- Email: [email protected]
Use of the Yeast Three-Hybrid System as a Tool to Study Caspases
ANALYTICAL BIOCHEMISTRY ARTICLE NO.
263, 62– 66 (1998)
AB982823
Use of the Yeast Three-Hybrid System as a Tool to Study Caspases Wim Van Criekinge,1 Maria van Gurp,1 Els Decoster, Peter Schotte, Marc Van de Craen, Walter Fiers, Peter Vandenabeele, and Rudi Beyaert2 Department of Molecular Biology, Flanders Interuniversity Institute for Biotechnology and University of Gent, B-9000 Gent, Belgium
Received July 1, 1998
Caspases are a family of heteromeric (p20/p10) cysteine proteases with important functions in the regulation of apoptosis and inflammation. Up to now, tools to identify new substrates for caspases have mostly been limited to the random screening of in vitro translated proteins that are known, or assumed, to play a role in apoptosis. We describe the use of a yeast three-hybrid approach as a tool that adapts the classical two-hybrid system to the needs of heteromeric caspases for functional dissection of known interactions or screening for physiological substrates and inhibitors. Functional heteromeric caspase-1 was obtained by coexpression of p20(Cys285Ser) and p10 caspase-1 subunits that were each fused to the Gal4 DNA-binding domain. Upon coexpression of a third hybrid of the Gal4 activation domain and the viral caspase-1 pseudosubstrate inhibitors CrmA or p35, or the prototype physiological caspase-1 substrate prointerleukin-1b, a functional Gal4 transcription factor could be reconstituted. In contrast, no interaction was found between CrmA or p35 and the immature p45 or p30 precursor forms of caspase-1. Therefore, the three-hybrid system might allow screening for new physiological substrates and inhibitors of heteromeric caspases. © 1998 Academic Press Key Words: caspase substrates; protein–protein interactions; two hybrid; three hybrid; apoptosis.
Caspase-1 has been identified as the main protease responsible for the maturation in vivo and in vitro of prointerleukin-1b (proIL-1b).3 It is synthesized as a pre1
These authors contributed equally to this study. To whom correspondence should be addressed at Department of Molecular Biology, K. L. Ledeganckstraat 35, B-9000 Gent, Belgium. Fax: 32-9-264-53-48. E-mail: [email protected]. 3 Abbreviations used: proIL-1b, prointerleukin-1b; DB, DNA-binding domain; AD, activation domain. 2
62
cursor molecule of 45 kDa and subsequently processed by removal of an N-terminal prodomain and of an internal linker sequence that separates the p20 and p10 subunits (1, 2). The active site of caspase-1 is composed of amino acid residues from both p20 and p10 subunits, with Cys285 and His237 forming a catalytic dyad in the active site (2). Apart from its role in inflammation, caspase-1 was shown to belong to a larger gene family, many members of which play an important role in apoptosis (3– 6). The number of intracellular substrates for caspases that are identified is growing fast (3, 5, 6). However, the functional significance of the cleavage of most of these substrates is still unclear since many of them are only cleaved during late stages of the apoptotic process. Up to now, tools to identify new substrates for caspases have mostly been limited to the random screening of in vitro translated proteins that are known, or assumed, to play a role in apoptosis. Alternatively, sequencing of proteins that are found to be processed in apoptotically dying cells might also result in the isolation of new substrates. The yeast two-hybrid technique is a well-established method to identify and clone genes encoding proteins that interact with a protein of interest (7–10). However, its application in the screening for proteins which interact with caspases is limited by the multimeric nature of active caspases. Therefore, a caspase substrate or another protein that binds specifically with the active heteromeric p20/p10 form of caspases will escape detection in a twohybrid approach with an unprocessed caspase precursor as bait. Recently, a number of so-called three-hybrid systems to analyze more complex macromolecular interactions have been developed (reviewed in 11). We describe the use of a three-hybrid approach adapted to the needs of caspases to detect and analyze the interaction of caspase-1 with the viral pseudosubstrate caspase inhibitors CrmA and p35 (12, 13), as well as with the prototype caspase-1 substrate proIL-1b (1, 2). The usefulness of 0003-2697/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
63
THREE-HYBRID SYSTEM AS A TOOL TO STUDY CASPASES
such a three-hybrid approach is demonstrated by the observation that CrmA and p35 specifically interact with the mature p20/p10 heteromeric form of caspase-1 and not with its immature precursor forms. MATERIALS AND METHODS
Yeast two-hybrid and three-hybrid systems. The two-hybrid system was purchased from Clontech Laboratories (Palo Alto, CA) and used as previously described (14). To perform a three-hybrid approach, a second Gal4 DNA-binding domain (DB) expression plasmid (pAS3) was created by blunt ligation of an ade2 gene under its own promotor (generously provided by Dr. S. Chavez) into an EcoRV opened pAS2 plasmid. As a result, the trp gene of pAS2 was disrupted and the cycloheximide sensitivity was eliminated. Transformation of the Saccharomyces cerevisiae strain HF7c was done by the lithium acetate method according to the manufacturer’s directions. Yeast colonies carrying putative interacting proteins were selected by growth on synthetic minimal media lacking Trp, Leu, Ade, and His in the presence of 5 mM 3-amino-1,2,4,-triazole, purchased from Sigma Chemical Co. (St. Louis, MO), and by screening for b-galactosidase activity in a filter assay using 5-bromo-4-chloro-3indolyl-b-D-galactopyranoside obtained from Saxon Biochemicals (Hannover, FRG) as substrate (14). Construction of the three-hybrid expression vectors. Murine p45-caspase-1 was cloned in frame as a NdeI–BamHI fragment after the Gal4DB (pAS2) and Gal4AD (pGAD424) coding sequences. Cloning