A high-coverage construction method of a database of gut microbiota-associated metabolites

Abstract

This invention discloses a method for constructing a high-coverage database of gut microbiota-related metabolites. First, reported gut microbiota-related metabolites are collected through literature retrieval, and duplicates are removed using PubChem CID to establish a database based on prior knowledge. Second, this database is expanded: using sex- and age-matched sterile / antibiotic-treated / specific pathogen-free mice, non-targeted metabolomics analysis is performed on their plasma, urine, and feces. Statistical methods are then used to identify novel gut microbiota metabolites not yet covered in the knowledge base. This invention combines knowledge-driven and experimentally measured gut microbiota-related metabolites from biological samples, providing a framework for constructing a high-coverage database of microbiota metabolites and laying the foundation for large-scale annotation of gut microbiota-related metabolites in metabolomics data and the study of the molecular mechanisms of microbiota-host interactions.

Description

Technical Field

[0001] This invention relates to the fields of analytical chemistry and gut microbiota metabolomics, and is a method for constructing a high-coverage database of gut microbiota-related metabolites.

[0002] Research Background

[0003] The gastrointestinal tract of mammals houses a complex and dynamic microbial ecosystem known as the "gut microbiota." It produces bioactive molecules that mediate physiological processes such as host energy uptake, immune cell development, and intestinal epithelial homeostasis. Metabolic disturbances in the gut microbiota are closely related to the development and progression of many diseases. For example, trimethylamine oxide derived from gut microbes promotes thrombosis, and imidazole propionic acid, a histidine metabolite produced by gut microbiota, can directly impair glucose tolerance and insulin signaling. Therefore, high-coverage analysis of gut microbiota-related metabolites is crucial for a comprehensive understanding of the interaction between gut microbes and the host, providing new insights into the molecular mechanisms of diseases.

[0004] Gut microbiota-related metabolites exhibit structural and chemical diversity, primarily encompassing low-molecular-weight fatty amines, nucleosides and nucleotides, vitamins, amino acids and their derivatives, organic acids, phenolic acids, indole derivatives, polyamines, bile acids, and fatty acids. The wide concentration distribution (spanning several orders of magnitude) of these various gut microbiota-related metabolites in real-world biological samples limits their simultaneous, high-coverage analysis.

[0005] Currently, non-targeted metabolomics methods based on high-performance liquid chromatography-high-resolution mass spectrometry (HPLC-HSMS) are widely used for large-scale detection of gut microbiota-related metabolites. Analyzing the metabolic profiles of real biological samples using this method will detect thousands of metabolite signals, but screening for compounds related to the gut microbiota from these signals remains a major challenge. Limited commercial standards, a lack of dedicated databases for microbiota metabolites, and insufficient coverage of microbiota-related metabolites in public databases all make large-scale annotation of gut microbiota-related metabolites extremely difficult. Therefore, there is an urgent need to establish a dedicated database containing as many gut microbial metabolites as possible to play an irreplaceable role in the broad annotation of metabolomics data.

[0006] In view of this, this invention proposes a high-coverage method for constructing a database of gut microbiota-related metabolites. This method combines extensive collection of prior knowledge with experiments using sterile / antibiotic-treated / specific pathogen-free mouse models to mine gut microbiota-related metabolites reported in the literature and experimentally measured in biological samples, thereby constructing a high-coverage microbiota-metabolomics-specific database. To date, there are no reports on applying the method described in this invention to the high-coverage construction of gut microbiota-related metabolite databases in metabolomics. Summary of the Invention

[0007] This invention provides a method for constructing a high-coverage database of gut microbiota-related metabolites. To achieve the objectives of this invention, firstly, reported gut microbiota-related metabolites are collected through extensive literature searches, and duplicates are removed by manual verification using PubChem CID, establishing a gut microbiota-related metabolite database based on prior knowledge. Secondly, novel gut microbiota-related metabolites not previously reported in the literature are identified from plasma, urine, and fecal samples using germ-free mice, antibiotic-treated mice, and specific pathogen-free mouse models, and these are added to the self-built database to expand its coverage. This method combines knowledge-driven and experimentally measured gut microbiota-related metabolites from actual biological samples to establish a comprehensive database, offering advantages such as high coverage and strong specificity, overcoming the problem of insufficient coverage of gut microbiota-related metabolites in existing publicly available metabolomics databases. The technical steps employed in this invention are as follows:

[0008] (1) Literature searches were conducted in ISI Web of Science (https: / / www.webofscience.com / wos / alldb / basic-search) and PubMed (https: / / pubmed.ncbi.nlm.nih.gov / ) using the keywords “gut microbial metabolites”, “gut microbiota / gut flora&metabolomics / metabonomics”, and “gut microbiota-host co-metabolism”. Measured rather than predicted gut microbiota-related metabolites explicitly reported in research papers and reviews, with the host limited to humans and mice, were collected. Their corresponding PubChem CID numbers were retrieved in PubChem (https: / / pubchem.ncbi.nlm.nih.gov / ). Then, the metabolites were deduplicated according to the PubChem CID to establish a gut microbiota-related metabolite library based on prior knowledge.

[0009] (2) Plasma, urine, and fecal samples were collected from sex-matched (same male-to-female ratio in all three groups) and age-matched (same average age in all three groups) germ-free mice, antibiotic-treated mice, and specific-pathogen-free mice (at least 6 mice in each group) using a standardized procedure. Once collected, samples were immediately and rapidly frozen and stored at -80°C for later use. Germ-free mice (GF) are mice raised in an isolation system that show no detectable microorganisms or parasites in their bodies or externally. Antibiotic-treated mice (ABX) are normal mice that have been treated with antibiotics (usually by gavage / drinking water) for a period of time (usually 1-2 weeks). Specific-pathogen-free mice (SPF) are mice without specific microorganisms or parasites in their bodies; generally, they refer to healthy mice without infectious diseases.

