Prediction and characterization of enzymatic activities guided by sequence similarity and genome neighborhood networks

Substrate Specificity of a Methyltransferase Involved in the Biosynthesis of the Lantibiotic Cacaoidin

Modification of the N- and C-termini of peptides enhances their stability against degradation by exopeptidases. The biosynthetic pathways of many peptidic natural products feature enzymatic modification of their termini, and these enzymes may represent a valuable pool of biocatalysts. The lantibiotic cacaoidin carries an N,N-dimethylated N-terminal amine group. Its biosynthetic gene cluster encodes the putative methyltransferase Cao4. In this work, we present reconstitution of the activity of the enzyme, which we termed CaoSC following standardized lanthipeptide nomenclature, using a heterologously produced peptide as the model substrate. In vitro methylation of diverse lanthipeptides revealed the substrate requirements of CaoSC. The enzyme accepts peptides of varying lengths and C-terminal sequences but requires dehydroalanine or dehydrobutyrine at the second position. CaoSC-mediated dimethylation of natural lantibiotics resulted in modestly enhanced antimicrobial activity of the lantibiotic haloduracin compared to that of the native compound. Improved activity and/or metabolic stability as a result of methylation illustrates the potential future application of CaoSC in the bioengineering of therapeutic peptides.

Proteome-Wide Bioinformatic Annotation and Functional Validation of the Monotopic Phosphoglycosyl Transferase Superfamily

Phosphoglycosyl transferases (PGTs) are membrane proteins that initiate glycoconjugate biosynthesis by transferring a phospho-sugar moiety from a soluble nucleoside diphosphate sugar to a membrane-embedded polyprenol phosphate acceptor. The centrality of PGTs in complex glycan assembly and the current lack of functional information make these enzymes high-value targets for biochemical investigation. In particular, the small monotopic PGT family is exclusively bacterial and represents the minimal functional unit of the monotopic PGT superfamily. Here, we combine a sequence similarity network (SSN) analysis with a generalizable, luminescence-based activity assay to probe the substrate specificity of this family of monoPGTs in a bacterial cell-membrane fraction. This strategy allows us to identify specificity on a far more significant scale than previously achievable and correlate preferred substrate specificities with predicted structural differences within the conserved monoPGT fold. Finally, we present the proof-of-concept for a small-scale inhibitor screen (eight nucleoside analogs) with four monoPGTs of diverse substrate specificity, thus building a foundation for future inhibitor discovery initiatives.

Genome mining of sulfonated lanthipeptides reveals unique cyclic peptide sulfotransferases

Although sulfonation plays crucial roles in various biological processes and is frequently utilized in medicinal chemistry to improve water solubility and chemical diversity of drug leads, it is rare and underexplored in ribosomally synthesized and post-translationally modified peptides (RiPPs). Biosynthesis of RiPPs typically entails modification of hydrophilic residues, which substantially increases their chemical stability and bioactivity, albeit at the expense of reducing water solubility. To explore sulfonated RiPPs that may have improved solubility, we conducted co-occurrence analysis of RiPP class-defining enzymes and sulfotransferase (ST), and discovered two distinctive biosynthetic gene clusters (BGCs) encoding both lanthipeptide synthetase (LanM) and ST. Upon expressing these BGCs, we characterized the structures of novel sulfonated lanthipeptides and determined the catalytic details of LanM and ST. We demonstrate that SslST-catalyzed sulfonation is leader-independent but relies on the presence of A ring formed by LanM. Both LanM and ST are promiscuous towards residues in the A ring, but ST displays strict regioselectivity toward Tyr5. The recognition of cyclic peptide by ST was further discussed. Bioactivity evaluation underscores the significance of the ST-catalyzed sulfonation. This study sets up the starting point to engineering the novel lanthipeptide STs as biocatalysts for hydrophobic lanthipeptides improvement.

Divergent Biosynthesis of Bridged Polycyclic Sesquiterpenoids by a Minimal Fungal Biosynthetic Gene Cluster

The bridged polycyclic sesquiterpenoids derived from sativene, isosativene, and longifolene have unique structures, and many chemical synthesis approaches with at least 10 steps have been reported. However, their biosynthetic pathway remains undescribed. A minimal biosynthetic gene cluster (BGC), named bip, encoding a sesquiterpene cyclase (BipA) and a cytochrome P450 (BipB) is characterized to produce such complex sesquiterpenoids with multiple carbon skeletons based on enzymatic assays, heterologous expression, and precursor experiments. BipA is demonstrated as a versatile cyclase with (-)-sativene as the dominant product and (-)-isosativene and (-)-longifolene as minor ones. BipB is capable of hydroxylating different enantiomeric sesquiterpenes, such as (-)-longifolene and (+)-longifolene, at C-15 and C-14 in turn. The C-15- or both C-15- and C-14-hydroxylated products are then further oxidized by unclustered oxidases, resulting in a structurally diverse array of sesquiterpenoids. Bioinformatic analysis reveals the BipB homologues as a discrete clade of fungal sesquiterpene P450s. These findings elucidate the concise and divergent biosynthesis of such intricate bridged polycyclic sesquiterpenoids, offer valuable biocatalysts for biotransformation, and highlight the distinct biosynthetic strategy employed by nature compared to chemical synthesis.

Genome Mining of Linaridins Provides Insights into the Widely Distributed LinC Oxidoreductases

Linaridins are a family of underexplored ribosomally synthesized and post-translationally modified peptides despite the prevalence of their biosynthetic gene clusters (BGCs) in microbial genomes, as shown by bioinformatic studies. Our genome mining efforts reveal that 96 putative oxidoreductase genes, namely, LinC, are encoded in linaridin BGCs. We heterologously expressed two such LinC-containing linaridin BGCs, yan and ydn, from Streptomyces yunnanensis and obtained three new linaridins, named yunnanaridins A-C (1-3). Their structures are characterized by Z-configurations of the dehydrobutyrines and the presence of a variety of epimerized amino acid residues. Yunnanaridin A (1) is the sixth member of the family of type-B linaridins, whereas yunnanaridins B (2) and C (3) represent the first examples of expressed type-C linaridins. Interestingly, heterologous expression of the same BGCs with LinC in-frame knockouts produced the same compounds. This work expands the structural diversity of linaridins and provides evidence for the notion that the widespread LinCs may not be involved in linaridin biosynthesis.

