Egl512-819 is sufficient for RNA binding.

(A) Schematic representation of the architecture of full-length Egl and the truncated constructs used. Solid boxes represent domains predicted to be folded, while lines are regions predicted to be unstructured. A conservation plot based on the alignment of multiple Egl homologs is shown below (Scores based on Consurf(Ashkenazy et al., 2016)). Sequences used to generate the alignment plot are shown in Figure S1. In the plot, residues positions are represented in the x-axis and percent identities (50-100%) are indicated in the y-axis. (B) Kd values were determined by FA. 5′-FAM-labeled K10 TLS was incubated with increasing concentrations of Egl constructs and the data were fitted to the Hill equation to obtain the Kd. (C) Column graph with mean Kd values shown as bars, standard deviation as black error bars and individual measurements as black dots. Data are represented as mean ± standard deviation (SD). (D) Table of the K10 TLS RNA binding affinities to various Egl truncations. n is the number of independent measurements. (E) SEC profiles of purified Egl512-819 (left), K10 TLS alone (middle) and Egl512-819 preincubated with a 1.2-fold excess of K10 TLS (right). The positions of the peaks are indicated by dotted lines in different shades of gray: the elution volume of Egl512-819 in isolation is in blue, while that of K10 TLS RNA is in black

Data collection and refinement statistics of the crystal structures of Egl512-819-K10 TLS complex and of MBP- (GSAM)- L-EHD and of K10-TR.

Structure of the Egl512-819-K10 TLS complex.

(A) Schematic representation of the Egl truncated constructs using for structural studies. (B) Crystal structure of Egl512-819 (with incomplete L and EHD regions) in complex with K10 TLS. (C) Crystal structure of L-EHD. (D) Electrostatic potential plotted on the surface of Egl512-819 with a gradient from red (negative) to blue (positive) from – to +5K T/e, respectively. (E) Model of Egl512-819 in complex with K10 TLS in three different views related by 90° and 180° rotation. The model is obtained by superposition of the structures in (B) and (C). The structures are shown as cartoon with Egl and K10 TLS colored in a gradient in blue and black, respectively.

Protein-RNA interactions within the Egl512-819-K10 TLS complex.

(A) Schematic representation of the architecture of Egl512-819-K10 TLS complex. Egl512-819 recognizes the K10 TLS with several contacts in the EXO, Linker and EHD. (B) Schematic representation of the interactions between Egl512-819 and K10 TLS. (C-H) The zoom-in panels show residues involved in the RNA binding. In (C-H), Egl512-819 and K10 TLS are shown as cartoon, and the figures in (A-G) share the same color-coding.

Mutagenesis analysis on the residues of Egl512-819 involved in RNA binding.

(A) Schematic representation of the Egl truncated constructs. (B) The residues of Egl involved in the binding of K10 TLS are highlighted in red and labeled on the protein surface rendering. On the right side, interacting residues are labelled. Underlined in blue are conserved residues mutated in the assays. (C) and (F) Kd values were determined by FA. 5′-FAM-labeled K10 TLS was incubated with increasing concentrations of Egl mutants and the data were fitted to the Hill equation to obtain the Kd. Data are represented as mean ± SD. (D) and (G) Column graph with mean Kd was shown as bars in the same color-coding, standard deviation as black error bars and individual Kd values as black dots. Weblogo representation of the residues involved in RNA binding was shown below the column graph. (E) and (H) Kd values of Egl mutants for K10 TLS are shown in the table ±SD.

RNA binding specificity of Egl512-819.

(A) and (D), Egl512-819 was preincubated with 5′-FAM-labeled K10 TLS RNA and unlabeled RNAs before measurement. The change in anisotropy was plotted and fitted to sigmoidal curve to obtain the IC50. (B) and (E), Column graph with mean IC50 was shown as bars, standard deviation as black error bars and individual IC50 values as black dots. (C) and (F), IC50 values of unlabeled RNAs are shown in the table. Data are represented as mean ± SD.

Egl residues required for RNA binding in vitro are required for function in vivo.

(A) Images of germaria and early-stage egg chambers for the indicated genotypes stained for the oocyte markers Orb (red; cytoplasmic marker) and Corolla (green; marker of the meiotic synaptonemal complex), as well as DNA (blue; DAPI). Arrowheads label oocytes in the control. In each of the mutant genotypes, all egg chambers fail to differentiate an oocyte (images are representative of at least 200 ovarioles visualized per genotype). PR29 and WU50 are previously characterized egl null alleles(Mach and Lehmann, 1997). (B) Left: stage 4/5 egg chambers of the indicated genotypes stained for Orb, Corolla and DNA, illustrating the partially penetrant oocyte differentiation defect in K659E/PR29 ovarioles. Right: quantification of oocyte differentiation phenotypes in the indicated genotypes (oocytes scored using Orb immunostaining; n = number of egg chambers examined). Control genotypes in the images in A and B are L-Triple/+ and PR29/+, which had phenotypes indistinguishable from the wild-type. Scale bars in A and B = 20 μm.

