An orthogonal metabolic framework for one-carbon utilization

Abstract

Systems and methods are provided for converting C1 substrates to products containing more than one carbon without generating central metabolic building blocks as intermediates. In one embodiment, the system / method may include a biochemical pathway that enables an orthogonal platform for C1 utilization based on the Formyl-CoA Elongation (FORCE) reaction. In one embodiment, the system / method may include an acyloin condensation between formyl-CoA and a carbonyl-containing molecule. In one embodiment, the system / method may include a reaction catalyzed by the enzyme 2-hydroxyacyl-CoA lyase (HACL).

Description

[Technical field]

[0001] REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 63 / 233,589, filed August 16, 2021, the contents of which are incorporated by reference in their entirety herein.

[0002] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT This invention was made with Government support under Award Nos. DE-AR0001508 and DE-EE0008499 awarded by the U.S. Department of Energy and the Joint Genome Institute. The Government has certain rights in this invention. [Background technology]

[0003] background 1. Field of disclosure The present disclosure relates to the field of bioconversion of carbon-containing substrates to products.

[0004] 2. Background Description Metabolism is one of the conserved features of all living cells that has evolved to optimize the survival and reproductive potential of the organism. All living organisms, from single-celled bacteria to organisms as complex as humans, display a canonical metabolic system that has evolved with a central metabolism that serves as a link between catabolic (degradation) and anabolic (synthesis) pathways in which carbon and energy sources are utilized for the biosynthesis of cellular components and metabolic products. This structure, often referred to as the "bow-tie" or "hourglass" system, can be exemplified by the metabolism of one-carbon ("C1") substrates.

[0005] However, the robustness and plasticity that most possess presents substantial engineering difficulties. For example, out of the need to balance the production and use of precursor metabolites, many cells have evolved complex and ordered mechanisms to control pathway flux. Furthermore, as metabolic pathways have been shaped by evolution and given fitness advantages as opposed to maximizing the utilization or synthesis of desired compounds, natural pathways often limit the achievable product titers, rates and yields due to inherent metabolic inefficiencies and negative crosstalk between product formation and growth-sustaining reactions. This is again summarized by the C1 substrate, as current approaches toward biological product synthesis from these molecules, including both natural and synthetic pathways, must simultaneously engineer C1 assimilation, central metabolism and product synthesis pathways for the efficient production of target chemicals. Thus, although C1 molecules can be utilized by microorganisms for growth, the biological production of industrial chemicals from C1 feedstocks and the efficient growth of industrial organisms on C1 substrates remain open challenges, regardless of the incentives and opportunities for such processes. To date, attempts to engineer C1 bioconversion, whether developing synthetic pathways or de novo enzyme designs, have relied on central carbon metabolism to utilize various C1 substrates as carbon sources and / or for product synthesis. These designs have required rewiring the host metabolic network to accommodate C1 bioconversion, which has proven to be a challenge. Summary of the Invention

[0006] overview The creation of products such as industrial chemicals from C1 substrates can be facilitated by the use of orthogonal biological systems. Orthogonal systems are less likely to affect or be affected by their biological context, allowing for more predictable and consistent behavior compared to modifying or repurposing natural parts. Orthogonal metabolic paradigms can be a powerful alternative to genetic engineering within bowtie systems, since routes to target products that bypass central metabolism can be more direct (i.e., fewer steps) and avoid many inherent regulatory mechanisms and carbon and energy inefficiencies.

[0007] Specifically for C1 substrates, since they do not have any pre-existing carbon-carbon bonds, there is in principle no reason to first create the central metabolic building blocks on the way to the desired product. Thus, given the appropriate biochemistry, an opportunity exists to establish an orthogonal metabolic framework based on the direct synthesis of products from C1 molecules.

[0008] This disclosure relates to the conceptualization and design of biochemical pathways that enable an orthogonal platform for C1 utilization based on the formyl-CoA elongation (FORCE) reaction. This approach relies on acyloin condensation between formyl-CoA and carbonyl-containing molecules. These reactions can be catalyzed by the enzyme 2-hydroxyacyl-CoA lyase (HACL). HACL (and the related enzyme oxalyl-CoA decarboxylase, OXC17) can use carbonyl-containing acceptors of a wide range of chain lengths and functionalizations, such as the C1 compound formaldehyde, to generate acyl-CoAs amenable to a wide range of biochemical transformations. In addition to product synthesis, the FORCE pathway can serve as a basis for enabling growth on unnatural C1 substrates through the generation of multicarbon compounds that are natural substrates for the host microorganism.

[0009] In one aspect of the disclosure, the recombinant 2-hydroxyacyl-CoA synthase is capable of enzymatic formation of 2-hydroxyacyl-CoA from a carbonyl-containing compound and formyl-CoA at a rate at least two-fold, alternatively three-fold, higher than the Rhodospiralle bacterial URHD0017 2-hydroxyacyl-CoA synthase.

[0010] In some embodiments, the carbonyl-containing compound is selected from the group consisting of an aldehyde and a ketone. In some embodiments, the 2-hydroxyacyl-CoA synthase is selected from Table 1 or Table 2. In some embodiments, the 2-hydroxyacyl-CoA synthase has 90% or greater identity to an enzyme in Table 1 or Table 2.

[0011] Another aspect of the present disclosure is an enzyme that further comprises producing formyl-CoA by contacting a one-carbon substrate with the enzyme catalyst. In another aspect, the enzyme further comprises converting the substrate to a carbonyl-containing compound by contacting the substrate with the enzyme catalyst. In another aspect, the enzyme further comprises converting 2-hydroxyacyl-CoA to an organic chemical product. In yet another aspect, the carbonyl-containing compound is an aldehyde having at least one substituent.

[0012] In some embodiments, the aldehyde substituent can be a hydroxyl, carbonyl, alkyl, or amine. In some embodiments, the carbonyl-containing compound is a ketone, and the ketone is a methyl ketone.

[0013] In some embodiments, the one-carbon substrate is formaldehyde and the formyl-CoA generating enzyme catalyst is an acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA.

[0014] In some embodiments, the one-carbon substrate is methanol and the enzyme catalyst that produces formyl-CoA is methanol dehydrogenase, which catalyzes the conversion of methanol to formaldehyde; and acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA.

[0015] In some embodiments, the one-carbon substrate is methane and the enzyme catalysts that generate formyl-CoA are methane monooxygenase, which catalyzes the conversion of methane to methanol; methanol dehydrogenase, which catalyzes the conversion of methanol to formaldehyde; and acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA.

[0016] In some embodiments, the one-carbon substrate is formate and the enzyme catalyst that generates formyl-CoA is an acyl-CoA synthase that catalyzes the conversion of formate to formyl-CoA or a formate kinase that catalyzes the conversion of formate to formyl-phosphate and a phosphate formyl-transferase that catalyzes the conversion of formyl-phosphate to formyl-CoA.

[0017] In some embodiments, the one-carbon substrate is carbon dioxide and the enzyme catalyst that generates formyl-CoA is carbon dioxide reductase, which catalyzes the conversion of carbon dioxide to formate; and acyl-CoA synthase, which catalyzes the conversion of formate to formyl-CoA; or formate kinase, which catalyzes the conversion of formate to formyl-phosphate, and phosphate formyl-transferase, which catalyzes the conversion of formyl-phosphate to formyl-CoA.

[0018] In some embodiments, the product derived from 2-hydroxyacyl-CoA is an aldehyde and the enzyme catalyst that converts 2-hydroxyacyl-CoA to the product is an acyl-CoA reductase, which catalyzes the conversion of 2-hydroxyacyl-CoA to an aldehyde.

[0019] In some embodiments, the product derived from 2-hydroxyacyl-CoA is an alcohol, and the enzyme catalyst that converts 2-hydroxyacyl-CoA to the product is an acyl-CoA reductase, which catalyzes the conversion of 2-hydroxyacyl-CoA to an aldehyde; and an alcohol dehydrogenase (aldehyde reductase), which catalyzes the conversion of the aldehyde to an alcohol.

[0020] In some embodiments, the product derived from 2-hydroxyacyl-CoA is a carboxylic acid, and the enzyme catalyst that converts 2-hydroxyacyl-CoA to the product is a thioesterase that catalyzes the conversion of 2-hydroxyacyl-CoA to a carboxylic acid.

[0021] In some embodiments, the enzyme catalysts are contained in a recombinant microorganism that has the genes for expressing the respective enzymes.

[0022] In some embodiments, the substrate is contacted with a recombinant microorganism containing the enzyme catalyst in an aqueous medium that optionally contains buffers, salts, vitamins, and minerals.

[0023] In some embodiments, the substrate is contacted with the enzyme catalyst by addition to each of them of an aqueous reaction mixture that optionally contains buffers, salts, vitamins, minerals and cofactors.

[0024] In some embodiments, the enzyme catalyst is provided as a crude cell extract, hi some embodiments, the enzyme catalyst is provided as a purified protein.

[0025] One aspect of the disclosure is a modified organism comprising at least one heterologous enzyme described herein, wherein the organism is capable of growing on a one-carbon substrate, and the organism is incapable of growing on the one-carbon substrate without the heterologous enzyme.

[0026] In some embodiments, the modified organism is a bacterium. In some embodiments, the organism comprises two or more modified enzymes. [Brief description of the drawings]

