Purification , active site interactions and partial primary structure of serine hydroxymethyltransferase from seedlings of trigonella foenum graecum

Abstract

SHMT catalyzes the conversion of Ser to Gly by the transfer of a hydroxymethyl group to H?-folate to yield 5,10-CH?-H?-folate. This compound serves as a carrier of one-carbon units for the biosynthesis of a number of biologically important products, such as thymidylate, purines, Met, etc. (Fig. 1-4).

Recently, it was shown that SHMT also catalyzes the hydrolysis of 5,10-CH?-H?-folate in the presence of Gly to form 5-HCO-H?-folate, and it was suggested that this reaction was the source of cellular 5-HCO-H?-folate. The observation that 5-HCO-H?-folate inhibits many of the enzymes in one-carbon metabolism suggests that it may play a role as a regulator of one-carbon metabolism.

In highly proliferating tissues, the activity of SHMT was shown to be selectively retained to facilitate channeling of Ser towards nucleotide biosynthesis, while the activity of enzymes utilizing Ser for the biosynthesis of glucose was decreased (Fig. 6). Studies from our laboratory showed that SHMT is a regulatory protein, and these regulatory properties were lost in highly proliferating tissues. These observations led to the postulation that SHMT could be an alternate target for cancer chemotherapy.

In eukaryotic cells, SHMT exists in the cytosol as well as in mitochondria. In plants, it was also localized in plastids (Fig. •). It was suggested that isoforms of SHMT helped maintain separate pools of folate coenzymes in different organelles of the cell. The role of plant mitochondrial SHMT was shown to be in photorespiration. However, the function of plant cytosolic SHMT is not clearly understood.

SHMT was purified to homogeneity from several mammalian and bacterial sources (Table 1). The enzyme catalyzed a number of reactions, all of which required PLP (Table 2).

SHMT from plants was purified to homogeneity from mung bean seedlings, soybean root nodules, pea leaf mitochondria, and spinach leaf mitochondria. The mung bean enzyme required PLP as a stabilizing agent. It was recently shown that an N-terminal pyruvoyl moiety replaced PLP in this enzyme. The role of PLP in several plant SHMTs was not unambiguously established.

The understanding of the role of SHMT in cell proliferation and the details of the catalytic mechanism of the enzyme were obtained from studies carried out on mammalian and E. coli SHMTs. The regulation of the glyA gene coding for SHMT was studied in detail in E. coli and S. typhimurium.

The present investigation was carried out to understand the structure and function of a plant SHMT. Fenugreek, a legume, was chosen as the source of the enzyme. The objectives of the present study were: (a) To purify SHMT from fenugreek seedlings and study its physico-chemical properties. (b) To identify the coenzyme present and establish its role in catalysis. (c) To examine the amino acid residues essential for the reaction. (d) To compare the structural features of fenugreek SHMT with SHMTs from other sources.

The materials and methods used in this study, such as material and medium for plant tissue culture (Table 3), preparation of H?-folate, estimation of enzyme activity, subcellular fractionation by differential centrifugation, electrophoretic methods like native and SDS-PAGE, immunological methods like ELISA, Ouchterlony double immunodiffusion, Western blotting and dot blot assay, spectral measurements including visible absorbance, CD and fluorescence spectroscopy; preparation of apoenzyme; assay for the exchange of ?-proton of Gly; amino acid analysis; reduction and carboxymethylation of the protein; fragmentation of the protein by CNBr digestion and tryptic digestion; separation of peptides by HPLC and tricine SDS-PAGE; and sequencing of peptides using automated gas-phase sequenator, are described in Chapter II.

In order to study the role of plant SHMT in cell proliferation, plant tissue culture was used as a model system. Callus was obtained from seedling explants (Fig. 11). SHMT activity in callus paralleled the growth of callus (Fig. 12, Table 5). Attempts at obtaining a differentiated plantlet from callus were unsuccessful. Hence, a comparative study of SHMT between callus and differentiated tissue could not be carried out. Attempts at purifying SHMT from callus were also not successful due to the instability of the enzyme in the callus.

Most of the activity of SHMT in fenugreek seedlings was present in the cytosolic fraction (Table 6), similar to the localization observed in mung bean seedlings, while the enzyme was predominantly localized in the mitochondrial fraction in leaves. The purification of SHMT from the soluble fraction of 48 h germinated seedlings was standardized. The purification procedure involved homogenization of the seedlings, ammonium sulfate precipitation of the 40–80% crude extract, chromatography on CM Sephadex, Mono S, Mono Q ion exchange, and finally on Blue Sepharose, a group-specific affinity column (Table 7). The use of 2-ME, EDTA, glycerol, and PLP prevented the inactivation of the enzyme during purification.

