A (S)-3-Hydroxybutyrate Dehydrogenase Belonging to the 3-Hydroxyacyl-CoA Dehydrogenase Family Facilitates Hydroxyacid Degradation in Anaerobic Bacteria

ABSTRACT

Ketone bodies, including acetoacetate, 3-hydroxybutyrate, and acetone, are produced in the liver of animals during glucose starvation. Enzymes for the metabolism of (R)-3-hydroxybutyrate have been extensively studied, but little is known about the metabolism of its enantiomer (S)-3-hydroxybutyrate. Here, we report the characterization of a novel pathway for the degradation of (S)-3-hydroxybutyrate in anaerobic bacteria. We identify and characterize a stereospecific (S)-3-hydroxylbutyrate dehydrogenase (3SHBDH) from Desulfotomaculum ruminis, which catalyzes the reversible NAD(P)H-dependent reduction of acetoacetate to form (S)-3-hydroxybutyrate. 3SHBDH also catalyzes oxidation of d-threonine (2R, 3S) and l-allo-threonine (2S, 3S), consistent with its specificity for β-(3S)-hydroxy acids. Isothermal calorimetry experiments support a sequential mechanism involving binding of NADH prior to (S)-3-hydroxybutyrate. Homologs of 3SHBDH are present in anaerobic fermenting and sulfite-reducing bacteria, and experiments with Clostridium pasteurianum showed that 3SHBDH, acetate CoA-transferase (YdiF), and (S)-3-hydroxybutyryl-CoA dehydrogenase (Hbd) are involved together in the degradation of (S)-3-hydroxybutyrate as a carbon and energy source for growth. (S)-3-hydroxybutyrate is a human metabolic marker and a chiral precursor for chemical synthesis, suggesting potential applications of 3SHBDH in diagnostics or the chemicals industry.

IMPORTANCE (R)-3-hydroxybutyrate is well studied as a component of ketone bodies produced by the liver and of bacterial polyesters. However, the biochemistry of its enantiomer (S)-3-hydroxybutyrate is poorly understood. This study describes the identification and characterization of a stereospecific (S)-3-hydroxylbutyrate dehydrogenase and its function in a metabolic pathway for the degradation of (S)-3-hydroxybutyrate as a carbon and energy source in anaerobic bacteria. (S)-3-hydroxybutyrate is a mammalian metabolic marker and a precursor for chemical synthesis and bioplastics, suggesting potential applications of these enzymes in diagnostics and biotechnology.

INTRODUCTION

During glucose deprivation in the human body, the ketone bodies acetoacetate, 3-hydroxybutyrate (3HB), and acetone are generated in the mitochondria of liver cells from acetyl coenzyme A (acetyl-CoA) derived from the beta oxidation of fatty acids, serving as an alternative fuel for the heart, brain, and other tissues (1). The major physiological isomer of 3HB is (R)-3-hydroxybutyrate (3RHB), which is produced by the stereospecific mitochondrial 3RHB dehydrogenase (2) and thought to play diverse physiological roles in regulating the immune barrier as a signaling metabolite (3), as well as in oxidative stress (4), neuroprotection (5), histone modification (6), anti-aging (7), and tumor-inhibition (8). (S)-3-hydroxybutyrate (3SHB) is also present in humans, but its levels in plasma are lower than those of 3RHB (9), and its physiological significance is not understood. In 1980, Reed et al. (10) presented evidence that the liver was a site of 3SHB production in rats. However, the work of Scofield et al. (11) two years later contradicted these findings. As a result, the biosynthetic origin of 3SHB in mammals remains elusive.

3RHB also plays important roles in diverse bacteria as a component of polyhydroxybutyrates (PHBs), which function in carbon and energy storage (12). In contrast, 3SHB is not known to be present in bacterial PHBs (13). The CoA thioesters of 3RHB and 3SHB and their stereospecific dehydrogenases are involved in distinct metabolic pathways (14, 15). 3RHB is involved in PHB biosynthesis in bacteria and fatty acid biosynthesis in both bacteria and eukaryotes (13). On the other hand, 3SHB is involved in butyrate fermentation in Clostridia and in fatty acid degradation through the beta oxidation pathway in both bacteria and eukaryotes. Bacterial PHB degradation involves isomerization of 3RHB-CoA to 3SHB-CoA catalyzed by crotonase, followed by degradation through the beta oxidation pathway (16, 17). In bacteria, free 3RHB is generated by PHB oligomer hydrolase (18). Natural bacterial production of free 3SHB is not known, but both 3RHB and 3SHB have been produced in engineered strains of Escherichia coli by hydrolysis of their respective CoA thioesters, with 3RHB generated by the 3RHB-CoA dehydrogenase PhaB from the PHB biosynthetic pathway, and 3SHB generated by the 3SHB-CoA dehydrogenase Hbd from the clostridial butyrate fermentation pathway (19).