[0027] BRIEF DESCRIPTION OF THE DRAWINGS [Figure 1] FIG. 1 is a diagram of the canonical (a) and synthetic (b, c) metabolic schemes for biological C1 utilization and a flow chart therefor. [Diagram 2] FIG. 2 is a diagram of and a flow chart for the FORCE pathway for product synthesis from a C1 substrate. [Diagram 3] FIG. 3 shows a graph of the analysis of the FORCE pathway. [Figure 4] FIG. 4 shows a graph of in vitro evaluation of the core module of the FORCE pathway using purified enzymes. [Diagram 5] FIG. 5 shows a graph of specific cell-free prototyping of the alpha-reduced variant of the FORCE product synthesis pathway. [Figure 6] FIG. 6 shows diagrams and graphs of stationary cell bioconversion of the C1 substrate formaldehyde using the aldose elongation and alpha reduction variant of the FORCE pathway. [Figure 7] FIG. 7 shows diagrams and graphs of FORCE pathway implementation in growing cell cultures using methanol as a C1 substrate. [Figure 8] FIG. 8 shows diagrams of simulated flux maps from a genome-scale E. coli model for growth using the FORCE pathway variants: a) (form)aldehyde elongation, b) alpha reduction, c) aldose elongation. [Figure 9] FIG. 9 shows a diagram of and graphs for a two-lineage system to assess the ability of the FORCE pathway to enable growth on a C1 substrate. [Figure 10] FIG. 10 is a diagram of an integrated representation of the orthogonal C1 pathway concept. [Figure 11] FIG. 11 is a diagram of the alternative FORCE pathway based on dehydration and alpha-reduction of 2-hydroxyacyl-CoA. [Figure 12]FIG. 12 is a graph of the effect of the NADH / NAD+ ratio on formaldehyde (top) and methanol (bottom) conversion to glycolate or acetate via the FORCE pathway. [Figure 13] FIG. 13 is a graph of the effect of termination on the repeat aldose elongation pathway. [Figure 14] FIG. 14 is a graph of the production of glycolate from formate by engineered E. coli using the formate activation pathway. [Figure 15] FIG. 15 is a graph of paraformaldehyde solubilization rates and resting cell biotransformation by paraformaldehyde. [Figure 16] FIG. 16 shows images of profiles for glycolate, formate and formaldehyde concentrations and a graph of cell growth of the sensor strain in a two-line system with 5 mM paraformaldehyde. [Figure 17] FIG. 17 shows images of profiles for glycolate, formate and formaldehyde concentrations and a graph of cell growth of the sensor strain in a two-line system with 500 mM methanol. [Figure 18] FIG. 18 shows images of the profiles for glycolate and formaldehyde concentrations in a two-line system with 1 mM formaldehyde and 10 mM formate. [Figure 19] FIG. 19 shows the bioprospecting strategy used for the identification of 34 2-hydroxyacyl-CoA synthase (HACS) variants using the Rhodospirillales bacterium URHD0017 (RuHACL) as the starting reference gene. [Figure 20]FIG. 20 shows (a) glycolate production from formaldehyde via 2-hydroxyacyl-CoA synthase (HACS) and acyl-CoA reductase (LmACR) activity, (b) a high-throughput stationary cell bioconversion platform for screening HACS variants, (c) screening results of the initial 29 HACS variants and RuHACL against glycolate productivity per cell density (uM / OD) and level of expression. [Figure 21] Figure 21 shows (a) glycolate production from formaldehyde via 2-hydroxyacyl-CoA synthase (HACS) and acyl-CoA reductase (LmACR) activity. 5 mM formaldehyde is used as the sole carbon source. (b) Two inducible promoters, HACS under the control of IPTG-inducible T7 promoter and LmACR and EcAldA under the control of cumate-inducible T5 promoter, are used for analysis. (c) Glycolate productivity (μM / OD) using RuHACLG390N as HACS with varying IPTG and cumate concentrations, (d) Glycolate productivity (μM / OD) using JGI15 as HACS with varying IPTG and cumate concentrations. [Figure 22] FIG. 22 shows (a) glycolate production from co-feeding of formaldehyde (0.5 mM or 5 mM) and formate (20 mM). Expression of HACS and FAE (AbfT) is independently controlled by inducible promoters (IPTG and cumate). (b) HACS screening results for the starting reference, RuHACL and the two best variants from the initial round. C1-C1 condensation reaction activity is represented by glycolate productivity under 0.5 mM or 5 mM formaldehyde (FALD). [Diagram 23]FIG. 23 shows (a) the protein structure as modeled in JGI15 using AlphaFold2, (b) the protein structure as modeled in JGI20 using AlphaFold2, and (c) two ligands (thiamine diphosphate and formyl-CoA) bound to the active site aligned with the crystal structure of OfOXC (PDB code: 2JI8). [Figure 24] FIG. 24 shows (a) active site residue identification by selecting amino acid residues within 3.5 Å from thiamine diphosphate (TPP), (b) active site residue identification by selecting amino acid residues within 3.5 Å from formyl-CoA (CoA), (c) alanine screening results scanning active site residues represented by glycolate productivity (μM / OD / h) for wild type JGI20 using formaldehyde (0.5 or 5 mM) and formate (20 mM) systems with AbfT co-expression. [Diagram 25] Figure 25 shows (a) sequence analysis of first round JGI variants at the c-terminus. Active variants (with asterisks) show the highly conserved RKPQQF-W residues. (b) Screening results of alanine scanning for conserved c-terminal residues represented by glycolate productivity (μM / OD / h) for wild type JGI20 using formaldehyde (0.5 or 5 mM) and formate (20 mM) system with AbfT co-expression. [Figure 26] Figure 26 shows (a) the alignment of JGI15 and JGI20 AlphaFold structural alignment represented by aligned amino acid residues. N461 of JGI20 and R493 of JGI15 are the two residues that are not aligned between the two variants. (b) Screening results of structural hybrids of the two proteins with single residue insertions / deletions represented by glycolate productivity (μM / OD / h) for wild type JGI20 using formaldehyde (0.5 or 5 mM) and formate (20 mM) systems with AbfT co-expression. [Figure 27]FIG. 27 shows (a) the sequence and structure of the c-terminal tail "covering loop" between JGI15 and JGI20, (b) the screening results of structural hybrids by replacing the JGI20 c-terminus with JGI15, represented by glycolate productivity (μM / OD / h) relative to wild-type JGI20 using a formaldehyde (0.5 or 5 mM) and formate (20 mM) system with AbfT co-expression. [Figure 28] Figure 28 shows (a) glycolate production from formaldehyde via 2-hydroxyacyl-CoA synthase (HACS) and acyl-CoA reductase (LmACR) activity. 5 mM formaldehyde is used as the sole carbon source. (b) Two inducible promoters, HACS under the control of IPTG-inducible T7 promoter and LmACR and EcAldA under the control of cumate-inducible T5 promoter, are used for analysis. (c) Glycolate productivity of RuHACL, three best first round mutants and AcHACL under 5 mM formaldehyde. [Figure 29] Figure 29, Identification of second round JGI HACS homologs based on AcHACL, JGI19, JGI15 and JGI20 as starting references. The phylogenetic tree includes all first and second round HACS variants as well as HACL and OXC available in the literature. [Diagram 30] Figure 30 shows (a) glycolate production from co-feeding of formaldehyde (0.5 mM) and formate (20 mM). Expression of HACS and FAE (AbfT) are independently controlled by inducible promoters (IPTG and cumate). (b) HACS screening results of promising second round variants represented by % change in glycolate productivity with respect to JGI15 as reference. [Diagram 31] Figure 31 shows (a) 2-hydroxyacid production from co-feeding of aldehyde and formate. (b) Expression of HACS and FAE (CaAbfT) is independently controlled by inducible promoters (IPTG and cumate). [Diagram 32]FIG. 32 shows (a) lactic acid (lactate) production from co-feeding acetaldehyde (5 mM) and formate (20 mM), (b) HACS screening results of promising first round variants represented by lactate productivity (μM / OD), (c) HACS screening results of promising second round variants represented by % change in lactate productivity with respect to JGI15 as reference. [Diagram 33] FIG. 33 shows (a) 2-hydroxybutyrate (2HB) production from co-feeding of propionaldehyde (5 mM) and formate (20 mM), (b) HACS screening results of promising first round variants represented by 2HB productivity (μM / OD), (c) HACS screening results of promising second round variants represented by % change in 2HB productivity with respect to JGI15 as reference. [Diagram 34] FIG. 34 shows (a) glyceric acid (glycerate) production from co-feeding of glycolaldehyde (5 mM) and formate (20 mM), (b) HACS screening results of promising first round variants represented by glycerate productivity (μM / OD). [Diagram 35] FIG. 35 shows (a) tartronic acid (tartronate) production from co-feeding of glyoxylate (5 mM) and formate (20 mM), (b) HACS screening results of promising first round variants represented by tartronate productivity (μM / OD). [Diagram 36] FIG. 36 shows (a) 2,4-dihydroxybutyrate (DHB) production from co-feeding of 3-hydroxypropionaldehyde (5 mM) and formate (20 mM), (b) HACS screening results of promising first round variants represented by DHB productivity (μM / OD). [Figure 37]Figure 37 shows the first round HACS screening with acetone and formate. (a) 2HIB production from co-feeding of 100 mM acetone and 20 mM formate. Expression of HACS and FAE (CaAbfT) is independently controlled by inducible promoters (IPTG and cumate). (b) HACS screening results of promising HACS variants. [Figure 38] Figure 38 shows methyl ketones as substrates for condensation with formyl-CoA using purified enzymes. (a) Pathway for condensation of methyl ketones and formyl-CoA from formate to 2-hydroxy-2-methyl acid; (b) GC-MS results of in vitro assay with different methyl ketones and formates. The desired product is indicated by a blue arrow. [Figure 39] FIG. 39 shows the production of 2-hydroxyacids, 3-hydroxyacids, alcohols, 1,2-diols, and α,β-unsaturated acids from the condensation of carboxylic acid-derived ketones and formyl-CoA. [Diagram 40] FIG. 40 shows the production of 2-hydroxyisobutyric acid, 3-hydroxyisobutyric acid, isobutanol, isobutene glycol, and methacrylic acid from the condensation of lactic acid derived acetone and formyl-CoA. [Diagram 41] FIG. 41 shows the production of 2-hydroxy-2-methylbutanoic acid, 3-hydroxy-2-methylbutanoic acid, 2-methylbutan-1-ol, 2-methylbutane-1,2-diol, and 2-methylbut-2-enoic acid from the condensation of 2-hydroxybutanoic acid-derived butanone and formyl-CoA. [Diagram 42] FIG. 42 shows the production of 2-hydroxy-2-methylpentanoic acid, 3-hydroxy-2-methylpentanoic acid, 2-methylpentan-1-ol, 2-methylpentane-1,2-diol, and 2-methylpent-2-enoic acid from the condensation of 2-hydroxypentanoic acid-derived pentanone and formyl-CoA. [Diagram 43]FIG. 43 shows the production of 2-hydroxy-2-methylheptanoic acid, 3-hydroxy-2-methylheptanoic acid, 2-methylheptan-1-ol, 2-methylheptane-1,2-diol, and 2-methylhept-2-enoic acid from the condensation of 2-hydroxyheptanoic acid-derived heptanone and formyl-CoA. [Diagram 44] FIG. 44 shows the production of 2,3-hydroxy-2-methylpropanoic acid, 2-methylpropane-1,3-diol, 2-methylpropane-1,2,3-triol, and 3-hydroxy-2-methylacrylic acid from the condensation of 2,3-hydroxypropanoic acid-derived hydroxyacetone and formyl-CoA. [Diagram 45] FIG. 45 shows the production of 2-hydroxy-2,3-dimethylbutanoic acid, 3-hydroxy-2,3-dimethylbutanoic acid, 2,3-dimethylbutan-1-ol, 2,3-dimethylbutane-1,2-diol, and 2,3-dimethylbut-2-enoic acid from the condensation of 2-hydroxy-3-methylbutanoic acid-derived 3-methyl-2-butanone and formyl-CoA. [Figure 46] FIG. 46 shows the production of 2-hydroxy-2-methyl-3-oxopropanoic acid, 2-methylpropane-1,3-diol, and 2-methylheptane-1,2,3-triol from the condensation of 2-hydroxy-3-oxopropanoic acid-derived methylglyoxal and formyl-CoA. [Figure 47] FIG. 47 shows the production of 2-hydroxy-2-methyl-4-oxopentanoic acid, 3-hydroxy-2-methyl-4-oxopentanoic acid, 5-hydroxy-4-methylpentan-2-one, 4,5-dihydroxy-4-methylpentan-2-one, and 2-methyl-4-oxopent-2-enoic acid from the condensation of 2-hydroxy-4-oxopentanoic acid derived pentane-2,4-dione and formyl-CoA. [Figure 48]FIG. 48 shows (a) glycolate production from formaldehyde via 2-hydroxyacyl-CoA synthase (RuHACL) and acyl-CoA reductase (ACR) variants, (b) screening results of initial ACR variants identified from the literature. [Figure 49] Figure 49 shows (a) a diagram of a phylogenetic tree of ACR variants identified using LmACR as the starting reference. (b) Screening of ACR variants was performed by measuring formaldehyde consumption under ACR variant overexpression. (c) Screening results of ACR variants under 0.5 mM and 3 mM formaldehyde concentrations represented by % change in formaldehyde consumption activity with respect to LmACR as the reference. [Figure 50] Figure 50 shows (a) glycolate production from formaldehyde and formate via 2-hydroxyacyl-CoA synthase (JGI15) and formate-activating enzyme variants. PTA-ACK makes formyl-CoA via a formyl-phosphate intermediate, and ACT converts CoA directly to formate from a CoA donor molecule. (b) Screening results of initial ACT & PTA-ACK variants identified from the literature. [Figure 51] FIG. 51 shows (a) a diagram of a phylogenetic tree of ACT variants identified using AbfT as a starting reference, and (b) a diagram of a phylogenetic tree of ACT variants identified using CcAck-Pta as a starting reference. [Figure 52] Figure 52 shows (a) glycolate production from formaldehyde and formate via 2-hydroxyacyl-CoA synthase and formate-activating enzyme variants. PTA-ACK makes formyl-CoA via a formyl-phosphate intermediate, and ACT converts CoA directly to formate from a CoA donor molecule. (b) Screening results of promising ACT and ACK-PTA variants for glycolate productivity (μM / OD / h). [Diagram 53]Figure 53 shows (a) the engineering of a glycolate autotrophic strain with glycolate as the sole source for glycine synthesis. The strain grows only when either supplemented with glycine or provided with glycolate along with the appropriate enzymes to synthesize glycine. (b) The engineered glycolate autotrophic strain can grow only with glycolate supplementation showing higher growth rates with increasing glycolate concentrations. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] Detailed Description The term "about," as used herein, refers to variations in numerical quantities that may occur through, for example, typical measuring and manufacturing procedures used for articles of footwear or other articles of manufacture that may include embodiments disclosed herein; inadvertent errors in these procedures; differences in manufacture, source, or purity of ingredients used to make a composition or mixture or to carry out the method; and the like. Throughout this disclosure, the terms "about" and "approximately" refer to a range of ±5% of the value of the numerical value that the term precedes.

[0029] In the canonical "bow-tie" system of metabolism, substrates are centralized in a central metabolism with the resulting biosynthetic building blocks and products of interest derived from the central metabolite. To date, attempts to engineer C1 biotransformations have required a central carbon metabolism for utilization of C1 substrates and their conversion to the products of interest. These designs, which exhibit minimal orthogonality, require optimizing the host metabolic network to accommodate C1 biotransformations, which has proven to be a challenge.

[0030] However, the implementation of the formyl-CoA elongation (FORCE) pathway, which allows C1 utilization and bioconversion in an orthogonal manner to the host metabolism, may resolve these challenges. The FORCE pathway is based on the use of formyl-CoA as an anabolic metabolite, which is enabled by the acyloin condensation reaction between formyl-CoA and carbonyl-containing substrates catalyzed by 2-hydroxyacyl-CoA lyase (HACL). Product synthesis is achieved with relatively high orthogonality to central metabolism compared to other approaches. Our analysis of pathway thermodynamics suggested favorable driving forces for the FORCE pathway conversion of formate, formaldehyde and methanol to glycolate or acetate as exemplary products. Self-contained orthogonal pathways have been shown to be potentially feasible both in vitro (purified enzymes and cell extracts) and in vivo (quiescent and growing cells) implementations, where diverse functional products (e.g. glycolate, glycolaldehyde, ethylene glycol, ethanol, glycerate) can be made in a growth- and host metabolism-dependent manner using formaldehyde, formate or methanol as the sole C1 substrate. The product synthesis shown here completely bypasses central metabolism, which is distinct from all other approaches reported to date. Potential bioprocesses can be envisioned in which growth and biocatalyst maintenance are carried out with multicarbon substrates, and biocatalysts are used for C1 bioconversion. Bioprocesses of this nature based on multienzyme cascades and two-stage fermentation have been the subject of recent reviews.

[0031] Design of orthogonal metabolic schemes for C1 utilization and product synthesis The FORCE pathway that can provide the bioconversion of some embodiments of C1 substrate to desired product is discussed below and shown in the figures.With specific reference to the example shown in Figure 1a and 1b, some embodiments of the system can have three main features of an orthogonal metabolic system: 1) activation of C1 substrate to building blocks suitable for carbon chain elongation; 2) repeated carbon chain elongation by one carbon per cycle; and 3) the end of the pathway that results in the accumulation of the desired product.For example, in embodiments, an orthogonal metabolic system with these features can be implemented using formyl-CoA as the activated C1 unit for repeated carbon chain elongation.

[0032] In the existing literature, there are scattered reports of the production of formyl-CoA from C1 molecules. However, acyl-CoA is a convenient intermediate between the carboxylate and aldehyde forms. As a result, it is possible to generate formyl-CoA from both oxidized and reduced C1 substrates, as shown in the one-carbon activation panel of Figure 2. From formaldehyde, formyl-CoA can be generated by the activity of acyl-CoA reductase (ACR), and methanol can be converted to formaldehyde by methanol dehydrogenase (MDH).

[0033] Formyl-CoA can be made from formate using CoA transferase. Formyl-CoA transferase is one such enzyme known to involve formate and formyl-CoA in CoA thioether transfer. Activation of formate to formyl-CoA by the promiscuous activity of acetyl-CoA synthetase (ACS) from Escherichia coli (EcACS) is all possible. The reaction catalyzed by EcACS is AMP formation (consuming 2 ATP equivalents), but evidence for an ADP formation pathway exists via the intermediate formyl-phosphate. In this pathway, formate is converted to formyl-phosphate by formate kinase (FOK), and phosphotransacylase (PTA) converts formyl-phosphate to formyl-CoA. It may also be possible to convert formate to formyl-CoA in an ATP-dependent manner via the direct reduction of formate to formaldehyde by formaldehyde dehydrogenase (FaldDH). Although such a conversion is thermodynamically challenging, as shown in Figure 2, this reaction has been demonstrated in a cell-free system and could be potentially useful for both in vitro and in vivo implementations. CO can also be converted to formate by the reverse activity of formate dehydrogenase (or carbon dioxide reductase), and methane can be converted to methanol by methane monooxygenase, which when coupled to the above reactions can result in the formation of formyl-CoA.

[0034] Orthogonal de novo construction of diverse carbon skeletons by extension using C1 units requires an iterative pathway similar to that found in nature that builds carbon skeletons from C2-C5 metabolites, but outside of the central metabolism. 2-Hydroxyacyl-CoA lyase (HACL) has a wide range of carbon chain length specificity, making it a good candidate for establishing an iterative pathway. There are many potential reaction pathways that could enable iteration by converting the product of the HACL catalyzed reaction, 2-hydroxyacyl-CoA, to an aldehyde that can be further extended by formyl-CoA. As shown in Figure 11, dehydration at the α-carbon is possible, similar to the mechanism of the well-established acrylate pathway, which converts 2-hydroxyacyl-CoA to 2-enoyl-CoA. Since 2-enoyl-CoA is involved in β-oxidation, the production of 2-enoyl-CoA is also convenient, potentially enabling the use of the enzyme toolkit and knowledge established for the platform pathway β-oxidation reversal. However, dehydration of 2-hydroxyacyl-CoA is much more difficult than dehydration of 3-hydroxyacyl-CoA in β-oxidation reversal and therefore requires an oxygen-sensitive radical mechanism, which also requires the presence of a β-carbon, limiting the pathway to implementation of intermediates of 3 carbons or larger.

[0035] These issues make the conversion of thioesters a more promising route. As shown in Figure 2 and the Formyl-CoA Elongation panel, reduction of CoA-thioesters produces 2-hydroxyaldehydes, which is possible due to the non-specific activity of certain acyl-CoA reductases (ACRs). Furthermore, ligation of 2-hydroxyaldehydes with formyl-CoA by HACL produces polyhydroxyacyl-CoAs and further polyhydroxyaldehydes, commonly known as aldoses. Polyhydroxyaldehydes can in principle serve as substrates for HACL-catalyzed reactions that can be referred to as aldose elongation, an example of which is shown in Figure 2 with the Formyl-CoA Elongation panel).

[0036] Further reduction of 2-hydroxyaldehydes to give 1,2-diols is possible through the activity of diol oxidoreductases (DORs). E. coli FucO is an example of a DOR that catalyzes the interconversion of 1,2-diols and 2-hydroxyaldehydes34. However, E. coli is only one example of a DOR, and in some instances other suitable DORs may be used instead. For example, in some embodiments, the DOR may be another prokaryotic bacterium. Alternatively, in some embodiments, the DOR may be a eukaryotic bacterium or fungus. Dehydration of 1,2-diols may be catalyzed by the activity of diol dehydratases (DDRs) to give aldehydes, effectively achieving alpha reduction. Diol dehydration also requires a radical mechanism, but B12-dependent diol dehydratases are oxygen tolerant. Further elongation of the aldehyde with formyl-CoA, which may be referred to as aldehyde elongation, results in an alkyl chain elongation similar to the two-carbon elongation in fatty acid biosynthesis or the reverse β-oxidation pathway. These pathways, including aldose elongation, may be collectively referred to as the α-reduction, and aldehyde elongation, formyl-CoA elongation (FORCE) pathways, which facilitate the use of formyl-CoA as a carbon chain elongation unit, as shown in FIG. 2 in the formyl-CoA elongation panel.

[0037] As shown in Figure 2, various product classes can be made as intermediates or from derivatives of FORCE pathway intermediates, some of which can also support microbial growth (shown in Figure 1c). For example, aldose sugars are a direct result of the 2-hydroxyaldehyde node. Diols, including major industrial chemicals such as ethylene glycol, are a result of the 1,2-diol node. Derivatives of the 2-hydroxyacyl-CoA node include 2-hydroxy acids such as the industrial products glycolic acid and lactic acid, which are made by reactions catalyzed by thioesterases. Many chemical classes can be derived from the aldehyde node, such as carboxylic acids, alcohols, and acyl-CoAs that can serve as precursors for other products.

