Expression of heterologous genes in Saccharomyces cerevisiae from vectors utilizing the glyceraldehyde-3-phosphate dehydrogenase gene promoter

Gene, 32 (1984) 263-274

263

Elsevier GENE

1155

Expression of heterologous genes in Saccharomyces cerevisiae from vectors utilizing the glyceraldehyde3-phosphate dehydrogenase gene promoter (Recombinant DNA; eukaryotic promoter; synthetic DNA; translation initiation; HBsAg; interferon)

Grant A. Bitter and Kevin M. Egan Amgen, 1900 Oak Terrace Lane, Thousand Oaks, CA 91320 (U.S.A.) (Received

June 4th, 1984)

(Revision

received

(Accepted

August

August

Tel. (805)499-5725

7th, 1984)

1 lth, 1984)

SUMMARY

The promoter region from the cloned glyceraldehyde-3-phosphate dehydrogenase (GPD) gene of Saccharomyces cerevisiue (Musti et al., 1983) has been characterized. A 653-bp TaqI restriction fragment with a 3’ border 24 bp upstream from the ATG initiation codon was isolated and demonstrated to contain all sequences necessary for promoter function in vivo. This DNA segment was converted to a portable promoter by cloning it into M13mp9, and the entire nucleotide sequence of the portable promoter was determined. Two generalized yeast expression vectors have been constructed utilizing the GPD portable promoter. The expression vectors include the yeast 2~ origin of replication and amplification functions, such that the plasmids are maintained at high copy number in cir” yeast hosts. These vectors direct synthesis of a consensus a-interferon (IFN-&on,) as 1y0 of total cell protein. Hepatitis B surface antigen (HBsAg) was also expressed from these vectors. The 5’ end of the HBsAg gene was replaced with a synthetic DNA segment which restored the deleted GPD untranslated leader and utilized optimal yeast codons for the first 30 amino acids. The partially synthetic gene resulted in a lo- to 15fold increased expression level from GPD vectors yielding HBsAg polypeptide as 2-4% of total cell protein.

INTRODUCTION

The yeast S. cerevisiue offers an attractive eukaryotic alternative to Escherichia coli as a host for the production of foreign polypeptides. It has no Abbreviations:bp,

base pairs;

gen; HBV, hepatitis hyde-3-phosphate GPD;

dehydrogenase;

kb, 1000 bp; PAGE,

RF, replicating

HBsAg,

hepatitis

B virus; IFN, interferon;

anti-

GPD, yeast gene coding

polyacrylamide

form; SD, see MATERIALS

section b; SDS, sodium

dodecyl

0378-l 119/84/$03.00

1984 Elsevier

0

B surface

GPD, glyceraldefor

gel electrophoresis; AND METHODS,

sulfate.

Science

Publishers

pathogenic relationship with man, is free of endotoxins and has been used in industrial fermentations for centuries. In addition, S. cerevisiae is capable of glycosylating proteins. This modification may be required for the activity and/or stability of certain polypeptides. Because of the complex regulation of eukaryotic gene expression, parameters that determine high-level gene expression in yeast have not been fully characterized. Hitzeman et al. (1981) originally demonstrated that heterologous genes could be expressed in yeast utilizing the alcohol

264

gene (ADHI) promoter.

dehydrogenase ly, expression utilizing and

of foreign genes in yeast was reported

promoter

phoglycerate

Subsequent-

segments

from the cloned

phos-

kinase gene (PGK; Derynk et al., 1983)

acid phosphatase

gene (PHO5;

Miyanohara

S. cerevisiae contains

three nontandemly

repeated

genes for glyceraldehyde-3-phosphate

de-

hydrogenase

(GPD) per haploid

and Holland,

1979a). These genes have been cloned

and sequenced Holland

(Holland

and Holland,

et al., 1983) and

experiments

genome

Sl nuclease

have demonstrated

(Holland

and Carbon,

1980) or JM 103 (Messing,

1983) as host for CaCl,-mediated (Cohen et al., 1972) or transfection 5’. cerevisiae 20B-12 1976) or RH218 et al.,

et al., 1983). structural

(Tschumper

1978) was

method of Hinnen were

selected

nitrogen containing

and

(MATa

transformation (Messing,

pep4-3

trpl;

1983). Jones,

(MATa gaZ2 trpl cir”; Miozzari transformed

according

to the

et al. (1978). Yeast transformants cultured

base without

in SD (0.67%

amino

0.5% Casamino

acids,

yeast

2% dextrose)

acids.

1976; 1980; protection

(c) Vector constructions

that all three genes

are transcribed in vegetatively growing yeast (Holland et al., 1983; Musti et al., 1983). GPD accounts for up to 5 % of the dry weight of commercial bakers

Recombinant DNA manipulations were performed using standard methodologies (Maniatis et al., 1982). Plasmids ppby (Musti et al., 1983) YRp7

yeast (Krebs, 1953) and the mRNA encoding this enzyme represents 2-5% of the total yeast poly(A)mRNA (Holland and Holland, 1978). One of the

(Struhl et al., 1979), pGT40 and pGT4 1 (Tschumper and Carbon, 1983) have been described. pHBV-8 was obtained from Dr. L. Overby, Abbott Labora-

three isozymes, the one encoded by the gene on pgap491 (Holland and Holland, 1979a), accounts for most of the cellular GPD enzyme (Jones and

tories. Restriction enzyme analysis and partial nucleotide sequence determinations demonstrated that

Harris, 1972; Holland et al., 1983). It is likely, therefore, that this gene is controlled by a highly efficient promoter. In the present report, we describe the isolation of a functional promoter segment from the cloned yeast GPD gene. Generalized expression vectors have been constructed utilizing this promoter, and high-level expression of IFN-c&on, and hepatitis B surface antigen have been achieved.

