Purification and cDNA Cloning of Maize Poly(ADP)-Ribose Polymerase

Abstract

Poly(ADP)-ribose polymerase (PADPRP) has been purified to apparent homogeneity from suspension cultures of the maize (Zea mays) callus line. The purified enzyme is a single polypeptide of approximately 115 kD, which appears to dimerize through an S-S linkage. The catalytic properties of the maize enzyme are very similar to those of its animal counterpart. The amino acid sequences of three tryptic peptides were obtained by microsequencing. Antibodies raised against peptides from maize PADPRP cross-reacted specifically with the maize enzyme but not with the enzyme from human cells, and vice versa. We have also characterized a 3.45-kb expressed-sequence-tag clone that contains a full-length cDNA for maize PADPRP. An open reading frame of 2943 bp within this clone encodes a protein of 980 amino acids. The deduced amino acid sequence of the maize PADPRP shows 40% to 42% identity and about 50% similarity to the known vertebrate PADPRP sequences. All important features of the modular structure of the PADPRP molecule, such as two zinc fingers, a putative nuclear localization signal, the automodification domain, and the NAD⁺-binding domain, are conserved in the maize enzyme. Northern-blot analysis indicated that the cDNA probe hybridizes to a message of about 4 kb.

PADPRP (EC 2.4.2.30) is an enzyme that transfers the ADP-ribosyl moiety from NAD⁺ to a protein acceptor substrate (Chambon et al., 1963). This enzyme is present in all higher eukaryotes studied so far (Ueda and Hayaishi, 1985). Recent studies also report the presence of PADPRP in lower eukaryotes such as Dictyostelium discoideum (Kofler et al., 1993). However, PADPRP activity has not been detected either in yeast (Collinge and Althaus, 1994) or in prokaryotes (Ueda and Hayaishi, 1985; Collinge and Althaus, 1994). Extensive research has been carried out on the mammalian enzyme (for review, see Ueda and Hayaishi, 1985; Lindahl et al., 1995; Shah et al., 1995). PADPRP is a nuclear protein that is tightly associated with chromatin (Mosgoeller et al., 1996). Most of the acceptor substrates of PADPRP are also nuclear proteins (Althaus et al., 1995; Mosgoeller et al., 1996). PADPRP itself is a substrate for poly(ADP)-ribosylation (Ueda and Hayaishi, 1985). Initially, the enzyme synthesizes an ester linkage preferentially between the glutamyl (γ) or sometimes the C-terminal (α) carboxyl group on the acceptor protein and the 1′-OH of the ribosyl group of ADP-Rib. Subsequently, up to 45 to 50 ADP units are added via a 2"-1" phosphodiester bond. Branching of the poly(ADP)-ribosyl chains via the 2"-1‴ phosphodiester linkages has also been observed (Ueda and Hayaishi, 1985).

Another important property of PADPRP is its very high affinity for naked single- or double-stranded DNA. Furthermore, the enzyme is rapidly activated by DNA ends (Ueda and Hayaishi, 1985; Shah et al., 1995). In fact, any perturbation in the cellular morphology and/or physiology leading to DNA-strand breakage results in a rapid and robust increase in PADPRP activity (Auer et al., 1995; Kupper et al., 1995; Rosenthal et al., 1995). Consequently, many nuclear proteins are ADP-ribosylated (Althaus et al., 1995). This, in turn, is believed to affect such fundamental cellular processes as DNA replication (Eki, 1994; Yoshida et al., 1994; Simbulan-Rosenthal et al., 1996), DNA repair (Malanga and Althaus, 1994; Smulson et al., 1994; Shall, 1994), and DNA recombination (Auer et al., 1995; Kupper et al., 1995; Rosenthal et al., 1995; Waldman et al., 1996). The molecular mechanism(s) by which PADPRP modulates these basic biological events remains unelucidated despite the extensive studies of this enzyme in the animal kingdom (Auer et al., 1995; Kupper et al., 1995; Rosenthal et al., 1995; Kleczkowska and Althaus, 1996). We reasoned that detailed characterization of PADPRP and its cognate gene(s) from plants could shed additional light on the possible role of PADPRP in plant growth, metabolism, and defense. However, only limited information is available on plant PADPRP (Whitby et al., 1979; Bocher and Szopa, 1982; Wielgat et al., 1987; Chen et al., 1994). Therefore, we have isolated, purified, and characterized PADPRP from suspension cultures of a maize (Zea mays) callus cell line. Microsequencing of the purified enzyme has been carried out to obtain amino acid sequences of three peptides. Polyclonal antibodies generated against one of these peptides specifically cross-reacted with the plant enzyme but not with the human PADPRP. We have isolated and characterized a maize cDNA clone that encodes the full-length PADPRP. Although there are some differences in the cDNA and the derived amino acid sequence of the maize PADPRP compared with other PADPRPs, there are regions with a high degree of homology to the known PADPRP sequences, especially in the C-terminal region of the protein. The N-terminal region contains two zinc fingers, a signature sequence present in all known PADPRP molecules.

