The Reversible unfolding of the Tetrameric serine hydroxymethyltransferase

Abstract

(1) The discovery, just over 3 decades ago by Anfinsen and coworkers, that proteins can be refolded under in vitro conditions, suggested that the information necessary for proper folding resides in the primary sequence of the polypeptide chain. This classical work on ribonuclease A also paved the way for the emergence of a new field in protein chemistry, namely, "the study of protein folding," which is considered to be the "second half of the genetic code." The understanding of the rules that govern protein folding is important not only from an academic point of view but also has widespread implications in biotechnology and is necessary to comprehend some of the disorders caused by the misfolding of proteins. (2) The central problem of protein folding is to understand the principles that drive the nascent polypeptide chain to its correct native state in spite of the availability of millions of alternate conformations, and this dilemma is the ‘Levinthal’s paradox.’ Several theories have been put forward to explain this paradox, the most recent being the New View theory. The ‘new view’ theory explains the folding in terms of ‘folding funnels (Fig. 1.3)’ and ‘energy landscapes (Fig. 1.4).’ The lateral area of an ‘energy landscape’ at a given depth represents the number of conformations having a specified intrachain free energy. Upon initiation of refolding, the polypeptide chain begins to slide down the ‘funnel,’ and with time, the conformational freedom becomes narrower, thereby channeling the protein towards its native structure within the biologically relevant time scale. The folding of monomeric proteins has been studied extensively, but the folding and association of oligomeric proteins still remains to be understood. The present thesis concerns the folding and assembly of SHMT, a tetrameric, PLP-dependent enzyme. (3) The enzymes of the ‘thymidylate cycle’ (thymidylate synthase, DHFR, and SHMT; Fig. 1.1) have attracted widespread attention as targets for cancer chemotherapy. Hitherto, the most widely used chemotherapeutic drugs have been directed towards thymidylate synthase and DHFR. Because of its involvement in the production of precursors for DNA synthesis and its altered regulation in the case of neoplastic cells, SHMT has been suggested as an alternate target for the development of anti-cancer drugs. (4) SHMT, a PLP-dependent enzyme, catalyzes the transfer of the hydroxymethyl group of serine to H4folate to yield glycine and 5,10-methylene-H4-folate. The enzyme, in addition, catalyzes a variety of reactions characteristic of PLP enzymes, such as transamination, racemization, and decarboxylation. (5) SHMT isolated from sheep liver is a homotetramer (subunit Mr 53,000) and requires PLP for catalysis. Extensive investigations from this laboratory have resulted in the elucidation of its primary structure, catalytic mechanism, identification of active site residues involved in catalysis, and interactions of a few potential chemotherapeutic agents at the active site. Recently, the gene coding for this enzyme has been cloned and overexpressed in E. coli. Site-directed mutagenesis and the study of deletion mutants have confirmed the function of a few of the active site amino acid residues identified earlier and the role of the NH2-terminal arm (14 aa) in maintaining the oligomeric structure of the enzyme. (6) The objectives of the present investigation were: a) To study the pathway for the reversible unfolding of tetrameric SHMT. b) To examine the role of cofactor, PLP, in the folding process. (7) The Materials and Methods used in this thesis, such as purification of sheep liver SHMT, its physicochemical and kinetic characterization, methods for monitoring the unfolding and refolding of the enzyme with urea or GdnHCl (far UV CD, fluorescence, size exclusion chromatography, etc.), production and characterization of monoclonal antibodies, epitope mapping, etc., are described in Chapter II of this thesis. (8) Sheep liver SHMT was purified using a combination of cation exchange chromatography, ammonium sulfate precipitation, and size exclusion chromatography. This procedure resulted in a 194-fold purification and an 18% recovery (Table III.1). The yield of the enzyme, starting from 0.75 kg of sheep liver, was 100-150 mg, sufficient for detailed physicochemical studies. The purity of the enzyme used in this study was confirmed by non-denaturing PAGE, activity staining (Fig. III.1a,b), and SDS-PAGE (Fig. III.2). The specific activity of the enzyme used in this study varied from 5.5–8 (1 unit = nmoles of HCHO formed/min at 37°C and pH 7.4). (9) To understand the folding pathway of a protein, the unfolding should be reversible under in vitro conditions. Chapter III describes the attempts to standardize conditions for the reversible unfolding of tetrameric SHMT. Urea and GdnHCl were chosen as denaturants in this study. Aggregation during refolding is often, especially in the case of oligomeric proteins, a competing reaction that leads to low or no recovery of the active protein. SHMT, being a tetramer, had a tendency to aggregate at intermediate concentrations of the denaturant (3 M urea or 1 M GdnHCl). The 8 M urea-induced unfolding was partially reversible (30-40%), as monitored by the recovery of activity, when the refolding was initiated by a 20-fold dilution into 50 mM potassium phosphate buffer, pH 7.3, containing 1 mM 2-ME, 1 mM EDTA, 10% glycerol, 0.1 M KCl, and 50 µM PLP (Table 10.2). The presence of aggregates at 3 M urea in the above buffer was established using biophysical techniques such as far UV CD (Fig. III.7), light scattering (Fig. B18), intrinsic fluorescence (Fig. III.9), fluorescence quenching (Fig. III.10), SEC (Fig. III.11). It was also shown that the aggregation at intermediate concentrations of the denaturant was the ‘bottleneck’ during refolding (Fig. III.12). (10) The formation of aggregates can be minimized by using detergents and stabilizing agents in the refolding buffer. Also, since aggregation is a diffusion-controlled phenomenon, enhanced recovery of activity can be achieved at low protein concentrations and specific temperatures. The presence of a non-ionic detergent, Brij-35, and polyethylene glycol (PEG 3350) in the refolding buffer greatly enhanced the recovery of active protein from 8 M urea or 6 M GdnHCl denatured states (Table III.3). The recovery was maximum (>95%) when the final protein concentration was 12.5 µg/ml and the refolding temperature was 30°C (Table D14). The spectral properties and kinetic parameters of the refolded enzyme were comparable to those of the native protein, which was not denatured but incubated under identical conditions (Tables III.5 & IV.2). (11) It was observed that the choice of detergent was important since an anionic detergent like CTAB inhibited the refolding (Table III.3). A strong reducing agent such as DTT was found to be essential, both during unfolding and refolding, presumably to prevent oxidation of SH groups. (12) MAbs have been used as sensitive tools to monitor the local conformational changes in complex molecules. Four MAbs (FgD7, G3B4, EUG5, and E2G8) against SHMT were obtained using classical hybridoma technology. These MAbs were characterized with respect to their isotype and epitope specificity. It was found that while FgD7 and G3B4 belonged to IgG1, E2G8 and EUG5 belonged to the IgG3 subclass. Epitope specificity, monitored by the method of additivity index, indicated that at least 3 MAbs recognized distinct epitopes, while E2G8 and EUG5 recognized the same or overlapping epitopes (Table IV.2). (13) Epitope mapping by identifying and sequencing the peptides generated by enzymatic digestion of SHMT, which cross-reacts with the MAb, showed that the epitope for FgD7 resided between residues 145-160 in the primary amino acid sequence of SHMT (Fig. IV.3 & IV.4). All attempts to identify the epitopes for the other 3 MAbs were unsuccessful, suggesting that these MAbs might recognize discontinuous epitopes. Hence, using partial proteolysis (Fig. IV.5) and an NH2-terminal deletion mutant, pSN6, available in the laboratory (Fig. IV.6), the epitopic regions recognized by these three MAbs were found to be within the NH2-terminal 75-280 residues. (14) Earlier studies on the holo- and apoenzymes had shown that the two forms of the enzyme were structurally similar. However, while all the four MAbs were able to recognize native (Fig. IV.1) and SDS-denatured (Fig. IV.2) antigen, FgD7 and G3B4 tended to react with the apoenzyme (Fig. IV.7a). EUG5 and E2G8 recognized both forms of the enzyme but with a lesser affinity towards the apoenzyme (Fig. IV.7b). These results suggested that the biophysically indistinguishable holo- and apoenzymes differ in the local conformations recognized by these MAbs. (15) In a direct ELISA, these MAbs reacted equally well with the control and the refolded enzymes, suggesting that the local conformation recognized by these MAbs is completely restored in the case of the refolded enzyme also (Fig. IV.8). (16) After standardizing the conditions for reversible unfolding (Table EDL3 and HL4) as well as establishing the similarity of the refolded enzyme with that of the native enzyme (Table III.5, VL2 and Fig. V.9), urea and GdnHCl induced folding pathways were studied in detail. To examine the role of PLP in the process, three different forms of the enzyme were used, i.e., holo, apo, and reduced holoSHMT, in which the cofactor was either present at the active site, respectively, as an internal aldimine, absent, or as a stable phosphopyridoxyl lysine residue at the active site. (17) Chapter V describes the urea-induced reversible unfolding pathway of SHMT monitored using a number of biochemical and biophysical techniques. The unfolding of the tetrameric SHMT (M?) proceeded via dissociation into dimers (M?) (Figs. V.4,5 & 6). The dissociation into dimers paralleled the loss of PLP (Fig. V.2) from the active site and of enzyme activity (Fig. V.1a). The unfolding of holoSHMT, monitored using far UV CD (Fig. V.1b), fluorescence (Fig. V.1c), and SEC (Figs. V.4,5), was found to be a multistate process with a ‘predenaturation transition’ (0-2 M urea; Table VL3) followed by a ‘global unfolding’ (2-6 M urea). The predenaturation transition (midpoint 1.2 M urea) was characterized by the dissociation of the tetramer into dimers and further unfolding of dimers into a putative intermediate (I) form. The apparent midpoint of dissociation was found to be independent of the protein concentration range used in the study (12.5 - 125 ?g/ml). The loss of secondary and tertiary structure during the predenaturation transition was minimal, and the proposed intermediate showed increased binding of a hydrophobic fluorescence probe, ANS (Figs. V.1d & V.2d). Global unfolding, monitored by CD and fluorescence, had similar midpoints (3.1±0.2 and 3.6 M urea, respectively) for both apo and holoSHMTs, and it represents the unfolding of this intermediate to the completely unfolded state (U). (18) The cofactor, PLP, had a profound effect on the predenaturation transition (Fig. V.1a,b) and SEC of holo and apoSHMT (midpoint of dissociation of 1.2 and < 0.5 M urea, respectively, Table VL3). This suggested that PLP, apart from its role in catalysis, has an important role in stabilizing the quaternary structure of the enzyme (Fig. V.4). This suggestion was confirmed by using reduced holoSHMT, which was stable towards urea-induced dissociation at low concentrations of the denaturant (Fig. V.8), and PLP, in the form of a phosphopyridoxyl lysine residue at the active site, afforded significant protection against urea-induced unfolding (Figs. V.7a,b). The presence or absence of PLP during the initial stages of refolding did not affect the recovery of active enzyme and the regain of native-like secondary structure and tryptophan fluorescence, suggesting that PLP probably was not required for the initiation of refolding (Fig. V.9). The oligomeric status of the protein during refolding, monitored using SEC, indicated that the refolding in the absence of externally added PLP resulted in the formation of only dimers. The addition of PLP shifted the equilibrium towards the formation of active tetramers (Fig. V.10). The reduced holoSHMT, in which the PLP was covalently linked even in the completely unfolded form, could be refolded and assembled into tetramers even in the absence of any externally added PLP, thus emphasizing the importance of PLP in achieving the final oligomeric structure (Fig. V.10c). (19) In an attempt to stabilize the putative unfolding intermediate, the GdnHCl-induced unfolding of SHMT was monitored (Chapter VI). Unlike during urea denaturation, in the case of GdnHCl-induced