Violet bioluminescent Polycirrus sp. (Annelida: Terebelliformia) discovered in the shallow coastal waters of the Noto Peninsula in Japan

Terebellidae worms have large numbers of tentacles responsible for various biological functions. Some Terebellidae worms whose tentacles emit light are found around the world, including exceptional violet-light-emitting Polycirrus spp. found in Europe and North America. However, there is no video-recorded observation of the luminous behavior of such unique species in nature, and the genetic information related to their ecology are lacking. Here, for the first time, we video-recorded the violet-light-emitting behavior of an undescribed Japanese worm in its natural habitat. The worm was designated as Polycirrus sp. ISK based on morphological observations, and the luminescence spectrum showed a peak at 444 nm, which is an exceptionally short wavelength for bioluminescence in a shallow coastal water environment. An analysis of differentially expressing genes based on separate RNA-Seq analysis for the tentacles and the rest of body revealed the specific expression of genes that are probably involved in innate immunity in the tentacles exposed to predators. We also found a Renilla luciferase homologous gene, but coelenterazine was not detected in the worm extract by analyses using a liquid chromatography and a recombinant Renilla luciferase. These results will promote an understanding of the ecology and luminescence mechanisms of luminous Polycirrus spp.

www.nature.com/scientificreports/ Thelepus spp. revealed an antimicrobial compound localized in their tentacles 14 . Recently, the Japanese Thelepus japonicas, which emits light at λ max 508 nm, was studied with a focus on the molecular mechanism underlying light emission 15 , while the molecular bases of luminous Polycirrus spp. remain to be clarified. Luminous Polycirrus spp. have long been known in various places around the world 16,17 , including Polycirrus auranticus from the coast of England 9 and Polycirrus preplexus from California 17 . The former species is reported to show a rather weak violet-blue light flashing out at the tips of tentacles when the worm is disturbed 9 . The latter is reported to be a nonsecretion flash-type light emission, and an analysis using a charge-coupled device (CCD) spectrophotometer revealed that the emitting light had a 445 nm emission peak 17 . However, almost all descriptions of luminous Polycirrus spp. in the literature are more than 30 years old and lack clear photographs or videos that would suggest these species' ecological behaviors.
In this study, for the first time, we video-recorded the violet-light-emitting behavior of an undescribed worm in the shallow coastal waters of the Noto Peninsula, Ishikawa, Japan. The worm was morphologically identified and named Polycirrus sp. ISK. In addition, we successfully collected the light-emission spectrum with a peak at 444 nm, which was very similar to that of P. preplexus found in California. Our RNA-Seq analysis showed that the existence of a gene coding for fucolectin, which is a fucose-binding lectin related to an innate immunity response, was significantly enriched in the tentacles. The RNA-Seq data included a homologous gene of Renilla luciferase, which is the enzyme responsible for coelenterazine-dependent bioluminescence, but coelenterazine was not detected in the worm extract by analyses using ultra performance liquid chromatography (UPLC) and a recombinant Renilla luciferase.

Results and discussion
Morphology and light-emitting behavior of the undescribed Japanese Terebellidae worm. In 2016, some of the present authors were exploring shallow coastal waters (depth less than 1 m) to observe the ecological behaviors of marine animals in the Noto Peninsula, when they discovered unknown violet-light-emitting worms. At the sampling point, the worms were living in small holes (a few centimeters in diameter) or in cracks in rocks covered by sand at the shallow sea bottom ( Supplementary Fig. S1). We successfully video-recorded their emission of violet light from the whole tentacle stretching into sea water when stimulated by air bubbling at night (Fig. 1A-C; Supplementary Videos S1 and S2). The violet-light emission consisted of rapid flashes with variable duration in the order of milliseconds (Supplementary Video S3), as observed for the worm P. perplexus in response to stimulation 17 . From our morphological observation, we identified the violet-light-emitting worm The worm with light emission at the tentacles. This worm was stimulated by an electric shock. Scale bars = 100 mm for A and B, 10 mm for C and D. Each photograph was extracted from the videos recorded with the following settings: sensitivity, ISO 51200 or 11 lx; white balance, 4300 K or 5800 K; shutter speed, 1/30 s or 1/60 s; iris, F1.8-3.5; frame rate, 29 Fig. 1D; Supplementary Video S3), and the luminescence spectrum showed that its λ max was 444 nm or slightly longer, depending on the individual ( Fig. 2A). We also found that light emission was efficiently induced by the addition of KCl solution and observed the time course of light emission with rapid fluctuations with variable duration in the order of milliseconds for up to 30 s (Fig. 2B). The flash pattern was similar to that observed in a study of P. perplexus 17 . In the genus Polycirrus, P. medius and P. nervosus in Japan have been described 18,19 . However, the morphological features of the species in the present study differed from these species on the basis of our observations described above. Japanese Polycirrus spp. have not been described as luminous worms according to our review of the literature and web pages. In addition, the number of reports for new Polycirrus spp. from all over the world has been increasing, but a limited number of species are known to emit light 13,17,18 . Our finding of KCl-induced light emission from Polycirrus sp. ISK suggested that we can easily test the light-emitting ability of Polycirrus spp. by luminescence measurement just after adding KCl solution. A spectrum pattern has been reported for only one species, P. perplexus collected in California 17 , and it would be necessary for further understanding of these species to examine the light-emitting abilities and to compare light-emitting behaviors and spectrum patterns. The color of bioluminescence is often related to habitat, and light in the blue range is typical for pelagic species 20 . Thus, one of the points to be focused on is the ecological function of the violet-light emission of this worm inhabiting in a shallow coastal water environment. In P. perplexus, deterring predation is a possible function of luminescence based on that species' habitat and its violet-light emission 17,21 . As shown in Supplementary Videos S1, S2, which are the first video records of in situ light emission of a Polycirrus species, the air bubble-stimulated luminescence of Polycirrus sp. ISK in its natural habitat also seemed to deter predation, but this explanation is still speculative.

