*Note: To structure these comments we included the text of the reviewer’s comments in plain text and then inserted our responses in bold beneath each comment or section of the reviewer’s comments.

In “All about Nitrite: Exploring Nitrite Sources and Sinks in the Eastern Tropical North Pacific Oxygen Minimum Zone,” Tracey et al extensively examine N cycling rates in the ETNP. They performed rates for denitrification, anammox, nitrite oxidation and ammonia oxidation. They also examined nitrite oxidation rates with a constantly purging system that regulated oxygen and used actual sensors to detect the oxygen. To me this addition is a big deal because one of the oxygen treatments was 1 nM. Thus, the authors showed significant nitrite oxidation rates at 1 nM oxygen. This work definitely adds to the literature, but I don’t think that they do a good job of selling themself.

I should start this review by admitting that I have not been a believer in anaerobic nitrite oxidation, so this review is written through that lens. However, I think this paper has the first convincing evidence that nitrite oxidation is occurring without oxygen. I think the authors need to shift the focus of their abstract and introduction to highlight this.

Right now, in the introduction, the paper acts like anaerobic nitrite oxidation rates in Babbin et al 2020 are sufficient. If so, why go through so much effort to measure rates a much more complicated manner? The previous rates were not sufficient to convince me because there still could be significant oxygen contamination despite the He bubbling in those types of rates. If I were the authors, I would lay it out in the introduction something like this:

Oxygen contamination is nearly impossible to avoid on deck because 1) Niskin bottles contaminate water at 1-2 uM oxygen before they are ever opened (Garcia-Robledo et al. 2016, 2021). Low oxygen adapted Nitrite oxidizers have a Km for oxygen at something like 4 nM O2 (Bristow et al. 2016). This contrasts to oxygen levels that inhibit denitrification, which are in the hundreds of nanomolar (Dalsgaard et al. 2014).

[Side Note: if the Km for Low oxygen adapted Nitrite oxidizers is 3-4 nM, then 3 nM is not functionally anoxic for nitrite oxidizers. ]

Originally, nitrite oxidation rates were obtained without purging contaminated oxygen (e.g Peng et al 2015, 2016). These rates did match metagenomic abundances of Nitrospina quite well (see Fuchsman et al 2017 Figure 9), so they weren’t crazy, but they certainly couldn’t be considered anaerobic rates.

In Sun et al 2017, they used a pump system which likely has lower oxygen contamination than Niskin bottles.

Then, rates were obtained after purging with helium in Babbin et al 2020. This was a big step forward. However, oxygen concentrations were not measured in these incubations. Just because oxygen has been reduced to levels where denitrification is not inhibited, does not mean that small amounts of oxygen do not exist. (see differences in important oxygen levels above)

Now here in the current paper, you 1) used a sensor to check oxygen concentrations! And 2) had a crazy set up to purge continuously. This is cool. It needs to be highlighted both in the introduction and the abstract. You have made changes so that more people will believe your results. But they will only read the paper if they know about these changes right away. When I look at the abstract, I see that you are trying to do this. But what you say there could also have been said in Babbin et al 2020—be more specific to make this paper stand out.

We are very grateful for these suggestions on how to more clearly articulate the significance of our oxygen manipulation experiments. We will edit the relevant lines of the introduction (93 – 114)  along the lines of your suggested logic to more clearly articulate how our experiments address the issues you raised with previous nitrite oxidation measurements’ oxygen contamination.

Second issue: The ETNP ODZ (at least offshore) generally has <10 nM ammonium (Widner et al. 2018). You are adding 2 uM ammonium. That is ok, you have to do that to see a signal in the mass spec. But it is more ammonium than these bugs have ever seen before. Additionally, at the top of the ODZ, nitrite is at low concentrations. You are adding 2 uM nitrite. Can you please look at your data and say whether adding ammonium and/or nitrite is stimulating anammox rates? To do this properly, you need to separate rates taken at depth with large in situ nitrite concentration from rates taken at tiny in situ nitrite concentrations and look at the separately.

I will admit that I did this with the rates in Babbin’s thesis (Sorry!) and I thought anammox rates were stimulated at the top of the ODZ by nitrite and maybe everywhere by ammonium. If you take this stimulation into account, the anammox rates get significantly smaller. Please make clear the enrichment factor for nitrite and ammonium in the experiments (like ammonium was 200x ambient ammonium). Figure 5 clearly shows that anammox is stimulated by nitrite.

