Secretory synthesis of human interleukin-2 by Streptomyces lividans

Geae. 86 (1990) 227-232 Elsevier

227

GENE 03371

Secretory synthesis of humnn interleukin-2 by Streptomyces lividaas (Recombinant DIqA; tendamistat; lymphokine; signal peptide; Gram-positive; Streptomyces tendae)

Eekhard Bender", Klaus-Peter Koner ~' nnd Joachim W. Engels" ° InsHtut~r Organische Chemic. Um'versit~t Frankfurt, D-6000 FrankfunlMain 50 (F.R.G.) Tel. 06915800-9150 and b H o e c ~ AG, D-6230 Frankfurt~Main 80 (F.R.G.) Tel. 069/305-3650 Received by K.F. Chater: 16 April 1989 Revised: 18 July 1989 Accepted: 19 July 1989

SUMMARY

To study the ability of Streptomyces lividansto produce heterologous proteins by secretion, we directly fused DNA encoding the leader peptide of the 0e-amylaseinhibitor, tendamistat, produced by Stmptomyces tendae, with DNA encoding the mature part of interleukin-2 (IL-2). Such cloned fusion constructs are translated in S. iividans, in spite of the quite different codon usage. The active 11-2is secreted into the culture broth, though the amounts are much less than that of the 0t-amylaseinhibitor. The presence oflL-2 in the supernatants could be demonstrated both by an activity assay and by immunoblotting. In addition to the secreted form, three different species of II-2 antibody immunoreactive proteins, with different Mrs, are either present in the cells or attached to the cells. This indicates that inefficient processing and translocation of the precursor is a major reason for the low activities found in the supernatant.

INTRODUt,WION

Streptomyces spp. are useful producers of antibiotics, but have not until now received much attention as an expression system for proteins. However, there are excellent arguments for investigating their potential in that direction. As soil bacteria, most ofthe Streptomycetes are usually unable to colonize in man. Therefore, they appear to be safe Correspondence to: Dr, J.W. Engels, lnstitut fdr Orgnnische Chemie, UniversitAt Frankfurt, D.6000 Frankfurt/Main 50 (F.R.G.) Tel. 069/5800-9146; Fax 069/5800-9494. Abbreviations: aa, amino acid(s); AI, 0~-amylaseinhibitor (tendamistat); AI, gene encoding AI; Ap, ampicillin, aph, aminogiycoside phosphotransferase; bp, base pair(s); BSA, bovine serum albumin; CP-linker, Clal.Pmll.linker; CTLL, cytotoxic mouse T-cell line; Is, immunoglobulin; IL-2, interleukin-2; lacZ, p-D-galactosidase; m.p., mature part; nt, nucleotide(s); oligo, oligodeoxyribonucleotide; or/, origin of DNA replication; PAGE, polyacrylamide-gelelectrophoresis; PB S, phosphatebuffered saline (see Fig. 4, legend); s, signal peptide; SDS, sodium dodecyl sulfate; 7',terminator; TCA, trichloracetic acid; Th, thiostrepton; u, unit(s); YEME and YPG, 2YT, see Fig. I, legend. 0378-1! 19/90/$03.50 © 1990Elsevier Science Publishers B.V.(BiomedicalDivision)

organisms for the synthesis ofmedically important proteins. As Gram + microorganisms they are attractive candidates for the secretory expression of peptides, especially because many of them, due to their sedentary way of life, naturally secrete large amounts of proteins into their environment. One of these is the 0c-amylaseinhibitor tendamistat (AI), a peptide of 74 aa, secreted by S. tendae(Vertesy et al., 1984). After cloning the gene, it was successfully expressed by S. lividans (Koller, 1986) and the nt sequence revealed a typical leader peptide at the N terminus (K.-P. K., unpublished). This presented a suitable Streptomycetes secretion system for our investigation. As the heterologous part in our experiment we used human IL-2, a lymphokine naturally secreted by T-cells. It acts via a specific receptor interaction stimulating growth of T-cells (Clark et al., 1984; Smith, 1988; Taniguchi et al., 1983) and appears to be a promising candidate for treating melanomas (Rosenberg et al., 1988). To our knowledge, there is until now only one report about secretory expression oflL-2 by bacteria (Miwa et al., 1987). Protein secretion is known to be mediated by signal peptides (Blobel and Dobberstein, 1975), whereby the aa

228

sequence following the leader must be compatible with translocation through the membrane (Moreno et al., 1980). Furthermore, the interaction of leader and mature part seems to play an important role (Himeno et al., 1986). Until now, however, we do not know the rules governing the successful combination of homologous leader and heterologous protein, although several reports about heterologous, secretory protein expression by Gram ÷ bacteria are available (Chang, 1987; Schein et at., 1986).

