Oncogene (2000) 19, 4706 ± 4712 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Identi®cation and characterization of JunD missense mutants that lack menin binding

JI Knapp1, C Heppner1, AB Hickman2, AL Burns1, SC Chandrasekharappa3, FS Collins3, SJ Marx1, AM Spiegel1 and SK Agarwal*,1

1Metabolic Diseases Branch, NIDDK, NIH, Bethesda, Maryland 20892, USA; 2Laboratory of Molecular Biology, NIDDK, NIH, Bethesda, Maryland 20892, USA; 3Genetics and Molecular Biology Branch, NHGRI, NIH, Bethesda, Maryland 20892, USA

Menin, the product of the MEN1 tumor suppressor gene, 1991). All family members share a well-conserved basic binds to the AP1 transcription factor JunD and represses DNA binding domain and a leucine zipper dimeriza- JunD transcriptional activity. The e€ects of human or tion motif (bZIP). Dimerization is required for binding mouse JunD missense mutations upon menin interaction to a speci®c DNA sequence known as the TPA (12-O- were studied by random and alanine scanning mutagen- tetradecanoylphorbol-13-acetate) responsive element esis of the menin binding region of JunD (amino acids (TRE). TRE sequences have been identi®ed in the 1 ± 70). JunD mutant proteins were tested for menin regulatory regions of many genes important for cell binding in a reverse yeast two-hybrid assay, and for proliferation, e.g., collagenase, stromelysin, metal- transcriptional regulation by menin in AP1-reporter lothionin IIA, TGF b (Angel et al., 1987; Lee et al., assays. Random mutagenesis identi®ed two di€erent 1987; Kerr et al., 1990, Kim et al., 1990) and in the mutations that disrupted menin interaction at mouse enhancers and promoters of numerous other genes JunD amino acid 42 (G42E and G42R). Mutation G42A (Wisdom et al., 1999). generated by alanine scanning did not a€ect menin JunD is the only AP1 family member to which binding, likely re¯ecting the conserved nature of this menin binds directly (Agarwal et al., 1999). JunD has amino acid substitution. Furthermore, by size exclusion other characteristics di€erent from those of the chromatography menin co-migrated with wild type JunD remaining jun proteins. JunD is the only AP1 protein but not with the JunD mutant tested (G42E). Alanine upregulated during cell quiescence (Pfarr et al., 1994). scanning mutagenesis of residues 30 ± 55 revealed two JunD levels are higher in cells in the G0/G1 phases of di€erent amino acids, P41 and P44, of mouse JunD that the cell cycle (Pfarr et al., 1994; Lallemand et al., were critical for interaction with menin. Mouse JunD 1997). When overexpressed, JunD slows growth in missense mutants P41A, G42R, G42E and P44A failed various cell lines (Metivier et al., 1993; Pfarr et al., to bind menin and also escaped menin's control over their 1994; Lallemand et al., 1997), and also antagonizes transcriptional activity. At lower amounts of transfected growth in RAS-transformed cells (Pfarr et al., 1994). In menin, the transcriptional e€ect of menin on the mutants contrast, c-jun is an immediate early gene, promoting P41A, G42R and G42E was changed from repression to entry into S/G2 and M phases of the cell cycle (Pfarr et activation, similar to that with c-jun. In conclusion, a al., 1994; Lallemand et al., 1997), can transform a small N-terminal region of JunD mediates a key variety of cells when overexpressed (Metivier et al., di€erence between JunD and c-jun, and a component 1993), and when overexpressed acts cooperatively with of this di€erence is dependent on JunD binding to menin. RAS to transform cells (Alani et al., 1991; Pfarr et al., Oncogene (2000) 19, 4706 ± 4712. 1994). Homozygous disruption of c-jun or JunB results in embryonic lethality (Hilberg et al., 1993; Johnson et Keywords: MEN1; menin; JunD; mutagenesis; two- al., 1993; Schropp-Kistner et al., 1999), while JunD7/7 hybrid; tumor suppressor mice are viable (Thepot et al., 2000). Additionally, menin represses JunD transcriptional activity, whereas it augments the transcriptional activity of c-jun Introduction (Agarwal et al., 1999). The jun proteins share a high degree of homology Menin, the protein encoded by the MEN1 (multiple except in their N-terminal 95 amino acid region (Berger endocrine neoplasia type 1) tumor suppressor gene, and Shaul, 1991). It is also within this region (amino resides mainly in the nucleus (Guru et al., 1998) and acids 8 ± 70 of JunD) that JunD binds menin (Agarwal interacts with JunD, a member of the AP1 (activation et al., 1999). Evidence that the 8 ± 70 amino acid region protein 1) family of transcription factors (Agarwal et of JunD is critical for its putative tumor suppressor al., 1999; Gobl et al., 1999). The AP1 family consists of function includes a study which demonstrates that both jun (JunD, JunB, c-jun and v-jun) and fos (c-Fos, chimeric JunD proteins containing amino acid 1 ± 79 of FosB, Fra-1 and Fra-2) proteins (Angel and Karin, c-jun develop the ability to transform cells when overexpressed (Metivier et al., 1993). Thus, it is conceivable that the amino acid 1 ± 79 portion contains unique binding sites and/or properties that confer upon *Correspondence: SK Agarwal, National Institutes of Health, c-jun the ability to transform cells. Similarly, the ability Building 10, Room 9C-101, 10 Center Dr., MSC 1802, Bethesda, of JunD to antagonize cell growth and transformation MD 20892-1802, USA Received 6 March 2000; revised 27 July 2000; accepted 1 August could depend upon its ability to bind menin in