT acetylation scaffold. We use our classification to predict the function of uncharacterized sequences and verify by in vitro protein assays that two fungal genes encode NAT enzymes targeting specific protein N-terminal sequences, showing that even slight changes on the GNAT fold can lead to change in substrate specificity. In addition to providing a new map of the relationship between eukaryotic acetyltransferases the classification proposed constitutes a tool to improve functional annotation of GNAT acetyltransferases.     

Navigating within thiamine diphosphate-dependent decarboxylases: Sequences,structures, functional positions,and binding sites

Thiamine diphosphate-dependent decarboxylases catalyze both cleavage and formation of C C bonds in various reactions, which have been assigned to different homologous sequence families. This work compares 53 ThDP-dependent decarboxylases with known crystal structures. Both sequence and structural information were analyzed synergistically and data were analyzed for global and local properties by means of statistical approaches (principle component analysis and principal coordinate analysis) enabling complexity reduction. The different results obtained both locally and globally, that is, individual positions compared with the overall protein sequence or structure, revealed challenges in the assignment of separated homologous families. The methods applied herein support the comparison of enzyme families and the identification of functionally relevant positions. The findings for the family of ThDP-dependent decarboxylases underline that global sequence identity alone is not sufficient to distinguish enzyme function. Instead, local sequence similarity, defined by comparisons of structurally equivalent positions, allows for a better navigation within several groups of homologous enzymes. The differentiation between homologous sequences is further enhanced by taking structural information into account, such as BioGPS analysis of the active site properties or pairwise structural superimpositions. The methods applied herein are expected to be transferrable to other enzyme families, to facilitate family assignments for homologous protein sequences.     

碳水化合物活性酶数据库( CAZy) 及其研究趋势简

碳水化合物活性酶数据库（ CAZy）是关于能够合成或者分解复杂碳水化合物和糖复合物的酶类的一个数据库资源,其基于蛋白质结构域中的氨基酸序列相似性,将碳水化合物活性酶类归入不同蛋白质家族。 CAZy数据库中包含了碳水化合物酶类的物种来源、酶功能EC分类、基因序列、蛋白质序列及其结构等信息。而随着宏基因组学技术的快速发展,CAZy数据库中家族内序列数据量剧增,这为家族内进一步进行亚家族分类奠定了基础;而蛋白质家族内新一层精细分类的引入可提高亚家族中酶分子功能预测的准确度,进而可指导酶分子理性设计来提高特定功能酶组分设计的成功概率,从而推动生物质转化产业的发展。     

Crystal structure of the ω‐aminotransferase from Paracoccus denitrificans and its phylogenetic relationship with other class III amino‐ transferases that have biotechnological potential

Apart from their crucial role in metabolism, pyridoxal 5′‐phosphate (PLP)‐dependent aminotransferases (ATs) constitute a class of enzymes with increasing application in industrial biotechnology. To provide better insight into the structure‐function relationships of ATs with biotechnological potential we performed a fundamental bioinformatics analysis of 330 representative sequences of pro‐ and eukaryotic Class III ATs using a structure‐guided approach. The calculated phylogenetic maximum likelihood tree revealed six distinct clades of which the first segregates with a very high bootstrap value of 92%. Most enzymes in this first clade have been functionally well characterized, whereas knowledge about the natural functions and substrates of enzymes in the other branches is sparse. Notably, in those clades 2‐6 members of the peculiar class of ω‐ATs prevail, many of which have proven useful for the preparation of chiral amines or artificial amino acids. One representative is the ω‐AT from Paracoccus denitrificans (PD ω‐AT) which catalyzes, for example, the transamination in a novel biocatalytic process for the production of L ‐homoalanine from L ‐threonine. To gain structural insight into this important enzyme, its X‐ray analysis was carried out at a resolution of 2.6 Å, including the covalently bound PLP as well as 5‐aminopentanoate as a putative amino donor substrate. On the basis of this crystal structure in conjunction with our phylogenetic analysis, we have identified a generic set of active site residues of ω‐ATs that are associated with a strong preference for aromatic substrates, thus guiding the discovery of novel promising enzymes for the biotechnological production of corresponding chiral amines. © Proteins 2013. © 2012 Wiley Periodicals, Inc.     

