Sspabcd–Sspe Is a Phosphorothioation-Sensing Bacterial Defence System with Broad Anti-Phage Activities

Articles https://doi.org/10.1038/s41564-020-0700-6

SspABCD–SspE is a phosphorothioation-sensing bacterial defence system with broad anti-phage activities

Xiaolin Xiong1,2,5,7, Geng Wu3,7, Yue Wei1,7, Liqiong Liu1,3, Yubing Zhang1,3, Rui Su4, Xianyue Jiang1, Mengxue Li1, Haiyan Gao1,3, Xihao Tian1, Yizhou Zhang1,5, Li Hu1,5, Si Chen1, You Tang1,5, Susu Jiang1,5, Ruolin Huang1,5, Zhiqiang Li5, Yunfu Wang2, Zixin Deng1,3, Jiawei Wang4, Peter C. Dedon 6, Shi Chen 1,2,5 and Lianrong Wang 1,2,5 ✉

Bacteria have evolved diverse mechanisms to fend off predation by bacteriophages. We previously identified the Dnd system, which uses DndABCDE to insert sulfur into the DNA backbone as a double-stranded phosphorothioate (PT) modification, and DndFGH, a restriction component. Here, we describe an unusual SspABCD–SspE PT system in Vibrio cyclitrophicus, Escherichia coli and Streptomyces yokosukanensis, which has distinct genetic organization, biochemical functions and phenotypic behaviour. SspABCD confers single-stranded and high-frequency PTs with SspB acting as a nickase and possibly introducing nicks to facilitate sulfur incorporation. Strikingly, SspABCD coupled with SspE provides protection against phages in unusual ways: (1) SspE senses sequence-specific PTs by virtue of its PT-stimulated NTPase activity to exert its anti-phage activity, and (2) SspE inhibits phage propagation by introducing nicking damage to impair phage DNA replication. These results not only expand our knowledge about the diversity and functions of DNA PT modification but also enhance our understanding of the known arsenal of defence systems.

s a result of intense competition for environmental resources This function is attributable to a unique feature of PT modification and coevolution of bacteriophages and their hosts, pro- in which the replacement of an oxygen atom by sulfur endows the Akaryotes have developed a spectrum of diverse defensive PT-modified DNA backbone with resistance to nucleases15. With a systems. These include restriction-modification (R-M), toxin–anti- loss of the dnd modification genes, which leads to PT deficiency, toxin, abortive infection, prokaryotic Argonaute and the clustered the unrestrained activity of DndFGH causes toxic double-stranded regularly interspaced short palindromic repeats (CRISPR) and breaks of the DNA16,17. The dndABCDE–dndFGH system is thus CRISPR-associated gene systems1–5. The relatively recent discovery regarded as a prokaryotic innate defence system with functional of defence systems such as BREX6, DISARM7 and Zorya8 highlights similarity to methylation-based R-M systems—that is, it restricts the fact that our knowledge of the arsenal of anti-phage measures invading DNA lacking PT modifications18,19. residing in prokaryotic genomes is incomplete and that many cur- Methylation-based R-M systems have consensus sequences that rently unknown defence mechanisms probably await discovery. are typically fully modified to avoid the attack of resident DNA by As a form of prokaryotic epigenetics, R-M systems involve sequence- cognate restriction endonucleases. However, only 12% of the 40,701 specific DNA modification, traditionally a nucleobase methylation, 5′-GAAC-3′/5′-GTTC-3′ consensus sequences in the Escherichia and a restriction endonuclease to discriminate and destroy unla- coli B7A genome are protected with PTs despite the presence of active belled invading DNA9,10. Despite the diversity of the four major restriction by DndFGH. In contrast to the double-stranded DNA types of R-M systems, all associated DNA modifications—such as (dsDNA) PT in B7A, PT modification in Vibrio cyclitrophicus FF75 6 4 5 N -methyl-adenine, N -methyl-cytosine, C -methyl-cytosine and occur on only one strand at 5′-CPSCA-3′ (PS, phosphate–sulfur link- 7-deazaguanine derivatives—were believed to occur exclusively on age) motifs and only 14% of the 160,541 genome-wide 5′-CCA-3′ nucleobase moieties until our discovery of phosphorothioate (PT) sites are PT modified20. Moreover, PTs do not occur consistently at modifications of the DNA sugar-phosphate backbone, in which the a given site in a population of bacterial genomes of both B7A and non-bridging oxygen is replaced by sulfur11–13. FF75, reflecting PT heterogeneity and implying that unusual DNA- DndABCDE proteins catalyse the sulfur substitution to form target-selection mechanisms are used by the Dnd modifying and 20,21 PTs in a sequence-selective and RP stereospecific manner in diverse restricting components . In addition to being a single-stranded bacteria, whereas DndFGH proteins, located in the vicinity of dnd- DNA (ssDNA) PT modification, the PT levels in FF75 are three- to ABCDE, utilize PT to discriminate and restrict foreign DNA11,12,14. tenfold higher than those in B7A, Salmonella enterica serovar Cerro

1Ministry of Education Key Laboratory of Combinatorial Biosynthesis and Drug Discovery, School of Pharmaceutical Sciences, Wuhan University, Wuhan, China. 2Taihe Hospital, Hubei University of Medicine, Shiyan, China. 3State Key Laboratory of Microbial Metabolism, School of Life Sciences and Biotechnology, The Joint International Research Laboratory of Metabolic and Developmental Sciences, Shanghai Jiao Tong University, Shanghai, China. 4State Key Laboratory of Membrane Biology, Beijing Advanced Innovation Center for Structural Biology, School of Life Sciences, Tsinghua University, Beijing, China. 5Department of Neurosurgery, Zhongnan Hospital, Wuhan University, Wuhan, China. 6Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA. 7These authors contributed equally: Xiaolin Xiong, Geng Wu, Yue Wei. ✉e-mail: [email protected]

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

87, Hahella chejuensis KCTC2396 and Pseudomonas fluorescens sulfur mobilization pathway. Moreover, SspC possessed an ATPase pf0-1 (refs. 20,22,23). activity of 8.58 ± 3.12 U per mg of protein and feasibly drove PT Based on an exploration of the unusual PT behaviour in FF75, modification by providing energy in a manner similar to that pre- we started with a comparative genomic analysis and characterized dicted for DndD28 (Supplementary Fig. 3). a ssDNA PT modification system—which we termed ssp—with The gene sspB poses a different problem. It is predicted to encode genetic organization, biochemical functions and phenotypic behav- a 321-amino-acid protein of unknown function (DUF4007), which iour distinct from the dnd system. SspABCD and SspE constitute exhibits no sequence similarity to any Dnd proteins, suggesting its a defensive barrier against a diverse array of phages via anti-phage critical role in the ssDNA PT formation. However, the insolubil- activity, which depends on the sensing of sequence-specific PT ity of SspB from FF75 impeded the exploration of its function. motifs and the introduction of nicking damage to phage DNA, Instead, we determined the crystal structure of full-length SspB from highlighting an unusual mechanism of defence. Streptomyces clavuligerus ATCC 27064 to a resolution of 1.75 Å using the single-wavelength anomalous dispersion (SAD) method and a Results selenomethionine (SeMet) derivative (Supplementary Table 1). The Discovery of the Ssp PT modification system. Despite mass- sspABCD homologues in ATCC 27064 are closely clustered and con- spectrometric evidence for PTs in V. cyclitrophicus FF75 (PT-linked fer PT-modified d(CPSC) in a Streptomyces lividans HXY6 strain lack- 3 d(CPSC) dinucleotides, 2.6 per 1 × 10 nucleotides), a BLAST search ing endogenous ssp genes (Fig. 1b and Supplementary Fig. 4). The of the FF75 genome for dnd gene homologues proved negative. We SspB structure contained an unprecedented fold, with the main body extended this analysis to several related Vibrio species for which we largely consisting of an α-helical structure and three β-sheets on the had PT-level data22. Two other Vibrio species with PT levels similar periphery (Supplementary Results and Extended Data Fig. 1). to those of FF75 were also found to lack dnd genes: Vibrio breoganii A DALI search indicated that SspB exhibited similarities to 1C10, with 3.1 PT per 1 × 103 nucleotides, and V. breoganii ZF-29, the endonucleases FokI (Z-score = 6.8; matched residues, 8%) and with 2.2 PT per 1 × 103 nucleotides. In contrast, Vibrio tasmaniensis R.BSPD6I (Z-score = 5.5; matched residues, 9%), inspiring the per- 1F-267 possessed a dndBCDE gene cluster and a previous analysis formance of a nuclease assay. A time-course experiment established revealed fivefold fewer PTs in d(GPSA)/d(GPST) motifs in this organ- that SspB rapidly converted covalently closed circular DNA to an ism than in FF75 (0.6 PT per 1 × 103 nucleotides)22. These differ- open circular form, which was followed by a slower conversion to ences between 1F-267 and the other three strains led us to perform linear DNA in the presence of Mg2+ or Mn2+ (Fig. 2a). Mg2+ ions a comparative analysis of their genomes on the RAST server24. This were indeed found bound directly to E18, and the E18R single-point analysis yielded homologues of three clustered genes—M565_ mutation remarkably impaired the nicking activity of SspB (Fig. 2b ctg1P1907 to M565_ctg1P1909—in FF75, 1C10 and ZF-29 but not and Extended Data Fig. 2). These findings implied that SspB acted in 1F-267 (Fig. 1a,b). The three genes, which bore no sequence as a nickase capable of introducing two or more nicks into pUC19 similarity to DndB, DndD or DndE, were co-transcribed as a single because dsDNA breaks only when two nicking sites on opposite operon in FF75, implying functional linkage (Supplementary Fig. 1). strands are close to each other. However, regular agarose gels cannot The M565_ctg1P1907 gene encoded a protein with unknown func- differentiate open circular plasmid DNA with one nick from that tion, whereas remote homology analysis with HHpred showed that with multiple nicks. Alkaline denaturing agarose gel electrophore- M565_ctg1P1908 had similarity to a domain denoted DUF499, sis was thus exploited to determine the nicking pattern on single a distant version of the AAA + ATPase domain25,26. Although strands. Although SspB-treated pUC19 was predominantly open M565_ctg1P1909 possessed a domain homologous to phospho- circular and linear DNA on a native gel, electrophoresis of DNA adenosine phosphosulfate reductase, it shared no significant under denaturing conditions revealed fragments of varying lengths sequence similarity to DndC in 1F-267 (Fig. 1b). (Fig. 2a). An identical DNA smear still occurred when the DNA We performed an in-frame deletion of M565_ctg1P1907 to substrates were devoid of 5′-CCA-3′ and 5′-CC-3′ motifs or con- M565_ctg1P1909 and the flanking DNA fragments to confirm the tained PT modifications, suggesting that SspB introduced massive involvement of M565_ctg1P1907 to M565_ctg1P1909 in d(CPSC) nicks onto both DNA strands in a non-sequence-specific manner modification in FF75 (Fig. 1c). Subsequent analysis using liquid (Supplementary Fig. 5). Mutations of the dimer-interface residues— chromatography-tandem quadrupole mass spectrometry (LC–MS/ including R23A, E204R, Y103A/E204R and F24R/Y103A—resulted MS) of DNA from the mutant strains confirmed that the individual in a simultaneous deficiency of the nicking activity of SspB in vitro deletion of M565_ctg1P1907 to M565_ctg1P1909 completely abol- and the d(CPSC) modification in vivo, indicating an essential role of ished d(CPSC) modification, as did the deletion of an upstream SspB-mediated DNA nicking before sulfur incorporation to yield gene, M565_ctg1P1902, which encodes a cysteine desulfurase sequence-specific single-stranded PT modifications (Fig. 2c–f and (Fig. 1b,c). The heterologous expression of M565_ctg1P1902 and Supplementary Results). M565_ctg1P1907 to M565_ctg1P1909 in E. coli resulted in the formation of d(CPSC) (Supplementary Fig. 2). Given the distinctive SspABCD and SspE constitute a defence barrier with broad genetic organizations between the dnd systems and the M565_ anti-phage activities. Located immediately downstream of ctg1P1902 and M565_ctg1P1907 to M565_ctg1P1909 genes, we con- sspBCD, the M565_ctg1P1910 gene—encoding the DUF262 and cluded that the latter group constitute an unusual PT modification DUF1524 domains—was found to be involved in growth defects system, which we termed ssp. Analogous to the dnd nomenclature, of PT-deficient Vibrio mutants at low temperatures by quanti- the ssp genes were designated sspA (M565_ctg1P1902) and sspBCD tative proteomic analysis and was designated sspE (Fig. 1 and (M565_ctg1P1907 to M565_ctg1P1909; Fig. 1c). Supplementary Results). The DUF262 domain is a member of the ParB-like superfamily; members of this superfamily (for example, The nicking activity of SspB is essential to ssDNA PT modifi- ParB, Osa, sulfiredoxin and DndB) have been identified in biologi- cation. Predicted to catalyse the initial step in PT modification, cally diverse contexts29–32. The DUF1524 domain possesses a con- DndA possesses cysteine desulfurase activity, which transfers sulfur served EHxxP motif, which is seen in protein groups in the His-Me from l-cysteine into the Fe–S cluster of DndC, an ATP pyrophos- finger endonuclease superfamily33,34. Given the growth defects phatase27. This relationship parallels the requirement for SspA and correlated with SspE and the positioning of sspE relative to ssp- SspD, which exhibited obvious cysteine desulfurase and ATP pyro- ABCD, we speculated that SspABCD–SspE might constitute a sin- phosphatase activities, respectively (Supplementary Fig. 3). This gle-stranded PT-based defence system. To test this, we challenged finding suggested that the ssp and dnd systems share the same initial E. coli DH10B cells that harboured plasmids pWHU732 and

