R E S E A R C H A R T I C L E

M I C R O B I O L O G Y

Identification of Fic-1 as an enzyme that inhibitsbacterial DNA replication by AMPylating GyrB,promoting filament formationCanhua Lu,1,2 Ernesto S. Nakayasu,3* Li-Qun Zhang,1† Zhao-Qing Luo2†

hD

ownloaded from

The morphology of bacterial cells is important for virulence, evasion of the host immune system, andcoping with environmental stresses. The widely distributed Fic proteins (filamentation induced by cAMP)are annotated as proteins involved in cell division because of the presence of the HPFx[D/E]GN[G/K]Rmotif. We showed that the presence of Fic-1 from Pseudomonas fluorescens significantly reduced theyield of plasmid DNA when expressed in Escherichia coli or P. fluorescens. Fic-1 interacted with GyrB,a subunit of DNA gyrase, which is essential for bacterial DNA replication. Fic-1 catalyzed the AMPylation ofGyrB at Tyr109, a residue critical for binding ATP, and exhibited auto-AMPylation activity. Mutation of theFic-1 auto-AMPylated site greatly reduced AMPylation activity toward itself and toward GyrB. Fic-1–dependent AMPylation of GyrB triggered the SOS response, indicative of DNA replication stress orDNA damage. Fic-1 also promoted the formation of elongated cells when the SOS response wasblocked. We identified an a-inhibitor protein that we named anti–Fic-1 (AntF), encoded by a gene imme-diately upstream of Fic-1. AntF interacted with Fic-1, inhibited the AMPylation activity of Fic-1 for GyrBin vitro, and blocked Fic-1–mediated inhibition of DNA replication in bacteria, suggesting that Fic-1 andAntF comprise a toxin-antitoxin module. Our work establishes Fic-1 as an AMPylating enzyme thattargets GyrB to inhibit DNA replication and may target other proteins to regulate bacterial morphology.

ttp

on M

ay 31, 2020://stke.sciencem

ag.org/

INTRODUCTION

The morphology of many bacteria changes in response to environmentalcues (1). One of the most frequently observed changes for some rod-shapedbacteria is the transition between a bacillary (free-living) and a filamentousform during the different phases of the life cycles (1). For example, uro-pathogenic Escherichia coli (UPEC), the predominant causative agent ofurinary tract infection (2), forms intracellular bacterial communities afterinvasion of the superficial epithelial cells of the bladder, where a subpop-ulation develops into filamentous cells (3, 4). The generation of filamen-tous bacteria is believed to confer resistance to phagocytosis by neutrophilsmigrating into the infection site (5, 6). In mice defective in Toll-like recep-tor 4 (the receptor for lipopolysaccharide), UPEC no longer forms fila-mentous bacteria, indicating that the host innate immune system triggersthe development program that leads to filamentation (3). Transition be-tween filamentous and rod shape has also been observed in Legionellapneumophila, the causative agent of Legionnaires’ disease, in responseto temperature as well as other environmental stimuli (7, 8). Filamentous L.pneumophila appears to be more resistant to phagocytosis, but whether theimmune system induces this development is unknown (7). Similarly, fila-mentous Salmonella enterica serovar Typhimurium cannot invade hostcells (9) despite inducing membrane ruffling, indicating the importanceof cell morphology in bacterial virulence.

In a model organism, such as E. coli, cell filamentation is caused byincomplete cell division. The segregation of completely replicated chro-mosomes is accompanied by the formation of a site for cell division, whereFtsZ, the guanosine 5′-triphosphate (GTP)–binding tubulin homolog,

1Department of Plant Pathology, China Agricultural University, Beijing 100193,China. 2Department of Biological Sciences, Purdue University, West Lafayette,IN 47907, USA. 3The Bindley Bioscience Center, Purdue University, West Lafay-ette, IN 47907, USA.*Present address: Biological Sciences Division, Pacific Northwest NationalLaboratory, Richland, WA 99352, USA.†Corresponding author. E-mail: zhanglq@cau.edu.cn (L.-Q.Z.); luoz@purdue.edu (Z.-Q.L.)

www

polymerizes into a ring-like structure to form the multicomponent com-plex that mediates cell division (10, 11). Inhibition of FtsZ activity byfactors such as SulA causes incomplete cell division and thus filamentation(12). The expression of sulA is regulated by LexA, the master regulator ofthe “SOS” response that senses cellular stresses, such as DNA damageand DNA replication inhibition (13, 14).

Cell filamentation occurs in an E. coli mutant grown at 43°C in amedium containing cyclic adenosine monophosphate (cAMP) (15).Further studies revealed that this mutant harbored a G55R mutationin the gene fic (filamentation induced by cAMP) (16, 17). The Fic motifhas a core structure of HXFX(D/E)GNGRXXR (x, any amino acid) and ispresent in thousands of proteins from organisms of all taxonomic orders(18–20).

The discovery that VopS, a type III effector from Vibrio parahaemo-lyticus, catalyzes the transfer of the AMP moiety from adenosine 5′-triphosphate (ATP) to GTPases involved in host cytoskeleton structure(for example, Rho and Rac) revived the study of Fic proteins (19). Otherstudies revealed that members of this protein family catalyze diverse reac-tions, including AMPylation (19, 21), phosphorylcholination (6, 22),UMPylation (23), and phosphorylation (24, 25). These proteins are in-volved in diverse cellular processes, such as bacterial virulence (6, 19, 21, 22),protein translation in prokaryotes (24, 25), neurotransmission in fly (26),and the unfolded protein response in eukaryotic cells (27, 28). The activityof some Fic proteins is reversibly regulated by specific enzymes. For ex-ample, phosphorylcholination conferred by AnkX from L. pneumophilais enzymatically reversed by Lem3, a dephosphorylcholinase (29). TheL. pneumophila effector SidD de-AMPylates the Rab1 small GTPase(30), which is AMPylated by a Fic-independent mechanism (31).

Whereas Fic proteins involved in bacterial virulence appear to beconstitutively active, the activity of most Fic proteins in bacteria and thosein eukaryotes is regulated by a motif with a consensus sequence of(S/T)XXXEG found either within the Fic proteins themselves or in a smallprotein called the a-inhibitor encoded by a gene that is often adjacent to

.SCIENCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 1

R E S E A R C H A R T I C L E

the fic gene. In bacteria, Fic proteins and their cognate a-inhibitors con-stitute typical toxin-antitoxin (TA) modules (32).

Whereas the functions of Fic proteins in bacterial virulence and the cellbiology of eukaryotic cells have been relatively well documented, ourunderstanding of the function of “housekeeping” Fic proteins from diversebacteria is limited. Many of these proteins are annotated as cell divisionproteins, but little is known about their physiological role beyond thefilamentation phenotype originally described for the FicG55R E. coli mu-tant (16). Here, we showed that Fic-1, a Fic protein from a pseudomonad,AMPylated and inhibited the DNA gyrase subunit B (GyrB), whichinhibited bacterial DNA replication. Fic-1 also arrested bacterial growthand induced cell filamentation. Our results indicated that Fic-1 targetedGyrB to inhibit bacterial division and suggested that additional targetsare important for the effects of Fic-1 on cell morphology.

http://D

ownloaded from

RESULTS

A Fic protein from Pseudomonas fluorescens interfereswith plasmid DNA replication and E. coli growthMost of the Fic proteins are found in taxonomically diverse bacteria. Sucha wide distribution suggests that they function in fundamental cellular pro-cesses. To determine the function of housekeeping Fic proteins, we ex-amined potentially discernable phenotypes in E. coli by expressing suchproteins from various bacteria. Among them, Pseudomonas fluorescensstrain 2P24, which is used for biocontrol of certain plant diseases (33),encodes three predicted Fic proteins, which we designated Fic-1, Fic-2,

on May 31, 2020

stke.sciencemag.org/

and Fic-3. These proteins are 199, 342,and 395 residues long, respectively (fig.S1), with Fic-1 having a length similar toFic proteins found in diverse bacteria (20)(fig. S2).

We studied the function of Fic-1, Fic-2,and Fic-3 proteins by expressing them inE. coli by individually introducing thecoding sequence and an upstream sequencecontaining the putative promoter into aColE1 plasmid. E. coli expressing Fic-1 from this plasmid produced significantlyless plasmid DNA than E. coli expressingFic-2 or Fic-3 (Fig. 1A). The inhibition ofplasmid DNA yield depended on the Ficmotif because a mutation in His135, whichis critical for catalytic activity, eliminated theinhibition (Fig. 1B). Expression of Fic-1 incells harboring other types of plasmids alsosignificantly decreased their yield, suggest-ing that Fic-1 inhibits bacterial DNA repli-cation (fig. S3). In contrast, expression ofconstitutively active forms of Fic-2 andFic-3, which contain mutations in the pre-dicted intramolecular inhibitory motif knownto abolish the inhibitory effects (32), did notdetectably affect plasmid yield (fig. S4).

