The C-terminal TPRs of IFT172 constitute a binding site for IFT-A subunits.

(A) Schematic representation of Homo sapiens (Hs) IFT172 protein domain organization. β-propeller and TPR (tetratricopeptide repeat) domains are indicated. Solid lines represent previously identified IFT57 and IFT80 binding sites. Dashed lines show various truncated constructs of Chlamydomonas reinhardtii (Cr) IFT172 and HsIFT172 used in this study.

(B) Cartoon representation of the AlphaFold predicted structural model for a complex between CrIFT172968-C and CrIFT144. The interaction interface is highlighted.

(C) Predicted Aligned Error (PAE) plot for the AlphaFold model in panel B. X- and Y-axes show indexed residues from each protein used for PAE calculation. Low PAE scores indicate high confidence in the predicted interaction interface.

(D) Cartoon representation of the AlphaFold predicted structural model for a complex between CrIFT172968-C and CrIFT140, highlighting the interaction interface.

(E) PAE plot for the AlphaFold model in panel D, showing high confidence (low PAE values) for the interaction interface residue pairs.

(F) Left: Structural superposition of interaction interfaces from AlphaFold models in panels B and D. Subunits are colored as indicated. IFT144 and IFT140 are predicted to associate with an identical binding site formed by IFT172 TPR helices αA and αB. The L1615P temperature-sensitive mutation identified in the fla11 strain of C. reinhardtii maps onto helix αA. Right: Surface amino acid conservation map for the IFT144/140 binding site in IFT172, color-coded as indicated.

(G) PAE plot for an AlphaFold predicted structure of the CrIFT172968-C-CrIFT139 complex. High PAE scores for all inter-chain residue pairs suggest that IFT172 and IFT139 do not interact directly.

IFT172 contains a U-box like domain distal to the IFT-A binding site.

(A) Cartoon representation of the AlphaFold-predicted structure of the C-terminal residues 1485-1755 of CrIFT172. This globular domain comprises several TPR helices, including the IFT140/144 binding αA and αB helices, followed by a loop region (L0) and a U-box/RING-like motif.

(B) Cartoon representation of the 2.1Å resolution X-ray crystal structure of HsIFT172C2. The structure exhibits a similar fold to the CrIFT1721485-1755 AlphaFold model, composed of TPR helices followed by a loop region (L0) and a U-box/RING-like motif.

(C) Structural comparison of the U-box/RING-like motif in HsIFT172 with canonical RING domains of HsRNF4 (PDB: 4PPE), RING1 domain of PARKIN (PDB: 6HUE), and U-box domain of ScPRP19 (PDB: 6BAY), indicating the corresponding RMSD after superposition with the IFT172 motif. The U-box/RING-like motif in HsIFT172 superposes well with several structural components of U-box/RING domains, including the first loop (L1), the following beta strands (β1 and β2), and the helix (α1). A major difference in this HsIFT172 motif is the replacement of the characteristic second loop region (L2) found in U-box/RING domains with an alpha helix (α2).

(D) (Left) Structural superposition of the IFT172 U-box/RING-like domain with the RING domain of RNF4, identifying corresponding Zn2+ binding sites. Zn2+ ions in RNF4 are depicted as grey spheres. (Right) Anomalous density map represented as a magenta mesh contoured at 5σ, obtained from HsIFT172C2 selenium-methionine substituted protein crystals. Anomalous density depicted in proximity to the U-box/RING-like domain. Cys residues and neighboring Met residues in the predicted Zn2+ binding site of IFT172 are represented as sticks.

(E) (Left) Structure of the U-box domain in D. rerio CHIP bound to D. rerio UbcH5a (PDB: 2OXQ), indicating the E2 binding site on the CHIP U-box domain. (Right) Comparison of HsIFT172C2 crystal structure with PDB: 2OXQ indicates the putative E2 binding site on IFT172 U-box is occluded by the IFT172 TPR domain.

(F) Sequence alignment of U-box domains in HsIFT172 and CrIFT172 with several canonical U-box/RING domains. Zn2+ coordinating residues on RING domains are highlighted. Selected functionally relevant residues on U-box domains are depicted in boxes, and corresponding residues are also observed in the IFT172 U-box domain sequence.

X-ray diffraction data collection and refinement statistics (PDB: 9H2D).

IFT172 exhibits ubiquitin conjugation activity in the presence of UbcH5a.

(A) Western blot analysis of an in vitro auto-ubiquitination assay containing M. musculus Ube1(E1), HsUbcH5a (E2), Ubiquitin (Ub) and ATP in the presence of various C-terminal IFT172 constructs as potential E3 ligases. Reactions were visualized by immunostaining with anti-Ubiquitin antibody.

(B) Western blot analysis of in vitro autoubiquitination assays containing Ube1 (E1), Ub, ATP and HsIFT172C1, in the presence of 11 different ubiquitin-conjugating E2 enzymes (Abcam #ab139472). Reactions were visualized by immunostaining with (top) anti-Ubiquitin antibody and (bottom) Coomassie staining. Reactions were conducted under non-reducing conditions, accounting for the visualization of significant amounts of E1∼Ub conjugate in the blot.