Oxygenolytic sulfoquinovose degradation by an iron-dependent alkanesulfonate dioxygenase

Sulfoquinovose (6-deoxy-6-sulfo-D-glucose, SQ), the polar head group of sulfolipids in plants, is abundant in nature. Many bacteria degrade SQ through pathways termed sulfoglycolysis producing C3 or C2 sulfonates, while certain bacteria degrade SQ through direct oxygenolytic cleavage of the SQ C-S bond, catalyzed by a flavin-dependent alkanesulfonate monooxygenase (sulfo-ASMO pathway). Here we report bioinformatics and biochemical studies revealing an alternative mechanism for oxygenolytic cleavage of the SQ C-S bond, catalyzed by an iron and α-ketoglutarate-dependent alkanesulfonate dioxygenase (SqoD, sulfo-ASDO pathway). In both the ASMO and ASDO pathways, the product 6-dehydroglucose is reduced to glucose by NAD(P)H-dependent SquF. Marinomonas ushuaiensis, a marine bacterium, which harbors the sulfo-ASDO gene cluster is shown utilizing SQ as a carbon source for growth, accompanied by increased transcription of SqoD. The sulfo-ASDO pathway highlights the range of microbial strategies for degradation of this ubiquitous sulfo-sugar, with potential implications for sulfur recycling in different biological environments.

Isethionate is an intermediate in the degradation of sulfoacetate by the human gut pathobiont Bilophila wadsworthia

The obligately anaerobic sulfite-reducing bacterium Bilophila wadsworthia is a common human pathobiont inhabiting the distal intestinal tract. It has a unique ability to utilize a diverse range of food- and host-derived sulfonates to generate sulfite as a terminal electron acceptor (TEA) for anaerobic respiration, converting the sulfonate sulfur to H2S, implicated in inflammatory conditions and colon cancer. The biochemical pathways involved in the metabolism of the C2 sulfonates isethionate and taurine by B. wadsworthia were recently reported. However, its mechanism for metabolizing sulfoacetate, another prevalent C2 sulfonate, remained unknown. Here, we report bioinformatics investigations and in vitro biochemical assays that uncover the molecular basis for the utilization of sulfoacetate as a source of TEA (STEA) for B. wadsworthia, involving conversion to sulfoacetyl-CoA by an ADP-forming sulfoacetate-CoA ligase (SauCD), and stepwise reduction to isethionate by NAD(P)H-dependent enzymes sulfoacetaldehyde dehydrogenase (SauS) and sulfoacetaldehyde reductase (TauF). Isethionate is then cleaved by the O2-sensitive isethionate sulfolyase (IseG), releasing sulfite for dissimilatory reduction to H2S. Sulfoacetate in different environments originates from anthropogenic sources such as detergents, and natural sources such as bacterial metabolism of the highly abundant organosulfonates sulfoquinovose and taurine. Identification of enzymes for anaerobic degradation of this relatively inert and electron-deficient C2 sulfonate provides further insights into sulfur recycling in the anaerobic biosphere, including the human gut microbiome.

A Variant of the Sulfoglycolytic Transketolase Pathway for the Degradation of Sulfoquinovose into Sulfoacetate

Sulfoquinovose (SQ, 6-deoxy-6-sulfo-glucose) constitutes the polar head group of plant sulfolipids and is one of the most abundantly produced organosulfur compounds in nature. Degradation of SQ by bacterial communities contributes to sulfur recycling in many environments. Bacteria have evolved at least four mechanisms for glycolytic degradation of SQ, termed sulfoglycolysis, producing C3 sulfonate (dihydroxypropanesulfonate and sulfolactate) and C2 sulfonate (isethionate) by-products. These sulfonates are further degraded by other bacteria, leading to the mineralization of the sulfonate sulfur. The C2 sulfonate sulfoacetate is widespread in the environment and is also thought to be a product of sulfoglycolysis, although the mechanistic details are yet unknown. Here, we describe a gene cluster in an Acholeplasma sp., from a metagenome derived from deeply circulating subsurface aquifer fluids (GenBank accession no. QZKD01000037), encoding a variant of the recently discovered sulfoglycolytic transketolase (sulfo-TK) pathway that produces sulfoacetate instead of isethionate as a by-product. We report the biochemical characterization of a coenzyme A (CoA)-acylating sulfoacetaldehyde dehydrogenase (SqwD) and an ADP-forming sulfoacetate-CoA ligase (SqwKL), which collectively catalyze the oxidation of the transketolase product sulfoacetaldehyde into sulfoacetate, coupled with ATP formation. A bioinformatics study revealed the presence of this sulfo-TK variant in phylogenetically diverse bacteria, adding to the variety of mechanisms by which bacteria metabolize this ubiquitous sulfo-sugar. IMPORTANCE Many bacteria utilize environmentally widespread C2 sulfonate sulfoacetate as a sulfur source, and the disease-linked human gut sulfate- and sulfite-reducing bacteria can use it as a terminal electron receptor for anaerobic respiration generating toxic H2S. However, the mechanism of sulfoacetate formation is unknown, although it has been proposed that sulfoacetate originates from bacterial degradation of sulfoquinovose (SQ), the polar head group of sulfolipids present in all green plants. Here, we describe a variant of the recently discovered sulfoglycolytic transketolase (sulfo-TK) pathway. Unlike the regular sulfo-TK pathway that produces isethionate, our biochemical assays with recombinant proteins demonstrated that a CoA-acylating sulfoacetaldehyde dehydrogenase (SqwD) and an ADP-forming sulfoacetate-CoA ligase (SqwKL) in this variant pathway collectively catalyze the oxidation of the transketolase product sulfoacetaldehyde into sulfoacetate, coupled with ATP formation. A bioinformatics study revealed the presence of this sulfo-TK variant in phylogenetically diverse bacteria and interpreted the widespread existence of sulfoacetate.

Decipher enzymes from human microbiota for drug discovery and development

The human microbiota plays an important role in human health and contributes to the metabolism of therapeutic drugs affecting their potency. However, the current knowledge on human gut bacterial metabolism is limited and lacks an understanding of the underlying mechanisms of observed drug biotransformations. Despite the complexity of the gut microbial community, genomic and metagenomic sequencing provides insights into the diversity of chemical reactions that can be carried out by the microbiota and poses new challenges to functionally annotate thousands of bacterial enzymes. Here, we outline methods to systematically address the structural and functional space of the human microbiome, highlighting a combination of in silico and in vitro approaches. Systematic knowledge about microbial enzymes could eventually be applied for personalized therapy, the development of prodrugs and modulators of unwanted bacterial activity, and the further discovery of new antibiotics.

