Exhaustive Study Guide: Microbial Metabolism, Secondary Metabolite Genome Mining, and Glycan Degradation

Overview of Microbial Metabolism and Gene Clusters

• Microbial Ubiquity and Functional Context:

  • Microbes inhabit virtually every environmental niche, including human internal and external microbiomes, fermented food products (such as kimchi and yogurt), terrestrial soils, aquatic environments, and atmospheric air.
  • Research in microbial biochemistry integrates multiple scientific disciplines: bioinformatics, microbiology, protein biochemistry, synthetic chemistry, structural biology, and functional genomics.

• Primary vs. Secondary Metabolism:

  • Catabolism involves the sequential enzymatic breakdown of complex organic biomolecules—primarily sugars, lipids, and proteins—to yield metabolic energy and basic chemical precursor building blocks.
  • Anabolism encompasses biosynthetic processes that utilize cellular energy and precursors to construct essential biological macromolecules necessary for cellular growth, maintenance, and reproduction (e.g., nucleic acids, structural proteins, and membrane lipids).
  • Secondary Metabolism yields specialized small-molecule metabolites that are non-essential for basic vegetative growth or reproduction, but provide critical ecological and physiological advantages.

• Organization of Microbial Gene Clusters:

  • Enzymatic pathways in bacteria operate via sequential reactions where successive chemical modifications are carried out by distinct, coordinated enzymes rather than a single multifunctional protein.
  • Genes encoding enzymes participating in a common metabolic pathway (catabolic or anabolic) are co-localized within defined contiguous chromosomal regions termed gene clusters.
  • Biosynthetic Gene Clusters (BGCs) are localized genomic regions containing all co-located genes required for the complete enzymatic assembly, tailoring, regulation, and export of secondary metabolites or natural products.

• Biological and Practical Significance of Microbial Metabolism:

  • Basic biological understanding of fundamental microbial physiology, global biogeochemical cycles, and environmental ecosystem dynamics.
  • Elucidation of host-microbe interactions, including how commensal bacterial species colonize host tissues and how pathogenic bacterial species orchestrate virulence mechanisms.
  • Translational applications in biotechnology, such as the generation of bioethanol and advanced biofuels.
  • Pharmaceutical drug discovery through the isolation and characterization of bioactive small molecules.

Natural Product Discovery and Genome Mining Strategies

• Ecological and Medicinal Value of Natural Products:

  • Natural products function ecologically as secondary metabolites that mediate interspecies competition (e.g., antimicrobial competitive inhibition for nutrient niches) and signaling.
  • Possess diverse chemical structures and distinct pharmacological bioactivities, functioning as antibiotics, anti-inflammatory agents, and anti-cancer therapeutics.
  • Analysis of FDA-approved pharmaceuticals spanning the years $1981$\text{--}$2014$ demonstrates that a substantial majority of approved drugs are unmodified natural products, natural product derivatives, or synthetic molecules designed to mimic natural product pharmacophores.

• Limitations of Classical Discovery Protocols:

  • Classical discovery relies on bioactivity-guided fractionation: crude extraction of produced metabolites, bioactivity screening against target organisms, and iterative chromatographic isolation.
  • Modern limitation: Classical methods hit a rediscovery bottleneck, repeatedly isolating known natural product scaffolds or minor structural analogs (e.g., variants containing simple additions or removals of methyl or functional groups).
  • While the cumulative count of total identified natural products increases, the rate of novel chemical scaffold discovery has declined significantly.

• Modern Genome Mining and Bioinformatics Approaches:

  • Reductions in high-throughput DNA sequencing costs have produced an abundant repository of microbial genomes across global databases, enabling genome-guided discovery.
  • Targeted Approaches: Focus on specific enzymatic pathways utilizing genome mining, sequence similarity networks (SSNs), and defined gene cluster identification.
  • Untargeted Approaches: Focus on unbiased chemical discovery using environmental bacterial isolation, untargeted mass spectrometry-based metabolomics, and broad genomic screening.

