Mechanistic Insights into RNA Chaperoning by Ro60 and La Autoantigens

SUMMARY

Although ATP-independent chaperones assist RNA folding, the mechanisms by which they function remain elusive. Here, we demonstrate how two RNA chaperones collaborate to unfold misfolded noncoding RNAs (ncRNAs). The ring-shaped Ro60 protein binds the ends of misfolded ncRNAs in its cavity, while La stabilizes nascent ncRNAs and assists their folding. Using cryo-electron microscopy to resolve the structure of a misfolded RNA complexed with Ro60 and La, we show that La cradles the Ro60 RNP, with its N-terminal domain binding the RNA 3’ end after it passes through the Ro60 cavity, while its C-terminal domain destabilizes structures in the misfolded RNA body. Using SHAPE-MaP, we show that La and Ro60 function synergistically to unfold non-native structures. As the RNAs bound by Ro60 and La include both ncRNA precursors and ncRNAs with oligouridine tails, this RNA chaperone machine may function widely to recognize misfolded and otherwise aberrant ncRNAs and assist their unfolding.

Graphical Abstract

IN BRIEF

This study elucidates the structural and molecular mechanisms by which two ATP-independent RNA chaperones, Ro60 and La, act synergistically to unfold misfolded RNAs. As numerous noncoding RNAs are associated with Ro60 and La, this RNA chaperone machine may function widely to resolve kinetic traps that impede correct folding.

INTRODUCTION

Many RNAs must fold into complex three-dimensional structures to carry out their functions. However, because RNAs contain only four nucleotide bases, they are susceptible to becoming kinetically trapped in non-native structures with stabilities comparable to the native forms^1,2. To prevent misfolding, some RNA binding proteins act as RNA chaperones, defined here as proteins that, through transient binding and release, facilitate the correct fate of bound RNAs in vivo³. RNA chaperones can protect newly synthesized RNAs from nucleases, favor formation of correctly folded RNAs, and/or assist annealing of two RNA strands^1,2,4,5.

Similar to protein chaperones, RNA chaperones are divided into ATP-dependent and ATP-independent classes. The best characterized ATP-dependent chaperones are members of the DEAD-box and DEAH-box protein families that unwind RNA and displace bound proteins using energy derived from ATP hydrolysis^6–8. These proteins assist numerous processes, including ribosome assembly, pre-mRNA splicing, and formation of RNP condensates^6–9. In contrast, the extent to which ATP-independent RNA chaperones contribute to RNA folding and function is less understood. In bacteria, two RNA chaperones, Hfq and ProQ/FinO, stabilize small regulatory RNAs and promote their base pairing with mRNA targets^10–12. Binding of Hfq to sRNAs and mRNAs can also alter or disrupt their secondary structure^13–18. In addition, cold shock proteins are proposed to destabilize RNA structures that form or become hyperstable at low temperatures¹⁹. Many RNA-binding proteins, when tested in assays designed to identify proteins that can disrupt RNA structures or assist correct folding of an RNA substrate, display RNA chaperone activity^1,20. However, the extent to which eukaryotic ATP-independent chaperones assist correct RNA folding and/or unfold misfolded RNAs as part of their normal functions, as well as the mechanisms by which they function, remains unexplored.

Two proteins that are proposed to function as ATP-independent RNA chaperones are the La and Ro60 proteins. Both proteins are clinically important targets of autoantibodies in patients with systemic autoimmune disease²¹. La, which is ubiquitous in eukaryotic cells, binds the UUU_OH that is the 3’ end of all newly synthesized RNA polymerase III transcripts, stabilizing the RNAs against exoribonucleases^22–24. La also binds noncoding RNA (ncRNA) precursors made by RNA polymerase II that end in UUU_OH^25–28. Because one or more of these uridylates are normally removed during maturation, La is usually not bound to the mature RNAs.

In addition to protecting the 3’ ends of ncRNA precursors from exoribonucleases, La assists RNA folding and RNP assembly. In budding yeast, binding of La to a pre-tRNA with a fragile anticodon stem prevents formation of an incorrect structure²⁹. In addition, binding by La to a precursor of the U4 spliceosomal snRNA that ends in UUU_OH enhances its assembly with core Sm proteins²⁶. In mammalian cells, some pre-tRNAs misfold to form hairpin structures when La is depleted, resulting in cleavage by Dicer and entry of pre-tRNA fragments into the miRNA pathway³⁰. All La proteins consist of a highly conserved N-terminal domain (NTD) containing a winged helix “La motif” and a canonical RNA recognition motif (RRM), which together form the “La module”, and a less-conserved C-terminal domain (CTD) (Figure 1A)^31,32. In most organisms, the CTD contains both a second RRM and a disordered C-terminus^31,32. Although the UUU_OH binds in a cleft between the La motif and the first RRM^33,34, both the mechanism(s) by which La assists RNA folding and the role(s) of the CTD in La function are unknown.

Figure 1. Architecture of a Ro60/La/misfolded RNP.

(A) Domain organization of La and Ro60.

(B) Secondary structures of X. laevis 5S⁸¹ (left), misfolded pre-5S⁴¹ (middle), and the predicted and observed secondary structures of the minimal misfolded pre-5S rRNA (right). The misfolded RNA contains mutations (red box, circles) that cause misfolding^36,41. Canonical stems are labeled with Roman numerals and loops with letters⁸². New stems and loops in misfolded RNAs are labeled I′, IV′, A′ and E′⁴¹. In the minimal RNA, nts 22–56 and 72–110 are replaced by loops (black boxes)⁴¹. The new base pair in the observed structure is red.

(C and D) EMSAs comparing binding of full-length (C) and minimal misfolded pre-5S (D) to Ro60, La, or both proteins. Lane 4, Ro60 was added before La.

(E) Cryo-EM density map of the Ro60/La/misfolded RNA/Fab complex at 2.7 Å resolution.

(F) Atomic model of the Ro60/La/misfolded RNA complex. The green sphere marks the Mg²⁺.

See also Figure S1 and Table S1.

In some animal cells, Ro60 is complexed with misfolded ncRNA precursors^35–37. This was first described in Xenopus laevis oocytes, where both Ro60 and La are complexed with a large class of variant 5S rRNAs that contain nucleotide changes that cause them to misfold^35,36. These RNAs are also longer at the 3’ end due to read-through of the first RNA polymerase III termination site³⁵. Structural studies revealed that Ro60 is a toroid formed by a series of HEAT repeats^38,39. The ring is closed by a von Willebrand factor A (VWA) domain, which are often sites of protein-protein interaction⁴⁰ (Figure 1A). Within the VWA domain is a divalent cation binding site called a metal-ion-dependent adhesion site (MIDAS) that can serve as a ligand-binding site⁴⁰. A crystal structure of Ro60 bound to a misfolded RNA fragment revealed that the single-stranded 3’ end of the RNA inserted through the Ro60 cavity, while an adjacent helix bound a large basic platform on the Ro60 surface⁴¹. Because the interactions were not strongly sequence-specific, it was proposed that Ro60 may scavenge ncRNAs that fail to bind their specific partners⁴¹. However, both the ways in which Ro60 affects the fates of the bound misfolded RNAs, and the role of La in the complex, remain unknown.

Here, we report that La and Ro60 collaborate to destabilize misfolded pre-5S rRNAs. Cryo-electron microscopy (cryo-EM) structures reveal that La cradles the Ro60-bound RNA, with the La NTD binding the RNA 3’ end as it emerges from the Ro60 cavity and the CTD interacting with the RNA body. Using the RNA structure probing method selective 2’-hydroxyl acylation analyzed by primer extension and mutational profiling (SHAPE-MaP)⁴², we show that La and Ro60 function synergistically to unfold non-native RNA elements. Our data support a model in which Ro60 carries out initial recognition of misfolded RNAs. Binding of Ro60 disrupts tertiary interactions and weak helices in bound RNAs, enabling the La CTD to bind and destabilize additional non-native structural elements in the RNA body. As diverse ncRNAs are bound by Ro60 and La in human cells, the roles of this RNA chaperone machine in recognizing misfolded RNAs and assisting their unfolding is likely widespread.

RESULTS

Hierarchical formation of a Ro60/La/misfolded pre-5S rRNA complex

To understand how Ro60 and La interact with misfolded ncRNAs, we assembled ternary complexes of the human proteins bound to their best characterized substrate, X. laevis misfolded pre-5S rRNA⁴¹. We examined binding to both the full-length RNA and a 64 nt “minimal misfolded RNA” that binds Ro60 with similar affinity as the full-length RNA⁴¹ (Figure 1B). After incubating Ro60 and La with RNA, RNPs were separated from unbound RNA in native gels.

Efficient ternary complex formation only occurred when Ro60 bound the RNA before La (Figures 1C and 1D, lane 4). When Ro60 and La were added together, several complexes formed with low efficiency (lane 5), consistent with multiple conformations and/or modes of interaction. Since structural studies revealed that the single-stranded 3’ end of a misfolded RNA fragment inserts through the Ro60 cavity⁴¹, our finding that Ro60 must bind first supports a model in which the 3’ end must pass through this cavity before binding La. If La were to bind first, the La-bound 3’ end would be sterically hindered from entering the Ro60 cavity.

La cradles the Ro60-bound misfolded RNA

We used cryo-EM to determine the architecture of the ternary complex. We examined ternary complexes containing the minimal misfolded pre-5S rRNA, since these complexes (Figure 1D) were more discrete than those with full-length RNA (Figure 1C). After purification (Figures S1A–S1B), initial experiments revealed that the flexibility of La posed a challenge to obtaining a high-resolution map. To reduce flexibility, we included a Fab fragment that binds the La motif (Figures S1C–S1D). With this approach, we solved the structure of the complex at 2.7 Å resolution (Figures 1E, 1F, S1E–S1G, and Table S1).

In the ternary complex, La cradles or clamps the Ro60 RNP, with the NTD and CTD of La each interacting with different parts of the RNA (Figures 1E and 1F). As in a structure of Ro60 complexed with a misfolded RNA fragment⁴¹, the 3’ tail of the RNA inserts through the Ro60 cavity (Figures 1E and 1F). The body of the RNA lays across the Ro60 outer surface. Stem I is positioned close to the Ro60 cavity, while stems I’ and II are splayed over the outer edge of the Ro60 surface (Figure 1F). As expected from our finding that Ro60 must bind first, the La NTD interacts with the 3’ end of the RNA that has passed though the Ro60 cavity. As in previous La NTD/RNA structures^33,34, the UUU_OH binds in a cleft between RRM1 and the La motif. A long helical linker separating the La NTD and CTD curves around Ro60, positioning La RRM2 such that it faces and contacts the misfolded RNA body. Notably, we did not detect protein-protein interactions between Ro60 and La. Instead, all interactions appear to be mediated by RNA.

The structure of the minimal misfolded RNA in the ternary complex differed from the secondary structure that was expected to form based on nuclease probing of the full-length misfolded RNA⁴¹. Specifically, most nucleotides predicted to form stem IV’ are unpaired in the cryo-EM structure, while G40 base pairs with C6 (Figure 1B).

Binding of Ro60 to misfolded RNA results in conformational changes

To distinguish conformational changes due to binding of Ro60 to misfolded pre-5S RNA from alterations due to La, we solved cryo-EM structures of free Ro60 and Ro60 complexed with the minimal misfolded pre-5S at 3.5 Å and 2.7 Å, respectively (Figure S2 and Table S1). Comparison of the apo- and RNA-bound structures revealed significant conformational changes, with an RMSD of 5.3 Å (Figure 2A). In apo-Ro60, the cavity is wider, with the edges of the H6 and H16 helices about 13.0 Å apart, while in the RNA-bound form, these helices are separated by about 9.2 Å. Upon RNA binding, α-helices and β-strands that bridge the VWA and HEAT repeat domains undergo rearrangement. Specifically, several helices near the N-terminus extend, while helices H16-H20 shift inward to narrow the cavity (Figures 2A and S3A). Helix H2, which is 8 amino acids in apo-Ro60, lengthens to 18 residues when bound to the RNA, while helix H6 increases from 13 to 18 amino acids (Figures 2A, S3A, and S3B). Residues T45-Y47, which form a β-strand in apo-Ro60, are part of a loop between helices H1 and H2 in the RNA-bound form, such that Y47 points toward the MIDAS (Figure 2A). Although densities corresponding to the N-terminal 23 amino acids were not observed in apo-Ro60, suggesting this region is unstructured, amino acids 16–18 and 21–24 form β-strands β1 and β2 in the RNA-bound form (Figure 2A). Together, these rearrangements reposition the VWA domain with respect to the HEAT repeat domain, resulting in an RNA-bound Ro60 with a narrower cavity.

Figure 2. Conformational changes on Ro60 binding to misfolded RNA.

(A) Apo-Ro60 (left) and RNA-bound Ro60 (middle), colored from blue at the N-terminus to red at the C-terminus. Overlay shows repositioning of the VWA domain (right). Due to insufficient resolution, the Mg²⁺ was not modeled in apo-Ro60. Green sphere, Mg²⁺ in RNA-bound Ro60.

