Review Reports - Estradiol-Induced Modulation of Clindamycin Susceptibility in Mono- and Dual-Species Biofilms of <i>Lactobacillus gasseri</i> and <i>Cutibacterium acnes</i>: An In Vitro Model Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript focuses on a highly atypical community: the vaginal resident L. gasseri and the
allochthonous C. acnes. Despite the clear significance of the topic, reading the manuscript raised a
number of questions and concerns. For clarity, I have divided these comments and questions into
sections.
I) Concerning mechanisms of interaction with a surface and biofilm dynamics
1. In the L. gasseri MA4 and C. acnes EAB1 co-culture, clindamycin significantly stimulated biofilm
growth despite both strains demonstrating sensitivity in monoculture. Could this stimulation be
partially driven by cross-feeding mechanisms utilizing nutrients from lysed cells, or by a specific
stress-induced overproduction of the extracellular polymeric matrix?
2. Regarding the divergence observed between CV, CFU, and MTT assay results, the manuscript
logically suggests the induction of a viable but non-culturable (VBNC) state. Could this
discrepancy also be attributed to a hyperproduction of extracellular polymeric substances (EPS)
in response to stress, which would increase CV staining while masking a decline in the viable cell
count?
3. The data indicate that pre-adsorbed estradiol frequently exerts a more pronounced effect,
particularly in mixed-species biofilms. Beyond specific receptor-like interactions, is it possible
that the pre-adsorption process alters the physicochemical properties of the polystyrene surface
(e.g., modifying local hydrophobicity), thereby creating a microenvironment more conducive to
initial cellular adhesion?
4. The working concentration of estradiol (2.2x10 -10 M) was selected based on systemic blood
plasma levels. Given that local tissue concentrations within the vaginal mucosa may differ—and
potentially be higher due to localized accumulation—could this anatomical context be
addressed in the Discussion to further support the robust effects observed with surface-bound
estradiol?
II) About methodology
5. In Section 2.4, optical density (OD 540 ) measurements were recorded under static conditions.
Because OD in a liquid medium reflects total culture turbidity (including the planktonic fraction),
would it be more precise to refer to this parameter as "static growth kinetics" rather than
"biofilm formation kinetics"? If the latter is retained, how was the specific contribution of the
adhered biomass distinctly delineated from planktonic proliferation?
6. For the enumeration of CFUs in dual-species biofilms, MRS and RCM media were utilized. Given
that lactobacilli can often proliferate on rich media such as RCM, and cutibacteria may survive
on MRS, what specific controls or selective inhibitors were employed to prevent cross-growth,
or was differentiation strictly based on distinct colony morphology?
7. The study investigates a sub-inhibitory clindamycin concentration of 0.5 µg/mL. Would it be
beneficial to explicitly state the Minimum Inhibitory Concentration (MIC) values for each of the
four specific strains (ATCC 33323, MA4, HM514, EAB1) to establish a definitive baseline
demonstrating the "sub-inhibitory" nature of this specific dosage?
8. In the agar block diffusion assay used to assess antibacterial activity, is it possible that the
synthesized bacteriocins are being sequestered or neutralized by the robust extracellular
polymeric matrix of the mature biofilm rather than freely diffusing into the agar? Have the
authors considered evaluating cell-free supernatants extracted directly from the biofilm matrix
to quantify actual bacteriocin release?

Also, to definitively address the discrepancies between biomass (CV staining) , metabolic activity
(MTT), and viability (CFU), and to visually validate the VBNC hypothesis, the incorporation of
Confocal Laser Scanning Microscopy (CLSM) would significantly elevate the manuscript. Using
differential fluorescent staining—such as a Live/Dead viability kit or utilizing DAPI (optimally at a
concentration of 0.2 μg/ml to ensure sharp nuclear counterstaining while minimizing
background fluorescence in the dense EPS matrix) combined with a lipophilic membrane
probe—would provide crucial 3D spatial context regarding how L. gasseri and C. acnes co-
localize and physically restructure under estradiol and clindamycin stress.
III) Gene expression studies
9. Given that the bacterial cells were subjected to significant multiphasic stress (antibiotics,
hormones, and solvents), was the stable expression of chosen reference housekeeping gene
validated under these experimental conditions?
10. The manuscript notes that transcriptional changes in bacteriocin genes do not consistently
correlate with phenotypic antagonism. Could post-transcriptional regulatory mechanisms, or the
rapid degradation of synthesized peptides by proteases within the competitive mixed-species
environment, also contribute to this observed uncoupling?
IV) Data Completeness and clarification
11. The MTT assay results for dual-species biofilms appear to be heavily dominated by the
metabolic profile of the C. acnes components. Were any preliminary normalization steps
performed to account for potential intrinsic differences in the baseline reduction rates of the
tetrazolium dye between L. gasseri and C. acnes?
Additionally, in Section 3.1.1, the text states that the hormonal effect is "manifested mainly
exclusively on biofilm cultures and is not detected in planktonic growth forms". Since this is a
critical distinction that highlights the unique nature of the biofilm phenotype, it would be highly
beneficial to include the comparative planktonic growth data as a Supplementary Figure rather
than relying solely on a literature citation.

Finally, a few words about clinical context. C. acnes is described as an adventitious member of
the vaginal microbiota. Given the profound strain-specific differences observed between the
skin-derived (HM514) and vaginal (EAB1) isolates under estradiol exposure, might the vaginal
isolate possess specific genomic adaptation markers (e.g., altered lipid metabolism genes or
specialized adhesion factors) that explain its distinct physiological response? A brief mention of
this in the Discussion could contextualize why these two isolates behave so differently.

Author Response

REVIEWER 1

I) Concerning mechanisms of interaction with a surface and biofilm dynamics

Question 1. In the L. gasseri MA4 and C. acnes EAB1 co-culture, clindamycin significantly stimulated biofilm growth despite both strains demonstrating sensitivity in monoculture. Could this stimulation be partially driven by cross-feeding mechanisms utilizing nutrients from lysed cells, or by a specific stress-induced overproduction of the extracellular polymeric matrix?

