Mice Expressing Only Covalent Dimeric Heparin Binding-deficient Lipoprotein Lipase

Lipoprotein lipase (LpL) hydrolyzes triglycerides of circulating lipoproteins while bound as homodimers to endothelial cell surface heparan sulfate proteoglycans. This primarily occurs in the capillary beds of muscle and adipose tissue. By creating a mouse line that expresses covalent dimers of heparin-binding deficient LpL (hLpLHBM-Dimer) in muscle, we confirmed in vivo that linking two LpL monomers in a head to tail configuration creates a functional LpL. The hLpLHBM-Dimer transgene produced abundant activity and protein in muscle, and the LpL was the expected size of a dimer (∼110 kDa). Unlike the heparin-binding mutant monomer, hLpLHBM-Dimer had the same stability as nonmutated LpL. The hLpLHBM-Dimer transgene prevented the neonatal demise of LpL knockout mice; however, these mice were hypertriglyceridemic. Postheparin plasma LpL activity was lower than expected with the robust expression in muscle and was no longer covalently linked. Studies in transfected cells showed that Chinese hamster lung cells, but not COS cells, also degraded tandem repeated LpL into monomers. Thus, although muscle can synthesize tethered, dimeric LpL, efficient production of this enzyme leading to secretion, and physiological function appears to favor secretion of a noncovalent dimer composed of monomeric subunits.

Triglycerides (TG) 1 in circulating lipoproteins are hydrolyzed by lipoprotein lipase (LpL). This enzyme is synthesized primarily in adipose and muscle, is transferred to the luminal surface of endothelial cells, and associates with heparan sulfate proteoglycans. This association with heparan sulfate proteoglycan is thought to localize LpL and direct liberated free fatty acids to the tissues where LpL is expressed (1). By creating transgenic mice expressing a heparin-binding mutant of human LpL (hLpL HBM ), we showed that mice expressing this protein have an alteration in tissue delivery of fatty acids derived from lipoprotein TG (2).
The association of a protein with heparin is thought to have several biochemical implications. In the case of LpL, heparan sulfate proteoglycan association fixes LpL to the endothelial surface (3). In addition, heparin stabilizes LpL activity (4). A similar process occurs for other heparin-binding proteins such as basic fibroblast growth factor. Like basic fibroblast growth factor, LpL is a homodimeric or oligomeric molecule, and its dissociation into monomers is thought to be part of its regulation (5,6). High affinity heparin binding could stabilize LpL by maintaining LpL secondary and tertiary structure, thereby modulating the dissociation of LpL into monomeric units.
In vitro studies have shown that active LpL can be produced when a short hinge is used to create a head to tail dimer of two monomeric subunits (7). Using this technique, we created a tandem repeat of hLpL HBM and expressed this protein in transgenic mice. This allowed us to study the role of heparin affinity in LpL stability and the importance of monomer to dimer assembly in the secretion of LpL from muscles.

