Quantitative determination of α s1 –casein in goat ’ s milk using reversed - phase high performance liquid chromatography (RP–HPLC)

The chemical, technological and allergy properties of goat ’ s milk are significantly affected by the level of α s1 - casein. Despite indications that more than 90% of cow ’ s milk protein allergy (CMPA) children reacted similarly to goat's milk in typical IgE - mediated forms, the availability of goat's milk without α s1 - casein appeared to be a potential alternative for patients with cow's milk allergy. In Malaysia, there is a lack of quantitative data on α s1 - casein in raw and commercial goat ’ s milk. To address this problem, this study aimed to quantify α s1 - casein in goat ’ s milk using reversed - phase high performance liquid chromatography (RP–HPLC). The procedure was developed using a total of twenty samples comprised of commercial and raw goat ’ s milk. For the calibration curve, a linear relationship with R² > 0.999 was obtained using cow ’ s α - casein standard and peak areas were observed over the concentration range, with a limit of detection (LOD) of 0.025 mg/ mL and limit of quantification (LOQ) of 0.075 mg/mL, respectively. The amount of α s1 - casein in the twenty samples ranged from 0.08±0.01 to 1.45±0.17 mg/mL. Three samples which were claimed as goat ’ s milk infant formula were not found to contain α s1 - casein. Repeatability and reproducibility were satisfactory for both retention times and peak areas. The RSD of peak areas ranged from 0.01 - 8.07% within an analytical day and from 0.61 - 1.77% across analytical days. This study will contribute as a reference method for analysis and surveillance data of α s1 - casein, in private or national laboratories, as well as milk and milk product manufacturers.


Introduction
Goat's milk industry is growing at a significant rate in the last decade and is considered an important agricultural commodity.From 1961 to 2017, there is a remarkable rising in global goat's milk production by over 27.9% in Asia alone (Miller and Lu, 2019).Despite the lack of officially recorded data in Malaysia about goat's milk production and consumption, in 2013, there were 8,195 heads of dairy goats in Peninsular Malaysia with 50% located in Johor (Shahudin et al., 2018).The significant growth in demand for goat's milk among the public is due to its health benefits contributed by the unique features of goat's milk compared to milk from other ruminants (Miller and Lu, 2019).
The casein profile of goat milk differs from that of other ruminants' milk, which is one of its distinguishing characteristics (Lima et al., 2017).This mainly draws down to the concentration of α s1 -casein, a casein subunit encoded by the CSN1S1 gene in goat's milk (Mangia et al., 2019).Twenty-three alleles of CSN1S1 gene were classified into strong, intermediate, weak, and null alleles, producing four expression levels of α s1 -casein, i.e., 3.5, 1.1-1.8,0.45-0.6 and 0 g/L, respectively (Mangia et al., 2019).Previous studies have shown that a small presence of α s1 -casein in goat's milk contributed to lower allergenicity as compared to cow's milk (Bellioni-Businco et al., 1999;Bevilacqua et al., 2001;Lara-Villoslada et al., 2005;Ah-Leung et al., 2006;Ballabio et al., 2011;Hodgkinson et al., 2012).From these findings, goat's milk can serve as protein source for hypoallergenic formulas (Clark and Mora-García, 2017).The high variation of α s1 -casein in milk is not limited to allergenicity.It was also found to significantly impact the cheese-making process.Talach (2013) reported that a higher concentration of α s1 -casein caused lower pH of milk, which provided firmer curd and shorter coagulation process.Furthermore, it is crucial to have the data on the  (Caroli et al., 2009).
Due to the variation in α s1 -casein concentration, which can vary up to 10-folds, the need for an analytical tool with robust, sensitive, and good working range to measure α s1 -casein in goat's milk is vital.Reversedphase high performance liquid chromatography (RP-HPLC) allows the quantification of the milk proteins (whey and casein) within a single run and without laborious sample preparation (Ostertag et al., 2021).Hence, this study aimed to determine and quantify α s1casein in goat's milk using RP-HPLC.The data from this study would be valuable for many parties including health practitioners, consumers, and others with an interest in health outcomes especially in Malaysia as a scientific reference.

Sampling
Sample collection was carried out according to method ISO 707:2008/IDF 50:2008(ISO, 2008).A total of twenty samples were used for the analysis.Details of the samples are shown in Table 1.Commercial goat's milk refers to goat's milk purchased from commercial online or physical stores that have been pasteurized prior to commercialization.Raw goat's milk refers to goat's milk purchased directly from the goat's farm right after the milking session (without pasteurization).For liquid commercial and raw goat's milk, the samples were transferred to the laboratory in a cooler bag with a temperature of 2-4°C.Upon arrival at the lab, the samples were aliquoted at 500 µL and kept frozen at -20°C until further analysis.Powdered samples were kept at room temperature until further analysis.

