Dear Editor,

This MS presents an assessment of aerosol oxidative potential in a number of environments in central Europe. The MS is well written and follows a straightforward structure. It presents a comprehensive analysis covering from the quantification of major and trace aerosol components to source contribution with PMF and finally apportioning the relationship between sources and OP results. The OP results are interpreted in terms of seasonality and links with emission sources in the study area. This is an interesting study which in principle merits publication. The main issue in my view is the methodology applied (acellular assays, AA and DTT), their lack of comparability and different responses obtained across studies. While this is simply the current state of the art, the authors should expand on these limitations, what the advantages with regard to other current methods are and how these limitations impact their results.

Some specific comments:

- lines 40-52, please add references and/or criteria for the selection of these 8 factors

- lines 71 and below: the authors discuss the increasing interest in methods to determine OP using  in vivo, in vitro cellular and in vitro acellular assays, but then they move on to only discuss the acellular assays (lines 75 and below) by stating that “The OP is frequently measured by acellular assays for exogenous ROS”. At least one paragraph should be dedicated to discussing the limitations/advantages of the acellular vs cellular (in vitro or in vivo) methods, and their implications regarding the results presented in this work. Some (non exhaustive) examples of authors working on cellular methods are 

Janssen, N. A. H., Yang, A., Strak, M., Steenhof, M., Hellack, B., Gerlofs-Nijland, M. E., Kuhlbusch, T., Kelly, F., Harrison, R., Brunekreef, B., Hoek, G., & Cassee, F. (2014). Oxidative potential of particulate matter collected at sites with different source characteristics. Science of the Total Environment, 472, 572–581. https://doi.org/10.1016/j.scitotenv.2013.11.099

Gerlofs-Nijland, M. E., Bokkers, B. G. H., Sachse, H., Reijnders, J. J. E., Gustafsson, M., Boere, A. J. F., Fokkens, P. F. H., Leseman, D. L. A. C., Augsburg, K., & Cassee, F. R. (2019). Inhalation toxicity profiles of particulate matter: a comparison between brake wear  with other sources of emission. Inhalation Toxicology, 31(3), 89–98. https://doi.org/10.1080/08958378.2019.1606365

Bessa, M. J., Brandão, F., Fokkens, P., Cassee, F. R., Salmatonidis, A., Viana, M., Vulpoi, A., Simon, S., Monfort, E., Teixeira, J. P., & Fraga, S. (2020). Toxicity assessment of industrial engineered and airborne process-generated nanoparticles in a 3D human airway epithelial in vitro model. Nanotoxicology, 15(4):542-.

Stone, V., Miller, M. R., Clift, M. J. D., Elder, A., Mills, N. L., Møller, P., Schins, R. P. F., Vogel, U., Kreyling, W. G., Jensen, K. A., Kuhlbusch, T. A. J., Schwarze, P. E., Hoet, P., Pietroiusti, A., Vizcaya-Ruiz, A. de, Baeza-Squiban, A., Teixeira, J. P., Tran, C. L., & Cassee, F. R. (2017). Nanomaterials Versus Ambient Ultrafine Particles: An Opportunity to Exchange Toxicology Knowledge. Environmental Health Perspectives, 125(10).

- line 98: “It is important to extend the studies on this emerging health-related metric”, wouldn’t it be necessary first to agree on a comparable method? The authors have just discussed the lack of comparability between the AA and DTT methods, and concluded that “they exhibit different responses to various groups of ROS-generating compounds and their bioavailability”. What can be concluded from this? What is the point of extending results to other regions if they cannot be compared? Please discuss how the authors plan to approach this.

-  line 121, how many samples were collected per location and per season? The numbers described here are relatively low to apply PMF. If all the samples pooled together for the PMF analysis please discuss the limitations, e.g., different emission profiles in rural vs city sites. Line 186 confirms that a multisite approach was applied. Please discuss. Section 3.6 discusses these limitations; it might be more useful for the reader if the discussions on limitations are distributed and found in the respective sections they refer to. For example, lines 543-551 could be moved to the paragraph containing line 121. 

- line 148, please reduce the number of self-citations; e.g., surely an earlier reference can be provided for the EC tracer method

- line 158, the OP of the extracts was measured without filtration, how can the authors be sure their results did not suffer from interference from quartz fibres extracted unintentionally from the filters by vortex agitation? This is a known artefact linked to the toxicity of quartz materials, especially fibrous materials.

- line 227, “steadily increasing towards the city centre”, as in the case of PM2.5? Then the OPDDT and PM2.5 results  would be aligned, but this doesn’t seem evident from Figure 1. Please clarify. 

- line 282, the authors state “Their (AA and DTT results) comparison to our OP data is hindered by important experimental details such as the extracted amount of PM from filters”. Aren’t the OP results normalised by PM mass? An in depth assessment of the uncertainties of the OP results should be presented, taking into account that the amount of PM extracted is another source of uncertainty. Line 284, “It can be roughly identified that our median OP values are somewhat larger”, please remove the subjective terms (roughly, somewhat larger…”. It is unclear whether the comparisons reported are reliable, based on the authors’ previous statements regarding comparability.

