The in vivo functional significance of PUF hub partnerships in C. elegans germline stem cells

Abstract

PUF RNA-binding proteins are conserved stem cell regulators. Four PUF proteins govern self-renewal of Caenorhabditis elegans germline stem cells together with two intrinsically disordered proteins, LST-1 and SYGL-1. Based on yeast two-hybrid results, we previously proposed a composite self-renewal hub in the stem cell regulatory network, with eight PUF partnerships and extensive redundancy. Here, we investigate LST-1-PUF and SYGL-1-PUF partnerships and their molecular activities in their natural context - nematode stem cells. We confirm LST-1-PUF partnerships and their specificity to self-renewal PUFs by co-immunoprecipitation and show that an LST-1(AmBm) mutant defective for PUF-interacting motifs does not complex with PUFs in nematodes. LST-1(AmBm) is used to explore the in vivo functional significance of the LST-1-PUF partnership. Tethered LST-1 requires this partnership to repress expression of a reporter RNA, and LST-1 requires the partnership to co-immunoprecipitate with NTL-1/Not1 of the CCR4-NOT complex. We suggest that the partnership provides multiple molecular interactions that work together to form an effector complex on PUF target RNAs in vivo. Comparison of LST-1-PUF and Nanos-Pumilio reveals fundamental molecular differences, making LST-1-PUF a distinct paradigm for PUF partnerships.

Keywords: CNOT complex; Intrinsically disordered proteins; Network hub; RNA repression.

Fig. 1.

PUF partnerships in the PUF hub and their biochemical analysis in nematodes. (A) GSC regulatory network. This simplified diagram depicts major regulatory hubs and how they relate to each other via repression (blunted line) or activation (arrow). The hubs control many hundreds to >1000 RNAs and thus promote GSC self-renewal (PUF hub) or differentiation (GLD-1, GLD-2 and FOG-1 hubs). See Kershner et al. (2013) and Hubbard and Schedl (2019) for more complete views. (B) PUF hub model. LST-1 and SYGL-1 are central to a composite regulatory hub: each is proposed to partner with any of four PUF proteins (FBF-1, FBF-2, PUF-3 and PUF-11; gray) and to repress differentiation RNAs for maintenance of the GSC pool. GLP-1/Notch signaling activates lst-1 and sygl-1 transcription (black arrows) at the distal end of the gonad, which restricts LST-1 and SYGL-1 expression to the GSCs. (C) LST-1^V5 distribution expands in glp-1(gf ts) mutants. Representative confocal z-projections of extruded gonads stained with a V5 antibody to detect LST-1^V5 (yellow) and with DAPI (cyan) for DNA. Red arrow marks spatially restricted LST-1^V5 at permissive temperature, as in wild type (Haupt et al., 2019; Shin et al., 2017); red line marks expanded LST-1^V5 in a germline tumor, formed at restrictive temperature. Asterisk indicates distal end of the gonad, and dotted line marks its boundary. (D) Subcellular distribution of LST-1^V5 in glp-1(gf ts) germline. Representative images of single confocal z-slices from the middle plane of distal region of extruded gonads stained with V5 antibody to detect tagged LST-1^V5 (yellow). LST-1^V5 is detected in both perinuclear puncta (white arrows) and the cytoplasm (red arrows). Inset shows higher magnification of the boxed area. (E) LST-1^V5 retains stem cell regulatory function when assayed in the absence of SYGL-1, both in a normal germline (row 2) and when expanded in glp-1 (gf ts) germline tumors (row 4), but not when it lacks its PUF-interacting motifs (row 6). (F) LST-1 protein architecture. LST-1 possesses an N-terminal ‘self-renewal’ domain composed of multiple IDRs (black lines along protein axis) and a C-terminal ‘spatial regulation’ domain with additional IDRs and a zinc finger (ZnF; ultramarine blue). Within the self-renewal region are two PUF interaction motifs (A and B), shown with wild-type (top) and mutant (bottom) sequences.

Fig. 2.

LST-1 associates specifically with PUF hub proteins in nematodes. (A-D) Western blots of input lysate and eluted samples after immunoprecipitation of epitope-tagged LST-1 from whole worms, after formaldehyde cross-linking. Blots were probed with relevant antibodies to detect epitope-tagged versions of LST-1, FBF-1 (A), FBF-2 (B), PUF-11 (C) and PUF-8 (D) as well as actin to see the loading control; 2% of input lysates and 20% of IP-eluted samples were loaded. Exposure times were different for input and IP lanes, so band intensities are not comparable. Arrows mark LST-1 and co-immunoprecipitated proteins. Gray dotted arrow indicates that PUF-8 did not co-immunoprecipitate with LST-1. Each coIP was repeated twice with similar results for the different replicates.

Fig. 3.

