Viral ADP-ribosyltransferases attach RNA chains to host proteins

The mechanisms by which viruses hijack their host’s genetic machinery are of enormous current interest. One mechanism is adenosine diphosphate (ADP) ribosylation, where ADPribosyltransferases (ARTs) transfer an ADP-ribose fragment from the ubiquitous coenzyme nicotinamide adenine dinucleotide (NAD) to acceptor proteins . When bacteriophage T4 infects Escherichia coli, three different ARTs reprogram the host’s transcriptional and translational apparatus . Recently, NAD was identified as a 5’-modification of cellular RNA molecules in bacteria and higher organisms . Here, we report that bacteriophage T4 ARTs accept not only NAD, but also NAD-RNA as substrate, thereby covalently linking entire RNA chains to acceptor proteins in an “RNAylation” reaction. One of these ARTs, ModB, efficiently RNAylates its host protein target, ribosomal protein S1, at arginine residues and strongly prefers NAD-RNA over NAD. Mutation of a single arginine at position 139 abolishes ADP-ribosylation and RNAylation. Overexpression of mammalian ADP-ribosylarginine hydrolase 1 (ARH1), which cleaves arginine-phosphoribose bonds, shows a decelerated lysis of E. coli when infected with T4. Our findings not only challenge the established views of the phage replication cycle, but also reveal a distinct biological role of NADRNA, namely activation of the RNA for enzymatic transfer. Our work exemplifies the first direct connection between RNA modification and post-translational protein modification. As ARTs play important roles in different viral infections, as well as in antiviral defence by the host , RNAylation may have far-reaching implications.

RNA (site-specifically 32 P-labelled) and tested it as substrate. No radioactive band appeared (Extended Data Fig. 3c), providing no support for spontaneous ADP-ribosylation.

ModB modifies specific arginines in rS1
To identify the amino acid residues in protein rS1 to which RNA chains are covalently linked during RNAylation, we took advantage of tools developed to analyse protein ADP-ribosylation. The radioactive signal of RNAylated protein rS1 (as prepared in Fig. 2b) did not change upon treatment with HgCl2 (which cleaves S-glycosides resulting from Cys), NH2OH (which hydrolyses O-glycosides) (Extended Data Fig. 4a) and recombinant enzyme ARH3 (which hydrolyses O-ADPr glycosides specifically at serine residues) (Extended Data Fig. 4b), while it was efficiently removed by treatment with human ARH1 24 (Fig. 3a-d). These findings indicate that the major product(s) of the ModBcatalysed RNAylation reaction are linked as N-glycosides via arginine residues (as shown in Fig. 3a,b). To identify the amino acid residues which are targeted by ModB, in vitro modified rS1 was subjected to tryptic digest, chromatographic purification, and mass-spectrometric analysis. This LC/MS/MS analysis revealed three specific modification sites in rS1, namely R19, R139, and R426 (Extended Data Fig. 5).
To establish the biological significance of RNAylation by T4 ARTs in vivo, we isolated endogenous (untagged) protein rS1 from non-infected and T4-infected E. coli, respectively. E. coli contains significant amounts of endogenous NAD-RNAs 4,6 . Ribosomes were isolated, and rS1 was pulled down by poly-U-sepharose and subjected to LC/MS/MS analysis (Fig. 3e). This experiment confirmed the in vitro data and revealed the same three sites, namely R19, R139 and R426, at which phosphoribose modifications were abundant only in the T4-infected sample. (Fig. 3f). Site-directed mutagenesis further confirmed the modified residues: R139K and R139A mutants of protein rS1 were expressed in T4-infected E. coli, purified and analysed, revealing that these mutations abolish the modification (Extended Data Fig. 6).