-EcoRt ~mau-D=tql largefragment

LI

L2

II

i.

Iigatlon l

420.bp Ilaetl.J:U.odlll fragment[

Sph[

•

We have fused the last aa of the AI leader directly onto the first of IL-2. Comparing the secretion of two different, naturally secreted proteins linked to the same Streptomyces leader sequence should tell us something about the specific mechanism governing this transport. At any rate the proposed construct should yield, if correctly processed, an authentic peptide, comparable with the natural molecule.

p~l Snbt., ~ R I small fragment

~d/LL~ fusion J

RESULTS AND DISCUSSION fJal

pTIet E~oRZ

~d/u,:z fusion~

pBCI

Lt~ll

]..J CP.llnker

Exull ~ R ~ I f,lal fkehl

llILEII F~IRI

large fragment ~

..... pU680 Clnl.~.v.ull + CP.llnker large fragment Clal.P~ull

Fig, I. Cloning scheme leading to 11,-2/,41 fusion construct pT! ! and to its expression vector pTIEI in $treptomyces.The £. cell strain used for cloning was BMH7902 (hsdM, 4sdR, lucZ Y, leu, pro, s ~ , thr, [F'traD3'6, lad e, lacZdMl$, pro*), For cloning and expression studies with $~ptomyces we used $. II~dumTK24 (Hopwood et at., 1985).The E. coil plasmid used for constructions were a pUCi2 derivative carrying a synthetic gene based on the £. co//codon usage for the human 11.-2 (Engels et at., 1985), a pUCI8 derivative with the gene of the z-amylase inhibitor tendamistat (K.-P. K., unpublished) and a pUCI8 plasmid carrying the CP-linker (Clai, Pmll)(see Fig. 2). The starting $treptemyces plasmid used for our expression construct was plJ680 (Kieser, T,, unpubfished; Hopwood et at,, 1985).E. coilBMH7902 was Brownin 2YT (I.6~ tryptone/l.0% yeast extract/0,5% NaC! (all ~ in w/v), pH 7.3) liquid medium or agar, supplemented with 50 pg/ml ampicillin when needed. Culturing of 8~ptomyces for genetic manipulations was done according to Hopwood et at. (1985). Shaking cultures for expression studies were inoculated i:200 from 10-12-day-old YEME (Hopwood et at., 1985) precultures; the medium used was YPG (Z$~o starch/5~o 81ucose/0.6~o peptone/0,5% yeast extract/0.Syo 81ycine/0.2% (NH~)2SO4/0.05% NaC1/0.056% MgCla, all % in w/v).The medium was phosphate-buffered to pH 7.3 and supplemented with I pg/ml of neomycin. Cultures were

(a) Plasmid constructions With the help of two linkers (LI and L2, Fig. 2), a synthetic gene encoding IL-2, based on the Esche~chia coli codon usage, was iigated into the Al gene. The result is a direct fusion between the C-terminal aa of the Al-leader peptide and the N-terminal aa of the mature part of IL-2, whose two consecutive stop codons terminate translation. For the expression in 8tmptomyces, the fusion was placed into the polylinker of pBCI (Fig. 2), a plJ680 derivative, upstream From a chemically synthesized aph terminator

