Reference: Biol. Bull. 189: 347-355. (December. 1995) A Transient Exposure to Symbiosis-Competent Bacteria Induces Light Organ Morphogenesis in the Host Squid JUDITH A. DOING AND MARGARET J. McFALL-NGAI* Department ofBiological Sciences, University ofSouthern California, LosAngeles, California 90089-0371 Abstract. Recent studies of the symbiotic association reviews see Schwemmler. 1989; Hirsch, 1992; Saffo, between the Hawaiian sepiolid squid Euprymna scolopes 1992). In some cases, bacterial symbioses may even be and the luminous bacterium Vibriofischeri have shown required for normal host development or survival. For that colonization ofjuvenile squid with symbiosis-com- example, enteric bacteria provideessential enzymes and petentbacteria inducesmorphogeneticchangesofthelight vitamins to their mammalian hosts, and associations organ. These changes occur over a 4-day period and in- with bacteria are required for normal development of clude cell death and tissue regression ofthe external cil- the mammalian immune system (Gordon and Pesti. iatedepithelium. In theabsenceofbacterialcolonization, 1971). In other cases, though essential only under morphogenesisdoesnotoccur. Todetermine whetherthe nutrient-poorconditions, the association is highly ben- bacteria must be present throughout the morphogenetic eficial to the host's fitness in its natural environment, process, we used the antibiotic chloramphenicol to clear such as the symbioses between leguminous plants with the light organ of bacteria at various times during the nitrogen-fixingbacteria orbetween theweevil Silophi/iis We initial colonization. provide evidence in this study oryiae and its associated gram-negative bacteria. In thatatransient, 12-hourexposuretosymbiosis-competent these partnerships, the bacteria provide nutritional bacteriaisnecessaryand sufficientto inducetissue regres- metabolites to their host (Nardon and Grenier, 1991) sion in thelight organ overthe next several days. Further, as well as influence its development. we show that successful entrance into the light organ is Ofthe known prokaryote-eukaryote associations, much necessary to induce morphogenesis, suggesting that in- progress has been made toward the understanding ofthe duction results from bacterial interaction with internal development ofplant-bacterial symbioses, both because crypt cells and not with the external ciliated epithelium. the plant hostsare easily maintained and manipulated in Finally, nodifference indevelopment wasobserved when the laboratory and the bacterial symbiontsareculturable. the light organ was colonized by a mutant strain of I', An animal-bacterial association offering similar experi- fischerithat did not produce autoinducer, a potential light mental benefits is the highly specific association between organ morphogen. the Hawaiian sepiolid squid Euprymna scolopes and the bioluminescent bacterium I'ibriofischeri. Thissymbiosis Introduction provides an experimental system to study the effect of Prolonged associations with bacterial symbionts are bacterial symbiontson hostanimal development (McFall- now recognized as important phenomena in the devel- Ngai and Ruby, 1991; Ruby and McFall-Ngai, 1992). In opmental program ofmany plant and animal hosts(for the host squid, the bacteria are always contained within epithelia-lined crypts inside the light organ, which is housed within the mantle cavity. However, the mor- R*eCcoerirveesdpo6nFdeibnrguaaurtyho1r9.95;accepted 10October 1995. phology ofthe light organ in juvenile squid is much dif- Abbreviations:CEA, ciliated epithelial appendages: CSW. California ferent from that ofthe adult (McFall-Ngai and Montgo- seawater;Cm,chloramphenicol. mery, 1990). and the light organ undergoes complex de- 347 348 J. A. DOING AND M. J. McFALL-NGAI velopmental changes following bacterial colonization (Montgomery and McFall-Ngai, 1994). Upon hatching,juvenilesquid areaposymbiotic(with- outbacterial symbionts) and normally acquire free-living V. fischeri from the surrounding seawater within hours (Wei and Young, 1989; McFall-Ngai and Ruby, 1991). Asubstantialportionofthejuvenilelightorganepithelium is microvillous and ciliated, bearing two lateral pairs of