Integrated Metabolomics and Targeted Gene Expression Profiling Reveal the Arginine–Anthocyanin Axis in Pomegranate Aril Paleness Disorder

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions in track changes in the re-submitted files

. Point-by-point response to Comments and Suggestions

Comments 1: The experimental design suffers from a fundamental confounding between apple proliferation (AP) severity, cultivar, and environmental conditions that undermines the causal interpretation of the results. Specifically, DN, KN, and KW samples were collected in Mayami, whereas TW samples originated from Torud, a climatically distinct and warmer region. Nevertheless, throughout the manuscript TW is treated as representative of the “severe AP” stage, and the observed metabolic and transcriptomic differences are largely attributed to disease progression. Because TW simultaneously differs in genotype and environment, it is not possible to disentangle whether the reported changes are driven by AP severity or by uncontrolled confounding factors. This limitation directly affects the central hypothesis of the study (the arginine–anthocyanin axis) and, in its current form, prevents the proposed mechanistic conclusions from being supported.

Response 1: We understand the reviewer’s concern regarding the potential confounding of genotype and environment. However, we would like to clarify that all four cultivars used in this study are naturally red-ariled with high anthocyanin content under normal conditions. Historically, even the TW cultivar produced fully pigmented fruits in the Torud region before the emergence of AP. In recent years, however, the same orchard exhibited severe AP symptoms, leading to its removal—confirming TW’s genetic susceptibility. Similarly, the KN trees in Mayami are from a 15-year-old orchard that initially yielded normal red fruits. Thus, the gradient of AP symptoms across cultivars reflects differential genetic susceptibility, further exacerbated in warmer climates (as in TW). This genotype-by-environment interaction is consistent with field observations of AP progression, and the inclusion of TW provides a realistic representation of severe AP under predisposing conditions. We have added a clarifying note in the Methods section (lines 113–117) to address this point.

Comments 2: (lines 113–114 and 122–123). There is a serious methodological inconsistency in the description of the LC–MS instrumentation, with two different platforms (AB Sciex 3200 and API 6500 QTRAP) reported without clarification. These instruments differ substantially in analytical performance and acquisition capabilities, and this discrepancy raises major concerns regarding the reproducibility of the metabolomic data. As the core results of the manuscript rely heavily on these analyses, the study cannot be properly evaluated in its current state without a clear, unambiguous, and technically coherent correction of this section.

Response 2: We sincerely apologize for this oversight, which occurred during the preparation of the manuscript, and we confirm that all metabolomic data were generated using a single, consistent LC-MS platform, ensuring reproducibility and reliability of the results. The corrected description now accurately reflects the methodology employed.

Comments 3: (lines 85–90, 215–217, 26–29).

The definition of experimental groups and AP severity is internally inconsistent throughout the manuscript. KW is described as moderately affected in the plant material section, yet later appears as a severe AP group in figures and results. In addition, the sampling location “Torud” is spelled inconsistently in several forms, including in the abstract. These issues go beyond minor editorial errors, as they directly affect the interpretation of group comparisons and reflect insufficient conceptual control of the experimental framework. Such inconsistencies must be resolved before the manuscript can be reconsidered.

Response 3: Thank you for highlighting the inconsistencies in the definition of AP severity groups and the spelling of sampling locations. We have carefully reviewed the entire manuscript and corrected these errors.

Comments 4: (lines 135–140).

The criteria used to define differential metabolites are not supported by an adequate statistical description or validation of the underlying multivariate model. The use of VIP scores without specifying the model type, number of components, or validation procedures (e.g., permutation testing) is problematic, particularly in high-dimensional metabolomic analyses prone to overfitting. Furthermore, no correction for multiple testing is reported at the univariate level. In the absence of these safeguards, the robustness of the selected metabolites and subsequent pathway enrichment analyses is questionable.

