Differential Impacts of the Head on Platynereis dumerilii Peripheral Circadian Rhythms

The marine bristle worm Platynereis dumerilii is a useful functional model system for the study of the circadian clock and its interplay with others, e.g., circalunar clocks. The focus has so far been on the worm’s head. However, behavioral and physiological cycles in other animals typically arise from the coordination of circadian clocks located in the brain and in peripheral tissues. Here, we focus on peripheral circadian rhythms and clocks, revisit and expand classical circadian work on the worm’s chromatophores, investigate locomotion as read-out and include molecular analyses. We establish that different pieces of the trunk exhibit synchronized, robust oscillations of core circadian clock genes. These circadian core clock transcripts are under strong control of the light-dark cycle, quickly losing synchronized oscillation under constant darkness, irrespective of the absence or presence of heads. Different wavelengths are differently effective in controlling the peripheral molecular synchronization. We have previously shown that locomotor activity is under circadian clock control. Here, we show that upon decapitation worms exhibit strongly reduced activity levels. While still following the light-dark cycle, locomotor rhythmicity under constant darkness is less clear. We also observe the rhythmicity of pigments in the worm’s individual chromatophores, confirming their circadian pattern. These size changes continue under constant darkness, but cannot be re-entrained by light upon decapitation. Our works thus provides the first basic characterization of the peripheral circadian clock of P. dumerilii. In the absence of the head, light is essential as a major synchronization cue for peripheral molecular and locomotor circadian rhythms, while circadian changes in chromatophore size can continue for several days in the absence of light/dark changes and the head. Thus, in Platynereis the dependence on the head depends on the type of peripheral rhythm studied. These data show that peripheral circadian rhythms and clocks should also be considered in “non-conventional” molecular model systems, i.e., outside Drosophila melanogaster, Danio rerio, and Mus musculus, and build a basic foundation for future investigations of interactions of clocks with different period lengths in marine organisms.


Introduction 33
Extensive research focusing on drosophilids and mice showed that the daily behavioral, physiological and 34 metabolic cycles in animals arise from coordination of central circadian clocks located in the brain and 35 peripheral clocks present in multiple tissues (1-3). In Drosophila, several peripheral tissues and appendages 36 (e.g. Malpighian tubules, fat bodies and antennae) have autonomous peripheral clocks that are directly 37 entrained by environmental cycles independent of the central clock, while others, such as oenocytes, are 38 regulated by the circadian clock located in the brain (4,5). The mammalian circadian system is highly 39 hierarchically organized. The master central clock in the suprachiasmatic nucleus (SCN) of the brain (often 40 referred to as a "conductor") synchronizes internal clock timing to the environmental solar day by passing the 41 information to the peripheral clocks via endocrine and systemic cues (6,7). These peripheral clocks also have 42 self-sustained circadian oscillators, with the master clock coordinating their phase to prevent 43 desynchronization among peripheral tissues, rather than acting as a pacemaker responsible for the periodicity 44 of the cycling itself (8). Besides being phase-controlled by the "SCN conductor", several mammalian peripheral clocks (e.g. in liver and kidney) have been shown to directly respond to non-photic entrainment cues, like food 46 or exercise (9). 47 We next tested if the peripheral circadian clock transcript oscillations would continue synchronously under 93 constant darkness for three days. All tested transcripts dampened strongly, with tr-cry still exhibiting weak 94 oscillations (Fig.1C). 95 Light signals are sufficient to maintain circadian clock transcripts in the trunk, independent of the head 96 To analyze if the peripheral circadian clock transcript oscillations in the worm's body were dependent on the 97 circadian clock in its head, we took advantage of the fact that bodies of decapitated animals survive in seawater 98 for two weeks (53). Adult animals were decapitated and placed under standard light-dark cycles for three days 99 before RNA extraction. Relative transcript levels of bmal, period and tr-cry exhibited overall similar levels and 100 continued a clear diel cycling of expression in trunks of decapitated animals ( Fig. 2A), indicating that the 101 peripheral circadian clock gene expression continues to synchronously run even in the absence of the brain 102 circadian clock. When we tested core circadian clock transcript oscillations in trunks of decapitated animals 103 under three days of constant darkness, no significant oscillations were detectable (Fig. 2B). 104 By placing decapitated animals on an inverted light cycle, we next tested for the capacity of peripheral clocks 105 to be re-synchronized by light, again using the transcript oscillations of bmal, per and tr-cry as readout. Cycling 106 of bmal and per was re-entrained to the inverted cycle when exposed to white, red or blue light (Fig. 3A,B, for 107 spectra and intensity see in Additional file 2: Figure S2). tr-cry transcript oscillations differed from this, in that 108 white and blue light could re-entrain its peripheral oscillations as in the case of bmal and per, whereas red light 109 did not (Fig. 3C). Overall, these results demonstrate that the peripheral circadian clock transcripts directly 110 respond to changes in the light cycle, independent of the head. 111

