N-Acetyl-d-glucosaminylphosphatidylinositol De-N-acetylase from Entamoeba histolytica

Abstract

PIG-L/GPI12 proteins are endoplasmic reticulum-resident membrane proteins involved in the second step of glycosylphosphatidylinositol anchor biosynthesis in eukaryotes. We show that the Entamoeba histolytica PIG-L protein is optimally active in the acidic pH range. The enzyme has an intrinsic low level of de-N-acetylase activity in the absence of metal and is significantly stimulated by divalent cations. Metal binding induces a large conformational change in the protein that appears to improve catalytic rates while not altering the affinity of the enzyme for its substrate.

Introduction

Entamoeba histolytica is an intestinal pathogen largely responsible for incidences of amoebic dysentery. Infection by this organism is known to be primarily mediated by the glycosylphosphatidylinositol (GPI)⁵-anchored Gal/GalNAc lectin on its cell surface (1). Indeed, GPI-anchored proteins are also known to be involved in infection and virulence of several pathogens, including Trypanosoma (2) and Leishmania (3).

GPI anchoring of cell surface proteins to the membrane is particularly important in eukaryotes, and defects in GPI anchor biosynthesis can affect function as well as the viability of the cells (4). Thus, the GPI biosynthetic pathway is an important therapeutic target, provided the species-specific differences of this pathway can be appropriately exploited. As the shorter biosynthetic intermediates have recently been shown to be able to influence cell membrane organization as well as composition, the early steps of the pathway are particularly attractive targets for designing therapies (5).

The biosynthesis of GPI anchor is initiated by the transfer of N-acetyl-d-glucosaminylphosphatidylinositol (GlcNAc) from UDP-GlcNAc to phosphatidylinositol (PI) (6). This is followed by a second essential step, the de-N-acetylation of GlcNAc-PI to yield glucosamine-PI (GlcN-PI), a step catalyzed by PIG-L/GPI12 (7, 8). PIG-L proteins have been identified in almost all eukaryotic organisms. The yeast PIG-L homologue, GPI12, is essential for cell viability as is its mammalian counterpart (8). GPI12 shows 24% identity with the rat PIG-L protein (8) and is known to be localized only on the ER membrane (8). Human PIG-L is a 252-amino-acid-long protein with a large cytoplasmic tail (7) showing 77% homology with rat PIG-L (8). Unlike GPI12, it has two ER localization signals and is uniformly localized on the ER as well as in the mitochondria-associated ER membrane (9). Among PIG-L proteins, there appears to be species-specific differences not only in terms of location and conformation but also in terms of specificity of the active sites. For example, the yeast and human homologues of PIG-L are inhibited by a naturally occurring terpenoid, whereas the trypanosomal PIG-L is not (10). The Trypanosoma PIG-L enzyme, unlike the human PIG-L, recognizes and catalyzes the deacetylation of GlcNAc-[L]-PI, GlcNAc-β-PI, and GlcNAc-(2-O-alkyl)PI, a fact that has been exploited for the design of Trypanosoma-specific inhibitors (11). Likewise, the Plasmodium falciparum GlcNAc-PI de-N-acetylase is distinct from the human PIG-L counterpart, leading to the synthesis of Plasmodium-specific inhibitors (12).

The role of the PIG-L gene from E. histolytica has been reported using an inducible antisense RNA approach (13). Using this approach, it was shown that the total of GPI-anchored proteins on the cell surface of the pathogen dropped by 85–90%, which had a deleterious effect on cell proliferation. However, detailed characterization of the E. histolytica PIG-L has not so far been carried out.

We present the characterization of the cytoplasmic catalytic domain of the PIG-L protein (EhΔTMPIG-L) from E. histolytica. We show that the de-N-acetylase activity of the purified domain is significantly different from other PIG-L proteins reported in literature. EhΔTMPIG-L has a GlcNAc-PI-de-N-acetylase activity with a pH optimum of 5.5. The protein is active even in the absence of metal. Lineweaver-Burk plot analyses revealed that the divalent cation influenced V_max values but not the K_m of the enzyme, suggesting that the metal ion did not directly participate in ligand binding or recognition but was important for the catalytic efficiency of the enzymatic site.

