2.1. Optimization of Experimental Conditions
In our previous research [30
], LC–MS (Agilent Technologies, Wilmington, DE, USA) was used for the simultaneous separation and detection of multiple lipophilic algal toxins under alkaline mobile phase conditions. On this basis, the LC–HR–MS (Agilent Technologies, Wilmington, DE, USA) and LC–MS/MS (Agilent Technologies, Wilmington, DE, USA) conditions for profiling the extracellular DSP compounds in the culture medium of marine algae were determined by further optimization of the gradient elution programs.
Developing a valid sample pretreatment method for the simultaneous analysis of all DSP compounds in the culture medium of marine toxigenic algae is critical. The applicability of three pretreatment methods, namely, direct sampling, SPE, and MRPA, was compared based on relevant reports [26
]. Two typical DSP toxins, OA and DTX1 (Figure 1
), were detected in the test solution obtained through these pretreatment methods, which indicated that these three methods can be used to characterize OA and DTX1 in the culture medium of toxigenic algae. However, the peak intensities and signal-to-noise ratio (S/N) of OA and DTX1 in Figure 1
b,c are apparently higher than those in Figure 1
a, which imply that SPE and MRPA can improve the detection sensitivity of extracellular DSP compounds in the culture medium. The extracted ion chromatograms (EICs) of OA at m
827.5 [M + Na]+
described in Figure 2
b,c show distinct isomeric peaks of OA, whereas those in Figure 2
a are unclear. These findings further demonstrate that the detection efficiency for low-concentration DSP compounds in the culture medium can be increased by SPE or MRPA. Compared with MRPA (Figure 1
c), SPE (Figure 1
b) has a better cleanup for the characterization of extracellular DSP compounds by LC–MS.
Furthermore, the applicability of SPE and MRPA for the quantitative analysis of extracellular DSP toxins was investigated. As shown in Figure 3
, the recoveries of OA and DTX1 for SPE were 103.6% and 94.8%, respectively, with RSD (n
= 3) ≤ 5.4%. For MRPA, the recoveries of OA and DTX1 were 110.3% and 100.6%, respectively, with RSD (n
= 3) ≤ 9.3%.
The results above indicate that direct sampling can be used as a pretreatment method for the screening and identification of high-concentration extracellular DSP toxins in the culture medium of toxigenic algae. This method has several advantages, such as simplicity, no consumption of organic solvents, and low cost. In addition, direct sampling allows the identification of main extracellular DSP toxins in the culture medium through LC–MS even when the culture medium volume is very small (only a few milliliters or tens of milliliters). Aside from the main DSP toxins, low-content extracellular DSP compounds can also be screened and identified in the culture medium through SPE and MRPA. Compared with MRPA, SPE requires much lesser time and organic solvents and can obtain a better cleanup effect for the enrichment of the extracellular DSP compounds in the culture medium. Therefore, SPE was selected as the preferred sample pretreatment method for the profiling and identification of extracellular DSP compounds in the culture medium of harmful marine algae.
2.2. Method Performance
Results from the instrumental precision of LC–HR–MS for two typical DSP toxins, OA and DTX1, are summarized in Table S1 (Supplementary Materials)
. The RSDs of peak area and retention time were ≤ 2.61% and ≤ 0.31%, respectively. The mass error of the measured exact mass for both OA and DTX1 was ≤ 5 ppm, which indicated satisfactory results in terms of precision and the exact mass of the method. The limits of detection (LODs) of OA and DTX1 were 12 and 25 pg, respectively. In general, sensitivity was able to completely meet the requirement for the screening and identification of extracellular DSP compounds in the real culture medium of harmful marine algae.
Matrix effects (ion suppression or ion enhancement) may affect the quantitative accuracy for LC–MS analysis, especially when electrospray ionization (ESI) source is used [36
]. Thus, the matrix effect on the LC–MS/MS for DSP toxin quantification was determined. Ion suppression or ion enhancement of OA (+15.06%) and DTX1 (−13.35%) was observed (Table 1
), indicating that the matrix effect cannot be ignored. Therefore, matrix-matched calibration standard curves were used for the quantification of the extracellular DSP compounds in the culture medium to ensure the accuracy of the quantitative results.
The LC–MS/MS method was validated, and the results for precision, linearity, regression equation, correlation coefficient, LOD, and limit of quantification (LOQ) are shown in Tables S2 and S3
. Good precision was obtained for OA and DTX1, with RSDs for peak areas and retention time of less than 4.20% and 1.16%, respectively. The matrix-matched calibration standard curves of target toxins showed good linear relationships with the coefficients of determination R2
≥ 0.9990, and the LODs and LOQs of this established method were within 0.22–0.47 and 0.56–0.93 pg/mL, respectively. The results indicate that the proposed method has a considerable sensitivity and quantitative linearity, and meets the requirements for the quantitative or semi-quantitative detection of extracellular DSP compounds in the real culture medium of harmful marine algae.
