Decision letter

VERDICT: minor

Summary of Evaluation

This manuscript extends activity-based protein profiling (ABPP) to acutely dissociated adult mouse brain cells ("brainocytes"), reports stereoprobe ligandability maps for a set of tryptoline and azetidine acrylamides, and follows up four CNS-enriched hits in depth. Two of these follow-ups are substantial: the complexoform-restricted liganding of DPYSL2 at C504, which requires assembly with DPYSL5 and is validated by co-expression, cysteine mutagenesis, competition, and CRISPR disruption in two independent neuroblastoma backgrounds; and a covalent ligand series (WX-02-679) that engages a conserved cysteine in the cyclic nucleotide-binding domain of HCN1/2/4 and blocks cAMP-dependent activation shifts in heterologous cells, in inside-out patches, and — with the caveat discussed below — in native CA1 neurons. The panel was unanimous that the chemistry, the proteomic design (enantiomeric controls, competition arms, pre-specified thresholds, 4–6 biological replicates) and the electrophysiology are sound, and that no result is a single-assay artifact. Compliance and citation integrity are clean.

I have weighed the debate's one genuinely contested item carefully, because it is the item most likely to change how a reader uses this compound. The manuscript's Highlight and Summary state that the stereoprobes block cAMP modulation "while sparing basal activity of HCN channels." In HEK293T cells this is directly and cleanly supported: WX-02-679 alone gives HCN2 V₁/₂ = −95.0 ± 1.8 mV versus vehicle −92.1 ± 2.0 mV, while abolishing the cAMP-induced shift to −83.5 mV. But in hippocampal slices, with no exogenous cAMP, the same compound stereoselectively reduces sag ratio and rebound depolarization and shifts RMP, input resistance and firing rate — and the Discussion itself explains this as "loss of the tonic depolarizing Ih." The scientific_validity reviewer flagged this as an internal contradiction, and the advocate in the debate conceded the Highlight-level wording is broader than the slice data support. I agree with the reviewers and the advocate that this is a claim-scope problem rather than a data problem. The heterologous experiment is the correct controlled test of basal sparing, and it holds; the slice result is a real and important boundary condition in tissue where a fraction of channels is tonically cAMP-bound. Both readings are consistent with the same molecular mechanism. What is not acceptable is a headline that generalizes the heterologous result to native tissue while the paper's own native-tissue data and Discussion say otherwise. Resolving this requires rewording and an explicit reconciliation in the text — not new experiments — and so does not by itself lift the decision to major.

Two other items required judgement. First, the skeptic argued that possible pseudoreplication in the slice recordings (N = 8–10 cells, animal count not stated) could undermine the very experiment bearing on the basal-activity question, and that if the cells came from few animals new recordings might be needed. I do not think this rises to a major revision. The slice experiment is not the load-bearing support for the basal-sparing claim (it is the counterweight to it), the effects are stereoselective against a matched enantiomer within the same preparation, and the required action is to state the animal-level n and, if cells are clustered within animals, to report the analysis at the animal level or acknowledge clustering. If that re-analysis materially weakens the slice effects, the appropriate consequence is a weaker statement about native-tissue activity — which the revision will already be making conservative. Second, the concern that brainocyte-restricted liganding could be a dissociation or detection artifact was raised independently by four reviewers, but every one of them described the remedy as quantification of data already in Dataset S2 (how many of the 28 brainocyte-unique hits were tested in slices and how many replicated) plus explicit acknowledgement of the uptake confound in Figure S2A. That is a reporting fix, and it is required below.

The compliance audits surface three HARD gaps that must be closed: absent mutagenesis and cloning primer sequences (the Methods promise "the primers shown below" and no table appears), an absent flow-cytometry gating strategy, and an unresolved reference 133. None obstructs evaluation of the central claims, but all three block replication of specific constructs and QC steps and belong in the revision.

Weighed against In Silico's bar, this is a well-controlled, multi-modal study whose evidence supports its findings once two claims are scoped to the data that actually support them. I would recommend it to a colleague working on covalent chemical biology or HCN pharmacology. Every demand below is achievable in text, figures and files.

