SEARCH

SEARCH BY CITATION

Keywords:

  • Photobacterium;
  • luminous bacteria;
  • symbiosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Photobacterium leiognathi is a facultative bioluminescent symbiont of marine animals. Strains of P. leiognathi that are merodiploid for the luminescence genes (lux-rib operon) have been previously obtained only from Japan. In contrast, strains bearing a single lux-rib operon have been obtained from all the areas sampled in Japan and the western Pacific. In this study, we tested whether distribution of merodiploid P. leiognathi is limited by physical barriers in the environment, or because fish in the western Pacific preferentially form symbiosis with bacteria bearing a single lux-rib operon. We collected light organ symbionts from Secutor indicius, a fish species that is typically found in the western Pacific and has only recently expanded its geographic range to Japan. We found that all S. indicius specimens collected from Japan formed symbiosis only with single lux-rib operon-bearing strains, although fish from other species collected from the same geographic area frequently contained merodiploid strains. This result shows that S. indicius were preferentially colonized by bacteria bearing a single lux-rib operon and suggests that the limited geographic distribution of merodiploid P. leiognathi can be attributed to preferential colonization of fish species found in the western Pacific by strains bearing only a single lux-rib operon.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Bioluminescence, that is, production and emission of light, is a frequently observed phenotype in marine animals (Widder, 2010). Bioluminescence has been observed in over 460 species of marine fish, including agriculturally important species (Widder, 2010; Urbanczyk et al., 2011). The animals use bioluminescence during mating, for prey detection or predator avoidance (Johnsen et al., 2004; Widder, 2010). In marine animals, bioluminescence frequently depends on mutualistic symbiosis with luminescent bacteria (Dunlap & Kita-Tsukamoto, 2006; Dunlap, 2009; Widder, 2010). Symbiotic bacteria colonize a specialized light organ inside the host and produce luminescence that is diffused from the animal body (Dunlap, 2009; Chakrabarty et al., 2011; Dunlap & Nakamura, 2011). All of the currently known light organ symbionts belong to the family Vibrionaceae (Gammaproteobacteria) with species belonging to the genera Aliivibrio and Photobacterium (Dunlap & Kita-Tsukamoto, 2006; Dunlap, 2009).

The genus Photobacterium includes three closely related species of light organ symbionts, that is, Photobacterium leiognathi, Photobacterium mandapamensis, and Photobacterium kishitanii (Boisvert et al., 1967; Reichelt & Baumann, 1973, 1975; Ast et al., 2007b; Urbanczyk et al., 2011). Bacteria from these three species have been studied in-depth, leading to a better understanding of their evolutionary relationship with nonsymbiotic species from this genus (Ast & Dunlap, 2005; Dunlap & Ast, 2005; Urbanczyk et al., 2011), diversity within each symbiotic species (Ast & Dunlap, 2004; Dunlap et al., 2004; Ast et al., 2007b), coevolution between host animals and symbionts (Dunlap et al., 2007; Kaeding et al., 2007), specificity of host–symbiont associations (Kaeding et al., 2007) and morphology of fish light organs colonized by the bacteria (Chakrabarty et al., 2011; Dunlap & Nakamura, 2011). The results of these studies allowed for a more detailed understanding of the evolution of the symbiotic Photobacterium and provided much of the current knowledge of fish bioluminescent symbionts ecology.

Certain strains of P. leiognathi were previously found to naturally contain multiple phylogenetically different copies of the lux-rib operon, which is responsible for the bacterial luminescent phenotype. The P. leiognathi lux-rib operon, luxCDABEG-ribEBHA, consists of genes that encode for bacterial luciferase and other enzymes involved in light production, as well as genes coding for riboflavin synthesis (Meighen, 1991; Lin et al., 2001; Ast et al., 2007a; Dunlap, 2009). Merodiploid strains contain two or more copies of the operon – one vertically inherited copy of lux-rib1 and one or more copies of the horizontally transferred operon lux-rib2 (Ast et al., 2007a; Urbanczyk et al., 2008). Discovery of the merodiploidy was made after multiple chromatogram peaks were found on sequencing chromatograms of luxA PCR amplicons of some P. leiognathi strains (Ast et al., 2007a). Cloning of the luxA PCR amplicons revealed that individual clones carry two different luxA sequences. Later, Ast et al. (2007a) showed that some strains of P. leiognathi carry two different copies of the lux-rib operon and developed a robust, PCR-based method to determine lux-rib copy number in the bacteria. Ast et al. (2007a) also found evidence that the lux-rib merodiploidy develops after horizontal transfer of the complete lux-rib2 operon via a mobile genetic element. Later studies showed that the lux-rib2 operon is frequently transferred between P. leiognathi strains, but horizontal transfer of the operon to other Vibrionaceae species is rare (Urbanczyk et al., 2008).