Chromatophore size follows a circadian pattern and free-runs under constant darkness 112
In order to assess how our findings on core circadian clock transcripts oscillations might relate to physiology 113 and behavior, we investigated possible outputs of peripheral circadian clocks, starting with changes in 114 chromatophore size in the trunk. Chromatophores are located along the dorsal part of the segmented body of 115 P. dumerilii. Based on light microscopy analyses it had previously been shown that the worm's chromatophores exhibit a segment-autonomous, diel contraction-expansion rhythm with increasing size 117 during the day and decreasing size at night (50-52). 118 We first aimed at identifying a possibility to automatize the recording of chromatophore changes. We found 119 that chromatophores exhibit a well-detectable autofluorescence under 488 nm light (Fig. 4A,B), which can be 120 used for automatic detection by any image software. In order to characterize the pattern of 121 contraction/expansion of the chromatophores, animals were photographed every three hours over the course 122 of 24h using a fluorescence microscope. We found a clear circadian pattern with higher chromatophore 123 expansion between ZT2 and ZT11 (Fig. 4C), and a sharp drop on chromatophore size from ZT11 to ZT14, before 124 lights go off and the subjective night period starts, already suggestive of an autonomous clock-driven process 125 and not a direct light response, again consistent with previous observations (55). 126 We next focused on sampling points corresponding to ZT/CT2 and ZT/CT14, during which a ~60% drop in 127 chromatophore size was evident (Fig. 4C, Fig. 5A,B), and used these two time points as a reference for the 128 study of circadian cycling of pigmentation over multiple days. Chromatophores expansion/contraction 129 continued to cycle in animals placed under constant darkness for five consecutive days (Fig. 5C) 130 For evidence that the light used for measuring chromophore size is not causing re-entrainment in this case, 131 see below. 132 Circadian pattern of chromatophore size free-runs, but cannot be re-entrained by light in decapitated 133 animals 134 In order to assess if the regulation of the cycling on chromatophore size is governed by peripheral clocks, 135 decapitated animals were used. The same animals were photographed at ZT2 and ZT14 prior to decapitation 136 as a starting reference point. Following decapitation, worms were placed in constant darkness for four days, 137 and subsequently exposed again to a normal LD cycle for additional three days. Chromatophores sizes were 138 registered along the experiment from day 0 to day 4, and again on day 7 to test for possible re-entrainment. 139 Upon decapitation, animals under DD conditions initially continue to exhibit clear rhythms of chromatophore size changes (Fig. 5D, for individual replicas see Additional file 3: Figure S3). Starting with the second day in 141 DD, the rhythm will however start to dampen and become statistically non-significant by day 3 (Fig.5D). 142 Subsequent exposure to a normal LD cycle did not lead to a re-synchronization of the chromatophore rhythm 143 (Fig. 5D). Consistently, cycling of chromatophore size does not get re-entrained on decapitated animals under 144 an inverted LD cycle applied for 5 days (Fig. 5E). 145 In order to rule out that we may have missed phase-shifts on decapitated animals due to the exposure to the 146 488nm light during the measurement procedure (due to too low sampling frequency), we also performed a 147 more densely spaced 24-hour sampling on day 5 in LD (post decapitation). This experiment confirmed our 148 interpretation of a dampening of the chromatophore size rhythm and inability to re-entrain in the absence of 149 the head (Fig. 5F). All together, these results suggest a circadian pattern of chromatophore size governed by a 150 peripheral clock, which however requires the head to maintain extended synchronization and for re-151 entrainment by light. 152 Circadian locomotor activity follows the light-dark cycle, but does not free-run under constant darkness in 153 decapitated worms 154 We next turned to locomotor activity as a read-out for circadian clock activity. We have previously shown that 155 P. dumerilii during new moon (circalunar phase of its circalunar clock) exhibits nocturnal circadian locomotor 156 activity, which free-runs under constant darkness for at least three days (14,48). We meanwhile established 157 an automated worm locomotor behavioral tracking system, which measures worm activity as distance moved 158 of the worm´s center point within 6 min time bins (56-58). It should be noted that this new type of analysis 159 measures relative distance moved, compared to the binary (movement: yes or no) manual scoring done by 160 Zantke et al. (14). Intact worms showed a significant circadian rhythmic locomotor activity under LD and DD 161 conditions (Fig. 6A,B, for individual actograms see Additional file 4: Figure S4). In contrast decapitated worms 162 exhibit an overall severe reduction in movement and rhythmicity (Fig.6C,D, for individual actograms see 163 Additional file 5: Figure S5). No significant circadian rhythmicity can be observed under constant darkness in headless worms (Fig.6C,D), with one exception. Overall, these data suggest that lack of signals from the brain 165 lead to a general suppression of movement and lack of circadian information for the locomotor output. 166 Headless worms are not generally paralyzed, as they can show spontaneous bursts of movement (Additional 167 file 5: Figure S5) and-depending on their stage at decapitation-can proceed to maturation and the associated 168 behavioral changes (59). Similar to the rhythmic transcript oscillations of the core circadian clock genes in the 169 trunk, acute light functions as a synchronization cue to the periphery of headless worms, but without head or 170 light stimuli a circadian locomotor pattern cannot be maintained by the trunk alone in 97% of the worms 171 analyzed. 172

