Unveiling the Chemical Composition of Sulfur-Fumigated Herbs: A Triple Synthesis Approach Using UHPLC-LTQ-Orbitrap MS—A Case Study on Steroidal Saponins in Ophiopogonis Radix

Ophiopogonis Radix (OR) is a traditional Chinese medicine. In recent years, in order to achieve the purpose of drying, bleaching, sterilizing and being antiseptic, improving appearance, and easy storage, people often use sulfur fumigation for its processing. However, changes in the chemical composition of medicinal herbs caused by sulfur fumigation can lead to the transformation and loss of potent substances. Therefore, the development of methods to rapidly reveal the chemical transformation of medicinal herbs induced by sulfur fumigation can guarantee the safe clinical use of medicines. In this study, a combined full scan-parent ions list-dynamic exclusion acquisition-diagnostic product ions analysis strategy based on UHPLC-LTQ-Orbitrap MS was proposed for the analysis of steroidal saponins and their transformed components in sulfur-fumigated Ophiopogonis Radix (SF-OR). Based on precise mass measurements, chromatographic behavior, neutral loss ions, and diagnostic product ions, 286 constituents were screened and identified from SF-OR, including 191 steroidal saponins and 95 sulfur-containing derivatives (sulfates or sulfites). The results indicated that the established strategy was a valuable and effective analytical tool for comprehensively characterizing the material basis of SF-OR, and also provided a basis for potential chemical changes in other sulfur-fumigated herbs.


Introduction
Traditional Chinese medicine (TCM) needs to be processed before it can be used to treat various diseases.This is one of the characteristics of the clinical use of Chinese medicine.As a unique pre-processing method, sulfur fumigation (SF) is a highly efficient and vital traditional post-harvest handling process for foods, agriculture products, and TCM [1,2].The mechanism of SF is that sulfur burns at high temperatures to generate SO 2 , which prevents pest infestation, mold, and bacterial contamination and provides a favorable appearance [3].In a humid environment, SO 2 combines with water to produce reductive components, which play a role in reducing the coloring components to facilitate the drying of the material [4].However, residual SO 2 can induce respiratory symptoms such as cough, chest tightness, and throat irritation [5,6].Therefore, SO 2 content is proposed as an official standard requirement for the quality control of sulfur-fumigated medicinal herbs in China [7] and many other countries.In addition, SF can trigger chemical transformations of original bioactive components to generate characteristic sulfate and sulfite derivatives in fumigated herbs [8][9][10].However, the evaluation of Chinese medicinal materials for SF remains at the level of sulfur dioxide residues (including sulfite derivatives) in the Pharmacopoeia of the People's Republic of China.This does not truly reflect the transformations in the process of SF.Thus, it is essential to develop a rapid and sensitive approach to ascertain the SF state of a given medicinal herb for TCM quality control.
Traditionally, the detection of common constituents with predictable molecular weights is accomplished by acquiring full-scan LC/MS data followed by the generation of extracted ion chromatograms corresponding to their mass-to-charge (m/z) values.However, not all constituents, especially microconsitutents, can be detected in full-scan MS data because of their differing amounts and poor chromatographic separation.Their MS/MS acquisitions cannot be triggered when coeluted with the constituents of a relative higher content [11].Therefore, a new strategy to enhance the constituent detection and identification capacities of LC-MS/MS was established.Since the multiple constituents contained in a specific traditional herb are derived from one or more certain biosynthetic pathways, these constituents could usually be structurally classified into several chemical families with the same carbon skeletons or substructures.So, it is easily understood that their formulas and molecular weights are predictable.Additionally, constituents with the same carbon skeletons will undergo similar fragmentation pathways in collision-induced dissociation (CID) mode, and thus generate similar diagnostic product-ions (DPIs) from their common carbon skeletons.In other words, a series of DPIs representing a specific parent nucleus or substitution groups can be used as the characteristic peaks to select out the corresponding chemical family [12][13][14].
Ophiopogonis Radix (Maidong in Chinese, OR), a widely used TCM published initially in the Herbal Canon of Shen Nong, originates from the dried roots of Ophiopogon japonicus (L.f.) Ker-Gawl.According to traditional Chinese medicine theory, OR nourishes the yin, promotes body fluid production, moistens the lung, eases the mind, and clears heart fire [7].Phytochemical studies have revealed the presence of various biologically active compounds, including steroidal saponins, homoisoflavonoids, and polysaccharides.These compounds have therapeutic effects against acute and chronic inflammation, diabetes, cardiovascular diseases, and other disorders [15].However, there are limited studies focused on SF-OR.Due to the complexity of the steroidal saponin composition, the full chemical transformation of SF-OR has not been obtained so far.Whether sulfur fumigation alters the chemical composition of OR and the identification of new sulfur-containing derivatives have become important issues for the effectiveness and safety control of OR.Therefore, there is an urgent need to develop a rapid, generally reliable, and accurate analytical method to fully characterize the chemical transformation of OR induced by sulfur fumigation.
In this study, we established a laboratory simulation method to obtain SF-OR samples.Then, a UHPLC-LTQ-Orbitrap MS combined with parent ions list-dynamic exclusion (PIL-DE) acquisition and a DPIs analysis was developed as a strategy for the comprehensive screening and identification of the steroidal saponin constituents of SF-OR.Conversely, a case study was used to validate the effectiveness and feasibility of the proposed strategy.It provides a basis for the identification of other sulfur-fumigated herbal components.