[0038] Thermodynamic analysis of the FORCE pathway for C1 utilization Although the standard Gibbs free energies of the pathway reactions shown in Figure 2 readily reveal potential reactions, Gibbs energies are insufficient to predict whether a pathway as a whole is viable and will operate in the intended direction. For this, a holistic approach to pathway thermodynamics is required that considers the ability of pathway reactions to influence each other. To achieve this, a "Max-min Driving Force" (MDF) approach was applied to these reactions.

[0039] The MDF of the FORCE pathway for the creation of the C2 metabolites glycolate and acetate from only one C1 substrate was evaluated, but only the MDF of soluble C1 substrates was evaluated, since mass transfer limitations would likely significantly limit CO2 and methane utilization. Glycolate and acetate were selected as representative C2 products that are both pathway products and growth substrates, with glycolate requiring the shortest pathway and acetate requiring an entire sequence of aldehyde elongation reactions. As shown in Figure 3a, for each substrate, there is a driving force toward the creation of glycolate that is greater than acetate. This is due to the thermodynamically favorable hydrolysis of glycolyl-CoA by thioesterase, whereas the creation of acetate requires the thermodynamically challenging reduction of glycolyl-CoA. Using standard limitations of metabolite concentrations, formaldehyde allows for the greatest MDF, as it does not require the challenging NAD+-dependent methanol dehydrogenase reaction or the formyl-CoA reduction reaction to formaldehyde. It is also clear that despite the thermodynamic difficulties of the MDH reaction, there is sufficient driving force in the favorable direction for the net production of glycolate and acetate.

[0040] The driving force for formate utilization is minimal, where ATP hydrolysis supports the activation of formate. Hydrolysis of 2 ATP equivalents by ACS provides just enough driving force for the net production of acetate, and utilization of 1 ATP equivalent provides only enough driving force for the production of glycolate. ATP-dependent pathways are not viable under these conditions.

[0041] The above analysis estimates a standard constraint for metabolite concentrations between 1 μM and 10 mM, but in practice the C1 substrate concentration can be higher or lower than this upper limit based on the ability to supply it exogenously. Then, more strict / realistic values ​​for the substrate concentration were used based on physical limitations such as the toxicity of the C1 substrate, as shown in Figure 3. While some organisms can survive and consume formaldehyde concentrations of about 10 mM42, most organisms, e.g., the bacterium E. coli, cannot. When the upper limit of formaldehyde was adjusted to a more reasonable 0.1 mM, the MDF of the pathway dropped as expected. On the other hand, methanol, which is considerably less toxic than formaldehyde, was supplied to the E. coli growth medium at a concentration of about 100 mM43. By increasing the upper limit of the methanol concentration, the MDF of methanol conversion was increased. Interestingly, at these concentrations, the driving force for methanol utilization is slightly larger than that for formaldehyde.

[0042] Similarly, E. coli has the ability to grow in the presence of formate concentrations of about 100 mM. 9 In other embodiments, other DORs that grow in the presence of formate concentrations could also be used. Increasing the formate concentration limit had no effect on the MDF in the 1 or 2 ATP consumption scenarios, but had a large effect on the MDF of the 0 ATP pathway. With 100 mM formate, the net production of glycolate rather than acetate is possible without the need for ATP hydrolysis. This analysis can inform cell-free biotransformation systems and provide valuable insights regarding substrate uptake for in vivo implementation.

[0043] In addition to substrate concentration, the NADH / NAD+ ratio is the other major constraint on pathway thermodynamics. The constraint on NADH / NAD+ used previously was 0.141, reflecting E. coli growth under aerobic conditions, but physiological NADH / NAD+ can vary, reaching values ​​close to or even higher than 1 under anaerobic conditions. Even higher ratios can be achieved in in vitro runs. To evaluate the effect of the NADH / NAD+ ratio on pathway driving force, the NADH / NAD+ ratio was varied. As shown in Figure 12, in the physiological range (taken here from 0.1 to 1), the pathway driving force remained positive for formaldehyde and methanol as substrates. As expected, low ratios were favored when the pathway was redox producing (methanol vs. acetate / glycolate and formaldehyde vs. glycolate), whereas high ratios were favored when the pathway was redox consuming (formate vs. acetate / glycolate) or redox balanced (formaldehyde vs. acetate - presumably because the reduction reaction is thermodynamically more difficult). As shown in Figure 3b, the NADH / NAD+ ratio can be crucial to the driving force of the formate utilization pathway.

[0044] For the conversion of formate to glycolate, pathways requiring 1-2 ATP equivalents retain a positive driving force throughout most of the physiological range. However, for the conversion of formate to acetate, the NADH / NAD+ ratio must be at the upper end of the physiological range to have a positive driving force with the consumption of 1 ATP equivalent. At 10 mM formate, neither driving force for the creation of glycolate or acetate is positive in the physiological range. If the concentration of formate is increased to 100 mM, the driving force for the creation of glycolate or acetate can be positive within the physiological range of NADH / NAD+ ratios even without ATP hydrolysis. Overall, the conversion of formate to a more reduced product such as acetate is challenging both thermodynamically and based on net redox balance.

[0045] As an example of an embodiment method of converting C1 substrate to product, the ability of the FORCE pathway to support iteration due to its intermediate redox state was further evaluated using formaldehyde as an exemplary substrate. As shown in FIG. 3c, the thermodynamics of both aldose and aldehyde elongation pathways support iteration up to 4 carbons. Also, the driving force of the pathway decreases with the number of iterations. After 4 carbons, the aldose elongation mode becomes unfavorable, likely due to the cumulative effect of successive acyl-CoA reduction reactions. Meanwhile, the aldehyde elongation mode remains favored despite requiring the same acyl-CoA reduction, likely due to the thermodynamically favorable reactions catalyzed by DOR and DDR. Different C1 activation and termination pathways have an impact on the MDF of the overall elongation cycle when the number of iterations is low. As shown in FIG. 13, as the number of iterations increases, the thermodynamics of the elongation cycle reactions become dominant. Based on the previous analysis, the MDF of the aldose and aldehyde elongation pathways can be expected to be similar or lower when methanol or formate are the C1 substrates because substrate utilization is more thermodynamically constrained.

[0046] In vitro pathway validation A prerequisite for the FORCE pathway is the production of formyl-CoA and formaldehyde. To verify the functionality of these reactions, a purified enzyme system was developed to follow the formation of both formyl-CoA and the HACL condensation product glycolyl-CoA from different C1 substrates, as shown in Figure 4. As shown in Figure 2, formyl-CoA can be produced from formaldehyde by acyl-CoA reductase (ACR). We observed the formation of both formyl-CoA and glycolyl-CoA in reactions involving both Listeria monocytogenes ACR (LmACR) and HACL (RuHACL) from the Rhodospirillales bacterium URHD0017. Formyl-CoA can also be derived from oxidized C1 substrates by activation of formate. Using formyl-CoA transferase (OfFrc) from Oxalobacter formigenes and succinyl-CoA as the CoA donor, we observed the activation of formate to formyl-CoA, and the addition of formaldehyde resulted in the formation of glycolyl-CoA. We further tested whether formaldehyde could be created in situ by reduction of formyl-CoA using LmACR. Indeed, glycolyl-CoA was observed, but its abundance was lower than when formaldehyde was added directly. Together, these results suggest that in this enzyme system, limitations are imposed either by constraints placed on the ACR reaction or the need for a suitable form of NAD(H) for the activity of the enzyme. In support of the latter, in the oxidative direction (i.e., formaldehyde to formyl-CoA), the amount of glycolate observed after hydrolysis of the CoA thioester was roughly equivalent to the amount of NADH added to the reaction (1 mM). As shown in Figures 2 and 3b, in the opposite direction, a lower than equivalent amount of glycolate was observed, which is consistent with the thermodynamics of the reaction becoming less favorable as the NADH / NAD+ ratio decreases.

[0047] Using a cell-free metabolic engineering approach, the FORCE pathway for product synthesis was further prototyped. Extracts of E. coli expressing each pathway enzyme comprising the α-reduction FORCE pathway were combined in series to show pathway function in a stepwise manner. As shown in Figure 2 (in the formyl-CoA elongation panel), outside of the direct production of 2-hydroxycarboxylates (e.g., glycolate) via thioester cleavage of 2-hydroxyacyl-CoA generated by HACL, other C2 products (e.g., aldose or aldehyde elongation pathways) require reduction of this 2-hydroxyacyl-CoA (glycolyl-CoA in the case of formaldehyde and formyl-CoA linkage). As shown in Figure 5a, Listeria monocytogenes ACR (LmACR) could also work against glycolaldehyde to minimize the complexity of the engineered system, where LmACR was used in a bifunctional role to catalyze both the oxidation of formaldehyde to formyl-CoA and the reduction of glycolyl-CoA to glycolaldehyde. As shown in Figure 5b, LmACR alone only resulted in the conversion of formaldehyde to formate. Glycolate was observed, including the previously identified HACL (RuHACL) from the Rhodospirillales bacterium URHD0017. However, glycolaldehyde was not significantly detected as a product, likely due to the presence of endogenous oxidoreductases in the cell extract system, which catalyzed the oxidation of glycolaldehyde to glycolate (e.g. AldA, AldB, PuuC, PatD) or, to a lesser extent, the reduction to ethylene glycol (e.g. FucO, YqhD, AdhP, EutG and others).

[0048] As shown in Figures 5a and 5b, the synthesis of the next reduction product, ethylene glycol, was significantly increased by the addition of cell extracts of E. coli overexpressing E. coli FucO, a 1,2-diol oxidoreductase (2-fold increase, from 1.37 ± 0.1 mM to 2.73 ± 0.03 mM). Ethylene glycol could be further dehydrated to acetaldehyde by diol dehydratase (shown in Figure 5a). Upon further addition of E. coli cell extracts expressing diol dehydratase (DDR) from Klebsiella oxytoca together with coenzyme B12, ethanol was detected (1.90 ± 0.03 mM at 1 h: Figure 5b), probably due to the reduction of acetaldehyde by endogenous oxidoreductase, with a corresponding decrease in ethylene glycol. At later time points (2 h), an increase in acetate was observed, likely due to oxidation of ethanol and acetaldehyde, again due to the presence of endogenous oxidoreductase activity. The synthesis of altered products (i.e., glycolate, ethylene glycol, ethanol, glycerate, acetate) indicates that judicious selection of pathway enzymes can be used to control the synthesis of products with varying levels of reduction, chain length, and functionality from the acyl-CoA node, all independent of central metabolism.

[0049] In vivo implementation of the FORCE pathway The orthogonal nature of the FORCE pathway allows for not only rapid prototyping in cell-free systems, but also easy in vivo implementation. Figures 6 and 7 show key features of the designed platform and the synthesis of additional products and utilization of various C1 substrates using stationary and growing cultures of E. coli. A key feature of the FORCE pathway design is repetition, which can be achieved via aldose or aldehyde elongation (as shown in the formyl-CoA elongation panel in Figure 2 and Figure 3c). To demonstrate the feasibility of repetitive aldose elongation in vivo, we targeted the synthesis of the three-carbon product glycerate from formaldehyde, as shown in Figure 6a. A strain (AC440:MG1655(DE3) ΔfrmA ΔfdhF ΔfdnG ΔfdoG ΔglcD) with C1 catabolic and glycolate consumption knockouts and (over)expressing RuHACLG390N, LmACR, and EcAldA16 was used. To promote the accumulation of glycolaldehyde and its condensation with formyl-CoA, EcAldA was removed from the expression vector. Although formaldehyde consumption was significantly reduced, FIG. 6b shows that accumulation of glycolaldehyde and glycerate was observed, indicating an iterative aldose elongation pathway. In an attempt to increase production of these compounds, the genes encoding aldehyde dehydrogenase (ΔaldA ΔaldB ΔpatD ΔpuuC, collectively referred to as Δaldh) were further knocked out. Using this host, the concentration of the oxidation product, glycolate, was reduced, whereas glycolaldehyde concentration increased when EcAldA was not overexpressed. Regardless of this, the knockout did not appear to affect the accumulation of glycerate, indicating a limitation to the condensation reaction between glycolaldehyde and formyl-CoA, likely catalyzed by RuHACL. The knockout also had no effect on accumulation of the by-product formate, indicating that a likely pathway for by-product formation is via thioester hydrolysis of formyl-CoA.

[0050] The pathway was also extended beyond the production of glycolaldehyde to the next reduction product, ethylene glycol, by including E. coli fucO, known to catalyze the interconversion of glycolaldehyde and ethylene glycol, in the expression vector. As shown in FIG. 6b, this resulted in the accumulation of ethylene glycol in the extracellular medium. Further knockout of aldehyde dehydrogenase resulted in an increase of about 68% in ethylene glycol production. Interestingly, including EcFucO dramatically increased the consumption of formaldehyde, much of which was converted to formate. This could be explained by the net redox balance for ethylene glycol production from formaldehyde, which requires additional reducing equivalents in the form of NADH. This reduces the NADH / NAD+ ratio and provides additional electron acceptors for formaldehyde oxidation.

[0051] To verify that the observed products were derived from formaldehyde and not from residual multicarbon substrates or biomass components, 13C-labeled formaldehyde was used as a substrate for the engineered strain. As shown in Figure 6c, the products glycolic acid, ethylene glycol, and glyceric acid were found to be sufficiently 13C-labeled based on the characteristic [M-15]+ ion of the TMS derivatives of the products. Therefore, both the 2- and 3-carbon products could be made from formaldehyde using the FORCE pathway.

[0052] In addition to altered products, we also evaluated whether different substrates could be utilized. As shown in Figure 2, 7a, only the additional expression of methanol dehydrogenase (MDH) was required to convert methanol to formaldehyde, which is a convenient extension of the previously established formaldehyde utilization pathway. The well-tested MDH variant from Bacillus methanolicus MGA3 (BmMDH2MGA3) was expressed in combination with RuHACLG390N, LmACR and EcAldA in strain AC440. Unlike formaldehyde, whose toxicity requires the use of stationary cells, methanol can also be added directly to a growing E. coli culture. As shown in Figure 7b, when the engineered methanol-utilizing strain was grown in the presence of complex nutrients and 500 mM methanol, the formation of glycolate was observed, which was not the case in the strain not expressing RuHACL.

[0053] In an effort to improve the performance of this system, RuHACLG390N was replaced with a newly identified HACL obtained from the beach sand metagenome (UniProt accession: A0A3C0TX30), herein referred to as BsmHACL. As shown in Figure 7c, the use of BsmHACL substantially increased glycolate accumulation by about three-fold, reflecting a major bottleneck in the pathway. Despite the improved glycolate production, formate accumulation remained high. In an effort to address this issue, the termination enzyme EcAldA was replaced with a previously identified CoA-transferase (CaAbfT) from Clostridium aminobutyricum, which was found to have better properties than OfFrc51. CaAbfT serves both to release glycolate from glycolyl-CoA and to reactivate the observed by-product formate to formyl-CoA for further condensation. When CaAbfT was expressed, glycolate accumulation was further increased by about 33%, and formate accumulation was reduced by about 36%. Finally, with CaAbfT serving as a way to terminate the pathway via the release of glycolate, endogenous thioesterases were not predicted to be necessary and were presumed to be at least partially responsible for the observed formate. Therefore, a strain lacking thioesterase (ΔyciA ΔtesA ΔtesB ΔybgC ΔydiI ΔfadM) was constructed and tested with the pathway. Using a thioesterase-deficient background further reduced the observed formate, but did not eliminate it. There may be other pathways for formate production, such as direct oxidation of formaldehyde or thioesterases yet to be identified. To verify that the observed glycolate was derived from methanol, 13C-labeled methanol was further used. As shown in Figure 7d, the glycolate produced in these cultures was fully derived from 13C methanol, as observed from the [M-15]+ ion of the TMS derivative of glycolate.

