Sensitive high-throughput single-cell RNA-Seq reveals within-clonal transcript-correlations in yeast populations

Single-cell RNA-seq has revealed extensive cellular heterogeneity within many organisms, but few methods have been developed for microbial clonal populations. The yeast genome displays unusually dense transcript spacing, with interleaved and overlapping transcription from both strands, resulting in a minuscule but complex pool of RNA protected by a resilient cell wall. Here, we have developed a sensitive, scalable, and inexpensive yeast single-cell RNA-seq (yscRNA-seq) method that digitally counts transcript start sites in a strand- and isoform-specific manner. YscRNA-seq detects the expression of low-abundant, non-coding RNAs, and at least half of the protein-coding genome in each cell. Within clonal cells, we observed a negative correlation for the expression of sense/antisense pairs, while paralogs and divergent transcripts co-express. Combining yscRNA-seq with index sorting, we uncovered a linear relationship between cell size and RNA content. Although we detected an average of ~3.5 molecules/gene, the number of expressed isoforms are restricted at the single-cell level. Remarkably, the expression of metabolic genes is highly variable, while their stochastic expression primes cells for increased fitness towards the corresponding environmental challenge. These findings suggest that functional transcript diversity acts as a mechanism for providing a selective advantage to individual cells within otherwise transcriptionally heterogeneous populations.