Developmental and Microbiological Analysis of the ... · detailed microbiological analysis of early developmental stages has not been carried out with any wild-caught bacterially

APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Dec. 2008, p. 7471–7481 Vol. 74, No. 240099-2240/08/$08.00�0 doi:10.1128/AEM.01619-08Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Developmental and Microbiological Analysis of the Inception ofBioluminescent Symbiosis in the Marine Fish Nuchequula nuchalis

(Perciformes: Leiognathidae)�

Paul V. Dunlap,1* Kimberly M. Davis,1‡ Shinichi Tomiyama,2§ Misato Fujino,2 and Atsushi Fukui2

Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, Michigan 48109,1 and School ofMarine Science and Technology, Tokai University, Orido, Shimizu, Shizuoka 424-8610, Japan2

Received 15 July 2008/Accepted 22 October 2008

Many marine fish harbor luminous bacteria as bioluminescent symbionts. Despite the diversity, abundance,and ecological importance of these fish and their apparent dependence on luminous bacteria for survival andreproduction, little is known about developmental and microbiological events surrounding the inception oftheir symbioses. To gain insight on these issues, we examined wild-caught larvae of the leiognathid fishNuchequula nuchalis, a species that harbors Photobacterium leiognathi as its symbiont, for the presence,developmental state, and microbiological status of the fish’s internal, supraesophageal light organ. Nascentlight organs were evident in the smallest specimens obtained, flexion larvae of 6.0 to 6.5 mm in notochordlength (NL), a developmental stage at which the stomach had not yet differentiated and the nascent gasbladderhad not established an interface with the light organ. Light organs of certain of the specimens in this size rangeapparently lacked bacteria, whereas light organs of other specimens of 6.5 mm in NL and of all largerspecimens harbored large populations of bacteria, representatives of which were identified as P. leiognathi.Bacteria identified as Vibrio harveyi were also present in the light organ of one larval specimen. Light organpopulations were composed typically of two or three genetically distinct strain types of P. leiognathi, similar tothe situation in adult fish, and the same strain type was only rarely found in light organs of different larval,juvenile, or adult specimens. Light organs of larvae carried a smaller proportion of strains merodiploid for thelux-rib operon, 79 of 249 strains, than those of adults (75 of 91 strains). These results indicate that light organsof N. nuchalis flexion and postflexion larvae of 6.0 to 6.7 mm in NL are at an early stage of development andthat inception of the symbiosis apparently occurs in flexion larvae of 6.0 to 6.5 mm in NL. Ontogeny of the lightorgan therefore apparently precedes acquisition of the symbiotic bacteria. Furthermore, bacterial populationsin larval light organs near inception of the symbiosis are genetically diverse, like those of adult fish.

Members of 12 families of marine teleost fish, representingsix orders, form bioluminescent symbioses with luminous bac-teria. Many of these fish are abundant in the marine environ-ment and geographically widespread and play important rolesin marine ecosystems (14, 17, 22, 24). The fish maintain theirbacteria in gland-like tissue complexes called light organs, theanatomical diversity of which among different fish reflects theevolutionary diversity of these animals (17, 47). The fish isthought to provide its bacteria, which are extracellular andtypically a single species within each fish family, with oxygenand nutrients for luminescence and reproduction. The animaluses the bacterial light for signaling, avoiding predators, andattracting prey (e.g., see references 20, 24, 32–35, 37, and 46).Growth of the bacterial population within the light organ leadsto a release of excess cells from the light organ into the sea-water (23, 41), from which the bacteria colonize various otherhabitats (14, 24, 40). In the only experimental study, bacteria

could be passed via seawater from adults to aposymbiotic ju-veniles to reinitiate the symbiosis (51). To date, four bacterialspecies, Aliivibrio fischeri, Photobacterium kishitanii, Photobac-terium leiognathi, and “Photobacterium mandapamensis,” havebeen identified as bioluminescent symbionts of fish (17, 26, 48).

Despite the abundance and ecological importance of thesefish and their apparent dependence on luminous bacteria forsurvival and reproduction, very little is known about develop-mental and microbiological events surrounding the inceptionof their symbioses. Limited information from microscopic ex-aminations of a few wild-caught or artificially reared specimensof early developmental stages of some fish suggests the animalacquires its symbiont early in development, prior to or duringnotochord flexion, and that initiation of light organ ontogenyprecedes bacterial acquisition (6, 25, 28, 31, 38, 52). However,detailed microbiological analysis of early developmental stageshas not been carried out with any wild-caught bacterially lu-minous fish, due to the rareness with which early life historystages of these animals have been obtained. This situationleaves undefined many aspects of symbiont-host interaction atthe inception of the symbiosis.

To gain insight into these issues, we collected early devel-opmental stages of wild-caught specimens of the leiognathidfish Nuchequula nuchalis, a shallow-dwelling, coastal species,and examined them using microscopic and microbiologicalmethods. Leiognathids bear an internal, supraesophageal light

* Corresponding author. Mailing address: University of Michigan,Department of Ecology and Evolutionary Biology, 830 North Univer-sity Avenue, Ann Arbor, MI 48109-1048. Phone: (734) 615-9099. Fax:(734) 763-0544. E-mail: [email protected].

‡ Present address: Department of Microbiology, University of Penn-sylvania, Philadelphia, PA 19104.

§ Present address: Marine Science Museum, Social Education Cen-ter, Tokai University, Miho, Shimizu, Shizuoka 424-8620, Japan.

� Published ahead of print on 31 October 2008.

7471

on June 30, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

https://crossmark.crossref.org/dialog/?doi=10.1128/AEM.01619-08&domain=pdf&date_stamp=2008-12-15

http://aem.asm.org/

organ and typically harbor P. leiognathi as their light organsymbiont (7, 9, 12, 21, 26, 43). The goals of the study were toidentify the developmental stage at which the symbiosis beginsin N. nuchalis, to determine whether initiation of light organontogeny precedes bacterial colonization, and to characterizethe genetic diversity of bacteria initiating the association.

MATERIALS AND METHODS

Collection of fish specimens. Larval and juvenile specimens of N. nuchaliswere collected at Miho Kaigan, Suruga Bay, Honshu, Japan, on 29 July, 1August, and 3 August 2005, from the wave zone 2 to 5 m from shore (approxi-mately 1-m depth) using a hand-towed beach seine net (1 m deep by 3 m wide by1 m long; mesh size, 0.53 mm). The net sampled from the surface to 1 m belowthe surface. On 1 August, seawater temperature and salinity at the sampled siteswere 25.4°C and 28.1 to 28.8 ppt, respectively. Adult specimens of N. nuchaliswere collected by boat seine at Miho Kaigan and from gill nets in Suruga Bay.Specimens of N. nuchalis were identified by reference to the work of Okiyama(42) and Kimura et al. (27). Specimens in which formation of all fin rays was notcomplete, less than 10.0 mm in length (standard length [SL]), were consideredlarvae, whereas specimens in which formation of all fin rays was complete, whichwere 10.0 mm (SL) and larger, were considered juveniles. Furthermore, larvaewere divided into flexion stage, with a 6.0- to 6.7-mm notochord length (NL), andpostflexion stage, with a 7.2- to 9.4-mm SL, respectively, based on whether thetips of the hypural bones had fully assumed a vertical position. In certain cases,larvae smaller than 7.0 mm in SL had completed notochord formation; lengths ofthese specimens are given as SL. Specimens of 43 mm in SL and larger wereconsidered adults.