(B) Molecular surface representations of Ro60 in the wide (left) and narrow (middle) conformations colored by electrostatic potential. RNA helices bind the basic surface (right).

(C) Interactions of minimal misfolded pre-5S rRNA with Ro60 and La in the ternary complex. Ro60 and La residues are purple and green, respectively. Solid lines indicate possible hydrogen bonds (< 3.5 Å). Dotted lines represent potential electrostatic interactions (< 7 Å). Residues involved in π-stacking are highlighted yellow. Of the 64 nts, 52 are modeled. For clarity, interactions with ribose rings are not shown. Dotted boxes, nts not modeled.

(D) Stems I’ (left) and II (right) in the Ro60 RNP showing partial base pair distortions.

(E) Interactions of G30 and U31 with Ro60. Dotted lines, possible hydrogen bonds.

See also Figures S2, S3 and Table S1.

These rearrangements also alter the distribution of positive charges on the Ro60 surface. The wide conformation contains two clusters of basic residues, one towards the outer edge and the other near the cavity. In the RNA-bound form, some residues, including K236, R255 and K340, cluster closer to the center, forming an extensive basic surface surrounding the cavity and extending to the basic patch on the Ro60 outer edge (Figure 2B, middle panels)⁴¹. Although K172 and R174 are not modeled in apo-Ro60 due to poor resolution, they are adjacent to the cavity in the RNA-bound conformation. Helical portions of the misfolded RNA contact this basic surface, with the tail inserting through the cavity. Basic residues on the outer edge, such as K108, R184, and R237, mostly interact with the RNA backbone through hydrogen bonding and electrostatic interactions (Figures 2B, 2C and S3C).

The human apo-Ro60 conformation resembles the crystal structure of a Ro60 ortholog from the bacterium Deinococcus radiodurans³⁹, while the RNA-bound form resembles crystal structures of X. laevis Ro60 bound to RNA ligands^38,41. Our findings reveal that these two distinct conformations are not due to species differences but are instead driven by interactions of RNA with the Ro60 cavity. In the RNA-bound form, the cavity has a higher positive charge density (Figure 2B), and the smaller cavity allows closer contacts with single-stranded RNA. Notably, the Ro60 mutant K170A/R174A was found to have an ~ 7-fold decrease in binding affinity for misfolded pre-5S³⁸, confirming the importance of cavity residues for binding these RNAs.

Ro60 binding also alters the conformation of the misfolded RNA. Unlike typical A-form RNA with planar base pairs, many of the base pairs in stems I’ and II show partial distortion (Figure 2D). Although Ro60 interacts with both sides of the stem I helix, with most interactions to the RNA backbone, Ro60 largely interacts with the 3’ side of the RNA in stem II and loop A’, with some interactions involving hydrogen bonding and stacking with nucleotide bases (Figure 2C and S3C). These interactions include hydrogen bonding between the bases of G30 and U31 and the R149 sidechain, while their phosphate groups form hydrogen bonds with the backbones of G142, M143, and W144 (Figures 2E and S3D). These interactions partially perturb base pairing and pull the 3’ strand of stem II toward Ro60, resulting in distortions to base pairs. Moreover, since stems I and II bind the Ro60 surface, the RNA may be constrained to adopt conformations that contribute to base pair distortion.

Conformational changes on La binding to misfolded RNA

Superposition of the Ro60/misfolded pre-5S rRNA structure with the ternary complex structure revealed that La also captures an altered conformation of the misfolded RNA. Stem I’ shifts further toward the Ro60 outer edge, positioning the bulge between stems I and I’ near RRM2 (Figure 3A). This shift results in closer contact of nts C6 and C7 with Ro60, allowing their backbones to form additional hydrogen bonds with K172 and R237, respectively (compare Figures 2C and S3C). As in the presence of Ro60, the RNA in the ternary complex shows partial base-pair distortion, with stem II showing slightly greater distortion in the presence of La (Figure 3B).

Figure 3. Conformational changes on La binding.

(A) Superposition of the Ro60-bound RNA in the absence (tan) and presence of La (orange).

(B) Stems I’ and II in the ternary complex, showing partial base-pair distortion (left). Overlay of stem II reveals increased distortion with La (right).

(C) La RRM2 interacts with single-stranded nts between stems I’ and I while α3 inserts through the bulge (left). G39 and U41 flip toward α1 and β2, respectively (right).

(D) Conservation analysis of RRM2⁸³. Zoom panels show residues interacting with RNA.

(E) Interactions of G39 and U41 with α1, β2, and β2-β3 loop residues (left) and A5, C6 and U38 with α3 residues (right). Dotted lines, possible hydrogen bonds.

(F) Overlay of La RRM2 in the ternary complex with the NMR structure PDB 1OWX.

See also Figure S3.

Most alterations in RNA structure appear mediated by La RRM2, a noncanonical RRM containing an extra helix (α3) that partly occludes the β-sheet, the part of canonical RRMs typically bound by single-stranded RNA⁴³. In the ternary complex, RRM2 interacts with the misfolded RNA body, with the α3 helix inserting through the bulge at the helix I/I’ junction (Figure 3C). Conserved residues within α1, β2, and the α1-β2 and β2-β3 loops mainly interact with RNA through stacking of side chains or hydrogen bonding to the protein backbone (Figures 2C, 3D and 3E). The bases of G39 and U41 are flipped out toward RRM2 and interact with conserved residues. Specifically, the R246 side chain is sandwiched between these bases with R266 stacking on G39. In addition, N3 of U41 forms a hydrogen bond to the I262 backbone and its O4 forms a hydrogen bond to the F264 backbone, while O6 of G39 forms a hydrogen bond to the R246 backbone (Figures 3E and S3E, left).

Although the ends of the RRM2 α3 helix are less conserved, it contains multiple positively charged amino acids (Figure 3D). Some of these residues interact closely with the RNA. The R334 side chain is sandwiched between A5 and C6 and forms a hydrogen bond to the C6 phosphate, while the R335 side chain forms a hydrogen bond to the U38 phosphate (Figures 3E and S3E, right). K328 and K332 are within 7 Å of the U38, G39, U41, and U42 phosphates, consistent with electrostatic interactions (Figure 2C).

Comparison of our full-length La structure to structures of the isolated NTD and RRM2 revealed that La also undergoes conformational changes. While the interactions of the NTD with the RNA 3’ end resemble crystal structures of La bound to short RNA oligomers or partial duplexes^33,34, with RMSDs between 0.9 and 1.2 Å (Figure S3F), RRM2 undergoes moderate structural changes upon RNA binding, with an RMSD of 2.8 Å (Figure 3F). In the ternary complex, the α3 helix contains 11 additional amino acids compared to an NMR structure of the isolated apo-RRM2⁴³ (Figure 3F, left), suggesting helix lengthening occurs on RNA binding. Moreover, the conserved loop between β2-β3, which contains F264 and R266, is pulled toward the flipped-out bases (Figure 3F, right).

We also compared our structure to those of similar RRMs complexed with RNA. La is the founding member of the La-related protein (LARP) superfamily, all of which contain an N-terminal La motif, usually adjacent to an RRM^31,32. LARP7 subfamily members also contain a domain that resembles RRM2, called xRRM⁴⁴. In mammals LARP7 is required for the assembly and stability of the 7SK RNP, a transcriptional regulator⁴⁵. In ciliates and fungi, LARP7 orthologs are crucial for the assembly and function of the telomerase RNP^46–48. In telomerase and 7SK RNPs, the bases of single-stranded nts point towards xRRM β2, resembling interactions between La RRM2 and misfolded RNA, and their α3 helices extend on RNA binding^49,50. Notably, the interaction of U41 with I262 and R246 in our structure (Figure 3E) resembles that of 7SK G314 with V484 and R468 in the LARP7 xRRM (Figures S3G). However, the α3 helices of these xRRMs bind across the RNA major groove, which for telomerase bends the RNA to promote RNP assembly^49,50. In contrast, the La RRM2 α3 helix threads through a bulge at the two-helix junction, which could serve to destabilize the adjacent helices.

Ro60 enhances binding of the La CTD to RNA

Our finding that the La CTD interacts with sequences in the misfolded RNA suggested that it might bind the RNA independently of the La NTD. Moreover, although all La CTDs end in a disordered region^43,51 (Figure S3H), we did not observe this density in our cryo-EM map, most likely due to its flexibility. To determine if the isolated CTD binds RNA, and whether the disordered region is important, we performed electrophoretic mobility shift assays (EMSAs). We compared the full CTD (La 222–408) with RRM2 (La 222–336; ending at the last residue in the cryo-EM structure) and La 335–408 (Figure 4A), which was shown using NMR to be disordered⁴³. Since the La CTD can interact with the misfolded RNA in the form of a Ro60 RNP, we also determined if Ro60 affects CTD binding.

Figure 4. Ro60 enhances binding of the La CTD to RNA.

(A) Schematics showing full-length and truncated La proteins.

(B and C) Binding of the La CTD and truncations to the minimal misfolded RNA. In (C), RNA was pre-bound to Ro60.

(D) To allow Ro60 to bind wild-type 5S RNA, 9 gene-encoded nts (red) were added. Only the central portion of the RNA is shown. For the full RNA, see Figure 5A.

(E and F) EMSAs comparing binding of the CTD to Y1 and full-length wild-type and misfolded pre-5S. In (F), RNAs were pre-bound to Ro60.

(G-N) EMSAs using ³²P-labeled minimal misfolded RNA and the indicated concentrations of La (G), La NTD (H), La CTD (I-J), and CTD truncations (K-N). In J, L and N, RNA was pre-bound to Ro60. Dissociation constants (± s.d.) are below gels.

(O) Binding curves of the minimal misfolded RNA with the indicated proteins, determined using MST. Error bars, s.d. (n=3).

(P) Binding curves of minimal misfolded RNA with the indicated proteins. Normalized fluorescence (ΔF_norm) is plotted against protein concentration. Error bars, s.d. (n=3).

In the absence of Ro60, the full CTD formed a discrete complex with the minimal misfolded pre-5S rRNA. This complex was not detected with either RRM2 alone or the disordered region (Figure 4B). Together with findings that the isolated NTD binds UUU_OH-containing RNAs^43,52,53, we conclude that La contains two independent RNA binding domains.

Interestingly, when the RNA was presented as a Ro60 RNP, the ternary complex formed at lower concentrations of the CTD and appeared more discrete in native gels (Figure 4C, lanes 2–4). Moreover, although RRM2 alone did not bind the Ro60 RNP, the disordered region was sufficient to form a ternary complex (lanes 8–10). Our finding that the disordered region complexes with the Ro60-bound RNA, but not with the RNA alone, supports a model in which Ro60 alters the RNA structure such that the disordered region can bind.

We also determined whether the isolated CTD discriminates correctly folded from misfolded full-length 5S rRNAs. Because Ro60 does not bind wild-type mature 5S rRNA, but binds when the 3’ end is extended to resemble that of the misfolded pre-5S RNA⁴¹, we used a 3’ extended RNA in these studies (Figure 4D). When the CTD was incubated with wild-type pre-5S rRNA, complex formation was not detected (Figure 4E). The CTD also failed to bind Y1 RNA, a member of a class of ncRNAs that bind Ro60 and tether it to effector proteins such as exoribonucleases^54,55. However, the CTD did bind the full-length misfolded pre-5S RNA, indicating that some feature of this RNA allows binding (Figure 4E, lanes 5–8). Notably, when the RNAs were presented as Ro60 RNPs, the CTD formed ternary complexes with both wild-type and misfolded pre-5S rRNAs (Figure 4F). Thus, Ro60 may alter the structure of the wild-type RNA, allowing the La CTD to bind.

We determined relative binding affinities by performing quantitative EMSAs. As expected for an RNA ending in UUU_OH⁵³, both full-length La and the isolated NTD bound the minimal misfolded RNA with low nM affinity [apparent dissociation constants (K_d) of 19 ± 3 nM and 8 ± 3 nM, respectively; Figures 4G and 4H]. The affinity of the CTD for this RNA was 36 to 90-fold lower (K_d = 720 ± 70 nM; Figure 4I), consistent with studies showing the CTD does not contribute significantly to the RNA binding affinity of La^43,53. When the same RNA was present as a Ro60 RNP, the band corresponding to the ternary complex was more discrete and the affinity of the CTD for this RNP increased (K_d = 480 ± 60 nM; Figure 4J). No binding was detected with RRM2 alone (Figures 4K and 4L). Although we were unable to measure a dissociation constant for the disordered region, discrete complexes only formed when the RNA was in the form of a Ro60 RNP (Figures 4M and 4N).

Since EMSAs require that complexes withstand gel fractionation⁵⁶, we also measured binding using microscale thermophoresis (MST). With MST, the affinities of full-length La, the NTD, and the CTD to the minimal misfolded RNA were similar to those measured by EMSAs (K_ds of 17 ± 1 nM; 27 ± 6 nM and 710 ± 70 nM, respectively; Figure 4O). However, Ro60 enhanced the affinity of the CTD for the RNP by nearly four-fold (K_d = 180 ± 70 nM; Figure 4O), compared to the 1.5-fold measured by EMSA. As with EMSAs, we did not detect measurable binding of RRM2 or the disordered region (Figure 4P). We conclude that Ro60 enhances binding of the CTD to misfolded RNA.

