Bioelectroanalytical Detection of Lactic Acid Bacteria

Round 1

Reviewer 1 Report

Bioelectroanalytical detection of lactic acid bacteria

Brief summary

From a manuscript of Evelina J. Y. Hana and colleagues, the authors explained how to develop

extracellular electron transport (EET) for detection of a member of lactic acid bacteria (LAB), Enterococcus faecalis. The authors performed experiments and improved this method to detect LAB which is different from other bacteria described by previous studies.

General comments

The manuscript was organized and in order. However, a lack of important detail must be added to improve scientific soundness and reliability of this study, especially in the method and result sections. I have some suggestions and comments as follows.

• Please also spell out all acronyms the first time when they appear in the manuscript such as BHI and CFU.
• Please include the city and country of all used machines and substances. Some of them are missing.
• Major English language modification is needed. Please keep the sentence clear and straightforward instead of using long and complicated sentences.

 

Specific comments

Title

• The title is too broad. The title should be more specific based on the aim or result.

Abstract

• Briefly important methods should be described.

Introduction

• The introduction shows the important and basic knowledge according to this study.

Methods

The method section lacks a lot of detail and information, making this manuscript not transparent.

• Importantly, there is no statistical analysis and statistical parameter in this manuscript that makes this study unreliable.
• No information about the number of experiments, measurements, replicates has been described.
• Line 85 (2.1. Growth conditions for faecalis OG1RF and other lactoacid bacilli (LABs)): there is no information of how authors obtained the strain of bacteria, especially “lactoacid bacilli strain.”
• Line 90: What is the “desire inoculum”?
• Line 102-103: What was the endpoint for pH measurement? What kind of pH meter did the author use for measurement? How much “sufficient” volume did the authors need for pH measurement? Please provide the detail.

Results and discussion

Because of the lack of statistical analysis, the authors explained all results without any statistical parameters. The reader would not be confident with the result and the interpretation. In addition, the result part should be in accordance with the method part. The description of each figure is also not clear.

• The scientific sounding of this part is poor since there is no statistical explanation wether which group compare to any groups and how significant of the result.
• No P values for any comparisons.
• The amount of variation or dispersion of the data or result is not indicated whether it is a standard deviation (SD) or other parameters. What do the error bars in all figures represent?
• There is no information of how the data (mean values) were calculated from.
• Figure 2 does not show any error bars.
• From the method point 2.1, it seems the author has used LABs strain but there is no result about them.
• From line 97-98, the author mentioned that OD600 were recorded at 0, 24, 48 h. Where is the result? This also leads to the question of what is the time point of OD600 measure in figure 2 panel a?

Conclusion

• Authors should emphasize the significance and application of the finding or development of this method to the real-life situation, either clinical or industrial field.

Author Response

We thank the reviewers for their valuable feedback which has helped improve our contribution considerably. We have made major revisions to the manuscript based on their comments. Please see our response to each question below:

Reviewer 1

Because of the lack of statistical analysis, the authors explained all results without any statistical parameters. The reader would not be confident with the result and the interpretation. In addition, the result part should be in accordance with the method part. The description of each figure is also not clear.

We respectfully disagree that the results lack statistical analysis. For all the main the results, we have presented, basic but appropriate statistical procedures have been applied and all experiments have been replicated or, for the pH experiment, pseudo replicated (see below for how it was necessary to pool these samples due to their low volume). In this manuscript, the dataset presented does not require nor will it benefit from more in-depth statistical tests. For example, take figure one, which is the key piece of data in the manuscript, it shows a dataset with triplicate analysis with 100s of data points expressed as the mean of the triplicates and with the standard deviation expressed as error bars (1a) or shaded points (1b). The key take home from figure one is what is the likely concentration of bacteria (±error) of inoculum size at a given time after incubation. However, we realise this may not have been communicated well in the original manuscript and we have made the following changes:

At ln 18 we have added:

“…where the values in parentheses represents the standard deviation of three replicates”

The full sentence beginning at ln 15 now reads:

“For every log fold increase in inoculum size ranging from 10,000 to 10,000,000 CFU mL^-1, the detection time increased linearly yielding mean detection times of 235 (± 16), 148 (± 12), 62 (± 12), and 32 (± 8) min for 10⁴, 10⁵, 10⁶, and 10⁷ CFU mL^-1 respectively where the values in parentheses represents the standard deviation of three replicates.”

The scientific sounding of this part is poor since there is no statistical explanation wether which group compare to any groups and how significant of the result.

No P values for any comparisons.