Ast et al. (2007a) also noted that P. leiognathi strains carrying multiple lux-rib operons were only found in a specific geographical area near the Honshu Island in northern Japan. Analysis of the bioluminescent symbionts of fish species distributed in northern Japan showed that these fish formed a symbiotic relationship with either merodiploid or single lux-rib bearing P. leiognathi strains (Ast et al., 2007a). In contrast, fish species distributed south of Honshu Island in Japan were colonized exclusively by P. leiognathi strains with a single lux-rib operon. Similarly, analysis of P. leiognathi strains isolated directly from the seawater detected merodiploid strains only in samples from northern Japan, while all the seawater isolates from the western Pacific had only a single lux-rib operon. Ast et al. (2007a) hypothesized that the geographic distribution of merodiploid P. leiognathi could be limited by physical barriers, which would prevent the movement of strains bearing multiple lux-rib operons to the western Pacific. For example, the northward flow of the Kuroshio Current (Fig. 1; Tomczak & Godfrey, 2003) could prevent or delay the southern movement of merodiploid P. leiognathi. An alternative hypothesis is that bioluminescent fish found in Okinawa, Taiwan, the Philippines and other regions of the western Pacific are preferably colonized by P. leiognathi strains bearing a single lux-rib operon. This preferential colonization of some fish species by single lux-rib bearing strains would limit the distribution of merodiploid strains to northern Japan, where they encounter fish species that can form symbiosis with merodiploid P. leiognathi. Differentiating between these two hypotheses is of importance for our understanding of luminous bacteria ecology and will help us to understand the mechanism underlying the generation and maintenance of luminous bacteria diversity.

image

Figure 1. Map of the western Pacific showing the distribution of merodiploid and single lux-rib bearing Photobacterium leiognathi. The scale bar is approximately 500 km. Enlarged map of the islands of Japan is shown in the inset. The black arrow shows the approximate flow of the Kuroshio Current. The numbers shown in the circles indicate the number of strains with a single lux-rib operon (white background) or indicate merodiploid strains (gray background). The data include results from the current study and additional information from Ast et al. (2007a). Geographic areas: (A) Fukui prefecture, Honshu, Japan; (B) Shimane prefecture, Honshu, Japan; (C) Kanagawa prefecture, Honshu, Japan (D); Shizuoka prefecture, Honshu, Japan; (E) Mie prefecture, Honshu, Japan; (F) Kochi prefecture, Shikoku, Japan; (G) Miyazaki prefecture, Kyushu, Japan; (H) Secutor indicius symbionts, Kagoshima prefecture, Kyushu, Japan; (I) Other P. leiognathi isolates from Kagoshima prefecture, Kyushu, Japan; (J) Ryukyu Islands, Japan; (K) Taiwan; (L) Luzon, the Philippine Islands; (M) Panay, the Philippine Islands; (N) Palawan, the Philippine Islands; (O) Gulf of Thailand.

