Racemases and epimerases operating through a 1,1-proton transfer mechanism: reactivity, mechanism and inhibition.

Racemases and epimerases catalyse changes in the stereochemical configurations of chiral centres and are of interest as model enzymes and as biotechnological tools. They also occupy pivotal positions within metabolic pathways and, hence, many of them are important drug targets. This review summarises the catalytic mechanisms of PLP-dependent, enolase family and cofactor-independent racemases and epimerases operating by a deprotonation/reprotonation (1,1-proton transfer) mechanism and methods for measuring their catalytic activity. Strategies for inhibiting these enzymes are reviewed, as are specific examples of inhibitors. Rational design of inhibitors based on substrates has been extensively explored but there is considerable scope for development of transition-state mimics and covalent inhibitors and for the identification of inhibitors by high-throughput, fragment and virtual screening approaches. The increasing availability of enzyme structures obtained using X-ray crystallography will facilitate development of inhibitors by rational design and fragment screening, whilst protein models will facilitate development of transition-state mimics.

Urinary l-erythro-ß-hydroxyasparagine-a novel serine racemase inhibitor and substrate of the Zn2+-dependent d-serine dehydratase.

In the present study, we identified l-erythro-ß-hydroxyasparagine (l-ß-EHAsn) found abundantly in human urine, as a novel substrate of Zn2+-dependent d-serine dehydratase (DSD). l-ß-EHAsn is an atypical amino acid present in large amounts in urine but rarely detected in serum or most organs/tissues examined. Quantitative analyses of urinary l-ß-EHAsn in young healthy volunteers revealed significant correlation between urinary l-ß-EHAsn concentration and creatinine level. Further, for in-depth analyses of l-ß-EHAsn, we developed a simple three-step synthetic method using trans-epoxysuccinic acid as the starting substance. In addition, our research revealed a strong inhibitory effect of l-ß-EHAsn on mammalian serine racemase, responsible for producing d-serine, a co-agonist of the N-methyl-d-aspartate (NMDA) receptor involved in glutamatergic neurotransmission.

Human serine racemase is inhibited by glyceraldehyde 3-phosphate, but not by glyceraldehyde 3-phosphate dehydrogenase.

Murine serine racemase (SR), the enzyme responsible for the biosynthesis of the neuromodulator d-serine, was reported to form a complex with glyceraldehyde 3-phosphate dehydrogenase (GAPDH), resulting in SR inhibition. In this work, we investigated the interaction between the two human orthologues. We were not able to observe neither the inhibition nor the formation of the SR-GAPDH complex. Rather, hSR is inhibited by the hGAPDH substrate glyceraldehyde 3-phosphate (G3P) in a time- and concentration-dependent fashion, likely through a covalent reaction of the aldehyde functional group. The inhibition was similar for the two G3P enantiomers but it was not observed for structurally similar aldehydes. We ruled out a mechanism of inhibition based on the competition with either pyridoxal phosphate (PLP) - described for other PLP-dependent enzymes when incubated with small aldehydes - or ATP. Nevertheless, the inhibition time course was affected by the presence of hSR allosteric and orthosteric ligands, suggesting a conformation-dependence of the reaction.

Potent Inhibition of Mandelate Racemase by Boronic Acids: Boron as a Mimic of a Carbon Acid Center.

Boronic acids have been successfully employed as inhibitors of hydrolytic enzymes. Typically, an enzymatic nucleophile catalyzing hydrolysis adds to the electrophilic boron atom forming a tetrahedral species that mimics the intermediate(s)/transition state(s) for the hydrolysis reaction. We show that para-substituted phenylboronic acids (PBAs) are potent competitive inhibitors of mandelate racemase (MR), an enzyme that catalyzes a 1,1-proton transfer rather than a hydrolysis reaction. The Ki value for PBA was 1.8 ± 0.1 µM, and p-Cl-PBA exhibited the most potent inhibition (Ki = 81 ± 4 nM), exceeding the binding affinity of the substrate by ∼4 orders of magnitude. Isothermal titration calorimetric studies with the wild-type, K166M, and H297N MR variants indicated that, of the two Brønsted acid-base catalysts Lys 166 and His 297, the former made the greater contribution to inhibitor binding. The X-ray crystal structure of the MR·PBA complex revealed the presence of multiple H-bonds between the boronic acid hydroxyl groups and the side chains of active site residues, as well as formation of a His 297 Nε2-B dative bond. The dramatic upfield change in chemical shift of 27.2 ppm in the solution-phase 11B nuclear magnetic resonance spectrum accompanying binding of PBA by MR was consistent with an sp3-hybridized boron, which was also supported by density-functional theory calculations. These unprecedented findings suggest that, beyond substituting boron at carbon centers participating in hydrolysis reactions, substitution of boron at the acidic carbon center of a substrate furnishes a new approach for generating inhibitors of enzymes catalyzing the deprotonation of carbon acid substrates.

