Mechanism of interaction of carbonyl-directed reagents at the active site of sheep liver serine hydroxymethyltransferase

Abstract

SHMT catalyzes the first step in the pathway for the interconversion of folate coenzymes. In this reaction, Ser is converted to Gly and the hydroxymethyl group of Ser is transferred to Hfolate to yield 5,10CHHfolate. SHMT catalyzes a variety of reactions characteristic of PLP enzymes, such as transamination, racemization, decarboxylation, etc. Recently, it was shown that the enzyme-Gly complex catalyzes the conversion of 5,10CHHfolate to 5CHOHfolate, probably accounting for the high concentration of 5CHOHfolate found in tissues. This step occurs at a branching point in onecarbon metabolism, and the product serves as a source of onecarbon units for several metabolically important compounds such as thymidylate, purines, Met, etc.

The enzyme exhibits homotropic interactions with Hfolate and heterotropic effects with nicotinamide nucleotides. It was earlier demonstrated that regulation of the enzyme was absent in extracts from neoplastic tissues. A proteinaceous factor present in the tumour extracts abolished the cooperativity seen in the enzyme from normal tissues. A role for Ser metabolism in neoplastic tissues was suggested by the observation that in proliferating tissues the pathway was directed toward Ser biosynthesis, and the enzymes involved in degradation were selectively inactivated.

The alteration in the pathway of Ser metabolism, the location of the enzyme at a pivotal point in the pathway, the alteration in its substratesaturation patterns in neoplastic tissues, and its participation in the “thymidylate synthase cycle” suggested that SHMT could be an additional target for attack by cancer chemotherapeutic agents.

Slowbinding competitive inhibitors have proved to be very useful as drugs, since inhibition of target enzymes by such compounds cannot be reversed by accumulation of substrates. The interactions of aminooxy compounds with sheepliver SHMT indicated that these compounds could act, at least in part, as slowbinding inhibitors. The literature on slowbinding inhibitors, especially of PLP enzymes, has been briefly and selectively reviewed.

The objectives of the present investigation were: (a) To elucidate the mechanism of inhibition of the enzyme by MA, the smallest substituted derivative of hydroxylamine, and define the minimal structural requirements for the specific interaction of aminooxy compounds with SHMT; (b) To examine the inhibition of the enzyme by LCS, LSH, and DSH; (c) To study the interactions of TSC, SC, TCH, AG, and other substituted hydrazides with the enzyme, since no information is available on the interaction of these compounds with this enzyme. A systematic examination of carbonyldirected agents would assist in the rational design of inhibitors for this enzyme.

The second chapter of the thesis describes the materials and methods used in this study. Largescale purification of the enzyme (Table VI), estimation of enzyme activity, spectral measurements including CD, fluorescence and stoppedflow spectrophotometry, and separation techniques such as HPLC and Centricon filtration are described.

The purity of the enzyme used in this study was checked by PAGE, activity staining, and SDSPAGE (Figs. 12-14). The enzyme preparations had a specific activity ranging from 7.5-10 units (mol of HCHO/min/mg protein) (Table VI).

Earlier studies from the laboratory had established the mechanism of interaction of aminooxy compounds OADS, AAA, hydroxylamine, and canaline with SHMT. MA, the smallest substituted aminooxy compound, was a reversible noncompetitive inhibitor of SHMT (Ki = 25 M) when Ser was the varied substrate (Figs. 15-18). PLP reversed the inhibition by MA (Table VIII). Based on the experimental observations reported in this thesis, a minimal kinetic mechanism was proposed for the interaction of MA with sheepliver SHMT (Fig. 33). The first step in the kinetic mechanism was the formation of an intermediate absorbing at 388nm generated by disruption of the internal aldimine absorbing at 425nm (Fig. 19). This step was very rapid and could not be monitored by conventional spectrophotometry (see insets in Figs. 19 and 20). Therefore, stoppedflow methods were used, and a rate constant of 1.2 × 10² M¹ s¹ was estimated (Figs. 26 and 27; Table IX). Timedifference spectra (Figs. 28-30) established that no intermediate absorbing at 425nm preceded formation of the 388nm species. The intermediate was generated at the enzyme active site (Fig. 20 inset B) and remained enzymebound (Fig. 25). It was converted in a unimolecular step to the final product (the oxime) with a rate constant of ~1 × 10³ s¹ (Table X). The formation of PLP-MA oxime was confirmed by fluorescence spectroscopy (Fig. 22), HPLC (Fig. 23), and studies with model compounds (Figs. 31 and 32).

