Structure-Function Mapping of Interleukin 1 Precursors

‘ r H E vJOURNA1. OF 8

~

Val. 266, No. 11, Issue of April 16, pp. 7081-7086, 1991 Printed in U.S . A.

BIOLOGICAL CHEMISTRY

1991 by The American Society for Biochemistry and Molecular Biology, Inc. ’

Structure-Function Mapping of Interleukin 1 Precursors CLEAVAGE LEADS T O A CONFORMATIONALCHANGEIN

T H E MATUREPROTEIN* (Received for publication, November 16, 1990)

Daria J. Hazuda#$, James Stricklery, Philip SimonII**,and Peter R. Young$ $$ From the Departments of $Molecular Genetics, TMacromolecular Sciences, and 11 Immunology, SmithKline Heecham Pharmaceuticals, Kingof Prussia, Pennsylvania 19406-0939

The two interleukin 1 (IL-1)genes (IL-la and B) encode 31-kDa precursor molecules, which are cleaved upon secretion to generate the mature, active, carboxyl-terminal 17-kDa proteins. The IL-1B precursor is inactive, whereas the IL-la precursor is asactive as the mature IL-la. In this report, we demonstrate that when either of the recombinant precursors is processed to the mature form, the mature region undergoes a conformational change from a proteinase K-sensitive structure to one that is proteinase K-insensitive. In addition, cysteine residues that are exposed to solvent in the IL-la precursor become buried in the mature protein. Limited structure-activity mapping ofthe IL18 precursor indicates that the amino-terminal 76 residues are responsible for the conformational change, whereas the most dramatic change in biological activity occurs after further removal of residues 77-94. These findings suggest that the altered structure of the mature region in precursor IL-1s has been conserved for some function. Denaturation/renaturation experiments implicate the precursor domain in protein folding, and byanalogy with signal-directedsecretory proteins, the unique conformation of the precursors may play a role in IL-1 secretion.

cytesandarethoughtto playa key role in immuneand inflammatory responses due to a multitude of activities involving many different targetcells. Included among these are the stimulation of T cells, B cells, neutrophils, and immature immune cells, as well as stimulation of acute phase protein synthesis in hepatocytes and release of proteases and prostaglandins from fibroblasts and epithelial cells (see Oppenheim et al., 1986 for review).These activities are mediated through binding to a specific receptor on the target cell (Dower and Urdal, 1987, and references therein). Both a and p forms of IL-1 are synthesized and accumulate as intracellular 31-kDa precursor proteins (Giri et al., 1985; Limjuco et al., 1986; Auron et al., 1987; Bomford et al., 1987; Young et al., 1988a; Hazuda et al., 1988). The IL-1 precursor proteins lack a hydrophobic leader sequence to direct their secretion through the endoplasmic reticulum (Lomedico et al., 1984; Auron et al., 1984; March et al., 1985; Young and Sylvester, 1989).Moreover,pulse-chase studies (Hazuda et al., 1988), electron microscopy (Singer et al., 1988), in vitro translation/translocation experiments (Giri et al., 1985; Suttles et al., 1990), transfection and fractionation studies (Young et al., 198813; Fuhlbrigge et al., 1988), and theobserved lack of glycosylation despite the presence of potential sites within the two IL-1 sequences (Baldari et al., 1987; Livi et al., 1990)? substantiate that IL-1 is not secreted through the established secretory apparatus (endoplasmic reticulum and Golgi). ProcA number of gene products existwhich act outsidecells and essing to the mature 17-kDa carboxyl-terminal form occurs to a requirea hydrophobic concomitantly with secretion and appears not be yetaresynthesizedintracellularlywithout signal sequence for secretion. Included among these are IL-1’ ment for export, since both precursor and mature forms are ( a and p ) (Auron et al., 1984; Lomedico et al., 1984; March et released (Hazuda et al., 1988; Suttles et al., 1990). The fact that IL-1 can be secreted asa precursor suggests al., 1985), fibroblast growth factors (acidic and basic) (Abraham et al., 1986; Jaye et al., 1986), yeast a factor (Schafer et that this form may have a functional role outside of the cell. al., 1989), and the bacterialhemolysins (Felmlee et al., 1985). The IL-1p precursor ismuch less active than the mature 17While the mechanism for the secretion of the latter two has kDa form invarious T cell assays (Mosley et al., 1987a, 198713; been recently characterized (McGrath and Varshavsky,1989; Black et al., 1988; Hazuda et al., 1989). We recently purified Koronakis et al., 1989), thatfor the mammalian cell products precursor IL-lB expressed in Escherichia coli to homogeneity remains unclear. Hence it isof interest to study the structuraland were able to show that it had less than 10”-fold of the activity of mature, recombinant IL-lp (Hazuda et al., 1989). characteristics of the secreted proteins in order to understand This differential activitywas not an artifact of expression or how they are secreted. The IL-1s are releasedpredominantly from activated mono- isolation, as the recombinant precursor could be made fully active following proteolysis to the 17-kDa form with chymo* The costs of publication of this article were defrayed in part by trypsin. The lack of activity was due to an inability to bind the payment of page charges. This article must therefore be hereby the IL-1 receptor. In contrast, other studies have indicated marked “aduertisement” in accordance with 18 U.S.C. Section 1734 that the IL-laprecursor is fully active (Mosley et al., 1987b). solely to indicate this fact. difference in Q Presentaddress:Dept. of Virus and Cell Biology, WP16-215, The question therefore arises as to whether this or conformational effects activity is due to stereochemical Merck Sharp and Dohme Research Laboratories, West Point, PA imposed on the IL-lP matureregion by its precursor domain, 19486. ** Presentaddress:CIBA-GEIGYPharmaceuticals, 556 Morris which are absent in precursor IL-10. Ave., Summit, NJ 07901. In this paper we examine the structure and activityof the $$ To whom correspondence should he addressed. ’ Theabbreviations used are:IL-1,interleukin 1; HIV,human IL-1s by examining the proteinaseK sensitivity andcysteine immunodeficiency virus; SDS, sodium dodecyl sulfate; DTNB, 5,.5’dit hiohis(%-nitrobenzoic acid).

