| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 7, 6171-6181, February 13, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Veterans Affairs Greater Los Angeles Healthcare System, Los Angeles, California 90073 and the Department of Medicine, UCLA, Los Angeles, California 90095
Received for publication, September 10, 2003 , and in revised form, October 24, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
= 2 h) and is rapidly secreted. The remaining inactive mass is degraded. HL maturation occurs in the ER and is strongly dependent on binding to calnexin in the early co-/post-translational stages. Later stages of HL maturation occur without calnexin assistance, although inactive HL at all stages appears to be associated in distinct complexes with other ER proteins. Thus, unlike other lipases in the gene family, HL maturation is the rate-limiting step in its secretion as a functional enzyme. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Despite the disparate anatomical location and distinct physiological functions of the lipases in this gene family, sequence homologies and gene structure strongly suggest a common folding pattern. HL, LPL, and EL are particularly closely related, whereas PL, PS-PLA1, and lipase H appear to have branched out earlier from the primordial progenitor (1-3, 5, 13). In particular, HL, LPL, and EL share homologous heparin-binding domains and also exhibit two conserved N-linked glycosylation sites, located in the N-terminal and in the C-terminal regions, respectively. At least for LPL and HL, it has been shown that the presence of the conserved N-terminal glycan was obligatory for the formation of active lipase (14, 15). Human HL is distinct from HL in other species, since it contains two additional N-linked glycan chains, whereas EL displays three other potential, but nonhomologous, glycosylation sites.
The requirement for glycosylation at the conserved N-terminal site for optimal HL and LPL enzymatic activity implies that these lipases associate with the lectin chaperones calnexin and/or calreticulin. Nascent proteins bind to these chaperones via the innermost glucose residue of the oligosaccharide chain, following removal of the outer two glucose residues by ER glucosidases I and II (16). Indeed, both HL and LPL activity were inhibited by drugs that prevent trimming of the outer glucose molecules (17-20), and direct evidence of binding of HL to calnexin was also demonstrated (21).
Besides association with calnexin, we have recently investigated additional aspects of LPL maturation (i.e. the process of acquisition of enzymatic activity by newly synthesized lipase) (22). We have demonstrated the concurrent formation of two different LPL pools. One pool became fully active; its maturation occurred in the ER and was rapid, to the extent that no inactive precursor to the active form could be detected at steady state or in pulse-chase experiments. The second LPL pool consisted of inactive, misfolded lipase molecules that associated into large aggregates via interchain disulfide bonds. In accordance with strict quality control, these aggregates were degraded in the ER, whereas the active LPL molecules were readily secreted.
The close homology between LPL and HL suggested that the latter might exhibit similar maturation properties, a hypothesis that was investigated in the present study. Our results demonstrate that, like LPL, part of nascent HL remains misfolded and is degraded in the ER, whereas the pool of HL that reaches a functional conformation is secreted. However, in stark contrast to the rapid maturation of LPL, HL achieves an active conformation after prolonged residence in the ER, during which it remains associated with calnexin and with an apparent array of other ER factors in complexes of varied size.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Expression Constructs—The cDNA of human HL and PL was inserted into the pcDNA6/V5-His expression vector (Invitrogen) as previously described for the human LPL construct (22). The termination codons (nucleotides 1,498-1,500 of HL and 1,396-1,398 of PL) were replaced with a BamHI site and cloned into the expression vector as a HindIII/BamHI fragment. This placed the 3'-end of human HL and PL in frame with the V5 epitope tag, which was subsequently used for antibody detection of the expressed lipase proteins. For HL, the in-frame cloning replaced the natural termination codon with glycine followed by a 50-amino acid C-terminal adduct: 501STSPVWWNSADIQHSGGRSSLEGPRFEGKPIPNPLLGLDSTRTGHHHHHH550 (the V5 epitope tag is underlined). For PL, the natural termination codon was replaced with leucine followed by the 50-amino acid C-terminal adduct: 467GDPLVQCGGILQISSTVAAARGHPFEGKPIPNPLLGLDSTRTGHHHHHH516. The constructs were driven by a cytomegalovirus promoter and contained a bovine growth hormone polyadenylation site.
Transfection, Selection, and Harvesting of Cells—CHO cells were stably transfected with the pcDNA6/V5-His vector containing the cDNA of HL, LPL, and PL, using the calcium phosphate kit and instructions provided by Invitrogen. The selection medium contained 10 µg/ml blasticidin. Lipase-expressing colonies were identified by assaying the medium for HL activity (23), after incubating the cells for 16 h in the presence of 10 units/ml heparin.
Cells were subcultured onto 100-mm plates and kept under blasticidin selection. Prior to experiments, cells were incubated for 6-12 h in medium without blasticidin, containing 10 units/ml heparin.
