Discovery of Borosin Catalytic Strategies and Function through Bioinformatic Profiling.

Borosins are ribosomally synthesized and post-translationally modified peptides (RiPPs) containing backbone α-N-methylations. These modifications confer favorable pharmacokinetic properties including increased membrane permeability and resistance to proteolytic degradation. Previous studies have biochemically and bioinformatically explored several borosins, revealing (1) numerous domain architectures and (2) diverse core regions lacking conserved sequence elements. Due to these characteristics, large-scale computational identification of borosin biosynthetic genes remains challenging and often requires additional, time-intensive manual inspection. This work builds upon previous findings and updates the genome-mining tool RODEO to automatically evaluate borosin biosynthetic gene clusters (BGCs) and identify putative precursor peptides. Using the new RODEO module, we provide an updated analysis of borosin BGCs identified in the NCBI database. From our data set, we bioinformatically predict and experimentally characterize a new fused borosin domain architecture, in which the modified natural product core is encoded N-terminal to the methyltransferase domain. Additionally, we demonstrate that a borosin precursor peptide is a native substrate of shewasin A, a reported aspartyl peptidase with no previously identified substrates. Shewasin A requires post-translational modification of the leader peptide for proteolytic maturation, a feature not previously observed in RiPPs. Overall, this work provides a user-friendly and open-access tool for the analysis of borosin BGCs and we demonstrate its utility to uncover additional biosynthetic strategies within the borosin class of RiPPs.

Darobactin Substrate Engineering and Computation Show Radical Stability Governs Ether versus C-C Bond Formation.

The Gram-negative selective antibiotic darobactin A has attracted interest owing to its intriguing fused bicyclic structure and unique targeting of the outer membrane protein BamA. Darobactin, a ribosomally synthesized and post-translationally modified peptide (RiPP), is produced by a radical S-adenosyl methionine (rSAM)-dependent enzyme (DarE) and contains one ether and one C-C cross-link. Herein, we analyze the substrate tolerance of DarE and describe an underlying catalytic principle of the enzyme. These efforts produced 51 enzymatically modified darobactin variants, revealing that DarE can install the ether and C-C cross-links independently and in different locations on the substrate. Notable variants with fused bicyclic structures were characterized, including darobactin W3Y, with a non-Trp residue at the twice-modified central position, and darobactin K5F, which displays a fused diether ring pattern. While lacking antibiotic activity, quantum mechanical modeling of darobactins W3Y and K5F aided in the elucidation of the requisite features for high-affinity BamA engagement. We also provide experimental evidence for ß-oxo modification, which adds support for a proposed DarE mechanism. Based on these results, ether and C-C cross-link formation was investigated computationally, and it was determined that more stable and longer-lived aromatic Cß radicals correlated with ether formation. Further, molecular docking and transition state structures based on high-level quantum mechanical calculations support the different indole connectivity observed for ether (Trp-C7) and C-C (Trp-C6) cross-links. Finally, mutational analysis and protein structural predictions identified substrate residues that govern engagement to DarE. Our work informs on darobactin scaffold engineering and further unveils the underlying principles of rSAM catalysis.

Genome Mining for New Enzyme Chemistry.

A revolution in the field of biocatalysis has enabled scalable access to compounds of high societal values using enzymes. The construction of biocatalytic routes relies on the reservoir of available enzymatic transformations. A review of uncharacterized proteins predicted from genomic sequencing projects shows that a treasure trove of enzyme chemistry awaits to be uncovered. This Review highlights enzymatic transformations discovered through various genome mining methods and showcases their potential future applications in biocatalysis.

Enterolyin S, a Polythiazole-containing Hemolytic Peptide from Enterococcus caccae.

The ß-hemolytic factor streptolysin S (SLS) is an important linear azol(in)e-containing peptide (LAP) that contributes significantly to the virulence of Streptococcus pyogenes. Despite its discovery 85 years ago, SLS has evaded structural characterizing owing to its notoriously problematic physicochemical properties. Here, we report the discovery and characterization of a structurally analogous hemolytic peptide from Enterococcus caccae, termed enterolysin S (ELS). Through heterologous expression, site-directed mutagenesis, chemoselective modification, and high-resolution mass spectrometry, we found that ELS contains an intriguing contiguous octathiazole moiety. The discovery of ELS expands our knowledge of hemolytic LAPs by adding a new member to this virulence-promoting family of modified peptides.

