Solubilization and Partial Purification of an Enzyme Involved in Rat Liver Microsomal Fatty Acid Chain Elongation: /?-Hydroxyacyl-CoA Dehydrase*

The solubilization and partial purification of fi-hy- droxyacyl-CoA dehydrase from rat liver microsomes has been accomplished through deoxycholate solubilization, ammonium sulfate fractionation, and ion ex- change chromatography. A purification of about 90- fold based on total soluble activity was achieved, with an overall yield of 40%. However, the initial solubiliza- tion is accompanied by the loss of the secondary portion of the U/S curve observed with intact microsomes. The enzyme requires detergent during the purifica- tion procedure to remain “soluble,” and is strongly activated by the inclusion of Triton X-100 at concentra- tions above its critical micelle concentration in the assay mixture. In addition a preference for micelles has been inferred based on discontinuities in the V/S curves relative to the measured critical micelle concentration of the substrates in the absence of Triton X-100. Kinetic parameters calculated on the basis of micelle-specific activity indicated that P-hydroxyacyl-CoA substrates possessing even-numbered alkyl chains from 14 to 20 carbon atoms differed little in V,,, but had progres- sively larger K,, as the chain


The solubilization
and partial purification of fi-hydroxyacyl-CoA dehydrase from rat liver microsomes has been accomplished through deoxycholate solubilization, ammonium sulfate fractionation, and ion exchange chromatography.
A purification of about 90fold based on total soluble activity was achieved, with an overall yield of 40%. However, the initial solubilization is accompanied by the loss of the secondary portion of the U/S curve observed with intact microsomes. The enzyme requires detergent during the purification procedure to remain "soluble," and is strongly activated by the inclusion of Triton X-100 at concentrations above its critical micelle concentration in the assay mixture.
In addition a preference for micelles has been inferred based on discontinuities in the V/S curves relative to the measured critical micelle concentration of the substrates in the absence of Triton X-100. Kinetic parameters calculated on the basis of micelle-specific activity indicated that P-hydroxyacyl-CoA substrates possessing even-numbered alkyl chains from 14 to 20 carbon atoms differed little in V,,, but had progressively larger K,, as the chain length increased. The partially purified preparation was also active with flhydroxy-8,11-eicosadienoyl-CoA; and with 2-trans-enoyl-CoA substrates in a reverse (hydration) reaction.
The membrane association of enzymes involved in fatty acid chain elongation imparts a number of important advantages to the cell both in view of the lipid nature of the metabolites involved and the potential juxtaposition of enzymes in the reaction sequence (1). However, progress in elucidating the nature of chain elongation is limited by the particulate and complex nature of microsomal preparations. Although considerable progress has been achieved in the past few years in the solubilization and purification of the fatty acid desaturases (2,3), the microsomal fatty acid chain elongation enzyme(s) have so far resisted such efforts (4-6). Thus relatively little is known of the organization, lipid requirements, and specificity of the enzymes involved in this process. Since the overall chain elongation process has proven to be intractable to solubilization attempts, studies of partial reaction steps followed ultimately by reconstitution may prove to be more fruitful.  ative (7, 8), which is in turn reduced to the @saturated product by microsomal enoyl-CoA reductase (9). The enzyme catalyzing the dehydration reaction behaves somewhat differently than do others in the sequence in that it is not inhibited by high substrate concentrations, and it is relatively stable in the presence of bile salt detergents (8). These properties have led us to initiate chain elongation purification studies through an examination of this enzyme. This paper reports the solubilization and partial purification of ,&hydroxyacyl-CoA dehydrase from rat liver microsomes. The partially purified preparation has been found to act on a variety of long chain saturated /I-hydroxyacyl-CoA substrates, as well as representative polyunsaturated and 2-trans derivatives. Evidence suggesting an integral protein nature (10) for the enzyme is also presented. After mincing the preparation, it was gently homogenized with a loose-fitting Teflon pestle in a Potter-Elvehjem vessel using 10 to 12 strokes. The homogenate was centrifuged at 500 x g for 10 min, after which the supernatant was centrifuged at 17,300 x g for 15 min. The second supernatant was then centrifuged at 100,000 x g for 1 h. The drained microsomal pellet was washed free of the clear glycogen pellet with sucrose buffer, gently homogenized, and resedimented at 100,000 x g for 50 min. The washed microsomal pellets were resuspended in sucrose buffer at 2.5 mg of protein ml-' and lyophilized.
