Mutating PARG reveals differential localization and binding of nucleolar proteins to pADPr

To further evaluate nucleoli structural integrity, we used Transmission Electron Microscopy (TEM) analysis to detect any nucleolar abnormalities associated with disrupting PARP1 activity. TEM revealed significant changes in nucleoli in both wild-type and Parg^27.1 mutant nuclei (Figure 3A and 3B). Wild-type nucleoli are typically located close to the middle part of nuclei and appear to be almost homogenic (Figure 3A). In contrast, nucleoli of Parg^27.1 mutants contain heavily condensed areas positioned close to nuclear lamina (Figure 3B). Such structural changes may impact nucleolar functions and affect the steady state equilibrium of ribosomal rRNA processing. Therefore, we tested localization of nucleolar proteins reported to participate in rRNA processing and maturation in wild-type and Parg^27.1 mutant nuclei using immunostaining.

Open in a separate window

Figure 3

Mutating PARG disrupts nucleoli and reveals differential localization of nucleolar proteins.

A–B. The structure of wild-type nucleolus detected by EM microscopy (A) is affected in Parg^27.1 mutants (B). In contrast to clear homogeneous (grey) content of wild-type nucleolus, Parg^27.1 mutant nucleoli accumulate “holes” and “aggregates” of proteins or chromatin (black). Arrows indicate the nucleolus. Arrowheads are pointing to the nuclear envelope. C. Nucleolar proteins colocalize with PARP1 (red) in wild-type nucleoli. Fibrillarin (green) protein is shown. D–G. Mutating PARG displaces PARP1 protein from chromatin to Cajal Bodies (arrowheads) and “traps” PARP1 within condensed nucleolar blocks (arrows). One class of nucleolar proteins completely co-localizes with PARP1 in CB and nucleoli (D–E). Another group of nucleolar proteins (F–G) behave independently from PARP1. This group of proteins has a homogeneous localization in nucleoli (arrows) and could barely be detected in CBs (arrowheads).

In addition to Fibrillarin, we evaluated colocalization using Nucleolin and dNop5, both of which have been shown to be involved in nucleolar rRNA processing and ribosomal maturation [34], [35]. Additionally, we used Casein Kinase II α (CKIIα as a nucleolar marker, since previous reports have shown that it participates in rDNA transcription by phosphorylating components of RNA polymerase I [36]. Our results from this analysis shows that, in wild-type nucleoli, all tested nucleolar proteins demonstrate similar homogenic staining and co-localization with PARP1 protein (Figure 3C, only Fibrillarin protein is included). In contrast, nucleolar proteins in Parg^27.1 nuclei accumulate in two distinct locations. This suggests that, their exclusive localization within the nucleolus requires the activity of a functional PARP1 protein.

Previously, we demonstrated that in Parg^27.1 mutants, most of PARP1 protein is automodified, and, together with all other poly(ADP-ribosyl)ated proteins and pADPr binding proteins, it is either targeted to Cajal Bodies (CBs) or arrested inside abnormal condensed nucleolar substructures [20], [32]. Based on this observation, we proposed that pADPr modification would lead to co-localization with automodified PARP1 in CBs. One group of proteins, Fibrillarin and Nucleolin, co-localized with PARP1 inside condensed blocks of the nucleolus and in CBs (Figure 3D and 3E). However, another set of proteins, including nucleolar markers CC01311, dNop5 and CKIIα were preferentially antagonistic to PARP1 localization and did not accumulate in CBs (Figure 3F and 3G; Figure S4). These observations suggested a direct interaction of the first group of proteins with automodified PARP1. This interaction likely affects the location of these proteins in the nucleolus and thus may impact nucleolar activity. In contrast, CC01311, dNop5 and CKIIα do not interact with automodified PARP1. As a result, their positioning in the nucleolus is determined by PARP1-independent mechanisms. The existence of a selected group of nucleolar proteins requiring PARP1 enzymatic activity for their localization suggested to us that PARP1 may function in the nucleolus by binding these proteins through pADPr attachment.

