Communication Vol. 269, No. 4, Issue of January 28, pp. 2361-2364. 1994 THE JOURNU OF B I O ~ I C A L CHEMISTRY

0 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Mammalian AMP-activated Protein Kinase Shares Structural and Functional Homology with the Catalytic Domain of Yeast Snfl Protein Kinase*

(Received for publication, November 5, 1993)

Ken I. MitchelhillS, David StapletonS, Guang GaoB, Colin House, Belinda Michell, Frosa Katsis, Lee A. Wittersi, and Bruce E. Kempll From St. Vincent's Znstitute of Medical Research, 41 Victoria Parade, Fitzroy, Victoria 3065, Australia and the §Endocrine-Metabolism Division, Dartmouth Medical School, Hanouer, New Hampshire 03755

The AMP-activated protein kinase is responsible for the regulation of fatty acid synthesis by phosphoryla- tion of acetyl-coA carboxylase. It may also regulate cho- lesterol synthesis via phosphorylation and inactivation of hormone-sensitive lipase and hydroxymethylglutaryl- CoA reductase. We have purified the "activated pro- tein kinase 14,000-fold from porcine liver. The 63-kDa catalytic subunit co-purifies with two proteins of 40 and 38 kDa that may function as subunits. Partial amino acid sequence of the 63-kDa subunit revealed a striking homology with the catalytic domain of the yeast protein kinase transcriptional regulator Snfl and its plant homologs. The Snf l (72 kDa) and Snf4 (36 kDa) complex was also purified and found to phosphorylate the AMP- activated protein kinase peptide substrate, HMRSAMS- GLHLVKRR-amide, but was not activated by AMP. Both Snfl/4 and the "activated protein kinase phospho- rylate and inactivate yeast acetyl-coA carboxylase in vitro. These results indicate that during evolution the catalytic domain sequences of the Snf l protein kinase subfamily have been exploited in the control of mamma- lian lipid metabolism and raise the possibilities that the AMP-activated protein kinase may have other sub- strates involved in regulating gene expression path- ways, as well as Snfl homologs participating in the con- trol of lipid metabolism in many eukaryotic organisms.

The AMP-activated protein kinase is thought to play a key regulatory role in the synthesis of fatty acids and cholesterol (reviewed in Ref. 1). It has been recognized for a number of years that hydroxymethylglutaryl-CoA reductase is regulated by protein phosphorylation (2 ) , leading to a reduction in en-

the National Health and Medical Research Council, and the National * This work was supported in part by the National Heart Foundation,

Institutes of Health (Grant DK35712 to L. A. W.). The costs of publica- tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

to the SwissProt Data Bank. The amino acid sequencefs) reported in this paper has been submitted

S These authors contributed equally to this work. ll Fellow of the National Health and Medical Research Council. To

whom correspondence should be addressed: St. Vincent's Institute of Medical Research, 41 Victoria Parade, Fitzroy, Victoria 3065, Australia. Tel.: 61-3-2882480; Fax: 61-3-4162676.

zyme activity. AMP has been found to stimulate hydroxymeth- ylglutaryl-CoA reductase phosphorylation (3). While the physi- ological significance of hydroxymethylglutaryl-CoA reductase regulation by phosphorylation remains equivocal, there is strong evidence that liver acetyl-coA carboxylase, the rate- limiting enzyme in fatty acid synthesis, is regulated by phos- phorylation involving several protein kinases including the AMP-activated protein kinase (4-7) and a Mn2+-dependent pro- tein phosphatase (8). The AMP-activated protein kinase has also been shown to phosphorylate the hormone-sensitive lipase (9), leading to the concept that this enzyme plays a coordinat- ing role in lipid metabolism (1). In addition to allosteric control by AMP, the AMP-activated protein kinase is also thought to be regulated by phosphorylation by another protein kinase (4). We have observed that insulin treatment of Fao hepatoma cells leads to the simultaneous inhibition of the AMP-activated pro- tein kinase and activation of acetyl-coA carboxylase (10).

