Inhibitors of Bacterial Swarming Behavior

Abstract Bacteria can migrate in groups of flagella‐driven cells over semisolid surfaces. This coordinated form of motility is called swarming behavior. Swarming is associated with enhanced virulence and antibiotic resistance of various human pathogens and may be considered as favorable adaptation to the diverse challenges that microbes face in rapidly changing environments. Consequently, the differentiation of motile swarmer cells is tightly regulated and involves multi‐layered signaling networks. Controlling swarming behavior is of major interest for the development of novel anti‐infective strategies. In addition, compounds that block swarming represent important tools for more detailed insights into the molecular mechanisms of the coordination of bacterial population behavior. Over the past decades, there has been major progress in the discovery of small‐molecule modulators and mechanisms that allow selective inhibition of swarming behavior. Herein, an overview of the achievements in the field and future directions and challenges will be presented.


Introduction
Bacteria display numerousw ell-regulated forms of population behavior to colonize ecological niches, cope with adversec onditions, and adapt to competitive or collaborative interactions with other species. Population behaviors range from the formation of sessile biofilms to variousf orms of cellularm otility. One form of motility-the rapid movement of groups of flagellated cells across surfaces-is termeds warming. [1] This behavior is driven by flagellai nat hin-liquid film on semi-solids urfaces. Hereby,s warmer cellsu sually undergo cell differentiation leadingt oe longated snake-or rod-shaped cells with multiple polar or peritrichousf lagella. [2] Other forms of bacterial motility include swimming behavior in three-dimensional liquid space, pili-driven twitching, or appendage-independent forms of active gliding and passive sliding. [3] Althoughm echanistically related,s wimming involves movement of individualc ells instead of the coordinated population behavior of groups of cells in swarming behavior. [1] In some species, the types of flagella used for swarming motility are distinct from that used for swimming and adjustment of gel strength allows for the study of both forms of motilitys eparately. [3b] Swarming represents maybe the most dynamic form of coordinated microbial behaviorst hat is controlled by multiple regulatory layers and consequently may be targeted in diverse ways by chemical modulators. These include globalr egulatory networks like for example master regulators, quorum sensing,t wo-component systems, surfaces ensing, and protease activity and also sens-ing of environmental factorss uch as temperature and salt concentration. [4] For most bacterial species, surfacemotility is facilitated by the production of surfactants, which also enable them to successfully colonize the host environments. [3b] In this review article, we will primarily focus on the connection between swarming motility and small molecules and mechanisms allowing to control swarming.
So far,m any questions such as why someb acteria swarm under certain conditions remain enigmatic. Followingalag phase, swarming colonies can reach expansion rates of about 5-36 mm h À1 and thereby cover an entire agar plate within severalh ours to af ew days. [2, 3b] This rapid colonization of new area may be one of the ecological functions of swarming. Many human pathogens display swarming behavior and swarming also has biomedical relevance. [1] Swarming was first described in 1885 for the urinary tract infective pathogen Proteus mirabilis and regarded as an undesired phenotype preventingt he isolationo fc linical strains from agar plates. [5] Hence, the need for suppressings warming behavior in cultures for diagnostic purposes was recognized early on. However,t he relevance of swarming motility for the infection process itself was only discovered much later.E ver since, swarming motility has been associated with virulence of variousi mportant humanp athogens such as Pseudomonas aeruginosa, [6] Escherichia coli, [7] P. mirabilis, [8] Vibrio cholerae, [9] Salmonella typhimurium, [10] and Clostridium septicum. [11] Many of thesep athogens experience major shifts in the expressionl evels of virulence factors and other pathogenicity related traits correlating with formation of swarm cells. For example, swarming P. mirabilis displays increased virulence by hemolysin, ureolytic and proteolytica ctivities, and invasion behavior in comparison with nonmotile cells. [12] The swarming phenotype also contributed to pathogenicity of P. mirabilis in infection models, [8] and similarly in uropathogenic E. coli expression of flagella was found to be important for the colonization of the upper urinary tract. [7] In P. aeruginosa,v irulence is enhanced under swarming conditions by upregulationo fg ene expression of the type III secretion system as well as numerous virulence factorsi ncluding extracellular proteases and the biosynthesis for siderophores and phenazines. [6] Swarmingb ehaviorm ay furtheri ncrease pathogenicity by facilitation of host attachment and colonization in variouso rganismsr anging from humans to fungi and plants. [13] In addition to increased virulence, swarming bac-teria in many cases exhibit enhanced tolerance against different antibiotics compared with their planktonic counterparts. [6,14] High cell densities of swarming bacteria protected S. typhimurium even from several orders of magnitude higher concentrations of antibiotics than swimming cells which only move at low cell densities. [15] Mixed species swarmsa lso allow the transport of nonmotile bacterial speciesw ith mutual benefits, whereby ac argo speciesm ay contribute with antibiotic resistancem echanisms to the detoxification of the environment. [16] Due to its impact on virulence and antibiotic tolerance, swarming motility is an importantp athogenicity related trait. Inhibiting bacterials warming behavior may thus have medical potentialf or treating or preventing infectious diseases.H owever,t he molecularm echanismsi nvolved in the regulationo f swarming fundamentally differ from species to species and their detailed understanding is in many cases still incomplete. [17] Surfacem otility requires thec ells to overcome biophysicalc hallenges such as surfacew etting, friction, and surface tension. [18] Also aw ide range of environmentalc onditions, nutrients, and physical parameters influence swarming motility and diverse physicaland chemical signals integrate into its regulation. [19] Thus, swarmingi nvolves intertwined regulatory networks operating on metabolic, signal transduction, and geneexpression level. [18,19] Consequently, strategies for swarmingi nhibition are diversea nd involve aw ide variety of different compound classes and modes of action. The literature on swarming modulation by smallm olecules is vast and dispersed across different research fields. Althoughm any excellent reviews on bacterialm otility and its biological regulation exist, [1,[17][18][19] no informative and comprehensive overview on the chemistry of controlling swarming behavior has been reported so far.I nt his article we will review the current status and highlight new developments of swarming-inhibitory compounds as well as provide mechanistic insights into their mode of action.

Swarming and Bacterial Signaling
One wayb acteria regulate their swarming behavior is through chemicals ignals. Different types of signaling pathways exist, the most prominent of which are quorum-sensing systems. Quorum sensing is ac ell-to-cell signaling strategy inducing gene expression in dependence of bacterialp opulation density.T he corresponding small-molecule signals are produced and accumulate during population growth. Ar eceptor sensing these signals positively regulates transcription of variousg enes including genes for the biosynthesis of the signal itself-hence also called autoinducer. This synchronizes gene expression in a population-density dependent manner and allows the coordinated production of virulencef actors such as toxins,e nzymes, or specific metabolites. [20] Examples for signaling molecules are the widely distributed autoinducer2( AI-2), the highly diverse class of N-acyl-homoserine lactones (N-acyl-HSLso rA HLs) in gram-negative bacteria, [21] as well as various autoinducing peptides (AIPs)i ng ram-positiveb acteria. [22] Although in some species quorum-sensing signals directly control swarmer cell dif-ferentiation, they regulate in others the production of biosurfactants that contribute to swarming motility by lowering surface tension. Examples of the latter are rhamnolipid of P. aeruginosa or surfactin of Bacillus subtilis. [23] Given that quorum sensingh as important impactso ns warming behavior,i nterference with its signaling can be applied to suppress swarming motility.