The enzymes and metabolic pathway for 3RHB degradation in mammals and bacteria has been extensively studied (20–25). In mammals, 3RHB is degraded in the mitochondria by 3RHB dehydrogenase, 3-oxoacid CoA-transferase, and acetoacetyl CoA thiolase to form acetyl-CoA (2, 26). In addition, mammals also possess a cytosolic hydroxyacid dehydrogenase belonging to the superfamily of NADPH-dependent short-chain dehydrogenase/reductases (SDRs) which oxidize 3RHB (27) and cis-4-hydroxy-l-proline (28). Certain bacteria, including the PHB-producing Cupriavidus necator, contain 3RHB dehydrogenase involved in 3RHB degradation through a similar pathway (29). Much less is known about the enzymes and pathways responsible for 3SHB degradation. The human lambda-crystallin homolog CRYL1 is a cytosolic 3S-hydroxyacid dehydrogenase that catalyzes NAD⁺-dependent oxidation of gluconate, 3SHB, and l-threonine (30). In 1994, an NADP⁺-dependent (S)-3-hydroxycarboxylate dehydrogenase was discovered and characterized in Clostridium tyobutyricum (31), and its N-terminal 45-amino acid motif was identified (NCBI reference no. AAB33485.1). However, further details are lacking.

As part of a bioinformatics study aimed at discovering new enzymes and metabolic pathways in sulfate- and sulfite-reducing bacteria (SSRB), we became interested in the enzyme Desru_2141 from Desulfotomaculum ruminis, which occurs in proximity to a glycyl radical enzyme and is annotated as a (S)-3-hydroxyacyl-CoA dehydrogenase (belonging to the 3HCDH Pfam family, PF00725). However, our preliminary studies showed that it lacks the predicted activity. Here, we demonstrate that Desru_2141 is in fact a 3SHB dehydrogenase (Dr3SHBDH). During the course of our work, a literature search revealed that Dr3SHBDH is closely related to the (S)-3-hydroxycarboxylate dehydrogenase from C. tyobutyricum, which, despite being described in the literature, was misannotated in sequence databases. We describe the involvement of this enzyme in dedicated pathways for the degradation of 3SHB as a carbon and energy resource in anaerobic bacteria.

RESULTS

Purification and characterization of recombinant Dr3SHBDH.

Dr3SHBDH was recombinantly produced and purified to near homogeneity (Fig. 1A). The purified enzyme was subjected to gel filtration chromatography, which showed a single symmetric peak at 230 mL (Fig. 1B), corresponding to a molecular mass of 38 kDa. The calculated molecular weight for Dr3SHBDH monomer is 35 kDa, indicating that the enzyme likely exists as a monomer in solution. Preliminary assays showed that purified Dr3SHBDH does not catalyze reduction of acetoacetyl-CoA.

FIG 1.

Dr3SHBDH is an NAD(P)+-dependent (S)-3-hydroxybutyrate dehydrogenase. (A) Purified Dr3SHBDH on 10% SDS gel with (lane 1) molecular weight marker and (lanes 2 to 4) 1, 2, and 4 μg of Dr3SHBDH. (B) Elution profile of the purified Dr3SHBDH using Superdex 200 size exclusion chromatography to determine its molecular weight. (C) Superposition of the homology of Dr3SHBDH (blue) with (S)-3-hydroxybutyryl-CoA dehydrogenase from Clostridium butyricum E4 (tan, PDB ID: 4KUH [14]) and Clostridium acetobutylicum (gray, PDB ID: 6AA8 [32]). (D) Superposition of the Dr3SHBDH (blue) with 4KUH active site (gray). Essential residues involved in substrate binding are displayed and labeled. (E) Acetoacetate reductase assay of Dr3SHBDH. (F) 3SHB dehydrogenase assay of Dr3SHBDH. (G) Enzyme activity assay of Dr3SHBDH with 3SHB or (RS)-3HB-CoA.

To gain insight into the substrate of Dr3SHBDH, we compared the AlphaFold structure of this enzyme to those of its closest structurally characterized homologs, the (S)-3-hydroxybutyryl-CoA dehydrogenases from Clostridium butyricum (14) (CbHbd) and Clostridium acetobutylicum (32) (CaHbd) (Fig. S1 in the supplemental material, Fig. 1C and D). These three enzymes share a conserved N-terminal Rossmann fold domain that contains residues for binding NADPH (Fig. 1C). Additionally, the catalytic residues and substrate-binding residues are conserved among these enzymes (H138, S117, and N188 in CbHbd; H145, S124, and N196 in Dr3SHBDH) (Fig. S1 and Fig. 1D). However, the cationic and hydrophobic residues involved in the interaction with the CoA moiety are not conserved in Dr3SHBDH (Fig. S1 and Fig. 1D). This suggests that the substrate for Dr3SHBDH could be the free organic acid instead of the CoA thioester.