EXPERIMENTAL PROCEDURES Purification-The AMP-activated protein kinase was purified from

porcine liver using a modification of the procedure reported for rat liver (11) using the same buffers. Liver (1 kg) was homogenized in 4,000 ml of buffer. A 2.5-7.0% (w/v) PEG' 6000 fraction was prepared and the resultant fraction batched onto 1,500 ml of DEAE-cellulose (Whatman) and eluted with buffer containing 0.25 M NaCl (2,000 ml). The eluate was chromatographed on 150 ml of Blue Sepharose (Pharmacia) and the AMP-activated protein kinase eluted with buffer containing 1 M NaCl. The enzyme fraction was concentrated and desalted by 10% (w/v) PEG 6000 precipitation prior to chromatography by peptide substrate affin- ity chromatography (12).2 The peptide substrate affinity column was washed with the same buffer containing 0.1% (v/v) Triton X-100 and 0.5 M NaCl and the AMP-activated protein kinase eluted with this buffer containing 2 M NaCl and 30% (v/v) ethylene glycol.

Acetyl-coA carboxylase was prepared from membrane-free yeast ex- tracts using avidin affinity chromatography and gel chromatography on Sepharose 6B. The purified acetyl-CoAcarboxylase migrated as a single band on SDS-PAGE (Fig. 1C). Acetyl-coA carboxylase was assayed by the incorporation of [14ClHC0, into acid-stable products as described (8). Purified acetyl-coA carboxylase (3.5 pmol, specific activity 9.1 pmol.rnin-'.mg-') was incubated with various concentrations of Snfll4 kinase or AMP-activated protein kinase in buffer containing 50 mM HEPES buffer, pH 7.5, 1 m~ dithiothreitol, 10 mM MgC12, 2 mM sodium citrate, 1 mg/ml bovine serum albumin in the presence or absence of 200 1.1~ AMP. The reactions were initiated by adding 200 p~ ATP and at the completion of the 60-min incubation at 30 "C, sodium [l4C1HCO3, ace- tyl-coA, additional ATP, buffer, and bovine serum albumin were added to achieve the concentrations described under the acetyl-coA carboxyl- ase assay conditions. For 32P radiolabeling, acetyl-coA carboxylase (3.5 pmol) was phosphorylated in the same buffer in the absence of bovine serum albumin with 0.2 m~ [Y-~~PIATP (1,000 cpm/pmol).

The Snfl/4 complex was purified from fresh compressed bakers' yeast (Saccharomyces cereuisiae) following freezing in liquid nitrogen and homogenizing with 100 g of 0.2-mm glass beads in a Waring blender for approximately 5 min. Liquid nitrogen was added during the homogeni- zation at 1-min intervals. The dry, frozen yeast powder was extracted for 3 min in 300 ml of 20 m~ Tris.HC1 buffer, pH 8.0,2 M NaCl, 0.2% (v/v) Tween 20, 10 mM 2-mercaptoethanol, 1 m~ EDTA, 1 m~ EGTA, 5 mM sodium pyrophosphate, 50 mM NaF, 5 pg/ml soybean trypsin inhibitor, 0.1 mM phenylmethylsulfonyl fluoride, and 1 mM benzamidine. The homogenate was centrifuged at 17,700 x g for 20 min and the pellet discarded. The supernatant was adjusted to 7% (w/v) PEG 6000 (using 28% (w/v) aqueous PEG) and stirred at 4 "C for 30 min. Following centrifugation at 4,500 x g for 20 min, the pellet was dissolved in 300 ml

The abbreviations used are: PEG, polyethylene glycol; PAGE, poly-

D. Stapleton, K. I. Mitchelhill, G. Gao, C. House, B. Michell, F. acrylamide gel electrophoresis.

Katsis, L. A. Witters, and B. E. Kemp, manuscript in preparation.

2361

2362 AMP-activated Protein Kinase Snfl Homolog

A

fie. 1. SDS-PAGE of the AMP-acti- vated protein kinase Snfl and acetyl- CoA carboxylase. The gradient acryl- amide gel (515% (w/v)) was stained with A M p a p ~ Coomassie Blue. hne A is 80 pg of puri- fied AMP-activated protein kinase (pre- 63kDa+ II parative scale), lane B is 3 pg of the SnW4 kinase, and lane C is 1 pg of the yeast acetyl-coA carboxylase. 40kDa

