Differential Localisation of PARP-1 N-Terminal Fragment in PARP-1^+/+ and PARP-1^−/− Murine Cells

Abstract

INTRODUCTION

PARP-1 (EC 2.4.2.30) is a nuclear enzyme found in all eukaryotes except yeast with approximately 10⁶ molecules per cell. It is a prominent member of a family of 17 PARPs in humans. On limited proteolysis of human PARP-1 using papain and α-chymotrypsin three functional domains were identified: 46 kDa amino terminal DNA binding domain (DBD), 22 kDa central automodification domain and 54 kDa carboxy-terminal NAD⁺ binding domain (Kameshita et al., 1984). DBD encompasses two Cys-Cys-His-Cys zinc finger motifs (FI and FII) that are involved in DNA binding and a nuclear localisation signal (NLS) (de Murcia et al., 1994; Gradwohl et al., 1990; Lamarre et al., 1986; Molinete et al., 1993). PARP-1 has been implicated in a number of functions such as DNA replication and repair, cell death, maintenance of genome integrity, inflammation and cell proliferation (Barnes and Lindahl 2004; Bjelland and Seeberg, 2003; Caldecott, 2008; Dizdaroglu et al., 2002; Svilar et al., 2011).

The primary function of a cell is to preserve genomic stability when faced with challenges that cause DNA damage. Failure to repair oxidative DNA lesions can lead to cancer and accumulation of damaged DNA can contribute to ageing and neurodegeneration (Caldecott, 2008; Maynard et al., 2009). PARP-1 binds to sites of DNA break and using NAD⁺ it catalyzes the poly-(ADP-ribosyl)ation of PARP-1 itself as well as histones, nuclear proteins, DNA repair proteins, transcription factors and chromatin modulators (Burkle, 2006; Luo and Kraus, 2012; Ogata et al., 1981; Schreiber et al., 2006). The mechanism of DNA damage induced polymer synthesis by PARP-1 has been elucidated. It engages with DNA as a monomer leading to distortion and destabilisation of its catalytic domain that in turn facilitates its automodification (Langelier et al., 2012). Other studies show that the 40 kDa carboxy-terminal domain forms homo and hetero dimers in the absence of DNA (Mendoza-Alvarez and Alvarez-Gonzalez, 2004).

Excessive activation of PARP-1 can lead to exhaustion of NAD+ as observed in conditions including reperfusion injury, inflammation and heart attack leading to apoptosis or necrosis. Furthermore, during apoptosis PARP-1 undergoes proteolytic cleavage and generates a 24 kDa N-terminal fragment containing DBD and 89 kDa fragment containing both the automodification and catalytic domains. These fragments have been shown to persist in apoptotic cells several days after the onset of cell death (Lazebnik et al., 1994; Nicholson et al., 1995; Rosenthal et al., 1997; Soldani and Scovassi, 2002; Tewari et al., 1995). Deficiency or inactivation of PARP-1 can cause telomere shortening and reduced PARP-1 activity was observed in ageing human samples (Chevanne et al., 2007; Muiras et al., 1998; Serrano and Blasco, 2007; Tong et al., 2001). Hence for cell survival, maintenance of genomic integrity and longevity presence of intact PARP-1 in the nucleus is paramount. The objective of this study was to investigate the intracellular trafficking role of full-length PARP-1 by examining the subcellular distribution of exogenous PARP-1 N-terminal fragment in PARP-1 wild type and knock-out cells. To visualise the fragment by noninvasive fluorescence microscopy we generated a DNA construct of PARP-1 N-terminal fragment (750 bp) tagged to pEGFPN1 and introduced into PARP-1^+/+ and PARP-1^−/− mouse embryo fibroblasts (MEFs).

MATERIALS AND METHODS

RESULTS

Although localisation of full-length and truncated PARP-1 in PARP-1 proficient cells has been well documented, subcellular distribution of PARP-1 fragments in the absence of endogenous PARP-1 is not known. Therefore, to examine the localisation and pattern of distribution of exogenous PARP-1 N-terminal fragment in PARP-1 deficient cells, PARP-1^+/+ and PARP-1^−/− mouse embryo fibroblasts were transiently transfected with PARP-1 N-term-EGFP plasmid and cells expressing the tagged protein were examined live using CLSM. At least one hundred fields of cells were analysed and 57% were fluorescent, 43% non-fluorescent and within 100% of the fluorescent cells the tagged protein was present exclusively in the nucleus. EGFP was expressed within the entire cell (Figs. 1A and 1B). Imaging at high magnification revealed a granular pattern of distribution of the tagged protein (Fig. 2A). Interestingly, optical z sections showed heterogenous distribution through the depth of the cell (Fig. 2B). The intensity of fluorescence of the sections was quantified using ImageJ. MFI of outer sections ranged from 0–10, inner sections from 20–60 and the mid-section gave a peak value of 88.7 indicating dense packaging of the tagged protein in the middle of the nucleus (Fig. 2C).