this 2000 region. Interaction of JunD mutants with menin JI Knapp et al 4707 MEN1 is a disorder characterized by various and fused in-frame with the Gal4 activation domain combinations of tumors of parathyroid, anterior (Gal4AD). A pooled plasmid library of mutated JunD pituitary, enteropancreatic, foregut carcinoid, skin was screened with Gal4 DNA binding domain-menin and other tissues (Marx, 1998). It is caused by fusion (Gal4DBD-menin) as the bait using a reverse heterozygous inactivating germline mutations in the yeast two-hybrid screen (Vidal et al., 1996). This MEN1 tumor suppressor gene located on chromosome method uses a yeast strain (e.g. MaV203) that 11q13 and the subsequent inactivation of the normal possesses a URA3 gene under the control of GAL4 allele in the tumor tissue (Chandrasekharappa et al., upstream activating sequences (UAS) in addition to the 1997; Agarwal et al., 1997; Lemmens et al., 1997; GAL4-LacZ reporter gene. Addition of 5-¯uoroorotic Bartsch et al., 1998; Bassett et al., 1998; Giraud et al., acid (5-FOA) is toxic to yeast expressing the URA3 1998; Mutch et al., 1999; Poncin et al., 1999; Teh et al., gene product. Therefore, in a two-hybrid assay in the 1998; Hai et al., 1999). MEN1 mutations also occur as presence of 5-FOA, cells expressing a JunD or JunD somatic mutations accompanied by 11q13 LOH in mutant protein that interacts with menin will die sporadic MEN1-like tumors including gastrinomas, whereas interaction-defective clones will survive. This insulinomas, parathyroid adenomas, bronchial carci- allowed for the selection of JunD mutants that did not noid tumors, and angio®bromas (reviewed in Marx et interact with menin. Non-interaction was also con- al., 1999). ®rmed by the presence of white colonies in the b- Overexpressed menin can partially revert the proper- galactosidase assay in which the LacZ reporter was not ties of RAS-transformed NIH3T3 cells (Kim et al., activated due to lack of mutant JunD and menin 1999). This supports the tumor suppressor role for interaction. DNA sequence changes that lead to menin. disruption of menin interaction were obtained by Several naturally occurring menin missense muta- isolating the GAL4AD-fusion plasmids from the yeast tions, speci®cally in the region from amino acids 139 ± clones and sequencing. The presence of a Gal4AD- 242, interfere with its ability to bind JunD, suggesting fused wild type or mutant JunD protein in yeast was that the menin-JunD interaction may be crucial to analysed by Western blot analysis using an anti- menin's tumor suppressor function (Agarwal et al., Gal4AD antibody (data not shown). Some of the 1999). To understand further the importance of the fusion proteins were truncated due to nonsense codons menin-JunD interaction, we sought to identify JunD or deletions. Only those plasmids that encoded protein mutants that lack menin binding and to compare the of the expected size with likely missense mutation(s) e€ect of menin on the transcriptional activity of JunD were further analysed. Three independent plasmids that mutants that bind or that do not bind menin. expressed protein of the appropriate size had mutations in human JunD at amino acid residue 34 (G34R, G34E) along with one other missense mutation in each insert (P37S, A40T, and M44I respectively). These two Results human JunD G34 substitutions were engineered into the corresponding amino acid G42 (Figure 1) in full Random mutagenesis and screening of human JunD amino length mouse JunD via site-directed mutagenesis. acids critical for menin binding Because full length human JunD as PCR template The menin-interacting region of human JunD (amino rarely yields product, mouse JunD was used for further acids 1 ± 70) was randomly mutated by forced-PCR analysis. Full length mouse JunD possessing mutations

Figure 1 JunD protein and amino acid homology. (a) Structural organization of full length JunD (aa 1 ± 347 for human) is shown. The menin binding domain is shown as a grey shaded box (aa 8 ± 70). The blackened boxes denote the three transactivation regions (TA) in JunD (aa 2 ± 9, 95 ± 108, and 132 ± 146). The cross-hatched area (aa 262 ± 289) indicates the basic domain. The slashed box shows the leucine zipper (ZIP) domain (aa 290 ± 325). (b) ClustalW alignment of the N-terminal amino acids of JunD (mouse, human and rat) and mouse c-jun. Identical amino acids are marked with an asterisk and conserved amino acids are marked with aà . Note the paucity of homology between JunD and c-jun within their ®rst 70 amino acids. Amino acids in boldface were mutated using site-directed mutagenesis. The human counterpart of the underlined mouse JunD amino acid G42 (G34 in human), was found altered to G42E and G42R via random mutagenesis and reverse yeast two-hybrid selection. Missense mutations at amino acids labeled with a boldfaced `+' caused lack of interaction with menin (P41A, G42R, G42E, P44A)

