Description

The predominant and ubiquitously expressed C. elegans histone H3 lysine 79 methyltransferase DOT-1.1 and its interacting partner, ZFP-1, play important roles in negatively modulating widely and highly expressed genes through promoter binding (Cecere et al. 2013) and in creating permissive environment for expression of developmentally regulated genes through enhancers (Esse et al. 2019; Esse and Grishok 2020). However, the mechanistic aspects of the ZFP-1/DOT-1.1 complex function are not clear. This study was motivated by the observation that genome-wide promoter-associated ZFP-1 and DOT-1.1 chromatin localization peaks (Mansisidor et al. 2011; Cecere et al. 2013) overlap with those of the O-GlcNAc modification on chromatin observed in wild type C. elegans (Love et al. 2010) (Figure 1A). Notably, the identities of the O-GlcNAc-modified proteins contributing to the O-GlcNAc ChIP-chip signals at the select promoters remain unknown. We chose to investigate the C. elegans basic helix- loop-helix leucine-zipper (bHLH/ZIP) transcription factor MML-1 (Myc and Mondo-like 1) (Pickett et al. 2007; Grove et al. 2009) as a candidate O-GlcNAc-modified protein, because both c-Myc (Chou et al. 1995) and the Mondo B protein (i.e., the Carbohydrate Response Element Binding Protein (ChREBP)) (Guinez et al. 2011) carry O-GlcNAc modification. MML-1 is a 1009 amino acid protein with multiple conserved domains, including the N-terminal Mondo Conserved Regions (I-V) (Pickett et al. 2007), which form a glucose sensing module in ChREBP (Abdul-Wahed et al. 2017), a proline-rich region, similar to that in ChREBP (Ortega-Prieto and Postic 2019), and a C-terminal DNA-binding and MXL-2 heterodimerization region called bHLH/ZIP (Pickett et al. 2007; Grove et al. 2009) (Figure 1B).

We obtained several viable mml-1 mutant alleles generated by the Million Mutation Project (Figure 1B, C) (Thompson et al. 2013). These included the potential null mutant mml-1(gk402844), which lacks the MML-1 protein signal recognized by the N-terminus-specific anti-MML-1 antibody (Figure 1C) (Pickett et al. 2007) (a generous gift from Dr. Donald Ayer). Notably, the mml-1(ok849) deletion allele (Figure 1B), which had been considered null (Johnson et al. 2014; Botts et al. 2016; Nakamura et al. 2016), leads to an in-frame deletion and produces an internally truncated protein (Figure 1C, D). We found that the O-GlcNAc-specific antibody RL2 immunoprecipitated both the wild type MML-1 and MML-1(ok849) (Figure 1D), which is consistent with our hypothesis that MML-1 is O-GlcNAc-modified. Moreover, both the wild type MML-1 and the MML-1(ok849) binding was detected at the promoters of the previously annotated ChIP targets of ZFP-1 (Mansisidor et al. 2011; Cecere et al. 2013) (Figure 1E). Notably, mml-1(gk402844) showed a background ChIP signal, whereas an increase in ChIP signal compared to that in wild type worms has been observed in mml-1(ok849) (Figure 1E). We believe that this increased promoter occupancy by MML-1(ok849) is due to the increased protein stability (Figure 1C, D), which suggests a role of the deleted region in regulating MML-1 dynamics.

Our interest in MML-1 has been driven by the hypothesis that MML-1 and ZFP-1/DOT-1.1 co-regulate common target genes. Since mml-1(ok849) phenocopied mml-1(RNAi) and behaved like a strong loss-of-function in published functional studies (Johnson et al. 2014; Nakamura et al. 2016), we were surprised to detect the truncated product by western immunoblotting (Figure 1C). However, the current annotation of the mml-1(ok849) allele is correct and predicts an in-frame deletion in MML-1, which results in a 608 aa long MML-1(ok849) (~67 kDa). This predicted size is consistent with our western data (Figure 1C, D).

To gain additional evidence that mml-1(ok849) and mml-1(gk402844) (a null) shared loss-of-function phenotypes, we analyzed Hermaphrodite-Specific Neuron (HSN) migration. The HSN under-migration phenotype is observed in zfp-1(ok554) reduction-of-function mutant, and it is, at least in part, due to the mis-regulation of pdk-1 expression in the hypodermis (Kennedy and Grishok 2014). Indeed, both mml-1(ok849) and mml-1(gk402844) worms displayed defects in HSN migration, 8% (n=49) and 45% (n=49), respectively (Figure 1F), analogous to those that we previously described for zfp-1(ok554) (25-30% affected animals) (Kennedy and Grishok 2014). We note that the mml-1(ok849) mutant animals exhibited both a less severe and a less penetrant HSN migration phenotype compared to the mml-1(gk402844) null. Interestingly, the MML-1/MXL-2 complex has earlier been implicated in cell migration in the male tail, and it acted non-autonomously, in the hypodermis, to promote cell migration (Pickett et al. 2007). In sum, these results suggest that MML-1/MXL-2 and ZFP-1/DOT-1.1 may act together in regulating genes expressed in the hypodermis to non-autonomously promote cell migrations during C. elegans development.

The MML-1 region deleted in MML-1(ok849) is proline-rich and similar to the proline-rich region in ChREBP, whose function is poorly understood (Ortega-Prieto and Postic 2019). Our findings, together with the previous studies of lifespan control using mml-1(ok849) and mxl-2(tm1516), which showed similar phenotypes and behaved like strong loss-of-function mutations (Johnson et al. 2014; Nakamura et al. 2016), indicate the importance of the proline-rich domain in the biological functions of MML-1. Thus, the existing data indicate that the MML-1(ok849) protein is less active, despite its binding to DNA and its O-GlcNAc modification. Moreover, they suggest that a posttranslational activation step targeting the proline-rich region of MML-1, and possibly ChREBP, is required for these DNA-binding proteins to be fully functional in transcription regulation.