This concern is a good point. We will perform this analysis and include it in the next version of the manuscript at a relevant point (likely section 4.5.4).  We will also add a table, text, or a figure that explicitly describes the enrichment factors for nitrite and ammonium.

Third issue: It is known that denitrifiers are mostly on particles (Ganesh et al. 2014, 2015; Fuchsman et al. 2017). The way we take samples for rates (small volumes from Niskin bottles) is biased against particles because 1) of the hydrodynamics, CTD rosettes under-sample big particles and 2) the nipple of a Niskin bottle is above the bottom, so big particles sink below the nipple and are not sampled (see (Suter et al. 2017)). Thus, your rates are, for the most part, water column rates. It is a mistake to ignore this.  

But the different detection limits of anammox and denitrification is a very good point as well. The 30N2 background in the mass spec can be huge. I think this is the single biggest problem with measuring denitrification. Were different mass specs used between the 2012/2013 and 2018 datasets? Some mass specs are better at detecting 30N2 than others due to the background issue.

 Yes, this concern about particle based denitrification is also a good point. We will add this caveat to section 4.5.4 on caveats about our anammox and denitrification measurements and their interpretation. We used the same mass spectrometer for the 2012/13 and 2018 measurements.

Fourth Issue: the abstract. Your potential readers will chose to read or not read this paper based on the abstract, so you want it be compelling. Specifically talk about the upgrades to the nitrite oxidation rate measurements. Lines 43-45 are not very compelling—better to make the section on dismutation short and sweet. Lines 48-51: rewrite in a straight forward way.

We are grateful for these suggestions on how to make the abstract more compelling.  We will reframe the text to emphasize our nitrite oxidation oxygen manipulations, will shorten the section regarding dismutation, and will rewrite the last sentence.

Detailed comments

Line 72: The most recent/accurate citation for this would be 30-50% of N2 production (DeVries et al. 2013)

We will adjust this citation accordingly.

Line 105: If you look at (Fuchsman et al. 2019) Figure 2, the Nitrospina depth profile matches the Prochlorococcus depth profile perfectly, supporting the link between primary production and nitrite oxidation in the ODZ.

We will add this citation to the sentence presenting a cryptic oxygen cycle as an explanation for nitrite oxidation.  

Line 106-108: I completely disagree. The methods in Babbin et al 2020 etc, while endeavoring to remove contaminating oxygen, and doing a good job at it, cannot guarantee that oxygen concentrations are below those needed by Nitrite oxidizers. Nitrite oxidizers can use 3 nM O2. That is an order of magnitude lower than levels needed for denitrification to occur. Not faulting the authors. But this technique doesn’t guarantee zero oxygen. Best to make clear.

We will address this concern by rewriting this section of the introduction according to the suggested logic above.

Line 108: There are maybe three Sun et al 2021 papers. I am not sure which one you are referring to here. I think the one not in the references.

At this line we are referencing Figure 3 from the following paper:

https://www.nature.com/articles/s41396-020-00852-3

 which is the Sun, 2021 ISME Journal paper titled “Microbial niche differentiation explains nitrite oxidation in marine oxygen minimum zones.” This study is contained in the references but in reviewing the manuscript again we noticed that it is not clear when the 2021 ISME Journal study vs. the Sun, 2021 ISME Communications study is referenced. We will edit the citations to 2021a and 2021b to clarify this.

Line 110: Please make clear which ODZ you are referencing. Sun et 2019 includes data from the ETSP (~350 m thick ODZ), where the ODZ is hundreds of meters thinner than the ETNP. In the ETNP, where the ODZ was 720 m thick, Nitrospina did go away at depth (Fuchsman et al. 2019). I think it is a good idea to keep clear in your head which studies have data from which ODZ.

Yes, this is a good point. We were trying to argue that in the Sun, 2019 paper the authors observe that a Nitrospina-like MAG is abundant through the ETSP ODZ to 300m (Figure 4).  We will acknowledge that this pattern is not seen in other studies or other OMZs by citing the Fuchsman, 2019 paper and other relevant studies such as Stewart et al., 2012 and Glass et al., 2015.

Line 111: Oxygen doesn’t inhibit nitrite oxidation. HIGH concentrations of oxygen inhibit nitrite oxidation. Very big difference. Microaerophiles are a known phenomenon.

 To address this comment, we will add the quantitative results from the Sun, 2021 (ISME Journal) study.