MATERIALSANDMETHODS

(a) Biochemicals Nucleic acid modification enzymes were purchased either from BRL or New England Biolabs and used under conditions suggested by the manufacturer. [ 35S] Methionine, [ 35S] cysteine, and [p 32P]ATP were purchased from New England Nuclear. Auszyme II immunoassay kits were purchased from Abbott Laboratories. Media components were from Difco. (b) Strains, transformation

and culture conditions

Plasmids were constructed either E. coli HBlOl (Bolivar

and isolated using et al., 1977) JA300

the hepatitis genome cloned in pHBV-8 was serotype adw (not shown). The construction of pHBsl(GPD), p(GPD-HBs)-2, p(GPD-HBs)-3, p(GPDHBs)-4, pHBS-2, pGPD-1 and pGPD-2 are described in RESULTS, sections a, b and d. All vector constructions utilized either plasmid pBR322 (Bolivar et al., 1977; Sutcliffe, 1978) or coliphage M13mp9 (Messing, 1983) as a backbone. Plasmid pp6y (Musti et al., 1983) was digested with Hind111 and the 2. I-kb DNA fragment containing the GPD gene purified by agarose gel electrophoresis. The fragment was digested to completion with TaqI and the 650 bp fragment containing the GPD promoter region purified by PAGE. Utilization of this promoter segment in expression vectors is described in RESULTS, sections a and d. The 850-bp HBsAggene with resynthesized 5’ end was purified from pHBs-2 as a BamHI-EcoRI fragment, and blunted with Sl. This fragment was cloned, in the correct orientation, into either BamHIdigested pGPD-1 or pGPD-2, both of which had been similarly Sl-digested, to generate pGPDl(HBs) and pGPD-2(HBs), respectively. The IFNas a 509-bp XbaI&on 1 gene was isolated BamHI DNA fragment with the 5’ sequence CTAGAGAATG ~ followed by the coding region TCTTAC for IFN-&on, (Alton et al., 1983). This gene was

265

cloned into either pGPD- 1 or pGPD-2 by blunt-end ligation as described above to generate pGPD-l(Int a) and pGPD-2 (Int a), respectively. (d) Assay of protein products HBsAg synthesis was assayed by both immunoassay and SDS-PAGE. Yeast strains were cultured to an A,,, of 0.5, collected by centrifugation and resuspended in 25 mM Tris pH 7.5, 0.9 M sorbitol. Zymolase 60000 (Kirin Brewery) was added to 50 pg/ml and the suspension incubated at 22°C for 60 min. The spheroplasts were collected by centrifugation, resuspended in 25 mM Tris, pH 7.5 and sonicated for 3 x 15-s bursts using a Heat Systems Model W375 Sonitier. The extracts were assayed using the Auszyme II kit. The mass of HBsAg present in the yeast extracts was quantitated by comparison to a standard curve generated using human serum HBsAg of known concentration. Three or four dilutions of the yeast extract which yielded A,,, values within the linear range of the standard curve were averaged to calculate the HBsAg concentration. For SDS-PAGE assay of HBsAg, cells were cultured to an A,,, of 0.5-1.0, collected by centrifugation, washed once in 0.67% yeast nitrogen base, and resuspended in SD containing an amino acid “drop-out” solution lacking tryptophan, methionine and cysteine at a density of 2 x 10’ cells/ml. [ 35S] Methionine or [ 35S] cysteine (10 PCi) was added to each culture and incubation continued at 30” C for 60 min. The cells were collected by centrifugation, resuspended in sample buffer (100 mM Tris, pH 6.8, 2% SDS, 10% glycerol, 5 y0 P-mercaptoethanol, 0.17 y0 bromphenol blue) boiled for 5-10 min. and subjected to 12.5% PAGE (Laemmli, 1970). The gels were dried and exposed to Kodak XAR-2 film at -70°C with an intensifier screen. IFN synthesis levels were quantitated by an endpoint cytopathic effect inhibition assay as described (Alton et al., 1983). For SDSPAGE analysis of IFN synthesis, yeast cells were cultured and labeled with [ 35S] methionine as above. Whole-cell proteins were subjected to 15% PAGE (Laemmli, 1970) and the dried gel was autoradiographed as above.

RESULTS

(a) Isolation of functional GPD promoter Musti et al. (1983) used a cloned chicken GPD gene as a hybridization probe to screen an S. cerevisiue genomic library. Nucleotide sequence analysis of their pp6y clone indicated that it encodes the same GPD isozyme as does the pgap491 clone originally isolated by Holland and Holland (1979a). This gene contains a TuqI restriction site which is 24 bp 5’ to the ATG translation initiation codon of the GPD gene (Holland and Holland, 1979b) and 13-20 bp 3’ to the transcription start points (Musti et al., 1983; Holland et al., 1983). The next TuqI site 5’ to the site in the untranslated leader is approximately 650 bp upstream. We thus expected this 650-bp TuqI restriction fragment to contain all DNA sequences necessary for promoter function in vivo. We tested this anticipation by utilizing the hetatitis B surface antigen gene for expression studies. Hepatitis B surface antigen (HBsAg) is a 226 amino acid polypeptide encoded by a gene that is included on a 1400-bp BamHI fragment in the genome of HBV serotype a&v (Valenzuela et al., 1979). The ATG translation initiation codon for the mature HBsAg is 126 bp from the BumHI site and there is no other ATG in this sequence (Valenzuela et al., 1980; Burnette, W.N. and Bitter, G.A., unpublished results). We subcloned the BumHI fragment into the yeast shuttle vector YRp7 and subsequently deleted 442 bp of HBV DNA and 275 bp of pBR322 DNA, as indicated in Fig. 1A. Plasmid pHBs-1 thus has a unique BumHI site 5’ to the HBsAg coding region. Following the HBsAg termination codon are 131 bp of HBV DNA, 23 bp of pBR322, and the 1453-bp yeast TRPI-ARSI EcoRI (Tschumper and Carbon, 1980) fragment. In this construction, the sense strands of the HBsAg and TRPI genes are on the same strand of the plasmid. It has previously been demonstrated (Amerer et al., 198 1) that in such a construction (5’-heterologous gene-3’ + 5’TRPl gene-3’), the TRPl gene will supply transcription termination and/or polyadenylation signals to RNA polymerase II molecules that have read through the heterologous gene and into the yeast DNA segment. The 650-bp GPD TuqI fragment, which includes the transcription start site, was cloned into the BumHI site of pHBs-1 as indicated

pHBsAg

pHBs-1

pHBs-l(GPD)

p (GPD-HBs)-2 Fig. 1. Construction fragment

of plasmid

into YRp7 to generate

refers to the Klenow fragment

vectors.

was replaced isolated

site of pHBs-1,

pHBs-l(GPD).

and BumHI-cleaved

constructed

from pHBs-2

pGT40

pHBs-2

segment

This fragment

either San-

of pHBs-l(GPD).