Required Revisions

1. Scope the "spares basal activity" claim to the dataset that supports it, and reconcile it with the slice data in the main text. Revise the Highlight, Summary, Significance section and the corresponding Results sentence so the claim is explicitly attributed to heterologous expression systems (HEK293T, three isoforms, two species). In Results, state plainly that in CA1 pyramidal neurons WX-02-679 reduces sag ratio, rebound depolarization and related passive properties in the absence of exogenous cAMP, and give the interpretation you already offer in the Discussion (a tonically cAMP-modulated fraction of native channels) at the point where the slice data are first presented, not only later. The current arrangement, in which the Highlight asserts sparing and the Discussion explains a reduction, is the specific inconsistency to remove.

2. Report animal-level replication for the hippocampal slice recordings and, if cells are clustered, account for it. For Figures 6J–K and S6E–G, state the number of animals and slices contributing to each condition alongside the cell counts. If multiple cells derive from the same animal, either report the primary comparisons at the animal level (or with a model that accounts for clustering) or state explicitly that cells were treated as independent units and that this may understate variance. Adjust the strength of the native-tissue statements to match whatever the animal-level analysis shows.

3. Quantify the brainocyte-versus-lysate and brainocyte-versus-slice comparisons rather than describing them qualitatively. Replace "generally recapitulated" (Figure S1D) with numbers: how many of the 28 brainocyte-unique enantioenriched proteins were quantified in the acute-slice experiment, how many reproduced the stereoselectivity, and how many did not. Where a protein was not quantified in slices, say so. All of this is derivable from Dataset S2.

4. Address the probe-uptake confound explicitly. Figure S2A shows lower overall stereoprobe reactivity in brainocytes than in HEK293T cells but does not quantify it. Provide a quantitative statement of the uptake difference (e.g., integrated lane signal normalized to protein) and state in Results that a detection-threshold shift is an alternative explanation for a portion of the brainocyte-versus-lysate differences, indicating what the acute-slice data do and do not exclude. Where available, report abundance (from the unenriched proteomics used for Figure 2E) for brainocyte-unique versus lysate-unique hits.

5. Justify or bound the relaxed liganding thresholds. The 33% (rather than 50%) parent-competition criterion is defended by lower uptake; state how many of the 114 liganded proteins, and how many of the 28 CNS-enriched hits in Table 1, would survive the stricter 50% criterion. A single supplementary column in Dataset S2 flagging threshold class is sufficient and would let readers apply their own stringency.

6. Support the "not seen in prior ABPP studies" claim with a cross-reference table. For each CNS-enriched liganded protein in Table 1, indicate whether it was (i) quantified and liganded, (ii) quantified but not liganded, or (iii) not quantified in the prior datasets you cite (refs 13, 31, 37, 38; PXD042541). Table 1 currently carries a binary Y/N column that conflates (ii) and (iii); separating them determines whether the brainocyte advantage is new ligandability or new coverage, and the claim in Results should be worded accordingly.

7. Report multiple-comparison handling for the IP-MS analyses. For Figure 3I (and Figure S4D), state how many proteins were tested, whether FDR correction was applied, and if so the method and threshold. If uncorrected p-values are shown, label them as such in the caption and note that the highlighted interactors (PRKAR1A, AKAP8, AKAP8L) are nominated hypotheses.

8. Provide the missing primer sequences. The Methods state that site-directed mutagenesis used "the primers shown below," but no primer table is present. Supply mutagenesis and cloning primer sequences for all constructs used (at minimum PLP1 C6A/C7A/C6A-C7A, PDE7B C3A/C136A/C321A/C325A/C328A, DPYSL2 C504A, HCN1 C542A, HCN2 C611A/C611S, HCN4 C662A/C662S), together with plasmid identifiers (Addgene numbers or backbone sequences/maps where non-commercial).

