Download Rat sn-glycerol-3-phosphate acyltransferase - UNC

Biochimica et Biophysica Acta 1439 (1999) 415^423
www.elsevier.com/locate/bba
Short sequence-paper
Rat sn-glycerol-3-phosphate acyltransferase: molecular cloning and
characterization of the cDNA and expressed protein
B. Ganesh Bhat 1 , Ping Wang, Ji-Hyeon Kim, Tracy M. Black, Tal M. Lewin,
Frederick T. Fiedorek Jr. 2 , Rosalind A. Coleman *
Departments of Nutrition, Pediatrics, and Medicine, The University of North Carolina, Chapel Hill NC 27599-7400, USA
Received 29 March 1999; received in revised form 21 May 1999; accepted 2 June 1999
Abstract
Rat mitochondrial glycerol-3-phosphate acyltransferase (GPAT) cDNA was cloned and characterized. We identified a
cDNA containing an open reading frame of 828 amino acids that had an 89% homology with the coding region of the
previously characterized mouse mitochondrial GPAT and a predicted amino acid sequence that was 96% identical. The rat 5P
UTR was only 159 nucleotides, in contrast to the 926 nucleotide 5P UTR of the mouse cDNA and had an internal deletion of
167 nucleotides. GPAT was expressed in Sf21 insect cells, and specific inhibitors strongly suggest that, like the Escherichia
coli GPAT, the recombinant mitochondrial GPAT and the mitochondrial GPAT isoform in rat liver contain critical serine,
histidine, and arginine residues. ß 1999 Elsevier Science B.V. All rights reserved.
Keywords: Glycerol-3-phosphate acyltransferase; Diacylglycerol metabolism; Glycerolipid synthesis; Triacylglycerol
Mammalian glycerol-3-phosphate acyltransferase
(GPAT1 ) catalyzes the acylation of sn-glycerol-3phosphate to form 1-acyl-sn-glycerol-3-phosphate,
thereby providing the committed step for the formation of glycerolipids [1,2]. Regulation of GPAT may
Abbreviations: BSA, bovine serum albumin; DTT, dithiothreitol; EDTA, ethylenedinitrilo-tetraacetic acid; GPAT, glycerol-3-phosphate acyltransferase; NEM, N-ethylmaleimide; PCR,
polymerase chain reaction; RACE, rapid ampli¢cation of
cDNA ends; SDS, sodium dodecylsulfate; UTR, untranslated
region
* Corresponding author. Department of Nutrition, CB# 7400,
University of North Carolina, Chapel Hill, NC 27599-7400,
USA. Fax: +1-919-966-7216.
1
Current address: Monsanto Company 44B, 800 N. Lindbergh Ave., St. Louis, MO 63167, USA.
2
Current address: Metabolic Diseases Clinical Research,
Glaxo-Wellcome Inc., RTP, NC 27709, USA.
be critical during metabolic transitions between fasting and refeeding and in the development of complex
metabolic disorders such as obesity, diabetes mellitus, and atherosclerosis. Understanding the regulation of GPAT is complicated by the existence of
two GPAT isoenzymes, one located in the outer mitochondrial membrane and the other in the endoplasmic reticulum [2]. These isoenzymes can be distinguished by their resistance (mitochondrial) or
sensitivity (microsomal) to NEM inhibition [1].
Mitochondrial GPAT was cloned from mouse liver
[3] and identi¢ed by a 40% similarity to a 300 amino
acid stretch of E. coli GPAT [4]. The mouse cDNA
was expressed in Sf9 cells and puri¢ed, but was enzymatically inactive until reconstituted with crude
soybean phosphatidylcholine [5]. Mitochondrial
GPAT from rat liver has been puri¢ed [6] and was
partially cloned (GenBank accession numbers
1388-1981 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved.
PII: S 1 3 8 8 - 1 9 8 1 ( 9 9 ) 0 0 1 0 3 - 1
BBAMCB 50418 5-8-99
416
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
Fig. 1. Nucleotide sequence and predicted amino acid sequence of rat mitochondrial cDNA. The 5P UTR and coding region shown
were sequenced in both directions. The asterisk denotes the stop site for translation, and the v indicates a 167 nucleotide deletion of
a sequence reported in mouse GPAT cDNA. Underlined portions indicate primers described in the text and the orientation of the
primers is shown as S (sense) or AS (antisense). The GenBank accession number is AF021348.
