Transcriptional Regulation of p90 with Sequence Homology to Escherichia coli Glycerol-3-phosphate Acyltransferase”

We have previously isolated cDNA clones for several mRNAs that are dramatically increased in livers of fasted mice refed a high carbohydrate diet. We report here the sequence and regulation of one such mRNA; the 6.8-kilobase mRNA has an open reading frame of 2481 nucleotides, and the coded protein contains 827 amino acid residues (Mr of 90,000) with a 30% identity and an additional 42% similarity in an approximately 300-amino acid stretch to Escherichia coli glycerol-3- phosphate acyltransferase. The p90 mRNA is highly expressed in liver and in adipose tissue. When previ- ously fasted mice were refed a high carbohydrate, fat-free diet, the liver mRNA level for p90 was increased about 20-fold at 8 h. Administration of dibutyryl cAMP at the time of refeeding prevented the increase in the p90 mRNA by 70%. In addition, there was no increase in the p90 mRNA level when previously starved strep- tozotocin-diabetic mice were refed. In diabetic animals, the p90 mRNA level increased by 2-fold 1 h after insulin injection and reached a maximum of 19-fold after 6 h. The increase in transcription rate of the p90 gene preceded that of steady state mRNA level caused by fastingirefeeding, and cAMP abolished the increase in transcription. Transcription of the p90 gene was not detectable in either fasted or refed streptozotocin-dia-betic mice, but increased 4-fold 30 min after insulin administration and

Transcriptional Regulation of p90 with Sequence Homology to Escherichia coli  We have previously isolated cDNA clones for several mRNAs that are dramatically increased in livers of fasted mice refed a high carbohydrate diet. We report here the sequence and regulation of one such mRNA; the 6.8-kilobase mRNA has an open reading frame of 2481 nucleotides, and the coded protein contains 827 amino acid residues (Mr of 90,000) with a 30% identity and an additional 42% similarity in an approximately 300-amino acid stretch to Escherichia coli glycerol-3phosphate acyltransferase. The p90 mRNA is highly expressed in liver and in adipose tissue. When previously fasted mice were refed a high carbohydrate, fatfree diet, the liver mRNA level for p90 was increased about 20-fold at 8 h. Administration of dibutyryl cAMP at the time of refeeding prevented the increase in the p90 mRNA by 70%. In addition, there was no increase in the p90 mRNA level when previously starved streptozotocin-diabetic mice were refed. In diabetic animals, the p90 mRNA level increased by 2-fold 1 h after insulin injection and reached a maximum of 19-fold after 6 h. The increase in transcription rate of the p90 gene preceded that of steady state mRNA level caused by fastingirefeeding, and cAMP abolished the increase in transcription. Transcription of the p90 gene was not detectable in either fasted or refed streptozotocin-diabetic mice, but increased 4-fold 30 min after insulin administration and further increased up to 8-fold at 2 h. On-going protein synthesis was necessary for this increase.
In the course of cloning specific genes which are under hormonal and nutritional control, we have isolated cDNA sequences to a novel protein p90 (1). The p90 mRNA level was elevated dramatically in the liver of previously fasted mice which were refed a high carbohydrate diet. The dietary manipulation of fasting/refeeding is known to induce lipogenesis, associated with increased circulating insulin and decreased glucagon. We have also reported that insulin and cAMP independently affect the mRNA level for p90; treatment of mature 3T3-Ll adipocytes with insulin elicited a 3fold increase in p90 mRNA and dibutyryl cAMP decreased * This work was sponsored in part by National Institutes of Health Grant DK 36264. 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.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession numbeds) M77003. the level to 10% of control. Overall, the pattern of expression of the p90 mRNA was similar to that of fatty acid synthase. These observations led us to believe that p90 mRNA may code for a protein which is involved in lipogenesis.
