THE EFFECT OF Mucuna pruriens DETOXIFICATION , THROUGH SONICATION AND ALKALI OR ACID EXTRACTION , ON L-DOPA CONCENTRATION AND NUTRITIONAL VALUE

a Department of Animal Sciences, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL 32608 USA. Phone: +1 352 392 7527; Fax: +1 352 392 7851, P.O. Box 110910. Email: adesogan@ufl.edu; huisden@ufl.edu b Analytical Toxicology Core Laboratory, Department of Physiological Sciences, College of Veterinary Medicine, University of Florida, Gainesville, FL 32611, USA. *Corresponding Author


INTRODUCTION
Mucuna pruriens, a legume indigenous to tropical regions, can be used as a supplemental protein source in the diets of animals and people; it also has nutraceutical and anti-nutrient properties (Rai et al., 2017;Mohan et al., 2016).Mucuna pruriens beans have high concentrations of crude protein (CP, 25-38 g/100g) and starch (39-41 g/100g) (Adebowale et al., 2005;Ezeagu et al., 2003) and the bioavailability and amino acid concentrations of Mucuna protein isolates exceed levels recommended for humans by the Food and Agriculture Organization, the World Health Organization, and the United Nations for all amino acids except methionine and cysteine (Adebowale et al., 2007).The lysine concentration of Mucuna is particularly high (Bressani, 2002), therefore Mucuna is potentially a valuable supplementary protein source to cereal-based diets, which are known to be lysine deficient.However, Mucuna contains anti-nutritive factors (Mohan et al., 2016) the most potent of which is L-Dopa (Ukuchkwu et al., 2002).The concentration of Mucuna L-Dopa ranges from 3 to 7 g/100g on a dry basis (Szabo and Tebbett, 2002).Symptoms of Mucuna intake in humans and monogastric livestock include reduced feed intake, weight loss, diarrhea, vomiting, and skin lesions (Del Carmen et al., 2002;Szabo and Tebbett, 2002).Some processing techniques can reduce Mucuna L-Dopa concentration to the safe threshold of <0.4 g/100g (Eilitta et al., 2003).Mucuna L-Dopa is readily soluble in dilute solutions of hydrochloric acid (Daxenbichler et al., 1972).Acidification of water to pH 3 allows extraction of the L-Dopa in Mucuna beans at 1-mm particle size to safe levels in less than 8 hours (Teixeira et al., 2003).However, this treatment could result in protein loss because of the increased protein solubility at pH less than the isoelectric point (pH 4.0-5.0) of Mucuna protein (Adebowale et al., 2007).Ensiling Mucuna beans for 28 days reduced the L-Dopa concentration by 43 to 61%, but the concentration of residual L-Dopa in the bean exceeded safe levels (Huisden et al., 2014).Alkaline conditions may also facilitate inactivation of L-Dopa in Mucuna beans.Diallo et al. (2002) reported that a calcium hydroxide solution was more effective than water at removing L-Dopa from Mucuna bean.Soaking the beans in 4 g/100g calcium hydroxide solution for 48 hours reduced the L-Dopa concentration to 0.001 g/100g.Teixeira et al. (2003) also reported that extraction of Mucuna beans (1 mm particle size) in NaOH solution at pH 11 reduced L-Dopa to safe levels (<0.4 g/100g) in less than 8 hours.However, Teixeira et al. (2003) and Wanjekeche et al. (2003) reported that melanin is produced when Mucuna L-Dopa is extracted at alkaline pH and this makes the beans black.Melanin has been anecdotally associated with the formation of melanoma in some studies (Dollery, 1999;Letellier et al., 1999;Siple et al., 2000), but no evidence for this association was found in other studies (Weiner et al., 1993;Pfutzner and Przybilla, 1997;Fiala et al., 2002).Nevertheless, the black color of the alkali-extracted bean may reduce its acceptability (St-Laurent et al., 2002;Spence, 2018).Sonication is a method used in more recent laboratory L-Dopa extraction procedures and it does not discolor the beans.St-Laurent et al. (2002) reported 5 minutes to be the most effective duration for sonication.However, effects of sonication on the nutritive value of Mucuna are unknown.Successful removal of L-Dopa from Mucuna beans with solvents depends on the particle size with smaller particles generally promoting the rate of L-Dopa removal to a greater extent (Teixeira et al., 2003).
