Introduction

Safranal is a flavor component of saffron. The stigma of Crocus sativus L. Saffron is one of the most expensive spices, especially in Europe, and its quality criteria are listed in ISO3632-1:2011 [1]. The ISO states that the amount of safranal in saffron contributes to the quality of saffron. Safranal exhibits several physiochemical properties [2], including pharmaceutical actions such as antianxiety [3], hypnotic [3, 4], and antidepressant [5] effects. In addition, safranal is drawing attention to its potential to improve sleep quality. Saffron is used as a food, food additive, and a crude drug in Japan. "Saffron color" is a food additive used as a yellow coloring agent obtained by the ethanol extraction of saffron. Its principal coloring components are crocin and crocetin, and safranal is also present [6]. Since crocin and crocetin are the main components of "Gardenia yellow," another food additive, the presence or absence of safranal distinguishes Gardenia yellow from Saffron color. However, there is no official specification for products made of saffron distributed as food additives in Japan. As a crude drug, saffron is listed in The Japanese Pharmacopoeia 18th edition (JP18) and is mainly used by women to improve the feeling of cold, pale, and premenstrual symptoms [7]. In the JP18 specification, crocin is the characteristic quality control component. The specification requires a comparison of the absorption at 438 nm of the test solution with a standard solution to ensure the quality of the saffron. However, safranal is not referred to in JP18, although it is dealt with as a component contributing to the quality of saffron in ISO3632-1:2011.

On the other hand, in ISO3632-2:2010, a method for the quantitative determination of safranal is defined using a UV–Visible spectrometer [8], in which the absorption of the saffron extract is measured using a spectrometer at the maximum absorption wavelength of safranal. However, the safranal content cannot be accurately determined because it is challenging to distinguish it from other components with absorptions near the wavelength used to measure safranal. Many studies on methods to quantify safranal have been reported, such as using UV–visible spectrometers [9], HPLC [10–12], and gas chromatography (GC) [13–15]. Some of these reports used commercial safranal reagents or isolated safranal with purities of more than 88% [9–12, 15], whereas others used safranal reagents with no information regarding their purities [13, 14]. In all cases, safranal, whose purity was unknown accurately, was used as the standard. Generally, the purity of a standard is regarded as 100% to construct a calibration curve, leading to quantitative values that are likely to include a bias of at least 10%. The quantitative values include some bias when the purity of the analytical standard is unknown, but no solution to this problem has been presented. The lack of analytical standards of accurately known purity for quantitative determination, especially for the quality control of natural products, is a significant problem because it makes it difficult to establish quality tests and evaluate the results.

To address this, we have developed a method using relative molar sensitivity (RMS) called the RMS method [16]. RMS is the ratio of an analyte's response intensity per mole to an RMS reference. Using identical chromatographic operational conditions allows RMS to be specific for the analyte and the RMS reference. Using the RMS of the analyte to the RMS reference allows the analyte content in a sample solution, to which a known amount of RMS reference has been added, to be quantified using chromatography. Quantification is based on the relationship between the ratio of the detector response intensities and the RMS. The RMS method permits using the RMS reference as an external standard instead of being added to the sample solution. Notably, the RMS method does not require a standard reference material identical to the analyte. Moreover, combined with quantitative NMR (¹H-qNMR), one can accurately determine the RMS without traditional reference materials. As reported in our previous studies, we have applied the RMS method to quantitatively determine active compounds in various natural food additives [17–21].

Based on our previous studies, the RMS method effectively determines compounds for which commercially unknown purity reagents, such as safranal, are unavailable. In the study reported here, we focused on developing a quantitative method for safranal in saffron using a GC-flame ionization detector (GC-FID). However, there is currently no information regarding the influence of inter-instrument differences of GC-FIDs in the RMS method because few studies have used GCFIDs [21, 22]. However, differences in the manufacture of HPLC instruments might affect the quantitative values obtained using HPLC–PDA [16]. Thus, we also compared the RMSs and quantitative values obtained using two GCFID instruments from different manufacturers.

Materials and reagents

Saffron products distributed as food (sample no. C2247) and as a Japanese Pharmacopoeia crude drug (sample no. C2248) (Takasago Industry Co., Ltd., Gifu, Japan) were purchased and used. Safranal reagent was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA). 1,4-Bis(trimethylsilyl) benzene-d₄ (1,4-BTMSB-d₄) (Cat. 024–17,031, lot no. KCL2411, certified purity: 100.0 ± 0.5%, k = 2) was obtained from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan) as a certified reference material for qNMR and as the RMS reference. Deuterated methanol (methanol-d₄) was purchased from MERCK (Darmstadt, Germany). All solvents used in this study were of HPLC or particular grade.

