Protein turnover information provides new insights and may guide researchers to more accurate Alzheimer's disease models

In business, the word “turnover” can be negative, implying an inability to keep a team together over time. But turnover, and the changes that come with it, is not inherently bad; it is actually fundamental. Managing turnover requires balancing the expected and unexpected — something our cells constantly do with proteins.  

Proteins are cellular workhorses: They move to the job site, do their job and are eventually degraded. However, the timeline for this protein lifecycle varies greatly, with some jobs requiring longer-lived proteins. Turnover is a key component of the process used to keep proteins at the optimal levels to support cellular needs. Dips or surpluses in protein levels can contribute to disease. For this reason, Junmin Peng, PhD, St. Departments of Structural Biology and Developmental Neurobiology, has developed methodologies and tools to monitor protein turnover in cells and tissues. 

“To truly understand protein degradation, we need to measure three components: protein abundance, protein ubiquitination levels and, finally, the protein turnover rate,” explained Peng. “We’ve understood the first two components for a long time but didn’t quite fully grasp protein turnover.” 

Protein abundance (how many proteins there are) must be regulated to ensure a job is done at the appropriate time, extent and location. Protein ubiquitination is the process by which a ubiquitin tag is attached to a protein that is not needed, providing instructions for the cell’s recycling systems to degrade it. While protein abundance and ubiquitin levels can be directly measured, protein turnover is trickier.  

This is because degraded proteins’ amino acids are not discarded but recycled. “Once a protein is degraded into a pool of amino acids,” said Peng, “they are recycled internally because the cell doesn’t just throw away the valuable amino acid building blocks.”  

With this in mind, researchers estimate protein turnover through isotope labeling, where they incorporate an amino acid tracker into the protein supply chain. However, this only provides an apparent turnover rate, as the existing, recycled amino acids are still in the system. 

Peng wanted to improve this method to find the true rate, so he turned to mass spectrometry and mathematical modeling. His lab champions a novel mass spectrometry analysis pipeline called JUMPt, based on differential equations, that allows more accurate processing of protein turnover data. 

Finding the truth in protein turnover 

To thoroughly study protein turnover across different tissues, Peng turned to a collaborator, Yansheng Liu, PhD, Yale University School of Medicine, to bring a complementary skillset to the project. “In proteomics, there are two major approaches for protein profiling: data-independent acquisition (DIA) and tandem mass tagging (TMT),” Peng explained. “We did the TMT assays, and then we sent the exact same samples to Dr. Liu, who did the DIA assay. This gave us two independent data sets to validate each other’s work.”  

With all the pieces in place, Peng and Liu kicked off their collaboration. The teams measured protein abundance and turnover rates across eight mouse tissues and nine brain regions and compiled a comprehensive atlas with information for 11,171 unique proteins called Tissue-PPT. The work, recently published in Cell, triples the number of documented protein turnover rates in mammalian tissues.  

By analyzing correlations between RNA levels, protein abundance and protein turnover rates, the researchers determined that proteins that work together as part of a “complex” often have similar turnover rates. This revealed a level of coordination that had not been appreciated before. They also found that protein turnover regulates many processes, including making new proteins (translation). Finally, the researchers teased out the importance of phosphorylation (a protein modification that can affect protein function) in the turnover of many proteins. 

Peng believes this atlas will be a watershed moment for the field. “Currently, we are analyzing protein turnover at the bulk level, but this approach could be extended to study how protein turnover changes under different stress conditions or disease states, even at the level of individual cells,” he said. “This work is the first large-scale analysis of its kind and sets the foundation and tone for future research in this area.” 

Alzheimer’s disease models come up short 

The Tissue-PPT dataset showed that phosphorylation regulates the stability of proteins involved in neurodegeneration, such as tau. Neurodegenerative diseases such as Alzheimer’s disease are of particular interest to the Peng lab. Two proteins, β-amyloid peptide and tau, are found to accumulate in Alzheimer’s disease, but they do so by different mechanisms. The β-amyloid peptide accumulates due to slow turnover, whereas tau accumulates as a result of the aggregation of hyperphosphorylated forms. 

In a 2020 study published in Neuron, Peng published a comprehensive exploration of the proteomic landscape of the human brain in Alzheimer’s disease. Mouse models are invaluable to understanding human diseases and testing potential treatment avenues. A number of different mouse models have been developed to study Alzheimer’s disease, but none of them seem capable of fully capturing the disease. To understand how those different mouse models relate to the human disease, the Peng group used a proteomics approach in a new study published in Nature Communications. They found that each mouse model only recaptures about 20% or less of the molecular events during Alzheimer’s disease progression. When data from four mouse models were combined, the findings still only reflected 40% of the molecular events in the human disease. 

“Each mouse model has its own strengths and advantages, capturing specific molecular events that take place in Alzheimer’s disease, but even when combined, we are still missing over half of what occurs in humans,” Peng said. “This highlights the limitations of mouse models: they simply cannot fully replicate the complexity of human biology.” 

Protein turnover offers new angle to interpret RNA-protein inconsistency in Alzheimer’s research 

Another intriguing finding from this study is that protein levels cannot be accurately inferred from RNA levels. The Peng lab compared RNA and protein data in human and mouse samples. They found that 30% of the data was inconsistent between RNA and proteins. They proposed that the inconsistencies lay in how fast these proteins degraded.  

While RNA levels determine most protein abundance, protein turnover is also a powerful regulator, as demonstrated by the new Tissue-PPT atlas. “We analyzed protein turnover for all the proteins we could detect in the mouse models,” Peng said. “We found that some key components, which we believe accumulate in the mouse, are not regulated by RNA. Instead, they have slower protein turnover rates.” 

Many of these proteins were aggregated in amyloid plaques, the hallmark of Alzheimer’s disease. This led the team to name this collection of proteins the “amyloidome” — essentially the proteome of amyloid plaques. 

Rather than viewing this as bad news for Alzheimer’s research, Peng is optimistic that this will guide researchers in selecting the best model for their research. “If you want to study a certain pathway or specific component using a mouse model, our data will allow you to select the one that can accurately reflect that specific pathway,” he stated. 

The findings from this body of work will help ensure that the intricate dance of protein turnover — and its implications for disease — can be considered and studied more accurately moving forward.