Analysis of Bacteriohopanoids from Thermophilic Bacteria by Liquid Chromatography–Mass Spectrometry

Background: Hopanoids modify plasma membrane properties in bacteria and are often compared to sterols that modulate membrane fluidity in eukaryotes. In some microorganisms, they can also allow adaptations to extreme environments. Methods: Hopanoids were identified by liquid chromatography–mass spectrometry in fourteen strains of thermophilic bacteria belonging to five genera, i.e., Alicyclobacillus, Brevibacillus, Geobacillus, Meiothermus, and Thermus. The bacteria were cultivated at temperatures from 42 to 70 °C. Results: Regardless of the source of origin, the strains have the same tendency to adapt the hopanoid content depending on the cultivation temperature. In the case of aminopentol, its content increases; aminotetrol does not show a significant change; and in the case of aminotriol the content decreases by almost a third. The content of bacteriohopanetetrol and bacteriohopanetetrol glycoside decreases with increasing temperature, while in the case of adenosylhopane the opposite trend was found. Conclusions: Changes in hopanoid content can be explained by increased biosynthesis, where adenosylhopane is the first intermediate in the biosynthesis of the hopanoid side chain.


Introduction
The geographic unit known as Eger Graben, after the river Eger, is located in the NW part of the Bohemian Massif (Czech Republic). It is elongated in the SW-NE direction along the Ore Mts (Erzgebirge) mountain chain and forms a part of the European Cenozoic rift system. The origin of this large, asymmetric, tectonic graben was related to the Alpine orogeny, which was manifested by tectonic and volcanic activity in the orogen's foreland [1]. The volcanism began in the Late Cretaceous and reached its peak in the Tertiary period. Owing to this volcanic activity, Karlovy Vary (Carlsbad) has become a world-famous spa. There are several mineral springs in the town of Carlsbad, the main being "Vřídlo"-the official name for a spring with a yield of 2000 L per minute and a temperature of 73 • C. The presence of thermophilic bacteria, such as bacteria of the genera Thermus [2] or Brevibacillus and Geobacillus [3], was discovered.
Several different definitions have been used for thermophilic and thermotolerant microorganisms. The most commonly used one depends on the optimal growth temperature, i.e., an interval of 45 to 80 • C for thermophiles [4].
Prokaryotic microorganisms, such as bacteria, but also cyanobacteria, with a few exceptions, do not contain sterols commonly found in plants and animals. However, many of them contain bacteriohopanepolyols, which are complex molecules based on the hopane skeleton (pentacyclic triterpenoid). The analysis of bacteriohopanoids directly from were taken from thermal springs using 50 mL sterile Falcon tubes in three replicates for each sample from the same spring of the same location, and were transported to the laboratory at a temperature of at least 50 • C, ensured by a thermal box. The collected water samples were immediately used for enrichment cultivations as follows: volume of 15 mL of spring water was supplemented with 1.5 mL of diluted YPD (dextrose 20 g/L, peptone 20 g/L, and yeast extract 10 g/L, pH 7.0) and cultivated on a rotary shaker (150 rpm). Cultivations were performed at the temperature corresponding to the original spring source (i.e., 58 • C in Štěpánka, Mlýnský, and Vřídlo, and 42 • C in Sadový). After 30 days of cultivation, aliquots of the enrichment cultures were streaked on an agar medium to obtain separate colonies. Reasoner's 2A agar (R2A, HiMedia, Brno, Czech Republic) and Thermus 162 agar (HiMedia, Brno, Czech Republic) supplemented by micro-filtered origin spring water were used. Colonies with different morphologies were further cultivated and identified by 16S rRNA as described before [3]. Four bacterial isolates, designated Geobacillus stearothermophilus ST-YPD, Brevibacillus agri SA-1, Geobacillus kaustophilus ML-1, and Geobacillus stearothermophilus VR-1, were obtained from the springs. For pre-inoculum preparation, the strains were cultivated in Luria-Bertani broth (LB: tryptone 10 g/L, yeast extract 5 g/L, and NaCl 10 g/L, pH 7.0) on an orbital shaker (150 RPM) for 72 h. Cultivation of biomass for further analysis was carried out in LB broth. Two hundred milliliters of sterile LB broth was inoculated with pre-inoculum to achieve OD600 0.2 and incubated on an orbital shaker (150 RPM) for 48 h until the stationary phase was reached. The isolate Geobacillus stearothermophilus ST-YPD was cultivated at 42 • C and the remaining three isolates at 58 • C (in conformity with the temperature of the springs of origin). At the end of cultivation, cells were harvested by centrifugation (10,000× g, 10 min, 4 • C) and washed twice with sterile 0.9% NaCl. Cell mass was frozen at −70 • C and lyophilized.