Identification of Steroidal Saponins from SF-OR 2.1.1. The Establishment of an Analytical Strategy
An efficient and integrated strategy was established for the target identification of steroidal saponins and their sulfur-containing derivatives in SF-OR using a UHPLC-LTQ-Orbitrap MS coupled with post-acquisition data-mining processing techniques (Figure 1).First, the samples were injected into the UHPLC-LTQ-Orbitrap MS to gain highresolution mass spectrometry (HRMS) data with full MS scanning acquisition.Second, the reactions (hydroxylation, glucosylation, xylosation, and rhamnosylation) were predicted using information derived from structural characteristics, the literature, and compound databases.A preliminary screening of the candidate compounds was performed using the Thermo Xcalibur 2.1 software to obtain their retention times and accurate molecular weights.According to the PIL-DE scan mode, multi-stage mass spectrometry data were collected.Third, to guide the subsequent rapid analysis, these compounds' DPI and specific neutral loss filter (NLF) were summarized based on the mass spectrometric cracking rules reported in the literature and the cracking information of reference substances.

Construction of Ion Lists
The Thermo Xcalibur 2.1 software was used to accurately calculate the molecular weights of the above candidate formulas with an error of ±10 ppm.Ion peaks with an intensity greater than 1.0 × 10 4 in the full scan map (Figure 3) were extracted as potential steroidal saponins of OR.For example, taking compounds with a molecular weight at m/z 721.4157, seven chromatographic peaks were extracted from the HR-MS 1 .Only the secondary mass spectrum of peak-7 was obtained (Figure 4A), while, in the PIL-DE scan mode, the ESI-MS/MS spectra of these seven peaks were obtained through one data acquisition (Figure 4B).The PIL-DE scan mode dramatically increases the information acquisition efficiency of multi-stage mass spectra.

Construction of Ion Lists
The Thermo Xcalibur 2.1 software was used to accurately calculate the molecular weights of the above candidate formulas with an error of ±10 ppm.Ion peaks with an intensity greater than 1.0 × 10 4 in the full scan map (Figure 3) were extracted as potential steroidal saponins of OR.For example, taking compounds with a molecular weight at m/z 721.4157, seven chromatographic peaks were extracted from the HR-MS 1 .Only the secondary mass spectrum of peak-7 was obtained (Figure 4A), while, in the PIL-DE scan mode, the ESI-MS/MS spectra of these seven peaks were obtained through one data acquisition (Figure 4B).The PIL-DE scan mode dramatically increases the information acquisition efficiency of multi-stage mass spectra.
steroidal saponins of OR.For example, taking compounds with a molecular weight at m/z 721.4157, seven chromatographic peaks were extracted from the HR-MS 1 .Only the secondary mass spectrum of peak-7 was obtained (Figure 4A), while, in the PIL-DE scan mode, the ESI-MS/MS spectra of these seven peaks were obtained through one data acquisition (Figure 4B).The PIL-DE scan mode dramatically increases the information acquisition efficiency of multi-stage mass spectra.

Analysis of the Characteristic Fragmentation Mechanism of Steroidal Saponins
The DPIs and NLFs were speculated according to the summarized fragmentation behaviors of known compounds, which supplied tremendous help in identifying the secondary metabolites in Ophiopogonis Radix.

The Characteristic Fragmentation Mechanism of Type-I Steroidal Saponins
As shown in Figure 5, the core structure of Type-I steroidal saponins was spirostanol, which generally had only one carbohydrate chain a ached to the C3 position.In ESI − mode, carbohydrates were removed one by one until a monosaccharide remained on the core structure; hence, most of their aglycon fragment ions were not observed.Take ophiopogonin C′ as an example (Figure 6); it gave rise to a [M − H] − ion at m/z 721.4157 (C39H61O12, <5 ppm) in negative mode.In its ESI-MS/MS spectrum, the product-ion at m/z 575 ([M − H − Rha] − ) indicated the presence of the Rha group.According to the literature [21][22][23], there was a Glc group in ophiopogonin C′, while NLF of 162 Da was not observed in this experiment.This observation indicated that the Glc group should be directly attached to the core structure, and Rha was a ached to the Glc.Based on the above analysis, the molecular formula of the core structure was deduced to be C27H42O3 with a molecular

Analysis of the Characteristic Fragmentation Mechanism of Steroidal Saponins
The DPIs and NLFs were speculated according to the summarized fragmentation behaviors of known compounds, which supplied tremendous help in identifying the secondary metabolites in Ophiopogonis Radix.