Tryptophan-Centric Bioinformatics Identifies New Lasso Peptide Modifications.

Lasso peptides are a class of ribosomally synthesized and post-translationally modified peptides (RiPPs) defined by a macrolactam linkage between the N-terminus and the side chain of an internal aspartic acid or glutamic acid residue. Instead of adopting a branched-cyclic conformation, lasso peptides are "threaded", with the C-terminal tail passing through the macrocycle to present a kinetically trapped rotaxane conformation. The availability of enhanced bioinformatics methods has led to a significant increase in the number of secondary modifications found on lasso peptides. To uncover new ancillary modifications in a targeted manner, a bioinformatic strategy was developed to discover lasso peptides with modifications to tryptophan. This effort identified numerous putative lasso peptide biosynthetic gene clusters with core regions of the precursor peptides enriched in tryptophan. Parsing of these tryptophan (Trp)-rich biosynthetic gene clusters uncovered several putative ancillary modifying enzymes, including halogenases and dimethylallyltransferases expected to act upon Trp. Characterization of two gene products yielded a lasso peptide with two 5-Cl-Trp modifications (chlorolassin) and another bearing 5-dimethylallyl-Trp and 2,3-didehydro-Tyr modifications (wygwalassin). Bioinformatic analysis of the requisite halogenase and dimethylallyltransferase revealed numerous other putative Trp-modified lasso peptides that remain uncharacterized. We anticipate that the Trp-centric strategy reported herein may be useful in discovering ancillary modifications for other RiPP classes and, more generally, guide the functional prediction of enzymes that act on specific amino acids.

Non-modular fatty acid synthases yield distinct N-terminal acylation in ribosomal peptides.

Recent efforts in genome mining of ribosomally synthesized and post-translationally modified peptides (RiPPs) have expanded the diversity of post-translational modification chemistries. However, RiPPs are rarely reported as hybrid molecules incorporating biosynthetic machinery from other natural product families. Here we report lipoavitides, a class of RiPP/fatty-acid hybrid lipopeptides that display a unique, putatively membrane-targeting 4-hydroxy-2,4-dimethylpentanoyl (HMP)-modified N terminus. The HMP is formed via condensation of isobutyryl-coenzyme A (isobutyryl-CoA) and methylmalonyl-CoA catalysed by a 3-ketoacyl-(acyl carrier protein) synthase III enzyme, followed by successive tailoring reactions in the fatty acid biosynthetic pathway. The HMP and RiPP substructures are then connected by an acyltransferase exhibiting promiscuous activity towards the fatty acyl and RiPP substrates. Overall, the discovery of lipoavitides contributes a prototype of RiPP/fatty-acid hybrids and provides possible enzymatic tools for lipopeptide bioengineering.

Substrate Prediction for RiPP Biosynthetic Enzymes via Masked Language Modeling and Transfer Learning.

Ribosomally synthesized and post-translationally modified peptide (RiPP) biosynthetic enzymes often exhibit promiscuous substrate preferences that cannot be reduced to simple rules. Large language models are promising tools for predicting such peptide fitness landscapes. However, state-of-the-art protein language models are trained on relatively few peptide sequences. A previous study comprehensively profiled the peptide substrate preferences of LazBF (a two-component serine dehydratase) and LazDEF (a three-component azole synthetase) from the lactazole biosynthetic pathway. We demonstrated that masked language modeling of LazBF substrate preferences produced language model embeddings that improved downstream classification models of both LazBF and LazDEF substrates. Similarly, masked language modelling of LazDEF substrate preferences produced embeddings that improved the performance of classification models of both LazBF and LazDEF substrates. Our results suggest that the models learned functional forms that are transferable between distinct enzymatic transformations that act within the same biosynthetic pathway. Our transfer learning method improved performance and data efficiency in data-scarce scenarios. We then fine-tuned models on each data set and showed that the fine-tuned models provided interpretable insight that we anticipate will facilitate the design of substrate libraries that are compatible with desired RiPP biosynthetic pathways.

Computationally guided exploration of borosin biosynthetic strategies.

