Discovery of novel murine PML isoforms

Abstract

Promyelocytic leukemia protein (PML) orchestrates the formation of PML nuclear bodies (PML NBs), membraneless organelles with diverse regulatory roles. Despite their importance, the specific functions of individual PML splicing variants remain unclear, particularly in murine models. Here we study the repertoire of murine PML isoforms expressed in mouse tissues and cells. We demonstrate that in addition to canonical mPML1-3, mice express five predicted variants (mPMLX1, mPMLX2, mPMLX4-X6) and a novel isoform, mPMLX7, distinguished by unique RBCC domain splicing. All isoforms exhibit distinct turnover kinetics at endogenous PML NBs. In PML-knockout cells, all isoforms except mPMLX7 form NBs de novo and are degraded upon arsenic exposure. Molecular dynamics simulations suggest mPMLX7 adopts a stable conformation; furthermore, this isoform is enriched in the nucleoplasm, suggesting a specialized function. Altogether, this isoform-resolved PML system provides a relevant model for dissecting the wide spectrum of PML-associated processes.

Keywords: Arsenic trioxide; Mouse PML isoforms; PML; PML isoforms; PML nuclear bodies.

Figure 1.

Expression of mPML isoforms. (a) Schematic representation of the exon and domain structure of the mPML isoforms. Upper panel: the alternative splicing of nine exons produces mRNA encoding nine individual mPML isoforms. The first three (mPML1 to 3) have been previously confirmed, while the other six (mPMLX1 to X6) were computationally predicted. All isoforms share the N-terminus sequence (exons 1–3) harboring the RBCC/TRIM domain consisting of a RING finger, two B-boxes and a coiled-coil (cc) domain. The dashed vertical lines (on the right) link isoforms that differ only in the presence or absence of exon 5. Alternative splicing of exons 7–9 produces unique C-termini for each isoform, with mPMLX3 uniquely retaining intronic sequence: mPML1 includes exons 4, 6–9; mPML2, 4–9; mPML3, 4–8 with a shortened exon 9; mPMLX1, 4–8 with a lengthened exon 9 containing two additional segments; mPMLX2, 4, 6–8 with the same extended exon 9; mPMLX3, 4–7 with retained intron; mPMLX4, 4–7 with a shortened exon 9 distinct from mPML3; mPMLX5, 4, 6–8 with a shortened exon 9 matching mPML3; mPMLX6, 4, 6, 7 with a shortened exon 9 matching mPMLX4; and mPMLX7, a uniquely shortened exon 3 with the remaining exons as in mPML2. All isoforms contain three major SUMOylation sites at positions K70, K165, and K500, a nuclear localization signal (NLS), and a SUMO-interacting motif (SIM). Additionally, isoforms mPML1 and 2 and X1 and X2 share a conserved exonuclease III-like (Exo III) domain. All domains, motifs, signal sequences, and specific sites were reproduced based on previous studies or predicted based on the homologous human PML system [27,34–36,52] (Supplementary figure S10b). Bottom panel: the newly discovered isoform mPMLX7 contains a uniquely shortened exon 3 that results in a six-amino acid C-terminal deletion in the B2 box and the complete loss of the cc domain. Black arrows represent the positions of primers used in this study to amplify the sequences of specific isoforms, whereas gray arrows denote those used for sequencing each isoform. b–d) Endogenous transcription of mPML isoforms. b–c) RT-PCR was performed on RNA isolated from the indicated cell lines using isoform-specific primer pairs (as indicated by black arrows in (a)). The mHprt reference gene was amplified as an internal control for cDNA integrity, and a no-reverse transcriptase (no-RT) control was included for each gene. (d) Semi-quantitative analysis of mPML isoform transcripts using cDNA nanopore sequencing. Schematic representation of the abundance of individual isoform transcripts. The number of reads corresponding to mRNAs encoding each isoform was normalized to the total number of reads (100%). (e) Protein expression of the endogenous mPML isoforms in various cell lines. Whole-cell lysates were analyzed by western blotting with anti-PML antibody. GAPDH served as a loading control. (f) Validation of mPML-tRFP/mPML-S constructs. Full-length mPML protein isoforms with (right panel) or without (left panel) the tRFP tag individually expressed in Pml^−/− MEFs. The cells were transfected with plasmids expressing individual isoforms, and cell lysates were analyzed by western blotting using an anti-PML antibody. GAPDH was visualized as a loading control. Isoform-specific molecular masses (MM) were calculated using the protein molecular weight tool in sequence manipulation suite [53]. The asterisk (*) indicates the approximate molecular mass of the mPML isoforms. Amino-acid (aa) lengths and predicted molecular masses of the isoforms are as follows: PML1, 839 aa, 93 kDa; PML2, 885 aa, 98 kDa; PML3, 632 aa, 70 kDa; PMLX1, 943 aa, 105 kDa; PMLX2, 897 aa, 100 kDa; PMLX3, 645 aa, 72 kDa; PMLX4, 589 aa, 66 kDa; PMLX5, 586 aa, 66 kDa; PMLX6, 543 aa, 61 kDa; and PMLX7, 674 aa, 74 kDa.