U36771, U36772, U36773). A sequence for rat mitochondrial GPAT cDNA has been published in a review article [7].
With in vitro assays, puri¢ed mitochondrial GPAT
appears to prefer
rated acyl-CoAs
GPAT shows no
hypothesized that
BBAMCB 50418 5-8-99
saturated acyl-CoAs over unsatu[5,6], whereas the microsomal
preference [2]; thus, it has been
mitochondrial GPAT plays a crit-
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
417
Fig. 1. (continued)
ical role in esterifying saturated fatty acids at the sn-1
position of glycerolipids [2,8]. On the other hand,
changes in GPAT activity under di¡erent physiologic
conditions suggest that positioning of the sn-1-acyl
group cannot be the only function of mitochondrial
GPAT because both GPAT isoenzyme activities increase during adipocyte di¡erentiation [9], neonatal
development [10] in mammary epithelial cells during
lactation [11], and in di¡erent nutritional states
[12,13].
BBAMCB 50418 5-8-99
418
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
Fig. 1. (continued)
Acute regulation of GPAT, possibly by a phosphorylation/dephosphorylation mechanism, has also
been reported in adipose tissue, liver [2] and muscle
[14], although it has not been ascertained which
GPAT isoform is regulated. More recently, a tyrosine kinase, which was partially puri¢ed from adipose cytosol, inactivated microsomal GPAT [15].
GPAT inactivation was prevented by treatment
with tyrosine kinase inhibitors, and GPAT could be
reactivated by treatment with liver phosphatase. Our
recent studies show that mitochondrial GPAT, but
not microsomal GPAT, is inactivated by AMP-activated kinase [16]. These studies strongly suggest that
a phosphorylation/dephosphorylation mechanism
regulates one or both GPAT activities.
Information on mRNA expression is available
Table 1
Acyltransferase homology blocksa
Block I
b
E. coli
Mouse
Rat
303
VPCHRSHMDYLLL
LPVHRSHIDYLLL
LPVHRSHIDYLLL
227c
Block II
Block III
Block IV
348
GAFFIRR
GGFFIRR
GGFFIRR
271
382
YFVEGGRSRTGR
IFLEGTRSRSGK
IFLEGTRSRSGK
312
417
ITLIPIYI
ILVIPVGI
ILVIPVGI
347
Amino acid alignment of four conserved regions of GPAT from E. coli, mouse, and rat. Residues shown by site-directed mutagenesis
to be important for glycerol-3-phosphate binding and catalysis are in boldface type. E. coli GPAT residues H306, D309, F351, I352,
and P421 are important for GPAT catalysis, whereas R354, E385, and S389 are important for glycerol-3-phosphate binding [21].
a
Blocks of homology were identi¢ed based on amino acid alignments performed using the CLUSTAL algorithm.
b
Numbers indicate the amino acid residue at the beginning of each block in E. coli GPAT.
c
Numbers indicate the amino acid residue at the beginning of each block in rat GPAT.
BBAMCB 50418 5-8-99
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
only for the mitochondrial isoenzyme. Cloning of
mouse mitochondrial GPAT allowed investigators
to show that mitochondrial GPAT mRNA transcription is induced during adipocyte di¡erentiation [12],
by insulin treatment, in fasted mice refed a high carbohydrate diet [3], and by ADD1/SREBP [17]. Thus,
there is clearly an elegant mechanism of transcriptional control.
Both microsomal and mitochondrial GPAT activities change under a variety of hormonal, developmental and nutritional in£uences, however a
number of important questions remain unanswered
concerning the role of GPAT in glycerolipid biosynthesis. Do the two GPATs have di¡erent functions
regarding their use of saturated vs. unsaturated fatty
acyl-CoAs? Do they play speci¢c and independent
roles in (1) regulating the £ux of acyl-CoAs towards
glycerolipid synthesis or towards L-oxidation, (2) in
the synthesis of mitochondrial vs. microsomal phospholipids, and (3) in the synthesis of triacylglycerol
vs. phospholipid? How does the enzyme's position
within its membrane location relate to acute regulation of activity? In order to facilitate such studies
and to compare the mitochondrial GPAT with other
cloned acyltransferases, we cloned, sequenced, and
expressed mitochondrial GPAT cDNA from rat
liver.
