A machine learning-enabled open biodata resource inventory from the scientific literature.

Modern biological research depends on data resources. These resources archive difficult-to-reproduce data and provide added-value aggregation, curation, and analyses. Collectively, they constitute a global infrastructure of biodata resources. While the organic proliferation of biodata resources has enabled incredible research, sustained support for the individual resources that make up this distributed infrastructure is a challenge. The Global Biodata Coalition (GBC) was established by research funders in part to aid in developing sustainable funding strategies for biodata resources. An important component of this work is understanding the scope of the resource infrastructure; how many biodata resources there are, where they are, and how they are supported. Existing registries require self-registration and/or extensive curation, and we sought to develop a method for assembling a global inventory of biodata resources that could be periodically updated with minimal human intervention. The approach we developed identifies biodata resources using open data from the scientific literature. Specifically, we used a machine learning-enabled natural language processing approach to identify biodata resources from titles and abstracts of life sciences publications contained in Europe PMC. Pretrained BERT (Bidirectional Encoder Representations from Transformers) models were fine-tuned to classify publications as describing a biodata resource or not and to predict the resource name using named entity recognition. To improve the quality of the resulting inventory, low-confidence predictions and potential duplicates were manually reviewed. Further information about the resources were then obtained using article metadata, such as funder and geolocation information. These efforts yielded an inventory of 3112 unique biodata resources based on articles published from 2011-2021. The code was developed to facilitate reuse and includes automated pipelines. All products of this effort are released under permissive licensing, including the biodata resource inventory itself (CC0) and all associated code (BSD/MIT).

Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST): A web tool for generating protein sequence similarity networks.

The Enzyme Function Initiative, an NIH/NIGMS-supported Large-Scale Collaborative Project (EFI; U54GM093342; http://enzymefunction.org/), is focused on devising and disseminating bioinformatics and computational tools as well as experimental strategies for the prediction and assignment of functions (in vitro activities and in vivo physiological/metabolic roles) to uncharacterized enzymes discovered in genome projects. Protein sequence similarity networks (SSNs) are visually powerful tools for analyzing sequence relationships in protein families (H.J. Atkinson, J.H. Morris, T.E. Ferrin, and P.C. Babbitt, PLoS One 2009, 4, e4345). However, the members of the biological/biomedical community have not had access to the capability to generate SSNs for their "favorite" protein families. In this article we announce the EFI-EST (Enzyme Function Initiative-Enzyme Similarity Tool) web tool (http://efi.igb.illinois.edu/efi-est/) that is available without cost for the automated generation of SSNs by the community. The tool can create SSNs for the "closest neighbors" of a user-supplied protein sequence from the UniProt database (Option A) or of members of any user-supplied Pfam and/or InterPro family (Option B). We provide an introduction to SSNs, a description of EFI-EST, and a demonstration of the use of EFI-EST to explore sequence-function space in the OMP decarboxylase superfamily (PF00215). This article is designed as a tutorial that will allow members of the community to use the EFI-EST web tool for exploring sequence/function space in protein families.

The Pseudomonas aeruginosa antimetabolite L -2-amino-4-methoxy-trans-3-butenoic acid (AMB) is made from glutamate and two alanine residues via a thiotemplate-linked tripeptide precursor.

The Pseudomonas aeruginosa toxin L-2-amino-4-methoxy-trans-3-butenoic acid (AMB) is a non-proteinogenic amino acid which is toxic for prokaryotes and eukaryotes. Production of AMB requires a five-gene cluster encoding a putative LysE-type transporter (AmbA), two non-ribosomal peptide synthetases (AmbB and AmbE), and two iron(II)/α-ketoglutarate-dependent oxygenases (AmbC and AmbD). Bioinformatics analysis predicts one thiolation (T) domain for AmbB and two T domains (T1 and T2) for AmbE, suggesting that AMB is generated by a processing step from a precursor tripeptide assembled on a thiotemplate. Using a combination of ATP-PPi exchange assays, aminoacylation assays, and mass spectrometry-based analysis of enzyme-bound substrates and pathway intermediates, the AmbB substrate was identified to be L-alanine (L-Ala), while the T1 and T2 domains of AmbE were loaded with L-glutamate (L-Glu) and L-Ala, respectively. Loading of L-Ala at T2 of AmbE occurred only in the presence of AmbB, indicative of a trans loading mechanism. In vitro assays performed with AmbB and AmbE revealed the dipeptide L-Glu-L-Ala at T1 and the tripeptide L-Ala-L-Glu-L-Ala attached at T2. When AmbC and AmbD were included in the assay, these peptides were no longer detected. Instead, an L-Ala-AMB-L-Ala tripeptide was found at T2. These data are in agreement with a biosynthetic model in which L-Glu is converted into AMB by the action of AmbC, AmbD, and tailoring domains of AmbE. The importance of the flanking L-Ala residues in the precursor tripeptide is discussed.