Accelerated identification of serine racemase inhibitor from Centella asiatica.

Serine racemase (SR) converts the free form of L-serine into D-serine (DS) in the mammalian brain. The DS functions as a co-agonist of N-methyl D-aspartate (NMDA) receptor. The over- activation of NMDA receptor leads to many neurological disorders like stroke, amyotrophic lateral sclerosis, Alzheimer's disease and an effective inhibitor of SR could be a corrective method for the receptor over-activation. We report for the first time here a rapid way of purifying and identifying an inhibitor from medicinal plants known to have the neuro-protective effect. We have purified SR inhibitor from the methanolic extract of Centella asiatica by affinity method. High resolution mass spectrometry and infrared spectroscopy were used to identify the ligand to be madecassoside. We have shown the madecassoside binding in silico and its inhibition of recombinant human serine racemase in vitro and ex vivo.

Substrate-product analogue inhibitors of isoleucine 2-epimerase from Lactobacillus buchneri by rational design.

A rational approach that may be applied to a broad class of enzyme-catalyzed reactions to design enzyme inhibitors affords a powerful strategy, facilitating the development of drugs and/or molecular probes of enzyme mechanisms. A strategy for the development of substrate-product analogues (SPAs) as inhibitors of racemases and epimerases is elaborated using isoleucine 2-epimerase from Lactobacillus buchneri (LbIleE) as a model enzyme. LbIleE catalyzes the PLP-dependent, reversible, racemization or epimerization of nonpolar amino acids at the C-2 position. The enzyme plays an important role in the biosynthesis of branched-chain d-amino acids and is a potential target for the development of antimicrobial agents. 3-Ethyl-3-methyl-l-norvaline (Ki = 2.9 ± 0.2 mM) and 3-ethyl-3-methyl-d-norvaline (Ki = 1.5 ± 0.2 mM) were designed as SPAs based on the movement of the sec-butyl side chain of the substrate l-Ile during catalysis, and were competitive inhibitors with binding affinities exceeding that of l-Ile by 1.3- and 2.5-fold, respectively. Surprisingly, these compounds were not substrates, but the corresponding compounds lacking the 3-methyl group were substrates. Unlike serine, glutamate, and proline racemases, which exhibit stringent steric requirements at their active sites, the active site of LbIleE was amenable to binding bulky SPAs. Moreover, LbIleE bound the SPA 2,2-di-n-butylglycine (Ki = 11.0 ± 0.2 mM) as a competitive inhibitor, indicating that the hydrophobic binding pocket at the active site was sufficiently plastic to tolerate gem-dialkyl substitution at the α-carbon of an amino acid. Overall, these results reveal that amino acid racemases/epimerases are amenable to inhibition by SPAs provided that they possess a capacious active site.

Novel 2-arylthiopropanoyl-CoA inhibitors of α-methylacyl-CoA racemase 1A (AMACR; P504S) as potential anti-prostate cancer agents.

α-Methylacyl-CoA racemase (AMACR; P504S) catalyses an essential step in the degradation of branched-chain fatty acids and the activation of ibuprofen and related drugs. AMACR has gained much attention as a drug target and biomarker, since it is found at elevated levels in prostate cancer and several other cancers. Herein, we report the synthesis of 2-(phenylthio)propanoyl-CoA derivatives which provided potent AMACR inhibitory activity (IC50 = 22-100 nM), as measured by the AMACR colorimetric activity assay. Inhibitor potency positively correlates with calculated logP, although 2-(3-benzyloxyphenylthio)propanoyl-CoA and 2-(4-(2-methylpropoxy)phenylthio)propanoyl-CoA were more potent than predicted by this parameter. Subsequently, carboxylic acid precursors were evaluated against androgen-dependent LnCaP prostate cancer cells and androgen-independent Du145 and PC3 prostate cancer cells using the MTS assay. All tested precursor acids showed inhibitory activity against LnCaP, Du145 and PC3 cells at 500 µM, but lacked activity at 100 µM. This is the first extensive structure-activity relationship study on the influence of side-chain interactions on the potency of novel rationally designed AMACR inhibitors.

