Story about honest mistakes: The cyanobacterium Synechocystis has a promiscuous Entner-Doudoroff (ED) aldolase but no functional ED pathway

Reviewer #1 (Public review):

Strengths:

Thorough reanalysis of the experimental results obtained in previous studies, which led to the publication of the PNAS paper in 2016.

New experimental evidence to confirm that enzymes previously considered as participating in the ED, actually are not catalyzing the ED biochemical reactions, but are involved in other metabolic pathways. Also, the authors completely discarded the occurrence of the GDH/GK shunt in Synechocystis PCC 6803. Generally speaking, the manuscript is very clearly written, with a precise description of the previous findings, the mistakes which took place in the 2016 paper, and the strategies they have used to address those issues, in order to reach a thoroughly revised vision of the glucose metabolic pathways in Synechocystis PCC 6803. In this regard, the drawings shown in Figures 1 and 7 are very helpful for the reader to follow the story and understand the possible metabolic transformations depending on the working hypothesis.

Also, I commend the authors for openly describing previous mistakes. In this paper, they reassess past observations under the light of more recent findings, and to integrate the information in this manuscript. The scientific conclusions are solid and very interesting, and besides they use the opportunity to offer valuable advice to researchers. This is especially focused on the importance of careful biochemical characterization of enzymes, which should always be carried out when studying proteins which have been identified as a specific enzyme on the basis of sequence homology. In a similar way, they found that an insertional mutant was the cause for the absence of specific metabolites, which had been attributed to particularities of a metabolic pathway in that mutant, when it was actually due to a nucleotide insertion; given that currently, genome sequencing is an affordable technique, this kind of mistakes can now be easily prevented by confirming the correct generation of the mutant by DNA sequencing, as proposed by the authors in a recently published preprint (Theune et al, bioRxiv 10.64898/2026.04.08.717167).

Weaknesses:

The authors propose that EDA might be involved in the PEP-pyruvate-OAA node, or in the proline metabolism, but this requires further experimental work for clarification; what their results indicate clearly is that this enzyme is not actually catalyzing the transformation of KDPG to GAP, which is the second specific enzyme of the ED pathway. But the real physiological function in this cyanobacterium is still unconfirmed.

Another aspect which could be improved is that the recombinant expression of some genes was carried out in E. coli; even if this is a useful and valid research strategy, in studies like this (where there is a strong focus on the physiological function of enzymes in the original organism, Synechocystis PCC 6803), I think it would have been more appropriate to express the 6803 genes in another cyanobacterium easily amenable for genetic transformation and gene expression, which would produce the protein in a physiological environment more similar to another cyanobacterium (compared to E. coli, which is an heterotrophic bacterium). I am not sure this would change any of the obtained results, but certainly would confer additional robustness to the enzymatic results.

Comments on revised version.

The authors have provided satisfactory replies to all my suggestions and corrections, and I have no further changes to suggest.

Reviewer #2 (Public review):

Summary:

The study presents novel results on the presence of the Entner Doudoroff pathway in Synechocystis sp. PCC 6803. In contrast to an earlier study, compelling evidence is given that this strain lacks both an ED pathway and a glucose dehydrogenase/glucokinase bypass but contains a promiscuous aldolase, which also decarboxylates oxaloacetate and cleaves 2-keto-4-hydroxyglutarate (as it occurs in proline degradation). The study concludes with successfully reconciling data of different studies and with lessons learned from the previous misconception.

Strengths:

Solid biochemical data is presented to reconcile contradicting data of earlier studies and to serve as basis for disclosing possible functions of a promiscuous aldolase. Earlier misconceptions and lessons to be learned are well discussed.

Weaknesses:

The materials and methods section is rather lengthy, suffering from a lack of conciseness and repetitions, and nevertheless misses some specifications.

Comments on revised version.

The materials and methods section has been significantly improved. The revised manuscript is now recommended for publication as it is.

Author response:

The following is the authors’ response to the original reviews.

