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

J. Biol. Chem., Vol. 276, Issue 35, 33165-33174, August 31, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/35/33165    most recent M105644200v1 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 HighWire Citing Articles via Google Scholar Google Scholar Articles by Wallace, T. J. Articles by Grogan, W. M. Search for Related Content PubMed PubMed Citation Articles by Wallace, T. J. Articles by Grogan, W. M. Social Bookmarking                      What's this?

Mutation of Residues 423 (Met/Ile), 444 (Thr/Met), and 506 (Asn/Ser) Confer Cholesteryl Esterase Activity on Rat Lung Carboxylesterase

SER-506 IS REQUIRED FOR ACTIVATION BY cAMP-DEPENDENT PROTEIN KINASE^*

Timothy J. Wallace, Ehab M. Kodsi, Timothy B. Langston, Mervat R. Gergis, and William M. Grogan

From the Department of Biochemistry and Molecular Biophysics, School of Medicine, Virginia Commonwealth University, Richmond, Virginia 23298-0614

Received for publication, June 19, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Site-directed mutagenesis is used to identify amino acid residues that dictate reported differences in substrate specificity between rat hepatic neutral cytosolic cholesteryl ester hydrolase (hncCEH) and rat lung carboxylesterase (LCE), proteins differing by only 4 residues in their primary sequences. Beginning with LCE, the substitution Met⁴²³  Ile⁴²³ alone or in combination with other mutations increased activity with p-nitrophenylcaprylate (PNPC) relative to more hydrophilic p-nitrophenylacetate (PNPA), typical of hncCEH. The substitution Thr⁴⁴⁴  Met⁴⁴⁴ was necessary but not sufficient for expression of cholesteryl esterase activity in COS-7 cells. The substitution Asn⁵⁰⁶  Ser⁵⁰⁶, creating a potential phosphorylation site, uniformly increased activity with both PNPA and PNPC, was necessary but not sufficient for expression of cholesteryl esterase activity and conferred susceptibility to activation by cAMP-dependent protein kinase, a property of hncCEH. The 3 mutations in combination were necessary and sufficient for expression of cholesteryl esterase activity by the mutated LCE. The substitution Gln¹⁸⁶  Arg¹⁸⁶ selectively reduced esterase activity with PNPA and PNPC but was not required for cholesteryl esterase activity. Homology modeling from x-ray structures of acetylcholinesterases is used to propose three-dimensional models for hncCEH and LCE that provide insight into the effects of these mutations on substrate specificity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Carboxylesterases (EC 3.1.1.1) are a large family of highly conserved, broad specificity carboxylic acid esterases (EC 3.1.1), which are variously secreted or found in membrane fractions or cytosols of virtually all species (1-24). Typical mammalian carboxylesterases are glycoproteins (14) with molecular masses in the 57-66 kDa range. Although they form multimers, aggregation is apparently not essential for catalytic activity (2, 25). Carboxylesterases for which sequences are available are members of the superfamily of serine esterases, characterized by 3 highly conserved putative active site domains, which contribute the catalytic triad, Ser, His, and an acidic residue (2, 8, 26). They also share the sequence EDCLY, associated with a disulfide bridge that forms a variable lid-like loop over the active site in serine esterases for which tertiary structures are available (27-31). This loop contributes to substrate specificity and is apparently necessary for activity of lipases with lipophilic substrates.

Mammalian carboxylesterases are usually associated with metabolism of water-soluble xenobiotic compounds and drugs (1-3, 18, 20, 21, 32-39). However, most homologs also hydrolyze naturally occurring acyl esters and amides (1, 40), including choline esters, fatty acyl-CoAs (24), acylcarnitines, acylglycerols (29, 41), phospholipids (42), retinyl esters (43), and cholesteryl esters (44). Competing nomenclatures, most of which predate sequence information, are based on putative functions, order of discovery, differential substrate specificities, and inhibitor sensitivities or physical properties (41, 45-49). None of these systems is entirely satisfactory since they often fail to correlate with each other or with amino acid sequence information. Sequence comparisons reveal remarkable similarities among carboxylesterases but provide little additional insight into substrate specificities or functional roles of the individual enzymes (1-24, 26). Moreover, the homologous enzymes and their corresponding nucleic acid sequences can be difficult to discriminate by conventional immunoblotting and nucleic acid hybridization assays, complicating efforts to study their specific physiological roles.

One of the few carboxylesterases clearly assigned to a specific function is the hepatic neutral cytosolic cholesteryl ester hydrolase (hncCEH)¹ cloned and characterized in this laboratory (25, 29, 44, 50). The hncCEH is suppressed by cholesterol feeding and other treatments that augment hepatic-free cholesterol and induced by treatments that decrease hepatic free cholesterol, indicating a role in regulation of intracellular free cholesterol (51-53). This regulation is mediated by consensus sterol regulatory elements on the hncCEH gene promoter, similar to those present on the promoters of other enzymes that regulate cholesterol flux through the liver (54, 55). Like hormone-sensitive lipase/CEH, the hncCEH also has triglyceride lipase activity (29) and is activated by cAMP-dependent protein kinase (56), a property not reported for other carboxylesterases. However, hncCEH has little sequence similarity with hormone-sensitive lipase/cholesterol esterase (57), lysosomal acid lipase/cholesterol esterase (58), or other enzymes classified as lipases (59-61).

In contrast, the rat lung cytosolic carboxylesterase (LCE) is >99% identical to hncCEH in cDNA nucleotide and protein amino acid sequences (3). The active site and lid regions are identical and exactly aligned. Thus, these enzymes are indistinguishable from each other and a number of other homologous carboxylesterases by conventional immunoblotting techniques. However, LCE exhibited no measurable hydrolytic activity with cholesteryl esters. In order to understand the physiological functions of carboxylesterases, it is important to define structural properties that dictate their substrate specificities and susceptibility to post-translational regulation.