Mechanistic Studies on Dehydration in Class V Lanthipeptides

Lanthipeptides are ribosomally synthesized and post-translationally modified peptides characterized by lanthionine (Lan) and/or methyllanthionine (MeLan) residues. Four classes of enzymes have been identified to install these structures in a substrate peptide. Recently, a novel class of lanthipeptides was discovered that lack genes for known class I-IV lanthionine synthases in their biosynthetic gene cluster (BGC). In this study, the dehydration of Ser/Thr during the biosynthesis of the class V lanthipeptide cacaoidin was reconstituted in vitro. The aminoglycoside phosphotransferase-like enzyme CaoK iteratively phosphorylates Ser/Thr residues on the precursor peptide CaoA, followed by phosphate elimination catalyzed by the HopA1 effector-like protein CaoY to achieve eight successive dehydrations. CaoY shows sequence similarity to the OspF family proteins and the lyase domains of class III/IV lanthionine synthetases, and mutagenesis studies identified residues that are critical for catalysis. An AlphaFold prediction of the structure of the dehydration enzyme complex engaged with its substrate suggests the importance of hydrophobic interactions between the CaoA leader peptide and CaoK in enzyme-substrate recognition. This model is supported by site-directed mutagenesis studies.

Identification and Characterization of the Biosynthetic Pathway of the Sulfonolipid Capnine

Capnine (2-amino-3-hydroxy-15-methylhexadecane-1-sulfonate) and capnoids (N-fatty acylated capnine derivatives) are sulfonolipids present in the outer membrane of gliding bacteria in the phylum Bacteroidetes and play a role in their unique gliding motility. They are structurally similar to sphingolipids and are thought to be biosynthesized via a similar pathway. Here we report the identification and biochemical characterization of the capnine biosynthetic enzymes cysteate synthase (CapA) and cysteate-C-fatty acyltransferase (CapB) from the pathogenic gliding bacterium Capnocytophaga ochracea and NAD(P)H-dependent dehydrocapnine reductase CapC from the avian pathogen Ornithobacterium rhinotracheale. CapA catalyzes the formation of cysteate from O-phospho-l-serine and sulfite, and CapB catalyzes the formation of dehydrocapnine from cysteate and 13-methyl-myristoyl-CoA, followed by reduction by CapC. CapA is closely related to cystathionine-β-synthase but distantly related to the archaeal cysteate synthase. Close homologues of CapA, CapB, and the CapA isozyme archaeal cysteate synthase are present in many Bacteroidetes bacteria, including environmental, pathogenic, and human oral and intestinal microbiome bacteria, suggesting the widespread ability of these bacteria to biosynthesize capnine and related sulfonolipids.

The Birth of Genomic Enzymology: Discovery of the Mechanistically Diverse Enolase Superfamily

Enzymes are categorized into superfamilies by sequence, structural, and mechanistic similarities. The evolutionary implications can be profound. Until the mid-1990s, the approach was fragmented largely due to limited sequence and structural data. However, in 1996, Babbitt et al. published a paper in Biochemistry that demonstrated the potential power of mechanistically diverse superfamilies to identify common ancestry, predict function, and, in some cases, predict specificity. This Perspective describes the findings of the original work and reviews the current understanding of structure and mechanism in the founding family members. The outcomes of the genomic enzymology approach have reached far beyond the functional assignment of members of the enolase superfamily, inspiring the study of superfamilies and the adoption of sequence similarity networks and genome context and yielding fundamental insights into enzyme evolution.

Comparative genomic analysis of azasugar biosynthesis

Azasugars are monosaccharide analogs in which the ring oxygen is replaced with a nitrogen atom. These well-known glycosidase inhibitors are of interest as therapeutics, yet several aspects of azasugars remain unknown including their distribution, structural diversity, and chemical ecology. The hallmark signature of bacterial azasugar biosynthesis is a three gene cluster (3GC) coding for aminotransferase, phosphatase, and dehydrogenase enzymes. Using the bioinformatics platform Enzyme Similarity Tool (EST), we identified hundreds of putative three gene clusters coding for azasugar production in microbial species. In the course of this work, we also report a consensus sequence for the aminotransferase involved in azasugar biosynthesis as being: SGNXFRXXXFPNXXXXXXXLXVPXPYCXRC. Most clusters are found in Bacillus and Streptomyces species which typically inhabit soil and the rhizosphere, but some clusters are found with diverse species representation such as Photorhabdus and Xenorhabdus which are symbiotic with entomopathogenic nematodes; the human skin commensal Cutibacterium acnes, and the marine Bacillus rugosus SPB7, a symbiont to the sea sponge Spongia officinalis. This pan-taxonomic survey of the azasugar 3GC signature may lead to the identification of new azasugar producers, facilitate studies of their natural functions, and lead to new potential therapeutics.

Streptomyces: host for refactoring of diverse bioactive secondary metabolites

Microbial secondary metabolites are intensively explored due to their demands in pharmaceutical, agricultural and food industries. Streptomyces are one of the largest sources of secondary metabolites having diverse applications. In particular, the abundance of secondary metabolites encoding biosynthetic gene clusters and presence of wobble position in Streptomyces strains make it potential candidate as a native or heterologous host for secondary metabolite production including several cryptic gene clusters expression. Here, we have discussed the developments in Streptomyces strains genome mining, its exploration as a suitable host and application of synthetic biology for refactoring genetic systems for developing chassis for enhanced as well as novel secondary metabolites with reduced genome and cleaned background.

Mining and unearthing hidden biosynthetic potential

Genetically encoded small molecules (secondary metabolites) play eminent roles in ecological interactions, as pathogenicity factors and as drug leads. Yet, these chemical mediators often evade detection, and the discovery of novel entities is hampered by low production and high rediscovery rates. These limitations may be addressed by genome mining for biosynthetic gene clusters, thereby unveiling cryptic metabolic potential. The development of sophisticated data mining methods and genetic and analytical tools has enabled the discovery of an impressive array of previously overlooked natural products. This review shows the newest developments in the field, highlighting compound discovery from unconventional sources and microbiomes.

Natural products are an important source of bioactive compounds and have versatile applications in different fields, but their discovery is challenging. Here, the authors review the recent developments in genome mining for discovery of natural products, focusing on compounds from unconventional microorganisms and microbiomes.