Inhibition of AI-2 signaling
AlthoughA I-2 is the most common quorum-sensing signal used by many different speciesa nd produced by gram-negative as well as gram-positiveb acteria, only af ew approaches have been reported in whichA I-2 signaling has been targeted for swarming inhibition. For E. coli,s warming-cell differentiation has been shown to be regulated by the central FlhC 2 D 2 master regulator the transcription of which is presumablya ctivated by AI-2 through the two-component system QseBC ( Figure 1). The FlhC 2 D 2 regulator in turn activates the fliA gene which encodes as igma factor specific for flagellar operons. [4b] In pathogenic E. coli strains, AI-2 plays an important role for virulencea nd an anoemulsiono f2 .5 %l imonene was found to interfere with AI-2 quorum sensing of E. coli O157:H7 (EHEC). Hereby, both swimming and swarming motilities were repressed. [24] The biosynthesis of the AI-2 signal is carriedo ut through cleavage of S-ribosylhomocysteine by LuxS ( Figure 1). [25] For signal detection, AI-2 is phosphorylated and derepresses transcription of target genes through binding to LsrR. [26] Fimbrolides,ac lass of halogenated furanones, are important inhibitors of the LuxS signal synthase and therebyo f quorum sensing by AI-2. [27] Fimbrolides have been initially discovereda sn atural products from the marine red alga Delisea pulchra and ag reat diversity of natural and synthetic derivatives hasb een investigated. [28] Af uranone (1)i nhibited biofilm formationa nd swarming but not swimming motilityi nE. coli and strongly antagonized the quorum sensing by AI-2. [29] The same furanone also inhibited swarming of B. subtilis. [30]