Oncogene Interaction of JunD mutants with menin JI Knapp et al 4708 G42R and G42E did not interact with menin in the Effect of menin on transcriptional activity of JunD yeast two-hybrid analysis. mutants Each of the 15 JunD mutants engineered into Alanine scanning mutagenesis of mouse JunD amino acids pcDNA3.1 mouse JunD (13 alanine mutants and the critical for menin binding two mutants identi®ed by random mutagenesis) was Multiple sequence alterations were present in indivi- analysed for its ability to activate transcription from an dual fragments generated by random mutagenesis. This AP1-reporter plasmid (pAP1, containing seven copies made it dicult to discern which mutation/s was of TGACTAA as the consensus TRE upstream of a responsible for disruption of menin interaction. Since minimal TATA and luciferase as the reporter). the random mutagenesis method had identi®ed mouse HEK293 cells were transiently contransfected with JunD amino acid G42 as important for menin binding, pAP1 and a jun expression plasmid (JunD, c-jun or the region ¯anking residue 42 was studied by alanine JunD mutants), with or without menin expression scanning to identify any additional mouse JunD amino plasmid pCMV sportMenin (Figure 3a). All of the acids that could be required for menin binding. Most JunD mutants retained their ability to activate (13 out of 16) of the mouse JunD amino acids between transcription, albeit with signi®cant variations in their residues 30 ± 55 unique to JunD when compared to c- activity (0.2 ± 1.2-fold) compared to wild type JunD jun and conserved between mouse, human and rat indicating that regions outside the three transcriptional JunD, were included in the alanine scanning (Figure 1). activation domains (aa 2 ± 9, 95 ± 108, 132 ± 146) may Thirteen amino acids were changed to alanine using also be important for transcriptional activation. As site-directed mutagenesis of mouse JunD (Figure 1). Yeast two-hybrid analysis showed that eleven of the thirteen alanine mutants were capable of binding menin, as assessed by their inability to survive in 5- FOA containing media and blue colonies in b- galactosidase assay. P41A and P44A did not interact with menin. In contrast to the two non-interacting mutations at G42 previously identi®ed by random mutagenesis (G42R and G42E), the G42 to alanine (G42A) mutation did not inhibit menin interaction. This presumably re¯ects the conservative nature of the glycine to alanine change. Expression of each of the four menin non-interacting JunD mutant proteins in yeast cells was con®rmed by Western blot analysis (data not shown).

In vitro binding assays between menin and JunD mutants To con®rm the lack of interaction observed in the yeast two-hybrid assay between menin and the 4 JunD mutants P41A, G42R, G42E and P44A, in vitro binding was tested by FLAG pull-down assays (Figure 2a). In vitro translated (IVT) wild type or mutant JunD labeled with 35S-methionine were tested for interaction with FLAG-menin or FLAG-BAP (negative control) coupled to anti-FLAG beads. In agreement with the results obtained above with the yeast two-hybrid Figure 2 In vitro analysis of mouse JunD (wild type or mutant) experiments, wild type JunD and the mutant G42A interaction with menin. (a) FLAG pull-down assays. FLAG- were retained by FLAG-menin and not FLAG-BAP, menin or FLAG-BAP coupled to anti-FLAG agarose beads was 35 while the 4 JunD mutants were not retained by FLAG- incubated with S-methionine labeled wild type or mutant JunD. Wild type JunD and the JunD mutant G42A were retained by menin or FLAG-BAP. The integrity of the wild type FLAG-menin and not FLAG-BAP, while four JunD mutants and the mutant IVT products was also tested for failed to bind FLAG-menin or FLAG-BAP. Input lane contained dimerization to JunD by GST pull-down assays. Each 10% of IVT material incubated with the beads. (b) and (c) Size of the wild type and mutant JunD proteins was exclusion chromatography. Recombinant puri®ed menin was incubated with GST-wtJunD(1 ± 104) or GST-G42E(1 ± 104), and retained by GST-JunD but not by GST alone (data loaded onto a Superdex 200 column. The column was run at not shown). 50 ml/min and 50 ml fractions were collected. The fractions were Recombinantly-expressed and puri®ed menin and analysed by SDS ± PAGE. The fraction numbers (20 ± 29) indicate JunD proteins, GST-wtJunD (1 ± 104) or GST-G42E elution time in minutes. The void volume of the column is (1 ± 104), were studied by size-exclusion chromatogra- 0.92 ml, corresponding to fraction 18. Unbound GST-wtJunD(1 ± 104) and GST-G42E(1 ± 104) elute at about 24.5 min, correspond- phy on a Superdex 200 column. Menin and GST- ing to an approximate molecular weight of 300 kDa. Unbound wtJunD (1 ± 104) co-migrated in the same fractions menin elutes at about 27.5 min, corresponding to a molecular (Figure 2b) while menin and GST-G42E (1 ± 104) weight of about 145 kDa. The formation of a menin-JunD eluted as two distinct peaks (Figure 2c), indicating complex is indicated by a shift in migration position of both proteins to a higher molecular weight species eluting before that a single point mutation is sucient to decrease or 24 min in (b). This shift is not seen with the G42E point mutant eliminate binding in solution between menin and (c). The numbers on the left of both panels are SDS ± PAGE wtJunD (1 ± 104). molecular weight markers (kDa) that were loaded in the ®rst lane

Oncogene Interaction of JunD mutants with menin JI Knapp et al 4709