Expression of Fic-1 causesbacterial filamentationTo further determine the effects of Fic-1,we examined themorphology of cells express-ing Fic-1. In E. coli strain BL21(DE3),

www

transformants harboring a plasmid that directs the production of His6-tagged Fic-1 formed smaller colonies on LB medium even without theinducer isopropyl b-D-1-thiogalactopyranoside (IPTG) (fig. S5A). TheFic motif–dependent inhibition of E. coli growth by Fic-1 was also appar-ent in liquid LB medium. Whereas the strain harboring plasmid control orthe one expressing the H135A mutant reached the stationary phase ofgrowth within 12 hours, we observed little growth in cultures of the strainharboring the plasmid expressing wild-type Fic-1 in the first 14 hoursof incubation (fig. S5B). E. coli cells expressing wild-type Fic-1 grew8 hours after inoculation (fig. S5B). We speculate that such growth maybe due to accumulation of mutations in fic-1 or in the bacterial chromo-some, thus making the cells less sensitive to Fic-1, or a combination ofboth.

Microscopic analysis revealed that a fraction of cells expressingFic-1 were filamentous with multiple (>4) distinct nucleoids. Shorter cellswith two to three nucleoids were also present (Fig. 2A). In contrast, insamples of the strain expressing Fic-1H135A, no filamentous cells werefound, which is similar to the strain containing the plasmid control (Fig. 2A).

We examined the function of Fic-1 in P. fluorescens strain 2P24. De-letion of fic-1 did not result in any discernable phenotype. We could notexpress Fic-1 in P. fluorescens using several constitutive promoters;therefore, we used the arabinose-inducible promoter. Induction of Fic-1 inhibited the growth of P. fluorescens, and this inhibition required afunctional Fic domain (fig. S5C). Cell growth in cultures expressingFic-1 resumed 32 hours after inoculation (fig. S5C), which we again spec-ulate may arise through mutations in either fic-1 or its targets that renderedthem no longer sensitive to its activity. A portion of cells expressing Fic-1,

Fig. 1. Fic-1 affects plasmid DNAproduction in E. coli. (A) The effectof individual fic genes on the amountof plasmid DNA in E. coli. PlasmidDNA containing the indicated con-structs isolated from saturated cul-tures of E. coli was resolved byagarose gel electrophoresis before(left panel) or after digestion with re-striction enzymes Bam HI and Sal I(fic-1) or Xba I and Hind III (fic-2 andfic-3) (middle panel). In the middle

.SCIENCESIGNALING.org 26 Janua

panel, both bands represent the products of the digested plasmid DNA. Data are representative of threeexperiments. The intensity of the bands corresponding to the vector was measured from threeindependent experiments (right panel). ***P < 0.001. (B) The effect of mutating the Fic domain ofFic-1 on the amount of plasmid DNA in E. coli. Plasmid DNA containing wild-type fic-1 and fic-1 with aFic domain mutation (Fic-1H135A) was isolated from saturated cultures of E. coli, subjected to restrictionenzyme digestion, and separated by agarose gel electrophoresis (upper panel). Western blot analysisof isocitrate dehydrogenase (ICDH), a metabolic enzyme, served as a loading control. Data are rep-resentative of three experiments. The intensity of the bands corresponding to the vector was measuredfrom three independent experiments (right panel). ***P < 0.001.

ry 2016 Vol 9 Issue 412 ra11 2

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

but not the enzymatically inactive mutant, became filamentous (Fig. 2B),suggesting that Fic-1 targets pathways conserved between E. coli andP. fluorescens. Both the wild-type and the H135A mutant were detectedin E. coli (Fig. 2C); in P. fluorescens, the amount of wild-type Fic-1 waslower than that of the H135A mutant (Fig. 2D). Together, these resultsindicated that overexpression of Fic-1 regulates one or more cellular pro-cesses relevant to DNA replication, bacterial growth, or both.

Fic-1 interacts with GyrB, a subunit of DNA gyraseWe sought to determine the mechanism of action of Fic-1 by identifyingits cellular target. Because Fic-1 inhibited plasmid replication, we pre-dicted that one or more proteins involved in DNA replication may beinvolved. To this end, we examined the potential interactions of Fic-1 witheach of the 77 E. coli proteins known to participate in DNA replication (tableS1). Among these, GyrB, a subunit of DNA gyrase (type II topoisomerase)specifically interacted with Fic-1 in a bacterial two-hybrid assay (34). Wedetected significant b-galactosidase activity only in the E. coli strain co-expressing the appropriate domains of two-hybrid assay protein Cya fusedto Fic-1 and GyrB, respectively (Fig. 3A). When recombinant GyrB wasincubated with beads coated with Fic-1, the Fic-1H135A mutant, or bovineserum albumin (BSA), beads coated with Fic-1 or Fic-1H135A retainedmore GyrB (Fig. 3B). Thus, by interacting with GyrB, Fic-1 may regulatebacterial DNA replication.

Fic-1 AMPylates GyrB on a tyrosine residue important forbinding ATPBecause the Fic domain is capable of catalyzing diverse biochemical re-actions (20), we attempted to determine whether GyrB is a target of Fic-1 byconstructing an E. coli strain that coexpressed these two proteins taggedwith glutathione S-transferase (GST) and His6, respectively. Purified GyrB-

www.SCIENCESIGNALING.org 26

His6 protein was then separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), excised, and digested with trypsin.The resulting peptides were analyzed byliquid chromatography coupled with tandemmass spectrometry (LC-MS/MS). Spectraderived from MS/MS analysis of peptideswere inspected by searching for peptideswith potential change in molecular mass; amass shift of 329 daltons corresponding tothe addition of an AMPmoiety was detectedin the tryptic peptide -F104DDNSYK110- inGyrB coexpressed with wild-type Fic-1 butnot with the Fic-1H135A mutant (Fig. 4A,upper panel). In addition, an intense peakof the diagnostic ion at 136.06 mass/chargeratio (m/z) (Fig. 4A, lower panel), corre-sponding to adenine derived from the break-down of AMP, unambiguously indicatedAMPylation. The unmodified fragmentwas almost undetectable inGyrB coexpressedwith wild-type Fic-1, indicating that themodification was extensive (Fig. 4A, upperpanel, middle). As a reference, the abundanceof the tryptic peptide -G699LLEEDAFIER709-outside the modification site was very simi-lar in these two samples (Fig. 4A, upperpanel, right). Further analysis of the y andb series of the peptide backbone,more spec-ifically, y1 and y2 fragments, allowed us to

unambiguously assign the site of modification to the Tyr109 residue(Fig. 4A, lower panel).

To verify the results of the MS analysis, we analyzed Fic-1–mediatedAMPylation of GyrB in vitro. Fic-1–His6, Fic-1H135A–His6, GyrB-His6,and GyrBY109F-His6 were purified, and the activity of Fic-1 and the mu-tant was evaluated using 32P-a-ATP. In the reaction containing wild-typeFic-1 and GyrB, robust 32P-AMP labeling of GyrB was detected, with anexposure time of less than 1 min (Fig. 4B, upper panel, third lane); self-modification of wild-type Fic-1 was also detectable (Fig. 4B, upper panel,second to fourth lanes). In reactions containing Fic-1H135A, no

32P-AMP-GyrB was detected (Fig. 4B, upper panel, fifth to seventh lanes), indicat-ing a Fic motif–dependent activity. GyrB contains two distinct domains(35). We found that the N-terminal domain (residues 1 to 200), but notthe C-terminal domain, of GyrB was modified by Fic-1 (fig. S6), indicat-ing that this portion of the topoisomerase is sufficient to interact with theFic-1. In agreement with the high degree of conservation with its E. colicounterpart, GyrB from P. fluorescens was modified by Fic-1 at Tyr111

(fig. S7), which corresponds to Tyr109 of the E. coli GyrB. The multiplemodified GyrB products detected in these reactions may result from mod-ification of partially degraded GyrB or the degradation of modified GyrB(fig. S7).

Weak but readily detectable signals in the position of GyrB were de-tected in the sample containing only wild-type Fic-1 (Fig. 4B, upper panel,second lane). Because this signal was also detected in experiments inwhich this sample was loaded far from the sample with strong modifica-tion (fig. S6, sixth and eighth lanes), we predicted that this representsAMPylated native GyrB that copurified with Fic-1. The lack of such signalsin the reaction of wild-type Fic-1 with GyrBY109F may be due to the compe-titive effect of the excess GyrB mutant used in the reaction (Fig. 4B, upperpanel, fourth lane). These results further validated GyrB as the target of Fic-1.

Fig. 2. Induction of cell filamenta-tion by Fic-1. (A) Cells from E. colitransformants expressing SUMO-tagged Fic-1 or its H135A mutantgrown for 16 hours on LB agar re-suspended in phosphate-buffered

saline (PBS) were fixed and stained with Hoechst. (B) Cells of P. fluorescens grown in LB with glucosewere diluted in ABM medium containing 0.2% arabinose and grown for 12 hours. Cells were treated forimaging as described in (A). Images acquired with a fluorescence microscope were pseudocolored withIPLab software. (C and D) Cells grown as described in (A) and (B) were processed for SDS-PAGE. Pro-teins were probed for Fic-1 by immunoblotting. The metabolic enzyme ICDH was probed as a loadingcontrol (lower panels). Similar results were obtained from at least three independent experiments.

