Authors: Zijian Zhang (1Department of Chemical and Systems Biology, Stanford School of Medicine; Stanford, 94305, USA.; 2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Adele Xu (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Yunhao Bai (3Department of Chemistry, Stanford University; Stanford, 94305, USA.; 4Department of Pathology, Stanford School of Medicine; Stanford, 94305, USA.), Yuxiang Chen (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Kitra Cates (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Craig Kerr (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Abel Bermudez (5Department of Radiology, Stanford School of Medicine; Stanford, 94305, USA.), Teodorus Theo Susanto (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Kelsie Wysong (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.), Fernando J. García Marqués (5Department of Radiology, Stanford School of Medicine; Stanford, 94305, USA.), Garry P. Nolan (4Department of Pathology, Stanford School of Medicine; Stanford, 94305, USA.), Sharon Pitteri (5Department of Radiology, Stanford School of Medicine; Stanford, 94305, USA.), Maria Barna (2Department of Genetics, Stanford School of Medicine; Stanford, 94305, USA.)
Categories: Article
Source: Science (New York, N.Y.)
Authors: Zijian Zhang, Adele Xu, Yunhao Bai, Yuxiang Chen, Kitra Cates, Craig Kerr, Abel Bermudez, Teodorus Theo Susanto, Kelsie Wysong, Fernando J. García Marqués, Garry P. Nolan, Sharon Pitteri, Maria Barna
Millions of ribosomes are packed within mammalian cells, yet we lack tools to visualize them in toto and characterize their subcellular composition. Here, we present ribosome expansion microscopy (RiboExM) to visualize individual ribosomes and an optogenetic proximity-labeling technique (ALIBi) to probe their composition. We generated a super-resolution ribosomal map, revealing subcellular translational hotspots and enrichment of 60S subunits near polysomes at the endoplasmic reticulum (ER). We found that Lsg1 tethers 60S to the ER and regulates translation of select proteins. Additionally, we discovered ribosome heterogeneity at mitochondria guiding translation of metabolism-related transcripts. Finally, we visualized ribosomes in neurons, revealing a dynamic switch between monosomes and polysomes in neuronal translation. Together, these approaches enable exploration of ribosomal localization and composition at unprecedented resolution.
mRNA translation is a highly compartmentalized process at the subcellular level, as suggested by evidence of mRNA localization to the ER, mitochondria, peroxisomes, and neuronal processes (1-8). However, less is known about the subcellular localization and composition of the ribosomes that catalyze and regulate mRNA translation in these compartments, particularly in mammalian cells and specialized cell types. Substoichiometric core ribosomal proteins (RPs) (9-13), ribosome-associated proteins (RAPs) (14), RP post-translational modifications, RP splicing isoforms, and ribosomal RNA (rRNA) modifications (15-19) are emerging as mechanisms of translation regulation. However, little is known about the subcellular organization of ribosomes that bear these features, and whether they play a role in ensuring that mRNAs are translated in appropriate amounts and in the appropriate milieu.
Microscopy and biochemical isolation are frequently used to investigate subcellular phenomena, but both strategies have limitations. Because ribosomes are small and abundant, the diffraction limit of visible light impedes the use of traditional light microscopy to detect interactions of the small 40S and large 60S ribosomal subunits with each other or with other proteins or RNAs. Cryo-electron tomography (20-23) and single-molecule fluorescence resonance energy transfer (smFRET) (24, 25) have been used to observe translation dynamics of single ribosomes or small groups of ribosomes, but they are limited in their ability to image ribosomes within sufficiently large areas to understand the subcellular context of translation events and are not readily compatible with markers for organelles, mRNAs or proteins of interest. Among existing biochemical techniques for isolating subcellular samples, sequential fractionation and gradient centrifugation are labor- and time-intensive processes that cannot reliably separate some organelles (26). A substrate-specific biotin ligase has previously been used to label and isolate organelle-associated ribosomes; however, because this enzyme is constitutively active at physiologic biotin concentrations (see Supplementary Text), this technique is generally limited to cell types that tolerate biotin-depleted media (1, 2).
Here, we describe two techniques to characterize translational machinery at the subcellular level. First, we developed ribosome expansion microscopy (RiboExM) for imaging the spatial relations of individual ribosome subunits, RAPs and mRNAs to different cellular compartments and to each other. Second, we developed AviTag-specific Location-restricted Illumination-enhanced Biotinylation (ALIBi), an optically controlled tool to rapidly and specifically label and affinity purify organelle-associated ribosomes for analysis of their constituent proteins and nucleic acids. Using a combination of these methods, we explored the organelle-specific interplay between RAPs, specialized ribosomes, and localized mRNA.
Achieving single-molecule imaging of ribosomes within mammalian cells requires (i) labeling the ribosomes with bright yet small markers that do not perturb function, and (ii) detecting these markers at a resolution approximating the diameter of a single ribosome. We generated a set of engineered mammalian cell lines stably expressing RPs fused to spaghetti-monster fluorescent proteins (smFPs) (Figure 1A and S1A). A smFP is a green fluorescent protein (GFP)-like barrel structure into which multiple copies of an epitope tag (V5, HA, or Myc) are inserted (27), enhancing labeling efficiency and intensity for individual RPs when probed with antibodies against the epitope tags. We engineered a cell line in which two core RPs, RPL7a and RPS2, which are stoichiometrically present on the ribosome (10), were tagged with smFP-HA and smFP-V5 respectively (Figure 1A) to represent the 60S and 40S subunits. We confirmed stable and near-complete integration of smFPs into 40S, 60S, monosomes, and polysomes, representing the majority of mature ribosomes (Figure S1B).
Current super-resolution techniques often fall short in providing the necessary resolution or detection efficiency for a wide range of fluorophores (28, 29). Therefore, we developed ribosome expansion microscopy (RiboExM), a super-resolution imaging technique that enables standard confocal microscopy with ~25 nm sub-ribosomal resolution to match the diameter of a mammalian ribosome (30) within a single round of expansion (Figure 1B, S2A and S2B) (31-33). We adapted an optimized gel network structure to obtain a higher expansion ratio, coupled with compatible epitope anchoring, gel homogenization, and staining procedures. Notably, immunostaining after denaturation and expansion achieves greater exposure and enhanced detection of otherwise physically hindered antigens through "molecular de-crowding" (34) (Figure S2B). Given the abundance of ribosomes, which form polysomes during translation, RiboExM circumvents these challenges in labeling and visualizing efficiently (35). We verified that expansion in RiboExM is it preserves the native morphology and subcellular localization of all tested organelles and expands them to highly similar extents (~8.7 fold, Figure S2C-F).
By applying RiboExM to smFP-tagged cells, we generated a panoramic ribosome localization map (Figure S3A, Movie S1, Movie S2). Despite the abundance of ribosomes, distinct ribosomal populations can be clearly resolved as individual puncta. We then developed an automated co-localization-based analysis pipeline to discern free ribosomal subunits, monosomes, and polysomes (Figure 1C, D). First, we applied local thresholding on deconvolved images of tagged RPL7a and RPS2 to generate 60S and 40S binary masking maps, respectively. Each punctum represents a single ribosomal subunit, monosome, or polysome. The two masking maps were then processed together on a punctum-to-punctum basis, and puncta containing closely positioned tagged RPL7a and RPS2 were classified as assembled ribosomes (Figure 1C). Specifically, based on the predicted distance between smFPs on a labeled ribosome, any two masked puncta with an inter-centroid distance smaller or equal to two pixels were considered colocalized. Non-colocalized puncta were classified as free ribosomal subunits. The background-subtracted total fluorescence intensity within non-colocalized puncta showed little variation (Figure S3B); hence, we use the 85% percentile value of this distribution as a threshold for the maximum fluorescence intensity of an individual tagged subunit. In contrast, the intensity of colocalized puncta showed a considerably larger variation (Figure S3B), corresponding with the varying numbers of ribosomes within polysomes. The per-punctum intensity of RPS2-smFP and RPL7a-smFP were linearly correlated (Figure S3C), suggesting that fluorescence intensity robustly indicated the number of ribosomes within polysomes.
Based on the distribution histogram and using the threshold value for a single subunit, we designed an algorithm to distinguish monosomes (puncta with intensity below the threshold, indicating a single ribosome), light polysomes (two to five times the threshold), and heavy polysomes (over five times the threshold), thus constructing a panoramic ribosomal map of the cell (Figure 1E and F). As a control, when two smFPs were expressed without fusing them to RPs, the fluorescence intensity distribution of colocalized puncta was highly similar to that of non-colocalized puncta, suggesting that the higher-fluorescence colocalized puncta observed upon expression of RP-fused smFPs are indeed polysomes rather than random variations in fluorescence intensity (Figure S3D). To achieve enhanced resolution for visualizing individual ribosomes within polysomes, we employed stimulated emission depletion (STED) microscopy on expanded samples (36). With ~6 nm resolution, we confirmed that puncta identified as polysomes by RiboExM indeed consist of multiple closely positioned ribosomes, whereas identified monosomes mostly contain single ribosomes (Figure S3E). Per our analysis pipeline, approximately 4-6 million tagged ribosomes were detected across the cell, which was over 80% of all mature ribosomes (Figure S1B) and was comparable to previous mass-based estimations for the number of ribosomes per cell (37).
A classic method for assessing translation is separation of 40S, 60S, monosomes, and polysomes via sucrose gradient fractionation (38). Inspired by this method, we generated an imaging-based pseudo-polysome gradient by quantifying the number of 40S, 60S, monosomes, and polysomes in subcellular regions (Figure 1H-I), which revealed that the proportion of active translational machinery varied substantially throughout subcellular space (Figure S4A). Averaged across multiple cells, the proportion of 40S, 60S, monosomes, and polysomes according to our pseudo-gradient generally concurred with quantification via conventional sucrose gradient fractionation (Figure S4B). We then applied RiboExM directly to sucrose gradient purified fractions (Figure S4B). The fluorescence intensity per punctum in the polysome fractions was approximately four times the intensity in the monosome fraction, while the 40S fraction had similar per-punctum intensity as the monosome fraction (Figure S4C), suggesting that our technique can correctly distinguish ribosomal subunits, monosomes, and polysomes. One theoretical limitation of RiboExM is that a small proportion of the putative 40S and 60S subunits could be monosomes containing non-tagged RPs. To further validate our approach, we constructed a cell line in which two core 40S RPs, RPS2 and RPS8, were tagged with smFP-V5 and smFP-HA respectively (Figure S4D). We quantified the false detection rates for monosomes and polysomes to be approximately 12% and 4% respectively (Figure S4E-G). The tagged and untagged RPs exhibited no preferential association with specific organelles (Figure S4E); thus untagged RPs are not expected to compromise the analysis of subcellular ribosomal localization.
To confirm whether the identified polysomes are bona-fide actively translating polysomes, we treated cells with the translation inhibitor puromycin for varying durations to dissociate polysomes to different degrees (39). Following a 15-minute treatment, the number of detected polysome puncta rapidly diminished; after 2 hours, they were nearly absent (Figure S5A-B). The rapid effect of puromycin (40) is consistent with our findings. Cells treated with cycloheximide (CHX), which blocks eukaryotic translation elongation and thereby stabilizes polysomes (41), displayed a slight enrichment of detected polysomes after 15 minutes (Figure S5C-D). These results suggest that polysomes detected by RiboExM are indeed actively translating.
We next investigated the distribution of ribosome subunits and assembled ribosomes within subcellular space. Unsupervised spatial K-Nearest-Neighbors (KNN) clustering of the processed ribosome maps based on local 40S, 60S, and polysome density within a sliding window (42) (Figure 1J and Figure S5E-H) showed that polysomes form spatially clustered regions representing translation hotspots within the cells, and that these hotspots substantially shrank following puromycin treatment (Figure 1J-K and Figure S5I-J). Clusters enriched in 40S were smaller and more evenly distributed throughout cells (Figure 1J). 60S-enriched clusters had greater proximity to translating polysomes than 40S-enriched clusters (Figure 1K-L). The close association between non-translating 60S and translation hotspots suggests involvement of 60S in regulating localized translation. A potential hypothesis is that confinement of 60S to hotspots controls the formation of translating ribosomes, whereas 40S subunits diffuse more freely to bind mRNAs. The variable abundance of 40S, 60S, and assembled ribosomes may also represent subcellular variation in ribosome assembly. We hence used the local concentration of free subunits and assembled ribosomes to calculate the local ribosomal binding constant Kribo (Figure S5K). The variation in ribosomal assembly and disassembly suggests involvement of specific organelles in regulating localized translation.
