245

Biochimica et Biophysica Acta, 1096 (1991) 245-252 c' 1991 Elsevier Science Publishers B.V. 0925-4439/91/$03.50 ADONIS 092544399100075T

BBADIS 61032

N o v e l muteins of human tumor necrosis factor a Ryoji Ito 1, Hiroshi Matsumoto 1, Kumiko Uchida 1, Toshiyuki Kubo ~, Yuji Tsukii 1, Tohru Endo i and Akira Kaji 2 I Research Institute for Molecular Genetics, Tsumura and Company, Ibaraki (Japan) and 2 Department of Microbiology, School of Medicine, University of Pennsyloania, Philadelphia, PA (U.S.A.)

(Received 31 July 1990)

Key words: Tumor necrosis factor; Synthetic gene; Gene expression; Site-directed mutagenesis; Random mutagenesis. Mutein; (Human)

For chemical synthesis of a gene coding for human tumor necrosis factor a (TNF-a), DNA sequence predicted by the amino acid sequence of human TNF molecule was prepared. Codons were chosen according to the codon usage in Escherichia coli (E. coli). The 490 bp gene was assembled by enzymic ligation of 42 oligonucleotides and was cloned into a vector (pKK223-3) for high expression of active TNF-a in E. coli. With use of site-directed mutagenesis on this DNA, five different muteins of TNF-a were synthesized. TNF-M1 and TNF-M4 have deletions of His-73 and Gin-102, respectively. These deletions didn't cause loss of the cytotoxic activity against L929 cells. TNF-M5, which has a substitution of Asp-10 to Arg, had the similar cytotoxic activity to that of TNF-a. The cytotoxic spectra against several tumor cells were not changed by this substitution. TNF-M3 has an amino acid substitution of G l u - l l 6 to His which occupies this position in human TNF-fl. This substitution didn't change the cytotoxicity. In addition, evidence was presented that the change of the carboxyl terminal residue doesn't always influence the cytotoxic activity of TNF-a. Many different muteins were also isolated by random mutagenesis with hydroxylamine-HCl. One of the muteins, which carries a mutation of His-15 to Tyr, lost the cytotoxic activity almost completely.

Introduction Tumor necrosis factor a (TNF-a) is a secreted polypeptide that was discovered in the serum of animals (mouse, rat and rabbit) injected sequentially with Bacillus Calmette Guerin and endotoxin [1]. Its name came from the observation that the polypeptide caused necrosis of certain tumors in vivo. T N F - a is produced by macrophages and several other cells, and it has a wide array of biological activities: inhibition of lipoprotein lipase synthesis [2]; stimulation of fibroblast growth [3]; induction of MHC class I [4] and acute phase proteins [5]. Human T N F - a was purified to homogeneity from

Abbreviations: cDNA, complementary deoxynucleic acid; DEAE, diethylaminoethyl; ELISA, enzyme-linked irnmunosorbent assay; HIC, hydrophobic interaction chromatography; HPLC, high-performance liquid chromatography; IPTG, isopropyl-/3-D-thiogalactopyranoside; MHC, major histocompatibility complex ; MTT, 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. Correspondence: A. Karl, Department of Microbiology, School of Medicine, University of Pennsylvania, Philadelphia, PA, U.S.A.

cell culture fluid of HL-60 promyelocytic leukemia cell line stimulated by 4/3-phorbol 12-myristate 13-acetate [6]. Its cDNA and genome were cloned [7] and it was found that T N F - a is a non-glycosylated cytokine with a molecular mass of 17 kDa and exists functionally in a trimeric form. The mature polypeptide of 157 amino acids is preceded by a sequence of 76 residues which is cleaved during the maturation process. The technique for introducing site-directed mutagenesis in a given gene has provided a powerful tool for increasing our knowledge on the structure and function of proteins. Initial studies of structure-function relationships of human T N F - a molecule have been reported [8-11]. Their results indicated that the carboxy-terminal region, but not the amino-terminal region, was required for the cytotoxic activity and that the replacement of both Cys-69 and Cys-101 by other amino acids didn't result in great loss in the cytotoxicity. Recently, Yamamoto et al. [12] showed that His-15 played an important role in the cytotoxicity of TNF-a. Synthesis of genes coding for small proteins of which amino acid sequence is known such as human epidermal growth factor [13] and human growth hormone [14], was carried out easily by the automatic DNA synthesis

246 technique without isolating the natural gene. For expression of proteins, natural codons can be changed by this method to those used with high frequency in bacteria or other organisms, depending on where the expression vectors are placed. Taking advantage of this approach,

Met-Val.Arg.

Ser.

+

Ser.

1

Ser.Arg.Thr.

+

we chemically synthesized a gene coding for human TNF-e~ and expressed it in E. coli. In addition, muteins were constructed by site-directed and random mutageneses and their cvtotoxic effects were tested. It v,as found, contrary to the general concept, except for the

Pro.