MATERIALS AND METHODS
Generation of Transgenic Mice-Based on human heparin-binding site mutated LpL, hLpL HBM (2), a minigene was created that encodes for muscle-specific expression of two covalently linked hLpL HBM monomers in a head to tail configuration. First a PCR over the hLpL HBMminigene Bluescript vector (pBhLpL HBM ) was performed utilizing VENT polymerase (New England Biolabs, Beverly, MA), sense primer 1 (ctc cca cga gcg ctc cat tca), and antisense primer 2 (ttg ggg cga gcg ctc tcg agg cct gac ttc tta ttc aga gac tt). The PCR product contained the LpL HBM cDNA downstream from the unique Eco47/III restriction enzyme site. Furthermore, the stop codon was replaced by an XhoI site followed by an Eco47/III site. After Eco47/III digestion this fragment was inserted in the Eco47/III-digested pBhLpL HBM leading to the intermediate plasmid pBi1. A second PCR was performed utilizing the plasmid Xa-LpL-pcDNA 3 (7) as a template. By using the primers aac tat cgg cgg ccg cac tcg agg tcg aag gtc gtc tcg aag ccg acc aa (primer 3) and atg aat gga gcg ctc gtg gga (primer 4), a fragment was produced that contained a linker region (amino acids Val, Glu, Gly, Arg, Leu, and Glu), 3Ј-flanked by NotI and XhoI restriction enzymes sites and 5Ј followed by the human LpL cDNA starting with the coding region for the mature * This work was supported by Grants HL45095 (to I. J. G.), HL014990 (to A. B.), and HL028481 (to H. W.) from the National Institutes of Health and Sachbeihilfe LU855/2-1 (to E. P. L.) and Me1507/2-1 (to M. M.) from the Deutsche Forschungsgemeinschaft. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
LpL protein (gcc at position 256 in the hLpL cDNA open reading frame). By using the NotI and Eco47/III sites, this PCR product was inserted in pBhLpL HBM to produce the intermediate plasmid pBi2. This generated a fragment containing the linker and the mature hLpL sequence coupled to the hLpL HBM minigene downstream from the Eco47/III site including the heparin binding mutations and the 3Ј-untranslated region. Finally, the linker LpL fragment of pBi2 was inserted in pBi1 by using the XhoI and EcoRV restriction enzyme sites. The resulting hLpL HBM-Dimer minigene ( Fig. 1) was sequenced, purified, and used for production of transgenic mice. Founder animals were crossed with heterozygote LpL knockout mice that had been backcrossed seven times to C57BL/6. Pups heterozygote for both hLpL HBM-Dimer and the LpL knockout allele were crossed again with heterozygous LpL knockout mice mice. 12.5% of the subsequent pups were homozygous for the LpL knockout allele (LpL0) and heterozygote for the hLpL HBM-Dimer transgene. Mice expressing only hLpL HBM (2) and nonmutated hLpL (8) were bred in the same manner.
Genotyping of Transgenic Mice-Tail tip DNA was screened by PCR. The genotype at the mouse LpL locus was analyzed by the 3Ј PCR as described. The hLpL HBM-Dimer transgene was detected by oligonucleotides priming in the muscle creatine kinase promoter and the linker region. The hLpL HBM and hLpL transgenes were detected by utilizing the hLpL PCR (9).
Plasma Lipid and Lipoprotein Analysis-The mice were fed a chow diet (4.5% fat, w/w). The plasma samples were collected after 6 h of daytime fasting. Plasma TG and cholesterol were determined with kits (Sigma-Aldrich) in duplicate. For lipoprotein analysis fast performance liquid chromatography analyses of pooled plasma, the samples were performed as described (10). In addition, individual plasma samples (60 l) were ultracentrifuged twice in a Beckman TLA-100 rotor (Beckman Coulter, Fullerton, CA) as described previously (11).
LpL Mass and Activity Measurements-To obtain postheparin plasma (PHP), fasted mice were bled 5 min after a tail vein injection of 100 units of heparin/kg body weight (Elkins-Sinns, Cherry Hill, NJ). Human and murine LpL protein was measured by ELISA as described previously by Peterson et al. (12). LpL activity was measured by the method described by Hocquette et al. (13). To distinguish lipolysis mediated by human LpL from activity caused by mouse LpL and hepatic lipase, the mouse plasma samples were assayed in the presence of a monoclonal antibody against human LpL (14) and under high salt conditions (final concentration, 1 M NaCl).
LpL Stability-LpL stability was assessed using muscle homogenates of quadriceps muscles from transgenic mice. 100 mg of wet weight muscle was homogenized in 900 l of homogenization buffer (25 mM NH 4 -HCl, 5 mM EDTA, 0.8% (w/v) Triton X-100, 0.01% (w/v) SDS, 5 units/ml heparin, 1 g/ml pepstatin A, 10 g/ml leupeptin, 0.017 TIU/ml aprotinin, pH 8.2) as described (13). After a 30-min centrifugation at 20,000 ϫ g and 4°C, the supernatants were frozen in aliquots (Ϫ70°C) and then defrosted just prior to stability experiments. To assay LpL stability, the LpL containing muscle homogenates were incubated in a water bath at 37°C. At the indicated time points the aliquots were taken and immediately assayed for LpL activity.
