HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 276, Issue 40, 37230-37236, October 5, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/40/37230    most recent M104326200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Kosman, J. Articles by Jonas, A. Search for Related Content PubMed PubMed Citation Articles by Kosman, J. Articles by Jonas, A. Social Bookmarking                      What's this?

Deletion of Specific Glycan Chains Affects Differentially the Stability, Local Structures, and Activity of Lecithin-cholesterol Acyltransferase^*

Jeffrey Kosman and Ana Jonas

From the Department of Biochemistry, College of Medicine at Urbana-Champaign, University of Illinois, Urbana, Illinois 61801

Received for publication, May 14, 2001, and in revised form, July 18, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The enzymatic and interfacial binding activity of lecithin-cholesterol acyltransferase (LCAT) is affected differentially by the location and extent of its glycosylation. Two LCAT glycosylation-deficient mutants, N84Q and N384Q, were constructed, permanently expressed in Chinese hamster ovary cells, and purified to determine the effects of deleting individual glycan chains on its stability, structure, and function. These purified mutants were studied by spectroscopic structural methods and enzymatic and binding assays to develop a molecular rationale for the relationship between LCAT glycosylation and activity. The N84Q LCAT mutant did not possess measurable enzymatic activity or interfacial binding affinity for reconstituted high-density lipoproteins. In addition, in thermal and chemical denaturation studies, N84Q LCAT was found to be significantly less stable than wild-type LCAT. The N384Q variant was initially more enzymatically active than wild-type LCAT, but gradually lost activity within months; however, it retained full interfacial binding activity. Significant changes were detected over time by circular dichroism in the -helical content of N384Q LCAT and in the -sheet content of N84Q LCAT, compared with wild-type LCAT. Fluorescence measurements with the probe 1-anilinonapthalene-8-sulfonate suggested an alteration of the active site cavity in both mutants. In conclusion, both mutants lost catalytic activity, N84Q shortly after purification and N384Q more gradually, and were destabilized, probably because the deletion of the glycan chains altered local structural elements near the active site cavity and/or the interfacial binding regions.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Lecithin-cholesterol acyltransferase (LCAT)¹ is the enzyme responsible for the generation of the majority of the cholesterol esters present in human plasma. The conversion of cholesterol to cholesterol esters is necessary to facilitate reverse cholesterol transport, the process which removes excess cholesterol and phospholipids from peripheral cells and delivers them to the liver for disposal (1). The LCAT reaction occurs primarily on the surface of high-density lipoproteins (HDL), which contains phospholipids, cholesterol, and apolipoprotein A-I (apoA-I). Although no crystal structure exists for LCAT, the study of its structure-function relationships has been facilitated by the construction of a computer model of the catalytic core of LCAT (2-6). Based on its amino acid composition alone, LCAT has an intermediate hydrophobicity between integral membrane proteins and apolipoproteins (7). One modification commonly found in plasma proteins, including LCAT, that increases their solubility and hydrophilicity is glycosylation.

Four N-linked complex type glycans and two O-linked glycans are found in LCAT (8, 9) and constitute around 20% of the total enzyme mass of human plasma LCAT (10). Different cell types transfected with the human LCAT gene express the glycan structure of LCAT differently (11, 12), which leads to disparities between recombinant preparations of LCAT. Chinese hamster ovary (CHO) cells express LCAT bearing a similar glycosylation pattern to plasma LCAT (12), making this cell type a preferred recombinant host. The polar glycan chains help solubilize LCAT and prevent it from aggregating; however, glycosylation is known to have a variety of additional effects on other proteins, including regulation of intracellular processing, induction and stabilization of protein folding, determination of binding affinity, and modulation of enzymatic activity. While the removal of certain glycan chains has no detectable effect on some glycoproteins, the removal of others may modify any or all of the functions listed above. Because the effects of removing one or more glycan chains from glycoproteins cannot be predicted or generalized from the available examples, each glycoprotein represents a unique system for the investigation of structure-function interactions.