Ser-Asp-Lys.Pro.Val.Ala-His.Val'Val'A

+

3

+

+

5

+

AATTC.ATG.GTT.CGT.TCC-TCT.TCC.CGT.ACT-CCA.TCT.GAC.AAA.CCA.~TA'G~T'CAT.GTT'GTT'G G-TAC.CAA.GCA.AGG.AGA.AGG-GCA.TGA.GGT-AGA.CTG.TTT.~GT-CAT.CGA.GTA.CAA.CAA.C 2

la -Asn.

Pro

CT •AAC

•CCG.

GA •TTG

•GGC

.Gln + CAG, •GTC

.Ala

GCA, •CGT

4

•Glu.

GAA .tTT

Gly,

Gln-

7 + •GGC •CAG. •CCG

• GTC

Leu-Gln.

Trp.

TTA,

+ TC4~, CTT

•AAT

CAG, - GTC

•ACC

6

Leu.Asn.

•GAA,

•AA~,

+ CGT

TTG

•GCA,

8 la.Asn.Gly,

CG • AAC

Val,

ii + • C~CnC • G T A

GC • TTC~ •CCC~ •CAT

Glu.

Arg.Arg.Ala-Asn.Ala 9 -CGT

•GCA

+ .AAC,

GCA

•CGT

• TTG

Leu.Arg.Asp.Asn-Gln.

Leu

°Val,Val.

Pro-

Ser.Glu

+ • C~AA •CTC~ • CC~T • C~AT • AAC

13 •CAA

+ • CTC

+ •C~TA •C~TT •CCC~ •TCC

•CTT

•GTT

•GAG

•CAT

•C~AC •C.CA •CTA

•T~

•CAA

Gln-

Val

T~ • TAC

•AGC

+ •CAG

17 •GTT

AG • ATG

•TCG

• C~TC •CAA

• Leu"

•CTG

Phe

• Lys

-Gly.Gln.

+ • TTC •AAG

•C~AC • AAG

Ile.Ser.Arg-

•ATC

• TCC

° Gly,

•CCA

•TTC

Ile.Ala.Val.

•ATC

• _TC~A • T A C ~ • A C ~ G • C ~ C A • T A G

CTG,

G

•GAC

•C

•GCA

Leu

•Tyr.

Leu"

I

+ • CTC~ • A •GAC

•T

Gly.

Cys

• Pro.

Ser.

Thr-

His

-Val-

Leu"

Leu"

+

Thr"

•~AG

+ •GGT

• TGC

19 •CCG

+ •AGC

•ACT

•CAC

•GTT

• TTG

•CTG

•CCA

•C~TC • CCA

•ACG

•C~C

•TCG

•TGA

•GTG

•CAA

• AAC

• C~AC • TC~A •G

+ • TCT

• TAC

•CAA.

• CC~T •CAA

• AGA

• ATC~ •GTT

+ •C

22

Ser.Tyr.Gln.Thr.Lys-Val.Asn.

•GTT

21 •ACT

H

•GGT

+23 •CGT

•GAC

•C~AC •ATG

20

+ AT •ACT

•CGT

•CTG,

16

18 is.Thr.

A

+ GCA

+ 15 • C ~ A A • C~C~T • C T C . • T A C

• G C ~ C • AC~C~ • C T T

14

Ser.

• Leu-

i0

12

le- Tyr.

• Leu

Leu-Leu'Ser'Ala"

+

25 •CTG

+

ACT

•AAG

•GT~

•AAC

• TGA

• TTC

•CAC

• TTC~ • C~AC • C~AC • AC~A •CGA

24

•CT~

Ile'L

+ • TCT

•~T

•ATC

•A

• TAG

•T

26

ys" Ser. Pro-Cys .Gln.Arg-Glu.Thr-Pro.Glu.Gly.Ala.Glu-Ala. Lys.Pro.Trp.Tyr'Glu'Pro"I AA" TCC.

27 + CCG. TGC

T T • A C ~ G • C~C~C • A C G

•CAA. •GTT

CGC.

+ GAA.

•GCC~ •CTT

ACC. • TGG

CCA. •GGT

+29 GAA.GGC •CTT

28

.GCT

•CCC~ • CGA

•GAA" •CTT

+ GCT

"AAA"

•CC.A • TTT

CCG.

+ TGG.

