Viral mutations enhance the Max binding properties of the vMyc b-HLH-L
http://www.100md.com
《核酸研究医学期刊》
Biomedical Research Centre, University of Dundee, Ninewells Hospital and Medical School Dundee, DD1 9SY, UK 1Marie Curie Research Institute The Chart, Oxted, Surrey, RH8 OHL, UK 2VLA Laboratories New Haw, Addlestone, Surrey KT15 3NB, UK 3Beatson Institute for Cancer Research Garscube Estate, Switchback Road, Bearsden, Glasgow G61 1BD, UK
*To whom correspondence should be addressed. Tel: +44 1382 660111 ext. 33740; Fax: +44 1382 669993; Email: d.h.crouch@dundee.ac.uk
ABSTRACT
Interaction with Max via the helix–loop–helix/leucine zipper (HLH-LZ) domain is essential for Myc to function as a transcription factor. Myc is commonly upregulated in tumours, however, its activity can also be potentiated by virally derived mutations. vMyc, derived from the virus, MC29 gag-Myc, differs from its cellular counterpart by five amino acids. The N-terminal mutation stabilizes the protein, however, the significance of the other mutations is not known. We now show that vMyc can sustain longer deletions in the LZ domain than cMyc before complete loss in transforming activity, implicating the viral mutations in contributing to Myc:Max complex formation. We confirmed this both in vitro and in vivo, with loss of Max binding correlating with a loss in the biological activity of Myc. A specific viral mutation, isoleucine383>leucine (I383>L) in helix 2 of the HLH domain, extends the LZ domain from four to five heptad repeats. Significantly, introduction of I383>L into a Myc mutant that is defective for Max binding substantially restored its ability to complex with Max in vitro and in vivo. We therefore propose that this virally derived mutation is functional by significantly contributing to establishing a more hydrophobic interface between the LZs of Myc and Max.
INTRODUCTION
Retrovirally transduced oncogenes have acquired mutations that considerably potentiate their transforming activity by subverting their normal regulation in a cell. These mutations may result in altered regulation of the oncoprotein by key signalling pathways (e.g. phosphorylation), a reduction or interference in key protein:protein interactions, altered protein turnover or a combination of all these. Indeed, comparative studies between these virally derived oncogenes and their cellular counterparts have contributed greatly toward our current understanding of their molecular mechanism of action. vSrc, the transforming component of Rous sarcoma virus, and vErbB, isolated from the avian erythroblastosis virus, AEV, contain several that contribute to their constitutive tyrosine kinase activity (1,2).
Nuclear oncogenes have also been retrovirally transduced and sustained mutations that potentiate their transforming activity (3–5). Retrovirally transduced cJun and cFos, the two components of the AP-1 transcription factor complex, have sustained mutations which abrogate key phosphorylation events and contribute to an increased half life respectively (3,4).
cMyc, the transforming component of the avian leukosis virus, MC29, belongs to the basic/helix–loop–helix/leucine zipper (b-HLH-LZ) class of transcription factors (6). Somatic and virally derived point mutations in Myc potentiate its function as an oncoprotein, the majority of which are clustered within the transactivation domain at the N-terminus of the protein (7–10). Of the five mutations in MC29 vMyc (11), a functional consequence has only been attributed to threonine 61. This mutation, threonine61>methionine (T61>M) which is a known phosphorylation site (12), results in significant stabilization of the Myc protein (7,13,14).
Myc functions in association with a small unrelated protein, Max, dimerizing through the C-terminal HLH-LZ domain. Dimerization with Max is not, however, sufficient for Myc to function, since the complex must also be able to bind to a specific target DNA sequence and activate transcription through its N-terminus (6). The LZ domains of Myc and Max form a parallel two-stranded -helical coil and dictate the specificity of heterodimerization (15). A detailed sequence comparison between the LZs of different transcription factors shows that in contrast to Fos and Jun which each contain five leucine residues in a heptad repeat (denoted L1–L5 in Figure 1), cMyc contains only four (denoted L2–L5 in Figure 1). Strikingly, in MC29 vMyc, mutation of isoleucine to a leucine at position 383 (Figure 1) extends the cMyc LZ to five leucine residues in a heptad repeat. This led us to speculate that by increasing the length of the LZ, this virally derived mutation may contribute positively to the interaction with Max. The data we present in this manuscript is consistent with this and for the first time, we propose a functional consequence of this virally derived mutation in the C-terminus of vMyc.
Figure 1 Sequence alignment of the LZ domains of different transcription factors. The amino acids that correspond to position 7 in a heptad repeat are boxed. The leucine repeats are indicated below, with leucine 1 (L1) and leucine 5 (L5) being the most N-terminal and C-terminal of the heptad repeats, respectively. Sequences were taken from the following accession numbers: human cMyc (NM_002467 ), murine cMyc (NM_010849 ), feline cMyc (M22727 ), avian cMyc (J00889 ), avian MC29 vMyc (VO1173), avian MH2 vMyc (K02082 ), human Nmyc (NM_005378 ), murine Nmyc (NM_008709 ), human Max (NM_002382 ), human cJun (NM_002228 ), human cFos (BC004490 ) and human C/EBP (NM_005194 ).
MATERIALS AND METHODS
Cell culture and transfections
Cell culture and transfection of appropriate SFCV-Myc constructs (10 μg) together with RCAN (A) helper (4 μg) into secondary chick embryo fibroblasts (CEFs) was performed essentially as described (16). Following G418 selection, cultures were expanded and used to analyse alterations in cell behaviour which was mediated by overexpression of the Myc oncoprotein (16). Briefly, anchorage independent growth was determined by plating 2 x 105 cells into 0.35% agar and incubated at 41°C for 2 weeks prior to photography. Growth rate was measured by plating 2 x 105 cells in a 35 mm diameter dish, and cumulative cell counts performed each day.
Construction of retroviral vectors expressing mutant Myc alleles
The construction of SFCV-cMyc and SFCV-vMyc has been described (17). All the LZ mutants of vMyc were generated by site directed mutagenesis as described (18) using mp8-vMyc as the template. The mutagenic oligonucleotides were as follows: vMyc7, 5'-aaccttgagtagctaaggaag-3'; vMyc10, 5'-agttgaaacactaacttgagc-3' and vMyc14, 5'-gtgtttcaactattctctcctccgcctcaa-3'. To generate the isoleucine383>leucine (I383>L) mutant, mutagenesis was performed using the primer 5' gttctgtctccaatcggacgag 3'. To generate the I383>L10 mutant, mutagenesis was performed on the I383>L template using the primer 5'-gttgaaacactaacttgagc-3'. The underscored nucleotides encode the mutant amino acid.
The resulting mutants were retrieved from mp8 and cloned as HindIII fragments into SFCV-sa– (17) and pSPT19 (19). All mutant sequences were confirmed by double stranded sequencing.
Western blot analysis
Cell lysates were prepared by lysing cultures in SDS-sample buffer (20). Following sonication and protein estimation, 50 μg protein was loaded onto 7.5% SDS–PAGE gels. Transfer to nitrocellulose and western blotting was performed as described (18). Proteins were detected using specific rabbit antibodies and visualized using either NBT/BCIP (anti-cMyc 237) or enhanced chemiluminescence (anti-Max).
In vitro translation of Myc and Max proteins and immunoprecipitation
Dimerization between Myc and Max was determined using methionine-labelled in vitro translated proteins essentially as described (19). Briefly, following incubation with Myc proteins, Max or Max9 were specifically immunoprecipitated using an anti-cMyc antibody, 237. The immune complexes were recovered on protein A–Sepharose beads, washed thoroughly and resolved on 10% SDS–PAGE gels. The methionine labelled proteins were detected by fluorography using Amplify (Amersham).
Reporter and activator plasmids
The PHO5 UAS-CYC-LacZ reporter plasmid, pRS314-Max/Max9, PHO4-cMyc (Pho4-cMyc) and Pho4 cMyc LZ mutant hybrids have been described previously (19). The PHO4-vMyc (Pho4-vMyc), Pho4 vMyc LZ mutants and Pho4-I>L10 hybrids were made by PCR of the appropriate templates and were cloned into the BglII site of pMA132 (19). The integrity of all expression constructs was verified by sequencing.
The Pho5 UAS-CYC-Lac Z reporter plasmid encodes the Pho4 DNA-binding sequence upstream of ?-galactosidase. This sequence, CACGTG, has also been shown to be a consensus sequence for the Myc/Max complex (22). The Pho4-Myc plasmids express the Pho4 transactivation domain fused to the b-HLH-LZ domain (amino acids 327–415) of the different Myc isoforms. pRS315-Max/Max9 expresses the two different Max isoforms.
Dimerization between the b-HLH-LZ domains of Myc and Max was detected by transforming the above plasmids into yeast. The Pho4–Myc/Max complexes formed in vivo will bind to the Pho5 UAS-CYC consensus sequence upstream of the LacZ gene. Transcription of LacZ will then be initiated by the Pho4 transactivation domain. The level of ?-galactosidase activity in cell lysates of the transformants will therefore be a direct measure of the extent of dimerization between the b-HLH-LZ domains of Myc and Max (19).