[0054] Having established CaAbfT as a promising pathway for formate activation, we further evaluated whether CaAbfT could be used to enable the incorporation of exogenously supplied formate. In the engineered strains of E. coli, CaAbfT was expressed to activate formate, and LmACR was not expressed such that there was no interconversion of formaldehyde and formyl-CoA. Therefore, the observed glycolate should result from formate activation to formyl-CoA and further condensation of the resulting formyl-CoA with formaldehyde. As shown in Figure 14, in the engineered strain expressing BsmHACL, a 12-fold increase in glycolate was observed when formate was included in the medium in addition to formaldehyde compared to when formaldehyde was supplemented alone, and the total carbon accumulated as glycolate was greater than the amount initially added as formaldehyde. Thus, production of glycolate by this strain is formate dependent, serving as evidence that exogenously supplied formate can be used by the FORCE pathway as a C1 substrate.

[0055] Flux balance analysis of the FORCE pathway for synthetic methylotrophy Having demonstrated the ability of FORCE pathways to support product synthesis, their ability to enable synthetic methylotrophy in E. coli was evaluated in silico, as some of the products (e.g. glycolate, glycerate, acetate) can serve as growth substrates. A genome-scale model of E. coli, iML151552, was used to evaluate the growth of E. coli on organic C1 substrates by the addition of reactions to the model that included selected pathways reported or proposed to enable methylotrophy. All pathways were evaluated with reactions that allow the interconversion of C1 molecules at the different reduction levels present. The reactions sufficient to run each pathway are given in Table 3. [Table 1-1] [Table 1-2]

[0056] As shown in Table 1 below, the simulation results suggest that all pathways previously proposed to enable some form of methylotrophy in E. coli, both natural (ribulose monophosphate or RuMP, serine) and synthetic (formolase, synthetic acetyl-CoA or SACA, reducing glycine), may do so. [Table 2]

[0057] The FORCE pathways evaluated for the conversion of non-natural C1 substrates to the natural growth substrates glycolate, acetate and glyceraldehyde were without exception, showing another advantage of the orthogonal nature of the platform. By developing a direct route(s) to a compound(s) that represents a physiological substrate for E. coli or any other organism, the FORCE pathways can utilize the control of natural metabolism and substrate(s) utilization, as opposed to the need to integrate and engineer varying or multiple metabolic nodes. Interestingly, this in silico analysis revealed that the pathways that result in the creation of three-carbon metabolites (FORCE-glyceraldehyde, formolase, RuMP) are predicted to produce the highest biomass yields on a carbon and electron basis, as shown in Table 1 above.

[0058] Analysis of the flux distribution of the three modeled FORCE pathways is shown in FIG. 8 and provides some insight into why the creation of three-carbon metabolites may be advantageous. As shown in FIG. 8, the FORCE pathway leading to the formation of glycolate utilizes the carbon-inefficient glycolate utilization pathway present in E. coli, which requires the decarbonation condensation of two molecules of glyoxylate. The creation of more reduced C2 metabolites, such as glycolaldehyde or acetate, is favored over oxidizing C2 in the form of glycolate. The predicted metabolism of glycolaldehyde is particularly interesting, as the model suggests that the pathway for glycolaldehyde assimilation involves condensation with glycine and the reverse pyridoxal-5-phosphate biosynthetic pathway, ultimately resulting in pentose phosphate rearrangement to produce glyceraldehyde-3-phosphate (shown in FIG. 8b). Based on the predicted flux distribution, this pathway appears to be favored over the assimilation of acetyl-CoA via the glyoxylate bypass. As shown in FIG. 8c, the direct production of glyceraldehyde from the HACL-based pathway results in the conversion of glyceraldehyde to glycerol, followed by native glycerol metabolism. Consequently, pathways that generate C3 molecules such as glyceraldehyde or dihydroxyacetone can utilize glycolytic reactions that result in the net production of ATP, ultimately allowing for greater biomass yields. Overall, in the three scenarios discussed above (and shown in FIG. 8 and Table 1), the FORCE pathway allows for the conversion of unnatural C1 substrates to native multicarbon substrates, as illustrated in FIG. 1c. Table 4 shows that the FORCE pathway also has promising characteristics based on other metrics such as redox balance, ATP demand, and the number of reactions required. [Table 3]

[0059] A two-strain co-culture system for assessing synthetic methylotrophy The orthogonality of the FORCE pathway to E. coli metabolism also allows for the full separation of the C1 conversion pathway from growth, and therefore a unique design to evaluate the pathway's methylotrophy capacity. One potentially advantageous implementation could use division of labor by separating multicarbon compound production and cell growth into two hosts, which is not possible when the pathway is directly interfaced with central metabolism, for example, via aldose phosphate or acetyl-CoA, two common products of the C1 anabolic pathway. Modularizing the system in this way allows for easier analysis of potential limitations. Using this idea, the FORCE pathway was evaluated for its ability to support E. coli growth on C1 substrates (e.g., formaldehyde, formate, and methanol).

[0060] As shown in Figure 9a, a two-lineage system consisting of two engineered strains of E. coli was designed, constructed and conceived to work in co-culture. The first strain, designated producer strain, contained constructs for expression of the FORCE pathway for conversion of the non-natural C1 substrate to the natural C2 growth substrate glycolate, but lacked the ability to consume and grow on glycolate. The second strain, designated sensor strain, retained the ability to grow on glycolate and additionally expressed eGFP constitutively as a signal, but did not express the FORCE pathway for glycolate production. Thus, the producer and sensor strains could be distinguished both by selection on glycolate minimal medium plates and by detection of fluorescent colonies. To assess the viability of the different substrates, three different producer strains were devised. The producer strain for formaldehyde utilization expressed LmACR, BsmHACL and EcAldA. The producer strain for evaluating formate utilization with formaldehyde expressed BsmHACL together with CaAbfT. Finally, the producer strain for methanol utilization was a thioesterase-deficient background expressing BmMdhMGA3, LmACR, BsmHACL and CaAbfT (shown in FIG. 9a).

[0061] As shown in Figure 15a, formaldehyde was the first substrate tested, but paraformaldehyde was used to allow for growth conditions. Paraformaldehyde gradually depolymerizes to produce formaldehyde in aqueous media with the ability to control solubilization rates by selection of particle size and concentration (Figure 15). Now referring to Figure 15b, this in turn allows for a system where formaldehyde can be maintained at sub-millimolar concentrations to avoid accumulation to toxic levels, with significant glycolate production still observed in Figure 15. In minimal media with (para)formaldehyde (equivalent to 5 mM) as the sole carbon substrate, growth of the sensor strain was observed as indicated by an increase in colony forming units (CFU) compared to a control system where the producer strain did not express HACL, as shown in Figure 9b and Figure 16. Glycolate accumulated rapidly in the first 8 hours, with sustained exponential growth of the sensor strain occurring after an initial lag phase. The sensor strain was found to have undergone approximately 6.6 doublings in 30 hours.

[0062] Methanol was also evaluated as a substrate for the two-lineage system. Figure 9c and Figure 17 show that growth of the sensor strain was observed only when the producer strain expressed BsmHACL. However, in Figure 17, the growth rate of the sensor strain was different, reflecting an almost linear increase in CFU over time, compared to the case of paraformaldehyde utilization. The observed difference in kinetics could reflect a limitation imposed by the rate of glycolate production from methanol by the producer strain, similar to the phenomenon observed in constant feed-rate fed-batch cultures56. Methanol utilization was substantially slower than (para)formaldehyde utilization, resulting in approximately 4.6 doublings in 72 hours.

[0063] Similar experiments were performed using a 1 mM formaldehyde and 10 mM formate co-substrate system tested using resting cells. As shown in Figure 18 and observed above, more carbon was observed in glycolate than added as formaldehyde, indicating incorporation of formate, Figure 18. Figure 9c shows that growth of the sensor strain was faster than growth on methanol, but did not result in as many doublings as 5 mM (para)formaldehyde. At 27 hours, approximately 4.9 doublings were observed.

[0064] Consideration Although product synthesis from C1 substrates is a limiting feature of the FORCE pathway, the pathway also has the ability to allow growth on non-natural C1 substrates (e.g., synthetic methylotrophy) through the production of multicarbon compounds that are naturally consumed by heterotrophic organisms, such as glycolate, acetate, or glyceraldehyde. To this end, the efficiency of the FORCE pathway to achieve synthetic methylotrophy was evaluated by genome-scale modeling and flux balance analysis. This analysis revealed that the FORCE pathway is comparable or better than alternative approaches. Although the current pathway performance could not support the growth of a single strain of E. coli on C1 substrates, the orthogonal nature of the pathway allowed for growth, isolation, and evaluation of pathway limitations for growth on formate, formaldehyde, and methanol in another strain of E. coli. Producer strains needed to be added in excess, and cell-specific improvements in pathway efficiency were shown to allow for the strengthening of the FORCE pathway by growth on a single chassis. The ability of the FORCE pathway to enable methylotrophy allows the implementation of bioprocesses more similar to traditional fermentations based on C1 as the sole carbon source. In these approaches, substrates are used both for product synthesis and for biocatalyst generation and maintenance.

[0065] Since the FORCE pathway is a branching point of flux towards product synthesis and growth, there is significant potential for easy control over flux partitioning as shown in Figure 10. Good control over these fluxes can be crucial to achieve high yield bioconversion from C1, especially when carbon and energy are limiting, e.g., in the case of formate as the sole substrate.

[0066] Further development of the FORCE pathway should enable more efficient design for synthetic methylotrophy and more diverse product synthesis, especially through pathway iteration. As an example, the major bottleneck, the acyloin condensation reaction of formyl-CoA, was evaluated with aldehydes catalyzed by HACL. The observation of formate as a by-product throughout various runs using reduced substrates formaldehyde and methanol is likely due to an imbalance between the rate of making formyl-CoA and its utilization by HACL. Formyl-CoA hydrolysis was also observed, which may be exacerbated in vivo by the presence of endogenous thioesterases. One exemplary approach to address this limitation is to use a CoA-transferase to reactivate formate to formyl-CoA, as we did using the CoA-transferase CaAbfT. Identification or engineering of better characterized HACL enzymes should help address this limitation. One specific example of this approach is the identification of BsmHACL described herein. Other examples include host strain modifications such as deletion of endogenous aldehyde dehydrogenases, and thioesterases have been examined.

[0067] Because the HACL-catalyzed condensation reaction and enzymatic activity have only recently been described, further genome mining, bioprospecting, enzyme engineering and biochemical characterization are expected to result in the discovery of better-functioning variants, ultimately overcoming pathway bottlenecks. HACL variants with well-defined chain length and functional group specificity, in combination with compatible and specific termination enzymes, should also enable the generation of specific products similar to those described by other platform pathways.

[0068] method: The methods outlined below describe the procedures and materials used to generate the specific test examples disclosed herein.

[0069] thermodynamic calculations Standard Gibbs free energies of reactions were found either from database sources (MetaCyc) or using the eQuilibrator biochemical thermodynamics calculator. The minimum-maximum driving forces of the pathways were calculated using previously reported methods implemented using MATLAB (Mathworks). The scripts used to perform the analysis are provided in the supplementary files.

[0070] Flux balance analysis Flux balance analysis was performed using COBRA Toolbox66 for MATLAB (Mathworks) with Gurobi solver (Gurobi Optimization, LLC). Reactions enabling various methylotrophic pathways (outlined in Table 3) were added to or modified into the E. coli genome-scale model iML151552. Constraints for substrate exchange reactions were set at 10 mmol C / g DCW / hr for all C1 substrates. The scripts used to perform the analysis are provided in the supplementary files.

[0071] reagent All chemicals were obtained from Fisher Scientific Co. and Sigma-Aldrich Co. unless otherwise specified. Primers were synthesized by Integrated DNA Technologies or by Eurofins Genomics. Restriction enzymes were obtained from New England Biolabs unless otherwise specified.

[0072] Genetic methods Plasmids and strains were constructed according to methods previously described. Genes not native to E. coli were codon optimized and synthesized by GeneArt (Thermo Fisher). E. coli genes were amplified from chromosomal DNA according to standard methods. Plasmids and strains used in this study are listed in Table 2. [Table 4]

[0073] Assessment of core pathway modules using purified enzymes RuHACLG390N, LmACR and OfFrc were expressed and purified as previously described. To test utilization of formaldehyde as the sole C1 substrate, reactions consisted of 50 mM KPi pH 7.4, 5 mM MgCl2, 0.1 mM TPP, 1 mM NAD+, 2 mM CoASH, 1 μM RuHACLG390N, 1 μM LmACR and 100 mM FALD. To test utilization of formate and formaldehyde as co-substrates, reactions consisted of 50 mM KPi pH 7.4, 5 mM MgCl2, 0.1 mM TPP, 1 mM succinyl-CoA, 1 μM RuHACLG390N, 2 μM OfFrc, 100 mM sodium formate and 100 mM formaldehyde. To test formate utilization as the sole C1 substrate, reactions were composed of 50 mM KPi pH 7.4, 5 mM MgCl2, 0.1 mM TPP, 1 mM NADH, 2 mM succinyl-CoA, 1 μM RuHACLG390N, 2 μM OfFrc, 1 μM LmACR, and 100 mM sodium formate. As a control, reactions were composed of 50 mM KPi pH 7.4, 5 mM MgCl2, 0.1 mM TPP, 1 mM NADH, 1 mM NAD+, 2 mM succinyl-CoA, 2 mM CoASH, 2 μM BSA, 100 mM sodium formate, and 100 mM formaldehyde. The reaction volume was 200 μL, and reactions were carried out at room temperature on a rotisserie shaker for 30 minutes. GC-MS analysis of the free acid was performed as previously described after treating 200 μL of the reaction sample with 5 μL 10 M NaOH.

[0074] To analyze acyl-CoAs by LC-MS, reactions were stopped by adding 8 μL of formic acid to 200 μL of reaction sample, primed twice with 200 μL methanol, and desalted on 1 mL HyperSep C18 Cartridges (Thermo Scientific) equilibrated with 100 μL of 1 mM ammonium acetate pH 3.0. The column was washed once with 200 μL of 1 mM ammonium acetate pH 3.0, and acyl-CoAs were eluted in 200 μL methanol. LC-MS analysis was performed based on that described previously. An Agilent 6540 Q-TOF LC-MS system was equipped with a jet stream electrospray ionization source set in positive ionization mode and a 100 mm x 4.6 mm Kinetex 2.6 μm Polar C18 100 Å column (Phenomenex). LC conditions were: column oven set at 40 °C, injection volume of 5 μL and 50 mM ammonium formate and methanol as the mobile phase. Compound separation was achieved using the following gradient method: 0 min, 0% methanol; 1 min, 0% methanol; 3 min, 2.5% methanol; 9 min, 23% methanol; 14 min, 80% methanol; 16 min, 80% methanol; 17 min, 0% methanol, at a flow rate of 400 μL / min. MS conditions were: capillary voltage 3.5 kV, nozzle voltage 500 V, fragmentor voltage 150 V, and nitrogen was used for nebulization (25 psig), drying (5 L / min, 225 °C) and sheath gas (10 L / min, 400 °C). A scan range of 100-1000 m / z was used. Data was analyzed using MassHunter Qualitative Analysis B.05.00 (Agilent).

[0075] Cell-free metabolic engineering for pathway validation Enzyme expression and cell extract preparation were performed as previously described. Cell-free reactions contained 50 mM KPi pH 7.4, 4 mM MgCl2, 0.1 mM TPP, 2.5 mM CoASH, 5 mM NAD+, 50 mM formaldehyde and 0.1 mM coenzyme B12. Individual cell extract loading was approximately 4.4 g / L protein (1 / 8 of reaction volume) and the amount of protein added to each reaction was normalized to approximately 26 g / L protein (3 / 4 of reaction volume) with BL21(DE3) extract. Reactions were incubated at room temperature for the indicated times, at which point reactions were stopped by adding 1 / 4 of the reaction volume of saturated ammonium sulfate solution acidified with 1% sulfuric acid. Samples were centrifuged at 20817xg for 15 min and the supernatants were analyzed by HPLC as previously described.