Glycerol-3-phosphate acyltransferase (EC 2.3.1.15) catalyzes the committed step of triacylglycerol and phospholipid biosynthesis by generating lysophosphatidic acid in mammals (2,3). Glycerol-3-phosphate acyltransferase activity is present in the microsomal membrane fraction which is the principal site for glycerolipid synthesis and also in the outer mitochondrial membrane (4)(5)(6). In liver, 50% of total activity is found in the mitochondrial fraction, while in most other tissues microsomal glycerol-3-phosphate acyltransferase activity is about 10 times that of the mitochondrial fraction (4,7). The partitioning of the fatty acids for esterification from those for oxidation is partly carried out by glycerol-3-phosphate acyltransferase and is known to be under nutritional and hormonal control. The glycerol-3-phosphate acyltransferase activity is thought to be increased by insulin and decreased by starvation, and the effect is presumably greater on the Nethylmaleimide-insensitive mitochondrial enzyme (8,9). In spite of the important roles it may play in the regulation of triacylglycerol and phospholipid biosynthesis, mammalian glycerol-3-phosphate acyltransferase has not been purified or characterized. The structural gene for the 83-kDa E. coli snglycerol-3-phosphate acyltransferase (plsB), however, has been identified and sequenced (10,11).
In the present paper, we report cloning, structure, and hormonal regulation of a murine protein p90 with sequence homology to E. coli glycerol-3-phosphate acyltransferase. The deduced amino acid sequence of the cDNA for the 6.8-kb' mRNA revealed a 30% identity and an additional 42% similarity in the 322-amino acid residue stretch with Escherichia coli glycerol-3-phosphate acyltransferase. The p90 mRNA was expressed in high levels in lipogenic tissues, such as liver and adipose tissue. Insulin causes a marked and rapid induction of the p90 mRNA in diabetic mice. In addition, the hepatic p90 mRNA induction by fasting/refeeding is prevented by cAMP and by streptozotocin diabetes. Furthermore, we report here that these hormonal effects are on the transcription rate of the gene coding for p90.

EXPERIMENTAL PROCEDURES
Construction of cDNA Library and Screening-Poly(A+) RNA was prepared as described previously from livers of fasted mice which were refed a high carbohydrate, fat-free diet and from 3T3-Ll adipocytes (1). The mRNA (5 pg each) was employed as a template for first strand synthesis catalyzed by avian myeloblastosis virus reverse transcriptase after oligo(dT) priming. Second strand was synthesized by the Klenow fragment of DNA polymerase 1 and RNase H (12). The double strand cDNAs were inserted into the EcoRI site of X g t l O and Xgtll as previously described (1). The resultant libraries were The abbreviation used is: kb, kilobase(s). screened by plaque hybridization with the two EcoRI-PstI fragments of previously isolated 5.3-kb cDNA sequence, first labeled with apP by random priming (13).
cDNA Sequencing-Phage DNA inserts of the isolated clones were subcloned into pBluescript SK+ (Stratagene). Nucleotide sequence analysis was carried out on alkali-denatured double-stranded templates by the chain termination method (14) using Sequenase (U. S. Biochemical Corp.) and universal primers or oligonucleotides synthesized by standard phosphoamidite techniques on a BioSearch synthesizer. Also employed were the overlapping deletion subclones of pBluescript insert constructed by successive exonuclease 111 and mung bean nuclease digestions (15). Nucleotide sequence analysis was carried out from both strands or from multiple independent overlapping clones on one strand.
Animal Treatment, RNA Preparation, and Northern Blot Anulysis-Diabetes was induced in 6-8-week-old CD-1 male mice by three weekly intraperitoneal injections of streptozotocin (10 mg/100 g body weight) (17). Mice were used 7 days after the last injection. Diabetes was confirmed by high fasting blood glucose level (>250 mg/ml) and urinary glucose by Clinistix (Ames). Insulin was administered to diabetic mice a t a combined dose of regular insulin (3 units/100 g; Sigma) intraperitoneally and Lente Insulin (30 units/100 g; Lilly) subcutaneously. Dibutyryl cAMP (6 mg/100 g) and theophylline (3 mg/100 g) were given intraperitoneally to the previously fasted mice a t the start of refeeding of a high carbohydrate, fat-free diet. Streptozotocin-diabetic animals and normal animals injected with dibutyryl cAMP consumed similar amounts of food as controls. Total RNA was isolated from five pooled livers using the modified phenol extraction method (16). Poly(A+) RNA was prepared by oligo(dT) chromatography. Poly(A+) RNAs were electrophoresed in 0.7% (unless otherwise indicated) formaldehyde-agarose gels (5 pg/per lane) and blotted onto nitrocellulose filters as described previously (1). The filters were hybridized with p13 with a 5.3-kb insert labeled with R2P in 50% formamide, 5 X SSC a t 42 "C and were washed at 65 "C in 0.1 x SSC, 0.1% sodium dodecyl sulfate.