In the current study, hncCEH and LCE are used as models to identify structural features necessary or sufficient for expression of cholesteryl esterase activity by carboxylesterases and for their activation by protein kinase. Site-directed mutagenesis of LCE is used to substitute amino acid residues individually and cumulatively for identification of critical residues that dictate substrate specificities and sensitivity to activation by protein kinase. Homology modeling is used to construct a three-dimensional structure that provides insight into the observed effects of individual mutations.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Site-directed Mutagenesis-- Beginning with the lung carboxylesterase cDNA previously cloned into plasmid pCR3^TM, 6 nucleotides were mutated individually or cumulatively using the QuickChange^TM site-directed mutagenesis kit from Stratagene (La Jolla, CA) according to the manufacturer's instructions. As summarized in Table I, these point mutations encoded amino acid sequences intermediate between LCE and hncCEH. In the initial round, cumulative mutations were introduced in the order, SDM-557 (Gln¹⁸⁶  Arg¹⁸⁶), SDM-1474 (Lys⁴⁹²  Glu⁴⁹²), SDM-1331 (Thr⁴⁴⁴  Met⁴⁴⁴), SDM-1472 (Ser⁴⁹¹  Thr⁴⁹¹), and SDM-1269 (Met⁴²³  Ile⁴²³). The hncCEH cDNA was used in lieu of an additional mutation at nucleotide 1517 (Asn⁵⁰⁶  Ser⁵⁰⁶). SDM-1472 and SDM-1474 resulted from initial sequencing errors but had no effect on activity. These were not present in later mutants. In the second round, SDM-1269 and SDM-1517 were introduced individually, in combination (double mutant) and in combination with SDM-1331 (triple mutant). Each primer was designed to be >35 base pairs with a melting temperature >78 °C (T_m = 81.5 ± 0.41(%GC)  675/N  % mismatch, where N = primer length in base pairs). Primers were: SDM-557, ⁵⁴¹GGGGATGAACACAGCCGGGGCAACTGGGGTCACTTGG⁵⁷⁷; SDM-1331, ¹³¹²GCTGGAGCCCCCACCTTCATGTATGAATTTGAGTATCGCC¹³⁵¹; SDM-1474, ¹⁴⁵⁸GGAGACCAATCTCAGCCAAATGGTGATGAAATACTGGGCC¹⁴⁹⁷; SDM-1472, ¹⁴⁴⁸CCTCAGAAGAGACCAATCTCACCCAAATGGTGATGAAATACTGGGC¹⁴⁹⁶; SDM-1269, ¹²⁴⁸GGACTTGGTTGCAGATGTGATATTTGGTGTCCCATCAGTAATGG¹²⁹¹; SDM-1517, ¹⁴⁹⁵GCCAACTTTGCTCGGAATGGGAGCCCTAATGGGGGAGGGCTGCC¹⁵³⁸. The reaction contained 5 µl of 10 times reaction buffer (100 mM KCl, 100 mM (NH₄)₂SO₄, 200 mM Tris-HCl, pH 8.8, 20 mM MgSO₄, 1% Triton X-100, 1 mg/ml nuclease-free bovine serum albumin, 30 ng of plasmid DNA, 125 ng, each sense and antisense primer, 1 µl of dNTP mixture and double-distilled water to a final volume of 50 µl. Pfu Turbo DNA polymerase (2.5 units) was added and the reaction was overlaid with 30 µl of mineral oil. Thermocycling parameters were: 95 °C/30 s; 12 cycles, 95 °C/30 s; 55 °C/1 min; 68 °C/14 min. Reactions were cooled on ice for 2 min, centrifuged and digested with 10 units of DpnI restriction enzyme for 1 h at 37 °C. Epicurian Coli XL1-Blue supercompetent cells (50 µl) were incubated on ice with 1 µl of reaction mixture for 30 min, heat pulsed 45 s at 42 °C, and cooled on ice 2 min. NZY⁺ broth (500 µl at 42 °C) was added to the transformation reactions and incubated 1 h at 37 °C with shaking at 235 rpm. The entire reaction was then plated on LB/ampicillin agar plates and incubated 16 h at 37 °C.

Confirmation of the Mutated Lung Carboxylesterase Sequence-- Colonies were selected and product pCR3 plasmid DNA containing the mutated LCE sequence was prepared using alkaline lysis and ethanol precipitation. Aliquots of each mutant plasmid were then sequenced in both directions using the ABI PRISM^TM Dye Terminator Cycle Sequencing Ready Reaction Kit (PerkinElmer Life Sciences, Foster City, CA) according to the manufacturer's instructions, with primers designed from the LCE sequence to produce overlapping products about 250 base pairs apart. The reactions were analyzed on the Applied Biosystems DNA Sequencer 373A. These procedures were also used to correct the cDNA sequences for hncCEH and LCE.

Transfection of the Mutated Lung Carboxylesterase cDNAs-- As described in Wallace et al. (3), mutant plasmids were purified on a double-banded cesium chloride gradient. COS-7 cells were grown to 90% confluency on 60-mm transfection plates. 8 µg of mutant plasmid DNA was incubated with 12-24 µl of LipofectAMINE (Life Technologies, Inc., Gaithersburg, MD) in 1.5 ml of OPTI-MEM^TM I Reduced serum-free medium (Life Technologies, Inc., Gaithersburg, MD) for 30 min at room temperature, followed by an additional 1.5 ml of serum-free medium. 3 ml of this liposomal complex was then added to a monolayer of COS-7 cells seeded to 90% confluency on 60-mm transfection plates. Non-transfected controls were treated with LipofectAMINE without plasmid. Cells were incubated for 48 h in a 37 °C humidified incubator with 5.0% CO₂ with addition of 2 ml of complete Dulbecco's modified Eagle's medium with 20% fetal bovine serum after 6 h. Cells were harvested in 1 ml of 250 mM phosphate buffer, pH 7.4, with 5 mM -mercaptoethanol, 80 mM KCl, 250 mM sucrose and protease inhibitors, 2 mM 3-methyl propionate, 0.5 mM benzoyl arginine, and 5 mM benzamidine. The cell suspension was sonicated using Heat Systems Ultrasonic Processor, and the cell debris was removed by centrifugation at 10,000 × g for 30 min at 4 °C. The resulting post-mitochondrial supernatant was assayed for hydrolytic activity with p-nitrophenyl acetate, p-nitrophenyl caprylate, and cholesteryl oleate as we have described (44). For studies with protein kinase activation, the supernatant was further centrifuged for 90 min at 104,000 × g and the resulting supernatant assayed with these substrates as we have described (56).