Materials and reagents
HPLC-grade acetonitrile, BisTris buffer, dithiothreitol (DTT), guanidine hydrochloride (GdnHCl), sodium citrate, and trifluoroacetic acid (TFA) were purchased from Fisher Scientific (Waltham, MA, USA).All other chemicals were of analytical grade.Alphacasein from cow's milk standard was purchased from Sigma Aldrich (St. Louis, MO, USA).All buffers and solutions were prepared using ultrapure water (18.2MΩcm) from the ELGA water purification system (High Wycombe, UK).

Milk samples preparation
With modifications, milk samples were prepared according to Bobe et al. (1998) and Montalbano et al. (2014).A solution containing 0.1 M BisTris buffer (pH 6.8), 6 M GdnHCl, 5.37 mM sodium citrate, and 19.5 mM DTT (pH 7) was added directly to frozen aliquots milk sample in a 1:1 ratio (v:v) and allowed to thaw at room temperature (28°C).After thawing, each sample was shaken for 10 s, incubated for 1 hr at room temperature, and centrifuged for 10 mins at 8 000×g.The supernatant was diluted at a ratio of 1:3 (v:v) with a solution containing 4.5 M GdnHCl and solvent A. All samples were filtered through a 0.22 µm nylon filter.

Construction of calibration curve
Since there was no commercial standard available for α s1 -casein from goat's milk, lyophilized commercial α-casein cow's standard was used as the standard.The standard solution was solubilized in a solution containing 4.5 M GndHCl and solvent A at 0.025, 0.05, 0.10, 0.50, 1.00, and 5.00 mg/mL, filtered and further used for the construction of the calibration curve.The calibration curve was used for protein quantification, where the limit of detection (LOD) and the limit of quantitation (LOQ) were determined by 3:1 and 10:1 signal to noise (S/N), respectively (Shrivastava and Gupta, 2011).

HPLC analysis
The HPLC system consisted of an Agilent 1200 Series chromatography (Santa Clara, CA, USA) equipped with a quaternary pump (Agilent 1200 Series, G1322A).Two mobile phases were used in the gradient elution system, whereby mobile phase A consisted of 10% acetonitrile and 0.1% TFA (v/v) (pH 2), and pH was  TFA (v/v).α s1 -casein were detected at 204 nm variable wavelength with a diode array detector (Agilent 1200 Series, G1311A).The equipment was controlled by Agilent ChemStation software.Separation was performed on a C8 reversedphase analytical column (Zorbax 300SB-C8, Agilent Technologies) with a silica-based packing (3.5 µm, 300Ă, 4.6 × 150 mm).The sample was injected via an auto-sampler (Agilent 1200 series, G1329A) using an injection loop of 100 µL and an injection volume of 5 µL.Samples were chromatographed in a gradient elution system, according to Bonfatti et al. (2008) for 45 mins at a flow rate of 0.5 mL/min.

Precision
Repeatability and reproducibility were carried out to determine the precision of the method.The repeatability was established by three consecutive extractions of each sample within a day.The reproducibility, known as dayto-day repeatability, was determined by analyzing each sample on different days.

Statistical analysis
A calibration curve was established using Microsoft Excel Version 1908/2019.The statistical analyses were performed by one-way analysis of variance (ANOVA) using Minitab 18.0 Statistical Software (2017) (State College, PA, USA).Means were compared using the least significant difference (LSD) test at p<0.05.

