REDUCTION OF GLUCOSINOLATES CONTENT DURING SAYUR ASIN FERMENTATION [Penurunan Kandungan Berbagai Glukosinolat selama Fermentasi Sayur Asin]

Glucosinolates (GLSs), health promoting compounds commonly found in Brassica vegetables, were studied during sayur asin fermentation made from Indian mustard (B. juncea). The current preliminary study aims to investigate the changes of glucosinolates content during 3 and 7 days of fermentation in two different media, i.e. coconut water and tajin liquor, and salt concentrations of 2.5 and 10% . The glucosinolates were analysed by HPLC after sample extraction in hot methanol followed by purification and de-sulphation. Results show that s inigrin was the most dominant glucosinolate among others, i.e. gluconapin, glucobrassicin, 4-hydroxy-glucobrassicin, 4-methoxy-glucobrassicin, and neo glucobrassicin, accounting for about 1000 and 4000 μmol/100 g dw in raw Indian mustard. Unfortunately , fermentation has substantially reduced the glucosinolates content in sayur asin. After 3 days of fermentation the sinigrin content was reduced by 95% as compared to that in the raw vegetable. The indole GLSs 4-methoxy-glucobrassicin and neo-glucobrassic in concentration decreased to 80-90% of the fresh materials. However, the decreasing mechanisms as well as factors contributing to the decrease of the glucosinolates could not be explained yet.

Fermentation is one of the methods to prepare Brassica vegetables, e.g.sauerkraut (Ciska and Pathak, 2004;Uva et al. 2006).However, fermentation commonly leads to GLSs loss in the product (Daxenbichler et al. 1980;Ciska and Pathak, 2004;Martinez-Villaluenga et al. 2009).To our knowledge, there is no study on the GLSs behaviour along the sayur asin production.Sayur asin is prepared by spontaneous fer mentation of Indian mustard in the presence of salt.Either tajin liquor or coconut water is used as the medium.This paper aims to describe the changes of GLSs content in Indian mustard during sayur asin production and how these are affected by the fer mentation medium and salt concentration.

Preparation of sayur asin
Raw Indian mustard (Brassica juncea), local name: sawi pahit, was obtained from a local supplier in Bandungan, Central Java.Sampling was per for med from the pooled vegetables after randomly divided them into number of treatments and replication.Each treatment contained approximately 1500 grams of Indian mustard or six whole plants for three replications.Each vegetable was washed by running tap water followed by draining.Subsequently it was withered under sunlight for four hours.Each vegetable was added with salt (2.5 or 10% w/w) and rubbed gently by hand simultaneously.Subsequently, it was immersed into a glass jar containing either the tajin liquor (1:2 w/v) (Nugerahani et al. 2000) or coconut water (1:5 w/v).The ratio was estimated after interviewing two local sayur asin producer s.Eventually each sample was incubated at 25 30°C for 7 days.Each jar represented a replication of one treatment, containing two whole plants.Therefore, there were six samples per treatment for glucosinolates analysis, except for the samples of raw vegetable (n = 4) and coconut water-10% salt-day 3 (n = 5).
Tajin liquor was prepared by boiling rice (IR-64) in distilled water at a ratio 1:14 (w/v) for 15 min.Subsequently, salt was added at a ratio 1:40 (w/v of added water) and mixed.The liquid was separated and cooled at room temperature.Fresh coconut water was obtained from the local market.The glucosinolates were analysed in raw vegetable and after three and seven days of fer mentation based on common practice of local producers.Subsequent experiment was per formed to study glucosinolates during a shorter fer mentation period, i.e. three days.The withered vegetable was fer mented in coconut water (1:5 w/v) and 2.5% (w/w) of salt.The glucosinolates were analysed in raw and withered vegetables, one and three hours after immersion into the medium, and one to three days of fermentation.Each treatment was represented by six samples for analysis

Sample preparation
Both raw and processed vegetables were chopped into pieces of approximately less than 1 cm and mixed thoroughly.The chopped vegetable was frozen in liquid nitrogen followed by grinding in a Waring blender and stored in boxes at -20°C until fur ther lyophilisation (ThermoScientific LL150).Samples of fermentation media were stored in freezer at -20°C.

Extraction of glucosinolate
The extraction of glucosinolates (GLSs) was employed twice in each replication and used hot methanol (Merck) as described by Verkerk et al. (2001) with minor modifications.Lyophilised sample of 0.1 g was added with 2.4 mL hot methanol 70% and 1 mL glucotropaeolin 3 mM (provided by Food Quality and Design, WU) was added as the internal standard.Sample was incubated in waterbath at 75°C for 20 minutes and it was mixed every 5 minutes.Then, sample was centrifuged for 10 min at 6000 rpm (Hettich EBA20).The supernatant was collected while residue was re-extracted twice with 10 mL of hot methanol following similar procedure.Eventually all supernatants were combined.

Purification and de-sulphation
The extracted GLSs were purified and de-sulphated on a 1.5 cm DEAE Sephadex A-25 (Sigma-Aldrich) anion exchange in glass-wool (Merck) matrix column.The column was washed twice with 1 mL bi-distilled water, and loaded with 2 mL of the GLSs ex tract and washed twice with 1 mL sodium acetate (Merck) solution 20 mM.Sulphatase (Sigma-Aldrich) was added and it was incubated overnight at room temperature (23-27°C).Then, the de-sulpho-GLSs were eluted three times with 0.5 mL bi-distilled water.Finally, the elute was filtered by a 0.45 μm filter prior to HPLC analysis.