Dehydrase Preparation-Lyophilized microsomes were solubilized by incubating 10 mg of protein/ml with occasional mixing at 0-4°C in a medium consisting of 0.5% sodium deoxycholate, 0.5 M KCl, 0.1 M sucrose, and 4 mM potassium phosphate buffer, pH 7.4. After 10 min, the preparations were centrifuged at 100,000 x g for 1 h. The clear supernatant was made 0.5% in sodium cholate, then 20% saturated in (NH,),SO, by the gradual addition of the solid salt. After incubating for 15 min at 4'C, the suspension was centrifuged at 17,300 x g for 15 min and the pellet discarded.
The supernatant was then made 33% saturated in (NHI)lSOd. After incubation and centrifugation as before, the supernatant was discarded and the pellet taken up in a minimum volume of 0.02 M Tris. Cl buffer, pH 7.8, containing 2% Triton X-100. This material was applied to a Sephadex G-50 column (about 2.7 x 9.0 cm) and eluted with the same detergent-buffer mixture.
The void volume was collected and clarified by centrifugation at 100,000 x g for 1 h. Most of the dehydrase activity could be recovered in the resulting supernatant, although some activity was invariably lost in the pellet.
The supernatant was applied to a DE-52 column (1.6 X 15 cm) and the column developed with the same buffer. Under these conditions, most of the protein was adsorbed by the column while the dehydrase was not retained and could be recovered in an early peak corresponding to the void volume. The active fractions were pooled and adjusted to pH 6.00 by the dropwise addition of 0.2 M HCl, then applied to a CM-52 column (1.6 x 28 cm) which had been equilibrated with 0.01 M potassium phosphate buffer, pH 6.00, containing 0.3% Triton X-100. After washing the column with application buffer, a linear gradient was initiated consisting of equal volumes of application and limit buffer (0.20 M potassium phosphate, pH 6.00, 0.3% Triton X-100), and the column developed at a flow rate of about 9 ml h-'. The effluent was assayed for dehydrase activity and the active fractions were pooled and concentrated by ultrafiltration using a PM-10 membrane. This preparation was stable for several weeks when stored at o-4°C.

Solubilization of Enzymes Involved in Partial Reactions of Fatty Acid Chain Elongation-We
have previously reported that the ,8-hydroxyacyl-CoA dehydrase involved in malonyl-CoA-dependent fatty acid chain elongation may be solubilized by treatment of rat liver microsomes with a sodium deoxycholate-KC1 mixture (8). However, exposure of microsomes to deoxycholate is also accompanied by a loss in the secondary rise in activity of this enzyme (Fig. 1) which has been noted to occur at high substrate concentrations with intact microsomes (8, 19). As indicated in Fig. 1, this conversion of the U/S curve to one displaying relatively normal hyperbolic saturation kinetics could be demonstrated with both the 100,000 x g supernatant, and total (uncentrifuged) microsomes in the presence of deoxycholate.
Attempts to restore the biphasic u/s behavior through removal of the detergent on a Sephadex G-50 column at pH 7.8 followed by repelleting of the enzyme were unsuccessful, although essentially all of the detergent was removed by this procedure as monitored by the inclusion of [%]deoxycholate in the initial solubilization medium. Most of the dehydrase activity was recovered in the 100,000 X g pellet at this stage, which yielded a u/s curve similar to that of the 100,000 X g supernatant.
Of the chain elongation enzymes investigated, only the dehydrase could be recovered in good yield under the conditions employed (Table I). Both palmitoyl-CoA condensation and 2-trans-octadecenoyl-CoA reductase specific activities were greatly reduced following treatment with the solubilization medium, and these activities were only slightly aug- as substrate with intact microsomes (0) or following extraction of microsomes with deoxycholate. After a lo-min preincubation with the detergent, aliquots were assayed immediately (0). The preparation was then centrifuged at 100,000 x g for 1 h and supernatant (D) and pellet (A) assayed. All assays were conducted with 0.625 mg of protein. The data represent nanomoles of product formed min-' (mg of protein)-' and are corrected for blank (zero time) conversions. The number in parentheses is the percentage of activity in intact microsomes.
" Represents activity recovered in the 100,000 x g pellet obtained following gel filtration and ultracentrifugation of the original detergent-solubilized supernatant. a All assays were carried out with ,f?-hydroxy-18:OCoA using the spectrophotometric method and standard conditions as described under "Experimental Procedures." ' Although some of the activity was lost in the pellet at this stage, most of the activity lost was sacrificed during the Sephadex G-50 chromatography step due to tailing. mented in the pellet recovered after removal of deoxycholate by gel filtration and a second ultracentrifugation (column 3 of Table I).

