Dynamic Model for Biomass and Proteins Production by Three Bacillus Thuringiensis ssp Kurstaki Strains

Round 1

Reviewer 1 Report

Authors of the present manuscript evaluate growth and production criteria of a bioreactor using semi-synthetic culture medium and develop a dynamic model that can be used to optimize Bt toxin production from three Bacillus thuringiensis ssp kurstaki strains.

Authors have used standard technology to evaluate the Btk cultures and to develop models. However, a major flaw in the manuscript is found in the section describing the Bt delta-toxin quantification. While Bradford method and BSA standards are fine for total protein quantification, the method used in this work for estimating protein production overestimates the Bt delta-toxin . In our experience, the proteins extracted using the protocol described in this manuscript contains about 50% proteins other than delta endotoxins because there is no reliable way to eliminate contaminating vegetative cells and debris transferred from spore pellets. We always use FPLC to purify endotoxins and use Western blots or purified (quantified) toxin as the standard to estimate the toxin concentration. I believe authors should purify the Bt toxin and re-evaluate the models before this manuscript could be published. 

In addition, there should be at least one more replicate for the Btk HD1 strain as the failure of batch 07 left that experiment with only 2 replicates. 

In table 5, Y2 (g_Pro/g_Glucosa*h)  should be Y2 (g_Pro/g_Glucose*h).

Author Response

Thank you for all your time here and comments. We have carefully responded to each of these comments and suggestions. All modifications were highlighted in yellow in the revised version. Hereafter, please note that the reviewer comments are shown in bold type and our responses in plain type.

 

Response to comments by Reviewer 1

Authors of the present manuscript evaluate growth and production criteria of a bioreactor using semi-synthetic culture medium and develop a dynamic model that can be used to optimize Bt toxin production from three Bacillus thuringiensis ssp kurstaki strains.

Authors have used standard technology to evaluate the Btk cultures and to develop models. However, a major flaw in the manuscript is found in the section describing the Bt delta-toxin quantification. While Bradford method and BSA standards are fine for total protein quantification, the method used in this work for estimating protein production overestimates the Bt delta-toxin . In our experience, the proteins extracted using the protocol described in this manuscript contains about 50% proteins other than delta endotoxins because there is no reliable way to eliminate contaminating vegetative cells and debris transferred from spore pellets. We always use FPLC to purify endotoxins and use Western blots or purified (quantified) toxin as the standard to estimate the toxin concentration. I believe authors should purify the Bt toxin and re-evaluate the models before this manuscript could be published. 

Thank you for this comment; we fully agree that alternative methods (more specific) can be used.

The Bradford method resulted in statistically higher estimates than either ELISA, or HPLC or SDS-PAGE/densitometry but also provided the lowest coefficients of variation (CVs) according to the published paper by Crespo et al. 2008 (Crespo AL, Spencer TA, Nekl E, Pusztai-Carey M, Moar WJ, Siegfried BD. Comparison and validation of methods to quantify Cry1Ab toxin from Bacillus thuringiensis for standardization of insect bioassays. Appl Environ Microbiol. 2008 Jan;74(1):130-5. doi: 10.1128/AEM.01855-07. Epub 2007 Nov 2. PMID: 17981939; PMCID: PMC2223200.).

Moreover, Bradford method was selected based on previous works of our teams (CBS and USJ) and others:

-- Abdelmalek N, Sellami S, Kallassy-Awad M, Tounsi MF, Mebarkia A, Tounsi S, Rouis S. Influence of Ephestia kuehniella stage larvae on the potency of Bacillus thuringiensis Cry1Aa delta-endotoxin. Pestic Biochem Physiol. 2017 Apr;137:91-97. doi: 10.1016/j.pestbp.2016.10.004. Epub 2016 Oct 21. PMID: 28364809.

- Abdelmalek N, Sellami S, Ben Kridis A, Tounsi S, Rouis S. Molecular characterisation of Bacillus thuringiensis strain MEB4 highly toxic to the Mediterranean flour moth Ephestia kuehniella Zeller (Lepidoptera: Pyralidae). Pest Manag Sci. 2016 May;72(5):913-21. doi: 10.1002/ps.4066. Epub 2015 Jul 24. PMID: 26103535.

