NEW YORK (GenomeWeb) – Utrecht University researchers have developed an optimized workflow for middle-down proteomics, but in the absence of appropriate protein digestion methods, practical implementation of the approach remains out of reach.
So suggested Utrecht Professor Albert Heck, who led a study published last week in Analytical Chemistry that looked at how conventional mass spec workflows might be optimized for middle-down proteomics work.
In this, Heck said, he and his colleagues were successful. But, he noted, they were unable to make significant headway on the key problem facing middle-down proteomics — the lack of a digestion method capable of reproducibly generating peptides in the desired size range.
As the name implies, middle-down proteomics aims to stake out ground between conventional bottom-up approaches and the less widespread but growing practice of top-down, or intact protein, proteomic analysis. The argument for the approach is that it would achieve some of the top-down community's goals of improving study of protein isoforms and post-translational modifications while avoiding the technical challenges inherent in measuring intact proteins on a proteome-wide scale.
The ideal size for middle-down peptides would be around 10 kD, Heck said. This would make them around 10 times the size of the peptides generated by conventional trypsin digestion but keep them well below the 20 kD size at which mass spec analysis becomes more difficult.
The issue, though, is how to consistently generate peptides of this size.
In the Analytical Chemistry study, Heck and his colleagues tested several approaches, using both the proteases Asp-N and Glu-C, as well as a non-enzymatic method in which they incubated proteins at high temperatures in diluted formic acid. Formic acid, the authors wrote, has been reported to cleave proteins at the C-termini of aspartic acid, which, they noted, suggested it could prove an effective digestion method for middle-down work.
Ultimately, though, none of these approaches produced peptides of the desired size. With the two proteases, the researchers generated peptides of an average of 1.5 kD, while with the formic acid approach they produced peptides averaging 1.9 kD. Both represented improvements over trypsin digestion, which typically produces peptides of around 1 kD, but they remained far off the desired 10 kD range.
"In theory you would expect the [peptides] would become much longer because the [Asp-N and Glu-C] proteases have less specificity [than trypsin], they don't cleave after every arginine and lysine," Heck said. "We had expected that we would form large number of middle-down peptides in a reproducible manner."
The researchers did manage with these digestion methods to generate peptides in the 10 kD range, but these were only a small proportion of the total peptides produced.
They met with more success in the portion of the study focused on optimizing mass spec workflows for middle-down protein detection, Heck said. Here, the notion was that methods developed for smaller tryptic peptides were biased against larger peptides in ways that limited their detection.
One of the keys here was adjusting the chromatography used to avoid loss of larger peptides. Specifically, the researchers moved from LC columns packed with particles with pore sizes in the 80 to 120 Angstrom range to particles with pore sizes around 300 A.
"We showed in the work that, indeed we get a better separation [at the larger pore size], and we get larger peptides going through the column," Heck said.
He and his colleagues also explored the use of different fragmentation techniques, looking to complement collision-induced dissociation (HCD) with electron-transfer dissociation (ETD) and EThcD, which combines ETD and HCD. They found that EThcD performed particularly well, making a peptide match for 55 percent of MS/MS spectra acquired from the Asp-N and Glu-C-digested samples, compared to 36 percent for both ETD and HCD.
EThcd also provided superior sequence coverage, giving median coverage of 95 percent, compared to 82 percent for ETD and 65 percent for HCD.
"Our hypothesis was that in the fragmentation and the detection efficiency, there is a bias towards small peptides if you use HCD," Heck said. "We showed that you need these other fragmentation techniques, especially for the larger peptides."
The researchers also adjusted the search techniques they used to allow for isotope deconvoluting of the MS/MS spectra, which in some cases they found improved identification of larger precursor ions.
"We showed some of the problems [in middle-down workflows], and we showed some potential solutions," he said. However, he suggested that their inability to generate reproducibly digest proteins into peptides of the desired middle-down size overshadowed the progress made on other portions of the workflow.
Heck added that while he continues to look for middle-down digestion methods, he sees little in the way of obviously promising approaches at the moment.
Given that the researchers were able to generate on the order of hundreds of middle-down length peptides, they could simply focus on these molecules in their analysis. "We can do the digest and then we could use size exclusion chromatography to only enrich for the large peptides," Heck said. "But," he added, "then we sort of throw away half of our proteome, and that's not what we want to do."
He noted that the paper's reviewers suggested they try a partial digestion, letting the protease work only long enough to make a subset of the total expected cleavages. However, Heck said, this would make reproducible digestions, and, therefore, reproducible analyses, impossible.
"I still think the solution will be to find the protease or a chemical way to cut proteins into 5- to 15-Kd peptides," he said.
Northwestern University researcher Neil Kelleher, who specializes in top-down proteomics but has also looked into middle-down approaches, agreed with Heck that there isn't currently a clear path forward on the digestion front.
"If someone could create a protease that created peptides in a distinct size range, then it could be great," he said. "Right now, I don’t see that reagent."
In a 2012, Nature Methods paper, Kelleher put forth the outer membrane protease T, OmpT, as a potential protease for middle-down work. OmpT is a di-basic protease, meaning that it recognizes not one but two amino acids and will cut only at sites where both are present, which typically leads to production of longer peptides compared to trypsin.
However, as Yury Tsybin, formerly an assistant professor at Ecole Polytechnique Fédérale de Lausanne and currently CEO of mass spec firm Spectroswiss, noted in a 2014 interview with GenomeWeb, OmpT is not entirely specific for di-basic sites. Additionally, it does not consistently create peptides small enough for middle-down analysis and does not give complete proteome coverage."
Speaking this week, Tsybin noted that the quest for a suitable digestion technique continues, but added that he and his colleagues have had some narrower success using middle-down digestion methods tailored to specific research purposes. For instance, he said they have developed a method using the enzyme Sap9 for de novo sequencing of antibodies.
"We have developed methods for data analysis allowing us to get almost complete sequence coverage in de novo fashion from a single digestion with Sap9 and a single LC run," he said, adding that this compares favorably to current antibody sequencing methods that require five enzymes.
Kelleher likewise noted that he does see "some great reagents for [middle-down] of specific subclasses of proteins" like antibodies.
That said, as one of the leading proponents of top-down proteomics, Kelleher would prefer to avoid chopping up proteins in the first place.
"I would, of course, skip the protease altogether," he said.