MATERIALS AND METHODS

Chemicals and Reagents

All chemicals used in this study were of molecular biology grade. Tris, Hepes, EDTA, MgCl₂, β-ME, and urea were procured from Sigma. Analytical-grade glycerol was obtained from Baxter (West Chester, PA). PefablocSC, DTT, pepstatin, bestatin, all restriction enzymes, DNA and RNA purification kits, and molecular weight markers were purchased from Boehringer Mannheim. Immunodetection kits for western blots were from Amersham. All reagents used in SDS-PAGE experiments, including kits for western blots, silver staining, and colloidal Coomassie blue staining, were from Novex (San Diego, CA). All radioactive chemicals were purchased from NEN-DuPont. Chromatographic resins were purchased from Sigma, Bio-Rad, and Pharmacia. Benzamide, 3-aminobenzamide, 3-methoxybenzamide, thymidine, nicotinamide, and novobiocin were purchased from Sigma. Isoquinoline diol and aminonaphthalimide were obtained from Aldrich. Rabbit polyclonal antibody 419 (raised against an epitope of the automodification domain of bovine PADPRP) and mouse monoclonal antibody C-2-10 (raised against bovine PADPRP) were kindly provided by Dr. Guy Poirier (Laval University, Quebec, Canada). Polyclonal anti-human PADPRP antibodies were purchased from Bio-Rad.

Cell Culture

The enzyme was isolated from a Hi II embryogenic callus cell line, 612B4, developed by Dave Peterson and Grace St. Clair (Pioneer Hi-Bred International, Johnston, IA). Exponentially growing cultures of 612B4 cells were maintained in the dark at 28°C in Murashige and Skoog medium supplemented with 2,4-D (2.5 mg/L). Cells were grown for 1 week on a gyratory shaker at 150 rpm and harvested by decantation. Routinely, 80 to 100 g of cells were obtained from 800- to 900-mL cultures grown in 12 to 14 flasks.

Enzyme Isolation and Purification

Cells were harvested by filtration and used to prepare WCE from 612B4 cells using the BioNebulizer (Glas-Col, Terre Haute, IN). Procedures for preparation of WCE have been described previously (Manley et al., 1983; Shapiro et al., 1988; Mahajan, 1993). All operations were carried out at 4°C or on ice unless indicated otherwise.

Chromatography on Heparin-Agarose

About 300 mL of heparin-agarose (Sigma) was washed extensively with HGED buffer (20 mm Hepes-KOH, pH 7.9, 20% glycerol, 0.1 mm EDTA), packed into a 5.0 × 30-cm Econo column (Bio-Rad) and connected to the Econo System (Bio-Rad). The matrix was equilibrated with HGED buffer plus 100 mm KCl. Three batches of crude WCE (approximately 1.5–1.8 g of total pooled protein in 60–80 mL) were loaded onto the column at a flow rate of 25 to 30 mL/h. The column was washed extensively with equilibration buffer until the A₂₈₀ of the effluent was less than 0.1 unit (approximately 900 mL). Small aliquots of peak fractions were saved for PADPRP assays and all fractions (7–8 mL each) showing A₂₈₀ greater than 0.1 unit were pooled. Protein was precipitated by adding solid ammonium sulfate (0.4 g/mL). The mixture was centrifuged at 40,000g for 30 min, dissolved in a minimum volume of HGED buffer, and dialyzed against HGED buffer plus 100 mm KCl containing Pefabloc and DTT. This fraction was designated HA-1. The column was further washed with 900 mL each of HGED buffer plus 400 mm KCl followed by HGED buffer plus 1 m KCl. Fractions from both washes were processed as described above and designated HA-2 and HA-3. PADPRP assays were performed on HA-1, HA-2, and HA-3, and the active fraction (HA-2) was used for further purification.

Chromatography on DNA-Cellulose

DNA-cellulose (Sigma) was washed extensively with HGED buffer and packed in the Econo column (2.5 × 30 cm). The column was connected to the Econo system and equilibrated with HGED buffer plus 100 mm KCl. Partially pure PADPRP from three heparin-agarose column preparations was loaded onto the DNA-cellulose column. Unbound protein was removed by washing with HGED buffer plus 100 mm KCl (200 mL; designated DC-1). The bound protein was eluted with HGED buffer plus 1 m KCl (designated DC-2). All fractions were processed as described above for PADPRP activity and protein.