Identification of novel small-molecule inhibitors of α-methylacyl-CoA racemase (AMACR; P504S) and structure-activity relationships.

α-Methylacyl-CoA racemase (AMACR; P504S; EC 5.1.99.4) catalyses epimerization of 2-methylacyl-CoAs and is important for the degradation of branched-chain fatty acids and the pharmacological activation of ibuprofen and related drugs. It is also a novel drug target for prostate and other cancers. However, development of AMACR as a drug target has been hampered by the difficulties in assaying enzyme activity. Consequently, reported inhibitors have been rationally designed acyl-CoA esters, which are delivered as their carboxylate prodrugs. The novel colorimetric assay for AMACR based on the elimination of 2,4-dinitrophenolate was developed for high-throughput screening and 20,387 'drug-like compounds' were screened, with a throughput of 768 compounds assayed per day. Pyrazoloquinolines and pyrazolopyrimidines were identified as novel scaffolds and investigated as AMACR inhibitors. The most potent inhibitors have IC50 values of ~2 µM. The pyrazoloquinoline inhibitor 10a displayed uncompetitive inhibition, whilst 10j displayed mixed competitive inhibition. The pyrazolopyrimidine inhibitor 11k displayed uncompetitive inhibition. This is the first report of the identification of specific drug-like small-molecule AMACR inhibitors by high-throughput screening. Pyrazoloquinolines and pyrazolopyrimidines may also be useful as inhibitors of other CoA-utilizing enzymes.

Catalytic mechanism and properties of pyridoxal 5'-phosphate independent racemases: how enzymes alter mismatched acidity and basicity.

Covering: up to March 2019 Amino acid racemases and epimerases are key enzymes that invert the configuration of common amino acids and supply many corresponding d-isomers in living organisms. Some d-amino acids are inherently bioactive, whereas others are building blocks for important biomolecules, for example lipid II, the bacterial cell wall precursor. Peptides containing them have enhanced proteolytic stability and can act as important recognition elements in mammalian systems. Selective inhibition of certain amino acid racemases (e.g. glutamate racemase) is believed to offer a promising target for new antibacterial drugs effective against pathogens resistant to current antibiotics. Many amino acid racemases employ imine formation with pyridoxal phosphate (PLP) as a cofactor to accelerate the abstraction of the alpha proton. However, the group reviewed herein achieves racemization of free amino acids without the use of cofactors or metals, and uses a thiol/thiolate pair for deprotonation and reprotonation. All bacteria and higher plants contain such enzymes, for example diaminopimelate epimerase, which is required for lysine biosynthesis in these organisms. This process cannot be accomplished without an enzyme catalyst as the acidities of a thiol and the substrate α-hydrogen are inherently mismatched by at least 10 orders of magnitude. This review describes the structural and mechanistic studies on PLP-independent racemases and the evolving view of key enzymatic machinery that accomplishes these remarkable transformations.

Identification and characterization of the 4-epimerase AglW from the archaeon Methanococcus maripaludis.

Archaea are ubiquitous single-cell microorganisms that have often adapted to harsh conditions and play important roles in biogeochemical cycles with potential applications in biotechnology. Methanococcus maripaludis, a methane-producing archaeon, is motile through multiple archaella on its cell surface. The major structural proteins (archaellins) of the archaellum are glycoproteins, modified with N-linked tetrasaccharides that are essential for the proper assembly and function of archaella. The aglW gene, encoding the putative 4-epimerase AglW, plays a key role in the synthesis of the tetrasaccharide. The goal of our work was to biochemically demonstrate the 4-epimerase activity of AglW, and to develop assays to determine its substrate specificity and properties. We carried out assays using UDP-Galactose, UDP-Glucose, UDP-N-acetylglucosamine, UDP-N-acetylgalactosamine and N-acetylglucosamine/N-acetylgalactosamine-diphosphate - lipid as substrates, coupled with specific glycosyltransferases. We showed that AglW has a broad specificity towards UDP-sugars and that Tyr151 within a conserved YxxxK sequon is essential for the 4-epimerase function of AglW. The glycosyltransferase-coupled assays are generally useful for the identification and specificity studies of novel 4-epimerases.

