Biocnimica et Biophysica Acta, 1090(1991) 70-80 '~ 1991 ElsevierScience PublishersB.V. 0167-4781/91/$03.50 ADONIS 016747819100197Q
70
BBAEXP 92284
Expression and characterization of recombinant human ciliary neurotrophic factor from Escherichia coli J o h n R . M c D o n a l d , C h r i s t i n e K o , D r z i s l a v M i s m e r *, D a r i n J. S m i t h a n d F r a n k C o l l i n s Synergen, Inc., Boulder, CO (U.S.A.)
(Received22 January 1991) (Revised manuscriptreceived19 April 1991)
Keywords: Recombinanthuman CNTF; Gene expression;Neural cell; Bioactivity;(E.
coil)
The gepe for ciliary neurotrophic factor (CNTF) was doned from a human genomic DNA library by screening with a DNA fragment amplified h e m human genomic DNA using the polymerase chain reaction. A DNA sequence coding for human CNTF was placed under control of an regulatoble oromoter in the expression vector o,IUIO03 and transformed into F_,sg/m'/ck/ac0di strain BI21(DE3). Induction of expression in cultures of this transformant led to the accumulation of approx. 25 m g / I per A m unit of human CNTF. CNTF was purified to homogeneity from cell lysates via anion-exchange, cation-exchnage and Zn 2 +affinity chromatography. Pdrified CNTF contained less than 0.1% contaminating E. co/i proteins, as determined by sodium dodecyl sulfate-polyacrylamide gel e k c t m p h m t s i s (SDS-PAGE), Western blot analysis and reversed-phase high-pressure liquid chromatography (HPLC). The protein exhibited an ultraviolet absorption maximum at 279 nm with a calculated extinction coefficient of A ~ = 9.0. Peptide map and amino acid sequence analyses conflrme~ that the expressed protein has the amino acid sequence expected for human CN3Y, except for the absence of the amino-terminal methionine. Highly-purified recombinant human CNTF supported the survival of chick embtTo parasympathetic, sympathetic and sensory neurons in culture at low pjcomolar concentrations. These results indicate that the biological activities prevJolJsly ascribed to impure CNTF preparations indeed reside in one molecule. Introduction The survival of nerve cells and the expression of their differentiated functions are thought to depend on proteins termed neurotrophic factors [1]. The first described and best characterized of these factors is nerve growth factor (NGF) [2,3]. N G F is highly specific for certain neuronal populations and this has led investigators to seek other neurotrophic factors which may
* Deceased: to whom this manuscr:pt is dedicated. Abbreviations:NGF, nerve growth factor; CBB, Coomassiebrilliant blue; CNBr, cyanogenbromide; CNTF, ciliary neurotrophic factor; IDA, iminodiaceticacid; IPTG, isopropyl/3-t)-thiogalactopyranoside; PAGE, polyacrylamidegel electrophoresis; PEI, polyethylimine; RP-HPLC, reversed-phasehigh performanceliquidchromatography; SDS, sodium dodecylsulfate;TFA, trifluoroaceticacid. The sequence data in this paper have been submitted to the EMBL/Genbank Data Libraries under the accession number X60542 (human CNTF). Correspondence: J.R. McDonaid~~ynergen, Inc., 1885 33rd Street, Boulder, CO 80301, U.S,A.
serve different types of neurons. Several such factors have been cloned, including brain-derived neurotrophic factor (BDNF) [4], neurotrophin 3 (NT-3) [5,6] and ciliary neurotrophic factor (CNTF) [7,8]. NGF, BDNF and NT-3 are members of a family of structurally-related, presumably target-derived, neurotrophic factors which differ in their overall tissue distribution and neuronal specificity [5,6,9]. CNTF differs from this NGF-related family of factors in several ways. CNTF has no sequence homology with members of the N G F family [8,10]. CNTF does not appear to be target-derived in adults, but rather CNTF protein and messenger R N A are concentrated in peripheral nervous tissue itself [7,8,11,12]. In addition, CNTF increases at the site of a peripheral nerve injury [13] and promotes the s:trvival in vivo a n d / o r in vitro of all classes of peripheral neurons: motor [14], sensory, sympathetic and parasympathetic [15,16]. These observations suggest a role for CNTF in repair of the peripheral nervous system after injury. To determine whether exogenous CNTF would be beneficial in treating peripheral nerve damage, it is necessary to have available large amounts of pure CNTF for animal testing. To this end, we report the
71 cloning of the human gene for CNTF, expression of the gene in E. coli, and purification of substantial quantities of recombinant human CNTF. The purified recombinant protein has been characterized by amino acid composition, amino acid sequence, reversed-phase HPLC, SDS-PAGE and Western blot analysis. The biological activity of the highly-purified recombinant protein is compared to that previously reported for partially purified preparations of CNTF.
the rabbit CNTF coding sequence [7]. In addition, each open reading frame contained a segment identical to the fragment amplified from human genomic DNA by pCR. Restriction endonuclease fragments from the human genomic DNA clones corresponded to those cbserved on DNA (Southern) blot analysis of human genomic DNA, indicating that the clones were representative of the organization of the CNTF gene in genomic DNA.
Materials and Methods
Preparation of DNA for the ea'pression of CNTF
Cloning the human CNTF gene Fully degenerate oligonucleotides were synthesized corresponding to the amino acid sequence of rabbit CNTF [7]. The sense orientation of each oligonucleotide is given starting with the 5' end together with the corresponding rabbit protein sequence (N denotes A, C, G or T) CNTF-I: T A T / C GTN A A A / G CAT/C C A A / G GC (Tyr-Val-Lys-His-GIn-Gly); CNTF-2: A A T / C A A A / G A A T / C A T I ' / C / A A T I ' / C C/TI" (Asn-Lys-Asn-lle-Asn-Leu); CNTF-3a: A A A / G T I ' A / G TGG GGN T I ' A / G AA; CNTF-3b: A A A / G "ITA/G TGG GGN CTN AA, CNTF-3c: A A A / G CTN TGG GGN T I ' A / G AA; CNTF-3d: A A A / G CTN TGG GGN CTN AA (Lys-Leu-Trp-Gly-Leu-Lys). Oligonucleotides CNTF-3a to d were used in separate PCR reactions to reduce degeneracy. Oligonucleotides I (sense) and 3a to d (anti-sense) were used as primers in PCR with human genomic DNA. Polymerase chain reactions were performed as previously described [17] except that each reaction contained 1.75 mM MgCI2, 100 ng of each oligonucleotide and 0.5 ttg of human genomic DNA prepared from placenta [18]. To identify DNA bands amplified from the CNTF gvne, DNA (Southern) blots of the PCR products were probed with 32p-labeled oligonucleotide 2, which occurs just downstream of oligonucleotide 1 in the rabbit gene [7]. A single approx. 400 bp band of amplified DNA hybridized to this probe in Southern blots of the PCR products from human genomic DNA. This band was most intense in the reaction using CNTF-3d. This band was cloned and sequenced to give the DNA sequence of the human CNTF gene between oligonucleotides 1 and 3 in Fig. 1. The approx. 400 bp fragment amplified from human genomic DNA was labeled with 32p by random priming and used to screen a human genomic DNA library at high stringency. The human genomic DNA library was constructed by cloning genomic DNA [18], partially digested with Sau3Al, into the BamH! site of Charon 30 [19]. Out of 1 • 106 clones, nine positive clones were isolated. Two of these clones were sequenced and the rest appeared related to these based on DNA (southern) blot analysis. The sequenced clones contained an open reading frame (Fig. 1) that was 89% identical to