[0076] Resting cell biotransformation Bioconversions using resting cells were performed as previously described with minor modifications. The basal salts medium used was M9 (6.78 g / L Na2HPO4, 3 g / L KH2PO4, 1 g / L NH4Cl, 0.5 g / L NaCl, 2 mM MgSO4, 100 μM CaCl2 and 15 μM thiamine-HCl) further supplemented with Neidhardt's micronutrient solution. Overnight LB cultures of the respective strains were used to inoculate 250 mL flasks containing 50 mL of the above-mentioned medium further supplemented with 20 g / L glycerol, 10 g / L tryptone, 5 g / L yeast extract and the appropriate antibiotics (50 μg / mL carbenicillin, 50 μg / mL spectinomycin) (1%). Flask cultures were incubated at 30 °C and 250 rpm in an NBS I24 benchtop incubator shaker (New Brunswick Scientific Co.). After 2.5 h, gene expression was induced by the addition of 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and 0.04 mM cumate (0.2 mM IPTG and 0.1 mM cumulate were used for experiments with formaldehyde and formate).

[0077] Cells from the above cultures were harvested by centrifugation (5000xg, 22°C, 5 min) and washed twice with the above M9 medium without carbon source. The final cell pellet was resuspended in M9 with the appropriate carbon source (OD600 of approximately 10 with 10 mM formaldehyde or OD600 of approximately 5 with 1 mM formaldehyde and 10 mM formate). 5 mL of the cell suspension was added to a 25 mL Erlenmeyer flask (Corning Inc.) and topped with a foam plug. The flask was incubated at 30°C, 200 rpm in an NBS I24 benchtop incubator shaker (New Brunswick Scientific Co.). When formaldehyde was the only carbon source, an additional 10 mM formaldehyde was added after 1.5 hours. After 24 hours, samples were obtained for HPLC analysis as previously described. When 13C-labeled formaldehyde was used as the substrate, samples were analyzed by GC-MS after extraction and derivatization as previously described.

[0078] Fermentation experiment The growth medium used was M9 (6.78 g / L Na2HPO4, 3 g / L KH2PO4, 1 g / L NH4Cl, 0.5 g / L NaCl, 2 mM MgSO4, 100 μM CaCl2 and 15 μM thiamine-HCl) further supplemented with 500 mM methanol, 10 g / L tryptone, 5 g / L yeast extract and Neidhardt's micronutrient solution. Overnight LB cultures of each strain were used to inoculate 50 mL closed-cap conical tubes (Genesee Scientific Co.) containing 5 mL of the above-mentioned medium further supplemented with the appropriate antibiotics (50 μg / mL carbenicillin, 50 μg / mL spectinomycin) (1%). After approximately 3 h, gene expression was induced by the addition of 0.04 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and 0.04 mM cumate. The tubes were incubated at 30°C, 200 rpm in an NBS I24 benchtop incubator shaker (New Brunswick Scientific Co.). Samples (100 μL) were taken every 24, 48, 72 and 96 hours after inoculation for OD600 measurements and HPLC analysis as previously described. When 13C methanol was used as substrate, samples were analyzed by GC-MS after extraction and induction as previously described.

[0079] A two-strain Escherichia coli system for growth using formaldehyde as the sole carbon source. A two-line experiment was performed using strains cultured and induced as previously described using M9 medium. The induced cells were resuspended to an initial concentration of 3*109 CFU (colony forming units) / mL (equivalent to an OD600 of approximately 5) in M9 medium. 20 mL of the suspension was added to a 25 mL flask containing 3 mg paraformaldehyde (equivalent to 5 mM) or 10 mL of the suspension was added to a 25 mL flask supplemented with 500 mM methanol or 1 mM formaldehyde and 10 mM sodium formate. A second E. coli strain capable of consuming glycolate, AC763, was cultured at 5*10 6AC763 was added to an initial concentration of 100 CFU / mL (equivalent to an OD600 of approximately 0.005). AC763 additionally had a chromosomal copy of eGFP constitutively expressed to aid in distinguishing between the two strains. Prior to its addition to the culture, AC763 was pre-grown in 5 mL of the above-mentioned M9 minimal medium supplemented with 5 g / L glycolate and 2 g / L tryptone in a 25 mL Erlenmeyer flask (from a single colony inoculum) for 24 h at 200 rpm and 30°C. The cells were then centrifuged (5000xg, 22°C, 5 min), washed twice with medium supplemented with 5 g / L glycolate, and resuspended to an optical density of approximately 0.05. After 24 h of incubation at 200 rpm and 30°C (5 mL in a 25 mL Erlenmeyer flask), the cells were centrifuged (5000xg, 22°C), washed twice with medium without carbon source, and the appropriate volume was added to the two-lineage system. Flasks containing both strains were further incubated at 30°C at 200 rpm. Samples were taken at various times for HPLC and cell growth analysis. Colony forming units per mL of culture were used as a measure of cell growth. Appropriate volumes of cultures were diluted into the minimal medium described above without a carbon source, and 50 μL of various dilutions were plated onto minimal medium plates containing 2.5 g / L glycolate. After plate incubation at 37°C, colonies were counted manually, aided by visualization using a blue light transilluminator (Vernier, Beaverton, OR) illuminating the eGFP-expressing strain AC763.

[0080] As previously noted, while the present disclosure has been described above with respect to particular embodiments and examples, it will be understood by those skilled in the art that the disclosure is not necessarily so limited, and that numerous other embodiments, examples, uses, modifications, and departures from the embodiments, examples, and uses are intended to be encompassed within the scope of the claims appended hereto. EXAMPLES

[0081] Working Example Example 1: Strategies used to identify enzymes with similar structure and / or function based on sequence similarity The purpose of this example is to provide an overview of the workflow used to identify enzyme variants with desired activity, starting from a reference enzyme as a query. In this example, 2-hydroxyacyl-CoA lyase, HACL (RuHACL) from Rhodospirillales bacterium URHD0017, is used as the starting query for the identification of first-round 2-hydroxyacyl-CoA synthase (HACS) variants. Protein BLAST (pBLAST) is used with an E-value cutoff based on the E-value between RuHACL and oxalyl-CoA decarboxylase, OXC from Escherichia coli (EcOXC) and Oxalobacter formigenes (OfOXC) (Figure 19). To down select representative variants by clustering the query results into families with similar sequences, we used the CD-HIT web server (Huang et al. Bioinformatics, (2010). 26:680). A more lenient restriction of 70% identity threshold was imposed on genes from prokaryotic origin, whereas 50% was used for no taxonomic restriction (Figure 19). Clustering and screening representative genes using CD-HIT yielded 93 HACS variants similar to RuHACL. Further curation from the list included elimination of too long or too short sequences and variants from the animal kingdom that are unlikely to be well expressed in E. coli. Consequently, we determined that after curation, 34 remaining variants were first-round HACS variants for synthesis and testing in E. coli as a host (Figure 19). The selected genes were then codon-optimized for expression in E. coli and synthesized in collaboration with the Joint Genome Institute (JGI). Five variants failed during synthesis, yielding 29 first-round JGI HACS variants (JGI1-JGI29) (Table 5).

[0082] Table 5. List of 2-hydroxyacyl-CoA (HACS) variants (JGI) identified by selecting representative genes from gene clusters with sequence similarity using RuHACL as the reference enzyme. [Table 5]

[0083] Example 2: Establishment of a high-throughput platform to screen first-round HACS variants for C1-C1 condensation The purpose of this example is to demonstrate a high-throughput platform for screening 2-hydroxyacyl-CoA synthase (HACS) variants in vivo. We used glycolic acid (glycolate) productivity per cell density (μM glycolate / OD600) as an indicator of HACS activity. Glycolate can be made from formaldehyde as the sole carbon source in the presence of active HACS variants and acyl-CoA reductase from Listeria monocytogenes (LmACR) (Figure 20A). It is shown that LmACR can catalyze the oxidation reaction of formaldehyde to formyl-CoA (Chou, A., et al. Nat. Chem. Biol. 15:900-906 (2019)). HACS condenses formaldehyde and formyl-CoA to form glycolyl-CoA, which can then be hydrolyzed to glycolate by the native thioesterase activity (Figure 20A).

[0084] To prototype a pathway for glycolate production from formaldehyde in vivo, we constructed vectors overexpressing various HACS candidates and LmACR, both under the control of an IPTG-inducible T7 promoter in pCDFDuet-1 and pETDuet-1, respectively (Figure 20C). As hosts for these vectors, we used genetically engineered strains of E. coli based on MG1655(DE3) with knockouts for formaldehyde (ΔfrmA) and formate (ΔfdhF ΔfdnG ΔfdoG) oxidation, as well as for glycolate utilization (ΔglcD), which we predicted could compete or interfere with our pathway analysis.

[0085] In vivo product synthesis was performed using M9 minimal medium (6.78 g / L Na2HPO4, 3 g / L KH2PO4, 1 g / L NH4Cl, 0.5 g / L NaCl, 2 mM MgSO4, 100 μM CaCl2 and 15 μM thiamine-HCl) unless otherwise specified. Cells were initially grown in 96-deep-well plates (USA Scientific, Ocala, FL) containing 0.2 mL of the above-mentioned medium further supplemented with 20 g / L glycerol, 10 g / L tryptone and 5 g / L yeast extract. A single colony of the desired strain was cultured overnight (14-16 h) in LB medium with the appropriate antibiotic and used as inoculum (1%). Antibiotics (100 μg / mL carbenicillin, 100 μg / mL spectinomycin) were included where appropriate. Cultures were then incubated at 30° C. at 1000 rpm in a Digital Microplate Shaker (Fisher Scientific) until an OD600 of approximately 0.4 was reached, at which point appropriate amounts of inducer(s) (isopropyl β-D-1-thiogalactopyranoside (IPTG)) were added. Plates were incubated for a total of 24 hours after inoculation (FIG. 20B).

[0086] Cells from the precultures described above were then centrifuged (4000 rpm, 22° C.), washed with the above-mentioned minimal medium without carbon source, and resuspended in 1 mL of the above-mentioned minimal medium with the indicated amount of carbon source. 5 mM formaldehyde was added at time 0 and incubated at 30° C., 1000 rpm in a Digital Microplate Shaker (Fisher Scientific). After 3 hours of incubation at 30° C., cells were pelleted by centrifugation and the supernatant was analyzed by HPLC or GC-MS as described below. Cell pellets harvested after bioconversion were resuspended to 20 OD in B-PER® Bacterial Protein Extraction Reagent (Thermo Fisher) supplemented with 0.1 mg / mL hen egg white lysozyme (Fisher) and 5 U / mL Benzonase® nuclease (Sigma) for cell lysis. After 15 min incubation at room temperature, 100 μL of each cell lysate was transferred to a 1.5 mL microcentrifuge tube for centrifugation at 15,000×g for 5 min. The soluble cell lysates obtained from the supernatants were analyzed using SDS-PAGE. Relative HACS expression was estimated by band area in the protein gel images.

[0087] Quantification of product and substrate concentrations (formic acid, formaldehyde, and glycolic acid) was determined by HPLC using a Shimadzu Prominence SIL 20 system (Shimadzu Scientific Instruments, Inc., Columbia, MD) equipped with a refractive index detector and an HPX-87H organic acid column (Bio-Rad, Hercules, CA) under operating conditions that optimized peak separation (0.3 ml / min flow rate, 30 mM H2SO4 mobile phase, column temperature 42°C). Compound identification and analysis were performed by GC-MS using an Agilent 7890B Series Custom Gas Chromatography system equipped with a 5977B Inert Plus Mass Selective Detector Turbo EI Bundle (for identification) and an Agilent HP-5-ms capillary column (0.25 mm internal diameter, 0.25 μm film thickness, 30 m length).

[0088] First-round HACS variant screening shows that three variants out of 29 (JGI15, 19, 20) show better glycolate productivity and relative HACS expression than the starting reference RuHACL (Figure 20C). JGI15 and JGI20 are the two best candidates showing a greater than three-fold improvement in glycolate productivity and are selected for further analysis.

[0089] Example 3: Testing of high performance variants under different C1-C1 fusion platforms The purpose of this example is to show the analysis for two high-performing HACS variants (JGI15 and JGI20) in comparison with the reference enzyme RuHACL. We used glycolic acid (glycolate) productivity per cell density (μM glycolate / OD600) as an indicator of HACS activity. Two different enzymatic pathways for glycolate synthesis were investigated. The first pathway (pathway 1) is similar to the pathway originally used for screening in example 2 by the addition of an extra gene, aldehyde dehydrogenase aldA from E. coli (EcAldA), which is overexpressed to drive the flux from glycolaldehyde to glycolate (Figure 21A). Also, HACS and LmACR-AldA are controlled under independent inducible promoters to examine the effect of varying the relative gene expression (Figure 21B). The second pathway (pathway 2) involves independent fluxes of formaldehyde and formyl-CoA, allowing for the assessment of enzyme activity in response to changes in formaldehyde concentration alone, while keeping the formyl-CoA flux constant. Formyl-CoA is made from formic acid (formate) catalyzed by acyl-CoA transferase (CaAbfT) from Clostridium aminobutyricum (Figure 22A).

[0090] For in vivo prototyping, we engineered vectors that independently controlled the expression of HACS variants and LmACR-EcAldA (pathway 1, FIG. 21A) or CaAbfT (pathway 2, FIG. 22A), where HACS is under the control of an IPTG-inducible T7 promoter in pCDFDuet-1, and LmACR-EcAldA or CaAbfT is under the control of a cumate-inducible T5 promoter in pETDuet-1 (FIG. 21B). As hosts for these vectors, we used engineered strains of E. coli based on MG1655(DE3) with knockouts for formaldehyde (ΔfrmA) and formate (ΔfdhF ΔfdnG ΔfdoG) oxidation, as well as for glycolate utilization (ΔglcD), which we predicted might compete or interfere with our pathway analysis.

[0091] In vivo product synthesis was performed using M9 minimal medium (6.78 g / L Na2HPO4, 3 g / L KH2PO4, 1 g / L NH4Cl, 0.5 g / L NaCl, 2 mM MgSO4, 100 μM CaCl2 and 15 μM thiamine-HCl) unless otherwise specified. Cells were initially grown in 96-deep-well plates (USA Scientific, Ocala, FL) containing 0.2 mL of the above-mentioned medium further supplemented with 20 g / L glycerol, 10 g / L tryptone and 5 g / L yeast extract. A single colony of the desired strain was cultured overnight (14-16 h) in LB medium with the appropriate antibiotic and used as inoculum (1%). Antibiotics (100 μg / mL carbenicillin, 100 μg / mL spectinomycin) were included when appropriate. Cultures were then incubated at 30° C., 1000 rpm in a Digital Microplate Shaker (Fisher Scientific) until an OD600 of approximately 0.4 was reached, at which point appropriate amounts of inducer(s) (isopropyl β-D-1-thiogalactopyranoside (IPTG) and cumate) were added. Plates were incubated for a total of 24 hours after inoculation (FIG. 20B).

[0092] Cells from the above precultures were then centrifuged (4000 rpm, 22° C.), washed with the above minimal medium without carbon source, and resuspended in 1 mL of the above minimal medium with the indicated amount of carbon source. 5 mM formaldehyde alone for LmACR-EcAldA co-expression (FIG. 21A), or 0.5 / 5 mM formaldehyde and 20 mM formate for CaAbfT co-expression (FIG. 22A) were added at 0 h and incubated at 30° C., 1000 rpm in a Digital Microplate Shaker (Fisher Scientific). Cells were harvested by centrifugation after 3 h for LmACR-EcAldA co-expression and 1 h for CaAbfT co-expression, and the supernatants were analyzed by HPLC or GC-MS as described in Example 2.

[0093] When 5 mM formaldehyde is used as the sole carbon source, JGI15 (Figure 21C) performs 2.5-fold better than RuHACL (Figure 21D) based on glycolate productivity (μM / OD600) at 3 h under optimal inducer concentration (relative gene expression). When formaldehyde and formate are co-fed, JGI15 outperforms RuHACL and JGI20 at a practical limit (7-fold and 1.5-fold, respectively) under low formaldehyde availability (0.5 mM), reflecting the good affinity of JGI15 with formaldehyde (low K m ) (Figure 4B). On the other hand, JGI20 showed good glycolate productivity under high formaldehyde concentration (5 mM), indicating good metabolic turnover (high k cat ) is shown.

[0094] Example 4: Kinetic characterization of high-performing HACS variants The purpose of this example is to show the kinetic characterization of the high-performing HACS variants (JGI15, JGI20, JGI23 and JGI24) derived from the first round of homologs using in vitro kinetic assays with purified enzymes. Kinetic assays were performed using a coupled reaction providing formyl-CoA from formate catalyzed by CoA transferase CaAbfT using acetyl-CoA as the CoA donor.