Download figure to PowerPoint

In this study, we investigated whether the limited geographic distribution of merodiploid P. leiognathi can be attributed to physical barriers or whether this pattern is observed because fish typically found in the western Pacific preferentially form symbiosis with strains bearing a single lux-rib operon. Differentiating between the two hypotheses was possible after discovering that Secutor indicius, a bioluminescent fish species typically found in southern China and Taiwan, had recently expanded its distribution range to Kyushu Island in Japan (Kimura et al., 2005). Since 2008, the fish was frequently observed in the Kyushu region, especially in the Miyazaki and Kagoshima prefectures (Y. Iwatsuki, pers. obs.). So far, S. indicius was not found near Ryukyu Islands (Y. Sakurai, Okinawa Environmental Research Co. Ltd., pers. commun.); therefore, S. indicius found near Kyushu Island most likely moved to the area from Taiwan. We hypothesized that if the S. indicius specimens obtained from Japan contained merodiploid P. leiognathi in their light organs, it would prove that S. indicius, and presumably other fish species found in the western Pacific, could form symbiosis with merodiploid symbionts. Therefore, absence of merodiploid P. leiognathi in the western Pacific could be explained by physical barriers in the environment, which prevent the southward movement of the bacteria. However, if the S. indicius specimens from Japan formed symbiosis only with strains bearing a single lux-rib operon, even though merodiploid strains are present in the environment, it would indicate that the fish is unable to form symbiosis with the merodiploid P. leiognathi. As S. indicius is primarily found in the western Pacific, we could infer that fish in the region are unable to form symbiosis with the merodiploid symbionts; therefore, we could conclude that absence of merodiploid P. leiognathi in the western Pacific was caused by the merodiploid strains inability to form symbiosis with fish typically found in the region.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Collection of fish specimens and isolation of bacterial strains

Fish were caught from shallow coastal waters. The fish were handled in accordance with the animal handling regulations of the University of Miyazaki and the Shimane Prefectural Fisheries Technology Station. After the fish were caught, they were kept on ice until taxonomic identification and light organ dissection. Fish identification follows Sparks et al. (2005) and Kimura et al. (2003). The light organs were dissected according to the procedure described by Kaeding et al. (2007). From each fish specimen, ten individual light organ symbionts were picked at random for use in the study. The list of fish specimens, geographical areas where the fish specimens were obtained, and the strains designations for the isolated bacteria are listed in Supporting information, Table S1. The bacteria were isolated from seawater according to the procedure described by Urbanczyk et al. (2008). Screening for bioluminescent bacteria in seawater samples usually results in the isolation of various luminous Vibrionaceae species. Photobacterium leiognathi were distinguished from other bacterial species using sequences of the luxA gene.

Bacterial culture conditions and DNA isolation

The bacterial strains used in the study are listed in Table S1. The bacteria were grown in LSW-70 broth, which contained (per liter) 10 g tryptone, 5 g yeast extract, 350 mL double-strength artificial seawater, 650 mL deionized water, and 1.5 g agar for solid medium (Dunlap & Kita-Tsukamoto, 2006). For isolating bacteria from seawater, 4 g of agar was used per liter. Bacterial genomic DNA was purified from LSW-70 cultures grown overnight using a DNeasy blood and tissue kit (Qiagen), following the manufacturer's protocol for Gram-negative bacteria.

