Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe...

5
Proc. Natl. Acad. Sci. USA Vol. 87, pp. 142-146, January 1990 Biochemistry Expression of a cloned streptavidin gene in Escherichia coli (gene expreson/blotfin-binding protein/17 promoter/T7 RNA polymerase) TAKESHI SANO AND CHARLES R. CANTOR Department of Genetics and Development, College of Physicians and Surgeons, Columbia University, New York, NY 10032 Contributed by Charles R. Cantor, August 24, 1989 ABSTRACT We describe the construction of systems for expressing the cloned streptavidin gene in Escherichia coli. Although the streptavidin gene is extremely lethal to the host cells, because of the strong biotin binding of the gene product' the gene was expressed efficiently in E. coil by using 17 RNA polymerase/T7 promoter expression systems. The expressed streptavidin accumulated to more than 35% of the total cell protein. The expressed streptavidin was insoluble in the cell. However, after solubilization by dialysis against 6 M guanidine hydrochloride (pH 1.5) and removal of guanidine hydrochlo- ride by dialysis, the protein became soluble and renatured. This simple procedure yielded streptavidin purified almost to ho- mogeneity. The purified streptavidin bound 3.5-3.9 molecules of biotin per molecule, indicating that it had almost full biotin-binding ability. Some of the purified streptavidin mol- ecules aggregated into ofigomers, suggesting that the C- terminal region of the molecule, present in our material but absent in typical preparations, may be responsible for the aggregation. Streptavidin, a protein produced by Streptomyces avidinii, specifically binds a water-soluble vitamin, D-biotin (1, 2). The strong and specific affinity for biotin (Kd - 10-15 M) (2, 3) has made streptavidin, as well as egg white avidin, a frequently used protein for detection and characterization of various biological substances (3-7). Streptavidin possesses no car- bohydrate chain (2) and has a lower isoelectric point, whereas avidin is a glycoprotein (8). Thus streptavidin is more useful for most applications than avidin because of its reduced nonspecific binding. In addition, streptavidin has no sulfur- containing amino acids (9), whereas avidin contains a disul- fide-linkage as well as two methionine residues per subunit (3, 8). This also offers some advantages for a variety of applications. We are interested in constructing expression vectors to produce streptavidin-containing fusion proteins. Such pro- teins could be detected and purified easily by their biotin binding. We are also interested in modifying the streptavidin gene to produce mutants that would expand the applications of the streptavidin-biotin system. As the first step in such studies, we attempted to express the cloned streptavidin gene from S. avidinii (9) in Escherichia coli. Although the strepta- vidin gene is extremely lethal to the host cells, it can be expressed efficiently by using T7 RNA polymerase/T7 pro- moter expression systems (10, 11). Here we describe several expression vectors and expression systems for the cloned streptavidin gene and also show some properties of the expressed streptavidin. MATERIALS AND METHODS Bacteria, Bacteriophages, and Lysogens. E. coli strains HMS174 (F- hsdR recA RifR) (12) and BL21 (F- ompT hsdS gal) (10, 13) were used for cloning and expression. E. coli strain ED8739 (F- metB hsdS supE supF) (10) was used as the host for growing CE6. The bacteriophage A derivative CE6 (10) was used for delivering T7 RNA polymerase. CE6 carries the T7 RNA polymerase gene (T7 gene 1) (14) undei the PL and p, promoters. BL21(DE3) and HMS174(DE3) are lysogens with the A derivative DE3 integrated into the chromosome (10). DE3 carries the T7 RNA polymerase gene under the lacUV5 promoter in the int gene. These strains were gifts from F. W. Studier (Brookhaven National Labo- ratory). Plasmids. pUC-SAI (a gift from M. Uhlen and R. Johansson Royal Institute of Technology, Sweden) has the Fsp I-Sma I fragment [580 base pairs (bp)] of the streptavidin gene (9) inserted in the Sma I site of pUC8, in the same orientation as the lacZ gene. pT7-7 (a gift from S. Tabor, Harvard Medical School) is a derivative of pT17-5 (15) and has the l10 promoter (one of six strong T7 promoters that is located upstream of the T7 gene 10) (14) and the translation initiation site for the T7 gene 10 protein (bp 22,857-22,972 of T7 DNA) (14) inserted into bp 2065-4360 of pBR322 (16). This plasmid also has the polylinker region of pUC12 downstream of the initiation codon. pET-3a (17) (a gift from F. W. Studier) derived from pBR322 carries the 10 promoter and the first 11 codons of T7 gene 10 protein (bp 22,880-22,998 of T7 DNA) (14) followed by a BamHI linker. In addition, the transcription terminator TO (bp 24,106-24,228 of T7 DNA) (14) is placed downstream of the BamHI linker. pLysS and pLysE (11) (gifts from F. W. Studier) carry the T7 lysozyme gene (T7 gene 3.5, bp 10,665-11,296 of T7 DNA) (14) in pACYC184 (18). T7 lysozyme binds T7 RNA polymerase and inhibits its activity (19). Therefore, the presence of T7 lysozyme in the cell reduces the basal level of Ti7 RNA polymerase activity in the uninduced state (11). pLysE has the T7 lysozyme gene under the tet promoter of pACYC184, whereas pLysS has the gene in the opposite orientation, and thus the T7 lysozyme level with pLysS should be much lower than with pLysE (11). Construction of Expression Vectors. Construction of the expression vectors was carried out using standard techniques (20). Restriction endonucleases, T4 DNA polymerase, and T4 DNA ligase were obtained from New England Biolabs or Boehringer Mannheim. pUC-SAI was digested with Bgl I and treated with T4 DNA polymerase to generate blunt ends. The plasmid was then digested with BamHI, and the DNA fragments were electrophoresed on agarose gels. A 480-bp DNA fragment, which had the coding region for the 15th amino acid to the C terminus of mature streptavidin, was extracted and purified from the agarose using a DEAE- membrane (Schleicher & Schuell). The purified DNA frag- ment was ligated between the Sma I and BamHI sites of pT7-7 [2.46 kilobases (kb)] with T4 DNA ligase. The resulting plasmid is referred to as pTSA-1 (2.95 kb; Fig. 1). pTSA-1 was digested with Nde I and BamHI, and a DNA fragment of 500 bp was purified using a DEAE-membrane after agarose gel electrophoresis. The purified DNA fragment was ligated between the Nde I and BamHI sites of pET-3a 142 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Transcript of Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe...

