Cell, Vol. 75, 187-197, October 8, 1993, Copyright © 1993 by Cell Press

SREBP-1, a Basic-Helix-Loop-Helix-Leucine Zipper Protein That Controls Transcription of the Low Density Lipoprotein Receptor Gene Chieko Yokoyama,* Xiaodong Wang,* Michael R. Briggs,*$ Arie Admon,t Jian Wu,* Xianxin Hua,* Joseph L. Goldatein,* and Michael S. Brown* *Department of Molecular Genetics University of Texas Southwestern Medical Center Dallas, Texas 75235 tHoward Hughes Medical Institute Department of Molecular and Cell Biology University of California, Berkeley Berkeley, California 94720

Summary

Sterol regulatory element I (SRE-1), a dscamer (5"ATC- ACCCCAC-3~ flanking the low density llpoprotein (LDL) receptor gene, activates transcription In sterol-deplated cells and is silenced by sterols. We report the cDNA cloning of human SREBP-1, a protein that binds SRE-1, activates transcription, and thereby mediates the final regulatory step in LDL metabolism. SREBP-1 contains a basic-helix-loop-helix-leucine zipper (bHLH-ZIP) motif, but It differs from other bHLH-ZIP proteins in its larger size (1147 amino acids) and target sequence. Instead of an inverted repeat (CANNTG), the target for all known bHLH-ZIP proteins, SRE-1 contains a direct repeat of CAC. Overexpression of SREBP-1 activates transcription of reporter genes containing SRE-1 in the absence (15-fold) and presence (90-fold) of sterols, abolishing sterol regulation. We suggest that SREBP-1 is regulated by an unknown factor that is overwhelmed when SREBP-1 is overexpressed. Understanding the regulation of SREBP-1 may be crucial for understand- ing the control of plasma cholesterol in humans.

Introduction

The low density lipoprotein (LDL) receptor mediates the endocytic uptake of cholesterol-carrying lipoproteins, thereby controlling cholesterol levels in cells and plasma. Feedback suppression of the LDL receptor gene is pro- posed to account for the high plasma cholesterol levels in industrialized populations (Brown and Goldstein, 1986). Transcription of the LDL receptor gene is controlled by a 10 bp sequence in the 5' flanking region, designated sterol regulatory element 1 (SRE-1) (Smith et al., 1990). When cellular sterol stores are depleted, this element is acti- vated, the gene is transcribed, and the cellular uptake of LDL increases. When cellular sterols rise, the activity of this element is abolished, and transcription is reduced

SPresent address: Ligand Pharmaceuticals, Incorporated, La Jolla, California 92121.

(Goldstein and Brown, 1990). This classic feedback con- trol mechanism allows cells to supply themselves with cho- lesterol derived from LDL while avoiding cholesterol over- accumulation. Although this mechanism is efficient for individual cells, it can have negative consequences at the whole-body level. Ingestion of a high cholesterol diet re- presses LDL receptors in the liver, thereby lowering the rate at which LDL is removed from the circulation. The resultant high LDL levels lead ultimately to atherosclerosis and myocardial infarction (Brown and Goldstein, 1986).

Elucidation of the molecular mechanism for the feed- back regulation of the LDL receptor gene is important for medical as well as scientific reasons. Recently, we purified a set of SRE-binding proteins (SREBPs) that bind to the SRE-1 sequence in a fashion that correlates with transcrip- tional activity. Nine point mutations in SRE-1 that abolish transcriptional activity also abrogate binding of SREBPs in vitro. In contrast, seven point mutations in SRE-1 and surrounding sequences that preserve transcriptional ac- tivity likewise preserve the ability to bind SREBPs in vitro (Briggs et al., 1993; Wang et al., 1993). On the basis of these data, we concluded that SREBPs mediate the final regulatory step that controls LDL receptor gene expression.

The SREBPs are present at extremely low concentra- tions in high salt nuclear extracts of HeLa cells and other animal cells. Their activity could not be detected by DNA binding assays until the proteins had been purified several hundred-fold by chromatographic techniques. Once the proteins were identified, complete purification (>38,000- fold) was attained through DNA affinity chromatography (Wang et al., 1993). When subjected to electrophoresis in SDS-polyacrylamide gels, purified preparations of SREBP migrated as several closely spaced bands clustered in the range of 59-68 kd. All of these proteins could be cross- linked to the SRE-1 sequence by ultraviolet light, sug- gesting that all were able to bind specifically to DNA either independently or through formation of a protein complex (Wang et al., 1993).

In the current paper, we report the isolation and func- tional analysis of a cDNA encoding one of the SREBPs (SREBP-1), which is identified as a member of the basic- helix-loop-helix-leucine zipper (bHLH-ZIP) family of DNA- binding proteins (Murr~ and Baltimore, 1992; Pabo and Sauer, 1992; Ferr~-D'Amar~ et al., 1993). When ex- pressed in animal cells by transfection, SREBP-1 activates transcription of a reporter gene driven by the SRE-1 se- quence. We postulate that SREBP-1 is one member of a set of SREBPs that mediates sterol-regulated transcrip- tion of the LDL receptor gene and hence plays a crucial role in the control of plasma cholesterol levels in normal and diseased states. The availability of a cDNA for SREBP-1 provides a long-sought tool for elucidaton of the signal transduction pathway by which cellular cholesterol derived from the receptor-mediated endocytosis of LDL regulates gene expression.

Cell 188

Table 1. Sequences of Tryptic Peptides from Human SREBP

Amino Acid Position Peptide Amino Acid Sequence in cDNA Sequence

1 ^I,KPEQRPSLHSR 469-480 2 AAG LSPLVSGTTVQTG PLPTLVS/m 264-286 3 FLQHSNQKLKQENLSL 372-387

4 SFTQVTLPSFSPSAASPQA Not found 5 LAAGSKAPASAQSRGEX/,R 308-325 6 SSINDKIIELK 337-347

Sequences were obtained from high pressure liquid chromatography-purified tryptic peptidas isolated from purified SREBP. Each peptide represents a pure species from a single high pressure liquid chromatography peak. Amino acid sequences corresponding to 5 of the 6 peptides were found in the cDNAs for SREBP-1. The sequence corresponding to peptide 4 was not found in the cDNA (Figure 1). Three positions of ambiguity in the amino acid sequences are noted, and the upper residue denotes the amino acid found in the cDNA-derived sequence. The solid and dotted underlining for peptide 3 corresponds to amino acids used to derive oligonucleotide primers 1 and 2, respectively, for PCRs in cDNA cloning, as described in Experimental Procedures.

A Plasmid Constructs

E¢e~7]ll AA.~}' AAA

pSP-~P-~ • s ' ~ i~////~. ............ .x pcYs

pCY2I

• E¢o47[I[

pSREBP-Ib 5' I¢¢¢~ ~ PoIy(A) pCY5

pCY22

pSREBP-I¢ 5' Im I ~ PoIy(A)

1ooo 2ooo ~ooo 4000 ;=~ I I t I

B 5' Sequences

pelt5

$' Sequences pCY21

Figure 1. SREBP-1 Plasmid Constructs (A) cDNAs encoding SREBP-1 sequences (pCY5, pCY21, and pCY22) are denoted by solid lines. To generate plasmid constructs pSREBP-la and pSREBP-lb, we used the Eco47111 site at nucleotide position 1167 of pSREBP-la to fuse the 5' end of pCY5 to the 3' end of either pCY21 or pCY22, as indicated. Plasmid pSREBP-lc was generated as the Sail fragment of pCY22. Identical sequences in all three plasmid con- structs are denoted by the open boxes. The 5' sequences of pCY22 and pCY5 differ in the regions denoted by the closed and hatched boxes shown in the plasmid constructs (see [B]). The 3' sequences

R e s u l t s

SREBP was purif ied from pooled extracts of HeLa cell nuclei as previously descr ibed (Wang et al., 1993). The final preparation showed a cluster of bands between 59 and 68 kd on silver-stained SDS-polyacry lamide gels. The entire fraction from the final DNA affinity chromatography step was subjected to tryptic digestion, and peptides were fractionated by high pressure liquid chromatography and sequenced as described in Experimental Procedures. Six unambiguous sequences were obtai ned (Table 1). Degen- erate ol igonucleot ides encoding a portion of peptide 3 were used to pr ime polymerase chain reactions (PCRs) with a HeLa cell cDNA library as template. This yielded a unique DNA sequence that was used to probe two HeLa cDNA libraries. Three overlapping partial cDNAs, desig- nated pCY5, pCY21, and pCY22, were characterized. These three cDNAs had the identical sequence over most of the coding region (indicated by the open box in Figure 1A), but they differed at the 5' and 3' ends, suggesting the possibil ity of alternative splicing or cloning artifact (Fig- ure 1A).

