Eur. J. Biochem. 196,623-629 (1991) 0FEBS 1991 001429569100185K
Production and purification of recombinant human interleukin-5 from yeast and baculovirus expression systems Evan INGLEY ',Robert L. CUTLER', Ming-Chiu FUNG', Colin J. SANDERSON3 and Ian G. YOUNG'
' Division of Biochemistry and Molecular Biology and
Division of Clinical Sciences, John Curtin School of Medical Research, Australian National University, Canberra, Australia National Institute for Medical Research, Mill Hill, London, England
(Received November 13, 1990) - EJB 90 1336
A cDNA for human interleukin-5 (hIL-5) was created from the hIL-5 gene using site-directed mutagenesis to splice out the introns in vitro. This cDNA was expressed in yeast and baculovirus systems, utilizing in both cases an in-frame fusion to the pre sequence of the a-mating-type factor to direct secretion. The highest level of production was achieved from Sf9 cells using a baculovirus vector in serum-containing medium (2.7 mg/l), whereas in serum-free medium ten times less hIL-5 was produced. In the yeast system much lower levels of hIL-5 were produced (12.5 pg/l). Recombinant hIL-5 was purified to homogeneity from serum-free baculovirus cultures. The rhIL-5 consisted of a 30-kDa homodimer linked by disulfide bridging. The purified recombinant protein had a specific activity on murine BCLl cells of 1.5 x lo4 U/mg, of 3 x lo5 U/mg in the murine eosinophil differentiation factor assay, and 2.4 x lo7 U/mg in a human peripheral eosinophil maintenance assay.
Interleukin 5 (IL-5) is a T-lymphocyte-derived lymphokine which stimulates the maturation of eosinophils, a bone-marrow-derived granulocyte which is produced as part of the cellular response to parasitic infections and in allergic reactions. Mouse and human IL-5 are about 70% similar in their amino acid sequences and will cross react [l, 21. Mouse IL-5 has been found to have several biological activities, B-cell growth factor I1 activity [3], eosinophil differentiation factor activity [l, 41, IgA enhancing activity [5, 61 and also up-regulates functional IL-2 receptors on splenic B-cells [7]. Human IL-5 does not appear to have analogous activity on human Bcells [2], although it is very potent in the activation of eosinophils [8]. Human IL-5 has not yet been purified from natural systems but the gene encoding hIL-5 has recently been cloned and characterized based on similarity with the murine cDNA [91. Based on the predicted protein sequence, mature hIL-5 consists of 115 amino acids, contains two cysteine residues and two potential N-glycosylation sites. Our interest in the structure and biological activites of hIL-5 led us to seek an expression system in which hIL-5 would be produced in an active mature form. Yeast and baculovirus systems have sucCorrespondence to I. G. Young, Medical Molecular Biology Group, Division of Biochemistry and Molecular Biology, John Curtin School of Medical Research, Australian National University, P. 0. Box 334, Canberra City, Australian Capital Teritory 2601, Australia Abbreviations. hIL-5, human interleukin-5; rhIL-5, recombinant IL-5; NaC1/Pi, phosphate-buffered saline (120 mM NaCl and 10 mM KCI in 10 mM phosphate pH 7.4); mEDF, mouse eosinophil differentiation factor; hCFU, human colony forming units; AcMNPV, wildtype baculovirus; AcIL-5, recombinant hIL-5-producing baculovirus. Note. The novel nucleotide sequence data published here has been deposited with the EMBL/Genbank sequence data banks and is available under accession number 502971.
cessfully been used to produce many recombinant human and murine interleukins [10 - 121. Here we report the expression of recombinant hIL-5 (rhIL5) in yeast and baculovirus expression systems, the purification of rhIL-5 to homogeneity and the characterization of the recombinant protein.
MATERIALS AND METHODS Strains and transformations
Bacterial strains Escherichia coli. JM101, RZ1032 and DH5 were used and transformed as described [13]. Saccharomyces cerevisiae strain JNY5 (Mat a, trpl, leu2, pep4: ura3, suc2) was obtained from J. Ngsee (University of British Columbia, Canada). The strain was maintained and spheroplasts were prepared and transformed as previously described [14]. The mouse tumor cell line BCLl (in vivo) was maintained and prepared as described [15]. The insect cell line Sf9 and the wild-type baculovirus (AcMNPV) were kindly provided by M. D. Summers (Department of Entomology, Texas Agricultural & Medical University, USA). The baculovirus system was used as described in the manual [16]. Cotransfection of Sf9 cells and recombinant virus purification were essentially as described [17]. Serum-free insect medium was prepared as described [121. Recombinant D N A technology
General genetic engineering techniques including plasmid, M13 RF and single-stranded DNA isolation, enzymatic reactions, and hybridization experiments were essentially per-
624 formed as described by Maniatis et al. [18]. DNA sequencing was performed using the dideoxy-termination method [I 91. DNA oligonucleotide synthesis was by phosphoramidite chemistry using an Applied Biosystems Inc. 380 B synthesizer. All oligonucleotides were gel-purified as described [20]. Sitedirected mutagenesis using oligonucleotides was performed essentially as described by Zoller and Smith [21], except that M13 single-strand template DNA was grown in E. coli RZ1032 so as to increase the mutagenesis efficiency [22], by incorporating uracil into the DNA. Yeast cultures The yeast strain JNY5 was maintained on medium containing 1% yeast extract, 2% peptone and 2% glucose. Transformants were grown on selective medium in the absence of leucine, consisting of 0.7% yeast nitrogen base (Difco), 2% glucose, 0.5% ammonium sulphate, essential amino acids [23] (not including leucine), 20 pg adenine/ml and 20 pg uracil/ml. Cultures were grown at 30°C for 3 days in a 14-1 New Brunswick Scientific fermentor with aeration through a ring sparger at 10 l/min and agitation at 700 rpm using turbine propellers. The pH of the cultures was maintained at 7.0 by addition of ammonium hydroxide using a New Brunswick Scientific pH controller. Cultures were inoculated with 1 1 of over-night culture grown in selective medium, described above. Culture medium contained 20% glucose, 0.7% yeast nitrogen base, 0.5% ammonium sulphate, 25 mM sodium phosphate pH 7.0, 20 pg adenine/ml, 20 pg tryptophan/ml, 20 pg uracil/ml, 1% yeast extract ( < 10-kDa filtrate), 2% peptone (< 10-kDa filtrate). A solution of 30% glucose was continuously added after the first 6 h of fermentation at a rate of about 100 ml/h. Yeast cultures routinely reached an A600 of z 50. Biological assays All samples were equilibrated in phosphate-buffered saline (NaCl/P, = 120 mM NaCl and 10 mM KC1 in 10 mM phosphate pH 7.4) containing 0.02% Tween-20, using NAP-5 columns (Pharmacia) and filtered through 0 . 2 2 - ~ mpore-size sterile membranes before being assayed. BCLl assay. Human IL-5 was assayed on BCLl cells according to Dutton et al. [24]. This assay was routinely used in the protein purification experiments using recombinant mIL5 (supernatants from COS cells transfected with an mIL-5producing plasmid [25]) as a standard. mEDF u.ssuy. Human IL-5 was assayed on mouse bone marrow from mice that had been parasitised with Mesocestoidees corti according to Strath et al. [26]. Human periplzeral eosinophil maintenance assay. Human peripheral blood eosinophils were isolated according to Owen et al. [27] and various dilutions of hIL-5 were incubated with 5 x lo4 eosinophils in 200 p1 medium (RPMI-1640, antibiotics, 10% foetal calf serum, 2 mM glutamine). After two days of incubation. the number of surviving eosinophils was counted in a haemocytometer and differential stains of the cells were made. Protein unulysis