(C) In vitro auto-ubiquitination assays containing M. musculus 6XHis-Ube1, 6XHis-TEV-HsUbcH5a, Ub and 6XHis-TEV-HsIFT172C1. Reactions were visualized by immunostaining with an anti-His tag antibody.

(D) Western blot analysis of In vitro auto-ubiquitination reactions containing M. Musculus Ube1(E1), HsUbcH5a (E2) WT/C85S mutant, Ubiquitin (Ub), HsIFT172C1 and ATP. As specified in each reaction, a reaction component was omitted (indicated by Δ) or a mutant component was used instead of the corresponding WT component (indicated by +). Reactions were visualized by immunostaining with (top) anti-Ubiquitin antibody and (bottom) Coomassie staining.

(E) Analysis of the putative E2 binding site in the (center) HsIFT172 U-box domain and (Left) U-box domain of ScPRP19. Both U-box domains are shown facing the E2 binding site. The PRP19 residues I5, Y31, and P39 whose mutagenesis leads to the loss of its ubiquitin ligase activity are represented as sticks (also highlighted in boxes within the sequence alignment in Fig. 2E). The equivalent residues in the putative E2 binding site of IFT172 are also depicted as sticks. (Right) Surface amino acid conservation map for the E2 binding site in IFT172 U-box domain, color-coded as indicated.

(F) Western blot analysis of in vitro ubiquitination assay with HsIFT172C1 WT and the specified HsIFT172C1 U-box variants. Reactions were visualized by immunostaining with (top) anti-Ubiquitin antibody and (bottom) Coomassie staining.

IFT172 U-box domain is a binding site for ubiquitin.

(A) Pulldown of purified UbcH5aC85S∼Ub conjugates with GST-tagged HsIFT172C2 immobilized on GSH beads. Samples after elution were analyzed by western blotting with (top) anti-ubiquitin antibody and (bottom) Coomassie staining. Input shown is 2% for the western blot and 7.8% for the Coomassie staining.

(B) PAE plot for an AlphaFold-generated structural model for a complex between HsIFT172C3 and HsUbcH5a. High error values are observed for interchain residue pairs, suggesting no high confidence prediction for an interaction between the two chains.

(C) Pulldown of tetra-ubiquitin with GST-tagged HsIFT172C2 constructs immobilized on GSH beads. The D1605R mutation on the IFT-A binding site of HsIFT172C2 does not impact the binding of tetra-ubiquitin to HsIFT172C2. Reactions were visualized by immunostaining with (top) anti-ubiquitin antibody and (bottom) Coomassie staining.

(D) Pulldown of tetra-ubiquitin with various GST-tagged HsIFT172 constructs immobilized on GSH beads. Both HsIFT172C2 and HsIFT172C3 pulldown tetra-ubiquitin at similar levels. The prominent lower molecular weight band in the HsIFT172C3 sample is a degradation/proteolytic cleavage product obtained upon HsIFT172C3 expression and purification from E. coli. The two lanes showing a pulldown of tetra-ubiquitin with HsIFT172C3 represent technical replicates. Reactions were visualized by immunostaining with (top) anti-Ubiquitin antibody and (bottom) Coomassie staining.

(E) AlphaFold-predicted structural model for a complex between HsIFT172C3 and HsUbiquitin. The model suggests that ubiquitin binds to the predicted E2∼Ub binding site of HsIFT172C3. The predicted E2 binding residues of the HsIFT172 U-box domain are depicted in the model.

(F) PAE plot for the AlphaFold structural model shown in panel E. Moderate PAE scores are observed for the residue pairs corresponding to the interaction interface (panel E) between the two chains.

Truncation of the IFT172 U-box domain impairs ciliogenesis and leads to altered TGFB signaling response in RPE1 cells.

(A) Schematic representation and nomenclature of RPE1 cell lines generated by CRISPR/Cas12a-mediated exon targeting of the IFT172 gene.

(B) Quantification of ciliogenesis (percentage of ciliated cells, upper panel) and ciliary length (lower panel) in all RPE1cell lines. Error bars in the upper panel represent standard error of the mean (*: p<0.05).

(C) IFM analysis of IFT172-GFP (green) localization to primary cilia (acetylated tubulin (Ac-tub., red) in RPE1 cell lines. Nuclei are stained with DAPI (blue). The top row displays whole-cell views, while the bottom row panels show zoomed in insets of the cilium (arrows). Asterisks indicate ciliary base region.

(D) WB analysis of IFT172 expression in IFT172-FL (heterozygous) and IFT172ΔU-box (heterozygous) RPE1 cell lines.

(E) WB analysis of phosphorylation levels of SMAD2 (p-SMAD2) and AKT (p-AKT; p-AKTT308) in IFT172-FL (heterozygous) and IFT172ΔU-box (heterozygous) RPE1 cell lines treated with 2 ng/mL TGFB-1 ligand for the indicated time points.