of p10-, p20-, and p30-caspase-1, CrmA, p35, and murine proIL-1b in frame of the Gal4DB and/or Gal4AD was achieved by introducing an additional NcoI and SalI restriction site by PCR at their N- and C-terminal end, respectively. The endogenous NcoI site in proIL-1b was deleted. Site-directed mutagenesis was carried out with a kit from Clontech Laboratories (Palo Alto, CA). Cloning and mutations were verified by DNA sequencing. Quantification of b-galactosidase activity. Quantitative assays were performed as previously described (15), with some minor modifications. Briefly, 3-ml liquid cultures were prepared from individual yeast transformant colonies and grown until mid-log phase (OD600 5 0.5– 0.8). Cell pellets were washed and resuspended in buffer 1 [100 ml 100 mM Hepes, 150 mM NaCl, 4.5 mM L-aspartic acid hemimagnesium salt, 1% bovine serum albumin, 0.1% Tween 20 (v/v) at pH 7.2]. Cells were opened by two freeze–thaw cycles in liquid nitrogen, and 900 ml of 25 mM chlorophenyl-red-b-Dgalactopyranoside (CPRG; Boehringer-Mannheim, FRG) in buffer 1 was added. Samples were mixed and centrifuged to pellet cell debris. Supernatants were transferred to fresh tubes and incubated at room temperature. When the samples showed a color change from
TABLE 1
Reconstitution of a p20/p10 Caspase-1 Heteromer in a Two-Hybrid Approach pAS2
pGAD424
b-Galactosidase (units)
p20-caspase-1 / p20-caspase-1 p53 p53 /
/ p10-caspase-1 p10-caspase-1 SV40-LT / SV40-LT
— — 47/39/23 7208/6314 — —
Note. b-Galactosidase activity was determined by the colony lift method and quantified using a CPRG assay. “—” represents the absence of His prototrophy and b-galactosidase activity of yeast cells cotransformed with pAS2 and pGAD42 plasmids encoding the indicated fusion proteins. Values are the averages of duplicate CPRG assays of three independent His1 transformants (experimental errors ,10%) and are given for two or three separate two-hybrid experiments. Caspase-1 was expressed as an inactive Cys285Ser mutein.
orange to red, 500 ml of 3 mM ZnCl2 was added to stop color development and the OD578 of the samples was measured. A b-galactosidase unit corresponds to 104 3 OD578/(t 3 V 3 OD600), where t is reaction time in minutes and V is volume of culture (3 ml). Under these conditions, cells transformed with a single plasmid or irrelevant non-interacting plasmids contained ,1 unit. RESULTS
Reconstitution of active caspase-1 by coexpression of its p20 and p10 subunits. Because mature active caspase-1 is known as a heteromer of p20 and p10 subunits, we first verified whether p20 and p10 could associate in a two-hybrid approach. Therefore, we cloned the p20 and p10 subunits as a carboxy-terminal fusion to Gal4DB and Gal4AD, respectively, and transformed them to S. cerevisiae strain HF7c that has Gal4-dependent His- and bGal-reporter genes in its genome. Cotransformation of the yeast expression vectors pVA3 and pTD1, which encode Gal4 fusion proteins of the strongly interacting p53 and SV40-LT oncogen, respectively, served as positive controls. The ability of transformed cells to grow on His-deficient plates in the presence of 5 mM 3-amino-1,2,4,-triazole and the expression of b-galactosidase activity were used as parameters for interaction. Coexpression of Gal4DBp20 and Gal4ADp10 was able to reconstitute a functional Gal4 transcription factor, indicative for the formation of a heteromeric interaction between the p20 and p10 subunit of caspase-1 (Table 1). In the latter case, blue colonies were detectable within 5–7 h, whereas the p53/SV40-LT interaction was already visible within a couple of minutes. This difference is also reflected by the relative amounts of b-galactosidase produced upon p20/p10 or p53/SV40-LT interaction,
64
VAN CRIEKINGE ET AL.
respectively. It should be noted that such measurements by two- or three-hybrid analysis are subject to many variables as discussed previously (16) and that they only reflect strongly versus weakly interacting protein pairs. Moreover, the observation that coexpression of Gal4DBp20 and Gal4ADp10 resulted in very low transformation efficiencies and severe growth retardation suggests that caspase-1 expression is toxic for yeast. This could be avoided by using an inactive caspase-1 mutein in which Cys285 in the catalytic site was mutated to Ser. For this reason, a caspase-1 (Cys285Ser) mutein was used in all subsequent experiments. Moreover, the use of an inactive mutein might be advantageous for trapping substrates of these proteases since fast release of the substrate upon cleavage would be avoided. CrmA, p35, and proIL-1b interact with the heteromeric p20/p10 form of caspase-1 in a three-hybrid approach. To examine the potential of a three-hybrid approach to detect proteins that interact with heteromeric caspases, we analyzed the interaction between a caspase-1 p20/p10
TABLE 2
CrmA, p35, and proIL-1b Associate with a p20/p10 Caspase-1 Heteromer in a Three-Hybrid Approach pAS2
pAS3
pGAD424
b-Galactosidase (units)
p20-caspase-1 / p20-caspase-1
/ p10-caspase-1 p10-caspase-1
/ / /
— — —
p20-caspase-1 p20-caspase-1 p20-caspase-1
/ / /
CrmA p35 proIL-1b
— — —
/ / /
p10-caspase-1 p10-caspase-1 p10-caspase-1
CrmA p35 proIL-1b
— — —
/ / /
/ / /
CrmA p35 proIL-1b
— — —
p20-caspase-1 p20-caspase-1 p20-caspase-1
p10-caspase-1 p10-caspase-1 p10-caspase-1
CrmA p35 proIL-1b
48/69/76 37/27 80/66/78
p53 p53 p53
p10-caspase-1 p10-caspase-1 p10-caspase-1
CrmA p35 proIL-1b
— — —
p20-caspase-1
p10-caspase-1
SV40-LT
—
Note. b-Galactosidase activity was determined by the colony lift method and quantified using a CPRG assay. “—” represents the absence of His prototrophy and b-galactosidase activity of yeast cells cotransformed with pAS2, pAS3, and pGAD424 plasmids encoding the indicated fusion proteins. Values are the averages of duplicate CPRG assays of three independent His1 transformants (experimental errors ,10%) and are given for two or three separate three-hybrid experiments. Caspase-1 was expressed as an inactive Cys285Ser mutein.