A human genomic DNA clone for CNTF in phage Charon 30 was digested with the restriction enz%nrnes Sail and Hindlll and a 4.3 kb fragment was subcloned into the Blnescript KS MI3(-) phagemid vector (Stratagene). This fragment contains the CNTF coding sequences upstream of the Hindlil site in the coding sequence (Fig. 1). This 4.3 kb fragment also contains a single, approx. 1.3 kb intron (Fig. 1). To allow expression in bacterial cells, the intron was removed by site-directed mutagenesis in vitro [20] using the synthetic oligonucleotide 5 ' - G A T G T F C q ' T G T I ' C A ~ CCI'GATC,-CTI'CACATAG G A T T C C G T A A G A ~ GTCAGGTCTGAACGAATCTTCC-Y to produce phagemid 1. The 5' end of the CNTF coding sequence in phagemid 1 was reconstructed for cloning into the expression vector and to make changes found to increase the efficiency of expression in E. co//. Phagemid 1 contains a single Nhel site at amino acids 22-23 in the human CNTF coding sequence (Fig. I). Partially overlapping complementary oligonucleotides, (5'GATCCGATCTI'GGAGGATGATI'AAAT~TCACTGAACA~CTCCGCTGAC(~CCJCACCGTCGAGATCTGTGCAGCCGCTCTATCWGG-3'/5'CTAGCCAGATAGAGCGGC'TGCACAGATCTCGACGGTGC~TCAGCGGAGAGTGTI~AGTGAAAGCCA TI'TAATCATCCTCCAAGATCG-3') containing a 3' Nhei overhang were synthesized, annealed together and ligated to Nhel-cut phagemid 1 to produce phagemid 2. These oligonucleotides alter the human co(Ion usage to that used preferentially by E. coil [21] without changing the amino acid sequence and contai,i a 5' BamHl overhang that creates a BamHi site in phagemid 2. Oligonucleotides 2 and 3 also contain a translational coupler to promote effective translation in E. co/i [22]. Phagemid 2 DNA was then digested with BamHi and Hindlll to release the DNA fragment referred to as CNI'F-Synl which contains DNA s..~luences suitable for expression in E. co/," and encoding human CNTF upstream of the Hindlli site (Fig. 1). To prepare the 3' end of the expression construct, a human genomic DNA clone for CNTF in phage Charon 30 was cut with the restriction enzyme H/ndlll and a 2.1 kb fragment, containing the CNTF coding se-
72 quences downstream of the H/ndlll site (Fig. 1), was suheloned into Hindlll-cut plasmid pEMBL8 [23). A Spel site was inserted into the 2.1 kb insert DNA by oligonucleotide directed mutagenesis 13 base pairs downstream of the stop codon ending the CNTF sequence using the synthetic oligonucleotide 4 (5'-ATG TAG CAG T r A GTC ACT AGT CTC T I C CTF GC'T-3'). The mutated plasmid was cut with H/ndlll and Spel to release the DNA referred to as CNTFSyn2. CNTF-Synl and CNTF-Syn2 were ligated at the H/ndlll overhangs to produce CNTF-Synl/2, which was subcloned into the BomHl- and Spel-cut phagemid expression vector pJUI003 [24] to produce pJU1003huCNTF, which was transformed into E. co/i strain BL21(DE3) [25]. This places expression of the CNTF insert under control of the T7 phage promoter upon induction with isopropyl ~-D-thio-galactopyranoside (IPTG) [24]. One transformant' CNTF-A, producing CNTF after induction with IPTG was selected.
Expression of recombinant human CNTF Overnight cultures of CNTF-A were prepared in Luria broth [26] supplemented with 10 # g / m i of tetracycline. These cultures were diluted (1 to 50) with the same medium and grown until the Aea0 reached 1.0 (3-4 h). Expression of CNTF was achieved by adding IPTG to a final concentration of 0.5 mM and incubating for 4 h. Cells were harvested by centrifngation (9000 × g, 5 min), washed with 50 mM sodium phosphate (pH 8.0) and recentrifuged. Cell pastes were either used immediately or stored frozen at -80°C.
Purification of recombinant human CNTF All purification steps were carried out on ice o,, at 4°C and fractions from the various chromatography columns were analyzed by SDS-PAGE. Step 1: Preparation of cell free extracts. A cell paste (4-5 g wet weight) was suspended in 3-4 volumes of buffer A (50 mM sodium phosphate (pH 8.0), containing 5 mM EGTA and 5 mM EDTA) ano passaged through a French pressure cell at 18000 lb/in 2. The resultant mixture was centrifuged at 48000 × g for 20 rain and the supernatant filtered through glass wool. Step 2" Removal of nucleic acids. PEI was added to the supematant to a final concentration of 0.25% (v/v) to facilitate removal of nucleic acids [27]. Without this treatment, the nucleic acid contained in the supernatant would bind to the anion-e~change resin and decrease the number of times that the Q-Sepharose could be regenerated and reused. After incubating for 10 rain, the mixture was centrifuged as above and the resultant supematant filtered through glass wool.
Step J: f l - ~
ion~
c~,,mn,gra~y.
Cell extract was loaded onto a column (1.5 × 20 cm) of Q-Sei~mose previously equih~rated with buffer A.
After loading, the column was washed with buffer A until the A2g:. reached baseline. CNTF was detected in the column flow-through/wash (see Results). The CNTF pool was dialyzed twice against 10 volumes of buffer B (5 mM sodium phosphate (pH 8.0), containing 10 mM NaCI, 1 mM EGTA and 1 mM EDTA).
Step 4: Q-Sepharose ion-exchange chromatography. The above CNTF pool was loaded onto a column (1.5 × 15 cm) of Q-Sepharose previously equil~rated with buffer 13. After loading, the column was washed with buffer B until the Azso reached baseline. Bound proteins were eluted with a gradient (150 ml) of 10 to 80 raM NaCI in buffer B. The CNTF pool was dialyzed twice against 10 volumes of buffer C (5 mM sodium phosphate (pH 7.1), containing 0.1 mM EGTA and 0.1 mM EDTA).
Step 5: S-Sepharose ion-exchange chromatography. Fhe above CNTF pool was loaded onto a column (1 × 10 cm) of S-Sepharose previously equilibrated with buffer C. After loading, the column was washed with buffer C until the A2s0 reached baseline. Bound proteins were eluted with a gradient (60 ml) of 0 to 0.5 M NaG m buffer C. The CNTF pool was dialyzed twice against 10 volumes of buffer D (10 mM Hepes (pH 7.5), containing 50 mM NaCI, 0.1 mM EGTA and 0.1 mM EDTA).
Step 6: ZnZ+-affinity chromatography. The pool was loaded onto a column (1 × 10 cm) of Zn 2+IDA agarose previously equilibrated with buffer D without the metal ion chelators, EGTA and EDTA. After loading, the column was washed with the same buffer until the A2s0 reached baseline. Bound proteins were eluted with a gradient (50 ml) of 0 to 50 mM histidine in buffer D (without chelators). The final, purified CNTF pool was dialyzed twice against 10 volumes of 10 mM phosphate (pH 8.0), containing 50 mM NaCI, 0.1 mM EGTA and 0.1 mM EDTA and stored at - 8ff'C.
