Mice
The Cornell College Institutional Animal Care and Use Committee (IACUC) authorized all animal protocols and experiments have been carried out in compliance with its institutional pointers. For skeletal muscle samples, grownup feminine C57BL/6J mice have been obtained from Jackson Laboratories (catalog no. 000664) and have been used at 6 months of age. For coronary heart samples, confirmed pregnant feminine C57BL/6J mice have been ordered from Jackson Laboratories to be delivered at embryonic stage E14.5.
Viral an infection
Litters weighing 3 g per pup have been orally gavaged utilizing intramedic tubing (Becton Dickinson, calalog no. 427401) with 50 μl with 107 plaque-forming models of REOV kind 1-lang (T1L) pressure in 1× phosphate buffered saline (PBS) containing inexperienced meals colour (McCormick) by way of a 1 ml tuberculin slip tip syringe (BD, catalog no. 309659) and 30G × 1/2 needle (BD, catalog no. 305106). Litters handled with 1× PBS containing inexperienced meals colour alone on the identical day have been used as mock controls for the respective an infection teams. The mock-infected and REOV-infected pups have been monitored and weighed every day till the timepoints used within the research (7 dpi). After dissection, samples have been embedded in OCT Compound (Tissue-Tek) and frozen recent in liquid nitrogen.
Muscle harm
To induce muscle harm, each tibialis anterior muscle groups of 6-month-old C57BL/6J mice have been injected with 10 µl notexin (10 µg ml–1; Latoxan). Both earlier than harm or 2, 5 or 7 dpi, mice have been sacrificed and tibialis anterior muscle groups have been collected. After dissection, samples have been embedded in OCT Compound (Tissue-Tek) and frozen recent in liquid nitrogen.
In situ polyadenylation and STRS
STRS was carried out utilizing a modified model of the Visium protocol. Tissue sections (10 μm thick) have been mounted onto the Visium Spatial Gene Expression v1 slides. For coronary heart samples, one tissue part was positioned into every 6 × 6mm seize space. For skeletal muscle samples, two tibialis anterior sections have been positioned into every seize space. After sectioning, tissue sections have been mounted in methanol for 20 min at −20 °C. Subsequent, H&E staining was carried out in response to the Visium protocol, and tissue sections have been imaged on a Zeiss Axio Observer Z1 Microscope utilizing a Zeiss Axiocam 305 colour digicam. H&E photographs have been shading corrected, stitched, rotated, thresholded and exported as TIFF recordsdata utilizing Zen v.3.1 software program (Blue version). After imaging, the slide was positioned into the Visium Slide Cassette. In situ polyadenylation was then carried out utilizing yeast poly(A) polymerase (yPAP; Thermo Scientific, catalog no. 74225Z25KU). First, samples have been equilibrated by including 100 µl 1× wash buffer (20 µl 5× yPAP Response Buffer, 2 µl 40 U µl–1 Protector RNase Inhibitor, 78 µl nuclease-free H2O) (Protector RNase Inhibitor; Roche, catalog no. 3335402001) to every seize space and incubating at room temperature for 30 s. The buffer was then eliminated. Subsequent, 75 µl yPAP enzyme combine (15 µl 5× yPAP response buffer, 3 µl 600U µl–1 yPAP enzyme, 1.5 µl 25 mM ATP, 3 µl 40U µl–1 Protector RNase Inhibitor, 52.5 µl nuclease-free H2O) was added to every response chamber. STRS was additionally examined with 20 U µl–1 SUPERase-In RNase Inhibitor (Thermo Fisher Scientific, catalog no. AM2694), however we discovered that SUPERase was not in a position to stop degradation of longer transcripts throughout in situ polyadenylation (Supplementary Fig. 12). The response chambers have been then sealed, and the slide cassette was incubated at 37 °C for 25 min. The enzyme combine was then eliminated. Earlier than working STRS, optimum tissue permeabilization time for each coronary heart and skeletal muscle samples was decided to be 15 min utilizing the Visium Tissue Optimization Equipment from 10x Genomics. Following in situ polyadenylation, the usual Visium library preparation was adopted to generate cDNA and last sequencing libraries. The libraries have been then pooled and sequenced in response to pointers within the Visium Spatial Gene Expression protocol utilizing both a NextSeq 500 or NextSeq 2000 (Illumina).