SDSjPAGE was performed according to Laemmli [28], except that the cross-linker used was piperazine diacrylamide (Bio-Rad). Isoelectric focusing was performed according to Poehling and Neuhoff [29]. Gels were silver-stained using the
method of Morrissey [30]. Protein concentrations were measured using the Bio-Rad protein assay kit, with y-globulin (Bio-Rad) as a standard. Amino acid analysis was performed on a Hewlett Packard 1090 Amino Quant system. N-terminal amino acid sequence analysis was carried out on an Applied Biosystems 477A protein sequencer. Deglycosylation experiments were performed using N-glycosidase F (Boehringer Mannheim), according to the manufacturers instructions. Protein pur fication All solutions contained 0.02% Tween-20 and 0.02% NaN3 and were filtered through 0.22-pm membranes and degassed for 30 min before use. Phenyl-Sepharose. Hydrophobicity chromatography was performed on a column (1.5 cm x 20 cm) of Phenyl-Sepharose CL-4B (Pharmacia) equilibrated in 1 M ammonium sulphate, 100 mM phosphate pH 7.3. Samples were adjusted to the above solution before being loaded onto the column at a rate of 0.5 ml/min. The column was washed with 50 ml equilibration solution before a linear descending gradient of 10 M ammonium sulphate was applied at a rate of 0.5 ml/min over 200 min. Fractions of 4 ml were collected. Subsequently 50 ml phosphate buffer and 50 ml double deionized and distilled water were applied to the column before a linear gradient of 0 - 30% ethylene glycol in double deionized and distilled water was applied. The column was controlled by an FPLC (Pharmacia) controller. Gel filtration. A Sephadex G-I00 super-fine (Pharmacia) column (1.2 cm x 60 cm) was equilibrated in 20 mM Tris/HCl pH 8.0. Samples (1 -2.5% of bed volume) were applied at a flow rate of 0.05 ml/min and 1-ml fractions were collected. Gel filtration chromatography was performed at 4°C. Anion exchange. A Mono Q HR 5 / 5 (Pharmacia) FPLC anion-exchange column was equilibrated in 20 mM Tris/HCl pH 8.0. Samples adjusted to the above solution were applied at a rate of 1 ml/min. The column was washed with one bed volume of buffer before a linear gradient of 0- 1 M NaCl in buffer was applied over 20 min; 1-ml fractions were collected. The column was controlled by an FPLC (Pharmacia) controller. RESULTS Construction of an rhIL-5 cDNA yeast expression vector The 2.83-kb EcoRI - PstI fragment of the hIL-5 gene, which includes all of the coding region, was prepared from the plasmid pEDFH-1 [9] and subcloned between the EcoRI and PstI sites of M13mp19 creating M13-IL5G. This construct (M13-IL5G) was used as a template for site-directed mutagenesis to splice out the introns in vitro (Figs 1 and 2A, B). Since the resultant construct (M13-IL5CI) contained a potential yeast transcription termination site (codons 93 103) it was modified in this region by replacing these codons with yeast-preferred codons using site-directed mutagenesis, producing M13-IL5C2 (Fig. 2C). The 0.37-kb CluI -BglII fragment from M13-IL5C2 was subcloned into a yeast expression construct, which contained a 1.59-kb HindIII BamHI fragment carrying the 3-phosphoglycerate kinase promoter [31], a 0.24-kb PstI - ClaI fragment of the yeast a-factor pre-pro leader sequence [32], a BarnHI - Pstl synthetic linker (35 bp) joining the 3-phosphoglycerate kinase promoter and the a-factor leader sequence and a 0.38-kb BgnI -Hind111 fragment carrying the 3-phosphoglycerate kinase gene terminator [31] in M13mp19 (HindIII site), to produce M13-IL5C3.
625
3
I - ,
. .-5'
, ,TTTRTGTAACTGCCGGTTTT~~TTTTTCACACCTCTTCT?TCT.
5 -RTTGRCGGCCRRRARRRRRRGTGTGGRGAR-3
Exon-3
Exon-4
A
1
Intron-1
3
Intron-2 980 bp
I
c T i:iGAG R C T c c T AR GG R C R A G G R,CAT G T A T T T T TSI C,T G G T T G R c A c G TG A c T T c T T . , , -5' 5'-CTGRTRGCCRATGAGRCTCTGRGGRTTCCTGTTCCTGTRCATRAAAATCflCC~RCTGTGCACT-3 I I I I
- , , , T GR G A CG R C T A T CG G T T A
Exon- 1
Exon-2
B
Exon-3
E m
Fig. 1 . Summary of the specific binding of the oligonucleotides E34 and El23 to M13-IL5G. The oligonucleotide E34 (A) was used to splice out the first intron and El23 (B) was used to splice out the second and third introns in vitro by site-directed mutagenesis
The BamHI site in the 3-phosphoglycerate kinase promoter fragment was created by site-directed mutagenesis using the oligonucleotide 5'-TTTACAACGGATCCAAAAACAATG3'. The ClaI site in the a-factor leader was created by sitedirected mutagenesis using the oligonucleotide 5'-GAAGGGGTATCGATGGATAAAAG-3'. Since it has been shown that the pre section alone of the a-factor is sufficient and more efficient at directing secretion in the case of expressing hIL1B [lo], we subsequently eliminated the pro section of the afactor leader sequence by site-directed mutagenesis using the oligonucleotide 5'-GCATTAGCTGCTCCAGTCGAAATTCCCACAAGTGC-3'. The first and third codon of the mature rhIL-5 were also altered by site-directed mutagenesis (using the above oligonucleotide) in order to mimic the efficient signal peptidase cleavage site of the pre-a hIL-1B construct [lo]. This rhIL-5 yeast expression unit (2.4-kb HindIII fragment) was subcloned into the yeast shuttle vector YEpl3, HindIII site (Fig. 3), to create YEpIL-5.
cotransfected into Sf9 cells together with purified wild-type baculovirus (AcMNPV) DNA. The culture supernatant was collected after 11 days and the recombinant virus purified by limiting dilution and dot-blot hybridization (the hIL-5 probe used wasa0.31-kbStyI-BglIIfragment ofM13ILSC1, bluntend repaired and subcloned into the SmaI site of pUC18, making pUCIL-5). The first round of screening produced four positives at dilution of which one was further purified as it contained wild-type virus. The second round of screening produced 12 positives at l o p 6 dilution and no positives at l o p 8dilution. One positive was further purified as it contained some wild-type virus. The third round of screening produced 16 positives at dilution and three at lo-' dilution, two of the three lo-' positives contained no wild-type virus and one of these was chosen to make virus stocks and to be used subsequently in rhIL-5 production. The recombinant virus was designated AcIL-5. Expression of rhIL-5 by the baculovirus system
Expression of rhIL-5 in yeast Although the 3-phosphoglycerate kinase promoter used can direct the production of high levels of this kinase in yeast ( z 50% of cellular protein), disappointingly low levels of hIL5 ( z 12.5 pg/l) were produced in JNY5 containing the plasmid YEpIL-5. Accordingly the hIL-5-coding region was transferred from the yeast expression construct to a baculovirus expression vector. Construction of an rhIL-5-secreting baculovirus The 0.43-kb BamHI - BglII fragment of MI 3-IL5C3 which encodes the prea hIL-5 was subcloned into the BamHI site of pVL941 [I61 producing pBVIL-5. This plasmid was
Recombinant hIL-5 was produced at 0.27 Fg/ml in largescale serum-free baculovirus cultures and at 2.7 pg/ml in serum-containing cultures. The rhIL-5 represented about 10% of the secreted protein in serum-free cultures. Purijication of rhIL-5 Recombinant hIL-5 was purified from a 250-ml serumfree culture of Sf9 cells infected with the rhIL-5-producing baculovirus (AcIL-5) and the supernatant collected 54 h after infection. Pheizyl-Sepharose. The majority of the rhIL-5 eluted from the column between 0.6 M and 0 M ammonium sulphate, with a small amount eluting in the double deionised and distilled
626 Mature hIL-5
F hla h o ' l b r 3 , - , , . C T FRC GRTGCRCRT CGGTRGGGGTGTCTTRTGGGTFTTCR.. .-5' -GkTRCGTGTRfGCCRTCCCCRCRGRRRTTCC-3
!
