Biosynthesis of Lipoprotein Lipase in Cultured Mouse Adipocytes I. CHARACTERIZATION OF A SPECIFIC ANTIBODY AND RELATIONSHIPS BETWEEN THE INTRACELLULAR AND SECRETED POOLS OF THE ENZYME*

Polyclonal antibodies have been raised in rabbits against homogeneous lipoprotein lipase (LPL) purified from the media of adipose 3T3-F442A cells. The antibody is able to inhibit the apolipoprotein C-II-depend- ent activity of LPL, to immunoprecipitate LPL under nondenaturating conditions from media and cellular extracts. A dot-blot immunoassay of secreted LPL is also described (range 0.1-0.7 milliunits). The secretion potential p, taken as the ratio of total releasable activity or antigen to initial cellular activity or antigen, was determined. This was shown in cells treated with heparin and cycloheximide to be equal to 1 for LPL antigen but significantly greater than 1 for LPL activity assayed under standard conditions. No LPL was actually degraded within the cells. A dramatic en-hancement of the intracellular activity was induced by a mere dilution of detergent-treated cell lysates with no change in LPL antigen. The total intracellular ac- tivity reached a plateau at a value which now became identical to that obtained in the medium of cells ex- posed to heparin and cycloheximide. The existence of an inhibitor of LPL activity has been excluded as well as that of an

Polyclonal antibodies have been raised in rabbits against homogeneous lipoprotein lipase (LPL) purified from the media of adipose 3T3-F442A cells. The antibody is able to inhibit the apolipoprotein C-II-dependent activity of LPL, to immunoprecipitate LPL under nondenaturating conditions from media and cellular extracts. A dot-blot immunoassay of secreted LPL is also described (range 0.1-0.7 milliunits). The secretion potential p, taken as the ratio of total releasable activity or antigen to initial cellular activity or antigen, was determined. This was shown in cells treated with heparin and cycloheximide to be equal to 1 for LPL antigen but significantly greater than 1 for LPL activity assayed under standard conditions. No LPL was actually degraded within the cells. A dramatic enhancement of the intracellular activity was induced by a mere dilution of detergent-treated cell lysates with no change in LPL antigen. The total intracellular activity reached a plateau at a value which now became identical to that obtained in the medium of cells exposed to heparin and cycloheximide. The existence of an inhibitor of LPL activity has been excluded as well as that of an increase in the catalytic activity of LPL during its secretion, before or after exposure to heparin. Our results indicate a systematic underestimation of LPL intracellular activity and suggest that LPL is present within intracellular cisternae in a cryptic state. This potential activity can be fully unmasked in vitro. In agreement with other data (Vannier, C., and Ailhaud, G., (1989) J. Biol. Chem. 264, 13206-132 16), our results appear to exclude the existence of a reservoir of catalytically inactive LPL molecules within adipose cells.
Lipoprotein lipase (LPL),' synthesized in parenchymal cells of tissues of mesodermal origin, is the key enzyme responsible for the hydrolysis of plasma triglycerides from apolipoprotein * This work was supported by Grant CRE 857014 from the "Institut National de la Sant6 et de la Recherche M6dicale" (to C. V.) and by the "Centre National de la Recherche Scientifique" (UPR 7300). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertlsement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The abbreviations used are: LPL, lipoprotein lipase; apo, apolipoprotein; PMSF, phenylmethylsulfonyl fluoride; DMEM, Dulbecco's modified Eagle's medium; PBS, phosphate-buffered saline; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; FBS, fetal bovine serum. C-11-containing lipoproteins at the capillary endothelium (1). It has long been known that rapid regulatory tissue-specific changes in its activity occur during fasting and refeeding, illustrating its inverse and coordinate expression in muscle and adipose tissue. This regulation allows fatty acids to be directed according to the metabolic requirements of the parenchymal cells (1)(2)(3). It is well established that, in response to nutritional changes, LPL activity at the endothelium can vary dramatically (4)(5)(6), whereas it is not greatly changed in adipocytes (1,7).
Numerous studies performed on intact fat pads as well as on isolated adipocytes and cells of preadipocyte clonal lines have documented the role of hormones and other factors in the control of LPL synthesis and expression in these cells (8).