Focused pulldown enrichment of viral fragments
We carried out hybridization-based enrichment of viral fragments on the Visium and STRS libraries for REOV-infected hearts utilizing the xGen Hybridization and Wash Equipment (IDT; 1080577)4. On this strategy, a panel of 5′-biotinylated oligonucleotides was used for seize and pulldown of goal molecules of curiosity, which have been then PCR amplified and sequenced. We designed a panel of 202 biotinylated probes tiled throughout all the REOV T1L genome to selectively sequence viral molecules from the sequencing libraries (Supplementary Desk 1). After fragmentation and indexing of cDNA, 300 ng of the ultimate Visium or STRS sequencing libraries from REOV-infected hearts have been used for xGen hybridization seize utilizing the xGen NGS Goal Enrichment Equipment protocol offered by the producer. One spherical of hybridization seize was carried out for the STRS library adopted by 14 cycles of PCR amplification. Due to the diminished variety of captured molecules, two rounds of hybridization have been carried out on the Visium libraries. Enriched Visium libraries have been PCR amplified for 18 cycles after the primary spherical of hybridization and by 5 cycles after the second spherical of hybridization. Postenrichment merchandise have been pooled and sequenced on the Illumina NextSeq 500.
Single-nucleus whole RNA-sequencing
C2C12 cells have been grown to 90% confluence and picked up with 0.25% TrypLE (Thermo Fisher Scientific). Nuclei have been remoted much like Petrany et al.41. Cells have been pelleted by centrifugation at 500g, at 4 °C, for five min, and resuspended in 6 ml chilled homogenization buffer (0.25 M sucrose, 1% bovine serum albumin, 1× PBS, 0.2 U µl–1 SUPERase•In RNase Inhibitor, nuclease-free H2O). Then, 1 ml chilled 2.5% Triton-X100 diluted in 1x PBS was added. Cells have been incubated on ice for five min, then pelleted by centrifugation at 1,000g at 4 °C for five min. Nuclei have been then resuspended in 1× PBS and counted utilizing Trypan blue. A complete of 5 million nuclei have been suspended in 200 µl 1× PBS, then 800 µl ice-cold methanol was added dropwise to repair. Nuclei have been then saved at −20 °C in a single day.
On the day of the experiment, nuclei have been faraway from −20 °C and incubated on ice for five min. Nuclei have been then pelleted by centrifugation at 1,000g, at 4 °C, for five min and resuspended in 200 µl wash resuspension buffer (0.04% bovine serum albumin, 1 mM DTT, 0.2 U µl–1 SUPERase•In RNase Inhibitor, 3× SSC buffer (Thermo Fisher Scientific, catalog no. 15557044), nuclease-free H2O). Nuclei have been then pelleted by centrifugation at 1,000g, at 4 °C, for five min and washed in 200 µl 1× wash buffer (40 µl 5× yPAP response buffer, 4 µl 20U µl–1 SUPERase•In RNase Inhibitor, 156 µl nuclease-free H2O). In situ polyadenylation was then carried out by suspending nuclei in 50 µl yPAP enzyme combine (10 µl 5× yPAP Response Buffer, 2 µl 600 U µl–1 yPAP enzyme, 1 µl 25 mM ATP, 2 µl 20 U µl–1 SUPERase•In RNase Inhibitor, 35 µl nuclease-free H2O) and incubating at 37 °C for 25 min with out agitation. Nuclei have been then washed with 500 µl of nuclei suspension buffer42 (10 mM Tris-HCl pH 7.5, 10 mM NaCl, 3 mM MgCl2, 1% bovine serum albumin, 0.2 U µl–1 SUPERase•In RNase Inhibitor, nuclease-free H2O) and pelleted by centrifugation at 1,000g, at 4 °C, for five min. Nuclei have been lastly resuspended in 200 µl nuclei suspension buffer, counted utilizing a Countess 3 (Thermo Fisher Scientific) and the LIVE/DEAD Viability/Cytotoxicity Equipment (Thermo Fisher Scientific, catalog no. L3224), then diluted to the right focus. Nuclei for normal single-nucleus RNA-sequencing have been processed equally, however with no in situ polyadenylation step (counted instantly after wash buffer was added). A complete of three,300 nuclei have been then loaded onto the Chromium controller (10x Genomics) for a focused seize of two,000 nuclei. Libraries have been generated utilizing the usual Chromium v.3 protocol. Remaining libraries have been sequenced utilizing the Illumina MiniSeq and Illumina NextSeq 500.