We assessed the enzymatic activity of Dr3SHBDH using a NADPH spectrophotometric assay. Incubation of Dr3SHBDH with acetoacetate and NADPH resulted in a time-dependent decrease in absorbance at 340 nm (NADPH, ε₃₄₀ [extinction coefficient at 340 nm] = 6,220 M⁻¹ cm⁻¹), corresponding to NADPH consumption. No reaction was observed when acetoacetyl-CoA replaced acetoacetate in the reaction system (Fig. 1E). Furthermore, incubation of Dr3SHBDH with 3SHB and NAD⁺ led to a time-dependent increase in absorbance at 340 nm (A₃₄₀) (Fig. 1F), corresponding to NADH production. Furthermore, no reaction was observed when (RS)-3HB-CoA replaced 3SHB in the reaction system (Fig. 1G). This demonstrates that Dr3SHBDH is a NADPH-dependent (S)-3-hydroxybutyrate dehydrogenase.

Substrate specificity and kinetics of Dr3SHBDH.

Assays of Dr3SHBDH were carried out with 1 mM NAD⁺, 50 mM CAPSO [3-(cyclohexylamino)-2-hydroxy-1-propanesulfonic acid] buffer (pH 10.0), and 1 mM of either isomer of 3HB. The results showed that Dr3SHBDH was only active with 3SHB, not with 3RHB (Fig. 2A and B). The production of acetoacetate from 3SHB was confirmed by mass spectrometry (MS) analysis. No reaction was observed in negative controls without 3SHB. Furthermore, no activity was detected when 3RHB was substituted for 3SHB (Fig. 2C). Acetone-DNPH was also detected in the reaction with 3SHB by derivatization with 2,4-dinitrophenylhydrazine (DNPH) followed by liquid chromatography-MS (LC-MS) analysis, indicating spontaneous decarboxylation of acetoacetate during the reaction or workup (Fig. 2D and E).

FIG 2.

UV-Vis Spectrometry, MS, and LC-MS assays of Dr3SHBDH-catalyzed 3SHB oxidation. (A) Reaction scheme for 3SHBDH-catalyzed 3SHB oxidation, spontaneous acetoacetate decarboxylation, and DNPH-derivatization. (B) Spectrophotometric assay for Dr3SHBDH-catalyzed 3SHB oxidation, showing reaction requirements and stereospecificity. (C) ESI (–) m/z spectra of underivatized reaction products, showing formation of acetoacetate from 3SHB. (D) Elution profile of the DNPH-derivatized products for the LC-MS analysis, monitoring absorbance at 360 nm. (E) The ESI (–) m/z spectrum of the acetone-DNPH peak in panel D.

Dehydrogenase activity of Dr3SHBDH was also detected with d-threonine (2R, 3S) and l-allo-threonine (2S, 3S), as well as with 3-hydroxylpropionate, d-Ser, and l-Ser, but not with l-threonine (2S, 3R), d-allo-threonine (2R, 3R), or racemic 2-hydroxybutyrate. This suggests that Dr3SHBDH has a high specificity for (S)-3-hydroxyacids, with tolerance for substitutions at the 2-position. The apparent Michaelis-Menten kinetic parameters were measured using the spectrophotometric assay (Table 1 and Fig. S2).

TABLE 1.

Spectrophotometric assay data^a

Substrate	kcat (s−1)	KM (mM)	kcat/KM (M−1 s−1)
3RHB	NA
3SHB	29.25 ± 0.64	0.08 ± 0.00	3.7 × 105
Sodium 2-hydroxybutyrate	NA
l-Threonine (2S, 3R)	NA
d-Threonine (2R, 3S)	8.25 ± 0.23	0.59 ± 0.06	1.4 × 104
l-Allo-threonine (2S, 3S)	24.86 ± 0.73	2.71 ± 0.20	9.2 × 103
d-Allo-threonine (2R, 3R)	NA
3-Hydroxylpropionate	0.15 ± 0.00	0.76 ± 0.07	0.2 × 103
l-Ser	0.04 ± 0.00	8.02 ± 1.66	4.9
d-Ser	0.08 ± 0.00	2.93 ± 0.34	27.3
NAD+	21.89 ± 0.36	0.33 ± 0.02	6.6 × 104
NADP+	2.72 ± 0.07	0.03 ± 0.00	9.1 × 104
Acetoacetate	14.78 ± 0.48	0.03 ± 0.00	4.9 × 105
NADH	12.03 ± 0.50	0.07 ± 0.01	1.7 × 105
NADPH	7.32 ± 0.67	0.07 ± 0.02	1.1 × 105

^aNA, no activity.

The substrate-binding properties of Dr3SHBDH.

