Untargeted metabolite profiling on the water-soluble metabolites of edible bird’s nest through liquid chromatography-mass spectrometry

Background and Aim: Edible bird’s nest (EBN) is the nutrient-rich salivary bioproduct produced by swiftlets in Southeast Asia. Currently, researchers are exploring the therapeutic effects of EBN, such as cell growth promotion, antioxidant content, antiviral effects, bone strengthening, eyes care, and neuroprotection bioactivities. The therapeutic effects of EBN have been studied through different extraction methods but the metabolites profile of the EBN in each extract has not yet been elucidated. This study aimed to profile the water-soluble metabolites of EBN prepared in different extraction methods. Subsequently, an extraction method will be selected as an ideal extraction method for untargeted metabolite profiling on the water-soluble metabolites in EBN. Materials and Methods: In this study, water-soluble metabolites of EBN extracted by the four extraction methods were subjected to metabolite profiling through liquid chromatography-mass spectrometry (LC-MS). The extraction methods were acid extraction(ABN), pancreatic extraction (EzBN), eHMG extraction, and spray drying of HMG extraction (pHMG). The metabolite profiles, such as the number of metabolites and their identities in each extraction method, were evaluated through LC-MS analysis. Results: The identity of metabolites present in the four extraction methods is inconsistent. Based on LC-MS analysis, only one and six metabolites were extracted differently through EzBN and ABN, respectively, in the first pre-screening. Through the second LC-MS screening on pHMG and eHMG extraction methods, eHMG was selected as an ideal extraction method due to the highest numbers of water-soluble metabolites with an amount of 193 was detected. Besides, eHMG extraction method was able to extract sialic acid and a high percentage of secondary metabolites. Conclusion: This study suggests that eHMG is the ideal extraction method for extracting higher number of water-soluble metabolites from EBN and could be further developed as an extraction method for industry application. In addition, this study also has identified the types of primary and secondary metabolites present in EBN.


Introduction
Edible bird's nest (EBN) is a well-known bioproduct made from the saliva secretion of swiftlet, specifically from the two genera of Aerodramus and Collocalia. The swiftlet from the two genera is mostly habitat in Southeast Asia [1,2]. The main constituents of EBN are proteins, carbohydrates, lipids, and a group of minerals such as calcium, sodium, potassium, magnesium, phosphorus, iron, zinc, copper, chromium, and selenium [2][3][4][5]. EBN has been regarded as traditional Chinese medicine by the practitioners in Qing dynasty due to its recuperative properties [1,6]. The recuperative properties of EBN are highlighted with the effect of boosting immune system, treating malnutrition, improving metabolism, enhancing skin complexion and alleviating asthma, helping in phlegm clearance, relieving cough, nourishing children, libido raising, enhancing renal function, recovery from illness and surgery, as well as improving concentration [7]. Recently, EBN is further demonstrated for its properties on suppressing the virus, inflammation and oxidative stress, strengthening bone, eye caring, and neuroprotective properties [8][9][10][11][12][13][14]. On the other hand, Roh et al. [15] and Kong et al. [16] have reported the proliferative effects of EBN on human adipose-derived stem cells and normal human fibroblasts with the presence of epidermal growth factor-like activity. In summary, EBN acts as a dual function bioproduct with both its nutritional and therapeutic values.
To study the constituents of EBN and its therapeutic effects, the development of an ideal extraction Available at www.veterinaryworld.org/Vol.13/February-2020/12.pdf methodology of EBN is very important. Several extraction methodologies were developed and used for studying the bioactivities of EBN. The study by Guo et al. [9] documented strong inhibition of influenza viruses by EBN extract that is pre-treated with pancreatin. Besides, Abidin et al. [11] also reported that the EBN extract prepared by eHMG extraction method successfully stimulated and enhanced the proliferation of corneal keratocytes in wound healing without altering their functionality. Chua et al. [17] prepared EBN extracts by the water extraction method (HMG). These extracts exhibited strong chondroprotective effects on osteoarthritis (OA). In addition, Aswir and Wan Nazaimoon [18] have documented acid-extracted EBN exhibited an anti-inflammation effect by significantly reducing the production of the inflammatory protein, tumor necrosis factor-alpha. In view of all the works, it is observed that different EBN extract obtained through different extraction methods showed different therapeutic effects. One possible explanation is because the extraction of an active component is highly dependent on the extraction method employed. Thus far, the identity of the metabolites in each of these extractions has not yet been further studied for the underlying mechanism of actions for their therapeutic effects. Hence, future study could be carried out to confirm the therapeutic effects of the metabolites.
Metabolite profiling is a powerful scientific tool for a complete investigation of a group of small molecules. This approach often used in analyzing biological components for the identification of potential biomarkers for certain diseases [19]. Recently, metabolite profiling has gained fame in food classification [20,21]; this is due to its untargeted analysis approach with the potential to cover the whole or the maximum metabolomics molecular information of foods. One of the examples of using the metabolite profiling approach on EBN has successfully demonstrated in the study done by Chua et al. [22]. The metabolites of the EBN were extracted through the chloroform/methanol solvent extraction, which was then successfully identified through gas chromatography-mass spectrometry (MS) and liquid chromatography-MS (LC-MS) techniques.
Since water is commonly used to prepare EBN essence for consumption and the metabolites of EBN are not fully established yet, this study aimed to preliminary profile the water-soluble metabolites of EBN prepared in different extraction methods. Subsequently, an extraction method will be selected as an ideal extraction method for untargeted metabolite profiling on the water-soluble metabolites in EBN.

