Capillary electrophoresis-mass spectrometry (CE-MS) has long occupied an intriguing position within proteomics. Its strengths are well established: exceptional separation efficiency, low sample consumption, and an ability to distinguish closely related proteoforms that can challenge conventional chromatographic approaches. Yet despite decades of development, concerns around robustness, reproducibility, and practical implementation have often limited its adoption.
Kevin Jooß, Assistant Professor of Bioanalytical Chemistry at Vrije Universiteit Amsterdam, believes the CE-MS conversation may be starting to change. In a recent international benchmarking study spanning 12 laboratories, he and colleagues assessed whether the technique has reached the robustness and reproducibility needed for broader adoption in top-down proteomics. The results suggest that many long-standing barriers to implementation are being overcome, while highlighting CE-MS as a complement to established liquid chromatography-mass spectrometry (LC-MS) approaches.
Here, Jooß discusses the rationale behind the study, the lessons from coordinating an international laboratory comparison, and the road ahead for CE-MS in top-down proteomics.
For readers less familiar with top-down proteomics, could you briefly explain what intact proteoform analysis enables – and why separation performance remains such a bottleneck in this space?
Top-down proteomics analyzes proteins in their intact form, allowing you to observe complete proteoforms directly – including combinations of post-translational modifications, sequence variants, and truncations. This avoids the “inference problem” of bottom-up proteomics, where proteins must be reconstructed from peptide fragments and important relationships between modifications can be lost. Because fragmentation occurs at the intact protein level, all fragment ions, in principle, can be directly traced back to their parent proteoform, providing unequivocal molecular information on the exact proteoform present.
A major challenge, however, is separation. Intact proteoforms often differ by only very small chemical changes, such as a single phosphorylation, while still having very similar physicochemical properties. At the same time, intact proteins are larger, more complex, and more prone to adsorption and peak broadening than peptides. Without effective separation before MS, overlapping spectra and ion suppression can quickly reduce sensitivity and make confident proteoform identification much more difficult.
What have been the main concerns holding CE-MS back, and why did you feel now was the right moment to reassess its readiness?
CE-MS has long been attractive because of its high separation efficiency, low sample consumption, and suitability for charged analytes such as proteins and peptides. However, adoption has lagged behind LC-MS due to a combination of practical and technical limitations. Historically, these included difficulties in coupling capillary electrophoresis (CE) robustly and reproducibly to mass spectrometers, maintaining stable electrospray at ultra-low flow rates, managing electrical decoupling, and working with lower loading capacity than liquid chromatography. The lack of standardized platforms, together with more fragile operation in earlier implementations, also made CE-MS harder for non-specialist laboratories to implement reliably.
Over the past decade, many of these barriers have been substantially reduced. Sheathless and improved sheath-flow interfaces, more stable capillary coatings, better voltage and injection control, and more robust commercial implementations have all strengthened reproducibility and ease of use. At the same time, increases in MS sensitivity mean that CE’s lower absolute loading capacity is much less limiting than it once was.
We felt it was the right moment to reassess CE-MS because the bottlenecks are shifting. For challenging applications such as top-down proteomics and proteoform-resolved analysis, separation performance is increasingly the limiting factor rather than raw sample throughput. In that context, CE’s intrinsic efficiency and ability to resolve closely related proteoforms become a major advantage rather than a niche feature.
Could you briefly walk us through how this inter-laboratory study was designed?
At the beginning of the project, Alexander Ivanov, Liangliang Sun and I had extensive discussions about how to design the study. Rather than performing a more traditional inter-laboratory comparison where every group tries to replicate an almost identical setup, we deliberately chose a different approach. We wanted to embrace the diversity of CE-MS platforms and demonstrate that reproducible, information-rich proteoform data can be obtained across a range of instrument configurations, interface designs, and laboratory environments.
Single-laboratory studies are valuable because they show what is possible under highly optimized conditions, but multi-laboratory benchmarking provides a much better sense of what is achievable in practice. By working across different CE-MS setups and levels of user experience, we could look beyond analytical performance alone and ask whether the technology is mature, transferable, and usable in real laboratory settings.
The study also provides the community with a practical reference point for future work, including protocols and benchmarking metrics that can be transferred across laboratories. We hope this helps build confidence in capillary zone electrophoresis–mass spectrometry (CZE-MS) workflows, supports future standardization, and makes CE-MS easier for more laboratories to adopt in top-down proteomics and proteoform analysis.