Instruments

•      Ultramicrobalance: XP2U (Mettler Toledo International Inc., Columbus, OH, USA.)

•      NMR: JNM-ECZ600 (600 MHz) (JEOL Ltd., Tokyo,

•   GC-FID: GC-2014 (Shimadzu Corporation, Kyoto, Japan) (GC1) and 7890A GC System (Agilen Technologies, Inc., Santa Clara, CA, USA) (GC2).

•      Ultrasonic cleaner: ASU-3 (oscillating frequency: 42 kHz) (AS ONE Corporation, Osaka, Japan).

•      Ball mill: Mixer mill MM 400 (Verder Scientific Co., Ltd., Tokyo, Japan).

Operating conditions

H‑qNMR analyses

Spectral width, − 5 to 15 ppm; data points, 60,018; pulse angle, 90°; relaxation delay, 60 s; scans, 8; pre-scans, 2; 13C decoupling, on (MPF8); spinning, off; temperature 25 ℃. Measurements were conducted in triplicate for each NMR standard solution.

GC‑FID analyses

Column, DB-5 (60 m × 0.25 mm, 0.25 µm film thickness, Agilent Technologies, Inc.); oven temperature program, starting at 60 ℃ (1 min hold) and increased to 150 ℃ at a rate of 3 ℃/min, then increased at 15 ℃/min to 240 ℃ (5 min hold); injector temperature, 220 ℃; split ratio, 1:10; carrier gas and flow, He 1.0 mL/min detector (FID) temperature, 240 ℃; makeup gas and flow, He or N₂ 30 mL/min (GC1) or 25 mL/min (GC2). He and N₂ were used as makeup gases to determine the RMS, while only He was used for the RMS method. Measurements were performed once for each solution prepared for GC-FID analyses.

Determination of the RMS of safranal to 1,4‑BTMSB‑d₄

Purity determination of safranal by ¹H‑qNMR

A sample solution for ¹H-qNMR (NMR standard solution) was prepared as follows: 10 mg of safranal reagent and 2 mg of 1,4-BTMSB-d₄ were accurately weighed in a vial, then dissolved in 2 mL of methanol-d₄. This solution was transferred to an NMR test tube, which was then sealed using a torch burner and assessed immediately after preparation by ¹H-qNMR under the conditions given in Sect. "¹H-qNMR analyses". The NMR standard solution was prepared three times.

¹H-qNMR spectra were analyzed using the software package Delta V5.3.3 (JEOL Ltd.). Using the H-3 signal (dt, CH) at δ5.75 ppm and the H-4 signal (m, CH) at δ5.97 ppm of safranal, and the methyl group signal of 1,4-BTMSB-d₄ at δ0 ppm (s, CH₃ × 6), safranal purity (P_s (w/w%)) was calculated as shown in Eq. (1):

Where P_rs is the certified purity of 1,4-BTMSB-d₄ (100%), I is the ¹H signal area, H is the number of ¹H nuclei in one molecule contributing to I, M is the molar mass (g/mol), and W is the amount (mg) of reagent used in the assay. The subscripts s and rs refer to safranal and 1,4-BTMSB-d₄.

Construction of calibration curves for safranal and 1,4‑BTMSB‑d₄ to determine the RMS

The NMR standard solution was prepared as described in Sect. "Purity determination of safranal by ¹H-qNMR" was diluted 50 to 2000-fold with ethanol to obtain GC traditional solutions 1–9 for RMS determination. The concentrations of safranal and 1,4-BTMSB-d₄ in GC standard solutions 1–9 for RMS determination were about 0.556–0.0139 and 0.0944–0.00236 µmol/mL, respectively. These GC traditional solutions were analyzed immediately after preparation by GC-FID using the conditions given in Sect. "GC-FID analyses." Measurements were performed with two GC-FIDs: GC1 and GC2. GC standard solutions 1–9 were prepared per the NMR standard solutions, and GC traditional solutions 1–9 were designed in three series.