The construction of pGV256-lex-OP was as described (23). To generate pRS315-lex-Max9, the Max 9 coding sequence was first inserted as a BglII fragment into pV44ER-lex (23). The entire cassette comprising the GAL UAS, CYC promoter, Lex-Max9 and the CYC terminator was then cloned as an Sst1–Kpn1 fragment into pRS315.
These plasmids were used in an alternate dimerization assay that independently measures Myc/Max complex formation. pGV256-lex-OP contains the Lex operator upstream of CYC-LacZ (23). Lex A, the bacterial repressor, binds to the Lex operator. pRS315-lexA-Max 9 encodes a fusion protein between the bacterial Lex A repressor and Max 9, which will bind to the Lex operator. These plasmids were then transformed into yeast along with Pho4-Myc. Any Lex A-Max9:Pho4-Myc complexes that are formed will bind to the Lex operator sequence via LexA-Max9. The ?-galactosidase activity that is detected in the transformants will be the direct result of the complex formation between LexA-Max 9 and Pho4-Myc, since the Pho4 transactivation domain will mediate transcriptional activation. This assay therefore provides a dimerization assay which is independent of DNA binding mediated by the Myc:Max complex.
Yeast culture and ?-galactosidase assays
Reporter assays were performed in the yeast strain Y700 (, his3–11, ade2–1, leu2–3, –112, ura3, trp1–1 and can1–100) essentially as described (19). Following selection on yeast glucose minimal agar plates supplemented with the appropriate amino acids, the transformants were grown in liquid culture and ?-galactosidase activity measured (22). Activities represent the average of at least two independent duplicate cultures within the same experiment and units of ?-galactosidase activity were calculated as described (19).
RESULTS
vMyc can sustain longer truncations in its LZ domain than cMyc before loss of biological activity
Previous work from our laboratory showed that the integrity of the LZ domain was essential for Myc to function (18). Deletion of seven amino acids (removal of L5) in cMyc resulted in a mutant, cMyc7, which retained partial transforming activity (18). Deletion of a further three amino acids (removal of L4a and L5), however, resulted in a non-transforming mutant of Myc, cMyc10. We reasoned that vMyc might be able to sustain longer truncations in the LZ domain than cMyc before complete loss of biological activity. To address this, we generated a series of point mutants of vMyc which successively truncated leucine residues from the C-terminus. These mutants, designated vMyc7, vMyc10 and vMyc14, were cloned into avian retroviral vector, SFCV-sa– (24) and transfected into CEF. It should be noted that these constructs only encode the vMyc portion of MC29, and not the p110 gag-Myc encoded by the original MC29 virus (25). Following G418 selection, expression of all Myc-containing constructs was confirmed by western blot analysis (Figure 2). All were expressed in CEF, although the vMyc14 and cMyc14 mutants did appear to be expressed at lower levels than the wild-type. These low levels were unexpected. These mutants should bind Miz-1, but not Max (Figures 4 and 5). Since the former stabilizes Myc, and the latter has no affect on its half life (13), we would expect at least equal levels of the vMyc14 and cMyc14 mutants. Chicken Miz-1 has not, however, been characterized, and its levels in CEF are not known. The significance of the reduced expression of these two mutants is therefore not known. The level of Max9, the major Max isoform in CEF (17,26), does not change in the Myc-infected cells (Figure 2B).
Figure 2 LZ mutants of MC29 vMyc and avian cMyc. (A) Schematic representation showing location of MC29 vMyc mutations. Five mutations are contained within MC29 vMyc (11). These are found in the transactivation domain (T61>M), adjacent to the basic region (B) (serine325>leucine—S325>L), within helix 1 (serine350>arginine—S350>R) and helix 2 (I383>L) of the HLH domain, and within the LZ domain (lysine407>arginine—K407>R). LZ deletion mutants of MC29 vMyc and avian cMyc are shown. The leucine repeats (L1–L5) within the LZs of MC29 vMyc and avian cMyc are also indicated. Leucine L4a is shown offset to indicate its internal location (position 3) within the most C-terminal heptad repeat. Premature translation termination codons (indicated by an asterisk) were introduced into MC29 v-myc and avian c-myc (18) by site-directed mutagenesis. The 7 and 14 mutants truncate specifically at L5 and L4 within the heptad, respectively, whilst the 10 truncation removes L5 and L4a. (B) Retrovirally-expressed cMyc, vMyc and their respective LZ mutants in CEFs were detected by western blot of total cell lysates using an anti-Myc antibody (upper panel). Equivalence of loading was shown by Max9, the main isoform of Max in CEF (17,26), which was detected using an anti-Max antibody (21) (lower panel). Vector control is shown by a dash.
Figure 4 Complex formation between Max/Max9 and the LZ mutants of v- and c-Myc in vitro. cMyc, vMyc, LZ mutants and Max/Max9 were produced in rabbit reticulocyte lysate and Myc/Max complex formation determined by immunoprecipitation of Max using an anti-myc specific antibody. -labelled proteins were resolved by SDS–PAGE and visualized by fluorography. Central lane was Max/Max9 alone.
Figure 5 Complex formation between Max/Max9 and the LZ mutants of v- and c-Myc in yeast. Two different experimental systems based on sensitive yeast two hybrid assays were used to measure complex formation in vivo. (A) Max/Max9-dependent dimerization and DNA binding were determined in vivo. ?-Galactosidase activity was quantitatively measured as a result of transcriptional activation resulting from the co-transfection of Pho4-cMyc, Pho4-vMyc or their respective LZ mutants into yeast, together with the PHO5 UAS-CYC-LacZ reporter. This assay has been shown previously to a reliable assay for Myc/Max complex formation (28). (B) Dimerization which is independent of Myc/Max DNA binding was measured by co-transfecting Pho4-cMyc, Pho4-vMyc or their respective LZ mutants into yeast, along with the reporter, pGV256-lex-OP (23) and pRS315-lexA-Max9. ?-Galactosidase activity was then determined.
CEF which overexpress Myc undergo extensive changes in cell morphology, grow more rapidly and acquire the ability to grow in an anchorage independent manner (16,18). Therefore, having confirmed the appropriate retroviral expression, we determined changes in cell morphology (Figure 3A), calculated the growth rate (Figure 3B) and determined the ability to grow in agar of CEF overexpressing each Myc mutant (16,18).
Figure 3 vMyc can sustain longer deletions in the LZ than cMyc before loss of biological activity (A) Cell morphology of CEF expressing cMyc, vMyc and their respective LZ mutants are shown. (a) cMyc, (b) cMyc7, (c) cMyc10, (d) cMyc14, (e) vMyc, (f) vMyc7, (g) vMyc10, (h) vMyc14 and (i) vector control. (B) Growth rate of CEF expressing cMyc, vMyc and their respective LZ mutants. Growth rate was measured by cumulative cell counts over 4 days. These data are representative of at least two different experiments. (C) The ability of retrovirally-expressed cMyc, vMyc and their respective LZ mutants to induce anchorage-independent growth was determined by plating infected CEF into soft agar. Colony counts were taken after 14 days. Vector control is shown by a dash. These data are representative of at least two different experiments.
Control chick cells are spindle shaped, non-refractile and contact inhibited, lining up in a parallel manner on the dish (Figure 3A, i). In contrast, cMyc-infected CEF are rounder, more refractile and do not undergo contact inhibition (Figure 3A, a). cMyc7 (Figure 3A, b) had an intermediate phenotype, being partially morphologically-transformed. Deletion of 10 amino acids or more from the cMyc LZ, however, resulted in mutants that were non-transforming (Figure 3A, c and d). vMyc (Figure 3A, e) was more highly transformed than cMyc (Figure 3A, a), highlighting a major role for the viral mutations in potentiating the biological activity of Myc (9). vMyc (Figure 3A, e) and vMyc7 (Figure 3A, f) were morphologically indistinguishable from each other and were highly transformed. vMyc10 (Figure 3A, g) was partially transformed, however, vMyc14 (Figure 3A, h) was non-transformed. Thus, deletion of 14 amino acids from the vMyc LZ was required before loss in biological activity, whilst deletion of only 10 amino acids was required before cMyc became functionally inert. These data clearly show that vMyc can sustain longer truncations in its LZ than cMyc before loss of morphological transformation.
Another feature of Myc-transformed cells is an accelerated cell growth (18). The growth rate of Myc-infected CEF, determined by cumulative cell counts >4 days, is shown graphically (Figure 3B). As previously reported, cMyc and cMyc7 have accelerated growth rates, but further truncation of the LZ domain resulted in growth rates comparable with uninfected CEF (18). In contrast, vMyc, vMyc7 and vMyc10 grew more rapidly than control cells, whilst vMyc14 was indistinguishable from the control.
