Physiological control of bioluminescence in a deep-sea planktonic worm, Tomopteris helgolandica

Planktonic organisms can achieve almost perfect invisibility via transparency, making this one of the most valuable and fascinating adaptations to pelagic environments (Johnsen and Widder, 1998). Paradoxically, at first sight, some of these cryptic organisms are also brightly luminous (Haddock and Case, 1999; Poupin et al., 1999).

The tomopterid holoplanktonic polychaetes belong to a particularly diverse family that is extremely transparent (Johnsen and Widder, 1998) and includes at least 11 bioluminescent species (Poupin et al., 1999). Although the spectral distribution has been measured in only two species (λ_max=565 nm for Tomopteris nisseni and λ_max=570 nm for T. septentrionalis) (Dales, 1971; Latz et al., 1988), these pelagic worms are often described as yellow emitters, in contrast to most bioluminescent marine organisms, which emit a blue–green light. Additionally, given the species-specific luminous organ distribution observed in this family, this unusual luminescence has been suggested to be an intraspecific communication signal (Harvey, 1952; Dales, 1971; Latz et al., 1988). Since the review of Dales (Dales, 1971), this elegant hypothesis is often highlighted despite the lack of evidence, given basic experimental data are missing such as emission pattern and associated physiological control mechanisms. Here, we focused on these fundamental aspects and present a pharmacological approach to investigating the luminescence control of T. helgolandica Greeff 1879, a widespread, previously established bioluminescent East Atlantic species (Harvey, 1952).

Tomopteris helgolandica specimens (1.5–6.0 cm in length), were collected from October 2010 to November 2012 at a 200–300 m depth from two connected fjords, Raunefjorden and Korsfjorden (Western Norway), using two types of towed net samplers: the Isaacs–Kidd midwater trawl (1.75 m wide×1.30 m high mouth, 6.5 m long, 500 μm mesh aperture) or ring plankton nets (1.5 m diameter mouth, 300 μm mesh aperture) depending on the weather and the vessel's equipment. Live specimens were housed in classical aquariums or in Kreisel tanks commonly used for maintaining gelatinous zooplankton (Baker, 1963; Raskoff et al., 2003) in a permanent dark cold room (6–8°C) at the Espegrend Marine Biological Station of the University of Bergen, Norway.

Several anaesthetics that have previously been used on polychaetes (Smaldon and Lee, 1979; Costa-Paiva et al., 2007; Ross and Ross, 2008; Cooper, 2011), including magnesium chloride, propylene phenoxetol and tricaine mesylate, were unsuccessful with T. helgolandica. However, preliminary experiments demonstrated that menthol efficiently anaesthetised the animals within 30 min and reduced the interindividual variability of the emitted light compared with non-treated specimens without affecting light emission parameters. Thus, before each experiment, the organisms were relaxed for 30 min in a menthol solution applied at increasing concentrations (0.25–2.5 g l⁻¹).

The organisms were dissected from the head to the tail into serial preparations that comprised three parapod pairs. Each preparation was placed into 50 μl of artificial seawater (ASW: 400.4 mmol l⁻¹ NaCl, 9.6 mmol l⁻¹ KCl, 52.3 mmol l⁻¹ MgCl₂, 9.9 mmol l⁻¹ CaCl₂, 27.7 mmol l⁻¹ Na₂SO₄, 20 mmol l⁻¹ Tris, final pH 8.3). Next, light production was triggered by adding 50 μl of a given test solution. For each studied specimen, one preparation was treated with the control stimulus (200 mmol l⁻¹ KCl or, depending on the tested effect and based on preliminary results, 1 mmol l⁻¹ cholinergic agonist) whereas the other preparations were treated with different pharmacological substances (Table 1). The light responses were standardised and expressed as a percentage of the control light response. All of the experiments were designed following the Latin square principle: from one specimen to another, the preparation from an identical position is never treated twice by the same solution to eliminate possible interindividual or interpreparation variability. The pharmacological solutions were prepared in ASW buffered at pH 8.3 just before the experiments were performed. The light intensity was measured for 10–20 min with a single tube luminometer (FB12, 2005, Berthold Technologies, Bad Wildbad, Germany). To avoid light stress or artefactual measurements, all handling and experiments were performed in partial darkness or under red lighting.

All of the statistical analyses were performed with JMP software (v10.0.0, 2012, SAS Institute Inc., Cary, NC, USA) on log-transformed data [∑log(x_(1→n))/n] for relative values >1. Variance normality and equality were first tested by the Shapiro–Wilk test and Levene's test, respectively. When these parametric assumptions were not met, a one-way ANOVA was replaced by a non-parametric Kruskal–Wallis ANOVA to assess the significant difference between more than two groups. All of the pairwise comparisons were tested using a post hoc Student's t-test (each pair), Dunnett's test (with control) or a post hoc Dunn's test, as appropriate. Each difference was considered to be significant at 0.05. For clarity, the graphically illustrated values are expressed as the geometric means [√(x₁×x₂×…×x_n)] with the corresponding s.e.m.