RP-HPLC TFA and acetonitrile were added to protein samples to final concentrations of 0.1% (v/v) and 5% (v/v), respectively, prior to injection. RP-HPLC was performed using a 250 × 4.6 mm SynChropak RP-8 column (SynChrom, Lafayette, IN) with 0.1% aqueous TFA as solvent A and 0.1% TFA in acetonitrile as solvent B.
Electmphore.~ and blotting techniques Electrophoresis was performed in 12.5% polyacrylamide slab gels (1.5 nun thick), with a 5% acrylamide stacking gel, in the presence of 0.1% (w/v) SDS at 40 mA, with the discontinuous buffer system of Laemmli [28]. Gels were stained with CBB as described previously [29] or silver-stained using a Rapid-,M-Stain ~ t (ICN Radiochemieals, It'vine, CA). Gels to be used to
73 separate proteins prior to Western blotting and protein sequencing were pre-electrophoresed for 16 h at 15 mA in the presence of 25 mM thioglycolic acid and 10 mM DTT. This prevents blockage of amino-terminal amino acid groups during electrophoresis of protein samples [30]. Western blotting was performed as previously described [31] using Immobilon-P (Millipore, Bedford, MA) or nitrocellulose (Schleier and Schuell, Keene, NH) membranes. Immobilon-P membranes were stained with CBB [30] and the appropriate protein bands excised for sequencing. Nitrocellulose membranes were subjected to treatment with antibodies to CIqTF and, subsequently, with goat anti- (rabbit IgG) conjugated to alkaline phosphatase (CappeD. The secondary antibody was detected using a kit with 5-bromo4-chioro-indol-3-yl phosphate and nitroblue tetrazolium supplied by Promega (Madison, Wl).
Prepalation of antibodies to CNTF Highly-purified recombinant human CNTF in 10 mM sodium phosphate (pH 8.0), containing 50 mM lqaCl, 0.1 mM EGTA and 0.1 mM EDTA was emulsified with 2 volumes of Freund's complete adjuvant and injected subcutaneously into multiple dorsal sites of two New Zealand rabbits (100 ttg of CNTF per rabbit). Booster injections of 100 /tg of CNTF emulsified in Freund's incomplete adjuvant were then given at 2 to 3-week intervals. Sera were prepared from blood collected 7 weeks after the initial injection and again at 2-week intervals and stored at -70"C. The titer of the antiserum was 3000 as determined by ELISA [32]. B/oassays Bioassays for CNTF activity were performed as described by Lin et al. [10]. Briefly, the in vitro assay for CIqTF activity [33] measures the survival of chick embryo ciliary ganglion (E8), sympathetic chain (Ell) or dorsal root ganglion (El0) neurons. 2000 purified neurons were placed into each well of a 96-well dish and serial dilutions of samples to be assayed were added. After 20 h (ciliary ganglion neurons) or 44 h (sympathetic chain and dorsal root ganglion neurons), neuronal survival was estimated by the ability of live cells to reduce the vital dye MTT (3-4[,5-dimethylthiazol-2yl]-2,5-diphenyltetrazolium) (Sigma). The titer of bioactivity in trophic units (TU) per mi was defir.ed as the dilution that gave 50% of the maximal neuronal survival in the MTT assay. For example, if a dilution of 1 : 1000 was required to give 50% survival, the titer was defined as 1000 TU/ml. Peptide mapp/ng and prote/n sequenc/ng Generation of C-terminal peptides of CNTT: was achieved by first digesting the protein with CNBr overnight at room temperature in hexafluoroacetone hydrate. Peptides were separated on a narrow-bore C8
RP-HPLC column (Brownlee, Santa Clara, CA) with 0.085% aqueous TFA as solvent A and 0.085% TFA in 80% acetonitrile as solvent B. The C-terminal peptide was then subdigested with endoproteinase ASP-N [7] and the peptides separated as above. Amino acid analysis and protein sequencing were carried out as described by Armes and Forney [34].
Other methods Protein concentrations were determined by the BioRad microassay based on CBB dye-binding [35] or by amino-acid analysis. Ultraviolet-absorption spectra were recorded in a Beckman D-50 spectrophotometer and densitometu performed using an LKB UItrascan X L laser densitometer.
Materials S- and Q-Sepharose were purchased from Pharmacia; IDA-agarose from Pierce Chemical and IPTG, PEI, M, markers and tetracycline from Sigma. Results and Discuss~m
Human CNTF gene The genomic DNA sequence and inferred amino acid sequence encoding human CNTF are shown in Fig. 1. The human DNA and protein sequences are 89% and 86% identical to the rabbit CNTF [7] and 85% and 83% identical to the rat CNTF [8] DNA and protein sequences, respectively. Alignment of the inferred amino acid ~ q u ~ of human, rabbit and rat CNTF is shown in Fig. 2. Only a single band that hybridized to CNTF-specific probes was observed in DNA (Southern) blots of human genomic DNA digested with various restriction endonucleases (not shown), consistent with only a single gene in human genomic DNA hybridizing at high stringency.
Purification of recombinant CNTF Upon induction with IPTG, cultures of the bacterial transformant pJUI003-CN-FF-A synthesized recombin=~t human CNTF. At the end of the culture peril., C N I F accounted for approx. 13% of the soluble protein in cell extracts (25 rag/! per A~0 unit) as judged by laser densitometer analysis of CBB-stained gels (Fig. 3, lane 1). When the soluble material from crude cell extract was subjected to anion-exchange chromatography on O-Sepharose in a relatively high ionic strength buffer at pH 8.0, CNTF was slightly retarded on the column and emerged in the flow-through and column wash (F'tg. 3, lanes 6-22) just after the passage of cellular debris and other proteins (Fig. 3, lanes 3-5). The bulk of E. co/i proteins ,~ere retained on the column under these conditions. When the CNTF pool from the first column was dialyzed into a low ionic strength buffer at
74 (;~(;.~.~ACAT~T~TTATT~AC~A~TATA6A~AGAAGTAAA~C~A~T~A~TTGTTTC~T~G~A~A~TT~AGTTAAG~C AT(; GeT TTC ACA GAG CAT TCA CCG CTG Ace CCT CAC CGT eGG GAC CTC T~T AGC CGC TCT }let Ala Phe Thr Glu His Set Pro Leu Thr Pro His Arg Arg Asp Leu Cys Set Arg Set
20
Nhel ollgo .I AIC T(;G eTA GCA AGe AAG ATT CGT TCA GAC CTG ACT GeT CTT ACe CAA TCC TAT GTG AAG 11o Trp Leu A l a Arg Lys I l e Arg Ser Asp Leu Thr A l a Leu T h r Clu S e r T y r Val Lys
40