The substrate-binding properties of Dr3SHBDH were investigated through isothermal titration calorimetry (ITC) assays. In the absence of NADH, Dr3SHBDH did not bind 3SHB (Fig. 3A). However, when NADH was included in the premix, Dr3SHBDH was able to bind 3SHB, with a dissociation constant (K_d) of 23 μM, ΔH (change in enthalpy, kcal/mol) = −3.81, and ΔG (change in Gibbs free energy, kcal/mol) = −6.83. The N value is estimated to be 0.4, indicating sub-stoichiometric binding, possibly because a fraction of the protein was in an inactive conformation (Fig. 3B). In the absence of 3SHB, Dr3SHBDH was still able to bind NADH, with a K_d of 24 μM, ΔH (kcal/mol) = –70.1, and ΔG (kcal/mol) = −6.21. The N value is estimated to be 1.1, indicating stoichiometric binding (Fig. 3C). When 3RHB was used as the substrate instead of 3SHB, no interaction was observed even when NADH was included in the premix (Fig. 3D). These observations suggest a strict order of binding of NADH prior to the carboxylic acid substrate, and strict stereoselectivity for 3SHB (Fig. 3E).

FIG 3.

ITC assay and substrate binding model of Dr3SHBDH. Assays show (A) 500 μM 3SHB titrated to 50 μM Dr3SHBDH; (B) 500 μM 3SHB titrated to 50 μM Dr3SHBDH which was premixed with 0.2 mM NADH; (C) 500 μM NADH titrated to 50 μM Dr3SHBDH; and (D) 500 μM 3RHB titrated to 50 μM Dr3SHBDH premixed with 0.2 mM NADH. (E) Dr3SHBDH binds the nucleotide prior to 3SHB, and does not bind 3RHB.

Enzymes for 3SHB degradation in Clostridium pasteurianum.

A BLAST search of the UniProt database showed that Dr3SHBDH has close homologs in many strict anaerobic fermenting bacteria within the genus Clostridium. A literature search revealed that C. tyobutyricum NADP⁺-dependent (S)-3-hydroxycarboxylate dehydrogenase was first described in 1994 (31). The reported N-terminal 45-amino acid sequence (NCBI reference no. AAB33485.1) was found to be identical to the UniProt entry A0A4P8A4W5 (overall 53% identity to Dr3SHBDH), which had not been manually annotated. This confirms the presence of 3SHBDH in Clostridium species, although the metabolic function of this enzyme remains undetermined.

To investigate the metabolic function of 3SHBDH, we focused on the enzyme from the model bacterium Clostridium pasteurianum DSM 525 (Cp3SHBDH). Using the Enzyme Function Initiative Genome Neighborhood Tool (EFI-GNT) (33), we found that the 3SHBDH gene cluster in Clostridium species contains a short-chain fatty acid transporter (SCFAT) and a putative acetate CoA-transferase (YdiF) (34) (Fig. 4A), suggesting a pathway for 3SHB import, oxidation to acetoacetate, and conversion to acetoacetyl-CoA, which is an intermediate in the butyrate fermentation pathway in Clostridia (Fig. 4B). To test this hypothesis, the C. pasteurianum enzymes Cp3SHBDH, CpYdiF, and CpHbd were recombinantly produced and purified (Fig. S3), and their enzyme activities were measured with NADPH spectrophotometric assays.

FIG 4.

C. pasteurianum utilizes 3SHB as an energy and carbon source. (A) Putative enzymes involved in 3SHB metabolism in anaerobic bacteria; gene ID (UniProt accession number) is shown in the gene synteny map. (B) Proposed pathway for 3SHB degradation by C. pasteurianum. (C) Comparison of the growth of C. pasteurianum on 3RHB, 3SHB,and glucose. (D) SDS-PAGE analysis of C. pasteurianum grown on 3SHB and glucose. Source data are provided as supplemental data sets S1 to S4.

Cp3SHBDH catalyzed the NADP⁺-dependent oxidation of 3SHB, with a higher activity for NADP⁺ compared to NAD+ (Fig. S4A and B). Meanwhile, CpHbd catalyzed the NADPH-dependent reduction of acetoacetyl-CoA, with similar activities for NADP⁺ and NAD⁺ (Fig. S4C and D). Three distinct reactions catalyzed by CpYdiF were detected using NADPH coupled assays with Cp3SHBDH or CpHbd, and NADPH formation/consumption was detected from their UV absorption spectra. CpYdiF catalyzed the formation of acetoacetyl-CoA from acetoacetate and acetyl-CoA, as detected in a coupled assay with CpHbd (Fig. S5A). CpYdiF also catalyzed the reverse reaction, formation of acetoacetate from acetoacetyl-CoA and acetate, as detected in a coupled assay with Cp3SHBDH (Fig. S5B). In addition, CpYdiF also catalyzed the formation of (S)-3-hydroxybutyryl-CoA from with 3SHB and acetyl-CoA, as detected in a coupled assay with CpHbd (Fig. S5C).

