Synergic interplay of the La motif, RRM1 and the interdomain linker of LARP6 in the recognition of collagen mRNA expands the RNA binding repertoire of the La module

Abstract

The La-related proteins (LARPs) form a diverse group of RNA-binding proteins characterized by the possession of a composite RNA binding unit, the La module. The La module comprises two domains, the La motif (LaM) and the RRM1, which together recognize and bind to a wide array of RNA substrates. Structural information regarding the La module is at present restricted to the prototypic La protein, which acts as an RNA chaperone binding to 3' UUUOH sequences of nascent RNA polymerase III transcripts. In contrast, LARP6 is implicated in the regulation of collagen synthesis and interacts with a specific stem-loop within the 5' UTR of the collagen mRNA. Here, we present the structure of the LaM and RRM1 of human LARP6 uncovering in both cases considerable structural variation in comparison to the equivalent domains in La and revealing an unprecedented fold for the RRM1. A mutagenic study guided by the structures revealed that RNA recognition requires synergy between the LaM and RRM1 as well as the participation of the interdomain linker, probably in realizing tandem domain configurations and dynamics required for substrate selectivity. Our study highlights a considerable complexity and plasticity in the architecture of the La module within LARPs.

Figure 1.

The LARPs. (A) Domain organization of human La and LARPs showing the conserved La module formed by the LaM and RRM1. Other domains/motifs are labelled as follows: RNA recognition motif 2 (RRM2); nuclear retention element (NRE); nuclear localization signal (NLS); short basic motif (SBM), DM15 box domain (DM15); variant PABP-interacting motif 2 (PAM2w); LaM and S1-like proteins associated motif (LSA) (1,2,11). The indicated domains/motifs are not in scale. (B) Multiple sequence alignment of the La modules of human LARPs performed with ClustalW2. The secondary structure elements of HsLa as well as the boundaries of HsLa LaM and RRM1 are indicated. The six highly conserved residues of the LaM, which in human La are involved in oligoU RNA binding, are labelled with asterisks.

Figure 2.

Structures of HsLARP6 LaM and RRM1. (A and B) Superposition of the backbone traces of the 20 lowest-energy structures for (A) the LaM and (B) the RRM1. The N- and C-termini, and the secondary structures are annotated. (C and D) Representative structures of (C) HsLARP6 LaM and (D) RRM1. (E and F) Structure of (E) HsLa LaM (PDB 1S7A) and (F) RRM1 (PDB 1S79) (3) displayed in the same orientation as HsLARP6 LaM and RRM1, respectively. α-helices are coloured in dark red and β-strands in green. β-strands are numbered in panels C–E. The novel α-helices in HsLARP6 RRM1 are reported in orange. All structure representations were generated using PyMOL.

Figure 3.

Alignment of LARP6 La modules from different species. The La module sequence of HsLARP6 was aligned with 15 LARP6 proteins from 11 species, including vertebrates-eutherians (Equus caballus, Canis familiaris, Ailuropoda melanoleuca, Mus musculus, Rattus norvegicus), vertebrates (Gallus gallus, Danio rerio), invertebrates (Nematostella vectensis, Drosophila melanogaster), plants (Arabidopsis thaliana) and protists (Phytophtora sojae). Secondary structure elements for HsLARP6 LaM/RRM1 and their domain boundaries (this study) are reported above and below the sequences, respectively. The six conserved residues on the LaM are indicated with asterisks. Boxes numbered from I to V indicate regions of structural/sequence dissimilarity between HsLARP6 and HsLa (see text). Species codes are the following: Ps, Phytophtora sojae; At, Arabidopsis thaliana; Nv, Nematostella vectensis; Dm, Drosophila melanogaster; Dr, Danio rerio; Gg, Gallus gallus; Aim, Ailuropoda melanoleuca; Cf, Canis familiaris; Mm, Mus musculus; Rn, Rattus norvegicus; Ec, Equus caballus; Hs, Homo sapiens.

Figure 4.

The wing 2 configuration is different in HsLARP6 and HsLa. (A) Alignment of selected amino acids encompassing the LaM wing 2 and interdomain linker for HsLARP6 and HsLa. The secondary structure elements as well as structured/flexible regions are indicated. For HsLa, this refers to the apo protein (4). (B and C) The conformation of the wing 2 highlighted in cyan is shown for HsLARP6 and HsLa LaM, respectively, in three different orientations. (D and E) Close-up view of the residues involved in forming and stabilizing the wing 2 relative to the rest of the LaM, for (D) HsLARP6 and (E) HsLa (see text).

Figure 5.

Interaction of HsLARP6 with RNA targets. (A) Expected secondary structures of the 48 nt and 32 nt SL RNAs obtained from mfold. (B–H) ITC experiments showing the thermal effect of mixing 48 nt RNA with (B) HsLARP6 La module (70–300), (C) HsLARP6 LaM, (D) HsLARP6 RRM1, (E) triple mutant R244E/R245E/R249E, (F) Loop1 chimera, (G) Interlinker chimera and (H) RRM1 chimera (see text). For each interaction the raw data and the normalized binding curve are reported. Black squares indicate the normalized heat of interaction obtained per each injection, while the grey curve represents the best fit obtained by a non-linear least-squares procedures based on an independent binding sites model. When measurable, dissociation constants are reported.

Figure 6.

Sequence and structure divergence in the La modules from LARPs. (A) Structure of the La module of HsLa in complex with oligoU RNA (PDB 2VOD) (4). The LaM is coloured in yellow, the RRM1 in brown, the interdomain linker in green and the RNA is shown as sticks. (B) HsLARP6 LaM (in cyan - showing residues 85–178) and RRM1 (in magenta - residues 181–292) have been oriented with respect to one another analogously to the arrangement of the equivalent domains of HsLa in complex with oligoU (HsLARP6 LaM residues 90–119,125–169 were superposed to HsLa LaM residues 11–40,45–91; HsLARP6 RRM1 residues 185–189, 228–232, 258–263, 287–290 were superposed to HsLa RRM1 residues 112–116, 138–142, 153–158, 181–184). In this configuration, the distance from the last structured residue in the LaM to the first structured residue in the RRM1 in HsLARP6 is ∼17 Å, which will necessitate an inter-connecting linker in the order of 5–6 amino acids in a fully extended conformation. In absence of significant structural rearrangement of the individual domains upon RNA binding, the interdomain linker of HsLARP6 is only two residues long, making such tandem arrangement in HsLARP6 highly improbable. (C) Selected sequences encompassing the LaM wing 2 and the interdomain linker of HsLa were aligned with 25 LARP proteins from 6 different species including vertebrates-eutherians (Homo sapiens), vertebrates (Gallus gallus), invertebrates (Drosophila melanogaster), plants (Arabidopsis thaliana) and protists (Dictyostelium discoideum, Phytophtora sojae). Stretches experimentally demarcated as wing 2 and interdomain linker are indicated with a yellow and cyan box, respectively, revealing poor sequence alignment for these regions. Structural information was obtained from the following PDBs: HsLa (1S7A; 1S79; 2VON); TbLa (1S29); DdLa (2M5W); HsLARP6 (this study, 2MTF and 2MTG); HsLARP4 (MRC, unpublished results). Species codes are as for Figure 3 with the following addition: (Dd) dictyostelium discoideum.