NEW YORK (GenomeWeb News) – In a paper appearing online today in Science, researchers from Harvard University and the University of Toronto report that they have assessed single molecule proteome and transcriptome profiles in individual Escherichia coli cells.
"This article is where single molecules meet systems biology," senior author Sunney Xie, a chemistry and chemical biology researcher at Harvard University, told GenomeWeb Daily News. "This is the first time that a proteome has been characterized with single molecule sensitivity for any organism."
Using single-molecule imaging of E. coli strains from a yellow fluorescent protein fusion library, the team measured and compared protein and mRNA levels in individual E. coli cells. They found that mRNA levels for specific genes did not correlate with protein levels for products of the same genes in single E. coli cells — apparently due to the stochastic nature of gene expression and differences in how long mRNA and proteins last in the cell.
The study relied on a yellow fluorescent protein, or YFP, library created by tweaking an E. coli mass spectrometry library developed by University of Toronto proteomics and bioinformatics researcher Andrew Emili.
By fusing a different gene to a yellow fluorescent coding sequence in each E. coli strain, the researchers could systematically look at protein levels by imaging individual cells from each strain and quantifying the fluorescence levels.
Meanwhile, they simultaneously measured mRNA levels using single molecule FISH targeting the YFP gene itself — an approach that bypasses potential complications introduced by differences in the secondary structure of bacterial gene transcripts, Xie noted.
"We have provided quantitative analyses of both abundance and noise in the proteome and transcriptome on a single-cell level for [the] Gram-negative bacteria E. coli," the team wrote. "Given that some proteins and most mRNAs of functional genes are present at low copy numbers in a bacterial cell, the single-molecule sensitivity afforded by our measurements is necessary for understanding stochastic gene expression and regulation."
When the researchers characterized the distribution or "noise" in the system, they found that this noise was inversely proportional to the gene expression level at low levels but that this noise remains constant at high levels.
At the lower expression levels, Xie explained, the system is dominated by intrinsic noise — in other words, the stochastic nature of individual molecules' biochemical reactions.
Overall, the team found that all of the E. coli genes tested were expressed to some extent — some because they are functionally necessary in the cell and some due to "leakage" in the system, Xie said. In single cells, he explained, it wasn't possible to compare gene expression and protein levels for genes expressed at extremely low levels in single cells.
But they could compare these levels for more highly expressed genes, showing that mRNA levels and protein levels for each gene did not correlate with one another in single cells.
"What we found, surprisingly, is that that in the same cell, the mRNA copy number and protein copy number for the same gene are not at all correlated," Xie said. "This has something to do with the fact that mRNA is short-lived, at least in bacteria, whereas proteins are long-lived."
Because gene expression tends to be a stochastic process, he explained, mRNA shows up in bursts in the cell. And since mRNAs have relatively short lifetimes compared to proteins, for example, protein levels may be high after mRNA transcripts have been degraded. In other words, Xie said, mRNA levels reflect a cell's most recent history, while protein levels reflect previous cellular events.
"This result highlights the disconnect between proteome and transcriptome analyses of a single cell," he and his co-authors wrote, "as well as the need for single-cell proteome analysis."
Based on the findings in E. coli, Xie also cautioned that care is needed when trying to decipher protein levels based on mRNA levels in single cells. Such problems can be overcome by averaging many cells or making a movie of the cells over time, he added.
"This provides a cautionary note for people who want to do single cell mRNA-profiling, at least for bacteria, by RNA-Seq," Xie said. "A large number of cells, that's no problem — if you average many cells."
The study may have other implications as well. In a perspectives article appearing in the same issue of Science, for instance, Sanjay Tyagi, a researcher with the University of Medicine and Dentistry of New Jersey's Public Health Research Institute, noted, "Now that there is reliable knowledge of the levels of expression and the underlying variation of mRNAs and proteins for a considerable portion of the E. coli genome, it is possible to explore how noise propagates along gene expression pathways, in which the amount of one protein can influence the expression of another."
"Investigators can also investigate how cells coordinate the expression of proteins that need to work together," Tyagi added, "such as multi-subunit proteins or proteins that serve within metabolic cycles."