A calibration curve for safranal was constructed using concentrations corrected for purity, as determined by ¹H-qNMR, while that of 1,4-BTMSB-d₄ was made with fixed concentrations due to its certified purity. The horizontal axes of the plots showed the molar concentration (µmol/mL) of the GC standard solutions, and the vertical axes showed the peak area obtained from the corresponding GC-FID chromatograms. Using the slopes of the calibration curves, the RMS of safranal to 1,4-BTMSBd₄ was determined as follows:

Where a is the slope of the calibration curve, the subscripts s and rs refer to safranal and 1,4-BTMSB-d₄. Since GC standard solutions 1–9 were prepared in three series, three calibration curves were constructed, and three RMSs were obtained. In the RMS method in this study, the average of the three RMSs was used as the RMS of safranal to 1,4-BTMSB-d₄.

Safranal determination in test solutions with known safranal concentrations

Test solutions with known safranal concentrations (test solution S) were prepared twice daily for analysis by the RMS and the absolute calibration curve methods for five days. The resultant safranal concentrations between these two determination methods were compared, and their repeatability and reproducibility were calculated by one-way analysis of variance.

Preparation of test solution S

Ten mg of safranal reagent was accurately weighed and dissolved in methanol to make precisely 2 mL. The solution was diluted 1000-fold with ethanol to obtain test solution S (0.0515 µmol/mL safranal). The safranal concentration was corrected for purity as determined by ¹H-qNMR as described in Sect. "Purity determination of safranal by ¹H-qNMR".

Safranal determination in test solution S by the RMS method

To prepare solutions containing the RMS reference, 2 mg of 1,4-BTMSB-d₄ was accurately weighed and dissolved in methanol to make precisely 2 mL. The solution was stepwise diluted with ethanol to prepare three 1,4-BTMSB-d₄ concentrations (0.0906–0.0181 µmol/mL) to use as standard solutions B1–B3. These traditional solutions and test solution S are designed in Sect. "Preparation of test solution S" was analyzed by GC-FID under the conditions described in Sect. "GC-FID analyses" to obtain the peak area (A_rs) of 1,4-BTMSB-d₄ in standard solutions B1–B3 and that (A_s) of safranal in test solution S. Safranal molar concentration C_s (µmol/mL) in test solution S was determined as follows:

Where C_rs is the molar concentration of 1,4-BTMSB-d₄ in standard solutions B1–B3 (µmol/mL), RMS is the value specified in Sect. "Construction of calibration curves for safranal and 1,4-BTMSB-d₄ to determine the RMS". The subscripts s and rs refer to safranal and 1,4-BTMSB-d₄.

Safranal determination in test solution S by the absolute calibration curve method

Ten mg of safranal reagent was accurately weighed and dissolved in methanol to make precisely 2 mL. The solution was stepwise diluted with ethanol to prepare five safranal concentrations (0.309–0.0257 µmol/mL) as standard solutions S1–S5. These traditional solutions and test solution S are designed in Sect. "Preparation of test solution S" was analyzed by GC-FID under the conditions described in Sect. "GC-FID analyses." A calibration curve for safranal was constructed using concentrations corrected for purity, as determined by ¹H-qNMR in Sect. "Purity determination of safranal by ¹H-qNMR". Its horizontal axis showed the molar concentration (µmol/mL), and the vertical axis gave the peak area. This calibration curve determined the molar concentration of safranal in test solution S.

Determination of safranal in saffron by the RMS method

Preparation of saffron extract

Saffron was kept in a desiccator with silica gel at room temperature for 24 h, then ground using a ball mill at 28.5 Hz for 15 s. Powdered saffron was accurately weighed (5 mg) into a 50 mL vial [the area of the base of the vial was about 1256 mm² (SV-50, NICHIDEN-RIKA GLASS Co., Ltd., Kobe, Japan)], to which 2 mL of ethanol was added to the vial, followed by sonication for 30 min and centrifugation at 5000 × g for 5 min. The supernatant was collected, and 2 mL of ethanol was added to the residue, followed by sonication for another 30 min. After centrifuging at 5000 × g for 5 min, the obtained supernatant was combined with the first supernatant, and then ethanol was added to make exactly 5 mL of total solution. The solution was filtered through a 0.45 µm pore size membrane filter, and the filtrate was used as the saffron extract. The saffron extract was immediately analyzed after preparation, as shown in Sect. "Safranal determination in the saffron extract by the RMS method."