TEM. Light organs for examination by transmission electron microscopy(TEM), which was carried out by staff of the University of Michigan Microscopyand Image Analysis Laboratory, were dissected from larval specimens of N.nuchalis that had been preserved in Karnovsky’s fixative (2% paraformaldehyde,2.5% glutaraldehyde, 0.1 M sodium phosphate buffer; EM Sciences, Hatfield,PA) and stored at 4°C. Light organs were washed in phosphate buffer, postfixedin buffered osmium tetroxide (1%) for 1 h, and then rinsed, dehydrated inascending strengths of ethanol, infiltrated with propylene oxide, infiltrated withpolyembed 812 epoxy resin, and polymerized. Ultrathin sections were mountedon slotted grids with a supporting membrane, double stained with lead citrate-uranyl acetate, and examined with a Philips CM-100 transmission electron mi-croscope.

Bacterial isolations. Specimens of larval, juvenile, and adult N. nuchalis to beused for bacterial isolations were kept on ice following collection and dissectedaseptically to remove the light organ, which was then homogenized in sterilebuffered (25 mM HEPES buffer, pH 7.25) 70% artificial seawater (39) (100 or500 �l for larval and juvenile specimens and 500 �l or 1 ml for adults). Ten to fivehundred microliters of the light organ homogenates, or dilutions thereof, werespread on LSW-70 agar plates (16), which contained (per liter) 10 g tryptone, 5 gyeast extract, 350 ml double-strength artificial seawater (39), 650 ml deionizedwater, and 15 g agar. Bacterial strains were purified on LSW-70 agar plates,grown overnight in LSW-70 broth at room temperature, and stored in cryopro-tective medium at �75°C (14).

Strain typing. Genomic DNA was purified from 1-ml cultures of strains grownovernight at room temperature in LSW-70 broth, using the Qiagen (Valencia,CA) DNeasy tissue extraction kit.

DNA fingerprint analysis (genomotyping) was carried out by repetitive ele-ment palindromic PCR (rep-PCR) (49), essentially as described previously (13,26). The reaction mixture for rep-PCR contained the following (per 25-�l reac-tion volume): 6.125 �l of tissue culture-grade water (Sigma, St. Louis, MO), 5 �lof 5� Gitschier buffer, 0.2 �l of bovine serum albumin (10 mg ml�1), 2.5 �l ofdimethylsulfoxide (100%), 3.125 �l of deoxynucleoside triphosphates (10 mM),1.25 �l (each) of primers REPIR-1 (5�-IIIICGICGICATCIGGC-3�) andREP2-1 (5�-ICGICTTATCIGGCCTAC-3�) (50 nM), 2.5 �l of MgCl2 (25 mM),and 1 �l of Taq polymerase (5 U �l�1) (Eppendorf). For individual reactions, 23�l of the mixture was combined with 1.5 �l of template DNA (or 2 �l dependingon the concentration of the DNA), prepared as described above. PCR wascarried out using the Bio-Rad (Hercules, CA) iCycler, with the following con-ditions for rep-PCR: hot start and 2 min of denaturation at 95°C; 30 cycles of 30 sat 92°C (denaturation), 1 min at 40°C (primer annealing), and 8 min at 65°C(polymerase extension); a single additional extension step of 8 min at 65°C; andsnap cooling to 4°C. Products of the rep-PCRs were separated on 1.75% agarosegels in 1� TA (0.04 M Tris-acetate) buffer containing 5 �l of 1% ethidiumbromide at 80 V for approximately 4.5 h. Digital images of gels were captured

with a Bio-Rad Fluor S Multi-imager and were examined visually. Repeatedrep-PCR analysis of the same strain consistently gave the same banding pattern.Strains were identified as distinct if their DNA fingerprints differed by one ormore bands.

Amplification of luxA region. The primers CWLAforPl (GTTTTAGATCAACTGTCTAAAGGRCG) and CWLArevPl (TCAGAACCATTCGCTTCAAATCCAAC), designed for amplification of the luxA region from P. leiognathi (26),were used to amplify the luxA region from genomic DNA of N. nuchalis symbi-onts. For amplifications, Taq polymerase and reagents of the Eppendorf (Ham-burg, Germany) MasterTaq kit were used with the following protocol: hot startand 2-min denaturing at 95°C; 35 cycles of 20 s at 94°C (denaturing), 15 s at 50°C(primer annealing), and 1 min at 72°C (polymerase extension); a single addi-tional extension step of 7 min at 72°C; and snap cooling to 4°C.

DNA sequencing and merodiploidy analysis. Sequencing of the luxA ampli-cons, approximately 0.8 kb in length, was carried out by staff of the University ofMichigan Sequencing Core using the amplification primers and dye terminatorcycle sequencing on a Perkin-Elmer (Wellesley, MA) ABI 3730 or 3700 DNAanalyzer. Bacterial identifications were based on an extensive GenBank data setfor luxA genes of luminous bacteria (e.g., see references 2–5, 16, 26, 48, and 49).The sequencing chromatograms were examined for single or multiple discreteluxA sequence peaks to identify the presence of single versus multiple (mero-diploid) lux-rib operons (4).

RESULTS

Structure of nascent light organ in N. nuchalis larvae. Larvaland juvenile specimens of N. nuchalis, ranging in size from 6.0mm in NL to 12.1 mm in SL, were collected by beach seine andexamined for the presence and developmental state of an in-ternal, supraesophageal light organ and the extent to which thelight organ was colonized by luminous bacteria. Tissue consis-tent with a light organ, partially wrapping around the esopha-gus and covered dorsally by a layer of pigment, was evident ineven the smallest specimens obtained, 6.0 to 6.5 mm in NL(Fig. 1). The pigment layer presumably functions to preventlight from passing from the light organ dorsally into andthrough the transparent epaxial musculature of the larval fish.In specimens of this size, the stomach has not yet differenti-ated, and the nascent gasbladder is small and has not yetestablished the interface with the light organ that is character-istic of juvenile and adult leiognathids (1, 15, 19, 32, 33).

The bulk of the light organ in larvae is composed of clustersof host cells among which open spaces are evident (Fig. 2 and3). This architecture is generally similar to the morphology oflight organ tubules of the adult fish, which form the core of thelight organ and house the bacteria extracellularly within thetubule lumina. The tubule lumina in larvae, however, varywidely in diameter within a single light organ, from approxi-mately 2 to 30 �m, with smaller and larger tubules often inclose association. Bacterial cells are present in many of thetubules, whereas many tubules of the same light organ appearto lack bacteria (Fig. 2 and 3).

