Characterization of Nucleosides and Nucleobases in Natural Cordyceps by HILIC–ESI/TOF/MS and HILIC–ESI/MS

A method combining hydrophilic interaction chromatography (HILIC) and electrospray ionization mass spectrometry (ESI-MS) was developed for the characterization and determination of natural Cordyceps. Separation was achieved on a Waters Xbridge Amide column with gradient elution. Identification of 15 target nucleosides and nucleobases was based on retention time, UV spectra and mass measurements of the protonated molecules ([M+H]+) and main fragment ions (ESI-TOF/MS). Eight non-target compounds were tentatively identified by ESI-TOF/MS. The 15 target compounds were quantified by HILIC-ESI-MS/MS using time-programmed selective ion monitoring or multiple reaction monitoring in positive-ion mode under optimized mass conditions. This technique showed good linearity, repeatability and recovery. This approach was also successfully implemented in the analysis of nucleosides and nucleobases in 12 batches of natural Cordyceps samples that were collected from different regions in China. The developed HILIC-ESI-MS method exhibited clear advantages in identifying and determining highly polar bioactive components in Cordyceps, as well as their quality control.


Optimization of HILIC-ESI-MS/MS Conditions
In this study, a HILIC column was used to separate the high-polarity components in Cordyceps because of its poor retention on the reverse-phase column [29]. Chromatographic conditions were optimized using both nucleoside and nucleobase standards and a real natural Cordyceps sample. Firstly, three types of HILIC column, a Waters Xbridge TM BEH Amide column (250 × 4.6 mm I.D., 3.5 μm), Waters Xbridge TM HILIC column (150 × 2.1 mm I.D., 3.5 μm) and a Merck ZIC ® HILIC column (150 × 2.1 mm I.D., 3.5 μm) were compared. The results indicated that the former had better resolution for these hydrophilic components than the other two with the same mobile phase. Acetonitrile and water were selected for the mobile phase in the HILIC method based on a previous study [25]. Considering the complexity of the Cordyceps ingredients, gradient elution was selected to improve separation in the subsequent study. Besides, different mobile phase additives were compared for good resolution, peak shapes and compatibility with the ESI source. The result showed that the use of 0.2% acetic acid and 10 mmol/L ammonium acetate as mobile phase additives could provide much improved resolution and good peak shapes for these compounds. Figure 1 shows the HILIC chromatogram for the 60% (v/v) aqueous methanol extract of Cordyceps.  The mass spectrometry conditions were optimized to quantify 15 nucleosides and nucleobases, as well as to achieve maximum sensitivity. Preliminary direct infusion studies on the 15 standards confirmed that the positive ion mode was more sensitive. MS parameters such as nebulizer gas pressure, dry gas flow rate and dry temperature were optimized by injecting a mixture of standards. However, the results indicated that they had little effect on the sensitivity of the protonated molecules. Additionally, in tandem mass spectrometry, parameters such as cone voltage and collision energy were also adjusted for the most abundant product ion by using the fluxes and concentrations mentioned above. In this study, different collision energies determined by optimization experiments were applied to the 15 nucleosides and nucleobases. In addition, [M+H] + was used as a precursor ion for most compounds. However, [M+Na] + was selected for 2'-deoxyuridine. MRM was used to quantify the nucleoside, thereby minimizing background noise and improving sensitivity. However, the abundance of product ions for nucleobases was too low for detection using MRM; hence, SIM was used for the quantitative determination of these nucleobases. Table 1 shows the optimum parameters for MS/MS.

Qualitative Analysis for Cordyceps Extracts by HILIC-DAD-ESI-TOF/MS
The crude extract of Cordyceps was analyzed by HILIC-DAD-ESI-TOF/MS under optimized conditions, and the total ion chromatogram (TIC) of the Cordyceps sample is shown in Figure 2. It's difficult to distinguish the peaks of the Cordyceps sample based on the TIC. However, 15 target nucleosides and nucleobases (peaks 1 to 15 in Figure 1) were unambiguously identified by their retention time, UV spectrum and accurate mass measurements in natural Cordyceps. Second, the molecular formula information of eight non-target compounds were also provided by ESI-TOF/MS. Table 2 lists 15 compounds found in the Cordyceps extract.    For example, peak 11 (35.2 min) in the chromatograms in Figure 1 is identified by retention time, UV spectra and ESI-TOF/MS spectrum, as shown in Figure 3. First, it exhibits a maximum absorbance wavelength (λ max ) of 265 nm, which characterizes nucleosides and nucleobases [24]. The exact masses of peak 11 are m/z 269.0886 and m/z 137.0462, as indicated in Figure 3. These measurements resulted from the [M+H] + and the [M-C 5 H 8 O 4 +H] + ions, which were consistent with the elemental compositions of C 10 H 13 N 4 O 5 and C 5 H 5 N 4 O, respectively. After comparison of its retention time, UV spectra and MS spectrum with a reference standard, it was identified as inosine.