Figure 1B shows the divergent sequences. The 5' end of pCY5 predicts a 5' untranslated region of 166 nt that contains an in-frame stop codon (single underl ining in Fig- ure 1 B). The first methionine (double underl ining in Figure 1 B) occurs within a good consensus for translational initia- tion as defined by Kozak (1984). The 5' end of pCY22 is shorter than that of pCY5, and the open reading frame continues to the 5' end. At the 3' end, the sequences of pCY21 and pCY22 are identical up to nucleotide 3268

of pCY22 and pCY21 differ in the regions denoted by the closed and hatched boxes shown in the plasmid constructs; the sequences di- verge at nucleotide position 3269 in pSREBP-la (see IBI). (13) The 5' and 3' regions of the nucleotide and deduced amino acid sequences of pCY5, pCY21, and pCY22 that differ from each other are shown. Numbers for the nucleotide residues (right) and amino acid residues (left) refer to the numbers for pSREBP-la (see Figure 2). An in-frame nonsense codon in the 5' untranslated region of pCY5 is denoted by single underlining. Putative initiator methioninee in pCY5 and pCY22 are denoted by double underlining. Putative polyadenyla- tion signals in pCY21 and pCY22 are boxed.

Sterol Regulatory Element-Binding Protein 1 189

i ~ V ~ P P ~ S ~ I ~ L E O ^ 41 S P F p C L F D p ~ y A C S 81 S S L E A F b S C P ~ A A P ~ n C ~ S S S V ~ S h ~ 161 e V L G Y e S e p G G F T

241 Q ~ H I A D S L TA 2~1 L P T L V S C G T I ~ A T V 321 R ~ E K R T A H N A I E K R ~ 1 x v 4Ol v s 441 L s 4e l ~ . 5 n v Y 5el c v

~ I v o

~oI v T 841 Q **I A V

~I x s lOOl ~ ~o11 T ^ lOal P T

¢ S G C N ~ D V ~ H R C S G S G G S G ~ R L ~ C T ~ V

S G R N V L G T ~ S L V Y G E p V T R P

W L A L R A L C R X C R X ~ G ~ Q O ~ C C . ~ T X T N

c P v u . ~ v ~ L r z . ~ z e x ~ N C S D A A G A P A

v r W L R R D ~ A X K A A R A L L G C A K S I D K A V ~ L F L C P ~ s ~ ~ e a v L A G A S B T R T H

~ G G T D P A P D T S S P G L S P B A T 5 80 P L S P P ~ P P T P L K M Y P S M P A F S P C 120

S P G N T Q p L ~ L p L S p p c P P V 2O0 V T A A P T A P V T T T V T S Q I Q Q V P V L 240

~ s s P ~ s P

S e ~ A V Y L ~ T S H ~ P C A L R V D K

F ~ T R F ~ L G D W S V ~ S T

C V T ~ P N P • s ~ s s s s

R g C p U V E ^ S S C ~ S ~ D L L L V V R T Q R D L S S L R

M t , ¢ L R T A V H K S S L K D 400 • L T P P e S ~ A C S F O S S 440 F E D S K ~ p ~ n . P R , . " ~ 180 A S ~ L G A R C L P S S D T T 520 L L p p V V W L L N e L V L V S~0 R N R K Q A D L D ~ A G D F A Q 6 0 0

A S A R D A A VYH L O H680 ~ C X G ~ X V ~ V & ~ ~ I Y n O

we~ e s = ~ N~ D P ~ Ioo P G S A D G D K E F S A ~ G Y b 8 4 0

LA~ P A M R R V ? L EAI040

vTsAPG~ R V G N L A E , A R T L i1471120

Figure 2. Amino Acid Sequence of Human SREBP-la as Predicted from the cDNA Se- quence Amino acid residues are numbered on the left and right. Amino acid residue 1 is the putative initiator methionine. Five tryptic peptides that were found in purified SREBP (Table 1) are underlined. The amino acid sequences of pCY5 and pCY21 diverge from pCY22 after residues 29 and 1034, respectively (see Figure 1). Both strands of the cDNA were sequenced.

of the composite sequence (Figure 1A), after which they diverge completely, pCY21 encodes an additional 113 amino acids followed by a stop codon and a 3' untranslated region of 544 bp with a potential polyadenylation signal 16 bp from the 3' end of the clone. No poly(A) sequence is present, pCY22 encodes only 37 amino acids followed by a stop codon, 433 bp of the 3' untranslated region, including a poly(A) sequence preceded by a potential poly- adenylation signal.

For functional studies, we produced an expressible com- posite cDNA encompassing the 5' end of pCY5, the invari- ant region, and the 3' end of pCY21 (designated pSREBP- l a in Figure 1A). We chose this combination on the basis of information obtained from two different cDNA clones isolated from two different Chinese hamster ovary cell cDNA libraries. These clones, apparently of full length, contained 5' ends that closely resembled the coding region of pCY5 and 3' ends whose coding regions resembled that of pCY21. Both hamster cDNAs encode a protein of 1133 amino acids that shows 81% identity to the predicted se- quence encoded by the human pSREBP- la (unpublished data). We also prepared human pSREBP-lb (5' end of pCY5 and 3' end of pCY22) and pSREBP-lc (5' and 3' ends of pCY22).

Figure 2 shows the predicted amino acid sequence en- coded by human pSREBP-la. The cDNA encodes a pro- tein of 1147 amino acids, which includes 5 of the 6 peptides whose sequences were obtained (indicated by underlines in Figure 2). The sixth peptide (peptide 4 in Table 1) was not found in the cDNA sequence, and we believe that it is derived from another protein in the purified SREBP prep- aration (see Discussion).

Searches of the GenBank, PIR, and SwissProt data banks revealed only one region of SREBP-1 that shows a close match with other proteins. Amino acid residues 324-394 of SREBP-1 conform to the consensus sequence for the bHLH-ZIP class of DNA-binding proteins (Figure 3A). In the bHLH region, the closest overall match to SREBP-1 is a transcription factor designated TFE3 that activates the immunoglobulin p.E3 motif (Beckmann et al., 1990). The sequence identity within this region was 44%. Sequence identity with other bHLH proteins was in the range of 33°/0. The putative ZIP region of pSREBP-1 con- sists of one alanine, three leucines, and one serine spaced at intervals of seven residues (Figur e 3B). This sequence begins within the second helix of the HLH motif, a feature that is also observed in several other transcription factors,

including AP-4, CBF1, TFE3, and TFEB (Ferrd-D'Amar~ et al., 1993). Helical wheel analysis of this sequence con- firms the segregation of hydrophobic residues on one face, suggesting a true ZIP (Landschulz et al., 1988; Pabo and Sauer, 1992).