38kDa

B

yACC +B

snf4 + -

C

2OOkDa + 0

l l6kDa + - 97.4kDa + -

66kDa + 0

45kDa + I

3lkDa + - 21.5kDa + -

of buffer A (20 nm HEPES buffer, pH 7.0, containing 0.01% (v/v) 'heen 20, 10 m~ 2-mercaptoethanol containing 0.1 nm phenylmethylsulfonyl fluoride and 1 nm benzamidine) and the insoluble material removed by centrifugation at 4,500 x g for 20 min. The supernatant was chromato- graphed on 50 ml of DE-52-cellulose (Whatman) equilibrated with buffer A. The resin was washed on a sintered glass funnel with 200 ml of equilibration buffer and washed with 100 ml of buffer A containing 0.1 M NaCl, and then the enzyme activity was eluted with 200 ml of buffer A containing 0.5 M NaCl. The eluate was adjusted to 5 m~ imid- azole and stirred for 30 min with 20 ml of nickel-nitrilotriacetic acid- agarose (Qagen) equilibrated with 20 nm HEPES buffer, pH 7.0,lO m~ 2-mercaptoethanol, 0.01% (v/v) Tween 20, 250 m~ NaCl, and 5 m~ imidazole. The resin was washed with 100 ml of the equilibration buffer on a sintered glass funnel, transferred to a column, and washed with 50 ml of 20 nm HEPES buffer, pH 7.0, containing 10 nm 2-mercaptoetha- nol, 0.05% (v/v) Tween 20,0.05% (v/v) Triton X-100, 1 M NaCl, and 25 m~ imidazole and then with 50 ml of 20 nm HEPES buffer, pH 7.0, con- taining 10 nm 2-mercaptoethanol, 0.5 M NaCl, and 50 m~ imidazole a t 1 ml/min. The enzyme activity was eluted at 0.2 ml/min with 20 m~ HEPES buffer, pH 7.0, containing 10 2-mercaptoethanol, 0.01% (v/v) Tween 20, 2 M NaCl, and 150 m~ imidazole; 1-ml fractions were collected and assayed for peptide kinase activity. The active fractions were pooled, made up to 30% (v/v) with ethylene glycol, and stored at -20 "C.

Protein Kinase Assays-The AMP-activated protein kinase was as- sayed as described previously (13) wing the SAMs peptide substrate, HMRSAMSGLHLVKRR-amide. The enzyme was diluted in diluting buffer (20 nm HEPES, pH 7.0, 0.1% (v/v) Triton X-100) prior to assay, and the reactions were initiated by adding 10 pl of diluted enzyme to the reaction mixture containing peptide substrate. The reactions were stopped by withdrawing 30-pl aliquota and applying to P-81 papers as previously described (14).

Protein Sequencing"The AMP-activated protein kinase was precipi- tated using 6% (w/v) trichloroacetic acid, and then the precipitate was washed with 20% (v/v) ethanol in ether, dried, and dissolved in SDS sample buffer. Samples were subject to 12.5% (w/v) PAGE, stained with Coomassie Blue, destained, and dried from 5% (v/v) glycerol in water. Stained protein bands were either excised from the dried gels and rehydrated in 500 pl of 200 nm Tris.HC1 buffer, pH 8.5, 10% (v/v) acetonitrile with 2 pg of modified trypsin (sequencing grade, Promega) and incubated overnight with shaking at 37 "C or the digestions were done in a cartridge as described previously (15) and applied directly to the reversed phase column. In the former case, the supernatant was removed and the gel pieces were washed sequentially with 500 pl of 6 M guanidine hydrochloride, 500 pl of 5% (v/v) trifluoracetic acid (Pierce Chemical Co.), and 500 pl of 6 M guanidine hydrochloride. Each wash was achieved by sonication for 30 min (samples floating in a sonicating water bath), followed by centrifugation and removal of the supernatant. Supernatants were pooled, filtered through a 0 .45-p filter and chro- matographed on a Hewlett Packard model 1090 liquid chromatograph with monitoring at 214 and 280 nm and manual peak collection. Initial chromatography was performed on a Brownlee RP-300 C, column (2.1 x 250 mm; Applied Biosystems) and selected peaks were re-chromato- graphed on a 2.1 x 30-mm Hypersil column ( 5 - p ODS CIS; Phenome- nex). Both chromatographies were developed with a 040% Wv) aceto- nitrile (Mallinckrodt) gradient in 0.1% (v/v) trifluoracetic acid at 100

flmin. Purified peptides were subject to pulsed liquid phase protein sequencing on an Applied Biosystems model 471A Protein Sequencer.