One organelle that potentially regulates subcellular ribosome localization is the ER, which can be visualized by overexpressing smMyc-tagged Sec61β, allowing simultaneous visualization of Myc and smFP-tagged RPs (Figure 1G). Indeed, by quantifying the occupancy of each ribosomal population within a 4 pixel (~100 nm) range of the endoplasmic reticulum (ER) network, we observed enrichment of polysomes but not monosomes near the ER (Figure 1M). 60S is significantly enriched near the ER, while 40S is slightly depleted compared to non-ER regions (Figure 1M). Consistent with the spatial clustering, 60S subunits also tended to localize near translating polysomes at the ER, suggesting that the close proximity between 60S and translating polysomes may be caused by their association with the ER.
Motivated by this finding of 60S enrichment at the ER, we sought to further characterize ribosomes from the ER and other subcellular compartments. We developed an optically controlled substrate-specific proximity biotinylation tool to label and affinity purify ribosomes associated with subcellular compartments of interest. This tool is derived from the E. coli biotin ligase BirA, which conjugates biotin to a lysine in the 15-amino acid AviTag peptide (43). However, we sought to control the timing of biotinylation (see Supplementary Text) without using biotin depletion, which can alter cellular physiology (44). We developed a BirA that is minimally active in standard culture media by engineering a split enzyme in which the N- and C-terminal fragments of BirA are individually inactive. By fusing each fragment to enhanced Magnet (eMag) optogenetic switch domains that heterodimerize under blue light (45, 46), BirA activity is reconstituted upon illumination (Figure 2A). One or both BirA fragments can be fused to an organelle-targeting domain, enabling biotinylation of AviTagged ribosomes associated with the organelle of interest. During illumination, cells are dosed with supplemental biotin to accelerate biotinylation and cycloheximide to halt translation elongation (Figure 2B). Biotinylated ribosomes are then affinity purified on streptavidin-coated beads (Figure 2A). A tobacco etch virus (TEV) protease cleavage site proximal to the AviTag allows specific and non-denaturing elution (Figure 2A) (1, 2), yielding samples of organelle-specific ribosome populations suitable for a broad range of downstream assays. We named this split BirA system “AviTag-specific Location-restricted Illumination-enhanced Biotinylation” (ALIBi).
We AviTagged ribosomes by fusing a FLAG epitope tag, TEV cleavage site, and AviTag (FTA) to the C-terminus of the 60S protein Rpl31 homozygously at its endogenous genomic locus in mouse embryonic stem cells (mESCs) (Figure S6A). Rpl31 resides near the peptide exit tunnel, which contacts organelle-specific translational machinery. Using this Rpl31-FTA cell line, we screened candidate BirA split sites selected via free energy analysis (47) and review of previously used splits for non-AviTag-specific biotin ligases that are structurally related to BirA (48-50) (Figure S6B, C). We also screened all orientations of fusing the eMags to the N- and C- fragments (Figure S6D, E). The N73-eMag/eMag-C74 combination (Figure S6B-E) afforded the most practical fold-increase and magnitude of biotinylation upon activation (Figure 2C, Figure S6F). Neither BirA fragment biotinylated if expressed alone, and no biotinylation of non-AviTagged proteins was detected (Figure 2C). Affinity purified eluates from cells expressing cytosolic-ALIBi were compared against a no-ALIBi negative control expressing GFP via relative quantitative mass spectrometry using tandem mass tags (TMT-MS). Of the 80 core RPs, 79 were detected and all were enriched in the cytosolic-ALIBi samples (Figure S7A). Furthermore, 492 non-RPs were enriched (Figure S7A, Table S2, Table S3), including 121 previously identified RAPs (14). To control for potential AviTag-dependent artifacts, these samples were compared against wild-type mESCs expressing cytosolic-ALIBi and GFP-FTA. Few proteins were enriched in GFP-FTA samples compared to GFP samples (Figure S7B, Table S3), and a nearly identical set of RPs and RAPs were enriched when comparing cytosolic-ALIBi and GFP-FTA samples (Figure S7C, Table S3).
We next validated ALIBi for isolating subcellular populations of translational machinery by targeting ALIBi fragments to the nucleus and nucleolus (Figure S8A, B and Table S1), where many proteins are known to interact with pre-ribosomes. Compared to cytosolic-ALIBi samples, nuclear- and nucleolar-ALIBi samples were enriched for dozens of non-RPs such as Nol12, Pelp1, Rsl24d1, and Llph (Figure 2D), which are well-established nuclear or nucleolar ribosome biogenesis factors that bind pre-60S complexes (51). Others such as the RNA splicing factor Srsf5, DNA deaminase Apobec3, and peptidyl-prolyl isomerase Nktr, have no known ribosome biogenesis role, but their catalytic functions suggest that they could manipulate rRNA or RPs during ribosome biogenesis (52-54). Thus, ALIBi can isolate organelle-specific translational machinery, and has uncovered previously unknown candidate proteins for future investigation of nucleolar and nuclear ribosome biogenesis.
We used ALIBi to isolate ER-associated translational machinery by co-expressing a cytosolic N-fragment and an ER-targeted C-fragment (Figure S8). Expressing ER-ALIBi did not significantly alter whole-cell RNA levels for any genes compared to cytosolic-ALIBi (Figure S9A). RNA-seq of affinity-purified ribosomes revealed 2497 enriched RNAs in ER- versus cytosolic-ALIBi samples (Figure S9B, C and Table S4), of which 98% localize to cellular components whose proteins are co-translationally translocated at the ER (Figure S9D). Individual genes that are cytosolically translated (i.e. Gapdh) and ER-translated (i.e. Canx) exemplify this differential enrichment (Figure S9E). This dataset correlated well with previous ER transcriptome and translatome studies (2, 6) (Figure S9F, G). A known ER-specific splicing isoform of the gene Naxd (55) was ER-enriched (Figure S10A), as well as previously unreported ER-enriched isoforms of other genes (Figure S10B, C and Table S5). Thus, ALIBi successfully isolates ER-ribosomes.
We next examined differences in translational machinery composition between ER- and cytosolic ribosomes. Proteins enriched in ER-ribosomes compared to cytosolic ribosomes included Ufm1, Sec61a1, and Hspa5 (Figure 2E), which participate in ER co-translational translocation, protein folding, and translation quality control (56-58). Of note, given the detergent sensitivity of other complexes involved in ER translation such as the oligosaccharyltransferase complex and the translocon-associated protein complex (58, 59), we would not expect these to remain bound to ribosomes under these lysis conditions. Compared to the cytosolic-ALIBi sample, the ER-ALIBi sample was enriched for 60S RPs and depleted of 40S RPs (Figure 2F). We considered whether 60S might be more readily biotinylated than monosomes or polysomes near membranes for steric reasons. To test this, we targeted ALIBi to the cytosolic face of the plasma membrane (Figure S8, Table S1, Figure S11A) and found no significant 60S RP enrichment there (Figure 2F). These data suggest that a greater proportion of 60S are free rather than assembled into 80S at the ER compared to the cytoplasm, consistent with the earlier RiboExM finding that 60S is enriched at the ER. The degree of 60S RP enrichment as measured by ALIBi was smaller than the degree of 60S enrichment measured by RiboExM, which is expected because ALIBi does not distinguish between RPs present in 60S from those in monosomes or polysomes.
We also observed that four of the most significantly enriched proteins co-purifying with ER-ribosomes were the late 60S ribosome biogenesis factors eIF6, Nmd3, Znf622, and Lsg1 (Figure 2E). During 60S biogenesis, dozens of proteins sequentially associate and dissociate from pre-60S as it progresses through the nucleolus, nucleus, and cytoplasm. eIF6, Nmd3, Znf622, and Lsg1 bind to pre-60S at different steps and are collectively some of the last factors to depart as pre-60S completes maturation in the cytoplasm (Figure 2G) (51, 60). They play key roles in pre-60S nuclear export, peptidyl transferase center maturation, peptide exit tunnel maturation, and prevention of premature joining of pre-60S with 40S (61-64). It is unlikely that this apparent enrichment of eIF6, Nmd3, Znf622, and Lsg1 could be caused by a subpopulation of nuclear pre-60S among the captured ER ribosomes, since nuclear and nucleolar translational machinery (Figure 2D) contains markedly different RAPs compared to ER translational machinery (Figure S11B).
Some proteins that preferentially co-purified with ER translational machinery (Sec61a1, Hspa5, Colgalt1, Col18a1) are also encoded by ER-localized mRNAs, according to the ER-ALIBi RNA-seq described above (Figure S12A). To investigate whether the apparent enrichment of these proteins in ER translational machinery could be due to detection of ER-translated nascent peptides rather than mature ribosome-associated proteins, we used ALIBi to isolate cytosolic and ER-ribosomes after treating cells with puromycin to dissociate nascent peptides (Figure S12B) (14). Puromycylated peptides were present in puromycin-treated cell lysate, and no detectable puromycylated peptides remained in the corresponding eluate (Figure S12C). Sec61a1, Colgalt1, and Col18a1 were not detected in a sufficient number of replicates to be assessed, while Hspa5 was detected but no longer enriched after puromycin treatment. Since Hspa5 (BiP) is an ER lumenal chaperone that helps move nascent peptides through the translocon (65), its sensitivity to puromycin suggests that its ER enrichment reflects either a nascent peptide or a nascent peptide-binding protein. Meanwhile, eIF6, Nmd3, Znf622, and Lsg1 remained among the most ER-enriched proteins even after puromycin treatment (Figure S12D). Alongside the observation that the mRNAs for these proteins are not ER-localized (Figure S12A), these data corroborate our proposed model of an ER-specific population of 60S bound by mature eIF6, Nmd3, Znf622, and Lsg1.
We next investigated whether these 60S biogenesis factors play a mechanistic role in localizing 60S to the ER. Among the four factors, we opted to further characterize Lsg1, since it has been reported to play an essential role in ER structure and function (66, 67). To match the mESCs used for ALIBi, we performed RiboExM with Lsg1 immunostaining on a mouse embryo-derived fibroblast line that stably expresses RPL7a-smGFP and RPS2-smRuby2. Lsg1 displayed nuclear and cytoplasmic localization, with notable enrichment at the ER membrane and colocalization with 60S (Figure 2H-J). Approximately half of all detected cytosolic Lsg1 was associated with ER-bound 60S (Figure 2K). To further explore the role of Lsg1, we used siRNA to decrease its expression by over 50% at 48h (Figure S13A). Via RiboExM, we observed substantial depletion of 60S at the ER, similar to the effects observed after 2h treatment with thapsigargin (Tg), a potent ER stress inducer (68, 69) (Figure 3A-C). Notably, Lsg1 knockdown does not induce ER stress at either the mRNA or protein level (Figure S13A-B), suggesting that Lsg1 helps maintain 60S enrichment at the ER through an ER-stress-independent mechanism. Depletion of polysomes from the ER compared to cytoplasm was evident following Tg treatment, but was significantly less evident after Lsg1 knockdown (Figure 3D-F), indicating that 60S depletion resulting from Lsg1 knockdown is not due to global disruption of ER translation.
We next investigated whether Lsg1 has transcript-specific functions in ER translation by using ER-ALIBi to enrich for ER-ribosome-bound mRNAs after Lsg1 knockdown (Figure 3G, Figure S14A). At 24h, Lsg1 knockdown minimally affected whole-cell RNA levels of genes besides Lsg1 (Figure S14B, Table S6). In the ER-ribosome-bound mRNA fraction after Lsg1 knockdown, 48 genes increased in abundance and 29 genes decreased compared to the ER-ribosome-bound mRNA fraction treated with non-targeting siRNA (Figure 3H, Figure S14C), which may reflect changes in the translation of these mRNAs at the ER. Multi-pass membrane proteins such as the cGMP-gated calcium channel Cnga3 (70), single-pass type I/II membrane proteins such as the developmental signaling receptor Robo4 (71), and secreted or lumenal proteins such as the angiogenesis factor Ccn1 (72) were overrepresented among the increased genes compared to the unchanged genes (Figure 3I-K and Table S6). These protein topologies are usually clients of Sec61-containing translocons, the most common route for ER co-translational translocation (73, 74). Secreted proteins such as the developmental signaling ligand Lefty1 (75) and single-pass membrane proteins such as the chondroitin-modifying enzyme Csgalnact2 (76) were also overrepresented among the 29 genes with decreased ER-ribosome association upon Lsg1 knockdown, but multi-pass membrane proteins were entirely absent (Figure 3I-K and Table S6). Among the genes with the greatest decrease in ribosome association upon Lsg1 depletion is the tail-anchored membrane protein Vamp3, a rare topology (3-5% of membrane proteins) (77) in which the C-terminus is post-translationally inserted into the ER membrane by the GET (guided entry of tail-anchored proteins) complex or EMC (endoplasmic reticulum membrane protein complex) rather than the Sec61 translocon (78, 79) (Figure 3J). Western blot revealed changes in Vamp3 and Robo4 protein expression upon Lsg1 knockdown that is consistent with the change in ribosome-bound mRNA level (Figure S14D).