Detection of RNAylation in vivo
The mass spectrometric pipeline detected ADP-ribosylation and RNAylation in the same way, namely as ribose-5'-phosphate or ADPr fragment. To distinguish between the two modifications, we considered an immunoblotting assay with an antibody-like ADP-ribose binding reagent ("pan-ADPr"). The specificity of pan-ADPr was investigated by Western blotting with in vitro-prepared ADPribosylated or RNAylated proteins, respectively (Extended Data Fig. 7a). As expected, rS1-ADPr and ModB-ADPr were both recognised by pan-ADPr and produced bands with high intensities, while no signal was observed for rS1-RNA, suggesting that pan-ADPr does not tolerate 3'-extensions of the ribose moiety. However, when rS1-RNA was digested with nuclease P1 prior to pan-ADPr treatment, thereby degrading the RNA and likely leaving rS1-ADPr, a strong signal, comparable to authentic rS1-ADPr, appeared in the blot (Fig. 4a). We applied this immunoblotting assay to investigate ADP-ribosylation and RNAylation in vivo. We expressed a plasmid-borne copy of rS1 in non-infected or T4-infected E. coli. Subsequently, rS1 was affinity-purified and its ADP-ribosylation analysed by pan-ADPr blotting (Extended Data Fig. 7b). In agreement with our mass-spectrometric data, this experiment revealed extensive ADP-ribosylation of rS1 only in the T4-infected sample. After nuclease P1-treatment, the pan-ADPr signal intensity of the rS1 band increased significantly (Fig. 4b), indicating that ~30 % of the modified rS1 was RNAylated in vivo (measured as the difference between P1-treated and nuclease untreated sample). Moreover, the signal for ADP-ribose disappeared upon ARH1 treatment, again confirming the nature of the RNAprotein linkage (Extended data Fig. 7b).
. CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted June 4, 2021. ; https://doi.org/10.1101/2021.06.04.446905 doi: bioRxiv preprint A recognition motif for ModB How ModB identifies its targets remains a puzzle. Target protein rS1 contains oligonucleotide-binding (OB) domains 22 . One structural variant of OB folds is the S1 domain, present in rS1 in six copies that vary in sequence (Extended Data Fig. 8a,b). We speculated that the S1 domain might be important for substrate recognition by ModB. To characterise ModB's specificity for different rS1 domains, we individually cloned, expressed and purified each S1 domain of rS1 and applied them in our RNAylation assay (Fig. 4 c,d and Extended Data Fig. 8c,d). For rS1 domains D2 and D6 we determined high RNAylation signals. In comparison, rS1 D1, D3, D4 and D5 domains were modified to a much lesser extent. Alignment of D2 and D6 of rS1, and the S1 domain of PNPase, another protein in E. coli that possess an S1 domain, revealed that these S1 motifs share an arginine residue as part of the loop connecting strands 3 and 4 of the -barrel 25 (Extended Data Fig. 8b). This loop is packed on the top of the β-barrel, thereby likely accessible for ModB. For rS1 D2, this particular residue is R139, which we had shown to be modified by mass spectrometry (Fig. 3f). Mutation analysis confirmed that the ADPribosylation level of D2 is dramatically reduced if R139 is substituted by alanine or lysine (Extended Data Fig.9). Based on these findings, we screened for other E. coli proteins that harbour an S1 domain with an arginine in the loop between strands 3 and 4, and identified RNase E. In our in vitro assays, RNase E, which carries the S1 motif in its active site, was efficiently modified by ModB, while control proteins without S1 domain (BSA, NudC inactive quadruple mutant) were not, supporting the identification of the subgroup of S1 domains with an embedded arginine as the RNAylation target motif (Figure 4e,f).

