A quality FAC assay presents certain column design constraints. Most researchers' experience with affinity chromatography involves the use of gravity-fed sample flows through large-bore column beds constructed of loosely packed agarose-based particulate supports. Large excess column capacities are usually designed into such systems, specifically to maximize the capture of a ligand, and little attention is paid to the surface area occupied by the capture molecule. The first objective in designing a new FAC assay, in contrast, is to optimize the amount of

2) b-Galactosidase recognizes b-Gal compounds, whereas the lectin GSIB4 selectively recognizes a-Gal compounds. All six mixtures were doped with low-micromolar concentrations of the respective indicators (1S,4S isomer of 4-hydroxy-2,2-

dimethyl-cyclopent-1-yl 1-thio-b-D-galactopyranoside as a known binder for b-galactosidase [15] and 4-phenyl-2-butanone-4-thio-a-D-galactopyranoside [racemate] as a binder for GSIB4).

immobilized capture molecule per unit column volume. In most situations, this should be as high as possible without introducing unwanted interactions between adsorbed molecules due to molecular crowding. This serves to minimize nonspecific binding. Current commercially available microparticulate solid phase supports have not been developed with this in mind, as opposed to chip surfaces in SPR. In a FAC experiment, any retention mechanism in the system will directly contribute to the measured breakthrough volume. In the worst case, nonspecific binding can completely override specific associations and, while we have shown that indirect methods can overcome this limitation, this does erode the utility of mixture screening. These are not new problems, as assay developers have long struggled with nonspecific associations. In our laboratory, we prefer to use passivated silica where possible. Passivation usually involves bonded hydroxyl coatings (glycerol or polyvinyl alcohol) as well as physisorbed protein, the goal being the masking of inevitable ''hot-spots'' in most particulate materials. Nonspecific binding is impossible to ''solve'' but stationary phases can be adapted to the analysis based on the class of molecules targeted. For example, we have found that oligosaccharide-based compounds can be successfully analyzed through cartridges constructed from beaded, polyol-coated polystyrene-divinylbenzene, whereas neutral heterocyclic organic compounds respond better to silica-based supports. Ultimately, the requirements for preserving protein activity usually dictates the freedom of design; for example immobilization of the enzyme GnT-V on silica led to its rapid inactivation, whereas immobilization on a polymeric support actually increased the stability of the enzyme compared to the enzyme in solution [19].

Some degree of nonspecific binding can be tolerated in FAC. In most situations nonspecific binding is characterized by weak retention mechanisms, which means that the contribution of such binding to the measured breakthrough volume is independent of ligand concentration and easily corrected in a concentration-series experiment. For discovery-based applications where large mixtures are processed, operating at maximal mixture dilution ensures that breakthrough volumes are dominated by the specific interaction [19].

In any case, high densities of capture molecule ensure that unwanted nonspecific binding sites are minimized and that the breakthrough curves can be interpreted from the standpoint of the specific interaction. If high densities are achieved, this presents an interesting design problem: the amount of stationary phase required for a FAC column is exceedingly small, on the order of 100 nL or less for a cartridge presenting 1-5 pmol of protein (assuming a 100-nm pore diameter material as a model). Most commercially available porous affinity supports are based on large beads (>30 mm) that cannot be packed into small capillary cartridges of 100 mm inner diameter or less. A larger system can be designed and used successfully in FAC experiments, but this is wasteful of sample. Most recently, we have implemented nonporous 5 mm spheres for FAC, which are easier to pack into small cartridges. These nonporous supports present a reduced surface area over the porous ones (> a factor of ten) and their only drawback is that they present a larger pressure drop over the cartridge length, but this is man ageable with newer fluidic systems we have designed. The excellent diffusion characteristics of these nonporous particles are an added advantage. While it is true that accurate breakthrough curve measurements can be made with low efficiency columns [11], sharper breakthrough curves support greater precision in multi-ligand mixture analysis and will be essential for measuring a wide range of rate constants.

Recent developments in the Brennan laboratory offer exciting possibilities for streamlining the generation of FAC columns. This laboratory has adapted in situ sol-gel technology for the capture of proteins directly from solution. Briefly, silica sols are prepared from diglycerylsilane and (3-aminopropyl)triethoxysilane with various additives to preserve protein function, in the presence of protein. In one example, dihydrofolate reductase (DHFR) was added in a ''single-pot'' solgel preparation and cast within fused silica capillaries [28]. The resulting monolithic column was successful in trapping DHFR within mesopores and preserving the activity of approximately 25% of this protein. The monolith could sustain pressure-driven flow and appears to support sufficiently rapid mass transfer (Figs. 6.12, 6.13). The issues with column reusability were attributed to a nonoptimum buffer rather than an inherent limitation of the entrapment procedure. This sol-gel procedure is an elaboration on well established entrapment methods [29], but with the added advantage of stability and better flow properties. Interestingly, none of the examples presented thus far demonstrate competitive behavior between multiple ligands (i.e. displacement); in the FAC analysis of trimethoprim and pyrimethamine a reversed order of elution based on Kd is described, but this could simply be due to the shift towards an on-rate limited situation for higher affinity compounds, as described earlier. Erosion of dynamic competition between ligands could occur if the sol-gel allows convective mixing of the entrapped protein; however the bimodal pore structure of these materials would

Fig. 6.12 Scanning electron microscopy images of a sol-gel derived column material (A) a rigid rod and (B) a magnification of the bimodal pore structure in the resulting monolithic material [28]. Adapted with permission from the American Chemical Society.

Fig. 6.12 Scanning electron microscopy images of a sol-gel derived column material (A) a rigid rod and (B) a magnification of the bimodal pore structure in the resulting monolithic material [28]. Adapted with permission from the American Chemical Society.

■(A) —' ^vT

Was this article helpful?

0 0

Post a comment