The protein encoded by pSREBP- la contains several other noteworthy features. At the extreme NH2-terminus, there is a stretch of 42 residues that contains 12 negatively charged amino acids and no positively charged amino acids (the first positively charged residue occurs at posi- tion 108). The acidic region is predicted to be largely a helical as determined by the method of Gamier et al. (1978). The acidic region (residues 1-42) is followed by a serine/proline-rich region (28% proline and 18% serine) extending from residues 61-178. Following the bHLH-ZlP

Ac,d,C

1147

- C O O H

A Hellx-Loop-Helix Domain

Max ~S D~PS~-~- ~ H C-Myc NTE EUK~ *PEL~ k3/TFE 3 ~Z-S/E47

E~.2 L y l - 1

. . . . . . . . . |-1-1"--F-'T . . . . . . l - - | - - r - ' f " - - I - - BaBie Helix i Helix 2

B Leucine Zipper Region

Helix

Figure 3. Domain Structure of SREBP-Ia

A schematic representation of SREBP-la is shown by the horizontal bar, with the numbers corresponding to the amino acid residues in Figure 2. The HLH region is denoted by the closed box, the ZIP region by the rippled box, and two serine/proline-rich regions by the open boxes. The NH=-terminal acidic region is also indicated. (/%) The amino acid sequence of the HLH region of SREBP-la is com- pared with that of other HLH family members (Beckmann et el., 1990; Prendergast et al., 1991). Residues identical to the consensus se- quence are boxed in black. The boundaries of the basic, helix 1, and helix 2 regions are indicated below the consensus sequence. Asterisks denote amino acids in Max that contact specific nucleotide bases in its DNA recognition sequence, as determined by X-ray crystallography (Ferr6-D'Amar~ et al., 1993; see Figure 9). (B) The amino acid sequence of the putative ZIP region of SREBP-la is shown,

Cell 190

SREBP-1

GAPDH

A

o

kb 7 . 5 -

4 . 4 -

2.4--

1 . 4 -

< -r"

B 1 2 3 4 5 6

205 - o?, o 1 1 6 -

x 8 0 -

5 0 -

= = _ ~ - > , ~

Figure 4. Expression of Human SREBP-1 (A) Tissue distribution of mRNA for SREBP-I. pSREBP-lc was hybridized to poly(A)* RNA from the indicated tissue (2 I~gllane in the left and right panels and 2.5 p.g/lane in the middle panel) as described in Experimental Proce- dures. The filters were exposed to Kodak XAR-5 film with an intensifying screen at -70°C for 16 hr. The same filters were subse- quently hybridized with rat glyceraldehyde-3- phosphate dehydrogenase (GAPDH) and ex- posed to film for 4 hr. The samples in the left and middle panels are from human adult tis- sues, and those in the right panel are from hu- man fetal tissues. (B) Immunoblot analysis of SREBP-1 overex- pressed in transfected 293 cells. Samples (see below) were subjected to electrophoresis on an SDS-polyacrylamide (7.5%) gel and trans- ferred to nitrocellulose filters. The filters were incubated with 5 p.glml rabbit anti-SREBP-1 im- munoglobulin G, followed by incubation with anti-rabbit immunoglobulin G conjugated to horseradish peroxidase using the ECL West- ern kit (Amersham). The filter was exposed to Kodak XAR-5 film for 10 s at room temperature. The position of migration of prestained molecu- lar size markers is indicated. Lanes 1-5 were loaded with 5 p.g of protein of total extract from 293 cells trensfected with 3 p.g of the following plasmids: lane 1, pCMV7; lanes 2 and 5, pSREBP-la; lane 3, pSREBP-lb; lane 4, pSREBP-lc. Lane 6 was loaded with 0.1 p.g of partially purified SREBP from HeLa cells (step 6 in Wang et al., 1993).

region, there is a stretch of 36 amino acids (residues 427- 462) that contains 33% serine, 19% glycine, and 19% proline (71% combined total).

Northern blots of human poly(A) ÷ RNA indicate that pSREBP-1 is expressed in a wide variety of tissues and is most abundant in liver and adrenal gland (Figure 4A, left and middle panels), the two tissues that express the highest concentration of LDL receptors (Brown and Goldstein, 1986). Among fetal tissues (Figure 4A, right panel), the liver and lung showed high expression (the fetal adrenal was not tested). In all tissues there was a predominant messenger RNA (mRNA) of - 4000 nt. The diffuse nature of the band raises the possibility of alter- nately spliced variants of similar size.

A polyclonal antibody was raised against amino acid residues 470-479 of SREBP-la (part of peptide 1 in Ta- ble 1), and this reacted with purified SREBP on immuno- blots (Figure 4B). The antibody also reacted with SREBP-1 that was produced in human embryonic kidney 293 cells after transfection with an expression vector containing pSREBP-la (Figure 4B, lanes 2 and 5), pSREBP-lb (lane 3), and pSREBP-lc (lane 4). As expected, pSREBP-1 b and pSREBP-lc produced proteins that were slightly smaller than the one encoded by pSREBP-ia, owing to shorter coding regions at the 5' and 3' ends (see Figure 1). Control cells transfected with the vector alone did not contain suffi- cient endogenous SREBP-1 to be recognized by this anti- body (Figure 4B, lane 1). All of the cDNA-encoded versions of SREBP-1 were considerably larger than the protein that

was purified from HeLa cells (Figure 4B, lane 6), indicating that the purified protein was proteolyzed. We do not know whether this proteolysis occurred by a physiologic process within the cell or whether it occurred during the purification procedure.

To test the DNA binding activity of the protein encoded by pSREBP-1, we performed gel retardation assays (Fig- ure 5) with three ~P-labeled oligonucleotide probes: probe H, which contains the wild-type human SRE-1 sequence; probe M, which is transcriptionally active and contains a T substituted for a C at position 10 in the SRE-1 sequence (corresponding to the wild-type mouse sequence); and probe *, which contains an A for C at position 10 and is transcriptionally inactive (Briggs et al., 1993; Wang et al., 1993). The partially purified SREBP from HeLa cells bound to probes H and M, but not * (Figure 5, lanes 7-9). Full- length SREBP-lc translated in a reticulocyte lysate sys- tem bound probes H and M, but not probe * (Figure 5, lanes 4-6). The retarded band migrated more slowly than the band produced by the purified HeLa SREBP (Figure 5, lanes 7-9), which is the expected result based on the larger size of the in vitro translated product. The reticulo- cyte lysate mixture without pSREBP-lc gave no retarded band (Figure 5, lanes 1-3). In the same assay, we tested the binding activity of a small fragment containing the bHLH-ZIP domain of SREBP-1 (amino acids 301-407) that was produced by expression of a fragment of pSREBP-lc in Escherichia coli. This peptide also bound probes H and M, but not * (Figure 5, lanes 10-12), indicat-

Sterol Regulatory Element-Binding Protein 1 191

In Vitro Translatsd HeLa Recombinant

Protein -pSREBP-1 +pSREBP-1 SREBP bHLH-Zip Probe H M * H M * H M * H M * Lane 1 2 3 4 5 6 7 8 9 10 1112

< P - a

< - - b

~ - - C ~- - -X

Figure 5. Gel Mobility Shift Assays of In Vitro Translated SREBP-lc, DNA Affinity-Purified SREBP from HeLa Cells, and Recombinant bHLH-ZlP Domain of SREBP-1 Aliquots of the indicated proteins (see below) were incubated in the standard gel shift assay (final volume of 20 p.I) for 20 min at room temperature with the indicated :=P-lapeled, PCR-derived DNA probe (94 bp). Each probe ( -4 x 104 cpm per reaction) contained two tan- dem copies of repeats 2 and 3 with the following versions of repeat 2: probe H, wild-type human SRE-1; probe M, wild-type mouse SRE-1, which differs by 1 bp (C-T) from the human sequence at position 10 in SRE-1 (see Figure 6); and probe *, mutant version of human SRE-1 containing a substitution of A for C at the same position, which abol- ishes transcriptional activity (Briggs et al., 1993; Wang et al., 1993). The following proteins were used in the assays: lanes 1-3, 3 p.I of in vitro transcription-translation mixture lacking plssmid pSREBP-lc; lanes 4-6, 3 Id of in vitro transcription-translation mixture with pSREBP-lc; lanes 7-9, 0.1 I~g of partially purified SREBP from HeLa cells (step 6 in Wang et al., 1993); lanes 10-12, 0.1 p.g of purified recombinant bHLH-ZlP domain of SREBP-1. Arrow a denotes SRE-1 bound to full-length SREBP-lc; arrow b denotes SRE-1 bound to puri- fied SREBP; arrow c denotes SRE-1 bound to the recombinant bHLH- ZIP domain; arrow x denotes a contaminating protein that binds the mutant probe as well as the two wild-type probes. After electrophoresis, lanes 1-9 were exposed to Kodak XAR-5 film for 16 hr at -800C with an intensifying screen; lanes 10-12 were exposed for 2 hr with an intensifying screen at -80°C.

ing that the bHLH-ZIP domain contains the specific DNA binding activity. As expected, this retarded band showed fast mobility, owing to the small size of the bHLH-ZIP fragment (107 amino acids).