PhotoModPlus: A web server for photosynthetic protein prediction from genome neighborhood features

A new web server called PhotoModPlus is presented as a platform for predicting photosynthetic proteins via genome neighborhood networks (GNN) and genome neighborhood-based machine learning. GNN enables users to visualize the overview of the conserved neighboring genes from multiple photosynthetic prokaryotic genomes and provides functional guidance on the query input. In the platform, we also present a new machine learning model utilizing genome neighborhood features for predicting photosynthesis-specific functions based on 24 prokaryotic photosynthesis-related GO terms, namely PhotoModGO. The new model performed better than the sequence-based approaches with an F1 measure of 0.872, based on nested five-fold cross-validation. Finally, we demonstrated the applications of the webserver and the new model in the identification of novel photosynthetic proteins. The server is user-friendly, compatible with all devices, and available at bicep.kmutt.ac.th/photomod.

Discovery of new enzymatic functions and metabolic pathways using genomic enzymology web tools

The continuing expansion of protein and genome sequence databases is an opportunity to identify novel enzymes with biotechnological applications. Whether applied to enzymology, chemical biology, systems biology, and microbiology, database mining must be 'user-friendly' so that experimentalists can devise focused strategies to discover the in vitro activities and in vivo functions of uncharacterized enzymes. We developed a suite of genomic enzymology tools (https://efi.igb.illinois.edu/) to (1) generate sequence similarity networks (SSNs) for exploration of sequence-function space in protein families (EFI-EST) and (2) provide genome context for members of protein families (EFI-GNT). Integrated analysis of this complementary information allows to generate testable hypotheses about new functions. After a brief overview of EFI-EST and EFI-GNT, we describe applications that illustrate their use.

Discovery of novel pathways for carbohydrate metabolism

Closing the gap between the increasing availability of complete genome sequences and the discovery of novel enzymes in novel metabolic pathways is a significant challenge. Here, we review recent examples of assignment of in vitro enzymatic activities and in vivo metabolic functions to uncharacterized proteins, with a focus on enzymes and metabolic pathways involved in the catabolism and biosynthesis of monosaccharides and polysaccharides. The most effective approaches are based on analyses of sequence-function space in protein families that provide clues for the predictions of the functions of the uncharacterized enzymes. As summarized in this Opinion, this approach allows the discovery of the catabolism of new molecules, new pathways for common molecules, and new enzymatic chemistries.

A transaldolase-dependent sulfoglycolysis pathway in Bacillus megaterium DSM 1804

Sulfoquinovose (6-deoxy-6-sulfoglucose, SQ) is a component of sulfolipids found in the photosynthetic membranes of plants and other photosynthetic organisms, and is one of the most abundant organosulfur compounds in nature. Microbial degradation of SQ, termed sulfoglycolysis, constitutes an important component of the biogeochemical sulfur cycle. Two sulfoglycolysis pathways have been reported, with one resembling the Embden-Meyerhof-Parnas (sulfo-EMP) pathway, and the other resembling the Entner-Doudoroff (sulfo-ED) pathway. Here we report a third sulfoglycolysis pathway in the bacterium Bacillus megaterium DSM 1804, in which sulfosugar cleavage is catalyzed by the transaldolase SqvA, which converts 6-deoxy-6-sulfofructose and glyceraldehyde 3-phosphate into fructose -6-phosphate and (S)-sulfolactaldehyde. Variations of this transaldolase-dependent sulfoglycolysis (sulfo-TAL) pathway are present in diverse bacteria, and add to the diversity of mechanisms for the degradation of this abundant organosulfur compound.

A ferredoxin-dependent dihydropyrimidine dehydrogenase in Clostridium chromiireducens

Dihydropyrimidine dehydrogenase (PydA) catalyzes the first step of the reductive pyrimidine degradation (Pyd) pathway in bacteria and eukaryotes, enabling pyrimidines to be utilized as substrates for growth. PydA homologs studied to date catalyze the reduction of uracil to dihydrouracil, coupled to the oxidation of NAD(P)H. Uracil reduction occurs at a flavin mononucleotide (FMN) site, and NAD(P)H oxidation occurs at a flavin adenine dinucleotide (FAD) site, with two ferredoxin domains thought to mediate inter-site electron transfer. Here, we report the biochemical characterization of a Clostridial PydA homolog (PydAc) from a Pyd gene cluster in the strict anaerobic bacterium Clostridium chromiireducens. PydAc lacks the FAD domain, and instead is able to catalyze uracil reduction using reduced methyl viologen or reduced ferredoxin as the electron source. Homologs of PydAc are present in Pyd gene clusters in many strict anaerobic bacteria, which use reduced ferredoxin as an intermediate in their energy metabolism.

Two radical-dependent mechanisms for anaerobic degradation of the globally abundant organosulfur compound dihydroxypropanesulfonate

2(S)-dihydroxypropanesulfonate (DHPS) is a microbial degradation product of 6-deoxy-6-sulfo-d-glucopyranose (sulfoquinovose), a component of plant sulfolipid with an estimated annual production of 10¹⁰ tons. DHPS is also at millimolar levels in highly abundant marine phytoplankton. Its degradation and sulfur recycling by microbes, thus, play important roles in the biogeochemical sulfur cycle. However, DHPS degradative pathways in the anaerobic biosphere are not well understood. Here, we report the discovery and characterization of two O₂-sensitive glycyl radical enzymes that use distinct mechanisms for DHPS degradation. DHPS-sulfolyase (HpsG) in sulfate- and sulfite-reducing bacteria catalyzes C-S cleavage to release sulfite for use as a terminal electron acceptor in respiration, producing H₂S. DHPS-dehydratase (HpfG), in fermenting bacteria, catalyzes C-O cleavage to generate 3-sulfopropionaldehyde, subsequently reduced by the NADH-dependent sulfopropionaldehyde reductase (HpfD). Both enzymes are present in bacteria from diverse environments including human gut, suggesting the contribution of enzymatic radical chemistry to sulfur flux in various anaerobic niches.