HepG2 cells were transiently transfected with the pcDNA6/V5-HL expression vector using the FuGene 6 transfection agent (Roche Applied Science) with the protocol provided by the manufacturer. Six hours prior to harvesting, fresh medium containing 10 units/ml heparin was added. The cells were harvested 24 h after transfection.
For harvesting, cell monolayers were washed twice with phosphate-buffered saline (PBS) and scraped off in 1 ml of PBS. The cell slurry was centrifuged at 1000 x g for 5 min at 4 °C, and the resulting cell pellets were stored at -80 °C. For analysis, pellets from individual plates were suspended in 0.6 ml of lysis buffer (10 mM Tris-HCl, pH 7.5, containing 0.2% sodium deoxycholate and 10 units/ml heparin) and sonicated for 6 s at a force of 2 g. Cellular debris was removed by centrifugation at 1,000 x g for 5 min at 4 °C.
Chromatography and Sucrose Gradient Centrifugation—For heparin-Sepharose chromatography, 1 ml of heparin-Sepharose (Amersham Biosciences) was packed into a 1 x 10-cm column and equilibrated with column buffer (10 mM Tris-HCl, pH 7.5, containing 0.1% Triton X-100). Combined cell lysates from six 100-mm plates were applied to the column, and the flow-through was collected. After washing the column with 10 bed volumes of column buffer, HL was eluted in two stages: the addition of 0.5 M NaCl, followed by 1.0 M NaCl in column buffer. Heparin was added to all eluted fractions to a final concentration of 10 units/ml, and samples were maintained at -80 °C until analyzed.
-Ricin chromatography of the active, high affinity HL fraction obtained from heparin-Sepharose chromatography was performed after removing the terminal sialic acid of the glycan chains by neuraminidase digestion (see below). 0.5 ml of Ricinus communis agglutinin (RCA120) linked to agarose (Sigma) were washed with 10 mM Tris-HCl, pH 7.5, containing 0.1% Triton X-100 and 1.0 M NaCl. After application of the sample, unbound HL was collected in the flow-through, and the bound HL was eluted with 0.2 M galactose (Sigma) in the column buffer.
Rate zonal centrifugation was carried out using a 12-ml, 5-20% linear sucrose gradient, topped by 0.55 ml of sample. The molecular markers used were ovalbumin (45 kDa), malic dehydrogenase (74 kDa), glucose 6-phosphate dehydrogenase (114 kDa), and catalase (240 kDa). Centrifugation was carried out at 200,000 x g for 22 h at 4 °C. Fractions of 0.48 ml were manually collected from the top, as described (22).
Glycosidase Digestion—For endoglycosidase H (endo H) digestion, samples were denatured by adding SDS to 0.5% and heated at 95 °C for 2 min. After chilling, 10 milliunits of endo H (Roche Applied Science) were added, and the samples were incubated at 37 °C for 16-18 h.
For neuraminidase digestion, 50 milliunits of neuraminidase (Calbiochem) were added to 1 ml of HL eluate obtained from heparin-Sepharose chromatography. Incubation was carried out for 90 min on ice prior to -ricin chromatography.
Lipase Assays and Determination of Specific Activity—The activity of HL and LPL was measured using the respective triolein substrates, prepared by sonication (24, 25). PL was assayed with the substrate used for HL, with the following modifications. A mixture of 0.5 µg of colipase (Sigma), 0.8 µmol of taurodeoxycholate (Sigma), and 0.2 µmol of CaCl2 was prepared in a final volume of 50 µl in water. Enzyme source and 10 mM Tris-HCl, pH 7.5, were then added to complete the volume to 100 µl, and the reaction was initiated by the addition of 100 µl of substrate. One milliunit of lipase activity is defined as 1 nmol of free fatty acid (FFA) hydrolyzed per min.
Lipase specific activity was calculated by dividing the activity loaded onto electrophoretic gels by the densitometric value of the band obtained by Western blotting. Since densitometry does not provide an absolute concentration, we report the specific activity relative to that of a sample chosen from the blot that was arbitrarily set to a value of 1.0.
