Overall summary:

This was a lovely paper to review chock-full of interesting fungal ecology. The authors explore the establishment of fungal communities into mesh-bags with sand installed at a Norwegian boreal forest from control and nitrogen addition plots. They explored changes in hydrophobicity of in-growth bags and how this was linked to the ectomycorrhizal fungal community over three years of sampling. They provide evidence that the proliferation of particular ectomycorrhizal fungi, like Piloderma olivaceum, could be what causes hydrophobicity to increase over-time in the mesh bags. Because it is not only a long-distance explorer but also possesses interesting mycelial chemistry, such as mycelium coated with calcium oxalate, there is good theoretical support for this idea. To my knowledge, this is quite a nice and novel finding, with important implications for soil C storage. They also show that N additions prevented increases in hydrophobicity, shifted fungal communities at later stages, and prevented P. olivaceum from establishing. It’s always nice to review a paper clearly written by people who are knowledgeable and care about fungal ecology and who had a clear question to ask, which makes it quite unique from the lion’s share of ITS-based DNA-metabarcoding studies.

I have no high-level concerns for the paper but a number of important points regarding specific sections below in the line-by-line comments. I hope these improve the paper even further.

Sincerely,

Mark Anthony

Line-by-line comments:

Line 56: change ‘to’ to ‘for’

Line 57: I have been curious of this framing of these results because ericoid mycorrhizal fungi include many very strong decomposers (e.g. Burke and Cairney 2002,  Mycorrhiza; Kohler et al. 2015, Nature Genetics)

Line 61: Though a great study, I would not say that Lindahl et al. (2021) could conclude causality in their work, and thus I would not say that C. acutes ‘resulted in...’.  Rather, I would say ‘was linked to’.

Line 65-68: These results by Lilleskov et al. (2011) are very important, but some of the summaries at the genus level need to be reconsidered. For example, increasing evidence suggests that members of the Russula and Lactarius genera include species that respond both negatively and positively to N additions (e.g. Morrison et al. 2016, Fungal Ecology; Van der Linde et al. 2018, Nature; Moore et al. 2021, GCB; Anthony et al. 2021, ELEMENTA).

Line 78-83: Very interesting and nice expectations for the results plus clearly stated. Nice!

Line 94-95: Each fertilisation treatments includes a 50 kg N ha-1 yr-1 range, why is this? Maybe adding one additional clause to clarify. Because you ultimately consider fertilisation a single treatment, it is easily defendable but it would be good to stand lone in the paper versus needing to read Bergh et al. (2008) first. Can you also describe how long the N addition treatment was fertilized prior to installing the mesh bags?

Line 95-96: Can you provide more information on the fertilisation of other macro- and micronutrients? Because micronutrient loss is also hypothesised to influence how fungi respond to N additions (e.g. Whalen et al. 2018, GCB).

Line 116: It would be interesting to know why in November versus a time when tree growth and belowground C allocation is presumably higher.

Line 154: Please add one sentence about how sequences were denoised, given this is 454 data it is especially important bioinformatic detail.

Line 169-172: Was this part of the work done manually?

Lien 174: I realise the bioinformatics was done many years ago, but because the submission is current, I encourage updating language around the Zygomycota to be consistent with current taxonomic consensus (see Spatafora et al. 2016, Mycologia).

Line 175: Does ‘unknown ectomycorrhizal status’ also refer to non-ectomycorrhizal taxa or just taxa thought to be ecto but not confirmed?

Line 180-182: Maybe just say ‘relative abundance’?

Line 184-198: I am not well-qualified to evaluate this method but it seems rather intuitive and appropriate.

Line 215: Can you also provide technical details on the C/N measurements and the ergosterol analysis?

Line 221: Was this a PCA of all these variables together or was some type of vector fitting used to fit the non-fungal values (e.g. envfit function in vegan)? I am also a bit concerned of using PCA on relative abundance data given how many zeros are probably in the data and co-linearity among some of the taxa. Is there a reason why PCoA was not used with Bray-Curtis distance or a distance-based redundancy analysis also using Bray-Curtis dissimilarity? Both of these would be more suitable non-parametric alternatives. I am also guessing from Figure 4 that this is a PCA of both fungal relative abundances and organic matter properties together. Thus, was some type of transformation used to put everything on the same scale? Additionally, can you define what was the criteria for being the most abundant fungal OTU. Was there a cutoff based on sequence proportion and/or occurrence across the sampling units? I am also guessing this is at the OTU level?