Photosynthetic protein classification using genome neighborhood-based machine learning feature

Identification of novel photosynthetic proteins is important for understanding and improving photosynthetic efficiency. Synergistically, genome neighborhood can provide additional useful information to identify photosynthetic proteins. We, therefore, expected that applying a computational approach, particularly machine learning (ML) with the genome neighborhood-based feature should facilitate the photosynthetic function assignment. Our results revealed a functional relationship between photosynthetic genes and their conserved neighboring genes observed by 'Phylo score', indicating their functions could be inferred from the genome neighborhood profile. Therefore, we created a new method for extracting patterns based on the genome neighborhood network (GNN) and applied them for the photosynthetic protein classification using ML algorithms. Random forest (RF) classifier using genome neighborhood-based features achieved the highest accuracy up to 87% in the classification of photosynthetic proteins and also showed better performance (Mathew's correlation coefficient = 0.718) than other available tools including the sequence similarity search (0.447) and ML-based method (0.361). Furthermore, we demonstrated the ability of our model to identify novel photosynthetic proteins compared to the other methods. Our classifier is available at http://bicep2.kmutt.ac.th/photomod_standalone, https://bit.ly/2S0I2Ox and DockerHub: https://hub.docker.com/r/asangphukieo/photomod.

A Pathway for Degradation of Uracil to Acetyl Coenzyme A in Bacillus megaterium

Bacteria utilize diverse biochemical pathways for the degradation of the pyrimidine ring. The function of the pathways studied to date has been the release of nitrogen for assimilation. The most widespread of these pathways is the reductive pyrimidine catabolic pathway, which converts uracil into ammonia, carbon dioxide, and β-alanine. Here, we report the characterization of a β-alanine:pyruvate aminotransferase (PydD2) and an NAD⁺-dependent malonic semialdehyde dehydrogenase (MSDH) from a reductive pyrimidine catabolism gene cluster in Bacillus megaterium Together, these enzymes convert β-alanine into acetyl coenzyme A (acetyl-CoA), a key intermediate in carbon and energy metabolism. We demonstrate the growth of B. megaterium in defined medium with uracil as its sole carbon and energy source. Homologs of PydD2 and MSDH are found in association with reductive pyrimidine pathway genes in many Gram-positive bacteria in the order Bacillales Our study provides a basis for further investigations of the utilization of pyrimidines as a carbon and energy source by bacteria.IMPORTANCE Pyrimidine has wide occurrence in natural environments, where bacteria use it as a nitrogen and carbon source for growth. Detailed biochemical pathways have been investigated with focus mainly on nitrogen assimilation in the past decades. Here, we report the discovery and characterization of two important enzymes, PydD2 and MSDH, which constitute an extension for the reductive pyrimidine catabolic pathway. These two enzymes, prevalent in Bacillales based on our bioinformatics studies, allow stepwise conversion of β-alanine, a previous "end product" of the reductive pyrimidine degradation pathway, to acetyl-CoA as carbon and energy source.

Targeting Bacterial Genomes for Natural Product Discovery

Bacterial natural products (NPs) and their analogs constitute more than half of the new small molecule drugs developed over the past few decades. Despite this success, interest in natural products from major pharmaceutical companies has decreased even as genomics has uncovered the large number of biosynthetic gene clusters (BGCs) that encode for novel natural products. To date, there is still a lack of universal strategies and enabling technologies to discover natural products at scale and speed. This review highlights several of the opportunities provided by genome sequencing and bioinformatics, challenges associated with translating genomes into natural products, and examples of successful strain prioritization and BGC activation strategies that have been used in the genomic era for natural product discovery from cultivatable bacteria.

An extended bacterial reductive pyrimidine degradation pathway that enables nitrogen release from β-alanine

The reductive pyrimidine catabolic pathway is the most widespread pathway for pyrimidine degradation in bacteria, enabling assimilation of nitrogen for growth. This pathway, which has been studied in several bacteria including Escherichia coli B, releases only one utilizable nitrogen atom from each molecule of uracil, whereas the other nitrogen atom remains trapped in the end product β-alanine. Here, we report the biochemical characterization of a β-alanine:2-oxoglutarate aminotransferase (PydD) and an NAD(P)H-dependent malonic semialdehyde reductase (PydE) from a pyrimidine degradation gene cluster in the bacterium Lysinibacillus massiliensis Together, these two enzymes converted β-alanine into 3-hydroxypropionate (3-HP) and generated glutamate, thereby making the second nitrogen from the pyrimidine ring available for assimilation. Using bioinformatics analyses, we found that PydDE homologs are associated with reductive pyrimidine pathway genes in many Gram-positive bacteria in the classes Bacilli and Clostridia. We demonstrate that Bacillus smithii grows in a defined medium with uracil or uridine as its sole nitrogen source and detected the accumulation of 3-HP as a waste product. Our findings extend the reductive pyrimidine catabolic pathway and expand the diversity of enzymes involved in bacterial pyrimidine degradation.

The hidden enzymology of bacterial natural product biosynthesis

Bacterial natural products display astounding structural diversity, which, in turn, endows them with a remarkable range of biological activities that are of significant value to modern society. Such structural features are generated by biosynthetic enzymes that construct core scaffolds or perform peripheral modifications, and can thus define natural product families, introduce pharmacophores and permit metabolic diversification. Modern genomics approaches have greatly enhanced our ability to access and characterize natural product pathways via sequence-similarity-based bioinformatics discovery strategies. However, many biosynthetic enzymes catalyse exceptional, unprecedented transformations that continue to defy functional prediction and remain hidden from us in bacterial (meta)genomic sequence data. In this Review, we highlight exciting examples of unusual enzymology that have been uncovered recently in the context of natural product biosynthesis. These suggest that much of the natural product diversity, including entire substance classes, awaits discovery. New approaches to lift the veil on the cryptic chemistries of the natural product universe are also discussed.

EvoMining reveals the origin and fate of natural product biosynthetic enzymes

Natural products (NPs), or specialized metabolites, are important for medicine and agriculture alike, and for the fitness of the organisms that produce them. NP genome-mining aims at extracting biosynthetic information from the genomes of microbes presumed to produce these compounds. Typically, canonical enzyme sequences from known biosynthetic systems are identified after sequence similarity searches. Despite this being an efficient process, the likelihood of identifying truly novel systems by this approach is low. To overcome this limitation, we previously introduced EvoMining, a genome-mining approach that incorporates evolutionary principles. Here, we release and use our latest EvoMining version, which includes novel visualization features and customizable databases, to analyse 42 central metabolic enzyme families (EFs) conserved throughout Actinobacteria, Cyanobacteria, Pseudomonas and Archaea. We found that expansion-and-recruitment profiles of these 42 families are lineage specific, opening the metabolic space related to ‘shell’ enzymes. These enzymes, which have been overlooked, are EFs with orthologues present in most of the genomes of a taxonomic group, but not in all. As a case study of canonical shell enzymes, we characterized the expansion and recruitment of glutamate dehydrogenase and acetolactate synthase into scytonemin biosynthesis, and into other central metabolic pathways driving Archaea and Bacteria adaptive evolution. By defining the origin and fate of enzymes, EvoMining complements traditional genome-mining approaches as an unbiased strategy and opens the door to gaining insights into the evolution of NP biosynthesis. We anticipate that EvoMining will be broadly used for evolutionary studies, and for generating predictions of unprecedented chemical scaffolds and new antibiotics. This article contains data hosted by Microreact.

