كتابة تعليق

In Cisse et al., the authors identified the BTB domains of a subclass of ZBTB proteins as interactors and efficient substrates for SUMO conjugation by the E2 UBC9 in the absence of SUMO E3 ligases. Previously, ZBTB proteins were identified as SUMO-modified substrates by proteomics, but their E2-dependent SUMO conjugation has not been demonstrated using recombinant proteins in vitro. This study combines crystallography, AlphaFold3 modeling, and biochemical assays that support the conclusion that the BTB domain of ZBTB38 interacts directly with UBC9 and is SUMO-modified efficiently because it contains a structural mimetic of a SUMO consensus motif that positions Lys43 of ZBTB38^BTB into the UBC9 catalytic site in an analogous manner to previously structurally characterized UBC9 substrates. This finding should be interesting to the field because it is unique among most other substrates, as its SUMO modification site lies within a beta-sheet of the BTB domain rather than in an unstructured region. Moreover, the site does not lie within a linear consensus motif; therefore, it is not readily identifiable by sequence alone. The structural models support the authors’ hypothesis that ZBTB38^BTB is an efficient substrate because it pre-pays the entropic cost of binding a flexible linear SUMO consensus motif. While the biochemical assays are consistent with their model, the findings would be strengthened by optimizing the assay conditions and conducting additional analyses of key residues within the ZBTB38^BTB-UBC9 interface, especially residues on the E2 surface that are predicted to be important for ZBTB38^BTB binding. The following suggestions are intended to help clarify the evidence supporting the proposed mechanism.

Major points

1. The authors assess substrate SUMOylation under a single assay condition, using relatively high concentrations of all components. In Figures 3C and 3E, the authors conclude that ZBTB38^BTB and RANGAP1^CTD have similar catalytic rates and efficiencies; a similar conclusion is drawn for ZBTB33^BTB in Figure 4D. While this is true to a first approximation with substrate present at 15 µM, it may not be true at lower or limiting concentrations. The authors are encouraged to include additional assays with a concentration range between 0.1 and 10 µM, with analysis of data within the linear range of the assay (first 10 minutes). In addition, SUMO was largely consumed and became limiting in these assays, which also weakens the conclusions at later time points. The authors could titrate the protein components to ensure that the measured rates truly reflect substrate SUMOylation by UBC9. Finally, claims that ZBTB38 is as efficient as RANGAP1 could be tested directly by putting both substrates into a single reaction at comparable concentrations to see if they can compete for UBC9 when in excess.

2. The discussion of key residues within the ZBTB38^BTB-UBC9 interface, illustrated in Figure 1F, and the corresponding mutagenesis analysis in Figures 2 and 3 could be expanded. This reviewer used AlphaFold3 to analyze the ZBTB38^BTB-UBC9 structure, as the PDB coordinates have not yet been released. There are two additional interactions that appear interesting and could be included. First, S71 of ZBTB38^BTB lies within hydrogen-bonding distance of Y134 of UBC9. Second, E75 of ZBTB38^BTB interacts with the N-terminal end of a UBC9 helix, forming a hydrogen bond with the A131 backbone. Therefore, S71 and E75 could be tested and included in Figures 2 and 3. Furthermore, mutational analysis of selected residues on the surface of UBC9 could further test the proposed interface. Candidates include Y87, S89, T91, D127, and A131. Y87 is included given its previously characterized role in E2-dependent conjugation assays with other substrates.

3. The authors assess the interaction between ZBTB38^BTB and UBC9 using size exclusion chromatography (Figures 2A-C). However, the SDS-PAGE band intensities do not correlate well with the chromatograms, and evidence for co-elution is limited to subtle shifts in band intensities in the same fractions rather than movement of bands into adjacent fractions. The authors could attempt to collect smaller fraction sizes, or they could use a more sensitive reagent for protein staining to see if that augments any co-elution signal.

4. The authors measure the dissociation constant of the ZBTB38^BTB-UBC9 complex (Figure 2D). This is a direct binding assay that could alleviate the issues related to the use of size exclusion chromatography. It would also afford the authors the opportunity to directly test different ZBTB38^BTB mutants that they presume disrupt binding. This would greatly strengthen any claims that interfacial residues on the BTB surface contribute to binding of UBC9.

5. The authors suggest that ZBTB38 is SUMOylated in cells, as shown by western blots of cell lysates that they treated with NEM to inactivate proteases to preserve the modification (Figure 3G). The authors labeled the upper band as SUMOylated ZBTB38, but the current data do not definitively establish the identity of this band. ZBTB38 can also be ubiquitinated, according to previous literature, and NEM treatment may augment the ubiquitination signal as well. As such, the authors could immunoprecipitate ZBTB38 from lysate and probe with anti-SUMO or anti-Ub antibodies to strengthen the claim.

Minor points

1. The authors discuss the sequence conservation of ZBTB38^BTB mapped onto the structure (Figure 1C). However, it is unclear which sequences were used as input for the multiple sequence alignment and what thresholds were applied. The authors may consider providing a table or list of accession codes for the sequences aligned in this panel, also specifying thresholds in the Methods section. As noted above, ZBTB38^BTB E75 makes a seemingly important contact to UBC9; the position of this residue and the two hydrophobic residues I38 and V73 could be indicated on this surface conservation map.

2. The buried surface area and estimated binding energy reported on page 3 would benefit from supporting references or a description of the calculation methods in the Methods section.

3. In the SUMO conjugation assay (Figures 3A-D), the authors state that SUMO and ZBTB38^BTB co-migrate on SDS-PAGE but conclude that the band above 25 kDa is SUMOylated ZBTB38^BTB. While this is likely to be true based on mutagenesis results, an orthogonal method, such as western blot or mass spectrometry, would help confirm the identity of the conjugate. Another option might be to include full-length SUMO in the assays, as it would migrate higher than ZBTB38^BTB.

4. The MBP-P53^CTD was used in the SUMO conjugation assay (Figure 3D). Because the modification efficiency appears substantially lower than that reported in previous studies (e.g., PMID: 16732283), the authors could consider removing the P53^CTD peptide from the MBP fusion tag to assess whether it impairs UBC9 recognition of P53^CTD.

5. The authors overexpressed GFP-ZBTB38 in HEK293 cells (Figure 3H). Why a GFP-tagged construct is used in the assay is not explained (although perhaps obvious to some). The authors may explain the rationale in the main text to improve clarity. Also, the entire, uncropped western blot for each panel (3F, G, H) could be included in the supplement to help readers better assess antibody specificity.

6. The authors refer to kinetics in several places; this is typically the word used to characterize enzyme behavior as it pertains to rate and binding. Since this was not done, the authors could consider changing the word to “time course” or similar. Referring to k_cat/K_M on page 12 in reference to Figure 3C and E is also unwarranted given the limitations of the assays used in this study. This also pertains to the Methods section titled “SUMOylation kinetics”; this is a time course using a single condition (E1, E2, substrate, SUMO, etc.) where k_cat and K_M are not determined.

7. Figure 5 includes summary graphics for multiple studies and would appear to this reviewer more suitable for a review, as the most relevant portion of this figure related to the presented data is Figure 5E. This will keep the article focused on the contribution of these findings to the field.

8. The X-ray crystallographic table appears reasonable (Table 1), but without the full validation report from the RCSB, this reviewer is unable to provide additional comments on the structure validation.

I reviewed this preprint as part of the Transparent and Accountable Peer Review training program. This report was written with the support of my research mentor.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.