HPLC Analysis of Very Polar Compounds in Bioanalysis

Several high performance liquid chromatography (HPLC) approaches can be used in the analysis of polar small molecules that are classified as pharmaceuticals, metabolites, or biomarkers. Many of these polar compounds can be challenging for traditional reversed-phase separation methods. One approach for analyzing these types of samples is the use of columns having polar retention capabilities such as ion-exchange, hydrophilic interaction liquid chromatography (HILIC), polar-embedded, or polar-endcapped columns. Another approach is aqueous normal phase (ANP) chromatography. This mode has the advantage of having both reversed-phase and normal-phase retention. In a recent web seminar, Joseph Pesek, a professor of Chemistry at San José State University (San José, California), explained the mechanisms, advantages, and disadvantages of these approaches to the analysis of polar compounds. Below, he answers questions raised during the web seminar. A recording of the web seminar is available for free at: www.chromatographyonline.com/HPLCPolarCompounds

Column Types and Their Properties

If hydrophilic interaction liquid chromatography (HILIC) works now, why did it not work 30 years ago?

Pesek: HILIC was not well understood 30 years ago. However, some of the shortcomings of HILIC were noted at that time and for that reason it was not pursued vigorously.

How is HILIC different from working with plain silica columns?

Pesek: Silica is just one column material used in HILIC. There are many others, each having somewhat different properties. Therefore, as in reversed-phase chromatography, in HILIC it is important to select the right type of column to match the properties of the analyte.

What is the difference between polar-endcapped and HILIC columns?

Pesek: Polar-endcapped columns are similar to traditional reversed-phase columns but instead of using a nonpolar endcapping reagent like a trimethyl group, a more polar group is used to cover some of the remaining silanols. HILIC columns generally are significantly more polar since they are based on bare silica or have bonded polar groups that cover a significant fraction of all of the available bonding sites.

Is silica really a HILIC stationary phase? Several studies suggest that it is very different from bonded phases.

Pesek: It is if you believe that the mechanism for HILIC retention is the presence of an adsorbed water layer on the surface that serves as a medium of analyte partition. It may be different from bonded phases but for most of these, the presence of a layer of water at or near the surface is used to describe retention.

What is your experience with the robustness of most HILIC columns?

Pesek: Most HILIC columns are less robust than traditional reversed-phase columns. This results in part from the fact that the surface can be easily contaminated and is hard to clean. In other cases, the bonded group is not very robust and is susceptible to being cleaved from the surface in mobile phases used for HILIC retention.

Are HILIC or similar columns compatible with ultrahigh-pressure liquid chromatography (UHPLC)?

Pesek: All of the columns discussed in the seminar, including aqueous normal phase (ANP) and HILIC columns, are compatible with UHPLC.

Can I use silica hydride columns for mixed analysis of polar and nonpolar compounds or amphiphilic compounds? If so, is there a possibility that the retention times of polar and nonpolar compounds might be the same (that is, their respective affinities are equal), thus yielding the same retention times?

Pesek: This is relatively unlikely because that would require that both mechanisms be operating at equal efficiency for those particular compounds. In most cases one mechanism would be more predominant than the other and retention would be different. However, even in the unlikely event that both were equal, that would only occur at one mobile-phase composition. Therefore, you could switch to a different isocratic composition or a different gradient and that would shift the relative contributions of the two mechanisms, allowing separation.

Why are there different bonded phases using silica hydride if it is possible to retain both polar and nonpolar analytes on any silica hydride column?

Pesek: The degree of retention in reversed-phase and aqueous normal phase modes on silica hydride columns depends on the degree and type of modification. For columns with no or minimal modification the ANP mode is stronger than the reversed-phase mode. As the degree of modification becomes greater — that is, in columns modified with larger groups or that have a greater degree of surface coverage — the reversed-phase properties increase.

What type of columns can be used to separate both polar and nonpolar compounds?

Pesek: The most versatile columns for this type of separation are silica hydride columns because the amount of retention in both modes can be adjusted by the type of modification and the mobile-phase composition. Polar-endcapped and polar-embedded columns have this same capability because they are also classified as mixed-mode stationary phases. The degree of polar retention depends on the endcapping group or the embedded group.