[0010] (3) Simple, non-selective pretreatment of plasma, urine, and feces from germ-free mice, antibiotic-treated mice, and specific pathogen-free mice was performed to achieve effective extraction of various intestinal flora-related metabolites. The specific process is as follows: 50-150 μL of plasma was placed in a 1.5 mL centrifuge tube, and 200-600 μL of methanol was added for protein precipitation and metabolite extraction, followed by vortexing for 60 s; 50-150 μL of urine was placed in a 1.5 mL centrifuge tube, and 200-600 μL of methanol was added, followed by vortexing for 60 s; 10 ± 1 mg of lyophilized 48... After h, fecal matter was placed in a 2 mL centrifuge tube, and zirconia grinding beads, 300 μL of ultrapure water, and 300 μL of methanol were added sequentially. The mixture was homogenized at 30 Hz for 1 min using a mixer, and repeated 5 times. Then, it was incubated at room temperature for 5 min. The samples extracted from plasma, urine, and feces were placed in a centrifuge and centrifuged at 18,000 g for 15 min at 4 °C. All supernatants were collected. The fecal supernatant was passed through a nylon filter with a pore size of 0.22 μm to obtain 280 μL of filtrate. The filtrate was then placed in a vacuum freeze dryer and freeze-dried at 4 °C.

[0011] (4) Reconstitute the lyophilized plasma, urine and feces samples with 50-150 μL acetonitrile / water (20 / 80, v / v), 50-150 μL acetonitrile / water (5 / 95, v / v) and 70 μL acetonitrile / water (20 / 80, v / v), respectively; vortex until completely dissolved, place in a centrifuge, centrifuge at 18,000g for 10-15 min at 4℃, and take the supernatant for sample injection analysis.

[0012] (5) A non-targeted metabolomics analysis of mouse samples was performed using a high-performance liquid chromatography (HPLC) method based on a pentafluorophenyl column combined with a high-resolution mass spectrometry (HS-MS) detection method based on a full-scan / iterative data-dependent secondary mass spectrometry acquisition mode. This method effectively separated intestinal microbiota-related metabolites and achieved high-coverage acquisition of secondary mass spectrometry information, obtaining the HPLC retention time, primary mass spectrometry information (mass-to-charge ratio and intensity of primary ions), and secondary mass spectrometry information (mass-to-charge ratio and intensity of secondary ions) of the metabolites. The HPLC-MS operating conditions are as follows:

[0013] Liquid chromatography conditions: Supelco Discovery HS-F5 column, specifications: inner diameter 2.1 mm × column length 150 mm, particle size 3 μm; column temperature 40℃, flow rate 0.25 mL / min, injection volume 5 μL, injection chamber temperature 6℃; in electrospray ionization (ESI) positive ionization mode. + Under the following conditions, mobile phase A is an aqueous solution containing 0.1% formic acid, and mobile phase B is an acetonitrile system containing 0.1% formic acid; in electrospray negative ionization mode (ESI)... - Under these conditions, mobile phase A was an aqueous solution containing 10 mM ammonium formate, and mobile phase B was an acetonitrile / water (95 / 5, v / v) solution containing 10 mM ammonium formate. A gradient elution program was used, initially 0% B phase, held for 2 min; then linearly increased to 60% B over 18 min; then linearly increased to 100% B over 3 min, and the system was flushed for 3 min; finally switched back to 0% B and equilibrated for 4 min.

[0014] Mass spectrometry conditions: electrospray ionization source, combined positive and negative ion modes; spray voltage at ESI + The voltage is 3.5kV in the mode, at ESI. - The voltage was 3.0 kV; the capillary temperature was 320 °C, and the auxiliary gas heater temperature was 350 °C; the sheath gas and auxiliary gas flow rates were 45 and 10, respectively; the S-lens RF level was 50. A full-scan / iterative data-dependent secondary mass spectrometry acquisition mode was used, with 4 iterations; the full-scan mode resolution was 60,000, and the scan range was 58-870; the data-dependent secondary mass spectrometry acquisition mode resolution was 30,000; secondary mass spectrometry acquisition was triggered by the top 10 ions with the strongest response in each full-scan cycle, with a dynamic exclusion duration of 10 s; normalized collision energies of 15, 30, and 45 were used. The automatic gain control target (AGC target) and maximum injection time (maximum IT) were 3 × 10⁻⁶ in the full-scan setting. 6 The ion capacity and 100ms are 1×10⁻⁶ under the secondary mass spectrometry acquisition settings. 5 Ion capacity and 50ms.

[0015] (6) Preprocess the metabolomics data, including peak detection, peak alignment, and 80% rule removal of missing values, to obtain a table of characteristic peaks of detected metabolites (including precursor ions, retention times, and peak areas of metabolites); perform internal standard correction on the peak areas of metabolites in the table to eliminate systematic errors in the preprocessing and instrument analysis process, to obtain a table of stable detected metabolite peaks (including precursor ions, retention times, and corrected peak areas of metabolites); perform fecal quality correction on fecal samples and creatinine level correction on urine samples; finally, obtain the relative abundance matrix of each metabolite in different biological samples.

[0016] (7) Compare the relative abundance differences of metabolites between germ-free / antibiotic-treated mouse groups and specific pathogen-free mouse groups in three types of biological samples. If the following two conditions are met, the metabolite is considered to be related to the gut microbiota:

[0017] Condition 1: Using the non-parametric Wilcoxon and Mann-Whitney U test, the relative abundance of this metabolite showed a significant difference between the germ-free and specific pathogen-free mouse groups, with an asymptotic significance p-value less than 0.05 (the null hypothesis was that the abundance of this metabolite did not differ significantly between the germ-free and specific pathogen-free mouse groups, and a p-value less than 0.05 indicates that the experimental results can reject the null hypothesis at the 5% significance level); Condition 2: Relative to the specific pathogen-free mouse group, the abundance of this metabolite increased or decreased simultaneously in both the germ-free and antibiotic-treated mouse groups, with no limit on the magnitude.

[0018] (8) Using the open-source software MS-DIAL (http: / / prime.psc.riken.jp / compms / msdial / main.html / ) and MetEx (http: / / www.metaboex.cn / MetEx / ), the publicly available metabolomics databases HMDB (The Human Metabolome Database, https: / / hmdb.ca / ) and MyCompoundID were analyzed.

