Biology subjects

Pereira, J. H.

Publications and source records attributed to Pereira, J. H..

9 recordsLinked to original sources

bioRxiv2026

De novo Rubisco design with protein language models

Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) fixes the majority of carbon dioxide globally but is challenged with low specificity for CO2 versus O2 and low catalytic efficiencies. Traditional engineering efforts have remained difficult because folding, assembly, specificity, and catalysis are tightly coupled, hampering efforts to explore sequence space. Therefore, we leveraged recent advances in protein large language models (PLMs) to generate sequences beyond those observed in nature, using both ProGen-2 that was fine-tuned on a limited dataset of non-Form I Rubiscos and an ESM-2 discriminator. With this approach, we generated 5.6 million novel Rubisco-like sequences and identified 21 highly diverse candidates predicted to be active that occupy regions of Rubisco phylogenetic space not previously observed in nature. Six designs were soluble in Escherichia coli, and five were shown to produce quantifiable 3PGA. One design produced an apparent CO2/O2 specificity estimate beyond the range of the natural representative Rubiscos assayed. We also solved the crystal structure of one de novo design that reproduced the predicted dimer and active-site geometry with sub-angstrom C agreement. Sequence-only generation followed by independent structural filtering therefore recovered soluble, active Rubiscos from regions of sequence space that are not represented in genomic databases. Together, these results establish a scalable strategy for accessing previously unexplored Rubisco sequence space, providing a broadly accessible path toward generating de novo Rubiscos that may have activity and specificity parameters needed to address longstanding limitations in biological carbon fixation.

biochemistry↗

bioRxiv2025

Diversity-driven biochemical survey reveals dimeric structural origin of rubisco

Rubisco is the entry point of nearly all organic carbon into the biosphere and is present in all domains of life. Despite its global importance, biochemical studies of this enzyme superfamily have been limited to a relatively narrow set of subclades. Recent advances in metagenomics have dramatically reshaped our understanding of both microbial and rubisco diversity; however, biochemical characterization of these sequences has not kept pace with the exponential growth in sequence data. To better survey the functional and structural diversity of rubisco, we systematically sampled and synthesized a library of diverse rubisco sequences with an emphasis on clades that have previously not been characterized. Our updated phylogenetic analysis reveals that many deep-branching rubiscos assemble as dimers, supporting a dimeric origin for the superfamily -- in contrast to the ecologically dominant hexadecameric form I. Additionally, we discover and structurally characterize the largest rubisco described to date, originating from a cryptic, early-branching subclade with novel structural folds that have previously not been observed in the rubisco superfamily. By integrating biochemical data with an updated phylogenetic framework, we propose a revised nomenclature for the rubisco protein family that reflects current insights and will better accommodate future discoveries.

biochemistry↗

bioRxiv2025

Biochemical and structural characterization of glycosyltransferase family 61 proteins reveal a key determinant of sugar donor specificity

Xylan, the most abundant non-cellulosic polymer in plant cell walls, is structurally diverse, especially in grasses where it is heavily substituted with arabinofuranose and further modified by various residues. Common substitutions across species include glucuronic and 4-O-methyl-glucuronic acid. Arabinose and xylose sidechains are synthesized by glycosyltransferase family 61 (GT61) proteins, many of which remain uncharacterized in plants, with limited structural and mechanistic understanding. In this study, we identified two novel GT61 enzymes in Sorghum bicolor, functioning as xylan arabinosyltransferase (SbXAT) and xylan xylosyltransferase (SbXXT). We resolved the crystal structure of SbXAT, which exhibits a GT-B fold with two Rossmann-like domains linked by a cleft that accommodates the catalytic site. Structural comparison with a predicted SbXXT model revealed a substrate-binding residue critical for sugar donor specificity, validated through site-directed mutagenesis and enzymatic assays. These findings enhance understanding of xylan biosynthesis and provide a foundation for engineering glycosyltransferases and predicting their functions.

biochemistry↗

bioRxiv2024

A polyketide-based biosynthetic platform for diols, amino alcohols and hydroxyacids