Ethical approval
The study did not involve any live animals, so no ethical approval was required.

Chemicals
LC-MS grade formic acid and acetonitrile were purchased from Fisher Scientific (Waltham, MA, USA). Deionized water was obtained from a Barnstead GenPure water purification system (Thermo Fisher Scientific Inc., Waltham, MA, USA).

Sample collection, preparation, and extraction
Raw unclean EBNs samples were collected collectively from different swiftlet premises located in Johor, Malaysia. The feathers and impurities were manually removed with forceps, and the raw unclean EBN was ground with mortar and pestle. Ground EBN was sieved through a 0.4 mm wire mesh to further separate the smaller pieces of feathers and impurities. The unclean EBN powder was then placed in an air force oven at 50-55°C overnight to reduce the moisture content.
There were four extraction methods selected for the comparison in this study, namely, eHMG, pHMG, ABN, and EzBN extraction methods. The raw unclean EBN was extracted with the proprietary methods of eHMG [11] and pHMG (the spray-dried of HMG extract) [17] that were innovated and standardized by School of Chemical and Energy Engineering in Universiti Teknologi Malaysia (UTM). These methods were modified based on the methods presented by Oda et al. [23] and Goh et al. [24]. Besides, another acid extraction (ABN) and pancreatin extraction (EzBN) were developed by the team of Universiti Tunku Abdul Rahman (UTAR) in 2016 [25] with some modification from the methods presented by Aswir and Wan Nazaimoon [18] and Goh et al. [9].

eHMG and pHMG
Due to the proprietary issue on these two extraction methods, the details of these two methods were unable to be described in this report.

Acid extraction (ABN)
The EBN powder was suspended in deionized water at 0.2% (w/v) and left for 24 h. The mixture was then boiled at 80°C with 2% (v/v) of 0.4 M sulfuric acid for 4 h. The extract was allowed to cool down and centrifuged at 2716 g (5000 rpm) for 15 min. The pH of the supernatant collected was neutralized to pH 7.0. The white precipitated formed was removed through centrifugation with 2716 g (5000 rpm) for 15 min at 4°C. The supernatant was collected and kept at 4°C for further analysis.

