The PAMP C-Di-AMP Is Essential for Listeria Monocytogenes Growth in Rich but Not Minimal Media Due to a Toxic Increase in (P)Ppgpp

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Article

The PAMP c-di-AMP Is Essential for Listeria monocytogenes Growth in Rich but Not Minimal Media due to a Toxic Increase in (p)ppGpp

Graphical Abstract Authors Aaron T. Whiteley, Alex J. Pollock, Daniel A. Portnoy

Correspondence [email protected]

In Brief c-di-AMP is a nucleotide second messenger recognized by the innate immune system and is predicted to be essential in many bacteria. Whiteley et al. show that depletion of c-di-AMP within Listeria monocytogenes leads to accumulation of (p)ppGpp, which inhibits bacterial growth but is also necessary for virulence.

Highlights d c-di-AMP is predicted to be essential in many pathogens and is thus an ideal PAMP

d DdacA mutants lacking c-di-AMP contain suppressor mutations within a (p)ppGpp synthase

d Decreased c-di-AMP increases (p)ppGpp, which inhibits growth via inactivation of CodY

d In minimal medium, which favors CodY inactivation, dacA is no longer essential

Whiteley et al., 2015, Cell Host & Microbe 17, 788–798 June 10, 2015 ª2015 Elsevier Inc. http://dx.doi.org/10.1016/j.chom.2015.05.006 Cell Host & Microbe Article

The PAMP c-di-AMP Is Essential for Listeria monocytogenes Growth in Rich but Not Minimal MediaduetoaToxicIncreasein(p)ppGpp

Aaron T. Whiteley,1 Alex J. Pollock,2 and Daniel A. Portnoy1,2,* 1Graduate Group in Infectious Diseases and Immunity, School of Public Health 2Department of Molecular and Cell Biology University of California, Berkeley, Berkeley, CA 94720, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.chom.2015.05.006

SUMMARY required for the expression of many determinants of pathogenesis, but it also contributes to metabolic adaptations. For Cyclic di-adenosine monophosphate (c-di-AMP) is instance, transcription of hpt, encoding a hexose phosphate a widely distributed second messenger that ap- transporter, is PrfA regulated, allowing growth on glucose-1- pears to be essential in multiple bacterial species, phosphate in the host cell cytosol (Chico-Calero et al., 2002). including the Gram-positive facultative intracellular L. monocytogenes also uses non-PrfA-mediated mechanisms pathogen Listeria monocytogenes. In this study, the for remodeling its metabolism. For example, pyruvate carbox- only L. monocytogenes diadenylate cyclase gene, ylase (PycA) is dispensable in rich medium but required for pathogenesis and is regulated allosterically by cyclic di-adenosine dacA, was deleted using a Cre-lox system activated D monophosphate (c-di-AMP) (Scha¨ r et al., 2010; Sureka et al., during infection of cultured macrophages. All dacA 2014). strains recovered from infected cells harbored one c-di-AMP is a member of the cyclic dinucleotide family of sec- or more suppressor mutations that allowed growth ond messengers that includes cyclic di-guanosine monophos- in the absence of c-di-AMP. Suppressor mutations phate (c-di-GMP) and cyclic AMP-GMP produced by bacteria in the synthase domain of the bi-functional (p)ppGpp and cyclic [G(20–50)pA(30–50)p] (cGAMP) produced by some meta- synthase/hydrolase led to reduced (p)ppGpp levels. A zoans (Danilchanka and Mekalanos, 2013). The role of c-di-AMP genetic assay confirmed that dacA was essential in during infection was identified as a result of biochemical charac- wild-type but not strains lacking all three (p)ppGpp terization of L. monocytogenes mutants that triggered diminished synthases. Further genetic analysis suggested that or enhanced activation of the cytosolic surveillance pathway c-di-AMP was essential because accumulated (CSP) (Crimmins et al., 2008; Woodward et al., 2010). The CSP is characterized by the robust induction of host type I interferon (p)ppGpp altered GTP concentrations, thereby inacti- and has implications for both innate and adaptive immunity vating the pleiotropic transcriptional regulator CodY. (Archer et al., 2014; O’Riordan et al., 2002). L. monocytogenes We propose that c-di-AMP is conditionally essential secretes c-di-AMP through multidrug efflux pumps; however, for metabolic changes that occur in growth in rich me- the effect of secreted c-di-AMP on the bacterium is not known dium and host cells but not minimal medium. and remains an active area of investigation (Kaplan Zeevi et al., 2013; Tadmor et al., 2014). c-di-AMP differs from c-di-GMP, the most extensively characterized bacterial cyclic dinucleotide, in INTRODUCTION that bacteria usually encode a single diadenylate cyclase (DAC) that is often essential (Corrigan and Gru¨ ndling, 2013). The DAC Listeria monocytogenes is a hardy and ubiquitous Gram-posi- protein domain (Pfam: DisA_N, PF02457) is the only identified pro- tive, facultative intracellular, foodborne pathogen that thrives tein domain capable of c-di-AMP synthesis in vivo and is widely as an environmental saprophyte, yet is capable of causing distributed among Archaea, Gram-positive bacteria, and some serious, often fatal, disease in a wide range of animals including Gram-negative bacteria (Witte et al., 2008). L. monocytogenes en- humans (Cossart, 2011). Its broad growth range is accompanied codes only one DAC, dacA, which cannot be deleted by conven- by a largely prototrophic metabolism, capable of utilizing an tional methods and is therefore also predicted to be essential array of carbon and nitrogen sources and requiring only a few (Witte et al., 2013). In addition, high-throughput and targeted essential vitamins and amino acids (Tsai and Hodgson, 2003). studies have identified c-di-AMP as essential in Bacillus subtilis, Remarkably, L. monocytogenes grows in the cytosol of host cells Staphylococcus aureus, Streptococcus pyogenes, Strepto- at a similar rate to rich medium (doubling in approximately coccus pneumoniae, Mycoplasma genitalium, and Mycoplasma 40 min) (Joseph and Goebel, 2007). To accomplish such eff- pulmonis (Corrigan and Gru¨ ndling, 2013). icient intracellular growth, L. monocytogenes remodels its tran- Many c-di-AMP-associated phenotypes have been observed scriptional program upon entering host cells by activation of in bacterial mutants containing inactivating mutations in c-di- the master virulence regulator PrfA (Freitag et al., 2009). PrfA is AMP-degrading phosphodiesterases. Genetic screens in multiple

788 Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. A L. monocytogenes dacAfl P -cre B Figure 1. Cre-lox Deletion of dacA and actA DdacA Suppressor Mutations (A) Schematic representation of the 10403S dacA lmo2119 L. monocytogenes strain used to delete dacA. The Genome loxP loxP 1 DdacA mutants were recovered after cre was

P 600 induced during infection of BMDMs from a Ts actA WT

OD D Ts cre dacAfl plasmid that could be cured from dacA mutants. Plasmid dacA.1 (B) BHI broth growth curve of wild-type, floxed- dacA.2 fl D 0.1 dacA.3 dacA (dacA ), and dacA mutants (for genotypes dacA.4 see Table S1). Data are representative of three dacA.5 independent experiments. In vitro ∆dacA mutant (C) Sensitivity to the toxic tripeptide bialaphos as 0246810 BMDM infection recovery measured by disk diffusion. Data are mean ± SEM Time (hr) of three independent experiments. p-oppB rep- C D 1.76% relA resents oppB complemented under its native 3.87% other ) promoters. The dotted line represents the L.o.D. 2 2.0 and *p % 0.01 by two-tailed Student’s t test as Total=284 compared to wild-type. 1.5 (D) Frequency of opp and relA suppressor muta- D 94.37% opp tions within a collection of 284 dacA mutants. 1.0 (E) Depiction of RelA protein with conserved ***** hydrolase, synthase, ThrRS-GTPase-SpoT (TGS), 0.5 RelA and Aspartokinase-Chorismate-mutase-TyrA (ACT) E domains. Annotation of specific amino acid Zone of ClearenceZone (cm 0.0 N Hydrolase Synthase TGS ACT C changes as a result of suppressor mutations: R295S WT- p-oppB 12345 was the first identified in DdacA.3 and R216P was R216P R295S Q326K ∆oppB ∆dacA (2) E324V identified twice.

