The Myc-Max-Mad/Mnt network of transcription factors has been implicated in oncogenesis and the regulation of proliferation in vertebrate cells. The identification of Myc and Max homologs in Drosophila has demonstrated a critical role for dMyc in cell growth control. The third member of this network, dMnt is the sole fly homolog of the mammalian Mnt and Mad family of transcriptional repressors. dMnt possesses two regions characteristic of Mad and Mnt proteins: a basic helix-loop-helix-zipper domain, through which it dimerizes with dMax to form a sequence-specific DNA binding complex, and a Sin-interacting domain, which mediates interaction with the dSin3 corepressor. Using the upstream activation sequence/GAL4 system, it was show that expression of dMnt results in an inhibition of cellular growth and proliferation. Furthermore, a dMnt null allele has been generated, that results in flies with larger cells, increased weight, and decreased life span compared to wild-type flies. These results demonstrate that dMnt is a transcriptional repressor that regulates Drosophila body size (Loo, 2005).

The Myc-Max-Mad/Mnt network is comprised of a group of conserved transcription factors of the basic helix-loop-helix-zipper (bHLHZ) class that are thought to function together as a molecular module to transcriptionally regulate cell growth, proliferation, and differentiation. The bHLHZ domains common to these transcription factors mediate dimerization of Myc or Mad family proteins with Max, thereby permitting Myc-Max and Mad-Max heterodimer binding to the E-box sequence CACGTG. An important aspect of the network is that it is capable of transcriptional activation and repression of multiple gene targets through recruitment of chromatin-modifying complexes. Myc associates with the coactivators TRRAP and p300/CBP, which bind or possess histone acetyltransferase activities. Histone acetyltransferase recruitment is generally associated with augmented gene expression. In contrast, Mad and Mnt family proteins associate with the Sin3 corepressor, which recruits histone deacetylases, leading to transcriptional repression. Myc-Max complexes have also been shown to repress the expression of several genes indirectly by binding and inactivating the Miz-1 transcription factor (Loo, 2005 and references therein).

Much research has focused on mammalian Mad family proteins and Mnt, since they appear to antagonize Myc activity and could function, at least in principle, as tumor suppressors. The mammalian Mad family of transcriptional repressors is encoded by four paralogs: mad1, mxi1, mad3, and mad4. The other characterized Max-binding repressor, Mnt (also known as Rox), possesses the two conserved domains common to all Mad family members: the N-terminal Sin-interacting domain (SID), which interacts with the Sin3 corepressor, and the bHLHZ domain required for heterodimerization with Max (Hurlin, 1997; Meroni, 1997). However, Mnt is considerably larger than any of the Mad family proteins and contains other regions, including proline- and proline/histidine-rich sequences, that are unique to Mnt (Loo, 2005 and references therein).

Attempts to understand the physiological roles of mammalian Mad and Mnt proteins have entailed both overexpression and targeted gene deletion studies. In general, this work has provided support for the notion that Mad and Mnt antagonize Myc (Loo, 2005 and references therein).

Surprisingly, given the capacity for overexpressed mad to inhibit proliferation, targeted deletions of mad family genes in mice do not result in dramatic phenotypes relating to differentiation and development. Mice homozygous for a mad1 null mutation display no detectable differences in viability, fertility, size, behavior, or incidence of neoplasia compared to controls. However, their granulocyte progenitor cells, when cultured in vitro, exhibit a delay in differentiation due to an inhibition of cell cycle exit (Foley, 1997). mxi1 knockout mice showed generalized hyperplasia in certain tissues and an increased incidence of carcinogen-induced tumors in older animals (Schreiber-Agus, 1998). The only detectable phenotype in mad3 homozygous null mice is an increased sensitivity to gamma irradiation in neural progenitor cells and thymocytes (Queva, 2001). These subtle phenotypes may be the result of redundancy with other Mad family members or with other cell cycle regulatory proteins. In fact, an apparently compensatory increase in Mxi1 and Mad3 expression was observed in the thymi of mad1 knockout mice (Foley, 1997). Therefore, as suggested for Myc family gene deletions, functional redundancy may obscure the developmental roles of mad genes in targeted deletion studies. This may be less true for Mnt, whose targeted deletion in mice results in craniofacial abnormalities and perinatal lethality (Toyo-oka, 2004). Conditional loss of mnt in murine breast epithelium leads to adenocarcinomas (Hurlin, 1997). Furthermore, loss of Mnt function in fibroblasts results in enhanced proliferation and upregulation of Myc target genes (Hurlin, 1997, Nilsson, 2004). Thus, mammalian Mnt is a tumor suppressor that presumably functions to antagonize Myc activity. However, because these studies of Mnt have been carried out in settings where wild-type Mad1 to Mad4 proteins are expressed, it is difficult to sort out the extent and consequences of the overlapping functions of these proteins (Loo, 2005 and references therein).