Figure 4 Menin dose response on transcriptional activity of mouse JunD mutants P41A, G42A, G42E, G42R and P44A. (a) 293 cells were transiently transfected with 0.5 mg of the luciferase reporter plasmid pAP1, 0.5 mg of plasmid expressing either c-jun, Figure 3 Functional analyses of mouse JunD mutants. (a) E€ect JunD or JunD mutants alone (labeled 1) or together with various of menin on transcriptional activity of c-jun, JunD and JunD amounts of menin expression plasmid pCMVsportMenin (0.2, 0.6, mutants. The luciferase reporter plasmid pAP1 (0.5 mg) was 1 and 2 mg) (labeled 2 ± 5 respectively). The empty vector cotransfected into 293 cells with 0.5 mg of plasmid expressing pcDNA3.1 was added to maintain equal amount of DNA either c-jun, JunD or JunD mutants alone (open bars) or together transfected. Results from a representative experiment are shown with 2 mg of menin expression plasmid pCMV sportMenin (®lled as average arbitrary luciferase units with standard error. Bars bars). The empty vector pcDNA3.1 was added to maintain equal corresponding to wild type JunD plus menin (4 and 5) are not amount of DNA transfected. Results from a representative seen since they are below the resolution of this ®gure. (b) Levels experiment are shown as average arbitrary luciferase units with of expression of menin, JunD and JunD mutants in a. Equal standard error. (b) Levels of expression of JunD. Equal amounts amounts (30 mg) of total cell lysates were analysed by Western (30 mg) of total cell lysates were analysed by Western blot analysis blot analysis using anti-menin antibody (SQV) or anti-JunD using an anti-JunD antibody (SantaCruz). The faster migrating antibody. The order of samples on the blot corresponds to the band below 45 kDa corresponds to an alternate form of JunD order in which the data are presented in the graph above. Lane generated due to an internal initiation of translation (Okazaki et labeled `0' contains lysate from untransfected 293 cells al., 1998). Lane labeled `293' contains lysate from untransfected 293 cells previously reported, menin coexpression inhibited wild repressed to levels below their baseline activity. Menin type JunD-mediated transcription but augmented c-jun- was signi®cantly less e€ective in repressing the activity mediated transcription (Agarwal et al., 1999). Menin of the mutant P44A compared to wild type JunD. As repressed the transcriptional activity of 11 of the JunD observed for wild type JunD, the transcriptional mutants that retained menin binding activity, while activity of the mutant G42A which can interact with menin did not e€ectively repress the activity of the four menin was repressed with increasing amounts of menin-interaction de®cient mutants (P41A, G42R, transfected menin, but the level of repression was G42E, P44A). Transactivation from the G42A mutant, lower than that for wild type JunD. The level of that retained its ability to bind menin, was repressed protein expression of JunD and JunD mutants detected moderately by menin. All of the JunD mutant proteins by Western blot was una€ected by changes in amounts were expressed at equal levels when transfected alone of cotransfected menin (Figure 4b). or together with menin (Figure 3b). Having identi®ed four JunD mutants whose tran- scriptional activities were not e€ectively repressed by Discussion menin (P41A, G42R, G42E and P44A), we tested the responses of the mutants to various amounts of The N-terminal region of JunD is necessary for cotransfected menin in 293 cells (Figure 4a). We interaction with menin and for menin-mediated inhibi- expected the transcriptional activity of the mutants to tion of JunD transcriptional activity (Agarwal et al., increase with increasing concentrations of menin as 1999). So far there are no known naturally occurring observed for c-jun. This, however, was not the case. JunD mutations that disrupt menin-JunD binding and When 293 cells were transfected with pAP1 and mutant cause MEN1 or associated spontaneously occurring JunD expressing plasmid, cotransfection of low tumors. amounts of menin expression plasmid caused an initial In the present study, we have employed two increase in transcriptional activity of P41A, G42R, and complementary strategies, random mutagenesis and G42E. With higher amounts of menin expression alanine scanning, to identify some amino acids in the plasmid, the augmentation e€ect of menin was N-terminal region of JunD responsible for JunD diminished (Figure 4a). Despite this diminished interaction with menin. Mouse JunD amino acid augmentation, the transcriptional activities of three residues P41, G42 and P44 were found to be critical mutants (P41A, G42R, and G42E) were never actually for JunD interaction with menin. Certain mutations

Oncogene Interaction of JunD mutants with menin JI Knapp et al 4710 (P41A, G42R, G42E and P44A) of these residues thus producing transcriptional augmentation. In the disrupted menin binding. In functional assays, trans- presence of higher amounts of menin, the JunD activation from three of the non-interactors (P41A, dependent e€ects may take over and menin is unable G42R and G42E) was not inhibited by menin. This to augment transcriptional activity. In this regard, the result con®rms that these mutations prevent e€ective JunD mutants (that lack menin interaction) might JunD binding to menin. In addition, similar to the e€ectively be Km mutants which present JunD e€ect of menin upon c-jun, cotransfection of lower dependent e€ects of menin in the presence of very amounts of menin increased the transcriptional activity large amounts of menin. This notion is supported by of three JunD mutants that lacked menin binding. the behavior of the P44A mutant which could be more P44A did not show interaction in yeast or in vitro, and sensitive to menin concentration and likely capable of its transcriptional activity was repressed less e€ectively interacting with menin at higher concentrations of by menin. The