Ro60 and La function synergistically to destabilize misfolded RNAs

To determine how Ro60 and La alter RNA structure, we performed SHAPE-MaP⁴². This technique probes the accessibility of RNA 2′ hydroxyls to chemical modification, with single-stranded nucleotides more accessible than those involved in base pairing or tertiary interactions. Probing of the wild-type and misfolded pre-5S rRNAs revealed higher reactivity in regions predicted to be single-stranded (Figure 5), consistent with the proposed structures. We compared the reactivity of the protein-free RNAs to their Ro60-, La- and Ro60/La-bound forms, in each case calculating the difference in SHAPE reactivity. Positive difference values indicate increased flexibility, likely due to disruption of base pairs or tertiary interactions.

Figure 5. Ro60 and the La CTD disrupt weak elements.

(A) Nts in wild-type pre-5S rRNA that show increased reactivity in SHAPE-MaP on binding (left to right) Ro60, La, Ro60 followed by La, and Ro60 followed by the CTD, are colored. Box, tertiary interactions in the 5S rRNA three-way junction (PDB 7OYC). Arrowheads, nts with >3-fold increased reactivity. Asterisk, nt with >10-fold increase.

(B and (C) Nts in full-length (B) and minimal misfolded pre-5S rRNAs (C) that show increased reactivity upon binding (left to right) Ro60, La, Ro60 followed by La, and Ro60 followed by the CTD are colored. Arrowheads, nts with >3-fold increases.

See also Figure S4 and Table S2.

Upon binding of either Ro60 or La to the full-length wild-type pre-5S rRNA, the largest increases in reactivity occurred in or near the three-way junction (Figures 5A and S4A, Table S2). With Ro60, SHAPE reactivity increased strongly (difference >1) for A13, C14, C15, C68, U109, G110 and G113, and was enhanced at A16, C63, G66, C67, and G107. The most affected nucleotides were U109 and G110, with increases of 10.1 and 3.9, respectively (Figures 5A, S4A, and Table S2). Notably, in cryo-EM structures of the ribosome, G66 forms tertiary interactions with A11, A13 and U109 through hydrogen bonding (Figure 5A, box)⁵⁷. Our results indicate that Ro60 disrupts at least some of these tertiary interactions. Upon La binding, changes in reactivity also centered around the three-way junction, but tended to affect different nucleotides than those affected by Ro60. SHAPE reactivity increased strongly at A11, C15, C68 and G113, and was enhanced at C9, C10, C63 and U109. Thus, although Ro60 and La each increase accessibility of nucleotides near the three-way junction, they differ in the extent to which they increase the reactivity of specific nucleotides.

When Ro60 binding was followed by binding of either full-length La or the CTD, the reactivity of some nucleotides increased, compared to Ro60 or La alone. These included nucleotides at or near the three-way junction, such as C14 and G108, and nucleotides in loops B and C. Nucleotides showing strongly enhanced reactivity in the presence of Ro60 and the CTD were largely restricted to the three-way junction, while nucleotides in loops B and C also increased strongly in reactivity when Ro60 was followed by full-length La.

For the full-length misfolded pre-5S rRNA, most increases in reactivity upon Ro60 or La binding were in helices and loops that differ from those in the wild-type RNA (Figures 5B and S4B, Table S2). On Ro60 binding, A13, U15 (both loop A’), U69, A74, A83, U84, and U109 (stems IV’ and V) became more accessible to acylation, consistent with at least transient disruption of misfolded helices. With La, nucleotides in these regions were also more accessible (U69, A74, C79, U80, G82-G85, A88) (Figure 5B). Notably, U69 corresponds to U38 in the minimal misfolded RNA, which is also strongly reactive on La binding (Figure 5C). Consistent with a role for La in altering RNA structure, this nucleotide was predicted to be base paired but is single-stranded in our structure (Figure 1B). In the presence of Ro60 and La, or Ro60 and the La CTD, some nucleotides in misfolded regions (A13, U15, G68, A113) showed increased accessibility compared to when only one protein was bound. The CTD had a stronger effect on nucleotides at the ends of helices in the three-way junction than full-length La (Figures 5B and S4B), suggesting that binding of the RNA 3’ end to the NTD allows the CTD to associate less strongly with the RNA body than when the CTD is the only RNA-binding domain. Results for the minimal misfolded RNA were similar, although nucleotides adjacent to the UUU_OH also became accessible on La binding (Figure 5C). Together, our data support a model in which Ro60 and La function synergistically to destabilize three-way junctions and misfolded helices.

Many ncRNAs ending in oligouridine are bound by Ro60 and La in human cells

The only RNAs currently described to be present in Ro60/La RNPs are misfolded pre-5S rRNAs³⁵ and Y RNAs⁵⁸. To identify additional RNAs, we purified Ro60/La RNPs from RO60^−/− human HaCaT cells that stably express Ro60 fused to the Twin-Strep-tag⁵⁹ (TS-Ro60; Figures S5A and S5B). After selecting TS-Ro60 RNPs on Strep-Tactin beads and eluting with biotin, La RNPs were immunoprecipitated with anti-La antibodies or control IgG. Following cDNA sequencing, nearly 90% of reads derived from the four human Y RNAs (Figure S5C), which are major constituents of these RNPs⁵⁸. Of the remaining RNAs, the most abundant (at least 200 reads) and enriched (at least 8-fold compared to the control library) were mostly ncRNAs transcribed by RNA polymerase III, such as 5S and tRNAs (Figure S5D). Consistent with findings that Ro60 binding to misfolded RNAs requires at least 5 single-stranded 3’ nucleotides⁴¹ and that La binds RNAs ending in UUU_OH²², many 5S RNAs contained additional 3’ nucleotides due to read-through of one or more termination sites and some tRNAs were precursors (Figures S5E and S5F).

We also recovered ncRNAs containing non-templated oligouridine tails. These tails, which are added by terminal uridylyltransferases⁶⁰, promote degradation by two exoribonucleases, DIS3L2^61,62 and ERI1⁶³. Tailed RNAs included wild-type 5S rRNAs, 3’ fragments of 12S mitochondrial rRNA (mt-rRNA), and precursors to microRNA let-7 (pre-let-7i) (Figures S5G–S5J). Mt-rRNAs are made as polycistronic transcripts and cleaved from flanking tRNAs by endonucleases⁶⁴. All 12S mt-rRNA fragments in our cDNA were truncated or extended relative to the mature 3′ end, with oligouridine tails often following non-templated adenylates (Figure S5I). As adenylation and uridylation occur when mt-rRNA decay factors are inactivated^65,66, these RNAs could represent decay intermediates. For pre-let-7i, 3’ oligouridylation prevents DICER cleavage and targets the RNA for DIS3L2 degradation⁶⁷.

Since La protects nascent ncRNAs from some exoribonucleases^24–26, we tested whether La protects La/Ro60-bound RNAs from DIS3L2 and ERI1. Although both enzymes were able to degrade a U-tailed form of the X. laevis minimal misfolded RNA, binding by either La alone or La and Ro60 prevented degradation (Fig. S5K–S5M). One possibility is that La-mediated protection, combined with La/Ro60-assisted unfolding of U-tailed misfolded RNAs, gives these RNAs additional chances to refold.

A structure of a Ro60/La/human misfolded pre-5S RNP reveals La flexibilities

Since the 5S rRNAs in human Ro60/La RNPs were wild-type, rather than variants, with many containing extra 3’ nucleotides, we confirmed that these RNAs formed Ro60/La RNPs in vitro. We tested two 5S rRNAs, one containing ten 3’ uridylates (“U-tailed 5S rRNA”), and a second containing a genome-encoded 3’ extension (“pre-5S rRNA”) (Figure S6A). Although both RNAs formed Ro60/La RNPs, the major ternary complex containing the U-tailed 5S rRNA formed irrespective of the order of Ro60 and La addition (Figure S6B). With the pre-5S rRNA, both the Ro60 RNP and the ternary complex migrated as two bands in native gels (Figure S6B), suggesting the RNA bound in more than one conformation. Formation of only one ternary complex band required that Ro60 bind before La (Figure S6B, lane 14, asterisk). Consistent with containing misfolded RNA, this complex migrated similarly in native gels to the ternary complex formed with the X. laevis misfolded pre-5S RNA (Figure S6B), and its levels increased when magnesium was omitted or the initial Ro60 and La binding occurred at room temperature, rather than 4°C (Figure S6C).

To determine the architectures of these complexes, we undertook structural studies. For the U-tailed 5S rRNPs, attempts to characterize the ternary complex were unsuccessful because La dissociated during purification. However, a cryo-EM map of the Ro60/RNA complex at 3.5 Å resolution revealed that, as expected, the RNA did not enter the cavity (Figures S6D and S7, Table S1). Consistent with our model that binding of RNA in the Ro60 cavity results in conformational changes, the structure resembled that of apo-Ro60 (Figure S6D).

For the pre-5S rRNPs, we focused on the ternary complex that required initial Ro60 binding. Because the RNA dissociated during purification, we mutated the RNA to force it to adopt a secondary structure similar to that of the X. laevis misfolded RNA (Figure S6A). With this approach, we obtained a single ternary complex, formation of which required sequential Ro60 and La binding (Figure 6A). However, due to the flexibility of the complex, we were unable to obtain clear densities corresponding to either the La NTD or CTD.

Figure 6. Structure of a Ro60/La/human misfolded RNP reveals additional flexibilities.

(A) EMSAs comparing binding of full-length wild-type and misfolded human pre-5S rRNAs to the indicated proteins. Lanes 4 and 9, Ro60 was added before La. Asterisk, ternary complex requiring initial Ro60 binding.

(B) Truncated misfolded pre-5S rRNA used for cryo-EM (left). Nts 22–56 and 81–105 were replaced by loops (black boxes). The RNA contains mutations (red box and circle) that cause it to misfold. Right, nts with increased reactivity in SHAPE-MaP.

(C) EMSAs comparing binding of the truncated RNA with the indicated proteins. Lane 4, Ro60 was added before La.

(D) Cryo-EM density map of Ro60/La/RNA/Fab complex at 3.1 Å resolution.

(E) Close up of the cryo-EM map (left) and the atomic model (right) showing interaction of the La CTD with the Ro60-bound RNA.

(F) Secondary structure of the RNA in the atomic model. Nts in gray were not modeled.

(G) Superposition of Ro60 and RNA in the two ternary complexes.

(H) Superposition of the helical linkers. Planes were defined using residues 8–182 (La module) and 230–308 (RRM2). Angles between planes were calculated with ChimeraX⁸⁴.

See also Figures S5–S8 and Tables S1–S2.

Notably, using a truncated form of this misfolded RNA, we obtained a stable ternary complex and solved its structure at 3.1 Å resolution (Figures 6B–6E, S8, and Table S1). The RNA differs from the X. laevis RNA in containing an extended stem IV’ and a 16 nt tail. As expected, the architectures of the ternary complexes were similar (Figures 6D and 6E). Although we did not observe a clear density for the 3’ tail that has passed through the Ro60 cavity, most likely due to its increased flexibility, we observed densities corresponding to three uridines in the cleft of the La module, indicating the RNA 3’ end is bound similarly (Figure 6D). As in the complex with the X. laevis RNA, G40 and C6 of the human RNA form a base pair (Figure 6F), while G39 and U41 flip toward α1 and β2 of the La RRM2, respectively (Figure 6E).

Importantly, due to the larger RNA, we were able to visualize the interaction of the RRM2 α3 helix with the three-way junction. Although we could not fully resolve the structural model of stem IV’, the cryo-EM density map clearly shows the α3 helix inserting through the remodeled junction (Figures 6D–6F, box). We consider it likely that this interaction contributes to the destabilization of three-way junctions that we detected with SHAPE-MaP.

Superimposing the structures of the two ternary complexes revealed striking flexibilities in La binding. In the structure with human RNA, the helical linker between the La NTD and CTD moves ~23° to reposition the NTD, allowing it to bind the longer 3’ extension (Figure 6G). Superimposing the La helical linkers in the two ternary complexes also revealed significant differences in the angles of the NTD and CTD (Figure 6H). Thus, as with the structural flexibility observed for Ro60, the flexibility of La may allow it to accommodate diverse RNAs.

DISCUSSION

Although it has long been proposed that RNA-binding proteins act as ATP-independent chaperones to destabilize misfolded RNAs^1,68, identifying proteins that carry out this role in cells and elucidating their mechanisms has been difficult. Here, we demonstrate how Ro60 collaborates with La to destabilize weak structural elements in misfolded RNA. Both proteins exhibit striking conformational flexibility, which likely enables them to interact with diverse RNAs. As the RNAs bound by Ro60 and La include multiple ncRNA precursors, La- and Ro60-mediated unfolding may function widely to resolve kinetic traps and promote correct folding of structured ncRNAs.

