Methyl Ether-Derivatized Sterols and Coprostanol Produced via Thermochemolysis Using Tetramethylammonium Hydroxide (TMAH)

Sterols are widely distributed in nature from lipids in organisms to sediments. As a conventional method, extraction and derivatization with TMS have been applied for sterol analysis, requiring a long preparation time for gas chromatography–mass spectrometry analysis. In this study, for sterol analysis, thermochemolysis using tetramethylammonium hydroxide (TMAH) was applied. This method performs hydrolysis and methylation simultaneously; thus, free and ether-bonding sterols can be analyzed as sterol methyl ethers in a relatively short time period. A sediment sample from a tideland (the Yatsu tideland, Japan) was analyzed using the TMAH method, and we detected more than 10 sterols, which include cholest-5-en-3β-ol (cholesterol), 24-ethylcholest-5-en-3β-ol (sitosterol), 24-methylcholesta-5,22E-3β-ol (brassicasterol), 24-ethylcholesta-5,24(28)Z-dien-3β-ol (isofucosterol), 4α,23,24-trimethyl-5α(H)-cholest-22E-en-3β- ol (dinosterol), and 5β(H)-cholestan-3β-ol (coprostanol). The detection of the various sterols can be attributed to multiple natural and artificial sources around the Yatsu tideland. In this paper, the mass spectra of these sterols are provided together with an interpretation of their fragmentation patterns. Additionally, the fecal pollution in the Yatsu tideland is discussed in the context of the detection of coprostanol.


Introduction
Sterols exist in the membranes of eukaryotic organisms, including microorganisms and terrestrial plants [1,2]. Thus, sterols have been detected in various natural environments, such as in lake and ocean waters and in sediments [3][4][5][6][7]. These natural sterols can be used to estimate ecological sources because sterol structures differ depending on the types of microorganisms [1,8]. Moreover, they have been used as environmental tracers, such as in degradation processes and for the reconstruction of paleoenvironments [9][10][11].
Sterols can be extracted with organic solvents, such as chloroform and methanol. A hydrolysis step is required if necessary. For sterol analysis by gas chromatography (GC)-mass spectrometry (MS), the extracted sterols must be derivatized to obtain enough volatility to pass through the GC. By derivatization, in the case of sterols, hydroxyl functional groups will be changed to more low-polarity functional groups. Typically, derivatization techniques are categorized into three types, namely, silylation, acylation, and alkylation. Although acylation is also used, silylation is most commonly used for derivatization.
Silylation is effective for various functional groups, including carboxylic acids, amides, and alcohols, and it is performed using silylation reagents, such as N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) and N,O-bis(trimethylsilyl)acetamide (BSA). The mass spectra of sterols determined via these conventional derivatizations have been reported in earlier studies [12][13][14][15][16]; thus, the interpretation of the mass spectra of sterols has become common, which allows us to easily identify sterols. Sterol trimethylsilyl (TMS) ether spectra can be divided into two patterns: ions that are accompanied by the loss of the side chain (SC) and ions that are not accompanied by the loss of the side chain ( Figure 1). The loss of the side chain has five main variations: SC, SC1, SC2, SC + C 2 H 3 , and SC + C 3 H 5 . SC, SC1, and SC occur via the cleavage of C 17 -C 20 , C 20 -C 22 , and C 22 -C 23 bonds, respectively. SC + C 2 H 3 represents the cleavage of the side chain followed by the cleavage of C 13 -C 17 and C 15 -C 16 bonds (C 2 H 3 ) on the D-ring. SC + C 3 H 5 is from the cleavage of SC and C 13 -C 17 and C 14 -C 15 bonds (C 3 H 5 ). However, the ions that are not accompanied by the loss of the SC only have two main variations: [HO-Si(CH 3 ) 3 ] + and [(CH 3 ) 3 Si-O-C 3 H 4 ] + . Sterols can be identified by these fragmentations because these fragmentation patterns are different for each sterol. Meanwhile, silylation is very sensitive to water, causing degradation of the derivatized functional groups. Additionally, because Si in TMS reagents can remain in the inlet of the GC, there is a possibility that the inlet of the GC will be contaminated. Moreover, sample preparation before derivatization (e.g., the extraction step) requires a relatively long time. Since it is a rapid method, thermochemolysis with tetramethylammonium hydroxide (TMAH) was applied for sterol analysis in this study. This method involves hydrolysis and methylation simultaneously, which allow the analysis of both ester and ether compounds and nonbonding compounds that have hydroxyl and carboxyl functional groups [17]. Thus, the TMAH method can provide various compounds, including lignin phenols, amino acids, sugars, and sterols, in one analysis in a relatively short time period [18][19][20][21][22][23]. Additionally, a recent study on the TMAH method extended the analysis to nucleobases [24]. Typically, thermochemolysis with TMAH includes using a pyrolyzer [25][26][27]. Several milligrams of the sample (the optimal weight is varied with the organic carbon content in the sample) was wrapped in a pyrofoil, and it was heated in a pyrolyzer that was directly connected to a GC-MS. Alternatively, thermochemolysis with TMAH can be carried out in a glass ampule [28]. Using this approach, thermochemolysis is performed on an ampule that contains a sample and the TMAH reagent that is sealed under vacuum in an oven.
Upon the TMAH reaction, sterols are provided as sterol methyl ethers. The application of the TMAH method for sterols can produce different types of sterols, including cholesterol, sitosterol, brassicasterol, and stanols [5,6,29]. The efficiency of methylation in sterols reaches >90% [29], indicating that sterols are able to be stably analyzed. Using the TMHA method, both free and bonding (e.g., cholesteryl ester) sterols can be analyzed ( Figure 2). However, there are few reports on the interpretation of mass spectra of sterol methyl ethers, which could help detect them. Therefore, in this study, we provide the mass spectra of sterol ethers obtained using the TMAH method from a sediment sample and also provide an interpretation of these mass spectral patterns to show how to detect them.

