Quantitative profiling of sphingolipids in wild Cordyceps and its mycelia by using UHPLC-MS

In the present study, 101 sphingolipids in wild Cordyceps and its five mycelia were quantitatively profiled by using a fully validated UHPLC-MS method. The results revealed that a general rank order for the abundance of different classes of sphingolipids in wild Cordyceps and its mycelia is sphingoid bases/ceramides > phosphosphingolipids > glycosphingolipids. However, remarkable sphingolipid differences between wild Cordyceps and its mycelia were observed. One is that sphingoid base is the dominant sphingolipid in wild Cordyceps, whereas ceramide is the major sphingolipid in mycelia. Another difference is that the abundance of sphingomyelins in wild Cordyceps is almost 10-folds higher than those in most mycelia. The third one is that mycelia contain more inositol phosphorylceramides and glycosphingolipids than wild Cordyceps. Multivariate analysis was further employed to visualize the difference among wild Cordyceps and different mycelia, leading to the identification of respective sphingolipids as potential chemical markers for the differentiation of wild Cordyceps and its related mycelia. This study represents the first report on the quantitative profiling of sphingolipids in wild Cordyceps and its related mycelia, which provided comprehensive chemical evidence for the quality control and rational utilization of wild Cordyceps and its mycelia.

In 1994, myriocin, a natural sphingolipid (SPL), was isolated from the culture broth of Isaria sinclairii (the imperfect stage of Cordyceps sinclairii) as a potent immunosuppressive constituent 11 . Starting from myriocin, FTY720 was synthesized and finally developed into a drug (Fingolimod) for the treatment of multiple sclerosis and was studied in phase III clinical trial of organ transplantation. This implies that SPLs might be active constituents of wild Cordyceps and its mycelia.
SPLs are a complex family of compounds that share a common structural feature, a sphingoid base backbone which is biosynthesized from serine and a long-chain fatty acyl-CoA, and then converted into ceramides, phosphosphingolipids, glycosphingolipids and other species 12 . Substantial studies have demonstrated crucial role of endogenous SPLs in various biological procedures. Meanwhile, numerous studies have revealed significant bioactivities of natural and chemically synthesized SPLs. For instance, sphingoid bases have been regarded as potential anticancer agents, as represented by safingol 13,14 and 1-deoxysphinganine 15,16 , both of which are being evaluated in phase I clinical trials. Recent study also suggested that structural analogues of ceramide (C16-serinol) and exogenous natural ceramide exhibited promising anticancer effects 17,18 . Additionally, evidence showed that sphingomyelin has effects on the post-initiation development of preneoplastic lesions in the rat colon 19 . These evidence strongly suggested that natural SPLs are pharmacologically active constituents of natural medicines. Therefore, comprehensive study on the SPLs in wild Cordyceps and its mycelia is desperately needed for the clarification of substantial basis of their pharmacological activities and therapeutic effects. We herein carried out a quantitative profiling of the SPLs of wild Cordyceps and its five mycelia by using an improved sphingolipidomic approach that has been established in our lab 20 .

Results
Identification of SPLs in Cordyceps mycelia. Using the method established in our lab 20 , SPLs in samples were identified on the basis of high-resolution MS and MS/MS data, matching of SPLs with comprehensive SPL database, and confirmation of SPL standards. As the result, a total of 101 SPLs were identified from a pooled sample of five mycelia (Supplementary material, Table S1), including 10 sphingosines (So), 6 sphinganines (Sa), one sphingoid base 1-phosphate (S1P), one lysosphingomyelin (S1Po), 51 ceramides (Cer), 8 hexosyl ceramides (HexCer), 19 sphingomyelins (SM), 3 inositol phosphorylceramides (PI-Cer) and 2 mannosylinositol phosphorylceramides (MIPC). It can be seen that Cers are the most diverse SPLs in mycelia, followed by SMs and Sos. Among the 101 identified SPLs, 88 species were also identified in wild Cordyceps.
Method validation for quantitation. To quantitatively profile the SPLs, a MRM-based method established in our lab was further validated for the quantification of SPLs in wild Cordyceps and its mycelia.
Linearity, limit of detection (LOD) and limit of quantitation (LOQ). As shown in Table 1, all 6 calibration curves exhibited good linearity (correlation coefficients r 2 ≥ 0.9977) over wide dynamic ranges which spanned more than 2 orders of magnitude. The LOD ranged between 0.005 and 3.34 nM, whereas LOQ ranged between 0.0167 and 16.7 nM.
Injection precision, intra-and inter-day precision. As shown in Table 2, the relative standard deviations (RSD) of the levels of all quantified endogenous SPLs in samples were less than 6% for injection precision tests. Satisfactory RSD median values (3.79% and 3.96%) were achieved for intra-and inter-day precisions.
Stability. As shown in Table 2, The RSDs for the levels of approximately 90% endogenous SPLs in samples were less than 10% for stability tests. The RSDs for the levels of over 90% endogenous SPLs in QC samples were less than 15% across the whole process.