Chromatography on Histone-Agarose

About 10 mL of histone-agarose (Sigma) was washed extensively with HGED buffer and packed into a clean glass column (1.0 × 5 cm). The column was equilibrated in HGED buffer plus 100 mm KCl. Active fraction from DNA-cellulose (DC-2) was further fractionated on histone-agarose by washing the column successively with HGED buffer containing 100 mm, 400 mm, and 1 m KCl. All fractions were processed as described above and dialyzed against 20 mm Tris-HCl buffer, pH 7.9, containing 100 mm KCl.

Chromatography on a Mini Q Column

A Mini Q column (Pharmacia) was connected to the Smart-LC (Pharmacia) and equilibrated by washing with 5 bed volumes of Tris-Cl, pH 8.0, containing 14 mm β-ME followed by 5 bed volumes of the same buffer containing 100 mm KCl. Active PADPRP from the histone-agarose step was loaded onto the column. The column was washed with 3 bed volumes of equilibration buffer and 400-μL fractions were collected. The column was further developed using a step gradient of KCl at 400 mm, 600 mm, and 1 m in Tris-HCl, pH 8.0. All fractions were tested for PADPRP activity, as described below.

Enzyme Assays

Catalytic activity of PADPRP was assayed following published protocols (Shah et al., 1995) with modifications suitable for the plant enzyme. The enzyme (in a total volume of 25 μL of 20 mm Hepes, pH 7.9, 100 mm KCl) was incubated with 2.5 to 5 μCi of [α-³²P]NAD⁺, 2 μg/mL final concentration of bovine histone (fraction IV), 2 μg of activated calf thymus DNA, and 0.5 mm DTT. The reactions were carried out at 6°C unless stated otherwise. At the end of the appropriate time intervals, the labeled protein was precipitated with 20% TCA. The precipitate was collected by centrifugation at 16,000g for 10 min, washed two times with 5% TCA, and counted in a liquid-scintillation counter. Protein heated at 65°C for 5 min was used as a negative control. Similar assay protocols were used for inhibitor studies.

Microsequencing

Protein samples obtained from the Mini Q column purification step were electrophoresed in duplicate on 10% polyacrylamide gels using 0.1% SDS in the running buffer (Shah et al., 1995). One-half of the gel was used to detect protein bands with a colloidal Coomassie blue staining kit (Novex) following the manufacturer's instructions. The other half was used in the activity-blot assay to confirm the position of the active PADPRP on the gel. The stained protein band corresponding to active PADPRP was cut out from the gel and used. Microsequencing was carried out at the W.M. Keck Foundation Biotechnology Resource Laboratory of Yale University (New Haven, CT). In-gel tryptic digestion of the protein, matrix-assisted laser desorption MS of the isolated peptides, and amino acid sequencing of representative peptides were carried out following published protocols (Stone et al., 1990, 1991; Williams et al., 1995; Williams and Stone, 1995).

Antipeptide Antibodies

Synthesis of the peptide antigens and antibody generation were carried out at Research Genetics (Huntsville, AL). Peptide P-1 was synthesized as a MAP following published protocols (Tam, 1988). Two New Zealand White rabbits (4–6 months old) were used for immunization. The antigen was prepared by dissolving 500 μg of MAP in 500 μL of saline mixed with 500 μL of complete Freund's adjuvant, and injected subcutaneously at three to four dorsal sites. Animals were boosted with equivalent antigen preparations at 2, 4, and 6 weeks after the first immunization. Animals were bled from the auricular artery to collect 30 to 50 of mL of blood on d 0, 27, 57, and 69. Blood samples were allowed to clot at room temperature for 15 min and serum was isolated from each sample by centrifugation at 5000g for 10 min. Cell-free serum was decanted gently into a clean tube and stored at −20°C until further use. The antibodies were purified using a HighTrap protein A-agarose column (Pharmacia) according to the manufacturer's instructions.

cDNA Cloning

Isolation of RNA, cDNA synthesis, and preparation of libraries were conducted under the direction of Dr. Xun Wang of the Genome Research Group as part of Pioneer's maize genome initiative. Total RNA was isolated from maize tissues with TRIzol Reagent (Life Technologies) using a modification of the guanidine isothiocyanate/acid phenol procedure, as described by Chomczynski and Sacchi (1987).

The selection of poly(A⁺) RNA from total RNA was performed using the PolyATract system (Promega). Synthesis of the cDNA was performed and unidirectional cDNA libraries were constructed using the SuperScript Plasmid System (Life Technologies). The selected cDNA molecules were ligated into the pSPORT1 vector between the NotI and SalI sites. The first strand of cDNA was synthesized by priming an oligo(dT) primer containing a NotI site. Individual colonies were picked and DNA was prepared either by PCR with M13 forward and reverse primers or by plasmid miniprep isolation. All of the cDNA clones were sequenced using M13 reverse primers.

