Journal of Neuroscience Research 29:251-260 (1991)

Rapid Communication Synthesis, Purification, and Characterization of Human Ciliary Neuronotrophic Factor FromE. coli A. Negro, G. Corona, E. Bigon, I. Martini, C. Grandi, S.D. Skaper, and L. Callegaro Advanced Technology Division (A.N., G . C . , E.B., I.M., L.C.) and Fidia Research Laboratories (S.D.S.), Fidia S.p.A., Abano Terme, and Department of Organic Chemistry, University of Padua ( C . G . ) ,Padua, Italy

The cDNA for human ciliary neuronotrophic factor (CNTF) has been cloned into an expression vector under the control of the T7 promoter. The BL21 strain of E. coli was transformed with this vector. Human CNTF accounted for about 30% of the total bacterial protein after induction with isopropy1-BD-thiogalactopyranoside. This human CNTF was purified to homogeneity from inclusion bodies by a combination of ion exchange chromatography and reverse-phase high performance liquid chromatography. The amino-terminal amino acid sequence of the purified protein was identical to the deduced amino acid sequence; however, the methionyl residue has been removed. On SDS-PAGE gels, human CNTF displayed a molecular weight of about 24 kDa, in accord with its deduced molecular mass; a PI of 5.8 indicates the acidic nature of the molecule. A proposed structure for human CNTF includes major alpha helical regions. The ED50 of purified human CNTF was approximately 30 pM, using cultured embryonic day 10 chicken dorsal root ganglion neurons; no activity was observed with neurons from embryonic day 8 ganglia. Polyclonal antibodies prepared against both a synthetic peptide of CNTF and the entire human CNTF protein recognized a single 24 kDa band on Western blots, corresponding to human CNTF. However, only the antibodies against intact CNTF blocked its biological activity. This represents the first molecular expresson and purification of human CNTF. Key words: ciliary neuronotrophic factor, human, recombinant, antibodies, neuropathologies INTRODUCTION Soluble trophic factors act as important determinants of neuron phenotypic development and survival in 0 1991 Wiley-Liss, Inc.

many systems (Barde, 1989). In the peripheral nervous system, evidence for a target tissue-derived trophic signal mediating nerve cell survival during development has been established for nerve growth factor (NGF) (LeviMontalcini, 1987). In the central nervous system, NGF is believed to play a role in the development of certain cholinergic neurons of the basal forebrain (Dreyfus, 1989). Much less is known about neuronotrophic factors other than NGF. Ciliary neuronotrophic factor (CNTF), originally described as a survival factor for cultured chicken embryonic parasympathetic neurons, also displays trophic activity for sympathetic and sensory ganglion neurons (Barbin et al., 1984; Manthorpe et al., 1986; Ernsberger et al., 1989; Lin et al., 1990). CNTF induces cholinergic differentiation in cultured newborn rat sympathetic neurons (Saadat et al., 1989) and promotes the differentiation of sympathetic neuroblasts (Ernsberger et al., 1989). In addition, CNTF participates in type-2 astrocyte development in vitro (Lillien et al., 1988). In vivo, CNTF is reported to spare motoneurons from degeneration after axotomy (Arakawa et a1. , 1990). Unlike NGF, which is present in milligram amounts in mouse salivary glands, CNTF is relatively less abundant in tissue, although sciatic nerve may yield a good natural source (Lin et al., 1990). Even so, extensive purification is required to rule out copurifying proteins as a source for the observed effects. Recombinant DNA technology provides an alternative source for neuronotrophic molecules. Recently, Lin et al. (1989) and Stockli et al. (1989) reported the cloning and expression

Received February 15, 1991; revised April 11, 1991; accepted April 12, 1991. Address reprint requests to Dr. A. Negro, Advanced Technology Division, Fidia S.p.A., via Ponte della Fabbrica 3/A, 35031 Abano Terme (PD), Italy.