BlockingAHL receptors
Halogenated furanones have been additionally described to target the LuxE subunit of the luciferase complex of Vibrio and N-acyl-homoserine lactones (AHL)-based quorum sensing through destabilization of homologues of the LuxR-regulator. [27,31] AHLs are the largestc lass of quorum-sensing signals in gram-negative bacteria that are produced through N-acylation of S-adenosyl-l-methionine( SAM) andc yclization to g-lactones by homologues of the synthase LuxI ( Figure 2, left). The signals are detected by binding to homologues of the transcription factor LuxR. [32] In many species, AHLs have major impact on swarming regulation because they are regulators of, for exam-ple, the biosynthesis of the surfactant serrawettin through LuxR in Serratia spp. (Figure2,l eft). Serrawettin promotes swarming motilityb yr eduction of surfacet ension. Consequently, targeting AHL-based quorum sensing has been of central interest for swarmingi nhibition. Twod ifferently brominated furanones( 1)a nd (2)o fD. pulchra inhibited AHL-dependent swarming motility of the enterobacterium Serratial iquefaciens which was restored in an AHL-negative mutant by supplementation with N-butanoyl-l-homoserine lactone (C4-HSL). [28b] The mechanismo fs warming inhibition involves the blockage of the biosynthesis of the surfactant serrawettin W2 as mentioned above through binding to LuxR. [33] Surprisingly, only one of four brominated furanones isolatedf rom D. pulchra inhibited swarming of the uropathogen P. mirabilis. [34] All four furanones (1-4)i nhibited swarming of different uncharacterized environmental strains of bacteria isolated from rock surfaces as well as from samples of D. pulchra. [35] Ta rgeting AHL receptors( LuxR homologues) has been maybet he most frequently employed strategy fori nterfering with AHL-based quorum sensing. Especially AHL signal analogs that mimic the native AHLs are promising candidates for inhibitors. For example, AHL signalingc an be inhibited by synthetic N-acylc yclopentylamides (Figure 2, left). [36] Am utant strain of enterobacterium Serratia marcescens that was unable to produce AHLs was nonmotile in as warming assay.E xogenous supply of N-hexanoyl-l-homoserine lactone (C6-HSL) restored the swarming phenotypea nd competitionw ith 50 mm N-nonanoyl cyclopentylamide( 5)r esulted in complete swarming inhibition. [37] Some speciess uch as the human pathogen P. aeruginosa even comprise more than one AHL-based quorum sensing system.I nP. aeruginosa,t he LuxI/LuxR homologues RhlI/RhlR and LasI/LasR utilize the signals N-butanoyl-l-homoserine lactone (C4-HSL) and N-(3-oxododecanoyl)-l-homoserine lactone (3-oxo-C12-HSL), respectively (Figure 2, right). These AHL-based quorum-sensing systems are hierarchically interconnected by the master regulator LasR with furtherq uorum-sensing and two-components ystems to control virulence in P. aeruginosa. [38] Recently discovered clinical isolates of P. aeruginosa from cystic fibrosis patients revealed an exceptional plasticityi nt he hierarchical regulation of quorum sensing whereby the RhlI/  RhlR system could compensate the loss of functional LasR. [39] The production of the swarming surfactant rhamnolipid which Pseudomonas requires to lower surface tension is RhlR regulated by transcriptiono ft he rhl genes. TheM eijler group developed synthetic AHLs with an isothiocyanate (ITC) warhead mimicking 3-oxo-C12-HSLo fP. aeruginosa. [40] These compounds and especially a b-fluorinated derivative ICT-F (6)c ovalently blocked the LasR receptor at Cys79 and inhibited swarming motilityb y4 4% at 150 mm and by 34 %a t2 0mm and also reducedp yocyanin production( Figure 3a). In contrast, the bro-minatedI TC-Br (7)d id not bind covalently and was aL asR agonist that increased swarming motilityu pt o2 .5-fold at 20 mm of ITC-Br in P. aeruginosa PA14. [40] High-throughput screening of ac ompound library against reporter strains revealed the plant-produced flavonoids phloretin, chrysin, and naringenin as potent inhibitors of the LasR and RhlR quorum-sensing receptors of P. aeruginosa. [41] Additionally,a lso flavonoids like quercetin (8), baicalein, and pinocembrin exhibited inhibitory activity whereby the presence of as pecific pattern of two hydroxyl-groups on the flavonoid Aring appeared to be required for activity ( Figure 2).
Flavonoids were found to be allosterici nhibitors of these quorum-sensing receptors and prevented their binding as transcriptionf actors to DNA. Twoo ft he most active compounds, phloretin (9)a nd 7,8-dihydroxyflavone weref inally tested on quorum-sensing-controlled behaviors of P. aeruginosa and completely abrogated swarming at 100 mm. [41] The flavonoid quercetin (8)c onsiderably reduceds warming motility of P. aeruginosa and Yersinia enterocolitica at 132 mm. [42] In Proteusv ulgaris,5 0mm of quercetin (8)n ot only inhibited the production of N-octanoyl-l-homoserine lactone (C8-HSL) by 81 %a nd caused an almost equal reduction in swarming area,b ut also supposedly interfered withs warming by binding to the sigma factor FliA which regulates flagellar operons (Figure 2, left). [43] Av irtual docking-approach against the AHL receptor LasR identified salicylic acid and chlorzoxazonea sp otential quorum-sensing inhibitors of P. aeruginosa which was confirmed biochemically through LasR and additionally RhlR and resultedi ni nhibition of swarming of S. liquefaciens in the millimolar range. [44] Also complex natural-product mixtures and ex-tracts have been found to exhibit quorum-sensingi nhibiting activities affecting swarming behavior.F or example, propolisbee glue-antagonized AHL-based quorum-sensing signaling in RhlR-and LasR-dependent reporter strains and reduced swarming activity of P. aeruginosa. [45] Some signals may even lead to crosstalk between different quorum-sensing systems. An example are diketopiperazines (DKPs), cyclic dipeptidesi nvolved in trans-kingdom interactions of bacteria with eukaryotes [46] and inter-species signaling between gram-negative and gram-positive bacteria. [47] DKPs such as cyclo(DAla-l-Val), cyclo(L-Pro-l-Tyr) (10), and cyclo(l-Phe-l-Pro) were isolated from culture supernatantso fv arious gram-negative bacteria including Pseudomonads, P. mirabilis, Citrobacter freundii,a nd Enterobacter agglomerans and recombinant LuxR-based AHL biosensor assay revealed that they competew ith the site of AHL binding and thereby antagonize quorum sensing.C yclo(l-Pro-l-Tyr) (10)r educed swarming of wild type S. liquefaciens as well as of a DswrI mutant for which swarming motilityd ependso ne xternals upply of N-butanoyl-l-homoserine lactone (C4-HSL) ( Figure 2, left). [48] In many cases, however,t he cellular targetso fq uorum-sensing inhibitors or their compound classes have not yet been clearlyi dentified.H ereby,p henotypic or transcriptional analyses have often tentatively pointed to interference with AHLbased quorum sensing as likely mechanismofswarming inhibition. An AHL-derived N-decanoyl-l-homoserine benzyle ster (11)f or example inhibiteds warm expansion and dendritic swarming pattern between5 0a nd 100 mm and reduced expression of both las and rhl genes as well as production of virulencef actors including rhamnolipids (Figure 2, right). [49] At 136 mm and higherc oncentrations, curcumin (12)i nhibited swarming motility of E. coli, P. aeruginosa PAO1, P. mirabilis,a nd S. marcescens and interfered with AHL-based quorum sensing in av iolaceina ssay ( Figure 3b). [50] At high concentrationso f around1 .5 mm,c affeine inhibited AHL production in P. aeruginosa andr educed swarming motility [51] and zingerone inhibited swarming, swimming, and twitching motility at 5mm and also decreased the production of AHLs. [52] Many further natural products and synthetic compounds have been postulated to inhibit quorum sensingo fP. aeruginosa at relatively high concentrationst hrough LasR whereby swarming motility, but not growth, was inhibited. Examples are, trans-anethole with ar eductiono fs warming motilityb y6 4% at 6mm [53] or pyridoxal lactohydrazone with ar eductiono fs warmingm otility by about 35 %a t3 2mm and % 70 %at1 26 mm. [54] The non-methylated version of the pyrrolidin alkaloid (R)norbgugaine superficially resembles 3-oxo-C12-HSLa nd inhibited swarming motility and production of virulence factors of P. aeruginosa. [55] The anti-inflammatory drugs diclofenac and also ketoprofen were shown to inhibit swarming motility of P. aeruginosa at 5mm concentration withouta ny growth inhibition. Reduced production of virulence factors as wella sa ctivity in an AHL-quorum-sensingi nhibition screen suggested that these compounds inhibited swarming through the quorum-sensing circuits with the molecular targets yet to be identified. [56] Adiazaborine-based copolymer with quorum-sensing inhibitory activity in av iolaceina ssay showeds warming inhibition by about 50 %a gainst P. aeruginosa PAO1 at ac oncentration of 100 mgmL À1 ,w hereas the MIC (minimal inhibitory concentration) was determined to be 10 times higher. [57] At relatively high concentrations of 10-12 mm,t he food additives diallyl disulfide (DADS) andm ethyl 2-methyl-3-furyl disulfide (MMFDS) inhibitedC 6-HSL production of the enterobacterium Hafnia alvei,r educed expression levels of luxI and luxR and inhibited swarming by more than 70 %. [58] In S. marcescens,p roduction of its red pigmentp rodigiosin is under control of AHL-based quorum sensing. Methanolic extracts of the benthic brown alga Padina gymnospora inhibited production of this pigment and activity guided fractionation led to a-bisabolol as active compound. Furthermore, a-bisabolol inhibitede xtracellular protease, biofilm formation and swarming motilitya ta nd above 450 mm suggesting interference with AHL-basedq uorum sensing as mechanism. Swarming was abolished completelya t1 .8 mm without inhibiting growth. [59] At much lower concentrations between 17 and 34 mm,p hytol (13)r educed virulence factor production of S. marcescens and strongly inhibiteds warming motility ( Figure 3c). [60] The activity of phytol was presumably mediated through quorum-sensingi nhibition because it resulted in transcriptional down-regulation of many quorum-sensing-controlled genes including the swarming differentiation masterregulator genes flhC and flhD. Finally,t reatment of rats with phytol in an acute pyelonephritis model even ameliorated the infection with S. marcescens. [60] 2.3. Interspecies activity of alkyl quinolone signals P. aeruginosa comprises am ulti-layered network of intertwined quorum-sensing systems regulating its virulence and population behaviors like swarming. In addition to the two AHLbased quorum-sensing systems introducedp reviously, P. aeruginosa also utilizes an alkyl quinolone-based system as wella s the more recently discovered integrated quorum-sensing (IQS) system. [38] The alkyl quinolone-based systems signal through congeners of the Pseudomonas Quinolone Signal (PQS, (14)) and itsb iosynthetic precursor HHQ (15)a nd the receptor PqsR (also known as MvfR) and possibly many further interaction partners( Scheme 1). [61] In P. aeruginosa,P QS as well as C4-HSL are known to regulate the transcription of rhlRg enes, thus modulating rhamnolipid production.I na ddition, HHQ andP QS have been implicated in interspecies and even interkingdom interactions. [62] For example, PQSa t5 0mm inhibited swarming of Pseudomonasputida and reduced biofilm formation by interferencew ith signaling and iron-uptake. [63] HHQ and PQS also repressed swarming and flagella-independent forms of motility in other gram-negative andg ram-positive bacteria.
[62b] Althought he mechanism of motility reduction by HHQ and PQS remained obscure in this studyi tw as presumably unrelated to their role as quorum sensing signals since homologs of the PQS signaling system are restricted to only af ew species of Pseudomonas and Burkholderia. [64] P. aeruginosa shares a common environment with Bacillus atrophaeus in soil and PQS completely abrogated swarming of B. atrophaeus at 10 mm, whereas HHQ at the same concentration only led to minor reductiono fs warming. [62b, 65] Developmento fs ynthetic HHQ derivatives with substitutions at the anthranilate-derived ring of the quinolone core and variations of the alkyl chain resulted in several potent compounds with enhanced anti-swarming activity.T wo of them (16 and 17)e ven completely abrogated swarming motility of B. atrophaeus (Scheme 1). [65] 2.4. Enzymatic quenching of the signal In addition to disrupting AHL signaling through inhibition of its production or blocking of the signalr eceptor,a lso enzymatic degradation of the signal itself is leading to quorum quenching and altereds warming behavior. [66] This can be accomplished by lactonases which hydrolyze the g-lactoner ing of AHLs. An example is provided by the mammalian paraoxonase enzyme family that degraded and thusq uenched AHL-based quorum sensingo fP. aeruginosa whereby swarming was significantly reduced already at concentrations of 3 mgmL À1 of human serum paraoxonase 1. [67] Another lactonase Ahl-1 from Bacillus weihenstephanensis isolate-P65 at 0.5 mg mL À1 also inhibited AHL accumulation and reduced virulence-factor productionand swarming of P. aeruginosa. [68] Screening of am etagenomic library revealed HqiA as novel AHL lactonase family enzyme that quenched AHL signals and the hiqA gene introduced in the swarming plant pathogen Pectobacterium carotovorum reduced its motility and production of virulence-related maceration enzymes. [69] Given that HHQ and PQS inhibitswarming of severalbacterial species, enzymatic quenching of these molecules by other bacterials peciesm ay affect motility in interspecies interactions. For example, the dioxygenase Hod from Arthrobactern itroguajacolicus and the enzymeA qd from Mycobacterium abscessus have been described as PQS-degrading enzymes. [70] So far,h owever,e ffects of these enzymes on HHQ-and PQS-mediated swarming inhibition still remain to be demonstrated.

