The manuscript by le Paige et al. is an insightful view on solid-state NMR studies of the nucleosome and how to profitably analyze its interactions with partner proteins. 

This is a very interesting article that examines the properties of sedimented samples of nucleosome particles, obtained by ultracentrifugation into ssNMR rotors. The properties of the sample are evaluated by ssNMR and SAXS, and in addition to nucleosome samples, also a complex with a weakly binding protein (PHD2) is studied.

While many groups have been using sedimentation, often without reporting the details of the sample, this study reports their properties, and this will certainly be useful for others that use similar procedures.

The study is rigorously done and I recommend publishing it, after considering the following points.

Page 5, line 138: “weighting factor of the 15N chemical shift differences (in ppm) of 6.51 “. Actually, the weighting factor is 1/6.51=0.154. This is in the range of commonly used value, reviewed here: Williamson, M.P. (2013). Using chemical shift perturbation to characterise ligand binding. Prog. Nucl. Magn. Reson. Spectrosc. 73, 1–16.

The methods indicate that the rotors were closed, but no rubber spacer was used, nor was the cap glued into our rotor. In our hands, the samples can lose water, often excessively, without rubber spacer or gluing. This observation has also been reported in one case: Asami, S., Szekely, K., Schanda, P., Meier, B.H., and Reif, B. (2012). Optimal degree of protonation for 1H detection of aliphatic sites in randomly deuterated proteins as a function of the MAS frequency. J. Biomol. NMR 54, 155–168; see Figure 11.

The spectra of the 1H water signal change in shape (Fig 2a); there are clearly several components in the first measurement, which disappear after some time. This could be related to loss of bulk water, as described here: Böckmann, A., Gardiennet, C., Verel, R., Hunkeler, A., Loquet, A., Pintacuda, G., Emsley, L., Meier, B.H., and Lesage, A. (2009). Characterization of different water pools in solid-state NMR protein samples. J. Biomol. NMR 45, 319–327.

My interpretation is that:

• There is bulk water, probably in the center of the rotor, in the beginning – this actually means that the calculated packing density is underestimated, see below.
• The bulk water is lost
• The hydration water is largely (but not totally) retained.

Can the authors comment on this? Have the caps been glued? Probably not, which may explain the losses. For other samples, such water loss can have dramatic consequences on the sample, as we reported for the case of SH3 crystals (reference above, Asami et al).

I am not convinced that the following statement (page 9) is totally true: “Taken together these data demonstrate the nucleosomes in the sediment remain well-folded and hydrated through the measurements without evidence for direct nucleosome-nucleosome contacts” The spectra of day 1 and day 34 are clearly different, and a change in hydration is likely.

On page 7, the authors state that the packing density (61%) is lower than in crystals. They state that in crystals the “packing coefficients of ~67% and a solvent content of ~54% “. I did not find out what is meant by “packing coefficient”. I assumed that it is the volume occupied by the protein relative to the total volume. But if it was so, then the solvent content should by the complement to the packing coefficient, i.e. 33% in this case. Could the authors clarify this?

As eluded to above, it is likely that the packing density in the rotor during MAS is higher than after ultracentrifugation into the rotor. In typical samples, at least crystalline ones, we always find that bulk water accumulates in the center of the rotor. The water 1H peak in the beginning of the measurement shows an additional component (on the right), which most likely is bulk water, located in the center of the rotor. Thus, I believe that the actual packing density of the species that is detected by NMR is higher than the authors’ estimate.

Figure 4e: it is surprising that some peaks (res. 20-25) have much higher intensity in the complex (INEPT based). This suggests that these residues become more mobile. Can the authors comment on this?

