Folic acid and FBP (from bovine milk) were obtained from Sigma-Aldrich (St. Louis, MO, USA). (6R, S)-5,10-Methenyl-5,6,7,8-tetrahydrofolic acid (5,10-CH=THF) chloride, (6R, S)-5,6,7,8-tetrahydrofolic acid (THF) trihydrochloride, (6R, S)-5-formyl-5,6,7,8-tetrahydrofolic acid (5-CHO-THF) calcium salt, and (6S)-5-methyl-5,6,7,8-tetrahydrofolic acid (5-CH${}_3$-THF) calcium salt were purchased from Schircks Laboratories (Zurich, Switzerland) and used as folate standard compounds. 10-CHO-THF was prepared according to the method published previously [15]. Type II porcine kidney acetone powder (Sigma-Aldorich) was used as folate conjugases. $\alpha$-amylase (from Aspergillus oryzae), protease (type XIV, from Streptomyces griseus), and certified reference material BCR-485 (from mixed vegetables) were obtained from Sigma-Aldrich. Ultracel$\mathrm{\circledR }$-10K (Amicon$\mathrm{\circledR }$ Ultra-2 mL) was purchased from Merck Millipore Ltd. (Tullagreen, Ireland) and used for centrifugal filter units for separation from FBP to folate compounds. Purple laver (Neopyropia yezoensis, previously Porphyra yezoensis) products (dried, toasted, and seasoned and toasted), dried kombu (Saccharina japonica) products, dried wakame (Undaria pinnatifida) products, and boiled and dried hijiki (Sargassum fusiformis) products were obtained from local markets in Tottori City, Japan, on October 25, 2022.

To remove the endogenous folate compounds, porcine kidney (0.12 g) powder was dissolved in 20 mL of 0.1 mol/L sodium phosphate buffer (pH 6.1), treated with activated charcoal powder (2.0 g), stirred for 1 h at 4 °C, and centrifuged at 900 $\times$g for 10 min at 4 °C. The supernatant fraction was treated with a membrane filter (25AS020AS; ADVANTEC$\mathrm{\circledR }$ Tokyo Roshi Kaisha Ltd., Tokyo, Japan) and used as the folate conjugase.

Total folate compounds were extracted from commercially available dried purple laver products according to the tri-enzyme method [11]. Five grams of the laver products were homogenized with a mortar and pestle. An aliquot (0.2 g) of the homogenate was dissolved in 2.0 mL of 0.1 mol/L sodium phosphate buffer (pH 6.1) containing 2% (v/v) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol and diluted with distilled water to a final volume of 5.0 mL. Octanol (100 $\mu$L) was then added, and the homogenates were autoclaved at 100 °C for 10 min. After cooling to room temperature (25 °C), 1 mL of 0.1 mol/L sodium phosphate buffer (pH 6.1) containing 2% (v/v) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol and 100 $\mu$L of protease solution (7.0 $\times$ 10${}^{-4}$ U) were added to each homogenate, and the homogenates were then incubated at 37 °C for 3 h. To stop the protease enzyme reaction, the homogenates were heated at 100 °C for 3 min. After cooling with ice, they were treated with 100 $\mu$L of $\alpha$-amylase. solution (0.3 U) for 2 h at 37 °C. Thereafter the homogenates were further treated with 400 $\mu$L of porcine kidney folate conjugase for 16 h at 37 °C. To stop the enzyme reactions, the treated homogenates were heated at 100 °C for 3 min and then cooled to room temperature (25 °C). The homogenate was diluted with distilled water to a final volume of 10 mL, filtered through filler paper (type 2, 90 mm, ADVANTEC®), and used as a food folate extract.

Folate assays were performed using a polypropylene tube (13 $\times$ 100 mm, Bio-Rad Laboratories, Hercules, CA, USA), which contain the food extract (50 $\mu$L), 0.1 mol/L sodium phosphate buffer (pH 7.0, 200 $\mu$L), and Lactobacillus rhamnosus folate assay medium (1 mL). The prepared assay mixture was diluted with distilled water to give a final volume of 2.0 mL and then autoclaved at 121 °C for 5 min. After cooling to room temperature (25 °C), the tube was inoculated aseptically with 50 $\mu$L of L. rhamnosus ATCC 7469 pre-cultured in Lactobacilli inoculum broth as described above, washed three times with 5 mL of saline buffer, and dissolved in saline buffer to achieve 92% light transmittance at 660 nm. After incubating the tube at 37 °C for 22 h, its optical density at 660 nm was determined using a UV-2550 Spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Each food sample was assayed for folate contents in triplicates, and this was repeated at least thrice.