EXPERIMENTAL PROCEDURES

Materials

Yeast strains (YPH501 and YPH500 (supplemental Table I) were purchased from Institute of Microbial Technology (Chandigarh, India), and DH5α cells were purchased from Bangalore Genei. The vectors pTZ18R and pTZ57R/T were purchased from Fermentas and New England Biolabs. YEpHIS vector was a kind gift from Dr. Marwan Al-Shawi, whereas EhPIG-L clone in pET30a was a kind gift from Prof. Alok Bhattacharya. Acetic anhydride was from Sigma, UDP[6-³H]GlcNAc was from American Radiolabeled Chemicals or Sigma, amylose resin was from New England Biolabs, and factor Xa was from Novagen. Restriction enzymes and DNA polymerases were from Fermentas, Bangalore Genei, or New England Biolabs, and all other materials were purchased from Merck, Qualigens, Himedia, or Sisco Research Laboratories. All the primers were synthesized by Sigma (supplemental Table II).

Cloning of E. histolytica PIG-L

The E. histolytica full-length PIG-L (EhFLPIG-L) as well as its transmembrane (TM)-deleted mutant (EhΔTMPIG-L) were cloned into pMAL-c2X such that the expressed protein contained an MBP tag at the N terminus (supplemental Experimental Procedures).

Complementation of ScGPI12 by EhPIG-L

To ascertain whether the EhPIG-L gene is capable of functionally complementing the ScGPI12 gene, a haploid yeast (YPH500) (supplemental Table I) mutant was generated in which the native promoter of ScGPI12 was replaced by the regulatable GAL1 promoter (YPH-GAL1-ScGPI12 (supplemental Table I) (17). This strain grew normally in galactose (Gal) but was defective in growth in glucose (Glc) medium (a condition where the expression from GAL1 promoter is shut down) because ScGPI12 is essential for yeast growth (8). This mutant was then transformed with EhPIG-L cloned into YEpHIS, a yeast expression plasmid (14), and YEpHIS alone (as vector control) and checked for the ability of the vector used for transformation to rescue the growth defect of this mutant in the Glc medium (supplemental Table I). Additionally, microsomes were prepared from all of these strains, in the presence of Gal and Glc, and assayed for the levels of de-N-acetylase activity (supplemental Table I).

Expression in Escherichia coli and Purification of MBP-tagged Proteins

TB1 strain of E. coli transformed with either pMALEhΔTMPIG-L or pMALEhFLPIG-L were grown at 37 °C in LB medium and protein expression induced by isopropyl-1-thio-β-d-galactopyranoside. The cell lysate, containing the protein of interest, was loaded on an amylose column, and the MBP-tagged proteins that bound to the column were eluted with maltose (supplemental Experimental Procedures).

Cleavage of MBP from MBPEhΔTMPIG-L

Details are given in the supplemental Experimental Procedures.

Preparation of ER Microsomes from Yeast

Yeast mixed membranes were prepared using standard protocols (15) with minor modifications (supplemental Experimental Procedures).

GPI-N-Acetyl Glucosamine Transferase (GPI-GnT) and De-N-acetylation Assays Using Yeast Microsomes

The early GPI intermediates were generated using standard protocols (8, 16) (supplemental Experimental Procedures). Radiolabeled glycolipids were detected by a Bioscan AR-2000 TLC scanner, and the area under the peak was used to quantify the amount of [6-³H]GlcNAc-PI and [6-³H]GlcN-PI present after each assay (supplemental Experimental Procedures). The activity of the microsomal fraction is reported here as the percentage of the fraction of [6-³H]GlcN-PI produced (supplemental Experimental Procedures). The [6-³H]GlcN-PI produced in such an assay could also be acetylated as described below to generate substrate for the recombinant proteins.

Acetylation of [6^-3H]GlcN-PI to [6^-3H]GlcNAc-PI

[6-³H]GlcNAc-PI for the GlcNAc-PI de-N-acetylase assay was generated by converting [6-³H]GlcN-PI formed in the GPI-GnT reaction back to [6-³H]GlcNAc-PI using acetic anhydride (8). The products were extracted with w-butanol, dried, and used for the GlcNAc-PI de-N-acetylase assays.

Substrate Concentration Determination

A standard curve was plotted by measuring different known amounts of UDP[6-³H]GlcNAc using a liquid scintillation counter (Packard Biosciences). A known amount of substrate ([6-³H]GlcNAc-PI) was then quantified using this standard curve, assuming that all radioactive counts arose from this species alone (TLC plate runs confirmed a single spot with R_f corresponding to GlcNAc-PI). It must be pointed out that the method overlooks the endogenously present GlcNAc-PI in the assay and hence only provides a rough estimate of the quantity of substrate.