Enzyme Assays-- As described in Wallace et al. (3) hydrolysis of p-nitrophenyl acetate (PNPA) and p-nitrophenyl caprylate (PNPC) (Sigma) was measured in an aliquot of post-mitochondrial homogenate or cytosol in a reaction mixture containing 20 µg of protein, 200 mM phosphate buffer, 5 mM -mercaptoethanol, and 80 mM KCl, in a final volume of 1 ml. The reaction was started by the addition of p-nitrophenyl acetate or p-nitrophenyl caprylate to a final concentration of 200 µM and incubated at 37 °C for 5 min for p-nitrophenyl acetate or 15 min for p-nitrophenyl caprylate. Under these conditions, the reaction rate was linear with respect to time and protein concentration. The release of p-nitrophenol was monitored by measurement of absorbance at 400 nm in a double beam spectrophotometer. Specific activities were corrected for endogenous activity by subtracting activities of non-transfected controls and ratioed to specific activity of lung carboxylesterase.

Cholesteryl esterase activity was measured as described in Ghosh et al. (44). Cholesteryl [1-¹⁴C]oleate (75 µM) (PerkinElmer Life Sciences, Boston, MA) was added in 12 µl of acetone to 500 µl of assay mixture containing 200 mM, pH 7.4, phosphate buffer, 80 mM KCl, 5 mM -mercaptoethanol, and 100 µg of protein. The reaction mixture was incubated for 30 min at 37 °C and then stopped by the addition of 3.25 ml of methanol:choroform:hexane (3.85:3.42:2.73; v/v/v) and 50 µl of 1 M NaOH. The tubes were vortexed and phases separated by centrifugation. Radioactivity was measured in a 1-ml aliquot of the aqueous phase by liquid scintillation counting. Under these conditions, the reaction is linear with respect to time and protein concentration. Assays of individual transfections were carried out in triplicate on each of 2-12 separate transfections. Since LCE expressed no marginal activity with this substrate (3), cholesteryl esterase specific activities were expressed as a percentage of those for non-transfected controls.

Activation by cAMP-dependent Protein Kinase-- Assays were carried out as described above with and without addition of 1 mM MgCl₂, 5 mM ATP, and 100 µM cyclic 3',5'-adenosine monophosphate (cAMP) to activate endogenous cytosolic protein kinase. The 104,000 × g supernatants were preincubated with the co-factors for 1-2 min prior to addition of substrate and continuation of the assay as described above. We have found these conditions to give optimal activation of hncCEH by endogenous protein kinase (56).

Protein Estimation-- Protein was estimated by Bio-Rad dye binding assay according to the manufacturer's instructions using bovine serum albumin as standard.

Statistical Analysis-- Statistically significant differences were determined by analysis of variance with the Bonferroni and the Tukey post-tests, which gave equivalent results. To assure definitive results, all assays were carried out at least in triplicate on individual transfections. Transfections were repeated to confirm significant findings. SDM-557 and SDM-557/1474 were each transfected in duplicate. SDM-557/1474/1331/1472 and SDM-557/1474/1331/1472/1269 were transfected in triplicate. SDM-1269/1331/1517 was transfected in quadruplicate. SDM-557/1474/1331 was transfected 6 times. SDM-1269, SDM-1517, and SDM-1269/1517 were transfected 3 times in quadruplicate.

Protein Homology Modeling-- The primary amino acid sequences translated from the cDNA sequences for hncCEH and LCE and the various mutants were submitted to the META II-PredictProtein server (www.EMBL-Heidelberg.DE/predictprotein) for automated distribution to the various modeling programs cited below. Secondary structures were modeled and predicted by JPRED, a program that predicts secondary structure from primary sequence using a neural network based analysis that seeks consensus among several modeling programs (62). O-Glycosylation and O-phosphorylation sites were predicted by NetOGlyc 2.0 (63) and NetPhos 2.0 (64), programs that produce neural network predictions based on analyses of sequence context and surface accessibility. First attempts to model three-dimensional structure by homology were made by submission of amino acid sequences to SWISS-MODEL (65, 66) for automated modeling. The EMBL Nucleotide Sequence Data base and SWISS-PROT/TrEMBL Protein Sequence Data base were searched for similar sequences and available three-dimensional structures were downloaded for manual sequence alignments. The homology modeling program Swiss-PdbViewer v.3.7b2 (65, 66) was used to fit manually aligned sequences to three-dimensional models of similar sequences for which crystal structures were available. The loop building feature was used to fill gaps in coil regions and the compute energy and energy minimization routines were used to minimize free energy of the resulting three-dimensional models.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Sequence Confirmation-- A major aim of this study was to identify structural differences between hncCEH and lung carboxylesterase responsible for their differential substrate specificities. As indicated in Table I the previously reported cDNA sequences (3, 44) produced translated amino acid sequences that differed in only 6 amino acid residues. None of these differences involved any of the consensus active site sequences for esterases, which were identical and exactly aligned in both enzymes (see Fig. 1). Subsequent resequencing of the hncCEH cDNA in expression (PCR3) and cloning (pUC19) vectors established a corrected sequence that differed from LCE by only 8 nucleotides in the coding region and a translated amino acid sequence that differed by only 4 amino acid residues (see Fig. 1). Thus, 2 of the previously reported differences in amino acid sequence were attributable to sequencing errors. Each mutant cDNA was sequenced in forward and reverse directions to confirm identity with the cDNA for LCE except for mutations detailed in Table I.