Public Reviews:

Reviewer #1 (Public review):

Summary:

Some of the authors proposed in a PNAS paper in 2016 the occurrence of the EntnerDoudoroff (ED) pathway in cyanobacteria and plants, on the basis of several lines of biochemical and genetic evidence. However, more recent results indicated that one of the two specific enzymes of the ED pathway (EDD) is missing in Synechocystis PCC 6803. The authors carried out additional experiments, which demonstrated that EDD is missing, and one of the enzymes (ED aldolase) is a promiscuous enzyme which seems to be involved in proline metabolism and is not actually participating in the ED pathway as initially believed. The results described in this paper are strong evidence that this new interpretation is appropriate, and therefore, it corrects the previous proposal, providing an honest description of the reasons why the authors had reached the wrong conclusion about the existence of the ED pathway in cyanobacteria and plants.

We thank Reviewer 1 for the summary and comments. Based on our finding that EDA is a promiscuous aldolase that, in addition to the cleavage of KDPG to GAP and pyruvate (a reaction of the ED pathway) catalyzes other reactions in vitro, we proposed potential in vivo functions of EDA, including its involvement in proline metabolism. However, these assumptions require further experimental testing. We do not yet have definitive findings regarding the function of the promiscuous aldolase EDA in Synechocystis in vivo, but respective studies are currently underway.

Strengths:

Thorough reanalysis of the experimental results obtained in previous studies, which led to the publication of the PNAS paper in 2016.

New experimental evidence to confirm that enzymes previously considered as participating in the ED actually are not catalyzing the ED biochemical reactions, but are involved in other metabolic pathways. Also, the authors completely discarded the occurrence of the GDH/GK shunt in Synechocystis PCC 6803. Generally speaking, the manuscript is very clearly written, with a precise description of the previous findings, the mistakes which took place in the 2016 paper, and the strategies they have used to address those issues, in order to reach a thoroughly revised vision of the glucose metabolic pathways in Synechocystis PCC 6803. In this regard, the drawings shown in Figures 1 and 7 are very helpful for the reader to follow the story and understand the possible metabolic transformations depending on the working hypothesis.

Also, I commend the authors for openly describing previous mistakes. In this paper, they reassess past observations in light of more recent findings and to integrate the information in this manuscript. The scientific conclusions are solid and very interesting, and besides, they use the opportunity to offer valuable advice to researchers. This is especially focused on the importance of careful biochemical characterization of enzymes, which should always be carried out when studying proteins which have been identified as a specific enzyme on the basis of sequence homology. In a similar way, they found that an insertional mutant was the cause of the absence of specific metabolites, which had been attributed to particularities of a metabolic pathway in that mutant, when it was actually due to a nucleotide insertion; this could have been easily prevented by confirming the correct generation of the mutant by DNA sequencing.

We thank the reviewer for this kind comment. We agree that biochemical characterization of enzymes as well as DNA sequencing to check deletion mutants, are important and valuable tools. As outlined in the manuscript and additionally in more detail in a recently submitted article, which is available at bioRxiv (Theune et al. 2026, doi: https://doi.org/10.64898/2026.04.08.717167) and is currently under review at PLOS One, we suggest that genome sequencing of deletion mutants in combination with complemented strains as controls are required to minimize the risk of misinterpretation based on secondary mutations (1). During the early stages of our research on the ED pathway, and later as well when we were already trying to resolve the conflicting results that had accumulated concerning the ED pathway, genome sequencing for Synechocystis mutants was not affordable as a routine procedure (2-4). Therefore, we could not have easily prevented this misconception based on this technique at that time. However, we strongly encourage genome sequencing of deletion mutants (in combination with complemented strains) as routine procedures these days (1).

Weaknesses:

The authors propose that EDA might be involved in the PEP-pyruvate-OAA node, or in the proline metabolism, but this requires further experimental work for clarification; what their results indicate clearly is that this enzyme is not actually catalyzing the transformation of KDPG to GAP, which is the second specific enzyme of the ED pathway. But the real physiological function in this cyanobacterium is still unconfirmed.