Molecular peripheral clock 174
Here we examine peripheral circadian clock transcript changes and diel rhythms in chromatophore size and 175 locomotor behavior in P. dumerilii in the absence and presence of its head. An endogenous circadian rhythm 176 driving body pigmentation change in P. dumerilii had been previously proposed based on photographic 177 recordings of isolated groups of body segments during the middle of the day and the night (51,52). With the 178 exception of gills in oysters (60), there is no information on fluctuation of circadian clock genes on peripheral 179 tissues or appendages in marine organisms. We document the expression of core circadian clock genes in the 180 periphery of P. dumerilii, arguing in favor of a functional circadian peripheral clock and opening an avenue to 181 study the molecular mechanisms of peripheral clocks in marine invertebrates. We show that light/dark cycles 182 can (at least in part) substitute for the head as major synchronizer for continuous peripheral core circadian 183 clock transcript oscillations. We confirm previous work on chromatophore rhythms, which -in contrast to 184 locomotion and transcript oscillations-exhibit free-running rhythms in trunks of headless worms. 185 As previously reported for the central clock (14), period and bmal transcripts cycle in antiphase from each 186 other, while tr-cry transcripts are neither directly in-phase nor in anti-phase with any of them. Our results on 187 trunks show that cycling of bmal and period continues under LD conditions, but not under DD independently of the head being present or absent, and can be entrained to an inverted LD cycle on decapitated animals. 189 These results suggest that the peripheral clocks are light dependent (through a yet to be identified set of 190 photoreceptors in the trunk) and get out of synchrony in the different peripheral cells and tissues on the 191 transcriptional level (at least as fast as three days in DD) in the absence of the head. A system with peripheral 192 clocks independent from the central circadian clocks in the brain and entrained directly by environmental 193 signals is reminiscent of Drosophila melanogaster (5). In that sense the peripheral clock in P. dumerilii 194 resembles that of insects, and light reception in the trunk likely occurs via Cryptochromes and/or Opsins. 195 Candidates include Go-Opsin1 and rOpsin1 (56,61). 196 Light exhibits different effects on the different readouts. In the cases of transcript oscillations and locomotor 197 activity the head is not required for its impact, suggesting that peripheral photoreceptors mediate this signal. 198 Interestingly, different wavelengths appear to have differential peripheral effects on transcript oscillations (red 199 light being able to re-entrain per and bmal trunk oscillations, but not tr-cry), which already indicates the 200 involvement of more than one photoreceptor. It will be interesting for future studies to understand why tr-cry 201 transcripts behave differently from bmal and per transcripts under different light conditions. These differences 202 could be the result of tr-cry only being expressed in a subset of tissues/cell types that do not desynchronize as 203 quickly. 204 As in the case of insects (5) and mammals (62), questions regarding how peripheral clocks are entrained and 205 the actual mechanisms that peripheral clock use to drive transcriptional changes on various tissues are still 206 questions to be answered in P. dumerilii, as is the case for the specific functions of the peripheral clocks. 207 It should be noted that our experiments were performed on the complete trunk (or at least several segments), 208 which also leaves open questions regarding the peripheral circadian cycling and their synchronizations in 209 specific segmentally repeated organs and tissues. 210