Figure 2.

Formation of PML NBs for individual mPML isoforms. Confocal microscopy visualization of individual mPML isoforms expressed as fusion proteins with C-terminal tRFP tag (cyan) in Pml^+/+ or Pml^−/− MEFs. Endogenous mPML proteins (yellow) and SUMO-1 proteins (magenta) were immunostained using specific antibodies. Cell nuclei were visualized by DAPI staining (borders of the nuclei indicated by a dashed line in the merge). Importantly, the anti-mPML antibody also detects the transiently expressed mPML-tRFP isoforms. Scale bar, 5 μm.

Figure 3.

The effect of arsenic treatment on mPML NB dynamics. Pml^−/− MEFs transiently expressing individual mPML isoforms or control Pml^+/+ MEFs or 3T3 cells were mock-treated or treated with 5 µM ATO for 1 h. Subsequently, the cells were fixed and stained using specific antibodies against mPML (green) and SUMO-1 (magenta) proteins. Cell nuclei were visualized by DAPI staining (blue). Representative fluorescence microscopy images are shown. Scale bar, 10 μm.

Figure 4.

The effect of arsenic treatment on mPML protein levels in the cells. Pml^−/− MEFs transiently expressing individual mPML isoforms or control Pml^+/+ MEFs or 3T3 cells were mock-treated (C) or treated with 5 µM ATO for the indicated times. Whole-cell lysates were analyzed by western blotting using an anti-PML or anti-GAPDH antibody. Asteriks (*) indicates the predicted molecular mass (MM) of the isoforms. ATO/C quantification represents the ratio of PML levels in ATO-treated cells (normalized to GAPDH) relative to untreated controls.

Figure 5.

Exchange of individual mPML isoforms at endogenous PML NBs. (a) FRAP experiments were performed using mPML-tRFP constructs expressed in Pml^+/+ MEFs. Circular ROIs containing individual PML NBs were selected, and photobleaching was carried out by point-scanning following pre-bleach time-lapse imaging. Post-bleach fluorescence recovery was monitored over 10 min. Nuclear localization of PML NBs was confirmed via DIC microscopy. Representative isoforms are shown. Scale bar, 5 μm. (b) Representative isoform-averaged fluorescence recovery traces. Black circles indicate the mean across all recorded traces for the isoform, with the gray shaded area denoting ± one standard deviation. The red curve represents a single-exponential fit using the averaged parameters from individual fits. (c) Combined violin and swarm plot comparing recovery times across different isoforms.

Figure 6.

AlphaFold3 structural prediction for isoforms of mPML. Full structures are given individually and are overlayed in the central image, with low-confidence regions hidden (residues 1–49 and 403–570). Additional hidden residues are 766–885 for mPML2; 867–943 form mPMLX1; 821–897 for mPMLX2; and 239–408 for mPMLX7. Confidence scores per residue, based on predicted local distance difference test (pLDDT) values, are provided by AlphaFold and color-coded as described in the legend.

Figure 7.

MD simulation results of mPMLX7. (a) Coordination of the Zn²⁺ cation in the two AlphaFill-suggested positions Zn₍₁₎ (left) and Zn₍₂₎ (right). The structures were extracted from the last frame of their respective production runs. (b) Superposition of the post-MD (left) Zn-free (red), (middle) Zn₍₁₎ (blue), and (right) Zn₍₂₎ (orange) structures over the AlphaFold prediction (gray). Left: the disordered regions have been labeled on the AlphaFold prediction. Middle-right: the Zn²⁺ positions are labeled and shown with blue balls.