An Ultracel$\mathrm{\circledR }$-10K centrifugal filter unit was treated with 1 mL of 25% (v/v) methanol solution and centrifuged at 7000 $\times$g for 10 min to wash its membrane. Then, 1 mL of 50 mmol/L potassium phosphate buffer containing 2% (v/v) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol was added to the filter unit, which was centrifuged under the same conditions. Aliquots (100 $\mu$g) of lyophilized FBP was dissolved in 0.5 mL of 50 mmol/L potassium phosphate buffer containing 2% (v/v) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol. The FBP solution was added to each solution (0.5 mL) of standard folate compounds (100 ng/mL) in a microtube, mixed well, and incubated on ice for 30 min. Then, the mixture was transferred into the washed Ultracel$\mathrm{\circledR }$-10K centrifugal filter unit and centrifuged at 7000 $\times$ g for 10 min at 4 °C (Supplementary Fig. 1). The centrifugal filter unit was washed twice with 1 mL of 50 mmol/L potassium phosphate buffer containing 2% (v/v) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol to purify the formed FBP–folate complex. Folate compounds were liberated from the FBP complex by the denaturation of FBP and treated with 300 $\mu$L of 40 mmol/L trifluoroacetic acid containing 2% (v/v) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol, removed by centrifugation at 7000 $\times$g for 10 min at 4 °C. The remaining folate compounds on the membrane of the filter unit was recovered to wash twice with 400 $\mu$L of 25% (v/v) methanol solution containing 2% (%) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol. The filtrate fractions were combined and used as HPLC samples.

After the denatured FBP was recovered with 50 mmol/L potassium phosphate buffer from the filter unit, it was treated in 50 mmol/L potassium phosphate buffer for 24 h at 4 °C until the denatured FBP was completely converted to the renatured form. FBP could be reused approximately 10 times to purify folate compounds in this system.

After the pH of the purple laver extracts prepared as described above was adjusted to pH 7.4 by the addition of 1 mol/L NaOH, the treated extracts were diluted 1.5 times with distilled water. An aliquot (1.5 mL) of the diluted purple laver extracts was mixed with 1.0 mL of 200 $\mu$g/mL FBP solution at 4 °C for 20 min to purify folate compounds. Then, the solution was then transferred to an Ultracel$\mathrm{\circledR }$-10K centrifugal filter unit and subjected to centrifugation at 7000 $\times$g at 4 °C for 15 min. Subsequently, the FBP–folate complex was washed twice with 1 mL of 50 mmol/L potassium phosphate buffer (pH 7.4) containing 2% (%) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol.

After the washed FBP–folate complex was treated with 500 $\mu$L of 40 mmol/L trifluoroacetic acid solution containing 2% (%) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol to denature FBP, folate compounds were recovered by centrifugation under the same conditions. After the Ultracel$\mathrm{\circledR }$-10K unit was washed with 400 $\mu$L of 25% (v/v) methanol solution containing 2% (%) ascorbic acid and 0.2% (v/v) 2-mercaptoethanol, the filtrate and washing fraction were combined and used as an HPLC sample.

The prepared sample was analyzed by HPLC using a Shimazu HPLC apparatus (SCL10A system controller, LC-10Ai pump, CT-20A column oven, fluorescence detector Shimazu RF-530, and electrochemical detector GL Science ED 723) and CDS version 5 chromato-data processing system (LAsoft, Ltd., Chiba, Japan) (Supplementary Fig. 2). A 500-$\mu$L aliquot of each sample was placed on an IntertSusuain AQ-C18 HPLC column (5 $\mu$m, 4.6 $\times$ 100 mm, GL Sciences) at 40 °C and eluted equilibrated with 50 mmol/L potassium dihydrogen phosphate solution (pH 2.0) containing 7% (v/v) acetonitrile at a flow rate of 1.0 mL/min. Folate compounds were detected at 292-nm excitation/362-nm emission (fluorescence detector) and at 1000 mV versus Ag/AgCl (electrochemical detector). Authentic THF, 5-CH${}_3$-THF, and 5-CHO-THF were detected during monitoring at fluorescence detection, and the retention times were 6.6 min, 7.7 min, and 17.5 min, respectively (Supplementary Fig. 3). Authentic 5,10-CH=THF (with a retention time of 10.9 min) and folic acid (with a retention time of 21.4 min) were completely separated from other compounds contained in the sample such as ascorbic acid and 2-mercaptoethanol during monitoring at electrochemical detection; however, THF, 5-CH${}_3$-THF, and 5-CHO-THF could not. Since 10-CHO-THF and 5,10-CH=THF could not be separated under these conditions, the peak fraction with a retention time of 10.9 min represents the sum values of 10-CHO-THF and 5,10-CH=THF.