In the analyses presented, there is no need to compare between groups. Taking figure 1 as an example, the data between different inoculum densities, are quite clearly different from one another and neither was not our aim to make a statistical comparison which would be of limited utility for an analytical procedure. We have not reported p values as we opted not to apply any statistical tests in which the p value should be reported. We hope this satisfies the reviewer.

The amount of variation or dispersion of the data or result is not indicated whether it is a standard deviation (SD) or other parameters. What do the error bars in all figures represent?

Variability has been reported at the bottom of each figure. For example, at ln 392: ‘Mean values are shown, error bars correspond to the standard deviation of three replicates and are represented by bars (A) or by the shaded area (B).’

However, and in the interest of clarity, we acknowledge that this information should be incorporated into body of the text. We have made the following changes:

At ln 132 we have added:

“COD measurements were conducted in triplicate and the mean value was used in the to calculate CE.”

Reviewer 2 Report

The article can be accepted after the major revision

 

This manuscript presents a study related to electrochemical detection of lactic acid bacteria (Enterococcus faecalis) 18 resorufin-β-D-galactopyranoside as substrate. The detection range is rapid ranging from 34 to 235 minutes for inoculum 19 sizes between 107 and 104 CFU ml⁻¹.

 

1. Authors must mention which type of SPE (working and counter electrode names).
2. Why authors have not used macro electrodes like glassy carbon for detection they are more cost-effective.
3. Authors must give a pictorial description of external electron transfer.
4. What is the electron transfer region on the bacteria surface?
5. Why authors are not using methods like cyclic voltammetry or impedance to confirm electron transfer results.
6. Authors should use any other method to support their results.

Author Response

We thank the reviewers for their valuable feedback which has helped improve our contribution considerably. We have made major revisions to the manuscript based on their comments. Please see our response to each question below:

Reviewer 2

This manuscript presents a study related to electrochemical detection of lactic acid bacteria (Enterococcus faecalis) 18 resorufin-β-D-galactopyranoside as substrate. The detection range is rapid ranging from 34 to 235 minutes for inoculum 19 sizes between 107 and 104 CFU ml⁻¹.

 Authors must mention which type of SPE (working and counter electrode names).

The reviewer is correct and offers a good piece of advice, we have included the specifications of the screen-printed electrodes in our manuscript. Thanks for the suggestion.

 

We have added this information at line 110 which previously read:

 

“The SPEs consist of a circular carbon working electrode with a diameter of 4 mm, a carbon counter electrode and a silver reference electrode (DropSens).”

 

But now reads thus:

 

“The SPEs consist of a circular carbon working electrode with a diameter of 4 mm, a carbon counter electrode, and a silver reference electrode on a ceramic support (DropSens, DRP-C110, Metrohm, Switzerland).”

 

Additionally, and in response to the reviewers request for a diagram below, we have graphically represented the SPE along with the electron transfer mechanism in a new figure, Figure 1.

 

 

1. Why authors have not used macro electrodes like glassy carbon for detection they are more cost-effective.

 

There is a lot of work on electrodes and their compatibility with bacteria – glassy carbon is typically not considered an ideal material for microbial bioelectrochemical systems. Besides, the intention here is to have a very simple ‘plug and play device’ that can be used by biologists or lay people to estimate bacterial numbers without the need to carry out any complicated electrochemical reactor set up or analyses. The screen-printed system is the nearest we could get to such a system. We agree with the reviewer that the research grade electrodes used here are relatively costly. However, screen printed electrode technology is inherently scalable and can be applied cheaply. There are numerous commercial examples where SPEs that are used routinely and at very low cost, for example glucometer test strips retail for around 0.35 USD: Qoo10 - Easy Touch Blood Glu : Nutritious Items

 

We have qualified this notion by making the following change at line 114:

 

“SPE technology is convenient for non-specialist users and can be commercially implemented at very low cost”.

 

2. Authors must give a pictorial description of external electron transfer.

 

At the reviewer’s request, we have provided a pictorial description, please see the new figure added to the manuscript, Figure 1.