In instances where HL was fully active, 1 milliunit of active lipase was considered to represent 1 ng of lipase protein. This value was supported by the reported specific activity of purified HL as 55,000 µmol of FFA/mg/h (26), corresponding to 0.92 nmol of FFA/ng/min, or nearly 1 milliunit/ng. This correlation of 1 milliunit/1 ng of lipase protein was applied to HL that was secreted or to active intracellular HL with high affinity to heparin (see Figs. 3C and 4A). Specifically, in Fig. 3C, the densitometric value of HL mass secreted after 2 h was converted to the value of its milliunits of enzymatic activity, and the mass of the other samples on the plot was reported relative to that value. Similarly, in Fig. 4A, the densitometric mass value of the peak active fraction eluted at 1.0 M NaCl was converted to the value of milliunits of activity, and the mass of all other samples eluted from the column was reported relative to that value.
|
|
|
PAGE and Western Blotting—Unless otherwise stated, cell lysates, medium, or proteins released from co-immunoprecipitates were SDS- and heat-denatured (95 °C for 2 min) in the presence of 0.4 M -mercaptoethanol (Sigma) and subjected to PAGE using 7% Tris-glycine gels. For cross-linked lysates containing multicomponent complexes, electrophoresis was carried out without reducer in 3-8% gradient Tris acetate gels (Novex). Western blotting was performed as previously described (24), using the horseradish peroxidase-conjugated anti-V5 antibody (Invitrogen; 1:2,500). Quantitation of the lipase bands was carried out by densitometric scanning of the Western blots using the NIH Image analysis program for the Macintosh.
Cross-linking of HL—Cells in 100-mm plates were washed twice with PBS at room temperature, followed by incubation for 20 min on ice with cold PBS containing 20 mM N-ethylmaleimide (NEM; Calbiochem). After an additional wash with cold PBS without NEM, cells were harvested, resuspended in PBS (0.45 ml/cell pellet obtained from one plate), and kept on ice. 50 µl of a freshly prepared solution containing 20 mM dithiobis(succinimidylpropionate) (DSP; Pierce) in Me2SO were added to the cells (2 mM final concentration), and the slurry was incubated on ice for 1 h with intermittent shaking. In control plates, an equivalent amount of Me2SO without DSP was added, and the cells were incubated in a similar manner. The cross-linking was quenched by the addition of 1 M glycine (50 mM final concentration) and 1 M NEM (40 mM final concentration) and an additional incubation on ice for 15 min. The cells were pelleted by centrifugation at 12,000 x g for 3 min.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Cells Exhibit Massive Amounts of Inactive HL Protein—Fully mature HL is a homodimer, with a predicted molecular mass of 120 kDa. To compare the assembly state of intracellular HL with that of the mature, secreted form, cell lysates and medium containing similar amounts of HL activity were subjected to density gradient centrifugation. As expected, HL from both sources displayed a superimposable distribution of enzymatic activity, sedimenting at a region corresponding to a dimer (103 kDa; see Fig. 1, graph). However, the similarity between cells and medium did not extend to the HL mass. As shown in Fig. 1 (bottom panels), while the medium exhibited barely detectable amounts of HL protein that overlapped with the activity (fractions 10-12), comparable amounts of activity in the corresponding fractions of the cell lysate exhibited massive amounts of HL protein. Thus, the intracellular HL specific activity (activity/units of HL mass) was very low. In addition, rather than sedimenting exclusively as a dimer, intracellular HL protein was spread almost continuously across the gradient, including large aggregates at the bottom (fraction 26). Clearly, the vast majority of HL mass in the cells was not functional and exhibited a diverse range of molecular weights.
|
In contrast to HL, the relatively high specific activity of PL and LPL suggested that the intracellular lipase protein contained in large part mature, fully functional molecules. This distinction was evident in the sedimentation pattern of the three lipases. As shown in Fig. 2B, a perfect correlation existed in the sedimentation profile of PL enzymatic activity and protein, corresponding to that of an expected monomer (11). The molecular size of the peak fraction was 
59 kDa, in close agreement with the predicted PL molecular mass of 53 kDa. LPL intracellular protein, as was previously described (22), was separated into the fully mature dimer form that overlapped the enzymatic activity and an inactive aggregate at the bottom of the gradient. In contrast, the sedimentation pattern of HL was not so clearly defined. Although the majority of HL protein migrated to a region that could loosely be interpreted as a homodimer, it was associated with very little enzymatic activity. Therefore, protein accumulation at this sedimentation range could at best represent a mixture of some homodimeric, active HL, and inactive HL that is either dimeric or monomeric in association with proteins of similar molecular weight. In addition, as seen here and in Fig. 1, HL protein trailed into larger molecular forms. Clearly, these cells contained very little HL with a fully mature conformation.