o l i g o #2 (;AT C ~ C CTG AAC AAG AAC ~TC AAC CTC GAC TCT GCG OAT COG ATe CCA GTG CCA AGe Hi~ (;it* ely Leu Asn Lys Asn Ile Asn Leu Asp Set Ala Asp ely Met Pro Val Ala Set
60
Hindlll ACT CAT CAC TGG ACT GAG CTG ACC GAG GCA GAG CGA CTC CAA GAG AAC CTT CAA GeT TAT Thr Asp Gin Trp Set Clu Leu Thr Clu Ala Clu Arg Leu Cln Clu Ash Leu Cln Ala Tyr
80
CCT Ace TTC CAT GTT TTG TTG CCC AGe CTC TTA CAA GAC CAC CAC GTG CAT TTT Ace CCA Arg Thr Phe His Val Leu Leu A l a A r g Leu Leu C l u Asp Cln C l n Val H i s Phe T h r Pro
i00
AC(; t;;~ GGT GAC TTC CAT CAA GeT ATA CAT ACC CTT CTT CTC CAA GTC GeT GCC TTT CCA "[Lr (;lu ely Asp Phe His Gin Ala lie His Thr Leu Leu Leu Gin Val Ala Ala Phe AIa
120
TAC CAG ATA GAG GAG TTA ATG ATA CTC CTC GAA TAC AAG ATC CCC CGC AAT GAG GeT CAT Tvr Gln lle Clu Clu Leu Met lle Leu Leu elu Tyr Lys lie Pro Arg Asn Glu Ala Asp
140
o l i g o #3 G(;C AqG CCT ATT AAT GTT GGA GAT GeT GeT CTC TTT GAG AAG _AAG CTG TGG GGC CTA AAG ely :-!et Pro lle Asn Val C-ly Asp ely ely Leu P|,e Clu Lys Lys Leu Trp ely Leu Lys
160
L'I'C CTC CAC GAG CTT TeA CAG TGG ACA eTA AGe TCC ATe CAT GAG CTT CGT TTC ATT TCT Val Leu Gin Glu Leu Ser Gin Trp Thr Val Arg Ser n e Hls Asp Leu Arg Phe lie Set
180
TCT CAT CAC ACT eGG ATC CCA CCA CGT eGG AGe CAT TAT ATT GCT AAC AAC AAG AAA ATG Ser ttis Gin Thr e l y I l e Pro A l a A r g e l y S e r H i s T y r I l e A l a Asn Asn Lys Lys Met
200
TA(;CACTTAGTCCCTTCTCTCTTCCTTGCTTTCTCTTCTAATGGAAT
Fig. I. The DNA and inferred amino acid sequence of human CNTF. Tile human CNTF coding sequence is interrupted by a single approx. 1.3 kb intron located between amino acids 38 and 39. The splice accepter/donor sequences at this site are: [ G T ~ A G T . . . 1.3 K b . . . T T T C C T G T A T C CTCGGCCAG]. These sequences fulfill the splice consensus requirements reported by Mount, 1982. The inter~ d Hindlll and Nhel sites used in construction of the expression vector are underlined as are the oligonucleotides used for cloning.
pH 8.0, CNTF now bound to a second column of Q-Scpharose and could be eluted by application of a salt gradient as a peak at 55-60 mM NaCI (Fig. 4). The resultant CNTF pool (Fig. 4, inset) was dialyzed into a low ionic strength buffer at pH 7.1 and subjected to cation-exchange chromatography on S-Sepharose. CNTF bound to the resin and was eluted by application of a salt gradient as a peak between 125-250 mM NaCI (Fig. 5). The CNTF pool (Fig. 5, inset) was then subjected to a final affinity chromatography step on a Zn2*-lDA-agarose column. CNTF bound to the column, probably ";a an interaction between zinc and histidi.~ residuc.~, of which CN'IF: possesses ten per molecule ¢Table !!). CNTF was ctutcd from the column by application of a hi,:tidinc gr~t:lient at 30-35 mM histidine (Fig. 6). A summary of the purification of recombinant hu-
man CNTF is show, ht Table I. The average yield of C N T F was 19 + 1.5% (n = 4). The percent C N T F in protein pools collected after each chromatography step wa:~ determined by laser densitometric analysis of e B B stained gels. These percentages were used to calculate the total amount of CNTF in any given pool and, from this, the fold purification and yields (Table !). Silverstained gels were not used for this purpose, since the intensity of staining with silver was not proportional to the amount of protein and this method, therefore, was not reliably quantitative. However, silver-stained gels were used to qualitatively as~ss the degree of purity of CNTF contained in the various pools (insets, Figs. 4, 5, and 6). The pl of human CNTF, calculated from the amino acid composition, is 6.4. This is significantly higher than that calculated for rabbit ( p / = 5.8) or rat ( p i =
75 MArTBHSPLTPHRRDLCSRSIWLARKIRSDLT~~ NAFAEQTPLTLHRRDLCSRSINLARKIRIDLT~~
s°
Rab
HArMIHSALTPRRnELCSnTIULAIKIllnLTAIT~~
Hum Rat Rab
DSADGMPVASTDQNSELTmABKLOBNLOAYRTrHVL~AR#~~ DSVDGVPVASTDRWS|MTBASRLQ|NLQAT/TVQG~T~ DSVDGVPMASTDQWS|LTEABRL~|NLOAT/TVHIMLAR~~
~°°
Hum Rat aab
TBGDFH.AIHTLLL.VAAFAY.IEILMI~LIYKIPRNE~M~INV~DG~ Is° TIGDFH~/LIBTLML.VSAIrAY.Lm~LMVLI~QKX~ENEqM~ATV~D 4 AEGDFHQAIHTLLL.VAAF~I~RLIIVL~CNX~PK~~V-I~
Hum Rat
Hum
loo
Rat Rab Fig. 2. The inferred amino acid sequences of human, rabbit and rat CNTF. The amino acid sequences are presented in single letter code. Numbers to the right indicate position in the human sequence. Regions in which the sequence is identical in all three species are shaded. The human sequence is presented here for the first time: the rabbit sequence is from Lin el al.. [7]: the rat sequence is from St6ckli el al+. [8]. since the inferred rabbit protein is one amino acid shorter than either human or rat CNTF. a gap (indicated by a dash) has been introduce into the rabbit sequence to maximize alignment.
5.7) CNTF [7,8]. This difference in calculated p / s u g gests that the above purification protocol, which relies heavily on ion exchange chromatography at carefully
controlled pH and ionic strengths, might require modification in order to be used to purify recombinant rat or rabbit CNTFs from bacterial expression systems.
66 ,,.I.~
emm,
m
m
m
omm
m
i,.k~.
45 36 20 24
.
.
.
.
.
.
.
.
JL i ~ "
20.1
14.2
1
5
10
15
20
25
Fit;. 3. SDS-PAGE analysis of ~ l e c t c d fractions elated from i.itial O-Sepharose ion-exchange chromatography column (step 3). Cell extract was chromatograpbed on a column of O-Sepharose (I.5 x 20 cm) as de~ribed under Materials and M*.thods. The chromalogram ~as developed at 2 m l / m i n and 2 ml fractions were collected. Selected fractions were subjected to SDS-PAGE and the gels stained with CBB. For ele~'lrophoresis. samples (15 pl. lanes I and 2 or 39 #1. lanes 3-28)were diluted in SDS-sample buffer (final concentrations: 10rl" glycerol, le~ D'FF. 0.Y;- SDS. 0.002% bromoph,:nol blve ~md 25 mM Tris-HCI ~.pH 6.8) and boiled for 2 rain. Key to gel lanes: I and 2. crude extract before and after PEI ,leatn~t,.,*,L ~'csFcciively: 3-28, even-numbered fractions from 12 to 62. Numbers in the left-hand margin indicate M r values ( × I0-~) of protein standards electrophoresed stmultal,eously. CNTF co-migrates with the trypshlogen standard (M, = 24(kX)) and ,he CNTF pool represents fractions 18-50.