• Automated BGC Prediction Tools and Limitations:

  • antiSMASH (Antibiotic and Secondary Metabolite Analysis Shell) represents the benchmark software for genome mining.
  • Function: Inputs genomic DNA sequences, scans for core biosynthetic machinery genes, identifies predicted BGCs, assigns natural product class predictions, and calculates similarity confidence scores against established, characterized BGCs.
  • Confidence Assessment:
  • High similarity outputs (e.g., BGCs matches to natamycin) indicate production of known or nearly identical compounds.
  • Low or medium similarity outputs (e.g., BGC matches to hydroxystreptomycin) indicate potential production of related derivatives.
  • Unmatched outputs (zero similarity scores) represent prime targets for novel scaffold discovery.
  • Structural Limitations: antiSMASH relies strictly on recognized rules derived from known core biosynthetic enzymes; non-canonical or novel core architectures are missed.

• Alternative Targeted Genome Mining Strategies:

  • Self-Resistance Gene Mining: Antimicrobial producers require self-immunity mechanisms to avoid autotoxicity. Genes encoding target-site modifications, efflux pumps, or inactivating enzymes co-localize within the producer's BGC and can serve as genomic targets.
  • Tailoring Enzyme Mining: Rather than searching for canonical core scaffold assembly enzymes (e.g., polyketide synthases or non-ribosomal peptide synthetases), mining focuses on accessory enzymes that append functional groups to core structures.
  • Glycosyltransferase (GT) Targeting: The Higgins lab specifically mines for GT enzymes that transfer activated sugar moieties onto natural product scaffolds.

Bioinformatics Pipeline for Glycosyltransferase-Targeted BGC Discovery

• Rationale for Glycosyltransferase Selection:

  • Over $20\%$ of characterized natural products contain glycosylated moieties.
  • Glycosylation modifies diverse core natural product structures without being restricted to a single class of chemical scaffolds.
  • Discrimination Challenge: Most GT enzymes are involved in essential cell wall/glycan biosynthesis rather than secondary metabolism.
  • Sequence Family Selection: Evaluation of six GT protein families identified family PF06722. All characterized enzymes within family PF06722 are dedicated exclusively to natural product secondary biosynthesis rather than structural glycan assembly.
  • The family PF06722 dataset initially comprised approximately $2000$ candidate protein sequences.

• Multi-Step Computational Pipeline Construction:

  • Developed by David Rosado in collaboration with Peter Doring at the Joint Genome Institute (JGI).
  • Step 1: Mapping Accession IDs. Protein sequences from UniProt IDs were mapped to corresponding genomic accession IDs (a complex step due to non-standardized database cross-referencing).
  • Step 2: Automated antiSMASH Execution. Full genomes were parsed through antiSMASH, followed by execution of custom Python scripts to verify if the candidate GT sequence resided within the predicted BGC boundary.
  • Step 3: Initial Classification. Out of the initial dataset:
  • Approximately $1800$ GT sequences were localized within antiSMASH-predicted BGCs ($117$ high similarity, $112$ medium similarity, $25$ low similarity, and $46$ zero similarity to known clusters).
  • $270$ GT sequences were not categorized into BGCs by standard antiSMASH parameters.
  • Step 4: Custom Boundary Definition. Standard antiSMASH identification rules failed to capture non-canonical BGCs containing the $270$ outlier GTs. Custom computational rules were authored to define BGC genomic boundaries for these unassigned sequences.
  • Step 5: Dereplication. Identical genomic BGC duplicates were consolidated, reducing the total dataset to $770$ unique BGCs.