View this table: [in this window] [in a new window]   Table I Differences between nucleotide coding and amino acid sequences of wild type lung carboxylesterase (LCE), hepatic neutral cytosolic cholesteryl ester hydrolase (hncCEH), and intermediate sequences produced by site-directed mutagenesis Numbering is based on exact alignment of corresponding coding sequences at the translational start sites (+1) for LCE and hncCEH. Although all nucleotide differences are shown, mutations were confined to those that altered the amino acid encoded. Mutations 1472 (G  C) and 1474 (A  C) reflect errors in the reported sequence for hncCEH that were later corrected by repeated sequencing. Neither of these mutations altered CEH or PNPA/PNPC esterase activities from those of the precursor proteins.

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Aligned amino acid sequences of hepatic neutral cytosolic cholesteryl ester hydrolase and lung carboxylesterase translated from the corresponding cDNAs. Arrows indicate mutated residues. High potential N-glycosylation or O-glycosylation sites are indicated by g. High potential O-phosphorylation sites are indicated by p. Boxes indicate active site triad (GESAG, SES, GDHGD) and consensus sequence involved in loop formation (SEDCLY).

Modeling of the Primary Sequences-- The translated primary amino acid sequences were submitted to the EMBL META II PredictProtein server to obtain homologous sequences and analyses for secondary structure, consensus glycosylation sites, and phosphorylation sites. As indicated in Fig. 1, each of the translated proteins contained 3 potential glycosylation sites of equivalent calculated potential (data not shown). Whereas 25 potential O-phosphorylation sites were identified (Fig. 1), the mutants differed only in residue 506, which is Asn in LCE and mutants lacking SDM-1517, and Ser in hncCEH and mutants with SDM-1517 (data not shown). Predictions of secondary structures from the primary sequences of hncCEH and LCE revealed no significant differences in distribution of helix, sheet, or coil structures, even in the immediate vicinity of the mutated residues (JPRED data not shown), suggesting that the differential substrate specificities of these enzymes may be attributable to subtle differences in tertiary structure. A search of protein data bases revealed no x-ray crystal structures for hncCEH or LCE or for proteins sufficiently similar to hncCEH or LCE to permit rigorous homology modeling (SWISS- MODEL). The best sequence match with a protein for which a three-dimensional structure was available was obtained with murine acetylcholinesterase (67, 68). This protein is 31% identical and 48% similar in amino acid sequence with hncCEH (Fig. 2). The proteins are homologous in consensus active site regions and their secondary structural domains are substantially aligned. Energy minimized three-dimensional models were constructed from translated sequences of hncCEH (Fig. 3) and LCE (not shown) using the alignments with acetylcholinesterase shown in Fig. 2. Models based on alignments with bile salt stimulated lipases (69, 70), which exhibited lower sequence similarities, shared similar structural features (not shown). The structural properties of these models and their implications for the effects of mutations will be discussed at length below (see "Discussion").

View larger version (57K): [in this window] [in a new window]   Fig. 2.   Sequence and secondary structure alignments of hncCEH and acetylcholinesterases from mouse and electric ray used in modeling the three-dimensional structure of hncCEH and LCE. The acetylcholinesterase sequences were obtained from the Swiss-Prot protein data base. Secondary structure of hncCEH was predicted by JPRED and alignments were made by Swiss-Pdb Viewer, as described under "Experimental Procedures." Lowercase indicates extended coil (loop) regions; uppercase indicates -sheet; bold face indicates helical regions. Asterisk indicates amino acid identity; period indicates amino acid similarity. Homologous regions around active site residues, loop structure, and putative phosphorylation site at Ser506 are boxed.

View larger version (48K): [in this window] [in a new window]   Fig. 3.   Proposed three-dimensional structure of hepatic neutral cytosolic cholesteryl ester hydrolase obtained by homology modeling with x-ray crystal structures of acetylcholinesterases. Amino acid residues of active site triad are shown in green; mutated residues are shown in blue; Cys residues involved in formation of the lid-loop structure are shown in yellow. The entire structure is shown in panel A. Panel B isolates aa87-aa116, the N-terminal lid-loop, aa203-aa235, which contains the active site Ser221 and aa339-aa565, the C-terminal domain that includes the mutated residues. Panel C is the same as Panel B with the model rotated 90 degrees in the horizontal plane.

Effects of Mutations on Differential Substrate Specificity-- As described under "Experimental Procedures," the cDNA for LCE was mutated to encode proteins with amino acid sequences intermediate between LCE and hncCEH (see Table I), with reference to the respective sequences shown in Fig. 1. The cDNAs were overexpressed in COS-7 cells and the catalytic activities of the resulting proteins were measured with the water-soluble substrates PNPA and PNPC and with cholesteryl oleate (see Fig. 4). For water-soluble substrates, specific activities were corrected by subtraction of non-transfected controls (10-25% of transfected values) and normalized to specific activities for LCE as described under "Experimental Procedures." Of necessity, specific activities with cholesteryl oleate were normalized by expression as % control, since LCE and most of the mutants expressed no measurable activity with this substrate.

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Hydrolysis of ester substrates, PNPA, PNPC, and cholesteryl oleate by LCE, hncCEH, and mutant proteins expressed in COS-7 cells. A series of point mutations were introduced individually and cumulatively into LCE to produce hncCEH by stages. Specific activities with PNP esters are corrected for activity of non-transfected controls and expressed as % of the corresponding baseline LCE activity. Activities with cholesteryl oleate are expressed as % of the activity of non-transfected controls. +, also included SDM-1474, which had no effect on activity (data not shown). ++, also included SDM-1474 and SDM-1472, neither of which had any effect on activity by themselves. *, specific activity was different from LCE, p < .001. **, specific activity was different from LCE, p < .05.