Small RNA-sequencing
For skeletal muscle samples, following the harm time course, tibialis anterior muscle groups have been dissected and snap frozen with liquid nitrogen. The Norgen Complete RNA Purification Equipment (catalog no. 17200) was used to extract RNA from 10 mg tissue for every pattern. For coronary heart samples, following the an infection time course, hearts have been dissected, embedded in OCT, and frozen in liquid nitrogen. RNA was extracted with Trizol (Invitrogen, catalog no. 15596026) and glycogen precipitation for a small fraction of every of the guts samples. RNA high quality was assessed by way of Excessive Sensitivity RNA ScreenTape Evaluation (Agilent, catalog no. 5067-5579) and all samples had RNA integrity numbers higher than or equal to seven.
Small RNA sequencing was carried out on the Genome Sequencing Facility of Greehey Kids’s Most cancers Analysis Institute on the College of Texas Well being Science Middle at San Antonio. Libraries have been ready utilizing the TriLink CleanTag Small RNA Ligation equipment (TriLink Biotechnologies). Libraries have been sequenced with single-end 50× utilizing a HiSeq2500 (Illumina).
Preprocessing and alignment of STRS, single-nucleus whole RNA-sequencing, Good-Seq-Complete and VASA-seq information
All code used to course of and analyze these information will be discovered at https://github.com/mckellardw/STRS. A top level view of the pipelines used for preprocessing and alignment is proven in Supplementary Fig. 1a.
Reads have been first trimmed utilizing cutadapt v.3.4 (ref. 43) to take away the next sequences: (1) poly(A) sequences from the three′ ends of reads, (2) the template swap oligonucleotide sequence from the 5′ finish of reads derived from the Visium Gene Expression equipment (sequence: CCCATGTACTCTGCGTTGATACCACTGCTT), (3) poly(G) artifacts from the three′ ends of reads, that are produced by the Illumina two-color sequencing chemistry when cDNA molecules are shorter than the ultimate learn size and (4) the reverse complement of the template switching oligonucleotide sequence from the 5′ ends of reads (sequence: AAGCAGTGGTATCAACGCAGAGTACATGGG). Subsequent, reads have been aligned utilizing both STAR v.2.7.10a44 or kallisto v.0.48.0 (ref. 45). Workflows have been written utilizing Snakemake v.6.1.0 (ref. 46).
For STAR, the genomic reference was generated from the GRCm39 reference sequence utilizing GENCODE M28 annotations. For STAR alignment, the next parameters, based mostly on work by Isakova et al. 6, have been used: outFilterMismatchNoverLmax=0.05, outFilterMatchNmin=16, outFilterScoreMinOverLread=0, outFilterMatchNminOverLread=0, outFilterMultimapNmax=50. Aligned reads have been deduplicated for visualization utilizing umi-tools v.1.1.2 (ref. 47). Aligned and deduplicated reads have been visualized with Built-in Genome Viewer v.2.13.0 (ref. 48). Normalized gene place plots and genomic loci profiles have been generated utilizing Qualimap v.2.2.2.a49.
