Glucose Oxidase from Aspergillus niger

THE JOURNAL OF BIOL.OGICAL CHEMISTRY 0 1990 by The American Society for Biochemistry

Vol. 265, No. 7, Issue of March 5, pp. 3793-3802,199O and Molecular

Printed in U.S.A.

Biology, Inc.

Glucose Oxidase from Aspergillus CLONING, GENE SEQUENCE, SECRETION FROM AND KINETIC ANALYSIS OF A YEAST-DERIVED

niger SACCHAROMYCES

CEREVISIAE

ENZYME* (Received

Katherine R. Frederick, Scott H. Chamberlain, From

the Chiron

Sumita From

Research

Chakraborty, the Department

James Tung, Amit Vasavada, Labs,

Chiron

Lawrence of Biological

Richard S. Emerick, Frank and Steven Rosenberg*

Corporation,

Emeryville,

M. Schopter,

Chemistry,

The

and Vincent

University

Glucose oxidase (@-D-glucose:oxygen l-oxidoreductase, EC 1.1.3.4) catalyzes the oxidation of (?-D-gluCOse to glucono-dlactone and the concomitant reduction of molecular oxygen to hydrogen peroxide: @-D-glucose + O2 + glucono+lactone + H202. The enzyme activity was first reported by Muller (1928) in extracts of Aspergillus niger, and subsequently, the enzyme has been purified from both Aspergillus (Pazur and Kleppe, 1964; Swoboda and Massey, 1965) and Penicillium species (Kusai et al., 1960). The fungal enzyme consists of a dimer of molecular weight 150,000 containing two tightly bound FAD cofactors (Pazur and Kleppe, 1964). The mecha-

of Michigan

October

16, 1989)

R. Masiarz,

94608

Massey Medical

School,

Ann

Arbor,

Michigan

48106

nism of action of the enzyme has been analyzed in some detail (Gibson et al., 1964; Bright and Appleby, 1969). Glucose oxidase has also been tested as the basis for glucose sensors (Degani and Heller, 1987, 1988), in glucose detection kits, and as a source of hydrogen peroxide in food preservation (Banks et al., 1986). As a first step in trying to elucidate the residues necessary for catalysis and to improve the properties of glucose oxidase by protein engineering techniques, the cloning and expression in yeast of the A. niger enzyme are described. MATERIALS

AND

METHODS

Protein Chemistry, Deglycosylation, and Analysis-Glucose oxidase from A. niger (EC 1.1.3.4) was obtained from Sigma (type 5). Sequence analysis was performed on the protein after dialysis into 0.2 M Nethylmorpholine acetate (pH 8.5) buffer containing 6 M guanidine HCl and 3 mM EDTA followed by reduction with P-mercaptoethanol and either carboxymethylation with iodoacetic acid (Truett et al., 1985) or pyridylethylation with 4-vinylpyridine (Freidman et al., 1970). Tryptic fragments of glucose oxidase were prepared by digestion with the enzyme in 2 M urea buffers after blockage of lysine residues with citraconic anhydride (Truett et al., 1985). The peptide mixtures were decitraconylated in acid and resolved by successive chromatography on Vydac Cl8 and C4 reverse phase columns under acidic conditions at elevated temperature (40 “C) and re-chromatography at room temperature under neutral conditions using gradients of acetonitrile. Cyanogen bromide fragments of the protein were resolved by successive chromatography using Bio-Gel P-10 in 30% formic acid and re-chromatography by reverse phase HPLC’ using Vydac Cl8 columns and n-propanol gradients in the presence of trifluoroacetic acid. Edman degradations were performed on the intact protein and peptide fragments using an Applied Biosystems 470A gas phase protein sequencer. The phenylthiohydantoin-derivatives were identified using either the reverse phase resolution system of Hawke et al. (1982) or that employed in the Applied Biosystems 120A phenylthiohydantoin analyzer (Hunkapiller, 1985). N-Linked carbohydrate was removed using either endoglycosidase H or N-glycanase (Boehringer Mannheim). SDS gel electrophoresis was done according to Laemmli (1970). Glucose Oxidase Assays-Glucose oxidase activity was determined using either A. niger glucose oxidase (Sigma, type 5) or purified yeastderived glucose oxidase as standard. The assay was a modification of the method of Kelley and Reddy (1986) as follows: assays were performed in a volume of 1.0 ml in 0.1 M NaP04 (pH 7) containing 0.2 mM o-dianisidine (Sigma), 10 rg of horseradish peroxidase (Boehringer Mannheim), and 9.5 mM D-gluCOSe. Assays were initiated by the addition of glucose oxidase (l-30 ng), incubated at room temperature for 20 min, and quenched by the addition of 0.1 ml of 4 N 1 The abbreviations used are: HPLC, high performance liquid chromatography; EndoH, endoglycosidase H; GO, glucose oxidase; kb, kilobase; PTH, phenylthiohydantoin; YEP, yeast extract/peptone; SDS, sodium dodecyl sulfate; bp, base pair(s).

3793

Downloaded from www.jbc.org by guest, on July 10, 2011

The gene for Aspergillus niger glucose oxidase (EC 1.1.3.4) has been cloned from both cDNA and genomic libraries using oligonucleotide probes derived from the amino acid sequences of peptide fragments of the enzyme. The mature enzyme consists of 583 amino acids and is preceded by a 22-amino acid presequence. No intervening sequences are found within the coding region. The enzyme contains 3 cysteine residues and 8 potential sites for N-linked glycosylation. The protein shows 26% identity with alcohol oxidase of Hansenuela polymorpha, and the N terminus has a sequence homologous with the AMP-binding region of other flavoenzymes such as p-hydroxybenzoate hydroxylase and glutathione reductase. Recombinant yeast expression plasmids have been constructed containing a hybrid yeast alcohol dehydrogenase II-glyceraldehyde-3phosphate dehydrogenase promoter, either the yeast (Yfactor pheromone leader or the glucose oxidase presequence, and the mature glucose oxidase coding sequence. When transformed into yeast, these plasmids direct the synthesis and secretion of between 75 and 400 pg/ml of active glucose oxidase. Analysis of the yeast-derived enzymes shows that they are of comparable specific activity and have more extensive Nlinked glycosylation than the A. niger protein.

* This work was supported by the Chiron Corporation and Labofina, s.a., Feluy, Belgium and by Grant GM11106 (to V. M.) from the National Institute of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 505242. $ To whom correspondence should be addressed: Protos Corp., 4560 Horton St., Emeryville, CA 94608.

California

for publication,

3794

Glucose Oxidase, Cloning and Expression the polyA tail was digested with Bg12 and Hd3, and the resulting mixture of fragments was subcloned between the BamHI and Hd3 sites of pBR322 yielding plasmid pBRX2a. This plasmid contains the entire GO coding sequence, the untranslated region, and approximately 2 kb of flanking XgtlO sequences. A second cDNA clone (X4a) containing most of the GO cDNA, but truncated immediately before the polyA addition site, was digested with EcoRI and a llOO-bp fragment, comprising the 3’-half of the GO coding sequence and the 3’-untranslated, was subcloned into the EcoRI site of pBR322 yielding plasmid pBRX4allOO. Plasmid pAGAP (Malcolm et al., 1989) was digested with NcoI and Bg12, and the synthetic duplex shown below as a single strand encoding the GO preprosequence was inserted, yielding plasmid pAGSGO-1. NC01 C*TGCAGACTCTCCTTGTCTcGAGccTTGTGGTcT &!I2 CCCTCGCTGCGCCCTGCCACACTACATAGATCT