Differentially expressing genes between the tentacles and the rest of body. A few years after
discovering this worm, we found it difficult to collect enough of them to conduct common biochemical and chemical analyses because we did not find a place densely inhabited by hundreds of the worms whose wet weight was a few tens of milligrams (e.g. 16.5, 29.8, or 31.8 mg). Next, we conducted RNA-Seq analysis. In luminous animals with strong light emission, such as firefly or syllid polychaetes (Syllidae), luciferase expression is high especially at the luminous organ or in the whole body 22,23 . On the other hand, the light emission of Polycirrus sp. ISK was not so strong compared to that of fireflies, and the light-emitting area was limited to the tentacles. In addition, the genetic information related to the tentacles responsible for various ecological functions is still limited. Thus, in the present study we decided to purify RNA from the tentacles and the rest of body separately (Fig. 1C) and performed RNA-Seq analysis followed by a computational analysis using the MASER pipeline 24 . By de novo assembly, 110,775 contigs were predicted; 26.1% of them showed more than twice the expression level in the tentacles than in the rest of body, whereas 20.8% showed more than twice the expression in the rest of body than in the tentacles. When we performed a blastX search to the NCBI nr database for the contigs longer than 300 bp, 35.6% showed significant homology with registered genes with e-values of less than 1e −10 . The average length for these contigs was 1384 bp, and half of them were in the range of 463-1863 bp ( Supplementary Fig. S2). In the assembled sequence, we found the cytochrome oxidase subunit I (COI) gene and tried to construct a phy- www.nature.com/scientificreports/ logenetic tree. However, the obtained phylogenetic tree was unreliable due to the low bootstrap values as shown in Supplementary Fig. S3.
To focus on the tissue-specific genes, we first picked up genes with high expression levels based on high fpkm values (over 1000) and then ranked these genes based on the tissue-specificity judged by the comparison of fpkm values in tentacles and the rest of body. In tentacle-specific genes, we found that some genes coding for lectin(like) domains were ranked in the top eight as shown in Supplementary Table S1. Of the top eight genes in the rest of body-specific genes (Supplementary Table S2), seven exhibited no similarity to any genes, and the remaining gene exhibited significant similarity to a hypothetical protein of Capitella teleta, which is a Polychaetes species with whole-genome information available 25 . Recently, TPM is preferably used to normalize expression level, and the value is used for statistical differential expression analysis 26 , and we also calculated TPM for tissue-specific genes (Supplementary Table S3).
As we were unable to conduct statistical differential expression analysis due to no biological/technical replication resulted from difficulties in the sample collection, we simply compared TPM value between the tentacle and the rest of body samples. The ratio of TPM (tentacle/rest of body) was calculated, and then top 100 genes (Fig. 3A), which were highly expressing in the tentacle, were selected. Similarly, top 100 genes highly expressing in the rest of body were selected using the ratio of TPM (rest of body/tentacle) (Fig. 3B). These gene lists were annotated by gene ontology (GO) terms and analyzed using WEGO program 27 . WEGO results showed different expression patterns for the tentacle and the rest of body. In the tentacle, GO terms including cell adhesion, biological adhesion, small molecular binding, positive regulation of biological process, regulation of response to stimulus, carbohydrate binding, and immune response were significantly higher (Fig. 3C, D). In the rest of www.nature.com/scientificreports/ body, GO terms including hydrolase activity, catalytic activity, localization, and establishment of localization were significantly higher. In the top 100 genes highly expressing in the tentacle, we found 21 genes annotated as a gene coding for fucolectin by blast search (Supplementary Table S4). Fucolectin is a fucose-binding lectin involved in the innate immunity of diverse invertebrate species 28 . However, its function in invertebrates remains unclear, and no information is available for Terebellidae, including sequence information. Fucolectin was first identified in eel with mRNA distribution mainly in liver and gill 28 . In sea cucumber, expression of the fucolectin gene is confirmed in respiratory trees, muscle, and tentacle 29 . We were not able to see whether this gene was expressed in the respiratory organ of Polycirrus sp. ISK because a characteristic of the genus Polycirrus is the absence of branchiae 18 . Nevertheless, the tentacle-specific expression of fucolectin was consistent with the observation in sea cucumber, and the high expression of such proteins involved in innate immunity seemed reasonable because tentacles stretching out of their bodies can be damaged by attack of predators and thus are threatened by infectious bacteria and other pathogens 11 , as is the respiratory organ. In addition, localization of antimicrobial compounds in Terebellidae worms is suggested to be of antiseptic importance in damage by predation 14 . This study would provide indispensable information about the ecological meaning of Polycirrus sp. ISK's life in future genetic studies.
