2020 Volume 68 Issue 12 Pages 1233-1237
The aim of this study was to investigate appropriate analytical conditions for hydrophilic nucleosides and nucleotides (monophosphates and triphosphates) by HPLC methods using a mixed-mode AX-C18 column with anion-exchange and hydrophobic interactions by quaternary ammonium and C18, respectively, and a reversed-phase pentabromobenzyl (PBr) column with dispersion force and hydrophobic interactions by PBr group. The higher compound polarity led to stronger retention on AX-C18 (triphosphates > monophosphates > nucleosides). AX-C18 demonstrated feasible retention of nucleotides via anion–exchange interaction by increasing the salt and methanol concentrations. In contrast, on PBr, the lower compound polarity led to stronger retention. On PBr, feasible retention of both nucleosides and nucleotides was obtained via dispersion interactions with purine and pyrimidine rings by increasing the methanol concentration. Regarding the pH of phosphate buffer used as the mobile phase, pH 7.0 should be used in measuring nucleoside triphosphates on AX-C18, whereas pH 2.5 is better suited for measuring nucleotides on PBr. In terms of selectivity to highly hydrophilic nucleotides, the mixed-mode AX-C18 column had an advantage over the reverse-phase PBr column. In contrast, PBr column was more versatile than the AX-C18 column. Taken together, HPLC analyses of nucleosides and nucleotides should be carried out by optimizing the interactions between the stationary phase and nucleic acids.
Nucleic acid components, such as nucleosides and nucleotides, which regulate many biological functions, are regarded as potential biomarkers for physiological functions, including cardiovascular system and oxidative stress.1,2) In addition, nucleic acid analogues have been utilized for chemotherapy against cancer and viral pathogens.3) Therefore, their quantification is needed for effective therapy because blood concentrations of drugs are often related to their efficacy and toxicity.4) Furthermore, convenient measurement methods of nucleic acids are also indispensable in numerous research fields, including biochemistry, physiology, and pathology.
Hydrophilic nucleic acids are poorly retained on conventional reversed-phase columns. Ion-exchange LC, ion-pair reversed-phase LC, or hydrophilic interaction liquid chromatography (HILIC) have been applied to increase the retention of hydrophilic nucleic acids.5) However, these methods have several disadvantages.5–7) Another method that may be a valuable alternative is mixed-mode chromatography with multiple separation modes. It is not a novel concept and has been applied to solid phase extraction.8,9) In HPLC applications, the versatility of columns with mixed-mode stationary phases has been reported.9) For example, mixed-mode columns with both reversed-phase and anion-exchange groups are a powerful tool for analyzing nucleic acids.6,10) Among the mixed-mode columns available, the AX-C18 column has a mixed-mode stationary phase chemically bonding octadecylsilyl (C18) and quaternary ammonium (strong anion-exchange) groups on silica, and separates nucleic acids by the difference in their hydrophobicity and charge.
Another alternative is reversed-phase columns with functional groups such as halogenated stationary phases instead of C18. The pentafluorophenyl (PFP)-bonded silica column was first developed and offers alternative selectivity to the conventional C18 column.11) A pentabromobenzyl (PBr)-bonded silica column was recently introduced into the market.12) In addition to hydrophobic interactions, the PFP column uses dipole-dipole and π–π interactions, whereas the PBr column uses dispersion force interactions for separation. Dispersion force is a weak intermolecular force known as instantaneous dipole-induced dipole force. The PBr column improves the retention of hydrophilic compounds, such as nucleic acids, with heterocyclic rings that have a strong dispersion force.
Thus far, the retention behavior of nucleic acids has not been compared between mixed-mode and reversed-phase PBr columns. The aim of this study was to investigate appropriate analytical conditions for different nucleic acids and their analogues on the two columns, AX-C18 and PBr columns, which use anion-exchange and dispersion force interactions, respectively, in addition to hydrophobic interactions to separate compounds. The novelty of this study is the development of a simple method to measure nucleic acids by HPLC analyses using AX-C18 and PBr columns, which has been challenging by conventional methods.
