Data for: Bioprocess optimization of nutritional parameters for enhanced anti-leukemic L-asparaginase production by Aspergillus candidus UCCM 00117: a sequential statistical approach

Published: 9 November 2020| Version 3 | DOI: 10.17632/bdhc5xcb6h.3
Maurice Ekpenyong,


This dataset describes the sequence of experiments conducted to optimize nutrients required to formulate a fermentation medium for production of a glutaminase-near-free L-asparaginase by a strain of Aspergillus candidus. It also presents data for effects of temperature, pH and metal ions on L-asparaginase activity and stability, as well as the therapeutic anti-cancer potential of the enzyme. Research hypotheses: Medium composition significantly influences L-asparaginase activity. Significant differences exist among levels of temperature, pH and metal ion effects on enzyme activity and stability. Findings: One-factor-at-a-time (OFAT) approach was valuable to select the major variables for L-asparaginase activity. The raw data as well as the 95% confidence interval statistic of the experiments are presented as Tables DT1 to DT4 and Figs DF1 to DF4 for carbon and nitrogen sources, spore density and metal ions respectively. Plackett-Burman design (PBD) helped to identify significant predictors towards enhanced L-asparaginase activity. Table DT5 presents the 2-level factorial design matrix for the PBD screening. Table DT6 is the analysis of variance (ANOVA) for full regression model of PBD analysis. Table DT7 is the coefficient table from which significant terms were extracted to build the first-order model. The path of steepest ascent (PSA) set up experiments to move the levels of significant variables close to optimum using coefficients of significant variables from first-order model. The raw data and 95% confidence interval statistic are presented as Table DT8 and Fig DF5. Factor levels for experiment 7 which produced maximum L-asparaginase activity served as center points for response surface methodology (RSM). Table DT9 is the design matrix for central composite design (CRD) of the RSM. Tables DT10 and DT11 are the ANOVA for full quadratic models for biomass concentration (Y1) and total protein (Y2) respectively, while Table DT12 is predictor coefficient summary with p-values. The diagnostic plots to test model adequacy are presented as Figs DF6a-c (Y1), Figs DF10a-c (Y2) and Figs DF14a-c (Y3). The surface (3-D) plots of significant two-way interactions of factors are presented as Figs DF7a-h (Y1), DF11a-j (Y2) and DF15a-n for L-asparaginase activity (Y3). Individual factor and cube plots are presented as Figs DF8 and DF9 (Y1), Figs DF12 and DF13 (Y2) and Figs DF16 and DF17 (Y3). A multi-objective optimization approach using desirability function was employed for the optimization experiment. L-asparaginase activity was maximized while total protein and biomass concentration were left at their minimum to obtain maximum specific activity of the enzyme and its yield, as presented in Fig DF18. Data for the anti-cancer cytotoxicity assay and the effects of levels of temperature, pH and metal ions on enzyme activity and stability are presented as Tables DT13 to DT19. Suitable controls and statistical analyses helped to interpret data.


Steps to reproduce

Data was obtained through sequential optimization experiments; from one-factor-at-a-time approach (Ekpenyong et al. 2017a) to Plackett-Burman design (Ekpenyong et al. 2017b). The regression coefficients of the first-order model obtained from PBD were used to calculate step sizes of significant predictors for path of steepest ascent (PSA) experiments. Nine sequential experiments were conducted and significant (p < 0.05) increments in L-asparaginase activity values were obtained as response variable. L-asparaginase activity value peaked at experiment 7 and the predictor levels at that point were employed as center points in a central composite rotatable design (CCRD) of a response surface methodology (RSM). The Design Expert version 12 recommended a Min-Run-Res V type fractional factorial experimental design of 6 predictors comprising 40 experimental runs including 6 center points for lack-of-fit (LOF) test. Three response variables were evaluated including biomass concentration (Banerjee et al. 1993), total protein (Bradford, 1976) and L-asparaginase activity (Imada et al. 1973). Models were significant at p < 0.05 and model adequacy checked with diagnostic plots, adjusted R2, predicted R2 and the Lack-of-fit (LOF) p-values. Since the study desired to obtain maximum specific activity and maximum specific yield of the enzyme, the multi-objective function maximized L-asparaginase activity while minimizing biomass concentration and total protein. Optimum factor settings were confirmed by conducting the production in a fermentation medium formulated with the optimized settings. A less than 5% difference between predicted L-asparaginase activity and that obtained in the confirmation experiment supported the predicted factor settings for L-asparaginase production by the mold. The anti-cancer potential of the therapeutic enzyme was investigated using four cancer cell lines namely HL-60, HeLa, MCF-7, HCT-116 and HEK-383 T (normal cells), using the in vitro cytotoxicity assay protocol by Skehan et al. (1990). Selectivity indices (SI) were calculated from IC50 value ratios of normal to tumor cell. Exponentially growing cell lines were prepared according to the protocol described in Ekpenyong et al. (2020). Analysis and IC50 determination for each cell line were performed using GraphPad Prism 8. The effects of temperature, pH and metal ions on L-asparaginase activity and stability were evaluated using the protocols described by El-Naggar et al. (2018), Li et al. (2018) and Mostafa et al. (2019). Reproducibility of data: Fermentation conditions and strain of Aspergillus candidus ( are critical towards reproducibility of results. The software employed for analysis included GraphPad Prism 8, Minitab 17 and Design Expert 12. Spectrophotometer calibration, centrifuge model, dialysis membrane, quality and sterility of cell lines, chemical grade and experimenter's tact are critical points for research reproducibility.


University of Calabar, University of Calabar Teaching Hospital, Cross River University of Technology, University of Uyo


Microbial Biotechnology, Bioprocess Optimization, Statistical Modeling, Bioprocess Modeling, Biochemical Toxicology, Aerobic Fermentation