Page 1: Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe streptavidin geneis extremely lethal to the host cells, ... gal) (10, 13) were used for

Proc. Natl. Acad. Sci. USAVol. 87, pp. 142-146, January 1990Biochemistry

Expression of a cloned streptavidin gene in Escherichia coli(gene expreson/blotfin-binding protein/17 promoter/T7 RNA polymerase)

TAKESHI SANO AND CHARLES R. CANTORDepartment of Genetics and Development, College of Physicians and Surgeons, Columbia University, New York, NY 10032

Contributed by Charles R. Cantor, August 24, 1989

ABSTRACT We describe the construction of systems forexpressing the cloned streptavidin gene in Escherichia coli.Although the streptavidin gene is extremely lethal to the hostcells, because of the strong biotin binding of the gene product'the gene was expressed efficiently in E. coil by using 17 RNApolymerase/T7 promoter expression systems. The expressedstreptavidin accumulated to more than 35% of the total cellprotein. The expressed streptavidin was insoluble in the cell.However, after solubilization by dialysis against 6 M guanidinehydrochloride (pH 1.5) and removal of guanidine hydrochlo-ride by dialysis, the protein became soluble and renatured. Thissimple procedure yielded streptavidin purified almost to ho-mogeneity. The purified streptavidin bound 3.5-3.9 moleculesof biotin per molecule, indicating that it had almost fullbiotin-binding ability. Some of the purified streptavidin mol-ecules aggregated into ofigomers, suggesting that the C-terminal region of the molecule, present in our material butabsent in typical preparations, may be responsible for theaggregation.

Streptavidin, a protein produced by Streptomyces avidinii,specifically binds a water-soluble vitamin, D-biotin (1, 2). Thestrong and specific affinity for biotin (Kd - 10-15 M) (2, 3) hasmade streptavidin, as well as egg white avidin, a frequentlyused protein for detection and characterization of variousbiological substances (3-7). Streptavidin possesses no car-bohydrate chain (2) and has a lower isoelectric point, whereasavidin is a glycoprotein (8). Thus streptavidin is more usefulfor most applications than avidin because of its reducednonspecific binding. In addition, streptavidin has no sulfur-containing amino acids (9), whereas avidin contains a disul-fide-linkage as well as two methionine residues per subunit(3, 8). This also offers some advantages for a variety ofapplications.We are interested in constructing expression vectors to

produce streptavidin-containing fusion proteins. Such pro-teins could be detected and purified easily by their biotinbinding. We are also interested in modifying the streptavidingene to produce mutants that would expand the applicationsof the streptavidin-biotin system. As the first step in suchstudies, we attempted to express the cloned streptavidin genefrom S. avidinii (9) in Escherichia coli. Although the strepta-vidin gene is extremely lethal to the host cells, it can beexpressed efficiently by using T7 RNA polymerase/T7 pro-moter expression systems (10, 11). Here we describe severalexpression vectors and expression systems for the clonedstreptavidin gene and also show some properties of theexpressed streptavidin.

MATERIALS AND METHODSBacteria, Bacteriophages, and Lysogens. E. coli strains

HMS174 (F- hsdR recA RifR) (12) and BL21 (F- ompT hsdS

gal) (10, 13) were used for cloning and expression. E. colistrain ED8739 (F- metB hsdS supE supF) (10) was used asthe host for growing CE6. The bacteriophage A derivativeCE6 (10) was used for delivering T7 RNA polymerase. CE6carries the T7 RNA polymerase gene (T7 gene 1) (14) undeithe PL and p, promoters. BL21(DE3) and HMS174(DE3) arelysogens with the A derivative DE3 integrated into thechromosome (10). DE3 carries the T7 RNA polymerase geneunder the lacUV5 promoter in the int gene. These strainswere gifts from F. W. Studier (Brookhaven National Labo-ratory).