appendages(ciliatedepithelial appendages; CEA)thatap- peartofacilitateinoculation ofbacteriaintothelightorgan (Fig. la; McFall-Ngai and Ruby, 1991; Montgomery and McFall-Ngai, 1993). Microscopy and high-speed cine- matography have revealed that the two appendages on each side of the light organ form a ring, at the base of which are three pores leading into three independent crypts (Fig. Ib). Beating ofthe cilia entrains passing sea- waterwithinthering, potentiallyincreasingtheprobability B that symbionts within the water will enter the pores (M. McFall-Ngai and R. Emlet, unpub. results). When the lceilglhtdoeragtahnishaosbsbeerevnedsuicncetshsefuClElyAcoalnodnirzeegdrebsysio}'n. foifscthheersi.e woiftFthihgecuoarmepppIle.entdeaSgcCeaEsnnAiinnagvn/devloeacatsnrcdohnetmhmeaitctirhcorgederraappowhrie(nsAg)i(noBf)tahsehhacotiwcliihanltigendgthelepiigphttohsoeirltgiiauonmn appendages occurs over a period offour days. Four-day- (arrow) with their associated internal crypts (approximated by dashed old squid that are not infected with I', fischeri do not line). A, anterior; P, posterior; h. hindgut. Scalebar = 100^m. show cell death nor regression ofthe CEA (Montgomery and McFall-Ngai, 1994). Therefore, the presence ofsym- biosis-competent bacteria somehow induces host tissues might beassociated with theeggclutch, squid weretrans- that are several cell layers away to initiate light organ ferred immediately upon hatching through three rinses morphogenesis. Cell death and the resulting regression of with California coastal seawater (CSW), which does not theCEAarethefirstobservableeventsoflightorgan mor- contain infective strains of r.fischeri (McFall-Ngai and phogenesis and therefore the first developmental evidence Ruby, 1991). Juveniles were used for infection studies that induction has occurred. within 6 h ofhatching. In this study we have asked whether the presence of bacteria within the light organ is required continuously Inoculation ofsquid with V. fischeri bacteria for4 daysto induceCEA regression. Additionally, weused Bacteria were grown to log phase in a seawater-based noninfective strains of I", fischeri to determine whether minimal medium (Ruby and Asato, 1993) and diluted to colonization ofthe light organ is necessary for induction. between 103 and 10s cells/ml for inoculation of squid. Finally, we tested whether I', fischeri autoinducer, a cell After inoculation, squid were rinsed in CSW and trans- density-dependent factor secreted by the bacteria and in- CSW CSW ferred to either or chloramphenicol-treated volved in the production of light, is required to induce (as described below). For all experiments, positive (in- light organ morphogenesis. fected) controls were exposed to symbiosis-competent bacteria in CSW for the entire 4 days and negative (un- Materials and Methods CSW infected) controls were exposed to alone. Animalcareand maintenance Monitoring bacterial colonization Adultsquid werecollectedatnight from Kaneohe Bay, Oahu, HI, with dipnets and were transported back to the Because I". fischeriis luminous in the light organ, suc- UniversityofSouthern California, LosAngeleswithin one cessful colonization of the organ can be monitored by week ofcollection. Animals were maintained in a 265- measuring the bioluminescence ofthe squid with a pho- liter recirculating aquarium at 23C, and females were tomultiplier tube (Luminescence Photometer, Model matedapproximatelyonceaweek. Eggclutches, attached 3600. Biospherical Instruments. Inc.). For these mea- tocoral rocksorother hard surfaces by the females, were surements, individual squid were kept in 5 mlofseawater transferred for hatchingto smaller temperature-controlled in glass scintillation vials. Seawater in the vials was 23C aquaria. To ensure thatjuvenile squid did not be- changeddailythroughoutthe4-dayexperiments,justprior come prematurely infected with any residual bacteria that to each luminescence measurement. SYMBIOSIS-SPECIFIC MORPHOGENESIS 349 2d 3d 4d 4dFaiygsur(de).2.SymTbiiomteicsearniiesmaolfsCwEeAreriengorceuslsaitoendinwistyhmb~io1t0i4cce(lslysm/)mlaonfdIa'.pfoisscyhmehriioEtSc1(14apwoi)tjhuivne6nihleofsqhuaitdchoivneg.r Thesym-4d panel representsa fully regressedCEA. Onlythe