•C~qC •ACC

31 TAC. GAA. • ATC~ •CTT

30

CCG"

+ A

• C~qC • T

32

•e.Tyr•Leu•G•y•G•y•Va••Phe•G•n•Leu•G•u•Lys•G•y.Asp•Arg•Leu•Ser•A•a•G•u•I•e•Asn•A +

+

33

+

+

35 +

+

TC•TAT.CTC•GGT•GGT•GTA.TTC•CAG•CTC.GAA.AAG•GGT•GAT•CGT•CTG•TCC•GCA•GAG•ATC°AAC°C

A~ATA~A~C~A~CA.CA~AA~C~A~C~T~TT~TA~GCA~AC~AGG~C~T~CTC'TA~'TT~G 34

36

rg.Pro.Asp.Tyr.Leu.Asp.Phe.Ala.Glu.Ser.Gly,Gln,Val.Tyr.Phe.Gly. I l e - I l e ' A l a "Leu- * GC •CCG

•GAT

CO. • C~C~C • C T A

+37 • TAC •CTG

•GAC

+ • TTT •GCT

• ATC~ •C~AC •C~C. • AAA

38

•GAG

•CC~A •CTC

40

• TCT

+39 •GGT

• AC~A • CCA

•CAA

•GTT

• C~T~ •C~A"

+ •TAC AT~

+ •TTC

oGGT

• A Ac~ • c c A

•ATC

•ATC

41 •GCT

•CTG

+ •T

• ~AC. • ~AC., CC~A •C~AC • A

42

A~,T~A,A TC'ACT'TTCGA Fig. l. Nucleotidesequenceofthesynthetichuman TNF-a gene. The positionsofthesyntheticoligonucleotidesl-42 w i t h i n t h e g e n e a n d t h e aminoacidsequenceofTNF-acodedbythegeneareshown.+ and *indicateeveryl0thnucleotideandthestopcodons, respectively.

247 results obtained by Kamijo [15], that the change of amino acid in carboxy terminus didn't necessarily lead to loss of the cytotoxicity. In addition, the change of position 116 to the amino acid (His) which exists in TNF-/~ at the same position, didn't influence the cytotoxic activity of TNF-a.

method [20] with the automatic synthesizer ABI 380A. They were purified by the combination of reverse-phase HPLC and anion-exchange HPLC. These oligonucleotides, except for fragments 1 and 42, were phosphorylated by T4 polynucleotide kinase. Four groups of oligonucleotides (A-D), which consisted of 10 or 12 fragments, were synthesized separately. Ligated products of the expected size corresponding to A, B, C and D fragments were isolated by 10% preparative polyacrylamide gel electrophoresis under non-denaturating conditions followed by elution with 0.3 M sodium acetate and ethanol precipitation. The fragments synthesized (A, B, C and D) were further ligated in a similar manner to give the complete TNF-a gene.

Materials and Methods

Recombinant DNA techniques E. coli strain LE392 [16] was used in this work. Bacteria were grown at 37°C in M9 medium supplemented with 0.2% glucose, 0.1% cassamino acids, 5 ffg/ml thiamine and 12.5 ffg/ml carbenicillin. Plasmids pSP64 [17], pKK223-3 [18] and pTZ18U [19] were obtained from Promega, Pharmacia and Toyobo, respectively. General cloning techniques have been described previously [16].

Expression of human TNF-a in E. coli and its purification The synthetic TNF-a gene from plasmid pSPTNF-a was recloned into plasmid pKK223-3 at the EcoRI and HindIII sites for expression of human TNF-a gene in E. coil. The plasmid generated, phTNF-a, was introduced into LE392 as described by Hanahan [21]. E. coli, LE392/phTNF-a, was grown at 37°C for 16 h in M9 medium. The cells were pelleted by centrifugation at 4000 × g for 5 min and suspended in the one-fiftieth

Design and construction of synthetic TNF-a gene The sequence of the synthetic TNF-a gene is shown in Fig. 1. The scheme for its chemical and enzymic synthesis is summarized in Fig. 2. Individual oligonucleotides were synthesized by the phosphoramidite

Kination

1__& ~ .7._7& ~ a" ~" ~" ~'

_11 ~

~

19

21

23 25 2 7 ~

31

33 35 37 39 41

Ligation(1 ) A

B

C

D

a

L

Ligation(2)

I

•

I

+

1

TNF~

EcoRI HindIII

~

SP6

EcoRI HindlII

Fig. 2. Construction strategy of the synthetic human TNF-a gene and its insertion into plasmids. Four groups of chemically synthesized oligonucleotides were separately ligated. Four fragments (A-D), were isolated from polyacrylamide gel and further ligated to give the TNF-a gene. The gene was cloned into plasmid pSP64 and recloned into plasmid pKK223-3 as indicated in this figure.