[0019] Automated searches were performed using databases including (https: / / www.mycompoundid.org / ), MoNA (MassBank of North America, https: / / mona.fiehnlab.ucdavis.edu / ), mzCloud (https: / / www.mzcloud.org / ), and GNPS (The Global Natural Product Social Molecular Networking, https: / / gnps.ucsd.edu / ). Structural annotations were performed on the identified gut microbiota-related metabolites, retaining those that simultaneously met the following two conditions (metabolite name, molecular formula, structural formula, and PubChemCID number):

[0020] Condition 1: The difference between the detected value of the exact mass number of the precursor ion and the theoretical value is less than 10 ppm; Condition 2: The matching score between the experimentally obtained secondary mass spectrometer and the secondary mass spectrometer in the open-source database is greater than 0.7.

[0021] (9) For the annotated gut microbiota-related metabolites, search for them in the knowledge base established in step (1) based on their PubChem CID, and add any novel gut microbiota-related metabolites that are not found to the self-built database.

[0022] This invention first collects gut microbiota-related metabolites reported in the literature through literature retrieval, and then removes duplicates of the metabolites using PubChem CID to establish a gut microbiota-related metabolite library based on prior knowledge. Secondly, the database was expanded: at least six sex- and age-matched germ-free (GF), antibiotic-treated (ABX), and specific-pathogen-free (SPF) mice were used in each group. Their plasma, urine, and feces were non-selectively pretreated to effectively extract different types of gut microbiota metabolites. A non-targeted metabolomics analysis of the mouse samples was performed using a high-resolution mass spectrometry (HS-MS) method based on pentafluorophenyl column liquid chromatography followed by a data-dependent secondary mass spectrometry acquisition mode based on full scan / iteration. After preprocessing the raw data including peak detection, peak alignment, missing value removal, internal standard correction, fecal quality correction, and urinary creatinine correction, significant differential metabolites between germ-free / antibiotic-treated mice and SPF mice were screened using non-parametric tests and fold change analysis. Open-source metabolomics databases were searched using automated software MS-DIAL and MetEx to annotate the differential metabolites structurally. Finally, gut microbiota-related metabolites not included in the knowledge base were added to the database. The method of this invention combines knowledge-based and experimentally measured gut microbiota-related metabolites in biological samples, providing a framework for constructing high-coverage microbiota metabolite databases and laying the foundation for large-scale annotation of gut microbiota-related metabolites in metabolomics data and the study of the molecular mechanisms of microbiota-host interactions.

[0023] Appendix Explanation

[0024] Table 1. Gut microbiota-related metabolites reported in 968 literatures.

[0025] Table 2. 282 newly defined gut microbiota-associated metabolites in sterile / antibiotic-treated / specific pathogen-free mouse models.

[0026] Table 3. 229 representative gut microbiota-related metabolite standards. Attached Figure Description

[0027] Figure 1 Procedure for experiments using sterile / antibiotic-treated / specific pathogen-free mouse models. FC (fold change): fold change rate.

[0028] Figure 2 The number and trends of actual detected and significantly differentially expressed metabolites in three types of mouse samples.

[0029] Figure 3 Venn diagram of newly annotated gut microbiota-related metabolites in three types of mouse samples.

[0030] Figure 4 The database contains 1250 gut microbiota-related metabolites with diverse chemical categories.

[0031] Figure 5 Comparison of chromatographic separation performance of two candidate columns.

[0032] Figure 6 Extraction ion chromatograms of 229 gut microbiota-related metabolite standards. Detailed Implementation

[0033] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings: the embodiments are implemented based on the technical solution of the present invention, and detailed implementation methods and specific operation processes are given, but the protection scope of the present invention is not limited to the following embodiments.

[0034] Example 1

[0035] High-coverage construction of a database of gut microbiota-related metabolites

[0036] (1) Literature searches were conducted in ISI Web of Science (https: / / www.webofscience.com / wos / alldb / basic-search) and PubMed (https: / / pubmed.ncbi.nlm.nih.gov / ) using the keywords “gut microbial metabolites”, “gut microbiota / gut flora&metabolomics / metabonomics”, and “gut microbiota-host co-metabolism”. Measured rather than predicted gut microbiota-related metabolites reported in research papers and reviews, with the host limited to humans or mice, were collected. Their corresponding PubChem CID numbers were then searched in PubChem (https: / / pubchem.ncbi.nlm.nih.gov / ). The metabolites were then deduplicated based on their PubChem CIDs, ultimately establishing a database containing 968 knowledge-oriented gut microbiota-related metabolites. The names of these 968 gut microbiota-related metabolites and their corresponding PubChem CID numbers are listed in Appendix 1;

[0037] (2) Germ-free mice, antibiotic-treated mice, and specific pathogen-free mouse models were used to define gut microbiota-related metabolites in actual biological samples, thereby expanding the self-built database. See Appendix for the specific process. Figure 1Germ-free mice (GF) are mice raised in an isolation system that have been tested and found to be free of any microorganisms and parasites, both inside and outside the body. Antibiotic-treated mice (ABX) are normal mice that have been treated with antibiotics (usually by gavage / drinking water, in this case, free access to water) for a period of time (usually 1-2 weeks, in this case, 2 weeks). Specific-pathogen-free mice (SPF) are mice in which no specific microorganisms or parasites are present in the body; they generally refer to healthy mice without infectious diseases (Reference: State Science and Technology Commission, Regulations on the Management of Laboratory Animals, 1988).

[0038] The construction process of the above three types of mouse models is as follows: Twelve 8-week-old wild-type C57BL / 6 male mice were housed under specific pathogen-free conditions; antibiotic-treated mice (the other six mice were specific pathogen-free mice) had their sterile drinking water supplemented with an antibiotic aqueous solution containing neomycin (final concentration in drinking water of 100 μg / mL), streptomycin (final concentration of 50 μg / mL), penicillin (final concentration of 100 U / mL), vancomycin (final concentration of 50 μg / mL), metronidazole (final concentration of 100 μg / mL), bacitracin (final concentration of 1 mg / mL), ciprofloxacin (final concentration of 125 μg / mL), ceftazidime (final concentration of 100 μg / mL), and gentamicin (final concentration of 170 μg / mL), and were allowed to drink freely for 2 weeks; another 6 10-week-old germ-free C57BL / 6 male mice were placed in a strictly sterile room (C57BL / 6 germ-free mice).

