Condensation of the fusion focus by the intrinsically disordered region of the formin Fus1 is essential for cell-cell fusion

Abstract

Secretory vesicle clusters transported on actin filaments by myosin V motors for local secretion underlie various cellular processes, such as neurotransmitter release at neuronal synapses,¹ hyphal steering in filamentous fungi,^2,3 and local cell wall digestion preceding the fusion of yeast gametes.⁴ During fission yeast Schizosaccharomyces pombe gamete fusion, the actin fusion focus assembled by the formin Fus1 concentrates secretory vesicles carrying cell wall digestive enzymes.^5,6,7 The position and coalescence of the vesicle focus are controlled by local signaling and actin-binding proteins to prevent inappropriate cell wall digestion that would cause lysis,^6,8,9,10 but the mechanisms of focusing have been elusive. Here, we show that the regulatory N terminus of Fus1 contains an intrinsically disordered region (IDR) that mediates Fus1 condensation in vivo and forms dense assemblies that exclude ribosomes. Fus1 lacking its IDR fails to concentrate in a tight focus and causes cell lysis during attempted cell fusion. Remarkably, the replacement of Fus1 IDR with a heterologous low-complexity region that forms molecular condensates fully restores Fus1 focusing and function. By contrast, the replacement of Fus1 IDR with a domain that forms more stable oligomers restores focusing but poorly supports cell fusion, suggesting that condensation is tuned to yield a selectively permeable structure. We propose that condensation of actin structures by an IDR may be a general mechanism for actin network organization and the selective local concentration of secretory vesicles.

Keywords: FUS; IDR; actin cytoskeleton; cell-cell fusion; condensate; cryptochrome CRY2; formin; fused in sarcoma; intrinsically disordered region; myosin V Myo52; optogenetics; yeast Schizosaccharomyces pombe.

Graphical abstract

Figure 1

Fus1N is essential for fusion and has localization and self-association properties (A) Formin chimeras tagged C-terminally with sfGFP. (B) DIC and GFP images ∼16 and ∼8 h post starvation of fus1Δ cells expressing the chimeric formins shown in (A). Yellow dashed lines outline mating pairs. (C) Percentage of cell pair fusion and lysis 24 h post starvation in WT and strains as in (B). p values relative to WT. (D) Interphase cells expressing Myo52-tdTomato and full-length Fus1-sfGFP from the nmt1 promotor. Cells were either untreated (left), treated with 200 μM latrunculin A (middle), or with 20% 1,6-hexanediol (right) for 5 min. White arrows mark resistant fusion focus-like structure; yellow arrowheads indicate labile Myo52 dots. (E) Interphase cells expressing Myo52-tdTomato and Fus1N-sfGFP (Fus1^1–792) from the nmt1 promoter. Cells were either untreated (left) or treated with 20% 1,6-hexanediol for 5 min (right). (F) Scheme of Fus1N with predicted domain organization. The top graph shows the disorder index of 3 prediction tools.^,, Fragments were C-terminally tagged with sfGFP. The localization summary is shown on the right. (G) GFP-fluorescence images of constructs as in (F). (H) Boxplot of Fus1 clusters mean fluorescence intensity of cells as in (E). The p value relative to untreated condition. (I) Average Fus1N FRAP recovery curves normalized to pre-bleach values in cells as in (E). The mean recovery half-time and standard deviation are indicated. N = 3 independent experiments, with n > 17 cells each (n > 54 cells in total). The shaded area shows the standard error. Scale bars, 5 μm. See also Figure S1.

Figure 2

Fus1 assemblies exclude ribosomes (A and B) Virtual z-slices through electron tomograms taken at the contact site of (A) WT and (B) acp2Δ cell pairs during the fusion process. The transparent cyan shape outlines regions devoid of ribosomes. Images on top show the transmitted light image and fluorescence of Fus1-sfGFP (green) and Myo52-tdTomato (magenta). (C) Vesicle density at the contact zone. The p value is shown. (D) Cross-section area of the ribosome-free zone in cells as in (A) and (B). (E–H) Virtual z-slices through electron tomograms of vegetative cells at the position of Fus1-sfGFP (E and F) or Fus1N^93–792-sfGFP (G and H). Images on the left show tomograms with and without the correlated fluorescence image (green) and fiducial beads (yellow and arrows in G and H). Scale bars, 100 nm, except for (E)–(H) (left), 500 nm. See also Video S1.

Figure 3

Fus1 IDR concentrates Fus1 and is essential for fusion (A) Merge and GFP images ∼8 h post starvation of Myo52-tdTomato and Fus1N-sfGFP (Fus1^1–792) expressed in fus1Δ or WT cells. (B) Normalized Fus1N-sfGFP fluorescence profiles perpendicular to the mating pair axis at the time of cell fusion, in strains as in (A). (C) Width at half-maximum (D50) of the profiles shown in (B). (D) Fluorescence images ∼8 h post starvation of fus1Δ or WT cells expressing the indicated Fus1N-sfGFP allele. (E) D50 of GFP-fluorescence profiles in strains as in (D), in the WT background. (F) DIC and fluorescence images ∼16 and ∼8 h post starvation of cells expressing the indicated Fus1-sfGFP allele from the native fus1 locus. The white arrow points to a lysed pair. Lysis is under-represented, as it mostly happens at later time points. (G) D50 of GFP-fluorescence profiles in strains as in (F). (H) Boxplot of WT and fus1^ΔIDR FRAP half-times. (I) Percentage of cell pair fusion and lysis 24 h post starvation in strains as in (F). p values relative to left-most strain. Scale bars, 5μm.

Figure 4

Fus1 IDR can be functionally replaced by self-assembling domains (A) DIC and fluorescence images ∼16 and ∼8 h post starvation of cells expressing the indicated Fus1 allele from the native fus1 locus either tagged with sfGFP and in combination with Myo52-tdTomato (upper panels) or untagged and in combination with mNeonGreen-Cdc8. FUS and CRY2 variants were introduced in Fus1^Δ492–791. Cells were exposed to blue light every 5 min for several hours. (B) Width at half-maximum (D50) of Fus1-sfGFP-fluorescence profiles in strains as in (A). fus1^Δ492–791 is shown for comparison. (C) Percentage of cell pair fusion and lysis 24 h post starvation under continuous white light (+) or in the dark (−) in strains with Fus1 or Fus1-CRY2 alleles. (D) Boxplot of Fus1-sfGFP focus fluorescence intensity at fusion time. (E) Percentage of cell pair fusion and lysis 24 h post starvation in strains with Fus1 or Fus1-FUS alleles. (F) Boxplot of fusion times in strains with Fus1 or Fus1-FUS alleles. (G) Average Fus1 FRAP recovery curves normalized to the maximal recovery value. The mean recovery half-time and the standard deviation are indicated. N = 4, 3, and 2 experiments for the WT, fus1^Δ492–791, and the 4 other alleles, respectively, with n > 9 cells each (n > 47 cells in total). The shaded area shows the standard error. (H) Boxplot of Myo52 focus fluorescence intensity at fusion time in strains as indicated. (I) Boxplot of Cdc8 focus fluorescence intensity at fusion time in strains as indicated. Scale bars, 5 μm. Black p values aligned with bars are relative to WT; gray ones to fus1^Δ492–791; p values between bars compare the two conditions. See also Figure S2 and Video S2.