The individual glycans of LCAT have been investigated by other laboratories to determine their role in LCAT function. Two of the glycosylation-deficient mutants, N20Q and N272Q, were similar to WT LCAT in enzymatic activity. However, the N84Q mutant retained little activity compared with WT LCAT, whereas the N384Q mutant was twice as active as WT (13-16). These studies involved recombinant LCAT present in cell media; purified mutants of LCAT are required to study structure-function relationships and to clarify some of the ambiguity regarding the effects of glycosylation on LCAT activity. In particular, determination of the reasons for the decrease of activity in the N84Q mutant and the increase of activity in the N384Q mutant may uncover significant facts regarding the catalytic mechanism of LCAT. Therefore, the goals of this study were to elucidate the effects of the deletion of glycan chains at positions Asn⁸⁴ and Asn³⁸⁴ in purified LCAT, using several experimental approaches to determine differences between the glycosylation-deficient mutants and WT LCAT in enzymatic activity, interfacial binding, and structure and stability.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Construction of LCAT Glycosylation Mutants-- Asparagine residues 84 and 384 were altered to glutamine to prevent glycosylation at those sites. Mutagenic primers were constructed by the Genetic Engineering Facility at the University of Illinois, Urbana-Champaign. The sequence of the N84Q mutagenic primer was: 5'-GGTTGTCTACCAGCGGAGCTCCGGACTCGTGTCCAAC-3', which also introduced a BspEI restriction site. The N384Q mutagenic primer was: 5'-GGTCTTCAGCCAGCTGACGCTGGAG-3', which introduced a PvuII restriction site. Underlined bases are those changed from the human LCAT sequence. The QuikChange Site-directed Mutagenesis Kit (Stratagene) was employed to generate the LCAT mutants (17), which were subcloned into the pSVDNA expression vector, described by Lee et al. (18). Constructs were sequenced by the Genetic Engineering Facility.

Expression and Purification of LCAT Glycosylation Mutants-- Overexpression of the LCAT glycosylation mutants was carried out in dhfr CHO cells as described by Jin et al. (19). Cells were stably transfected by the calcium phosphate method. The selective pressure of increasing methotrexate concentrations in cell medium produced concomitant amplification of both dhfr and LCAT genes in the cells. Colonies overexpressing LCAT were isolated and the level of expression was analyzed by Western blotting (17). Eventually, cell lines expressing 2.5 mg/liter of cell medium of N84Q LCAT and 5.0 mg/liter of cell medium of N384Q LCAT were produced. WT LCAT (10 mg/liter) was expressed by a CHO cell line generated by an identical procedure by Jin et al. (19).

Purification of WT LCAT and the glycosylation mutants was carried out (19) over phenyl-Sepharose and Affi-Gel Blue columns. Protein concentration was determined by A₂₈₀ and purity was evaluated on Coomassie or silver-stained SDS-PAGE Phast gels (Amersham Pharmacia Biotech). Pure protein fractions were concentrated and dialyzed into storage buffer of 10 mM Tris, 5 mM EDTA, pH 7.6, and were stored on ice, in capped plastic tubes under nitrogen. Irreversible aggregation of LCAT was monitored by nondenaturing Phast gels and by Western blots.

Enzymatic Assays-- Acyltransferase assays for LCAT were performed as described previously (18) with a minor modification. Because of the relative inactivity of the LCAT mutants, up to 10 times more mutant LCAT was used than WT LCAT in order to obtain kinetic data of sufficient accuracy.

LCAT esterase assays were performed based on the procedures of Bonelli and Jonas (20). For the same reason as in the acyltransferase assays, samples required higher mutant LCAT concentrations to obtain measurable activity. However, in neither enzymatic assay did the mutant LCAT concentration exceed that which would induce aggregation. Irreversible aggregation is observed at concentrations greater than 0.5 mg/ml of pure WT LCAT (19).

Interfacial Binding to rHDL-- The surface plasmon resonance method was used to measure the interfacial binding kinetics of LCAT with reconstituted HDL particles (rHDL). These particles were prepared as described in previous work (21) except that cholesterol was not included and 1-palmitoyl-2-oleoyl phosphatidylcholine (POPC) was substituted for egg phosphatidylcholine. Thus, the molar ratio of the components of rHDL was 99:0.7:1 of POPC (Sigma)/Cap-Biotinylated dipalmitoyl phosphatidylethanolamine (Avanti Polar Lipids)/apolipoprotein A-I (purified from human plasma). Control dipalmitoyl phosphatidylcholine (DPPC) rHDL were made in the same proportions, substituting DPPC for POPC. The small proportions of biotinylated lipid used did not interfere with any properties of the rHDL (21). The incorporation of biotinylated lipid was later determined by the Pierce Biotinylation Kit to be 1.1 mol of biotin/mol of rHDL. The lack of cholesterol and the ambient temperature used in these experiments (25 °C) ensured that little or no enzymatic activity took place on the rHDL during the short exposure time to LCAT.

Association and dissociation rates of LCAT on the immobilized rHDL were measured on a Biacore 3000 instrument (Biacore, Inc.) maintained by the Immunological Resources Center (University of Illinois, Urbana, IL). The SA (streptavidin) chip (Biacore) was used to immobilize the biotin-labeled rHDL. Biotinylated POPC and DPPC rHDL were injected through separate flow cells until around 900 response units of mass had bound to each cell. LCAT does not bind appreciably to DPPC rHDL under the present reaction conditions, so a flow cell coated with this particle provides a negative control for POPC rHDL binding. LCAT was diluted to 2 µM to 62 nM and passed though the cells using KINJECT mode at 30 µl/min, for 1 min contact time, and 5 min dissociation time (21). LCAT dissociated from rHDL completely during the dissociation period, making regeneration of the surface unnecessary. Corrected sensorgrams were analyzed by Biacore evaluation software, using the algorithm for a 1:1 Langmuir binding isotherm with global fitting kinetics, compensated for drifting baseline. Finally, a mass transfer control experiment was performed, verifying that binding constants were independent of flow rate.

