The fungal RNA-binding protein Rrm4 mediates long-distance transport of ubi1 and rho3 mRNAs

Abstract

Cytoskeletal transport promotes polar growth in filamentous fungi. In Ustilago maydis, the RNA-binding protein Rrm4 shuttles along microtubules and is crucial for polarity in infectious filaments. Mutations in the RNA-binding domain cause loss of function. However, it was unclear which RNAs are bound and transported. Here, we applied in vivo RNA binding studies and live imaging to determine the molecular function of Rrm4. This new combination revealed that Rrm4 mediates microtubule-dependent transport of distinct mRNAs encoding, for example, the ubiquitin fusion protein Ubi1 and the small G protein Rho3. These transcripts accumulate in ribonucleoprotein particles (mRNPs) that move bidirectionally along microtubules and co-localise with Rrm4. Importantly, the 3' untranslated region of ubi1 containing a CA-rich binding site functions as zipcode during mRNA transport. Furthermore, motile mRNPs are not formed when the RNA-binding domain of Rrm4 is deleted, although the protein is still shuttling. Thus, Rrm4 constitutes an integral component of the transport machinery. We propose that microtubule-dependent mRNP trafficking is crucial for hyphal growth introducing U. maydis as attractive model for studying mRNA transport in higher eukaryotes.

Figure 1

Pab1 and Rrm4 co-localise in shuttling particles. (A) Growth of S. cerevisiae strain HKY171 carrying the mutation pab1-53^ts conferring reduced growth at 28°C. Proteins expressed are indicated on the left (black triangle, 10-fold serial dilutions). (B) Filament of AB33pab1G. Rectangles indicate regions that are magnified below. Frames are taken from Supplementary Video 1 (upper part). To document motility of particles, overlays of two frames that are 0.4 s apart and coloured in red and green are shown. Arrowheads indicate moving particles (upper and lower size bar, 10 and 2 μm, respectively). (C) Filament of AB33pab1R/rrm4G. Inverted frames are taken from Supplementary Video 2 recorded with dual-colour detection. Simultaneous Gfp and Rfp images are shown juxtaposed. Red and green arrowheads mark Pab1R- and Rrm4G-containing particles, respectively. Elapsed time in seconds is indicated (size bars, 10 and 2 μm, respectively). (D) Western blots after tandem affinity purification of Rrm4GT or GT. Input fraction (after mechanical destruction of filaments) and output fraction (after TEV protease cleavage and second affinity purification step) are shown on the left and right, respectively. In the upper and lower part, proteins (given on the right) were detected using α-Gfp and α-Myc antibodies (size marker in kD on the left; ^*, note that α-Myc incubation detects residual amounts of Rrm4GT because of the protein A domain). (E) Filament of AB33pab1G/rrm4Δ (labelling as in B; size bars, 10 and 2 μm, respectively; frames of Supplementary Video 1, lower part).

Figure 2

Rrm4 binds distinct mRNAs containing a CA-rich motif. (A) CLIP analysis of AB33rrm4GT filaments. Covalently bound RNA is detected as radioactively labelled protein-RNA complexes that are larger in size than Rrm4GT (117 kD after TEV cleavage; size markers on the left; asterisk, unspecific band). UV-C irradiation and RNase T treatment is indicated above the lanes. Equal loading was verified by western blot using α-Gfp antibodies on the same membrane (bottom). (B) Graphic representation of the CA-rich consensus sequence. The height of each column represents the level of conservation, the relative height of each letter depicts its frequency at the respective position (Crooks et al, 2004). (C) Graph indicating significant enrichment of Rrm4 target genes (black bars) within the respective functional categories (Ruepp et al, 2004) relative to the frequency within the genome (white bars). One or two asterisks symbolise P-value of enrichment <0.05 or <0.01, respectively (P-values were determined using hypergeometric distribution, Ruepp et al, 2004).

