Temporal regulation of foregut development by HTZ-1/H2A.Z and PHA-4/FoxA

Abstract

The histone variant H2A.Z is evolutionarily conserved and plays an essential role in mice, Drosophila, and Tetrahymena. The essential function of H2A.Z is unknown, with some studies suggesting a role in transcriptional repression and others in activation. Here we show that Caenorhabditis elegans HTZ-1/H2A.Z and the remodeling complex MYS-1/ESA1-SSL-1/SWR1 synergize with the FoxA transcription factor PHA-4 to coordinate temporal gene expression during foregut development. We observe dramatic genetic interactions between pha-4 and htz-1, mys-1, and ssl-1. A survey of transcription factors reveals that this interaction is specific, and thus pha-4 is acutely sensitive to reductions in these three proteins. Using a nuclear spot assay to visualize HTZ-1 in living embryos as organogenesis proceeds, we show that HTZ-1 is recruited to foregut promoters at the time of transcriptional onset, and this recruitment requires PHA-4. Loss of htz-1 by RNAi is lethal and leads to delayed expression of a subset of foregut genes. Thus, the effects of PHA-4 on temporal regulation can be explained in part by recruitment of HTZ-1 to target promoters. We suggest PHA-4 and HTZ-1 coordinate temporal gene expression by modulating the chromatin environment.

Figure 1. Enhancement of pha-4(ts) by mys-1, ssl-1, and htz-1

(A) Feeding dsRNA to wild-type (WT) or pha-4(ts) [18,25] worms at the permissive temperature of 24 °C. WT worms generate viable progeny with mys-1, ssl-1 or htz-1 RNAi. In the pha-4(ts) background, L1 arrest increased with mys-1, ssl-1, or htz-1 RNAi compared to control GFP(RNAi) (grey bars). Embryonic lethality remained unchanged (black bars). Effectiveness of RNAi feeding was manifest through viable, but sterile, progeny for mys-1 and ssl-1 [14], as well as repeated enhancement of L1 lethality for pha-4(ts), performed in parallel. n = 100 worms/plate, three plates per column. Error bars indicate the standard deviation. (B) Alignment of C. elegans htz-1 (R08C7.3) with human H2A.Z, yeast Htz1, and one of the core H2A genes from yeast, Hta1. Extended acid patch region essential for H2A.Z function is indicated by the bar [50,60,67].

Figure 2. mys-1, ssl-1, and htz-1 Enhance Pharyngeal Defects of pha-4(ts)

(A–C) Feeding dsRNA to pha-4(ts) worms at the permissive temperature of 24 °C. (A) A pha-4(ts); GFP(RNAi) L1 with a wild-type pharynx (arrowheads). (B) A pha-4(ts); ssl-1(RNAi) L1 with an unattached pharynx (arrowheads). (C) Quantitation of pha-4(ts) animals exhibiting a normal pharynx (WT), an unattached or incomplete pharynx (Pun), or no detectable pharynx. htz-1, ssl-1, or mys-1 RNAi significantly increased the number of Pun pha-4(ts) animals (htz-1: p=0.0290; ssl-1 and mys-1: p < 0.0001, Fisher exact test). (D–F) Feeding dsRNA to pha-4(ts) animals at the intermediate temperature of 20 °C. (D) A pha-4(ts); GFP(RNAi) L1 at the intermediate temperature of 20 °C with a morphologically wild-type pharynx (arrowheads). (E) A pha-4(ts); htz-1(RNAi) worm at 20 °C missing a detectable pharynx. (F) Quantitation of pha-4(ts) animals exhibiting a normal pharynx (WT), an unattached or incomplete pharynx (Pun), or no detectable pharynx. htz-1, ssl-1, or mys-1 RNAi significantly increased the number of worms with no detectable pharynx (htz-1 and ssl-1: p < 0.0001; mys-1: p = 0.0015; Fisher exact test). WT worms with reduced htz-1, ssl-1, or mys-1 activity had a wild-type pharynx at 24 °C and 20 °C (unpublished data). RNAi was conducted by feeding dsRNA [29].

Figure 3. Specificity of mys-1, ssl-1, and htz-1 Synergy with pha-4(ts)

The indicated worm strains were fed dsRNA for GFP (negative control) or mys-1, ssl-1, or htz-1 at 20 °C (or 24 °C for WT and pha-4(ts)). Lethal embryos (black bars) or lethal L1 progeny (grey bars) were scored for each strain. Effectiveness of RNAi feeding was manifest through viable, but sterile, progeny for mys-1 and ssl-1 [14], as well as repeated enhancement of L1 lethality for pha-4(ts). n =1 00 worms/plate, three plates per column. Error bars indicate the standard deviation.

Figure 4. Association of YFP::HTZ-1 with Pharyngeal Promoters

(A) Extrachromosomal target arrays were visualized by LacI::CFP (red) bound to the Lac operator (LacO). YFP::HTZ-1 (green) was excluded (arrowheads) from arrays with no-target promoter (row 1), but associated with target arrays containing promoters for pharyngeal genes myo-2 or R07B1.9 in pharyngeal cells (rows 2 and 3). Merge is yellow. YFP::HTZ-1 was excluded from arrays containing the myo-2 promoter in non-pharyngeal cells (row 4). Cartoons depict interpretation of data. (B) Percentage of embryos containing one or more co-localized LacI::CFP and YFP::HTZ-1 dots in the pharynx (black) or outside of the pharynx (grey). Association was significantly higher in the three myo-2 promoter lines (p < 0.0001) and two R07B1.9 lines (p = 0.0014). No significant difference in association was found when comparing the no-target lines to the heat-shock promoter target at 15 °C, 24 °C, or heat-shock at 33 °C for 30 min (p > 0.095). More than 60 embryos were scored for each no-target, heat-shock, and R07B1.9 line. More than 120 embryos were scored for each myo-2 line. YFP::HTZ-1 association was predominantly pharyngeal for myo-2 (p < 0.0001) and R07B1.9 (p = 0.0407) target arrays, but not control arrays (p = 1.00). An equivalent number of images were taken at the 1.5-, 2-, and 3-fold stages of embryogenesis for each line. The p-values were calculated using Fisher exact test.