Thermal Denaturation-- Thermal denaturation experiments were performed by determining enzymatic activity as a function of temperature. Purified LCAT is not stable at temperatures above 4 °C, and does not retain activity for long at temperatures above 37 °C. However, LCAT in plasma or cell media is much more stable and can be used to judge the relative stability of WT compared with mutant enzymes as a function of temperature. Aliquots of cell media were incubated in water baths from 37 to 55 °C for 1 h prior to use in the assay. The decrease in percent conversion of substrate was assumed to correspond to enzyme denaturation. Denaturation curves and the T_M, or midpoint of denaturation, were calculated by Origin software.

Circular Dichroism and Fluorescence Measurements-- All circular dichroism (CD) measurements were performed on a Jasco J-720 spectropolarimeter using parameters given previously (19). Pure LCAT samples (0.05 mg/ml) contained progressively increasing concentrations of the chemical denaturant guanidine hydrochloride (GdnHCl). Spectra were analyzed by the SELCON 3 algorithm. The intrinsic fluorescence of LCAT due mostly to tryptophan residues was also measured as described previously (19). All LCAT samples were identical to those used in CD measurements.

ANS Fluorescence-- The fluorescent probe 1-anilinonapthalene-8-sulfonate (ANS) (Molecular Probes) increases in quantum yield and blue shifts upon binding to a hydrophobic protein site. Thus, ANS is a useful indicator of hydrophobic regions in a protein accessible to the probe. A common application of ANS is the detection of folding intermediates or unfolded states of a protein (22). In this case, it is used to assess qualitatively the differences in structure between WT LCAT and the glycosylation mutants. Experimental conditions were modeled after studies on apoA-I (22). LCAT was added to a 1-cm path length fluorescent cuvette at a concentration of 0.05 mg/ml. ANS was dissolved in methanol as a 5 mg/ml solution. To ensure complete saturation of available binding sites, ANS was added in a 100-fold molar excess over LCAT. The fluorometer was a PerkinElmer Life Sciences LS-50B set to the following parameters: emission slit, 5 nm; excitation slit, 5 nm; excitation wavelength 395 nm; emission scan range, 405-560 nm; scan speed 200 nm/min; and 8 accumulations. In addition, carbonic anhydrase (Sigma) was tested at the same concentration and conditions as LCAT as an example of a globular protein that has very little affinity for ANS.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Characterization of LCAT Glycosylation Mutants-- The glycosylation mutants of LCAT were examined in terms of their purity, molecular mass, and tendency toward aggregation. The peak of spectral mass (Protein Sciences Facility, University of Illinois) for WT LCAT was 58.0 kDa, N84Q LCAT was 56.2 kDa, and N384Q LCAT was 57.0 kDa as determined by mass spectrometry analysis. This decrease in mass observed in the two mutants corresponds closely to the theoretical decrease for loss of a triantennary and biantennary complex-type N-linked glycan chain, respectively. Thus, the predominant glycan chain structure is similar at these two sites between CHO-expressed LCAT and human LCAT (8). Similar relative masses were observed using SDS-PAGE gels and high molecular weight markers (Fig. 1A). This gel also shows the high degree of purity (>95%) of the WT and mutant LCAT preparations. Gels, after eight months of storage of the LCAT variants, gave identical results.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Characterization of purified LCAT variants. A, SDS-PAGE, Coomassie stained, showing the purified recombinant LCAT species. Lane 1, molecular weight protein standards; lane 2, WT LCAT; lane 3, N84Q mutant; lane 4, N384Q mutant. Note the slight differences in migration due to the distinct glycosylation states. B, SDS-PAGE and Western blot analysis of purified LCAT variants. Different amounts of each LCAT species were used to attain comparable band intensities. Lanes 1 and 4, WT LCAT (16 ng); lane 2, N84Q mutant (80 µg); lane 3, N384Q mutant (80 ng). C, native, Coomassie-stained 12.5% polyacrylamide gel showing the LCAT bands near the 67-kDa molecular marker. No aggregated, high molecular bands are present. Lane 1, WT LCAT; lane 2, N84Q mutant; lane 3, N384Q mutant; lane 4, molecular weight (Stokes radius) protein standards.