As stated above and in the manuscript, we agree that the in vivo role of EDA requires further experimental work which is in progress. However, our results demonstrate that EDA splits KDPG into GAP and pyruvate in vitro, but we assume that this reaction does not play a role in vivo due to the absence of its substrate.

Another aspect which could be improved is that the recombinant expression of some genes was carried out in E. coli; even if this is a useful and valid research strategy, in studies like this (where there is a strong focus on the physiological function of enzymes in the original organism, Synechocystis PCC 6803), I think it would have been more appropriate to express the 6803 genes in another cyanobacterium easily amenable for genetic transformation and gene expression, which would produce the protein in a physiological environment more similar to another cyanobacterium (compared to E. coli, which is an heterotrophic bacterium). I am not sure this would change any of the obtained results, but it certainly would confer additional robustness to the enzymatic results.

Synechocystis is easily amenable to genetic manipulation, and we agree that expression and purification of all enzymes from this host would have been ideal. However, the first characterization of Synechocystis EDA was performed with proteins that were purified from Synechocystis and showed activity on KDPG at comparable rates as proteins that were purified from E. coli in this study (2). Moreover, most biochemical characterizations of EDAs from archaea, bacteria and plants were performed after recombinant expression in E. coli and yielded highly active enzyme as in the case of Synechocystis is this study (5-7). Therefore, we currently have no reason to worry that the expression in E. coli might affect the enzymatic activity of EDA. The main reason for utilizing E. coli as an expression strain in this study was to gain higher yields of protein for in-depth analyses.

Bibliography:

I think the list of papers used in this manuscript is complete and up to date. However, I do miss recent papers which addressed one aspect that was proposed in the original 2016 PNAS paper: the authors wrote, "We therefore suggest that Prochlorococcus might oxidize glucose via the ED pathway under mixotrophic conditions, as shown for Synechocystis." Recent studies checked this hypothesis and have shown that the ED pathway seems to be also missing in Prochlorococcus and marine Synechococcus, and I think this manuscript is a good place to cite them, since these results are consistent with the findings of this paper.

We will include a references from Moreno-Cabezuelo et a. 2023 (DOI: 10.1128/spectrum.03275-22) in which the proteomes of three marine Prochlorococcus and three marine Synechococcus strains were investigated upon exposure to glucose (8). Protein levels of EDA were either downregulated or not affected while proteins involved in OPP pathway and CBB cycle were upregulated. The authors of this study conclude that this indicates that the latter processes rather than the ED pathway are involved in photomixotrophy in these strains. However, flux analyses are still missing.

Reviewer #2 (Public review):

Summary:

The study presents novel results on the presence of the Entner-Doudoroff pathway in Synechocystis sp. PCC 6803. In contrast to an earlier study, compelling evidence is given that this strain lacks both an ED pathway and a glucose dehydrogenase/glucokinase bypass but contains a promiscuous aldolase, which also decarboxylates oxaloacetate and cleaves 2-keto-4-hydroxyglutarate (as it occurs in proline degradation). The study concludes with successfully reconciling data from different studies and with lessons learned from the previous misconception.

Strengths:

Solid biochemical data are presented to reconcile contradicting data of earlier studies and to serve as a basis for disclosing possible functions of a promiscuous aldolase. Earlier misconceptions and lessons to be learned are well discussed.

Weaknesses:

The materials and methods section is rather lengthy, suffering from a lack of conciseness and repetition, and nevertheless misses some specifications.

We thank Reviewer 2 for the kind summary and comments and will improve the materials and methods part accordingly in a revised version.

Recommendations for the authors:

Reviewer #1 (Recommendations for the authors):

Some additional aspects that could be improved:

(1) L182-184: Are there any known in vitro attempts to determine whether some DHADs can accept 6PG as substrate? If not, did the authors check this possibility in the lab?

To the best of our knowledge, as mentioned in the manuscript, some DHADs are tested for their activity towards gluconate and some other substrates but not 6PG. It was suggested that since gluconate is smaller it might fit in the catalytic site of DHADs normally occupied by DHIV (7). We discussed this point in the manuscript (Line 177-185).

