Pith. sign in

REVIEW 4 major objections 4 minor 4 references

Nanoscale Protein Diffusion in Supercooled Cryoprotectant Solutions

T0 review · 4 major / 4 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Proteins keep moving in supercooled cryoprotectant when macroscopic viscosity says they should be frozen.

desk verdict A credible new XPCS dataset showing ferritin out-running the Stokes–Einstein baseline below 230 K; the effect is likely real, but the baseline and the model claim need tightening. read the letter →

arxiv 2512.02742 v2 pith:UJ4PRZUG submitted 2025-12-02 cond-mat.soft

Maddalena Bin , Anita Girelli , Mariia Filianina , Mario Reiser , Sharon Berkowicz , Milla {AA}hlfeldt , Michelle Dargasz , Sonja Timmermann
show 10 more authors
Jaqueline Savelkouls Takeshi Kawasaki Shinji Saito Federico Zontone Yuriy Chushkin Fajun Zhang Frank Schreiber Michael Paulus Christian Gutt Fivos Perakis
This is my paper · ORCID
classification cond-mat.soft
keywords proteindiffusionStokes–Einsteindeviationsupercooledglycerol–waterX-rayphotoncorrelationspectroscopydynamicalheterogeneityferritinvitrificationfluctuatingfriction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper measures how ferritin proteins diffuse through a glycerol–water cryoprotectant as the solution is cooled from room temperature to 210 K. It finds that below about 230 K, the protein moves faster than the Stokes–Einstein relation predicts from the solvent's bulk viscosity, by up to a factor of 2.7 at 210 K. The authors argue this shows that molecular-scale protein mobility is controlled by local, spatially heterogeneous solvent friction rather than by the average macroscopic viscosity. If correct, this means that cryopreservation protocols based on bulk glass-transition temperatures may underestimate how mobile proteins remain during vitrification, which is directly relevant to designing safer freeze–thaw procedures for biological and pharmaceutical samples.

What carries the argument

The central object is the ratio D/D0 between the measured ferritin diffusion coefficient and a Stokes–Einstein reference built from larger silica nanoparticles in the same solvent. The identity D0 = (Rh,NP/Rh,P)·DNP converts nanoparticle diffusion into the SE prediction for ferritin. The mechanism carrying the argument is the fluctuating-friction model of Rozenfeld–Luczka–Talkner, which in the slow-fluctuation limit gives D/D0 = 1/(1−δ²), directly converting the measured enhancement into a relative amplitude of local friction fluctuations δ = Δγ/γ0. Together these allow the paper to translate a factor-of-2.7 mobility excess into a statement about ~80% local friction fluctuations at 210 K.

What would settle it

Measure the diffusion of the same ferritin in the same glycerol–water mixture using a probe that does not rely on the SE baseline—for example, fluorescence correlation spectroscopy or pulsed-field-gradient NMR—and check whether the diffusion coefficient at 210 K is still ~2.7 times higher than the value extrapolated from the macroscopic viscosity, or measure the diffusion of a series of silica nanoparticles of different sizes to see whether the nanoparticle reference itself obeys SE down to 210 K.

Watch

Extended reading notes

Core claim

The central claim is that ferritin diffusion in a 23 mol% glycerol–water mixture deviates from Stokes–Einstein behavior below T≈230 K, with the measured diffusion coefficient exceeding the SE prediction by up to ∼2.7 at T=210 K. This deviation is quantified by comparing ferritin (Rh=7.3 nm) to larger silica nanoparticles (Rh=50 nm) assumed to follow SE, giving a reference D0 = (Rh,NP/Rh,P)·DNP. The two datasets overlap down to 230 K and diverge below, and VFT fits yield an arrest temperature T0=85±11 K for ferritin versus T0=122±4 K for the nanoparticles. A minimal fluctuating-friction model, D/D0 = 1/(1−δ²), links the enhancement to local friction fluctuations δ = Δγ/γ0, which grow to ∼0.79

Load-bearing premise

The entire comparison rests on the assumption that the 50 nm silica nanoparticles obey the Stokes–Einstein relation across the whole temperature range, so that the rescaled nanoparticle diffusion gives the correct SE baseline for ferritin; if the nanoparticles themselves deviate from SE or sample a different local viscosity, the size of the reported protein enhancement changes.