248 vol. of 10 mM Tris-HC1 (pH 8.0). The suspended cells were disrupted by sonication for 10 min and the cell extracts were centrifuged at 4000 × g for 15 min. The supernatant was collected and polyethyleneimine was added to a final concentration of 0.2% for eliminating D N A and RNA. After the supernatant was obtained by centrifugation at 4000 × g for 15 min, a m m o n i u m sulfate was added to a final concentration of 45% saturation. The precipitates, which contained most of TN F-a, were collected by centrifugation at 4000 × g for 15 rain and were resuspended in 10 mM Tris-HC1 (pH 7.0). After dialysis against the same solution, the fraction was loaded onto a DEAE-Sepharose CL-6B column (1.6 × 20 cm) pre-equilibrated with 10 m M TrisHC1 (pH 7.0). After the column was washed with the same solution, proteins were eluted with 10 mM TrisHCI (pH 7.0) buffer containing 0-500 mM linear gradient of NaCI. Fractions, which were relatively rich in TNF-c~, were pooled and were applied to H I C column (TSK gel phenyl-5PW, Tosoh, 7.5 mm × 75 mm) as described by Lin and Y a m a m o t o [22]. Fractions were analyzed with 15% SDS-PAGE for 4 h at 150 V [23] and in vitro cytotoxicity assay was performed as described below. T N F cytotoxicity assay The cytotoxic activities of T N F - a and its muteins were measured on actinomycin D-treated murine L929 cells as described [10]. Cell viability was determined by measuring the cellular metabolic activity with the M T T assay [24], with an automatic micro ELISA reader. Site-directed mutagenesis of TNF-a gene Site-directed mutagenesis of synthetic T N F - a gene was carried out with the uracil-containing template method developed by Kunkel [25]. The primers which were used to generate TNF-M1, TNF-M2, TNF-M3, T N F - M 4 and T N F - M 5 , were 5 ' - T G C C C G A G CACTGTTTTGCTGACT, 5'-GGTATCATCGCTATGTAGTG, 5'-CCGTGGTACCATCCGATCTATCTCGGTGGT, 5'-AATCCCCGTGCC G C G A A A C C C C and 5 ' - T C C T C T T C C C G T A C T C C A T C T C G C A A A , respectively. D N A sequences were determined by the dideoxynucleotide chain termination reaction method [26]. Random mutagenesis Hydroxylamine-HC1 was chosen for introducing random mutagenesis to T N F - a gene. Plasmid p T Z T N F - a was constructed by inserting T N F - a gene between EcoRI site and HindIII site of plasmid pTZ18U. p T Z T N F - a was mutagenized in 250 m M hydroxylamine-HC1 (pH 6.0), 200 m M phosphate buffer (pH 6.0) at 3 7 ° C for 10 min. The reaction was stopped by an addition of 1000-fold vol. of M9 medium supplemented with 12.5 # g / m l carbenicillin. E. coli JM101

was infected with mutagenized pTZTN F-a and infected bacteria were grown in the same medium at 3 7 ° ( to, 12 h. D N A sequencing and an expression of T N F - a muteins were carried out as described above. Results and Discussion

Synthesis and expression of human TNf'-cx ,~ene. Pur!fication of the gene product Both nucleotide and amino acid sequences of the synthetic human T N F - a gene are shown in Fig. 1. Codons used in the chemical synthesis of the T N F - ~ gene were chosen mainly according to the codon usage of E. coli [27]. In selecting codons, homologous or complementary sequences were examined with a computer and codons were altered to avoid mismatches during the aligning process of oligonucleotides. The chemical synthesis and enzymic ligation of the oligonucleotides for human TNF-c~ gene is summarized ill Fig. 2 and was carried out as described in Materials and Methods. The complete T N F - a gene was inserted into plasmid pSP64. D N A was prepared from this plasmid and the sequence was confirmed by the dideoxynucleotide chain termination procedure [26]. The cloned gene was expressed with use of tac promoter and the gene product was purified. Fig. 3 shows SDS-PAGE pattern of whole cell extracts of E. coli harboring phTNF-¢~. A strong extra band of a protein corresponding to TN F-a with a molecular mass of 17 kDa was observed (Fig. 3, lanes 3 and 4). Although the synthetic TN F-c~ gene was designed for high expression in E. coli, the natural human TNF-c~ gene was also expressed to a similar extent of that observed in Fig. 3 lane 3 (data nol shown). A comparison of lanes 3 and 4 indicates that T N F - a expression in LE392/phTNF-c~ was not influenced by IPTG. This was predicted because LE392 didn't have a lacI q gene. Overnight growth of LE392 cells with plasmid phTNF-c~ was identical to those with plasmid pKK223-3, suggesting that the expression of phTNF-c~ didn't influence cell growth (data not shown). H u m a n TNF-c~ was purified from the crude extract of LE392/phTNF-c~ as shown in Fig. 3, lane 5. The specific activity of purified T N F - a against routine L929 cells was 5.5 - 10 v cytotoxic u n i t s / m g and the cytotoxic activity was completely neutralized by an antiserum prepared against human TNF-c~ kindly supplied by Dr. Aggarwal (data not shown). The 24 amino acid residues at N-terminus were completely identical with those of natural human TNF-c~. Although the initiation codon was added to the synthetic TNF-c~ gene, 90% of methionine of the gene product were already removed in E. coli (data not shown). Novel muteins of TNF-e~ and their cytotoxic activities Davis et al. [28] reported that human T N F - a structure contained many /~-sheets but very few c~-helices.

249

1 2 3 4 5

(×1o -3 )

~ 3'1

!i iii!i~ *"" .......