The full-length GO cDNA plasmid pBRX2a was digested with SalI, and another synthetic duplex encoding amino acids 1-21 of the mature GO sequence was ligated to it. The ligase was heat-inactivated, and the mixture was treated with polynucleotide kinase and then digested with PstI. The digest was precipitated with ethanol, run on a 1% agarose gel, and the 980-bp Bgl2-PstI fragment, corresponding to the 5’-half of the mature GO coding sequence, was isolated. Plasmid pBRX4allOO was digested with EcoRI, treated with Klenow polymerase and the 4 deoxynucleotide triphosphates to create blunt ends, and then ligated with Bg12 linkers of the sequence GGAGATCTCC. The mixture was subsequently heated at 65 “C for 15 min and then digested with Bgl2 and PstI. The digest was run on a 1% agarose gel and the 915-bp PstI-BgQ fragment, corresponding to the 3’-half of the GO coding region and the 3’-untranslated sequence, was isolated. The two halves of the GO gene were then ligated together, the mixture digested with Bg12 and then ligated to plasmids DAGSGO-1 and DCBR which had been treated with 8212 and alkaline phosphatase. P&mid pCBR is analogous to pAGAP (Malcolm et al., 1989), but it contains, in addition, the yeast a-factor leader (Brake et al., 1984) inserted between the yeast alcohol dehydrogenase IIglyceraldehyde-3-phosphate dehydrogenase hybrid promoter (Cousens et al., 1987) and glyceraldehyde-3-phosphate dehydrogenase terminator sequences. The Bgl2 site in this plasmid lies at the sequence Lys-Arg-Ser corresponding to the cleavage site of the KEX2 enzyme (Julius et al., 1984). Restriction map analysis was used to isolate plasmids with the GO gene Bg12 fragment in the correct orientation for expression yielding plasmids containing the yeast alcohol dehydrogenase II-glyceraldehyde-3-phosphate dehydrogenase promoter, either GO or a-factor prepro sequences, the GO gene, and glyceraldehyde-3-phosphate dehydrogenase terminator as BamHI bounded expression cassettes. These were then excised and transferred to the BamHI site of the yeast-E. coli shuttle vector, pAB24 (Barr et al., 1987). These expression plasmids are pSGO2 (GO preprosequence) and paGO(a-factor preprosequence). Yeast Transformation and Growth-Yeast transformations were done according to Hinnen et al. (1978). The yeast strain GRFl81 (Mata, leu2-3, leu2-212, his3-11, his3-15, ura3A, CAN, [cir”]) was derived from GRF180 (Malcolm et al., 1989) by plasmid-directed deletion of the ura3 gene using 5-fluoro-erotic acid as selection for uracil auxotrophs (Boeke et al., 1984). Transformants were initially obtained on uracil selective plates containing 8% glucose and were then streaked onto minimal plates lacking leucine to select for the high copy number leu2-D allele (Beggs, 1978; Erhart and Hollenberg, 1983). Innoculae were grown in leucine selective media containing 8% glucose for 48 h and diluted 1:lOO into YEP medium containing 4% glucose for expression. Expression cultures (25 ml) were grown for 144 h at 30 “C and harvested by centrifugation. Secreted GO activity was determined after diluting media in 0.1 M NaPOl (pH 7), whereas intracellular enzyme activity was measured after lysis of the cells with glass beads (Rosenberg et al., 1984). Purification of Yeast-derived Glucose O&me--The yeast-derived enzymes were purified from the conditioned medium by a modification of the method of Pazur and Kleppe (1964). Yeast cells were removed by centrifugation, and the conditioned YEP medium was diluted lo-fold with 0.01 M sodium acetate (pH 4.5). This material was applied to a DEAE-Sepharose Fast Flow column (20 ml, Pharmacia LKB Biotechnology Inc.) equilibrated in the same buffer. The column was then washed with 3 volumes of the equilibration buffer and the enzyme eluted with 0.1 M sodium acetate (pH 3.7). Fractions


H,SO,. The reduced o-dianisidine was then measured at 400 nm on a Shimadzu Model UV-160 spectrophotometer or at 405 nm on an enzyme-linked immunosorbent assay reader (Titertek Multiscan). Enzyme amounts were calculated as nanograms of glucose oxidase relative to a standard curve of absorbance uersus enzyme amount. An alternate continuous spectrophotometric assay was also employed for kinetic characterization of the purified enzyme. This was a modification of a procedure described by Lockridge et al. (1972) for the assay of H,O, with lactate oxidase. In 1 ml of total volume was placed 0.8 ml of 0.3 M phosphate buffer (pH 5.6), 0.1 ml of 1 M glucose, 20 ~1 of horseradish peroxidase (1 mg/l ml), and 40 ~1 of a Triton-stabilized o-dianisidine solution. The latter was Drenared fresh daily by mixing 0.8 ml of 10 mM o-dianisidine with 0.2 ml of 20% Triton X-100. After equilibration at 25 “C, the reaction was started by the addition of glucose oxidase, and the absorbance increase at 460 nm followed with time. Molecular activity was calculated on the basis of EdM) of 11.6 mM-’ cm-’ for the o-dianisidine oxidation product (Worthington and Teller, 1972). Complete steady state analysis at pH 5.6,4 “C, was also determined by enzyme-monitored turnover, using a stopped flow spectrophotometer, and analzyed as described earlier (Gibson et al., 1964). The stopped flow instrument was interfaced with a Nova II (Data General) computer and has been described in detail earlier (Beaty and Ballou, 1981). Absorption spectra were determined with a Hewlett-Packard 8452A Diode Array Spectrophotometer equipped with a thermostatted cell holder. Growth of A. niger and Isolation of Nucleic Acids-A. niger (ATCC strain 9029) was grown in YEP medium (1% yeast extract, 2% peptone; Difco) containing 2% glucose at 25 “C. RNA was isolated by the method of Chirgwinet al: (1979) as modified by Turpen and Griffith (1986) from mvcelia of an overnight culture of strain 9029 grown to a density of i g/liter. Enrichment of poly(A+) RNA was obtained by chromatography on oligo-dT-cellulose (Aviv and Leder, 1972). Genomic DNA was isolated from mycelia grown to 40 g/liter in YPD as described (Boel et al., 1984). Construction of cDNA and Genomic Libraries-All recombinant DNA manipulations were done essentially according to Maniatis (1982). The cDNA library was constructed in the XgtlO vector (Huyn et al., 1985) using poly(A+) RNA which had been purified by two cycles on oligo-dT-cellulose as starting material (Aviv and Leder, 1972). A total of 8 x lo5 clones were obtained. The genomic library was prepared as follows: A. niger genomic DNA (50 pg) was digested with 0.5 units of Sau3a (New England Biolabs) for 50 min to maximize DNA in the molecular weight range of 5-20 kb. The digest was resolved on a 1% agarose gel and the region corresponding to 7-10kb fragments was excised; the DNA was electroeluted, extracted with phenol/chloroform, and concentrated by ethanol precipitation. This material was ligated with pBR322 which had been digested with BamHI, treated with calf intestinal phosphatase (Boehringer Mannheim), and isolated from an agarose gel. The ligated DNAs were used to transform Escherichia coli DH5 (Bethesda Research Labs; F-, recA1, endAl, hsd R17(rk-, mk-) supE44 l-, thil-, gyrA, relA1) to ampicillin resistance. A total of 7000 transformants were obtained, subsequent analysis showed that >90% of the colonies were recombinant and the average insert size was 8 kb. Library Screening and DNA Sequencing-All oligonucleotides were made using standard phosphoramidite chemistry on an Applied Biosystems Model 380A DNA synthesizer or as described (Warner et al., 1984). Oligonucleotides were 5’-end-labeled with T4 polynucleotide kinase (New England Biolabs) and [-y-32P]ATP (Amersham Carp). DNA fragments were labeled using Amersham Corp. nick translation or random priming kits (Feinberg and Vogelstein, 1983). Oligonucleotide screening of the cDNA library was done using the tetramethylammonium chloride washing method of Wood et al. (1985). Tetramethylammonium chloride was from Aldrich. Nick translated or random primed probes were hybridized to duplicate filters at 42 “C in Wallace mix (Wallace et al., 1979,198l) containing 50% formamide and 10% dextran sulfate. Potential positives were replated and screened again with the same probe and then were analyzed using Southern blotting as described (Maniatis et al., 1982). Digests of hgtl0 clones were directly subcloned into plasmid vectors-(either pBR322 or M13) and positives identified by colony hybridization. DNA sequencing was done by subcloning in Ml3 using the method of Sanger et al. (1977). Construction of Yeast Expression Plasmids-All restriction enzymes and other reagents foi recombinant DNA manipulations were obtained from New England Biolabs unless otherwise noted. A GO cDNA clone in XgtlO (Xia) containing the entire coding sequence and

Glucose Oxidase, Cloning and Expression

The sequences cleavage method

3795

TABLE I Peptide sequences derived from glucose oxidase of peptide fragments of glucose oxidase determined as described used to produce them.

sequence

in the text

are shown

Location”

