Fig 2: Differential HDX-MS analysis of His-CTNNBL1[77–563], CDC5L[1–195] and CWC15[80–229] complexes. CTNNBL1 surface model displaying areas of interaction with (A) CWC15 and (B) CDC5L, deduced from above-threshold hydrogen/deuterium content decrease (blue) and increase (orange) (>10%, >0.5D, t-test P < 0.05, see Materials and Methods). The view shows the N-terminal region on the left and a 180° horizontal rotation. The mean of maximum exchange (?HDX) values plotted for each amino acid corresponding to each protein are shown (see Materials and Methods) in (C) for His-CTNNBL1[77–563] in complex with CWC15[80–229] (purple line); CWC15[80–229] as a dimer with CDC5L[1–195] (red line) or His-CTNNBL1[77–563] (blue line) in D); and for CDC5L[1–195] in association with CWC15[80–229] (purple line) in (E). In all (C)–(E), the ?HDX profiles associated with the ternary complex are shown as a green line. Regions with a decrease in exchange above the significance threshold indicated by the dotted lines are shadowed grey. Peptide coverage is indicated by yellow bars at the bottom of each graph on the x-axis which gives the aa numbering. The position of NLS1 in CDC5L is indicated by a black bar.

Fig 3: CWC15 depends on CTNNBL1 for efficient association with the Prp19 complex in vivo. (A) Abundance of the endogenous Prp19 complex proteins relative to ß-actin in resting B cells from two control mice (+/-) and two CD19-Cre Ctnnbl1-/flx (-/-) mice. (B) Abundance of CDC5L, CWC15 and CTNNBL1 in cell lysates of human U2OS cells 24 h after transduction with siRNA pools specific for either CDC5L or CWC15. Non-targeting siRNA pools were used as control. A representative western is shown, with abundance of the proteins relative to ß-actin quantified below showing the average ±SEM from three independent experiments. (C) Immunoprecipitation (IP) of mouse CDC5L and PRPF19 in retrovirally transduced B cells expressing human HA-CWC15 or empty vector (ctrl). Transduced B cells from either CD19-Cre Ctnnbl1+/flx control mice (+/-) or CD19-Cre Ctnnbl1-/flx (-/-) were stimulated with LPS and IL-4 for 4 days prior to lysis and purification on anti-HA resin. Lysates show 1.5% of the input used for IP. The amount of lysate used in the IPs was adjusted to account for the reduced levels of CDC5L and PRPF19 in the (-/-) samples. Size markers (kDa) are indicated. (D) Splicing efficiency estimated from the ratio of first intron inclusion versus spliced messenger RNA measured by q-PCR on total RNA extracted from resting or activated B cells from either CD19-Cre Ctnnbl1+/flx control (+/-) or CD19-Cre Ctnnbl1-/flx (-/-) mice at the indicated times post activation. The average ±SEM ratios are obtained from three independent mice. Statistical P values of an unpaired one-tail t-test for each time point are shown. The right histograms show the relative abundance mRNA relative to ß-2 Macroglobulin (B2M) in control (+/-) and CTNNBL1 deficient (-/-) B cells. Statistical analysis using a two-tail unpaired t-test reported no P-values below 0.1.

Fig 4: Architecture of the CWC15, CTNNBL1 and CDC5L sub-complex in relation to the core Prp19 complex. (A) Architecture of the mammalian Prp19 complex. Based on biochemical and structural information, including negative staining electron microscopy of purified proteins, the core Prp19 complex is thought to consist of an extended structure formed by a tetramer of PRPF19 with the coiled-coil regions in its centre and the WD40 domains on the outside. A single copy of CDC5L is associated with the coiled-coil region of the PRPF19 tetramer through its less conserved carboxy-terminal region, designated as the spliceosomal interacting domain, whereas the more conserved N-terminal portion containing the MYB domains and NLS region is thought to associate with CTNNBL1. One copy of SPF27 is also found in association with the central portion of the PRPF19 tetramer which also contains the N-terminal region of PLRG1. CWC15 and CTNNBL1 have been described as a separate stable complex outside the Prp19 core, but are also found in association with the Prp19 complex through interaction with the N-terminal region of CDC5L. The precise configuration of the CTNNBL1/CWC15/CDC5L sub-complex (dotted ellipse) within the core Prp19 complex had not been fully defined but was expected to adopt a linear conformation. The positioning of HSP7C, a sub stoichiometric component has not been characterized (Model based on Grote et al. (30)). (B) Changes in the interactions between CWC15, CTNNBL1 and CDC5L in relation to the different spliceosomal complexes. CTNNBL1 and CWC15 form a stable dimer with the N-terminal of CTNNBL1 bound to the C-terminus of CWC15 (orange). In the B and Bact forms of the spliceosome, the Prp19 complex is stabilized by the presence of CTNNBL1, enhancing the otherwise weak interactions between CWC15 and CDC5L. CTNNBL1 acts as bridge, binding the N-terminal NLS region of CDC5L (green squiggle) through its CTD region and the C-terminus of CWC15 through its N-terminal HEAT-like region. This allows for a three-way interaction facilitating binding between the MYB domain of CDC5L (green triangular forms) and a region in the central portion of CWC15 (red squiggle). This stable three-way interaction maintains the normal levels of the Prp19 complex in B and Bact spliceosomes. At the later C stage of the spliceosome complex, CTNNBL1 is not present, and the NLS region of CDC5L can interact with the C terminal region of CWC15.

Fig 5: CTNNBL1 interacts differently with CDC5L or CWC15 but all three form a stable sub-complex. (A) Pull downs of His-CTNNBL1[77–563] on Ni-NTA in the presence of increasing potassium acetate (KAc) concentrations (0, 0.05, 0.1, 0.15, 0.2, 0.3, 0.4, 0.6, 0.8 and 1 M) and either MBP-CDC5L[1–195] on the top or GST-CWC15[80–229] bottom. A control pull down (-ctrl) using beads was performed in 0.11 M KAc. Asterisks indicate bait of pull downs and size markers (kDa) are shown. For each experiment, the proteins used for pull down are labelled input and shown below. (B) SDS-PAGE analysis of recombinant human His-CTNNBL1[77–563], CDC5L[1–195] and CWC15[80–229] protein mix after Superdex-200 gel filtration. The elution volume (ml) of the fractions analysed and the fractions eluting molecular weight markers of known size (alcohol de-hydrogenase (ADH) and albumin) are indicated. (C) As in B, a protein mixture of purified His-CTNNBL1[77–563], CDC5L[1–195] and CWC15[80–229] analysed by SEC-MALS. The grey curve shows the refractive index (arbitrary units) against the elution volume (ml) (right y-axis). The horizontal black lines indicate the molar mass distribution in g/mol (left y-axis). The calculated molar mass for the two observed peaks are indicated, with a large peak corresponding to a protein complex of 99 kDa and a second small protein complex of 25 kDa which corresponds to the predicted molecular mass of monomeric CWC15[80–229] fragment. Based on the UV absorbance of the eluted main peak (thin line) a 102 kDa protein complex with a 1:1:1 stoichiometry could be inferred based on the extinction coefficient of the three protein fragments in the mix.