The HBsAg gene from pHBV-8

The 3’ region of HBV DNA and a portion

of_!?. coli DNA polymerase

with the 131-bp synthetic

from pHBs-l(GPD).

(A) Construction

pHBsAg.

c) was cloned into the unique BumHI to HBV DNA was designated

p(G PD-H Bs)-3

or pGT41

and a clone with the untranslated is identical

to pHBs-1

(Fig. 5). (B) Construction

includes

by the same procedures

650-bp GDP Tag1fragment (MATERIALS

I). The

the GPD promoter,

to generate

to generate pHBs-l(GPD). These vectors were introduced into S. cerevisiue RH218, and whole cell extracts from transformants were tested for the presence of HBsAg, by immunoassay. HBsAg was detected in cells carrying pHBs-l(GPD) (GPD transcription start points adjacent to HBV DNA), but not in transformants carrying a similar plasmid in which the GPD fragment was in the opposite orientation (not shown). No HBsAg was detected in cells harboring pHBs-1. These results demonstrate that this 650-bp DNA fragment, in the correct orientation, contains all sequences necessary for promoter function in vivo.

of 2p-based

(DNAPase

AND METHODS,

leader region of the GPD segment

with the exception

HBsAg-coding

p(GPD-HBs)-2

used to construct

was cloned as a BamHI

of pBR322 were deleted as indicated

that the 218-bp BumHI-XbaI

vectors.

adjacent fragment

fragment

was

region and TRPI gene, and was cloned

into

and p(GPD-HBs)-3,

The 2.8-kb San-BglII

I

section

respectively.

p(GPD-HBs)-4

was

p(GPD-HBs)-3.

(b) Construction of self amplifying yeast vectors Quantitation of the immunoassay results indicate that transformants carrying pHBs-l(GPD) synthesize 0.5 pg HBsAg per liter culture at an A600 of 1.0. Two additional plasmids were constructed using the GPD-HBsAg-TRPI expression segment from pHBsl(GPD) and replication functions from the yeast 2~ plasmid (Fig. 1B). p(GPD-HBs)-2 contains the 2242-bp EcoRI fragment of 2p DNA (B form) which includes the origin of replication. p(GPD-HBs)-3 contains the entire 2~ plasmid (B form; Hartley and Donelson, 1980) cloned in the EcoRI site in the large unique region. In this construction, the REP1 and REP2 genes, as well as the REP3 locus (Jayaram

267

et al., 1983), are intact. The 2~ replicon-containing plasmids were transformed into S. cerevisiue RH218 (a cir” isolate lacking endogenous 2~ plasmid) and HBsAg expression monitored by immunoassay (Fig. 2). Transformants carrying p(GPD-HBs)-2 synthesize HBsAg at approximately the same level as pHBs-l(GPD) transformants, consistent with the plasmid being present at low copy number. In contrast, p(GPD-HBs)-3 transformants synthesize lo-20 fold more HBsAg in this cir” host (approx. - 10 pg HBsAg/l culture at an A,,, of 1) indicating that this vector includes all functions necessary for self-amplification to high copy numbers. Utilization of self-amplifying expression vectors in the absence of endogenous 2~ DNA avoids potential complications due to recombination between endogenous and recombinant plasmids. (c) Nucleotide

sequence of the GPD promoter

The above results demonstrate that the isolated 650-bp GPD TaqI fragment is a functional yeast promoter. To construct generalized expression vectors utilizing the GPD promoter, this fragment was cloned into the AccI site of M13mp9 from which it

may be excised as a HindIII-BumHI portable promoter utilizing restriction sites supplied by the phage polylinker segment. The entire nucleotide sequence of the GPD portable promoter was determined and is presented in Fig. 3. The 3’ terminus of the GPD segment (TuqI site) is at position -24 if the A of the ATG initiation codon is assigned position + 1. Musti et al. (1983) demonstrated by Sl nuclease protection and primer extension experiments that there are two 5’ ends of the GPD mRNA which map at position -37 and -38. Holland et al. (1983) mapped the major transcript 5’ end at position -44. The reason for the discrepancy between the results of these investigators is not clear. The nucleotide sequence from position -150 to -24 is identical to that previously reported by Holland and Holland (1979). By isolating the GPD portable promoter from the phage RF as a HindIII-BumHI fragment, the sequence of the GPD untranslated leader is altered from TCGAATAAACA to TCGACGGATCC (TaqI restriction site underlined). Additional structural features of the GPD promoter are analyzed below (see DISCUSSION). (d) Generalized expression

vectors

Two generalized expression vectors were constructed utilizing the GPD portable promoter (Fig. 4). Both vectors incorporate the entire 2~ form B plasmid cloned in the EcoRI site in the large unique region, and both vectors include the TRPl gene for selection in yeast. pGPD-1 also utilizes the TRPl gene for transcription termination/polyadenylation signals, while pGPD-2 incorporates the 3’ region (BgflI-HindIII) of the yeast PGK gene (Hitzeman et al., 1982) for this function. Both vectors contain a unique BumHI site between the promoter and terminator segments into which heterologous genes may be cloned for expression. mg Protein/mL

Fig. 2. HBsAg synthesis in yeast transformants. Extracts were prepared from strain RH218 transformed with pHBs-l(GPD) (O-O), p(GPD-HBs)-2 (O-O) or p(GPD-HBs)-3 (A-A), as described in MATERIALS AND METHODS, section d. The extracts were diluted to the indicated protein concentrations and HBsAg measured by Auszyme II immunoassay. Background values were determined by assaying an equivalent protein concentration of a nontransformed yeast cell extract; these values (approx. 0.01 A,,, units) were subtracted from the value obtained for the transformed yeast cell extracts.