Biochemical and structural investigation of sulfoacetaldehyde reductase from Klebsiella oxytoca

Sulfoacetaldehyde reductase (IsfD) is a member of the short-chain dehydrogenase/reductase (SDR) family, involved in nitrogen assimilation from aminoethylsulfonate (taurine) in certain environmental and human commensal bacteria. IsfD catalyzes the reversible NADPH-dependent reduction of sulfoacetaldehyde, which is generated by transamination of taurine, forming hydroxyethylsulfonate (isethionate) as a waste product. In the present study, the crystal structure of Klebsiella oxytoca IsfD in a ternary complex with NADPH and isethionate was solved at 2.8 Å, revealing residues important for substrate binding. IsfD forms a homotetramer in both crystal and solution states, with the C-terminal tail of each subunit interacting with the C-terminal tail of the diagonally opposite subunit, forming an antiparallel β sheet that constitutes part of the substrate-binding site. The sulfonate group of isethionate is stabilized by a hydrogen bond network formed by the residues Y148, R195, Q244 and a water molecule. In addition, F249 from the diagonal subunit restrains the conformation of Y148 to further stabilize the orientation of the sulfonate group. Mutation of any of these four residues into alanine resulted in a complete loss of catalytic activity for isethionate oxidation. Biochemical investigations of the substrate scope of IsfD, and bioinformatics analysis of IsfD homologs, suggest that IsfD is related to the promiscuous 3-hydroxyacid dehydrogenases with diverse metabolic functions.

Crystal structure of substrate-bound bifunctional proline racemase/hydroxyproline epimerase from a hyperthermophilic archaeon

The hypothetical OCC_00372 protein from Thermococcus litoralis is a member of the ProR superfamily from hyperthermophilic archaea and exhibits unique bifunctional proline racemase/hydroxyproline 2-epimerase activity. However, the molecular mechanism of the broad substrate specificity and extreme thermostability of this enzyme (TlProR) remains unclear. Here we determined the crystal structure of TlProR at 2.7 Å resolution. Of note, a substrate proline molecule, derived from expression host Escherichia coli cells, was tightly bound in the active site of TlProR. The substrate bound structure and mutational analyses suggested that Trp241 is involved in hydroxyproline recognition by making a hydrogen bond between the indole group of Trp241 and the hydroxyl group of hydroxyproline. Additionally, Tyr171 may contribute to the thermostability by making hydrogen bonds between the hydroxyl group of Tyr171 and catalytic residues. Our structural and functional analyses provide a structural basis for understanding the molecular mechanism of substrate specificity and thermostability of ProR superfamily proteins.

Discovery of caerulomycin/collismycin-type 2,2'-bipyridine natural products in the genomic era

2,2'-Bipyridine (2,2'-BP) is the unique molecular scaffold of the bioactive natural products represented by caerulomycins (CAEs) and collismycins (COLs). CAEs and COLs are highly similar in the chemical structures in which their 2,2'-BP cores typically contain a di- or tri-substituted ring A and an unmodified ring B. Here, we summarize the CAE and COL-type 2,2'-BP natural products known or hypothesized to date: (1) isolated using methods traditional for natural product characterization, (2) created by engineering the biosynthetic pathways of CAEs or COLs, and (3) predicted upon bioinformatics-guided genome mining. The identification of these CAE and COL-type 2,2'-BP natural products not only demonstrates the development of research techniques and methods in the field of natural product chemistry but also reflects the general interest in the discovery of CAE and COL-type 2,2'-BP natural products.

Resistance to Enediyne Antitumor Antibiotics by Sequestration

The enediynes, microbial natural products with extraordinary cytotoxicities, have been translated into clinical drugs. Two self-resistance mechanisms are known in the enediyne producers-apoproteins for the nine-membered enediynes and self-sacrifice proteins for the ten-membered enediyne calicheamicin. Here we show that: (1) tnmS1, tnmS2, and tnmS3 encode tiancimycin (TNM) resistance in its producer Streptomyces sp. CB03234, (2) tnmS1, tnmS2, and tnmS3 homologs are found in all anthraquinone-fused enediyne producers, (3) TnmS1, TnmS2, and TnmS3 share a similar β barrel-like structure, bind TNMs with nanomolar KD values, and confer resistance by sequestration, and (4) TnmS1, TnmS2, and TnmS3 homologs are widespread in nature, including in the human microbiome. These findings unveil an unprecedented resistance mechanism for the enediynes. Mechanisms of self-resistance in producers serve as models to predict and combat future drug resistance in clinical settings. Enediyne-based chemotherapies should now consider the fact that the human microbiome harbors genes encoding enediyne resistance.

Functional assignment of multiple catabolic pathways for D-apiose

Colocation of the genes encoding ABC, TRAP, and TCT transport systems and catabolic pathways for the transported ligand provides a strategy for discovering novel microbial enzymes and pathways. We screened solute-binding proteins (SBPs) for ABC transport systems and identified three that bind D-apiose, a branched pentose in the cell walls of higher plants. Guided by sequence similarity networks (SSNs) and genome neighborhood networks (GNNs), the identities of the SBPs enabled the discovery of four catabolic pathways for D-apiose with eleven previously unknown reactions. The new enzymes include D-apionate oxidoisomerase, which catalyzes hydroxymethyl group migration, as well as 3-oxo-isoapionate-4-phosphate decarboxylase and 3-oxo-isoapionate-4-phosphate transcarboxylase/hydrolase, which are RuBisCO-like proteins (RLPs). The web tools for generating SSNs and GNNs are publicly accessible ( http://efi.igb.illinois.edu/efi-est/ ), so similar 'genomic enzymology' strategies for discovering novel pathways can be used by the community.