Safranal determination in the saffron extract by the RMS method

Standard solution B1 (0.091 µmol/mL) and saffron extract prepared in Sects. "Safranal determination in test solution S by the absolute calibration curve method" and "Preparation of saffron extract" were analyzed using GC-FID under the conditions described in Sect. "GC-FID analyses." Using the peak area of 1,4-BTMSB-d₄ (A_rs) in standard solution B1 and that of safranal (A_s) in the saffron extract, the safranal molar concentration C_conc. (%) in saffron was determined as follows:

C_rs is the molar concentration of 1,4-BTMSB-d₄ in standard solutions B1 (µmol/mL), M is the molar mass (safranal, 150.22 g/mol), and RMS is the value determined in Sect. "Construction of calibration curves for safranal and 1,4-BTMSB-d₄ to determine the RMS", V is the volume of saffron extract (5 mL), and W is the weight of saffron used to prepare the saffron extract (mg). The subscripts s and rs refer to safranal and 1,4-BTMSB-d₄.

Results and discussion

Establishing the safranal extraction and GC‑FID conditions

The purity of the commercial safranal reagent purchased for this study was 76.8%, as determined by ¹H-qNMR. Due to the instability of safranal in methanol, the purity of safranal in methanol solution stored at four ℃ gradually decreased by about 30% over 75 days (Fig. 1), indicating that the standard solutions should be prepared just before the construction of the calibration curve. Moreover, the purity of safranal reagent stored at − 30 ℃, which is the reagent manufacturers' recommended storage temperature, decreased by about 5% over 180 days after opening the package (Fig. 1). Thus, safranal quantitative determination may vary widely, depending on the manufacturer of the safranal reagents used for the assay and their length and conditions of storage. In this study, we applied the RMS method to determine safranal in saffron quantitatively.

Establishing conditions for the extraction of safranal from saffron

The conditions for extracting safranal from saffron reported by Bononi et al. [15] require a large amount of saffron. We followed their extraction procedure but attempted to reduce the amount necessary for quantitative determination since saffron is costly. The amount of saffron and the volume of ethanol used as extraction solvent were reduced to one-tenth of that in the previous report [15] (5 mg and 2 mL, respectively) using a 10 mL vial (SV-10, NICHIDEN-RIKA GLASS Co., Ltd.) for extraction rather than a flask, then sonication was performed for 60 min to obtain the extract solution and analysis by GC-FID. To ensure that all safranal was extracted, 2 mL of ethanol was added to the residue, and extraction was repeated twice. However, GC-FID still detected safranal in the third extract solution, indicating that a simple scale-down of the previously reported previously reported safranal extraction procedure cannot extract safranal effectively from saffron.

Therefore, the safranal extraction conditions were examined. First, ethanol, acetone, and diethyl ether were selected as extraction solvent candidates. Safranal extraction from saffron was performed under the same conditions described above, except for the solvent. The exact amount of safranal was extracted using acetone or ethanol, while only one-fourth of the amount of safranal was removed when diethyl ether was used. We thus selected ethanol as the extraction solvent, as in the previous report [15].

 

Fig. 1 Time-dependent changes in the purity of safranal reagent. A Change in the purity of safranal reagent stored at – 30 ℃ after opening the package, B change in the purity of safranal dissolved in methanol (5 mg/mL) and held at four ℃

Second, we examined the balance between the amount of sample (saffron) and the volume of solvent (ethanol) used per extraction. When the amount of saffron was 5 mg or 10 mg, the volume of ethanol was fixed at 2 mL, while if the amount of saffron was set at 5 mg, then the volume of ethanol was 1 mL or 2 mL respectively. There was no difference in the amount of extracted safranal using these conditions.

Third, we examined whether the time for extraction (length of sonication) affects the amount of extracted safranal. We varied the sonication length (30, 60, and 90 min); the longer the sonication step, the more significant the amount of safranal extracted. However, more safranal was removed if extraction was conducted thrice, each with 30 min of sonication, compared to one 90-min sonication step.

Finally, we investigated the effect of the base area size of the vial used for extraction on the amount of safranal released by conducting the extractions in two different-sized vials: a 50 mL vial (base area: about 1256 mm²) and a 10 mL vial (base area: about 452.2 mm²). In both cases, safranal was extracted from 5 mg of saffron with 2 mL of ethanol using 30 min of sonication. Safranal was detected in the Third ethanol extraction. Peak 1: safranal

Fig. 2 GC-FID chromatograms of saffron extracts by ethanol using a 50 mL vial. A is the first ethanol extraction, B is the second ethanol extraction, and C is the first and second extracts by GC-FID.

