To all articles

Applications of Proteomics

Single-molecule analysis of tau proteoforms expands possibilities in neurodegenerative disease research

Jennifer Tsang

November 12, 2026


The tau protein is a microtubule associated protein with critical roles in Alzheimer’s disease and related dementias. Its normal function is to stabilize microtubules in axons and it is also involved in neuronal development and maintenance. But things go awry when tau aggregates and forms tangles. It’s a complicated protein with six isoforms, which can contain multiple post-translational modifications that can co-occur on multiple residues at the same time, producing a potentially very large number of unique proteoforms. What’s largely unknown is which proteoforms are produced, which are associated with healthy tau function, and which proteoforms are associated with disease. 

In a recent Nature Methods publication, “Large-scale single-molecule analysis of tau proteoforms,” the Nautilus team joined collaborators at Genentech, the Neural Stem Cell Institute, and The Ronald Loeb Center for Alzheimer’s Disease at Mt. Sinai to untangle some of the biology behind tau using Iterative Mapping. Below, find a synopsis of the key takeaways from the paper as discussed in a recent Translating Proteomics podcast episode and discover what this work may mean for the future of neurodegenerative disease research. 

Takeaway #1: The ability to probe billions of intact proteins at the single-molecule level reveals previously hidden biology 

Methods like Iterative Mapping enable researchers to study proteins in ways that have been difficult before.

“For decades, the field of Alzheimer’s has really measured either total tau, individual isoforms, or the average of abundances of certain phosphorylation sites,” says Andreas Huhmer, Senior Director of Scientific Affairs and Alliance Management at Nautilus. “But those measurements do not reveal modifications that occur together on the same tau molecule. This study shows that combination matters.” By probing intact proteins at the single-protein level, Iterative Mapping can help researchers characterize individual proteins at scale. “This is now not just a theoretical construct, but this is a very practical, accessible technology,” says Parag Mallick, Co-Founder and Chief Scientist at Nautilus.  

One example of how single-molecule analysis reveals a previously hidden aspect of Alzheimer’s disease is that it shows there is likely more remodeling in brain tissue rather than a progressive increase in phosphorylation as disease progresses. The field has generally viewed the accumulation of phosphorylation as a sign of disease progression. In the publication, the Voyager™ Platform uncovered signs that this wasn’t exactly the case. Abundant proteoforms with 1-2 phosphorylated sites decreased as the disease progressed. At the same time, rarer species with three or four phosphorylations were enriched in cognitively impaired populations. A finding like this could be obscured by measurements of total tau or individual phosphorylation sites, which do not show when multiple modifications occur on the same molecule. 

Takeaway #2: Co-ocurrence shows tau modifications are not random 

Methods like bottom-up mass spectrometry break proteins into peptides before analysis. This prevents researchers from seeing if specific modification found on different peptides co-occur on the same protein molecules. In this study, the Voyager Platform helped scientists identify co-occurring modifications on tau.  

The paper found that a substantial fraction of molecules had 2-6 co-occurring phosphorylations on a single molecule. Being able to see this information and quantify phosphorylations on intact proteins is helpful because it shows that modification is not random: some combinations of phosphorylation occurred disproportionately.

Quantifying proteoforms generated in a single patient, the authors found evidence that a quadruply phosphorylated proteoform likely arose from a distinct subset of previously doubly and triply phosphorylated proteoforms. If phosphorylation were random, one would expect all possible doubly phosphorylated precursors to come in equal proportions and all triply phosphorylated percusors to come in equal proportions. This was not the case. Instead, three of the four triply phosphorylated proteins and two of the six double phosphorylated isoforms had higher than expected abundance. This indicates that phosphorylation is an ordered process where a modification at one site may allow modifications to occur at a different site, for example, by changing the protein shape. 

Takeaway #3: Different tau biology captured in different model systems 

The paper also looked at how tau proteoforms behaved in different model systems: cellular models, organoid models, mouse brains, and adult human brains and found that they have different proteoform landscapes. For example, induced neurons, which are created by converting non-neuronal cells into neurons, contained more immature forms of tau while tau from human brains contained less phosphorylation in general but had a more mature isoform distribution. “The takeaway here, clearly, is we need to stop treating every neural model system [as if it has] the same tau biology,” says Huhmer.  

Being able to examine proteoform distributions and characteristics at the single-protein level across different model systems provides researchers with the ability to assess whether current models faithfully capture human tau biology. It can also help biologists develop new models benchmarked against known tau proteoform biology, with the best models recapitulating proteoform profiles found in patients. 

Understanding selective vulnerability 

With technologies like the Voyager Platform, scientists can see single proteins in a more accessible way and answer questions that weren’t approachable before. While not part of the study, another avenue of exploration opened by Iterative Mapping is the study of selective vulnerability, where specific brain regions are affected by neurodegenerative disease while others are spared.

In Alzheimer’s disease, tau pathology usually begins in the entorhinal cortex and then spreads. This spreading isn’t solely about physical proximity as studies have shown that neural connectivity plays a role. It’s also possible that different brain regions contain different distributions of tau proteoforms that are more or less prone to adopting disease-associated profiles.

For example, the proteoforms normally found in vulnerable regions may be just one or two phosphorylations away from more aggregation-prone forms. Environmental stresses, inflammation, and metabolic changes may shift their phosphorylation dynamics resulting in tau proteoform pools dominated by aggregation-prone species. Understanding this on a single-protein level can help scientists develop ways to intercept early or come up with preventative therapies, for example by blocking a phosphorylation step so that aggregation-prone tau modification patterns are not generated. 

What does this mean for studies that look at tau role in neurodegenerative diseases? 

In addition to Alzheimer’s disease, tau is involved in other neurodegenerative diseases such as frontotemporal degeneration (FTD) and progressive supranuclear palsy (PSP). Patients often have more than one of these tauopathies at the same time which can complicate diagnosis and treatment. For instance, a treatment meant for Alzheimer’s disease may not be effective for FTD as shown in a 100-patient clinical trial that found that the Alzheimer’s drug memantine did not benefit patients with FTD. With an increased and more comprehensive understanding of the distribution of modifications found on tau proteins, scientists may better differentiate between different tauopathies, helping them develop better diagnostics and treatments.  

“I firmly believe that in the proteoform landscape is the information to differentiate these tauopathies.” – Parag Mallick 

Moving beyond neurodegenerative diseases 

While this work has laid the foundation for other studies of tau in neurodegeneration, its results suggest the Voyager Platform can help biologists study other proteins that may have many proteoforms such as oncology proteins like AKT1, EGFR, and P53. The human genome contains about 20,300 genes, with alternative splicing, sequence variations, and post-translational modifications generating potentially millions of distinct proteoforms. Being able to analyze them as intact proteins on the single-molecule level may be a game changer for our understanding of functional biology. 

Interested in using Iterative Mapping in your work? Join our Early Access Program.

Share this Article

Stay up-to-date on all things Nautilus

World-class articles, delivered weekly

MORE ARTICLES

Stay up-to-date on all things Nautilus

Subscribe to our Newsletter