In our model, Ro60 carries out initial recognition of misfolded RNAs (Figure 7A). This is supported by data showing that Ro60 binds structured RNAs containing protein-free single-stranded 3’ ends⁴¹, by our finding that ternary complex formation is more efficient when Ro60 binds the RNA before La, by our cryo-EM structure demonstrating that the 3’ end of the misfolded RNA inserts through the Ro60 cavity to engage the La module, and by our demonstration that the La CTD only binds stably to wild-type pre-5S rRNA when the RNA is prebound to Ro60. Some RNAs may be 3’ extended precursors, while others may undergo uridylation before Ro60 binding. Although RNAs can also bind Ro60 and La in other configurations, some of which do not require prior Ro60 binding, we do not know the extent to which these configurations favor RNA unfolding.

Figure 7. Model for Ro60 and La function.

(A) Ro60 carries out initial recognition of misfolded RNA.

(B) Ro60 binding destabilizes tertiary interactions and weak helices.

(C) Ro60 and La collaborate to unfold the RNA.

(D) Some RNAs may refold, allowing specific RNA-binding proteins to displace Ro60 and La.

(E) Modifications to the 3’ end may eliminate La binding, resulting in degradation.

On binding, Ro60 disrupts tertiary interactions and increases the accessibility of nucleotides in weakly base paired helices (Figure 7B). Ro60 may destabilize these elements by capturing transiently exposed RNAs as they become available during normal fluctuations. Such a role would resemble that of the DEAD box family of ATP-dependent RNA chaperones, which disrupt tertiary interactions and enhance unwinding of short helices through ATP-independent binding to transiently exposed stretches of RNA⁶⁹. Bases at helix termini may also become more accessible if steric constraints imposed by Ro60 binding cause increased fraying of the ends. As proposed for an ATP-dependent RNA chaperone⁷⁰, if misfolded elements are less likely than correctly base paired structures to form stable tertiary interactions with other parts of the RNA, these elements may be preferentially disrupted. Notably, on Ro60 binding, the 3’ end that has passed through the cavity will be accessible to terminal uridylyltransferases, which could generate a binding site for the La NTD.

Upon binding of the La NTD to the RNA 3’ end, the CTD is positioned such that it faces the misfolded RNA body (Figure 7C). In this configuration, it binds exposed single-stranded RNA, using portions of the RRM, the extended α3 helix and the disordered region. Insertion of the RRM α3 helix through multi-helix junctions may contribute to destabilizing and remodeling these junctions, which are critical architectural elements that influence further RNA folding⁷¹.

Our finding that the disordered region is required for La CTD binding, and that the isolated region binds Ro60-bound RNA, suggests that, as demonstrated for some DEAD-box helicases^72,73, one role of this region is to tether the CTD to RNA. As the isolated disordered region preferentially binds structured RNAs presented as Ro60 RNPs, stable binding may require single-stranded RNA. Moreover, since the disordered region of S. cerevisiae La is required for correct folding of a pre-tRNA with a fragile anticodon stem⁵¹, this region may contribute to destabilizing weak helices. This region could also contribute to folding by shielding charges and preventing other molecules from accessing the RNA⁷⁴. Consistent with shielding newly made ncRNAs from nucleases, some ncRNA precursors require the disordered region for stable accumulation⁵¹.

Although we do not know the full fates of the RNAs bound by Ro60 and La, we favor a model in which RNAs can either refold or be targeted for degradation. As binding of La to fragile pre-tRNAs enhances their folding^29,30, some RNAs may refold while bound to Ro60 and La. Formation of the correct fold would favor binding by specific RNA-binding proteins that displace Ro60 and La (Figure 7D). Other Ro60/La-bound RNAs may eventually undergo degradation by exoribonucleases (Figure 7E). In this case, La would first need to be displaced by other RNA-binding proteins and/or enzymes that modify the 3’ end. The extent to which specific RNAs undergo folding versus degradation will likely differ for each target, and similar to protein chaperones⁷⁵, may involve kinetic partitioning between downstream events such as binding of partner proteins to correctly folded RNAs and elimination of the La binding site by nucleases or modification.

Our studies also give insights as to how Ro60 and La function as chaperones when bound individually to RNA targets. In D. radiodurans, Ro60 is tethered by Y RNA to a ring-shaped 3’ to 5’ exoribonuclease, forming a double-ringed RNP machine specialized for structured RNA decay⁷⁶. Although threading of the RNA 3’ end through the Ro60 ring may assist ATP-independent unwinding⁷⁶, our finding that Ro60 can disrupt tertiary interactions and weak helices makes it likely that Ro60 also assists degradation via these mechanisms. For La, iterative binding by the CTD may disrupt incorrect helices, giving them additional opportunities to form correct structures. Although S. cerevisiae and most other fungal La proteins lack an RRM2, the sole RRM in these species is predicted to contain an α3 helix⁷⁷, which may function similarly to the RRM2 α3 helix. Moreover, human La binds some internal ribosome entry sites to stimulate translation of specific mRNAs^32,78. For those IRESs that require RRM2 for La recognition⁷⁹, RRM2-mediated remodeling of the IRES may assist ribosome binding.

Limitations of the study

Although our study reveals how Ro60 and La unfold misfolded RNA, we do not know the extent to which the fates of specific RNAs depend on these proteins, or the extent to which other factors, such as helicases, specific protein partners, and nucleases contribute. Moreover, as a Y RNA regulates access of a bacterial Ro60 to RNA substrates⁸⁰ and tethers this Ro60 to an exoribonuclease to alter its substrate specificity⁷⁶, Y RNAs could contribute to the fate of Ro60/La-bound RNAs in animal cells. Future experiments in which we determine how Ro60 and La affect the fates of their targets in vivo, and identify the other components involved, will be required to address these questions. This information may allow the development of assays that recapitulate the downstream effects of Ro60 and La on their targets in vitro.

RESOURCE AVAILABILITY

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Sandra L. Wolin (sandra.wolin@nih.gov)

Materials availability

All unique reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

Cryo-EM maps and models have been deposited to the EMDB and PDB, respectively, and will be publicly available at the time of publication. For the ternary complex containing X. laevis minimal misfolded RNA, accession numbers are EMD-49357 (consensus map), EMD-49358 (N-terminal La focused map), EMD-49359 (C-terminal La focused map) and EMD-49360, PDB 9NFA (composite map and model). Other accession numbers are EMD-49328 and PDB 9NEN (apo-Ro60); EMD-49329 and PDB 9NEP (Ro60/minimal misfolded RNP); EMD-49327 (Ro60/U-tailed 5S RNP); EMD-49353 (ternary complex containing truncated misfolded human pre-5S RNA, consensus map); EMD-49354 (ternary complex containing truncated misfolded human pre-5S RNA, focused map) and EMD-49355 and PDB 9NF8 (ternary complex containing truncated misfolded human pre-5S RNA, composite map and model).

Raw sequencing data and processed files were deposited at NCBI GEO under accession number GSE290478.

STAR Methods

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Insect cells

Spodoptera frugiperda Sf21 cells were a gift of Eugene Valkov (National Cancer Institute, National Institutes of Health). They were maintained in SF900II media (ThermoFisher Scientific 10902104) at 27°C.

Bacterial cells

E. coli DH10EMBacY cells, a gift of Eugene Valkov (National Cancer Institute, National Institutes of Health), were used for bacmid production to produce recombinant baculovirus. E. coli BL21(DE3)pLysS cells (Promega, L1195) were used to express full-length and truncated forms of human La protein. All strains were grown at 37°C in Luria-Bertani broth (LB; 10 g/L tryptone, 10 g/L NaCl, 5 g/L yeast extract). Where appropriate, kanamycin was used at 50 μg/ml and ampicillin was used at 100 μg/ml.

Human cell lines

HaCaT keratinocyte-derived cells⁸⁵ (Cytion 300493) were maintained in DMEM (Gibco 11965118), 10% fetal bovine serum (FBS) (Gemini Bio-Products 100-106-500), 2 mM L-Glutamine (Gibco 25030081) and 100 U/ml penicillin-streptomycin (Gibco 15140122) at 37°C. To generate RO60^−/− HaCaT cells, guide RNAs targeting RO60 (1F: 5’-CACCGAAACTTACAAGACAGGTCA-3’ and 1R: 5’-AAACTGACCTGTCTTGTAAGTTTC-3’; 2F: 5’-CACCGCTAATTCAGTACTTGAACC-3’ and 2R: 5’-AAACGGTTCAAGTACTGAATTAGC-3’) were cloned into pX459 [pSpCas9(BB)-2A-Puro (PX459) V2.0 plasmid⁸⁶, a gift from Feng Zhang (Addgene plasmid 62988)]. After transfection with Lipofectamine 3000 (ThermoFisher L3000015), cells were selected with 1 μg/ml puromycin (Sigma Aldrich P8833) for 48 hours and clonal cell lines established by screening for loss of Ro60 protein. To generate RO60^−/−;TS-Ro60 cells, RO60^−/− cells were transfected with PiggyBac vector⁸⁷ expressing Ro60 with a N-terminal Twin-Strep tag⁵⁹ and (GGGGS)₃ linker. After selecting with 5 μg/ml blasticidin S (InvivoGen ant-bl-05) for one week, cells were transferred into 96 well plates and clones screened for Twin-Strep-Ro60 expression by western blotting. A clone that expressed Twin-Strep-Ro60 at near wild-type levels was found to contain wildtype levels of Y RNAs, consistent with rescue by the tagged protein (Figure S6A and B).

METHOD DETAILS

Protein expression and purification

To express human Ro60, cDNA encoding the protein sequence (Accession ID J04137) was codon-optimized for insect cells and synthesized as gBlocks (Integrated DNA Technologies). The cDNA was inserted into the EcoRI/HindIII sites of pLIB⁸⁸ (a gift from Jan-Michael Peters, Addgene 80610), and integrated into baculovirus by transformation into DH10EMBacY cells (a gift from E. Valkov, National Cancer Institute, NIH). Bacmid DNA was isolated and transfected into Sf21 insect cells (a gift from E. Valkov, National Cancer Institute, NIH) to produce baculovirus as described⁸⁹. For protein expression, Sf21 cells were infected with recombinant virus at a cell density of 1.0 × 10⁶ cells/ml. Cells were split after 48 hr and harvested at 96 hr. Cells were resuspended in 25 mM HEPES pH 8.0, 100 mM NaCl, 3 mM MgCl₂, 10% glycerol, 2 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP), and protease inhibitor cocktail (Millipore Sigma 11836170001) and lysed by sonication. Ro60 purification was as described³⁸, except that HEPES pH 8.0 and TCEP were used instead of Tris and DTT, respectively. Briefly, the cleared lysate was applied to a HiTrap Heparin HP column (Cytiva 17040601) and eluted with a linear salt gradient in 20 mM HEPES pH 8.0, 1 M NaCl, 10% glycerol, 2 mM TCEP. Fractions containing Ro60 were loaded onto a HiLoad 16/600 Superdex 200 pg gel filtration column (Cytiva 28989335). Ro60 was eluted in 20 mM HEPES pH 8.0, 200 mM NaCl, 10% glycerol, 2 mM TCEP, concentrated to 1 mg/ml, and stored at −80°C.

To express full-length and truncated forms of human La, the cDNA⁹⁰ was amplified with oligonucleotides La_5p and La_3p (full-length La); La_5p and La194_3p (La 1–194); La CTD_5p and La CTD_3p (La 222–408); La tail_5p and La CTD_3p (La 335–408). The cDNAs were then inserted into the NcoI/XhoI sites of pET-28a(+) (Millipore Sigma 69864), resulting in a 6X His tag at the C-terminus. For La 222–336, aa 337–408 were deleted from La 222–408 using KLD enzyme mix (New England Biolabs M0554S) and oligonucleotides “La336_5p” and “La336_3p” according to the manufacturer’s instructions. The proteins were expressed in E. coli BL21(DE3)pLysS cells (Promega L1195) and induced with 400 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for 3 hr at 37°C. For purification, cells were resuspended in buffer A (50 mM HEPES pH 8.0, 100 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 10% glycerol, 1 mM DTT) with 1 mg/ml lysozyme and protease inhibitor cocktail and lysed by sonication. Cell lysates were cleared by spinning 50 min at 284,600 × g and applied to His Pur Ni-NTA resin (Thermo Scientific 88222) pre-equilibrated with buffer A. After washing the resin with buffer A containing 25 mM imidazole, the protein was eluted with buffer A containing 250 mM imidazole. The eluate was applied to a HiTrap Heparin HP column (Cytiva 17040601) and eluted with a linear salt gradient in buffer B (50 mM HEPES pH 8.0, 1 M NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 10% glycerol, 1 mM DTT). Fractions containing La were pooled and the buffer exchanged for buffer C (25 mM HEPES pH 8.0, 100 mM NaCl, 3 mM MgCl₂, 0.1 mM EDTA, 10% glycerol, 1 mM DTT) using a PD-10 desalting column (Cytiva 17085101), concentrated to between 1 and 4 mg/ml, and stored at −80°C.