[0039] The rearing conditions were as follows: ambient temperature 22±1℃, filtered air ventilation, constant humidity (40%-70%), 12-hour light / dark cycle, and provision of purified and sterile standard customized feed and autoclaved distilled water (Jiangsu Xietong Pharmaceutical Biotechnology Co., Ltd.). Six sex- (male) and age- (10 weeks) matched C57BL / 6 germ-free mice, six antibiotic-treated mice, and six specific pathogen-free mice were obtained. After euthanasia in a carbon dioxide chamber, plasma, urine, and fecal samples were collected, rapidly frozen, and stored at -80℃ for later use.

[0040] (3) Plasma, urine, and fecal samples from germ-free mice, antibiotic-treated mice, and specific pathogen-free mice were thawed on ice and then subjected to simple, non-selective pretreatment. The specific process was as follows: 150 μL of plasma was placed in a 1.5 mL centrifuge tube, and 600 μL of methanol was added for protein precipitation and metabolite extraction. The tube was vortexed for 60 s. 50 μL of urine was placed in a 1.5 mL centrifuge tube. Add 200 μL of methanol to a centrifuge tube and vortex for 60 s. Weigh 10 ± 1 mg of lyophilized feces (after 48 h) into a 2 mL centrifuge tube, add zirconium oxide grinding beads, 300 μL of ultrapure water, and 300 μL of methanol in sequence, homogenize with a grinder at 30 Hz for 1 min, repeat 5 times, and then incubate at room temperature for 5 min. Place the above plasma, urine, and fecal samples into a centrifuge and centrifuge at 18,000 g for 15 min at 4 °C. Collect all the supernatant after centrifugation. The fecal supernatant needs to be passed through a nylon filter with a pore size of 0.22 μm to obtain 280 μL of filtrate. Then place it in a vacuum freeze dryer and freeze dry at 4 °C.

[0041] (4) Reconstitute the lyophilized plasma, urine and feces samples with 50 μL acetonitrile / water (20 / 80, v / v), 150 μL acetonitrile / water (5 / 95, v / v) and 70 μL acetonitrile / water (20 / 80, v / v), respectively; vortex until completely dissolved, place in a centrifuge, centrifuge at 18,000g for 10-15 min at 4℃, and take the supernatant for sample injection analysis.

[0042] (5) A non-targeted metabolomics analysis of mouse samples was performed using a high-performance liquid chromatography (HPLC) method based on a pentafluorophenyl column combined with a high-resolution mass spectrometry (HMS) detection method based on a data-dependent secondary mass spectrometry acquisition mode of full scan / iteration. This method effectively separated gut microbiota-related metabolites and achieved high coverage acquisition of secondary mass spectrometry information. The HPLC-HMS operating conditions are as follows:

[0043] Liquid chromatography conditions: Supelco Discovery HS-F5 column (Merck, Germany), specifications: inner diameter 2.1 mm × column length 150 mm, packing particle size 3 μm; column temperature 40℃, flow rate 0.25 mL / min, injection volume 5 μL, injection chamber temperature 6℃; in electrospray ionization (ESI) positive ionization mode. + Under the following conditions, mobile phase A is an aqueous solution containing 0.1% formic acid (by volume), and mobile phase B is an acetonitrile system containing 0.1% formic acid (by volume); in electrospray negative ionization mode (ESI)... -Under these conditions, mobile phase A was an aqueous solution containing 10 mM ammonium formate, and mobile phase B was an acetonitrile / water (95 / 5, v / v) solution containing 10 mM ammonium formate. A gradient elution program (v / v) was used, initially 0% B phase, held for 2 min; then linearly increased to 60% B phase over 18 min; then linearly increased to 100% B phase over 3 min, and the system was flushed for 3 min; finally switched back to 0% B phase and equilibrated for 4 min.

[0044] (6) Mass spectrometry conditions: Electrospray ionization source, combined positive and negative ion modes; spray voltage at ESI + The voltage is 3.5kV in the mode, at ESI. - The voltage was 3.0 kV; the capillary temperature was 320 °C, and the auxiliary gas heater temperature was 350 °C; the sheath gas and auxiliary gas flow rates were 45 and 10, respectively; the S-lens RF level was 50. A full-scan / iterative data-dependent secondary mass spectrometry acquisition mode was used, with 4 iterations; the full-scan mode resolution was 60,000, and the scan range was 58-870; the data-dependent secondary mass spectrometry acquisition mode resolution was 30,000; secondary mass spectrometry acquisition was triggered by the top 10 ions with the strongest response in each full-scan cycle, with a dynamic exclusion duration of 10 s; normalized collision energies of 15, 30, and 45 were used. The automatic gain control target (AGC target) and maximum injection time (maximum IT) were 3 × 10⁻⁶ in the full-scan settings. 6 The ion capacity and 100ms are 1×10⁻⁶ under the secondary mass spectrometry acquisition settings. 5 Ion capacity and 50ms. Metabolomics data were preprocessed using Compound Discoverer (version 3.1, Thermo Fisher Scientific) for peak detection and alignment. Missing values ​​were removed using the 80% rule (removing ions with a frequency less than 80% in all groups), resulting in a characteristic peak table of detected metabolites. The peak areas of metabolites in the table were corrected using internal standards (standards added to the test samples) to eliminate systematic errors from preprocessing and instrument analysis, retaining stably detected metabolites. Fecal samples were also corrected for fecal mass (weight of feces), and urine samples were corrected for creatinine levels (relative abundance of the metabolite creatinine in urine). Finally, the corrected relative abundance of each metabolite in different biological samples was obtained. A total of 5777 (ESI) was stably detected in fecal samples. + ) and 4282 (ESI - ) metabolites, a total of 2563 (ESI) were stably detected in plasma samples. + ) and 2498 (ESI) - ) metabolites, a total of 6951 (ESI) were consistently detected in urine samples. +) and 6286 (ESI - ) metabolites (see appendix) Figure 2 a).