252

Negro et al. ttcatatggcttttactgagcattcac 1

tgqqacaqttgaqttaaqqg~GCTTTCACAGAGCATTCACCGCTGACC

51

CCTCACCGTCGGGACCTCTGTAGCCGCTCTATCTGGCTAGCAAGGAAGAT

101

TCGTTCAGACCTGACTGCTCTTACGGAATCCTATGTGAAGCATCAGGGCC

151

TGAACAAGAACATCAACCTGGACTCTGCGGATGGGATGCCAGTGGCAAGC

201

ACTGATCAGTGGAGTGAGCTGACCGAGGCAGAGCGACTCCAAGAGAACCT

251

TCAAGCTTATCGTACCTTCCATGTTTTGTTGGCCAGGCTCTTAGAAGACC

301

AGCAGGTGCATTTTACCCCAACCGAAGGTGACTTCCATCAAGCTATACAT

351

ACCCTTCTTCTCCAAGTCGCTGCCTTTGCATACCAGATAGAGGAGTTAAT

401

GATACTCCTGGAATACAAGATCCCCCGCAATGAGGCTGATGGGATGCCTA

4 51

TTAATGTTGGAGATGGTGGTCTCTTTGAGAAGAAGCTGTGGGGCCTAAAG

501

GTGCTGCAGGAGCTTTCACAGTGGACAGTAAGGTCCATCCATGACCTTCG

551

TTTCATTTCTTCTCATCAGACTGGGATCCCAGCACGTGGGAGCCATTATA

601

TTGCTAACAACAAGAAAATGucaqttaqtccctctctcttccttqct cgt caat cagggagagacgt cggg

Fig. 1. Nucleotide sequence of cDNA clones for CNTF obtained from human retina and brain cDNA libraries. The adjoining sequences in small letters indicate the oligonucleotidesused for PCR amplification. of rabbit and rat CNTF, respectively. However, no purification schemes or protein yields were given in these two studies. Here we describe, for the first time, the synthesis, purification, and characterization of human CNTF. Immunoreactive antibodies were prepared against both synthetic CNTF peptides and intact CNTF, the latter also blocking human CNTF biological activity in vitro.

MATERIALS AND METHODS Bacterial Strains and Reagents The restriction endonuclease and T4 DNA ligase were purchased from Bethesda Research Laboratories. Complementary DNA (cDNA) was prepared using a cDNA kit available from Stratagene. Deoxyoligonucleotide primers were synthesized using phosphoramidite chemistry and an applied Biosystems 380B automated DNA synthesizer. Polymerase chain reaction (PCR) was performed using the Amplitag kit (Cetus-Perkin Elmer). The host cell line BL21(D3) and plasmid pT7.7 were generously provided by Dr. Stan Tabor (Harvard Medical School). DNA manipulaton, transformation, and plasmid purification were performed according to published procedures (Sambrook et al., 1989). Cloning of Human CNTF cDNA Two Xgtl 1 cDNA Iibraries from human retina and brain (Clontech) were independently screened for human

CNTF. The probe used for screening corresponded to the second exon of human CNTF obtained by the screening of a genomic EMBL3 library (Negro et al., unpublished data). The nucleotide sequence obtained from two independent cDNA clones is shown in Figure 1.

Construction of the Expression Plasmid Two oligonucleotides derived from the 5‘ and 3‘ regions of the CNTF gene were used as primers for amplification of the coding region by PCR (Saiki et al., 1988). Some codons were changed in order to decrease the structural stability of mRNA (Zucher and Stiegler, 1981), without changing the corresponding amino acid sequence. Restriction sites NdeI and SalI were incorporated into the primers. The sequences of the deoxyoligonucleotide primers were (see also Fig. 1): NdeI (forward) S’TTCATATGGCTTTTACTGAGCATTCAC 3’ SalI (reverse) S’GGGCTGCAGAGAGGGACTAACTGC 3’

The amplified sequence was cut at the NdeI and SalI sites with endonuclease and cloned directly into the expression vector pT7.7 at the same endonuclease sites. Construction of the recombinant plasmid, termed pT7.7hCNTF, was verified by double-stranded dideoxysequences (Sanger et al., 1977).