To further test our hypothesis that a specific sub-group of nucleolar proteins preferentially interacts with pADPr, we performed co-immunoprecipitation experiments (Co-IP) using anti-pADPr antibody. Indeed, we found that Fibrillarin, AJ1, Nucleolin, and Nucleophosmin co-precipitate with pADPr in wild-type animals and Parg^27.1 mutants (Figure 4A and 4B), while other nucleolar proteins, including CC01311, dNop5 and CKIIα, do not show interaction with pADPr in wild-type nuclei (Figure 4A). Interestingly, a slight interaction of CC01311 protein with pADPr was detected in Parg^27.1 mutant (Figure 4B), although this weak interaction is likely indirect and may be mediated by other nucleolar components.

Open in a separate window

Figure 4

Nucleolar proteins show differential interaction with pADPr.

Immunoprecipitation assays using mouse anti-pADPr antibody. Wild-type (A) and Parg^27.1 mutant (B) third-instar larvae were used. The following antibodies were used for Western blot analysis: rabbit anti-Fibrillarin; rabbit AJ1; rabbit anti-Nucleolin; rabbit anti-GFP (to detect nucleolar CC01311 marker); rabbit anti-Nucleophosmin; rabbit anti-dNop5; rabbit anti-CK2α; and mouse anti-Dlg (as a control).

This result suggests that by binding a specific set of proteins to pADPr, PARP1 may determine the order of steps that occur during the process of ribosome biogenesis. Therefore, we further tested if mutating PARP1 or PARG affects any specific steps involved in ribosome production.

Inhibition of PARP1 activity disrupts rRNA processing

Resident nucleolar proteins, such as Fibrillarin, have critical roles in rRNA processing, modification, and maturation [31], [37]–[39]. Our observation of fragmented localization of nucleolar-specific proteins, some of which are critical for rRNA maturation, prompted us to investigate the effect of inhibiting PARP1 activity on rRNA processing and maturation. The transcription of rDNA produces a precursor molecule which is subjected to multiple rounds of exonucleolytic and endonucleolytic cleavages by resident nucleolar protein complexes to generate mature 18S, 5.8S, and 28S rRNA final products [40], [41]. We reasoned that the displacement of nucleolar proteins involved in rRNA processing will have a disruptive effect on the generation of mature rRNA products. To evaluate our hypothesis, we isolated RNA from both wild-type and mutant larvae and analyzed precursor rRNA intermediates using Northern blot. Surprisingly, we found no quantifiable difference between mature rRNA amounts in wild-type, Parg^27.1, and Parp^C03256 mutants (Figure 5, 18S and 28S). Instead, a significant increase in rRNA intermediates was detected in Parg^27.1 and Parp^C03256 mutants (Figure 5, Figure S5).

Open in a separate window

Figure 5

Production of rRNA intermediates increases upon disruption of PARP1 or PARG activity.

A. Diagrammatic representation of mammalian rRNA transcript processing. The final 18S rRNA product becomes part of the 40S small ribosomal complex, while the 5.8S and 28S rRNA transcripts are incorporated into the large 60S ribosomal complex. Red bar above the scheme indicates internal transcriber spacer (ITS) probe, which was used to detect intermediates of rRNA processing on Northern blots. B. Northern blot analysis of rRNA intermediates. Disruption of PARG or PARP1 activity enhances the production of rRNA intermediates (right lanes) compared to the heterozygous (left lanes), which has normal PARP1 and PARG activity. The total level of mature rRNA does not increase (18S and 28S). Labeled probe to Drosophila Tubulin mRNA was used as a loading control.

When taken together with the observations reported above, this suggest that displacement of key components of ribosomal biogenesis, which occurs in Parg^27.1 and Parp^C03256 mutants, results in rRNA processing delays and over-accumulation of immature intermediates. In contrast, the initial 47S rRNA precursor product and the intermediate 36S product were effectively processed in Parg^27.1 and Parp^C03256 heterozygotes, and accumulation of these two precursor products was observed in larvae lacking PARP1 or PARG function (Figure 5, red asterisks). The last finding indicates that mutating pADPr pathway does not affect the rate of rDNA transcription, but blocks specific steps of precursor rRNA maturation and accumulation of immature ribosomes. To test this hypothesis, we then proceeded to compare ribosomal content of wild-type, Parp1, and Parg mutants.