Analysis

Protein concentration was estimated by the Bradford method (Bradford, 1976) using bovine γ-globulin as a standard. Activity blots, western blots, and product analysis were performed essentially following published protocols (Panzeter and Althaus, 1990; Simonin et al., 1991; Desnoyers et al., 1994; Shah et al., 1995), except that all assays were carried out at 6°C. Isolation of mRNA and northern-blot analysis were performed according to published protocols (Luehrsen, 1994).

RESULTS

Enzyme Purification and Characterization

Table I summarizes the procedure and the results of the purification steps for maize PADPRP. Disruption of the cell walls was achieved using the BioNebulizer. Crude extracts were prepared by a combination of methods used for preparation of S100, and nuclear extracts from mammalian cells (Manley et al., 1983; Shapiro et al., 1988; Mahajan, 1993). This procedure allowed us to process about 100 g of cells to obtain 0.5 to 0.6 g of crude WCE that was active for PADPRP. The crude extract was fractionated on successive chromatographic steps as indicated. Heparin-agarose was very effective in quantitative concentration of the enzyme. Chromatography on DNA-cellulose resulted in a more than 140-fold purification. When the partially pure enzyme from the histone-agarose step was used for micropurification on a Mini Q column (under the conditions described above), the PADPRP activity appeared in the flow-through fraction. Protein from this fraction showed a band of approximately 115 kD on silver-stained gel after SDS-PAGE (Fig. 1A). In addition, a minor band corresponding to about 15 kD was detected in some of the preparations. Figure 1B shows the results of a typical activity-blot assay performed using the purified PADPRP from the Mini Q flow-through fraction. A single band of approximately 115 kD was labeled with [³²P]NAD⁺ (Fig. 1B, lane M). In this experiment we also used the human PADPRP as a positive control (Fig. 1B, lane H). These results further substantiate the identity of the approximately 115-kD protein purified from maize extracts as PADPRP. Overall recovery of the enzyme activity was less than 3%. Our attempts to improve recovery of the purified enzyme by affinity chromatography using aminobenzamide-agarose (Burtscher et al., 1986; Ushiro et al., 1987; Shah et al., 1995) failed because maize PADPRP did not bind to the affinity-chromatographic resins. Use of conventional chromatographic methods resulted in rapid inactivation of the enzyme, presumably owing to the thermolability of the maize enzyme. As shown in Figure 2A, we consistently observed a loss of catalytic activity of maize PADPRP by simply leaving the sample at 15°C for 15 min. Incubation at temperatures above 30°C caused rapid inactivation. Therefore, all purification steps were performed at 4°C or on ice and all assays were performed at 6°C.

Table I.

Purification of PADPRP from maize

Step	Protein	Activitya		Purification	Recovery
		Total Activity	Specific Activity
	mg	units	units/μg	-fold	%
WCE	3319.56	265,825.01	0.05	1.00	100.00
Heparin agarose	503.62	67,073.84	0.14	3.13	25.23
DNA cellulose	3.96	73,730.16	6.52	141.50	27.74
Histone agarose	0.37	27,927.05	9.63	208.84	10.51
Mini Q	0.13	6,629.83	13.67	296.45	2.49

^a One unit of PADPRP activity is defined as the amount of enzyme required to incorporate 1 pmol of [³²P]NAD per min at 6°C. 

Figure 1.

SDS-PAGE analysis of purified maize PADPRP. A, About 300 ng of active PADPRP from the Mini Q column flow-through fraction was electrophoresed on a 4% to 20% precast SDS-PAGE gel. Protein was detected by silver staining. Numbers on the left indicate the positions and sizes of Bio-Rad silver-stain high-range molecular-mass markers. The arrow indicates the position of the PADPRP band at 115 kD. B, For the activity blot, purified maize PADPRP (M) was electrophoresed on a 4% to 20% SDS-PAGE gel and electrotransferred to a nitrocellulose filter. The assay was carried out following published protocols with modifications described in Methods. Human PADPRP (H) was used as a positive control. Autoradiography was performed at −80°C for 4 h.

Figure 2.

A, Temperature inactivation of maize PADPRP. Assays were carried out at the designated temperatures for 15 min, as detailed in the text. Data are the average of three independent experiments. B, High-resolution PAGE analysis of maize PADPRP reaction products. ADP-ribosylation of histones was performed using [³²P]NAD. Reaction products were analyzed as described by Panzeter and Althaus (1990). Autoradiography was performed at −80°C for 30 min. Gel origin (O) and positions for xylene cyanol (XC) and bromphenol blue (BPB) are indicated.

