Supporting Information
Rapid and sensitive analysis of 27 underivatized free amino acids, dipeptides and tripeptides in fruits of Siraitia grosvenorii Swingle using HILIC-UHPLC-QTRAPÒ/MS2 combined with chemometrics methods
Guisheng Zhou × Mengyue Wang × Yang Li × Ying Peng × Xiaobo Li ()
G. Zhou × M. Wang × Y. Li × Y. Peng × X. Li ()
School of Pharmacy,
Shanghai Jiao Tong University,
Shanghai 200240, China
e-mail:
The stability of FAAs and small peptides
The temperature stability of amino acids and small peptides were very important in this study. On the one hand, temperature was an important storage factor of amino acids and small peptides; on the other hand, temperature was a parameter of ultrasound-assisted extraction. The stability of analytes in 50% (v/v) acetonitrile were evaluated by analyzing high (100 ng/mL) and low (10 ng/mL) concentrations of 27 mixture standards (n = 6) exposed to different conditions (room temperature 25°C, 12 h; 4°C, 72 h). From the results of analysis, the analytes were considered stable in 4°C and 25°C because the response of stored samples and fresh samples, or the measured analyte concentration and its corresponding theoretical value, differed by less than 5% (Table S2). The thermal stability of amino acids was reported in many papers (Bada et al. 1995; Duke et al. 1994; Yan et al. 2009). From the previous reports, we could obtain a conclusion that the investigated amino acids in this study were stable at relatively low temperature conditions (4°C and 25°C).
In some previous thermodynamic studies it had been shown that the addition of cosolutes such as electrolytes, surfactants, or other biomolecules to aqueous small peptide and amino acids solutions would have a strong effect on the hydration of these solutes (Pałecz et al. 2010; Yan et al. 2009), and consequently invoking important changes in their ability to bind other molecules (Singh et al. 2015; Yan et al. 2009). Temperature might change this processes of hydration (Duke et al. 1994; Singh et al. 2015; Yan et al. 2009). Based on the previously reported, a mixture of amino acids was relatively stable and had litter changes of thermodynamic parameters in the low temperature (less than 328.15 k). Besides, in this study, 27 investigated compounds were dissolved in 50% (v/v) acetonitrile without additives (electrolyte, non-electrolytes, surfactants, etc.), and their thermal stability was also studied at relatively low temperature conditions (4°C and 25°C). Hence, a mixture of the target amino acids was stable in our research condition.
Optimization and verification procedures
In this study, to optimize UAE parameters, a screening design of PB was built to identify the main factors affecting the responses (the total content of 24 FAAs and 3 active small peptides from LHG) among 6 variables (such as ultrasonic power, frequency, extraction time, times, temperature, and the solid–liquid ratio). The screening experiments indicated that ultrasonic power, extraction time and solid–liquid ratio were the most effective parameters to the yield of target compounds. In preliminary experiment, the effects of ultrasonic power, extraction time and solid–liquid ratio on UAE were respectively studied to observe the yields of total contents of FAAs (24 amino acids) and peptides (3 active small peptides) from LHG. On the one hand, we paid attention on the total content of peptides using different ultrasonic treatments; on the other hand, we also added the relevant experiments to explore the relationship between the ultrasonic treatments and each peptide cleavage.
The PB screening experiments indicated that ultrasonic power, time and solid–liquid ratio were the most effective ultrasonic treatments to the yield of target compounds. In this study, the above 3 most effective ultrasonic treatments were used to further investigate the relationship between the ultrasonic treatments and each target peptide cleavage.
Firstly, mixture standards containing 100 ng/mL each active small peptide (GSH, Ala-Gln and Cyst) were used to investigate the relationship between the ultrasonic power and each target peptide cleavage. Generally speaking, ultrasonic power was a fundamental parameter in the process of UAE. In this study, five different ultrasonic power (100, 175, 250, 325 and 400 W) were selected to evaluate the influence of ultrasonic power on the cleavage of each small peptide. The other ultrasonic treatments such as ultrasonic time (43 min) and ultrasonic temperature (25°C) were fixed. The results obtained by the HILIC-UHPLC–QTRAPÒ/MS2 analysis in Fig. S2(A) indicated that the detected concentrations of GSH, Ala-Gln and Cyst did not significantly change with the increase of ultrasonic power from 100 to 325 W. However, a slight decrease of the concentrations of GSH and Cyst when the ultrasound power was increased from 325 to 400 W. Therefore, the variable of ultrasonic power will not be a factor to degrade GSH and Cyst when ultrasonic power was less than 325 W. In our study, the optimum ultrasonic power 280 W, which could not degrade the 3 target peptides, was selected for the whole UAE extraction.