Structure-activity relationships of rationally designed AMACR 1A inhibitors.

α-Methylacyl-CoA racemase (AMACR; P504S) is a promising novel drug target for prostate and other cancers. Assaying enzyme activity is difficult due to the reversibility of the 'racemisation' reaction and the difficulties in the separation of epimeric products; consequently few inhibitors have been described and no structure-activity relationship study has been performed. This paper describes the first structure-activity relationship study, in which a series of 23 known and potential rational AMACR inhibitors were evaluated. AMACR was potently inhibited (IC50 = 400-750 nM) by ibuprofenoyl-CoA and derivatives. Potency was positively correlated with inhibitor lipophilicity. AMACR was also inhibited by straight-chain and branched-chain acyl-CoA esters, with potency positively correlating with inhibitor lipophilicity. 2-Methyldecanoyl-CoAs were ca. 3-fold more potent inhibitors than decanoyl-CoA, demonstrating the importance of the 2-methyl group for effective inhibition. Elimination substrates and compounds with modified acyl-CoA cores were also investigated, and shown to be potent inhibitors. These results are the first to demonstrate structure-activity relationships of rational AMACR inhibitors and that potency can be predicted by acyl-CoA lipophilicity. The study also demonstrates the utility of the colorimetric assay for thorough inhibitor characterisation.

Potent dialkyl substrate-product analogue inhibitors and inactivators of α-methylacyl-coenzyme A racemase from Mycobacterium tuberculosis by rational design.

Rational approaches for the design of enzyme inhibitors furnish powerful strategies for developing pharmaceutical agents and tools for probing biological mechanisms. A new strategy for the development of gem-disubstituted substrate-product analogues as inhibitors of racemases and epimerases is elaborated using α-methylacyl-coenzyme A racemase from Mycobacterium tuberculosis (MtMCR) as a model enzyme. MtMCR catalyzes the epimerization at C2 of acyl-CoA substrates, a key step in the metabolism of branched-chain fatty acids. Moreover, the human enzyme is a potential target for the development of therapeutic agents directed against prostate cancer. We show that rationally designed, N,N-dialkylcarbamoyl-CoA substrate-product analogues inactivate MtMCR. Binding greatly exceeds that of the substrate, (S)-ibuprofenoyl-CoA, up to ∼250-fold and is proportional to the alkyl chain length (4-12 carbons) with the N,N-didecyl and N,N-didodecyl species having competitive inhibition constants with values of 1.9 ±â€¯0.2 µM and 0.42 ±â€¯0.04 µM, respectively. The presence of two decyl chains enhanced binding over a single decyl chain by ∼204-fold. Overall, the results reveal that gem-disubstituted substrate-product analogues can yield extremely potent inhibitors of an epimerase with a capacious active site.

Human serine racemase is nitrosylated at multiple sites.

Serine racemase is a pyridoxal 5'­phosphate dependent enzyme responsible for the synthesis of d­serine, a neuromodulator of the NMDA receptors. Its activity is modulated by several ligands, including ATP, divalent cations and protein interactors. The murine orthologue is inhibited by S-nitrosylation at Cys113, a residue adjacent to the ATP binding site. We found that the time course of inhibition of human serine racemase by S-nitrosylation is markedly biphasic, with a fast phase associated with the reaction of Cys113. Unlike the murine enzyme, two additional cysteine residues, Cys269, unique to the human orthologue, and Cys128 were also recognized as S-nitrosylation sites through mass spectrometry and site-directed mutagenesis. The effect of S-nitrosylation on the fluorescence of tryptophan residues and on that of the pyridoxal phosphate cofactor indicated that S-nitrosylation produces a partial interruption of the cross-talk between the ATP binding site and the active site. Overall, it appears that the inhibition results from a conformational change rather than the direct displacement of ATP.

Design, synthesis, and evaluation of novel inhibitors for wild-type human serine racemase.

