Generation and Characterization of Recombinant Antibody-Like ADP-Ribose Binding Proteins

Abstract

ADP-ribosylation is an enzyme-catalyzed post-translational modification of proteins in which the ADP-ribose (ADPR) moiety of NAD⁺ is transferred to a specific amino acid in a substrate protein. The biological functions of ADP-ribosylation are numerous and diverse, ranging from normal physiology to pathological conditions. Biochemical and cellular studies of the diverse forms and functions of ADPR require immunological reagents that can be used for detection and enrichment. The lack of a complete set of tools that recognize all forms of ADPR [i.e., mono-, oligo-, and poly(ADP-ribose)] has hampered progress. Herein, we describe the generation and characterization of a set of recombinant antibody-like ADP-ribose binding proteins, in which naturally-occurring ADPR binding domains, including macrodomains and WWE domains, have been functionalized by fusion to the Fc region of rabbit immunoglobulin. These reagents, which collectively recognize all forms of ADPR with different specificities, are useful in a broad array of antibody-based assays, such as immunoblotting, immunofluorescent staining of cells, and immunoprecipitation. Observations from these assays suggest that the biology of ADPR is more diverse, rich, and complex than previously thought. The ARBD-Fc fusion proteins described herein will be useful tools for future exploration of the chemistry, biochemistry, and biology of ADP-ribose.

Introduction

ADP-ribosylation is an enzyme-catalyzed post-translational modification of proteins in which the ADP-ribose (ADPR) moiety of NAD⁺ is transferred to a specific amino acid in a substrate protein, with release of the nicotinamide moiety.¹ ADP-ribosylation reactions are catalyzed by a variety of ADP-ribosyltransferase (ART) enzymes, including (1) bacterial toxins, such as Cholera toxin and Diphtheria toxin^2,3, (2) Ecto-ADP-ribosyltransferases (ectoARTs), which are extracellular, membrane-bound, or secretory proteins that share homology in the catalytic domain with Clostridium C2 and C3 toxins (ARTCs)⁴, (3) members of the sirtuin family of enzymes in eukaryotes and prokaryotes^5–7, and (4) members of the poly(ADP-ribose) polymerase (PARP) family of enzymes, which share homology in the catalytic domain with Diphtheria toxin (ARTDs).^8–10 With the exception of some members of the PARP family, ARTs generally catalyze mono(ADP-ribosyl)ation reactions.^2,4,8,11,12 Most bacterial ARTs transfer ADP-ribose onto arginine residues, although asparagine and other amino acids are targeted as well.^{2, 3} PARP family mono- and poly(ADP-ribosyl)transferase enzymes transfer ADP-ribose primarily onto glutamate, aspartate, serine, and lysine residues, generating mono-, oligo-, and poly(ADPR-ribose) (MAR, OAR, and PAR, respectively).^8,11–15

The biological functions of ADP-ribosylation are numerous and diverse, ranging from normal physiology to pathological conditions, such as bacterial toxicity, aging, and cancer.^16–18 Recent advances in PARP-directed therapeutics have shown promise for the treatment of cancers.^19,20 While the cellular targets of bacterial toxins and the molecular effects of the ADP-ribosylation reactions that they catalyze have been well characterized^2,3, the specific targets and effects of PARP-mediated ADP-ribosylation are less well understood.¹⁶ The few examples of the latter include: (1) NFAT (ADP-ribosylation increases DNA binding²¹), (2) C/EBPβ (ADP-ribosylation inhibits DNA binding and transcriptional activity²²), (3) p53 (ADP-ribosylation inhibits nuclear export²³), and (4) NELF-E (ADP-ribosylation inhibits RNA binding and NELF-dependent promoter-proximal pausing by RNA polymerase II²⁴). ADP-ribosylation of proteins can alter the biochemical activity of the ADP-ribosylated protein or create new interaction surfaces that drive protein-protein interactions.⁸

Interestingly, nature has devised protein modules that specifically recognize and bind to various forms of ADP-ribose.^25,26 These modules are found in a variety of proteins with diverse functions, including a number of PARP family members, and are likely to mediate many of the biological functions of ADP-ribosylation.^8,25,26 Some well characterized ADP-ribose binding domains (ARBDs) include macrodomains and WWE domains.^25,26 Macrodomains recognize MAR, as well as the terminal ADPR moieties in OAR and PAR, allowing them to bind to all three forms of ADPR (i.e., MAR, OAR, and PAR) (Figure 1, A and B).^27–29 In contrast, WWE domains recognize the iso-ADPR linkages joining ADPR monomers, restricting their binding to OAR and PAR (Figure 1, A and B).^30–32 In addition to the biology that they provide, these and other types of ARBDs can be functionalized to serve as useful research tools for the molecular recognition of various forms of ADPR.^{22,24,29,33–36}

Figure 1. Design of antibody-like ADP-ribose binding reagents.

(A and B) The chemical structure of (A) a poly(ADP-ribosyl)ated amino acid or (B) a mono(ADP-ribosyl)ated amino acid, with the amino acids shown in heteroatomic colors and the ADP-ribose units colored in grey (proximal to the amino acid) and black. The blue box with a dashed line highlights the chemical moiety in the ADP-ribose modification recognized by the WWE domain (the iso-ADP-ribose linkage between two ADP-ribose monomers). The green boxes with dashed lines highlight the chemical moiety in the ADP-ribose modification recognized by the macrodomains. The WWE and macrodomains used in this study are indicated (species and protein).

(C) Ribbon diagram depicting the X-ray crystal structure of a monoclonal IgG antibody (PDBID:1IGY), with the variable fragment (Fv) in green and the homodimerized Fc fragment in purple and pink.

(D) Schematic diagram of the plasmid constructs used to express the ADP-ribose binding domain-Fc (ARBD-Fc) fusion proteins in bacteria. The constructs contain DNA segments encoding: (1) 10xHis and/or Strep[II] tags, (2) an ADP-ribose binding domain (green), (3) a flexible glycine and serine linker, and (4) rabbit IgG constant fragment (Fc) (pink).

Although recent developments in the mass spectrometry-based identification of ADPR-modified amino acids have enhanced the study of specific ADP-ribosylation events on target proteins³⁷, the lack of a complete set of immunological tools that recognize the diverse forms of ADPR has hampered progress. Anti-ADPR polyclonal antibodies have been reported, but the specificity and utility of these antibodies has not been assessed broadly and, like other polyclonal antibodies, they are difficult to produce in a constant supply^38–44 Instead, the PARP field has relied on the anti-PAR monoclonal antibody 10H, which is thought to bind to PAR with a lower limit for the length of polymers detected of 10 to 20 ADPR units.^45,46 Although useful, this antibody has left the field blind to mono- and oligo(ADP-ribosyl)ation, as well as their biological importance. Herein, we describe the generation and characterization of a set of recombinant antibody-like ADP-ribose binding proteins, in which natural ARBDs have been functionalized with the Fc region of rabbit immunoglobulin. These reagents are useful in a broad array of antibody-based assays, such as immunoblotting, immunofluorescent staining of cells, and immunoprecipitation.