Fig. 1B shows Western blots of the purified LCAT variants adjusted in the mass applied to the gel to give comparable staining intensity. Evidently the N84Q mutant is about 500-fold less immunoreactive than WT LCAT, and the N384Q mutant is about 5-fold less reactive. This illustrates the importance and specificity of glycosylation in the expression of glycoprotein epitopes.

The aggregation tendencies of the two LCAT mutants did not differ significantly from that of WT LCAT, as tested by native, nondenaturing PAGE (Fig. 1C). At the concentrations used for storage and for these experiments (<0.35 mg/ml), no aggregation was detected for any of the LCAT species. Concentrations in excess of 0.5 mg/ml WT LCAT have been shown in the past to result in irreversible aggregation (19).

Enzymatic Activity-- The acyltransferase assay is an important tool for the characterization of the LCAT reaction, for it encompasses all steps of LCATs enzymatic reaction with a physiologically relevant substrate, the rHDL, and reflects the interfacial binding affinity for rHDL in the apparent K_m value. Analysis of the results of the acyltransferase assay using a Lineweaver-Burk plot allowed calculation of the apparent k_cat and K_m of LCAT (Table I). The deletion of glycan chains Asn⁸⁴ or Asn³⁸⁴ resulted in a relatively rapid loss of activity for both glycosylation-deficient mutants. Interestingly, N84Q LCAT was inactive immediately following its purification from cell media, whereas freshly purified N384Q LCAT initially possessed higher activity than WT LCAT, but lost activity steadily over the course of months. The decrease in N384Q activity with time is shown as a percentage of WT LCAT activity in Fig. 2. Shortly after purification, N384Q was about 50% more active than WT LCAT, in agreement with previous studies performed on the mutant in cell media (13-16). Seven months later, N384Q had almost no measurable activity while WT LCATs activity was unchanged. Yet, during that same period, the apparent K_m of N384Q remained constant relative to WT LCAT apparent K_m (Fig. 2 and Table I). Thus, these results indicate that the removal of the glycan chain has drastically affected N84Q LCAT, rendering it enzymatically inactive. In the case of the N384Q mutant, removal of the glycan chain has destabilized N384Q such that its enzymatic activity decreases far more rapidly than WT LCAT, while its interfacial binding affinity remains unchanged.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Changes in enzymatic activity of N384Q LCAT with time. Enzymatic parameters of N384Q LCAT (kcat-acyltransferase assay (), Km-acyltransferase assay (), kcat-esterase assay (), Km-esterase assay ()) were measured at each time point in parallel with wild-type LCAT and expressed as a percent of the corresponding wild-type LCAT parameter.

The LCAT esterase assay measures the catalytic rate of ester bond hydrolysis for a water-soluble substrate, para-nitrophenylbutyrate. Because the reaction is independent of interfacial binding and activation steps, this assay is used mainly to detect changes in the catalytic core of LCAT. The same trends were observed in the esterase assay results as in the acyltransferase assay (Fig. 2 and Table I). The N84Q mutant was catalytically inactive less than a month after purification. The N384Q mutant initially had a greater activity than WT LCAT, but lost all measurable activity in a matter of months. WT LCAT, as usual, retained a constant activity over the same number of months. The K_m values of WT LCAT and N384Q LCAT remained relatively constant. Evidently, the Asn⁸⁴ glycan chain is essential for LCAT activity after purification, but the N84Q mutant possesses some activity while still in cell media (see thermal denaturation studies). The Asn³⁸⁴ glycan seems to be important in stabilizing LCAT activity for a period of longer than a few days.

Interfacial Binding Kinetics-- Surface plasmon resonance is a powerful method to study LCAT interfacial binding to rHDL. Interestingly, the N84Q mutant displayed no binding affinity at all for the rHDL. Coupled with the negative enzymatic results, this result confirms that the purified mutant is not functional. In contrast, the N384Q mutant binds to rHDL with similar affinity to WT LCAT. In fact, the binding of N384Q LCAT to rHDL held constant regardless of the age of the enzyme preparation, in agreement with the enzymatic assay results that the apparent K_m value did not change over time. These results applied equally to association and dissociation rate binding constants, and the resultant equilibrium dissociation constant, as seen in Table II.