Specific activities were measured in the linear range with respect to protein concentrations and assay times for hncCEH and LCE. The baseline enzyme, LCE (100% in Fig. 4) gave similar activities with PNPC (40 ± 7 µmol/min/mg of protein) and PNPA (27 ± 5.3 µmol/min/mg of protein). Differences between these values were not statistically significant. As reported earlier, LCE gave no measurable activity with cholesteryl oleate (3). The first mutation (SDM-557; Gln¹⁸⁶  Arg¹⁸⁶) had little or no effect on the relative hydrolytic activities with PNPC and PNPA. This mutant also expressed no activity with cholesteryl oleate. A second cumulative mutation (SDM-1474; Lys⁴⁹²  Glu⁴⁹²), based on an error in a published sequence (data not shown), and the third cumulative mutation (SDM-1331; Thr⁴⁴⁴  Met⁴⁴⁴), representing a real difference in the model enzymes (Fig. 2), also had no significant effect on the activities with any of the substrates. Neither mutation produced any activity with cholesteryl oleate. A fourth cumulative mutation (SDM-1472; Ser⁴⁹¹  Thr⁴⁹¹), also based on a sequencing error, produced no change in the relative activities with PNPA or PNPC and no activity with cholesteryl oleate (data not shown). The fifth and last cumulative mutation (SDM-1269; Met⁴²³  Ile⁴²³), one mutation short of the native hncCEH (44), selectively increased activity with the more hydrophobic PNPC relative to activity with the more hydrophilic PNPA, but still produced no activity with cholesteryl oleate. These results suggested that the final cumulative mutation shifted the substrate specificity toward more hydrophobic substrates but that the cholesteryl esterase activity is dependent on the single remaining amino acid difference from hncCEH at residue 506.

The amino acid residues were then mutated and expressed individually and in various combinations to further define the effects of critical residues on substrate specificities. SDM-1269 (Met⁴²³  Ile⁴²³) alone increased specific activity 8-fold with PNPC (p < .001). In contrast, PNPA activity was not significantly increased, confirming the shift in preference toward more hydrophobic substrates seen with the cumulative mutation at nucleotide 1269. Moreover, as with the cumulative mutation, SDM-1269 produced no significant change in cholesteryl esterase activity in the cell homogenate.

SDM-1517 (Asn⁵⁰⁶  Ser⁵⁰⁶) alone stimulated activity with both PNPA and PNPC, 8.7- and 6-fold, respectively, but had no significant effect on cholesteryl esterase activity, which was 45% greater (not significant) than LCE but actually declined as ratios to activities with PNPA and PNPC (ratios not shown). Thus, this mutation alone markedly increases general esterase activity but apparently does not alter substrate specificities.

In contrast, the combination of mutations SDM-1269/1517 selectively boosted activity with PNPC to its highest level, 18.8-fold that of LCE, while increasing PNPA activity only 4-fold. Although the addition of SDM-1269 shifted specificity toward the more hydrophobic substrate, the double mutant expressed no cholesteryl esterase activity, suggesting that either or both SDM-557 and SDM-1331 were also essential for this activity.

This was confirmed by expression of the triple mutant, SDM-1269/1331/1517. This mutant expressed 3.6-fold the LCE activity with PNPA and 7.5-fold the LCE activity with PNPC, thus retaining the preference for PNPC conferred by SDM-1269 and the enhanced esterase activity conferred by both SDM-1269 and SDM-1517, alone or in combination. More importantly, the triple mutant expressed 2.7-fold the cholesteryl esterase activity of non-transfected controls, not significantly different from the activity seen with native CEH in this and previous studies. These results confirm that these 3 mutations are both necessary and sufficient for expression of CEH activity.

Since the hncCEH itself would result from the combination of all four mutations, it was not necessary to add SDM-557 (Gln¹⁸⁶  Arg¹⁸⁶) to the triple mutant. The hncCEH hydrolyzed cholesteryl oleate and showed a preference for more hydrophobic substrates. Its activity with the hydrophilic substrate, PNPA was only 6% of the corresponding value for PNPC and 10% of the activity of LCE with PNPA. Moreover, hncCEH and mutants with SDM-557 expressed substantially lower activities with PNP esters than those without this mutation. Most notably, this mutant expressed lower activity with all 3 substrates than the triple mutant SDM-1269/1331/1517, despite identity of these enzymes in the 3 critical residues. Although SDM-557 alone did not significantly diminish LCE activity with any of the substrates, the activities of LCE were already relatively low in comparison with mutants that lacked SDM-557. These data suggest that presence of a basic residue at this site may limit hydrolytic activity.

Effects of cAMP-dependent Protein Kinase on Catalytic Activity of Mutant SDM-1517-- Previous studies have shown that hncCEH is activated by cAMP-dependent protein kinase (56). As seen in Fig. 1, the mutation, SDM-1517 (Asn⁵⁰⁶  Ser⁵⁰⁶) confers an additional putative consensus phosphorylation site on LCE. To test the hypothesis that this mutation may enable cholesteryl esterase activity by providing an additional phosphorylation site, activity of the triple mutant protein (SDM-1269/1331/1517) was measured in the cytosolic fraction from transfected COS-7 cells with and without cAMP and cofactors previously shown to activate hncCEH in liver cytosol (56). As shown in Fig. 5, transfection increased cytosolic activity with PNPA (10 ± 1.2 to 133 ± 3 µmol/h/mg of protein), PNPC (4 ± 1 to 286 ± 56 µmol/h/mg of protein), and cholesteryl oleate (7 ± 3 to 101 ± 30 pmol/h/mg of protein), as expected. PKA activators did not increase activity with either of the water-soluble substrates in transfected cells or non-transfected controls. However, PKA activators increased cholesteryl esterase activity to 4-fold (395 ± 115 pmol/h/mg of protein) that of untreated transfected cells. Inasmuch as PKA activators had no measurable effect on cholesteryl esterase activity in non-transfected controls, the contribution of endogenous CEH to this activation was negligible. Inasmuch as Ser⁵⁰⁶ appears to be essential for expression of cholesteryl esterase activity by hncCEH, this data suggested that phosphorylation of this site selectively stimulates and may be required for hydrolysis of cholesteryl esters.

