Different as night and day: Behavioural and life history responses to varied photoperiods in Daphnia magna

Abstract Nearly all animal species have utilized photoperiod to cue seasonal behaviours and life history traits. We investigated photoperiod responses in keystone species, Daphnia magna, to identify molecular processes underlying ecologically important behaviours and traits using functional transcriptomic analyses. Daphnia magna were photoperiod‐entrained immediately posthatch to a standard control photoperiod of 16 light/ 8 dark hours (16L:8D) relative to shorter (4L:20D, 8L:16D, 12L:12L) and longer (20L:4D) day length photoperiods. Short‐day photoperiods induced significantly increased light‐avoidance behaviours relative to controls. Correspondingly, significant differential transcript expression for genes involved in glutamate signalling was observed, a critical signalling pathway in arthropod light‐avoidance behaviour. Additionally, period circadian protein and proteins coding F‐box/LRR‐repeat domains were differentially expressed which are recognized to establish circadian rhythms in arthropods. Indicators of metabolic rate increased in short‐day photoperiods which corresponded with broadscale changes in transcriptional expression across system‐level energy metabolism pathways. The most striking observations included significantly decreased neonate production at the shortest day length photoperiod (4L:20D) and significantly increased male production across short‐day and equinox photoperiods (4L:20D, 8L:16D and 12L:12D). Transcriptional expression consistent with putative mechanisms of male production was observed including photoperiod‐dependent expression of transformer‐2 sex‐determining protein and small nuclear ribonucleoprotein particles (snRNPs) which control splice variant expression for genes like transformer. Finally, increased transcriptional expression of glutamate has also been shown to induce male production in Daphnia pulex via photoperiod‐sensitive mechanisms. Overall, photoperiod entrainment affected molecular pathways that underpin critical behavioural and life history traits in D. magna providing fundamental insights into biological responses to this primary environmental cue.


| INTRODUC TI ON
The cladoceran Daphnia magna is a keystone ecological model species commonly utilized for freshwater ecology research and standardized toxicological testing. Daphnia magna integrates environmental sensing and response mechanisms to enable behavioural and life history modifications commensurate with environmental challenges and seasonal changes. Changes in behaviour and life history strategy of Daphnia species are often studied in context with multiple, combinational cues, such as temperature, photoperiod, resource availability and population density (Alekseev & Lampert, 2006;Altshuler et al., 2011;Gust et al., 2016;Heugens, Hendriks, Dekker, Van Straalen, & Admiraal, 2001;Heugens et al., 2006;Jiang et al., 2014;Korpelainen, 1986). In a study of interactive effects between temperature and photoperiod (Armitage and Landau, 1982), it was concluded that (a) interactive effects occurred impacting lifespan, size, reproductive output and broods; (b) temperature had independent effects; and (c) photoperiod alone was unlikely to have independent effects. The present study was conducted to begin untangling responses to environmental cues by investigating the most basal and consistently rhythmic environmental cue, photoperiod. Across evolutionary time, nearly all animal species have harnessed photoperiod to initiate various physiological and developmental processes that ultimately influence behaviour and life history traits (Bradshaw & Holzapfel, 2007). Our investigations assessed the influence of photoperiod on the immediate behavioural responses to external stimuli and lifelong changes to D. magna life history strategy. The molecular mechanisms underlying behavioural and life history responses to photoperiod are not well understood in D. magna; therefore, a global transcriptomic expression investigation was conducted to identify functional responses involved in organism-level outcomes.
Daphnia species exhibit negative phototaxis, a behaviour that is generally hypothesized to facilitate predator avoidance (Dodson, Tollrian, & Lampert, 1997;Van Gool & Ringelberg, 1995). Deviations from negative phototaxis have been observed in Daphnia species (Michels, Leynen, Cousyn, Meester, & Ollevier, 1999;Ringelberg, 1964) that may be attributed to environmental modifiers, such as light wavelength (Storz & Paul, 1998) and food availability (Jiang et al., 2014;Neary, Cash, & Mccauley, 1994;van Gool & Ringelberg, 1995). Despite the broad evidence that photoperiod can affect animal behaviour, the influence of photoperiod on phototaxis remained to be investigated in D. magna prior to the present study. Further, the direct influence of photoperiod on critical life history parameters in D. magna also remained unclear. For example, D. magna normally reproduce via asexual clonal parthenogenesis, however, combinations of environmental factors such as temperature, shortday length photoperiod, crowding and food availability can stimulate male production enabling genetic recombination in response to a changing environment (Carvalho & Hughes, 1983;Hobaek & Larsson, 1990;Kleiven, Larsson, & Hobaek, 1992). Short-day photoperiod has been reported to play a critical role in triggering male production in Daphnia pulex (Toyota, Gavin, Miyagawa, Viant, & Iguchi, 2016;Toyota, Miyakawa, Yamaguchi, et al., 2015) and recently also in D. magna when in conjunction with increasing temperature.
Ultimately, varying investments to sexual reproduction are known to have important ecological and evolutionary consequences (Tessier & Cáceres, 2004). These observations compelled us to investigate the system-level responses to photoperiod in order to evaluate its effect(s) across behavioural, metabolic, survival, growth, reproduction and male production responses in D. magna under controlled conditions.
We conducted experiments to test the null hypothesis that photoperiod alone had no effect on specific behavioural and life history outcomes in D. magna. The experiments were conducted in custom-constructed photoperiod chambers where <24-hr-old D. magna were reared in five different photoperiods ranging from short-day, equinox and long-day states. Phototaxis responses were observed ex situ using digital tracking of velocity and duration spent in light/ dark zones (Noldus daNioVisioN ® software), and carbohydrate metabolic rate was quantified using IQ Toxicity Test™ (Aqua Survey, Inc.) to estimate overall metabolic rate, activity and health. These short-term behaviours/strategies were compared against long-term life history parameters including survival, growth, reproduction and reproductive strategy (female vs. male neonate production).
Microarray-based global gene expression assays were conducted to examine the functional genomic processes underlying both the behavioural and life history responses of D. magna to photoperiod.
Overall, the approach yielded unique system-level insights into the influence of photoperiod from the molecular to lifetime responses in D. magna.