Modification and T4 replication cycle
rS1 is an important RNA-binding protein required for the translation of virtually all cellular mRNAs in E. coli. To investigate the biological consequences of rS1 modification by ModB, we analysed rS1 levels during T4 infection using an E. coli strain that contains a chromosomal fusion of rS1 with a FLAG-tag ( Fig. 4g and Extended Data Fig. 10a). Immediately after infection, rS1 levels dropped steeply, whereas they increased moderately over 20 min in the absence of T4. We thus speculated that ADP-ribosylation and/or RNAylation might influence the stability of rS1. To test this hypothesis, we overexpressed human ARH1 in E. coli during T4 infection, thought to remove ADP-ribose and linked RNA. As a control, we overexpressed a largely inactive ARH1 D55,56A mutant. Indeed, with active ARH1, the ADPribosylation signal was dramatically reduced (Extended Data Fig. 10b), while the mutant showed a pattern similar to the parent strain (Extended Data Fig. 10a). Using these constructed E. coli strains, we analysed the influence of ADP-ribosylation and RNAylation on rS1 levels during phage infection. Indeed, the strain expressing active ARH1 showed an increase in rS1 levels over time, like the uninfected sample, whereas the mutant strain exhibited declining levels, like the T4-infected sample without ARH1 (Extended Data Fig. 10b,c). Thus, the removal of ADPr and RNA chains during phage infection coincides with a stabilisation of the rS1 level. To investigate if these modifications are important for the lysogenic behaviour of the phage, we infected E. coli strains expressing either ARH1 or its inactive mutant with T4 and monitored the optical density over time (Fig. 4i). In the inactive mutant strain, bacterial lysis started 50 min post-infection, while delayed lysis (120 min) was observed when active ARH1 was overexpressed (Fig. 4i). Collectively, these data indicate that ADP-ribosylation and/or RNAylation interfere with protein stability and modulate the course and efficiency of T4 infection.
. CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

DISCUSSION
To date, all interactions between RNAs and proteins are described to be based on non-covalent interactions 26 . In contrast, we show here that ADP-ribosyltransferases can attach NAD-capped RNAs to target proteins in a covalent fashion. This finding represents a distinct biological function of the NAD-cap on RNAs in bacteria, namely activation of the RNA for enzymatic transfer to an acceptor protein. We discovered that RNAylation of target proteins, a novel post-translational protein modification, plays a role in the infection of the bacterium E. coli by bacteriophage T4. Our data indicate that T4 ART ModB modifies proteins that possess an S1 RNA binding domain. We identified specific arginine residues to be modified, thereby increasing molecular weight and negative charge of the target protein and undoubtedly causing major changes of the properties and functions of the modified proteins. The post-translational modification of crucial players in bacterial translation and transcription demonstrates the importance of the known ADP-ribosylation and the newly discovered RNAylation reaction for bacteriophage pathogenicity. Introduction of the human ADP ribosylhydrolase ARH1, which removes these modifications, into E. coli, caused a significant delay in bacterial lysis upon phage infection. Why do phage ARTs attach RNAs to proteins involved in translation? One possibility may be that these RNAs help (e.g., by base pairing) to preferentially recruit mRNAs encoding for phage proteins to the ribosomes and thereby guarantee their biosynthesis. Likewise, the observation that RNase E, the major player in RNA turnover in E. coli, is RNAylated at its catalytic centre by ModB may suggest that the T4 phage, after reprogramming transcription by Alt and ModA, shuts down RNA degradation in the host to ensure a long half-life of phage mRNAs. We are working vigorously on methods for identifying the RNAs attached to target proteins, which will allow the elucidation of their biochemical mechanisms. ARTs are known to occur not only in bacteriophages, and ADP-ribosylated proteins have been detected in hosts upon infections by various viruses, including influenza, corona, and HIV. In addition to viruses using ARTs as weapons, the mammalian antiviral defence system applies host ARTs to inactivate viral proteins. Moreover, mammalian ARTs and poly-(ADP-ribose) polymerases (PARPs) are regulators of critical cellular pathways and are known to interact with RNA 27 . Thus, ARTs in different organisms might catalyse RNAylation reactions, and RNAylation may be a phenomenon of broad biological relevance. Finally, RNAylation may be considered as both a post-translational protein modification and a posttranscriptional RNA modification. Our findings challenge the established views of how RNAs and proteins can interact with each other. The discovery of these new RNA-protein conjugates comes at a time when the structural and functional boundaries between the different classes of biopolymers become increasingly blurry 28,29 . . CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is . CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is  Here, the mechanism of ADPribosylation is shown exemplarily for arginine. Initially, the N-glycosidic bond between the ribose and nicotinamide is destabilised by a glutamate residue of an ART. This leads to the formation of an oxocarbenium ion of ADP-ribose. Nicotinamide serves as the leaving group. This electrophilic ion is attacked by a nucleophilic arginine residue of the acceptor protein after glutamate-mediated proton abstraction. This leads to the formation of an N-glycosidic bond 30 . b. Analogous to ADP-ribosylation in the presence of NAD, we propose that ARTs might use NAD-RNA to catalyse an "RNAylation" reaction, thereby covalently attaching an RNA to an acceptor protein.