Amplification and sequencing of lux genes

PCR amplification was performed using the Promega GoTaq Green Master Mix and the following protocol: Initial denaturation was performed at 95 °C for 2 min. This was followed by 30 cycles of denaturation at 95 °C for 1 min; annealing for 30 s at temperatures between 45 and 50 °C, depending on the primer; followed by an extension step at 72 °C for 1 min 30 s or 1 min 45 s, depending on the expected product length. The 30 cycles were followed by a 5 min final extension step at 72 °C and a snap cooling to 4 °C. We used CWLAPlfor (GTTTTAGATCAACTGTCTAAAGGRCG) and CWLAPlrev (TCAGAACCATTCGCTTCAAATCCAAC; Ast et al., 2007a) as general luxA primers that could amplify any type of P. leiognathi luxA. To test for the vertically inherited lux-rib1 operon, we used the specific primers luxAforsec#3 (ATCACTTATTTAGCTGAACGTGGTTTA) and luxBrev#2 (TGCTTCGAATTGTTGCTGTTGAGAGC; Ast et al., 2007a). To test for the horizontally transferred lux-rib2 operon, we used the luxAforprim#3 (ATCATTTATCTAGCTAAACGTGGATAT) and luxBrev#2 primers (Ast et al., 2007a). This primer pair infrequently also amplified the sequence of the vertically inherited lux-rib1 operon. In select strains, we amplified additional fragment of the lux-rib operon, namely complete sequence of the luxB gene and a partial sequence of the luxE gene. A fragment of the luxB sequence and portion of a luxE were amplified with the luxBforsec (TTCTTTAAATCAAGTGATTACCACC) and luxErev1 (TAATATCATCTATTTCTGTACTCAC) primers or luxBforprim (CTCTCTGAATCAGGTAATTACAACT) and luxErev1 primers (Ast et al., 2007a). For amplifying the remaining luxE fragment, we used the luxBforsecNO.2 (AATATTGACCACCAATTCCCACTGCTC) and luxErevNo.2 (AAATAAAATACGGAGGGCCAATTAAAC) primers or the luxBforprimNO.2 (AATGTTGATCATCAATTCCCGCTACTG) and luxErevNo.2 primers (Ast et al., 2007a). The two seawater isolates used in the study were initially identified as P. leiognathi by amplifying and sequencing a fragment of luxA using the CWLAf (CTACTGGATCAAATGTCAAAAGGACG) and CWLAr (TCAGAACCGTTTGCTTCAAAACC) primers (Wimpee et al., 1999). PCR amplified DNA was sequenced using an ABI Prism 3100 Genetic Analyzer (Applied Biosystems) with an ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems). Nucleotide sequences determined in the study were deposited in the GenBank (see Table S1 for the accession numbers).

Phylogenetic analysis

luxABE sequences were manually aligned by inferred amino acid sequence. Phylogeny was reconstructed by parsimony analysis using paup* (Swofford, 2003), using branch and bound algorithm. Jackknife support was calculated with paup* using 34% deletion and 1000 replicates, emulating Jac resampling. The tree was visualized using FigTree v.1.3.1. Additional sequences of luxABE genes were obtained from GenBank (see Table S1 for the accession numbers).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Isolation of P. leiognathi strains

In previous studies, merodiploid P. leiognathi strains were found only in the Honshu Island area, whereas strains bearing a single lux-rib operon were found throughout the geographic sampling area (Fig. 1). To determine limits of the geographic distribution of P. leiognathi strains bearing a mobile genetic element, additional strains were isolated from fish and seawater samples obtained from Shimane prefecture (western coast of Honshu Island), Miyazaki prefecture (eastern coast of Kyushu Island), and Kagoshima prefecture (southern part of Kyushu Island). Thirty strains of P. leiognathi were isolated from the light organs of three specimens of Nuchequula nuchalis obtained from the coast of the Shimane prefecture. A total of 71 strains of P. leiognathi were obtained from the coast of Kagoshima prefecture, including 40 strains isolated from the light organs of four S. indicius specimens, 30 strains from the light organs of three Equulites rivulatus specimens, and a single isolate from seawater. A single P. leiognathi strain was also isolated from seawater sample obtained in the coastal region of Miyazaki prefecture.

Analysis of the lux-rib copy number

To determine if the newly isolated P. leiognathi strains contained single or multiple lux-rib sequences, a method similar to that previously described by Ast et al. (2007a) was used. Initially, partial sequences of luxA were amplified using the general primers CWLAPlfor and CWLAPlrev, which can amplify any of the known luxA genes of P. leiognathi (see 'Materials and methods' for details). If the analyzed strain contains multiple lux-rib operons with different nucleotide sequences, sequencing of the PCR product usually results in multiple chromatogram peaks (Ast et al., 2007a; see Fig. S1 for example of double chromatogram peaks). In contrast, the absence of multiple chromatogram peaks is a strong indicator that the strain contains only a single lux-rib operon. Strains that did not produce multiple chromatogram peaks when general luxA primers were used were additionally tested using specific primers that preferentially amplify the luxA gene from either a vertically inherited (lux-rib1) or horizontally transferred (lux-rib2) operon. Strains isolated from the light organ of the same fish that had the same sequence of luxA (without visible multiple peaks) were treated as a single strain type, and only one isolate was used for further testing. In this study, strains were classified as bearing a single lux-rib operon only if all of the PCR products obtained using general and specific luxA primers had the same sequence and if none of the chromatograms had any multiple peaks. We noted that the light organ symbionts of E. rivulatus specimen 9 did not have visible multiple peaks when using the CWLAPlfor and CWLAPlrev primers, but a double peak was observed later on sequencing luxA amplified using the luxAforprim#3 and luxBrev#2 primers. The same double peak was visible in all the strains isolated from E. rivulatus specimen 9. Therefore, all the ten strains from E. rivulatus specimen 9 were classified as merodiploid.

