Chemicals and standards
Methanol and acetonitrile were purchased from J.T. Baker (USA). All standards of MC-LA, MC-LF, MC-LR, MC-WR, MC-LW, MC-LY, MC-RR, and MC-YR (95% purity) were obtained from BePure, China. The internal standard leucine enkephalin was purchased from Zhenzhun Biologicals, China. Formic acid was obtained from Waters, USA. Ultrapure water was produced using a Milli-Q water purification system (USA).
MC standard solutions were diluted to 1 mg L–1 with methanol, placed in a brown glass bottle, and stored at − 20 °C. The 1 mg L–1 MC solution was diluted to the required concentration before each experiment. The leucine enkephalin internal standard solution was diluted to 50 μg L–1 with methanol/water (1:1, v/v) and stored at − 3 °C.
Mixed calibration standard solutions were serially diluted to 1000, 500, 200, 100, 50, 20, and 10 ng L–1 with methanol before use, and then stored at − 20 °C.
Operating conditions of UPLC–MS/MS
The sample was introduced into the injection loop and transferred to an online-SPE column for analyte preconcentration. The online aqueous mobile phase continued to flow after sample loading to ensure matrix and salt removal. The analyte was eluted by back-flushing the online-SPE column using the UPLC mobile phase and separated by the chromatographic column prior to MS/MS detection.
Eight different MCs were analyzed using an ACQUITY UPLC system coupled to a Xevo TQ-MS (triple-quadrupole MS/MS) mass spectrometer (Waters, Manchester, UK). The analytical column and SPE column were ACQUITY UPLC BEH C18 (1.6 μm, 2.1 mm × 50 mm) and XBridgeC8 Direct Connect HP (10 μm, 2.1 mm × 30 mm), respectively. The system was operated in electrospray positive mode (ESI+), with a capillary voltage of 3.70 kV, source and desolvation temperatures of 150oC and 500 °C, respectively, desolvation gas flow rate of 1000 L h–1, cone hole backflush gas pressure of 30 V, cone hole backflush gas flow rate of 50 L h–1, and collision gas flow rate of 0.06 mL min–1. All parameters were optimized to provide the best sensitivity for all analytes. Detection and quantification were achieved using targeted analysis via positive ion scanning and multiple-reaction monitoring. The other mass spectrometer parameters are listed in Table 1.
Table 1 Compound-dependent MS/MS parameters of eight MC congeners Optimization of chromatographic conditions
Several gradients were investigated to optimize the peak resolution and sensitivity and to minimize the run time, which involved altering the flow rate, gradient, and concentration of acetonitrile. The MC congeners were detected on a triple-quadrupole mass spectrometer using multiple-reaction monitoring with transitions optimized manually to achieve maximum sensitivity.
SPE column
The separation efficiency of online-SPE columns depends on their retention of MCs. The fillers used in the online-SPE column are usually silica gel as the matrix and a bonded phase with relatively weakly polar functional groups. The more polar components of the sample elute from the online-SPE column first, whereas the less polar components are retained on the column. Therefore, the online-SPE column with reverse-phase bonded, weakly polar silica gel adsorbent is optimal for the separation of MCs. Therefore, we evaluated the separation efficiency of XBridge C8 and Oasis HLB C18 by comparing the chromatographic resolution of eight different MCs (20 ng L–1). Compared with the commonly used Oasis HLB Direct Connect HP (20 µm, 2.1 mm × 30 mm), XBridge C8 is used for reversed-phase extraction, which is more suitable for nonpolar to medium-polarity target compounds. The online-SPE column used in this study was the XBridge C8 Direct Connect HP (10 μm, 2.1 mm × 30 mm) to retain the strongly polar substances of eight different MCs eluted in the void volume of the Oasis HLB C18 column.
Water spiking with internal standard and filter membranes
An appropriate internal standard can balance the change in the signal response intensity of the analyte caused by matrix interference under certain conditions and reduce the interference of the analyte loss during sample pretreatment [27, 28]. Leucine enkephalin is used as the internal standard for the determination of MC. Xu et al. showed that MCs and leucine enkephalin can be separated well, and their recovery rates are similar owing to their similar structures [29].
Ultrapure water samples need to be filtered before analysis to protect the instruments and reduce matrix effects. However, MCs can be lost because hydrophilic filter membranes can partially absorb MCs through hydrogen bonds. Therefore, the recovery of three different membranes was evaluated in this study: polyethersulfone (PES), polytetrafluoroethylene (PTFE), and mixed cellulose ester (MCE). During the membrane evaluation experiment, the MC concentration was 50 ng L–1, and each membrane was evaluated four times in parallel.