GlcNAc-PI De-N-acetylase Assay for E. histolytica PIG-L Proteins

The EhPIG-L protein variants were tested for their ability to de-N-acetylate the [6-³H]GlcNAc-PI, generated as described above. Glycolipids were extracted, resolved by HPTLC, and analyzed as before. The specific activity was calculated as (F_P × pmol of substrate)/(time of assay in h × μg of protein) (supplemental Experimental Procedures).

Steady State Assays

Because our method for substrate estimation does not account for the endogenously present GlcNAc-PI in the assay, the values we report here are the “apparent” rather than “absolute” K_m and V_max for the catalytic activity and are intended only for comparing different assay conditions. A single batch of substrate was used for all assays to minimize errors in data analysis. A time course for the enzymatic activity suggested that it was linear in the time scale being studied (supplemental Fig. 1).

Far-UV Circular Dichroism (CD) of EhΔTMPIG-L

The CD spectra of purified cleaved EhΔTMPIG-L were recorded as described in the supplemental Experimental Procedures.

RESULTS

The presence of a conserved core in the GPI anchor of all eukaryotic systems studied so far points to the conserved nature of the pathway and its importance for the organism. PIG-L proteins are involved in the second step of this pathway, converting GlcNAc-PI to GlcN-PI. Because PIG-L is a transmembrane protein, in vitro characterization of the enzymatic activity of the protein has presented a major challenge. We present here the in vitro characterization of the overexpressed PIG-L protein from E. histolytica without its TM domain. We show that there are differences in the nature of the de-N-acetylase activity of the protein that could potentially be used to distinguish and specifically target this pathogen for therapeutic purposes.

EhPIG-L Is Able to Functionally Complement a Conditionally Lethal gpi12 Mutant

To determine whether EhPIG-L was the functional equivalent of the yeast GPI12, the conditionally lethal YPH-GAL1-ScGPI12 strain was made. Although Gal promotes growth of this strain, Glc represses it. When this strain was transformed with YEp-EhPIG-L, the growth repression in glucose was significantly relieved, suggesting that EhPIG-L is able to functionally complement yeast GPI12 (Fig. 1A). This was confirmed by assaying for the GPI-GnT and de-N-acetylation activity. As can be seen from Fig. 1B, microsomes from the strain grown in galactose show significant GlcN-PI production. The strain complemented with the vector control alone and grown in glucose showed very low GlcN-PI production. In comparison, the complemented strain grown in glucose showed over 2-fold recovery of de-N-acetylase activity.

FIGURE 1.