- Azzouz H, Kebaili-Ghribi J, ben Farhat-Touzri D, Daoud F, Fakhfakh I, Tounsi S, Jaoua S. Selection and characterisation of an HD1-like Bacillus thuringiensis isolate with a high insecticidal activity against Spodoptera littoralis (Lepidoptera: Noctuidae). Pest Manag Sci. 2014 Aug;70(8):1192-201. doi: 10.1002/ps.3661. Epub 2013 Oct 22. PMID: 24124020.

- Azzouz, H., Kebaili‐Ghribi, J., Daoud, F., Abdelmalak, N., Ennouri, K., Belguith‐Ben Hassan, N., ... & Rouis, S. (2015). Selection and characterization of Bacillus thuringiensis strains toxic against pyralid stored‐product pests. Journal of Applied Entomology, 139(9), 690-700.

- Elleuch J, Tounsi S, Ben Hassen NB, Lacoix MN, Chandre F, Jaoua S, Zghal RZ. Characterisation of novel Bacillus thuringiensis isolates against Aedes aegypti (Diptera: Culicidae) and Ceratitis capitata (Diptera: Tephridae). J Invertebr Pathol. 2015 Jan;124:90-7. doi: 10.1016/j.jip.2014.11.005. Epub 2014 Nov 26. PMID: 25433312.

- Ennouri K, Ben Ayed R, Triki MA, Ottaviani E, Mazzarello M, Hertelli F, Zouari N. Multiple linear regression and artificial neural networks for delta-endotoxin and protease yields modelling of Bacillus thuringiensis. 3 Biotech. 2017 Jul;7(3):187. doi: 10.1007/s13205-017-0799-1. Epub 2017 Jun 29. PMID: 28664374; PMCID: PMC549143

- Fayad N, Patiño-Navarrete R, Kambris Z, Antoun M, Osta M, Chopineau J, Mahillon J, El Chamy L, Sanchis V, Kallassy Awad M. Characterization and Whole Genome Sequencing of AR23, a Highly Toxic Bacillus thuringiensis Strain Isolated from Lebanese Soil. Curr Microbiol. 2019 Dec;76(12):1503-1511

- Micheline EL Khoury, Hichem Azzouz, Chavanieu A, Abdelmalak N, Chopineau J and Kallassy Awad M. Isolation and characterization of a new Bacillus thuringiensis strain Lip highly toxic to Ephestia kuehniella (Lepidoptera: Pyralidae) larvae harboring a new cry1Aa gene. Archives of Microbiology 2014 Jun;196(6):435-44.

- Rahbani Mounsef J, Salameh D, Kallassy Awad M, Brandam C, Lteif R. Evaluation of a cereal milling by-product for the low- cost production of Bacillus thuringiensis kurstaki in submerged fermentation. European Journal of Biotechnology and Bioscience 2014; 1 (6): 10-16

 

In addition, there should be at least one more replicate for the Btk HD1 strain as the failure of batch 07 left that experiment with only 2 replicates. 

Based on reviewer 1 comment, as it was mentioned in section 2.5 “Between 3 to 4 batch cultures per strain were carried out (Table 2)”. Each experiment was conducted independently since mathematically independent experiments are necessary to calibrate and to validate our models.

For batch 07 we mentioned “It is relevant to indicate that in batches 07 and 08, cell biomass was estimated by OD600nm instead as an alternative to dry matter measurement...” since correlations (OD600nm vs dry matter measurement) were validated for all strains during oxidative phase. To make it more clear, we modified accordingly section 2.5 “It is relevant to indicate that in batches 07 and 08, dry matter measurement was estimated by OD600nm for exponential growth phase”.

In table 5, Y2 (g_Pro/g_Glucosa*h)  should be Y2 (g_Pro/g_Glucose*h).

We agree with this comment, new version was corrected (misspelling error because of Spanish native co-authors).

Reviewer 2 Report

This is s well-written manuscript. However, some minor edits are required to improve the English. Otherwise, it looks okay to me.

Author Response

Thank you for all your time here and comments.  All modifications were highlighted in yellow in the revised version. Hereafter, please note that the reviewer comments are shown in bold type and our responses in plain type.

This is s well-written manuscript. However, some minor edits are required to improve the English. Otherwise, it looks okay to me.

We agree with this comment, new version include English double check as well as typographic corrections.