76 0.25 - -
0.4 I
.--
7 BO
...... - ~
0.3 4 60
:~
.
'
/
40 ~" 30
70
BBAEXP 92284
Expression and characterization of recombinant human ciliary neurotrophic factor from Escherichia coli J o h n R . M c D o n a l d , C h r i s t i n e K o , D r z i s l a v M i s m e r *, D a r i n J. S m i t h a n d F r a n k C o l l i n s Synergen, Inc., Boulder, CO (U.S.A.)
(Received22 January 1991) (Revised manuscriptreceived19 April 1991)
Keywords: Recombinanthuman CNTF; Gene expression;Neural cell; Bioactivity;(E.
coil)
The gepe for ciliary neurotrophic factor (CNTF) was doned from a human genomic DNA library by screening with a DNA fragment amplified h e m human genomic DNA using the polymerase chain reaction. A DNA sequence coding for human CNTF was placed under control of an regulatoble oromoter in the expression vector o,IUIO03 and transformed into F_,sg/m'/ck/ac0di strain BI21(DE3). Induction of expression in cultures of this transformant led to the accumulation of approx. 25 m g / I per A m unit of human CNTF. CNTF was purified to homogeneity from cell lysates via anion-exchange, cation-exchnage and Zn 2 +affinity chromatography. Pdrified CNTF contained less than 0.1% contaminating E. co/i proteins, as determined by sodium dodecyl sulfate-polyacrylamide gel e k c t m p h m t s i s (SDS-PAGE), Western blot analysis and reversed-phase high-pressure liquid chromatography (HPLC). The protein exhibited an ultraviolet absorption maximum at 279 nm with a calculated extinction coefficient of A ~ = 9.0. Peptide map and amino acid sequence analyses conflrme~ that the expressed protein has the amino acid sequence expected for human CN3Y, except for the absence of the amino-terminal methionine. Highly-purified recombinant human CNTF supported the survival of chick embtTo parasympathetic, sympathetic and sensory neurons in culture at low pjcomolar concentrations. These results indicate that the biological activities prevJolJsly ascribed to impure CNTF preparations indeed reside in one molecule. Introduction The survival of nerve cells and the expression of their differentiated functions are thought to depend on proteins termed neurotrophic factors [1]. The first described and best characterized of these factors is nerve growth factor (NGF) [2,3]. N G F is highly specific for certain neuronal populations and this has led investigators to seek other neurotrophic factors which may
* Deceased: to whom this manuscr:pt is dedicated. Abbreviations:NGF, nerve growth factor; CBB, Coomassiebrilliant blue; CNBr, cyanogenbromide; CNTF, ciliary neurotrophic factor; IDA, iminodiaceticacid; IPTG, isopropyl/3-t)-thiogalactopyranoside; PAGE, polyacrylamidegel electrophoresis; PEI, polyethylimine; RP-HPLC, reversed-phasehigh performanceliquidchromatography; SDS, sodium dodecylsulfate;TFA, trifluoroaceticacid. The sequence data in this paper have been submitted to the EMBL/Genbank Data Libraries under the accession number X60542 (human CNTF). Correspondence: J.R. McDonaid~~ynergen, Inc., 1885 33rd Street, Boulder, CO 80301, U.S,A.
serve different types of neurons. Several such factors have been cloned, including brain-derived neurotrophic factor (BDNF) [4], neurotrophin 3 (NT-3) [5,6] and ciliary neurotrophic factor (CNTF) [7,8]. NGF, BDNF and NT-3 are members of a family of structurally-related, presumably target-derived, neurotrophic factors which differ in their overall tissue distribution and neuronal specificity [5,6,9]. CNTF differs from this NGF-related family of factors in several ways. CNTF has no sequence homology with members of the N G F family [8,10]. CNTF does not appear to be target-derived in adults, but rather CNTF protein and messenger R N A are concentrated in peripheral nervous tissue itself [7,8,11,12]. In addition, CNTF increases at the site of a peripheral nerve injury [13] and promotes the s:trvival in vivo a n d / o r in vitro of all classes of peripheral neurons: motor [14], sensory, sympathetic and parasympathetic [15,16]. These observations suggest a role for CNTF in repair of the peripheral nervous system after injury. To determine whether exogenous CNTF would be beneficial in treating peripheral nerve damage, it is necessary to have available large amounts of pure CNTF for animal testing. To this end, we report the
71 cloning of the human gene for CNTF, expression of the gene in E. coli, and purification of substantial quantities of recombinant human CNTF. The purified recombinant protein has been characterized by amino acid composition, amino acid sequence, reversed-phase HPLC, SDS-PAGE and Western blot analysis. The biological activity of the highly-purified recombinant protein is compared to that previously reported for partially purified preparations of CNTF.
the rabbit CNTF coding sequence [7]. In addition, each open reading frame contained a segment identical to the fragment amplified from human genomic DNA by pCR. Restriction endonuclease fragments from the human genomic DNA clones corresponded to those cbserved on DNA (Southern) blot analysis of human genomic DNA, indicating that the clones were representative of the organization of the CNTF gene in genomic DNA.
Materials and Methods
Preparation of DNA for the ea'pression of CNTF
Cloning the human CNTF gene Fully degenerate oligonucleotides were synthesized corresponding to the amino acid sequence of rabbit CNTF [7]. The sense orientation of each oligonucleotide is given starting with the 5' end together with the corresponding rabbit protein sequence (N denotes A, C, G or T) CNTF-I: T A T / C GTN A A A / G CAT/C C A A / G GC (Tyr-Val-Lys-His-GIn-Gly); CNTF-2: A A T / C A A A / G A A T / C A T I ' / C / A A T I ' / C C/TI" (Asn-Lys-Asn-lle-Asn-Leu); CNTF-3a: A A A / G T I ' A / G TGG GGN T I ' A / G AA; CNTF-3b: A A A / G "ITA/G TGG GGN CTN AA, CNTF-3c: A A A / G CTN TGG GGN T I ' A / G AA; CNTF-3d: A A A / G CTN TGG GGN CTN AA (Lys-Leu-Trp-Gly-Leu-Lys). Oligonucleotides CNTF-3a to d were used in separate PCR reactions to reduce degeneracy. Oligonucleotides I (sense) and 3a to d (anti-sense) were used as primers in PCR with human genomic DNA. Polymerase chain reactions were performed as previously described [17] except that each reaction contained 1.75 mM MgCI2, 100 ng of each oligonucleotide and 0.5 ttg of human genomic DNA prepared from placenta [18]. To identify DNA bands amplified from the CNTF gvne, DNA (Southern) blots of the PCR products were probed with 32p-labeled oligonucleotide 2, which occurs just downstream of oligonucleotide 1 in the rabbit gene [7]. A single approx. 400 bp band of amplified DNA hybridized to this probe in Southern blots of the PCR products from human genomic DNA. This band was most intense in the reaction using CNTF-3d. This band was cloned and sequenced to give the DNA sequence of the human CNTF gene between oligonucleotides 1 and 3 in Fig. 1. The approx. 400 bp fragment amplified from human genomic DNA was labeled with 32p by random priming and used to screen a human genomic DNA library at high stringency. The human genomic DNA library was constructed by cloning genomic DNA [18], partially digested with Sau3Al, into the BamH! site of Charon 30 [19]. Out of 1 • 106 clones, nine positive clones were isolated. Two of these clones were sequenced and the rest appeared related to these based on DNA (southern) blot analysis. The sequenced clones contained an open reading frame (Fig. 1) that was 89% identical to