Loss of anchorage dependence by transformed cells was measured by the ability to grow in soft agar. Myc-infected CEF were plated in agar and after 2 weeks, plates were photographed. A minimum of eight frames were taken for each and the average number of colonies calculated (Figure 3C). It can clearly be seen that in contrast to vMyc7, cMyc7 only partially retains the ability to grow in agar. In contrast to cMyc10, vMyc10 was however still partially able to grow in agar.
Collectively, these data show that the integrity of the LZ domains of vMyc and cMyc are required for biological activity. More importantly, these results highlight a major functional difference in the behaviour of the vMyc and cMyc LZ domains, since vMyc can sustain longer truncations than cMyc before complete loss in its biological activity (Table 1). Indeed, deletion of 10 amino acids from the C-terminus of vMyc results in a mutant that still retains biological activity, whilst deletion of 10 amino acids from cMyc results in a mutant that is functionally inactive. This difference in the behaviour of the LZ domains must be a direct consequence of the virally derived mutations.
Table 1 LZ mutants of avian cMyc and vMyc are functionally different
Biologically active LZ mutants of vMyc retain the ability to dimerize with Max in vitro and in vivo
Since Myc requires Max to function (27), we set out to establish whether the vMyc LZ mutants retained the ability to dimerize with Max. -labelled Myc and Max were produced in an in vitro translation system, and Myc–Max complexes allowed to form at 37°C. Complexes were recovered by immunoprecipitation using an anti-Myc antibody that does not recognize Max (Figure 4). The central lane in both panels, indicated by a dash, contained Max or Max9 alone demonstrating that the Myc antibody does not immunoprecipitate Max. From these data, it can be seen that in contrast to vMyc, vMyc7 and vMyc10 which can complex with Max (upper panel) and Max9 (lower panel), only cMyc and cMyc7 complex with Max. Therefore, these data confirm that vMyc can sustain greater truncations to the LZ domain than cMyc before losing the ability to complex with Max. Importantly, all the transforming mutants of both vMyc and cMyc (Figure 3) bound to Max and Max9.
To confirm these data, we tested the ability of the LZ mutants of Myc to complex with Max in vivo using a yeast assay designed to measure dimerization and DNA binding of Myc:Max complexes. This assay has been used previously as a reliable and sensitive assay for Myc:Max function in vivo (19,28). In this assay, the bHLH-LZ domain of avian Myc or vMyc (amino acids 330–417) or their respective LZ mutants, were fused to the transactivation domain of the yeast transcription factor Pho4. When Pho4-Myc was transfected into yeast together with a PHO5 UAS-CYC-lacZ reporter plasmid, transcription through the PHO5-UAS occurred only in the presence of Max or Max9 (19). As expected, only Pho4-cMyc and Pho4-cMyc7 were able to activate transcription, whilst Pho4-cMyc10 and Pho4-cMyc14 did not. In contrast to cMyc, co-expression of Max or Max9 with Pho4-vMyc, Pho4-vMyc7 and Pho4-vMyc10 activated transcription through the PHO5-UAS, albeit at a reduced level for Pho4-vMyc10 (Figure 5A). These in vivo data confirm the in vitro results (Figure 4) clearly showing that vMyc can sustain a larger truncation to the LZ domain than cMyc before losing the ability to bind to Max.
As an independent measure of dimerization, we used an assay which asks whether Myc function was mediated by direct complex formation with Max, independently of its DNA binding activity. To achieve this, we fused the bacterial protein, LexA, to Max9, generating pRS315-lexA-Max9. When transfected into yeast, LexA binds to a specific DNA sequence, LexOP, which lies upstream of ?-galactosidase, however, no ?-galactosidase activity will be detected in the absence of Pho4-Myc. As expected, no activity was detected when pRS315-lexA-Max9 was transfected into yeast along with pGV256-lex-OP (23). When they were co-transfected with Pho4-Myc plasmids, however, ?-galactosidase activity was recorded. As can be seen from Figure 5B, significant ?-galactosidase activity was recorded with all the transforming mutants of cMyc and vMyc.
Collectively, these data (Figures 4 and 5) clearly show that the LZ domains of vMyc and cMyc differ significantly in their ability to bind to Max both in vitro and in vivo.
Mutation of I383>L in cMyc10 background partially restores binding to Max in vitro and in vivo
cMyc10 does not bind Max/Max9 (19) (Figures 4 and 5). Theoretically, the I383>L mutation of vMyc could extend the LZ domain from four to five heptad repeats and stabilize the Myc/Max interaction (Figure 1). To directly test this, we introduced this mutation into a cMyc10 background. This mutant, cMycI>L10 (Figure 6A) was tested for the ability to complex with Max both in vitro (Figure 6B) and in vivo (Figure 6C). In both these assays, replacement of isoleucine 383 with a leucine significantly restored binding to both Max and Max9. This mutant did not, however, significantly induce a biological phenotype when overexpressed in CEF, suggesting that the levels of cMycI>L10:Max complex may be below the threshold required to mediate a biological response. These data do, however, show that this single mutation is sufficient to contribute positively to the interaction of Myc and Max.
Figure 6 I383>L mutation in cMyc10 partially restores interaction with Max/Max9 in vitro and in vivo. (A) Schematic representation of cMyc, cMyc10 and cMycI>L10 mutants. I383>L mutation was introduced into cMyc10, which does not bind to Max (18). (B) Complex formation between Max/Max9 and cMyc, cMyc10, and cMyc I>L10 co-translated in vitro was determined by immunoprecipitation of Max/Max9 using an anti-Myc antibody, and -labelled proteins resolved by SDS–PAGE and visualized by fluorography. (C) Complex formation between Max/Max9 and Pho4-cMyc, Pho4cMyc10 and Pho4-cMycI>L10 was measured in vivo by co-transfecting the appropriate plasmids into yeast. ?-Galactosidase activity was then determined.
DISCUSSION
Somatic or virally derived mutations that potentiate the function of a protein can be viewed as naturally occurring protein engineering. Currently, many examples of these exist, which can result in proteins with altered properties, such as different half lives, differential responses to signalling pathways through loss in phosphorylation sites or altered protein:protein interactions. In the data presented here, viral mutations clearly enhance the Max binding properties of the b-HLH-LZ domain, since the vMyc LZ can tolerate greater loss of its dimerization interface than cMyc before loss of Max binding.
We propose that the I383>L mutation in MC29 vMyc represents an example of naturally occurring protein engineering. Given that HLH-LZ heterodimers are stabilized by hydrophobic and polar interactions involving the -helices, H1, H2 and LZ (29), the function of the I383>L mutation in the C-terminus of vMyc would be to contribute to establishing a more hydrophobic interface which stabilizes the Myc–Max complex.
Our data are wholly consistent with this proposal, since the introduction of a single amino acid change, I383>L, into cMyc10, a LZ mutant which is defective for Max binding, could significantly restore its ability to complex with Max both in vitro and in vivo. This amino acid substitution did not, however, restore biological activity, although small colonies were detected in agar (data not shown). Therefore, although this mutation contributed positively to the Myc:Max interaction, it is most likely that the levels of cMyc I>L10:Max complexes formed in CEF were below the threshold required to elicit any biological response (17).
The I383>L mutation is located in the C-terminal HLH domain, a region which is known to bind other proteins (30). This mutation may therefore regulate the interaction of Myc with these factors. This is thought highly unlikely given the extremely tight correlation between the ability of the different LZ mutants to bind to Max and transform cells, suggesting that it is the Myc:Max complex which is regulated by these mutations (19,27).
The behavioural differences between the respective LZ mutants of vMyc and cMyc clearly show that the virally derived mutations in vMyc impinge on the structure and function of its b-HLH-LZ domain. Whilst the integrity of the LZ domain of vMyc was still required for its biological activity, the v- and cMyc LZ mutants differ significantly with respect to both their biological activity and their ability to bind to Max (Figures 3–5). Given that the vMyc LZ can tolerate larger truncations from its LZ zipper domain than cMyc before loss in its biological activity and binding to Max, this must reflect a stronger interaction between the LZ domains of vMyc and Max than the LZ domains of cMyc and Max.
These findings could have major implications in the rational design of peptide inhibitors for cancer treatment, since a knowledge of the avidity of protein:protein interactions mediated through HLH-LZ dimerization could dictate the design of peptide inhibitors for use in a therapeutic context. Indeed, one such strategy was recently described which was based on the use of helix 1 (H1) peptides to inhibit the protein:protein interactions between Myc and Max resulting in a block to Myc-mediated cell proliferation (31). Although this study focussed on H1 rather than LZ peptides to block Myc function, it highlights the general applicability of targeting any essential dimerization interface with a view to inhibiting function. Our data could therefore have important implications if the target is a cancer cell which contains a mutated Myc that has a stronger dimerization interface than the wild-type Myc, since therapeutic targeting of this interface with a relatively weak peptide inhibitor would obviously be counterproductive.