The bacterial cells are approximately 0.9 by 1.8 �m andexhibit the double-track membranes of gram-negative bacteria(Fig. 2C). The bacterial cells often contain one or more smallspherical refractile inclusions, presumably granules of poly-�-hydroxybutyrate, a carbon storage product formed by P. leiog-nathi in culture but not typically seen for P. leiognathi in sym-biosis with adult fish (7, 8, 11). The polar flagella formed by P.leiognathi grown culture (8) are not evident here, as also notedfor P. leiognathi in symbiosis with adult leiognathids (12), andbacteria experimentally released from the nascent light organinto seawater were not initially motile. Although most bacterialcells show electron opacity consistent with viability, semitrans-

7472 DUNLAP ET AL. APPL. ENVIRON. MICROBIOL.


.asm.org/

Dow

nloaded from

http://aem.asm.org/

parent “ghost” cells are present at low frequency (Fig. 2C);bacterial “ghost” cells have been noted previously in light or-gans of various species of leiognathid fish (7, 11). Evident herealso are bacterial cells within gaps between host cells (Fig. 2Aand B).

Granular material fills many of the tubules lacking bacteria(Fig. 2A and 3A), and this material is less evident or lacking inthose tubules containing high numbers of bacteria. In certainof the small tubules lacking bacteria, extensive microvillus-likeextensions are evident (Fig. 3B and C); in others, dense gran-ular material is present; and in larger tubules that lack bacteria,the material appears more evenly dispersed (Fig. 3A). Overall,these observations suggest that for larvae of N. nuchalis of 6.0to 6.7 mm in NL, the light organ is at an early stage of devel-opment and is undergoing bacterial colonization.

Developmental stage at inception of symbiosis. Consistentwith this interpretation, light organs of some larval specimensin this size range apparently lacked bacteria. To test for thepresence of bacteria, we aseptically dissected the light organ,homogenized it to release bacteria, and then spread all or aportion of the homogenate on plates of seawater-based agarmedium to allow bacteria present to reproduce and form col-onies. Light organs of certain of the smallest specimens exam-ined, 6.0 to 6.5 mm in NL, yielded no colonies, luminous ornonluminous, whereas light organs of other specimens of 6.5mm in NL and all specimens of 6.6 mm in NL and largerharbored large populations of luminous bacteria (Table 1; Fig.

2). These results indicate that inception of the symbiosis ap-parently occurs in flexion larvae of 6.0 to 6.5 mm in NL.Therefore, initiation of light organ ontogeny apparently pre-cedes the host’s acquisition of symbiotic bacteria.

The bacterial colonies arising from the light organ homog-enates of larval and juvenile specimens were luminous, as werethose of adults, and in all cases except one (described below),the luminescence and morphology of the bacterial colonieswere characteristic of P. leiognathi, the bacterial symbiont typ-ically present in light organs of leiognathids (12, 16, 26, 43).Nonetheless, to confirm the identity of these bacteria, we se-quenced the luxA gene of strains representative of each straintype (see below) from each larval, juvenile, and adult speci-men. The luxA gene sequence distinguishes among all knownspecies of luminous bacteria (e.g., 4, 5, 26, 48, 49) and there-fore provides a rapid and effective means of confirming speciesidentifications. All examined strains (with the exception de-scribed below) were confirmed by this method to be membersof P. leiognathi (Table 1).

Bacterial population diversity in larval light organs. Popu-lations of P. leiognathi in light organs of adult leiognathids arecomposed typically of a few to several genetically distinct straintypes (16). Whether this diversity is acquired at the inceptionof the symbiosis, by primary colonization of the nascent lightorgan by genetically distinct strains, or results from secondarycolonization events over time by different strains is not known.To distinguish between these possibilities, we examined the

FIG. 1. Sketch of the position and structure of the internal, supraesophageal light organ of a flexion larva of N. nuchalis (6.3 mm in NL).(A) Lateral view of fish. The light organ (not shown) is located internally, slightly above and interior to the pectoral fin base. (B) Lateral view oflight organ and associated tissues with body wall cut away (a, esophagus; b, light organ; c, gasbladder; d, gut) showing its supraesophageal,pregastric location. (C) Dorsal view of light organ and associated tissues. Note the nascent gasbladder. Bar, 1 mm (A), 0.1 mm (B), or 0.1 mm (C).

VOL. 74, 2008 INCEPTION OF SYMBIOSIS IN THE FISH N. NUCHALIS 7473


.asm.org/

Dow

nloaded from

http://aem.asm.org/

population structure of bacteria in light organs of N. nuchalislarvae.

To identify the number of distinct strain types in each lightorgan population, strains from each larval specimen (Table 1)were typed by rep-PCR, a DNA fingerprinting (genomotyping)method (13, 50). An example of DNA fingerprinting of strains,for bacteria from the light organ of the larval specimenlnuch.40, is shown in Fig. 4. This population analysis revealedthat light organs of larval N. nuchalis carried from one to sixdifferent strain types (n � 15 larval specimens), an average of2.8 strain types per specimen. These values are similar to thosereported previously for symbiont populations of adult leiog-nathid fish, from one to five different strain types (n � 26), anaverage of 2.6 types per specimen (16). Light organs of adultspecimens of N. nuchalis apparently contained a somewhatgreater diversity of strains, from 3 to 7 different strain types(n � 5 adult specimens), an average of 4.0 types per specimen(Table 1), but this difference might be due to the smallersample size. Regardless, bacterial populations in light organsof N. nuchalis near inception of the symbiosis commonly arecomposed of multiple genetically distinct strains, and the ex-tent of symbiont population diversity in light organs of larvaeis similar to that in adult fish. Primary colonization of thenascent light organ by multiple genetically distinct strainstherefore apparently accounts for the symbiont population di-versity seen in adult light organs.

Limited incidence of symbiont strain “sharing.” Previously,for adult leiognathids, we found that the same symbiont straintype was only rarely present in light organs of different fish,

even for specimens of the same host species collected at thesame time and location (16). This limited symbiont “sharing”suggests that the different strain type or types are not specificto individual host species. However, relatively few host speci-mens of the same species from the same time and place havebeen analyzed, leaving open the possibility that the appearanceof limited strain “sharing” can be explained by small samplesize. It is also possible that the different specimens of the fishexamined previously acquired their symbiotic bacteria fromdifferent locations, each with different P. leiognathi strain typespresent.

With the many specimens of larval N. nuchalis collectedfrom the same location over the course of a few days, we wereable to avoid these caveats. The DNA fingerprinting analyses(rep-PCR gels) of symbiont populations of individual larval N.nuchalis, as well as those of juveniles and adults, were exam-ined visually to identify strains from different host specimensthat appeared similar. The genomic DNA of those strains wasthen reanalyzed together on a single rep-PCR gel to facilitateside-by-side comparisons between pairs or groups of strainsdeemed similar to each other (Fig. 5). This analysis revealedthat symbiont strain “sharing” was rare; each symbiont popu-lation was composed almost always of completely differentstrains. Only a single strain type was found to be present inlight organs of more than one fish, represented by strainslnuch.31.1, lnuch.34.16, and lnuch.36.1 from larvae caught on 1August and lnuch.28.12 from an adult fish caught using a gillnet on 30 July. This strain type accounted for 45% of the 55examined strains from these four host specimens. In all other

FIG. 2. TEM of ultrastructure of the light organ of a flexion larva of N. nuchalis (6.7 mm in NL). (A) TEM section through the light organ,showing the pigment layer and host cells forming tubules that either contain or lack bacteria. (B) Light organ tubules containing bacteria.(C) Bacterial cells, some with refractile granules that are likely to be poly-�-hydroxybutyrate, and including a “ghost” cell, in a light organ tubule.Bar, 20 �m (A), 5 �m (B), or 1 �m (C).