Although Lsg1 has a known function in ribosome biogenesis, no hallmarks of nucleolar stress (80) were observed in the whole-cell transcriptome upon Lsg1 knockdown that would suggest impaired ribosome biogenesis (Table S6). Nevertheless, we examined whether the effects of Lsg1 knockdown on ER ribosome-mRNA association may be due to a ribosome biogenesis defect. We performed siRNA-mediated knockdown of Nmd3 (Figure S15A) as an example of a 60S ribosome biogenesis factor that is also enriched at the ER (Figure 2E), and Gnl3 (81-83) as an example that is not ER-enriched (Figure S15B). Distinct sets of mRNAs were perturbed in the ER-ribosome fraction upon knockdown of Lsg1, Nmd3, and Gnl3. For example, of the 48 genes that increased in the ER-ribosome fraction upon Lsg1 knockdown, 70% were not affected by Nmd3 or Gnl3 knockdown (Figure S15F, J). This comparison suggests that most of the observed effects of Lsg1 knockdown are Lsg1-specific, revealing a mechanism of selective translational control at the ER by Lsg1 that is likely independent from its role in ribosome biogenesis.
We next targeted ALIBi to the cytosolic surface of the outer mitochondrial membrane (OMM) (Figure S8A-B, Table S1) to characterize the ribosome interactome there. Previous evidence suggests that co-translational translocation of proteins synthesized by cytosolic ribosomes occurs in yeast (84), meanwhile it remains more controversial if similar mechanisms exist in mammalian cells. Compared to cytosolic ribosomes, OMM-ribosomes were enriched in several proteins known to participate in translation at the OMM. These include Dnaja1, Dnaja2, Hspa8, and Hspa9, which are chaperone and co-chaperone proteins that assist in protein translocation across the OMM (85-88); and Pabpc1 and Larp1, poly-A binding proteins that interact with the integral OMM protein Akap1 (89, 90) (Figure 4A). Indeed, RNA binding was the most overrepresented molecular function gene ontology term among the mitochondria-enriched proteins (106 genes observed, 64 genes expected, false discovery rate = 2.19 x 10^-5^). Several of these RNA-binding proteins, such as Pum2, Nxf1, and Acin1, have been shown to regulate mitochondrial function (91-93). These findings suggest that they may interact directly with OMM-associated ribosomes or mRNA.
Several core RPs, including RPS25 and RPL29, were depleted in OMM-associated translational machinery (Figure 4A). This was intriguing in light of previous work showing that RPS25 is a substoichiometric RP, and that several mitochondria-related transcripts are depleted from RPS25-lacking ribosomes (10, 11). Although they have been reported in multiple biological contexts, specialized ribosomes have never been visualized within cells (94). MitoTracker, a mitochondria-specific fluorescent dye that is retained after the expansion microscopy process (33), was utilized to visualize mitochondria following RiboExM. To directly visualize these potential specialized ribosomes, we generated cells expressing different combinations of smFP-tagged RPs. For ribosomes lacking RPS25, we generated a cell line stably expressing tagged RPS2 and RPS25 (Figure S16A-C), both of which are 40S RPs. A considerable number of specialized ribosomes lacking tagged RPS25 and RPL29 were detected, exhibiting preferential localization toward the mitochondria (Figure 4B-E). No such enrichment was observed when a combination of two other core 40S RPs, RPS2 and RPS8, was tagged (Figure 4F). Notably, these specialized ribosomes not only exhibited enrichment in close proximity to the mitochondria but were also significantly more abundant on the mitochondrial membrane, as defined as fully colocalizing with MitoTracker staining, suggesting that they may dock on the OMM for mitochondria-specific translation (Figure 4D and E).
To achieve mRNA imaging compatible with RiboExM, we introduced 24 copies of short hairpin sequence into the 3’ untranslated region (UTR) of TagBFP and co-expressed PP7 coat protein fused with mCherry (Figure 4G) (40, 95), which binds the hairpin sequence with high affinity and transforms the mRNA signal into an mCherry signal that can be immunostained and visualized by RiboExM. After overexpressing the reporter plasmids, we observed bright and distinct mRNA signals alongside the different ribosomal populations (Figure 4H). Notably, there was almost no background PP7 signal post-expansion in the absence of the reporter mRNA (Figure S16D), indicating that these non-amplified signals did not produce substantial background that would compromise mRNA imaging. Setting the distance threshold at 2 pixels (~50 nm), quantification revealed that nearly half of the detected mRNAs were associated with assembled polysomes (Figure 4I). Meanwhile, approximately 30% of the mRNAs were not associated with any RPs, while the remaining minor fraction was bound by monosomes or single ribosomal subunits (Figure 4I).
We next combined mRNA imaging with RiboExM imaging of RPS25-lacking specialized ribosomes, focusing on a subset of transcripts implicated in mitochondria-related metabolism that were either enriched or depleted in ribosome profiling of RPS25-containing ribosomes (10). Despite a relatively low basal expression level in the cytoplasm, the majority of transcripts were found bound to ribosomes, suggesting an active translation state (Figure 4J). The vitamin B12-dependent enzyme methylmalonyl-Coenzyme A mutase (Mut) and the enzyme propionyl-Coenzyme A carboxylase (Pcc) were significantly more associated with the RPS25-lacking ribosomes, whereas the vitamin B12 transporter transcobalamin 2 (Tcn2) was primarily associated with the RPS25-containing ribosomes (Figure 4J,K). The two genes preferentially translated by the specialized ribosomes, Mut and Pcc, are directly involved in vitamin B12 utilization and may therefore require more regulated translation on the outer mitochondrial surface (10) (Figure 4L). Our results hence reveal mitochondrial enrichment of specialized ribosomes and suggest a role for ribosome heterogeneity in translational control of metabolism-related transcripts.
Thus far, RiboExM has successfully visualized 40S, 60S, monosomes, polysomes, specialized ribosomes, RAPs and mRNAs within engineered mammalian cells. We aimed to extend our investigations to neurons, where localized translation has been reported to play a critical role in polarized neuronal processes (4, 5). Given the limited division capacity of mature neurons, we engineered human embryonic stem cells (hESCs) to stably express smFP-tagged ribosomal constructs (Figure 5A) and differentiated them to yield a highly homogeneous population of induced neurons (iNs) (96) (Figure 5A, B).
Mature neurons displayed relatively similar, yet smaller cell bodies compared to fibroblast cells, where ribosomes accumulated and could be visualized at a high density (Figure 5C). In contrast, ribosomes appeared sparser in neuronal processes, despite filling the entire length of these processes, which could extend up to one millimeter (Figure 5C). Notably, there were substantially more free ribosomal subunits within the neuronal processes, especially in the middle neurites (Figure 5C, D). These subunits may represent a dormant population that can be rapidly activated for translation in the distal region far from the cell body. Accordingly, the proportion of assembled ribosomes in neuronal processes was substantially lower compared to the cell body (Figure 5D). Assembled ribosomes were also enriched in the distal region where synapses form, indicating active translation at synapses compared to middle neurites (Figure 5D). While polysomes represented the primary translation machinery in the neuronal cell body, monosomes were substantially more abundant in distal neurites (Figure 5D). The majority of these monosomes were actively translating, as demonstrated by their reduction following puromycin treatment (Figure 5E). Upon treatment with brain-derived neurotrophic factor (BDNF) or arachidonyl-2-chloroethylamide (ACEA) to enhance local translation in distal neurites (4), significantly more polysomes formed while monosomes maintained a similar proportion (Figure 5F). While it has been reported that monosomes are responsible for translating synaptic mRNAs (5), our results suggest that acutely induced neuronal translation may still involve newly formed polysomes. Monosomes, on the other hand, may be responsible for a basal level of translation distant from the cell body.
Recent studies have found that ribosome composition can be remodeled within neurons (97). We therefore investigated the localization of specialized ribosomes within neuronal processes. Specifically, we focused on RPS25-lacking ribosomes and observed their presence in neurites located distant from the cell body (Figure 5G). These specialized ribosomes exhibited lower abundance in minor neuronal branches and were primarily localized within the main neurite shaft, where mitochondria were more often observed (98, 99) (Figure 5H). These findings suggest a potential involvement of specialized ribosomes in translational regulation within distal neurites.
This study provides tools for investigating translational machinery by visualizing individual ribosomal populations and characterizing their composition in subcellular space, yielding major conclusions regarding organelle-specific translational control. An observation made here through both RiboExM and ALIBi is that free 60S subunits are enriched at the ER compared to the cytoplasm. Monosomes are typically thought to dock to the ER translocon only after a signal peptide emerges from the peptide exit tunnel. Our observation of 60S enrichment at the ER suggests mechanisms for regulating ER translation that involve 60S and associated proteins.
Our results suggest that at least one of these proteins, Lsg1, functions outside its known role in ribosome biogenesis (100). We found that Lsg1 non-uniformly impacts translation of proteins with certain membrane topologies. A recent study reported that the protein Vapa tethers Lsg1 to the ER in an orientation that would allow Lsg1 to simultaneously occupy its known binding site on the 60S intersubunit interface (62), and that this feature of Lsg1 is uncoupled from its role in ribosome biogenesis (101). We speculate that, by tethering 60S to the ER, Lsg1 positions it to promote translation of select transcripts. Such a mechanism of translational control guided by ribosome biogenesis factors acting as organelle-localized RAPs merits further investigation. It is possible that these factors have not been detected in previous characterization of ER translation because most studies have focused on 80S and sometimes deliberately filter out free 60S (22, 102), whereas Lsg1 cannot bind 80S (62). One recent work that characterized ER-localized 60S in the context of ER ribosome stalling (103) detected a subpopulation of 60S bound to Lsg1, Nmd3, and Znf622, but did not further investigate it. Evidently, detection of organelle-specific translational machinery depends not only on purification conditions but also on which ribosome subunits are investigated, highlighting the value of multiple orthogonal approaches.
Beyond uncovering the role of Lsg1 in ER translation, our survey of subcellular ribosome populations via ALIBi identified other proteins that were differentially ribosome-associated between the cytoplasm, ER, mitochondria, plasma membrane, nucleus, and nucleolus in our datasets and have no known role in organelle-specific translation or ribosome biogenesis. Some of these have structural features suggesting possible interaction with translational machinery or RNA, and may present promising future avenues of investigation.
Although a small subpopulation of ribosomes was first observed decades ago under electron microscopy as dense particles near the mitochondria (104, 105), electron microscopy cannot reveal ribosomal composition in detail, and thus direct visualization of specialized ribosomes has not previously been achieved. In this study, we found that RPS25-lacking and RPL29-lacking ribosomes are not only enriched in the mitochondria periphery but are more prevalent among ribosomes that directly contact the OMM. Whether specialized ribosomes translate different pools of mRNAs is another essential question. Being capable of simultaneously visualizing RPs and mRNAs using RiboExM, we observed a preference for these specialized ribosomes to translate two genes involved in vitamin B12 utilization. As these specialized ribosomes were also observed in distal neuronal processes, it is possible that they also play a role in translating specific neuronal transcripts and potentially regulating energy consumption far from the resources present in the soma. It will be valuable to investigate whether OMM translocons or chaperones exist, and whether these preferentially bind to specialized ribosomes.
Here, we have demonstrated two orthogonal approaches for investigating localized variations in translational machinery. Future uses of RiboExM and ALIBi may include characterizing perturbations of localized translational control in disease states. More broadly, these tools are applicable to any scientific question that calls for studying interactions of macromolecules in crowded subcellular spaces at high resolution.
U2OS cells (ATCC HTB-96), C3H/10T1/2 cells (ATCC CCL-226) and HEK293T cells (ATCC CRL-11268) were maintained in high glucose DMEM medium (Gibco 11965-092), supplemented with 10% FBS (Hyclone), and penicillin–streptomycin mix (Gibco) at 37 °C with 5% CO2. The cell lines were purchased from ATCC and used within 20 passages.
H7-hESCs were a gift from Kyle Loh’s lab (Stanford University) and were cultured in mTeSR1 media (StemCell Technologies, 85850) on plastic dishes coated with Geltrex (Gibco, A1413302). H7-hESCs were passaged 10 using Accutase (Gibco, A1110501) and cultured overnight in mTeSR1 supplemented with 2 μM thiazovivin (Tocris, 3845) to promote cell survival. Media was changed daily and cells were passaged every two days.