A human genomic DNA clone for CNTF in phage Charon 30 was digested with the restriction enz%nrnes Sail and Hindlll and a 4.3 kb fragment was subcloned into the Blnescript KS MI3(-) phagemid vector (Stratagene). This fragment contains the CNTF coding sequences upstream of the Hindlil site in the coding sequence (Fig. 1). This 4.3 kb fragment also contains a single, approx. 1.3 kb intron (Fig. 1). To allow expression in bacterial cells, the intron was removed by site-directed mutagenesis in vitro [20] using the synthetic oligonucleotide 5 ' - G A T G T F C q ' T G T I ' C A ~ CCI'GATC,-CTI'CACATAG G A T T C C G T A A G A ~ GTCAGGTCTGAACGAATCTTCC-Y to produce phagemid 1. The 5' end of the CNTF coding sequence in phagemid 1 was reconstructed for cloning into the expression vector and to make changes found to increase the efficiency of expression in E. co//. Phagemid 1 contains a single Nhel site at amino acids 22-23 in the human CNTF coding sequence (Fig. I). Partially overlapping complementary oligonucleotides, (5'GATCCGATCTI'GGAGGATGATI'AAAT~TCACTGAACA~CTCCGCTGAC(~CCJCACCGTCGAGATCTGTGCAGCCGCTCTATCWGG-3'/5'CTAGCCAGATAGAGCGGC'TGCACAGATCTCGACGGTGC~TCAGCGGAGAGTGTI~AGTGAAAGCCA TI'TAATCATCCTCCAAGATCG-3') containing a 3' Nhei overhang were synthesized, annealed together and ligated to Nhel-cut phagemid 1 to produce phagemid 2. These oligonucleotides alter the human co(Ion usage to that used preferentially by E. coil [21] without changing the amino acid sequence and contai,i a 5' BamHl overhang that creates a BamHi site in phagemid 2. Oligonucleotides 2 and 3 also contain a translational coupler to promote effective translation in E. co/i [22]. Phagemid 2 DNA was then digested with BamHi and Hindlll to release the DNA fragment referred to as CNI'F-Synl which contains DNA s..~luences suitable for expression in E. co/," and encoding human CNTF upstream of the Hindlli site (Fig. 1). To prepare the 3' end of the expression construct, a human genomic DNA clone for CNTF in phage Charon 30 was cut with the restriction enzyme H/ndlll and a 2.1 kb fragment, containing the CNTF coding se-
72 quences downstream of the H/ndlll site (Fig. 1), was suheloned into Hindlll-cut plasmid pEMBL8 [23). A Spel site was inserted into the 2.1 kb insert DNA by oligonucleotide directed mutagenesis 13 base pairs downstream of the stop codon ending the CNTF sequence using the synthetic oligonucleotide 4 (5'-ATG TAG CAG T r A GTC ACT AGT CTC T I C CTF GC'T-3'). The mutated plasmid was cut with H/ndlll and Spel to release the DNA referred to as CNTFSyn2. CNTF-Synl and CNTF-Syn2 were ligated at the H/ndlll overhangs to produce CNTF-Synl/2, which was subcloned into the BomHl- and Spel-cut phagemid expression vector pJUI003 [24] to produce pJU1003huCNTF, which was transformed into E. co/i strain BL21(DE3) [25]. This places expression of the CNTF insert under control of the T7 phage promoter upon induction with isopropyl ~-D-thio-galactopyranoside (IPTG) [24]. One transformant' CNTF-A, producing CNTF after induction with IPTG was selected.
Expression of recombinant human CNTF Overnight cultures of CNTF-A were prepared in Luria broth [26] supplemented with 10 # g / m i of tetracycline. These cultures were diluted (1 to 50) with the same medium and grown until the Aea0 reached 1.0 (3-4 h). Expression of CNTF was achieved by adding IPTG to a final concentration of 0.5 mM and incubating for 4 h. Cells were harvested by centrifngation (9000 × g, 5 min), washed with 50 mM sodium phosphate (pH 8.0) and recentrifuged. Cell pastes were either used immediately or stored frozen at -80°C.
Purification of recombinant human CNTF All purification steps were carried out on ice o,, at 4°C and fractions from the various chromatography columns were analyzed by SDS-PAGE. Step 1: Preparation of cell free extracts. A cell paste (4-5 g wet weight) was suspended in 3-4 volumes of buffer A (50 mM sodium phosphate (pH 8.0), containing 5 mM EGTA and 5 mM EDTA) ano passaged through a French pressure cell at 18000 lb/in 2. The resultant mixture was centrifuged at 48000 × g for 20 rain and the supernatant filtered through glass wool. Step 2" Removal of nucleic acids. PEI was added to the supematant to a final concentration of 0.25% (v/v) to facilitate removal of nucleic acids [27]. Without this treatment, the nucleic acid contained in the supernatant would bind to the anion-e~change resin and decrease the number of times that the Q-Sepharose could be regenerated and reused. After incubating for 10 rain, the mixture was centrifuged as above and the resultant supematant filtered through glass wool.
Step J: f l - ~
ion~
c~,,mn,gra~y.
Cell extract was loaded onto a column (1.5 × 20 cm) of Q-Sei~mose previously equih~rated with buffer A.
After loading, the column was washed with buffer A until the A2g:. reached baseline. CNTF was detected in the column flow-through/wash (see Results). The CNTF pool was dialyzed twice against 10 volumes of buffer B (5 mM sodium phosphate (pH 8.0), containing 10 mM NaCI, 1 mM EGTA and 1 mM EDTA).
Step 4: Q-Sepharose ion-exchange chromatography. The above CNTF pool was loaded onto a column (1.5 × 15 cm) of Q-Sepharose previously equil~rated with buffer 13. After loading, the column was washed with buffer B until the Azso reached baseline. Bound proteins were eluted with a gradient (150 ml) of 10 to 80 raM NaCI in buffer B. The CNTF pool was dialyzed twice against 10 volumes of buffer C (5 mM sodium phosphate (pH 7.1), containing 0.1 mM EGTA and 0.1 mM EDTA).
Step 5: S-Sepharose ion-exchange chromatography. Fhe above CNTF pool was loaded onto a column (1 × 10 cm) of S-Sepharose previously equilibrated with buffer C. After loading, the column was washed with buffer C until the A2s0 reached baseline. Bound proteins were eluted with a gradient (60 ml) of 0 to 0.5 M NaG m buffer C. The CNTF pool was dialyzed twice against 10 volumes of buffer D (10 mM Hepes (pH 7.5), containing 50 mM NaCI, 0.1 mM EGTA and 0.1 mM EDTA).
Step 6: ZnZ+-affinity chromatography. The pool was loaded onto a column (1 × 10 cm) of Zn 2+IDA agarose previously equilibrated with buffer D without the metal ion chelators, EGTA and EDTA. After loading, the column was washed with the same buffer until the A2s0 reached baseline. Bound proteins were eluted with a gradient (50 ml) of 0 to 50 mM histidine in buffer D (without chelators). The final, purified CNTF pool was dialyzed twice against 10 volumes of 10 mM phosphate (pH 8.0), containing 50 mM NaCI, 0.1 mM EGTA and 0.1 mM EDTA and stored at - 8ff'C.