In this study we tested the DHAD Slr0452 from Synechocystis and DHAD from Synechococcus with 6PG as substrate but the enzyme did not catalyze 6PG dehydration. The result for Slr0452 is shown in Figure S3.

(2) L234-241: This paragraph shows how important it is to avoid relying only on sequence alignments to assign functions to proteins (and enzymes in particular). The authors stress this idea elsewhere in the paper, but I think it should receive even more attention in a manuscript like this. Physiological characterization of the protein function is paramount, especially in the case of enzymes. Also, the GDH1 overexpression mutant was done in E. coli; this adds an additional layer of uncertainty, since the protein processing in E. coli might not be entirely identical to that carried out in cyanobacteria, as I mentioned above. This, in turn, could be one of the reasons for not finding the expected enzymatic activity. This comment is also relevant for the results shown in Table 1 (page 12).

As outlined following lines L234-241 we tested crude cell extracts from Synechocystis WT, a Synechocystis strain overexpressing putative GDH1 (Sll1709) and a Dzwf deletion mutant which can be assumed to upregulate a GDH/GK bypass if present in Synechocystis for GDH activity but did not find any. This strongly indicates that sll1709 does not code for an active GDH in Synechocystis. We thereafter also overexpressed GDH1 in E. coli and again could not detect any GDH activity. In case of GDH1, enzyme activity was therefore tested both in Synechocystis and E. coli and yielded similar results.

(3) L309-310: "we currently have no explanation for the gluconate that was detected in previous IC-ESI-MSMS measurements in Synechocystis". This is a serious issue, given how important this evidence was for the conclusions of the 2016 PNAS paper and the hypothesis of the ED pathway in cyanobacteria. I would suggest that the authors provide some possible explanations, even if it is based on studies from other teams.

It would be very speculative and therefore in our eyes not helpful to search for explanations in this case as indeed the measurements were done in another lab. One explanation might be that 6P gluconate got dephosphorylated and yielded gluconate as an artifact. However, as we have no experimental validation for this idea and furthermore cannot test it. We therefore prefer to not further comment on this aspect.

(4) L347: The presence of an insertion in the sequence of zwf in the ∆gnd mutant is a welcome explanation for the observed results: abolition of 6PG production in this mutant. This is another important message to stress in this manuscript: construction of mutants should always be double checked by DNA sequencing of the relevant genomic regions, to ensure that these kinds of side problems do not appear, leading to confusing results.

We agree with the reviewer and would get even one step further and rather suggest combining whole-genome sequencing with complemented mutants as it is difficult to know all relevant genomic regions that should be sequenced. We discuss this issue in even more detail in another manuscript that is currently available as online preprint in bioRxiv and is under review (9). This work was now added as a citation in line 566 and the end of the following statement (line 564-566): Routine complementation of deletion mutants and sequencing of selected genes or the entire genome are effective means of identifying secondary mutations that can lead to misleading phenotypes (9).

(5) L418-419 and Table 2: The results observed for the authors (i.e., that EDA could also catalyze reactions with OAA and KHG, albeit with substantially lower catalytic efficiency than with KDPG), is a matter for concern: if their hypothesis is correct (meaning that this EDA is fundamentally involved in the PEP-pyruvate-OAA node and/or proline metabolism, and not with the ED pathway), then why should it keep in the evolution of these organisms such a strong preference for KPDG, when it is not being used physiologically for the ED pathway)? Furthermore, the Km for KDPG is lower than for OAA or KHG, leading to a very big difference in the Kcat/Km values.

To solve these questions further respective studies are underway. As EDD is absent from Synechocystis no KDPG should be available in the cells so that catalytic activity on KDPG should be irrelevant in vivo. The in vivo role of Eda requires further clarification.

Hereafter, I will mention some aspects, following the instructions of eLife, which are related to suggestions for improved experiments/data/analyses, improvements of writing and presentation, and minor corrections to text/figures.

(1) L81: I would modify the text to "Accordingly, this raises further questions...".

Thanks for this hint. We modified the text accordingly.

(2) L189: Add "pages" after "following".

Thanks for pointing this out. We replaced “following” by “below”.