The Characteristic Fragmentation Mechanism of Type-I Steroidal Saponins
As shown in Figure 5, the core structure of Type-I steroidal saponins was spirostanol, which generally had only one carbohydrate chain attached to the C 3 position.In ESI − mode, carbohydrates were removed one by one until a monosaccharide remained on the core structure; hence, most of their aglycon fragment ions were not observed.Take ophiopogonin C ′ as an example (Figure 6); it gave rise to a [M − H] − ion at m/z 721.4157 (C 39 H 61 O 12 , <5 ppm) in negative mode.In its ESI-MS/MS spectrum, the product-ion at m/z 575 ([M − H − Rha] − ) indicated the presence of the Rha group.According to the literature [21][22][23], there was a Glc group in ophiopogonin C ′ , while NLF of 162 Da was not observed in this experiment.This observation indicated that the Glc group should be directly attached to the core structure, and Rha was attached to the Glc.Based on the above analysis, the molecular formula of the core structure was deduced to be C 27  S2).The molecular formula of the core structure was eventually identified as C 27 H 42 O 5 (446 Da), which had one more hydroxyl group than that of ophiopogonin B. Due to the polyhydroxy substitution, these compounds generated the fragment ion of the core structure in the ESI − mode, which could be used as a particular fragmentation behavior to provide a reference for subsequent identification.8).According to the literature [24], a Glc group indicated that the Glc group should be directly a ached to the core structure, and Rha was a ached to the Glc.It was concluded that the molecular formula of the core was C27H40O3 with a molecular weight of 412 Da.   8).According to the literature [24], a Glc group indicated that the Glc group should be directly a ached to the core structure, and Rha was a ached to the Glc.It was concluded that the molecular formula of the core was C27H40O3 with a molecular weight of 412 Da.  8).According to the literature [24], a Glc group indicated that the Glc group should be directly a ached to the core structure, and Rha was a ached to the Glc.It was concluded that the molecular formula of the core was C27H40O3 with a molecular weight of 412 Da.The Characteristic Fragmentation Mechanism of Type-IV Steroidal Saponins The core structure of Type-IV was deformed furostanol with a C5-C6 double bond and two carbohydrates at C3 and C26 (shown in Figure 9).It is worth noting that the carbohydrate chain at position C26 was generally substituted by Glc, which appeared as a specific NLF of 180 Da in the ESI-MS 2     The Characteristic Fragmentation Mechanism of Type-IV Steroidal Saponins The core structure of Type-IV was deformed furostanol with a C5-C6 double bond and two carbohydrates at C3 and C26 (shown in Figure 9).It is worth noting that the carbohydrate chain at position C26 was generally substituted by Glc, which appeared as a specific NLF of 180 Da in the ESI-MS    The Characteristic Fragmentation Mechanism of Type-VI Steroidal Saponins The core structure of Type-VI was furostanol with two carbohydrate chains at C3 C26.In negative mode, the carbohydrates were removed one by one until no monosac rides remained on the core structure; hence, their aglycon fragment ions could be served.The carbohydrate chain at C26 generally contained 1-2 Glc, which would ap as a specific NLF of 180 Da in MS 2 spectra.The characteristic fragmentation mechanis Type-V steroidal Saponins is shown in Figure 12.Ophiofurspiside M generated the H] − ion at m/z 917.4741, with the molecular formula of C45H73O19 and a mass error w 5 ppm.In the ESI-MS 2

The Characteristic Fragmentation Mechanism of Type-VI Steroidal Saponins
The core structure of Type-VI was furostanol with two carbohydrate chains at C 3 and C 26 .In negative mode, the carbohydrates were removed one by one until no monosaccharides remained on the core structure; hence, their aglycon fragment ions could be observed.The carbohydrate chain at C 26 generally contained 1-2 Glc, which would appear as a specific NLF of 180 Da in MS 2 spectra.The characteristic fragmentation mechanism of Type-V steroidal Saponins is shown in Figure 12  The different core structures of the six types of steroidal saponins, combined with the relevant information reported in the literature [15], allow the fragmentation mechanism of these compounds in negative ion mode to be analyzed and the NLFs and DPIs to be summarized.The potential NLFs of the OR Steroidal Saponins are summarised in Table 1.The different core structures of the six types of steroidal saponins, combined with the relevant information reported in the literature [15], allow the fragmentation mechanism of these compounds in negative ion mode to be analyzed and the NLFs and DPIs to be summarized.The potential NLFs of the OR Steroidal Saponins are summarised in Table 1.Congeneric compounds generally have a similar MS fragmentation regularity, thereby generating characteristic DPIs that can represent the structure of such compounds.However, it is difficult to accurately identify the structure of natural compounds with a large molecular weight and relatively complex structures based on one DPI.Therefore, in this experiment, the concept of pDPI was proposed to provide meaningful guidance for the rapid identification of OR steroidal saponins.According to the components identified and reported in the literature, the pDPIs of the six types of core structures are summarized, as shown in Table 2.A total of 9 pDPIs of Type-I, 2 pDPIs of Type-II and Type-III, 1 pDPI of Type IV, 1 pDPI of Type-V, and 3 pDPIs of Type-VI were found.pounds.However, it is difficult to accurately identify the structure of natural compounds with a large molecular weight and relatively complex structures based on one DPI.Therefore, in this experiment, the concept of pDPI was proposed to provide meaningful guidance for the rapid identification of OR steroidal saponins.According to the components identified and reported in the literature, the pDPIs of the six types of core structures are summarized, as shown in Table 2.A total of 9 pDPIs of Type-I, 2 pDPIs of Type-II and Type-III, 1 pDPI of Type IV, 1 pDPI of Type-V, and 3 pDPIs of Type-VI were found.pounds.However, it is difficult to accurately identify the structure of natural compounds with a large molecular weight and relatively complex structures based on one DPI.Therefore, in this experiment, the concept of pDPI was proposed to provide meaningful guidance for the rapid identification of OR steroidal saponins.According to the components identified and reported in the literature, the pDPIs of the six types of core structures are summarized, as shown in Table 2.A total of 9 pDPIs of Type-I, 2 pDPIs of Type-II and Type-III, 1 pDPI of Type IV, 1 pDPI of Type-V, and 3 pDPIs of Type-VI were found.pounds.However, it is difficult to accurately identify the structure of natural compounds with a large molecular weight and relatively complex structures based on one DPI.Therefore, in this experiment, the concept of pDPI was proposed to provide meaningful guidance for the rapid identification of OR steroidal saponins.According to the components identified and reported in the literature, the pDPIs of the six types of core structures are summarized, as shown in Table 2.A total of 9 pDPIs of Type-I, 2 pDPIs of Type-II and Type-III, 1 pDPI of Type IV, 1 pDPI of Type-V, and 3 pDPIs of Type-VI were found.3.  3.  3.