Transcripts coding for luciferase-like genes in the worm. To find genes similar to the known luciferase, which is an enzyme oxidizing a specific compound called luciferin to emit light, from related species in polychaetes, we performed a blastX analysis against the Odontosyllis luciferase sequence using our RNA-Seq data. We found a gene coding for a protein that exhibited similarity to Odontosyllis luciferase, but the e-value was more than 1e −10 ( Supplementary Fig. S4). In addition, the top hit for this gene analyzed by blastX was annotated to code an uncharacterized protein of Saccoglossus kowalevskii (Hemichordata), and its specific function was not predicted. Other hits were for genes from Chordata, Mollusca, and other phyla but there was no hit from Annelida. This result would suggest that the light-emission system of Polycirrus sp. ISK differs from that of the genus Odontosyllis, although further experiments using high purity Odontosyllis luciferase and the substrate will be necessary to confirm this. In further blastX analyses of representative luciferases, photoproteins, and a putative luciferase [luciferases from the ostracod Cypridina noctiluca (Accession number: BAD08210.1), the copepod Gaussia princeps (AAG54095.1), the deep-sea shrimp Oplophorus gracilirostris (BAB13775.1 and BAB13776.1), the firefly Photinus pyralis (AAA29795.1), the sea pansy Renilla reniformis (AAA29804.1); photoproteins from the hydrozoan jellyfish Aequorea victoria (AAA27720.1), the hydroid Clytia gregaria (CAA49754.1), the hydroid Obelia geniculate (AAL86372.1); a putative luciferase from the tunicate Pyrosoma atlanticum 30 sequences using our RNA-Seq data], we found some tissue-nonspecific genes whose sequences exhibited similarity to firefly luciferase (FLuc) or Renilla luciferase-like protein (RLuc-like) sequences with an e-values of less than 1e −10 and percent identity of more than 50%. FLuc is a member of the acyl-adenylate-forming superfamily of enzymes responsible for firefly luciferin-dependent bioluminescence, which is found in terrestrial luminous beetles emitting light ranging from green to red 31 . Previously, a putative acyl-CoA synthetase protein was found in the luminous organ of firefly squid emitting blue light 32 , but there is no clear biochemical evidence that such protein is responsible for firefly squid's bioluminescence. On the other hand, RLuc is responsible for coelenterazine-dependent bioluminescence, which is found in marine luminous organisms belonging to various taxa. An RLuc-like protein is found to be localized in luminous organs of the brittle star Amphiura filiformis, as revealed by taking advantage of the cross reactivity of anti-RLuc antibody to A. filiformis RLuc-like protein 33 . A recent study reported that recombinant RLuc-like protein found in P. atlanticum exhibited luciferase activity to coelenterazine 30 . However, an RLuc-like protein from sea urchin Strongylocentroutus purpuratus is confirmed to exhibit dehalogenase activity to various substrates but no luciferase activity to coelenterazine 34 . Therefore, it is suspected that Polycirrus sp. ISK possesses a luminescence system using an RLuc-like enzyme.