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

a PAPS reductase

ParB-like superfamily AAA domain

V. tasmaniensis 1F-267 dndB dndC dndD dndE dndH dndG dndF d(GPSA)/d(GPST)

M565_ctg1P1910 b M565_ctg1P1908 DUF262 (ParB-like superfamily) AAA + ATPase DUF1524 M565_ctg1P1902 M565_ctg1P1907 M565_ctg1P1909 cysteine desulfurase DUF4007 PAPS reductase

V. cyclitrophicus FF75 d(CPSC)

71% 77% 72% 89%

V. breoganii 1C10 d(CPSC) 71% 77% 72% 89%

V. breoganii ZF-29 d(CPSC) 26% 23% 23% 34%

S. clavuligerus ATCC 27064 d(CPSC)

27% 23% 23% 34% 24%

S. yokosukanensis DSM 40224 d(CPSC)

26% 31% 45% 19%

E. coli 3234/A d(CPSC)

c d(CPSC) d(CPSC) FF75 sspA sspB sspC sspD sspE +

NH2 ΔsspA – N ΔsspB –

– HO N O ΔsspC O NH ΔsspD – 2 – S O N XXL-6 + P O N O ΔsspE + O O

ΔsspBCDE – OH

Fig. 1 | Organization of the ssp PT system in the genome of V. cyclitrophicus FF75. a, Typical dndBCDE–dndFGH system in V. tasmaniensis 1F-267. The dndA gene is located either adjacent to dndBCDE or elsewhere in the genome, which is consistent with the function of DndA as a cysteine desulfurase that is functionally replaced by an IscS orthologue. b, Comparison of ssp PT system in different species. This DNA region includes the genes M565_ctg1P1902

(sspA) and M565_ctg1P1907 to M565_ctg1P1910 (sspBCDE), whose protein products confer ssDNA PT modifications in d(CPSC). The genes coloured in grey are not conserved in the ssp systems, which is consistent with the observation that their deletion (mutant XXL-6) did not alter the level of d(CPSC) modifications. Homologous genes were aligned for comparison and the numbers between the dotted lines indicate the percentage of amino-acid sequence identity between the related Ssp proteins of FF75. Similar to dndA in dnd systems, sspA is not always located in the vicinity of other ssp genes—for example, the sspBCD–sspE module in E. coli 3234/A—suggesting similar functional substitution by a cysteine desulfurase orthologue elsewhere in the genome. c, Schematic of generated deletion mutants. +, confers DNA PT modification in d(CPSC); −, PT deficiency. The white blocks represent in-frame deletions in the corresponding genes. PT-linked d(CPSC) is shown in the structural inset (right). pWHU733 (which express SspABCD and SspE of V. cyclitrophicus phage (Fig. 3a). Although the cold induction of SspE caused a 10- to FF75, respectively) with seven coliphages spanning all three mor- 100-fold stronger protection against T4 and T7 phages at 15 °C than phological families of the Caudovirales—that is, the siphophages that observed at 28 °C (Supplementary Fig. 6), the overall moderate T1, JMPW1, JMPW2, T5 and EEP; the myophage T4; and the levels of phage resistance were presumably attributed to relatively podophage T7—at 28 °C (Supplementary Table 2). Following phage poor expression of Vibrio ssp genes in the heterologous E. coli host infection, DH10B(sspABCD, sspE) exhibited resistance against DH10B. E. coli DH10B cells harbouring the homologous sspBCDE six of the seven phages to varying degrees. Plaque assays showed cassette from E. coli 3234/A conferred 10- to 100-fold increased that SspABCD–SspE conferred approximately 1,000-fold protec- resistance against all phages tested (except for T5) at 28 °C and even tion against infection by T4 phage and 10- to 100-fold protection provided approximately 100-fold protection against the infection by against T1, JMPW1, JMPW2, EEP and T7 but no resistance to T5 single-stranded bacteriophage ϕX174 (Fig. 3b).

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

a SspB b

Uncut L Nt.BspQI0.2 0.5 1 3 6 9 12 16 Time (h) Divalent metal ions

OC Neutral 2+ 2+ 2+ 2+ 2+ 2+ L Uncut L Nt.BspQI EDTA Mg Mn Zn Ca Cu Ni

CCC OC L

CCC Denaturing

c d

Conservation SspB SspB R23 SspB SspB E105′ ′

Y103′ R203 Motif 1 E204′ F24 H28 ′ F24 H28′ E204 R203 Motif 2 Y103 ′

E105 Motif 3 R23 SspB′ ′

SspB′ f

d(C C) e PS Standard

E204R R23A Y103A/E204R F24R/Y103A Uncut L Nt.BspQI SspB SspB SspB SspB SspB S. lividans HXY6 OC L S. lividans HXY6 (SspABCD)

100 S. lividans HXY6 (SspAB CD) CCC R23A 75 S. lividans HXY6 (SspABE204RCD)

50 S. lividans HXY6 (SspABF24R/Y103ACD) 25 S. lividans HXY6 (SspABY103A/E204RCD)

Relative abundance (%)

5 10 Time (min) 45

Fig. 2 | Pseudodimerization of SspB and its effects on nickase activity and DNA PT modification. a, Time course for SspB-treated pUC19 in the presence 2+ 2+ of MgCl2. The DNA-nicking result was identical when Mg was replaced with 10 mM Mn . Following the nicking reaction, aliquots of the reaction mixture were withdrawn, denatured if required and resolved on 1% agarose gels under both neutral (top) and denaturing conditions (bottom). Nt.BspQI-treated pUC19 was used as a control. The results are representative of three independent experiments. b, Divalent cation requirement tests for the nicking activity of SspB. Results are representative of three independent experiments. c,d, Conserved motif 3 and the C-terminal half of motif 1 are involved in pseudodimerization of SspB. d, The two SspB protomers are coloured magenta and green. The nitrogen and oxygen atoms are coloured blue and red, respectively. The hydrogen bonds are shown as magenta dashed lines. e, Nicking activity assays for the wild-type and SspB variants E204R, R23A, Y103A/ E204R and F24R/Y103A on pUC19 DNA. Results are representative of two independent experiments. f, Mutations of these residues in the plasmid pWHU1811, which expressed the sspABCD cluster of S. clavuligerus ATCC 27064, strongly impaired d(CPSC) modification in S. lividans HXY6. Chemically synthesized d(CPSC) in the RP configuration was used as a reference. Data are representative of three independent experiments. CCC, covalently closed circular DNA; OC, open circular DNA; L, linear DNA.