E14Tg2a.4 mESCs (106) were a gift from Thom Saunder’s lab (University of Michigan). Cell line identity was verified via short tandem repeat profiling performed by the American Type Culture Collection (ATCC, 137-XV, Manassas, VA). Cells were cultured in Knockout DMEM (ThermoFisher 10829-018) supplemented to a final concentration of 15% ES-qualified fetal bovine serum (MilliporeSigma ES-009-B), 1% non-essential amino acids (MilliporeSigma TMS-001-C), 2 mM L-glutamine (MilliporeSigma TMS-002-C), 1% penicillin/streptomycin (ThermoFisher 15140-122), 55 μM beta-mercaptoethanol (ThermoFisher 21985-023), and 1000 U/mL mouse leukemia inhibitory factor (mLIF, Gemini 400-495 10^7). Media was changed daily and cells were passaged every two days. Mycoplasma testing was performed according to the manufacturer’s instructions using the PromoKine PCR Mycoplasma Test Kit (PK-CA91-1096, Heidelberg, Germany).
All constructs for the RiboExM experiments were expressed using the PBX vectors allowing PiggyBac-based transposition. For constitutive expression, full-length coding sequences (CDS) were expressed under the CAG promoter (CMV enhancer with chicken β-actin minimal promoter), allowing strong expression in all cell types tested.
The sfGFP and mRuby2-based smFP sequences were generated by Dr. Loren L. Looger from UCSD and synthesized by Twist Bioscience. Ribosomal proteins (RP), Sec61β and Lsg1 sequences were cloned from the mESC genome. The plasmids with PP7 tandem stem loops and the coat protein were gifts from Xiaowei Yan from Stanford University.
To clone ALIBi constructs, enhanced Magnet domain sequences “eMagA^F^” and “eMagB^F^” (46) were synthesized on a BioXP 3200 platform (Codex DNA, San Diego, CA). A codon-optimized BirA sequence was synthesized by IDT. Epitope tags, linkers, and the NES, NLS, MyrPalm, and MAVS organelle-targeting sequences were added via PCR using oligonucleotides synthesized by IDT. The murine Fbl targeting sequence was synthesized on a BioXP 3200 and the human Sec61β targeting sequence was cloned from pJW1512, a gift from Jonathan Weissman (Addgene plasmid # 62366 ; http://n2t.net/62366; RRID:Addgene_62366) (2). ALIBi constructs were expressed under a CAG promoter using the pCAGEN backbone, a gift from Connie Cepko (Addgene plasmid # 11160 ; http://n2t.net/11160; RRID:Addgene_11160) (107).
Individual parts of constructs were first cloned using the KAPA HiFi PCR kit (Roche) and then assembled using the NEBuilder HiFi DNA assembly cloning kit (New England Biolabs).
The cell lines expressing RP-smFP were generated in U2OS, C3H/10T1/2 and hES cells. For U2OS and C3H/10T1/2 cell lines, reverse transfection of the RP-smFP constructs and the PiggyBac plasmid with Lipofectamine 3000 (Life Technologies) was carried out according to the manufacturer’s protocol. For hESC cell lines, nucleofection of the constructs using the P3 Primary Cell 4D-Nucleofector^™^ X Kit S (Lonza, V4XP-3032) was used according to the manufacturer’s protocol. In approximately 400k cells per nucleofection with 6 μg of PiggyBac plasmid mix were used per sample, using CB-150 program for nucleofection using 4D-Nucleofector (Lonza, AAF-1003X). Immediately afterwards, cells were cultured in mTeSR1 + 2 μM thiazovivin + 10 μM Y-27632 (Selleckchem, S1049) on a 6-well plate coated with Geltrex for 2 days to improve survivability. Cells were cultured and passaged continuously for 7 days prior to Fluorescence-Activated Cell Sorting (FACS) to select for population with the highest expression of RP-smFPs.
The CRISPR strategy for inserting FLAG-TEV-AviTag C-terminally at the endogenous Rpl31 locus in mESCs via exon replacement was designed as first, two guide RNA (gRNA) recognition sites were identified that flanked the exon of the target gene that contained the stop codon, using on- and off-target gRNA site scoring algorithms (108, 109) implemented in Benchling. gRNA sites were only used if they had no other high-probability predicted cut sites throughout the mouse genome. Each 20 nt gRNA sequence was cloned into a PX459 backbone (a gift from Feng Zhang; Addgene plasmid # 62988 ; http://n2t.net/62988; RRID:Addgene_62988) (110) digested with BbsI (ThermoFisher FD1014), with a single upstream G nucleotide preceding the gRNA sequence since this has been shown to improve cutting efficiency (110). To construct the homology-directed repair template, the sequence between the two gRNA cut sites was cloned along with 300 bp homology arms on each end. Immediately preceding the stop codon in the repair template, a sequence was inserted to encode the FLAG-TEV-AviTag peptide. The gRNA recognition sites or protospacer adjacent motifs (PAM) on the repair template were modified at silent coding positions or non-coding positions with low evolutionary conservation to prevent cutting of the repair template or of the repaired genomic DNA. An unmodified PAM and gRNA recognition sequence were appended to each end of the repair template, distal to the homology arms. Distal to the appended gRNA and PAM sequences on each side, 10 additional bases were appended to buffer against small deletions that occur with TOPO cloning. This construct was then inserted into a non-expressing pCR4Blunt-TOPO backbone (ThermoFisher 450031) such that the Cas9-gRNAs expressed from the PX459 plasmids would cleave the repair template as a linear dsDNA from the circular pCR4Blunt-TOPO plasmid (111).
At passage number 28, 10^6^ mESCs were transfected with the two PX459-based plasmids harboring a Cas9-puromycin fusion construct and gRNAs flanking the targeted exon (0.5 μg each), and one pCR4Blunt-TOPO-based plasmid harboring the linearizable repair template (2 ug). The three combined plasmids were diluted in 100 μL Opti-MEM Reduced Serum Medium (ThermoFisher 11058021). In parallel, 7.5 μL of Lipofectamine 2000 (ThermoFisher 11668-019) was diluted in 100 μL Opti-MEM. The plasmid/Opti-MEM and Lipofectamine/Opti-MEM were combined and incubated for 20 minutes at room temperature. Cells were trypsinized as described above, resuspended in 250 μL Opti-MEM, added to the plasmid-Lipofectamine complexes for 10 minutes at room temperature, and plated into one 12-well. Media was changed after 4 hours. At 24 hours after transfection, cells were treated with media containing 1 μg/mL puromycin (Millipore P8833). At 48 hours after transfection, fresh media containing puromycin was changed in. At 72 hours after transfection, puromycin-free media was changed in. At 96 hours after transfection, cells were washed, trypsinized, dissociated, and plated at 1000 cells per 10 cm plate to form colonies derived from single cells. At 7 days after sparse plating, individual colonies were lifted using a pipet tip, dissociated at 37 °C in 0.025% trypsin-EDTA in DPBS, replica plated in gelatinized 96-well plates in regular mESC media, and screened by genomic DNA PCR, Sanger sequencing, and Western blot for desired edits. Importantly, PCR primers were designed such that at least one primer bound distal to the 300 bp homology arms, to avoid amplifying residual repair template.
This method has been adapted from a previous publication (112). The lentiviral constructs used for generating hiN cells included TetO-Neurog2-T2A-Puro, TetO-Neurog2-T2A-Puro-P2A -eGFP and Ubi-rtTA. HEK293T cells were infected at 70% confluency on 10-cm dishes using polyethylenimine with three helper plasmids (5 μg of pRSV-REV, 5 μg of pMDLg/pRRE and 2.5 μg of pVSVg) and 10 μg of the lentiviral vector DNA. The supernatant was collected 46 h after infection and concentrated by ultracentrifugation at 21,000 rpm for 2 h at 4 °C. The lentivirus was then resuspended in the DMEM medium, aliquoted, snap-frozen in liquid N2 and stored at −80 °C for no more than six weeks before use. The production and usage of lentivirus were all performed in a dedicated biosafety level 2+ laboratory.
The CD1 mice were housed in the animal facility at the Veterinary Service Center (VSC) of Stanford University under standard housing conditions (~20–24 °C, ~50% humidity, 12-h light/12-h dark cycle). Newborn postnatal (P0) pups were euthanized by decapitation. The cortex was gently dissected from the brain and digested with papain (Worthington Biochemicals) for 30 min. The digested cells were then dissociated by harsh trituration and plated onto T75 flasks (Thermo Fisher) in high glucose DMEM supplemented with 10% FBS. The cells were passaged upon reaching confluency. The glial cells were ready to use for co-culture within 14 days after dissociation.
Prior to plating cells for iN differentiation, the cell culture plate was coated with 200 Matrigel (Corning, 356234) for at least 2 hours. On day 0, approximately 500k of H7-hESCs stably expressing smFP-tagged RPs were plated per well of a 6-well plate in 1 mL of mTeSR1 + 2 μM thiazovivin + 2 μg/mL polybrene (Tocris, 7711). 2 μL each of Ngn2-Puro and rtTA lentivirus were added per well. On day 1, after approximately 18 hours of infection, media was replaced to 2 mL of N3 media (DMEM/F12 1 (Gibco, 11320-033) + 1× N2 supplement (Gibco, 17502-048) + 1× non-essential amino-acids (Millipore, TMS-001-C), 5 mg of insulin (dissolved in 10 mM NaOH) (Sigma-Aldrich, I6634-100MG), 0.5x Pen/Strep (Gibco, 15140163)) + 2 μg/mL doxycycline (Dox) (Fisher Scientific, BP26535). On day 2 and 3, media was replaced to 2 mL of N3 media + 2 μg/mL Dox + 2 μg/mL puromycin (Sigma-Aldrich, P8833-25MG). On day 4, media was replaced to 2 mL of N3 media + 2 μg/mL Dox + 2 μg/mL puromycin + 4 μM Cytosine β-D-arabinofuranoside (AraC) (Sigma-Aldrich, C1768).
To prepare iN for imaging, on day 4 glass coverslips were first placed in a 24-well plate, sterilized by washing 3× with 70% ethanol for 5 mins each, and then UV-irradiated for 30 minutes as the ethanol evaporated. Afterwards, the coverslips were coated with 200 matrigel overnight. On day 5 morning, approximately 50k mouse glial cells from newborn wild-type CD1 mice were plated on each coverslip with NBP media (Neurobasal (Gibco, 21103-049), 1× Glutamax (Gibco, 35050061), 1× Gem21 NeuroPlex^™^ Serum-Free Supplement (Gemini, 400-160), 0.5× Pen/Strep (Gibco, 15140163), 5% iN grade serum - Cytiva HyClone^™^ (Cytiva, SH30396.03)). On day 5 evening, premature iN cells were detached gently with Accutase, and approximately 150k cells per well were re-plated with 800 μL NBP media + 2 μg/mL doxycycline. On day 7, wash gently with pre-warmed NB zero media (NBP without 5% serum) to remove any dead cells, and replace media with 800 μL NB-2% (NBP with 2% serum instead) + 2 μg/mL Dox + 4 μM AraC. On day 10, partially replace spent media with 300 μL NB-2% + 2 μg/mL Dox + 4 μM AraC. On day 14, stop adding Dox and AraC, and partially replace spent media with 300 μL NB-2%. Cells were then prepared for imaging on day 20.
All chemicals used were purchased from Sigma-Aldrich and used without further purification, unless otherwise stated. If ER staining was required, the cells were first transfected with a plasmid encoding Sec61β-smMyc for 2 days. If mitochondria staining was required, the cells were first incubated with 400 nM Mitotracker Deep Red FM (Thermo, M22426) for 45 min. Coverslips with adhered cells were later rinsed with 1× PBS and then fixed with fixation solution (3% PFA+0.1% GA in 1× PBS) at room temperature (RT) for 10 min. Without permeabilization, coverslips with cells were washed with 1× PBS for 5 min. Subsequently, they were transferred to the anchoring buffer (0.7% PFA+1% AA in 1× PBS), incubated at 37 °C for 5 h. Coverslips with cells were then washed with 1× PBS for 10 min before embedding into gel.
After the PBS wash, 100 μL of monomer solution without accelerator and initiator (1.1 M sodium acrylate, 2.0 M acrylamide, 30 ppm N,N’-methylenebisacrylamide (Bis-AA) in 1× PBS) was first added onto coverslips and incubated for 15 min (32). Meanwhile, a gelation chamber for a thin gel was prepared using two strips of #1.5 coverslips as spacers along with one glass slide. The monomer solution was carefully aspirated from the coverslips, and the coverslips were flipped and placed between the two strips of #1.5 coverslips to create the gelation chamber as previously described (113). TEMED, then APS (both final concentrations of 0.15% (w/w)) were mixed with a new tube of monomer solution on ice. Approximately 30-50 μL of this mixture was added through the gap between the glass slide and the coverslip with cells, taking care to avoid the formation of bubbles. The gelation chamber was then incubated in a humidity chamber at 37 °C for 1 hour to allow hydrogel formation.