Answer: This is a plausible explanation, but the scope of the paper was to describe the effects of estradiol and clindamycin on the bacterial communities of L. gasseri and C. acnes. Further experiments are planned to determine the specific mechanisms of action of the active compounds. Indeed, clindamycin stimulated biofilm growth in the dual-species biofilms of L. gasseri MA4 and C. acnes EAB1, despite both strains individual antibiotic sensitivity. Although this stimulation was observed, the specific mechanisms underlying it, such as cross-feeding from lysed cells or stress-induced excess production of extracellular polymeric matrix, were not fully elucidated in our study. Based on literature data on similar mechanisms of defense against stress conditions in other bacterial biofilm communities, it can be hypothesized that similar mechanisms may work in our community. However, in the future, the following experiments should be performed to confirm this. First, it is necessary to measuring the degree of cell lysis in monocultures and cocultures exposed to clindamycin using. For example, live/dead cell staining, using of labeled nutrients to monitor their uptake and metabolism by surviving cells in the dual-species community after clindamycin treatment can help here. Next, to determine the change in the amount of EPS produced under stress conditions, the full study of extracellular matrix biochemistry might be conducted. Next, it will be necessary to quantify the total amount of biomass in dua-lspecies biofilms treated with clindamycin compared to untreated control samples. Finally, it will be useful to analyze the expression level of genes involved in the EPS synthesis pathways in both L. gasseri MA4 and C. acnes EAB1 under clindamycin-induced stress in monocultures. Thus, this should be a completely new study.

We added in the discussion the following text (LINES 780-783): “Furthermore, the key factors underlying this alteration in the active substance's effect are likely the interactions among microorganisms within the community and the resulting change in the composition of the extracellular polymeric matrix, which warrants further investigation.”

Question 2. Regarding the divergence observed between CV, CFU, and MTT assay results, the manuscript
logically suggests the induction of a viable but non-culturable (VBNC) state. Could this
discrepancy also be attributed to a hyperproduction of extracellular polymeric substances (EPS)
in response to stress, which would increase CV staining while masking a decline in the viable cell count?

Answer: Thank you for the remark. Actually, many hormones and antimicrobials including estradiol and clindamycin were shown to alter production and composition of biofilm matrix. In this paper we wanted firstly to prove any effect of estradiol on clindamycin efficacy. And of course, further experiments are planned to determine specific changes and molecular mechanisms that can explain observed results.

Indeed, both the induction of a VBNC state and EPS hyper production are bacterial stress responses that can lead to discrepancies between different viability assessment methods. Induction of a VBNC state is typically observed in most bacterial cultures. An increase in EPS can contribute to more intense CV staining (which marks total biomass), potentially masking a decrease in metabolically active or culturable cells, as measured by the MTT assay or CFU assay, respectively. However, not all stress conditions necessarily lead to increased EPS production. (For instance, methylparaben, as an environmental contaminant, compromises water disinfection under real conditions. https://doi.org/10.1016/j.chemosphere.2025.144685).

Therefore, to evaluate the contribution of the extracellular matrix to the observed response, the biofilm matrix could be extracted and its total amount quantitatively assessed as it was mentioned above. Changes in matrix production could then be compared between control samples and samples exposed by substances inducing stress in the studied cells.

We added in the discussion the following text (LINES 837-839): “Furthermore, the aforementioned changes in the production of the extracellular polymeric matrix may also account for this discrepancy in results, which requires separate investigation in the future.”

Question 3. The data indicate that pre-adsorbed estradiol frequently exerts a more pronounced effect,
particularly in mixed-species biofilms. Beyond specific receptor-like interactions, is it possible
that the pre-adsorption process alters the physicochemical properties of the polystyrene surface
(e.g., modifying local hydrophobicity), thereby creating a microenvironment more conducive to
initial cellular adhesion?

Answer: There is indeed evidence in the literature that adsorbed molecules and surface modifications can alter the properties of a surface, in our case, polystyrene, and influence cell adhesion. (for instance https://doi.org/10.1163/016942409X12598231567862). Polystyrene is known to accumulate estradiol (https://doi.org/10.1016/j.envres.2025.120977), potentially this can make the plastic more bioavailable at the interface between bacteria and plates. However, little is known about changes in local hydrophobicity caused by hormone accumulation. It is plausible that an alteration in the local microenvironment could partially explain the observed results; however, the purpose of the pre-absorption part was to rule out that the accumulation of estradiol on the surface of polystyrene played a minor role in the observed results. Therefore, the proposed mechanism is consistent with known principles of surface-cell interactions; further studies of its interaction with polystyrene and changes in its properties are needed to confirm this specific effect of estradiol.

We added in the discussion the following text (LINES 820-824): “Polystyrene is known to accumulate estradiol (https://doi.org/10.1016/j.envres.2025.120977), potentially this can make the plastic more bioavailable at the interface between bacteria and plates. However, little is known about changes in local hydrophobicity caused by hormone accumulation. It is plausible that an alteration in the local microenvironment could partially explain the observed results”

Question 4. The working concentration of estradiol (2.2x10 ^-10M) was selected based on systemic blood
plasma levels. Given that local tissue concentrations within the vaginal mucosa may differ—and
potentially be higher due to localized accumulation—could this anatomical context be
addressed in the Discussion to further support the robust effects observed with surface-bound
estradiol?

Answer: Thank you! This concentration was chosen not only based on its similarity to blood hormone levels but also on preliminary experiments with the hormone and bacterial cultures. The primary criterion for selecting this concentration among all those tested was the significant effect of the hormone at this concentration on the studied cultures. This concentration demonstrated the best results, which could be used for further development. However, areas with increased concentrations of adsorbed estradiol may indeed appear in the vagina. Its adsorption to the surface may increase the local concentration of the hormone available to bacterial cells during the initial stages of attachment. This localized effect may enhance the adhesion and biofilm formation of vaginal microorganisms. Furthermore, literature data demonstrating strong local pharmacokinetic and tissue-specific effects of vaginal estradiol support the idea that surface-bound estradiol may exert more pronounced biological effects than estradiol uniformly distributed throughout the environment. (Differences in Local and Systemic TFV PK Among Premenopausal Versus Postmenopausal Women Exposed to TFV 1% Vaginal Gel 2018, A. Thurman et. al).