.asm.org/

Dow

nloaded from

http://aem.asm.org/

cases, the apparent similarity of strains from different fishspecimens was not supported by direct side-by-side compari-sons of the genomic fingerprints (Fig. 5); although similar inDNA fingerprints, the other compared strains were genomo-typically distinct. These results confirmed the earlier finding oflimited symbiont “sharing” among adult leiognathids (16) andextended that finding to multiple specimens of a single hostspecies near inception of the symbiosis and collected from thesame general location within a span of a few days. These resultstherefore affirm the view that colonization of the fish’s lightorgan appears to be random with respect to symbiont straintype (16).

Incidence of lux-rib merodiploid strains. In addition to thegenetic diversity in P. leiognathi described above, many strainsof this species also carry multiple copies of the lux-rib operon,an unusual case of merodiploidy in bacteria (4). Most of thestrains examined previously and found to be merodiploid forthe lux-rib operon were isolated from light organs of adultleiognathid fish, but whether merodiploidy might relate insome way to symbiosis is not known. To gain insight on thisissue, we compared the incidence of merodiploid strains inlight organs of larval, juvenile, and adult N. nuchalis. Thechromatogram of the luxA gene sequence for each strain typewas examined for multiple discrete peaks, which are diagnosticof lux-rib merodiploidy (4). We found that 170 of the 249strains from the 16 examined larvae carried a single lux-riboperon whereas 79 strains were merodiploid (Table 1), a ratioof 2.2 to 1. Light organs of juveniles and adults also carriedmerodiploid strains. Adult specimens of N. nuchalis, however,had a much higher proportion of merodiploid strains thanlarvae. Of 91 strains from five adult specimens, 16 carried a

single lux-rib operon whereas 75 were merodiploid (Table 1), aratio of 1 to 4.7. These results suggest that the proportion ofstrains carrying multiple lux-rib operons increases as the sym-biosis matures.

Occurrence of Vibrio harveyi in light organ symbiosis. Lightorgans of most specimens of leiognathid fish carry only P.leiognathi (12, 16, 43; this study). Although highly specific,however, the association is not strictly exclusive to P. leiognathi;some leiognathids also harbor P. mandapamensis together withP. leiognathi (26). Here we found that platings of the lightorgan homogenate of one of the larval N. nuchalis specimens,lnuch.41, yielded luminous colonies of two different kinds, sug-gesting the presence of two bacterial species. The more-numerous kind was typical of the bacteria from light organs ofother larval, juvenile, and adult leiognathid fish in colony mor-phology, luminescence color and intensity, lack of distinctivepigmentation, and absence of a strong odor; representatives ofthis kind were identified as P. leiognathi by luxA sequenceanalysis (Table 1). The less-numerous kind, present at approx-imately 15% of the total number of colonies, formed large,rapidly growing colonies of a bluer color of luminescence(compared to the more blue-green color of luminescence of P.leiognathi colonies). When strains of this kind were grown inpure culture, the colonies developed a diffusible brown pig-ment over time and produced a strong, putrid odor. The traitsof rapid growth, blue luminescence, brown pigment, and putridodor are characteristic of V. harveyi and related bacteria (8;P. V. Dunlap, personal observation). Twenty strains of thiskind were isolated and characterized by rep-PCR DNA finger-printing; all were genomotypically identical (data not shown).The luxA gene from two of these strains, lunch.41.H1 and

FIG. 3. TEM of tubules lacking bacteria in the light organ of the flexion larva of N. nuchalis (6.7 mm in NL) shown in Fig. 2. (A) Larger tubuleslacking bacteria, showing the dispersed and dense granular material and microvillus-like extensions in different tubules (see the text). (B) Closeupof smaller tubule (lower left in panel A) with microvillus-like extensions filling the tubule. (C) Closeup (from panel B) of microvillus-likeextensions, showing their contiguity with host cell membranes. Bar, 4 �m (A), 1 �m (B), or 400 nm (C).



.asm.org/

Dow

nloaded from

http://aem.asm.org/

lunch.41.H2, was sequenced, and analysis of those sequencesidentified the strains as V. harveyi (Table 1). The high inci-dence of strains of this kind, their luminescence, and theirgenomotypic uniformity are characteristic of bacteria fromlight organs (e.g., see references 13 and 16); they therefore arenot likely to be chance contaminants arising from the gut tract.This is the first indication that V. harveyi, a bacterial species notknown to be a bioluminescent symbiont, apparently can occurin the light organ habitat.

DISCUSSION

The developmental and microbiological events surroundinginception of symbiosis are central to understanding how lumi-nous fish establish their bioluminescent associations. Here, forthe leiognathid fish N. nuchalis, we show that light organ on-togeny is initiated early in development of the fish and appar-ently precedes acquisition of the bacteria. Microbiologically,we show that the symbiosis is initiated by multiple, genetically

TABLE 1. Population structure of light organ symbionts from larval-, juvenile-, and adult-stage specimens of N. nuchalis

Specimen Size(mm)a

Light organpopulation sizeb

No. of strainsexamined

No. ofgenomotypes Strain(s) for which the luxA gene was sequencedc S or M lux-rib

operonse

Larval stagelnuch.L1 6.0 0lnuch.31 6.2 1 1d 1 lnuch.31.1 (1) M (1)lnuch.L2 6.5 0lnuch.32 6.5 �1.7 � 104 17 4 lnuch.32.2 (4), lnuch.32.3 (3), lnuch.32.4 (5),

lnuch.32.7 (5)S (17)

lnuch.42 6.5 �2.0 � 103 17 1 lnuch.42.2 (17) M (17)lnuch.44 6.5 �2.0 � 103 16 2 lnuch.44.14 (5), lnuch.44.15 (11) M (16)lnuch.45 6.6 �2.0 � 103 17 3 lnuch.45.5 (12), lnuch.45.13 (4) S (16)

lnuch.45.10 (1) M (1)lnuch.43 6.7 �2.0 � 103 17 1 lnuch.43.2 (17) S (17)lnuch.40 7.2 �2.0 � 103 17 5 lnuch.40.1 (7), lnuch.40.9 (2), lnuch.40.10 (1),

lnuch.40.12 (2)S (12)

lnuch.40.3 (5) M (5)lnuch.41 7.3 �2.0 � 103 3 2 lnuch.41.1 (2) M (2)

lnuch.41.2 (1) S (1)20 1 lnuch.41.H1, lnuch.41.H2c

lnuch.33 7.6 �2.3 � 103 16 1 lnuch.33.1 (16) S (16)lnuch.34 8.2 �2.5 � 104 17 2 lnuch.34.13 (13), lnuch.34.16 (4) M (17)lnuch.25 8.4 �2.4 � 102 18 3 lnuch.25.11 (8), lnuch.25.15 (7), lnuch.25.17 (3) S (18)lnuch.23 9.1 �5.4 � 104 19 5 lnuch.23.1 (7), lnuch.23.2 (6), lnuch.23.3 (4),

lnuch.23.11 (1), lnuch.23.15 (1)S (19)

lnuch.26 9.1 �2.9 � 103 20 1 lnuch.26.8 (20) S (20)lnuch.35 9.2 �2.3 � 104 17 3 lnuch.35.9 (2), lnuch.35.10 (11), lnuch.35.14 (4) S (17)lnuch.24 9.4 �2.0 � 103 18 2 lnuch.24.9 (4), lnuch.24.13 (14) M (18)lnuch.36 9.4 �1.5 � 104 19 6 lnuch.36.1 (9), lnuch.36.2 (4), lnuch.36.12 (1),

lnuch.36.14 (2), lnuch.36.15 (1)S (17)

lnuch.36.3 (2) M (2)