[0045] (7) Compare the relative abundance differences of metabolites between germ-free / antibiotic-treated mouse groups and specific pathogen-free mouse groups in three types of biological samples. Metabolites are considered to be related to gut microbiota if they meet the following two conditions:

[0046] Condition 1: Using SPSS Statistics 20.0 software, a nonparametric Wilcoxon and Mann-Whitney U test was performed. The relative abundance of this metabolite showed a significant difference between the germ-free mouse group and the specific pathogen-free mouse group, as indicated by a p-value (asymptotic significance) of less than 0.05 (the null hypothesis was that the abundance of this metabolite did not differ significantly between the germ-free and specific pathogen-free mouse groups, and a p-value of less than 0.05 indicates that the experimental results can reject the null hypothesis at the 5% significance level); Condition 2: Compared with the specific pathogen-free mouse group, the abundance of this metabolite increased or decreased simultaneously in both the germ-free mouse group and the antibiotic-treated mouse group, with no limit on the magnitude.

[0047] Ultimately, 3312 (ESI) was identified in feces. + ) and 2776 (ESI) - ) gut microbiota-related metabolites, 447 of which were identified in plasma (ESI) + ) and 624 (ESI) - ) gut microbiota-related metabolites, with 2458 (ESI) identified in urine. + ) and 2059 (ESI - ) gut microbiota-related metabolites (see appendix) Figure 2 a). Furthermore, some of these gut microbiota-related metabolites decreased relative to specific pathogen-free mice in germ-free / antibiotic-treated mice (due to gut microbiota-specific production or transformation), while others increased (due to gut microbiota-specific degradation). The distribution of these trends is shown in the appendix. Figure 2 b.

[0048] (9) Using the open-source software MS-DIAL (http: / / prime.psc.riken.jp / compms / msdial / main.html / ) and MetEx (http: / / www.metaboex.cn / MetEx / ), the public metabolomics databases HMDB (The Human Metabolome Database, https: / / hmdb.ca / ) and MyCompoundID were analyzed.

[0049] Automated searches were performed using databases including (https: / / www.mycompoundid.org / ), MoNA (MassBank of North America, https: / / mona.fiehnlab.ucdavis.edu / ), mzCloud (https: / / www.mzcloud.org / ), and GNPS (The Global Natural Product Social Molecular Networking, https: / / gnps.ucsd.edu / ). Structural annotations (metabolite name, molecular formula, structural formula, and PubChem CID number) were performed on the identified gut microbiota-related metabolites. Metabolite structural annotations that simultaneously met the following two conditions were retained:

[0050] Condition 1: The difference between the detected value of the exact mass number of the precursor ion and the theoretical value is less than 10 ppm; Condition 2: The matching score between the experimentally obtained secondary mass spectrometer and the secondary mass spectrometer in the open-source database is greater than 0.7.

[0051] A total of 540 gut microbiota-related metabolites were annotated from the three types of mouse samples. Using their PubChem CID numbers, a search was performed on the database containing 968 knowledge-guided gut microbiota-related metabolites constructed in step (1). 258 were found to be already included in this knowledge base, while the remaining 282 were not. The detection status of these 282 newly defined gut microbiota-related metabolites in the three types of samples is shown in the appendix. Figure 3 The names of their metabolites and their corresponding PubChem CID numbers are shown in Appendix 2.

[0052] Many gut microbiota-related metabolites from the knowledge base were not detected in the three mouse samples mentioned above. This may be due to species differences between mice and humans, or because the abundance of these metabolites in actual mouse samples is low, below the detection limit. Future research could potentially expand the coverage of gut microbiota-related metabolites by incorporating more diverse sample types, such as tissues and various body fluids, or by improving pretreatment conditions to concentrate the target metabolites in the samples.

[0053] (10) Finally, combining 968 previously reported gut microbiota-related metabolites from plasma, urine, and feces obtained through sterile / antibiotic-treated / specific pathogen-free mouse models, the self-built database contains a total of 1250 metabolites. The chemical diversity of these metabolites is shown in the appendix. Figure 4 .

[0054] The results show that the method of this invention combines knowledge-based and experimentally measured gut microbiota-related metabolites in biological samples, providing a framework for constructing a high-coverage microbiota metabolite database. This database can be used for extensive annotation and identification of gut microbiota-related metabolites in different biological samples, increasing the possibility of discovering new microbiota metabolites and providing a basis for subsequent research on the molecular mechanisms of gut microbiota-host interactions.

[0055] Example 2

[0056] The chromatographic performance of a high-performance liquid chromatography (HPLC) method based on a pentafluorophenyl column (Supelco Discovery HS-F5, 2.1 mm inner diameter × 150 mm column length, 3 μm packing particle size, Merck, Germany) was characterized using commercially available standard compounds of 229 representative bacterial metabolites from different categories included in a self-built database of gut microbiota-related metabolites. The chromatographic performance was compared with that of a commonly used HPLC method based on an ACUITY HSS T3 column (2.1 mm inner diameter × 100 mm column length, 1.8 μm packing particle size, Waters, USA).

[0057] (1) Preparation of mixed standards: Acetonitrile / water (20 / 80, v / v) was used as the solvent to prepare mixed standards containing 229 representative intestinal flora metabolites from different categories (see Appendix Table 3), all at a concentration of 1 μg / ml. The distribution of the oil-water partition coefficient (ClogP) of these 229 metabolites is shown in Appendix Table 3. Figure 5 a.