right halfofeach light organ isshown. Scale bar = 100fim. Differences in sizereflect individual variation. Determination ofCiliatedEpithelialAppendage (CEA) CEA regression ofsymbiotic juveniles was compared regression to that ofaposymbioticjuveniles for 4 days following in- Regression ofthe CEA was determined at the end of ovcaruilaabtlieonexwiptohsubraectteirmieas(,FiCg.E2A).rFeogrreesxspieonriamtednatsyi4nvwoalsvidnig- 4 days for each experiment. Squid were fixed for 24 h in vided into five stages(0, 1, 2, 3, and 4), which correspond seawater containing either 5% formalin or 2% parafor- to the regression seen at (Fig. 1A), 1, 2, 3, and 4 days maldehyde/2% glutaraldehyde. Samples were subse- ofuninterrupted symbiosis (Fig. 2). CEA oflight organs quently rinsed twice forM30 min in 50 mA/ sodium phos- were considered regressed ifthey were at or beyond stage phate buffer with 0.45 NaCl (pH 7.2), followed by a 3. Due to the high variability between individual squid, dehydration serieswith 15%-100% ethanol. Sampleswere even within control groups, regression is reported as an critical-point dried with liquid CO2, or desiccated with E/C index, which is defined as the percentage ofexperi- hexamethyldisilazane (Polysciences, Inc.). Dried squid mental animals with regressed CEA divided by that of were mounted onto aluminum stubsand the ventral por- thesymbioticcontrols in agiven experiment(experimen- tions of the mantle and siphon were dissected away to tal andcontrolanimalsforagiven experimentwerealways reveal thejuvenile light organs. These samples were then from the same clutch ofeggs). sputter coated with gold and the light organs were ex- amined with a Cambridge 360 scanning electron micro- Manipulation ofexposure times andcolonization levels scope(SEM). Thepresence orabsenceofCEAwasscored and recorded, and photomicrographs were taken ofrep- Transient vs. continuous exposure to V. fischeri strain resentative samples. ESI14. To determine firstwhether acontinuousexposure 350 J. A. DOING AND M. J. McFALL-NGAI to competent bacteria is necessary for initiating morpho- fischeri(MJ1 1),which isnotnormallyassociated withthe genesis ofthe light organ, we used the bacteriostatic an- E. scolopeslight organ but iscapable ofcolonization, was tibiotic chloramphenicol (Cm) to clear the light organ of also tested for its ability to induce morphogenesis. Fol- viable symbionts (Fig. 3a). Squid were exposed to one of lowing the inoculation period, squid were transferred to twosymbiosis-competent strainsof I'ibriofischeri:ESI 14, CSW for the remainder ofthe 4 days. Possible coloniza- achloramphenicol-sensitive lightorgan isolate(Boettcher tion ofsquid exposed to noninfective strains was deter- and Ruby, 1990) or ESI 14-U2, a spontaneous chloram- minedbybothluminescence measurementsandbacterial phenicol-resistant mutant ofESI 14 (donated by J. Graf)- plate counts. Colonization ofpositive and negative con- The inoculations were performed for two time periods: trols was determined by luminescence only. (1)continuously for4 days, or(2) for 12 h. Those exposed Noninfective strains. Strain M101 was produced by continuously were inoculated with I", fischeri and then transposon (Mu dl-1681) mutagenesis ofsymbiosis-com- transferred to CSW after 12 h for the remainder of the petent strain ESI 14, resulting in a nonmotile mutant. 4 days (Fig. 3a, top bar). These squid remained infected Nonmotile mutants of I'. fischeri have previously been fortheduration oftheexperiment. Followingincubation shown to be noninfective in E. scolopes(Grafetal.. 1994). with V. fischeri. squid exposed for only 12 hours were Squid were exposed to ~104 M101 cells/ml for 12 h. transferredtoCSWtreatedwith 10 ng/mlCm inseawater Strain MdR12 is a non-symbiotic wild type isolate from fortheremainderoftheexperiment (Fig. 3a, second bar). Southern California coastal seawater. Strain MJ1 was The transient time period of 12 h was chosen because originally isolated from the light organ of the Japanese successful colonization ofthe light organ by bacteria can pinecone fish Monocentrisjaponica, but has been in cul- beconfirmedbytheappearanceofluminescencebetween ture for 21 years(Ruby and Nealson, 