FIG. 1. Schematic representation of the interaction of CrmA with a p20/p10 caspase-1 heteromer in a yeast three-hybrid system.
heteromer and known substrates, in casu the viral caspase-1 pseudosubstrate inhibitors CrmA and p35 as well as the prototype physiological caspase-1 substrate proIL-1b. Therefore, a third expression plasmid (pAS3) which allows expression of a Gal4DB fusion protein was constructed by replacing the trp selection marker of pAS2 with the ade2 selection marker. Association was subsequently analyzed by transforming HF7c with pAS2-p20-caspase-1(Cys285Ser), pAS3-p10-caspase-1, and pGAD424-CrmA, pGAD424-p35, or pGAD424-proIL1b. Single plasmid transformations or cotransformation using the appropriate empty plasmid pGAD424, pAS2, or pAS3 were used as negative controls. The specificity of the interactions was verified by cotransforming cDNAs coding for irrelevant non-interacting proteins, viz. p53 or SV40-LT. Reporter gene activity was specifically detected in yeast cells expressing Gal4DB-p20caspase-1 and Gal4DB-p10-caspase-1 in combination with Gal4AD-CrmA, Gal4AD-p35, or Gal4AD-proIL-1b (Table 2). These results clearly demonstrate that CrmA, p35, and proIL-1b are able to interact with a p20/p10 caspase-1 heteromer in a three-hybrid approach (Fig. 1). Quantification of b-galactosidase showed similar levels for all three interactions, although slightly lower amounts were found in the case of p35. CrmA and p35 do not bind the zymogen form of caspase-1. To analyze whether CrmA and p35 can also interact with immature caspase-1, we cloned the p45-caspase-1 precursor as well as a p30-caspase-1 intermediate form lacking the N-terminal prodomain in frame with the Gal4DB and tested their interaction with Gal4AD-CrmA or Gal4AD-p35 in a two-hybrid analysis. The previously described interaction of p45caspase-1 with itself was used as a positive control (14). Although specific homomerization of p45-caspase-1
THREE-HYBRID SYSTEM AS A TOOL TO STUDY CASPASES TABLE 3
CrmA and p35 Do Not Associate with p45-Caspase-1 or p30-Caspase-1 in a Two-Hybrid Approach pAS2
pGAD424
b-Galactosidase (units)
p45-caspase-1 p45-caspase-1 p45-caspase-1 p45-caspase-1
/ p45/caspase-1 CrmA p35
— 121/139 — —
p30-caspase-1 p30-caspase p30-caspase
/ CrmA p35
— — —
/ / /
p45-caspase-1 CrmA p35
— — —
p53 p53 /
SV40-LT / SV40-LT
7208/6314 — —
Note. b-Galactosidase activity was determined by the colony lift method and quantified using a CPRG assay. “—” represents the absence of His prototrophy and b-galactosidase activity of yeast cells cotransformed with pAS2 and pGAD424 plasmids encoding the indicated fusion proteins. Values are the averages of duplicate b-galactosidase assays of three independent His1 transformants (experimental errors ,10%) and are given for two separate two-hybrid experiments. Caspase-1 was expressed as an inactive Cys285Ser mutein.