Pancreatin extraction (EzBN)
The EBN powder was suspended in deionized water at 0.2% (w/v) and left for 24 h. The EBN mixture was boiled at 100°C for 30 min. An amount of 1 ml of 0.5 mg/ml pancreatin was added into EBN mixture and was allowed for the reaction at 45°C for 4 h with pH 8.5-9.0. The enzyme was inactivated by heating at 90°C for 10 min. The supernatant was collected after centrifugation at 2716 g (5000 rpm) for 15 min. The extract was kept at 4°C.
Before subjecting the extracts to LC-MS analysis, all the four extracts were centrifuged at 9660 g Available at www.veterinaryworld.org/Vol.13/February-2020/12.pdf (12,000 rpm) for 10 min and the supernatant of the extracts was filtered through 0.2 µm polytetrafluoroethylene membranes.

Quadrupole time-of-flight (QTOF) LC-MS analysis
The four EBN extracts were qualitatively analyzed using Agilent 6560 Ion Mobility QTOF (IM-QTOF) LC-MS system that coupled with the Agilent 1290 ultra-high-performance liquid chromatography (Agilent Technologies, USA). The metabolites present in the EBN extracts were separated through POROSHELL 120 EC-C18 (4.6×100 mm; 2.7 μ; Agilent Technologies, USA) chromatographic column with the mobile phase that consisted of (A) 0.1% formic acid in water and (B) 0.1% formic acid in acetonitrile. All the four EBN extracts were undergone the first pre-screening evaluation with the elution of 5-95% B (0.0-1. The other setting parameters for IM-TOF analysis remained the same throughout the analysis process. The injection volume was 1 µl and the column temperature was maintained at 40°C. The acquisition of the metabolites was performed in positive (ES+) mode. The mass spectra were recorded over an m/z range from 100 to 1000. Deionized water was used as the background blank. Whereas, the operating conditions of the mass spectrometer were set as follows: Capillary voltage of 4000 V, nozzle voltage of 500 V, and fragmentor voltage of 365 V were maintained. Nebulizer pressure (N 2 ) was kept at 20 psi, drying gas temperature was maintained at 225°C. Drying gas flow was 13 L/min and sheath gas flow was 12 L/min at 400°C.

Data mining and metabolites identification
The metabolite features from the acquired MS spectral raw data were extracted with the untargeted molecular feature extraction algorithm in Agilent MassHunter Workstation -Qualitative Analysis software B.07.00 (Agilent Technologies, USA). The algorithm filtered off the peak height with 100 counts to avoid the noise spectral picking, as well as the mass of internal reference ions with 121.0967 and 922.1389. Then, the algorithm locates the covariant ions in the chromatogram and grouped them as a single metabolite feature using the information of mass, isotopic distribution with common organic elements (C, H, O, N, P, Cl, F, and S), charge-state and adducts of sodium, potassium, and ammonium. The extracted metabolite features were characterized by retention time (RT) and intensity.
The identity of the extracted metabolite features was searched against METLIN Personal Metabolite Database in the MassHunter software based on the accurate mass and RT (optional). The mass and RT tolerance of the compound identity matching was restricted to ±5 ppm and ±0.1 min (optional), respectively. The accuracy of the identity of each metabolite was calculated as a score. The metabolites list of each extract was retained if the identity of the metabolite fulfilled the threshold score of 80, and the error of database matching was less than ±5 ppm.

