Review of Mahaffey et al., 2025 Zonal gradients in phosphorus and nitrogen acquisition and stress revealed by metaproteomes of Prochlorococcus and Synechococcus

This manuscript utilises a protein biomarker approach to assess nutrient stress in marine picocyanobacterial populations across a transect in the North Atlantic Ocean. I have long been a fan of utilising protein biomarkers as a means of telling us directly what cells are experiencing in their in situ environment but I also know that you need to interpret this data very carefully. Hence whilst I acknowledge the tremendous amount of work that has gone into this manuscript I also believe the authors need to be very clear with both the benefits and limitations of these studies – and with some mention of this in the abstract warranted at least (since the authors allude to differences in regulation of specific genes between Synechococcus and Prochlorococcus, but this can also be extended to differences in regulation within a genus (i.e. between Pro ecotypes or Syn clades) and potentially between strains (as well as importantly whether there is indeed differential presence/absence of genes in ecotypes/clades which can complicate interpretation of subsequent metaproteomics data). Please see my specific comments on this manuscript below:

Line 39: microbial metabolism

Line 53 please delete ‘and’ at the end of the line

Line 62: patterns

Line 100: Ostrowski et al., 2010 ISME J should also be included here

Line 110: The Lomas et al., 2012 reference is either missing or incorrect

Line 115: please delete for i.e. encoding a …

Line 117: please delete for

Line 117 and Line 119: a new high affinity phosphatase (psip) was recently described and should be cited in this section (i.e. line 119) as well as Ostrowski et al 2010 since this is regulated by PtrA

Line 128: I think the authors should be up-front here about the fact that PhoA for Prochlorococcus and Synechococcus is not only not biochemically characterised but also atypical compared to the E. coli PhoA. The implication that PhoA requires zinc like the E. coli enzyme might be wrong. Indeed, the atypical alkaline phosphatase present in freshwater Synechococcus (PCC7942), which has sequence similarities to SYNW2390 and SYNW2391 in Syn WH8102, is inhibited by zinc! (see Ray et al., J. Bacteriology 1991)

Line 133: what is this submitted paper?

Line 136-7: Again, this is a suggestive sentence since this implies PhoA from picocyanobacteria requires zinc (which we don’t know) – see comment above

Line 220: 25 mM

Line 253: I’m wondering if you have any evidence that the PhoX you are quantifying from Prochlorococcus actually is a bona fide alkaline phosphatase (it is quite a different sequence (25% identity to the Syn WH8102 PhoX which has been biochemically characterised).

Lines 195-246: There is a detailed proteomic protocol here but there is little or no information on how protein abundance is normalised. The only place I could find something was in the legend to Fig 4. With this in mind please can you write a more detailed normalisation protocol in the methods section (and explain in detail what you mean by nSC). You state this represents the spectral counts normalized to the maximum value of each protein across 6 stations, but it is unclear to me what this actually means. Clearly for the data presented in Table 2 and used throughout the manuscript, you want fold change in protein abundance not just to reflect cell number of Syns/Pros or total Syn/Pro protein abundance (otherwise changes in abundance of proteins merely reflect changes in the abundance of organisms rather than reflecting potential expression levels as a function of the environment). With this in mind for normalisation, please can you also inform readers how you account for changes in gene presence/absence in Pro and Syn strains (clades/ecotypes) as you move along the transect. Some alkaline phosphatases are sporadically present in these organisms and if there is no gene present clearly you won’t encode protein to detect. In other words does the taxonomic composition (clade/ecotype of Syn/Pro) across the transect significantly change? You could assess this from the metagenomes you have from the cruise. (I’m guessing that Po populations are all HLII and Syn mostly clade III but some evidence for this across the transect would certainly be useful, otherwise your proteomic data could be very hard to interpret.

Lines 257-276: Please indicate here the length of time the nutrient bioassays were carried out for. Also, please indicate when samples for proteomics were taken (I assume 48 hours).

Line 336 (Table 2): Please can you indicate a statistical significance to these fold changes e.g. is a 1.3 fold change statistically significant?

Line 340-1: should read between west and east (i.e. delete west after east).

Line 349-350: I don’t follow this statement. Fig S2 shows data as far as I can tell only for the Pro HLII ecotype and not any other clades. Also, what is the difference between the hatched and filled box for the HLII ecotype? Additionally, what is HIII? MIT9314 is a strain not a clade.