Secondly, the mixture standards containing 100 ng/mL each active small peptide (GSH, Ala-Gln and Cyst) were also used to investigate the relationship between the ultrasonic time and each target peptide cleavage. The detected concentration of each peptide over different ultrasonic time from 30 to 50 min is shown in Fig.S2(B), when the other factors were as follows: ultrasonic power 280 W and ultrasonic temperature 25°C. The results indicated that the detected concentrations of three target peptides began to slowly decline when ultrasonic time was 45 min, which began to decrease due to the degradation of peptides over longer ultrasonic time. From the Fig.S2(B), three target peptides were relatively stable at the selected optimum ultrasonic time 43 min in this study.
Thirdly, to investigate the influence of different concentrations on the peptides cleavage, the five different concentrations of mixture standards (50, 100, 150, 200 and 300 ng/mL) were used in this study when the ultrasonic time (43 min), power (280 W) and ultrasonic temperature (25°C) were fixed. As shown in Fig. S2(C), there was no significant difference between the detected concentration and the corresponding theoretical concentration of each peptide. Therefore, the variable of concentration will not be a decisive factor when investigating peptides cleavage in the stated conditions.
In conclusion, the obvious cleavage of peptides did not present using the selected ultrasonic treatments (such as ultrasonic time 43 min and power 280 W) in this study.
Fig. 1A–C and 1a–c are the 3D surface plots and planar contour plots between every two independent variables on the basis of Eq. (5). Fig. 1A and Fig. 1a show the effects of ultrasonic power and extraction time on the yield of total content of FAAs and small peptides (Y(Tc)). When the ultrasonic power fixed, Y(Tc) increased with the increase of extraction time until reaching a maximum and then decreased. Similarly, ultrasonic power caused an initial increase and then decrease in the Y(Tc). This result indicated that both ultrasonic power and extraction time were important variables for FAAs and small peptides extraction from LHG. Fig. 1B and Fig. 1b show the effects of ultrasonic power and the solid–liquid ratio on the Y(Tc). When the solid–liquid ratio was fixed, the Y(Tc) rapidly increased with the increase of ultrasonic power until reaching a maximum and then slowly decreased. However, the solid–liquid ratio had less of an effect on changing the content of the Y(Tc). From Fig. 1C and Fig. 1c, it could be seen that the effect of the solid–liquid ratio on the extraction rate of the Y(Tc) was not very obvious at a given value of the extraction time as the surface was relatively flat. When the solid–liquid ratio was at a certain value, the extraction rate of the Y(Tc) also increased and then decreased. The extraction of the Y(Tc) depended largely on the ultrasonic power and extraction time.
The maximum yield of Y(Tc) was calculated as 5682.64 mg/g in the following optimum UAE conditions: ultrasonic power of 279.54 W, extraction time of 42.83 min and the solid–liquid ratio of 302.15 mL/g.
A desirability function test was performed using an optimizer procedure in Design–Expert 8.5 software. This approach consisted in first converting each response variable into a desirability function di, that varied from 0 to 1 (Wu and Hamada 2011). The idea was that this desirability function acts as a penalty function that leaded the algorithm to regions where we could find the desired response variable values. The factor levels that taken to a maximum or a minimum of the response variable were called “optimum points”. Eq. (1) expressed the global desirability function, D, defined as the geometric mean of the individual desirability functions. The algorithm should search for response variable values where D tended to 1 (Pourfarzad et al. 2014).
D = (d1 ´ d2 ´ ... ´ dn)1/n……………………………………………………………(1)
where d1, d2...dn were responses and n was the total number of responses in the measure.
The numerical optimization found a point that maximizes the desirability function using the professional statistical software of Design–Expert 8.5. Verification of the model was carried out by T-test using the statistical software of SPSS 16.0 to compare the mean actual values of the responses with the predicted value.