The identification of Myc and Max homologs in Drosophila melanogaster has greatly facilitated genetic analysis of candidate functional pathways and targets. The Drosophila homologs of Myc and Max (dMyc and dMax) heterodimerize, bind E-box sequences, activate transcription, and in the case of dMyc, recapitulate mammalian Myc functions. Furthermore, both loss-of-function and gain-of-function studies with dMyc have found it to be a positive regulator of cell growth (i.e., cell mass). The third member of the Max network in Drosophila is dMnt, the sole homolog of mammalian Mad and Mnt proteins. dMnt is thought to be a negative regulator of cell growth that also plays a role in regulating Drosophila body size and life span (Loo, 2005).

dMnt possesses both the basic helix-loop-helix-zipper and the SID domain characteristic of mammalian Mnt and the four Mad family proteins. Both these domains are functional -- the dMnt bHLHZ mediates association with Max and DNA binding, while the dMnt SID interacts directly with the Sin3 corepressor. In contrast, dMnt lacks the proline-rich and proline/histidine-rich domains characteristic of mammalian Mnt. Because at this time there is no data indicating that these domains are required for Mnt function, their absence in Drosophila Mnt cannot be taken as evidence of functional divergence. It is surmised that dMnt represents a progenitor of the vertebrate Mad family and Mnt proteins. This notion is consistent with the developmental patterns of expression observed for dMnt that appear similar to those for mammalian Mad proteins (expressed during cell cycle exit) as well as Mnt (expressed in proliferating cells). The numerous vertebrate paralogs are likely a reflection of the more complex and diverse regulation imposed on these genes during vertebrate development (Loo, 2005).

It is probably not a coincidence that the alternatively spliced forms of dMnt that were observed observed involve functionally significant alterations in the SID and bHLHZ domains. In contrast to dMnt, which functions as a transcriptional repressor, neither the dMnt DeltaZip protein nor the dMnt DeltaSID protein displays positive or negative transcriptional activity on a synthetic reporter gene containing E-box binding sites. These forms could potentially act as dMnt dominant interfering proteins. However, both the dMnt DeltaZip and dMnt DeltaSID forms contain unique sequences introduced by alternative splicing and not found in dMnt. It is speculated that it is the polypeptide regions that are unique to the DeltaZip and DeltaSID isoforms that confer distinct E-box- and Sin3-independent functions on these proteins. For example, overexpression of dMnt, but not dMnt DeltaZip or dMnt DeltaSID, has strong effects on wing and eye formation. However, while all three isoforms reduced the size of random clones in the wing disc, dMnt has the most dramatic effects and also influences cell size and cell cycle phasing. By contrast, the DeltaZip and DeltaSID isoforms had no detectable effects on cell size or cell cycle. It will be of interest to further explore potential protein interactions mediated by the alternatively spliced sequences in the DeltaZip and DeltaSID isoforms and to delineate in greater detail their expression patterns and biological roles (Loo, 2005).

An imprecise excision screen of an existing allele containing a P-element insertion in the dmnt gene generated dmnt¹, a null allele, and dmnt², an allele that expresses only the dMnt DeltaSID form. Surprisingly, both mutant alleles are viable with no obvious pattern of defects or effects on male or female fertility. It was, however, observed that dmnt¹ and dmnt² mutants were 20% and 12% heavier than their wild-type controls, respectively. Furthermore, dmnt¹ and dmnt² alleles had an average of 18% and 10% fewer cells per unit area, respectively, than the wild-type controls, indicative of increased cell size. The intermediate effect observed with dmnt², which expresses only dMnt DeltaSID, again suggests that this alternatively spliced form functions in a pathway related to the dMnt pathway. The augmentation of cell and organismal size in the dMnt mutant flies is generally consistent with dMnt acting as an antagonist of dMyc. Flies with dMyc loss-of-function alleles are smaller and have increased trichome density, while dMyc overexpression generates larger cells and flies. Mice with single deletions of murine mad family genes mad1, mxi1, and mad3 are viable, although hyperplasia, sensitivity to radiation damage, and a delay in differentiation is observed. The relatively mild phenotypes have been generally ascribed to redundancy between mad family members. The data derived from the dmnt null mutant suggest that mad genes are not essential for survival. However, a strong synthetic effect in mice bearing both mad1 and p27^KIP1 null alleles indicates that mad genes may be redundant with other genes that act to limit cell growth and proliferation. Interestingly, no obvious genetic interaction or developmental or adult phenotypes were seen in mutant flies that were hemizygous for dmnt¹ and heterozygous for dacapo, the sole Drosophila p21^CIP1/p27^KIP1 ortholog (Loo, 2005).

The finding that both the dmnt loss-of-function alleles result in a 24% decrease in life span and a diminished life expectancy is surprising. Although it is possible that dmnt loss of function might have a nonspecific toxic effect, this is thought unlikely because no sensitivity to starvation or bacterial infection was observed in the dmnt¹ mutant flies. Nonetheless, increased longevity was not found using a dMnt transgene and a motor neuron-specific GAL4 driver used previously to overexpress superoxide dismutase and increase life span. However, a negative result with the transgene may well be due to a failure to target expression to the appropriate cell type or developmental time. While this work was in progress, the mad/mnt homolog, mdl-1 in Caenorhabditis elegans (Yuan, 1998) was identified as a positively regulated target of the DAF-16 transcription factor (Drosophila homolog: Foxo), which is known to influence the rate of aging during early adulthood. In addition, mdl-1 small interfering RNA results in a ~10% decrease in the life span of C. elegans. Perhaps Drosophila Mad/Mnt homologs function as part of a conserved pathway involved in longevity. Clearly, further studies will be required to define dMnt's possible role in aging. Interestingly, mammalian c-Myc has been recently shown to have the ability to attenuate cellular senescence and promote transformation through its direct regulation of the Werner syndrome gene, also known to function in aging (Loo, 2005).