Sample separation and extraction
Since the commercial standard of α s1 -casein was not available as a single protein, this study utilized cow's αcasein standard, which was composed of α s1 -and α s2caseins.A clear separation of α s1 -casein and cow's α s2casein in standard is shown in Figure 1(a).The α s2 -casein showed multiple peaks due to the phosphorylated form of α s2 -casein (Bordin et al., 2001).At 0.10 mg/mL of αcasein, the peaks of α s2 -casein were non-identified, which supported the basis of a 4:1 proportion of α s2casein in milk (Campbell and Marshall, 2016).Figure 1 (b) shows the calibration curve of peak area versus concentration of α s1 -casein with R² > 0.99.The values of peak area were derived from the chromatographic profiles of the cow's α-casein standard at different concentrations (Figure 1a).(2014).In this study, no peak was observed after the sample was stored for 1 hr after the addition of that solution which could be contributed to the instability of the milk proteins.To address this, the extraction of α s1casein where the addition of 4.5 M GdnHCl and Solvent A solution in the samples was done freshly before separating the HPLC column.This modification resulted in an improved resolution of α-casein.
The method was modified because casein exists in milk protein as large, colloidal particles and due to the aggregation nature of the protein, it is necessary to fully dissociate and denature the casein before being subjected to HPLC (Gaspard and Brodkorb, 2019).Pre-treating samples with denaturing or reducing agents can disrupt the non-covalent interactions and disulfide bonds.In this study, there were few reducing agents used, i.e., guanidine hydrochloride and DTT.However, the amphipathic nature of casein molecules has made casein displays a strong tendency to associate (Melnikova et al., 2019).This can cause a formation of a gel-like structure in α-casein which was shown to be the reversible action.
During the gel formation, disulfide bonds were not involved.There is only minimal effect on the intermolecular interaction between α-casein molecules when disulfide bonds were disrupted as it only changes the local electrostatics (Melnikova et al., 2019).Thus, this shows α-casein tends to self-association.Therefore, as storage time increased, α-casein could not be detected, which might be due to the casein being self-associated to form a micelle.A similar condition was reported by Dumpler et al. (2017), in which the milk protein was unstable especially when whey protein was dissolved in guanidine buffer and left at room temperature for several hours.This was due to the incomplete denaturation of whey proteins or subsequent refolding at lower guanidine concentrations (Dumpler et al., 2017).

HPLC analysis
As shown in Figure 2, retention times of α s1 -casein from commercial goat's milk of liquid type (sample A), raw goat's milk of liquid type (sample T), and commercial cow's milk of liquid type (sample S) were confirmed with the retention times of the cow's α-casein standard.All peaks were set at a fixed time,±0.47 min (calculation not showed) based on the cow's α-casein standard.
In Figure 2, samples A, S, and T had similar retention time with cow's α-casein standard.This result was supported by Montalbano et al. (2016) who reported that in homozygous condition, it was possible to identify and observe not only major peaks but some minor peaks in the goat's milk chromatogram.Therefore, cow's standards can be used to identify casein in goat's milk.An in-depth insight into α-casein structure was elaborated by Ingham et al. (2018), in which goat's milk had a strong similarity in internal structure with cow's milk according to the resonant soft X-ray scattering (RSoXS) and small-angle X-ray scattering (SAXS).In our study, the detection wavelength was set at 204 nm because it showed the most significant peak during the screening of standard α -casein using a UV/Vis spectrophotometer.