1. SNGIEASLLT 2. TVDYIIAGGGLTGLTTA 3. LTENPN’ISVLVIESGSYESD 4. GGFHX*TTALLIQYENY 5. PKE 6. SAVEDRGVPTKKDFGQdGDPHGVS 7. FPNTLUEDQV B.‘ISDAILEDYASM

with

Cleavage

l-10 ‘I-9-A= L%D= @Ky3= p525-E52,

the

method

None Trypsin Trypsin Trypsin CNBr CNBr CNBr Trvnsin

’ The locations indicated are derived from translation of the cDNA sequence. “These positions are potential N-linked glycosylation sites. The asparagine at position 43 was present in reasonable quantity which suggests that it is not glycosylated, position 5 in peptide 4 yielded a blank cycle in the Edman degradation, suggesting that the asparagine in that position (388) is a site for glycosylation. ’ The difference at position 206 is due to an error in amino acid identification due to the use of the Applied Biosystems 120A resolution system to identify the PTH-derivative derived from a carboxymethyl peptide; during the course of this study, the PTH-derivative resolution system was changed from that of Hawke et al. (1982) to the on-line 120A PTH-analyzer system of Applied Biosystems (Hunkapiller, 1985). ‘-‘These sites showed differences between the amino acid sequence shown and that derived from the cDNA: 206 is CZ” and L22O is H**‘L

Q

‘Peptide 8 is derived from a cleavage intact protein; K57” was not susceptible

with trypsin which was obviously not precluded by citraconylation to chemical modification under the conditions employed in this

of the study.


FIG. 1. The sequence of the A. niger glucose oxidase gene. The DNA sequence derived from both cDNA and genomic clones is shown. The TATAA and AAACAA sequences in the 5’- and 3’-untranslated sequences are indicated by a solid line. The translation of the entire sequence is shown beginning at amino acid -22 with respect to the N terminus of the mature protein. Potential sites for N-linked glycosylation are indicated by the letters CHO.

containing GO activity were pooled and concentrated by ultrafiltration. Analysis of Protein Sequences-Computer analysis of protein sequences was done initially with the program DFASTP (Lipman and Pearson, 1985) to screen the Dayhoff protein sequence database. Further analysis and alignment was done using the program MA-

LIGN (Biomathematics Computation Lab, San Francisco) and by visual examination.

University

of California,

RESULTS Determination

Oxidase-Glucose

of

Peptide

Sequences

oxidase from A. niger

from

A.

migrates

niger

Glucose

as a protein

3796

Glucose Oxidase, Cloning and Expression A 12

3

4

5

6 -23 -

9.7 6.6

-

23 20

-

EcoRl

1353 Hd3

2783

.EcoRl

3380

Xhol

3980

with an apparent molecular mass of 75-80 kDa on SDSpolyacrylamide gel electrophoresis. Treatment with either Nglycanase or endoglycosidase H decreases the apparent molecular mass by 5-10 kDa, suggesting that a few sites for Nlinked glycosylation are used, consistent with the observations of others (Pazur et al., 1965, Swoboda and Massey, 1965). When the protein is subjected to automated Edman degradation, the N-terminal sequence: Ser-Asn-Gly-Ile-Glu-AlaSer-Leu-Leu-Thr is found. Treatment with cyanogen bromide or digestion with trypsin after treatment with citraconic anhydride, and isolation of individual peptides by HPLC followed by gas phase sequencing, yielded the collection of sequences shown in Table I. Isolation of cDNA and Genomic Clones Encoding Glucose Oxidase-CNBr peptide 6 was used to design two long unique probes for the glucose oxidase gene. The probes, long7 and long& were designed with a GC bias in the third position as has been found for other filamentous fungi genes (Corrick et al., 1987). long’i GCTCACACCGTGGGGATCACCCTGGCCGAAATCTTTCTTGGT long8

ATCTTTCTTGGTGGGCACGCCGCGATCCTCCACAGCGCTCAT

These probes were used initially to screen duplicate filters of a cDNA library of 8 X lo5 recombinant phage in XgtlO. Four double positives were obtained which upon secondary screening remained positive. Subsequent DNA sequence analysis of one of these phage, X4a, showed 33142 matches with probe long7 and 36/42 matches with probe long8. The single open reading frame agreed with the peptide sequence in all positions except one, Cys instead of Gln, at position 17. In addition, the sequence of CNBr peptide 7 (Table I) was encoded immediately 3’ of the peptide 6 match. Three cDNA clones were sequenced on both strands in

their entirety; all of the overlapping sequences, covering more than three-fourths of the coding sequence, were the same. The longest clone extended 22 bp beyond a likely start codon, whereas a second stopped at that methionine codon. A 600bp EcoRI-NcoI fragment from one of the cDNA clones was then used to screen a genomic library in pBR322. Several positive clones were obtained and one was analyzed in detail. A combination of DNA sequence and restriction analysis showed that the genomic sequence was contiguous with the cDNA clones, indicating there are no intervening sequences in the coding region. A composite of the coding region from the cDNA clones and the 5’- and 3’-flanking regions from the genomic clone is shown in Fig. 1. The indicated methionine starts an open reading frame comprised of 605 amino acids, encoding a potential protein of 65,700 molecular weight. The mature N terminus as determined by protein sequencing is found at amino acid 23, suggesting the first 22 amino acids are proteolytically removed. This sequence has some characteristics of a signal peptide, but is likely more complex, since the final processing is at a single arginine residue. The mature protein of 583 amino acids contains 3 cysteine residues and 8 consensus sites for N-linked glycosylation. All of the peptide sequences in Table I are found in the cDNA sequence with only three differences out of 117 residues and one difference due to a known resolution artifact. These differences are likely due to incorrect assignment of amino acids during peptide sequencing. To determine if there are multiple GO genes and to confirm the genomic clone structure, the entire cDNA was used to probe a genomic blot of A. niger DNA. The results (Fig. 2A) show that GO is most likely encoded by a single gene and is consistent with the map shown of the genomic clone (Fig. 2B).


FIG. 2. A, southern blot of A. niger DNA probed with a glucose oxidase gene probe. B, map of the A. niger glucose oxidase gene. Aliquots of A. niger DNA (10 pg) were digested with a series of restriction enzymes, electrophoresed on a 1% agarose gel, and transferred to nitrocellulose. A 1.9-kb Bg12 fragment of the GO gene from plasmid pSG02 was labeled with Alp by random priming, hybridized to the blot, washed at 60 “C in 0.1 X SSC (8.75 g NaCl, 4.4 g sodium citrate/liter, pH 7.0), 0.1% SDS, and exposed to film for 16 h. The map shown was derived by standard methods using restriction analysis and Southern blotting. The solid bar represents the region of the cDNA clones isolated and the arrow the direction of translation. The digests are: lane 1, EcoRI; lane 2, Hind3; lane 3, BarnHI; lane 4, SalI; lane 5, PstI, lane 6, XhoI.

3797

Glucose Oxidase, Cloning and Expression ind3

TABLE

II

Glucose oxidase produced by recombinant yeast Transformants were grown in YEP medium containing 4% glucose for I44 h at 30 “C at 300 rpm. Extracellular activity was determined directly on conditioned media. Intracellular activity was measured after lysis of cells with glass beads in 0.1 M NaPOl (pH 6.8). Glucose oxidase activitf paGO 1 (16.l)b

pSGO2

plasmid 2 (17.3)b

1 (22.V

plasmid 2 (17.l)b

dml

GO

pSGO-2

Extracellular Intracellular

90 3

83 1.8

233 57

379 42

Total % secreted

93 97

85 98

290 80

421 90

’ Glucose oxidase activity was determined using a purified sample of yeast derived glucose oxidase as standard. b Transformant is indicated by 1 or 2; the numbers in parentheses are the optical densities at 650 nm.

GO

\

p@GO-I

FIG.

oxidase

3. Structure in yeast.

“Materials

of plasmids

for

the

expression

of glucose

Plasmids were constructed as described under and Methods.” The first nlasmid. nSG02, contains the

yeast alcohol dehydrogenase II-glyceraldehyde-i-phosphate dehydrogenase promoter, glucose oxidase presequence, GO coding sequence, and glyceraldehyde-3-phosphate dehydrogenase terminator inserted into the BamHI site of the yeast-E. coli shuttle vector pAB24. The second plasmid, potGO-1, has the yeast a-factor leader substituted for the GO presequence.