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Effects of Ser506 (SDM-1517) on activation of cytosolic esterase activity by endogenous protein kinase. Mutated proteins with (SDM-1517, SDM-1269/1331/1517) and without (SDM-1269) Ser506 were expressed in COS-7 cells and assayed for esterase activity with and without treatment by protein kinase cofactors. Specific activities are reported as a ratio with basal activities of non-transfected controls. Panel A, cholesteryl esterase activity; panel B, PNPA esterase activity; panel C, PNPC esterase activity. p values indicate statistical significance of differences between transfected and non-transfected activities without protein kinase cofactors or between transfected with and without protein kinase cofactors.

To test this hypothesis, the effects of PKA activators were determined with the single mutant SDM-1517, encoding a protein that includes Ser⁵⁰⁶ and expresses little or no CEH activity. As shown in Fig. 5, transfected cells expressed much higher levels of activity than non-transfected controls with the water-soluble substrates and equivalent cholesteryl esterase activity, consistent with results from post-mitochondrial homogenates (Fig. 4). However, in contrast to results seen with the triple mutant, PKA activators stimulated PNPA activity 18% and PNPC activity 86% in transfected cells but had no effect on cholesteryl esterase activity. These data indicate that phosphorylation of the mutant containing Ser⁵⁰⁶ selectively stimulated hydrolysis of the more hydrophobic of the water-soluble substrates but was not sufficient to confer cholesteryl esterase activity on this homolog that expressed no basal activity.

Finally, to confirm the importance of Ser⁵⁰⁶ to activation of esterase activity, SDM-1269 was expressed with and without protein kinase activators. As expected, cytosolic esterase activity was increased in transfected cells 5.9-fold with PNPA and 15.3-fold (132 ± 132 to 2017 ± 109 µmol/h/mg of protein) with PNPC, confirming the disproportionate increase in activity with the more hydrophobic substrate, seen with this mutation in the post-mitochondrial homogenates. Cholesteryl esterase activity also increased 1.8-fold (0.31 ± .02 to 0.57 ± .04 nmol/h/mg of protein), although this did not approach the 14-fold increase seen with the triple mutant. Nevertheless, none of the activities were stimulated by protein kinase activators. Thus, whereas other sites may play a role in altering substrate specificity, activation by protein kinase seems to be uniquely associated with the putative phosphorylation site Ser⁵⁰⁶.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

General Observations and Conclusions-- These studies establish that the substrate specificities of carboxylesterases, long regarded as broad specificity general esterases, are quite sensitive to minor differences in amino acid sequence distal to the consensus active site regions. The primary amino acid sequences of hepatic neutral cytosolic cholesteryl ester hydrolase and lung carboxylesterase differ in only 4 residues (Fig. 1). Beginning with the sequence of LCE, each of these residues was mutated alone or cumulatively to the corresponding hncCEH residue and tested for hydrolytic activity with PNPA, PNPC, and cholesteryl oleate in COS-7 cells, a mammalian cell line that efficiently expresses hncCEH and LCE activities (3, 44). As depicted in Fig. 4 and detailed under "Results," 3 mutations, Met⁴²³  Ile⁴²³, Thr⁴⁴⁴  Met⁴⁴⁴, and Asn⁵⁰⁶  Ser⁵⁰⁶ were necessary and sufficient to confer CEH activity on LCE in post-mitochondrial homogenates, whereas the fourth, Gln¹⁸⁶  Arg¹⁸⁶, was not required. Although an increase in cytosolic CEH activity was detectable with Met⁴²³  Ile⁴²³ (Fig. 5), it was much smaller than that seen with the PNP esters in any of the mutants or with cholesteryl oleate in the triple mutant. This increase may reflect increased sensitivity of the assay to small changes in specific activity, due to lower basal CEH activity in cytosol than in post-mitochondrial homogenates.

The individual mutations produced major shifts in relative specific activities with the various substrates, as detailed under "Results." All constructs with Ile⁴²³, including hncCEH itself, exhibited substantially greater activities with the more hydrophobic PNPC than with PNPA, indicating that this residue accounts for the corresponding differences between hncCEH and LCE.

On the other hand, constructs with Ser⁵⁰⁶ produced higher levels of esterase activity than those with Asn⁵⁰⁶, without altering substrate selectivity. Although Met⁴⁴⁴ was apparently necessary for CEH activity, this residue did not alter activity with the water-soluble substrates. Together, these observations suggest that these residues are associated with fundamentally different functions of the enzymes. This concept will be further developed in the context of the three-dimensional structure.

Activation by protein kinase apparently requires the putative phosphorylation site at Ser⁵⁰⁶, inasmuch as 2 mutants with this residue were activated by protein kinase, whereas none of the esterase activities of SDM-1269, which lacks Ser⁵⁰⁶, were activated by protein kinase; this includes the small amount of CEH activity that this mutant expressed in the cytosol (Fig. 5). Although these data suggest that phosphorylation of Ser⁵⁰⁶ is necessary for optimal hncCEH activity, it is not sufficient for CEH activity, since SDM-1517 expressed no measurable activity with cholesteryl oleate with or without protein kinase. It should be noted that protein kinase activation of the mutants with Ser⁵⁰⁶ occurred with different substrates, a phenomenon that may give insight into the mechanism of activation, as discussed further below.

Modeling Approaches-- Each mutation introduced an amino acid of substantially different polarity and, in the case of SDM-557, substituted polar-charged Arg¹⁸⁶ for polar uncharged Gln¹⁸⁶. However, none of the mutated residues are located in the primary sequence in close proximity to any of the consensus active site or lid sequences (see Fig. 3). Modeling of mutated sequences revealed no meaningful alterations in predicted secondary structure that might explain changes in activity or substrate specificity. Moreover, only one of the mutations influences the predicted glycosylation or phosphorylation potential at any site. Specifically, Asn⁵⁰⁶  Ser⁵⁰⁶ introduces a high potential phosphorylation site into the mutated proteins, increasing enzyme activity but not conferring CEH activity by itself. It is apparent that these effects can only be understood in the context of three-dimensional structure.