was again observed, we could not demonstrate any interaction of CrmA or p35 with the p45- or p30caspase-1 zymogens (Table 3). Together with the data described in the previous section, these results clearly demonstrate that CrmA and p35 specifically interact with mature p20/p10 caspase-1. DISCUSSION
The number of cellular substrates for caspases that are identified is growing fast (3, 6). Cleavage of many of these substrates can often be linked to the morphological changes that occur in apoptotic cells, such as chromatin condensation, DNA degradation, cytoskeleton breakdown, membrane blebbing, and formation of apoptotic bodies. In contrast to the proteolytic maturation of proIL-1b and proIL-18 by caspase-1 (17), the regulatory importance of the cleavage of most substrates is still unclear since many of them are only cleaved during late stages of the apoptotic process. Complete understanding of the function of caspases awaits the identification of cellular substrates or inhibitors of caspases that act early during signaling pathways. In this paper we describe the potential use of a three-hybrid approach to detect proteins that specifically interact with the mature form of caspases. Indeed, we were able to demonstrate that the viral pseudosubstrate caspase inhibitors CrmA and p35 specifically interact with the p20/p10 heteromeric form of caspase-1 and not with
65
the caspase-1 zymogen. This clearly shows that CrmA and p35 do not inhibit caspase-1 by preventing its maturation, but by blocking the enzymatic activity of the mature caspase. Moreover, in a similar threehybrid approach we could demonstrate the interaction between p20/p10 caspase-1 and its prototype physiological substrate proIL-1b, indicating that a threehybrid screening with caspases is also applicable to detect transient enzyme–substrate interactions. An important difference of our three-hybrid system compared to those that have recently been described (11, 18) is the fact that both baits (viz. p20 and p10 subunits of caspase-1) are expressed as a fusion with Gal4DB. The nuclear localization sequence of the latter assures that these fusion proteins will be properly transported into the nucleus which is a prerequisite for reporter gene activation. Moreover, the use of two Gal4DB fusion proteins allows to isolate also proteins that would specifically bind with one of the two Gal4DB fusion proteins, thus combining two-hybrid and three-hybrid screenings in a single experiment. Finally, the use of ade2 as an additional auxotrophic selection marker makes our three-hybrid system compatible with the existing yeast strains and libraries that are currently used in Gal4-based two- and threehybrid systems and which make use of trp1, leu2, his3, and ura3 as selection markers. In principle even a four-hybrid system could be made by combining these systems. Our data clearly show that three-hybrid screening of a cDNA library might be a valuable new approach to identify potential substrates or other interacting partners of caspases. Indeed, three-hybrid screening of a cDNA library with p20/p10 caspase-1 as bait already resulted in several candidate caspase-1 interacting proteins (unpublished observations), including actin that was recently shown to be a substrate for caspase-1 (19). Further characterization of these proteins is underway. ACKNOWLEDGMENTS Dr. S. Chavez is thanked for providing us with a plasmid containing the ade2 selection marker. W. Van Criekinge is a research assistant, and R. Beyaert and P. Vandenabeele are postdoctoral researchers, respectively, with the FWO-Vlaanderen. P. Schotte and M. van Gurp are Fellows of the IWT. This research was supported by the IUAP, the FWO-Vlaanderen, an EC-BIOMED2 grant (BMH4CT96-0300), and an EC-TMR grant (ERBFMRXCT970153).
REFERENCES 1. Walker, N. P., Talanian, R. V., Brady, K. D., Dang, L. C., Bump, N. J., Ferenz, C. R., Franklin, S., Ghayur, T., Hackett, M. C., Hammill, L. D., Herzog, L., Hugunin, M., Houy, W., Mankovich, J. A., McGuiness, L., Orlewicz, E., Paskind, M., Pratt, C. A., Reis, P., Summani, A., Terranova, M., Welch, J. P., Xiong, L., Mo¨ller, A., Tracey, D. E., Kamen, R., and Wong, W. W. (1994) Cell 78, 343–352.
66
VAN CRIEKINGE ET AL.
2. Wilson, K. P., Black, J. A., Thomson, J. A., Kim, E. E., Griffith, J. P., Navia, M. A., Murcko, M. A., Chambers, S. P., Aldape, R. A., Raybuck, S. A., and Livingston, D. J. (1994) Nature 370, 270 –275. 3. Cohen, G. M. (1997) Biochem. J. 326, 1–16. 4. Van de Craen, M., Vandenabeele, P., Declercq, W., Van den brande, I., Van Loo, G., Molemans, F., Schotte, P., Van Criekinge, W., Beyaert, R., and Fiers, W. (1997) FEBS Lett. 403, 61– 69. 5. Earnshaw, W. C., and Martins, L. M. (1997) Trends Cell Biol. 7, 111–114. 6. Porter, A. G., Ng, P., and Janicke, R. U. (1997) BioEssays 19, 501–507. 7. Fields, S., and Song, O. K. (1989) Nature 340, 245–246. 8. Chien, C. T., Bartel, P. L., Sternglanz, R., and Fields, S. (1991) Proc. Natl. Acad. Sci. USA 88, 9578 –9582. 9. Fields, S., and Sternglanz, R. (1994) Trends Genet. 10, 286 – 292. 10. Luban, J., and Goff, S. P. (1995) Curr. Opin. Biotechnol. 6, 59 – 64. 11. Brachmann, R. K., and Boeke, J. D. (1997) Curr. Opin. Biotechnol. 8, 561–568.
12. Ray, C. A., Black, R. A., Kronheim, S. R., Greenstreet, T. A., Sleath, P. R., Salvesen, G. S., and Pickup, D. J. (1992) Cell 69, 597– 604. 13. Bump, N. J., Hackett, M., Hugunin, M., Seshagiri, S., Brady, K., Chen, P., Ferenz, C., Franklin, S., Ghayur, T., Li, P., Licari, J., Mankovich, L., Shi, A., Greenberg, H., Miller, L. K., and Wong, W. W. (1995) Science 269, 1885–1888. 14. Van Criekinge, W., Beyaert, R., Van de Craen, M., Vandenabeele, P., Schotte, P., De Valck, D., and Fiers, W. (1996) J. Biol. Chem. 271, 27245–27248. 15. Legrain, P., Chapon, C., and Galisson, F. (1993) Genes Dev. 7, 1390 –1399. 16. Estojak, J., Brent, R., and Golemis, E. A. (1995) Mol. Cell. Biol. 15, 5820 –5829. 17. Ghayur, T., Banerjee, S., Hugunin, M., Butler, D., Herzog, L., Carter, A., Quintal, L., Sekut, L., Talanian, R., Paskind, M., Wong, W., Kamen, R., Tracey, D., and Allen, H. (1997) Nature 386, 619 – 623. 18. Zhang, J., and Lautar, S. (1996) Anal. Biochem. 242, 68 –72. 19. Mashima, T., Naito, M., Fujita, N., Noguchi, K., and Tsuruo, T. (1995) Biochem. Biophys. Res. Commun. 217, 1185–1192.