In this study, the lux-rib copy number of 102 P. leiognathi isolates was tested. Forty-one strains were found to be merodiploid, and 61 had a single lux-rib operon (see Table S1 for a complete list of the bacterial strain). Twenty strains from the light organs of N. nuchalis specimens obtained from Shimane prefecture were merodiploid (strains 170910FA1 to 170910FA10 and 170910FC1 to 170910FC10), and ten strains were found to have a single lux-rib operon (strains 170910FB1 to 170910FB10). All ten strains of N. nuchalis specimen B have the same sequence of the luxA gene, indicating that they represent the same strain type. All the 40 symbionts isolated from S. indicius caught in Kagoshima prefecture had a single lux-rib operon (strains 220710F2A to 220710F2J, 220710F3A to 220710F3J, 220710F4A to 220710F4J, and 220710F5A to 220710F5J). We also analyzed the diversity of symbionts within each light organ. Light organs of S. indicus specimen 5 had two strain types, that is, strains 220710F5F and 220710F5G different luxA sequence from the remaining eight symbionts of the fish. All strains isolated from the S. indicius specimen 5 had a single lux-rib operon. S. indicius specimens 2, 3, and 4 each had a single strain type in their light organ. Analysis of the light organ symbionts of E. rivulatus specimens obtained from Kagoshima prefecture showed that 10 strains from specimen 10 had a single lux-rib operon (strains 220710F10A to 220710F10J, two strain types were found in the light organs of specimen 10), and 20 strains were found to be merodiploid (strains 220710F8A to 220710F8J, and 220710F9A to 220710F9J). The seawater isolate from Kagoshima prefecture had a single lux-rib operon (strain 130710G), whereas the seawater isolate from Miyazaki prefecture had multiple copies of the lux-rib genes (140711A).

The results of the current analysis of 102 strains were combined with those reported by Ast et al. (2007a) about the lux-rib copy number in 171 P. leiognathi strains (Fig. 1) and showed that merodiploid P. leiognathi strains were only found in Japan, near Honshu and Kyushu islands.

Phylogenetic diversity of S. indicius symbionts

The results of the study showed that all of the S. indicius symbionts contained a single copy of the lux-rib operon. This was in contrast to the symbionts of E. rivulatus collected from the same geographic area and at the same time, two-thirds of which were merodiploid. The absence of merodiploid symbionts in the S. indicius light organs – despite the presence of merodiploid strains in the environment – suggests preferential colonization of the fish light organs by strains bearing only a single lux-rib operon. However, two alternative explanations for the absence of merodiploid P. leiognathi in S. indicius light organs were also tested. One alternative explanation could be the host behavior during symbiosis inception. If S. indicius larvae would preferentially form symbiosis with bacteria released by the adult fish and if there was no secondary colonization of the fish light organs, it would reduce the likelihood of S. indicius forming symbiosis with bacteria from the Kyushu region. Instead, the fish would preferentially form symbiosis with P. leiognathi strains brought to the Kyushu region by S. indicius individuals migrating from the south. Furthermore, the absence of merodiploid strains in S. indicius light organs could be explained if all the fish specimens sampled in this study acquired their symbionts south of the Kyushu Island where only single lux-rib bearing P. leiognathi are available, and only later moved to the Kyushu region. To test these two alternative scenarios, phylogenetic analysis of luxABE genes was performed using gene sequences from select S. indicius light organ symbionts and other P. leiognathi strains (Fig. 2). If any of the two alternative scenarios would be true, lower diversity of all examined S. indicius symbionts would be expected. Conversely, the resulting tree provides strong support for the diversification of S. indicius symbiontss, with the luxABE genes placing S. indicius symbionts in clades with P. leiognathi strains isolated from various geographic locations, including strains isolated near Kyushu Island. In particular, strain 220710F3A shows close phylogenetic relationship to two strains, namely 220710F10A and 220710F10B, isolated from the light organs of E. rivulatus obtained from the Kyushu Island region. This result implies that S. indicius forms symbiosis with P. leiognathi strains that are present in the Kyushu Island area. Furthermore, the diversity of S. indicius symbionts as well as presence of different strain types in the light organs of S. indicisu specimen 5 indicates that the fish behavior during symbiosis inception does not prevent the colonization of its light organ by the merodiploid P. leiognathi. Instead, the absence of merodiploid strains in the S. indicius light organs should be attributed to preferential colonization of the fish by single lux-rib bearing strains.