Materials and Methods

Antibodies and ADPR Binding Reagents

Alexa Fluor 488- and HRP-conjugated goat anti-rabbit IgG antibodies (A-11008 and 31346, respectively) and HRP-conjugated goat anti-mouse IgG antibody (31430) were purchased from Thermo Fisher Scientific. The anti-poly(ADP-ribose) mouse monoclonal antibody 10H was purchased from Millipore (MAB3192). The custom anti-PARP-1 rabbit polyclonal antiserum, which was generated by using an antigen comprising the amino-terminal half of PARP-1⁴⁷, is now available from Active Motif (39559). The ARBD-Fc fusion proteins described herein are now available through Millipore (MABE1016, RRID AB_2665466; MABE1031, RRID AB_2665467: MABE1075, RRID AB_2665468; MABE1076, RRID: AB_2665468).

Construction of Plasmid Vectors for the Expression of ADPR Binding Domain-Fc Fusion Proteins in Bacteria

WWE(RNF-146)-Fc.

DNA encoding the WWE(RNF146)-Fc reagent, comprising (1) the WWE domain from H. sapiens RNF146^31,32,48,49 (UniProt ID Q9NTX7), which we refer to as WWE(RNF146), (2) a flexible linker, and (3) the constant or “Fc” region of rabbit immunoglobulin, was synthesized as three overlapping gene blocks by Integrated DNA Technologies (IDT). These three fragments were assembled and amplified into a single double stranded DNA fragment using PCR, and were subsequently cloned into the pET19b vector (Novagen) between the Ndel and BamHI restriction endonuclease sites.

Macro(AF1521)-Fc and Macro(mH2A1.1)-Fc.

DNA encoding the macrodomain from A. fulgidus AF1521^27,50 (UniProt ID O28751) or H. sapiens macroH2A1.1^28,29 (UniProt ID A0A0D2UG83) [i.e., Macro(AF1521) and Macro(mH2A1.1), respectively] were amplified using PCR from plasmid DNA (kindly provided by from M. Bycroft and Y. Yu, respectively). The WWE cassette from the WWE(RNF146)-Fc/pET19b plasmid (described above) was excised using the NdeI and SalI sites (the latter is within the linker region) and replaced with PCR amplified DNA encoding Macro(AF1521) and Macro(mH2A1.1), which were also digested using NdeI and SalI.

Macro3X(PARP14)-Fc and Macro2/3(PARP14)-Fc.

DNA encoding either the triple macrodomain cassette [i.e. Macro3X(PARP14)] or macrodomains 2 and 3 [i.e. Macro2/3(PARP14)] from H. sapiens PARP-14⁵¹ (UniProt ID Q460N5) was amplified from cDNA prepared as previously described.²⁴ The WWE cassette from the WWE(RNF146)-Fc/pET19b plasmid (described above) was excised using the NdeI and SalI sites, and replaced with PCR amplified DNA encoding Macro3X(PARP14) or Macro2/3(PARP14), which were also digested using NdeI and SalI.

Expression and Purification of the Antibody-like ADPR Binding Reagents in Bacteria

Expression.

The ADPR binding reagents were expressed in bacteria using the pET19b-based vectors described above. Due to expression issues with the Macro3X(PARP14)-Fc, a different protocol was used. E. coli strain BL21(DE3) Rosetta2 pLysS was made competent using a CaCl₂ protocol and transformed with the pET19b-based plasmids encoding one of the ADPR binding reagents described above. For WWE(RNF146)-Fc, Macro(AF1521)-Fc, Macro(mH2A1.1)-Fc, and Macro2/3(PARP14)-Fc, the transformed bacteria were grown in LB containing ampicillin and chloramphenicol at 37°C until reaching an OD_{595 nm} of 0.4–0.6. Recombinant protein expression was induced by the addition of 1 mM IPTG for 2 hours at 37°C. For Macro3X(PARP14)-Fc, the transformed bacteria were grown in LB containing ampicillin and chloramphenicol at 37°C until reaching an OD_{595 nm} of 0.2. The culture was cooled to 18°C and grown to an OD_{595 nm} of <0.8. Recombinant protein expression was then induced by the addition of 0.1 mM IPTG for 18 hours at 18°C. In all cases, the cells were collected by centrifugation, and the cell pellets were flash frozen in liquid N₂ and stored at −80°C.

Purification.

The frozen cell pellets were thawed on wet ice and lysed by sonication in Ni-NTA Lysis Buffer (10 mM Tris•HCl pH 7.5, 0.5 M NaCl, 0.1 mM EDTA, 0.1% NP-40, 10% glycerol, 10 mM imidazole, 1 mM PMSF, 1 mM β-mercaptoethanol). The lysates were clarified by centrifugation at 15,000 RPM using an SS34 rotor (Sorvall) at 4°C for 45 minutes. The supernatant was incubated with 1 mL of Ni-NTA resin equilibrated in Ni-NTA Equilibration Buffer (10 mM Tris•HCl pH 7.5, 0.5 M NaCl, 0.1% NP-40, 10% glycerol, 10 mM imidazole, 1 mM β-mercaptoethanol) at 4°C for 2 hours with gentle agitation. The resin was collected by centrifugation at 4°C for 10 minutes at 1000 x g and the lysate (supernatant) was removed. The resin was washed three times with Ni-NTA Wash Buffer (10 mM Tris•HCl pH 7.5, 1 M NaCl, 0.2% NP-40, 10% glycerol, 10 mM imidazole, 1 mM PMSF). The recombinant proteins were then eluted using Ni-NTA Elution Buffer (10 mM Tris•HCl pH 7.5, 0.2 M NaCl, 0.1% NP-40, 10% glycerol, 500 mM imidazole, 1 mM PMSF, 1 mM β-mercaptoethanol). The eluates were collected by centrifugation (4°C for 10 minutes at 1000 x g) and dialyzed in Ni-NTA Dialysis Buffer (10 mM Tris•HCl pH 7.5, 0.2 M NaCl, 10% glycerol, 10 mM imidazole, 1 mM PMSF, 1 mM β-mercaptoethanol). The dialyzed proteins were quantified using a Bradford protein assay (Bio-Rad), aliquoted, flash frozen in liquid N₂ and stored at −80°C. To assess the purity and quality of each purified ADPR binding reagent, 2 μg of purified protein was subjected to SDS-PAGE and stained with Coomassie brilliant blue.

Expression and Purification of PARP-1 and PARP-3 in Insect Cells

PARP-1 (UniProt ID P09874) and PARP-3 (UniProt ID Q9Y6F1) were expressed in Sf9 insect cells and purified as described previously²⁴, with minor modifications.

Expression.

Briefly, Sf9 insect cells cultured in SF-II 900 medium (Invitrogen) were transfected with 1 μg of bacmid driving expression of FLAG epitope-tagged wild-type PARP-1 or PARP-3 using Cellfectin transfection reagent (Invitrogen) as described by the manufacturer. After three days, the medium was supplemented with 10% FBS, penicillin and streptomycin, and collected as a baculovirus stock. After multiple rounds of amplification of the viral stock, the resulting high titer baculovirus was used to infect fresh Sf9 cells in 500 mL suspension cultures to induce expression of PARP-1 or PARP-3 for two days. The PARP-expressing Sf9 cells were then collected by centrifugation, flash frozen in liquid N₂, and stored at −80°C.

