Microbore protein digestor: Hollow fibre reactor for online proteolysis

Enzyme reactors

The routine analysis of protein mixtures by the bottom-up proteomics technique requires that they are digested into their component peptides which are subsequently analysed by mass spectrometry. The digestion agents are generally some kind of enzyme, such as trypsin, chymotrypsin, pepsin or endoproteinase Lys-C. In some experiments, more than one enzyme is employed.

Conventional enzyme digestion can take several hours, following which the peptide products have to be collected from the mixture and analysed. In recent years, a number of research groups have devised quicker ways, mostly involving some type of custom-built enzyme reactor. These have included systems based on membranes and chips.

A new enzyme reactor has now been introduced by Korean researchers which they claim will perform better than previous types. Myeong Hee Moon and colleagues from Yonsei University, Seoul, and the Korea Research Institute of Standards and Science, Daejeon, designed a microbore system based on a hollow fibre in which the proteins are digested, separated from the reaction mixture and directed to a trap for online analysis by liquid chromatography-tandem mass spectrometry.

Porous hollow fibre

The key to the reactor is the porous hollow fibre, just 4.8 cm long with an internal diameter of 500 µm and a tiny volume of about 10 µL. The fibre was positioned inside a PEEK tube and one end of the fibre was blocked with epoxy resin so that peptides could only exit through the pores into the outer tube.

A protein solution was passed into the fibre. Any free peptides, salt and impurities present passed through the pores and were directed towards an external trap. The peptides were trapped here while the other compounds passed through. This initial step serves to concentrate the protein within the fibre and desalt the solution.

In the next step, a solution of the enzyme, also containing dithiothreitol to reduce the disulphide bonds in the protein, was passed into the fibre to begin the digestion. In optimisation experiments with bovine serum albumin, the best protein:enzyme ratio was 1:1. The reaction was allowed to proceed for 30 minutes while a basic buffer solution was passed continuously through to remove peptides as they were formed, thereby favouring further reaction.

The trapped peptides were eluted from the trap and directed towards the LC/MS system fitted with a nano-LC column. While the analysis proceeded, the waste enzyme and any residual protein were washed out of the fibre to prepare for the next digestion.

The system was also tested for the online enrichment of glycopeptides, using α-1-acid glycoprotein as an example. The protein was first digested offline and the resultant solution was mixed with concanavalin A, a lectin with which glycans containing multiple mannose groups form complexes. This mixture was injected into the hollow fibre membrane and any free peptides and uncomplexed glycopeptides passed through the membrane to be caught in the trapping column.

The complexed glycopeptides were then digested by injecting a solution of PNGase F which deglycosylates the N-linked glycopeptides. The peptides formed from this reaction exited through the pores and were collected in a second trap. The contents of each trap were analysed separately by LC-tandem-MS.

Accelerated online digestion and separation

One way to test the efficacy of the new reactor is through the sequence coverage. Using the optimum enzyme:protein ratio of 1:1, the sequence coverage for bovine serum albumin reached 91.3% with a total of 83 peptides detected. This was accomplished in the absence of urea in the reaction mixture, a reagent that normally improves coverage. The best comparative figure for conventional in-solution digestion was 81.7%.

The digestion efficiency was improved by raising the reaction temperature up to 60°C. Heat has the same effect as a reducing agent, unfolding the protein to allow better access for the enzyme. At higher temperatures, the sequence coverage fell sharply because the enzyme became deactivated.  

The sequence coverage began to fall as the amount of protein was decreased, while maintaining the same enzyme:protein ratio. For 100 ng protein, coverage was about 90% but it fell sharply after that. However, this should allow the detection of biomarkers that are present at low abundances in biological media like plasma and urine.

The test with the glycoprotein demonstrated that the enzyme reactor could be used to isolate glycopeptides from non-glycosylated peptides, allowing both sets to be analysed in turn when they were eluted from their respective traps.

The conditions set for the glycopeptides were also used to compare the glycoproteins in urine samples from patients with prostate cancer with those of healthy patients. In all, glycopeptides and N-linked glycopeptides common to both sets of patients were detected corresponding to 55-N-linked glycoproteins. One of these was zinc α-2-glycoprotein which has been proposed as a biomarker for prostate cancer and was 2.17-fold more abundant in the cancer patients than the healthy people.

Operation of the enzyme reactor can be fully automated. Even the glycopeptide analysis could be modified by fitting a preliminary reactor for glycoprotein digestion, followed by an online mixer where the lectin is added, then a second reactor to digest the glycopeptides.

This adaptable system brings about rapid protein digestion and separation, with reduced carryover by back-flushing it from the system. It provides an inexpensive way of simplifying shotgun proteomics experiments that could aid researchers looking for disease biomarkers.