REFERENCES
. CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is

General
Reagents were purchased from Sigma-Aldrich and used without further purification. Oligonucleotides were purchased from Integrated DNA Technologies, Inc. DNA and RNA concentrations were determined by measurements with the NanoDrop ND-1000 spectrophotometer. Radioactively labelled proteins or nucleic acids were visualised using storage phosphor screens (GE Healthcare) and a Typhoon 9400 imager (GE Healthcare).

Preparation of 5'PPP-/5'P-/5'-NAD-RNA by in vitro transcription
DNA template for Qβ-RNA was amplified by PCR. PCR products were analysed by 1 % agarose gel electrophoresis and purified using the QIAquick PCR purification kit (QIAGEN).  Table 1.
. CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted June 4, 2021. ; https://doi.org/10.1101/2021.06.04.446905 doi: bioRxiv preprint

Cloning of ADP-ribosyltransferases, ADP-ribose hydrolases and target proteins
To amplify bacteriophage T4 genes modA, modB and alt, a single plaque from T4 bacteriophage revitalisation was resuspended in Millipore water and used in a "plaque"-PCR, analogous to bacterial colony PCR. The gene encoding for the ADP-ribosylhydrolase ARH1 was purchased from IDT as gblocks and amplified by PCR. E. coli genes coding for rS1 and PNPase were PCR-amplified from genomic DNA of E. coli K12, which was isolated via GenElute Bacterial Genomic DNA Kit (Sigma-Aldrich). Nucleotide sequences are listed in Extended Data Table 2. XhoI and NcoI restriction sites were introduced during amplification using appropriate primers (Extended Data Table 3). The resulting PCR product was digested with XhoI and NcoI and cloned into pET-28c vector (Merck Millipore). After Sanger sequencing, the resulting plasmids were transformed into E. coli One Shot BL21 (DE3) (Life Technologies). The ARH1 D55,56A and rS1 mutants were generated by site-directed mutagenesis using a procedure based on the Phusion Site-Directed Mutagenesis Kit (Thermo Scientific). The resulting plasmids were sequenced and transformed into E. coli One Shot BL21 (DE3). AHR1, ARH1 D55A D56A and rS1 were transferred from pET28 to pTAC-vector using specific primers with restriction sites for XhoI and SphI (Extended Data Table 3). The resulting PCR product was digested with XhoI and SphI and cloned into pTAC-MAT-2 (Sigma-Aldrich). The resulting plasmids were transformed into E. coli strain B and E. coli Ced64 (kind gift from Gerhart E. Wagner, Uppsala University). All strains used and generated in this work are summarised in Extended Data Table 4 Purification of rS1 domains and variants, PNPase Domain, RNase E (1-529), Alt, NudC V157A, E174A, E177A, E178A IPTG-induced E. coli One Shot BL21 (DE3) containing the respective plasmid were cultured in LB medium at 37 °C. Protein expression was induced at OD600 = 0.8, bacteria were harvested by centrifugation after 3 hours at 37°C and lysed by sonication (30 s, 50 % power, five times) in HisTrap buffer A (50 mM Tris-HCl pH 7.8, 1 M NaCl, 1 M Urea, 5 mM MgSO4, 5 mM -mercaptoethanol, 5 % glycerol, 5 mM imidazole, 1 tablet per 500 ml complete EDTA-free protease inhibitor cocktail (Roche)). The lysate was clarified by centrifugation (37,500 g, 30 min, 4 °C) and the supernatant was applied to a 1 mL Ni-NTA HisTrap column (GE Healthcare). The protein was eluted with an imidazole gradient using an analogous gradient of HisTrap buffer B (HisTrap buffer A with 500 mM imidazole) and analysed by SDS-PAGE. Further protein purification was achieved by size exclusion chromatography (SEC) through a Superdex TM 200 10/300 GL column (GE Healthcare) using a buffer containing 0.5 M NaCl and 25 mM Tris-HCl, pH 8. Fractions of interest were analysed by SDS-PAGE, pooled and concentrated in Amicon Ultra-4 centrifugal filters (MWCO 10 kDa, centrifugation at 2,000 rpm., 4 °C). Protein concentration was measured with the NanoDrop ND-1000 Spectrophotometer. Proteins were finally stored in buffer supplemented with 50 % glycerol at -20 °C.
Purification of ARH1 and ARH1 D55,56A E. coli BL21 DE3 pET28-ARH1 and BL21-pET28-ARH1 D55A D56A were grown to an OD600 = 0.6 at 37 °C, 175 rpm. Afterwards, bacteria were allowed to cool to room temperature for 30 minutes. Expression was induced with 1 mM IPTG, and bacteria were finally grown overnight at room temperature, 175 rpm. Bacteria were harvested by centrifugation, and proteins were purified analogously to rS1 variants.
Purification of ModB E. coli BL21 DE3 pET28-ModB were grown to an OD600 = 2.0 at 37 °C, 185 rpm and cooled down to 4 °C while shaking at 160 rpm for at least 30 min. Protein expression was induced by the addition of 1 mM IPTG. The cultures were then incubated for 120 min at 4 °C, 160 rpm and bacteria harvested by centrifugation (4,000 rpm, 4 °C, 25 min). The ModB protein was purified from the supernatant as described for rS1 variants.