Synthesis and Characterization of Human CNTF

Extraction and Purification of Human CNTF From E . coli Two liters of E . coli (strain BL21) carrying the pT7.7hCNTF plasmid were grown in ZB medium at 37°C with 200 pgiml of ampicillin, until reaching an O.D. of 0.8 (590 nm). The cells were induced to produce the recombinant protein as previously described (Studier and Moffatt, 1986) by addition of isopropyl-B-D-thiogalactopyranoside (IPTG) for 3 hr. After disrupting the cells by sonication, CNTF was recovered in the particulate fraction. Washing the pellet (spheroplasts) three times with 4 M urea, 0.1% Triton X-100, 50 mM Tris/ HCI, 1 mM EDTA (pH 7.4) provided for an easy, partical purification of CNTF. The recombinant protein at this stage comprised greater than 7 5 4 0 % of the total protein, as determined by optical density tracing. The insoluble pellet was then treated with 8 M urea, 50 mM Tris/HCl, 5 mM EDTA, 1 mM PMSF (pH 7.4) for 8 hr at room temperature. The soluble fraction obtained after centrifugation at 900g for 20 min at 4°C was dialyzed against 1 M urea, 50 mM TrisiHCl (pH 7.4) and loaded on a fast flow Mono Q column ( 1 X 5 cm) equilibrated with the same buffer. Fractions eluting with 100 mM NaCl, 1 M urea, 50 mM Tris/HCI (pH 7.4), which contained CNTF, were dialyzed against 50 mM ammonium acetate and concentrated in an Amicon ultrafiltration unit (10 kDa cut-off). Final purification was achieved by reverse-phase high-performance liquid chromatography (RP-HPLC) using a Vydac 300A column (0.46 x 26 cm) and a two-part gradient generated from 0.05% trifluoroacetic acid (TFA) (A) and 0.05% TFA in acetonitrile (B). The column was equilibrated in solvent A plus 10% solvent B and the sample, 500 pg of protein concentrated from the ion-exchange step, applied. The column was first washed with a linear gradient of 40% (B) over 10 min and CNTF eluted with a gradient of 100% (B) in 25 min. The flow rate was 1.25 ml/min; protein was monitored spectrophotometrically at 220 nm. Analytical SDS-PAGE, Western Blot Analysis, and Isoelectric Focusing Bacterial culture samples were analyzed on 12% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) (Laemmli, 1970) stained with Coomassie brilliant blue. For preparative runs, the gel was soaked with 1 M KCl and the band, visualized under ultraviolet light, was cut out and placed in an eppendorf tube with phosphate-buffered saline (PBS). The gel slices were crushed and agitated for 4 hr, and the gel pieces removed by centrifugation. Electroblotting was performed onto nitrocellulose membranes (BioRad) overnight at 2 mA constant current, blocked in 5% bovine serum albumin (BSA) for 1 hr, and incubated for 2

253

hr with affinity-purified rabbit antisera diluted 1 :50. All sera were diluted in PBS containing 0.1% BSA. After washing three times with PBS containing 0.1% BSA, the blots were incubated with goat anti-rabbit IgG for 1 hr. Blots were washed and immunoreactivity visualized with Auroprobe Immunogold Silver stain (Janssen). Isoelectric focusing was performed with the PhastSystem (Pharmacia) following the manufacturer's instructions. All operations were carried out at 25°C. Protein concentrations were determined by the BCA protein assay (Pierce).

Amino-Terminal Sequencing The amino-terminal sequence of the purified human CNTF was determined by Edman degradation using an Applied Biosystems model 477A protein sequencer. Released phenylthiohydantoin derivatives of amino acids were identified on line. Circular Dichroism (CD) Measurements Measurements of CD were carried out using a Jasco J-500 spectropolarimeter equipped with a Jasco DP-501 data processor, at a sensitivity of 1 mdeg/cm and with an instrument time constant of 8 sec. Rectangular cuvettes with a 1-mm path length were used. Spectra were averaged from at least ten scans, and are presented as mean residue ellipticity, [el, in units of deg . cm2/ dmol-', based on the equation: [el = Mw/L C . 10, where 8 is the observed ellipticity, Mw is the mean residue weight taken as 120, L is the path length, and C is the protein concentration. Estimates of helical secondary structure from far-ultraviolet CD spectra were determined (Greenfield and Fasman, 1969) according to the following empirical formula: %helix

=

[~1208nrn + -

4000 29000

Preparation of Antibodies Peptides A (amino acids 143-160): PINVGDGGLFEKKLWGLK and B (amino acids 44-59): LNKNINLNSADGMPVA from the deduced human CNTF sequence were synthesized using Fmoc chemistry with a Milligen 9050 automated peptide synthesizer (Milford, MA). Purification was achieved by RP-HPLC on a BioRad CIS HP semipreparative column (0.46 X 26 cm) matched to a waters 600E HPLC using a linear 0.1% TFA-water/acetonitrile gradient. Antisera were produced by injecting subcutaneously in multiple dorsal sites two female New Zealand white rabbits with an emulsion containing 0.5 mg of peptide in 2.5 ml of PBS (pH 7.5) and 3.5 ml of complete Freund's adjuvant. A booster injection of 0.25 mg peptide in 1 ml of PBS and 1.5 ml of incomplete Fruend's adjuvant was given intramuscularly

254

Negro et al.