Expression of the Glucose Oxidase cDNA in S. cerevisiaeWe inserted the mature GO coding sequence into two vectors for expression in Saccharomyces cerevisiae. These vectors contain expression cassettes of the yeast alcohol dehydrogenase II-glyceraldehyde-3-phosphate dehydrogenase promoter for regulated transcription (Cousens et al., 1987), the a-factor or GO preprosequences for secretion and/or protein sorting, the GO cDNA, and the yeast glyceraldehyde-3-phosphate dehydrogenase terminator. As shown in Fig. 3, the cassettes are inserted in the BamHI site of the yeast-E. coli shuttle


ind3

vector pAB24 (Barr et al., 1987). When grown under derepressing conditions, transformants of these plasmids in yeast strain GRF181 secrete large amounts of active GO into the medium. Typical shake flask data are summarized in Table II. The levels observed (~300 pg/ml) of secreted GO activity from the pSGO2 transformants are among the highest observed for secreted proteins from yeast (Pentilla et al., 1988, Tschopp et al., 1987). Characterization of Yeast-derived Glucose Oxidase and Amino Acid Composition and Sequence Data-In order to begin to examine the GO secreted from yeast, we analyzed samples from both pSG02 and paGOtransformants on 8% polyacrylamide gels. The results (Fig. 4) show that both yeast derived proteins migrate more slowly than the A. niger enzyme. Treatment with endoglycosidase H, however, results in the yeast and A. niger enzymes migrating as doublets of similar if not identical mobilities. Thus, the yeast proteins have more extensive N-linked glycosylation. The yeast-derived enzymes were concentrated and purified using a modification of the method described for the A. niger protein (Pazur and Kleppe, 1964). They were then subjected to Nterminal sequence and amino acid composition analysis. The enzyme purified from paGOconditioned medium had the same serine N terminus as the A. niger enzyme. Analysis of the pSGO2 derived enzyme showed a majority of the authentic N-terminal serine and a minor sequence: Leu-Pro-X-Tyr-XArg-Ser-Asn-Gly, which corresponds to enzyme which has only been processed at position -6. This is the most likely signal peptidase cleavage site as suggested by the analysis of von Hejne (1985). Both enzymes had amino acid compositions in agreement with that predicted from the cDNA and observed previously by other workers (Pazur et al., 1965). The pSG02derived enzyme was utilized for detailed kinetic analysis. Spectral Characteristics-The absorption spectrum of the pSGO2 yeast-derived enzyme is shown in Fig. 5 and has the same characteristics as those reported for the enzyme isolated from A. niger (Swoboda and Massey, 1965), with X,., of 278, 382, and 452 nm and absorbance values in the ratio of 12.7:0.92:1.0. However, whereas the wild type enzyme was reported to have an extinction coefficient at 452 nm of 14.1 the pSG02 yeast-derived enzyme has an Ed52 mM-’ cm-‘, value of 12.83 mM-’ cm-‘. This was determined from the change in absorption spectrum due to release of FAD from denatured protein on diluting native enzyme into guanidinium HCl to give a final concentration of 7.6 M guanidine HCl (pH 7.3) and a determined value of Ed50 of FAD under the same conditions of 12.05 mM-’ cm-’ (results not shown). As with

3798

Glucose Ozidase, Cloning and Expression M

1

2

3

4

5

6

7

8

9

10

11

0.24

12

-Oxidized

1

Enyme

0.20

200g e

a

0.16 0.12

0.2

43-

FIG. 6. Enzyme-monitored cose oxidase. Enzyme (9.35 M phosphate (pH 5.6), 3 “C, of glucose in the same buffer and the reaction followed at path length of the stopped-flow

-

--

-

---

c

400

Wavelength

500

600

(nm)

FK;. 5. Absorption spectra of yeast-derived glucose oxidase. EFI,,,, oxidized enzyme in 0.1 M phosphate (pH 5.6), 25 “C; EFlSO:, after addition of 20 mM NaHSO:$; EFl,,dH-, after subsequent addition of 20 mM glucose (the same spectrum is obtained in the absence of bisulfite).

the wild type enzyme, the pSG02 yeast-derived enzyme is devoid of flavin fluorescence and does not appear to contain the blue fluorophore which has been reported to be present in the enzyme isolated from A. niger (Swoboda and Massey, 1966a, 1966b). Kinetics Analysis-The pSG02 yeast-derived enzyme, like the wild type, has very high catalytic activity. Under the standard continuous spectrophotometric assay conditions described under “Materials and Methods,” at pH 5.6, 25 “C, the yeast-derived enzyme has a turnover number of 17,000-20,000 min-‘, compared with a value of 16,200 min-’ reported for wild type enzyme (Gibson et al., 1964). It also gives a series

0.6

0.6

Time

(set)

1 .o

1.2

turnover

of yeast-derived glumixing) in air-equilibrated 0.1 was mixed with different concentrations to give the final concentrations shown, 450 nm. The AlsO scale is for the 2-cm spectrometer. pM after

of parallel Lineweaver-Burk plots on systematic variations of both the glucose and OZ concentrations. The latter results were obtained by stopped flow enzyme-monitored turnover as shown in Fig. 6, where 9.35 FM enzyme in 0.1 M phosphate (r)H 5.6) 3 “C! was reacted with various concentrations of glucose at an initial O2 concentration of 456 /*M (air saturated at 3 “C). Immediately after mixing, a steady state is established, determined by the relative rates of reduction of the enzyme flavin by glucose and oxidation of reduced flavin by oxygen. The steady state persists until practically all the oxygen is exhausted by catalytic turnover, whereupon the enzyme becomes fully reduced. Analysis of these curves as described by Gibson et al. (1964) yields the set of parallel Lineweaver-Burk plots shown in Fig. 7 (left-hand panel), and a replot of the intercepts uersus l/[oxygen] gives the secondary Lineweaver-Burk plot of Fig. 7 (right-handpanel). These data yield the following kinetic constants: V,,,,, = 500 s-‘, K m(glucoaej= 0.14 M, K,,,c,,ySenj = 2.6 X low4 M. Under similar conditions, but at 0 “C, the corresponding values for the A. niger enzyme have been reported: V,,,., = 235 s-l, Kmcglucosej= 0.12 M, Kmcoxygen~= 2.1 x 1O-4 M (Gibson et al., 1964). While the V,,,,, value determined for the yeast-expressed enzyme appears to be significantly higher than the wild type enzyme, it must be emphasized that the steepness of the secondary plots (cf. Fig. 7, right-hand panel) can lead to significant errors in the estimation of the extrapolated l/V,,,,,. Similar enzyme-monitored turnover analyses have been performed using 2-deoxyglucose, a much poorer substrate for the enzyme. With this substrate, much closer agreement is obtained with the values reported previously for the wild type enzyme. The following values were obtained at pH 5.6, 3 “C, with values reported for the wild type enzyme (Gibson et al., 1964) being given in brackets: V,,,,, = 20 s-l (15 s-l), K,,,~P.~eoxyC~ueose~ = 2.4 X lo-:’ M (2.5 X lo-” M). With 2-deoxyglucose, in both cases the K, for On was so low as to be difficult to measure with accuracy (-2 x lo-” M). It should be noted that in these enzyme-monitored turnover experiments it was not necessary to add KCN to inhibit contaminating catalase (as was done with the wild type enzyme; Gibson et al., 1964), since the yeast-expressed enzyme is completely devoid of catalase activity.

Formation of Enzyme Flavin N-(5)-Sulfite Complex-One of the characteristics of flavoproteins of the oxidase class, first recognized with glucose oxidase (Swoboda and Massey, 1966a) is the ability to undergo nucleophilic attack by sulfite (or bisulfite) to form a flavin N-(5)-sulfite adduct in a thermodynamically reversible equilibrium (Massey et al., 1969, Muller and Massey, 1969). The yeast-derived glucose oxidase, as expected, also displays the same properties, and the spec-


FIG. 4. SDS-polyacrylamide gel electrophoresis of glucose oxidase samples with and without treatment with endoglycosidase H. Plasmid transformants of yeast strain GRFlSl were grown and samples assayed as described under “Materials and Methods.” lml samples of supernatant fractions were precipitated with trichloroacetic acid containing 2 mg/ml deoxycholate after denaturation of the samples by heating at 100 “C for 5 min as described by Swoboda and Massey (1965). These samples were redissolved in 75 ~1 of 0.5 M citrate (pH 6) after washing the pellets three times with acetone. To each was added 75 ~1 of water, 20 ~1 of 0.1% SDS, and they were heated at 65 “C for 15 min to dissolve the pellets. 2 ~1 of phenylmethanesulfonyl fluoride added. The samples were split in half and to one set 4 ~1 (4 milliunits) of EndoH was added. Both sets were incubated O/N at 37 “C. They were then trichloroacetic acid-precipitated, and the equivalent of 0.1 ml of yeast culture supernatant was loaded for the EndoH treated samples and 0.2 ml for the untreated controls on an 8% gel which was subsequently stained with Coomassie Blue. The samples are: lanes l-6, no EndoH; lane 1, pAB24 vector; lane 2, 5 pg of A. niger GO; lanes 3 and 4, pnGO-1; lanes 5 and 6, pSG02; lanes 7-12, +EndoH; lane 7, pAB24; lane 8, 2.5 fig of A. niger GO; lanes 9 and IO, paGO-1; lanes II and 12, pSG02.