organisms established that inactivating mutations in homologs tive method to delete dacA based on an inducible Cre-lox system of the conserved phosphodiesterase gdpP increase intracellular (Reniere et al., 2015). loxP sites were inserted into the c-di-AMP levels, increase resistance to acid stress, suppress mu- L. monocytogenes chromosome flanking dacA (dacAfl). Codon- tations in lipotechoic acid biosynthesis, and increase resistance optimized cre recombinase was expressed from the actA to b-lactam antibiotics (Corrigan et al., 2011; Luo and Helmann, promoter (PactA-cre) and cloned into a temperature-sensitive 2012; Rallu et al., 2000; Witte et al., 2013). Likewise, bacterial mu- (Ts) plasmid. The actA promoter was chosen because it is tants depleted for DAC expression exhibit increased sensitivity to not expressed in broth but is highly active during growth b-lactam antibiotics (Mehne et al., 2013; Witte et al., 2013). A inside mammalian cells (Shetron-Rama et al., 2002). The dacAfl diverse set of proteins interact with c-di-AMP, and a conserved PactA-cre strain grew normally in broth but resulted in deletion c-di-AMP interacting riboswitch regulates translation of a wide of dacA upon infection of cultured macrophages (Figure 1A). array of genes in many organisms (Corrigan et al., 2013; Nelson Wild-type bacteria from infected macrophages formed colonies et al., 2013; Sureka et al., 2014). However, none of the identified on rich medium agar in approximately 14 hr, whereas DdacA mu- c-di-AMP receptors are conserved among all c-di-AMP-produc- tants formed visible colonies between days 2 and 5. The DdacA ing organisms, despite conservation of many c-di-AMP-related mutants, cured of the cre expressing plasmid, were verified by phenotypes (Corrigan and Gru¨ ndling, 2013). Here we report the PCR using primers internal to the dacA gene and external to the isolation and characterization of suppressor mutations that allow dacA-locus. L. monocytogenes to grow in the absence of c-di-AMP. Our find- Five mutants, numbered DdacA.1–DdacA.5, were chosen for ings may help unify divergent c-di-AMP-related phenotypes and initial characterization. As expected, these mutants grew poorly support a model in which L. monocytogenes requires intracellular in brain-heart infusion (BHI) broth, a rich medium commonly c-di-AMP for metabolic adaptations during growth in rich medium used for cultivating L. monocytogenes (Figure 1B). We hypo- and in host cells. thesized that dacA was essential but that these DdacA strains contained suppressor mutations that bypassed the RESULTS essential functions of c-di-AMP. Genome sequencing of strains DdacA.1–DdacA.5 and the parent dacAfl confirmed that dacA Generation of L. monocytogenes DdacA Mutants and was absent and revealed two groups of mutations not found in Identification of Suppressor Mutations the parent strain (Table S1). Four strains contained mutations There is mounting evidence that c-di-AMP is an essential mole- in the 5-gene operon oppABCDF encoding a previously identi- cule in many Firmicutes, including L. monocytogenes (Corrigan fied oligopeptide permease (Opp) (Borezee et al., 2000). These and Gru¨ ndling, 2013). Accordingly, we were unable to generate four mutants displayed decreased sensitivity to killing by the DdacA mutants in wild-type L. monocytogenes but were suc- toxic tripeptide bialaphos that is transported exclusively by cessful in generating a dacA deletion in a strain that contained the Opp (Figure 1C) (Borezee et al., 2000). The only strain without a second copy of dacA expressed from an inducible promoter an opp mutation (DdacA.3) encoded a point mutation (R295S) in (Witte et al., 2013; Woodward et al., 2010). We sought an alterna- the synthase domain of the previously identified bi-functional

Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. 789 A 16.6 kb B dacA Figure 2. dacA Essentiality Assay 80 ∆dacA (A) Schematic of dacA essentiality assay depict- dacA ermRdacA kanR ing erythromycin resistance genes (ermR), kana- 0 60 mycin resistance genes (kanR)3 of dacA and DdacA, erythromycin (Erm)/kanamycin (Kan)- ∆dacA ermR∆ kanR containing medium agar, and resistant trans- 40 ductants (ermR/kanR). See text for description. (B) Genetic linkage of dacA or DdacA with co- Linkage (%) 20 transduced antibiotic resistance marker in wild- Listeria Phage transduce type or mutants merodiploid for dacA (p-dacA). Recipient marked dacA locus <1% D 0 (C) dacA essentiality assay of dacA suppressor D R295S WT WT p-dacA mutants, dacA.3 (encoding a relA mutation) was not determined due to an inability to obtain C dacA R Patch 80 erm transductants in this background. The Erm Kan ∆dacA R295S Colonies relA mutation was interrogated instead by reconstructing the mutation in a wild-type back- 60 ground. Dotted line indicates L.o.D., and all data # of kanRermR colonies are mean ± SEM of at least three independent = Genetic linkage (%) 40 experiments. See also Figure S1. # of Total ermR colonies Linkage (%) 20 dacA dacA linkage ∆dacA linkage ≠ essential 0 <1% mycin resistance gene (kanR)3 to the 0 dacA R295S dacA locus. A himar1 transposon (ermR) dacA linkage ∆dacA linkage WT 1 2 4 5 relA ≈ not essential encoding an erythromycin resistance gene was present 16.6 kb 50 of the dacA locus (Figure 2A). The assay was per- guanosine penta- and tetraphosphate ((p)ppGpp) synthase/hy- formed by (1) lysogenizing recipient L. monocytogenes strains drolase relA (Bennett et al., 2007; Taylor et al., 2002) and with phage derived from either of the two drug-resistant donor remained bialaphos sensitive (Figures 1C and 1E). These data strains, (2) selecting for transduction on erythromycin, and (3) indicated that the opp nucleotide changes were loss-of-function analyzing genetic linkage by scoring transductants for kana- mutations, consistent with the disruptive nature of these poly- mycin resistance (Figure 2A). In a wild-type recipient, the morphisms (frameshifts and a premature stop codon) and that dacA-kanR allele displayed approximately 35% linkage with the relA mutation in the DdacA.3 strain did not affect Opp the ermR gene, while co-transduction of the DdacA-kanR allele activity. was below the limit of detection (L.o.D.) (Figure 2B). This linkage To further investigate dacA essentiality, an additional 284 disequilibrium was ameliorated in recipients merodiploid for DdacA mutants were selected for characterization. All of the dacA (WT p-dacA), indicating the difference in linkage between DdacA strains isolated encoded mutations. Of these mutants, dacA-kanR and DdacA-kanR alleles was specific to deletion of 94.37% were resistant to bialaphos, suggesting mutations in the dacA gene (Figure 2B). Moreover, the genetic linkage anal- opp genes, and 1.76% harbored relA mutations as determined ysis produced similar results when performed with an alternative by Sanger sequencing of the synthase domain of the relA gene himar1 transposon 10 kb 50 of the dacA locus, demonstrating (Figure 1D). The additional relA alleles identified are depicted in that the location of the himar1 transposon had no effect on the Figure 1E. Genome sequencing of the remaining 11 mutants, results (Figure S1). DdacA.6–DdacA.16, that were sensitive to bialaphos and did The DdacA phage lysate used for the essentiality assay was not harbor relA synthase domain mutations revealed that each derived from a strain that presumably harbors suppressor muta- strain contained more than one mutation (Table S1). These addi- tions as a result of the Cre-lox-mediated deletion of dacA. Given tional mutations often recurred in the same genes (three of which the results of the linkage experiments, these suppressor muta- share cystathionine b-synthase [CBS] domains), appeared in tions were not linked to the dacA locus. To further assess the genes encoding identified c-di-AMP binding proteins (Sureka impact of suppressor mutations, DdacA.1–5 mutants generated et al., 2014) and included opp and relA mutations that escaped through Cre-lox recombination were subjected to the identical detection (Table S1). Characterization of opp and other suppres- linkage analysis described above (Figure 2A). Linkage was sor mutations will be the subject of future studies. unaffected in DdacA.1, 2, 4, and 5, verifying the existence of suppressor mutations and that the dacA gene was no longer Development of a dacA Essentiality Assay essential in these strains (Figure 2C). Linkage analysis was not The above results suggested that dacA was essential and that determined for DdacA.3 (encoding a relAR295S mutation) owing each of the DdacA mutants had accumulated one or more sup- to an inability to obtain erythromycin-resistant transductants in pressor mutations. To expand upon these studies, we developed this background for an unknown reason. This limitation was a rapid essentiality assay based on the method of co-transduc- overcome by reconstructing the relAR295S mutation in a wild- tion of linked genetic markers used to show gene essentiality type background. dacA was also not essential in this back- in Escherichia coli (De Las Pen˜ as et al., 1997). Two donor ground, confirming the relAR295S mutation suppresses dacA L. monocytogenes strains were constructed, dacA and DdacA essentiality (Figure 2C). These data were consistent with dacA using Cre-lox, in which the dacA locus was marked by a kana- being essential to wild-type L. monocytogenes and established