(F,G) Quantification of p-SMAD2 (F) and p-AKT (G) levels from panel E, normalized to DCTN1 and GAPDH. Error bars represent standard error of the mean (*: p<0.05; **:p<0.01).

List of primary and secondary antibodies used in western blots

Uncovering interactors of the CrIFT172 C-terminus

(A) Size exclusion chromatography (SEC) elution profile for CrIFT172968-C (top). Protein composition of denoted elution fractions analyzed by Coomassie staining after SDS-PAGE separation (bottom).

(B) Chlamydomonas reinhardtii CC1690 cells visualized by light microscopy before flagella isolation (top) and purified flagella fraction after de-flagellation at the same magnification (bottom).

(C) Volcano plot showing distribution of mass spectrometry (MS) analysis hits for flagellar proteins pulled down by CrIFT172968-C in C. reinhardtii CC1690, compared against the tobacco etch virus (TEV) protease control.

(D) Table of the 10 significant CrIFT172968-C flagellar interactors identified in panel C. Gene name(left) and protein annotations from Uniprot (right) are shown.

(E) Alphafold predicted structural model for a complex between CrIFT172968-C and the UBX domain containing protein (CHLRE_06g293900v5) identified as an interaction partner (Fig. S1C-D). A linker region that lies between the UBX and SEP domains is predicted to form contacts with IFT172. The IFT-A binding helices described in Fig. 1F are denoted as αA and αB.

(F) PAE plot for the AlphaFold predicted structure shown in panel E.

Purification and structure determination of HsIFT172 C-terminal domain.

(A) SEC elution profile for HsIFT172C2 shown on top. The protein composition of the denoted elution fractions was analyzed by Coomassie staining on an SDS-PAGE (bottom).

(B) Representative electron density map for HsIFT172C2 crystals. The MR-SAD map is shown as a blue mesh contoured at 1σ and SeMet anomalous density is shown as a magenta mesh contoured at 5σ.

(C) Top 10 hits obtained from a search of structural homologs for HsIFT172C3 against the PDB database using the DALI server49. Z-score and RMSD of corresponding structural alignment are indicated.

(D) Phosphorylation sites on HsIFT172C2 identified from curated databases of known phosphorylation sites76. The phosphorylation sites on HsIFT172C2 are exclusively present at the TPR/U-box interface, suggesting phosphorylation as a potential mechanism for relieving the structural inhibition of the U-box E2 binding site.

(E) Representative hydrophobic and polar contacts that allow the IFT172 U-box domain to pack against the TPR helices.

Purification and auto-ubiquitination assays with HsIFT172C constructs.

(A) In vitro ubiquitination reactions performed with HsIFT172C1, followed by incubation with the specified concentrations of the deubiquitinase enzyme USP2 (R&D Systems # E-504-050) or buffer control. Reactions were visualized by immunostaining with (top) anti-Ubiquitin antibody and (bottom) Coomassie staining.

(B) Western blot analysis of in vitro ubiquitination reactions containing HsIFT172C1 after separation on a non-reducing SDS-PAGE (left) and reducing SDS-PAGE gel (right). Reactions were visualized by immunostaining with (top) anti-Ubiquitin antibody and (bottom) anti-His tag antibody. Δ indicates reaction components that were omitted in the specific reaction.

(C-F) SEC elution profiles (top) and protein composition of the denoted elution fractions analyzed by Coomassie staining post separation on SDS-PAGE (bottom) for: (C) HsIFT172C1 WT. (D) HsIFT172C1 P1725A. (E) HsIFT172C1 C1727R. (F) HsIFT172C1 F1715A. (1) and (2) denotes two different HiLoad Superdex 200 columns used to run the samples.

Purification of GST-tagged HsIFT172 constructs and of the E2∼Ub conjugate.

(A) SEC elution profile (top) and protein composition of the denoted elution fractions analyzed by Coomassie staining after separation on SDS-PAGE (bottom) for His-GST-TEV-HsIFT172C2.

(B) SEC elution profile (top) and protein composition of the denoted elution fractions (bottom) for His-GST-TEV-HsIFT172C3. A significant amount of proteolytically cleaved protein fragments was observed upon His-GST-TEV-HsIFT172C3 purification.

(C) Reaction products of an upscaled in vitro ubiquitin charging reaction for UbcH5aC85S WT conjugate separated on a SEC column (top). The denoted elution fractions were analyzed by Coomassie staining on an SDS-PAGE gel (bottom).

Truncation of the IFT172 U-box domain impairs ciliogenesis and leads to altered TGFB signaling response in RPE1 cells.

(A) WB analysis of p-AKT levels (p-AKTT308 and pAKTS473) in IFT172-FL (heterozygous) and IFT172ΔU-box (heterozygous) RPE1 cell lines treated with PDGF-DD ligand for the indicated time points.

(B) Quantification of p-AKT levels in panel A normalized to DCTN1 and total AKT. Error bars represent standard error of the mean.