Phosphate buffer and methanol were HPLC grade (Nacalai Tesque, Kyoto, Japan). Adenosine, thymidine, guanosine, uridine, cytidine, gemcitabine (Gem), ATP and uridine triphosphate (UTP) were purchased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Cytarabine (Ara-C) and thymidine triphosphate (TTP) were purchased from Sigma-Aldrich, Inc. (St. Louis, MO, U.S.A.). AMP, thymidine monophosphate (TMP), GMP, uridine monophosphate (UMP) and guanosine triphosphate (GTP) were purchased from Nacalai Tesque Inc. Cytidine monophosphate (CMP) was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Cytidine triphosphate (CTP) was purchased from Cayman Chemical (Ann Arbor, MI, U.S.A.). Ara-C monophosphate (Ara-CMP) and Gem monophosphate (Gem-MP) were purchased from Toronto Research Chemicals (Ontario, Canada). Ara-C triphosphate (Ara-CTP) was purchased from Jena Bioscience (Jena, Germany). The stock solutions (10 mM) of each nucleic acid were prepared by dissolving in deionized water. Other chemicals used were of reagent grade.
HPLC AnalysisColumns (150 × 4.6 mm i.d., 5-µm particle size) used in this study were ReDual™ AX-C18 (Shimadzu GLC, Tokyo, Japan) and COSMOSIL PBr (Nacalai Tesque). For single measurement, samples of 15 µL (100 µM) diluted by the mobile phase were analyzed by HPLC (Shimadzu LC-20 A system). The HPLC conditions were as follows: column temperature, 35 °C; flow rate, 1.0 mL/min; detection, UV 270 nm. The mobile phase was composed of 20 mM phosphate buffer/methanol (90 : 10 v/v) for the AX-C18 column and 20 mM phosphate buffer for the PBr column (Table 1, Supplementary Table 1). To make the elution of analyte faster, the mobile phase composition was changed to 50 mM phosphate buffer/methanol (90 : 10 or 90 : 20 v/v) for the AX-C18 column (Table 2, Supplementary Table 2) and 20 mM phosphate buffer/methanol (90 : 10 v/v) for the PBr column (Table 3, Supplementary Table 3). The pH of phosphate buffer was adjusted to 2.5 or 7.0. Samples were measured by three replicate injections.
| Compound | Log P valuea) (descending order) | Acidic pKa b) | Ratio (ionic to mocelular form) | Basic pKa c) | Ratio (ionic to molecular form) | Retention | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| AX-C18 (20 mM PB : MeOH = 90 : 10) | PBr (20 mM PB) | |||||||||
| pH 2.5 | pH 7.0 | pH 2.5 | pH 2.5 | pH 7.0 | pH 2.5 | pH 7.0 | ||||
| (A) Nucleoside | ||||||||||
| Adenosine | −0.76 | 3.82 | 20.89 | −− | ± | + | ++ | |||
| Thymidine | −0.84 | ± | ± | ++ | ++ | |||||
| Guanosine | −1.47 | 3.12 | 4.17 | − | − | ++ | ++ | |||
| Uridine | −1.58 | −− | −− | ± | ± | |||||
| Cytidine | −1.81 | 4.26 | 57.54 | −− | −− | ± | ± | |||
| Ara-C | −1.81 | 4.26 | 57.54 | −− | −− | ± | ± | |||
| Gem | −2.22 | 4.26 | 57.54 | −− | − | ± | + | |||
| (B) Monophosphate (nucleotide) | ||||||||||
| TMP | −3.04 | 1.86 | 4.36 | 1.38 × 105 | + | ± | + | ± | ||
| AMP | −3.19 | 1.86 | 4.36 | 1.38 × 105 | 3.82 | 20.89 | − | ± | ± | ± |
| UMP | −3.45 | 1.86 | 4.36 | 1.38 × 105 | ± | − | ± | −− | ||
| GMP | −4.07 | 1.86 | 4.36 | 1.38 × 105 | 2.42 | 0.83 | ± | ± | ± | ± |
| Gem-MP | −4.08 | 1.86 | 4.36 | 1.38 × 105 | 4.26 | 57.54 | − | ± | ± | −− |
| CMP | −5.13 | 1.86 | 4.36 | 1.38 × 105 | 4.26 | 57.54 | −− | − | − | −− |
| Ara-CMP | −5.13 | 1.86 | 4.36 | 1.38 × 105 | 4.26 | 57.54 | −− | − | ± | −− |
| (C) Triphosphate (nucleotide) | ||||||||||
| TTP | −3.50 | 0.80 | 50.12 | 1.58 × 106 | − | ++ | ++ | ± | −− | |
| ATP | −3.65 | 0.97 | 33.88 | 1.07 × 106 | 3.82 | 20.89 | ++ | ++ | ± | ± |
| UTP | −3.91 | 0.97 | 33.88 | 1.07 × 106 | − | ++ | ++ | − | −− | |
| GTP | −4.53 | 0.97 | 33.88 | 1.07 × 106 | 2.42 | 0.83 | ++ | ++ | ± | −− |
| CTP | −5.60 | 0.97 | 33.88 | 1.07 × 106 | 4.26 | 57.54 | ++ | + | − | −− |
| Ara-CTP | −5.60 | 0.97 | 33.88 | 1.07 × 106 | 4.26 | 57.54 | ++ | ++ | − | −− |
a–c) Calculated by SciFinder, b) phosphate residue, c) nitrogen atom in base. PB, phosphate buffer. Symbols: ++, excessive retention (20 ≤ k); +, great retention (10 < k < 20); ±, favorable retention (2 < k ≤ 10); −, little retention (1 < k ≤ 2); −−, insufficient retention (k ≤ 1).