21 ~ 14 ~

~

~

---TNF

......

Fig. 3. SDS-polyacrylamide gel analysis of h u m a n T N F - a synthesized in LE392 carrying plasmid p h T N F - a . Cells (1.107 cells/ml) were grown for 4 h at 37°C, with (lanes 2 and 4) and without (lanes 1 and 3) 1 m M IPTG. The cell pellets were suspended in the sample buffer containing 5% 2-mercaptoethanol, 3% SDS and 65 m M Tris-HC1 (pH 6.8) and heated at 1 0 0 ° C for 3 min. The samples were subjected to 15% SDS-PAGE for 4 h at 150 V, followed by staining with Coomassie brilliant blue R-250. Lanes ] and 2 represent L E 3 9 3 / p K K 2 2 3 - 3 . Lanes 3 and 4 represent L E 3 9 2 / p h T N F - a . Lane 5 represents T N F - a prepared as described in Materials and Methods.

Human T N F - a has 35% and 79% amino acid sequence homology with human TNF-/3 (lymphotoxin) and mouse TNF-a, respectively (Fig. 4) [29,30]. Structure analysis of human T N F - a molecule with the method by Chou and Fasman [31] revealed that there were three/3-turns at positions 69, 99 and 138 (Fig. 5A). These plots were compared with the hydrophobicity plot according to Hopp and Woods [32]. It can be seen from Fig. 5B that these regions containing /{-turn correspond to high hydrophilic regions. This suggests that these regions are exposed on the surface of T N F - a molecule. The homologous regions between human T N F - a and TNF-/~ (for example, 59 to 62 and 93 to 95) are also highly conserved in T N F - a sequences of mouse and rabbit (Fig. 4). The conserved regions are likely to be important for the structure and the function of TNF-a. Recent report by Jones et al. [33] on the X-ray analysis of human T N F - a molecule represents an initial elucidation of T N F - a structure. However, since they didn't describe the residue numbers corresponding to each structure, an exact comparison of the structure analysis mentioned above with their results isn't possible at the present moment. There have been several reports on the relationships between the primary sequence of T N F - a and its biological activities [8-11,15]. Those muteins lacking the amino-terminal regions or having other amino acids in place of two cysteine residues were found to be biologically active. On the other hand, substitutions of His-15 to other amino acids or changes of the carboxy-terminal regions caused loss of the cytotoxicity. One of our

¥ i0 Human Mouse Rabbit Human

TNF- a : TNF- a : TNF-a: TNF-fl :

TNF-

30

40

LTL~SSS~N~SDKPVAHWANHQVEEQLEWLSQRANALLANGM~LI VT~RAI~SDKP~AHWANPQVEGQLQWLSQRANALLANG~L

50 Human

20

60

70

T

80

j

90

= :~DNQLVVPSEGLYLIYSQVLFKGQGCPS[H~---~LTHTISRIAVSY~T~VNLLSAIKS

Mouse TNF- ~ : ~ D N Q L V V P A D G L Y L V Y S Q V L F K G Q G C P D ~ - ~ - - - ~ L T H T V S R F A I S Y ~ E I K V N L L S A V K S Rabbit T N F - a : ~ N Q L V V P A D G L Y L I Y S Q V L F ~ G Q G C ~ S ~ - - - ~ L L T H T V S R F A V S Y ~ K V N L L S A I K S H u m a n TNF-fl :S T I F YSS~KAYSPKATSSPL~PF~LLS~K_~

oo¥

¥

Human T ~ F - ~ :

P----~_~TPEGAE~PWYEPIYLGGVrQLE~GDRLSA~qyLD~AESGQVYrGIIAL

Mouse

pqpK_~TPEGA~L~WYEPIYLGGVFQLEKGDQLSAEVI~I~YLDFAESGQVYFRVIALI

TNF-

a :

R a b b i t TNF- ~: H u m a n TNF-fl :

P_~H~ETP E ~ A ~ P M ~ W Y E P I Y L G G V F Q L E K G D R L S ~ E V N ~ Y L D L A E S V ...... YP~Q P HS F S H-~P

F

GQVYF G i IAL I ST AL

Fig. 4. Comparison of TNF-a and TNF-fl sequences. Conserved amino acids are boxed. The numbers above each row refer to amino acid positions of h u m a n TNF-a. The arrows indicate the positions of site-directed mutagenesis.

250 A 2.0

0.0005 t ,

I

0.0004

0.0003 i,

1.0

H ii

,,

ii

i

ii

,~ -,

tl

ii

,,,,-, it~

÷~

0.0

,,, i

.

l

"~'.

i

",i

,,

,,

i I

J i

"

' i i~

~'i/. ',-

.

0.0002

II

i, ,Ji iii i1! ill

i i i 0.0001

I;,~ It

i Ii

in

i

II

I

iI I

~'' '~

V

I1~

I I

I I

'"

.

I

'

.