Conversely, the stability of LZ interactions could be exploited to therapeutically intervene with Myc:Max complexes in cancer cells. An example of this was described by Jean-Francois et al. (29), who introduced two point mutations in Max, His82>Leu and Asn78>Val, with a view to increasing the hydrophobic interface between the two contributing Max monomers. This situation is similar to that described here, since the His82>Leu mutation increased the number of heptad repeats in the LZ from three to four. Their rationale was that the stabilized mutant Max homodimers would compete with Myc:Max heterodimers, block their binding to target E-boxes and as such, would act effectively as an anti-Myc drug (29). Indeed, this mutant dimer was subsequently shown to have improved thermodynamic stability and form more stable E-box complexes. Therefore, rather than titrating out Max to abolish Myc function (29), Myc peptides that provide a stronger dimerization interface could also theoretically be used to dominantly interfere with the formation of Myc:Max complexes. Obviously, this would only be applicable in a cancer cell that contained wild-type Myc.
In summary, we have shown for the first time a major difference between the C-termini of vMyc and cMyc, since vMyc can sustain greater truncations of the C-terminal LZ before loss in biological activity. Furthermore, we show that a single point mutation in helix 2 from vMyc can in isolation positively contribute to the interaction with Max. We therefore propose that this mutation provides a more hydrophobic surface between the dimerization interfaces of Myc and Max. Together with data showing the functional significance of a mutation in the transactivation domain (7), these data highlight the importance of virally derived mutations as examples of naturally occurring protein engineering.
ACKNOWLEDGEMENTS
We wish to thank Trevor Littlewood for help and reagents. This work was funded by Cancer Research UK and Association for International Cancer Research. Funding to pay the Open Access publication charges for this article was provided by JISC.
REFERENCES
Frame, M.C. (2004) Newest findings on the oldest oncogene; how activated Src does it J. Cell Sci., 117, 989–998 .
McCubrey, J.A., Shelton, J.G., Steelman, L.S., Franklin, R.A., Sreevalsan, T., McMahon, M. (2004) Effects of a conditionally active v-ErbB and an EGF-R inhibitor on transformation of NIH-3T3 cells and abrogation of cytokine dependency of hematopoietic cells Oncogene, 23, 7810–7820 .
Acquaviva, C., Salvat, C., Brockly, F., Bossis, G., Ferrara, P., Piechaczyk, M.P., Jariel-Encontre, I. (2001) Cellular and vial Fos proteins are degraded by different proteolyic systems Oncogene, 20, 942–950 .
Dai, T., Rubie, E., Franklin, C.C., Kraft, A., Gillespie, D.A., Avruch, J., Kyriakis, J.M., Woodgett, J.R. (1995) Stress-activated protein kinases bind directly to the delta domain of c-Jun in resting cells: implications for repression of c-Jun function Oncogene, 10, 849–855 .
Kanei-Ishii, C., Nomura, T., Tanikawa, J., Ichikawa-Iwata, E., Ishii, S. (2001) Differential sensitivity of v-Myb and c-Myc to Wnt-1-induced protein degradation J. Biol. Chem., 279, 44582–44589 .
Cole, M.D. and McMahon, S.B. (1999) The Myc oncoprotein: a critical evaluation of transactivation and target gene regulation Oncogene, 18, 2916–2924 .
Gavine, P.R., Neil, J.C., Crouch, D.H. (1999) Protein stabilization: a common consequence of mutations in independently derived v-Myc alleles Oncogene, 18, 7552–7558 .
Hoang, A.T., Lutterbach, B., Lewis, B.C., Yano, T., Chou, T.Y., Barrett, J.F., Raffeld, M., Hann, S.R., Dang, C.V. (1995) A link between increased transforming activity of lymphoma-derived Myc mutant alleles, their defective regulation by p107, and altered phosphorylation of the c-Myc transactivation domain Mol. Cell. Biol., 15, 4031–4042 .
Petropoulos, C.J., Givol, I., Hughes, S.H. (1996) Comparative analysis of the structure and function of the chicken c-myc and v-myc genes: v-myc is a more potent inducer of cell proliferation and apoptosis than c-myc Oncogene, 12, 2611–2623 .
Henriksson, M., Bakardjiev, A., Klein, G., Luscher, B. (1993) Phosphorylation sites mapping in the N-terminal domain of c-Myc modulate its transforming potential Oncogene, 8, 3199–3209 .
Walther, N., Jansen, H., Trachmann, C., Bister, K. (1986) Nucleotide sequence of the CMII v-myc allele Virology, 154, 219–223 .
Lutterbach, B. and Hann, S.R. (1994) Hierarchical phosphorylation at N-terminal transformation-sensitive sites in c-Myc protein is regulated by mitogens and in mitosis Mol. Cell. Biol., 14, 5510–5522 .
Salghetti, S., Kim, S.Y., Tansey, W.P. (1999) Destruction of Myc by ubiquitin-mediated proteolysis: cancer-associated and transforming mutations stabilize Myc EMBO J., 18, 717–726 .
Flinn, E.M., Busch, C.M., Wright, A.P. (1998) Myc boxes, which are conserved in myc family of proteins, are signals for protein degradation via the proteasome Mol. Cell. Biol., 18, 5961–5969 .
Muhle-Goll, C., Nilges, M., Pastore, A. (1995) The leucine zippers of the HLH-LZ proteins Max and c-Myc preferentially form heterodimers Biochemistry, 34, 13554–13564 .
Crouch, D.H., Gallagher, R., Goding, C.R., Neil, J.C., Fulton, R. (1996) Multiple phenotypes associated with Myc-induced transformation of chick embryo fibroblasts can be dissociated by a basic region mutation Nucleic Acids Res., 24, 3216–3221 .
La Rocca, S.A., Crouch, D.H., Gillespie, D.A.F. (1994) c-Myc inhibits myogenic differentiation and MyoD expression by a mechanism which can be dissociated from cell-transformation Oncogene, 9, 3499–3508 .
Crouch, D.H., Lang, C., Gillespie, D.A.F. (1990) The leucine zipper domain of avian cMyc is required for transformation and autoregulation Oncogene, 5, 683–689 .
Crouch, D.H., Fisher, F., Clark, W., Jayaraman, P.S., Goding, C.R., Gillespie, D.A.F. (1993) Gene-regulatory properties of Myc helix–loop–helix/leucine zipper mutants: max-dependent DNA-binding and transcriptional activation in yeast correlates with transforming capacity Oncogene, 8, 1849–1855 .
Batchelor, C.L., Woodward, A.M., Crouch, D.H. (2004) Nuclear ERM (ezrin, radixin, moesin) proteins: regulation by cell density and nuclear import Exp. Cell Res., 296, 208–222 .
Littlewood, T.D., Amati, B., Land, H., Evan, G.I. (1992) Max and c-Myc/Max DNA-binding activities in cell extracts Oncogene, 7, 1783–1792 .
Fisher, F., Jayaraman, P.-S., Goding, C.R. (1991) c-Myc and the yeast transcription factor PHO4 share a common CACGTG-binding motif Oncogene, 6, 1099–1104 .
Jayaraman, P.-S., Hirst, K., Goding, C.R. (1994) The activation domain of a basic helix–loop–helix protein is masked by repressor interaction with domains distinct from that required for transcription regulation EMBO J., 13, 2192–2199 .
Fuerstenberg, S., Beug, H., Introna, M., Khazaie, K., Munoz, A., Ness, S., Nordstrom, K., Sap, J., Stanley, I., Zenke, M., et al. (1990) Ectopic expression of the erythrocyte band 3 anion exchange protein, using a new avian retrovirus vector J. Virol., 64, 5891–5902 .
Lee, C.M. and Reddy, E.K. (1999) The v-myc oncogene Oncogene, 18, 2997–3003 .
Sollenberger, K.G., Kao, T.-L., Taparowsky, E.J. (1994) Structural analysis of the chicken max gene Oncogene, 9, 661–664 .
Luscher, B. and Larsson, L.-G. (1999) The basic region/helix–loop–helix zipper domain of Myc proto-oncoproteins: function and regulation Oncogene, 18, 2955–2966 .
Fisher, F., Crouch, D.H., Jayaraman, P.S., Clark, W., Gillespie, D.A.F., Goding, C.R. (1993) Transcription activation By Myc and Max: flanking sequences target activation to a subset of CACGTG motifs in vivo EMBO J., 12, 5075–5082 .
Jean-Francois, N., Frederic, G., Raymund, W., Benoit, C., Lavigne, P. (2003) Improving the thermodynamic stability of the leucine zipper of Max increases the stability of its b-HLH-LZ:E:box complex J. Mol. Biol., 326, 1577–1595 .
Sakamuro, D. and Prendergast, G.C. (1999) New Myc-interacting proteins: a second Myc network emerges Oncogene, 18, 2942–2954 .