January 2016 Vol 9 Issue 412 ra11 3

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

AMPylation by Fic-1 inactivates the ATPase activity ofGyrB and induces the SOS responseGyrB is a subunit of DNA gyrase (DNA gyrase is also known as DNAtopoisomerase II), the main activity of which is to introduce negativesupercoiling to relieve the strain caused by helicase during DNA replica-tion, in a process powered by ATP hydrolysis (36). In E. coli GyrB, Tyr109

is critical for binding ATP by forming a hydrogen bond with the N3 atomof the adenine ring (35). To examine the effects of Tyr109 AMPylation onthe activity of GyrB, we determined the ATPase activity (37, 38) of GyrBthat had been co-incubated with Fic-1. Whereas untreated GyrB or GyrBthat had been co-incubated with Fic-1H135A exhibited readily detectableATP hydrolysis (Fig. 4C, left panel), neither Fic-1 nor its H135 mutanthydrolyzed ATP, nor did GyrB with a mutation in Tyr109 (Fig. 4C, leftpanel). As expected, GyrB that had been co-incubated with Fic-1 andATP displayed significantly lower ATPase activity (Fig. 4C, left panel).Thus, AMPylation of GyrB by Fic-1 abolished the ATPase activity of GyrB.

Inhibition of GyrB activity causes DNA replication arrest, which re-sults in the exposure of single-strand DNA and the subsequent formationof RecA filaments (39, 40). This facilitates the autocatalytic cleavage ofthe LexA repressor and the subsequent induction of the SOS response(41). Consistent with this pattern, the amount of RecA increased signifi-cantly in cells expressing Fic-1, but not in those expressing the Fic-1H135Amutant, whereas the amount of LexA, the SOS repressor, decreased in aFic-1–dependent manner (Fig. 4C, right and lower panels). Novobiocin,

www

which selectively inactivates GyrB, served as a positive control forinduction of the SOS response.

Self-AMPylation of Fic-1 is important for its activityon GyrBWe consistently observed self-AMPylation of Fic-1, particularly in reac-tions lacking its substrate GyrB (Fig. 4B, upper panel, second lane). Toexplore the potential role of self-modification on Fic-1 function, weidentified the modified site by MS analysis. Tyr5 was AMPylated inFic-1 incubated with ATP (Fig. 5A). Substitution of Tyr5 with an Alaresidue resulted in no detectable self-AMPylation (Fig. 5B and fig.S8). The Fic-1Y5A mutant failed to AMPylate GyrB (Fig. 5B). In agree-ment with the loss of AMPylator activity toward GyrB, Fic-1Y5A no lon-ger affected the yield of plasmid DNA in E. coli (Fig. 5C). Together,these results indicate that self-AMPylation is critical for this functionof Fic-1.

The activity of Fic-1 is regulated by a specific inhibitorAnalysis of the DNA sequence upstream of the fic-1 open reading frameidentified a gene potentially encoding a protein of 56 residues (fig. S1).These two genes overlap for one nucleotide. Sequence analysis revealedthat the protein encoded by this gene harbors an -S24LRLEG29- motif,which is found in a-inhibitors of Fic proteins (32). When expressed inE. coli, this gene rescued the inhibition of plasmid DNA replicationconferred by Fic-1 (fig. S9A), suggesting that the protein functioned asan a-inhibitor. We designated this gene as antF (anti–Fic-1). AntF inter-acted with Fic-1 in a bacterial two-hybrid assay; Cya fusions of AntF andFic-1 drove the expression of the LacZ reporter to levels comparable tothat of leucine zipper proteins used as positive controls (34) (fig. S9B). Sim-ilar results were obtained with a reciprocal fusion orientation (fig. S9B). Inagreement with the genetic results, inclusion of recombinant AntF in theAMPylation reaction with Fic-1 reduced the modification of GyrB in adose-dependent manner (Fig. 6, A and B). Even a 1:27 molar ratio ofFic-1 to AntF1 significantly inhibited the AMPylation activity, and a1:3 ratio completely abolished activity.

We attempted to examine the potential phenotypes of mutants lackingthe a-inhibitor. Despite repeated attempts, we could not delete the inhibitor-encoding gene without compromising the expression of fic-1 in P. fluorescensstrain 2P24 (Fig. 7A), suggesting that this bacterium cannot tolerate un-controlled activity of Fic-1 expressed from its cognate promoter. Theseresults are consistent with the observation that ectopically expressedFic-1 in P. fluorescens was lower than the mutant, although the inductionof filamentation and growth arrest were apparent (Fig. 2B and fig. S5C).

When the SOS response is blocked, Fic-1 triggerscell elongationActivation of the SOS pathway involves the formation of RecA filaments(40, 42), the auto-cleavage of the SOS repressor LexA (43), and the in-duction of sulA, which binds FtsZ, thus blocking the completion of celldivision and leading to the formation of filamentous cells (12). To exam-ine whether cell filamentation induced by Fic-1 required the SOS pathway,we introduced constructs harboring arabinose-inducible Fic-1 into anE. coli strain defective in recA and sulA. Expression of Fic-1, but notFic-1H135A, caused morphological changes. In the strain expressing Fic-1,cells longer than 10 mm were readily detectable and more than 50% of cellswere longer than 2.5 mm, whereas most of the cells expressing theFic-1H135A mutant displayed lengths between 1 and 1.75 mm, which weresimilar to those in the vector control (Fig. 7, B and C). Whereas some ofthe cells expressing Fic-1 were filamentous, most cells displayed anelongated morphology with a single nucleoid (Fig. 7, B and C). These

Fig. 3. The interactions between Fic-1 and GyrB. (A) Fic-1 interacts withGyrB as measured by a bacterial two-hybrid assay. E. coli strains derivedfrom BTH101 harboring plasmids expressing the indicated fusion proteinpairs were assayed for b-galactosidase activity indicative of protein-proteininteractions. Experiments were performed in triplicate, and similar resultswere obtained from three independent experiments. V, control plasmid. (B)Interactions between GyrB and Fic-1 in vitro. GyrB-His6 was incubated withAffigels coated with Fic-1, Fic-1H135A, or BSA. After extensive washing,bound proteins eluted with SDS sample buffer were resolved by SDS-PAGE and detected by Coomassie brilliant blue staining. Ten percent ofthe GyrB used for the binding assay was loaded as reference. Data shownare representative of three experiments with similar results.

.SCIENCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 4

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

results suggested that in cells defective in the SOS pathway, inhibition ofDNA replication by Fic-1 did not block cell elongation.

Despite repeated attempts, we were unable to construct a derivative ofthe ∆sulA strain of P. fluorescens expressing Fic-1 (no transformants wereobtained in experiments aiming at introducing the plasmid that directs theexpression from the arabinose-inducible promoter into the ∆sulA strain).This observation further indicates that P. fluorescens is more sensitive tothe Fic protein than E. coli is.

DISCUSSION

Fic proteins are widespread across all domains of life, with the largestnumber present in taxonomically diverse bacteria (19). Despite the factthat high-resolution structures of a number of bacterial Fic proteins areavailable, our understanding of the biological activities of this familycomes primarily from bacterial virulence factors (24) and the Doc proteinof phage P1 that targets host protein translation as a Fic domain–mediatedkinase (24, 44). Our results establish that a housekeeping Fic protein AM-Pylates GyrB, a subunit of the bacterial type II DNA topoisomerase, thusinterfering with DNA replication. The modification by Fic-1 abolishes theactivity of GyrB and induces filamentation.

www

Several lines of evidence show that Fic-1 specifically targets GyrB.First, Fic-1 but neither Fic-2 nor Fic-3 from the same bacterium interfereswith bacterial growth or plasmid replication. Second, Fic-1 is able to in-duce cell filamentation, a phenotype often associated with the inhibition ofGyrB activity (38, 45). Third, GyrB is the only protein that interacts withFic-1 among the 77 DNA replication proteins tested. In addition, nativeGyrB was copurified with Fic-1–His6 (Fig. 4B and fig. S6), consistentwith the sequence of its Fic domain (H135PFREGNGR143), which predictsan adenylyl transferase activity (24, 44). Fic-1 targets the GyrB subunit oftype II topoisomerase by adding an AMP moiety to the highly conservedtyrosine residue Tyr109 essential for binding ATP. Antibiotics of thecoumarin and cyclothialidine families inhibit the ATPase activity ofDNA gyrase by blocking the binding of ATP to GyrB (46). Our resultsare consistent with a recent report showing the targeting of DNA gyraseand topoisomerase IV by several Fic proteins (47). It appears that interfer-ence with the ATP binding of GyrB is a common mechanism for inhibit-ing its activity. This is most likely because ATPase activity is essential forGyrB function, and inhibition of ATP binding probably is the most effec-tive means to block its activity.

