Cyclophilin A Isomerisation of Septin 2 Mediates Abscission during Cytokinesis

Abstract

The isomerase activity of Cyclophilin A is important for midbody abscission during cell division, however, to date, midbody substrates remain unknown. In this study, we report that the GTP-binding protein Septin 2 interacts with Cyclophilin A. We highlight a dynamic series of Septin 2 phenotypes at the midbody, previously undescribed in human cells. Furthermore, Cyclophilin A depletion or loss of isomerase activity is sufficient to induce phenotypic Septin 2 defects at the midbody. Structural and molecular analysis reveals that Septin 2 proline 259 is important for interaction with Cyclophilin A. Moreover, an isomerisation-deficient EGFP-Septin 2 proline 259 mutant displays defective midbody localisation and undergoes impaired abscission, which is consistent with data from cells with loss of Cyclophilin A expression or activity. Collectively, these data reveal Septin 2 as a novel interacting partner and isomerase substrate of Cyclophilin A at the midbody that is required for abscission during cytokinesis in cancer cells.

Keywords: cancer; cell division; cyclophilin; cytokinesis; prolyl isomerase.

Figure 3

(A) Schematic representing the full-length sequence of Sept2, displaying the GTPase domain (yellow box) and the four putative CypA isomerisation sites at Pro24, Pro69, Pro162 and Pro259. PBR: phosphoinosite-binding polybasic region, G1–4: guanine nucleotide-binding motifs, SUD: septin-unique domain [48]. (B) Molecular modelling of Sept2 using Molegro Molecular Viewer (http://molexus.io/molegro-molecular-viewer/,) (accessed on 13 November, 2019) with the trimer Sept2-Sept6-Sept7 crystal structure determined by Sirajuddin, M (PDB ID: 2QAG) [49]. Pro24 is part of an exposed α-helix (i); Pro69 is located within an external loop between two α-helices (ii); Pro162 is located within an α-helix (iii) and Pro259 is located within a loop (iv). (C) K562 cells were transfected with pEGFP-Sept2^WT, pEGFP-Sept2^P24A, pEGFP-Sept2^P69A, pEGFP-Sept2^P162A or pEGFP-Sept2^P259A and enriched in mitosis by nocodazole treatment (160 nM) for 16 h followed by release into complete media for 60 min prior to cytocentrifugation and immunostaining. All cells were immunostained with anti-α-tubulin primary antibody, followed by incubation with AlexaFluor 594 secondary antibody. All cells were counter-stained with DAPI and visualised using a 60× oil objective lens (NA 1.4) on an Olympus Fluoview FV1000 confocal microscope. Images are representative of the percentage of 50 cells per condition over 3 independent experiments. Scale bar: 5 μm. (D) Bar chart highlighting the percentage (%) of each midbody phenotype observed within the five cell populations. Differences observed between the flanking and central midbody phenotypes were compared, and p-values were derived from one-sample, two-tailed Z-tests in comparison to EGFP-Sept2^WT cells. Asterisks are indicative of statistical significance (* = p-value < 0.05).

Figure 1

CypA interacts with Sept2, and loss of CypA expression significantly delays Sept2 midbody flanking-to-central transition during cytokinesis. (A) Table summarising the results obtained for Sept2 by label-free quantification (LFQ) MaxQuant analysis of the CypA interactome. The coverage of the protein is 8.6%. ‘CONT’ refers to the values obtained for the EGFP immunoprecipitation control, while ‘CypA’ refers to the values obtained for the EGFP-CypA^WT immunoprecipitation. Fold change was calculated by dividing the average ‘CONT’ values by the average ‘CypA’ values. p-values were derived by t-tests using values obtained for three experimental replicates and two technical replicates. K562 cells were transfected with pEGFP and pEGFP-CypA^WT (B) and lysed 48 h post-transfection, and exogenous EGFP-tagged proteins were immunoprecipitated using GFP-Trap beads and resolved by SDS-PAGE. Resolved proteins were analysed by Western blot using an anti-Sept2 primary antibody. Whole-cell extracts (WCE) and post-binds (P-B) collected after immunoprecipitation were probed with anti-GFP to confirm the presence and absence of exogenous EGFP. Jurkat^CypA+/+ (C) and Jurkat^CypA−/− (D) cells were enriched in mitosis by treatment with nocodazole (160 nM) for 16 h before being released into complete media until they reached the midbody stage. Cells were collected by cytocentrifugation and immunostained with anti-Sept2 and anti-α-tubulin primary antibodies followed by incubation with AlexaFluor 488 and 594 secondary antibodies, respectively. Cells were counterstained with DAPI and visualised using a 60× oil objective lens (NA 1.4) on an Olympus Fluoview FV1000 confocal microscope. The images are representative of n = 150 cells over 3 independent experiments. White arrows indicate the location of the midbody in each image and the location of each region magnified in the left-most panels of (C,D). Scale bar: 10 µm. (E) Bar graph representing the percentage of each Sept2 phenotype observed at the midbody in Jurkat^CypA+/+ and Jurkat^CypA−/− cells following midbody enrichment using nocodazole treatment. The graph represents the percentage of 150 cells analysed per condition. Differences observed between the flanking and central midbody phenotypes were compared, and p-values were derived from one-sample, two-tailed Z-tests. Asterisks are indicative of statistical significance (** = p-value < 0.01, *** = p-value < 0.001).