Each sample (2 g) of the commercially available dried purple laver products was homogenized using a mortar and pestle. Total B${}_{12}$ compounds were extracted from each homogenate by boiling with 40 mL of distilled water and 10 mL of 0.57 mol/L acetate buffer (pH 4.5) containing 0.05% (w/v) KCN for 30 min in the dark to convert B${}_{12}$ compounds into the CN forms. The extraction procedures were performed in a draft chamber (Dalton Co., Tokyo, Japan). An aliquot (20 mL) of each extract (100 mL) prepared above was placed on a Sep-Pak ® plus C18 cartridge (Waters Corp., Milford, MA, USA) that was activated with 5 mL of 100% ethanol and equilibrated with 10 mL of distilled water. B${}_{12}$ compounds were eluted from the C18 cartridge with 2 mL of 75% (v/v) ethanol. After the eluate was filtered with a DISMIC-25JP membrane filter (Toyo Roshi Kaisya, Ltd, Tokyo, Japan), the remaining B${}_{12}$ compounds on the filter were recovered with 1 mL of 75% (v/v) ethanol. The filtrates were combined and evaporated to dryness under reduced pressure. The residual fraction was dissolved in 1.0 mL of distilled water. B${}_{12}$ compounds were purified from the solution using an immunoaffinity column (EASI-EXTRACT B${}_{12}$ P80; $\phi$ 8.0 $\times$ 60 mm, R-Biopharm, Darmstadt, Germany) according to the manufacturer’s protocol. The elute was evaporated to dryness under pressure is reduced, dissolved in 100 $\mu$L of Milli-Q water, and used as an HPLC sample, as described previously [16]. The HPLC apparatus (SPD-10AV UV-Vis detector, SCL-10A VP system controller, DGU-20A${}_3$ degasser, LC-10Ai liquid chromatograph, CTO-20A column oven) and a reversed-phase HPLC column (Wakosil-II 5C18RS, $\phi$4.6 $\times$ 150 mm; FUJIFILM Wako Pure Chemical Corp., Osaka, Japan) were used. B${}_{12}$ compounds were isocratically eluted with 20% (v/v) methanol containing 1% (v/v) acetic acid at a flow rate of 1.0 mL/min at 40 °C and monitored by measuring the absorbance at 361 nm.

Total B${}_{12}$ was extracted from various edible seaweed products and determined using HPLC. Dried purple laver (nori) products contained substantial amounts (approximately 30–60 $\mu$g B${}_{12}$/100 g dry weight) of B${}_{12}$. These values were much higher than those in other edible seaweeds (0.5 $\mu$g/100 g dry weight) (Table 1). Moreover, nori products did not contain pseudovitamin B${}_{12}$ (Fig. 1) mostly found in edible cyanobacteria because only a single peak of B${}_{12}$ was detected during HPLC (Fig. 4) (Supplementary Fig. 4). These results coincide with previously reported values determined using a microbiological assay method [17].

Fig. 4.

HPLC chromatograms for authentic B${}_{\mathrm{12}}$ and corrinoid compounds present in dried purple laver products. (A) Authentic B${}_{12}$. (B) Corrinoid compounds present in dried purple laver products. These are typical HPLC chromatograms of authentic B${}_{12}$ and corrinoid compounds present in dried purple laver products for three independent experiments.

Miyamoto et al. [18] reported that the B${}_{12}$ contents of the seasoned and toasted purple laver products were reportedly about half of the values of the dried purple laver products. Similar results were obtained in this study (Table 1). No B${}_{12}$ content was reduced in dried purple laver products during the toasting process [18], suggesting that the decreased B${}_{12}$ contents in the seasoned and toasted laver products may be due to B${}_{12}$ destruction caused by the interaction of various seasonings rather than the toasting process.

Total folate compounds were extracted, treated with the tri-enzyme (proteinase, $\alpha$-amylase, and conjugase) method, and determined using microbiological assay. The total folate content was much higher in dried purple laver (nori) products (approximately 880–1300 $\mu$g/100 g dry weight) than in other edible seaweeds (230 $\mu$g/100 g dry weight). These values determined from the tri-enzyme-treated extracts of seaweed products were similar to those from the extracts treated with di-enzymes (proteinase and conjugase) that were adopted as an official method of food composition analysis in Japan [19]. These results indicated that dried purple laver (nori) products contain high levels of B${}_{12}$ and folate compounds compared with other seaweed products tested.