We previously described a series of 16 point mutations in the SRE-1 and flanking sequences that give clear-cut results with regard to transcriptional activity (Briggs et al., 1993). Seven of the mutants are induced normally by sterol depletion after transfection into cultured cells, and nine are induced poorly, if at all. Purified SREBP from HeLa cells bound only to the seven mutants that were efficiently induced (Wang et al., 1993). Figure 6 shows that the same pattern of binding was found for the recombinant truncated bHLH-ZIP domain of SREBP-1 that was produced in E. coli. The bHLH-ZIP domain bound to the wild-type SRE-1

SRE-1 0~ 1 2 3 4 5 6 7 8 9 10 Q.

AAIA T C A C C C c A C l T GC

• t t t t g a t a g g a t a g t a

Figure 6. BindingoftheRecombinantbHLH-ZIPDomainofSREBP-1 to Wild-Type and Mutant Forms of Repeat 2 Aliquots of the purified recombinant bHLH-ZIP domain of SREBP-1 (0.2 p.g) were incubated in the standard gel shift assay (final volume of 20 p.I) for 20 min at room temperature with the indicated wild-type or mutant a=P-labeled, PCR-derived DNA probe (45 bp). Each a=p probe (4 x 104cpm per reaction) contained one copy of repeats 2 and 3 with the indicated point mutation in repeat 2 (Wang et aL, 1993). After electrophoresis, the gel was exposed to Kodak XAR-5 film for 2 hr at -80°C with an intensifying screen. The SRE-1 sequence contained within repeat 2 is boxed.

sequence and to all seven of the mutants that retained transcriptional induction, including the highly diagnostic sequence that contains a G in place of the C at position 6 of SRE-1 (Wang et al., 1993). On the other hand, the bHLH-ZIP domain protein failed to bind to any of the oligo- nucleotides that had point mutations at any of the other nine positions in the SRE-1 sequence. This finding indi- cates that the bHLH-ZIP domain, which contains only 107 amino acids, contains the structural information sufficient for specific DNA binding.

To test the functional activity of SREBP-1 in vivo, we employed a reporter plasmid (plasmid K) that contains two copies of the SRE-1 sequence upstream of a gene encod- ing bacterial chloramphenicol acetyltransferase (CAT). In the native LDL receptor promoter, the 10 bp SRE-1 se- quence is contained within a 16 bp element, designated repeat 2, which is followed by a distantly related sequence, designated repeat 3. Although repeat 3 binds transcription factor Spl , it is insufficient to give high level transcription on its own, and it requires a contribution from SRE-1 within repeat 2 (Smith et al., 1990; Briggs et al., 1993). The latter element is active only when cells are deprived of sterols. Plasmid K contains two copies of the sequence of repeats 2 and 3 in tandem followed by a TATA box from adenovirus E lb (Briggs et al., 1993). When transfected into ceils, this construct gives rise to high levels of CAT enzyme activity in the absence of sterols, and activity is suppressed by sterols, owing to removal of the positive contribution of the SRE-1 sequences (Briggs et al., 1993).

Cell 192

C

O.

C E O

3;

O

2.5 -- A Simian CV-1 Cells

2.0

1.!

1.,

+ Sterols 0.5

I I I I 50 :- B Human 293 Cells

2O No Sterols

10 5.1-fold / -

01 S I [ I [ 0 0.25 0.5 0.75 1.0

pSREBP-la (pg)

Figure 7. Effect of SREBP-la on Expression of CAT Activity under Control of Synthetic LDL Receptor Promoter Elements in Transfected Simian CV-1 Cells and Human 293 Cells (A) Simian CV-1 cells. (B) Human 293 cells. Cells were transiently cotransfected with a synthetic LDL receptor promoter-CAT gene driven by two tandem copies of repeats 2 and 3 (plasmid K) (10 pg in [A] and 1 p.g in [B]) and the indicated amounts of pSREBP-la, as described in Experimental Procedures. After incubation for 48 hr (A) or 40 hr (B) in the absence (closed circles) or presence (open circles) of 10 pg/ml cholesterol plus 1 pglm125-hydroxycholesterol, triplicate dishes of cells were harvested for measurement of CAT activity. The fold value in (B) refers to the difference in CAT activity in cells incubated with or without stercls in the absence of pSREBP-la.

We cotransfected plasmid K into simian CV-1 cells (Fig- ure 7A) and human 293 cells (Figure 7B) together with varying amounts of a plasmid encoding pSREBP-la. In the absence of pSREBP-la, the CV-1 cells produced 7-fold more CAT activity in the absence as compared with the presence of sterols. Increasing amounts of pSREBP-la increased the amount of CAT activity under both condi- tions of incubation, but the relative effect in the presence of sterols was more profound. As a result, at high levels of SREBP, CAT activity was the same in the absence and presence of sterols. Similar results were obtained with the 293 cells. In the absence of SREBP, these cells also showed a 5-fold higher CAT activity when sterols were depleted. When pSREBP-la was present, CAT activity was increased 15-fold in the absence of sterols. Again, the relative stimulation was greatest in the presence of sterols (90-fold) so that at high levels of pSREBP-la, CAT activity was actually higher in the presence of sterols than in their absence.

To confirm that the transcriptional stimulation by pSREBP-la was dependent on SRE-1, we repeated the cotransfection experiment with reporter plasmids that con- tain inactivating mutations in both copies of SRE-1 (Figure 8, top panels). One of these plasmids (Q) contains an A in place of C at position 3 of SRE-1 (the 5' C of the first CAC trimer), and the other plasmid (X) contains an A in place of C at position 10 (the 3' C of the second CAC trimer) (see Figure 6). Transcription from the plasmid con- taining the wild-type SRE-1 sequence (plasmid K) was

~_ A Plasmld K - B Pla~crHd Q - C Flasmkl X - "-- Mutant SRE-1 Mutant SRE-I S wt SRE*I (C = ~ A ) ( C I °~A )

c

F-

~o m mmq pCMW pSREBP-Ia I~MV7 p6REBP-la pCMV7 pSREBP-Ia

~ " 2 5 - D Pl~mldK "E HMG CoASynlhas~ -F HMGCOA Reduclase r "~ (wt SRE-1) (-527 to +39) (-Z77 Io +231) 0'71"G 1-11~47R~) ~

. 0.5

-,~ 15 0.4

10 013

0.2

.~ 0.1

o 0 0 ~CMV7 pSRESP-la pCMV7 pSREBP-Ia pCMV7 pSREBP-la pCMV7 pSREBP-la

Figure 8. Effect of SREBP-la on Expression of CAT Activity under Control of Synthetic Promoter Elements Containing Wild-Type and Mutant SRE-1 Sequences and Native Promoter Elements from Three SteroI-Regulated Genes in Transfected 293 Cells (Top panels) Wild-type and mutant SRE-1 sequences. (Bottom panels) Native promoter elements from three sterol-regulated genes. 293 cells were transiently cotransfected with 0.2 pg of either pCMV or pSREBP- la, as indicated, and 1 pg of a reporter CAT gene driven by one of the following promoter elements: synthetic wild-type SRE-1 sequence containing two tandem copies of repeats 2 and 3 (plasmid K) (A and D); mutant SRE-1 sequence containing two tandem copies of repeats 2 and 3 with a point mutation in repeat 2 (plasmid Q) (B); mutant SRE-1 sequence containing two tandem copies of repeats 2 and 3 with a point mutation in repeat 2 (plasmid X) (C); native HMG CoA synthase, nucleotides -527 to +39 (plasmid J) (E); native HMG CoA reductase, -277 to +231 (pRedCAT-1) (F); or native LDL receptor, -1471 to +36 (p1471) (G). After incubation for 40 hr in the absence (closed bars) or presence (hatched bars) of 10 p.g/ml cholesterol plus 1 pg/ml 25- hydroxycholesterol, triplicate dishes of cells were harvested for mea- surement of CAT activity. The data in the top and bottom panels are from two separate transfection experiments.

stimulated 10-fold by cotransfection with pSREBP-la, and sterol suppression was nearly eliminated (Figure 8A). In contrast, transcription driven by plasmid Q was stimulated only slightly by pSREBP-la (Figure 8B), and plasmid X was not stimulated at all (Figure 8C).

