This process is regulated at many levels, yielding a dynamic steady-state mRNA population that is maintained by synthesis and turnover, at varying rates for each individual transcript [ 5 ]. The rate of transportation from nucleus to cytoplasm can affect the amount of transcript detected in both the total and the cytoplasmic fractions, and hence might bias measurements of transcript levels. It has previously been shown that mRNA molecules that are not of immediate need to produce proteins are retained in the nucleus [ 6 , 7 ].
In addition to nuclear retention, the gene level is also regulated by other mechanisms and one of them is the degradation of mRNA by the exosome complex [ 8 , 9 ].
It is known that the levels of mRNA and protein abundance in cells are modestly correlated [ 10 — 12 ]. Validation of this argument will require studies that assess how well the levels of total RNA and cytoplasmic RNA are correlated with protein abundance. To investigate the impact of nuclear transcripts present in total RNA, we compared the expression levels of genes obtained from the total fraction with those obtained from the cytoplasmic fraction. We investigated the effect of the length and structure of untranslated regions and the length of the coding sequences on the transcript levels in total and cytoplasmic RNA.
We present here an extensive study of RNA-Seq that compares gene expression levels from poly A isolated total and cytoplasmic RNA as well as their relation to protein levels. Each extraction was replicated four times. To ensure that the cytoplasmic fraction was pure from nuclear contamination, all extractions were analyzed using capillary electrophoresis Additional file 1 : Figure S1. The nuclear and cytoplasmic preparations had, in addition to the ribosomal peaks, discriminatory signature profiles in which the nuclear fractions contained an additional peak, which were present only in the total RNA preparation [ 3 , 13 ].
All of the total RNA samples displayed the signature peak, whereas the cytoplasmic fractions did not except for one cytoplasmic U-2 OS sample, which was removed from further processing. The samples were then sequenced using massively parallel sequencing, and RPKM Reads Per Kilobase of exon model per Million mapped reads values were calculated [ 14 ].
Distribution of gene expressions for the total and cytoplasmic preparation. A : Heatmap of sample preparation and cell lines. The DESeq algorithm was used to find sets of genes detected at different levels in cytoplasmic and in total RNA [ 15 ], hereafter referred to as differentially detected DD genes.
A number of DD genes were identified between the total and cytoplasmic fractions within each cell line Figure 2 A—C. Number of differentially detected genes between the preparation methods for each cell line. Messenger RNAs vary in sequence and length and this can affect their rate of transportation to the cytoplasm. To investigate this, genes that were detected differentially—in one, two, or all three cell lines—were selected and classified into two groups: genes that had a higher number of copies in the total RNA fraction and genes that had a lower number of copies in the total RNA fraction and plotted separately Figure 2 A and B.
Differential detection of genes in total or cytoplasmic RNA fractions relies on that total RNA fraction would contain all mature polyadenylated transcripts whether they were in the cytoplasm or in the nucleus of the cell, whereas the cytoplasmic fractions only contain transcripts already transported to the cytoplasm.
To study whether the lengths of untranslated regions UTRs could affect the transportation rate of transcripts, we compared the UTR and coding sequence lengths of differentially detected genes with those of genes exhibiting no differential detection. This trend was consistent for genes that were differentially detected in one or more cell line Figure 3 and Additional file 3 : Figure S3A and B.
A more negative fold energy corresponds to a more structured sequence. Figure 3 D—E and Additional file 4 : Figure S4 show that genes that were detected at higher levels in total RNA had lower UTR fold energies were more structured than those with no differential detection. Boxplot showing length and fold energies of UTRs and coding sequence for all cell lines. C : Coding sequence length. Transcripts that are degraded in the cytoplasm in high rates will also contribute to the differential detection since those degraded in cytoplasm will be detected at lower levels in the cytoplasmic RNA fraction compared to total RNA fraction.
To investigate whether these genes have a higher number of micro-RNA miRNA targets, hence resulting in a higher probability for degradation when exported into the cytoplasm, an analysis comparing the number of miRNA targets per gene was performed. The same method to classify differentially detected genes described in Figure 2 was used for the analysis. As described by Akan et al. Interestingly, the cell line with a more pronounces difference between the groups, U-2 OS, is also the cell line that has more uniquely expressed miRNAs, and this could be the explanation for the slightly higher number of miRNA targets per gene seen in U-2 OS.
Overall, the data suggests that the miRNA may be one of the contributing factors for differential detection of genes. Boxplot showing the number of microRNA targets per gene for all three cell lines separately. A : A B : U-2 OS. C : U MG. A ratio-based correlation analysis Spearman was performed between protein abundance levels detected by mass spectrometry for approximately proteins [ 10 ] and the corresponding total and cytoplasmic mRNA levels, for each cell line.