A high-resolution size analysis of the ADP-Rib polymers was performed essentially as described by Panzeter and Althaus (Fig. 2B). The presence of reaction products corresponding to 40 to 45 ADP-Rib residues is clearly seen based on the mobility of the marker dyes (Fig. 2B, XC) in this 30-min exposure. Longer exposure of this gel (not shown here) indicated that the maize enzyme added more than 60 ADP-ribosyl residues under the reaction conditions used in this study. No significant difference in the enzyme activity was observed when different purified fractions of histone were used as the acceptor substrates (data not shown). Various compounds known to inhibit mammalian PADPRP (Banasik et al., 1992; Banasik and Ueda, 1994) were tested for their effect on the activity of maize PADPRP (Fig. 3A). Except for novobiocin, all compounds tested showed significant inhibition at micromolar concentrations. Isoquinoline diol and aminonaphthalimide were equally potent and were the most potent inhibitors, whereas nicotinamide and thymidine were the least potent inhibitors. Novobiocin, a specific inhibitor of mono(ADP)-ribosyltransferases (Banasik and Ueda, 1994), did not inhibit maize PADPRP activity (Fig. 3A).

Figure 3.

A, The effect of various inhibitors on maize PADPRP activity. Reactions were carried out in the presence of different concentrations of each compound, as described in Methods. An equal volume of each solvent was added to the control reactions and used to calculate percent inactivation. The data represent the average of three experiments. The compounds tested were aminonaphthalimide (−), aminobenzamide (+), benzamide (*), 3-methoxybenzamide (○), isoquinoline diol (▵), thymidine (•), nicotinamide (⋄), and novobiocin (□). B, Disulfide-mediated dimerization of maize PADPRP. Maize WCE was prepared using the protocol described in the text, except that the buffers were devoid of DTT or β-ME. About 50 μg of protein was mixed with an equal volume of 2× SDS-sample buffer without (−) or with (+) β-ME and electrophoresed on 8% SDS-PAGE gels. Activity-blot assays were carried out as described previously. Autoradiography was performed at −80°C for 48 h. Arrows indicate the positions of the PADPRP bands corresponding to monomeric (approximately 115 kD) and dimeric (approximately 220 kD) species of the enzyme.

During our initial activity-blot experiments with crude maize PADPRP, when β-ME was omitted inadvertently from the SDS-PAGE sample buffer, we observed a faint band of approximately 220 kD. This indicated the possibility of dimerization of PADPRP. However, all of our buffers contained DTT as a reducing agent (essential to reduce the inactivation of PADPRP). Therefore, we specifically prepared three batches of maize WCE in the complete absence of added reducing agents such as DTT or β-ME. These samples were analyzed by activity-blot assays after electrophoresis in the absence or presence of β-ME. Representative results from these experiments are shown in Figure 3B. In addition to the 115-kD band, we consistently observed an approximately 220-kD band when samples were not treated with the reducing agents (Fig. 3, lane −). Inclusion of β-ME in the sample buffer caused the disappearance of the higher-molecular-mass band (Fig. 3, lane +). The intensity of the lower-molecular-mass band also appeared to be approximately twice that of the same band (Fig. 3, lane −), indicating a disulfide-mediated autodimerization of the maize enzyme. Similar experiments with the HeLa extracts did not show the presence of a dimer (data not shown).

Active PADPRP protein obtained from the Mini Q column purification step was used for microsequencing, as outlined in Methods. Amino acid sequence of the peptide P-1 was obtained initially (Table II). At that time, this peptide sequence did not show homology to any known sequences in the GenBank or SwissProt databases. Nonetheless, the P-1 sequence was used to make a synthetic peptide as an antigen for generating polyclonal antibodies in rabbits (Tam, 1988). Subsequently, the amino acid sequences of two more peptides were obtained (Table II).

Table II.

Amino acid sequence of maize PADPRP peptides

Peptide	Sequence
1	Leu-Arg-Phe-Ser-Val-Val-Gly-Gln-Ser-Lys
2	Gly-Leu-Tyr-Phe-Ala-Asp-Leu-Val-Ser-Lys
3	Glu-Ala-Ala-Asn-Glu-Trp-Ile-Glu-Lys

Isolation and microsequencing of peptides was done as detailed in Methods. Peptide 1 did not exhibit homology to any known sequence in the GenBank database. Peptides 2 and 3 exhibited homology to known animal PADPRP sequences.

Immunochemical Analysis

Anti-P-1 antibodies were synthesized by immunizing rabbits with MAP P-1 (see above). For western-blot analysis, duplicate samples of partially pure PADPRP were electrophoresed on 4% to 20% SDS-PAGE and transferred onto a PVDF membrane. One-half of the membrane was used for the activity-blot assays and the other half was used for immunoblotting with anti-P-1 antibodies. The activity-blot data establish the identity of a 115-kD band as PADPRP (Fig. 4A), and the western-blot analysis clearly shows cross-reactivity of anti-P-1 antibodies with maize PADPRP (Fig. 4B). In another set of western-blot experiments, the human (HeLa) and maize PADPRP samples were electrophoresed in duplicate and electroblotted as described above. One-half of the blot was treated with anti-human PADPRP antibodies (Bio-Rad) and the other half was treated with anti-P-1 antibodies. As seen in Figure 5, the anti-maize PADPRP antibodies cross-reacted with the maize enzyme but not with the human enzyme (A), and the anti-human PADPRP polyclonal antibodies cross-reacted with the human enzyme but not with the maize enzyme (B). Similar experiments were performed with polyclonal (C) and monoclonal (D) antibodies against the bovine PADPRP. Neither antibody cross-reacted with the maize enzyme, but both antibodies cross-reacted with the human enzyme.