Borosins are ribosomally synthesized and post-translationally modified peptides containing backbone α- N -methylations. Identification of borosin precursor peptides is difficult because (1) there are no conserved sequence elements among borosin precursor peptides and (2) the biosynthetic gene clusters contain numerous domain architectures and peptide fusions. To tackle this problem, we updated the genome mining tool RODEO to automatically evaluate putative borosin BGCs and identify precursor peptides. Enabled by the new borosin module, we analyzed all borosin BGCs found in available sequence data and assigned precursor peptides to previously orphan borosin methyltransferases. Additionally, we bioinformatically predict and experimentally characterize a new fused borosin domain architecture, in which the modified core is N-terminal to the methyltransferase domain. Finally, we demonstrate that a borosin precursor peptide is the native substrate of shewasin A, a previously characterized pepsin-like aspartic peptidase whose native biological function was unknown.

High-Yield Lasso Peptide Production in a Burkholderia Bacterial Host by Plasmid Copy Number Engineering.

The knotted configuration of lasso peptides confers thermal stability and proteolytic resistance, addressing two shortcomings of peptide-based drugs. However, low isolation yields hinder the discovery and development of lasso peptides. While testing Burkholderia sp. FERM BP-3421 as a bacterial host to produce the lasso peptide capistruin, an overproducer clone was previously identified. In this study, we show that an increase in the plasmid copy number partially contributed to the overproducer phenotype. Further, we modulated the plasmid copy number to recapitulate titers to an average of 160% relative to the overproducer, which is 1000-fold higher than previously reported with E. coli, reaching up to 240 mg/L. To probe the applicability of the developed tools for lasso peptide discovery, we targeted a new lasso peptide biosynthetic gene cluster from endosymbiont Mycetohabitans sp. B13, leading to the isolation of mycetolassin-15 and mycetolassin-18 in combined titers of 11 mg/L. These results validate Burkholderia sp. FERM BP-3421 as a production platform for lasso peptide discovery.

Biosynthesis of macrocyclic peptides with C-terminal ß-amino-α-keto acid groups by three different metalloenzymes.

Advances in genome sequencing and bioinformatics methods have identified a myriad of biosynthetic gene clusters (BGCs) encoding uncharacterized molecules. By mining genomes for BGCs containing a prevalent peptide-binding domain used for the biosynthesis of ribosomally synthesized and post-translationally modified peptides (RiPPs), we uncovered a new class involving modifications installed by a cytochrome P450, a multi-nuclear iron-dependent non-heme oxidative enzyme (MNIO, formerly DUF692), a cobalamin- and radical S-adenosyl-L-methionine-dependent enzyme (B12-rSAM), and a methyltransferase. All enzymes encoded by the BGC were functionally expressed in Burkholderia sp. FERM BP-3421. Structural characterization with 2D-NMR and Marfey's method on the resulting RiPP demonstrated that the P450 enzyme catalyzed the formation of a biaryl C-C crosslink between two Tyr residues with the B12-rSAM generating ß-methyltyrosine. The MNIO transformed a C-terminal Asp residue into aminopyruvic acid while the methyltransferase acted on the ß-carbon of the α-keto acid. Exciton-coupled circular dichroism spectroscopy and microcrystal electron diffraction (MicroED) were used to elucidate the stereochemical configurations of the atropisomer that formed upon biaryl crosslinking. The conserved Cys residue in the precursor peptide was not modified as in all other characterized MNIO-containing BGCs; However, mutational analyses demonstrated that it was essential for the MNIO activity on the C-terminal Asp. To the best of our knowledge, the MNIO featured in this pathway is the first to modify a residue other than Cys. This study underscores the utility of genome mining to discover new macrocyclic RiPPs and that RiPPs remain a significant source of previously undiscovered enzyme chemistry.

Benzylic Radical Stabilization Permits Ether Formation During Darobactin Biosynthesis.