[0095] Expression of selected enzyme variants was achieved using plasmid-based gene expression constructed either by the Joint Genome Institute (JGI) for HACS variants (JGI15, 20, 23 and 24) or by cloning the desired gene(s) into pCDFDuet-1 (Novagen, Darmstadt, Germany) digested with appropriate restriction enzymes and utilizing In-Fusion cloning technology (Clontech Laboratories, Inc., Mountain View, CA). Linear DNA fragments for insertion were generated by gene synthesis of codon-optimized genes. Genes were synthesized by GeneArt (Life Technologies, Carlsbad, CA) or Twist (Twist Biosciences). The resulting In-Fusion reaction products were used to transform E. coli Stellar cells (Clontech Laboratories, Inc., Mountain View, CA) and clones identified by PCR screening were further confirmed by DNA sequencing.

[0096] Overnight cultures of expression strains were grown in LB and used to inoculate 25 mL TB medium in 250 mL baffled flasks at 1% v / v (250 μL). Cultures were grown at 30 °C and 250 rpm in a rotary shaker until the OD550 reached 0.4-0.6, at which point expression was induced with 0.1 mM IPTG. 24 h after inoculation, cells were harvested by centrifugation. Cell pellets were washed once with cold 9 g / L NaCl solution and stored at -80 °C until required. Antibiotics were included where appropriate at the following concentrations: carbenicillin (50 μg / mL) and spectinomycin (50 μg / mL).

[0097] For protein purification, E. coli cell pellets expressing the desired his-tagged enzymes were prepared as described above. Frozen cell pellets were resuspended in cold lysis buffer (50 mM NaPi pH 7.4, 300 mM NaCl, 10 mM imidazole, 0.1% Triton-X 100) to an OD550 of approximately 40, to which 1 mg / mL lysozyme and 250 U of Benzonase nuclease were added. The mixture was further processed by sonication on ice using a Branson Sonifier 250 (5 min at 25% duty cycle and output control set at 3) and centrifuged at 7500×g for 15 min at 4° C. The supernatant was applied to a chromatography column containing 1 mL TALON metal affinity resin (Clontech Laboratories, Inc., Mountain View, CA) and pre-equilibrated with lysis buffer. The column was then washed first with 10 mL of lysis buffer, then twice with 20 mL of wash buffer (50 mM NaPi pH 7.4, 300 mM NaCl, 20 mM imidazole). The his-tagged protein of interest was eluted by 1-2 applications of 4 mL elution buffer (50 mM NaPi pH 7.4, 300 mM NaCl, 250 mM imidazole). The eluate was collected and applied to a 10,000 MWCO Amicon ultrafiltration centrifugation device (Millipore, Billerica, MA) and the concentrate (approximately 100 μL) was washed twice with 4 mL of 50 mM KPi pH 7.4 for desalting. Protein concentration was estimated by the Bradford method. The purified protein was stored at -80 °C in 20 μL aliquots until needed.

[0098] SDS-PAGE was performed using NuPAGE 12% Bis-Tris Protein Gels with SDS running buffer and stained with SimplyBlue SafeStain (ThermoFisher Scientific, Waltham, MA) according to the manufacturer's protocol.

[0099] The in vitro kinetic assay consisted of 100 mM KPi pH 6.9, 10 mM MgCl2, 0.15 mM TPP, 2 mM acetyl-CoA, 1 μM CaAbfT, 0.25 μM HACS variant and 20 mM sodium formate. The reaction was incubated at room temperature for 3 min to convert formate to formyl-CoA, and then a specific concentration of aldehyde (specifically here acetaldehyde or propionaldehyde) was added to the reaction. After another 3 min of incubation, the reaction was terminated by adding 1 / 20 of the reaction volume of 10 M NaOH solution. After 30 min of hydrolysis, the pH was neutralized by adding 1 / 20 of the reaction volume of 10 N H2SO4. The samples were centrifuged at 20817xg for 15 min and the supernatant was analyzed by GC-MS as described below.

[0100] For this analysis, 0.15 reaction volumes of the internal standard methyl succinate were added to the samples. The resulting samples were extracted into 4 mL ethyl acetate by vigorous vortexing for 20 min. The organic phase was separated and evaporated to dryness under a stream of nitrogen. The residue was dissolved in 30 μL pyridine and 30 μL N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and incubated at 60 °C for 15 min. Compounds were identified and analyzed by GC-MS using an Agilent 7890B Series Custom Gas Chromatography system equipped with a 5977B Inert Plus Mass Selective Detector Turbo EI Bundle (for identification) and an Agilent HP-5-ms capillary column (0.25 mm internal diameter, 0.25 μm film thickness, 30 m length). Samples were analyzed by GC (1 μL injection, 20:1 separation ratio) using helium as the carrier gas at a flow rate of 1.5 mL / min and the following temperature profile: initial 90° C., 3 min; ramp 15° C. / min to 170° C.; ramp 20° C. / min to 300° C. and hold for 8 min. Injector and detector temperatures were 250° C. and 350° C., respectively.

[0101] As can be inferred from Example 3 (FIG. 22B), JGI15 has a lower k with respect to formaldehyde compared to JGI20. cat Low K m With longer chain aldehydes, JGI15 and JGI20 had lower K for acetaldehyde and propionaldehyde compared to JGI23 and JGI24. m (millimolar or less) and has a lower K cat This indicates that they have stronger affinity for aldehydes and lower enzymatic activity. JGI23 and JGI24 have better activity (higher K cat ) and by improving the affinity for the substrate (lower K m ), have a better ability to work better in vivo. Similarly, AcHACL has a lower Km with acetone and JGI15 has a better Kcat (Table 8).

[0102] Table 8. Apparent kinetic parameters for 2-hydroxyacyl-CoA synthase (HACS) variants using various aldehydes and ketones as substrates. [Table 6]

[0103] Example 5: Modeling the protein structure of high-performance HACS variants and understanding key catalytic residues through structural analysis This example shows the analysis of recombinant high-performance HACS variants (JGI15 and JGI20) using protein structure analysis and alanine scanning methods. The complete dimer structures of JGI15 (Figure 23A) and JGI20 (Figure 23B) are modeled using AlphaFold (Jumper et al. Nature 596:583-589 (2021)) in the ColabFold platform (Mirdita et al. Nature Methods 19:679-682 (2022)). The model is aligned with the crystal structure of oxalyl-CoA decarboxylase from Oxalobacter formigenes in complex with formyl-CoA (PDB code: 2JI8) (Berthold et al. Structure 15: 853-861 (2007)) to understand the orientation of two important ligands, thiamine diphosphate (TPP) and formyl-CoA, in the active site. The structures are highly similar with root mean square deviation (RMSD) values ​​of 1.185 Å and 0.981 Å for JGI15 and JGI20, respectively. Both the active site (where the formyl residues of TPP and formyl-CoA interact for catalysis) and the CoA binding site are exposed to the solvent, indicating the correct orientation of the ligand docking into the JGI15 and JGI20 structures (FIG. 23C).

[0104] To understand the specific amino acid (AA) residues responsible for catalytic activity and substrate binding, we used JGI20 as a reference protein and selected all AA residues within 3.5 Å of either TPP (FIG. 24A) or formyl-CoA (FIG. 24B) (Table 7). We then selected AA residues that were not conserved among all first-round JGI variants (30 including RuHACL) and were unique to variants with C1-C1 condensation activity. Three AA residues from the TPP binding site (H80, Q113 and Y367 from JGI20) and six from the CoA binding site (F112, V354, M392, T397, Q544 and W548) were consequently selected (FIG. 24C). Q544 and W548 are located at the c-terminus of JGI20, which has been shown to form a closed loop covering the active site in other similar proteins, such as 2-hydroxyacyl-CoA lyase (AcHACL) from Actinomycetospora chiangmaiensis (Zahn et al. J. Biol. Chem. 298(1) 101522 (2022)). It was found that the active HACS variants have the conserved residues of "RKPQQF-W" in this region, but not the others (Figure 25A). Consequently, we decided to use an alanine scanning method (Morrison and Weiss, Curr Opin Chem Biol. 5 (3): 302-7 (2001)) to examine the importance of conserved residues among only those variants that are active near the active site and the c-terminal closing loop, and to identify key catalytic residues among them.

[0105] Table 7. Active site residues of high-performing variants for C1-C1 condensation (within 3.5 Å of TPP and formyl-CoA based on the AlphaFold structure of JGI20) and corresponding residues. Bold residues indicate non-conserved residues between active variants. Residues marked with an asterisk (*) indicate key catalytic residues predicted to distinguish HACS from OXC. [Table 7]

[0106] JGI15 and JGI20 mutants were prepared by cloning wild-type JGI15 and JGI20 into vector pUC19 (Clontech Laboratories, Inc., Mountain View, CA). Primers containing the desired mutations were designed according to the "in vivo assembly" (IVA) protocol for mutagenesis (Garcia-Nafria et al., Sci. Rep. 6, 12. 2016). PCR products containing the mutations were generated according to the IVA protocol and used to transform E. coli Stellar cells (Clontech Laboratories). The desired mutation sequences were confirmed by DNA sequencing. The mutant genes were then cloned into the final expression vector (pCDFDuet-1) using restriction enzyme digestion and ligation. The HACS activity of the mutants is tested in a format identical to Pathway 2 (Figure 22A) described in Example 3, using 0.5 mM formaldehyde and 20 mM formate as carbon sources.

[0107] Alanine scanning results for active site residues indicate that glutamine 113 (Q113) and tyrosine 367 (Y367) from TPP binding and phenylalanine 112 (F112) and methionine 392 (M392) from CoA binding are important residues for HACS activity for formaldehyde-formyl-CoA condensation (Figure 24C). Q113 has been shown to have important catalytic functions in other 2-hydroxyacyl-CoA synthases, such as AcHACL (Zahn et al. J. Biol. Chem. 298(1) 101522 (2022)). Mutagenesis of F112 and Q113 also showed abolition of HACS activity in previous tests for RuHACL (Chou, A., et al. Nat. Chem. Biol. 15:900-906 (2019)). Known oxalyl-CoA decarboxylase (OXC) variants, including genes from E. coli, O. formigenes and Methylorubrum extorquens, were also found to have a conserved tyrosine-glutamic acid (YE) residue at the F112 Q113 position. All variants from the first round HACS variants (JGI4, 6, 7, 9, 10, 11) that have an OXC-like "YE" residue do not have any glycolate production from formaldehyde and cluster together with OXC in the phylogenetic analysis (Figure 29). Therefore, "FQ" and "YE" could be important catalytic residues that distinguish the HACS and OXC forms of the enzyme. Interestingly, two further residues (Y367 and M392) (Figure 24C) that were found to be important for catalytic function are also conserved only among HACS but not OXC (Table 7). OXC-type enzymes have a non-conserved residue at the position corresponding to Y367 and a conserved leucine (L) residue at the position of M392, so these two residues could also play important roles in the catalytic function that distinguishes HACS from OXC activity.

[0108] Except for Q545A of JGI20, none of the c-terminal residues in JGI15 and JGI20 abolished activity from point mutagenesis to alanine (Figure 25B). According to the literature, it is expected that the c-terminus of HACS serves as a closed loop that stabilizes substrate binding without catalytic function (Zahn et al. J. Biol. Chem. 298(1) 101522 (2022)). Therefore, it was presumed that the c-terminus plays an important role in restricting the substrate size of the binding pocket. Based on the alanine scanning results for this region for both JGI15 and JGI20, we may find a trend toward decreased overall activity, especially under high formaldehyde concentration (5 mM), possibly due to decreased stability of the binding pocket. Interestingly, however, some mutations, such as P547A (JGI15), P543A (JGI20), Q549A (JGI15) and Q544A (JGI20), show a significant improvement in activity at low formaldehyde (0.5 mM) (Figure 25B), which could be a result of altering the substrate binding affinity with the least aldehyde, formaldehyde.

[0109] Example 6: Improving HACS activity by creating a hybrid protein of two high-performance variants This example shows the engineering of recombinant high-performance HACS variants (JGI15 and JGI20) by creating hybrid proteins based on structural analysis. Based on kinetic characterization (Table 8), JGI20 has a higher k cat However, it has a high K with formaldehyde. mThe inventors speculated that either the affinity of JGI20 or the turnover of JGI15 could be improved by creating a hybrid protein between the two. To identify the structural differences between the two proteins, the inventors used the "pairwise structural alignment" function in the Protein Data Bank (PDB) website (www.rcsb.org). The JGI15 and JGI20 structures modeled by AlphaFold were used for structural comparison, and the results show that there are two residues that are not aligned between the two protein structures (Figure 26A). The JGI15-20 hybrid protein was constructed by inserting or deleting AA residues to completely align the two structures. Consequently, JGI15 N465ins, R493del and N465 R493del are constructed to make JGI15 "JGI20-like", while JGI20 N461del, R480ins and N461del R480ins are constructed to make JGI20 "JGI15-like" (Figure 26B).

[0110] An alternative approach is to genetically engineer the active site of JGI20 to mimic JGI15, thereby decreasing the substrate binding affinity (K m ) was based on improving the K α -dependent cleavage potential of JGI20. Comparing the active site residues of the two enzymes (Table 11), the only non-conserved residues are A253 and P254, where JGI15 has two consecutive glycine residues at the corresponding positions. As a result, a "JGI15-like" JGI20 A253G P254G was constructed. Another target region was the c-terminus, where alanine scanning results indicate altered activity. JGI15 and 20 have highly conserved sequences in the c-terminal tail, except for the last 4-5 residues (Figure 27A). Based on the protein structures of JGI15 and 20 modeled by AlphaFold, the difference in the c-terminus indicates a slightly different orientation of the closed loop (Figure 27A). We believe that this difference may explain the K α -dependent cleavage potential between JGI15 and JGI20. m We hypothesized that this could contribute to the difference in C-terminus of JGI20, and constructed a “JGI15-like” C-terminus of JGI20: JGI20 L549H T550G R551del.

[0111] Table 11. List of acyl-CoA kinase (ACK) and phosphoacyltransferase (PTA) variants (JGIKs) identified by selecting representative genes from gene clusters with sequence similarity using CcAck and CcPta as reference enzymes. [Table 8-1] [Table 8-2]

[0112] Construction and testing of the mutants is carried out in an identical format to that described in Example 5.

[0113] The JGI15-20 hybrid based on the structural alignment showed a significant improvement of JGI15 at high formaldehyde and JGI20 at low formaldehyde, which interestingly shows a positive effect in both variants (Figure 26B). Aligning the two variants may affect the orientation of catalytic residues that specifically alters activity under high or low formaldehyde concentrations. Mutagenesis on the JGI20 active site and C-terminus to mimic JGI15 showed a more significant improvement in glycolate productivity under low formaldehyde concentrations of 39% and 61%, respectively (Figure 27B). We also attempted to combine beneficial mutations from the JGI15-20 structural hybrid and active site hybrids, and found that JGI20 R480 L549H T550G R551del, combining AlphaFold structural hybrid and c-terminal hybrid, showed the best improvement in activity by up to 70% with only a minimal decrease in activity at high formaldehyde concentrations (Figure 27B). m This is close to a 50% improvement from JGI15 at 0.5 mM FALD, showing a significant reduction in, which is contemplated from the hybrid approach.

[0114] Example 7: Identification, synthesis and screening of second round HACS variants for activity with formaldehyde This example shows the identification, synthesis and screening of second round HACS variants using formaldehyde as a substrate. From the first round variants, we found JGI15, JGI19 and JGI20 that were active for glycolyl-CoA synthase activity over the starting reference enzyme, RuHACL (Figure 28C). Also, the potential glycolyl-CoA synthase (C1-C1 condensation) activity of 2-hydroxyacyl-CoA lyase (AcHACL) from Actinomycetospora thiangumaiensis was shown in the literature (Rohwerder et al. Front. Microbiol. 11:691 (2020)). Based on the in vivo screening using formaldehyde as the sole carbon source (Figure 28A) (Example 3, Pathway 1), AcHACL showed better glycolate productivity than RuHACL under 5 mM formaldehyde. AcHACL is distantly related to RuHACL and other first-round JGI variants as seen from the phylogenetic tree (Figure 29). Therefore, we decided to use JGI15, JGI19, JGI20 and AcHACL from the first-round variants as reference enzymes to identify the second-round HACS variants (JGIH).