To purify the La CTD and CTD truncations, lysates were applied to the Ni-NTA resin as above. After washing the resin with buffer A containing 400 mM NaCl and 25 mM imidazole, the protein was eluted with buffer A containing 250 mM imidazole. To purify La 222–408, the Ni-NTA eluate was applied to a HiTrap Heparin HP column (Cytiva 17040601) and eluted with a linear salt gradient in buffer B. Fractions containing La 222–408 were pooled and the buffer exchanged for buffer C using a PD-10 desalting column (Cytiva 17085101). For La 222–336 and La 335–408, the eluate from the Ni-NTA column was applied to a Superdex 200 Increase 10/300 GL gel filtration column (Cytiva 28990944) and the protein eluted in buffer C. Proteins were concentrated to 1 mg/ml and stored at −80°C.

To express human ERI1, cDNA encoding the protein sequence (Accession ID BC035279.1) containing a C-terminal 6X His tag was codon-optimized for insect cell expression and synthesized as gBlocks (Integrated DNA Technologies). After inserting the cDNA into the SalI/PstI sites of pLIB and integrating into baculovirus by transformation into DH10EMBacY cells, bacmid DNA was isolated and transfected into Sf21 insect cells to produce baculovirus as described⁸⁹. To produce protein, Sf21 cells were infected with recombinant virus at a cell density of 1.0 × 10⁶ cells/ml and harvested at 72 hr. Cells were resuspended in buffer D (20 mM Tris pH 8.0, 50 mM NaCl, 1 mM PMSF) and protease inhibitor cocktail and lysed by sonication. After applying the cleared lysate to His Pur Ni-NTA resin pre-equilibrated in buffer D and washing the resin with 20 mM Tris pH 8.0, 300 mM NaCl, 40 mM imidazole, the protein was eluted with 20 mM Tris pH 8.0, 300 mM NaCl, and 250 mM imidazole. After loading the eluate onto a Superdex 200 Increase 10/300 GL gel filtration column, the protein eluted in 20 mM Tris pH 8.0, 500 mM NaCl, 5 mM DTT, and 10% glycerol. The protein was concentrated to ~ 0.3 mg/ml and stored at −80 °C.

RNA preparation

For in vitro transcription, sequences corresponding to wild-type X. laevis pre-5S rRNA³⁵, the mutant misfolded (stem mt 1 + 2) pre-5S rRNA³⁶, the 64 nt minimal misfolded pre-5S rRNA⁴¹, the minimal misfolded pre-5S rRNA with 15 additional Us at the 3’ end, the human pre-5S RNA, the human U-tailed 5S rRNA, and the human misfolded pre-5S rRNA were cloned into the HindIII/XhoI sites of p2RZ (Addgene 27644) containing an upstream hammerhead ribozyme and a downstream HDV ribozyme⁹¹ under control of a T7 promoter. DNAs consisting of the T7 promoter, the hammerhead ribozyme, the 5S rRNAs and the HDV ribozyme were amplified using oligonucleotides WT5S_5p and WT5S_3p (wild-type pre-5S); 5Smin_5p and 5Smin_3p (full-length misfolded pre-5S, minimal misfolded pre-5S rRNA, human pre-5S rRNA and U-tailed 5S rRNA); h5Smut_5p and h5Smut_3p (human misfolded pre-5S rRNA) and 5Smin_5p and WT5S_3p (minimal misfolded pre-5S rRNA with 15 Us at the 3’ end). Reactions contained 40 mM Tris-HCl pH 8.0, 10 mM DTT, 0.1 mM spermidine, 0.01% Triton X-100, 10 mM rNTPs, 10–14 mM MgCl₂, 10 μg/ml amplified DNA, 0.1 mg/ml T7 RNA polymerase⁹² (made in house) and 0.5 U/ml inorganic pyrophosphatase (New England Biolabs M0361L) and were incubated at 37°C for 3 to 4 h. Afterwards, the RNAs were purified from denaturing gels and treated with T4 polynucleotide kinase (PNK; New England Biolabs M0201L) to produce RNAs ending in 5’ monophosphate and 3’ hydroxyl. PNK reactions contained 12 μM RNA, 1 mM ATP, 200 U/ml T4 PNK and 500 U/ml RNase inhibitor (Promega Corporation N515) in the buffer supplied with the T4 PNK enzyme and were incubated for 2 hrs at 37°C. After phenol:chloroform:isoamyl alcohol (50:49:1) extraction, the RNA was exchanged into water using PD-10 columns (Cytiva 17085101) and precipitated with ethanol. For the human Y1 and truncated human misfolded pre-5S rRNAs, DNAs were amplified using oligos “hY1–5p”, “hY1–3p”, “sub1mt_tr_5p” and “sub1mt_tr _3p”, respectively. The resulting DNAs were transcribed with T7 RNA polymerase and the RNAs purified as described above. Oligonucleotide sequences are listed in Table S3.

For EMSAs and some SHAPE-MaP experiments, the X. laevis full-length misfolded pre-5S rRNA sequence was amplified using oligonucleotides 5Smt_5p and 5Smt_3p (Table S3) and transcribed with T7 RNA polymerase. Afterwards, RNA was purified on denaturing gels, extracted with phenol:chloroform:isoamyl alcohol (50:49:1) and precipitated with ethanol.

Assembly of complexes for cryo-EM

To assemble the Ro60/La/minimal misfolded pre-5S rRNA complex, 35 nmol of RNA was resuspended in buffer E (20 mM HEPES pH 8.0, 125 mM NaCl, 1.5 mM MgCl₂, 1 mM EDTA, 0.05% Tween 20, 5% glycerol) and refolded by heating to 95 °C for 2 min, freezing on dry ice for 3 min and thawing on ice for 20 min. Next, 40 nmol of Ro60 was added and incubated for 20 min at room temperature and 20 min at 4°C, followed by addition of 35 nmol La for the same incubation times and temperatures. The final concentration of Ro60, La and RNA were 8 μM, 7 μM and 7 μM, respectively. The complex was concentrated using a 100 kDa molecular weight cutoff (MWCO) Pierce Protein Concentrator PES (VWR International P188503) and injected onto a Superdex 200 Increase 10/300 GL gel filtration column (Cytiva 28990944). The column flow rate was 0.4 ml/min and the complex was eluted in buffer E without Tween 20 and glycerol. Fractions containing the complex were pooled and the concentration estimated by quantitating the amount of La using SDS-PAGE with protein standards. For ~ 25 μM complex, Fab [Creative Biolabs HPAB-0883-FY-F(E)] was added to 0.2 mg/ml and incubated for 20 min at room temperature and 20 min at 4°C. After concentrating using a 100 kDa MWCO concentrator, the final concentration of the complex was ~ 25 – 50 μM.

To assemble the Ro60/minimal misfolded pre-5S rRNP, 25 nmol of RNA was resuspended in buffer E and refolded as above. Afterwards, the RNA was incubated with 25 nmol of Ro60 for 20 min at room temperature and 20 min at 4°C. The final concentration of Ro60 and RNA were each 5 μM. The complex was concentrated using a 10 kDa MWCO concentrator (Millipore Sigma UFC901024) and injected onto a Superdex 200 Increase 10/300 GL gel filtration column. The column flow rate was 0.4 ml/min and the complex was eluted in buffer E without Tween 20 and glycerol. Fractions containing the complex were pooled and concentrated to ~ 25 – 50 μM using a 10 kDa MWCO concentrator.

To prepare free Ro60, 18 nmol Ro60 was loaded onto a Superdex 200 Increase 10/300 GL gel filtration column. The flow rate was 0.4 ml/min, and the complex was eluted in 20 mM HEPES pH 8.0, 125 mM NaCl, 1.5 mM MgCl₂, 1 mM EDTA, and 2 mM TCEP. Fractions containing Ro60 were pooled and concentrated to ~ 20 – 60 μM using a 10 kDa MWCO concentrator.

To assemble the Ro60/U-tailed human 5S rRNA complex, 14.5 nmol RNA was resuspended in buffer F (20 mM HEPES pH 8.0, 125 mM NaCl, 1 mM EDTA, 0.05% Tween 20, 5% glycerol) and incubated on ice for 30 min. Afterwards, 20 nmol of Ro60 was added and incubated for 30 min at 4°C and 10 min at room temperature. The final concentration of Ro60 and RNA were 4 μM and 3 μM, respectively. The complex was concentrated using a 100 kDa MWCO concentrator and injected onto a Superdex 200 Increase 10/300 GL gel filtration column. The flow rate was 0.4 ml/min and the complex was eluted in buffer F without Tween 20 and glycerol. Fractions containing the complex were pooled and concentrated to ~ 35 – 50 μM using a 100 kDa MWCO concentrator.

To assemble the Ro60/La/human truncated misfolded pre-5S rRNA complex, 10 nmol of RNA was resuspended in buffer F and incubated on ice for 30 min. Afterwards, 11 nmol of Ro60 was added and incubated for 50 min at room temperature, followed by addition of 7 nmol of La for the same incubation time and temperature. The final concentration of Ro60, La and RNA were 1.1 μM, 1 μM and 0.7 μM, respectively. The assembled complex was concentrated using a 50 kDa MWCO concentrator (Millipore Sigma UFC205024) and injected onto a Superdex 200 Increase 10/300 GL gel filtration column. The column flow rate was 0.4 ml/min and the complex was eluted in buffer F without Tween 20 and glycerol. Fractions containing the complex were pooled and the molar concentration estimated as above. For ~ 7 μM of complex, the Fab [HPAB-0883-FY-F(E), Creative Biolabs] was added to a final concentration of 0.14 mg/ml and incubated for 20 min at room temperature and 20 min at 4°C. After concentrating with a 50 kDa MWCO concentrator, the final concentration was between 15 and 35 μM.

Cryo-EM sample preparation and imaging

The Ro60/La/minimal misfolded pre-5S rRNA/Fab, free Ro60, Ro60/minimal misfolded pre-5S rRNA, Ro60/U-tailed human 5S rRNA and Ro60/La/truncated misfolded human pre-5S rRNA/Fab samples used for data collection were approximately 25, 20, 45, 35 and 35 μM, respectively, and adsorbed onto glow-discharged holey carbon-coated grids (Quantifoil Carbon R 1.2/1.3 on Gold 300, N1-C14nAu30-01) for 10 s. Grids were blotted for 3 s at 10°C, 100% humidity and frozen in liquid ethane using a Vitrobot Mark IV (Thermo Fisher Scientific).

Movies were collected in counted mode, in Electron Event Representation (EER) format, on a CFEG-equipped Titan Krios G4 (Thermo Fisher Scientific) operating at 300 kV with a Selectris X imaging filter (Thermo Fisher Scientific) and slit width of 10 eV, at 165,000x magnification on a Falcon 4 or Falcon 4i direct detection camera (Thermo Fisher Scientific), corresponding to a calibrated pixel size of 0.693 Å (Ro60/La/minimal misfolded pre-5S rRNA/Fab; free Ro60; Ro60/minimal misfolded pre-5S rRNA) or 0.732 Å (Ro60/U-tailed 5S rRNA; Ro60/La/human truncated misfolded pre-5S rRNA/Fab). Movies were collected at a total dose of 49.8 to 60.3 e⁻/Ų (Table S1), fractionated to ~ 1.0 e⁻/Ų per fraction for motion correction.

Cryo-EM data processing

Patched motion correction, CTF parameter estimation, particle picking, extraction, and initial 2D classification were performed in SIMPLE 3.0⁹³. All downstream processing was carried out in cryoSPARC⁹⁴ or RELION 3.1⁹⁵, using the csparc2star.py script within UCSF pyem⁹⁶ to convert between formats. Global resolution was estimated from gold-standard Fourier shell correlations (FSCs) using the 0.143 criterion and local resolution estimation was calculated within cryoSPARC.

The cryo-EM processing workflow for Ro60/La/minimal misfolded pre-5S rRNA/Fab is outlined in Figure S1E–G. Briefly, particles were subjected to two rounds of reference-free 2D classification (k=200 each) using a 160 Å soft circular mask within cryoSPARC. Selected particles (586,912) were subjected to multi-class ab initio reconstruction using a maximum resolution cutoff of 7 Å, generating four volumes. Particles from the most populated and structured class containing Fab were selected and non-uniform refined against their corresponding volume lowpass-filtered to 30 Å, generating a 2.9 Å map. Bayesian polishing in RELION further improved map quality to 2.7 Å after non-uniform refinement. Particle subclassification was subsequently performed on this consensus refinement to further improve density for the N- and C-terminal domains of La. To improve N-terminal domain density of La, heterogeneous refinement was performed against four identical copies of the consensus refinement map lowpass-filtered to 20 Å. Particles (96,901) corresponding to a class with strong density for the Fab and N-terminal domain of La were selected and non-uniform refined against their matching volume, lowpass-filtered to 30 Å, to generate a 2.8 Å volume. To improve C-terminal domain density of La, a four-class alignment-free 3D classification (RELION) was performed on particles from the same 2.7 Å consensus refinement and volume, using a soft mask around the C-terminal domain of La. Particles (95,850) exhibiting strongest density for this region were selected and non-uniform refined against the consensus volume, lowpass-filtered to 30 Å, yielding a 2.9 Å map. A soft mask was generated and used to subtract Fab and N-terminal domain of La density from the particles. Local refinement using a soft mask focused on the C-terminal domain of La on a subtracted volume lowpass-filtered to 6 Å yielded an improved 3.0 Å volume. Map sharpening for the consensus and subclassified volumes was performed using the “highres” model of deepEMhancer⁹⁷. Sharpened volumes for the improved N- and C-terminal La domain volumes were re-fit to the consensus density using the fitmap function and a composite map of Ro60/La/minimal misfolded pre-5S rRNA/Fab was generated from all three sharpened volumes in ChimeraX⁸⁴.