Most of the endogenous free d-serine (about 90%) in the brain is produced by serine racemase (SR). d-Serine in the brain is involved in neurodegenerative disorders and epileptic states as an endogenous co-agonist of the NMDA-type glutamate receptor. Thus, SR inhibitors are expected to be novel therapeutic candidates for the treatment of these disorders. In this study, we solved the crystal structure of wild-type SR, and tried to identify a new inhibitor of SR by in silico screening using the structural information. As a result, we identified two hit compounds by their in vitro evaluations using wild-type SR. Based on the structure of the more potent hit compound 1, we synthesized 15 derivatives and evaluated their inhibitory activities against wild-type SR. Among them, the compound 9C showed relatively high inhibitory potency for wild-type SR. Compound 9C was a more potent inhibitor than compound 24, which was synthesized by our group based upon the structural information of the mutant-type SR.

Human serine racemase structure/activity relationship studies provide mechanistic insight and point to position 84 as a hot spot for ß-elimination function.

There is currently great interest in human serine racemase, the enzyme responsible for producing the NMDA co-agonist d-serine. Reported correlation of d-serine levels with disorders including Alzheimer's disease, ALS, and ischemic brain damage (elevated d-serine) and schizophrenia (reduced d-serine) has further piqued this interest. Reported here is a structure/activity relationship study of position Ser84, the putative re-face base. In the most extreme case of functional reprogramming, the S84D mutant displays a dramatic reversal of ß-elimination substrate specificity in favor of l-serine over the normally preferred l-serine-O-sulfate (∼1200-fold change in kcat/Km ratios) and l (l-THA; ∼5000-fold change in kcat/Km ratios) alternative substrates. On the other hand, the S84T (which performs l-Ser racemization activity), S84A (good kcat but high Km for l-THA elimination), and S84N mutants (nearly WT efficiency for l-Ser elimination) displayed intermediate activity, all showing a preference for the anionic substrates, but generally attenuated compared with the native enzyme. Inhibition studies with l-erythro-ß-hydroxyaspartate follow this trend, with both WT serine racemase and the S84N mutant being competitively inhibited, with Ki = 31 ± 1.5 µm and 1.5 ± 0.1 mm, respectively, and the S84D being inert to inhibition. Computational modeling pointed to a key role for residue Arg-135 in binding and properly positioning the l-THA and l-serine-O-sulfate substrates and the l-erythro-ß-hydroxyaspartate inhibitor. Examination of available sequence data suggests that Arg-135 may have originated for l-THA-like ß-elimination function in earlier evolutionary variants, and examination of available structural data suggests that a Ser84-H2O-Lys114 hydrogen-bonding network in human serine racemase lowers the pKa of the Ser84re-face base.

A novel serine racemase inhibitor suppresses neuronal over-activation in vivo.

Serine racemase (SRR) is an enzyme that produces d-serine from l-serine. d-Serine acts as an endogenous coagonist of NMDA-type glutamate receptors (NMDARs), which regulate many physiological functions. Over-activation of NMDARs induces excitotoxicity, which is observed in many neurodegenerative disorders and epilepsy states. In our previous works on the generation of SRR gene knockout (Srr-KO) mice and its protective effects against NMDA- and Aß peptide-induced neurodegeneration, we hypothesized that the regulation of NMDARs' over-activation by inhibition of SRR activity is one such therapeutic strategy to combat these disease states. In the previous study, we performed in silico screening to identify four compounds with inhibitory activities against recombinant SRR. Here, we synthesized 21 derivatives of candidate 1, one of four hit compounds, and performed screening by in vitro evaluations. The derivative 13J showed a significantly lower IC50 value in vitro, and suppressed neuronal over-activation in vivo.

A novel colorimetric assay for α-methylacyl-CoA racemase 1A (AMACR; P504S) utilizing the elimination of 2,4-dinitrophenolate.

α-Methylacyl-CoA racemase (AMACR; P504S) regulates branched-chain fatty acid degradation, activates Ibuprofen and is a recognised cancer drug target. A novel, facile colorimetric assay was developed based on elimination of 2,4-dinitrophenolate. The assay was used to test 5 known inhibitors, determining IC50 and Ki values, reversibility and characterizing irreversible inhibition.

Differential regulation of NMDA receptors by d-serine and glycine in mammalian spinal locomotor networks.