Mechanistic diversity in the RuBisCO superfamily: RuBisCO from Rhodospirillum rubrum is not promiscuous for reactions catalyzed by RuBisCO-like proteins.

d-Ribulose 1,5-bisphosphate carboxylase/oxygenases (RuBisCOs) are promiscuous, catalyzing not only carboxylation and oxygenation of d-ribulose 1,5-bisphosphate but also other promiscuous, presumably nonphysiological, reactions initiated by abstraction of the 3-proton of d-ribulose 1,5-bisphosphate. Also, RuBisCO has homologues that do not catalyze carboxylation; these are designated RuBisCO-like proteins or RLPs. Members of the two families of RLPs catalyze reactions in the recycling of 5'-methylthioadenosine (MTA) generated by polyamine synthesis: (1) the 2,3-diketo-5-methylthiopentane 1-phosphate (DK-MTP 1-P) "enolase" reaction in the well-known "methionine salvage" pathway in Bacillus sp. and (2) the 5-methylthio-d-ribulose 1-phosphate (MTRu 1-P) 1,3-isomerase reaction in the recently discovered "MTA-isoprenoid shunt" that generates 1-deoxy-d-xylulose 5-phosphate for nonmevalonate isoprene synthesis in Rhodospirillum rubrum. We first studied the structure and reactivity of DK-MTP 1-P that was reported to decompose rapidly [Ashida, H., Saito, Y., Kojima, C., and Yokota, A. (2008) Biosci., Biotechnol., Biochem. 72, 959-967]. The 2-carbonyl group of DK-MTP 1-P is rapidly hydrated and can undergo enolization both nonenzymatically and enzymatically via the small amount of unhydrated material that is present. We then examined the ability of RuBisCO from R. rubrum to catalyze both of the RLP-catalyzed reactions. Contrary to a previous report [Ashida, H., Saito, Y., Kojima, C., Kobayashi, K., Ogasawara, N., and Yokota, A. (2003) Science 302, 286-290], we were unable to confirm that this RuBisCO catalyzes the DK-MTP 1-P "enolase" reaction either in vitro or in vivo. We also determined that this RuBisCO does not catalyze the MTRu 1-P 1,3-isomerase reaction in vitro. Thus, although RuBisCOs can be functionally promiscuous, RuBisCO from R. rubrum is not promiscuous for either of the known RLP-catalyzed reactions.

1-methylthio-D-xylulose 5-phosphate methylsulfurylase: a novel route to 1-deoxy-D-xylulose 5-phosphate in Rhodospirillum rubrum.

Rhodospirillum rubrum produces 5-methylthioadenosine (MTA) from S-adenosylmethionine in polyamine biosynthesis; however, R. rubrum lacks the classical methionine salvage pathway. Instead, MTA is converted to 5-methylthio-d-ribose 1-phosphate (MTR 1-P) and adenine; MTR 1-P is isomerized to 1-methylthio-d-xylulose 5-phosphate (MTXu 5-P) and reductively dethiomethylated to 1-deoxy-d-xylulose 5-phosphate (DXP), an intermediate in the nonmevalonate isoprenoid pathway [Erb, T. J., et al. (2012) Nat. Chem. Biol., in press]. Dethiomethylation, a novel route to DXP, is catalyzed by MTXu 5-P methylsulfurylase. An active site Cys displaces the enolate of DXP from MTXu 5-P, generating a methyl disulfide intermediate.

A RubisCO-like protein links SAM metabolism with isoprenoid biosynthesis.

Functional assignment of uncharacterized proteins is a challenge in the era of large-scale genome sequencing. Here, we combine in extracto NMR, proteomics and transcriptomics with a newly developed (knock-out) metabolomics platform to determine a potential physiological role for a ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO)-like protein from Rhodospirillum rubrum. Our studies unraveled an unexpected link in bacterial central carbon metabolism between S-adenosylmethionine-dependent polyamine metabolism and isoprenoid biosynthesis and also provide an alternative approach to assign enzyme function at the organismic level.