Purification of Dehydrase
Actiuity-Deoxycholate-solubilized dehydrase activity was partially purified by ammonium sulfate fractionation followed by DEAE-and CM-cellulose chromatography as summarized for a typical preparation in Table II. The overall purification varied somewhat among preparations, but was in the range of 80-to loo-fold based on the total soluble activity. Recovery of activity following DEAE-chromatography was generally 65 to 70% of that applied to the column. Attempts to recover the remaining activity with a buffer concentration gradient, followed by washing the column with limit buffer which was 1 M in KC1 did not yield any additional active fractions. Thus a portion of the applied activity is apparently either very tightly bound, or inactivated during the course of development on the column. When the dehydrase pool from the DEAE-step was passed through a carboxymethyl-cellulose column, excess Triton X-100 washed through in the void volume after which two protein peaks were recovered in the early part of the chromatogram.
As indicated in Fig. 2, dehydrase activity was associated with the first of these peaks and was recovered in a broad, nonsymmetrical band. The enzyme recovered at this stage is either a very large protein, or more likely exists as an aggregate even in the presence of Triton X-100, since it is excluded from a Sephadex G-200 column. When the preparation was analyzed on a Sepharose GB-CL column (1.6 x 36 cm) equilibrated with 0.05 M Tris.Cl, pH 7.40, 0.05 M NaCl, and 0.3% Triton X-100, the enzyme was included and recovered in a single peak with a K,, = 0.4. However, no further increase in specific activity was achieved.
Effect of Enzyme Concentration-Initial attempts to determine the protein dependence for the dehydrase reaction resulted in non-linear curves which were concave to the ordinate. This response was found to reflect a requirement for Triton X-100 in the assay medium which was partially provided by the detergent present in the stock enzyme preparation. When optimum detergent concentrations were included in the assay medium, a linear response up to at least 2.4 pg of protein was obtained (Fig. 3).
Activation of Dehydrase by Detergent-The effect of Triton X-100 on dehydrase activity is indicated in Fig. 4. The addition of up to 100 pg of the detergent to the amount already present in the assay medium from the enzyme aliquot (160 pg) was without effect on the rate of dehydration.
However, additional Triton led to a rapid rise in rate until a plateau was achieved when 400 pg (560 pg total) of detergent were included.
The CMC of Triton X-100 in aqueous solution is about 0.24 mM (20-22). The arrow in Fig. 4  The column was developed as described under "Experimental Procedures" and fractions assayed for dehydrase activity by the standard spectrophotometric procedure. and protein by the method of Lowry et al. (17) (310 pg), assuming a mean molecular weight of 645 for detergent monomers (20, 23). Thus, the results of Fig. 4 suggest that the reaction is augmented by detergent micelles, with the monomers having little if any effect. This interpretation is supported by the results of analysis conducted under the same conditions but in the presence of 0.25 M KC1 (not shown), in which the rise in dehydration rate occurred at lower Triton concentrations, consistent with a reduced CMC'. Although the effect is more pronounced with ionic detergents, the CMC of non-ionic detergents may also decrease with increasing salt concentration (20). Effect of pH- Fig.  5 shows the effect of pH on the initial velocity of the dehydrase reaction. The enzyme was found to be active over a broad range of pH values from 5.5 to 11.0. In addition, three optima were observed at 6.5, 8.5, and 10. Although the variation in activity between pH 6 and 10.5 was not large, the pattern indicated in Fig. 5 was observed in two separate preparations.
It is possible that three enzymes were represented in these preparations, but if so, they appear to be tightly associated since the relative activities at each maximum were nearly identical in the two preparations examined. Only one band of activity could be detected in all column chromatograms whether the assays were conducted at pH 6.5 or 7.8.