Activation of N-methyl-d-aspartate receptors (NMDARs) requires the binding of a coagonist, either d-serine or glycine, in addition to glutamate. Changes in occupancy of the coagonist binding site are proposed to modulate neural networks including those controlling swimming in frog tadpoles. Here, we characterize regulation of the NMDAR coagonist binding site in mammalian spinal locomotor networks. Blockade of NMDARs by d(-)-2-amino-5-phosphonopentanoic acid (d-APV) or 5,7-dichlorokynurenic acid reduced the frequency and amplitude of pharmacologically induced locomotor-related activity recorded from the ventral roots of spinal-cord preparations from neonatal mice. Furthermore, d-APV abolished synchronous activity induced by blockade of inhibitory transmission. These results demonstrate an important role for NMDARs in murine locomotor networks. Bath-applied d-serine enhanced the frequency of locomotor-related but not disinhibited bursting, indicating that coagonist binding sites are saturated during the latter but not the former mode of activity. Depletion of endogenous d-serine by d-amino acid oxidase or the serine-racemase inhibitor erythro-ß-hydroxy-l-aspartic acid (HOAsp) increased the frequency of locomotor-related activity, whereas application of l-serine to enhance endogenous d-serine synthesis reduced burst frequency, suggesting a requirement for d-serine at a subset of synapses onto inhibitory interneurons. Consistent with this, HOAsp was ineffective during disinhibited activity. Bath-applied glycine (1-100 µM) failed to alter locomotor-related activity, whereas ALX 5407, a selective inhibitor of glycine transporter-1 (GlyT1), enhanced burst frequency, supporting a role for GlyT1 in NMDAR regulation. Together these findings indicate activity-dependent and synapse-specific regulation of the coagonist binding site within spinal locomotor networks, illustrating the importance of NMDAR regulation in shaping motor output.NEW & NOTEWORTHY We provide evidence that NMDARs within murine spinal locomotor networks determine the frequency and amplitude of ongoing locomotor-related activity in vitro and that NMDARs are regulated by d-serine and glycine in a synapse-specific and activity-dependent manner. In addition, glycine transporter-1 is shown to be an important regulator of NMDARs during locomotor-related activity. These results show how excitatory transmission can be tuned to diversify the output repertoire of spinal locomotor networks in mammals.

Serine racemase inhibition induces nitric oxide-mediated neurovascular protection during cerebral ischemia.

There are no effective neuroprotectant drugs for acute cerebral ischemia. Serine racemase (SR) synthesizes d-serine, which is involved in N-methyl-d-aspartate (NMDA) receptor-induced neurotoxicity. Recently, SR deletion was reported to protect against focal cerebral ischemia. However, regulatory mechanisms controlling SR-activity in the neurovascular unit (NVU) during cerebral ischemia remain to be clarified. We investigated the effects of SR inhibition on neurovascular protection after ischemia. The SR inhibitor phenazine methosulfate (PMS) alleviated neuronal damage in an ex vivo ischemic model (oxygen glucose deprivation [OGD]) using primary neuronal cultures, and in an in vivo mouse model of ischemia (middle cerebral artery occlusion [MCAO]). Ischemic preconditioning (IP) and PMS-treatment inhibited SR phosphorylation after ischemia ex vivo. In addition, SR phosphorylation after MCAO was also decreased in PMS-treated mice. Reductions in regional cerebral blood flow (CBF) after MCAO were improved by administration of PMS. Treatment with PMS increased phosphorylation of endothelial nitric oxide synthase (eNOS) in the ischemic core and penumbra region. In neuron-endothelial cell co-cultures, PMS promoted nitric oxide production after OGD. These findings indicate that SR inhibition acts as a neuroprotectant in the NVU and ameliorant of CBF abnormalities post-stroke. Thus, pharmacologic SR inhibition has potential clinical applications.

Rational design and synthesis of substrate-product analogue inhibitors of α-methylacyl-coenzyme A racemase from Mycobacterium tuberculosis.

2,2-Bis(4-isobutylphenyl)propanoyl-CoA and 2,2-bis(4-t-butylphenyl)propanoyl-CoA are rationally designed, gem-disubstituted substrate-product analogues that competitively inhibit α-methylacyl-coenzyme A racemase from Mycobacterium tuberculosis with Ki values of 16.9 ± 0.6 and 21 ± 4 µM, respectively, exceeding the enzyme's affinity for the substrate by approximately 5-fold.