In that respect, insulin and corticosteroids are now recognized to stimulate the synthesis of the enzyme (g-ll), whereas catecholamines seem to be involved, as are other lipolytic agents, in the acceleration of its degradative process (12,13). These various hormones could thus directly or indirectly control the movement of LPL between adipocytes and endothelial cells. However, the mechanisms whereby these events take place have received less attention. With few exceptions (14,15), adipose cells exhibit a slow constitutive secretion of active LPL. Heparin is known to stimulate the secretion of active LPL without affecting its cellular activity (12,16,17), and this stimulation can also occur in the absence of protein synthesis; so far, heparin remains the most powerful effector of this process. It has been hypothesized that, independently of protein synthesis, a hormone-stimulated intracellular translocation of LPL, leading in turn to its secretion, is accompanied by an increase in its catalytic activity (7,18,19). In adipocytes, Robinson et al. (10) have suggested that LPL exists before secretion partly as an inactive precursor. This putative precursor pool or storage form of inactive LPL molecules has not yet been characterized.
Our previous studies have shown that the effect of heparin on LPL secretion from differentiated Ob17 adipose cells, among the few which do not exhibit a constitutive enzyme secretion, was to induce a rapid mobilization of an intracellular store of active molecules (9,20). In full agreement with other data from subcellular fractionation and immunotitration studies (21), we have provided compelling morphological and biochemical evidences for LPL being mainly localized in the Golgi apparatus (20, 22), thus establishing its status as a secretory protein in the adipocyte. These investigations in Ob17 cells have also shown that LPL acquired its enzymatically active conformation in the proximal cisternae of the Golgi apparatus (20) and that no further increase in its catalytic activity occurred during secretion.
The aim of the present and accompanying papers (23) was to delineate more precisely the relationships between catalytic activity and protein content during the export of LPL, using a combination of immunological and biochemical approaches. These studies have been carried out with Ob17SA16 and 3T3-F442A adipose cells which display both a constitutive and a heparin-stimulated secretion. In this first paper, we report the characterization of a rabbit antibody raised against mouse LPL and its use in immunoassays of the cellular and secreted enzyme. It is shown that no inactive precursor is present within the cells but instead it is proposed that active LPL molecules are stored as a pool whose activity is masked upon condensation. In the second paper, the main features of the biosynthesis and intracellular processing of LPL, as well as its oligomerization, are reported. The existence of a regulated pathway for LPL secretion in adipocytes is documented.

Lipoprotein Lipase Activity and Lipoprotein Lipase Antigen in Adipose Cells and in Medium-Our previous investigations
on LPL secretion had been focused on the determination of the true LPL activity releasable under heparin stimulation by Ob17 cells (15,20). This activity was accurately measured only by using the continuous flow technique by means of the combined knowledge of the secretion rate and the rate constant of enzyme inactivation (15). A typical time course of LPL secretion from Ob17SA16 cells is shown in Fig. 3a. The first phase of secretion, which terminated 40-45 min after heparin addition, was independent of protein synthesis as it was not altered by cycloheximide. This phase thus corresponds to the secretion of a preexisting pool of intracellular molecules and occurred at a higher rate than that of newly synthesized molecules which are released during the second phase of secretion. The curves indicate that the constitutive secretion of LPL occurred at a constant rate which was similar to that found for the release of newly synthesized enzyme molecules in cells exposed to heparin. The conditions of Fig.  3a allow the comparison of the actual sizes of the initial cellular and releasable pools of LPL activity. It was found that the secretion potential p (see Miniprint, Table I), measured in the absence of protein synthesis and under heparin stimulation was higher than 1. Therefore, two hypotheses could be proposed to explain this increase of LPL activity upon secretion. First, some cellular or operational factor(s) may have led to an underestimation of LPL intracellular activity. Second, a large majority of the intracellular LPL molecules might be enzymatically inactive but were activated upon secretion.
The first hypothesis was excluded because cell lysates, depleted of LPL by affinity chromatography, were unable to inhibit the activity of the purified enzyme. Additional control experiments excluded also the possibility of lipids and other intracellular components as well as Triton X-114 molecules interfering with LPL assays (not shown). Moreover, the linearity of LPL assays is maintained for activities up to 20-25 milliunits/ml, corresponding to the release of 12-15% of the total fatty acyl chains. Because the amounts of LPL protein in the cell lysates and in the secretion media were adjusted to be similar and because the activities of LPL determined in the secretion media were below 20-25 milliunits/ml, under-  calculated from the secretion rates obtained by assaying the enzymatic activity secreted every 5 min (dU/dt). b, secretion of LPL by 3T3-F442A cells exposed to heparin and cycloheximide at the above concentrations. Inset, intracellular LPL activity of control cells not exposed to heparin and cycloheximide. Cell lysate prepared at the standard dilution (see "Experimental Procedures") was further diluted with Buffer A containing 150 mM NaCl and 3 pg/ml heparin, in the absence of Triton X-114, up to the indicated final dilution. The measured activity as total milliunits/dish was corrected for dilution.
estimation of LPL activity due to a lack of substrate could also be excluded.