Liver cDNA was prepared from 7 days old rats
and used as the template for PCR screening with
degenerate oligonucleotide primers homologous to
putative rat monoacylglycerol acyltransferase. A
490 basepair length product was obtained and cloned
into the pCR2.1 vector (Invitrogen) and sequenced
by the UNC automated DNA sequencing facility. A
sequence similarity search using the BLAST program
at NCBI [18] revealed that the fragment had an 89%
homology to mouse mitochondrial glycerol 3-phosphate acyltransferase (GenBank accession number
M77003) at nucleotides 3090^3555.
Additional oligonucleotide primers (indicated in
Fig. 1) were generated using the known mouse
GPAT sequence and used to amplify additional
GPAT sequences from rat liver cDNA by PCR.
Two fragments of 1.4 and 1.7 kb were obtained
with primers 3S/1AS and 3S/2AS respectively. The
two fragments were homologous to nucleotides
1711^3124 and 1711^3538 of the mouse GPAT
cDNA sequence. The 1.7 kb fragment was digested
419
Fig. 2. Expression of recombinant GPAT in Sf21 insect cells.
Cell lysates were electrophoresed in the presence of SDS on a
10% polyacrylamide gel and stained with Coomassie. Lane 1,
uninfected Sf21 cells; lane 2, Sf21 membranes containing overexpressed L-galactosidase; lane 3, Sf21 cells containing overexpressed GPAT (94 kDa).
with EcoRI to give two fragments of 0.8 and 0.9 kb.
The 0.8 kb fragment, which contained DNA homologous to the coding region from the mouse GPAT
cDNA sequence, was random prime labeled with
[32 P]dCTP and used to probe a liver cDNA library
prepared from 7 days old rats [19,20]. A 4.6 kb clone
was isolated and sequenced from both ends. One end
was found to have 89% homology to mouse GPAT
starting at nucleotide 2036. The other end contained
a poly(A) tail and was homologous to nucleotides
5931^6634 in the mouse GPAT 3P UTR.
The 4.6 kb rat GPAT cDNA clone did not contain
the start of the coding region. In order to obtain the
5P end of the rat GPAT cDNA, additional oligonucleotide primers were designed based on the mouse
GPAT cDNA sequence. Only primers 9S/10AS (Fig.
1) resulted in successful PCR ampli¢cation from rat
liver cDNA. The 9S/10AS PCR product was homologous to the mouse GPAT cDNA sequence encompassing nucleotides 1064^2085. Other primers designed to be homologous to nucleotides 1^20, 23^
BBAMCB 50418 5-8-99
420
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
Fig. 3. Inhibition of GPAT activity by DEPC and reversal with hydroxylamine. To isolate total particulate preparations, rat liver or
Sf21 cells was homogenized with 10 up-and-down strokes with a Te£on-glass homogenizer in 0.25 M sucrose, 1 mM EDTA, 10 mM
Tris-HCl, pH 7.4 and 1 mM DTT, then centrifuged at 100 000Ug for 1 h. To isolate microsomes, the homogenate was centrifuged at
15 000 rpm for 15 min at 4³C. The supernatant was centrifuged at 100 000Ug for 1 h at 4³C. The resulting total particulate or microsomal pellets were resuspended in the same bu¡er and frozen in aliquots at 380³C until use. Total particulate or microsomal membranes (40 Wg protein) were incubated with 300 WM (or 200 WM for recombinant GPAT) DEPC in acetonitrile (10% of the incubation
volume) for 10 min at 23³C. Controls contained 10% acetonitrile alone. After 10 min 75 mM (or 200 mM for recombinant GPAT)
hydroxylamine was added and samples were removed at intervals for assay of GPAT activity. GPAT activity was assayed using 75
mM Tris-HCl (pH 7.5), 1 mM DTT, 2 mg/ml BSA, 4 mM MgCl2 , 8 mM NaF, 300 WM [3 H]glycerol-3-P, and 112.5 WM palmitoylCoA in a total volume of 200 Wl [25] after a 15 min incubation on ice in the presence or absence of 1 mM NEM. Microsomal activity
was calculated by subtracting the NEM-resistant activity from the total activity. [3 H]glycerol-3-phosphate was synthesized enzymatically [26]. Protein was measured using BSA as the standard [27]. A: GPAT activity in rat liver microsomes. B: GPAT activity after
treatment of a rat liver total particulate preparation with 2 mM NEM on ice for 15 min to inhibit microsomal GPAT. C: Recombinant GPAT activity from Sf21 cells (infected for 3 days) after treatment of total particulate preparation with 2 mM NEM on ice for
10 min to inhibit endogenous microsomal GPAT. Control activities were 1.3, 0.4, and 3.2 nmol/min/mg protein for A, B, and C, respectively.