Extraction and Isolation of Hopanoids
The extraction procedure was based on the method of Bligh and Dyer [22]. Briefly, the lyophilized cells (approximately 10 mg) were suspended in a chloroform-methanol mixture (2:1) (~1 mL) for 30 min at 20 • C with stirring, after which chloroform and water were added and the insoluble material was separated by centrifugation. The aqueous phase was aspirated off, and the chloroform phase was evaporated to dryness under reduced pressure. Lipid extracts were acetylated with acetic anhydride in pyridine (0.5 mL, 1:1 v/v) for one hour at 50 • C and left at room temperature overnight. The solvent was removed under a vacuum. The acetylated extract was dissolved in acetonitrile (ACN):propan-2-ol (iPrOH) (1:1 v/v) and used for LC-MS analysis.
The lyophilized cells of the previously cultured strains listed in Table S2 were extracted and further analyzed in a similar manner as the four strains mentioned in Table S2.

Analysis of Hopanoids by LC-MS
The LC equipment consisted of a 1090 Win system, PV5 ternary pump, an automatic injector (HP 1090 series, Agilent, Santa Clara, CA, USA), and three Luna Omega 1.6 µm, reverse-phase (RP) C18, 100 Å, LC Columns L × I.D. 150 × 2.1 mm, connected in series. Although the efficiency stated by the manufacturer was not achieved, i.e., 350,000 theoretical plates/meter (see manufacturer's literature), the efficiency achieved in our case was onlỹ 115,000 theoretical plates/meter. Pregnane-3,20-diol 3,20-diacetate with a retention time (tR) 23.9 min was used as an external standard, with a flow rate of 0.95 mL/min, an injection volume of 10 µL, and a column temperature of 25 • C. Peracetylated hopanoids were separated using a solvent gradient program with ACN and iPrOH: initial ACN/iPrOH (85:15, vol/vol), a linear change to 15:85 ACN/iPrOH (vol/vol) from 0 min to 60 min, and a hold for 10 min. The composition was returned to the initial conditions over 10 min. For ESI-MS analysis, an LTQ Orbitrap Velos was used.
An LTQ Orbitrap Velos mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA), a high-resolution hybrid mass spectrometer equipped with a heated electrospray interface (H-ESI), was operated in the positive ionization mode. The MS scan range was performed in the Fourier transform mode and recorded within a window between 150-1500 m/z. Mass spectra were acquired with a target mass resolution of R = 110,000 at m/z 800, and the ion spray voltage was set at 3.5 kV in the positive ionization mode. For positive ionization mode the following parameters were used: sheath gas flow, 18 arbitrary units (AU); auxiliary gas flow, 7 AU; ion source temperature, 250 • C; capillary temperature, 230 • C; capillary voltage, 50 V; and tube lens voltage, 170 V. Helium was used as a collision gas for collision-induced dissociation (CID) experiments. The CID normalization energy of 35% was used for the fragmentation of parental ions. The MS/MS product ions were detected in high-resolution FT mode. The calibration of the MS spectrometer was conducted using a Pierce LTQ Orbitrap Positive Ion Calibration Solution (Thermo Fisher Scientific, San Jose, CA, USA). The mass accuracy was better than 0.9 ppm. The chemical structures of compounds were confirmed with the help of the spectral database LIPID MAPS ® Lipidomics Gateway (http://www.lipidmaps.org/, (accessed on 14 September 2021)).

Quantification
Pregnane-3,20-diol-3,20-diacetate was obtained from commercially available pregnanediol and was acetylated at room temperature overnight with pyridine/acetic anhydride (1:1; 2 mL). A semi-quantitative estimate of hopanoid abundance was calculated from the characteristic base peak ([M + H-CH 3 COOH] + ) of individual hopanoids relative to the m/z 345.2788 (base peak) of the acetylated 5α-pregnane-3β,20β-diol internal standard. The limit of detection (LOD) was~40 pg (0.1 pM) at S/N = 3 and the limit of quantification (LOQ) at S/N = 10 was about 132 pg. The content of additional bacteriohopanepolyols was then calculated from the relative amount in the HPLC analysis. Without standards for each hopanoid, a significant potential error in quantification requires caution when comparing these values with those in the literature.