For kallisto, a transcriptomic reference was additionally generated utilizing the GRCm39 reference sequence and GENCODE M28 annotations. The default k-mer size of 31 was used to generate the kallisto reference. Reads have been pseudoaligned utilizing the ‘kallisto bus’ command with the chemistry set to ‘VISIUM’ and the ‘fr-stranded‘ flag activated to allow strand-aware quantification. Pseudoaligned reads have been then quantified utilizing bustools v.0.41.0. First, spot barcodes have been corrected with ‘bustools right‘ utilizing the ‘Visium-v1’ whitelist offered within the Area Ranger software program from 10x Genomics. Subsequent, the output bus file was sorted and counted utilizing ‘bustools kind’ and ‘bustools depend,’ respectively. To estimate the variety of spliced and unspliced transcripts, reads pseudoaligned utilizing kb-python v.0.26.0, utilizing the ‘lemanno’ workflow.
Spots have been chosen manually based mostly on the H&E photographs utilizing Loupe Browser from 10x Genomics. Spatial areas for every spot have been assigned utilizing the Visium coordinates offered for every spot barcode by 10x Genomics within the Area Ranger software program (‘Visium-v1_coordinates.txt’). Downstream analyses with the output depend matrices have been then carried out utilizing Seurat v.4.0.4 (refs. 50,51). Along with guide choice, spots containing fewer than 500 detected options or fewer than 1,000 distinctive molecules have been faraway from the evaluation. Counts from multimapping options have been collapsed right into a single function to simplify quantification. Gene biotype percentages have been computed in response to gene biotypes offered within the GENCODE M28 annotations.
Single-nucleus information have been preprocessed and aligned as described above, with a special barcode whitelist matching the 10x Genomics Chromium v.3 chemistry. Rely matrices have been filtered for cells with greater than 750 distinctive molecules and fewer than 5% of reads mapping to mitochondrial genes. Counts have been then log-normalized with Seurat. Cells have been merged and differential gene expression evaluation was carried out between the usual and in situ polyadenylated nuclei utilizing the ‘FindMarkers()‘ perform. A two-sided Wilcoxon Rank-Sum take a look at was used for differential gene expression evaluation.
Uncooked fastq recordsdata for VASA-drop7 samples have been downloaded from GEO (GSE176588) utilizing parallel-fastq-dump (v.0.6.5). Reads have been trimmed utilizing cutadapt v.3.4 to take away poly(A) and poly(G) sequences. Reads have been then aligned and quantified utilizing kallisto/BUStools as described above. The ‘–know-how’ flag for ‘kallisto bus’ was set to ‘0,6,22:0,0,6:1,0,0’ for cell barcode and UMI identification to replicate the modified fastq recordsdata authors uploaded to GEO. Gene counts from multimapping options have been collapsed right into a single function.
Uncooked fastq recordsdata for Good-Seq-Complete6 samples have been downloaded from GEO (GSE151334) utilizing parallel-fastq-dump (v.0.6.5). Reads have been then pseudoaligned utilizing ‘kallisto quant’ with the ‘–fragment-length’ flag set to 75 and the ‘–sd’ flag set to 10. Transcript counts have been transformed to gene counts in response to GENCODE M28 gene symbols, then counts from multimapping options have been collapsed right into a single function.
Rarefaction evaluation of Visium and STRS information
Uncooked fastq recordsdata for every library have been randomly downsampled 4 instances utilizing seqtk v.1.2 to last learn counts totaling between 100,000 and 50,000,000 reads. Remaining libraries have been then pseudoaligned utilizing the kallisto pipeline described above.
Annotation-free quantification of transcriptionally energetic areas in single-nucleus RNA-sequencing information
The ‘from_STARsolo’ model of the TAR-scRNA-seq9 pipeline was used with the outputs from reads aligned with STAR for single-nucleus RNA-sequencing information. Default parameters have been used for ‘MERGEBP’ (500) and ‘THRESH’ (10,000,000) for TAR merging and filtering, respectively. Rely matrices generated by TAR-scRNA-seq have been subset based mostly on cell barcodes that remained after customary high quality management.