Northern and Western Blot Analysis of LpL from Quadriceps Muscles-For Northern blots, total RNA was prepared from quadriceps muscles and hearts of 12 week-old hLpL HBM /LpL0 and hLpL HBM-Dimer / LpL0 male mice using a kit (TRIzol™ reagent; Invitrogen). 15 g of total RNA were applied to a 1% agarose gel and blotted to a Hybond-N Nylon Membrane (Amersham Biosciences) using standard techniques. The blot was hybridized with a probe spanning the ϳ500-bp RsrII/ HindIII fragment from the hLpL minigene and then autoradiographed. To confirm that equal amounts of RNA were loaded to each lane, the blot was probed for glyceraldehyde-3-phosphate dehydrogenase.
For Western blot analyses muscle homogenates of 20 week-old hLpL/ LpL0 and hLpL HBM-Dimer /LpL0 male mice were obtained exactly as described above. To increase sensitivity, the homogenates were partially purified by heparin-Sepharose essentially as described (15,16). Briefly, 10 l of resuspended heparin-Sepharose was added to 1 ml of muscle homogenate and incubated for 3 h at 4°C. The Sepharose was washed with a Tris-buffer containing 0.75 M NaCl. LpL then was eluted with 50 l of a Tris buffer containing 1.5 M NaCl. SDS-PAGE under reducing conditions with 5 l of eluates and blotting to nitrocellulose membranes were performed using standard techniques. The membrane was incubated with a polyclonal rabbit anti-bovine LpL antibody (SA1357; 1:2500 dilution) and an anti-rabbit IgG (Sigma-Aldrich; 1:5000 dilution). For SA1357 production, bovine LpL was kindly provided by Dr. G. Olivecrona (Umea, Sweden). The antibody was produced by Eurogentec (Herstal, Belgium). The bands were visualized by ECL (Amersham Biosciences).
Expression of Dimeric LpL in Cell Culture-A full-length cDNA for a tandem repeat dimeric LpL, LpL TR , has been reported (7). The cDNA was purified, digested, and inserted into the pcDNA3 expression vector (Invitrogen). COS-7 and Chinese hamster lung (CHL) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics. Co-precipitates of plasmid DNA and CaPO 4 were prepared according to the manufacturer's instructions to mediate cell transfections. The calcium phosphate-DNA mixtures were incubated at room temperature for 30 min prior to addition to 50 -60% confluent cells. For transient transfection of COS cells, the calcium phosphate-DNA mixture was left overnight on the cells and then removed by a rinse with phosphate-buffered saline. The cells were then treated with 3 ml of 10% (v/v) Me 2 SO in phosphate-buffered saline for 2.5 min. After removal of the Me 2 SO solution, fresh Dulbecco's modified Eagle's medium supplemented with serum substitute (Nutridoma; Roche Applied Science) and 20 units/ml heparin was added to each dish. The medium was harvested daily for a 3-day period and stored at Ϫ80°C. Stably transfected CHL cells were selected by growth in the presence of Geneticin (G418, 500 g/ml; Sigma-Aldrich). The surviving colonies were expanded and assayed for lipase activity. The cell clones with the highest levels of lipolytic activity were used for expression in T-225 flasks containing Dulbecco's modified Eagle's medium, supplemented with 1% Nutridoma and 10 units/ml heparin. The medium was harvested daily for up to 30 days and stored at Ϫ80°C.
Heparin-Sepharose Chromatography-Affinity chromatography was performed by using a fast performance liquid chromatography system (Amersham Biosciences) with a 1-ml Hi-Trap heparin-Sepharose column (Amersham Biosciences) at 4°C essentially as described (17). The cell culture supernatants of hLpL HBM-Dimer or native hLpL expressing cells were loaded to the column. The LpL mass concentrations in the fractions were determined by ELISA as described (17).