Detection and Structural Elucidation of OR Steroidal Saponins
Based on the established pDPIs and NLFs, 191 steroidal saponins were quickly identified from OR.There were 105 Type-I, 12 Type-II and Type-III, 13 Type-IV, 18 Type-V, and 42 Type-VI OR steroidal saponin compounds screened and identified.The detailed MS data information is shown in Table 3.

Screening of the Candidate Molecular Weight of Sulfur-Containing Derivatives of Steroidal Saponins
The accurate [M − H] − of the candidate molecular formula was calculated in Section 2.2.1, the chromatographic peak from the ESI-MS spectra was extracted, and the peaks with an intensity > 1.0 × 10 4 were selected as the potential sulfur-containing derivatives of Steroidal Saponins.According to the established PIL-DE scan mode, MS 2 data collection was performed.

Identification of Sulfur-Containing Derivatives of Steroidal Saponins
Based on the established structure identification strategy with pDPIs and NLFs, rapid screening of the steroidal saponins' sulfur-containing derivatives was carried out.In addition, the isotope peak [M − H+2] − of the sulfur-containing derivatives under ultra-high resolution split into two peaks with a mass difference of 0.01 Da.Therefore, 12   The reported pathways of ginsenosides during sulfur fumigation are sulfation and sulfite [25].In addition, sulfur generates SO2 at high temperatures to lower the pH, making the glycosides easily hydrolyzed.Therefore, to fully characterize the sulfur-containing derivatives of steroidal saponins in OR, six types of steroidal saponins were used as the core structure, and Rha (0-2), Fuc (0-2), Xyl (0-2), Glc (0-4), Ara (0-1), Ac (0-2), SO2 (0-1), and SO3 (0-1) were used as substituents for molecular design.

Screening of the Candidate Molecular Weight of Sulfur-Containing Derivatives of Steroidal Saponins
The accurate [M − H] − of the candidate molecular formula was calculated in Section 2.2.1, the chromatographic peak from the ESI-MS spectra was extracted, and the peaks with an intensity > 1.0 × 10 4 were selected as the potential sulfur-containing derivatives of Steroidal Saponins.According to the established PIL-DE scan mode, MS 2 data collection was performed.

Identification of Type-IV Sulfur-Containing Derivatives of Steroidal Saponins
There were no Type-IV sulfur-containing derivatives of steroidal saponins detected in SF-OR.14A).In the 100,000 FWHM @ 400 m/z ultra-high resolution mode, its isotope peak at [M − H+2] − showed two peaks at m/z 803.3691 and m/z 803.3810 due to the existence of element "S".In the MS 2 spectrum (Figure 14B), the DPI at m/z 653 ([M − H-146] − ) indicated the presence of a Rha or a Fuc.The pDPI at m/z 799/653 of SS30 was 79.95 Da more than the pDPI at m/z 719/573 of S106, indicating that these two compounds had the exact fragmentation mechanism.Therefore, SS30 was identified as the sulfated product of S106.According to the characteristic fragmentation mechanism of Type-II and Type-III steroidal saponins and the specificity of the sulfur-containing compound isotope peaks, a total of three Type-II and Type-III sulfur-containing derivatives of steroidal saponins were identified and targeted from SF-OR.They were all sulfated products, as shown in Table 4.  Based on the mass characteristic fragmentation mechanism of Type-V steroidal saponins, a total of eight Type-V steroidal saponins' sulfur derivatives were identified from the targeted screening of OR, all of which belonged to sulfated products, as shown in Table 4. Based on the mass characteristic fragmentation mechanism of Type-V steroidal saponins, a total of eight Type-V steroidal saponins' sulfur derivatives were identified from the targeted screening of OR, all of which belonged to sulfated products, as shown in Table 4.   S9), meaning that their fragmentation mechanism was consistent with that of Trigoneoside IVa.Therefore, SS72-SS74 was eventually identified as Trigoneoside IVa sulfite.

Identification of
A total of 55 Type-VI steroidal saponins' sulfur derivatives were identified from the targeted screening of OR based on the fragmentation mechanism of Type-VI steroidal saponins, in which 10 compounds were sulfite products and 45 were sulfate products, as shown in Table 4.

