Since the introduction of ACQUITY UPLC (Ultra Performance Liquid Chromatography) by Waters (Milford, Mass.) over a decade ago, there has been increased interest in the effects of high pressure on HPLC samples. Prior to the UHPLC era, the effect of pressure on HPLC was studied, but generally for small molecules and pressures lower than 400 bar. Increasing pressure did affect retention behavior, but not by much. The effect of temperature was studied more thoroughly, since it had a large effect and was useful in elucidating the mechanism of retention by Van’t Hoff plots, etc. Early UHPLC systems had a maximum pressure (Pmax) of 1000 bar, which has since been extended to over 1500 bar and has the potential to go even higher, especially if submicron column packings prove useful. Since proteins potentially can react to changes in the temperature and pressure encountered during UHPLC conditions, what should the analyst be aware of as the separation proceeds?
Effects of high pressure on proteins
The following examples illustrate the effects of pressure, temperature and heating on proteins.
Figure 1 – Waters Ethylene Bridged Hybrid (BEH) particle technology, available in numerous particle sizes and bonded phases for HPLC and UPLC, offers excellent peak shape and efficiency for basic analytes, a rational array of chromatographic selectivity and improvements in chemical stability at mobile phase extremes, particularly at elevated pH.Professors Szabolcs Fekete and Davy Guilllarme (University of Lausanne, Geneva, Switzerland) studied four proteins—lysozyme, myoglobin, filgrastim and interferon alpha-2A—using a C4 RPLC 1.7-μm Waters BEH 300, 50 × 2.1 mm packed column (see Figure 1).1 This short butyl surface is often used to improve column efficiency for proteins under reversed-phase liquid chromatographic (RPLC) conditions. Earlier studies showed a 3000% increase in retention with a 1000-bar increase in pressure for myoglobin, which was much larger than the increase for a range of small-molecule probes.2 C4 stationary phases are less commonly used in small-molecule separations, but longer hydrophobic phases, such as C18, are often too retentive for proteins.
Bridgman studied the effect of high pressure on egg whites, which are rich in lysozyme,3 and found that proteins in egg whites are completely denatured at a pressure of 7 kbar.
Figure 2 – View of a steaming lake in Yellowstone National Park. The lake and exiting streams cover mats of colored living bacteria that are responsible for the blue and other colors.High pressure and elevated temperature may not always be denaturing or deadly, as we see in extremophiles. Unusual fish in particular inhabit ocean depths where pressure can be in the kilobar range.4 They may have proteins that function only at high pressure. Acre-size mats of bacteria thrive in hot water at Yellowstone National Park (Figure 2).
BaroFold Inc. (Boulder, Col.), makers of PreEMT, or Pressure Enabled Protein Manufacturing, seeks to improve the tolerability, efficacy and safety of a variety of protein therapeutics. The company denatures proteins by increasing and then decreasing pressure, which can lead to refolding.5 The pressure ranges of UHPLC and the PreEMT process overlap. This raises the question: Are protein structure and function dependent on ambient pressure?
Pressure-induced protein denaturation in liquid chromatography
An LC sample is generally prepared and stored at atmospheric pressure until it is switched into the flow stream and transported to the column. Upon injection, the pressure on the sample increases rapidly to the system pressure, which can range from 10 to 2000 bar. Small molecules and some peptides and smaller proteins are robust and are not affected by pressure. However, robustness decreases with increasing protein size; for example, insulin (small) is more robust than hemoglobin, which in turn is more robust than multimeric proteins such as IgMs (very large, with several subunits).
Pressure effectively denatures proteins since interruption of the native structure usually leads to a reduction in system volume. The presence of internal cavities within the folded protein structure particularly favors pressure-induced unfolding.
The column packing and mobile phase can interact with the protein and may alter the mechanism of unfolding by lowering the activation energy of the unfolding reaction. Fekete and Guillarme studied RPLC, with retention based on hydrophobic interactions. Some proteins have hydrophobic patches on the surface. These are often necessary for their location and function. By way of contrast, other chromatographic modes, such as ion exchange, prefer interactions (attractive or repulsive) with ionophores. Steric exclusion chromatography (SEC) should have the lowest interaction with the stationary phase. Studies in SEC mode would be an interesting extension of Fekete and Guillarme’s work.
Chromatographic procedure
The chromatographic process can be explained in three parts.
Top of the column, including the inlet frit
Upon injection, the sample encounters the frit followed by the top of the chromatographic bed. A protein analyte endures the full system pressure. With sub-2-μm column packings, the pressure can be in the low-kbar range.
Rates of unfolding are rarely measured, however. A paper by Broom et al.6 (University of Waterloo, Ontario, Canada) reports that protein-unfolding rates correlate strongly with folding rates, which restores the native structure. In a related report by Yanxin Liu et al.7 (University of Illinois, Urbana, Ill.), pressure-jump was used to study the denaturation and refolding of a lambda receptor. The formation of helices proceeded stepwise and was complete in 19 μsec. A second folding phase of about 1.5 msec resolved errors in the μsec helix formation.