The objective of this study was to examine the effects of the method of extraction on the L-Dopa concentration and nutritional composition of finely (1 mm) or coarsely (6 mm) ground Mucuna beans.Methods examined included extraction in either acetic acid (pH 3) or sodium hydroxide (pH 11) for 8 hours or extraction by sonication (SON) in water (pH 7) for 5 minutes.This study is the second in a series of studies (Huisden et al., 2010(Huisden et al., , 2014(Huisden et al., and 2018) ) aimed at evaluating the efficacy of Mucuna detoxification methods and their effects on the nutritive value of the detoxified bean.
Extraction methods.The beans were crushed (Roller Mill model 10004, Peerless International, Missouri, USA) and either sieved to pass through a 6-mm screen (USA Standard Testing Sieve, Fisher Scientific, Pittsburgh, PA, USA) or ground in a Wiley mill to pass through a 1-mm screen (Arthur H. Thomas Company, Philadelphia, PA, USA).Of 24 representative 50-g samples of fine (1 mm) and coarse (6 mm) particles, each of 8 replicate samples was individually subjected to sonication (SON) in water (neutral pH) or soaked in acidic (ACD) or alkaline (ALK) solutions.The ACD solution was brought to pH 3 by adding 25 ml/100ml (v/v) acetic acid solution to distilled water, up to a final volume of 2 L. The alkaline solution of pH 11 was produced by adding sodium hydroxide to distilled water, up to a final volume of 2L.Each replicate suspension was shaken (Eberbach shaker, Michigan, USA) at room temperature for 8 hours, then sieved through four layers of cheesecloth and a Whatman #1 filter paper (1001-240, Fisher Scientific, Pittsburgh, PA, USA).The residues were subsequently rinsed with 1 L of distilled-deionized water.Eight replicate samples were also submerged in 2 L of water (pH 7) within a sonication bath (Branson Ultrasonics, Connecticut) and sonicated for 5 minutes at room temperature.For each treatment, pairs of replicate residues were composited for chemical analysis (n=4).
Chemical analysis.Sonicated and solvent-extracted residues were dried at 55 ºC to constant weight and ground to pass through a 1-mm screen in a Wiley mill (Arthur H. Thomas Company, Philadelphia, PA).Dry matter concentration was determined after drying at 60 o C for 72 hours and ash was measured by combustion in a muffle furnace at 550 o C overnight.Gross energy was determined with an adiabatic bomb calorimeter (1261 isoperibol bomb calorimeter, Parr Instrument Company, Moline, Illinois, USA), using benzoic acid as a standard.The following analyses were also conducted: aNDF (Van Soest et al., 1991), EE (AOAC, 1984; Method 24.005), WSC (MAFF, 1986), L-Dopa (Siddhuraju and Becker, 2001), CP (Noel and Hambleton, 1976), amino acids (AOAC, 1994; Method 994.12) and starch (Hall, 2001).Tannin concentration quantification was based on the Butanol-HCl procedure of Terrill et al. (1992) and the Radial Diffusion Assay for Tannins (Hagerman, 1987).Purified quebracho tannin extract was used as the standard and results were expressed as quebracho tannin equivalents.

Statistical analysis.
The experiment had a 2 (particle sizes) × 3 (extraction methods) factorial treatment structure with a control and it was analyzed as a completely randomized design with 4 replicates per treatment.Data were analyzed with the MIXED procedure (SAS, 2005) and a model containing treatment effects.Significance was declared at P<0.05 and means were separated with Fisher's F-protected least significant different test.This approach was used instead of contrasts to detect differences between all particle sizes by extraction method combinations.Means reported are least square means.