790 Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. A Figure 3. DdacA Suppressor Mutations in relA Affect Starvation-Induced (p)ppGpp GTP TLC of 32P-labeled intracellular nucleotides from bacterial mutants. Bacteria were grown in low- phosphate defined medium plus tryptone, and ppGpp where indicated, starvation was induced using SHX. Wild-type E. coli (CF1943) and relA251::kan E. coli (CF1944) are included as controls; all other pppGpp strain are L. monocytogenes. (A–C) A representative TLC is shown (A), and quantification of the ratio ðpppGpp + ppGppÞ= Origin ðpppGpp + ppGpp + GTPÞ as a percent of wild- type is shown for E. coli (B) and L. monocytogenes SHX ---+ - + - + - + (C). All data are representative of n = 10 indepen- WT relA251::kan WTrelAR295S WT ∆relA dent experiments, graphed data are mean ± SEM E. coli E. coli of pooled data, *p % 0.05 by two-tailed Students t test, and ns denotes not significant (p > 0.05). BC- SHX +SHX * * 300 - SHX +SHX 150

oli lor et al., 2002). Amino acid starvation c ns

200 T was simulated using serine hydroxamate E.

ns W 100 (SHX) and (p)ppGpp was quantiﬁed as a f proportion of (p)ppGpp + GTP levels. WT f

%o Control experiments demonstrated that o 100 50 wild-type E. coli (CF1943) accumulated % (p)ppGpp in response to starvation while E. coli carrying a disrupted relA gene 0 0 (CF1944) did not (Figures 3A and 3B). WT relA251::kan WTrelAR295S ∆relA Wild-type L. monocytogenes also accumulated (p)ppGpp in response to starvation; however, mutants expressing relAR295S did not (Figures 3A and 3C), an assay whereby comparison of dacA and DdacA genetic link- supporting the supposition that this mutation disrupted RelA age is a measure of dacA essentiality. synthase activity. In related Firmicutes, two proteins (RelP and RelQ) in addition Suppressor Mutations in relA Decreased (p)ppGpp to RelA are capable of synthesizing (p)ppGpp, although RelA is Accumulation in Response to Starvation the only synthase predicted to respond to starvation (Nanamiya We chose to characterize the suppressor mutations in relA et al., 2008). The other two small alarmone synthases, identified because these mutations appeared sufficient to ablate dacA es- here as relP (lmo0802) and relQ (lmo0967) based on their homol- sentiality and because of the previously documented nucleotide ogy to B. subtilis, S. aureus, and Streptococcus mutans (Geiger cross-talk between c-di-AMP and (p)ppGpp (Corrigan et al., et al., 2014; Lemos et al., 2007; Nanamiya et al., 2008), were 2015; Rao et al., 2010; Sureka et al., 2014). The DdacA suppres- likely responsible for the basal levels of (p)ppGpp observed in sor mutations in relA clustered within or near the synthase the untreated condition (Figures 3A and 3C). RelA is unique domain of the RelA protein (Figure 1E). RelA (encoded by a ho- because it is the only identified (p)ppGpp hydrolase in molog of the E. coli relA/spoT gene) synthesizes (p)ppGpp in L. monocytogenes. The hydrolase function of RelA was revealed response to starvation during the ‘‘stringent response’’ by trans- by increased (p)ppGpp levels in a DrelA mutant as compared to ferring two phosphates from ATP to either GTP or GDP to pro- wild-type (Figures 3A and 3C). The difference between the levels duce pppGpp or ppGpp, respectively (collectively referred to of (p)ppGpp in the untreated DrelA and relAR295S strains was as (p)ppGpp). In Firmicutes, RelA is also a hydrolase that de- therefore due to the functional hydrolase component of the grades (p)ppGpp when nutrients are abundant (Mechold et al., RelAR295S protein, which can degrade (p)ppGpp synthesized 1996). The hydrolase and synthase enzymatic activities can be by RelP and RelQ. These data demonstrated that the suppressor separated by point mutations in their respective domains mutation in relA encoded a hydrolase-only form of the protein. (Hogg et al., 2004). The impact of the relAR295S suppressor mutation was interrogated by reconstructing the mutation in Accumulation of (p)ppGpp Is Toxic to DdacA Mutants the chromosome of wild-type L. monocytogenes and measuring In Firmicutes, (p)ppGpp inhibits DNA primase and enzymes that 32P-labeled intracellular nucleotides by thin-layer chromatog- catalyze GTP synthesis (Kriel et al., 2012; Wang et al., 2007). The raphy (TLC). These experiments were performed in low-phos- net effect of increased (p)ppGpp is both a transcriptional and phate defined medium supplemented with tryptone, which translational response that results in a decreased growth rate mimicked rich medium and stimulated uptake of added 32P(Tay- (Dalebroux and Swanson, 2012). We hypothesized that the

Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. 791 ACB Figure 4. (p)ppGpp Accumulates during ns dacA 200 * 80 Depletion of c-di-AMP Leading to dacA GTP ∆dacA Essentiality and Decreased Growth Rate 32 150 (A and B) TLC analysis of P-labeled intracellular 60 ppGpp nucleotides from bacteria grown in low-phosphate defined medium plus tryptone without starvation. 100 40 Chromosomal dacA was deleted in a strain har- boring an IPTG-inducible dacA gene to construct %ofWT pppGpp Linkage (%) a conditional dacA depletion strain (cDdacA) 50 20 (Witte et al., 2013). A representative TLC is shown (A), and quantification of the ratio <1% <1% <1% <1% Origin 0 0 ðpppGpp + ppGppÞ=ðpppGpp + ppGpp + GTPÞ is WT c∆dacA c∆dacA WT - p-relA p-relP p-relQ shown as a percent of wild-type (B). Data are WT c∆dacA c∆dacA +IPTG -IPTG +IPTG -IPTG ∆relAPQ representative of n = 11 independent experiments, D E graphed data are mean ± SEM of pooled data, *p % 0.05 by two-tailed Students t test, and ns denotes not significant. (C) dacA essentiality assay. Complemented genes 1 1 indicated by (p-) were introduced at a neutral site 600 600 using their native promoter. Dotted line indicates L.o.D.; data are mean ± SEM of at least three in- OD OD WT dependent experiments. WT c∆dacA(+) 0.1 (D and E) BHI broth growth curves, with or without ∆relAPQ 0.1 c∆dacA(-) c∆dacA(+) ∆relAPQ (+/) IPTG in (E). Data are representative of three ∆dacA∆relAPQ c∆dacA(-) ∆relAPQ independent experiments. 02468 0246810