Following gelation, the gelation chamber was disassembled using a clean razor blade, and the coverslips with cell-embedded gels were transferred into the denaturing buffer (200 mM SDS, 200 mM NaCl, and 50 mM Tris in ddH2O water, pH=9.0) and incubated for 18 h at 70 °C in a water bath, followed with an incubation for 1 h at 95°C. Once the solution cooled to room temperature, the denaturing buffer was replaced with 1× PBS containing 1% Triton X-100. The gels were then incubated for 30 min at room temperature with rotation, and this process was repeated three times to remove residual SDS. By this time, the gels should have gone through a first round of expansion in 1x PBS.
The gels were then incubated in 1× blocking medium (0.5% BSA, 0.05% (wt/vol) NaN3 in 1× PBS) for 30 min. Meanwhile, the antibody cocktail was prepared by diluting antibodies into antibody diluent (3% normal donkey serum with 0.5% BSA in 1× TBS IHC wash buffer with Tween 20 (Cell Marque, 935B-09)). The cell-embedded gels were then incubated in a humidity chamber with primary antibody cocktails at 37 °C for 18 h while rotating at 25 rpm. After that, the gel was washed three times with the washing buffer (0.5% BSA, 0.05% (wt/vol) NaN3 in 1× PBS) for 30 min each wash at 37 °C with rotation. Immunostaining was conducted with the following primary mouse anti-V5 (Invitrogen, R960-25, RRID:AB_2556564), rat anti-HA (Roche, 3F10, RRID:AB_2314622), rabbit anti-Myc (CST, 2278, RRID:AB_490778), and rabbit anti-mCherry (Abcam, ab167453, RRID:AB_2571870).
In the meantime, a secondary antibody cocktail was prepared by diluting antibodies into the antibody diluent. Immunostaining was conducted with select combinations from the following secondary goat anti-mouse IgG Alexa Fluor 488, anti-mouse IgG Alexa Fluor 546, anti-rat IgG Alexa Fluor 488, anti-rat IgG Alexa Fluor 546, and anti-rabbit IgG Alexa Fluor 647 (Thermo Fisher Scientific, A-11001, RRID:AB_2534069; A-11030, RRID:AB_2737024; A-11006, RRID:AB_2534074; A-11081, RRID:AB_2534125; A-21245, RRID:AB_2535813). The cell-embedded gels were incubated with the secondary antibody cocktail at 37 °C for at least 3 hours with rotation at 25 rpm. After the incubation, the gel was washed three times with the washing buffer. Each wash lasted for 30 min at 37 °C with rotation. The gels were stained with DAPI in the washing buffer during the second wash.
Immediately following the immunofluorescence staining, the cell-embedded gels were transferred into at least 5 mL of ddH2O to undergo a second round of expansion (approximately an additional two-fold expansion) for 30 min. The water was then exchanged until the gel no longer changed in size, typically requiring three exchanges. Fully expanded gels were then cut to approximately 15 mm × 15 mm regions and packed into 35mm FluoroDish glass bottom dishes (World Precision Instruments) for examination under a confocal microscope. A glass coverslip was placed on top of the gel and ddH2O was filled beneath to prevent gel dehydration.
Images were acquired using a custom-built inverted spinning disk microscope as described previously unless otherwise stated (112). A Zeiss Axio Observer Z1 microscope was coupled to the Perkin Elmer UltraVIEW Vox spinning disk confocal microscopy system with an encoded ASI MS2000 motorized piezo stage equipped with a Plan-Apochromat ×63/1.20 water immersion objective (Carl Zeiss microscopy). The UltraVIEW Vox system utilized 405-nm (for BFP), 488-nm (for sfGFP), 561-nm (for mKate and mScarlet) and 640-nm (for iRFP) solid-state laser lines paired with emission filters (Semrock) that were specifically selected to minimize crosstalk across various fluorescent proteins (452W25 for BFP and DAPI, 510W20 for sfGFP, 588W21 for mScarlet, 632W22 for mKate and 692LWP for iRFP). Images were captured in the gain mode of a back-thinned electron-multiplying charge-coupled device camera (Hamamatsu ImageEM C91003). Acquisition of the images was accomplished with the Volocity Acquisition suite 6.3 (Perkin Elmer) for multichannel time-lapse confocal recordings. The laser power and exposure time were adjusted to minimize phototoxicity during acquisition. The laser power measured at the entrance of the spinning disk was less than 2.5 mW for all excitation wavelengths.
STED microscopy was performed on a Leica TCS SP8 3xGated STED microscope using an ×100/1.20 water immersion objective. Standard confocal microscopy using a 488 nm laser was performed with the same microscope. A 592 nm depletion laser was used at 50-100% power for STED. 2D Images were acquired with line averaging. STED images and confocal images are then aligned in FIJI for comparison.
All scale bars presented in this work are calibrated and corrected based on the expansion factor, unless otherwise stated.
Acquired images and movies were processed using the Volocity 6.3 ‘Visualization and Quantification’ Suite (Perkin Elmer) unless otherwise stated. Three-dimensional volume images were stacked and displayed using maximum intensity projection if required.
Custom Python codes were used for analyzing acquired RiboExM images. In brief, the images were first background-corrected and deconvolved. Unless otherwise stated, all images were analyzed and presented in 2D as individual z-sections (z-stacked images may be used for visualization purposes). We then applied local adjusted thresholding to these pre-processed images to generate binary ribosomal, organellar or other protein and mRNA masks. Colocalization analysis was later conducted on these masks on a punctum-to-punctum basis. For each masked punctum, a disk-shaped periphery was created based on a fixed distance threshold as stated in the text or figure legends. The colocalization of these peripheries was used to indicate the large and small ribosomal subunits in proximity indicated assembled ribosomes, while the non-colocalized puncta represented free ribosomal subunits. The total fluorescence intensity within each non-colocalized puncta was calculated to represent the baseline of single RPs. Monosomes (smaller or equal to the baseline), light polysomes (2-5 times the baseline) and heavy polysomes (larger than 5 times the baseline) were distinguished and annotated based on the total fluorescence intensity within each colocalized puncta. The locations of the 40S, 60S, monosomes, light polysomes and heavy polysomes were then assembled into a single composite ribosomal map. Similar strategies were applied for analyzing ribosome heterogeneity and the proximity of ribosomes to organelles, proteins, or mRNAs. For displayed images, individual protein and mRNA puncta were dilated by a disk with a radius of 1 pixel to enhance visualization only. Total ribosome number was estimated based on captured z-stacked (from top of the cell to the bottom) and stitched whole cell images, where ribosome numbers from individual z sections were used for the estimation of the total detected ribosome number.
Unsupervised KNN spatial clustering (K = 8, 10, 12, and 16) was conducted based on local concentration of the 40S, 60S, monsomes, and polysomes within a 100-pixel × 100-pixel sliding window. A matrix with the row representing the concentration, and the column representing the location of the sliding window was created and used for KNN clustering.
For the ribosomal assembly kinetics analysis in Figure S5K, the binding constant Kribo is estimated Kribo=[80S]+L∗[Lightpolysome]+H∗[Heavypolysome][40S][60S] , where L and H represent the average number of ribosomes per light or heavy polysome puncta.
The parameters used in these analyses were first manually finetuned and obtained from control experiments serving as the ground truth.
Cells were first seeded onto coverslip, then transfected with Sec61β-smMyc plasmid by TransIT-LT1 reagent (Mirus Bio, MIR 2304) for 48 hours, then before harvesting and fixation, the cells were stained with 200 nM Mitotracker Red FM (Thermo, M7512) for 45 min, then washed and proceeded as standard RiboExM protocol stated above with only difference that, to acquire the pre-expansion images, the cells were stained with antibodies both before and after RiboExM. Rat anti-Myc (Abcam, ab206486, RRID:AB_2861226) and anti-rat IgG-Alexa488 (Thermo, A21208, RRID:AB_141709) were used as primary and secondary antibodies with 100 and 500 dilution, along with 0.2 μg/mL DAPI (Thermo, 62247). Fluorescent images were acquired with an Andor Dragonfly 200 spinning disk confocal microscope equipped with Nikon Plan Apo 60×/1.4 oil and Plan Apo 10×/0.45 objectives, Fusion 2.4.0.14 software was used to perform imaging and stitching.
Pairing the same cells on coverslips before gelation and after the RiboExM process is essential to validate the isotropy during the RiboExM process. To facilitate this, specific asymmetric patterns were created by selectively removing cells from the coverslip prior to gelation. A large area surrounding these patterns was then imaged using a 60× objective and documented as a reference map for the next steps. Following the RiboExM process, the fully expanded gel derived from the coverslip was first imaged at lower magnification. Matching the arrangement of neighboring cells made it possible to locate the same cells. These cells were then re-imaged with a 60× objective for downstream analysis. After image acquisitions, the matched confocal z-stacks were first z-projected by max intensity, then removed the background (“sliding paraboloid” on, radius = 1000 pixels) by built-in function of ImageJ.
two strategies were used on three different for ER (Myc-staining), Mitochondria (mitotracker) and Nucleus (DAPI), roughly 10 distance measurements between landmarks were taken by ImageJ tool from 4 or 7 images each, the distance pairs were individually divided to give a population of expansion fold; Additionally, for DAPI nucleus staining, two large areas containing ~400 cells each were roughly matched first, the nucleus were segmented with a deep-learning based algorithm to estimate the nucleus size and the expansion ratio (114). As the optical resolution of the microscope is constant, the imaging resolution is calculated by dividing the imaging resolution of the confocal microscope by the expansion ratio.
the processed FOVs were then first roughly aligned manually, then went through the isotropy analysis pipeline as previously reported (115). Briefly, a rigid registration was firstly performed to correct for minor lateral mismatch, then, a non-rigid registration (B-Spline) were performed between the resulting image and the matched FOV, with the up-sampled pre-expansion images as the moving images, and post-expansion image as fixed images, to compensate the sparsity of post-expansion images, Gaussian blur with radius = 4 pixels were performed before matching. The distortion detected through this process was used to calculate the root mean square error (RMSE) to showcase the isotropy preservation during expansion.
10T1/2 cells were cultured in 15-cm dishes until reaching approximately 70-80% confluency. 45 minutes before harvesting, the culture medium was replaced with fresh DMEM (Gibco, 11965118). The cell harvest and lysis protocol was adapted and modified from a previous study (116). Specifically, after the media change, cells were treated with 100 μg/mL cycloheximide (Sigma Aldrich, C7698-1G) for 30 minutes to inhibit translation. Following treatment, cells were washed with 10 mL of pre-warmed DPBS (Gibco, 14190-250) containing 100 μg/mL cycloheximide and subsequently dissociated with 5 mL of 0.05% trypsin-EDTA (Gibco, 15400-054) containing 100 μg/mL cycloheximide. The cells were pelleted by centrifugation at 1,000 rpm for 3 minutes at room temperature. Per one 15-cm dish cell pellet, 300μl ice-cold lysis buffer (20 mM Tris-HCl, pH 7.5, 10 mM KCl, 5 mM MgCl2, 100 μg/mL cycloheximide, 1 mM DTT (Pierce, A39255), 5 mM putrescine (Fisher Scientific, 50213628), 200 U/mL SUPERase.In^™^ RNase inhibitor (Ambion, AM2696), and 1× Halt^™^ protease and phosphatase inhibitor (Thermo Scientific, 78443) in nuclease-free water) was used to resuspend and incubated on ice for 5 minutes. NP-40 (0.5% final concentration), sodium deoxycholate (0.5% final concentration), and 20 U/mL TURBO DNase (Ambion, AM2238) were added to the lysate, following by rocking at 4 °C for 20 minutes. Cleared lysates were obtained by sequential centrifugation at 800 × g for 5 minutes (twice), 8,000 × g for 5 minutes, and 21,300 × g for 5 minutes. RNA concentrations were measured using Nanodrop.
The lysates were then layered onto linear 10%-45% (w/v) sucrose gradients prepared in gradient buffer (20 mM Tris-HCl, pH 7.5, 15 mM MgCl2, 150 mM NaCl, 1 mM DTT, 100 μg/mL cycloheximide in nuclease-free water) and centrifuged using a Beckman SW41 rotor at 40,000 rpm for 2.5 hours at 4°C. The sucrose gradients were fractionated using an automated density gradient fractionation system, with continuous A260 monitoring. The resulting fractions were precipitated using the ProteoExtract^™^ Protein Precipitation Kit (Millipore, 539180).
