Structural and Functional Insights into Human Nuclear Cyclophilins

Abstract

The peptidyl prolyl isomerases (PPI) of the cyclophilin type are distributed throughout human cells, including eight found solely in the nucleus. Nuclear cyclophilins are involved in complexes that regulate chromatin modification, transcription, and pre-mRNA splicing. This review collects what is known about the eight human nuclear cyclophilins: peptidyl prolyl isomerase H (PPIH), peptidyl prolyl isomerase E (PPIE), peptidyl prolyl isomerase-like 1 (PPIL1), peptidyl prolyl isomerase-like 2 (PPIL2), peptidyl prolyl isomerase-like 3 (PPIL3), peptidyl prolyl isomerase G (PPIG), spliceosome-associated protein CWC27 homolog (CWC27), and peptidyl prolyl isomerase domain and WD repeat-containing protein 1 (PPWD1). Each "spliceophilin" is evaluated in relation to the spliceosomal complex in which it has been studied, and current work studying the biological roles of these cyclophilins in the nucleus are discussed. The eight human splicing complexes available in the Protein Data Bank (PDB) are analyzed from the viewpoint of the human spliceophilins. Future directions in structural and cellular biology, and the importance of developing spliceophilin-specific inhibitors, are considered.

Keywords: NMR; X-ray crystallography; alternative splicing; nuclear cyclophilins; peptidyl prolyl isomerases; spliceophilins; spliceosomes.

Figure 1

Structure and annotation of the canonical cyclophilin PPIA. (A) Left panel, secondary structural elements are labeled for the PDB ID: 2CPL. Right panel, residues that comprise the S1 (proline-binding) and S2 (specificity) pockets are shown in stick representation and labeled. S1 pocket labels are on the right, S2 pocket labels on the left, and in bold are residues invariant across the nuclear cyclophilin family. Residue numbering also follows PPIA, by convention. (B) The residues that comprise the S1 pockets are largely invariant, while (C) the residues that comprise the S2 pockets are more variable.

Figure 2

Domain organization of the nuclear cyclophilins. Colors are consistent throughout all figures. Spliceosome association is modified from [17,18] and is further shown in Figure 3.

Figure 3

Cryo-electron microscopy structures of spliceosomal complexes. Center, a graphical representation of the canonical metazoan pre-mRNA substrate, including the 5′ and 3′ exons, the consensus 5′ and 3′ splice recognition sites, the presence of the branchpoint (BP) that forms the intermediate splicing lariat structure, and the polypyrimidine tract (Y(n)). The classical graphical representation of the splice cycle, shown in the center of the figure, follows the association and dissociations of uridine-rich small nuclear RNAs (U-snRNAs) to characterize the splicing intermediate (E, A, B, B_act, B*) and catalytic (C) complexes. Complexes discussed in the text are shown on the outside of the splice cycle, with spliceophilins labeled and highlighted in color. The key U snRNAs U2, U5, U6 are present in all structures, and their rough positions labeled in each structure. The structures are roughly aligned to the part of the splice cycle they are thought to represent: B complex (PDB ID: 5O9Z) [24], B_act (PDB IDs: 5Z57, 5Z58, 5Z56, and 6FF4) [25,26], or C (PDB IDs: 5YZG, 5XJC, and 5MQF) [27,28,29].

Figure 4

The interaction between PPIH and PRPF4 seen in solution is preserved in early spliceosomes. (A) The complex between a small peptide from PRPF4 and the PPIH isomerase domain is shown in cartoon representation (PDB ID: 2MZW) [40,41,42]. Key catalytic site residues are highlighted, and in the right panel the high-affinity interaction between Phe122 and the α1–β3 loop of PPIH is shown. In (B), a table summarizing the high- and low-affinity sites between PRPF4 and PPIH. Below, a model of the proposed function of the high- and low-affinity sites. Panels (A,B) are modified from their original form in [32]. (C) The neighborhood around PPIH in the B complex human spliceosome (PDB ID: 5O9Z). In addition to the previously known interaction with PRPF4, potential interactions with WBP4, PRPF6, and PRPF8 are highlighted. U4 snRNA is within 10 Å of PPIH and is labeled. Proteins in dark grey are further than 10 Å away from PPIH and are not candidates for direct interaction, so are not labeled. The region of PRPF4 forming the low-affinity interaction with PPIH (the extreme N-terminal 100 residues) is disordered in this structure. Ac: Acetylation; SUMO: Small Ubiquitin-Like Modifier.