Round 2

Reviewer 1 Report

Authors of this manuscript have responded to the comments by this reviewer with some arrogance. My first thought was to reject the paper because it falsely claimed have estimated the production of delta endotoxin and refused to address the concerns raised by this reviewer. But I decided to give them another chance. First, HD1 strain had three batches and batch 7 failed due to a technical error (Lines 194-195). Therefore, only two batches were available for analysis. How can authors claim that they have used 3-4 batches in the analysis?  If failed experiments can be used as replicates, we can publish a lot more papers in the Journal of Irreproducible Results. 

The quantification of the proteins is another point that authors have addressed amicably if not for their arrogance. Many researchers use crude proteins extracts in bioassays which are always compared to a susceptible strain to calculate the resistance ratios (RR). So, people in the field know not to put too much weight on the LC50, but to look for the RR. For example, an overlay assays will provide completely different LC50 values compared to diet incorporation assays of the same strain of insects. However, RR from both types of assays should be similar. However, in absolute quantifications like what is reported in this manuscript cannot report crude protein quantities and call them delta endotoxin because the methodology used in this manuscript will never produce 100% pure protein. Authors have used comparative studies done with bioassays or manuscript somehow fell through the cracks of imperfect review system to claim that “they did is that way so, we are doing it the same way”. This is how false information mushroom in the scientific literature. Authors need to learn to compare apples to apples, not apples to oranges. I have attached a picture of an FPLC run to demonstrate how much Cry1Ac and other proteins are present in the crude extract.  In this run, only a little more than 60% of the total proteins was Cry1Ac from HD1 strain.

Authors have listed a series of publications (mostly bioassays) that did not use a toxin purification method such as FPLC or sucrose gradients. Likewise, I can cite many publications that used proper methods to quantitate the toxin content and did not use the crude total protein and the “toxin concentration.” (e.g. 10.1038/s41598-017-02545-y, 10.1002/pro.2536). Just like authors cited publications that used crude protein as the “purified toxin”, I can cite 100’s of publications that have used toxin quantification accurately but do not want to waste my time on it.

Authors could have addressed this comment easily by admitting their technical shortcoming (or not having an FPLC to do the work) by simply stating that they are measuring the total protein production and NOT the delta endotoxin production. However, even that contradicts the purpose of this manuscript (comparing new strains to HD1) because we would not know the fraction of Cry1Ac in each strain.  Therefore, either the authors will have to change the theme of the paper and compare the total protein production among three strain or purify the Bt toxin and compare actual toxin production. Alternatively, authors can run a discontinuous sucrose gradient to purify the delta endotoxins (e.g. 10.1002/pro.2536) PAGE and demonstrate that Cry1Ac s the ONLY protein present in the crude protein extract (which I believe is impossible). Finally, although the authors state in the conclusions that “Our overarching aim to estimate -endotoxin bioproduction was successfully achieved” (lines 459-460), I do not believe the authors have estimated the delta-endotoxin production accurately.  Rather, they have estimated the total protein production in all three Bt strains.

Comments for author File: Comments.jpg

Author Response

Thank you for all your time here and comments. We have carefully responded to each of these comments and suggestions. Please note that your comments are shown in bold type and our responses in plain type.

Authors of this manuscript have responded to the comments by this reviewer with some arrogance. My first thought was to reject the paper because it falsely claimed have estimated the production of delta endotoxin and refused to address the concerns raised by this reviewer. But I decided to give them another chance.

We apology if our answers may have been misunderstood or appeared not cordial. It was not at all the objectives of our scientific exchanges. In this revised version R2, authors have carefully considered initial and new comments from reviewer 1. Consequently, the title and the materials and methods were modified to clarify that the performed measurements are the total protein after a specific separation procedure. Considering this methodology and bibliography, the endotoxin constitutes the dominant protein fraction (Bradford).

Changes in text :

• Title :

“Dynamic model for biomass and proteins production by three Bacillus thuringiensis ssp kurstaki strains.”

• Abstrat:

Line 27-28: “were depicted with an overarching aim to estimate total protein productivity”

• M&M:

Line 142: “Quantification of proteins production”

Lines 143-144: “In order to estimate the concentration of total proteins (mainly composed of d-endotoxin) produced during the fermentation”

Lines 152-153: “The concentration of the proteins in this supernatant was deter-mined by Bradford assay”.