The correlation coefficients for the other two cell lines UMG and A were very similar. The correlations were similar whether differentially detected genes were included or excluded, see Additional file 5 : Table S1 for correlation coefficients between protein abundance and total and cytoplasmic RNA, respectively, for genes detected differentially in all three cell lines.
When designing a gene expression experiment with the goal of measuring steady-state levels of mRNA, care should be taken to isolate RNA from the correct cellular compartment.
However, isolating the cytoplasmic RNA instead of total RNA is feasible when working with cell cultures, but for many other biological models are total RNA the only choice. Despite the proposed advantage of sequencing only cytoplasmic RNA for cells in suspension, it is still not clear whether the cytoplasmic fraction represents the full complexity of the steady-state RNA of whole cells. One argument against using cytoplasmic RNA could be that the translation levels of certain transcripts might be regulated by their transportation rate from nucleus to cytoplasm [ 6 , 7 ].
Moreover, the transportation rate of transcripts from nucleus to cytoplasm could depend on particular properties of the transcript such as length or sequence. Here, we investigated how the representations of transcripts differ between the cytoplasmic and total RNA fractions. Multiplexing is also the best way to minimize potential lane-to-lane sequencing variation, as all of your samples are subject to the same sequencing conditions. For example, if you require two sequencing lanes for six samples we recommend 6-plexing and sequencing over two lanes, instead of 3-plexing per lane.
Libraries containing different indexed adapters are then constructed, quantified, pooled in equimolar amounts, and sequenced. Deconvoluting the barcodes informatically allows multiple libraries to be sequenced in a single lane at a potential cost and time saving. To date, two methods have been exploited for this: using the commercially available indexing kits Illumina TruSeq, Nextera, or Bioo Scientific or synthesizing your own adapter oligos with your own barcodes.
When sequencing on the HiSeq and especially on the NovaSeq it is highly recommended to use uniquely-dual-indexed adapters UDI adapters to avoid index hopping artifacts. Please see this FAQ. How many cells will I need? Do you have recommendations for the isolation of plant total RNA samples? How should I purify my samples? Other Library Considerations Library Indexing and Pooling Indexing, also called barcoding, allows for the sequencing of multiple libraries in a single lane, i.
Recent Posts. Abstract Single-cell analysis enables detailed molecular characterization of cells in relation to cell type, genotype, cell state, temporal variations, and microenvironment. Keywords: cell heterogeneity, sarcoma, single-cell analysis, total mRNA level, transcriptome size. Introduction Gene expression profiling is widely used in both research and medicine for the characterization of different biological and pathological conditions.
Materials and Methods 2. Single-Cell Collection Cells were detached using 0. Results 3. Open in a separate window. Figure 1. Figure 2. Individual Sarcoma Cells Reveal Heterogeneity in Total Polyadenylated Transcriptome Levels Sarcoma includes many entities with specific cellular phenotypes and unique genotypes, all with mesenchymal origin.
Figure 3. Discussion We developed a method to quantify the amount of polyadenylated RNA in single cells, which can be used to profile global transcript differences among cell types as well as to monitor the effects of intrinsic and extrinsic factors.
Acknowledgments The authors wish to thank Malin Nilsson for valuable experimental assistance. Click here for additional data file. Author Contributions Conceptualization, E. Conflicts of Interest A. References 1. Kubista M. The secrets of the cell. Bengtsson M. Gene expression profiling in single cells from the pancreatic islets of Langerhans reveals lognormal distribution of mRNA levels. Genome Res. Hedlund E. Single-cell RNA sequencing: Technical advancements and biological applications.
Han X. Rodriguez-Fraticelli A. Clonal analysis of lineage fate in native haematopoiesis. Avraham R. Miyamoto D. Patel A. Single-cell RNA-seq highlights intratumoral heterogeneity in primary glioblastoma. Coate J. Variation in transcriptome size: Are we getting the message? Islam S. Characterization of the single-cell transcriptional landscape by highly multiplex RNA-seq.
Marguerat S. Coordinating genome expression with cell size. Trends Genet. Mitchison J. Growth During the Cell Cycle. Nucleosome loss leads to global transcriptional up-regulation and genomic instability during yeast aging. Genes Dev. Lin C. Nie Z. Global transcriptional and translational repression in human-embryonic-stem-cell-derived Rett syndrome neurons. Cell Stem Cell. Revisiting global gene expression analysis. Dolatabadi S. Karlsson J.