Juvenile stagelnuch.37 10.2 �1.0 � 104 15 2 lnuch.37.12 (5), lnuch.37.13 (10) M (15)lnuch.38 12.1 �6.0 � 104 18 1 lnuch.38.2 (18) M (18)

Adult stagelnuch.27 98 �4.9 � 107 20 7 lnuch.27.1 (2), lnuch.27.2 (6), lnuch.27.3 (3),

lnuch.27.10 (2), lnuch.27.12 (1),lnuch.27.16 (3)

M (17)

lnuch.27.9 (3) S (3)lnuch.28 94 NDf 18 3 lnuch.28.11 (4) S (4)

lnuch.28.12 (11), lnuch.28.13 (3) M (14)lnuch.29 92 ND 16 4 lnuch.29.7 (3), lnuch.29.10 (1), lnuch.29.11 (9),

lnuch.29.12 (3)M (16)

lnuch.30 87 ND 17 3 lnuch.30.14 (7), lnuch.30.15 (9), lnuch.30.16 (1) M (17)lnuch.39 111 ND 20 3 lnuch.39.1 (7), lnuch.39.3 (4) M (11)

lnuch.39.5 (9) S (9)

a NL for flexion larvae (6.0 to 6.7 mm) or SL for postflexion larvae (7.2 to 9.4 mm), juveniles, and adults.b Estimates of population size are based on plate counts from platings of light organ homogenates.c All listed strains were identified as P. leiognathi based on luxA gene sequence analysis, with the exception of lnuch.41.H1 and lnuch.41.H2, which were identified

as V. harveyi. Numbers in parentheses following designations for specific strains for which the luxA gene was sequenced are the numbers of strains of the samegenomotype as the designated strain.

d Not used in counts of strain types found in each fish.e Numbers in parentheses following S (single) or M (multiple) designation are the numbers of strains bearing single or multiple lux-rib operons, based on genomotypic

identity with sequenced strains.f ND, not determined; for estimates of light organ population sizes in adult leiognathids, see reference 11.



.asm.org/

Dow

nloaded from

http://aem.asm.org/

distinct strains of P. leiognathi, that the same strain types areonly rarely found in light organs of different host specimens,and that the association, although highly specific for P. leiog-nathi, is not completely exclusive of other bacteria. Becausethis study is based on wild-caught specimens of the fish, theresults and observations reported here also provide insight intolife history traits of the fish that may be critical for continuityof the association.

For other bacterially luminous fish, analysis of early devel-opmental stages has been limited to microscopy of a few ex-perimentally reared or wild-caught specimens. Larvae fromexperimentally fertilized eggs of Monocentris japonica (Beryci-formes: Monocentridae) were reared for 3 weeks, to a 6-mmtotal length, at which point they had not yet developed theirmandibular light organs (52). Incipient subocular light organswere observed in 4.1-mm-NL flexion and 5.8-mm-SL postflex-ion larvae of Anomalops katoptron (Beryciformes: Anomalopi-dae) from the plankton (28) and in a wild-caught 6.2-mm-NLflexion larva of Kryptophaneron alfredi (Beryciformes: Anoma-lopidae) (6). A perianal light organ was evident in preflexionand flexion larvae of Paratrachichthys and Aulotrachichthys(Beryciformes: Trachichthyidae) collected from the plankton(25). In these studies, the possible presence of bacteria was not

reported. Bacteria were mentioned in other studies, however.A nascent supraesophageal light organ with a low number ofbacteria and an interface with the gasbladder was evident in a9.8-mm-total-length wild-caught specimen of Equula (Perci-formes: Leiognathidae) (19). Bacterial colonization of the nas-cent ventral light organ of larvae of Siphamia versicolor (Per-ciformes: Apogonidae) from the plankton was not evident in2.4-mm- and 2.8-mm-NL preflexion specimens but was notedfor a 3.5-mm-SL flexion specimen (31). In the nascent escallight organs of two wild-caught specimens of Melanocetus mur-rayi (Lophiiformes: Melanocetidae), bacteria were apparentlyabsent from the light organ of the smaller specimen (18.5 mmin SL) but were present in the light organ of the larger speci-men (36.0 mm in SL), and the presence of bacteria in the largerspecimen correlated with formation of a duct connecting thelight organ to the environment (38). Juveniles of N. nuchalis,reared from experimentally fertilized eggs, apparently wereaposymbiotic and were able to initiate the symbiosis whenpresented with P. leiognathi (51). In general, these studiesindicate that ontogeny of light organs in various bacteriallyluminous fish begins early in fish development, in most casesapparently during preflexion or flexion stages. They suggestalso that initiation of light organ ontogeny precedes coloniza-

FIG. 4. DNA fingerprint (rep-PCR) analysis of bacteria from the light organ of a postflexion larva of N. nuchalis (lnuch.40, 7.2 mm in NL).Visual inspection identified five strain types from symbiosis with this fish, represented by lnuch.40.1, lnuch.40.2, lnuch.40.7, lnuch.40.8, lnuch.40.11,lnuch.40.15, and lnuch.40.16 (first type); lnuch.40.3, lnuch.40.4, lnuch.40.5, lnuch.40.6, and lnuch.40.17 (second type); lnuch.40.9 and lnuch.40.14(third type); lnuch.40.12 and lnuch.40.13 (fourth type); and lnuch.40.10 (fifth type). See Table 1 for more information. Left and right flanking lanesare 1-kb and 100-bp size markers, respectively.



.asm.org/

Dow

nloaded from

http://aem.asm.org/

tion by the symbiotic bacteria, which presumably are acquiredfrom the environment.

The limited information on fish contrasts with an extensiveliterature from laboratory studies of the bacterially luminoussquid, Euprymna scolopes (Sepiolida: Sepiolidae), in which de-velopmental events both during bacterial colonization and inthe absence of bacterial colonization have been investigated indetail (e.g., see references 10, 36, 44, and 45). Nonetheless, theresults presented here on N. nuchalis indicate major differ-ences with E. scolopes. The squid, which harbors A. fischeri asits symbiont, undergoes direct development, with embryoshatching as juveniles, which are morphologically similar toadults (10, 36), whereas the fish undergoes indirect develop-ment, passing through a series of developmental steps, i.e.,postembryo and preflexion, flexion, and postflexion larvalstages, before attaining the juvenile stage. Shortly after hatch-ing, the squid acquires its symbiotic bacteria, whereas the workpresented here indicates that the fish apparently initiates sym-biosis several days after hatching, after substantial posthatch-ing development has occurred. Another difference is ecologi-cal; the squid hatches into the habitat where the adult squid

live and therefore encounters bacteria presumably releasedfrom the local population of adult squids (29, 30, 45). In con-trast, the fish apparently hatches some distance from locationswhere adults live and to initiate symbiosis must migrate in-shore, to a habitat where its symbiotic bacteria are more abun-dant (discussed below).