The Gram-negative selective antibiotic darobactin A has attracted interest owing to its intriguing fused bicyclic structure and unique mode of action. Biosynthetic studies have revealed that darobactin is a ribosomally synthesized and post-translationally modified peptide (RiPP). During maturation, the darobactin precursor peptide (DarA) is modified by a radical S-adenosyl methionine (rSAM)-dependent enzyme (DarE) to contain ether and C-C crosslinks. In this work, we describe the enzymatic tolerance of DarE using a panel of DarA variants, revealing that DarE can install the ether and C-C crosslinks independently and in different locations on DarA. These efforts produced 57 darobactin variants, 50 of which were enzymatically modified. Several new variants with fused bicyclic structures were characterized, including darobactin W3Y, which replaces tryptophan with tyrosine at the twice-modified central position, and darobactin K5F, which displays a fused diether ring pattern. Three additional darobactin variants contained fused diether macrocycles, leading us to investigate the origin of ether versus C-C crosslink formation. Computational analyses found that more stable and long-lived Cß radicals found on aromatic amino acids correlated with ether formation. Further, molecular docking and calculated transition state structures provide support for the different indole connectivity observed for ether (Trp-C7) and C-C (Trp-C6) crosslink formation. We also provide experimental evidence for a ß-oxotryptophan modification, a proposed intermediate during ether crosslink formation. Finally, mutational analysis of the DarA leader region and protein structural predictions identified which residues were dispensable for processing and others that govern substrate engagement by DarE. Our work informs on darobactin scaffold engineering and sheds additional light on the underlying principles of rSAM catalysis.

Bioinformatics-guided discovery of biaryl-linked lasso peptides.

Lasso peptides are a class of ribosomally synthesized and post-translationally modified peptides (RiPPs) that feature an isopeptide bond and a distinct lariat fold. A growing number of secondary modifications have been described that further decorate lasso peptide scaffolds. Using genome mining, we have discovered a pair of lasso peptide biosynthetic gene clusters (BGCs) that include cytochrome P450 genes. Using mass spectrometry, stable isotope incorporation, and extensive 2D-NMR spectrometry, we report the structural characterization of two unique examples of (C-N) biaryl-linked lasso peptides. Nocapeptin A, from Nocardia terpenica, is tailored with a Trp-Tyr crosslink, while longipepetin A, from Longimycelium tulufanense, features a Trp-Trp linkage. Besides the unusual bicyclic frame, a Met of longipepetin A undergoes S-methylation to yield a trivalent sulfonium, a heretofore unprecedented RiPP modification. A bioinformatic survey revealed additional lasso peptide BGCs containing P450 enzymes which await future characterization. Lastly, nocapeptin A bioactivity was assessed against a panel of human and bacterial cell lines with modest growth-suppression activity detected towards Micrococcus luteus.

Non-modular Fatty Acid Synthases Yield Unique Acylation in Ribosomal Peptides.

Recent efforts in genome mining of ribosomally synthesized and post-translationally modified peptides (RiPPs) have expanded the diversity of post-translational modification chemistries 1, 2 . However, RiPPs are rarely reported as hybrid molecules incorporating biosynthetic machineries from other natural product families 3-8 . Here, we report lipoavitides, a class of RiPP/fatty acid hybrid lipopeptides that display a unique, membrane-targeting 4-hydroxy-2,4-dimethylpentanoyl (HMP)-modified N -terminus. The HMP is formed via condensation of isobutyryl-CoA and methylmalonyl-CoA catalyzed by a 3-ketoacyl-ACP synthase III enzyme, followed by successive tailoring reactions in the fatty acid biosynthetic pathway. The HMP and RiPP substructures are then connected by an acyltransferase exhibiting promiscuous activity towards the fatty acyl and RiPP substrates. Overall, the discovery of lipoavitides contributes a prototype of RiPP/fatty acid hybrids and provides possible enzymatic tools for lipopeptide bioengineering.

McrD binds asymmetrically to methyl-coenzyme M reductase improving active-site accessibility during assembly.

Methyl-coenzyme M reductase (MCR) catalyzes the formation of methane, and its activity accounts for nearly all biologically produced methane released into the atmosphere. The assembly of MCR is an intricate process involving the installation of a complex set of posttranslational modifications and the unique Ni-containing tetrapyrrole called coenzyme F430. Despite decades of research, details of MCR assembly remain largely unresolved. Here, we report the structural characterization of MCR in two intermediate states of assembly. These intermediate states lack one or both F430 cofactors and form complexes with the previously uncharacterized McrD protein. McrD is found to bind asymmetrically to MCR, displacing large regions of the alpha subunit and increasing active-site accessibility for the installation of F430-shedding light on the assembly of MCR and the role of McrD therein. This work offers crucial information for the expression of MCR in a heterologous host and provides targets for the design of MCR inhibitors.