Editorial extensions

If this is right

  • If proteins remain mobile below the solvent's apparent glass transition, cryopreservation protocols that rely solely on bulk Tg may be insufficient to arrest protein diffusion and aggregation during storage.
  • The measured T0 = 85 K for ferritin, far below the solvent Tg ≈ 165 K, implies that molecular-scale solutes can diffuse at temperatures where the macroscopic solvent is effectively arrested.
  • VFT analysis of the two tracer sizes gives a concrete size dependence of the apparent arrest temperature, offering a benchmark for theories of how dynamical heterogeneity couples to probe size in supercooled liquids.
  • The fluctuating-friction parameter δ, reaching ~0.8 at 210 K, provides a direct experimental estimate of the amplitude of dynamical heterogeneity in a cryoprotected solution, which can be compared with simulation and neutron-scattering studies.
  • XPCS is demonstrated as a viable probe of single-particle diffusion in deeply supercooled cryoprotectant solutions, opening a route to test other proteins and formulations under actual vitrification conditions.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension would be to vary protein size continuously to map how the SE deviation onset temperature and magnitude scale with tracer radius, testing whether the enhancement peaks when the probe radius matches the correlation length of slow solvent domains (~10 nm).
  • The fluctuating-friction model's prediction D/D0 = 1/(1−δ²) could be tested at even lower temperatures: if δ approaches 1, the model predicts a divergence in the enhancement, which is unphysical; the actual behavior would reveal a crossover to hopping or other transport mechanisms.
  • The paper's assumption that silica nanoparticles follow SE could be checked with a second independent reference tracer of a different chemistry (e.g., gold or polymer beads) to rule out probe-specific surface effects in the measured deviation.
  • Because the ferritin concentration is relatively high (volume fraction 0.047), an experimental check with a dilution series could confirm that the reported enhancement is not influenced by interparticle interactions or collective diffusion effects in the XPCS signal.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper reports XPCS measurements of ferritin diffusion in 23 mol% glycerol–water mixtures from ambient conditions down to 210 K, complemented by SAXS showing no structural changes. The measured diffusion coefficients are compared with those of 50 nm silica nanoparticles through a Stokes–Einstein (SE) baseline, D0 = (Rh,NP/Rh,P)·DNP, and the ratio D/D0 is reported to increase below ~230 K, reaching ~2.7 at 210 K. VFT fits yield T0 = 85±11 K for ferritin versus 122±4 K for the nanoparticles. A minimal fluctuating-friction model with Equation (6), D/D0 = 1/(1−δ²), is used to attribute the enhancement to local friction fluctuations, with δ ≈ 0.57 at 220 K and ≈0.79 at 210 K.

Significance. If the central claim holds, this is a valuable direct measurement of protein mobility in a cryoprotectant at deeply supercooled temperatures, with implications for vitrification storage. The experimental strengths include direct XPCS measurements with linear Γ(q) = Dq², SAXS control of structural stability, and explicit flux-dependent controls. The D/D0 trend is internally consistent between ferritin and nanoparticle data down to ~230 K. However, the quantitative magnitude (2.7×) and the model interpretation rest on two load-bearing assumptions: that the 50 nm silica tracers obey SE over the full temperature range, and that the model prediction is not circularly derived from the same data. These issues require attention before the claims can be considered established.