(3) Page 7: The whole beginning of the Results section is actually more discussion than description of results, and the first mention of figures appears in L207 of page 8. Given the content of the paper, I think the authors might reconsider using "Results and Discussion" rather than different, specific sections for Results and Discussion. This is one of the papers where I think the combined use of both makes sense and will allow an easier understanding of the message.

We thank the reviewer for this valuable suggestion and changed the heading to Results and Discussion.

(4) L181: Add reference regarding the llvD/EDD superfamily.

We added the following references in lines 172-176 and 185-189:

(1) Melse, O., Sutiono, S., Haslbeck, M., Schenk, G., Antes, I., & Sieber, V. (2022). Structure Guided Modulation of the Catalytic Properties of [2Fe− 2S]-Dependent Dehydratases. ChemBioChem, 23(10), e202200088.

(2) Ren, Y., Vettenranta, E., Penttinen, L., Jänis, J., Rouvinen, J., & Hakulinen, N. (2025). The engineered dimer of L-arabinonate dehydratase from Rhizobium leguminosarum bv. trifolii: The role of intersubunit interactions in IlvD/EDD family. Biochemical and Biophysical Research Communications, 757, 151610.

(3) Ahmed, H., Ettema, T. J., Tjaden, B., Geerling, A. C., Van Der Oost, J., & Siebers, B. (2005). The semi-phosphorylative Entner–Doudoroff pathway in hyperthermophilic archaea: a reevaluation. Biochemical Journal, 390(2), 529-540. ff

(4) Bräsen, C., Esser, D., Rauch, B., & Siebers, B. (2014). Carbohydrate metabolism in Archaea: current insights into unusual enzymes and pathways and their regulation. Microbiology and Molecular Biology Reviews, 78(1), 89-175.

(5) Figure 2: The data shown in column plots in Fig 2A, B and C, and 4B, could be presented in tables, which would save space while providing the same information: basically, very little/no activity in some cases vs high levels of activity in others.

We would like to keep the column plots as we still think that they visualize our data well.

(6) L255 "Unfortunately, we were not able to overexpress putative GDH2". It would be interesting to give more details about the possible reasons for this fact.

We tested different growth conditions for recombinant GDH2 expression. The expression culture was incubated at 37 °C for 3 hours as well as overnight at 18 °C for overnight after induction. Both experiments did not resolve the expression problem.

(7) L409-410: I think this sentence should include a brief part explaining the kind of essay used to test this activity.

We added now that the LDH-coupled continuous assay was used (see line 403).

(8) Figure 6E: Please give the specific activity in U/mg, as in Fig 6F, instead of percents.

100% is given in U/mg units in the figure legend as “control without effector (100 %; specific activity of 4.3 U/mg)”. For easy comparison of effectors, the relative activity (%) is often used. We would therefore prefer to keep the current data presentation.

(9) L576: Provide the origin of the utilized PCC 6803 strain, given there is a certain level of variability in this strain (glucose tolerance, etc). Also, even if there are some methods which are very widely used, I think the Materials and Methods section should either properly describe them or else cite the source. For instance, BG11 medium is mentioned, but no further information is given.

We included the information that the glucose-tolerant Synechocystis strain was utilized and added the receipt of and a citation for BG11 medium (10).

(10) L582 Generation of mutants: This section mentions the Gibson Assembly cloning method, but I miss further information to allow the reader to reproduce the methodology with as many details as possible, or at least cite papers which do so.

We added a reference in which Gibson Assembly is described (11). Together with the primers listed in Table S3 the generation of mutants is now reproducible.

(11) L609 Please give information in g, not rpm, for centrifugation. Also, mention the model and brand of the centrifuge and rotors used. Also, immunoblotting is very loosely described. This is also valid for other sections, for instance, L618, L636.

We now added the following information: Cells were harvested by centrifugation at an RCF (relative centrifugal force) of 3,992 x g in a Beckman Coulter with a JLA-8.1000 Rotor for 20 minutes at 4°C. We now added a reference (12) in which immunoblotting is described in more detail.