A

Nde I

Sol I

D T ~7. hCNTF

kDa94

-

67

-

43

-

30

-

20

-

14

-

B

C

D

E

F

I

Fig. 2. Expression vector pT7.7hCNTF with the coding sequence of the human CNTF gene.

3 weeks later. A second booster injection was given 16 weeks after immunization was started, and then monthly. The titers of peptide antisera were determined by ELISA using peptide-coated microwells. The IgG fraction was precipitated by caprylic acid treatment and further purified by affinity chromatography (Steinbuch and Audran, 1969). The protein concentration of the two IgGs was 50 pglml and 75 pg/ml, respectively, for peptides A and B. Antibodies to intact CNTF were obtained with the same technique, injecting 0.25 mg of protein purified after SDS-PAGE. Antisera were heated at 56°C for 30 min prior to use in culture.

Assay of CNTF Biological Activity The activity of recombinant CNTF was determined based on the known ability of CNTF to maintain neurons from embryonic day 10 (E10) chicken dorsal root ganglia (DRG), but not those from E8 DRG (Barbin et al., 1984; Manthorpe et al., 1986; Ernsberger et al., 1989; Lin et al., 1990). Cells were prepared as described (Skaper et a]., 1990). Tissue culture wells (0.16 cm2/well, Falcon) were precoated with polylysine and type I collagen (Vitrogen). Each well received varying dilutions of CNTF in 50 pl of culture medium (Dulbecco’s modified Eagle’s medium with 100 units/ml penicillin, 2 mM Lglutamine, and 10% heat inactivated fetal calf serum) and 50 pl of cell suspension diluted with culture medium (2,000 enriched neurons/well). After 48 hr, cultures were fixed with glutaraldehyde and examined under phase-contrast microscopy. Two perpendicular strips across each well (approximately 20% of total well area) were assessed. The number of cells with processes

Fig. 3. SDS-PAGE (A-B) and Western blot (C-F) analysis of cell lysates from plasmid pT7.7hCNTF or control plasmid pT 7.7. See text for details. Lane A: total proteins obtained from E . coli BL21 transformed with plasmid pT7.7 and induced with IPTG. Lane B: total proteins obtained from E. coli BL21 transformed with plasmid pT7.7hCNTF and induced as in lane A. Western blots were developed using lane B material and the following antibodies: lane C: antiserum against a CNTF nonrelated peptide; lane D: antiserum against CNTF peptide B; lane E: anti-CNTF peptide A affinity-purified IgGs; lane F: antiserum against CNTF purified by SDS-PAGE. greater than two soma1 diameters per strip was counted and the values from the two strips averaged. For antibody studies, human CNTF at a constant concentration (50 ng/ml) was preincubated for 2 hr at 37°C with increasing dilutions of antisera.

RESULTS The coding sequence of the human CNTF gene was cloned into the expression vector pT7.7 (Fig. 2). This sequence was obtained by the screening of a h g t l l human cDNA library (Fig. 1). The E. coli strain BL21(D3) transformed with the recombinant vector pT7.7hCNTF expressed high levels (30% by gel scanning) of a single protein having a molecular weight expected for CNTF (Stockli et al., 1990; Lin et al., 1990). Since CNTF was not expressed prior to induction of transcription with

Synthesis and Characterization of Human CNTF

255

C

I

I

20

40

6'0

80

Fraction number Fig. 4. Fast protein liquid chromatography of human CNTF using a fast flow Mono Q column. Material ( 2 mg) from the inclusion bodies extract was loaded and eluted as described in the text. Bar indicates region of CNTF biological activity. A:

IPTG, or by the same bacterial cell host transformed with plasmid pT7.7, the protein produced was clearly encoded by the recombinant plasmid (Fig. 3A,B). The eluate from such gels maintained E l 0 DRG neurons in culture. The ability of antibodies prepared against synthetic human CNTF peptides to recognize on Western blots the protein synthesized by transformed E. coli (Fig. 3C-F) further confirmed its identity as CNTF. The peptide sequences chosen were based on the hydrophobic profile methods of Hopp and Woods (1981) obtained by computer analysis. The few epitopes available on these small peptides may limit the diversity and intensity of the antibody response, for example, in blocking biological activity. However, their ability (but not that of a peptide unrelated to CNTF) to recognize the native CNTF protein would appear to justify the peptide sequences selected. Human CNTF is insoluble in E. coli, being contained within inclusion bodies (data not shown). This characteristic has been used to advantage to purify the recombinant protein, using several washings and subsequent exposure to 8 M urea for solubilization. Because this material contained large (75-80%) amounts of CNTF, two chromatographic steps were sufficient to achieve purification. The ion exchange elution profile is shown in Figure 4, and that for RP-HPLC in Figure 5 .