In flexion and postflexion larvae of N. nuchalis, the lightorgan system is at an early stage of development. We estimatedthe age of a 6.5-mm-SL postflexion larva at approximately 19days posthatch based on the presence of 19 incremental ringson its otoliths and assuming the rings are formed daily. Thenascent light organ in fish of this size is rudimentary in overallstructure compared to that of adult fish (1, 7, 19), and tissueforming it appears to be undergoing active growth, with newtubules being formed, many of which apparently are not yetcolonized by bacteria. Organs closely associated with the lightorgan also are early in development. The stomach, which de-velops just posterior to the light organ, has not yet differenti-ated in fish of this size, and the nascent gasbladder, a keycomponent of the leiognathid light organ system, is small andhas not established the interface with the light organ that is

FIG. 5. DNA fingerprint (rep-PCR) analysis of strains deemed genomotypically similar from light organs of different N. nuchalis specimens.White bars indicate pairs or groups of strains identified by visual comparison of DNA fingerprints of individual larval, juvenile, and adult symbiontpopulations (see Fig. 4 and Table 1 for an example and additional details) as possibly similar to each other and reexamined here in side-by-sidecomparisons to test for identity. Only strains lnuch.31.1, lnuch.34.16, lnuch.36.1 (from larvae), and lnuch.28.12 (from an adult fish) were confirmedhere as identical in DNA fingerprint; all other examined strains differed by one or more bands. Left and right flanking lanes are 100-bp sizemarkers.



.asm.org/

Dow

nloaded from

http://aem.asm.org/

characteristic of the mature symbiosis (1, 15, 19, 32, 33). Theapparent absence of bacteria in light organs of the smallerflexion larvae of N. nuchalis indicates that inception of thesymbiosis most likely occurs at a host size of 6.0 to 6.5 mm inNL and that initiation of light organ ontogeny precedes bac-terial colonization.

The high degree of species-level specificity of P. leiognathifor leiognathid fish contrasts with the substantial populationgenetic diversity of this bacterium and the absence of bacterialstrain-host species specificity (12, 16, 26, 43; this study). Sym-biont populations in light organs of individual leiognathid fishare typically composed of multiple genetically distinct straintypes. However, the same strain type occurs only rarely in lightorgans of different fish specimens, even for specimens of thesame host species collected close to inception of the symbiosisand from the same location within a few days (16; this study).The absence of host species-related strain specificity is consis-tent with the lack of congruence between the phylogenies ofleiognathid fish and their light organ bacteria (17), and therareness with which the same strain type is found in lightorgans of different fish specimens (16; this study) suggests thepresence of extensive genetic diversity in the global populationof P. leiognathi. In this regard, the similar levels of symbiontdiversity in adults and in larvae indicate that the symbiosiscommonly is initiated by multiple, genetically distinct bacteria;it is not necessary to invoke secondary colonization to explainsymbiont population diversity in light organs of adult fish (16).Colonization of the nascent N. nuchalis light organ thereforeappears to be random with respect to P. leiognathi strain type,with aposymbiotic flexion larvae typically picking up two, three,or more different representatives of the many different straintypes present in their local environment. These strains serve asfounders of the symbiont population.

Whether changes occur over time in the strain compositionof light organ populations of individual specimens of N. nu-chalis is an open question. The complex microarchitecture ofthe light organ of adults, with heavily colonized tubules, effluxof excess bacterial cells from the tubules into collecting cham-bers, and ducts connecting the chambers to the esophagus (1,7, 11, 19), suggests that secondary colonization may not occur.Competition among founding strains could alter the straincomposition of the light organ over time, but the diversity ofstrains in light organs of adults, the similar levels of straindiversity in larvae and adults, and the lack of host species-related strain specificity do not appear to be consistent with acompetitive succession of strain types. Clonal divergence couldlead to a change in population structure, but an experimentaltest for clonal divergence of a strain of P. leiognathi underlaboratory conditions found no detectable change in the DNAfingerprint after approximately 400 generations (4). In thisregard, the higher proportion of lux-rib merodiploid strains inadults compared to larvae is intriguing, since it might resultfrom transposon-mediated transfer of the second lux-riboperon among strains in the symbiosis, leading to an increasingproportion of strains that are merodiploid as the symbiosismatures (4). Alternatively, merodiploid strains might have acompetitive advantage in the symbiosis (4), coming to propor-tionally dominate a population over time. A small sample sizemight also be a factor, and additional work will be necessary todistinguish among these possibilities.

The identification of V. harveyi in the light organ of a spec-imen of N. nuchalis is unexpected. This bacterial species hadnot been isolated from the light organ habitat previously, andits interactions with marine animals often are pathogenic (8,14, 22). Whether this instance represents a rare occurrence ora more common interaction is not known, but the presence ofV. harveyi in coastal waters where larvae of N. nuchalis occur(49) indicates the opportunity for this bacterium to encounterfish whose nascent light organs are in the process of beingcolonized. One possibility is that V. harveyi is acquired onlyinfrequently by larvae of N. nuchalis and that the pathogenicityof this bacterium leads to death of the fish; as a consequence,larvae that have taken up this bacterium into their nascent lightorgans would be short-lived and therefore only rarely found.An alternative possibility is that the light organs of larval N.nuchalis sometimes are colonized by V. harveyi, which gains atemporary foothold in the light organ but is outcompeted by P.leiognathi as the light organ develops and the symbiosis ma-tures. Regardless, the finding of V. harveyi in the light organ ofa larval specimen of N. nuchalis is a further indication (26) thatleiognathid light organs are not the exclusive, species-specifichabitat of P. leiognathi.

Although little is known about the early life history of anybacterially luminous fish, information of this kind is central tounderstanding the ecological interactions between host andsymbiont that ensure continuity of the association in new hostgenerations. Particularly important is knowledge of the habitatin which the symbiosis begins, because larval stages of the fishmay acquire their species and strains of symbiont in a habitatvery different microbiologically from where adults are col-lected. In this regard, information on the life history of N.nuchalis in Suruga Bay is accumulating and leads to the fol-lowing scenario. Adult N. nuchalis fish are demersal, occurringalong the bottom in relatively shallow areas of the bay, typicallydown to an approximately 20-m depth. The adults spawn in thishabitat. Currents then disperse the fertilized eggs, which areplanktonic and separate (18), out into the open waters of thebay, up to 1 to 2 kilometers offshore, where the embryos hatchand larval development begins. The preflexion larvae, whichpresumably are aposymbiotic, then migrate inshore over thenext 15 to 20 days as they develop. When they reach the wavezone along the shoreline, as flexion larvae of approximately 6to 6.5 mm in NL, they acquire their symbiotic bacteria andinitiate the symbiosis. They continue to develop in this habitatinto the postflexion and early juvenile stages. As larger juve-niles, they migrate along the bottom away from the wave zone,ultimately taking up demersal existence as adults in deeperwater.