1976) and does not 10 and 12 h after exposure. The Cm-resistant strain infect E. scolopes. Squid were exposed to ~105 cells/ml ESI 14-U2 was used as a control for any inhibitory phar- ofthis strain for 24 h. macologicaleffectsthatCm mayhaveonCEA regression. Infectivestrain. Strain MJ1 1 wasisolated fromthelight Squid were monitored for luminescence before exposure organ ofM. japonica in 199~2 and is infective to E. scol- to bacteria, every 2 h during initial colonization andevery opes. Squid wereexposedto 1 5 MJ1 1 cells/ml for 12 h. 12 hthereafter. Uninfected controlswereexposedtonon- This strain was ofinterest because, although it is capable infective CSW with or without Cm (Fig. 3a, third and ofcolonization, bacterial numbers inside the light organ fourth bars) and monitored for luminescence every 12 h. reach only 10% ofthe levels seen with ESI 14 (K.H. Lee To insure that Cm treatment was effectively clearing and E.G. Ruby, pers. comm.). the light organ of viable bacteria, the decrease of both bacteria colony forming units (CPU) and luminescence Exposureto an aittoinducer mutant strain of V. fischeri was monitored in squid treated with 10 ng Cm/ml CSW following exposure to bacteria for 12 h. Symbiont bioluminescence in the E. scolopes light or- VariableexposuretimestoV. fischeristrainESI14. To gan is induced via a well studied mechanism involving determine the minimum time period required to induce the secreted I', fischeri molecule autoinducer (VAI), a morphogenesis, hatchlingsquid wereexposedto I'.fischeri homoserine lactone. Normally VAI is expressed consti- for variable lengths oftime (Fig. 3b). At time 0, all squid tutively at a low level, but when cell densities become were placed in a single bowl with ofCSW containing ~5 high, such as in the light organ (Boettcher and Ruby, X 103 ESI 14 cells/ml. Groups of 10-20 animals were re- 1990), the build up of VAI in the extracellular medium moved from the bowl at 1,4, 8, and 12 h. Upon removal positively regulates VAI gene expression and in turn ac- from the bowl at each time period, halfofthe squid were tivates expression of the lux operon, which encodes for CSW rinsedtwiceandtransferred to vialswith Cm-treated thosegenes responsible forbacterial light production (for (Fig. 3b,top),whiletheotherhalfweretransferredtovials review, see Dunlap and Greenberg, 1991). To determine with Cm-free CSW (Fig. 3b. bottom). Groupstransferred whether VAI was a morphogen ofthe squid light organ, to Cm-free CSW became infected within 12 h. Lumines- we used a mutant strain (3100) of I', fischeri (provided cence was measured immediately before and after expo- by Kendall Gray) containing an insertion in the autoin- sure to bacteria and once per day thereafter. ducer gene, which renders the cells incapable ofmaking autoinducer. One-day-old squid were exposed to symbiosis-compe- Exposureto other strains of V. fischeri tent V.fischeristrain 310fiortoESl 14ataconcentration To determine whether colonization ofthe light organ of ~103 cells/ml for approximately 20 h. Because the by the bacteria is necessary to induce CEA regression, 31OS2 strain is nonluminous, successful colonization of squid were exposed to three noninfective strains of I . the squid could not be monitored with a photometer. In- fischeri (M101, MdR12, and MJ1). A fourth strain of T. stead, at theend of4 days, two ofthe squid that had been SYMBIOSIS-SPECIFIC MORPHOGENESIS 351 12 96h Wll/lllllllllllllllllllllh \llllllllllllllllllllllllllh B EXPOSURETOBACTERIA Cm TREATMENT ] CSW Figure3. Experimentaldesignfortransientand variableexposuretobacteria, n = 5-20perexperiment for each treatment group in A and B. At the e~nd of4days (96 h), squid were fixed for SEM to score regression. A. Hatchling squid were exposed to 10"cells/ml ofESI14 or ESI14-U2 for 12 h, at which timethe bacteria-containingseawaterwaschanged toCSW(top bar)toallowthe infection toensue, orto Cm-treated seawater(second bar)to stopthe infection and cure the light organ. Controlswere exposed to CS~Walone(thirdbar)ortoCSWfollowedbyCm-treatedCSW(fourthbar).B.Hatchlingsquidwereexposed to 104 cells/ml ESI14 for 1. 