relAR295S mutation suppressed dacA essentiality by decreasing conditional dacA depletion strains cDdacADrelAPQ and cDdacA (p)ppGpp that may have accumulated as a consequence of dele- grew similarly to wild-type. In the absence of IPTG (when dacA is tion of dacA. We tested this hypothesis by first measuring depleted), the cDdacADrelAPQ strain displayed an increased (p)ppGpp under non-starvation conditions in a dacA conditional growth rate compared to the cDdacA strain (Figure 4E). These depletion strain (cDdacA), which expressed dacA under the data are consistent with a role for c-di-AMP in maintaining low control of an IPTG inducible promoter (Witte et al., 2013). In com- (p)ppGpp levels that are otherwise detrimental for growth. parison to wild-type, conditional depletion of dacA led to an increase in (p)ppGpp levels in non-starvation conditions (Figures A Screen for Mutations that Rescue the Virulence Defect 4A and 4B). of a DrelAPQ Mutant Reveals a Critical Role for CodY To further evaluate the role of increased (p)ppGpp in dacA es- We next sought to understand the function(s) of (p)ppGpp in sentiality, we constructed a L. monocytogenes strain lacking L. monocytogenes. relA mutants are attenuated for pathogen- (p)ppGpp by sequentially deleting the relP, relQ, and relA genes esis (Bennett et al., 2007), although the role of (p)ppGpp in (DrelAPQ) and subjected this strain to the dacA essentiality infection is still unclear, since our data indicated that relA mu- assay in BHI (Figure 2A). The dacA gene was no longer essential tants have elevated levels of (p)ppGpp (Figures 3A and 3C). in the DrelAPQ background (Figure 4C). Complementation of The DrelAPQ mutant grew similarly to wild-type in rich medium, DrelAPQ with any of the three (p)ppGpp synthases using their despite lacking all sources of (p)ppGpp (Figure 4D); however, it native promoters restored the essentiality of the dacA gene was severely attenuated in a plaque assay, an in vitro infection (Figure 4C). These data indicated that dacA was essential due model that serves as a surrogate for virulence (Figure 5A). In to accumulation of the nucleotide (p)ppGpp rather than an this assay, confluent mammalian fibroblasts are infected with interaction with any single (p)ppGpp synthase. Additionally, L. monocytogenes, and intracellular growth and cell-to-cell relA was sufficient to render dacA essential, which suggested spread of the bacteria produce a zone of clearance (plaque) that although RelA is a bifunctional synthase/hydrolase, in the that is quantifiable and high-throughput (Sun et al., 1990). absence of c-di-AMP, RelA functioned as a synthase. Small plaques often correlate to virulence defects in vivo, and The dacA gene was not essential in a DrelAPQ background, the DrelAPQ strain produced small plaques that were 30% the although the DdacADrelAPQ mutant grew slowly compared to area of wild-type plaques (Figure 5A). wild-type (Figure 4D). These data established a role for (p)ppGpp These results suggested that (p)ppGpp was necessary for a in dacA essentiality. Additionally, we hypothesized that the accu- productive infection. We speculated that the contribution of mulation of (p)ppGpp observed after depletion of dacA (Figures (p)ppGpp to virulence and toxicity to DdacA mutants were related 4A and 4B) might be partially responsible for the growth defect and performed a transposon mutagenesis screen for mutations of the cDdacA strain (Witte et al., 2013). This hypothesis that rescued the small plaque phenotype of the DrelAPQ mutant. was tested by measuring growth rate of a conditional dacA We identified 98 mutants from over 10,000-screened that dis- depletion strain constructed in a wild-type or DrelAPQ back- played increased plaque size. DNA sequencing of the region ground (cDdacADrelAPQ) in BHI. In the presence of IPTG, the adjacent to the transposon insertions and phage transduction

792 Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. A B * Figure 5. A Screen for Suppressor Muta- 109 tions of the DrelAPQ Virulence Defect Re- 100 * * * veals a Critical Role for Inactivation of CodY 108 (A) A murine ﬁbroblast (L2 cell) monolayer was 80 107 infected with L. monocytogenes mutants from Table 1, producing plaques. The reduced plaque 6 10 area correlates with reduction in virulence. Data 60 5 are mean ± SEM of pooled data from least three

Spleen 10

/ independent experiments; the dotted line repre- 40 104 sents DrelAPQ plaque area for reference; P pre- CFU ceding a gene name indicates the transposon is in que Area (% of WT) 103 a l the promoter; *p % 0.05 by one-way ANOVA

P 20 102 compared with DrelAPQ. (B) Recovered CFU at 48 hr post-infection from 0 10 CD-1 mice intravenously infected via tail vein with 8 8 7 4 5 T Q Y 7 6 rR 43 WTcodY::spc - p-relA codY::spc 105 CFU of each indicated strain. Data are pooled W cod pu rsbX spoVGguaB2 relAP P P lmo000 P lmo0974 lmo080lmo0948 lmo021lmo07lmo188 lmo211 lmo18 ∆relAPQ results from at least two independent experiments, bars indicate median value, p-relA represents relA complemented under its native promoter, and *p % 0.05 using two-tailed Student’s t test. led to the identification of 14 genes, that when disrupted, signifi- is consistent with previous reports demonstrating a virulence cantly increased the DrelAPQ plaque size (Figure 5A; Table 1). defect for a codY mutant (Lobel et al., 2014; 2012), and muta- Transposon insertions in codY produced the most significant tions in codY suppress the virulence defect of other pathogenic increase in the DrelAPQ plaque size (Figure 5A; Table 1). CodY Firmicutes with decreased (p)ppGpp (Geiger and Wolz, 2014). is activated by high GTP levels and branch-chain amino Our data suggested that the principle role of (p)ppGpp during acids (BCAAs) to promote DNA binding that transcriptionally re- infection was the inhibition of GTP synthesis leading to inactiva- presses a large regulon of genes, but also is capable of transcription of CodY. tional activation of a few genes (Geiger and Wolz, 2014). In B. subtilis, (p)ppGpp inhibits GTP synthesis at multiple enzy- (p)ppGpp-Dependent Inactivation of CodY Is Necessary matic steps and the subsequent decrease in GTP leads to for the Essentiality of dacA CodY deactivation (Kriel et al., 2012). The DrelAPQ mutant is We speculated that the function of (p)ppGpp during infection incapable of modulating GTP levels via (p)ppGpp, and thus overlaps with the role of (p)ppGpp in dacA essentiality. Accord- CodY remains constitutively activated. In L. monocytogenes, ingly, dacA may not be essential in the DrelAPQ mutant, because codY mutations likely rescue the plaque defect of the DrelAPQ in the absence of (p)ppGpp, GTP remains elevated, and CodY is mutant by phenocopying (p)ppGpp-dependent inhibition of highly active. We examined the role of CodY in the essentiality of GTP synthesis that takes place in wild-type bacteria and dacA by comparing the DrelAPQ and DrelAPQ codY::spc mu- demonstrates a critical role for CodY deactivation during infec- tants in the genetic assay for dacA essentiality (Figure 2A). While tion. Other identified suppressor mutations (such as guaB2, dacA was not essential in the DrelAPQ mutant, addition of a lmo1884, and purR) in purine nucleotide synthesis/acquisition codY mutation returned dacA to its original essential phenotype might have affected CodY activity by modulating intracellular (Figure 6A). Addition of the spoVG mutation to the DrelAPQ GTP levels. We identified which mutations were epistatic to mutant strain did not alter dacA essentiality (Figure 6A). These re- mutations in codY by constructing a marked deletion in codY sults suggested that among the diverse functions of (p)ppGpp, (codY::spc) and transducing the isolated transposons into the inactivation of GTP synthesis and thus inactivation of CodY DrelAPQ codY::spc background (Table 1). Only mutations in was selectively toxic to DdacA mutants. Further, these findings the promoter of spoVG and lmo0948 further increased the imply that elements of the CodY regulon, which are necessary DrelAPQ codY::spc plaque size. However, lmo0948::himar1 for infection, may be toxic in the absence of c-di-AMP. also increased the plaque area of wild-type and thus was likely not specific to (p)ppGpp (Table 1). Although spoVG is regulated dacA Is Not Essential in Minimal Medium by (p)ppGpp in B. subtilis (Tagami et al., 2012) and has been We hypothesized that dacA might no longer be essential in identified with divergent phenotypes in multiple organisms (Ju- growth conditions that favored inactivation of CodY. The best tras et al., 2013; Matsuno and Sonenshein, 1999; Meier et al., example of such a condition is in minimal medium, where a 2007), it is unclear how this mutation contributed to the virulence B. subtilis strain unable to produce (p)ppGpp cannot grow of the DrelAPQ mutant. without a codY mutation (Kriel et al., 2012; 2014). Similarly in In a mouse model of infection, the DrelAPQ mutant was L. monocytogenes, the DrelAPQ mutant does not grow on mini- approximately 100-fold less virulent compared to wild-type or mal medium (data not shown), prompting us to examine these the DrelAPQ strain complemented with relA under its native growth conditions. Unlike rich medium, dacA was no longer promoter (Figure 5B). The codY::spc mutation suppressed the essential in a defined minimal medium (Figure 6B) (Phan-Thanh virulence defect of the DrelAPQ strain to the level of a codY::spc and Gormon, 1997). Remarkably, in-frame DdacA deletions mutation alone, approximately 10-fold less virulent than wild- were readily obtainable by allelic exchange when bacteria were type (Figure 5B). The virulence defect of the codY::spc mutant cultivated in minimal medium. Genome sequencing of DdacA

Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. 793 Table 1. Transposon Mutations that Suppress the DrelAPQ Plaque Defect Plaque Area (%WT ± SEM) Gene Namea Annotationb himar1 Locationc DrelAPQd DrelAPQ codY::spcd wild-typed codY (lmo1280) GTP-responsive transcriptional regulator LMRG_00730::119 84 ± 2.3* N/A 89 ± 1.2*