Thermal dissociation to alkyl derivatives
Generation processes for sterol methyl ether from cholesteryl stearate via the tetramethylammonium hydroxide (TMAH) reaction. This figure was drawn based on Asperger et al. [29] and Challinor [17].

A Tideland Surface Sediment Sample
In this study, the Yatsu tideland, Japan, was selected as the research location ( Figure 3). The Yatsu tideland is located in the innermost location of Tokyo Bay, and is surrounded by industrial and residential areas that have expanded through reclamation works. In 1993, for the conservation of ecosystems and landscapes in the natural environment, the Yatsu tideland was registered in the Ramsar Convention. Various organisms, such as lugworms, gobies, clams, sea lettuces, and migratory birds inhabit the Yatsu Tideland [30][31][32][33], although the area is small (40 hectares). Therefore, various organic compound inputs are expected. A sediment sample was taken from the edge of the Yatsu tideland using an Ekman-Birge grab ( Figure 3). The sediment sample was freeze-dried immediately after sampling and then powdered well. The powdered sample was stored in a freezer at −20 • C until the sterols were analyzed.

Thermochemolysis with TMAH
Several milligrams of the sample were added into a 20-mL glass ampule with 150 µL (25% in methanol) of TMAH reagent (Sigma-Aldrich Co., St. Louis, MO, USA) and with nonadecanoic-d37 acid (99.1%, CDN isotopes Co., Pointe-Claire, QC, Canada) as an internal standard. The glass ampule was placed into a desiccator to dry the solvents for approximately 15 min. To remove the solvents more sufficiently, a 40 • C hot plate with a nitrogen stream was applied for approximately 10 min. After the solvents were dried, the glass ampule was sealed under vacuum conditions and was heated in an oven at 300 • C for 30 min. After the ampule was cooled to room temperature, the reacted products were extracted with ethyl acetate thrice. The solvent was removed once, and the extracted products were re-resolved with ethyl acetate to an accurate volume of 100 µL. Several microliters within the re-resolved products were injected into the GC-MS system.

GC-MS Conditions
An Agilent GC-MS (GC, 6890N; MS, 5973 inert) was used for this study. A fused silica capillary GC column for the DB-5 ms (inner diameter, 0.25 mm; column length, 30 m; film thickness, 0.25 µm) was installed on the GC system with helium gas (purity: >99.9 vol.%) flowing as the carrier gas. The spitless injection mode was employed. The injection room temperature was set to 310 • C. The setting for the oven was as follows: 60 • C (20 min), 310 • C (6 • C/min), 310 • C (20 min). The mass ion source and quadrupole were programmed at 230 • C and 150 • C, respectively. MS data were recorded using the full scanning mode from 50 to 650 Da (Dalton). Ionization was achieved using the electron impact mode at 70 eV.

Identification of Sterols
The mass spectrum of the sterols obtained using the TMAH method was interpreted by comparing it with reported mass spectra from previous studies on methyl ether derivatization [34] and TMS derivatization [12][13][14][15]. The National Institute of Standards and Technology (NIST)/ United States Environmental Protection Agency (EPA)/ National Institutes of Health (NIH) Mass Spectral Library (NIST 05) was also employed for the interpretation of the sterols. The epicoprostanol-to-coprostanol and coprostanol-to-cholesterol ratios were calculated from their concentrations, which were obtained by comparison with the internal standard.