RESULTS
All processing methods reduced (P<0.05)L-Dopa concentrations of fine Mucuna particles from 2.8 g/100g to less than 0.2 g/100g (Figure 1).
Acid and alkali treatments made the solvents and extracted bean residues darker (Figure 2).All methods also reduced (P<0.05)CP and WSC concentrations of fine particles by 24-31% and 78-81%, respectively (Table 1) and increased (P<0.05)their aNDF and starch concentrations by at least 62 and 14%, respectively.Ether extract concentration of fine particles was reduced (P<0.05)from 5.5 g/100g to 4.2 g/100g by SON, whereas ACD and ALK reduced (P<0.05)GE values of fine particles by approximately 10%.The ash concentration of fine particles was increased (P<0.05) by 88% and 35% by ALK and SON, respectively.No tannins were detected in the Mucuna bean samples with either of the analytical methods that were used.The ACD treatment increased (P<0.05) the aNDF concentration of coarse particles by 35% but decreased (P<0.05)their WSC and EE concentrations by 51% and 31%.Ash concentration and GE of coarse particles were not affected (P>0.05) by any of the treatments.The amino acid profile of coarsely ground Mucuna was not measured because the detoxification methods had little effect on the L-Dopa concentration of the coarsely ground beans.

DISCUSSION
Removal of anti-nutrient properties, such as L-Dopa, may be crucial to the safe use of legumes as a food source (Mohan et al., 2016).Safe L-Dopa levels in Mucuna beans destined for monogastric livestock consumption are considered to be 0.4 g/100g or less (Eilitta et al., 2003).All extraction methods were equally effective in reducing the L-Dopa concentration of fine Mucuna particles to <0.2 g/100g and thus making them safe for consumption by monogastrics.However, the L-Dopa concentration of coarse Mucuna particles was not decreased by ACD and SON and the 29% reduction by ALK treatment was inadequate to make the bean safe for consumption by monogastrics.These particle-size dependent responses are in agreement with Teixeira et al. (2003), who showed that L-Dopa removal at pH 3 or 11 depended on particle size.It was more effective in beans ground to a 1-mm particle size versus those ground to 2-, 4-, and 8-mm sizes.The efficacy of the ACD and ALKtreatment of fine particles also agrees with the observations of Teixeira et al. (2003).
The CP concentration of fine particles was reduced by 24-31% by all detoxification methods but that of coarse particles was only reduced by SON.The latter was likely because of the cell rupturing effect of sonication, which would have exposed more of the protein to the solubilization.Most (67.5 g/100g) of the protein in Mucuna is watersoluble (Adebowale et al., 2007).Greater CP and WSC losses in finer particles were because of the greater surface area exposure.This is in agreement with Myhrman (2002) and Teixeira et al. (2003) who reported CP losses of 11% and up to 50%, respectively due to leaching after soaking of finely ground bean samples in acid or alkaline solutions.Losses of CP from ACD and ALK-treated fine particles were also facilitated by the solubility of Mucuna protein Crude protein losses due to increased solubility at low and high pH (Adebowale et al., 2007) are consistent with our findings of CP loss with the ACD and ALK treatments.Other than discoloration, no precipitation, haze or clogging of filter pores was evident observed in the beans.
Sonication may have decreased the amino acid concentration of the beans because its' cell rupturing effect may have increased their solubility.The contrasting increases in concentrations of amino acids in ACD-and ALK-treated beans are probably evidence of increased proteolysis and peptidolysis by these treatments.
The increased starch concentration of fine particles due to ACD, ALK or SON treatments agrees with responses to acid or alkali extraction of Mucuna reported by Siddhuraju and Becker (2005).The increased starch concentration was due to partial loss of soluble components including WSC, protein, and L-Dopa, all of which decreased with solvent extraction relative to CON.The reduction in concentration of these components and the energy value of the bean, and the concomitant increases in starch and fiber concentration imply that solvent extraction and sonication modified the nutritive composition of the bean and resulted in losses of key components.Nevertheless, except for SON, the extraction methods did not reduce the concentration of amino acids in the beans.The pH of the solvent is an important factor that can affect the success of L-Dopa removal from Mucuna and the residual nutritional quality.Several authors mentioned that due to formation of melanin, Mucuna beans are darker after acid or alkali extraction (Teixeira et al., 2003;Wanjekeche et al., 2003).
Melanin is a metabolite of L-Dopa characterized by its dark color and conversion of L-Dopa into melanin is most evident at alkaline pH (Teixeira et al., 2003;Wanjekeche et al., 2003).The darker color of the alkali versus acid extracts in this study (Figure 2) agrees with these observations.Diallo et al. (2002) successfully reduced the L-Dopa concentration to 0.001 g/100g after 48 hours of soaking cracked Mucuna beans in calcium hydroxide solution but noted the remarkably dark coloration of the treated beans.Beans cooked in acid solutions are lighter in color than beans cooked in alkaline solutions (Wanjekeche et al., 2003).Adebowale et al. (2007) noted that darker colors occurred in sodium hydroxide solutions of pH 11 relative to less alkaline solutions.The association of such dark colorations with melanin is should be investigated because the effects of melanin on health are controversial (Ziyang et al., 2018).Melanin has been anecdotally associated with the formation of melanoma in some studies (Dollery, 1999;Letellier et al., 1999;Siple et al., 2000) but not others (Weiner et al., 1993;Pfutzner and Przybilla, 1997;Fiala et al., 2002).The darker color of the solvent-extracted bean in this study highlights the need for further investigation of concentrations of melanin residues in the acid-or alkali-treated bean.The effect of these treatments on Mucuna's food or feed safety (Huisden et al., 2018) and nutraceutical properties (Rai et al., 2017;Tavares et al., 2017;Mohamed et al., 2018;), also require further investigation; especially since it has been suggested that L-Dopa plays a role in its ethnopharmacological effects (Huisden et al., 2010).