A, complementation of ScGPI12 by EhPIG-L. YPH-GAL1-ScGPI12 cells grow on 1% Gal plates (panel 1) but not on 1% Glc plates (panel 2). When complemented with EhPIG-L, the same cells were able to grow on both Gal (compare panel 3 with vector control in panel 4) as well as Glc plates (compare panel 5 with vector control in panel 6). B, recovery of de-N-acetylase activity. Microsomes (∼1 mg of total protein) isolated from the various yeast strains were assayed for their ability to generate [6-³H]GlcN-PI in 2 h at 30 °C. UDP[6-³H]GlcNAc (1 μCi; 60 Ci/mmol) was used in these assays. YPH-GAL1-GPI12 strain grown in 1% Gal, Sc(Gal), shows a significant percentage of the fraction of [6-³H]GlcN-PI formation, whereas those grown in 1% Glc, Sc(Glc), show about one-third of this activity. Expectedly, microsomes from YPH-GAL1-ScGPI12 containing EhPIG-L, Eh(Gal), or the vector alone grown in Gal, Vector (Gal), show activity similar to Sc(Gal). Microsomes from cells containing the YEpHIS vector alone when grown in Glc, Vector (Glc), show low de-N-acetylase activity, similar to Sc(Glc). Microsomes from YPH-GAL1-ScGPI12 containing EhPIG-L and grown in Glc, Eh(Glc), show 2-fold higher activity in comparison with Sc(Glc) or Vector(Glc). C, SDS-PAGE showing purification of MBPEhΔTMPIG-L. Total protein (3.6 μg) was loaded in each lane. Lane 1, mock-purified MBPlacZ; lane 2, eluted MBPEhΔTMPIG-L (∼66 kDa); M, molecular weight markers. D, factor Xa cleavage of MBPEhΔTMPIG-L. The cleavage and removal of MBP tag from EhΔTMPIG-L were done as explained under “Experimental Procedures.” Lane 1, purified MBPEhΔTMPIG-L; lane 2, after cleavage and dialysis (MBP has a molecular mass of ∼45 kDa, whereas EhΔTMPIG-L has a molecular mass of ∼28 kDa); lanes 3–5, after the first, second, and third round of passage through amylose (1 ml); M, molecular weight markers. E, purification of MBPEhFLPIG-L was done as explained under “Experimental Procedures.” Lane 1, eluted MBPEhFLPIG-L (upper band in the doublet) co-purified with E. coli GroEL (lower band); lane 2, mock-purified MBPlacZ; M, molecular weight markers. F, acetylation of [6-³H]GlcN-PI. [6-³H]GlcN-PI formed after the GPI-GnT assay was converted back to [6-³H]GlcNAc-PI with (CH₃CO)₂O as explained under “Experimental Procedures.” Glycolipids were analyzed before (top panel) and after (lower panel) acetylation using HPTLC. The identities of GlcNAc-PI and GlcN-PI are shown. G, the GlcNAc-PI de-N-acetylase activity of EhPIG-L variants in vitro. [6-³H]GlcNAc-PI (∼1 nmol) generated after acetylation was incubated with 4 μg of protein at 37 °C for 2 h in the presence of 5 mm each of MgCl₂ and MnCl₂. The glycolipids were analyzed by HPTLC. H, radiolabeled glycolipids were detected, and the specific activity was calculated as mentioned under “Experimental Procedures.” C-1, no protein control; MBP, MBPLacZ; Cyt., MBPEhΔTMPIG-L; FL, MBPEhFLPIG-L; FL+n-OG, MBPEhFLPIG-L in the presence of 0.06% n-OG; C-2, control for FL+n-OG (MBPEhFLPIG-L in absence of detergent, different batch of protein; all other conditions identical). I, comparison of the activity of EhΔTMPIG-L with MBPEhΔTMPIG-L. GlcNAc-PI de-N-acetylase assay was done as above followed by quantification of glycolipids. Fusion, MBPEhΔTMPIG-L; Cleaved, EhΔTMPIG-L. The lower activity for MBPEhΔTMPIG-L as compared with that in H may be explained by the fact that the sample was incubated at 20 °C for 16 h as control. In panels B, H, and I, error bars indicate S.D.

Overexpression and Purification of PIG-L from E. histolytica

The full-length as well as TM-deleted EhPIG-L were cloned into pMAL-c2X vector. The corresponding MBP-tagged proteins, MBPEhFLPIG-L and MBPEhΔTMPIG-L, respectively, were found to be expressed in TB1 cells. However, the expression level of MBPEhFLPIG-L was low as compared with the MBPEhΔTMPIG-L and was barely visible on a Coomassie Brilliant Blue-stained gel. The soluble MBP-EhΔTMPIG-L was purified as a single band of an approximate molecular mass of 66 kDa (Fig. 1C). The factor Xa site between MBP and EhΔTMPIG-L was accessible, and EhΔTMPIG-L could be cleaved from the MBPEhΔTMPIG-L (Fig. 1D). The cleaved MBP could be efficiently removed by passing this sample through a column of amylose (Fig. 1D). In contrast, the full-length MBPEhFLPIG-L co-purified with the E. coli chaperone protein GroEL (Fig. 1E), as was confirmed by the peptide mass fingerprint of GroEL (data not shown). The GroEL could not be separated from the protein to get reasonable yields of soluble protein for analysis. Hence, the de-N-acetylase activity of this protein was assayed in the presence of bound GroEL in both the absence and the presence of detergent as in Ref. 18.

Substrate for De-N-acetylase Activity

From the complementation analysis, we concluded that the EhPIG-L is capable of using the yeast GlcNAc-PI as a substrate. Hence, phospholipids extracted after the GPI-GnT assay using yeast microsomes were acetylated and used for the de-N-acetylase assay. As can be seen from the TLC profiles of glycophospholipids before and after acetylation (Fig. 1F), the faster moving GlcN-PI present before acetylation almost completely disappears after acetylation, yielding higher concentrations of the GlcNAc-PI substrate for the activity assay.