Purification.

PARP-1- or PARP-3-containing Sf9 cell pellets were thawed on wet ice. The cells were resuspended in FLAG Lysis Buffer [20 mM HEPES pH 7.9, 0.5 M NaCl, 4 mM MgCl₂, 0.4 mM EDTA, 20% glycerol, 250 mM nicotinamide, 2 mM β-mercaptoethanol, 2x protease inhibitor cocktail (Roche)] and lysed by Dounce homogenization (Wheaton). The lysate was clarified by centrifugation at 15,000 RPM using an SS34 rotor (Sorvall) at 4°C for 10 minutes, mixed with an equal volume of FLAG Dilution Buffer (20 mM HEPES pH 7.9, 10% glycerol, 0.02% NP-40), sonicated for 15 seconds, and clarified by centrifugation again. The clarified lysate was mixed with anti-FLAG M2 agarose resin (Sigma), washed twice with FLAG Wash Buffer #1 (20 mM HEPES pH 7.9, 150 mM NaCl, 2 mM MgCl₂, 0.2 mM EDTA, 15 % glycerol, 0.01% NP-40, 100 mM nicotinamide, 0.2 mM β-mercaptoethanol, 1 mM PMSF, 1 μM aprotinin, 100 μM leupeptin), twice with FLAG Wash Buffer #2/High Salt (20 mM HEPES pH 7.9, 1 M NaCl, 2 mM MgCl₂, 0.2 mM EDTA, 15% glycerol, 0.01% NP-40, 100 mM nicotinamide, 0.2 mM β-mercaptoethanol, 1 mM PMSF, 1 μM aprotinin, 100 μM leupeptin), and twice with FLAG Wash Buffer #3 (20 mM HEPES pH 7.9, 150 mM NaCl, 2 mM MgCl₂, 0.2 mM EDTA, 15% glycerol, 0.01% NP-40, 0.2 mM β-mercaptoethanol, 1 mM PMSF). The FLAG-tagged PARP proteins were eluted from the anti-FLAG M2 agarose resin using FLAG Wash Buffer #3 containing 0.2 mg/mL FLAG peptide (Sigma). The eluted proteins (~0.5 mg/mL) were quantified using a Bradford protein assay (Bio-Rad), aliquoted, flash frozen in liquid N₂ and stored at −80°C. To assess the purity and quality of the purified proteins, 1 μg of purified protein was subjected to SDS-PAGE and stained with Coomassie brilliant blue.

Expression and Purification of ARH3 in Bacteria

Expression.

6xHis-tagged human ARH3 (UniProt ID Q9NX46) was expressed in E. coli strain BL21(DE3) using a pET28b-based bacterial expression vector. The transformed bacteria were grown in LB containing ampicillin at 37°C until reaching an OD_{595 nm} of 0.4–0.6. Recombinant protein expression was induced by the addition of 1 mM IPTG for 2 hours at 37°C. The cells were collected by centrifugation, and the cell pellets were flash frozen in liquid N₂ and stored at −80°C.

Purification.

Recombinant 6xHis-tagged ARH3 was purified using Ni-NTA affinity chromatography as described above for the antibody-like ADPR binding reagents.

In Vitro Auto(ADP-ribosyl)ation Reactions with PARP-1 and PARP-3 to Generate Mono-, Oligo-, and Poly(ADP-ribose) Standards

In Vitro Auto(ADP-ribosyl)ation Reactions.

Five hundred ng of purified recombinant PARP-1 or PARP-3 were incubated at 25°C in a 100 μL reaction volume [20 mM HEPES pH 8.0, 5 mM MgCl₂, 5 mM CaCl₂, 0.01% NP-40, 25mM KCl, 1 mM DTT, 0.1 mg/mL sonicated salmon sperm DNA (Stratagene), 0.1 mg/mL BSA (Sigma)] under the following conditions: (1) PARP-1 with 250 μM NAD⁺ for 5 minutes for poly(ADP-ribose), (2) PARP-1 with 3 μM NAD⁺ for 30 minutes for oligo(ADP-ribose), and (3) PARP-3 with 250 μM NAD⁺ for 30 minutes for mono(ADP-ribose). All reactions were stopped by the addition of one third of a reaction volume of 4x SDS-PAGE Loading Buffer, followed by heating to 100°C for 5 min.

Deproteinization and Analysis of Poly(ADP-ribose) Synthesized In Vitro.

Purified recombinant human PARP-1 was incubated with NAD⁺ and sheared salmon sperm DNA under the conditions described above to generate free poly(ADP-ribose). The PAR product was deproteinized, precipitated, pelleted, dissolved, and subjected to PAGE analysis with subsequent silver staining based on protocols described previously.^52,53 Additional details are provided in the Supporting Methods and Materials.

Cell Culture

293T transformed human embryonic kidney cells, HeLa S3 cervical cancer cells, 3T3-L1 mouse pre-adipocytes, and MCF-7 breast cancer cells were obtained from the ATCC. The 293T cells and HeLa S3 cells were maintained in Dulbecco’s Modified Eagle’s Medium (Sigma) supplemented with 10% fetal bovine serum (Atlanta Biologicals) and 1% penicillin/streptomycin (Sigma). The MCF-7 cells were maintained in Minimum Essential Medium Eagle (Sigma) supplemented with 5% calf serum (Sigma). Cell stocks were passaged at regular intervals and plated separately for experiments.

Preparation of Extracts from Mammalian Cells for Immunoblotting

Mammalian cell lines (e.g. HeLa S3, MCF-7) were grown to ~80–90% confluence, treated for 1 hour with or without 20 μM of the PARP inhibitor PJ34, and then treated for an additional 5 minutes with or without 2 mM H₂O₂. Treated cells were gently washed and collected in ice cold PBS, and then pelleted by centrifugation at 450 x g for 5 minutes.

Preparation of Nuclear Extracts.

The packed cell volume (PCV) was estimated, and the cell pellets were resuspended to homogeneity in 5x PCV of Isotonic Lysis Buffer [10 mM Tris•HCl pH 7.5, 2 mM MgCl₂, 3 mM CaCl₂, 0.3 M sucrose, 1 mM DTT, 1x protease inhibitor cocktail (Roche), 20 μM PJ34 (Sigma), and 500 nM ADP-HPD (a PARG inhibitor; Millipore)] and incubated on ice for 15 minutes. NP-40 was added from a 10% solution in Isotonic Lysis Buffer to a final concentration of 0.6%, and the mixture was vortexed vigorously for 10 seconds. The lysate was subjected to a short burst of centrifugation for 30 seconds at 11,000 rpm in a microcentrifuge at 4°C to collect the nuclei. The pelleted nuclei were resuspended in ice cold Nuclear Extraction Buffer [20 mM HEPES pH 7.9, 1.5 mM MgCl₂, 0.42 M NaCl, 0.2 mM EDTA, 25% (v/v) glycerol, 1 mM DTT, 1x protease inhibitor cocktail, 20 μM PJ34 inhibitor, and 500 nM ADP-HPD) and incubated with gentle mixing for 15 minutes at 4°C. The mixture was subjected to centrifugation at maximum speed in a microcentrifuge for 10 minutes at 4°C to remove the insoluble material. The supernatant was collected as the soluble nuclear extract. Protein concentrations for the nuclear extracts were determined using a Bradford protein assay (Bio-Rad). The extracts were aliquoted, flash frozen in liquid N₂ and stored at −80°C.

