A high-resolution Orbitrap Mass spectral library for trace volatile compounds in fruit wines

The overall aroma is an important factor of the sensory quality of fruit wines, which attributed to hundreds of volatile compounds. However, the qualitative determination of trace volatile compounds is considered to be very challenging work. GC-Orbitrap-MS with high resolution and high sensitivity provided more possibilities for the determination of volatile compounds, but without the high-resolution mass spectral library. For accuracy of qualitative determination in fruit wines by GC-Orbitrap-MS, a high-resolution mass spectral library, including 76 volatile compounds, was developed in this study. Not only the HRMS spectrum but also the exact ion fragment, relative abundance, retention indices (RI), CAS number, chemical structure diagram, aroma description and aroma threshold (ortho-nasally) were provided and were shown in a database website (Food Flavor Laboratory, http://foodflavorlab.cn/). HRMS library was used to successfully identify the volatile compounds mentioned above in 16 fruit wines (5 blueberry wines, 6 goji berry wines and 5 hawthorn wines). The library was developed as an important basis for further understanding of trace volatile compounds in fruit wines. Measurement(s) volatile compounds Technology Type(s) GC-Orbitrap-MS Measurement(s) volatile compounds Technology Type(s) GC-Orbitrap-MS


Background & Summary
Among the hundreds of volatile compounds detected in fruit wines, only a small percentage of them could play key roles in the contribution of characteristic aroma 1 . Currently, the gas chromatograph-mass spectrometer has been widely used for the identification and quantification of aroma compounds. The quadrupole mass spectrometer (qMS) could be the most common mass spectrometer for analysis [2][3][4][5][6] . However, some trace analytes were difficult to be detected using qMS due to their low resolution and sensitivity 4,7-12 . These trace compounds needed to be identified by other detectors. The aldehydes and ketones could be detected in Syrah wines 13 and model wine solution by flame ionization detector (FID) 14,15 . The flame photometry (FPD) was used to identify sulfur compounds in Cabernet Sauvignon wines 16,17 . Besides, sulphur chemiluminescence (SCD) 13,18,19 and pulsed flame photometry (PFPD) 20,21 also could be used for the analysis of sulfur compounds in grape wines. The pyrazines could be identified in wines 22 and oak woods 23 by nitrogen-phosphorous detection (NPD). The triple-quadrupole mass spectrometer (QqQ-MS) in selected-reaction-monitoring (SRM) could identify lactones 24 , terpenes 25 and sulfur compounds 26 in wines. Thus, multiple methods had to be used for the detection of various aroma compounds 14,16 . Meanwhile, the use of multiple instruments is time-consuming and costly. And it is also difficult to have so many instruments in a same laboratory. And it is an urgent challenge to identify trace aroma volatile compounds mentioned above simply and effectively in fruit wines.
In recent years, high-resolution mass spectrometry, such as quadrupole-time-of-flight-MS (Q-TOF), could improve the accuracy of identification 22,23,27 . Since Orbitrap-MS technology invented by Alexander Makarov was first commercially available in 2005, this new technique of high resolution and high sensitivity mass spectrometry has been shown great advantages for qualitative and quantitative analysis of compounds [28][29][30] , and therefore many studies have been focused on metabolomics using liquid chromatography coupling [31][32][33][34] . After GC was coupled with Orbitrap-MS in 2015, its resolution could reach 60,000 (219 m/z, FWHM), mass accuracy could reach 1 ppm, and sensitivity could reach femtogram level, which provided more possibilities to advance the  35,36 . At present, GC-Orbitrap-MS began to be used to detect pesticide residues 37 , nitrosamines in children's products 38 , persistent organic pollutants in the environment 39 , soluble and extractable substances in package materials 40 , stimulants and banned substances in urine 41 and metabonomics 42 . GC-Orbitrap-MS can provide accurate qualitative quantification of benzene compounds in chili peppers 43 . In summary, the GC-Orbitrap-MS could be a potential technique for the determination of aroma volatile compounds in fruit wines due to its high resolution and high sensitivity. At present, the NIST library is widely used for the identification of aroma volatile compounds analyzed by gas chromatography-mass spectrometry 7,8,44,45 . However, the mass spectrums in the NIST library were mostly obtained by low-resolution mass spectrometry. There were differences in ion fragments and ion abundance between high-resolution mass spectrums obtained by GC-Orbitrap-MS and low-resolution mass spectrums obtained by GC-Quadrupole-MS 46 , which led to the qualitative inaccuracy. The high-resolution mass spectrometry (HRMS) spectrums of aroma compounds analyzed by GC-Orbitrap-MS need to be established for accurate identification. In addition, the basic information of aroma compounds, such as CAS number, chemical structure diagram, aroma description and aroma threshold (ortho-nasally), need to be acquired by a large collection of literature. Thus, there is an urgent need to establish a library of HRMS spectrum and basic information to facilitate analyzing and consulting by scholars all over the world.