4, 8, or 12hours (started at time0), at which time the bacteria-containing seawaterwaschangedtoCm-treated seawater(topbar)ortoCSW(lowerbar). Cm exposed to 31012 were homogenized and plated to verify i.e., ifthe treatment was lifted after4 daysand a new thatthey had been infected. Therestofthatgroup(n = 8) inoculum of V.fischeri was introduced into the seawater was scored for CEA regression. ESI 14 and negative con- the squid became luminescent within 24 h, indicatingthat trol groupswere monitored forcolonization by measuring they were still capable ofbeing infected. luminescence only. /uuuui Results Transient vs. continuous exposure to V. fischeri strain ESI14 Thenumberofviablebacteriainthelightorgandeclines sharplyafteronly2 hoursinCm-treatedCSWto487CPU (approximately 1% oftheinitial value), concomitantwith Cm a decline in luminescence (Fig. 4). After 10 hours of treatment, all ofthe squid monitored had no viable bac- teria detectable in their light organs. Additionally, ifthe Cm treatment wasremovedafter4 daysand replaced with CSW alone, the light organs ofthe squid did not become reinfected, confirmingthattherewere no viable J'.fischeri Cm in the light organ after treatment with for 4days. Cm Squid exposed to appeared as healthy as those not Cm exposed to and there was no adverse effect on the ability ofthesquidtoinfectaftera4-dayexposuretoCm; 352 J. A. DOING AND M. J. McFALL-NGAI Figure 5. Scanning electron micrographs oflight organs (right halfonly) of4-day old squid exposed upon hatchingto: (A)CSW for4days;(B) ESI14 for 12 h followedbytreatmentwithCm for3.5days;(C) ESI14-U2 (Cm resistant strain) for 12 h followed by treatment with Cm for 3.5days. See Figure 3A for experimental design. Scalebar = 100urn. Squidexposed to I'.fischerifor 12 h showed regression Variabletransient exposure to V. fischeri strain ESI14 ofthe CEA similar to that of squid exposed for 4 days C(niFnoSihgirW.beig5t(brFo)eir.sgys.Nieeof5gnfaae)cottfioovrCenECAcmCo-.EnttArArdeordaleitgtearindeoisnsCmaialSollnyWs,,ea(Csxnpeomovtsidesdidhedonfcwnoeonrdt)4bhsdyahavcoyeoswmeatd-no pttoioosEdnexedopt=foetsrCoumEbriaAencetorefterhgsieraqeufmsiosdirinotanotimblraeeuacvsmteteartl1iei2admhfetohsrrahet1o,qowu4n,ielrd8ye,Cmt1ehE2no,Atsoerfrose1rgq4ruieihnsdodsuuiercox-sn- rpelseitsetraentgrsetsrsaiionnEoSfIC1E4A-Uf2roamndsqturiedatiendfewcittehdCwimth(tFihge.C5mc)-. c(loEun/mtCirnoolu0ss.q9ua6ti)dtch(oeFmitpgi.amr6ea:boTlfaeCbtlmoettIh)ra.etaTothmfeecsnoetn.tsiqSnuquiuodiudswleeyrxepeoxvspieosdisbetldoy bacteria for 1 or 4 h were not luminous and showed no CEA regression. Those individuals exposed for 8 h were only occasionally luminous and the E/C ratio was only 0.24. Although someanimalsdiedduringtheexperiments, the death rate was not greater than that normally seen in animals 4 days post-hatching (averaging less than 10%) and the incidence ofdeath appeared random with respect toexperimental groups. Animals that died were not used in thecalculation ofCEA regression percentages. Thelevel ofinfection, measured by platinglightorgan homogenates afterexposure to bacteria, wassignificantly higherat 12 h Table I Pooleddatafromfourseparateexperimentsasdescribedin Figure3b. E/Crepresentsratioofo E\P(Experimental) to % CONT(Control) hrs exposed Figure 6. Relative percentage of light organs in 4-day-old squid showingstage 3 regression oftheCEA (see text). Hatchling squid were exposed to ESI 14 for 1. 4. 8,and 12 hours, followed bytreatmentwith Cm (see Fig. 3B-for experimental design). Valuesare reported as E/C. the percentage ofCm-treated squid (experimental) showing regression dividedby thepercentageofinfectedcontrolsshowingregression. Each point representsvaluespooled from fourseparateexperiments(seeTable I).Verticalbarsrepresentthefullrangeofdataforthefourexperiments. SYMBIOSIS-SPECIFIC MORPHOGENESIS 353 than at 8 h (Fig. 7). Although CEA regression was higher Table 11 at 12 h than at 14 h (Fig. 6), the values at 14 h are within Observedcapabilityofvariousstrains <>/Vibriofischeri toinfect the range ofvalues for 12 h. These results suggest that the Euprymnascolopesandinducelightorgan morphogenesis, n = 5 minimum exposure time for complete regression ofthe squidforeach strainandalllivesquidineachgroupshowed CEA must lie between 8 and 12 h. thesameresults Exposureto