The presence of a specific and effective inhibitor of Fic-1 indicatestight regulation of its activity. The fact that we were unable to delete

Fig. 4. Fic-1 AMPylates GyrB at Tyr109. (A) Identification of AMPylationof GyrB by MS. The upper panel is the extracted ion chromatograms

of the unmodified and adenylylated versions of the peptide -F105DDNSYK110- derivedfrom GyrB coexpressed with Fic-1 (left panel) or mutant Fic-1H135A (middle panel). Theabundance of the randomly chosen peptide -G699LLEEDAFIER709- is almost identical inthese two samples (right panel). The lower panel is the tandem mass spectrum ofpeptide -F105DDNSYK110-. The asterisks represent neutral losses of the adenosine moietyof the adenylylation modification. F, immonium ion of phenylalanine residue. (B) Invitro AMPylation of GyrB by Fic-1. After incubation at 35°C for 30 min, 32P-a-AMP–labeled signals and total proteins were detected by autoradiography for 30 sec(upper panel) or by staining (lower panel). (C) AMPylation of GyrB abolishes its

.SCIE

ATPase activity. Samples from 30-min reactions were examined for ATP hydrolysis

by measuring released phosphate (left panel). Induction of the SOS pathway by Fic-1 or Fic-1H135A was evaluated by the levels of LexA and RecA withICDH as a loading control (middle panel). The band intensities of RecA (upper right) and LexA (lower right) were measured by LI-COR. All results arefrom three independent experiments. ***P < 0.0001.

NCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 5

R E S E A R C H A R T I C L E

http://stke.scienceD

ownloaded from

Fig. 5. Self-AMPylation is important for the activity of Fic-1.(A) Self-modification by Fic-1 occurs on Tyr5. Fic-1 co-incubated with ATP was analyzed by MS after trypsin di-

gestion. The upper panel is the extracted ion chromatograms of the unmodified andadenylylated versions of the tryptic peptide containing Tyr5. The abundance of therandomly chosen peptide -N179GVMEPMEQVFEK191- is almost identical in thesetwo samples (right panel). The lower panel is the tandem mass spectrum of theAMPylated peptide -Y5GVGEDAYCYPGSTVLR21-. CAM, carbamidomethylation.(B) A mutation in Tyr5 abolishes self-AMPylation of Fic-1. Reactions containing theindicated proteins were incubated at 35°C for 30 min, AMPylation was detected byautoradiography for 2 min (upper panel), and total proteins were visualized by stain-ing (lower panel). WT, wild type. (C) The Y5A mutation abolishes the ability of Fic-1 to inhibit plasmid DNA yield. DNA of plasmids carrying fic-1 or the indicated mu-tants isolated from identical amounts of E. coli cells was digested with restrictionenzymes Bam HI and Sal I and was separated using agarose gels (upper panel).

www.SCIENCESIGNA

m

ag.o

An identical set of samples was processed for immunoblotting to detect the levels of Fic-1 (lower panel). The images in (B) and (C) are representative ofthree independent experiments with similar results.

on May 31, 2020

rg/

Fig. 6. Inhibition of the AMPylation activity of Fic-1 byAntF. (A) Dose-dependent inhibition of Fic-1 by AntF.The indicated amounts of His6-SUMO-AntF were addedto a series of identical reactions containing Fic-1. After30 min of incubation, equal amounts of a mixture con-taining GyrB and 32P-a-ATP were added, and the re-actions were allowed to proceed for 30 min at 35°C.32P-a-GyrB and total proteins were detected by auto-radiography (upper panel) and Coomassie blue stain-

ing (lower panel). (B) Quantification of 32P-a-GyrB signals. The strength of autoradiography signals from three independent experiments doneunder the same conditions was measured and analyzed by ImageJ. **P < 0.01.

LING.org 26 January 2016 Vol 9 Issue 412 ra11 6

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

the inhibitor gene in strain 2P24 further indicates the importance of suchregulation, which is consistent with the observation that ectopic expressionof Fic-1 inhibits bacterial growth. Apparently, the inhibitory effects ofAntF need to be eliminated when cell filamentation or cell division arresthas become necessary. We propose that such elimination can be achievedby at least two mechanisms, both involved in the induction of a signalingcascade in response to appropriate environmental cues (Fig. 8). In the firstscenario, the signaling events induce the expression of a protease that de-grades AntF, whereas in the second scenario, the titration of the inhibitor ismediated by a protein to which it has a higher affinity than to Fic-1 (Fig. 8).

Many bacteria, including E. coli and Pseudomonas spp., become fila-mentous under stress conditions such as DNA damage caused by factorslike ultraviolet (UV) radiation or partial inhibition of cell wall biosynthesisby antibiotics (1). UPEC forms filaments only in immune competent hosts(3). Filamentous bacteria are more resistant to engulfment by phagocytesand other predators (1). In parallel, inhibition of DNA replication will mostlikely lead to slower metabolic rates, making bacteria resistant to some anti-biotics and to certain detrimental environmental conditions. Regulation ofcell morphology by targeting DNA replication has been described for theSocA-SocB TA module in Caulobacter crescentus (48). Similarly, Fic-in-duced filamentation has been reported, although themechanism is unknown(32). The identification of GyrB as a target of Fic-1 demonstrates thefunction of at least a subset of widely distributed Fic proteins in the regula-tion of bacterial cell division. Our results suggest that in response to Fic-1 activity, a yet unidentified SOS-independent pathway contributes to cellmorphology changes (Fig. 8). SOS-independent induction of cell fil-amentation has been documented inC. crescentus (49). It will be interestingto determine whether the SOS-independent pathways involved in the regu-lation of cell morphology are conserved among bacterial species.

Self-modification occurs in many Fic proteins, which has been used tomonitor their biochemical activity when a substrate is not available (32).Our finding that the Fic-1Y5A mutant defective in self-AMPylation hadalmost completely lost its activity against GyrB implies an important role

www

of self-AMPylation in its function. In its complex with the a-inhibitorVbhA, the N-terminal end of VbhT (its fourth residue is a Tyr) doesnot seem to participate in the formation of the catalytic pocket (32). Itis possible that in the absence of the inhibitor, this Tyr assumes an impor-tant role in the overall conformation of these Fics or in the formation ofthe catalytic pocket. Determining the role of this Tyr4 will likely reveal

Fig. 7. Expression of Fic-1 in P. fluorescens strains andinduction of cell filamentation by Fic-1 in E. coli defectivein sulA and recA. (A) Bacterial strains were grown in LB toan optical density at 600 nm (OD600) of 2.4, and proteinsamples prepared from 300-ml cultures were resolved bySDS-PAGE. Fic-1 was probed with a specific antibody. Aprotein nonspecifically recognized by the antibody atabout 30 kD served as a loading control. Note that Fic-1 is detectable only in the WT. (B) Morphology of cellsexpressing Fic-1. Cells of transformants grown for 16hours on LB agar resuspended in PBS were fixed andstained with Hoechst. Images acquired with a fluores-cence microscope were pseudocolored with IPLabsoftware. Images shown are representative of three ex-

.SCIENCESIGNA

periments with similar results. (C) Distribution of cells with different lengths. The length of 500 cells was measured from each of three samples, and theirdistribution was plotted. Data shown are representative from three independent experiments.

Fig. 8. A model of Fic-1–mediated induc-tion of bacterial cell filamentation and itsregulation. Fic-1 and the a-inhibitor AntF

form a dynamic complex under normal conditions. Inducing signals fromthe environment activate a cascade that leads to the production of a se-questering protein (Seq) that competes for AntF or the activation of aprotease that degrades the a-inhibitor. Freed or activated Fic-1 then inacti-vates GyrB by AMPylation, leading to the induction of SulA and the forma-tion of filamentous cell. Alternatively, AMPylated GyrB may induce cellfilamentation through an SOS response–independent pathway (dashedarrows).

LING.org 26 January 2016 Vol 9 Issue 412 ra11 7

R E S E A R C H A R T I C L E

whether this residue is generally important for the activity this groups ofFic proteins.

Pressure from the immune system and metabolic fluctuations often re-sult in persister cells that are resistant to immune killings as well as toantibiotics (46). In E. coli, TA modules are involved in the formation ofpersisters (50). FicTAs may participate in the establishment of persisters orin the regulation of the transition between the bacillary and filamentousforms during infection. One future research avenue is the identification ofthe pathway that parallels the SOS cascade in regulating bacterial mor-phology. Similarly, it will be of great interest to determine whether anyof these stresses, such as antibiotic, UV radiation, and immune suppres-sion, cause Fic-1 activation, and if so, the mechanism underlying suchderepression.

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

MATERIALS AND METHODS

Bacterial strains, plasmids, and mediaBacteria and plasmids used in this study are listed in table S2. P. fluor-escens strain 2P24 and its derivatives were grown in LB medium orAgrobacterium mannitol (ABM) minimal medium (51) at 28°C. E. colistrains were incubated at 37°C with the exception of strain BTH101 usedfor bacterial two-hybrid assays (34), which was cultured at 30°C. For plas-mid selection, antibiotics were supplemented in media at the followingconcentrations: ampicillin, 50 mg/ml; kanamycin, 50 mg/ml; gentamicin,10 mg/ml; chloramphenicol, 20 mg/ml; and tetracycline, 20 mg/ml. Whenneeded, 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside (X-gal) was usedat 20 μg/ml. To induce SOS in E. coli, novobiocin was added to a finalconcentration of 25 mg/ml to bacterial cultures established by 10-fold di-lution of overnight cultures. The cells were incubated at 37°C in a shaker(200 rpm/m) for 6 hours before processing for SDS-PAGE and immuno-blotting.