(12) L613: Describe the "small scale purification".

We now added the information that the small-scale purification was performed using a 50-ml aliquot of the large culture which was treated as described below for the remaining sample.

(13) L619-620: Describe the composition of the lysis buffer.

The composition of the lysis buffer is already described as follows: lysis buffer (50 mM NaPO₄ pH=7.0; 250 mM NaCl; 1 tablet complete protease inhibitor EDTA-free

(Roche) per 50 mL)

(14) L691: Specify which amounts of auxiliary enzymes in coupled enzymatic assays were used.

Thanks for pointing this out. We have now integrated the information that 1U of each of the auxiliary enzymes was utilized in the coupled enzymatic assays.

(15) L716: The authors mention several times using a "double beam spectrophotometer". Please provide the model and brand.

Model and brand were now added for the double-beam spectrophotometer (Uvikon 810, Kontron, Augsburg, Germany).

(16) L717 and 718: define "∆absorption".

In line 715, the information is given that absorption was measured at 340 nm, "∆absorption" is accordingly the ∆absorption at 340 nm. This information was added.

(17) L723: "Synechocystis" should be in italics.

Synechocystis is now written italics.

(18) L749-759: This section should be described in more detail: preparation of protein extracts, SDS, immunoblotting, etc, or cite references of the same team where these methods were properly described.

In this section the listed methods are already described in detail.

(19) 798-799: "frozen cell pellets". Please provide numbers to specify the amount of material used.

Thanks for pointing this out. We now added the information that frozen cell pellets with a wet weight of 2.4 g wet weight were resuspended.

(20) L871-872: "It was ensured that auxiliary enzymes were not rate-limiting. One unit (1 U) of enzyme activity is defined as 1 µmol substrate consumed or product formed per minute" is repeated several times in the manuscript (L 907-909, L936-938). I would advise using it the first time, and on other occasions, refer to the same conditions as described above.

We have accordingly circumvented the repetition of 1 U definition from the manuscript.

Reviewer #2 (Recommendations for the authors):

(1) Interpunctuation, especially comma placement, should be improved.

We improved interpunctuation, especially comma placement to the best of our knowledge.

(2) Line 63: delete the first "which".

“Which” was deleted.

(3) Lines 81/82: revise sentence.

We revised the sentence to: Accordingly, this raises further questions about the presence of the ED pathway in cyanobacteria and plants.

(4) Line 164: "presumed" instead of "presumes".

The word was changed accordingly.

(5) Line 228: by others? especially in reference 1?

We deleted by others as the reference is given.

(6) Line 240: "or" instead of "no".

“no” was replaced by “or”

(7) Figure 2: The axes are not well visible, and the explanation for the positive control in panel B is missing in the legend.

Axes from figures 2, 3 and 4 were enlarged. For Figure 2B the following information was added in the legend: As a positive control, 0.05 U glucose dehydrogenase from Pseudomonas sp. was added to Δzwf cultures and to purified putative GDH1 (Sll1709). Axes from figures 2, 3 and 4 were enlarged.

(8) The investigation on the general absence/presence of the GDH/GK bypass in cyanobacteria may not be necessary for this study.

We included this data in this manuscript as the mistaken assumption that the GDH/GK bypass exist in Synechocystis lead among other observations to the misinterpretation of an existing ED pathway in Synechocystis. We would therefore prefer to keep these data in the manuscript.

(9) Line 316: delete "on".

“on” was deleted.

(10) Line 352: delete "or".

“or” was deleted.

(11) Lines 352/353: ZWF expression level appears to be reduced accordingly. This should be stated.

We agree that Zwf expression might be lower, however, we are not entirely sure if this is truly valid and would rather test this assumption further as described in the following lines.

(12) Figure 4: The axes are not well visible.

Axes from figures 2, 3 and 4 were enlarged.

(13) Line 363: values are not only normalized to protein content, but also to activity found for the WT.

We now added: The values are normalized to Zwf enzyme activity found in the WT based on protein content.

(14) Line 408: no separate subsection required.

The title for a new subsection was deleted.