Figure 3

FISH reveals accumulation of rho3 mRNA in Rrm4-dependent particles. (A) Fixed filaments of AB33 and AB33P_otefrho3G (upper and lower part, respectively). Inverted pictures were obtained after FISH analysis with probes directed against rho3 as well as rho3 and gfp (upper and lower part, respectively). mRNA particles (filled arrowheads) and septa (asterisks) are highlighted (size bar, 10 μm). (B) Epifluorescence image of strain AB33P_otefrho3G (asterisk, septum; size bar, 10 μm). (C) Quantitative real-time PCR showing that the relative amount of rho3 and rho3G mRNA is not altered in rrm4Δ strains (the relative amount in AB33 [rrm4/rho3] was set to 1; error bars, s.d.). (D) PIA analysis of FISH-treated filaments of AB33 and AB33P_otefrho3G as well as respective rrm4Δ strains. Image example of AB33 analysed with rho3 probes (inverted image labelled as in A) followed by the corresponding fluorescence intensity graph. Exemplary graphs of filaments of strains AB33rrm4Δ as well as AB33P_otefrho3G and AB33P_otefrho3G/rrm4Δ are given below. Relative fluorescence signals were plotted against the longitudinal axis of filaments (relative distance, rear pole was set to 0). Detected peaks are indicated by filled arrowheads (see Materials and methods). (E) Bar diagram of mean particle numbers determined by PIA analysis. Particles were quantified in AB33 and AB33P_otefrho3G (upper and lower part, respectively) as well as respective rrm4Δ strains (as labelled below; error bars, s.e.m., n⩾3).

Figure 4

λN-Gfp RNA reporter system reveals microtubule-dependent transport of ubi1 and rho3 mRNAs. (A) Schematic representation of the constructs used for the adapted version of the λN-Gfp RNA reporter system. Top: the arabinose-regulated promoter P_crg1 drives expression of a fusion protein consisting of the λN peptide (amino acid 2–22) fused to eGfp in triplicate (bent arrow, transcriptional start); T_nos, heterologous transcriptional terminator of the nopaline synthase gene from Agrobacterium tumefaciens (Zarnack et al, 2006). Bottom: constructs carry 16 copies of the λN-binding site boxB inserted in the 3′ UTR (CA, CA-rich sequence; ubi1 3′, 311 nt fragment of the ubi1 3′ UTR; rho3 3′, 159 nt fragment of the rho3 3′ UTR; m, 40 nt fragment of the mfa2 3′ UTR; Res, resistance cassette; brackets, presence of multiple copies of the transgene in ubi1_B-containing strains). (B) Filament of SB1ubi1_B. Inverted frames are taken from Supplementary Video 3. Rectangles indicate regions that are magnified below. Arrow in upper panel indicates a static particle (top), open arrowheads show a stochastically moving particle, and filled arrowheads point towards directed particles (lower panels). Elapsed time is given in the left corner (upper and lower size bars, 10 and 2 μm, respectively). (C) Filaments of SB1 derivatives expressing different boxB-containing mRNAs. Inverted frames are taken from Supplementary Video 2 and 3 (ubi1_B, rho3_B* and mfa2_B, mfa2_ubiB; respectively). Movement of directed particles is tracked by red arrows (dark and bright colours are used in regions of overlap). (D) Graphs representing the average particle number and the average distance covered by particles. Treatment with inhibitors benomyl and CCCP as well as deletion of rrm4 was performed using strain SB1ubi1_B (⩾29 filaments analysed, n=3; error bars, s.e.m.; P<0.001, ANOVA test compared with ubi1_B). (E) Graphs with labelling as in D. For strains generated by ectopic insertion of the boxB-containing transgenes (rho3_B, mfa2_B, and mfa2_ubiB) the results of two independent transformants were combined (>60 filaments; n=6; error bars, s.e.m.; asterisk, P<0.005; by paired, two-tailed t-test).

Figure 5

Rrm4 and ubi1 mRNA are part of directed mRNPs in vivo. Filaments of SB1ubi1_B/rrm4R 20 h (left) and 12 h (right) after induction. Inverted frames are taken from time-lapse video recorded with excitation filters on a spinning wheel. Black and white arrowheads indicate Rrm4R- and ubi1_B mRNA-containing (λNG³-labelled) particles, respectively. Elapsed time in seconds is given on the right. Exposure times were 225 ms (left) and 300 ms (right). Switching excitation filters from red to green fluorescence was faster than the reciprocal switch because of the experimental setup (see Materials and methods; size bars, 10 μm).

Figure 6

Rrm4 is required for the formation of directed ubi1_B mRNA particles. (A) Filament of SB1ubi1_B/rrm4Δ. Inverted frames are taken from Supplementary Video 5. Rectangles indicate regions that are magnified below. Arrows show a static particle (left) and open arrowheads point towards a stochastically moving particle (right). No directed particles are detectable. Elapsed time is given on the right (upper and lower size bars, 10 and 2 μm, respectively). (B) Filament of SB1ubi1_B/rrm4^ΔRRMR. Frames are taken from Supplementary Video 6 (upper part, labelling as in A). Rrm4^ΔRRMR-containing particles that are moving bidirectionally along microtubules are indicted by white arrowheads (note that no directed ubi_B mRNA-containing particles are detectable, see Supplementary Video 6, lower part).