Development of Isocratic RP-HPLC Method for Separation and Quantification of L-Citrulline and L-Arginine in Watermelons

Watermelons (Citrullus lanatus) are known to have sufficient amino acid content. In this study, watermelons grown and consumed in Malaysia were investigated for their amino acid content, L-citrulline and L-arginine, by the isocratic RP-HPLC method. Flesh and rind watermelons were juiced, and freeze-dried samples were used for separation and quantification of L-citrulline and L-arginine. Three different mobile phases, 0.7% H3P04, 0.1% H3P04, and 0.7% H3P04 : ACN (90 : 10), were tested on two different columns using Zorbax Eclipse XDB-C18 and Gemini C18 with a flow rate of 0.5 mL/min and a detection wavelength at 195 nm. Efficient separation with reproducible resolution of L-citrulline and L-arginine was achieved using 0.1% H3P04 on the Gemini C18 column. The method was validated and good linearity of L-citrulline and L-arginine was obtained with R2 = 0.9956, y = 0.1664x + 2.4142 and R2 = 0.9912, y = 0.4100x + 3.4850, respectively. L-citrulline content showed the highest concentration in red watermelon of flesh and rind juice extract (43.81 mg/g and 45.02 mg/g), whereas L-arginine concentration was lower than L-citrulline, ranging from 3.39 to 11.14 mg/g. The isocratic RP-HPLC method with 0.1% H3P04 on the Gemini C18 column proved to be efficient for separation and quantification of L-citrulline and L-arginine in watermelons.


Introduction
Citrullus lanatus (Thunb.) Matsum. and Nakai, commonly known as watermelon, is a nonseasonal fruit which is cultivated abundantly in Malaysia and other tropical regions [1]. It belongs to the Cucurbitaceae plant family, which originated from the African Kalahari Desert [1]. Watermelons have high content of phytonutrients and are rich in dietary antioxidants such as carotenoids (lycopene and -carotene), polyphenolics, ascorbic acid, and significant amino acids [2]. Watermelons are usually consumed by juicing the flesh, beneficial in the prevention and improvement of health problems, such as cardiovascular diseases, erectile dysfunction, hypertension, and cancers [3]. Figueroa et al. [4] demonstrated that watermelon juice supplementation improves aortic hemodynamics by reducing the reflected wave amplitude in prehypertensive individuals. A study by Poduri et al. [5] reported that watermelon attenuated hypercholesterolemiainduced atherosclerosis in mice. Commercial watermelon juices provide enormous marketing potential and nutritious drinks for individuals to maintain a healthy lifestyle.
Amino acids, particularly L-citrulline and L-arginine, are regarded as major types of phytonutrients present in watermelons which may contribute to their reputed and diversified health benefits [6]. L-citrulline, C 6 H 13 N 3 O 3 (IUPAC name: 2-amino-5-(carbamoylamino)pentanoic acid) (Figure 1), is a nonessential amino acid firstly identified from watermelon, Citrullus vulgaris Schrad. [7,8]. L-citrulline is a physiological endogenous amino acid to most living systems involved in protein metabolism and removal of excess metabolic ammonia [9]. It serves as a precursor for L-arginine and product of nitric oxide (NO) cycle [10]. L-arginine, C 6 H 14 N 4 O 2 (IUPAC name: (S)-2-amino-5-guanidinopentanoic acid) (Figure 2), is a semiessential and free form physiological amino acid that functions as one of 20 building block proteins for biological processes such as cell division, ammonia removal, wound healing, and hormone release [7,8]. Wu et al. [11] demonstrated that supplementation of L-citrulline and Larginine from watermelon juice improved serum levels of NO metabolites and aortic endothelial-mediated vasodilation in diabetic rats. L-citrulline and L-arginine are present in all parts of watermelon fruits including flesh, rind, and seed [7]. A study done by Rimando and Perkins-Veazie [12] reported that the rind of red watermelon and yellow watermelon contains more L-citrulline at a concentration ranging from 15.6 to 29.4 mg/g than flesh, 7.9-28.5 mg/g. Similar to the above finding, Jayaprakasha et al. [13] reported that rinds of Citrullus vulgaris varieties such as petite treat and jamboree watermelon and also yellow crimson watermelon contained slightly higher L-citrulline ranging from 13.95 to 28.46 mg/g than flesh, 11.25-16.73 mg/g. These findings suggested that watermelon rind has an abundance of L-citrulline content in comparison to its content in flesh.
Analyses of L-citrulline and L-arginine were routinely conducted using capillary electrophoresis and quantification by a spectrophotometric method; however, the method is less sensitive, leading to discrepancies in the outcomes [6]. L-citrulline and L-arginine are polar, nonvolatile, and devoid of chromophores; thus analysis by reverse-phase high performance liquid chromatography (RP-HPLC) commonly employed a derivatization method using pre-or postcolumn derivatization [14][15][16][17]. Jayaprakasha et al. [13] stated that precolumn derivatization such as orthophthalaldehyde (OPA), naphthalene-2,3-dicarboxaldehyde, or 4-dimethylaminoazobenzene-4 -sulfonyl chloride (dabsyl chloride) was able to provide accurate and stable chromatography baseline, but the reactions were unstable and affected by the sample matrix [18]. Postcolumn derivatization by ninhydrin is tedious due to long analysis time up to 72 hours and instability of derivatization reagents that may cause poor compound recovery [19]. Analysis of underivatized L-citrulline and Larginine is warranted for rapid and effective quantification of these compounds. Given that no amino acids content of Lcitrulline and L-arginine in Malaysia watermelons has been reported so far, we have developed an isocratic RP-HPLC method for separation and quantification of L-citrulline and L-arginine in watermelons.