Discussion
UHPLC-LTQ-Orbitrap has become a powerful tool for drug analysis with its high sensitivity, accuracy, and separation ability [27].However, the spectral information contained in existing chemical standards and databases is minimal.The LC-MS technique alone cannot satisfy the structural characterization of complex and diverse TCM [28].Therefore, more new techniques and research strategies are needed to meet this challenge.Based on this, the study proposes a combined full scan-parent ions list-dynamic exclusion acquisition-diagnostic product ions analytical strategy for revealing the chemical composition of sulfur-fumigated TCM.The core technique of the strategy is to predict the possible molecular structures based on the structural characteristics of the compounds.A UHPLC-LTQ-Orbitrap high-resolution mass spectrometer is then utilized for data acquisition.After the initial screening of candidate compounds, multi-stage mass spectrometry data are targeted based on the PIL-DE scanning mode [12].By studying the mass spectral  S9), meaning that their fragmentation mechanism was consistent with that of Trigoneoside IVa.Therefore, SS72-SS74 was eventually identified as Trigoneoside IVa sulfite.
A total of 55 Type-VI steroidal saponins' sulfur derivatives were identified from the targeted screening of OR based on the fragmentation mechanism of Type-VI steroidal saponins, in which 10 compounds were sulfite products and 45 were sulfate products, as shown in Table 4.

Discussion
UHPLC-LTQ-Orbitrap has become a powerful tool for drug analysis with its high sensitivity, accuracy, and separation ability [27].However, the spectral information contained in existing chemical standards and databases is minimal.The LC-MS technique alone cannot satisfy the structural characterization of complex and diverse TCM [28].Therefore, more new techniques and research strategies are needed to meet this challenge.Based on this, the study proposes a combined full scan-parent ions list-dynamic exclusion acquisition-diagnostic product ions analytical strategy for revealing the chemical composition of sulfur-fumigated TCM.The core technique of the strategy is to predict the possible molecular structures based on the structural characteristics of the compounds.A UHPLC-LTQ-Orbitrap high-resolution mass spectrometer is then utilized for data acquisition.After the initial screening of candidate compounds, multi-stage mass spectrometry data are targeted based on the PIL-DE scanning mode [12].By studying the mass spectral cleavage behavior of the representative compounds, the characteristic cleavage patterns of com-pounds with different parent nucleus structures are summarized, and the DPIs and NLFs of such compounds are inferred.In addition, for natural products with a high molecular weight and relatively complex structures, it is often difficult to accurately characterize their structures based on one DPI [29].Therefore, the concept of pDPI is proposed in the strategy.Accordingly, the rapid identification of chemical components is performed.
The proposed strategy was extensively investigated for the steroidal saponins of SF-OR.The successful implementation of the strategy accelerated the identification of the steroidal saponin components, expanded the search scope, and ensured the accuracy of the component identification.Although the strategy has many advantages, there are still some limitations.First, complex matrices and impurities in the samples may interfere with the analytical results, affecting accuracy and sensitivity.Second, the parent ion scanning technique has certain requirements for the sample concentration, as too high or too low will affect the detection sensitivity.Nevertheless, this innovative measure has broad application value in scientific research and medical diagnosis, while also providing a basis for the identification of compounds and their pharmacological activity research.
In this study, we summarized 6 types of 191 steroidal saponins of SF-OR based on the identified components and research reports [30][31][32][33].Most of them were dominated by Type-I spirostanol matric structures, which generally had only one sugar chain.The representative compounds Ophiopogonin A, B, C, and D have a wide range of pharmacological effects, such as hypoglycemia [34], antitumor [35], the protection of myocardial ischemia [36], regulation of immunity [37], and resistance to myocardial infarction [38].It was found that the steroidal saponin constituents are susceptible to the partial removal of sugar groups during SF and the further formation of corresponding sulfur-containing derivatives.A total of 95 sulfurcontaining derivatives of steroidal saponin were found in the identification of SF-OR.It has been shown that SF decreases the total content of polysaccharides and increases the content of oligosaccharides and free monosaccharides in TCM.Although SO 2 residues decreased during storage, chemical transformations of non-saccharide and sugar components continued to occur in the TCM [39].In addition, SF induces chemical transformations of the bioactive components in TCM.A reduction in the bioactive components may affect their efficacy, and the generation of new sulfur-containing derivatives may also affect their safety [40].For example, three highly toxic and carcinogenic components to humans were found in the volatile oil of sulfurfumigated Radix Angelicae Dahuricae [41].The anti-inflammatory and anti-tumor effects of sulfur-fumigated Astragalus membranaceus were substantially reduced [42].The analgesic effect of Paeoniae Radix Alba was reduced after SF [43].Therefore, when applying sulfurfumigated TCM, attention should be paid to the transformation of its chemical composition and safety evaluation.
So far, there is very limited information on SF-induced changes in the steroidal saponin composition of OR, which can be analyzed quickly and efficiently using our proposed strategy.It provides a model for the identification and pharmacological activity studies of other sulfur-fumigated TCM in the future.

Chemicals and Reagents
The NSF-OR sample was purchased from Beijing Tong Ren Tang in China.The voucher specimen of NSF-OR was deposited in the School of Chinese Pharmacy, Beijing University of Chinese Medicine, Beijing, China.
Methanol and acetonitrile were MS-grade and purchased from Fisher Scientific (Fair Lawn, NJ, USA).Formic acid (MS grade) was provided by Sigma Aldrich (St. Louis, MO, USA).All the other analytical-grade chemicals were available at the workstation of Beijing Chemical Works (Beijing, China).The deionised water used throughout the experiment was purified using the Milli-Q Gradient A 10 System (Millipore, Billerica, MA, USA).Grace Pure SPE C 18 -Low solid-phase extraction cartridges (200 mg/3 mL, 59 µm, 70 Å) were purchased from Grace Davison Discovery Science (Deerfield, IL, USA).