Concerning anatomy, according to the review by Carmina and Lobo [https://doi.org/10.1016/B978-0-323-47912-7.00034-2], the blood plasma estradiol concentration in women during the mid-luteal phase ranges from 0.22 to 0.73 nM. However, the correlation between estradiol concentrations in tissues and serum appears to be marginal. A number of studies have assessed estradiol levels in endometrial tissue (see, e.g., [https://journals.lww.com/greenjournal/abstract/2002/04000/estradiol_absorption_from_vaginal_tablets_in.8.aspx; https://doi.org/10.1097/GME.0000000000001463]) and examined the efficacy of various hormonal preparations for estradiol administration [https://doi.org/10.1016/S0002-9378(99)70042-6; https://doi.org/10.1016/S0378-5173(01)00900-0; https://doi.org/10.1016/j.contraception.2004.12.015; https://doi.org/ 10.1097/GME.0000000000000790]. According to one review, the experimentally measured baseline estradiol concentration in endometrial tissue of older women was found to be between 1.3 and 3.85 × 10⁻¹¹ M [https://doi.org/10.1097/GME.0000000000001463], while that in blood plasma was reported to be considerably higher, at 3.7–7.3 × 10⁻¹¹ M [https://doi.org/10.1016/B978-0-323-47912-7.00034-2]. Some studies have even reported baseline serum concentrations in postmenopausal women below 1.3 × 10⁻¹¹ M [https://journals.lww.com/greenjournal/abstract/2002/04000/estradiol_absorption_from_vaginal_tablets_in.8.aspx]. In younger women, estradiol concentrations are significantly higher and fluctuate depending on the phase of the menstrual cycle and pregnancy status [https://doi.org/10.1016/B978-0-323-47912-7.00034-2], making it challenging to select a baseline concentration for experiments involving bacteria. Therefore, to maximize the uniformity of the resulting data, the estradiol concentration used in this study was based on the lower limit reported for blood plasma of women in the mid-luteal phase (0.22 nM).

We added some text in the discussion (LINES 824-831): “It is also known that the concentration of estradiol in vaginal tissues differs from that in blood plasma[50]. In postmenopausal women, it is typically an order of magnitude lower [51-53]. In younger women, the concentration varies depending on the menstrual cycle [54]. Taken together, these findings provide indirect, preliminary evidence, that potential decrease may play a role in the fact that the concentration of estradiol decreases upon its pre-adsorption onto the surface, thereby altering the effect of the hormone on biofilms, although definitive mechanistic studies will be required.”

II) About methodology
Question 5. In Section 2.4, optical density (OD₅₄₀) measurements were recorded under static conditions. Because OD in a liquid medium reflects total culture turbidity (including the planktonic fraction), would it be more precise to refer to this parameter as “static growth kinetics”; rather than “biofilm formation kinetics”? If the latter is retained, how was the specific contribution of the adhered biomass distinctly delineated from planktonic proliferation?

Answer: When measuring optical density (OD₅₄₀) directly in a liquid medium under static conditions, this indicator reflects the overall turbidity of the culture, including both planktonic (free-floating) cells and cells associated with the forming biofilm. Therefore, the use of the term "static growth kinetics" for this type of measurement appears appropriate and does not alter the original purpose of the experiment, which was to determine the antibiotic concentration that significantly affects the growth rate and generation time of microorganisms.

Regarding the specific contribution of the planktonic suspension to the total optical density in the well, this is a rather complex issue, since the microorganisms studied here are non-motile and their cells settle to the bottom of the wells. The true ratio of planktonic to biofilm cells is difficult to determine in this case, because the cells tend to aggregate in the near-bottom liquid layer, also forming biofilm aggregates. Accordingly, one may refer to a certain "suspension" fraction containing both biofilms and truly planktonic cells, as well as a "settled" fraction consisting solely of biofilms. Moreover, a portion of the "settled" fraction is inevitably detached during pipetting.

In the present study, we did not focus specifically on the distinction between these fractions. However, based on our experience, including the study by Mosolova et al., the optical density ratio of the settled fraction to the suspension fraction (the one collected by pipetting) can vary from 50:50 to 10:90, depending on the microorganism. In our case, we did not aim to examine this distinction specifically; rather, we focused on the biofilms at the bottom of the wells and, in case of kinetics, on the whole biomass.

We changed the 2.4 paragraph title to “Static growth kinetics study” (LINE 181), and the paragraph 3.2 title to “Static growth kinetic study of L. gasseri and C. acnes” (LINES 385-386).

Question 6. For the enumeration of CFUs in dual-species biofilms, MRS and RCM media were utilized. Given that lactobacilli can often proliferate on rich media such as RCM, and cutibacteria may survive on MRS, what specific controls or selective inhibitors were employed to prevent cross-growth, or was differentiation strictly based on distinct colony morphology?

Answer: To more easily differentiate lactobacilli and cutibacteria, two culture media were used: MRS and RCM. Both microorganisms were capable of growing on either medium, but lactobacilli demonstrated more pronounced growth on MRS, while cutibacteria grew preferentially on RCM. When counting colony-forming units in mixed cultures, differences in colony morphology served as an additional selective criterion: cutibacteria colonies were characterized by a pinkish color. The results were further confirmed using light microscopy, which is used to analyze cell morphology from each variant of colony (randomly). Throughout the experiment, no Lactobacillus was found on RCM plates with cutibactria, and no Cutibacterium was found on MRS with Lactobacilli. A plausible explanation for the observed results might be differences in growth rates, which affect resource availability and the production of antimicrobials that inhibit the growth of other bacteria.

We added the following text in the paragraph 3.5.1 (LINE 576-581): Lactobacilli exhibited more pronounced growth on MRS, whereas cutibacteria grew preferentially on RCM. When enumerating colony-forming units in mixed cultures, dif-ferences in colony morphology served as an additional distinguishing criterion: cutibacte-rial colonies were characterized by a pinkish coloration. The results were further verified by light microscopy, which was employed to examine cell morphology from each colony variant.”

Question 7. The study investigates a sub-inhibitory clindamycin concentration of 0.5 µg/mL. Would it be beneficial to explicitly state the Minimum Inhibitory Concentration (MIC) values for each of the four specific strains (ATCC 33323, MA4, HM514, EAB1) to establish a definitive baseline demonstrating the “sub-inhibitory” nature of this specific dosage?

Although 0.5 µg/mL clindamycin cannot be considered strictly sub-inhibitory for all strains tested, this concentration was selected as the working concentration for the following reasons. First, it represented the lowest level at which a consistent effect on microbial growth dynamics and/or biofilm formation was observed across all studied strains; depending on the strain, this effect manifested as either growth inhibition or stimulation. Second, the aim was to identify a concentration capable of reducing the growth rates of the strains by approximately 30–60%, as this degree of attenuation was expected to render the protective or sensitizing effect of estradiol more pronounced. In this range, the antibiotic impairs growth without fully suppressing it, thereby allowing the hormone's influence on antibiotic susceptibility to be assessed. It should be noted that the true sub-inhibitory concentration varies among the strains used; therefore, 0.5 µg/mL is referred to throughout the manuscript as the working concentration rather than a universally sub-inhibitory one. To improve clarity, the MIC values for each strain (ATCC 33323, MA4, HM514, and EAB1) will be mentioned in the revised manuscript.