In support of this scenario, preflexion larvae of N. nuchalisranging in size from 1.6 to 4.3 mm in NL have been collectedby plankton tows offshore in open waters of Suruga Bay,whereas flexion and postflexion larvae and early juveniles arefound in the wave zone and typically range from 6 to 10 mm inlength (rarely less than 6 mm in NL and infrequently greaterthan 10 mm in SL) (M. Fujino, S. Tomiyama, and A. Fukui,unpublished data). Furthermore, P. leiognathi is readily foundin the wave zone along the shoreline in Suruga Bay (49), but itsincidence in offshore waters of the bay appears to be very low(K. Davis, P. V. Dunlap and A. Fukui, unpublished data).Thus, migration of aposymbiotic larvae from open waters of



.asm.org/

Dow

nloaded from

http://aem.asm.org/

the bay to the wave zone along the shoreline, where the fishapparently acquires its symbiont, may be a part of the lifehistory of this fish that is critical to continuity of the biolumi-nescent symbiosis. However, the ontological status and processof development of the nascent light organ in preflexion larvaeof the fish have not yet been described. Therefore, althoughthis scenario is supported by some preliminary data, testing itwill require a more detailed understanding of the incidence ofpreflexion larvae in Suruga Bay and the developmental andmicrobiological status of the nascent light organs of these hostspecimens.

ACKNOWLEDGMENTS

We thank H. Ichikawa and K. Watanabe for assistance with beachseining, K. Kimura for the gift of adult specimens of N. nuchalis, J. Astfor advice and guidance in DNA sequence analysis, and S. Meshinchifor carrying out the TEM. DNA sequencing was carried out by staff ofthe University of Michigan Sequencing Core.

This work was supported by grant DEB 0413441 from the NationalScience Foundation.

REFERENCES

1. Ahrens, G. 1965. Untersuchungen am Leuchtorgan von Leiognathus klunz-ingeri (Steindachner). Z. Wiss. Zool. 173:90–113.

2. Ast, J. C., and P. V. Dunlap. 2004. Phylogenetic analysis of the lux operondistinguishes two evolutionarily distinct clades of Photobacterium leiognathi.Arch. Microbiol. 181:352–361.

3. Ast, J. C., and P. V. Dunlap. 2005. Phylogenetic resolution and habitatspecificity of members of the Photobacterium phosphoreum species group.Environ. Microbiol. 7:1641–1654.

4. Ast, J. C., H. Urbanczyk, and P. V. Dunlap. 2007. Natural merodiploidy ofthe lux-rib operon of Photobacterium leiognathi from coastal waters of Hon-shu, Japan. J. Bacteriol. 189:6148–6158.

5. Ast, J. C., I. Cleenwerck, K. Engelbeen, H. Urbanczyk, F. L. Thompson, P. DeVos, and P. V. Dunlap. 2007. Photobacterium kishitanii sp. nov., a luminousmarine bacterium symbiotic with deep-sea fish. Int. J. Syst. Evol. Microbiol.57:2073–2078.

6. Baldwin, C. C., and G. D. Johnson. 1995. A larva of the Atlantic flashlightfish, Kryptophaneron alfredi (Beryciformes: Anomalopidae), with a compar-ison of beryciform and stephanoberyciform larvae. Bull. Mar. Sci. 56:1–24.

7. Bassot, J.-M. 1975. Les organes lumineux a bacteries symbiotiques dequelques Teleosteens leiognathides. Arch. Zool. Exp. Gen. 116:359–373.

8. Baumann, P., and R. H. Schubert. 1984. Family II. Vibrionaceae Veron1965, p. 516–548. In N. R. Kreig and J. G. Holt (ed.), Bergey’s manual ofsystematic bacteriology, 1st ed., vol. 1. Williams & Wilkins, Baltimore, MD.

9. Boisvert, H., R. Chatelain, and J.-M. Bassot. 1967. Etude d’un Photobacte-rium isole de l’organe lumineux des poissons Leiognathidae. Ann. Inst.Pasteur (Paris) 112:520–524.

10. Claes, M. F., and P. V. Dunlap. 2000. Aposymbiotic culture of the sepiolidsquid Euprymna scolopes: role of the symbiotic bacterium Vibrio fischeri inhost animal growth, development, and light organ morphogenesis. J. Exp.Zool. 286:280–296.

11. Dunlap, P. V. 1984. Physiological and morphological state of the symbioticbacteria from light organs of ponyfish. Biol. Bull. 167:410–425.

12. Dunlap, P. V. 1985. Osmotic control of luminescence and growth in Photo-bacterium leiognathi from ponyfish light organs. Arch. Microbiol. 141:44–50.

13. Dunlap, P. V., and J. C. Ast. 2005. Genomic and phylogenetic characteriza-tion of luminous bacteria symbiotic with the deep-sea fish Chlorophthalmusalbatrossis (Aulopiformes: Chlorophthalmidae). Appl. Environ. Microbiol.71:930–939.

14. Dunlap, P. V., and K. Kita-Tsukamoto. 2006. Luminous bacteria, p. 863–892.In M. Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer, and E. Stack-ebrandt (ed.), The prokaryotes: a handbook on the biology of bacteria, 3rded., vol. 2. Ecophysiology and biochemistry. Springer-Verlag, New York,NY.

15. Dunlap. P. V., and M. J. McFall-Ngai. 1987. Initiation and control of thebioluminescent symbiosis between Photobacterium leiognathi and leiognathidfish. Ann. N. Y. Acad. Sci. 503:269–283.

16. Dunlap, P. V., A. Jiemjit, J. C. Ast, M. M. Pearce, R. R. Marques, and C. R.Lavilla-Pitogo. 2004. Genomic polymorphism in symbiotic populations ofPhotobacterium leiognathi. Environ. Microbiol. 6:145–158.

17. Dunlap, P. V., J. C. Ast, S. Kimura, A. Fukui, T. Yoshino, and H. Endo. 2007.Phylogenetic analysis of host-symbiont specificity and codivergence in bio-luminescent symbioses. Cladistics 23:507–532.

18. Fujita, S. 1960. Egg development and prelarval stages of development of asilver belly, Leiognathus nuchalis. Bull. Jpn. Soc. Sci. Fish. 26:1091–1094.

19. Harms, J. W. 1928. Bau und entwicklung eines eigenartigen leuchtorgans beiEquula spec. Z. Wiss. Zool. 131:157–179.

20. Hastings, J. W. 1971. Light to hide by: ventral luminescence to camouflagethe silhouette. Science 173:1016–1017.

21. Hastings, J. W., and G. Mitchell. 1971. Endosymbiotic bioluminescent bac-teria from the light organ of pony fish. Biol. Bull. 141:261–268.

22. Hastings, J. W., and K. H. Nealson. 1981. The symbiotic luminous bacteria,p. 1332–1345. In M. P. Starr, H. Stolp, H. G. Truper, A. Balows, and H. G.Schlegel (ed.), The prokaryotes: a handbook on habitats, isolation, andidentification of bacteria. Springer-Verlag, Berlin, Germany.

23. Haygood, M. G., B. M. Tebo, and K. H. Nealson. 1984. Luminous bacteria ofa monocentrid fish (Monocentris japonicus) and two anomalopid fish (Photo-blepharon palpebratus and Kryptophaneron alfredi): population sizes andgrowth within the light organs, and rates of release into the seawater. Mar.Biol. 78:249–254.

24. Herring, P. J., and J. G. Morin. 1978. Bioluminescence in fish, p. 273–329. InP. J. Herring (ed.), Bioluminescence in action. Academic Press, London,United Kingdom.