Although PL, LPL, and HL are closely related lipases and their tertiary structure (especially that of LPL and HL) is presumably very similar (31, 32), we nevertheless considered the possibility that CHO cells specifically lacked the ability to fold HL. Therefore, we examined the expression of HL in HepG2, a human hepatoma cell line that endogenously expresses HL (18). In order to analyze HL mass, HepG2 cells were transiently transfected with the V5-tagged HL construct. The expression levels of the transfected protein were low, and little activity above that of endogenous HL was detected. Nevertheless, sucrose gradient centrifugation revealed even in these cells the familiar widespread distribution of the HL protein (Fig. 2B, bottom panel). Similar sedimentation patterns of HL protein were also observed in mouse cell lines derived from fetal liver or from skin explants, under different levels of expression (data not shown). These results clearly indicate that, regardless of cell type, maturation of HL per se is distinctively different from that of other members of the lipase gene family.
Inactive HL Is a Precursor of the Native Enzyme—The presence of vast amounts of inactive HL alongside minimal amounts of active, mature HL suggested that the inactive component represents, at least in part, intermediate(s) in the maturation of the nascent polypeptide. To test this hypothesis, protein synthesis was arrested by incubating cells in the presence of cycloheximide (Cx), and the fate of the preexisting HL activity and mass was followed.
Even after a prolonged incubation in Cx (8 h), little change occurred in the small amounts of intracellular HL activity (Fig. 3A). Surprisingly, for the first 1 h or so, the secretion of active HL from these cells was also comparable with that of control cells. The effect of Cx became apparent only during the second hour, as secretion of activity gradually tapered off. After about 4 h, no more protein was secreted, and the amount of HL activity in the medium remained constant for the rest of the incubation. By that time, however, total HL activity in the cells and medium generated in the presence of Cx had increased almost 6-fold over the initial activity and was only 40% lower than the total activity in control cells.
The HL activity generated de novo under conditions of arrested protein synthesis must have originated from preexisting inactive HL. This was confirmed by following the fate of the HL protein in Cx-treated cells and medium (Fig. 3B). The gradual disappearance of HL protein from the cells was detectable even after the first 15 min of incubation in Cx and continued throughout the experiment (upper panel). Concomitantly, HL protein appeared in the medium (lower panel), indicating a precursor-product relationship.
An accurate view of the dynamics of HL secretion and turnover was obtained by plotting, alongside the activity, the amounts of HL mass shown in Fig. 3B. As shown in Fig. 3C, it became apparent that 50% of the inactive mass present in the cells at time 0 was secreted (see closed and open squares), the remainder of the inactive mass undergoing degradation. Further, the secretion rates of HL activity and mass were closely correlated (open squares and circles), indicating that HL mass was released from the cells only after the inactive protein matured into an active, secretable form. Finally, the rates of HL disappearance from the cell and appearance in the medium were similar, with a t of 2 h (117 versus 123 min, respectively). Since the intracellular activity (i.e. mature HL) remained at very low steady state levels (see dark circles), we conclude that HL maturation is a slow, rate-limiting process relative to secretion. When the protein reaches an active conformation, it is rapidly secreted, whereas HL protein that does not acquire a secretable conformation is rapidly degraded within the cell.
Maturation of HL Occurs in the ER—Considering that the ER is the main compartment for quality control, we hypothesized that the immature HL precursor undergoes folding and assembly in the ER. Thus, at least part of the small HL enzymatic activity detected in cell lysates was derived from HL located in the ER.
To test this hypothesis, it was necessary to isolate the active HL from the massive amounts of inactive protein residing in the cell and to establish its intracellular location. We utilized the fact that native HL has a high affinity to heparin and is eluted at ionic strength concentrations exceeding 0.75 M (33). Thus, the first step in the isolation of active HL was heparin-Sepharose chromatography. As shown in Fig. 4A, the vast majority of HL protein (black circles) did not attach to the heparin-Sepharose matrix and remained in the flow-through. Increasing the ionic strength to 0.5 M NaCl resulted in further release of HL protein. In the final step, when the ionic strength was increased to 1.0 M NaCl, a very small amount of HL protein emerged. However, HL activity (gray triangles) relative to its mass (i.e. enzyme-specific activity) was low in the flow-through, somewhat higher in the 0.5 M NaCl wash, and highest in the 1.0 M NaCl eluate (0.02, 0.05, and 1.00, respectively). Based on its high affinity to heparin and the high specific activity, the 1.0 M NaCl eluate was assumed to represent the mature, functional form of HL.
The intracellular localization of the fractions eluted from the heparin column was determined by endo H sensitivity. Glycans sensitive to endo H digestion contain a high mannose structure, characteristic of glycoproteins residing in the ER. As expected, all fractions contained high mannose glycan chains, indicative of their localization within the ER (Fig. 4A, see panels below peak fractions). However, prolonged exposure of the film revealed that only the 1.0 M NaCl eluate contained, in addition to the predominate endo H-sensitive pool, a small amount of protein containing endo H-resistant glycans (see separate strips below respective fractions). Thus, as evidenced by the very low levels of HL glycans exhibiting endo H-resistant forms, only very small amounts of HL protein reside within Golgi or post-Golgi compartments.