The mechanism of MA interaction with SHMT was similar to that of OADS and AAA. However, AAA was nearly three orders of magnitude faster than MA in the initial phase of interaction (Table XI). These results highlighted the importance of the carboxyl group in enhancing the reactivity of substituted aminooxy compounds.

Preliminary examination of LCS interaction with the enzyme showed that LCS inhibited the enzyme in a time and concentrationdependent manner (Table XIII). No spectrally detectable intermediate was observed (Fig. 36), nor was an intermediate detected by CD studies (Fig. 37). The data indicated an inhibition mechanism similar to DCS.

Hydrazine and its derivatives showed concentrationdependent inhibition of enzyme activity (Fig. 35). Spectral studies showed a biphasic decrease at 425nm. Initial rapid phases were captured by stoppedflow (Tables XIV and XV). LSH was more reactive than DSH.

TSC, a systemic convulsant, is known to inhibit PLP enzymes by forming PLPthiosemicarbazone and the apoenzyme. No detailed study existed on its interaction with SHMT; therefore a full investigation was undertaken.

TSC inhibited SHMT in a timedependent, biphasic manner (Figs. 41 and 43). The initial phase was apparently noncompetitive (Ki = 50 M; Fig. 44). Spectral studies (Figs. 45-54) indicated formation of a novel intermediate absorbing at 464 and 440nm, the first report of such species for TSC-PLP interactions. This intermediate formed at the active site (Fig. 55) and remained bound (Fig. 57). The initial phase proceeded with a secondorder rate constant of 11 M¹ s¹ (Table XX) and was reversible (koff 5 × 10 s¹) (Fig. 58). PLP caused timedependent reactivation (Table XIX). The intermediate slowly converted to PLP-thiosemicarbazone (k 4 × 10 s¹) (Figs. 60 and 61). Reaction with HCHO suggested a resonancestabilized structure (Fig. 62).

A minimal kinetic mechanism for TSC interaction with SHMT is proposed (Fig. 64). Structure-activity relationships were examined using related compounds.

SC was a potent inhibitor (Tables XXII and XXV). Its interaction produced a shortlived intermediate absorbing at 400nm (Fig. 67 A, B), which rapidly converted to PLP-semicarbazone.

AG was a poor inhibitor (Table XXXI), reacting very sluggishly (Figs. 72 and 73) to form the PLP-hydrazone product (Figs. 72 and 75).

2MTSC was also a poor inhibitor (Table XXXI), forming 2Mthiosemicarbazone of PLP (Figs. 72, 73, 75).

4MTSC interacted similarly to TSC but less effectively (Fig. 66). Formation of the 464/440nm intermediate occurred with a secondorder rate constant of ~3M¹s¹ (Table XXXII) and koff of 4 × 10s¹ (Fig. 76 B).

TCTC was the most potent inhibitor studied (Table XXVI) with an apparent Ki of 3µM (Fig. 65). An intermediate absorbing at 464/440nm was formed (Figs. 68 and 69).

Substitutions that interfere with formation of a resonancestabilized structure involving the 2N and 3thioketone positions reduced inhibitory potency.

The unique interaction of TSC with SHMT is explained by the minimal mechanism (Fig. 64). A slowbinding step yields a resonancestabilized E-TSC-PLP intermediate (Fig. 79). The intermediate slowly yields PLP-thiosemicarbazone; a faster alternate pathway also exists.