TM Deletion Results in a Functionally Folded Protein Capable of De-N-acetylase Activity

MBPEhΔTMPIG-L was found to be an active de-N-acetylase as seen by the appearance of the GlcN-PI peak in Fig. 1G and the disappearance of the GlcNAc-PI peak concomitantly. In contrast, MBPEhFLPIG-L and MBP alone do not show significant de-N-acetylase activity in the absence of detergent (Fig. 1, G and H). However, MBPEhFLPIG-L does show a low level of activity in the presence of 0.06% octyl-β-glucoside (n-OG) (Fig. 1H), presumably due to better solubilization and therefore greater availability of the substrate to the active site of the enzyme.

On the other hand, the MBPEhΔTMPIG-L was found to be over 3-fold more active even in the absence of detergent than MBPEhFLPIG-L in the presence of n-OG (Fig. 1, G and H). The EhΔTMPIG-L was also active after cleavage of MBP, thus demonstrating that the fusion tag does not interfere with the functional conformation of the cytoplasmic domain (Fig. 1I).

Given the significantly higher activity of the cytoplasmic domain as compared with full-length protein, we focused our studies on the TM-deleted protein. Further, as yields were low, for all experiments, unless stated otherwise, the MBP fusion tag was not removed because the tag did not appear to significantly alter the de-N-acetylase activity.

Early cell-free studies using mixed microsomes had indicated that the de-N-acetylase activity in mammalian cells was stimulated by GTP, but later studies indicated that the role of GTP could be in stimulating membrane fusion in microsomes used for these studies (19). We too did not observe any stimulation of the GlcNAc-PI de-N-acetylase activity of MBPEhΔTMPIG-L by GTP (data not shown).

Divalent Cations Stimulate De-N-acetylation

Previous reports have suggested that the catalytic domain of rat PIG-L is a metalloprotein that lost activity irreversibly when treated with metal chelators for long periods of time (8, 18). Hence, we tested whether MBPEhΔTMPIG-L was also a metal-dependent de-N-acetylase.

As can be seen in Fig. 2A, MBPEhΔTMPIG-L was active in the absence of any externally added metal. However, the de-N-acetylase activity of MBPEhΔTMPIG-L was 2-fold higher in the presence of 5 mm each of Mg⁺² and Mn⁺², suggesting that the activity of the enzyme was stimulated by divalent metals. To determine whether the EhPIG-L had an intrinsically bound metal, the protein was treated with EDTA (5 mm). Treatment with EDTA consistently caused up to ∼2% reduction in activity (Fig. 2A). The time period of incubation with the metal chelator (between 5 and 60 min) did not appear to affect the recovery of the de-N-acetylase activity of the protein upon subsequent supplementation with metal ion. In this regard, the E. histolytica PIG-L appears to differ from rat PIG-L (18).

FIGURE 2.

A, effect of EDTA and divalent metal ions on the GlcNAc-PI de-N-acetylase activity. MBPEhΔTMPIG-L (0.38 mg/ml) was incubated without (−Ions) or with (+EDTA) 5 mm EDTA at 37 °C for different times as indicated in figure, and then 10 μl of this sample was tested for activity (40-μl reaction) in either the absence or the presence of 5 mm MnCl₂ (+Metal) in 50 mm acetate buffer (pH 5.5) at 37 °C for 2 h using ∼1 nmol of substrate. Inset, MBPEhΔTMPIG-L (0.4 mg/ml) was incubated without (−Ions) and with (+1,10-PO) 1 mm 1,10-PO for 60 min and assayed for activity as before in the absence or presence (+Metal) of 5 mm MnCl₂. B, effect of different divalent cations (5 mm) on the activity of MBPEhΔTMPIG-L. The assay used ∼4 μg of enzyme and ∼1 nmol of substrate. C, binding of Mn²⁺ and Zn²⁺ causes similar conformational changes in EhΔTMPIG-L. Far-UV CD spectra are shown for the protein in the absence (curve 1) and presence of 5 mm MnCl₂ (curve 2) or ZnCl₂ (curve 3). D, effect of pH on the GlcNAc-PI de-N-acetylase activity. The activity of MBPEhΔTMPIG-L (0.4 μg) was assayed for 2 h at the indicated pH values in either 50 mm acetate (pH 3.5, 4.5, 5.5, 6.5) or 50 mm HEPES (pH 7.5, 8.5) buffers using ∼140 pmol of substrate in the presence of 5 mm MnCl₂. The assay was carried out for 4 h in the absence of metal. E, pH dependence of the global conformation of EhΔTMPIG-L. No significant difference in overall conformation was seen in far-UV CD spectra at the pH indicated. F, Lineweaver-Burk plots for MBPEhΔTMPIG-L. Assays were carried out at pH 5.5 using ∼0.4 μg of enzyme and varying amounts of substrate, as described under “Experimental Procedures,” in the absence (■) and presence (○) of 5 mm MnCl₂. Results are the average of two sets of experiments carried out in duplicate. In panels A, B, and F, error bars indicate S.D. In panels C and E, MRE indicates Mean Residue Ellipticity.