263, 62– 66 (1998)
AB982823
Use of the Yeast Three-Hybrid System as a Tool to Study Caspases Wim Van Criekinge,1 Maria van Gurp,1 Els Decoster, Peter Schotte, Marc Van de Craen, Walter Fiers, Peter Vandenabeele, and Rudi Beyaert2 Department of Molecular Biology, Flanders Interuniversity Institute for Biotechnology and University of Gent, B-9000 Gent, Belgium
Received July 1, 1998
Caspases are a family of heteromeric (p20/p10) cysteine proteases with important functions in the regulation of apoptosis and inflammation. Up to now, tools to identify new substrates for caspases have mostly been limited to the random screening of in vitro translated proteins that are known, or assumed, to play a role in apoptosis. We describe the use of a yeast three-hybrid approach as a tool that adapts the classical two-hybrid system to the needs of heteromeric caspases for functional dissection of known interactions or screening for physiological substrates and inhibitors. Functional heteromeric caspase-1 was obtained by coexpression of p20(Cys285Ser) and p10 caspase-1 subunits that were each fused to the Gal4 DNA-binding domain. Upon coexpression of a third hybrid of the Gal4 activation domain and the viral caspase-1 pseudosubstrate inhibitors CrmA or p35, or the prototype physiological caspase-1 substrate prointerleukin-1b, a functional Gal4 transcription factor could be reconstituted. In contrast, no interaction was found between CrmA or p35 and the immature p45 or p30 precursor forms of caspase-1. Therefore, the three-hybrid system might allow screening for new physiological substrates and inhibitors of heteromeric caspases. © 1998 Academic Press Key Words: caspase substrates; protein–protein interactions; two hybrid; three hybrid; apoptosis.
Caspase-1 has been identified as the main protease responsible for the maturation in vivo and in vitro of prointerleukin-1b (proIL-1b).3 It is synthesized as a pre1
These authors contributed equally to this study. To whom correspondence should be addressed at Department of Molecular Biology, K. L. Ledeganckstraat 35, B-9000 Gent, Belgium. Fax: 32-9-264-53-48. E-mail: [email protected]. 3 Abbreviations used: proIL-1b, prointerleukin-1b; DB, DNA-binding domain; AD, activation domain. 2
62
cursor molecule of 45 kDa and subsequently processed by removal of an N-terminal prodomain and of an internal linker sequence that separates the p20 and p10 subunits (1, 2). The active site of caspase-1 is composed of amino acid residues from both p20 and p10 subunits, with Cys285 and His237 forming a catalytic dyad in the active site (2). Apart from its role in inflammation, caspase-1 was shown to belong to a larger gene family, many members of which play an important role in apoptosis (3– 6). The number of intracellular substrates for caspases that are identified is growing fast (3, 5, 6). However, the functional significance of the cleavage of most of these substrates is still unclear since many of them are only cleaved during late stages of the apoptotic process. Up to now, tools to identify new substrates for caspases have mostly been limited to the random screening of in vitro translated proteins that are known, or assumed, to play a role in apoptosis. Alternatively, sequencing of proteins that are found to be processed in apoptotically dying cells might also result in the isolation of new substrates. The yeast two-hybrid technique is a well-established method to identify and clone genes encoding proteins that interact with a protein of interest (7–10). However, its application in the screening for proteins which interact with caspases is limited by the multimeric nature of active caspases. Therefore, a caspase substrate or another protein that binds specifically with the active heteromeric p20/p10 form of caspases will escape detection in a twohybrid approach with an unprocessed caspase precursor as bait. Recently, a number of so-called three-hybrid systems to analyze more complex macromolecular interactions have been developed (reviewed in 11). We describe the use of a three-hybrid approach adapted to the needs of caspases to detect and analyze the interaction of caspase-1 with the viral pseudosubstrate caspase inhibitors CrmA and p35 (12, 13), as well as with the prototype caspase-1 substrate proIL-1b (1, 2). The usefulness of 0003-2697/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
63
THREE-HYBRID SYSTEM AS A TOOL TO STUDY CASPASES
such a three-hybrid approach is demonstrated by the observation that CrmA and p35 specifically interact with the mature p20/p10 heteromeric form of caspase-1 and not with its immature precursor forms. MATERIALS AND METHODS
Yeast two-hybrid and three-hybrid systems. The two-hybrid system was purchased from Clontech Laboratories (Palo Alto, CA) and used as previously described (14). To perform a three-hybrid approach, a second Gal4 DNA-binding domain (DB) expression plasmid (pAS3) was created by blunt ligation of an ade2 gene under its own promotor (generously provided by Dr. S. Chavez) into an EcoRV opened pAS2 plasmid. As a result, the trp gene of pAS2 was disrupted and the cycloheximide sensitivity was eliminated. Transformation of the Saccharomyces cerevisiae strain HF7c was done by the lithium acetate method according to the manufacturer’s directions. Yeast colonies carrying putative interacting proteins were selected by growth on synthetic minimal media lacking Trp, Leu, Ade, and His in the presence of 5 mM 3-amino-1,2,4,-triazole, purchased from Sigma Chemical Co. (St. Louis, MO), and by screening for b-galactosidase activity in a filter assay using 5-bromo-4-chloro-3indolyl-b-D-galactopyranoside obtained from Saxon Biochemicals (Hannover, FRG) as substrate (14). Construction of the three-hybrid expression vectors. Murine p45-caspase-1 was cloned in frame as a NdeI–BamHI fragment after the Gal4DB (pAS2) and Gal4AD (pGAD424) coding sequences. Cloning of p10-, p20-, and p30-caspase-1, CrmA, p35, and murine