"

.

i I

I

I

I

'

ii I

"

.

I

i

50

100

150

50

I O0

150

0.0000

B 2.0-

1.0-

0.0" -I

.0

-2.0 ~ 0

Fig. 5. /~-turn and hydrophobicity plots of human TNF-a. Computer plots were made according to the methods of Chou and Fasman (1978) (A) and Hopp and Woods (1981) (B). The horizontal axes represent sequence position. (A) The right and left vertical axes represent Pt and pt bend values, respectively.Solid and dashed lines correspond to Pt and pt values, respectively.(B) The vertical axis represents hydrophilicity value.

p r i m a r y aims has been to investigate the relationships between the structure a n d the f u n c t i o n of TNF-c~. A n ultimate aim is to prepare a d v a n t a g e o u s m u t e i n s of T N F - a in the clinical applications because T N F - a has been k n o w n to have some severe side effects [34]. However, almost n o t h i n g is k n o w n on f u n c t i o n a l d o m a i n s or receptor b i n d i n g regions of T N F - a . Accordingly, we i n t r o d u c e d some deletions or s u b s t i t u t i o n s t h r o u g h o u t the TNF-c~ molecule. F o r designing several muteins, His-73 a n d G l n - 1 0 2 were chosen because these a m i n o acids are peculiar to h u m a n T N F - a as shown in Fig. 4. The a m i n o acid c o r r e s p o n d i n g to His-73 d o e s n ' t exist in mouse or rabbit TNF-c~. G l n - 1 0 2 of h u m a n T N F - a corresponds to Phe a n d His of m o u s e a n d r a b b i t T N F - a , respectively. As discussed above, these positions c o r r e s p o n d to or n e a r b y /3-turn regions. G e n e s of two novel T N F - a muteins, T N F - M 1 (His-73 deletion) a n d T N F - M 4 ( G i n 102 deletion), were m a d e by site-directed mutagenesis a n d they were expressed in E. coll. Each m u t e i n together with other m u t e i n s ( T N F - M 2 , T N F - M 3 a n d T N F - M 5 ) was purified in a similar fashion to TNF-ct. S D S - P A G E of various m u t e i n s after final purification is shown in Fig. 6. T h e specific activity of each m u t e i n is

indicated in T a b l e I. It shows that His-73 a n d G l n - 1 0 2 are dispensable a n d that these regions are also very flexible. It is noted from Fig. 4 that a m i n o acids are missing in TNF-c~ at positions c o r r e s p o n d i n g to 7 5 - 7 8 of TNF-/3. In addition, a m i n o acids are missing in TNF-/3 positions c o r r e s p o n d i n g to 101 -106 of T N F - a . F r o m these facts, it appeared that further deletion of

1

2

3

4

5

Fig. 6. SDS-PAGE of purified TNF-a muteins, lane l, TNF-MI; lane 2, TNF-M2; lane 3, TNF-M3; lane 4, TNF-M4: lane 5. TNF-M5. Each fraction after the HIC step was used for electrophoresis as described in Fig. 3. The arrow indicates the position of muteins.

251

T A B L E II

to loss of the cytotoxicity [15]. It is noted that all three T N F - a (human, mouse and rabbit) have Glu at the position 116 and human TNF-fl has His at this position. It was designed, therefore, to change G l u - l l 6 to His, which has positive polarity while Glu has negative polarity. TNF-M3 thus prepared, didn't lose the cytotoxic activity (Table I). This result indicates that positive polarity of the amino acid residue at the position 116 doesn't appear to influence greatly the cytotoxic activity. It would be of great interest to examine if TNF-M3 behaves like TNF-fl with respect to side effects, such as lowering blood pressure [37,38]. Such studies are in progress. To effectively isolate many different muteins, random mutagenesis with hydroxylamine-HC1 was used. Human T N F - a gene was treated with 250 mM hydroxylamine-HC1 (pH 6.0) at 37 ° C for 10 min. From the result of D N A sequencing analysis, it was found that these conditions are optimum for introducing single or double point mutations within T N F - a gene and that the G : C to A : T substitution took place as expected. Although the transitions were distributed over many different sites, mutations at positions Pro-70 and Pro106 often occurred and these positions appeared to be a hot spot for transition mutations. DNA sequences of nine novel muteins were determined and they were purified as described in Materials and Methods. The cytotoxic activities of the muteins for mouse L929 cells are shown in Table II. TNF-R3 and TNF-R2 had 4. 10 _4 and 3- 10 -~ times lower activity than TNF-a, respectively. On the other hand, no significant loss of activity was observed with the remaining muteins. These results demonstrate that His-15 and Pro-ll7 are important for the cytotoxic activity of TNF-a. While this manuscript was being prepared, Ashman et al. [39] reported chemical synthesis and expression of a human T N F - a gene. Their approach for the gene assembly and expression of T N F - a gene was similar to our strategy. However, codons were not exactly the same as ours and there was no attempt to produce muteins in their paper. We are now studying anti-tumor activities and side effects of our muteins.

