| Reduced glutathione(GSH)is a tripeptide composed of y-glutamate,cysteine and glycine,which has a wide range of clinical applications including preservation and restoration of liver function,pigmentation prevention,side effects of radiation and chemotherapy,hypoxemia,early senile cataract,corneal inflammation,allergies,adjuvant treatment for intoxication,etc.Recently it has been found that GSH possesses a broad market since the changes of GSH content and GSH/GSSG ratio are closely related to many diseases.Nowadays,the most common method at home and abroad of GSH industrial production is yeast fermentation.The critical quality attributes of GSH in the production process mainly include:GSH content,moisture content and related substances such as acyl glycine,cysteine,GSSG,L-pancreaticacyl-4-nitroaniline,etc.Among them,GSH content in the culture,moisture of dried GSH and GSSG content in GSH are three critical quality attributes that can affect the quality of GSH products.The fermentation termination and GSSG content are mainly detected by HPLC method,and moisture content can be determined via oven drying method.Those above detection methods have become the bottleneck of rapid release for GSH product quality in terms of complex procedures and long operation time.Therefore,it is fairly meaningful to draw near infrared spectroscopy(NIRS)into industrial production to achieve quality control for GSH.In this study,we used NIRS to realize rapid and nondestructive analysis of critical quality attributes for GSH industial production,and established analytical models of critical quality attributes for GSH.Our work mainly includes rapid detections of GSH in the fermentation process,moisture in GSH products and GSSG in GSH products.Study contents were as follows:(1)NIRS application in GSH rapid detection in the fermentation processZymotic fluid was collected every 5 h during fermentation process of GSH,and 6 batches were collected at each time point,and 6 samples in each batch.Four batches of GSH samples were separated as the calibration,while the other two batches were set apart as the validation.Transmission mode and diffuse mode were applied respectively to establish GSH content detection models for centrifugated zymotic fluid or original zymotic fluid.At the same time,GSH contents in zymotic fluid were determined by HPLC as the reference.The models of GSH rapid detection were obtained via exploration of spectral preprocessing methods and selection for spectra ranges in fermentation production.Optimal preprocessing methods were FD-SG(7,2)smoothing and SD-Norris(51,5)smoothing,and Rc,Rcv,Rp,RMSEC,RMSECV,RMSEP of the models were 0.9957,0.9787,0.9372,0.0497 g/L,0.111 g/L,0.119 g/L and 0.9464,0.9110,0.9082,0.166 g/L,0.212 g/L,0.279 g/L respectively.Preliminary study showed that NIRS was expected to replace HPLC for the rapid detection of terminal GSH fermentation content.NIR-TS model was superior to NIR-DRS model for broken bacteria centrifugal supernant detection and original zymotic fluid detection.Subsequently,accessories and conditions of sampling will be further optimized.(2)NIRS application in moisture rapid detection of GSH productsIn this study,NIR-DRS was used to establish and evaluate rapid moisture detection models of GSH products.Freeze-drying method was applied to make 44 GSH samples containing different moisture and collected NIRS.KS classification method was used to separate 29 samples as the calibration and 15 samples as the validation.Moisture of GSH samples determined by infrared moisture analyzer was used as a reference.NIR-DRS model was obtained for moisture rapid detection of GSH products via exploration of spectral preprocessing methods and selection for spectra ranges.Optimal preprocessing method was FD-SG(13,2)smoothing-Autoscale,and Rc,Rcv,Rp,RMSEC,RMSECV,RMSEP of the models were 0.9798,0.9638,0.9072,0.1113%,0.1490%,0.1254%respectively.The results showed that the model with the good predictive ability could be used in the rapid detection of moisture in GSH samples.(3)NIRS application in GSSG rapid detection of GSH productsIn this study,NIR-DRS was used to establish and evaluate rapid GSSG detection models of GSH products.Thirty-three GSH samples containing different special ratios of GSSG were prepared,and NIRS results were collected.Gradient classification method was used to separate 23 samples as the calibration and 10 samples as the validation.Optimal preprocessing method was FD-SG(15,2)smoothing-Autoscale,iPLS interval was set as 30 to acquire proper spectra ranges,and Rc,Rcv,Rp,RMSEC,RMSECV,RMSEP of the models were 0.9958,0.9734,0.9333,0.0459%,0.1192%,0.1832%respectively.The results showed that the model with the good predictive ability could be used in the rapid detection of GSSG in GSH samples.The innovation points were listed below:(1)It was the first time to apply NIRS to the rapid and nondestructive analysis of GSH in fermentation process,which was expected to replace HPLC to analyze terminal GSH fermentation contents.(2)It was the first time to put forward the thought to realize GSH online monitoring and timely intervention by NIRS in fermentation process(3)It was the first time to apply NIRS to the rapid and nondestructive analysis of key quality parameters such as moisture and GSSG to achieve rapid quality control of GSH products.(4)A variety of methods for model optimization were carried on,which served as useful reference for complex system models. |