Response 4: We appreciate the reviewer’s concern regarding statistical rigor. In fact, OPLS-DA was used to generate the VIP scores, with model validity assessed via 200 permutation tests (all p-values for Q² < 0.05). The model used two predictive and one orthogonal component, selected based on cross-validation. We regret that these details were not included in the original manuscript and will add them to the revised Methods section.

Comments 5: (lines 321–327, 133–134).

The manuscript reports the tentative identification of a large number of metabolites but does not provide objective criteria to assess the confidence of these annotations. No information is given regarding mass accuracy, MS/MS fragmentation, use of authentic standards, or internationally accepted levels of identification. Because the biological conclusions rely on specific metabolic pathways, this lack of transparency represents a substantial weakness that limits the overall reliability of the study.

Response 6: We fully acknowledge the reviewer's concern regarding the confidence level of metabolite annotations. We agree that without MS/MS fragmentation data or authentic standards, identifications remain tentative. However, we wish to clarify the following points:

First, all metabolite annotations were performed by matching accurate mass and retention time against the KNApSAcK database, with an additional filter restricting matches to metabolites previously reported in pomegranate (Punica granatum) based on published literature [19,20]. This targeted approach reduces the risk of false-positive annotations.

Second, our primary objective was not to discover novel compounds, but to compare relative metabolite abundance between healthy and AP-affected samples under identical analytical conditions. All comparisons were made within the same LC-MS run, using consistent peak integration and alignment parameters. Thus, although absolute identification confidence is limited, the differential patterns reported are robust and reproducible across biological replicates. We can made the raw metabolomic data of detected metabolites and their peak intensities for all across the four pomegranate type as suplementary data

Comments 6: (lines 188–208).

The interpretation of the FTIR results is overly conclusive relative to the evidence presented. Spectral preprocessing steps are not described, and no multivariate chemometric analysis is applied to support group discrimination. The inferences drawn regarding nitrogen metabolism, antioxidant capacity, and biochemical composition therefore lack appropriate statistical support and, in their current form, should be considered speculative. This section requires substantial methodological and interpretative revision.

Response 6: We appreciate the reviewer’s concern regarding the FTIR analysis. Spectral preprocessing steps and multivariate statistical modeling were not included, as the FTIR analysis was intended only to provide initial, qualitative insights into gross compositional differences between healthy and AP-affected samples. Nonetheless, we believe the FTIR data offer useful supporting information that aligns with the metabolomic findings—particularly regarding reductions in phenolic and carbohydrate-related functional groups in affected samples. The spectra were generated from three biological replicates per group, and the presented trends were reproducible. However, we fully respect the reviewer’s judgment. If the reviewer considers this section insufficiently supported or not essential to the core conclusions of the study, we are prepared to remove it from the manuscript upon request.

Comments7: (lines 91–92, 173–175, 183–185).

The gene expression analysis contains significant ambiguities regarding the number of biological replicates and the statistical approach employed. The discrepancy between the reported number of replicates and the use of multiple t-tests across genes and experimental groups raises concerns about the validity of the results. Without an appropriate statistical framework (e.g., ANOVA with correction for multiple comparisons) and a complete description of technical controls, the qPCR data do not provide a sufficiently solid basis for the transcriptomic conclusions.

Response 7: We appreciate the reviewer’s concern regarding the statistical analysis of gene expression data. As stated in the Materials and Methods section (lines 200–203) , three independent biological replicates were used for each cultivar, and qPCR was performed with three technical replicates per biological sample. Relative expression was calculated using the 2^–ΔΔCt^ method, and Student’s t-test was employed to compare each affected cultivar (KN, KW, TW) individually with the healthy control (DN). This pairwise comparison approach is commonly used and widely accepted in plant gene expression studies, including those on pomegranate fruit, and we believe it provides valid statistical support for our conclusions. However, we fully respect the reviewer’s judgment. If the reviewer considers it necessary, we are prepared to re-analyze the data using one-way ANOVA with a post-hoc test (e.g., Tukey’s HSD) and apply appropriate multiple comparison corrections. We will update the manuscript and figures accordingly and leave the decision to the reviewer’s discretion

Comments 8: (lines 434–436).