Nieddu, E., Melchiori, A., Pescarolo, M.P., Bagnasco, L., Biasotti, B., Licheri, B., Malacarne, D., Tortolina, L., Castagnino, N., Pasa, S., et al. (2005) Sequence specific peptidomimetic molecules inhibitors of a protein–protein interaction at the helix 1 level of c-Myc FASEB J., 19, 632–634 .(D. H. Crouch*, F. Fisher1, S. A. La Rocc)
*To whom correspondence should be addressed. Tel: +44 1382 660111 ext. 33740; Fax: +44 1382 669993; Email: d.h.crouch@dundee.ac.uk
ABSTRACT
Interaction with Max via the helix–loop–helix/leucine zipper (HLH-LZ) domain is essential for Myc to function as a transcription factor. Myc is commonly upregulated in tumours, however, its activity can also be potentiated by virally derived mutations. vMyc, derived from the virus, MC29 gag-Myc, differs from its cellular counterpart by five amino acids. The N-terminal mutation stabilizes the protein, however, the significance of the other mutations is not known. We now show that vMyc can sustain longer deletions in the LZ domain than cMyc before complete loss in transforming activity, implicating the viral mutations in contributing to Myc:Max complex formation. We confirmed this both in vitro and in vivo, with loss of Max binding correlating with a loss in the biological activity of Myc. A specific viral mutation, isoleucine383>leucine (I383>L) in helix 2 of the HLH domain, extends the LZ domain from four to five heptad repeats. Significantly, introduction of I383>L into a Myc mutant that is defective for Max binding substantially restored its ability to complex with Max in vitro and in vivo. We therefore propose that this virally derived mutation is functional by significantly contributing to establishing a more hydrophobic interface between the LZs of Myc and Max.
INTRODUCTION
Retrovirally transduced oncogenes have acquired mutations that considerably potentiate their transforming activity by subverting their normal regulation in a cell. These mutations may result in altered regulation of the oncoprotein by key signalling pathways (e.g. phosphorylation), a reduction or interference in key protein:protein interactions, altered protein turnover or a combination of all these. Indeed, comparative studies between these virally derived oncogenes and their cellular counterparts have contributed greatly toward our current understanding of their molecular mechanism of action. vSrc, the transforming component of Rous sarcoma virus, and vErbB, isolated from the avian erythroblastosis virus, AEV, contain several that contribute to their constitutive tyrosine kinase activity (1,2).
Nuclear oncogenes have also been retrovirally transduced and sustained mutations that potentiate their transforming activity (3–5). Retrovirally transduced cJun and cFos, the two components of the AP-1 transcription factor complex, have sustained mutations which abrogate key phosphorylation events and contribute to an increased half life respectively (3,4).
cMyc, the transforming component of the avian leukosis virus, MC29, belongs to the basic/helix–loop–helix/leucine zipper (b-HLH-LZ) class of transcription factors (6). Somatic and virally derived point mutations in Myc potentiate its function as an oncoprotein, the majority of which are clustered within the transactivation domain at the N-terminus of the protein (7–10). Of the five mutations in MC29 vMyc (11), a functional consequence has only been attributed to threonine 61. This mutation, threonine61>methionine (T61>M) which is a known phosphorylation site (12), results in significant stabilization of the Myc protein (7,13,14).
Myc functions in association with a small unrelated protein, Max, dimerizing through the C-terminal HLH-LZ domain. Dimerization with Max is not, however, sufficient for Myc to function, since the complex must also be able to bind to a specific target DNA sequence and activate transcription through its N-terminus (6). The LZ domains of Myc and Max form a parallel two-stranded -helical coil and dictate the specificity of heterodimerization (15). A detailed sequence comparison between the LZs of different transcription factors shows that in contrast to Fos and Jun which each contain five leucine residues in a heptad repeat (denoted L1–L5 in Figure 1), cMyc contains only four (denoted L2–L5 in Figure 1). Strikingly, in MC29 vMyc, mutation of isoleucine to a leucine at position 383 (Figure 1) extends the cMyc LZ to five leucine residues in a heptad repeat. This led us to speculate that by increasing the length of the LZ, this virally derived mutation may contribute positively to the interaction with Max. The data we present in this manuscript is consistent with this and for the first time, we propose a functional consequence of this virally derived mutation in the C-terminus of vMyc.
Figure 1 Sequence alignment of the LZ domains of different transcription factors. The amino acids that correspond to position 7 in a heptad repeat are boxed. The leucine repeats are indicated below, with leucine 1 (L1) and leucine 5 (L5) being the most N-terminal and C-terminal of the heptad repeats, respectively. Sequences were taken from the following accession numbers: human cMyc (NM_002467 ), murine cMyc (NM_010849 ), feline cMyc (M22727 ), avian cMyc (J00889 ), avian MC29 vMyc (VO1173), avian MH2 vMyc (K02082 ), human Nmyc (NM_005378 ), murine Nmyc (NM_008709 ), human Max (NM_002382 ), human cJun (NM_002228 ), human cFos (BC004490 ) and human C/EBP (NM_005194 ).
MATERIALS AND METHODS
Cell culture and transfections
Cell culture and transfection of appropriate SFCV-Myc constructs (10 μg) together with RCAN (A) helper (4 μg) into secondary chick embryo fibroblasts (CEFs) was performed essentially as described (16). Following G418 selection, cultures were expanded and used to analyse alterations in cell behaviour which was mediated by overexpression of the Myc oncoprotein (16). Briefly, anchorage independent growth was determined by plating 2 x 105 cells into 0.35% agar and incubated at 41°C for 2 weeks prior to photography. Growth rate was measured by plating 2 x 105 cells in a 35 mm diameter dish, and cumulative cell counts performed each day.
Construction of retroviral vectors expressing mutant Myc alleles
The construction of SFCV-cMyc and SFCV-vMyc has been described (17). All the LZ mutants of vMyc were generated by site directed mutagenesis as described (18) using mp8-vMyc as the template. The mutagenic oligonucleotides were as follows: vMyc7, 5'-aaccttgagtagctaaggaag-3'; vMyc10, 5'-agttgaaacactaacttgagc-3' and vMyc14, 5'-gtgtttcaactattctctcctccgcctcaa-3'. To generate the isoleucine383>leucine (I383>L) mutant, mutagenesis was performed using the primer 5' gttctgtctccaatcggacgag 3'. To generate the I383>L10 mutant, mutagenesis was performed on the I383>L template using the primer 5'-gttgaaacactaacttgagc-3'. The underscored nucleotides encode the mutant amino acid.
The resulting mutants were retrieved from mp8 and cloned as HindIII fragments into SFCV-sa– (17) and pSPT19 (19). All mutant sequences were confirmed by double stranded sequencing.
Western blot analysis
Cell lysates were prepared by lysing cultures in SDS-sample buffer (20). Following sonication and protein estimation, 50 μg protein was loaded onto 7.5% SDS–PAGE gels. Transfer to nitrocellulose and western blotting was performed as described (18). Proteins were detected using specific rabbit antibodies and visualized using either NBT/BCIP (anti-cMyc 237) or enhanced chemiluminescence (anti-Max).
In vitro translation of Myc and Max proteins and immunoprecipitation
Dimerization between Myc and Max was determined using methionine-labelled in vitro translated proteins essentially as described (19). Briefly, following incubation with Myc proteins, Max or Max9 were specifically immunoprecipitated using an anti-cMyc antibody, 237. The immune complexes were recovered on protein A–Sepharose beads, washed thoroughly and resolved on 10% SDS–PAGE gels. The methionine labelled proteins were detected by fluorography using Amplify (Amersham).
Reporter and activator plasmids
The PHO5 UAS-CYC-LacZ reporter plasmid, pRS314-Max/Max9, PHO4-cMyc (Pho4-cMyc) and Pho4 cMyc LZ mutant hybrids have been described previously (19). The PHO4-vMyc (Pho4-vMyc), Pho4 vMyc LZ mutants and Pho4-I>L10 hybrids were made by PCR of the appropriate templates and were cloned into the BglII site of pMA132 (19). The integrity of all expression constructs was verified by sequencing.
The Pho5 UAS-CYC-Lac Z reporter plasmid encodes the Pho4 DNA-binding sequence upstream of ?-galactosidase. This sequence, CACGTG, has also been shown to be a consensus sequence for the Myc/Max complex (22). The Pho4-Myc plasmids express the Pho4 transactivation domain fused to the b-HLH-LZ domain (amino acids 327–415) of the different Myc isoforms. pRS315-Max/Max9 expresses the two different Max isoforms.
Dimerization between the b-HLH-LZ domains of Myc and Max was detected by transforming the above plasmids into yeast. The Pho4–Myc/Max complexes formed in vivo will bind to the Pho5 UAS-CYC consensus sequence upstream of the LacZ gene. Transcription of LacZ will then be initiated by the Pho4 transactivation domain. The level of ?-galactosidase activity in cell lysates of the transformants will therefore be a direct measure of the extent of dimerization between the b-HLH-LZ domains of Myc and Max (19).