Kinetics of RNAylation in the presence of high NAD concentrations
Kinetic experiments for RNAylation in the presence of different NAD concentrations were performed using conditions for RNAylation and ADP-ribosylation as mentioned above. The kinetics were performed in 0.15 μM 32 P-NAD-RNA 8-mer with either 0.125 μM, 12.5 μM, 125 μM or 1.25 mM NAD corresponding to an approximate molar excess of NAD over 32 P-NAD-RNA 8-mer of 1-, 100-, 1,000-and 10,000-fold. The reactions temperatures were pre-adjusted to 15 °C. 5 μL samples were taken after 0 (before addition of ModB), 1, 5, 10, 30, 60 and 120 minutes and mixed with 5 μl 2x Laemmli buffer to stop the reaction. Samples were analysed by 12 % SDS-PAGE analysis. 10 μL of a reference in 1x Laemmli buffer were applied to each SDS-polyacrylamide gel containing 1.91 pmol of protein rS1 . CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted June 4, 2021. ; https://doi.org/10.1101/2021.06.04.446905 doi: bioRxiv preprint modified by 6.74 pmol ModB in 1x transferase buffer and 0.15 μM 32 P-NAD-RNA 8-mer. Intensities of radioactive bands were quantified using ImageQuant 5.2. Relative RNAylation values were computed by the division of the individual intensity values by the intensity of the reference.

Investigation of self-RNAylation of ribosomal protein rS1
In scales of 20 μL reactions, 3.6 µM of 32 P-ADPr-8mer were either incubated with 2.6 µM of protein rS1, 3.9 µM of ModB or both 2.59 µM of protein rS1 and 3.89 µM of ModB in 1x transferase buffer. As a positive control, equal amounts of protein rS1 and ModB were incubated with 0.6 µM of 32 P-NAD-8mer. All reactions were incubated at 15 °C for 60 minutes. Samples were taken before the addition of ModB (0 minutes), after 60 minutes of incubation and stopped by the addition of 1 volume 2x Laemmli buffer each. The reactions were analysed by 12 % SDS-PAGE and autoradiography imaging. Preparation of RNAylated and ADP-ribosylated rS1 for enzymatic treatments ADP-ribosylation or RNAylation reactions performed with radio-labelled substrates were washed and equilibrated in 1x transferase or 1x degradation buffer for further enzymatic treatment. Therefore, the reactions were washed with 4 column volumes of the corresponding buffer via centrifugation at 10,000 x g, 4 °C in 10 kDa Amicon filters.
Nuclease P1 digest of protein rS1 RNAylated with Qβ-RNA (rS1-Q-RNA) 19 μL of rS1/rS1-Qβ-RNA ( 32 P) mixture equilibrated in 1x transferase buffer were incubated with either 1 μL of nuclease P1 or 1 μL of Millipore water at 37 °C for 60 minutes. Samples were taken before the addition of enzyme (0 minutes) and after 60 minutes of incubation and reactions stopped by addition of 1 volume 2x Laemmli buffer. Reactions were analysed by 10 % SDS-PAGE using rS1-ADPr ( 32 P) as a reference and autoradiography imaging.
His-tagged-protein rS1 from T4-infected or uninfected E. coli B strain pTAC-rS1 was washed with two filter volumes of 1x degradation buffer (12.5 mM Tris-HCl, pH 7.5, 25 mM NaCl, 25 mM KCl, 5 mM MgCl2) by centrifugation in 10 kDa Amicon Ultra-4 centrifugal filters at 5,000 g, 4 °C and concentrated to a final volume of 120 μL. The fractions were analysed by 12 % SDS-PAGE analysis, and the gel was stained in Instant Blue solution for 10 min and imaged immediately.