Bacterial strain constructionIn-frame deletion mutants of P. fluorescens strain 2P24—2P24∆fic-1,2P24∆fic-1∆antF, 2P24∆aantF, and 2P24∆sulA—were constructed by atwo-step gene replacement procedure as described previously (33). Thesuccess of deletion was verified by colony polymerase chain reaction(PCR) with primers flanking the genes of interest.

Site-directed mutagenesisMutations in specific site of the genes of interest were introduced byQuikChange II (Agilent Technologies) with the PfuUltra High-FidelityDNA polymerase and overlapping primer pairs bearing the desired muta-tions. The specificity of the mutations was verified by double-strandedDNA sequencing.

Evaluation of plasmid DNA yield in E. coliThe coding regions of fic-1, fic-2, and fic-3 amplified by PCR with ap-propriate primers (table S2) from genomic DNA of P. fluorescens strain2P24 (33) were digested with appropriate restriction enzymes. Bam HIand Sal I were used for fic-1 and Hind III and Xba I were used for fic-2 and fic-3 for inserting into pHSG399 (52) to give pCL001, pCL002, andpCL003, respectively. The expression of the fic genes in these plasmidsis driven by the lac promoter. The histidine residue in the Fic motif offic-1 was mutated to an alanine by site-directed mutagenesis to givepCL004 (Fic-1H135A). To test the effects of the predicted a-inhibitor,the region containing both the inhibitor and fic-1 genes was insertedpHSG399 to make pCL005. These plasmids were individually trans-formed into E. coli strain DH5a. Bacteria were grown overnight at 37°C

www

in 3 ml of LB broth containing chloramphenicol (20 mg/ml). On the basisof the cell density (OD600), two identical samples with the same concen-tration of cells were withdrawn. One set of samples was subjected to plas-mid DNA isolation with the Zyppy Plasmid Miniprep kit (ZymoResearch), eluting twice, each with 50 ml of TE buffer (pH 8.0). DNAsamples (10 ml) were digested with Bam HI and Sal I or Hind III andXba I, the two restriction enzymes used for subcloning of the genes.The DNA fragments were separated by agarose gel electrophoresis, andthe gels were stained with ethidium bromide (1 mg/ml) for 15 min. Imageswere acquired with a Bio-Rad gel documentation system. The second sam-ple set was used for immunoblotting to detect ICDH as a loading control.

To tightly regulate gene expression, fic-1 and the fic-1H135A mutantswere subcloned into pBAD22 containing the arabinose-regulatable PBADpromoter (53). To express wild-type fic-1 in P. fluorescens, a tightly regulatedplasmid was developed by inserting the PBAD promoter into pBBR1MCS-2(54) to give pCL008, which was used to express fic-1 and the fic-1H135Amutant genes. The relevant plasmids were introduced into P. fluorescensstrains PM933, PM937, and PM938 (table S2). Induction of expressionfrom the PBAD promoter was performed in ABM minimal medium with0.2% arabinose.

Bacterial two-hybrid and b-galactosidase activity assaysThe Cya-based bacterial two-hybrid system (34) was used to examine in-teractions of proteins. Briefly, Fic-1 was fused to the T25 fragment of theCya in pKT25 (34), and proteins known to be involved in DNA replication(table S1) were individually fused to the T18 fragment of adenylyl cyclasein pUT18C (34). Plasmid pairs were introduced into strain BTH101 (34),and the interactions were first evaluated on LB media containing X-gal.Strains exhibiting positive interactions indicated by the hydrolysis of X-galto form blue colonies were retained for further analysis by quantitativemeasurement of b-galactosidase activity in liquid cultures.

To measure b-galactosidase activity, overnight bacterial cultures werediluted into freshmedia at a ratio of 1:20. Cells (100 ml) from saturated cul-tures were collected for enzymatic assays with ONPG (o-nitrophenol-b-D-galactoside) following the standard protocol (55). All assayswere per-formed in triplicate, and the activity was expressed in Miller units.

Purification of recombinant proteinsUnless otherwise stated, E. coli strain BL21(DE3) was used to expressrecombinant protein. To express Fic-1 or Fic-1H135A, the coding regionsamplified by PCR (primer information in table S2) were digested withNde I and Sal I before inserting into similarly digested pET22b(+) to givepET22b–Fic-1 and pET22b–Fic-1H135A. To purify potential substrates ofFic-1 and other proteins such as AntF, their coding regions amplified byPCR were inserted into appropriately digested pET-SUMO that allows theproduction of His6-SUMO–tagged proteins (56). The integrity of eachgene in the expression vector was verified by double-stranded sequencing.

For protein production, strains derived from BL21(DE3) harboring theappropriate expression plasmids were cultured overnight in 5 ml of LBbroth at 37°C, which were then transferred to 500 ml of broth and grownat 37°C to an OD600 of 0.6 to 0.8. After adding IPTG to a final con-centration of 250 mM, incubation was continued at 18°C in a shaker(200 rpm) for 16 hours. The cells were harvested by centrifugation at6000g for 10 min at 4°C. Unless otherwise indicated, the pellets weresuspended in PBS containing 20 mM imidazole and protease inhibitors(1 mM phenylmethylsulfonyl fluoride and 5 mM benzamidine) before lysisby sonication. Cell debris and unbroken cells were removed by centrifu-gation twice at 12,000g for 15 min at 4°C. The cleared lysates were in-cubated with 2 ml of Ni-NTA beads (Qiagen) for 2 hours at 4°C. Beadswith bound proteins were loaded onto a column. After washing the column

.SCIENCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 8

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

with 3 × column volume of PBS buffer containing 20 mM imidazole,protein was eluted with PBS containing 250 mM imidazole. Proteins weredialyzed in PBS buffer with the addition of 20% (v/v) glycerol and 0.5 mMdithiothreitol (DTT). When needed, the SUMO-specific protease Ulp1 (56)was used to cleave the His6-SUMO tag from the recombinant proteins byincubating at 30°C for 3 hours. Proteins in solution were dialyzed in ap-propriate buffers.

To purify recombinant GyrB and its mutants, the coding regions wereamplified by PCR with primers (table S2), which introduced six histidineresidues at the carboxyl termini of the proteins. The products were di-gested with Bam HI and Xho I and inserted into Bam HI/Sal I digestedpGEX-6P1 to give pCL009. Plasmids were introduced E. coli strainXL1Blue and cells induced to express the proteins were prepared as de-scribed above. For purification, cells were lysed in tris buffer [50 mM tris-HCl (pH 7.4), 150 mM NaCl], and the same buffer containing 20 mMimidazole was used to wash the column loaded with the protein-boundbeads. The target protein was eluted with tris buffer (50 mM tris-HCl,pH 7.4) containing 200 mM imidazole. After dialysis against tris buffer[50 mM tris-HCl (pH 7.4), 0.5 mM DTT, and 20% (v/v) glycerol], theGST tag was removed with the PreScission protease by incubationovernight at 4°C. The GST tag and the protease were removed withGlutathione Sepharose 4 Fast Flow beads (GE Healthcare). The purityof all proteins was 95% or higher as evaluated by Coomassie brilliant bluestaining after SDS-PAGE.

Bacterial growth analysisTo evaluate the effects of Fic-1 on bacterial growth, E. coli strain BL21(DE3)was transformed with pHisSUMO, pHisSUMO–Fic-1, or pHisSUMO–Fic-1H135A. Transformants were selected on LB plates containing ka-namycin. Growth was documented by acquiring the images of the coloniesappearing on the same plate after 16 hours of incubation at 37°C. Toevaluate the growth in liquid medium, transformants patched onto LBplates were grown overnight at 37°C; the cells on patches were thencollected and washed twice with PBS. The densities of the cell suspensionswere measured by optical density after appropriate dilutions. Cultures ofidentical density were established in 150 ml of LB broth. After incubationat 37°C for 30 min, each culture was split into three subcultures and incu-bated in a shaker at 200 rpm. The growth of the cells was monitored bymeasuring OD600 at 2-hour intervals. The averages of the readings fromthe three independent cultures were plotted against the incubation time.

For P. fluorescens, derivatives of strain 2P24 harboring the appropriateplasmids were grown to saturation for 14 hours in LB medium containing0.2% glucose. The cultures were 1:100 in ABM medium (51) containing0.2% arabinose, and growth was monitored by measuring OD600 at theindicated time points.

Coexpression of GyrB and Fic-1Plasmid pCL001 (pHSG399::fic-1) (table S2) was cotransformed withpGyrB-His6 (pCL009) into strain XL1Blue and the resulting strain wasgrown in 5 ml of LB broth to saturation. A 200-ml culture establishedby a 1:20 dilution of the overnight culture was grown to an OD600 of0.6 to 0.8, IPTG was added to 100 mM, and the culture was further incu-bated at 16°C in a shaker for 14 hours to induce expression of the proteins.Cells were used to purify GyrB-His6 as described above.