Identified Sterols in the Samples
From the Yatsu sediment, seven types of sterols from three ∆ 5 -sterols, four 5α(H)-stanols, two 5β(H)-stanols, one ∆ 5,22 -sterol, two ∆ 22 -sterols, two ∆ 5,24(28) -sterols, and one 4α-Me-∆ 22 -sterol were determined (Table 1 and Figure 4). These multiple differences indicate that there are various input sources present for the Yatsu tideland, which can result from the rich ecosystem and human-living area being adjacent to each other. * The symbols are given as n(m)∆ p , where n is the carbon number of a sterol, m is the position of the methyl or ethyl group in the side chain, and p is the position of a double bond (s). ** m/z: bold and italic denote the base peak and molecular ion, respectively.   Table 1.

∆ 5 -Sterols
The compounds of cholest-5-en-3β-ol (cholesterol), 24-methyl-cholest-5-en-3β-ol (campesterol), and 24-ethylcholest-5-en-3β-ol (sitosterol) were detected from the tidal sediment, and they were grouped as ∆ 5 -sterols (Figures 4 and 5). C 27 and C 28 sterols, including cholesterol and campesterol, are widely present in plankton [1,8]. However, C 29 sterols, such as sitosterol, are present in terrestrial plants [35,36]. Therefore, the distributions of C 27 , C 28 , and C 29 sterols have been used for the estimation of sources in sedimentary sterols [8]. However, the results must be interpreted carefully, because C 29 sterols are also found in phytoplankton [1].   A peak at m/z 71, a characteristic ion for ∆ 5 -sterol, was observed for all ∆ 5 -sterols, representing [CH 3 -O-C 3 H 4 ] + . Peaks at m/z 255, 229, and 213 were due to an ion accompanying the loss of the SC ( Figure 5). These peaks (m/z 71, 255, 229, and 213) did not change with the type of derivatization because these peaks appear together with the loss of a derivatized functional group at the C 3 position. Therefore, these peaks have also been found in ∆ 5 -sterol TMS derivatives [12][13][14]. The peak at m/z 255 is derived from the losses of the side chain and HO-CH 3 (Figures 4 and 6). Typically, 5α(H)-stanols are produced under anoxic conditions via bacterial conversion from sterols [7,37,38]. Thus, the ratio of 5α(H)-stanol/∆ 5 -sterol can be applied as a tracer for redox conditions in the field of organic geochemistry [5,6,39]. However, the 5α(H)-stanol/∆ 5 -sterol ratio could be large due to the preferential degradation of sterols in terrestrial organic matters, even under oxic conditions [35].

5α(H)-Stanols
The Although campestanol was also confirmed, the spectrum is not shown in Figure 6, since it was ambiguous.