Western Blots of PHP and Cell Culture LpL-PHP samples were subjected to sucrose gradient centrifugation and immunoprecipitation exactly as described (18). For Western blot, the samples were mixed with a half-volume of buffer containing 2% SDS, 0.1 M Tris-HCl, pH 6.8, 50% glycerol, 10% ␤-mercaptoethanol, 0.05% bromphenol blue. The mixtures were placed in boiling water for 5 min prior to loading onto a 7% acrylamide gel. The gels were electroblotted onto nitrocellulose and subsequently blocked for 1 h in Tris-buffered saline containing 3% bovine serum albumin. After blocking, the membrane was incubated overnight at 4°C with a monoclonal antibody that detects human, but not mouse, LpL (5D2, kindly provided by J. Brunzell). The 5D2 epitope is in the C-terminal portion of native LpL (19). Immunoblots were developed with anti-mouse IgG conjugated to biotin. After washing, the blot was incubated with streptavidin conjugated to horseradish peroxidase. Immunoreactive bands were visualized with chemiluminescent reagents (Pierce) and exposure to x-ray film.
Histological Analysis-Quadriceps muscles were dissected from 6-month-old male hLpL HBM-Dimer /wild type mouse LpL and wild type littermates and fixed in formaldehyde. Histological analyses were performed as described (20).  (2,8), a construct encoding for a covalent dimeric LpL HBM was made. The stop codon of the first LpL HBM cDNA was replaced by a linker region followed by a second LpL HBM cDNA lacking the sequence for the signal and the native 3Ј-untranslated region. The position of the primers used is indicated by arrows (P1, P2, P3, and P4). The primer sequence can be found under "Materials and Methods." The sequence for the hinge region was identical to that described by Wong et al. (7) except that an isoleucine in position 2 was changed to a valine. hLpL HBM-Dimer was chromatographed. Elution of native LpL from heparin-Sepharose leads to two peaks: one peak eluting at Ͼ1 M NaCl that contains active dimeric enzyme (Fig. 2) and, commonly, a second peak that elutes with ϳ0.75 M NaCl and is not associated with activity. Presumably this second peak is inactive monomeric or misfolded protein. LpL HBM primarily is found in the latter peak, as has been previously reported (2). LpL HBM-Dimer protein eluted with a single peak at ϳ0.9 M NaCl. Thus, the construct produced an enzyme whose heparin affinity was lower than that of nonmutated LpL but greater than that of monomeric LpL.

In Vitro
Dimer Expression in Muscle-Our studies of the transgenic mice focused on the actions of the hLpL HMB Dimer transgene on the LpL-deficient background because this allowed measurements of LpL without interference from the native enzyme. Northern blots were performed to verify that the hLpL HBM-Dimer construct was expressed in muscles from mice. As shown in Fig.  3A, the mRNA of the hLpL HMB Dimer /LpL0 mice expressing the dimeric construct led to a larger transcript than that found in hLpL HBM /LpL0 or hLpL/LpL0 mice. The line of mice that was bred also had much greater expression of the transgene. Moreover, these mice produced a larger LpL protein of ϳ110 kDa in muscle tissue (Fig. 3B). This is the expected size of the dimeric protein.
Enzyme Activity and Protein in Muscles-Homogenates of left quadriceps and hearts from mice expressing hLpL, hLpL-HBM , and hLpL HBM-Dimer constructs on the LpL0 background (n ϭ 3 each line) were assayed for LpL activity and mass. The average activity found in the hLpL HBM-Dimer /LpL0 muscles (Fig. 4A) was similar to that in the hLpL HMB transgenics and greater that that found in mice expressing nonmutated human LpL (8).
LpL protein was assayed in muscle homogenates from these three lines of mice. In concert with the RNA and activity data, hLpL HBM-Dimer muscles contained ϳ1.5 times the LpL protein of hLpL HBM and more than 10 times as much protein as the line of hLpL-expressing mice (Fig. 4B). The specific activities in muscle tissues were 0.7 for hLpL HBM-Dimer versus 1.1 for hL-pL HBM and 2.8 mol of free fatty acid/g/h for hLpL-expressing mice on LpL0 background (Fig. 4C).