Validation of the Method
The linearity and calibration curves of 15 analytes were determined using the developed HILIC-MS/MS technique. All correlation coefficients obtained are higher than 0.9965, as presented in Table 3. Most linear dynamic ranges of the calibration curves for HILIC-MS/MS determination exceed two orders of magnitude. The limit of detection and the limit of quantification were defined at a signal-to-noise ratio of approximately 3 and 10, respectively. The detection limits of HILIC-MS/MS techniques range from 0.15 ng/mL to 10.00 ng/mL for the 15 nucleosides and nucleobases. This finding indicates that the developed method exhibits the same sensitivity as that in a previous report using HPLC-ESI-MS/MS [28] and a higher sensitivity than those reported using IP-RPLC-MS [19].  Sample stability was tested by periodically analyzing a sample solution maintained at room temperature over varying periods (0, 2, 4, 6, 12 and 24 h). The storage stability of the measurements for the 15 nucleosides and nucleobases ranges from 3.25% to 5.69% (RSD, n = 6) for HILIC-MS/MS determination, as indicated in Table 4. Intra-and inter-day variations were chosen to determine the precision of the method. For the intra-day variability test, six replicate samples prepared independently were analyzed within one day. For the inter-day variability test, the replicate samples were analyzed in duplicate for three consecutive days. Total intra-and inter-day variations (RSD, n = 6) for the 15 analytes were less than 5.23% and 6.73%, respectively ( Table 4). As shown in Table 4, the recoveries range from 83.6% to 117.3%, demonstrating the good recovery possible by using the method.

Quantitative Analysis of Investigated Compounds in Natural Cordyceps by HILIC-ESI/MS/MS
The newly developed HILIC-ESI-MS/MS method was applied for the simultaneous determination of 15 nucleosides and nucleobases (thymine, uracil, thymidine, 2'-deoxyuridine, cordycepin, uridine, hypoxanthine, adenine, adenosine, xanthine, inosine, cytosine, guanine, cytidine and guanosine) of 12 batches of Cordyceps samples that were collected from different regions in China. Table 5 summarizes the content of the investigated compounds in Cordyceps.
The results in Table 5 indicate that almost all Cordyceps samples are rich in nucleosides and nucleobases. The total content of the 15 nucleosides and nucleobases range from 388.43 μg/g to 1622.84 μg/g. The No. 10 sample from Suohumari Town, Qinghai Province has the highest content. Thymidine, uridine and adenine are rich in 12 natural Cordyceps samples. In addition, among the 12 samples, the No. 5 sample from Xiamagong Town, Qinghai Province exhibits the highest cordycepin content. Besides, the result also indicated that there were differences between the nucleosides and nucleobases contents of Cordyceps from different sources. The samples from different sources have various chemical profiles of nucleosides and nucleobases, which could be selected as the chemical markers for the quality control of Cordyceps.

Materials and Standards
HPLC-grade acetonitrile was obtained from Merck (Darmstadt, Germany). Acetic acid and ammonium acetate were purchased from Fluka (Buchs, Switzerland). All other reagents such as methanol were of analytical grade (Tianjin Chemical Factory, Tianjin, China). Water was prepared using a Millipore Milli Q-Plus system (Millipore, Bedford, MA, USA).
Twelve batches of natural Cordyceps samples were collected from different regions in China ( Table 6). The species of these natural Cordyceps was identified by Prof. Luqi Huang (Institute of Chinese Material Medical, China Academy of Chinese Medical Sciences, Beijing, China). The standards of thymine, uracil, thymidine, 2'-deoxyuridine, cordycepin, uridine, hypoxanthine, adenine, adenosine, xanthine, inosine, cytosine, guanine, cytidine, guanosine and 2-chloroadenosine were purchased from Sigma-Aldrich (St. Louis, MO, USA). Figure 4 shows the chemical structures of these reference compounds.

Sample Preparation
Cordyceps powder (0.3 g) was mixed with 60% (v/v) aqueous methanol (25 mL) in a glass tube with a stopper and then immersed in a conventional stainless steel ultrasonic cleaning bath. The average power used for extraction was 320 W. A single frequency of 40 kHz, as well as applied pulsed frequency, was used. Extraction was controlled at 20 °C. Samples were extracted in 15 min for three consecutive times. The extracts were combined and then filtered, and the filtrate was dried using a rotary evaporator at 55 °C. The residue was then dissolved in 5 mL of 90% (v/v) aqueous acetonitrile. The concentration of analytes was adjusted between linearity ranges according to the preliminary determination. 2-Chloroadenosine (1 µg/mL) was used as an internal standard for the quantitative determination of analytes by using a calibration curve. The solution was filtered through a 0.22 μm nylon filter membrane. A 20 μL aliquot of the filtrate was subsequently injected into the HPLC system.