25. Jordan, A. R., and B. D. Bruce. 1993. Larval development of three roughyspecies complexes (Pisces: Trachichthyidae) from southern Australian wa-ters, with comments on the occurrence of orange roughy Hoplostethus atlan-ticus. Fish. Bull. U. S. 91:76–86.

26. Kaeding, A. J., J. C. Ast, M. M. Pearce, H. Urbanczyk, S. Kimura, H. Endo,M. Nakamura, and P. V. Dunlap. 2007. Phylogenetic diversity and co-sym-biosis in the bioluminescent symbioses of Photobacterium mandapamensis.Appl. Environ. Microbiol. 73:3173–3182.

27. Kimura, S., R. Kimura, and K. Ikejima. 2008. Revision of the genusNuchequula with descriptions of three new species (Perciformes: Leiognathi-dae). Ichthyol. Res. 55:22–42.

28. Konishi, Y., and M. Okiyama. 1997. Morphological development of fourtrachichthyid larvae (Pisces: Beryciformes), with comments on trachichthyoidrelationships. Bull. Mar. Sci. 60:66–88.

29. Lee, K.-H., and E. G. Ruby. 1994. Competition between Vibrio fischeri strainsduring initiation and maintenance of a light organ symbiosis. J. Bacteriol.176:1985–1991.

30. Lee, K.-H., and E. G. Ruby. 1995. Symbiotic role of the viable by noncultur-able state of Vibrio fischeri in Hawaiian coastal seawater. Appl. Environ.Microbiol. 61:278–283.

31. Leis, J. M., and S. Bullock. 1986. The luminous cardinalfish Siphamia(Pisces, Apogonidae): development of larvae and the luminous organ, p.703–714. In T. Uyeno, R. Arai, T. Taniuchi, and K. Matsuura (ed.), Indo-Pacific fish biology: proceedings of the second international conference onIndo-Pacific fish. Ichthyological Society of Japan, Tokyo.

32. McFall-Ngai, M. J. 1983. Adaptation for reflection of bioluminescent light inthe gasbladder of Leiognathus equulus (Perciformes: Leiognathidae). J. Exp.Zool. 227:23–33.

33. McFall-Ngai, M. J. 1983. The gasbladder as a central component of theleiognathid bacterial light organ symbiosis. Am. Zool. 23:907.

34. McFall-Ngai, M. J., and P. V. Dunlap. 1983. Three new modes of lumines-cence in the leiognathid fish Gazza minuta (Perciformes: Leiognathidae):discrete projected luminescence, ventral body flash and buccal luminescence.Mar. Biol. 73:227–237.

35. McFall-Ngai, M. J., and J. G. Morin. 1991. Camouflage by disruptive illu-mination in leiognathids, a family of shallow-water, bioluminescent fish. J.Exp. Biol. 156:119–137.

36. McFall-Ngai, M. J., and E. G. Ruby. 1991. Symbiont recognition and sub-sequent morphogenesis as early events in an animal-bacterial mutualism.Science 254:1491–1494.

37. Morin, J. G., A. Harrington, K. Nealson, N. Krieger, T. O. Baldwin, andJ. W. Hastings. 1975. Light for all reasons: versatility in the behavioralrepertoire of the flashlight fish. Science 190:74–76.

38. Munk, O., K. Hansen, and P. J. Herring. 1998. On the development andstructure of the escal light organ of some melanocetid deep sea anglerfish(Pisces: Ceratioidei). J. Mar. Biol. Assoc. U. K. 78:1321–1335.

39. Nealson, K. H. 1978. Isolation, identification and manipulation of luminousbacteria. Methods Enzymol. 57:153–166.

40. Nealson, K. H., and J. W. Hastings. 1991. The luminous bacteria, p. 625–639.In A. Balows, H. G. Truper, M. Dworkin, W. Harder and K.-H. Schleifer(ed.), The prokaryotes. Springer-Verlag, New York, NY.

41. Nealson, K. H., M. G. Haygood, B. M. Tebo, M. Roman, E. Miller, and J. E.McCosker. 1984. Contribution of symbiotically luminous fish to the occur-rence and bioluminescence of luminous bacteria in seawater. Microb. Ecol.10:69–77.

42. Okiyama, M. (ed.) 1988. An atlas of early stage fish in Japan. Tokai Uni-versity Press, Tokyo, Japan.

43. Reichelt, J. L., K. Nealson, and J. W. Hastings. 1977. The specificity ofsymbiosis: pony fish and luminescent bacteria. Arch. Microbiol. 112:157–161.

44. Ruby, E. G. 1996. Lessons from a cooperative, bacterial-animal association:the Vibrio fischeri-Euprymna scolopes light organ symbiosis. Annu. Rev. Mi-crobiol. 50:591–624.

45. Ruby, E. G., and L. M. Asato. 1993. Growth and flagellation of Vibrio fischeri



.asm.org/

Dow

nloaded from

http://aem.asm.org/

during initiation of the sepiolid squid light organ symbiosis. Arch. Microbiol.159:160–167.

46. Sasaki, A., K. Ikejima, S. Aoki, N. Azuma, N. Kashimura, and M. Wada.2003. Field evidence for bioluminescent signaling in the pony fish, Leiog-nathus elongatus. Environ. Biol. Fish. 66:307–311.

47. Sparks, J. S., P. V. Dunlap, and W. L. Smith. 2005. Evolution and diversificationof a sexually dimorphic luminescent system in ponyfish (Teleostei: Leiognathi-dae), including diagnoses for two new genera. Cladistics 21:305–327.

48. Urbanczyk, H., J. C. Ast, M. J. Higgins, J. Carson, and P. V. Dunlap. 2007.Reclassification of Vibrio fischeri, Vibrio logei, Vibrio salmonicida and Vibriowodanis as Aliivibrio fischeri gen. nov., comb. nov., Aliivibrio logei comb. nov.,Aliivibrio salmonicida comb. nov., and Aliivibrio wodanis comb. nov. Int. J.Syst. Evol. Microbiol. 57:2823–2829.

49. Urbanczyk, H., J. C. Ast, A. J. Kaeding, J. D. Oliver, and P. V. Dunlap. 2008.Phylogenetic analysis of the incidence of lux gene horizontal transfer inVibrionaceae. J. Bacteriol. 190:3494–3504.

50. Versalovic, J., J. Schneider, F. J. de Bruijn, and J. R. Lupski. 1994. Genomicfingerprinting of bacteria using repetitive sequence-based polymerase chainreaction. Methods Mol. Cell Biol. 5:25–40.

51. Wada, M., N. Azuma, N. Mizuno, and H. Kurokura. 1999. Transfer ofsymbiotic luminous bacteria from parental Leiognathus nuchalis to offspring.Mar. Biol. 135:683–687.

52. Yamada, K., M. Haygood, and H. Kabasawa. 1979. On fertilization and earlydevelopment in the pine-cone fish, Monocentris japonicus. Ann. Rep.Keikyu Aburatsubo Mar. Park Aquar. 10:31–38.



.asm.org/

Dow

nloaded from

http://aem.asm.org/

Documents

Developmental and Microbiological Analysis of the ... · detailed microbiological analysis of early developmental stages has not been carried out with any wild-caught bacterially