(e) Expression

of HBsAg

We sought to increase the translation efficiency of the HBsAg mRNA in yeast by optimizing the sequence at the 5’ end of the gene. We chemically synthesized a 131-bp DNA segment with BumHI and X&I cohesive ends (Fig. 5). This segment was cloned into BumHI and J&I-digested pHBs-1 to generate pHBs-2 (not shown). pHBs-2 is thus

268

-650 TCAATACTCG CCATTTCAAA AGTTATGAGC GGTAAAGTTT

GAATACGTAA CTTATGCATT

AAGCTTGGCT TTCGAACCGA

GCAGGTCGAG CGTCCAGCTC

TTTATCATTA AAATAGTAAT

ATAATTAATA TATTAATTAT

GTAGTGATTT CATCACTAAA

TCCTAACTTT AGGATTGAAA

-600 ATTTAGTCAA TAAATCAGTT

AAAATTAGCC TTTTAATCGG

TTTTAATTCT AAAATTAAGA

GCTGTAACCC CGACATTGGG

GTACATGCCA CATGTACGGT

-550 AAATAGGGGG TTTATCCCCC

CGGGTTACAC GCCCAATGTG

AGAATATATA TCTTATATAT

ACACTGATGG TGTGACTACC

TGCTTGGGTG ACGAACCCAC

-500 AACAGGTTTA TTGTCCAAAT

TTCCTGGCAT AAGGACCGTA

CCACTAAATA GGTGATTTAT

TAATGGAGCC ATTACCTCGG

CGCTTTTTAA GCGAAAAATT

-450 GCTGGCATCC CGACCGTAGG

AGAAAAAAAA TCTTTTTTTT

AGAATCCCAG TCTTAGGGTC

CACCAAAATA GTGGTTTTAT

TTGTTTTCTT AACAAAAGAA

-400 CACCAACCAT GTGGTTGGTA

CAGTTCATAG GTCAAGTATC

GTCCATTCTC CAGGTAAGAG

TTAGCGCAAC AATCGCGTTG

-350 'J'ACAGAGAAC AGGGCACAAA ATGTCTCTTG TCCCGTGTTT

CAGGCAAAAA GTCCGTTTTT

ACGGGCACAA TGCCCGTGTT

CCTCAATGGA GGAGTTACCT

GTGATGCAAC CACTACGTTG

-300 CTGCCTGGAG GACGGACCTC

TAAATGATGA ATTTACTACT

CACAAGGCAA GTGTTCCGTT

TTGACCCACG AACTGGGTGC

CATGTATCTA GTACATAGAT

-250 TCTCATTTTC AGAGTAAAAG

TTACACCTTC AATGTGGAAG

TATTACCTTC ATAATGGAAG

TGCTCTCTCT ACGAGAGAGA

GATTTGGAAA CTAAACCTTT

-200 AAGCTGAAAA TTCGACTTTT

AAAAGGTTTA TTTTCCAAAT

AACCAGTTCC TTGGTCAAGG

CTGAAATTAT GACTTTAATA

TCCCCTACTT AGGGGATGAA

A CGGTAGGTAT T GCCATCCATA

TGATTGTAAT ACTAACATTA

TCTGTAAATC AGACATTTAG

-1nrl TATTTCTTAA ATAAAGAATT

-150

ACTTCTTAAA TGAAGAATTT

TTCTACTTTT AAGATGAAAA

AGTTTCGACG TCAAAGCTGC

GATCC CTAGG

Fig. 3. Nucleotide

ATAGTTAGTC TATCAATCAG

of the GPD portable

sequence

promoter.

section d) was cloned into the AccI site of M13mp9. to the phage Hind111 site while in M13/CPD-3 were sequenced

by the dideoxy

the determined

sequence,

used as a primer transcription

to resequence

the central

start site is adjacent

M 13 polylinker

are underlined.

numbering

system + 1.

is relative

region

by the asterisks,

to the translation

the GPD segment

start site was adjacent

(Sanger

(see MATERIALS

to the phage BarnHI

site is presented.

Nucleotides

AND METHODS,

sequence

start site adjacent

site. Both phage isolates

primer (Messing,

-261 to -248 of the coding

The nucleotide

CCAAGAACTT GGTTCTTGAA

with the transcription

et al., 1977) using the Ml3 universal

to nucleotide

of the fragment.

to the phage BarnHI

The TATA homology

by Musti et al. (1983) are indicated position

method

(complementary

TagI GPD fragment

contained

the transcription

chain termination

an oligonucleotide

The 650-bp

M13/GPD-1

**

-5n TTTTTTTTAG TTTTAAA~CA AA-AAAAAATC AAAATTTTGT

strand)

of M13/GPD-3,

of the GPD portable

1983). From

was synthesized in which

promoter

derived

is enclosed in a box. The position of the 5’ ends of the yeast GPD transcript while the 5’ end mapped

initiation

codon

(Hohand

by Holland

and Holland,

et al. (1983) is indicated

and

the GPD from the mapped

by the dot. The

1979a) with the A of the ATG

assigned

269

Eco RI

Eco RI

/ Eco RI Fig. 4. Construction M 13/GPD-3 pGPD-1

of generalized

was constructed

Tschumper

and Carbon,

the HindIII-BumHI EcoRI-BgflI

GPD expression

(Fig. 3) as a HindHI-BarnHI from pGT41 1980). pGPD-2

PGK

fragment)

promoter

fragment

was constructed

segment

of these vectors

are indicated

region, open segment;

terminator

regions

are indicated

promoter

was isolated

with the GPD portable

and the yeast

Both pGPD-1

promoter.