Rasheed S. Characterization of a newly derived human sarcoma cell line HT Cancer. For details, please refer to the Methods section. For the boxplots in b and d , the center line, and lower and upper bounds of each box represent the median, and first and third quartiles, respectively.
The lower upper whisker extends to smallest largest values no further than 1. All conditions were adjusted, and 10 million reads were used in d and e. We calculated the number of detected transcripts that exhibited expression changes of less than twofold against rdRNA-seq.
We also plotted the squared coefficient of variation CV 2 against expression levels to examine reproducibility. Consistently, we also confirmed that the read coverage of RamDA-seq against relative transcript length was most similar to that of rdRNA-seq Supplementary Fig. This result was consistent with previously reported results 24 , In addition, the fraction of exonic regions covered by the reads indicated that RamDA-seq covered a higher fraction of exonic regions than did the other methods in all length bins Fig.
PE: data from paired-end reads. The transcripts were sorted into bins represented by the number at the top of each panel according to transcript length. Each row represents a histone transcript. Each column represents a sample using the indicated scRNA-seq method. The points and error bars represent means and SDs, respectively. Each line represents a scRNA-seq method. The numbers in parentheses represent the number of transcripts. We first confirmed that RamDA-seq could specifically detect the expression of differentially expressed non-poly A transcripts, which were identified by bulk RNA-seq, at the single-cell level Supplementary Note 6 and Supplementary Fig.
Diffusion map analysis revealed variability within cells even at the same time points Fig. We identified such transcripts, including non-poly A transcripts Fig. The dynamically regulated non-poly A transcripts were spread in all clusters with various expression patterns, suggesting that non-poly A transcripts are involved in various cell functions.
Furthermore, reasoning that transcripts with similar expression patterns should share biological functions, we attempted to infer the potential functions of these dynamically regulated non-poly A transcripts by performing functional enrichment analysis of each cluster Supplementary Data 1 ; see Supplementary Note 8 for further discussion.
RamDA-seq analyses of cell differentiation. DC, diffusion component. The numbers in parentheses represent the number of cells. Rows are ordered and colored by clusters. Columns are ordered by pseudotime and colored by sampling time points. Smoothed values are transformed to Z -scores for each row.
Raw values are scaled from 0 to 1 for each row. The x -axis represents pseudotime. Thin, colored areas represent SDs. The numbers before and after the slash in the parenthesis represent the numbers of non-poly A transcripts and all transcripts included in each cluster, respectively.
Right Expression profile of the representative transcript for each cluster. Each black curve represents a fitted generalized additive model. The x -axes represent pseudotime. The upper heat map represents the read coverage at the single-cell level. The middle plots represent the coverage averaged for cells at each time point as well as those of rdRNA-seq and paRNA-seq.
Gene models are shown at the bottom. The arrowhead indicates the position of the qPCR primer. The read coverage was normalized to the average of all cells. The center line, and lower and upper bounds of each box represent the median, and first and third quartiles, respectively. Consistently, mapping data of single cells using RamDA-seq showed full-length transcript coverage for Neat Fig.
To assess whether the observed decrease was specific to the long isoform or common to the two isoforms, we compared the read coverage of the region common to both isoforms and the region specific to the long isoform.
Further studies are necessary to elucidate the potential biological significance of the observed dynamics of Neat1 isoforms. Collectively, these results indicate that many non-poly A transcripts are dynamically regulated and highlight the utility of full-length single-cell total RNA-seq for studying the dynamic regulation and potential functions of non-poly A RNAs.
RS is a multistep process of intron removal using cryptic splice sites within long introns 14 and was recently observed in vertebrates Due to the large number of intronic reads Supplementary Fig. The height of the sawtooth wave pattern was associated with the expression level of host genes at all time points Supplementary Fig.
Single-cell analysis of recursive splicing. The upper heat maps represent the RamDA-seq read coverage for each cell. The middle tracks represent the averaged RamDA-seq coverage for each sampling time point.
Gene models and nucleotide sequences around recursive-splicing sites are shown at the bottom. The p -values of F-tests are indicated.
We fitted linear regression models against the RamDA-seq read coverages of each single cell in Cadm1 , Robo2 , and Magi1. RS was detected in a subpopulation of cells 71 of cells for Cadm1 , 12 of 54 cells for Magi1 , and 1 of 1 cell for Robo2 although many cells in which RS was not detected also appeared to show the sawtooth pattern Supplementary Fig.
The monotonically decreasing pattern was also observed even when we filtered cells with a more stringent threshold of intronic reads Supplementary Fig. These observations raise the possibility of cell-to-cell heterogeneity in RS. Therefore, further investigation is needed to reveal the mechanisms and significance of the observed heterogeneity in RS.