LST-1–FBF interaction is PIM dependent and RNA independent. (A) LST-1–FBF interaction requires LST-1 PIMs, PIM-A and PIM-B. Shown are western blots of input lysate and eluted sample after immunoprecipitation of epitope-tagged LST-1 from whole worms, after formaldehyde cross-linking. Blots were probed with anti-V5 antibody to see LST-1^V5, anti-FLAG for FBF-1^FLAG and FBF-2^FLAG, and anti-actin-4 for the loading control actin; 2% of input lysates and 20% of IP-eluted samples were loaded. Each coIP was repeated at least twice with similar results for different replicates. The red box highlights presence or absence of FBFs in the LST-1 immunoprecipitate. (B) LST-1–FBF interaction is independent of RNA. Shown are western blots of input lysate and eluted sample after immunoprecipitation of LST-1^V5 from whole worms, without formaldehyde cross-linking and with or without Benzonase. Blots were probed as described in A; 2% of input lysates and 20% of IP-eluted samples were loaded. Each coIP was repeated twice with similar results for the different replicates. The red box highlights FBF-1 in the LST-1 immunoprecipitate.

Fig. 4.

LST-1 repressive activity is PIM dependent in tethering assay. (A) Schematic of the tethering assay. LST-1^V5-λN carries a C-terminal V5 (yellow) and N-terminal λN22 (red). LST-1^V5-λN binds to BoxB hairpins for recruitment to reporter mRNA. (B) Quantitation of the effect of tethered LST-1 on reporter expression. GFP intensity was compared in the distal germline (1-40 µm from the DTC), where LST-1 is expressed at a high level, to GFP intensity more proximally (80-120 µm from the DTC), where LST-1 is expressed at a vanishingly low level. (C,D) Tethering results. Representative confocal images (maximum z-projection) of extruded gonads stained with anti-GFP (top) and anti-V5 (middle) antibodies and DAPI (bottom). GFP is green and LST-1 is yellow when tagged; DAPI marks all gonadal nuclei. An asterisk marks the distal end. (C) Tethering LST-1^V5-λN. Left: control, no tag and no reporter; middle: untethered LST-1, V5 and reporter but no λN22; right: tethered LST-1, V5 and λN22 plus reporter. An asterisk marks the distal end. (D) Tethering LST-1(A^mB^m)^V5-λN. Columns same as C. LST-1(A^mB^m)^V5 and LST-1(A^mB^m)^V5-λN are both restricted to distal end; LST-1(A^mB^m)^V5-λN does not repress reporter expression. Asterisk marks distal end. Dashed lines mark gonad boundary. (E,F) Boxplots of distal:proximal GFP intensity ratios. Each dot represents a separate sample. Boxes represent 25th-75th quantile; middle line, median; blue plus sign, mean; whiskers, minimum and maximum values. ***P<0.0001 (two-tailed Student's t-test). n.s., not significant. (Difference between LST-1^V5 and LST-1^V5-λN: P=1.25×10⁻¹⁸; difference between LST-1(A^mB^m)^V5 and LST-1(A^mB^m)^V5-λN: P=0.63). Sample sizes: LST-1^V5, n=35; LST-1^V5-λN, n=35; LST-1(A^mB^m)^V5, n=26; LST-1(A^mB^m)^V5-λN, n=26.

Fig. 5.

LST-1 association with NTL-1 is PIM dependent. (A) LST-1 colocalizes with NTL-1 in vivo. Representative deconvolved single confocal z-slices from middle plane of the distal region of an extruded gonad. Left: strain carrying both LST-1^V5 and NTL-1^FLAG; right: strain carrying both LST-1(A^mB^m)^V5 and NTL-1^FLAG. Row 1, V5 antibody detects LST-1 (magenta); row 2, FLAG antibody detects NTL-1 (green); row 3, DAPI highlights nuclei (cyan); row 4, merged images show co-staining with LST-1/NTL-1 overlap seen as white; insets, magnification of co-staining. Dashed line marks gonad boundary and asterisk marks distal end. (B) Variable colocalization of LST-1 and NTL-1. Images show representative examples of different degrees of overlap, taken from staining in A with further magnification. Code, shown below, is used in pie charts to show varying percentages of overlap with LST-1^V5 (left) and LST-1 (A^mB^m)^V5 mutant (right). Data in pie charts was generated from imaging ten gonads of each strain, with 200 LST-1 foci scored in the same region of each gonad (1-30 µm from the distal tip). (C) LST-1^V5 and NTL-1^FLAG co-immunoprecipitation. Western blots were probed with V5 antibody for LST-1, FLAG antibody for NTL-1, and actin-4 antibody for the loading control; 2% of input lysates and 20% of eluted samples were loaded. Exposure times of input and IP lanes are different, so band intensities are not comparable. The coIPs were repeated twice with similar results for the different replicates. The red box highlights presence or absence of NTL-1 in the LST-1 immunoprecipitate.

Fig. 6.