Peak Selections and Data Processing
A Thermo Xcalibur 2.1 workstation was used for the data acquisition and processing.Peaks with an intensity over 10,000 in negative ion mode were selected as fragment ions.The chemical formulas attributed to the selected peaks were calculated using a formula predictor by setting the parameters as follows: C , H , O , S [0-2], and the ring double-bond (RDB) equivalent value .The maximum mass errors between the measured and calculated values were fixed within 5 ppm.All the relevant data, including the peak number, retention time, accurate mass, predicted chemical formula, and corresponding mass error, were recorded.

Conclusions
In this study, an analytical strategy for PIL-DE acquisition combined with DPIs and NLFs based on UHPLC-LTQ-Orbitrap MS was proposed.Steroidal saponins and their transformed components in SF-OR were analyzed according to the method.As a result, 191 steroidal saponins were quickly screened and identified from SF-OR.There were 105 Type-I, 12 Type-II and Type-III, 13 Type-IV, 18 Type-V, and 43 Type-VI.In addition, the occurrence of sulfate or sulfite esterification reactions led to the emergence of many new sulfur-containing derivatives in the SF process.Based on the summarized MS cleavage patterns and specific isotope peaks of the steroidal saponins, 95 sulfated and sulfite-esterified components of steroidal saponins were identified from the SF-OR, including 29 Type-I, 3 Type-II and Type-III, 8 Type-V, and 55 Type-VI.The implementation of this new strategy comprehensively characterized the compositional profile of the steroidal saponins in SF-OR.Meanwhile, this case also confirmed the feasibility of the new strategy for the identification of the chemical compositions of SF-herbs, which provides a reference for the compositional identification of other Chinese herbal medicines.

Figure 1 .
Figure 1.Summary diagram of the developed strategy and methodology.Figure 1. Summary diagram of the developed strategy and methodology.

Figure 1 .
Figure 1.Summary diagram of the developed strategy and methodology.Figure 1. Summary diagram of the developed strategy and methodology.

Figure 2 .
Figure 2. Core structures of six steroidal saponins.Type-I: spirostanol; Type-II and Type-III: deformed spirostanol with a C5-C6 double bond and a carbohydrate at C3; Type-IV: deformed furostanol with a C5-C6 double bond and two carbohydrates at C3 and C26; Type-V: deformed furostanol with ring opening producing two carbonyls in the E-ring.

Figure 2 .
Figure 2. Core structures of six steroidal saponins.Type-I: spirostanol; Type-II and Type-III: deformed spirostanol with a C 5 -C 6 double bond and a carbohydrate at C 3; Type-IV: deformed furostanol with a C5-C6 double bond and two carbohydrates at C3 and C26; Type-V: deformed furostanol with ring opening producing two carbonyls in the E-ring.

Figure 3 .Figure 3 .Figure 4 .
Figure 3.The total ion chromatograms of SF-OR were obtained in full scan mode.

Figure 4 .
Figure 4. Data acquisition efficiency of compound m/z 721.4157 in full scan mode (A) and PIL-DE mode (B).

Figure 7 .
Figure 7.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of 14-hydroxydiosgenin 3-O-α-Lrha-(1→2)-β-D-glc in negative ion mode.The Characteristic Fragmentation Mechanism of Type-II and Type-III Steroidal Saponins The core structures of Type-II and Type-III were deformed spirostanol with a C 5 -C 6 double bond and a carbohydrate at C 3 .Similarly, most of their aglycon fragment ions were not observed.(1β,3β)-3-hydroxyspirost-5, 25(27)-dien-1-yl-O-6-deoxy-α-L-Rha-(1→2)-β-Dglc presented the [M − H] − ion at m/z 719.4008 (C 39 H 59 O 12 , <5 ppm).In the ESI-MS/MS spectrum, the fragment ion at m/z 573 ([M − H − Rha] − ) indicated the presence of the Rha group (Figure8).According to the literature[24], a Glc group indicated that the Glc group should be directly attached to the core structure, and Rha was attached to the Glc.It was concluded that the molecular formula of the core was C 27 H 40 O 3 with a molecular weight of 412 Da.

Figure 9 .
Figure 9.The characteristic fragmentation mechanism of Type-IV steroidal Saponins.

Figure 9 .
Figure 9.The characteristic fragmentation mechanism of Type-IV steroidal Saponins.

Figure 9 .
Figure 9.The characteristic fragmentation mechanism of Type-IV steroidal Saponins.

Figure 11 .
Figure 11.The characteristic fragmentation mechanism of Type-V steroidal Saponins.

Figure 12 .
Figure 12.The characteristic fragmentation mechanism of Type-VI steroidal Saponins.