Quantitative comparison of SPLs in wild Cordyceps and its mycelia.
By using the validated method, totally 101 SPLs in wild Cordyceps and its mycelia were quantitatively profiled. To view the overall distribution of SPLs, the total contents of four classes of SPLs were calculated (Fig. 1), which showed a rank order of sphingoid bases > ceramides > phosphosphingolipids ≫ glycosphingolipids in wild Cordyceps, and a different order (ceramides > sphingoid bases ≫ phosphosphingolipids/glycosphingolipids) in most mycelia. In mycelia, some special distribution features were observed for Cordyceps sinensis and Hirsutella sinensis. In Cordyceps sinensis, sphingoid bases were relatively rich and their total content was higher than that of ceramides, which is a reversed rank order of sphingoid bases and ceramides for the other mycelia. In Hirsutella sinensis, however, the total content of sphingoid bases was notably low whereas the abundance of phosphosphingolipids was relatively high, which led to a reversed rank order of sphingoid bases and phosphosphingolipids for the other mycelia. The quantitative results also revealed distribution pattern of individual SPLs of wild Cordyceps and its mycelia, which was described by classes as below.
Quantification of sphingoid bases. Totally 18 sphingoid bases were quantified, including 10 Sos, 6 Sas, one S1P and one S1Po (Fig. 2). Levels of the four subclasses in both wild Cordyceps and its mycelia were generally in an order of So > Sa > S1P > S1Po, except for Hirsutella sinensis and Gliocadium roseum, in which Sa was more abundant than So and the content of S1P was extremely low (as low as 0.66 and 1.56 pmol/mg, vs about 7.17 pmol/mg in other samples). Structurally, sphingoid bases with carbon chain length > = 18 were more abundant than those species with carbon chain length < 18 in terms of both number of species and content in all samples. When comparing different samples, the total content of sphingoid bases was found to be highest in Cordyceps sinensis, followed by wild Cordyceps (including individual parts fruiting body of wild Cordyceps and sclerotium of wild Cordyceps), Cephalospovium sinensis, Mortierella SP, Gliocadium roseum and Hirsutella sinensis. Of note, the total content of sphingoid bases in Cordyceps sinensis was more than 3-folds of that in Hirsutella sinensis.
Quantification of ceramides. Ceramides were the most structurally diverse species. A total of 51 ceramides were quantitatively determined ( Figure S2). It can be seen that ceramides with d18 sphingoid base backbone (18 carbons of sphingoid base backbone with two hydroxyl groups) were the most structurally diverse (28-38 species) and dominant species, followed by those species with t18 sphingoid base backbone (18 carbons of sphingoid base backbone with three hydroxyl groups, also named phytosphingosine backbone) (9-10 species, with content of approximately 30% of that of d18 species). Of note, two uncommon Cers with odd-numbered sphingoid base backbone [Cer (d19:2/16:0) and Cer (d19:2/16:0(OH))] and one Cer with very short chain length sphingoid base backbone (d14) were found as low-abundance species in most samples. The d19 species was relatively rich in Hirsutella sinensis and wild Cordyceps. Comparison of the total content of Cers among samples suggested that Cephalospovium sinensis has the highest content of Cers, followed by other mycelia, whereas wild Cordyceps has the lowest content of Cers. Fig. 3, a total of 24 phosphosphingolipids belonging to three subclasses were measured, including 19 SMs, 3 PI-Cers and 2 MIPCs. In both wild Cordyceps and its mycelia, SMs and/or PI-Cers were dominant species whereas MIPCs were minor SPLs. Even though, the levels of SMs and PI-Cers varied remarkably. In Cordyceps sinensis, the abundance of SMs (0.69 pmol/mg) was about 14-folds lower than that of PI-Cers (10.32 pmol/mg), while in Cephalospovium sinensis, Gliocadium roseum and Mortierella SP, the total content of SMs (about 1.5-8.3 pmol/mg) was comparable to that of PI-Cers (2.4-10.2 pmol/mg). However, different from both Cordyceps sinensis and Cephalospovium sinensis etc, the level of SMs in Hirsutella sinensis and wild Cordyceps was notably high (about 35-42 pmol/mg), while PI-Cers was much less (about 0.1-1.4 pmol/mg).