Muscle Histology-Histological analysis of quadriceps muscles from hLpL HBM-Dimer transgenic mice revealed no major pathological changes. A slight increase in lipocytes in the endomysium and subsarcolemmal enhancement of periodic acid Schiff-positive material, minor histopathologic changes, were noted in hLpL HBM-Dimer mice. Most significantly, no lipolysisinduced myopathy like that in mice overexpressing hLpL in muscles (21) was seen in these animals.
Stability of LpL Activity from Muscles of hLpL HBM-Dimer Mice-To determine whether the dimeric configuration altered the stability of the mutated LpL, muscle homogenates from each line of mice were incubated at 37°C, and the activity was assessed over time (n ϭ 3 each line). A slow decay of activity was found in muscles expressing both hLpL and hLpL HBM-Dimer (Fig. 4D). In contrast, hLpL HBM was less stable. Therefore, the instability associated with the mutation in heparin binding was corrected by creating the tethered dimer. were separated on a 1% agarose gel, blotted to a nylon membrane, and hybridized with the ϳ500-bp RsrII/HindIII fragment of the hLpL minigene. B, Northern blot of heart muscle: Also 15 g of total RNA of hearts of hLpL HBM-Dimer /LpL0 (lane1), hLpL HBM /LpL0 (lane 2), and hLpL/LpL0 mice (lane 3) were processed as described above. C, Western blot of quadriceps muscle homogenates of hLpL and hLpL HBM-Dimer expressing mice. LpL from muscle homogenates was partially purified by heparin-Sepharose chromatography. hLpL muscles show a major immunoreactive band of ϳ55 kDa. hLpL HBM-Dimer muscle contains an LpL of ϳ110 kDa. Bovine LpL, run as a control, also shows a ϳ55-kDa immunoreactive band. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Lipids and Lipoproteins in hLpL HBM-Dimer Transgenic
Mice-Plasma lipids and lipoproteins were assessed in the hLpL HMB Dimer mice and were compared with the other transgenic lines. Although the hLpL HMB Dimer transgene rescued the LpL knockout mice, plasma TG in these mice was more than twice that found in animals expressing either the native hLpL or the hLpL HBM transgene (Table I). This was due to an increase in very low density lipoprotein TG as noted both by ultracentrifugation (Table I) and by gel filtration chromatography (Fig. 5). Plasma cholesterol was not different between these mice. Thus, despite the robust expression of the hLpL HBM-Dimer transgene, these mice have higher TG. This suggested that either the dimer construct was less active in the plasma or that the amount of this LpL in the PHP did not parallel the muscle expression (Fig. 4). LpL Activity and Mass in PHP-The PHP LpL activities and masses of the three lines of mice are shown in Fig. 6. Surprisingly, hLpL HBM-Dimer /LpL0 mice had less LpL activity than mice expressing either hLpL HBM or hLpL despite the greater expression of this protein in the muscle. The specific activities in PHP were 0.7 for hLpL HBM-Dimer versus 1.5 for hLpL HBM and 4.5 mol of free fatty acid/g/h for hLpL expressing mice on LpL0 background. Thus, activity of the dimeric protein was unexpectedly low, which is consistent with the increased TG in these mice.
PHP LpL Western Blots-Because muscle, but not PHP, from hLpL HBM-Dimer mice contained more LpL activity, we questioned whether the dimeric protein was altered either prior to or after its secretion. To test this the protein was first isolated using sucrose gradients; active LpL sedimented as a dimeric protein in all three samples (data not shown). Surprisingly, Western blots of PHP LpL from the hLpL HBM-Dimer transgenic mice produced a band the same size as bovine LpL and wild type human LpL, i.e. PHP did not contain covalently linked LpL dimers (Fig. 7). Thus, either the muscle did not secrete the covalent dimer LpL or the protein was degraded after its secretion.