Nuclear Run-on Transcription
Assays-Livers from three mice were homogenized in 5 volumes of buffer containing 0.32 M sucrose, 3 mM MgC12, 5 mM Hepes (pH 6.9), and 0.5 mM p-mercaptoethanol. Nuclei collected by centrifugation were washed once by centrifugation through a 2.1 M sucrose cushion at 20,000 rpm for 60 min in a Beckman SW 28 rotor. The nuclei were stored in liquid nitrogen in 50 mM Tris (pH 7.9), 5 mM MgCI2, 0.5 mM 8-mercaptoethanol, and 40% glycerol. Run-on transcription was carried out a t 25 "C for 45 min in a reaction mixture containing 10' nuclei and 100 pCi of [a-"PIUTP (3000 Ci/mmol) in a final volume of 0.5 ml. Labeled RNA was isolated and hybridized to 5 pg of plasmids fixed on nitrocellulose as described previously (17).

RESULTS AND DISCUSSION
Isolation of Overlapping cDNA Clones Coding for p9O"We have previously described the isolation of a cDNA clone (p13) coding for a novel mRNA which is induced by refeeding (1). We utilized differential hybridization screening of a partial cDNA library constructed with mRNA fractions larger than 28 S RNA in sucrose density gradient centrifugation (1).
When we used a synthetic poly(A)-tailed RNA ladder ranging from 0.24 to 9.5 kb in size instead of the ribosomal RNAs as standards in Northern blot analysis, this 5.3-kb cDNA sequence hybridized to an mRNA with a size of 6.8 kb (Fig. 1). The DNA sequence determination showed that the 5.3-kb insert of the cDNA clone p13 contains a poly(A) tail. To obtain sequences corresponding to the 5' region of this mRNA and also to verify the sequences in the 3' region, two additional libraries were constructed from fasted/refed liver and 3T3-Ll adipocyte (day 5) mRNAs by a modified Gubler and Hoffman method. These libraries were screened with the 0.7kb most 5' sequence of p13 and the 5.3-kb total p13 insert. Overlapping cDNA clones were isolated and characterized. The sequence at the 5'-end of the p90 mRNA, however, was present only in a single cDNA clone (Fig. 2).
cDNA Sequence of p9O"The 6650-nucleotide cDNA sequence for the 6. vector and is presented in Fig. 3. The longest open reading frame of 2481 nucleotides precedes the 3240 nucleotides of the 3"untranslated sequence. This open reading frame starts at the Met codon of nucleotide position 1 designated in Fig.  3. The sequence immediately preceding the assigned initiator ATG exhibits characteristics of translation start sites in eukaryotic mRNAs; ATG is preceded by CTGCC with one nucleotide diverged from the consensus sequence of CC(G/ A)CC (18). The sequence of 926 nucleotides upstream from the ATG codon was also established from the available cDNA clones. The unusually long leader sequence may participate in the regulated expression of this gene. Although it is likely that the initiator ATG is correctly assigned, final proof will require purification and amino acid sequence analysis of p90.