Mature microRNA quantification
For STRS information: after trimming (see above), barcode correction with STAR v.2.7.10a and UMI-aware deduplication with umi-tools v.1.1.2, reads have been break up throughout all 4,992 spot barcodes and analyzed utilizing miRge3.0 v.0.0.9 (ref. 24). Reads have been aligned to the miRbase reference offered by the miRge3.0 authors. MiRNA counts have been log-normalized in response to the full variety of counts detected by kallisto and scaled utilizing a scaling issue of 1,000. For small RNAseq information: reads have been first trimmed utilizing trim_galore v.0.6.5. Reads have been then aligned and counted utilizing miRge3.0 v.0.0.9. Uncooked fastq recordsdata for all skeletal muscle and coronary heart datasets from the Small RNA Atlas52 have been downloaded from GEO (GSE119661) and processed equally.
Unsupervised clustering and differential gene expression evaluation of spot transcriptomes
Spot UMI counts as generated by kallisto have been used. First, counts have been log-normalized and scaled utilizing default parameters with Seurat. Principal part evaluation was then carried out on the highest 2.000 most variable options for every tissue slice individually. Lastly, unsupervised clustering was carried out utilizing the ‘FindClusters()‘ perform from Seurat. The highest principal elements which accounted for 95% of variance throughout the information have been used for clustering. For skeletal muscle samples, a clustering decision was set to 0.8. For coronary heart samples, clustering decision was set to 1.0. Default choices have been used for all different parameters. Lastly, clusters have been merged in response to comparable gene expression patterns and based mostly on histology of the tissue underneath every subcluster.
Differential gene expression evaluation was carried out utilizing the ‘FindAllMarkers()‘ perform from Seurat. Default parameters have been used, together with using the two-sided Wilcoxon rank-sum take a look at to determine differentially expressed genes. To determine options enriched within the skeletal muscle STRS datasets, all Visium and STRS have been first merged and in contrast in response to the strategy used (Visium versus STRS). To determine cluster-specific gene expression patterns, skeletal muscle samples have been first clustered as described above individually. STRS samples have been then merged, and differential gene expression evaluation was carried out throughout the three harm area teams.
Cell-type deconvolution of Visium and STRS datasets
Cell-type deconvolution of skeletal muscle Visium and STRS information was carried out as beforehand5 utilizing BayesPrism19 (beforehand generally known as ‘Tumor microEnvironment Deconvolution’, TED, v.1.0; github.com/Danko-Lab/TED). We used the ‘scMuscle’ dataset generated in McKellar et al.5,53 as a single-cell transcriptomic reference for skeletal muscle. For coronary heart samples, we used all mock and contaminated single-cell RNA-sequencing samples generated by Mantri et al.4. Extremely and differentially expressed genes throughout cell varieties have been recognized with differential gene expression evaluation utilizing Seurat (FindAllMarkers, utilizing two-sided Wilcoxon rank-sum take a look at). The ensuing genes have been filtered based mostly on common log2-fold change (avg_logFC > 1) and the proportion of cells throughout the cluster that categorical every gene (pct.expressed > 0.5), yielding round 1,000 genes in each single-cell references. Mitochondrial and ribosomal protein genes have been faraway from this record, in keeping with suggestions from the BayesPrism authors. For every of the cell varieties, imply uncooked counts have been calculated throughout the round 1,000 genes to generate a gene expression profile for BayesPrism. Uncooked counts for every spot have been then handed to the run.Ted perform, utilizing the ‘GEP’ possibility for enter.kind and default parameters for the remaining inputs. Remaining Gibbs theta values have been used as estimates for the fraction of transcripts from every spot that have been derived from every of the cell varieties. In plots (Supplementary Figs. 8 and 11), a minimal threshold worth for theta of 0.01 was used. For skeletal muscle, after deconvolution all spots have been merged. Principal part evaluation was carried out on the nonthresholded BayesPrism theta values utilizing Seurat.