Brownwith rotary shaking of 300 rev/min at 28°C. Genetic manipulations of $treptomyceswere done according to Hopwood et al. (1985), except for plasmid minipreparations which were performed as described by MacNeil (1986). General procedures of DNA manipulations and E. coil molecular genetics were performed as described in Maniatis et el. (1982). All constructions with £. ~ plasmids were verified by the dideoxy chain.termination method (Sanger et at., 1980), $~eptomyces plasmids were analysed by restriction enzyme analysis of minipreparation plasmid DNA. All restriction enzymes and ligese were obtained from New England Biolabs, Klanow fragment of £. cell DNA polymerase ! was obtained from Boehringer-Mannheim. All enzymes were used according to the manufacturer's recommendations. First a silent leader mutant of the A! gene was generated, which carries an Apal site at the Y-and of the DN& coding for the signal peptide (pUCI8/AI^p.) (E.B., unpublished). The mutation was introduced by the gapped duplex method of Kramer et at. (1984). The ApaI-B$tEll large fragment of pUCI8/AI^p, was ligeted with the help of the two linkers LI and L2 (Fig. 2) to the 420-bp Haell.Htadlll fragment of pUCI2/IL.2. For expression in $~ptomyces we made use of plasmid pBCI, which was generated by ligation of the Pvull-C/al large fragment from plJ680 and the ll$-bp CP-linker (Fig. 2), which consists of a polySnker at the Y-end of a chemically synthesized aph terminator. Plasmid pBCI was opened by an EcoRI-Sphl cut and iigated to the EcoRl.$phl fragment ofpTl I, yielding pTIEI.

229 eA Q

¢

A-T G-e G-C '~"G ¢-G T-A T-A ~-G ¢-G ¢-G G-C G-¢ G A-T C-G G-¢ ¢-G ¢-G

Se

~~'TGAO{."T

'

t d , ' m nlxzx

IUIPz ' ' p ~ x x e.xaz

nl o.v •I +o,x<-- m m - t . r . t n . t o ,

••• N

n

V

1

A

L

R

L

SnhX

A

A

L

V

G

g

G

A

A

L

A

, Ixz

---~

L

8

z

&GGQD.~"4~'~G P L A A

~X-8

~ m • • • mmam~mam,mwHunm~mluDmem.mm,m~manblm~lHum~memm~mannmiDm~mmm, mmmmiquunBmam

~ |

~

euml~mmm~lmo~mulomunqumman~ml*

!..¢TO&GQ'Jm&T&O'JL'~m'~'¢TGQ&~~~'OTOI~O~L'Q@

•

.

..

L T

stop

.::

I

stop

~,c . . . . . . . .

...--

zz~

gene

.............

$'

--------'U--

IIAG4xxx ..............

3,

I,

ElSmxx c-

s,

naeax ax

.....

--

...-I,

Fig. 2. Representation of the junctions between the Ai and the IL.2 genes and of the CP-linker sequence in which the fusion construct was inserted. Oligos were synthesized by solid-phase cyanoethyi-phosphoramidite chemistry (Sinha et al., 1983)with an Applied Biosystems DNA Synthesizer 380A. The resulting crude product was purified by electrophoresis on a 12Yodenaturing (6 M urea) polyacrylamide gel. The band corresponding to the correct molecular weight was cut out, and the oligns were eluted and finally desalted by Sephadex (Pharmacia) G$0 and Biogel P6 chromatography. The upper part shows the sequence of the CP-iinker (Clal, Pmll) with the spa terminator, the latter drawn as stem-loop structure. The stop codon was included to avoid readthrough translation from the Th R gene. The £coR! and H/ndlll sites used for the insertion are marked. Below, the nt and 8a sequences at the junctions between the AI and the IL-2 genes are shown. The two linkers LI and 1.2 used for the fusion are in hatched boxes. The restriction sites at the ends oftbe gene fragments are indicated, with exception of the Apal at the 5' and the Haell at the 3' end of Ll, because they were destroyed by the design of LI, as to restore the original nt sequence. The arrow indicates the fusion point between the AI signal peptide and 11.-2.

(Thompson and Gray, 1983). The construction followed the work of Pulido and Jimenez (1987) and the corresponding cloning scheme is shown in Fig. 1. (b) Detection of IL-2 by the activity assay With the help of the activity assay for IL-2 it is possible to detect and to quantify IL-2 in respect to its biological activity, which should be a measure for its correct structure.