Homology models guide discovery of diverse enzyme specificities among dipeptide epimerases in the enolase superfamily.

The rapid advance in genome sequencing presents substantial challenges for protein functional assignment, with half or more of new protein sequences inferred from these genomes having uncertain assignments. The assignment of enzyme function in functionally diverse superfamilies represents a particular challenge, which we address through a combination of computational predictions, enzymology, and structural biology. Here we describe the results of a focused investigation of a group of enzymes in the enolase superfamily that are involved in epimerizing dipeptides. The first members of this group to be functionally characterized were Ala-Glu epimerases in Eschericiha coli and Bacillus subtilis, based on the operon context and enzymological studies; these enzymes are presumed to be involved in peptidoglycan recycling. We have subsequently studied more than 65 related enzymes by computational methods, including homology modeling and metabolite docking, which suggested that many would have divergent specificities;, i.e., they are likely to have different (unknown) biological roles. In addition to the Ala-Phe epimerase specificity reported previously, we describe the prediction and experimental verification of: (i) a new group of presumed Ala-Glu epimerases; (ii) several enzymes with specificity for hydrophobic dipeptides, including one from Cytophaga hutchinsonii that epimerizes D-Ala-D-Ala; and (iii) a small group of enzymes that epimerize cationic dipeptides. Crystal structures for certain of these enzymes further elucidate the structural basis of the specificities. The results highlight the potential of computational methods to guide experimental characterization of enzymes in an automated, large-scale fashion.

The Enzyme Function Initiative.

The Enzyme Function Initiative (EFI) was recently established to address the challenge of assigning reliable functions to enzymes discovered in bacterial genome projects; in this Current Topic, we review the structure and operations of the EFI. The EFI includes the Superfamily/Genome, Protein, Structure, Computation, and Data/Dissemination Cores that provide the infrastructure for reliably predicting the in vitro functions of unknown enzymes. The initial targets for functional assignment are selected from five functionally diverse superfamilies (amidohydrolase, enolase, glutathione transferase, haloalkanoic acid dehalogenase, and isoprenoid synthase), with five superfamily specific Bridging Projects experimentally testing the predicted in vitro enzymatic activities. The EFI also includes the Microbiology Core that evaluates the in vivo context of in vitro enzymatic functions and confirms the functional predictions of the EFI. The deliverables of the EFI to the scientific community include (1) development of a large-scale, multidisciplinary sequence/structure-based strategy for functional assignment of unknown enzymes discovered in genome projects (target selection, protein production, structure determination, computation, experimental enzymology, microbiology, and structure-based annotation), (2) dissemination of the strategy to the community via publications, collaborations, workshops, and symposia, (3) computational and bioinformatic tools for using the strategy, (4) provision of experimental protocols and/or reagents for enzyme production and characterization, and (5) dissemination of data via the EFI's Website, http://enzymefunction.org. The realization of multidisciplinary strategies for functional assignment will begin to define the full metabolic diversity that exists in nature and will impact basic biochemical and evolutionary understanding, as well as a wide range of applications of central importance to industrial, medicinal, and pharmaceutical efforts.

N-acylation during glidobactin biosynthesis by the tridomain nonribosomal peptide synthetase module GlbF.

Glidobactins are hybrid NRPS-PKS natural products that function as irreversible proteasome inhibitors. A variety of medium chain 2(E),4(E)-diene fatty acids N-acylate the peptidolactam core and contribute significantly to the potency of proteasome inhibition. We have expressed the initiation NRPS module GlbF (C-A-T) in Escherichia coli and observe soluble active protein only on coexpression with the 8 kDa MbtH-like protein, GlbE. Following adenylation and installation of Thr as a T-domain thioester, the starter condensation domain utilizes fatty acyl-CoA donors to acylate the Thr(1) amino group and generate the fatty acyl-Thr(1)-S-pantetheinyl-GlbF intermediate to be used in subsequent chain elongation. Previously proposed to be mediated via acyl carrier protein fatty acid donors, direct utilization of fatty acyl-CoA donors for N-acylation of T-domain tethered amino acids is likely a common strategy for chain initiation in NRPS-mediated lipopeptide biosynthesis.

Activation of the pacidamycin PacL adenylation domain by MbtH-like proteins.