An economically and environmentally acceptable synthesis of chiral drug intermediate l-pipecolic acid from biomass-derived lysine via artificially engineered microbes

Deficiency in petroleum resources and increasing environmental concerns have pushed a bio-based economy to be built, employing a highly reproducible, metal contaminant free, sustainable and green biomanufacturing method. Here, a chiral drug intermediate l-pipecolic acid has been synthesized from biomass-derived lysine. This artificial bioconversion system involves the coexpression of four functional genes, which encode l-lysine α-oxidase from Scomber japonicus, glucose dehydrogenase from Bacillus subtilis, Δ1-piperideine-2-carboxylase reductase from Pseudomonas putida, and lysine permease from Escherichia coli. Besides, a lysine degradation enzyme has been knocked out to strengthen the process in this microbe. The overexpression of LysP improved the l-pipecolic acid titer about 1.6-folds compared to the control. This engineered microbial factory showed the highest l-pipecolic acid production of 46.7 g/L reported to date and a higher productivity of 2.41 g/L h and a yield of 0.89 g/g. This biotechnological l-pipecolic acid production is a simple, economic, and green technology to replace the presently used chemical synthesis.

Specialized plant biochemistry drives gene clustering in fungi

The fitness and evolution of prokaryotes and eukaryotes are affected by the organization of their genomes. In particular, the physical clustering of genes can coordinate gene expression and can prevent the breakup of co-adapted alleles. Although clustering may thus result from selection for phenotype optimization and persistence, the impact of environmental selection pressures on eukaryotic genome organization has rarely been systematically explored. Here, we investigated the organization of fungal genes involved in the degradation of phenylpropanoids, a class of plant-produced secondary metabolites that mediate many ecological interactions between plants and fungi. Using a novel gene cluster detection method, we identified 1110 gene clusters and many conserved combinations of clusters in a diverse set of fungi. We demonstrate that congruence in genome organization over small spatial scales is often associated with similarities in ecological lifestyle. Additionally, we find that while clusters are often structured as independent modules with little overlap in content, certain gene families merge multiple modules into a common network, suggesting they are important components of phenylpropanoid degradation strategies. Together, our results suggest that phenylpropanoids have repeatedly selected for gene clustering in fungi, and highlight the interplay between genome organization and ecological evolution in this ancient eukaryotic lineage.

Prediction of enzymatic pathways by integrative pathway mapping

The functions of most proteins are yet to be determined. The function of an enzyme is often defined by its interacting partners, including its substrate and product, and its role in larger metabolic networks. Here, we describe a computational method that predicts the functions of orphan enzymes by organizing them into a linear metabolic pathway. Given candidate enzyme and metabolite pathway members, this aim is achieved by finding those pathways that satisfy structural and network restraints implied by varied input information, including that from virtual screening, chemoinformatics, genomic context analysis, and ligand -binding experiments. We demonstrate this integrative pathway mapping method by predicting the L-gulonate catabolic pathway in Haemophilus influenzae Rd KW20. The prediction was subsequently validated experimentally by enzymology, crystallography, and metabolomics. Integrative pathway mapping by satisfaction of structural and network restraints is extensible to molecular networks in general and thus formally bridges the gap between structural biology and systems biology.

Panning for gold in mould: can we increase the odds for fungal genome mining?

A perspective on existing and emerging strategies for the prioritisation of secondary metabolite biosynthetic gene clusters (BGCs) to increase the odds of fruitful mining of fungal genomes.

The Metallophore Staphylopine Enables Staphylococcus aureus To Compete with the Host for Zinc and Overcome Nutritional Immunity

During infection, the host sequesters essential nutrients, such as zinc, to combat invading microbes. Despite the ability of the immune effector protein calprotectin to bind zinc with subpicomolar affinity, Staphylococcus aureus is able to successfully compete with the host for zinc. However, the zinc importers expressed by S. aureus remain unknown. Our investigations have revealed that S. aureus possesses two importers, AdcABC and CntABCDF, which are induced in response to zinc limitation. While AdcABC is similar to known zinc importers in other bacteria, CntABCDF has not previously been associated with zinc acquisition. Concurrent loss of the two systems severely impairs the ability of S. aureus to obtain zinc and grow in zinc-limited environments. Further investigations revealed that the Cnt system is responsible for the ability of S. aureus to compete with calprotectin for zinc in culture and contributes to acquisition of zinc during infection. The cnt locus also enables S. aureus to produce the broad-spectrum metallophore staphylopine. Similarly to the Cnt transporter, loss of staphylopine severely impairs the ability of S. aureus to resist host-imposed zinc starvation, both in culture and during infection. Further investigations revealed that together staphylopine and the Cnt importer function analogously to siderophore-based iron acquisition systems in order to facilitate zinc acquisition by S. aureus Analogous systems are found in a broad range of Gram-positive and Gram-negative bacterial pathogens, suggesting that this new type of zinc importer broadly contributes to the ability of bacteria to cause infection.IMPORTANCE A critical host defense against infection is the restriction of zinc availability. Despite the subpicomolar affinity of the immune effector calprotectin for zinc, Staphylococcus aureus can successfully compete for this essential metal. Here, we describe two zinc importers, AdcABC and CntABCDF, possessed by S. aureus, the latter of which has not previously been associated with zinc acquisition. The ability of S. aureus to compete with the host for zinc is dependent on CntABCDF and the metallophore staphylopine, both in culture and during infection. These results expand the mechanisms utilized by bacteria to obtain zinc, beyond Adc-like systems, and demonstrate that pathogens utilize strategies similar to siderophore-based iron acquisition to obtain other essential metals during infection. The staphylopine synthesis machinery is present in a diverse collection of bacteria, suggesting that this new family of zinc importers broadly contributes to the ability of numerous pathogens to cause infection.

Genomic Enzymology: Web Tools for Leveraging Protein Family Sequence-Function Space and Genome Context to Discover Novel Functions

The exponentially increasing number of protein and nucleic acid sequences provides opportunities to discover novel enzymes, metabolic pathways, and metabolites/natural products, thereby adding to our knowledge of biochemistry and biology. The challenge has evolved from generating sequence information to mining the databases to integrating and leveraging the available information, i.e., the availability of “genomic enzymology” web tools. Web tools that allow identification of biosynthetic gene clusters are widely used by the natural products/synthetic biology community, thereby facilitating the discovery of novel natural products and the enzymes responsible for their biosynthesis. However, many novel enzymes with interesting mechanisms participate in uncharacterized small-molecule metabolic pathways; their discovery and functional characterization also can be accomplished by leveraging information in protein and nucleic acid databases. This Perspective focuses on two genomic enzymology web tools that assist the discovery novel metabolic pathways: (1) Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST) for generating sequence similarity networks to visualize and analyze sequence–function space in protein families and (2) Enzyme Function Initiative-Genome Neighborhood Tool (EFI-GNT) for generating genome neighborhood networks to visualize and analyze the genome context in microbial and fungal genomes. Both tools have been adapted to other applications to facilitate target selection for enzyme discovery and functional characterization. As the natural products community has demonstrated, the enzymology community needs to embrace the essential role of web tools that allow the protein and genome sequence databases to be leveraged for novel insights into enzymological problems.