However, almost no safranal was detected in the third extract when a 50 mL vial was used (Fig. 2), while safranal was still detected when a 10 mL vial was used. These results suggest that selecting vials with the appropriate base area size for the amount of saffron and the solvent volume for safranal extraction from saffron by sonication is essential. Based on these results, the conditions for preparing saffron extracts were determined as shown in Sect. "Preparation of saffron extract".

Establishing the GC‑FID conditions

Saffron extract was prepared and analyzed using GC-FID under the conditions described in the previous report [15]. To shorten the analysis time, modified conditions (different oven temperature programs, split ratio, and injection volume) were also investigated while ensuring that the peak shape, the area of safranal, and separation of the safranal peak from other components were not affected by the modified analysis conditions, compared with the original conditions in the previous report [15]. The conditions for GC-FID analysis of saffron extracts were thus determined as shown in Sect.' GC-FID analyses".

Selection of the RMS reference

Next, a compound for use as the RMS reference was

selected. Selecting an appropriate standard compound whose purity is known accurately, chemically stable, and readily available is essential. Furthermore, choosing a compound with low volatility and hygroscopicity is advisable for easy weighing and a retention time close to the analytes. We considered 1,4-BTMSB-d₄ to be a suitable candidate as the RMS reference. Certified reference 1,4-BTMSB-d₄ is available commercially and was analyzed by GC-FID under the conditions shown in Sect. "GC-FID analyses." It eluted at 32.9 min (Safranal, the quantitative analyte in this study, eluted at 27.5 min, Figure 3).

Determination of the RMS of safranal to 1,4‑BTMSB‑d₄

To determine the exact RMS of safranal to 1,4-BTMSBd₄, the purity of the safranal reagent was determined by ¹H-qNMR (Fig. 4). The signal used for quantification should be well separated from other signals derived from the analyte and impurities in the reagents, allowing facile integration of the area of the signal. The purity of safranal was determined using the H-3 signal at δ5.75 ppm and the H-4 signal at δ5.97 ppm of safranal.

As mentioned above, the purity of the safranal reagent decreased as the storage time increased; thus, NMR standard solution and GC standard solutions 1–9 for RMS determination were prepared and measured on the same day. The ¹H-qNMR spectral and GC-FID chromatographic results were used to prepare calibration curves for safranal and 1,4-BTMSB-d₄. The GC-FID analyses were performed using two GC-FID instruments (GC1 and GC2) made by different manufacturers. The resultant RMSs of safranal to 1,4-BTMSB-d₄ determined using GC1 and GC2 were 0.813 and 0.769, respectively, and the relative standard deviation (RSD%) of these RMS values was 3.18% (Table 1A).

Fig. 3 GC-FID chromatograms of a mixture of safranal and 1,4-BTMSB-d₄. Peak 1: safranal, peak 2: 1,4-BTMSB-d₄

Fig. 4 ¹H-Quantitative NMR (¹H-qNMR) spectrum of an NMR standard solution prepared using the safranal reagent. Quantitative signals are expanded in the insets. 

 

Table 1 Relative molar sensitivity (RMS) of safranal to 1,4-BTMSBd_4. A Comparison of RMSs by the makeup gas. B RMS was determined using two different GC-FID instruments.

 

Makeup gas GC1 GC2   N2 He 1 0.811 0.770 2 0.814 0.768 Average 0.791   RSD (%) 3.18   Makeup gas GC1 GC2   He He 1 0.777 0.765 2 0.772 0.765 3 0.772 0.767 Average 0.770   RSD (%) 0.624

In our previous experiments regarding RMS methods using several HPLC–PDA instruments from different manufacturers, the difference in the quantitative values obtained was 0.49–4.2% [16–19], in line with the RSD% value of 3.18% obtained in this study. However, empirically, some factors contributing to the differences in the quantitative values obtained using different HPLC–PDA instruments are due to the principle of detection used by the PDA detector, such as wavelength discrepancies between detectors, the effect of stray light, and differences in individual lamps. These factors were irrelevant in GC-FID analysis, suggesting that the differences in RMS observed in this study are due to other factors. The analysis conditions used with GC1 and GC2 involved different makeup gases: GC1 and GC2 used N₂ and He, respectively. Regardless of the makeup gas, the calibration curves showed good linearity, with a coefficient of determination greater than 0.999. However, in

Table 2 Comparison of the RMS method with the calibration curve method for the determination of safranal in test solution