RESULTS Porcine "activated protein kinase was purified 14,000-

fold using three chromatography steps, including DEAE-cellu- lose ion-exchange chromatography, Blue Sepharose chromatog- raphy and peptide substrate affinity chromatography (12). The yield of purified enzyme vaned in the range 100-400 pg from 1 kg of porcine liver and had a specific activity of approximately 11 pmol.min".mg" using the SAMS-peptide substrate (13). This specific activity is similar to the catalytic subunit of the cAMP-dependent protein kinase with Kemptide as substrate (15 pmol-min".mg') and is therefore unlikely to be due to a minor component of the purified protein preparation. SDS- PAGE of the purified enzyme preparation revealed that it con- tained three proteins of 63,40, and 38 kDa (Fig. lA). The 40- and 38-kDa proteins were present in approximately equimolar amounts to the 63-kDa protein based on visual comparison of the Coomassie Blue staining and were present in the same relative abundance in all preparations. In our attempts to pu- rify the rat liver "activated protein kinase using the pro- cedure previously reported (111, the purity and quantity of en- zyme was insufficient to definitely recognize that the 38- and 40-kDa proteins were specifically associated with the 63-kDa protein. Reinspection of the polyacrylamide gels reveals that they were always present along with many other proteins. The 63-kDa porcine protein was isolated by SDS-PAGE and partial amino acid sequence analysis obtained using peptides derived from tryptic and Asp-N protease digestions. The protein se- quence obtained showed a striking homology to the yeast Snfl protein kinase described previously (16) and plant homologs of Snfl. The resulting peptide sequences could be aligned over the entire catalytic domain of Snfl (Fig. 2, Snfl(62-294)). Sequence comparison with Snfl(47-325) revealed 66% identity and 75% if conservative substitutions were included. In the core of the catalytic domain the identity over 54 residues was 81%. Since Snfl is 633 residues and migrates on SDS-PAGE with an ap- parent mass of 72 kDa (Fig. 1B ) and the AMP-activated protein kinase is 63 kDa, we expect it to have approximately 554 resi- dues, closer in length to the plant homologs of Snfl that vary in size from 502 (RKIN1 from rye grass; Ref. 17) to 512 ( h b i - dopsis homolog; Ref. 18) or 513 residues (barley homolog; Ref. 19). Outside the catalytic core only two peptide sequences could be confidently aligned and additional peptide sequences total- ing 120 residues could not be unequivocally aligned. For these peptides a search of the protein sequence data bases revealed there was no strong homology to other known mammalian pro-

AMP-activated Protein Kinase Snfl Homolog

10 20 30 40 50 snf l 1 MSSNNNTNTA PANANSSHHH HHHHHHHHHH GHGGSNSTLN NPKSSLADGA

AMP-PK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . EGRV

s n f l 51 HIGNYQIVKT LCECSFCKVK LAYH-QK VALKIINKKV LAKSDMQGRI ... ...... ... ...... AMP-PK K v ................................ s n f l 101 EREISYLRLL RHPHIIKLYD VIKSKDEIIM VIEYAGNELF DYIVQRDKMS

AMP-PK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D

s n f l 151 EQEARRFFQQ IISAVEYCHR HKIVHRDLKP ENLLLDEHLN VKIADFCLSN . . . . . . . . . .................... . . . . . . . . . .................... AMP-PK EKESRRLFQQ ILSG.. . . . . . . . . . .DLW ENVLLDAHMN AKIADFCLSN

s n f l 201 IMTEGNFLKT SCGSPNYAAP EVISGKLYAG PEVDWSCCV ILYVMLCRRL ..... ................... . . . . . ................... AMP-PK -X XAXXPNYAAP EVISGRLYAG PEVD.. . . . . . . . . . . . . . . s n f l 251 PFDDESIPVL FKNISNGVYT LPKFLSPGAA GLIKRMLIVN PLNRISIHEI

: AMP-PK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DI

s n f l 301 HQDDWFKVDL PEYLLPPDLK PHPEEENENN DSKKKSSPD NDEIDDNLVN . . . . . . . . ........ AMP-PK XEHEXFKODL EI(yLEEE. ..

s n f l 351 ILSSTMGYEK / / 561 KIPDLMKMVI QLFQIETNNI' LVDFKFKWE

AMP-PK I / . . . . . . . . . . . . . . . QLYOWSR'PI LLDFR ..... . . . . .... . . . . ....