Sample Preparation. Citrullus lanatus (Thunb.) Matsum.
& Nakai of red watermelon and yellow crimson watermelon was obtained from Selangor Fruit Valley, Selangor. Seeds were removed manually and the edible part was cut into cubes. Watermelon flesh and rind were juiced and frozen at −80 ∘ C for at least 2 days. The frozen juices were put in a freeze-drier (Labconco, USA) for 4 days until completely dried. The dried juice powders were kept at −20 ∘ C. For the analysis, samples were prepared in the form of juice extract and methanol extract. Juice extract was prepared directly by dissolving the dried juice powder in dH 2 O. For methanol extract, a known quantity of dried juice powders was extracted with 30 mL of MeOH and 1 mL of 1 N HCl, vortexed, and sonicated for 30 minutes. The samples were macerated by cold maceration for a period of 72 hours in an orbital shaker. Methanol extracts were then filtered using Whatman filter paper. The residues were reextracted twice using fresh solvent and the three methanol extracts were pooled. The obtained methanol extracts were evaporated to dryness using a rotary vacuum evaporator at 60 ∘ C and stored at 4 ∘ C until analysis.  , and Gemini C 18 , 250 × 4.6 mm, 110Å, 3 m (Phenomenex, Torrance, CA)] were used. The analysis proceeded for quantification of both compounds, L-citrulline and L-arginine in watermelons juice extracts and methanol extracts using the chosen column and mobile phase: Gemini C 18 eluted by 0.1% H 3 P0 4 with a flow rate of 0.5 mL/min at 195 nm. Chromeleon software was used for quantification of L-citrulline and L-arginine. The concentration of L-citrulline and L-arginine content was quantified based on the linear curve of standards. The content of compounds was expressed as milligrams per gram (mg/g) of sample extracts.

Method Validation.
The validation of the isocratic RP-HPLC method was performed for linearity of calibration curve, limit of detection (LOD), limit of quantification (LOQ), accuracy, and precision. The linearity of the isocratic RP-HPLC method for quantification of compounds was constructed using the concentration range of 0.1-1000 g/mL for L-citrulline and 0.1-500 g/mL for L-arginine. The regression equation was calculated in the form of = + , where is the concentration and is the peak area of compounds. Linearity was established by the coefficient of determination ( 2 ). LOD and LOQ were measured based on signal-tonoise ratio (S/N) method. LOD is the lowest concentration of analyte that can be detected with signal-to-noise ratio of 3 : 1 and LOQ is the lowest concentration that can be quantified with acceptable precision and accuracy with signal-to-noise ratio of 10 : 1. S/N of 3 is considered acceptable for LOD, while LOQ is established at S/N of 10. Precision of the method was determined as percentage relative standard deviation (%RSD) of peak area of intraday and interday analysis data. Intraday (three times in a day operation under the same conditions) and interday (three different days) studies were performed at three different concentrations (Level 1: 20 g/mL; Level 2: 60 g/mL; Level 3: 150 g/mL). The resulting peak area was used to calculate SD and the relative standard deviation (%RSD). Accuracy of the method by recovery study was done by adding a known amount of reference standard solution (three concentrations) to test samples. The spiked extract solutions were injected three times, and the recovery was calculated with the value of detected versus added amounts.