Preparation of Whole Cell Extracts.

The cell pellets were resuspended in 1x Lysis Buffer [10 mM HEPES pH 8.0, 2 mM MgCl₂, 1% SDS, 250 U Universal Nuclease (Pierce), 1x protease inhibitor cocktail (Roche)] and incubated for 5 minutes at room temperature with mixing to generate whole cell protein extracts. Protein concentrations for the whole cell extracts were determined using a BCA protein assay (Pierce). The extracts were aliquoted, flash frozen in liquid N₂ and stored at −80°C.

Aliquots of nuclear or whole cell extract were mixed with one third of a volume of 4x SDS Loading Buffer (200 mM Tris•HCl pH 6.8, 8% SDS, 40% glycerol, 4% β-mercaptoethanol, 50 mM EDTA, 0.08% bromophenol blue), followed by heating to 100°C for 5 min. The cell extracts were subjected to immunoblotting as described below.

Dot Blotting and Immunoblotting

Preparation of Nitrocellulose Membranes for Dot Blotting and Immunoblotting.

For dot blotting, 3-fold dilutions of the PARP-1 or PARP-3 ADP-ribosylation reaction products described above were spotted in 0.5 μL amounts to a nitrocellulose membrane and dried to bind the ADP-ribosylated proteins to the membrane. For immunoblotting, the PARP-1 or PARP-3 ADP-ribosylation reaction products, 10 to 30 μg of nuclear or whole cell extract, or aliquots of immunoprecipitated proteins, were resolved on a 10% PAGE-SDS gel and transferred to a nitrocellulose membrane.

Dot Blotting and Immunoblotting.

The membranes were blocked for 1 hour at room temperature in Tris-Buffered Saline with 0.05% Tween (TBST) containing 5% non-fat dry milk. Primary antibodies/detection reagents [e.g., ADPR binding reagents, 10H monoclonal antibody, PARP-1 polyclonal antibody) were diluted in 1% non-fat dry milk and incubated with membranes for 1 hour at room temperature. After extensive washing with TBST, the membranes were incubated with an appropriate HRP-conjugated secondary antibody (Pierce; goat anti-rabbit for the PARP-1 polyclonal antibody and ADPR binding reagents, and goat anti-mouse for the 10H monoclonal antibody) diluted in 1% non-fat dry milk for 1 hour at room temperature. The membranes were washed extensively with TBST before chemiluminescent detection using SuperSignal™ West Pico substrate (Thermo Scientific) and a ChemiDoc system (Bio-Rad).

Blocking or Removal of ADP-ribose Prior to Immunoblotting

Blocking with free ADP-ribose.

Immunoblotting was performed as described above, except that the ADPR detection reagents were diluted in TBST and pre-incubated with 10 mM adenosine 5’-diphosphoribose sodium salt (ADP-ribose, Sigma) or 10 mM NaCl for 1 hour at room temperature prior to incubation with the membranes. Detection was performed as described above.

Removal of ADP-ribose with Hydroxylamine.

Immunoblotting was performed as described above, except that the membranes were incubated with or without 1 M NH2OH in TBST containing 2% non-fat dry milk at room temperature for 8 hours. After extensive washing with TBST, blotting with the ADP-ribose detection reagents and subsequent signal detection were performed as described above.

Removal of ADP-ribose with ARH3.

Thirty μg of whole cell extract from HeLa S3 cells treated with 2 mM H₂O₂ were incubated with 1 μM ARH3 in a 50 μL reaction under the following conditions: 50 mM potassium phosphate (pH 7.5), 10 mM MgCl₂, 5 mM DTT at 37 °C for 2 hours. The reaction was stopped by the addition of one third of a reaction volume of 4x SDS-PAGE loading buffer, followed by heating to 100°C for 5 min. The samples were then subjected to immunoblotting as described above.

Immunofluorescent Staining of Cells

HeLa S3 cells were grown on sterile glass coverslips in a 12-well plate, treated for 1 hour with or without 20 μM of PJ34, and then treated for an additional 5 minutes with or without 2 mM H₂O₂. The cells were washed twice with PBS at room temperature and fixed with an ice cold 7:3 solution of methanol:acetone (v/v) at 4°C for 15 minutes. The fixed cells on the cover slips were washed with ice cold PBS and blocked with 5% non-fat milk in PBS containing 0.05% Tween-20 (PBST) for 30 minutes at room temperature. After blocking, the samples were incubated overnight at 4°C with ADPR binding reagent diluted 1:100 (10 ng/μL) in 1% non-fat milk in PBST. The samples were washed with PBS and incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG (Thermo Fisher) diluted 1:500 in 1% non-fat milk in PBST for 30 minutes at room temperature in the dark. The samples were then washed with PBS, and the genomic DNA was stained with 1 μM TO-PRO-3 (Thermo Fisher) in PBS for 2 minutes at room temperature. The samples were washed with PBS, the coverslips were mounted onto glass slides, and the staining was visualized using a Leica confocal microscope.

Assaying the Binding of ADPR Binding Reagents by Protein A-Agarose

One μg of ADPR binding reagent or rabbit IgG (Pierce) was diluted in 200 μL of TBST and immobilized on 25 μL (packed bed volume) of protein A-agarose resin (Millipore) by incubation for 2 hours with gentle mixing at 4°C. The resin was washed three times with 1 mL of ice cold TBST to remove unbound ADPR binding reagent or rabbit IgG. The bound proteins were eluted by boiling the protein A-agarose resin for 5 minutes in two bed volumes of 2x SDS Loading Buffer. Inputs and eluates were analyzed on 10% PAGE-SDS acrylamide gels stained with Coomassie brilliant blue.

Immunoprecipitation of ADP-Ribosylated Proteins Using the ADPR Binding Reagents

293T cells were grown to ~80–90% confluence and then treated 5 minutes with 2 mM H 2O2. The cells were washed and collected in ice cold PBS, and pelleted by centrifugation at 800 RCF in a benchtop centrifuge for 5 minutes at 4°C. The packed cell volume was noted. The cells were resuspended to homogeneity in 1 packed cell volume of IP Lysis Buffer (25 mM Tris•HCl pH 7.5, 450 mM NaCl, 0.3% NP-40, 2 mM MgCl₂, 1 mM DTT, 2.5 μM APD-HPD, 1x protease inhibitor cocktail) and incubated for 15 minutes on ice. The samples were centrifuged for 10 minutes at maximum speed in a microcentrifuge at 4°C. The supernatant containing cellular proteins was collected and diluted 3-fold in IP Dilution Buffer (25 mM Tris•HCl pH 7.5, 2.5 μM APD-HPD, 1x protease inhibitor cocktail) to yield cellular proteins in IP Binding Buffer (25 mM Tris•HCl pH 7.5, 150 mM NaCl, 0.1% NP-40, 0.67 mM MgCl₂, 0.33 mM DTT, 2.5 μM APD-HPD, 1x protease inhibitor cocktail).