s n f l 601 SSYGDDl'l'VS NISEDEMSTF SAYPFLHLTT KLIMELAVNS QSN

FIG. 2. Sequence comparison between Snfl kinase and pep tides derived from the AMP-activated protein kinase 63-kDa subunit. The 63-kDa fragment was proteolyzed and the resultant pep- tides purified and sequenced as described under "Experimental Proce- dures." Where peptide sequences of independent peptide isolates were determined two or three times, they are underlined. The yield of phen- ylthiohydantoin-derivatives vaned from 10-40 pmol with repetitive se- quencing yields of greater than 90%. Approximately half the peptides were not readily aligned to the catalytic core and are presumed to be part of the COOH-terminal sequence.

tein kinases. The sequence relationship between the 63-kDa catalytic domain of the AMP-activated protein kinase and Snfl raised the question of whether Snfl was able to phosphorylate and inactivate yeast acetyl-coA carboxylase. Since Snfl has a naturally occurring 11-residue His repeat, we purified the Snfl/4 complex using nickel chelation chromatography. Fol- lowing purification, the Snfl preparation contained two major protein bands of 72 and 36 kDa, as well as a variable amount of a 50-kDa protein minor component (Fig. lB). The 72- and 36-kDa proteins corresponded to the masses of Snfl (16) and Snf4 (20) reported previously. The identities of Snfl and Snf4 were confirmed by Edman sequence analysis of tryptic pep- tides. Following tryptic digestion of the 72-kDa band, the pep- tide, NLGAEXAKPSEE, corresponding to residues 527-539 (KNLGAEWAKPSEE) was obtained from two independent preparations of Snfl. Two other peptides, LIMELAVNS and FLSPGAAGLIK, were sequenced from the 72-kDa band, corre- sponding to Snfl residues 622-630 and 274-284, respectively. The Snf4 tryptic peptides obtained were TSYDV (Snf4(3&34)) and YILLGSN (Snf4(316-322)). Furthermore, both co-purified through the peptide substrate affinity column designed for the AMP-activated protein kinase. Previously, Celenza et al. (20) had shown that Snfl and Snf4 are associated on the basis of co-immunoprecipitation experiments. Our results demonstrate that they both co-purify using at least three chromatography methods and indicate that they form a stable heteromeric com- plex. The NH2 terminus of the AMP-activated protein kinase was not identified from the partial peptide sequence, but it seems unlikely to contain the His repeat sequence found in Snfl (Fig. 21, since it was not retained on a nickel-nitrilotriace- tic acid-agarose affinity column.

A search of the protein data bases suggests that fragments of the AMP-activated protein kinase may have been fortuitously sequenced previously. A tryptic peptide fragment isolated from

A

10

I ;D +

2363

B

7 __.

0 3 5 10 20

Preincubation Time lminutesl

FIG. 3. Effect of phosphorylation on acetyl-coA carboxylase activity. Panel A, acetyl-coA carboxylase was preincubated with either 100 ng of the AMP-activated protein kinase or 175 ng of SnW4 kinase in the presence or absence ofAMP. Panel B, time course of inactivation of acetyl-coA carboxylase by Snfl. Results shown are the average t range for duplicate determinations, and the experiment has been re- peated three times with similar results. The incorporation of PPIphos- phate into acetyl-coA carboxylase is shown in the autoradiograph inset.

a 60-kDa protein band present in a placental scatter factor preparation (i.e. hepatic growth factor) (21), had the sequence IETYHNLLEGGQEDFE that aligns to Snfl(359-375) shown in Fig. 2. Furthermore, a polymerase chain reaction fragment generated during amplification of the Drosophila P-adrenergic receptor kinase that had poor homology to the receptor kinase family contained the sequence DLKPKNLLLDHNMHAKIAD- FGLSNMMLDGEFLR, which aligns to Snfl(177-209) (17) and may represent a part of the Drosophila homolog of Snfl.

The partial AMP-activated protein kinase sequence obtained encompassed the GXGXYG motif of the ATP-binding domain and extended beyond the end of the catalytic domain to the HEWFK sequence adjacent to the I a-helix in the CAMP-de- pendent protein kinase structure (22). We have not tried to model the Snfl protein kinase sequence or the AMP-activated protein kinase sequence on the three-dimensional structure of the CAMP-dependent protein kinase (23), but inspection of the sequence reveals that both protein kinases contain Glu resi- dues corresponding to Glu-127 and Glu-170, involved in recog- nition of Arg residues at the P-2 and P-3 positions, but do not contain an acidic residue equivalent to Glu-230. These results suggest the possibility that at least one basic residue in prox- imity to the phosphorylation site may be important for sub- strate recognition. Furthermore, they do not contain an equiv- alent residue to Glu-203 and therefore would not be expected to recognize a more distal Arg in the P-6 position. Compared to the CAMP-dependent protein kinase, Snfl and the AMP-activated protein kinase have more polar residues in the binding pockets for the the P+l and P-11 positions that are important for high affinity hydrophobic interactions between the CAMP-depend- ent protein kinase and its inhibitor peptide (24). These differ- ences are consistent with the distinct substrate specificities of the AMP-activated protein kinase and the CAMP-dependent protein kinase apparent from their protein and peptide sub- strate specificities (1).