Protein precipitates were resuspended in 60 μL of 2× Laemmli buffer (Fisher scientific, 50-196-784), boiled at 95 °C in a Thermomixer at 1,000 rpm for 5 minutes, and then loaded in equal volumes onto a 4-20% Criterion^™^ TGX^™^ Precast Midi Protein Gel (Bio-Rad, 5671094). Electrophoresis was performed at 120 V until adequate separation was achieved. Proteins were transferred to PVDF membranes (Bio-Rad, 1704273) using a Trans-Blot Turbo^™^ Transfer System (Bio-rad, 1704150). Membranes were blocked in 5% milk in 1× PBST (0.1% Tween-20 in PBS (Fisher Scientific, BP2944100)) at room temperature for 1 hour, followed by three 10-minute washes in 1× PBST. The membranes were incubated overnight at 4°C with primary antibodies. Ribosomal protein (RP) primary antibodies (Rps2 (Proteintech, 15562-1-AP, RRID:AB_3085465), Rps5 (Abcam, ab168823, RRID:AB_3668637), Rpl29 (Proteintech, 15799-1-AP, RRID:AB_2878187)) were used at a 1000 dilution, while HA (Roche, 3F10, RRID:AB_2314622) and V5 (Invitrogen, R960-25, RRID:AB_2556564) primary antibodies were used at a 20000 dilution in 5% BSA with 0.02% NaN3 in PBST. Following three 10-minute washes in 1× PBST, the membranes were incubated with HRP-conjugated secondary antibodies(Anti-rabbit (Cytiva, NA934V, RRID:AB_772206), Anti-mouse (Cytiva, NA931V, RRID:AB_772210), Anti-Rat (Santa Cruz Biotechnology, SC-2006, RRID:AB_1125219)): anti-RP secondary antibodies at a 10000 dilution except for Rps2 which were used at a 5000 dilution, and anti-HA/V5 secondary antibodies at a 20000 dilution in 5% milk in 1× PBST, for 1 hour at room temperature. Finally, the membranes were washed three times for 10 minutes in 1× PBST and developed using Clarity^™^ Western ECL Substrate (Bio-Rad, 1705061).
The cell pellet collected from cultured cells was lysed in lysis buffer (20 mM HEPES-KOH, pH = 7.6, 15 mM Mg(OAc)2, 60 mM NH4Cl, 1 mM DTT, 100 μg/ml cycloheximide, 1% Triton X-100, 0.5% deoxycholate, 8% glycerol, 0.02 U/μl Turbo DNase (Ambion, AM2238), 0.2 U/μl SUPERase Inhibitor (Ambion, AM2696), 1x Combined Protease and Phosphatase Inhibitor (Thermo Scientific, 78443)) at 4 °C for 30 min with occasional vortexing. Cell lysate was then cleared by a series of 800g x 5 min, 800g x 5 min, 8000g x 5 min and 21,300g x 5 min. Cleared cell lysate was loaded onto a 10% - 50% sucrose gradient (20 mM Tris pH 7.5, 100 mM NaCl,15 mM MgCl2, 100 μg/ml cycloheximide, made on a Biocomp Model 108 Gradient Master) and centrifuged in a Beckman SW41 rotor at 40,000 rpm at 4 °C for 2.5 h. Gradients were further fractionated with continuous A260 measurements (Brandel, BR-188).
For ExM purpose, collected gradients were diluted in lysis buffer and applied onto glass coverslips to dry at RT before proceeding directly for ExM. Polysome fractions were pooled from multiple fractions if required.
Western blot lysis 25 mM Tris pH 7.5 (ThermoFisher 15567027), 150 mM NaCl (ThermoFisher AM9760G), 15 mM MgCl2 (ThermoFisher AM9530G), 1% v/v Triton-X 100 (MilliporeSigma X100-500ML), 8% v/v glycerol (MilliporeSigma G6279), 1 mM DTT (MilliporeSigma 43815), 0.5% w/v sodium deoxycholate (MilliporeSigma S1827), 1X Halt EDTA-Free Protease Inhibitor (ThermoFisher 78425), 0.02 U/μL Turbo DNAse (ThermoFisher AM2239), 0.2 U/μL Superase RNAse Inhibitor (ThermoFisher AM2696) in Ultrapure distilled water (ThermoFisher 10977-015).
Cells were washed, trypsinized, and neutralized with media as described above. The cells were pelleted at 200 g for 3 min at 4 °C, washed with ice-cold DPBS, pelleted again, and washed again in DPBS. Western blot lysis buffer (above) was added. Cells and the lysis buffer were vortexed for 30 seconds and rested on ice for 30 seconds for three cycles, then incubated at 4 °C for 15 minutes. Lysates were clarified by centrifuging at 7000 g for 5 minutes at 4 °C. Unless otherwise stated, total protein in each sample was quantified by bicinchoninic acid assay (ThermoFisher 23225) per manufacturer protocol and samples were normalized to equal total protein. Samples were resolved on a 4%–20% Tris-glycine gradient SDS-PAGE gel (Bio-Rad 5671095) and transferred to a polyvinylidene difluoride membrane (Bio-Rad 1704273).
For streptavidin probing, membranes were blocked overnight at 4 °C in 5% bovine serum albumin (BSA, Fisher BP1600-100) in phosphate-buffered saline (PBS) (Fisher BP2944100) with 0.1% Tween-20 (MilliporeSigma P9416) (PBST). Membranes were then incubated at room temperature for 10 minutes with 5000 streptavidin-HRP (ThermoFisher, N100) in 5% BSA in PBST. Membranes were washed 3 times for 10 min each in 5% BSA in PBST, then 3 times for 10 min each in PBST before detection using Clarity Western ECL Substrate (Bio-Rad 170-5061) and imaging on a ChemiDoc MP (Bio-Rad 17001402).
For other probes, membranes were blocked for 1 hour at room temperature or overnight at 4 °C with 5% milk in PBST. Membranes were incubated for 16 h at 4 °C with the following antibodies in 5% BSA in PBST with 0.02% sodium anti-V5 (1:5000, Thermo R960-25, RRID:AB_2556564), anti-FLAG (1:1000, MilliporeSigma F3165, RRID:AB_259529), anti-B-actin (1:1000, Cell Signaling 3700S, RRID:AB_2242334), anti-GAPDH (1:5000, ThermoFisher AM4300, RRID:AB_437392), anti-puromycin (1:10000, EMD Millipore MABE343-AF488, RRID:AB_2736875). Membranes were washed 3 times for 10 min in PBST before incubation for 30 min at room temperature with secondary antibodies coupled to horseradish peroxidase at 10000 dilution in 5% milk/PBST: donkey anti-mouse (Cytiva NA931V, RRID:AB_772210). Membranes were washed 3 times for 10 min in PBST before detection using Clarity Western ECL Substrate or SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher 34095) and imaging on a ChemiDoc MP. Band intensity was quantified in ImageLab 6.1 (Bio-Rad) with manual lane detection and automatic band detection. For quantification of biotinylation by ALIBi compared to whole BirA, a two-fold serial dilution of each sample was used to generate a standard curve via linear regression in Microsoft Excel, and the relative band intensity was interpolated.
For western blots in Figure S13 and S14, a modified RIPA lysis buffer was 20 mM Tris pH 8.0 (ThermoFisher AM9855G), 15 mM MgCl2 (ThermoFisher AM9530G), 150 mM NaCl (ThermoFisher AM9760G), 1% Triton X-100 (MilliporeSigma X100), 0.5% sodium deoxycholate (MilliporeSigma S1827), 0.1% SDS (Promega V6551), 1 mM DTT (MilliporeSigma 43815), 1X HALT Protease + Phosphatase inhibitor (ThermoFisher 78443), 0.125 U/μL universal nuclease (Pierce 88700) in Ultrapure water (ThermoFisher 10977-015). Cells were washed twice with cold DPBS and 500 μL chilled modified RIPA buffer was added per 6-well. The plates were rocked at 4 °C for 15 minutes. Lysate was pipetted six times per well with a P1000 to dislodge adherent cells and transferred to a 1.5 mL microcentrifuge tube. Tubes were centrifuged at 17,000 rpm for 3 minutes at 4 °C. Supernatant was transferred to a fresh microcentrifuge tube. Western blot was performed as described in the preceding section, with the following anti-Lsg1 (1:1000, ThermoFisher PA5-58069, RRID:AB_2643577), anti-Chop (1:1000, Cell Signaling Technology 2895, RRID:AB_2089254), anti-Atf4 (1:1000, Cell Signaling Technology 11815, RRID:AB_2616025), anti-Actb (1:1000, Cell Signaling Technology 3700, RRID:AB_2242334), anti-Nmd3 (1:1000, ProteinTech 16060-1-AP, RRID:AB_2282830), anti-Gnl3 (1:1000, ProteinTech 67169-1-Ig, RRID:AB_2882465), anti-Vamp3 (1:2000, ProteinTech 10702-1-AP, RRID:AB_2212628), anti-Robo4 (1:1000, ProteinTech 20221-1-AP, RRID:AB_10665944), anti-Atp1a1 (1:20,000, ProteinTech 14418-1-AP, RRID:AB_2227873), and anti-GAPDH (1:5000, ThermoFisher AM4300, RRID:AB_437392). Band intensity was quantified in ImageLab 6.1 (Bio-Rad) with manual lane detection and automatic band detection. Data are represented as mean ± standard error of the mean, with dots representing technical replicates. Statistical differences between comparison groups were evaluated using unpaired t-tests (GraphPad Prism v10; GraphPad Software).
All steps were performed in glass containers pre-rinsed in methanol or plastic tissue-culture plates, using HPLC-grade reagents. After gel electrophoresis performed as described above for western blots, the gel was incubated on an orbital shaker for 30 minutes in 100 mL fix solution (50 mL methanol, 7 mL glacial acetic acid, adjusted with water to a total volume of 100 mL) twice. It was then incubated in SYPRO Ruby stain (MilliporeSigma S4942) overnight on an orbital shaker, covered in foil. The gel was then washed for 30 minutes on an orbital shaker in 100 mL wash solution (10 mL methanol, 7 mL glacial acetic acid, adjusted with HPLC-grade water to a total volume of 100 mL), rinsed for 5 minutes three times in water, and imaged on a ChemiDoc MP.
Permeabilization buffer (cytoplasmic): 0.2% Triton X-100 (MilliporeSigma X100), 10% goat serum (Fisher ICN2939249) in DPBS (ThermoFisher 14190-250).
Permeabilization buffer (ER): 0.1% Saponin (MilliporeSigma 47036-50G-F), 10% goat serum (Fisher ICN2939249) in DPBS (ThermoFisher 14190-250).
Blocking 10% goat serum (Fisher ICN2939249) in DPBS (ThermoFisher 14190-250).
Mouse 10T1/2 fibroblasts were pre-seeded at 100,000 cells per 12-well onto glass coverslips that had been pre-coated overnight with 0.1% gelatin (MilliporeSigma ES-006-B). The media consisted of DMEM (ThermoFisher 11965-118) with 10% FBS (MilliporeSigma TMS-013-B) and 1% penicillin/streptomycin (ThermoFisher 15140-122). After 24 h of reattachment, each well was transfected with one plasmid encoding an ALIBi N- or C-fragment: 500 ng of plasmid was diluted in 100 μL Opti-MEM Reduced Serum Medium (ThermoFisher 11058021). 1.5 μL of Lipofectamine 2000 (ThermoFisher 11668-019) was added to 100 μL Opti-MEM. The plasmid/Opti-MEM and Lipofectamine/Opti-MEM were combined and incubated for 20 minutes at room temperature. Media was aspirated from each well and 200 μL plasmid/Lipofectamine/Opti-MEM mix was pipetted onto the cells. The cells were incubated at room temperature for 1 min in the transfection mix, then 1 mL of media without antibiotics was added per well. After 48 h, coverslips were washed twice with 1 mL of ice-cold DPBS (ThermoFisher 14190-250) and fixed for 10 min at room temperature in 4% paraformaldehyde in PBS (dilution of Fisher 43368-9M in 1.33X PBS (Fisher BP2944100)). After fixation, coverslips were washed again as above, and incubated in the respective permeabilization buffer (see above) for 10 min at room temperature. Coverslips were washed again as above, blocked for 1 h at room temperature, incubated with primary antibodies overnight at 4 °C in blocking buffer, washed three times as above, incubated with secondary antibody in blocking buffer for 1 h at room temperature in the dark, washed once, stained with 1000 DAPI (Thermo Fisher 62248) in PBS at room temperature for 10 min in the dark, washed twice, and mounted onto glass slides for confocal imaging. The following primary antibodies were anti-V5 (1:500, ThermoFisher R960-25, RRID:AB_2556564), anti-calnexin (1:250, Cell Signaling 2679S, RRID:AB_2228381). The following secondary antibodies were anti-mouse AF568 (1:500, ThermoFisher A9 A10037, RRID:AB_11180865), anti-rabbit AF647 (1:500, ThermoFisher A21244, RRID:AB_2535812).