The construction of pGV256-lex-OP was as described (23). To generate pRS315-lex-Max9, the Max 9 coding sequence was first inserted as a BglII fragment into pV44ER-lex (23). The entire cassette comprising the GAL UAS, CYC promoter, Lex-Max9 and the CYC terminator was then cloned as an Sst1–Kpn1 fragment into pRS315.
These plasmids were used in an alternate dimerization assay that independently measures Myc/Max complex formation. pGV256-lex-OP contains the Lex operator upstream of CYC-LacZ (23). Lex A, the bacterial repressor, binds to the Lex operator. pRS315-lexA-Max 9 encodes a fusion protein between the bacterial Lex A repressor and Max 9, which will bind to the Lex operator. These plasmids were then transformed into yeast along with Pho4-Myc. Any Lex A-Max9:Pho4-Myc complexes that are formed will bind to the Lex operator sequence via LexA-Max9. The ?-galactosidase activity that is detected in the transformants will be the direct result of the complex formation between LexA-Max 9 and Pho4-Myc, since the Pho4 transactivation domain will mediate transcriptional activation. This assay therefore provides a dimerization assay which is independent of DNA binding mediated by the Myc:Max complex.
Yeast culture and ?-galactosidase assays
Reporter assays were performed in the yeast strain Y700 (, his3–11, ade2–1, leu2–3, –112, ura3, trp1–1 and can1–100) essentially as described (19). Following selection on yeast glucose minimal agar plates supplemented with the appropriate amino acids, the transformants were grown in liquid culture and ?-galactosidase activity measured (22). Activities represent the average of at least two independent duplicate cultures within the same experiment and units of ?-galactosidase activity were calculated as described (19).
RESULTS
vMyc can sustain longer truncations in its LZ domain than cMyc before loss of biological activity
Previous work from our laboratory showed that the integrity of the LZ domain was essential for Myc to function (18). Deletion of seven amino acids (removal of L5) in cMyc resulted in a mutant, cMyc7, which retained partial transforming activity (18). Deletion of a further three amino acids (removal of L4a and L5), however, resulted in a non-transforming mutant of Myc, cMyc10. We reasoned that vMyc might be able to sustain longer truncations in the LZ domain than cMyc before complete loss of biological activity. To address this, we generated a series of point mutants of vMyc which successively truncated leucine residues from the C-terminus. These mutants, designated vMyc7, vMyc10 and vMyc14, were cloned into avian retroviral vector, SFCV-sa– (24) and transfected into CEF. It should be noted that these constructs only encode the vMyc portion of MC29, and not the p110 gag-Myc encoded by the original MC29 virus (25). Following G418 selection, expression of all Myc-containing constructs was confirmed by western blot analysis (Figure 2). All were expressed in CEF, although the vMyc14 and cMyc14 mutants did appear to be expressed at lower levels than the wild-type. These low levels were unexpected. These mutants should bind Miz-1, but not Max (Figures 4 and 5). Since the former stabilizes Myc, and the latter has no affect on its half life (13), we would expect at least equal levels of the vMyc14 and cMyc14 mutants. Chicken Miz-1 has not, however, been characterized, and its levels in CEF are not known. The significance of the reduced expression of these two mutants is therefore not known. The level of Max9, the major Max isoform in CEF (17,26), does not change in the Myc-infected cells (Figure 2B).
Figure 2 LZ mutants of MC29 vMyc and avian cMyc. (A) Schematic representation showing location of MC29 vMyc mutations. Five mutations are contained within MC29 vMyc (11). These are found in the transactivation domain (T61>M), adjacent to the basic region (B) (serine325>leucine—S325>L), within helix 1 (serine350>arginine—S350>R) and helix 2 (I383>L) of the HLH domain, and within the LZ domain (lysine407>arginine—K407>R). LZ deletion mutants of MC29 vMyc and avian cMyc are shown. The leucine repeats (L1–L5) within the LZs of MC29 vMyc and avian cMyc are also indicated. Leucine L4a is shown offset to indicate its internal location (position 3) within the most C-terminal heptad repeat. Premature translation termination codons (indicated by an asterisk) were introduced into MC29 v-myc and avian c-myc (18) by site-directed mutagenesis. The 7 and 14 mutants truncate specifically at L5 and L4 within the heptad, respectively, whilst the 10 truncation removes L5 and L4a. (B) Retrovirally-expressed cMyc, vMyc and their respective LZ mutants in CEFs were detected by western blot of total cell lysates using an anti-Myc antibody (upper panel). Equivalence of loading was shown by Max9, the main isoform of Max in CEF (17,26), which was detected using an anti-Max antibody (21) (lower panel). Vector control is shown by a dash.
Figure 4 Complex formation between Max/Max9 and the LZ mutants of v- and c-Myc in vitro. cMyc, vMyc, LZ mutants and Max/Max9 were produced in rabbit reticulocyte lysate and Myc/Max complex formation determined by immunoprecipitation of Max using an anti-myc specific antibody. -labelled proteins were resolved by SDS–PAGE and visualized by fluorography. Central lane was Max/Max9 alone.
Figure 5 Complex formation between Max/Max9 and the LZ mutants of v- and c-Myc in yeast. Two different experimental systems based on sensitive yeast two hybrid assays were used to measure complex formation in vivo. (A) Max/Max9-dependent dimerization and DNA binding were determined in vivo. ?-Galactosidase activity was quantitatively measured as a result of transcriptional activation resulting from the co-transfection of Pho4-cMyc, Pho4-vMyc or their respective LZ mutants into yeast, together with the PHO5 UAS-CYC-LacZ reporter. This assay has been shown previously to a reliable assay for Myc/Max complex formation (28). (B) Dimerization which is independent of Myc/Max DNA binding was measured by co-transfecting Pho4-cMyc, Pho4-vMyc or their respective LZ mutants into yeast, along with the reporter, pGV256-lex-OP (23) and pRS315-lexA-Max9. ?-Galactosidase activity was then determined.
CEF which overexpress Myc undergo extensive changes in cell morphology, grow more rapidly and acquire the ability to grow in an anchorage independent manner (16,18). Therefore, having confirmed the appropriate retroviral expression, we determined changes in cell morphology (Figure 3A), calculated the growth rate (Figure 3B) and determined the ability to grow in agar of CEF overexpressing each Myc mutant (16,18).
Figure 3 vMyc can sustain longer deletions in the LZ than cMyc before loss of biological activity (A) Cell morphology of CEF expressing cMyc, vMyc and their respective LZ mutants are shown. (a) cMyc, (b) cMyc7, (c) cMyc10, (d) cMyc14, (e) vMyc, (f) vMyc7, (g) vMyc10, (h) vMyc14 and (i) vector control. (B) Growth rate of CEF expressing cMyc, vMyc and their respective LZ mutants. Growth rate was measured by cumulative cell counts over 4 days. These data are representative of at least two different experiments. (C) The ability of retrovirally-expressed cMyc, vMyc and their respective LZ mutants to induce anchorage-independent growth was determined by plating infected CEF into soft agar. Colony counts were taken after 14 days. Vector control is shown by a dash. These data are representative of at least two different experiments.
Control chick cells are spindle shaped, non-refractile and contact inhibited, lining up in a parallel manner on the dish (Figure 3A, i). In contrast, cMyc-infected CEF are rounder, more refractile and do not undergo contact inhibition (Figure 3A, a). cMyc7 (Figure 3A, b) had an intermediate phenotype, being partially morphologically-transformed. Deletion of 10 amino acids or more from the cMyc LZ, however, resulted in mutants that were non-transforming (Figure 3A, c and d). vMyc (Figure 3A, e) was more highly transformed than cMyc (Figure 3A, a), highlighting a major role for the viral mutations in potentiating the biological activity of Myc (9). vMyc (Figure 3A, e) and vMyc7 (Figure 3A, f) were morphologically indistinguishable from each other and were highly transformed. vMyc10 (Figure 3A, g) was partially transformed, however, vMyc14 (Figure 3A, h) was non-transformed. Thus, deletion of 14 amino acids from the vMyc LZ was required before loss in biological activity, whilst deletion of only 10 amino acids was required before cMyc became functionally inert. These data clearly show that vMyc can sustain longer truncations in its LZ than cMyc before loss of morphological transformation.
Another feature of Myc-transformed cells is an accelerated cell growth (18). The growth rate of Myc-infected CEF, determined by cumulative cell counts >4 days, is shown graphically (Figure 3B). As previously reported, cMyc and cMyc7 have accelerated growth rates, but further truncation of the LZ domain resulted in growth rates comparable with uninfected CEF (18). In contrast, vMyc, vMyc7 and vMyc10 grew more rapidly than control cells, whilst vMyc14 was indistinguishable from the control.
Loss of anchorage dependence by transformed cells was measured by the ability to grow in soft agar. Myc-infected CEF were plated in agar and after 2 weeks, plates were photographed. A minimum of eight frames were taken for each and the average number of colonies calculated (Figure 3C). It can clearly be seen that in contrast to vMyc7, cMyc7 only partially retains the ability to grow in agar. In contrast to cMyc10, vMyc10 was however still partially able to grow in agar.