To investigate the involvement of the aforementioned enzymes in 3SHB degradation, C. pasteurianum DSM 525 was cultivated anaerobically in a defined medium with glucose, 3RHB, or 3SHB as the sole carbon source. The growth of C. pasteurianum cells was robust in the presence of glucose or 3SHB, and a high optical density (at 600 nm, OD₆₀₀) was reached within 3 days (OD₆₀₀ of 1.8 and 1.0, respectively) (Fig. 4C). SDS-PAGE analysis revealed the induction of several prominent protein bands in 3SHB-grown cells (Fig. 4D). Through protein mass spectrometry, the predominant protein bands were identified as CpHbd, Cp3SHBDH, CpYdiF, and Cpthiolase respectively (corresponding to data sets S1 to S4). This is consistent with the proposed pathway for conversion of 3SHB to acetoacetyl-CoA, and suggests that acetoacetyl-CoA is further metabolized by reduction to 3SHB-CoA during butyrate formation or cleaved into acetyl-CoA during acetate formation (Fig. 4B).

Bioinformatics analysis of the 3HCDH family in Clostridia.

The realization that many 3-hydroxyacid dehydrogenases in the 3HCDH family might be erroneously annotated as 3-hydroxyacyl-CoA dehydrogenase led us to carry out a more comprehensive bioinformatics study to investigate the diversity of 3HCDH enzymes in Clostridia. Using the Enzyme Function Initiative Enzyme Similarity Tool (EFI-EST), we constructed a sequence similarity network (SSN) for 2,352 enzymes belonging to the 3HCDH family in bacteria from the class Clostridia, which includes C. pasteurianum and D. ruminis (Fig. 5A). The E value threshold was adjusted to 10⁻⁶⁶, which separates functionally related sequence clusters. The largest cluster, which includes CbHbd and CaHbd, corresponds to (S)-3-hydroxybutyryl-CoA dehydrogenases. One of the clusters contains DeSHBDH, CpSHBDH, and the recently discovered (S)-3-hydroxy-5-oxohexanoate dehydrogenase from Clostridium scatologenes (CsTfD, UniProt ID: A0A0E3GQA2), which is involved in phloroglucinol degradation (our unpublished data). Another cluster contains homologs of l-carnitine dehydrogenase form, which is involved in l-carnitine degradation.

FIG 5.

Bioinformatics analysis of 3HCDH enzymes in Clostridia. (A) Sequence similarity network of 2,352 3HCDH enzymes from bacteria in the class Clostridia, with an edge E value threshold of 10⁻⁶⁶. The SSN is colored as heat map, with node color reflecting the length of the C-terminal sequence in excess of CbHbd. (B) Overlay of the crystal structure of CbHbd in complex with acetoacetyl-CoA (yellow), with AlphaFold structures of Cp3HBDH (tan), CsTfD (forest green), and NtLcdH (cornflower blue), showing the predicted C-terminal helix occluding the CoA binding site.

A comparison of the crystal structure of CbHbd in complex with acetoacetyl-CoA (PDB ID: 4KUH) with the AlphaFold structures of Cp3HBDH, CsTfD, and Natranaerobius thermophilus LcdH (NtLcdH, UniProt ID: B2A0X9), obtained from UniProt, suggest that the three 3-hydroxyacid dehydrogenases contain a C-terminal helical extension that occludes CoA binding (Fig. 5B). This helix was also present in the crystal structure of human 3S-hydroxyacid dehydrogenase CRYL1 (PDB ID: 3F3S), which further validates the AlphaFold structures. The presence of this C-terminal helix could thus serve as a marker to distinguish 3-hydroxyacid dehydrogenases from 3-hydroxyacyl-CoA dehydrogenases.

To investigate the occurrence of this C-terminal extension for entries in the SSN, we constructed a multiple sequence alignment using Clustal Omega and then plotted the number of amino acids in excess of the CbHbd sequence for each entry as a heat map on the SSN (Fig. 5A). This analysis revealed that two of the clusters contain sequences lacking C-terminal extensions, suggesting that these are 3-hydroxyacyl-CoA dehydrogenases. Additionally, three of the clusters containing long (>60-amino acid) extensions are multi-domain fusion proteins. Of these, one is a tandem dimer and two are fused to enoyl-CoA hydratase domains (ECH_1, PF16113 or ECH_2, PF00378), suggesting that they are also 3-hydroxyacyl-CoA dehydrogenases. Most of the remaining sequences contain C-terminal extensions of 20 to 40 amino acids, which are predicted by AlphaFold to be helices occluding the CoA-binding site, suggesting that they are 3-hydroxyacid dehydrogenases. The genome neighborhoods of the entries in the SSN were examined using EFI-GNT, revealing that many of the putative (S)-3-hydroxyacid dehydrogenases are associated with enzymes in the beta-keto acid cleavage enzyme family (BKACE, PF05853) (Fig. 5A and Fig. S6A). BKACE enzymes catalyze the acetyl-CoA-dependent cleavage of beta-keto acids to form an acyl-CoA and acetoacetate (35) (Fig. S6B). This implies that the products of BKACE-associated (S)-3-hydroxyacid dehydrogenases are unlikely to be acetoacetate, and therefore their substrates are unlikely to be 3SHB, but rather other (S)-3-hydroxyacids that are not yet identified.