R TCG RQ-GUC
a n
A
3'-.,.TRT
R A G m
M=tAvLy7
U Clo I
KEX II cleavage site
*
3 gT- ~ RT TRTRCGRTR~CGT CRTT- CG TR TRTRTRGCRTC GC RGRTRTCTRGR-~~G T C , .,-5' I
I
Bgl I1
B
BQID
3 - - , , , RRGTTTTTGRACRGGRRTTRTTTCTTTRTGTRACTGCCGGTTTTTTTTTTCRCRCCTCTTCTTTC ; -5' 5 ' - C fl R RR RC T TGTCC J+'q&G.& GT&ETGJC~~ CRR g~f!f+RFlG TG T G G fl GR fl G fi fl -3 Leu Ile Lys Lys 5 1 Ile Asp Gly Gln Lys Lys *93 +94 +95 +96 t97 +98 tw ti01 +im I
cm3-103
C
Fig. 2. Summary ojtlie specific binding of the oligonucleotides BGII and CLI to M13-IL5G, and CD93 - 103 to M13-IL5CI. The oligonucleotide CL1 (A) was used to insert a ClaI site and a KEXII protease site at the beginning of the mature hIL-5 coding region; BGII (B) was used to insert a BgnI site juxtaposed 3' to the termination codon of the hIL-5. The oligonucleotide CD93 - 103 (C) was used to eliminate the potential yeast transcription termination site by replacing the codons 93 - 103 with 3-phosphoglycerate kinase preferred codons
*
PGK Promoter 5' 3'.
I
Pre -a Factor
I
hIL-5
~ ~ ~ V bm ----------------
Met&rPkh&U& ~
TTRTTCGCRGCRTCCTCCGCRTTRGCTGCTCCRGTCGRRRTTCCCRCfl RATRRGCGTCGTRGGRGGCGTRRTCGACGAGGTCRGCTTTRRGGGTGT
TTTTTRCRRCG RARARTGTTGC
Bum H1
Linker
~
~
..3' . . .5 '
Pst 1
PGK Terminator
PGK Promoter
Fig. 3. Summary o j t h e yeast expression vector YEpIL-5. The construct contained the 3-phosphoglycerate kinase (PGK) promoter, an in-frame fusion of the pre a-factor to the mature coding region of hIL-5, the PGK terminator, the yeast LEU2 chromosomal gene, a section of pBR322, and a section of the yeast 2-wm episome. The linker was created by annealing the oligonucleotides T-GATCCAAAAAAAATGTCTTTCCCATCTATCTTCACTGCA-3' (5'-phosphorylated) and 5'-GTGAAGATAGATGGGAAAGACATTTTTTTTG-3'. The asterisk (*) indicates where the second amino acid of the pre-x-factor was changed from Arg to Ser so as to resemble more conventional signal sequences
water fractions (Fig. 4). The positive fractions 18 - 34 were pooled and used for further purification. Gel filtration. The pooled fractions from the PhenylSepharose column were concentrated to 2.5 ml using Centricons (Amicon) and then loaded onto the Sephadex G-100 column. Two peaks of hIL-5 activity eluted from the column. The major peak corresponded to a size of about 30 kDa (frac-
tion numbers 52 - 63) and the minor peak corresponded to about 60 kDa (Fig. 5). The z 30-kDa positive fractions were pooled and used for further purification. Anion exchange. The positive fraction pool from the gel filtration column was loaded onto the Mono Q column. The rhIL-5 did not bind to the column but this step removed other contaminating proteins.
M
627
.
Glycol %
0
10
20
30
40
50
60
70
80
z
Fraction Number
Fig. 4. Phenyl-Sepharose chromatography of rhZL-5. The column (1.5 x 20 cm) of Phenyl-Sepharose C L 4 B (Pharmacia) was equilibrated in 1 M ammonium sulphate/100 mM sodium phosphate pH'7.3. The Sfp supernatant (54 h after infection with AcIL-5) was adjusted to the above solution before being loaded onto the column at a rate of 0.5 ml/min. The column was washed with 50 ml equilibration solution before a linear gradient of 1-0 M ammonium sulphate was applied at a rate of 0.5ml/min over 200 min. Fractions of 4 m l were collected. Subsequently, 50 ml phosphate pH 7.3 and 50 ml double deionized and distilled water (ddHzO) were applied before a linear gradient of 030% ehtylene glycol in ddHzO was applied. The column was controlled by an FPLC controller. Fractions were assayed on BCL, cells according __ to Duttdn et ai.-[24]
-.I_
,
, 2000 6 7
4 3 25 ' 3
................
--"--
--b
20000
66 42 I
Q
0.1 0 0
kDa
0
m
a-
1
31
21
14
I
I
I
4
3 2 1
9000
1
Fraction Number Fig. 5. Gelfiltration chromatography rhIL-5. A Sephadex G-100 superfine (Pharmacia) column (1.2 x 60 cm) was equilibrated in 20 mM TrisjHCl pH 8.0. The sample (positive fractions from the PhenylSepharose column concentrated to 2 ml) was applied at a flow rate of 0.05 ml/min, and I-ml fractions were collected. Fractionation was performed at 4°C. Fractions were assayed on BCLl cells according to Dutton et al. [24]. Numbers at the top indicate the molecular mass (in kDa) of marker proteins Gel Slice Number
Physico-chemical characterization of rhIL-5 The purity of the rhIL-5 was checked on a 12.5% SDS/ PAGE gel under reducing and non-reducing conditions. This showed that the rhIL-5 was 95% pure after passage through the Mono Q column (Fig. 6), and consisted of a homodimer linked together by disulfide bridging, as is the case with natural mIL-5 [33]. The predicted protein sequence of human IL-5 contains two cysteine residues, as does mIL-5, which could participate in dimer formation. The rhIL-5 appeared as two bands of similar intensity at 30 and 32 kDa under non-reducing conditions. Under reducing conditions, however, the smaller-sized monomer band (14 kDa) was less intense than the larger monomer band (16 kDa). In gel elution experiments (Fig. 6) the biological activity could only be eluted under nonreducing conditions from the z 30-kDa area and not under reducing conditions, even after the reducing agent was removed by extensive dialysis. The rhIL-5 was shown to have
Fig. 6. SDSjPAGE analysis of purified rhIL-5 and gel elution analysis. Purified recombinant hIL-5 (40 ng) was analysed on a 12.5% SDSj PAGE gel under reducing (lane 3) and non-reducing (lane 1) conditions. Buffer only was run in lane 2. Molecular mass markers were in lane 4 (values at top). The gel was silver-stained. Duplicate samples of reduced (+) and non-reduced (m) hIL-5 were run in other lanes and, after electrophoresis, 2-mm gel slices were taken from these two lanes and placed into 0.5 ml NaCI/P, containing 0.02% Tween-20 and 1 mg/ml bovine serum albumin. The eluted material was extensively dialysed against NaCl/P, before testing in the BCLl assay
an isoelectric point of about 7.2. Deglycosylation of the rhIL5 had no effect on its biological activity. Deglycosylated rhIL5 migrated as an approximately 13-kDa band under reducing conditions (data not shown). The N-terminal sequence of the first 11 residues of the rhIL-5 was Ala-Pro-Val-Glu-Ile-ProThr-Ser-Ala-Leu-Val, as expected, showing that the signal
628 Table 1. Summary of the purfication of rhIL-5 f r o m the baculovirus system Protein quantification for steps 1, 2 and 3 were by Bio-Rad protein assay kit and for step 4 by amino acid analysis Purification step
1. Supernatant 2. PhenylSepharose 3. Gel filtration 4. Anion exchange
Total protein
Total activity
Specific activity
Recovery
mg
U
U/Pg
Yo
32.5
8500
0.26
100
the biological activity of the rhIL-5. The monomer is not biologically active and does not readily form into biologically active dimers after removal of the reducing agent. The purified rhIL-5 has a low specific activity in the BCLl and mEDF assays, and a high specific activity in a human peripheral eosinophil maintenance assay, and a calculated high specific activity in the hCFU assay. The rhIL-5 is biologically inactive below pH 7.0 (which appears to be due to aggregation) and above pH 9.0.