Substrate Concentration Dependences--u/s curves for the dehydrase reaction are presented in Fig. 6. With the three longer chain substrates, the curves approached saturation at about the same substrate concentration, and a progressive increase in rate as the chain length decreased could be noted. Although the curves with the 18 and 20 carbon substrates approximated normal rectangular hyperbolae, the curve obtained with the 16 carbon substrate displayed a slight discontinuity at substrate concentrations less than 10 to 15 pM. While both ,&hydroxy-12:0-CoA and /3-hydroxy-14:0-CoA were active at higher substrate concentrations, marked deviation from expected behavior was obtained at low substrate levels.  low at 15 pM, then rose rapidly at substrate concentrations above 20 pM; while no activity could be detected with the 12 carbon substrate at concentrations below 25 to 30 pM. Long chain acyl-CoA thioesters are effective detergents which are known to form micelles at relatively low concentrations in aqueous solutions (24,25). Since incubations always contained Triton X-100 above its CMC it might be predicted that mixed Triton X-lOO-acyl-CoA micelles would be present at any concentration of substrate. The observed negligible activity with low concentrations of substrates was thus unexpected. The u/s data in Fig. 6 might be explained if the dehydrase acts only on substrate micelles above their own CMC. Alternatively Triton X-100 might not form mixed micelles with the substrate until the CMC of the acyl-CoA derivatives is approached.
In either case there should be a rapid rise in specific activity when the concentration of the active substrate equals its CMC. The results in Fig. 6 would be consistent with a progressively increasing CMC of the active substrate micellar complex as the substrates alkyl chain was shortened which is a characteristic of CMC values for such a series (20, 26). The active substrate micelle with the 18 and 20 carbon substrates might be expected to possess CMC too low to detect in this fashion since velocity determinations at very low substrate concentrations were not reliable in these assays.
The CMC of a variety of saturated and unsaturated acyl-CoA thioesters have been reported by Cleland et al. (24,25). However, no data exist for the P-hydroxy derivatives which might be expected to yield somewhat higher values due to the presence of hydroxyl function in the chain. Also, since the observed CMC may be influenced by the conditions of analysis including pH temperature and ionic strength it was important to determine the behavior of these substrates under the same conditions employed in the enzymatic analyses. Thus CMC determinations were conducted in 0.1 M potassium phosphate buffer, pH 6.50, at room temperature, using the dye binding technique of Zahler et al. (24). Since the presence of pinacyanol chloride in the medium tends to promote micelle formation, the determination employed either 1 cm or 10 cm path cells with a dye concentration of 5.0 or 0.5 pM, respectively, dependent on the substrates expected CMC', in order to maintain a low mole fraction of dye in the region of the measured CMC (24).
Values obtained under these conditions with the five substrates are reported in Table III, and represent the intersection point of lines drawn through the linear regions of the plots (24) at acyl-CoA concentrations well below and above the CMC'. In all cases, the data were fitted to regression lines which were solved simultaneously to obtain the substrate concentration reported as the apparent CMC. It should be noted that the relatively high ionic strength used in these analyses might be expected to promote micelle formation (20). This could be demonstrated with palmitoyl-CoA which yielded a CMC' of approximately 2.6 pM in this system, while a value of 3.2 PM in agreement with the report of Zahler et al. (24) was obtained in 6.7 InM potassium phosphate buffer, pH 6.9. Comparison of the data in Table III with Fig. 6 suggests that a reasonable correlation exists between observed CMC regions and U/S curve discontinuities.
The relationship is most apparent with the two shorter chain substrates which possess CMC' high enough to fall within the accessible substrate concentration range on the u/s curves. A number of investigators (27)(28)(29)(30) have attempted to provide a theoretical basis for kinetic analysis of reactions involving lipid substrates. Gatt and Bartfai (27) note that the kinetic parameters of enzymatic reactions involving "soluble" amphiphilic lipids such as long chain acyl-CoAs in their micellar form may be estimated by plotting the reciprocal of the concentration of substrate in micellar form (St -S,,,.) uersus the reciprocal of observed velocity minus the reaction rate at S = CMC' (V -V,,,). Fig.  7 represents the v/s data replotted on this basis and assuming micelle-specific activity (i.e. V,,, = 0) in each case. It must be noted that these plots assume that the CMC of the active The assays were conducted by the dye binding method of Zahler et al. (24) as described in the text. of Triton X-100 is similar to that measured in the absence of this detergent.