Other signaling systems
In addition to its multiple quorum-sensing systems, P. aeruginosa also comprises al arge diversity of distinct two-component systemsr egulating virulence. [71] Each of them is composed of a histidinek inase (HK) sensinge xternal stimuli and ar esponseregulator protein that alters gene expression upon phosphory- lation by the kinase. The many two-component systems for Pseudomonas have been shown to be intricately involved in swarming regulation for example through the action of the response regulator GacA,w hich is activated by the HK GacS. GacA is connected to swarming throught he RhlI/RhlR system through several regulatory steps. Benzothiazole-based histidine kinase inhibitors (Rilu-1 (18), , and Rilu-12 (20)) reduced PQS signaling, decreased rhamnolipid production and drastically impaired swarming motility at 200 mm (Figure 4).
Gene-expression analysis suggested that these benzothiazoles inhibited the sensory kinase GacS wherebyt he transcription of the response regulator gacA and also the flagellar regulator fleQ was decreased. [72] In some cases also chemoattractants may be important for swarming motility.T his was demonstrated for P. mirabilis on minimal medium, in which swarming depended on the amino acid l-glutamine as signall ead to swarmer-cell differentiation and up-regulation of the expression of flagellin (fliC)a nd hemolysin (hpmA). The glutamine-analogue g-glutamyl hydroxamate interfered with this signaling and inhibited swarming at 10 mm. [73] Consequently,c hemical signaling and the modulation of its activity by small molecules is ap romising strategy for controlling swarming and other population behaviors in different species. The diversity of signaling pathways even within as ingle speciess uch as P. aeruginosa and the manifoldi nteractions of microbial signals across species give rise to al arge and yet only partially exploredc hemical space for specific and selective inhibitors of swarming behavior.