conservative change of mouse JunD menin. JunD may well have menin-independent glycine 42 to alanine did not disrupt menin interaction functions, some of which may be relevant to its role or transcriptional repression by menin. We cannot in oncogenesis. The mutant JunD proteins described in exclude the possibility that amino acid substitutions this report could prove useful in separating the menin- other than the alanines we created might a€ect menin dependent and -independent roles of JunD. binding, or menin's ability to regulate transcriptional In conclusion, we have discovered critical amino activity. It is perhaps important that the critical JunD acids within the N-terminus of JunD that are essential region for menin interaction (mouse amino acids 41 ± for menin binding and for menin's e€ect on the 44), identi®ed herein is conserved among the JunD regulation of transcription. This information will be proteins and absent in c-jun (Figure 1b). The crystal useful in determining whether elimination of the JunD- structure of the N-terminal region of JunD or c-jun has menin interaction is enough to cause cellular transfor- not been deciphered, therefore the variability in mation and perhaps even cause tumors. These mutant transcriptional activity of the mutants compared to JunD proteins will be useful in understanding the wild type JunD cannot be related to their positional importance of the menin-JunD interaction for the anti- e€ect on the transcriptional activation domains. mitogenic e€ects of JunD and also for the tumor c-jun and JunD have been reported to play at times suppressor function of menin. opposite roles in cell proliferation and transformation with overexpression of JunD counteracting and c-jun overexpression promoting cell transformation (Caste- Materials and methods lazzi et al., 1991; Pfarr et al., 1994; Lallemand et al., 1997). The JunD mutants described here (except P44A) Random mutagenesis and yeast reverse two-hybrid library that have lost menin interaction also show augmented screening transactivation in the presence of menin (like c-jun). The N-terminal 70 amino acid region of human JunD was We have proposed that menin and JunD could ampli®ed by forced PCR (Fromant et al., 1995) using synergistically exert their tumor suppressor function pACT2-JunD (1 ± 70) (Agarwal et al., 1999) as template via menin-JunD interaction (Agarwal et al., 1999). and Taq gold (Perkin Elmer, Foster City, CA, USA). Brie¯y, Since the mutant JunD N-terminus has lost the ability to decrease the ®delity of the Taq polymerase and induce to interact directly with menin and therefore is not base mismatches, a ®nal concentration of 9.5 mM MgCl2 and repressed by menin, it may have gained c-jun-like 0.5 mM MnCl2 were added to a 50 ml PCR reaction mix. With the aforementioned conditions, four di€erent PCR properties of cell transformation and proliferation. In reactions were performed, each containing 0.2 mM of dNTPs preliminary experiments, we therefore generated and 3.4 mM of one of the following: dATP, dTTP, dCTP, or NIH3T3 stable cell lines with two di€erent JunD dGTP. Thermocycler conditions were as follows: 928C for mutants P41A and G42R. The levels of expression of 10 min, 30 PCR cycles (948C for 30 s, 608C for 30 s, 728C for mutant JunD obtained in the stable cells was increased 1 min), and a ®nal extension for 5 min at 728C. PCR only two fold. No signi®cant di€erence in morphology, products were shotgun cloned into the pACT2 yeast vector proliferation or colony formation in soft agar was using the restriction sites incorporated in the primer observed in three di€erent stable cell lines transfected sequences. Transformants were pooled as a library, and with each of the mutants P41A or G42R when plasmid was puri®ed with the Maxi Prep Kit (Qiagen, compared to wild type NIH3T3 cells (not shown). It Valencia, CA, USA). MaV203 yeast competent cells (Life Technologies, Gaithersburg, MD, USA) were co-transformed will be important to perform further analysis with with 1 mg each of puri®ed pooled plasmid DNA and pAS2-1 stable cell lines that express higher amounts of menin (menin fused to the Gal4 DNA binding domain) as transfected protein. per manufacturer's protocol. Transformed yeast was plated In the present study, the e€ect of menin on JunD on -trp-leu+5FOA agar plates to select for colonies in which mutants vs c-jun, both of which lack direct menin the mutant JunD and menin did not interact. b-galactosidase interaction, di€ered in that increasing amounts of assays (Clontech, Palo Alto, CA, USA) were performed to transfected menin induced a linear increase in tran- con®rm the lack of interaction. Random colonies growing on scriptional activity with c-jun, this was not the case -trp-leu+5FOA agar plates were used for plasmid isolation with the JunD mutants. These results raise questions (Stratagene, La Jolla, CA, USA), and were sequenced to regarding the role of menin in transcriptional regula- identify the mutations. Western blot was performed as described (Agarwal et al., 1999) to con®rm expression of tion. Menin could have both JunD dependent and the Gal4AD-mutant JunD fusion proteins. JunD independent e€ects on transcription as evidenced by its di€erent e€ects on wild type JunD and c-jun activity. In the case of the JunD mutants that lack Site-directed mutagenesis and alanine scanning menin interaction, the JunD independent menin e€ects Fifteen sequence alterations ± G30A, G31A, F33A, G37A, may dominate in the presence of low menin amounts, P41A, G42A, G42E, G42R, P44A, P45A, T46A, S48A,

