Poly(ADP-ribose) Has a Branched Structure in Vivo*

We have searched for the presence of branching in the chromosomal polymer poly(ADP-ribose) as it occurs in vivo. Treatment of the polymer with phospho- diesterase and phosphomonoesterase results in the conversion of internal residues to the nucleoside ribo- syladenosine and the conversion of points of branching to diribosyladenosine. We have detected diribosylad- enosine in digests of the polymer derived from carcin-ogen-treated SV40 virus-formed 3T3 cells and in nor- mal rat liver, kidney, and spleen. The frequency of residues involved in branching varied from 0.8 to 1.6 mole 8 over a 50-fold range of total levels of poly(ADP-ribose). Thus, branching seems to be a general feature of poly(ADP-ribose) as it occurs in vivo. repair

800 x g for 10 min, the supernatant was extracted 5 times with an equal volume of diethylether. The resulting solution containing diribosyladenosine was then adjusted to 200 mM sodium citrate buffer, pH 4.5, 20 m~ chloracetaldehyde. The fluorescent derivative I,N6ethenodiribosyladenosine (E-diribosyladenosine) was obtained by incubating this mixture at 60 "C for 8 h. Excess chloracetaldehyde was removed by extraction 5 times with an equal volume of diethylether. This preparation was used as the e-diribosyladenosine standard.
Analysis of SVT2 Cells"SVT2 cells were grown as described previously (16). Ten tissue culture dishes (55 cm2) containing 3.5 x 10' cells/dish were treated with 50 pg/ml MNNG' for 20 min. Medium was removed and cells were rapidly washed on the dish with 10 ml of cold phosphate-buffered physiological saline followed by the addition of 5 ml of ice-cold 20% trichloroacetic acid. Dishes were kept on ice for 15 min and the resulting precipitate was collected by centrifugation at 800 X g for 10 min and washed twice with 20% trichloroacetic acid and once with diethylether. The pellet was dissolved'in 50 ml of 0.1 M potassium phosphate buffer, pH 5.0, containing 5 M guanidine-HC1. This solution was divided into 5 aliquots of 10 ml each containing an amount of material equivalent to 7.0 X 10' cells. All 5 aliquots were subjected to our procedure for poly(ADP-ribose) determination (17) with modifications indicated below. One aliquot was spiked with 50,000 cpm of ['4C]poly(ADP-ribose) synthesized in permeabilized cells to determine recovery. Recovery yields were approximately 65%. Adenosine deaminase treatment was accomplished as follows. The dihydroxyboryl column eluate (2 m l ) was concentrated in uucuo to 100 pl of 100 m~ Tris-HC1, pH 7.4, and 138 units of adenosine deaminase from calf intestine (Type I, Sigma). After incubation, protein was precipitated by adjusting to 20% trichloroacetic acid. After centrifugation, the supernatant was extracted 5 times with an equal volume of diethylether. The sample was adjusted to 1 ml and made 200 mM in sodium citrate buffer, pH 4.5, incubated with chloroacetaldehyde, and prepared for chromatography as before (16). High pressure liquid chromatography was performed with a Beckman llOA liquid chromatograph equipped with an Altex Ultrasphere-ODS reversed phase column (250 mm X 4.6 mm inner diameter X %-inch outer diameter). Fluorescence was detected with a Varian Fluorichrom Filter Fluorometer equipped with a deuterium light source. Excitation was performed using a 220-1 Varian interference fdter (220 nm band pass) and a Varian 3-75 fiiter (370 nm cutoff) was used for emission. The sensitivity of the fluorometer was set at 20 mV full scale at the time of injection and increased to 2 mV full scale approximately 5 min after elution of E-ribosyladenosine ( Fig. 2, B , C, D, and E ) . This sensitivity change allowed the detection of €-&bosyladenosine. Sample injection volume was 2 ml in all cases in 200 mM sodium citrate buffer, pH 4.5. Elution was performed isocratically at room temperature with 7 m~ ammonium formate buffer, pH 5.8/ 100% methanol, 973 (v/v). The flow rate was 1.4 ml/min.
Analysis of Rat Tissues-For determination of rat tissues, 3 adult male Sprague-Dawley rats were killed by decapitation; liver, kidneys, and spleen were quickly removed, and each tissue was blended in icecold 20% (w/v) trichloroacetic acid using a Waring Blendor. The mixture was kept cold during blending by adapting an ice water bath to the blending chamber. The resulting suspension was centrifuged at 800 X g for 20 min and the precipitate was washed twice with 20% trichloroacetic acid and once with diethylether. The pellet was subsequently dissolved in 100 ml of 0.1 M potassium phosphate buffer, pH 8.6, containing 6 M guanidine-HCl. Each sample was then divided in two aliquots of 50 ml each. One of the aliquots in each group was spiked with 50,000 cpm of ["C]poly(ADP-ribose) to determine recovery. Recovery values varied between 60 and 70% and the values shown are corrected for recovery. The remainder of the procedure for Eribosyladenosine and 6-diribosyladenosine determination was as described before except that it was scaled up by a factor of 10.