Sub-Inhibitory Concentrations of Antibiotics
Antibiotics are highly important drugs against pathogenic bacteria that contribute immensely to human health. Many antibiotics are naturally produced by soil microbes and it has been proposed that somea ntibiotics may even have roles in the ecosystem beyond inhibiting growth of competitors. [74] These antibioticsa re regarded to serve at sub-lethal concentrations, that is, below MIC as cell-cell communication signals and regulate transcriptiono fc ertain genes, including that of important virulence factors. [75] Accordingly,s omea ntibiotics control at low concentrationsm icrobial behavior and also affect swarming motility. The macrolide azithromycin, for example, showed swarming-inhibitory effects against P. aeruginosa and P. mirabilis in various studies. Hereby, the best inhibition with azithromycin (21)w as at ac oncentration of about 21 mm (1/16 MIC) with more than 80 %i nhibition of swarming of P. aeruginosa PAO1 (Scheme 2). [76] In another study,1 1 mm azithromycin inhibited the swarming of 15 clinical isolateso fP. aeruginosa from 18 to 73 %, whereas swarming of all clinicali solates of P. mirabilis was already completely inhibited at 5 mm. [77] Swarming inhibition by azithromycin correlated with suppressede xpression of flagellin in P. aeruginosa and P. mirabilis. [78] Azithromycin also reduced expression of lasI/lasR and rhlI/rhlR in P. aeruginosa and inhibited AHL production. [76,79] Some macrolide antibiotics like erythromycin and clarithromycina lso inhibited swarming and flagellin expression, [78] whereas for example the macrolide rokitamycin had no effect on the expression of flagellin and consequently did not inhibit swarming. [78] Also, b-lactam antibiotics as inhibitors of cell-wall biosynthesis affect virulence and population behavior at concentrations below the MIC.F or example, the third-generation cephalosporin ceftazidime (22)i nhibited virulence of P. aeruginosa PAO1 and PAF97 and reduced swarming motilityb ya round8 0% at 0.9 and 3.7 mm,r espectively (Scheme 2). [80] The antibioticsc efotaxime (23), ciprofloxacin (24), chloramphenicol, and trimethoprim completely blocked swarming of the gram-negative pathogen Salmonella enterica (ser.T yphimurium) at subgrowth inhibitory concentrations of 3.5, 0.02, 6, and 3 mm,r espectively.I nc ontrast, amikacin, colistin, kanamycin,a nd tetracycline did not inhibit swarming of S. enterica (ser.T yphimurium). While cefotaxime (23), ciprofloxacin (24)a nd trimethoprim inhibited polar-chemoreceptora rray assembly of S. enterica (ser.T yphimurium) that is essential for swarming, chloramphenicol inhibited swarming by ad ecrease in flagellation (Scheme 2). [81] Many furthera ntibiotic classes have been linked to regulatory effects on bacterial behavior at sublethal concentrations. [75c] For example, also the aminoglycoside gentamicin, like azithromycin (21), reduced lasI/lasR and rhlI/rhlR expression and AHL productioni nP. aeruginosa and considerably impaired swarming motility at approximately 0.2 mm (1/16 MIC) by over 70 %. [76] At 1 = 4 of the MIC, gentamicin (MIC % 0.06-0.2 mm)a nd amikacin (MIC = 1.7-3.4 mm)r esulted in 30-60 %s warming inhibition of variousc linicali solateso fP. mirabilis. [82] The gyrase inhibitors nalidixic acid and novobiocin completely inhibited swarming of E. coli at 20 and 200 mm,r espectively. [83] In the lower micromolar range, also sulfonamides such as sulfamethazin blocked swarming of the majority of 250 strains of P. mirabilis and P. vulgaris tested. [84] Doxycycline was reportedt oi nhibits warming of P. aeruginosa PAO1 in the lower micromolar range with more than 60 %i nhibition at 4.5 mm likely through targeting of quorum sensing. [85] In addition, differenta ntibiotic peptides inhibited swarming at sublethal concentration. For example, the naturally occurring pseudopeptidea ntibiotic actinonin at below MIC concentrationsb etween 0.05 and 0.5 mm reduced swarming motility of S. enterica (ser.T yphimurium) and Vibrio vulnificus. [86] As mall cationic peptide( KRFRIRVRV-NH 2 )w ith weak antibiotic activity considerably inhibited (by > 70 %) at sub-MICc oncentration of 4 mm the swarming motilities of P. aeruginosa PAO1 and PA14 and Burkholderia cenocepacia. Hereby,t he transcription of several flagellar genes and rhlB for rhamnolipidp roduction was downregulated. [87] As eries of cationic antimicrobial peptides with repeatingt ryptophan-arginine motif was tested against the swarming of E. coli. In this study,t he hexapeptide (RW) 3 -NH 2 showedt he strongest swarming inhibition with almost complete blockage of swarming at ac oncentration of 25 mm. [88] Cationicp eptides are known to exhibit their antimicrobial activity by targeting cell membranes [89] and may thusalso disrupt flagellar integrity. [88] The inhibition of swarming motilitya tl ow concentrations appearst ob eacommont heme for many but not all antibiotics. In some cases, like amikacin, swarming inhibition even seems to be species specific. [81,82] Although the mechanismsb y which antibiotics in low concentrations inhibit swarming behavior are so far not conclusively understood, targeting of quorum sensing as well as direct interference with the regulation of flagellar gene expression or flagellar integrity are likely central concepts.A ntibiotics also may lead to long-term regulatory changes in bacterial cells which have been pre-exposed for extended time to sublethal concentrationso fa ntibiotics. For example, pretreatment of E. coli with approximately 1 mm ( 1 = 2 MIC)g entamicin through downregulation of succinate dehydrogenase (sdh)g enes inhibited swarming but not swimming motility in af umarate-dependent manner. [90] Fumarate metabolism also was found to be important for swarming motility of P. mirabilis. [91] Also ac ontinuousl ow-dose pre-exposure of P. aeruginosa to erythromycin (2 mm)a nd clarithromycin (1 mm)f or 2-18 months led to approximately 70 %r eduction of swarming motilitya nd attenuated virulence although it did not affect the MIC value. [92] Whether these effectsa re caused or facilitated by geneticm utations or entirely rely on regulatory changes that prevailf or several generations after antibiotic exposure has so far not been investigated.

Secondary Plant Metabolites
Plants produce an enormous diversity of secondarym etabolites and great deal of research has focusedo nn atural products and their effects on bacterial population behaviors including swarming motility. For example, different plant extracts inhibiteds warming of E. coli O157:H7 (EHEC) whereby extracts of the sedge grass Carex dimorpholepis containing high concentrationso ft he phytoalexine trans-resveratrol were the most potent. Swarming of EHEC was inhibited by 44 mm trans-resveratrol (25)w hich correlated with transcriptional repression of the motilityg enes flhD, fimA, fimH,a nd motB (Scheme 3). [93] At 263 mmtrans-resveratrol completely inhibited swarming of P. mirabilis and significantly reduced swarming already at 66 mm. Am utant of the gene rsbA restored swarming of P. mirabilis in presence of trans-resveratrol with preserved flagellin production and elongated-cell phenotype, suggesting that the regulatory protein RsbA mediates inhibition of swarmer cell differentiation by trans-resveratrol. [94] Resveramax, af ormulation of trans-resveratrol further inhibited swarming of P. aeruginosa and global effects on quorum-sensing-related phenotypes were observed. [95] In another study,a lso trans-oxyresveratrol (26)a nd trans-piceatannol (27)a lmost completely abolished swarming of P. aeruginosa between 100 and 200 mm without inhibiting growth (Scheme 3). Transcription analysis revealed downregulation of the las and rhl quorum-sensing regulatory circuits. [96] The structurally related chlorogenic acid only slightly inhibited swarming of P. aeruginosa but also exhibitedg lobal effects on quorum-sensing-controlled virulence factors. [97] The compound (Z,Z)-5-(trideca-4',7'-dienyl)-resorcinol that was isolated from the plant Lithream olleoides significantly inhibited swarming motility of P. mirabilis at 28 mm and completely abolished swarming at 433 mm. [98] Furthermore, many similarp lantderived phenolic compounds including caffeica cid, cinnamic acid, ferulica cid, and vanillic acid have been reported to inhibit swarming of P. aeruginosa at 4mm. [99] Also, tanninss uch as proanthocyanidins are important phenolic compounds produced by many plant species. Cranberry proanthocyanidin extracts and pomegranate extractsc ontaining the related punicalagin completely abolished swarming of P. aeruginosa at 100 mgmL À1 without inhibiting growth. Both extracts did not affect swimming motility.A ddition of rhamnolipid partially restored swarming, suggesting that the mechanism involved repression of biosurfactant production. [100] Cranberry products also transiently impaired swarming of urinary tract infective P. mirabilis. [101] More defined tannins such as pure epigallocatechin gallate and tannic acid (28)b locked swarming of P. aeruginosa down to approximately 20 and3mm, respectively (Scheme 3). [102] In contrast, methyl gallate, which corresponds to as tructural motif of tannic acid only exhibited low swarming inhibitory activity against P. aeruginosa in the range of several hundred micromolar. [103] Neutralized tannic acid at 12 mm (0.02 %( w/v)) also inhibitedt he swarming of all 27 strains of P. mirabilis tested. [104] Many furtherp lant metabolite classes inhibit swarming. Examplesa re terpenes of which citronellol poorly inhibited swarming of P. mirabilis at 1.9 mm [105] and the related citral ( 29) which considerably inhibited swarming motility of the foodborne pathogen Cronobacter sakazakii already at 113 mm and repressed various virulence genes. [106] At millimolar concentrations also the red pigment brazilin from the wood of the Caesalpinia family, [107] cinnamaldehyde, [108] and 2-phenethylamine [109] inhibited swarmingm otilityo fd ifferent species. A 10'(Z),13'(E)-heptadecadienylhydroquinone (HQ17-2) isolated from the lacquert ree inhibited swarming motility of P. mirabilis between 36 and 145 mm through the two-components ystem RcsB which controlst he flhDC genes encoding the flagellar master regulator FlhD 2 C 2 . [110] With exception of tannins, plant metabolites exhibitedc omparably low activity on swarming bacteria.T he mechanisms hereby may be as diversea st he compound classes and range from inhibition of surfactant production to regulatory effects on flagellar gene expression.