What type of gradient do you suggest for ANP analysis of hydrophilic and hydrophobic molecules in the same run? I have a large hydrophobic active and a prodrug, as well as five small polar metabolites.

Pesek: The usefulness of ANP separations is that you have two options in selecting the gradient. It would be best to try both a reversed-phase gradient (from 80% to 20% aqueous) and then an ANP gradient (from 20% to 80% aqueous) as an initial screening process. From the preliminary results you can determine which mode is most likely to produce the best separation. Once the direction has been selected, then the gradient can be modified to reach the desired retention and separation of the components in your mixture.

How much efficiency do ANP columns have? Does efficiency have detrimental effects for biological matrices for liquid chromatography–mass spectrometry (LC–MS) analysis?

Pesek: Normal-phase retention generally has lower efficiency than reversed-phased methods. It is highly dependent on the compound being retained. Efficiency doesn't have detrimental effects for biological matrices for LC–MS analysis, but sometimes lower ionization of the compounds is observed in MS detector when compounds are being analyzed in biological matrices. The use of internal standards (usually deuterated compounds) will help to assess how much lower the peaks are.

Are there other bonded phases for ANP chromatography?

Pesek: Yes. As stated in the seminar, polar-embedded, polar-endcapped, and fluorinated phases have some ANP capabilities. Also, graphitized carbon phases have demonstrated some ANP behavior.

Can you tell us about carbon columns?

Pesek: Carbon columns, or porous graphitized carbon columns, also possess dual retention capabilities. Their structure is porous particles composed of flat sheets of hexagonal carbon. Selectivity in the reversed-phase mode is somewhat different than with traditional C18 columns. The absorptive nature of the carbon surface results in the polar retention observed.

What manufacturers make low-carbon silica hybride columns?

Pesek: The current suppliers of low-carbon silica hydride carbons are Microsolv Technology and VWR in the United States and Hichrom Ltd. in Europe and the UK.

How stable or durable are silica hydride columns?

Pesek: Silica hydride columns are just as durable, and in some cases more durable, than traditional chemically modified stationary phases based on organosilane chemistry. This is due to the direct silicon-carbon at the surface. For columns that are used primarily for polar retention in the ANP mode, these materials are more durable than typical HILIC columns because the hydride surface is not as easily contaminated as ordinary silica and the surface is less susceptible to attack by aggressive mobile phases because of its more hydrophobic nature.

Can you compare silica hydride and ionic liquids? Are there any important advantages of one over the other for retaining polar analytes?

Pesek: It is assumed you are referring to ionic liquid stationary phases. All of these materials contain a charged group that is part of the bonded material. The charged site is usually within the bonded chain, but not always. It is most similar to the polar-embedded phases discussed in the presentation. They have been shown to retain both polar and nonpolar compounds. To date, the range of compounds analyzed with ionic liquid stationary phases is not as broad as with silica hydride phases so the same versatility has not been demonstrated. A good review of ionic liquids in chromatography can be found in the book by Mun and Sim (1).

General

Why are polar compounds retained more strongly when they are ionized and not when they are neutral, as in reversed-phase chromatography?

Pesek: This occurs because the highest polarity is generally found in ionized compounds, as a result of the fixed charge on the compound. Therefore, in both aqueous normal phase chromatography and HILIC, maximum retention is attained for compounds with the highest polarity, in other words, those that are ionized.

Have you had any issues with ion suppression when using ion-pairing agents? If so, how did you remedy the issue?

Pesek: The issue of ion suppression for the analysis of polar compounds when using ion-pairing reagents in reversed-phase chromatography is a prime reason to not use this approach. In general, ANP chromatography with silica hydride stationary phases offers the most versatile approach for the analysis of polar compounds. It has been demonstrated that ANP chromatography can analyze positively or negatively charged compounds having molecular weights below 100 to larger molecules, such as peptides. Polar uncharged molecules can also be analyzed by this approach. In every case the mobile phases do not contain more than 10 mM of an additive, such as acetic or formic acids, or ammonium acetate or formate, all of which are MS-compatible.

