BGPT: Paper Review: Inactivation of SARS-CoV-2 at acidic pH is driven by partial unfolding of spike

Explore by Goal

Quick Answer Copied

Mechanism claim (paper):

Acidic pH (below ~3) rapidly reduces SARS‑CoV‑2 infectivity primarily by partial unfolding of spike (S), which prevents viral attachment to target cells—while gross virion envelope integrity remains largely intact.

Long Answer

Paper review: Inactivation of SARS‑CoV‑2 at acidic pH is driven by partial unfolding of spike

Communications Biology (2025); DOI: 10.1038/s42003-025-08514-w

What the paper sets out to determine

Kinetic threshold: what pH and timescale are required for fast SARS‑CoV‑2 inactivation in acidic conditions?
Mechanism: whether inactivation is driven by membrane disruption vs spike conformational damage, and whether the attachment step is the bottleneck.
Specificity across coronaviruses and variants: compare SARS‑CoV‑2 with HCoV‑229E and test two SARS‑CoV‑2 strains (BetaCoV/Germany/BavPat1/2020 and JN.1).

Core findings (mechanism-first)

1) Acidic threshold

SARS‑CoV‑2: stable at pH 7.1; substantial infectivity drop at pH 2.8 (≥10 min exposure) and rapid near-complete loss at pH 2.2 (≈10 s).

2) Virion envelope integrity mostly preserved

RNase protection: low pH did not strongly change genomic copy protection relative to infectious titer, arguing against envelope disruption as the main cause of loss of infectivity.
Cryo-ET: morphological readouts indicate largely intact membrane/surface spikes for conditions such as pH 2.8 (with limitations in quantification at low post-neutralization particle counts).

3) Attachment is abolished after acidic exposure

Super-resolution attachment assay: after neutralization, virions treated at pH 2.8 show a pronounced loss of cell-associated virions compared with pH 7.1 controls, while short exposures (when no inactivation is detected) show bound virions similar to untreated.

4) Spike undergoes irreversible partial unfolding

Limited proteolysis (LiP) on recombinant spike: LiP-LC-MS detects significant peptide abundance changes at pH 2.0 vs pH 7.0, interpreted as conformational changes increasing proteinase K accessibility or enabling trimer dissociation, albeit with limited S coverage (~35%).
Antibody conformation readout on intact virions: conformation-biased antibodies show reduced binding for antibodies recognizing native (neutralizing) RBD epitopes after pH 2.0 exposure and increased binding for antibodies recognizing denatured/linear epitopes (NTD/S2 preference), plus quantification by intensity ratios and Manders overlap.

5) Species specificity

HCoV‑229E: virions remain infectious at pH 2.0 for at least 1 hour, and spike conformation changes are weaker/absent relative to SARS‑CoV‑2 by their antibody-based readouts.

Mechanistic model (what is known vs inferred)

Observed (high confidence in this paper)

Acidic exposure below ~3 produces rapid/titrable loss of SARS‑CoV‑2 infectivity in seconds–minutes regimes, and HCoV‑229E is much more resistant at pH 2.0.
Attachment of virions to ACE2/TMPRSS2-expressing cells is strongly reduced after acidic treatment.

Inferred mechanism (with stated uncertainties)

The paper infers partial unfolding of spike as the primary driver because conformation-biased antibody binding patterns on virions after pH pulses are consistent with disruption of native RBD epitopes and exposure of linear/denatured epitopes; they also use LiP-MS and argue it is irreversible.
Alternative/remaining contributors: spike shedding and subtle structural damage are not fully excluded by available assays, and cryo‑ET limitations prevent definitive quantification of spike loss at some conditions.

Critical appraisal (skeptical, test-focused)

Strengths

Mechanism triangulation: infectivity assays + single-cycle infection + RNase protection + cryo‑ET + attachment microscopy + conformation-biased antibody readouts collectively converge on “attachment blocked by spike unfolding” rather than “envelope rupture.”
Variant + cross-species comparison: the inclusion of HCoV‑229E and two SARS‑CoV‑2 strains supports the idea of spike-protein-specific pH sensitivity rather than a generic “low pH denatures everything.”

Limitations & possible blind spots

Non-aerosol simplification: experiments occur in bulk solutions with controlled pH and neutralization steps, not full respiratory aerosol microenvironments (matrix, viscosity, salt crystallization/phase behavior, RH history). The paper explicitly frames this as a step to isolate acidic pH effects.
Spike unfolding readout depends on epitope accessibility: antibody-based inference assumes “binding loss/gain = native disruption/linear exposure,” which is plausible but can be confounded by epitope masking, quaternary rearrangements, or partial shedding. The authors discuss shedding as a possible additional mechanism.
Cryo‑ET quantification under power constraints: for some conditions (e.g., pH 2.8 after neutralization), low virion counts limit conclusions about spike number changes, which weakens any envelope/spike-count-based elimination of alternative mechanisms.
Recombinant LiP coverage gap: LiP-MS coverage (~35%) suggests some spike regions may remain undetected, making the antibody assay especially important for RBD/NTD claims; the paper notes this coverage limitation.
Neutralization step as a modeling choice: the setup neutralizes after pH exposure to approximate re-equilibration when inhaled again; however, this choice shapes what “irreversible” means operationally and could bias the apparent recovery/irreversibility.

What would most decisively falsify the paper’s main mechanism?

Direct structural evidence of unfolding without attachment loss (e.g., if spike conformation reports indicate unfolding but virion attachment remains normal at pH 2.8), which would weaken the mechanistic coupling between “unfolding” and “attachment abrogation.”
Evidence that infectivity loss is primarily due to envelope damage despite RNase protection results—i.e., if independent membrane-accessibility assays (orthogonal to RNase) show envelope permeabilization at the relevant pH/time while spike native epitope loss is minimal.
High-resolution mapping of S shedding vs trimer disruption in the same pH regime, where distinguishing shedding (S1 release) from unfolding (conformational collapse) would refine or overturn the attribution of “partial unfolding” as primary.

Data transparency & reproducibility (what is available)

LiP LC‑MS data are deposited to ProteomeXchange/PRIDE with dataset identifier PXD064612.
Additional data (e.g., source data for graphs and exact p-values) are provided in supplementary information files.

Author reviews (browse by author)

Click a button for a bespoke author-centered review on BGPT.

Feedback:

Updated: April 29, 2026