RP-HPLC TFA and acetonitrile were added to protein samples to final concentrations of 0.1% (v/v) and 5% (v/v), respectively, prior to injection. RP-HPLC was performed using a 250 × 4.6 mm SynChropak RP-8 column (SynChrom, Lafayette, IN) with 0.1% aqueous TFA as solvent A and 0.1% TFA in acetonitrile as solvent B.
Electmphore.~ and blotting techniques Electrophoresis was performed in 12.5% polyacrylamide slab gels (1.5 nun thick), with a 5% acrylamide stacking gel, in the presence of 0.1% (w/v) SDS at 40 mA, with the discontinuous buffer system of Laemmli [28]. Gels were stained with CBB as described previously [29] or silver-stained using a Rapid-,M-Stain ~ t (ICN Radiochemieals, It'vine, CA). Gels to be used to
73 separate proteins prior to Western blotting and protein sequencing were pre-electrophoresed for 16 h at 15 mA in the presence of 25 mM thioglycolic acid and 10 mM DTT. This prevents blockage of amino-terminal amino acid groups during electrophoresis of protein samples [30]. Western blotting was performed as previously described [31] using Immobilon-P (Millipore, Bedford, MA) or nitrocellulose (Schleier and Schuell, Keene, NH) membranes. Immobilon-P membranes were stained with CBB [30] and the appropriate protein bands excised for sequencing. Nitrocellulose membranes were subjected to treatment with antibodies to CIqTF and, subsequently, with goat anti- (rabbit IgG) conjugated to alkaline phosphatase (CappeD. The secondary antibody was detected using a kit with 5-bromo4-chioro-indol-3-yl phosphate and nitroblue tetrazolium supplied by Promega (Madison, Wl).
Prepalation of antibodies to CNTF Highly-purified recombinant human CNTF in 10 mM sodium phosphate (pH 8.0), containing 50 mM lqaCl, 0.1 mM EGTA and 0.1 mM EDTA was emulsified with 2 volumes of Freund's complete adjuvant and injected subcutaneously into multiple dorsal sites of two New Zealand rabbits (100 ttg of CNTF per rabbit). Booster injections of 100 /tg of CNTF emulsified in Freund's incomplete adjuvant were then given at 2 to 3-week intervals. Sera were prepared from blood collected 7 weeks after the initial injection and again at 2-week intervals and stored at -70"C. The titer of the antiserum was 3000 as determined by ELISA [32]. B/oassays Bioassays for CNTF activity were performed as described by Lin et al. [10]. Briefly, the in vitro assay for CIqTF activity [33] measures the survival of chick embryo ciliary ganglion (E8), sympathetic chain (Ell) or dorsal root ganglion (El0) neurons. 2000 purified neurons were placed into each well of a 96-well dish and serial dilutions of samples to be assayed were added. After 20 h (ciliary ganglion neurons) or 44 h (sympathetic chain and dorsal root ganglion neurons), neuronal survival was estimated by the ability of live cells to reduce the vital dye MTT (3-4[,5-dimethylthiazol-2yl]-2,5-diphenyltetrazolium) (Sigma). The titer of bioactivity in trophic units (TU) per mi was defir.ed as the dilution that gave 50% of the maximal neuronal survival in the MTT assay. For example, if a dilution of 1 : 1000 was required to give 50% survival, the titer was defined as 1000 TU/ml. Peptide mapp/ng and prote/n sequenc/ng Generation of C-terminal peptides of CNTT: was achieved by first digesting the protein with CNBr overnight at room temperature in hexafluoroacetone hydrate. Peptides were separated on a narrow-bore C8
RP-HPLC column (Brownlee, Santa Clara, CA) with 0.085% aqueous TFA as solvent A and 0.085% TFA in 80% acetonitrile as solvent B. The C-terminal peptide was then subdigested with endoproteinase ASP-N [7] and the peptides separated as above. Amino acid analysis and protein sequencing were carried out as described by Armes and Forney [34].
Other methods Protein concentrations were determined by the BioRad microassay based on CBB dye-binding [35] or by amino-acid analysis. Ultraviolet-absorption spectra were recorded in a Beckman D-50 spectrophotometer and densitometu performed using an LKB UItrascan X L laser densitometer.
Materials S- and Q-Sepharose were purchased from Pharmacia; IDA-agarose from Pierce Chemical and IPTG, PEI, M, markers and tetracycline from Sigma. Results and Discuss~m
Human CNTF gene The genomic DNA sequence and inferred amino acid sequence encoding human CNTF are shown in Fig. 1. The human DNA and protein sequences are 89% and 86% identical to the rabbit CNTF [7] and 85% and 83% identical to the rat CNTF [8] DNA and protein sequences, respectively. Alignment of the inferred amino acid ~ q u ~ of human, rabbit and rat CNTF is shown in Fig. 2. Only a single band that hybridized to CNTF-specific probes was observed in DNA (Southern) blots of human genomic DNA digested with various restriction endonucleases (not shown), consistent with only a single gene in human genomic DNA hybridizing at high stringency.
Purification of recombinant CNTF Upon induction with IPTG, cultures of the bacterial transformant pJUI003-CN-FF-A synthesized recombin=~t human CNTF. At the end of the culture peril., C N I F accounted for approx. 13% of the soluble protein in cell extracts (25 rag/! per A~0 unit) as judged by laser densitometer analysis of CBB-stained gels (Fig. 3, lane 1). When the soluble material from crude cell extract was subjected to anion-exchange chromatography on O-Sepharose in a relatively high ionic strength buffer at pH 8.0, CNTF was slightly retarded on the column and emerged in the flow-through and column wash (F'tg. 3, lanes 6-22) just after the passage of cellular debris and other proteins (Fig. 3, lanes 3-5). The bulk of E. co/i proteins ,~ere retained on the column under these conditions. When the CNTF pool from the first column was dialyzed into a low ionic strength buffer at
74 (;~(;.~.~ACAT~T~TTATT~AC~A~TATA6A~AGAAGTAAA~C~A~T~A~TTGTTTC~T~G~A~A~TT~AGTTAAG~C AT(; GeT TTC ACA GAG CAT TCA CCG CTG Ace CCT CAC CGT eGG GAC CTC T~T AGC CGC TCT }let Ala Phe Thr Glu His Set Pro Leu Thr Pro His Arg Arg Asp Leu Cys Set Arg Set
20
Nhel ollgo .I AIC T(;G eTA GCA AGe AAG ATT CGT TCA GAC CTG ACT GeT CTT ACe CAA TCC TAT GTG AAG 11o Trp Leu A l a Arg Lys I l e Arg Ser Asp Leu Thr A l a Leu T h r Clu S e r T y r Val Lys
40