One possible explanation for the activity of the protein in the absence of any added metal ions could be that the protein purifies with an intrinsically tightly bound metal ion. Hence, MBPEhΔTMPIG-L, purified in the absence of ions, was treated with 1,10-phenanthroline (1,10-PO) for 60 min and assayed for de-N-acetylase activity (Fig. 2A, inset). As can be seen from the figure, the activity was not significantly affected by treatment with the stronger metal chelator. Further, the protein was stimulated ∼2-fold by the addition of 5 mm MnCl₂/MgCl₂. Additionally, it was observed that protein purified in the absence of any added metal could be stimulated to a level of activity that roughly corresponded to the activity of the protein purified in the presence of added metal. These results suggest that PIG-L from E. histolytica is active in the absence of metal but is stimulated by metal.

Further, several divalent cations were found to stimulate the activity of the protein (Fig. 2B). It is interesting to note that although Mg²⁺, Co²⁺, and Mn²⁺ could stimulate activity to similar extents, Zn²⁺ at the same concentration did not stimulate the de-N-acetylase activity of MBPEhΔTMPIG-L to the same extent, although it did bind to the protein and induce a conformational change very similar to Mn²⁺ (Fig. 2C).

Kinoshita and co-workers (8) had reported similar results in their study using full-length rat PIG-L protein, which co-purified along with GroEL. The de-N-acetylase activity of the protein was not affected by metal chelators but was stimulated, among others, by Ni²⁺, Mn²⁺, Co²⁺, and Mg²⁺ but not by Zn²⁺. Ferguson and co-workers (18), on the other hand, showed that the cytoplasmic domain of rat PIG-L was a 1,10-PO-sensitive metalloenzyme with preference for Zn²⁺ over Cu²⁺, Ni²⁺, Co²⁺, or Mg²⁺.

Metal Does Not Influence the pH Profile of the Enzyme Activity

Does the presence of the metal ion influence the pH profile of the enzymatic activity of the protein? To address this question, we next analyzed the pH dependence of the enzymatic activity of the protein in the absence and presence of ions. The pH profile of the de-N-acetylase activity of MBPEhΔTMPIG-L (as well as the EhΔTMPIG-L) indicated that the enzyme was maximally active at pH 5.5 in the absence as well as in the presence of metal ions (Fig. 2D). To rule out the possibility that the choice of buffers, in particular the presence of acetate, was in any way influencing the assay, we performed the assay at pH 6.5 using both acetate and phosphate buffers and saw no significant difference in activity between the two cases (supplemental Fig. 2). To determine whether the loss in activity at higher pH was due to a conformational change, CD spectra of EhΔTMPIG-L were recorded. As can be seen from Fig. 2E, the overall conformations of the protein at pH 5.5, pH 7.5, and pH 8.5 are nearly identical, suggesting that the global conformation of the protein is not altered by change in pH. Thus, pH appears to be directly modulating the functioning of the active site residues of the protein.

Metal Alters Rate of De-N-acetylation without Influencing Substrate Affinity

We next investigated whether the metal ion had a direct role in substrate binding in the active site. For this purpose, we analyzed the activity of the enzyme, in the absence and presence of metal, as a function of different substrate concentrations. Lineweaver-Burk plots of the resultant data (Fig. 2F) indicated that in the absence of metal, the protein bound to its substrate with an apparent K_m of 1.95 ± 0.24 μm and a V_max of 55.15 ± 2.96 pmol·h⁻¹·μg of protein⁻¹. Interestingly, the presence of metal did not significantly alter the K_m (2.05 ± 0.18 μm) of the enzyme for its substrate although it greatly enhanced V_max (130.08 ± 5.62 pmol·h⁻¹·μg of protein⁻¹). Thus, the effect of the metal is on the catalytic efficiency of the enzyme due to alteration of k_cat rather than K_m.

