Proteomic analysis of the SMN complex reveals conserved and etiologic connections to the proteostasis network

Abstract

Introduction: Molecular chaperones and co-chaperones are highly conserved cellular components that perform a variety of duties related to the proper three-dimensional folding of the proteome. The web of factors that carries out this essential task is called the proteostasis network (PN). Ribonucleoproteins (RNPs) represent an underexplored area in terms of the connections they make with the PN. The Survival Motor Neuron (SMN) complex is an assembly chaperone and serves as a paradigm for studying how specific RNAs are identified and paired with their client substrate proteins to form RNPs. SMN is the eponymous component of a large complex, required for the biogenesis of uridine-rich small nuclear ribonucleoproteins (U-snRNPs), that localizes to distinct membraneless organelles in both the nucleus and cytoplasm of animal cells. SMN protein forms the oligomeric core of this complex, and missense mutations in the human SMN1 gene are known to cause Spinal Muscular Atrophy (SMA). The basic framework for understanding how snRNAs are assembled into U-snRNPs is known. However, the pathways and mechanisms used by cells to regulate their biogenesis are poorly understood.

Methods: Given the importance of these processes to normal development as well as neurodegenerative disease, we set out to identify and characterize novel SMN binding partners. We carried out affinity purification mass spectrometry (AP-MS) of Drosophila SMN complexes using fly lines exclusively expressing either wildtype or SMA-causing missense alleles.

Results: Bioinformatic analyses of the pulldown data, along with comparisons to proximity labeling studies carried out in human cells, revealed conserved connections to at least two other major chaperone systems including heat shock folding chaperones (HSPs) and histone/nucleosome assembly chaperones. Notably, we found that heat shock cognate protein Hsc70-4 and other HspA family members preferentially associated with SMA-causing alleles of SMN.

Discussion: Hsc70-4 is particularly interesting because its mRNA is aberrantly sequestered by a mutant form of TDP-43 in mouse and Drosophila ALS (Amyotrophic Lateral Sclerosis) disease models. Most important, a missense allele of Hsc70-4 (HspA8 in mammals) was recently identified as a bypass suppressor of the SMA phenotype in mice. Collectively, these findings suggest that chaperone-related dysfunction lies at the etiological root of both ALS and SMA.

Keywords: AP-MS; Amyotrophic lateral sclerosis (ALS); affinity purification coupled with mass spectrometry; chaperone mediated autophagy; proteostasis networks; ribonucleoprotein (RNP) biogenesis; spinal muscular atrophy (SMA).

FIGURE 1

Experimental setup, workflow and overview of results. (A) Ideogram of fruitfly chromosome 3 (Chr3) showing cytogenetic band locations of the endogenous Smn gene at 73A9 and the Flag-Smn rescue transgene (Tg) at a PhiC31 landing site located at 86F8. Cartoon above shows the features of the rescue transgene driven by the native Smn promoter and control region. The SMN coding region was tagged with a 3x-FLAG epitope at the N-terminus. In addition to the WT Smn line, stocks expressing four different missense alleles (D20V, G73R, I93F and G210C) were also generated. (B) Diagram of experimental workflow, beginning with embryo collection and ending with identification of peptides and protein predictions. Panels created using Biorender.com. (C) Western blot of immunoprecipitation (IP) experiment for Oregon-R (OreR) control or the various stable transgenic stocks described above. The upper blot was probed with anti-Flag and the lower blot was probed with anti-SMN, verifying the presence of the untagged endogenous protein in the OreR input lane, but not in any of the other lanes. (D) Venn diagram of the total number of proteins identified in Flag-IPs from Smn^WT transgenic animals, comparing the four biological replicates (WT1–4) generated in this study with a fifth one from a previous dataset, WT0 (Gray et al., 2018). (E) Graphical heat representation of the top 100 protein hits as determined by Log2 fold-change (LFC) from the WT samples vs. the OreR controls. Heatscale is at bottom right. Diameters of the circles are proportional to calculated LFC values for each protein. Panel was created using Cytoscape.com.

FIGURE 2

Analysis of AP-MS data. (A) Volcano plot of the full Smn^WT dataset. Dots represent individual proteins. The SMN complex is shown in dark blue, known SMN clients (Sm proteins) are shaded light blue. Other notable proteins are shaded black. Vertical dotted line represents a fold-change cutoff of 1.5x (LFC ≥0.58) for enriched proteins (shaded in red). Horizontal dotted line is shown for display purposes, see text for details regarding significance cutoffs. (B) Heatmap of fold-change ratios for well-known SMN binding partners, comparing the data from the WT pulldowns to those of the G210C, D20V and Tud mutants. Tud = combined results for G73R and I93F. AllMut = combined results for all of the mutants. (C) Cartoon of known intermediate in spliceosomal snRNP assembly pathway, showing the seven canonical Sm proteins (B, D3, D1, D2, F, E and G), Gemin2 (Gem2), and SMN (with its three differentially shaded domains corresponding to those in Figure 1A). See text for details.

FIGURE 3

Functional enrichment analysis. (A) Comparative gene ontology (GO) terms associated with the top 100 proteins identified in the Smn^WT and Smn^Tud AP-MS datasets. For each GO term listed, the size of the dot is proportional to the number of genes contained within that term (gene count), and the fraction of those genes scoring significantly (gene ratio) is represented using a heatmap (legend at right). Adjusted p-values (−log10 transformed) for each GO term were calculated and plotted separately for the WT and Tud results. (B) A volcano plot of the full Smn^Tud dataset. Dots represent individual proteins, shaded as per Figure 2A and shown in the color key (inset). Heat shock proteins are highlighted in orange. Hsc70-4 is circled. See text for details.

FIGURE 4

Comparison of previously published (Youn et al., 2018) BioID proximity labeling data for human stress granule proteins. Fold-change scatter plots of GEMIN3/DDX20 versus (A) STRAP/UNRIP and (B) LSM2, are shown. Color key is inset in panel (A). The inset in panel B provides a table listing the top ten BirA-tagged baits that included SMN as a prey. Fold-change (FC) values shown for comparison. See text.

FIGURE 5

Diagram of protein-protein interactions between three key molecular chaperone systems: SMN (RNP assembly), Hsc70-4 (protein folding) and Nap1 (nucleosome assembly). Solid lines indicated known interactions. Factors listed in bold text are were identified in this work; those in green are innate immune signaling factors that were shown to interact genetically with SMN (Garcia et al., 2024).