Oncogene Interaction of JunD mutants with menin JI Knapp et al 4711 K51A, D53A, L55A ± were generated in full length mouse anion exchange chromatography. The His-tag was removed JunD using pcDNA3.1-mouse-JunD as template with the by thrombin cleavage at 48C (Hickman et al., 1999) using 10 Quikchange Mutagenesis kit (Stratagene). Two of these units thrombin (Sigma) per mg of menin. mutants, G42R and G42E, were originally found through The N-terminal 104 amino acids of mouse JunD or mouse the above library screen in human JunD (G34R and G34E), JunD mutant G42E were PCR ampli®ed and cloned in-frame and were engineered into corresponding amino acid G42 of with the C-terminus of GST into pGEX3X, and called GST- mouse JunD. All primer sequences are available upon wtJunD(1 ± 104) and GST-G42E(1 ± 104). The GST-fusion request. Mutagenesis was performed as per the manufac- proteins were expressed in BL21(DE3) cells and puri®ed by turer's protocol. DNA was sequenced with the AmpliCycle anity chromatography (Hickman et al., 1999) on glu- Sequencing Kit (Perkin Elmer), using 33P-end labeled tathione Sepharose 4B (Pharmacia). primers, to ensure that that the proper mutations had been The puri®ed proteins were dialyzed against 20 mM Tris created. pH 7.5, 0.5 M NaCl, 1 mM DTT, 1 mM EDTA, 10% (w/v) glycerol prior to size-exclusion chromatography. The puri®ed GST-wtJunD(1 ± 104) or GST-G42E(1 ± 104) proteins at Yeast two-hybrid analysis of JunD mutants approximately 1 mg/ml were mixed with a 1.5 molar excess For yeast two-hybrid interaction with Gal4DBD-menin of menin and incubated at 48C for 30 min. The mixture was fusion, each of the 15 site-directed mutated JunD inserts applied to an analytical gel ®ltration Superdex 200 PC 3.2/ was PCR ampli®ed with appropriate primers and cloned in- 30 column (Pharmacia) run at 48C. The eluted fractions frame with the Gal4-AD into the pACT2 yeast two-hybrid were analysed by SDS ± PAGE on 4 ± 12% NuPAGE gels vector. The same procedure used above for randomly (Novex). mutated human JunD was used to test expression and interaction of the site-directed mouse JunD mutants with Cell culture menin. HEK293 cells were grown in DMEM (Dulbecco's modi®ed eagle medium) supplemented with 10% fetal calf serum and FLAG and GST pull-down assays 2 mM glutamine, 100 mg/ml penicillin-streptomycin at 378C Menin was fused at its N-terminus with the FLAG epitope and 5% CO2. Transient transfection was carried out with by subcloning the 610 amino acid menin coding region into Superfect reagent (Qiagen) as per the manufacturer. the vector FLAG-MAC (Sigma, St. Louis, MO, USA). The plasmid construct expressing FLAG epitope-tagged bacterial AP1 reporter assays alkaline phosphatase (FLAG-BAP) was purchased from Sigma as a negative control. FLAG-menin or FLAG-BAP HEK293 cells were seeded in 6-well plates 24 h prior to fusion protein was puri®ed on anti-FLAG M2-agarose beads transfection and grown to 60 ± 75% con¯uency. Cells were (Sigma) as per the manufacturer in bu€er containing 16 transfected with 0.5 mg of the reporter plasmid pAP1, 0.5 mg TBS, 10% glycerol, 0.1% Igepal (Sigma) and complete of either pcDNA3.1 mouse JunD, pcDNA3.1 mouse JunD protease inhibitor cocktail (Boehringer Mannheim, Ger- mutants, or pcDNA3.1 c-jun, and 2 mg of either pCMVsport- many). Menin or empty pCMVsport vector. Cells were lysed 2 days In vitro translated (IVT) wild type and mutant JunD post-transfection and Luciferase activity was measured with proteins were made using the TNT T7 Coupled Reticulocyte Luciferase Assay Reagent as per the manufacturer (Prome- Lysate System (Promega, Madison, WI, USA) in the presence ga). The results were normalized to correct for di€erences in of [35S]Methionine. Puri®ed FLAG-menin and FLAG-BAP protein content. Results from each experiment were plotted beads (2 mg each) were incubated with 8 ml of IVT mouse as average luciferase activity with standard error. JunD or the ®ve mutants with rotation at 48C overnight. To study the dose response of menin on c-jun, JunD or Beads were washed eight times in wash bu€er (500 mM NaCl, JunD mutant activity, the same protocol as above was 0.1% Igepal, 20 mM Tris-HCl pH 8.0), with centrifugation followed, with varying amounts of pCMVsportMenin ± 0.2, (3000 r.p.m. for 5 min) between each wash. Beads were 0.6, 1 and 2 mg. resuspended in 30 ml of SDS ± PAGE loading bu€er and Equivalent levels of DNA in all transfections were heated for 10 min at 958C. Samples (30 ml) were separated in maintained with an appropriate amount of empty 10% Tris-Glycine gels (Novex, San Diego, CA, USA). Input pCMVsport or pcDNA3.1 vector. All transfections were lanes of IVT protein contained 10% of the amount incubated performed at least twice in duplicate or triplicate with at least with the beads. The gels were removed on Whatman ®lter two independent DNA preparations. paper, dried and exposed to X-ray ®lm. Equal amounts of cell lysate were analysed by Western blot GST pull-down assays were performed as described with anti-JunD (Santa Cruz Biotechnology, Santa Cruz, CA, (Agarwal et al., 1999). Radiolabeled IVT JunD, or JunD USA) and anti-menin (SQV) antibodies (Guru et al., 1998), mutants were incubated with GST alone or GST-JunD to compare expression levels. coupled to glutathione sepharose beads and processed as for the FLAG pull-down assay.