4500 4000 4000
1.13 6.67 15.0
53 47 47
We thank Gary Mayo for synthesizing the oligonucleotides and Bern Presnell for N-terminal sequence analysis and amino acid analysis.
4.0 0.600 0.2667
REFERENCES peptide from the yeast a-factor had been correctly processed. The amino acid analysis of the rhIL-5 showed the predicted composition for all amino acids, except for aspartic acid which was higher than predicted, apparently due to a low-molecularmass contaminant in the buffer used. Based on the amino acid analysis, an accurate specific activity was calculated (see above). A summary of the purification is presented in Table 1. Purified rhIL-5 was stable in the pH range 7.0-9.0, with and without salt (0.15 M NaCl), while below pH 7.0 and above pH 9.0 there was a great reduction in its biological activity (data not shown). A pH of 7.4 would appear to be optimal for maintaining the biological activity. A gel filtration column of rhIL-5 at pH 5.0 showed the rhIL-5 protein eluted in the breakthrough fractions, suggesting that the loss of activity observed at low pH values may be due to aggregation.
Biological u s ~ u y . ~ The purified rhIL-5 showed a specific activity of 1.5 x lo4 U/mg in the BCLl assay, 3.0 x lo5 U/mg in the mEDF assay, and 2.4 x lo7 U/mg in a human peripheral eosinophil
maintenance assay. A specific activity in the human colonyforming units (hCFU) assay of 5 x 10’ U/mg was calculated by comparison of the activity of the rhIL-5 in the mEDF assay with that of a standard rhIL-5 of known activity in both the mEDF and hCFU assays. DISCUSSION We describe here the expression of mature, biologically active rhIL-5 in yeast and baculovirus expression systems. The highest levels of rhIL-5 were obtained from the baculovirus system in serum-containing medium (2.7 mg/l) while ten times less rhIL-5 was produced in serum-free medium. The yeast produced the lowest level of rhIL-5 (12.5 pg/l). An efficient purification procedure was developed for rhIL-5 giving good recoveries of protein with high biological activity. After three purification steps, the rhIL-5 appeared to be pure. Interestingly the insect cells correctly processed the yeast a-factor pre sequence, as N-terminal sequence analysis revealed the expected sequence of mature hIL-5 in the secreted protein. Physico-chemical analysis of the rhIL-5 indicated that it exists as a homodimer linked together by disulphide bonds, the dimer being about 30 kDa in mass, with an isoelectric point of about 7.2. Deglycosylation of the rhIL-5 produced a protein with a molecular mass in agreement with that predicted from the encoded polypeptide sequence of 115 amino acids ( z13 kDa, for the monomer). Deglycosylation did not affect
1. Metcalf, D., Cutler, R. L. & Nicola, N. A. (1983) Blood61,9991005. 2. Clutterbuck, E. J., Shields, J. G., Gordon, J., Smith, S. H., Boyd, A., Callard, R. E., Campbell, H. D., Young, I. G. & Sanderson, C. J. (1987) Eur. J. Immunol. 17, 1743-1750. 3. Swain, S. L. & Dutton, R. W. (1982) J . Exp. Med. 156, 18211834. 4. Ydmaguchi, Y., Suda, T., Suda, J., Eguchi, M., Miura, Y., Harada, N., Tominagd, A. & Takatsu, K . (1988) J . Exp. Med. 167,43- 56. 5. Coffman, R. L., Shrader, B., Carty, J., Mosmann, T. R. & Bond, M. W. (1987) J. Immunol. 139, 3685-3690. 6. Harriman, G. R., Kunimoto, D. Y., Elliot, J. F., Paetkau, V. & Strober, W. (1988) J . Immunol. 140,3033-3039. 7. Loughan, M. S., Takatsu, K., Harada, N. & Nossal, G. J. V. (1987) Proc. Nut1 Acad. Sci. U S A 84, 5399- 5403. 8. Saito, H., Hatake, K.. Dvorak, A. M., Leiferman, K. M., Donnenberg, A. D., Arai, N., Ishizaka, K . &Ishizaka, T. (1988) Proc. Nut1 Acad. Sci. U S A 85, 2288 - 2292. 9. Campbell, H. D., Tucker, W. Q. L., Hort, Y., Martinson, M., Mayo, G., Clutterbuck, E. J., Sanderson, C. J. & Young, I. G. (1 987) Proc. Natl Acad. Sci. U S A 84, 6629 - 6633. 10. Ernst, J. F. (1988) D N A 7, 355-360. 11. Price, V.. Mochizuki, D., March, C. J., Cosman, D., Deeley, M. C., Klinke, R., Clevenger, W., Gillis, S., Baker, P. & Urdal, D. (1987) Gene 55,287-293. 12. Maiorella, B., Inlow, D., Shauger, A. & Hardno, D. (1988) Biotechnology 6, 1406- 1410. 13. Hanahan, D. (1983) J . Mol. Bid. 166, 557-580. 14. Hinnen, A,, Hicks, J. B. & Fink, G. R. (1978) Proc. Natl Acad. Sci. U S A 75, 1929-1933. 15. Isakson, P. C., Krolick, K . A,, Uhr, J. W. & Vitetta, E. S. (1980) J . Immunol. 125, 886-892. 16. Summers, M. D. & Smith, G. E. (1987) A manual of methods f o r haculovirus vectors and insect cell culture procedures, Texas A& M University, College Station TX. 17. Fung, M.-C., Chiu, K. Y. M., Weber, T., Chang, T.-W. & Chang, N. T. (1988) J . Virol. Methods 19, 33-42. 18. Maniatis, T., Fritsh, E. F. & Sambrook, J. (1982) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory, Cold Spring Harbor NY. 19. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Nut1 Acad. Sci. U S A 74, 5463 - 5461. 20. Atkinson, T. C., Smith, M. & Zoller, M. J. (1984) Solid phase phosphite triester synthesis of oligodeoxyribonucleotides. Manual f o r advanced techniques in molecular cloning course, Cold Spring Harbor Laboratory, Cold Spring Harbor NY. 21. Zoller, M. J. & Smith, M. (1983) Methods Enzymol. 100, 468500. 22. Kunkel, T. A. (1985) Proc. Natl Acad. Sci. U S A 82,488-492. 23. Spencer, J. F. T., Spencer, D. M. & Smith, A. R. W. (1980) Y m s t genetics, p. 533, John Wiley, New York. 24. Dutton, R. W., Wetzel, G. D. & Swain, S . L. (1984) J . Immunol. 132.2451 -2456.
25. Campbell, H. D., Sanderson, C. J., Wang, Y., Hort, Y., Martinson, M. E., Tucker, W. Q. J., Stellwagen, A,, Strath, M. & Young, I. G. (1980) Eur. J. Biochem. 174,345-352. 26. Strath, M., Warren, D. J. & Sanderson, C. J. (1985) J. Zmmunol. Methods 83,209 - 215. 27. Owen, W. F., Rothenberg, M. E., Peterson, J., Weller, P. F., Silberstein, D., Sheffer, A. L., Stevens, R. L., Soberman, R. J. & Austen, K. F. (1989) J. Exp. Med. 170, 343-348. 28. Laemmli, U. K . (1970) Nature 227, 680-685.
29. Poehling, H.-M. & Neuhoff, J. (1980) Electrophoresis I , 90102. 30. Morrissey, J. H. (1981) Anal. Biochem. 117, 307-310. 31. Hitzeman, R. A,, Hagie, F. E., Hayflick, J. S., Chen, C. Y., Seeburg, P. H. & Derynck, R. (1982) Nucleic Acids Res. 10, 7791 - 7808. 32. Kurjan, J. & Herskowitz, I. (1982) Cell 30, 933 -943. 33. McKenzie, D. T., Filutowicz, H. I., Swain, S. L. & Dutton, R. W. (1987) J. Immunol. 139,2661 -2668.