The manuscript proposes a mechanistic model in which reduced PgADS expression leads to arginine accumulation; however, no direct quantitative evidence is presented to demonstrate such accumulation. Given that this inference represents a central pillar of the proposed model, its current support is largely theoretical rather than data-driven. Without explicit and statistically validated measurements of arginine and related metabolites, this conclusion remains insufficiently substantiated.

Response 8: We agree with the reviewer that direct evidence of arginine accumulation should be presented. In the pathway enrichment analysis of downregulated metabolites, arginine was among the metabolites significantly reduced in AP-affected cultivars. This finding presented explicitly stated in the Results section. The reduced expression of PgADS is therefore consistent with disrupted arginine catabolism, and we have revised the Discussion to reflect this alignment more clearly.

Comments 9: (lines 153–161, 390–410).

The use of Arabidopsis-based gene interaction networks to infer functional relationships in pomegranate introduces a significant level of extrapolation that is not adequately justified. Criteria for orthology assignment, interaction confidence thresholds, and evidence types are not reported, yet the results are discussed as if they reflected direct biological processes in the studied system. While potentially useful as an exploratory approach, this strategy cannot, on its own, support the mechanistic claims presented.

Response 9: We appreciate the reviewer's concern regarding the use of Arabidopsis interaction networks. We fully acknowledge that this approach represents an exploratory, predictive step rather than direct evidence in pomegranate.  Since a species-specific protein-protein interaction network for pomegranate is not currently available in public databases, we used Arabidopsis thaliana as a well-annotated model plant to generate hypotheses about potential functional connections between arginine metabolism and anthocyanin biosynthesis. This is a widely accepted strategy in non-model plants. Importantly, all predicted candidate genes were verified against the pomegranate genome, and only confirmed pomegranate orthologs (the seven genes analyzed in this study) were selected for experimental validation via qPCR. Thus, the network analysis was used only as a discovery tool, and all mechanistic conclusions are based on direct gene expression data from pomegranate tissues, not on the Arabidopsis network itself.

Comments 10: (lines 362–363, 386–389, 412–419).

The inconsistent numbering of figures creates substantial confusion and hinders critical evaluation of the manuscript. The reuse of figure numbers for different analyses suggests a lack of basic editorial revision that must be corrected prior to any further consideration of the work.

Response 10: We apologize for this oversight and appreciate the reviewer's careful reading. We have carefully reviewed all figures and their citations throughout the text and corrected the numbering accordingly.

Minor comments

Comments 1: (lines 16–17, 28–29). Several typographical and encoding errors are present in the author information and abstract, including corrupted characters and duplicated words, which should be carefully corrected.

Response 1: Thanks, The typo errors was corrected

Comments 2: (lines 177–179). The term “cNDA” appears instead of “cDNA,” constituting a technical error that should be corrected to avoid confusion.

Response 2: The term "cNDA" has been corrected to "cDNA" 

Comments 3: (lines 487–488, 568–569).

Minor formatting issues, including stray characters and incorrect word spacing, are present and require editorial correction.

Response 3: It was corrected.

Comments 4: (lines 422–424, 514–521, 539–545). Inconsistencies are observed in gene nomenclature and associated enzymatic functions, including confusion between dehydratase and dehydrogenase activities. Given the importance of these genes to the proposed model, nomenclature and functional assignments must be carefully reviewed and harmonized across the text and figures.

Response 4: The nomenclature and functional assignments was carefully checked and corrected

Comments 5: (lines 223–224, 329–334). The relationship between detected “features” and the number of metabolites reported per group is not clearly explained. Clarification is needed regarding metabolite inclusion criteria and the handling of missing values

Response 5: We have added these clarifications to the Results sections accordingly.

Comments 6: (lines 554–555). The claim that the results enable the identification of pre-symptomatic diagnostic markers is not supported by the experimental design. This statement should be removed or substantially tempered.

Response 6: The sentence has been revised