Similarity graphing and enzyme-reaction database: methods to detect sequence regions of importance for recognition of chemical structures

We developed a new method which searches sequence segments responsible for the recognition of a given chemical structure. These segments are detected as those locally conserved among a sequence to be analyzed (target sequence) and a set of sequences (reference sequences). Reference sequences are the sequences of functionally related proteins, ligands of which contain a common chemical substructure in their molecular structures. 'Similarity graphing' cuts target sequences into segments, aligns them with reference sequence pairwise, calculates the degree of similarity for each alignment, and shows graphically cumulative similarity values on target sequence. Any locally conserved regions, short or long in length and weak or strong in similarity, are detected at their optimal conditions by adjusting three parameters. The 'enzyme-reaction database' contains chemical structures and their related enzymes. When a chemical substructure is input into the database, sequences of the enzymes related to the input substructure are systematically searched from the NBRF sequence database and output as reference sequences. Examples of analysis using similarity graphing in combination with the enzyme-reaction database showed a great potentiality in the systematic analysis of the relationships between sequences and molecular recognitions for protein engineering.     

Genome scale prediction of substrate specificity for acyl adenylate superfamily of enzymes based on active site residue profiles

Background  

Enzymes belonging to acyl:CoA synthetase (ACS) superfamily activate wide variety of substrates and play major role in increasing the structural and functional diversity of various secondary metabolites in microbes and plants. However, due to the large sequence divergence within the superfamily, it is difficult to predict their substrate preference by annotation transfer from the closest homolog. Therefore, a large number of ACS sequences present in public databases lack any functional annotation at the level of substrate specificity. Recently, several examples have been reported where the enzymes showing high sequence similarity to luciferases or coumarate:CoA ligases have been surprisingly found to activate fatty acyl substrates in experimental studies. In this work, we have investigated the relationship between the substrate specificity of ACS and their sequence/structural features, and developed a novel computational protocol for in silico assignment of substrate preference.     

Sequence and structure classification of kinases

Kinases are a ubiquitous group of enzymes that catalyze the phosphoryl transfer reaction from a phosphate donor (usually ATP) to a receptor substrate. Although all kinases catalyze essentially the same phosphoryl transfer reaction, they display remarkable diversity in their substrate specificity, structure, and the pathways in which they participate. In order to learn the relationship between structural fold and functional specificities in kinases, we have done a comprehensive survey of all available kinase sequences (>17,000) and classified them into 30 distinct families based on sequence similarities. Of these families, 19, covering nearly 98% of all sequences, fall into seven general structural folds for which three-dimensional structures are known. These fold groups include some of the most widespread protein folds, such as Rossmann fold, ferredoxin fold, ribonuclease H fold, and TIM beta/alpha-barrel. On the basis of this classification system, we examined the shared substrate binding and catalytic mechanisms as well as variations of these mechanisms in the same fold groups. Cases of convergent evolution of identical kinase activities occurring in different folds are discussed.     

The 2.1 Å crystal structure of an acyl‐CoA synthetase from Methanosarcina acetivorans reveals an alternate acyl‐binding pocket for small branched acyl substrates

The acyl‐AMP forming family of adenylating enzymes catalyze two‐step reactions to activate a carboxylate with the chemical energy derived from ATP hydrolysis. X‐ray crystal structures have been determined for multiple members of this family and, together with biochemical studies, provide insights into the active site and catalytic mechanisms used by these enzymes. These studies have shown that the enzymes use a domain rotation of 140° to reconfigure a single active site to catalyze the two partial reactions. We present here the crystal structure of a new medium chain acyl‐CoA synthetase from Methanosarcina acetivorans. The binding pocket for the three substrates is analyzed, with many conserved residues present in the AMP binding pocket. The CoA binding pocket is compared to the pockets of both acetyl‐CoA synthetase and 4‐chlorobenzoate:CoA ligase. Most interestingly, the acyl‐binding pocket of the new structure is compared with other acyl‐ and aryl‐CoA synthetases. A comparison of the acyl‐binding pocket of the acyl‐CoA synthetase from M. acetivorans with other structures identifies a shallow pocket that is used to bind the medium chain carboxylates. These insights emphasize the high sequence and structural diversity among this family in the area of the acyl‐binding pocket. Proteins 2009. © 2009 Wiley‐Liss, Inc.     