However, when cell lysates were diluted above the standard 2-fold dilution routinely used, the total activity of LPL increased in parallel and as a function of the dilution factor. At final dilutions ranging from 32 to 128, a plateau was reached. This plateau gave the true intracellular LPL activity as its value (determined before heparin addition) was equal to that found for the total heparin-releasable LPL activity ( Fig. 3b  and inset). Thus, as determined for LPL antigen, the secretion potential p determined for activity had now become equal to 1. The fact that this value was simply reached upon dilution of cell lysates prepared in detergent-containing buffer, rather than supporting the hypothesis of a precursor of LPL being activated upon secretion, argued for a systematic underestimation of the cellular activity.
In order to support. this tentative conclusion, further experiments were carried out using the solubilized membrane fraction of the post-nuclear supernatant obtained from cell homogenates of 3T3-F442A cells ( Fig. 4) because (i) this fraction contains more than 90% of the cellular activity in Ob17 cells (22), and (ii) immunoprecipitation assays indicated that it contained the total antigen pool of 3T3-F442A cells (not shown). As indicated by the results of Fig. 4a, very low LPL activity was detected in the concentrated membrane lysate. This activity was dramatically increased upon dilution. A plateau was obtained, corresponding to a 50-fold increase in the total activity when compared to the initial value obtained at the standard dilution. This increase could vary from experiment to experiment according to the activities of LPL determined before dilution and which were within the limits of sensitivity of the assay. When heparin was added to the dilution buffer at a concentration of 3 pg/ml (maximal heparin concentration used in secretion media and in experiment of Fig. 3b), a further %fold increase in the plateau value was observed. This increase remained unchanged with higher concentrations of heparin (Fig. 4a). The effect of heparin was not specific since it could be mimicked by adding heat-inactivated LPL (Fig. 4a). Additional experiments showed that ( i ) the higher the LPL concentration in the lysate, the higher was The post-nuclear membrane fraction of 3T3-F442A cells, grown and differentiated in 60-mm dishes, was lysed with Triton X-114 as described under "Experimental Procedures." The lysate at the standard dilution was used as the starting material. LPL activity was assayed after serial dilution (up to 128-fold final dilution), in Buffer A containing 150 mM NaCl, with or without the indicated supplementation. Cellular activities were calculated after correcting the measured activity for dilution. a, effect of dilution in buffer alone (e); or in buffer supplemented either with 3 pg/ml (A) or 75 pg/ml (A) heparin; or with heat-inactivated, purified mouse LPL added at concentrations corresponding to 22.4 milliunits/ml ( . ) and 89.7 milliunits/ml (0) native enzyme. Results are expressed as percent of the maximal activity obtained at the plateau in the presence of 3 pg/ml heparin (122 milliunits/dish). b, LPL activity was measured in the heparin-containing secretion medium (0) and in the lysed membrane fraction (e) of two parallel sets of cells, after dilution in Buffer A containing 150 mM NaCl and 3 pg/ml heparin. The measured activity was corrected for dilution and was expressed as total milliunits/dish. Inset, activity measured, and not corrected for dilution, in the diluted samples of secreted medium (0) and cell lysate (e) and expressed as milliunits/ml. Secretion of LPL was studied in ITT medium using the continuous flow technique (see "Experimental Procedures"). After equilibration of medium flow through the chamber ( t = 0 ) , the effluent was collected every 5 min for 55 min (constitutive secretion) before the addition (arrow) of 3 pg/ml heparin in the medium entering the chamber (stimulated secretion). In each fraction (700 p1 final volume), 200-pl aliquots were used for the LPL activity (0) and for the quantitative dot-blot assays (e). Plotted are the integral curves of activity and antigen secretions.