2.3. Screening and Identification of Extracellular DSP Compounds in the Culture Medium of P. lima
The testing solution of P. lima
culture medium was analyzed using the developed LC–HR–MS method in both positive and negative ion modes to obtain the crude total ion chromatogram (TIC) data (Figure 4
a). Then, the exact mass data of all 93 known DSP compounds (Table S4
) were entered to extract MS signals from the LC–HR–MS TICs. As shown in Figure 4
b,c, several suspected EIC peaks were obtained through signal extraction, whether in positive or negative ion mode. The HR–MS spectra of each suspected peak are provided in Figure 5
. The process shown below exemplifies the identification procedure for typical peaks by LC–HR–MS and LC–MS/MS.
For the DSP toxins without a commercially available reference standard, peak 8 was taken as an example to introduce the specific process for screening and identification. The 8th peak corresponds to m
977.5630 obtained by accurate mass extraction (extraction window of m
977.53–977.58). The MS spectrum close-up of this peak is provided in Figure 6
b. The possible molecular formula of this compound was deduced as C54
using Masshunter software, which was consistent with the chemical composition of OA-D10a/b. As shown in Table 2
, a relative mass error of −3.53 ppm was encountered for this compound. Moreover, the measured isotopic value clearly matched the theoretically calculated value (Figure 6
b). Thus, the molecular formula of this compound was confirmed as C54
, whereby the compound was tentatively identified as OA-D10a/b. The above findings indicate that HR–MS combined with accurate mass information can meet the requirements for the rapid screening and preliminary identification of DSP compounds in the culture medium.
LC–MS/MS was employed to further verify the compound. Figure 6
c shows the MS/MS spectrum of the suspected OA-D10a/b. Compared with the MS/MS spectrum of OA standard (Figure 7
), a typical fragment ion (m
827.7) of peak 8 was consistent with the quasi-molecular ion ([M + Na]+
) of OA. This result indicates that peak 8 is a member of the OA family. Other fragment ions including m
809.6, 791.5, 723.5, and 705.5 in Figure 6
c were in agreement with the fragment ions of OA. Thus, the compound was ultimately confirmed as OA-D10a/b. Li et al. [18
] also characterized OA-D10a/b by LC–MS/MS and obtained different MS fragmentation information, which might be due to the different types of mass spectrometer used.
For the DSP toxins with commercially available reference standards, retention time and fragmentation information were applied as evidence for verification by comparing with the values of the reference standard. The process described below exemplifies the confirmation of peaks 1, 2, 3, 4, and 5, as shown in Figure 4
b. The MS spectra (Figure 5
) of peaks 1, 2, and 3 contained the base peak of m
827.45, suggesting that the three compounds are OA suspects. Similarly, the exact molecular weights of peaks 4 and 5 were consistent with that of DTX1. However, we could not determine the peaks of OA and DTX1. Thus, LC–MS/MS was applied to verify the suspected toxins. By comparing with the retention times of standard OA and DTX1, peaks 3 and 5 were able to be preliminarily identified as OA and DTX1 (Figure 8
), respectively. Subsequently, the fragment ions of peak 3 and OA, and peak 5 and DTX1 were found to be coincident by comparing the MS/MS spectra in Figure 7
. As a consequence, peaks 3 and 5 (Figure 7
c,e) were ultimately determined to be OA and DTX1, respectively.
In addition, MS/MS analysis was performed on peaks 1, 2, and 4 (Figure 7
c,d,f). The characteristic fragment ions of peaks 1, 2 and peak 4 were in agreement with the OA and DTX1 reference standards, respectively. This finding indicates that peaks 1, 2 and peak 4 are the isomeric peaks of OA and DTX1, respectively. Due to the absence of reference standards, the isomers of OA and DTX1 can only be determined by combining with the relevant literature [20
]. Peaks 1 (Figure 7
c) and 2 (Figure 7
d) may be one or both of 19-epi
-OA, DTX2, DTX2b, or DTX2c. Although peak 4, OA methyl ester, and 35S DTX1 are isomers, the characteristic fragment ions of peak 4 (Figure 7
f) were consistent with that of DTX1. Therefore, peak 4 was speculated to be 35S DTX1 instead of OA methyl ester.
The screening and identification of other DSP compounds in the culture medium of P. lima
was similar to the procedure described above. In this study, nine extracellular DSP compounds were successfully identified in the culture medium of P. lima
. The retention time, molecular formula, detection ion, and mass deviation of all DSP compounds are listed in Table 2
. The relative deviation between the measured and calculated values of the accurate molecular weight of each compound was smaller than 6 ppm, which was in line with the accuracy requirements for screening by HR–MS. In addition, the MS/MS characteristics of all DSP compounds were similar because of their similar molecular structures. During MS analysis, the DSP compounds became prone to [M + Na]+
ion, with higher abundance in ESI positive mode, and were accompanied by [M + K]+
ion in trace amounts. In ESI negative mode, [M − H]−
ion was easily generated. However, all DSP compounds had lower sensitivity in negative mode except for OA and DTX1, as well as their isomers. During MS/MS analysis, the main fragment ions were m
809, 791, 723, and 705 for the compounds of the OA group and m
823, 805, 737, and 719 for the compounds of the DTX1 group, which were produced by a series of losses of H2
O, HOCONa, and H2
O from the precursor [M+Na]+
ion. As mentioned in the literature [4
], only OA and DTX1 were identified in the culture medium of P. lima
. In this study, the profile features of DSP compounds in the culture medium of P. lima
were elucidated by LC–HR–MS and LC–MS/MS, laying a good foundation for research on extracellular DSP toxins released by marine toxigenic algae.