A search of the GenBank and EMBL protein data bases with the deduced amino acid sequence revealed homology with glycerol-3-phosphate acyltransferase from E. coli as shown in Fig. 4. The alignment of deduced amino acid sequence of this novel murine protein p90 to the 806-residuelong E. coli glycerol-3-phosphate acyltransferase (10,11) starts with amino acid residue 153 and ends with residue 444 with a 1 residue gap at position 417, two %residue gaps at positions 210 and 396, and a 6-residue gap at position 281 of '-926 *ttggagtctgattcCc~g=tgtg~tgggt=t~~ttgttgtgttt=g*tg -876 r q l C .~t g a C C t t l l t C t g C t a c . l l g g~~g g~~g~~t~~~= = = = q~~~~= =~~~t g g = = = = -816 C c a g t t c~g~g g t t

g 1 c t g a c c a t c t t~t t g = g = t t c t g t =~t t t t g t t t~~= =~= t~g = -756 t t g t t c s a g a t t t g g s t a t g g t g =~g = g~t t t t t~* =~~= = t~= t = t~t = g t~~t t = = t~
-636 a C t C C t~. t t t C E~g C g a g t g c a t t t t t =~t~~g t t a g g g g~g~t g~g =~t t = t = t t g t c t = -696 CLCCcgag.CtgtgaC~tgCagcaggttttt=t=t~tgt=*=~=tt=~t~ttg===gtt*

-576 a a g c a t s a g a a g c a t t t g a a~t =~g~t =~= t =~g g a a a a t~t = t t t~t t t t t~t = =~~~q -516 aattgatgataagtttcatcsrcctgttcacttcagcttgg~~gg=g~~~tttt==g~tt -456 t t t c t c t~i a g t t c t c t t a t c t g t g t g~t g~g t t t~~=~=~t = = =~= t g~~t = g~t t = t t~= t -396
CtCttggaCtgttggCtgCtgCSgCggtgCttCtCtggggtt~=tg=tgtgt~==~g~=g -336 tctccagcttccagctacacaaltaggcctctgg~gg~g~tt==*~gt=~~~===~==t~ -276 g g g~t g t~t t~t t~t g a l g~t g t = t g g q a c a g~r t t t g = t~g t =~= = t g~t t~= = =~t~=~ -156 g g a a a g a a a a g a t a t t a a a g c a q c g g c t c t c r c a q q q a g a g~g~~t~g g g g~= t~= t~g~   -216 t c c a g a t a t c a c a t c a a q g a t a c a g a g c c t q g c a a s a g a t~~= = = t g~~~~~=~t =~~~g   -96 a g c c a c a a g c c a a = g a g c t a g a l l g c c g a t t t g~~t~t t t t g g g g~=~*~t t~t t~~~=~g -36 g l g t t g e c a c~t g g t t t g g g . c t t g c a c p t t c t~t g =~ ATG GAG GAG TCT TCA GTG p90. In this stretch of 292 residues, there is a 30% identity and an additional 42% similarity in amino acid sequence.
There is no homology outside this region between p90 and E. coli glycerol-3-phosphate acyltransferase. Although we have tentatively suggested p90 as a mammalian glycerol-3-phosphate acyltransferase-like protein, it is also a possibility that p90 may be a protein with only a similar functional domain of glycerol-3-phosphate acyltransferase. Structure/function 150 160 170    Poly(A+) RNA (5 pg) were prepared from mouse brain, lung, liver, kidney, and skeletal muscle and were subjected to electrophoresis in 0.7% agarose gels under denaturing conditions. The resolved RNAs were transferred to nitrocellulose filters and hybridized with ?'Plabeled p13. studies have not been reported in E. coli glycerol-3-phosphate acyltransferase. The hydropathy profile and the predicted secondary structure of the open reading frame obtained by computer analysis did not show any recognizable transmembrane domains or mitochondrial target signal (data not shown). It has been suggested that glycerol-3-phosphate acyltransferase activity may be regulated by phosphorylation and dephosphorylation (20). The glycerol-3-phosphate acyltransferase activity in rat adipocyte microsomes was reported to decrease when incubated with CAMP-dependent protein kinase in vitro, and the heat-stable CAMP-dependent protein kinase inhibitor prevented inactivation. At the near C-terminal region of p90, a putative phosphorylation site for CAMP-dependent protein kinase (Lys-Arg-Val-Ser) is present. Further studies are needed for the positive identification of p90.

S.SEO1 E A I A~~L N P D G S A O O O S K A I O K~R K~K I L O~A T V S P G~I R L T G L K L F N S F PLSBSE K L L A S R A I~A L V E D E~S K K I S H E K A O O N A I~L L W E I M N F S Y E~I R L T D R I L
Tissue Distribution of p90 mRNA-A11 mouse tissues we examined, liver, muscle, kidney, brain, and lung, expressed p90 mRNA (Fig. 5). Liver showed the highest concentration while muscle and kidney were intermediate, and lung and brain showed the lowest concentration of p90 mRNA. Since poly(A+) mRNAs used for the Northern blot were prepared from refed mice, the different mRNA levels may be due to tissue differences in inducibility by fasting/refeeding. Overall, the pattern of p90 mRNA expression was quite similar to that of fatty acid synthase mRNA we previously reported (1,17). We have also observed a high level of p90 mRNA in adipose tissue comparable to that in liver (data not shown).