Line 106-114: I think you need to address oxygen intrusions here. It is kind of dismissed at line 105, but there is real data to support oxygen intrusions into the ODZs. For the ETNP see (Margolskee et al. 2019; maybe Monreal et al. 2022).

We will acknowledge these papers’ arguments for oxygen intrusions into the ODZ in this section.

Line 111-114: This sentence is very confusing. You are talking about your experiments, but you cite Berg et al 2022, which is a review paper.

 We cited Berg et al., 2022 because that is the source of our threshold of what constitutes functional anoxia. However, you are correct that we should cite the primary literature that establishes this. We will also rewrite the sentence to split apart into two sentences how our experiments build on previous results and our sources of what constitutes functionally anoxic oxygen concentrations.

Line 124: awkward sentence—uncertainty….there is uncertainty??

 We will rewrite this sentence along the following lines:

“and the enzyme hypothesis’ inability to account for structural and phylogenetic differences in the NXRs of the four NOB genera”

Lines 146-166: Almost all denitrifiers are found on particles (Ganesh et al. 2014, 2015; Fuchsman et al. 2017). The way we take samples for rates (small volumes from Niskin bottles) is biased against particles because 1) of the hydrodynamics, CTD rossettes undersample big particles and 2) the nipple of a Niskin bottle is above the bottom, so big particles sink below the nipple and are not sampled. Thus, your rates are, for the most part, water column rates. It is a mistake to ignore this.

See comment above for how we will address this by discussing it in section 4.5.4.  

Lines 146-166: I felt like this section could be reduced.

I think it is particularly important for the authors to understand that the earlier literature on anammox and denitrification rates were all contaminated by oxygen in the stoppers. That is why we now have to keep our stoppers/septa under anaerobic conditions. This was not discovered until Dalsgaard et al 2012—well it was presented at meetings before then. But the 2005-2009 exetainer literature is all contaminated by oxygen. Oxygen inhibits denitrification at lower levels than it inhibits anammox. I think it is best to step carefully here.

 Either here or in section 4.5 we will state that anammox and denitrification literature from 2005 – 2009 was influenced by oxygen contamination.

Line 167 and 170: the (1) and (2) in different sentences is awkward.

 We will rephrase these sentences to remove the (1) and (2).

Line 183-187: my favorite current paper related to this is (Ito et al. 2017) which shows that the oxygen content of the Pacific has been decreasing since the 1980’s. Also Horak et al is about current changes that actually have happened rather than expected changes. I think that citing current changes (Ito, Horak) is a powerful motivator. Clarify.

 We will rephrase this section to emphasize that changes have been observed (Horak and Ito) and that additional future deoxygenation is predicted by models.

Line 205 etc: Where the measurement performed on the ship or frozen and brought to the lab? Where the samples pre-filtered (if frozen)? This is normal for nitrate concentrations.

Ammonium and nitrite measurements were performed onboard the ship. Nitrate samples were frozen onboard immediately after collection. Upon return to the Ward laboratory, nitrate samples were thawed and then immediately measured.

Line 222: What psi of He?

 We will add this information in the next draft of the manuscript.

Line 252: Don’t write out the 2 in 2 uM.

 We will make this change.

Annoying comment: Was there N2 gas left in your control samples? This technique assumes that all N2 is purged from the sample before the incubation occurs. Sometimes that is untrue. This is something that you have to data to check and tell the reader.

In our data processing we accounted for the possibility that a few exetainers contained residual N2 gas by performing an outlier check for each time-course. Outliers above or below three times the mean of the ²⁸N₂ calculated for each time-course were not used in our calculation of anammox and denitrification rates. The details of this procedure are described on page 1 of the supplemental materials. We will change the text describing the outlier calculation to make it clear that we used this procedure to remove points with some initial N₂ from our calculation.

Line 375-376: Please use the word qPCR when talking about gene abundance measurements to inform the reader about the technique used.

 We will make this change.

Line 385: Calling samples ODZ core that are in the deep oxycline is not acceptable. If they are in the deep oxycline, then call them deep oxycline.

It would make more sense to label depths based on the measured oxygen at those depths.

 The naming scheme for shallow boundary waters and ODZ core waters came from Babbin, 2020 and is based on a global average of the relationship between potential density and oxygen. In line 385 we were trying to state that because the potential density threshold is based on a global average a few depths that in the SR1805 oxygen depth profiles are clearly deep oxycline are labelled as “ODZ core.” We will edit the sentence to make this clearer. However, we do think that keeping this naming scheme is helpful since it allows comparisons to the Babbin, 2020 paper.

