This story originally ran on Dec. 8.
By Tony Fong
A US-Chinese team of researchers has identified differences in expression levels between proteins and mRNA between hibernating and non-hibernating arctic squirrels, suggesting "substantial post-transcriptional regulation of proteins during torpor-arousal cycles of hibernation."
The results, which appear in the Nov. 20 issue of Molecular & Cellular Proteomics, could help develop new "preventive and therapeutic approaches" for ailments such as trauma, cardiovascular disease, stroke, and ischemia/reperfusion injury.
In the study, the researchers wrote that hibernation "involves complex mechanisms of metabolic reprogramming and tissue protection." Studies into the process have looked primarily into changes at the mRNA level, but while these studies have identified "significant differences" in mRNA levels in hibernating and non-hibernating ground squirrels, the transcripts of mRNA "are protected while translation is inhibited during torpor" in the animals, the researchers wrote.
"Therefore protein variety and abundance can be very different from corresponding gene expression at the mRNA level, and differential protein expression may more directly reflect regulatory changes related to hibernation."
But because a protein database specific to hibernating species has not been available, large-scale proteomic studies into the area have lagged behind. Leveraging the recently sequenced genome of the closely related 13-lined ground squirrel, however, the authors built a protein database of the arctic ground squirrel and identified 3,104 unique proteins from liver tissue. These included 517 with "significant" differential expression in animals sampled after at least eight days of "continuous torpor" within five hours of a spontaneous arousal episode, and one to two months after the end of hibernation.
In their approach, the researchers used a label-free LC-MS/MS shotgun strategy based on spectral counting, which they said is the first time such a strategy has been applied to a hibernating species.
Prior proteomics studies into squirrels have relied on 2D gel electrophoresis, which enabled researchers to identify proteins that express differentially in active animals compared to those in hibernation. But 2D gel technology, according to the authors of the MCP study, has "intrinsic problems such as limited coverage, low sensitivities, and unidentifiable spots."
By applying a high-throughput approach, the team screened the differentially expressed proteins and genes, which may begin to explain how animals that hibernate are protected during that state. And that, in turn, can provide insight into human disease and novel therapies, Jun Yan, the corresponding author on the study, told ProteoMonitor by e-mail.
"We hope to test the functions of the interesting proteins that we identified in our study and their application in human treatment in the future," Yan said. "But before that, one has to have a basic understanding of the molecular mechanism of hibernation because hibernation involves coordinated changes of hundreds or thousands of proteins."
Yan is an independent junior group leader at the CAS-MPG Partner Institute for Computational Biology at the Shanghai Institutes of Biological Sciences.
In their research, the scientists studied liver tissue from 12 arctic ground squirrels in three physiological stages: late torpor, or LT, defined as after eight to 12 days of continuous torpor during the winter; early arousal from torpor, or EA, during the winter; and as a non-hibernating control, post-reproduction in "euthermic, post-hibernation, and post-reproductive animals,” or PR, sampled in May and June.
The three stages were chosen because in previous gene-expression studies, they showed the most significant differences in across-group comparisons. The researchers chose to study liver "due to its important role in energy metabolism, and the large amplitude of change in metabolic rate that hibernators display," they wrote.
Key to their work was the construction of their protein database for the ground squirrel, based on the newly available sequence of the 13-lined ground squirrel genome. The genome from the Ensembl database provided the researchers with 14,830 unique 13-lined ground squirrel proteins. They then mapped all human and mouse ReqSeq mRNA sequences onto the 13-lined ground squirrel genome. Removing overlaps with the Ensembl genome, they were able to obtain 624 additional ground squirrel proteins.
By mapping onto the 13-lined ground squirrel all available expressed sequence tag sequences of three closely related ground squirrel species — the 13-lined ground squirrel, the golden-mantled ground squirrel, and arctic ground squirrel — they obtained 473 more proteins.
Finally, they clustered the expressed sequenced tags that "had failed to map" onto the 13-lined ground squirrel and obtained 3,370 additional proteins, for a total of 19,297 ground squirrel proteins for their database.
Using Sequest v. 3.1, the researchers identified 3,594 proteins with at least two unique peptide hits to arctic ground squirrels in their protein database. Of those proteins, 3,537 had homologous human gene symbols. Removing redundant proteins resulted in 3,104 unique proteins with at least two peptide hits.