Isolated phages showed increased tolerance to the SspABCD– mode of action of the Ssp system, we conducted Southern hybrid- SspE defence when they had been passaged through E. coli DH10B ization analyses on total DNA extracted from T4 phage-infected cells expressing SspABCD and SspE of FF75; this resistance was DH10B cells at various time points following infection at 28 °C. In subsequently abolished after the propagation of such phages on contrast to the extensive replication of T4 phage DNA in DH10B DH10B devoid of the ssp system, consistent with the classical R-M cells lacking sspBCD–sspE, its DNA replication was remarkably phenotype (Supplementary Fig. 7). To gain further insight into the inhibited in DH10B(sspABCD, sspE; Fig. 3c). Surprisingly, T4 DNA

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

a Ssp system of V. cyclitrophicus FF75

DH10B DH10B DH10B DH10B (empty vectors) (SspABCD, SspE) (SspE) (SspAC314SBCD, SspE)

10–1–10–6 10–1–10–6 10–1–10–6 10–1–10–6 T4 T4gt T1 JMPW1 JMPW2 T5 T7 EEP

DSM 13127 DSM 13127 (empty vectors) (SspABCD, SspE)

φX174

b Ssp system of E. coli 3234/A DH10B DH10B (empty vectors) (SspABCD, SspE)

–1 –6 10–1–10–6 10 –10 DSM 13127 DSM 13127 T4 (empty vectors) (SspBCD, SspE) T4gt –1 –6 T1 10–1–10–6 10 –10 JMPW1 φX174 JMPW2 T5 T7 EEP d

– + c SspABCDE(FF75)– SspABCDE(FF75)+ SspBCDE(3234/A) SspBCDE(3234/A)

φX174 ssDNARFI dsDNARFII dsDNA T4 DNA U 0 5 10 20 30 60 U 0 5 10 20 30 60 Time U 5 15 30 60 U 5 15 30 60 Time (min) (min)

Nicked dsDNA Phage T4 ssDNA Supercoiled dsDNA

Fig. 3 | The Ssp system provides protection against phages in a PT-dependent manner. a, Phage plaque assays on E. coli strains harbouring empty vectors, SspABCD or SspABCD–SspE from V. cyclitrophicus FF75 using 2 μl of each serial tenfold dilution (10−1–10−6) of the phage indicated on the left (top). No phage resistance was observed when an individual SspE was expressed in E. coli DH10B. A C314S point mutation on SspA not only abolished d(CPSC) PT modification but also impaired the anti-phage activity of the Ssp defence system. Phage ϕX174 was propagated in the specific host E. coli DSM 13127 (bottom). b, E. coli DH10B harbouring the Ssp system from E. coli 3234/A showed a higher level of resistance to phages than that of V. cyclitrophicus FF75. a,b, The plaque assays were carried out at 28 °C to compare the anti-phage activities of the Ssp systems from FF75 and 3234/A. c,d, Southern blot analyses of phages T4 and ϕX174 DNA during the infection of Ssp-containing and Ssp-lacking E. coli cells. Probes were designed to match positions 148,913–149,433 and 573–1,167, respectively, in the T4 and ϕX174 genomes. Each lane contains either 1.5 μg total DNA digested with PacI (c) or 1 μg undigested total DNA (d). U, uninfected. All images are representative of at least three independent experiments. seemed to not be cleaved or degraded with increasing infection time, PT-sensing anti-phage defence mechanism of the SspABCD–SspE even in the presence of SspABCD–SspE (Fig. 3c). These results system. DH10B cells harbouring a single sspE were still sensitive to indicate that SspABCD–SspE prevents phage propagation in a all phages tested, implying the dependence of SspE on SspABCD manner other than direct DNA cutting or degradation, point- proteins or PT modification (Fig. 3a). To address this possibility, ing to a mechanism of defence unlike those observed in typical we constructed a plasmid (pWHU3481) to express SspAC314SBCD methylation-based R-M systems or DndFGH-mediated Dnd sys- with a single cysteine (C314) replaced by a serine in SspA. The C314 tems. Moreover, the existence of Ssp systems in phylogenetically in SspA is equivalent to C327 in DndA from S. lividans, which is diverse bacteria highlights a diversity of previously unappreciated essential to nucleophilic attack on the l-cysteine substrate to initiate PT systems and expands our knowledge of the arsenal of anti- PT modification35. Although the single-point mutation C314S abol- phage defensive measures (Supplementary Results and Extended ished PT modification in DH10B(sspAC314SBCD, sspE), the struc- Data Fig. 3). tural integrity of SspABCD proteins was retained (Supplementary

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

Fig. 2). Consequently, SspAC341SBCD–SspE failed to provide protec- Ssp-containing cells. In combination with the delayed generation of tion against all phages tested, implicating unusual PT-dependent both RFI and RFII DNA in the presence of the Ssp system, it is plau- anti-phage activity (Fig. 3a). sible that SspE exerted anti-phage activity in the early stage of infec- We further noticed that SspE contained a DGQQR motif in the tion—for example, during complementary strand DNA synthesis. DUF262 domain. This motif was predicted to be an equivalent of the sulfiredoxin signature sequence F(G/S)GCHR36, a part of a Structural characteristics of SspE. To provide insights into the nucleotide-binding pocket, which prompted us to determine the molecular mechanism of defence by SspE, we endeavoured to NTPase activity. Due to the instability of SspE from FF75, we instead determine the crystal structure of SspE (Supplementary Results). expressed and purified SspE (locus_tag AQI95_RS39285) from Among the available SspE proteins we screened, the crystal struc- S. yokosukanensis DSM 40224 to homogeneity, thereby enabling us to ture of SspE from S. yokosukanensis DSM 40224 was determined. dissect its function. The sspABCD–sspE orthologues of DSM 40224 The DUF262 domain in SspE exhibited an unprecedented globu- on pWHU3658 conferred S. lividans HXY6 defence against JXY1, a lar shape, with nine α-helices as the main body and a sheet of five bacteriophage of the Podoviridae family (Fig. 4a and Supplementary β-strands at the side, whereas the DUF1524 domain was bi-lobed Fig. 8). In addition, the sspE of DSM 40224 on pWHU735 provided and consisted completely of α-helices (Fig. 5a). The linker between E. coli DH10B with resistance to T4 phage when coupled with the the DUF262 and DUF1524 domains was expected to be flexible, and sspABCD of FF75 on pWHU732 (Fig. 4a). A colorimetric inorganic there was no extensive packing between the two domains (Fig. 5a). phosphate (Pi) release assay revealed that SspE acted as an NTPase The DGQQR motif lay exactly at the junction of the third β-sheet capable of hydrolyzing GTP, CTP, UTP and ATP, with enzymatic and the third α-helix (Fig. 5b). activities ranging from 7.45 ± 1.00 to 24.95 ± 2.00 nmol Pi min−1 mg−1 Replacement of the glycine in the DGQQR motif with alanine (Fig. 4b). Interestingly, the hydrolysis activities towards GTP, CTP unexpectedly abolished the PT-sensing GTPase function in SspEG97A and UTP were remarkably stimulated following the addition of a but enhanced its nicking activity (Fig. 4c,d). However, the overall 40-bp DNA fragment that harboured the PT-modified 5′-CPSCA-3′ shape of the circular-dichroism spectrum of SspEG97A was essentially consensus20 (Fig. 4b and Supplementary Table 3). Conversely, DNA identical to that of wild-type SspE (Supplementary Fig. 9), implying with 5′-GPSAAC-3′, 5′-GPSTTC-3′ and 5′-GPSGCC-3′—the most that the altered nicking activity was not the result of a significant common PT motifs conferred by Dnd systems—exerted no such structural perturbation of SspEG97A and prompting us to analyse stimulatory effect, as exemplified by GTPase activity, suggesting the the enzymatic activities of separately expressed DUF262 and necessity of both PT structure and its sequence contexts (Fig. 4c DUF1524 domains. Interestingly, although the individual DUF262 and Supplementary Table 3). A single-point mutation at G97 in the domain completely lacked GTPase activity, the individual DUF1524 DGQQR motif resulted in completely impaired GTPase activity in domain still retained nicking activity at an enhanced level similar

SspEG97A, even in the presence of CPSCA-containing DNA; mean- to that of SspEG97A on T4 DNA (Fig. 4c,d). Considering the neces- while, the mutation of the G97 residue on plasmids pWHU735 and sity of both NTPase and nicking endonuclease for SspE-mediated pWHU3658 abolished protection from infection by phages T4 and anti-phage activity, it is thus conceivable that the two domains JXY1, respectively (Fig. 4a). These data highlight the critical role of coordinate to ensure the overall structure, enzymatic activi- PT-sensing NTPase activity in the Ssp phage-defence mechanism. ties and phage resistance of full-length SspE. This was reinforced by the observation that the single-point mutation of R100A Nicking endonuclease activity of SspE. The possession of in the DGQQR motif resulted in an altered circular-dichroism

DUF1524 prompted us to assess the nuclease activity of SspE. In spectrum of SspER100A and the deficiencies in PT-sensing GTPase, parallel with the ssDNA PT modification, SspE was identified as nicking nuclease activities and phage resistance (Fig. 4, Extended a nicking endonuclease based on two lines of evidence: (1) when Data Fig. 4 and Supplementary Fig. 9). In terms of the involvement incubated with supercoiled pUC19, a time-course assay showed that of NTPase activity, we speculated that PT-enhanced NTP hydroly- the first and predominant products were open circular plasmids sis may help with SspE translocation, conformational changes or and no specific nicking site was determined by run-off sequenc- movement along the DNA substrate to ensure the efficiency of anti- ing, as observed with the sequence-specific nicking endonuclease phage activity.

Nt.BspQI; and (2) the SspEH639A variant, harbouring a single-point Collectively, our results established the nature of the defensive mutation of H639 in the EHxxP endonuclease catalytic site, dis- action of the Ssp system, in which SspABCD accomplishes ssDNA played strongly impaired nicking activity on pUC19 (Extended PT modification by virtue of the nicking activity of SspB, and SspE Data Fig. 4). In support of this observation, the replacement of is a dual-function protein exerting both PT-sensing NTPase and H639 with an alanine on pWHU735 and pWHU3658 resulted in DNA-nicking endonuclease activities to damage invasive phage

SspABCD–SspEH639A providing no phage protection (Fig. 4a), which DNA and eventually disturb phage propagation early in infection implies that the interference of phage-DNA replication by the Ssp (Extended Data Fig. 5). system was attributable to nicking damage to the phage DNA. We thus assessed the integrity of the T4 DNA on an alkaline Discussion denaturing agarose gel. SspE caused negligible damage to phage Here, we have mapped out the similarities and differences of the DNA on a native agarose gel, agreeing well with the observation that two PT-based ssp and dnd systems, with the discovery of a role the phage DNA appeared intact in the presence of SspABCD–SspE for ssp genes in ssDNA PT modification and PT-sensing defence in Southern blot analyses, whereas the T4 DNA degraded severely against phages. The genes and proteins involved in the dnd and ssp under denaturing conditions, suggesting that massive nicks were modification systems are different, which points to an evolution- introduced in the phage genome by SspE (Fig. 4d). In an attempt ary divergence of previously unappreciated PT-based systems and to capture the possible SspE-induced nicks in vivo, we analysed the opens the door to the discovery of other PT-based epigenetic sys- ssDNA, and replicative form I (RFI; double-stranded, supercoiled) tems. Moreover, the divergence between dnd and ssp genes increases and II (RFII; double-stranded, nicked) DNA of ϕX174 by Southern when the other systems partnering with the modification proteins hybridization. Similar to T4 phage, the replication of ϕX174 was are considered. SspE is capable of sensing sequence-specific PTs to sensitive to the inhibitory action of the Ssp system, whereas no provide protection against phage infection, expanding our knowl- signs of DNA degradation were observed (Fig. 3d). However, we edge of the arsenal of anti-phage defensive mechanisms. did not detect a reduction in supercoiled RFI DNA or a simultane- At the biochemical level, the functional similarity between DndA ous increase in nicked RFII DNA at any time point of infection in and SspA and that between DndC and SspD indicate that the dnd