Chromatophore size and the peripheral clock
Daily changes in chromatophore size on P. dumerilii can be easily used as a visual read out of the circadian 212 clock, adding up to its locomotor activity, circalunar spawning and clock-related genes as means to study 213 chronobiology in this model system. 214 Remarkably, the first studies on the cyclic changes of body pigmentation in P. dumerilii date back to 1939 (55). 215 Based on this and further classical work, these changes in body pigmentation were already attributed to an 216 endogenous circadian rhythm present in each segment of the worm's body (50-52), pointing at the existence 217 of peripheral circadian clocks long before their molecular mechanism had been unraveled and circadian 218 peripheral oscillations were proven to exist in the peripheral tissues in mammals (63-66). Our analyses support 219 the classical studies on Platynereis, in that chromatophore size in the body of P. dumerilii is higher during day 220 time, with a major drop before the night begins; which argues in favor of a clock-driven manifestation and not 221 a direct light response. The average magnitude of this change corresponds with previous quantitative reports 222 by Fischer (50). We report that individuals placed in DD for five days still exhibit a significant daily difference 223 on chromatophore size, although its maximum value decreases by 25% compared to the initial LD conditions. 224 It has been reported that such cycling starts to fade on individual chromatophores after seven days in DD (50), 225 but we did not test for the long term stability of the cycle for individual chromatophores. Noticeably, a similar 226 circadian cycling of chromatophore size on DD and inverted LD cycles has been shown for the marine isopod 227 Eurydice pulchra (23), posing the questions if this regulation is similar. 228 Changes in chromatophore size with a circadian cycling is commonly seen in crustaceans (23,67-69). There is 229 usually an increase in size during the day thought to be related to UV protection (70), but an inverted pattern 230 of bigger chromatophore size during the night, to possibly enable individuals to camouflage with the substrate, 231 can also be found (71). The fact that P. dumerilii is mostly active during the night (14), when pigmentation is 232 lower, does not argue in favor of a camouflage role; especially since pigmentation does not respond to changes 233 in background brightness (50). The most parsimonious option is therefore a role of pigmentation in protecting 234 the animal from UV light. It should be noted, however, that it has also been often proposed, but not tested, 235 that circadian changes in pigmentation might be a mechanism related to energy saving (72).
Remarkably, in our hands the re-entrainment of chromatophore rhythms by light requires the presence of the 237 head. This might be either because the required photoreceptor(s) are located in the head or because hormonal 238 feedback signals, such as Pdu-PDF (pigment dispersing factor) (73), are required for the synchronization 239 process. It is likely that pigmentation in P. dumerilii is controlled by a hormonal process, as in some crustaceans 240 (68,69,74), which in turn is governed by the central clock. The hormonal nature of the cycling on 241 chromatophore size has been further supported by the immediate reaction of chromatophores, present on 242 isolated skin tissue, when coelomic fluid from P. dumerilii during a given time point (e.g. day or night) is added 243 Finally, while we overall confirm previous work on the chromatophore rhythms in the trunk of P. dumerilii, 245 there is one clear difference between our findings and that of Röseler (52). Her work shows that 246 chromatophore rhythms can still be re-entrained by white light even in the absence of the prostomium (head), 247 whereas in our hands decapitated animals do not re-entrain their chromatophore rhythm in response to white 248 light. We identified two main reasons that might explain this discrepancy. The materials and methods of her 249 paper do not state the light intensity and spectrum. It is thus possible that this strongly deviates from our 250 conditions. The other difference is the extent of head removal. In our study we removed the head including 251 the jaw piece, whereas Röseler (52) specifies prostomium-removal, which implies that her worms still 252 possessed the jaw and the surrounding tissue. This region possesses multiple neurons and neurosecretory 253 cells, which could be important for proper re-entrainment. Further work will certainly help to disentangle these 254 differences. 255