0.1 M NaCI; B: 0.2 M NaCl; C: 0.4 M NaCI; D: 1 M NaCl. Each step was performed in 1 M urea, 50 mM TrisiHCI, pH 7.4.

80

la

-

6o m c

I

-m>

u

0 u)

2

-

40 aQ

,

0.5

I

I

20

Fig. 5 . Reverse-phase HPLC of human CNTF from the active fractions in Figure 4. Material was loaded and eluted as described in the text. The peak between 27 min and 29 min contained CNTF biological activity.

256

Negro et al.

A

B

C

D

E

F

kDa

43

0 -

30

20

I

II

I I

II

II

200

210

220

230

240

250

nm

14 Fig. 6. SDS-PAGE analysis of recombinant human CNTF during purification. Lane A: induced bacterial cells carrying the plasmid pT7.7; lane B: induced bacterial cells carrying the plasmid pT7.7hCNTF; lane C: pellet (from inclusion bodies) in 6 M guanidiniurdHC1; lane D: soluble protein from lane C material; lane E: mono Q column (Fig. 4) pool from lane D proteins; lane F: RP-HPLC pool (Fig. 5 ) from lane E proteins. See Materials and Methods for preparation of gels. The material obtained after the last step was essentially homogeneous, the yield being 10 mg CNTF protein per liter cell broth. SDS-PAGE analysis for the proteins obtained from each purification step is illustrated in Figure 6. The CNTF on SDS-PAGE had a molecular weight of approximately 24 kDa, and a PI of 5.8 (Fig. 6). The following amino acid sequence for the NH2 terminal region of human CNTF was obtained: NH2-Ala-PheThr - Glu - His - Ser- Pro- Leu-Thr- Pro- His- Arg- Arg- Asp. This sequence corresponds to the amino acid sequence (Fig. 8) deduced from the cDNA nucleotide sequence of Figure 1, indicating that the recombinant protein was correctly synthesized. The CD spectrum for human CNTF is illustrated in Figure 7 and indicates that this molecule, in large part, is in alpha-helical form. A putative three-dimensional structure was obtained by applying PC: GENE computer software (IntelliGenetics, Inc.), the secondary structure

Fig. 7. Far-ultraviolet CD spectra of recombinant human CNTF were recorded at 25°C in 50 mM Tris-HC1 buffer, pH 7.5. Concentration of CNTF: 0.2 mg/ml.

program of Garnier et al. (1978), and the antigenic (hydropathic) program of Hopp and Woods (1981). Following these criteria, the more hydrophilic portions of the molecule should residue at the periphery, with the core of the protein being more hydrophobic and alpha-helical in nature (Fig. 8). The human CNTF protein expressed in E. coli was biologically active, having an ED50 of approximately 30 pM (based on a molecular weight of 24 kDa) after the last purification step, based on the survival of E l 0 chicken sensory ganglion neurons (Fig. 9). This value suggests that human CNTF exerts its trophic effects via highaffinity receptors, and is similar to the activity of purified recombinant rat CNTF reported using E8 chicken ciliary ganglion neurons (Squint0 et al., 1990). Extracts from pT7.7-transformed E . coli displayed no trophic activity (data not shown). The neuronal specificity of purified human CNTF was verified using cultures of E8 chicken DRG neurons. As Table I shows, the recombinant protein maintained E10, but not E8 DRG neurons, whereas mouse NGF supported both E8 and E l 0 DRG neurons (Barbin et al., 1984; Manthorpe et al., 1986; Ernsberger et al., 1989; Lin et al., 1990). The CNTF protein recovered from SDS-PAGE gels was also biologically active. Furthermore, the trophic activity of human CNTF was blocked

257

Synthesis and Characterization of Human CNTF

K N I N L _.