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

a JXY1 T4

–1 –6 10–1–10–6 10 –10

Vectors

SspABCD, SspE

SspABCD, SspEG97A

SspABCD, SspEH639A

SspABCD, SspER100A

SspABCD–SspE of SspABCD of V. cyclitrophicus FF75 and S. yokosukanensis DSM 40224 SspE of S. yokosukanensis DSM 40224

No DNA 3 3 3 3 CCA

) CPSCA –1 2 2 2 2

Velocity

(µM Pi min 1 1 1 1

0 0 0 0 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 ATP (µM) GTP (µM) CTP (µM) UTP (µM)

c d

G97A R100A H639A

SspE SspE SspE SspE DUF262 H639A R100A G97A 3 P = 0.0043 Uncut SspE SspE SspE SspE DUF1524 **

P = 0.00045 Neutral *** 2 P = 0.67 P = 0.75 P = 0.38 NS NS NS

Relative GTPase activity Denaturing

CA CA CA CA CA CCA AAC TTC GCC CCA CCA CCA CCA PS GTTC PS PS PS PS C GAACPS PS GGCC C C C C No DNA G G G PS No DNA No DNA No DNA No DNA

Fig. 4 | PT-sensing NTPase activity of SspE and its involvement in phage resistance. a, Plaque assays of phages JXY1 (left) and T4 (right) on S. lividans HXY6 and E. coli DH10B strains expressing SspABCD–SspE and variants at 28 °C. All results are representative of two independent experiments. b, NTPase activity was assessed in the absence of DNA or in the presence of DNA containing or lacking the 5′-CPSCA-3′ motif. Data are shown as the mean ± s.d. and are based on three independent experiments (n = 3). c, GTPase activity was enhanced by the addition of DNA fragments with 5′-CPSCA-3′ but not by those with 5′-GPSAAC-3′, 5′-GPSTTC-3′ or 5′-GPSGCC-3′. Data represent the mean ± s.d. (unpaired two-sided Student t-tests; n = 3 total data points from three independent experiments). **P < 0.01, ***P < 0.001; NS, not significant. d, T4 phage DNA was treated with SspE and variants at 28 °C for 9 h, followed by electrophoresis under both neutral and denaturing conditions. Images are representative of two independent experiments.

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

a b N domain

Q98 Q99

G97 R100 D96

C domain

R101 R100 R100 R101 ATP

Mg2+ ADP

Fig. 5 | Structural characteristics of SspE from S. yokosukanensis DSM 40224. a, Structure of full-length SspE. The α-helices, β-sheet and the loop are coloured cyan, red and purple, respectively. The amino and carboxy termini are connected by the loop area. b, Amino-terminal structure of SspE. The α-helix, β-sheet and the loop are coloured cyan, red and purple, respectively. The DGQQR motif is marked in yellow with side chains. c, Structure alignment of the NTPase-active sites from hSrx and SspE. The aligned structures of SspE with hSrx–ATP–Mg2+ (left) and SspE with hSrx–ADP (right) are shown. The hSrx, Mg2+ and ADP or ATP are coloured yellow, magenta and red, respectively, and SspE is coloured cyan. The side chains of the conserved arginine residues R100 (SspE) and R101 (hSrx) are labelled. and ssp PT systems probably share the initial sulfur mobilization In conclusion, we have described an unusual type of PT-based step but follow that with divergent steps for DNA-target selection defence system, in which SspABCD confers ssDNA PT modifica- and sulfur incorporation. This divergence is probably attributable tion and SspE functions as a nicking endonuclease to impair phage to SspB, a DNA nickase, whose nicking activity has not yet been replication in a PT-sensing manner. Considering the occurrence of observed for Dnd proteins but has been proposed to be a criti- the Ssp system in phylogenetically diverse bacteria, PT-mediated cal step in PT formation12. Although SspB exhibited no sequence Ssp defences may be widely adopted in bacteria, highlighting the preference for 5′-CCA-3′ in vitro, defects in SspB nicking activity diversity of PT systems and expanding the known arsenal of defence resulted in the simultaneous reduction or abolishment of PT modi- mechanisms employed by prokaryotes for protection against phage fication, which raises two possibilities: (1) the introduction of nicks infections. is a requisite for ssDNA PT, possibly followed by activation of DNA chemical groups at the nicking sites to facilitate sulfur incorpora- Methods tion; or (2) other elements—such as Ssp proteins, preferential rec- Bacterial strains, plasmids, phages and media. All of the strains, plasmids ognition sequences or DNA secondary structures—are essential to and phages used in this study are listed in Supplementary Table 2. Typically, the anchor SspB at target DNA sites to avoid unwanted DNA nicking Vibrio strains were grown at 28 °C in tryptic soy broth (TSB) or on agar plates and degradation. It is currently unclear which Dnd protein parallels supplemented with 2% NaCl, whereas the E. coli strains were grown in Luria– SspB in function but it is reasonable to deduce that an analogous Bertani (LB) broth or on agar plates at 37 °C unless otherwise indicated. step introduces breaks on both strands of consensus motifs to yield Construction of V. cyclitrophicus mutants. In-frame deletion of V. cyclitrophicus dsDNA PT modifications. Although further work is needed, the FF75 genes was carried out using the pJC4 suicide plasmid and a two-step observations made here with SspB represent a major step forward homologous recombination procedure. First, fragments up- and downstream in defining the biochemical pathways involved in PT modification of a target gene were generated via PCR using two pairs of primers (listed in by the ssp and dnd systems. Supplementary Table 3). Second, the recombinant fragment was amplified using a mixture of the up- and downstream PCR products (with an overlap of Another striking observation in this study is that the Ssp mech- approximately 40 bp) as templates and then inserted into pJC4. In this manner, we anism of action is not consistent with simple R-M activity. Two generated the plasmids pWHU722–9 for the construction of ssp mutants. After features of anti-phage activity are unusual: SspE represses phage sequencing, the resultant plasmids were transformed into the diaminopimelic replication by introducing nicks into phage DNA rather than per- acid auxotroph E. coli WM3064 as a donor, which was then conjugated to forming typical dsDNA degradation or cutting and the defensive V. cyclitrophicus FF75 as the recipient. Transconjugants were picked from TSB (with 2% NaCl) plates containing 2.5 μg ml−1 chloramphenicol without activity against phages depends on the stimulation by sequence- diaminopimelic acid and purified by serial streaking. The purified transconjugants specific PTs. An expanded discussion of SspE-mediated defence is were used for double crossovers on an LB agar plate containing 5% sucrose at provided in the Supplementary Discussion. 28 °C. In this way, we obtained in-frame deletion mutants of XXL-6, ΔsspA, ΔsspB,