Inset, secretion rates of LPL activity (0, dU/dt) and antigen (e, dA/ dt) expressed, respectively, as milliunits and as optical density units/ fraction. the increase due to the addition of heparin or heat-inactivated LPL and (ii) the higher the LPL concentration in the cell lysate, the higher was the half-maximally effective concentration of heparin. The same phenomenon of "masked" intracellular activity is further illustrated in the experiment of Fig.   4b which shows clearly a 300-fold increase in the LPL activity after a 64-fold dilution of the cell lysate whereas extrapolation of the curve indicates that no LPL activity should be detectable in the absence of dilution, i.e. before homogenization, that is to say intact cells. The results of Fig. 4b also clearly indicate that no effect of dilution was observed on the total activity of the secreted enzyme. For LPL concentrations in the range of 0.1-6.5 milliunits/ml, not surprisingly because determinations were performed in the linear part of the assay curve (see above), the measured activity is a hyperbolic function of the dilution factor. By contrast, at low dilutions a bellshaped curve, instead of a plateau, is obtained for the cellular activity (Fig. 4b, inset).
Secretion of Lipoprotein Lipase in the Presence or Absence of Heparin-A dot-blot immunoassay was used to study the kinetics of LPL secretion carried out in the absence and then in the presence of heparin (Fig. 5 ) . This was performed using the continuous flow technique, by taking advantage of the irreversible adsorption of LPL to nitrocellulose in the absence or in the presence of very low concentrations of contaminant proteins, which are conditions occurring in I T T medium (see "Experimental Procedures"). The integral secretion curves show that, before and after heparin addition, there was a parallel increase in both enzyme activity and titratable antigen illustrated by the dramatic increase in the secretion rates of both parameters (Fig. 5, inset). For the fractions collected from 0 to 90 min a linear correlation ( r = 0.97) was obtained for the two parameters. Whether or not heparin was present, the specific activity of the secreted LPL was found to be 0.71 k 0.04 absorbance unit/milliunit, in agreement with the values reported in Fig. 2. It must be stressed that the experiment of Fig. 5 failed to detect any change in the molecular activity of LPL. This result demonstrates unambiguously that LPL did not become more active during secretion into the medium in the presence of heparin.

DISCUSSION
The elucidation of the mechanisms involved in the synthesis and secretion of functional LPL at the molecular level was facilitated by the preparation of an antibody specific to the mouse enzyme permitting the use of a homologous antigenantibody interaction. Changes in the catalytic properties of LPL during its processing can only be interpreted accurately by means of an antibody allowing access to the antigenic structure specific to inactive and/or precursor forms. This goal has been reached in the present work, which represents the first successful attempt to immunize a rabbit against mouse LPL. Owing to its monospecificity and high titre, this antibody, which interacts specifically with the mature form of LPL secreted by the cells, has proven to be suitable for immunoprecipitation, immunotitration and immunoblotting experiments (see Ref. 23).
So far, a few polyclonal antibodies have been successfully raised against LPL from rodents. The anti-rat LPL antisera obtained by Jansen et al. (35) and Etienne et al. (36) have allowed us in previous studies to delineate the induction, the intracellular activation and the localization of LPL in mouse adipose Ob17 cells (9,20,22), since these anti-rat LPL antisera raised in goats were able, via a cross-reaction with the mouse enzyme, to inhibit the enzymatic activity and to recognize LPL antigenic sites in formaldehyde-fixed cells. Unfortunately, these antibodies proved not to be suitable for the immunological studies reported herein and performed with mouse Ob17SA16 and 3T3-F442A cells. On the other hand, Al-Jafari and Cryer (14,21) have recently used antisera directed against rat heart LPL, raised in chicken or guinea pig, to develop an immunoassay for LPL in rat adipocytes. Recently, Schotz and colleagues (37) also reported an enzymelinked immunoassay for rat LPL using an antibody raised in chicken against the bovine enzyme. Other antisera have been prepared by Olivecrona and colleagues by immunizing rabbits and hens with bovine milk LPL (38), and goats with guinea pig milk LPL (39). The cross-reactivity displayed by these antibodies with mouse LPL, checked by their ability to inhibit the enzyme activity, did not seem to have an evolutionary basis. In one of these cases only, advantage was taken of the inhibitory effect of the antiserum to detect LPL in mouse adipose cells by immunoprecipitation (34, 40).