The unaltered apparent K_m also implies that the presence of the metal does not influence substrate binding, indicating that the metal may not be directly contacting the substrate in the active site. However, as also evidenced in the CD spectra (Fig. 2C), metal binding induces a very significant change in conformation of the protein. It is likely, therefore, that the role of the metal is in inducing a conformational change in the protein that results in formation and/or stabilization of a catalytically efficient active site in the protein, bringing the catalytic residue of the active site within optimum distance/orientation with respect to the substrate for the de-N-acetylase activity.

DISCUSSION

In the present study, we found that the EhPIG-L functionally complemented a conditionally lethal yeast gpi12 mutant. Thus, EhPIG-L appeared capable of recognizing the yeast GlcNAc-PI as substrate.

Therefore, using substrate generated from yeast microsomes, we looked for the activity of the recombinant protein expressed as an MBP-tagged fusion protein. In the absence of detergent, the full-length protein was inactive. However, in the presence of 0.06% n-OG, the full-length protein exhibited a low level of activity. Thus, MBPEhFLPIG-L appeared to be capable of enzymatic function. The recombinant MBPEhΔTMPIG-L, on the other hand, was significantly more active even in the absence of detergent.

Most interestingly, we found that the de-N-acetylase activity of EhΔTMPIG-L was stimulated by several divalent cations. The only reported crystal structure (PDB:1UAN) of a putative PIG-L protein from Thermus thermophilus HB8, TT1542, did not contain a bound metal ion in the active site (20). However, it is not clear whether this represents a functional protein. Our results indicate that even in the absence of metal ions, the PIG-L protein from E. histolytica was active. Thus, it is possible that the crystal structure of the archaeal putative PIG-L represents a functional enzyme.

Early reports using recombinant full-length rat PIG-L protein had indicated that the enzyme was not inhibited by EDTA, although it was stimulated by several divalent cations (8). More recent reports on the active domain from rat PIG-L have suggested that it is an extremely 1,10-PO-sensitive metalloenzyme that does not recover activity upon long incubation with the chelator (18). However, the E. histolytica PIG-L appears to be different from rat PIG-L because a similar sensitivity to either EDTA or 1,10-PO was not observed in our assays.

That the divalent metal induces a conformational change in the domain is clear from our CD spectroscopic data. Further, the metal influences neither the optimum pH of the de-N-acetylase activity nor the apparent affinity of the enzyme for the substrate. Thus, the role of the metal in either coordinating the ligand directly or participating in the catalysis by lowering the pK_a of H₂O for a nucleophilic attack on the amide carbon is ruled out. Hence, it would appear that the main role of the metal is in inducing and/or stabilizing a conformational change in the enzyme such that it is catalytically more efficient than the apo-form of the enzyme.

It is interesting to note that the optimum pH of the enzymatic activity was 5.5. Milne et al. (21) in an early report on PIG-L from African trypanosomes showed that the pH optimum of the protein was 7.4 and that the enzyme was active in the range of pH 6.5–8.0. Although very detailed analyses of the pH profiles of other PIG-L proteins studied so far are not available, all studies use pH 7.4 for their analyses, leading us to the inference that this might be the optimum pH for these PIG-L proteins. If so, the very fact that the amoebic PIG-L has a different optimum pH for activity is interesting, although the physiological significance of pH 5.5 is not clear. We do not know, for example, how the protein interacts with the membrane itself and whether this influences the active site, besides controlling substrate presentation and solvent accessibility to this site. We also do not know whether the N-terminal transmembrane domain directly influences the structure of the protein and thereby, the activity/conformation of the active site of the protein in vivo.

An acidic pH for the catalytic activity would suggest the role of one or more acidic residues in the active site of the protein. The Ferguson group (18) had proposed an elegant model for de-N-acetylation by rat PIG-L, which involved a role for residues from the conserved HPDDD and HXXH motifs in the catalytic activity of the enzyme. Indeed, ClustalW alignment also suggests the presence of a corresponding HADDD and a HXXH motif in EhPIG-L (supplemental Fig. 3). Although the roles of these residues in EhPIG-L may not be identical to that in rat PIG-L, the conserved acidic residues could perhaps be important for de-N-acetylation by the EhPIG-L. Site-directed mutagenesis studies are currently in progress to identify the residues that play a direct role in catalysis and to develop a model for the mechanism of de-N-acetylation by EhPIG-L.