What was the greatest analytical or practical challenge in coordinating and executing a study of this scale – and how did you overcome it?
One of the greatest challenges came right at the beginning, in designing the study itself. We had extensive discussions about which sample sets to include, and how to balance practical feasibility with reproducibility and scientific rigor. A key priority was to minimize variability introduced during sample preparation, while still ensuring that the study captured a sufficiently diverse and challenging set of proteoform targets.
In the end, we decided to use commercially available samples, including protein mixes and intact protein cell lysates. This helped reduce pretreatment steps and improve reproducibility across laboratories, while still preserving enough analytical complexity to make the comparison meaningful.
From a practical standpoint, coordinating a study of this scale also depended on clear and continuous communication. Regular update meetings were important for keeping the project aligned, tracking progress, addressing problems early, and maintaining consistency across the different stages of the work.
Overall, I think the combination of careful experimental design and sustained coordination was crucial to making the study work.
Which results surprised you most, or challenged long-standing assumptions about CE-MS robustness?
We were surprised initially by the level of reproducibility achieved after applying a relatively simple internal standard-based migration time correction. We had expected variability to remain more pronounced – especially given the diversity of CE-MS platforms involved, which covered most commercially available systems. However, after normalization, we consistently observed migration-time relative standard deviations, or RSDs, well below one percent across the majority of participants. This clearly demonstrated that, with an appropriate correction strategy, CE-MS can achieve a very high degree of inter-laboratory reproducibility – more than is often assumed.
Another interesting observation concerned proteoform identification performance. Although run-to-run repeatability was somewhat lower for CE-MS than for LC-MS, three CE-MS injections per sample were enough to surpass LC-MS in the total number of identified proteoforms. This highlights the strong potential of CE-MS as a complementary platform for deep proteoform characterization.
How should researchers think about integrating CE-MS alongside established LC-based workflows, rather than viewing them as competitors?
I would strongly advocate for treating CE-MS and LC-MS as complementary rather than competing techniques, as both offer distinct and valuable advantages for proteoform analysis.
While there have been efforts to develop online 2D configurations combining LC and CE separations, these approaches are still not mature enough for routine implementation in standard laboratory workflows. An alternative is to use LC-based fractionation, such as SEC, followed by reinjection into a CE-MS system. This can provide additional depth, but it also adds workflow complexity.
In practice, the most effective and broadly applicable strategy today is to analyze samples using both CE-MS and LC-MS in parallel, when possible. This leverages the orthogonality of the two techniques without requiring complex integrated setups. A key practical advantage of CE-MS in this context is its extremely low sample consumption – typically only a few nanoliters per injection. This makes it particularly well-suited to complementary measurements when sample availability is limited, allowing CE-MS to be performed before LC-MS analysis while consuming very little material.
Taken together, this complementarity can significantly expand proteome coverage and increase confidence in proteoform identification, providing insights that neither technique could deliver alone.
What still needs to happen for CE-MS to become a routine tool in top-down proteomics?
In my view, we are not quite there yet for CE-MS to be considered a fully routine tool, but we are clearly much closer than we were a few years ago.
The main remaining barriers are not necessarily fundamental – they’re more practical and ecosystem-related. On the technical side, continued improvements in robustness, long-term interface stability, and ease of operation are still needed, particularly to reduce the level of expert intervention required during routine use. Standardization is just as important, especially around data processing workflows, alignment strategies such as migration time normalization, and consistent reporting of performance metrics. Our study showed that once harmonization strategies are applied, reproducibility across laboratories can already be very strong, highlighting standardization as a key enabler of broader adoption.
On a community level, inter-laboratory studies such as ours play an important role in making CE-MS more visible and approachable to the broader scientific community. But education is just as critical. Many students and early-career scientists, who will soon become the next generation of researchers in academia and industry, are simply not exposed to the capabilities of CE-MS for proteoform separation and characterization. Increasing the visibility of CE-MS in undergraduate and graduate curricula would help build a stronger foundation of trained users.
Hands-on training is also important. Dedicated events and summer schools, such as the CE-MS summer school organized by Liangliang’s group at Michigan State University – now in its third iteration – are extremely valuable for giving researchers practical experience and building confidence with the technique.
Overall, I would say the technical foundation is already strong, but routine implementation will depend on continued improvements in usability and harmonization, alongside sustained efforts to educate the next generation of users.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