7081

’ P. R. Krasney and P. R. Young, unpublished studies.

Structure-Function Mappingof IL-1 Precursors accessibility of precursor IL-Is, mature IL-Is, and several intermediate-length proteolyt,ic fragments. Our results demonstrate that the precursor domains of hoth IL-1s a k e r t h e conformation of their corresponding mature regions, despite the fact that theIL-ltu precursor retains full activitv and the IL-1I.I precursor does not. Addit.ional mapping indicates that t h e region which suppresses the activityof precursor IL-I/j is distinct from that which causes the conformational change. MATERIALSANDMETHODS ('loninp, /:'.~prrssion. nnd I'urificntion of Hrcomhinnnt I'rrcursor 11,-

/(t-'l'he cloning nfa partial human I L I o c D N A has heen descrihed I,! nl.. 1989). 'To ohtain a c D N A encoding the f u l l prc~iously (Hassell lenkqh precursor, we screened the activated human monocyte cDNA lihraryin p13lCW'L originally used fortheisolation o f thepartial c.I)Sh (Meyers r t nl., 1987) hy hyhritlization with a nick-translated / ~ , ' ( ~ ~ ~ l ~ l / / ~ restriction i ~ t ( l l l l fragmentfromthepartial cDNA. 'This (.I)NA was then engineered for expression in E . coli hv isolating an XfiO-hase pair / l n l l / N d r l fragment containing the complete coding region. ligating i t t o a self-complementnry synthetic oligonucleotide linker ('('A'I'(;(;'I'A('('A'I'(;(;, recleaving with NcoI/HincII. antl lipat ing into t h r Nu,I/I.,'collV sites of t h e pOTSNco /

0

2

~"

"