Off-Target Effects of Synthetic Compounds
Off-target activities of drugs, pesticides, and other xenobiotics have in some cases also led to inhibition of swarming behavior. The gastrointestinal drug solfacone (30)f or example, whichi s also present in herbs used in traditional Chinese medicine, sig-nificantlyi nhibited Heliobacter pylori swarming at ac oncentration of 22 mm withouta ny growth inhibition (Scheme 4). [111] 3-Amino 1,8-naphthalimide (31), an analogue of virstatin, ac ompound targeting the cholera toxin regulator To xT,w as highly effective against swarming of V. cholerae at ac oncentrationo f about 12 mm withouta ny effect on the growth of the bacteria. This effect could be attributed to an inhibition of chemotaxis, but the secondary target was not further identified. [112] The drug ambroxol, commonly used in asthma andc hronic bronchitis, completely inhibited swarming motility of P. mirabilis at high concentrations of 2.4 mm. [113] Furthermore, the effects of ar ange of psychotropic drugs was tested against another Proteus and Proteus-related strains. Of these compounds, the antihistamine promethazine (32)e xhibitedt he best inhibition effects against P. vulgaris at 150 mm, which was several times lower than the MIC value (Scheme 4). [114] Swarming inhibition could be antagonized by K + + and Na + + ions, suggesting that interference of promethazine with ion homeostasis would adversely impact flagellar motility. [114] Different psychotropic drugs were also tested against P. vulgaris, P. mirabilis,a nd Morganella morganii,w hereby the antidepressant sertraline inhibited all strain's swarming motilitya t about 100 mm independent of its MIC which was 2-16 times higher. [115] Swarming of P. mirabilis and P. vulgaris wase fficiently blocked by the synthetic compound p-nitrophenyl glycerol, which completely abolished swarming at 0.1 and 0.2 mm for more than 24 ha nd depending on culture conditions even for > 80 h, whereas growth was only affected above 0.5 mm.H owever,s warming cells exposed to p-nitrophenyl glycerols eemed to have developedr esistancea nd resumed swarming motility soonert han unexposed cells. [116] p-Nitrophenylg lycerolh as been used in clinicall aboratories to block swarming for bacterial isolation and also other studies reported complete swarming inhibition for Proteus between 0.2 and 0.7 mm as well as downregulation of virulence factors. [117] Many different mycotoxins,f ungicides, insecticides, and herbicides affected in the upper micro-to millimolar range the swarming motilities of P. mirabilis and Azospirillum brasilense. [118] The chromogenic b-galactosidase substrate 5-bromo-4-chloro-3-indolyl-b-d-galactopyranoside (X-Gal) reduced or inhibited swarming of different Vibrio species,i ncluding V. cholerae, Vibrio mimicus, V. vulnificus, Vibrio alginolyticus,and Vibrio parahaemolyticus at 235 mm without affecting viabilityb ut facilitated swarming motility of P. mirabilis and S. marcescens. [119] Although the mode of swarming inhibition of most of these compounds remains unexplored, their pharmacophore properties, as well as their rela-tively potent activities, suggest specific interference with cellular processes required for bacterial motility that warrantf urther investigation.

Fatty Acids
Swarming motility is dependent on many factors like for example the population density and the concentrationo fs odium ions. Furthermore, the surfacew etness of the solid mediumi s fundamentally important and ac hallenge for standardizing swarming assays. [3b, 120] In many species, swarming relieso nt he control of surfacet ension and wetness by the secretiono fs urfactants. Modulating the secretion of surfactants is am echanism that can stall swarming colonies and this mechanism has been reported for the swarming inhibitory activity of various fatty acids. For example, the branched-chain fatty acid 12methyltetradecanoic acid selectively and completely inhibited swarming motility of P. aeruginosa PAO1 at ac oncentration of 41 mm withouta ffecting growth. [121] The effect could be assigned to ag eneral repression of secreted surfactants which also included surface-activep recursors of rhamnolipids. [122] Surfactantp roduction of P. aeruginosa has been also blockedb y the supplementation of swarming plates with halogenated alkanoic acids. These compounds directly inhibit the biosynthesis of polyhydroxyalkanoic acid (PHA) and rhamnolipids throughi nhibition of the enzymes PhaG and RhlA, respectively, and thus block surfactant-mediated swarming motility.2 -Bromohexanoic acid was hereby found to be the most potent congener inhibiting swarming at 2mm. [123] The swarming inhibition by fatty acids can be further attributed to the modulation of regulatory systems associatedw ith swarming motility. The saturated fatty acids dodecanoic and tetradecanoic acid completely blocked swarming motilityo fa clinically isolated S. marcescens strain at 0.01 %( wt/vol) supplemented to swarming plates. The effect, which turned out to be dose-dependent, resulted mainly in ad elay in the swarming lag time. Swarming inhibition was hereby associated with the saturated fatty acid-regulated two-component regulatory system RssAB. [124] Another non-QS-regulatedm echanism was found to be responsible for the swarming inhibition of S. marcescens by petroselinic acid (cis-6-octadecenoic acid) at 0.7 mm which was associated with a0 .8-fold downregulation of the swarming motility master regulator genes flhDC. [125]