We added some text in the paragraph 3.1.2 (LINES 356-359): All strains used in this study can be considered clinically relevant. Specifically, the minimum inhibitory concentration (MIC) of clindamycin is 10 µg/mL for L. gasseri ATCC 33323 and 20 µg/mL for L. gasseri MA4. Both cutibacterial strains exhibit MIC values exceeding 20 µg/mL.

Question 8. In the agar block diffusion assay used to assess antibacterial activity, is it possible that the synthesized bacteriocins are being sequestered or neutralized by the robust extracellular
polymeric matrix of the mature biofilm rather than freely diffusing into the agar? Have the
authors considered evaluating cell-free supernatants extracted directly from the biofilm matrix
to quantify actual bacteriocin release?

Answer: We thank the Reviewer for this remark. Indeed, the question of bacteriocin synthesis is highly relevant, as bacteriocins can be considered a hallmark of probiotic lactobacilli. In particular, L. gasseri produces several such compounds, and a novel one has been discovered relatively recently [https://doi.org/10.1007/s00253-020-10493-3]. Therefore, cell-free supernatants undoubtedly warrant investigation in a separate dedicated study. However, there is a certain complication here.

If cell-free supernatants in the classical sense are considered, the cultures must be grown in a liquid medium. If, on the other hand, one pursues a more specialized approach — as the Reviewer has quite rightly pointed out — and works exclusively with biofilms, specifically with the extracellular matrix, several issues arise that must be addressed. First, bacteriocins may diffuse into the surrounding medium; therefore, an inherent limitation would be that only the fraction retained within the matrix can be assessed. In this case, one would have to choose between this approach and collecting the entire solid medium beneath the biofilm to extract the total pool of bacteriocins, in order to determine what is released into the medium versus what remains in the matrix. Second, and logically following from the first point, bacteriocin production would need to be normalized per unit of biomass. This would require measuring either total organic carbon, the exact number of cells, or the amount of extracellular matrix. The latter is challenging, since even with our protocol [https://doi.org/10.3389/fmicb.2019.01284] there is no guarantee that the entire matrix is recovered down to the last molecule. Finally, there is a purely technical difficulty: the biomass yield of lactobacilli is low. In our experience, approximately 1–3 g of wet biomass can be obtained from 30 plates, each containing 30 mL of solidified MRS, which makes such experiments rather resource-intensive, particularly given the multitude of sample types involved. This does not preclude the necessity of conducting these experiments; however, it strongly argues for separating this work into an entirely new study.

Also, we added this text into the Discussion section (LINES 860-865): Finally, the agar-block diffusion assay should be clarified as measuring net diffusible an-tibacterial activity released from agar-grown bacterial communities, rather than total bac-teriocin synthesis. Reduced or absent inhibition zones, therefore, do not necessarily indi-cate absence of bacteriocin production, as antimicrobial compounds may be retained or neutralised within the colony biofilm matrix, associated cells, or agar block.

Question 8.1. Also, to definitively address the discrepancies between biomass (CV staining), metabolic activity (MTT), and viability (CFU), and to visually validate the VBNC hypothesis, the incorporation of Confocal Laser Scanning Microscopy (CLSM) would significantly elevate the manuscript. Using differential fluorescent staining—such as a Live/Dead viability kit or utilizing DAPI (optimally at a concentration of 0.2 μg/ml to ensure sharp nuclear counterstaining while minimizing background fluorescence in the dense EPS matrix) combined with a lipophilic membrane probe—would provide crucial 3D spatial context regarding how L. gasseri and C. acnes co-localize and physically restructure under estradiol and clindamycin stress.

Answer: We greatly appreciate this comment. Indeed, confocal laser scanning microscopy (CLSM) can, in certain cases, serve as a highly valuable tool for assessing the three-dimensional structure of biofilms. However, we chose not to perform such experiments within the scope of the present study for several reasons.

First, biofilms are prone to considerable background fluorescence — the matrix contains numerous organic substances that bind dyes. DAPI primarily stains nucleoids; however, increasing its concentration entails the risk of strong background fluorescence from the matrix. Conversely, decreasing the concentration carries the risk that an insufficient amount of the dye will penetrate the cells. As for lipophilic dyes, such as Dil, they present the same limitation, with the difference being that they stain only membranes and extracellular lipids or other hydrophobic compounds. Thus, the issue of background fluorescence would invariably persist. Consequently, a preferable approach would be to assemble binary communities from strains carrying, for instance, plasmids encoding GFP or other fluorophores. However, generating such recombinant strains of Gram-positive bacteria constitutes a separate, non-trivial task (involving vector selection, cloning, recombination, etc.).

Finally, microscopy techniques such as CLSM require a large number of samples and images for the results to be considered statistically reliable. In this context, it is far more practical to employ more robust methods when the three-dimensional organization of biofilms is not the primary objective of the study.

III) Gene expression studies
Question 9. Given that the bacterial cells were subjected to significant multiphasic stress (antibiotics, hormones, and solvents), was the stable expression of chosen reference housekeeping gene validated under these experimental conditions?

Answer: We sincerely thank the Reviewer for this pertinent question. The stability of reference gene expression under the multiphasic stress conditions imposed in our study (antibiotics, hormones, and solvents) is indeed a critical consideration for the accuracy of qPCR data. We fully agree that such validation is essential. In the present study, we employed 16S rRNA, which is commonly used as a housekeeping gene. While we did not perform an exhaustive validation across all stress combinations, we assessed its expression stability across several representative conditions and observed no significant variation (data not shown).
16S rRNA is a vital gene. Its critical function may require relatively stable expression, a desirable property for a reference gene. The selected reference gene was based on its widespread use in other studies, which reported relatively stable expression under various experimental conditions. In our study, it’s choice was based on data from previous studies and their use in experimental studies, rather than on de novo stability testing for each specific scaffold.

We added the following sentence in the text (LINES 704-705): “We assessed the expression of 16S rRNA gene stability across several representative conditions and observed no significant variation (data not shown)”

Question 10. The manuscript notes that transcriptional changes in bacteriocin genes do not consistently correlate with phenotypic antagonism. Could post-transcriptional regulatory mechanisms, or the rapid degradation of synthesized peptides by proteases within the competitive mixed-species environment, also contribute to this observed uncoupling?

Answer: The lack of a consistent correlation between transcriptional changes in bacteriocin-encoding genes and the observed phenotypic antagonistic activity may be explained by additional regulatory layers beyond transcription. In particular, post-transcriptional and post-translational regulatory mechanisms may significantly influence the final amount of active bacteriocin produced. Also, bacteriocins can be subject to degradation by endogenous proteolytic enzymes or by general instability in the growth medium, which may reduce their detectable activity despite increased transcription of the corresponding genes. It will be included in the revised manuscript as a possible explanation for the observed discrepancies between transcriptional and phenotypic data.