Although elevated levels of DNA synthesis of cells (CTLL) could be detected by stimulation with supernatant from a two day culture of $. lividans containing pTIEI, significant amounts were seen first at the third day of incubation. On the fourth day there was a maximal level of activity, dropping during ~further cultivation. The A1 gene had the same onset of production, but the activity was not reduced until the eighth day of cultivation (Fig. 3). The decline of activity for the lymphokine during the later growth phase

230 analysis from different fractions ofthe culture. The immuneblots from culture supernatants of IL-2-secreting Streptomyces cells (Fig. 4) showed that the protein responsible for

IL-2 lu/ml) 1000. 800 600 +,00

lq

200

64

t

4

0:

- 20

Al

I,uglmll 200,

- 16

.a

150 a

100, 50,

2

&

6 - 8 " 16time(days)

Fig. 3. Production kinetics of IL-2 (panel a) and A! (panel b). $. IlvaPdans clone transformed with pTIEI (Tpxv~t),served as source of IL-2, which was assayed in supernatants by measuring IL-2 activity with cells from a C'rLL according to Gillis etal. (1978). (Panel b)Time-dependent synthesis of Ai, cloned into the same vector and the same host strain 5. livldans TK24. The individual symbols represent the results of independent experiments, whose mean values (x) are connected by a solid line. A! was assayed according to Bergmeyer (1970). may have been due to proteolytic breakdown, which was not counterbalanced by secretion in contrast to the steadystate activity of the proteinase-stable AI. Assuming, based on the results of Weigel et al. (1989), that I mg of IL-2 corresponds to 2 × 107 u, the maximal yields in the supernatants were about 30 ng/ml, this is less than one thousandth of the homologous gene product secreted by the control strain. There could be several reasons for this difference: for instance, poor translation of the E. coil codons which are rarely used in Streptomyces whose D N A is characterized by a high G + C-content (Benigni et al., 1975) or difficulties with translocation or processing of the precursor. The last hypothesis is supported by the result that disrupted cells from I ml of a four-day culture, whose secreted IL-2 activity was 7.1 × 102 u/ml, contained 47 u/ml IL-2. (Cell disruption was done as described in the legend to Fig. 4, except that the cells were resuspended in phosphate buffer before ultrasonic treatment). (¢) A n a l y s i s o f IL-2 activity by immunoblots

To get some insight into the secretory pathway of our fusion construct in S. iividans, we made an immunoblot

b

c

d

•

f

g

h

i

j

k

Fig. 4. lmmunoblot analysis of 11.-2 and derivatives. All samples were dissolved in reducing loading buffer, unless indicated otherwise. Aliquots of the culture were taken from the fourth day of fermentation. Lanes: a, disrupted cells equivalent to 50 #i of TpTIE t culture; h, disrupted cells equivalent to 50 #1 of TK24 culture mixed with prestained protein Mr standards (BRL); c, again TK24, but mixed with I00 ng E. coil derived IL.2; d, disrupted cells equivalent to $0/d of TK24 culture (control); e, insoluble particles from 50/d of TpTIB t culture supernatant in nonreducing loading buffer; f, insoluble particles from 50/410fTpTP~tculture supernatant; g, insoluble varticles from 50+ui of TK24 culture supernatant (control); h, TCA precipitated soluble proteins of 1250#! of TK24 culture supernatant (control); i, TCA precipitated soluble proteins of 1250 ~1of Tpxtt+,culture supernatant; j, TCA-precipitated control, mixed with prestained marker proteins; k, TCA-precipitated control, mixed with 100 ng of £. ¢oli derived 11.-2. Cells destined for disruption were collected by 5 min centriFugation at 30 )
231

