Artificial Neural Network modelling for yield enhancement of L-Glutaminase free L-Asparaginase from Bacillus sp. associated with marine red alga Gracilaria dura

doi:10.21203/rs.3.rs-1764638/v1

Download PDF

Research Article

Artificial Neural Network modelling for yield enhancement of L-Glutaminase free L-Asparaginase from Bacillus sp. associated with marine red alga Gracilaria dura

https://doi.org/10.21203/rs.3.rs-1764638/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The bacterial flora associated with Gracilaria dura, the sole producer of natural agarose was obtained. Screening for the presence of various enzymes and bioactive compounds resulted in discerning different clinically and industrially important enzymes. Three isolates were identified to be producing L-asparaginase (LA), a potent anticancer drug. The strain 4T1-C26E registered highest LA activity (25.2 IU ml^− 1) in the enzyme assay, was selected for yield improvement by optimizing process parameters through Response Surface Methodology (RSM). Various parameters like glucose concentration, L-asparagine concentration, pH of the media, and incubation period were optimized through Central Composite Design (CCD) for enhanced yield of LA by 4T1-C26E, identified as Bacillus licheniformis. The CCD table was exploited for application in artificial neural networking (ANN) for the development of optimum method of cultivation. In this study, we obtained 2.5-fold (88.4 IU ml^− 1) increase in LA production by this strain through RSM optimization. Further, Artificial Neural Networking Particle Swarm Optimization (ANN-PSO) predicted the input combinations. The experimental validation of LA production was found to be 4-fold (126.4 IU ml^− 1) higher with respect to the unoptimized conditions. This illustrates the effectiveness of ANN-PSO based optimization technique, the error percent between the predicted and experimental output was 4.03%. The most important benefit of LA from 4T1-C26E is that it possesses no L-glutaminase activity. Further, it is extracellularly produced which is suitable for easier downstream processing thereby reducing chances of immunogenic responses by endotoxins. However, characterization of kinetic parameters and cytotoxic studies on cell cultures are mandatory to confirm its pharmaceutical application.

Microbial L-asparaginase

Anti-cancer

Seaweed

Central Composite Design

Artificial Neural Networking

Table 1: Enzyme activity of three bacteria identified positive for LA production in primary screening.

No	Isolate Name	Enzyme units (IU ml^-1)
1	4T1-C8E	15.2
2	4T1-C26E	25.2
3	5T1-C10E	20.2

Table 2: Range and levels of the selected physiochemical parameters

Independent Variables	Level
Independent Variables	-2	-1	0	1	2
pH	4.5	6	7.5	9	10.5
Glucose (%)	0.2	0.4	0.6	0.8	1
L-asparagine (%)	0	0.5	1.25	2	2.75
Incubation time (d)	4	8	12	16	20

Table 3: Central composite design displaying different combinations of critical physicochemical factors with their corresponding output and RSM and ANN predictions to the responses.

Run	Factor1A: pH (Number)	Factor 2B: Glucose (Percentage)	Factor 3C: L-asparagine (Percentage)		Factor 4D: Incub. period (Number)	Response1: Enzyme Activity (IU ml^-1)	RSM predicted	ANN Predicted
1	7	0.6		1.25	12	61	58	62.33
2	9	0.4		2	8	43	35.37	43.00
3	7	0.6		1.25	12	51	58	52.33
4	7	1		1.25	12	57	56.8	57.30
5	7	0.2		1.25	12	47	37.5	47.00
6	9	0.4		2	16	39	34.9	39.00
7	7	0.6		1.25	12	59	58	58.33
8	7	0.6		1.25	4	9	26.7	9.00
9	10.5	0.6		1.25	12	27	36.6	27.00
10	9	0.4		0.5	8	0	-4.1	0.00
11	7	0.6		1.25	20	35	29.7	35.00
12	6	0.8		0.5	8	29	28.6	28.01
13	7	0.6		1.25	12	51	58	52.33
14	6	0.4		2	16	53	54.4	53.00
15	9	0.8		2	8	41	36.1	41.00
16	9	0.4		0.5	16	41	37.4	41.00
17	6	0.8		2	8	88.4	91.3	88.40
18	9	0.8		0.5	8	7	5	6.99
19	9	0.8		2	16	25	37.5	25.37
20	6	0.4		0.5	8	27	13.8	27.00
21	7	0.6		0.25	12	0	24.6	0.86
22	4.5	0.6		1.25	12	84.4	79.8	85.69
23	6	0.8		2	16	63	62.7	63.00
24	7	0.6		1.25	12	55	58	54.33
25	7	0.6		1.25	12	53	58	52.33
26	9	0.8		0.5	16	59	48.4	58.99
27	6	0.4		2	8	78.9	85	78.00
28	6	0.8		0.5	16	35	42.05	34.99
29	7	0.6		2.75	12	67	71.3	67.00
30	6	0.4		0.5	16	25	25.4	24.99

Table 4: ANN – Validation experiments

pH	Glucose (%)	L-asparagine (%)	Incubation time (days)	Enzyme activity (IU/ml)		% Error between predicted and experimental output
pH	Glucose (%)	L-asparagine (%)	Incubation time (days)	Predicted values	Experimental values	% Error between predicted and experimental output
10	0.7	2.5	14	31.08	29.8 ± 1.4	4.29
8	0.5	1.5	10	46.16	46.2 ± 0.8	0.08
5	0.8	1	9	41.0	40.4 ± 0.9	1.48

Table 5: Biochemical characteristics of Bacillus endophyticus strain 4T1C8E, Bacillus licheniformis strain 4T1C26E, Bacillus velezensis strain 5T1C10E. + indicates Positive (more than 90%), - indicates Negative (more than 90%) and V indicates a 11-89% Positive result.

No.	Test	Reaction of Isolates
	Biochemical	4T1C26E	5T1C10E	4T1C8E
1	ONPG	+	+	-
2	Lysine utilization	-	-	-
3	Ornithine utilization	-	-	-
4	Urease	-	-	-
5	Phenylalanine Deamination	-	-	-
6	Nitrate reduction	+	-	-
7	H₂S production	-	-	+
8	Citrate utilization	+	+	+
9	Voges Proskauer's	+	+	-
10	Methyl red	+	+	-
11	Indole	-	-	-
12	Malonate utilization	+	-	-
	Carbohydrate utilization
13	Esculin hydrolysis	+	+	+
14	Arabinose	+	V	-
15	Xylose	-	-	+
16	Adonitol	-	-	-
17	Rhamnose	-	-	+
18	Cellobiose	V	V	+
19	Melibiose	V	-	+
20	Saccharose	+	V	+
21	Raffinose	-	-	-
22	Trehalose	V	+	+
23	Glucose	+	+	+
24	Lactose	-	-	-
25	Oxidase	-	+	-

No competing interests reported.

Supplementaryinformation.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Artificial Neural Network modelling for yield enhancement of L-Glutaminase free L-Asparaginase from Bacillus sp. associated with marine red alga Gracilaria dura

Status:

Version 1

Abstract

Figures

Full Text

Tables

Additional Declarations

Supplementary Files

Status:

Version 1