ADP-ribosylated proteins from whole cell lysates in IP Binding Buffer were applied to protein A-agarose resin pre-bound with an ADPR binding reagent or rabbit IgG (as described above) and immunoprecipitated by incubation with gentle mixing overnight at 4°C. The resin with bound proteins were then washed three times with 1 mL of ice cold TBST. The immunoprecipitated ADP-ribosylated proteins were eluted from the protein A-agarose resin by boiling the resin in two bed volumes of 2x SDS-PAGE Loading Buffer for 5 minutes. The elute was subjected to immunoblotting.

Results

Expression and Purification of Recombinant ARBD-Fc Fusion Proteins

Natural ARBDs can be used as tools for the molecular recognition of various forms of ADPR (i.e., MAR, OAR, PAR) when functionalized with different polypeptide “tags”.^{22,24,29,33–36} We reasoned that functionalizing ARBDs with the Fc region of rabbit IgG (Figure 1C) would produce antibody-like proteins that could bind ADPR, with the specificity for the different forms of ADPR determined by the particular the ARBD used. With this in mind, we generated cDNAs encoding N-terminally 10xHis- and Strep[II]-tagged fusion proteins with an ARBD located N-terminal to the Fc region of rabbit IgG (Figure 1D). For the ARBDs, we used the WWE domain from H. sapiens RNF146^31,32,48,49, or the macrodomains from A. fulgidus AF1521^27,50, H. sapiens macroH2A1.1^28,29, or H. sapiens PARP-14⁵¹ The latter has three macrodomains in tandem, and we initially included sequences encoding all three in the fusion cDNAs. We then cloned the ARBD-Fc fusion cDNAs into the bacterial expression vector pET19b (Novagen) (Figure 1D).

We expressed the ARBD-Fc fusion proteins in E. coli and then purified them using nickel-NTA affinity chromatography. When analyzed by denaturing SDS-PAGE and Coomassie Brilliant Blue staining, the fusion proteins ran at the expected molecular sizes (Figure 2; compare the top and bottom panels). The preparations of WWE(RNF146)-Fc, Macro(AF1521)-Fc, and Macro(mH2A1.1)-Fc were consistently >90 percent pure (Figure 2, top). In contrast, most preparations of Macro3X(PARP14)-Fc had two major contaminating bands (Figure 2, top), which did not affect the utility of the protein in ADPR detection (see below). The purified proteins have expected specificities for ADPR as follows: (1) WWE(RNF146)-Fc: OAR and PAR; (2) Macro(AF 1521)-Fc: MAR, OAR, and PAR; (3) Macro(mH2A1.1)-Fc: MAR, OAR, and PAR; and (4) Macro3X(PARP14)-Fc: MAR.^27–32

Figure 2. Expression and purification of ADP-ribose binding domain-Fc (ARBD-Fc) fusion proteins.

ARBD-Fc fusion proteins were expressed in E. coli and purified using Ni-NTA affinity chromatography. (Top) The purified proteins were separated by SDS-PAGE and stained with Coomassie brilliant blue. The arrows indicate protein bands with the expected molecular weight of the ARBD-Fc fusion proteins. The asterisks indicate breakdown products or contaminating E. coli proteins, which do not alter the functionality of the ARBD-Fc fusion protein. (Bottom) List of four ARBD-Fc fusion proteins and their expected molecular weights in kilodaltons (kDa).

Macro3X(PARP14)-Fc had a lower yield compared to the other reagents, thus we developed a modified version of the PARP14 macrodomain-based reagent to produce larger amounts. In the modified construct, we included macrodomains 2 and 3 of PARP14 only, excluding macrodomain 1, yielding Macro2/3(PARP14)-Fc. In this regard, a previous study indicated that macrodomains 2 and 3 are the critical ADPR readers in PARP-14.⁵¹ We expressed Macro2/3(PARP14)-Fc in E. coli and purified it as described above. The yields were improved and the ratio of full-length protein to lower molecular weight contaminants was increased compared to Macro3X(PARP14)-Fc (Figure S1A). As described below, both Macro3X(PARP14)-Fc and Macro2/3(PARP14)-Fc behaved similarly in all assays and were used interchangeably as indicated.

Generation of Protein-linked Mono-, Oligo-, and Poly(ADP-ribose) Standards

To test the ability of the ARBD-Fc fusion proteins to bind to and recognize specific forms of ADPR, we required a source of MAR, OAR, and PAR. To generate different forms of ADPR, we used purified recombinant PARP-1 (which can automodify with OAR and PAR) and PARP-3 (which can automodify with MAR) in biochemical reactions with NAD⁺. To stimulate the catalytic activity of PARP-1 and PARP-3, we added sonicated salmon sperm DNA. We controlled the extent of automodification of PARP-1 by altering the concentration of NAD⁺ in the reactions and the time of incubation (i.e., 3 μM NAD⁺ for 30 min. for OAR and 250 μM NAD⁺ for 5 min. for PAR) (Figure 3A). Reactions lacking NAD⁺ were used as a control. For an initial confirmation that the PARP-1 reactions yielded the expected products, we performed an immunoblot analysis of the reactions with the 10H anti-PAR monoclonal antibody, which recognizes PAR, but not MAR or OAR (Figure 3B). As expected, the reaction in which PAR was generated showed a signal at the expected size of PARylated PARP-1 (Figure 3B, lane 5), while the other reactions and controls showed no signal. The PAR chains generated in this way are on average about 35 ADPR units in length (Figure S2, right). In contrast, the OAR chains are on average about 10 ADPR units in length (Figure S2, left).

Figure 3. Immunoblot and dot blot analyses of mono-, oligo-, and poly(ADP-ribosyl)ated PARP proteins using ARBD-Fc fusion proteins.

(A) Purified recombinant PARP-1 and PARP-3 were incubated with or without NAD⁺ under different reaction conditions to promote auto(ADP-ribosyl)ation. The yield was mono(ADP-ribosyl)ated PARP-3 (red), oligo(ADP-ribosyl)ated PARP-1 (green), or poly(ADP-ribosyl)ated PARP-1 (blue).

(B through F) Immunoblot analyses. The mono, oligo, and poly(ADP-ribosyl)ated PARP proteins were separated by SDS-PAGE, transferred to nitrocellulose, and subjected to immunoblotting using the 10H anti-PAR monoclonal antibody, as well as four ARBD-Fc fusion proteins, as labeled. Each lane contained approximately the same number of terminal ADP-ribose units, to our best approximation. The molecular weights in kilodaltons (kDa) are indicated.