[0058] (2) The above mixed standard was analyzed by liquid chromatography-high resolution mass spectrometry based on a pentafluorophenyl column:

[0059] Liquid chromatography conditions: Supelco Discovery HS-F5 column, 2.1 mm inner diameter × 150 mm column length, 3 μm particle size; column temperature 40 °C, flow rate 0.25 mL / min, injection volume 5 μL, injection chamber temperature 6 °C; in electrospray ionization (ESI) positive ionization mode. + Under the following conditions, mobile phase A is an aqueous solution containing 0.1% formic acid, and mobile phase B is an acetonitrile system containing 0.1% formic acid; in electrospray negative ionization mode (ESI)... - Under these conditions, mobile phase A was an aqueous solution containing 10 mM ammonium formate, and mobile phase B was an acetonitrile / water (95 / 5, v / v) solution containing 10 mM ammonium formate. A gradient elution program was used, initially with 0% B phase for 2 min; then linearly increased to 100% B phase over 18 min, and the system was flushed for 5 min; finally, the system was switched back to 0% B phase and equilibrated for 5 min.

[0060] Mass spectrometry conditions: electrospray ionization source, combined positive and negative ion modes; spray voltage at ESI+ The voltage is 3.5kV in the mode, at ESI. - The system operates at 3.0 kV; capillary temperature is 320 °C, and auxiliary gas heater temperature is 350 °C; sheath gas and auxiliary gas flow rates are 45 and 10, respectively; the S-lens RF level is 50. A full-scan / data-dependent secondary mass spectrometry acquisition mode is used; the full-scan mode has a resolution of 60,000 and a scan range of 58–870; the data-dependent secondary mass spectrometry acquisition mode has a resolution of 30,000; secondary acquisition is triggered by the top 10 most responsive ions in each full-scan cycle, with a dynamic exclusion duration of 10 s; normalized collision energies of 15, 30, and 45 are used. The automatic gain control target (AGC target) and maximum injection time (maximum IT) are 3 × 10⁻⁶ in the full-scan setting. 6 The ion capacity and 100ms were set at 1×10⁻¹⁰ ms, respectively, while in the secondary mass spectrometry settings, these values ​​were 1×10⁻¹⁰ ms. 5 Ion capacity and 50ms.

[0061] (3) The above mixed standard was analyzed using liquid chromatography-high resolution mass spectrometry based on a Waters ACUITY HSS T3 column:

[0062] Liquid chromatography conditions: Waters ACQUITY HSS T3 column, 2.1 mm inner diameter × 100 mm column length, 1.8 μm particle size; column temperature 55 °C, flow rate 0.35 mL / min, injection volume 5 μL, injection chamber temperature 6 °C; in electrospray ionization (ESI) positive ionization mode. + Under the following conditions, mobile phase A is an aqueous solution containing 0.1% formic acid, and mobile phase B is an acetonitrile system containing 0.1% formic acid; in electrospray negative ionization mode (ESI)... - Under these conditions, mobile phase A was an aqueous solution containing 6.5 mM ammonium bicarbonate, and mobile phase B was a methanol / water (95 / 5, v / v) solution containing 6.5 mM ammonium bicarbonate. A gradient elution program was used, initially with 0% B phase for 2 min; then linearly increased to 100% B phase over 18 min, and the system was flushed for 5 min; finally, the system was switched back to 0% B phase and equilibrated for 5 min.

[0063] The mass spectrometry conditions are set in the same way as in step (2).

[0064] (4) Using Xcalibur 2.2SP1.48 software (Thermo Fisher Scientific), the precursor ions (ESIs) of 229 gut microbiota-related metabolites were analyzed. + [M+H] + ESI - [MH] -Feature extraction was performed on the raw data, with a mass error set to 10 ppm, to obtain the retention time, primary ion mass-to-charge ratio, and secondary mass spectrum (secondary ion mass-to-charge ratio and intensity) of metabolites that met the conditions. By comparing the experimental secondary spectrum with secondary spectra in the publicly available metabolomics databases HMDB (The Human Metabolome Database, https: / / hmdb.ca / ) and MoNA (Mass Bank of North America, https: / / mona.fiehnlab.ucdavis.edu / ), the detection of the standard under the two liquid chromatography separation conditions was determined.

[0065] (5) Analysis of standard sample detection:

[0066] When using a pentafluorophenyl column for liquid chromatography separation, a total of 227 standards were detected, with a total peak intensity of 1.08 × E0.05. 10 When using a T3 column for liquid chromatography separation, a total of 219 standard samples were detected, with a total peak intensity of 1.05 × E. 10 (See appendix) Figure 5 b). This indicates that the pentafluorophenyl column has a broader coverage and higher sensitivity for gut microbiota-related metabolites. Furthermore, the T3 column generally exhibits worse retention of test standards than the pentafluorophenyl column, characterized by high-density chromatographic elution peaks between 0 and 1 minute (see Appendix). Figure 5 c). Regarding isomer separation performance, the pentafluorophenyl column achieved baseline separation of N,N-dimethylglycine and 4-aminobutyric acid, while the chromatographic peaks of these two isomers overlapped on the T3 column (see appendix). Figure 5 d) indicates that the pentafluorophenyl column has a stronger isomer separation capability. Based on the above considerations, the Supelco Discovery HS-F5 (2.1×150mm, 3μm) column can achieve high coverage and effective separation analysis of the gut microbiota metabolome.

[0067] (6) The chromatographic elution gradient was optimized based on the peak distribution of the standards to ensure that the chromatographic peaks of the standards were distributed as evenly as possible along the retention time axis. The final optimized elution gradient for the Supelco Discovery HS-F5 column (2.1 mm inner diameter × 150 mm column length, 3 μm packing particle size) was as follows: initially 0% B phase, held for 2 min; then linearly increased to 60% B phase within 18 min; then linearly increased to 100% B phase within 3 min, followed by a 3 min system flush; finally, switched back to 0% B phase and equilibrated for 4 min. Under this gradient, 175 standards were detected in electrospray positive ion mode, and their chromatographic retention behavior is shown in the attached figure. Figure 6As shown in Figure a; 181 standard samples were detected in the electrospray negative ion mode, and their chromatographic retention behavior is shown in the attached figure. Figure 6 As shown in b.

[0068] The above experimental results show that the liquid chromatography-mass spectrometry method based on pentafluorophenyl columns can effectively separate intestinal flora-related metabolites, and is suitable for high-coverage analysis of flora metabolites in mouse plasma, urine and feces in the sterile / antibiotic-treated / specific pathogen-free mouse model-limited experiments of the present invention.