Figure 12 .
Figure 12.The characteristic fragmentation mechanism of Type-VI steroidal Saponins.2.1.5.Determination and Verification of NLFs and DPIs of OR Steroidal Saponins

6 . 6 .
Detection and Structural Elucidation of OR Steroidal Saponins Based on the established pDPIs and NLFs, 191 steroidal saponins were quickly identified from OR.There were 105 Type-I, 12 Type-II and Type-III, 13 Type-IV, 18 Type-V, and 42 Type-VI OR steroidal saponin compounds screened and identified.The detailed MS data information is shown in Table 3. VI 755(C 39 H 63 O 14 )/ 575(C 33 H 51 O 8 ) Molecules 2024, 29, x FOR PEER REVIEW 13 Detection and Structural Elucidation of OR Steroidal Saponins Based on the established pDPIs and NLFs, 191 steroidal saponins were quickly identified from OR.There were 105 Type-I, 12 Type-II and Type-III, 13 Type-IV, 18 Type-V, and 42 Type-VI OR steroidal saponin compounds screened and identified.The detailed MS data information is shown in Table

6 .
771(C 39 H 63 O 15 )/591(C 33 H 51 O 9 ) Molecules 2024, 29, x FOR PEER REVIEW 13 Detection and Structural Elucidation of OR Steroidal Saponins Based on the established pDPIs and NLFs, 191 steroidal saponins were quickly identified from OR.There were 105 Type-I, 12 Type-II and Type-III, 13 Type-IV, 18 Type-V, and 42 Type-VI OR steroidal saponin compounds screened and identified.The detailed MS data information is shown in Table

6 .
Detection and Structural Elucidation of OR Steroidal Saponins Based on the established pDPIs and NLFs, 191 steroidal saponins were quickly identified from OR.There were 105 Type-I, 12 Type-II and Type-III, 13 Type-IV, 18 Type-V, and 42 Type-VI OR steroidal saponin compounds screened and identified.The detailed MS data information is shown in Table
Identification of Type-I Sulfur-Containing Derivatives of Steroidal Saponins SS9, with the [M − H] − ion at m/z 801.3714 (C 39 H 61 O 15 S, <5 ppm), was 79.95 Da (SO 3 ) more than that of S2-S8.In the 100,000 FWHM @ 400 m/z ultra-high resolution mode, its isotope peak at [M − H+2] − formed two peaks at m/z 803.3691 and m/z 803.3810, further confirming that SS9 was a sulfur compound.In its ESI-MS/MS spectrum, the DPI at m/z 655 ([M − H − Rha] − ) was consistent with S2-S8 (Figure 13).The pDPI it generated at m/z 801/655 was 79.95 Da more than that of m/z 721/575 of S2-S8.Therefore, the compound SS9 was identified as the sulfated product of the compound S2-S8.Molecules 2024, 29, x FOR PEER REVIEW 21 of 34 the DPIs at m/z 787 ([M − H − Rha] − ) and 607 ([M − H − Rha-180] − ) indicated the core structure was furostanol, and Glc was present at C26.The pDPI at m/z 771/591 could also be applied to identify Type-VI steroidal saponins rapidly.2.2.Identification of Sulfur-Containing Derivatives of Steroidal Saponins from SF-OR 2.2.1.Molecular Design of Sulfur-Containing Derivatives of Steroidal Saponins Identification of Type-I Sulfur-Containing Derivatives of Steroidal Saponins SS9, with the [M − H] − ion at m/z 801.3714 (C39H61O15S, <5 ppm), was 79.95 Da (SO3) more than that of S2-S8.In the 100,000 FWHM @ 400 m/z ultra-high resolution mode, its isotope peak at [M − H+2] − formed two peaks at m/z 803.3691 and m/z 803.3810, further confirming that SS9 was a sulfur compound.In its ESI-MS/MS spectrum, the DPI at m/z 655 ([M − H − Rha] − ) was consistent with S2-S8 (Figure13).The pDPI it generated at m/z 801/655 was 79.95 Da more than that of m/z 721/575 of S2-S8.Therefore, the compound SS9 was identified as the sulfated product of the compound S2-S8.

Figure 13 .Figure 13 .
Figure 13.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS9 in negative ion mode.As shown in Figure S8, SS10-SS14 displayed the [M − H] − ion at m/z 817.3663 (C39H61O16S, 5 ppm).This was 79.95 Da greater than S11-S15 in negative ion mode, whichFigure 13.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS9 in negative ion mode.As shown in Figure S8, SS10-SS14 displayed the [M − H] − ion at m/z 817.3663 (C 39 H 61 O 16 S, 5 ppm).This was 79.95 Da greater than S11-S15 in negative ion mode, which implied that SS10-SS14 might be the sulfated product of S11-S15.Two isotope peaks at m/z 819.3615 and m/z 819.3772 further indicated that SS10-SS14 were sulfur-containing compounds.The DPI at m/z 671 ([M − H − Rha] − ) was consistent with S11-S15.The pDPI at m/z 817/671 was 79.95 Da greater than the pDPI at m/z 721/575 of S11-S15.Therefore, the compounds SS10-SS14 were presumed to be the sulfated products of S11-S15.According to the characteristic fragmentation mechanism of Type-I steroidal saponins and the specificity of the sulfur-containing compound isotope peaks, a total of 24 Type-I sulfurcontaining derivatives of steroidal saponins were identified from SF-OR.They were all sulfated products, as shown in Table4.

Figure 15 .
Figure 15.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS40 in negative ion mode.