The enzyme isolated by the above procedure was homogeneous as determined by PAGE (Fig. 17), SDS-PAGE (Fig. 16, 19), and gel filtration (Fig. 18). The enzyme was a homotetramer with a native Mr of 240,000 (Fig. 18) and a subunit Mr of 61,000 (Fig. 19). The antibodies raised against the purified SHMT were specific to the enzyme (Fig. 20, 22). An ?-helical content of 26.2% and 37% were calculated by an analysis of the far-UV-CD spectrum (Fig. 25). The fluorescence spectrum of the enzyme showed that it was a typical Trp-containing protein (Fig. 26).

The purified enzyme had a specific activity between 7–8 (Table 7), comparable to the enzyme isolated from E. coli, mammalian livers (Table 1), tobacco root nodules, and spinach leaf mitochondria (Table 9). The saturation pattern of the enzyme for Ser and H?-folate was hyperbolic (Fig. 23 & Fig. 24). Km values of 1.2 mM and 0.3 mM were obtained for Ser and H?-folate, respectively.

The visible absorbance (Fig. 27) and CD spectrum (Fig. 28) of the purified enzyme had maxima at 420 and 430 nm, respectively, characteristic of PLP-proteins. Reduction of the enzyme by NaBH? changed the spectral (Fig. 30) and fluorescent characteristics (Fig. 31) of the enzyme. Similar changes were observed when PLP-enzymes such as AATase were treated similarly.

Cys inhibited the enzyme (Fig. 33), and the spectra of the Cys-treated enzyme showed disappearance of the 420 nm peak corresponding to the PLP-Schiff base and appearance of a new peak at 330 nm corresponding to the thiazolidine complex (Fig. 32, 34).

LCS inhibited the enzyme more efficiently compared to DCS (Fig. 35A & 35B). OADS inhibited the enzyme completely at concentrations as low as 2.5 ?M (Fig. 37). The enzyme in the presence of OADS showed fluorescence properties similar to that observed for the free PLP-OADS mixture (Fig. 38). The inhibition caused by Cys, semicarbazide, DCS, LCS, and OADS was reversed by PLP (Table 10, 11). Similar observations with other enzymes have been interpreted to indicate the requirement of PLP for catalysis. Earlier studies showed that the mung bean enzyme, which contains an N-terminal pyruvoyl moiety instead of PLP, was inactivated by carbonyl reagents, but reactivation was not observed upon the addition of coenzyme.

The apoenzyme prepared by treatment of the holoenzyme with Cys was specifically reactivated by PLP in a concentration-dependent manner. Half-maximal activity was obtained at 4 ?M PLP concentration (Fig. 39).

The inactive apoenzyme was reactivated upon incubation with PMP and pyruvate, where PMP alone was ineffective (Table 12). The reactivation of the apoenzyme due to reconstitution to the holoenzyme was reflected in the appearance of the absorbance peak at 420 nm (Fig. 40). These observations indicated that fenugreek SHMT, like other mammalian and E. coli SHMT, was able to bring about the transamination of PMP with pyruvate to yield PLP, which then reactivated the apoenzyme.

Fenugreek SHMT interacted with Gly and H?-folate to form intermediates with characteristic spectral properties (Fig. 41). H?-folate caused accumulation of the quinonoid intermediate, reflected by a large increase in absorbance at 495 nm (Fig. 41) and also increased incorporation of ?-proton of Gly into solvent (Fig. 42). This observation indicated that the catalytic mechanism of fenugreek SHMT was similar to that proposed for mammalian SHMT (Fig. 8). The observations reported in Chapter IV clearly showed that fenugreek SHMT is a PLP protein.

The identification of the amino acid residues required for activity was examined by chemical modification studies. PG, a reagent specific for Arg residues, was found to inactivate the enzyme in a concentration- and time-dependent manner (Fig. 44). From a replot of pseudo first-order rate constants against log [PG] concentration (Fig. 45), an 'n' value of 0.82 was obtained, suggesting that modification of a single Arg residue per active site of the enzyme was sufficient to cause complete loss of activity.

The substrates Ser and H?-folate afforded almost complete protection against inactivation by PG (Table 14), suggesting that the Arg that was modified was probably present at the active site of the enzyme.

Spectra of the PG-modified as well as the native enzyme in the absence and presence of substrates indicated that H?-folate was unable to increase the concentration of the quinonoid intermediate in the PG-modified enzyme-Gly complex (Fig. 46, 47, 48). This observation was further supported by the inability of H?-folate to increase the exchange of ?-proton of Gly with solvent protons (Table 15).