Figure 4.

Anti-P-1 antibodies specifically cross-react with maize but not human PADPRP. About 25 μg of partially pure maize PADPRP was electrophoresed on 4% to 20% SDS-PAGE. Samples were run in duplicate and blotted onto a PVDF membrane following the manufacturer's instructions. After blocking, the membrane was cut into two halves. One-half was used for activity blots (A) as described previously. The other half, used for western blots (B), was incubated with primary (1:10,000 dilution) and secondary (1:2,000 dilution) antibodies, washed extensively, and developed using the enhanced chemiluminescence kit (Amersham). This image was made by superimposing the autoradiographs and PVDF membranes. Numbers denote the molecular size of the SeeBlue markers.

Figure 5.

Immunological analysis of maize PADPRP. About 25 μg of partially pure human (H) or maize (M) PADPRP was electrophoresed on 4% to 20% SDS-PAGE gels run in parallel and blotted onto PVDF membranes following the manufacturer's instructions. After blocking, membranes were incubated with individual antibodies, washed extensively, incubated with suitable secondary antibodies, and developed using the enhanced chemiluminescence kit (Amersham). A, Anti-P-1 polyclonal antibody; B, anti-human PADPRP polyclonal antibody; C, anti-bovine PADPRP polyclonal antibody; and D, anti-bovine PADPRP monoclonal antibody. The arrow indicates the position of the specific PADPRP band. B and C were overexposed to allow detection of any cross-reactivity of antibodies to the maize PADPRP. Numbers denote positions of the SeeBlue molecular mass markers.

cDNA Characterization

Analysis of several expressed sequence tags recovered from our Maize Genome Project showed significant homology to animal PADPRP sequences. One of these clones, containing a 3.45-kb insert, was analyzed further. A partial nucleotide sequence of this clone encompassing the complete coding region of maize PADPRP is shown in Figure 6. This sequence contains an open reading frame beginning at nucleotide 149 and ending at nucleotide 3127. Within this reading frame is the 2943-nucleotide-long cDNA for maize PADPRP, which encodes the 980-amino acid protein (indicated by the one-letter code for the amino acid sequence). Regions of the sequence corresponding to zinc fingers, the putative nuclear localization signal, and the NAD-binding region are indicated. Amino acid sequences of the three peptides obtained via microsequencing of the maize protein are also shown (Fig. 6).

Figure 6.

Nucleotide and deduced amino acid sequences of maize PADPRP cDNA. The partial nucleotide sequence of a 3.45-kb cDNA clone (obtained from the ear-leaf collars cDNA library) is shown in the top line. Numbers denote nucleotide positions of the cDNA insert from the 5′ to the 3′ direction. The deduced amino acid sequence of the open reading frame is given in single-letter code. Important features of the modular structure of maize PADPRP shown here include zinc fingers I and II (double underline), the putative nuclear localization signal (underline), and the C-terminal catalytic domain (in italics). The consensus NAD-binding sites are shown in bold. Lys-861 and Glu-957 (*) are the conserved residues found to be important for catalytic activity of human PADPRP (de Murcia et al. 1994). The amino acid sequence of the three peptides obtained by microsequencing is also highlighted (dotted underline). The position of the stop codon is marked (⊗). The 33-nucleotide insert (encoding part of zinc finger I) that is missing in the GenBank sequence no. AJ 222589 is underlined.

Comparison of the maize PADPRP amino acid sequence indicated a 40% to 42% identity and about 50% similarity with most vertebrate PADPRP sequences. The extent of identity and similarity of maize PADPRP with the two invertebrate sequences analyzed (Sarcophaga peregrina and Drosophila melanogaster) was slightly lower (37% and 48%, respectively). The maize enzyme appears to be the smallest of all of the reported full-length PADPRPs. Alignment of the maize PADPRP sequence with the amino acid sequence of the human, bovine, avian, frog, flesh fly, and fruit fly enzyme indicates three major gaps of about 12, 9, and 36 amino acid residues, all of which are in the automodification domain (de Murcia et al., 1994; Smulson et al., 1994; Uchida and Miwa, 1994; Masson et al., 1995; Ruf et al., 1996). In addition, the vertebrate enzyme sequence has four to five additional amino acids at the C terminus. The maize PADPRP sequence has two seven-residue insertions, one in the automodification domain and one in the catalytic domain. Furthermore, several one- to four-residue insertions are seen throughout the length of the maize PADPRP molecule. Despite these differences in all of these sequences, the catalytic domain (Fig. 6) and the DNA-binding domain containing two zinc fingers (Fig. 7) are very similar. Moreover, the critical residues for zinc-finger formation and NAD⁺ binding (see legends to Figs. 6 and 7) are conserved in all of these sequences.