Conclusions 256
We find that the overall circadian clock transcript oscillations of the trunk are under strong control of the light-257 dark cycle and do not show synchronized oscillation under constant darkness, irrespective of the absence or 258 presence of heads. In the absence of heads, locomotor activity is also strongly coordinated by the light-dark 259 cycle. In contrast, circadian changes of body pigmentation in the trunk free-run over several days in constant darkness, even in the absence of the head. Jointly these data indicate that autonomous peripheral clocks exist 261 in the trunk of the bristle worm, coordinating for instance pigmentation. However, the synchronization of 262 rhythmic circadian oscillations in other peripheral tissues and their respective output are more strongly 263 coordinated by light than by the circadian oscillator positioned in the head of the worm. Our data build a basis 264 for future analyses of the multiple clocks of the bristle worm, but also suggest that peripheral clocks should be 265 taken into consideration when studying other organisms with circadian and non-circadian oscillators. 266 They were consistently performed between ZT 13.5 and ZT14. For head samples, the region containing the 277 pharynx and the posterior end of the head was removed (see (14)). 278

Circadian re-entrainment under white, blue and red light conditions 279
When testing for circadian re-entrainment (i.e. an inverted light cycle) under white, blue or red light, animals 280 were exposed to the new conditions for seven days before sampling. Light spectra and intensities of white, 281 blue and red LEDs (ProfiLuxSimu-L from GHL advanced technology gmbh, Germany) used for circadian re-282 entrainment were measured using an ILT950 spectrometer (International Light Technologies Inc., Peabody, USA). Special care was taken to account for the standard conditions where the worms were housed, i.e. 22cm 284 away from light source and with a transparent plastic lid positioned between the spectrometer and the light 285 source. Measured light intensities for white, red and blue lights were 8,2 x 10 13 photons/cm 2 /s, 3,8 x 10 13 286 photons/cm 2 /s and 2,4 x 10 13 photons/cm 2 /s, respectively (for spectra see Additional file 2: Figure S2A). 287

Total RNA extraction and RT-qPCR 288
Total RNA was extracted from heads or trunks (i.e. decapitated animals) using the RNeasy Mini Kit (QIAGEN). 289 Reverse transcription was carried out using 0.4 μg of total RNA as template (QuantiTect Reverse Transcription 290 kit, QIAGEN). RT-qPCR analyses were performed using a Step-One-Plus cycler. The expression of each test gene 291 was normalized by the amount of the internal control gene cdc5. The relative expression was calculated using 292 the formula 1/2 ΔCt . Primers and PCR program used are listed in Zantke et al. (14). 293