Structural Studies-- Thermal denaturation studies were used to estimate the stability of the LCAT glycosylation mutants compared with WT LCAT. A decrease in enzymatic activity at progressively higher temperatures generally corresponds to the denaturation of the protein structure. The denaturation curves are shown in Fig. 3A. The T_m for WT LCAT was 48.1 °C, for N384Q LCAT was 47.5 °C, and for N84Q LCAT was 44.5 °C. The N384Q mutant denatured similarly to the WT LCAT, but the N84Q mutant demonstrated a decreased thermal stability compared with either the N384Q mutant or WT LCAT. The T_m of N84Q LCAT has shifted down by about 3.5 °C. This difference in T_m corresponds to a difference of around 0.5 kcal/mol of free energy change in calorimetry studies of other proteins (23). It is important to note that the N84Q mutant does have some activity at this point, although the same mutant has no activity after the protein purification procedure, as shown by the enzymatic activity assays. The complete lack of activity in the purified form of N84Q is likely related to the decreased stability of the enzyme, since the purification procedure is the same as that used for WT LCAT and N384Q LCAT.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Denaturation of WT LCAT and glycosylation mutants. Wild-type LCAT, solid line (); N84Q LCAT, dashed line (); N384Q LCAT, dotted line (). Panel A, thermal denaturation, monitored by acyltransferase activity measurements on cell culture media. Activity at 37 °C was assumed to correspond to native enzyme structure. Panel B, chemical denaturation of purified proteins by GdnHCl, monitored by fluorescence. The denaturation curves of N384Q LCAT at 1 week and 6 months were nearly superimposable, and so were combined for clarity. Panel C, chemical denaturation of purified proteins by GdnHCl, monitored by CD. The denaturation curves of N384Q LCAT at 1 week and 6 months old were combined. All data points are the average of at least two measurements. Plots and calculations were performed by Origin software.

The denaturation of LCAT was also followed by isothermal spectroscopic means. By adding progressively higher concentrations of GdnHCl, CD can be used to observe the disappearance of secondary structure and to determine the stability (or G_D°) of LCAT. In a complementary study, fluorescence spectroscopy was used to estimate the average environment of the tryptophan residues of LCAT, which reflect the tertiary structure of the protein, and to follow LCAT denaturation. As high concentrations of GdnHCl denature LCAT, the tryptophans become increasingly exposed to solvent, altering their fluorescence emission spectra and thus allowing observation of the progress of the denaturation.

According to the CD spectra of LCAT in buffer, the N84Q mutant possesses less -sheet structure but slightly more -helical structure compared with WT LCAT (Table III). The possible location of these altered structural motifs is considered in the discussion section. A fresh sample of the N384Q mutant also possesses more -helical structure than does WT LCAT. However, an older sample of N384Q LCAT undergoes a marked decrease in -helix content with a concurrent increase in the amount of unordered structure. From this data, it appears that N384Q LCAT, over a period of months, loses much of its helical structure, which unravels to become more disordered.

From the GdnHCl denaturation experiments, WT LCAT was found to have a G_D° of 2.0 kcal/mol. This value is low for globular proteins, which typically have G_D° values of 5-15 kcal/mol, but is consistent with previous measurements (24) and underscores LCATs flexibility and ability to undergo conformational changes. Both LCAT glycosylation mutants were found to have even lower stabilities, as shown in Fig. 3B and Table IV.

Little difference is seen in the wavelength of maximum fluorescence (WMF) and peak intensity between the fluorescence spectra of WT LCAT and N384Q LCAT, regardless of the age of the preparation (Table III). Evidently, the localized effects that alter the -helical structure of this mutant do not extend to regions containing tryptophan. The N84Q mutant, however, did display a significant blue shift in its WMF, indicating a more sheltered and hydrophobic environment for its tryptophan residues. In general, a blue shift suggests that a region of the protein containing tryptophan residues has become more compact or more sequestered from solvent exposure. The denaturation studies, as monitored by fluorescence, confirm much of what has been discovered by the other techniques. The N384Q mutant does not change in stability with age, but is consistently slightly less stable than WT LCAT (Fig. 3C and Table IV). The N84Q mutant, as expected, is significantly less stable than WT LCAT.

Interaction with ANS permitted a qualitative evaluation of the probe-accessible hydrophobic sites on LCAT. This experiment was intended to examine further the structural changes observed in the glycosylation mutants and especially the blue shift observed in the N84Q fluorescence results. ANS is a hydrophobic, fluorescent probe whose quantum yield increases and WMF decreases (blue shifts) upon noncovalently binding to a hydrophobic region in a protein. In buffer, ANS has little fluorescent intensity. The intensity and WMF do not change when incubated with tightly folded globular proteins such as carbonic anhydrase, which lacks specific binding sites for ANS (22). However, the emission maximum of ANS incubated with WT LCAT shifts to 470 nm; this value is typical of other lipid-binding proteins such as apolipoprotein A-I (22) and bovine serum albumin (25). As shown in Fig. 4, a relatively fresh preparation of N384Q is similar to WT LCAT in its interaction with ANS. However, as the N384Q preparation ages, the ANS spectrum shifts toward that of carbonic anhydrase. Interestingly, the N84Q ANS spectrum is also more similar to the spectrum of carbonic anhydrase than to the WT LCAT spectrum. These observations indicate that a hydrophobic region of WT LCAT accessible to ANS has become inaccessible or altered in the glycosylation mutants.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   ANS fluorescence spectra. ANS was added in an excess over protein of 100:1 mol/mol and incubated with the indicated proteins for 1 h. The "ANS" curve indicates the probe in buffer in the absence of any proteins. Fluorescence intensity (y axis) is uncorrected and is given in arbitrary units.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