Figure 7.

The zinc-finger domain is conserved in animal and plant PADPRP. The comparison of the deduced amino acid sequence of the zinc-finger domain of the maize PADPRP with some animal enzyme sequences was performed using the PileUp program (Genetics Computer Group, Madison, WI). A, Finger I; B, finger II. Cys and His residues (underlined) involved in zinc binding are conserved in both fingers. Other conserved residues are indicated in boxes. Gaps in the sequence are shown by dashes.

The cDNA clone described above was used as a probe for northern-blot analysis of poly(A⁺) RNA from maize leaves. As shown in Figure 8, the probe hybridized to an mRNA species corresponding to approximately 4 kb. These results were further confirmed using two separate PCR probes of about 500 bp synthesized using the cDNA as a template (data not shown). These results strongly indicate that the cDNA clone we have obtained encodes the maize PADPRP.

Figure 8.

Expression of PADPRP in maize leaves. mRNA was prepared from maize leaves as described in Methods. For northern blots, 10 μg of RNA was electrophoresed on formaldehyde-agarose gels, transferred to nitrocellulose membranes, and hybridized to a ³²P-labeled maize PADPRP cDNA probe. Autoradiography was performed at −80°C for approximately 48 h.

DISCUSSION

In this paper we present a comprehensive biochemical, immunological, and molecular analysis of a plant PADPRP. To the best of our knowledge, this is the first report describing purification of a plant PADPRP, generation of antibodies specifically cross-reacting with a plant PADPRP, and a cDNA encoding a protein exhibiting all of the known characteristic features of a PADPRP molecule. Several reports describing purification and detailed characterization of PADPRP from a variety of animal sources have appeared during the past three decades (for review, see Ueda and Hayaishi, 1985; Uchida and Miwa, 1994; Shah et al., 1995). In contrast, plant PADPRP has been difficult to purify, primarily because of the lack of milder methods to break plant cell walls. We overcame this problem by using the BioNebulizer, which allows the removal of plant cell walls under milder conditions. Furthermore, we employed a combination of protocols used for preparing extracts from animal cells to obtain large amounts of enzymatically active plant cell extracts. We have successfully used this protocol to prepare extracts from a variety of plant tissues of different species (P. Mahajan and Z. Zuo, unpublished results). The crude WCE was concentrated and purified by a series of chromatographic steps using group-specific ligands such as heparin, DNA, and histone immobilized on agarose. Affinity chromatography using aminobenzamide-agarose (Ushiro et al., 1987; Panzeter et al., 1994) was unsuccessful because the maize enzyme failed to bind to the affinity matrix. This is particularly surprising because aminobenzamide and several other specific inhibitors of animal PADPRP (Banasik et al., 1992; Banasik and Ueda, 1994) also inhibit the maize enzyme (Fig. 2). Kofler et al. (1993) have also made similar observations for the enzyme from Dictyostelium discoideum.

Further biochemical analysis of the maize enzyme showed important similarities to known PADPRP in properties such as molecular size and ADP-Rib polymer formation in vitro (Figs. 1 and 2), as well as specificity for acceptor and donor substrates and kinetic constants (P. Mahajan and Z. Zuo, unpublished data). Nonetheless, maize PADPRP also exhibits biochemical and immunological properties that are distinct from its animal counterpart. For example, the maize PADPRP exhibited disulfide-mediated dimer formation, whereas the human PADPRP prepared under similar conditions did not. Based on their kinetic studies, Mendoza-Alvarez and Alvarez-Gonzalez (1993) proposed that PADPRP exists as a catalytic dimer. Macromolecular self-association of PADPRP has also been shown using chemical cross-linking (Bauer et al., 1990). Dimerization of PADPRP has been implicated as important for automodification. Our data presented in Figure 3 provide a direct demonstration of this phenomenon. These results do not rule out the possibility that the high-molecular-mass species may represent a heterodimer of PADPRP with other cellular protein. Additional experiments are under way to answer this question. Nonetheless, it seems likely that the association of the animal enzyme is not mediated through a disulfide bond, reflecting differences in amino acid composition of the plant and animal enzymes.