Comparison of the c),totoxic activity of muteins on mouse L929 cell in vitro

Acknowledgements

TABLE I Comparison of the ~ytotoxic activity of muteins on various tumor cell lines in vitro

Mutein

Mutation

Cell line

Specific activity (U/mg)

TNF-c~ TNF-M1 TNF-M2 TNF-M3 TNF-M4 TNF-M5 TNF-M5 TNF-M5 TNF-M5

wild type His-73-deletion Leu-157-Met Glu-l16-His Gln-102-deletion Asp-10-Arg Asp-10-Arg Asp-10-Arg Asp-10-Arg

L929 L929 L929 L929 L929 L929 A375 K562 T24

5.5.10 v 3.4.107 7.2.107 4.6.107 4.6. ] 0 7 2.1.10 v -

amino acids from T N F - a and TNF-fl in these regions may be possible without losing their cytotoxicities. To examine the effect of mutations in conserved regions of TNF-c~ molecules, three muteins, TNF-M5 (Asp-10 to Arg), TNF-M2 (Leu-157 to Met) and TNFM3 (Glu-116 to His) were similarly synthesized. Soma et al. [35] reported that a substitution of basic amino acid residues in the amino-terminal region resulted in novel TNF-c~ muteins which gave better clinical results. TNF-M5, which carries a substitution of Asp-10 to Arg, showed approximately the same specific cytotoxicity as that of TNF-c~. However, it didn't have the cytotoxic activity against human A375 cells, human K562 ceils or human T24 cells (Table I). Further tests are required to determine whether TNF-M5 has a better clinical effect or not. As discussed in the preceding paragraph of this section, mutations around the carboxyl terminal region, in general, have been reported to cause loss of the cytotoxicity [8]. It was, therefore, of interest to change carboxyl terminal Leu to other amino acids. We chose Met because the molecular volume and polarity of Met are close to those of Leu [36]. Such a mutein (TNF-M2) maintained the cytotoxic activity indicating that a modification at the carboxy terminus does not always lead

Mutein

Mutation

Specific activity (U/mg)

TN F-c~ TNF-R1 TNF-R2 TNF-R3 TN F-R4 TNF-R5 TNF-R6 TNF-R7 TNF-R8 TNF-R9

wild type Ser-95-Phe, Gly-153-Val Pro-ll7-Ser His-15-Tyr Arg-44-Cys, Thr-105-Ile Pro-106-Ser, Arg-131-Cys Arg-131-Cys Arg-138-Cys, Ser-147-Phe Pro-70-Leu Pro-106-Ser

5.5-107 2.8.107 1.8.106 2.2.103 8.8.10 v 2.3.10 v 2.5.107 7.5.107 3.5.107 8.6.107

We wish to thank Toshimi Morita and Noriko Matsuyama for D N A sequencing analysis, Yoshiaki Ishima for protein sequencing analysis, Kyoko Yamaguchi for T N F cytotoxicity assay and Masahiro Yamamoto for assistance in preparing the manuscript.

References 1 Carswell, E.A., Old, L.J., Kassel, R.L., Green, S., Fiore, N. and Williasmson, B. (1975) Proc. Natl. Acad. Sci. U.S.A. 72, 3666-3670. 2 Beutler, B. and Cerami, A. (1986) Nature 320. 584-588.