0.4

Glucose

FIG. 7. Kinetic analysis of enzyme-monitored turnover experiments. Left-hand graph, LineweaverBurk plots of the data of Fig. 6 (and a similar experiment with 12.5 mM glucose) analyzed as described by Gibson et al. (1964). Right-hand graph, secondary plot showing the intercepts from the primary plot as a function of the reciprocal of the glucose concentration.

3 s P T-

Oxidase,

Cloning

3799

and Expression

25 mM 0.02-

I 5,000

10,000

14021

EFlSO,

JZFl,,

+ glucose

% EFl,, k on %

EFl,~H-

+ HSO;

(1)

+ glucono&lactone

(2)

As Kred is known to be very large (Gibson et al., 1964), it was possible to obtain an estimate of k,rr by following the spectral change with time. This was done by adding 20 mM glucose to the enzyme-sulfite complex and following the reaction by taking repetitive scans at 10-20-s intervals with a diode-array spectrophotometer. The spectral change occurred with a tl,* of 90-100 s, independent of the initial bisulfite concentration (5-20 mM), yielding a tentative value for K,,rr of 7.3 X 10e3 s-‘. This was confirmed by measuring the rates of formation of the enzyme-sulfite complex at low concentration of bisulfite, again using the diode-array spectrophotometer to record spectra every 3 s until equilibrium was reached. At concentrations of bisulfite of 0.1, 0.2, 0.3, and 0.4 mM, the observed pseudo first order rate constants were 1.41 X lo-‘s-i, 2.04 X lo-’ s-‘, 2.57 X low2 s-l, and 3.33 X lo-’ s-‘, respectively. When these koobsvalues are plotted uersus the bisulfate concentration, a linear plot is obtained (cf. Strickland et al., 1975) with slope = k,, = 66 M-’ s-l and intercept = k,rr = 7.3 X 1O-3 s-‘. Thus, Kd = k,fJk,, = 1.1 X 10m4 M, in excellent agreement with the values obtained from static equilibrium titration measurements. In addition, it should be noted that the values of k,,ff determined by the two methods described above are also in remarkable agreement and provide convincing evidence for the validity of Equations 1 and 2. DISCUSSION The gene for A. niger glucose oxidase has been cloned from both cDNA and genomic libraries in this study. This conclusion is based on the following: first, the sequences of several

(M-l)

20

I 40

l/ [Glucose]

I

I 60

I

I 60

(M-1)

peptides isolated from the protein are found encoded in the same translational reading frame in the gene sequence. Second, the single open reading frame encodes a protein of 605 amino acids, consistent with the mass of glucose oxidase of -75 kDa, as determined by SDS gel electrophoresis and sedimentation analysis (Kelley and Reddy, 1986; Jones et al., 1982). The native molecular weight of 150,000-185,000 determined by others (Swoboda and Massey, 1965; Pazur and Kleppe, 1964) is consistent with the protein being a dimer of identical subunits, each containing a FAD cofactor. Third, these clones have been used to engineer yeast strains to secrete proteins having glucose oxidase activity which react with antibodies to the Aspergillus protein on Western blotting (this work) .2 We have sequenced a total of 2821 bp of Aspergillus DNA including 485 bp of 5’-untranslated and 517 bp of 3’-untranslated sequence. Two cDNA clones covering the coding region were sequenced in their entirety. The 463-bp 5’ of the longest cDNA and 365-bp 3’ of the polyA addition site were determined from a genomic clone. The coding and flanking sequences show substantially different base compositions; the coding region is 56% G:C in keeping with other fungal genes (Innis et al., 1985), whereas the 5’-untranslated is 52% G:C and the 3’-untranslated is 60% A:T. The putative promoter sequence contains a TATAA sequence, at -59 with respect to the longest cDNA clone isolated, and at -81 with respect to the initiation codon. The sequence between the TATAA and the likely ATG is highly pyrimidine rich (74%). Similar regions are seen in the promoter regions of highly expressed yeast genes (Dobson et al., 1982). The assignment of the initiating ATG is based upon the A at -3 (Kozak, 1984) and the fact that the subsequent 15 amino acids comprise a likely signal peptide sequence (von Hejne, 1985). In addition, there are no other in-frame ATGs within 60 bp. Two sets of direct repeats are found in the 5’-untranslated region: GGATTAT at -245 and -232 and GGAGGATG at -194 and -152. The 3’-flanking region does not contain a consensus polyA addition site (AATAAA). A comparison of the sequence of a cDNA and genomic clone shows that the GO mRNA can be polyadenylated at the sequence AAACAA at +151. Other fungal genes such as ulcR of Aspergillus niduluns (Felenbok et al., 1988) do not contain a canonical polyA site. The sequence between the TGA and polyA site contains several repeated sequences: A/T GGGGA at +9,31, and 51 and AGGTTACAT at +86 and 98. An additional inverted repeat TTGGTAG: CTACCAA (+81 and 147) is of possible interest because it overlaps the polyA addition site at +151. The possible role of * K. Frederick,

unpublished

data.


trum of the N-(5)sulfite adduct is shown in Fig. 5. From the spectrophotometric titration of the enzyme with bisulfite at pH 5.6, 25 “C, Kd values of 1.06 X 10m4 M and 1.23 X 10T4 M were obtained in two separate experiments, values somewhat lower than that reported for wild type enzyme of 2.3 x 10e4 M under the same conditions (Swoboda and Massey, 1966a). The Kd value was confirmed by kinetic measurements. As shown in Fig. 5, in agreement with results initially reported for A. niger enzyme purified from A. niger (Swoboda and Massey, 1966a), the addition of glucose to the enzyme-sulfite complex leads to the spectrum of fully reduced enzyme. As glucose is only able to reduce the free oxidized enzyme, the spectral change shown presumably involves the dissociation of the enzyme-sulfite complex followed by rapid reduction of the free oxidized enzyme.

0

15,000

3800

Glucose

Oxidase,

and Expression

Cloning GO A0

10 20 30 40 50 .____ SNGIEASLLTDPKDVSGRTVDYIIAGGGLTGLTTAARLT--ENPNISVLVIESGSYESDR : :..::: :: ,:.::_ . . . . . . . .::.:_ MAIPDEFDIIWGGGSTGCCIAGRLANLDDQNLTVALIEGGENNINN 10 20 30 40

.,

60 70 80 90 100 110 GO G-PIIEDLNAYGDIFGSSVDHAYET-VELATNNQTAIRSGNGLGGSTLVNGGTWTRPHK .. ..:... :.. : :., :....,: :::.. _: .::. A0 PWVYLPGVYPRNMRLDSKTATFYSSRPSKALNGRWlIVPCANIIL;GGSSINFLMYTRASA 50 60 70 80 90 100 GO A0

120 130 140 150 160 170 AQVDSWETVFGNEGWNWDNVAAVAAYSLQAERARAPNAKQIAAGHYFNASCHGVNGTVHAGPR :.:: .:::. sDYDDWE----SEGWS~----------TDELL~L~~~~~~---~~R~~~~R~LHGF~~~~ 110 120 130 140

180 190 200 210 220 230 DTG.-DDYS-PIVKALMSAVEDRGVPTKKDF~GCGDPHGVSMFPNTLHED-QVRSDAAREW __:_ :, . . . . . . . . . . . . . . .:. __::__ : :::.:... A0 KVSFtiNYTYPTCQDFLRAAESQGIPVVDDLEDFKTSHGAEHWLKWINRDLGRRSDSAHAY 150 160 170 180 190 200