The amino acid sequence obtained through microsequencing of the maize PADPRP also supports the argument for the similarities and differences in the biochemical properties. Thus, peptide P-1 does not show similarity to any known sequences in the database, but P-2 shows high homology to all known PADPRP sequences. We have also presented data to show immunological differences between the maize and animal enzymes. The antibodies against maize PADPRP peptides do not cross-react with the human enzyme. Similarly, antibodies against the human or bovine enzyme do not cross-react with the maize PADPRP. This, to some extent, may also have contributed to the difficulty in cloning plant PADPRP by homology. Earlier, Lepiniec et al. (1995) published the sequence of a clone recovered from an Arabidopsis cDNA library. (Recently, the sequence for a maize homolog of this Arabidopsis clone was deposited in GenBank, accession no. AJ222588). Sequence of the Arabidopsis clone contained an open reading frame encoding a protein of 539 amino acids, which showed significant homology to the NAD⁺-binding domain of animal PADPRP. However, it was not clear from these studies if this protein showed any activities associated with PADPRP, such as DNA binding and/or hydrolysis of NAD⁺. Furthermore, this clone lacked the entire zinc-finger domain characteristic of PADPRP sequences published to date. Although the sequence was shown to contain a helix-loop-helix domain, it was not clear if this putative homolog had any DNA-binding activity. Thus, the functional relevance of the Arabidopsis clone to PADPRP activity was unclear. We have presented biochemical and immunological analyses of the maize enzyme along with the molecular characterization.

In our efforts to clone a plant PADPRP, we have analyzed several expressed sequence tags obtained through our Maize Genome Project. One of these clones contained an open reading frame encoding the full-length maize PADPRP. Based on the deduced amino acid sequence, maize PADPRP is a protein of 980 residues. It must be emphasized here that the deduced amino acid sequence of this clone confirms the sequence of the three peptides obtained independently through microsequencing of the protein purified from maize cell extracts. Furthermore, the derived amino acid sequence also shows the modular structure typically observed in all known PADPRP sequences. Important features, including the zinc fingers, two dinucleotide-binding domains (also known as the Rossman fold characterized by the sequences GXXXGKG and GXGKT), the Lys residue at position 861 (comparable to residue 893 in the human sequence), and the catalytically important Glu residue at position 957 (comparable with residue 988 in the human sequence), are conserved. Based on these results, we conclude that the plant cDNA described here encodes a protein that is similar structurally and functionally to the animal PADPRP. While this paper was in preparation, a sequence for maize PADPRP (accession no. AJ222589) was deposited in the GenBank. The nucleotide sequence and the deduced amino acid sequence presented here are essentially identical to the sequence deposited in the GenBank except for one important difference. The sequence deposited in the GenBank lacks a stretch of 33 nucleotides (underlined nucleotides in Fig. 6), which is present in our sequence as well as in other PADPRP sequences. These amino acids are from the part of zinc finger 1 involved in forming the DNA-binding domain, and many of these 11 residues have been conserved across different species (Fig. 7A). Furthermore, we have confirmed the presence of these 11 amino acids by reverse-transcriptase PCR using maize leaf mRNA as a template (data not shown). Therefore, we believe that deletion of these amino acids will drastically affect the function of the maize PADPRP molecule. It is also possible that the difference in the two maize PADPRP sequences is a result of differential splicing. If this is true, it might provide a mechanism for regulating PADPRP levels in maize. Experiments are currently under way to test this prediction. When the maize cDNA was used as a probe for a northern-blot analysis of maize leaf mRNA (Fig. 8), we detected a band of 4 kb, which corresponds to the size of the mature message for PADPRP. Thus, this cDNA clone would be a valuable authentic reagent for studying the biological role of plant PADPRP.

The biological function of PADPRP has remained a topic of debate (Auer et al., 1995; Kupper et al., 1995; Rosenthal et al., 1995; Jeggo, 1997). Recent studies using PADPRP- knockout mice (Heller et al., 1995; Wang et al., 1995; Menissier de Murcia et al., 1997) indicated that animals null for PADPRP were viable and fertile but may have elevated levels of genomic instability. Similar observations have been reported for dominant-negative or mutant cell lines (Chatterjee and Berger, 1994; Schreiber et al., 1995; Kupper et al., 1996). A role for PADPRP in programmed cell death (Bernardi et al., 1995; Tewari et al., 1995) and oxidative stress response in animals (Heller et al., 1995) as well as plants (Berglund and Ohlsson, 1995) has also been proposed. In this connection, we note that the specific sequence EVD/G, where animal PADPRP is cleaved during programmed cell death (Lazebnik et al., 1994), is absent in the maize PADPRP sequence. It is tempting to speculate that our observation may lead to a novel role (and/or mechanism) for PADPRP in plant cell apoptosis. The availability of the full-length cDNA and specific antibodies to a plant PADPRP should also prove helpful in elucidating the role of PADPRP in plant growth, metabolism, and defense.

Abbreviations:

β-ME

β-mercaptoethanol

MAP

multiple antigenic peptide

PADPRP

poly(ADP)-Rib polymerase

WCE

whole-cell extract