M

~ ~ D E L L ~ A L L ~ ~ F T ~ Y A ~ L N L O L R L A E T L E S W ~

' i " 'P 'p -%00 1 E

G

D

F

~

"S 199 H Y

O

A

I

1&l

~

T

L

L

L

O

V

A

A

F

A

Y

170

~

I

E

L

L

160

M

I

L

L

"'"I': E

Y

K

I

P

~

N

E

A

O

G

~

150

T

chhhhhh/ I

A N N K K M

Fig. 8. A proposed three-dimensional structure for human CNTF. The amino acid sequence was deduced from the nucleotide sequence of the cDNA clone in Figure 1. a-helic ( w ); @-turn ( WV); extended (-); coil (.).

by an antiserum prepared against the purified protein (Table 11) but not by affinity-purified anti-mouse NGF IgC. Such anti-mouse NGF antibodies blocked the activity of mouse NGF and human NGF (Bigon et al., 199 1). In contrast, antibodies against CNTF peptides that recognized the CNTF protein on immunoblots (Fig. 3) failed to inhibit CNTF biological activity (data not shown), suggesting that the antigenic determinants constituted by these peptides may not be involved in the CNTF active site.

DISCUSSION We report here the transformation of E . coli with a recombinant vector for human CNTF and its functional expression. The human CNTF was purified to greater than 97% homogeneity. A partial amino acid sequence indicated that the recombinant protein was correctly synthesized. Antibodies generated to synthetic CNTF peptides based on the complete amino acid sequence recognized the intact CNTF on immunoblots. In general agreement with previous studies showing CNTF to be an

acidic protein (Barbin et al., 1984) with a molecular weight of 22-24 kDa (Lin et al., 1989, 1990; Stockli et al., 1989), the molecular weight of recombinant human CNTF by SDS-PAGE and Western blots was approximately 24 kDa. The heterogeneity of purified natural mammalian CNTF observed on Western blots using antibodies raised against synthetic CNTF peptides (Lin et a]., 1989, 1990) could result from degradation during purification. The CNTF polypeptide obtained here may represent the mature sequence. It is possible that a protease is involved in the normal release mechanism of CNTF; if true, the present recombinant CNTF might be used for such studies. The human CNTF migrated on SDS gels at an apparent molecular weight slightly larger than that predicted from the primary sequence. This is in accord, however, with the natural rabbit CNTF (Lin et al., 1990) and recombinant rat CNTF (Squinto et al., 1990), and may reflect the hydrophobic nature of CNTF. The high concentration of acetonitrile needed to elute CNTF by RP-HPLC is also indicative of a rather hydrophobic molecule. In fact, this characteristic of CNTF has been used

258

Negro et al.

Fig. 9. Concentration-dependent survival of E l 0 chicken sensory neurons in the presence of human CNTF. Neurons (2,000 per well) were grown with the indicated amounts of purified

(RP-HPLC) CNTF for 48 hr. Numbers represent surviving S.D.) for triplicate wells. The ED50 is apneurons (mean proximately 0.65 ngiml.

TABLE I. Biological Specificity of Human CNTF*

tively resistant to harsh chemical treatments such as 8 M urea, exposure to pH 3.6, and the acid organic solvent used in RP-HPLC. This biological stability of CNTF proved to be advantageous when carrying out the described purification scheme. These particular chemical behaviors of human CNTF will discussed below, in relation to a proposed three-dimensional structure for this protein. The recombinant human CNTF promoted the survival of E 1 0 , but not E8 chicken DRG neurons, as expected from its reported biological specificity (Barbin et al., 1984; Manthorpe et al., 1986; Ernsberger et al., 1989; Lin et al., 1990). The ED50 for human CNTF was 30 pM using E l 0 DRG neurons, comparable to that found for recombinant rat CNTF with E8 ciliary ganglion neurons (Squinto et al., 1990). In addition, human CNTF activity was blocked by a polyclonal antiserum against the mature polypeptide, but not by antibodies to mouse NGF. Similarly, the CNTF antiserum did not block the trophic activity of mouse or human NGF. Such antibodies will be valuable in elucidating the sites of CNTF synthesis in the nervous system. The circular dichroism data for human CNTF (Fig. 7) and the preponderance of alpha-helicity (Fig. 8) probably account for the stability of CNTF in the presence of SDS, thus permitting the molecule to refold into its biologically active form upon removal of SDS. The structure for human CNTF proposed in Figure 8 is, of course, by no means complete. However, the ability to generate large amounts of the pure protein will provide material

Neurons per well Neuron type

NGF

CNTF

*

40 20 770 t 90

EX DRG El0 DRG

920 t 11 850 -C 120

~

*Culture wells were seeded with 2,000 enriched neurons in medium containing purified human CNTF (50 ngiml) or mouse NGF (20 ng/ ml). Cell numbers were assessed 48 hr later. Values are mean t S.D. for triplicate wells. Background levels were approximately 50-100 neurons per well.