The calibration curve method The RMS method   Standard solution B1 Standard solution B2 Standard solution B3 Accuracy (%) 98.8 99.4 101 144 Repeatability (%) 0.81 0.81 0.81 0.74 Reproducibility (%) 0.82 0.81 1.3 39

Accuracy means the percentage agreement between the safranal quantitative value determined experimentally and the theoretical value in test solution S

the case of GC1, which used N₂ as the makeup gas, the calibration curve curved slightly upwards compared with that of GC2, which used He as the makeup gas, leading to differences in the slopes of the two calibration curves (for GC1 and GC2) (see Eq. (2)). Therefore, we changed the makeup gas of GC1 from N₂ to He and re-determined the RMS of safranal to 1,4-BTMSB-d₄ and determined that the RMSs of GC1 and GC2 were 0.774 and 0.767, respectively. The RSD% of these RMSs was 0.624% (Table 1B). Based on these results, an RMS of 0.770 (the average value obtained using the two GC-FID instruments) seems appropriate for the RMS method for quantifying safranal in saffron. These results also showed that a difference in the makeup gas could affect the quantitative value determined by the RMS method.

Comparing the RMS method with the calibration curve method

To verify the RMS method for safranal quantitation, test solution S, with a known safranal concentration, was analyzed using the RMS and the calibration curve methods (Table 2). In the RMS method, 0.770 was used as the RMS, determined as described in the previous section, and standard solutions B1–B3 containing three levels of 1,4-BTMSB-d₄ concentration were used as the RMS reference. When common solution B3, which had the lowest concentration of 1,4-BTMSB-d₄ of the three traditional solutions, was used as the RMS reference, the accuracy (the percentage agreement between the safranal quantitative value determined experimentally and the theoretical value in test solution S) was 144%, and the reproducibility was 39%. Using standard solutions B1 and B2 as the RMS references, the accuracy was 99.4% and 101%, and the reproducibility was 0.81% and 1.3%, respectively. In our previous study, we reported that the variation in quantitative value determined by the RMS method using HPLC instruments was within 1.5% when the signal-to-noise ratio of the peaks of the analyte and the RMS reference was more than 50 [23]. The signal-to-noise ratios of 1,4-BTMSB-d₄ in standard solutions B1, B2, and B3 (average of 5 days) were 296, 145, and 44, respectively, consistent with our previous report [23].

On the other hand, when safranal in test solution S was determined by the calibration curve method, the accuracy was 98.8%. The reproducibility was 0.82%, comparable with those of the RMS method using standard solution B1 or B2 as the RMS reference (the signal-to-noise ratios of 1,4-BTMSBd₄ were more significant than 130). Notably, in the calibration curve method, the calibration curve's horizontal axis (the safranal concentration) was adjusted to account for the purity of the safranal reagent, determined by ¹H-qNMR. As shown in Figure 1, the purity of the safranal reagent used in this study varied depending on the length of storage, from 72% to 77%. If the purity of the safranal reagent was regarded as 100% and this value was used for generating the calibration curve, the resulting quantitative value would be 23–28% larger than the actual value. It would vary depending on how many days of analysis were performed after opening the package. Using the RMS method for safranal determination, the quantitative values were equivalent to those obtained by the calibration curve method, constructed using a safranal reagent whose purity was corrected according to the purity determined by ¹H-qNMR. Moreover, since a stable compound is used as the RMS reference, a few factors should affect the quantitative values, such as instability of the safranal reagents and differences in analysis dates and analysts. In these respects, the RMS method developed in this study is superior to the calibration curve method using safranal reagent as the standard without correcting for purity.

Determination of safranal in saffron by the RMS method

Safranal in saffron samples distributed as food or JP18 crude drug (C2247 and C2248, respectively, Fig. 5) was determined by the RMS method developed in this study. Saffron extracts were prepared three times, and the safranal content was measured by GC1 using the conditions described in Sect. "GC-FID analyses." The safranal content was 0.56–0.62% in C2247, and 0.57–0.63% in C2248, and no differences were observed by application.

Fig. 5 GC-FID chromatogram of the saffron extract [the ethanol extract solution of saffron (C2248)]. Peak 1: safranal

 Conclusions

In this study, we developed an RMS method based on GCFID using 1,4-BTMSB-d₄ as the RMS reference to determine safranal, a compound that contributes to the quality of saffron. Safranal reagent is liquid at room temperature and highly volatile, so it isn't easy to weigh accurately. Safranal gradually degraded even if stored at the temperature recommended by the manufacturer. The purity of the safranal reagent used in this study was determined by ¹H-qNMR shortly after purchase and was about 77%. When this safranal reagent is used to assess safranal in saffron using the calibration curve method, the quantitative value would be at least 20% larger than true. Moreover, since the safranal reagent is unstable and its purity decreased during storage, the quantitative value of safranal would likely vary greatly depending on the age of the standard solutions.