otherstrains of V. fischeri When squid were exposed to the nonmotile strain of V.fischeri. M101, neither colonization nor CEA regres- sion was observed, supporting the above evidence that the bacteria must be within the light organ to induce morphogenesis. Additionally, of the natural isolates tested, only the infective strain, MJ11, induced CEA regression (Table II). Exposureto an aiiloinducer mutant of V. fischeri Squid exposed to theautoinducer mutant of I'.fischeri, 31012, were infected and showed complete regression of CEAafter4 days(TableII),thuseliminatingthepossibility that autoinducer is required for light organ morpho- genesis. Discussion The results of this study show that light organ mor- phogenesis ofthe squid Euprymna sco/opes in response to the presence ofsymbiotic bacteria (1) requires only a 12-h exposure to symbiosis-competent bacteria; (2) re- quirescolonization ofthe lightorgan bybacteria; (3)does not require I',fischeri autoinducer. The finding that a transient exposure to symbiosis- competent bacteria is sufficient to induce morphogenesis ofthe squid light organ (i.e.. the bacteria are not required throughoutthe4-day morphogeneticprocess)suggeststhat 50000- 354 J. A. DOING AND M. J. McFALL-NGAI and morphological changes in the plant without coloni- teract with animal cells (Lin el ai, 1993) and may be zation ofthe host by the symbiont (Long, 1989; Appel- potential morphogens. For example, cholera toxin has baum. 1990; Hirsch, 1992), morphogenetic induction in been shown experimentally to induce metamorphosis in the I'ihrio-Euprymnasymbiosis requiresthatthe bacteria certain marine larvae (Hofmann and Brand, 1987). In be within the confined space ofthe light organ. Studies addition. Reich and Schoolnik recently found that V.fis- using strain MJ1 1, which produces a colonization con- cheri carries a gene homologous to toxR (1994), which sisting ofonly about 10% ofthe typical cell number yet regulates cholera toxin production in I', cholerae, and induces morphogenesis, indicate that the actual bacterial also synthesizes a cholera toxin-like ADP-ribosylating volume is probably not exerting a physical pressure or protein (1995). However, while commercially available stretching ofthe light organ to induce morphogenesis. It choleratoxin mimicssomeaspectsofthesymbioticstate, is unclearat thistime whether the signal is secreted from by itself it does not cause morphogenesis in E. scolopes thebacteria intothe light organ crypt lumen orthe signal (Small and McFall-Ngai, 1993), suggesting that ifan en- ispresenteddirectlyon thesurfaceofthebacteria. Ineither dotoxin-like molecule is the squid morphogen, it is sig- case,theobservedtimewindowofbetween 8 and 12 hours nificantly different from cholera toxin, or that additional necessary to induce CEA regression, may reflectthe need molecules(perhapson thesurfaceofthebacteria)arealso for an accumulation of bacteria, or their products, to a required. Other bacterial factors that have been demon- critical density within the light organ. Ifthe morphogenic strated to affect metamorphosis ormorphogenesis in var- signal is secreted, there are a few possible scenarios: (1) ious host organisms include oligopeptides(Hofmann and secretion ofthe morphogen is induced and only occurs Brand. 1987), phorbol esters (Mullen 1985). diacylglyc- within the environment ofthe light organ, (2) the mor- erols (Leitz and Muller, 1987), and lipo-oligosaccharides phogen is produced constitutively and light organ crypts (Lerouge el a/., 1990; van Brussel el al, 1992), any of provide a barrier to diffusion of bacterial products such which may prove important in our system. that within the light organ the signal reaches a critical In conclusion, we have shown that a transient coloni- concentration required for transduction, or (3) the pres- zation off. scolopeswith symbiosis-competent V. fischeri ence ofbacteria within the light organ (perhaps through induces morphogenesis ofthe squid light organ. Trans- direct cell-cell contact) renders the host cells competent duction ofthe morphogenic signal requires the presence to "accept" the secreted signal from the bacteria. ofthe bacteria within the light organ for