Peak shape is usually the main problem I encounter with HILIC, mixed-mode, or ion chromatography, even when I'm matching the strength of the sample solvent with the mobile phase. Do you have any tips on improving peak shape?

Pesek: It is generally better to have the sample solvent strength stronger that the mobile-phase solvent, especially at the beginning of a gradient. In addition, silica hydride phases are less susceptible to peak distortions than HILIC. With the proper gradient, most compounds will give good peak shape. Certain compounds like polyprotic acids and phosphate-containing species have special considerations because they can be affected by metal ions in the high performance liquid chromatography (HPLC) system. Sometimes the addition of 0.5% formic acid or ammonia to the sample (depending on the compound) can improve peak shape.

Given that ANP media do not appear to form a water layer, does the ionic strength of the mobile phase have a significant impact on the retention characteristics of ANP chromatography, as it does in HILIC?

Pesek: The ionic strength of the mobile phase has a smaller effect in ANP chromatography than in HILIC because it appears that the mechanism is some type of competitive adsorption on the surface. Ionic strength would be expected to have some impact on this process.

Up to what percent of acid can you use to acidify silica hydride columns?

Pesek: Generally speaking there is not much need to go above 0.5% acetic or formic acids for most ANP applications. If it is trifluoroacetic acid for biological applications, the usual limit is about 0.1% when using UV detection. Trifluoroacetic acid is not recommended when using MS detection because of ion suppression.

Can you inject aqueous samples using ANP columns?

Pesek: Yes, it is possible to inject aqueous samples for many applications. Each specific analysis would have to be tested to determine if it is possible in that case.

Are all of these techniques (HILIC, ANP, and ion-exchange chromatography, and reversed-phase chromatography with polar-embedded or polar-endcapped columns) compatible with mass spectrometry?

Pesek: Yes, to some degree. It all depends on the additive and the concentration of the additive in the mobile phase. In ANP chromatography, most applications reported to date use mobile phases that are MS-compatible.

Could you please provide some guidance on how to prepare samples for HILIC chromatography? For example, what diluent should I use?

Pesek: Mobile phase in approximately the same ratio as the mobile phase.

Is the use of mobile phases with high amounts of methanol on phenyl columns considered ANP chromatography?

Pesek: In ANP separations, methanol is usually too strong of a solvent for retention of hydrophilic compounds. It does not matter what the column is. Only a few very polar compounds will be retained using methanol. However, if retention increases as the amount of methanol is increased in the mobile phase, then it would be considered a normal-phase mechanism.

With HILIC, are typical validation parameters such as accuracy and precision comparable to those when using reversed-phase chromatography?

Pesek: No. In most cases the average levels of accuracy and precision in a HILIC analysis are lower than typical values obtained by reversed-phase separations. The accuracy and precision of a HILIC analysis is strongly influenced by the equilibration time if a gradient method is used. It is generally found that comparable analyses by ANP chromatography have a higher level of accuracy and precision.

How long is the post-run equilibration time for a typical HILIC column? How does it compare to the post-run equilibration time of the other columns you mentioned in the presentation?

Pesek: Post-run equilibration times for HILIC columns can be anywhere from 5 to 45 min. For ANP silica-hydride columns, the equilibration never takes more than 5 min.

Are there any special precautions one should take regarding water quality when carrying out ANP experiments?

Pesek: As demonstrated, in ANP and HILIC the mobile phase must be free of metal ion contamination. The presence of metal ions, particularly of copper and iron, will have an adverse effect on peak shape for certain acids and phosphate-containing compounds.

What is peak 6 in the right hand chromatogram of the five bases on the slide "Compatibility of hydride surface for chromatographically challenging compounds"?

Pesek: Actually, there was a mistake on this figure. Peaks 1–5 on the left side of the slide are correct. However, peak 1 on the right figure is a solvent peak and peaks 2–6 should actually be peaks 1-5 to correspond to the peaks on the left side. You will see that if this is done, the retention times will be very close for each of the five compounds on the two columns.