The phosphopantetheinyl transferase superfamily: phylogenetic analysis and functional implications in cyanobacteria

Phosphopantetheinyl transferases (PPTs) are a superfamily of essential enzymes required for the synthesis of a wide range of compounds including fatty acid, polyketide, and nonribosomal peptide metabolites. These enzymes activate carrier proteins in specific biosynthetic pathways by the transfer of a phosphopantetheinyl moiety to an invariant serine residue. PPTs display low levels of sequence similarity but can be classified into two major families based on several short motifs. The prototype of the first family is the broad-substrate-range PPT Sfp, which is required for biosynthesis of surfactin in Bacillus subtilis. The second family is typified by the Escherichia coli acyl carrier protein synthase (AcpS). Facilitated by the growing number of genome sequences available for analyses, large-scale phylogenetic studies were utilized in this research to reveal novel subfamily groupings, including two subfamilies within the Sfp-like family. In the present study degenerate oligonucleotide primers were designed for amplification of cyanobacterial PPT gene fragments. Subsequent phylogenetic analyses suggested a unique, function-based PPT type, defined by the PPTs involved in heterocyst differentiation. Evidence supporting this hypothesis was obtained by sequencing the region surrounding the partial Nodularia spumigena PPT gene. The ability to genetically classify PPT function is critical for the engineering of novel compounds utilizing combinatorial biosynthesis techniques. Information regarding cyanobacterial PPTs has important ramifications for the ex situ production of cyanobacterial natural products.     

Identification of novel restriction endonuclease-like fold families among hypothetical proteins

Restriction endonucleases and other nucleic acid cleaving enzymes form a large and extremely diverse superfamily that display little sequence similarity despite retaining a common core fold responsible for cleavage. The lack of significant sequence similarity between protein families makes homology inference a challenging task and hinders new family identification with traditional sequence-based approaches. Using the consensus fold recognition method Meta-BASIC that combines sequence profiles with predicted protein secondary structure, we identify nine new restriction endonuclease-like fold families among previously uncharacterized proteins and predict these proteins to cleave nucleic acid substrates. Application of transitive searches combined with gene neighborhood analysis allow us to confidently link these unknown families to a number of known restriction endonuclease-like structures and thus assign folds to the uncharacterized proteins. Finally, our method identifies a novel restriction endonuclease-like domain in the C-terminus of RecC that is not detected with structure-based searches of the existing PDB database.     

Post-Messinian evolutionary relationships across the Sicilian channel: Mitochondrial and nuclear markers link a new green toad from Sicily to African relatives

Background

The major birch pollen allergen, Bet v 1, is a member of the ubiquitous PR-10 family of plant pathogenesis-related proteins. In recent years, a number of diverse plant proteins with low sequence similarity to Bet v 1 was identified. In addition, determination of the Bet v 1 structure revealed the existence of a large superfamily of structurally related proteins. In this study, we aimed to identify and classify all Bet v 1-related structures from the Protein Data Bank and all Bet v 1-related sequences from the Uniprot database.

Results

Structural comparisons of representative members of already known protein families structurally related to Bet v 1 with all entries of the Protein Data Bank yielded 47 structures with non-identical sequences. They were classified into eleven families, five of which were newly identified and not included in the Structural Classification of Proteins database release 1.71. The taxonomic distribution of these families extracted from the Pfam protein family database showed that members of the polyketide cyclase family and the activator of Hsp90 ATPase homologue 1 family were distributed among all three superkingdoms, while members of some bacterial families were confined to a small number of species. Comparison of ligand binding activities of Bet v 1-like superfamily members revealed that their functions were related to binding and metabolism of large, hydrophobic compounds such as lipids, hormones, and antibiotics. Phylogenetic relationships within the Bet v 1 family, defined as the group of proteins with significant sequence similarity to Bet v 1, were determined by aligning 264 Bet v 1-related sequences. A distance-based phylogenetic tree yielded a classification into 11 subfamilies, nine exclusively containing plant sequences and two subfamilies of bacterial proteins. Plant sequences included the pathogenesis-related proteins 10, the major latex proteins/ripening-related proteins subfamily, and polyketide cyclase-like sequences.