[0069] Table 1. Gut microbiota metabolites reported in 968 studies

[0070]

[0071]

[0072] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0073]

[0074]

[0075] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0076]

[0077]

[0078] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0079]

[0080]

[0081] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0082]

[0083]

[0084] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0085]

[0086]

[0087] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0088]

[0089]

[0090] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0091]

[0092]

[0093] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0094]

[0095]

[0096] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0097]

[0098]

[0099] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0100]

[0101]

[0102] Table 1. Gut microbiota metabolites reported in 968 studies. (Continued)

[0103]

[0104]

[0105] Table 2. Newly defined gut microbiota metabolites from 282 mouse models

[0106]

[0107]

[0108] Table 2. Newly defined gut microbiota metabolites from 282 mouse models. (Continued)

[0109]

[0110]

[0111] Table 2. Newly defined gut microbiota metabolites from 282 mouse models. (Continued)

[0112]

[0113]

[0114] Table 2. Newly defined gut microbiota metabolites from 282 mouse models. (Continued)

[0115]

[0116]

[0117] Table 3. 229 representative gut microbiota metabolite standards

[0118]

[0119]

[0120] Table 3. 229 representative gut microbiota metabolite standards. (Continued)

[0121]

[0122]

[0123] Table 3. 229 representative gut microbiota metabolite standards. (Continued)

[0124]

[0125]

Claims

1. A method for constructing a high-coverage database of gut microbiota-related metabolites, characterized in that: (1) Conduct literature searches in ISI Web of Science and PubMed to collect gut microbiota-related metabolites that have been measured but not predicted and whose host is limited to humans and mice, and remove duplicates of metabolites by manually checking PubChem CID to establish a database of gut microbiota-related metabolites based on prior knowledge. (2) Expand the gut microbiota-related metabolite database established in step (1), identify novel gut microbiota-related metabolites not reported in the literature from actual biological samples using germ-free mice, antibiotic-treated mice and specific pathogen-free mouse models, and add them to the knowledge-guided gut microbiota-related metabolite database to obtain a high-coverage gut microbiota-related metabolite database. Step (2) involves the following steps: First, plasma, urine, and fecal samples are collected from germ-free mice, antibiotic-treated mice, and specific pathogen-free mice of the same sex and age in the three groups of mice with the same sex-to-male ratio. Non-selective extraction methods are used to extract various intestinal flora-related metabolites from the mouse plasma, urine, and fecal samples. The supernatants are then centrifuged and freeze-dried. The freeze-dried samples are reconstituted in acetonitrile / water. A high-resolution mass spectrometry detection method based on a pentafluorophenyl column-based high-performance liquid chromatography separation-based full-scan / iterative data-dependent secondary mass spectrometry acquisition mode is used to perform non-targeted metabolomics analysis on the reconstituted biological samples, obtaining rich liquid chromatography retention time, primary mass spectrometry information (i.e., mass-to-charge ratio and intensity of primary ions), and secondary mass spectrometry information (i.e., secondary...). The mass-to-charge ratio and intensity of the ions were determined; peak detection, peak alignment, and missing value removal were performed on the metabolomics data to obtain a table of characteristic peaks of the detected metabolites; further data preprocessing was performed, including internal standard correction, fecal quality correction, and creatinine response correction in urine; the differences in metabolite abundance between the germ-free / antibiotic-treated mouse group and the specific pathogen-free mouse group were compared in three types of biological samples, and the characteristics of gut microbiota-related metabolites were defined by combining nonparametric tests and statistical analysis of fold change; finally, the gut microbiota-related metabolites defined were annotated by automatically searching public metabolomics databases using open-source software MS-DIAL and MetEx; and the gut microbiota-related metabolites not included in the knowledge base established in step (1) were expanded into the database.

2. The method according to claim 1, characterized in that, The non-selective extraction method uses methanol / water as the extractant and employs vortex or bead milling extraction. The metabolite characteristic peak table includes the precursor ion, retention time, and peak area of ​​the metabolite.

3. The method according to claim 1, characterized in that, In step (1), in order to establish a high-coverage knowledge base of gut microbiota-related metabolites, literature searches were conducted in ISI Web of Science and PubMed using the keywords "gut microbial metabolites", "gutmicrobiota / gut flora & metabolomics / metabonomics", and "gutmicrobiota-host co-metabolism". The collected measured rather than predicted gut microbiota-related metabolites reported in research papers and reviews, with the host limited to humans and mice, were collected. The corresponding PubChemCIDs were then searched in PubChem. The metabolites were deduplicated according to the PubChem CIDs to establish a knowledge base of gut microbiota-related metabolites based on prior knowledge.

4. The method according to claim 1, characterized in that, In step (2), three mouse models were used: Germ-free mice (GF) refer to mice raised in an isolation system that have no detectable microorganisms or parasites in or outside the body; Antibiotic-treated mice (ABX) refer to normal mice that have been treated with antibiotics for 1-2 weeks; Specific-pathogen-free mice (SPF) are mice in which no specific microorganisms or parasites are present in the body, generally referring to healthy mice without infectious diseases. Sex-matched and age-matched germ-free mice, antibiotic-treated mice and specific-pathogen-free mice models were constructed, with at least 6 mice in each group, and their plasma, urine and fecal samples were collected using a standardized procedure.