image

Figure 2. Phylogenetic hypothesis of the relationship between luminous bacteria based on the sequences of the luxA,luxB, and luxE genes. The noncoding spacer regions of all the taxa and the luxF sequence of some Photobacterium species were excluded. The inferred gaps were treated as informative data. The total number of aligned nucleotide positions in the data set is 2334, including 905 parsimony informative characters. One of eight equally parsimonious trees (length, 2218; consistency index, 0.7317; retention index, 0.8142). The jackknife resampling values are shown at the nodes; some have been omitted for clarity. The representative strains isolated from Secutor indicius light organs are shown in bold. Strains 220710F5A and 220710F5F were isolated from a light organ of the same fish specimen.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Here, we present the results of a study designed to test whether limited geographic distribution of merodiploid P. leiognathi strains can be attributed to physical barriers in the environment or whether it can be explained by preferential colonization of some fish species by P. leiognathi strains bearing only a single copy of the lux-rib operon. If the physical barriers in the environment limit the geographic distribution of merodiploid P. leiognathi strains and if there is no preferential colonization of some fish species by bacteria with a single lux-rib operon, then all the fish in the Kyushu Island region should be able to form symbiosis with merodiploid strains. However, we did not find any merodiploid strains among forty P. leiognathi symbionts isolated from the light organs of four S. indicius specimens caught near Kyushu Island. In contrast, analysis of only thirty symbionts from the light organs of E. rivulatus, caught at the same time and in the same geographic area, found twenty strains that had multiple copies of the lux-rib genes. Similarly, twenty merodiploid strains were found after analyzing as few as thirty symbionts of N. nuchalis. This striking difference in the lux-rib copy number between S. indicius symbionts and the bacteria isolated from other fish species allows us to confidently predict that the fish species does not, or very rarely form symbiosis with merodiploid P. leiognathi strains. Furthermore, we found two different strain types in the light organ of S. indicius specimen 5, all of which had a single copy of the lux-rib operon. This result indicates that even in case of multiple independent symbiont acquisition events, only bacteria with a single lux-rib operon colonize S. indicius. The absence of merodiploid P. leiognathi strains in S. indicius light organs, even though the strains were present in the environment, suggests that the fish preferentially formed symbiosis with strains bearing a single lux-rib operon. As S. indicius is usually found in the western Pacific region, and has only recently expanded its geographic distribution to Kyushu Island (Kimura et al., 2005), and considering that merodiploid strains were never obtained south of Kyushu Island, we predict that other fish species found in the western Pacific also do not form symbiosis with merodiploid P. leiognathi. Therefore, the limited geographic distribution of merodiploid P. leiognathi can be attributed to fish species typically found south of the Honshu and Kyushu islands preferentially forming symbiosis with P. leiognathi bearing a single lux-rib operon.