Plasmids. pUC-SAI (a gift from M. Uhlen and R. JohanssonRoyal Institute of Technology, Sweden) has the Fsp I-SmaI fragment [580 base pairs (bp)] of the streptavidin gene (9)inserted in the Sma I site of pUC8, in the same orientation asthe lacZ gene. pT7-7 (a gift from S. Tabor, Harvard MedicalSchool) is a derivative of pT17-5 (15) and has the l10promoter(one of six strong T7 promoters that is located upstream oftheT7 gene 10) (14) and the translation initiation site for the T7gene 10 protein (bp 22,857-22,972 of T7 DNA) (14) insertedinto bp 2065-4360 of pBR322 (16). This plasmid also has thepolylinker region of pUC12 downstream of the initiationcodon. pET-3a (17) (a gift from F. W. Studier) derived frompBR322 carries the 10 promoter and the first 11 codons ofT7 gene 10 protein (bp 22,880-22,998 of T7 DNA) (14)followed by a BamHI linker. In addition, the transcriptionterminator TO (bp 24,106-24,228 of T7 DNA) (14) is placeddownstream of the BamHI linker. pLysS and pLysE (11)(gifts from F. W. Studier) carry the T7 lysozyme gene (T7gene 3.5, bp 10,665-11,296 of T7 DNA) (14) in pACYC184(18). T7 lysozyme binds T7 RNA polymerase and inhibits itsactivity (19). Therefore, the presence of T7 lysozyme in thecell reduces the basal level of Ti7 RNA polymerase activity inthe uninduced state (11). pLysE has the T7 lysozyme geneunder the tet promoter ofpACYC184, whereas pLysS has thegene in the opposite orientation, and thus the T7 lysozymelevel with pLysS should be much lower than with pLysE (11).

Construction of Expression Vectors. Construction of theexpression vectors was carried out using standard techniques(20). Restriction endonucleases, T4 DNA polymerase, andT4 DNA ligase were obtained from New England Biolabs orBoehringer Mannheim. pUC-SAI was digested with Bgl I andtreated with T4 DNA polymerase to generate blunt ends. Theplasmid was then digested with BamHI, and the DNAfragments were electrophoresed on agarose gels. A 480-bpDNA fragment, which had the coding region for the 15thamino acid to the C terminus of mature streptavidin, wasextracted and purified from the agarose using a DEAE-membrane (Schleicher & Schuell). The purified DNA frag-ment was ligated between the Sma I and BamHI sites ofpT7-7 [2.46 kilobases (kb)] with T4 DNA ligase. The resultingplasmid is referred to as pTSA-1 (2.95 kb; Fig. 1).pTSA-1 was digested with Nde I and BamHI, and a DNA

fragment of 500 bp was purified using a DEAE-membraneafter agarose gel electrophoresis. The purified DNA fragmentwas ligated between the Nde I and BamHI sites of pET-3a

142

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Page 2: Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe streptavidin geneis extremely lethal to the host cells, ... gal) (10, 13) were used for

Proc. NatL. Acad. Sci. USA 87 (1990) 143

Abla

I TO

itrept-avid in

010 tio

BpTSA-1 and pTSA-2

pT7-7 Slreptavidin gene

... ATG (ACT AGA ATT CGC GCC CAG GCC GGT ATC ...

5 I IMet Ala Arg lie Arg Ala Gin Ala Gly Ile ...

Streptavidin gene of S. avidinil* . . GCC CAGGTC TCG GCCGCC GAGGCC GGT ATC

10 1SAla Gin Val Ser Ala Ala Glu Ala Gly lie ...

FIG. 1. (A) Expression vectors for cloned streptavidin gene,pTSA-1 and pTSA-2. (B) DNA sequence and deduced amino acidsequence of the N-terminal region encoded in the expression vectorsand in the natural streptavidin gene of S. avidinii (13).

(4.5 kb) with T4 DNA ligase. The resulting plasmid is referredto as pTSA-2 (5.0 kb; Fig. 1).pTSA-1 and pTSA-2 have the same coding sequence,

which carries most of the coding region for the maturestreptavidin, under the promoter with some truncation of

the N-terminal region. For pTSA-2, the transcription termi-nator TO follows the coding region. The streptavidin subunitencoded in the plasmids consists of 140 amino acids and itsmolecular weight is estimated to be 15,900 from the deducedamino acid sequence.

Expression by Infection with Bacteriophage CE6. Expres-sion was carried out as described by Studier and Moffatt (10).HMS174 carrying a target plasmid was grown at 370C withshaking in LB medium supplemented with 0.4% maltose andampicillin (50 Ag/ml). When the A6w of the culture reached1.0, a purified CE6 bacteriophage stock was added to a finalconcentration of 2-4 x 109 infective bacteriophage particlesper ml, producing a multiplicity of infection of 5-10. Theinfected cells were incubated at 370C with shaking.

