Nobel Laureate Linus Pauling, Ph.D., predicted in 1939 that hydrogen bonding would prove to be more significant in the field of biology than any other type of chemical bond. His prediction has been fulfilled abundantly for the most part, except in chromatography. Hydrogen bonding has been acknowledged as a contributing factor to protein retention on ion exchangers since the mid-1950s and more recently suggested as a selectivity modifier for many mixed-mode chromatography products.1,2 But it has not yet been exploited as a primary adsorption mechanism.
There are various reasons why it may have been overlooked. The mechanism is less intuitive than methods like ion exchange, and individual hydrogen bonds are known to be weaker than ionic bonds. However, it is also true that potential hydrogen bonding partners on biomolecule surfaces are far more abundant than charged residues. Recent data show that hydrogen bond chromatography is not only feasible but offers unique practical attributes that ion exchange, hydrophobic interaction, and other modes of chromatography cannot. Among them: a high degree of salt tolerance and an ability to sort retained biomolecules by size.
Hydrogen Bonding Monolith Stronger at Acidic pH
Figure 1 compares retention of a series of purified test samples on a hydrogen bonding monolith (CIMmultus™ H-Bond™ ADC) versus a strong anion exchanger (CIMmultus™ QA).3 The ligand for H-Bond ADC consists of a terminal series of hydrogen donors grafted to a root series of hydrogen acceptors. It has a weak positive charge, but no cation exchange or hydrophobic groups. The QA (quaternary amino) anion exchanger provides an ideal experimental control because it is also positively charged but it cannot participate in hydrogen bonding. The hydrogen atoms covalently bonded to its methyl carbon atoms are unable to act as hydrogen donors, and its quaternary nitrogen atoms lack the lone pair of free electrons required to accept hydrogen.
Consistent with expectations, retention is weaker for all biomolecules on the strong anion exchanger under acidic conditions than it is under alkaline conditions (Figure 1). Biomolecules become more electropositive at low pH, creating electrostatic repulsion from the positively charged exchanger surface. Directly contrary to those results, retention on the H-Bond column becomes stronger at acidic pH—despite its positive charge. This is understood to result from increasing protonation, which creates a net increase in hydrogen bonding potential.
Another important distinction is that biomolecule retention on the H-Bond column increases with molecular size, and the effect is compounded at acidic pH. At pH 6.0, H-Bond retention of bovine serum albumin (BSA, ~66 kDa) is about twice as strong as the strong anion exchanger. Retention of IgM (~960 kDa) is about four times stronger and the trend continues with the virus at about 16.7 MDa. There was no size discrimination among these species at any pH on the strong anion exchanger. Using the QA monolith as a reference to indicate the contribution of ionic interactions, the implication is that hydrogen bonding is responsible for about 80% of the binding energy achieved at pH 6.0 on the H-bond monolith.
Van der Waals forces could hypothetically contribute to this differential but experiments conducted in the presence of sorbitol suggest that most of the binding energy is attributable to hydrogen bonding. Sorbitol is a monosaccharide with six hydrogen donors and acceptors. As such, sorbitol is to hydrogen bond chromatography as what salts are to ion exchange chromatography. A sorbitol solution of 200 mM reduced binding strength for all tested samples by about half. Experiments with higher concentrations were limited by viscosity of the sugar.
These findings lead to the question of practical utility and there are several features of note. The high salt requirements for elution translate into high tolerance for salts in applied samples. An IgM requiring a NaCl concentration less than 20 mM to bind a strong anion exchanger at pH 6.0 bound effectively to the H-Bond monolith in 200 mM NaCl at pH 6.0. High salt tolerance can also be exploited to minimize or eliminate the need for a separate buffer exchange step in multi-step chromatography methods.