Catalytic Site Proximity Profiling for Functional Unification of Sequence-Diverse Radical S-Adenosylmethionine Enzymes.

The radical S-adenosylmethionine (rSAM) superfamily has become a wellspring for discovering new enzyme chemistry, especially regarding ribosomally synthesized and post-translationally modified peptides (RiPPs). Here, we report a compendium of nearly 15,000 rSAM proteins with high-confidence involvement in RiPP biosynthesis. While recent bioinformatics advances have unveiled the broad sequence space covered by rSAM proteins, the significant challenge of functional annotation remains unsolved. Through a combination of sequence analysis and protein structural predictions, we identified a set of catalytic site proximity residues with functional predictive power, especially among the diverse rSAM proteins that form sulfur-to-α carbon thioether (sactionine) linkages. As a case study, we report that an rSAM protein from Streptomyces sparsogenes (StsB) shares higher full-length similarity with MftC (mycofactocin biosynthesis) than any other characterized enzyme. However, a comparative analysis of StsB to known rSAM proteins using "catalytic site proximity" predicted that StsB would be distinct from MftC and instead form sactionine bonds. The prediction was confirmed by mass spectrometry, targeted mutagenesis, and chemical degradation. We further used "catalytic site proximity" analysis to identify six new sactipeptide groups undetectable by traditional genome-mining strategies. Additional catalytic site proximity profiling of cyclophane-forming rSAM proteins suggests that this approach will be more broadly applicable and enhance, if not outright correct, protein functional predictions based on traditional genomic enzymology principles.

Genome mining unveils a class of ribosomal peptides with two amino termini.

The era of inexpensive genome sequencing and improved bioinformatics tools has reenergized the study of natural products, including the ribosomally synthesized and post-translationally modified peptides (RiPPs). In recent years, RiPP discovery has challenged preconceptions about the scope of post-translational modification chemistry, but genome mining of new RiPP classes remains an unsolved challenge. Here, we report a RiPP class defined by an unusual (S)-N2,N2-dimethyl-1,2-propanediamine (Dmp)-modified C-terminus, which we term the daptides. Nearly 500 daptide biosynthetic gene clusters (BGCs) were identified by analyzing the RiPP Recognition Element (RRE), a common substrate-binding domain found in half of prokaryotic RiPP classes. A representative daptide BGC from Microbacterium paraoxydans DSM 15019 was selected for experimental characterization. Derived from a C-terminal threonine residue, the class-defining Dmp is installed over three steps by an oxidative decarboxylase, aminotransferase, and methyltransferase. Daptides uniquely harbor two positively charged termini, and thus we suspect this modification could aid in membrane targeting, as corroborated by hemolysis assays. Our studies further show that the oxidative decarboxylation step requires a functionally unannotated accessory protein. Fused to the C-terminus of the accessory protein is an RRE domain, which delivers the unmodified substrate peptide to the oxidative decarboxylase. This discovery of a class-defining post-translational modification in RiPPs may serve as a prototype for unveiling additional RiPP classes through genome mining.

Genome mining unveils a class of ribosomal peptides with two amino termini.

The era of inexpensive genome sequencing and improved bioinformatics tools has reenergized the study of natural products, including the ribosomally synthesized and post-translationally modified peptides (RiPPs). In recent years, RiPP discovery has challenged preconceptions about the scope of post-translational modification chemistry, but genome mining of new RiPP classes remains an unsolved challenge. Here, we report a RiPP class defined by an unusual ( S )- N 2 , N 2 -dimethyl-1,2-propanediamine (Dmp)-modified C -terminus, which we term the daptides. Nearly 500 daptide biosynthetic gene clusters (BGCs) were identified by analyzing the RiPP Recognition Element (RRE), a common substrate-binding domain found in half of prokaryotic RiPP classes. A representative daptide BGC from Microbacterium paraoxydans DSM 15019 was selected for experimental characterization. Derived from a C -terminal threonine residue, the class-defining Dmp is installed over three steps by an oxidative decarboxylase, aminotransferase, and methyltransferase. Daptides uniquely harbor two positively charged termini, and thus we suspect this modification could aid in membrane targeting, as corroborated by hemolysis assays. Our studies further show that the oxidative decarboxylation step requires a functionally unannotated accessory protein. Fused to the C -terminus of the accessory protein is an RRE domain, which delivers the unmodified substrate peptide to the oxidative decarboxylase. This discovery of a class-defining post-translational modification in RiPPs may serve as a prototype for unveiling additional RiPP classes through genome mining.