Quantitative analysis of α s1 -casein
Table 2 shows the mean of retention time, area, and calculated concentration of α s1 -casein in the twenty samples of goat's milk.The concentrations ranged from 0.08±0.01mg/mL to 1.45±0.17mg/mL.In contrast, a total of six samples showed a concentration of α s1 -casein lower than the LOQ, while α s1 -casein was not detected in three samples.There was a significant variation in the concentration of α s1 -casein in all goat's milk samples obtained in this study.Based on the theory discussed earlier, different concentrations of α s1 -casein could be correlated with genetic polymorphisms, however, the genotypes were not analyzed in this study (Mangia et al., 2019).A clear vision of this can be observed in samples B and C. Both samples were a mixture of goat's milk from Saanen and Toggenburg breeds but derived from different farms.However, both samples showed significant differences (p<0.05) in terms of concentration of α s1 -casein.This can be explained further by Mohsin et al. (2019), where variation in chemical composition can occur even when similar breed and environmental settings were applied.Other factors that can influence the chemical composition of milk are breed, individuality, stage of lactation, diet, parity, feeding regime, management practices, environment, locality, health and nutritional status of the animal (Sonu and Basavaprabhu, 2020).
The α s1 -casein in samples P, Q, and R was considered as not detected because their chromatographic profiles showed no visible peak of α s1casein (Figure 3).All three samples were commercial goat's milk powder and from the labelling, they were intended for infant formula.Furthermore, the manufacturer for sample R had claimed that their sample was absent from α s1 -casein.Although these samples might be derived from goat's milk with a null variant of α s1 -casein, however, to postulate that genetic polymorphism affects the quantification of α s1 -casein in these samples was inconclusive.As referred by Geiselhart et al. (2021), processing technologies have been applied to prevent and eliminate milk allergies.This correlated to the three samples, in dried powder form and the processing of milk was involved.A few processing technologies which were proven to mitigate allergenicity include thermal processing (pasteurization, sterilization, ultrahigh temperature, spray drying) and nonthermal processing (high pressure, homogenization, ultrasonic, enzymatic hydrolysis, fermentation) (Geiselhart et al., 2021).Due to a lack of information on the type of processing used for the samples as well as the genetic polymorphism of the goat's breed, it is difficult to determine the factors which influenced the concentration of α s1 -casein in these samples.Sample P, Q, and R were powdered goat's milk that has been claimed as infant formula, in which the target market is children who could not tolerate cow's milk.However, substituting cow's milk with goat's milk for children with cow's milk allergy should be avoided without any investigation by the paediatrician as proteins in both types of milk share high homology attributes, which could elicit similar types of allergic reactions.Cross-reactivity of goat's milk in children was investigated by Bellioni-Businco et al. (1999), who reported that 24 of the 26 children tested reacted to goat's milk, however, a 5-fold concentration of goat's milk is needed to induce a reaction.Goh et al. (2019) also found that it was unusual to have patients with isolated goat's milk allergy without cow's milk allergy.This was further supported by Pham and Wang (2017), who reported that goat's milk and cow's milk had a high protein identity with more than 84%.Hence, crossreactivity between both milk is expected even though goat's milk may have low levels of α s1 -casein.Repeatability and reproducibility of HPLC analysis were assessed to determine the method precision (Montalbano et al., 2014).The RSD values for the retention times and peak areas are given in Table 2. Montalbano et al. (2014) reported that RSD values for retention times were below 0.22% (repeatability) and 0.60% (reproducibility), while RSD for peak areas were below 0.77% (repeatability) and 5.00% (reproducibility).Based on Table 3, the RSD values for retention times were below 0.96% (repeatability) and 1.77% (reproducibility).The RSD values for peak areas were below 8.07% (repeatability) and 15.01%(reproducibility).The RSD values for both retention times and peak areas showed a higher percentage than Montalbano et al. (2014).Pre-column (guard column) conditions might have affected this variation, which was not used in our study.The significant effect of guard column application in the HPLC system was reported by Scott (1992).Nevertheless, injecting 1.00 mg/mL of cow's α-casein standards in a two-days interval assisted in the improvement of reproducibility and quantitative precision.A blank injection was also used between each sample to have a better reproducibility of quantification as described in Montalbano et al. (2014).According to Cox et al. (2011) and European Medicines Agency (2012), RSD values less than 20% were considered satisfactory for HPLC analysis.Therefore, it can be concluded that the method used in this provides good precision to determine α s1 -casein in goat's milk.

Conclusion
In the present work, we have quantified α s1 -casein in 20 samples of goat's milk with a quantification range from 0.08±0.01 to 1.45±0.17mg/mL.Three samples of commercial goat's milk infant formula were not found to contain any α s1 -casein.Repeatability and reproducibility were satisfactory for both retention times and peak areas with RSD < 20%.HPLC analysis showed good sensitivity, repeatability, and reproducibility for the detection and quantification of α s1 -casein in goat's milk.This method can be applied in further studies to quantify α s1 -casein in other types of milk with considerations of precaution steps highlighted in this study.Overall, the data on the level of α s1 -casein in goat's milk from this study can be used as a scientific reference and would be beneficial for many including testing laboratories, health practitioners and milk product manufacturers.More importantly, the general consumer can have a better understanding and awareness of the incidence and level of α s1 -casein in raw and commercial goat's milk for better management of allergenicity.
For extraction of α s1 -casein from goat's milk, the duration of storage after solubilization of the sample with a solution containing 4.5 M GdnHCl and Solvent A was not discussed inBobe et al. (1998) andMontalbano et al.

Table 1
. Details of the 20 goat's milk samples used in the study Sample A to S were obtained from commercial sources while sample T was raw milk obtained from a farm in Selangor FULL PAPER adjusted with 0.1 N of HCl.Mobile phase B consisted of 90% of acetonitrile and 1%

Table 2 .
Retention time, peak area, and calculated concentration of α s1 -casein in goat's milk samples Values are presented as mean±SD of triplicates (n = 3).Values with different superscript within the same column are significantly different (p<0.05).nd: not detected, LOQ: Limit of quantification.Figure 3. Peak of α s1 -casein was not shown in samples P, Q, and R within the range of 20.31±0.47 mins https://doi.org/10.26656/fr.2017.7(2).364© 2022 The Authors.Published by Rynnye Lyan Resources FULL PAPER

Table 3 .
Precision of retention times and peak areas of α s1 -casein from cow's α-casein standard a Three aliquots of three consecutive extraction in each sample (n = 9) b Three aliquots of three consecutive extraction in each sample in different days.Concentration of 0.5 mg/mL was used as a daily indicator for reproducibility