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

ΔsspC, ΔsspD, ΔsspE and ΔsspBCDE. The deletion of these genes was verified by Construction of pWHU733. A 3.4-kb fragment including sspE and its promoter PCR and sequencing. was amplified from the genomic DNA of FF75 using the primer pair 184sspE- 1/184sspE-2 (Supplementary Table 3). The fragment was then digested with Construction of ΔsspA. First, 706-bp (left arm) and 733-bp (right arm) XbaI and SphI, and ligated with pACYC184, which had been digested using the chromosomal regions were amplified using the primer pairs 1902L1/1902L2 and same enzymes. 1902R1/1902R2, respectively (Supplementary Table 3). The two fragments were then combined into a single 1.4-kb fragment using 1902L1 and 1902R2. This Construction of pWHU734. A 3.1-kb fragment, including sspE and its promoter, fragment was sequenced and inserted into a SpeI- and SacI-digested pJC4 plasmid, was amplified from the genomic DNA of FF75 using the primer pair 184sspE-1/ generating pWHU727. The pWHU727 plasmid was then transformed into sspE-2-his (Supplementary Table 3). The fragment was subsequently cloned into E. coli WM3064 and conjugated into FF75 to generate a two-step double- pEASY-blunt zero. crossover mutation. Construction of pWHU735. The ORF region of the sspE gene (2.3-kb) was Construction of ΔsspB. First, 729-bp (left arm) and 755-bp (right arm) amplified from the genomic DNA of DSM 40224 using the primer pair 40224e-F- chromosomal regions were amplified using the primer pairs 1907L1/1907L2 and pql25/40224e-R-pql25 (Supplementary Table 3), whereas the synthetic promoter 1907R1/1907R2, respectively (Supplementary Table 4). The two fragments were PQL25 (ref. 37) was PCR amplified using the primer pair pql25-F-40224e/pql25- then combined into a single 1.4-kb fragment using 1907L1 and 1907R2. This R-40224e (Supplementary Table 3). The two PCR products of PQL25 and sspE fragment was sequenced and inserted into SpeI- and SacI-digested pJC4 plasmid, were combined using pql25-F-sspe40224 and 40224e-R-pql25. The resulting generating pWHU724. The pWHU724 plasmid was then transformed into PQL25-sspE fragment was used to replace the original sspE gene of FF75 on the E. coli WM3064 and conjugated into FF75 to generate a two-step double- plasmid pWHU733 via an in vitro recombination method using the Basic seamless crossover mutation. cloning and assembly kit (TransGen Biotech). For recombination, the plasmid pWHU733 was linearized via PCR using the primers 733-F-pql2540224e and Construction of ΔsspC. First, 728-bp (left arm) and 764-bp (right arm) 733-R-40224epql25 (Supplementary Table 3). chromosomal regions were amplified using the primer pairs 1908L1/1908L2 and 1908R1/1908R2, respectively (Supplementary Table 3). The two fragments were Construction of pWHU1811. A 7.1-kb fragment containing sspABCD was amplified then combined into a single 1.4-kb fragment using 1908L1 and 1908R2. This from the genomic DNA of S. clavuligerus ATCC 27064 using the primer pair cc-1/ fragment was sequenced and inserted into SpeI- and SacI-digested pJC4, generating cc-2. The 7.1-kb fragment was digested by XbaI and StuI and ligated into pSET152, pWHU722. The pWHU722 plasmid was then transformed into E. coli WM3064 which had been treated with the same enzymes. and conjugated into FF75 to generate a two-step double-crossover mutation. Construction of pWHU3481. A 1,867-bp fragment (sspA) with the C314S single- Construction of ΔsspD. First, 748-bp (left arm) and 765-bp (right arm) point mutation was amplified from the genomic DNA of FF75 using the primer chromosomal regions were amplified using the primer pairs 1909L1/1909L2 and pairs sspA-L1/sspA-L2 and sspA-R1/sspA-R2 (Supplementary Table 3). The two 1909R1/1909R2, respectively (Supplementary Table 3). The two fragments were fragments were then combined into a single 1.8-kb fragment using sspA-L1 and then combined into a single 1.5-kb fragment using 1909L1 and 1909R2. This sspA-R2. This fragment was digested with XbaI and SacII and replaced the original fragment was sequenced and inserted into a SpeI- and SacI-digested pJC4 plasmid, sspA in pWHU732, generating pWHU3481. generating pWHU725. The pWHU725 plasmid was then transformed into E. coli WM3064 and conjugated into FF75 for mutation. DNA-nicking assays. DNA-nicking activity was assessed with 1.5 μg pUC19 or phage DNA in 50 mM potassium acetate, 20 mM Tris–acetate (pH 7.9), 10 mM Construction of ΔsspE. First, 763-bp (left arm) and 764-bp (right arm) magnesium acetate or other metal ions (MnAc2, CuSO4, ZnCl2, CaCl2 and NiSO4), chromosomal regions were amplified using the primer pairs 1910L1/1910L2 and 0.1 mg ml−1 BSA and 2 μM SspB, SspE or derivatives in a total volume of 50 μl. 1910R1/1910R2, respectively (Supplementary Table 3). The two fragments were The reactions were carried out at 28 °C for a time course or as indicated. A 10-μl then combined into a single 1.5-kb fragment using 1910L1 and 1910R2. This reaction aliquot was subsequently quenched by adding 6×gel loading dye (purple; fragment was sequenced and inserted into a SpeI- and SacI-digested pJC4 plasmid, New England Biolabs), followed by 1% native agarose gel electrophoresis. The generating pWHU723. The pWHU723 plasmid was then transformed into E. coli DNA in the remaining 40-μl aliquot of the reaction was precipitated overnight at WM3064 and conjugated into FF75 for mutation. −80 °C by adding 4 μl of 3 M acetic acid (pH 5.2) and 88 μl ethanol before alkaline denaturing gel electrophoresis. Construction of ΔsspBCD. First, 700-bp (left arm) and 728-bp (right arm) chromosomal regions were amplified using the primer pairs 1907-1909L1/1907- Alkaline denaturing gel electrophoresis. Alkaline denaturing gel electrophoresis 1909L2 and 1907-1909R1/1907-1909R2, respectively (Supplementary Table 3). was performed following the protocol by Colombo and colleagues38. Enzyme- The two fragments were then combined into a single 1.4-kb fragment using 1907- treated DNA was precipitated with ethanol, washed once with 70% ethanol and 1909L1 and 1907-1909R2. This fragment was sequenced and inserted into a SacI- then resuspended in 1×alkaline loading buffer (50 mM NaOH, 1 mM EDTA, 2.5% digested pJC4 plasmid, generating pWHU3490. The pWHU3490 plasmid was then Ficoll (type 400) and 0.025% bromophenol blue) for 2 h. Meanwhile, the agarose transformed into E. coli WM3064 and conjugated into FF75 to generate a two-step gel (made in ddH2O) was equilibrated in 1×alkaline electrophoresis buffer (50 mM double-crossover mutation. NaOH and 1 mM EDTA) for at least 30 min. Electrophoresis was carried out at 2 V cm−1. Construction of ΔsspBCDE. The 701-bp (left arm) and 718-bp (right arm) chromosomal regions were amplified using the primer pairs 1910-BCD-L1/1910- Proteomics analysis. Overnight cultures of Vibrio strains were diluted 100-fold in BCD-L2 and 1910-BCD-R1/1910-BCD-R2, respectively (Supplementary Table 3). 20 ml fresh TSA (with 2% NaCl) medium and then cultured at 28 °C or 22 °C. The The two fragments were then combined into a single 1.4-kb fragment using 1910- cells (5 ml) were washed and collected when the cultures reached an optical density

BCD-L1 and 1910-BCD-R2. This fragment was sequenced and inserted into a at 600 nm (OD600) of 0.8. Protein extraction, digestion, isobaric tags for relative SacI-digested pJC4 plasmid, generating pWHU3491. The pWHU3491 plasmid was and absolute quantitation (iTRAQ) labelling and nano-LC–MS/MS analysis were then transformed into E. coli WM3064 and conjugated into ΔsspBCD to generate performed according to a protocol described by Shi and colleagues39. Trypsin/P ΔsspBCDE. was specified as a cleavage enzyme allowing up to two missing cleavages. The length of most peptides ranged from 7 to 16 amino acids, which is consistent with Construction of XXL-6. First, 742-bp (left arm) and 762-bp (right arm) the properties of tryptic peptides. The mass error was set to 10 ppm for precursor chromosomal regions were amplified using the primer pairs 1906-1903L1/1906- ions and 0.02 Da for fragment ions. Carbamidomethyl on cysteine, TMT-6plex 1903L2 and 1906-1903R1/1906-1903R2, respectively (Supplementary Table 3). (N-term) and TMT-6plex (K) were specified as fixed modifications. Oxidation of The two fragments were then combined into a single 1.5-kb fragment by a further Met was specified as a variable modification. The false discovery rate was adjusted round of PCR using 1906-1903L1 and 1906-1903R2. This fragment was sequenced to <1%, and the peptide ion score was set to ≥20. A total of 2,012 proteins were and inserted into a SpeI- and SacI-digested pJC4 plasmid, generating pWHU726. identified, of which 1,677 were quantified. Proteins with an average fold change The pWHU726 plasmid was then transformed into E. coli WM3064 and >1.30 or <0.77 were considered up- or downregulated, respectively (P < 0.05)40. conjugated into FF75 to generate a two-step double-crossover mutation. A Gene Ontology-annotated proteome was derived from the UniProt-GOA database (http://www.ebi.ac.uk/GOA/), and the proteins were classified by Construction of pWHU732. First, 5,828-bp (sspBCD) and 1,867-bp (sspA) Gene-Ontology annotation based on three categories: biological process, cellular fragments were amplified from the genomic DNA of FF75 using the primer pairs component and molecular function. sspBCD1/sspBCD-2 and sspA-1/sspA-2 (Supplementary Table 3), respectively. The 5,828-bp fragment was digested with KpnI and BamHI, and ligated with Protein expression and purification. The gene encoding the full-length SspB pBluescript II SK(+), which had been digested using the same enzymes. The protein from S. clavuligerus ATCC 27064 was subcloned into the pET28a vector resultant plasmid was then digested with XbaI and SacII, and ligated with the along with an N-terminal 6×His tag. The resulting plasmid was transformed into 1,867-bp (sspA) fragment after its digestion with the same enzymes. E. coli BL21(DE3) competent cells. The transformed cells were cultured at 28 °C