Review of the manuscript “Characterization of nucleosome sediments for protein interaction studies by solid-state NMR spectroscopy” (mr-2021-21) by Ulric B. le Paige et al., by Lars Nordenskiöld and Xiangyan Shi.

le Paige and co-authors presented a thorough discussion about packing nucleosome samples by sedimentation for ssNMR studies. It revealed that the nucleosomes in the ssNMR rotor are highly compacted, well hydrated and stable, meanwhile have sufficient disorder of nucleosome arrangement to allow proteins interacting with nucleosomes. This method allows studying the interactions between nucleosomes and binding factors, as demonstrated by the ssNMR study of a co-sedimented PHD2-nucleosome complex. The contribution is very helpful to practitioners interested in studying nucleosome interactions with NMR.

There are several minor comments.

1. The authors added 2 mM Mg2+ in the nucleosome solutions prior to sedimentation. What is the reason to choose 2 mM Mg2+? Does it affect the sedimentation efficiency, final concentration or ssNMR data quality if adding less or adding more Mg2+? See also below on the effect of Mg2+ concentration on order as observed in SAXS X-ray diffraction spectra.
2. When assessing the final concentration of nucleosomes packed in the ssNMR rotor, have the authors also tried quantifying this by other methods, such as quantifying the 13C or 1H signals using an external standard? It’s not uncommon that considerable sample loss can happen during packing, clearing top for caps, closing caps etc, which would lead to overestimating the material quantity in the rotor by checking the supernatant.
3. In the sedimented PHD2-nucleosome complex sample, 20:1 ratio was used. Does the free PHD2 remain in the supernatant or the sediment? If it is mostly sedimented, is it possible to prevent this by optimizing sedimentation time, or buffer or other conditions? It will be useful to include such information in the manuscript.
4. In the introduction, line 60 "mainly via interactions mediated by the histone tails" is simplified, they interact by stacking in columns at various degree of order and the stacking for a condensed system is governed by not only close packing considerations, but the contacts (stacking) between the histone octamer core surfaces are mediated in addition to tails, by charge-charge interactions by charged aa on the two surfaces as well as presence of divalent or higher charge salt counterions that screen electrostatic repulsion.
5. In many places, the space between a value and the unit is missing, for example, Line106 - 10 mM; Line156, 10 bp; Line 148, 15 s; Line 117, 298 K; and many others.
6. Line 76, “contacts” -> “contact”
7. Line 120, “monitored” -> “were monitored”
8. How were the error bars in Figure 4 derived? 

9. Regarding SAXS data. "The sediments are devoid of pronounced long-range ordering of nucleosomes". Here they use 2 mM Mg2+. As the authors note later on, the long range order depends highly on the Mg2+ concentration used when precipitating the NCPs. As shown by Berezhnoy et al, the most pronounced order was achieved at 20 mM Mg2+, while there is no long range order at 2 mM.
10. Line 265: “the sediments seem to primarily consist of heterogeneously packed nucleosomes with mean inter-particle distance of 78 nm.” Should not this read “7-8 nm”? Line 259 reads: “pronounced peak at q* ~0.08, corresponding to a characteristic distance of ~7-8 nm”.
11. Line 348: “we estimate that the length scale of the regular structure in the sediment is ~1520 nm, corresponding to stacks of two to three nucleosomes, without a significant preference in relative orientation between the stacks”. How did the authors arrive at this conclusion? Is it based on FWH in SAXS spectra, if so details should be given? The peaks seem too broad and ill-defined for such an analysis.
12. It would be good if the authors try to adhere to the recent guidelines for the description of SAXS results described in Acta Crystallogr. D Struct. Biol., 73(Pt 9), 710 (2017).
13. Figure 3c (167 bp nucleosome SAXS solution data) lacks indication of the nucleosome concentration and details of calculation of the form factor profile (nucleosome structure used for calculation, program applied).
14. For the condensed (precipitated) samples of the 167bp nucleosome, the authors obtained SAXS profiles that are different from data reported for condensed NCPs (nucleosome without linker DNA) and also from the recently published SAXS results for the 177bp nucleosome ( Soft Matter, 14, 9096 (2018); Sci. Rep., 11, 380 (2021)). It would be good to make some discussion and comparison with the indicated data. Is the difference due to the low Mg2+ concentration in the samples?