20 mM Tris pH 7.5 (ThermoFisher 15567027), 15 mM MgCl2 (ThermoFisher AM9530G), 150 mM NaCl (ThermoFisher AM9760G), 0.5 mM DTT (MilliporeSigma 43815), 2% n-dodecyl b-d-maltoside (MilliporeSigma W53094), 100 μg/mL cycloheximide, 1X Halt EDTA-Free Protease Inhibitor (ThermoFisher 78425), 0.1 U/μL Superase RNAse Inhibitor (ThermoFisher AM2696), 0.014 U/μL Turbo DNAse (ThermoFisher AM2239), 1 U/mL apyrase (New England Biolabs M0398L) in Ultrapure distilled water (ThermoFisher 10977-015).
20 mM Tris pH 7.5 (ThermoFisher 15567027), 15 mM MgCl2 (ThermoFisher AM9530G), 150 mM NaCl (ThermoFisher AM9760G), 0.5 mM DTT (MilliporeSigma 43815), 1% v/v Triton-X 100 (MilliporeSigma X100-500ML), 0.5% w/v deoxycholate (MilliporeSigma S1827), 5% v/v glycerol (MilliporeSigma G6279), 100 μg/mL cycloheximide, 1X Halt EDTA-Free Protease Inhibitor (ThermoFisher 78425), 0.1 U/μL Superase RNAse Inhibitor (ThermoFisher AM2696), 0.01 U/μL Turbo DNAse (ThermoFisher AM2239), 1 U/mL apyrase (New England Biolabs M0398L) in Ultrapure distilled water (ThermoFisher 10977-015).
100 mM NaOH (MilliporeSigma 72068), 50 mM NaCl (ThermoFisher AM9760G) in Ultrapure distilled water (ThermoFisher 10977-015).
100 mM NaCl (ThermoFisher AM9760G) in Ultrapure distilled water (ThermoFisher 10977-015).
20 mM Tris pH 7.5 (ThermoFisher 15567027), 15 mM MgCl2 (ThermoFisher AM9530G), 150 mM NaCl (ThermoFisher AM9760G), 0.5 mM DTT (MilliporeSigma 43815), 0.05% n-dodecyl b-d-maltoside (MilliporeSigma W53094), 100 μg/mL cycloheximide in Ultrapure distilled water (ThermoFisher 10977-015).
20 mM Tris pH 7.5 (ThermoFisher 15567027), 15 mM MgCl2 (ThermoFisher AM9530G), 150 mM NaCl (ThermoFisher AM9760G), 0.5 mM DTT (MilliporeSigma 43815), 0.1% v/v Triton-X 100 (MilliporeSigma X100-500ML), 5% v/v glycerol (MilliporeSigma G6279), 100 μg/mL cycloheximide in Ultrapure distilled water (ThermoFisher 10977-015).
20 mM Tris pH 7.5 (ThermoFisher 15567027), 15 mM MgCl2 (ThermoFisher AM9530G), 300 mM NaCl (ThermoFisher AM9760G), 0.5 mM DTT (MilliporeSigma 43815), 0.05% n-dodecyl b-d-maltoside (MilliporeSigma W53094), 100 μg/mL cycloheximide in Ultrapure distilled water (ThermoFisher 10977-015).
20 mM Tris pH 7.5 (ThermoFisher 15567027), 15 mM MgCl2 (ThermoFisher AM9530G), 300 mM NaCl (ThermoFisher AM9760G), 0.5 mM DTT (MilliporeSigma 43815), 0.1% v/v Triton-X 100 (MilliporeSigma X100-500ML), 5% v/v glycerol (MilliporeSigma G6279), 100 μg/mL cycloheximide in Ultrapure distilled water (ThermoFisher 10977-015).
1X AcTEV buffer, 1 mM DTT, 1% v/v AcTEV protease in Ultrapure distilled water (ThermoFisher 10977-015).
once transfected, cells were handled in a dark room with lights from the ceiling, biosafety cabinets, windows, and equipment turned off or obscured. Red headlamps were worn to minimally illuminate the workspace. The time spent handling plates outside the 37 °C incubator was minimized, all media was warmed thoroughly to 37 °C before use, and plates were carried on a styrofoam plate or warmed gel pack to minimize cooling. Lysis, centrifugation, pulldown, and wash steps were carried out in a dark room with samples and buffers chilled to 4 C. Protein low-binding microcentrifuge tubes were used unless otherwise specified.
For experiments comparing cytosolic, nuclear, and nucleolar ALIBi samples by mass Rpl31-FTA mESCs were transfected with plasmids encoding the appropriate N- and C-terminal ALIBi fragments. 3.75 μg of each ALIBi plasmid was diluted in 250 μL Opti-MEM (ThermoFisher 11058021). For the GFP no-ALIBi control condition, 7.5 μg of a plasmid encoding V5-GFP was used instead. For the GFP-FTA no-ALIBi control condition, wild-type mESCs were used instead of Rpl31-FTA mESCs, and an additional 0.16 μg of plasmid encoding GFP-FTA was co-transfected with the cytosolic N- and C-terminal ALIBi fragments. In parallel, 18.75 μL of Lipofectamine 2000 (ThermoFisher 11668-019) was diluted in 250 μL Opti-MEM. The plasmid/Opti-MEM and Lipofectamine/Opti-MEM mixes were combined and incubated for 20 minutes at room temperature. For each condition, 7.5 million trypsinized cells were resuspended in 750 μL Opti-MEM, mixed with the plasmid-Lipofectamine complexes for 10 minutes at room temperature, and plated into one 10 cm plate containing 7.5 mL warmed media. Plates were incubated at 37 °C for 4 hours to allow reattachment before changing to fresh media. Media was changed 24 hours after transfection. At 48 hours after transfection, cells were changed to media containing 100 μg/mL cycloheximide (CHX, MilliporeSigma C7698) and 50 μM biotin (ThermoFisher B20656). For the experiment in which cells were treated with puromycin, cells were changed to media containing 100 μg/mL puromycin (ThermoFisher J67236.8EQ) and 50 μM biotin instead. Plates were immediately placed on a blue light transilluminator (Clare Research Chemical DR46B or DR196) inside a 37 °C incubator for 15 minutes. Plates were washed twice with ice-cold DPBS containing 100 μg/mL CHX, before adding 2 mL ice-cold lysis buffer. After incubating plates with lysis buffer on ice for 5 minutes, lysate was pipetted up and down six times in the plate with a P1000 pipet to fully dislodge cells and transferred to a 5 mL microcentrifuge tube. Each sample was sonicated for 45 seconds with a Microson Ultrasonic Cell Disruptor XL (Misonix XL-2000) at 4 °C on a power setting of 3 watts, then transferred to a 2 mL microcentrifuge tube. The lysate was further incubated for 15 minutes on ice, then clarified by centrifuging at 17,000 g for 3 minutes at 4 C. A 1200 μL aliquot of the clarified lysate was added to pre-washed Dynabeads MyOne Streptavidin C1 beads (ThermoFisher 65002) for affinity purification. For each sample, a 450 μL aliquot of beads was pre-washed as the storage buffer was removed and the beads were washed twice with a double bead volume of RNAse removal buffer, once with a double bead volume of salt wash, and twice with a double bead volume of Wash I. After adding clarified lysate, beads were rotated at 4 °C for 1 h, then washed twice with a double bead volume of Wash I and once with a double bead volume of Wash II. The beads were then rotated in one bead volume of elution buffer at room temperature for 1 h.
For experiments comparing cytosolic, ER, plasma membrane, and mitochondrial ALIBi samples by mass spectrometry, the above procedure was performed without the sonication step.
For experiments comparing cytosolic and ER ALIBi by RNA-seq, a similar procedure was performed at 6-well scale, using buffers marked “for RNA-seq” as listed 1.5 μg of each ALIBi plasmid was diluted in 100 μL Opti-MEM. In parallel, 7.5 μL of Lipofectamine 2000 was diluted in 100 μL Opti-MEM. The plasmid/Opti-MEM and Lipofectamine/Opti-MEM mixes were combined and incubated for 20 minutes at room temperature. For each condition, 3 million trypsinized cells were resuspended in 300 μL Opti-MEM, mixed with the plasmid-Lipofectamine complexes for 2 minutes at room temperature, and plated into one well of a 6-well plate containing 1.5 mL warmed media. Plates were incubated at 37 °C for 4 hours to allow reattachment before changing to fresh media. Media was changed 24 hours after transfection. At 48 hours after transfection, cells were changed to media containing 100 μg/mL cycloheximide and 50 uM biotin. Plates were immediately placed on a blue light transilluminator inside a 37 °C incubator for 15 minutes. Plates were washed once with ice-cold DPBS containing 100 μg/mL CHX, before adding 500 μL ice-cold lysis buffer per well. After incubating plates with lysis buffer on ice for 5 minutes, lysate was pipetted up and down six times in the plate with a P1000 pipet to fully dislodge cells and transferred to a 1.5 mL microcentrifuge tube. The lysate was further incubated for 15 minutes on ice, then clarified by centrifuging at 17,000 g for 3 minutes at 4 C. A 300 μL aliquot of the clarified lysate was added to pre-washed beads for affinity purification. For each sample, a 150 μL aliquot of beads was pre-washed as the storage buffer was removed and the beads were washed twice with 500 μL of RNAse removal buffer, once with 500 μL of salt wash, and twice with 500 μL of Wash I. After adding clarified lysate, beads were rotated at 4 °C for 1 h, then washed once with a double bead volume of Wash I and twice with a double bead volume of Wash II. The beads were then incubated in one bead volume of elution buffer at room temperature for 1 h on a thermomixer at 800 rpm.
For ER ALIBi with Lsg1, Nmd3, or Gnl3 siRNA knockdown and subsequent RNA-seq, a modified transfection was performed at 6-well 0.625 μg of each ALIBi plasmid and 26.7 pmol of Lsg1 siRNA (Horizon Discovery M-057491-01-0005), Nmd3 siRNA (Horizon Discovery M-051297-01-0005), Gnl3 siRNA (Horizon Discovery M-051664-01-0005), or control siRNA (Horizon Discovery D-001206-14-05) were diluted in 41.7 μL Opti-MEM. In parallel, 3.13 μL of Lipofectamine 2000 was diluted in 41.7 μL Opti-MEM. The plasmid/Opti-MEM and Lipofectamine/Opti-MEM mixes were combined and incubated for 20 minutes at room temperature. For each condition, 1.25 million trypsinized cells were resuspended in 125 μL Opti-MEM, mixed with the plasmid-Lipofectamine complexes for 10 minutes at room temperature, and plated into one well of a 6-well plate containing 2 mL warmed media. The remaining steps were performed as described above for ALIBi RNA-seq, with the modification that ALIBi activation, cell lysis, and affinity purification were performed 24 h after transfection.
Prior to affinity purification, an aliquot of cell lysate was reserved and mixed at a volume ratio of 9 with TRIzol (Thermo 15596-018) in an RNA non-stick tube. After affinity purification, eluate was mixed at a volume of 3 with TRIzol LS (Thermo 10296010) in an RNA non-stick tube. RNA was extracted using the DirectZol kit (R2060), following the manufacturer protocol without the on-column DNAse treatment. Extracted RNA was treated with Turbo DNAse (AM2239) per the manufacturer protocol and re-purified with an RNA Clean & Concentrator-5 column (R1016) per the manufacturer protocol. RNA concentration was estimated by Nanodrop 2000 (Thermo), and RNA integrity number (RIN) was assessed on an Agilent 2100 Bioanalyzer (total RNA pico) at the Stanford Protein and Nucleic Acid Facility. For each sample, 750 ng of RNA was used as input for the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (New England Biolabs E7760S) with the NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs E7490) and indices from the NEBNext Multiplex Oligos for Illumina (Index Primers Set 1 and 2) (New England Biolabs E7335S, E7500S), per the manufacturer protocol with 9 cycles of PCR amplification. Libraries were assessed on an Agilent 2100 Bioanalyzer (DNA 1000) to have an amplicon size of ~300 bp. Libraries were sequenced by Novogene on a Novaseq X with paired-end 150 bp reads to a total target depth of 20 - 26 million read pairs per sample.
For all RNA-seq analysis, the reference genome was generated using STAR/2.7.9a (117) on the Mus musculus GRCm39 primary assembly with annotations from GENCODE vM27 (118) and --sjdbOverhang 149. Alignment was performed with adapter sequences AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC and AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTAGATCTCGGTGGTCGCCGTATCATT, and --clip3pAdapterMMp 0.1 0.1. Read counting was performed with htseq-count from htseq/0.11.2 (119) -m intersection-nonempty -s reverse -i gene_id with annotations from GENCODE vM27. Genes were included in downstream analysis if the mean raw count across all samples was at least 5.
For RNA-seq analysis comparing cytosolic and ER ALIBi, differential abundance of read counts was performed using DESeq2 (120) with a design matrix of “~ organelle + fraction + replicate + fraction” where “organelle” has values of “cytoplasm” or “endoplasmic reticulum” and “fraction” has values of “lysate” (whole-cell RNA fraction) or “eluate” (ribosome-bound RNA fraction). GO-Slim Cellular Component gene ontology terms were mapped using the PANTHER database (121). For comparison to existing ER transcriptome and translatome datasets from HEK293 cells, mouse gene IDs were converted based on the corresponding ENSEMBL gene ID or UCSC ID in each species, and mouse genes with no human homologs were excluded.