Analytes analyzed by GC-FID are often volatile and complex to weigh accurately. Once an appropriate compound is selected as the RMS reference and the RMS of the analyte to the RMS reference is determined, analysts can accurately determine the quantitative value by weighing only the RMS reference instead of the analyte. Thus, there is less likely variation in quantitative values due to the analytical skills of the analyst preparing the standard solutions. The RMS method developed here should reduce the factors that could cause variation and deviation from the actual value.

Inter-instrument differences of GC-FIDs were investigated for the first time in this study. The difference was relatively small compared with using HPLC–PDA. However, the RMS was affected by the makeup gas. There are few reports on the effects of differences in conditions and GC-FID manufacturers on RMS. The findings in the present study are valuable for developing the RMS method as a universal quantitative method, such as official methods.

Please send your full-text request to our email address info[at]flonita.com

Note: in your mail, please put the paper title in the subject or mail body section or use the share link button provided at the end of that post to share your look-up page with us.

Original paper reference;  Masumoto N, Ohno T, Suzuki T, Togawa T, Sugimoto N. Application of the relative molar sensitivity method using GC-FID to quantify safranal in saffron (Crocus sativus L.). Journal of Natural Medicines. 2023 Sep;77(4):829-38.

1.      International Organization for Standardization (2011) ISO 3632-1 International standard. Spices-Saffron (Crocus sativus L.) -Part 1: Specification. ISO 3632-1:2010

2.      Rezaee R, Hosseinzadeh H (2013) Safranal: from an aromatic natural product to a rewarding pharmacological agent. Iran J Basic Med Sci 16:12–26

3.      Hosseinzadeh H, Noraei NB (2009) Anxiolytic and hypnotic effect of Crocus sativus aqueous extract and its constituents, crocin and safranal, in mice. Phytother Res 23:768–774. https:// doi. org/ 10. 1002/ ptr. 2597

4.      Liu Z, Xu XH, Liu TY, Hong ZY, Urade Y, Huang ZL, Qu WM (2012) Safranal enhances non-rapid eye movement sleep in pentobarbital-treated mice. CNS Neurosci Ther 18:623–630. https://d oi. org/ 10. 1111/j. 1755- 5949. 2012. 00334.x

5.      Hosseinzadeh H, Karimi G, Niapoor M (2004) Antidepressant effects of Crocus sativus stigma extracts and its constituents, crocin and safranal, in mice. J Med Plants 3:48–58

6.      Japan Food Additives Association (1999) Kisontenkabutsu meibo syuusai hinmoku risuto tyuukaisyo explanatory notes of list of existing additives. Tokyo, Japan Food Additives Association (in Japanese)

7.      Notification No. 1225-6 (Dec. 25, 2015) Director-General of Evaluation Division, Pharmaceutical Safety and Environmental Health Bureau, Ministry of Health, Labour and Welfare (in Japanese)

8.       International Organization for Standardization (2010) ISO 3632-2 International standard. Spices-Saffron (Crocus sativus L.) -Part 2: Test methods. ISO 3632-2:2010

9.      Maggi L, Sánchez AM, Carmona M, Kanakis CD, Anastasaki E, Tarantilis PA, Polissiou MG, Alonso GL (2011) Rapid determination of safranal in the quality control of saffron spice (Crocus sativus L.). Food Chem 127:369–373. https://d oi.org/1 0 . 1016/j. foodc hem. 2011. 01. 028

10.     Lage M, Cantrell CL (2009) Quantification of saffron (Crocus sativus L.) metabolites crocins, picrocrocin and safranal for quality determination of the spice grown under different environmental Moroccan conditions. Sci Hortic 121:366–373. https://d oi.org/10.1016/j. scien ta. 2009. 02. 017

11.    García-Rodríguez MV, Serrano-Díaz J, Tarantilis PA, López-Córcoles H, Carmona M, Alonso GL (2014) Determination of saffron quality by high-performance liquid chromatography. J Agric Food Chem 62:8068–8074. https:// doi. org/ 10. 1021/ jf501 9356