the activity had the same Mr as the reference IL-2 derived from E. coli. This indicated, in agreement with the activity assays, the presence of secreted and processed 11.-2. We also analyzed cells disrupted by sonication, where we found essentially three different immunoreactive protein bands. The protein with the lowest Mr correlated well with the processed molecule in the TCA-precipitated supematant, the largest protein showed a relative mobility close to the unprocessed preprotein consisting of the Al-leader and the mature part of 11.-2. Surprisingly, there was still another band between these two, which seemed to have about l0 additional aa in comparison to mature IL-2, whose cleavage by signal peptidase is difficult to explain with our current models. Taken together these three species amounted to about 20-fold the quantity of the secreted IL-2, that is approx. 0.5/tg per cell of I ml of culture as judged by the immunoblots. The presence of processed IL-2 attached to the cells indicates problems with the translocation into the culture medium, and the appearance of wrongly processed or unprocessed precursors suggests that the fusion point is perhaps not the optimal substrate for the signal peptidase. Following secondary structure predictions (Chou and Fasman, 1978), we argue that our fusion construct has a decline in the probability for a/hum at the junction of leader and mature part. Recent observations of other groups (Andrews et al., 1988; Duffaud et al., 1988), suggest some impact of the/I turn on processing by the signal peptidase. On the other hand it may also be that bad translocation through the membrane is responsible for the pi"oblems occurring with the signal-peptide cleavage. There was also a relatively large amount of these three proteins in the insoluble particles ofthe culture supernatant, which were cleared from cells as described in the legend to Fig. 4. These proteins show apparently no intermolecular disulfide bridges, because the mobility of the protein was not influenced by the addition of reducing agents to the loading buffer, therefore we think that they are not derived from inclusion bodies. Judged from the banding pattern of the silver stained gels after electroblotting, the insoluble material consists ofintraceilular proteins (data not shown). We found that the amounts of these three proteins were higher in 50/Al supematant than in washed cells corresponding to 50 ~tl shaking culture (Fig. 4).

cell lysis. Although a Western blot is only a half-quantitative method, it is obvious that there is a good agreement between the determined activities and the amount of immunereactive, processed 11..-2 in the supernatant. We therefore conclude that the secreted 11-2has the correct,biologically active tertiarystructure. (2) The presence of the cellular IL-2 species demonstrates that the precursor is not a good substrate for the translocation machinery. Interestingly,Chang and Chang (I988) also found three bands, corresponding to the precursor, the mature part and a -4 form in theirimmunoblots of cellular proteins, during the analysis of tumor-necrosisfactor secretionby $. h'vidans.They interpretedtheirresults as consecutive steps on the secretory pathway, including a compartment for the final release of the protein on the outside of the cell.In contrast, we would explain our IL-2 species of intermediate Mr as an incompletely processed precursor, which is unable to traverse the membrane, whereas the mature part is secreted, but at a reduced rate making it detectable in the cellularfraction. (3) The insoluble particlesin the supematant are interpreted by us as products of cell lysis:the relativelylarge amount of the three 11.-2 species may reflecttheir attachment to the membrane, making them amenable for precipitation. (4) Poor translation or m R N A instabilityseems not to

be the main reason for the poor yields in the supernatants, because there is about 0.5 ~g/ml of immunoreactive material inside the cells. (5) During the last few years other groups have reported intracellular (Gray et al., 1984; Mufloz et al., 1985; Pulido and Jimenez, 1987) or, more recently, secretory (Lichtenstein et al., 1988; Noack et al., 1988) heterologous protein production by Streptomyces too. Further studies will show whether Streptomyces is becoming a valuable organism for the production of medically useful proteins. Therefore we are now making changes of the leader and the mature part, to see whether the IL-2 has a negative influence on transport or whether some necessary information from the sequence of the original gene is missing.

ACKNOWLEDGEMENTS

(d) Conclusions (1) The processing of the IL-2 precursor indicates that the observed activity is the result of secretion. Further, the parallel onset of production for the homologous and the heterologous proteins, together with the absence of the two species of higher Mr from the soluble proteins in the supernatant, indicates that the processed IL-2 is not released by

We wish to thank Dr. D.A. Hopwood for kindly providing us with the necessary Streptomyces plasmids and strains, Dr. R. Obermeier (Hoechst AG) for IL-2 activity assays, M. Wimmer for recombinant IL-2 (E. col~), Dr. Kurrle (Behring-Werke) for kindly sending us the monoclonal IL-2 antibody. The dedicated technical assistance of I. Claes, S. Scholz and R. Vogel is gratefully acknowledged.