Production and purification of recombinant human interleukin-5 from yeast and baculovirus expression systems Evan INGLEY ',Robert L. CUTLER', Ming-Chiu FUNG', Colin J. SANDERSON3 and Ian G. YOUNG'
' Division of Biochemistry and Molecular Biology and
Division of Clinical Sciences, John Curtin School of Medical Research, Australian National University, Canberra, Australia National Institute for Medical Research, Mill Hill, London, England
(Received November 13, 1990) - EJB 90 1336
A cDNA for human interleukin-5 (hIL-5) was created from the hIL-5 gene using site-directed mutagenesis to splice out the introns in vitro. This cDNA was expressed in yeast and baculovirus systems, utilizing in both cases an in-frame fusion to the pre sequence of the a-mating-type factor to direct secretion. The highest level of production was achieved from Sf9 cells using a baculovirus vector in serum-containing medium (2.7 mg/l), whereas in serum-free medium ten times less hIL-5 was produced. In the yeast system much lower levels of hIL-5 were produced (12.5 pg/l). Recombinant hIL-5 was purified to homogeneity from serum-free baculovirus cultures. The rhIL-5 consisted of a 30-kDa homodimer linked by disulfide bridging. The purified recombinant protein had a specific activity on murine BCLl cells of 1.5 x lo4 U/mg, of 3 x lo5 U/mg in the murine eosinophil differentiation factor assay, and 2.4 x lo7 U/mg in a human peripheral eosinophil maintenance assay.
Interleukin 5 (IL-5) is a T-lymphocyte-derived lymphokine which stimulates the maturation of eosinophils, a bone-marrow-derived granulocyte which is produced as part of the cellular response to parasitic infections and in allergic reactions. Mouse and human IL-5 are about 70% similar in their amino acid sequences and will cross react [l, 21. Mouse IL-5 has been found to have several biological activities, B-cell growth factor I1 activity [3], eosinophil differentiation factor activity [l, 41, IgA enhancing activity [5, 61 and also up-regulates functional IL-2 receptors on splenic B-cells [7]. Human IL-5 does not appear to have analogous activity on human Bcells [2], although it is very potent in the activation of eosinophils [8]. Human IL-5 has not yet been purified from natural systems but the gene encoding hIL-5 has recently been cloned and characterized based on similarity with the murine cDNA [91. Based on the predicted protein sequence, mature hIL-5 consists of 115 amino acids, contains two cysteine residues and two potential N-glycosylation sites. Our interest in the structure and biological activites of hIL-5 led us to seek an expression system in which hIL-5 would be produced in an active mature form. Yeast and baculovirus systems have sucCorrespondence to I. G. Young, Medical Molecular Biology Group, Division of Biochemistry and Molecular Biology, John Curtin School of Medical Research, Australian National University, P. 0. Box 334, Canberra City, Australian Capital Teritory 2601, Australia Abbreviations. hIL-5, human interleukin-5; rhIL-5, recombinant IL-5; NaC1/Pi, phosphate-buffered saline (120 mM NaCl and 10 mM KCI in 10 mM phosphate pH 7.4); mEDF, mouse eosinophil differentiation factor; hCFU, human colony forming units; AcMNPV, wildtype baculovirus; AcIL-5, recombinant hIL-5-producing baculovirus. Note. The novel nucleotide sequence data published here has been deposited with the EMBL/Genbank sequence data banks and is available under accession number 502971.
cessfully been used to produce many recombinant human and murine interleukins [10 - 121. Here we report the expression of recombinant hIL-5 (rhIL5) in yeast and baculovirus expression systems, the purification of rhIL-5 to homogeneity and the characterization of the recombinant protein.
MATERIALS AND METHODS Strains and transformations
Bacterial strains Escherichia coli. JM101, RZ1032 and DH5 were used and transformed as described [13]. Saccharomyces cerevisiae strain JNY5 (Mat a, trpl, leu2, pep4: ura3, suc2) was obtained from J. Ngsee (University of British Columbia, Canada). The strain was maintained and spheroplasts were prepared and transformed as previously described [14]. The mouse tumor cell line BCLl (in vivo) was maintained and prepared as described [15]. The insect cell line Sf9 and the wild-type baculovirus (AcMNPV) were kindly provided by M. D. Summers (Department of Entomology, Texas Agricultural & Medical University, USA). The baculovirus system was used as described in the manual [16]. Cotransfection of Sf9 cells and recombinant virus purification were essentially as described [17]. Serum-free insect medium was prepared as described [121. Recombinant D N A technology
General genetic engineering techniques including plasmid, M13 RF and single-stranded DNA isolation, enzymatic reactions, and hybridization experiments were essentially per-
624 formed as described by Maniatis et al. [18]. DNA sequencing was performed using the dideoxy-termination method [I 91. DNA oligonucleotide synthesis was by phosphoramidite chemistry using an Applied Biosystems Inc. 380 B synthesizer. All oligonucleotides were gel-purified as described [20]. Sitedirected mutagenesis using oligonucleotides was performed essentially as described by Zoller and Smith [21], except that M13 single-strand template DNA was grown in E. coli RZ1032 so as to increase the mutagenesis efficiency [22], by incorporating uracil into the DNA. Yeast cultures The yeast strain JNY5 was maintained on medium containing 1% yeast extract, 2% peptone and 2% glucose. Transformants were grown on selective medium in the absence of leucine, consisting of 0.7% yeast nitrogen base (Difco), 2% glucose, 0.5% ammonium sulphate, essential amino acids [23] (not including leucine), 20 pg adenine/ml and 20 pg uracil/ml. Cultures were grown at 30°C for 3 days in a 14-1 New Brunswick Scientific fermentor with aeration through a ring sparger at 10 l/min and agitation at 700 rpm using turbine propellers. The pH of the cultures was maintained at 7.0 by addition of ammonium hydroxide using a New Brunswick Scientific pH controller. Cultures were inoculated with 1 1 of over-night culture grown in selective medium, described above. Culture medium contained 20% glucose, 0.7% yeast nitrogen base, 0.5% ammonium sulphate, 25 mM sodium phosphate pH 7.0, 20 pg adenine/ml, 20 pg tryptophan/ml, 20 pg uracil/ml, 1% yeast extract ( < 10-kDa filtrate), 2% peptone (< 10-kDa filtrate). A solution of 30% glucose was continuously added after the first 6 h of fermentation at a rate of about 100 ml/h. Yeast cultures routinely reached an A600 of z 50. Biological assays All samples were equilibrated in phosphate-buffered saline (NaCl/P, = 120 mM NaCl and 10 mM KC1 in 10 mM phosphate pH 7.4) containing 0.02% Tween-20, using NAP-5 columns (Pharmacia) and filtered through 0 . 2 2 - ~ mpore-size sterile membranes before being assayed. BCLl assay. Human IL-5 was assayed on BCLl cells according to Dutton et al. [24]. This assay was routinely used in the protein purification experiments using recombinant mIL5 (supernatants from COS cells transfected with an mIL-5producing plasmid [25]) as a standard. mEDF u.ssuy. Human IL-5 was assayed on mouse bone marrow from mice that had been parasitised with Mesocestoidees corti according to Strath et al. [26]. Human periplzeral eosinophil maintenance assay. Human peripheral blood eosinophils were isolated according to Owen et al. [27] and various dilutions of hIL-5 were incubated with 5 x lo4 eosinophils in 200 p1 medium (RPMI-1640, antibiotics, 10% foetal calf serum, 2 mM glutamine). After two days of incubation. the number of surviving eosinophils was counted in a haemocytometer and differential stains of the cells were made. Protein unulysis
