Bioluminescence and Fluorescence in Littoral Earthworms

Pontodrilus litoralis is a cosmopolitan littoral earthworm known to exhibit bioluminescence. Recently, a congeneric species Pontodrilus longissimus from Thailand was described. These species are sympatric but their burrowing depths on Thai beaches are different. In this study, we examined the in vivo and in vitro bioluminescence properties of P. longissimus and P. litoralis. Mechanical stimulation induced in vivo luminescence in P. litoralis, as reported previously, but not in P. longissimus. In vitro cross-reaction tests between these species revealed the absence of luciferin and luciferase activities in P. longissimus. P. litoralis had strong uorescence in a coelomic uid that matches to the spectral maximum of its bioluminescence, but P. longissimus did not. These results suggest that P. longissimus does not have luminescence ability due to the lack of all bioluminescent components, luciferin, luciferase, and light emitter, despite its close relationship to the luminous P. litoralis. The presence of both luminous and non-luminous species in a single genus is uncommon, and our present ndings will shed insight on the possible functions of bioluminescence in the earthworm, such as avoiding predation by littoral earwigs.


Introduction
The earthworm genus Pontodrilus Perrier, 1874 displays various unique characteristics. The littoral earthworm P. litoralis (Grube, 1855) distributes in the tropical and sub-tropical coasts of the Atlantic, Indian and Paci c Oceans [1][2][3], and is known to be bioluminescent [4][5][6][7]. The luminescent system of P. litoralis has been shown to be a luciferin-luciferase type reaction triggered by hydrogen peroxide, with a uorescence compound acting as a light emitter [7], although the chemical structure of the luciferin remains uncertain and the luciferase gene has not been determined. Recently, the littoral earthworm Pontodrilus longissimus Seesamut & Panha, 2018 was described from the coastal areas in Thailand and Peninsular Malaysia [8] based on the morphological differences in the size of the body, the number of segments and the diverticulum from other congeners.
In the present study, the bioluminescence and uorescence of P. litoralis and P. longissimus were examined in vivo and in vitro, and the results suggested that P. longissimus lacks luminescence ability despite its genetically close relationship to P. litoralis. Based on these ndings, we discuss the biological function of earthworm bioluminescence and a convenient parataxonomy method for Pontodrilus species.

In vivo and in vitro bioluminescence
After the live specimens of both Pontodrilus species were stimulated by electricity or rough handling, P. litoralis exuded a green luminescent uid, whereas the uid exuded from P. longissimus was not luminescent (Fig. 1A). Under a handheld long-wave UV lamp (365 nm), almost the entire body of P. litoralis emitted strong yellow uorescence, which was most conspicuous at the rows of setae, whereas P. longissimus did not emit uorescence under the same condition (Fig. 1B).
The cross-reactivities of the crude luciferase and crude luciferin in P. litoralis and P. longissimus were examined ( Fig. 2A). For certainty of the results, we used a concentration of P. longissimus extract that was 5-fold higher than the concentration of P. litoralis extract. The results showed that signi cant luminescence was detected only when mixing the luciferin extract from P. litoralis with the luciferase extract from P. litoralis. On the other hand, both luciferin and luciferase activities in the extracts of P. longissimus were negative (the levels of both activities were almost the same as in a negative control). This nding suggested that P. longissimus is non-luminous due to the lack of both luciferin and luciferase.
Fluorescence and luminescence spectra of luminous P. litoralis Fluorescence spectra were measured using a crude coelomic uid extract of P. litoralis (Fig. 2B). The peaks of the excitation spectra were 370 and 453 nm, whereas the emission peaks were 450 and 523 nm, indicating the presence of at least two uorescence compounds in P. litoralis. The luminescence spectrum of P. litoralis had a maximum wavelength of 528 nm in vivo and 524 nm in vitro (Fig. 2C). Wampler and Jamieson showed that the spectral maximum of bioluminescence (540 nm) in the Bermudan P. bermudensis (which is now considered synonymous with P. litoralis [2]) matched the uorescence maximum of the coelomic uid, and suggested the uorescence substance is a light emitter [7]. Although the spectral maximum values of our present study were different from their results, probably due to genetic differences in the specimens examined or differences in the spectrophotometers used, our results also showed a spectral match between uorescence and bioluminescence in vitro. The small redshift of the in vivo spectrum might have been due to a re ection effect from the reddish earthworm body.