Statistical Analysis
The statistical analysis was performed using the IBM SPSS Statistics (Statistical Package for the Social Sciences; IBM Corp, 2013) software, version 26 (IBM ® Corporation, Armonk, NY, USA).

Characterization and Identification of Bacterial Isolates
Molecular classification of the four strains collected from hot springs of Carlsbad was performed by 16S rRNA gene sequence analysis, as shown in Table S3. The results of the molecular analysis were consistent with the phenotypic characteristics [3]; therefore, the strains were designated Brevibacillus agri SA-1, Geobacillus kaustophilus ML-1, Geobacillus stearothermophilus ST-YPD, and Geobacillus stearothermophilus VR-1.

Analysis of Total Hopanoids by Liquid Chromatography-Mass Spectrometry Using Precursor Ion and Neutral Ion Scans
According to Bligh and Dyer, total lipids obtained from fourteen bacteria cultured at different temperatures [22] were acetylated and analyzed by RP-UPLC/high-resolution positive tandem ESI. Hopanoids were derivatized (acetylated) mainly for reasons of reducing the polarity of individual compounds on the basis of already-published papers [13][14][15][16][17][18]. In essence, there is only one paper [19] which describes the analysis of non-derivatized hopanoids. Conversely, many more papers (see publications from Talbot lab) describe the analysis of derivatized hopanoids. Bacteriohopanoid standards are not commercially available, and they have therefore not been tested. Based on our previous experience [23], we used a completely different elution mixture than that described in various papers [15,16,19]. This was mainly due to the fact that the elution mixture with acetonitrile has a lower viscosity than the commonly used methanol-water-isopropanol mixture. Lower viscosity of the elution mixture used by us made it possible to use a combination of three columns in a series and thus achieve an efficiency of~115,000 theoretical plates. This value is at least an order of magnitude higher than the column efficiency, for example, for N-palmitoleyl of peracetylated bacteriohopanetriol ( Figure 4 in paper [15]). Due to this resolution, it was not a problem to separate individual molecular species and isomers, see for example the compounds with tR 35.06 vs. 35 (Table S4). The presence of stereoisomers of hopanoids is quite common in bacteria [18,21,24]. The high-resolution NLS and PIS of the peracetylated total lipid extract made it possible to identify twenty-one molecular species (see Figure 1). Furthermore, the use of tandem MS confirmed the structure of N-acyl-substituted bacteriohopanepolyols, as shown in Table S4. mercially available, and they have therefore not been tested. Based on our previous experience [23], we used a completely different elution mixture than that described in various papers [15,16,19]. This was mainly due to the fact that the elution mixture with acetonitrile has a lower viscosity than the commonly used methanol-water-isopropanol mixture. Lower viscosity of the elution mixture used by us made it possible to use a combination of three columns in a series and thus achieve an efficiency of ~115,000 theoretical plates. This value is at least an order of magnitude higher than the column efficiency, for example, for N-palmitoleyl of peracetylated bacteriohopanetriol ( Figure 4 in paper [15]). Due to this resolution, it was not a problem to separate individual molecular species and isomers, see for example the compounds with tR 35.06 vs. 35 (Table S4). The presence of stereoisomers of hopanoids is quite common in bacteria [18,21,24]. The high-resolution NLS and PIS of the peracetylated total lipid extract made it possible to identify twenty-one molecular species (see Figure 1). Furthermore, the use of tandem MS confirmed the structure of N-acyl-substituted bacteriohopanepolyols, as shown in Table S4.  Figure 1 shows the PIS at m/z 205.1951 and NLS at m/z 206.2035. If methyl-substituted hopanoids were to be present, then both PIS and NLS would have to show peaks. As shown in Figure 1, no peaks are present; hence, their concentration is at least three orders of magnitude lower than the concentration of non-methyl hopanoids. The presence of   Figure 1, no peaks are present; hence, their concentration is at least three orders of magnitude lower than the concentration of non-methyl hopanoids. The presence of genes encoding the production of enzymes, i.e., the two methylases required to produce 2-or 3-methylhopanoids, has been shown in the literature [25]. The presence of both hopanoid C-2 methylase (EC: 2.1.1.-) and/or hopanoid C-3 methylase (EC: 2.1.1.-) in the studied bacterial species was analyzed using the BLASP tool from the NCBI. All species, except for Meiothermus ruber and Thermus aquaticus, possess a protein sequence of high similarity to the enzymes mentioned above; see Table S5 for more information. Nevertheless, the presence of a gene does not require its expression or metabolite biosynthesis.