Figure 5

Structures of PPIE in solution and in spliceosomes. In (A), the cyclophilin and RRM domains of PPIE are shown in cartoon representation. The isomerase domain is represented by an overlay of PDB IDs: 2R99, 1ZMF, and 3UCH. Selected catalytic residues and protein–protein interaction regions are labeled. The RRM domain is represented by an overlay of PDB IDs: 2CQB, 2KU7, 2KYX, 3LPY, and 3MDF (chain A only). Secondary structure elements are labeled. (B) The RRM of PPIE was previously shown to interact with the PHD domain of MLL through an extensive interface involving multiple β-strands. (C) The neighborhood of the PPIE PPI (left) and RRM (right) domains in the mature B_act spliceosome is shown (PDB ID: 5Z56). The region around Pro83 of SF3B4 is pointed towards the active site of PPIE (marked with an asterisk). Other interactors are in color and labeled. (D) The neighborhood of the PPIE PPI (bottom center) and RRM (top center) domains in the late B_act spliceosome is shown (PDB ID: 5Z57). The protein network surrounding PPIE is unchanged relative to those shown in (C), but in this view SF3A3 is visible, which is just within 10 Å of PPIE isomerase domain. (E) The neighborhood of the PPIE PPI (left) and RRM (right) domains in the C complex spliceosome (PDB ID: 5YZG). Highlighted on the left panel are interactions between the isomerase domain and snRNPA’, B/B’, and SmD3 along with ISY1. Visible in this structure is the position of the pre-mRNA substrate, which is interacting with the RRM motif. In the right panel, other RRM interactors including SYF1, AQR, and snRNPA’. The isomerase domain is visible in the lower right corner for reference. (F) The neighborhood of the PPIE PPI (left) and RRM (right) domains in the C* complex spliceosome (PDB ID: 5MQF). In this model, the interaction with SYF1 and AQR is mediated through the isomerase domain of PPIE rather than the RRM motif, and the model of U2 snRNA is placed extremely close to PPIE. In this structure the orientation of PPIE seems to be flipped, with the isomerase domain interacting with AQR and SYF1, and the RRM motif close to PRPF17.

Figure 6

Structures of PPIL1 in solution and in spliceosomes. In (A) PPIL1 is shown in cartoon representation (overlay of PDB IDs: 1XWN, 2K7N, and 2X7K). Selected catalytic residues and protein–protein interaction regions are labeled, including the regions proposed to interact with SNW1/SKIP in solution. (B,C) Mature B_act complexes contain PPIL1 (PDB ID: 5Z56). (B) The view from the catalytic face of PPIL1. Proline 95 from PRPF17, which is centered in the S1 pocket, is highlighted. (C) The view from the back-face of PPIL1. The interactions between PPIL1 and SNW1, RBM22, CRNKL1, and CDC5L are more clearly seen. Note the extensive ordering of the disordered region of SNW1 around the β4–β5 and β7–α2 loop of PPIL1. The PPIL1 region in the late B_act complex (PDB ID: 5Z57) is identical to that of PDB ID: 5Z56, and is not shown here. (D) The PPIL1 neighborhood in the B_act complex (PDB ID: 6FF4). While many of the interactions are similar to those shown in (B,C), many of the proteins seen in that complex are missing from this model. Left panel, the view from the catalytic face; right panel, the view from the back face. Models of CRNKL1, CDC5L, and SPF27 are not included, and the ordered region of SNW1 is decreased. (E) The PPIL1 neighborhood in the C complex (PDB ID: 5YZG) centered on the catalytic face (left) and the back-face (right). The PPIL1 interaction neighborhood is similar to that shown in (B,C) with the addition of the SYF2 protein to the model, which interacts with the back-face of PPIL1 near α2. The models of C* complex in 5XJC and 5MQF are not shown, as the interaction environment around PPIL1 is very similar to that in (E).

Figure 7

Validation of interactions between PPIL2 and the spliceosomal protein ZNF830. In (A) PPIL2 is shown in cartoon representation (PDB ID: 2ZKC). Selected catalytic residues and protein–protein interaction regions are labeled. (B) The U-box containing PPIL2 protein is a functional E3 ligase. Increasing concentrations of PPIL2 (0.1–10 µM) in an in vitro assay and in the presence of ubiquitin, E1, and E2 proteins lead to the production of poly-ubiquitin chains [54]. (C) Surface plasmon resonance of full-length PPIL2 (ligand) and ZNF830 (analyte) indicates nanomolar-affinity binding between the two proteins. (D) left panel, concentration dependence of isolated PPIL2 U-box binding to full-length ZNF830. Right panel, sequential addition of PPIL2 U-box followed by a mix of PPIL2 U-box and PPIL2 isomerase domain results in increasing response upon addition of ZNF830 zinc finger domain. This is interpreted as an ability for both domains of PPIL2 to interact simultaneously with ZNF830 ((C) and (D) are from [61]). Methods are summarized in Appendix A.