We investigated whether the absence of symbiotic merodiploid strains in S. indicius could be caused by fish behavior at the early stages of light organ colonization or by symbiosis inception in the area south of Kyushu Island where only strains bearing a single lux-rib operon are present. Very little is known about the early stages of P. leiognathi-fish symbioses, and it is possible that the aposymbiotic larvae of S. indicius are more likely to form symbiosis with bacteria released by adult fish (Dunlap et al., 2008; Dunlap & Nakamura, 2011). It is also not clear whether bioluminescent symbioses between fish and P. leiognathi allow for secondary light organ colonization in the adult fish (Dunlap et al., 2011). Therefore, host behavior may result in S. indicius forming symbiosis with the same type of bacteria as those found in the western Pacific, instead of with the bacteria found in the Kyushu region. If S. indicius larvae obtained their symbionts from the adult fish, it would result in formation of a P. leiognathi subpopulation with very low diversity. However, phylogenetic analysis performed using the sequences of luxABE genes showed that S. indicius symbionts have high phylogenetic diversity (Fig. 2), which indicates that S. indicius larvae are unlikely to form symbiosis only with bacteria released by adult fish. In addition, presence of multiple strain type in S. indicius specimen 5 light organ suggests multiple independent acquisitions of symbionts by the fish also indicating that S. indicius can be repeatedly colonized by different types of symbionts. Furthermore, little is known about the migration patterns of S. indicius and other ponyfish species. It was therefore possible that the light organs of the four S. indicius specimens analyzed in this study were colonized south of Kyushu Island, a region without merodiploid P. leiognathi, before the fish moved north. However, the results presented in Fig. 2 show that some S. indicius symbionts form a well-supported clade with the other bacteria obtained from the Kyushu region, which suggests that S. indicius light organs can be colonized in the Kyushu Island area. Therefore, the results of the phylogenetic analysis indicate that the absence of certain P. leiognathi strains in S. indicius light organs should not be attributed to fish behavior or the geographic area of symbiosis inception.

The preferential colonization of some fish species by P. leiognathi bearing a single lux-rib operon appears to have created a specific population subdivision (P. leiognathi containing the lux-rib2 operon on a mobile genetic element), which is found exclusively in northern Japan. Apparently, immigrant strains with the mobile genetic element that disperse south of Kyushu Island are unable to persist and establish in the environment, because they cannot colonize and reproduce inside light organs of fish found in the western Pacific. This result shows that symbiosis with marine animals is very important for the survival and reproductive success of P. leiognathi. Even though P. leiognathi is a facultative symbiont and can survive outside the host, it appears to have a limited ability for growing and establishing a population without forming symbiosis with an animal host. The light organs provide an excellent environment for the survival and reproduction of the symbiont, allowing for a very high density of bacterial cells (Dunlap, 1984), and P. leiognathi seems to be better adapted to the light organ environment than to the environment outside the host.

Even though each species of facultative bioluminescent symbiont colonize several species of host animals (Dunlap et al., 2007; Kaeding et al., 2007), results of the current study as well as work by Mandel et al. (2009) suggest that within each bioluminescent symbiont species, only a portion of the global population will form symbiosis with particular host species. These results appear to contradict environmental congruence hypothesis (Dunlap et al., 2007), which proposed that environmental congruence of host and symbiont dictates which bioluminescent bacterial species will colonize a specific host species, and suggested that bioluminescent bacteria are primarily opportunistic colonizers of animal light organs. However, our results suggest a more complex scenario, which involves selection of the bacterial symbiont in addition to environmental congruence between the symbiont and host. Considering that presence of a specific mobile genetic element in P. leiognathi limited host species for the bacterial symbiont, the selection could be against presence of a specific bacterial genotype, rather than for presence of specific symbiosis-related genes. Bacterial adaptations to bioluminescent symbioses could therefore involve loss of specific genes, rather than acquisition of additional symbiosis-related genes.