Expression Using Lysogen Systems. Expression was per-formed as described by Studier et al. (11). BL21(DE3) orHMS174(DE3) carrying a target plasmid was grown at 370Cwith shaking in LB medium supplemented with 0.4% glucose,ampicillin at 50 Ag/ml, and chloramphenicol at 25 Ag/ml (forcells carrying pLysS or pLysE). When the A6w reached 1.0,100 mM isopropyl ,3-D-thiogalactopyranoside was added to afinal concentration of 0.4 mM to induce the T7 RNA poly-merase gene under the lacUV5 promoter. For cells trans-formed with pLysE, the induction was carried out at an A6wof 0.6, because more time is required to titrate out T7lysozyme by the newly synthesized T7 RNA polymerase thanfor the cells without the T7 lysozyme gene or those carryingpLysS. After induction, the cells were incubated at 370C withshaking.

Protein Composition. Cells from 5 ml of culture werecollected by centrifugation at 1600 x g for 10 min at 40C anddissolved in 600 Al of SDS sample buffer consisting of 2%(wt/vol) SDS, 50 mM TrisHC1 (pH 6.8), and 10 mM 2-mercaptoethanol. The sample was heated in boiling water for5 min and subjected to SDS/PAGE analyses.Western Blotting (Immunodetection). Proteins after SDS/

PAGE were transferred onto a nitrocellulose membrane(Schleicher & Schuell) using a semi-dry electroblotter (Nov-aBlot, LKB) (21). Rabbit anti-streptavidin (ICN ImmunoBi-

ologicals) antiserum and goat anti-rabbit IgG antibodies con-jugated to alkaline phosphatase (Bio-Rad) were used as theprimary and the secondary antibodies, respectively. Thealkaline phosphatase activity was detected using 5-bromo-4-chloro-3-indolyl phosphate and p-nitroblue tetra-zolium (Bio-Rad).

Purification of Expressed Streptavidin. All procedures werecarried out at 40C or on ice, unless otherwise stated. Theculture (100 ml) incubated at 370C after the induction wascentrifuged at 1600 x g for 10 min. The cell pellet was washedwith 100 ml of 100 mM NaCI/1 mM EDTA/10 mM Tris'HCI,pH 8.0 and centrifuged as above. The cells were suspendedin 10 ml of 2 mM EDTA/30 mM Tris HC1, pH 8.0, containing0.1% Triton X-100. Lysis occurred under these conditionsbecause of the presence of T7 lysozyme in the cells. The celllysate was stored at -700C until used.The frozen cell lysate stock was thawed at room temper-

ature (-220C). MgSO4 (1.0 M), DNase I (1 mg/ml, Sigma),and RNase A (1 mg/ml, Sigma) were added to final concen-trations of 12 mM, 10 Ag/ml, and 10 ug/ml, respectively, andthe mixture was allowed to stand for 15 min at room tem-perature to reduce the viscosity of the lysate. The cell lysatewas centrifuged at 39,000 x g for 15 min, and the precipitatewas washed again with 10 ml of 2 mM EDTA/30 mMTris HCI, pH 8.0/0.1% Triton X-100 followed by centrifuga-tion as above. The precipitate was dissolved in 5 ml of 6 Mguanidine hydrochloride (pH 1.5) with gentle stirring anddialyzed against the same solution. The dialysate was thendialyzed against 0.2 M NaHCO3 (pH 8.7) to remove guanidinehydrochloride. The dialysate was centrifuged at 39,000 X gfor 15 min and the supernatant was used as the expressedstreptavidin fraction.

Protein Characterization. Gel-filtration chromatographywas conducted at room temperature using a Sephacryl S-200HR column (Pharmacia). Detailed conditions are given in thelegend to Fig. 6. The biotin-binding ability of streptavidin wasdetermined by the bentonite adsorption method (22) usingD-[carbonyl-14C]biotin (52 mCi/mmol; 1 Ci = 37 GBq; Am-ersham). SDS/PAGE analyses were performed with a dis-continuous buffer system (23) using a 15% polyacrylamidegel. The proteins were stained with 0.15% Coomassie brilliantblue (Sigma) dissolved in 45% (vol/vol) methanol/10o (vol/vol) acetic acid. Protein concentration was determined by themicrobiuret method (24) using bovine serum albumin as thestandard or by using E9°' = 3.4 for streptavidin (25).

RESULTS AND DISCUSSIONExpression by Infection with Bacteriophage CE6. The effi-

cacy of the streptavidin-biotin detection system originatesfrom the high and specific binding affinity of streptavidin forbiotin. However, this characteristic is a severe disadvantagein attempts to express the cloned streptavidin gene. When thestreptavidin gene is expressed, even in small amounts, thegene product binds biotin molecules, which are essential forcell metabolism; this causes the death of the host cell.We first constructed the expression vector pT7-SAIIH,

which contained the entire streptavidin gene including thesignal peptide and also encoded 25 vector-derived aminoacids. pT7-SAIIH could transform HMS174(DE3), a lysogencarrying the T7 RNA polymerase gene, and a protein of Mr22,400, consistent with the deduced amino acid sequence,was expressed efficiently after induction of the 37 RNApolymerase gene (data not shown). However, the expressedprotein did not show any biotin-binding ability, suggestingthat the signal peptide region or the 25 vector-derived aminoacids inactivated the protein. These results encouraged us toconstruct expression vectors for the streptavidin gene with-out the signal peptide-i.e., mature streptavidin.