GO

240 250 260 270 280 LLPNY-QRPNLQVLTGQWGKVLLSQ-NGTTPRAVGVE-FGTHKGNTH~AKH~I :. . . . . . _.:. .::.... :.: __..: ,,. : A0 VHPTMRNKQSLFLITSTKCDKVIIEDGKAVAVRTVPMKPLNPKKPVSRTFRARKQIVISC 210 220 230 240 250 260 GO

290 300 310 320 330 340 GSAVSPTILEYSGIG~SILEPU;IDTVVDLP-VGLNLQDQTTATSRITSAGAG~GQA :. :: .:. :::: : ..:....:::: :: :.::. A0 GTISSPLVLQRSGIGAAHHLRSVGVKPIVDLPGVGENFQDHYCFFTPY~VKPDVPTFDDF 270 280 290 300 310 320 co

350 360 370 380 390 400 AWPATFNE--TFGD~YSEKAHELLNTKLEQWAEEAVARGGFHNTTWIQ-YENY----R . .:,. ::.:_ : .:. :..: A0 VRGDPVAQKAAFDQWYSNKDPLTTNGIEAGVKIRPTEEELATADEDFRRGYAEYFENKP 330 340 350 360 370 380 GO

410 420 430 440 450 DWIVNHNVAYSELFLDTAGVASFDVWDLL--PFTRGWHILDKDPYLHHFAYDPQYF : ___..: ::.::,:,: A0 ~KPLM~YSVI~G,,G,H,K;,NGKFMTMFHFLEYPFSRGF 390 400 410 420 430 440 GO

460 470 480 490 LNELDLLGQAAATQLARNISNSGAMQTYFAGETIPGDNL-------------------AY .: :: : .:..... :__ ::::... : A0 NDERDLWPMWJAYKKSRETARR--MES-FAG~TSHHPL-F~SPA~RDLDLETCSAY 450 460 470 480 500 490 GO

GO A0

500 510 DRDLSA--------WTE-----------------------------~~~~~~-y~pyHFR :: :: : AGPKHLTANLYHGSWTVPIDKPTPKNDFKVTSNQVQLHSDIEYTEEDDEAIVNYIKEHTE 510 520 530 540 550 560 520

GO A0

GO A0

::

530

540

550

560

PNYHGVGTCSMMPKEM------GGVVDNAARWGVoGLRVIDGSIPPTQMSSHVMTVFYA _: . '...' ::: :::_: ::::::.:::.:. . . TTWHCLGTCSMAPREGSKIAPKGGVLDARLNWGVQNLKDLSVCPDNVGCNTYSTALT 570 580 590 600 610 620 570 580 MALKISDAILEDYASMQ :._ . ...: IGEKAATLVAED 630

FIG. 8. Homology polymorpha.

of glucose

oxidase

with

alcohol

oxidase

of

The sequences of the two proteins are shown in the single letter code. The original alignment was done using the program DFASTP; the symbol: indicates an identity, whereas.is a conservative replacement. The final alignment shown was done using the program MALIGN which incorporated the insertion in the alcohol oxidase sequence from 485 to 558. The major regions of homology described in the text are indicated with solid lines. H.

The N-terminal homology is clearly a representative of the pa-8 motif involved in AMP binding in p-hydroxybenzoate hydroxylase and human glutathione reductase (Hofsteenge et al., 1980; Wierenga et al., 1979; Thieme et al., 1981; KrauthSiegel et al., 1982). An alignment of eight flavoproteins in this region is shown in Fig. 9. The glycines at positions 26, 28, and 31 in GO are analogous to those pointed out by Hofsteenge and co-workers (1980) for p-hydroxybenzoate hydroxylase and are conserved, as is E4’, which may be hydrogenbonded to the 2’-OH of ribose as in p-hydroxybenzoate hydroxylase. The substantial degree of homology of GO with glutathione reductase and p-hydroxybenzoate hydroxylase in this region suggests that the AMP portion of the FAD cofactor binds near the N terminus of the GO subunits. The second


these repeats in mRNA stability or polyA site selection awaits further work. The amino acid composition of glucose oxidase derived from the cDNA sequence is in close agreement with that measured by other workers (Pazur et al., 1965; Jones et al., 1982) and with that determined on the proteins expressed in yeast (this work). This and other observations require that glucose oxidase be a dimer of identical subunits. The presequence is somewhat similar to that found for A. niger glucoamylase, except that a single Arg residue is found prior to the mature sequence as opposed to a dibasic Lys-Arg sequence (Innis et al., 1985). The enzyme which processes the GO presequence in yeast is unknown. Several groups have studied glucose oxidase based upon its properties as an acidic (polyanionic) protein. These include the effect of ionic strength on catalysis (Voet et al., 1981), inhibition by putrescine (Voet and Andersen, 1984), and resistance to inactivation by SDS at neutral pH (Jones et al., 1982). From the cDNA sequence a charge of -58 on the glucose oxidase dimer at neutral pH is predicted, ignoring any ionic contributions from histidines, tyrosines, and the carbohydrate. Previous work measured a charge of -77 by titration using an isoelectric point of 4.05 (Voet et al., 1981). The difference between this result and that predicted from the cDNA sequence is unknown, but may be due to a slightly higher isoelectric point of 4.2 reported by others (Swoboda and Massey, 1965) or to contributions from the FAD cofactor and covalent non-cofactor phosphate (James et al., 1981). The mature enzyme contains eight potential N-linked glycosylation sites and 3 cysteine residues. Pazur and co-workers (Pazur et al., 1964) isolated a glycopeptide from GO by Pronase digestion and determined its amino acid composition. The closest correspondence to their results in the cDNA sequence is at positions 88-92 (Asn-Asn-Gln-Thr-Ala). From sequencing of peptide fragments of GO (Table I), we found that Asn-43 is not glycosylated, whereas Asn-388 likely is modified. Thus, at least two of the eight potential sites are utilized in Aspergillus. Of the potential sites, five contain the sequence Asn-X-Thr and three the alternative Asn-X-Ser. Although the data are very limited, the former sequence may be preferred in Aspergillus, as it is in the yeast, S. cereuisiae (Moehle et al., 1987). Swoboda and Massey (1965) showed clearly that GO contains two disulfide bonds/dimer and they suggested that only a single free cysteine was present, making a dimer of identical subunits unlikely. The results presented here show that if there are two disulfide bonds then there are 2 free cysteines/ dimer, since a single gene encodes the two subunits. A comparison of the mature GO sequence with the Dayhoff protein sequence database yielded a single homologous protein, alcohol oxidase of Hansenuela polymorpha. The initial scan showed 23% identity over almost the entire GO sequence (122/583 amino acids). Subsequent work showed this to be an underestimate as a highly homologous region near the C terminus of the proteins was missed, due to a large insertion of about 70 amino acids in the alcohol oxidase as compared to the GO sequence. The best alignment of the sequences is shown in Fig. 8. The three overlined regions of the proteins show relatively high degrees of sequence identity: amino acids 21-52 versus 9-41 (48%); 294-328 versus 274-308 (53%); and 519-556 versus 571-614 (51%), where the former numbers indicate the GO sequence and the latter those of alcohol oxidase. Using this alignment, the proteins are 26% identical (148/583). Subsequent analysis of these regions shows that two are recognizably related to motifs found in other flavoenzymes.

Glucose

Oxidase,

Cloning

Sequencer and Predicted Secondary Slruclures

PrOkill ._.. p

..

GO

VDYIIAGGGL”

AOh

FDIIVVGGGSTGCCIA

. .

. a ....

. . .