Line 257: I think the alphabetic labels are missing on the figure (e.g., a, b, & c).

Line 268: I would also be interesting to have the Pearson correlation coefficient here.

Line 271: I think Figure 2 could be cleaned up a bit so the legend and axis labels look nicer and do not contain underscores.

Line 285: Why is it total fungal EMF and saprotrophic fungal communities? Is this because all other trophic groups and non-assigned to trophic group fungi were removed? Did you also look at EMF alone and saprotrophs alone?

Line 301-304: These details seem like they should be the first part of section 3.4, where the molecular results are first introduced.

Line 305-310: Already stated verbatim in section 3.3?

Line 311-316: Already stated at lines 290-294.

Line 321. Missing period after ‘(Fig 3b)’

Line 331-333: What linear model was used? Please add these details into the methods section.

Line 333-334: I do not believe you mean to say the ‘proportion of EMF to total Basidiomycota’, as it sounds like you calculated a ratio of the two, but rather, I think you mean to to say, ‘the proportion of EMF and the proportion of EMF increased over-time…’. I would also write Basidiomycota and Ascomycota versus ‘mycetes’

Section 3.5: You only need to site Fig. 4 one time.

Line 358: The idea of fungal succession is important, but it need not be independent from changes in environmental conditions across time. I personally do not feel this qualification is necessary here. It derails the momentum of your story so early on! If you feel it is essential to already provide a caveat in this first paragraph of the discussion, I encourage you make one that does not derail the traction of the main conclusion of this work. You could say something more powerful like: ‘Whether shifts in EMF were due to selection of later succession fungal taxa as the forest aged versus variation in climatic conditions remains unclear, but is ultimately not particularly important in terms of understanding how shifts in EMF relate to soil organic matter cycling’.

Line 359-363: Very nice and strong result! I find it very interesting.

Line 394: I believe this result is robust around T. asterphora responding positively to N additions, but I am willing to hedge it is geographically unique. Van der Linde et al. (2018. Nature) suggest an N depo optimum around 9 kg N ha yr-1 for this taxon, which is quite low for many parts of Central Europe where the work was conducted. Since your work was more northern and in a boreal forest, I bet the results are quite different from more southern, temperate forests. This is just a comment for consideration. I do not think there is need to comment on this in the text.

Line 401: Could you explain why the N tolerance component of this sentence helps to explain this result in little more detail?

Line 404: You found the hydrophobicity did not change until the third year in the control mesh bags, which is also when proportions of EMF dramatically increased to make this group dominant. This supports the idea that the increase in hydrophobicity was due to EMF accumulation. It could be worth stating this also in this paragraph.

Line 421-441: Such an interesting paragraph! The explanation around yeasts and their ergosterol content is particularly compelling.

Line 442-473: Another very interesting paragraph! Good ideas to explain the role of Piloderma species in soil C cycling.

Figure 1. Is this the average contact angle at 1 and 5 s combined?

Figure 2. It could help to provide a sentence describing what CA at 1 s and 5 s means; otherwise, the figure is a little tricky to interpret as a stand-alone component.

Figure 3b: It is hard to differentiate between Tomentella stuposa and Tylospora asterophora and then between Piloderma olivaceum and Amphinema byssoides and then between Suillus lutes and Amphinema sp 5. Could you select colors with more contrast?

Figure 4: Are the vectors the coefficients of the linear combos of the initial variables used to make the PCA? I would also write OTU instead of species.