DISCUSSION

Although bacterial degradation of 3RHB in the context of PHB degradation is well studied, knowledge of the bacterial degradation of 3SHB is limited. Our studies revealed that in many Gram-positive fermenting bacteria, 3SHBDH is associated with a SCFAT and/or YdiF; and that in C. pasteurianum, these genes are involved in the degradation of 3SHB as a carbon and energy source through the butyrate fermentation pathway. In contrast, C. pasteurianum did not utilize 3RHB as a growth substrate (Fig. 4C). The ability of butyrate-forming fermenting bacteria to utilize 3SHB is noteworthy given the importance of 3SHB-CoA as an intermediate in the butyrate fermentation pathway. Biochemical assays suggest that 3SHB may be oxidized to acetoacetate prior to conversion to acetoacetyl-CoA by YdiF or, alternatively, that 3SHB may be directly converted to 3SHB-CoA by YdiF. Gene clusters containing 3SHBDH and SCFAT are present in several Clostridium and Sporomusa species isolated from various anaerobic environments, as well as Megasphaera species from animal sources and two strains from human sources, namely, Megasphaera sp. ASD88 from feces (36) and Megasphaera from the genital tract (37), indicating the relevance of 3SHB as a substrate in these environments (Fig. S7). However, close homologs of 3SHBDH in other fermenting bacteria and SSRB are not associated with SCFAT, and further research is required to understand their metabolic function in these bacteria (Fig. S7).

Our investigation of Dr3SHBDH and Cp3SHBDH has revealed that many enzymes annotated as (S)-3-hydroxyacyl-CoA dehydrogenases in sequence databases are actually (S)-3-hydroxyacid dehydrogenases. The enzymes studied are close homologs of the previously described (S)-3-hydroxyacid dehydrogenase from Clostridium tyrobutyricum (CtS3HBDH), for which the N-terminal sequence has been described. These enzymes exhibit a preference for NADP⁺ over NAD⁺ as the redox cofactor, and have a high degree of stereoselectivity for the (S)-3-hydroxyacid isomer, while tolerating substituents at other positions. For instance, Dr3SHBDH tolerates 2-amino groups of d-threonine and l-allo-threonine, while CtS3HBDH was reported to oxidize longer chain substrates such as 3-hydroxyvalerate, 3-hydroxycaproate, and 3-hydroxyisocaproate (31). These enzymes are also close homologs of TfD, which oxidizes (S)-3-hydroxy-5-oxohexanoate as part of the phloroglucinol degradation pathway (our unpublished data).

Comparison of crystal structures of 3SHB-CoA dehydrogenase with AlphaFold models of 3-hydroxyacid dehydrogenases in UniProt showed that the 3-hydroxyacid dehydrogenases contain an additional C-terminal helix predicted to occlude CoA binding. The same helix is present in the AlphaFold model of l-carnitine dehydrogenases, indicating that it may be a distinguishing feature between 3-hydroxyacid and 3-hydroxyacyl-CoA dehydrogenases. Our bioinformatics study of 3HCDH family enzymes in Clostridia suggests that many 3-hydroxyacid dehydrogenases have yet to be discovered, and further studies may reveal new enzymatic activities and biochemical pathways. Animal models suggest that 3SHB in plasma or urine could serve as a biomarker for diseases such as diabetes ketoacidosis (38) or kidney damage (39), and therefore 3SHBDH could have applications in medical diagnostic assays. In addition, the broad substrate range and strict stereoselectivity of these enzymes make them excellent candidates for various biotechnological applications, particularly in enzymatic synthesis of precursors for bioplastics and chiral feedstocks (40).

MATERIALS AND METHODS

General.

Lysogeny Broth (LB) medium was purchased from Oxoid Limited (Hampshire, United Kingdom). Acetonitrile was obtained from Concord Technology (MN, USA). TALON resins were purchased from Clontech Laboratories, Inc. (CA, USA). HIS*BIND RESIN (69670, 10 mL) was ordered from EMD Millipore Corp. (MA, USA). All protein purification chromatographic experiments were performed on an ÄKTA pure FPLC (fast protein liquid chromatography) system (GE Healthcare, IL, USA). 3SHB, 3RHB, l-threonine, d-threonine, l-allo-threonine, d-allo-threonine, 3-hydroxylpropionate, l-Ser, d-Ser, acetoacetate-Li⁺, acetyl-CoA, acetoacetyl-CoA, DL-β-hydroxybutyryl coenzyme A [(RS)-3HB-CoA], NADPH, and NADP⁺ were purchased from Sigma-Aldrich (MO, USA) or Solarbio (Beijing, China).