 

3. What is the electron transfer region on the bacteria surface?

 

Electron transfer on bacterial cell surfaces typically occurs through ‘direct’ processes such as outer membrane cytochromes or appendages know as nanowires. In this instance the electron transfer is mediated by the additive electron shuttle, resorufin, which is released and activated by the bacterial enzymes from resorufin-β-D-galactopyranoside. The electron transfer occurs between reduced bacterial electron carriers, such as quinoids, which donate electrons to the resorufin to form hydroxyresorufin which in turn become oxidised at the electrode. In Fig 3a, the diminutive current collected from the unmediated system demonstrates the utility of resorufin in electron transfer. Mediated electron transport with the type of carriers used here (i.e. resorufin) is not controversial in microbiology. We invite the reviewer to read one of the many reviews on the subject: Kato S. Biotechnological Aspects of Microbial Extracellular Electron Transfer. Microbes Environ. 2015;30(2):133-139. doi:10.1264/jsme2.ME15028

 

4. Why authors are not using methods like cyclic voltammetry or impedance to confirm electron transfer results.

 

The intention of this study was to see if a non-respiring organism can be used to set up a detection signal in a very simple device that can be used by laypeople. The type of detection has been described before, in for example, (Hinks, J., Han, E. J., Wang, V. B., Seviour, T. W., Marsili, E., Loo, J. S., & Wuertz, S. (2016). Naphthoquinone glycosides for bioelectroanalytical enumeration of the faecal indicator Escherichia coli. Microbial biotechnology, 9(6), 746-757.) We are interested only in the detection time and the respiratory behaviour of the organism and never intended to use such techniques. We hope our method paves the way for in depth testing of this organism in both applied and basic science studies going forward.

 

We have qualified this notion by adding the following text at ln 34:

 

“Simple techniques for enumerating LAB that could operated by non-specialists would, therefore, find application in numerous settings.”

 

5. Authors should use any other method to support their results.

 

We thank the reviewer for recognising that we have used any other method to support our results, We have used the detection assay itself and supported the results with growth assays and a respiratory analysis. This work was interrupted by COVID-19 and has not resumed and is unlikely to be continued du to working and manpower restrictions.

Reviewer 3 Report

The authors used resorufin-β-D-galactopyranoside to detect and enumerate Enterococcus faecalis bioelectroanalytically, and tested the response of E. faecalis to heme and menaquinone. In addition, the authors showed that it is only necessary to capture a little of a microbe’s metabolic activity to set up an effective detection signal, and suggested building a new consensus about what EET should mean and propose. Here, I’d like to encourage the authors to address the following concerns.

Major:

1. It seems that the research of response to heme and MK-4 is not clearly related to the detection described in Line 139-161.
2. The E.coli with different detection times mentioned in Lines 179-184 are two strains, while in this study only analyzed one strain of E. faecalis, it’s not necessary to compared this work with that research.
3. In Lines 191-196, the authors suggested that residual heme or quinoids in the BHI medium may have an effect. However, the authors did not change the medium. If there were indeed residual heme or quinoids in the medium, the subsequent experimental results could not be valid.
4. As a scientific research work, please elaborate on the innovation of this research.

Minor:

5. There is no error bar in figure 2.
6. Sometimes, the word "heme" is used to describe heme in this article, but sometimes "haem" is used, and there is even a "heam" used in Line 194. Please use the same word consistently.
7. If the electroactive glycosides are exogenously added electron carriers, the control mentioned in Line 224 should not be called "wild-type".
8. It appears that all species producing glucosidases can respond to electroactive glycosides. Hence the detection method described in this article could not specifically detect E. faecalis.
9. Is there any more realistic model describing the relationship between log CFU and Time besides the linear model?

 

Author Response

We thank the reviewers for their valuable feedback which has helped improve our contribution considerably. We have made major revisions to the manuscript based on their comments. Please see our response to each question below:

Reviewer three

 

1. It seems that the research of response to heme and MK-4 is not clearly related to the detection described in Line 139-161.

 

EET is the mechanism whereby bacteria communicate with an electrode, and it is usually described as a respiratory process. LAB do not respire but instead conserve energy by fermentation, the main difference between respiration and fermentation in the context of EET is that in respiration, electron balance of respiratory electrons is achieved by the reduction of a terminal electron acceptor such as oxygen or metal oxides. This is not the case in fermentative organisms and where the carbon source acts as both the electron donor and the acceptor. Hence conceptually, in organisms that do not respire, like the LAB, it is hard to reconcile their metabolism to electrode reduction and they are typically overlooked in such bio electrochemical studies. Supplementing LAB with MK4 and heme is a common strategy to induce respiration in LAB which we did. The author is correct that the detection is not related to these these but they were necessary to show that there is not much stimulation of the LAB in question which leads us to the conclusion that there LAB in this study are NOT respiring the electrode and that the current detected is incidental (ln 276). This is an important observation and while it is not as emotive or impactful as claiming the organism to be electrogenic, it nonetheless shows that it is possible to detect non respiring organisms by this means which we think is a finding of utility and of longevity and which will give researchers confidence to pursue the detection of these organism.