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

until the OD600 reached 0.8, and protein expression was induced with 0.2 mM The crystals belonged to the P212121 space group and contained two molecules of isopropyl β-d-1-thiogalactopyranoside for 14–18 h at 16 °C. After harvesting and SspB in one asymmetric unit. The final refined model had Rwork and Rfree values of resuspension, the cells were lysed using a cell homogenizer (JNBIO). The lysate was 17.48 and 22.45%, respectively. The quality of the structural model was checked cleared by centrifugation at 24,000g for 30 min, and the supernatant was subjected with the CCP4 program PROCHECK44, which showed good stereochemistry to Ni2+-NTA-affinity chromatography (GE Healthcare). The eluted protein was according to a Ramachandran plot. loaded onto a Superdex 200 gel filtration column (GE Healthcare) in a solution Crystallization trials for wild-type SspE were initially screened by Index, containing 10 mM Tris–HCl, pH 8.0, 100 mM NaCl and 2 mM dithiothreitol. The Crystal screen 1, Crystal screen 2, Salt RX 1, Salt RX 2, PEG/Ion 1 and PEG/Ion peak fractions were combined and concentrated to 10 mg ml−1 for crystallization. 2 kits from Hampton Research and Wizard I, II, III and IV kits from Rigaku. The The gene encoding SspE from S. yokosukanensis DSM 40224 hanging-drop method by vapour-diffusion crystallization in 24-well plates was was amplified by high-fidelity PCR using the following primers: carried out at a temperature of 14 °C. Typically, 1 μl reservoir solution was mixed 5′-GGGAATTCCATATGATGGAGACTAAAGAGATCTTCG-3′ with 1 μl protein solution and equilibrated against 200 ml reservoir solution. After (forward; NdeI restriction site indicated in bold font) and the initial screening, small SspE crystals were obtained from condition no. 63 of 5′-ATAAGAATGCGGCCGCTTACGGCTCGAATCCCAG-3′ (reverse; NotI the Index kit, which consisted of 10% polyethylene glycol monomethyl ether 5000, restriction site indicated in bold font). The full-length sspE was inserted into the 5% Tacsimate pH 7.0 and 0.1 M HEPES pH 7.0. We finely tuned the precipitant expression vector pET28a (Novagen) through the NdeI and NotI sites, yielding concentration by fine-tuning, concentration range of 4–12%, to optimize the pWHU3641. Successfully cloned plasmids were verified by sequencing. The crystallization conditions. The concentration and pH of the salt Tacsimate and plasmids pWHU3641 and pWHU732 (carrying sspABCD from V. cyclitrophicus HEPES buffer were then both adjusted. Finally, a well-formed single crystal was FF75) were co-transformed into the E. coli strain BL21 (DE3). The SspE protein grown in a solution containing 6% (wt/vol) polyethylene glycol monomethyl ether constructs were expressed with a His tag on the N terminus. When the OD600 5000, 5% (vol/vol) Tacsimate pH 7.0 and 0.09 M HEPES pH 7.0 using the hanging- reached 0.6–0.8, the bacterial cultures were induced with 0.5 mM isopropyl β-d- drop vapour-diffusion method in 24-well plates at 14 °C. 1-thiogalactopyranoside and cultured overnight at 15 °C. The bacterial cells were Se-SAD X-ray data of SspE at a resolution of 3.31 Å were collected on the then collected by centrifugation at 3,000g for 20 min at 4 °C. The cell pellets were SSRF beamline BL19U1. The data were integrated and scaled using HKL2000 homogenized in buffer containing 25 mM Tris–HCl pH 8.0, 300 mM NaCl, 20 mM (ref. 47). Further processing was carried out using programs from the CCP4 suite48. imidazole and 0.05 mM phenylmethylsulfonyl fluoride, and lysed by sonication. Phaser-EP43,49 was used for SAD experimental phasing, and RESOLVE from the The cell debris was removed by centrifugation at 24,000g for 50 min at 4 °C. The PHENIX50 tools suite was used for model building. One distinct solution was supernatant was applied twice to a Ni2+-NTA-affinity chromatography column identified. The resulting electron density was of sufficiently good quality and and then 5-ml volumes of the sample were eluted with 40 mM, 80 mM, 110 mM, most of the side chains were clearly shown. Additional missing residues were 300 mM and 500 mM imidazole. The target protein products were digested manually added in COOT51. The structure was refined using PHENIX50. Model overnight by thrombin at 4 °C and further purified by Source 15Q ion-exchange validation was performed using PROCHECK52 and the WHATCHECK routine and gel-filtration chromatography using a Superdex 200 column (GE Healthcare). of WHAT IF53. The Superdex 200 buffer used was 10 mM Tris–HCl pH 8.0, 100 mM NaCl and 2 mM dithiothreitol. The purity of the protein was verified by SDS–PAGE and Phage isolation. Soil samples (10–20 g) were mixed with 1 ml of a spore suspension Coomassie-blue staining. The purified protein from Superdex 200 gel-filtration of S. lividans HXY6 and incubated in 50 ml TSBY broth (30 g l−1 TSB, 340 g l−1 −1 −1 chromatography was concentrated to 10 mg ml , flash-frozen in liquid nitrogen sucrose, 5 g l yeast extraction) supplemented with 1 mM CaCl2 overnight at 28 °C. and stored at −80 °C until use. The supernatant was collected by centrifugation at 3,000g for 10 min at 4 °C and filtered through a 0.22-μm membrane (Millipore). Serially diluted filtrates were Southern blot assays. An overnight culture of bacterial cells was diluted 1:100 plated onto a double-layer MS plate (20 g l−1 soybean powder, 20 g l−1 d-mannitol, in 100 ml LB medium and cultured at 37 °C to an OD600 of approximately 0.5, 2% bottom agar and 0.8% top agar that mixed with 50 μl of a spore suspension of followed by infection with either phage T4 or ϕX174 at an multiplicity of infection S. lividans HXY6). After 36 h of incubation at 28 °C, single plaques were picked into of 6 or 10, respectively. Total DNA was extracted from the cells and collected 1 ml TSBY broth and incubated at 4 °C for 24 h. Three single-plaque isolation cycles throughout the phage infection time course using the chloroform–phenol method. were conducted to obtain pure phage stocks. The phage produced on S. lividans A total of 1.5 μg T4 DNA was digested by PacI for 6 h before electrophoresis on HXY6 was designated JXY1. The morphology of phage JXY1 was observed using a 1% agarose gel. For phage ϕX174, 1 μg undigested total DNA was subjected to an electron microscope (1200EX JEOL) at 60 kV. The genomic DNA of JXY1 was hybridization. Electrophoresis was performed in 1×TAE buffer at 4 V cm−1. The gel extracted and sequenced by the Nanopore PromethION platform by NextOmics. was transferred to an Amersham Hybond-N+ nylon membrane (GE Healthcare) in The sequence reads were assembled into a 45,385-bp circular DNA using the Canu, 20×SSC buffer (0.3 M trisodium citrate and 3 M NaCl, pH 7.0), and the subsequent Pilon and Circlator54–56 programs with an average coverage of 11,742×. The FASTA- blotting, washing and signal-detection steps were carried out using the DIG- formatted complete sequence was annotated using the RASTtk pipeline. high prime DNA labeling and detection starter kit II (Roche) according to the manufacturer’s protocol. Phage plaque assays. E. coli cells were cultured overnight with shaking, diluted

1:100 in LB medium and cultured to an OD600 of 0.8. An aliquot of the cell culture Western blot analysis. A Trans1-T1 E. coli strain harbouring pWHU734 was (500 µl) was added to 10 ml LB broth with melted 0.75% agar, and then overlaid on cultured at 28 °C to the mid-log phase, which was followed by a temperature a pre-poured 1.5% LB agar plate. Phage stocks were serially diluted in SM buffer reduction from 28 °C to 15 °C. Cells were collected before and after cold treatment (100 mM NaCl, 8 mM MgSO4 and 50 mM Tris–HCl pH 7.5). After the top agar (15 °C, 1.5 h) for western blot analysis. The proteins were separated by SDS–PAGE solidified, 2 μl of the diluted phages were spotted onto the E. coli lawns. Plaques and wet-transferred onto polyvinylidene fluoride membranes. The membranes were imaged after overnight growth at the indicated temperatures. were incubated in 5% (wt/vol) non-fat milk in Tris-buffered saline containing 0.1% Tween and then incubated with either a primary mouse anti-6×His antibody Circular-dichroism spectroscopy. The circular-dichroism spectra of the (Proteintech) followed by a secondary goat anti-mouse horseradish peroxidase- indicated proteins (0.2 mg ml−1) were collected on a Chirascan circular-dichroism tagged antibody or a primary rabbit anti-RpoB antibody (Abcam) followed by a spectroscopy instrument (Applied Photophysics) at 25 °C in a low-salt buffer secondary goat anti-rabbit horseradish peroxidase-tagged antibody. Finally, the (10 mM NaCl and 1 mM Tris–HCl pH 8.0). Circular-dichroism spectra covering membranes were imaged using a ChemiDoc XRS+ system (Bio-Rad), and images 190–280 nm were acquired with a bandwidth of 5 nm, response time of 2 s, were analysed using the Image Lab 3.0 software (Bio-Rad). scanning speed of 200 nm min−1 and step size of 1 nm. The results were corrected by subtracting the buffer baseline to avoid a signal contribution from the buffer. Crystallization, data collection and structure determination. Crystals of native Each spectrum was calculated and plotted as the average of three runs. and SeMet-substituted SspB protein from S. clavuligerus ATCC 27064 were grown at 14 °C using the hanging-drop vapour-diffusion method, with 1 l protein mixed Molecular graphics. All protein structure figures were generated using the PyMOL μ 57 with an equal volume of reservoir solution containing 25% polyethylene glycol program . The sequence conservation of SspB was mapped onto the surface of its 58 (PEG) 3350 and 0.2 M magnesium chloride (Hampton Research). The crystals crystal structure was generated using the ConSurf server (http://consurf.tau.ac.il) . were cryoprotected in crystallization buffer supplemented with 25% ethylene glycol. One set of SAD data for SeMet–SspB at a resolution of 1.95 Å and one set Reporting Summary. Further information on research design is available in the of data for native SspB at a resolution of 1.75 Å resolution were collected using Nature Research Reporting Summary linked to this article. the BL17U1 beamline at the Shanghai Synchrotron Radiation Facility (SSRF) at 100 K. Data reduction was performed using the HKL2000 software41. The phases Data availability for SeMet–SspB were determined by the SAD method using the Autosol module The data supporting the findings of this study are available from the corresponding of PHENIX42, and the phases for native SspB were determined by the molecular author on request. The atomic coordinates and structure factors of SspB and the replacement method using the Phaser program43 in the CCP4 package44 with the magnesium ion-bound SspB from S. clavuligerus ATCC 27064 have been deposited structure of SeMet–SspB as the search model. After automatic model building in the Protein Data Bank with the accession numbers 6JUF and 6LB9, respectively, using the AutoBuild module of PHENIX42 and manual model building in COOT45, and those of SspE from S. yokosukanensis DSM 40224 have the accession number refinement was performed using the REFMAC5 program46 in the CCP4 package44. 6JIV. The genome sequence of phage JXY1 has been deposited in GenBank under