Purification of endogenous rS1 from infected E. coli strain B
For isolation of endogenous protein rS1, ribosomes were first isolated as described in 7 . Bacterial pellets of T4 infected and non-infected E. coli B strain were resuspended in resuspension buffer (20 mM MgCl2, 200 mM NH4Cl, 20 mM Tris-HCl pH 7.5, 0.1 mM EDTA, 6 mM -mercaptoethanol) and cells were lysed by sonication. The lysate was cleared by ultracentrifugation at 30,000 x g, 4 °C, 30 min. 15 mL of cleared supernatant were loaded onto 1 volume sucrose cushion (20 mM MgCl2, 200 mM NH4Cl, 20 mM Tris-HCl pH 7.5, 0.1 mM EDTA, 6 mM -mercaptoethanol, 30 % (w/v) sucrose) in glass ultracentrifuge tubes, which were centrifuged at 67,000 x g, 18 °C for 24 hours. The supernatant was discarded, the ribosomal pellets washed with rS1 wash buffer (20 mM Mg(OAc)2, 1 M NH4Cl, 10 mM Tris-HCl pH 7.6, 7 mM -mercaptoethanol) carefully and resuspended in 4 mL of rS1 wash buffer. Endogenous rS1 was then purified from resuspended ribosomes using self-synthesised poly(U)-Sepharose 8 . Synthesis of poly(U)-Sepharose was performed according to 9 with slight modifications. 700 mg of cyan bromide-activated Sepharose beads were resuspended in 30 mL 0.1 M NaCl, 2 mM Tris-HCl pH 7.5 and incubated at 4 °C for 2 hours under rotation. Beads were pelleted by centrifugation at 15,567 x g for 3 minutes and the supernatant removed. Beads were then resuspended in 30 mL of 0.2 M 4-morpholine-ethanesulfonic acid (MES) and pelleted again to discard the supernatant. For coupling of poly-uridine (poly(U)) material to the beads, 10 mg of poly(U) material (Santa Cruz) resuspended in 5 mL of 0.2 M MES pH 6.0 were added to the equilibrated, activated Sepharose beads and the suspension incubated at 4 °C overnight. The resulting poly(U)-Sepharose was equilibrated in rS1 wash buffer by washing with 4 polypropylene column volumes (10 mL). 6 mL of poly(U)-Sepharose suspension were obtained for isolation of endogenous rS1 from ribosomes. To specifically isolate rS1, 3 mL of the beads were mixed with 4 mL of the isolated ribosomes and incubated at 4 °C for 30 minutes. The bead-protein mixture was subjected to gravity affinity purification with 10 mL polypropylene columns. Ribosomes were washed with 4 column volumes of rS1 wash buffer and endogenous rS1 eluted with 10 mL rS1 elution buffer (20 mM Mg(OAc)2, 1 M NH4Cl, 10 mM Tris-HCl pH 7.6, 7 mM mercaptoethanol, 7 M urea). Afterwards, beads were re-equilibrated in rS1 wash buffer and the washing and elution steps were repeated once. Eluates of the same sample were pooled, concentrated and analysed by SDS-PAGE.