We next tested the ability of pSREBP-la to stimulate transcription of reporter constructs containing the native promoters of three sterol-regulated genes (Figure 8, bot- tom panels). As a positive control in this experiment, we cotransfected pSREBP-la together with plasmid K, and again we observed a marked stimulation of transcription and an abolition of sterol suppression (Figure 8D). A simi- lar effect was observed when transcription of the CAT gene was driven by a fragment of the native LDL receptor pro- moter (Figure 8G), which contains repeats 2 and 3 as well as repeat 1, another weak Spl-binding site (Dawson et al., 1988). The absolute level of CAT activity was much lower with the native promoter construct than with plasmid K (note the different scales for Figures 8D and 8G). The promoter for 3-hydroxy-3-methylglutaryl coenzyme A

Sterol Regulatory Element-Binding Protein 1 193

(HMG CoA) synthase was much more active than the LDL receptor promoter in this assay, and it, too, was stimulated by pSREBP-la with an abolition of sterol regulation (Figure 8E). In contrast, the promoter for HMG CoA reductase was not affected by pSREBP-la (Figure 8F). The promoter for HMG CoA synthase, but not HMG CoA reductase, con- tains a sequence that fits the requirements for binding SREBP-1 (Wang et al., 1993; Osborne et al., 1992).

Discussion

SREBP-1, s Regulator of Cholesterol Levels The current paper reports the cloning and functional ex- pression of SREBP-1, a transcription factor of the bHLH- ZIP family that activates transcription of the LDL receptor gene through binding to SRE-1. The following lines of evi- dence indicate that SREBP-1 is a crucial molecule in con- trolling LDL receptor gene expression and is thus a key element in controlling plasma cholesterol levels in hu- mans. First, when overproduced in animal cells by trans- fection, SREBP-1 activates transcription of reporter genes with artificial promoters that contain multiple copies of SRE-1 as well as natural promoters containing the 5' flank- ing region of the LDL receptor gene or the HMG CoA syn- thase gene. Second, the requirements for the activity of SREBP-1 correlate with the known requirements for tran- scription of the LDL receptor gene. Third, in vitro the pro- tein produced from pSREBP-1 bound to sequences con- taining SRE-1 as determined by gel mobility shift assays. Fourth, the sequence requirements for binding matched precisely the requirements for transcriptional activation by SRE-1.

SREBP-1 Is a bHLH-ZIP Protein The sequence of SREBP-1 reveals it to be a member of the bHLH-ZIP family of transcription factors. This growing family includes the oncogenic protein Myc and its modula- tors Max, Mad, and Mxil (Ferr~-D'Amar(~ et al., 1993; Zervos et al., 1993; Ayer et al., 1993). It includes a variety of proteins, such as TFE3, that regulate transcription of immunoglobulin genes (Beckmann et al., 1990): AP-4, which activates transcription from the SV40 enhancer (Hu et al., 1990), and certain transcription factors that regulate differentiation in mammals and insects (Murr~ and Balti- more, 1992). The bHLH-ZIP proteins form homodimers and heterodimers through interactions between the paired helices on adjacent monomers. Dimerization is reinforced by the adjacent ZIPs that form coiled coils. Dimerization orients the nearby basic regions of the two monomers so that they can make specific contacts with palindromic sequences in the major groove of the target DNA (Ferr~- D'Amar(~ et al., 1993).

All of the classic bHLH-ZIP proteins recognize palin- dromic sequences containing the so-called E box (CANNTG). These proteins can be divided into two classes, depending on the two central nucleotides of the E box (Dang et al., 1992). Class B proteins recognize CACGTC, whereas class A proteins recognize CAGCTG. Recently, Ferr~-D'Amar(~ et al. (1993) reported the three- dimensional structure of.Max, a class B bH LH-ZIP protein,

in a complex with DNA. The two basic regions form ex- tended a helices that straddle the DNA in a chopstick-like fashion. Three residues from the basic region of a single monomer make specific contacts with one of the CAC tri- mers and its opposing GTG (Figure 9, top panel). These residues are His-28, Glu-32, and Arg-36 (in the numbering scheme of Ferr~-D'Amar~ et al., 1993). The second protein monomer is attached to the first in a symmetrical head-to- head fashion, thereby allowing it to bind to the second CACIGTG sequence in inverted orientation.

Novel Features of SREBP-1 SREBP-1 differs from other bHLH-ZIP proteins in two cru- cial respects. First, SREBP-la, which contains 1147 amino acids, is much larger than the other bHLH-ZIP pro- teins, which contain 160-536 amino acids. The extra length is at the COOH-terminal end of the protein. Second, SREBP-1 does not recognize a classic E box. The SRE-1 target sequence has no dyad symmetry, but instead it con- tains a direct repeat of the target CAC sequence on the same DNA strand separated by two Cs (Figure 9, bottom panel). The two CAC sequences are spaced five residues apart, and thus they occur on opposite sides of the double helix. Binding of SREBP-1 requires both copies of CAC: point mutations (transversions) in either copy abolish bind- ing as well as the transcriptional activity of SRE-1 (Fig- ure 6).

SREBP-1 shares with Max the three amino acids in the basic domain that contact the CAC/GTG sequence. In SREBP-1 these are His-328, Glu-332, and Arg-336 (see asterisks in Figure 3; see Figure 9). It is likely, therefore, that a monomer of SREBP-1 binds to a single CAC/GTG

Max

Dyad Axis

Glu ~ i His 28 Arg~

/ il i s'-C A CiG T G-a' 3'-G T GiC A C-5'

I Arg ~ His 28 Glu 32

Glu~ Glu 332

BR B -I A A 5'-A T C A C c c C A C - 3 '

3'-T A G T G G GG T G-5' I I I L

Arg~ His328 Arg336 His328

Direct Repeat

Figure 9. Schematic Diagram of Contacts between Amino Acids of bHLH-ZIP Proteins and Nucleotide Bases of Their Target DNA (Top panel) Contacts between the Max homodimer and its palindromic recognition sequence as determined by X-ray crystallography (Ferr6- D'Amar~ et al., 1993). (Bottom panel) Possible contacts between the analogous amino acids of SREBP-1 and its direct repeat recognition sequence.

Cell 194

in a fashion similar to that previously described for Max (Figure 9). But the binding of SREBP-1 requires two se- quential CAC sequences as well as the flanking nucleo- tides. How can a second SREBP-1 monomer bind to the second CACIGTG? This would not be possible if two SREBP-1 monomers were joined in a symmetrical head- to-head dimer such as that described for Max. Perhaps two monomers of SREBP-1 bind to DNA independently and in the same orientation. More likely, SREBP-1 forms higher order multimers through nonsymmetrical interac- tions between helices, as postulated by Farmer et al. (1992) for bHLH proteins without ZIPs. Such a multimer might wind around the DNA so as to allow two basic re- gions in the same orientation to attach to two CAC se- quences on the same DNA strand, but on opposite sides of the helix. The two monomers would have to make some- what different contacts with DNA since the two half-sites have a different sequence (ATCAC and CCCAC). More- over, the requirement for the terminal C in the second CAC is not absolute. A transition to a T is tolerated (wild-type mouse SRE-1 sequence M in Figure 5), but a transversion to an A is not (mutant * sequence in Figure 5). Resolution of these issues will require detailed physiochemical char- acterization of SREBP-1-DNA complexes.

The predicted a-helical acidic region at the NH2- terminus of SREBP-la is likely to be a transcriptional acti- vation domain. Acidic regions that activate transcription are found at the NH2-termini of some, but not all, bHLH- ZIP proteins, including Myc and TFE3 (Kato et al., 1990; Beckmann et al., 1990). The proline-rich region adjacent to the acidic region of SREBP-1 may also play a role in transactivation (Mermod et al., 1989).

Alternatively Spliced Forms of SREBP-1 We do not know the significance of the pCY22 cDNA whose coding sequence differs from pCY5 and pCY21 at the NH2-and COOH-termini, respectively (Figure 1). We believe that pCY5 and pCY21 reflect the major form of SREBP-1 because we isolated two full-length cDNAs with ends corresponding to these sequences from two different Chinese hamster cell libraries (see Results). The pCY22 sequence might be produced by alternative splicing, or it may result from a rearrangement during cloning. In experi- ments not shown, we have found that pSREBP-lc, which corresponds to pCY22, binds to the same SRE-1 se- quence as pSREBP-la, and it activates transcription driven by plasmid K in transfected cells. Moreover, tran- scription stimulated by pSREBP-lc, like that driven by pSREBP-la, is not suppressed by sterols. Thus, to date we have found no evidence for a functional difference be- tween pSREBP-la and pSREBP-lc. It should be noted that the NH2-terminus of the SREBP-lc protein lacks some of the acidic amino acids found in SREBP-la. However, the NH2-terminus of SREBP-lc retains an overall acidic character with seven negatively charged residues and no positives up to the residue corresponding to residue 42 of SREBP-la.