In Vivo Hormonal Regulation of p90 mRNA in Liver-We have examined the hormonal and nutritional regulation of the p90 mRNA level in mouse liver (Fig. 6). The p90 mRNA level increased 20-fold after 8 h and further increased to 30-fold after 16 h of refeeding previously starved mice with a high carbohydrate, fat-free diet, a state in which lipogenesis would be high. Since fastinglrefeeding causes an increase in insulin secretion and a decrease in circulating glucagon level, the effects of diabetes and cAMP in the nutritional induction of p90 mRNA were examined. When previously starved streptozotocin-diabetic mice were given a high carbohydrate diet for 8 h, there was no significant increase in the p90 mRNA level. In addition dibutyryl CAMP, administered at the time of refeeding to normal starved animals, inhibited by 70% the increase in the p90 mRNA level caused by feeding. This indicates that the very low level of p90 mRNA observed in fasting probably was caused by increased circulating glucagon. Similar to the regulation of fatty acid synthase mRNA, we have previously reported (1,17) positive regulation by insulin and negative regulation by cAMP are indicated for p90 mRNA expression in liver. We have previously reported that in 3T3-L1 adipocytes, dibutyryl cAMP decreased the p90 mRNA level by 80% while insulin increased the level 2.5-fold (1). Although the effects of these counter-regulatory hormones on the p90 mRNA level are much greater in liver in uivo, we conclude that the p90 mRNA expression is under similar control in liver and in adipocytes.
Effect of Fasting/Refeeding and cAMP on Transcription Rate of the Gene Coding for p9O"We have measured the transcription rate for the p90 gene in liver of normal mice previously fasted and then refed a high carbohydrate, fat-free diet. The transcription rate increased 2.5-fold at 6 h, 7-fold at 9 h, and reached 22-fold after 16 h of refeeding (Fig. 7).  in streptozotocin-diabetic mice. Poly(A+) RNA was prepared from livers of streptozotocin-diabetic mice killed at the indicated times after insulin administration. Saline-injected mice were also used at 8 h as a control (8C). Northern blot analyses were carried out as described for Fig. 5 and under "Experimental Procedures." The autoradiogram of the Northern blot was scanned densitometrically. The data were normalized to the values obtained prior to insulin administration and are presented as -fold increases. The level of actin mRNA remained the same through the experiment. Similar results were obtained from two separate experiments.
Increase in mRNA level and transcriptional rate for p90 caused by fasting/refeeding was similar but somewhat slower than that for the fatty acid synthase gene we have reported previously (17). When dibutyryl CAMP was given at the start of refeeding of normal animals, there was no increase in transcription, indicating a negative transcriptional regulation FIG. 9. Effect of insulin on the transcriptional rate for p90 in streptozotocin-diabetic mouse liver. Nuclei were isolated from pooled livers of three diabetic mice and diabetic mice treated with insulin as described under "Experimental Procedures." Run-on transcription and hybridization with p90 (p13, +), actin (pAM91, W), cfos, and pBR322 fixed on nitrocellulose were carried out as described under "Experimental Procedures." Exp. 1, 1/2, 2, and 12 h after insulin injected intraperitoneally; a-amanitin, run-on transcription carried out in the presence of 1 pg/ml a-amanitin with liver nuclei prepared from mice treated with insulin for 2 h; CHX, run-on assays of nuclei prepared from livers of mice treated with insulin for 2 h and injected with cycloheximide intraperitoneally (1 mg/100 g body weight) 30 min prior to insulin administration. Exp. 2, same as in Exp. 1, but run-on assays were carried out with the liver nuclei prepared from mice 1/2, 1, 2, and 6 h after insulin injection. The autoradiogram of Exp. 2 was used for quantitative densitometric scanning. The results were normalized to the value obtained at time zero (before insulin administration) and are presented as -fold change. of p90 mRNA by CAMP. Furthermore, we did not detect any appreciable transcription for p90 in fasted or refed streptozotocin-diabetic animal liver, indicating the requirement of insulin for transcriptional induction by fasting/refeeding. Regulation ofp90 mRNA by Insulin-As shown above, when we carried out fastinglrefeeding experiments with streptozotocin-diabetic animals to examine the effect of insulin on p90 mRNA, there was no detectable mRNA or transcription for p90 in livers of refed animals indicating an insulin requirement for the induction of p90 by fasting/refeeding (Figs. 6  and 7). To determine whether insulin plays a direct physiological role in regulating the p90 mRNA level, the time course