Before the HPLC analysis, folate compounds from food extracts should be purified. Affinity chromatography with FBP attached to agarose gel has been used to remove impurities more effectively. The preparation of affinity columns is time-consuming, and the folate-binding capacity of FBP is slightly lost during the preparation. Thus, we developed a simple method for purifying folate compounds from food extracts using FBP and Ultracel®-10K centrifugal filter unit (Supplementary Fig. 1). To evaluate the recovery rates of folate compounds during purification, various standard compounds such as THF, 5-CHO-THF, 5,10-CH=THF, and 5-CHO-THF were used. As shown in Supplementary Table 1, more than 85% of the recovery rate were obtained at each folate compound. To evaluate whether this method can be applied into food folate analysis, certified reference material BCR-485 (from mixed vegetables) was used as a sample. Our HPLC analysis indicated that 5-CH${}_3$-THF (229 $\mu$g/100 g) is the predominant folate compound in BCR-485, which contained 244 $\mu$g of total folate compounds per 100 g weight (Supplementary Table 2). These values coincided those reported previously, indicating that this method can be applicable to food folate analysis.

This HPLC method was performed to determine folate compounds found in some purple laver (nori) products. Fig. 5A,B showed the elution profiles of the FBP-purified compounds of a dried purple laver product in fluorescence and electrochemical detectors, respectively. The retention times of peaks 1 (6.7 min) and 2 (7.7 min) were identical to those of authentic THF and 5-CH${}_3$-THF, respectively, in fluorescence detection. Whereas in electrochemical detection, the retention times of peaks 1 (6.7 min), 3 (11.7 min), 4 (16.6 min), and 5 (20.1 min) were identical to those of authentic THF, 5,10-CH=THF (or 10-CHO-THF or both), 5-CHO-THF, and folic acid (FA), respectively. Similar elution patterns of FBP-purified compounds were found in all purple laver products. Table 2 summarizes the levels of folate compounds determined from commercially available purple laver products. 5-CH${}_3$-THF (163–253 $\mu$g/100 g dry weight) was the major folate compound in all purple laver products. 5-CHO-THF was found in the similar level (approximately 73–175 $\mu$g/100 g dry weight). 5,10-CH=THF and 10-CHO-THF (approximately 57–123 $\mu$g/100 g dry weight), THF (approximately 17–28 $\mu$g/100 g dry weight), and FA (9.8–52 $\mu$g/100 g dry weight) were the minor folate compounds found in purple laver products. FA would be due to the oxidation of the reduced folate compounds such as 5-CH${}_3$-THF during food processing and storage. The sum of these folate compounds identified in HPLC was approximately 40% of total folate compounds determined by the microbiological assay in each purple laver product, suggesting that some of the unidentified peaks may be derived from certain folate compounds. However, no information is available on whether these unknown 1–3 fractions contain some folate compounds because the antioxidant reagent 2-mercaptoethanol was eluted in the fraction of unknown peak 1. Although information on the levels of reduced folate compounds in edible seaweeds is very limited, Porphyra spp. (100 g, dry weight) reportedly contained approximately 61 $\mu$g of folate compounds, which consist of 5-CH${}_3$-THF (approximately 34 $\mu$g/100 g, dry weight) and FA (approximately 36 $\mu$g/100 g, dry weight) [20].

Fig. 5.

HPLC chromatograms for folate compounds present in dried purple laver products. (A) Fluorescence detector; and (B) electrochemical detector. The retention times of peaks 1, 2, 3, 4, and 5 were identical to those of authentic THF, 5-CH${}_3$-THF, and 5,10-CH = THF or 10-CHO-THF, 5-CHO-THF, and folic acid (FA), respectively.

Because the weight of one sheet (20 $\times$ 20 cm) of dried purple laver products is approximately 3 g [21], two sheets (approximately 6 g) of dried purple laver products would be sufficient to meet the recommended dietary allowance (RAD) of adults for B${}_{12}$ (2.4 $\mu$g/day). In RAD for folates (400 $\mu$g/day), ingestion of two sheets of the products would provide approximately 80 $\mu$g of folates per day [22].

It is concerning that the excessive consumption of edible seaweed products may lead to the ingestion of harmful amounts of iodine because substantial amounts of iodine (approximately 200 mg/100 g dry weight) have been found in dried kombu products [23, 24]. However, dried purple laver products reportedly contained less iodine (approximately 4–8 mg/100 g dry weight) [21]. Therefore, consuming two sheets (approximately 6 g/day) of dried purple laver products would not lead to an excessive intake of iodine (approximately 0.2–0.5 mg/day).