Line 351: Please clarify what you mean by zonal trend. (I agree generally the spectral counts follow the increase in Pro cell abundance but they don’t capture the oscillations in Pro cell abundance at the east end of the transect).

Lines 353-4: This statement about Pro-PstS and Pro-PhoA is also only true if all Pro cells across the transect possess the gene in question.

Lines 397_399: This explanation is probably fine but most Syns and some Pros also possess PtrA (as well as the PhoBR system) (Ostrowski et al., 2010). Whilst PtrA appears to be controlled by PhoBR it’s also possible PtrA could respond independently (& hence perhaps to DOP directly).

Line 400-1: It’s probably worth to more explicitly explain what you mean here - so is this a hint that you think Pro PhoX may require zinc (not Fe)?

Line 402: total Pro abundance increased across the transect from west to east (Fig 3b)..(it didn’t decrease).

Line 403: The Pro and Syn cell number data in Table S5 is a single value (we would need to see the data for cell abundance across the time course of the bioassay). Please can this info be added to the Table (i.e. flow cytometry data for the initial time point and the 48 hour time point - which is when I assume the proteomics samples are taken).

Line 406: you mention about targeting specific clades here – but HLII refers to an ecotype (are you saying you can target specific sub-clades within the HLII ecotype?

Line 422: please delete psip1 here – it is not a regulator but shown to be a high affinity alkaline phosphatase (Torcello-Requena et al., 2024)

Lines 423-426: It might be worth mentioning that Pro MED4 possesses both PhoBR and PtrA (though not sure if these genes are present in all HLI Pros).

Line 431: You mention that HLI abundance typically increases with depth but also HLI abundance also generally increases in slightly higher latitude cooler waters (see AMT transect data in Johnson et al., 2006; Zwirglmaier et al., 2007, 2008 Env Micro).

Lines 438-440: It is surprising that Syn-PstS was not detected in the metaproteomes since PstS is generally highly expressed in these organisms in P-deplete conditions. It’s thus also odd that in contrast PstS can be detected in the bioassays – especially since from Table S3 Syn PstS is only detecting PstS from clade X which is generally not an abundant lineage in this water type (especially compared to clade III). Any thoughts on this contradiction would be useful to add.

Line 451: Regarding the fold change in abundance of PstS in the bioassays do you detect the PstS peptide initially (i.e. before nutrient addition) or are you interpolating a number here to give these fold change values? Please add some info here to explain this.

Line 455: You are stating here that Syn PhoA possesses Zn or Co but we don’t know this – I would change considering to assuming.

Line 468-469: The Waterbury et al., 1986 ref doesn’t show any data for alkaline phosphatase so please change this citation to one that supports this statement.

Line 469: in the presence

Lines 501-506: please delete these lines from the text since the info here relates to a different submitted paper and no cobalt additions were made in the nutrient bioassays reported here. (As an aside nutrient additions only indirectly report metal requirements of enzymes. To prove this the protein(s) would need purifying and the precise metals required for activity determined).

Line 522: as well as the different depths sampled you are also not capturing the whole Syn population in the bioassay experiment (since the peptide targets clade X) and this needs mentioning here.

Line 534: I would be tempted to be slightly less robust that PhoA is regulated by zinc  - given that the Zur regulon in WH8102 (the same strain for the Cox and Saito manuscript) does not include phoA (Mikhaylina et al., Nature Chem Biology 2022) - so there appears not to be direct regulation by a zinc responsive sensor. That said, BmtA is regulated by Zur in WH8102 and it was proposed by Mikhaylina et al., that this may provide zinc for PhoA. The authors also need to cite Ostrowski et al., 2010 since PhoA (SYNW2390) is regulated by PtrA.

Lines 537-541: As far as I can see there is no cobalt nutrient addition data included in this manuscript but it seems the submitted manuscript contains this info. Can the latter manuscript alone not discuss the cobalt data otherwise it’s a bit confusing where to access this info?

Line 549: which transporter are you alluding to here? – Please can you give an example locus ID or perhaps better a cyanorak cluster number here.

Line 567: I am very curious why the authors speculate that PhoX may be regulated by organic complex forms of Co or Zn. I think this may be dangerous to say given that Kathuria and Martiny Env Micro (2011) showed neither Co or Zn stimulated PhoX activity from Synechococcus sp. WH8102 (though it is a shame that Fe was not assessed in this latter study).