Alignment of Egl homologs and Egl RNA binding region, related to Figure 1.

(A) Sequences used to generate the alignment include: [Insecta, order Diptera] D. melanogaster (Dme), D. sechellia, D. simulans, D. erecta, D. yakuba, D. ananassae, D. persimilis, D. pseudoobscura pseudoobscura, D. willistoni, D. virilis, D. mojavensis, D. grimshawi, A. gambiae (Aga), C. quinquefasciatus; [Lepidoptera] B. mori (Bmo), H. melpomene, D. plexippus; [Coleoptera] T. castaneum (Tca); [Hymenoptera] A. cephalotes, A. mellifera (Ame); [Phthiraptera] P. humanus corporis; [Hemiptera] A. pisum (Api); [Crustacea] D. pulex (Dpu); [Chelicerata] I. scapularis (Ixo); [Nematoda] C. elegans (Cgi), C. briggsae, P. pacificus; [Plathyelmintes] S. mansoni (Dre); [Echinodermata] S. purpuratus. Numbering refers to the Dme sequence. The secondary structure of Dme Egl512-819 and L-EHD is schematized above the alignment (beta-sheets: ; alpha-helices: ; residues not ordered in the structure are shown as dotted lines; gaps indicate regions for which no structural information is available). Residues interacting with the 5′ or 3′ strand of the K10 stem-loop RNA are marked with a filled or with an empty dot above the Dme sequence, respectively; conserved residues are highlighted in blue according to 80% sequence identity in Jalview(Waterhouse et al., 2009). Catalytic residues in the exonuclease domain: ; mutated residues: ; conserved catalytic residues are highlighted in gray. The alignment was generated with MUSCLE(Edgar, 2004), visualized with Jalview (Waterhouse et al., 2009) and edited in Adobe Illustrator. (B) RNA binding properties of GST-tagged and untagged Egl512-819. Kd values were determined by FA. 5′-FAM-labeled K10 TLS was incubated with increasing concentrations of GST-Egl512-819 or Egl512-819 and the data were fitted to the Hill equation to obtain the Kd. Data are represented as mean ± SD. (C) Column graph with mean Kd was shown as bars in the same color-coding, standard deviation as black error bars and individual Kd values as black dots. (D) Kd values of GST-Egl512-819 or Egl512-819 for K10 TLS are shown in the table ±SD. RALS profiles of Egl512-819 unbound (E) and in complex with K10 TLS RNA (F). For each plot, the calculated molecular weight (MW) at the peak is indicated in black; the predicted molecular weight is shown in parenthesis. RALS and UV curves are colored in red and black, respectively.

Egl512-819 multimerization state in the complex and quality of the electron density, related to Figure 2.

(A) The crystal lattice of Egl512-819-K10 TLS includes two complexes in the ASU. In molecule A, Egl512-819 is rendered as cartoon in blue, while K10 RNA is in black. In molecule B, the protein is in light blue and the RNA in gray. A table displaying the extension of the interface between the two complexes in the ASU is shown below. (B) A zoomed in view of the interactions at the interface between Egl512-819 molecule A and Egl512-819 molecule B. The side chain of Lys682A interacts with the backbone carboxyl of Gln649 B. (C) Interactions between Egl512-819 molecule B and K10 TLS molecule A. K10 TLS RNAA backbone contacts His643B and Gln647B. With the exception of Lys726, all other interacting residues are evolutionarily conserved. (D) Stereo view of the electron density of the 2Fo-DFc maps for Egl512-819-K10 TLS complex structures after refinement.

Conformational changes of K10 TLS in the complex and in isolation, related to Figure 3.

(A) From left to right: crystal structures of K10 TLS (in red), of the Egl512-819 bound K10 TLS (in black), ideal A-form RNA K10 TLS lacking the two bulges C35 and A41 (Δbulges) (in gray) and NMR solution structure of K10 TLS (PDB ID.: 2KE6; Bullock et al. 2010; in gold) in a similar view. The width of major/minor grooves and RMSD are indicated. (B) Schematic representation of both Egl512-819 bound K10 TLS and NMR solution structure of K10 TLS. (C) Structural alignments of the Egl512-819 bound K10 TLS with K10 TLS (left), with an ideal A-form RNA K10 TLS (Δbulges) (middle) and with the NMR solution structure of K10 TLS. (D) K10 TLS structural parameters calculated with w3DNA (Zheng et al., 2009). Major groove widths: direct P-P distances, not accounting for Van der Waals radii of the phosphate groups. Graphical illustrations of the parameters are reproduced from http://x3dna.org (Lu and Olson, 2003).