(15) Lines 437-439: These are results descriptions, which should not be placed in the legend, but in the main text, as is partially done.

We deleted these result descriptions in the legend.

(16) Lines 441/442: formatting: one or no bracket pair.

The brackets were corrected.

(17) Lines 443/444: refer to Table 2 instead of giving the values in the legend to avoid duplication.

We deleted the values and now refer to Table 2.

(18) Line 503: delete "identified".

We deleted the second “identified” in the sentence and changed the wording to: Apart from four identified cyanobacteria that possess potential EDDs. In addition, we also added the names of the four cyanobacteria that were found including the sequence IDs of the putative EDDs.

(19) Lines 529/530: revise sentence and format.

We added one sentence and revised the following sentence: In contrast to Synechocystis EDA, EDA from Synechococcus prefers OAA over KDPG. The catalytic efficiency of Synechococcus EDA on oxaloacetate is rather low (OAA 0.437 s^-1 mM^-1), however, its activity can be enhanced by NADP⁺(13).

(20) Line 542: revise sentence.

We revised the sentence to: It remains to be investigated whether this reaction could play a role in vivo, with KDPG potentially acting as a regulatory metabolite at low concentrations.

(21) Line 577: The glass tubes used for cultivation should be specified.

We now added the following information: Custom-made glass tubes were placed in a photobioreactor (manufactured by Willi Hilke, Uslar, Germany).

(22) Lines 584-585: unclear, was the resistance cassette not placed in the gene to be deleted?

Yes, the resistance cassette was placed in the gene to be deleted and was fused for homologous recombination to two DNA fragments approximately 200 bp directly upstream and downstream of the gene. This information was now added.

(23) Line 607: cultivation equipment to be specified.

The following information was now added: For the purification of GDH1 from Synechocystis, a 6 L photoautotrophic culture of the P3-His-GDH1 overexpression strain was grown in a 10 L glass flask at 28°C, illuminated with constant light (50 µmol m^-2 s^-1) and gassed with filter sterilized ambient air to an OD₇₅₀ of about 1.

(24) Line 670: GTS should be specified.

Thank you for this hint. This was a typo. GST was meant not GTS. This was now corrected and GST was specified as Glutathione S-Transferase.

(25) Line 679: delete "gluconate kinase (GK) and".

The second gluconate kinase (GK) was deleted and sentence was revised to:

For gluconate kinase (GK) activity measurements in Synechocystis crude cell extracts the GK reaction was enzymatically coupled to 6-phosphogluconate dehydrogenase (GND) reaction, the latter providing NADP⁺ reduction, which was monitored photometrically at 340 nm.

(26) Line 686: again GK activity? Difference unclear. Was the previously described procedure for GND activity determination?

GK activity measurements in Synechocystis crude cell extracts and GK activity measurements with recombinant enzyme that was expressed in E. coli were done in two different labs with different protocols. Therefore, the first description refers to measurements with Synechocystis while the second measurement refers to measurements with E.coli. This is now specified more clearly.

(27) Type/supplier of spectrophotometers and centrifuges used should be given.

Model and brand were now added for the double-beam spectrophotometer (Uvikon 810, Kontron, Augsburg, Germany). As this study was performed in two different labs over a period of 8 years including one lab moving to a new location, it is now impossible to specify all centrifuges that were utilized. Even though we agree that it would be good to provide this information, we now would have difficulties to be specific.

(28) Consider the referencing of published methods to streamline the materials and methods section.

We now streamlined the materials and methods section by deleting repetitions as outlined below. However, as protein expression, protein purification and enzymatic tests were performed in different labs, in some cases several protocols are given.

(29) Line 757: give specifics of anti-rabbit antibody and define PBS-T and PBS-T Cytiva.

Specifics were added to the text.

(30) Lines 761ff: It is not given for all genes used how they were derived. All synthesized?

We now added detailed information for all genes.

(31) Line 762: codon-optimized for? E. coli?

The information was added that genes that were expressed in E. coli were codon-optimized for E. coli.