Proposed Regulatory Role of SREBP-1 in SteroI-Mediated Regulation Overexpression of SREBP-1 stimulates transcription from

SRE-1, but it abolishes sterol sensitivity. Several explana- tions are possible. First, SREBP-1, like other bHLH-ZIP proteins, may be controlled by homo- and heterooligomeri- zation. It is possible that homooligomers of SREBP-1 are constitutively active, whereas heterooligomers with an- other bHLH-ZlP protein are active only in the absence of sterols. Overexpression of SREBP-1 may force the forma- tion of homooligomers, thereby producing constitutive transcription. In this regard it is interesting that purified SREBP appears as a cluster of proteins of 59-68 kd on SDS-polyacrylamide gels (Wang et al., 1993) and that we failed to find a coding sequence In any of the pSREBP-1 cDNAs for 1 of 6 peptides sequenced from purified SREBP (Table 1). Preliminary evidence indicates that this se- quence comes from a separate protein of the bHLH-ZlP family that copurifies with SREBP (X. H. et al., unpublished data). The functional significance of this second bHLH- ZIP protein, which is encoded by a distinct but related gene, is unknown.

Other mechanisms of regulation in addition to heterooli- gomerization are possible, including sterol-mediated regu- lation of SREBP-1 entry or retention in the nucleus by a process that becomes overwhelmed when SREBP-1 is overexpressed. Regulated nuclear entry or retention may be a property of the long COOH-terminal half of SREBP-1, which does not have a counterpart in other bHLH-ZIP proteins.

Experimental Procedures

General Methods end Materials Standard molecular biology techniques were used (Sambrook et al., 1989). cDNA fragments were subcloned into pBluescript II SK(+) (Stra- tagene) or pCRII vectors (Invitrogen) and sequenced by the dideoxy chain termination method (Sanger et el., 1980) using M13 universal and reverse sequencing primers; TT, T3, and SP6 promoter primers; or specific internal primers. Sequencing reactions were performed on an Applied Biosystems Model 373A DNA sequencer. Selected regions of the cDNA were manually sequenced with a Sequenase kit (U. S. Biochemicals). Randomly primed probes were synthesized using a random primer labeling kit (Pharmacia). Reporter CAT genes (plas- raids K, J, Q, and X [Briggs et al., 1993], p1471 [Smith et el., 1990], and pRedCAT-1 [Osborne et al., 1988]) were described in the indicated references.

Amino Acid Sequence of Human SREBP SREBP was purified from nuclear extracts prepared from 1500 I of HeLa cells by repetitions of a previously described 8-step procedure (Wang et al., 1993). About 6 p.g of the purified protein, consisting of a cluster of bands at 59-68 kd on SDS-polyacrylamide gels, was precipitated with 10% (v/v)trichloroacetic acid, washed with acetone, and dissolved in a buffer containing 0.4 M ammonium bicarbonate, 8 M urea, and 5 mM dithiothreitol at 55°C with vortexing. After treat- ment with 10 mM iodoacetamide, the protein was digested with se- quencing grade trypsin (Boehringer Mannheim) as described (Stone et el., 1989). The resulting peptides were separated by reverse-phase high pressure liquid chromatography using a 0.1% (v/v) trifluoroacetic acid and acetonitrile gradient on a 1 mm x 25 cm RP-300 column (Applied Biosystems). The resolved peptides were collected and se- quenced by automated Edman degradation on an Applied Biosystems Model 477A sequencer using standard chemistry, a microreaction car- tridge, and fast cycles.

cDNA Cloning of Human SREBP.1 An aliquot (10 ng) of purified phage template DNA from a HeLa ;~gtl0 cDNA library (Clontech) was amplified with 300 pmol of a degenerate oligonucleotide (primer 1) from the COOH-terminal end of peptide 3 (Table 1), 5"TT(TIC)TC(TIC)TG(TIC)TTIAG(TIC)'I-r(TIC)TG-3', and 20

Sterol Regulatory Element-Binding Protein 1 195

pmol of a ~,gtl0 vector reverse primer, 5'-GG CTTATGAGTATTTCTTC- CAGGG-3'. The primers were removed by passage over a Qiagen Tip-5 column, and 1120 of the PCR product was subjected to a second round of PCR with 160 pmol of a degenerate oligonucleotide (primer 2) from the NH2-terminal end of peptide 3, 5'-TT(TIC)TG(A/G)TTIGA(A/ G)TG(TIC)TG-3', and 20 pmol of the ~,gtl0 vector reverse primer. PCR products were isolated by egarose gel electrophoresis and subcloned into the pCRII vector using the TA Cloning Kit (Invitrogen). The DNA sequence of the resulting 345 bp PCR product encoded peptides 2, 5, and 6 of SREBP (Table 1). The 345 bp PCR product was randomly labeled with [=P]dCTP (4 x 10 s cpmlml) and used to screen HeLa cDNA libraries as described below.

Poly(A)* RNA from HeLa $3 cells cultured in 2.5% (v/v) newborn calf serum (Wang et al., 1993) was isolated with Oligotex-(dT)30 resin (Qiagen) and used to construct a T7-driven ~.EXIox vector-based cDNA library (Novagen). First-strand synthesis was carried out with a combi- nation of oligo(d'l') and random primers. After second-strand synthesis, the double-stranded cDNA was ligated to EcoRI-Notl adapters con- taining a Sail restriction site (GIBCO BRL), inserted into ZEXIox-ECORI arms (Novegen), and packaged with Gigapack II packaging extract (Stratagene).

The HeLa ;Lgtl 0 and HeLa ;~EXIox libraries were screened by hy- bridizing duplicate filters with the 345 bp PCR probe (see above) at 60°C overnight in 3 x SSC containing 0.1% (w/v) SDS, 5 x Denhardt's solution, and 50 ~glml salmon sperm DNA and washed three times with 3 x SSC, 0.1% SDS at room temperature for 5 rain and twice with 0.1 x SSC, 0.1% SDS at 60°C for 20 rain. Of 6 x 105 plaques screened in the ;~EXIox library, seven positive clones were identified and subcloned into the NOd site of pBluescript II SK(+). The two longest clones (pCY21 and pCY22) were characterized. Of the 5 x 105 plaques screened in the ;~gtl0 library, six positive clones were identified and subcloned into the EcoRI site of pBluesoript II SK(+). The clone con- taining the most 5' sequence (pCY5) was characterized.

Construction of Expression Vectors for pSREBP-1 Vectors for expression of human SREBP-1 in cultured animal cells were constructed in pCMV7, a modified version of pCMV5 (Andersson et al., 1989) that contains a hybrid adenovirus-immunoglobulin intron (provided by D. W. Russell, University of Texas Southwestern Medical Center). Plasmid pSREBP-la was constructed by digesting pCY5 with Sail and Eco47111 to obtain the vector and 5' cDNA sequences and by ligating them into the 2989 bp Eco47111-Sall fragment encoding the 3' end of pCY21 (Figure 1). This pBluescript II SK-based vector was digested with BamHI and Hincll, and the cDNA fragment (4183 bp) was cloned into a BamHI-Smal-digested pCMV7 vector, pSREBP-lb was constructed by digesting pCY5 with Xbel and Eco471il (1204 bp) to obtain the 5'cDNA sequence and by ligating it into an XbaI-Eco47111 (5615 bp) fragment encoding the 3' end of pCY22. This cDNA was subcloned into pCMV7 as described for pSREBP-la, pSREBP-lc was constructed by inserting the Sail fragment of pCY22 into the Sail site of pCMV7, and the orientation was confirmed by restriction mapping. All restriction sites used in the above constructions, except Eco47111, are present in the linker sequences.