GuHCI = 6 M punnidinium hydrochloride, 2 mxf F:I)Th. " Mevers et al.. 1987. "

and can only be reacted with DTNH upon denaturation in 6 M guanidinium chloride (Meyerset al., 1987). Together, these data demonstrate that the mature domain of precursor IL-IIj has a more exposed, protease-sensitive tertiary structure than matureIL-I&Similarchemicalmodificationdataarenot availablefor IL-ICY,but it appears that the proteinase Ksensitive precursor structure has been conserved between 11,In and [ j , suggesting a significant role in 11,-1 hiology. Structure-Function Mapping of II,- I B Precursor-To assess . is required to maintainthe if theentireprecursorregion alternat.ive conformationof the mature domain,we examined FIG. 4. Proteinnse K sensitivity of recombinant precursor the proteinase K sensitivity of two proteolytic fragments of and mature ll,-ls. 0 . 1 mg:/ml purified rccomhinant precursor and precursor IL-1/9, which have some or most of the precursor mature l l d - l ( ~( p o n d A ) and if ( p n n d H ) were treatedwith (+) or without (-) 0 . 0 1 mg/ml proteinase K for 5 min at 22 "C and analyzed region removed. Tr-ypsin and chymotrypsin remove 76 and a s in Fig. 2 . 1 1 3 amino acids from the 116-amino acid precursor region. respectively (Black e l al., 1988; Hazuda et al., 1989) (Fig. 3 a n d Fig. 5 , lanes .? a n d 5 ) . In conditions where the entire precursor was completely sensitive to proteinase K, the chymotr-yptic fragment (residues 114-269) was resistant, much likemature lL-l[j (Fig. 5 , lane 6).In contrast, the tryptic fragment (residues 77-269) was degraded to a molecule close insize to mature IL-l/j and to thechymotrypsin-cleaved precursor IL-18 (Fig. 5 , lane 4 ) . These data indicate that the carboxyl termini of both the tr-ypsin and chymotrypsin fragments were ina proteinase K-resistant conformation like that of mature IL-I/j. In the caseof the tr-yptic fragment, only the amino-terminal extension remained in a proteinase K-sensitive conformation, similar to the tryptic fragment of l I , - l c r described earlier (Fig.44 ). These data suggest that removalof the amino-terminal 76 amino acids of 11,-113 precursor(domain I inFig. :K') was sufficient to allow the mature region to fold into its proteaseinsensitiveconformation. A comparableregionmaybeinvolved for t h e 11,-lcu precursor, since Kobayashi et a / . (1990) were only able to obtain stable precursor I L - I ( r in E . coli if theydeletedtheamino-terminal 64 residues.Interestingly, within domain I was found the most highly conserved region between the I L - l t r a n d / j precursors, residues 1-19. These are predicted to form an amphipathic cr-helix (Young and Sylvester, 1989). Further experiments will be necessary to map the precise residues responsible for the conformational alteration of the mature domains. The altered conformation of the I L I d mature region induced by domain I is not sufficient to account for the inactivdifferential sensitivity of precursor and mature IL-Is to pro- ity of the precursor. The most substantial reduction of its teinase K digestion supports that the mature region of both activity is attributable to domainI1 (residues 77-94), suggestIL-1s was in a different. conformation prior to cleavage from ing that the 39 residues adjacent to the amino terminus of t h e precursors. mature IL-lLj may sterically hinder accessihility to the recepFurther evidence for an alternative conformation for the is consistent with studiesof site-specific tor binding site. This mature region of purified precursor IL-ld was obtained by mutants that implicate the amino terminus of 11,-I in receptor examining the accessibility of its cysteine residues to chemicalbinding and activity (Horuk et al., 1987; Huang et 01.. 1987: of Masui ef al., 1989).' Furthermore,thefindingfrom modification(Table 11). Titration of thefourcysteines x-ray precursor IL-l/j with the sulfhydryl reagent DTNH indicated crystallography that the I L - I n and I L - l J structuresdiffer that all four were reactive in the native structure and not most in the orientation of their amino termini (l'reistle ct al., h i e d or involved in disulfide linkages. In contrast, the two cysteine residues retained in mature IL-l/j are buried, though ' 1'. 1.. Simon. iV. Fenderson. S. 1,d'astro. .I, Silvwtri. . I . S. not disulfide-linked (I'riestle et a/., 1988; Finzel ef al., 1989), l,illquist, 1'. K . Hhatnagar. and 1'. I