• Algorithmic Innovation for Prioritization:

  • Limitation of BiG-SCAPE: Standard comparative clustering tools rely on pairwise binary gene presence/absence analyses. When applied to the $770$ unique BGCs, BiG-SCAPE failed to form distinct clusters, yielding a fully interconnected, unprioritizable network.
  • Probabilistic Adjacency Model (EFE Model): David Rosado authored a multidimensional probabilistic model calculating the conditional probability of gene adjacency (the relative probability of gene A co-localizing with genes B, C, and D).
  • Quantifying Unique Architecture:
  • The model calculated a scalar "Dissimilarity Score" measuring structural variance among BGCs within the dataset.
  • MIBiG Integration: The model ingested all known, characterized BGCs from the MIBiG repository, which naturally mapped into two primary multidimensional nodes corresponding to Polyketide Synthase (PKS) and Non-Ribosomal Peptide Synthetase (NRPS) pathways.
  • Novelty Score Calculation: Overlaying the $770$ candidate GT BGCs onto the MIBiG spatial distribution allowed calculation of a quantitative "Novelty Score" relative to all published secondary metabolite pathways.

Experimental Validation, BGC Induction, and Natural Product Pathways

• Systematic Prioritization Filtering Workflow:

  1. Initial dataset: $770$ unique BGCs.
  2. AntiSMASH similarity filter: Removal of medium and high similarity hits yielded $541$ BGCs.
  3. Intra-dataset dissimilarity filter: Selection of structural outliers yielded $165$ BGCs.
  4. Novelty filter: Evaluation against MIBiG reference space yielded $68$ novel candidate BGCs.
  5. Resistance gene filter: Screening for co-localized self-resistance genes yielded $9$ final high-priority candidate BGCs targeted for antimicrobial discovery.

• Proof-of-Concept Validation (Dr. Shruti Sharma):

  • Targeted a BGC displaying $53\%$ sequence similarity to the known phogacin biosynthetic pathway.
  • Phogacin variants include Phogacin A (aglycone structure), Phogacin B (glycosylated structure), and Phogacin C (structurally altered aglycone requiring an additional $10$\text{--}$12$ biosynthetic genes).
  • Experimental Isolation: Strain cultivation and compound extraction yielded Phogacin A and Phogacin B.
  • Mechanistic Insight: The missing $47\%$ sequence similarity between the target BGC and the reference phogacin BGC corresponded precisely to the absence of the $10$\text{--}$12$ extra biosynthetic genes required for Phogacin C assembly.

• Inducible Expression of Silent BGCs (Bryce and Kristen Thompson):

  • Target: An uncharacterized, low-similarity BGC containing biosynthetic genes, sugar modification genes, and two transcriptional regulators, but lacking antiSMASH core similarity match.
  • Challenge: BGCs are frequently transcriptionally silent under standard laboratory culture conditions.
  • Induction Strategy: Bryce identified a co-localized SoxR transcriptional regulator within the BGC, known to activate gene expression in response to redox-active molecules.
  • Experimental Result: Master's student Kristen Thompson treated bacterial cultures with menadione (a redox-active compound), successfully inducing transcription of the BGC and eliciting the secretion of a bioactive antimicrobial compound.

• Amino Cyclitol Pathway Characterization (Michael):

  • Investigation of tigromycin A biosynthesis, a pharmaceutical candidate natural product.
  • Focuses on characterizing the unique enzymatic mechanisms of two biosynthetic enzymes, Pig17 and Pig7, involved in building the amino cyclitol functional group.

Microbial Glycan Degradation and Glycoside Hydrolase Mechanisms

• Distribution and Biological Roles of Glycans:

  • Glycans exist as cell-surface glycoproteins, glycolipids (such as glycosylinositolphytoceramides), and secreted glycans (e.g., human milk oligosaccharides [HMOs] and immunoglobulin G [IgG] N-linked glycans essential for antibody function).
  • Physiological rationale for microbial glycan cleavage:
  • Primary energy acquisition through carbohydrate catabolism.
  • Host cell tissue adherence and colonization.
  • Bacterial cell wall peptidoglycan remodeling and autolysis (e.g., Streptococcus pneumoniae).
  • Immune evasion and host immune modulation via enzymatic stripping of host antibody or cell surface glycans.