SDSjPAGE was performed according to Laemmli [28], except that the cross-linker used was piperazine diacrylamide (Bio-Rad). Isoelectric focusing was performed according to Poehling and Neuhoff [29]. Gels were silver-stained using the
method of Morrissey [30]. Protein concentrations were measured using the Bio-Rad protein assay kit, with y-globulin (Bio-Rad) as a standard. Amino acid analysis was performed on a Hewlett Packard 1090 Amino Quant system. N-terminal amino acid sequence analysis was carried out on an Applied Biosystems 477A protein sequencer. Deglycosylation experiments were performed using N-glycosidase F (Boehringer Mannheim), according to the manufacturers instructions. Protein pur fication All solutions contained 0.02% Tween-20 and 0.02% NaN3 and were filtered through 0.22-pm membranes and degassed for 30 min before use. Phenyl-Sepharose. Hydrophobicity chromatography was performed on a column (1.5 cm x 20 cm) of Phenyl-Sepharose CL-4B (Pharmacia) equilibrated in 1 M ammonium sulphate, 100 mM phosphate pH 7.3. Samples were adjusted to the above solution before being loaded onto the column at a rate of 0.5 ml/min. The column was washed with 50 ml equilibration solution before a linear descending gradient of 10 M ammonium sulphate was applied at a rate of 0.5 ml/min over 200 min. Fractions of 4 ml were collected. Subsequently 50 ml phosphate buffer and 50 ml double deionized and distilled water were applied to the column before a linear gradient of 0 - 30% ethylene glycol in double deionized and distilled water was applied. The column was controlled by an FPLC (Pharmacia) controller. Gel filtration. A Sephadex G-I00 super-fine (Pharmacia) column (1.2 cm x 60 cm) was equilibrated in 20 mM Tris/HCl pH 8.0. Samples (1 -2.5% of bed volume) were applied at a flow rate of 0.05 ml/min and 1-ml fractions were collected. Gel filtration chromatography was performed at 4°C. Anion exchange. A Mono Q HR 5 / 5 (Pharmacia) FPLC anion-exchange column was equilibrated in 20 mM Tris/HCl pH 8.0. Samples adjusted to the above solution were applied at a rate of 1 ml/min. The column was washed with one bed volume of buffer before a linear gradient of 0- 1 M NaCl in buffer was applied over 20 min; 1-ml fractions were collected. The column was controlled by an FPLC (Pharmacia) controller. RESULTS Construction of an rhIL-5 cDNA yeast expression vector The 2.83-kb EcoRI - PstI fragment of the hIL-5 gene, which includes all of the coding region, was prepared from the plasmid pEDFH-1 [9] and subcloned between the EcoRI and PstI sites of M13mp19 creating M13-IL5G. This construct (M13-IL5G) was used as a template for site-directed mutagenesis to splice out the introns in vitro (Figs 1 and 2A, B). Since the resultant construct (M13-IL5CI) contained a potential yeast transcription termination site (codons 93 103) it was modified in this region by replacing these codons with yeast-preferred codons using site-directed mutagenesis, producing M13-IL5C2 (Fig. 2C). The 0.37-kb CluI -BglII fragment from M13-IL5C2 was subcloned into a yeast expression construct, which contained a 1.59-kb HindIII BamHI fragment carrying the 3-phosphoglycerate kinase promoter [31], a 0.24-kb PstI - ClaI fragment of the yeast a-factor pre-pro leader sequence [32], a BarnHI - Pstl synthetic linker (35 bp) joining the 3-phosphoglycerate kinase promoter and the a-factor leader sequence and a 0.38-kb BgnI -Hind111 fragment carrying the 3-phosphoglycerate kinase gene terminator [31] in M13mp19 (HindIII site), to produce M13-IL5C3.
625
3
I - ,
. .-5'
, ,TTTRTGTAACTGCCGGTTTT~~TTTTTCACACCTCTTCT?TCT.
5 -RTTGRCGGCCRRRARRRRRRGTGTGGRGAR-3
Exon-3
Exon-4
A
1
Intron-1
3
Intron-2 980 bp
I
c T i:iGAG R C T c c T AR GG R C R A G G R,CAT G T A T T T T TSI C,T G G T T G R c A c G TG A c T T c T T . , , -5' 5'-CTGRTRGCCRATGAGRCTCTGRGGRTTCCTGTTCCTGTRCATRAAAATCflCC~RCTGTGCACT-3 I I I I
- , , , T GR G A CG R C T A T CG G T T A
Exon- 1
Exon-2
B
Exon-3
E m
Fig. 1 . Summary of the specific binding of the oligonucleotides E34 and El23 to M13-IL5G. The oligonucleotide E34 (A) was used to splice out the first intron and El23 (B) was used to splice out the second and third introns in vitro by site-directed mutagenesis
The BamHI site in the 3-phosphoglycerate kinase promoter fragment was created by site-directed mutagenesis using the oligonucleotide 5'-TTTACAACGGATCCAAAAACAATG3'. The ClaI site in the a-factor leader was created by sitedirected mutagenesis using the oligonucleotide 5'-GAAGGGGTATCGATGGATAAAAG-3'. Since it has been shown that the pre section alone of the a-factor is sufficient and more efficient at directing secretion in the case of expressing hIL1B [lo], we subsequently eliminated the pro section of the afactor leader sequence by site-directed mutagenesis using the oligonucleotide 5'-GCATTAGCTGCTCCAGTCGAAATTCCCACAAGTGC-3'. The first and third codon of the mature rhIL-5 were also altered by site-directed mutagenesis (using the above oligonucleotide) in order to mimic the efficient signal peptidase cleavage site of the pre-a hIL-1B construct [lo]. This rhIL-5 yeast expression unit (2.4-kb HindIII fragment) was subcloned into the yeast shuttle vector YEpl3, HindIII site (Fig. 3), to create YEpIL-5.
cotransfected into Sf9 cells together with purified wild-type baculovirus (AcMNPV) DNA. The culture supernatant was collected after 11 days and the recombinant virus purified by limiting dilution and dot-blot hybridization (the hIL-5 probe used wasa0.31-kbStyI-BglIIfragment ofM13ILSC1, bluntend repaired and subcloned into the SmaI site of pUC18, making pUCIL-5). The first round of screening produced four positives at dilution of which one was further purified as it contained wild-type virus. The second round of screening produced 12 positives at l o p 6 dilution and no positives at l o p 8dilution. One positive was further purified as it contained some wild-type virus. The third round of screening produced 16 positives at dilution and three at lo-' dilution, two of the three lo-' positives contained no wild-type virus and one of these was chosen to make virus stocks and to be used subsequently in rhIL-5 production. The recombinant virus was designated AcIL-5. Expression of rhIL-5 by the baculovirus system
Expression of rhIL-5 in yeast Although the 3-phosphoglycerate kinase promoter used can direct the production of high levels of this kinase in yeast ( z 50% of cellular protein), disappointingly low levels of hIL5 ( z 12.5 pg/l) were produced in JNY5 containing the plasmid YEpIL-5. Accordingly the hIL-5-coding region was transferred from the yeast expression construct to a baculovirus expression vector. Construction of an rhIL-5-secreting baculovirus The 0.43-kb BamHI - BglII fragment of MI 3-IL5C3 which encodes the prea hIL-5 was subcloned into the BamHI site of pVL941 [I61 producing pBVIL-5. This plasmid was
Recombinant hIL-5 was produced at 0.27 Fg/ml in largescale serum-free baculovirus cultures and at 2.7 pg/ml in serum-containing cultures. The rhIL-5 represented about 10% of the secreted protein in serum-free cultures. Purijication of rhIL-5 Recombinant hIL-5 was purified from a 250-ml serumfree culture of Sf9 cells infected with the rhIL-5-producing baculovirus (AcIL-5) and the supernatant collected 54 h after infection. Pheizyl-Sepharose. The majority of the rhIL-5 eluted from the column between 0.6 M and 0 M ammonium sulphate, with a small amount eluting in the double deionised and distilled
626 Mature hIL-5
F hla h o ' l b r 3 , - , , . C T FRC GRTGCRCRT CGGTRGGGGTGTCTTRTGGGTFTTCR.. .-5' -GkTRCGTGTRfGCCRTCCCCRCRGRRRTTCC-3
!