One of the conclusions from the current studies is that the N84Q LCAT mutant is a less stable protein than WT LCAT. This conclusion is confirmed by the thermal denaturation study and both chemical denaturation studies. The decrease in stability (G_D°) is under 1 kcal/mol, which is not large in absolute value, but is a significant proportion of the measured stability of WT LCAT (1.6-2.0 kcal/mol). Probably this instability is the main cause of the absence of activity after purification of the N84Q mutant. In several previously published reports (13-16) and in the thermal denaturation studies, the N84Q mutant possessed weak but detectable enzymatic activity in cell media. However, after undergoing the same purification procedures as WT LCAT and the N384Q mutant, the purified form of N84Q was no longer active.

A change in structure may be responsible for the destabilization and loss of activity in the N84Q mutant. One possible explanation for the differences in N84Q LCAT structure compared with WT LCAT is that the mutant protein is misfolded. Glycosylation plays a role in the quality control processes in the endoplasmic reticulum and Golgi, so the lack of a glycan at position Asn⁸⁴ may allow a bypass of the usual folding inspection mechanism. However, the secreted protein is capable of normal, albeit reduced, enzymatic functions in cell media, which would not be likely if the protein were extensively misfolded.

Instead, the changes in structure may be a localized effect due to the absence of the Asn⁸⁴ glycan. One distinct possibility is that the proposed ₂-sheet of LCAT has become destabilized. In the LCAT computer model published by Rosseneu and colleagues (2), the ₂-strand, a structurally conserved element of the / hydrolase fold superfamily, spans residues Thr⁷⁹-Asn⁸⁴ (Fig. 5). The most striking difference in N84Q structure from WT LCAT, as reported by CD spectra, is the loss of -sheet content. While the ₂-strand has not been shown to play any significant role in LCAT activity, the region N-terminal to it has. The residues within the disulfide bridge of Cys⁵⁰-Cys⁷⁴ constitute LCATs lid region, a structural element conserved among lipases. This feature covers the entrance to the active site of lipases and facilitates interaction between substrate and enzyme by participating in interfacial binding and active site interactions with lipid substrates (26). Alteration of this region in LCAT, whether through deletion (27) or mutation (3), completely abolishes binding affinity for rHDL particles (21). Clearly, the lid region of LCAT is essential for both binding and activity with rHDL. Significantly, even as the N384Q mutant activity faded to zero over a period of months, it retained full binding affinity for rHDL, whereas the N84Q mutant was unable to bind rHDL at all, according to the surface plasmon resonance results. Therefore, the lack of a glycan chain at position Asn⁸⁴ may destabilize the ₂-sheet of LCAT, repositioning or altering the neighboring lid region and blocking interfacial binding. The explanation for the absence of N84Q activity in the esterase assay is less apparent, but the deletion of the lid region has also been shown to partially decrease esterase activity (27). Therefore, the lid region may also have some role in binding lipid substrates and analogs such as para-nitrophenylbutyrate and/or orienting them in the active site.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Sequences of LCAT in the N-terminal and C-terminal regions. The Asn84 and Asn384 glycosylation sites are boxed. In the N-terminal sequence, tryptophan residues are in bold, the lid region has a dashed underline, and the proposed 2-sheet is underlined and in bold. In the C-terminal sequence, the catalytic histidine is marked with an asterisk and in bold, the proposed -His helix is underlined and in italics, and O-glycosylation sites are marked with an "o" and are in bold.

An alternative explanation for the lack of activity in the N84Q mutant involves the positioning of the active site residues. According to the computer model, the catalytic Asp³⁴⁵ residue is located only 3.4 Å from Arg⁸⁰ in the ₂-strand (4). Thus, even a slight rearrangement of the ₂-structure might lock Asp³⁴⁵ into a salt bridge with Arg⁸⁰, making it unavailable for catalysis.

Three observations about N84Q LCAT indicate that it may have a more compact structure than WT LCAT. Fluorescence results show that N84Q has a blue-shifted spectrum compared with WT LCAT, meaning that, on average, tryptophan residues are buried deeper in the hydrophobic interior of the protein and are less accessible to solvent. Unfortunately, LCAT contains 12 tryptophans, and it is not possible to distinguish which have been affected by the N84Q mutation. However, it is worth noting that the sequence close to the lid region of LCAT contains three tryptophans, making it a possible candidate for the proposed structural changes. A second observation relates to the shape of the denaturation curves of the N84Q mutant. As monitored by both CD and fluorescence, the N84Q denaturation curves are not only shifted in terms of [GdnHCl]_1/2, but also slightly steeper in slope (m). A study by Byrne and Stites (23) proposes that a change in the slope of a denaturation curve may be related to the amount of hydrophobic area exposed during the course of denaturation of a protein. A steeper slope indicates that a protein exposed more hydrophobic area during denaturation, implying that it started with less of that hydrophobic area exposed. Finally, N84Q LCAT bound ANS less effectively than WT LCAT did. A consideration of the principles of ANS binding would be helpful before elaborating on this topic.