In contrast to the above data, PARP-1 N-terminal fragment was present exclusively in the cytoplasm of PARP-1^−/− MEFs (Fig. 3A). EGFP was localised in the entire cell. At least 100 fields of cells were examined and 62% were fluorescent, 38% non-fluorescent and 100% of the fluorescent cells displayed fluorescence exclusively in the cytoplasm (Fig. 3B). RGB images obtained at high magnification as well as the processed contrast enhanced 8-bit gray scale images confirmed the localisation of the fragment in the cytoplasm. Furthermore, the images revealed a granular pattern of distribution similar to PARP-1^+/+ cells (Fig. 4A). Fluorescence intensity plot along the line selection across the cytoplasm and nucleus as depicted in Fig. 4A, Panel 1 was obtained using ImageJ. According to the plot profile, average gray values were 98.6% higher in the cytoplasm compared to the nucleus (Fig. 4B). Thus, both imaging and analytical data provided conclusive evidence for the cytoplasmic distribution of the N-terminal fragment in PARP-1 deficient cells. Optical z sectioning of the cells was not possible due to rapid photobleaching.

DISCUSSION

In addition to DNA damage response and cell survival, members of PARP family play a role in many functions including cell division and regulation of chromatin structure and transcription (reviewed in Krishnakumar and Kraus, 2010; Lord and Ashworth, 2012). Both full-length PARP-1 and truncated DNA binding domain are distributed in the nuclei of PARP-1 proficient cells. PARP-1-DsRed, PARP-1-EYFP and PARP-1-24 kDa-DsRed plasmids are distributed in the nucleoli and nucleoplasm of live COS-7 and HeLa S3 cells (Yung et al., 2004). Early studies have also reported extranuclear PARP activity in the ribosomal and mitochondrial fractions (Kun et al., 1975; Roberts et al., 1975). Our findings of PARP-1 N-terminal fragment in the nucleus of wild type MEFs exhibiting a granular pattern of distribution with patchy areas of strong intensity is consistent with previous reports of PARP-1 in HeLa cells (Fig. 2) (Mosgoeller et al., 1996). Furthermore, in wild type MEFs the fragment was more densely packaged in the mid-sections of the nucleus (Figs. 2B and 2C). We have reported a similar pattern of heterogenous distribution through the depth of the nucleus in wild type IRK cells (rat embryo fibroblasts) expressing the N-terminal fragment (Rajiah, 2013).

In contrast to wild type cells the fragment localised exclusively to the cytoplasm in PARP-1 knock-out cells and presented a granular pattern of distribution (Figs. 3 and 4). Recently full-length PARP-1 plasmid was expressed in PARP-1^−/− MEFs and localisation was observed in the nucleus by immunocytochemical method (Patel et al., 2012). This PARP-1 has in addition to NLS, a central automodification and C-terminal catalytic domains which may have facilitated proper folding of the molecule required for nuclear entry. In PARP-1^+/+ MEFs used in our study, these domains would have been provided by endogenous PARP-1 to enable nuclear entry of the N-terminal fragment whereas in PARP-1^−/− MEFs this facilitation was not available. Recent findings highlight a novel role for PARP-1 in the regulation of intracellular trafficking of antiapoptotic p53 and the key immune response protein, nuclear factor kappa B (NF-κB). PARP-1 inhibition by gene knock-out, knock-down or by chemical inhibitors in smooth muscle cells prevents NF-κB translocation into the nucleus (Zerfaoui et al., 2010). Following the discovery of post-translational modification of NF-κB by PARP-1 and its nuclear translocation during inflammation, the mechanism of NF-κB nuclear retention promoted by PARP-1 leading to modification of its gene regulatory effects has been proposed. In brief, despite the presence of NLS in NF-κB, it requires PARP-1 for nuclear translocation.

Interestingly, cytoplasmic molecules lacking NLS have also been localised in the nucleus. PTEN (phosphatase and tensin homolog) and Ferritin are traditionally cytoplasmic proteins and lack a consensus NLS. However, accumulating evidences show that PTEN and Ferritin are transported into the nucleus and Ferritin is able to bind and protect nuclear DNA from oxidative damage (Alkhateeb and Connor, 2010; Cai et al., 1998; Lian and Di Cristofano, 2005; Thompson et al., 2002). While researchers suggest the existence of a specific nuclear chaperone, what regulates Ferritin entry into the nucleus remains unclear. In conclusion, our GFP tagged PARP-1 N-terminal fragment has allowed for non-invasive visualisation of its subcellular distribution in live cells. Our findings provide confirmatory evidence for the requirement of endogenous PARP-1 for nuclear transport of the fragment. This pilot study highlights an intracellular trafficking role for PARP-1 independently of its enzymatic activity and future work will focus on investigating the requirement of intact PARP-1 for nuclear translocation of its siblings and other molecules.

Article information

References