PH Domain-Arf G Protein Interactions Localize the Arf-GEF Steppke for Cleavage Furrow Regulation in Drosophila

Abstract

The recruitment of GDP/GTP exchange factors (GEFs) to specific subcellular sites dictates where they activate small G proteins for the regulation of various cellular processes. Cytohesins are a conserved family of plasma membrane GEFs for Arf small G proteins that regulate endocytosis. Analyses of mammalian cytohesins have identified a number of recruitment mechanisms for these multi-domain proteins, but the conservation and developmental roles for these mechanisms are unclear. Here, we report how the pleckstrin homology (PH) domain of the Drosophila cytohesin Steppke affects its localization and activity at cleavage furrows of the early embryo. We found that the PH domain is necessary for Steppke furrow localization, and for it to regulate furrow structure. However, the PH domain was not sufficient for the localization. Next, we examined the role of conserved PH domain amino acid residues that are required for mammalian cytohesins to bind PIP3 or GTP-bound Arf G proteins. We confirmed that the Steppke PH domain preferentially binds PIP3 in vitro through a conserved mechanism. However, disruption of residues for PIP3 binding had no apparent effect on GFP-Steppke localization and effects. Rather, residues for binding to GTP-bound Arf G proteins made major contributions to this Steppke localization and activity. By analyzing GFP-tagged Arf and Arf-like small G proteins, we found that Arf1-GFP, Arf6-GFP and Arl4-GFP, but not Arf4-GFP, localized to furrows. However, analyses of embryos depleted of Arf1, Arf6 or Arl4 revealed either earlier defects than occur in embryos depleted of Steppke, or no detectable furrow defects, possibly because of redundancies, and thus it was difficult to assess how individual Arf small G proteins affect Steppke. Nonetheless, our data show that the Steppke PH domain and its conserved residues for binding to GTP-bound Arf G proteins have substantial effects on Steppke localization and activity in early Drosophila embryos.

Fig 1. The PH domain of Step is necessary but not sufficient for its membrane localization and activity.

(A) Localization of GFP-Step constructs at early cellularization. N values indicate the number of embryos analyzed quantitatively. Amphiphysin (Amph) staining indicates the furrows. GFP-Step[WT] had a higher furrow membrane:cytoplasm ratio than the other constructs, as shown in the micrographs and quantifications of the ratios (means ± SD). GFP-Step[∆PH] also localized to structures that appeared to be centrosomes (yellow arrows). (B) GFP-Step[WT] over-expression produced a range of effects on furrows at early cellularization, from no effect (black) to sporadic membrane loss (red, arrows) to a general disruption of furrows including membrane loss and other disorganization (blue). Quantification of furrow defects (right) showed that the GFP-Step[∆PH] construct had a weaker effect on furrows than GFP-Step[WT] (N represents embryo numbers).

Fig 2. Conserved sequence and lipid binding properties of the PH domain of Step.

(A) The sequence of the PH domain and PB region of Step compared to that of mouse Grp1. The di-glycine sequence implicated in PIP3 binding is marked with asterisks. The Ile and Lys residues implicated in GTP-bound Arf small G protein binding are marked with pound symbols. (B) The amino acid sequences in (A) were expressed as GST fusions and exposed to lipid arrays at 0.08 μg/ml. Both showed similar preferential binding to PIP3 (red asterisks). (C) The PIP3 binding of the GST Step PH+PB protein was unaffected by amino acid residue changes expected to disrupt GTP-bound Arf small G protein binding and was lowered by conversion of the di-glycine sequence to a tri-glycine sequence (red asterisks). The proteins were exposed to lipid arrays at 0.08 μg/ml and the arrays were processed and imaged side-by-side (shown as a single photograph of all four arrays). Similar amounts and stabilities of the proteins were confirmed by separating 3 μg of each protein by SDS-PAGE and performing a Coomassie stain (D).

Fig 3. Amino acid residue changes affecting the localization of Step.

GFP-Step[WT] and GFP-Step[3G] had indistinguishable furrow localization (assessed as in Fig 1A, and additionally with line scans). Over-expression of either protein also induced sporadic furrow loss (red asterisks). GFP-Step[I319E] and GFP-Step[K319L] showed a lower furrow membrane:cytoplasm ratio (quantified as in Fig 1A), and also localized to structures that appeared to be centrosomes (yellow arrows).

Fig 4. Amino acid residue changes affecting the maintenance of Step at furrow membranes.

(A) Time points from FLIP analyses of GFP-Step[WT] and GFP-Step[I319E] at early cellularization. 0s shows the embryo before the first photobleaching. In the second column, the white bar shows where the embryo was photobleached, as well as cell compartments in Row 1 and Row 2 away from the cell row being bleached. The same positions were bleached at 60s intervals (arrows), and the embryos are shown 45s after each bleaching. The red asterisks indicate example Row 2 furrows that maintain their signal for GFP-Step[WT] or become depleted of signal for GFP-Step[I319E]. The images are also shown with an inverted Fire look up table (Image J) to show that a gradient of both cytosolic and membrane signal depletion arises with the repeated GFP-Step[I319E] photobleaching, but not for GFP-Step[WT]. (B) Quantification of the responses at Row 2 furrows as averages of three sites per embryo normalized to the signals at 0s (each line is data from one embryo). (C) Quantification of the same responses at Row 2 furrows as in B, but as averages of each single site over time points 225-405s normalized to the signals at 0s (each circle is data from one furrow). For B and C, note that Step[I319E] displayed the greatest increase in mobility versus Step[WT].

Fig 5. Amino acid residue changes affecting the activity of Step.

GFP-Step[WT] and GFP-Step[3G] had indistinguishable disruptive effects on furrows (assessed as in Fig 1B). GFP-Step[I319E] and GFP-Step[K351L] each showed a lower degrees of furrow disruption, despite similar furrow levels as GFP-Step[WT]. The difference was greater for GFP-Step[I319E].

Fig 6. Localization of GFP-tagged Arf and Arf-like small G proteins at early cellularization.

Live imaging revealed Arf1-GFP, Arf6-GFP and Arl4-GFP localization to furrows (arrows). Arf1-GFP also localized to strong cytosolic puncta resembling Golgi, and Arf4-GFP did as well.

Fig 7. The effects of Arf or Arf-like small G protein removal are distinct from those of Step removal.

Embryos with peripherally dividing nuclei at metaphase were identified by phospho-histone H3 staining (not shown). (A) Amph staining shows the intact furrows of a control (mCherry) RNAi embryo and the expansion of the furrow base (brackets) that occurs with step RNAi. arf1 RNAi resulted in furrow loss (arrows; note that the X-Y section showing the base of furrows is closer to the embryo surface in this case). arf6 null mutants showed no apparent defects, and arl4 RNAi embryos displayed no apparent defects despite an ability of the RNAi to effectively deplete Arl4-GFP (shown for early cellularization). (B) A quantification of the furrow phenotypes with Arf and Arf-like small G protein depletion. Two different arf1 shRNA lines produced similar defects with different strengths (data shown separately). Two different arl4 shRNA that each effectively depleted Arl4-GFP had indistinguishable effects on furrows (data combined). N represents embryo numbers.