This plasmid

and pGPD-2

as follows:

2~ DNA,

segment.

hatched

segment;

The GPD transcription

pBR322 DNA sequences.

TRPZ gene (852-bp

contain

contains

EcoRI-BgLII

in preparation)

GPD portable

fragment;

by replacing

TRPl gene (852-bp

the yeast

the entire yeast 2~ plasmid

ofthe 2~ plasmid relative to pBR322 are opposite

TRPl gene, stippled

from the RF form of phage

The thin line indicates

from pPG70 (Jones, M. and Koski, R., manuscript

cloned in the Safl site of pBR322.

terminator

promoter

vector construction.

using the GPD portable

EcoRI site in the large unique region. The orientations components

The GPD portable

vectors.

for subsequent

cloned at the

in the two vectors. The functional promoter,

start point and polarity

black

segment;

PGK

of the TRPZ and PGK

by the arrows.

identical to pHBs-1 (Fig. 1A) with the exception that the 218-bp BamHI-X&I fragment of HBV DNA (Valenzuela et al., 1980) has been replaced by the

described in MATERIALS to generate pGPD-l(HBs) shown).

synthetic DNA segment. The 27 bp immediately 5’ to the ATG initiation codon is the GPD untranslated

The synthesis of HBsAg was assayed by SDSPAGE of whole yeast cell proteins (Fig. 6A). A new protein is expressed when the partially synthetic

leader (Holland and Holland, 1979), which was deleted from the GPD portable promoter. The 89 bp from the XbaI site encodes the same amino acids as does the native HBsAg gene. However, the codons utilized in this segment are those which are preferentially utilized in highly expressed yeast genes (Bennetzen and Hall, 1981). The BamHI-EcoRI 850-bp fragment of pHBs-2 was purified and cloned into the unique BamHI site of pGPD-1 and pGPD-2 as

Fig. 5. Nucleotide the HBsAggene. segment

sequence

of the resynthesized

Oligonucleotides

was assembled

1984). The sequence

essentially

shown

was cloned into pHBs-1

were synthesized as described

is the BamHI-XbaI

to generate

pHBs-2

5’ segment

of

and the gene (Bitter

et al.,

fragment

that

(legend to Fig. 1).

HBsAg

gene is present

METHODS, section c, and pGPD-2(HBs) (not

AND

in pGPD-1

and pGPD-2.

This protein has been identified as HBsAg based on the following criteria: (i) it has an M, of 22000; (ii) its expression is dependent on the presence of the HBsAg gene cloned in the correct orientation in the expression vectors (compare lanes 1,3 to lanes 2,4); (iii) it is preferentially labelled with cysteine over methionine (compare lanes 5, 7 with lanes 1, 3) consistent with HBsAg being a cysteine-rich protein (Valenzuela et al., 1980). Cells transformed with either pGPD-l(HBs) or pGPD-2(HBs) synthesize HBsAg as 4% of the total cell protein. This estimate is based on densitometric scanning of autoradiograms of [ 35S] methionine-labeled proteins in Fig. 6A. Consistent with this determination, silverstained gels and quantitative Western blots indicate that HBsAg comprises approx. 2% of the whole cell protein (not shown). The effect of modification of the 5’ end of the

270

43 -

25.7

18.4 14.3

-

(Fig. 6)

(Fig. 7) Fig. 6. Synthesis pGPD-2

of HBsAg

d. The proteins

electrophoresed

with [3’S] cysteine. lysozyme,

in S. cerevisiae. (A) S. cerevisiae 20B-12 transformed

(HBs) (lanes 3,7) or pGPD-2 The positions

(lanes 4, 8) were cultured

in lanes

l-4 were from cultures

of the A4, standards

[35S] cysteine

and electrophoresed

Fig. 7. Synthesis

of IFN-aCon,.

(lane 3) or pGPD-2 MATERIALS

as described

(ovalbumin,

in MATERIALS

section

with pGPD-l(Inta)

d. The M, 19000 IFN-c&on,

HBsAg gene on expression levels was quantitated. Cells carrying either p(GPD-HBs)-3 (native HBsAg gene) or p(GPD-HBs)-4 (resynthesized HBsAg gene; legend to Fig. 1) were labeled with [35S] cysteine and whole cell proteins analyzed by SDSPAGE (Fig. 6B). HBsAg polypeptide is undetectable in whole cell extracts from p(GPD-HBs)-3 transformants. However, the protein is clearly evident in extracts from p(GPD-HBs)-4 transformants (lane 2), indicating an increased expression of more than tenfold. Quantitation of HBsAg levels by immunoassay indicate that the resynthesized gene is expressed lo-15 times more efficiently than the native gene (not shown).

d. HBsAg

25.7; /I-lactoglobulin,

is indicated

(f) Expression

18.4;

(lane 2) were labeled with is indicated

(lane I), pGPD-l(lane

and whole cell proteins

protein

section

while those in lanes 5-8 were labeled

43; a-chymotrypsinogen, section

(lanes 2, 6),

AND METHODS,

(lane 1) or p(GPD-HBs)-4

AND METHODS,

and labeled with [35S] methionine

(lanes 1,5), pGPD-1

in MATERIALS

with [35S] methionine

with p(GPD-HBs)-3

S. cerevisiae 20B-12 cells transformed

(lane 4) were cultured

AND METHODS,

labeled

are indicated

14.3 kDa1). (B) S. cerevisiae RH218 transformed

with pGPD-l(HBs)

and labeled as described

electrophoresed

by the arrow. 2), pGPD-2

(Inta)

as described

in

by the arrow.