Next, we examined whether RamDA-seq could be used to detect eRNAs in single cells with differentiation time-series data. Similar trends were observed when we used non-poly A eRNAs with enhancer-like histone modifications Supplementary Fig. Consistently, bimodal peaks are observed around enhancers in the read coverage of total RNA-seq 7.
On the other hand, the distribution of the read coverage around random genomic regions was steadily low across all time points Fig. Single-cell analysis of enhancer RNAs. The detection of each enhancer is called when the TPMexceed 0. The shaded areas represent SDs. The number of detected enhancers top left and enriched known DNA motifs of transcription factors top right are shown in Venn diagrams.
The numbers in parentheses represent the number of eRNAs. Previous studies have demonstrated an enrichment for cell type- and condition-specific transcription factor DNA-binding motifs at active eRNA loci 10 , 30 , which prompted us to search for enrichment of motifs of cell type-specific transcription factors.
In parallel, the same analysis was performed using rdRNA-seq. Having validated our genome-wide eRNA analysis with RamDA-seq, we searched for eRNAs showing variations according to pseudotime of the cells and performed hierarchical clustering Methods section. Notably, GATA4, a late PrE marker 34 , was enriched in one transiently upregulated cluster cluster 3 and in the late upregulated cluster cluster 5 , which suggests that these clusters represent enhancers that function in the differentiation into PrE.
Altogether, we conclude that RamDA-seq can detect eRNA activity associated with cell type-specific regulation as well as the potential regulator of eRNAs within a subpopulation of single cells. RamDA-seq revealed many known and unannotated non-poly A transcripts that were dynamically regulated as differentiation progressed, including Neat Fig. Moreover, RT-RamDA improves sensitivity and reproducibility by eliminating the necessity for PCR amplification, which often results in amplification bias.
NSRs contributes to the full-length transcript coverage and high efficiency of capturing poly A and non-poly A RNAs by multiple priming. These characteristics contribute to RT efficiency and the cost reduction of oligo primer synthesis. There are some limitations to this method.
Therefore, it is difficult to perform pre-indexing high-throughput sequencing and molecule counting using UMIs. To address this issue, modifying the NSRs is necessary, for example, by adjusting the annealing temperature of NSRs to prevent misannealing or removing the complementary sequences annotated as rRNAs in RepeatMasker.
It is also important to achieve strand-specific sequencing in RamDA-seq to distinguish overlapping transcripts. Full-length total RNA-seq from single cells will be valuable to many studies using rare cells. Many biologically and clinically important cell types are rare and are often found in heterogeneous cell populations.
Thus, these cell types require single-cell approaches, and accumulating evidence suggests the importance of full-length total RNA-seq in single cells. Enhancers account for cell type-specific expression 28 , 29 and diseases, and their activity and potential regulators can be inferred by eRNAs 10 , Based on our results, single-cell analyses using RamDA-seq could be useful for identifying novel biomarkers and drug targets, non-canonical and aberrant RNA-processing events, and active enhancers and their potential regulators in rare cells.
Neat1 is an architectural component of paraspeckle nuclear bodies 38 , which regulate gene expression via capture of A-to-I edited mRNAs 39 and transcription factors 40 , and is required for corpus luteum formation and establishment of pregnancy in mice The long non-poly A isoform Neat , not the short poly A isoform Neat , is essential for the formation of paraspeckles Although the two isoforms are transcribed from the same promoter, they show different expression patterns, and Neat is expressed only in a small subpopulation of cells in adult mouse tissues Therefore, distinguishing the expression of the two isoforms of Neat1 at the single-cell level is critical for studying their functions.
These results suggest that RamDA-seq could be beneficial for investigation of temporal and spatial expression patterns of long non-poly A RNAs in single cells. Unexpectedly, we observed cell-to-cell heterogeneity in read coverage patterns around RS sites, suggesting that some cells showed RS, and other cells showed normal splicing Supplementary Fig.
These results indicate that RamDA-seq can detect cell-to-cell heterogeneity in RS and could help to address the mechanisms and relationship between transcription and splicing. Toward these goals, several important challenges remain.
Given that some cells in which RS was not detected showed weak sawtooth patterns, RamDA-seq highlights the limitation of the current linear regression model used to detect RS in this study and the need for further improvement in computational methods to robustly detect RS using single-cell data.
Another challenge is to experimentally and computationally distinguish biological and technical variabilities in RS at the single-cell level. We will address these challenges in the future. Recently, droplet-based scRNA-seq methods, which can sequence a very high number of cells at once, have been proposed 43 ,
0コメント