Figure 14 .
Figure 14.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS30 in negative ion mode.
Identification of Type-V Sulfur-Containing Derivatives of Steroidal Saponins SS40 provided the [M − H] − ion at m/z 979.4192 (C 45 H 71 O 21 S) in the ESI-MS spectrum (Figure 15A).It was 79.95 Da greater than S131, indicating that it probably was the sulfated product of S131.Two isotope peaks at m/z 981.4125 and m/z 981.4264 proved that element "S" existed.The pDPI at m/z 817 ([M − H-146] − )/m/z 637 ([M − H-146-180] − ) was 79.95 Da more than that of S106 (Figure 15B), indicating that SS40 was the sulfonation product of S106.Molecules 2024, 29, x FOR PEER REVIEW 27 of 34 Identification of Type-II and Type-III Sulfur-Containing Derivatives of Steroidal Saponins SS30, with the [M − H] − ion at m/z 799.3563 (C39H59O15S), was 79.95 Da greater than S106 (Figure

Figure 14 .
Figure 14.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS30 in negative ion mode.

Figure 15 .
Figure 15.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS40 in negative ion mode.Figure 15.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS40 in negative ion mode.

Figure 15 .
Figure 15.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS40 in negative ion mode.Figure 15.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS40 in negative ion mode.
Type-VI Sulfur-Containing Derivatives of Steroidal Saponins SS61, with the [M − H] − ion at m/z 997.4303 (C 45 H 73 O 22 S), was 79.95 Da (SO 3 ) greater than Ohiopojaponin B and showed two isotope peaks at m/z 999.4235 and m/z 999.4373 (Figure 16A).In the ESI-MS 2 spectrum (Figure 16B), the pDPIs at m/z 851 ([M − H-146] − )/m/z 671 ([M − H-146-180] − ) were, respectively, attributed to the neutral losses of Rha and Glc.The core structure was inferred as furostanol or pseudo-furostanol.The pDPI at m/z 851/671 was 79.95 Da more than m/z 771/591 of S106, confirming that SS61 and Ohiopojaponin B had the exact fragmentation mechanism.Ultimately, SS61 was identified as the sulfated product of Ohiopojaponin B. Molecules 2024, 29, x FOR PEER REVIEW 28 of 34 Type-VI Sulfur-Containing Derivatives of Steroidal Saponins SS61, with the [M − H] − ion at m/z 997.4303 (C45H73O22S), was 79.95 Da (SO3) greater than Ohiopojaponin B and showed two isotope peaks at m/z 999.4235 and m/z 999.4373 (Figure 16A).In the ESI-MS 2 spectrum (Figure 16B), the pDPIs at m/z 851 ([M − H-146] − )/m/z 671 ([M − H-146-180] − ) were, respectively, a ributed to the neutral losses of Rha and Glc.The core structure was inferred as furostanol or pseudo-furostanol.The pDPI at m/z 851/671 was 79.95 Da more than m/z 771/591 of S106, confirming that SS61 and Ohiopojaponin B had the exact fragmentation mechanism.Ultimately, SS61 was identified as the sulfated product of Ohiopojaponin B.

Figure 16 .
Figure 16.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS61 in negative ion mode.

Figure 16 .
Figure 16.The ESI-MS spectrum (A) and ESI-MS/MS spectrum (B) of SS61 in negative ion mode.For compound SS72-SS74, the ESI -fragment ion with m/z 1127.4726 was designated as the quasi-molecular ion [M − H] − , which indicated that the possible elemental composition was C 51 H 83 O 25 S.In the 100,000 FWHM @ 400 m/z ultra-high resolution mode, two isotope peaks at m/z 1129.4956 and m/z 1129.5038split from the ion [M − H+2] − of element "S", and DPIs at m/z 997 ([M − H − Rha] − ), m/z 981 ([M − H-Glc] − ), m/z 917 ([M − H-Glc-SO 2 ] − ), and m/z 801 ([M − H-Glc-180] − ) were observed (FigureS9), meaning that their fragmentation mechanism was consistent with that of Trigoneoside IVa.Therefore, SS72-SS74 was eventually identified as Trigoneoside IVa sulfite.A total of 55 Type-VI steroidal saponins' sulfur derivatives were identified from the targeted screening of OR based on the fragmentation mechanism of Type-VI steroidal saponins, in which 10 compounds were sulfite products and 45 were sulfate products, as shown in Table4.

Table 3 .
Identification results of prototype components of Steroid Saponins in SF-OR.

Table 3 .
Identification results of prototype components of Steroid Saponins in SF-OR.

Table 3 .
Identification results of prototype components of Steroid Saponins in SF-OR.

Table 3 .
Identification results of prototype components of Steroid Saponins in SF-OR.

Table 3 .
Identification results of prototype components of Steroid Saponins in SF-OR.

Table 4 .
Identification results of sulfur-containing derivatives of Steroid Saponins in SF-OR.

Table 4 .
Identification results of sulfur-containing derivatives of Steroid Saponins in SF-OR.

Table 4 .
Identification results of sulfur-containing derivatives of Steroid Saponins in SF-OR.
Unable to determine the name of the compound.Identification of Type-II and Type-III Sulfur-Containing Derivatives of Steroidal Saponins SS30, with the [M − H] − ion at m/z 799.3563 (C 39 H 59 O 15 S), was 79.95 Da greater than S106 (Figure NO.Mass (m/z) Formula [M − H] −MS/MS Fragment IonsNote: "_": DPI; "?": Unable to determine the binding position; "-":