Construction of transgenic Drosophila

To construct the anti-Parp RNAi transgene we cloned an 1839-bp fragment of Parp-e cDNA (from GM10715 clone) in direct and inverted orientation within the pUASt vector. As a spacer between inverted repeats we used a 720-bp fragment of EGFP sequence (Figure S3). Transformation was as described [57], with modifications [58].

Western blot

Protein extracts were separated on a 4–12% gel (Invitrogen), transferred onto a nitrocellulose membrane and detected using Amersham/GE Healthcare (#RPN2106) kit, according to manufacturer's instructions. The following primary antibodies were used: anti-Fibrillarin (rabbit, 14000, gift form Dr. J. Gall), anti-Nucleolin (rabbit, 14000), nucleolar AJ1 antibody (rabbit, 11000, gift form Dr. J. Gall), anti-Nucleophosmin (rabbit, 11000), anti-dNop5 (rabbit, 12000, gift form Dr. G. Vorbruggen), anti-GFP (rabbit, Torrey Pines Biolabs, #TP401, 11000), anti-CKIIα (rabbit, 1150, Stressgen, # KAP-ST010), anti-DLG (rabbit, 15000, gift from Dr, F. Roegiers) and anti-RPS6 (mouse mAb, 11000, Cell Signalling #2317).

Electron microscopy

Electron microscopy was performed essentially as described [20].

Immunoprecipitation

Immunoprecipitation experiments were performed as described [20], with little modifications. Briefly, 30 ul of Protein-G Sepharose 4B were added to the protein lysates and incubated overnight at 4°C with rotation. Beads were washed 4 times for 5 min in 1 ml of lysis buffer. Bound proteins were eluted using 60 ul of 1× Laemli with heating at 95°C for 5 min. The following antibodies were used for immunoprecipitation: anti-pADPr (Mouse mAb, H10 120, Tulip, #1020).

Northern blot

Total RNA was isolated using Trizol reagent (Gibco BRL), precipitated twice with 3 M LiCl, treated with Amplification Grade Dnase I (Gibco BRL). The following primers were used to produce ITS probe detecting intermediates of pre-rRNA: ITSf (5′-ataacaaaatgattccatgg-3′) and ITSr (5′-aaaaatacaccattttactgg-3′); for 18S rRNA probe: 18Sf (5′-aaaagtgaaaccgcaaaagg-3′) and 18Sr (5′-taatgatccttccccgcagg-3′). For Northern blot analysis, at least of 2.5 ug of total nuclear RNA from third instar larvae was used per lane. A Tubulin probe was used as a loading control.

Sucrose density gradient analysis

Analysis of ribosomes by sucrose density gradient centrifugation was carried out as described [6], with modifications. Briefly, 3^rd instar larvae were picked, washed 2× in distilled water, followed by 20 min incubation at room temperature in 200 ul of Buffer A (20 mM HEPES [pH 7.5], 10 mM KCl, 2.5 mM MgCl₂, 1 mM EGTA, 100 ug/ml cycloheximide, 1 mM DTT). 450 ul of Buffer A was then added, and larvae were then lysed on ice and centrifuged for 5 min at 10,000 rpm at 4°C. The absorbance of the supernatant was measured at 260 nm and 400 ug of each sample was carefully loaded onto a 10.5 ml 10–55% sucrose gradient in Buffer A without cycloheximide and DTT, centrifuged for 8 hrs at 27,000 rpm in a SW41 rotor. 1 ml fractions were collected using a Foxy Jr gradient collector (ISCO) with a UV detection system for recording profile. 25% of selected fractions were TCA-precipitated and analyzed by Western blot for the presence of S6 ribosomal protein (Cell Signaling). 380 ul of the remaining samples was used for RNA isolation. 30 ul of 10% SDS/Tris (pH 7.5) and 1 ml of combined phenolChloroformIsoamyl Alcohol solution (invitrogen cat # 15593-031) was added to each sample and heated at 65°C for 2 min. Samples were centrifuged and the supernatant transferred to a new tube. Extraction was repeated again using the phenolChloroformIsoamyl Alcohol solution, followed by transfer of the supernatant. 1/10 the volume of 3 M sodium acetate and 2× the volume of cold ethanol was added to each sample followed by incubation at −20°C for 20 min. Samples were centrifuged at 4°C for 4 min and the resulting pellet dissolved in appropriate volume of RNase free water to be used for qPCR analysis.