approximately Rather than secreted, the bacterial signal may be a 12 h. Further investigations are necessary to determine molecule presented on the bacterial cell surface that in- the nature ofthe bacterial signal, the role ofcolonization teracts directly with a receptor on the animal cell mem- in thegeneration ofthesignal, andthetransduction path- brane. Direct interactions viaglycan-adhesin bindinghave way within the host squid. been implicated in many symbiosesand there isevidence for a mannose lectin in E. scolopes: when squid are in- Acknowledgments oculated with bacteria in the presence of mannose. col- onization is significantly inhibited (V. Weis, K. Brennan We thank Ned Ruby, Mary Montgomery, and Alicia and M. McFall-Ngai. unpub. data). In the Rhizobium- Thompson for technical advice and Angel Lemus for legume symbiosis, plant lectins that recognize specific graphicsassistance. WethankJorgGrafand Kendall Gray bacterial surface oligosaccharides have been suggested to fortheirgenerousdonationsofmutant I'.fischeristrains. playa majorrole in attachment and invasion mechanisms We also thank Andrea Small, Katie Brennan. Jamie Fos- (Dazzo and Truchet, 1983). Further, in pathogenic as- ter. Wes Toller, and Karen Visick for helpful comments sociations, bacterial adhesins on pili often are involved on the manuscript. Thisis HIMB contribution #991. This in recognizing specific sugar receptors on the animal cell work was supported by NSFGrant No. IBN 9220482 (to surface (Finlay and Falkow, 1989). MM-N and EG Ruby) and ONR Grant No. N00014-91- Recently it has been shown that various other autoin- J-1357(toMM-N). ducer molecules regulate the production of exoenzyme virulence determinants in Pseudomonas aeruginosa and Literature Cited Erwinia carolovora (Jones el ai. 1993). Also, these au- toinducersare structural analogs ofactinomycetes A-fac- Appelbaum, E. 1990. The Rhizobium/Bradyrhizobium-lesame sym- tor, which hasbeen implicated asan autoregulatorofcel- biosis. Pp. 131-155inMolecularBiologyofSymbioticNitrogenFix- lular differentiation between different Streptomyces spe- ation. P. M. Gresshoff. ed. CRC Press, Boca Raton. FL. cies (Beppu, 1992). However, the results of this study Beppu, T. 1992. Secondary metabolitesaschemical signals for cellular differentiation. Gene115: 159-165. indicate that I", fischeri autoinducer is not required for Boettcher, K. J., and E. G. Ruby. 1990. Depressed light emission by light organ morphogenesis. Other secreted bacterial mol- symbiotic I'ibrio fixcheri ofthe sepiolid squid Euprymna scolopes. ecules such as Vibrio spp. endotoxins are known to in- ./. Bactenol 172: 3701-3706. SYMBIOSIS-SPECIFIC MORPHOGENESIS 355 Dazzo, K.B.,andG. I,.Truchet. 1983. Interactionsoflectinsandtheir McKall-Ngai, M.J., and Montgomery, M. K. 1990. The anatomy of saccharide receptors in the Rhizobium-\egume symbiosis. J. Mem- theadult bacterial light organ ofEitprymnascolopesBerry(Cephal- braneBiol. 73: 1-16. opoda: Sepiolidae). Biol. Bull 179: 332-339. Drake, D., and T. C. Montic. 1988. Flagella, motility and invasive McEall-Ngai, M.J.,andE.G. Ruby. 1991. Symbiontrecognitionand virulenceofPseiidomonas aeruginosa. J Gen Microbiol. 134: 43- subsequent morphogenesis as early events in an animal-bacterial 52. mutualism. Science 254: 1491-1494. Dunlap.P.Y.,andE.P.Greenberg. 1991. Roleofintercellularchemical Montgomery, M. K., and M. J. McEall-Ngai. 1993. Embryonic de- communication in the Vibriofischeri-monocentnd fish symbiosis. velopmentofthelightorganofthesepiolidsquidEitprymnascolopes Pp. 219-252 in Microbial Cell-Cell Intentions. M. Dworkin. ed. Berry.Biol. Bull. 184: 296-308. Amer. Soc. Microbiol. Washington, DC. Montgomery,M.K.,andM.J.McEall-Ngai. 1994. Bacterialsymbionts Dunn. B. E. 1993. Pathogenic mechanisms of Helicobacter priori. induce host organ morphogenesis during early postembryonic de- Gastroenlerol. Clin. N. Ai. 22:43-57. velopmentofthesquid Eitprymnascolopes. Development 120: 1719- Eaton, K. A., D. R. Morgan, and S. Krakawa. 1989. Campylobacter 1729. pylori virulence factors in gnotobiotic piglets. Inject. Immun. 