ANS will bind to a protein only under certain conditions. First, the protein may contain a hydrophobic cavity of sufficient size to accommodate ANS. The fatty acid-binding sites of bovine serum albumin (25) and the heme-binding site of apomyoglobin (28) are both examples of cavities which bind ANS. Second, the protein may be in the molten globule state. ANS is often used to detect folding intermediates or follow the denaturation of a protein. It will also bind apoA-I, which, due to its flexibility and clustering of hydrophobic residues (29), bears some resemblance to a molten globule (30). Third, the protein may possess an extremely hydrophobic surface. Although this mode of binding has been demonstrated with erythropoietin (31), it has not been commonly observed, even among other hydrophobic proteins.

The fact that N84Q LCAT was unable to interact with ANS suggests that a hydrophobic environment found in WT LCAT has been made unavailable for ANS binding. In principle, the removal of a glycan chain might expose the hydrophobic surface area of a protein to solvent (31) or decrease the stability of that region (32). In both of these cases, however, ANS would tend to bind N84Q LCAT to a greater degree than WT LCAT, if its binding changed at all. Instead, the lesser degree of binding observed indicates that a hydrophobic cavity in WT LCAT is no longer accessible to ANS in the N84Q mutant. The Asn⁸⁴ glycosylation site has already been shown to be proximal to the lid region, which is located at the entrance to the active site. Therefore, the removal of the Asn⁸⁴ glycan chain may have altered the active site cavity such that it either has become inaccessible to ANS or has collapsed. This hypothesis would explain both the lack of enzymatic activity and the more compact structure observed in N84Q LCAT compared with WT LCAT.

The denaturation curves of all species of LCAT are unusually shallow for a globular protein, indicating a noncooperative unfolding mechanism. In other words, many regions of the protein may unfold independently, in contrast to the abrupt collapse of structure typical of many other globular proteins. Because of this independence, one region of LCAT may be destabilized or unfolded, but may not be detected by a technique which focuses on secondary structure or tryptophan environment changes. This phenomenon may help explain the differences in the structural results obtained for the aging N384Q mutant. Both the WMF of N384Q LCAT and the denaturation profile of N384Q LCAT, monitored by fluorescence, remain constant, regardless of the age of the preparation. However, if structural changes were occurring near Asn³⁸⁴, they would not be detected by fluorescence methods, for this region is devoid of tryptophan residues.

The CD results show first a slightly increased -helix content in a fresh sample of N384Q protein, and then a dramatic decline in helical content months later. This trend mimics the trend in the enzymatic activity for N384Q, which is initially higher than for WT LCAT in cell media and in fresh preparations of enzyme, and then falls steadily to near zero within months. A modification of the -His helix of LCAT may be sufficient to rationalize these results. The -His helix, which has been modeled to be proximal to the Asn³⁸⁴ glycosylation site of LCAT (Fig. 5), is a conserved element of / hydrolases and is located near the catalytic core. This helix has already been shown to be essential for LCAT activity, because C-terminal truncations which progress further than residue 398 bring about a total loss of enzymatic activity (18, 33). Therefore, a progressive disruption of this helix seems likely to have an effect on LCAT activity.

According to ANS spectra, a 2-month-old preparation of N384Q LCAT resembles WT LCAT, but a 6-month-old preparation resembles more closely the compact structure of the N84Q mutant, which has little binding affinity for ANS. As argued for the N84Q mutant, the decrease in ANS binding may indicate that the active site cavity of N384Q LCAT is becoming inaccessible to ANS or is collapsing. If the catalytic residues in the active site were inaccessible to ANS, the same may also be true for lipid substrates, and the enzymatic activity of N384Q would correspondingly decrease. Therefore, the removal of the Asn³⁸⁴ glycan may modify the -His helix and surrounding structure, restricting access of substrates to the active site.

In addition, one of the regions affected by the deletion of the Asn³⁸⁴ glycan may include the catalytic histidine (His³⁷⁷). The decline in esterase activity mirrors the decline in acyltransferase activity, suggesting that the catalytic mechanism of LCAT is affected. The involvement of His³⁷⁷ is plausible, for this catalytic residue is only seven amino acids away from the glycosylation site (Fig. 5). If the structure of LCAT were altered upon removal of the glycan at Asn³⁸⁴ such that this key residue was repositioned, the catalytic activity of LCAT would drop accordingly.