R TCG RQ-GUC
a n
A
3'-.,.TRT
R A G m
M=tAvLy7
U Clo I
KEX II cleavage site
*
3 gT- ~ RT TRTRCGRTR~CGT CRTT- CG TR TRTRTRGCRTC GC RGRTRTCTRGR-~~G T C , .,-5' I
I
Bgl I1
B
BQID
3 - - , , , RRGTTTTTGRACRGGRRTTRTTTCTTTRTGTRACTGCCGGTTTTTTTTTTCRCRCCTCTTCTTTC ; -5' 5 ' - C fl R RR RC T TGTCC J+'q&G.& GT&ETGJC~~ CRR g~f!f+RFlG TG T G G fl GR fl G fi fl -3 Leu Ile Lys Lys 5 1 Ile Asp Gly Gln Lys Lys *93 +94 +95 +96 t97 +98 tw ti01 +im I
cm3-103
C
Fig. 2. Summary ojtlie specific binding of the oligonucleotides BGII and CLI to M13-IL5G, and CD93 - 103 to M13-IL5CI. The oligonucleotide CL1 (A) was used to insert a ClaI site and a KEXII protease site at the beginning of the mature hIL-5 coding region; BGII (B) was used to insert a BgnI site juxtaposed 3' to the termination codon of the hIL-5. The oligonucleotide CD93 - 103 (C) was used to eliminate the potential yeast transcription termination site by replacing the codons 93 - 103 with 3-phosphoglycerate kinase preferred codons
*
PGK Promoter 5' 3'.
I
Pre -a Factor
I
hIL-5
~ ~ ~ V bm ----------------
Met&rPkh&U& ~
TTRTTCGCRGCRTCCTCCGCRTTRGCTGCTCCRGTCGRRRTTCCCRCfl RATRRGCGTCGTRGGRGGCGTRRTCGACGAGGTCRGCTTTRRGGGTGT
TTTTTRCRRCG RARARTGTTGC
Bum H1
Linker
~
~
..3' . . .5 '
Pst 1
PGK Terminator
PGK Promoter
Fig. 3. Summary o j t h e yeast expression vector YEpIL-5. The construct contained the 3-phosphoglycerate kinase (PGK) promoter, an in-frame fusion of the pre a-factor to the mature coding region of hIL-5, the PGK terminator, the yeast LEU2 chromosomal gene, a section of pBR322, and a section of the yeast 2-wm episome. The linker was created by annealing the oligonucleotides T-GATCCAAAAAAAATGTCTTTCCCATCTATCTTCACTGCA-3' (5'-phosphorylated) and 5'-GTGAAGATAGATGGGAAAGACATTTTTTTTG-3'. The asterisk (*) indicates where the second amino acid of the pre-x-factor was changed from Arg to Ser so as to resemble more conventional signal sequences
water fractions (Fig. 4). The positive fractions 18 - 34 were pooled and used for further purification. Gel filtration. The pooled fractions from the PhenylSepharose column were concentrated to 2.5 ml using Centricons (Amicon) and then loaded onto the Sephadex G-100 column. Two peaks of hIL-5 activity eluted from the column. The major peak corresponded to a size of about 30 kDa (frac-
tion numbers 52 - 63) and the minor peak corresponded to about 60 kDa (Fig. 5). The z 30-kDa positive fractions were pooled and used for further purification. Anion exchange. The positive fraction pool from the gel filtration column was loaded onto the Mono Q column. The rhIL-5 did not bind to the column but this step removed other contaminating proteins.
M
627
.
Glycol %
0
10
20
30
40
50
60
70
80
z
Fraction Number
Fig. 4. Phenyl-Sepharose chromatography of rhZL-5. The column (1.5 x 20 cm) of Phenyl-Sepharose C L 4 B (Pharmacia) was equilibrated in 1 M ammonium sulphate/100 mM sodium phosphate pH'7.3. The Sfp supernatant (54 h after infection with AcIL-5) was adjusted to the above solution before being loaded onto the column at a rate of 0.5 ml/min. The column was washed with 50 ml equilibration solution before a linear gradient of 1-0 M ammonium sulphate was applied at a rate of 0.5ml/min over 200 min. Fractions of 4 m l were collected. Subsequently, 50 ml phosphate pH 7.3 and 50 ml double deionized and distilled water (ddHzO) were applied before a linear gradient of 030% ehtylene glycol in ddHzO was applied. The column was controlled by an FPLC controller. Fractions were assayed on BCL, cells according __ to Duttdn et ai.-[24]
-.I_
,
, 2000 6 7
4 3 25 ' 3
................
--"--
--b
20000
66 42 I
Q
0.1 0 0
kDa
0
m
a-
1
31
21
14
I
I
I
4
3 2 1
9000
1
Fraction Number Fig. 5. Gelfiltration chromatography rhIL-5. A Sephadex G-100 superfine (Pharmacia) column (1.2 x 60 cm) was equilibrated in 20 mM TrisjHCl pH 8.0. The sample (positive fractions from the PhenylSepharose column concentrated to 2 ml) was applied at a flow rate of 0.05 ml/min, and I-ml fractions were collected. Fractionation was performed at 4°C. Fractions were assayed on BCLl cells according to Dutton et al. [24]. Numbers at the top indicate the molecular mass (in kDa) of marker proteins Gel Slice Number
Physico-chemical characterization of rhIL-5 The purity of the rhIL-5 was checked on a 12.5% SDS/ PAGE gel under reducing and non-reducing conditions. This showed that the rhIL-5 was 95% pure after passage through the Mono Q column (Fig. 6), and consisted of a homodimer linked together by disulfide bridging, as is the case with natural mIL-5 [33]. The predicted protein sequence of human IL-5 contains two cysteine residues, as does mIL-5, which could participate in dimer formation. The rhIL-5 appeared as two bands of similar intensity at 30 and 32 kDa under non-reducing conditions. Under reducing conditions, however, the smaller-sized monomer band (14 kDa) was less intense than the larger monomer band (16 kDa). In gel elution experiments (Fig. 6) the biological activity could only be eluted under nonreducing conditions from the z 30-kDa area and not under reducing conditions, even after the reducing agent was removed by extensive dialysis. The rhIL-5 was shown to have
Fig. 6. SDSjPAGE analysis of purified rhIL-5 and gel elution analysis. Purified recombinant hIL-5 (40 ng) was analysed on a 12.5% SDSj PAGE gel under reducing (lane 3) and non-reducing (lane 1) conditions. Buffer only was run in lane 2. Molecular mass markers were in lane 4 (values at top). The gel was silver-stained. Duplicate samples of reduced (+) and non-reduced (m) hIL-5 were run in other lanes and, after electrophoresis, 2-mm gel slices were taken from these two lanes and placed into 0.5 ml NaCI/P, containing 0.02% Tween-20 and 1 mg/ml bovine serum albumin. The eluted material was extensively dialysed against NaCl/P, before testing in the BCLl assay
an isoelectric point of about 7.2. Deglycosylation of the rhIL5 had no effect on its biological activity. Deglycosylated rhIL5 migrated as an approximately 13-kDa band under reducing conditions (data not shown). The N-terminal sequence of the first 11 residues of the rhIL-5 was Ala-Pro-Val-Glu-Ile-ProThr-Ser-Ala-Leu-Val, as expected, showing that the signal
628 Table 1. Summary of the purfication of rhIL-5 f r o m the baculovirus system Protein quantification for steps 1, 2 and 3 were by Bio-Rad protein assay kit and for step 4 by amino acid analysis Purification step
1. Supernatant 2. PhenylSepharose 3. Gel filtration 4. Anion exchange
Total protein
Total activity
Specific activity
Recovery
mg
U
U/Pg
Yo
32.5
8500
0.26
100
the biological activity of the rhIL-5. The monomer is not biologically active and does not readily form into biologically active dimers after removal of the reducing agent. The purified rhIL-5 has a low specific activity in the BCLl and mEDF assays, and a high specific activity in a human peripheral eosinophil maintenance assay, and a calculated high specific activity in the hCFU assay. The rhIL-5 is biologically inactive below pH 7.0 (which appears to be due to aggregation) and above pH 9.0.