Blot Hybridization of RNA Poly(A)* RNA from human adrenal gland and Northern blots containing 2 p.g of poly(A)* RNA from multiple human adult and fetal tissues were purchased from Clontech. Poly(A) + RNA from HeLa $3 cells cultured in 2.5% (v/v) newborn calf serum (Wang et al., 1993) was prepared with Oligotex-(dT)30 beads (Qiagen). Blots were hybridized with a ran- dom-primed probe synthesized from the 3.6 kb NOd fragment of pSRESP-lc. Hybridizations were performed in 5 x SSPE containing 50% formamide at 2 x 106 cpmlml for 18 hr at 42°C. After hybridiza- tion, each filter was washed once with 1 x SSC and 0.05% SDS for 30 min at room temperature and twice with 0.1 x SSC and 0.1% SDS for 20 min each at 50°C. After exposure to Kodak XAR-5 film for the indicated time, each filter was reprobed with a =P-labeled cDNA (1.2 kb) for rat glyceraldehyde-3-phosphate dehydrogenase (Chen et al., 1991) at 4 x 106 cpm/ml.

In Vitro Translation of SREBP-1 mRNA pSRESP-lc was cloned into the Sail site of pBluascript SK(+), purified by CsCI bending, and translated in the TNT T7 coupled reticulocyte lysate system (Promega). Each coupled transcription-translation reac-

tion contained 1 I~g of plasmid DNA in a final volume of 50 id and was incubated at 30°C for 2 hr. For radiolabeling, pSREBP-lc was translated in a methionine-free amino acid mixture supplemented with [~S]methionine according to the instructions of the manufacturer. An aliquot (1 p J) of the reaction was added to 20 Id of SDS sample buffer (Laemmli, 1970) and subjected to electrophorasis on a 4%-15% gradi- ent gel. The gel was fixed in 50% (v/v) methanol and 10% (v/v) acetic acid for 20 min, treated with Enlightening enhancer solution (New England Nuclear-DuPont) for 15 min at room temperature, dried, and exposed to Kodak X-ray film at -80°C for 16 hr. For gel mobility shift assays, pSREBP-lc was translated with unlabeled amino acids. An aliquot (3 p.I) of the reaction was assayed directly in the gel mobility shift assay.

Gel Mobility Shift Assays SREBP was incubated with a PCR-generated ~P-labeled DNA probe containing one or two tandem copies of wild-type or mutant elements of repeats 2 and 3 of the LDL receptor promoter as described in the figure legends. Probes were synthesized, and electrophoresis was performed as previously described (Briggs et al., 1993; Wang et al., 1993).

Production of Recombinant bHLH-ZIP Domain of SREBP-1 pSREBP-I(bHLH-ZIP) is a recombinant plasmid that encodes a fusion protein of 130 amino acids. The COOH-termina1107 amino acids com- prise the bHLH-ZIP domain of SREBP-la (residues 301-407). The NH2-termina123 amino acids includes an initiator methionine followed by three functional elements: His-His-His-His-His-His, the site for NF ÷ binding; Asp-Asp-Asp-Asp-Lys, the site for a specific endopeptidase (enterokinase); and Arg-Arg-Ala-Ser-Val, the site for the catalytic do- main of heart muscle kinase (Blanar and Rutter, 1992). pSREBP- l(bH LH-ZIP) was constructed by PCR with pSREBP-t c as a template and the following two primers: 5' oligonucleotide, CGCGGATCCG- ATGACGATGACAAACGTCGTGCATCTGTTGAGAAGCTGCCTAT- CAACCGG (corresponding 5'~3' to the BamHI cloning site, followed by sequences for the enterokinase site, heart muscle kinase site, and SREBP-1); 3' oligonucleotide, CTAATTAAGCTTACTATCCACTGC- CACAGGCCGACAC (corresponding 3'45 , to the Hindlll cloning site and part of the SREBP-1 sequence). The amplified PCR product was digested with BamHI and Hindlll and inserted into pQE-30, a bacterial expression vector that contains sequences encoding an initiator methi- onine and six consecutive histidine residues preceding a polylinker cloning site (Qiagen). The resulting pSREBP-I(bHLH-ZIP) was trans- formed into E. coil host strain M15(pREP4) (Qiagen), grown in 1 I cultures at 37°C, and induced with isopropyl ~-D-thiogalactopyrano- side for 5 hr. The cells were collected by centrifugation; disrupted by stirring at room temperature for 1 hr in 6 M guanidine-HCI, 0.1 M sodium phosphate, 10 mM Tris-chloride at pH 8.0; and centrifuged at 10,000 x g for 15 min at 4°C. The supernatant solution was sub- jected to Ni~-Sepharose affinity chromatography under conditions recommended by the manufacturer (Qiagen). The bHLH-ZIP fusion protein was eluted with 250 mM imidazole, 8 M urea, 0.1 M sodium phosphate, 10 mM Tris-chloride at pH 6.3 and then dialyzed overnight at 4°C into buffer A (25 mM HEPES-KOH at pH 7.5, 12 mM MgCI2, 10% [v/v] glycerol, 0~5 mM phenylmethylsulfonyl fluoride, 0.5 mM dithi- othreitol, 0.1 M KCI). The dialyzed fusion protein ( - 0 . 2 mg/ml) was stored in multiple aliquots at -80°C.

Transfectlon and Reporter CAT Assays For immunoblot analysis, monolayers of human embryonic kidney 293 cells were set up (3 x 105 cells per 60 mm dish) in Dulbecco's modified Eagle's medium containing 10%o fetal bovine serum, 100 Ulml penicil- lin, and 100 p.g/ml streptomycin. After incubation for 24 hr, the cells were transfected with 3 p.g of either the indicated pSREBP-1 or the control vector pCMV7 plus 0.3 p.g of pVA (a plasmid encoding adenovi- rus virus-associated RNA1) (Akusj&rvi et ai., 1987) using the calcium- phosphate method (Sambrook et al., 1989). The cells were washed twice with phosphate-buffered saline 4 hr after transfections and refed with fresh medium. After 40 hr, cells were washed once with ice-cold phosphate-buffered saline and harvested in phosphate-buffered sa- line. The resulting cell pellets were lysed with 0.3 ml of lysis buffer containing 1% (w/v) SDS (Gil et al., 1985). Cell extracts were sheared through a 20-gauge needle and subjected to electrophoresis and im- munoblot analysis.

Cell 196

For in vivo functional analysis, monolayers of human 293 cells and simian CV-1 cells were set up for experiments as described above and by Briggs et al. (1993), respectively. Varying amounts of pSREBP- la (0-1 p.g) were cotransfected with the indicated reporter CAT gene (Briggs at el., 1993). In Figure 7, the total amount of DNA was adjusted to 3.3 p.g (293 cells) or 12 p.g (CV-1 cells) by the addition of pCMV7. After transfaction, the medium was changed to Dulbecco's modified Eagle's medium containing 10% calf lipoprotein-deficient serum in the absence or presence of sterols (10 p.glml cholesterol plus 1 p,g/ml 25-hydroxycholssterol added in 20 Id of ethanol). After incubation for 40-48 hr, cells were harvested, and CAT activity was measured by the xylenes extraction method (Briggs et al., 1993). Protein concentra- tion was determined by the method of Bradford (1976).

Antibodies end Immunoblotting An antibody directed against SREBP-1 was produced by immunizing rabbits with a synthetic peptide (synthesized by S. Stradley, University of Texas Southwestern Medical Center), (C)KPEQRPSLHS, corre- sponding to amino acids 470-479 in SREBP-la (Figure 2). The paptide was coupled to keyhole limpet hemocyanin using m-maleimidobenzoic acid N-hydroxysuccinimide ester (Harlow and Lane, 1963). New Zealand white rabbits were immunized with 500 p.g of coupled peptide in Fraund's complete adjuvant, and immunoglobulin G fractions were prepared (Harlow and Lane, 1988). Immunoblot analysis was per- formed with anti-rabbit horseradish peroxidase-conjugated immuno- globulin G using the Enhanced Chemiluminescence (ECL) Western Blotting Detection System kit (Amersham). Gels were calibrated with prostained high and low range molecular weight markers (Bio-Rad).