[0115] Table 6. List of 2-hydroxyacyl-CoA (HACS) variants (JGIHs) identified by selecting representative genes from gene clusters with sequence similarity using AcHACL, JGI15, JGI19 and JGI20 as reference enzymes. [Table 9-1] [Table 9-2] [Table 9-3] [Table 9-4]

[0116] For each of the starting reference enzymes, the method described in Example 1 (Figure 19) was used. A total of 99 enzymes closely related to AcHACL (AcHACL cluster), distantly related to AcHACL (AcHACL cluster), JGI19 cluster, JGI15 cluster and JGI20 cluster are identified (Figure 29). We also identified 9 extra enzymes structurally similar to AcHACL, JGI15, JGI19 or JGI20 from I-TASSER without considering sequence similarity (Yang et al. Nature Methods, 12: 7-8 (2015)). A total of 108 genes (JGIH1-JGIH108) are codon-optimized and synthesized in collaboration with the Joint Genome Institute, and 99 variants are successfully constructed into the pCDFDuet-1 expression vector for testing.

[0117] Second round variants are tested for glycolyl-CoA synthase activity using high throughput screening, co-fed with formaldehyde (0.5 mM) and formate along with the formate-activating enzyme (Figure 30A). Results show that six variants (JGIH25, 26, 30, 41, 61 and 65) perform better than wild type JGI15, with JGIH25 and 65 showing over a 50% increase in glycolate productivity. There are five additional candidates that perform at a similar level to JGI15 (Figure 30B).

[0118] Based on phylogenetic tree analysis, JGIH25, 26 and 30 belong to JGI20 cluster, 41 and 61 belong to JGI15 cluster, and JGI65 belongs to JGI19 cluster. Couple variants from AcHACL cluster (JGIH5 and 12) also show significant glycolate productivity, about 80% of JGI15. When comparing the residues of the six best variants aligned with the active site residues of JGI20 identified from Example 5, we can find that most of them are highly conserved, except for two residues (A253 P254 of JGI20) that were not conserved between JGI15 and JGI20 (Table 7). JGIH61 and 65 are phylogenetically the most distant from JGI15 and 20, so there are multiple non-conserved residues in the active site other than the two previously identified ones. After further characterization of the two variants, such as their affinity with formaldehyde and formyl-CoA and turnover rate, construction of novel hybrid proteins could be considered based on learnings from previous JGI15 and 20 hybrid approaches. The c-terminal residues of the six best variants are also well conserved, except that JGIH61 and 65 have two and three extra AA residues at the c-terminus. None of the six variants have the same c-terminal residues as JGI15, which could be another target for hybrid protein approach for closed loops.

[0119] Example 8: Screening of first and second round variants for activity with aldehydes The purpose of this example is to demonstrate a high-throughput platform for screening first and second round 2-hydroxyacyl-CoA synthase (HACS) variants using various aldehydes as substrates in vivo. We used 2-hydroxyacid productivity per cell density (μM / OD600) as an indicator of HACS activity. 2-hydroxyacids can be produced by co-feeding various aldehydes and formic acid (formate) as carbon sources in the presence of active HACS variants and acyl-CoA transferase from Clostridium aminobutyricum (CaAbfT). It is shown that CaAbfT can catalyze the reaction of formate to formyl-CoA (Nattermann, M., et al. ACS Catal 11(9):5396-5404 (2021)). HACS condenses aldehydes and formyl-CoA to form 2-hydroxyacyl-CoA, which can then be hydrolyzed to 2-hydroxyacids by the native thioesterase activity (Figure 31A).

[0120] For in vivo prototyping, we engineered vectors to independently control the expression of the HACS variants and CaAbfT, where HACS is under the control of an IPTG-inducible T7 promoter in pCDFDuet-1 and CaAbfT is under the control of a cumate-inducible T5 promoter in pETDuet-1 (Figure 31B), and transformed the vectors into the engineered strains of E. coli described in Example 3.

[0121] By co-feeding with 5 mM aldehyde and 20 mM formate, HACS variants were screened for 2-hydroxyacid production using a high-throughput screening platform as described in Example 3. Cells were harvested after 1 h by centrifugation and the supernatants were analyzed by HPLC (as described in Example 2) or the SoGO method.

[0122] In the SoGO method, glycolate oxidase from spinach (SoGO) is used to catalyze the oxidation of 2-hydroxyacids to produce 2-oxoacids and hydrogen peroxide (H2O2). Amplex UltraRed (Invitrogen) reagent is then used as a fluorescent substrate for horseradish peroxidase (HRP) (Sigma), which reacts with H2O2 in a 1:1 stoichiometric ratio to produce the brightly fluorescent and strongly absorbing reaction product Amplex UltroxRed (excitation / emission maxima approximately 568 / 581 nm). 2-hydroxyacid concentrations were calculated based on calibration of the fluorescence readings measured by Amplex UltroxRed using a BioTek Synergy HT plate reader (BioTek Instruments).

[0123] Example 9: Screening of first and second round HACS variants for activity with acetaldehyde This example shows the first and second round screening of HACS variants using acetaldehyde as a substrate in vivo. We used lactic acid (lactate) productivity per cell density (μM lactate / OD600) as an indicator of HACS activity. HACS variants were screened for lactate production using the high-throughput screening platform described in Example 3 by co-feeding 5 mM acetaldehyde and 20 mM formate. HACS condenses acetaldehyde and formyl-CoA to form lactoyl-CoA, which can then be hydrolyzed to lactate via native thioesterase activity (Figure 32A).

[0124] The first round screening of HACS variants shows that 6 variants out of 29 produce significant lactate productivity (Figure 32B). JGI15 and JGI20 are the two best candidates that show more than 2-fold higher lactate productivity compared to other HACS variants.

[0125] Quantification of product concentration (lactate) for the second round HACS variants was determined by the SoGO method described in Example 8. The results show that one variant, JGIH48, performs better than wild type JGI15 with over a 20% increase in lactate productivity (Figure 32C). Furthermore, JGIH28 performs at a similar level to JGI15.

[0126] Example 10: Screening of first and second round HACS variants for activity with propionaldehyde This example shows the first and second round screening of HACS variants with propionaldehyde as a substrate in vivo. We used 2-hydroxybutyrate (2HB) productivity per cell density (μM 2HB / OD600) as an indicator of HACS activity. HACS variants were screened for 2HB production using the high-throughput screening platform described in Example 3 by co-feeding with 5 mM propionaldehyde and 20 mM formate. HACS condenses propionaldehyde and formyl-CoA to form 2-hydroxybutyryl-CoA, which can then be hydrolyzed to 2HB by native thioesterase activity (Figure 33A).

[0127] First round screening of HACS variants shows that 10 variants out of 29 produce significant 2HB productivity (Figure 33B). JGI20, JGI23 and JGI24 are the three best candidates that show 3-fold higher 2HB productivity compared to other HACS variants.

[0128] Quantification of product concentration (2HB) for the second round HACS variants was determined by the SoGO method described in Example 8. Results show that three variants (JGIH25, JGIH28 and JGIH48) performed better than JGI23, with JGIH28 having an increase in 2HB productivity of over 40% (Figure 33C).

[0129] Example 11: Screening of 1st and 2nd round HACS variants for activity with glycolaldehyde This example shows the first and second round screening of HACS variants with glycolaldehyde as a substrate in vivo. We used glyceric acid (glycerate) productivity per cell density (μM glycerate / OD600) as an indicator of HACS activity. HACS variants were screened for 2HB production using the high-throughput screening platform described in Example 3 by co-feeding 5 mM glycolaldehyde and 20 mM formate. HACS condenses glycolaldehyde and formyl-CoA to form glyceryl-CoA, which can then be hydrolyzed to glycerate by native thioesterase activity (Figure 34A).

[0130] The first round screening of HACS variants shows that 9 variants out of 29 produce significant glycerate productivity (Figure 34B). JGI15 and JGI20 are the two best candidates that show more than 2-fold higher glycerate productivity compared to other HACS variants.

[0131] Based on the phylogenetic tree analysis (Figure 29), since JGI15 and JGI20 are the best performing candidates from the first round HACS screening, we expect that variants belonging to the JGI15 and / or JGI20 clusters will show comparable performance (glycerate productivity) for the second round HACS variant screening.

[0132] Example 12: Screening of first and second round HACS variants for activity with glyoxylic acid This example shows the first and second round screening of HACS variants using glyoxylic acid (glyoxylate) as substrate in vivo. We used tartronic acid (tartronate) productivity per cell density (μM tartronate / OD600) as an indicator of HACS activity. HACS variants were screened for tartronate production using the high-throughput screening platform described in Example 3 by co-feeding 5 mM glyoxylate and 20 mM formate. HACS condenses glyoxylate and formyl-CoA to form tartronyl-CoA, which can then be hydrolyzed to tartronate by native thioesterase activity (Figure 35A).

[0133] The first round of screening of HACS variants shows that 6 variants out of 29 produce significant tartronate productivity (Figure 35B). JGI20 is the best candidate, showing 30% better tartronate productivity compared to other HACS variants.

[0134] Based on the phylogenetic tree analysis (Figure 29), since JGI20 is the best performing candidate from the first round HACS screening, we expect that variants belonging to the JGI20 cluster will show comparable performance (tartronate productivity) for the second round HACS variant screening.

[0135] Example 13: Screening of first and second round HACS variants for activity with 3-hydroxypropionaldehyde This example shows the first and second round screening of HACS variants with 3-hydroxypropionaldehyde (3HP) as a substrate in vivo. We used 2,4-dihydroxybutyrate (DHB) productivity per cell density (μM DHB / OD600) as an indicator of HACS activity. HACS variants were screened for DHB production using the high-throughput screening platform described in Example 3 by co-feeding with 5 mM 3HP and 20 mM formate. HACS condenses 3HP and formyl-CoA to form 2,4-dihydroxybutyryl-CoA, which can then be hydrolyzed to DHB by native thioesterase activity (Figure 36A).

[0136] First round screening of HACS variants indicates that 3 variants out of 29 (JGI15, JGI20 and RuHACL) result in significant DHB productivity (Figure 36B).

[0137] Based on the phylogenetic tree analysis (Figure 29), JGI15, JGI20 and RuHACL are the best performing candidates from the first round HACS screening, so we expect that variants belonging to these clusters will show comparable performance (DHB productivity) for the second round HACS variant screening.

[0138] Example 14: Screening of first round variants for activity with ketones This example shows the first round 2 screening of HACS variants using various ketones as substrates for branched compound production. HACS variants are tested using the high-throughput screening platform described in Example 3, Pathway 2 by co-supplying 100 mM acetone and 20 mM formate with the formate-activating enzyme CaAbfT (Figure 37A).

[0139] The results show that JGI15, JGI19 and JGI20, together with AcHACL, perform better than other HACLs, with JGI15 performing the best. Kinetic characterization of JGI15 and AcHACL with acetone and formate was performed using the methods described in Example 4. JGI15 showed significantly better activity (higher K cat ), but has a fairly high K m (Table 8). AcHACL has even worse activity, but a significantly lower K m (FIG. 37B). Better HACL for 2HIB production by condensation of acetone and formyl-CoA is expected in the second round HACS variants using AcHACL or JGI15 as reference.

[0140] Example 15: Methyl ketones as substrates for condensation with formyl-COA by in vitro assay This example shows the implementation of condensation of methyl ketone with formyl-CoA using purified enzymes. Formyl-CoA production catalyzed by CoA transferase CaAbfT and condensation catalyzed by HACS JGI15 are the same as the above example (Figure 38A). Methyl ketone can be produced by fatty acid synthesis and β-oxidation pathways shown in the literature (Goh EB, et al., Appl Environ Microbiol 78:70-80(2012);Nies SC, et al., Metab Eng 62:84-94(2020)).

[0141] The enzymes CoA transferase CaAbfT and HACS JGI15 were overexpressed and purified as described above. In vitro purified enzyme reactions for the condensation of methyl ketones and formyl-CoA consisted of 100 mM KPi pH 6.9, 10 mM MgCl2, 0.15 mM TPP, 2 mM acetyl-CoA, 1 μM JGI15, 2 μM CaAbfT, 20 mM formate and 100 mM of the tested methyl ketone. Reactions were incubated at 30°C for 24 hours unless otherwise specified.

[0142] For this analysis, samples containing acyl-CoA were first treated with 10 M NaOH solution in 1 / 20 of the reaction volume, which was added to terminate the reaction. After 30 min of hydrolysis, 10 N H2SO4 in 1 / 20 of the reaction volume was added to improve the efficiency of the acid extraction. The resulting samples were extracted into 4 mL ethyl acetate by vigorous vortexing for 90 s. The organic phase was separated and evaporated to dryness under a stream of nitrogen. The residue was dissolved in 50 μL pyridine and 50 μL N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and incubated at 60°C for 15 min. Compound identification and analysis were performed by GC-MS using an Agilent 7890B Series Custom Gas Chromatography system equipped with a 5977B Inert Plus Mass Selective Detector Turbo EI Bundle (for identification) and an Agilent HP-5-ms capillary column (0.25 mm ID, 0.25 μm film thickness, 30 m length). Samples were analyzed by GC (1 μL injection with a separation ratio of 20:1) using helium as the carrier gas at a flow rate of 1.5 mL / min and the following temperature profile: start 90° C. for 3 min; ramp 15° C. / min to 170° C.; ramp 20° C. / min to 300° C. and hold for 8 min. Injector and detector temperatures were 250° C. and 350° C., respectively.

[0143] Condensation may be carried out using, but is not limited to, acetone, methyl ethyl ketone (C nJGI15 can catalyze the condensation of ketones tested as shown in Figure 38B, indicating that other identified HACS can condense other ketones with formyl-CoA to make 2-hydroxy-2-methyl acids and derivatives (Figure 39). For condensation, examples include, but are not limited to, acetone, methyl ethyl ketone (C 2 -ketones, n>3, e.g., butanone, pentanone, and heptanone), hydroxylated ketones (hydroxyacetone), and other functionalized ketones (acetylacetone, branched chain ketones, methylglyoxal). ... n -ketones, n>3, e.g. butanone, pentanone and heptanone), hydroxylated ketones (hydroxyacetone) and other functionalized ketones (acetylacetone, branched chain ketones, methylglyoxal) can be used (Figures 40-47).

[0144] Example 16: Identification, synthesis and screening of ACR variants This example shows the identification, synthesis, and screening of acyl-CoA reductase (ACR) variants for specific acylation of the formaldehyde oxidation (formaldehyde to formyl-CoA) reaction. From an initial screening of known ACRs as measured by glycolate productivity coupled with RuHACL (Figure 48A), we identified acyl-CoA reductase (LmACR) from Listeria monocytogenes as the most active (Figure 48B) and selected it as a starting reference for identifying novel enzymes with sequence similarity.

[0145] Using the methods described in Example 1 (Figure 19), 44 novel ACR variants (JGIR1-44) were identified, and 40 final constructs will be synthesized in collaboration with the Joint Genome Institute (Figure 49A, Table 9).

[0146] Table 9. List of acyl-CoA reductase (ACR) variants (JGIR) identified by selecting representative genes from gene clusters with sequence similarity using LmACR as the reference enzyme. [Table 10-1] [Table 10-2]

[0147] To reduce the complexity of the overall reaction scheme, ACR variants are tested in the same resting cell format as described in Example 2, without the presence of HACS (Figure 49B). Consequently, formaldehyde reduction to formyl-CoA catalyzed by ACR variants in vivo was measured using a high-throughput NASH colorimetric assay (Nash, Biochem J. 55(3):416-421 (1953)). Formaldehyde consumption (OD600) per cell density was calculated and compared with LmACR activity as a reference (Figure 49C). There are three variants (JGIR2, 5 and 14) that show 10-20% improved formaldehyde consumption activity under both low formaldehyde (0.5 mM) and high formaldehyde (3 mM). JGIR10 exhibits a significant improvement in activity (40%) under high formaldehyde compared to LmACR, but exhibits reduced activity at low formaldehyde, possibly due to a higher k cat but with a higher K m There are several other variants with activity comparable to LmACR that merit further investigation in other conditions.