Comparison of the coelomic uid cells and protein bands between the two littoral earthworms
The coelomic cells of these littoral earthworm species were observed under a uorescence microscope ( Fig. 3). The results showed that the P. litoralis coelomic cells emitted uorescence but the P. longissimus did not. The size of the coelomic cells was approximately 15 µm in diameter, and numerous small uorescent particles were detected in the coelomic cells of P. litoralis. SDS-PAGE of the coelomic uids showed different protein constitutions between the two species (Fig. 4).

Discussion
In this study, we con rmed that P. longissimus is non-bioluminescent, despite its close relationship to the luminous P. litoralis. The presence of both luminous and non-luminous species in a single genus is uncommon; in general, bioluminescence is shared among all members of the same genus, sometimes at the family level. For example, the family Lampyridae ( re ies) consists of over 67 genera and 2,000 species around the world, and all are considered to be bioluminescent, at least in the larval stage, which uses the same luciferin molecule and homologous luciferase [9]. In contrast, we can list only a few exceptions, such as Vibrio and Photobacterium (marine bacteria) [10], Epigonus (deep-sea shes) [11] and Eisenia (terrestrial earthworms) [12]; these genera have been reported to contain both luminous and non-luminous species. P. litoralis and P. longissimus are easily collectible at the same beach [8] and rearable in a laboratory; thus they are suitable materials for studying the ecology and evolution of bioluminescence.
In vitro luciferin-luciferase cross-reaction tests of P. longissimus and P. litoralis con rmed that the lack of luminescence ability of P. longissimus is due to the absence of all bioluminescent components, i.e., luciferin, luciferase and the light emitter in coelomic uid. It has previously been suggested by crossreaction tests that the luminous earthworms in the genera Pontodrilus (Megascolecidae), Microscolex and Diplocardia (Acanthodrilidae) share the same basic bioluminescence mechanisms [5,7,13,14], in spite of their far-distant relationship to each other [15,16]. It is expected that the ancestral state of Pontodrilus is non-bioluminescent, because the nearest extant relatives of Pontodrilus belong to the genus Plutellus Perrier, 1873, and all members of this group are non-bioluminescent [6,17]. These ndings suggested that P. litoralis secondarily acquired the bioluminescent properties as a parallel evolution, similar to the case of bioluminescence in lampyrid and elaterid beetles [18]. We detected a clear difference in protein composition of the secreted uid between P. litoralis and P. longissimus. The luciferase and other bioluminescent components of luminous earthworms were not determined, and further comparative analyses between the proteins and substances of these secreted uids will be useful to understand the mechanism of bioluminescence and its parallel evolution.
In Thailand, P. longissimus was found sympatrically with P. litoralis at the beaches along the coast, but the microhabitats of the two congeners are different; P. litoralis was collected on the beach surface (under trash or leaf litter on sandy beaches), whereas P. longissimus was found at a greater depth than P. litoralis, i.e., a depth of more than 10 cm, where trash and leaves are scarce [8] (Fig. 5A-5D). It has been hypothesized that the biological function of bioluminescence in Annelida, including P. litoralis, is to stun or divert attention as an anti-predator defense [19][20][21][22][23][24][25], but experiments and observations of the prey are limited. Sivinski & Forrest [25] reported the luminescence of Microscolex phosphoreus deterred from predation of the mole cricket Scapteriscus acletus (but the specimen was nally consumed) under laboratory conditions. A British television program [26] presented by David Attenborough showed that the French luminous earthworm Avelona ligra glowed when attacked by the carabid beetle, but the beetle consumed the luminescent worm without any hesitation. We consider that the absence of bioluminescence in P. longissimus may correlate with the habitats with low-predation pressure, whereas P. litoralis acquired a bioluminescence property during evolution that enables it to enter the surface environment of the beach, which is rich in nutrition and food sources [3,27] as well as in potential predators.