5. The method according to claim 1, characterized in that, In step (2), to maximize the coverage of various gut microbiota-related metabolites and minimize losses during extraction, the pretreatment of mouse plasma, urine, and feces was kept as simple and non-selective as possible. Specifically: 50-150 μL of plasma was placed in a 1.5 mL centrifuge tube, and 200-600 μL of methanol was added for protein precipitation and metabolite extraction, followed by vortexing for 60 s; 50-150 μL of urine was placed in a 1.5 mL centrifuge tube, and 200-600 μL of methanol was added, followed by vortexing for 60 s; 10 ± 1 mg of lyophilized feces (after 48 h) was weighed into a 2 mL centrifuge tube, and zirconium oxide grinding beads, 300 μL of ultrapure water, and 300 μL of methanol were added sequentially. The mixture was homogenized at 30 Hz for 1 min using a mixing grinder, repeated 5 times, and then incubated at room temperature for 5 min; the extracted plasma, urine, and feces samples were then placed in a centrifuge and centrifuged at 18,000 g for 15 minutes at 4 ℃. After centrifugation, collect all the supernatant. The fecal supernatant needs to be passed through a nylon filter with a pore size of 0.22 μm to obtain 280 μL of filtrate. Then, place it in a vacuum freeze dryer and freeze dry at 4 °C.

6. The method according to claim 1, characterized in that, In step (2), the lyophilized samples of plasma, urine and feces were reconstituted with 50-150 μL of acetonitrile / water with a volume ratio of 20 / 80, 50-150 μL of acetonitrile / water with a volume ratio of 5 / 95, and 70 μL of acetonitrile / water with a volume ratio of 20 / 80, respectively. After vortexing until completely dissolved, the samples were placed in a centrifuge and centrifuged at 18,000 g for 10-15 min at 4 ℃. The supernatant was then collected for analysis.

7. The method according to claim 1, characterized in that, Step (2) uses high-performance liquid chromatography (HPLC) based on a pentafluorophenyl column combined with high-resolution mass spectrometry (HS-MS) detection based on full-scan / iteration data-dependent secondary mass spectrometry acquisition to perform non-targeted metabolomics analysis on mouse samples, in order to achieve effective separation of gut microbiota-related metabolites and high coverage acquisition of secondary mass spectrometry information; the HPLC-MS operating conditions are as follows: (1) Liquid chromatography conditions: Supelco Discovery HS-F5 column, with an inner diameter of 2.1 mm and a column length of 150 mm, and a particle size of 3 μm; column temperature 40 ℃, flow rate 0.25 mL / min, injection volume 5 μL, and injection chamber temperature 6 ℃; in electrospray positive ionization mode (ESI) + Under the following conditions, mobile phase A is an aqueous solution containing 0.1% formic acid, and mobile phase B is an acetonitrile system containing 0.1% formic acid; in electrospray negative ionization mode (ESI)... - Under these conditions, mobile phase A was an aqueous solution containing 10 mM ammonium formate, and mobile phase B was an acetonitrile / water solution containing 10 mM ammonium formate at a volume ratio of 95 / 5. A gradient elution program was used, initially 0% B phase, held for 2 min; then linearly increased to 60% B phase over 18 min; then linearly increased to 100% B phase over 3 min, and the system was flushed for 3 min; finally switched back to 0% B phase and equilibrated for 4 min. (2) Mass spectrometry conditions: electrospray ionization source, combination of positive and negative ion modes; spray voltage at ESI + The voltage is 3.5 kV in the mode, at ESI. - The voltage was 3.0 kV; capillary temperature was 320 °C, and auxiliary gas heater temperature was 350 °C; sheath gas and auxiliary gas flow rates were 45 and 10, respectively; S-lens RF level was 50; a full-scan / iterative data-dependent secondary mass spectrometry acquisition mode was used, with 4 iterations; the full-scan mode resolution was 60,000, and the scan range was 58-870; the data-dependent secondary mass spectrometry acquisition mode resolution was 30,000, and secondary mass spectrometry acquisition was triggered by the top 10 ions with the strongest response in each full-scan cycle, with a dynamic exclusion duration of 10 s and normalized collision energies of 15, 30, and 45; the automatic gain control target (AGC target) and maximum injection time (maximum IT) were 3 × 10⁻⁶ in the full-scan setting. 6 The ion capacity and 100 ms, under secondary mass spectrometry acquisition settings, are 1 × 10⁻⁶. 5 Ion capacity and 50 ms.

8. The method according to claim 1, characterized in that, Step (2) preprocesses the metabolomics data, including peak detection, peak alignment, and 80% rule removal of missing values, to obtain a table of characteristic peaks of the detected metabolites; internal standard correction is performed on the peak areas of the metabolites in the table to eliminate systematic errors in the preprocessing and instrument analysis process, to obtain a table of stably detected metabolite peaks; fecal samples also need to be corrected for fecal quality, and urine samples also need to be corrected for creatinine levels; finally, the relative abundance of each metabolite in different biological samples is obtained.

9. The method according to claim 1, characterized in that, Step (2) Compare the relative abundance differences of metabolites in the germ-free / antibiotic-treated mouse group and the specific pathogen-free mouse group in the three types of biological samples. If the following two conditions are met, the metabolite is considered to be related to the gut microbiota: Condition 1: Using the non-parametric Wilcoxon and Mann-Whitney U test, the relative abundance of this metabolite showed a significant difference between the germ-free mouse group and the specific pathogen-free mouse group, with an asymptotic significance p-value less than 0.

05. This indicates that the null hypothesis was that the abundance of this metabolite did not differ significantly between the germ-free and specific pathogen-free mouse groups, and a p-value less than 0.05 means that the experimental results can reject the null hypothesis at the 5% significance level. Condition 2: Compared with the specific pathogen-free mouse group, the abundance of this metabolite increased or decreased simultaneously in both the germ-free mouse group and the antibiotic-treated mouse group, with no limit on the magnitude.

10. The method according to claim 1, characterized in that, Step (2) Automated searches were performed in the publicly available metabolomics databases HMDB, MyCompoundID, MoNA, mzCloud, and GNPS using the open-source software MS-DIAL and MetEx. Structural annotations were performed on the identified gut microbiota-related metabolites, and annotation results that met both of the following conditions were retained: Condition 1: The difference between the detected precise mass number of the precursor ion and the theoretical value is less than 10 ppm; Condition 2: The matching score between the experimentally obtained secondary mass spectrometer and the secondary mass spectrometer in the open-source database is greater than 0.7; For the gut microbiota-related metabolites annotated in step (2), search for novel gut microbiota-related metabolites that were not found in the knowledge base established in step (1) based on their PubChem CID.

Images