Co-migration analysis Menin coding region was cloned in-frame with the C- terminus of a six histidine tag in the EpiTag/His vector Acknowledgments (Invitrogen, Carlsbad, CA, USA). His-tagged menin was We are grateful to Dr Regina Collins for support with cell expressed in E. coli BL21(DE3) cells and puri®ed by anity culture. We thank Dr George Poy for DNA sequencing. chromatography on Chelating Sepharose (Amersham Phar- We also thank Sylvia Park and Karl Bilimoria for technical macia, Piscataway, NJ, USA), according to the manufac- protocols. J Knapp is a Howard Hughes Medical Institute turer's protocols. Further puri®cation was carried out by ± National Institutes of Health Scholar.

Oncogene Interaction of JunD mutants with menin JI Knapp et al 4712 References

Agarwal SK, Guru SC, Heppner C, Erdos MR, Collins RM, Kim SJ, Angel P, Lafyatis R, Hattori K, Kim KY, Sporn Park SY, Saggar S, Chandrasekharappa SC, Collins FS, MB, Karin M and Roberts AB. (1990). Mol. Cell. Biol., Spiegel AM, Marx SJ and Burns AL. (1999). Cell, 96, 10, 1492 ± 1497. 143 ± 152. Kim YS, Burns AL, Goldsmith PK, Heppner C, Park SY, Agarwal SK, Kester MB, Debelenko LV, Heppner C, Chandrasekharrapa SC, Collins FS, Spiegel AM and Emmert-Buck MR, Skarulis MC, Doppman JL, Kim Marx SJ. (1999). Oncogene, 18, 5936 ± 5942. YS, Lubensky IA, Zhuang Z, Green JS, Guru SC, Lallemand D, Spyrou G, Yaniv M and Pfarr CM. (1997). Manickam P, Olufemi S, Liotta LA, Chandrasekharappa Oncogene, 14, 819 ± 830. SC, Collins FS, Spiegel AM, Burns AL and Marx SJ. Lee W, Mitchell P and Tjian R. (1987). Cell, 49, 741 ± 752. (1997). Hum. Mol. Genet., 6, 1169 ± 1175. LemmensI,VandeVenWJ,KasK,ZhangCX,GiraudS, AlaniR,BrownP,BinetruyB,DosakaH,RosenbergRK, Wautot V, Buisson N, De Witte K, Salandre J, Lenoir G, Angel P, Karin M and Birrer MJ. (1991). Mol. Cell Biol., Pugeat M, Calender A, Parente F, Quincey D, Gaudray P, 11, 6286 ± 6295. De Wit MJ, Lips CJ, Hoppener JW, Khodaei S, Grant AL, Angel P, Baumann I, Stein B, Delius H, Rahmsdorf HJ and Weber G, Kytola S, Teh BT, Farnebo F, Phelan C, Herrlich P. (1987). Mol. Cell. Biol., 7, 2256 ± 2266. Hayward N, Larsson C, Pannett AAJ, Forbes SA, Duncan Angel P and Karin M. (1991). Biochim. Biophys. Acta., 1072, Bassett JH and Thakker RV. (The European Consortium 129 ± 157. on MEN1). (1997). Hum. Mol. Genet., 6, 1177 ± 1183. Bartsch D, Kopp I, Bergenfelz A, Rieder H, Munch K, Jager Okazaki S, Ito T, Ui M, Watanabe T, Yoshimatsu K and Iba K, Deiss Y, Schudy A, Barth P, Arnold R, Rothmund M H. (1998). Biochem. Biophys. Res. Commun., 250, 347 ± and Simon B. (1998). Eur. J. Endocrinol., 139, 416 ± 420. 353. Bassett JH, Forbes SA, Pannett AA, Lloyd SE, Christie PT, Marx SJ. (1998). In: The Genetic Basis of Human Cancer. Wooding C, Harding B, Besser GM, Edwards CR, Vogelstein B & Kinzler KW eds. McGraw Hill: New York, Monson JP, Sampson J, Wass JA, Wheeler MH and pp. 489 ± 506. Thakker RV. (1998). Am.J.Hum.Genet.,62, 232 ± 244. Marx SJ, Agarwal SK, Kester MB, Heppner C, Kim YS, Berger I and Shaul Y. (1991). Oncogene, 6, 561 ± 566. Skarulis MC, James LA, Goldsmith PK, Saggar SK, Park CastellazziM,SpyrouG,LaVistaN,DangyJP,PiuF, SY, Spiegel AM, Burns