Figure 5. Association of YFP::HTZ-1 Depends on pha-4 Activity

(A) Two PHA-4 binding sites within the myo-2 promoter were mutated in the myo-2 Mut construct [18]. (B) YFP::HTZ-1 association in the pharynx (black) and outside of the pharynx (grey). n = 384 embryos for three wild-type (WT) lines and 192 embryos for three mutant (Mut) lines. p-value indicates a significant difference for pharyngeal association of YFP::HTZ-1 in WT versus Mut. (C) Association of YFP::HTZ-1 with the wild-type myo-2 promoter (line #1) decreased in pha-4(RNAi) embryos within pharyngeal cells. An equivalent number of images were taken at the 1.5-, 2-, and 3-fold stages of embryogenesis and >80 embryos were imaged for each trial. p-values calculated using Fisher's exact test.

Figure 6. YFP::HTZ-1 Association Peaks at the Onset of myo-2 Expression

(A) Association of YFP::HTZ-1 with the myo-2 promoter in the pharynx during the comma, 1.5-, 2-, and 3-fold stages of development (two trials: black, dark grey). YFP::HTZ-1 association in the pharynx does not peak at the 2-fold stage for a myo-2 promoter with both PHA-4 binding sites mutated (Mut, light grey). n > 60 embryos imaged at each stage, for each experiment. Multiple lines were used for each trial. (B) Pharyngeal YFP::HTZ-1 association with myo-2 target array #1 decreased at the 2-fold stage when pha-4 activity was reduced by RNAi. n > 20 embryos imaged at each stage for each of the two wild-type (WT) and two pha-4(RNAi) experiments. The p-values for (A) and (B) indicate the significance of the WT peak at the 2-fold stage as calculated by a repeated measures analysis of variance (ANOVA).

Figure 7. htz-1 Influences the Onset of Late Pharyngeal Gene Expression

(A) Onset of R07B1.9::GFP and myo-2::GFP after the comma stage for wild-type (black) or htz-1(RNAi) (grey) embryos. Activation of the R07B1.9 reporter was delayed after injection of htz-1 dsRNA (p < 0.0001, Student t-test). No RNAi: n = 18, htz-1 RNAi: n = 12 embryos. myo-2 activation was scored in embryos carrying either a wild-type myo-2::GFP reporter (WT) or a myo-2::GFP reporter with both PHA-4 binding sites mutated (Mut) [18]. Activation of the WT reporter was delayed after injection of htz-1 dsRNA (p < 0.0001, Student t-test), whereas onset of the Mut reporter was unchanged (p = 0.2068). No RNAi: n = 15 WT, n = 33 Mut Embryos; htz-1 RNAi: n = 9 WT, n = 21 Mut embryos. Error bars indicate the standard deviation. (B) Expression of R07B1.9::GFP at 90 and 180 min after the comma stage with htz-1 RNAi or without (No RNAi). Pharyngeal R07B1.9::GFP is indicated by arrowheads. (C) Expression of myo-2::GFP (WT) at 150, 200, and 250 min after the comma stage with htz-1 RNAi or without (No RNAi). (D) Expression of myo-2::GFP lacking PHA-4 binding sites [18] at 150, 200, 250 min after the comma stage with htz-1 RNAi or without (No RNAi).

Figure 8. htz-1 Influences the Onset of Early PHA-4–Dependent Expression

(A) Onset of 3XPRE::GFP after the two-cell stage for wild-type (black) or htz-1(RNAi) (grey) embryos. 3XPRE::GFP is a reporter construct with three copies of a high-affinity PHA-4 response element upstream of the Δpes-10 promoter that reproducibly activates early pharyngeal expression [19]. Activation of the 3XPRE::GFP reporter is delayed after injection of htz-1 dsRNA (p = 0.0289, Student t-test). No RNAi: n = 23, htz-1 RNAi: n = 9 embryos. Error bars indicate the standard deviation. (B) Expression of 3XPRE::GFP at 150 and 200 min after the two-cell stage with htz-1 RNAi or without (No RNAi).

Figure 9. Summary and Model: HTZ-1 Synergizes with PHA-4 to Establish the Foregut

(A) Summary of the data presented in this paper and in [18]. myo-2 expression initiates at the 2-fold stage, and onset at this stage requires pha-4 and htz-1 because mutation of PHA-4 binding sites (Mut) or htz-1 RNAi lead to a delay in myo-2 activation. HTZ-1 association with the myo-2 promoter peaks at the onset of myo-2 transcription, and this association requires pha-4. (B) Model to explain how HTZ-1 synergizes with PHA-4. PHA-4 association with the myo-2 promoter leads to exchange of H2A-containing nucleosomes for one or more nucleosomes carrying HTZ-1/H2A.Z at the 2-fold stage. Based on data from other organisms [1], we propose MYS-1 and/or SSL-1 function in a complex that performs the exchange reaction. HTZ-1–containing nucleosomes promote transcriptional activation by the 2-fold stage.