Separation of L-Citrulline and L-Arginine by Isocratic RP-HPLC Method.
The initial isocratic RP-HPLC method for separation of mixed standard, L-citrulline, and L-arginine was performed using selected mobile phases according to previous literatures with slight modifications [8,13,20]. Interaction between mobile phase and stationary phases in isocratic RP-HPLC is important for the determination of solutes' retention time [21].
In this study, separation for determination of mixed standard, L-citrulline, and L-arginine was performed using a hydrophilic anionic ion-pairing reagent with different concentrations of phosphoric acid (H 3 P0 4 ) or addition of acetonitrile (ACN) as mobile phases: 0.7% H 3 P0 4 , 0.1% H 3 P0 4 , and 0.7% H 3 P0 4 : ACN (90 : 10). The mobile phase at the concentration of 0.7% H 3 P0 4 : ACN (90 : 10) resulted in L-citrulline and L-arginine were unretained and coeluted ( value close to 0) as shown in Figure 3(a). The mixture of 0.7% H 3 P0 4 : ACN (90 : 10) is highly hydrophilic, leading to rapid elution of L-citrulline and L-arginine with poor separation. Peaks of L-citrulline and L-arginine were slightly retained and partially separated using 0.7% H 3 P0 4 (Figure 3(b)). However, optimum resolution was not achieved by 0.7% H 3 P0 4 as value between L-citrulline and L-arginine is close to 1. The mobile phase of 0.1% H 3 P0 4 resulted in efficient separation with reproducible peaks of L-citrulline and Larginine although all chromatograms showed stable baseline (Figure 3(c)). This finding is in agreement with Fekete et al. [22] who noted that 0.1% H 3 P0 4 acts as a good separation agent by increasing the polarity and improving the retention time of zwitterionic molecules including amino acids. Dolan [23] supported the notion that 0.1% H 3 P0 4 adequately provides reasonable buffering for amino acids separation by RP-HPLC. This showed that a concentration less than 1.0% H 3 P0 4 as mobile phase provides efficient separation of amino acids, peptides, or proteins as demonstrated by Shibue et al. [24]. Thus, the mobile phase 0.1% H 3 P0 4 is proven to provide efficient separation and the best resolution of mixed standard, L-citrulline, and L-arginine.
The study also evaluated separation of mixed standard in two different columns, Zorbax Eclipse XDB-C 18 and Gemini C 18 using 0.1% H 3 P0 4 . Zorbax Eclipse XDB-C 18 did not provide good separation and resolution of L-citrulline and L-arginine as shown in Figure 4(a). A study by Barber and Joseph [25] showed that polar compounds were less separated and not well resolved using Zorbax Eclipse XDB-C 18 with a longer analysis time of 54 minutes. Efficient separation and resolution of L-citrulline and L-arginine from  mixed standard were achieved using Gemini C 18 as shown in Figure 4(b). L-citrulline and L-arginine are eluted at a short retention time with L-arginine, 4.773 min, followed by L-citrulline at 5.787 min ( Figure 5). Efficient separation of Lcitrulline with a retention time of about 4 min was achieved on the Gemini C 18 column due to the high degree similarity of column with polar compounds [13]. Gemini C 18 is a new generation hybrid column end-capped with porous silica as base core and polymer media coated on top of the silica core which exhibit silica-like mechanical properties of base material while similarly decreasing the number of residual silanols [26]. This result demonstrated that Gemini C 18 is the most suited column for efficient separation of mixed standard, L-citrulline and L-arginine.
International Journal of Analytical Chemistry The result from chromatography separation of Lcitrulline and L-arginine shown in Figure 5(b) demonstrated that the reverse-phase (RP) mode provided efficient separation and substantial retention achieved on both polar compounds without the need for derivatization. RP mode can efficiently be applied in this study although Brown et al. [27] proposed the use of hydrophilic interaction liquid chromatography (HILIC) mode after cyano-and pentafluorophenylpropyl stationary phases failed to retain target compounds including L-citrulline and L-arginine. HILIC mode is primarily used when separation of very polar compounds is needed or for incomplete chromatographic separation in RP mode [28]. However, HILIC mode required an expensive and robust system equipped with tandem mass spectrometric (MS) detection for L-citrulline and L-arginine separation in high biological matrix samples such as serum and plasma [16,27]. Complete chromatographic separation of L-citrulline and L-arginine by RP mode in this study eliminates the use of HILIC mode.

Validation of Isocratic RP-HPLC Method.
Method validation of isocratic RP-HPLC method was performed by determination of linearity, LOD, LOQ, recovery, and intraday and interday analysis. The results of linearity, LOD, and LOQ were summarized in Table 2. Good linear regression equations of L-citrulline and L-arginine standard are displayed between corresponding peak areas versus concentrations of compounds based on the correlation coefficients ( 2 = 0.9956, = 0.1664 + 2.4142 and 2 = 0.9912, = 0.4100 + 3.4850, resp.). The LOD and LOQ for L-citrulline were 0.42 g/mL and 1.28 g/mL while those for L-arginine were 0.88 g/mL and 2.66 g/mL, which demonstrated that isocratic RP-HPLC method was efficiently sensitive. The method had good accuracy that showed efficient recoveries for both 6 International Journal of Analytical Chemistry    (Table 3). The RSD (%) for intraday and interday precision ranged from 0.37% to 1.09% for L-citrulline and 0.33% to 2.05% for Larginine, in which both RSD ≤ 2% (Table 4). These validation results confirmed that the isocratic RP-HPLC method is precise, accurate, and sensitive for simultaneous quantification of L-citrulline and L-arginine.