AL, Debelenko LV, Zhuang Z, Yaniv M and Brun G. (1991). Proc. Natl. Acad. Sci. USA, Lubensky IA, Liotta LA, Emmert-Buck MR, Guru SC, 88, 8890 ± 8894. Manickam P, Crabtree J, Erdos MR, Collins FS and Chandrasekharappa SC, Guru SC, Manickam P, Olufemi S, Chandrasekharappa SC. (1999). Recent Prog. Horm. Res., Collins FS, Emmert-Buck MR, Debelenko LV, Zhuang Z, 54, 397 ± 439. Lubensky IA, Liotta LA, Crabtree JS, Wang Y, Roe BA, Metivier C, Piu F, Pfarr CM, Yaniv M, Loiseau L and Weisemann J, Boguski MS, Agarwal SK, Kester MB, Kim Castellazzi M. (1993). Oncogene, 8, 2311 ± 2315. YS, Heppner C, Dong Q, Spiegel AM, Burns AL and Mutch MG, Dilley WG, Sanjurjo F, DeBenedetti MK, Marx SJ. (1997). Science, 276, 404 ± 407. Doherty GM, Wells Jr SA, Goodfellow PJ and Lairmore Fromant M, Blanquet S and Plateau P. (1995). Anal. TC. (1999). Hum. Mutat., 13, 175 ± 185. Biochem., 224, 347 ± 353. Pfarr CM, Mechta F, Spyrou G, Lallemand D, Carillo S and Gobl AE, Berg M, Lopez-Egido JR, Oberg K, Skogseid B Yaniv M. (1994). Cell, 76, 747 ± 760. and Westin G. (1999). Biochim. Biophys. Acta., 1447, 51 ± Poncin J, Abs R, Velkeniers B, Bonduelle M, Abramowicz 56. M, Legros JJ, Verloes A, Meurisse M, Van Gaal L, Giraud S, Zhang CX, Serova-Sinilnikova O, Wautot V, Verellen C, Koulischer L and Beckers A. (1999). Hum. Salandre J, Buisson N, Waterlot C, Bauters C, Porchet N, Muta., 13, 54 ± 60. Aubert JP, Emy P, Cadiot G, Delemer B, Chabre O, Schorpp-Kistner M, Wang ZQ, Angel P and Wagner EF. Niccoli P, Leprat F, Duron F, Emperauger B, Cougard P, (1999). EMBO J., 18, 934 ± 948. Goudet P, Sarfati E, Riou JP, Guichard S, Rodier M, Teh BT, Kytola S, Farnebo F, Bergman L, Wong FK, Weber Meyrier A, Caron P, Vantyghem M-C, Assayag M, Peix J- G, Hayward N, Larsson C, Skogseid B, Beckers A, Phelan L, Pugeat M, Rohmer V, Vallotton M, Lenoir G, Gaudray C, Edwards M, Epstein M, Alford F, Hurley D, P, Proye C, Conte-Devolx B, Chanson P, Shugart YY, Grimmond S, Silins G, Walters M, Stewart C, Cardinal Goldgar D, Murat A and Calender A. (1998). Am. J. Hum. J, Khodaei S, Parente F, Tranebjaerg L, Jorde R, Menon Genet., 63, 455 ± 467. J, Khir A, Tan TT, Chan SP, Zaini A, Khalid BAK, Guru SC, Goldsmith PK, Burns AL, Marx SJ, Spiegel AM, Sandelin K, Thompson N, Brandi M-L, Warth M, Stock J, Collins FS and Chandrasekharappa SC. (1998). Proc. Leisti J, Cameron D, Shepard JJ, Oberg K, Nordenskjold Natl. Acad. Sci. USA, 95, 1630 ± 1634. M and Salmela P. (1998). J. Clin. Endocrinol. Metab., 83, Hai N, Aoki N, Matsuda A, Mori T and Kosugi S. (1999). 2621 ± 2626. Eur. J. Endocrinol., 141, 475 ± 480. Thepot D, Weitzman JB, Barra J, Segretain D, Stinnakre Hickman AB, Klein DC and Dyda FD. (1999). Mol. Cell., 3, MG, Babinet C and Yaniv M. (2000). Development, 127, 23 ± 32. 143 ± 153. Hilberg F, Aguzzi A, Howells N and Wagner EF. (1993). Vidal M, Brachmann RK, Fattaey A, Harlow E and Boeke Nature, 365, 179 ± 181. JD. (1996). Proc. Natl. Acad. Sci. USA, 93, 10315 ± 10320. Johnson RS, van Lingen B, Papaioannou VE and Spiegel- Wisdom R. (1999). Exp. Cell. Res., 253, 180 ± 185. man BM. (1993). Genes. Dev., 7, 1309 ± 1317. Kerr LD, Miller DB and Matrisian LM. (1990). Cell, 61, 267 ± 278.

Oncogene