Chromatophore size 294
Three consecutive segments located towards the middle of the body were selected on each animal to evaluate 295 changes in chromatophore size. In order to precisely re-identify the same segments over the course of the 296 experiments, animals were anesthetized with MgCl2 and a parapodium, two segments away from the region 297 of interest, was removed with a sterile surgical blade. When required, one animal at a time was placed on a 298 glass cover without water and extended carefully. An epifluorescence stereoscope (Zeiss Lumar) with a 488 299 nm laser and a FITC filter was used to take pictures making sure to always use the same magnification across 300 animals and sampling points. Animals were placed again in seawater until the next sampling point. Image 301 analysis was done using Adobe Photoshop. On fluorescent images, the RGB channels red and blue were 302 lowered to zero, and the three segments of interest were extracted by erasing the unwanted area 303 (chromatophores on each segment have a specific pattern, which makes their visual identification easier). A 304 new layer was generated by using the magic wand tool to single out the bright green chromatophores from 305 the background fluorescence. Using the image's histogram, the number of colored pixels was used as a proxy 306 of chromatophore size. Animals had between 20 and 40 chromatophores along the three segments, but no effort was done to quantify size of individual chromatophores. Instead, the sum of all the chromatophores of 308 interest were used as chromatophore size value at each time point. Absolute pixel number was expressed as 309 a percentage of the maximum value for each animal across all time points of the experiment (i.e. xi/(Max{xi, … 310 , xj}*100). Average among biological replica and SEM were further calculated. 311

Locomotor Activity Assay 312
Immature worms of comparable size were starved for three days before the start of the assay. After 313 decapitation, worms were placed in individual hemispherical concave wells (diameter = 35 mm, depth = 15 314 mm) of a custom-made 36-well clear plastic plate (as described in (56)). Intact worms were also treated with 315 MgCl2 for 5min prior to locomotor recording to ensure proper comparisons to decapitated worms. Video 316 recording of worm's behavior over several days was accomplished as described previously (14)

Statistical Analyses 326
Statistical analyses were performed using either the Data Analysis plug-in in Microsoft Excel using an alpha 327 value of 0.05 (molecular and chromatophore data) or GraphPad Prism (locomotor activity data). For changes 328 in chromatophore size, each two-sample two-tailed student's t-Test was preceded by an F-Test to check if the 329 variances of the two groups were equal or not. To test if transcriptional changes in gene expression over time 330 oscillated to a statistically significant difference, fold change data was analyzed for each sampling point using single factor ANOVA. In order to ease the logistic process of analyzing a considerable amount of data, RNA 332 samples and chromatophore size images were not analyzed blindly but chronologically as experiments were 333 being performed. Statistical differences in locomotor activity across treatments were estimated using repeated 334 measures ANOVA followed by Sidak´s multiple comparison test. To identify the free-running period length of 335 intact worms under DD conditions Lomb-Scargle periodogram analysis was done using the ActogramJ plugin 336 for Fiji (75). 337 Ethics approval and consent to participate: All animal work was conducted according to Austrian and 338 European guidelines for animal research. 339

Consent for publication: Not applicable. 340
Availability of data and material: All sequence resources referred to here were already published previously 341 and submitted to public databases. All other data that support the findings of this study are available from the 342 corresponding author upon reasonable request. 343 Competing interests: The authors declare that they have no competing interests.     circles (prominent green autofluorescence of the jaws can also be see in the anterior end). (C) Average chromatophore size based on autofluorescence (see materials and methods) of the same group of individuals 557 followed over a 24h period (n=15) under standard LD conditions. Pairwise student's t-Tests (preceded by F-558 Tests) were performed comparing each of the ZT hours. Individual averages from ZT2 to ZT5 and ZT14 to ZT23 559 were statistically similar within them but not between them in all permutations possible (alpha=0.05). Error 560 bars denote SEM. 561 Average distance moved of intact (B) and headless worms (D) during day and nighttime, as well as during 577 different time periods of constant darkness (i.e. CT4-CT16 for subjective day, and CT16-CT4 for an extended 578 subjective night). The subjective night period (CT16-CT0) was extended to CT4 because under DD conditions 579 activity levels cycle with a ca. 25h ± 0.22 period (mean ± SEM, see Additional file 4: Figure S4 B), and therefore activity during DD1-3 darkness is expected to run into the early phase of the subjective day. For day and night 581 time activity, only data from LD3 and and LD4 were pooled, because during the first two days after head 582 removal worms hardly moved. Bars indicate mean ± SEM. p-values were calculated using repeated measures 583 ANOVA followed by Sidak´s multiple comparison test with ****p < 0.0001, **p<0.01, *p<0.