• Catalytic Action of Glycoside Hydrolases (GHs):

  • GHs catalyze the hydrolytic cleavage of glycosidic bonds linking monomeric sugar units, requiring the addition of a water molecule ($\text{H}_2\text{O}$).
  • Substrate Specificity Criteria:
  1. Non-reducing terminal sugar identity being cleaved.
  2. Underlying adjacent sugar moiety attached to the target sugar.
  3. Regiochemistry and stereochemistry of the glycosidic linkage (e.g., $\alpha\text{-1,2}$, $\alpha\text{-1,3}$, $\alpha\text{-1,4}$, or $\alpha\text{-1,6}$ linkages joining carbon-1 of the terminal sugar to carbon-2, 3, 4, or 6 of the secondary sugar).
  • CAZy Classification: Glycoside hydrolases are organized into sequence-based families ($195$ distinct GH families currently defined in the Carbohydrate-Active EnZymes database).
  • Structure-Function Conservation: Enzymes within a single GH family preserve the overall tertiary protein fold and catalytic mechanism, but can exhibit widely divergent substrate specificities.

• Industrial, Energy, and Clinical Applications of GH Enzymes:

  • Food Biotechnology: Application of lactase (a GH enzyme) to hydrolyze lactose into glucose and galactose for lactose-free dairy products.
  • Bioenergy Production: Enzymatic depolymerization of plant cell wall cellulose into monomeric glucose for bioethanol fermentation.
  • Biomedical Diagnostics: Use of specific GHs to map aberrant glycosylation patterns associated with disease states.

Substrate Specificity and Structural Biology of Fucosidases (GH151)

• Biological Importance of Fucose:

  • L-fucose is a terminal modification on mammalian N- and O-linked glycans, ABO blood group antigens, and Lewis human tissue antigens.
  • Fucosidases remove terminal fucose units from complex glycans.

• Mammalian Fucosidase Families:

  • GH29 Family: Most thoroughly characterized family ($152$ characterized members), exhibiting broad substrate specificities across multiple linkage types.
  • GH95 Family: Highly restricted specificity, cleaving exclusively $\alpha\text{-1,2}$ linkages (fucose $\alpha\text{-1,2}$ galactose).
  • GH151 Family: Historically contained only a single characterized enzyme (PTGH151) displaying broad activity across all four primary fucose linkages ($\alpha\text{-1,2}$, $\alpha\text{-1,3}$, $\alpha\text{-1,4}$, and $\alpha\text{-1,6}$ disaccharides).

• Sequence Similarity Network (SSN) Analysis of GH151:

  • Constructed an SSN where individual nodes represent distinct GH151 protein amino acid sequences and connecting edges represent sequence similarity.
  • Sequence nodes segregated into distinct structural clusters, hypothesized to correspond to divergent functional substrate specificities.
  • Representative GH151 targets were synthesized, recombinantly expressed in Escherichia coli, and biochemically profiled.

• Biochemical Profiling of GH151 Family Members:

  • PTGH151 (Reference Enzyme): Demonstrates broad cleavage capability across $\alpha\text{-1,2}$, $\alpha\text{-1,3}$, $\alpha\text{-1,4}$, and $\alpha\text{-1,6}$ fucosidic linkages.
  • CIGH151 (isolated from the gut commensal Bifidobacterium infantis): Cleaves $\alpha\text{-1,2}$, $\alpha\text{-1,3}$, and $\alpha\text{-1,4}$ linkages, but lacks catalytic activity against $\alpha\text{-1,6}$ linkages.
  • CHGH151 (identified from specific cluster mapping): Displays strict specificity exclusively for $\alpha\text{-1,6}$ fucosidic linkages.
  • Sequence Outliers (Distant Purple Cluster Nodes): Hydrolyze synthetic p-nitrophenyl-alpha-L-fucopyranoside substrates (confirming fucosidase function) but fail to cleave standard mammalian disaccharide linkages, indicating specificity for novel, unmapped glycan structures.

• Structural Biology Insights:

  • X-ray crystal structures were solved for both distant GH151 outlier enzymes, including an enzyme-fucose ligand complex structure.
  • Active Site Analysis: Comparative structural alignment demonstrates strict conservation of essential catalytic residues across PTGH151 and outlier structures, but reveals distinct amino acid substitutions surrounding the binding pocket that govern altered linkage specificities.