Collectively, these data show that the integrity of the LZ domains of vMyc and cMyc are required for biological activity. More importantly, these results highlight a major functional difference in the behaviour of the vMyc and cMyc LZ domains, since vMyc can sustain longer truncations than cMyc before complete loss in its biological activity (Table 1). Indeed, deletion of 10 amino acids from the C-terminus of vMyc results in a mutant that still retains biological activity, whilst deletion of 10 amino acids from cMyc results in a mutant that is functionally inactive. This difference in the behaviour of the LZ domains must be a direct consequence of the virally derived mutations.
Table 1 LZ mutants of avian cMyc and vMyc are functionally different
Biologically active LZ mutants of vMyc retain the ability to dimerize with Max in vitro and in vivo
Since Myc requires Max to function (27), we set out to establish whether the vMyc LZ mutants retained the ability to dimerize with Max. -labelled Myc and Max were produced in an in vitro translation system, and Myc–Max complexes allowed to form at 37°C. Complexes were recovered by immunoprecipitation using an anti-Myc antibody that does not recognize Max (Figure 4). The central lane in both panels, indicated by a dash, contained Max or Max9 alone demonstrating that the Myc antibody does not immunoprecipitate Max. From these data, it can be seen that in contrast to vMyc, vMyc7 and vMyc10 which can complex with Max (upper panel) and Max9 (lower panel), only cMyc and cMyc7 complex with Max. Therefore, these data confirm that vMyc can sustain greater truncations to the LZ domain than cMyc before losing the ability to complex with Max. Importantly, all the transforming mutants of both vMyc and cMyc (Figure 3) bound to Max and Max9.
To confirm these data, we tested the ability of the LZ mutants of Myc to complex with Max in vivo using a yeast assay designed to measure dimerization and DNA binding of Myc:Max complexes. This assay has been used previously as a reliable and sensitive assay for Myc:Max function in vivo (19,28). In this assay, the bHLH-LZ domain of avian Myc or vMyc (amino acids 330–417) or their respective LZ mutants, were fused to the transactivation domain of the yeast transcription factor Pho4. When Pho4-Myc was transfected into yeast together with a PHO5 UAS-CYC-lacZ reporter plasmid, transcription through the PHO5-UAS occurred only in the presence of Max or Max9 (19). As expected, only Pho4-cMyc and Pho4-cMyc7 were able to activate transcription, whilst Pho4-cMyc10 and Pho4-cMyc14 did not. In contrast to cMyc, co-expression of Max or Max9 with Pho4-vMyc, Pho4-vMyc7 and Pho4-vMyc10 activated transcription through the PHO5-UAS, albeit at a reduced level for Pho4-vMyc10 (Figure 5A). These in vivo data confirm the in vitro results (Figure 4) clearly showing that vMyc can sustain a larger truncation to the LZ domain than cMyc before losing the ability to bind to Max.
As an independent measure of dimerization, we used an assay which asks whether Myc function was mediated by direct complex formation with Max, independently of its DNA binding activity. To achieve this, we fused the bacterial protein, LexA, to Max9, generating pRS315-lexA-Max9. When transfected into yeast, LexA binds to a specific DNA sequence, LexOP, which lies upstream of ?-galactosidase, however, no ?-galactosidase activity will be detected in the absence of Pho4-Myc. As expected, no activity was detected when pRS315-lexA-Max9 was transfected into yeast along with pGV256-lex-OP (23). When they were co-transfected with Pho4-Myc plasmids, however, ?-galactosidase activity was recorded. As can be seen from Figure 5B, significant ?-galactosidase activity was recorded with all the transforming mutants of cMyc and vMyc.
Collectively, these data (Figures 4 and 5) clearly show that the LZ domains of vMyc and cMyc differ significantly in their ability to bind to Max both in vitro and in vivo.
Mutation of I383>L in cMyc10 background partially restores binding to Max in vitro and in vivo
cMyc10 does not bind Max/Max9 (19) (Figures 4 and 5). Theoretically, the I383>L mutation of vMyc could extend the LZ domain from four to five heptad repeats and stabilize the Myc/Max interaction (Figure 1). To directly test this, we introduced this mutation into a cMyc10 background. This mutant, cMycI>L10 (Figure 6A) was tested for the ability to complex with Max both in vitro (Figure 6B) and in vivo (Figure 6C). In both these assays, replacement of isoleucine 383 with a leucine significantly restored binding to both Max and Max9. This mutant did not, however, significantly induce a biological phenotype when overexpressed in CEF, suggesting that the levels of cMycI>L10:Max complex may be below the threshold required to mediate a biological response. These data do, however, show that this single mutation is sufficient to contribute positively to the interaction of Myc and Max.
Figure 6 I383>L mutation in cMyc10 partially restores interaction with Max/Max9 in vitro and in vivo. (A) Schematic representation of cMyc, cMyc10 and cMycI>L10 mutants. I383>L mutation was introduced into cMyc10, which does not bind to Max (18). (B) Complex formation between Max/Max9 and cMyc, cMyc10, and cMyc I>L10 co-translated in vitro was determined by immunoprecipitation of Max/Max9 using an anti-Myc antibody, and -labelled proteins resolved by SDS–PAGE and visualized by fluorography. (C) Complex formation between Max/Max9 and Pho4-cMyc, Pho4cMyc10 and Pho4-cMycI>L10 was measured in vivo by co-transfecting the appropriate plasmids into yeast. ?-Galactosidase activity was then determined.
DISCUSSION
Somatic or virally derived mutations that potentiate the function of a protein can be viewed as naturally occurring protein engineering. Currently, many examples of these exist, which can result in proteins with altered properties, such as different half lives, differential responses to signalling pathways through loss in phosphorylation sites or altered protein:protein interactions. In the data presented here, viral mutations clearly enhance the Max binding properties of the b-HLH-LZ domain, since the vMyc LZ can tolerate greater loss of its dimerization interface than cMyc before loss of Max binding.
We propose that the I383>L mutation in MC29 vMyc represents an example of naturally occurring protein engineering. Given that HLH-LZ heterodimers are stabilized by hydrophobic and polar interactions involving the -helices, H1, H2 and LZ (29), the function of the I383>L mutation in the C-terminus of vMyc would be to contribute to establishing a more hydrophobic interface which stabilizes the Myc–Max complex.
Our data are wholly consistent with this proposal, since the introduction of a single amino acid change, I383>L, into cMyc10, a LZ mutant which is defective for Max binding, could significantly restore its ability to complex with Max both in vitro and in vivo. This amino acid substitution did not, however, restore biological activity, although small colonies were detected in agar (data not shown). Therefore, although this mutation contributed positively to the Myc:Max interaction, it is most likely that the levels of cMyc I>L10:Max complexes formed in CEF were below the threshold required to elicit any biological response (17).
The I383>L mutation is located in the C-terminal HLH domain, a region which is known to bind other proteins (30). This mutation may therefore regulate the interaction of Myc with these factors. This is thought highly unlikely given the extremely tight correlation between the ability of the different LZ mutants to bind to Max and transform cells, suggesting that it is the Myc:Max complex which is regulated by these mutations (19,27).
The behavioural differences between the respective LZ mutants of vMyc and cMyc clearly show that the virally derived mutations in vMyc impinge on the structure and function of its b-HLH-LZ domain. Whilst the integrity of the LZ domain of vMyc was still required for its biological activity, the v- and cMyc LZ mutants differ significantly with respect to both their biological activity and their ability to bind to Max (Figures 3–5). Given that the vMyc LZ can tolerate larger truncations from its LZ zipper domain than cMyc before loss in its biological activity and binding to Max, this must reflect a stronger interaction between the LZ domains of vMyc and Max than the LZ domains of cMyc and Max.
These findings could have major implications in the rational design of peptide inhibitors for cancer treatment, since a knowledge of the avidity of protein:protein interactions mediated through HLH-LZ dimerization could dictate the design of peptide inhibitors for use in a therapeutic context. Indeed, one such strategy was recently described which was based on the use of helix 1 (H1) peptides to inhibit the protein:protein interactions between Myc and Max resulting in a block to Myc-mediated cell proliferation (31). Although this study focussed on H1 rather than LZ peptides to block Myc function, it highlights the general applicability of targeting any essential dimerization interface with a view to inhibiting function. Our data could therefore have important implications if the target is a cancer cell which contains a mutated Myc that has a stronger dimerization interface than the wild-type Myc, since therapeutic targeting of this interface with a relatively weak peptide inhibitor would obviously be counterproductive.