- line 341, another example of subjective terminology “definitely underline”, please remove. Line 342, “wholistic” should be “holistic”

- lines 345-348, please elaborate on the oil source, what is the authors’ interpretation, precisely? This source accounts for almost 70% of PM mass in summer and spring in the rural area, which is somewhat surprising. Unless the site is located in close proximity to an industrial activity (in which case the site should be renamed), it is likely that the source refers to long-range transported secondary aerosols, as opposed to direct oil combustion.  This would also be consistent with the decreasing relevance of this source towards the city centre, where contributions from primary sources (e.g., traffic) are higher in relative terms. Please review.

- lines 464 and below: the authors state that “shares from vehicles (i.e. joint sources of road traffic and vehicle metal wear) in the non-heating period to OP prevailed”, and that “points to the remarkable role of primary traffic emissions in causing oxidative stress in spring and summer”, why only in the non-heating season? If OP is driven by particle chemical composition (I.e., sources) then the effect of traffic aerosols should also be present during the winter months, even if the impact of BB aerosols on OP is even larger during the winter months. Is this the case?  Can the impact of BB and traffic on OP be effectively disaggregated, or is there a compound effect?

- line 481, please define “proximity metric”

The OP of three regions in and around Budapest; urban, suburban and rural, over all four seasons is investigated in this paper.  The DTT and AA assays are used to measure OP, and source apportionment and linear regression employed to determine the sources affecting these assays at the different locations and for different seasons.  The method has been widely applied in other European locations by some of the co-authors, so although the methods are not new, the results are since the location is novel. Overall, the results are interesting, if OP can be/is linked to health endpoints, since they increase the knowledge of factors affecting PM2.5 OP, which are consistently showing the importance of biomass burning and non-tailpipe vehicle emissions.  The analysis is very detailed and for the most part the paper is well organized and clearly written. This paper covers an important topic and is suitable for publication in ACP with the following revisions to consider.

Last line of Abstract, does this imply use of OP or just sources in developing regulations?

Line 40, how are ambient particles biologically complex?  Is this referring to point 8) in the following lines?  Please clarify.

Lines 66 to 68, this is an incomplete list of current studies linking OP to health effects. Eg, the Weichenthal group has published a number of papers on this topic in the last few years. There may be more, but that could be a place to start an expanded literature search. A better literature review is needed here since this is a critical argument on why these assays are useful, ie, that there is a link to health effects.

Lines 128 to 130, how was water on the filters controlled or not controlled as part of the mass measurement (ie, equilibrium reached as some low RH)?

Line 139, if metal ions, such as Cu(II), Fe(II), Mn(II) play a role in OP (see line 92), why were total metals analyzed. How does this influence the interpretation of the data, ie, how can the metals data be linked to OP?  One might consider the solubility of the metals.  Low soluble metals, such as iron may have little relation to the total iron concentration, whereas for copper with higher solubility, the issue may be different.

Line 158, are insoluble species included in the analysis? Specifically, what fraction of insoluble species are expected to be included in the measured OP.  Does soot/EC, which may have surface active species, contribute?  More quantitative details are needed here.

Line 331, how are the intercepts on the DTT vs AA in both plots of Fig 2 explained?

Lines 335 to 343, the paragraph is not very clear. Why are the assays coherent based on both the mass and volume normalized data?  Also, I assume the statement that AA is more sensitive than DTT is because the slope of DTT vs AA is less than 1 (might state this explicitly). But what if DTT is just more sensitive to a more chemical species? Reference to sources here (SOC, BB) seems out of place, it has not been discussed yet.

Source apportionment does not show any secondary biogenic SOA?  Also, can PMF be performed here for a given location and season given the limited number of samples?

Line 424 and throughout.  Is intersect the proper term for the regression intercept?  Maybe either is acceptable?

Discussion of slopes of DTT or AA vs PM2.5 mass concentration.  Isn’t the slope in the plots of Figure 4 equal to DTTm and AAm? This could be explicitly discussed and aid in the interpretation. Also, contrast this DTTm and AAm to the data in Table 1 which give a different trend in AAm (I believe), likely due to the intercept not affecting AAm from the slope method.  This should be discussed.

Does it make sense that the “toxicity” based on AA (ie, the slope in Fig 4b) is the least for the city center and higher at the rural site?  The term toxicity is used loosely here since these acellular assays really do not provide insight on actual toxicity. One might consider this when using the term toxicity throughout the manuscript.  The AAm spatial trend is opposite other studies from Europe and most other regions; cities have higher mass normalized OP than rural areas. Are there studies that show the opposite, as found for AAm in this study?  Do the authors think this is actually true from a health perspective; ie, the toxicity of rural aerosol, which is largely due to the influence of long-range transported dust, is more toxic than urban aerosols?

On both scatter plots, why not make the symbols, regression line and the r2 colors match? This would make it easier to interpret.

Line 562, could one find a better word then pronouncedly?

Finally, two different acelluar assays were used. They respond differently to chemical species in the aerosol particles, and so respond differently to sources, which results in differing interpretations of the PM2.5 air quality in the various regions. It is concluded that more than one assay is needed to get a full picture of air quality. But more could be added.  Are these two assays optimal for that goal, or can the authors suggest other pairs based on published findings. If not, what test should be done to select optimal assays?  It seems like more interpretation could be gleaned from this study.