Thus, for some proteins, there is a possibility that pressure-induced denaturation quickly leads to a denatured state, which may expose hydrophobic surfaces to interaction with the hydrophobic stationary phase. Fekete and Guillarme studied this effect by increasing the flow rate from about 100 μL/min (atmospheric P) to 1400 μL/min (750 bar). They discovered that retention increased 240% for lysozyme and 480% for interferon alpha-2A. Since the stationary phase is hydrophobic, the retention may be due to increased exposure of hydrophobic patches of the analyte. Small molecules, on the other hand, show only a small (25–100%) increase when the pressure is increased from 100 to 1100 bar.2 In the gradient elution mode, the increase was smaller still.
Frictional heating at the column inlet
At the top of the column, frictional heating should be minimal, since the generation of heat from resistance to flow from passing from the injector to the column inlet is minimal.
Column midpoint
At the column midpoint, the pressure is reduced by half. The dissipation of pressure with length has been confirmed by coupling two identical columns in series, which doubles the pressure. For a pressure-sensitive protein analyte, the reduced pressure may pass through the pressure point, which favors refolding and renaturation. When this happens, the protein may start to refold. This can be very rapid.7
By column midpoint, the temperature would be higher than at the inlet due to resistive heating. The heating is not adiabatic, since narrow-column-diameter (2.1-mm) columns packed with sub-twos often show resolving power superior to fatter columns. This has been attributed to improved dissipation of heat to the column walls, which provides a more uniform radial and axial temperature profile.
Column terminus
The column terminus is most interesting—this is where the effects of pressure and temperature can be seen. Because it has been exposed to the most frictional heating, the temperature is highest here and the pressure is lowest, about equal to atmospheric pressure. If the pressure-induced denaturation is reversible, then the original protein could be refolded into its native form. Homologous multimeric proteins can also reassemble and renature. However, the retention times of the component part’s heterogeneous multimeric proteins is expected to be different, and thus would not be in the same volume for regeneration.
Temperature-induced denaturation is now a possibility. This may not be reversible. Power is dissipated via frictional heating as the liquid races through the packed bed. Fekete and Guillarme studied this by adding a variable-pressure restrictor after the column terminus. This allowed the pressure to be kept constant at 750 bar while the flow rate was varied from 100 to 1400 μL/min. Since the pressure drop was constant, any retention changes would be due to resistive heating. The retention of lysozyme and myoglobin was particularly sensitive to changes in temperature.
The researchers measured changes in retention factors as a function of column temperature. Van’t Hoff plots demonstrated a peculiar curved trajectory. At low temperatures (25 °C), pressure fluctuates, but as the temperature reaches about 50 °C, the plots show an apex followed by a decrease. This indicates that the proteins have undergone a significant change in conformation. For lysozyme, the pressure-induced changes dominate at low temperature (<30 °C). However, the temperature has a stronger influence on interferon alpha, even in the conventional HPLC range.
Conclusion
Proteins that have undergone a structural change during chromatography may not be representative of the starting material. If the identity of the original analyte is of interest, as it would be in discovery proteomics and preparative applications, the identity of the peaks should be confirmed post-chromatography. This can be done with a quick scan of the starting sample to see if post-column materials are present in the starting sample. Circular dichroism can often reveal if the purified proteins are similar to the reference standards. Running an electrophoresis gel, especially a native gel, with the pre-chromatography sample and collected fractions from a UHPLC should show if any unique bands appear in the collected UHPLC fraction. This would indicate they came from the chromatography.
In conclusion, transitioning a protein assay from HPLC to UHPLC may not be as simple as it is with small molecules. The effects of pressure on proteins separated by ion exchange and steric exclusion chromatography might not be as pronounced as those separated by RPLC. The added pressure raises new possibilities and hence questions that should be addressed as part of a method validation.
References
- Fekete, S.; Guillarme, D. Estimation of pressure-, temperature- and frictional heating-related effects on proteins’ retention under ultra-highpressure liquid chromatographic conditions. J. Chromatogr. A 2015, 1293, 73–80.
- Fekete, S.; Veuthy, J.-L. et al. The effect of pressure and mobile phase velocity on the retention properties of small analytes and large biomolecules in ultra-high pressure liquid chromatography. J. Chromatogr. A 2012, 1270, 127–38.
- Bridgman, P.W. The coagulation of albumin by pressure. J. Biol. Chem. 1914, 19, 511–12.
- http://list25.com/25-most-terrifying-deep-sea-creatures/
- http://www.barofold.com/
- Broom, A.; Gosavi, S. et al. Protein unfolding rates correlate as strongly as folding rates with native structure. Prot. Sci. 2015, 24, 580–7.
- Liu, Y.; Prigozhin, M.B. et al. Observation of complete pressure-jump protein refolding in molecular dynamics simulation and experiment. J. Am. Chem. Soc. 2014, 136, 4265–72.
Robert L. Stevenson, Ph.D., is Editor Emeritus, American Laboratory/Labcompare; e-mail: [email protected]