proIL-1b in frame of the Gal4DB and/or Gal4AD was achieved by introducing an additional NcoI and SalI restriction site by PCR at their N- and C-terminal end, respectively. The endogenous NcoI site in proIL-1b was deleted. Site-directed mutagenesis was carried out with a kit from Clontech Laboratories (Palo Alto, CA). Cloning and mutations were verified by DNA sequencing. Quantification of b-galactosidase activity. Quantitative assays were performed as previously described (15), with some minor modifications. Briefly, 3-ml liquid cultures were prepared from individual yeast transformant colonies and grown until mid-log phase (OD600 5 0.5– 0.8). Cell pellets were washed and resuspended in buffer 1 [100 ml 100 mM Hepes, 150 mM NaCl, 4.5 mM L-aspartic acid hemimagnesium salt, 1% bovine serum albumin, 0.1% Tween 20 (v/v) at pH 7.2]. Cells were opened by two freeze–thaw cycles in liquid nitrogen, and 900 ml of 25 mM chlorophenyl-red-b-Dgalactopyranoside (CPRG; Boehringer-Mannheim, FRG) in buffer 1 was added. Samples were mixed and centrifuged to pellet cell debris. Supernatants were transferred to fresh tubes and incubated at room temperature. When the samples showed a color change from
TABLE 1
Reconstitution of a p20/p10 Caspase-1 Heteromer in a Two-Hybrid Approach pAS2
pGAD424
b-Galactosidase (units)
p20-caspase-1 / p20-caspase-1 p53 p53 /
/ p10-caspase-1 p10-caspase-1 SV40-LT / SV40-LT
— — 47/39/23 7208/6314 — —
Note. b-Galactosidase activity was determined by the colony lift method and quantified using a CPRG assay. “—” represents the absence of His prototrophy and b-galactosidase activity of yeast cells cotransformed with pAS2 and pGAD42 plasmids encoding the indicated fusion proteins. Values are the averages of duplicate CPRG assays of three independent His1 transformants (experimental errors ,10%) and are given for two or three separate two-hybrid experiments. Caspase-1 was expressed as an inactive Cys285Ser mutein.
orange to red, 500 ml of 3 mM ZnCl2 was added to stop color development and the OD578 of the samples was measured. A b-galactosidase unit corresponds to 104 3 OD578/(t 3 V 3 OD600), where t is reaction time in minutes and V is volume of culture (3 ml). Under these conditions, cells transformed with a single plasmid or irrelevant non-interacting plasmids contained ,1 unit. RESULTS
Reconstitution of active caspase-1 by coexpression of its p20 and p10 subunits. Because mature active caspase-1 is known as a heteromer of p20 and p10 subunits, we first verified whether p20 and p10 could associate in a two-hybrid approach. Therefore, we cloned the p20 and p10 subunits as a carboxy-terminal fusion to Gal4DB and Gal4AD, respectively, and transformed them to S. cerevisiae strain HF7c that has Gal4-dependent His- and bGal-reporter genes in its genome. Cotransformation of the yeast expression vectors pVA3 and pTD1, which encode Gal4 fusion proteins of the strongly interacting p53 and SV40-LT oncogen, respectively, served as positive controls. The ability of transformed cells to grow on His-deficient plates in the presence of 5 mM 3-amino-1,2,4,-triazole and the expression of b-galactosidase activity were used as parameters for interaction. Coexpression of Gal4DBp20 and Gal4ADp10 was able to reconstitute a functional Gal4 transcription factor, indicative for the formation of a heteromeric interaction between the p20 and p10 subunit of caspase-1 (Table 1). In the latter case, blue colonies were detectable within 5–7 h, whereas the p53/SV40-LT interaction was already visible within a couple of minutes. This difference is also reflected by the relative amounts of b-galactosidase produced upon p20/p10 or p53/SV40-LT interaction,
64
VAN CRIEKINGE ET AL.
respectively. It should be noted that such measurements by two- or three-hybrid analysis are subject to many variables as discussed previously (16) and that they only reflect strongly versus weakly interacting protein pairs. Moreover, the observation that coexpression of Gal4DBp20 and Gal4ADp10 resulted in very low transformation efficiencies and severe growth retardation suggests that caspase-1 expression is toxic for yeast. This could be avoided by using an inactive caspase-1 mutein in which Cys285 in the catalytic site was mutated to Ser. For this reason, a caspase-1 (Cys285Ser) mutein was used in all subsequent experiments. Moreover, the use of an inactive mutein might be advantageous for trapping substrates of these proteases since fast release of the substrate upon cleavage would be avoided. CrmA, p35, and proIL-1b interact with the heteromeric p20/p10 form of caspase-1 in a three-hybrid approach. To examine the potential of a three-hybrid approach to detect proteins that interact with heteromeric caspases, we analyzed the interaction between a caspase-1 p20/p10
TABLE 2
CrmA, p35, and proIL-1b Associate with a p20/p10 Caspase-1 Heteromer in a Three-Hybrid Approach pAS2
pAS3
pGAD424
b-Galactosidase (units)
p20-caspase-1 / p20-caspase-1
/ p10-caspase-1 p10-caspase-1
/ / /
— — —
p20-caspase-1 p20-caspase-1 p20-caspase-1
/ / /
CrmA p35 proIL-1b
— — —
/ / /
p10-caspase-1 p10-caspase-1 p10-caspase-1
CrmA p35 proIL-1b
— — —
/ / /
/ / /
CrmA p35 proIL-1b
— — —
p20-caspase-1 p20-caspase-1 p20-caspase-1
p10-caspase-1 p10-caspase-1 p10-caspase-1
CrmA p35 proIL-1b
48/69/76 37/27 80/66/78
p53 p53 p53
p10-caspase-1 p10-caspase-1 p10-caspase-1
CrmA p35 proIL-1b
— — —
p20-caspase-1
p10-caspase-1
SV40-LT
—
Note. b-Galactosidase activity was determined by the colony lift method and quantified using a CPRG assay. “—” represents the absence of His prototrophy and b-galactosidase activity of yeast cells cotransformed with pAS2, pAS3, and pGAD424 plasmids encoding the indicated fusion proteins. Values are the averages of duplicate CPRG assays of three independent His1 transformants (experimental errors ,10%) and are given for two or three separate three-hybrid experiments. Caspase-1 was expressed as an inactive Cys285Ser mutein.
