This study reveals that Evi1 is an important transcriptional regulator of adipocyte differentiation. Specifically, Evi1 acts in conjunction with C/EBPβ to increase expression of Pparγ2 and thus initiate the developmental gene program of adipogenesis.
gene is expressed in two isoforms (PPARγ1 and PPARγ2) from separate promoter regions. PPARγ2 is expressed specifically in adipose cells, whereas PPARγ1 is broadly expressed in many cell types (7
). PPARγ2 contains an extra 30 amino acids at the N terminus that boosts its ligand-independent transactivation function relative to PPARγ1 (28
). Surprisingly, however, the molecular events controlling the distinct expression patterns of Pparγ1
have remained elusive. Our results in both gain- and loss-of-function analyses reveal that Evi1 selectively controls Pparγ2
expression in adipocytes. This Evi1-dependent transcription of Pparγ2
is mediated, at least in part, via a physical interaction between Evi1 and C/EBPβ.
MECOM was required for the induction of Pparγ2
expression and adipogenesis in preadipocytes ( and ). Interestingly, we found that loss of MECOM blocked the mitotic clonal expansion (MCE) of 3T3-L1 cells (see Fig. S3C in the supplemental material) that normally occurs at the onset of differentiation (40
). The lack of MCE likely contributed to the reduced differentiation of MECOM-depleted cells. However, ectopic expression of Evi1 in NIH 3T3 cells caused the precocious expression of Pparγ2
prior to differentiation (B; see also Fig. S8A). Moreover, MECOM and C/EBPβ bound to regulatory elements near the Pparγ2
promoter at day 1 and day 2 of 3T3-L1 differentiation (D and E) when Pparγ2
expression increased (B). Together, our results strongly suggest that Evi1 directly regulates Pparγ2
transcription during adipocyte differentiation. Whether MCE facilitates the recruitment of Evi1 to the Pparγ
gene will require further study.
C/EBPβ is widely expressed in many cell types and has been shown to bind to many sites at or near the Pparγ
), but the functional significance of many of these sites has not been studied. Interestingly, Evi1 was detected by chromatin immunoprecipitation at only two of the six C/EBPβ binding regions (kb +2.6 and −183) (D and E), with the association of Evi1 with the kb +2.6 site being particularly robust, and this was lost in the absence of C/EBPβ (G). Thus, the presence of Evi1 with C/EBPβ may define the functional enhancers for Pparγ2
transcription. Evi1 was not able to increase C/EBPβ function in transient, plasmid-based transcription assays when Evi1-C/EBPβ binding regions (kb +2.6 and −183) were cloned upstream of a luciferase reporter gene. One possible explanation for this is that Evi1 recruits histone-modifying enzymes to cause structural changes in the chromatin around Pparγ2
rather than directly stimulating transcription. A detailed analysis of long-range DNA interactions at this locus will be needed to test this idea.
Evi1 and its longer form, Mds1-Evi1, have been shown in other cell lineages to activate or repress transcription through several proposed mechanisms, including via direct DNA binding (47
) or through coregulatory effects (as shown here). In particular, Evi1 interacts with numerous coregulators, including the corepressor CtBP2 (26
), acetyltransferases CBP and p300/CBP-associated factor (P/CAF) (4
), chromatin remodelers Brg1 and BRM (5
), histone methyltransferases SUV39H1 (3
) and Polycomb complex (49
), and DNA methyltransferase (35
). Together, these interactions suggest that Evi1 may have a fundamental role in coordinating the restructuring of chromatin in multiple genetic programs. Another interesting question is what determines the recruitment of Evi1 to these specific C/EBPβ DNA-binding sites and not others. The structural and mechanistic details that mediate the coactivator function of Evi1 will be an important area for future study.
Evi1 exists in at least two distinct forms, Mds1-Evi1 and Evi1, expressed from the MECOM locus through different promoters. The Mds-Evi1 isoform, but not Evi1, includes an amino-terminal PR domain that characterizes proteins in the PrdI-BF1-Riz1 histone methyltransferase family (9
). This arrangement suggests that Mds-Evi1 and Evi1 could have distinct functions at a common set of target loci recognized through their two identical zinc finger domains. Interestingly, the shorter Evi1 isoform was substantially more potent than Mds-Evi1 in inducing adipogenesis (B). Furthermore, the expression of the two isoforms during differentiation (A and B) strongly suggests a prominent role for Evi1 at the earliest time points, whereas the expression of Mds1-Evi1 remains detectable throughout adipogenesis. We speculate that the absence of a PR domain in Evi1 allows for distinct coactivating functions that are required during early adipogenic induction. The PR domain-containing Mds1-Evi1 may compete with residual Evi1 at later stages or substitute for Evi1 to maintain expression of adipogenic genes. It will now be important to examine the role of Evi1 and Mds1-Evi1 in mature adipocyte function using gain- and loss-of-function studies in cells and animals.
Evi1 is most closely related to Prdm16 within the 17-member PR domain protein family. Prdm16 is expressed at high levels in brown adipocytes relative to white adipocytes, where it drives a brown fat-specific gene program (34
). The structural and sequence homology between Evi1 and Prdm16 suggests that these factors may have some common or similar actions in adipose cells. Brown and white adipocytes arise from separate developmental origins; therefore, Prdm16 and Evi1 may regulate similar processes in brown and white fat lineages, respectively. For instance, Evi1, like Prdm16, stimulates adipogenesis through a physical association with C/EBPβ (17
), albeit in different cell types. However, Prdm16 protein has not been localized to specific regulatory regions in Pparγ
, and it will be interesting to examine whether Prdm16 binds in the Pparγ
locus to the same sites in brown fat cells as Evi1 does in white fat cells. Despite inducing common basic features of the fat phenotype, Prdm16 potently induces brown adipose-related genes (34
), whereas Evi1 does not. This result suggests that the ability of Prdm16 to activate the brown fat genetic program is mediated via a domain that is not shared with Evi1. Notably, white adipose tissue can acquire molecular and functional features of brown fat in response to prolonged cold exposure or after treatment with β-adrenergic agents, but the emergence of this “beige” fat requires Prdm16 (33
). Any role that Evi1 might play in this process remains to be studied. Conceivably, a balance between the levels of Evi1 and Prdm16 may determine the relative phenotypic expression of the white or brown fat programs.
In summary, we have identified Evi1 as a key competency factor that allows preadipocytes to undergo differentiation. Evi1 likely regulates other critical gene programs in adipocytes, functioning as a coregulatory protein with C/EBPβ and other transcription factors or as a direct DNA-binding transcription factor. Elucidating how Evi1 controls adipose expansion and adipocyte function will be essential for our understanding of adipose biology in development and disease.