Bioinformatics-Guided Discovery of Biaryl-Tailored Lasso Peptides.

Lasso peptides are a class of ribosomally synthesized and post-translationally modified peptides (RiPPs) that feature an isopeptide bond and a distinct lariat fold. A growing number of secondary modifications have been described that further decorate lasso peptide scaffolds. Using genome mining, we have discovered a pair of lasso peptide biosynthetic gene clusters (BGCs) that include cytochrome P450 genes. Here, we report the structural characterization of two unique examples of (C-N) biaryl-containing lasso peptides. Nocapeptin A, from Nocardia terpenica, is tailored with Trp-Tyr crosslink while longipepetin A, from Longimycelium tulufanense, features Trp-Trp linkage. Besides the unusual bicyclic frame, longipepetin A receives an S-methylation by a new Met methyltransferase resulting in unprecedented sulfonium-bearing RiPP. Our bioinformatic survey revealed P450(s) and further maturating enzyme(s)-containing lasso BGCs awaiting future characterization.

Peptidase Activation by a Leader Peptide-Bound RiPP Recognition Element.

The RiPP precursor recognition element (RRE) is a conserved domain found in many prokaryotic ribosomally synthesized and post-translationally modified peptide (RiPP) biosynthetic gene clusters (BGCs). RREs bind with high specificity and affinity to a recognition sequence within the N-terminal leader region of RiPP precursor peptides. Lasso peptide biosynthesis involves an RRE-dependent leader peptidase, which is discretely encoded or fused to the RRE as a di-domain protein. Here we leveraged thousands of predicted BGCs to define the RRE:leader peptidase interaction through evolutionary covariance analysis. Each interacting domain contributes a three-stranded ß-sheet to form a hydrophobic ß-sandwich-like interface. The bioinformatics-guided predictions were experimentally confirmed using proteins from discrete and fused lasso peptide BGC architectures. Support for the domain-domain interface derived from chemical shift perturbation, paramagnetic relaxation enhancement experiments, and rapid variant activity screening using cell-free biosynthesis. Further validation of selected variants was performed with purified proteins. We developed a p-nitroanilide-based leader peptidase assay to illuminate the role of RRE domains. Our data show that RRE domains play a dual function. RRE domains deliver the precursor peptide to the leader peptidase, and the rate is saturable as expected for a substrate. RRE domains also partially compose the elusive S2 proteolytic pocket that binds the penultimate threonine of lasso leader peptides. Because the RRE domain is required to form the active site, leader peptidase activity is greatly diminished when the RRE domain is supplied at substoichiometric levels. Full proteolytic activation requires RRE engagement with the recognition sequence-containing portion of the leader peptide. Together, our observations define a new mechanism for protease activity regulation.

Bioinformatic prediction and experimental validation of RiPP recognition elements.

Ribosomally synthesized and post-translationally modified peptides (RiPPs) are a family of natural products for which discovery efforts have rapidly grown over the past decade. There are currently 38 known RiPP classes encoded by prokaryotes. Half of the prokaryotic RiPP classes include a protein domain called the RiPP Recognition Element (RRE) for successful installation of post-translational modifications on a RiPP precursor peptide. In most cases, the RRE domain binds to the N-terminal "leader" region of the precursor peptide, facilitating enzymatic modification of the C-terminal "core" region. The prevalence of the RRE domain renders it a theoretically useful bioinformatic handle for class-independent RiPP discovery; however, first-in-class RiPPs have yet to be isolated and experimentally characterized using an RRE-centric strategy. Moreover, with most known RRE domains engaging their cognate precursor peptide(s) with high specificity and nanomolar affinity, evaluation of the residue-specific interactions that govern RRE:substrate complexation is a necessary first step to leveraging the RRE domain for various bioengineering applications. This chapter details protocols for developing custom bioinformatic models to predict and annotate RRE domains in a class-specific manner. Next, we outline methods for experimental validation of precursor peptide binding using fluorescence polarization binding assays and in vitro enzyme activity assays. We anticipate the methods herein will guide and enhance future critical analyses of the RRE domain, eventually enabling its future use as a customizable tool for molecular biology.