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles accession number MN994275. Source data for Figs. 1b, 2a,b,e,f, 3a–d and 4a–d 28. Yao, F., Xu, T., Zhou, X., Deng, Z. & You, D. Functional analysis of spfD gene and Extended Data Figs. 2c, 3, 4a, 6a,b and 7a–e are included in this article and its involved in DNA phosphorothioation in Pseudomonas fuorescens Pf0-1. Supplementary information. FEBS Lett. 583, 729–733 (2009). 29. Xia, S. et al. Tight control of genomic phosphorothioate modifcation by the Code availability ATP-modulated autoregulation and reusability of DndB. Mol. Microbiol. 111, Custom codes or software used in this study are available from the authors on 938–950 (2018). request. 30. Maindola, P. et al. Multiple enzymatic activities of ParB/Srx superfamily mediate sexual confict among conjugative plasmids. Nat. Commun. 5, 5322 (2014). Received: 8 April 2019; Accepted: 2 March 2020; 31. Schumacher, M. A. & Funnell, B. E. Structures of ParB bound to DNA reveal Published: xx xx xxxx mechanism of partition complex formation. Nature 438, 516–519 (2005). 32. Basu, M. K. & Koonin, E. V. Evolution of eukaryotic cysteine sulfnic acid References reductase, sulfredoxin (Srx), from bacterial chromosome partitioning protein 1. Yamaguchi, Y., Park, J. H. & Inouye, M. Toxin-antitoxin systems in bacteria ParB. Cell Cycle 4, 947–952 (2005). and archaea. Annu. Rev. Genet. 45, 61–79 (2011). 33. He, X. et al. Expression and purifcation of a single-chain Type IV restriction 2. Tock, M. R. & Dryden, D. T. Te biology of restriction and anti-restriction. enzyme Eco94GmrSD and determination of its substrate preference. Sci. Rep. Curr. Opin. Microbiol. 8, 466–472 (2005). 5, 9747 (2015). 3. Dy, R. L., Przybilski, R., Semeijn, K., Salmond, G. P. & Fineran, P. C. A 34. Jablonska, J., Matelska, D., Steczkiewicz, K. & Ginalski, K. Systematic widespread bacteriophage abortive infection system functions through a Type classifcation of the His-Me fnger superfamily. Nucleic Acids Res. 45, IV toxin-antitoxin mechanism. Nucleic Acids Res. 42, 4590–4605 (2014). 11479–11494 (2017). 4. Marrafni, L. A. CRISPR–Cas immunity in prokaryotes. Nature 526, 35. Chen, F. et al. Crystal structure of the cysteine desulfurase DndA from 55–61 (2015). Streptomyces lividans which is involved in DNA phosphorothioation. 5. Swarts, D. C. et al. DNA-guided DNA interference by a prokaryotic PLoS ONE 7, e36635 (2012). Argonaute. Nature 507, 258–261 (2014). 36. Machnicka, M. A., Kaminska, K. H., Dunin-Horkawicz, S. & Bujnicki, J. M. 6. Goldfarb, T. et al. BREX is a novel phage resistance system widespread in Phylogenomics and sequence-structure-function relationships in the GmrSD microbial genomes. EMBO J. 34, 169–183 (2015). family of Type IV restriction enzymes. BMC Bioinform. 16, 336 (2015). 7. Ofr, G. et al. DISARM is a widespread bacterial defence system with broad 37. Liu, Q. et al. Engineering an iterative polyketide pathway in Escherichia coli anti-phage activities. Nat. Microbiol. 3, 90–98 (2018). results in single-form alkene and alkane overproduction. Metab. Eng. 28, 8. Doron, S. et al. Systematic discovery of antiphage defense systems in the 82–90 (2015). microbial pangenome. Science 359, eaar4120 (2018). 38. Colombo, C. V., Menin, L. & Clerici, M. Alkaline denaturing Southern blot 9. Willbanks, A. et al. Te evolution of epigenetics: from prokaryotes to humans analysis to monitor double-strand break processing. Methods Mol. Biol. 1672, and its biological consequences. Genet. Epigenet. 8, 25–36 (2016). 131–145 (2018). 10. Blow, M. J. et al. Te epigenomic landscape of prokaryotes. PLoS Genet. 12, 39. Shi, H., Wang, X., Tan, D. X., Reiter, R. J. & Chan, Z. Comparative e1005854 (2016). physiological and proteomic analyses reveal the actions of melatonin in the 11. Wang, L. et al. Phosphorothioation of DNA in bacteria by dnd genes. reduction of oxidative stress in Bermuda grass (Cynodon dactylon (L). Pers). Nat. Chem. Biol. 3, 709–710 (2007). J. Pineal Res. 59, 120–131 (2015). 12. Wang, L., Jiang, S., Deng, Z., Dedon, P. C. & Chen, S. DNA phosphorothioate 40. Satoh, K., Maniwa, T., Oda, T. & Matsumoto, K. Proteomic profling for the modifcation—a new multi-functional epigenetic system in bacteria. identifcation of serum diagnostic biomarkers for abdominal and thoracic FEMS Microbiol. Rev. 43, 109–122 (2018). aortic aneurysms. Proteome Sci. 11, 27 (2013). 13. Tiaville, J. J. et al. Novel genomic island modifes DNA with 7-deazaguanine 41. Otwinowski, Z. & Minor, W. Processing of X-ray difraction data collected in derivatives. Proc. Natl Acad. Sci. USA 113, E1452–E1459 (2016). oscillation mode. Methods Enzymol. 276, 307–326 (1997). 14. Zhou, X. et al. A novel DNA modifcation by sulphur. Mol. Microbiol. 57, 42. Adams, P. D. et al. PHENIX: building new sofware for automated 1428–1438 (2005). crystallographic structure determination. Acta Crystallogr. D 58, 15. Eckstein, F. & Gish, G. Phosphorothioates in molecular biology. Trends 1948–1954 (2002). Biochem. Sci. 14, 97–100 (1989). 43. McCoy, A. J. et al. Phaser crystallographic sofware. J. Appl. Crystallogr. 40, 16. Gan, R. et al. DNA phosphorothioate modifcations infuence the global 658–674 (2007). transcriptional response and protect DNA from double-stranded breaks. 44. Collaborative Computational Project, Number 4. Te CCP4 suite: programs Sci. Rep. 4, 6642 (2014). for protein crystallography. Acta Crystallogr. D 50, 760–763 (1994). 17. Cao, B. et al. Pathological phenotypes and in vivo DNA cleavage by 45. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. unrestrained activity of a phosphorothioate-based restriction system in Acta Crystallogr. D 60, 2126–2132 (2004). Salmonella. Mol. Microbiol. 93, 776–785 (2014). 46. Winn, M. D., Murshudov, G. N. & Papiz, M. Z. Macromolecular TLS 18. Xu, T., Yao, F., Zhou, X., Deng, Z. & You, D. A novel host-specifc restriction refnement in REFMAC at moderate resolutions. Methods Enzymol. 374, system associated with DNA backbone S-modifcation in Salmonella. Nucleic 300–321 (2003). Acids Res. 38, 7133–7141 (2010). 47. Otwinowski, Z. & Minor, W. Processing of X-ray difraction data collected in 19. Chen, C. et al. Convergence of DNA methylation and phosphorothioation oscillation mode. Methods Enzymol. 276, 307–326 (1997). epigenetics in bacterial genomes. Proc. Natl Acad. Sci. USA 114, 48. Winn, M. D. et al. Overview of the CCP4 suite and current developments. 4501–4506 (2017). Acta Crystallogr. D 67, 235–242 (2011). 20. Cao, B. et al. Genomic mapping of phosphorothioates reveals partial 49. McCoy, A. J., Storoni, L. C. & Read, R. J. Simple algorithm for a maximum- modifcation of short consensus sequences. Nat. Commun. 5, 3951 (2014). likelihood SAD function. Acta Crystallogr. D 60, 1220–1228 (2004). 21. Kellner, S. et al. Oxidation of phosphorothioate DNA modifcations leads to 50. Adams, P. D. et al. PHENIX: a comprehensive Python-based system for lethal genomic instability. Nat. Chem. Biol. 13, 888–894 (2017). macromolecular structure solution. Acta Crystallogr. D 66, 213–221 (2010). 22. Wang, L. et al. DNA phosphorothioation is widespread and quantized in 51. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and bacterial genomes. Proc. Natl Acad. Sci. USA 108, 2963–2968 (2011). development of Coot. Acta Crystallogr. D 66, 486–501 (2010). 23. Tong, T. et al. Occurrence, evolution, and functions of DNA 52. Laskowski, R. A., MacArthur, M. W., Moss, D. S. & Tornton, J. M. phosphorothioate epigenetics in bacteria. Proc. Natl Acad. Sci. USA 115, PROCHECK: a program to check the stereochemical quality of protein E2988–E2996 (2018). structures. J. Appl. Crystallogr. 26, 283–291 (1993). 24. Aziz, R. K. et al. Te RAST server: rapid annotations using subsystems 53. Vriend, G. WHAT IF: a molecular modeling and drug design program. technology. BMC Genomics 9, 75 (2008). J. Mol. Graph. 8, 52–56 (1990). 25. Zimmermann, L. et al. A completely reimplemented MPI bioinformatics 54. Hunt, M. et al. Circlator: automated circularization of genome assemblies toolkit with a new HHpred server at its core. J. Mol. Biol. 430, using long sequencing reads. Genome Biol. 16, 294 (2015). 2237–2243 (2018). 55. Walker, B. J. et al. Pilon: an integrated tool for comprehensive microbial variant 26. Anantharaman, V., Makarova, K. S., Burroughs, A. M., Koonin, E. V. & detection and genome assembly improvement. PLoS ONE 9, e112963 (2014). Aravind, L. Comprehensive analysis of the HEPN superfamily: identifcation 56. Koren, S. et al. Canu: scalable and accurate long-read assembly via adaptive of novel roles in intra-genomic conficts, defense, pathogenesis and RNA k-mer weighting and repeat separation. Genome Res. 27, 722–736 (2017). processing. Biol. Direct 8, 15 (2013). 57. Te PyMOL Molecular Graphics System, Version 1.8 (Schrödinger, 27. You, D., Wang, L., Yao, F., Zhou, X. & Deng, Z. A novel DNA modifcation by LLC, 2015). sulfur: DndA is a NifS-like cysteine desulfurase capable of assembling DndC 58. Landau, M. et al. ConSurf 2005: the projection of evolutionary conservation as an iron-sulfur cluster protein in Streptomyces lividans. Biochemistry 46, scores of residues on protein structures. Nucleic Acids Res. 33, 6126–6133 (2007). W299–W302 (2005).

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

Acknowledgements experiment. G.W., R.S., Z.L., Y.Wang, Z.D., J.W., P.D., Shi Chen and L.W. analysed the We thank C. Chen for his assistance in the preparation of the manuscript. We thank study data. G.W. designed the crystallization experiments. X.X., G.W., Y.Wei, P.C.D., M. Polz at the Massachusetts Institute of Technology for his gift of the Vibrio strains. Shi Chen and L.W. wrote the manuscript. We thank the staff at the BL17U1 beamline at the SSRF. The T4gt phage was provided by S. Xu (New England Biolabs). This work was supported by grants from the National Competing interests Natural Science Foundation of China (grant nos. 31720103906, 31925002, 31520103902 The authors declare no competing interests. and 31872627), China National Key Research and Development Program (grant no. 2019YFA0904300), Open Funding Project of State Key Laboratory of Microbial Metabolism, US National Science Foundation (grant no. CHE-1019990), US National Institute of Allergy Additional information and Infectious Disease (grant no. AI112711) and the Singapore–MIT Alliance for Research Extended data is available for this paper at https://doi.org/10.1038/s41564-020-0700-6. and Technology, sponsored by the National Research Foundation of Singapore. Supplementary information is available for this paper at https://doi.org/10.1038/ s41564-020-0700-6. Author contributions Correspondence and requests for materials should be addressed to L.W. L.W. supervised the study. X.X. and Y.Wei performed most of the biochemical and Reprints and permissions information is available at www.nature.com/reprints. genetic experiments and analyses. L.L., Yubing Zhang and H.G. performed the crystallization experiments and analyses. X.J., M.L. and R.H. performed the phage- Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in related assays. Si Chen conducted the bioinformatics analysis. X.T., Yizhou Zhang, published maps and institutional affiliations. L.H. and S.J. performed the mutation experiments. Y.T. performed the proteomics © The Author(s), under exclusive licence to Springer Nature Limited 2020

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

Extended Data Fig. 1 | Structural characterization of SspB (locus_tag SSCG_01550). a, Sequence alignment of SspB with residue numbers and secondary structure elements indicated. Several putative dimerization-mediating residues in SspB, for example, H28, Y103, and E105, are replaced by similar amino acid residues in SspB homologues from other species; therefore, they are considered conserved. The imidazole ring of H28 resembles the benzene ring of F/Y in the SspB homologues of other species, which could also be involved in the hydrophobic interactions and thus serve a similar function. At the Y103 position of SspB homologues, there exists either a Y or an F residue, both of which possess benzene rings and can form hydrophobic interactions. The homologues of SspB have either E or D at E105. Both E and D have negatively charged carboxyl groups in their side chains and would play similar roles in the formation of salt bridges at the dimeric interface. b, Overall structure of SspB from S. clavuligerus ATCC 27064. α-helices, β-sheets, and loops are coloured red, yellow, and green, respectively. c, Conserved motifs 1, 2 and 3 are coloured magenta, green, and blue, respectively. d, The surface of the SspB structure is coloured according to the conservation score calculated by the ConSurf server. The “front” side of the SspB surface is highly conserved, while the “back” side is much less conserved. Motifs 1, 2, and 3 are mapped to the most conserved surface on the SspB structure.