Although the three-dimensional structure has not been reported for any carboxylesterase, some tentative conclusions can be drawn from models, such as that depicted in Fig. 3, based on alignments with three-dimensional structures of serine esterases that share general structural features, as well as substantial sequence homologies. For example, murine acetylcholinesterase (Swiss Protein Database 1MAAC) is 31% identical and 48% similar to hncCEH in amino acid composition (67, 68) (see Fig. 2). The lid-associated sequence SEDCLYLN and the active site sequence VTIFGESAGG are identical in the 2 proteins, although only 6/28 residues in the lid-loops of LCE/hncCEH (Ser⁸⁸-Asp¹¹⁵) are similar to those of acetylcholinesterase, consistent with the view that this domain is one determinant of substrate specificity (27-31). Also very similar between the enzymes and among species is YWANFARG(N/S)PN (10/12 identical or similar substitution), corresponding to the region around mutated Asn⁵⁰⁶/Ser⁵⁰⁶ in LCE/hncCEH, consistent with the view that this site is important for esterase activity but not substrate specificity. Similar alignments can be made with corresponding domains of other serine esterases (24, 67, 69, 71-79), including some lipases that are less similar in overall sequence (see, for example, Table II).

Moreover, acetyl cholinesterase is virtually identical to hncCEH in distribution and size of corresponding helical and sheet regions (Fig. 2). Although the alignments of secondary structure differ slightly in the length of some intervening coils, the correspondence is almost exact in the aa³³³-aa⁵⁰⁸ region of hncCEH where the 3 critical mutations are located, indicating substantial similarity in three-dimensional structures. By contrast, the acetylcholinesterase from Pacific electric ray (Swiss Protein Database 1EA5A), 29% identical with hncCEH, was only 57% similar to the corresponding mouse sequence. Nevertheless, it is virtually identical in distribution of secondary structure and its three-dimensional structure corresponds well with the murine form. Available three-dimensional structures of these enzymes and other serine esterases (bile salt stimulated lipase/cholesteryl esterase) with substantial sequence homologies align well with each other and exhibit similar structural features, especially in the putative active site domains, validating their use as models for hncCEH, LCE, and the intervening mutants. Similar approaches have been used to model other enzymes after bile salt-stimulated lipase (27, 30, 31).

Modeling Insights-- As depicted in Fig. 3, the putative catalytic triad is located in a pocket (gorge) formed between distinct N-terminal (approximately aa¹-aa³⁵⁰) and C-terminal domains (approximately aa³⁵¹-aa⁵⁶⁵). A loop formed by a disulfide bridge between Cys⁸⁷ and Cys¹¹⁶ constitutes a lid structure that covers the substrate-binding pocket (see Fig. 3B for isolated view). This lid, which varies in size and amino acid composition among serine esterases, is known to be an important determinant of substrate specificity in lipases (27, 30). However, none of the mutations are located in close proximity to the lid structure. The results of this study establish that the differences in substrate specificity between hncCEH and LCE reside in 3 residues at positions 423, 444, and 506. The model locates the corresponding residues and/or their associated structural domains in a shallow arc around the C-terminal wall of the binding pocket in these enzymes (Fig. 3) and model enzymes for which three-dimensional structures have been unambiguously determined (not shown). Since these residues appear to be remote from the catalytic site in all models, it is probably safe to assume that they are involved in substrate binding, orientation or access to the binding site. Only Gln¹⁸⁶/Arg¹⁸⁶ (SDM-557) is located in the N-terminal domain, far removed from the binding pocket or lid domain. Although this mutation may affect the overall activity of the enzyme, it does not appear to have any effect on substrate specificity.

In contrast, SDM-1269 and SDM-1331 mutate residues found in the C-terminal wall of the binding pocket, which is also known to play a role in determining substrate specificity in lipases (31). SDM-1269 substitutes the substantially more hydrophobic Ile⁴²³ for Met⁴²³ in a short loop that protrudes into the binding pocket at the junction between 2 helices (aa⁴⁰⁸-aa⁴²² and aa⁴²⁶-aa⁴³⁷) in the C-terminal wall, about 11 Å from active site Ser²²¹ in the model. This substitution selectively increases activity with the more hydrophobic substrate PNPC, suggesting that Ile⁴²³ may be associated with a fatty acid-binding domain, since PNPC differs from PNPA only in its fatty acyl moiety.

SDM-1331 substitutes the hydrophobic residue Met⁴⁴⁴ for polar uncharged Thr⁴⁴⁴. This residue is located about 16 Å from active site Ser²²¹ and 17 Å from aa⁴²³, in a -sheet domain (aa⁴⁴¹-aa⁴⁴⁸) most of which is shielded from the binding pocket. Nevertheless, aa⁴⁴⁴ is in contact with the binding pocket and forms part of the C-terminal wall (Fig. 3). When Thr⁴⁴⁴ is present, its side chain forms the hydrophilic floor of a pit bounded by hydrophobic residues Phe⁴⁴³, Met⁴⁴⁴, Val³⁴⁷, Gly³⁴⁸, and Ile³⁴⁹ and roughly perpendicular to a line drawn through Ser²²¹ and aa⁴²³ in the plane common to the 3 residues. When Met⁴⁴⁴ is present, its side chain provides the floor to a hydrophobic pit bounded by the same residues. Although substitution of Met⁴⁴⁴ for Thr⁴⁴⁴ had little or no effect on activity with any of the substrates in the absence of other mutations (Fig. 4), it was required for activity of the triple mutant with cholesteryl oleate, suggesting that this residue may be associated with a cholesterol-binding domain. This would be consistent with the dimensions of the hydrophobic pit, its distance, about the length of a cholesterol molecule, from active site Ser²²¹, and the large change in polarity in the otherwise hydrophobic environment of the pit.