Gene syntheses, cloning, expression and purification of Dr3SHBDH, Cp3SHBDH, CpYdiF, and CpHbd.

The codon-optimized gene fragments of Dr3SHBDH (UniProt ID: F6DKL7), Cp3SHBDH (A0A0H3J7I0), CpYdiF (A0A0H3J9J8), and CpHbd (A0A0H3JAY2) were synthesized by General Biosystems Inc. (Anhui, China), and inserted into the pET-28a-HT vector at the SspI restriction site (41) using Gibson Assembly cloning protocol (42). Recombinant Dr3SHBDH was produced and purified as previously described protocol for Klebsiella oxytoca sulfoacetaldehyde reductase (41). Cp3SHBDH, CpHbd, and CpYdiF were purified by Ni-NTA (nitriloacetic acid) affinity chromatography. All proteins were examined on 10% SDS polyacrylamide gel, and the concentrations of purified proteins were determined at 280 nm using a NanoDrop One (Thermo Fisher Scientific, MA, USA).

Determination of the oligomeric state of Dr3SHBDH.

A 4 mL Dr3SHBDH (5 mg/mL) solution was injected into a Superdex 200 gel filtration column (~300 mL) and eluted over 150 min with buffer C at 2.0 mL/min. A solution of molecular weight markers (Sigma MWGF 1000-1KT) was analyzed under the same conditions. The molecular weight of Dr3SHBDH was calculated from its elution volume, using a second-degree polynomial for the relationship between log (molecular weight) and retention time.

Dr3SHBDH oxidation assays.

In a typical assay, 50 mM CAPSO buffer (pH 10.0), 1 mM 3SHB [3RHB, (RS)-3HB-CoA], and 1 mM NADP⁺ were premixed and 0.1 or 0.5 μg of Dr3SHBDH was added to initiate the reaction (total volume = 200 μL) in a 96-well plate. The A₃₄₀ was monitored using a Tecan M200 plate reader at 15-s intervals for 2 to 3 min at room temperature (RT). To measure the Michaelis-Menten kinetic constants, the concentration of the substrate was varied at a range of 0 to 0.3 mM for 3SHB (0 to 10 mM d-Thr or 3-hydroxylpropionate, 0 to 20 mM l-allo-Thr, 0 to 50 mM l-Ser or d-Ser) in the presence of a saturated concentration of 1 mM NAD⁺, and 0.1 to 7.5 μg of Dr3SHBDH was added to initiate the reaction. In another set of experiments, the concentration of 3SHB was fixed at 1 mM, while the NAD⁺ concentration varied from 0 to 1 mM, and the NADP⁺ concentration varied from 0 to 2 mM.

Dr3SHBDH reduction assays.

In a typical assay, 100 mM Tris-HCl (pH 7.5), 0.5 mM acetoacetate-Li⁺ (acetoacetyl-CoA), and 1 mM NADPH were premixed and 0.5 or 1 μg of Dr3SHBDH was added to initiate the reaction in a 96-well plate. The A₃₄₀ was monitored using a Tecan M200 plate reader at 15-s intervals for 2 to 3 min at RT. To measure the Michaelis-Menten kinetic constants, the concentration of the substrate was varied from 0 to 0.5 mM for acetoacetate-Li⁺ in the presence of a saturated concentration of 1 mM NADH. In another set of experiments, the concentration of acetoacetate-Li⁺ was fixed at 0.5 mM, while the NADH concentration was varied from 0 to 0.75 mM, and the NADPH concentration was varied from 0 to 0.4 mM. ΔA₃₄₀ and the ε₃₄₀ of NADPH (6,220 M⁻¹cm⁻¹) were used to calculate the rates of the reactions. GraphPad Prism version 6.0 was used to extract kinetic parameters.

MS and LC-MS assays for Dr3SHBDH.

A 400-μL reaction mixture containing 50 mM CAPSO (pH 10.0), 1 mM 3SHB or 3RHB, 1 mM NAD⁺, and 5 μg of Dr3SHBDH was incubated for 5 min at RT. A positive control containing 0.5 mM acetoacetate was also prepared. The reaction solution was mixed with the same volume of methanol to denature the enzyme after centrifugation (12,000 × g for 5 min at RT), then filtered prior to MS assay. MS analysis was carried out with a 10 μL sample volume and developed at a flow rate of 0.5 mL/min with methanol, performed using an Agilent 6420 Triple Quadrupole LC/MS instrument (Agilent Technologies, CA, USA). A decarboxylation product of acetoacetate, acetone, was detected by derivatization with 2,4-DNPH. A 200-μL reaction mixture containing 50 mM CAPSO (pH 10.0), 1 mM 3SHB or 3RHB, 1 mM NAD⁺, and 5 μg of Dr3SHBDH was incubated for 5 min at RT. One positive control, 0.5 mM acetoacetate, was also prepared. Derivatization and LC-MS analysis were conducted as described previously for sulfoacetaldehyde detection (41), except that the gradient was 25% to 85% CH₃CN (0.1% formic acid) for chromatography.