Biochemistry: Sano and Cantor

Page 3: Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe streptavidin geneis extremely lethal to the host cells, ... gal) (10, 13) were used for

144 Biochemistry: Sano and Cantor

We then constructed the expression vector pTSA-1, inwhich the coding region for the signal peptide was omitted.In addition, the coding sequence corresponding to severalamino acids in the N-terminal region of the mature strepta-vidin is truncated (Fig. 1B). Streptavidin preparations puri-fied from the culture medium of S. avidinii are usuallyproteolyzed at both the N and the C termini to produce a

minimal size molecule (9, 26-28), called core streptavidin.The N terminus encoded by pTSA-1 has almost the same

structure as that of the core streptavidin (26, 28) preceded byfive additional vector-derived amino acids (Fig. 1B).pTSA-1 could transform HMS174 and could be stably

maintained in the cell during growth. We tried to express thestreptavidin gene by infection with bacteriophage CE6,which delivers T7 RNA polymerase to the cell. The SDS/PAGE patterns of total cell protein (Fig. 2A) show that, withincreasing incubation time after infection, a protein band atMr 16,000 increases in intensity. This band could not beobserved at the time of infection. The molecular weight ofthis protein is consistent with that of the streptavidin subunitencoded in pTSA-1. In Western blots (Fig. 2B), only one bandwith Mr 16,000 cross-reacted with anti-streptavidin. Thus theMr 16,000 protein was the product of the cloned streptavidingene.HMS174 transformed with pTSA-1 grew well in enriched

medium, such as LB, but grew poorly in M9 minimummedium. Western blots (Fig. 2B) show that a small amount ofstreptavidin was expressed even before infection (not appar-ent in Fig. 2). This suggests that the target gene could betranscribed by E. coli RNA polymerase even if placed underthe T7 promoter. When expressed streptavidin binds biotin inthe cell, additional biotin molecules can apparently be takenup from the medium. This idea is supported by the fact thatthe cells can grow well in M9 medium supplemented with 8.2,uM biotin. Biotin supplementation is presumably indispens-able for the host cells to maintain stably the plasmid carryingthe streptavidin gene.The above method, where T7 RNA polymerase is delivered

by infection with bacteriophage CE6, successfully expressesthe cloned streptavidin gene in E. coli. This system isapplicable to many toxic genes that cannot be maintained incells carrying the T7 RNA polymerase gene (10, 11). How-ever, the method is inconvenient for routine expressionexperiments, because purified bacteriophage stock must beprepared each time. In addition, the multiplicity of infectionand the growth conditions affect the expression efficiency ofthe target gene considerably (10, 11). The expression effi-

A b B ba 01 2.3 c aO 3 c.L

FIG. 2. Expression of streptavidin by infection with bacterio-phage CE6. Total cell protein of HMS174(pTSA-1) was subjected toSDS/PAGE. (A) Stained protein. (B) Western blots (immunodetec-tion). In lanes b, the number above each lane is the time afterinfection in hr. Lanes: a, core streptavidin (Mr, 13,500); b,HMS174(pTSA-1) (each lane contains the total cell protein from 208u1l of culture for A and from 83 p.1 of culture for B); c Left, molecularweight standard proteins (Pharmacia); c Right, prestained molecularweight standard proteins (Bio-Rad). Horizontal lines indicate posi-

tions of the standard proteins (Mr = 94, 67, 43, 30, 20.1, and 14.4 x

103) in stained protein and those of the prestained standard proteins(Mr = 110, 74, 50, 34, 27, and 19 x 103) in Western blots.

ciency of the cloned streptavidin gene varied with everyexperiment. Thus we attempted, instead, to express thecloned streptavidin gene using lysogen systems.

Expression Using pTSA-1 in Lysogen Systems. pTSA-1could transform HMS174(DE3), a lysogen carrying the T7RNA polymerase gene; however, the plasmid was unstableand lost during growth. In addition, this plasmid could nottransform BL21(DE3), a lysogen containing the T7 RNApolymerase gene, even in the presence of pLysS or pLysEcarrying the T7 lysozyme gene. In contrast, HMS174(DE3)-(pLysS) or HMS174(DE3)(pLysE), lysogens carrying the T7RNA polymerase gene and transformed with pLysS or pL-ysE, respectively, could be transformed with pTSA-1 andmaintained the plasmid stably during the growth, suggestingthat streptavidin is one of a small fraction of target genes thatare more stable in HMS174(DE3) derivatives than in theequivalent derivatives of BL21(DE3) (11). Therefore, theexpression of the streptavidin gene using the latter systemswas characterized.SDS/PAGE analysis of total cell protein (Fig. 3A) shows

that a protein at Mr 16,000 was expressed extensively in cellscarrying pTSA-1, but this protein was not observable in cellswithout pTSA-1. At the time of induction, little trace of thisband could be seen. However, the cellular content of thisprotein increased with incubation time after the induction. InWestern blots (Fig. 3B), only the Mr 16,000 protein cross-reacted with anti-streptavidin in the cells with pTSA-1. Theseresults show clearly that the streptavidin gene was expressedefficiently by using this lysogen system. The expressionefficiency was higher for HMS174(DE3)(pLysS)(pTSA-1)than for HMS174(DE3)(pLysE)(pTSA-1). The expressedstreptavidin content in HMS174(DE3)(pLysS)(pTSA-1) wasestimated to be >25% of the total cell protein 3 hr afterinduction.