. ..

p . . . . . ~.~~

GLTTAARLTE~~~ISVL--VIE GRLANLDDQNLTVALIE

AOP

F3ILVLGGGFSGSCISGRLANLDHSLKVG-LIE

Giux

YDYLVIGGGSGGLASARRAAELGARAA---VVE

GRDZ

YUYIAIGGGSGG;ASINRRAMYGQKCA---LIE

DAAOX

MR”““IGAGV~GLSTALCIEERYnSVLO--PLD

pHBH

TQVAIIGAGPSGLLL

LPDH

TQ”“VLGAGPAGYSAAFRCACLGLFTV---IVE

;C,LHKAGIND”----LE

Protan

SUplNX 521

GO

NYHGVGTCSW

---

be*-K

AOh

TWHCLGTCSK

---

AP-R

ADP

TWHCLGTCSI

---

GP-R

GRDe

HKELGGTCVN

IVGCIVPKK

MRT

RGTIGGTCVNlVGClVPSK

LPDH

YNTLZGVCLNIVGCJIPSX

FIG. 10. Homology of the active oxidoreductases with the sequence oxidase. The homology of GO with

site region of the disulfide around C-521 of glucose

the alcohol oxidases and three of the disulfide oxidoreductases is shown. Conserved residues are shown in bold, and the sequence around the second of the active site cysteines (VGC) in the disulfide oxidoreductases, which is absent in the oxidases, is boxed. The enzymes are as in Fig. 9 with the addition of mercuric reductase (MRT; Brown et al., 1983). region of homology between GO and alcohol oxidase has no clear counterpart in other enzymes in the database. Hydropathic and secondary structural analysis shows that this region is quite hydrophobic and is likely to be mostly /3 sheet. The C-terminal region contains one of the 3 cysteine residues in GO, C5’l. An analysis of this region shown in Fig. 10, suggests that it is related to the active site disulfide domain in the disulfide oxidoreductases such as glutathione reductase, lipoamide dehydrogenase, and mercuric reductase. The 02dependent flavin oxidases (GO and alcohol oxidase) appear to have deleted a small region of 3 amino acids (VGC) including the second of the active site cysteine residues found in the other enzymes. Cysteine 521 in GO has been shown to be nonessential for enzymatic activity or synthesis in yeast by its replacement with serine.3 The homology found suggests an evolutionary relationship between these two different classes of flavoenzymes. 3 S. Rosenberg and K. Frederick, unpublished

results.

This work describes the first example of the secretion of an active flavoenzyme from S. cereuisiae. Glucose oxidase is both one of the largest and most efficiently secreted glycoproteins engineered to be secreted from yeast, as the native molecular weight is >150,000 and more than 100 mg/liter of active enzyme are obtained. Both yeast derived enzymes show substantially more N-linked carbohydrate than the A. niger protein, and the degree of carbohydrate is dependent upon the secretion signal used (Fig. 4). A detailed kinetic comparison of the various enzymes shows no effect of the additional carbohydrate on enzyme activity, although the most hyperglycosylated material synthesized with the GO leader appears to be more thermostable than the A. niger protein.4 For some other heterologous proteins which are hyperglycosylated by yeast, such as tissue plasminogen activator, the extra carbohydrate substantially reduces the activity of the enzyme (MacKay, 1987). The GO proteins from both yeast and A. niger migrate as doublets on SDS gel electrophoresis after EndoH treatment. The cause of this is unknown but may be due to O-linked glycosylation, the anomalous SDS binding properties of the protein (Jones et al., 1982), or the noncoenzyme phosphate found at least in the A. niger enzyme (Swoboda and Massey, 1966b; James et al., 1981). At present, the reason for the differential glycosylation of glucose oxidase between yeast and A. niger and in yeast using the two different leaders is unknown. Yeast is known to hyperglycosylate some secreted foreign proteins (Schultz et al., 1987; MacKay, 1987), especially those using the a-factor leader for secretion. This is most likely due to increased transfer of outer chain mannose to the same sites used by other organisms (e.g., Aspergillus) and, less likely, to the glycosylation of additional sites. Recent experiments suggest that KEX2 and mannosyltransferase I reside in different postendoplasmic reticulum compartments in yeast (Cunningham and Wickner, 1989). The intermediate level of glycosylation seen in the pcrGO-1 transformants could be due to more rapid transit through a compartment or an alternate route through the secretory pathway for the a-factor-GO fusion rather than the GO protein. Alternatively, it is known that the three sites for N-linked glycosylation in the a-factor leader are utilized very efficiently (Julius et al., 1984). Thus, in the case of the a-factor-GO fusion protein, the o-factor sites may compete with the GO sites as substrates for the glycosyl transferases, leading to reduced glycosylation of the GO sites. This could occur either before or after cleavage of the a-factor-GO fusion protein, as exogenously added acceptors for N-linked glycosylation as small as tripeptides have been shown to compete efficiently for glycosylation sites in yeast (Rothblatt et al., 1987). Acknowledgments-We thank Doris Coit for cDNA libraries, B. Irvine for DNA sequencing, K. Chu for protein sequencing, F. Zavrl, M. Brauer, and J. Allen for library screening, other colleagues at Chiron for many helpful discussions, and Annie de Baetselier at the International Institute of Cellular and Molecular Pathology in Brussels for help in process development. REFERENCES Aviv, H., and Leder, P. (1972) Proc. Natl. Acad. Sci. U. S. A. 69, 1408-1412 Banks, J. G., Board, R. G., and Sparks, N. H. C. (1986) Biotechnol. Appl.

Biockm.

8,103-147

Barr, P. J., Gibson, H. L., Enea, V., Arnot, D. E., Hollingdale. M. R., and Nussenzweig, V. (1987) j. l&p. Med. 165, 1160-l-171 Beaty, N. B., and Ballov, D. P. (1981) J. Biol. Chem. 256,4611-4618 Beggs, J. D. (1978) Nature 275, 104-109 ’ K. Frederick and S. Rosenberg, unpublished

data.


FIG. 9. N-terminal homology of glucose oxidase with other flavoenzymes. The sequences of eight flavoenzymes homologous to glucose oxidase in the region of ammo acids 26-50 are shown. Seauences were initiallv defined using the program DFASTP. The sequences shown are:” GO, glucose oiidase,-amino acids 20-50 (this work); AOh, alcohol oxidase from H. polymorpha, Ledeboer et al. (1985), from amino acid 8; AOp, alcohol oxidase from Pichiapmtoris, Ellis et al. (1985), from amino acid 8; GRDh, glutathione reductase (human), Krauth-Siegel et al. (1982), from amino acid 21; GRDe, zlutathione reductase (E. coli). Greer and Perham (1986). from amino acid 5; DAAOX, n-amino aciioxidase (porcine), Ronchiet al. (1982), from amino acid 1; pHBH, p-hydroxybenzoate hydroxylase from Pseudomonas fluorescens, Hofsteenge et al. (1980), from amino acid 3; LPDH, Iipoamide dehydrogenase from E. coli, Stephens et al. (1983), from amino acid 6. The predicted secondary structures indicated are based on homology with the structures of human glutathione reductase and p-hydroxybenzoate hydroxylase. The completely conserved glycines between the first p strand and the a helix are shown in bold type as is the acidic residue at the end of the second sheet.

3801

and Expression

3802

Glucose Oxidase, Cloning and Expression

Chem.

239,

3927-3934

Greer, S., and Perham, R. N. (1986) Biochemistry 25, 2736-2742 Hawke, D., Yuan, P. M., and Shively, J. E. (1982) Anal. Biochem. 120,302-311 Hinnen, A., Hicks, J. B., and Fink, G. R. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, 1929-1933 Hofsteenge, J., Vereijken, J. M., Weijer, W. J., Beintema, J. J., Wierenga, R. K., and Drenth, J. (1980) Eur. J. Biochem. 113,141-

---

1.50

Hunkapiller, M. W. (1985) Applied Biosystems-User Bulletin No. 14, Anulied Biosvstems, Foster City, CA Huyn, ‘i?: A., Young, R. A:, and Davi&R. W. (1985) in DNA Cloning, A Practical Approach (Glover, D. M., Ed) pp. 49-78, IRL Press, Oxford. Great Britain Innis, M.’ A,, Holland, M. J., McCabe, P. C., Cole, G. E., Wittman, V. P., Tal, R., Watt, W. K., Gelfand, D. H., Holland, J. P., and Meade, J. H. (1985) Science 228, 21-26 James, T. L., Edmondson, D. E., and Husain, M. (1981) Biochemistry 20.617-621 Jones, M. N., Manley, P., and Wilkinson, A. (1982) Biochem. J. 203, 285-291 Julius, D., Brake, A., Blair, L., Kunisaw, R., and Thorner, J. (1984) Cell 37, 1075-1089 Kelley, R. L., and Reddy, C. A. (1986) J. Bacterial. 166, 269-274 Kozak. M. (1984) Nucleic Acids. Res. 12.857-872 Krauti-Siegel, i., Blatterspiel, R., Sale& M., Schiltz, E., Schirmer, R. H., and Untucht-Grau, R. (1982) Eur. J. Biochem. 121, 259267 Kusai, K., Sekuzu, I., Hagihara, B., Okunuki, K., Yamauchi, S., and Nakai. M. (1960) Biochem. RioDhvs. Acta 40, 555-557 Laemmii, U. k. (1970) Nature 2!?!7; 680-685 Ledeboer, A. M., Edens, L., Maat, J., Visser, C., Bos, J. W., Verrips, C. T.. Janowicz, Z., Eckart, M., Roggenkamp, R., and Hollenberg, C. P.‘(1985) Nucleic Acids kes. .13,%63-3081 Lipman, D. J., and Pearson, W. R. (1985) Science 22, 1435-1441 Lockridge, O., Massey, V., and Sullivan, P. A. (1972) J. Biol. Chem. 247,8097-8106 Mackay, V. L. (1987) in Biological Research on Industrial Yeasts