major comments (4)
  1. [Eq. (5) and Fig. 4] The central D/D0 ratio is built entirely on the assumption that the 50 nm silica nanoparticles follow SE over the full temperature range. No independent validation is provided (e.g., comparison of DNP with known glycerol-water viscosity or with a tracer independently verified to obey SE). The text asserts that any NP deviation 'would only reduce the apparent SE violation,' but this is a directional claim without support: positive SE decoupling of the NPs would indeed reduce the apparent enhancement, while negative deviation would inflate it. Please provide an independent baseline or quantify the plausible range of D0 from published SE-violation data for comparable tracers, and propagate this uncertainty to D/D0.
  2. [Eq. (6) and Fig. 4] The model 'prediction' is circular. The values δ=0.57 at 220 K and δ=0.79 at 210 K are obtained by inverting Eq. (6) from the measured D/D0 (δ = √(1 − D0/D)), so the solid line in Fig. 4 is drawn through the data by construction, not as an independent prediction. The model is currently a reparameterization of the observed enhancement. To claim that the model reproduces the enhancement, δ should be constrained by independent measurements or by a physical model of heterogeneity (e.g., domain-size distribution); otherwise, the text should explicitly label this as an illustrative mapping rather than a predictive test.
  3. [Fig. 2b and Methods] Ferritin XPCS was measured at q=0.1–0.43 nm⁻¹ and ϕ=0.047, where interparticle correlations and hydrodynamic interactions can produce a q-dependent collective diffusion coefficient D(q)=D_self/S(q). The paper states that the measured diffusion 'corresponds to the single-particle (self) motion' without presenting S(q) or demonstrating a low-q plateau. The linear Γ(q)=Dq² is consistent with both self-diffusion and collective diffusion with a constant effective D. Please provide evidence, such as measurements at a lower volume fraction or a demonstration that D is q-independent within error across the full q range, to support the self-diffusion interpretation.
  4. [Abstract, Fig. 3, and Conclusions] VFT parameters are quoted inconsistently: T0=85±11 K in the abstract/conclusions versus T0=85±12 K in the text and Fig. 3; the B values also differ between text and conclusion (B=26±8 vs. B=11±1). More importantly, the reported difference in T0 (85 K vs 122 K) inherits the SE assumption through Eq. (5). If D0 is recalibrated, both the T0 difference and the onset temperature may shift. Please report D/D0 with full error bars propagated from the individual fits, and state the sensitivity of T0 to the assumed SE behavior of the nanoparticles.
minor comments (4)
  1. [Front matter] The manuscript contains two different abstract texts: one at the very beginning and another after the author list. They should be merged or the duplication removed.
  2. [Methods] The text states 'the hydrodynamic ratio h=R_h/R_p = 0.075' with R_h=7.3 nm and R_p=6.25 nm; the ratio is 1.168, not 0.075. Please correct the definition or the value.
  3. [General] Typographical: 'an good fit' should be 'a good fit' (near Eq. 4). Also, the XPCS proposal number is given as SC-5375 in Methods but SC-5275 and SC-5359 in the Acknowledgements; please unify.
  4. [Fig. 4] The symbols in Fig. 4 show D/D0 without visible error bars, even though the individual D values have fitting uncertainties. Please add propagated error bars to the ratio.

Circularity Check

1 steps flagged · score 6.0 of 10

Fig. 4's 'model prediction' is the inverse of Eq. 6 with delta extracted from the measured D/D0; the model reproduces the enhancement by construction, although the underlying ferritin-vs-silica ratio is an independent measurement.

  1. fitted input called prediction [Figure 4 caption and the paragraph following Eq. 6]
    "the solid line represents the prediction of the fluctuating–friction model based on Eq. 6, where δ = ∆γ/γ0 corresponds to the relative amplitude of local friction fluctuations. ... At T=220 K, the extracted value is δ≈0.57, while at T=210 K it increases to δ≈0.79"

    Eq. 6 is a one-to-one map, D/D0 = 1/(1−δ²). The paper obtains δ by inverting the measured D/D0 values ('the extracted value is δ≈0.57...'), so the solid line in Fig. 4 is not an independent prediction. It is the same measured ratio re-expressed through Eq. 6, with no independent observable constraining δ (no direct measurement of friction heterogeneity). Thus the model's 'reproduction' of the enhancement is tautological; it adds interpretation but no predictive test.