PspoVG (Plmo0196) Promoter of genes similar to spoVG LMRG_02618::-144 72 ± 1.6* 89 ± 1.6* 105 ± 1.6 guaB2 (lmo2758) Inosine-5’-monophosphate (IMP) dehydrogenase LMRG_01938::220 71 ± 1.0* 77 ± 1.0 93 ± 1.4 lmo0808 Spermidine/putrescine ABC transporter LMRG_02789::1593 49 ± 0.9* 83 ± 2.3 108 ± 1.6* lmo0948 GntR family transcriptional regulator LMRG_02047::494 45 ± 1.5* 84 ± 1.2* 111 ± 2.4* lmo0006/7 Between DNA Gyrase subunits B/A LMRG_02435::-26 44 ± 1.2* 70 ± 1.1* 99 ± 1.3 lmo0218 S1 RNA binding domain protein similar to yabR LMRG_02640::121 43 ± 1.2* 72 ± 2.6 102 ± 2.7 lmo0767 Sugar ABC transporter permease LMRG_00455::222 42 ± 1.4* 80 ± 0.9 100 ± 1.3 lmo1884 Xanthine uptake transporter similar to pbuX LMRG_01031::319 41 ± 0.7* 77 ± 1.5 107 ± 2.9

Plmo0974 Promoter of D-alanine-poly(phosphoribitol) ligase LMRG_02073::-223 41 ± 1.1* 77 ± 1.6 100 ± 1.9 (dlt) operon purR (lmo0192) Purine associated transcriptional repressor LMRG_02614::471 39 ± 1.8* 70 ± 1.5* 93 ± 1.6 anrB (lmo2115) FtsX family ABC transporter permease associated LMRG_01269::1842 39 ± 1.1* 80 ± 1.9 104 ± 2.4 with nisin resistance rsbX Negative regulator of sigma-B (serine LMRG_02320::2103 38 ± 1.5* 59 ± 0.9* 72 ± 1.2* (lmo0896) phosphatase) lmo1843 RluA family 23S pseudouridylate synthase LMRG_00990::876 37 ± 0.9* 80 ± 1.1 96 ± 2.0 aAnnotated using EGD-e ordered loci and previously published name where appropriate. P indicates the transposon location is within a predicted promoter of the annotated gene. bGene similarity based on Bacillus subtilis genome annotation. cSequence-mapped transposon insertion site (10403S ordered genetic locus::nucleotides 30 of ORF start codon). dGenetic background of the transposon mutant; data represent the mean ± SEM for at least three independent experiments; asterisks (*) indicate plaque area was signiﬁcantly different from background strain (p % 0.05 by one-way ANOVA, Tukey test).

mutants constructed on minimal medium using a marked dacA genes (DrelAPQ) suppressed dacA essentiality. Although the deletion confirmed the absence of suppressor mutations. These growth defect of the DdacADrelAPQ strain implies additional data suggested a model in which c-di-AMP is essential for roles for c-di-AMP in bacterial physiology, these data suggested growth in rich medium, because in the absence of c-di-AMP, that DdacA mutants failed to grow in rich medium and in cells (p)ppGpp accumulated and indirectly inactivated CodY, which because (p)ppGpp levels were elevated, a condition known to facilitated transcriptional changes selectively toxic to DdacA mu- inhibit bacterial growth (Dalebroux and Swanson, 2012). tants. This work identified that mutations that decreased We reasoned that there might be an overlap between the (p)ppGpp or replacement with a medium favoring CodY inactiva- (p)ppGpp-regulated genes required for pathogenesis and those tion were sufficient to reverse the essentiality of dacA. that suppressed dacA essentiality. Therefore, we screened for transposon mutations that suppressed the plaque defect of a DISCUSSION strain lacking (p)ppGpp (DrelAPQ), which was nearly 100-fold less virulent in mice. Fourteen loci were identified and trans- There is an emerging consensus that c-di-AMP is an essential poson mutations disrupting codY produced the largest plaque. molecule in Firmicutes (Corrigan and Gru¨ ndling, 2013). However, Mutations in codY restored the virulence of L. monocytogenes here we report the isolation of L. monocytogenes dacA deletion strains lacking (p)ppGpp to that of the codY mutant alone. mutants that lack c-di-AMP. As conventional approaches of CodY is a GTP-responsive transcriptional regulator whose func- isolating mutations were unsuccessful, a strain was constructed tion is inextricably linked with (p)ppGpp levels (Geiger and Wolz, in which loxP sites were introduced into the L. monocytogenes 2014). During exponential growth in rich medium, CodY is GTP chromosome flanking the dacA gene, and Cre recombinase bound and represses dozens of biosynthetic operons (Sonen- was expressed from the PrfA-regulated actA promoter that is shein, 2007). CodY also enhances transcription of genes induced in host cells (Reniere et al., 2015). Upon infection of involved in GTP synthesis, most notably guaB, causing a feed- macrophages, Cre was expressed leading to deletion of dacA, forward regulatory loop that maintains CodY activation (Bennett thereby providing an unbiased selection for L. monocytogenes et al., 2007). During starvation, (p)ppGpp interrupts this feed- mutants able to grow in the absence of the c-di-AMP. All DdacA forward loop through allosteric inhibition of GTP-synthesis mutants isolated from the infected cells contained one or more enzymes, thereby allowing for the expression of many biosyn- suppressor mutations that promoted bacterial growth in rich me- thetic operons (Kriel et al., 2012). Other mutations that sup- dium. Mutations in the synthase domain of the bi-functional pressed virulence defects of the DrelAPQ mutant, such as RelA/SpoT homolog relA that retained (p)ppGpp hydrolase activ- guaB2 and lmo1884, were epistatic to codY because they likely ity or deletion of all three L. monocytogenes (p)ppGpp synthase recapitulate the role of (p)ppGpp, decreasing GTP abundance

794 Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. A dacA et al., 2010; Sureka et al., 2014). Depletion of c-di-AMP leads 80 to overactivity of PycA and thus increased levels of TCA cycle in- ∆dacA termediates, such as glutamate/glutamine, primarily because L. monocytogenes has an incomplete TCA cycle and lacks a-ke- 60 toglutarate dehydrogenase. Consequently, mutations in citrate synthase (citZ), the ﬁrst step of the TCA cycle, relieve the buildup of glutamate/glutamine and suppress the virulence defect of a conditional dacA depletion strain (Sureka et al., 2014). Interest- ge (%) 40 ingly, among our DdacA suppressor mutants that contained multiple mutations, we identiﬁed missense and promoter muta-

Linka tions in pycA and a premature stop codon in citZ, suggesting 20 that mutations which lower potentially toxic concentrations of glutamate/glutamine counter dacA essentiality. Although it is <1% 1% not clear how increased glutamate/glutamine levels might result 0 in toxicity, it may be noteworthy that in E. coli, glutamate func- + WT - ∆spoVG codY::spc tions as the principle counter-ion to K (McLaggan et al., 1990; 1994), which is an indispensible cation for balancing osmotic ∆relAPQ stress in bacteria (Epstein, 2003). In S. aureus, three of the four B identiﬁed c-di-AMP-binding proteins modulate intracellular po- dacA 80 tassium levels (Corrigan et al., 2013), and in a diverse set of or- ∆dacA ganisms, the c-di-AMP-binding ydaO-yuaA riboswitch regulates potassium transporters and osmoprotection genes (Nelson et al., 2013). We hypothesize that bacteria lacking c-di-AMP 60 accumulate both K+ and glutamate and are unable to regulate a subsequent lethal change in internal osmotic pressure. In support of this, we identiﬁed mutations in the glycine betaine/proline 40 osmoprotection transporter in some of our DdacA suppressors mutants that contained multiple mutations, although we were D