SYGL-1 has two PUF-interacting motifs and RNA repressive activity. (A) Diagram of SYGL-1 with its multiple IDRs (white lines internal to and along the axis of the rectangle representing the protein). Two candidate PIMs, PIM-A and PIM-B, were identified in the SYGL-1 amino acid sequence. (B) Conservation of PIM-A and PIM-B in SYGL-1 orthologs from related Caenorhabditid species. (C) Summary of SYGL-1 PIM effects on FBF binding, yeast two-hybrid assay. Superscript m denotes a mutant, with amino acid changes in red. +++, strong binding; ++, weaker binding; +, poor binding; −, no binding. (D) Summary of SYGL-1 PIM effects on GSC maintenance in nematodes. Mutation conventions as in C. SYGL-1 self-renewal activity was scored both in the presence of its LST-1 redundant counterpart as a control and in the absence of LST-1. (E) Spatial restriction of SYGL-1^V5 and SYGL-1(A^mB^m)^V5 to distal gonad. Representative confocal z-projections of extruded gonads stained with V5 antibody (yellow) and DAPI (cyan). Dashed line marks gonad boundary. (F) SYGL-1(A^mB^m)^V5 has lost self-renewal activity. Representative z-projected confocal images of extruded gonads stained with SP56 antibody (red) for sperm and DAPI (cyan). Dashed line marks gonad boundary and asterisks mark the distal end. Top: in the presence of wild-type LST-1, SYGL-1(A^mB^m)^V5 has no effect on GSC self-renewal; middle: in the absence of LST-1, SYGL-1(A^mB^m)^V5 cannot maintain GSCs because the germline is tiny and GSCs differentiated in early larvae to produce a few sperm: bottom: lst-1(ø) sygl-1(ø) germlines are similar to lst-1(ø) sygl-1(A^mB^m)^V5 germlines. (G) Number of total germ cells (GC) per animal in different strains. Total number of germ cells in lst-1(ø) sygl-1(A^mB^m)^V5 is more than lst-1(ø) sygl-1(ø), but fewer than lst-1(+) sygl-1(A^mB^m)^V5. (H) Tethered SYGL-1 reveals RNA repressive activity. Assay is same as in Fig. 4A, except λN22 is inserted at the C terminus of SYGL-1^V5. Images show representative z-projection of distal region in extruded gonads. Dashed line marks gonad boundary. Top: anti-GFP antibody detects GFP; middle, anti-V5 detects SYGL-1; bottom, DAPI highlights DNA within gonadal nuclei. Left: control, no tag and no reporter; middle: untethered SYGL-1, V5 and reporter but no λN22; right: tethered SYGL-1, V5 and λN22 plus reporter. (I) Boxplots of distal:proximal GFP intensity ratios. Conventions as described in Fig. 4E,F. ***P<0.0001 (two-tailed Student's t-test). P=1.296×10⁻¹⁴ between SYGL-1^V5 and SYGL −1^V5-λN. Sample sizes: SYGL-1^V5, n=23; SYGL-1^V5-λN, n=28.

Fig. 7.

Models for assembly and function of LST-1–PUF partnership in nematodes. (A) Assembly of LST-1–PUF partnerships. LST-1 (blue) can bind to PUF proteins (gray) via either of two PIMs (A and B) (Haupt et al., 2019). No RNA is depicted because LST-1–PUF assembly does not require RNA (this work). PUF proteins comprise an RNA-binding domain (RBD) and an N-terminal tail (wavy line) with IDRs (Fig. S6). LST-1 is largely intrinsically disordered (wavy line) and also has C-terminal zinc finger (ZnF; purple); LST-1 stem cell function resides in its IDR region and does not require the zinc finger. Conventions in A are also used in B-D. (B) LST-1 RNA repressive activity depends on PUF partnership. When tethered (left), the LST-1–PUF complex represses expression of the reporter RNA; a PUF protein is included in this diagram, because PIM-defective LST-1(A^mB^m) cannot repress reporter RNA expression when tethered (right). Tethering employs λN22 (red triangle) fused to LST-1 to bind BoxB stem loops in reporter RNA (see Fig. 4A). (C) LST-1–CNOT association depends on PUF partnership. Left: LST-1–PUF complex associates with CNOT complex (light yellow). Right: PIM-defective LST-1(A^mB^m) disrupts the LST-1–PUF complex and destabilizes CNOT association. Dark yellow and blue shapes represent CCF-1 and CCR-4 deadenylases. The model includes an interaction between the PUF protein and CCF-1, based on work with FBF-1 and FBF-2 (Suh et al., 2009). (D) The LST-1–PUF partnership brings together multiple interactions to form a stable complex with the CNOT complex and repress target RNAs. Left: the LST-1–PUF complex provides LST-1 IDRs and PUF IDRs; PUF protein contacts CCF-1 and also binds the PUF-binding element (PBE) in its target RNA. Right: PIM-defective LST-1(A^mB^m) disrupts LST-1–PUF and destabilizes the larger complex (LST-1–PUF–CNOT–target RNA), shown here as loss of all interactions for simplicity.