full rationale

The experimental determination of D/D0 is not circular: ferritin and silica nanoparticle diffusion are independent XPCS/DLS datasets, combined through Eq. 5, and the reported enhancement is a measured external comparison. The D0 baseline rests on the assumption that the 50 nm silica particles obey Stokes–Einstein, supported in part by a self-citation; that is a validation/correctness concern rather than an internal circularity. The clear circular step is the fluctuating-friction model: Eq. 6 is inverted to define δ from the same D/D0 data that the 'prediction' line is supposed to reproduce. This makes the model curve fit-by-construction, though it does not invalidate the measured ferritin-vs-nanoparticle ratio. The central quantitative claim (ferritin D/D0 up to 2.7 at 210 K) remains an experimental result, so the score is 6 rather than higher.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical entities. All quantitative claims rest on the VFT fit parameters (fitted to the measured D(T)), the per-temperature δ extracted from Eq. 6, and four domain assumptions: SE validity for the NP reference, the single-particle interpretation of XPCS at ϕ=0.047, the applicability of the VFT form for extrapolation, and the correctness of the fluctuating-friction model in the slow-fluctuation limit.

free parameters (3)
  • VFT parameters for ferritin (A, B, T0) = A not reported; B=26±8; T0=85±11 K
    Fit to D(T) (Fig. 3); T0 and B are extrapolated and used to claim ferritin stays mobile below the solvent Tg.
  • VFT parameters for silica nanoparticles (A, B, T0) = A not reported; B=11±1; T0=122±4 K
    Fit to D(T) for the reference tracer; used via Eq. 5 to define the SE baseline D0.
  • Friction-fluctuation amplitude δ = δ≈0.57 at 220 K; δ≈0.79 at 210 K
    Obtained by inverting Eq. 6 using the measured D/D0; the Fig. 4 model curve is not an independent prediction.
assumptions (4)
  • domain assumption Silica nanoparticles (Rh=50 nm) obey the Stokes–Einstein relation over the full measured range, so D0 in Eq. 5 is a valid SE reference for ferritin.
    Stated before Eq. 5: 'We assume that the nanoparticles, given the larger size and low concentration, follow the SE relation to a good approximation.' This anchors the magnitude of the SE violation.
  • domain assumption XPCS at c≈100 mg/mL (ϕ=0.047) reports single-particle (self) diffusion of ferritin, with no significant collective or interaction correction.
    Stated in the introduction: 'At the low protein concentrations considered here, this diffusion corresponds to the single-particle (self) motion of ferritin.' No S(q) data or concentration series is shown.
  • domain assumption The VFT relation (Eq. 4) is an appropriate description over 210–293 K, and its fitted T0 can be extrapolated well below the data range.
    Used to extract T0=85 K vs 122 K; the claim that proteins 'remain mobile far below the apparent Tg' depends on this extrapolation.
  • domain assumption Equation 6 (Ref. 40) is the correct slow-fluctuation limit for local friction fluctuations in this protein–solvent system.
    The model is imported from Ref. 40; its validity here is assumed, and δ is inferred by inverting the equation rather than measured independently.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Nanoscale Protein Diffusion in Supercooled Cryoprotectant Solutions." pith.science (2026). https://pith.science/paper/UJ4PRZUG

@misc{pith2026251202742,
  author       = {Pith},
  title        = {Pith review of: Nanoscale Protein Diffusion in Supercooled Cryoprotectant Solutions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UJ4PRZUG}},
  note         = {Machine review of arXiv:2512.02742}
}
read the original abstract

Vitrification during cryopreservation requires a quantitative understanding of protein transport in deeply supercooled cryoprotectant solutions, yet direct measurements at molecular length scales remain scarce. Here, we combine X-ray Photon Correlation Spectroscopy (XPCS) and small-angle X-ray scattering (SAXS) to investigate ferritin diffusion in glycerol-water mixtures from ambient conditions down to 210 K. The measured diffusion coefficients reveal that ferritin retains a higher mobility upon cooling than expected from hydrodynamic scaling based on measurements of larger silica reference tracers, with the difference emerging below approximately 230 K. A minimal fluctuating-friction model reproduces the observed relative enhancement in diffusion, illustrating how local variations in the effective friction can give rise to such behavior. These measurements provide direct experimental benchmarks for future theoretical and simulation studies aimed at understanding molecular transport in deeply supercooled liquids approaching the glass transition.