Mass spectrometryPurified proteins were resolved by SDS-PAGE and the bands correspond-ing to GyrB or Fic-1 were excised from acrylamide gels and digested withtrypsin (57). The resulting peptides were analyzed by LC MS/MS on anEkspert nanoLC 400 (Eksigent) connected to a 5600 TripleTOF mass

www

spectrometer (AB Sciex). Peptides were loaded into a C18 trap column(200 mm × 0.5 mm, ChromXP C18-CL, 3 mm, 120 Å, Eksigent) andthe separation was carried out in a capillary C18 column (75 mm ×15 cm, ChromXP C18-CL, 3 mm, 120 Å) at 200 nl/min using the follow-ing gradient: 5% solvent B [solvent A: 0.1% fatty acid (FA) and solvent B:80% acetonitrile/0.1% FA] for 1 min, 5 to 35% solvent B for 60 min, 35 to80% solvent B for 1 min, 80% solvent B for 6 min, 80 to 85% solvent Bfor 1 min, and hold at 5% for 11 min. Full MS spectra of eluting peptideswere collected in the range of 400 to 2000 m/z, and the 10 most intenseparent ions were submitted to fragmentation for 250 ms using rolling col-lision energy. The spectra were analyzed manually by de novo sequencing.

In vitro binding between Fic-1 and GyrBWater-activated Affigel 15 (Bio-Rad) beads were coated with 50 mg ofFic-1–His6 or Fic-1H135A–His6 at 4°C for 14 hours. Beads coated withthe same amount of BSA were also prepared. After washing with threebed volumes of PBS containing 1 mM DTT and 1% Triton X-100, thebeads were blocked with 20 mM tris (pH 8.0) at 4°C for 2 hours. Washedbeads were split into identical samples and 40 mg of GyrB-His6 was addedinto each reaction. Ten percent of the samples were withdrawn as inputcontrols, and the bead slurry was incubated at 4°C on a rotary shaker for4 hours. Unbound proteins were removed by washing the beads with fivebed volumes of PBS, and bound proteins were extracted with SDS loadingbuffer. Proteins separated by SDS-PAGE were detected with Coomassiebrilliant blue staining.

In vitro AMPylation assayThe AMPylation assay was performed as described previously (58). Brief-ly, 20-ml reactions containing 1.5 mg of Fic-1–His6, 10 mg of GyrB-His6,and 5 mCi of 32P-a-ATP (PerkinElmer) in a reaction buffer [25 mM tris-HCl (pH 7.5), 50 mM NaCl, 3 mMMgCl2, 0.5 mM EDTA] were allowedto proceed for 30 min at 35°C. The reaction was terminated with 5 ml of5× Laemmli buffer. After denaturing by boiling for 5 min, samples wereseparated by SDS-PAGE. The gels were first stained with Coomassie bril-liant blue to evaluate protein levels in the reactions and were then driedbefore x-ray film autoradiography to detect AMP-a–labeled molecules.

For dose-dependent AMPylation of GyrB by Fic-1, a 120-ml masterreaction containing 60 mg of GyrB-His6 was established in the AMPyla-tion buffer. Six subreactions were then established by aliquoting the masterreaction into one reaction of 30 ml, four reactions of 20 ml, and one reac-tion of 10 ml. Fic-1–His6 (6.75 mg) was added to the 30-ml reaction, andthe 10-ml mixed sample was added to one of the 20-ml reactions to performa threefold serial dilution. One of the 20-ml reactions was set as a Fic-1–free control. The reaction was started by adding 5 mCi of 32P-a-ATP intoeach test tube. After incubation at 35°C for 30 min, reactions were termi-nated, and the samples were subjected to SDS-PAGE and detection byCoomassie brilliant blue staining as described above.

To test the effects of the AntF on the activity of Fic-1 on GyrB, iden-tical subreactions were prepared from a 120-ml master reaction containing9 mg of Fic-1–His6, into which the indicated amounts of His6-SUMO-AntF were then added, and the tubes were incubated at room temperature(RT) for 30 min. A reaction that did not receive His6-SUMO-AntF wasused as a control. A solution containing 32P-a-ATP and 10 mg of GyrB-His6 was aliquoted into these reactions. After incubation at 35°C for 30 min,samples resolved by SDS-PAGE were analyzed for protein levels and for32P-a-ATP–labeled signals as described above.

Fluorescence microscopy analysisBacterial cells from transformants appearing on selective medium weresuspended in PBS and fixed with 4% paraformaldehyde following an

.SCIENCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 9

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

established protocol (59). Tween 20 was added to 0.5‰ diluted samples;10 ml from each sample was dropped onto coverslips and dried over aflame, and nonadherent cells were removed with PBS. To stain DNA,the coverslips were placed in 50 ml of Hoechst (10 mg/ml) at RT for30 min. After washing three times with PBS, the coverslips were mountedon glass slides with antifade reagent (Vector Laboratories). Samples wereinspected with an Olympus IX-81 fluorescence microscope, and imageswere acquired using a charge-coupled device camera with identical digitalimaging parameters (objectives, exposure duration, contrast ratios, etc.).The images were similarly processed using the IPLab software package(BD Biosciences).

ATPase assayTo measure ATPase activity of GyrB, all proteins were purified in trisbuffer. The ATPase/GTPase Activity Assay Kit (MAK113, Sigma) was usedin 96-well plates at RT (22°C). GyrB (5 mg) or GyrBY109F (5 mg) waspreincubated with the indicated amounts of Fic-1 or Fic-1H135A in 20-ml assaybuffer (40 mM tris, 80 mM NaCl, 8 mM MgAc2, and 1 mM EDTA, pH7.5), and the final volume was adjusted to 30 ml with deionized H2O. Si-milar reactions without any protein or only one of the proteins being testedwere set up as controls. To test the dose-dependent activity of Fic-1, 5 mgof GyrB was preincubated with increasing amounts of Fic-1. The reactionswere initiated by adding 10 ml of 4 mM ATP and were allowed to proceedfor 30 min before being terminated with 200 ml of malachite green. After30 min at RT, the intensity of the signal was measured by determining theabsorbance at 620 nm. The concentration (micromolar) of free phosphatein the reactions was calculated from a standard curve using phosphatestandard supplied in the kit.

Antibodies and immunoblottingFic-1–specific antibodies were generated with Fic-1–His6 by Pocono RabbitFarm and Laboratory following standard protocols. Antibodies were affinity-purified using Affigel 15 (Bio-Rad) coated with Fic-1–His6 following astandard protocol (29). The antibody was used at 1:30,000 for immuno-blotting. Antibody against a-ICDH (59) was used at 10,000. Antibodiesagainst RecA and LexA were purchased from Santa Cruz Biotechnologyand Abcam, respectively, and were used at 1:3000 and 1:5000.

For immunoblotting, proteins resolved by SDS-PAGE were transferredonto nitrocellulose membranes. The membranes were blocked with 5% (w/v)nonfat dried milk in PBST (PBS + 0.2% Tween 20) buffer. After beingwashed three times with the buffer, the membranes were incubated withthe primary antibody at RT for 1 hour. Similarly washed membranes wereincubated with an appropriate IRDye infrared secondary antibody (LI-COR Biosciences), and the signals were detected; if necessary, the intensityof the bands was quantitated by using the LI-COR Odyssey imaging system.

Data quantitation and statistical analysesThe images captured from DNA agarose gels by a Bio-Rad UniversalHood II Gel Doc (Bio-Rad Laboratories) were quantified by using theQuantity One 4.6.9 software (Bio-Rad Laboratories). ImageJ was usedto analyze autoradiograph films in AMPylation assays after scanning byArtix Scan M1 (Microtek). Replications of these experiments were scannedand analyzed with the same parameters. Immunoblots were scanned andquantified using the Odyssey 3.0 (LI-COR Biosciences). Student’s t testwas used to compare the mean levels between two groups, each with atleast three independent samples.

SUPPLEMENTARY MATERIALSwww.sciencesignaling.org/cgi/content/full/9/412/ra11/DC1Fig. S1. Diagrams of the three fic genes present in P. fluorescens strain 2P24.

www.

Fig. S2. Alignment of Fic proteins similar to Fic-1 from other bacteria.Fig. S3. Fic-1 affects the DNA yield of plasmids from different incompatibility groups.Fig. S4. Mutants of Fic-2 and Fic-3 defective in the intramolecular inhibitory motif did notaffect plasmid DNA yield.Fig. S5. Fic-1 inhibits the growth of both E. coli and P. fluorescens.Fig. S6. Fic-1 AMPylates the N-terminal domain of GyrB (1–200).Fig. S7. Fic-1 AMPylates GyrB from P. fluorescens on Tyr111.Fig. S8. The Fic-1Y5A mutant is defective in self-AMPylation and has low activity againstGyrB.Fig. S9. AntF inhibits the activity of Fic-1 by direct interactions.Table S1. Proteins analyzed for interactions with Fic-1.Table S2. Bacterial strains, plasmids, and primers used in the study.Reference (60)

REFERENCES AND NOTES1. K. D. Young, The selective value of bacterial shape. Microbiol. Mol. Biol. Rev. 70,

660–703 (2006).2. G. C. Ulett, M. Totsika, K. Schaale, A. J. Carey, M. J. Sweet, M. A. Schembri, Uro-

pathogenic Escherichia coli virulence and innate immune responses during urinarytract infection. Curr. Opin. Microbiol. 16, 100–107 (2013).

3. S. S. Justice, D. A. Hunstad, P. C. Seed, S. J. Hultgren, Filamentation by Escherichiacoli subverts innate defenses during urinary tract infection. Proc. Natl. Acad. Sci. U.S.A.103, 19884–19889 (2006).