• Conclusions:

Lines 460-462: “The objectives of approach were to model bioperformances (veg-etative cell, spore, substrate and protein) considering different B. thu-ringiensis ssp. kurstaki strains and successive culture phases”

Lines 463-465: “Our overarching aim to estimate total proteins production (mainly composed of d-endotoxin) was successfully achieved”.

First, HD1 strain had three batches and batch 7 failed due to a technical error (Lines 194-195). Therefore, only two batches were available for analysis. How can authors claim that they have used 3-4 batches in the analysis?  If failed experiments can be used as replicates, we can publish a lot more papers in the Journal of Irreproducible Results.

Thanks for this comment and appreciation. We wish toclarify these points by adding elements in :

• Section 2.5 (lines 166-167): “Two batches per strain were used to perform parameter calibration and between one and two batches were used to validate models.”
• Section 2.5.2 (lines 195-197): “However, in batch 07, no values were recorded for the biomass concentration after exponential growth phase due to technical misplaced measurements”

It is important to point out that model simulations use only initial conditions and parameters that were estimated previously during calibration. Results show that batch 07 can be simulated correctly during exponential growth phase for biomass and the model follows the dynamics of the other state variables.

The quantification of the proteins is another point that authors have addressed amicably if not for their arrogance. Many researchers use crude proteins extracts in bioassays which are always compared to a susceptible strain to calculate the resistance ratios (RR). So, people in the field know not to put too much weight on the LC50, but to look for the RR. For example, an overlay assays will provide completely different LC50 values compared to diet incorporation assays of the same strain of insects. However, RR from both types of assays should be similar. However, in absolute quantifications like what is reported in this manuscript cannot report crude protein quantities and call them delta endotoxin because the methodology used in this manuscript will never produce 100% pure protein. Authors have used comparative studies done with bioassays or manuscript somehow fell through the cracks of imperfect review system to claim that “they did is that way so, we are doing it the same way”. This is how false information mushroom in the scientific literature. Authors need to learn to compare apples to apples, not apples to oranges. I have attached a picture of an FPLC run to demonstrate how much Cry1Ac and other proteins are present in the crude extract.  In this run, only a little more than 60% of the total proteins was Cry1Ac from HD1 strain.

We are grateful to reviewer 1 to point out different approaches and somehow their contradictions to qualify or quantify proteins. In our work, we are aware that the methodology is not perfect and that a debate can be developed on this subject. Nevertheless, the initial objective of this article which is not to identify specifically the endotoxin but to model the whole bioperformances (substrate, biomass, sporulation, protein) with a semi-synthetic medium and using 3 strains.

Authors have listed a series of publications (mostly bioassays) that did not use a toxin purification method such as FPLC or sucrose gradients. Likewise, I can cite many publications that used proper methods to quantitate the toxin content and did not use the crude total protein and the “toxin concentration.” (e.g. 10.1038/s41598-017-02545-y, 10.1002/pro.2536). Just like authors cited publications that used crude protein as the “purified toxin”, I can cite 100’s of publications that have used toxin quantification accurately but do not want to waste my time on it.

This argumentation is perfectly acceptable but authors have answered here-above (idem).

Authors could have addressed this comment easily by admitting their technical shortcoming (or not having an FPLC to do the work) by simply stating that they are measuring the total protein production and NOT the delta endotoxin production. However, even that contradicts the purpose of this manuscript (comparing new strains to HD1) because we would not know the fraction of Cry1Ac in each strain.  Therefore, either the authors will have to change the theme of the paper and compare the total protein production among three strain or purify the Bt toxin and compare actual toxin production. Alternatively, authors can run a discontinuous sucrose gradient to purify the delta endotoxins (e.g. 10.1002/pro.2536) PAGE and demonstrate that Cry1Ac s the ONLY protein present in the crude protein extract (which I believe is impossible). Finally, although the authors state in the conclusions that “Our overarching aim to estimate d-endotoxin bioproduction was successfully achieved” (lines 459-460), I do not believe the authors have estimated the delta-endotoxin production accurately.  Rather, they have estimated the total protein production in all three Bt strains.

To avoid any confusion and to sustain Reviewer 1 remarks, we modified the text in M&M as follows :

Line 156-159: “Considering our protocol, protein concentration estimates the total protein production after separation but not specifically the d- endotoxin even if it is the dominant fraction “