Results and Discussion
The efficiency of EBN extraction methods The method of extraction is a crucial process that maximizes the extraction of the bioactive metabolites from EBN. To search for an ideal extraction method for the untargeted metabolite profiling of EBN, four different extraction methods with the therapeutic effects were assessed and evaluated. For example, pancreatin extraction with antiviral effect as reported by Guo et al. [9]; eHMG extraction with the effect of enhancing proliferation of corneal keratocytes by Abidin et al. [11]; HMG extraction showed chondroprotective effect on OA as documented by Chua et al. [17]; and finally the acid extraction with anti-inflammation bioactivities reported by Aswir and Wan Nazaimoon [18]. The approach of LC-MS is recognized with its high sensitivity, accuracy, and reproducibility [26][27][28]; thus, there was no technical replicate done in this untargeted metabolite profiling analysis.
The number of detected metabolites in each of the extraction method was analyzed by MassHunter software. Nearly 37-67% out of the total metabolites from the four different extracts were putatively identified by matching with the METLIN metabolites database. The complete information of all the identified metabolites in each extraction method is detailed in Table-1. The identities of the extracted metabolites are unique among the four different extracts, suggesting that there is no single extraction method that could extract all types of metabolites due to the differences in natural physicochemical properties of the metabolites [29][30][31][32].
Based on the mobile phase for compound separation in the first screening evaluation, there were significant differences in the number of extracted metabolites under each extraction method (Table-2a). The highest total number of metabolites obtained was from pHMG extract and followed by eHMG extract. The total number of metabolites detected in both of pHMG and eHMG extracts was greater than EzBN and ABN extracts, with approximately 20-30 times and 4-5 times, respectively. However, the LC-MS separation for each extract was not well defined by referring to the chromatograms obtained ( Figure-1). Therefore, the second screening evaluation was carried out with an improved LC-MS mobile phase.
Available at www.veterinaryworld.org/Vol.13/February-2020/12.pdf  Both of the eHMG and pHMG extraction methods were selected to undergo the second screening evaluation since they showed greater efficacy in extracting the higher number of metabolites from EBN in the first screening evaluation.
The second screening evaluation with an optimized LC-MS mobile phase for separating compounds has greatly improved the elution efficacy and increased the number of analyzed metabolites ( Figure-2a and b). The good separation in the liquid chromatography has broadened the range of eluted metabolites. Hence, the second screening evaluation has provided a better comparison between the eHMG and pHMG extraction methods. The eHMG extraction method has successfully recovered a significant number in total extracted metabolites as compared with pHMG (Table-2b). There were 193 metabolites detected from eHMG extraction method (Table-2b), which are more than 26 non-polar metabolites detected in the study done by Chua et al. [22]. Therefore, the eHMG extraction method was selected as the ideal extraction method because it provided the maximal recovery of the number of water-soluble metabolites present in EBN.