In this study, Almeida et al. used mesh ingrowth bags filled with maize amended (13C enriched) sand to understand how EMF biomass/necromass contributes to the hydrophobicity of SOM under N fertilization.  They found that N fertilization caused significant changes in the fungal communities colonizing the mesh ingrowth bags from so called ‘nitrophobic’ taxa that respond negatively to mineral N deposition to more ‘nitrophilic’ taxa. Nitrophobic taxa generally produce mycelia that composed of cords/rhizomorphs that are hydrophobic (Ho) while nitrophilic taxa generally have hydrophilic (Hi) mycelia. These changes in community composition corresponded to changes in hydrophobicity of soil organic matter (SOM) in the bags (measured by contact angle) suggesting that hydrophobic fungal necromass deposition (particularly from Piloderma) may be important to the formation of hydrophobic SOM, which is thought to be more resistant to decay and thus increase C storage. 

I found the manuscript to be interesting, novel, and improves our understanding of EMF and their influence on SOM hydrophobicity. I also found the study to be methodologically sound and have little in terms of major critiques. The one major comment I did have was on the interpretation of the results and mechanisms driving increased hydrophobicity of the bag SOM substrate: How do you tease apart the effects of fungal in-growth from hydrophobic compounds entering the bags from the surrounding litter/SOM (e.g. ingress of hydrophobic compounds via transport in water)? While I think the evidence presented in the manuscript strongly suggests  that the former is probably the major driver, I don’t think one can completely rule out the potential effects that the surrounding litter and SOM (and associated changes with the N treatment) may have on the substrate properties in the in-growth bags. I would suggest adding a few sentences in the discussion about plausible alternative mechanisms.

Line comments:

L3 For correct grammar change “fungi” to “fungal” OR  just omit “symbiosis”  

L37 revise to say “...this mycelium turns into necromass…”

L44 the authors might want to add hypothesized mechanisms behind differences in decomposition rates among hi and ho SOM here

L53 change “saprophyte” to “saprotroph” for consistency (and a more widely accepted term)

L60-63 I would suggest explicitly stating that this particular species of Cortinarius has retained the enzymatic capacity to breakdown complex SOM in order to access nutrients. 

L64-68 species are mentioned but genera are given as the examples 

L64-68 I would add a sentence stating that for Russula and Lactarius there is quite a bit of variability in response to N fertilization at the species level 

L66 missing a “,” in front of Suillus 

L92-96 Please provide what form the N fertilizer was

L174 The Zygomycota is no longer a recognized Phylum (now split into the Mucoromycota & Zoopagomycota; Spatafora et al. 2016)

Figure 1. This is a matter of style but I feel the data could be presented in a different way that is more intuitive and impactful (e.g. boxplots?). The information is there, for me it just was not conveyed immediately.  

Figure 2 I would be curious to see the N treatment treated as a covariate in an ANCOVA. Is the control slope steeper compared to the fertilized? Just looking at the plots it would appear so and may bolster the support for arguments made in the discussion about Ho and not Hi biomass that is contributing to SOM hydrophobicity. 

Figure 3b. Maybe in the key you could add the hydrophobicity of each of the taxa based on genus level classifications in Agerer 2001 and Lilleskov et al. 2011 for those that are not immediately familiar? Additionally, a third panel with the relative abundance of the two hydrophobicity groupings could be added. 

Figure 4. “saprotrobes” should be “saprotrophs”

Ectomycorrhizal fungi (EMF) have varied potential effects on SOM properties and decomposition. Some EMF produce large quantities of hydrophobic mycelia, which may increase the recalcitrance of SOM, enhancing soil C stocks. Not all EMF produce hydrophobic mycelia, however. N fertilization, for instance, has been shown to select against hydrophobic EMF, potentially reducing the effects of EMF fungal ingrowth on SOM hydrophobicity and recalcitrance. This study performs a number of analyses on sand-filled meshbags amended with maize incubated for 2 years and 8 months over a full fertilization experiment (~50-75 kg N/ha/yr + other macro-/micronutrients) to test the effects of the EMF community on the hydrophobicity of the meshbag contents. While I found the aims of this study compelling, several aspects give me pause. I am particularly interested in how this study parses between the effects of the full fungal community vs EMF on substrate hydrophobicity.