Nonribosomal peptide synthetase (NRPS) assembly lines are major avenues for the biosynthesis of a vast array of peptidyl natural products. Several hundred bacterial NRPS gene clusters contain a small (∼70-residue) protein belonging to the MbtH family for which no function has been defined. Here we show that two strictly conserved Trp residues in MbtH-like proteins contribute to stimulation of amino acid adenylation in some NRPS modules. We also demonstrate that adenylation can be stimulated not only by cognate MbtH-like proteins but also by homologues from disparate natural product pathways.

SylC catalyzes ureido-bond formation during biosynthesis of the proteasome inhibitor syringolin A.

Syringolins are a class of cyclic tripeptide natural products that are potent and irreversible inhibitors of the eukaryotic proteasome. In addition to being hybrid NRPS/PKS molecules, they also feature an unusual ureido-linkage (red) between two amino acid monomers. Here we report the first in vitro characterization of enzymatic ureido-linkage formation which is catalyzed by an NRPS, SylC. Using (13)C- and (18)O-labeling studies, we show that biosynthesis occurs via N-carboxylation to form an initial N-carboxy-aminoacyl-S-Ppant enzyme intermediate which undergoes intramolecular cyclization followed by condensation with a second amino acid to form the ureido-containing dipeptide product.

Mechanism of the orotidine 5'-monophosphate decarboxylase-catalyzed reaction: evidence for substrate destabilization.

The reaction catalyzed by orotidine 5'-monophosphate decarboxylase (OMPDC) involves a stabilized anionic intermediate, although the structural basis for the rate acceleration (k(cat)/k(non), 7.1 x 10(16)) and proficiency [(k(cat)/K(M))/k(non), 4.8 x 10(22) M(-1)] is uncertain. That the OMPDCs from Methanothermobacter thermautotrophicus (MtOMPDC) and Saccharomyces cerevisiae (ScOMPDC) catalyze the exchange of H6 of the UMP product with solvent deuterium allows an estimate of a lower limit on the rate acceleration associated with stabilization of the intermediate and its flanking transition states (>or=10(10)). The origin of the "missing" contribution, or=10(10)), is of interest. Based on structures of liganded complexes, unfavorable electrostatic interactions between the substrate carboxylate group and a proximal Asp (Asp 70 in MtOMPDC and Asp 91 in ScOMPDC) have been proposed to contribute to the catalytic efficiency [Wu, N., Mo, Y., Gao, J., and Pai, E. F. (2000) Proc. Natl. Acad. Sci. U.S.A. 97, 2017-2022]. We investigated that hypothesis by structural and functional characterization of the D70N and D70G mutants of MtOMPDC and the D91N mutant of ScOMPDC. The substitutions for Asp 70 in MtOMPDC significantly decrease the value of k(cat) for decarboxylation of FOMP (a more reactive substrate analogue) but have little effect on the value of k(ex) for exchange of H6 of FUMP with solvent deuterium; the structures of wild-type MtOMPDC and its mutants are superimposable when complexed with 6-azaUMP. In contrast, the D91N mutant of ScOMPDC does not catalyze exchange of H6 of FUMP; the structures of wild-type ScOMPDC and its D91N mutant are not superimposable when complexed with 6-azaUMP, with differences in both the conformation of the active site loop and the orientation of the ligand vis a vis the active site residues. We propose that the differential effects of substitutions for Asp 70 of MtOMPDC on decarboxylation and exchange provide additional evidence for a carbanionic intermediate as well as the involvement of Asp 70 in substrate destabilization.

Discovery of a dipeptide epimerase enzymatic function guided by homology modeling and virtual screening.

We have developed a computational approach to aid the assignment of enzymatic function for uncharacterized proteins that uses homology modeling to predict the structure of the binding site and in silico docking to identify potential substrates. We apply this method to proteins in the functionally diverse enolase superfamily that are homologous to the characterized L-Ala-D/L-Glu epimerase from Bacillus subtilis. In particular, a protein from Thermotoga martima was predicted to have different substrate specificity, which suggests that it has a different, but as yet unknown, biological function. This prediction was experimentally confirmed, resulting in the assignment of epimerase activity for L-Ala-D/L-Phe, L-Ala-D/L-Tyr, and L-Ala-D/L-His, whereas the enzyme is annotated incorrectly in GenBank as muconate cycloisomerase. Subsequently, crystal structures of the enzyme were determined in complex with three substrates, showing close agreement with the computational models and revealing the structural basis for the observed substrate selectivity.

Mechanistic diversity in the RuBisCO superfamily: a novel isomerization reaction catalyzed by the RuBisCO-like protein from Rhodospirillum rubrum.