Medium- and branched-chain diols and amino alcohols are important industrial solvents, polymer building blocks, cosmetics and pharmaceutical ingredients, yet biosynthetically challenging to produce. Here, we present a novel approach utilising a modular polyketide synthase (PKS) platform for the efficient production of these compounds. This platform takes advantage of a versatile loading module from the rimocidin PKS and NADPH-dependent terminal thioreductases (TRs), previously untapped in engineered PKSs. Reduction of the terminal aldehyde with specific alcohol dehydrogenases enables production of diols, oxidation enables production of hydroxy acids, and transamination with specific transaminases enables production of various amino alcohols. Furthermore, replacement of the malonyl-coenzyme A (CoA)-specific acyltransferase (AT) in the extension module with methyl- or ethylmalonyl- CoA-specific ATs enables production of branched-chain diols and amino alcohols. In total, we demonstrated production of nine 1,3-diols (including the difficult-to-produce insect repellent and cosmetic ingredient 2-ethyl-1,3-hexanediol), six amino alcohols, and two carboxylic acids using our PKS platform in Streptomyces albus. Finally, tuning production of the PKS acyl-CoA substrates enabled production of high titers of specific diols and amino alcohols (1 g/L diol titer in shake flasks), demonstrating high tunability and efficiency of the platform.

bioengineering↗

bioRxiv2023

Identification of a putative rhamnogalacturonan-II CMP-beta-Kdo transferase through a callus-based gene editing method which overcomes embryo lethality.

Rhamnogalacturonan II (RG-II) is a structurally complex and conserved domain of the pectin present in the primary cell walls of vascular plants. Borate crosslinking of RG-II is required for plants to grow and develop normally. Mutations that alter RG-II structure also affect crosslinking and are lethal or severely impair growth. Thus, few genes involved in RG-II synthesis have been identified. Here we developed a method using CRISPR/Cas9-mediated gene to generate callus carrying loss-of-function mutations in the MPG2 gene that encodes a putative family GT29 glycosyltransferase. Plants homozygous for this mutation do not survive. We show that in the callus mutant cell walls, RG-II does not crosslink normally because it lacks 3-deoxy-D-manno- octulosonic acid (Kdo) and thus cannot form the -L-Rhap-(1[->]5)--D-kdop-(1[->] sidechain. We suggest that MGP2 encodes an inverting CMP-{beta}-Kdo transferase (RCKT1). Our discovery provides further insight into the role of sidechains in RG-II dimerization.

plant biology↗

bioRxiv2022

A highly active Burkholderia polyketoacyl-CoA thiolase for production of triacetic acid lactone

Triacetic acid lactone (TAL) is a platform chemical biosynthesized primarily through decarboxylative Claisen condensation by type III polyketide synthase 2-pyrone synthase (2-PS). However, this reaction suffers from intrinsic energy inefficiency and feedback inhibition by and competition for malonyl-CoA. TAL production through non-decarboxylative Claisen condensation by polyketoacyl-CoA thiolase alleviates many of these disadvantages. We discovered five more thiolases with TAL production activity by exploring homologs of a previously reported polyketoacyl-CoA thiolase, BktB, from Cupriavidus necator. Among them, the BktB homolog from Burkholderia sp. RF2-non_BP3 has [~] 30 times higher in vitro and in vivo TAL production activity and led to [~]10 times higher TAL titer than 2-PS when expressed in Escherichia coli, achieving a titer of 2.8 g/L in fed-batch fermentations. This discovery of a novel polyketoacyl-CoA thiolase with superior TAL production activity paves the way for realization of total biomanufacturing of TAL.

bioengineering↗

bioRxiv2022

A membrane-associated light harvesting model is enabled by functionalized assemblies of gene-doubled TMV proteins