252 3 Sugarman, B.J., Aggarwal, B.B.. Hass, P.E., Figari, I.S., Palladmo. M.A. Jr. and Shepard H.M. (1985) Science 230, 943 -945. 4 Collins, T., Lapierre, L.A., Fiefs, W., Strominger, J.l,. and Pober. J.S. (1986) Proc. Natl. Acad. Sci. U.S.A. 83. 446 450. 5 Perlmutter, D.H., Dinarello, C.A., Punsal, P.I. and Colten, H.R. (1986) J. Clin. Invest. 78, 1349-1354. 6 Aggarwal, B.B., Kohr, W.J., Hass, P.E., Moffat, B., Spencer. S.A., Henzel, W.J., Bringman, T.S., Nedwin, G.E., Goeddel, D.V. and Harkins, R.N. (1985) J. Biol. Chem. 260, 2345 2354. 7 Pennica, D., Nedwin, G.E., Hayflick, J.S., Seeburg, P.H., Derynck, R. Palladino, M.A., Kohr, W.J., Aggarwal, B.B. and Goeddel. D.V. (1984) Nature 312 (1984) 724-729. 8 Korobko, V.G., Dobrynin, V.N., Shingarova, L.N., Filippov, S.A.. Maksimova, Y.N., Chumakov, A.M. and Davydov, I.V. (1989) Protein engineering of tumor necrosis factors. The proceedings of the second international conference on tumor necrosis factor and related cytokines, Napa, CA, 1989. 9 Tsujimoto, M., Tanaka, S., Sakuragawa, Y., Tsuruoka, N., Funakoshi, K., Butsugan, T., Nakazato, H., Nishihara, T., Noguchi, T. and Vilcek, J. (1987) J. Biochem. 101,919-925. 10 Creasey, A., Doyle, L.V., Reynolds, M.T., Jung, T., Lin. L.S. and Vitt, C.R. (1987) Cancer Res. 47, 145-149. 11 Narachi, M.A., Davis, J.M., Hsu, Y.R. and Arakawa, T. (1987) J. Biol. Chem. 262, 13107 13110. 12 Yamamoto, R., Wang, A., Vitt, C.R. and Lin, L.S. (1989) Protein Eng. 2, 553 558. 13 Engler, D.A., Matsunami, R.K., Campion, S.R.. Stringer, C.D., Stevens, A. and Niyogi, S.K. (1988) J. Biol. Chem. 263, 1238412390. 14 lkehara, M., Ohtsuka, E., Tokunaga, T., Taniyama, Y., lwai, S., Kitano, K., Miyamoto, S., Ohgi, T., Sakuragawa, Y., Fujiyama, K. lkari,T., Kobayashi, M., Miyake, T. Shibahara, S., Ono, A., Ueda, T., Tanaka, T., Baba, H., Miki, T., Sakurai, A., Oishi, T., Chisaka, O. and Matsubara, K. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 5956 5960. 15 Kamijo, R., Takeda, K., Nagumo, M., Konno, K., Hasegawa, A., Inaka, K. and Ikehara, M. (1989) Biochem. Biophys. Res. Commun. 160, 820-827. 16 Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) in Molecular Cloning: a laboratory manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 17 Melton, D.A., Krieg P.A., Rebagliati, M.R., Maniatis, T., Zinn, K. and Green, M.R. (1984) Nucleic Acids Res. (1984) 12. 7035-7056.

18 Brosius, .I. and t l o l x . A. (1984) Proc. N;~tl

19 20 21

22

23 24 25 26 27 28 29 30 31 32 33 34

35

36 37 38 39

\ c a d .~,.i 1 .S A

>.1

6929 6933. Tabor, S. and Richardson, C.('. (1985) lh,~c. \ a l l . . . \ c a d . Sck I'.S.A. 82, 1074 11)78. Beaucage, S.I,. and ('aruthers. M . H (1981) letrahedron l,ctt. 22. 1859 1862. Hanahan, D. (1985) in Techniques for transl'ormation of E. cM,' m D N A Cloning I (Glover, D.M., ed.), pp. 109 135 IRL Press, Oxford, U.K. Lin, L.S. and Yamamoto, R. (1987) in Purification of native and recombinant tumor necrosis factor. Protein Purification: Micro t,~ Macro. Alan R. Liss. Inc. pp. 409-419. Laemmli, U.K. (1970) Nature 227, 680 685. Mosmann, T. (1983).I. lmmunol. Methods 65, 55-63. Kunkel, T.A. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 488-492. Sanger, F., Nicklen, S. and Coulson, A.R. (1977) Proc. Natl. Acad. U.S.A. 74, 5463-5467. Gouy, M. and Gautier, C. (1982) Nucleic Acids Res. I(1, 7055 7074. Davis, J.M., Narachi, M.A., Alton. N.K. and Arakawa, T. (1987) Biochem. 26, 1322 - 1326. Gray, P.W. (1987) Lymphokines 13, 199 208. Pennica, D. and Goeddel, D.V. (1987) Lymphokines 13, 163 t80. Chou, P.Y. and Fasman, G.D. (1978) Annu. Rev. Biochem. 47, 251 276. Hopp, T.P. and Woods, K.R. (1981) Proc. Natl. Acad. Sci. LJ.S.A. 78, 3824-3828. Jones. E.Y., Stuart, D.l. and Walker, N.P.C. (1989) Nature 338, 225-228. Oliff, A., Jones, D.D., Boyer, M., Martinez, D., Kiefer, D., Vuocolo, G.. Wolfe, A. and Socher, S.H. (1987) Tumors secreting human T N F / c a c h e c t i n induce Cachexia in mice. Cell 50, 555-563. Soma, G., Tsuji. Y.. Tanabe, Y., Noguchi, K., Tanabe, N.K., Gatanaga, T., lnagawa, H., Kawakami, M. and Mizuno, D. 11988) J. Biol. Response Mod. 7, 587-595. Miyata, T., Miyazawa, S. and Yasunaga, T. (1979) J. Mol. f(vol. 12, 219 236. Kreil, E.A., Greene, E., Fitzgibbon, C . Robinson, I).R. and Zapol, W.M. (1989) Circulation Res. 65, 502-514. Ulich, T.R., Castillo J., Keys, M. and Granger G.A. (1987) Amer. J. Path. 128. 5 12. A s h m a n , K., Matthews, N. and Frank. R.W. (1989) Protein Engineering 2, 387 391.