232 REFERENCES Andrews, D.W., Perara, E., Lesser, C. and Lingnppa., V.R.: Sequeaces beyond the cleavage site influence signal peptide function. J. Biol. Chem. 263 (1988) 15791-15798. Benigni,R., Antonov, R.P. and Carere, L.A.:Estimate of the 8enome size by maturation studies in $~ptomyces. Appl. Microbiol. 30 (1975) 324-326. Bergmeyer,H.U.: Methoden der enzymatischen Analyse.2nd ed., Vol. !. Verla8 Chemie, Weinheim, 1970 pp. 394-395. Blobel, G. and Dobberstein, B.: Transfer of proteins across membranes, I. Presence of proteolytically processed and unprocessed nascent immunoglobulin fight chains on membrane-bound ribosomes of routine myeloma. I. Ced Biol. 67 (1975) 835-851. Chang, S.: Engineering for protein secretion in Gram-positive bacteria. Methods Enzymol. 153D (1987) 507-516. Chang, S.Y. and Chang, S.: Secretion of heterolngous proteins in Streptomyces li~dans. In Okami, Y., Beppu, T. and Ognwara, H. (Eds.), Biology of Actinomycetes 88, prec. 7th Int. Symp. on the Biology of Actinomycetes. Japan Sci. Soc. Press, Tokyo, 1988, pp. 103-107. Chou, P.Y. and Fasman, G.D.: Empirical predictions of protein conformation. Annu. Rev. Biochem. 47 (1978) 251-276. Clark, S.C., Arya, S.K., Wong-Staal, F., Matsumoto-Kobayashi, M., Kay, R.M., Kaufmann, R.J., Brown, E.L., Shoer~aker, C., Copeland, T., Oroszlan, S., Smith, K., Sarngadharan, M.G., Lindner, S.G. and Gallo, R.C.: Human T-cell growth factor: partial amino acid sequence, eDNA cloning and organization and expression in normal and leukemiccells. Prec. Natl. Acad. Sci. USA 81 (1984) 2543-2547. Duffaud, G. and lnouye, M,: Signal peptidases recognize a structural feature at the cleavage site of secretory proteins. J. Biol. Chem, 263 (1988) 10224-10228. Engels, J.W., Uhlmann, E., Wengenmeyer, F., M011ner, H., Winnacker, E.L., Mertz, R. and Okazaki, H.: Patent DR 3419995 (1985). Oillis, S., Ferm, M., Ou, W. and Smith, K.A.: T-cell growth factor: parameters of preduction and a quantitative microassay for activity. J, lmmunol. 120 (1978) 2027-2032. Gray, O., Seizer, O., Buell, G., Shaw, P., Esconez, S., Hofer, S., Voegli, P, and Thompson, CJ.: Synthesis of bovine growth hormone by $~ptemyces lJVldans,Oene 32 (1984) 21-30. Himeno, T., Imanaka, T. and Aiba, S,: Protein secretion in BacillussubHlis as influencedby the combination of signal sequence and the following mature part. FEMS Microbiol, Lett. 35 (1986) 17-21. Hopwood, D.A., Bibb, MJ., Chater, K,F., Kieser, T., Thompson, C.J., Kieser, H.M,, Lydiate, D,J. and Schrempf, H.: Genetic Manipulation of 8~ptomyces: a Laboratory Manual, The John Innes Foundation, Norwich, U.K., 1985, Keller, K.P,: Cloning and expression of the a-amylase gene from $~ptomyces tendae. In Szabo, G., Biro, S. and Goodfellow, M, (Eds.), Biological, Biochemical and Biomedical Aspects of Actinomycetes, Part A. Academiai Kiado, Budapest, 1986, p. 153. Kramer, W., Drutsa, V., Jansen, H.W., Kramer, B., Pflugfelder, M. and Fritz, H.J.: The gapped duplex approach to oli8onucleotide-directed mutation construction, Nucleic Acids Res, 12 (1984) 9441-9456. Laemmli, U.K.: Cleavage of structural proteins during assembly of bacteriophage T4. Nature 227 (1970) 680-685. Lichtenstein, H., Brawner, M.E., Miles, M., Meyers, C.A., Young, P.R., Simon, P.L and Eckhurdt, T.: Secretion of Interleukin.IB and Escherlc~a cell 8alactokinase by Streptomyces li~dans. J. Bacteriol. 170 (1988) 3924-3929.