Cytochromes P450 for natural product biosynthesis in Streptomyces: sequence, structure, and function

Covering: up to January 2017Cytochrome P450 enzymes (P450s) are some of the most exquisite and versatile biocatalysts found in nature. In addition to their well-known roles in steroid biosynthesis and drug metabolism in humans, P450s are key players in natural product biosynthetic pathways. Natural products, the most chemically and structurally diverse small molecules known, require an extensive collection of P450s to accept and functionalize their unique scaffolds. In this review, we survey the current catalytic landscape of P450s within the Streptomyces genus, one of the most prolific producers of natural products, and comprehensively summarize the functionally characterized P450s from Streptomyces. A sequence similarity network of >8500 P450s revealed insights into the sequence-function relationships of these oxygen-dependent metalloenzymes. Although only ∼2.4% and <0.4% of streptomycete P450s have been functionally and structurally characterized, respectively, the study of streptomycete P450s involved in the biosynthesis of natural products has revealed their diverse roles in nature, expanded their catalytic repertoire, created structural and mechanistic paradigms, and exposed their potential for biomedical and biotechnological applications. Continued study of these remarkable enzymes will undoubtedly expose their true complement of chemical and biological capabilities.

Characterization of a Novel cis-3-Hydroxy-l-Proline Dehydratase and a trans-3-Hydroxy-l-Proline Dehydratase from Bacteria

Hydroxyprolines, such as trans-4-hydroxy-l-proline (T4LHyp), trans-3-hydroxy-l-proline (T3LHyp), and cis-3-hydroxy-l-proline (C3LHyp), are present in some proteins including collagen, plant cell wall, and several peptide antibiotics. In bacteria, genes involved in the degradation of hydroxyproline are often clustered on the genome (l-Hyp gene cluster). We recently reported that an aconitase X (AcnX)-like hypI gene from an l-Hyp gene cluster functions as a monomeric C3LHyp dehydratase (AcnX_{Type I}). However, the physiological role of C3LHyp dehydratase remained unclear. We here demonstrate that Azospirillum brasilense NBRC 102289, an aerobic nitrogen-fixing bacterium, robustly grows using not only T4LHyp and T3LHyp but also C3LHyp as the sole carbon source. The small and large subunits of the hypI gene (hypI_S and hypI_L, respectively) from A. brasilense NBRC 102289 are located separately from the l-Hyp gene cluster and encode a C3LHyp dehydratase with a novel heterodimeric structure (AcnX_{Type IIa}). A strain disrupted in the hypI_S gene did not grow on C3LHyp, suggesting its involvement in C3LHyp metabolism. Furthermore, C3LHyp induced transcription of not only the hypI genes but also the hypK gene encoding Δ¹-pyrroline-2-carboxylate reductase, which is involved in T3LHyp, d-proline, and d-lysine metabolism. On the other hand, the l-Hyp gene cluster of some other bacteria contained not only the AcnX_{Type IIa} gene but also two putative proline racemase-like genes (hypA1 and hypA2). Despite having the same active sites (a pair of Cys/Cys) as hydroxyproline 2-epimerase, which is involved in the metabolism of T4LHyp, the dominant reaction by HypA2 was clearly the dehydration of T3LHyp, a novel type of T3LHyp dehydratase that differed from the known enzyme (Cys/Thr).IMPORTANCE More than 50 years after the discovery of trans-4-hydroxy-l-proline (generally called l-hydroxyproline) degradation in aerobic bacteria, its genetic and molecular information has only recently been elucidated. l-Hydroxyproline metabolic genes are often clustered on bacterial genomes. These loci frequently contain a hypothetical gene(s), whose novel enzyme functions are related to the metabolism of trans-3-hydroxyl-proline and/or cis-3-hydroxyl-proline, a relatively rare l-hydroxyproline in nature. Several l-hydroxyproline metabolic enzymes show no sequential similarities, suggesting their emergence by convergent evolution. Furthermore, transcriptional regulation by trans-4-hydroxy-l-proline, trans-3-hydroxy-l-proline, and/or cis-3-hydroxy-l-proline significantly differs between bacteria. The results of the present study show that several l-hydroxyprolines are available for bacteria as carbon and energy sources and may contribute to the discovery of potential metabolic pathways of another hydroxyproline(s).

AGeNNT: annotation of enzyme families by means of refined neighborhood networks

BACKGROUND

Large enzyme families may contain functionally diverse members that give rise to clusters in a sequence similarity network (SSN). In prokaryotes, the genome neighborhood of a gene-product is indicative of its function and thus, a genome neighborhood network (GNN) deduced for an SSN provides strong clues to the specific function of enzymes constituting the different clusters. The Enzyme Function Initiative ( http://enzymefunction.org/ ) offers services that compute SSNs and GNNs.

RESULTS

We have implemented AGeNNT that utilizes these services, albeit with datasets purged with respect to unspecific protein functions and overrepresented species. AGeNNT generates refined GNNs (rGNNs) that consist of cluster-nodes representing the sequences under study and Pfam-nodes representing enzyme functions encoded in the respective neighborhoods. For cluster-nodes, AGeNNT summarizes the phylogenetic relationships of the contributing species and a statistic indicates how unique nodes and GNs are within this rGNN. Pfam-nodes are annotated with additional features like GO terms describing protein function. For edges, the coverage is given, which is the relative number of neighborhoods containing the considered enzyme function (Pfam-node). AGeNNT is available at https://github.com/kandlinf/agennt .

CONCLUSIONS

An rGNN is easier to interpret than a conventional GNN, which commonly contains proteins without enzymatic function and overly specific neighborhoods due to phylogenetic bias. The implemented filter routines and the statistic allow the user to identify those neighborhoods that are most indicative of a specific metabolic capacity. Thus, AGeNNT facilitates to distinguish and annotate functionally different members of enzyme families.