We have reinvestigated in the present study the quantitative relationships between the intracellular and the releasable pools of LPL antigen and LPL activity by determining their respective size in mouse adipose cells. Two questions were posed ( i ) are some inactive LPL molecules present within the cells? ( i i ) do changes in the catalytic activity of the enzyme occur upon secretion?
The immunoprecipitation assay has allowed us to determine the actual amounts of LPL involved in the constitutive and the heparin-stimulated secretions exhibited by 3T3-F442A and Ob17SA16 cells. At the steady state, e.g. in the absence of heparin, only part of the protein pool is exported through constitutive secretion. Remarkably, when heparin-stimulated secretion occurred, the entire pool of preexisting molecules was released into the extracellular medium. The fact that this release remained possible even in the absence of protein synthesis ( p = 1, see Miniprint, Table I) is not surprising but strengthens the idea that, under appropriate stimulation by a secretagogue, adipocytes become able to secrete their entire LPL content in viuo. Moreover, no change in the specific activity of LPL was observed whether or not the enzyme secretion was constitutive or heparin-stimulated (Fig. 5). In that respect, immunoprecipitation data (see Miniprint, Table  I) have been confirmed by the dot-blot immunoassay of LPL (Fig. 5).
The most striking observation, with regard to intracellular LPL activity, remains the discrepancy between the values of the secretion potential p measured by the enzymatic assay and the immunoprecipitation assay. When intracellular activity was assayed using cell lysates containing LPL at a concentration near or equal to that found in the secretion media where the enzyme amount is accurately determined, the secretion potential appeared to be largely over-estimated ( p > 1). Until now such an observation, e.g. a secretion-coupled increase in enzymatic activity in the absence of protein synthesis, has represented the main argument in favor of the existence of an inactive precursor of LPL within the cells (10). Among hypotheses, the inactive form of LPL could be represented by that transiently present in the endoplasmic reticulum of Ob17 cells (20). Alternatively, the inactive forms of LPL described in 3T3-Ll cells by Olivecrona et al. (34) could also be candidates. However, because of the protocol used by these latter authors to extract and concentrate LPL from acetone-ether powders, it is not possible to decide whether these inactive molecules actually represented an inactive precursor form of LPL. Our results presented above and elsewhere (23) show that, in the presence of heparin and in the absence of protein synthesis, no enzyme degradation takes place; they also show that the enzyme activity being released is identical to that determined within the cells providing that the cell lysates are extensively diluted (Figs. 3 and 4).
Our findings indicate also that the maximal increase in LPL activity requires the synergistic effects of dilution and heparin addition. The structure in which potentially active LPL molecules are present remains unclear. However, two types of interactions could be responsible for the observed cryptic character of the intracellular LPL activity. We envision that in situ, a first interaction would take place between LPL and some intracellular molecules, giving rise to a specific complex (first level of interaction), and that a second interaction would take place between these complexes, giving rise to a network (second level of interaction). If this were so, one could postulate that ( i ) by mere dilution which lowers the concentration of the network components, the second interaction would be disrupted, leading in turn to a partial access of LPL molecules to the substrate; (ii) the first interaction would then be disrupted by displacing LPL from its binding site with either heparin or heat-inactivated LPL.
Some observations are in favor of the hypothesis of LPL being bound to intracellular proteoglycan structures. First, although proteoglycans in mammalian tissues are known to be components of the extracellular matrix, it has been demonstrated that heparan sulfate proteoglycans and sulfated glycosaminoglycans are part of the internal content of secretory organelles (41), including the chromaffin granule (42), cholinergic synaptic vesicles (43), the zymogen granule (44), the prolactin granule (45), and the adrenocorticotropin secretory granule (46). The fact that proteoglycans represent potential binding sites for LPL is well illustrated by the recent work of Klinger et al. (47) who used affinity chromatography on immobilized LPL to purify the heparan sulfate proteoglycans from rat brain. Second, some secretory proteins such as adrenocorticotropin (48), and in a more extreme form represented by the crystalline state of insulin in /3 granules (49) are stored intracellularly in a condensed state. It remains to be shown that granules containing LPL and proteoglycans are indeed present in adipose cells. If it were so, the condensed state of LPL within the cells would be compatible with the existence of a concentrative sorting process in the regulated secretory pathway (50, 51).