To rule out the possibility that the enzymatic activity measured in the 1.0 M NaCl eluate was attributable only to this small amount of HL protein in the Golgi, this fraction was separated from HL located in the ER by -ricin chromatography. As -ricin toxin specifically binds galactose (the penultimate sugar residue in Golgi-processed glycoproteins), the terminal sialic acid of HL glycans was removed by neuraminidase digestion. Following application of this sample onto a -ricin-agarose column, the flow-through was recovered, and the Golgi-derived HL was eluted with 0.2 M galactose. As shown in Fig. 4B, the majority of the protein remained in the flow-through of the -ricin column and was entirely endo H-sensitive. Again, no protein could be visualized in the 0.2 M galactose eluate unless the film was exposed for over 1 h. Thus, the vast majority of the HL with high affinity to heparin resided in the ER. To ascertain that this ER-residing high affinity fraction was fully active, we attempted to compare its relative specific activity with that of HL secreted from the cell. However, since processing of the four glycan chains of HL in the Golgi is not carried out to completion and some high mannose glycans are maintained, the migration of secreted HL during PAGE electrophoresis was heterogeneous, thus complicating its quantitation by densitometry. Therefore, we included in the analysis HL secreted from Lec1 cells, mutant CHO cells secreting glycoproteins that uniformly contain Man5GlcNac2 glycan chains (34). As shown in Fig. 4C, equal amounts of HL activity from the -ricin flow-through, CHO medium, and Lec1 medium displayed similar amounts of mass, corresponding to comparable specific activities. Thus, HL maturation occurs in the ER, and, since very little HL was detectable in post-ER compartments, we conclude that upon reaching maturity the functional HL is rapidly secreted.
Castanospermine Inhibits Maturation of Newly Synthesized HL—The maturation of HL in the ER, like that of other glycoproteins, depends on the partial glucose trimming of its carbohydrate side chains. Thus, in the presence of castanospermine or N-methyldeoxynojirimycin, drugs that inhibit the activity of glucosidases I and II, HL activity and secretion were shown to be impaired (17, 18, 21). However, considering that inactive HL protein acquired activity over a period of hours (Fig. 3), it was not clear whether the glucosidase inhibitors affected the maturation of the HL precursor preexisting at the time of their addition or only that of HL synthesized de novo in their presence.
To address this question, cells were incubated with castanospermine (Cst) for 6 h, and HL activity was assessed relative to control cells (Fig. 5). To monitor the normal maturation of the preexisting HL pool, HL secreted in the presence of Cx was included in the analysis. During the first 2 h in Cst, there was little change in HL activity compared with control cells, and under both conditions, HL activity was only slightly higher than that secreted from Cx-treated cells (compare dark circles and triangles). Thus, it appeared that Cst did not affect the maturation of the preexisting inactive HL pool. After 2 h, however, the inhibitory effect of Cst gradually increased. When the gain in activity derived from maturation of the precursor (i.e. activity secreted by Cx-treated cells; see open squares) was subtracted from the activity secreted by Cst and control cells, the inhibitory effect of Cst was significant, with less than 30% activity remaining. Thus, when glucose trimming of newly synthesized protein was prevented, maturation of this form was significantly compromised.
|
The unperturbed maturation of the preexisting HL precursor in the presence of Cst was confirmed when Cx-treated cells were incubated for 6 h in the absence or presence of Cst (Fig. 5C). The increase in activity due to HL maturation was similar under both conditions. Thus, glucose trimming must be crucial in the initial co- or post-translational stages but has no apparent effect on later stages of HL maturation.
Robust Interaction of HL with Calnexin—The requirement of newly synthesized HL to undergo glucose trimming for acquisition of enzymatic activity supported the notion that HL associates with the lectin chaperone calnexin (Cnx). However, although LPL is also dependent on glucose trimming for acquisition of an active conformation (35), we were unable to demonstrate a significant LPL-calnexin association unless the cells were under stress (22). It was of interest, therefore, to compare side by side the binding ability of the two lipases to calnexin. To show that association to calnexin was lipase-specific, PL was included in the analysis as a negative control, since this lipase is not dependent on mammalian calnexin for full expression of activity (36). Indeed, as shown in Fig. 6A, when comparable amounts of lipase protein were immunoprecipitated with anti-Cnx antibody, only HL associated in appreciable amounts to the chaperone. The robust binding of HL, but not of LPL, to Cnx is in agreement with the data shown above, suggesting that HL alone is undergoing slow maturation, presenting at steady state nonnative conformations that are associated with the chaperone.