Mobile phases
The two components of the mobile phase were termed as “A + B” in this study. Six different mobile phases were evaluated: (1) methanol + water; (2) acetonitrile + water; (3) methanol with 0.1% (v/v) formic acid + water with 0.1% (v/v) formic acid; (4) acetonitrile with 0.1% (v/v) formic acid + water with 0.1% (v/v) formic acid; (5) acetonitrile with 0.25% (v/v) formic acid + water with 0.1% (v/v) formic acid; (6) acetonitrile with 0.5% (v/v) formic acid + water with 0.1% (v/v) formic acid. The separation efficiency was evaluated by comparing the peak intensities of the eight different MCs with the column maintained at 35 °C.
Chromatographic resolution (R) was used to characterize the degree of separation of two adjacent chromatographic peaks, which is equal to the ratio of the difference between the retention times of adjacent chromatographic peaks (t1 and t2) and the average peak width of the two chromatographic peaks (w1 and w2), as shown in Eq. (1):
$$R=\frac{{t}_{2}-{t}_{1}}{\frac{1}{2}({w}_{1}+{w}_{2})}=\frac{2({t}_{2}-{t}_{1})}{{w}_{1}+{w}_{2}},$$
(1)
where t1 and t2 are the retention times of the first and second peaks, respectively, and w1 and w2 are the widths of the first and second peaks, respectively.
When R < 1, the two peaks overlap; when R = 1, the resolution can reach 98%, and when R = 1.5, the resolution can reach 99.7%. When R = 1.5, two adjacent components are usually considered to be completely separated.
Gradient elution procedures
Four gradient-elution procedures were employed: (1) the water phase was held at 100% for 4.1 min, followed by a decrease to 0% over 2.9 min, and then washed for 4 min at 100% before the next injection; (2) the water phase was held at 98% for 4.6 min, followed by a decrease to 25% over 5.4 min, and then washed for 2 min at 98% before the next injection; (3) the water phase was held at 95% for 4.1 min, followed by a decrease to 60% over 1.9 min and another decrease to 5% over 3 min, and then washed for 3 min at 95% before the next injection; and (4) the water phase was held at 95% for 6 min, followed by a decrease to 5% over 3 min, and then washed for 2 min at 95% before the next injection.
Method validation
Seven concentration sequences (10, 20, 50, 100, 200, 500, and 1000 ng L–1) of the standard MC solution were detected according to the finalized method. Leucine enkephalin (10 ng L–1) was added as the internal standard, and the authenticity and absolute recovery of the analyte were calculated. A linear regression and standard curve were applied with the injection concentration (x) corresponding to the peak area (y).
Three mixed standard solutions with high (500 ng L–1), medium (100 ng L–1) and low (20 ng L–1) concentrations were added to the blank water samples. After filtration by using a 0.22-μm MCE membrane, the concentrations of the target substances in the water samples were determined by online-SPE UPLC–MS/MS. Six samples of each concentration were taken for parallel experiments, and the recovery rate and relative standard deviation (RSD) were calculated. The LOD (ng L–1) and LOQ (ng L–1) were calculated using signal-to-noise ratios (SNRs) of 3 and 10 based on the lower-end calibration curve levels. The accuracy values were calculated as the average value of the recovery for concentrations of 20, 100, and 500 ng L–1, and the precision values were calculated as the average RSD for concentrations of 20, 100, and 500 ng L–1.
Water samples for method development
Twelve water samples (CH1, CH2, …CH12) collected from Chaohu Lake (Anhui Province, China) in August 2020 were used to validate the method. At least 12 zones in Chaohu Lake were chosen for water sample collection, and 500 mL of water was collected from depths of 0–50 cm in each zone after the surface scum was removed. Then, 100 mL of each water sample was filtered in situ using a 0.45-μm cellulose acetate filter membrane (JiuDing, China) in a 120-mL polypropylene bottle. The samples were then placed in a cooler with ice packs and transferred to the laboratory for further treatment. The 12 sampling points are shown in Fig. 1.
Each water sample of 20 mL was filtered using a disposable medical syringe coupled with a 0.22-μm filter in the injection vials; each sample was spiked with 10 ng L–1 of internal standard prior to injection. To quantify the MCs, a seven-point mixed standard calibration curve in the range of 10–1000 ng L–1 was created based on an initial sample size of 20 mL.