MacNeil, DJ.: A flexible boiling procedure for isolating plasmid DNA from Gram-positive microorganism. J. Microbiol. Methods 5 (1986) 115-123. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Miwa, T., Kobayashi, M., Matsui, Y., Ei, H. and Yamane, K.: Manufacture of 11-2by genetically engineered Bacillus. Patent JP 62/3796 A2 (8713796) (1987). Moreno, F., Fowler,A.V.,Hall, M., Silhavy,TJ., Zabin, J. and Schwartz, M.: A signal sequence is not sufficient to lead p-galactosidase out of the cytoplasm. Nature 286 (1980) 356-359. Muftoz, A., Perez-Aranda, A. and Barbero, J.L: Cloning and expression of human interleukin-2in Sn'eptomycesli~dansusing the Eschen'cAfa coil consensus promoter. Biochem. Biophys. Res. Commun. 133 (1985) 511-519. Noack, D., Geuther, R., Tonew, M., Breitling,R. and Behnke, D.: Expression and secretion of interferon-al by Slreptomyce$ Iividans: use of staphylokinase signals and amplification of the nee gene. Gene 68 (1988) 53-62. Pulido, D. and Jimenez, A.: Optimization of gene expression in Streptomyces lividans by a transcription terminator. Nucleic Acids Res. 15 (1987) 4227-4240. Rosenberg, S.A.,Packard, B.S.,Aebersold, P.M., Solomon, D., Topalian, S.L., Toy, S.T., Simon, P., Lotze, M.T., Yang, J.C., Seipp, C.A., Schwartzentruber, D., Wei, J.P. and White, D.E.: Use of tumor. infiltrating iymphocytes and interleukin-2 in the immunotherapy of patients with metastatic melanoma. N. Engl. J. Med. 319 (25) (1988) 1676-1680. ganger, F., Coulson, A.R,, Barnell, B.G., Smith, A.J.H. and Roe, B,A,: Single stranded bacteriophage as an aid to rapid DNA sequencing. J. Mol. Biol. 143 (1980) 161-178. Schein, C.H., Kashiwagi, K., Fujisawa, A. and Weissmann, C.: Secretion of mature IFN-a2 and accumulation of uncleaved precursor by Bacflus subt#e transformed with a hybrid a.amylase signai sequence. IFN-ot2 pne, Blotechnolagy 8 (1986) 719-725. Sinha, N.D., Blernat, J. and K0ster, H.: #.CyanoethylN,N.diaikylamino/ N-morphollno-monochloro phosphoamidites, new phosphorylating agerts facilitating ease of deprotection and work-up of synthesized oli8onucleotides. Tetrahedron Lett. 24 (1983) $843-$846. Smith, K.A.: lntarleukin-2: inception, impact and implications. Science 240 (1988) 1169-1176. Taniguchi, T., Matsui, H., Fujita, 1"., Takaoka, C., Kashima, N., Yoshimoto, R, and Hamuro, J.: Structure and expression of a cloned eDNA for human interleukin.2. Nature 302 (1983) 305-310. Thompson, CJ. and Gray, G.S.: Nucleotide sequence ofa streptomycete aminoglycosidephosphotransferase 8ene and its relationship to phosphotransferases encoded by resistance plasmids. Prec. Natl. Acad. Sci. USA 80 (1983) $190-$194. Towbin, H., Staehelin, T. and Gordon, J.: Electrophoretic transfer or p~oteins from polyacrylamidegels to nitrocellulose sheets: procedure and some applications. Prec. Nail, Acad. Sci. USA 76 (1979) 4350-43M, Vertesy, L., Oeding, V., Bender, R., Zepf, K, and Nesemann, G.: Tend~stat (HOFA67), a tight.binding a-amylase inhibitor from Streptomyces tendae 41S8. Bur. J. Biochem. 141 (1984) 505-512. Weigel, U., Meyer,M. and Sebald, W.: Mutant proteins of human inter. leukin 2. Bur. J. Biochem. 180 (1989) 295-300.