of the steady state mRNA level was followed after insulin administration in streptozotocin-diabetic mice (Fig. 8). The p90 mRNA level increased 2-fold 1 h after injection of insulin and further increased to attain a maximum level of 19-fold by 6 h. The magnitude and time course of p90 mRNA induction by insulin in diabetic animals was similar to our previous reports on that for fatty acid synthase, an enzyme crucial for lipogenesis (17). It is known that glucagon secretion is de-  /100 g body weight) was injected intraperitoneally 30 min prior to insulin administration. Insulin was given intraperitoneally to diabetic mice; after 6 h, poly(A+) RNA was prepared from livers and Northern blot analyses were carried out as described in Figs. 5  creased when insulin is administered to diabetic animals (21). It cannot be ruled out that the decreased circulating glucagon may have contributed to the increased mRNA level for p90 by insulin administration. However, it is likely that insulin plays a positive role in increasing the p90 mRNA level, considering the rapid increase in transcription shown in this experiment and the independent effect on the p90 mRNA level observed in 3T3-Ll adipocytes (1). The dramatic increase in mRNA level for p90 by insulin may be due to the insulin effect on the transcriptional rate or on post-transcriptional events. The insulin effect on the mRNA level for p90 in diabetic animals, however, was abolished by actinomycin indicating transcriptional regulation by insulin (data not shown). Therefore, we carried out the transcription run-on analysis in isolated liver nuclei from diabetic animals at various times after insulin administration (Fig. 9). There was a rapid and marked increase in the transcription rate for p90 when insulin was given to the streptozotocindiabetic mice; the transcription rate increased 4-fold after 30 min and attained a maximal increase of %fold a t 2 h (Fig. 9,  Exp. 2). The increase in transcription preceded and probably contributed to the increase in p90 mRNA level observed in Fig. 8. The time course of the increase in the transcription rate for p90 elicited by insulin was similar to that for the fatty acid synthase gene we previously reported (17). We tested whether the insulin effect on this specific transcription requires ongoing protein synthesis and/or a regulatory protein with a short half-life. As shown in Fig. 10, the increase in p90 mRNA level was not observed in the presence of cycloheximide after 6 h of insulin administration to streptozotocindiabetic mice liver. Also abolished by cycloheximide was the increase in transcription for p90 by insulin (Fig. 9, Exp. 1 ). We observed that cycloheximide also abolished the effect of insulin on the fatty acid synthase gene transcription (data not shown). It appears that a similar mechanism is involved in insulin regulation of these genes. Stumpo and Blackshear (22) reported a transient increase in c-fos mRNA in 3T3-Ll cells treated with insulin and proposed a possible role for cfos in the induction of lipogenic enzymes. In addition, these investigators showed that the insulin effect on c-fos is through the insulin receptor, and the serum-response element in the c-fos promoter is necessary for the increase in c-fos transcription by insulin (23). We did not detect any change in c-fos transcription from 30 min to 12 h after insulin administration in diabetic mice (Fig. 9, Exp. 1). The transcription rate measured in these experiments was by RNA polymerase I1 as indicated by the inhibition shown by a-amanitin. The p90 gives us a good model system for studying positive transcriptional control by insulin and negative control by cAMP at a molecular level.
p90, which is up-regulated by insulin and down-regulated by CAMP, is expressed highest in the tissues where and when lipogenesis is high. The marked effects of insulin and cAMP on the mRNA level and the transcription rate of this gene indicate the possibility that p90 may be the mitochondrial enzyme. There are conflicting reports on the hormonal regulation of glycerol-3-phosphate acyltransferase activities. Both mitochondrial and microsomal enzyme activities have been shown to decrease in adipose tissue by streptozotocin diabetes, and insulin administration restored them significantly (8,9). On the other hand, the glycerol-3-phosphate acyltransferase activity of the mitochondrial form, but not the microsomal form, has been shown to decrease in starvation and by antiinsulin serum treatment in liver (8). Nevertheless, the mitochondrial glycerol-3-phosphate acyltransferase is generally thought to be more sensitive to hormones (2, 3). Antibodies against this glycerol-3-phosphate acyltransferase-like protein will make it possible to characterize and to demonstrate positive identification and subcellular localization of p90.