The two contrasting procedures for discovering and identifying proteins in biological samples are known as the bottom-up and the top-down approaches. In the former, the mixture of proteins in the sample is treated with enzymes that break them up into their constituent peptides and the whole mixture is analysed by mass spectrometry. Although there are now some efficient methods for carrying this out, the production of so many peptides inevitably complicates protein identification.
The alternative approach is top-down, in which the proteins are analysed intact. To go down this route, most methods require a high degree of separation so that the individual proteins can be analysed separately by mass spectrometry. This is not without its challenges such as the incompatibility of the conventional buffers towards mass spectrometry and the poor overall resolution of some of the chromatographic procedures.
Ying Ge and colleagues from University of Wisconsin-Madison noted that all published separation protocols for mixtures of intact proteins involve up to two consecutive chromatographic steps and decided to see what difference a third dimension would make. They were encouraged by their earlier work using hydrophobic interaction chromatography (HIC) in which they identified ammonium tartrate as a novel salt which was compatible with mass spectrometry.
The three types of chromatography that they tested were HIC, ion-exchange chromatography (IEC) and reversed-phase chromatography (RPC). The latter two are often used together in conventional 2D-LC protocols for protein separation. For IEC a mixed-bed column containing equal amounts of weak cation-exchange and anion-exchange materials was employed as this type is known to provide better separation than a single-mode column.
The three separations were deployed in the order IEC, HIC and RPC and there were solid operational reasons for doing so. In IEC, there is an increasing salt gradient in the mobile phase during the separation and the relatively high final concentration is suitable for the start of the HIC run during which a high salt concentration is lowered. Coupling of HIC with RPC is a well-established protocol.
Using a tissue cell lysate, a 2D approach using IEC-RPC was compared with a 3D IEC-HIC-RPC approach, in which identical IEC and RPC/MS conditions were chosen in each method to aid the comparison.
In the 2D method, one particular IEC fraction was examined closely and it revealed some coeluting proteins when it was analysed by RPC/MS. However, in the 3D method, this same fraction was chromatographed by HIC to give 35 separate fractions, each of which was then analysed by RPC/MS. The overlapping proteins from the 2D process were now clearly separated and some additional proteins were detected that were not seen after 2D separation.
When the number of proteins identified using tandem mass spectrometry was taken into account, the improved performance of the 3D method was clear. The 35 HIC fractions produced from the IEC fraction yielded a total of 640 intact proteins. Some of the proteins were present in more than one fraction so the actual number of individual proteins with no repeats was 201.
The same IEC fraction analysed in the 2D approach by IEC-RPC/MS yielded only 47 proteins. Using longer gradients had some effect, taking the total number of non-redundant proteins from 47 with an 80-minute gradient to 63 and 67 with 3-hour and 7-hour gradients, respectively, but well short of the identification power of the 3D method.
Although the performance of the 3D system is good, the researchers said that it could become much better for several reasons. Firstly, the majority of the proteins identified had molecular masses lower than 30 kDa even though an initial SDS-PAGE screening had revealed larger ones. The failure was attributed to limitations in the mass spectrometer, particularly the falling signal-to-noise ratio with increasing mass. Initial separation of the high-molecular-mass proteins from those of lower mass by SEC would also help.
Secondly, it is well known that offline chromatography like that used in this system is susceptible to protein losses and switching to a fully online process would cut out the losses and increase sensitivity.
In addition, it would be more informative if all 35 IEC fractions were analysed subsequently by HIC-RPC/MS, but there often has to be a trade off between the extra information gained and the time that it would take. Combining some fractions would speed up the overall analysis time but lose some of the separation power.
However, even without any of these improvements, the 3D method provides a vast improvement over the 2D method for protein separation from biological samples and will help to expose new biomarkers of disease as hidden proteins come to light.