We added some information in the text (LINES 865-868): “Moreover, bacteriocins may be susceptible to degradation by endogenous proteolytic enzymes or to general instability in the growth medium, which can reduce their detectable activity despite increased transcription of the corresponding genes.”

IV) Data Completeness and clarification

Question 11. The MTT assay results for dual-species biofilms appear to be heavily dominated by the metabolic profile of the C. acnes components. Were any preliminary normalization steps
performed to account for potential intrinsic differences in the baseline reduction rates of the
tetrazolium dye between L. gasseri and C. acnes?
Additionally, in Section 3.1.1, the text states that the hormonal effect is “manifested mainly exclusively on biofilm cultures and is not detected in planktonic growth”. Since this is a
critical distinction that highlights the unique nature of the biofilm phenotype, it would be highly
beneficial to include the comparative planktonic growth data as a Supplementary Figure rather
than relying solely on a literature citation.

Answer, Indeed. the metabolic capacity and tetrazolium reduction efficiency between C. acnes and L. gasseri may influence the absolute MTT signal obtained in mixed cultures. In the present study, no species-specific normalization of MTT reduction rates was performed. However, the visual MTT reduction rate was quite similar in both species. Moreover, interpretation of the data was not based on absolute metabolic attribution, but on comparative differences between treated and untreated conditions, supported by complementary CFU and crystal violet assays.

In the present study, planktonic growth behavior was not the primary focus of the experimental design, which was centered on biofilm-associated responses. Therefore, planktonic-specific quantitative comparison was not systematically included. However, the absence of a detectable hormonal effect in planktonic cultures is consistent with previously published reports in similar systems. Nevertheless, due to no performing of a clear check, we deleted the text in the LINES 327-329 (It is noteworthy that the identified effect of the hormone is manifested mainly exclusively on biofilm cultures and is not detected in planktonic growth forms, which is consistent with literature data).

Question 12. Finally, a few words about clinical context. C. acnes is described as an adventitious member of the vaginal microbiota. Given the profound strain-specific differences observed between the skin-derived (HM514) and vaginal (EAB1) isolates under estradiol exposure, might the vaginal isolate possess specific genomic adaptation markers (e.g., altered lipid metabolism genes or specialized adhesion factors) that explain its distinct physiological response? A brief mention of this in the Discussion could contextualize why these two isolates behave so differently.
Answer: Thank you. We added the text into the discussion (LINES 767-774): Further, pronounced strain-specific differences observed between the vaginal isolate (C. acnes EAB1) and the skin-derived isolate (C. acnes HM514) under estradiol exposure may reflect niche-associated functional adaptations within the species. In this context, it is plausible that the vaginal isolate possesses distinct genomic or regulatory features that confer altered responsiveness to hormonal stimuli, particularly in pathways related to membrane composition, nutrient utilization, and cell surface interactions. Such adapta-tions may underlie the differential physiological responses observed between the two iso-lates in the present study.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting study that investigates if estradiol changes biofilm formation and clindamycin susceptibility in mono- and dual-species communities of Lactobacillus gasseri and Cutibacterium acnes, using both reference and vaginal isolates. The study is ambitious in scope and combines Crystal violet staining assays, kinetics, CFU counts, MTT, antagonism assays, and bacteriocin expression analysis. The main findings are certainly novel, especially the strain-dependent effects and the reported “inversion” of clindamycin activity in some dual-species settings. However, before publication, the authors should respond to the following points.

Major

Title and discussion frame C. acnes as part of the vaginal context, yet the introduction itself presents it mainly as a broader commensal that may be a transient vaginal component, citing only limited evidence for vaginal relevance. With only one skin-derived strain and one vaginal isolate, the manuscript should avoid broad statements about “the vaginal niche” (e.g., line 656) or “vaginal microbiota” unless those claims are explicitly restricted to this experimental model.
The figures are blurry and would benefit from higher resolution.
Estradiol was dissolved in ethanol, and ethanol controls were included. However, the methods report the final ethanol concentration as “0.6% (v/v) or 0.1 M” while Fig. 11 reports ethanol as 1.7 x 10^-5 M, which is inconsistent and needs correction or explanation.
The manuscript uses 0.5 ug/mL as a working concentration for interaction studies, but the results show that this concentration behaves very differently across strains: it inhibits L. gasseri, stimulates C. acnes HM514, and has mixed effects in dual-species communities. That makes it difficult to intrepret 0.5 ug/mL as a universally sub-inhibitory concentration. The authors should either justify this common-dose design much more explicitly or provide strain-specific MIC/MBIC-based reasoning.
The manuscript does not clearly state the biological replicate number for each assay in the main methods, despite presenting many pairwise comparisons across strains, treatments, and communities. The use of Mann–Whitney tests across many comparisons also requires a clearer multiple-testing strategy. The paper mentions “p < 0.05 or q < 0.05 where appropriate” but does not explain exactly when or how correction was applied. In addition, the text repeatedly interprets p < 0.2 as a “tendency”, which should not be used as evidence of an effect in a paper intended for publication. Please report exact n, define the multiple-testing procedure, and avoid inferential language for non-significant findings.
The pre-adsorption experiments are interesting, and they suggest that presentation of estradiol at the surface may matter. But the actual effects are often modest, system-dependent, and sometimes non-significant. Despite that, the discussion and conclusion describe the data as “compelling evidence” for direct hormonal interaction with the bacterial cell envelope. That language goes too far for the current experiments. The authors should present this as a plausible hypothesis supported by preliminay indirect evidence, not as a demonstrated mechanism.
Gene expression analysis: Methods mention ATP synthase as a reference gene while the supplementary table S1 lists 16S primers, so the normalization approach needs to be clarified. The discussion admits that paralog similarity complicates absolute quantification and that transcriptional patterns were “instable”. Consequently, conclusions about bacteriocin regulation should be softened and the analytical workflow described more transparently.

Minor

Supplementary Table S1 should read “forward” not “forvard”

Author Response

REVIEWER 2

Major

Question 1.Title and discussion frame C. acnes as part of the vaginal context, yet the introduction itself presents it mainly as a broader commensal that may be a transient vaginal component, citing only limited evidence for vaginal relevance. With only one skin-derived strain and one vaginal isolate, the manuscript should avoid broad statements about “the vaginal niche” (e.g., line 656) or “vaginal microbiota” unless those claims are explicitly restricted to this experimental model.