Conversely, the stability of LZ interactions could be exploited to therapeutically intervene with Myc:Max complexes in cancer cells. An example of this was described by Jean-Francois et al. (29), who introduced two point mutations in Max, His82>Leu and Asn78>Val, with a view to increasing the hydrophobic interface between the two contributing Max monomers. This situation is similar to that described here, since the His82>Leu mutation increased the number of heptad repeats in the LZ from three to four. Their rationale was that the stabilized mutant Max homodimers would compete with Myc:Max heterodimers, block their binding to target E-boxes and as such, would act effectively as an anti-Myc drug (29). Indeed, this mutant dimer was subsequently shown to have improved thermodynamic stability and form more stable E-box complexes. Therefore, rather than titrating out Max to abolish Myc function (29), Myc peptides that provide a stronger dimerization interface could also theoretically be used to dominantly interfere with the formation of Myc:Max complexes. Obviously, this would only be applicable in a cancer cell that contained wild-type Myc.
In summary, we have shown for the first time a major difference between the C-termini of vMyc and cMyc, since vMyc can sustain greater truncations of the C-terminal LZ before loss in biological activity. Furthermore, we show that a single point mutation in helix 2 from vMyc can in isolation positively contribute to the interaction with Max. We therefore propose that this mutation provides a more hydrophobic surface between the dimerization interfaces of Myc and Max. Together with data showing the functional significance of a mutation in the transactivation domain (7), these data highlight the importance of virally derived mutations as examples of naturally occurring protein engineering.
ACKNOWLEDGEMENTS
We wish to thank Trevor Littlewood for help and reagents. This work was funded by Cancer Research UK and Association for International Cancer Research. Funding to pay the Open Access publication charges for this article was provided by JISC.
REFERENCES
Frame, M.C. (2004) Newest findings on the oldest oncogene; how activated Src does it J. Cell Sci., 117, 989–998 .
McCubrey, J.A., Shelton, J.G., Steelman, L.S., Franklin, R.A., Sreevalsan, T., McMahon, M. (2004) Effects of a conditionally active v-ErbB and an EGF-R inhibitor on transformation of NIH-3T3 cells and abrogation of cytokine dependency of hematopoietic cells Oncogene, 23, 7810–7820 .
Acquaviva, C., Salvat, C., Brockly, F., Bossis, G., Ferrara, P., Piechaczyk, M.P., Jariel-Encontre, I. (2001) Cellular and vial Fos proteins are degraded by different proteolyic systems Oncogene, 20, 942–950 .
Dai, T., Rubie, E., Franklin, C.C., Kraft, A., Gillespie, D.A., Avruch, J., Kyriakis, J.M., Woodgett, J.R. (1995) Stress-activated protein kinases bind directly to the delta domain of c-Jun in resting cells: implications for repression of c-Jun function Oncogene, 10, 849–855 .
Kanei-Ishii, C., Nomura, T., Tanikawa, J., Ichikawa-Iwata, E., Ishii, S. (2001) Differential sensitivity of v-Myb and c-Myc to Wnt-1-induced protein degradation J. Biol. Chem., 279, 44582–44589 .
Cole, M.D. and McMahon, S.B. (1999) The Myc oncoprotein: a critical evaluation of transactivation and target gene regulation Oncogene, 18, 2916–2924 .
Gavine, P.R., Neil, J.C., Crouch, D.H. (1999) Protein stabilization: a common consequence of mutations in independently derived v-Myc alleles Oncogene, 18, 7552–7558 .
Hoang, A.T., Lutterbach, B., Lewis, B.C., Yano, T., Chou, T.Y., Barrett, J.F., Raffeld, M., Hann, S.R., Dang, C.V. (1995) A link between increased transforming activity of lymphoma-derived Myc mutant alleles, their defective regulation by p107, and altered phosphorylation of the c-Myc transactivation domain Mol. Cell. Biol., 15, 4031–4042 .
Petropoulos, C.J., Givol, I., Hughes, S.H. (1996) Comparative analysis of the structure and function of the chicken c-myc and v-myc genes: v-myc is a more potent inducer of cell proliferation and apoptosis than c-myc Oncogene, 12, 2611–2623 .
Henriksson, M., Bakardjiev, A., Klein, G., Luscher, B. (1993) Phosphorylation sites mapping in the N-terminal domain of c-Myc modulate its transforming potential Oncogene, 8, 3199–3209 .
Walther, N., Jansen, H., Trachmann, C., Bister, K. (1986) Nucleotide sequence of the CMII v-myc allele Virology, 154, 219–223 .
Lutterbach, B. and Hann, S.R. (1994) Hierarchical phosphorylation at N-terminal transformation-sensitive sites in c-Myc protein is regulated by mitogens and in mitosis Mol. Cell. Biol., 14, 5510–5522 .
Salghetti, S., Kim, S.Y., Tansey, W.P. (1999) Destruction of Myc by ubiquitin-mediated proteolysis: cancer-associated and transforming mutations stabilize Myc EMBO J., 18, 717–726 .
Flinn, E.M., Busch, C.M., Wright, A.P. (1998) Myc boxes, which are conserved in myc family of proteins, are signals for protein degradation via the proteasome Mol. Cell. Biol., 18, 5961–5969 .
Muhle-Goll, C., Nilges, M., Pastore, A. (1995) The leucine zippers of the HLH-LZ proteins Max and c-Myc preferentially form heterodimers Biochemistry, 34, 13554–13564 .
Crouch, D.H., Gallagher, R., Goding, C.R., Neil, J.C., Fulton, R. (1996) Multiple phenotypes associated with Myc-induced transformation of chick embryo fibroblasts can be dissociated by a basic region mutation Nucleic Acids Res., 24, 3216–3221 .
La Rocca, S.A., Crouch, D.H., Gillespie, D.A.F. (1994) c-Myc inhibits myogenic differentiation and MyoD expression by a mechanism which can be dissociated from cell-transformation Oncogene, 9, 3499–3508 .
Crouch, D.H., Lang, C., Gillespie, D.A.F. (1990) The leucine zipper domain of avian cMyc is required for transformation and autoregulation Oncogene, 5, 683–689 .
Crouch, D.H., Fisher, F., Clark, W., Jayaraman, P.S., Goding, C.R., Gillespie, D.A.F. (1993) Gene-regulatory properties of Myc helix–loop–helix/leucine zipper mutants: max-dependent DNA-binding and transcriptional activation in yeast correlates with transforming capacity Oncogene, 8, 1849–1855 .
Batchelor, C.L., Woodward, A.M., Crouch, D.H. (2004) Nuclear ERM (ezrin, radixin, moesin) proteins: regulation by cell density and nuclear import Exp. Cell Res., 296, 208–222 .
Littlewood, T.D., Amati, B., Land, H., Evan, G.I. (1992) Max and c-Myc/Max DNA-binding activities in cell extracts Oncogene, 7, 1783–1792 .
Fisher, F., Jayaraman, P.-S., Goding, C.R. (1991) c-Myc and the yeast transcription factor PHO4 share a common CACGTG-binding motif Oncogene, 6, 1099–1104 .
Jayaraman, P.-S., Hirst, K., Goding, C.R. (1994) The activation domain of a basic helix–loop–helix protein is masked by repressor interaction with domains distinct from that required for transcription regulation EMBO J., 13, 2192–2199 .
Fuerstenberg, S., Beug, H., Introna, M., Khazaie, K., Munoz, A., Ness, S., Nordstrom, K., Sap, J., Stanley, I., Zenke, M., et al. (1990) Ectopic expression of the erythrocyte band 3 anion exchange protein, using a new avian retrovirus vector J. Virol., 64, 5891–5902 .
Lee, C.M. and Reddy, E.K. (1999) The v-myc oncogene Oncogene, 18, 2997–3003 .
Sollenberger, K.G., Kao, T.-L., Taparowsky, E.J. (1994) Structural analysis of the chicken max gene Oncogene, 9, 661–664 .
Luscher, B. and Larsson, L.-G. (1999) The basic region/helix–loop–helix zipper domain of Myc proto-oncoproteins: function and regulation Oncogene, 18, 2955–2966 .
Fisher, F., Crouch, D.H., Jayaraman, P.S., Clark, W., Gillespie, D.A.F., Goding, C.R. (1993) Transcription activation By Myc and Max: flanking sequences target activation to a subset of CACGTG motifs in vivo EMBO J., 12, 5075–5082 .
Jean-Francois, N., Frederic, G., Raymund, W., Benoit, C., Lavigne, P. (2003) Improving the thermodynamic stability of the leucine zipper of Max increases the stability of its b-HLH-LZ:E:box complex J. Mol. Biol., 326, 1577–1595 .
Sakamuro, D. and Prendergast, G.C. (1999) New Myc-interacting proteins: a second Myc network emerges Oncogene, 18, 2942–2954 .
Nieddu, E., Melchiori, A., Pescarolo, M.P., Bagnasco, L., Biasotti, B., Licheri, B., Malacarne, D., Tortolina, L., Castagnino, N., Pasa, S., et al. (2005) Sequence specific peptidomimetic molecules inhibitors of a protein–protein interaction at the helix 1 level of c-Myc FASEB J., 19, 632–634 .(D. H. Crouch*, F. Fisher1, S. A. La Rocc)