Alignment of Egl-related EXO domains, interdomain interaction of Egl512-819 and K10 RNA binds at positively charged and conserved residues on the surface of Egl512-819, related to Figure 2 and 3.

(A) Sequence alignment of the EXO domains of D. melanogaster Egl (Egl_Dme), E. coli RNase D (RNaseD_Eco), T. brucei RRP6 (RRP6_Tbr) and B. mori EXO domain-containing 1 (Exd1_Bmo). Conserved residues are highlighted in red. Green boxes indicate exonuclease signature motifs, while signature catalytic residues are marked with . The alignment was generated with MUSCLE (Edgar, 2004), visualized with ESPript (Robert and Gouet, 2014) and edited in Adobe Illustrator. The catalytic sites are marked by light gray boxes. (B) Interdomain interaction of Egl512-819. (C-D) Surface representation of Egl in the same orientation as in (C) and rotated 90° along the x axis. In (C), the surface is colored according to electrostatic potential, with positively charged residues in blue. In (D), the surface is colored according to conservation, with a gradient from white to green indicating increasingly conserved residues. Conservation scores were calculated from the Egl sequences included in the multiple sequence alignment shown in Figure S1A, by using ConSurf (Ashkenazy et al., 2016).

Binding affinity of Egl alanine mutants to K10 TLS and Circular Dichroism spectra of Egl mutants used in this study, related to Figure 4.

(A) Circular Dichroism (CD) spectra of recombinantly purified Egl mutants used in FA. Wavelength scans are averages of five scans between 200 nm and 250 nm collected at 20°C in a quartz cuvette with a 1-mm path length with 0.2 mg/ml protein in buffer (20 mM Tris-HCl, pH 7.5, 300 mM NaCl, 10% glycerol, 1 mM DTT). Wavelength scans are shown as solid curves with Egl WT colored in black, the R753E/K754E/K755E triple mutant in dark gray, and the other mutants in light gray. (B) Column graph with mean Kd shown as bars, standard deviation as black error bars and individual Kd values as black dots. (C) Kd values between Egl alanine mutants and K10 TLS are shown in the table.

Binding affinity of Egl512-819 to RNA variants, related to Figure 5.

(A) and (B) Kd values were determined by FA. 5′-FAM-labeled RNA variants were incubated with increasing concentrations of Egl512-819 and the data were fitted to the Hill equation to obtain the Kd. Data are represented as mean ± SD. (A) Column graph with mean Kd was shown as bars, standard deviation as black error bars and individual Kd values as black dots. (B) Kd values between Egl512-819 and RNA variants are shown in the table. Data are represented as mean ± SD. (C) Secondary structures of RNA variants used for direct binding assay. (D) Secondary structures of RNA variants used in competition assay. Secondary structures of RNA variants were generated with mFold (Zuker 2003).

Additional characterisation of egl mutations in Drosophila, related to Figure 6.

(A) Female flies homozygous for the L-Triple mutation fail to produce mature oocytes (labelled in the control by an asterisk). In the control ovariole, developing oocytes within syncytial egg chambers are visualised by staining with an anti-Orb antibody (red). DNA is labelled with DAPI (cyan). In the absence of oocyte differentiation, mutant egg chambers eventually degenerate (arrowheads), which is consistent with previous observations in egl null mutants(Mach and Lehmann, 1997). Scale bar, 100 μm. Images are representative of > 200 ovarioles examined per genotype. Control genotype is PR29/+, which has indistinguishable oocyte development from the wild-type. (B) The mutations in egl that impair RNA binding do not prevent protein expression. Images are of germaria and early egg chambers. Note that Egl (green) accumulates in the oocyte in control egg chambers by microtubule-based transport(Mach and Lehmann, 1997). This accumulation is also seen in K659E/K659E egg chambers, which have no overt defects in oocyte differentiation (Fig. 6B). In contrast, homozygous L-Triple fail to differentiate an oocyte (Fig. 6A) and therefore do not accumulate Egl asymmetrically in the egg chamber. Thus, for this genotype, protein levels in the rest of the egg chamber should be compared to control. Control is wild type (w1118). Scale bar = 50 μm. Images are representative of > 75 ovarioles imaged per genotype.