Linkage (%) unable to rescue dacA essentiality by adding osmoprotectants 20 or altering salt concentrations (data not shown). In minimal medium, dacA may not be essential, because glutamate is required <1% for synthesis of many additional metabolites made under nutrient 0 stress (Sonenshein, 2007). In addition, CodY overactivation (in the absence of (p)ppGpp) could remedy dacA essentiality by repressing glutamate synthase or altering L. monocytogenes metabolism to provide decreased PycA precursors (Brinsmade Figure 6. dacA Is Essential in Rich Medium Due to CodY Inactivation but dacA Is Not Essential in Minimal Medium et al., 2014; Sonenshein, 2007). (A and B) dacA essentiality assay. The results of this study suggest that there is a signaling loop (B) The commonly used rich medium is BHI, and minimal medium for between (p)ppGpp and c-di-AMP, which is not surprising L. monocytogenes is detailed in the Supplemental Experimental Procedures. since c-di-AMP-speciﬁc phosphodiesterases are inhibited by Dotted line indicates L.o.D.; data are mean ± SEM of at least three independent (p)ppGpp (Corrigan et al., 2015; Huynh et al., 2015; Rao et al., experiments. 2010). What is surprising is that deletion of c-di-AMP phosphodiesterases pgpH and gdpP in L. monocytogenes and and therefore the activity of CodY. Most importantly however, S. aureus, respectively, led to increased (p)ppGpp during stress the codY mutation restored the essentiality of dacA in a DrelAPQ despite containing elevated levels of c-di-AMP (Corrigan et al., mutant. Therefore, c-di-AMP essentiality is likely caused by one 2015; Liu et al., 2006), the opposite of the phenotype predicted or more CodY-regulated genes that are inappropriately ex- by work presented here. We hypothesize that both high and pressed in rich medium due to elevated levels of (p)ppGpp. If low c-di-AMP can contribute to increased (p)ppGpp by altering this were correct, one would predict that dacA might not be central metabolism and amino acid biosynthesis, speciﬁcally essential in conditions favoring expression of CodY-regulated the levels of the BCAAs valine, leucine, and isoleucine. (p)ppGpp genes, such as minimal medium where CodY-repressed genes production is stimulated by low BCAA levels, which are uniquely are essential for growth (Kriel et al., 2014). Indeed, we were poised as sensors of nutrient stress because they require precur- able to construct supressorless DdacA mutants in minimal me- sors from carbon, nitrogen, and sulfur metabolism for their syn- dium using conventional methods of allelic exchange. thesis (Somerville and Proctor, 2009). Valine and leucine are It is not clear why c-di-AMP is essential in rich but not derived from pyruvate, whereas isoleucine is derived from oxalo- minimal media. However, c-di-AMP is an allosteric inhibitor of acetate (Sonenshein, 2007). In L. monocytogenes, c-di-AMP L. monocytogenes PycA, an enzyme that catalyzes the conver- may affect the abundance of these precursors by regulating sion of pyruvate to oxaloacetate and entry of carbon into the PycA activity (Scha¨ r et al., 2010; Sureka et al., 2014). When over- TCA cycle and is essential for growth in minimal medium (Scha¨ r active (low c-di-AMP), PycA activity leads to pyruvate depletion

Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. 795 and potentially low levels of valine and leucine. When under- standard error of the mean (SEM) for at least three independent experiments. active (high c-di-AMP), the bacterium is depleted of oxaloace- For wild-type merodiploid for dacA, an IPTG-inducible dacA was introduced tate. This hypothesis is consistent with the demonstrated toxicity into wild-type, and the experiment was performed in the presence of IPTG. of excess c-di-AMP (Huynh et al., 2015; Mehne et al., 2013) and Virulence Analysis places c-di-AMP as a key regulator of metabolic homeostasis. In vivo virulence analysis was performed as previously described (Reniere Both c-di-AMP and (p)ppGpp contribute to bacterial stress re- et al., 2015) with the following changes: female, 8- to 12-week-old CD-1 sponses; for example, in related Firmicutes, mutations in relP mice (Charles River) were injected via tail vein with 200 ml of PBS containing and relQ increase antibiotic sensitivity while mutations in c-di- 105 CFU of L. monocytogenes. Mice were euthanized 48 hr post-infection, liver AMP-dependent phosphodiesterases lead to increased antibi- and spleen removed, organs homogenized in filter-sterilized 0.1% NP40, and otic resistance (Abranches et al., 2009; Corrigan and Gru¨ ndling, the CFU of the liver and spleen enumerated by plating serial dilutions on LB-Agar containing streptomycin. Statistical significance was determined by 2013; Geiger et al., 2014). Importantly, there are fundamental dif- a two-tailed heteroscedastic Student’s t test. This study was carried out in ferences between (p)ppGpp and c-di-AMP: whereas the former strict accordance with the recommendations in the Guide for the Care and is only made during acute stress, the latter appears to be present Use of Laboratory Animals of the National Institutes of Health. All protocols during all growth conditions (Corrigan and Gru¨ ndling, 2013). were reviewed and approved by the Animal Care and Use Committee at the Even in minimal medium, where c-di-AMP is not essential, University of California, Berkeley (MAUP# R235-0815B). DdacA mutants grew slowly compared to wild-type (data not For detailed explanations of bacterial strains and culture conditions, DNA shown). Another fundamental difference is that c-di-AMP is manipulations and strain construction, genome sequencing, disk diffusions, phage transduction, (p)ppGpp quantification, plaque assay, DrelAPQ viru- secreted by L. monocytogenes during growth in media and in lence suppressor screen, a table of L. monocytogenes strains, a table of plas- cells, perhaps altering intracellular nucleotide concentrations mids and E. coli strains, and a table of oligonucleotides used in this study, or regulating extracellular processes (Woodward et al., 2010). please see Supplemental Experimental Procedures. There is also evidence that Chlamydia trachomatis and Myco- bacterium tuberculosis are capable of secreting c-di-AMP during SUPPLEMENTAL INFORMATION infection (Barker et al., 2013; Yang et al., 2014). Collectively, these properties make c-di-AMP an ideal pathogen-associated Supplemental Information includes one figure, one table, and Supplemental molecular pattern (PAMP); i.e., it is highly expressed, conserved, Experimental Procedures and can be found with this article online at http:// dx.doi.org/10.1016/j.chom.2015.05.006. essential for virulence, and secreted, thereby triggering STING a central hub of host innate immunity (Danilchanka and Mekala- AUTHOR CONTRIBUTIONS nos, 2013; Vance et al., 2009). A.T.W. and A.J.P. performed experiments; A.T.W. and D.A.P. designed the EXPERIMENTAL PROCEDURES study and wrote the manuscript; all authors participated in discussing results and commented on the manuscript. Generation of DdacA Suppressor Mutants by Cre-lox fl ACKNOWLEDGMENTS The dacA PactA-cre strain was grown at 30 C overnight without agitation, and bone-marrow-derived macrophages (BMDMs) were infected as previously described using gentamicin to kill extracellular bacteria (Witte et al., 2013). After The authors would like to thank Joshua Woodward (University of Washington) infection, bacteria were grown intracellularly for 4 hr to allow for adequate cyto- for helpful advice, Peter Lauer and Bill Hanson (Aduro BioTech) for sharing ge- solic access and actA induction. Infected BMDMs were then washed three netic tools, Michael Cashel (National Institutes of Health) for generously times with sterile PBS, lysed with 0.1% NP-40, and plated on media agar at providing E. coli strains, Michelle Reniere for helpful discussions and critical 37C, curing the cre-containing plasmid. DdacA mutants were verified by reading of the manuscript, and Regina Matthew for technical assistance. PCR using primers internal to dacA and primers external to the dacA locus. This work used the Vincent J. Coates Genomics Sequencing Laboratory at In some cases, BMDM lysates were initially plated at 30C on selective UC Berkeley, supported by NIH S10 Instrumentation Grants S10RR029668

BHI-agar to enrich for bacteria that retained the PactA-cre plasmid, prior to and S10RR027303. This work was supported by National Institutes of Health plasmid-curing at 37C. For generating 284 additional DdacA suppressor grants 1P01 AI063302 and 1R01 AI27655 to D.A.P.; A.T.W. is supported by fl mutants, 24 independent infections with dacA PactA-cre were used, dacA dele- the NSF GRFP DGE 1106400 and the UC Berkeley Center for Emerging and tion was confirmed by PCR, bacteria were grown in minimal medium with bia- Neglected Diseases Irving H. Wiesenfeld Graduate Fellowship. D.A.P. has a laphos to analyze Opp activity, the synthase domain of relA was sequenced consulting relationship with and a financial interest in Aduro BioTech, Inc., with primers relA-syn-F/R, and Opp activity was reanalyzed by disk diffusion. and both he and the company stand to benefit from the commercialization of the results of this research. dacA Essentiality Assay The dacA essentiality assay was performed by adapting previously described Received: December 22, 2014 methods (De Las Pen˜ as et al., 1997). Three transducing lysates were con- Revised: March 5, 2015 structed from dacAfl-kanR lmo2103/2104::himar1 (dacA lysate) and three Accepted: March 31, 2015 transducing lysates were constructed from DdacA-kanR lmo2103/2104:: Published: May 28, 2015 himar1 (DdacA lysate) produced by Cre-lox deletion of dacA. dacA essentiality REFERENCES in a recipient strain was analyzed by transducing with each of the six lysates and selecting for erythromycin-resistant (ermR) transductants on the indicated media. At 48 hr, 50 transductants were patched from each transduction onto Abranches, J., Martinez, A.R., Kajfasz, J.K., Cha´ vez, V., Garsin, D.A., and appropriate media agar containing either erythromycin or kanamycin. The pro- Lemos, J.A. (2009). The molecular alarmone (p)ppGpp mediates stress re- portion of kanamycin resistant colonies is a measure of genetic linkage, and sponses, vancomycin tolerance, and virulence in Enterococcus faecalis. one out of 50 colonies defined the L.o.D. at 2%. As a control, the essentiality J. Bacteriol. 191, 2248–2256. assay was also performed with a lmo2110::himar1 transposon insertion instead Archer, K.A., Durack, J., and Portnoy, D.A. (2014). STING-dependent type I of lmo2103/2104::himar1. The mean genetic linkage from the three transducing IFN production inhibits cell-mediated immunity to Listeria monocytogenes. lysates per genotype constituted one experiment; data represent the mean ± PLoS Pathog. 10, e1003861.