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

Extended Data Fig. 2 | Interaction of magnesium ions with SspB. a, Ribbon diagram of magnesium ion-bound SspB from S. clavuligerus ATCC 27064. Magnesium ions are shown as small blue spheres. b, A magnesium ion is bound to E18 in each monomer. The SspB-Mg2+ complex was prepared by soaking the apo SspB crystals in reservoir solution containing 25% polyethylene glycol (PEG) 3350 and 1.0 M magnesium chloride for 6 h. c, The single- point mutation E18R remarkably impaired the DNA nicking activity of SspB. Data shown are representative images of two independent experiments.

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

Extended Data Fig. 3 | Phylogenetic distribution of 234 homologues of the V. cyclitrophicus FF75 SspBCD–SspE in bacteria. Phylogenetic groups are coloured by group (see the legend). Representative strains are labelled. For clarity, other strain names are not shown but are provided in Supplementary Data 2.

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

Extended Data Fig. 4 | Sequence non-specific nicking activity of SspE. a, Time-course treatment of pUC19 DNA by SspE. In this assay, 0.3 µg of pUC19 DNA was untreated or treated with 2 µM SspE at 28 °C in CutSmart (New England Biolabs, 50 mM potassium acetate, 20 mM Tris-acetate, pH 7.9, 10 mM magnesium acetate, 1.5 μM BSA). At the indicated time points, aliquots of the reaction mixture were withdrawn and analysed on a 1% agarose gel. L, linear pUC19 DNA. Nt.BspQI-nicked pUC19 was used as a control. Data shown are representative images of three independent experiments. b, Run-off sequencing of gel-purified SspE- or Nt.BspQI-nicked pUC19 DNA. The peaks in the sequencing chromatogram diminish sharply at the sequence-specific nicking site of Nt.BspQI. The extra adenine (A) base at the end of the run-off sequence was incorporated by the Taq DNA polymerase. In contrast, no such sequencing peak drops were observed when SspE-nicked pUC19 DNA was used as template despite the presence of the 5′-CCA-3′ motif.

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

Extended Data Fig. 5 | A schematic model of the Ssp system and defence against phage infection.

Nature Microbiology | www.nature.com/naturemicrobiology Articles NATurE MICroBIoloGy

Extended Data Fig. 6 | Interaction of SspB and SspC. a, Pulldown of SspC with purified glutathione S-transferase (GST)-tagged SspB. BL21(DE3) cell lysate overexpressing 6xHis-SspC protein of S. clavuligerus ATCC 27064 was used as an input. A pulldown assay was performed by incubating the input with 56 pmol of purified GST-SspB of ATCC 27064 or GST, which was immobilized on glutathione (GSH) magnetic beads in PBS buffer at 4 °C overnight. Input and pulldown samples were examined via immunoblotting analysis. *, nonspecific bands. Data shown are representative images of three independent experiments. b, Nicking activity assays of SspB in the presence of SspC. Supercoiled pUC19 (0.3 µg) was incubated with 2 µM SspB, 2 µM SspB and 2 µM SspC or 2 µM SspBE204R and 2 µM SspC at 28 °C in CutSmart buffer. At the indicated time points, aliquots of the reaction mixture were withdrawn and analysed on a 1% agarose gel. Data shown are representative images of three independent experiments. c, Run-off sequencing results of gel-purified nicked pUC19 DNA. Nicked pUC19 DNA resulting from the treatment with SspB-SspC was gel-purified and subjected to run-off sequencing. The peaks in the sequencing chromatogram diminish sharply at the sequence-specific nicking site of Nt.BspQI. In contrast, no such sequencing peak drops were observed when SspB-SspC-nicked pUC19 DNA was used as template despite the presence of the 5′-CCA-3′ motif.

Nature Microbiology | www.nature.com/naturemicrobiology NATurE MICroBIoloGy Articles

Extended Data Fig. 7 | Effect of temperature on growth and ssp gene expression in Vibrio strains. a, Growth of the wild-type and ssp mutants was assessed at 28 °C, 15 °C and 4 °C, respectively, on TSB agar plates supplemented with 2% NaCl. M565_ctg1P1910 = SspE. Results are representative of two independent experiments. b, The transcription of sspBCD and sspE at 15 °C and 28 °C was compared by quantitative real-time PCR with the housekeeping gene gapA, encoding glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as a reference. Data are shown as the mean ± s.d. and based on three independent experiments (n = 3). Error bars are s.d. of the means. Two-sided Student t test was used for statistical analysis. **, p < 0.01. ns, not significant. In contrast to the enhanced transcription of sspE at 15 °C, sspBCD is constitutively expressed at all tested temperatures. c, E. coli Trans1-T1

(pWHU734) cells expressing His-tagged SspE under its native promoter were grown at 28 °C to an OD600 of 0.5 and were subsequently shifted to 15 °C for 1.5 h. The expression level of SspE was determined by Western blot. RpoB (RNA polymerase beta subunit) was used as an internal control. Results are representative of two independent experiments. d, Cell survival following a temperature downshift from 28 °C to 15 °C was assessed by determining the number of cells (colony-forming units, cfu) in the cultures at different time points. CFU numbers represent the mean values of three independent experiments (n = 3). Error bars are s.d. of the means. e, Cell survival after 6 h of antibiotic treatment (as a percentage on a log scale). Exponentially growing Vibrio cells were pretreated at 15 °C for 1.5 h, followed by exposure to 100 μg ml-1 ampicillin or 200 μg ml-1 streptomycin for 6 h. The concentrations of antibiotics used here are more than five times the minimum inhibitory concentration for V. cyclitrophicus FF75. Data are shown as mean ± s.d. and are representative of three independent experiments. Two-sided Student t test was used to measure significance. *, p < 0.05, ***, p < 0.001.

Nature Microbiology | www.nature.com/naturemicrobiology nature research | reporting summary

Corresponding author(s): Lianrong Wang

Last updated by author(s): Feb 13, 2020 Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see Authors & Referees and the Editorial Policy Checklist.

Statistics For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a Confirmed The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section. A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)

For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable.

For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated

Our web collection on statistics for biologists contains articles on many of the points above. Software and code Policy information about availability of computer code Data collection Both online and local NCBI BLAST (v2.2.30+) were used to search the NCBI database. The license of NCBI BLAST: Public domain For the assembly of phage genome sequencing, programs canu v1.7.11, pilon v1.22 and circlator v1.5.5 were used. The license of canu/ pilon/circlator: Public domain

Data analysis Multiple alignments and phylogenetic tree constructions were performed using MEGA (v5.22). ImageLab 3.0 software was used for imaging analysis.

The licenses for the above software were described as follow: MEGA: protected under the copyright law, but FREE for use in research and education. It is not allowed to redistribute the MEGA software and associated materials partially or fully in any form.

For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information. Data October 2018 Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: - Accession codes, unique identifiers, or web links for publicly available datasets - A list of figures that have associated raw data - A description of any restrictions on data availability

The atomic coordinates and structure factors of SspB and the magnesium ion-bound SspB from S. clavuligerus ATCC 27064 have been deposited in the Protein Data Bank with the accession numbers 6JUF and 6LB9, respectively. The atomic coordinates and structure factors of SspE from S. yokosukanensis DSM 40224 has the accession number as 6JIV.

1 List of the figures that have associated raw data: Figure 1-4, Extended Data Figure 2-4 and 6-7, Supplementary Figure 1-9 and 11-13.

The annotated genome data of phage JXY1 has been submitted to NCBI with accession number MN994275 and will be released at Feb 17, 2020. nature research | reporting summary

Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf

Life sciences study design All studies must disclose on these points even when the disclosure is negative. Sample size No sample-size calculation was performed. For gel imaging studies we perform at least 2 independent experiments and for assays with statistical analysis, at least three independent experiments were performed in order to get enough data for P value calculation.

Data exclusions No data exclusions.

Replication Values represent the mean ± SD of at least three analyses.

Randomization This is not relevant to our study. This study focuses on the discovery of novel bacterial defense systems

Blinding This is not relevant to our study. This study focuses on the discovery of novel bacterial defense systems

Reporting for specific materials, systems and methods We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Materials & experimental systems Methods n/a Involved in the study n/a Involved in the study Antibodies ChIP-seq Eukaryotic cell lines Flow cytometry Palaeontology MRI-based neuroimaging Animals and other organisms Human research participants Clinical data

Antibodies Antibodies used Anti-RNA polymerase beta antibody (ab191598) (Catalog number:ab191598, CloneNo: EPR18704, Lot: GR224168-11, dilution: 1/2000, Abcam, Cambridge, UK) 6*His, His-Tag Antibody (Catalog number: 66005-1-Ig, CloneNo.: 1B7G5, Lot: 10004365, dilution: 1/10000, Proteintech, Wuhan, China) GST Monoclonal Antibody (Mix) (Catalog number: YM3144, Lot: 411005100205, dilution: 1/5000, Immunoway, Plano, USA)

Validation Anti-RNA polymerase beta antibody has been validated by abcam as it is capable of detecting a approximately 150 kDa band in western-blot of E. coli RNA polymerase beta subunit with 10 μg E. coli whole cell lysate. 6*His, His-Tag Antibody is species-independent and has been validated by Proteintech in western-blot to detect a 50 kDa 6xHis- tagged fusion protein (20 μg/lane) at dilution folds from 1/5000 to 1/160000. GST Monoclonal Antibody (Mix) is species-independent and has been validated in western blot to detect GST fusion protein, diluted at 1/10000. October 2018