I think more explicit treatment of EMF community composition would strengthen the claim that the hydrophobicity unique to some EMF taxa is responsible for increased substrate hydrophobicity in the control plots. For instance, Almeida et al. could code for mycelial hydrophobicity and compare the relative sequence abundance of hydrophobic EMF taxa over time/across treatments. The effect of fertilization is particularly important given that the design of the study hinges on the idea that EMF with hydrophobic mycelia decline with fertilization. I would also like to see more information about the variability within the EMF community. For instance, instead of a stacked bar chart, Figure 3b could be broken out by treatment and each taxon could have an error bar. Further, much of the discussion centers on the physiology and ecology of the most abundant species of EMF in the control plots (Piloderma oliviceum), but little information is available regarding how consistently this taxon shows up in the meshbags. If P. oliviceum is indeed abundant across most samples, this would strengthen the discussion of its role in potentially enhancing SOM hydrophobicity. Also, if possible, regressing the relative sequence abundance of hydrophobic EMF against the averaged contact angle of the substrate in an ANCOVA would strengthen the claim that EcMF mycelial hydrophobicity is imparting increased hydrophobicity to meshbag contents. If this emerges across N fertilization treatments, this would be particularly impactful.

45-47: The hydrophobicity of living EMF mycelia is framed as a possible driver of SOM hydrophobicity here, but the subsequent analyses do not address how hydrophobic vs. hydrophilic EMF differ in abundance over the fertilization treatments. Is there a way to either change this framing, or address it with further analyses?

48-50: I’m unclear about how by removing N and P from SOM, EMF activity may reduce soil C stocks. By inhibiting saprotrophic activity by outcompeting them for nutrients, wouldn’t this enhance soil C stocks by reducing respiration by saprotrophic fungi (Gadgil effect)? Do you mean that EMF themselves are mineralizing C from SOM, decreasing soil C stocks?

81-83: Here, you write that you would expect higher hydrophobicity in the control vs. fertilized plots due to the higher proportion of hydrophobic species – where is this tested? Genus-level assignments on mycelial hydrophobicity are available in the literature, as well as exploration type assignments that could offer further resolution on the effect of EMF mycelial traits on substrate hydrophobicity.

235: Table 1 would benefit from indicators of significance, either between treatments or over time. I found myself asking questions like “is the decline in the contact angle in the fertilized plots over time significant?” and struggling to locate the relevant information in the text. Alternatively, it could be visualized, which would also make it easier to interpret.

258-260: Looks like you’re missing figure labels (a, b, c).

272-273: Figure 2, would it be possible to make this a little cleaner (axis titles, the legend title)? Also, when the proportion of EMF reads is so low during the first two harvests, how can you attribute new C (C3 ingrowth into C4 substrate) to EMF alone? Table 1 indicates that roughly half of the “new” carbon enters the substrate by the end of the first incubation, although the proportion of EMF reads is only ~ 10%.

296: Figure 3, could you break panel b out to provide more information about how variable the EMF community is? Also, could you provide visual information about how the traits of the EMF community (hydrophobicity and/or exploration type) may be shifting according to sequencing results?

300: How do different kinds of EMF respond to the fertilization treatment over time? Hydrophobic vs. hydrophilic genera?

359-363: Here you argue that overall EMF abundance is linked with higher hydrophobicity. This is different from your original framing, where you implicate EMF producing hydrophobic mycelia.

372-374: What is the mechanism of partner selection implied here? Reduced total C allocation to EMF fungi? How do you rule out environmental filtering and/or changes in EMF C sink strength?  The Defrenne study cited here is correlative – how does it support your causal claim?

375-376: This is where I think more information about the homogeneity vs. patchiness of the EMF community would be helpful – are Piloderma spp. abundant across all samples?

415-419: Refer to comment from line 235. How can you attribute new C to EMF when their relative abundance after the first incubation was so low? Also, the synthesis offered here (EMF necromass and biomass may contribute to SOM hydrophobicity) deviates from your original hypothesis, that EMF with hydrophobic mycelia in particular are contributing to SOM hydrophobicity.

437-439: This argument suggests that filamentous fungi contribute more than yeasts to SOM hydrophobicity – that is a much larger group than EMF fungi. How can you parse between the effects of filamentous fungi at large and EMF?

445-449: This discussion would be strengthened by a more robust quantitative analysis of the relative abundance of hydrophobic/hydrophilic EMF across treatments. Where is your final hypothesis addressed?