796 Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. Barker, J.R., Koestler, B.J., Carpenter, V.K., Burdette, D.L., Waters, C.M., Joseph, B., and Goebel, W. (2007). Life of Listeria monocytogenes in the host Vance, R.E., and Valdivia, R.H. (2013). STING-Dependent Recognition of cells’ cytosol. Microbes Infect. 9, 1188–1195. Cyclic di-AMP Mediates Type I Interferon Responses during Chlamydia tra- Jutras, B.L., Chenail, A.M., Rowland, C.L., Carroll, D., Miller, M.C., Bykowski, chomatis Infection. MBio. 4, http://dx.doi.org/10.1128/mBio.00018-13. T., and Stevenson, B. (2013). Eubacterial SpoVG homologs constitute a new Bennett, H.J., Pearce, D.M., Glenn, S., Taylor, C.M., Kuhn, M., Sonenshein, family of site-specific DNA-binding proteins. PLoS ONE 8, e66683. A.L., Andrew, P.W., and Roberts, I.S. (2007). Characterization of relA and Kaplan Zeevi, M., Shafir, N.S., Shaham, S., Friedman, S., Sigal, N., Nir Paz, R., codY mutants of Listeria monocytogenes: identification of the CodY regulon Boneca, I.G., and Herskovits, A.A. (2013). Listeria monocytogenes multidrug and its role in virulence. Mol. Microbiol. 63, 1453–1467. resistance transporters and cyclic di-AMP, which contribute to type I interferon Borezee, E., Pellegrini, E., and Berche, P. (2000). OppA of Listeria monocyto- induction, play a role in cell wall stress. J. Bacteriol. 195, 5250–5261. genes, an oligopeptide-binding protein required for bacterial growth at low Kriel, A., Bittner, A.N., Kim, S.H., Liu, K., Tehranchi, A.K., Zou, W.Y., Rendon, temperature and involved in intracellular survival. Infect. Immun. 68, 7069– S., Chen, R., Tu, B.P., and Wang, J.D. (2012). Direct regulation of GTP homeo- 7077. stasis by (p)ppGpp: a critical component of viability and stress resistance. Mol. Brinsmade, S.R., Alexander, E.L., Livny, J., Stettner, A.I., Segre` , D., Rhee, Cell 48, 231–241. K.Y., and Sonenshein, A.L. (2014). Hierarchical expression of genes controlled Kriel, A., Brinsmade, S.R., Tse, J.L., Tehranchi, A.K., Bittner, A.N., Sonenshein, by the Bacillus subtilis global regulatory protein CodY. Proc. Natl. Acad. Sci. A.L., and Wang, J.D. (2014). GTP dysregulation in Bacillus subtilis cells lacking USA 111, 8227–8232. (p)ppGpp results in phenotypic amino acid auxotrophy and failure to adapt to Chico-Calero, I., Sua´ rez, M., Gonza´ lez-Zorn, B., Scortti, M., Slaghuis, J., nutrient downshift and regulate biosynthesis genes. J. Bacteriol. 196, Goebel, W., and Va´ zquez-Boland, J.A.; European Listeria Genome 189–201. Consortium (2002). Hpt, a bacterial homolog of the microsomal glucose- 6- Lemos, J.A., Lin, V.K., Nascimento, M.M., Abranches, J., and Burne, R.A. phosphate translocase, mediates rapid intracellular proliferation in Listeria. (2007). Three gene products govern (p)ppGpp production by Streptococcus Proc. Natl. Acad. Sci. USA 99, 431–436. mutans. Mol. Microbiol. 65, 1568–1581. Corrigan, R.M., and Gru¨ ndling, A. (2013). Cyclic di-AMP: another second Liu, S., Bayles, D.O., Mason, T.M., and Wilkinson, B.J. (2006). A cold-sensitive messenger enters the fray. Nat. Rev. Microbiol. 11, 513–524. Listeria monocytogenes mutant has a transposon insertion in a gene encoding Corrigan, R.M., Abbott, J.C., Burhenne, H., Kaever, V., and Gru¨ ndling, A. a putative membrane protein and shows altered (p)ppGpp levels. Appl. (2011). c-di-AMP is a new second messenger in Staphylococcus aureus Environ. Microbiol. 72, 3955–3959. with a role in controlling cell size and envelope stress. PLoS Pathog. 7, Lobel, L., Sigal, N., Borovok, I., Ruppin, E., and Herskovits, A.A. (2012). e1002217. Integrative genomic analysis identifies isoleucine and CodY as regulators of Corrigan, R.M., Campeotto, I., Jeganathan, T., Roelofs, K.G., Lee, V.T., and Listeria monocytogenes virulence. PLoS Genet. 8, e1002887. Gru¨ ndling, A. (2013). Systematic identification of conserved bacterial c-di- Lobel, L., Sigal, N., Borovok, I., Belitsky, B.R., Sonenshein, A.L., and AMP receptor proteins. Proc. Natl. Acad. Sci. USA 110, 9084–9089. Herskovits, A.A. (2014). The metabolic regulator CodY links Listeria monocy- Corrigan, R.M., Bowman, L., Willis, A.R., Kaever, V., and Gru¨ ndling, A. (2015). togenes metabolism to virulence by directly activating the virulence regulatory Cross-talk between two nucleotide-signaling pathways in Staphylococcus gene prfA. Mol. Microbiol. 95, 624–644. aureus. J. Biol. Chem. 290, 5826–5839. Luo, Y., and Helmann, J.D. (2012). Analysis of the role of Bacillus subtilis s(M) Cossart, P. (2011). Illuminating the landscape of host-pathogen interactions in b-lactam resistance reveals an essential role for c-di-AMP in peptidoglycan with the bacterium Listeria monocytogenes. Proc. Natl. Acad. Sci. USA 108, homeostasis. Mol. Microbiol. 83, 623–639. 19484–19491. Matsuno, K., and Sonenshein, A.L. (1999). Role of SpoVG in asymmetric sep- Crimmins, G.T., Herskovits, A.A., Rehder, K., Sivick, K.E., Lauer, P., Dubensky, tation in Bacillus subtilis. J. Bacteriol. 181, 3392–3401. T.W., Jr., and Portnoy, D.A. (2008). Listeria monocytogenes multidrug resis- McLaggan, D., Logan, T.M., Lynn, D.G., and Epstein, W. (1990). Involvement of tance transporters activate a cytosolic surveillance pathway of innate immu- gamma-glutamyl peptides in osmoadaptation of Escherichia coli. J. Bacteriol. nity. Proc. Natl. Acad. Sci. USA 105, 10191–10196. 172, 3631–3636. Dalebroux, Z.D., and Swanson, M.S. (2012). ppGpp: magic beyond RNA poly- McLaggan, D., Naprstek, J., Buurman, E.T., and Epstein, W. (1994). merase. Nat. Rev. Microbiol. 10, 203–212. Interdependence of K+ and glutamate accumulation during osmotic adapta- Danilchanka, O., and Mekalanos, J.J. (2013). Cyclic dinucleotides and the tion of Escherichia coli. J. Biol. Chem. 269, 1911–1917. innate immune response. Cell 154, 962–970. Mechold, U., Cashel, M., Steiner, K., Gentry, D., and Malke, H. (1996). De Las Pen˜ as, A., Connolly, L., and Gross, C.A. (1997). SigmaE is an essential Functional analysis of a relA/spoT gene homolog from Streptococcus equisi- sigma factor in Escherichia coli. J. Bacteriol. 179, 6862–6864. milis. J. Bacteriol. 178, 1401–1411. Epstein, W. (2003). The roles and regulation of potassium in bacteria. Prog. Mehne, F.M.P., Gunka, K., Eilers, H., Herzberg, C., Kaever, V., and Stu¨ lke, J. Nucleic Acid Res. Mol. Biol. 75, 293–320. (2013). Cyclic di-AMP homeostasis in Bacillus subtilis: both lack and high-level Freitag, N.E., Port, G.C., and Miner, M.D. (2009). Listeria monocytogenes - accumulation of the nucleotide are detrimental for cell growth. J. Biol. Chem. from saprophyte to intracellular pathogen. Nat. Rev. Microbiol. 7, 623–628. 288, 2004–2017. Geiger, T., and Wolz, C. (2014). Intersection of the stringent response and the Meier, S., Goerke, C., Wolz, C., Seidl, K., Homerova, D., Schulthess, B., CodY regulon in low GC Gram-positive bacteria. Int. J. Med. Microbiol. 304, Kormanec, J., Berger-Ba¨ chi, B., and Bischoff, M. (2007). sigmaB and the 150–155. sigmaB-dependent arlRS and yabJ-spoVG loci affect capsule formation in Geiger, T., Ka¨ stle, B., Gratani, F.L., Goerke, C., and Wolz, C. (2014). Two small Staphylococcus aureus. Infect. Immun. 75, 4562–4571. (p)ppGpp synthases in Staphylococcus aureus mediate tolerance against cell Nanamiya, H., Kasai, K., Nozawa, A., Yun, C.-S., Narisawa, T., Murakami, K., envelope stress conditions. J. Bacteriol. 196, 894–902. Natori, Y., Kawamura, F., and Tozawa, Y. (2008). Identification and functional Hogg, T., Mechold, U., Malke, H., Cashel, M., and Hilgenfeld, R. (2004). analysis of novel (p)ppGpp synthetase genes in Bacillus subtilis. Mol. Conformational antagonism between opposing active sites in a bifunctional Microbiol. 67, 291–304. RelA/SpoT homolog modulates (p)ppGpp metabolism during the stringent Nelson, J.W., Sudarsan, N., Furukawa, K., Weinberg, Z., Wang, J.X., and response. Cell 117, 57–68. Breaker, R.R. (2013). Riboswitches in eubacteria sense the second messenger Huynh, T.N., Luo, S., Pensinger, D., Sauer, J.-D., Tong, L., and Woodward, J.J. c-di-AMP. Nat. Chem. Biol. 9, 834–839. (2015). An HD-domain phosphodiesterase mediates cooperative hydrolysis of O’Riordan, M., Yi, C.H., Gonzales, R., Lee, K.-D., and Portnoy, D.A. (2002). c-di-AMP to affect bacterial growth and virulence. Proc. Natl. Acad. Sci. USA Innate recognition of bacteria by a macrophage cytosolic surveillance 112, E747–E756. pathway. Proc. Natl. Acad. Sci. USA 99, 13861–13866.

Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc. 797 Phan-Thanh, L., and Gormon, T. (1997). A chemically defined minimal medium involved in LTA synthesis and triggering of innate immunity during infection. for the optimal culture of Listeria. Int. J. Food Microbiol. 35, 91–95. Front. Cell Infect. Microbiol. 4,16. Rallu, F., Gruss, A., Ehrlich, S.D., and Maguin, E. (2000). Acid- and multistress- Tagami, K., Nanamiya, H., Kazo, Y., Maehashi, M., Suzuki, S., Namba, E., resistant mutants of Lactococcus lactis : identification of intracellular stress Hoshiya, M., Hanai, R., Tozawa, Y., Morimoto, T., et al. (2012). Expression of signals. Mol. Microbiol. 35, 517–528. a small (p)ppGpp synthetase, YwaC, in the (p)ppGpp(0) mutant of Bacillus sub- Rao, F., See, R.Y., Zhang, D., Toh, D.C., Ji, Q., and Liang, Z.-X. (2010). YybT is tilis triggers YvyD-dependent dimerization of ribosome. Microbiologyopen 1, a signaling protein that contains a cyclic dinucleotide phosphodiesterase 115–134. domain and a GGDEF domain with ATPase activity. J. Biol. Chem. 285, Taylor, C.M., Beresford, M., Epton, H.A.S., Sigee, D.C., Shama, G., Andrew, 473–482. P.W., and Roberts, I.S. (2002). Listeria monocytogenes relA and hpt mutants Reniere, M.L., Whiteley, A.T., Hamilton, K.L., John, S.M., Lauer, P., Brennan, are impaired in surface-attached growth and virulence. J. Bacteriol. 184, R.G., and Portnoy, D.A. (2015). Glutathione activates virulence gene expres- 621–628. sion of an intracellular pathogen. Nature 517, 170–173. Tsai, H.-N., and Hodgson, D.A. (2003). Development of a synthetic minimal Scha¨ r, J., Stoll, R., Schauer, K., Loeffler, D.I.M., Eylert, E., Joseph, B., medium for Listeria monocytogenes. Appl. Environ. Microbiol. 69, 6943–6945. Eisenreich, W., Fuchs, T.M., and Goebel, W. (2010). Pyruvate carboxylase Vance, R.E., Isberg, R.R., and Portnoy, D.A. (2009). Patterns of pathogenesis: plays a crucial role in carbon metabolism of extra- and intracellularly repli- discrimination of pathogenic and nonpathogenic microbes by the innate im- cating Listeria monocytogenes. J. Bacteriol. 192, 1774–1784. mune system. Cell Host Microbe 6, 10–21. Shetron-Rama, L.M., Marquis, H., Bouwer, H.G.A., and Freitag, N.E. (2002). Intracellular induction of Listeria monocytogenes actA expression. Infect. Wang, J.D., Sanders, G.M., and Grossman, A.D. (2007). Nutritional control of Immun. 70, 1087–1096. elongation of DNA replication by (p)ppGpp. Cell 128, 865–875. Somerville, G.A., and Proctor, R.A. (2009). At the crossroads of bacterial meta- Witte, G., Hartung, S., Bu¨ ttner, K., and Hopfner, K.-P. (2008). Structural bolism and virulence factor synthesis in Staphylococci. Microbiol. Mol. Biol. biochemistry of a bacterial checkpoint protein reveals diadenylate cyclase ac- Rev. 73, 233–248. tivity regulated by DNA recombination intermediates. Mol. Cell 30, 167–178. Sonenshein, A.L. (2007). Control of key metabolic intersections in Bacillus sub- Witte, C.E., Whiteley, A.T., Burke, T.P., Sauer, J.-D., Portnoy, D.A., and tilis. Nat. Rev. Microbiol. 5, 917–927. Woodward, J.J. (2013). Cyclic di-AMP is critical for Listeria monocytogenes Sun, A.N., Camilli, A., and Portnoy, D.A. (1990). Isolation of Listeria monocyto- growth, cell wall homeostasis, and establishment of infection. MBio 4, genes small-plaque mutants defective for intracellular growth and cell-to-cell e00282–e13. spread. Infect. Immun. 58, 3770–3778. Woodward, J.J., Iavarone, A.T., and Portnoy, D.A. (2010). c-di-AMP secreted Sureka, K., Choi, P.H., Precit, M., Delince, M., Pensinger, D.A., Huynh, T.N., by intracellular Listeria monocytogenes activates a host type I interferon Jurado, A.R., Goo, Y.A., Sadilek, M., Iavarone, A.T., et al. (2014). The cyclic response. Science 328, 1703–1705. dinucleotide c-di-AMP is an allosteric regulator of metabolic enzyme function. Yang, J., Bai, Y., Zhang, Y., Gabrielle, V.D., Jin, L., and Bai, G. (2014). Deletion Cell 158, 1389–1401. of the cyclic di-AMP phosphodiesterase gene (cnpB) in Mycobacterium tuber- Tadmor, K., Pozniak, Y., Burg Golani, T., Lobel, L., Brenner, M., Sigal, N., and culosis leads to reduced virulence in a mouse model of infection. Mol. Herskovits, A.A. (2014). Listeria monocytogenes MDR transporters are Microbiol. 93, 65–79.

798 Cell Host & Microbe 17, 788–798, June 10, 2015 ª2015 Elsevier Inc.