Finally, SDM-1517 substitutes Ser⁵⁰⁶ for Asn⁵⁰⁶ in a loop that presents this residue on the external surface of the C-terminal domain, where it is accessible to phosphorylation by protein kinase but distant and completely shielded from the active site. Neither the Asn nor the Ser has a high glycosylation potential or belongs to a consensus glycosylation sequence. However, as seen in Fig. 3, the loop in question extends from the base of a large helix (aa⁴⁸¹-aa⁵⁰³) that comprises one end of the C-terminal wall in the binding pocket. This helix presents a hydrophobic face to the binding site. Moreover, the N-terminal end of this helix is attached to a loop (aa⁴⁷⁴-aa⁴⁸⁰)/helix (aa⁴⁷⁰-aa⁴⁷³)/loop (aa⁴⁴⁹-aa⁴⁶⁹) that extends out over and into the active site, contributing His⁴⁶⁶ of the catalytic triad. Whereas this substitution alone seems to have little potential to affect three-dimensional structure, introduction of an anionic phosphate at Ser⁵⁰⁶ is likely to generate substantial structural perturbations. For example, the model locates 2 basic amino acids, Arg⁵⁰³ and Lys⁴⁹⁶ in the proximal helix within 8-10 Å of Ser⁵⁰⁶. In the case of hncCEH, retraction of this component of the C-terminal wall by electrostatic interaction with a phosphate group on Ser⁵⁰⁶ may make the active site more accessible or improve alignment of the substrate with the active site.

Such an interpretation might account for the observation that protein kinase activation is seen with cholesteryl oleate but not PNPA or PNPC in SDM-1269/1331/1517, whereas protein kinase increases activity with both PNPA and PNPC in SDM-1517 (see Fig. 5), which expresses no CEH activity. This would be expected if protein kinase activation only occurs where steric effects are substantially limiting hydrolysis of a particular substrate. In the case of SDM-1517, steric properties that prevent productive binding of the bulky hydrophobic cholesteryl ester permit suboptimal levels of hydrolysis with the much smaller PNP esters. Here, conformational changes induced by phosphorylation enhance productive binding of the PNP esters but are insufficient to permit hydrolysis of cholesteryl esters. Moreover, stimulation of hydrolysis by phosphorylation is much greater with PNPC than with the substantially smaller PNPA, which should have easier access to a sterically hindered binding site. This provides additional support for the view that the effects of phosphorylation are mediated by a conformational change. By the same logic, substitutions in the binding pocket of the triple mutant, which are necessary for productive binding of cholesteryl oleate, should produce a more permissive binding site for smaller substrates, minimizing the effects of phosphorylation on hydrolysis of the much smaller PNP-esters.

As seen in Table II, the Ser⁵⁰⁶/Asn⁵⁰⁶ loop and its associated helix apparently serve an essential function, since it is remarkably conserved in a broad range of serine esterases from many species, irrespective of enzyme function. However, the Ser O-phosphorylation site associated with protein kinase activation is present in only 3 of the hundreds of homologous serine esterases from various species examined. Two of these are hncCEH and pI 6.1 esterase, which differ by a single amino acid substitution, the pI 6.1 esterase also expresses CEH activity and is activated by protein kinase.² The third is an adipose tissue homolog that is identical in sequence. Although some of the other homologs have cholesteryl esterase activity (pancreatic cholesteryl esterase/bile salt-stimulated lipase), none are known to be activated by protein kinase. Like LCE, most of the homologs with water-soluble substrates have Asn in this position, while most of the enzymes that utilize lipophilic substrates have Asp.

While these interpretations are consistent with the common features of crystal structures for similar esterases from several different species, they do not take into account bound water, hydrogen bonds, electrostatic interactions, or subunit interactions that would have specific localized effects on the three-dimensional structure. Nor do they attempt any analysis of the dynamics of substrate binding or catalysis. Thus, they should be viewed as tentative hypotheses to guide additional studies and to be tested more rigorously if and when crystal structures become available for hncCEH, LCE, or a more similar homolog.

	FOOTNOTES

^* This work was supported in part by National Institutes of Health Grant DK44613.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. Tel.: 804-828-9519; Fax: 804-828-1473; E-mail: grogan@hsc.vcu.edu.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) L46791 and L81144.

The amino acid sequences reported in this paper have been submitted to the Swiss Protein Database under Swiss-Prot accession numbers 1MAA, 1AKN, and 1EA5A.

Published, JBC Papers in Press, June 27, 2001, DOI 10.1074/jbc.M105644200

² T. J. Wallace, E. M. Kodsi, T. B. Langston, M. R. Gergis, and W. M. Grogan, unpublished data.

	ABBREVIATIONS

The abbreviations used are: hncCEH, hepatic neutral cytosolic cholesteryl ester hydrolase; LCE, lung carboxylesterase; PNPA, p-nitrophenylacetate; PNPC, p-nitrophenylcaprylate; aa, amino acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. Ito, U. Tchoua, M. Okamoto, and H. Tojo Purification and Properties of a Phospholipase A2/Lipase Preferring Phosphatidic Acid, Bis(monoacylglycerol) Phosphate, and Monoacylglycerol from Rat Testis J. Biol. Chem., November 8, 2002; 277(46): 43674 - 43681. [Abstract] [Full Text] [PDF]

				G. Haemmerle, R. Zimmermann, M. Hayn, C. Theussl, G. Waeg, E. Wagner, W. Sattler, T. M. Magin, E. F. Wagner, and R. Zechner Hormone-sensitive Lipase Deficiency in Mice Causes Diglyceride Accumulation in Adipose Tissue, Muscle, and Testis J. Biol. Chem., February 8, 2002; 277(7): 4806 - 4815. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 276/35/33165    most recent M105644200v1 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 HighWire Citing Articles via Google Scholar Google Scholar Articles by Wallace, T. J. Articles by Grogan, W. M. Search for Related Content PubMed PubMed Citation Articles by Wallace, T. J. Articles by Grogan, W. M. Social Bookmarking                      What's this?