We found that the porcine AMP-activated protein kinase 63- kDa protein was autophosphorylated following SDS-PAGE, transfer to nitrocellulose, and renaturation. Previously Snfl/4 was shown to autophosphorylated on the 72-kDa protein but not the 36-kDa protein (251, and we confirmed this in the pre- sent study. These results support the view that the 63-kDa protein contains the catalytic domain of the AMP-dependent

protein kinase and agree with the earlier observations (11) that the rat liver AMP-dependent protein kinase has a mass of 63 kDa. Comparison of the Snfl sequence with its plant homologs indicates that the COOH-terminal domain is not as highly conserved as the catalytic core, with the presence of both sub- stitutions and deletions. Peptides derived from the AMP-acti- vated protein kinase falling outside the homologous catalytic core did not align strongly to the COOH-terminal sequence of Snfl.

The tryptic peptide maps of the 40- and 38-kDa proteins were different from each other, as well as those of the 63-kDa protein (data not shown), indicating that neither were fragments of the 63-kDa protein. The 40-kDa protein was more rapidly phos- phorylated than was the 63-kDa kinase domain when the pu- rified preparation of the AMP-activated protein kinase was incubated with [Y-~~PIATP, and the rate of phosphorylation was sensitive to the salt c~ncentration.~ No incorporation of 32P was observed for the 38-kDa protein. The co-purification of the 38- kDa protein with the AMP-activated protein kinase together with the absence of phosphorylation on this protein raises the possibility that it is analogous to Snf4 and forms a heteromeric complex with the 63-kDa catalytic subunit of the AMP-acti- vated protein kinase.

The purified AMP-activated protein kinase inhibited yeast acetyl-coA carboxylase approximately %fold (Fig. 3 A ) as was found earlier with less pure preparations (26), verifying that the isolated porcine enzyme was indeed the counterpart of the previously reported rat liver enzyme (11). The Snfl/4 kinase also phosphorylated and inactivated yeast acetyl-coA carbox- ylase. Fig. 3A shows that both the AMP-activated protein ki- nase and Snfl/4 kinase inactivate yeast acetyl-coA carboxyl- ase. In contrast to the AMP-activated protein kinase, Snfl/4 was not stimulated by AMP when either peptide substrate or acetyl-coA carboxylase were used as substrate. Fig. 3B illus- trates the time-dependent phosphorylation and inactivation of yeast acetyl-coA carboxylase by SnflM kinase complex. While this is the only instance we know of where Snfll4 has been shown to alter the function of a protein, it is not yet known whether acetyl-coA carboxylase is regulated by Snfl/4 in vivo.

Our results show that both the AMP-activated protein kinase and SnflM kinase share similar specificities for both peptide and protein substrates. The SAMS-peptide was phosphorylated by the Snfll4 kinase with a K,,, of 38 8 and a V,,, of 4.9 * 0.6 pmol.min".mg" (mean of three independent determina- tions), compared to the AMP-activated protein kinase K,,, of 32

and a V,, of 10.8 pmol.rnin-l*mg-' measured in the pres-

3 D. Stapleton, K. I. Mitchelhill, G. Gao, C. House, B. Michell, F. Katsis, L. A. Witters, and B. E. Kemp, unpublished data.

2364 AMP-activated Protein Kinase Snfl Homolog

ence ofAMP. The V,, obtained with SnfU4 kinase is similar to the AMP-activated protein kinase V,, measured in the ab- sence of AMP (approximately 5.2 * 0.4 pmol.min-'.mg').

Few natural substrates have been identified for Snfl kinase. A 96-kDa protein, SIP1 (271, was identified as a putative sub- strate using the two-hybrid screening method of transcrip- tional activation with Snfl, but its function is not known. Pre- viously, Snfl had been recognized for its important role in the transcriptional control of invertase in adapting yeast to growth on non-glucose sugars. The sequence and functional relation- ships between Snfl and the AMP-activated protein kinase raise a vista of questions concerning the possible role of Snfl and its plant homolog kinases in lipid metabolic regulation, as well as the possibility that the AMP-activated protein kinase is in- volved in the control of transcription and translation of proteins in mammalian cells.