Amphiphilic Compounds
In addition to fatty acids, also others urface-active substances are known to inhibit swarming. The swarming inhibiting effect against P. mirabilis in the case of homologous sodium alkylsulfates increased with chain length from hexyl-(20-30 mm)t o tetradecyl sulfate (0.1-0.5 mm)w ithout impaired growth. [126] At 0.5 mm,s odium tetradecyl sulfate completely inhibited swarming of P. mirabilis and impaired swarming already at 0.1 mm supposedly either by inhibition of formation of flagella or lysis of existing flagella. [117c] The effect of 58 chemical substances including detergents and surfactants was tested against Bacillus swarming. [127] Sodium dodecyl sulfate and bile salts such as sodiumt aurocholate and sodium desoxycholate strongly inhibited or completely blocked swarming of different strains of B. subtilis, Bacillusa lvei, Bacillus coagulans,a nd Bacillus circulans in the lower millimolar range, whereas polysorbates( Tween 20-80) even promoted swarming. [127] Bile salts also inhibited swarming of enterobacteria such as P. mirabilis. [128] Rhamnolipids of P. aeruginosa are ac lass of native surfactants with dual roles in reducing surface tension andm odulating tendril formation. Although a rhlA mutantd eficienti nb iosynthesis of all rhamnolipidsa sw ell as their b-d-(b-d-hydroxyalkanoyloxy)alkanoic acid (HAA) precursor is unable to swarm, the rhlB and rhlC mutantse xhibit altered, irregular tendril patterns (Figure 5a). [129] Purified rhamnolipids even can inhibits warming of wild-type P. aeruginosa,d emonstrating their important roles in spatialm odulation of motilityi ns warming colonies. [129] Al ibrary of syntheticf arnesyl-modified disaccharides mimicking rhamnolipidso fP. aeruginosa PAO1 was explored for effects on swarming motilityand quorum sensing. [130] Many of these compounds promoted swarming at low concentrations and inhibited swarming at higher concentrations. While the farnesylated disaccharides SFbM( 33)a nd SFbC( 34)c ompletely inhibited swarming of wild type P. aeruginosa PAO1 already at 20 and 25 mm,r espectively,t he closely related compound DbC ( 35) with ad odecyl chain rescued a rhlA mutant at 20 mm and did not inhibit swarmingo fw ild type PAO1 up to 85 mm (Figure 5b). This indicates that also the lipid component has major impact for controlling motility.As ulfate functionalized saturated farnesol (36)e ven inhibiteds warming completely between 5a nd 10 mm (Figure 5b). It was proposed that different saccharide or lipid-binding receptorsi nt he outer membrane may have been responsible for these activities. [130] Similar to some fatty acids discussed before, these rhamnolipid mimetics may thus act on regulatory level.
The endosymbiont Burkholderia gladioli of the beetle Lagria villosa produces the antibiotic lipocyclopeptide icosalide which is an interesting example for the intraspeciesr egulation of swarming by amphiphilic compounds.Althoughlinear lipopeptides of B. gladioli promoted swarming, icosalidei nhibited swarming motilityi ndicating that their interplay may regulate host colonization and free-living lifestyles. [131]

Interference with Flagellar Motor Assembly and Function
Each bacterial flagellum consistso falong helical protein filament which connects through ah ook to the basal body in the cell envelope. Rotation of the motor complex in the membrane is powered by the transport of protons or sodium ions across the membrane. The rotor is surrounded by ar ing of membrane-anchored statorc omplexes that comprise the corresponding ion channels and their interactions with the rotor generatet he torque for the rotationo ft he flagellum ( Figure 6). Mostb acterial species possess multiple stator systems which can engage in highly dynamic rotor-stator interactions tuning the flagellar motor. [132] The incorporation and exchange of stators in the motor complex depends on diverse environmental factors like the level of viscous drag or sodiumion concentration but is also regulated by the intracellular second messenger cyclic diguanylate (c-di-GMP). [133] In P. aeruginosa,m otility is mediated by one rotor with two sets of stators, MotAB and MotCD. Although MotCD is required for swarming, the MotAB statorr epresses swarming motility. Under high c-di-GMP concentrations stator selection is in favor of MotAB and thereby c-di-GMP inhibits swarming. [134] Also in other species elevated c-di-GMPl evelsl ead to inhibition of motility. [135] Intracellular c-di-GMP levelsare controlled by multi-ple diguanylate cyclases (DGCs) which produce c-di-GMP from two molecules of GTP and phosphodiesterases (PDEs) that hydrolyze c-di-GMP ( Figure 6). Different DGCs and PDEs may hereby control c-di-GMP on local and globals cale in the cell and integrated iverse signals and stimuli. [136] In ap ositive feedback regulation, disengaged MotCD stators further stimulate DGC activity,t hereby block motilitya nd support biofilm formation. [137] Inhibitors of DGCs and PDEs can be designed to modulate c-di-GMPl evels.Z heng et al. reported ab enzoisothiazolinone derivative( 37)w hich was found by in silico screening against the structure of an E. coli PDE. [138] This compound inhibited selectively c-di-GMP hydrolysis of the locally acting PDE RocR of P. aeruginosa with a K i of 83 mm,b ut did not inhibit three other PDEs of P. aeruginosa whereby global cellular c-di-GMP levelsr emained unaffected ( Figure 6). Inhibition of RocR at 100 mm completelys uppressed swarmingb ut did not increase biofilm production. [138] Another strategy to interfere with swarming motility involvesd irect blocking of the corresponding flagellar motor. Phenamil (38)a nd amiloride (39)a re inhibitors of Na + + -driven motors and have been used to dissect motor functions in different bacterial modelssuch as Vibrio and Bacillus ( Figure 6). [139] Both compounds are pyrazine derivatives that block the Na + +channels of the stator complexes and thus prevent generating torque for flagellar rotation. [139a] High-throughput screening for swarming inhibitors of V. cholerae resultedi na2,4-diamino quinazoline (40)a nd derivatives which inhibited swarming with IC 50 values in the single-digit micromolar range ( Figure 6). These compounds blocked Na + + -driven flagellar motors of different Vibrio speciesb ut had no effect on the proton-driven flagellar motors of E. coli and the lateralf lagellao fV. parahaemolyticus. [140]

Phages Modulating Motility
Flagellar function can also be impaired by certain bacteriophages. Phages can infect bacteria either by the direct exploitation of their host resulting in phage replicationa nd host-cell lysis (lytic) or by integrating into the bacterial genomea nd being replicated along with bacterial-cell division (lysogenic). Althoughalysogenic infection as such typically has no effect on bacterial motility, P. aeruginosa PA14 lysogenized with the bacteriophage DMS3 was unable to swarm and form biofilms. This inhibition depended on CRISPRs as wella sf ive of the six cas genes of the host that, when deleted,r estored the swarming and biofilm-forming phenotype. [141] Flagellotrophic phages physically attach to their host's flagellaa nd have been found to infect only motile cells. [142] Yet, effectso nm otility of the host bacteria have been rarely reported. The flagellotrophic phage c 7 has ab road host range of various specieso fb acteria. By contact with P. mirabilis,t his phage rendered its host immediately nonmotile and swarming of more than 85 %o fc linical Proteus isolates was inhibited without killing of the bacteria. [143] Thus, specific bacteriophages are able to impairs warming possibly on regulatory level or by direct physical interactions. So far,t he detailed mechanismso fh ow phages interfere on regu-