Expression of Dimeric LpL in Vitro-To test the possibility that the LpL dimer construct is "clipped" by cells, we expressed dimeric LpL protein in two different cell lines. To rule out the possibility that clipping of the dimeric construct was due to the C-terminal heparin-binding mutations, we utilized nonmutated covalent dimeric LpL for cell culture expression (LpL TR ) (7). LpL TR was expressed in CHL cells and in COS cells. Western blots of medium from COS cells showed a dimeric-sized LpL protein (Fig. 8A). However, when the same construct was expressed in CHL cells, the medium contained only monomeric LpL. Thus, some cells and presumably tissues like muscle are unable to secrete dimeric forms of LpL. Western blot of cell fractions of CHL and COS cells expressing LpL TR showed that CHL cells contained intracellular products of the dimer of ϳ 85, 66, and 63 kDa (Fig. 8B).
To further investigate where clipping of the covalent dimeric protein occurs, we looked for protease activity of cell culture supernatants of CHL cells. We mixed media of nontransfected CHL cells with supernatants of COS cells expressing LpL TR and incubated the samples for 1 h at room temperature. Then of hLpL HBM-Dimer /LpL0 mice was as high as in hLpL HBM expressing muscles but about three times higher than in hLpL expressing muscles. B, immunoreactive LpL mass was the highest in hLpL HBM-Dimer expressing muscles. C, specific LpL activity was the lowest in hLpL HBM-Dimer expressing muscles. D, the activity stability was measured by incubating muscle homogenates at 37°C and assaying aliquots for LpL activity at the indicated time points. LpL activity in hLpL HBM-Dimer muscles (closed circles) was comparable with that of nonmutated hLpL (open triangles), whereas hLpL HBM (open squares) showed reduced stability. FIG. 4. LpL activity, mass, and stability of hLpL HBM-Dimer . Homogenates of quadriceps muscles of transgenic mice on the LpL knockout background were prepared as described. A, LpL activity in muscles we assayed for lipase activity and analyzed the mixes by LpL Western blot. Nontransfected CHL medium did not affect lipase activity of transfected COS cell media, and no additional monomeric sized LpL band was detected by immunoblot, i.e. no proteolysis occurred. Also when medium from CHL cells that were stably transfected with LPL TR was mixed with the LpL TR -COS medium, the activity and size in Western blot analysis of the covalent dimeric protein remained unaffected. These results are in concert with additional experiments in which protease inhibitors like benzamidine or phenylmethylsulfonyl fluoride did not inhibit the clipping of LpL TR expressed by CHL cells. In conclusion there is no evidence for the involvement of an extracellular protease in this process, at least in the in vitro setting.

DISCUSSION
Several processes are thought to be central to the regulation of LpL activity in vivo: 1) LpL is expressed as a monomeric  protein and then assembled into a dimer to allow its physiological action; this appears to be a complex process (20). 2) LpL actions are limited by LpL dissociation from heparan sulfate proteoglycan on capillaries and its clearance in the liver (21). 3) The dissociation of the dimeric protein into subunits in the bloodstream then prevents further lipolysis from occurring. In a previous study we showed that a mutation in a heparinbinding region of LpL led to instability of the dimeric LpL complex. Large amounts of this mutated LpL were found in the pre-heparin plasma; however, we were unable to determine whether defective heparin binding per se or conversion of active LpL to monomers was the primary defect leading to enzyme instability (2). By creating a hinged dimer of heparin-binding defective LpL, we hoped to better understand the roles of heparin association and dimer to monomer interconversion in LpL biology. Our studies showed the following: 1) hLpL HBM-Dimer associated with heparin-Sepharose; however, it was dissociated with ϳ0.9 M NaCl. Thus, its affinity for heparin was reduced compared with native LpL. 2) hLpL HBM-Dimer was expressed in high levels in muscles of transgenic mice, and the dimeric mRNA and protein were demonstrated. 3) Tissues from these mice had LpL activity with greater stability than that found with hLpL HBM . 4) Although the hLpL HBM-Dimer transgene rescued LpL knockout mice, the mice were hyperlipidemic. This suggested that the LpL was less active. 5) Despite the robust expression of this transgene in muscle, plasma LpL activity was less than in transgenic mice expressing hLpL and hLpL HBM , and the specific activity of the LpL was the lowest of the three lines of mice. Moreover, no covalently linked dimeric protein was found in the PHP of the hLpL HBM-Dimer transgenic mice. Thus, muscle inefficiently produces dimeric LpL, and conversion of dimer to monomers must occur either prior to or after secretion. 6) Similar clipping of the dimer occurred in CHL cells but not COS cells. Therefore the ability to secrete dimeric LpL is limited to only some types of cells. 7) Monomeric sized LpL fragments were found within these cells. Thus, it is likely that the clipping occurs intracellularly.