Additional Glycan Degradation Initiatives and Research Collaborations

• Specialized Glycoprotein Glycan Removal:

  • Masatana: Characterizing specialized GH enzymes involved in stripping complex glycans from host glycoproteins.
  • David Rosado and Molly: Investigating the GH101 family, targeting the degradation of mucin-type O-linked glycans.
  • Core N-Glycan Processing: Ongoing investigation of novel GH5 family enzymes involved in processing core N-glycan structures, executed by a team of undergraduate researchers.

• Novel Binding Domains and Ecological Isolations:

  • Janique: Mapped published RNA-sequencing data from Akkermansia muciniphila to discover a novel Carbohydrate Binding Module (CBM) non-covalently associated with a GH123 catalytic domain.
  • Ecological Mucin Degradation Discovery:
  • Isolation of specialized mucin-degrading bacteria from freshwater mussel mucus (in collaboration with Dr. Carla Atkinson).
  • Isolation of novel mucin-degrading strains from soil environments on the Shelby Quad via the LEAP program for high school students.
  • Research team: Janique, Sela, Kristen, Bryce, and Virginia (Accelerated Master's Program student).

• Collaborative Infrastructure and Support:

  • Intra-university collaborations across the University of Alabama Departments of Biological Sciences, Chemistry and Biochemistry, and Chemical Engineering.
  • Inter-institutional collaborations: UT Dallas, Oregon State University (Tyco Mountman), University of Calgary, and University of Mississippi.
  • Extramural Funding: NIH R35 MIRA Award and University of Alabama start-up research funds.

Questions & Discussion

• Substrate Discovery via In Silico Screening:

  • Question: Could virtual computational screening and docking approaches—similar to those used by Dr. Matt Thompson's group for inhibitor discovery—be utilized to identify native glycan substrates for orphan GH enzymes?
  • Answer: Yes. This methodology forms a core objective of an NSF CAREER grant application. Computational docking strategies (validated first using known enzyme-substrate models) are being applied to predict natural sugar substrates for uncharacterized outlier enzymes.

• Functional Role of Glycosylation on Natural Products:

  • Question: Does the attachment of sugar moieties directly drive natural product bioactivity or alters pharmacokinetic properties?
  • Answer: It depends on the specific natural product. Glycosylation can directly enhance pharmacodynamic bioactivity (enabling binding to cellular targets) or alter pharmacokinetic bioavailability (modulating solubility and cellular permeability).

• Glycosylation as a Self-Resistance Mechanism:

  • Question: Do producer host cells utilize glycosylation to protect themselves from their own toxic natural products?
  • Answer: Yes. Bacteria can synthesize an inactive, glycosylated precursor intracellularly to prevent autotoxicity. Upon secretion across the cell membrane, an extracellular glycoside hydrolase encoded within the same BGC cleaves the sugar moiety, releasing the active antimicrobial compound into the extracellular space.

• Host Glycan Specificity and Niche Colonization:

  • Question: Does GH substrate specificity dictate host range and bacterial colonization capability?
  • Answer: Yes. Enzymatic specificity directly dictates nutrient competition and host tissue tropism. For example, Bifidobacterium infantis expresses specialized fucosidases optimized to degrade human milk oligosaccharides (HMOs) present in infant milk, conferring a distinct competitive advantage to establish a healthy infant gut microbiome.

• Enzymatic Localization and Pharmacological Implications:

  • Question: How does the cellular localization (secreted vs. intracellular) of glycan-degrading enzymes impact drug delivery and development?
  • Answer: Glycan-degrading enzymes targeting large host cell surface structures possess N-terminal signal peptides that dictate secretion across the bacterial membrane. Small, transportable oligosaccharides are imported and cleaved internally. For natural products, bioavailability relies heavily on crossing cell membranes; poor membrane permeability, unwanted degradation by human digestive enzymes, or cleavage by gut microbiome hydrolases can diminish pharmaceutical efficacy.