Indeed, no possible predator was found to live alongside P. longissimus. In contrast, various carnivorous invertebrates, such as earwigs, robe beetles, carabid beetles, crabs, and hermit crabs, were found to live along with P. litoralis on the beaches in Thailand and Japan (Seesamut pers. obs.). In this context, we performed a feeding experiment using the maritime earwig sympatrically distributed in the P. litoralis habitat. The maritime earwig Anisolabis maritima (Dermaptera, Anisolabididae) was a cosmopolitan, also distributed in Japan. It has developed compound eyes (Fig. 5E) and is considered a carnivorous animal that forages its prey at night [28,29]. We found A. maritima (body length ≤ 30 mm) predominantly at the beach where P. litoralis was collected (Fig. 5F). Some robe beetles (Coleoptera, Staphylinidae) were also found at the same habitat, but they seemed to be too small (< 10 mm) for predation of P. litoralis, and under our laboratory observations, the robe beetle did not attack the worm. Thus, we think A. maritima is major potential predator of P. litoralis at the beach in Japan. Living P. litoralis and A. maritima were collected from the same beach on the same day, and we observed the predation behavior in the laboratory in a dark cage with beach sand spread on the bottom. Our observation of the predation of P. litoralis by the earwigs (Supplementary Video 1) may have provided an important insight into the function of bioluminescence in P. litoralis. The earwigs immediately started to aggressively attack the worm with their mandibles and abdominal cerci, a pair of scissors-like pincers; the worm secreted luminescent mucus from its wounds (Supplementary Video 1), and it appeared that the retention of the gluey luminescent mucus on the mouths and forelegs of the earwigs was unpleasant to them, since they struggled to remove the glue by frequent grooming (Fig. 5E, Supplementary Video 2). Indeed, after aggressive attacks, the earwigs nally abandoned their consumption of the worm, and thus the worm survived. To the best of our knowledge, this is the rst observation of earthworm bioluminescence by predation under almost natural conditions. Based on these observations, we hypothesized that the luminous glue of P. litoralis may function to deter and/or divert from the predation, and that the luminescence might even enhance the avoidance learning of the predator. Nevertheless, in terms of the function of luminescence, we consider that the global distribution of P. litoralis is a consequence of its adaptation to the beach surface, which provided the opportunities for dispersal by current, whereas P. longissimus is endemic to the coast of the Thai-Malay peninsula [8,30] due to its deeper inhabitation in sand.
Based on microscopic observations, we con rmed that both species secrete coelomic cells by stimulation, but neither bioluminescence nor uorescence was observed in P. longissimus. The presence and absence of uorescence in the same genus of earthworm was also reported in the terrestrial genus Eisenia; E. andrei showed uorescence in coelomic uid, while E. fetida did not. Although both species are non-bioluminescent and the uorescence emission maximum in E. andrei was in the UV region, 370 nm, they suggested the difference in uorescent characteristics was useful to deliminate these closely related species as a " uorescence ngerprint" by using a uorescent probe [31]. In the case of Pontodrilus, on the other hand, the uorescence emission maximum of P. litoralis was in the visible region and the uorescence intensity was strong enough to be observed by the naked eye under portable UV light, without using any additional uorescent probe. Therefore, the uorescence ngerprint method was also applicable to Pontodrilus. Moreover, we found that the protein band patterns by SDS-PAGE were clearly different between species, thus this may have been useful as a protein ngerprint for the taxonomic delimitation of these closely related species [32,33]. The littoral zones have rich species diversity of both macro-and microorganisms [34,35]. They comprise a front of human pressure in marine ecology and one of the most important zones for conservation [36,37]. Therefore, understanding of littoral fauna is unavoidable. Earthworms have principally strong effects on soil ecosystems [38][39][40]. Pontodrilus is a major "ecosystem engineer" [40] that inhabits the littoral habitat. Thus, species identi cation of P. litoralis and P. longissimus is signi cant to assess the littoral environment. They are actually distinguishable by the internal morphology of the spermathecal diverticulum, but special skills and equipment are necessary for the morphological analyses. In this study, we showed the differences in the bioluminescence, uorescence, and protein-ngerprinting characteristics between P. litoralis and P. longissimus, and demonstrated that the analysis of these differences provides an easy in situ methodology to identify these earthworms for marine ecological studies and conservation of littoral zones in Southeast Asia.