TABLE 11. Antibodies to Human CNTF Block CNTF Biological Activity* Neurons per well Antibodies used None Anti-human CNTF Anti-mouse NGF

CNTF

NGF

780 2 92 120 i- 30 740 t 100

790 -C 85 820 2 110 190 -C 50

"Culture wells were seeded with 2,000 enriched E l 0 DRG neurons in medium containing purified human CNTF (50 ngiml) or mouse NGF (20 ngiml). Anti-CNTF antibodies were added as a 1:40 dilution of a whole antiserum raised against the CNTF. Anti-NGF antibodies were added at 1 pgiml of affinity-purified anti-mouse NGF IgG. Cell numbers were assessed 48 hr later. Values are mean t S.D. for triplicate wells.

in the purification of rabbit sciatic nerve CNTF by hydrophobic interaction chromatography (Lin et al., 1990). Interestingly, the biological activity of CNTF was rela-

*

Synthesis and Characterization of Human CNTF

for X-ray crystallographic analysis, thereby defining more precisely the three-dimensional picture of CNTF. Some antibodies against selected CNTF peptide sequences (which are CNTF-immunoreactive) failed to block CNTF trophic activity. In addition, certain synthetic human CNTF peptides were unable to generate immunoreactive immunoglobulins. This antigenic approach, together with site-directed mutagenesis, may be used to delineate regions of the human CNTF molecule necessary for biological activity (Negro et al., in preparation). When CNTF is expressed in eukaryotic systems it is not secreted, but is retained in the cytoplasm, making cell lysis necessary to obtain a biologically active material (Lin et al., 1989; Stockli et al., 1989; Negro et al., unpublished observations). Such systems are time-consuming and are not practical for large-scale preparation of CNTF. The particular structural characteristics of CNTF, namely, no apparent preprosequence, glycosylation sites, or intrachain disulfide bonds thus make E. coli a good system. The removal of the amino-terminal MET residue of human CNTF in E. coli is not surprising, since in E. coli the first amino acid is always removed when the second amino acid in the protein is ALA (Hirel et al., 1989). This modification, however, does not influence the biological activity of human CNTF. The exceptionally high yield of CNTF in the E. coli system (30%of total cellular protein, 10 mg purified CNTF per liter cell broth) provides another advantage over eukaryotic systems. The physiological role(s) of CNTF in the nervous system remains to be clarified. In vitro, CNTF promotes cholinergic differentiation of embryonic chicken sympathetic neurons (Saadat et al., 1989), inhibits proliferation and induces differentiation of embryonic chicken sympathetic neurons (Emsberger et al., 1989), and plays a role in type-2 astrocyte differentiation (Lillien et al., 1988, 1990). It is not known if CNTF displays any of these activities in vivo. The survival activity of CNTF for avian embryonic motoneurons in culture (Arakawa et al., 1990), together with the observation that CNTF prevents lesion-mediated degeneration of motoneurons in newborn rats (Sendtner et al., 1990), makes this molecule a promising candidate for the treatment of degenerative disorders of motoneurons. The efficient production and purification of recombinant human CNTF reported here will provide an important tool for such studies.

ACKNOWLEDGMENTS We wish to thank Dr. Sonia Mattioli for affinity purification of CNTF antibodies, Dr. Carlo Moretto for the CD spectrum, Dr. Patrizia Polverino for protein se-

259

quencing and Roberta Drigo for typing and editing the manuscript.