o l i g o #2 (;AT C ~ C CTG AAC AAG AAC ~TC AAC CTC GAC TCT GCG OAT COG ATe CCA GTG CCA AGe Hi~ (;it* ely Leu Asn Lys Asn Ile Asn Leu Asp Set Ala Asp ely Met Pro Val Ala Set
60
Hindlll ACT CAT CAC TGG ACT GAG CTG ACC GAG GCA GAG CGA CTC CAA GAG AAC CTT CAA GeT TAT Thr Asp Gin Trp Set Clu Leu Thr Clu Ala Clu Arg Leu Cln Clu Ash Leu Cln Ala Tyr
80
CCT Ace TTC CAT GTT TTG TTG CCC AGe CTC TTA CAA GAC CAC CAC GTG CAT TTT Ace CCA Arg Thr Phe His Val Leu Leu A l a A r g Leu Leu C l u Asp Cln C l n Val H i s Phe T h r Pro
i00
AC(; t;;~ GGT GAC TTC CAT CAA GeT ATA CAT ACC CTT CTT CTC CAA GTC GeT GCC TTT CCA "[Lr (;lu ely Asp Phe His Gin Ala lie His Thr Leu Leu Leu Gin Val Ala Ala Phe AIa
120
TAC CAG ATA GAG GAG TTA ATG ATA CTC CTC GAA TAC AAG ATC CCC CGC AAT GAG GeT CAT Tvr Gln lle Clu Clu Leu Met lle Leu Leu elu Tyr Lys lie Pro Arg Asn Glu Ala Asp
140
o l i g o #3 G(;C AqG CCT ATT AAT GTT GGA GAT GeT GeT CTC TTT GAG AAG _AAG CTG TGG GGC CTA AAG ely :-!et Pro lle Asn Val C-ly Asp ely ely Leu P|,e Clu Lys Lys Leu Trp ely Leu Lys
160
L'I'C CTC CAC GAG CTT TeA CAG TGG ACA eTA AGe TCC ATe CAT GAG CTT CGT TTC ATT TCT Val Leu Gin Glu Leu Ser Gin Trp Thr Val Arg Ser n e Hls Asp Leu Arg Phe lie Set
180
TCT CAT CAC ACT eGG ATC CCA CCA CGT eGG AGe CAT TAT ATT GCT AAC AAC AAG AAA ATG Ser ttis Gin Thr e l y I l e Pro A l a A r g e l y S e r H i s T y r I l e A l a Asn Asn Lys Lys Met
200
TA(;CACTTAGTCCCTTCTCTCTTCCTTGCTTTCTCTTCTAATGGAAT
Fig. I. The DNA and inferred amino acid sequence of human CNTF. Tile human CNTF coding sequence is interrupted by a single approx. 1.3 kb intron located between amino acids 38 and 39. The splice accepter/donor sequences at this site are: [ G T ~ A G T . . . 1.3 K b . . . T T T C C T G T A T C CTCGGCCAG]. These sequences fulfill the splice consensus requirements reported by Mount, 1982. The inter~ d Hindlll and Nhel sites used in construction of the expression vector are underlined as are the oligonucleotides used for cloning.
pH 8.0, CNTF now bound to a second column of Q-Scpharose and could be eluted by application of a salt gradient as a peak at 55-60 mM NaCI (Fig. 4). The resultant CNTF pool (Fig. 4, inset) was dialyzed into a low ionic strength buffer at pH 7.1 and subjected to cation-exchange chromatography on S-Sepharose. CNTF bound to the resin and was eluted by application of a salt gradient as a peak between 125-250 mM NaCI (Fig. 5). The CNTF pool (Fig. 5, inset) was then subjected to a final affinity chromatography step on a Zn2*-lDA-agarose column. CNTF bound to the column, probably ";a an interaction between zinc and histidi.~ residuc.~, of which CN'IF: possesses ten per molecule ¢Table !!). CNTF was ctutcd from the column by application of a hi,:tidinc gr~t:lient at 30-35 mM histidine (Fig. 6). A summary of the purification of recombinant hu-
man CNTF is show, ht Table I. The average yield of C N T F was 19 + 1.5% (n = 4). The percent C N T F in protein pools collected after each chromatography step wa:~ determined by laser densitometric analysis of e B B stained gels. These percentages were used to calculate the total amount of CNTF in any given pool and, from this, the fold purification and yields (Table !). Silverstained gels were not used for this purpose, since the intensity of staining with silver was not proportional to the amount of protein and this method, therefore, was not reliably quantitative. However, silver-stained gels were used to qualitatively as~ss the degree of purity of CNTF contained in the various pools (insets, Figs. 4, 5, and 6). The pl of human CNTF, calculated from the amino acid composition, is 6.4. This is significantly higher than that calculated for rabbit ( p / = 5.8) or rat ( p i =
75 MArTBHSPLTPHRRDLCSRSIWLARKIRSDLT~~ NAFAEQTPLTLHRRDLCSRSINLARKIRIDLT~~
s°
Rab
HArMIHSALTPRRnELCSnTIULAIKIllnLTAIT~~
Hum Rat Rab
DSADGMPVASTDQNSELTmABKLOBNLOAYRTrHVL~AR#~~ DSVDGVPVASTDRWS|MTBASRLQ|NLQAT/TVQG~T~ DSVDGVPMASTDQWS|LTEABRL~|NLOAT/TVHIMLAR~~
~°°
Hum Rat aab
TBGDFH.AIHTLLL.VAAFAY.IEILMI~LIYKIPRNE~M~INV~DG~ Is° TIGDFH~/LIBTLML.VSAIrAY.Lm~LMVLI~QKX~ENEqM~ATV~D 4 AEGDFHQAIHTLLL.VAAF~I~RLIIVL~CNX~PK~~V-I~
Hum Rat
Hum
loo
Rat Rab Fig. 2. The inferred amino acid sequences of human, rabbit and rat CNTF. The amino acid sequences are presented in single letter code. Numbers to the right indicate position in the human sequence. Regions in which the sequence is identical in all three species are shaded. The human sequence is presented here for the first time: the rabbit sequence is from Lin el al.. [7]: the rat sequence is from St6ckli el al+. [8]. since the inferred rabbit protein is one amino acid shorter than either human or rat CNTF. a gap (indicated by a dash) has been introduce into the rabbit sequence to maximize alignment.
5.7) CNTF [7,8]. This difference in calculated p / s u g gests that the above purification protocol, which relies heavily on ion exchange chromatography at carefully
controlled pH and ionic strengths, might require modification in order to be used to purify recombinant rat or rabbit CNTFs from bacterial expression systems.
66 ,,.I.~
emm,
m
m
m
omm
m
i,.k~.
45 36 20 24
.
.
.
.
.
.
.
.
JL i ~ "
20.1
14.2
1
5
10
15
20
25
Fit;. 3. SDS-PAGE analysis of ~ l e c t c d fractions elated from i.itial O-Sepharose ion-exchange chromatography column (step 3). Cell extract was chromatograpbed on a column of O-Sepharose (I.5 x 20 cm) as de~ribed under Materials and M*.thods. The chromalogram ~as developed at 2 m l / m i n and 2 ml fractions were collected. Selected fractions were subjected to SDS-PAGE and the gels stained with CBB. For ele~'lrophoresis. samples (15 pl. lanes I and 2 or 39 #1. lanes 3-28)were diluted in SDS-sample buffer (final concentrations: 10rl" glycerol, le~ D'FF. 0.Y;- SDS. 0.002% bromoph,:nol blve ~md 25 mM Tris-HCI ~.pH 6.8) and boiled for 2 rain. Key to gel lanes: I and 2. crude extract before and after PEI ,leatn~t,.,*,L ~'csFcciively: 3-28, even-numbered fractions from 12 to 62. Numbers in the left-hand margin indicate M r values ( × I0-~) of protein standards electrophoresed stmultal,eously. CNTF co-migrates with the trypshlogen standard (M, = 24(kX)) and ,he CNTF pool represents fractions 18-50.
76 0.25 - -
0.4 I
.--
7 BO
...... - ~
0.3 4 60
:~
.
'
/
40 ~" 30