Quantitation of glycosphingolipids.
Glycosphingolipids observed in all samples belong to HexCers. Among which, species with sphingoid base bone of d19:2 accounted for majority of the structures (more than 80%). Totally 8 HexCers were quantified and the results were shown in Figure S3. It can be seen that in all samples, the total levels of HexCers in mycelia were generally 2-5 folds of that in wild Cordyceps, except for Gliocadium roseum of which the HexCer's abundance was even lower than wild Cordyceps.
Difference among sphingolipidomes of wild Cordyceps and its mycelia. The supervised partial least squares discriminant analysis (PLS-DA) model was employed to visualize the general classification of samples. As shown in Fig. 4A, QC samples employed in the analysis were clustered into one group in the PLS-DA score scatter 3D plot. It can be seen that wild Cordyceps was separated from its mycelia at PLS 1 vector, Hirsutella sinensis was separated from other mycelia at PLS 2 vector, and Cordyceps sinensis, Cephalospovium sinensis and other mycelia were separated at PLS 3 vector. PLS-DA was further performed for wild Cordyceps and its mycelia and the plot was shown in Fig. 4B. It can be seen that wild Cordyceps and its mycelia were separated from each other by using this PLS-DA model with a high R 2 Y value of 0.55-0.66 and Q 2 value of 0.94-0.97. Based on PLS-DA analysis, the SPLs with VIP ≥ 1 were selected as potential markers and further confirmed by using T-test. Finally, potential markers for the classification of wild Cordyceps and Hirsutella sinensis, Cordyceps sinensis, Cephalospovium sinensis, Mortierella SP, Gliocadium roseum were selected as three SPLs with the highest VIP values, respectively (Table S2). Significant differences of 5 representative markers [HexCer (t19:1/16:1), Sa (d16:0), SM (d18:2/16:0(OH)), Cer (d18:0/16:1) and Cer (d18:1/16:1)] between wild Cordyceps and its mycelia were shown in Fig. 4C. Additionally, potential markers for the classification of different mycelia were found by PLS-DA model analysis (Table S2), in which sphingoid bases was the highest contributing SPL for the classification of most mycelia.

Discussions
Our study represents the first report for the quantitative profiling of SPLs in wild Cordyceps and its mycelia. By using the fully validated UHPLC-MS method, 101 sphingolipids in wild Cordyceps and its five mycelia were quantitatively profiled. The results revealed a general rank order for the abundance of different classes of sphingolipids in wild Cordyceps and its mycelia. However, remarkable sphingolipid differences between wild Cordyceps and its mycelia were observed. Among five kinds of mycelia, Cordyceps sinensis and Hirsutella sinensis present quite different profiles as evidenced by particularly high abundance sphingoid bases in Cordyceps sinensis and high level of phosphosphingolipids in Hirsutella sinensis.
It offers a robust method for evaluating the quality of wild Cordyceps and its mycelia in the aspect of sphingolipidomes. Based on multivariate analysis, respective sphingolipids were found as potential chemical markers for the differentiation of wild Cordyceps and different mycelia. In addition, our study provided comprehensive chemical evidences for the pharmacological effects and clinical efficacy of wild Cordyceps and its mycelia.   carbon and number of unsaturation degree (e.g. in d18:1/24:0, 24 means number of total carbon; 0 means number of unsaturation degree). For some SMs, annotation of total carbon numbers and total number of unsaturation degree denotes the sum of those of sphingoid base backbone and N-acyl chain.