CONCLUSIONS
Particle size affected the efficacy of L-Dopa removal in the solvent extracts.Both acidic and alkaline solvents performed equally well at detoxifying fine (1 mm) particles of Mucuna bean to safe levels (<0.4 g/100g L-Dopa) but also reduced their WSC and CP concentrations and increased their starch and aNDF concentrations.However, these methods were not effective at detoxifying coarse (6 mm) Mucuna particles and they had less consistent effects on their nutritive value.Acidic and alkaline solvent extraction darkened the bean, suggesting that they increased the formation of melanin, a metabolite of L-Dopa characterized by its dark color.Future research should determine melanin concentrations in acid-or alkaliextracted beans.
Sonication did not cause discoloration of Mucuna, yet it was also an effective method of detoxifying fine but not coarse particles of Mucuna to safe levels (<0.4 g/100g L-Dopa).Sonication generally resulted in similar modifications to the nutritive value of the bean as acid or alkali solvent extraction but caused greater losses of EE from fine particles and unlike the other treatments, it decreased the concentrations of the amino acids.

Figure 1 .
Figure 1.Concentration of L-Dopa after the various treatments; (P < 0.05) and error bars denote standard error.

Figure 2 .
Figure 2. Color changes after detoxification of Mucuna bean through A) Alkaline extraction at 1-mm particle size, B) Alkaline extraction at 6-mm particle size, C) Acid extraction at 1-mm particle size, D) Acid extraction at 6-mm particle size.

Table 1 .
Effect of processing method on the chemical composition of fine (1 mm) and coarse (6 mm) Mucuna beans.
a-e Within a row, means without a common superscript letter differ (P<0.05);A untreated beans; B acid-treated beans; C alkali-treated beans; D sonicated beans; WSC = water-soluble carbohydrate; aNDF = neutral detergent fiber.

Table 2 .
Effect of processing method on the amino acid composition of finely ground (1-mm) Mucuna beans, DM basis.