Figure 2

(A) Chymotrypsin cleavage of the substrate N-succinyl-Ala-Ala-Pro-Phe-pNitroanilide was quantified in the absence of CypA (‘No CypA’), or presence of His-CypA^WT and His-CypA^R55A, with the addition of CsA (60 nM) to each reaction in (B) only. The absorbance of pNitroaniline was measured at 390 nm. The results shown are mean values obtained for each time point across three independent experiments. (C,D) Jurkat^CypA+/+ cells were enriched in mitosis following treatment with nocodazole (160 nM) for 16 h before being released into media supplemented with CsA (10 µM) for 60 min prior to collection by cytocentrifugation and immunostaining. All cells were immunostained with anti-Sept2 and anti-α-tubulin primary antibodies, followed by incubation with AlexaFluor 488 and 594 secondary antibodies, respectively. Cells were counterstained with DAPI and visualised using a 60× oil objective lens (NA 1.4) on an Olympus Fluoview FV1000 confocal microscope. The images are representative of n = 150 cells over 3 independent experiments. White arrows indicate the location of the midbody in each image and the location of each region magnified in the left-most panels of (C,D). Scale bar: 10 µm. (E) Bar graph representing quantification of Sept2 dynamics at the midbody with hourglass, dynamic, flanking and central phenotypes indicated in Jurkat^CypA+/+ or Jurkat^CypA−/− cells in the absence or presence of CsA (10 μM), as outlined. The graph represents the percentage of 150 cells analysed per condition. Differences observed between the flanking and central midbody phenotypes were compared, and p-values were derived from one-sample, two-tailed Z-tests in comparison to untreated Jurkat^CypA+/+ cells. p = <0.00001–1 indicates the range of p-values calculated relative to CypA^+/+.. Asterisks are indicative of statistical significance (** = p-value < 0.01, *** = p-value < 0.001).

Figure 4

Sept2-Pro259 is critical for the interaction between Sept2 and CypA. (A) Cartoon representation of the modelling performed for the interaction between CypA (green) and Sept2 (brown). (B) Zoomed in view of the active site of CypA (green) with regards to Pro259 of Sep2 (brown). K562 cells were transfected with pEGFP, pEGFP-Sept2^WT (C) or pEGFP-Sept2^P259A (D); lysed 48 h post-transfection and exogenous EGFP-tagged proteins were immunoprecipitated using GFP-Trap^® beads and resolved by SDS-PAGE. Resolved proteins were analysed by Western blot using anti-CypA (C,D) primary antibodies. Whole-cell extracts (WCE) and post-binds (P-B) collected after immunoprecipitation in (C,D) were probed with anti-GFP to confirm the presence and absence of exogenous EGFP.

Figure 5

Expression of EGFP-Sept2^P259A in HeLa cells impairs abscission. (A,B) HeLa cells were transfected with pmCherry-α-tubulin alone (A) or with pmCherry-α-tubulin and pEGFP-Sept2^P259A (B) before mitotic enrichment by nocodazole treatment (160 nM) for 16 h. Cells were released from nocodazole and imaged immediately using a Nikon Eclipse Ti-E microscope equipped with a temperature controlled, humidified chamber (37 °C, 5% CO₂) and an 100× oil-immersion objective lens (1.4) coupled with intermediate magnification (1.5×). Z-stacks (0.3 μm slices, 20 μm in total) were acquired from metaphase/late anaphase onwards, every 60 s to 80 s for up to 150 min post-telophase. The white arrow indicates the location of the ingressed midzone and subsequent midbody within the intercellular bridge. (i) to (v) indicate the stages of Sept2 dynamics depicted during cytokinetic progression. Scale bar: 15 μm. (C) Quantification of time from telophase to abscission of cells expressing Sept2^P259A relative to cells expressing endogenous Sept2 (date represents mean time from 3 independent live cell recordings for each condition). Differences observed in time to abscission were compared, and p-values were derived from a two-tailed, two-sample t-test assuming unequal variance. Asterisks are indicative of statistical significance (* = p-value < 0.05).