4. S. S. Justice, A. Harrison, B. Becknell, K. M. Mason, Bacterial differentiation, devel-opment, and disease: Mechanisms for survival. FEMS Microbiol. Lett. 360, 1–8(2014).

5. S. S. Justice, D. A. Hunstad, L. Cegelski, S. J. Hultgren, Morphological plasticity as abacterial survival strategy. Nat. Rev. Microbiol. 6, 162–168 (2008).

6. D. J. Horvath Jr., B. Li, T. Casper, S. Partida-Sanchez, D. A. Hunstad, S. J. Hultgren,S. S. Justice, Morphological plasticity promotes resistance to phagocyte killing of uro-pathogenic Escherichia coli. Microbes Infect. 13, 426–437 (2011).

7. A. Prashar, S. Bhatia, Z. Tabatabaeiyazdi, C. Duncan, R. A. Garduño, P. Tang, D. E. Low,C. Guyard, M. R. Terebiznik, Mechanism of invasion of lung epithelial cells by filamentousLegionella pneumophila. Cell. Microbiol. 14, 1632–1655 (2012).

8. A. Prashar, S. Bhatia, D. Gigliozzi, T. Martin, C. Duncan, C. Guyard, M. R. Terebiznik,Filamentous morphology of bacteria delays the timing of phagosome morphogenesisin macrophages. J. Cell Biol. 203, 1081–1097 (2013).

9. S. Humphrey, T. MacVicar, A. Stevenson, M. Roberts, T. J. Humphrey, M. A. Jepson,SulA-induced filamentation in Salmonella enterica serovar Typhimurium: Effects onSPI-1 expression and epithelial infection. J. Appl. Microbiol. 111, 185–196 (2011).

10. C. L. Kirkpatrick, P. H. Viollier, New(s) to the (Z-)ring. Curr. Opin. Microbiol. 14, 691–697(2011).

11. S. X. Sun, H. Jiang, Physics of bacterial morphogenesis. Microbiol. Mol. Biol. Rev. 75,543–565 (2011).

12. Y. Chen, S. L. Milam, H. P. Erickson, SulA inhibits assembly of FtsZ by a simplesequestration mechanism. Biochemistry 51, 3100–3109 (2012).

13. S. Mizusawa, D. Court, S. Gottesman, Transcription of the sulA gene and repressionby LexA. J. Mol. Biol. 171, 337–343 (1983).

14. W. L. Kelley, Lex marks the spot: The virulent side of SOS and a closer look at theLexA regulon. Mol. Microbiol. 62, 1228–1238 (2006).

15. R. Utsumi, H. Tanabe, Y. Nakamoto, M. Kawamukai, H. Sakai, M. Himeno, T. Komano,Y. Hirota, Inhibitory effect of adenosine 3',5'-phosphate on cell division of Escherichia coliK-12 mutant derivatives. J. Bacteriol. 147, 1105–1109 (1981).

16. M. Kawamukai, H. Matsuda, W. Fujii, T. Nishida, Y. Izumoto, M. Himeno, R. Utsumi,T. Komano, Cloning of the fic-1 gene involved in cell filamentation induced by cyclicAMP and construction of a delta fic Escherichia coli strain. J. Bacteriol. 170, 3864–3869(1988).

17. M. Kawamukai, H. Matsuda, W. Fujii, R. Utsumi, T. Komano, Nucleotide sequences offic and fic-1 genes involved in cell filamentation induced by cyclic AMP in Escherichiacoli. J. Bacteriol. 171, 4525–4529 (1989).

18. V. Anantharaman, L. Aravind, New connections in the prokaryotic toxin-antitoxinnetwork: Relationship with the eukaryotic nonsense-mediated RNA decay system.Genome Biol. 4, R81 (2003).

19. L. N. Kinch, M. L. Yarbrough, K. Orth, N. V. Grishin, Fido, a novel AMPylation domaincommon to fic, doc, and AvrB. PLOS One 4, e5818 (2009).

20. A. Garcia-Pino, N. Zenkin, R. Loris, The many faces of Fic: Structural and functionalaspects of Fic enzymes. Trends Biochem. Sci. 39, 121–129 (2014).

21. C. A. Worby, S. Mattoo, R. P. Kruger, L. B. Corbeil, A. Koller, J. C. Mendez, B. Zekarias,C. Lazar, J. E. Dixon, The fic domain: Regulation of cell signaling by adenylylation. Mol.Cell 34, 93–103 (2009).

22. S. Mukherjee, X. Liu, K. Arasaki, J. McDonough, J. E. Galán, C. R. Roy, Modulation ofRab GTPase function by a protein phosphocholine transferase. Nature 477, 103–106(2011).

SCIENCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 10

R E S E A R C H A R T I C L E

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

23. F. Feng, F. Yang, W. Rong, X. Wu, J. Zhang, S. Chen, C. He, J.-M. Zhou, A Xanthom-onas uridine 5'-monophosphate transferase inhibits plant immune kinases. Nature 485,114–118 (2012).

24. D. Castro-Roa, A. Garcia-Pino, S. De Gieter, N. A. van Nuland, R. Loris, N. Zenkin,The Fic protein Doc uses an inverted substrate to phosphorylate and inactivate EF-Tu.Nat. Chem. Biol. 9, 811–817 (2013).

25. J. W. Cruz, F. P. Rothenbacher, T. Maehigashi, W. S. Lane, C. M. Dunham, N. A. Woychik,Doc toxin is a kinase that inactivates elongation factor Tu. J. Biol. Chem. 289, 7788–7798(2014).

26. M. Rahman, H. Ham, X. Liu, Y. Sugiura, K. Orth, H. Krämer, Visual neurotransmissionin Drosophila requires expression of Fic in glial capitate projections. Nat. Neurosci.15, 871–875 (2012).

27. H. Ham, A. R. Woolery, C. Tracy, D. Stenesen, H. Krämer, K. Orth, Unfolded proteinresponse-regulated Drosophila fic (dFic) protein reversibly AMPylates BiP chaperoneduring endoplasmic reticulum homeostasis. J. Biol. Chem. 289, 36059–36069 (2014).

28. A. Sanyal, A. J. Chen, E. S. Nakayasu, C. S. Lazar, E. A. Zbornik, C. A. Worby, A. Koller,S. Mattoo, A novel link between Fic (Filamentation induced by cAMP)-mediatedadenylylation/AMPylation and the unfolded protein response. J. Biol. Chem. 290,8482–8499 (2015).

29. Y. Tan, R. J. Arnold, Z.-Q. Luo, Legionella pneumophila regulates the small GTPaseRab1 activity by reversible phosphorylcholination. Proc. Natl. Acad. Sci. U.S.A. 108,21212–21217 (2011).

30. M. R. Neunuebel, Y. Chen, A. H. Gaspar, P. S. Backlund Jr., A. Yergey, M. P. Machner,De-AMPylation of the small GTPase Rab1 by the pathogen Legionella pneumophila.Science 333, 453–456 (2011).

31. M. P. Müller, H. Peters, J. Blümer, W. Blankenfeldt, R. S. Goody, A. Itzen, The Legionellaeffector protein DrrA AMPylates the membrane traffic regulator Rab1b. Science 329,946–949 (2010).

32. P. Engel, A. Goepfert, F. V. Stanger, A. Harms, A. Schmidt, T. Schirmer, C. Dehio,Adenylylation control by intra- or intermolecular active-site obstruction in Fic proteins.Nature 482, 107–110 (2012).

33. H.-L. Wei, L.-Q. Zhang, Quorum-sensing system influences root colonization andbiological control ability in Pseudomonas fluorescens 2P24. Antonie Van Leeuwenhoek89, 267–280 (2006).

34. G. Karimova, J. Pidoux, A. Ullmann, D. Ladant, A bacterial two-hybrid system basedon a reconstituted signal transduction pathway. Proc. Natl. Acad. Sci. U.S.A. 95,5752–5756 (1998).

35. D. B. Wigley, G. J. Davies, E. J. Dodson, A. Maxwell, G. Dodson, Crystal structure ofan N-terminal fragment of the DNA gyrase B protein. Nature 351, 624–629 (1991).

36. R. J. Reece, A. Maxwell, J. C. Wang, DNA gyrase: Structure and function. Crit. Rev.Biochem. Mol. Biol. 26, 335–375 (1991).

37. J. A. Ali, A. P. Jackson, A. J. Howells, A. Maxwell, The 43-kilodalton N-terminal fragmentof the DNA gyrase B protein hydrolyzes ATP and binds coumarin drugs. Biochemistry 32,2717–2724 (1993).

38. S. C. Kampranis, A. Maxwell, Hydrolysis of ATP at only one GyrB subunit is sufficientto promote supercoiling by DNA gyrase. J. Biol. Chem. 273, 26305–26309 (1998).

39. J. W. Little, Mechanism of specific LexA cleavage: Autodigestion and the role of RecAcoprotease. Biochimie 73, 411–421 (1991).

40. M. Nöllmann, N. J. Crisona, P. B. Arimondo, Thirty years of Escherichia coli DNAgyrase: From in vivo function to single-molecule mechanism. Biochimie 89, 490–499(2007).