of IFN-don,

To demonstrate the general utility of pGPD-1 and pGPD-2, the gene for IFN-c&on, was cloned into the BamHI site of each plasmid to generate pGPD-1 (Into) and pGPD-2(Intcc) (see MATERIALS AND METHODS, section c). IFN-aCon, has an average amino acid sequence of all the naturally occurring leukocyte IFNs, and expression of this gene in E. coli yields a protein with a tenfold higher antiviral activity than any of the naturally occurring a-interferons (Alton et al., 1983). Extracts prepared from S. cerevisiae 20B-12 transformed with pGPD-1 (Into) or pGPD-2(Intcc) contained 4 x lo8 units interferon

271

per liter culture at an A,,, of 1 (not shown). Whole cell proteins from these transformants were analyzed by SDS-PAGE (Fig. 7). Consistent with the bioassay results, the consensus u-interferon represents approx. 1y0 of the total cell protein. We have also expressed a synthetic gene encoding IFN-y from these vectors at levels of 2-5x of the total cell protein (not shown). Thus, pGPD-1 and pGPD-2 are general vectors useful for expressing any cDNA in yeast. The level of protein production from these vectors is likely to depend on sequences immediately 5’ to the ATG initiating codon as well as codon utilization.

DISCUSSION

This report describes the isolation and characterization of a functional promoter from the cloned yeast GPD gene. Since this protein represents one of the most abundant yeast gene products, it was expected that the gene includes signals which promote highly efficient transcription initiation. Several expression vectors utilizing this promoter have been constructed which program the synthesis of heterologous gene products. It has previously been demonstrated that transcription termination and/or polyadenylation signals are required for the efficient expression of foreign genes in S. cerevisiae (Hitzeman et al., 1983). We have found that the level of heterologous protein synthesis is the same whether these signals are supplied by DNA sequences from either the efficiently expressed yeast PGK gene or the inefficiently expressed yeast TRPl gene (Figs. 6,7). Expression of heterologous genes from the GPD promoter has been increased both by utilizing self-ampliflying vectors and by optimizing DNA sequences surrounding the translation initiation codon. The nucleotide sequence of the GPD promoter several interesting structural region presents features. The DNA from -677 to-250 (numbering system is relative to the ATG initiation codon of the GPD structural gene) is 60% AT. In contrast, the region -250 to -24, which encompasses the transcription initiation site, is 7 1y0 AT. The 5 ’ boundary of the GPD promoter is not known, but it is likely that upstream sequences (adjacent to the Hind111 site of the portable promoter) could be deleted without

affecting promoter function. Several conserved DNA sequences in the 5’ region of eukaryotic genes have been implicated in promoter function. The TATA box has the consensus sequence 5’-TATA,AAT,-3’ and functions to direct RNA polymerase II to initiate transcripts a fixed distance downstream (Corden et al., 1980). The GPD promoter region contains the sequence TATATAAA beginning at position -141 and is thus centered 94 bp upstream from the transcription start point. This distance is similar to that observed for the yeast PGK gene (109 bp; Hitzeman et al., 1982) but much longer than the 25-30 bp separating the TATA box from the transcription initiation site observed for mammalian class II genes. The large distance between the TATA box homology and transcription start point may explain the inability of the mammalian transcription apparatus to initiate specific transcription on these yeast class II genes in vitro (Bitter, 1983). A second conserved sequence in the 5’ flanking region of eukaryotic class II genes is the CAAT box which has the consensus sequence 5’-GG,CCAATCT-3’ and is generally located 70-80 bp upstream from the transcription start site (Benoist et al., 1980). The function of this conserved sequence is not known and it is not present, or has only weak homology, in a number of yeast genes that have been sequenced (Dobson et al., 1982). The GPD 5’ flanking region includes the sequence GGCAATTC beginning at position -276, which is 232 bases upstream from the transcription initiation region. Whether this sequence homology to the mammalian CAAT box is a functional component of the yeast GPD promoter remains to be determined. Finally, there are two notable sequences in the AT-rich region of the GPD promoter which include long stretches of A’s or T’s in the coding strand. Upstream of the TATA box, the sequence (A),GCTG(A), begins at position -204 while downstream of the TATA box, the sequence (T),AG(T),(A), begins at position -61. The function, if any, of these sequences in promoter activity is not known. In addition to using the highly efficient yeast GPD

272

promoter

in developing

have incorporated

these expression

synthetic

timize

sequences

around

codon.

An analysis

DNA

vealed

that

segments

the translation

of the nucleotide

cloned yeast genes encoding the untranslated

these mRNAs

vectors, we

are markedly

abundant leader

sequences

of

proteins

re-

sequences

of

A-rich and G-deficient

(Amerer et al., 198 1), and this characteristic results in efficient translation

to op-

initiation

initiation.

probably

In addition,

in sequence,

No optimization have demonstrated

that the strength of the GPD and

PGK promoters Koski, reported modified

and

may provide a mechanism for efficient translation elongation. The native HBsAg gene encodes an untranslated leader which contains 6 A residues and 10 G residues in the 25 bp immediately preceding the translation initiation codon (Valenzuela et al., 1980). In contrast, the native yeast GPD untranslated leader contains 17 A residues and only 1 G residue in the 25 bases upstream from the ATG codon (Holland and Holland, 1979b). Additionally, the HBsAg gene has a codon usage that is not biased as are the highly expressed yeast genes and, in fact, has a codon bias index of 0.018 when calculated as in Bennetzen and Hall (198 1). The 5 ’ end of the native HBsAg gene was replaced with a chemically synthesized segment which restored the GPD untranslated leader region deleted from the GPD portable promoter and utilized optimal yeast codons for the first 30 amino acids. This gene was expressed lo- to 15-fold more efficiently than the native HBsAggene. Expression of the native HBsAg gene from p(GPD-HBs)-3 yields a transcript with a 156-base untranslated leader. This is much longer than native yeast untranslated leaders, which are usually 30-50 bases. The transcript expressed from p(GPD-HBs)-4, in contrast, has the native GPD untranslated leader with the exception of an insertion of CGGATCC 27 bp upstream from the ATG initiation codon. It is not clear whether the more than tenfold increased expression from the resynthesized HBsAg gene is due to optimization of either the untranslated leader or the first 30 codons, or a combination of both. Hitzeman et al. (1983) have also studied the expression of HBsAg in yeast ; utilizing the PGK promoter, these investigators report HBsAg expression levels comparable to those observed in this study. The vector utilized by Hitzeman et al. (1983) encodes an mRNA untranslated leader that is identical in size, and very similar

(Bitter, G.A. and

results).