| Compound | Retention | |||
|---|---|---|---|---|
| AX-C18 (50 mM PB : MeOH = 90 : 10) | AX-C18 (50 mM PB : MeOH = 90 : 20) | |||
| pH 2.5 | pH 7.0 | pH 2.5 | pH 7.0 | |
| TTP | ++ | ± | ++ | ± |
| ATP | ++ | + | ++ | ± |
| UTP | ++ | ± | ++ | ± |
| GTP | ++ | ± | ++ | ± |
| CTP | + | ± | + | ± |
| Ara-CTP | + | ± | + | ± |
PB, phosphate buffer. Symbols: ++, excessive retention (20 ≤ k); +, great retention (10 < k < 20); ±, favorable retention (2 < k ≤ 10).
| Compound | Retention | |
|---|---|---|
| PBr (20 mM PB : MeOH = 90 : 10) | ||
| pH 2.5 | pH 7.0 | |
| Adenosine | ± | + |
| Thymidine | − | ± |
| Guanosine | − | ± |
PB, phosphate buffer. Symbols: +, great retention (10 < k < 20); ±, favorable retention (2 < k ≤ 10); −, little retention (1 < k ≤ 2).
For simultaneous measurement, samples of 15 µL (10 µM) diluted by the mobile phase were used and HPLC conditions were changed as follows: column temperature, 40 °C; flow rate, 1.0 mL/min; detection, UV 270 nm, gradient elution of the mobile phase, 20 mM phosphate buffer/methanol (A, 90 : 10 v/v, pH 6.3) to 50 mM phosphate buffer/methanol (B, 85 : 15 v/v, pH 6.3) (0–100% B, 20–40 min) (Fig. 1).
Insert figure shows the chromatogram for the time range of 0 to 15 min.
The retention factor k was calculated by the following equation: k = (tR−t0)/t0, where tR is the retention time and t0 is the hold-up time (the time when an unretained compound passes through the column).13) t0 was calculated by dividing the hold-up volume V0 by the flow rate (1.0 mL/min). For a column of length L (150 mm) and inner diameter d (4.6 mm), V0 is ε(π/4)d2L (1.75 mL), where ε is the column packing porosity (a constant with an approximate value ≈0.7 for a particle size of 5 µm).13) The criteria of retention strength were defined as follows: 20 ≤ k, excessive retention (++); 10 < k < 20, great retention (+); 2 < k ≤ 10, favorable retention (±); 1 < k ≤ 2, little retention (−); k ≤ 1, insufficient retention (−−).14)
The relative retention strength of nucleic acids evaluated by the retention factor is shown in Table 1. Nucleosides and nucleotides (nucleoside monophosphates and triphosphates) are listed in descending order of log P-values. Phosphate groups in nucleotides form a chelate complex with active metal ions on the surface of the LC flow path, resulting in peak tailing.15) Thus, phosphate buffer was used as the mobile phase to competitively inhibit metal-phosphate interaction. In the mixed-mode AX-C18 column, at least 10% methanol was added to phosphate buffer because less than 10% methanol caused a gradual decrease in the retention time by repetitive measurements (data not shown). In contrast, the mobile phase without organic solvent was able to be used compatibly with the reversed-phase PBr column to reduce the elution of hydrophilic nucleic acids.