9. Provide the flow-cytometry gating strategy. For the ~90% viability figure, give the gating logic (scatter gates, LIVE/DEAD channel and threshold), the laser/detector assignment, the analysis software, and the number of independent preparations with a mean and range or SD rather than a single approximate value.

10. Resolve reference 133 and the incomplete in-press citations. Reference 133 is cited in the inside-out patch clamp Methods but does not appear in the reference list. Also update refs 14, 15, 75, 88 and 93 with final citations if now published, or label them clearly as preprints/in press.

11. Complete the mass-spectrometry search reporting. Add the ProLuCID version, precursor and fragment mass tolerances, the complete variable/fixed modification list, missed-cleavage allowance, and the DTASelect filter settings used to reach <1% peptide FDR. State whether protein-level FDR was also controlled.

12. Correct the numerical and notation errors flagged by the citation audit. The reported HCN2 V₁/₂ values include impossible negative error terms ("−83.5 ± −0.8 mV", "−84.0 ± −0.8 mV"); correct these. Harmonise the WX-02-46 IC₅₀ between text (14 µM) and Figure 5L (14.4 µM). For the gel-ABPP IC₅₀ fits, report the fitting model, Hill slope and goodness-of-fit alongside the confidence intervals.

13. Report n explicitly for figures currently labelled only "representative." For each gel-ABPP panel and for the inside-out recording in Figure 6H, state the number of independent experiments and, for Figure 6H, the number of patches and the variability across them, rather than showing a single trace with no replication figure.

14. State replicate structure and negative-control detail for the PDE7B activity assay. For Figure S3F, report n, the positive-control inhibition achieved by IBMX and BRL-50481 (as percentage or IC₅₀), and the assay's dynamic range, so that the null result for WX-03-57 can be distinguished from insufficient assay sensitivity.

Minor Suggestions

• Temper the DPYSL2 forward-looking language. The claim that these compounds "may offer chemical tools to study the specific functions of the DPYSL2:DPYSL5 complex" would be better supported with an estimate of complex abundance relative to total DPYSL2 (from your existing co-IP or IP-MS data, even semi-quantitatively). Absent that, "may enable" is preferable to any wording implying an established tool.
• Similarly, the structural interpretation in Figure 4H (C504 at the hetero-interface) is a reasonable inference from the homo-tetramer structure but is not demonstrated; label it explicitly as a model.
• Consider reporting whether an orthogonal check for off-target cysteine engagement was performed in the slice preparation, or noting its absence as a limitation. The ivabradine/RO-275 controls establish that the stereoprobes do not act at the pore site but do not speak to proteome-wide selectivity at 20 µM in tissue.
• Explain, even speculatively but with reference to structure, why the CNBD cysteine-to-alanine/serine mutants are basally dysfunctional (Figures S6A, S6C). Readers will ask whether the mutation perturbs cAMP binding, folding or trafficking, and the current single sentence leaves the alternative that C542 is structurally load-bearing unexamined.
• Report the cellular composition of the brainocyte suspension (neuron/astrocyte/oligodendrocyte/microglia fractions), even approximately from your flow data or marker immunoblots. This would considerably strengthen the interpretation of Figure 2G, which currently infers cell-type origin from public RNA-seq alone.
• Note in the CNS-enrichment methods that thresholds (Z > 4 in ≥2 datasets; signal > 150 bioGPS / > 10 GTEx) are post-hoc, and state whether the composition of Table 1 is robust to modest changes in them.
• The slice pre-incubation is described in oxygenated ACSF, but the acute-slice ABPP experiments used DMEM (Figure S1C). Briefly justify the DMEM choice or note it as a possible source of physiological divergence from the slice electrophysiology.
• Add RRIDs for antibodies and state STR authentication and mycoplasma status for the cell lines; add a code-availability statement covering the R/MATLAB analysis scripts.
• Consider adding a brief statement of whether recordings and slice treatments were blinded to compound identity. Given that the key controls are matched enantiomers, blinding would materially strengthen the stereoselectivity claims and its absence is worth acknowledging.