Quantification of L-Citrulline and L-Arginine Contents in
Two Different Watermelon Extracts. Consumption of watermelon extracts rich in L-citrulline and L-arginine is proven to be beneficial for diseases prevention. Thus, a rapid, reliable, and efficient isocratic RP-HPLC method is essential for simultaneous quantification of these amino acids in juice extracts and methanol extracts. The chromatographic profiles of both extracts in red watermelon and yellow crimson watermelon are presented in Figure 6. Quantification of Lcitrulline and L-arginine was performed using Chromeleon software. The content was calculated based on the calibration curve of L-citrulline and L-arginine standard achieved with good correlation coefficients and linear regression equations, 2 = 0.9956, = 0.1664 + 2.4142 and 2 = 0.9912, = 0.4100 + 3.4850, respectively. The results are tabulated in Table 5.
Red watermelon juice extract showed slightly high yield of L-citrulline in rind, 45.02 mg/g compared to flesh, 43.81 mg/g. Similar trends were shown in L-citrulline content in rind and flesh of yellow crimson juice extract, 16.61 mg/g and 15.77 mg/g, respectively. This finding is in accordance with the study by Jayaprakasha et al. [13] which found that rinds from C. vulgaris watermelon varieties of petite treat and jamboree watermelon and yellow crimson watermelon contained significantly high L-citrulline, 13.95 mg/g, 20.84 mg/g, and 28.46 mg/g, respectively, compared to flesh, 11.25 mg/g, 16.73 mg/g, and 14.74 mg/g, respectively, using RP-HPLC method. L-arginine content in red watermelon juice extract was higher in flesh, 11.10 mg/g, compared to rind, 3.39 mg/g. L-arginine content was approximately 3fold lower than L-citrulline in red watermelon flesh juice extract. Consumption of watermelon flesh juice extract aided in efficient conversion of significantly high L-citrulline, a potent endogenous precursor to L-arginine in the kidney, which resulted in increased plasma L-arginine concentration. Findings by Collins et al. [29] proved that plasma L-arginine concentration increased by 95.2 ± 3.5 M and 108.0 ± 4.1 M compared to normal plasma baseline, 86.4 ± 3.5 M after 3 weeks of consumption of 780 mL (∼1 g L-citrulline/day) to 1560 mL (∼2 g L-citrulline/day) of watermelon juices. Recently, a study by Bailey et al. [30] supported the notion of the increased plasma L-arginine concentration by 116 ± 9 M compared to placebo, 67 ± 13 M after 2 weeks of 300 mL/day watermelon juice consumption, which contains International Journal of Analytical Chemistry  ∼3.4 g L-citrulline/day. L-citrulline content in crude flesh and rind extract of red watermelon and yellow crimson watermelon varied in the range of 13.91-24.99 mg/g, while L-arginine content was in the range of 4.08-8.41 mg/g. Lcitrulline and L-arginine content is much lower in methanol extracts compared to juice extracts. Fish and Bruton [17] stated that methanol extracts may diminish the solubility of amino acids, thus marked reduction in amino acids yield. The quantitative results confirmed that watermelon juice extracts most effectively quantified higher yield of L-citrulline and Larginine, and this study's outcomes may possibly suggest that juice extraction method is best in optimizing amino acids yield.

Conclusion
The isocratic RP-HPLC method has been successfully developed for separation and quantification of L-citrulline and L-arginine content in both watermelon extracts of flesh and rind using the selected mobile phase (0.1% H 3 P0 4 ) in Gemini C 18 . The established isocratic RP-HPLC method provides evidence that L-citrulline and L-arginine are best retained using Gemini C 18 column. The validated method is robust, sensitive, accurate, and precise with good linearity ( 2 ≥ 0.99), low values of LOD and LOQ, recoveries within 98.88%-103.41%, and %RSD precision less than 2%. Juice extract effectively yielded higher L-citrulline and Larginine content by juice extraction method; thus it is potentially used for quantitative amino acids analysis. The present study procedure may provide a basis for separation and quantification of L-citrulline and L-arginine in local watermelons. The high content of L-citrulline and L-arginine suggested watermelons as a good source of nutraceutical and health benefits ingredients. However, further researches are necessary to explore biological activities such as aphrodisiac properties of these active constituents in watermelons to support their potential in human diet and prevention of health related diseases.

Conflicts of Interest
The authors declare that there are no conflicts of interest that would prejudice the impartiality of this scientific work.