Figures

Figures reproduced from arXiv: 2512.02742 by the authors.

Figure 1
Figure 1. Temperature-dependent SAXS intensity I(q) for ferritin solutions in glycerol–water mixtures at c ≈ 70 mg mL−1 (solid lines). The nearly invariant scattering profiles indicate structural stability and the absence of cold denaturation or nanocrystallization upon cooling. measures temporal fluctuations in the scattered X-ray intensity through the normalized intensity autocorrelation function, g2(q, t) = ⟨I(q, t0)I(q, t… view at source ↗
Figure 2
Figure 2. (a) Intensity autocorrelation functions g2(q, t) for ferritin solutions at q = 0.1 nm−1 and various temperatures. Solid lines represent exponential fits. (b) Relaxation rate Γ(q) as a function of q 2 for the same temperatures. Linear fits with Γ(q) = Dq2 yield the diffusion coefficients D(T). Error bars reflect fitting uncertainties propagated from the exponential model. for temperatures between T = 250 K and T = 21… view at source ↗
Figure 3
Figure 3. (a) Diffusion coefficient of ferritin solutions (red triangles) compared to nanopar [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Ratio of the measured ferritin diffusion coefficient [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

4 extracted references

  1. [1]

    A.; Gibson, M

    (1) Murray, K. A.; Gibson, M. I. Chemical approaches to cryopreservation.Nat. Rev. Chem. 2022,6, 579–593. (2) Caliskan, G.; Kisliuk, A.; Tsai, A. M.; Soles, C. L.; Sokolov, A. P. Protein dynamics in viscous solvents.J. Chem. Phys.2003,118, 4230–4236. (3) Dirama, T. E.; Carri, G. A.; Sokolov, A. P. Coupling between lysozyme and glycerol dynamics: Microscop...

    2022

  2. [162]

    X-ray scattering at beamline BL2 of DELTA: Studies of lysozyme-lysozyme interaction in heavy water and structure formation in 1-hexanol.J

    17 (42) Dargasz, M.; Bolle, J.; Faulstich, A.; Schneider, E.; Kowalski, M.; Sternemann, C.; Savelkouls, J.; Murphy, B.; Paulus, M. X-ray scattering at beamline BL2 of DELTA: Studies of lysozyme-lysozyme interaction in heavy water and structure formation in 1-hexanol.J. Phys. Conf. Ser.2022,2380, 012031. (43) Steinmann, R.; Chushkin, Y.; Caronna, C.; Chava...

    2022

  3. [1621]

    Ferritin-based nanomedicine for disease treatment.Med

    16 (31) Zhu, Y.; Zhu, Y.; Cao, T.; Liu, X.; Liu, X.; Yan, Y.; Shi, Y.; Wang, J.-C. Ferritin-based nanomedicine for disease treatment.Med. Rev.2021,3, 49–74. (32) Boutet, S.; Robinson, I. K. Precrystallization clusters of holoferritin and apoferritin at low temperature.Phys. Rev. E2007,75. (33) Chen, B.; Sigmund, E. E.; Halperin, W. P. Stokes-Einstein Rela...

    2021

  4. [5580]

    Dynamics of proteins in solution.Q

    (29) Grimaldo, M.; Roosen-Runge, F.; Zhang, F.; Schreiber, F.; Seydel, T. Dynamics of proteins in solution.Q. Rev. Biophys.2019,52. (30) Rodrigues, M. Q.; Alves, P. M.; Rold˜ ao, A. Functionalizing Ferritin Nanoparticles for Vaccine Development.Pharmaceutics2021,13,

    2019

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.