(Steward, G. G., Russell, I., Klein, R. D., and Hiebsch, R. R., Eds) Vol. 2, pp. 27-36, CRC Press, Inc. Boca Raton, FL Malcolm, B. A., Rosenberg, S., Corey, M. J., Allen, J. S., de Baetselier, A., and Kirsch, J. F. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 133137 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Snrina Harbor. NY Massey, b., I&ller, F:, Feldberg, R., Schuman, M., Sullivan, P. A., Howell, L. G., Mayhew, S. G., Matthews, R. G., and Foust, G. P. (1969) J. Biol. Chem. 244, 3999-4006 Moehle, C. M., Tizard, R., Lemmon, S. K., Smart, J., and Jones, E. W. (1987) Mol. Cell. Biol. 7,4390-4399 Miiller, D. (1928) Biochem. 2. 199, 136-170 Miiller, F., and Massey, V. (1969) J. Biol. Chem. 244,4007-4016 Pazur, J. H., and Kleppe, K. (1964) Biochemistry 3,578-583 Pazur, J. H., Klenpe, K., and Cepure, A. (1965) Arch. Biochem. Biojhys. lil, 351-357 Penttila. M. E.. A&r& L.. Lehtovaara. P.. Bailey, M., Teeri. T. T., and Knowles; J. K. d. (i988) Gene (&&.) 63,-iO3-i12 Ronchi, S., Minchiotti, L., Galliano, M., Curti, B., Swenson, R. P., Williams, C. H., Jr., and Massey, V. (1982) J. Biol. Chem. 257, 8824-8834 Rosenberg, S., Barr, P. J., Najarian, R., and Hallewell, R. (1984) Nature 312,77-80 Rothblatt, J. A., Webb, J. R., Ammerer, G., and Meyer, D. I. (1987) Embo.

J. 6, 3455-3463

Sanger, F., Nicklen, S., and Coulson, R. (1977) Proc. Natl. U. S. A. 74,

Acad.

Sci.

5463-5467

Schultz, L. D., Tanner, J., Hofmann, K. J., Emini, E. A., Condra, J. H., Jones, R. E., Kieff, E., and Ellis, R. W. (1987) Gene (Amst.) 54,113-123 Stephens, P. E., Lewis, H. M., Darlison, M. G., and Guest, J. R. (1983) Eur. J. Biochem. 135, 519-527 Strickland, S., Palmer, G., and Massey, V. (1975) J. Biol. Chem. 250, 4048-4052 Swoboda, B. E. P., and Massey, V. (1965) J. Biol. Chem. 240,22092215 Swoboda, B. E. P., and Massey, V. (1966a) J. Biol. Chem. 241,34093416 Swoboda, B. E. P., and Massey, V. (1966h) Fluvins & Flavoproteins (Slater, E. C., ed) pp. 263-278, Elsevier, Amsterdam Thieme, R., Pai, E. G., Schirmer, R. H., and Schultz, G. E. (1981) J. Mol.Biol. 152,763-782 Truett, M. A., Blather, R., Burke, R. L., Caput, D., Chu, C., Dina, D., Hartog, K., Kuo, C. H., Masiarz, F. R., Merryweather, J. P., Najarian, R., Pachl, C., Potter, S. J., Puma, J., Quiroga, M., Rall, L. B., Randolph, A., Urdea, M. S., Valenzuela, P., Dahl, H. H., Favalaro, J., Hansen, J., Nordfang, O., and Ezban, M. (1985) DNA 4,333-349 Tschopp, J. F., Sverlow, G., Kosson, R., Craig, W., and Grinna, L. (1987) Biotechnology 5,1305-1308 Turpen, T. H., and Griffith, 0. M. (1986) Biotechniques 4, 11-15 Voet, J. G., Coe, J., Epstein, J., Matossian, V., and Shipley, T. (1981) Biochemistry

20,7182-7185

Voet, J. G., and Andersen, E. C. (1984) Arch. Boichem. Biophys. 233, 88-92 von Hejne, G. (1985) J. Mol. Biol. 184,99-105 Wallace, R. B., Shaffer, J., Murphy, R. F., Bonner, J., Hirose, T., and Itakura, K. (1979) Nucleic Res. 6,3543-3557 Wallace. R. B., Johnson, M. J., Hirose, T., Miyake, T., Kawashima, E. H.,‘and Iiakura, K.‘(1981) Nucleic Acids ties. 9, 879-893 Warner. B. D.. Warner. M. E.. Karns, G. E., Ko, L., Brown-Shimer, S., anh Urdea, M. S. i1984) DNA 3; 401-411 Wierenga, R. K., De Jong, R. J., Kalk, K. H., Hol, W. G. J., and Drenth, J. (1979) J. Mol. Biol. 131,55-73 Wood, W. I., Gitschier, J., Lasky, L. A., and Lawn, R. M. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 1585-1588 Worthington, C. S., and Teller, J. D. (1972) Worthington Enzyme Manual. D. 43. Freehold, NJ


Boeke, J. D., Lacroute, F., and Fink, G. R. (1984) Mol. & Gen. Gertet. 197,345-346 Boel, E., Hansen, M. T., Jhort, I., Hoegh, I., and Fiil, N. P. (1984) Embo J. 3, 1581-1585 Brake, A. J., Merryweather, J. P., Coit, D., Heberlein, U., Masiarz, F. R., Mullenbach. G. T., Urdea. M. S., Valenzuela, P., and Barr, P. J. (1984) Proc. &tl. A&d. Sci. U. S. A. 81, 4642-4646 Bright, H. J., and Appleby, M. (1969) J. Biol. Chem. 244,3625-3634 Brown, N. L., Ford, S. J., Pridmore, R. D., and Fritzinger, D. C. (1983) Biochemist& 22, .4089-4095 ChirPwin. J. M.. Przvbvla, A. E.. MacDonald. R. J.. and Rutter, W. J. 71976) Biockem&& I&, 5294-5299 Corrick, C. M., Twomey, A. P., and Hynes, M. J. (1987) Gene (Am&) 53,63-71 Cousens, L. S., Shuster, J. R., Gallegos, C., Ku, L., Stempien, M. M., Urdea, M. S., Sanchez-Pescador, R., Taylor, A., and TekampOlson. P. (1987) Gene (Am&.) 61. 265-275 Cunninhham, K. w., and Wickner, %‘. T. (1989) Yeast 5,25-33 Degani,Y., and Heller, A. (1987) J. Phys. Chem. 91, 1285 Dezani. Y.. and Heller. A. (1988) J. Am. Chem. Sot. 110,2615-2620 DoLson, hi. J., Tuite, h. F:, Roberts, N. A., Kingsman, A. J., Kingsman, S. M., Perkins, R. E., Conroy, S. C., Dunbar, B., and Fothegill, L. A. (1982) Nucleic Acids Res. 10, 2625-2637 Ellis, S. B., Brust, P. F., Koutz, P. J., Waters, A. F., Harpold, M. M., and Gingeras, T. R. (1985) Mol. Cell. Biol. 5, 1111-1121 Erhart. E.. and Hollenberg. C. P. (1983) J. Bacterial. 156,625-635 Feinberg, k. P., and Vogelstein, B. (1983) Anal. Biochem. 132,6-13 Felenbok. B.. Seoueval. D.. Mathieu. M.. Siblev. S.. Gwvnne. D. I.. and D&iei, R. *W. (1988) Gene (A&st.)‘73,38&396 Friedman, M., Krull, L. G., and Cavins, J. F. (1970) J. Biol. Chem. 245,3868-3871 Gibson, Q. H., Swoboda, B. E. P., and Massey, V. (1964) J. Biol.