Figure 8

PPIL3 is shown in cartoon representation (PDB IDs: 2XWN and 2OK3). Selected catalytic residues and protein–protein interaction regions are labeled.

Figure 9

Structures of CWC27 in and out of the spliceosome. In (A) the isomerase domain of CWC27 is shown in cartoon representation (overlay of PDB IDs: 2HQ6 and 4R3E). Selected catalytic residues and protein–protein interaction regions are labeled. The substitution of Glu122 in the active site renders CWC27 inactive, although it still binds proline-containing peptides. In (B) the neighborhood around the isomerase domain of CWC27 in the mature B_act complex (PDB ID: 5Z56). Modeled interactions with U5 snRNP 200 kDa, BUD31, PRPF8, and RNF113 are highlighted. In (C), the neighborhood around CWC27 in the late B_act complex (PDB ID: 5Z58). The view is very similar to that in PDB ID: 5Z56, save for the absence of BUD31 and slight movement of PRPF8. In (D), the neighborhood around CWC27 in the B_act complex (PDB ID: 6FF4) is shown. Again, the modeled interactions are very similar to those in (B,C), with more of RNF113 modeled in 6FF4, including an additional, extensive interaction with β1–β2 of CWC27. The model includes BUD31 but U5 snRNP 200 kDa is missing. All models have only the isomerase domain of CWC27, with ≈200 additional residues uncharacterized.

Figure 10

PPWD1 structures in and out of the spliceosome. In (A), one of the three molecules of the isomerase domain of PPWD1 from the asymmetric unit of PDB ID: 2A2N is shown in cartoon representation. Selected active site residues are labeled, along with the α1–β3 loop and the β7–α2 region. In (B), the view is expanded to include all three PPWD1 molecules in the asymmetric unit. The proline from a neighboring molecule is shown in the active site. Modified from [30]. In (C), the PPWD1 neighborhood is shown in the C complex (PDB ID: 5YZG). Only PRPF8 is close enough to form contacts with the isomerase domain of PPWD1. (D) Over 90 Å away, PRPF8 along with U5 snRNP 200 kDa are modeled near the WD40 domain of PPWD1. For reference, the region of PRPF8 that interacts with the isomerase domain of PPWD1 is visible in the upper-right corner of the figure, in color. This represents the first, and as of yet only, model of the WD40-domain of PPWD1.

Figure 11

PPIG structures in and out of the spliceosomal context. In (A), the isomerase domain of PPIG is shown in cartoon representation (overlay of PDB IDs: 2GW2, 2WFI, and 2WFJ). Selected residues in the catalytic site, along with regions of protein–protein interactions, are labeled. In (B), the neighborhood around the isomerase domain of PPIG in the catalytic C spliceosome (PDB ID: 5YZG). Potential interactions between PPIG and SNW1, CWC15, and PRPF8 are highlighted. No model exists for the roughly 600 residues outside of the PPIG isomerase domain.

Figure 12

Spliceophilins impact alternative splicing patterns in human cells. (A) Knockdown of PPIH using RNAi results in changes in cassette exon, 5′ and 3′ splice sites, and intron retention, as measured by the Affymatrix HTA 2.0 array. (B) Compared to control cells (SCR, scrambled RNAi), PPIH knockdown cells (KD, PPIH-specific RNAi) exhibit altered alternative splicing of multiple genes (labeled at top of gels). Methods are summarized in Appendix B. When full-length PPIH is transiently expressed in knockdown cells (REV, PPIH ectopic expression), the splicing changes are reversed. When the W133A mutant of PPIH (MUT, PPIH mutant ectopic expression) is transiently expressed, the KD phenotype is seen, indicating that isomerase activity is dispensable for this function. (C) The knockdown of PPIL2 exhibits splicing patterns identical to that of PPIH knockdown, while PPIG only regulates a subset of genes, and PPIL3 does not regulate any of the common targets tested. Not shown, PPIH KD regulates all splicing events tested to date. All panels from [69]. Methods are summarized in Appendix B.