inactivation, the loss of activity and the release of PLP from the active site were two separate events (Fig. VL1). The inactivation at low concentrations of GdnHCl was due to competitive inhibition rather than removal of PLP or unfolding (Fig. VL1, inset). Far UV CD (Fig. VL2) and fluorescence measurements (Fig. VL3) suggested that the PLP present at the active site, unlike in the case of urea, failed to protect the enzyme against unfolding. The transition midpoints monitored using far UV CD and fluorescence were 1.6 ± 0.1 M and 2.0 ± 0.1 M GdnHCl, respectively. The non-coincidence of the melting curves obtained using different optical probes suggested the presence of intermediates in the unfolding pathway. The fluorescence data were analyzed according to the 3-state model. Though there was an excellent agreement between the experimental and theoretically determined melting curves (Fig. VL4), the thermodynamic parameters obtained (Table VL1) from such an exercise might not be a true representation of the free energy changes during unfolding, because of the multimeric nature of the protein and the presence of at least one unfolding intermediate, apart from the dimer. SEC indicated that the dissociation and unfolding in the case of holoSHMT was a concerted event. However, the apoenzyme unfolded via the formation of a dimer, further emphasizing the role of PLP in maintaining the quaternary structure of the enzyme (Fig. VL5a,b). At 0.45 M GdnHCl, the apoenzyme showed two peaks corresponding to dimer and an unfolding intermediate (I), which was less compact than the dimer and more compact than the completely unfolded enzyme. The equilibrium between the two species was slow enough to be separated on an SEC column (Fig. VL5a). The time course of change in the relative areas under the two peaks at 0.45 M GdnHCl suggested that the area under I increased at the expense of the dimer. Surprisingly, after 2-3 h of incubation, one more species, which was less compact than I and more compact than the completely unfolded form, was formed, and this peak increased at the expense of I (Figs. VL7,8). This might represent an independent intermediate or an oligomeric product of I. Both these forms showed increased ANS binding (Fig. VL9). (20) During refolding of GdnHCl-unfolded SHMT, PLP was not required for the initiation of refolding. But SEC of the refolding intermediates indicated that the protein refolded in the absence of PLP, eluted in the elution volume corresponding to the tetramer, and was inactive (Fig. VL10). This intermediate could be converted to the active form by the addition of PLP. This species eluting with an apparent Mr of the tetramer might represent the unfolding intermediate, which was seen at 0.45 M GdnHCl. (21) Based on the results obtained, the following minimum model for the reversible, equilibrium unfolding of tetrameric SHMT has been proposed. M? ? Aggregation (22) The similarities and differences between the folding pathways of dimeric (E. coli SHMT) and tetrameric (sheep liver SHMT) SHMTs are discussed in the light of the available literature with respect to the folding conditions, oligomeric status during (un)folding, role of PLP in the process, and the presence of possible intermediates in the pathway. Owing to its tetrameric nature, refolding of the unfolded sheep liver SHMT required the presence of Brij-35 and PEG-3350, unlike eSHMT, which could be refolded in the absence of these or any other additives. eSHMT has been shown to refold via the formation of monomer, which then formed ‘expanded apodimers.’ PLP binds to this intermediate, resulting in minor conformational changes to yield the active holodimer. On the other hand, the available evidence suggests that PLP can bind only to the tetrameric form and has an important role in stabilizing the quaternary structure of sheep liver SHMT. (23) The results presented in this thesis describe the complexities involved in the folding of an oligomeric protein like SHMT. The data obtained during the course of this investigation highlight the presence of at least one unfolding intermediate, apart from the dimer, during the reversible unfolding of tetrameric SHMT, and emphasize the role of cofactor PLP in maintaining the structural integrity of the enzyme