57: Muller, W. A. 1985. Tumor promoting phorbol esters induce meta- 1119-1125. morphosisand multiplehead formation in thehydroidHydractinia GIoirnldp.oantn.h,oBgI.eIn.Bi.Ac,.i,tayan.nddMiScL.r.oFbPaielsoktlio..w.R1e9v17.918.593.:T2h1Ce0o-g2mn3mo0to.onbiotthiecmaensimianlmaiscraobtioaoll NardwaeeocSelnovi,iumlraP.tce,aevaoonlDfduiEtfAvfi.oeornlMe:un.tttGiihroaeentaniroorilnyeerI2o.9nf:n1os29vy91am1tb6.ii-oo2ns2.i2Sse..Lr.iPaMpl.aerng1d5uol3si-ys1m6ab9nidoisnRiS.syFmtehbseitooerrs,yiseadnasd.s inthestudyofhost microbial relationships.Baeteriol. Rev 35: 390- MIT Press, Cambridge, MA. 429. Reich, K. A. and G. K. Schoolnik. 1994. The light organ symbiont Graf. J., P. V. Dunlap, and E. G. Ruby. 1994. Effect oftransposon- ('ibriofischeripossessesahomologofthe I'ibriocholeraetransmem- induced motility mutations on colonization ofthe host light organ branetranscriptionalactivatorToxR. / Bacterial 176: 3085-3088. by Vibriofischeri. J Baeteriol. 176: 6986-6991. Reich. K. A., and G. K. Schoolnik. 1995. Purification and character- Guentzel, M. N.,and I,.J. Berry. 1975. Motilityasavirulencefactor ization of an ADP-ribosyltransferase from the non-pathogenic light for Vibriocholerae. Infect. Immun. 11: 890-897. organ symbiont. Vibrio fischeri. Am. Soc. Microbiol. Abstr. B-199: Hirsch, A. M. 1992. Tansley Review No. 40. Developmental biology 200. oflegume nodulation. NewPhytol. 122: 211-237. Ruby,E.G.,and L.M.Asato. 1993. GrowthandflagellationofVibrio Hofmann, D. K.,and II. Brand. 1987. Inductionof metamorphosisin fischeriduringinitiation of thesepiolid squid light organ symbiosis. thesymbioticscyphozoanCassiopeaandromeda:roleofmarinebac- Arch. Microbiol. 159: 160-167. teriaandofbiochemicals. Symbiosis4: 99-116. Ruby, E.G.,and M.J. McFall-Ngai. 1992. A squidthatglowsinthe Jones,S.,B.Yu,N.J.Bainton,M.Birdsall,B.W.Bycroft,S.R.Chhabra, night: development ofan animal-bacterial mutualism. / Bacterial. A.J. R. Cox, P. Golby, P. J. Reeves, S. Stephens, M. K. Winson, 174:4865-4870. G.P.C.Salmond,G.S.A.B.Stewart,andP.Williams. 1993. The Ruby, E.G.,and K. H. Nealson. 1976. Symbioticassociation ofPho- IIL\ automducer regulates the production ofexoenzyme virulence tobacterium fischeri with the marine luminescent fish Monocentris determinants in Erwinia carotovoraand Pseiidomonas aeruginosa. japonica:amodelofsymbiosisbasedonbacterialstudies.Biol. Bull EMBOJ 12:2477-2482. "l5I: 574-586. Leitz, T., and W. A. Muller. 1987. Evidence for the involvement of Saffo, M. B. 1992. Invertebrates in endosymbiotic associations.Am. oPif-mseigtnaamlolrinpghoasnidsdiinactyhleglhyycderrooildsHeycdornadctmiensisaeencgheirnsatiunFtlheemiinnigt.iaDteivo.n SchwZoeom/m.l3e2r:,5W57.-516958.9. Insect endocytobiosis as a model system for egg cell differentiation. Pp. 37-53 in Insect Endocytobiosis: .Mor- Biol. 121: 82-89. phology, Physiology. Genetics. Evolution. W. Schwemmler and G. Lerouge, P., P.Roche,C. Eaucher, F.Maillet,G.Truchet,J.C. Prome, Gassner,eds. CRCPress. Boca Raton, FL. andJ.Denarie. 1990. SymbiotichostspecificityofRhiiobiitminel- Small, A. L., and M. J. McFall-Ngai. 1993. Changes in the oxygen iloti is determined by a sulphated and acylated glucosamine oligo- environmentofasymbioticsquidlightorganinresponsetoinfection saccharidesignal. Nature344: 781-784. by itsluminousbacterialsymbionts.Am. Zoo/. 33:61A. Lin, / , K. Kumagai, K. Baba, J. J. Mekelanos, and M. Nishibuchi. van Brussel,A.A. N., R. Bakhuizen, P.C.vanSpronsen,H. P.Spaink, 1993. !ibrioparahaemolyticushasahomologofthe Vibriocholerae T.Tak, B.J.Lugtenberg,andJ.W.kijne. 1992. Inductionofpre- fo.vRSoperon that mediatesenvironmentally induced regulation of infectionthreadstructureintheleguminoushostplantbymitogenic thethermostabledirecthemolysingene./ Baeteriol 175:3844-3855. lipo-oligosaccharidesofRlu:obium. Science257: 70-72. Liu,S.,T. Ezaki, H. Miura, K. Matsui, and E. Yabuchi. 1988. Intact Wei,S.L.,andYoung,R.E.1989. Developmentofasymbioticbacterial motilityasaSalmonellatyplninvasion-related factor. Infect Immun bioluminescenceinanearshorecephalopod.Euprymnascolopes.Mar. 56: 1967-1973. Biol. 103: 541-546. Long, S. R. 1989. Rhizobium-legume nodulation: life together in the Yancey, R.J., D. L.Willis,and L.J. Berry. 1978. Roleofmotilityin underground. Cell56:203-215. experimentalcholera in adult rabbits. Infect. Immun. 22: 387-392.