12.    García-Rodríguez MV, López-Córcoles H, Alonso GL, Pappas CS, Polissiou MG, Tarantilis PA (2017) Comparative evaluation of an ISO 3632 method and an HPLC-DAD method for safranal quantity determination in saffron. Food Chem 221:838–843. https:// doi. org/ 10. 1016/j. foodc hem. 2016. 11. 089

13.    Culleré L, San-Juan F, Cacho J (2011) Characterisation of aroma active compounds of Spanish saffron by gas chromatography– olfactometry: quantitative evaluation of the most relevant aromatic compounds. Food Chem 127:1866–1871. https:// doi. org/ 10. 1016/j. foodc hem. 2011. 02. 015

14.    Amanpour A, Sonmezdag AS, Kelebek H, Selli S (2015) GCMS-olfactometric characterization of the most aroma-active components in a representative aromatic extract from Iranian saffron (Crocus sativus L.). Food Chem 182:251–256. https://d oi.org/1 0 . 1016/j. foodc hem. 2015. 03. 005

15.    Bononi M, Milella P, Tateo F (2015) Gas chromatography of safranal as preferable method for the commercial grading of saffron (Crocus sativus L.). Food Chem 176:17–21. https:// doi. org/ 10. 1016/j. foodc hem. 2014. 12. 047

16.    Nishizaki Y, Masumoto N, Sugimoto N (2019) Role of relative molar sensitivity based on quantitative NMR in ensuring reliability of food analysis: development of chromatography without need for analyte reference material. Foods Food Ingredients J Jpn 224:123–130 (in Japanese)

17.     Masumoto N, Ishizuki K, Nishizaki Y, Ohtsuki T, Kuroe M, Yamazaki T, Numata M, Matsufuji H, Sugimoto N, Sato K (2021) Determination of mogroside V in Luohanguo extract for daily quality control operation using relative molar sensitivity to singlereference caffeine. Chem Pharm Bull 69:18–25. https://d oi.org/10.1248/ cpb. c20- 00245

18.     Masumoto N, Nishizaki Y, Maruyama T, Igarashi Y, Nakajima K, Yamazaki T, Kuroe M, Numata M, Ihara T, Sugimoto N, Sato K (2019) Determination of perillaldehyde in Perilla herbs using relative molar sensitivity to single-reference diphenyl sulfone. J Nat Med 73:566–576. https:// doi. org/ 10. 1007/ s11418- 019- 01306-7

19.    Nishizaki Y, Ishizuki K, Masumoto N, Tada A, Sugimoto N, Sato K (2020) HPLC determination of quercetin using relative molar sensitivity to methylparaben as a single reference. Jpn J Food Chem Saf 27:42–52. https:// doi. org/ 10. 18891/ jjfcs. 27.2_ 42

20.    Nishizaki Y, Sato-Masumoto N, Yokota A, Mikawa T, Nakashima K, Yamazaki T, Kuroe M, Numata M, Ihara T, Ito Y, Sugimoto N, Sato K (2018) HPLC/PDA determination of carminic acid and 4-aminocarminic acid using relative molar sensitivities with respect to caffeine. Food Addit Contam Part A Chem Anal 35:838–847. https:// doi. org/ 10. 1080/ 19440 049. 2018. 14406 42

21.    Masumoto N, Nishizaki Y, Nakajima K, Sugimoto N, Sato K (2021) Determination of allyl isothiosyanate in mustard and Horseradish extracts by single reference GC and HPLC based on relative molar sensitivities. Food Hyg Saf Sci 62:73–78. https:// doi. org/ 10. 3358/ shoku eishi. 62. 73. (inJapanese)

22.     Kitamaki Y, Saito N, Yamazaki T, Otsuka S, Nakamura S, Nishizaki Y, Sugimoto N, Numata M, Ihara T (2017) Determination of PAHs in solution with a single reference standard by a combination of 1H quantitative NMR spectroscopy and chromatography. Anal Chem 89:6963–6968. https://d oi.org/1 0 .1021 /a cs.analch em. 6b050 74

 

23.    Sakai Y, Masumoto N, Nishizaki Y, Ohtsuki T, Matsufuji H, Sugimoto N, Sato K (2020) Investigation of factors affecting on quantitative value by single-reference HPLC method with the relative molar sensitivity (RMS) and application to determination of lutein in foods with function Claims. Jpn J Food Chem Saf 27:123–134. https:// doi. org/ 10. 18891/ jjfcs. 27.3_ 123. (inJapanese)