Conclusion

The ubiquitous distribution of Bet v 1-related proteins among all superkingdoms suggests that a Bet v 1-like protein was already present in the last universal common ancestor. During evolution, this protein diversified into numerous families with low sequence similarity but with a common fold that succeeded as a versatile scaffold for binding of bulky ligands.     

Redesign of the substrate specificity of Escherichia coli aspartate aminotransferase to that of Escherichia coli tyrosine aminotransferase by homology modeling and site-directed mutagenesis.

Although several high-resolution X-ray crystallographic structures have been determined for Escherichia coli aspartate aminotransferase (eAATase), efforts to crystallize E. coli tyrosine aminotransferase (eTATase) have been unsuccessful. Sequence alignment analyses of eTATase and eAATase show 43% sequence identity and 72% sequence similarity, allowing for conservative substitutions. The high similarity of the two sequences indicates that both enzymes must have similar secondary and tertiary structures. Six active site residues of eAATase were targeted by homology modeling as being important for aromatic amino acid reactivity with eTATase. Two of these positions (Thr 109 and Asn 297) are invariant in all known aspartate aminotransferase enzymes, but differ in eTATase (Ser 109 and Ser 297). The other four positions (Val 39, Lys 41, Thr 47, and Asn 69) line the active site pocket of eAATase and are replaced by amino acids with more hydrophobic side chains in eTATase (Leu 39, Tyr 41, Ile 47, and Leu 69). These six positions in eAATase were mutated by site-directed mutagenesis to the corresponding amino acids found in eTATase in an attempt to redesign the substrate specificity of eAATase to that of eTATase. Five combinations of the individual mutations were obtained from mutagenesis reactions. The redesigned eAATase mutant containing all six mutations (Hex) displays second-order rate constants for the transamination of aspartate and phenylalanine that are within an order of magnitude of those observed for eTATase. Thus, the reactivity of eAATase with phenylalanine was increased by over three orders of magnitude without sacrificing the high transamination activity with aspartate observed for both enzymes.(ABSTRACT TRUNCATED AT 250 WORDS)     

A complementary bioinformatics approach to identify potential plant cell wall glycosyltransferase-encoding genes

Plant cell wall (CW) synthesizing enzymes can be divided into the glycan (i.e. cellulose and callose) synthases, which are multimembrane spanning proteins located at the plasma membrane, and the glycosyltransferases (GTs), which are Golgi localized single membrane spanning proteins, believed to participate in the synthesis of hemicellulose, pectin, mannans, and various glycoproteins. At the Carbohydrate-Active enZYmes (CAZy) database where e.g. glucoside hydrolases and GTs are classified into gene families primarily based on amino acid sequence similarities, 415 Arabidopsis GTs have been classified. Although much is known with regard to composition and fine structures of the plant CW, only a handful of CW biosynthetic GT genes-all classified in the CAZy system-have been characterized. In an effort to identify CW GTs that have not yet been classified in the CAZy database, a simple bioinformatics approach was adopted. First, the entire Arabidopsis proteome was run through the Transmembrane Hidden Markov Model 2.0 server and proteins containing one or, more rarely, two transmembrane domains within the N-terminal 150 amino acids were collected. Second, these sequences were submitted to the SUPERFAMILY prediction server, and sequences that were predicted to belong to the superfamilies NDP-sugartransferase, UDP-glycosyltransferase/glucogen-phosphorylase, carbohydrate-binding domain, Gal-binding domain, or Rossman fold were collected, yielding a total of 191 sequences. Fifty-two accessions already classified in CAZy were discarded. The resulting 139 sequences were then analyzed using the Three-Dimensional-Position-Specific Scoring Matrix and mGenTHREADER servers, and 27 sequences with similarity to either the GT-A or the GT-B fold were obtained. Proof of concept of the present approach has to some extent been provided by our recent demonstration that two members of this pool of 27 non-CAZy-classified putative GTs are xylosyltransferases involved in synthesis of pectin rhamnogalacturonan II (J. Egelund, B.L. Petersen, A. Faik, M.S. Motawia, C.E. Olsen, T. Ishii, H. Clausen, P. Ulvskov, and N. Geshi, unpublished data).