Extraction of the luminescent substance
To prepare the crude Pontodrilus luciferase and luciferin, the live earthworms were rinsed with distilled water and transferred for 24 h to Petri dishes in wet tissue paper moistened with arti cial seawater to avoid the contamination of their stomach content when extracting coelomic uid. All experiments were carried out on ice except for the measurements of light intensity and spectra. Coelomic uid was extracted as follows: twenty live worms of each species (2.72 g wet weight of P. litoralis and 7.4 g wet weight of P. longissimus) were put on a mortar and stimulated with a pestle to induce exudation of coelomic uid, then 10 ml of 50 mM Tris-HCl at pH 7.2 was added. After removing the specimens, the solution was centrifuged at 15,000 × g for 15 min at 4 °C in a TOMY MX-100 high speed refrigerated microcentrifuge, and the supernatants were collected as the crude extracts. The crude luciferin and luciferase fractions were prepared based on the method by Bellisario [41]. In brief, the crude extract was ltered using a 10K centrifugal lter device (Merck, Germany), and the rst ow through was used as a crude luciferin extract and the retentates on the membranes were collected as crude luciferase extract.

Cross-reaction experiment and spectral measurement
The total protein concentrations of crude luciferase extracts measured using a protein assay kit (Bio-Rad, USA) were 19.56 µg/ml in P. litoralis and 102.78 µg/ml in P. longissimus. The luminescent activity was monitored using a luminometer (Centro LB960, Berthold). Ten µl of crude luciferase was mixed with 40 µl of crude luciferin and 10 µl of 0.3% hydrogen peroxide was injected to initiate the luminescence reaction.
The luminescence was recorded in relative light units (RLUs) for 120 s accumulation after injection of hydrogen peroxide.

Spectral measurement
Luminescence and uorescence spectra were recorded with a spectro uorometer (JASCO, FP-777W). For the luminescence spectra measurements, the excitation light source was shut off. Smooth data were applied using the binomial method and the spectral response was not corrected. An in vivo luminescence spectrum was obtained using a single living specimen put into a quartz cuvette immediately after stimulation by rough handling. To obtain an in vitro luminescence spectrum, 100 µl of crude luciferase and 300 µl of luciferin were mixed with 400 ml of 50 mM Tris-HCl at pH 7.2 and 40 µl of 0.3% hydrogen peroxide, and immediately measured. Fluorescence spectra of coelomic uid in P. litoralis were obtained using crude extract suspended in 500 µl of 50 mM Tris-HCl at pH 7.2. The bandwidths used for the emission and excitation were 5 nm.

Coelomic cells photography and SDS-PAGE
Coelomic cells of Pontodrilus were isolated by stimulating earthworms on microscope slides, observed under a uorescence microscope (Nikon Eclipse E600, Japan) with a 60x objective lens (Nikon CFI Plan Fluor Series, Japan), and photographed using an attached digital camera (Nikon D5500, Japan). The uorescence excitation was 380 nm.
The protein of crude coelomic extract of both species was run by 15% SDS-PAGE gel using a 1D Gel Electrophoresis Mini Gel, AE-6530mPAGE (ATTO), followed by silver staining (Silver Stain MS Kit, FUJIFILM Wako Pure Chemical Corporation).

Video recording
Video recording of the live specimens was performed using a Nikon D500 and Micro NIKKOR 60 mm lens (Nikon) with the following settings: ISO 64000, F2.8, expose 1/60 s, under red light (LED Lenser T 2 QC).