A b c d ea01 301 301 3501 35i-

B b c d ea 01 301 3 030 5 01 35f

FIG. 3. Expression of streptavidin using pTSA-1 in lysogensystems. Total cell protein of HMS174(DE3)(pLysS) andHMS174(DE3)(pLysE) with or without pTSA-1 was subjected toSDS/PAGE. (A) Stained protein. (B) Western blots. In lanes b-e, thenumber above each lane is the time after the induction in hr. Lanes:a, core streptavidin; b, HMS174(DE3)(pLysS); c, HMS174(DE3)-(pLysS)(pTSA-1); d, HMS174(DE3)(pLysE); e, HMS174(DE3)-(pLysE)(pTSA-1); f Left, standard proteins; f Right, prestainedstandard proteins. For A, each lane contains the total cell proteinfrom 250 Al of culture except for the lanes at 3 hr for lanes b and 3hr and 5 hr for lanes d, where 83 ,ul of culture was used. For B, thetotal cell protein from 50 ,ul of culture was applied in each lane.

Proc. Natl. Acad. Sci. USA 87 (1990)

Page 4: Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe streptavidin geneis extremely lethal to the host cells, ... gal) (10, 13) were used for

Proc. Natl. Acad. Sci. USA 87 (1990) 145

Expression Using pTSA-2 in Lysogen Systems. Transcrip-tion by T7 RNA polymerase is very active (14), yielding longRNA molecules. IfRNA synthesis were terminated just afterthe coding region, production of coding region RNA shouldbecome more efficient. This would increase the expression ofthe target gene. To test this, we constructed the expressionvector pTSA-2, which has the T7 transcription terminator 40bp downstream from the coding region. Surprisingly, pTSA-2could transform both HMS174(DE3) and BL21(DE3) in thepresence of pLysS or pLysE, where pTSA-1 could nottransform BL21(DE3) even in the presence of pLysS orpLysE. The '10 promoter-containing fragment in pTSA-2lacks bp 22,857-22,879, which have significant promoteractivity for E. coli RNA polymerase (17), whereas thefragment in pTSA-1 retains this sequence. Therefore, smalleramounts of streptavidin should be expressed from pTSA-2than from pTSA-1 in the uninduced state, resulting in a lowertoxicity to the host cell.BL21(DE3)(pLysS) could not maintain pTSA-2 stably.

Cells that lost the target plasmid over-grew the culture duringthe incubation after the induction of the T7 RNA polymerasegene, and thus the expression of the Mr 16,000 protein waspoor (Fig. 4A). However, BL21(DE3)(pLysE) could maintainpTSA-2 stably, and the Mr 16,000 protein was expressedefficiently (Fig. 4A). Western blots (Fig. 4C) show that onlythe Mr 16,000 protein cross-reacted with anti-streptavidin in

A b c d ea 01 301 301 135 f

B g k i i1 01 3 0 30 T 35501 3

1303 _ 5 53035 5 -

FIG. 4. Expression of streptavidin using pTSA-2 in lysogensystems. Total cell protein of BL21(DE3)(pLysS), BL21(DE3)-(pLysE), HMS174(DE3)(pLysS), and HMS174(DE3)(pLysE) withor without pTSA2 was subjected to SDS/PAGE. (A and B) Stainedprotein. (C) Western blots. In lanes b-c and g-j, the number aboveeach lane is the time after the induction in hr. Lanes: a, corestreptavidin; b, BL21(DE3)(pLysS); c, BL21(DE3)(pLysS)(pTSA-2); d, BL21(DE3)(pLysE); e, 13L21(DE3)(pLysE)(pTSA-2); f Left,standard proteins; f Right, prestained standard proteins; g,HMS174(DE3)(pLysS); h, HMS174(DE3)(pLysS)(pTSA-2); i,HMS174(DE3)(pLysE); j, HMS174(DE3)(pLysE)(pTSA-2). For Aand B, each lane contains the total cell protein from 208 ,.l of cultureexcepnt for lane bat 3 hr lne d At 3 hrlneg at 3 hr, lne h at 3 hr,and lane i at 3 hr that contain 125 ,1, and for lane d at 5 hr and lanei at 5 hr that contain 83 ,lI. For C, the cell protein from 42 p1 of culturewas applied to each lane.

the cells with pTSA-2, indicating that the Mr 16,000 proteinwas the product of the streptavidin gene. For BL21(DE3)-(pLysE)(pTSA-2), the expressed streptavidin accounted for>35% of the total cell protein 5 hr after the induction.