Specific Analytes or Types of Analysis

We are trying to analyze putrescine by LC–MS. Can you recommend a protocol or a column that can do the job?

Pesek: Being a diamine compound, putrescine is very amenable to analysis by ANP chromatography using a low carbon-bonded silica hydride–based column (Diamond Hydride, MicroSolv Technology). A gradient using water with 0.1% formic acid and acetonitrile with 0.1% formic acid going from 90% organic to 40% organic would be a good start. It can be modified to shorten analysis time or improve peak shape.

What would be the column of choice to analyze catecholamines by LC–MS? In the past I had problems with ion suppression because the catecholamines are not retained on reversed-phase columns.

Pesek: This is another good application for the silica hydride columns in the ANP mode. You can use 0.1% acetic or formic acid as the additive in the mobile phase and will not get any ion suppression in your mass spectrometer.

What is the best way to perform quantitative analysis of glucuronides?

Pesek: Glucuronides are best retained by ANP or HILIC modes. For gradient analyses, ANP mode with a silica hydride phase will generally have faster re-equilibration. These are analyzed with MS detection in the negative ion mode with the organic component of the mobile phase 95:5 acetonitrile–water and 10 mM ammonium formate and the aqueous component is water with 10 mM ammonium formate; a starting gradient would be from 95% organic solvent to 30% organic solvent.

Paraquat is a very polar compound, strongly binds with soil matrix, and is difficult to detach from soil. How can it be analyzed?

Pesek: You might want to consider extraction with 0.5 M K₂SO₄ as proposed by Vance and colleagues (2). If you can get a reasonable level of extraction, with good MS or MS–MS detection you should be able to determine paraquat at low levels using ANP chromatography or HILIC. ANP mode has the advantage when using a gradient of faster equilibration.

In your opinion, what would be the best approach to separate sugar phosphates, particularly C5-sugar phosphates?

Pesek: Sugar phosphates can be separated by ANP or HILIC modes. Care must be exercised to be sure the system is free of metal ion contamination. For complex samples requiring gradient methods, ANP chromatography will generally have shorter equilibration times between runs.

Do you recommend a hydride material or HILIC material for separating cytidine nucleosides?

Pesek: There have been several reports in the literature for the separation of cytidine nucleosides using hydride materials in MS compatible mobile phases. One strategy might be similar to what is found in a recent article in the Journal of Separation Science (3).

What is your favorite method for analyzing sugar phosphates, such as those from central carbon metabolism (G6P, F6P, R5P, X5P, Ru5P)?

Pesek: Although it is possible to analyze these compounds by HILIC, the silica hydride material, particularly the low carbon-bonded silica hydride, is a very versatile material for separating these types of compounds. Care must be exercised to make sure that the HPLC system and the solvents are free of metal ion contamination.

Does ANP chromatography have important applications in large-scale metabolomics studies? Or is it more applicable to targeted approaches?

Pesek: It can be applicable to either type of approach. A group in New York has done a large-scale study of bacterial metabolites and another group in Australia has done a large-scale study on plant metabolites. There are many more reports in the literature for targeted studies based on a specific metabolite or small group of metabolites.

What percentage of pesticides can be detected in wines using present HPLC methods?

Pesek: This is like many generic analytical problems in that there is no definitive answer. Obviously, part of the success in indentifying pesticides will depend on the separation method used — reversed-phase mode, normal-phase mode, or a combination of both. Second, it will depend on the type of detection used. Some methods are more sensitive than others and then there is the matter of how many of the pesticides the detector responds to.

References

(1) J. Mun and H. Sim, Eds., Handbook of Ionic Liquids: Properties, Applications and Hazards (Nova Science Publishers, Hauppauge, New York, 2012).

(2) E.D. Vance, P.C. Brooks, and D.S. Jenkinson, Soil Biol. Biochem. 19(6), 703–707 (1987).

(3) M.T. Matyska, J.J. Pesek, J. Duley, M. Zamzami, and S.M. Fischer, J. Sep. Sci. 33, 930–938 (2010).