HILIC-DAD
All separation processes were performed on an Agilent 1200 HPLC system (Agilent Technologies, Santa Clara, CA, USA) equipped with a quaternary solvent delivery system, a vacuum degasser, an autosampler, a column oven, and a diode-array detector connected to an Agilent ChemStation software program. A Waters Xbridge Amide column (250 × 4.6 mm I.D., 3.5 μm) was used. The mobile phase consists of A (0.2% acetic acid and 10 mM ammonium acetic) and B (acetonitrile), which are used in gradient elution: 0 min to 15 min, 95% to 95% B; 15 min to 30 min, 95% to 90% B; 30 min to 40 min, 90% to 80% B; 40 min to 45 min, 80% to 75% B. The flow rate of the mobile phase was 0.8 mL/min, and the injection volume was 20 μL. Elution was monitored at 260 nm by using a diode-array detector, and the column was maintained at 20 °C.

HILIC-ESI-TOF/MS Analysis
A G6520 Q-TOF mass spectrometer (Agilent Technologies) with an electrospray ion source was coupled with the Agilent 1200 HPLC system (Agilent Technologies) equipped with a Waters Xbridge Amide column (250 × 4.6 mm I.D., 3.5 μm) for the qualitative analysis of nucleosides and nucleobases in Cordyceps. Data were processed using MassHunter software (A02.02) provided with the Agilent TOF/MS. The mobile phase and the gradient elution program described in Section 3.3 were employed. The outlet of the flow cell was connected to a split valve to divert a flow of 0.4 mL/min to the ESI source. The mass spectrometer conditions were optimized for nucleosides and nucleobases detection, as follows: drying gas temperature, 350 °C; drying gas flow, 10.0 L/min; nebulizing gas pressure, 45 psi; fragmentor, 120 V; and capillary voltage, 4,500 V. The mass spectrometer was scanned from m/z 50 to m/z 1000 in a full scan mode. Reference solution was used, and the ion with m/z 121.050873 was selected for mass calibration to eliminate system bias.

HILIC-ESI-MS/MS Analysis
Quantitative analyses of nucleosides and nucleobases were performed on an Agilent 1200 HPLC system (Agilent Technologies) equipped with a Waters Xbridge Amide column (250 × 4.6 mm I.D., 3.5 μm) and coupled to an Agilent G6420A mass spectrometer (Agilent Technologies) triple quadrupole instrument. The mobile phase and the gradient elution program described in Section 2.3 were employed. The outlet of the flow cell was connected to a split valve to divert a flow of 0.4 mL·min −1 to the ESI source via a short fused silica tubing. Peaks were detected by a positive-ion mode MS and MS/MS. Mass spectrometry was conducted in the scan mode from m/z 50 to m/z 600. Selective ion monitoring (SIM) or multiple reaction monitoring (MRM) mode was used for quantitative analysis. The ESI-MS conditions were as follows: drying gas temperature, 350 °C; drying gas flow, 11.0 L/min; nebulizing gas pressure, 45 psi; fragmentation voltage, 120 V; and capillary voltage, 4500 V.

Calibration Solutions
Standard stock solutions of all analytes, except for xanthine and guanine, were prepared by dissolving each compound in an aqueous methanol solution (20%, v/v) at 0.1 mg/mL and then stored at 4 °C. Xanthine and guanine (0.05 mg/mL) were prepared in aqueous acetic acid (2%, v/v) and then stored at 4 °C. The aforementioned standard solutions were prepared and diluted to appropriate concentrations with 90% (v/v) aqueous acetonitrile for the construction of calibration curves. Each calibration curve was generated by running samples at six different concentrations in triplicate. Relative peak areas were plotted against the concentrations of each analyte. The correlation coefficient was determined using a linear regression model.

Conclusions
In this study, a reliable technique for the simultaneous identification and determination of nucleosides and nucleobases in natural Cordyceps was developed by combination of HILIC-ESI-TOF/MS and HILIC-ESI-MS/MS approach. Fifteen nucleosides and nucleobases in 12 batches of natural Cordyceps were determined with the developed method. The results suggest that the developed method is specific and sensitive for the analytes examined, and can be used for the identification of non-target compounds. The analysis result showed that Cordyceps samples were rich in nucleosides and nucleobases, and they could be selected as markers for quality assessment of Cordyceps. Furthermore, the developed HILIC-ESI-MS method is a useful tool for the identification and determination of the highly polar bioactive components in Cordyceps as well as their quality control, and could also be used for the routine quantitation of nucleosides and nucleobases from other TCMs.