When applied on whole specimens, potassium chloride (200 mmol l⁻¹ KCl) triggered a maximal yellow light emission at specific parapodial sites and locally reached 10³ Mquanta s⁻¹ (where 1 megaquanta=10⁶ photons) (Fig. 1A,C). Because of the linear relationship between the total emitted light (L_tot) and the maximal light intensity (L_max) of the control light responses (Fig. 1B), only L_tot is presented.

The primary neurotransmitter families known to mediate bioluminescence throughout invertebrates were pharmacologically screened on T. helgolandica parapods (Nicol, 1960; Case and Strause, 1978; Anctil, 1979; Gardner and Walker, 1982; Anctil, 1987; Walker et al., 1996) (Table 2). Only carbachol elicited a luminescence higher than the ASW-induced emission (mechanical stimulus).

The amount of emitted light increased with increasing carbachol concentration (Fig. 2A), and different emission patterns were observed at various concentrations. At low carbachol concentrations, the pattern consisted of a series of weak intensity flashes (Fig. 2A, top left panel), which was in contrast to the monophasic shape of 1 mmol l⁻¹ carbachol light emission (Fig. 2A, top right panel), which was similar to the KCl-induced light response (Fig. 1A). However, the tissue preparations generally responded poorly to acetylcholine compared with its non-hydrolysable agonist carbachol. Thus, an eserine pre-treatment was tested. This cholinesterase inhibitor induced a weak light response (0.04±0.03% of KCl) and significantly facilitated acetylcholine-induced emission but did not affect carbachol-induced light emission (Fig. 2B). Lastly, the specificity of the cholinergic receptors was evaluated using nicotinic and muscarinic blocking agents, tubocurarine and atropine, respectively (Table 1). Only tubocurarine significantly inhibited the carbachol-induced emission (Fig. 2C). This observation was confirmed with the nicotinic agonist dimethylphenylpiperazinium (DMPP), which triggered an intense light emission (9519.55±8490.38% of KCl; not shown).

Given that the nicotinic control pathway suggests that Ca²⁺ is a second messenger, we aimed to investigate the calcium dependence of the reaction. Pre-incubating the tissue preparations in Ca²⁺-free ASW significantly inhibited the KCl- and carbachol-induced luminescence responses by 95 and 100%, respectively (Fig. 3A). The luminescence was also significantly reduced by different calcium channel blockers, verapamil (phenylalkylamines) and diltiazem (benzothiazepine), and by trifluoperazine, a calmodulin inhibitor (Fig. 3B).

Although some bioluminescent organisms produce a continuous glow, most light emission signals are transient events mediated by specific control mechanisms (Nicol, 1960). Two control levels are commonly distinguished: an extrinsic control represented by peripheral control pathways and an intrinsic control that includes the photogenic reaction and the related intracellular signalling pathways (Case and Strause, 1978). In self-luminescent metazoans characterised by differentiated photogenic structures, emission is controlled either by hormones (Claes and Mallefet, 2009) or via coupling mechanisms between photocytes and excitable cells, including neural, muscular or epithelial cells (Herring and Morin, 1978; Anctil, 1987; Hastings and Morin, 1991; Krönström et al., 2009). Although luminescence can originate in a nerve-free bioluminescent epithelium, such as in the conducting epithelia of some Hydrozoa (Bassot et al., 1978; Dunlap et al., 1987) and Anthozoa (Germain and Anctil, 1996), it is most frequently controlled by neural pathways (Nicol, 1960; Case and Strause, 1978). In addition to turning the light emission on and off, nervous control abilities can modulate and adjust the intensity, duration, frequency or angular distribution of a light signal and thus generate diversity and specificity. However, a large proportion of the functional diversity of the existing emission patterns and control systems is unknown, especially in annelids, where the most detailed bioluminescence control studies have been limited to polynoid and chaetopterid benthic species (Gardner and Walker, 1982; Anctil, 1987). Therefore, the luminescence control of pelagic worm species has been poorly documented (Harvey, 1952; Haddock et al., 2010).