|
The heterogeneous sedimentation of nonnative HL into increasingly larger molecular forms (Figs. 1 and 2) raised the question of whether HL associated to Cnx in solitary heterodimers (i.e. a single HL and a single Cnx molecule) or whether the two molecules also coexisted in higher molecular forms. To address this question, HL levels in calnexin co-immunoprecipitates were determined across the gradient fractions and expressed relative to molecular weight (Fig. 6C). The majority of HL in the Cnx immunoprecipitates sedimented at a relatively discrete peak centered at 110 kDa, closely corresponding to a 1:1 HL-Cnx association (53 + 66 kDa (37), respectively). In addition, HL was associated with Cnx in higher molecular forms, namely in the broad "shoulder" (150 kDa) descending from the main peak, in the 250-kDa region, and also in the largest aggregate at the bottom of the gradient. Thus, although most HL was bound to Cnx in solitary homodimers, Cnx was also a component of some of the larger HL-containing complexes.
HL Is a Component of Distinct Heterogeneous Complexes in the ER—The HL-Cnx association in complexes of relatively discrete sizes suggested that the broad distribution of immature HL in density gradients (Figs. 1 and 2) represented HL in an array of distinct, increasingly larger complexes. This was confirmed when the cells were treated with DSP, a membrane-permeable, thiol-cleavable cross-linker. As shown in Fig. 7 (upper panel), when lysates of DSP-treated cells were subjected to density centrifugation and analyzed in nonreduced SDS gels, the amount of monomeric HL in the fractions was significantly diminished. Instead, HL appeared in ever larger, distinct protein bands. Only toward the bottom of the gradient were the protein bands diffuse, possibly due to the difficulty of these large complexes in penetrating into the gel. In nonreducing gels, in the absence of cross-linker, HL migrated in the gradient fractions entirely as a monomer (lower panel), in a pattern indistinguishable from that displayed by HL in reduced gels (Figs. 1 and 2). Thus, interchain disulfide bonds did not play a role in the association of HL with other proteins in the array of complexes. Taken together, these results indicate that the slow maturation of HL is occurring through multiple associations of the lipase with other proteins, most likely chaperones, such as calnexin.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
The significantly lower specific activity of intracellular HL compared with the secreted fraction has been previously recognized (18, 38, 39). However, the relevance of the low specific activity to slow HL maturation was best understood when HL was compared with PL and LPL in terms of intracellular inactive mass and specific activity (Fig. 2).
PL expressed in CHO cells displays very high activity levels and a complete correlation in the sedimentation profile between mass and activity, suggesting that PL becomes fully active very soon after its synthesis. Moreover, PL displays full enzymatic activity even when expressed in lower eukaryotes, such as insect and yeast cells (36, 40). Additionally, PL activity is not affected by the murine cld/cld mutation, which virtually abolishes the activity of LPL and HL (35, 41). Thus, PL must acquire full catalytic activity with little assistance from the chaperones and folding factors employed by HL and LPL. On the other hand, the maturation of LPL is more discriminatory. Even in mammalian cells, whereas part of newly synthesized LPL protein acquires a functional conformation, the remainder is associated into misfolded aggregates that are degraded in the ER (22) (Fig. 2B). In insect cells, the expression of LPL activity, which is very low, increases only if it is co-expressed with calreticulin and, to a lesser degree, calnexin (42). Clearly, maturation of LPL requires the assistance of mammalian chaperones, and even in their presence the degree of successful folding and assembly is not complete. The maturation of HL appears to be even more complex. In all mammalian cell lines tested as well as in primary hepatocytes (38), only negligible amounts of intracellular HL protein were enzymatically active, whereas the majority of the lipase lacked activity and was improperly assembled (Figs. 1 and 2).
Slow Maturation of an Inactive HL Precursor in the ER—About one-half of inactive HL protein in the ER was slowly converted to an active form, which was recovered in the medium (Fig. 3). We found that HL maturation, and not secretion, was the rate-limiting step in this process (t = 2 h). Indeed, depending on the rate of folding, newly synthesized glycoproteins bind to ER chaperones for periods ranging from a few minutes to several hours (43). Most likely, HL belongs to the latter case. On the other hand, most glycoproteins spend about the same length of time in the Golgi, usually 5-15 min (43). These arguments, along with the fact that the ER, and not the Golgi, is the site of rigorous quality control (44), oppose the contention that HL maturation supposedly occurs in the Golgi, away from numerous chaperones that may block its full activation (38).