(G) Dot blot analyses. Purified recombinant PARP-1 and PARP-3 were incubated with or without NAD⁺ under different reaction conditions to promote auto(ADP-ribosyl)ation. The yield was mono(ADP-ribosyl)ated PARP-3 (red), oligo(ADP-ribosyl)ated PARP-1 (green), or poly(ADP-ribosyl)ated PARP-1 (blue). Serial dilutions of the auto(ADP-ribosyl)ated PARP proteins were applied to nitrocellulose membranes for dot blotting using the 10H anti-PAR monoclonal antibody, as well as four ARBD-Fc fusion proteins, as labeled. Each spot contained approximately the same number of terminal ADP-ribose units, to our best approximation.

Immunoblotting and Dot Blotting with Recombinant ARBD-Fc Fusion Proteins

To explore in more detail the specificity of the various ARBD-Fc fusion proteins for different forms of ADPR, we performed immunoblot analyses of automodification reactions containing PARP-3-MAR, PARP-1-OAR, and PARP-1-PAR, as well as control reactions lacking NAD⁺ (Figure 3, C–F, Figure S1B). The results with most of the reagents were as expected, based on the previously reported specificity of the ARBDs^27–32, with one exception (Table 1). We observed that (1) the WWE domain from H. sapiens RNF146 recognized OAR and PAR, but not MAR (Figure 3C); (2) the macrodomain from A. fulgidus AF1521 recognized MAR, OAR, and PAR (Figure 3D); (3) the macrodomain from H. sapiens macroH2A1.1 recognized MAR and PAR, but not OAR (Figure 3E); and (4) the macrodomains from H. sapiens PARP-14 recognized MAR, but not OAR or PAR (Figure 3F). Similar results were obtained in dot blot assays using the same sources of MAR, OAR, and PAR (Figure 3G). The recognition of MAR and PAR, but not OAR, by the macroH2A1.1 macrodomain was paradoxical, because we expected macrodomains to recognize any terminal ADPR moiety (i.e., a single protein-linked ADPR moiety or the terminal ADPR moiety of a protein-linked OAR or PAR chain) (Figure 1, A and B). This observation may provide information about the nature of H. sapiens macroH2A1.1 or OAR (discussed in more detail below).

Table 1.

Summary of the selectivity of the 10H anti-PAR monoclonal antibody, as well as four ARBD-Fc fusion proteins, for mono-, oligo-, and poly(ADP-ribose).

Detection Reagent	Mono (MAR)	Oligo (OAR)a	Poly (PAR)b
Anti-PAR (10H)	−	−	+
WWE(RNF146)-Fc	−	+	+
Macro(AF1521)-Fc	+	+	+
Macro(mH2A1.l)-Fc	+	−	+
Macro3X(PARP14)-Fc	+	−	−
Macro2/3(PARP14)-Fc	+	−	−

^aThe OAR chains generated for use in the assays described in this study had an average size of about 10 ADPR units in length (Figure S2, left).

^bThe PAR chains generated for use in the assays described in this study had an average size of about 35 ADPR units in length (Figure S2, right).

Testing the ARBD-Fc Fusion Proteins in Cell-based Assays

Antibodies are useful tools for biochemical and molecular assays performed using cells, such as immunoblotting of cellular extracts, immunofluorescent staining of intact cells, and immunoprecipitations. With this in mind, we tested the performance of the ARBD-Fc fusion proteins in a series of cell-based immunological assays.

Immunoblotting.

First, we tested the ADPR-binding reagents in immunoblotting assays using nuclear extracts prepared from HeLa and MCF-7 cells under basal growth conditions (Figure 4A). All of the reagents yielded robust signals in the assays, and each reagent produced a unique pattern of detection that differed from the other reagents and between the cell types (Figure 4A). We also tested the reagents in whole cell extracts from Hela cells treated with 2 mM hydrogen peroxide (H₂O₂; a DNA damaging agent that activates the catalytic activity of PARP-1 and possibly other nuclear PARPs), 20 μM PJ34 (PARPi; a broad-spectrum inhibitor of nuclear PARPs^54,55), or both, H₂O₂ and PJ34 combined (Figure 4B). These treatments allowed us to increase (i.e., H₂O₂) or decrease (i.e., PJ34) the levels of ADP-ribosylation in predictable ways to determine the effect on detection by the reagents. The signals from all four ARBD-Fc reagents were dramatically increased in response to treatment with H₂O₂, an effect that was blocked by co-treatment with PJ34 (Figure 4B). Interestingly, we observed inhibition by PJ34 of the H₂O₂- stimulated signal obtained using the PARP-14-based reagent, a PARPi that is thought to target primarily the nuclear polyenzymes PARPs 1 and 2⁵⁴ (Figure 4B, lower right). This may be due to the detection of initial protein-proximal PARP-1- and PARP-2-mediated mono(ADP-ribosyl)ation events, or perhaps PARP-3 mono(ADP-ribosyl)ation events that are modulated in some way by PARPs 1 and 2.

Figure 4. Immunoblot blot analysis of ADP-ribosylation in HeLa and MCF-7 cells using ARBD-Fc fusion proteins.

(A) Nuclear extracts were prepared from HeLa cells (left) and MCF-7 cells (right) maintained under standard culture conditions. Aliquots of the nuclear extracts containing equal total protein levels were separated by SDS-PAGE, transferred to nitrocellulose, and subjected to immunoblotting using the ARBD-Fc fusion proteins as indicated above each lane.

(B) Whole cell extracts were prepared from HeLa cells grown in culture following treatment with or without 2 mM H₂O₂ and 20 μM PARP inhibitor PJ34 (PARPi), as indicated. Aliquots of the whole cell extracts containing equal total protein levels were separated by SDS-PAGE, transferred to nitrocellulose, and subjected to immunoblotting using the ARBD-Fc fusion proteins as indicated.

To test the specificity of the signals detected in the immunoblotting assays with extracts from H₂O₂-treated HeLa cells, we used three different approaches. First, we blocked the ADPR-binding reagents using free ADPR prior to incubation with the membrane for immunoblotting. This eliminated the signal for all three MAR-binding reagents, but did not affect the signal with WWE(RNF146)-Fc, which binds to the linkage between ADPR units in OAR and PAR (Figure 5A). Second, we treated the membranes with hydroxylamine (NH₂OH), which cleaves ADPR from glutamate and aspartate residues⁵⁶, prior to immunoblotting. This eliminated the signal for all four of the ADPR-binding reagents (Figure 5B), but had no appreciable effect on signals detected by an acetyl-lysine antibody or a dimethyl-lysine antibody (Figure S3). The results with hydroxylamine treatment suggest that the bulk of the signal in H₂O₂-treated HeLa cells is from ADP-ribosylated glutamate or aspartate residues. Finally, we incubated the extracts with purified ARH3 (Figure S4), a glycohydrolase that cleaves PAR chains, as well as the proximal ADPR units from proteins^57–59. Again, this eliminated the signal for all four of the ADPR-binding reagents (Figure 5C). A previous report suggested that ARH3 is capable of cleaving the proximal ADPR from Ser residues, but not from Glu, Arg, or Lys residues.⁵⁷ However, given that HeLa cells are known to have a complex ADP-ribosylated Asp and Glu proteome²⁴, and the fact that hydroxylamine also eliminated the signal, our results suggest that ARH3 is capable of cleaving ADPR from Asp and Glu under the conditions used here. Collectively, our results indicate that the immunoblotting signals that we detect with the ADPR-binding reagents are dependent on the presence and recognition of ADPR.