Answer: We sincerely thank the Reviewer for this important and well-justified comment. We fully agree that, given the limited number of strains and the inclusion of only one vaginal isolate, any broad generalizations about the vaginal niche or vaginal microbiota are unwarranted. We have carefully revised the Title, Discussion, and all relevant sections of the manuscript to ensure that claims regarding the vaginal context are explicitly restricted to our experimental model and do not imply broader ecological conclusions. Specifically, we have replaced phrases such as "the vaginal niche" and "vaginal microbiota" with more precise formulations (e.g., "under conditions mimicking certain aspects of the vaginal environment," "in this experimental system"), and we have added qualifying language to emphasize the preliminary nature of any niche-related interpretations. We believe these revisions address the Reviewer's concern appropriately.

So, we rewrote the abstract (LINES 19, 20-21, 24-26, 28-30,33-34, 39-40, 41, 42), the introduction (LINES 115-116, 120-124), the discussion (LINES 767-774, 780-783, 795-796. 797-798, 801-802, 803-804, 805-806, 809-810, 811-813, 814-818, 820-831, 837-839, 845-846, 846-847, 848-849, 851-852, 855-856, 857-867, 869, 870, 871, 872-874, 874-875, 876-877, 878-880, 882) and conclusion (LINES 886-887, 888-889, 890-891, 893-898, 898-899, 900-902). For instance, the Abstract now sounds like:

This pilot study investigated the effects of estradiol and clindamycin on mono- and dual-species biofilms of selected reference and clinical isolates of Lactobacillus gasseri and Cutibacterium acnes, including one vaginal isolate of C. acnes. Our findings demonstrate complex, strain-dependent effects of both compounds and their combinations. Estradiol inhibited biofilm formation in L. gasseri strains but exhibited divergent impacts on C. acnes isolates, stimulating the skin-derived strain while suppressing the vaginal isolate. The observation that pre-adsorbed estradiol tended to enhance its biological activity is consistent with, though does not prove, the hypothesis of a direct hormonal interaction with the bacterial cell envelope. Crucially, estradiol modulated the susceptibility of both species to clindamycin. At the working concentration selected, clindamycin susceptibility varied considerably between strains, with the antibiotic stimulating biofilm growthin skin-derived C. acnes HM514 biofilms. In dual-species communities, an apparent inversion of clindamycin activity was observed, suggesting that estradiol may alter antibiotic efficacy in a manner dependent oncommunity composition and strain identity. Furthermore, while transcriptional changes in bacteriocin genes were evident under hormonal and antibiotic pressure, these shifts did not consistently correlate with observed phenotypic antagonistic activity. These results underscore the limitations of traditional monospecies assays and highlight the importance of considering hormonal background, community context, and the substantial phenotypic variabilityamong individual microbial isolates when evaluating antimicrobial interventions

The conclusion sounds like:

In summary, this pilot study demonstrates that the impact of estradiol and clindamycin on lactobacilli and cutibacteria is highly contingent on strain-specific physiology. While estradiol inhibited biofilm formation in L. gasseri ATCC 33323 and MA4, it exerted divergent effects on the two C. acnes isolates tested, stimulating the skin-derived HM514 but suppressing the vaginal isolate EAB1. The apparent inversionof clindamycin activity within L. gasseri MA4 – C. acnes EAB1 dual-species biofilms further underscores the limitations of traditional monospecies assays in predicting polymicrobial outcomes. Furthermore, the tendency of surface-bound estradiol to enhance biological activity is consistent with the hypothesis of a direct hormonal interaction with the bacterial cell envelope, although this interpretation requires further experimental validation. These findings highlight that improving antimicrobial strategiesrequires a multifactorial approach that accounts for hormonal background, community context, and the substantial phenotypic variability among individual isolates.

Question 2. The figures are blurry and would benefit from higher resolution.

Answer: Thank you! We fixed the images, we made them more fulfilled, and we enhanced their resolution up to 600 dpi.

Question 3. Estradiol was dissolved in ethanol, and ethanol controls were included. However, the methods report the final ethanol concentration as “0.6% (v/v) or 0.1 M” while Fig. 11 reports ethanol as 1.7 x 10^-5 M, which is inconsistent and needs correction or explanation.

Answer: Thank you for this consistent remartk. It was actually a typo. Actual ethanol concentration was 0.01 M, we changed the values everywhere (LINES 155-156, 344, 615, 616; Figures 5,6,7,8, 12, 13, 14, Suplementary figures S1, S3, S4).

Question 4. The manuscript uses 0.5 ug/mL as a working concentration for interaction studies, but the results show that this concentration behaves very differently across strains: it inhibits L. gasseri, stimulates C. acnes HM514, and has mixed effects in dual-species communities. That makes it difficult to intrepret 0.5 ug/mL as a universally sub-inhibitory concentration. The authors should either justify this common-dose design much more explicitly or provide strain-specific MIC/MBIC-based reasoning.

Answer: Although 0.5 µg/mL clindamycin cannot be considered strictly sub-inhibitory for all strains tested, this concentration was selected as the working concentration for the following reasons. First, it represented the lowest level at which a consistent effect on microbial growth dynamics and/or biofilm formation was observed across all studied strains; depending on the strain, this effect manifested as either growth inhibition or stimulation. Second, the aim was to identify a concentration capable of reducing the growth rates of the strains by approximately 30–60%, as this degree of attenuation was expected to render the protective or sensitizing effect of estradiol more pronounced. In this range, the antibiotic impairs growth without fully suppressing it, thereby allowing the hormone's influence on antibiotic susceptibility to be assessed. It should be noted that the true sub-inhibitory concentration varies among the strains used; therefore, 0.5 µg/mL is referred to throughout the manuscript as the working concentration rather than a universally sub-inhibitory one. To improve clarity, the MIC values for each strain (ATCC 33323, MA4, HM514, and EAB1) will be mentioned in the revised manuscript.

Question 5: The manuscript does not clearly state the biological replicate number for each assay in the main methods, despite presenting many pairwise comparisons across strains, treatments, and communities. The use of Mann–Whitney tests across many comparisons also requires a clearer multiple-testing strategy. The paper mentions “p < 0.05 or q < 0.05 where appropriate” but does not explain exactly when or how correction was applied. In addition, the text repeatedly interprets p < 0.2 as a “tendency”, which should not be used as evidence of an effect in a paper intended for publication. Please report exact n, define the multiple-testing procedure, and avoid inferential language for non-significant findings.