4500 4000 4000
1.13 6.67 15.0
53 47 47
We thank Gary Mayo for synthesizing the oligonucleotides and Bern Presnell for N-terminal sequence analysis and amino acid analysis.
4.0 0.600 0.2667
REFERENCES peptide from the yeast a-factor had been correctly processed. The amino acid analysis of the rhIL-5 showed the predicted composition for all amino acids, except for aspartic acid which was higher than predicted, apparently due to a low-molecularmass contaminant in the buffer used. Based on the amino acid analysis, an accurate specific activity was calculated (see above). A summary of the purification is presented in Table 1. Purified rhIL-5 was stable in the pH range 7.0-9.0, with and without salt (0.15 M NaCl), while below pH 7.0 and above pH 9.0 there was a great reduction in its biological activity (data not shown). A pH of 7.4 would appear to be optimal for maintaining the biological activity. A gel filtration column of rhIL-5 at pH 5.0 showed the rhIL-5 protein eluted in the breakthrough fractions, suggesting that the loss of activity observed at low pH values may be due to aggregation.
Biological u s ~ u y . ~ The purified rhIL-5 showed a specific activity of 1.5 x lo4 U/mg in the BCLl assay, 3.0 x lo5 U/mg in the mEDF assay, and 2.4 x lo7 U/mg in a human peripheral eosinophil
maintenance assay. A specific activity in the human colonyforming units (hCFU) assay of 5 x 10’ U/mg was calculated by comparison of the activity of the rhIL-5 in the mEDF assay with that of a standard rhIL-5 of known activity in both the mEDF and hCFU assays. DISCUSSION We describe here the expression of mature, biologically active rhIL-5 in yeast and baculovirus expression systems. The highest levels of rhIL-5 were obtained from the baculovirus system in serum-containing medium (2.7 mg/l) while ten times less rhIL-5 was produced in serum-free medium. The yeast produced the lowest level of rhIL-5 (12.5 pg/l). An efficient purification procedure was developed for rhIL-5 giving good recoveries of protein with high biological activity. After three purification steps, the rhIL-5 appeared to be pure. Interestingly the insect cells correctly processed the yeast a-factor pre sequence, as N-terminal sequence analysis revealed the expected sequence of mature hIL-5 in the secreted protein. Physico-chemical analysis of the rhIL-5 indicated that it exists as a homodimer linked together by disulphide bonds, the dimer being about 30 kDa in mass, with an isoelectric point of about 7.2. Deglycosylation of the rhIL-5 produced a protein with a molecular mass in agreement with that predicted from the encoded polypeptide sequence of 115 amino acids ( z13 kDa, for the monomer). Deglycosylation did not affect
1. Metcalf, D., Cutler, R. L. & Nicola, N. A. (1983) Blood61,9991005. 2. Clutterbuck, E. J., Shields, J. G., Gordon, J., Smith, S. H., Boyd, A., Callard, R. E., Campbell, H. D., Young, I. G. & Sanderson, C. J. (1987) Eur. J. Immunol. 17, 1743-1750. 3. Swain, S. L. & Dutton, R. W. (1982) J . Exp. Med. 156, 18211834. 4. Ydmaguchi, Y., Suda, T., Suda, J., Eguchi, M., Miura, Y., Harada, N., Tominagd, A. & Takatsu, K . (1988) J . Exp. Med. 167,43- 56. 5. Coffman, R. L., Shrader, B., Carty, J., Mosmann, T. R. & Bond, M. W. (1987) J. Immunol. 139, 3685-3690. 6. Harriman, G. R., Kunimoto, D. Y., Elliot, J. F., Paetkau, V. & Strober, W. (1988) J . Immunol. 140,3033-3039. 7. Loughan, M. S., Takatsu, K., Harada, N. & Nossal, G. J. V. (1987) Proc. Nut1 Acad. Sci. U S A 84, 5399- 5403. 8. Saito, H., Hatake, K.. Dvorak, A. M., Leiferman, K. M., Donnenberg, A. D., Arai, N., Ishizaka, K . &Ishizaka, T. (1988) Proc. Nut1 Acad. Sci. U S A 85, 2288 - 2292. 9. Campbell, H. D., Tucker, W. Q. L., Hort, Y., Martinson, M., Mayo, G., Clutterbuck, E. J., Sanderson, C. J. & Young, I. G. (1 987) Proc. Natl Acad. Sci. U S A 84, 6629 - 6633. 10. Ernst, J. F. (1988) D N A 7, 355-360. 11. Price, V.. Mochizuki, D., March, C. J., Cosman, D., Deeley, M. C., Klinke, R., Clevenger, W., Gillis, S., Baker, P. & Urdal, D. (1987) Gene 55,287-293. 12. Maiorella, B., Inlow, D., Shauger, A. & Hardno, D. (1988) Biotechnology 6, 1406- 1410. 13. Hanahan, D. (1983) J . Mol. Bid. 166, 557-580. 14. Hinnen, A,, Hicks, J. B. & Fink, G. R. (1978) Proc. Natl Acad. Sci. U S A 75, 1929-1933. 15. Isakson, P. C., Krolick, K . A,, Uhr, J. W. & Vitetta, E. S. (1980) J . Immunol. 125, 886-892. 16. Summers, M. D. & Smith, G. E. (1987) A manual of methods f o r haculovirus vectors and insect cell culture procedures, Texas A& M University, College Station TX. 17. Fung, M.-C., Chiu, K. Y. M., Weber, T., Chang, T.-W. & Chang, N. T. (1988) J . Virol. Methods 19, 33-42. 18. Maniatis, T., Fritsh, E. F. & Sambrook, J. (1982) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory, Cold Spring Harbor NY. 19. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Nut1 Acad. Sci. U S A 74, 5463 - 5461. 20. Atkinson, T. C., Smith, M. & Zoller, M. J. (1984) Solid phase phosphite triester synthesis of oligodeoxyribonucleotides. Manual f o r advanced techniques in molecular cloning course, Cold Spring Harbor Laboratory, Cold Spring Harbor NY. 21. Zoller, M. J. & Smith, M. (1983) Methods Enzymol. 100, 468500. 22. Kunkel, T. A. (1985) Proc. Natl Acad. Sci. U S A 82,488-492. 23. Spencer, J. F. T., Spencer, D. M. & Smith, A. R. W. (1980) Y m s t genetics, p. 533, John Wiley, New York. 24. Dutton, R. W., Wetzel, G. D. & Swain, S . L. (1984) J . Immunol. 132.2451 -2456.
25. Campbell, H. D., Sanderson, C. J., Wang, Y., Hort, Y., Martinson, M. E., Tucker, W. Q. J., Stellwagen, A,, Strath, M. & Young, I. G. (1980) Eur. J. Biochem. 174,345-352. 26. Strath, M., Warren, D. J. & Sanderson, C. J. (1985) J. Zmmunol. Methods 83,209 - 215. 27. Owen, W. F., Rothenberg, M. E., Peterson, J., Weller, P. F., Silberstein, D., Sheffer, A. L., Stevens, R. L., Soberman, R. J. & Austen, K. F. (1989) J. Exp. Med. 170, 343-348. 28. Laemmli, U. K . (1970) Nature 227, 680-685.
29. Poehling, H.-M. & Neuhoff, J. (1980) Electrophoresis I , 90102. 30. Morrissey, J. H. (1981) Anal. Biochem. 117, 307-310. 31. Hitzeman, R. A,, Hagie, F. E., Hayflick, J. S., Chen, C. Y., Seeburg, P. H. & Derynck, R. (1982) Nucleic Acids Res. 10, 7791 - 7808. 32. Kurjan, J. & Herskowitz, I. (1982) Cell 30, 933 -943. 33. McKenzie, D. T., Filutowicz, H. I., Swain, S. L. & Dutton, R. W. (1987) J. Immunol. 139,2661 -2668.