REFERENCES 1. Hardie, D. G . (1992) Biochim. Biophys. Acta 1123,231-238 2. Ferrer, A., and Hegardt, F. G. (1984) Arch. Biochem. Biophys. 230,227-237 3. Ferrer,A., Caelles. C., Massot, N., and Hegardt, F. G. (1985)Biochem. Bwphys.

4. Carling, D., Zammit, V. A,, and Hardie, D. G . (1987)FEBS Lett. 223,217-222 5. Davies. S . P., Sim, A. T. R., and Hardie, D. G. (1990) Eur. J . Biochem. 187,

6. Carling, D., and Hardie, D. G . (1989) Biochim. Biophys. Acta 1012,81436

8. Thampy, K. G., and Wakil, S . J. (1985) J. Biol. Chem. 260,63164323 7. Yeh, L. A,, Lee, K. W., and Kim, K. H. (1980) J. Biol. Chem. 265,2308-2314

9. Garton, A. J., Campbell, D. G., Carling, D., Hardie, D. G., Colbran, R. J., and

Res. Commun. 132,497304

183-190

10. Witters, L. A,, and Kemp, B. E. (1992) J. Biol. Chem. 267,2864-2867 11. Carling, D., Clarke, P. R., Zammit, V. A,, and Hardie, D. G. (1989) Eur. J.

12. Woodgett, J. R. (1991) Methods Enzymol. 200, 169-178 13. Davies, S . P., Carling, D., and Hardie, D. G. (1989) Eur. J . Biochem. 186,

14. Pearson, R. B., Mitehelhill, K. I., and Kemp, B. E. (1993) in Protein Phosphory- 123-128

Zation: A Practical Approach (Hardie, G . D., ed) pp. 265-291, Oxford Uni- versity Press

15. Tetaz, T., Bozas, E., Kanellos, J., Walker, I., and Smith, I. (1993) in lkchniques in Protein Chemistry (Hogue Angeletti, R., ed) Vol. IV, pp. 389-397, Aca- demic Press, New York

Yeaman, S . J. (1989) Eur. J . Biochem. 179, 249-254

Biochem. I S , 129-136

16. Celenza, J. L., and Carlson, M. (1986) Science 233, 1175-1180 17. Alderson, A,, Sabelli, P. A,, Dickinaon, J. R., Cole, D., Richardson, M., Keis, M.,

Shewry, P. R., and Halford, N. G . (1991) Proc. Natl. Acad. Sci. U. S . A 88. 18. LeGuen, L., Thomas, M., Bianch, M., Halford, N. G., and Kreis, M. (1992) Gene

860243605

19. Halford, N. G., Vincente Carbdosa. J., Sabelli, P. A., Shewry, P. R., Hannappel, (Amst. ) 120.249-254

U., and Kreis, M. (1992) R a n t J. 2,791-797 20. Celenza, 3. L., Eng, F. J., and Carlson, M. (1989) Mol. c e l l . BioZ. 9,5045-5054 21. Rosen, E. M., Meromsky, L., Romero, R., Setter, E., and Goldberg, I. (1990)

22. Knighton, D. R., Zheng, J., Ten Eyck, L. F., Ashford, V. A., Xuong, N.-H., Taylor,

23. Knighton, D. R., Pearson, R. B., Sowadski, J. M., Means, A. R.. Ten Eyck. L. F.,

24. Knighton, D. R., Zheng, J., Ten-Eyck, L. F., Ashford, V. A., Xuong, N.-H., Taylor,

25. Celenza, J. L., and Carlson, M. (1991) Methods Enzymol. 200,423-430 26. Witters, L. A,, and Watts, T. D. (1990) Biochern. Biophys. Res. Commun. 169,

27. Yang, X., Albert Hubbard, E. J., and Carlson, M. (1992) Science 267,68@482

Biochem. Biophys. Res. Commun. 188,1082-1088

S . S . , and Sowadski, J. M. (1991) Science 253,407414

Taylor, S . S. , and Kemp, B. E. (1992) Science 268,130-135

S. S., and Sowadski, J. M. (1991) Science 253,414-420

369-376