Given

and by Hitzeman

seems likely that the increased optimization

region. We

this con-

and the similar HBsAg expression

herein

yeast genes have an extreme codon species,

are comparable

R., unpublished

sideraction,

and Hall, 198 1) which corresponds tRNA

however, were made by

in the HBsAg-coding

highly expressed

isoaccepting

of codons,

leader.

these investigators

bias (Bennetzen to the abundant

to the native PGK untranslated

HBsAg

gene

expression

is due,

of the untranslated

Utilizing

levels

et al. (1983), it level of the

in large

part,

to

leader region.

the GPD promoter,

either

the

TRPl

orPGK transcription termination regions, multicopyvectors, and optimized translation initiation sequences, we have obtained expression of heterologous genes in S. cerevisiae as l-5 % of the total cell protein.

Since

certain

native

yeast

proteins

(e.g.,

GPD) are expressed at this level from single-copy genes, it is apparent that further modification of the expression vectors will yield even higher absolute expression

levels. Optimization

of sequences

around

the translation initiation region have profound effects on expression. It now seems likely that further optimization of translation elongation will increase production yields.

ACKNOWLEDGEMENTS

We are grateful to Dr. Richard A. Kramer for providing plasmid pp6y. We thank Pamela Foreman for assistance in early vector constructions, Dr. Ray Koski for plasmid pPG70, Dr. Frank Martin for assembly of the HBsAg 5’ gene segment, and Matt Jones for confirming the sequence of the cloned synthetic segment.

REFERENCES Alton, K., Stabinsky, L., Altrock, characterization derived

Y., Richards,

B., Miller,

human

and Schellekens,

and biological

Stebbing,

H. (Eds.),

expression

B., Goldstein, N.: Production,

effects of recombinant

IFN-c( and IFN-y

System. Elsevier, Amsterdam, Amerer, G., Hitzeman, R., Hagie, The functional

R., Ferguson,

L. and

analogs,

The Biology

DNA

in de Mayer,

E.

of the Interferon

1983, pp. 119-128. F., Barta, A. and Hall, B.D.:

of mammalian

genes in yeast,

in

273

Walton,

A.G.

Symposium

(Ed.),

Proceedings

on Macromolecules:

sevier, Amsterdam, Beggs,

plasmid.

Nature

Cleveland DNA.

El-

1981, pp. 185-197.

J.D.: Transformation

Bennetzen,

of the Third Recombinant

of yeast

by a replicating

hybrid

257 (1978) 104-108.

C., O’Hare,

ovalbumin

in yeast. J. Biol.

R. and Chambon,

of putative

control

P.: The

regions. Nucl.

Acids Res. 8 (1980) 129-142. Bitter, G.A.: Transcription in vitro

systems

mammalian

and yeast

in eukaryotes: utilizing

Analysis

components

cells. Mol. Gen. Genet.

of

from

191 (1983)

434-441. Bitter, G.A., Chen, K.K., Banks, A.R. and Lai, P.-H.: Secretion of foreign proteins

from Succharomyces

cerevisiue directed

by

gene fusions. Proc. Nat]. Acad. Sci. (USA) 8 1(1984)

a-factor

5330-5334. Bolivar,

F.,

Rodriguez,

Heyneker, Construction

cloning

J.R., Guarascio,

within

Greene,

the yeast

P.J.,

H.W., Crosa,

and characteristics

A multipurpose Broach,

R.L.,

H.L., Boyer,

M.C.,

of new cloning vehicles,

system.

S.: II.

2~ circle is site-specific.

B., Buchwalder,

C. and Chambon,

in new environment. Cohen,

S.N., Chang,

Cell 29

antibiotic

A., Sassone-Corsi,

P.: Expression

Science

resistance

P.,

of cloned genes

in bacteria:

Genetic

transformation

of

Sci. (USA) 69

R., Singh, A. and Goeddel,

human

interferon-y

cDNA

D.V.: Expression

in yeast.

Nucl.

of the

Acids

Res. 11

M.J., Tuite, M.F., Roberts,

Kingsman,

S.M.: Conservation

N.A., Kingsman,

A.T. and

of high efficiency

promoter

in Succhuromyces cerevisiue. Nucl. Acids Res. 10

sequences

J.L. and Donelson,

yeast plasmid. Henikoff,

Nature

J.E.: Nucleotide

in yeast

distal

sequence

of the

to a control

E.H.: Transcription sequence.

termi-

Cell 33 (1983)

A., Hicks, J.B. and Fink, G.R.: Transformation

Proc. Natl. Acad. Hitzeman,

R.A.,

Amerer, Hitzeman,

F.E.,

Levine,

H.L.,

G. and Hall, B.D.: Expression

interferon

in yeast. Nature

of a human

D.V., gene for

R.: The primary

structure

of the Succhurokinase.

Nucl.

R.A., Chen, C.Y., Hagie, F.E., Patzer,

E.J., Liu, C.-S.,

Eskell, D.A., Miller, J.F., Yaffe, A., Klein, D.G., Levinson, A.D. and Opperman, antigen

2745-2763.

H.: Expression

in yeast.

Nucl.

of hepatitis

Acids

Res.

structure

dehydrogenase

of a

gene from Suc-

M.J.: Structural

of two

genes. J. Biol. Chem. 255 (1980) 2596-2605.