Lines 569-570: Please can you explain a bit more what you mean when you say deconvoluting P strategies is complicated by other environmental and physiological factors that can influence their growth.

Line 593: regarding Syn abundance across the transect do you have any evidence that there might be increased grazing pressure (or viral infection/lysis) in the east?

Lines 602 and 649: picocyanobacteria

Line 689-691: PhoX in Syn WH8102 is SYNW1799 (see Kathuria and Martiny 2011) so SYNW0120 is not correct. Indeed, for SYNW1799 this is located next to potential iron transporters (which would be consistent with PhoX requiring iron as it does in other bacteria)! Please correct this.

Line 691: So in the lines above you mention single strains (MED4 for Prochlorococcus) or WH8102 for Syn. In this line you then state these genes are separated in the chromosome in the whole genus - did you check this is the case for all published Pro and Syn genomes? Otherwise you would need to qualify this statement.

Line 738: mechanistic

Mahaffey et al. conducted a proteomics analysis on the samples from the zonal transect of the North Atlantic subtropical gyre to investigate the distribution and nutrient acquisition strategies of Prochlorococcus and Synechococcus. The natural nutrient gradients across the zonal transect make this analysis interesting, as it may drive observable trends in the cell abundance of cyanobacteria and their nutrient-related protein expression (PstS, PhoA, PhoX, P-II, UrtA, and AmtB). The study tried to analyse all nutrient-related proteins, including phosphorus, nitrogen, and trace metals (e.g., Fe, Zn, Co), with a focus on phosphorus. The main finding may be that there is a transition from phosphorus stress in the west and nitrogen stress in the east of Prochlorococcus and Synechococcus. Overall, the study is interesting. However, there are several issues that need to be addressed prior to publication.

1. The manuscript is too lengthy to get the main idea. The author presented and discussed the protein changes associated with the nutrient strategies of Prochlorococcus and Synechococcus, including phosphorus, nitrogen, and trace metals (e.g., Fe, Zn, Co). Each section is discussed in detail, which makes the article appear unfocused. Also, it is vague what the focus of each section is. I strongly recommend restructuring the manuscript to highlight the most significant findings (specifically, the P stress and uptake strategies), and the other parts could be presented as auxiliary or as support for the main findings. In addition, the manuscript is so wordy that it is easy for readers to lose interest and patience. I suggest rewriting the Results and Discussion part to make it more concise.

2. The entire article is riddled with speculation, and the viewpoints are rarely supported by solid evidence. For example, starting from line 635, the authors discuss the impact of aerosol dust deposition on the distribution of Prochlorococcus and Synechococcus, building on many previous findings. However, the current study did not provide any evidence indicating the presence of aerosol dust during sampling. The discussion on copper toxicity is also suppositional. In fact, the whole discussion on the distribution of the cyanobacterial cell abundance is hardly convincing. The abundance of cyanobacterial cells is a result of various biological and ecological processes, including growth, competition, grazing, and mortality. Although nutrient uptake strategies and ambient nutrients significantly influence growth, the relationship between growth and cell abundance is not directly proportional. And the correlation results are not direct evidence. Therefore, I don’t like this part and suggest removing it from the discussion.

3. The conclusion is too lengthy. The first paragraph is entirely unnecessary as it is repeated in the introduction and discussion sections. The conclusion should highlight the most significant findings of the study and their potential implications. Please rewrite it.

4. The abstract is also wordy. I think it exceeds the required word count of most journals.

5. Many statements in the manuscript cannot stand up to scrutiny. For instance: in line 658, “the view that the growth of picocyanobacteria are insensitive to nutrient availability”. This statement is strange and misleading. The ability of cyanobacteria to thrive under ultralow nutrient conditions does not indicate nutrient insensitivity; rather, it highlights the challenge of quantifying nutrients at such trace concentrations. in line 603, “Prochlorococcus are major players in the microbial loop and thus the availability of recycled nutrients such as ammonium and urea, driven by basin-scale processes, may influence their biogeography.” The causality of this sentence is strange. The availability of recycled nutrients impacts the biogeography of Prochlorococcusshould be because they are nutrient sources for Prochlorococcus?

6. It is better to reorder the figures. Fig. 1g and 1h should be put to Fig. 3 as it is biology-related. And Fig. 1h was mentioned after the Fig. 2 in the main text.