Another consequence of our findings is that the significance of intracellular LPL activity now becomes questionable as this is probably minimal or absent in intact cells. For instance, our previous studies had shown that the catalytically active conformation of LPL was acquired at the time of its emergence in the cis-and/or medial-Golgi cisternae (20); this activity can now be regarded as a mere indicator of the functional maturation of LPL. In other words, apart from the crypticity of LPL within the subcellular compartments which can be unmasked by detergent solubilization (22), an additional level of crypticity appears to be due to the presence of LPL within a network. Thus the intracellular LPL activity so far described is an unavoidable reflection of the true potential value. In uiuo, this phenomenon of masking the activity of LPL is advantageous for the cell owing to the phospholipase AI activity of the enzyme on glycerophospholipids.
In conclusion, the proposal regarding the existence of a putative reservoir of inactive LPL molecules appears to be related to a systematic underestimation of its intracellular activity. In that respect, because active LPL is a homodimer (33), the monomer cannot be regarded as an inactive precursor form as, in vitro, activation by dimerization would not be favored by lowering the protein concentration. heparin. either in differentiation medium containing 0.1 or 7.5% fetal bovine serum or in I n -Secretion was measured in the absence or m the presence of medium as indicated. Two protocols were used as already described (IS). The fin1 protocol (batchwise secretion) was used for the preparation of LPL samples intended for affinity chromatography on Heparin-Ultrogel or quantitative immunoprecipitation. Cells were washed with DMEM at 37'C and funher incubated in the medium used for secretion for the indicated periods.
This medium was then wllected and rapidly chilled at 0°C. When it was intended for LPL assays, the secretion medium was diluted IS-fold with 5 mM sodtum barbital pH 7.4. 1 M glycerol (buffer A) to stabilize the enzymatic actlvity and maintained at O' C or stored at -70°C until use. Under these conditions, the final volume of the diluted secretion medium was 100-150 pl per cm2 of cell culture.  = 59.000). Iodination of this purified LPL was performed by the Iactoperoxidax/glucose oxidase procedure according IO Wallinder d. (27). The specific radioactivity of the trichloroacetic acid-precipitable material was 940 cpmlng and 97% of the radioactivity was recovered undcr he Mr = 59,000 species.
-dies aeainst mouse LPL from secretion media of 3T3-F442A cells. A solution of the protein (I50 pg in 150 p1) was -Polyclonal antibiles were raised in a rabbit against LPL purified emulsified with an equal volume of Freunds complete adjuvant. One half of the mixlure was injected into the popliteal lymph nodes (28) and the other half was injected intradermally at four different sites in the back. Three weeks later, 125 pg in incomplete adjuvant were injected in the armpits and the groins. Two weeks later, the same emulsion was injected into the right subscapular cavil (25 pg LPL) and subcutaneously at four various sites in the back (IOOpg). Two weeks later. an intramuscular injectton in the posterior legs was given using 100 pg of protein in PBS. After 8 days, injection. It was then bled 5 days after a last intramuscular inJection of 25 pg of protein. The 100 pg were injected intravenously. The rabbit was bled weekly for 6 weeks without booster antiserum. filtered through 0.45 prim-pore m e filten, was stored at -70°C.

m T B q / m o l ) . Ihe condmons for the enzymatic assay have been prevlously
-LPL activity is taken as the apo C-11-dependent hydrolysis of glycerol described (20) -cion of matwe LPL onto n>trocellolose m a k w x a NaCl at concentrations ranging from 0.15 to 2 M as well 1s LPL secreted by the cclls in ITT medium -Purified LPL in buffer A contammg (see above) bmd meversibly to nmocellulose. Taking advantage of this property, concentrative adsorpnon of native LPL from very dilute solutions could be achieved by filtrat~on on membranes composed of mixed cellulose acetate and nitrate (type HATF. Mlllipore Corporation). Elution of LPL did not occur ac pH values between 2.0 and 9.0. However. quantitative recovery of the protein was obtained by heating the filters at 96°C for 5 min in the presence of 180 mM Tris-HCI buffer pH 6 8 containing 2% SDS (dcnaturatlng buffer used for SDS-PAGE).
Prevaration of cell Ivsate5 -The procedure to obtain cell lysates in Triton X-I 14 suttable for LPL assays has already been descnbed (20) ; it was used with slight modifications. Briefly. washed cells were solubilized at 0-2°C in buffer A containing l5OmM NaCI. 0.4% (wlv) Triton X-I14 (5Opl per cm2 of cell culture ) and a cocktail of protease inhibitors (aprotinin 4 pg/ml, pepstaun 4 pglml,

PMSF ~x I O -~M ) .