Answer: Thank you very much. Indeed it was an editorial mistake. We changed q to p everywhere. Also, we added the N in the paragraph 2.10 (LINES 297-302):

All experiments were performed in at least three independent biological replicates. The experiments in 96-well plates were performed four times. The biofilm experiments on filters were performed six times. RNA isolation and differential gene expression experiments, as well as pre-adsorption of estradiol and antibacterial activity tests, were conducted in three independent replicates. Growth kinetics experiments were performed four times.

Question 6: The pre-adsorption experiments are interesting, and they suggest that presentation of estradiol at the surface may matter. But the actual effects are often modest, system-dependent, and sometimes non-significant. Despite that, the discussion and conclusion describe the data as “compelling evidence” for direct hormonal interaction with the bacterial cell envelope. That language goes too far for the current experiments. The authors should present this as a plausible hypothesis supported by preliminay indirect evidence, not as a demonstrated mechanism.

Answer: We sincerely thank the Reviewer for this fair and constructive assessment. We fully agree that our original Discussion overstated the strength of the conclusions that can be drawn from the pre-adsorption experiments, and we appreciate the opportunity to correct this. We have revised the relevant section thoroughly, replacing the term "compelling evidence" and similar assertive language with more measured phrasing. The surface-associated hypothesis is now presented as a plausible but preliminary interpretation, supported by indirect experimental observations. We have also explicitly acknowledged that the observed effects were often modest in magnitude and, in some cases, did not reach statistical significance, and that alternative explanations — such as altered local bioavailability of the hormone or changes in surface hydrophobicity — cannot be ruled out at this stage. We believe these revisions provide a more balanced and accurate representation of the current data.

We rewrote the discussion part (LINES 794-831): Furthermore, the fact that estradiol exerts its regulatory influence both when dis-solved in the medium and when pre-adsorbed onto the surface is consistent with the hy-pothesis supports the hypothesis of a mechanism involving receptor-like surface struc-tures in Gram-positive bacteria [26] although direct evidence for such structures remains lacking. As an alternative, the hormone could trigger a membrane stress response, as pre-viously demonstrated in P. aeruginosa [487], suggesting that estradiol may act by modu-lating membrane integrity or signaling pathways associated with surface sensing. Pre-adsorption of the hormone onto polystyrene surfaces tended to enhance its biological activitygenerally enhanced its biological activity compared to standard administration, particularly in dual-species biofilms, although the magnitude of these effects was often modest and varied. Notably, the finding that estradiol was more effective in a pre-adsorbed state strongly aligns with the surface-associated hypothesis, and raises the possibility thatsuggesting that its regulatory impacts are mediated, at least in part, are mediated through localized interactions at the cell-substrate interface. Adsorption from physiological saline produced the most stable effects, whereas MRS- or ethanol-based pre-adsorption increased the variability of the outcomesincreased the magnitude of effect fluctuations. Although in some cases the statistically significant differences were modest, the recurrence of this trend lends preliminary support to the idea that surface-bound es-tradiol may influencethe consistency of this phenomenon across various systems sup-ports the idea that surface-bound estradiol is a critical determinant of early adhesion and biofilm architecture, and in line with previous observationsthis correlated with previous data obtained in eukariotic models [49]. These data are suggestive of a direct signaling role for estradiol, but they do not exclude indirect mechanisms.Such interactions suggest that estradiol acts as a direct signaling effector rather than through indirect metabolic pathways. Moreover, surface-bound estradiol generally showed enhanced activity, partic-ularly in mixed-species biofilms, suggesting that these interactions are crucial for early adhesion and architecture. Finally, polystyrene is known to accumulate estradiol [50], which could increase its local bioavailability at the interface between bacteria and plates. However, little is known about changes in local hydrophobicity caused by hormone ac-cumulation. It is plausible that alterations in the local microenvironment could partly ac-count for the observed results. It is also known that the concentration of estradiol in vagi-nal tissues differs from that in blood plasma[50]. In postmenopausal women, it is typical-ly an order of magnitude lower [51-53]. In younger women, the concentration varies de-pending on the menstrual cycle [54]. Taken together, these findings provide indirect, pre-liminary evidence, that potential decrease may play a role in the fact that the concentra-tion of estradiol decreases upon its pre-adsorption onto the surface, thereby altering the effect of the hormone on biofilms, although definitive mechanistic studies will be required

Question 7: Gene expression analysis: Methods mention ATP synthase as a reference gene while the supplementary table S1 lists 16S primers, so the normalization approach needs to be clarified. The discussion admits that paralog similarity complicates absolute quantification and that transcriptional patterns were “instable”. Consequently, conclusions about bacteriocin regulation should be softened and the analytical workflow described more transparently.

Answer: We sincerely thank the Reviewer for this careful and constructive critique. We fully agree that the normalization strategy was not described with sufficient clarity and that some of our conclusions regarding bacteriocin regulation were phrased too definitively given the methodological constraints.

Regarding the reference gene: we apologize for the inconsistency between the Methods section and Supplementary Table S1. In the present study, 16S rRNA was used as the reference gene for normalization, as indicated in Table S1. The mention of ATP synthase in the Methods was an editorial error, which we have corrected in the revised manuscript. We acknowledge that 16S rRNA is not an ideal reference gene under all conditions; however, its expression remained relatively stable across our experimental treatments, and we have now explicitly stated this caveat in the Methods.

Concerning the analytical workflow and data interpretation: we have substantially revised the relevant Discussion paragraph to (i) describe the normalization and quantification procedure more transparently, (ii) explicitly acknowledge the limitations imposed by high paralog similarity on absolute quantification, and (iii) soften the claims regarding bacteriocin regulation by framing them as preliminary observations that require further validation at the protein and functional levels.

Now the discussion sounds like (LINES 843-868): Similarly, transcriptional changes in bacteriocin genes did not consistently correlate with phenotypic antagonism. Given the evolutionary conservation of bacteriocin systems in L. gasseri, the observed transcriptional shifts may reflect a broader regulatory response likely represent the broader regulatory response of the gassericin-producing loci, although this interpretation remains tentative. While high sequence similarity between certain paralogs complicates absolute quantification; therefore, the reported gene expression data should be viewed as indicative of overall trendsremains a challenge for absolute quantification, the consistent trends in gene modulation reflect a systemic metabolic shift under hormo-nal and antibiotic pressure. Furthermore, gene expression patterns showed considerable variability across biological replicatesgene expression patterns exhibited inherent insta-bility, likely reflecting the high spatial and physiological heterogeneity characteristic of mature biofilms. This transcriptional noise, in conjunction with the multilayered regula-tion of bacteriocins, suggests that clindamycinsupports the notion that clindamycin, be-yond its canonical role as a translation inhibitor, may act as a stress cuefunctions as a stress cue that elicits metabolic reprogramming rather than directly modulating bacteri-ocin peptide output. However, this hypothesis requires further experimental validation at the protein and functional levels.rather than directly regulating bacteriocin peptide out-put. Finally, the agar-block diffusion assay should be clarified as measuring net diffusible antibacterial activity released from agar-grown bacterial communities, rather than total bacteriocin synthesis. Reduced or absent inhibition zones, therefore, do not necessarily in-dicate absence of bacteriocin production, as antimicrobial compounds may be retained or neutralised within the colony biofilm matrix, associated cells, or agar block. Moreover, bacteriocins may be susceptible to degradation by endogenous proteolytic enzymes or to general instability in the growth medium, which can reduce their detectable activity de-spite increased transcription of the corresponding genes.