48, 188^208, or 558^581 in the 5P UTR of mouse
GPAT did not yield any PCR ampli¢ed products
using rat liver cDNA or total RNA as template.
To obtain the 5P UTR and start site, rapid ampli¢cation of cDNA ends (RACE; Clontech) was performed using rat liver cDNA and a primer which
annealed to both 8AS (Fig. 1) and the adapter sequence (Clontech). Sequencing of the RACE product
revealed that the rat and mouse 5P UTRs di¡ered
considerably. The mouse and rat 5P UTRs are 926
and 159 nucleotides, respectively, with the rat sequence lacking the ¢rst 607 nucleotides of the mouse
BBAMCB 50418 5-8-99
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
5P UTR and having an internal deletion of an additional 167 nucleotides (Fig. 1). The presence of a 600
nucleotide shorter 5P UTR in rat GPAT is consistent
with our inability to obtain PCR ampli¢ed products
from rat liver cDNA using primers homologous to
regions in the ¢rst 550 nucleotides of the mouse
GPAT cDNA. The presence of the deletion was con¢rmed by sequencing the product obtained using 24S
and 8AS primers in PCR ampli¢cation of rat liver
cDNA. The 3P UTR was only partially sequenced
but appeared to be about 84% homologous to the
3362 base mouse GPAT 3P UTR (data not shown).
The nucleotide sequences of the open reading frames
of rat and mouse mitochondrial GPAT are 89% homologous and the amino acid sequences are 96%
identical.
To verify that our cloned rat mitochondrial GPAT
cDNA encoded a functional GPAT enzyme, we employed the baculovirus expression system (BacPAK,
Clontech) to express our rat mitochondrial GPAT
clone in Sf21 insect cells. Nucleotides 160^2646 encompassing the rat mitochondrial GPAT ORF were
cloned into the pBacPAK8 vector (Clontech) following the polyhedrin promoter. Following infection of
Sf21 insect cells with control and recombinant GPAT
baculovirus for 3 days, NEM-resistant GPAT-specific activity measured in total particulate samples increased 24-fold from 0.08 nmol/min/mg in control
samples to 1.93 nmol/min/mg in mitochondrial
GPAT-containing samples. The intensity of the Coomassie-stained GPAT band at 94 kDa (Fig. 2, lane 3)
indicates abundant overexpression of mitochondrial
GPAT.
The results obtained with inhibitors are in strong
agreement with our observations with overexpressed
E. coli GPAT [21]. Since comparison of the rat and
mouse GPAT coding sequences with GPAT from E.
coli reveals the presence of conserved sequences that
are critical to glycerol-3-phosphate binding and catalysis in E. coli GPAT (Table 1) [21], we propose
that these acyltransferases function using a similar
catalytic mechanism. Through site-directed mutagenesis studies, we have shown that in E. coli GPAT,
residues H306, D311, F351, I352, and P421 are important for GPAT catalysis, whereas R354, E385,
and S389 are important for binding the glycerol-3phosphate substrate [21]. The inhibition data (see
below) and the sequence homologies suggest that
421
Fig. 4. Inhibition of GPAT activity by phenylglyoxal. Membrane proteins (0.2 mg) were incubated in 200 Wl in the absence
or presence of various concentrations of phenylglyoxal and then
assayed for GPAT activity at di¡erent time points. Phenylglyoxal was added to the incubation mixture (125 mM sodium
carbonate bu¡er, pH 7.5) to a ¢nal concentration of zero (E),
0.2 (b), 0.5 (7), 1 (R), or 2.5 mM (a). Data are presented as
percent of activity in control (sodium carbonate bu¡er only)
microsomes (0.96 nmol/min/mg), mitochondria (0.43 nmol/min/
mg), and Sf21 insect cells (2.75 nmol/min/mg). A: GPAT activity in rat liver microsomes. B: GPAT activity in rat liver total
particulate preparations treated with 2 mM NEM on ice for 15
min to inhibit microsomal GPAT. C: Recombinant GPAT
activity in Sf21 cell total particulate preparations treated with
2 mM NEM on ice for 10 min to inhibit endogenous microsomal GPAT.
