Although many credit Thermo's Orbitrap line as a “game-changer” for proteomics in general, Kelleher says the Elite is the first mass spectrometer in that line to have the speed and performance to make top-down practical. “The Elite changed everything about top-down on an Orbitrap. It really is a sea change.”
In part, he explains, that's because proteins in a mass spectrometer don't really behave like big peptides. They are harder to fragment and take longer to weigh accurately, for instance. “I'm here to tell you that a 40, 50-plus charge state, a big fat protein, is different than a little 2+ tryptic peptide,” he says.
Using an Orbitrap Elite (and a 12-T FTICR-MS), Kelleher's team in 2011 performed the most comprehensive top-down proteomics analysis to date: the identification, quantification, and characterization of some 1,043 proteins in 3,000 isoforms from human HeLa cells, along with similar-scale analyses on two other cell lines (1).We're looking for a few good columns
Kelleher's analysis required more than just a good mass spec though—the key was a comprehensive liquid-phase separation strategy that could ease the protein load on the mass spectrometer.
High-resolution instruments are relatively slow. The simpler the sample, the more time the instrument can devote to each protein. Liquid chromatography is the easiest way to do this, but unlike peptides, which are relatively simple to fractionate, intact proteins don't separate easily. Rather than resolving as sharp chromatographic lines, intact proteins tend to slowly extrude from columns, producing broad, complex peaks.
When it comes to top-down proteomics, says Catherine Fenselau, professor of chemistry and biochemistry at the University of Maryland, “the big hang-up” is chromatography. Steven Patrie, an assistant professor of pathology at the University of Texas Southwestern Medical Center, is pursuing one alternative. His team recently combined “superficially porous” reversed-phase resins with capillary columns for LC/MS (SPLC/MS). SP resins have a thin porous shell but a non-porous center. Such a configuration maintains protein surface interactions at the outer most layers of the resin, thus providing the speed of non-porous materials while maintaining the high loading capacity of porous ones.
Using one such column—a 75-micron capillary column packed with 5-micron C18 beads—Patrie and his team showed that hundreds of intact HeLa cell proteins from 10,000 to 50,000 cells could be characterized in as little as five minutes, with low attomole sensitivity (2). That's the kind of performance researchers get from bottom-up approaches, notes Patrie. “If peptides are being detected at an attomole level, and you can also detect a protein at an attomole level, why bother doing the digestion?” he says rhetorically.
Paša-Tolić's team devised what they call “a novel online two-dimensional liquid chromatography–tandem mass spectrometry platform” for histone analysis that couples reverse-phase HPLC with weak cation exchange hydrophilic interaction chromatography to separate histone populations first into families (eg, H2A, H2B, H3, and H4) and then by post-translational modification.
“[With] H4 it looks really pretty,” says Paša-Tolić. “You see zero to five different acetylation states. And then within each acetylation state, you have phosphorylations, methylations, and these are not particularly well separated, but you can still get MS/MS data on them.” From 7.5 micrograms of purified histones, her team identified some 700 unique post-translational isoforms.
Kelleher's analysis used an even more complicated, four-dimensional strategy, separating first by isoelectric point, then size, then hydrophobicity, and finally in the mass spectrometer itself.
The resulting data are beautifully resolved. In one figure, the team documented the shifting post-translational modifications in the HMGA1a protein during senescence. Phosphorylated forms slowly disappear while methylated forms multiply, and a dimethylated form of the protein arose during senescence that was not present in control cells.
Still, the overall study exhibited a strong bias against proteins greater than 50-kDa. “That's where top-down needs to go,” says Kelleher, “is to really do better at high mass.”