Alphafold2 modeling of KCTD10 interactions with RNAPII machinery
Data files
Sep 16, 2025 version files 6.31 GB
-
alphafold2_kctd10-dimer.zip
8.17 MB
-
alphafold2_kctd10-RNAPII.zip
6.30 GB
-
README.md
2.92 KB
Abstract
During DNA replication, the replisome must remove barriers and roadblocks including the transcription machinery. Transcription-replication conflicts (TRCs) occur when there are collisions between the replisome and transcription machinery, and are increasingly recognized as an important source of mammalian genome instability. How cells facilitate replisome bypass at sites of TRCs is incompletely understood. Here, we investigated the role of the CUL3-KCTD10 E3 ligase as a sensor for TRCs. We found that the substrate adaptor KCTD10 interacts with both the replisome and transcription machinery and regulates both in unstressed conditions. These bivalent interactions allow KCTD10 to detect co-directional TRCs and facilitate higher-order assembly of KCTD10 complexes that can recruit CUL3 to ubiquitinate the RNA polymerase factor TCEA2. In the absence of KCTD10 there is increased retention of TCEA2 and RNA polymerase complex, causing an accumulation of TRCs and increased DNA damage. These data report results from a screen we performed using Alphafold2 to look for potential protein-protein interactions between KCTD10 and transcriptional regulators.
Dataset DOI: 10.5061/dryad.15dv41p7z
Description of the data and file structure
Screening for KCTD10 interaction partners
FASTA sequences belonging to the gene ontology term “transcription by RNA polymerase II” (GO:0006366) were retrieved from the European Bioinformatics Institute (EMBL-EBI) Gene Ontology and GO Annotations database (www.ebi.ac.uk/QuickGO/). Using linux shell commands, the FASTAs were split, concatenated with the full length FASTA sequence for KCTD10 (UniprotKB ID: Q9H3F6), and then pooled into batches of 100 files. Each batch was run through ColabFold (v1.5.5)9–11 using Tesla T4 GPUs as a first pass, then larger complexes were completed using Tesla A100 GPUs. One protein could not be completed due to limitations of Alphafold2 (CHD7). To prioritize top hits, an iPTM score > 0.4 was chosen, corresponding to > 1 standard deviation above the average iPTM scores for the complex formed between KCTD10 and two copies of PCNA.
Files and variables
Folder Structure
alphafold2_kctd10-RNAPII.zip
|__inputs
|__dx00000
|__dx00001
|__dx00002
|__dx00003
|__results
|__dx00000
|__dx00001
|__dx00002
|__dx00003
alphafold2_kctd10-dimer.zip
|__input
|__results
Data Formats
.fasta
file with concatenated protein sequence for the indicated protein pair.
citations for proteins used in multiple sequence alignment.
configuration file with command options used in alphafold.
Graph of the sequence coverage for the multiple sequence alignment.
Graph of the predicted aligned error plots for all five models. Lower values indicate well-defined relative positions and orientations in the pair.
Graph of the per-residue measure of local confidence (plddt). It is scaled from 0 to 100, with higher scores indicating higher confidence and usually a more accurate prediction.
json file with the aligned error between each residue.
json file with the per residue measure of local confidence and the overall predicted aligned error for the model.
pdb file containing predicted structure of the indicated protein pair.
multiple sequence alignment file for protein against list of references.
Code/software
Protein structures can be visualized by loading pdb files into PyMOL (https://www.pymol.org/) or UCSF ChimeraX (https://www.rbvi.ucsf.edu/chimerax/). Error plots can be generated using the corresponding json file in ChimeraX.
Bibtex files can be opened with any citation manager software (e.g. Zotero, Endnote).