LC-MS/MS
To identify ADPr-modification sites on rS1, 100 µg of in vitro or in vivo modified-rS1 were digested with trypsin (sequencing grade modified, Promega) overnight at a 1:20 (w/w) ratio at 37 °C. The digest was stopped by adding 0.5 % trifluoroacetic acid and 5 % acetonitrile (v/v), followed by desalting using acommercially available pre-packed C18 column for acidic RP C18 desalting (Harvard Apparatus, Microspin C18 Column) and enriched using TiO2 as described elsewhere for protein-RNA cross-link enrichment 10 . Next, the sample was subjected to LC-MS analysis using a Dionex UltiMate 3000 UHPLC + focused system (Thermo Scientific) with a C18 analytical column (75 μm × 300 mm, ReproSil-Pur 120 C18-AQ, 1.9 μm, Dr. Maisch GmbH, packed in house). Peptides were separated by RP C18 chromatography on a 58 min multi-step gradient (flow rate 0.3-0.4 µl min −1 ). Eluting peptides were . CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted June 4, 2021. ; https://doi.org/10.1101/2021.06.04.446905 doi: bioRxiv preprint analysed with an orbritrap mass spectrometer (QExactive HF-X or Orbitrap Lumos Fusion, Thermo Scientific) with the following settings: MS1 spectra in profile mode (resolution of 120k), MS2 spectra in centroid mode (resolution of 30k); isolation window set to 1.6 m/z and dynamic exclusion set to 7 s. Raw data were searched in an open search using PD2.2 (SEQUEST search engine) to identify possible adduct masses that could correspond to RNA or NAD. Precursor mass tolerance was set to 700 Da and fragment mass tolerance was set to 0.02 Da. We only identified ribose-5-phosphate as a shorter peptide modification reliably with an adduct mass of mz = 212.0086 and validated results in stringent searches with precursor mass tolerance set to 7 ppm and fragment mass tolerance set to 10 ppm. Identified peptides were filtered on PSM level to an FDR of 1 %.

MALDI-TOF
To identify larger adduct masses than ribose-5-phosphate that escaped LC-MS detection, we performed MALDI-TOF experiments. Modified rS1 peptides were prepared as described above at basic pH. Formic acid was substituted with 10 mM ammonium hydroxide in all RP-C18 buffers. Eluted sample was mixed with 2,5-dihydroxy benzoic acid (DHB) and then spotted onto an MTP stainless steel plate in a double layer fashion. MS data was acquired on a Bruker Autoflex Speed MALDI-TOF instrument in positive reflectron mode for survey scans or LIFT mode for pseudo MS2. The instrument was calibrated against BSA peptides and then recalibrated against wt rS1 peptides to 20 ppm mass error. Record spectra were baseline corrected but not smoothened.