Not All Inactive HL Becomes Functional—The precursor-product relationship between the inactive HL in the cell and the active lipase that was secreted did not extend to the entire intracellular HL pool. As shown in Fig. 3C, the fraction that became active and was recovered in the medium was 52% of the total inactive HL pool. The remainder disappeared from the cells at a rate that was similar to the rate of maturation (t = 2 h). Since all inactive HL contained a high mannose glycan structure (Fig. 4, A and B), it follows that the HL fraction that did not attain a native conformation underwent ER degradation. The lability of the inactive HL fraction contrasted to the stability of active HL, since little decrease was seen in the enzymatic activity of cells and medium over the course of 8 h in Cx (see graph of HL activity in cells and medium, Fig. 3, A or C). Thus, inactive HL followed the fate of other noncompliant proteins that fail to reach a native state and are targeted to ER-associated protein degradation (ERAD; see Ref. 45).
We suggest that all newly synthesized HL molecules have the opportunity to acquire a native, functional conformation, aided and monitored by the chaperones and folding enzymes in the ER. Only those molecules that fail to fold are eventually degraded. It is tempting to speculate that the sorting of active and inactive HL toward secretion or degradation occurs through the action of ER 
-mannosidase I, the slowly acting enzyme that removes a mannose residue from the middle branch of the glycan chain to form the Man8GlcNAc2 isomer Man8B (46). This enzyme has been proposed to act as a "timer" of ER residency (47). If a glycoprotein reaches the Man8B form in its native state, it is recognized by the mannose-binding protein ERGIC53 and transported to the Golgi (48). Conversely, the Man8B isomer linked to a nonnative glycoprotein solicits its degradation in the ER (49).
HL-Calnexin Association: Robust and Specific—The detection of a robust amount of HL protein in association with calnexin provided additional evidence of the slow maturation of this lipase. Indeed, the rapid maturation of LPL did not allow detection of a similar LPL·Cnx complex (Fig. 6A), although calnexin has also been resolutely implicated in the maturation of this protein.
Calnexin and calreticulin interact with the glycan moieties of substrate glycoproteins only after they have been trimmed by glucosidases I and II to the monoglucosylated form (Glc1Man9-8GlcNAc2). Indeed, in the presence of the glucosidase inhibitor Cst, we demonstrated that HL maturation was severely impeded. However, interaction with calnexin was essential only in the early HL maturation stages (Fig. 5A). At steady state, cells contain mostly forms that, although immature, have passed the critical stage of calnexin association, and employ "later," glucose-independent chaperones or folding factors.
It has been postulated that if there is a glycosylation site within the first 50 amino acids of a nascent glycoprotein, interaction with calnexin or calreticulin begins cotranslationally (50). Although the conserved N-linked glycosylation site in the amino-terminal region of human HL is slightly downstream, at position 57, preventing glycosylation at this locus caused a 75% reduction in HL activity. Interestingly, when, in addition to this locus, glycosylation at the earlier, nonconserved N-linked locus at position 20 was also abolished, HL activity was virtually nonexistent (14). These findings suggest that HL indeed belongs to the category of proteins that bind to calnexin at co-translational stages. Likewise, in LPL, the conserved N-linked glycosylation site at position 43 was essential for expression of activity, implying that LPL also belongs to this category. On the other hand, human PL has one glycosylation site at position 167, far beyond the critical initial 50 amino acids, making it almost implicit that calnexin/calreticulin are not utilized for PL maturation.
HL binding to calnexin appears to occur specifically through the monoglucosylated glycan chain, although association of proteins through polypeptide-based interactions has also been reported (51). Thus, when abolition of glucosidase activity by Cst retained newly synthesized HL in the Glc3Man9GlcNAc2 form, Cnx bound only to the preexisting HL protein (Fig. 6B). The PAGE migration of this preexisting form, including its subpopulation precipitated by calnexin, was slightly slower than HL from untreated cells (compare lane 2 and lower band in lane 4 to lanes 1 and 3). We attribute the slower migration of this form to a difference in the degree of glucosylation. Specifically, in untreated cells, Cnx binds primarily to the monoglucosylated glycan at asparagine 57. Folding proceeds through release and binding triggered by deglucosylation/reglucosylation carried out by glucosidase II and UDP-glucose:glycoprotein glucosyltransferase (52). Since the other three glycan chains of HL are not essential for proper folding (14), they are poorer substrates for UDP-glucose:glycoprotein glucosyltransferase and presumably remain glucose-free after initial deglucosylation. However, in the presence of Cst, the preexisting unfolded protein lingers in the ER, and as glucosidases I and II are inhibited, all four chains may be reglucosylated by UDP-glucose:glycoprotein glucosyltransferase to the Glc1Man9-8GlcNac2 form.