Coelenterazine content in the worm. To investigate whether Polycirrus sp. ISK possesses not only a
Renilla luciferase homologous gene but also coelenterazine, we analyzed an ethanolic extract of Polycirrus sp. ISK by UPLC with a UV-visible detector (Fig. 4). The obtained UPLC chromatogram did not show a peak corresponding to that of authentic coelenterazine. When further checking the chromatogram, we found the peak at a retention time similar to those of authentic coelenteramide and coelenteramine, which can be formed from coelenterazine. However, the absorption spectrum obtained by UPLC analysis and the mass spectrum obtained by MS/MS analysis were not identical to those of authentic coelenteramide or coelenteramine ( Fig. 4 and Supplementary Figs. S5 and S6). In addition, when the worm extract was mixed with a recombinant RLuc, we did not detect luminescence using a luminometer. These results suggested that the luminescence system in the worm was independent of coelenterazine, although a RLuc homologous gene was found. Similarly, the existence of an RLuc homologous gene was reported in P. atlanticum, which has been suggested to use a coelenterazine-independent luminescence system relying on bacterial bioluminescent symbionts 30,35 . We also mixed the worm extract with a recombinant cypridinid luciferase, but we did not detect luminescence using a luminometer. This result was consistent with Harvey's observation for P. caliendrum 16 . To examine whether the luminescence system is based on luciferin-luciferase reaction, which is found in various luminous animals including some syllid Odontosyllis spp. 23,36-39 , we prepared two different extracts of the worm using 100 mM HEPES-NaOH buffer (pH 7.4) and methanol, and subsequently subjected a mixture of the two to luminescent measurement. As a result, no light emission was detected from the mixture of the buffer and methanolic extracts of the worm. This result was also consistent with Harvey's observation for P. caliendrum 16 . However, there is still a possibility that the light emission is based on luciferin-luciferase reaction, because luciferin-luciferase reaction found in fireflies or luminous mushrooms requires a cofactor such as ATP or NADPH, and we did not test all possible conditions due to the limitation of the number of collected specimens. In addition, extraction of luciferin and luciferase in the active www.nature.com/scientificreports/ form is sometimes difficult, as shown in previous studies 37 . Further studies using hundreds or more of the specimens must be performed to elucidate the mechanism underlying the violet-light emission.  www.nature.com/scientificreports/ Equipment for photography and video recording. Photographs and videos were taken by a mirrorless camera (α7S; Sony, Tokyo, Japan) with a SEL24F18Z lens (Sony) and an underwater camera housing (Nauticam NA A7; Nauticam, Hong Kong, China).

Measurement of luminescence emission spectrum.
The luminescence emission spectrum of the worm Polycirrus sp. ISK stimulated by an electric shock was measured using a high-sensitivity charge-coupled device (CCD) spectrophotometer, LumiFLspectrocapture (AB-1850; ATTO, Tokyo, Japan) with the following settings: measurement mode, single; measurement time, 1 min; slit width, 0.25 mm; camera gain, high; diffraction grating, 150 lines/mm; and shutter for measurement, automatic. An anode and a cathode were put into the tube, and an electric pulse was generated using a 9 V battery. Analysis of the worm extract using UPLC and mass spectrometry. The whole body of a single specimen stored in approximately 500 μL of ethanol at − 80 °C was homogenized in the storage ethanol on ice with a plastic pestle and centrifuged at 15,000×g for 5 min at 4 °C, after which 2 μL of the supernatant was subjected to luminescence analysis. A portion of the rest supernatant was five times diluted with 10% (v/v) acetonitrile in water and filtered through a centrifugal filter Ultrafree-MC (0.22 μm; Millipore, Billerica, MA, USA). Then, 10 μL of the filtrate was subjected to UPLC analysis and separation. UPLC analysis and separation were performed on a Waters ACQUITY UPLC H-Class system (Waters, Milford, MA, USA) equipped with an ACQUITY UPLC C18 column (ϕ2.1 × 100 mm, 1.7 μm; Waters) and an ACQUITY UPLC PDA eλ detector (Waters). The UPLC conditions were as follows: mobile phase, a linear gradient of acetonitrile in water from 10 to 100% for 20 min; flow rate, 0.3 mL/min; UV detection, 333 nm or 435 nm. The fraction eluted at a retention time of 9.5-10.0 min (panel a in Fig. 4A) was 20 times diluted with 1% formic acid and subjected to MS/MS analysis using an LCMS-9030 quadrupole time-of-flight mass spectrometer (Shimadzu, Kyoto, Japan). The parameters for MS/MS analysis were as follows: interface temperature, 300 °C; desolvation temperature, 250 °C; interface voltage, 4.5 kV; polarity, positive; collision energy, 30 V. Under the present UPLC condition, the detection limits were as follows: 15 pmol for coelenterazine, 0.1 pmol for coelenteramide, and 5 pmol for coelenteramine.
Luminescence analysis of the worm extract. To 2 μL of the ethanolic extract of the worm in a white 96-well plate (Eppendorf microplate 96/F-PP; Eppendorf, Hamburg, Germany) was added 100 μL of a solution of a recombinant Renilla luciferase or cypridinid luciferase in 200 mM Tris-HCl (pH 7.5) containing 100 mM NaCl using the auto injector in the Phelios luminometer (ATTO), followed by the immediate measurement of luminescence intensity at room temperature. The concentrations of Renilla luciferase and cypridinid luciferase were sufficient to detect 100 fmol of coelenterazine and 100 amol of cypridinid luciferin, respectively.