Acknowledgments

We thank Robert Tjian and Steve McKnight for helpful suggestions; Jeff Cormier and Amber Luong for DNA sequencing; Gloria Brunschede and Daphne Norsworthy for excellent technical assis- tance; Lavon Sanders and Edith Womack for invaluable help with tissue culture; and our colleagues Mark Evans, Ryuichiro Sato, and Jianxin Yang for sharing unpublished data on the hamster SREBP-1 cDNA sequence. This research is supported by grants from the Na- tional Institutes of Health (NIH) (HL-20948) and the Perot Family Foun- dation. X. W. is the recipient of a postdoctoral fellowship from the Damon Runyon-Walter Winchell Cancer Research Fund (#1156). M. R. B. was the recipient of a postdoctoral fellowship from NIH (5F32HL07863).

Received July 30, 1993; revised August 13, 1993.

References

Akusj~trvi, G., Svensson, C., and Nygard, O. (1987). A mechanism by which adenovirus virus-associated RNAI controls translation in a transient expression assay. Mol. Cell. Biol. 7, 549-551. Andersson, S., Davis, D. L., DahlbSck, H., JSrnvall, H., and Russell, D. W. (1989). Cloning, structure, and expression of the mitochondrial cytochrome P-450 sterol 26-hydroxylase, a bile acid biosynthetic en- zyme. J. Biol. Chem. 264, 8222-8229. Ayer, D. E., Kretzner, L., and Eisenman, R. N. (1993). Mad: a hetero- dimeric partner for Max that antagonizes Myc transcriptional activity. Cell 72, 211-222. Beokmann, H., Su, L.-K., and Kadesch, T. (1990). TFE3: a helix-loop- helix protein that activates transcription through the immunoglobulin enhancer p.E3 motif. Genes Dev. 4, 167-179. Blanar, M. A., and Rutter, W. J. (1992). Interaction cloning: identifica- tion of a helix-loop-helix zipper protein that interacts with c-Fos. Sci- ence 255, 1014-1018. Bradford, M. M. (1976). A rapid and sensitive method for the quantita- tion of microgram quantities of protein utilizing the principle of protein- dye binding. Anal. Bioohem. 72, 248-254. Briggs, M. R., Yokoyama, C., Wang, X., Brown, M. S., and Goldstein, J. L. (1993). Nuclear protein that binds sterol regulatory element of LDL receptor promoter. I. Identification of the protein and delineation of its target nucleotide sequence. J. Biol. Chem. 268, 14490-14496.

Brown, M. S., and Goldstein, J. L. (1986). A receptor-mediated pathway for cholesterol homeostasis. Science 232, 34-47. Chen, W.-J., Andres, D. A., Goldstein, J. L., Russell, D. W., and Brown, M. S. (1991). cDNA cloning and expression of the peptide-binding I~ subunit of rat p21 '=' farnasyltransferase, the counterpart of yeast DPRIlRAM1. Cell 66, 327-334. Dang, C. V., Dolde, C., Gillison, M. L., and Kato, G. J. (1992). Discrimi- nation between related DNA sites by a single amino acid residue of Myc-related basic-helix-loop-helix proteins. Proc. Natl. Acad. Sci. USA 89, 599-602.

Dawson, P. A., Hofmann, S. L., van der Wssthuyzen, D. R., Brown, M.S., and Goldstein, J. L. (1988). Sterol-dependent repression of low density lipoprotain receptor promoter mediated by 16-base pair se- quence adjacent to binding site for transcription factor Spl. J. Biol. Chem. 263, 3372-3379.

Farmer, K., Catala, F., and Wright, W. E. (1992). Alternative multimeric structures affect myogenin DNA binding activity. J. Biol. Chem. 267, 5631-5636.

Ferr6-D'Amar~, A. R., Prondergast, G. C., Ziff, E. B., and Burley, S.K. (1993). Recognition by Max of its cognate DNA through a dimeric b/HLHIZ domain. Nature 363, 38-45. Gamier, J., Osguthorpe, D. J., and Robson, B. (1978). Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. Mol. Biol. 120, 97-120. Gil, G., Faust, J. R., Chin, D. J., Goldstein, J. L., and Brown, M. S. (1985). Membrane-bound domain of HMG CoA reductase is required for sterol-enhanced degradation of the enzyme. Cell 41,249-258. Goldstein, J. L., and Brown, M. S. (1990). Regulation of the mevalonate pathway. Nature 343, 425-430. Harlow, E., and Lane, D. (1988). Antibodies: A Laboratory Manual (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press). Hu, Y.-F., Luscher, B., Admon, A., Mermod, N., and Tjian, R. (1990). Transcription factor AP-4 contains multiple dimerization domains that regulate dimer specificity. Genes Dev. 4, 1741-1752. Kato, G. J., Barrett, J., Villa-Garcia, M., and Dang, C. V. (1990). An amino-terminal c-Myc domain required for neoplastic transformation activates transcription. Mol. Cell. Biol. 10, 5914-5920. Kozak, M. (1984). Selection of initiation sites by eucaryotic ribosomes: effect of inserting AUG triplets upstream from the coding sequence for preproinsulin. Nucl. Acids Ras. 12, 3873-3893. Laemmli, U. K. (1970). Cleavage of structural proteins during the as- sembly of the head of bacteriophage T4. Nature 227, 680--685. Landschulz, W. H., Johnson, P. F., and McKnight, S. L. (1988). The leucine zipper: a hypothetical structure common to a new class of DNA binding proteins. Science 240, 1759-1764. Mermod, N., O'Neill, E. A., Kelly, T. J., and Tjian, R. (1989). The proline- rich transcriptional activator of CTF/NF-I is distinct from the replication and DNA binding domain. Cell 58, 741-753. Murr6, C., and Baltimore, D. (1992). The helix-loop-helix motif: struc- ture and function. In Transcriptional Regulation, S. L. McKnight and K. R. Yamamoto, ads. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press), pp. 861-879. Osborne, T. F., Gil, G., Goldstein, J. L., and Brown, M. S. (1988). Operator constitutive mutation of 3-hydroxy-3-mathylglutaryl coen- zyme A reductase promoter abolishes protein binding to sterol regula- tory element. J. Biol. Chem. 263, 3380-3387. Osborne, T. F., Bennett, M., and Rhee, K. (1982). Red 25, a protein that binds specifically to the sterol regulatory region in the promoter for 3-hydroxy-3-methylglutaryl--coenzyme A reductase. J. Biol. Chem. 267, 18973-18982. Pabo, C. O., and Sauer, R. T. (1992). Transcription factors: structural families and principles of DNA recognition. Annu. Rev. Biochem. 61, 1053-1095. Prendergast, G. C., Lawe, D., and Ziff, E. B. (1991). Association of Myn, the murine homolog of Max, with c-Myc stimulates mathylation- sensitive DNA binding and ras cotransformation. Cell 65, 395--407. Sambrook, J., Fritech, E. F., and Maniatis, T. (1989). Molecular Clon-

Sterol Regulatory Element-Binding Protein 1 197

ing: A Laboratory Manual, Second Edition (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press). Sanger, F., Coulson, A. R., Barrell, B. G., Smith, A. J. H., and Roe, B. A. (1980). Cloning in single-stranded bacteriophage as an aid to rapid DNA sequencing. J. Mol. Biol. 143, 161-178. Smith, J. R., Osborne, T. F., Goldstein, J. L., and Brown, M. S. (1990). Identification of nucleotides responsible for enhancer activity of sterol regulatory element in low density lipoprotein receptor gene. J. Biol. Chem. 266, 2306-2310. Stone, K. L., LoPresU, M. B., Crawford, J. M., DeAngelis, R., and Williams, K. R. (1989). Enzymatic digestion of proteins and HPLC peptide isolation. In A Practical Guide to Protein and Peptide Purifica- tion for Micmsequencing, P. T. Matsudairia, ed. (San Diego, California: Academic Press), pp. 33-47. Wang, X., Briggs, M. R., Hua, X., Yokoyama, C., Goldstein, J. L., and Brown, M. S. (1993). Nuclear protein that binds sterol regulatory element of LDL receptor promoter. II. Purification and characterization. J. Biol. Chem. 268, 14497-14504. Zervos, A. S., Gyuris, J., and Brant, R. (1993). Mxil, a protein that specifically interacts with Max to bind Myc-Max recognition sites. Cell 72, 223-232.

GenBank Accession Number

The accession number for the sequence reported in this paper is U00968.