The robust calnexin binding to the preexisting HL fraction during incubation with Cst (Fig. 6B, lane 2) is in apparent contradiction to the fact that its maturation proceeds unaffected (Fig. 5, A and C). Specifically, when glucosidase II is inactivated, HL release from calnexin through deglucosylation cannot occur, and the protein presumably remains trapped by the chaperone. If so, how could this Cnx-bound fraction become active and then be secreted?
We propose that, in the presence of Cst, the majority of Cnx-trapped HL (Fig. 6B, lane 2) represents the inactive fraction destined for degradation. Recent findings, which establish the active role of Cnx in ERAD, support this hypothesis. Specifically, besides its role in protein folding, Cnx also exists in complex with EDEM (ER degradation-enhancing -mannosidase-like protein), the recently discovered receptor responsible for directing malformed proteins toward degradation (53). EDEM extracts misfolded glycoproteins from calnexin, an action that requires the activity of glucosidase II. Thus, in the presence of Cst, the glycoprotein substrate remains attached to Cnx, and its transfer to EDEM is impeded. Taken together, we suggest that in untreated cells, HL associated to Cnx represents mostly newly synthesized lipase on its way to maturation, whereas in the presence of Cst, HL bound to Cnx represents predominately the fraction destined for degradation.
HL Forms Multiprotein Complexes in the ER—Despite the use of detergent for cell lysis, the sedimentation pattern of HL in density gradients clearly suggested its association into progressively larger molecular forms (Figs. 1 and 2). Thus, even under these conditions, HL remained bound to other ER proteins, such as, for instance, calnexin. Although the majority of the calnexin-bound HL in the cells appeared as a complex of a single calnexin and a single HL molecule, some higher molecular weight complexes were also present (Fig. 6C). Based on the unique organization and mode of function of the lectin chaperones, the additional components of these higher complexes might be inferred. Specifically, the peak trailing the HL-calnexin heterodimer might represent a ternary complex between HL, calnexin, and calreticulin (46 kDa) (54, 55), whereas the peak in the 250-kDa region might represent the former three with the addition of Erp57 (69 kDa). We attribute this hypothesis to the fact that Erp57, which has thiol oxidoreductase activity, forms stable complexes with calnexin and calreticulin (55, 56).
Besides association with calnexin, we confirmed that inactive HL was part of a number of distinct, heterogeneous components, by using the membrane-permeable cross-linker DSP (Fig. 7). The individual components of these increasingly larger HL complexes remain to be identified in an upcoming investigation. However, because of the slow kinetics of HL maturation, it is implicit that these complexes represent, at least in part, folding intermediates in association with various chaperones and other factors. The fact that, in the absence of DSP, HL migrated as a monomer even in nonreducing gels (Fig. 7, bottom panel) supports the notion that the larger forms are not terminally misfolded aggregates linked by interchain disulfide bonds, as was the case with inactive LPL (22).
The heterogeneous nature of the HL complexes is in agreement with the recent findings of distinct chaperone systems in the ER (57). Besides the calnexin/calreticulin system, a multiprotein network composed of BiP, GRP94, protein-disulfide isomerase, and other chaperones represents another major system involved in protein folding and quality control. The choice between chaperone systems is made during translocation of the nascent polypeptide, based on the presence of N-linked glycans in the initial 50 or so amino acids (50). As mentioned above, HL most likely utilizes the calnexin system at the initial stages. The possibility that partially folded substrates might "roll over" to the next chaperone system has been proposed (57), and the association of HL into various complexes could indicate this to be the case.
In addition to proteins undergoing productive folding, preformed chaperone complexes can stably associate with unfolded/misfolded proteins that will ultimately be degraded. Thus, in addition to intermediates in productive folding, at least some of the detected HL complexes undoubtedly represent that half of the lipase unable to fold correctly and directed toward degradation. Indeed, the future challenge will be to further define the factors involved in HL folding and their association into the multimolecular complexes organized in the ER.
|
FOOTNOTES |
|---|
To whom correspondence should be addressed: VA Greater Los Angeles Healthcare System, 11301 Wilshire Blvd., Bldg. 113, Rm. 312, Los Angeles, CA 90073. E-mail: markdool{at}ucla.edu.
1 The abbreviations used are: HL, hepatic lipase; Cnx, calnexin; Cst, castanospermine; Cx, cycloheximide; DSP, dithiobis[succinimidylpropionate]; EL, endothelial lipase; endo H, endoglucosidase H; ER, endoplasmic reticulum; FFA, free fatty acids; HDL, high density lipoprotein; LPL, lipoprotein lipase; NEM, N-ethylmaleimide; PBS, phosphate-buffered saline; PL, pancreatic lipase; PS-PLA1, phosphatidylserine phospholipase A1; CHO, Chinese hamster ovary.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