After heat treatment at 30°C to pellet Triton X-I14 (29). the detergent-depleted lysate was adjusted by a 2-fold dilution with the same buffer excluding Triton X-I 14 to a flnal volume of 1 0 0 pI per cm2 of cell cullure, as already described for secretion media. The lysates could be stored at -20°C before asssaying LPL activity.
Membranes of the post-nuclear supernatant containing more than 95% of latent LPL activity The 2-fold dtlution of cell lysates or post-nuclear membrane lysates will be referred to as the were obtained as already descnbed (22). For LPL assays, they were solubilized and diluted as above.
standard dilution of cellular LPL.
protocol described below.
The cell lysates used in quantitative immunopreclpitation were prepared according to the . .
filtratlon unit (Bio-dot microfiltration unit. BioRad Laboratones) where the rubber sealing gasket --A s s a y s were performed using a %-well formaued vacuum was replaced by a sheet of Whaunan 3MM paper. Before use. lhe nimcellulose and paper sheets were soaked ~n PBS. then mounted in the apparatus. LPL samples (200 pI) were applied to individual wells and allowed IO stand for 20 min at room temperature. A reduced pressure was applied to pull the liquld through the membrane withm 2 min and then to wash the wells four tunes with 400 VI of buffer A containing 1.5 M NaCl. The apparatus was daassembled and the nitrocelluiosc sheet was washed in PBS (2 min), saturated by a 0.5% (w/v) gelatm solution in PBS (Zh, 37'C). and washed in 0.25% (wlv) gelatin in PBS. The nitrocellulose was then treated either with anti-LPL antiserum (1,500-fold dilution) or with normal rabbit semm (same dilution) in Tween buffer (PBS Containing 0.25% (w/v) gelatin and 0.25% (wh.) Tween 20). for 16h at 4' C (200 pVcm2 of membrane). It was then washed six times (IO min each in Tween buffer at roam temperature). The nitrocellulose sheet was then treated for 2h with horseradish peroxidase-conjugated goat anti-rabbit I@ diluted 1 . W fold in Tween buffer. After washing SIX times as above. it was incubated In PBS for IO min at room temperature and 5 min at 37°C. Peroxidase was revealed, as described by Graham and Kamovsky (30). The wet membrane was scanned by reflectance densitometry using an integrating LKB laser microdensitometer.

--
T h e amount of anti-LPL antiserum to be added to secretion media or cell lysates to ensure LPL isolation was determmed in preliminary experiments carried out at 0-4T as follows, Secretion media, stabilized as above and containing [3Hlleucine-labelled LPL. were first assayed for enzymatic acnvity. Then media with activities ranging from IO to 80 mU/ml were diluted 1.3-fold with 80 mM NazHP04 pH 7.3 containing 150 mM NaCI. 20 mM EDTA. 4% w/v) Triton X-100, 1 M glycerol and the cocktail of protease inhibitors (see above). They were then that antiserum at a final 100-fold dilution was sufficient to ensure the precipitation of labelled LPL incubated (250 p1-I ml) for 16h at 4°C with increasing amounts of anti-LPL antiserum. It was found (40,mUlml) when an optimal ratio of 3 mg (lyophilized beads) of Protein A-Sepharose per ml antxserum was used. The titer of the antibodies. expressed as precipilable activity per pI antiserum was thus 4 mU/pl. Routinely. the antnemm was used at a 50to 100-fold dilution and was adjusted to give a 3-fold excess (3 pU4 mu) over the antigen. Since the ratio of volume IO cell culture area was kept constant for cell lysates and secretm media. both samples were processed in he same way.