5β(H)-Stanols (Coprostanol and Epicoprostanol)
From the sediment sample, 5β(H)-cholestan-3β-ol (coprostanol) and 5β(H)-cholestan-3α-ol (epicoprostanol) were determined (Figures 4 and 6). Coprostanol can be produced through the reduction of cholesterol by bacteria within the gut of higher animals including humans [40,41], which is ubiquitously present in their fecal matter [42,43]. Therefore, coprostanol has been used as an indicator of fecal pollution and has been widely applied to environments [30,[44][45][46][47][48][49]. The mass fragments that appeared for coprostanol and epicoprostanol were similar to those of the 5α(H)-stanols, implying that the same fragmentation as in 5α(H)-stanols occurs in 5β(H)-stanols ( Figure 6). However, these intensities are considerably different from those of 5α(H)-stanols. As with 5α(H)-stanols, the peaks at m/z 257, 248, 230, and 215 were confirmed from coprostanol and epicoprostanol as an ion accompanying the cleavage of the SC (Figure 6). An intense peak at m/z 248 represents C 3 H 5 in the D-ring according to the molecular weight followed by the loss of the side chain ([M -(SC + C 3 H 5 )] + ). The peak at m/z 257 is derived from the loss of HO-CH 3 15)] + ) were confirmed to be due to ions that are not accompanied by the loss of the SC (Figure 6).
The detection of coprostanol from sediment suggests that the Yatsu tideland is affected by some fecal pollution. This might be because industrial and residential areas were expanded by reclamation works, and they have enhanced the environmental load on the Yatsu tideland. A biplot of the coprostanol/cholesterol ratio with the epicoprostanol/coprostanol ratio in the sediment is shown in Figure 7. Epicoprostanol is typically converted from coprostanol via a bacterial reaction during sewage treatment processes, and therefore is observed from treated or old sewage samples [53]. Thus, a high epicoprostanol/coprostanol ratio is observed from "old or treated" sewage samples [54]. In our Yatsu tideland sediment sample, the epicoprostanol/coprostanol ratio shows approximately 0.4, indicating that the source is a mixture of new and old. The ratio of coprostanol/cholesterol can estimate the degree of fecal contamination [55][56][57]. According to a proposal by Grimalt and Albaigés [56], when the coprostanol/cholesterol ratio is greater than 0.2 in an environmental sample, the environment can be considered to have fecal contamination. The coprostanol/cholesterol ratio in our sediment sample was approximately 0.5 (Figure 7), which is over the value of the definition. Therefore, the Yatsu tideland is an environment that is affected by fecal pollution according to this definition. However, the value is not higher as with large contaminated areas (Figure 7). Additionally, our data is limited to one site. Thus, a more accurate investigation of the polluted area of the Yatsu tideland requires that different sample areas be analyzed.  Of note, coprostanol analysis via the TMAH method has methodological advantages because the process can be performed quickly with a small amount of solvent. These advantages can result in the treatment of a large number of samples and the reduction of solvents during analysis.
The peaks at m/z 316, 301, 287, 257, 229, 215, and 201 are due to ions that accompany the cleavage of the SC (Figure 9). The peak at m/z 316 results from the loss of the side chain (SC1) with a hydrogen atom from . This ion, m/z 257, is a characteristic ion that is used for the determination of ∆ 22 -sterols, as in the case of TMS derivatization [51,63,64], and it shows that the sterol skeleton is saturated (no double bond on the C 5 -C 6 position) as with 5α(H)-stanols [51]. The peaks at m/z 229, 215, and 201 can arise from the loss of
The spectra of both ∆ 5,24(28) -sterols show peaks at m/z 328, 313, 296, 285, 281, 255, 243, 229, and 213 ( Figure 10). Of these ions, the peaks at m/z 328, 313, 285, and 243 were also confirmed in the spectra of the ∆ 24(28) -sterols. The ion at m/z 328 arises from the loss of the side chain (SC2) with a hydrogen atom and is also a characteristic ion for ∆ 5,24(28) -sterols. The peaks at m/z 255, 253, 229, and 213 are also visible in the fragmentation pattern of other sterols, such as ∆ 24(28) -sterols and ∆ 5 -sterols. The peak at m/z 296 results from the losses of the side chain (SC1) and HO-CH 3 from
The retention time in the GC-MS results is also important for determining sterols. For example, coprostanol and cholestanol, isomers of the H atom at the C 5 position (i.e., 5β(H) or 5α(H)), elute at substantially different times even though the structures are similar (Figure 4). The retention time of coprostanol is earlier than that of cholestanol. Additionally, 5α(H)-stanols and ∆ 22 -stanols were obtained immediately after their corresponding ∆ 5 -sterols and ∆ 5,22 -stanols, which is useful for the determination of these compounds ( Figure 4). Additionally, the number of double bonds and their positions (C 22 and/or C 5 positions) appear at different retention times; for example, retention times become longer in the order of 28(24)∆ 5,22E (brassicasterol), 28(24)∆ 22E (brassicastanol), and 28(24)∆ 5 (campesterol) (Figure 4).

Conclusions
A tideland surface sediment from the Yatsu tideland, Japan, was analyzed using thermochemolysis with tetramethylammonium hydroxide (TMAH) for sterols. From the sediment, 10 sterols, including ∆ 5 -sterols, 5α(H)-sterols, ∆ 5,22 -sterols, ∆ 22 -sterols, ∆ 5,24(28) -sterols, 4α-Me-∆ 22 -sterols, and coprostanol, were determined. The Yatsu tideland sediment sample enabled the determination of these various sterols, because the Yatsu tideland is located where rich ecosystems and human-living areas are adjacent to each other. The MS results of these sterols helped determine them. Characteristic mass fragments at m/z 255, m/z 215, m/z 257, and m/z 330 were useful to detect ∆ 5 -type sterols (i.e., ∆ 5 -sterols, ∆ 5,22 -sterols, and ∆ 5,24(28) -sterols), 5α(H)-sterols, 5β(H)-sterols, ∆ 22 -sterols, and 4α-Me-∆ 22 -sterols from the extracted mass chromatograms. Additionally, some fecal pollution in the Yatsu tideland was suggested by the detection of coprostanol. The TMAH method can treat samples in a relatively short time with a small amount of sample and then provide sterol methyl ethers. Thus, this method is suitable for studies that analyze many samples. The provided sterol spectra in this study, including other such studies, will help determine the kinds of sterols present in a sample using the GC-MS result.