Correlation evaluation between REOV counts, host gene expression and spot cell-type fraction
We used a generative additive mannequin carried out in Monocle v.2.18.0 (ref. 54) to search out genes that adjust with viral UMI depend. A Seurat object for STRS information and viral UMI counts from the REOV-infected coronary heart was transformed to a CellDataSet object utilizing the ‘as.CellDataSet()’ command carried out in Seurat. For comparability between gene expression and REOV counts, the expression household was set to ‘detrimental binomial’ as urged for UMI depend information within the Monocle documentation. For comparability between cell-type fraction (theta, as computed by BayesPrism), a minimal theta worth of 0.001 was used and the expression household was set to ‘uninormal’. The CellDataSet objects have been then preprocessed to estimate measurement elements and dispersion (only for comparability with genes). Genes expressed in fewer than ten spots have been eliminated. We then used the generative additive mannequin carried out within the ‘differentialGeneTest()’ command in Monocle to determine genes or cell-type fractions that adjust with log2-transformed viral UMI counts. To seek out the path during which these genes diversified with viral UMI counts, we calculated the Pearson correlation for all genes with log2-transformed viral UMI counts.
RNA fluorescence in situ hybridization utilizing hybridization chain response HCR-V3
Single-molecule fluorescence in situ hybridization (smFISH) was carried out as described beforehand4,55. Probes have been designed utilizing NCBI primer-blast for two-step hybridization technique with break up probe design and hybridization chain response (HCR)-V3 (ref. 55) (Supplementary Desk 2). Break up probes for every gene goal have been blended and diluted in nuclease-free water to a last whole probe focus of 10 µM. Hairpin pairs labeled with two completely different fluorophores, specifically Alexa-488 and Alexa-546 (Molecular Devices), have been used for HCR-V3.
Slides with tissue sections have been warmed to room temperature after which mounted in 4% paraformaldehyde for 12 min at room temperature. After fixation, sections have been washed for five minutes in 1× PBS twice, incubated for 1 h in 70% ethanol for tissue permeabilization, washed once more for five minutes in 1× PBS, after which used for major hybridization. Hybridization buffer (HB) combine was ready with 2× SSC, 5× Denhardt’s resolution, 10% ethylene carbonate, 10% dextran sulfate, 0.01% SDS, 1 µM of probe pool combine per goal for the hybridization response. 20 µl of HB combine (with probes) per part was then placed on every slide to cowl the tissue part, coated with Parafilm and incubated in a single day at 37 °C inside a humidifying chamber for major hybridization. After major hybridization, Parafilm was eliminated and slides have been washed in hybridization wash buffer (0.215 M NaCl, 0.02 M Tris-HCl pH 7.5 and 0.005 M EDTA) for 20–30 min at 48 °C. Amplification buffer (AB) combine was ready with 2× SSC, 5× Denhardt’s resolution, 10% dextran sulfate, 0.01% SDS and 0.06 µM HCR hairpins for the amplification response. Then, 2 µl of every fluorophore labeled hairpins at 3 µM similar to the goal genes have been blended, incubated at 95 °C for 1.5 min, coated in aluminum foil and cooled to room temperature for 30 min to kind hairpins earlier than including to the AB combine. A 20 µl portion of AB combine per part was then positioned on every slide to cowl the tissue part, coated with Parafilm and incubated in a single day at room temperature at the hours of darkness for sign amplification. After sign amplification, Parafilm was eliminated and slides have been washed in 5× SSCT buffer twice for 30–40 min after which twice for 10 minutes. The slides have been then cleaned rigorously with Kimwipe and handled with Prepared Probes Auto-fluorescence Quenching Reagent Combine (Thermo Fisher, catalog no. R37630) for five min and washed 3 times in 1× PBS. Lastly, tissue sections have been then counterstained with 4,6-diamidino-2-phenylindole for 10 min at room temperature, washed for five min in 1× PBS twice, extra PBS was cleaned off utilizing Kimwipe, and sections have been instantly mounted on coverslips utilizing Slowfade antifade media, left in a single day for therapy and imaged the following day on a Zeiss Axio Observer Z1 Microscope utilizing a Hamamatsu ORCA Fusion Gen III Scientific CMOS digicam. FISH photographs have been shading corrected, stitched, rotated, thresholded and exported as TIFF recordsdata utilizing Zen v.3.1 software program (Blue version).
Reporting abstract
Additional info on analysis design is on the market within the Nature Analysis Reporting Abstract linked to this text.