The cryo-EM processing workflow for Ro60 is outlined in Figures S2A–S2C. Particles were subjected to one round of reference-free 2D classification (k=200) using a 130 Å soft circular mask within cryoSPARC. A 387,456 particle subset of 2D-cleaned particles was used to generate four volumes ab initio (maximum resolution cutoff of 8 Å). The structured volume derived from the most populated class was lowpass filtered to 8 Å and used as reference volume in a non-uniform refinement against the full 2D-cleaned particle set (450,220 particles), generating a 3.7 Å reconstruction. Bayesian polishing followed by a round of 2D classification (k=100, 130 Å mask) yielded a pruned subset of 194,142 particles. Multi-class ab initio reconstructions (maximum resolution cutoff of 6 Å) were performed on this particle set. Particles belonging to the most populated class were subsequently refined (non-uniform) against their corresponding reconstruction lowpass filtered to 15 Å, generating a 3.5 Å map which was sharpened in deepEMhancer using the “tighttarget” model.

The cryo-EM processing workflow for Ro60/minimal misfolded pre-5S rRNA is outlined in Figures S2D–S2F. Particles were subjected to one round of reference-free 2D classification (k=200) using a 100 Å soft circular mask within cryoSPARC. A 264,508 particle subset of 2D-cleaned particles was used to generate four volumes ab initio (maximum resolution cutoff of 7 Å). These volumes were lowpass filtered to 20 Å and used as references for heterogeneous refinement against all 314,355 2D-cleaned particles. Particles (190,106) from the most populated class were selected and non-uniform refined against the corresponding volume, lowpass filtered to 30 Å, generating a 2.8 Å volume. Bayesian polishing followed by an additional round of non-uniform refinement (using a 30 Å lowpass-filtered reference) yielded a 2.7 Å map which was sharpened in deepEMhancer using the “highres” model.

The cryo-EM processing workflow for Ro60/U-tailed human 5S rRNA is outlined in Figure S7. Particles were subjected to two rounds of reference-free 2D classification (k=200) using a 140 Å soft circular mask within cryoSPARC. 2D-cleaned particles (1,519,659) were then heterogeneously refined against three ab initio volumes previously generated from the Ro60/minimal misfolded pre-5S rRNA dataset, lowpass-filtered to 8 Å. Particles (549,555) from the class that yielded strongest Ro density were selected and subjected to a 3-class ab initio reconstruction (maximum resolution cutoff of 8 Å) to further clean the dataset. Particles (256,141) were selected from the most populated class that also demonstrated strongest Ro density and non-uniform refined against their corresponding volume (lowpass-filtered to 30 Å), generating a 3.8 Å volume. Heterogeneous refinement of these particles against Ro60 and Ro60/U-tailed human 5S rRNA volume references (lowpass-filtered to 20 Å) further purified the RNA-bound form, resulting in a 3.7 Å volume after non-uniform refinement from 191,412 particles. Bayesian polishing followed by another round of non-uniform refinement (using a 30 Å lowpass-filtered reference) yielded a 3.5 Å map which was sharpened in deepEMhancer using the “highres” model.

The cryo-EM processing workflow for Ro60/La/human truncated misfolded pre-5S rRNA/Fab is outlined in Figure S8. Particles were subjected to two rounds of reference-free 2D classification (k=200 each) using a 160 Å soft circular mask within cryoSPARC. A 403,109 particle subset of 2D-cleaned particles was used to generate four volumes ab initio (maximum resolution cutoff of 7 Å). These volumes were lowpass filtered to 8 Å and used as references for heterogeneous refinement against all 2,046,703 2D-cleaned particles. Particles (651,415) from the most populated and structured class containing clear secondary structure elements were selected and subjected to another round of heterogeneous refinement against 8 Å lowpass-filtered volumes generated from the previous heterogeneous refinement. Particles (414,104) from the most populated class were selected and non-uniform refined against the corresponding volume, lowpass filtered to 15 Å, generating a 3.2 Å volume. Bayesian polishing followed by a round of 2D classification (k=100, 160 Å mask) generated a cleaner 333,294 particle subset. Local CTF refinement followed by non-uniform refinement yielded a 3.1 Å map. Particle subtraction of the Fab fragment and N-terminal domain of La density was subsequently performed, resulting in a 3.0 Å reconstruction with improved density for the C-terminal domain of La after non-uniform refinement. Alignment-free 3D classification of the particles (k=8) using a soft mask focused on the C-terminal domain of La produced one class (45,104 particles) with improved connectivity and structural features. Particles belonging to this class were selected and non-uniform refined against their corresponding volume, lowpass-filtered to 15 Å. This resulted in the generation of a focused 3.3 Å volume with improved interpretability of the C-terminal domain of La. Map sharpening for the consensus and subclassified focused volume was performed using the “tighttarget” model of deepEMhancer. The sharpened volume for the subclassified focused volume was re-fit to the sharpened consensus density using the fitmap function and a composite map of Ro60/La/human truncated misfolded pre-5S rRNA/Fab was generated from both volumes in ChimeraX.

Model building and refinement

Atomic models were generated first by docking published or AlphaFold⁷⁷ models into their respective cryo-EM volumes, followed by manual rebuilding and real-space refinement in Coot v.0.9.8.8^98,98. Both models were further refined in real-space using PHENIX⁹⁹ with rotamer, Ramachandran, and secondary structure restraints against deepEMhancer post-processed maps, yielding the models described in Table S1. All models were validated using MolProbity¹⁰⁰ within PHENIX. C_a traces of atomic models were superposed and RMSD calculated using least square fit (LSQ) superpose in Coot⁹⁸. For La RRM2 (Figure 3F), residues 230–327 were selected for superposition. Figures were prepared using UCSF ChimeraX v.1.6.1⁸⁴.

Conservation analyses of the La CTD

ConSurf⁸³ was used to select 300 sequences that sample the list of La homologs and calculate conservation scores of La RRM2 (Figure 3D). The alignment of La disordered C-termini (Figure S3H) from H. sapiens (NP_001281074.1), X. laevis (P28048.1), C. elegans (CCD67305.1), D. melanogaster (AAA21776.1), A. thaliana (AEE86109.1), and S. cerevisiae (NP_010232.1) was performed with T-Coffee¹⁰¹.

Electrophoretic mobility shift assays

X. laevis full-length and minimal misfolded pre-5S rRNAs were refolded by heating to 95°C for 2 min, freezing on dry ice for 3 min and thawing on ice for 20 min in buffer E. Because the wildtype X. laevis 5S rRNA and human Y1 RNA were used without refolding, these RNAs were incubated on ice for 20 – 30 min in buffer E. Ro60 was added and the reaction incubated for 30 min at 4°C and 10 min at room temperature, followed by addition of La protein for the same incubation times and temperatures. For the experiments shown in Figures 4B–4F, the final RNA concentration was 1 μM and the La concentrations were 0.5, 1 and 2 μM. In Figures 4C and 4F, the final Ro60 concentration was 1.2 μM. Afterwards, reactions were fractionated in 4 or 5 % polyacrylamide (37.5:1 acrylamide:bisacrylamide)/5 % glycerol gels (pre-run at 18 V/cm for 10 min at 4°C). Gels were run at 4°C for 10 min at 3 V/cm, followed by 18 V/cm in 0.5 X TBE for 80 min, and stained with GelRed (Biotium).

For K_d measurements, 1.5 fmol of 5’ end-labeled minimal misfolded pre-5S rRNA was refolded in buffer E as described above. After refolding, 2 μg poly(deoxyinosinic-deoxycytidylic) acid (MilliporeSigma), and 1 μg bovine serum albumin (Thermo Fisher Scientific), and 10 U of RNase inhibitor (Promega) were added. Full length or truncated La proteins were added and incubated for 30 min at 4°C and 10 min at room temperature. The final volume of the reaction was 10 μl. For K_d measurements in the presence of Ro60, Ro60 was added to 38.4 nM and incubated 30 min at 4°C and 10 min at room temperature before adding La. After performing EMSAs as described, dried gels were scanned using a Typhoon FLA 7000 Phosphorimager (Cytiva), and fractions of bound RNA quantified with ImageJ (NIH). Subsequent data analysis is described under Quantification and Statistical Analyses, below.

Microscale thermophoresis (MST)

Complexes were assembled as described for EMSAs using 1 nM minimal misfolded pre-5S rRNA labeled at the 5’ end with Cy5 (Horizon Discovery; Table S3). The concentration of Tween 20 was reduced to 0.005% to avoid interference from detergent micelles. Twofold serial dilutions of full-length La, La 1–194, La 222–408, La 222–336, and La 335–408 were added, starting with 400 nM, 50 nM, 6.9 μM, 9.7 μM and 5.4 μM, respectively, and the reactions incubated for 30 min at room temperature. To measure binding affinities in the presence of Ro60, 100 nM of Ro60 was incubated with the RNA for 30 min at 4 °C and 10 min at room temperature before adding La proteins. Twofold serial dilutions of full-length La 222–408, La 222–336, and La 335–408 were added, beginning with 860 nM, 9.7 μM, 5.4 μM, respectively, and the reactions incubated for 30 min at room temperature. Samples were loaded into Monolith Premium Capillaries (Nanotemper Technologies MO-K025) and equilibrated to 30 °C for 10 min. Experiments were performed at 30 °C using the Monolith NT.115 Pico system (Nanotemper Technologies) with 35 % LED power and medium MST power with laser off/on times of 0 s and 10 s, respectively. All experiments were performed in triplicate. Dissociation constants (K_d) were determined as described under Quantification and Statistical Analyses.

SHAPE-MaP

RNA-protein complexes were formed as described above, with RNA, Ro60 and La each at a final concentration of 1 μM. The resulting RNAs or RNPs (90 μl each) were mixed with 10 μl of dimethylsulfoxide or 100 mM 1M7 (Tocris 6602) and incubated at 37°C for 2 min⁴². After fractionating in 5 % polyacrylamide (37.5:1 acrylamide:bisacrylamide)/5 % glycerol gels, RNA or RNP bands were excised and the eluted RNAs extracted with phenol:chloroform:isoamyl alcohol (50:49:1) and precipitated with ethanol. For the truncated human misfolded pre-5S, 90 μl of 1 μM RNA in buffer F was mixed with 10 μl of dimethylsulfoxide or 1 M 2A3 (Tocris 7376) and incubated at 37°C for 20 min. The reaction was quenched with 100 μl of 1 M DTT and the RNA extracted with phenol:chloroform:isoamyl alcohol (50:49:1) and precipitated with ethanol.

For library preparation, splint ligation was used to ligate the first adaptor to only those RNAs with full-length, correct 3’ ends¹⁰². The ligation reaction contained 0.65 μM RNA, 0.65 μM pre-adenylated L3-App¹⁰³, 0.65 μM 3’ blocked DNA bridge (5Smt_DNA for full-length and minimal misfolded pre-5S rRNAs, WT_DNA for X. laevis wildtype pre-5S rRNA, and h5S_DNA for human truncated misfolded pre-5S rRNA; Table S3), 1X RNA ligase buffer (New England Biolabs), 1 U/μl RNase inhibitor (Promega N2515), and 20 U/μl T4 DNA ligase (New England Biolabs M0202S) in a volume of 10 μl, and was incubated 1 h at 37°C. Afterwards, samples were extracted with phenol:chloroform:isoamyl alcohol (50:49:1), precipitated with ethanol, and resuspended in 11 μl of water. For reverse transcription, 5 μl of the RNA and RT oligo were used with SuperScript II (Invitrogen 18084022) as described⁴². Briefly, the RNA was mixed with 2 pmol of RT oligo in a volume of 11 μl, incubated at 90°C for 3 min, and then cooled on ice. Then, 8 μl of 2.5 X MaP buffer (125 mM Tris pH 8, 187.5 mM KCl, 15 mM MnCl₂, 25 mM DTT and 1.25 mM of each dNTP) was added and the mixture was incubated at 42°C for 2 min. Finally, 1 μl of SuperScript II reverse transcriptase was added, and the reaction was incubated for 3 hr at 42°C. Subsequent library preparation steps were as described¹⁰³. Briefly, the DNA was extracted with Dynabeads MyONE silane (Thermo Fisher 37002D). The second adaptor was then ligated in a final volume of 21 μl using 1 X RNA ligase buffer (New England Biolabs), 20 nmol ATP, 21 % PEG 8000, and 45 U of T4 RNA ligase 1, high concentration (New England Biolabs M0437M), at 25°C overnight. After extracting the cDNA with Dynabeads MyONE silane, it was amplified using optimized cycle numbers. Libraries were quantified by qPCR using KAPA library quantification kit (Roche 07960140001), checked for quality using a 4150 TapeStation (Agilent), and sequenced on a NextSeq 1000 sequencer (Illumina). Results were analyzed using Shapemapper2–2.2¹⁰⁴. All experiments were performed twice (Table S2). In Figures 5, S4 and 6, repeat 1 is shown.