After labelling, the cells were washed twice with ice-cold PBS and lysed at O°C for 30 min m buffer A (20 mM Na2HP04 pH 7.3, 150 mM NaCI. 1 M glycerol) containing 5 mM EDTA and 1% (wlv) Triton X-I00 ; 0.1 mM PMSF. 4 pg/ml pepslatin and 4 pg/ml aprotinin were added to buffer A. Insoluble material was pelleted by centrifugation at 12.000 g for I5 min. The supmatant was used after a 2-fold dilutton in buffer A containing 5 mM EDTA. The cell lysates and secretion media (1-2 mU6O-mm dish) were immunoprecipitated using the annserum at the dilution indicated above. Incubations were performedfor 15h at 4°C. A 1:l rluny of protein A.Sepharose in buffer A containing 0.5% (w/v) Triton X-100 and 2 mM EDTA (total volume. 250 pll was added and allowed to react for 2.5h at 4 T with cnd-over-end rotation of the lube. The beads were then washed three times in buffer A containing 0.5% (wlv) Tnton X-I00 and 2 mM EDTA, twice with PBS and once w~t h IO mM Tris-HCI pH 7.2. The immune complexes were dissociated by heating for 6 min at 9 6 T in 180 mM Tris-HCI pH 6.8, 2% (w/v) SDS 10% (w/v) glycerol and 0.003% (w/v) bromophenol blue in the absence of a reducing agent. The beads were then removed by two centrifugations (5 min, 2000 g) and an aliquot of the supemawt was used for SDS-PAGE according to Laemmlt (31). Fluorography of the gels was performed after impregmtion with salicylic acid (32). Fluorographs were scanned with an integrating LKB laser densitometer. When munoprecipitatnon was used for the direct determination of incorporated radioactivity. 0.2% (Wh. 1 gelatin was included in the two PBS washes and the last wash in Tris buffer was omitted.

RESULTS
bcificitv of anti-LPL antbody -The protocol described under "Expenmental Procedures" allowed purification of LPL from secretion mdia in a single step. It was used for the large scale purification of LPL. The immune response of rabbits injectcd with pure LPL was checked uslng final dilutions of xrum samples ranging from 400to 8.000-fold. The selected antibody was able to inhibit the apo C-11-dependent actavity of LPL but had no effect on its basal triacylglycerol hydrolase activity. Thls property was also noted in the case of he inhibitory effect of the anubody on LPL secreted by rat adipose cells (not shown). This polyclonal antibody could also bc used for immunoprecipitation experiments (Fig 1). A typical precipitatton curve of [3Hlleucine-labeUcd LPL from a secretion medium of 3T3-F442A cells, as a function of antiserum concentration. is shown in Figure la. The concentration of IgG needed to give a half-maximal precipitation was estunated to be was required to favor the antigen-antibody recognition and to obtain quantitative precipitatton. The less than 2 nM. It is worlh noting that. in contrast with another protocol (34). no denaturation step finding that the antibody bound strongly to native LPL suggests that a sub-population of IgG interacted with conformational determinants of the molecule. The specific immunolsolatton of  At this poinl. i t IS ncccssary to dcfinc the sccrction potential p . taken as the t h e ratio of total relearahlc activity or anttgen to initial cellular activlty or anttgen.dclermmed In cell% exposed simultnncourly to hcparin and cyclohcninlidc. Kesults of Tahle   dpm1dish. x 10-3 (70) pot-hlot immunoassav of SCCIcharactcristics of the assay. illustrate the ability of anti-LPL antiserum to quantitatively assay the -The data of Figure2. which summarize the antigen. Figure 2a shows a dot-blot obtained with serial dilutions of purified mouse and bovim milk LPL. No or little peroxidase staining was ohserved when non-immunc rabbu serum was used (Fig.2a. lane C). Ry cutting out and counting the dots contaimng radioactive antigens. il was first ascenained that the amount of adsorkd matenal was a linear function of that used for dotting onto the membrane the membrane was 92 f 4% when compared to that of the lnltlally soluhle antigen. over the 0.1 to (hg.2b. left). In the case of mousc LPL, the recovery of the I'tllleucine radioactwily associated wlth 3.6 mU range studied. 7hc comspondtng figure for bovme LPL. whlch was dhted such that the response was within the range of the standard curve.
was 94 t 5%. The mtcnsitlcs of the corresponding peroxidase signals given by the two antigens are shown in Frgure 2b (right). The ahsorbancc remained a IInear function (up to 0.7 optical density units) of the antigen concentration providing that the latter was telow 0.9 and 0.7 milliunits per 200 pl of initial sample for mouse and bovine LPL. respectively. The standard tmation curve of purified mouse LPL rcmaimd unchanged when the antigen was dotted after dilution in ITT medium instead of buffer A contaming IS0 m M NaCI.
The sensitivity and the specificity of the assay are established by the data of Figure 2c and c. the absorbance is ploned as optical density units.