HPLC analysis
HPLC set-up used LC 10 Avp system (Shimadzu) comprising a pump system with low pressure gradient valve, a degasser, a system controller, and a converter board.A UV Vis detector was perfor med at a wavelength of 229 nm.Results were analysed by using Class VP software.The de-sulpho-GLSs were separated by GraceSmar t RP-18 5μ column at 1 mL/min flow rate.The elution was a gradient system of water containing 0.05% tetramethyl-ammonium chloride (code A) and acetonitrile (Merck)/water (20:80, v/v) containing 0.05% tetramethyl-ammonium chloride (code B).Total elution time was 31 min and gradient as follows: 100% A for 1 min, then within 20 min to 100% B and within 5 min back to 100% A followed by maintaining 100% A for 5 min.The GLSs were identified by comparing with standards of sinigrin and glucotropaeolin (provided by Food Quality and Design WU), and known GLS profiles of broccoli, cauliflower, red cabbage, radish, and Brussels sprouts, and confirming literatures.Each GLS was quantified as response factor against glucotropaeolin as internal standard.Values of glucosinolates will be presented in table and figure as mean ± SEM.

Glucosinolates in Indian mustard
The glucosinolates (GLSs) identified in Indian mustard were aliphatics, i.e. sinigrin and gluconapin, and indoles, i.e. glucobrassicin, 4-hydroxy-glucobrassicin, and 4-methoxyglucobrassicin and neo glucobrassicin, as shown in the chromatogram (Figure 1).Moreover, the GLS sinigrin accounted for more than 90% of the total GLSs.In general, the GLSs identified are in accordance with previous repor ts (Font et al.  2004; He et al. 2003 a ).However, He et al. (2003 a ) reported that gluconapin and progoitrin were dominant GLSs in growing potherb mustard (B.juncea Coss).D iscrepancies in the GLS profile could be due to the differences in variety, agricultural conditions, and the analytical methods.

Effect of fermentation on the glucosinolates content
Fermentation substantially reduced the GLSs content in Indian mustard during fermentation (Table 1).At the third fer mentation day, there was a remarkable drop of sinigr in for more than 95% compared to sinigrin in raw vegetable.Meanwhile, the concentration of indole 4-methoxyglucobrassicin and neo-glucobrassicin decreased at 80-90% as compared to ones in raw vegetable.Subsequent experiment was per for med in a shor ter period of fer mentation.Sinigrin, as the most dominant GLS among others in the Indian mustard, i s chosen to describe the GLS changes during fer mentation (Figure 2).This study shows that initial sinigrin content in Indian mustard was about 4 times lower compared to the one in the fir st experiment.It is likely that biological variability of the vegetable may affect GLS concentrations (He et al. 2003 b ;  Krumbein et al. 2005; Verkerk et al. 2009).Interestingly, sinigrin content slightly increased after withering until three hours of the vegetable immersion into medium.Previous studies also reported an increase of GLSs level, e.g. after microwave processing of red cabbage (Verkerk and Dekker, 2004), boiling of cauliflower (D'Antuono et al. 2007), and steaming of broccoli (Gliszczyñska-šwiglo et al. 2006;Miglio et al. 2008).It has been shown previously that such treatments (in this case are withering and fer mentation) might increase the extractability of GLS from the sample matrix in analy tical methods.Afterward, sinigr in of the day 1 sample decreased drastically at about 80% as compared to the previous hour 3 sample.Subsequent decrease was observed at the day 2 and 3 of fer mentation resulting in 7-8% of sinigrin in raw vegetable.Previous studies reported that fer mentation significantly reduce GLSs in Brassica vegetables.During sauerkraut production, Tolonen et al. (2002) observed a very low quantity of 4-methoxy-glucobrassicin while, others repor ted no GLSs in fer mented product (Daxenbichler et al. 1980;Mar tinez-Villaluenga et al. 2009).Meanwhile, Suzuki et al. (2006) reported smaller quantities of the GLSs in nozawanazuke, a fer mented product of Brassica rapa L. in Japan, including glucobrassicin and gluconastur tiin as the major GLSs, relative to the fresh nozawana.The most recent study on white cabbage fer mentation repor ted that GLSs content can be retained by a thermal treatment prior to fer mentation.After 71 h of fer mentation the cabbage still contained 35% of total GLSs before fermentation (Sarvan et al. 2013).The contents of GLS der ived products depend on the fer mentation conditions as well as the concentrations of GLS of the raw material (Ciska and Pathak, 2004).Previous studies reported GLS changes during fermentation based on the changes of identified breakdown products, e.g.isothiocyanates (ITCs) and cyanides from aliphatic GLSs, indole-3-carbinol and ascorbigen from glucobrassicin, and 2-phenylethyl ITCs from gluconasturtiin (Ciska and Pathak, 2004;Daxenbichler, 1980).
In the context of significant drops of GLSs content, the current result shows th at possible effects of fer mentation medium and salt concentration could not be explained clearly.In general, the use of tajin liquor may affect to the lower sinigrin than the one using coconut water.While, higher salt concentration (i.e.10% ) may also affect to the lower sinigrin retention at the third day of fer mentation even though, this was not clearly indicated in the sayur asin in coconut water.Further studies are needed to confir m the finding and to see any interactions between variables influencing GLSs content.
Mechanisms affecting GLSs content dur ing fer mentation are different from other processing by applying heat; however, these are scarcely studied (Nugrahedi et al. 2013).Bacter ia and sodium chloride might take significant role to change the GLSs profile (Tolonen et al. 2002;Suzuki et al. 2006;Martinez-Villaluenga et al. 2009).While, leaching of GLSs into the medium is scarcely involved in the sayur asin.Accordingly, Suzuki et al. (2006) repor ted the GLSs retention in the tissues after fermentation.