Methods
Overview of the experimental design. Materials and methods. Chemical and reagentsThe information of standards was shown in Table 1. The individual stock solution of each standard is dissolved in ethanol and stored at −20 °C. Wine Samples collectionThree kinds of commercial fruit wines (blueberry wine, B, goji berry wine, G and hawthorn wine, H) purchased from retail stores in China were used for the establishment of HRMS library. All blueberry samples were with an alcohol content of 12% v/v (percent by volume). Three blueberry wines were received from Beiyushidai, including blueberry dry wine produced in 2019 (B1) and 2017 (B2) and blueberry semi-dry wine produced in 2019 (B3). A blueberry dry wine (B4) was produced by Shenghua in 2019. Another blueberry dry wine (B5) produced in 2019 was provided by Yicunshanye. Goji berry semi-dry wine (G1) was produced by Ningxiahong in 2019, with an alcohol content of 7% v/v. Four batches of goji berry dry wine (G2-G5) produced by Senmiao in 2017 were with an alcohol content of 11% v/v. G6 was made by our laboratory in 2016 with an alcohol content of 11% v/v. All hawthorn wine samples were semi-dry wines from Shengbali. H1 and H2 produced in 2019 were with an alcohol content of 12% v/v. The other H3-H5 were produced in 2020 with an alcohol content of 13% v/v from Shengbali. Preparation of the spiked mixtureThe direct liquid introduction method was used to determine the mass spectral information of the target compound. The standard mixtures (Mixture 1 with 24 esters, Mixture 2 with 6 carbonyl compounds and 8 lactones and 6 acids, Mixture 3 with10 high alcohols and 6 furans and 5 pyrazines, Mixture 4 with 5 terpenes and 4 benzenes and 4 volatile phenols and 1 sulfide) were prepared to extract. The mother solution of each compound was dissolved in ethanol at higher concentration. Each standard mixtures were mixed by the mother solution of compounds according to the concentrations (