FIG. 1. Schematic representation of the interaction of CrmA with a p20/p10 caspase-1 heteromer in a yeast three-hybrid system.
heteromer and known substrates, in casu the viral caspase-1 pseudosubstrate inhibitors CrmA and p35 as well as the prototype physiological caspase-1 substrate proIL-1b. Therefore, a third expression plasmid (pAS3) which allows expression of a Gal4DB fusion protein was constructed by replacing the trp selection marker of pAS2 with the ade2 selection marker. Association was subsequently analyzed by transforming HF7c with pAS2-p20-caspase-1(Cys285Ser), pAS3-p10-caspase-1, and pGAD424-CrmA, pGAD424-p35, or pGAD424-proIL1b. Single plasmid transformations or cotransformation using the appropriate empty plasmid pGAD424, pAS2, or pAS3 were used as negative controls. The specificity of the interactions was verified by cotransforming cDNAs coding for irrelevant non-interacting proteins, viz. p53 or SV40-LT. Reporter gene activity was specifically detected in yeast cells expressing Gal4DB-p20caspase-1 and Gal4DB-p10-caspase-1 in combination with Gal4AD-CrmA, Gal4AD-p35, or Gal4AD-proIL-1b (Table 2). These results clearly demonstrate that CrmA, p35, and proIL-1b are able to interact with a p20/p10 caspase-1 heteromer in a three-hybrid approach (Fig. 1). Quantification of b-galactosidase showed similar levels for all three interactions, although slightly lower amounts were found in the case of p35. CrmA and p35 do not bind the zymogen form of caspase-1. To analyze whether CrmA and p35 can also interact with immature caspase-1, we cloned the p45-caspase-1 precursor as well as a p30-caspase-1 intermediate form lacking the N-terminal prodomain in frame with the Gal4DB and tested their interaction with Gal4AD-CrmA or Gal4AD-p35 in a two-hybrid analysis. The previously described interaction of p45caspase-1 with itself was used as a positive control (14). Although specific homomerization of p45-caspase-1
THREE-HYBRID SYSTEM AS A TOOL TO STUDY CASPASES TABLE 3
CrmA and p35 Do Not Associate with p45-Caspase-1 or p30-Caspase-1 in a Two-Hybrid Approach pAS2
pGAD424
b-Galactosidase (units)
p45-caspase-1 p45-caspase-1 p45-caspase-1 p45-caspase-1
/ p45/caspase-1 CrmA p35
— 121/139 — —
p30-caspase-1 p30-caspase p30-caspase
/ CrmA p35
— — —
/ / /
p45-caspase-1 CrmA p35
— — —
p53 p53 /
SV40-LT / SV40-LT
7208/6314 — —
Note. b-Galactosidase activity was determined by the colony lift method and quantified using a CPRG assay. “—” represents the absence of His prototrophy and b-galactosidase activity of yeast cells cotransformed with pAS2 and pGAD424 plasmids encoding the indicated fusion proteins. Values are the averages of duplicate b-galactosidase assays of three independent His1 transformants (experimental errors ,10%) and are given for two separate two-hybrid experiments. Caspase-1 was expressed as an inactive Cys285Ser mutein.