(32) Lines 782-786: Rationals for experimental strategies do not belong to materials and methods sections, but to results sections.

The part was deleted in the materials and methods section and transferred to the results section.

(33) Lines 818/819: repetitive.

We removed the repetition and refer to the purification method as stated above in the materials and methods section.

(34) Lines 847-851: True for all EDA-type assays? Kinetic parameters are shown in Table 2 rather than Table 1.

Yes, true for all EDA-type assays. We changed the Table number to 2.

(35) Line 863: delete "in".

“in” was deleted

(36) Lines 888-893: sounds repetitive.

The lines were modified accordingly.

(37) Lines 908/909: repetitive.

The repetitive comment on the definition of 1U was deleted.

(38) Lines 928-938: repetition

The repetition was deleted.

References

(1) M. Theune et al., Easy-to-use whole-genome sequencing workflows and standardized practices to uncover hidden genetic variation in <em> Synechocystis </em> PCC 6803 wild-type and knock-out strains. bioRxiv 10.64898/2026.04.08.717167, 2026.2004.2008.717167 (2026).

(2) X. Chen et al., The Entner–Doudoroff pathway is an overlooked glycolytic route in cyanobacteria and plants. Proceedings of the National Academy of Sciences 113, 5441-5446 (2016).

(3) D. Schulze et al., GC/MS-based 13C metabolic flux analysis resolves the parallel and cyclic photomixotrophic metabolism of Synechocystis sp. PCC 6803 and selected deletion mutants including the Entner-Doudoroff and phosphoketolase pathways. Microbial Cell Factories 21, 69 (2022).

(4) A. Makowka et al., Glycolytic Shunts Replenish the Calvin–Benson–Bassham Cycle as Anaplerotic Reactions in Cyanobacteria. Molecular Plant 13, 471-482 (2020).

(5) V. Zaitsev et al., Insights into the Substrate Specificity of Archaeal Entner–Doudoroff Aldolases: The Structures of Picrophilus torridus 2-Keto-3-deoxygluconate Aldolase and Sulfolobus solfataricus 2-Keto-3-deoxy-6-phosphogluconate Aldolase in Complex with 2-Keto-3-deoxy-6-phosphogluconate. Biochemistry 57, 3797-3806 (2018).

(6) J. S. Griffiths et al., Cloning, isolation and characterization of the Thermotoga maritima KDPG aldolase. Bioorg Med Chem 10, 545-550 (2002).

(7) S. E. Evans et al., Plastid ancestors lacked a complete Entner-Doudoroff pathway, limiting plants to glycolysis and the pentose phosphate pathway. Nature Communications 15, 1102 (2024).

(8) J. Moreno-Cabezuelo, G. Gómez-Baena, J. Díez, J. M. García-Fernández, Integrated Proteomic and Metabolomic Analyses Show Differential Effects of Glucose Availability in Marine Synechococcus and Prochlorococcus. Microbiol Spectr 11, e0327522 (2023).

(9) M. Theune et al., Easy-to-use whole-genome sequencing workflows and standardized practices to uncover hidden genetic variation in Synechocystis sp. PCC 6803 wild-type and knock-out strains. bioRxiv 10.64898/2026.04.08.717167, 2026.2004.2008.717167 (2026).

(10) R. Y. Stanier, R. Kunisawa, M. Mandel, G. Cohen-Bazire, Purification and properties of unicellular blue-green algae (order Chroococcales). Bacteriol Rev 35, 171-205 (1971).

(11) D. G. Gibson et al., Enzymatic assembly of DNA molecules up to several hundred kilobases. Nature Methods 6, 343-345 (2009).

(12) M. Boehm et al., Comprehensive study on ferredoxin isoforms in the cyanobacterium Synechocystis sp. PCC 6803. bioRxiv 10.64898/2026.04.08.717189, 2026.2004.2008.717189 (2026).

(13) N. Xie, C. Sharma, K. Rusche, X. Wang, Phosphoketolase and KDPG aldolase metabolisms modulate photosynthetic carbon yield in cyanobacteria. The Plant cell 10.1093/plcell/koae291 (2024).