System

One of the benefits of FAC is the simplicity of the fluidic system; all that is required is to continuously infuse the solution of ligand(s) through a packed bed. Inspection of most breakthrough curves might suggest that the analyses reflect "poor" chromatography, because breakthrough curves can appear extremely diffuse. As was discussed, this is usually due to the actual binding event, but such curves can mask poor column performance as well. To capitalize on the strength of FAC for generating accurate binding data, proper attention should be paid to the fluidic system. Early systems used simple syringe pumps or conventional HPLC systems for higher flow rates. As the columns continue to diminish in capacity, neither system is appropriate as stable low flow is difficult to achieve. Syringe pumps in particular should not be used for obtaining binding data, as flow rate variations can be as high as +20% at low mL min-1 flow rates.

We have designed new pumps based on a 250 mL positive displacement pump, an integrated electronic inlet/outlet valve, and an inline flow sensor. The pump operates at nano- to micro-flow rates using feedback from the inline flow sensor. During operation, the feedback signal is used to dynamically clamp the output flow to a desired value, in this way achieving flow-rate stability of better than 1% even below 100 nl min-1 [19]. These nanofluidic modules are capable of operating at 5000 psi, suitable to drive our FAC cartridges incorporating 5 mm nonpo-rous particles. Because they dynamically respond to preserve constant flow rates, they are particularly well suited to conventional injection systems, where sample is loaded through a "superloop"; upon injection, the loop pressure is rapidly equilibrated to the system pressure and stable flow rates are generated. This allows us to make measurements of binding data with CVs of 2% or better [19].

Inclusion of a second pump for effluent dilution and transfer supports both online and offline MS analysis. As shown above, fraction collection followed by LC/ MS analysis significantly expands the performance characteristics of the FAC method, but the effluent can also be sampled for MALDI-based analysis. Advantages to this method include greater salt tolerance over the electrospray approach, extension to complex mixtures of protein and archiving of the run. MALDI is generally considered to possess higher peak capacity than electrospray (at least

Volume (jxl)

Fig. 6.14 Infusion of dilute human serum through a transferrin-binding protein B (TbpB) FAC assay. Insets demonstrate MALDI-TOF spectra acquired before (2 min) and after (7 min) the principal breakthrough curve for transferrin, shown as a solid black trace representing m/z @80000. The asterisk denotes the m/z of human transferrin [15].

Volume (jxl)

Fig. 6.14 Infusion of dilute human serum through a transferrin-binding protein B (TbpB) FAC assay. Insets demonstrate MALDI-TOF spectra acquired before (2 min) and after (7 min) the principal breakthrough curve for transferrin, shown as a solid black trace representing m/z @80000. The asterisk denotes the m/z of human transferrin [15].

for peptides and proteins). Neither ionization method when applied to the unfrac-tionated effluent is expected to have the peak capacity of the LC/MS approach, but with the MALDI method the effluent can be sampled more extensively, making it an excellent choice of less complex mixtures. We demonstrated the utility of FAC-MALDI/MS in the detection of transferrin binding to transferrin-binding protein B (TbpB) subunit, a peripheral outer membrane lipoprotein from bacteria essential for iron uptake direct from human serum [15] (e.g. Neisseriaceae spp). In this experiment, a soluble form of TbpB was expressed with a recombinantly introduced biotin tag, in place of the lipid anchor. This construct was immobilized on 5 mm nonporous streptavidin beads; approximately 1.7 pmol of active TbpB was bound. Dilute human serum was infused through the column and the effluent spotted on a MALDI plate in 15 s intervals. The breakthrough curve for transferrin could be readily detected by MALDI (Fig. 6.14). The double-plateau nature of the breakthrough likely reflects weakly bound apotransferrin, followed by the more strongly retained iron-loaded form. We have also shown that this effluent can be sampled for proteomics analysis and protein discovery, where fractions are digested with trypsin and the resulting peptides compared against each other using LC-MS/MS datasets. This should be an attractive alternative to conventional pathway discovery which uses bead-based pulldowns and washes. FAC supports the discovery of weaker or transient interactions, which most often go undetected in conventional pulldowns. A drawback to the approach is that each fraction may require laborious 2D-LC-MS/MS analysis to array the contents of the fractions.

More recently, Brennan has shown that FAC-MALDI-MS can be used to screen small molecules, relying upon MRM transitions to overcome the chemical noise generated by the matrix [16]. This is an acceptable approach for known compounds, but for ligand discovery from uncharacterized mixtures, ion selection will be difficult. More appropriate may be the DIOS (desorption/ionization on silica) surfaces that support matrix-free desorption of small molecules.

Was this article helpful?

0 0

Post a comment