2.4. Contents of Extracellular DSP Compounds in the Culture Medium of P. lima at Different Growth Stages
Present research on the excretion of DSP toxins for marine toxigenic algae is limited [2
]. In the present study, extracellular DSP compounds in the culture medium collected 8, 16, 22, and 25 days after the inoculation of P. lima
were identified and quantified. Different kinds of DSP compounds are collected from the culture medium of P. lima
at different incubation times (Table S5
). When the incubation time was 8 days, nine DSP compounds were detected in the algal cells and the culture medium of P. lima
. When the incubation time periods were 16, 22, and 25 days, eight DSP toxins were detected intracellularly. Furthermore, only 5, 7, and 6 DSP toxins, respectively, were detected in the culture medium.
Qantitative/semi-quantitative results of extracellular DSP compounds in the culture medium of P. lima
are shown in Table 3
. The content of each intercellular DSP compound increased gradually as the culture time was prolonged (e.g., OA content increased from 2.26 pg/cell to 14.45 pg/cell, and DTX1 content increased from 19.56 pg/cell to 66.25 pg/cell). Meanwhile, the content of extracellular DSP compounds initially decreased and then increased, which agreed with the research results of Nascimento et al. [16
]. In the initial stage of culture (8 days), nutrients in the culture medium were abundant, and the number of cells increased rapidly due to the rapid division in the exponential growth phase, which may have triggered the reduction in the content of extracellular toxins per unit cell. At the same period, the total amount of extracellular DSP compounds in the culture medium was also reduced by 2.43%–45.69%, which can probably be attributed to the reabsorption and reuse of DSP compounds at the rapid growth stage. When the rate of cell division slowed down, aging cultures and apoptosis appeared to promote a passive release of DSP compounds. Thus, the total content of extracellular DSP compounds increased significantly [2
], whereas the accumulation rate of intracellular toxins slowed down. The ratios of intracellular to extracellular contents of major DSP toxins depicted in Figure 9
initially increased, and then decreased, which was consistent with the explanation above. As shown in Table 3
, the content of intercellular DTX1 at the different growth stages of P. lima
was significantly higher (about 3–8 times) than that of OA. However, the contents of extracellular DTX1 and OA showed no significant difference. This finding might be the result of a very fast enzymatic hydrolysis into OA diol ester derivatives from OA sulfated diester (the initial form intracellularly), which was further hydrolyzed at a low rate to yield OA [27
The total content of DSP compounds per unit cell (intracellular and extracellular) increased continuously, suggesting that the production rate of DSP compounds was greater than the elimination rate throughout incubation. During the 25-day incubation period, the content of intracellular DSP compounds was essentially higher than that in the culture medium. However, the content of 5,7-dihydroxy-2,4-dimethylene-heptyl okadaate was the opposite, which might be caused by its different metabolic patterns, relative to other DSP compounds. The total concentration of extracellular DSP compounds in the culture medium was within 57.70–79.63 μg/L, and accounted for 4.81%–16.31% of the total toxin content. Hence, the content of extracellular DSP compounds in the culture medium should not to be underestimated. Li et al. [26
] examined the monthly variations of DSP toxins in the coastal seawaters of Qingdao, with an OA concentration range of 1.41–89.52 ng/L. Nevertheless, the concentration ranges of OA (19.9–34.0 μg/L) and DTX1 (15.2–27.9 μg/L) in the culture medium of P. lima
were much higher than those in seawater. Previous studies [41
] reported that the algal cell density is generally within 102
cells/mL during red tide outbreaks. Similarly, the cell density in this work was within the 1.28–2.58 × 104
cells/mL during P. lima
incubation, which corresponded to the density at the time of red tide. Although the reality in seawater and in vitro culture media is indeed different in terms of aspects such as nutrients, temperature, light and the stimulation of harsh environment, all of which will likely affect the production of toxins. The concentration of DSP compounds in vitro culture media can still provide a certain reference for assessing the environmental risk of extracellular toxins during red tide outbreaks. It can be deduced that the concentrations of DSP toxins could potentially reach tens of or even hundreds of micrograms per liter in the seawater during red-tide outbreaks of DSP-producing algae. If so, extracellular toxins would pose a huge threat to various marine life, especially to cultivated shellfish.