[ pagebreak ]
Their database created, the researchers then did a quantitative analysis of protein expression across the 12 arctic ground squirrel liver samples by selecting 1,209 proteins that had spectral counts greater than or equal to two in every animal sample in at least one physiological state.
Applying a statistical method for determining protein quantitation called the power law global error model, combined with a signal-to-noise method, on normalized spectral abundance factors, the research team identified 517 proteins that showed "significant" differences in "at least one pair-wise comparison: 188 proteins between LT and PR, 279 proteins between EA and PR, and 325 proteins between EA and LT."
An alternative statistical method, Qspec, which is based on a Poisson model with generalized linear mixed effects, identified 216 proteins with "significant" differences in at least one pair-wise comparison —117 proteins between LT and PR, 98 proteins between EA and PR, and 82 proteins between EA and LT "showing significant differences by default parameters," the researchers wrote.
They then performed an mRNA analysis combining earlier work they had performed with new analysis for real-time PCR measurements for a total of 188 genes and compared their mRNA results with the LC-MS/MS data.
The LC-MS/MS and real-time PCR data shared 104 genes, the researchers wrote. In a comparison of the Pearson correlation coefficients and "their statistical significances between protein and mRNA levels across the same set of animals for each gene," 30 genes were "significantly correlated" between the two datasets.
In a comparison between the protein levels measured by LC-MS/MS and their earlier mRNA work using Illumina's BeadArray platform, they found 85 genes shared between the two analyses. "Thus there is overall good correlation between mRNA and protein levels among AGS liver genes," the study's authors said.
Next, they analyzed their data to identify possible changes in protein expression affiliated with specific biological processes. Their analysis of proteins involved in metabolic shift and tissue protective changes fell largely in line with what was already known from mRNA-based observations.
As expected, many of the proteins that were differentially expressed during hibernation were metabolic proteins. According to Yan, HMGCoA synthase was of particular interest. An important enzyme in ketone body formation in mitochondria, it was overexpressed during hibernation, which suggests "hibernators might rely on ketone body as a metabolic resource during torpor," Yan said.
Yan also singled out a group of proteins related to tissue protection during hibernation. Regucalcin, or senescence marker protein-10 was "significantly overexpressed" during hibernation. Other research has proposed RGN regulates calcium homeostatis in cells, suppresses cell proliferation, and promotes cell survival. "RGN shows decreased expression in the livers of aging rats," Yan said, and "may play an important role for tissue protection against damages in liver during hibernation."
The researchers also observed a "large number of proteins" that showed differential expression between EA and LT, many of which were not observed at the mRNA level. Proteins involved in protein translation and degradation, mRNA processing, and oxidative phosphorylation "were significantly overexpressed in EA compared to LT during torpor-arousal cycle," while changes at the mRNA level were not significant, they reported.
These include seven eukaryotic translation initiation factors, 13 ribosomal proteins, and 12 subunits of proteasome.
Proteins involved in oxidative phosphorylation and electron transport chain including seven units of ATP synthases, two subunits of cytochrome c oxidase, and seven subunits of NADH dehydrogenase were also significantly overexpressed in EA versus LT, but the mRNA levels of these proteins, measured by real-time PCR and Illumina BeadArray did not show significant changes between EA and LT, the authors said.
"This may indicate significant post-transcriptional modifications leading to rapid protein turnover, mRNA processing, and oxidative phosphorylation during the torpor-arousal cycle," they added.
According to the research team, their work may have implications for human disease. During hibernation, drastic changes occur to an animal's physiology, many of which are similar to changes in humans with diseases such as cardiovascular disease and stroke.
"However hibernators have evolved a remarkable ability to sustain such changes and tolerate the enormous stress during hibernation," Yun said. "Using hibernation as a model, one can learn how nature has solved these problems. That is exactly why we have undertaken the effort to screen the differentially expressed genes and proteins during hibernation in large scale."
In continuing work, the researchers hope to test the "functions of the interesting proteins" identified in their study "and their application in human treatment in the future," Yan said.
The researchers also said that miRNAs may represent an important post-transcriptional regulatory mechanism in hibernation. The discrepancy between mRNA and protein changes for many changes during the torpor-arousal cycle may be due to miRNA regulation. "Now we are trying to extend our study to profile miRNA expression changes during hibernation," Yan said.