In contrast, both HMS174(DE3)(pLysS)(pTSA-2) andHMS174(DE3)(pLysE)(pTSA-2) could express the Mr 16,000protein after induction (Fig. 4B). In Western blots (Fig. 4C),only the Mr 16,000 protein cross-reacted with anti-streptavidin. The expression efficiency was higher forHMS174(DE3)(pLysS), consistent with what was observedwith expression using pTSA-1 (Fig. 3). The expression effi-ciency of HMS174(DE3)(pLysS)(pTSA-2) 3 hr after the in-duction seems almost the same as that ofBL21(DE3)(pLysE)-(pTSA-2) 5 hr after the induction. A slightly higher expres-sion efficiency could be found in the cells transformed withpTSA-2 than those with pTSA-1 (Fig. 3) in this system.We conclude that the cloned streptavidin gene is expressed

efficiently in E. coli by using the lysogen systems with theaddition ofT7 lysozyme to reduce the basal level of I7 RNApolymerase activity in the uninduced state.

Purification and Properties ofExpressed Streptavidin. BL21has a potential advantage as an expression system. This cellshould lack the Ion proteinase and the ompT proteinase (10,13). Because these proteinases may cause degradation ofexpressed proteins in the cell and during purification steps(13), we used BL21(DE3)(pLysE)(pTSA-2) incubated for 5 hrafter induction as the source for purification attempts.Most of the expressed streptavidin was insoluble in the

cell, probably forming an inclusion body as generally ob-served in E. coli overexpression systems. However, whenthe insoluble protein fraction was dissolved by dialyzingagainst 6 M guanidine hydrochloride (pH 1.5) and thenguanidine hydrochloride was removed by dialysis, the ex-pressed streptavidin became soluble. The resulting superna-tant was composed almost solely of the expressed strepta-vidin (Fig. 5), demonstrating that the expressed streptavidincould be purified almost to homogeneity by this simpleprocedure. Nucleic acid contamination was sometimesfound, but this could be removed easily by affinity chroma-tography using 2-iminobiotin as the ligand (29, 30) or bygel-filtration chromatography. The yield of the expressedstreptavidin ranged from 3.9 to 6.5 mg per 100 ml of culture.This yield was almost the same as that from S. avidinii, whichtakes 5-10 days to culture and requires chromatographicpurification steps (29-31).

a bcde f g

FIG. 5. SDS/PAGE analysis of the total protein of each fractionduring purification of the expressed streptavidin using BL21(DE3)-(pLysE)(pTSA-2). Lanes: a, core streptavidin; b, at the time ofinduction; c, at 5 hr after the induction; d, soluble fraction after thelysis; e, soluble fraction after solubilization of the insoluble proteinsby dialysis against 6 M guanidine hydrochloride (pH 1.5) followed bydialysis against 0.2 M NaHCO3 (pH 8.7); f, insoluble fraction afterthe solubilization in guanidine hydrochloride followed by the re-moval of guanidine hydrochloride; g, molecular weight standardproteins.

Biochemistry: Sano and Cantor

Page 5: Expression · PDF fileexpressing the cloned streptavidin gene in Escherichia coli. Althoughthe streptavidin geneis extremely lethal to the host cells, ... gal) (10, 13) were used for

146 Biochemistry: Sano and Cantor

40 60Elution volume (ml)

FIG. 6. Elution pattern of the purified streptavidin on gel-filtration chromatography using a Sephacryl S-200 HR column (1.5cm x 48 cm). The purified streptavidin (66 gg) was applied to thecolumn equilibrated with 0.2 M NaHCO3 (pH 8.7) containing 0.02%NaN3. The elution was carried out at room temperature with thesame solution at a flow rate of 13 ml/hr. Protein concentration wasdetermined by A280. Arrowheads indicate the positions where themonomer (tetramer of the subunit), dimer, trimer, and tetramershould elute (from right to left), judged from the calibration usingstandard proteins.

The purified streptavidin bound 3.5-3.9 molecules ofbiotinper tetramer of the subunit, indicating that the protein hasalmost full-biotin binding ability (4 molecules of biotin per

tetramer). When dialysis against 6 M guanidine hydrochlo-ride (pH 1.5) was eliminated from the purification proce-

dures, the biotin-binding ability was reduced by 20%, whichreveals that some of the expressed streptavidin had boundbiotin in the cell or during purification. Although solubiliza-tion in 6 M guanidine hydrochloride at neutral pH is sufficientto isolate the expressed streptavidin, dialysis against 6 Mguanidine hydrochloride (pH 1.5) is indispensable for a moreactive streptavidin preparation.The purified streptavidin showed a sharp peak at Mr 64,000

(tetramer of the subunit) with a broad shoulder at highermolecular weight range (Fig. 6) on gel-filtration chromatog-raphy. On SDS/PAGE analysis, only one band could beobserved at Mr 16,000 even for the fractions at the highermolecular weight range (data not shown). These resultsindicate that some of the expressed streptavidin aggregatedinto oligomers, which accounted for 30-50%o of the totalprotein. In a core streptavidin preparation (molecular weightof the subunit was 13,500; from Boehringer Mannheim),aggregated molecules were <5% of the total protein. Earlierreports suggested that the full-length gene product tends toaggregate (26, 28, 30) and both termini of the molecule are

disordered or relatively flexible (27, 32). Because the N-terminal region has been truncated in our streptavidin prep-aration by the deletion of the corresponding coding region, itis likely that the C-terminal region of the mature streptavidinis responsible for the aggregation, although participation ofthe N-terminal region cannot be excluded. Because bothterminal regions are rich in hydrophilic amino acid residues(13), ionic or hydrogen bonds might be responsible for theintermolecular interactions. With an efficient expressionsystem now in hand, it should be possible to explore this issuefurther along with many other outstanding issues relevant to

streptavidin's unusual properties and great usefulness innumerous biotechnology applications.