Western Blotting
Proteins were separated by 10 % SDS-PAGE and gels were equilibrated in transfer buffer (25 mM Tris, pH 8.3, 192 mM glycine, 20 % [v/v] methanol). 0.2 μm polyvinylidene difluoride (PVDF) membranes (GE Healthcare) were activated in methanol for 1 min and equilibrated in transfer buffer. Proteins were transferred from gels to PVDF membranes in a semi-dry manner at 300 mA for 1.5 h, if not indicated differently. After the transfer, membranes were dehydrated by soaking in methanol and washed 2x with TBS-Tween (TBS-T; 10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05 % [v/v] Tween® 20). Afterwards 10 mL blocking buffer (5 % [w/v] milk powder in TBS-T) were added to the membranes and incubated at room temperature for 1 h. In order to detect ADP-ribosylated proteins, membranes were incubated with 0.1 ng/mL primary antibody MABE1016 (Merck) in 10 mL washing buffer (1 % [w/v] milk powder in TBS-T) at 4 °C overnight 11 . Membranes were washed and incubated with 10 mL of a 1:10,000 dilution of the horseradish-peroxidase-(HRP)-goat-anti-rabbit-IgG secondary antibody (Advansta) in washing buffer at room temperature for 1 h. Afterwards, membranes were washed with PBS. Proteins with detected modifications were visualised by chemiluminescence using the SignalFire ECL Reagent or the SignalFire Elite ECL Reagent (Cell Signaling Technology) according to the manufacturer's instructions. If proteins in SDS-PAGE gels needed to be visualised before blotting, a 2,2,2-trichloroethanol (TCE) staining method 12 was used. Resolving gels were supplemented with 0.5 % (v/v) TCE. For visualisation, gels were activated by ultraviolet transillumination (300 nm) for 60 s. Proteins then showed fluorescence in the visible spectrum.
Quantification of RNAylation rS1 proteins were isolated from E. coli strain B pTAC rS1 infected or non-infected with bacteriophage T4. 1.5 µM rS1 were digested with 1 µM ARH1 in the presence of 12.5 mM Tris-HCl pH 7.5, 25 mM NaCl, 25 mM KCl and 5 mM MgCl2. 1.5 µM rS1 were digested with 0.5 U endonuclease P1 in 100 mM Mg(OAc)2, 220 mM NH4Cl, 500 mM HEPES pH 7.5, 10 mM EDTA, 100 mM β-mercaptoethanol and 10 % glycerol. Digests were incubated at 37 °C for 2 hours. Afterwards, digests were ethanol-precipitated by . CC-BY-NC-ND 4.0 International license made available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is The copyright holder for this preprint this version posted June 4, 2021. ; https://doi.org/10.1101/2021.06.04.446905 doi: bioRxiv preprint the addition of 9 volumes of ethanol and precipitated by centrifugation (14,000 rpm) at 4 °C for 1 h. Protein pellets were resuspended in 10 μL 1x Laemmli buffer and analysed via Western blot. ADPrmodifications were detected by the primary antibody MABE1016 (Merck) as described above.

Reprobing of PVDF membranes with sodium azide
Following the visualisation of the chemiluminescence signal, the PVDF membranes were washed with Dulbecco's PBS. Subsequently, the membranes were incubated in blocking solution (5 % [w/v] milk powder in TBS-T) with 1 mM sodium azide for 3 hours at 4°C. Afterwards, the membranes were washed in TBS-T 6 times, for 5 minutes each. Inactivation of HRP was visualised by chemiluminescence using the SignalFireTM ECL Reagent or the SignalFireTM Elite ECL Reagent.

Detection of the FLAG-tagged rS1
PVDF membranes were incubated with a 1:10,000 dilution of the anti-FLAG-M2 antibody (Thermo Fisher Scientific) in a washing solution at 4 °C overnight. The membranes were washed 3 times with washing solution and incubated in a 1:10,000 dilution of the HRP-goat-anti-mouse-IgG (Advansta) for 1 h at room temperature. FLAG-tagged rS1 was visualised by chemiluminescence as described above.
T4 Phage infection of E. coli strain B and Ced64 with ARH1 WT and ARH D55,56A plasmids E. coli B strain or E. coli Ced64 containing the plasmid pTAC ARH1 WT or pTAC ARH1 D55,56A were inoculated in 40 mL LB medium supplemented with 100 µg/mL ampicillin at OD600 = ~ 0.1 and incubated at room temperature, 150 rpm. At OD600 = 0.3, the expression of ARH1 WT or ARH1 D55,56A was induced with 1 mM IPTG. At OD600 = 0.8, cultures were infected with T4 at a MOI 10. After 0, 5, 10, 30, 60, 90, 110, 140 and 200 minutes, the OD600 values of the cultures were measured. At the indicated time points, 100 μL aliquots were taken, and bacteriophage T4 infections stopped by the addition of 2x Laemmli buffer. The aliquots were analysed via Western Blotting.

Visualisation of ARH1 WT and ARH1 D55,56A via Western Blotting
The proteins were transferred onto a PVDF membrane as described above. Following the exact same treatment, the membranes were incubated in a 1:2,000 dilution of 6x-His Tag Monoclonal Antibody (HIS.H8) in 10 mL of washing solution. The following day, horseradish-peroxidase-(HRP)-goat-antimouse-IgG (Advansta, San Jose, USA) was used as a secondary antibody. His-tagged proteins were visualised by chemiluminescence as described above.