According to the results calculated from the desirability function, the maximum D value of 0.896 was provided when the ultrasonic power was 279.54 W, extraction time 42.83 min and the solid–liquid ratio 302.15 mL/g. The optimum yield of total of FAAs and small peptides was calculated as 5682.64 mg/g in the above optimum condition. However, considering the operability in actual production, the optimal conditions could be modified as follows: ultrasound power of 280 W, extraction time 43 min and the solid–liquid ratio 302 mL/g. Under the modified conditions, the actual maximum yields of target compounds were detected as 5590.3 ± 76.2 mg/g (n = 3), respectively. Thus, the predicted extraction condition was similar to the experimental value.
Effect of gradient elution
In this study, 27 target compounds covered a wide range of polarities. Simultaneous chromatographic separation of them was not commonly realized using isocratic elution under the HILIC column. Therefore, the gradient elution was used to separate target compounds. In the literature, buffer type and salt concentration usually affected the HILIC separation (Cai et al. 2009). Ammonium acetate was the ideal salt because it provided the best results in selectivity and reproducibility, presented excellent solubility and was highly volatile, making it suitable for eventual further MS analysis (Zhou et al. 2014). In this study, different mobile phase additives (buffer and/or pH modifying agent) were used to improve HILIC separation. As a result, a mixed solution including A (water, 10 mmol/L ammonium acetate and 0.5% acetic acid) and B (acetonitrile, 1 mmol/L ammonium acetate and 0.1% acetic acid) was chosen as the optimized mobile phase using gradient elution. Factually, it was very important that the ionic strength was different and this could promote the overestimation or underestimation in the late eluting compounds when the buffer was changed during the chromatographic gradient. This problem will be comprehensively researched in our further study. The above mentioned effect was also considered in our presented study. To solve this problem, the general analytical method validation was researched in this study. Validation of the method strictly complied with the ICH regulations for confirmation analysis procedure. Several performance parameters were studied, including matrix effects, linearity, LOD, LOQ, precision, repeatability, stability and recovery. These results of method validation suggested the established method could provide sufficient accuracy for the quantification of LHG samples. The MS response was a stable value for each target compound with the given suitable concentration at a specific retention time point during the chromatographic gradient, and this response could be used for the quantification of each target compound from LHG based on the method validation. Therefore, the presented HILIC-UHPLC-QTRAP/MS2 using gradient elution was considered as a practical method for this study. Additionally, to analysis amino acids, mobile phase containing ammonium acetate using gradient elution (The detector was MS) was also reported in many published literatures (Guo et al. 2013; Zhou et al. 2013).
Retention time deviations in HILIC column
The literatures reported that retention time deviations were also a problem in HILIC analysis (Neville et al. 2012). The intra-day retention time deviation was assessed by injecting 27 standard compounds six times during the day, while the inter-day retention time deviation was assessed by injecting samples for three consecutive days. Table S4 shows the data for both inter- and intra-day experiments (n = 6). There was a small change, 0.0066 ± 0.0036 min, in the average retention times of 27 target compounds, and there was also a small change (0.0041 ± 0.0015 min) in the time difference between the inter- and intra-day experiments. The inter- and intra-day experiments in this study were conducted by referring to the related literature (Neville et al. 2012), which could provide evidence in support of the robustness of retention times with the development HILIC method.
Optimization of ESI modes
According to the previous reports, while most of amino acids could be monitored under both ESI+ and ESI− modes, the stronger response was observed under the ESI+ mode than ESI− mode and the ESI+ mode was often selected for the next study (Liu et al. 2013). In some papers, the MS response of Asn in some matrixes was highest in positive ion mode but signal-to-noise ratio was improved when ionisation was done in negative ion mode (Nielsen et al. 2006; Zhang et al. 2011). However, based on the other papers, positive ion mode, which was selected to detect Asn, had the higher MS response and better signal-to-noise ratio than negative ion mode in their reported matrixes or instruments (Buiarelli et al. 2013; Liu et al. 2013; Guo et al. 2013; Zhou et al. 2013). From these reports, matrixes, instruments and/or other factors might be important parameters to select suitable ion mode for detecting different compounds. For these reasons, to optimize the QTRAPÒ/MS2 conditions, Q1 full scans were conducted under both ESI− and ESI+ modes and the comprehensive information of each analyte was obtained operating in the above two ion modes for our present matrixes and instruments. The results revealed that 27 target compounds had higher sensitivity and preferable signal-to-noise ratios in the ESI+, which made it easier to detect analytes of lower content in LHG, and easier to confirm molecular ions or quasi-molecular ions in the identification of each peak. Thus, the ESI+ mode was selected in the following studies.