was again observed, we could not demonstrate any interaction of CrmA or p35 with the p45- or p30caspase-1 zymogens (Table 3). Together with the data described in the previous section, these results clearly demonstrate that CrmA and p35 specifically interact with mature p20/p10 caspase-1. DISCUSSION
The number of cellular substrates for caspases that are identified is growing fast (3, 6). Cleavage of many of these substrates can often be linked to the morphological changes that occur in apoptotic cells, such as chromatin condensation, DNA degradation, cytoskeleton breakdown, membrane blebbing, and formation of apoptotic bodies. In contrast to the proteolytic maturation of proIL-1b and proIL-18 by caspase-1 (17), the regulatory importance of the cleavage of most substrates is still unclear since many of them are only cleaved during late stages of the apoptotic process. Complete understanding of the function of caspases awaits the identification of cellular substrates or inhibitors of caspases that act early during signaling pathways. In this paper we describe the potential use of a three-hybrid approach to detect proteins that specifically interact with the mature form of caspases. Indeed, we were able to demonstrate that the viral pseudosubstrate caspase inhibitors CrmA and p35 specifically interact with the p20/p10 heteromeric form of caspase-1 and not with
65
the caspase-1 zymogen. This clearly shows that CrmA and p35 do not inhibit caspase-1 by preventing its maturation, but by blocking the enzymatic activity of the mature caspase. Moreover, in a similar threehybrid approach we could demonstrate the interaction between p20/p10 caspase-1 and its prototype physiological substrate proIL-1b, indicating that a threehybrid screening with caspases is also applicable to detect transient enzyme–substrate interactions. An important difference of our three-hybrid system compared to those that have recently been described (11, 18) is the fact that both baits (viz. p20 and p10 subunits of caspase-1) are expressed as a fusion with Gal4DB. The nuclear localization sequence of the latter assures that these fusion proteins will be properly transported into the nucleus which is a prerequisite for reporter gene activation. Moreover, the use of two Gal4DB fusion proteins allows to isolate also proteins that would specifically bind with one of the two Gal4DB fusion proteins, thus combining two-hybrid and three-hybrid screenings in a single experiment. Finally, the use of ade2 as an additional auxotrophic selection marker makes our three-hybrid system compatible with the existing yeast strains and libraries that are currently used in Gal4-based two- and threehybrid systems and which make use of trp1, leu2, his3, and ura3 as selection markers. In principle even a four-hybrid system could be made by combining these systems. Our data clearly show that three-hybrid screening of a cDNA library might be a valuable new approach to identify potential substrates or other interacting partners of caspases. Indeed, three-hybrid screening of a cDNA library with p20/p10 caspase-1 as bait already resulted in several candidate caspase-1 interacting proteins (unpublished observations), including actin that was recently shown to be a substrate for caspase-1 (19). Further characterization of these proteins is underway. ACKNOWLEDGMENTS Dr. S. Chavez is thanked for providing us with a plasmid containing the ade2 selection marker. W. Van Criekinge is a research assistant, and R. Beyaert and P. Vandenabeele are postdoctoral researchers, respectively, with the FWO-Vlaanderen. P. Schotte and M. van Gurp are Fellows of the IWT. This research was supported by the IUAP, the FWO-Vlaanderen, an EC-BIOMED2 grant (BMH4CT96-0300), and an EC-TMR grant (ERBFMRXCT970153).
REFERENCES 1. Walker, N. P., Talanian, R. V., Brady, K. D., Dang, L. C., Bump, N. J., Ferenz, C. R., Franklin, S., Ghayur, T., Hackett, M. C., Hammill, L. D., Herzog, L., Hugunin, M., Houy, W., Mankovich, J. A., McGuiness, L., Orlewicz, E., Paskind, M., Pratt, C. A., Reis, P., Summani, A., Terranova, M., Welch, J. P., Xiong, L., Mo¨ller, A., Tracey, D. E., Kamen, R., and Wong, W. W. (1994) Cell 78, 343–352.
66
VAN CRIEKINGE ET AL.
2. Wilson, K. P., Black, J. A., Thomson, J. A., Kim, E. E., Griffith, J. P., Navia, M. A., Murcko, M. A., Chambers, S. P., Aldape, R. A., Raybuck, S. A., and Livingston, D. J. (1994) Nature 370, 270 –275. 3. Cohen, G. M. (1997) Biochem. J. 326, 1–16. 4. Van de Craen, M., Vandenabeele, P., Declercq, W., Van den brande, I., Van Loo, G., Molemans, F., Schotte, P., Van Criekinge, W., Beyaert, R., and Fiers, W. (1997) FEBS Lett. 403, 61– 69. 5. Earnshaw, W. C., and Martins, L. M. (1997) Trends Cell Biol. 7, 111–114. 6. Porter, A. G., Ng, P., and Janicke, R. U. (1997) BioEssays 19, 501–507. 7. Fields, S., and Song, O. K. (1989) Nature 340, 245–246. 8. Chien, C. T., Bartel, P. L., Sternglanz, R., and Fields, S. (1991) Proc. Natl. Acad. Sci. USA 88, 9578 –9582. 9. Fields, S., and Sternglanz, R. (1994) Trends Genet. 10, 286 – 292. 10. Luban, J., and Goff, S. P. (1995) Curr. Opin. Biotechnol. 6, 59 – 64. 11. Brachmann, R. K., and Boeke, J. D. (1997) Curr. Opin. Biotechnol. 8, 561–568.
12. Ray, C. A., Black, R. A., Kronheim, S. R., Greenstreet, T. A., Sleath, P. R., Salvesen, G. S., and Pickup, D. J. (1992) Cell 69, 597– 604. 13. Bump, N. J., Hackett, M., Hugunin, M., Seshagiri, S., Brady, K., Chen, P., Ferenz, C., Franklin, S., Ghayur, T., Li, P., Licari, J., Mankovich, L., Shi, A., Greenberg, H., Miller, L. K., and Wong, W. W. (1995) Science 269, 1885–1888. 14. Van Criekinge, W., Beyaert, R., Van de Craen, M., Vandenabeele, P., Schotte, P., De Valck, D., and Fiers, W. (1996) J. Biol. Chem. 271, 27245–27248. 15. Legrain, P., Chapon, C., and Galisson, F. (1993) Genes Dev. 7, 1390 –1399. 16. Estojak, J., Brent, R., and Golemis, E. A. (1995) Mol. Cell. Biol. 15, 5820 –5829. 17. Ghayur, T., Banerjee, S., Hugunin, M., Butler, D., Herzog, L., Carter, A., Quintal, L., Sekut, L., Talanian, R., Paskind, M., Wong, W., Kamen, R., Tracey, D., and Allen, H. (1997) Nature 386, 619 – 623. 18. Zhang, J., and Lautar, S. (1996) Anal. Biochem. 242, 68 –72. 19. Mashima, T., Naito, M., Fujita, N., Noguchi, K., and Tsuruo, T. (1995) Biochem. Biophys. Res. Commun. 217, 1185–1192.