According to our results, T. helgolandica luminescence is under nervous control, as revealed by yellow luminescence induced by KCl depolarisation in nervous fibres, which directly or indirectly causes a photogenic structure response (De Bremaeker et al., 1996). Furthermore, the pharmacological screen revealed a dose-dependent carbachol sensitivity. In fact, carbachol and acetylcholine both induced light emission, but the tissue preparations demonstrated a low responsiveness to acetylcholine. Pharmacological carbachol concentrations (over 0.1 mmol l⁻¹) elicited a monophasic signal, similar to KCl-induced light emission, but 1 mmol l⁻¹ acetylcholine failed to elicit such a pattern. However, eserine pre-treatment significantly facilitated acetylcholine-induced light emission, which attained similar intensities to those emitted in response to low carbachol concentrations (<0.1 mmol l⁻¹, nearest physiological concentration) and suggested an involvement of cholinesterase activity. Lastly, the flash trains (L_max=400 Mquanta s⁻¹) observed at the lowest concentrations were probably more representative of the naturally expressed signal. The specific inhibitory effect of tubocurarine on carbachol-induced emission not only indicates an involvement of the cholinergic pathway but also demonstrates the nicotinic receptor prevalence. These observations were supported by the sensitivity of the samples to the nicotinic agonist DMPP and by their calcium-dependent light response, which suggested that L-type calcium channels were also involved. Tomopteris helgolandica produces yellow flashes from each parapod through neural control that activates nicotinic cholinergic receptors and a calcium-dependent intrinsic mechanism.

Numerous and widespread cholinergic control mechanisms exist in annelids (Gardner and Walker, 1982; Walker et al., 1996). The control of bioluminescence is relatively well known in Polynoidae and Chaetopteridae (Nicolas et al., 1978; Gardner and Walker, 1982; Anctil, 1987) and is reinforced by the present study of one Tomopteridae. However, because of its ubiquity, the specific mode, level and site of action of acetylcholine in the bioluminescent process remain unclear. Muscarinic cholinergic and serotoninergic mechanisms have been described in benthic scale-worms (Polynoïdae) as part of the excitatory pathway of elytral luminescence (Nicol, 1954; Nicolas et al., 1978; Miron et al., 1987; Anctil et al., 1989). The tube-worm Chaetopterus variopedatus (Chaetopteridae) produces a glowing blue mucus in response to the contractile action of the adjacent epithelio-muscular cells, which are controlled by muscarinic cholinergic and GABAergic pathways (Nicol, 1952; Anctil, 1981; Martin and Anctil, 1984; Anctil, 1987). Given that some photocytes in other organisms are not directly innervated but are controlled by adjacent supportive cells that trigger light emission by epithelial conduction (Anctil, 1987; Dunlap et al., 1987), the calcium entry via L-type channels we observed could act at both the neuro-photocyte level and an intermediate level. The presence of parapod nerve fibres, revealed by the histological studies of Greeff (Greeff, 1882; Greeff, 1885) and Bonhomme (Bonhomme, 1952) on T. mariana and T. keferteini, respectively, suggested that the bioluminescence of these worms was under nervous control. In particular, Greeff exhibited a scheme of photogenic organs with direct nerve connections. However, the characterisation of photogenic structures remains ambiguous (Malaquin and Carin, 1922; Bonhomme, 1952).

Despite the differences observed in the light emission pattern, its control and the bioluminescence characteristics between polynoid and chaetopterid worms, the luminescence in all cases has been associated with defensive functions. The same hypothesis has been suggested for the blue luminescence of the benthic polychaete Polycirrus perplexus (Terebellidae) (Huber et al., 1989) and for the ‘green bombs’ expelled by the recently described deep-sea pelagic specimens belonging to Acrocirridae (Osborn et al., 2011). Only syllid worms, whose behaviour has been well studied, use their green light emission for both deterrence and intraspecific communication during mating swarms (Wilkens and Wolken, 1981; Tsuji and Hill, 1983; Fischer and Fischer, 1995; Gaston and Hall, 2000; Deheyn and Latz, 2009). Given that the open ocean does not facilitate contact between planktonic organisms, an atypical emission wavelength would be highly advantageous for Tomopteridae. Although the emission of yellow light has been interpreted by numerous authors as a specific signal that involves a private communication channel (Harvey, 1952; Dales, 1971; Latz et al., 1988), the maximal wavelength of T. septentrionalis (λ_max=570 nm) does not match its spectral sensitivity, which is centred on blue (Buskey and Swift, 1985).

Nevertheless, it is likely that T. helgolandica's yellow light may play different roles, as suggested by the observation of different emission patterns – flash versus glow – according to the stimulus applied. Flash is often associated with a deterrent function, whereas glows are considered attractive, suggesting that the worms modulate the light output depending on the context that incites the use of bioluminescence. However, in the absence of further experimental data, this hypothesis remains speculative.

Our results strongly support the hypothesis that T. helgolandica's bioluminescence is under nervous control, revealing new insight into the pathways involved. The yellow light flashes are produced via activation of nicotinic cholinergic receptors and a calcium-dependent intracellular mechanism involving L-type calcium channels.

However, the understanding of tomopterid bioluminescence at an ecological level is beyond our current knowledge. In addition to studying the intrinsic mechanisms of light emission, an assessment of the mechanisms that govern their visual capabilities as well as their reproductive biology and behaviour will be performed in our future research.

We thank the Espegrend Marine Biological Station of the University of Bergen and the Technological Platform of Statistical Methodology and Computing Support of the Catholic University of Louvain.