41. M. D. Sutton, B. T. Smith, V. G. Godoy, G. C. Walker, The SOS response: Recentinsights into umuDC-dependent mutagenesis and DNA damage tolerance. Annu. Rev.Genet. 34, 479–497 (2000).

42. K. Drlica, X. Zhao, DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiol.Mol. Biol. Rev. 61, 377–392 (1997).

43. M. M. Cox, Regulation of bacterial RecA protein function. Crit. Rev. Biochem. Mol. Biol.42, 41–63 (2007).

44. J. W. Cruz, N. A. Woychik, Teaching Fido new modiFICation tricks. PLOS Pathog. 10,e1004349 (2014).

45. J. M. Diver, R. Wise, Morphological and biochemical changes in Escherichia coli afterexposure to ciprofloxacin. J. Antimicrob. Chemother. 18, 31–41 (1986).

www.

46. K. Lewis, Persister cells. Annu. Rev. Microbiol. 64, 357–372 (2010).47. A. Harms, F. V. Stanger, P. D. Scheu, I. G. de Jong, A. Goepfert, T. Glatter, K. Gerdes,

T. Schirmer, C. Dehio, Adenylylation of gyrase and topo IV by FicT toxins disrupts bac-terial DNA topology. Cell Rep. 12, 1497–1507 (2015).

48. C. D. Aakre, T. N. Phung, D. Huang, M. T. Laub, A bacterial toxin inhibits DNA replicationelongation through a direct interaction with the b sliding clamp. Mol. Cell 52, 617–628(2013).

49. J. W. Modell, T. K. Kambara, B. S. Perchuk, M. T. Laub, A DNA damage-induced,SOS-independent checkpoint regulates cell division in Caulobacter crescentus.PLOS Biol. 12, e1001977 (2014).

50. D. P. Pandey, K. Gerdes, Toxin–antitoxin loci are highly abundant in free-living butlost from host-associated prokaryotes. Nucleic Acids Res. 33, 966–976 (2005).

51. M.-D. Chilton, T. C. Currier, S. K. Farrand, A. J. Bendich, M. P. Gordon, E. W. Nester,Agrobacterium tumefaciens DNA and PS8 bacteriophage DNA not detected in crowngall tumors. Proc. Natl. Acad. Sci. U.S.A. 71, 3672–3676 (1974).

52. S. Takeshita, M. Sato, M. Toba, W. Masahashi, T. Hashimoto-Gotoh, High-copy-numberand low-copy-number plasmid vectors for lacZa-complementation and chloramphenicol- orkanamycin-resistance selection. Gene 61, 63–74 (1987).

53. L. M. Guzman, D. Belin, M. J. Carson, J. Beckwith, Tight regulation, modulation, andhigh-level expression by vectors containing the arabinose PBAD promoter. J. Bacteriol.177, 4121–4130 (1995).

54. M. E. Kovach, P. H. Elzer, D. S. Hill, G. T. Robertson, M. A. Farris, R. M. Roop II, K. M. Peterson,Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying differ-ent antibiotic-resistance cassettes. Gene 166, 175–176 (1995).

55. J. Miller, Experiments in molecular genetics. (Cold Spring Harbor Laboratory, ColdSpring Harbor, NY, 1972).

56. F. Hsu, W. Zhu, L. Brennan, L. Tao, Z.-Q. Luo, Y. Mao, Structural basis for substraterecognition by a unique Legionella phosphoinositide phosphatase. Proc. Natl. Acad.Sci. U.S.A. 109, 13567–13572 (2012).

57. A. Shevchenko, H. Tomas, J. Havlis, J. V. Olsen, M. Mann, In-gel digestion for massspectrometric characterization of proteins and proteomes. Nat. Protoc. 1, 2856–2860(2007).

58. Y. Tan, Z.-Q. Luo, Legionella pneumophila SidD is a deAMPylase that modifies Rab1.Nature 475, 506–509 (2011).

59. L. Xu, X. Shen, A. Bryan, S. Banga, M. S. Swanson, Z.-Q. Luo, Inhibition of hostvacuolar H+-ATPase activity by a Legionella pneumophila effector. PLOS Pathog.6, e1000822 (2010).

60. N. T. Keen, S. Tamaki, D. Kobayashi, D. Trollinger, Improved broad-host-range plas-mids for DNA cloning in Gram-negative bacteria. Gene 70, 191–197 (1988).

Acknowledgments: We thank A. Aronson for critical reading of the manuscript, and L. Csonkaand the Coli Genetic Stock Center at Yale University for bacterial strains. Funding: This workwas supported by NIH grants R56AI103168 and K02AI085403 to Z.-Q.L., National NaturalScience Foundation of China grants (31272082, 31572045) to L.-Q.Z., and the 111 ProjectB13006 to Y.-L. Peng (China Agricultural University). C.L. was supported by a fellowship fromthe China Scholarship Council. Author contributions: L.-Q.Z. and Z.-Q.L. designed thestudy and supervised the work. C.L. performed the genetic and biochemical experiments.E.S.N. performed the MS experiments and analyzed the data. C.L. and Z.-Q.L. wrote themanuscript. Competing interests: The authors declare that they have no competing in-terests. Data and materials availability: The MS data are available at ProteomeXchangerepository (www.proteomexchange.org/) under accession numbers PXD003077 andPXD003076. GenBank accession numbers for the genes described in this study: antF,KT020754; fic-1, KT020755; fic-2, KT020756; fic-3, KT020757; gyrB, KT020758.

Submitted 17 July 2015Accepted 7 January 2016Final Publication 26 January 201610.1126/scisignal.aad0446Citation: C. Lu, E. S. Nakayasu, L.-Q. Zhang, Z.-Q. Luo, Identification of Fic-1 as anenzyme that inhibits bacterial DNA replication by AMPylating GyrB, promoting filamentformation. Sci. Signal. 9, ra11 (2016).

SCIENCESIGNALING.org 26 January 2016 Vol 9 Issue 412 ra11 11

GyrB, promoting filament formationIdentification of Fic-1 as an enzyme that inhibits bacterial DNA replication by AMPylating

Canhua Lu, Ernesto S. Nakayasu, Li-Qun Zhang and Zhao-Qing Luo

DOI: 10.1126/scisignal.aad0446 (412), ra11.9Sci. Signal. 

the effect of Fic-1 on bacterial filament formation.identifies one molecular target for the inhibition of DNA replication by Fic-1 and suggests that additional targets mediateresulted in filament formation in bacteria through a mechanism independent of the SOS response. Thus, this study

expression alsoFic-1Fic-1 inhibited its activity. The authors also identified a Fic-1 antitoxin, which they named AntF. needed for bacterial DNA replication, GyrB, which inhibited GyrB activity and induced SOS response. Self-AMPylation ofplasmids. This suggested that Fic-1 is involved in inhibiting DNA replication. Fic-1 AMPylated the DNA topoisomerase,

Fic-3 or Fic-2 was reduced compared to DNA yield from bacteria encoding Fic-1cultures expressing plasmids encoding . found that DNA yield from bacterialet alUMPylation, and are involved in regulating bacterial filament formation. Lu

Bacterial Fic proteins catalyze diverse posttranslational modifications, such as AMPylation, phosphorylation, andAMPing down DNA replication

ARTICLE TOOLS http://stke.sciencemag.org/content/9/412/ra11

MATERIALSSUPPLEMENTARY http://stke.sciencemag.org/content/suppl/2016/01/22/9.412.ra11.DC1

CONTENTRELATED

http://science.sciencemag.org/content/sci/352/6288/993.fullhttp://stke.sciencemag.org/content/sigtrans/10/479/eaan5621.fullhttp://stke.sciencemag.org/content/sigtrans/10/479/eaal4501.fullhttp://stke.sciencemag.org/content/sigtrans/9/448/ec228.abstracthttp://stke.sciencemag.org/content/sigtrans/9/439/ec177.abstracthttp://science.sciencemag.org/content/sci/353/6298/492.fullhttp://stke.sciencemag.org/content/sigtrans/9/429/ec125.abstracthttp://stke.sciencemag.org/cgi/content/full/sigtrans;2/62/pe14http://science.sciencemag.org/content/sci/329/5999/1660.fullhttp://science.sciencemag.org/content/sci/329/5994/946.fullhttp://science.sciencemag.org/content/sci/334/6053/235.fullhttp://stke.sciencemag.org/content/sigtrans/2/66/ec131.abstracthttp://stke.sciencemag.org/content/sigtrans/2/95/ra71.fullhttp://stke.sciencemag.org/content/sigtrans/2/62/pe14.fullhttp://stke.sciencemag.org/content/sigtrans/4/194/pe42.full

REFERENCES

http://stke.sciencemag.org/content/9/412/ra11#BIBLThis article cites 59 articles, 19 of which you can access for free

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.Science SignalingYork Avenue NW, Washington, DC 20005. The title (ISSN 1937-9145) is published by the American Association for the Advancement of Science, 1200 NewScience Signaling

Copyright © 2016, American Association for the Advancement of Science

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.Science SignalingYork Avenue NW, Washington, DC 20005. The title (ISSN 1937-9145) is published by the American Association for the Advancement of Science, 1200 NewScience Signaling

Copyright © 2016, American Association for the Advancement of Science

on May 31, 2020

http://stke.sciencemag.org/

Dow

nloaded from