Minor

Question 8: Supplementary Table S1 should read “forward” not “forvard”

Answer: thank you. We changed it in the S1 supplementary materials

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors, thank you for the opportunity to read your interesting manuscript. The research was well conducted and a number of significant results were obtained.
However, before potential publication, some modification is needed.
General
The results are written too broadly. Fifteen figures related to the results, most of which have several parts, A - D, significantly burden the manuscript, reduce the fluency of the text and make it difficult for the reader to pay attention. I suggest that this part be significantly shortened and some of the figures and text be transferred to supplementary materials.

In the materials and methods, some additional information is needed regarding the names of the producers (lines 125, 127) and the full names of the agents (lines 129, 193)
The introduction, discussion and conclusion are excellently written.
Minor
Lines 53, 210: Sneathia, L. gasseri, C. acnes - italics
Line 218 and 237: lowercase letters
Delete lines 779 - 782

Author Response

REVIEWER 3

Dear authors, thank you for the opportunity to read your interesting manuscript. The research was well conducted and a number of significant results were obtained.

Answer: Thank you very much for such an evaluation of our work. We are appreciated.

Other reviewers had a lot of questions, and now the manuscript was rewritten significantly. We hope that now it become better.

However, before potential publication, some modification is needed.

General

The results are written too broadly. Fifteen figures related to the results, most of which have several parts, A - D, significantly burden the manuscript, reduce the fluency of the text and make it difficult for the reader to pay attention. I suggest that this part be significantly shortened and some of the figures and text be transferred to supplementary materials.

Answer: Thank you! We removed the section of antagonistic activity (paragraph 3.6.1.) to the Supplementary material. Now it contains Supplementary figure S4 (formerly figure 12). We added the following (LINES 688-690): “The bacterial strains studied also displayed a degree of mutual antagonism, which was observed even at the intra-species (strain) level. For a detailed account, see Supplementary Materials S4”.

In the materials and methods, some additional information is needed regarding the names of the producers (lines 125, 127) and the full names of the agents (lines 129, 193)

Answer: we corrected the issues mentioned (LINES 171-172, 173, 175,176, 179-180, 183-184, 195, 203, 205, 230-231, 236, 284-285).

The introduction, discussion and conclusion are excellently written.

Answer: Thank you! It is very pleasant!

Minor

Lines 53, 210: Sneathia, L. gasseri, C. acnes – italics

Answer: we fixed those issues. Thank you!

Line 218 and 237: lowercase letters

Answer: we fixed those issues. Thank you!

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors addressed the comments. It can be acceptable for publication.

Author Response

Thank you for the reviewing our manuscript!

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for your revised manuscript and for the effort invested in addressing the points raised. Most concerns have been resolved, however, the remaining issues below should be resolved before the manuscript can be considered for acceptance.

My original comment no. 5: The manuscript presents a large number of statistical comparisons across strains, treatments, and community types, yet the methods section still does not clearly specify whether and how a correction for multiple comparisons was applied. The authors should explicitly state if a correction procedure (e.g., Bonferroni, or another method) was applied, and if not, prominently acknowledge this as a limitation. Ideally, the respective applied statistical tests are mentioned in the figure legends.
In the provided pdf the replacement figures still look low resolution to me, but this might just be a pdf conversion artifact. Please ensure that the final manuscript includes high resolution figures.

Author Response

Response to the Reviewer. Round 2.

My original comment no. 5: The manuscript presents a large number of statistical comparisons across strains, treatments, and community types, yet the methods section still does not clearly specify whether and how a correction for multiple comparisons was applied. The authors should explicitly state if a correction procedure (e.g., Bonferroni, or another method) was applied, and if not, prominently acknowledge this as a limitation. Ideally, the respective applied statistical tests are mentioned in the figure legends.

Response: Thank you for the remark. We apologize for incorrect addressing in the Round 1. Indeed, we did not apply the Benjamini–Hochberg procedure or the Bonferroni correction to adjust the alpha level, and this can indeed be considered a limitation of our statistical analysis. However, the number of independent experiments was relatively small, and each experiment included a limited number of comparison groups. Specifically, the ethanol-treated samples and the antibiotic-treated samples were compared against the untreated control. The estradiol-treated samples and the samples treated with the combination of ethanol and clindamycin were compared against the ethanol-treated samples. The samples treated with the combination of estradiol and clindamycin were compared against the samples treated with clindamycin and ethanol. Furthermore, we aimed to avoid type II errors.

Also, we added information about Mann-Whitney U-test in every legend where it was appropriate (LINES 355-356; 393-394; 434; 455-456; 486-487; 503-504; 525-526; 542-543; 564-565; 588; 635-636; 715; 741; 760-761). The added fragment is “Statistical analysis was performed using the Mann–Whitney U test.”

We added some text in the Paragraph. 2.10 (LINES 311-319):

The lack of multiplicity adjustment, such as the Benjamini–Hochberg or Bonferroni correction, may be viewed as a weakness of the statistical approach. However, the number of independent experiments was relatively small, and each experiment included a limited number of comparison groups. Specifically, the ethanol-treated samples and the antibiotic-treated samples were compared against the untreated control. The estradiol-treated samples and the samples treated with the combination of ethanol and clindamycin were compared against the ethanol-treated samples. The samples treated with the combination of estradiol and clindamycin were compared against the samples treated with clindamycin and ethanol. Furthermore, we aimed to avoid type II errors.

In the provided pdf the replacement figures still look low resolution to me, but this might just be a pdf conversion artifact. Please ensure that the final manuscript includes high resolution figures.

Response: Thank you. Indeed, it was the PDF-conversion limitation, because there was a serious resolution decrease even in the template logo. We assure that all the figures are 600 DPI and have perfect resolution.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors responded to all queries and significantly clarified and improved the manuscript, which can be published in this form.

Author Response

Thank you for reviewing our manuscript!