Purification of Ro60/La RNPs from HaCaT cells

To isolate Ro60/La RNPs, RO60^−/−;TS-Ro60 cells were washed in Dulbecco’s Phosphate-Buffered Saline (PBS; Millipore Sigma D8537) and harvested in NET-2 (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 2.5 mM MgCl₂, 0.5 % NP-40) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and protease inhibitor cocktail. Cells were sonicated in a Bioruptor Plus (Diagenode B01020002) and the lysate cleared by centrifuging for 5 min at 10,000 x g at 4 °C. Lysates were incubated with MagStrep Strep-Tactin XT beads (IBA 2-4090-010) for 1.5 hrs at 4°C with rotation. After washing three times with NET-2, beads were incubated with NET-2 containing 50 mM D-(+)-Biotin (Sigma Aldrich 2031) for 30 min at 4°C to elute Twin-Strep-Ro60 RNPs. To isolate La-containing RNPs, half the eluate was incubated with Protein G Dynabeads (ThermoFisher 10004D) pre-coupled to a patient anti-La serum (a gift of John Harley, University of Cincinnati) at 4 °C for 1.5 hours with rotation. The remaining eluate was incubated with Protein G Dynabeads coupled to normal human sera. After three washes with NET-2, RNAs were extracted with phenol:chloroform:isoamyl alcohol (50:49:1) followed by addition of glycoblue (Invitrogen) and ethanol precipitation. Because the patient anti-La serum may no longer be available, we suggest mouse monoclonal anti-La/SSB antibody coupled to agarose beads (Santa Cruz Biotechnology sc-80656 AC) as an alternative. The cDNA library was prepared as described¹⁰³, except that T4 RNA ligase 2 (truncated KQ) (New England Biolabs M0373S) was used for the first adaptor ligation and SuperScript III (Thermo Fisher 18080044) was used for reverse transcription. Briefly, ligation was performed at 28°C for 1 hr, 25°C for 6 hr and incubated overnight at 4°C. After isolating the RNA using Dynabeads MyONE silane, the RNA was mixed with 0.5 pmol RT oligo and 10 nmol dNTP in a volume of 12 μl, incubated at 70°C for 5 min, and held at 25 °C. The reaction was then mixed with 1 X RT buffer (Thermo Fisher), 5 mM DTT, 0.5 U/μl SUPERase·In RNase inhibitor (Thermo Fisher AM2696), and 5 U/μl SuperScript III in a final volume of 20 μl. It was incubated at 25°C for 5 min, 42 °C for 20 min, 50°C for 40 min, 80°C for 5 min, and then held at 4°C. Subsequent library preparation was performed as described above for SHAPE-MaP. The library was sequenced on an Illumina NextSeq in 100 nt paired end mode.

After adapter trimming and demultiplexing, reads were mapped to the human genome (hg38) using NovoAlign (Novocraft) (Table S4) and deduplicated using UMI-tools¹⁰⁵. Multimapping was allowed to the maximum level since RNAs such as 5S rRNA and Y RNAs were expected to map to multiple loci. Counting of individual RNA species was performed with featureCounts using the --fraction option¹⁰⁶. Gene annotation files were downloaded from the UCSC genome browser.

Immunoblotting

After washing HaCaT cells with PBS and harvesting in NET-2 containing protease inhibitor cocktail and 1 mM PMSF, cells were lysed by sonication with a BioRuptor Plus on low power at 4 °C. Lysates were fractionated in NuPAGE 4–12% Bis-Tris gels (Invitrogen NP0302BOX) with NuPAGE MOPS SDS running buffer (Invitrogen NP0001). After transferring to nitrocellulose (Thermo Scientific 88018) at 30 V for 16 hr in NuPAGE Transfer Buffer (Invitrogen NP00061), membranes were blocked overnight in 5% nonfat dry milk in TBS-T (10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.1% TWEEN 20). After washing twice with TBS-T at room temperature, membranes were incubated with primary antibodies diluted in 5% milk in TBS-T for one hour, washed with TBS-T, and incubated with HRP-coupled secondary antibody diluted in 5% milk in TBS-T for 1 hour at room temperature. After washing with TBS-T, blots were visualized with WesternBright ECL HRP substrate (Advansta Inc. K-12045-D50) and imaged with a G:Box imaging system (Syngene). Primary antibodies were monoclonal anti-Ro60¹⁰⁷ and anti-GAPDH (Proteintech, 60004–1-Ig). The secondary antibody was Peroxidase AffiniPure goat anti-mouse IgG, light chain-specific (Jackson ImmunoResearch 115-035-174).

Northern blotting

After extracting RNA from HaCaT cells with Trizol (Invitrogen 15596018), RNA was resuspended in 8 M urea, 20 mM Tris-HCl pH 8.0, 1 mM EDTA, 0.05% w/v bromophenol blue, 0.05% w/v xylene cyanol and fractionated in 5% polyacrylamide, 8 M urea gels. After transferring to Hybond-N+ (Cytiva RPN303B) in 50 mM Tris, 45 mM boric acid, 0.5 mM EDTA, RNA was immobilized on the membrane by UV crosslinking. Blots were hybridized with [γ-³²P]-ATP-labeled oligonucleotides in Church and Gilbert hybridization buffer¹⁰⁸ at 35 °C overnight. To detect individual Y RNAs, membranes were probed simultaneously with multiple oligonucleotides to increase the signal. Specifically, oligonucleotides RNY1_1, RNY1_2, and RNY1_3 were used to detect Y1 RNA, RNY3_1, RNY3_2, RNY3_3 and RNY3_4 were used to detect Y3 RNA, RNY4_1, RNY4_2 and RNY4_3 were used to detect Y4 RNA, and RNY5_1, RNY5_2 and RNY5_3 were used to detect Y5 RNA (Table S3). Membranes were washed in 1X SSC (150 mM NaCl, 15 mM sodium citrate pH 7.0) containing 0.1% SDS for 30 min at 35° C, followed by 30 minutes in 0.5X SSC, 0.1% SDS at 35 °C. Membranes were imaged using a Typhoon FLA 7000 Phosphorimager.

Exoribonuclease assays

After mixing 9 fmol of 5’-end labeled RNA with 6 fmol of unlabeled RNA, the RNA was refolded as described above in 20 mM HEPES pH 8.0, 125 mM NaCl, 1.5 mM MgCl₂, 1 mM EDTA, 0.005% Tween 20, and 5% glycerol. To reconstitute the ternary complex, 37.5 fmol of Ro60 was added first and incubated for 30 min at 4°C and 10 min at room temperature. Next, 45 fmol of La was added to the Ro60/RNA and RNA samples and incubated for the same times and temperatures. To initiate degradation, 22.5 fmol of human DIS3L2 (Origene TP035512M) or 37.5 fmol of ERI1 was added and incubated at 37°C in a final volume of 15 μl. At each time point, aliquots were mixed with two volumes of 90% formamide, 1 mM EDTA, 0.25% SDS, 0.02% bromophenol blue, and 0.01% xylene cyanol and kept on dry ice until all reactions were complete. Samples were heated at 90°C for 5 min and fractionated in 8% polyacrylamide/8 M urea gels. Dried gels were scanned using a Typhoon FLA 7000 Phosphorimager.

QUANTIFICATION AND STATISTICAL ANALYSES

For the quantitative EMSAs, dried gels were imaged using a Typhoon FLA 7000 Phosphorimager (Cytiva), and the fraction of bound RNA was quantified with ImageJ (NIH). Data were analyzed with GraphPad Prism 10 and fitted by nonlinear regression using a specific binding model with Hill slope: $\text{Y}=\begin{array}{c}{\text{B}}_{\text{max}}\times {\text{X}}^{\text{h}}\\ {{\text{K}}_{\text{d}}}^{\text{h}}+{\text{X}}^{\text{h}}\end{array}$, where Y is the fraction of bound RNA, B_max is the maximum specific binding, X is the protein concentration, h is the Hill slope, and K_d is the equilibrium binding constant. All means and standard deviations were derived from the graphed data for three technical replicates.

For the MST measurements, dissociation constants (K_d) were determined with the Hill model using MO Control software v 2.6.5 (Nanotemper Technologies). All experiments were performed as three technical replicates (n=3). All graphs were generated with GraphPad Prism 10.

For SHAPE-MaP, normalized reactivity was obtained using ShapeMapper version 2.2.0. Reactivity differences were calculated by subtracting the reactivity of the unbound RNA from the reactivity for the protein-bound RNA.

Supplementary Material

1

Figure S1. Fab/ternary complex assembly and cryo-EM processing workflow, including local and global resolution estimates, related to Figure 1.

(A and B) After subjecting the concentrated ternary complexes to gel filtration (A), fractions containing the complex (B) were pooled.

(C) To determine the amount of Fab, increasing amounts were added to 25 μM ternary complex and the Fab-bound RNP separated from free Fab and the unbound RNP in native gels.

(D) To confirm that the Fab remained associated, the ternary complex and the Fab-bound complex were fractionated in native gels before and after concentration.

(E) Image processing workflow with consensus and focused maps annotated as I, II, or III.

(F) Gold-standard Fourier Shell Correlation (FSC) curves used for global resolution estimation of all maps (left; map I, middle; map II, right; map III).

(G) Local resolution estimate of the volumes (left; map I, middle; map II, right; map III).

2

Figure S2. Cryo-EM processing workflow for free Ro60 and the Ro60/minimal misfolded pre-5S RNP, including local and global resolution estimates, related to Figure 2.

(A) Image processing workflow for Ro60.

(B) Gold-standard Fourier Shell Correlation (FSC) curves used for global resolution estimation of Ro60.

(C) Local resolution estimate of the Ro60 volume.

(D) Image processing workflow of the Ro60/minimal misfolded pre-5S RNP.

(E) Gold-standard Fourier Shell Correlation (FSC) curves used for global resolution estimation of Ro60/minimal misfolded pre-5S RNP.

(F) Local resolution estimate of the Ro60/minimal misfolded pre-5S RNP volume.

3

Figure S3. Structural changes on binding of Ro60 and La to misfolded RNA, related to Figures 2 and 3.

(A) Schematic representations of Ro60 secondary structures in the wide and narrow conformations.

(B) Comparison of Ro60 helices H2 and H6 in the wide and narrow Ro60 conformations.

(C) Interactions of the misfolded RNA in the Ro60 RNP. Solid lines, possible hydrogen bonds (< 3.5 Å). Dotted lines, potential electrostatic interactions (< 7 Å). Residues involved in π-stacking are highlighted in yellow. For clarity, interactions with ribose rings are not indicated. Of the 64 nts, 44 are modeled. Dotted boxes, nts not modeled.

(D and E) Figures 2E (D) and 3E (E) overlayed with cryo-EM density map. Dotted lines, possible hydrogen bonds.

(F) Comparison of the La NTD in the ternary complex (left) and in a crystal structure bound to single-stranded RNA (middle, PDB 2VON). The RMSD was 0.92 Å (right panel).

(G) Interactions of the LARP7 ortholog p65 xRRM with telomerase RNA (left, PDB 8GAP) and the LARP7 xRRM with 7SK RNA (middle, PDB 7SLP). Right, close-up of the interaction of 7SK RNA G314 with LARP7 V484 and R468.

(H) Alignment of La C-terminal disordered regions. Residues with more than 70% conservation are red and those with 50 – 70% conservation are orange.

4

Figure S4. Ro60 and the La CTD disrupt weak RNA elements, related to Figure 5.

(A and B) Changes in SHAPE reactivity upon binding of Ro60, La, Ro60 followed by La, and Ro60 followed by the La CTD (amino acids 222–408), to full-length X. laevis wild-type (A) and misfolded pre-5S rRNAs (B).

5

Figure S5. Structured ncRNAs containing 3’ uridylates are present in Ro60/La RNPs, related to Figure 6.

(A) Ro60 levels in wild-type, RO60^−/− and RO60^−/−, TS-RO60 HaCaT cells were compared by immunoblotting. GAPDH, loading control.

(B) RNAs from the indicated cell lines were subjected to Northern blotting to detect Y1, Y3, Y4 and Y5 RNAs, which are all stabilized by Ro60. 7SL, loading control.

(C) After mapping sequenced reads to the hg38 human genome assembly, the fractions of reads in each category is shown.

(D) RNAs most enriched in Ro60/La RNPs compared to the control immunoprecipitate.

(E) Human 5S rRNAs are encoded antisense to Alu SINEs in tandem repeats. Three termination sites are boxed. 5S rRNAs with 3’ extensions are enriched in Ro60/La RNPs.