[0148] Example 17: Identification, synthesis and screening of formate-activating enzyme (ACT and ACK-PTA) variants This example shows the identification, synthesis and screening of acyl-CoA transferase (ACT) variants and acyl-CoA kinase (ACK) and phosphoacyltransferase (PTA) variants specifically for the formate activation (formate to formyl-CoA) reaction (Figure 50A). From the initial screening of known ACT and ACK-PTA variants measured by glycolate productivity coupled with JGI15 (Figure 50B), we identified the most active acyl-CoA transferase (CaAbfT) from Clostridium aminobutyricum from the ACT variants and the most active acyl-CoA kinase and phosphoacyltransferase combination (CcAck-Pta) from Clostridium cylindrosporum from the ACK-PTA variants (Figure 50B), and selected them as starting references for identifying novel enzymes with sequence similarity.

[0149] Using the methods described in Example 1 (Figure 19), 62 novel ACT variants (JGIT1-62) and 38 novel ACK-PTA variants will be identified and synthesized in collaboration with the Joint Genome Institute (Figure 51AB, Table 10).

[0150] Table 10. List of acyl-CoA transferase (ACT) variants (JGITs) identified by selecting representative genes from gene clusters with sequence similarity using CaAbfT and OfFrc as reference enzymes. [Table 11-1] [Table 11-2] [Table 11-3]

[0151] ACT and ACK-PTA variants are tested in the same resting cell format as described in Example 3 (pathway 2) using different formate-activating enzyme variants instead of JGI20 and CaAbfT as HACS (Figure 52A). The glycolate activity of the variants was measured by adding 2.5 mM formaldehyde and 20 mM formate as carbon source. The results show that two ACT variants (JGIRT45 and 51) show equivalent or improved CoA transferase activity, and several ACK-PTA variants show significant formyl-CoA making activity that was not observed from CcAck-Pta (Figure 52B). This is consistent with our previous observation that CcAck-Pta requires high formate concentration (50 mM) to show significant glycolate productivity indicative of the high Km of this enzyme (Figure 52B). Some ACK-PTA variants, such as JGIK1, 18 and 31, showed comparable activity to the high-performance ACT variants, which could provide more diverse pathways for making formyl-CoA even under relatively low formate concentrations (Figure 52B).

[0152] Example 18: Strategies for engineering and screening enzymes with improved catalytic efficiency This example shows a potential strategy to further engineer HACS, ACR, ACT and ACK-PTA enzymes for improved activity and selectivity towards desired substrate(s). The approach described in Examples 5 and 6 can be applied not only to the second round HACS variants, but also to other variants for other ACR, ACT and ACK-PTA variants. The structure modeled by AlphaFold following homology-guided alignment allows the identification of active site residues as shown in Example 5. These key residues can then be targeted for directed evolution by saturation mutagenesis. Both simultaneous and iterative mutagenesis can be considered for directed evolution. Alternatively, DNA shuffling of multiple variants with high expression, activity or substrate specificity can be shuffled to identify candidates with higher catalytic efficiency. Random mutagenesis of candidate genes by error-prone PCR is also an option.

[0153] To increase the throughput of the screening method, a selection-based screening method can be used. To screen ACR for formaldehyde oxidation activity, we can enhance the toxicity of formaldehyde. E. coli lacking the formaldehyde detoxification pathway (frmA) cannot survive under submillimolar concentrations of formaldehyde. High catalytic efficiency (k cat / K m Cells with ACR having the ribosomal ATPase 1 (RIA) can rapidly convert formaldehyde to the substantially less toxic formyl-CoA, allowing them to survive in the presence of other nutrients for cell maintenance and growth.

[0154] We also developed a selection-based screening platform for glycolate production via a glycine autotrophic strain. As a host for the selection platform, we engineered a glycine autotrophic strain of E. coli based on MG1655(DE3) with a knockout for glycine production and utilization (ΔaceA Δkbl ΔltaE ΔglyA), and grew this strain only with glycine supplementation (Figure 53A). Glycolate can be first oxidized to glyoxylate, catalyzed by E. coli glcD, followed by the promiscuous activity of a heterologous alanine dehydrogenase that catalyzes the reduction of glyoxylate to glycine (Figure 53A). Alanine dehydrogenase from Mycobacterium tuberculos and Bacillus subtilis has been shown to have the activity of using glyoxylate to generate glycine (J. Bacteriol. 194:1045-1054, 2012; Biochemistry 20:5650-5655, 1981).

[0155] For gene deletion, CRISPR is used based on the method developed in Appl. Environ. Microbiol. 81:2506-2514, 2015). First, the host strain is transformed with the vector for expression of plasmid pCas, Cas9 and λ-red recombinase. The resulting strain is grown under 30°C with L-arabinose for induction of λ-red recombinase expression, and when OD reaches about 0.6, competent cells are prepared and transformed with pTargetF (AddGene 62226) expressing sgRNA and N20 spacer that target the insertion locus of the target gene and template. The template is the deleted gene with about 500 bp sequence homologous to the upstream and downstream of the insertion locus, which is constructed by overlap PCR using Phusion polymerase or synthesized by GenScript (Piscataway, NJ). The method for replacing the N20 spacer of the pTargetF plasmid is inverse PCR with a modified N20 spacer dangling at the 5' end of the primer using Phusion polymerase, followed by self-ligation using T4 DNA ligase and T4 polynucleotide kinase (New England Biolabs, Ipswich, MA, USA). Transformants growing at 30°C on solid medium (LB+agar) supplemented with spectinomycin and kanamycin (or other appropriate antibiotics) are isolated and screened for chromosomal gene inserts by PCR. The sequences of gene inserts amplified from genomic DNA by PCR using Phusion polymerase are further confirmed by DNA sequencing. pTargetF can then be cured by IPTG induction and pCas can be cured by growth at higher temperatures such as 37-42°C.

[0156] The resulting glycine autotrophic strains were transformed with vectors constitutively expressing alanine dehydrogenase from Mycobacterium tuberculosis (MtAld) or Bacillus subtilis (BsAld). When the strains were incubated in minimal medium (M9) with 5 g / L glucose, they could not grow without glycine supplementation, indicating glycine autotrophy. Of the two candidates, the strain with BsAld started growing with glycolate instead of glycine supplementation (Figure 28B), indicating that glycolate was successfully converted to glycine by the native glcD and heterologously expressed BsAld genes. This also showed that a minimal supplement of 50 mg / L glycolate was sufficient to see growth, indicating a suitable platform for selection-based screening of HACS, ACR, ACT and ACK-PTA variants for glycolate production from C1 compounds.

[0157] Sequence information for certain examples and embodiments described herein is as follows: [Table 12-1] [Table 12-2] [Table 12-3] [Table 12-4] [Table 12-5] [Table 12-6] [Table 12-7] [Table 12-8] [Table 12-9]

Table 12-10

Claims

1. A recombinant microorganism expressing a 2-hydroxyacyl-CoA synthase, wherein the 2-hydroxyacyl-CoA synthase is capable of enzymatically forming 2-hydroxyacyl-CoA from a carbonyl-containing compound and formyl-CoA at a rate that is at least two-fold, and alternatively three-fold, faster than the Rhodospirillales bacterium URHD0017 2-hydroxyacyl-CoA synthase.

2. A recombinant microorganism as described in claim 1, wherein the 2-hydroxyacyl-CoA synthase is at least 90% identical to SEQ ID NO: 1 (JGI15) or SEQ ID NO: 3 (JGI20).

3. A recombinant microorganism as described in claim 2, wherein the 2-hydroxyacyl-CoA synthase has the sequence of SEQ ID NO:

1.

4. A recombinant microorganism as described in claim 2, wherein the 2-hydroxyacyl-CoA synthase has the sequence of SEQ ID NO:

3.

5. A recombinant microorganism as described in claim 2, wherein the 2-hydroxyacyl-CoA synthase contains one or more mutations relative to SEQ ID NO: 3, and optionally the mutations are N461del and R480ins relative to SEQ ID NO: 3, A253G and P254G relative to SEQ ID NO: 3, and / or L549H, T550G and R551del relative to SEQ ID NO:

3.

6. A recombinant microorganism described in claim 1 or 2, wherein the microorganism further expresses an enzyme catalyst that produces formyl-CoA from a one-carbon substrate.

7. 3. The recombinant microorganism of claim 1, wherein the carbonyl-containing compound is selected from the group consisting of aldehydes and ketones.

8. A recombinant microorganism as described in claim 1 or 2, further comprising an enzyme catalyst that converts a substrate to a carbonyl-containing compound.

9. 3. The recombinant microorganism of claim 1 or 2, further comprising an enzyme catalyst that converts 2-hydroxyacyl-CoA into an organic chemical product.

10. The one-carbon substrate is formaldehyde and the enzyme catalyst that produces formyl-CoA is: a. an acyl-CoA reductase (acylating aldehyde dehydrogenase) that catalyzes the conversion of formaldehyde to formyl-CoA, or The one-carbon substrate is methanol and the enzyme catalyst that produces formyl-CoA is: a. methanol dehydrogenase, which catalyzes the conversion of methanol to formaldehyde; and b. Acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA is, or The one-carbon substrate is methane and the enzyme catalyst that produces formyl-CoA is: a. methane monooxygenase, which catalyzes the conversion of methane to methanol; b. methanol dehydrogenase, which catalyzes the conversion of methanol to formaldehyde; and c. acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA is, or The one-carbon substrate is formate, and the enzyme catalyst that produces formyl-CoA is: a. an acyl-CoA synthase, which catalyzes the conversion of formate to formyl-CoA; or b. Formate kinase, which catalyzes the conversion of formate to formyl phosphate, and phosphate formyl transferase, which catalyzes the conversion of formyl phosphate to formyl CoA. is, or The one-carbon substrate is carbon dioxide and the enzyme catalyzed to produce formyl-CoA is: a. a carbon dioxide reductase, which catalyzes the conversion of carbon dioxide to formate; and b. an acyl-CoA synthase, which catalyzes the conversion of formate to formyl-CoA; or c. Formate kinase, which catalyzes the conversion of formate to formyl-phosphate, and phosphate formyltransferase, which catalyzes the conversion of formyl-phosphate to formyl-CoA That is, The recombinant microorganism of claim 6.

11. The method of claim 1, wherein the product is an aldehyde and the enzyme catalyst that converts 2-hydroxyacyl-CoA to said product is: a. Acyl-CoA reductase, which catalyzes the conversion of 2-hydroxyacyl-CoA to an aldehyde is, or The product is an alcohol and the enzyme catalyst that converts 2-hydroxyacyl-CoA to the product is: a. an acyl-CoA reductase, which catalyzes the conversion of 2-hydroxyacyl-CoA to an aldehyde; and b. Alcohol dehydrogenase (aldehyde reductase), which catalyzes the conversion of aldehydes to alcohols is, or The product is a carboxylic acid and the enzyme catalyst that converts 2-hydroxyacyl-CoA to the product is: a. Thioesterases, which catalyze the conversion of 2-hydroxyacyl-CoA to carboxylic acids That is, The recombinant microorganism of claim 9.

12. A recombinant microorganism as described in claim 1 or 2, wherein the modified organism is a bacterium.

13. The recombinant microorganism of claim 12, wherein the bacterium is Escherichia coli.

14. A method for forming 2-hydroxyacyl-CoA from a carbonyl-containing compound and formyl-CoA, wherein the formation of 2-hydroxyacyl-CoA is catalyzed by a 2-hydroxyacyl-CoA synthase, which is capable of enzymatically forming 2-hydroxyacyl-CoA from a carbonyl-containing compound and formyl-CoA at a rate that is at least two-fold, alternatively three-fold, faster than the Rhodospirillales bacterium URHD0017 2-hydroxyacyl-CoA synthase.

15. The method described in claim 14, wherein the 2-hydroxyacyl-CoA synthase is at least 90% identical to SEQ ID NO: 1 (JGI15) or SEQ ID NO: 3 (JGI20).

16. The method of claim 14 or 15, further comprising forming formyl-CoA from a one-carbon substrate, wherein the formation of formyl-CoA is catalyzed by an enzyme catalyst.

17. i) conversion of a substrate to a carbonyl-containing compound, wherein the conversion to the carbonyl-containing compound is catalyzed by an enzyme catalyst. ii) conversion of 2-hydroxyacyl-CoA to an organic chemical product, wherein the conversion of 2-hydroxyacyl-CoA to an organic chemical product is catalyzed by an enzyme catalyst.

18. A process according to claim 1, wherein the one-carbon substrate is formaldehyde and the enzyme catalyst that catalyzes the formation of formyl-CoA is: a. It is an acyl-CoA reductase (acylating aldehyde dehydrogenase) that catalyzes the conversion of formaldehyde to formyl-CoA; ii) The one-carbon substrate is methanol and the enzyme catalyst that catalyzes the formation of formyl-CoA is: a. methanol dehydrogenase, which catalyzes the conversion of methanol to formaldehyde; and b. Acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA That is, iii) The one-carbon substrate is methane and the enzyme catalyst that catalyzes the formation of formyl-CoA is: a. methane monooxygenase, which catalyzes the conversion of methane to methanol; b. methanol dehydrogenase, which catalyzes the conversion of methanol to formaldehyde; and c. acyl-CoA reductase (acylating aldehyde dehydrogenase), which catalyzes the conversion of formaldehyde to formyl-CoA That is, iv) The one-carbon substrate is formate and the enzyme catalyst that catalyzes the formation of formyl-CoA is: a. an acyl-CoA synthase, which catalyzes the conversion of formate to formyl-CoA; or b. Formate kinase, which catalyzes the conversion of formate to formyl phosphate, and phosphate formyl transferase, which catalyzes the conversion of formyl phosphate to formyl CoA. is, or v) The one-carbon substrate is carbon dioxide and the enzyme catalyst that catalyzes the formation of formyl-CoA is: a. a carbon dioxide reductase, which catalyzes the conversion of carbon dioxide to formate; and b. an acyl-CoA synthase, which catalyzes the conversion of formate to formyl-CoA; or c. Formate kinase, which catalyzes the conversion of formate to formyl-phosphate, and phosphate formyltransferase, which catalyzes the conversion of formyl-phosphate to formyl-CoA That is, 17. The method of claim 16.

19. The method of claim 14 or 15, wherein the carbonyl-containing compound is selected from the group consisting of aldehydes and ketones.

20. The method of claim 19, wherein the aldehyde has at least one substituent, the substituent being hydroxy, carbonyl, carboxyl, alkyl, alkenyl, alkynyl, or amine.

21. The method of claim 19, wherein the ketone is a methyl ketone.

22. A method as described in claim 14 or 15, wherein the enzyme catalyst is contained in a recombinant microorganism carrying genes for expressing each enzyme, and optionally the substrate is contacted with the recombinant microorganism containing the enzyme catalyst in an aqueous medium optionally containing a buffer, salt, vitamin, or inorganic substance.

23. A method as described in claim 14 or 15, wherein the substrate is contacted with the enzyme catalyst by addition to an aqueous reaction mixture, optionally containing a buffer, salts, vitamins, minerals and cofactors, and optionally the enzyme catalyst is provided as a crude cell extract or purified protein.

24. A recombinant 2-hydroxyacyl-CoA synthase, wherein the 2-hydroxyacyl-CoA synthase comprises one or more mutations relative to SEQ ID NO: 3 (JGI20), and wherein the 2-hydroxyacyl-CoA synthase is capable of enzymatically forming 2-hydroxyacyl-CoA from a carbonyl-containing compound and formyl-CoA at a rate that is at least two-fold, and alternatively three-fold, faster than the Rhodospirillales bacterium URHD0017 2-hydroxyacyl-CoA synthase.

25. The recombinant 2-hydroxyacyl-CoA synthase described in claim 24, wherein the mutations are N461del and R480ins relative to SEQ ID NO: 3, A253G and P254G relative to SEQ ID NO: 3, and / or L549H, T550G and R551del relative to SEQ ID NO: 3.

Images