The retention characteristics of nucleic acids can be explained by three interactions, hydrophobic, anion-exchange and dispersion force interactions. The retention of nucleosides was greater on the PBr column than on the AX-C18 column (Table 1A, Supplementary Table 1A). It was difficult for the AX-C18 column to retain compounds with relatively lower log P-values, such as uridine and cytidine (analogues), only via hydrophobic interactions. On the other hand, the PBr column was able to retain all nucleosides via dispersion force interactions in addition to hydrophobic interactions. Greater retention (k > 20) of some nucleosides (adenosine, thymidine and guanosine) on the PBr column was considered to be due to stronger dispersion force interactions with purine (adenosine and guanosine) than pyrimidine rings and stronger hydrophobic interaction due to larger log P-values. The retention of Gem and Gem-MP was greater than that of other cytidine derivatives, presumably due to dispersion force interactions between halogen (F in Gem) and PBr groups (Tables 1A, 1B, Supplementary Tables 1A, 1B).
The retention of nucleoside monophosphates was almost equal between the two columns (Table 1B, Supplementary Table 1B). The retention of nucleoside triphosphates was much greater on the AX-C18 column than on the PBr column (Table 1C, Supplementary Table 1C). On the AX-C18 column, the retention strength was in the descending order of polarity (nucleoside triphosphates > nucleoside monophosphates > nucleosides). In contrast, the retention strength of these compounds on the PBr column was in the reverse order of that on the AX-C18 column (Table 1, Supplementary Table 1). Significantly increased retention (k > 20) was observed with nucleoside triphosphates over monophosphates on the AX-C18 column (Tables 1B, 1C, Supplementary Tables 1B, 1C). The ratio of ionic form to molecular form of phosphate groups was larger in nucleoside triphosphates than in monophosphates (Tables 1B, 1C). In addition, the negative charge of nucleoside triphosphates is larger than nucleoside monophosphates,16) supporting the stronger retention of nucleoside triphosphates via anion–exchange interactions.
Effects of Mobile Phase Composition on the Retention of Nucleic Acids with Criteria k > 20We next examined whether excessive retention of nucleoside triphosphates on the AX-C18 column and nucleosides on the PBr column can be attenuated by changing the composition of the mobile phase. Methanol was used as the organic solvent because π-electrons within acetonitrile are considered to hinder dispersion force interactions between analytes and the PBr stationary phase.17) At pH 7.0, the increase in salt and methanol concentrations in the mobile phase reduced the retention times of nucleoside triphosphates on the AX-C18 column, suggesting that higher salt and methanol concentrations attenuate anion–exchange and hydrophobic interactions, respectively. However, at pH 2.5, the AX-C18 column exhibited excessive retention of nucleoside triphosphates, except for triphosphates of cytidine and Ara-C, even after increasing salt and methanol concentrations (Table 2, Supplementary Table 2). It was possible to adjust the retention for nucleoside triphosphates on the AX-C18 column by changing the mobile phase composition at pH 7.0 (Table 2, Supplementary Table 2), although the AX-C18 column excessively retained nucleoside triphosphates with the mobile phase composition presented in Table 1C. On the other hand, on the PBr column, the addition of methanol reduced the retention times of nucleosides at pH 2.5 and/or 7.0 (Table 3, Supplementary Table 3). Taken together with the results shown in Table 1, the PBr column is recommended for analyses of moderate hydrophilic nucleosides, whereas both AX-C18 and PBr columns are available for analyses of highly hydrophilic nucleotides. In the separation of structurally closely related oligonucleotides,6) the mixed-mode column had higher selectivity than ion-pair reversed-phase and single anion-exchange columns. Therefore, the mixed-mode AX-C18 column is more suitable for analyses of hydrophilic nucleotides (particularly nucleoside triphosphates) than the PBr column. On the other hand, the reversed-phase PBr column, which demonstrated feasible retention of both nucleosides and nucleotides, was more versatile than the AX-C18 column. For wide-range analyses of natural compounds, the PBr column is advantageous as a screening column, whereas the mixed-mode column sacrificed versatility for selective retention of highly polar compounds.18)