(F) Reads mapping to tRNA-Tyr-GUA. Arrow, transcription direction.

(G) 5S rRNAs containing U-tails are enriched in Ro60/La RNPs.

(H) Fragment derived from 12S mt-rRNA 3’ end.

(I) Some reads from 12S mt-rRNA 3’ ends contain non-templated tails. Genomic sequence is on top. Arrow; 12S mt-rRNA 3’ end. Non-templated nts are bold, with Ts in red.

(J) The secondary structure of pre-mir-let-7i is shown on top, with the 3’ ends of reads mapping to the mir-let-7i locus below. Green, sequences of encoded miRNAs. Red, non-templated Ts.

(K) RNA used in decay assays. Bracket, site of ERI1 stall.

(L and M) After assembling the RNA with the indicated proteins, DIS3L2 (L) or ERI1 (M) was added and aliquots removed at intervals. Nuclease was omitted in lanes 13–16. Limit oligonucleotides that accumulate during DIS3L2 degradation are indicated (L). The site at which ERI1 stalls is indicated by the brackets in (K) and (M).

6

Figure S6. RNPs formed by human 5S rRNAs containing 3’ extensions, related to Figure 6.

(A) Secondary structures of human U-tailed 5S rRNA, pre-5S rRNA, and pre-5S rRNA containing nt changes that force misfolding (red box and circle).

(B) X. laevis misfolded pre-5S and H. sapiens U-tailed 5S and pre-5S rRNAs were mixed with the indicated proteins and fractionated in native gels. In lanes 4, 9 and 14, Ro60 was added before La. Asterisks, ternary complexes that form when Ro60 is added first.

(C) H. sapiens pre-5S rRNA was mixed with the indicated proteins in the absence and presence of 1.5 mM MgCl₂ and fractionated in native gels. In lanes 4, 6, 10 and 12, Ro60 was added before La. In lanes 4 and 10, RNPs were incubated at 4°C, followed by room temperature incubation. In lanes 6 and 12, RNPs were incubated at room temperature, then at 4°C.

(D) Cryo-EM density map of the Ro60/human U-tailed 5S rRNP at 3.5 Å resolution.

7

Figure S7. Cryo-EM processing workflow for Ro60/human U-tailed 5S rRNA complex, including local and global resolution estimates, related to Figure 6.

(A) Image processing workflow.

(B) Gold-standard Fourier Shell Correlation (FSC) curves used for global resolution estimation.

(C) Local resolution estimate of the final volume.

8

Figure S8. Cryo-EM processing workflow for the Ro60/La/truncated pre-5S RNA/Fab complex, including local and global resolution estimates, related to Figure 6.

(A) Image processing workflow with consensus and focused map annotated as I or II, respectively.

(B) Gold-standard Fourier Shell Correlation (FSC) curves used for global resolution estimation of the consensus (left) or focused (right) maps.

(C) Local resolution estimate of the consensus (left) or focused (right) maps.

Document S1. Tables S1, S3 and S4

Table S2. SHAPE-MaP results, related to Figures 5, 6, and S4

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse Anti-La/SSB recombinant antibody; Fab fragment	Creative BioLabs	Cat# HPAB-0883-FY-F(E)
Patient anti-La serum	gift of J. Harley, U.S. Department of Veterans Affairs Medical Center, Cincinnati, OH	N/A
Anti-La/SSB antibody, conjugated to agarose	Santa Cruz Biotechnology	Cat# sc-80656 AC; RRID:AB 1124923
Monoclonal mouse-anti-Ro60	Xue et al.107	N/A
Anti-GAPDH	Proteintech	Cat# 60004-1-Ig; RRID:AB 2107436
Peroxidase AffiniPure goat anti-mouse IgG, light chain-specific	Jackson ImmunoResearch	Cat# 115-035-174; RRID:AB_2338512
Bacterial and virus strains
E. coli: DH10EMBacY cells	gift of Eugene Valkov, National Cancer Institute, Frederick, MD	N/A
E. coli: BL21(DE3)pLysS	Promega	Cat# L1195
Chemicals, peptides, and recombinant proteins
cOmplete™, Mini, EDTA-free Protease Inhibitor Cocktail	Millipore Sigma	Cat# 11836170001
T7 RNA polymerase containing P266L mutation	Homemade. Plasmid92 was a gift from Jennifer Miller, National Cancer Institute, Frederick MD	N/A
Inorganic Pyrophosphatase (E. coli)	New England Biolabs	Cat# M0361L
T4 Polynucleotide Kinase	New England Biolabs	Cat# M0201L
RNasin RNase Inhibitor, Recombinant	Promega Corporation	Cat# N2515
Poly(deoxyinosinic-deoxycytidylic) acid	Millipore Sigma	Cat# P4929
Ultrapure BSA Bovine serum albumin	ThermoFisher	Cat# AM2616
1M7	Tocris	Cat# 6602
2A3	Tocris	Cat# 7376
T4 DNA ligase	New England Biolabs	Cat# M0202S
Invitrogen Superscript II Reverse Transcriptase	Thermo Fisher	Cat# 18084022
Invitrogen Superscript III Reverse Transcriptase	Thermo Fisher	Cat# 18080044
Invitrogen Superase In RNase Inhibitor	Thermo Fisher	Cat# AM2696
Lipofectamine 3000	ThermoFisher	Cat# L3000015
Blasticidin S	InvivoGen	Cat# ant-bl-05
Puromycin	Sigma Aldrich	Cat# P8833
D-(+)-Biotin	Sigma Aldrich	Cat# 2031
Glycoblue Coprecipitant	ThermoFisher	Cat# AM9516
T4 RNA ligase 2 (truncated KQ)	New England Biolabs	Cat# M0373S
T4 RNA ligase 1, High concentration	New England Biolabs	Cat# M0437M
KLD Enzyme Mix	New England Biolabs	Cat# M0554S
DIS3L2 (NM_152383) Human Recombinant Protein	Origene	Cat# TP035512M
Critical commercial assays
KAPA Library Quantification Kit – Illumina-UNI;	Roche Diagnostics	Cat# 07960140001
Deposited data
Cryo-EM structure of Ro60/La/minimal misfolded pre-5S rRNA complex with Fab, consensus map	This paper	EMD-49357
Cryo-EM structure of Ro60/La/minimal misfolded pre-5S rRNA complex with Fab, N-terminal La/Fab focused map	This paper	EMD-49358
Cryo-EM structure of Ro60/La/minimal misfolded pre-5S rRNA complex with Fab, C-terminal La focused map	This paper	EMD-49359
Cryo-EM structure of Ro60/La/minimal misfolded pre-5S rRNA complex with Fab, composite map	This paper	EMD-49360, PDB 9NFA
Cryo-EM structure of human Ro60	This paper	EMD-49328, PDB 9NEN
Cryo-EM structure of Ro60/minimal misfolded pre-5S rRNA complex	This paper	EMD-49329, PDB 9NEP
Cryo-EM structure of Ro60/U-tailed human 5S rRNA complex	This paper	EMD-49327
Cryo-EM structure of Ro60/La/truncated misfolded human pre-5S rRNA complex with Fab, consensus map	This paper	EMD-49353
Cryo-EM structure of Ro60/La/truncated misfolded human pre-5S rRNA complex with Fab, focused map	This paper	EMD-49354
Cryo-EM structure of Ro60/La/truncated misfolded human pre-5S rRNA complex with Fab, composite map	This paper	EMD-49355, PDB 9NF8
Sequencing data: RNAs present in Ro60/La RNPs	This paper	GEO: GSE290478
Experimental models: Cell lines
Spodoptera frugiperda: Sf21 cells	gift of Eugene Valkov, National Cancer Institute, Frederick, MD	N/A
H. sapiens: HaCaT cells	Cytion	Cat# 300493; RRID:CVCL 0038
H. sapiens: RO60−/− HaCaT cells	This paper	N/A
H. sapiens: RO60−/−;TS-Ro60 HaCaT cells expressing	This paper	N/A
Ro60 with Twin-Strep Tag
Oligonucleotides
DNA and RNA oligonucleotides	This paper	Table S3
Recombinant DNA
human Ro60 cDNA containing codon-optimized cDNA (Accession ID J04137)	Integrated DNA Technologies (IDT)	N/A
pLIB	Addgene	Cat# 80610
Human Ro60 cDNA in pLIB	This paper	N/A
pET-28a(+) Novagen	Millipore Sigma	Cat# 69864
Human La cDNA	Gift of Jack Keene, Duke University; Chambers and Keene90	N/A
Human La cDNA in pET28	This paper	N/A
Human La CTD (amino acids 222-408) in pET28a(+)	This paper	N/A
Human La RRM2 (amino acids 222-336) in pET28a(+)	This paper	N/A
Human La disordered region (amino acids 335-408) in	This paper	N/A
pET28a(+)
p2RZ	Addgene	Cat# 27644
X. laevis wild-type pre-5S rRNA in p2RZ	This paper	N/A
X. laevis mutant misfolded pre-5S rRNA in p2RZ	This paper	N/A
X. laevis minimal misfolded pre-5S rRNA in p2RZ	This paper	N/A
X. laevis minimal misfolded pre-5S rRNA containing 15 additional uridylates at 3’ end in p2RZ	This paper	N/A
H. sapiens wild-type pre-5S RNA in p2RZ	This paper	N/A
H. sapiens U-tailed 5S rRNA in p2RZ	This paper	N/A
H. sapiens full-length mutant misfolded RNA in p2RZ	This paper	N/A
X. laevis mutant misfolded pre-5S rRNA under control of a T7 promoter in pSP64	Shi et al.36	N/A
pX459	Addgene	Cat# 62988
Piggybac transposon plasmid	Gift of Tian Xu, Westlake University, China, Ding et al.87	N/A
Ro60 with a N-terminal Twin-Strep tag in Piggybac transposon plasmid Software and algorithms	This paper	N/A
SIMPLE 3.0	Caesar et al.93	https://github.com/hael/SIMPLE
cryoSPARC	Punjani et al.94	https://cryosparc.com
RELION 3.1	Zivanov et al.95	https://github.com/3dem/relion
pyEM	Asarnow et al.96	https://github.com/asarnow//pyem
deepEMhancer	Sanchez-Garcia et al.97	https://github.com/rsanchezgarc/deepEMhancer
ChimeraX v.1.6.1	Pettersen et al.84	https://www.cgl.ucsf.edu/chimerax/
Alphafold	Jumper et al.77	https://alphafold.ebi.ac.uk
Coot v0.9.8.8	Emsley et al.98	https://www2.mrc-lmb.cam.ac.uk/personal/pemsley/coot/
PHENIX	Afonine et al.99	https://phenix-online.org/
MolProbity	Williams et al.100	https://phenix-online.org/documentation/reference/molprobity_tool.html
ConSurf	Ashkenazy et al.83	https://consurf.tau.ac.il/consurf_index.php
T-Coffee	Notredame et al.101	https://www.ebi.aauk/jdispatcher/msa/tcoffee?stype=protein
MO Control Software v2.6.5	NanotemperTechnologies	https://support.nanotempertech.com/hc/en-us/sections/17715198724753-Software
Shapemapper2-2.2	Busan and Weeks104	https://github.com/Weeks-UNC/shapemapper2
NovoAlign	Novocraft	https://www.novocraft.com/
UMI-tools	Smith et al.105	https://github.com/CGATOxford/UMI-tools
featureCounts	Liao et al.106	https://subread.sourceforge.net/
Other
HiTrap Heparin HP column	Cytiva	Cat# 17040601
HiLoad 16/600 Superdex 200 pg column	Cytiva	Cat# 28989335
His-Pur Ni-NTA resin	Thermo Scientific	Cat# 88222
PD-10 Desalting Column	Cytiva	Cat# 17085101
Superdex 200 Increase 10/300 GL column	Cytiva	Cat# 28990944
Pierce Protein Concentrators PES, 100kDa MWCO	VWR International, Inc.	Cat# P188503
Amicon Ultra Centrifugal Filter, 10 kDa MWCO	Millipore Sigma	Cat# UFC901024
Amicon Ultra Centrifugal Filter, 50 kDa MWCO	Millipore Sigma	Cat# UFC205024
Monolith Premium Capillaries	Nanotemper Technologies	Cat# MO-K025
MagStrep Strep-Tactin “type 3” XT beads	IBA	Cat# 2-4090-010
QUANTIFOIL® Carbon R 1.2/1.3 on Gold 300	Quantifoil	Cat# N1-C14nAu30-01
BioRuptor Plus	Diagenode	Cat# B01020002
Dynabeads Protein G for Immunoprecipitation	ThermoFisher	Cat# 10004D
Dynabeads MyONE silane	ThermoFisher	Cat# 37002D

HIGHLIGHTS.

• Cryo-EM structures reveal that La cradles the Ro60-bound misfolded RNA

• The La C-terminal domain destabilizes structures in the misfolded RNA body.

• La and Ro60 function synergistically to destabilize weak RNA structural elements

• RNAs bound by Ro60 and La include ncRNA precursors and RNAs with oligouridine tails