Some homologues of D-ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCO) do not catalyze carboxylation and are designated RuBisCO-like proteins (RLPs). The RLP from Rhodospirillum rubrum (gi:83593333) catalyzes a novel isomerization reaction (overall 1,3-proton transfer reaction; likely, two 1,2-proton transfer reactions) that converts 5-methylthio-D-ribulose 1-phosphate to a 3:1 mixture of 1-methylthioxylulose 5-phosphate and 1-methylthioribulose 5-phosphate. Disruption of the gene encoding the RLP abolishes the ability of R. rubrum to utilize 5'-methylthioadenosine as a sole sulfur source, implicating a new, as-yet-uncharacterized, pathway for sulfur salvage.

Prediction and assignment of function for a divergent N-succinyl amino acid racemase.

The protein databases contain many proteins with unknown function. A computational approach for predicting ligand specificity that requires only the sequence of the unknown protein would be valuable for directing experiment-based assignment of function. We focused on a family of unknown proteins in the mechanistically diverse enolase superfamily and used two approaches to assign function: (i) enzymatic assays using libraries of potential substrates, and (ii) in silico docking of the same libraries using a homology model based on the most similar (35% sequence identity) characterized protein. The results matched closely; an experimentally determined structure confirmed the predicted structure of the substrate-liganded complex. We assigned the N-succinyl arginine/lysine racemase function to the family, correcting the annotation (L-Ala-D/L-Glu epimerase) based on the function of the most similar characterized homolog. These studies establish that ligand docking to a homology model can facilitate functional assignment of unknown proteins by restricting the identities of the possible substrates that must be experimentally tested.

Mechanistic diversity in the RuBisCO superfamily: the "enolase" in the methionine salvage pathway in Geobacillus kaustophilus.

D-Ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCO), the most abundant enzyme, is the paradigm member of the recently recognized mechanistically diverse RuBisCO superfamily. The RuBisCO reaction is initiated by abstraction of the proton from C3 of the d-ribulose 1,5-bisphosphate substrate by a carbamate oxygen of carboxylated Lys 201 (spinach enzyme). Heterofunctional homologues of RuBisCO found in species of Bacilli catalyze the tautomerization ("enolization") of 2,3-diketo-5-methylthiopentane 1-phosphate (DK-MTP 1-P) in the methionine salvage pathway in which 5-methylthio-d-ribose (MTR) derived from 5'-methylthioadenosine is converted to methionine [Ashida, H., Saito, Y., Kojima, C., Kobayashi, K., Ogasawara, N., and Yokota, A. (2003) A functional link between RuBisCO-like protein of Bacillus and photosynthetic RuBisCO, Science 302, 286-290]. The reaction catalyzed by this "enolase" is accomplished by abstraction of a proton from C1 of the DK-MTP 1-P substrate to form the tautomerized product, a conjugated enol. Because the RuBisCO- and "enolase"-catalyzed reactions differ in the regiochemistry of proton abstraction but are expected to share stabilization of an enolate anion intermediate by coordination to an active site Mg2+, we sought to establish structure-function relationships for the "enolase" reaction so that the structural basis for the functional diversity could be established. We determined the stereochemical course of the reaction catalyzed by the "enolases" from Bacillus subtilis and Geobacillus kaustophilus. Using stereospecifically deuterated samples of an alternate substrate derived from d-ribose (5-OH group instead of the 5-methylthio group in MTR) as well as of the natural DK-MTP 1-P substrate, we determined that the "enolase"-catalyzed reaction involves abstraction of the 1-proS proton. We also determined the structure of the activated "enolase" from G. kaustophilus (carboxylated on Lys 173) liganded with Mg2+ and 2,3-diketohexane 1-phosphate, a stable alternate substrate. The stereospecificity of proton abstraction restricts the location of the general base to the N-terminal alpha+beta domain instead of the C-terminal (beta/alpha)8-barrel domain that contains the carboxylated Lys 173. Lys 98 in the N-terminal domain, conserved in all "enolases", is positioned to abstract the 1-proS proton. Consistent with this proposed function, the K98A mutant of the G. kaustophilus "enolase" is unable to catalyze the "enolase" reaction. Thus, we conclude that this functionally divergent member of the RuBisCO superfamily uses the same structural strategy as RuBisCO for stabilizing the enolate anion intermediate, i.e., coordination to an essential Mg2+, but the proton abstraction is catalyzed by a different general base.