For RNA-seq analysis of ER ALIBi with Lsg1, Nmd3, or Gnl3 siRNA knockdown, differential abundance of read counts was analyzed separately for the whole-cell RNA fractions and for the ER ribosome-bound RNA fractions from cells transfected with ER-ALIBi using DESeq2 with a design matrix of “~replicate + sirna” where “sirna” has values representing either the siRNA-targeted gene or “scramble” (non-targeting control siRNA). For the Nmd3 and Gnl3 whole-cell RNA fractions, one out of three replicates for each condition was excluded from analysis because it appeared as an outlier in principle component analysis (PCA). Uniprot subcellular location codes were used to identify genes encoding membrane proteins or secreted/lumenal proteins.
For all RNA-seq analysis, adjusted p-values reflect multiple comparison correction using the Benjamini-Hochberg method, as implemented in DESeq2.
To identify differentially abundant splicing isoforms, MAJIQ (build v2.4) (122) was executed using all default options and --simplify 10, and GENCODE vM27 annotations. Isoforms were visualized using ggsashimi (123).
All buffers were made in HPLC water (Fisher W5-4) unless otherwise stated. Eluates were precipitated using the ProteoExtract protein precipitation kit (MilliporeSigma 539180) per manufacturer protocol, and redissolved by vortexing in 50 μL of 6M urea (MilliporeSigma U6504), 2M thiourea (MilliporeSigma T8656), 50 mM ammonium bicarbonate (MilliporeSigma 5.3305). Dithiothreitol (DTT, Thermo PIA39255) was added to a final concentration of 5 mM, and the samples were incubated at room temperature for 30 min. Iodoacetamide (IAA, Thermo A39271) was added to a final concentration of 10 mM, and the samples were incubated at room temperature for 20 min in the dark. To quench residual IAA, additional DTT was added to a final concentration of 5 mM and incubated for 5 minutes at room temperature. Lyophilized LysC (Wako, 125-05061) was dissolved in 50 mM ammonium bicarbonate. 0.5 μg of LysC was added to each sample, and samples were incubated at 22 °C on a thermomixer at 400 rpm for 3 h. Samples were diluted with 190 μL of 50 mM ammonium bicarbonate. Lyophilized trypsin (Thermo 90057) was dissolved in water, and 1 μg was added per sample. Samples were incubated for 12 h at 22 °C on a thermomixer at 400 rpm. 50% heptafluorobutyric acid (HFBA, MilliporeSigma 52411) was added to a final concentration of 0.5%, and neat formic acid (MilliporeSigma A117) was added to achieve a pH of 1.5-2.5. Samples were desalted using OMIX C18 tips (Agilent A57003100) per the manufacturer protocol, and eluted twice with 50% acetonitrile (Fisher A955) with 0.1% formic acid, then once with 75% acetonitrile with 0.1% formic acid. Desalted samples were dried on a Speedvac (ThermoFisher).
For TMT labeling, samples were redissolved in 60 μL of 50 mM triethylammonium bicarbonate (TEAB, ThermoFisher T7408). TMT 6plex labels (ThermoFisher 90066) were dissolved in 100 μL of ethanol (MilliporeSigma 459828) per 0.8 mg vial. 20 μL of TMT label were added to each sample and incubated on a thermomixer in the dark at 22 °C, 400 rpm for 1 h. 4 μL of 5% v/v hydroxylamine (diluted from MilliporeSigma 467804) were added to each sample and incubated at 22 °C, 400 rpm for 15 minutes. Samples within the same replicate were pooled. 50% HFBA was added to a final concentration of 0.5%, and neat formic acid was added to achieve a pH of 1.5-2.5. Pooled samples were desalted using OMIX C18 tips per the manufacturer protocol, and eluted 3 times with 50% acetonitrile, 0.1% formic acid. Desalted samples were dried on a Speedvac (ThermoFisher), then fractionated on C18 columns at high pH (Thermo 84868) per the manufacturer protocol for TMT samples. Fractions were dried by SpeedVac and resuspended in 5% acetonitrile, 0.1% formic acid.
A Dionex Ultimate Rapid Separation Liquid Chromatography system (Thermo Fisher Scientific) was used to load 10 μL of the reconstituted tryptic peptides onto a C18 trap column (Thermo Fisher Scientific) with a flow rate set at 5 μL/min for 10 minutes. Tryptic peptides were separated by reversed-phase chromatography on a 25 cm long C18 analytical (New Objective) packed in-house with BEH C18, 130 Å, 1.7 μm particle size (Waters) encapsulated in a column heater (MSWIL) with a temperature set at 60° C. Peptides were eluted by changing the mixture of mobile phase A (0.1% formic acid in water) and mobile phase B (0.1% formic acid in acetonitrile). The gradient program consisted of holding mobile phase B at 2% for the first 6 minutes, slowly ramped up to 35% over the next 80 minutes, followed by an increase to 85% over 5 minutes with a 5-minute hold. The analytical column was re-equilibrated for 10 minutes prior to the next sample injection. The flow rate throughout the gradient was set to 0.3 μL/min. Eluted peptides were analyzed using an Orbitrap Eclipse Tribrid mass spectrometer (Thermo Fisher Scientific). The cycle time was set at top-speed for 3 seconds with an MS1 mass scan range of 375-2000 m/z and resolution of 120,000. The most abundant precursor ions were fragmented with Higher Energy Collisional Dissociation (HCD) and with a collisional energy set to 38%. Dynamic exclusion was enabled for 60 seconds and the normalized AGC target set to 250%. MS2 fragments were detected in the Orbitrap with a mass resolution of 30,000 with injection time set to auto and mass scan range between 110-1800 m/z.
Raw files were searched using MaxQuant 1.6.5.0 (124) against a Mus musculus Uniprot database containing canonical and isoform entries and the sequence of TEV protease (Uniprot ID P04517) as a reference for AcTEV, with matching between runs. Batch-specific TMT label correction factors were used, and reporter mass tolerance was set to 0.003 Da. Methionine oxidation, N-terminal acetylation, and asparagine/glutamine deamidation were allowed as variable modifications, and cysteine carbamidomethylation was allowed as a fixed modification. First search and main search peptide tolerance were both set to 20. Proteases were set to trypsin/P and LysC, with up to two missed cleavages allowed. Minimum peptide length was set to 6. PSM and protein FDR were both set to 0.01 with a minimum of 1 peptide, 1 razor+unique peptide, and 0 unique peptides.
Protein group IDs were filtered to remove reverse sequences, contaminant sequences, and those with a Q-value > 0.01. Normalization for total sample amount was performed using the quantity of TEV protease in each sample as an internal Ti,j,k=Ri,j,k×RTEV,cyto,kRTEV,j,k
Where Ri,j,k is the corrected reporter intensity for protein i in organelle j in replicate k, and T is the TEV-normalized value. Next, background subtraction to account for proteins that nonspecifically bind and elute from the streptavidin beads independently of biotinylation was performed using the GFP sample in each replicate as a no-ALIBi Bi,j,k=Ti,j,k−Ti,GFP,k
Where Bi,j,k is the background-subtracted value. Values of Bi,j,k<=0 were excluded from further analysis. To account for the amount of ribosomal material in each sample, the value for each protein was divided by the median value among RPs in that sample, and multiplied by the median value among RPs in the cytoplasmic sample in the same Ni,j,k=Bi,j,k×median[Bi,cyto,kBi,j,k]i∈RP
Where Ni,j,k is the RP-normalized value, cyto indicates a cytosolic ALIBi sample, and RP is the set of ribosomal proteins. A log-fold-ratio was calculated between organelle i and the reference cytosolic-ALIBi sample in each replicate after adding a pseudocount of 1 to avoid log(0) Li,j,k=log2(Ni,j,k+1)−log2(Ni,cyto,k+1)
Proteins with non-excluded values in at least 3 replicates were included. A one-sample t-test was performed against μ=0.
For mitochondrial samples, four replicates were initially collected but two were excluded due to an insufficient amount of detected protein after background subtraction. Due to a smaller total sample amount and fewer replicates, a modified analysis was performed for the mitochondrial after filtering protein groups as described above, a ratio of protein values between the mitochondrial ALIBi sample and the reference cytosolic ALIBi sample was calculated (with 1 added to the denominator to avoid divide-by-zero errors), and normalization for total sample amount was performed using the median value of these ratios within each Ci,j,k=Ri,j,kRi,cyto,k+1Mi,j,k=Ci,j,kmedian[Ci,j,k]i∈all
Where C is the ratio relative to the reference cytosolic ALIBi sample and M is the median-normalized value. Next, background subtraction was performed analogously as Bi,j,k=Mi,j,k−Mi,GFP,k
Values of Bi,j,k<0 were set to 0. Protein i in replicate k was excluded from further analysis if Bi,mitok=Bi,cyto,k=0. To account for the amount of ribosomal material in each sample, the value for each protein was divided by the median value among RPs in that Ni,j,k=Bi,j,kmedian[Bi,j,k]i∈RP
A log-fold-ratio was calculated as described above. Proteins with non-excluded values in both replicates were included. A one-sample t-test was performed against μ = 0. As an alternative to multiple hypothesis correction, which can be overly strict and impractical for mass spectrometry proteomics data (125), we threshold for likely organelle-enriched RAPs on both log-fold-change and unadjusted p-value.
A coding sequence encoding whole BirA fused to either a nuclear export sequence or Sec61β (Table S1) was cloned into a construct containing a T7 promoter and human hemoglobin (hHBB) 5’ and 3’ UTRs. A no-BirA control construct was also assembled, which contained a GFP coding sequence in lieu of whole BirA. These DNA templates for in vitro transcription were amplified using KAPA HiFi Hotstart PCR kit (Kapa Biosystems KR0369) per manufacturer instructions and gel purified from an agarose gel (New England Biolabs Monarch DNA Gel Extraction Kit T1020L). In vitro transcription was performed using the MEGAscript T7 transcription kit (ThermoFisher AM1334) and column purified using the MEGAclear transcription clean-up kit (ThermoFisher AM1908). m7G capping and polyadenylation were performed using the ScriptCap m7G Capping System (CellScript C-SCCE0625) and A-Plus Poly(A) Polymerase Tailing Kit (CellScript C-PAP5104H). The transcription and capping reactions were performed using the manufacturer’s recommended conditions, and the polyadenylation reaction was carried out by adding the requisite reagents to the capping reaction directly without an intermediate purification step. The polyadenylation was performed in a 66 μL reaction volume containing 30 μg input RNA (as quantified prior to capping) and 10 units of A-Plus PolyA polymerase, incubated for 30 minutes at 37 °C. The product was again purified on a MEGAclear column. After each column purification, the expected product length was confirmed by BioAnalyzer.
To transfect, 7.5 × 10^6^ Rpl31-FTA mESCs were plated on a pre-gelatinized 10 cm plate 4h prior to transfection. 24 μL of Lipofectamine MessengerMax (ThermoFisher LMRNA001) was diluted in 1 mL Opti-MEM Reduced Serum Medium (ThermoFisher 11058021). 12 μg of RNA was diluted in a separate aliquot of 1 mL Opti-MEM. The diluted Lipofectamine and RNA were combined, inverted several times to mix, and incubated for 5 minutes at room temperature. The media was aspirated from the plated mESCs and the RNA/Lipofectamine mix was added to the cells. The cells were incubated for 5 minutes at room temperature before fresh warmed media was added. After 2h of incubation, the cells were then incubated in fresh warmed media containing 50 μM supplemental biotin and 100 μg/mL cycloheximide for 15 minutes, then immediately harvested for affinity purification as described above.
For RiboExM, each expansion experiment was performed at least three times independently unless otherwise stated (technical replicates). Analysis results from the different replicates were pooled and recorded in GraphPad Prism 10. The significance level was determined using GraphPad Prism 10, assuming Gaussian distribution and equal variance between groups. Unpaired t test or ordinary one-way ANOVA was carried out unless otherwise stated. Graphs are shown as mean ± standard deviation unless otherwise stated. ns P > 0.05; * P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001; **** P ≤ 0.0001.
For ALIBi, the number of replicates and statistical tests used are stated in the respective figure legends describing each experiment. Each replicate was an independently performed experiment (technical replicates).
This work strives to adhere to relevant guidelines for experiment reporting, including Minimal Information about a high throughput SEQuencing Experiment (MINSEQE) and Minimum Information About a Proteomics Experiment (MIAPE).
Supplementary Text
Figs. S1 to S17
Tables S1 to S8
References (127-134)
Movies S1 to S2