Pyrazine
Continued www.nature.com/scientificdata www.nature.com/scientificdata/ mixtures were diluted with dichloromethane to volume in a 10-mL volumetric flask. 1 μL of each mixture was injected. The split mode was applied with a split ratio of 10:1. The liquid injection was performed using the TriPlus RSH autosampler (Thermo Fisher Scientific, Bremen, Germany). Extraction of volatile compounds in wine samplesHeadspace solid-phase microextraction (HS-SPME) was used to extract the volatile compounds from fruit wines. 5 mL of wine samples mixed with 1.00 g NaCl and 10 μL of internal standard (1.077 g/L 4-methyl-2-pentanol) were prepared in a 20 mL glass vial. The sample vials were stirred and heated at 60 °C for 30 min. Then the preconditioned fiber (50/30 μm Divinylbenzene/Carboxen/ Polydimethylsiloxane (DVB/CAR/PDMS)) was used to absorb the volatile compounds in the headspace of the sample via for 30 min at 60 °C. After absorption, the fiber was inserted into the GC injection port for desorbing at 250 °C for 10 min. Two technical replicates were performed for each sample. Automatic headspace solid-phase microextraction was performed on the TriPlus RSH autosampler. GC-Orbitrap-MS analysisA Thermo Scientific Trace 1300 gas chromatography equipped with a Thermo Scientific Q-Exactive Orbitrap mass spectrometer (GC-Orbitrap MS, Thermo Scientific, Bremen, Germany) was used for detection. The spiked mixture was performed under the following GC-Orbitrap-MS conditions. A TG-WAXMS 30 m × 0.25 mm × 0.25 μm (Thermo Scientific, Bremen, Germany) was used to separate analytes. Helium was used as the carrier gas (1.2 mL/min). The oven temperature program was set as follows: 40 °C held for 5 min, then heated to 180 °C at 3 °C/min, finally increased from 180 °C to 240 °C at 30 °C/min and hold 15 min. The wine samples were performed under the following GC-Orbitrap-MS conditions. A DB-WAX 30 m × 0.25 mm × 0.25 μm (J&W Scientific, Folsom, CA, USA) was used to separate the volatile compounds under a 1.2 mL/min flow rate of helium (carrier gas).The oven temperature program was set as follows: 40°C held for 5 min, then heated to 180°C at 3 °C /min, finally increased from 180 °C to 250 °C at 30 °C/min and hold 10 min.
The Orbitrap-MS operated in full-scan MS acquisition mode (m/z 33-350). The ion source was maintained at 280 °C with an MSD transfer line temperature of 230 °C. Positive ion-electron ionization (EI) was used at 70 electron volts (eV) in Orbitrap-MS. Identification of the compoundsRetention indices (RI) were calculated from the retention times of C6-C24 n-alkanes under the same chromatographic and mass spectrometric conditions. The high-solution mass spectrums of volatile compounds were collected in different standard mixtures. Then, the qualitative determination Compounds B1 B2 B3 B4 B5 G1 G2 G3 G4 G5 G6 H1 H2 H3 H4 H5 Table 3. The qualitative determination of target volatile compounds in goji berry wines, blueberry wines and hawthorn wines. 'B' represent blueberry wine, 'G' represent goji berry wine, 'H' represent hawthorn wine.
www.nature.com/scientificdata www.nature.com/scientificdata/ of target compounds in fruit wines was performed by the match of the retention time and ion fragments in samples and standards.The experimental design and analysis pipeline are shown in Fig. 1.

Data Records
A total of 36 original data files were stored in MetaboLights 47 , including 4 standard mixtures and 32 wine samples (two technical replicates).

technical Validation
Two technical replicates were performed on each wine sample. The qualitative determination of target volatile compounds in fruit wines was shown in Table 3.  www.nature.com/scientificdata www.nature.com/scientificdata/

Usage Notes
The HRMS library of volatile compounds was shown on the database website (http://foodflavorlab.cn/), including HRMS spectrum, exact ion fragment, relative abundance, RI, CAS number, chemical structure diagram, aroma description and aroma threshold (ortho-nasally). Table 1 showed CAS No., formula and RI of each target volatile compound. The information of standards and contents of spiked mixtures were shown in Table 1. Table 2 showed elemental composition judgments, exact ion fragments and error mass of each target volatile compound. Table 3 showed the qualitative determination of target volatile compounds in blueberry wine, goji berry wine and hawthorn wine. Figure 2 showed the web page of the database website (http://foodflavorlab.cn/) including the home page, upload page, search page and result page. Figure 3 showed the page view (PV) of the database website (http://foodflavorlab.cn/) from Nov. 2020 to May. 2022.

Code availability
The Processing setup, Quan browser and Qual browser (Thermo Fisher Scientific, Les Ulis, France) in Xcalibur version 4.1 and Thermo Scientific TraceFinder (version 4.1) were used for collecting the HRMS library of volatile compounds. The structures of the volatile compounds were drawn using ChemDraw Professional 17.0 (Cambridgesoft, USA). High-resolution mass spectrums are plotted using Python (version 3.7).