Kinetic parameters estimated from the data of Fig. 7 are summarized in Table IV. These data suggest that V,, is essentially constant for substrates of chain length 14 to 18, and declines only slightly with the 20 carbon derivative while Km, appears to decline progressively as the chain length decreases from 20 to 14 carbon atoms. ,&Hydroxy-12:0-CoA is atypical in all cases, and as is evident in Fig. 7 it still appeared to deviate from normal behavior when replotted in this manner. The reason for this behavior is unclear. It might also be noted that the apparent K,,, values estimated in this fashion are based on total substrate concentration in micellar form, rather than on micellar concentration which is dependent on Kinetics of enoyl-CoA hydration conducted under the same conditions as described in

Fraction
Relative specific activity" Total soluble 1.26 G-50/100,000 x g supernatant 1.25 DEAE-cellulose 1.10 CM-cellulose 1.21 n Ratio of specific activity with ,&hydroxy-8,11-20:2-CoA versus that obtained with P-hydroxy-18:0-CoA as determined by the standard assay procedure. the aggregation number. Thus differences in substrate aggregation number could have influenced these results.
As expected, the partially purified dehydrase preparation was also active in the reverse direction when either 2-trans-l&l-CoA or 2-tram-20:1-CoA were employed as substrate (Fig. 8). No atypical behavior was observed in these curves; however, the CMC' of these substrates would be expected to be at least as low as those of the corresponding chain length ,&hydroxy derivatives and thus unlikely to influence the u/s response. The enzyme was also active with a representative polyunsaturated substrate (P-hydroxy-&ll-eicosadienoyl-CoA). As indicated in Table V, the ratio of specific activity with this substrate relative to P-hydroxy-18:0-CoA remained essentially constant during purification, consistent with a common enzyme active with both saturated and polyunsaturated substrates as has previously been suggested for this reaction (8).

DISCUSSION
While deoxycholate treatment of microsomes was effective in solubilizing the P-hydroxyacyl-CoA dehydrase, neither condensation nor enoyl-CoA reductase activities could be solubilized in acceptable yields by this approach. The latter two enzymes also responded differently following the removal of detergent, and in neither case could original activity levels be restored. However, these results do provide support for previous suggestions that separate enzymes are involved in the partial reaction steps of chain elongation (8,31), and suggest that the sensitivity to substrate inhibition of both condensation and enoyl-CoA reductase reactions (8) may reflect the well-known detergency characteristics of acyl-CoA thioesters (24,25). By contrast, the dehydrase does not display substrate inhibition (8) consistent with its resistance to detergent inactivation.
It is interesting that while the fist and last steps of chain elongation are subject to substrate micellar inhibition and are facilitated by the presence of an acyl-CoA binding protein such as bovine serum albumin in the medium (8, 32), the intervening dehydrase reaction is relatively insensitive to excess substrate or the presence of albumin. The possibility exists that intermediate chain elongation reaction steps may utilize acyl-CoA substrates diffusing within the plane of the membrane, with only the initial and final steps involved in an exchange with "soluble" acyl-CoA in the cytoplasm. The potential importance of the incorporation of acyl-CoA into membranes, and their diffusion within the bilayer relative to model lipid synthetic enzyme systems has been discussed (33,34). This interpretation suggests that ,&ketoacyl-CoA reductase might be expected to respond in a manner similar to the dehydrase. Unfortunately, difficulties attendant to the assay of the former reaction (8) have so far precluded an analysis of its characteristics.
Although "solubilization" of dehydrase activity through the use of detergents was achieved as defined both by centrifugal and gel filtration criteria (1,35), a continuing requirement for detergent during purification suggests that a rather hydrophobic species was involved. This result, in conjunction with the close relationship between the percentage of activity solubilized and the detergent concentration in the solubilization medium, is consistent with an enzyme which exists at least partly embedded in the membrane; corresponding to an "integral" protein according to the classification of Singer and Nicolson (10). Whether a lipid requirement exists for this enzyme is unknown at this time. No exogenous lipid was required in the assays, but it is possible that some tightly associated lipid was carried throughout the purification procedures (20,36) since only moderate detergent concentrations were used. It is also possible that the Triton X-100 effect related at least in part to such a requirement since detergents may substitute for bound lipid in many instances (20).
In addition to an activation by detergent micelles, the partially purified dehydrase also appeared to require micelles for activity. The significance of this preference remains unclear. During the course of chain elongation, the major regulatory step is that of condensation (8). If multiple condensation enzymes contribute their respective products to a common subsequent pathway (8) a micellar preference at the dehydrase step could serve a secondary regulatory role, and might favor the utilization of longer chain substrates (possessing lower CMC) at low substrate concentrations.
An alternative explanation could involve a micellar requirement for substrate incorporation into the detergent. enzyme complex. It is, of course, possible that the apparent substrate micellar preference observed in this system is an allotopic response without physiological significance. However, a number of lipid enzyme systems active on substrate micelles have been described (25,29,(37)(38)(39).