The function of the horizontally transferred lux-rib genes that could be important for colonizing fish found south of Kyushu Island is currently not known. Merodiploidy of P. leiognathi does not result in visibly brighter luminescence or any other detectable phenotypic difference from strains bearing a single lux-rib operon (Ast et al., 2007a). However, even if the horizontally transferred lux-rib2 operons do not affect the P. leiognathi luminescent phenotype, it could have other functions during associations with host fish. For example, the lux genes play a role during Aliivibrio salmonicida infection of fish (Nelson et al., 2007), even though A. salmonicida strains bearing lux genes do not have a luminescent phenotype. Nelson et al. (2007) hypothesized that during host fish colonization, the cryptic lux operon of A. salmonicida may provide oxidized flavin at low oxygen tension (Bourgois et al., 2001), which would aid during the pathogenic colonization of some fish. Similarly, the lux-rib2 genes of P. leiognathi could play a role unrelated to the luminescent phenotype during light organ symbiosis. The horizontally transferred lux-rib2 operon in P. leiognathi could aid in the symbiosis with some fish species, but the additional luciferase activity could prevent the merodiploid strains from colonizing the light organs of fish species found in the western Pacific. Furthermore, luminescence levels of symbiotic bacteria depend on osmolarity of the host tissue (Dunlap, 1985; Stabb et al., 2004), but the effects of osmolarity on the symbiont luminescence vary considerably between different species of symbiotic bacteria. Considering that different host species of P. leiognathi likely differ in their tissue osmolarity levels, it is possible that different lux-rib operons of P. leiognathi are preferable during symbiosis with specific fish species. Another possibility is that the expression of the horizontally transferred lux-rib2 operon is differently regulated than that of the vertically inherited lux-rib1 genes, and changes in gene regulation could affect the ability of P. leiognathi to colonize light organs of a specific fish species. Mandel et al. (2009) have shown that the presence or absence of a single regulatory gene can change the host range of bioluminescent symbionts; therefore, it is not unreasonable to think that differential regulation of lux-rib gene expression could affect the P. leiognathi host range. Similarly, Bose et al. (2011) have shown that a highly divergent region between A. fischeri luxR and luxI genes, which are located upstream of the bacteria lux operon and regulate the luminescence genes expression, contributes significantly to A. fischeri luminescence output. Results of that study also suggested that specific levels of luminescence are selected for in squid associated A. fischeri (Bose et al., 2011). Considering that sequences upstream the lux-rib1 and lux-rib2 operons of P. leiognathi are also highly divergent (Ast et al., 2007a, 2007b), it is possible that expression of the two operons result in different levels of luminescence, some of which could be advantageous during symbiosis with specific species of fish. Much additional biochemical and molecular genetic work will be required to test these hypotheses and to determine the possible role of horizontally transferred lux-rib2 genes in light organ symbiosis.

At present, we cannot exclude the possibility that the absence of merodiploid strains in S. indicius light organs is because of genes other than the horizontally transferred lux-rib2 operon. The exact structure of the mobile genetic element in all analyzed merodiploid strains is not yet known; therefore, we cannot rule out the possibility that genes other than lux-rib2 from the mobile genetic element are important for P. leiognathi symbiosis with some fish species. However, we find this possibility unlikely, because the immediate surrounding of the mobile lux-rib2 operon was reported to differ significantly between three merodiploid strains (Ast et al., 2007a), which suggest high diversity in the genetic composition of the mobile genetic element. Furthermore, analysis of large fragments of the mobile genetic element found in a merodiploid strain suggests extensive rearrangements of the element involving insertion elements (H. Urbanczyk & T. Furukawa, unpublished data). Also, the mobile genetic element is frequently transferred between P. leiognathi strains (Ast et al., 2007a; Urbanczyk et al., 2008) and likely each analyzed merodiploid strain obtained their mobile genetic element from a different source, in an independent horizontal transfer event. Numerous transfers of the mobile genetic element between P. leiognathi strains combined with frequent rearrangements of the element likely result in different gene content of the mobile genetic element in all merodiploid strains analyzed in this study and make it unlikely that genes of the mobile genetic element other than the lux-rib2 genes are important for the colonization of certain species of fish.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank H. Semura and T. Miura (Shimane Prefectural Fisheries Technology Station) for the gift of N. nuchalis specimens. This work was supported by the Sasakawa Scientific Research Grant from the Japanese Science Society; by the Program to Disseminate Tenure Tracking System from the Japanese Ministry of Education, Culture, Sports, Science and Technology; as well as by a grant for Scientific Research on Priority Areas from the University of Miyazaki.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
fem1353-sup-0001-FigureS1.pptapplication/306KFig. S1. Example of sequencing chromatogram double peaks characteristic for merodiploid P. leiognathi.
fem1353-sup-0002-TableS1.xlsapplication/msexcel49KTable S1. Bacteria collected in this study.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.