Figure 5. Blocking or removing ADP-ribose eliminates detection by ARBD-Fc fusion proteins in immunoblotting assays using cell extracts.

Whole cell extracts were prepared from HeLa cells treated with 2 mM H₂O₂. Aliquots of the extracts containing equal total protein levels were separated by SDS-PAGE, transferred to nitrocellulose, and subjected to immunoblotting using the ARBD-Fc fusion proteins as indicated above each lane.

(A) Immunoblotting after blocking the ADPR-binding reagents with 10 mM free ADPR prior to incubation with the membrane

(B) Immunoblotting after treating the membranes with 1 M hydroxylamine, which cleaves ADPR from glutamate and aspartate residues⁵⁶, prior to immunoblotting.

(C) Immunoblotting after incubating the extracts with 1 μM purified ARH3, a glycohydrolase that cleaves PAR chains as well as the proximal ADPR units from proteins^57–59, prior to immunoblotting.

Immunofluorescent staining.

Next, we tested the ADPR-binding reagents in immunofluorescent cell staining assays using HeLa cells treated with 2 mM H₂O₂ or 20 μM PJ34 (Figure 6). The cells were stained using a conventional immunofluorescent staining protocol using the reagents with methanol fixation and co-staining for DNA using TO-PRO-3. Staining for ADPR was evident in the basal (vehicle-treated) condition with WWE(RNF146)-Fc, Macro(mH2A1.1)-Fc, and Macro3X(PARP14)-Fc, but not Macro(AF1521)-Fc (Figure 6; see the enlarged images in Figure 6B). In all cases where staining was observed in the basal condition, it was reduced by treatment with PJ34, and the background signal with all four reagents was low (Figure 6). The staining in the basal condition was primarily nuclear, although cytoplasmic ‘speckles’ were evident with WWE(RNF146)-Fc and perinuclear cytoplasmic staining was evident with Macro3X(PARP14)-Fc (Figure 6B). Interestingly, staining in the basal condition was excluded from the nucleoli in many cells stained with WWE(RNF146)-Fc and Macro(mH2A1.1)-Fc (Figure 6B; see the yellow arrows). With all four reagents, treatment with H₂O₂ increased the amount of staining, as well as pattern of staining to some degree (Figure 6B; see also Figure 4B). In particular, the exclusion of nucleolar staining observed with WWE(RNF146)-Fc and Macro(mH2A1.1)-Fc in the basal condition was reduced by H₂O₂ treatment (Figure 6B). Together, the results from the immunofluorescent cell staining assays demonstrate that (1) ARBD-Fc fusion proteins work well for detecting ADPR in this assay and (2) cells produce different forms of ADPR in different subcellular compartments and under different conditions.

Figure 6. Immunofluorescent staining of ADP-ribosylation in HeLa cells using ARBD-Fc fusion proteins.

HeLa cells grown in culture on glass cover slips were treated with vehicle, H₂O₂, or PJ34 (PARPi), as indicated. Following treatment, the cells were fixed with methanol, stained with TO-PRO-3 (a DNA stain), and immunostained for ADP-ribose using the ARBD-Fc fusion proteins, as indicated. The cover slips were affixed to glass slides, and the cells were imaged for fluorescence by laser scanning confocal microscopy.

(A) Fluorescence images for all conditions and all ADPR-Fc fusion proteins, as indicated. The ADPR (green), DNA (red), or merged images are indicated. Scale bar = 50 μm (same for all images).

(B) Nine-fold magnification of selected conditions from (A). Exclusion of staining from the nucleoli is indicated by the yellow arrows. Scale bar = 10 μm (same for all images).

Immunoprecipitation.

Finally, we tested the ADPR-binding reagents in immunoprecipitation assays using nuclear extracts from 293T cells treated with H₂O₂ (Figure 7). Many immunological assays are based on molecular interactions with the Fc region of the immunoglobulin (antibody) that lends specificity to the assay. For example, immunoprecipitation is typically performed using bacterial immunoglobulin binding proteins, such as Staphylococcus sp. Protein A or Protein G, which bind to the Fc region. As a first step in exploring the utility of the ARBD-Fc fusion proteins in immunoprecipitation assays, we performed a simple Protein A binding assay to determine if the ARBD-FC fusions could be recovered quantitatively from binding reactions. We incubated equivalent amounts of IgG heavy chain or one of the four ARBD-Fc fusion proteins with Protein A-agarose resin. After washing the resin, we eluted the IgG or ARBD-Fc fusion proteins and analyzed the eluate by SDS-PAGE (Figure 7A). We observed efficient binding in all cases, and the extent of binding for the ARBD-Fc fusion proteins was as good, if not better, than what was observed with IgG heavy chain (Figure 7A).

Figure 7. Immunoprecipitation of ADP-ribosylated PARP-1 from 293T cells using ARBDFc fusion proteins.

(A) Protein A-agarose binding assay with ADPR-Fc fusion proteins. The ADPR-Fc fusion proteins indicated were incubated with protein A-agarose. The efficiency of binding was assessed by SDS-PAGE analyses of the eluted material, with subsequent staining using Coomassie brilliant blue.

(B) 293T cells were treated with vehicle or H₂O₂, as indicated. Nuclear extracts prepared from the treated cells were subjected to immunoprecipitation using the indicated ADPR-Fc fusion proteins (the lane numbers above the gel correspond to the number labels for the ADPR-Fc fusion proteins shown in panel A). The immunoprecipitated material was separated by SDS-PAGE, transferred to nitrocellulose, and subjected to immunoblotting using an anti-PARP-1 antibody. The molecular weights in kilodaltons (kDa) are indicated.

After determining that the ARBD-Fc fusion proteins are bound efficiently by Protein A, we assayed their ability to immunoprecipitate automodified PARP-1 from nuclear extracts prepared from 293T cells with or without H₂O₂ treatment (Figure 7B). Automodified PARP-1 from these cells contains a mixture of MAR, OAR, and PAR modifications (Figure 7B). We observed that all of the reagents were able to immunoprecipitate automodified PARP-1 from the nuclear extract, and the patterns of recognition were largely as expected (Figure 7B). For example, WWE(RNF146)-Fc immunoprecipitated OAR- and PAR-modified PARP-1 efficiently (Figure 7B). The WWE domain that it contains binds the iso-ADPR linkage between ADPR monomers in PAR (Figure 1A), thus providing many binding sites per each molecule of automodified PARP-1. In contrast, Macro(AF1521)-Fc immunoprecipitated OAR- and PAR-modified PARP-1 less efficiently (Figure 7B). The macrodomain that it contains binds the terminal ADPR unit in PAR (Figure 1A), thus providing a limited number of binding sites per each molecule of automodified PARP- 1. Macro(mH2A1.1)-Fc and Macro2/3(PARP14)-Fc both immunoprecipitated MARylated PARP-1 efficiently (Figure 7B). The extent to which ARBD interactions with MAR are modulated by the linkage to the modified amino acid, or by nearby amino acids in the modified protein, remains an open question. Unexpectedly, Macro(mH2A1.1)-Fc, which recognized both MAR and PAR in the biochemical assays (Figure 3), was unable to immunoprecipitate OAR- and PAR-modified PARP-1 (Figure 7B). Taken together, these results demonstrate that this collection of ARBD-Fc fusion proteins can be used to immunoprecipitate MAR-, OAR-, and PAR-modified proteins.