The Absence of Methylhopanoids in the Studied Bacteria was Proven by Liquid Chromatography-Mass Spectrometry
The tandem mass spectra of both methyl derivatives (2β-methyl and 3β-methyl hopanoids, respectively) are characterized by a fragment ion m/z 205.1951 instead of 191.1794, and a series of fragments resulting from a characteristic neutral loss of 206.2035 Da instead of 192.1878 Da (as a result of C-9 cleavage in C and 11 and C-8 and 14) [17]. These m/z values are~14 units higher than those of non-methyl bacteriohopanepolyols.

Use of Two Tandem Mass Spectrometry Methods (PIS and/or NLS) for the Identification of Hopanoid Structures
The major ions in the tandem MS spectra of compounds (1b-e, 2e, and 3e) come from the partial or complete loss of up to five acetylated hydroxyl groups ( Figure 2 and Table 1). An ion of the structure, [M + H-CH 3 COOH] + , was identified as the base peak in all tandem MS. The key ion confirming the amide structure was the ion of the type [M + H-CH 3 COOH-RC=O] + ; depending on the number of OH groups, values at 728.5100, 668.4888, and 608.4677 Da were found. It was also shown by tandem MS that ring A is not substituted with a methyl group (see above).  Bacteriohopanetetrol (5)  Two N-acylaminotriols containing palmitic and palmitoleic fatty acids bound to aminotriol via an amide bond have been previously identified in Nitrosomonas europaea [15,34]. A total of five N-acyl derivatives were identified, i.e., two saturated (palmitic and stearic acids) and three unsaturated (hexadecenoic, octadecenoic, and nonadecenoic acids). Furthermore, unsaturated C18 N-acylaminotriol was preliminarily identified in the purple non-sulfur bacterium Rhodomicrobium vannielii [35]. Table 3 shows the content of all identified hopanoids depending on the culture temperature. An example is a strain of G. stearothermophilus, which has been cultured at three different temperatures (55, 58, and 70 • C). Regardless of the source of origin (two different hot springs or evaporated milk), all five strains show the same tendency to change the content of hopanoids. In the case of aminopentol (1a) its content increases, aminotetrol (2a) shows no substantial change, and in the case of aminotriol (3a), the content decreases by almost a third. Similar results were obtained with M. ruber, which was cultured at 55, 65, and 70 • C. Again, the content of aminopentol increased, did not change with aminotetrol and decreased with aminotriol. The remaining two strains, i.e., A. acidoterrestris and T. aquaticus, show essentially the same trend. The situation with N-acyl derivatives is far more complicated. In this case, the N-acyl chain changes the molecule's shape (see Figure S1), and thus, of course, the spatial arrangement of lipids in the membrane. Table 3. Content (relative %) of hopanoids from fourteen strains of thermophilic bacteria identified by reverse-phase liquid chromatography-positive electrospray mass spectrometry (see Materials and Methods). Bacterium  1a  1b  1c  1d  1e  2a  2b  2c  2d  2e  3a  3b  3c  3d  3e  4  Especially in all cultured thermophilic bacteria, the proportion of even and odd Nacyls is the same. In many cases, the molecular species having N-acyls with odd numbers of carbon atoms are more abundant. Furthermore, there is also a general trend that the proportion of N-acyl pentols increases, and the proportion of N-acyl triols decreases with increasing temperature. The three hopanoids, i.e., molecular species with structures 4, 5, and 6, differ from each other depending on the content and temperature. While the content of compounds 5 and 6 decreases with increasing temperature, the opposite is true for adenosylhopane (4). This large abundance can be explained, for example, by increased biosynthesis, where adenosylhopane is the first intermediate in hopanoid side chain biosynthesis [36].