the corresponding residues would be necessary for
mitochondrial GPAT activity.
In order to gain a better understanding as to which
amino acid residues might be important for activity,
to compare the responses of mitochondrial and microsomal GPAT activities, and to compare mitochondrial GPAT from rat liver with recombinant
mitochondrial GPAT, we tested several speci¢c amino acid reagents in isolated microsomes (for microsomal activity), in rat liver total particulate preparations treated with 2 mM NEM for 15 min on ice (for
mitochondrial activity), and in insect cell total particulate preparations treated with 2 mM NEM for 10
min on ice (for recombinant GPAT activity). Inhibition of both microsomal and mitochondrial GPAT
BBAMCB 50418 5-8-99
422
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
with diethylpyrocarbonate was reversible with
NH2 OH, consistent with the presence of a critical
histidine residue in each enzyme, however, the microsomal GPAT was more susceptible than the mitochondrial GPAT (67% vs. 40%) to inhibition by
300 WM DEPC (Fig. 3). Recombinant GPAT appeared more sensitive to DEPC, but activity was reversed only about 50% (Fig. 3C), similar to the reversal seen with the liver mitochondrial preparation
(Fig. 3B). As had been reported for GPAT from E.
coli [22], the arginine-speci¢c reagent phenylglyoxal
inhibited both microsomal and mitochondrial GPAT
activities (Fig. 4). Again, microsomal GPAT was
more sensitive to the inhibitor. Microsomal and mitochondrial GPAT activities were inhibited 88% and
59%, respectively, by 1 mM phenylglyoxal. A second
arginine-speci¢c inhibitor, 2,3-butanedione inhibited
GPAT activities similarly; at 5 mM, 2,3-butanedione
inhibited microsomal and mitochondrial GPAT activities 54% and 21%, respectively (data not shown).
Of interest was the ¢nding that 10-fold higher concentrations of 2,3-butanedione were required to inhibit the recombinant GPAT compared to mitochondrial GPAT from rat liver, but the inhibitory
concentrations of phenylglyoxal were similar for
each, perhaps due to the di¡ering abilities of each
inhibitor to access the critical arginine residue when
GPAT is greatly overexpressed. Aminophenylboronate, which inhibits serine-dependent proteases and
lipases [23], yielded a mixed type of inhibition with
the glycerol-3-phosphate substrate (data not shown).
With aminophenylboronate the apparent Ki values
were 2.9 mM for microsomal GPAT, 8.2 mM for
mitochondrial GPAT, and 10.3 mM for recombinant
GPAT activities.
Based on these inhibitor results, we speculate that
for rat mitochondrial GPAT, residues that play important roles in catalysis are equivalent to the residues that were shown by site-directed mutagenesis to
be critical for E. coli GPAT activity [21]. Thus, the
corresponding H231 and R277 residues in rat mitochondrial GPAT would be important (Table 1).
R277 in rat GPAT is equivalent to R211 in dihydroxyacetone-P acyltransferase (DHAP-AT). Of interest in this regard is a report that in four patients
with rhizomelic chondrodysplasia punctata type 2
due to de¢cient DHAP-AT, R211 was mutated to
histidine or cysteine [24].
Acknowledgements
This work was supported by United States Public
Health Service Grants DK56598 (to R.A.C.) and
HD08431 (to T.M.L.) from the National Institutes
of Health.
References
[1] R.M. Bell, R.A. Coleman, Annu. Rev. Biochem. 49 (1980)
459^487.