Discussion

Biochemical and cellular studies of the diverse forms and functions of ADPR require immunological reagents that can be used for detection and enrichment. Prior to this report, the state-of-the-art has been the anti-PAR monoclonal antibody 10H, which is thought to bind to PAR with a lower limit for the length of polymers detected of around 10 ADPR units.^45,46 Thus, 10H cannot detect MAR or OAR, limiting its utility and leaving the field blind to mono- and oligo(ADP-ribosyl)ation events. Herein, we have described the generation and characterization of recombinant antibody-like ADP-ribose binding proteins, in which natural ARBDs have been functionalized with the Fc region of rabbit immunoglobulin. Specific recognition of the diverse forms of ADPR by these reagents comes from the particular ARBD included in each fusion protein (e.g., various macrodomains and a WWE domain) (Table 1). Importantly, the set of ARBD-Fc fusion proteins that we have generated includes those that can recognize MAR and OAR, as well as one that can recognize all forms of ADPR (Table 1). As we have demonstrated herein and in recently published studies^22,24, these reagents are useful in a broad array of antibody-based assays, such as immunoblotting, immunofluorescent staining of cells, and immunoprecipitation.

Functionalization of Naturally-Occurring ARBDs as Research Tools

Our ARBD-Fc fusion proteins represent one example of the use of naturally-occurring ARBDs as tools for exploring the chemistry, biochemistry, and biology of ADP-ribose. Other examples include: (1) ARBD-GFP (green fluorescent protein) fusion proteins, which allow real­time tracking of localized MARylation and PARylation events in cells.^29,34,35 and (2) ARBD-GST (glutathione-S-transferase) fusion proteins, which can be used for enrichment in biochemical and molecular assays. With respect to the latter, the Af1521 macrodomain, which recognizes MAR and the terminal ADP-ribose of PAR, has been fused with a GST tag and used to enrich for ADP-ribosylated targets in genomic and proteomic screens.^33,36 Similarly, the RNF146 WWE domain has been fused with a GST tag and used to determine the genomic localization of PAR.³³ While our ARBD-Fc fusion proteins are not the first examples of functionalization of natural ADRBs for use as research tools, they are the first examples to include all of the key features of a monoclonal antibody. These include (1) monospecificity (i.e., monoclonality), (2) binding to protein A and G, (3) binding to Ig-directed secondary antibodies, and (4) renewable production. The Ig Fc region is the target of many antibody-directed tools, reagents, and methods, underlying the utility of the ARBD-Fc fusion proteins in a wide range of immunological assays.

Recent reports have indicated that some macrodomains possess intrinsic ADP-ribosylhydrolase activity.^60–65 In theory, such an activity might limit the utility of macrodomains as reagents for the detection of ADPR. In this regard, we note that only one of the three macrodomains that we tested herein, namely the macrodomain from AF1521, has ADP-ribosylhydrolase activity, while the macrodomains from PARP-14 and macroH2A1.1 do not.^61,63 Furthermore, the ADP-ribosylhydrolase activity reported for macrodomains, like AF1521, is generally weak.^60–65 or is undetectable under the conditions used in the assays described herein. In our hands, we have not found this to be a significant issue with the Macro(AF1521)-Fc reagent. If it becomes an issue for future macrodomain-based reagents, the activity can be blocked by site-specific mutation of the macrodomain, without loss of ADPR binding.^60–65

Using ARBD-Fc Fusion Proteins to Uncover the Nature of Different Forms of ADPR

While not a major focus of this work describing the generation and characterization of a set ARBD-Fc fusion proteins, our work has revealed some interesting aspects of the chemistry, biochemistry, and biology of ADP-ribose. For example, our results showing a dramatic loss of signal upon hydroxylamine treatment when using the ARBD-Fc fusion proteins in immunoblotting assays with cell extracts (Figure 5B) suggest that most of the ADP-ribosylation in HeLa cells occurs on glutamate and aspartate residues (the chemistry of ADP-ribosylation is such that hydroxylamine is only expected to cleave ADPR from these residues⁵⁶). Results from previous studies have suggested that ADP-ribosylation can occur on a variety of residues (e.g., Glu, Asp, Lys, Ser, Arg) and that the proportion of the amino acids that are modified may vary by cell or tissue type.^13,36,56 This approach using hydroxylamine treatment coupled with immunoblotting may be a useful way to get an initial assessment of the relative levels of glutamate and aspartate ADP-ribosylation in biological samples prior to more detailed analyses.

Furthermore, from previous studies, we know that macrodomains recognize MAR, as well as the terminal ADPR moieties in OAR and PAR, allowing them to bind to all three forms of ADPR (i.e., MAR, OAR, and PAR) (Figure 1, A and B).^27–29 In contrast, WWE domains recognize the iso-ADPR linkages joining ADPR monomers, restricting their binding to OAR and PAR (Figure 1, A and B).^30–32 The fact that the different macrodomains that we tested exhibited different specificities for MAR, OAR, and PAR (Table 1) suggest that they can recognize additional and distinct features of these molecules. The binding of MAR and PAR, but not OAR, by the macroH2A1.1 macrodomain is perplexing (Figure 3, E and G; Table 1), but it might suggest that binding by this macrodomain could be inhibited by branch points near the terminal ends of ADPR chains in OAR. Determining if this is indeed the case will require further tests. The binding of MAR, but not OAR or PAR, by the PARP-14 macrodomains (Figs. 3, F and G; Table 1), may indicate that (1) additional interactions occur between the PARP-14 macrodomains and ADP-ribosylated proteins or amino acids, or (2) the binding of the PARP-14 macrodomains to the terminal ADP-ribose units of poly(ADP-ribose) chains is inhibited by the penultimate ADP-ribose units. The extent to which ARBD interactions with MAR are modulated by the linkage to the modified amino acid or nearby amino acids in the modified protein remains an open question. Collectively, such observations suggest that the biology of MAR, OAR, and PAR is more diverse, rich, and complex than previously thought. In this regard, the immunofluorescent cell staining assays shown in Figure 6 shown a diverse array of staining patterns with the different ARBD-Fc fusion proteins, suggesting different biologies for ADPR in different subcellular compartments. The ARBD-Fc fusion proteins described herein will be useful tools for future exploration of the chemistry, biochemistry, and biology of ADP-ribose.

Abbreviations:

ADPR

ADP-ribose

ARBD

ADP-ribose binding domain

Fc

Immunoglobulin constant region

IgG

Immunoglobulin G

MAR

Mono(ADP-ribose)

OAR

Oligo(ADP-ribose)

PAR

Poly(ADP-ribose)

PARP

Poly(ADP-ribose) polymerase