 

1. The E.coli with different detection times mentioned in Lines 179-184 are two strains, while in this study only analyzed one strain of E. faecalis, it’s not necessary to compared this work with that research.

 

The work to which the reviewer refers did indeed use different strains of E. coli. E. coli can both ferment and respire. Using a similar detection technique, mutants of E. coli which were unable to respire, and which obtained energy by fermentation only could not be detected very well. This was the inspiration for this work which sought to see how well an obligate fermentative organism could be detected. We are not comparing with the work that the reviewer mentions but contrasting with it to highlight the surprise finding that the LAB are so rapidly detected.

 

We have made the following change to help reaffirm this message at ln 38:

 

“Even though LAB possess many of the components of respiratory chains, they do not respire on account of their inability to synthesise functional heme containing cytochromes that typically act as the terminal reductases in oxidative respiratory chains [6].”

 

1. In Lines 191-196, the authors suggested that residual heme or quinoids in the BHI medium may have an effect. However, the authors did not change the medium. If there were indeed residual heme or quinoids in the medium, the subsequent experimental results could not be valid.

The reviewer raises a good point, in an ideal world it would be good to have a completely defined medium. E. faecalis is a fastidious organism and it is difficult to grow on defined medium and therefore is most commonly grown with undefined components (soya bean or yeast extract etc). That is why we opted for the growth supplement studies and nonetheless it does not affect our conclusion, i.e. that very little of the organism metabolic power is being harvested at the electrode and that this technique can be applied to organisms that do not respire the electrode.

 

1. As a scientific research work, please elaborate on the innovation of this research.

 

We have combined the comments above and made the following points about innovation of this work:

 

At ln 36:

 

“Genera belonging to the LAB include Lactobacillus, Streptococcus, Vagococcus, and Enterococcus. Simple and rapid techniques for enumerating non respiring LAB and which could be operated by non-specialists would, therefore, find application in numerous settings.”

 

 

Minor:

1. There is no error bar in figure 2.

 

The reviewer is correct, for the Figure 2b does not have any error bars as the data are the aggregate of three pooled samples and in this sense are pseudo replicated. We have included the following explanation in the text:

At ln 102:

“End point pH was recorded with a pH meter from pooled samples to ensure sufficient volume for accurate pH measurements from the low volumes”

We have added the error bars to Fig 2a, we thank the author for spotting this oversight and apologise for this error.

 

1. Sometimes, the word "heme" is used to describe heme in this article, but sometimes "haem" is used, and there is even a "heam" used in Line 194. Please use the same word consistently.

 

Thanks again for your useful suggestions. We have made five replacements of ‘haem’ to ‘heme’ and one replacement of the erroneous ‘heam’ to ‘heme’. There are now 14 instances of the word ‘heme’

 

1. If the electroactive glycosides are exogenously added electron carriers, the control mentioned in Line 224 should not be called "wild-type".

 

Wild type refers to the genetic background of an organism and not the treatments. And in this case all the organisms are wild types, so we agree the use of the word in this instance is better replaced with the word ‘control’. Thus we have replaced ‘wild type’ with control at ln 224. We hope this clarifies the manuscript.

 

 

1. It appears that all species producing glucosidases can respond to electroactive glycosides. Hence the detection method described in this article could not specifically detect E. faecalis.

 

The reviewer is correct. For very specific detection of bacteria additional development of growth medium is required and is commonly applied in microbiology. The approach is usually to use growth inhibitors to supress the growth of non-target organism along with growth promoters for the target organism. However, this was not an aim of this work as it is beyond our means and the intended scope of the paper. The number of permutations in media development is mindboggling and of low impact and is typically conducted in industrial research settings. Specific but proprietary medium for E. faecalis does exist. See the entrolert kits from idexx for water testing which use specific growth media along with a glucoside that is conjugated to a chropophore in an analogous colorimetric assay which takes 24 hours for detection:

Enterolert - IDEXX US

 

We have added this following at ln 281:

 

“Nonetheless, the reducing power captured here is sufficient to achieve detection that may be more rapid than traditional techniques.”

 

 

1. Is there any more realistic model describing the relationship between log CFU and Time besides the linear model?

   

There may be a better or more realistic model but we have opted to use a simple approach that exploits the well know growth phenomenon of called the lag whose length is directly proportional to the number of bacteria in the original sample.  

Round 2

Reviewer 1 Report

Please describe the statistic parameters in the Fig 3A (OD₆₀₀): mean with standard deviation... like the Fig 2.

Reviewer 2 Report

The article can be acceptable in the present form