J.P., Labiniec,

Homologous

yeast

comparison

dehydrogenase Holland,

glyceraldehyde-3-phosphate

L., Swimmer,

nucleotide

C. and Holland,

sequences

at the

M.J.:

5’ termini

of

messenger RNAs synthesized from the yeast enolase and glyceraldehyde-3-phosphate dehydrogenase gene families. J. Biol. Chem. 258 (1983) 5291-5299. Jayaram,

M., Li, Y.-Y. and Broach,

encodes

components

required

E.: Proteinase

Genetics

J.R.: The yeast plasmid

2~

for its high copy propagation. of Succhuromyces

mutants

B virus 11 (1983)

cerevisiue.

85 (1976) 23-30.

G.M.T.

and

dehydrogenase

Harris,

J.I.:

Glyceraldehyde-3-phosphate

amino acid sequence

of enzyme from bakers’

yeast. FEBS Lett. 22 (1972) 185-189. E.G.: Yeast glyceraldehyde-3-phosphate of fractions

dehydrogenase

precipitated

by nucleic acid. J.

Biol. Chem. 200 (1953) 471-478. Laemmh,

U.K.: Cleaving of structural

Maniatis,

T., Fritsch,

A Laboratory Messing,

protein during the assemNature

227 (1970) 680-685.

E.F. and Sambrook,

Manual. Harbor,

Cold Spring

J.: Molecular Harbor

Cloning

Laboratories,

NY, 1982.

J.: New Ml3 vectors

for cloning.

Methods

Enzymol.

101 (1983) 20-78. Miozzari,

G., Neiderberger,

through

P. and

J. Bacterial.

A., Toh-e, A., Nozaki,

and Matsubara,

R.: Tryptophan

K.: Expression

A.M., Zehner,

C., Hamada, of hepatitis

K.A., Patterson, mapping

for glyceraldehyde-3-phosphate

lated by sequence

homology

F., Ohtomo,

N.

B surface antigen

Sci. USA 80 (1983) l-5.

Z., Bostian,

R.A.: Transcriptional

of the flux

134 (1978) 48-54.

gene in yeast. Proc. Natl. Acad. Kramer,

Huffer,

in Succhuromyces cerevisiue: control

the pathway.

B.M. and

of two yeast genes dehydrogenase

with the chicken

iso-

gene. Gene 25

(1983) 133-143. S:and Coulson,

chain termination

inhibitors.

A.R.: DNA sequencing

Proc. Natl. Acad.

with

Sci. USA 74

(1977) 5463-5467. Struhl, K., Stinchcomb, frequency

Acids Res. 10 (1982) 7791-7808.

surface

M.J.: The primary

repeated

Sanger, F., Nicklen,

J.S., Chen, C.Y., Seeburg,

myces cerevisiue gene for 3-phosphoglycerate Hitzeman,

J.P. and Holland,

coding

Goeddel,

293 (1981) 717-722.

R.A., Hagie, F.E., Hayflick,

P.H. and Derynk,

ofyeast.

Sci. USA 75 (1978) 1929-1933.

Hagie,

254

non-tandemly

Musti,

607-614. Hinnen,

dehydro-

churomyces cerevisiue. J. Biol. Chem. 254 (1979b) 9839-9845. Holland,

Miyanohara,

286 (1980) 860-865.

S., Kelly, J.D. and Cohen,

nates

J.P. and Holland,

biosynthesis

(1982) 2625-2637. Hartley,

Holland,

Cold Spring

(1983) 1819-1837. Dobson,

and characterization

from Succhuromyces cerevisiue. J. Biol. Chem.

bly of the head of bacteriophage.

(1972) 21 lo-21 14. Derynck,

M.J.: Isolation

for glyceraldehyde-3-phosphate

I. Electrophoresis

L.: Non chromosomal

DNA. Proc. Natl. Acad.

and phosphoglycerate

(1979a) 5466-5474.

Krebs,

209 (1980) 1406-1414.

A.C.Y. and Hsu,

E. coli by R-factors

J.P. and Holland,

of a gene coding

Jones,

J., Wasylyk,

Kedinger,

17 (1978) 4900-4907.

Jones,

M.: Recombination

(1982) 227-234. Corden,

dehydrogenase

klinase.

Biochemistry

of

Cell 34 (1983) 95-104.

Gene 2 (1977) 95-113.

V.R. and Jayaram,

plasmid

Betlach,

J.H. and Falkow,

and identification

acids coding for enolase, glycer-

glyceraldehyde-3-phosphate

initiation

heterologous

J.P.: Isolation

ribonucleic

aldehyde-3-phosphate

genase

K., Breathnach,

gene. Sequence

M.J. and Holland,

yeast messenger

Holland,

J.L. and Hall, B.D.: Codon selection

Chem. 257 (1982) 3026-3031. Benoist,

Holland,

D.T., Scherer,

transformation

hybrid DNA molecules.

ofyeast:

S. and Davis, R.W.: High autonomous

replication

of

Proc. Natl. Acad. Sci. USA 76 (1979)

1035-1039. Sutcliffe, J.G.: pBR322 sequence:

accurate

restriction

map derived

DNA size markers

from the DNA

up to 4361 nucleotide

pairs long. Nucl. Acids Res. 5 (1978) 2721-2728.

214

Tschumper, fragment

G. and containing

gene. Gene Tschumper,

Carbon,

of a yeast

replicator

DNA

and the TRPZ

10 (1980) 157-166.

G. and Carbon,

centromere.

J.: Sequence

a chromosomal

J.: Copy number

Valenzuela, H.M. coding

control

by a yeast

antigen.

P., Gray, and Rutter,

P., Quiroya,

for the major Nature

protein

by J. Carbon.

J., Goodman,

sequence

of hepatitis

280 (1980) 815-819.

Gene 23 (1983) 221-232. Communicated

M., Zaldivar,

W.J.: Nucleotide

of the gene

B virus

surface