Elucidating the greater activity of N384Q over WT LCAT in cell media was one of the motivations in performing these studies. This increased activity may be attributed to the consistently lower apparent K_m values in the acyltransferase assay of the N384Q mutant over WT LCAT (Table I), suggesting a more effective binding of the mutant to rHDL. A more effective binding could be due to a decrease in the steric hindrance of binding, once the bulky glycan chain is removed. However, the fact that this difference in apparent K_m was observed in the acyltransferase assay, but not in the equilibrium dissociation constant measured by the surface plasmon resonance studies raises doubts about the interfacial binding hypothesis. Instead, it suggests that the Asn³⁸⁴ glycan may have some inhibitory effect on the access or binding of the lipid substrates to the active site of LCAT. This possibility has some merit, for in an analogous enzyme, cholesterol esterase, the -His helix forms part of the cholesterol binding pocket (34). However, it should be stressed that apparent K_m values are not a direct measure of the binding of LCAT to the rHDL surface and should be interpreted with great caution.

One important property of the N384Q mutant which does not change with time is its binding to rHDL. As measured by the acyltransferase assay and surface plasmon resonance, the binding of N384Q LCAT to rHDL particles is the same whether it has maximal or minimal enzymatic activity. The independence of these two results indicates that the protein regions responsible for these two activities are well separated.

The general conclusion from this work is that the deletion of individual glycan chains destabilizes LCAT. However, the effects of this destabilization take different forms depending on the region of the protein. The N84Q mutant is destabilized in the sense that it is unable to withstand the procedures of purification, becoming inactive, and is more susceptible to thermal and chemical denaturation than WT LCAT. The destabilization of its structure affects both interfacial binding affinity and enzymatic activity, probably by altering key catalytic and binding elements near the glycosylation site. The result is a protein which has apparently collapsed upon itself to alter or eliminate the hydrophobic active site cavity, rendering the enzyme inactive. The N384Q mutant is destabilized only slightly relative to WT LCAT as observed by chemical and thermal denaturation. However, starting from a higher activity, it loses its activity much more rapidly than does WT LCAT, probably due to loss of key -helical elements involved in the active site cavity or due to rearrangement of catalytic residues. After months of storage, this mutant also loses its ability to bind ANS. After performing a literature search, no other comparable study could be found tracking the effects of destabilization over time. Although examples are plentiful of studies describing the effects of glycosylation on enzyme activity, structure, or stability, none follow up these observations after extended periods. Therefore, the N384Q LCAT mutant is a unique example of localized, gradual destabilization due to deglycosylation. Without utilizing more sophisticated techniques such as NMR or x-ray spectroscopy, it is not possible to describe the effects of glycosylation on LCAT at the atomic resolution level. However, the current studies have provided evidence that glycosylation affects the stability of localized protein structures which in turn affects enzyme functions such as catalysis or interfacial binding. They have also provided a rare example of how deglycosylation affects a protein's stability over time, independent of additional chemical, thermal, or pressure denaturation.

	ACKNOWLEDGEMENTS

We thank Dr. Shanthi Adimoolam for constructing the WT LCAT-expressing CHO cells and for assistance with cell culture techniques. We also thank Dr. Henry Pownall, Baylor College of Medicine, for the gift of anti-LCAT antibodies, Dr. Lihua Jin for the instruction in surface plasmon resonance measurements, and Adam Kennedy and Jeanne Danes for excellent technical assistance.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant HL-29939 (to A. J.) and by a predoctoral fellowship from the American Heart Association, Illinois Affiliate (to J. K.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Dept. of Biochemistry, College of Medicine at Urbana-Champaign, University of Illinois, 506 South Mathews Ave., Urbana, IL 61801. Tel.: 217-333-0452; Fax: 217-333-8868; E-mail: a-jonas@uiuc.edu.

Published, JBC Papers in Press, August 2, 2001, DOI 10.1074/jbc.M104326200

	ABBREVIATIONS

The abbreviations used are: LCAT, lecithin-cholesterol acyltransferase; rHDL, reconstituted high density lipoproteins; CHO, Chinese hamster ovary; GdnHCl, guanidine hydrochloride; ANS, 1-anilinonapthalene-8-sulfonate; WMF, wavelength of maximum fluorescence; G_D°, free energy of denaturation; POPC, 1-palmitoyl-2-oleoyl phosphatidylcholine; DPPC, dipalmitoylphosphatidylcholine; PAGE, polyacrylamide gel electrophoresis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This Article Abstract Full Text (PDF) All Versions of this Article: 276/40/37230    most recent M104326200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Kosman, J. Articles by Jonas, A. Search for Related Content PubMed PubMed Citation Articles by Kosman, J. Articles by Jonas, A. Social Bookmarking                      What's this?