Quantitative Reverse Transcription–PCR

10 ug of RNA was used for cDNA synthesis using the high capacity cDNA Reverse Transcription Kit (applied biosystems cat # 4368814). The following primer sequences: Tubulin: Forward (ccttcgtccactggtacgtt), Reverse (ggcgtgacgcttagtactcc); GAPDH: Forward (cgacaagttcgtgaagctga), Reverse (attctaccgcgccctaatct); Act5C: Forward (gtgcccatctacgagggtta), Reverse (agggcaacatagcacagctt) were utilized with Power SYBER Green master mix (cat # 4367659) for PCR using the Applied Biosystems StepOnePlus detection system with the following cycling conditions: 95°C for 10 min, followed by 40 (2-step) cycles (95°C, 15 s; 60°C, 60 s).

Probes for In Situ Hybridization

rDNA probes were generated by digesting full length Drosophila rDNA gene (18S, 5.8S, and 28S) containing intergenic sequences with the following enzymes: HaeIII, AluI, MspI, RsaI, and MseI (New England Biolabs). Labeled probes were generated using DIG DNA Labeling Kit (Roche; Cat # 11 175 033 910) essentially as outlined by manufacturer.

Fluorescence In Situ Hybridization (FISH)

FISH protocol was carried out as described [59], with modifications. Salivary glands were dissected from 3^rd instar larvae and fixed in for 10 min at room temperature (RT) in 4% formaldehyde in buffer A (15 mM PIPES, 80 mM KCl, 20 mM NaCl, 2 mM EDTA, 0.5 mM EGTA, 0.5 mM spermidine, 0.15 mM spermine, 1 mM DTT) prewarmed to 37°C. After fixing, salivary glands were washed three times (5 min each) in 2× SSCT (SSC in 0.1% Tween) followed by successive incubation at RT for 10 min in 20% formamide in 2× SSCT, 40% formamide in 2× SSCT, 50% formamide in 2× SSCT. Salivary glands were again incubated in fresh 50% formamide in 2× SSCT for 30 min at 37°C. The solution was aspirated and salivary glands were incubated with DNA probes diluted in 40 ul of hybridization solution (50% formamide, 3× SSCT, 20% dextran sulfate, 0.25% Tween). Probes and salivary glands were denatured (91°C for 2 min) together in a thermal cycler. Hybridization was carried out at 37°C for 24 hrs. Following hybridization, salivary glands were washed three times for 20 min each in 50% formamide in 2× SSCT at 37°C, and a wash in 25% formamide in 2× SSCT at RT. Salivary glands were washed again three times in 2× SSCT and 1× PBS/Tween at RT for 5 min each. Following washes, immunostaining protocol for Fibrillarin (outlined below) was carried out, starting by blocking with 10% BSA solution. The DIG labeled probes and Fibrillarin were detected by staining with rhodamine-conjugated anti-digoxigenin F(ab) fragments (Boehringer Mannheim) and Goat Anti-Mouse Alexa 488 diluted in 1× PBS/Tween at RT for 2 hrs. Salivary glands were washed with 1× PBS/Tween at RT twice, 20 min each time followed by incubation with DNA binding dye, Drag5, for 1 hr at RT.

We thank Drs. J. Gall, A. Veraksa, F. Roegiers, and G. Vorbruggen for providing materials. Alana O'Reilly, Ekaterina Pechenkina, and David Martin contributed valuable comments on the manuscript. We also thank members of the FCCC Fly Working Group for fruitful discussions on the paper.