Cultiv. Temp
Hot terrestrial springs provide suitable habitats for several photoautotrophic and chemolithoautotrophic microorganisms. The distribution of these organisms seems to be mostly affected by temperature. For example, temperatures above 73 • C often eliminate the growth of cyanobacteria and other phototrophic bacteria and, conversely, allow the growth of chemolithoautotrophs. The effect of environmental conditions on the production of bacteriohopanepolyols has been described in several papers (see below). These changes occur because hopanoids regulate membrane fluidity and induce the order in the phospholipid membrane [10,37,38], which helps to protect the microorganism from external stresses.
A study was published on symbiotic methanotrophs in Sphagnum moss [39,40], where the proportion of aminotriol decreased with increasing temperature (from 5 to 25 • C), while the abundance of both aminotetrol and aminopentol grew. Similarly, Osborne et al. [41] describe changes in the abundance of bacteriohopanepolyols as a function of temperature. It was shown that with increasing temperature (from 4 • C to 40 • C) aminotriol also decreased, while the presence of aminopentol increased and the aminotetrol content remained constant. Jahnke et al. [42] report that for the psychrotolerant methanotroph (considered Methylomonas methanica), the relative percentage of C31 hopanol decreased. In contrast, C30 hopanol increased with increasing temperature. Similarly, derivatives of hopanols after the oxidation of hopanoids have been analyzed in other methanotrophic bacteria of the genus Methylovulum [43]. Again, it was confirmed that the content of pentols increased, the triol content decreased, and the tetrol content remained more or less constant as the cultivation temperature increased. Unfortunately, all the publications mentioned above analyzed the degradation products formed by the oxidation of vicinal diols (Rohmer degradation) and, of course, quantified. Although this procedure dramatically simplifies the analysis of hopanoids, the information about the structure of native hopanoids is completely lost.
Previous studies show that the proportion of total hopanoids in the thermoacidophilic bacterium Alicyclobacillus acidocaldarius increases with culture temperature. As the temperature changed from 60 to 65 • C, the content of tetrol (1,2,3,4-tetrahydroxypentane-29-hopane) increased several times regardless of the pH of the culture. Studies [9,44] have shown that a change in culture temperature from 14 • C to 37 • C in ethanologenic Zymomonas mobilis was accompanied by an increase in 1,2,3,4-tetrahydroxypentane-29-hopane by several tens of percent. Additionally, Joyeux et al. [45] investigated the effect of culture temperature on hopanoid content in acetic acid bacterium Frateuria aurantia. It was found that the C31/C32 ratio of hopanoid degradation products was more than doubled in the culture performed at 36 • C compared to the culture at 16 • C.
Mono-and di-unsaturated bacteriohopanols were not identified in the thermophilic bacteria analyzed in this paper. One of the possible hypotheses of their absence is that they contribute to the increase in membrane fluidity, which is undesirable in thermophilic bacteria. These compounds were isolated from Gluconacetobacter xylinus, Burkholderia cepacia, methane-oxidizing Methylocaldum szegediense [46], or Rhodopseudomas palustris [47].

Conclusions
Although the enzyme squalene-hopene cyclase is cited more than 26,000 times in a GenBank search, only an estimated 10% of bacteria can produce hopanoids [48,49]. Even if many phyla of bacteria are potentially capable of producing hopanoids, their production was detected by either GC-MS or LC-MS in only a few dozen (see Table S1 of individual strains that were cultured).
Twenty-one intact bacteriohopanepolyols in extracts from fourteen strains of bacteria were identified by RP-UPLC/HR-ESI + -MS/MS. The distribution varied depending on the strain of bacteria and also the cultivation temperature. Regardless of the place of origin, all five strains of Geobacillus stearothermophilus have the same tendency to change the hopanoid content. In the case of aminopentol (1a) its content increases, aminotetrol (2a) does not show a significant change, and in the case of aminotriol (3a) the content decreases by almost a third. Similar results were obtained with Meiothermus ruber, which was cultured at 55, 65, and 70 • C. In the case of N-acyl derivatives, the proportion of N-acylpentols increases and the proportion of N-acyltriols decreases with increasing temperature. The content of bacteriohopanetetrol and bacteriohopanetetrol glycoside decreases with increasing temperature, while in the case of adenosylhopane it is exactly the opposite. This dependence can be explained, for example, by increased biosynthesis, where adenosylhopane is the first intermediate in the biosynthesis of the hopanoid side chain.
This study expands the knowledge of the high diversity of hopanoid-producing bacteria in various hot springs, and suggests that bacteriohopanepolyols can be detected in individual cultures. The analysis of bacteriohopanepolyols by LC-MS, especially by RP-UPLC/HR-ESI + -MS/MS, is a valuable tool for evaluating hopanoids in bacteria living in hot springs.