Effects of Acidity of the Mobile Phase on the Retention of Nucleic AcidsAmong nucleosides, the retention times of adenosine, guanosine, cytidine, Ara-C and Gem on both AX-C18 and PBr columns were shorter at pH 2.5 than at pH 7.0, although the criteria of retention were similar for guanosine, cytidine and Ara-C (Table 1A, Supplementary Table 1A). This is because bases in these nucleosides became ionic at pH 2.5, leading to weaker hydrophobic interactions (Table 1A). In the case of nucleotides, the retention times on the PBr column were shorter at pH 7.0 than pH 2.5, suggesting pH 2.5 for the analyses of nucleotides on the PBr column (Tables 1B, 1C, Supplementary Tables 1B, 1C). Phosphate groups in nucleotides became ionic more extensively at pH 7.0 than at pH 2.5, which leads to weaker hydrophobic interactions (Tables 1B, 1C).7) On the AX-C18 column, weaker anionic properties of nucleotides at pH 2.5 may attenuate anion–exchange interactions. However, at pH 2.5, the retention of nucleotides was greater than at pH 7.0 on the AX-C18 column except for AMP, Gem-MP, CMP and Ara-CMP (Tables 1B, 1C, 2, Supplementary Tables 1B, 1C, 2). The difference in retention behavior of nucleotides between pH 2.5 and 7.0 on AX-C18 can be explained by three types of charges, positive charge of the quaternary ammonium group in the stationary phase, negative charge of the phosphate group in nucleotides and positive charge of bases in nucleotides. The positive charge of the quaternary ammonium group becomes larger at pH 2.5, whereas the negative charge of the phosphate group in nucleotides becomes larger at pH 7.0. The effects of the positive charge of the quaternary ammonium group on the retention should be more substantial than that of the negative charge of the phosphate group. However, in the case of AMP, Gem-MP, CMP and Ara-CMP, the retention of which was reduced at pH 2.5, the positive charge of bases in these nucleotides attenuated the interaction between the anionic phosphate group and cationic quaternary ammonium group because the negative charge of monophosphates is smaller than that of triphosphates in nucleotides (Table 1B, Supplementary Table 1B). Therefore, when measuring nucleoside triphosphates on the AX-C18 column, pH 7.0 should be used as an ion-pair reagent such as strong basic quaternary ammonium, which is used at around pH 7.0 for analyses of acidic compounds.19)
Simultaneous Measurement of Nucleic AcidsUsing the AX-C18 column, 15 kinds of nucleosides and nucleotides (5 nucleosides, 5 nucleoside monophosphates and 5 nucleoside triphosphates) were measured simultaneously (Fig. 1). By applying the gradient of salt and methanol concentrations in the mobile phase at pH 6.3, 15 kinds of nucleosides and nucleotides were successfully measured simultaneously, although there was little difference in retention times of nucleic acids eluted within 10 min. First, nucleosides and nucleoside monophosphates were eluted with 20 mM phosphate buffer/methanol (90 : 10 v/v). Next, nucleoside triphosphates were eluted by increasing salt and methanol concentrations to 50 mM phosphate buffer/methanol (85 : 15 v/v). Regarding the mobile phase pH, pH 6.3 was better for the separation of nucleosides and nucleoside monophosphates, whereas pH 7.0 was better for the elution of nucleoside triphosphates. In addition, the gradient of mobile phase pH caused a baseline drift. Therefore, the mobile phase at pH 6.3 was selected for the AX-C18 column. On the other hand, it was difficult to simultaneously measure 15 kinds of nucleosides and nucleotides on the PBr column (data not shown) because the retention times between nucleoside monophosphates and triphosphates were similar (Supplementary Tables 1B, 1C). When measuring nucleic acids in biological samples, the mixed-mode AX-C18 column is suitable for separating nucleoside triphosphates from interfering biological substances because of the selectivity to nucleoside triphosphates. However, further studies are needed in order to simultaneously measure different nucleic acids in biological samples.
AX-C18 and PBr columns enabled the measurement of nucleic acids within a feasible retention time using a simple mobile phase composition, which has been challenging by conventional methods such as HILIC-based, ion-exchange or ion-pair reversed-phase HPLC. This study provides useful information not only for predicting compound-stationary phase interactions involved in separation, but also for optimizing analytical conditions for the simple measurement of nucleic acids.
The authors declare no conflict of interest.
The online version of this article contains supplementary materials.