Size Discrimination
The ability of hydrogen bond chromatography to support a degree of size discrimination also creates compelling opportunities to improve purification. Figure 2 illustrates a series of chromatograms conducted with an IgM mAb in cell culture harvest. All runs were eluted with linear NaCl gradients. Even at pH 8.0, the H-Bond monolith gave clearly better separation from contaminants than the strong anion exchanger. Separation improved further at pH 6.0 with the majority of small contaminants eluting before IgM and larger contaminants, including aggregates and DNA, eluting after.
Stronger binding of larger solutes particularly highlights the suitability of hydrogen bond chromatography for very large biologics, like virus particles, but these products also impose special requirements. Diffusion constants for viruses can be 5–15 times slower than for proteins like albumin and, in many cases, they are too big to enter the pores in traditional column-chromatography media. These factors limit binding capacity and productivity. Their large size also makes them highly vulnerable to shear forces such as created in the void volumes of particle columns.
Monoliths bypass these concerns. Mass transport is exclusively convective through highly interconnected 2–6 µm channels. Flow is laminar, thereby avoiding the turbulent shear stress created in particle columns. Capacity and separation performance are both independent of flow rate, so neither capacity nor productivity are compromised. Virus capacity on monoliths is commonly 10–100 times higher than on particle columns.4
Development of Hydrogen Bond Chromatography Methods
Salt gradients represent the simplest screening option but alternative gradient formats offer better overall contaminant reduction. Gradients with nonionic hydrogen donor-acceptors, like sorbitol, represent one option and they offer the additional benefit of stabilizing most biomolecules. Figure 3 illustrates a bacteriophage eluted in a gradient to 200 mM sorbitol at 100 mM NaCl, pH 6.0. In the absence of sorbitol, its elution required nearly 2.5 M NaCl.
Another option is to elute with an increasing pH gradient. A monoclonal IgM that required 1.2 M NaCl for elution at pH 6.0, eluted in a gradient from pH 6.0 to pH 8.0 at 50 mM NaCl. Subsequent analysis by size exclusion chromatography (SEC) showed a single IgM peak free of aggregates. These results are further noteworthy because they highlight the dominance of hydrogen bonding as the primary adsorption mechanism, and its distinction from anion-exchange chromatography. pH gradient elution from anion exchangers is achieved with decreasing pH gradients.
These results frame a general strategy for method development. Screen first with a salt gradient. If it is a particular objective to maximize size discrimination, begin at the lowest pH where the product of interest is known to be stable. Otherwise, pH 6.0 is a good starting point. Evaluate elution with a pH gradient or nonionic donor-acceptor gradient at the highest salt concentration where the product of interest does not elute. Hold salt concentration constant during pH or nonionic donor-acceptor gradients. As with other chromatography methods, design of experiments can be used to reduce the workload of optimizing the fine details.
Overall, expect hydrogen bond chromatography to be complementary to other chromatography methods, and in some cases to offer better performance than established methods. Situations where size discrimination is helpful will be prime applications. Being able to conduct separations in a chemically biocompatible low-shear environments will be especially attractive with very large biologics. Otherwise, the unique selectivity of hydrogen bond chromatography may provide the needed missing ingredient in any analytical or preparative situation
Pete Gagnon ([email protected]) is CSO at BIA Separations in Ajdovscina, Slovenia, and a downstream processing editorial advisor for GEN.
1. Boardman N., S. Partridge, Separation of neutral proteins in ion exchange resins, Biochem. J. 59 (1955) 543–552.
2. Cole R., The chromatography of insulin in urea-containing buffer, J. Biol. Chem. 235 (1960) 2294–2299.
3. Gagnon P., U. Cernigoj, S. Peljhan, R. Zabar, U. Vidic, K. Vrabec, Hydrogen bond chromatography: principles and methods, Oral presentation, 8th International Monolith Summer School and Symposium, Portoroz, Slovenia, June 15–20, 2018.
4. Gagnon P., The emerging generation of chromatography tools for virus purification, Bioprocess Intl. 6(Sup. 6) (2008) 24–30.