We thank F. W. Studier for providing us with E. coli cells,lysogens, bacteriophage, plasmids for the T7 expression system, andan unpublished paper and for many valuable suggestions: S. Taborfor pT7-7; M. Uhlen and R. Johansson for pUC-SAI; and W. A.Hendrickson, A. Plihler, and C. L. Smith for helpful suggestions anddiscussion. This work was supported by Grant CA 39782 from theNational Cancer Institute, National Institutes of Health.

1. Chaiet, L., Miller, T. W., Tausing, F. & Wolf, F. J. (1963)Antimicrob. Agents Chemother. 3, 28-32.

2. Chaiet, L. & Wolf, F. J. (1964) Arch. Biochem. Biophys. 106,1-5.

3. Green, N. M. (1975) Adv. Prot. Chem. 29, 85-133.4. Bayer, E. A. & Wilchek, M. (1980) Methods Biochem. Anal.

26, 1-46.5. Fuccillo, D. A. (1985) BioTechniques 3, 494-501.6. Buckland, R. M. (1986) Nature (London) 320, 557-558.7. Wilchek, M. & Bayer, E. A. (1988) Anal. Biochem. 171, 1-32.8. DeLange, R. J. & Huang, T.-S. (1971) J. Biol. Chem. 246,

698-709.9. Argarafia, C. E., Kuntz, I. D., Birken, S., Axel, R. & Cantor,

C. R. (1986) Nucleic Acids Res. 14, 1871-1882.10. Studier, F. W. & Moffatt, B. A. (1986) J. Mol. Biol. 189,

113-130.11. Studier, F. W., Rosenberg, A. H. & Dunn, J. J. (1989) Meth-

ods Enzymol., in press.12. Campbell, J. L., Richardson, C. C. & Studier, F. W. (1978)

Proc. NatI. Acad. Sci. USA 75, 2276-2280.13. Grodberg, J. & Dunn, J. J. (1988) J. Bacteriol. 170, 1245-1253.14. Dunn, J. J. & Studier, F. W. (1983) J. Mol. Biol. 166, 477-535.15. Tabor, S., Huber, H. E. & Richardson, C. C. (1987) J. Biol.

Chem. 262, 16212-16223.16. Tabor, S. & Richardson, C. C. (1985) Proc. Natl. Acad. Sci.

USA 82, 1074-1078.17. Rosenberg, A. H., Lade, B. N., Chui, D.-S., Lin, S.-W.,

Dunn, J. J. & Studier, F. W. (1987) Gene 56, 125-135.18. Chang, A. C. Y. & Cohen, S. N. (1978) J. Bacteriol. 134,

1141-1156.19. Moffatt, B. A. & Studier, F. W. (1987) Cell 49, 221-227.20. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Labora-tory, Cold Spring Harbor, NY).

21. Kyhe-Andersen, J. (1984) J. Biochem. Biophys. Methods 10,203-209.

22. Korenman, S. G. & O'Malley, B. W. (1967) Biochim. Biophys.Acta 140, 174-176.

23. Laemmli, U. K. (1970) Nature (London) 227, 680-685.24. Itzhaki, R. F. & Gill, D. M. (1964) Anal. Biochem. 9,401-410.25. Green, N. M. & Melamed, M. D. (1966) Biochem. J. 100,

614-621.26. Pahler, A., Hendrickson, W. A., Kolks, M. A. G., Argarafia,

C. E. & Cantor, C. R. (1987) J. Biol. Chem. 262, 13933-13937.27. Hendrickson, W. A., Pahler, A., Smith, J. L., Satow, Y.,

Merritt, E. A. & Phizackerley, R. P. (1989) Proc. Natl. Acad.Sci. USA 86, 2190-2194.

28. Bayer, E. A., Ben-Hur, H., Hiller, Y. & Wilchek, M. (1989)Biochem. J. 259, 369-376.

29. Hofmann, K., Wood, S. W., Brinton, C. C., Montibeller, J. A.& Finn, F. M. (1980) Proc. Natl. Acad. Sci. USA 77, 4666-4668.

30. Bayer, E. A., Ben-Hur, H., Gitlin, G. & Wilchek, M. (1986) J.Biochem. Biophys. Methods 13, 103-112.

31. Suter, M., Cazin, J., Jr., Butler, J. E. & Mock, D. M. (1988) J.Immunol. Methods 113, 83-91.

32. Weber, P. C., Ohlendorf, D. H., Wendoloski, J. J. & Sale-mme, F. R. (1989) Science 243, 85-88.

Proc. Natl. Acad. Sci. USA 87 (1990)