ITC assay of Dr3SHBDH.

Titrations consisted of 19 injections: a single 0.4-μL injection followed by 18 2-μL injections of 500 μM substrates into cells containing either 50 μM Dr3SHBDH (0.2 mM NADH) or blank buffer (20 mM Tris-HCl [pH 7.5], 0.1 M KCl, and 1 mM dithiothreitol). To make the heat capacity between two cells the same, a reference cell was filled with water. Experiments were performed at 20°C and a stirring speed of 750 rpm on a PEAQ-ITC instrument (Malvern Panalytical, United Kingdom). The data sets were analyzed using a single-site binding model using MicroCal PEAQ-ITC software. Fitting was performed to derive stoichiometry, binding affinity (K_d), and changes in enthalpy (ΔH) and entropy (ΔS).

Cp3SHBDH enzyme activity assays.

In a typical oxidation reaction assay, 50 mM Tris-HCl (pH 8.8), 2 mM 3SHB (or 3RHB), and 1 mM NADP⁺ were premixed, and 0.04 to 0.2 μM of Cp3SHBDH was added to initiate the reaction in a 96-well plate. The A₃₄₀ was monitored using a BioTek plate reader at 30-s intervals for 4 min at RT. In a typical reduction reaction assay, 50 mM Tris-HCl (pH 7.5), 1 mM acetoacetate (acetoacetyl-CoA), and 1 mM NADPH were premixed, and 0.1 μM of CpHbd was added to initiate the reaction in a 96-well plate. The A₃₄₀ was monitored using a BioTek plate reader at 30-s intervals for 4 min at RT.

CpYdiF enzyme activity assays.

In this study, Cp3SHBDH or CpHbd were used as coupling enzymes to detect CpYdiF activity. A 40-μL reaction mixture containing 50 mM Tris-HCl (pH 8.0), 1 mM acetoacetyl-CoA (or acetyl-CoA), 10 mM potassium acetate (or acetoacetate), 10 μM CpYdiF, 1 μM Cp3SHBDH (or CpHbd), and 1 mM NADH was incubated for 30 min at RT. Alternatively, a 40-μL reaction mixture containing 50 mM Tris-HCl (pH 8.0), 1 mM acetyl-CoA, 10 mM 3SHB, 4 μM CpYdiF, 1 μM CpHbd, and 1 mM NAD⁺ was incubated for 6 min at RT. The negative control was also prepared without substrates or CpYdiF. The absorbance from 190 to 850 nm was monitored using a NanoDrop One (Thermo Fisher Scientific).

Growth of Clostridium pasteurianum with 3SHB.

C. pasteurianum (DSM 525, ATCC 6013) was purchased from DSMZ (German Collection of Microorganisms and Cell Cultures GmbH), and cultured to investigate whether it can utilize 3SHB as a carbon and energy source. The rich growth medium for C. pasteurianum was prepared by dissolving 10 g beef extract, 15 g Casitone, 2.5 mL 1 N NaOH, 0.5 g yeast extract, 0.5 g K₂HPO₄, 50 μL 0.1% (wt/vol) Na-resazurin solution, 0.4 g d-glucose, 0.1 g cellobiose, 0.1 g maltose, 0.1 g starch (soluble), and 0.5 g/L l-cysteine hydrochloride in 100 mL distilled water and adjusting the pH to 7.0. The cells were initially inoculated into rich growth medium and cultured anaerobically at 37°C for 2 days. Next, they were centrifuged at 6,000 × g and resuspended in defined medium (43). A portion of each 100-μL starter culture was transferred into three anaerobic bottles containing 5.0 mL of defined medium. Carbon sources for cell cultures were also provided by 20 mM glucose, 3RHB, or 3SHB. After incubation at 37°C for 3 days, cultures of glucose or 3SHB became turbid, indicating bacterial growth.

Protein identification by SDS-PAGE and mass spectrometry.

Centrifugation was used to obtain cells, which were lysed by sonication and boiling in 1× protein loading buffer and then analyzed on a 12.5% SDS-PAGE gel. Prominent protein bands that were induced by growth on 3SHB were manually excised. The in-gel digested and extracted peptide mixtures were analyzed twice on an Orbitrap Fusion with EASY-nLC 1200 (Thermo Fisher Scientific), and these tandem mass spectra were searched against an appropriate protein database (retrieved from IMG) using Mascot (Matrix Science) and Proteome Discoverer version 1.3 (Thermo Fisher Scientific). Peptide hits were compared to GCF_001856695.1 from the C. pasteurianum (DSM 525) protein database by the Proteome Discoverer (version 1.3) searching algorithm. The Mowse scoring algorithm was applied to identify proteins with a 95% confidence level.

Data availability.

Source data underlying Fig. 1 to 5 are provided as a source data file. Other data are available from the corresponding authors upon reasonable request.