Line 429: Please tell us what the oxygen concentrations was at 850 m. usually it is still very low at that depth.

The oxygen concentration was 1.5 μM at 850m. We will include a table of the oxygen concentrations in the supplemental materials for the final manuscript.

Line 430: I am confused by this statement. In Figure S1, station 2 and station 9 (7/6) do not have rates reaching the bottom of the ODZ in the figure.

We were trying to argue that N loss rates also peaked near the oxycline in these three Falkor stations not that the Falkor stations also had a deep anammox peak. We will edit the wording of this sentence to make it clearer.

All Figures. In all your figures, including supplemental, can you add in minor tick marks or more depth numbers. It is very difficult to determine the depth of a rate from the graph. We should be able to estimates values from a graph.

Figure 3: Please indicate which rate goes with which cruise by shape of the symbol.

 We will add minor tick marks to all figures and will modify the shapes used in figure 3 to show which rate goes with which cruise.

Figure 4 or line 448: Can you clarify the ambient oxygen concentrations of the water before it was used for these oxygen varying experiments? Maybe in the Figure 4 caption?

 Yes, we will add this information to either line 448, Figure 4’s caption, or the discussion of these results (section 4.3).

Line 489: I think it should be noted that ambient nitrate concentrations are often 20 uM in the ETNP ODZ. So when you added nitrate, did you double the nitrate concentrations? It seems to me that the ambient nitrate concentrations are critical for thinking about these experiments. Maybe adding nitrate didn’t matter because there was already so much nitrate to begin with. Please tell use the nitrate and oxygen concentrations in the original water for each experiment.

 We will add this information either here or the discussion section relevant to these results (section 4.4).

Line 541: (Fuchsman et al. 2019) have productivity measurements in the ETNP offshore, showing it is oligotrophic.

 We will add this citation to this section.

Line 552: If you wanted to, you could mention migrating zooplankton (Bianchi et al. 2014). the biological N2 maximum is largest at similar depths—coincident with the vertical migration depth (Fuchsman et al. 2018). Bianchi is fixated on anammox, but zooplankton also provide organic matter. See (Escribano et al. 2009; Cram et al. 2022). I would guess that this explains your results.

 We will add a discussion of migrating zooplankton to this section.

Line 913: (Fuchsman et al. 2017) found very few reads for genes for DNRA in ETNP metagenomes

 We will this citation to this line.

Line 924: best citation is DeVries et al 2013

 We will add this citation here.

*Note: To structure these comments I included the text of the reviewer’s comments in plain text and then inserted our responses in bold beneath each comment or section of the reviewer’s comments. 

The manuscript by Tracey and colleagues presents a detailed study of nitrogen transformations right above and within oxygen deficient waters of the eastern tropical North Pacific. Their work is centered around the assessment of sources and sinks of nitrite as a key component of the nitrogen cycle. Considering pivotal role of nitrogen in the biogeochemistry of low oxygen regions, the topic is timely since the better we understand the spatial-temporal variability of its transformations, the better we can constrain the rather large uncertainties in the global marine nitrogen (and carbon) budgets.

I consider the study novel as it provides provide evidence for nitrite oxidation under anoxic conditions, which is an exciting finding that would explain previous observations within OMZ waters. Likewise, these results set a benchmark upon which future work can build up to refine our understanding of nitrogen cycling under ongoing deoxygenation. It is clear that the experimental approach has limitations in its applicability to in-situ conditions (as acknowledged by the authors in section 4.5.4.), but showing the importance of a nitrate-nitrite recycling loop is a signifficant finding that challenges our current view on the most important N-cycling processes in marine systems. 

The paper is very well written, its structure is clear and the connection between objectives, experimental methods and conclusions is logical and easy to follow. I would only have to minor recommendations with respect to form and content. With regards to the former, I noticed unconsistencies in the spelling of gene names, which by convention are written in italics. With regards to the latter, I did notice that the significance of the study for a wide biogeochemistry community (including colleagues working in observations and modeling) does not come across as clear as the authors might think. I know they mention one or two sentences abuout the general relevance of their findings for the understanding of the nitrogen cycle, but I missed a clear statement on how this exactly can / should be done. In my opinion this comprehensive work should have such a statement to reinforce its added value to the literature.

Thank you for your comments and suggestions, we are glad to hear that you think the paper is well-written, clear, and significant for the field. We will fix all gene names to the correct convention and add in a section that notes the significance of the study for the broader biogeochemistry community (modelers and observational scientists).