It was recently shown that in sheep liver SHMT, Arg 269 and 462 were involved in the interaction of H?-folate with the enzyme. In order to identify the Arg protected by Ser against modification by PG, the enzyme was first modified with cold PG in the presence of Ser, followed by removal of both Ser and PG, and a second modification of the enzyme with [¹?C]-PG in the absence of PG. The [¹?C]-PG-modified enzyme was digested with trypsin, and the peptides were separated by HPLC on a Vydac C8 column (Fig. 49, 50). In the elution profile of the PG-modified enzyme, the size of the peak eluting at 41.3 min was increased significantly compared to the control, and the maximum amount of radioactivity was also recovered in the same peak fraction (Fig. 50B). The sequence of the peptide carrying [¹?C]-PG-labeled Arg could not be obtained due to the small amount of enzyme used for modification and loss of the peptide carrying radioactivity during attempts to purify it.

Comparison of the sequence around Arg-269 of sheep liver SHMT and other SHMTs, including fenugreek SHMT (obtained during primary structure determination, Fig. 64, 66), showed that Arg-269 was conserved in all the eukaryotic SHMTs, and the sequence around Arg-269 was fairly well conserved. However, this stretch corresponds to one of the insertion regions of eukaryotic SHMTs and is absent in SHMT from prokaryotes (Fig. 66, 67). In E. coli SHMT, Lys 193, 194, and 197 were speculated to be involved in binding carboxyl groups of H?-folate.

The modification of His with DEPC showed a concentration- and time-dependent inactivation of the enzyme (Fig. 51). From a replot of the first-order rate constants against log [DEPC] concentration (Fig. 52), an 'n' value of 1 was obtained, indicating that modification of a single His residue per active site of the enzyme was sufficient to cause complete loss of activity.

Substrates Ser and H?-folate protected the enzyme against inactivation by DEPC (Tables 17, 18), suggesting that the residue modified was probably at the active site of the enzyme.

The observations reported in this thesis (Chapter V) could be correlated to the model for the active site of human liver SHMT (Fig. 55) proposed from this laboratory based on earlier observations.

The amino acid composition of fenugreek SHMT showed a higher content of Gly, Ala, Pro, Lys, Phe, and Met residues compared to SHMT from other sources (Tables 19, 20).

Fenugreek SHMT migrated further than sheep liver SHMT when electrophoresed at pH 8.5 under non-denaturing conditions, suggesting that it might contain more negative charges on the surface (Fig. 68). The dimeric E. coli SHMT moved ahead of both the eukaryotic SHMTs (Fig. 68), suggesting that it could be more anionic.

Antiserum raised against fenugreek SHMT formed a clear precipitin line with fenugreek SHMT, but not with sheep or E. coli SHMTs (Fig. 70). However, dot blot assays showed that antisera raised against fenugreek SHMT cross-reacted strongly with fenugreek SHMT, and weak cross-reactivity was also observed with sheep liver SHMT. The reactivity with E. coli SHMT and other non-specific proteins was not discernible (Fig. 71A). Antiserum raised against sheep liver SHMT showed good cross-reactivity with homologous antigen, weak cross-reactivity with fenugreek SHMT, but no reactivity with E. coli SHMT (Fig. 71B). These observations indicated that SHMT from eukaryotic sources had some similarity in their structure but were immunologically distinct from the E. coli enzyme.

In order to compare the structural similarities of fenugreek SHMT with various SHMTs, attempts were made to obtain sequences of some peptides of fenugreek SHMT. The tryptic peptides were separated by HPLC (Fig. 57, 58A–H), and sequences of 13 peptides were determined (Fig. 59, 60, Table 21). The enzyme was cleaved with CNBr, and the CNBr-peptides were separated by tricine SDS-PAGE, electroblotted onto PVDF membrane (Fig. 62A, 62B), and sequenced (Fig. 63, Table 22).

Fenugreek SHMT sequence (130 residues) from 7 tryptic and 4 CNBr peptides were placed on the sheep liver SHMT sequence based on homology (Fig. 64). Comparison of the sequences of various SHMTs corresponding to 130 residues of fenugreek SHMT indicated extensive homology (Fig. 66, Table 23) and showed that 34 out of 130 residues were invariant in all the sequences (Fig. 66).

The complete sequences of SHMTs thus far determined were compared (Fig. 67, Table 24), and considerable homology in their sequence was detected, indicating that this protein is highly conserved.

The results presented in this thesis demonstrate that fenugreek SHMT is a PLP-dependent protein, its mechanism of catalysis is similar to other PLP-dependent enzymes, and its primary structure is highly conserved among SHMTs from various taxa.