In our previous studies of mutated LpL, alteration of Cterminal basic amino acids led to two effects: 1) reduced association of the protein with heparin and 2) reduced stability of the protein. By introducing this mutation into an obligate dimer, we hoped to dissect the roles of each of the processes in LpL biology. We first showed that hLpL HBM-Dimer was defective in heparin binding. Like the active forms of hLpL HBM , this protein eluted from a heparin affinity column at 0.8 -0.9 M NaCl rather than Ͼ1.2 M NaCl, the elution of native LpL. When LpL HBM-Dimer was expressed within muscle, unlike hLpL HBM , dimeric protein had normal stability. Thus, tethering restored stability and suggested that defective heparin-association alone likely led to more rapid monomerization, which in turn reduced LpL activity.
Several observations suggested that the dimeric LpL was processed aberrantly in vivo. Despite the robust expression of the transgene and large amounts of activity in the muscle homogenates, hLpL HBM-Dimer mice had higher TG than mice expressing hLpL and even hLpL HBM . Moreover, the specific activity of PHP hLpL HBM-Dimer was lowest in these mice. This led us to question whether hLpL HBM-Dimer was altered in some manner. Western blot analysis of PHP demonstrated only monomeric LpL-sized enzyme. Thus, the protein was inefficiently secreted because the PHP did not have more protein despite more robust muscle expression, and the dimer was clipped into monomeric units either prior to or after secretion from the muscle.
Because culture of the myocytes from the transgenic mice was not a viable option, we studied other cells to determine whether they also degraded dimeric LpL into monomers. Moreover, we elected to study this effect using LpL that was not mutated so we could assess whether the clipping was due to the dimeric construct and not another mutation in the LpL. LpL TR was efficiently produced and secreted from COS but not from CHL cells. This surprising observation allowed us to understand the observations that we had in the hLpL HBM-Dimer mice. Skeletal muscles, like CHL cells, appear to be unable to secrete active dimeric LpL and clip the protein, which leads to only monomeric forms within the PHP.
Several experiments were performed to determine the site of dimer proteolysis. Most interesting, we found that CHL but not COS cells secreted clipped dimer. Thus, the ability to produce dimeric forms of LpL was cell-specific and conceivably might be tissue-specific in vivo. We attempted to define where the proteolysis occurred. Medium from CHL cells did not appear to have a protease, and lysosomes inhibitors did not prevent cleavage of the dimer. However, intracellular cleavage of dimer occurred in the CHO cells. Thus, we postulate that dimeric protein is clipped within cells prior to its secretion.
LpL is not unique in its requirement for dimeric association to optimize its biologic activity. Moreover, a number of dimeric molecules associate with heparin. Our data suggest that heparin maintenance of the dimer, rather than heparin association itself, stabilizes LpL because activity in muscles expressing mutated dimers was as stable as nonmutated LpL.
Because the requirements for dimerization do not appear to be the most efficient method to produce an important metabolic enzyme, there have been a number of hypothesizes to explain this mode of synthesis. One obvious reason might be to allow for inactivation of the enzyme, thus preventing excessive local lipolysis and accumulation of toxic levels of reactive lipid products. Our data suggest an entirely new paradigm for monomeric protein production. As we found, in vivo, myocytes are incapable of efficient production of tethered dimeric LpL. In contrast, it is well established that LpL monomers assemble as dimers prior to their secretion from several cells types (20). Either myocytes are different, or the requirements for secretion of the tethered dimer differ from that of the assembled homodimer. Although our studies are limited to LpL and its production in muscle, we hypothesize that the secretion pathways of other dimeric heparin-binding proteins are cell-specific. Maybe an intracellular or less likely an extracellular protease, in some cells, specifically prevents production of tethered dimeric proteins.