Interspecies Competition and the Microbiota
Competitivec hemical interactions of bacteria play an important role in multi-species communities in many different environments.T hus, many species may have evolved small molecules to modulate population behaviors of their competitors to their own benefit. This includes interference with swarming motility. For example, the marine bacterium Marinobacter litoralis inhibited swarming of P. aeruginosa by its lipopolysaccharide (LPS) whereas LPS from other speciesd id not affect motility. [144] In another study,t he methanol extracts of 72 Actinomycetes isolated from marine invertebrates were screened fora ctivity against P. aeruginosa. Extracts of two strainsi nhibited at 0.1 mg mL À1 swarming of P. aeruginosa by 90 and 85 %, the major active component of which was cinnamic acid. [145] In addition to smallm olecules, proteinsa lso may contribute to competitive interactions. This was observed for the soil bacterium and human pathogen Burkholderia pseudomallei that secreted aprotein factor to inhibitswarming of Burkholderia thailandensis by damaging or processing of its flagella. [146] Also the competition for resources can influence bacterial motility. Essentialt race elements such as ferric iron are highly embattled in the microbial world andb acteria competef or ferric iron by deployings iderophores as high-affinity iron chelators. Availability of ferric iron also controlss warming behavior of V. parahaemolyticus and V. alginolyticus. [147] Although in V. parahaemolyticus iron limitation is essential for swarmer-cell differentiation, [147a] V. alginolyticus requires bioavailability of ferric iron for swarming. To sequester ferric iron from the environment, V. alginolyticus encodes many different iron-siderophore receptors in its genome that allow the bacterium to engage in piracy of siderophores produced by other species. Astrain of Shewanella algae whichwas co-isolated with V. alginolyticus from the same seaweed sample evaded this siderophore piracy by producing avaroferrin (41)( Figure7a)-a chimera of the homodimeric macrocyclic hydroxamate siderophores putrebactin and bisucaberin. [148] In ad isc-diffusion assay on agar,a varoferrin (50 nmol) led to the formation of az one with inhibited swarming motility of V. alginolyticus whereas the homodimerics iderophores were considerably lessactive. [149] Other siderophores were inactive (> 500 nmol), whereas deferasirox,a na rtificially optimized iron chelator for which no receptor in V. alginolyticus is available wasapotent swarming inhibitor like avaroferrin. These results suggested that evasion of siderophore piracy by the chimeric siderophore of S. algae limited ferric iron uptake and thereby stalled swarming of V. alginolyticus. [149] This mechanism was confirmed by exploiting the promiscuity of the central NRPS-independents iderophore (NIS) synthetases giving access to non-naturalr ing-size engineered siderophores, which inhibited swarming of V. alginolyticus with potencyc omparable to avaroferrin. [150] In contrast, S. marcescens swarms only under limitation of ferric iron which is sensedb yatwo-component system through the endogenously produced iron chelator 2-isocyano-6,7-dihydroxycoumarin (42)( ICDH-Coumarin) (Figure 7b). [151] Competitive interactions within the microbiota of higher organisms may shape health andd isease of their eukaryotic host. [152] Particularly interesting hereby are the abilities of commensal and probiotic microbes to protect their hosts from pathogens. For example, the swarming pathogen S. marcescens causest he white-pox diseasei nc orals by colonizing and penetrating the coral'sm ucus layer.C ommensal bacteria were isolated from the coral Acropora palmata and investigated for their ability to competew ith S. marcescens. In co-culturing experiments, strains of Photobacterium damselae, Photobacterium leiognathi and Vibrio harveyi induced ac lear swarming inhibition zone of S. marcescens,t he active compounds, however, have not yet been identified. [153] Interactions betweenm icrobial species can also be found within the humanm icrobiota. For instance, culture supernatants of probiotic Lactobacillus acidophilus and Lactobacillus plantarum werea ctive against the swarming motilityo fS. marcescens andc ompletely inhibited swarming at 2% (v/v). [154] Lactic acid produced by ap robiotic Pediococcus strain inhibited at sub-MICc oncentrations the productiono fs hort-chain AHLs as well as swarming ands wimming motility of clinicali solates of P. aeruginosa. However, there was no evidence that short-chain AHL inhibition was causalf or inhibiting motility. [155] Various microorganisms share the ability to oxidize bicyclic aromatic compounds like naphthalenea nd indole. The oxidation products 1-naphtol as well as different hydroxyindoles completely blocked swarming motility of P. aeruginosa at 50 mm.T he activity wasf ound not to be related to changes in c-di-GMP levels or rhamnolipid pro- ductiona nd was restricted to inhibition of swarming but not swimming motility. [156] Humanp athogens also may compete with each other,w hichh as been for example reported by the ability of hemolytic E. coli but not P. aeruginosa or Acinetobacter baumannii to completely block swarming of P. mirabilis. [157] In addition to microbe-microbe interactions, swarming motility can be influenced by metabolites of the human host. This has been demonstrated for urea which inhibited at around 0.5-1 % swarming of the urinary tract-infective human pathogen P. mirabilis. [158] Humanu rine contains approximately 1.5 %o fu rea (250 mm)a nd may thusr epresent af irst line of defense against colonization by this pathogen. [159]

11.S ummary and Outlook
An enormous diversity of approaches has been reported that allows to control the swarming behavior of different bacterial species. Swarming inhibitors cover the wide range from simple fatty acids over structurally complex secondary plant metabolites to enzymesi ntercepting bacterial signals and phages that block flagellar motility.E qually diverse are the mechanisms involved in inhibition of swarming and inhibitors have already contributedl argely to our understanding of flagellar function and the differentl evels of regulatory control.M any swarming inhibitorsh ave been demonstrated or proposed to interfere on regulatory levels. However,m echanism-based inhibitors targeting signal productionw ith ac ovalent mode of action such as halogenated furanones represent only am arginal group. The majority of compoundss eems to interfere with signal receptors andt ranscriptionf actorsc ontrolling gene expression. Although indirecte ffects through quorum sensing cannot always be ruled out,a tl east severalc ompounds appear to directly interfere with flagellar gene expression or the flagellar master regulator.I na ddition to the regulation of flagellar genes, also inhibition of surfactant production is in some cases responsible for blocking motility.O ther compounds even may directly impair flagellar integrity or interferew ith motor function.
Microbe-microbei nteractions may still hold great potential for the discovery of novel swarmingi nhibitors.A lthough potent effects of extracts have been already reported, the active compounds have largely remained uncharacterized. Especially interactions within the human microbiota between commensala nd pathogenic microbes mayl ead to swarming inhibitors that could help to dissect the roles of swarming for health and disease of the human host. Understanding the corresponding chemistry and mechanismsc ould also allow to exploit microbialc ompetition for the customized control of microbial populations and interactions. Currently,i nv ivo application presents am ajor challenge which may require new generations of swarming inhibitors. So far potent anti-swarming activity has been rare. Particularly effective were antibiotics at sublethal concentrations and selected surfactants that inhibited swarming in the lower micromolar range. However,m any swarming inhibitors were of rather low efficacy and only partially reduced motilityo ro nly blocked swarming at substantially high concentrationso fs everalh undred micromolar or even millimolar.A lthough we tried to focus on compounds that genuinelyb lock swarming and do not simply reduce motility as as ide effect of growth inhibition, it is generally challenging to distinguish both effects. Especially when compounds are cytotoxic at higher concentrations, growth inhibition must be carefully evaluated. Also swarming inhibition of ac ompound was frequently overcome at longer incubation times. This limited number of highly active inhibitors may be explained by the altered physiological state of swarmer cells and cell-density effects whicha lso causei ncreased antibiotic tolerance of swarming bacteria. These challenges will have to be overcome for the development of customized high-efficacy swarming inhibitors to allow in vivo applications in animal models and finally also in humans. Blocking swarming motilitym ay exhibit future potentialf or use in combination therapies by decreasing virulence and host colonization while increasing antibiotic susceptibility of bacterial pathogens.