[2] R.M. Bell, R.A. Coleman, in: P.D. Boyer (Ed.), The Enzymes, Vol. XVI,, Academic Press, New York, 1983, pp.
87^112.
[3] D.-H. Shin, J.D. Paulauskis, N. Moustaid, H.S. Sul, J. Biol.
Chem. 266 (1991) 23834^23839.
[4] V.A. Lightner, R.M. Bell, P. Modrich, J. Biol. Chem. 258
(1983) 10856^10861.
[5] S.-F. Yet, Y.K. Moon, H.S. Sul, Biochemistry 34 (1995)
7303^7310.
[6] A. Vancura, D. Haldar, J. Biol. Chem. 269 (1994) 27209^
27215.
[7] A.V. Nikonov, T. Morimoto, D. Haldar, in: Recent Research Development in Lipids Research, Vol. 2, Transworld
Research Network Publishers, Trivandrum, 1998, pp. 207^
222.
[8] D. Haldar, A. Vancura, Methods Enzymol. 209 (1992) 64^
72.
[9] R.A. Coleman, B.C. Reed, J.C. Macball, A.K. Student,
M.D. Lane, R.M. Bell, J. Biol. Chem. 253 (1978) 7256^
7261.
[10] R.A. Coleman, E.B. Haynes, J. Biol. Chem. 258 (1983) 450^
465.
[11] M.R. Grigor, J.E. Allan, J.M. Carrington, A. Carne, A.
Geursen, D. Young, M.P. Thompson, E.B. Haynes, R.A.
Coleman, J. Nutr. 117 (1987) 1247^1258.
[12] S.-F. Yet, S. Lee, Y.T. Hahm, H.S. Sul, Biochemistry 32
(1993) 9486^9491.
[13] S.C. Jamdar, W.F. Cao, Biochim. Biophys. Acta 1255 (1995)
237^243.
[14] M. del C. Vila, G. Milligan, M.L. Standaert, R.V. Farese,
Biochemistry 29 (1990) 8735^8740.
[15] T.E. Lau, M.A. Rodriguez, Lipids 31 (1996) 277^283.
[16] D.M. Muoio, K. Seefeld, L.A. Witters, R.A. Coleman, Biochem. J. 338 (1999) 783^791.
[17] J. Ericsson, S.M. Jackson, J.B. Kim, B.M. Spiegelman, P.A.
Edwards, J. Biol. Chem. 272 (1997) 7298^7305.
[18] S.F. Altshul, W. Gish, W.M. Miller, E.W. Myers, D.J. Lipman, J. Mol. Biol. 215 (1990) 403^410.
[19] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning :
A Laboratory Manual, 2nd Edn., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989.
[20] F.M. Ausubel, R. Brent, R.E. Kingston, D.D. Moore, J.G.
Seidman, J.A. Smith, K. Struhl, Current Protocols in Mo-
BBAMCB 50418 5-8-99
B. Ganesh Bhat et al. / Biochimica et Biophysica Acta 1439 (1999) 415^423
lecular Biology, Vol. 1, John Wiley and Sons, New York,
1989.
[21] T.M. Lewin, P. Wang, R.A. Coleman, Biochemistry 38
(1999) 5764^5871.
[22] P.R. Green, R.M. Bell, Biochim. Biophys. Acta 795 (1984)
348^355.
[23] P.E. Fielding, C.J. Fielding, in: D.E. Vance, J. Vance (Eds.),
Biochemistry of Lipids, Lipoproteins and Membranes, Vol.
20, Elsevier Biomedical Press, Amsterdam, 1991, pp. 427^
459.
423
[24] R. Ofman, E.H. Hettema, E.M. Hogenhout, U. Caruso,
A.O. Muijsers, R.J.A. Wanders, Hum. Mol. Genet. 7
(1998) 847^853.
[25] D.M. Schlossman, R.M. Bell, Arch. Biochem. Biophys. 182
(1977) 732^742.
[26] Y.-Y. Chang, E.P. Kennedy, J. Lipid Res. 8 (1967) 447^455.
[27] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall,
J. Biol. Chem. 193 (1951) 265^275.
BBAMCB 50418 5-8-99