REFERENCES Arakawa Y, Sendtner M, Thoenen H (1990):Survival effect of ciliary neurotrophic factor (CNTF) on chick embryonic motoneurons in culture: Comparison with other neurotrophic factors and cytokines. J Neurosci 10:3507-3515. Barbin G, Manthorpe M, Varon S (1984): Purification of the chick eye ciliary neuronotrophic factor. J Neurochem 43 Barde YA (1989): Trophic factors and neuronal surv 1525-1534. Bigon E, Soranzo C, Minozzi C, Skaper SD, Callegaro L (1991): Large scale purification and immunological characterization of human placenta nerve growth factor. J Neurochem Res (in press). Dreyfus CF (1989): Effects of nerve growth factor on cholinergic brain neurons. Trends Pharmacol Sci 10: 145-149. Ernsberger U, Sendtner M, Rohrer H (1989): Proliferation and differentiation of embryonic chick sympathetic neurons: Effects of ciliary neurotrophic factor. Neuron 2: 1275-1284. Garnier J, Osguthorpe DJ, Robson B (1978): Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J Mol Bid 120:97-120. Greenfield N, Fasman GD (1969): Computed circular dichroism spectra for the elucidation of protein conformation. Biochemistry 10:4108-4116. Hirel P-H, Schmitter J-M, Dessen P, Fayat G , Blanquet S (1989): Extent of N-terminal methionine excision from Escherichia coli proteins is governed by the side-chain length of the penultimate amino acid. Proc Natl Acad Sci USA 86:824?-825 1. Hopp TP, Woods KR (1981): Prediction of protein antigenic determinants from amino acid sequences. Proc Natl Acad Sci USA 78:3824-3828. Laemmli UK (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 22?:680-685. Levi-Montalcini R (1987): The nerve growth factor thirty five years later. Science 237:1154-1162. Lillien LE, Sendtner M, Rohrer H, Hughes SM, Raff MC (1988): Type-2 astrocyte development in rat brain cultures is initiated by a CNTF-like protein produced by type-1 astrocytes. Neuron 1:485-494. Lillien LE, Sendtner M, Raff MC (1990): Extracellular matrix-associated molecules collaborate with ciliary neurotrophic factor to induce type 2-astrocyte development. J Cell Biol 1 I1:653644. Lin L-FH, Mismer D, Lile JD, Armes LG, Butler ET 111, Vannice JL, Collins F (1989): Purification cloning, and expression of ciliary neurotrophic factor (CNTF). Science 246: 1023-1025. Lin L-FH, Armes LG, Sommer A , Smith DJ, Collins F (1990): Isolation and characterization of ciliary neurotrophic factor from rabbit sciatic nerves. J Biol Chem 265:8942-8947. Manthorpe M, Skaper SD, Williams LR, Varon S (1986): Purification of adult rat sciatic nerve ciliary neuronotrophic factor. Brain Res 38?:282-286. Saadat S , Sendtner M, Rohrer H (1989): Ciliary neurotrophic factor induces cholinergic differentiation of rat sympathetic neurons in culture. J Cell Biol 108:1807-1816. Saiki RK, Gelfand DH, Stoffel S , Scharf SJ, Higchi R, Horn GT, Mullis B, Erlich HA (1988): Primer-directed emzymatic amplification of DNA with thermostable DNA polymerase. Science 239:487-49 I .

260

Negro et al.

Sambrook SD, Fritsch EF, Maniatis T (1989): “Molecular Cloning: A Laboratory Manual,” 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Press. Sanger F, Nicklen S , Coulson AR (1977): DNA sequencing with chain terminating inhibitor. Proc Natl Acad Sci USA 74:5465-5468. Sendtner M, Kreutzberg GW, Thoenen H (1990): Ciliary neurotrophic factor prevents the degeneration of motoneurons after axotomy . Nature 345:440-441. Skaper SD, Facci L, Milani D, Leon A, Toffano G (1990): Culture and use of primary and clonal neural cells. In Conn PM (ed): “Methods in Neurosciences,” vol. 2. New York: Academic Press, pp. 17-33. Squinto SP, Aldrich TH, Lindsay RM, Morrissey DM, Panayotatos N, Bianco SM, Furth ME, Yancopoulos GD (1990): Identifi-

cation of functional receptors for ciliary neurotrophic factor on neuronal cell lines and primary neurons. Neuron 5:757-766. Steinbuch M, Audran R (1969): The isolation of IgG from mammalian sera with the aid of caprylic acid. Arch Biochem Biophys 134: 279-284. Stockli KA, Lottspeich F, Sendtner M, Masiakowski P, Carroll P, Gotz R , Lindholm D, Thoenen H (1989): Molecular cloning, expression and regional distribution of rat ciliary neurotrophic factor. Nature 242:920-923. Studier FW, Moffatt BA (1986): Use of bacteriophage T7RNA polymerase to direct selective high-level expression of cloned genes. J Mol Biol 189:113-130. Zucher M , Stiegler P (1981): Optimal computer folding of large RNA sequences using thermodynamics and auxiliary information. Nucleic Acids Res 8:133-148.