Levels ofpoly(ADP-ribose) in SVT2 cells and rat tissues
Poly(ADP-ribose) was measured as described under "Experimental Procedures." The values shown represent the mean of duplicate determinations that aereed within 10% of the mean. experiments. It was important to show that the presence of peaks 4 and 5 was not due to the presence of endogenous fluorescent compounds. As shown in Fig. 2C, omission of chloracetaldehyde treatment resulted in the absence of all peaks, including peaks 4 and 5. The next step was to demonstrate that peaks 4 and 5 were indeed a product of a polymer sensitive to treatment with snake venom phosphodiesterase.

Fig. 2 D shows that when venom phosphodiesterase treatment
was omitted, peaks 4 and 5 were absent. Finally, peaks 4 and 5 were further characterized as adenine-containing compounds by treatment with adenosine deaminase just prior to treatment with chloracetaldehyde. Formation of an etheno derivative of the adenine ring requires the presence of the N6amino group and treatment with adenosine deaminase converts both ribosyladenosine and diribosyladenosine to the respective inosine derivatives (17).2 Fig. 2E shows that adenosine deaminase treatment did eliminate both peaks 4 and 5.
From these experiments, we conclude that peaks 4 and 5 from MNNG-treated SVT2 cells are E-ribosyladenosine and E-diribosyladenosine, respectively. To our knowledge, this represents the first demonstration that poly(ADP-ribose) has a branched structure in vivo. Fig. 1 were pretreated with the DNA damaging agent MNNG which causes a large increase in the intracellular levels of poly(ADP-ribose) (3). To determine if a branched structure exists in normal untreated tissue, we have analyzed different tissues of the rat. The values obtained from MNNG-treated SVT2 cells and adult rat tissues are compared in Table I. The presence of diribosyladenosine was detected in each of the tissues even though the total intracellular levels of poly(ADP-ribose) in the rat tissues were only 2 to 4% those of the cells treated with MNNG. It is also interesting that despite the great difference in absolute content of poly(ADPribose) the relative amount of diribosyladenosine was very similar, ranging from 0.8 to 1.6% of the total residues. The frequency of branching of the polymer in vivo is somewhat lower than the value of 3 to 5% reported from the in vitro system (14). We have also examined several tissues from H. Juarez-Salinas, V. Levi, E. L. Jacobson, and M. K. Jacobson, submitted for publication. turkey and chicken for the presence of diribosyladenosine and have obtained values in the same range as those reported here for rat tissue^.^ The picture that is emerging suggests a highly conserved poly(ADP-ribose) structure regardless of the type of animal tissue examined or the physiological conditions prevailing in the cell at the time of synthesis.

The cells analyzed in
The presence of branching in poly(ADP-ribose) raises a number of interesting questions related to the structure of this polymer. We do not know the percentage of chains that are branched or the location of branch points within a chain. Likewise, it will be of interest to determine whether the branch points are sites of branching of a single residue or are sites for the synthesis of long branches of polymer. The branching of poly(ADP-ribose) may play a role in the regulation of synthesis and/or degradation of the polymer. Alternatively, branching may be required to effect alterations in chromatin structure associated with one or more of the postulated functions of this polymer.