Photosynthetic light harvesting requires efficient energy transfer within dynamic networks of light harvesting complexes embedded within phospholipid membranes. Artificial light harvesting models are valuable tools for understanding the structural features underpinning energy absorption and transfer within chromophore arrays. Most artificial light harvesting complexes are static or in the solution phase, rather than in a two-dimensional fluid environment as in natural photosynthesis. We have developed a method for attaching a protein-based light harvesting model to a supported lipid bilayer (SLB), which provides an extended fluid membrane surface stably associated with a solid substrate. The protein model consisted of the tobacco mosaic viral capsid proteins (TMV) that were gene-doubled to create a tandem dimer (dTMV). Assemblies of dTMV were shown to break the facial symmetry of the double disk to allow for differentiation between the disk faces. Single reactive lysine and cysteine residues were incorporated into opposing surfaces of each monomer of the dTMV assemblies. This allowed for the site-selective attachment of both chromophores for light absorption and a peptide for attachment to the SLB. A cysteine modification strategy using the enzyme tyrosinase was employed for the bioconjugation of a peptide containing a polyhistidine tag for association with SLBs. The dual-modified dTMV complexes showed significant association with SLBs and exhibited mobility on the bilayer. The techniques used herein offer a new method for protein-surface attachment and provide a platform for evaluating excited state energy transfer events in a dynamic, fully synthetic artificial light harvesting system. Significance StatementHere we have constructed a model photosynthetic membrane containing proteins, chromophores, lipids, and aqueous components, all of which can be modified in their composition. This model is based on an asymmetric disk assembly consisting of engineered tandem dimers of the tobacco mosaic viral capsid protein (dTMV). We have developed methods to achieve dye conjugation and attachment of a supported lipid bilayers (SLB) site selectively on distinct protein surfaces. These dye-labeled protein complexes exhibit mobility on the SLB, resulting in a dynamic model of light harvesting membranes using entirely synthetic components. Additionally, this unique asymmetric assembly of TMV and the facile methods for protein functionalization are expected to expand the tunability of model light harvesting systems.

bioengineering↗

bioRxiv2021

Accurate prediction of protein structures and interactions using a 3-track network

DeepMind presented remarkably accurate protein structure predictions at the CASP14 conference. We explored network architectures incorporating related ideas and obtained the best performance with a 3-track network in which information at the 1D sequence level, the 2D distance map level, and the 3D coordinate level is successively transformed and integrated. The 3-track network produces structure predictions with accuracies approaching those of DeepMind in CASP14, enables rapid solution of challenging X-ray crystallography and cryo-EM structure modeling problems, and provides insights into the functions of proteins of currently unknown structure. The network also enables rapid generation of accurate models of protein-protein complexes from sequence information alone, short circuiting traditional approaches which require modeling of individual subunits followed by docking. We make the method available to the scientific community to speed biological research. One-Sentence SummaryAccurate protein structure modeling enables rapid solution of structure determination problems and provides insights into biological function.

bioinformatics↗

bioRxiv2020

An iron (II) dependent oxygenase performs the last missing step of plant lysine catabolism

Due to low abundance in many staple food crops, the essential amino acid lysine must be produced industrially to meet global food supply needs. Despite intensive study, manipulation, and agricultural importance, the steps of plant lysine catabolism beyond the 2-oxoadipate (2OA) intermediate remain undescribed. Recently we described a missing step in the D-lysine catabolic pathway of the bacterium Pseudomonas putida in which 2OA is converted to D-2-hydroxyglutarate (D2HG) via hydroxyglutarate synthase (HglS), an enzyme belonging to the previously uncharacterized DUF1338 protein family. Here we solve the structure of HglS to 1.1[A] resolution in the substrate-free form and in complex with 2OA. Structural similarity to hydroxymandelate synthase suggested a successive decarboxylation and intramolecular hydroxylation mechanism forming 2HG in a Fe(II)- and O2-dependent manner, which is validated experimentally. 2OA specificity was mediated by a single arginine (R74), highly conserved across nearly all DUF1338 family proteins, including in 76% of plant enzymes. In Arabidopsis thaliana, a DUF1338 homolog is coexpressed with known lysine catabolism enzymes, and mutants show significant germination rate defects consistent with disrupted lysine catabolism. Structural and biochemical analysis of the Oryza sativa homolog FLO7 revealed identical activity to HglS despite low sequence identity. Our results suggest that nearly all DUF1338 containing enzymes likely catalyze the same biochemical reaction, exerting the same physiological function across bacteria and eukaryotes. SignificanceTo meet human demands, millions of tons of lysine are produced by bacterial fermentation annually due to its low abundance in staple crops. Here, we show the last missing step of lysine catabolism in nearly all plant endosperms is likely catalyzed by an iron-dependant DUF1338-containing enzyme homologous to the bacterial hydroxyglutarate synthase. Structural and bioninformatic analyses of DUF1338-containing enzymes showed high conservation of critical catalytic and specificity-conferring residues across multiple domains of life despite low sequence identity. These results suggest that the DUF1338 family evolved a specific physiological function within lysine catabolism across multiple domains of life.

molecular biology↗