The metabolite profile of extraction methods
In the second screening evaluation, there were approximately more than half out of the total metabolites (60.39% and 57.14% of metabolites, respectively) from eHMG and pHMG extracts that were putatively identified. The information of the retained metabolites for both eHMG and pHMG extraction methods in the second screening evaluation are shown in Tables-3  and 4, respectively. Based on the comparison between eHMG and pHMG extraction methods in the second screening evaluation, 24 out of the total identified metabolites were found to be similar in each extract (Figure-2c). The result indicated that the eHMG extraction method not only extracted a greater number of metabolites but also there were approximately 57.14% of the metabolites from pHMG extraction method which were found to be similar to eHMG. The identities of the metabolites that found to be similar in both of the extraction methods are marked in Tables-3  and 4.
Sialic acid is known as the key component of EBN because it is served as the unique quantitative marker for grading the EBN. In this study, sialic acid was identified in the eHMG extraction method with the identity of 2,7-Anhydro-alpha-N-acetylneuraminic acid (Table-3  detected of sialic acid in eHMG extract has further convinced that eHMG extraction method is more suitable as the ideal extraction method. The type of metabolites present in eHMG and pHMG extracts (from the second screening) was further categorized into five groups based on the      macronutrient classification (Figure-2d). The five groups of macronutrients are comprised oligosaccharides, peptides, lipids, nucleosides, and secondary metabolites. There were 192 and 42 metabolites identified from eHMG and pHMG extracts (Tables-3  and 4), respectively. The differences in the type of metabolites between eHMG and pHMG extracts have further supported the preference of the type of metabolites toward each extraction method. Among the macronutrients, eHMG extraction method can extract mostly secondary metabolites, followed by peptides, oligosaccharides, lipids, and nucleosides (Figure-2d). The primary metabolites obtained from this study support the finding from the previous proximate analysis of EBN, which protein is the highest composition followed by carbohydrates and lipids [2,36,37]. The presence of secondary metabolites could most probably explain the recuperative and therapeutic effects of EBN. The secondary metabolite with the identity of O 2 -vinyl 1-(pyrrolidin-1-yl)diazen-1-ium-1,2-diolate (V-PYRRO/nitric oxide [NO]) was found in eHMG extract (Table-3). This secondary metabolite acts as NO donor and delivers NO specifically after metabolism by cytochrome P450 in hepatocytes without affecting the NO-sensitive tissues as well as systolic blood pressure [38]. The in vivo study done by Li et al. [39] showed that V-PYRRO/NO is able to protect the hindrance to renal congestion and lipid peroxidation from acetaminophen-induced nephrotoxicity in mice. In addition, V-PYRRO/NO can protect against high-fat diet (HFD)-induced liver steatosis and insulin resistance without affecting the mitochondria biogenesis [40]. Interestingly, Zhang et al. [41] showed that EBN could prevent HFD-induced insulin resistance by regulating the transcriptional changes in insulin signaling genes. Hence, the presence of V-PYRRO/NO in EBN may explain the protective effect of EBN against the HFDinduced damages. In short, from this study, it is believed that the study on secondary metabolites profiling in EBN in the future is crucial and not to be neglected.
A polysaccharide with an identity of chondroitin was identified from the first screening of eHMG extract Available at www.veterinaryworld.org/Vol.13/February-2020/12.pdf  [42]. Chondroitin is a glycosaminoglycan that acts as a chondroprotective agent for the treatment of OA. OA is the lesion of articular cartilage caused by trauma. Since chondroitin is an essential proteoglycan in cartilage, it acts on OA by stimulates the cartilage repair through enhancing the production of the extracellular matrix of cartilage. Besides, chondroitin helps to maintain the viscosity of the synovial fluid to lubricate the joint and therefore reducing the pain of the patient. Furthermore, chondroitin suppresses the inflammatory cytokines such as interleukin-1β that induce the release of matrix metalloproteinases and Available at www.veterinaryworld.org/Vol.13/February-2020/12.pdf aggrecanases which cause the degradation of the cartilage [43,44]. In an in vitro study done by Chua et al. on the effects of EBN to OA [17], the authors reported that EBN can protect articular cartilage from further deterioration by reducing inflammation and enzymatic lesions process and enhancing the cartilage formation simultaneously. Therefore, the effects of EBN on OA might be contributed by chondroitin.

Conclusion
There was no single extraction method could provide optimal conditions in extracting all the metabolites from EBN. Therefore, complementary extraction methods should be used in parallel when broader metabolite profiles are required. eHMG extraction method was selected as the ideal extraction method for untargeted profiling the type of polar metabolites in EBN. This is because the number and the type of metabolites detected are the highest in eHMG extracts among the four evaluated extraction methods. Furthermore, the presence of key metabolites of sialic acid has further defined the suitability of eHMG extraction method. Therefore, the findings in this study could offer great potential for enhancement in the industrial EBN extraction process and hence improve the overall EBN yield and bioactivities. Nevertheless, the validation of the structure elucidation and functional assays of interesting metabolites shall be carried out in the future.

Authors' Contributions
YML conceived the study design. SRT conducted all the designed experiments, data processing, and analysis. THL contributed to the sample collection and performed the in-house extraction method (eHMG and pHMG) for the study. SRT prepared the manuscript with critical feedback from the coauthors. THL, SKC, and YML supervised the study and provided input and advice in the project. All authors have read and approved the final manuscript.