| Experimental approach and rationale
The experimental approach focused on investigating the effects of photoperiod on key life history traits, metabolic rate, phototaxis responses and global transcriptomic expression in Daphnia magna to characterize foundational responses to light. Separate groups of D. magna were simultaneously exposed to five different photoperiods representing varying seasonal day lengths; this included a comparative control of 16 hr of light and 8 hr of dark (16L:8D) that is the standard for acute (OECD No. 20, 1984;USEPA, 2002;USEPA, 1996a) and chronic (ASTM E1193-97, 2012OECD No. 211, 2012; USEPA, 1996b) ecotoxicological methods, versus an array of alternative photoperiods including extreme short-day length (4L:20D),

K E Y W O R D S
Daphnia behaviour, Daphnia life history, male production, photoperiod, phototaxis short-day length (8L:16D), equinox (12L:12L) and extreme long-day length (20L:4D). Since changes in life history strategy are likely to involve changes in energy utilization (Ananthasubramaniam et al., 2015), an indicator of overall metabolic rate was also investigated.
Provided the inherently variable nature of life history and behavioural responses to stimuli, we repeated the photoperiod exposures in two temporally separated experiments on 17 October 2014 and 11 March 2015, hereafter referred to as experimental trial 1 and experimental trial 2. This was done to assess the reproducibility of all measures and endpoints from one experiment to the next, and assess if the two data sets could be combined for examination as one large data set.

| Experimental animals
Daphnia magna neonates (<24 hr old) obtained from in-house cultures reared for several years in a 16L:8D light cycle were used in the photoperiod experiments. Organisms were originally purchased from a commercial source in May 2009 (Aquatic Biosystems; EPA Ohio, AROF2, Lot No. 070092DM), and in-house cultures were started using one individual to ensure testing utilized a single genotype. All culturing and test methods used hard reconstituted water (HRW), formulated according to USEPA (2002). Test organisms in cultures and experiments were held at 25 ± 1°C and provided a daily ration of 2 × 10 5 algae cells/ml (Raphidocelis subcapitata, formerly Selenastrum capricornutum) and yeast-cerophyl-trout chow (YCT; Aquatic Biosystems) as prescribed by guidance (ASTM, 2012;USEPA, 2002). The algae used as food were cold stored (4°C) in the dark to inhibit growth and were directly spiked from cold storage into the experimental chambers that contained D. magna. Proliferation of the algae in the experimental chambers was not quantified, but proliferation was unlikely given rapid feeding by Daphnia immediately after spiking, tri-weekly water renewals flushing out old algae, static water conditions, test media devoid of nutrients and low relative light intensity in experimental chambers.

| General exposure methods
Specialized exposure chambers resembling ant farms were constructed from clear acrylic (JR's Custom Acrylics). The chamber dimensions were 35 × 17 × 2 cm ( Figure S1) and contained approxi-  Figure S3). Duplicate photoperiod experimental trials 1 and 2 were conducted for 21 days, adapted from the standard ASTM method E1193-97 (ASTM, 2012 duration, at 25 ± 1°C within Darwin Environmental Chambers (Darwin). Complete (100%) water renewals using HRW were performed three times a week (Monday, Wednesday, Friday); viable neonates were enumerated and removed from the experimental chambers at each water exchange. Thus, only the original individuals loaded into the chambers remained for the duration of the experiment. Water quality monitoring was performed, the values of which remained within test method specifications (EPA-821-R-02-012), with the following ranges: temperature (25 ± 2°C), pH (7.8-8.3), dissolved oxygen (>4 mg/L), specific conductivity (290 ± 20% µS/cm) and ammonia-N (<2 mg/L).

| Life history endpoints
Survival and the number of neonates produced were assessed three times per week, prior to each water change. Female and male neonates were enumerated by removing all neonates from exposure chambers, growing neonates in 2-L plastic beakers at the prescribed feeding ration for 5-7 days and identifying the sex under a dissecting microscope according to available literature and keys (Ebert, 2005;Olmstead & Leblanc, 2006

| Sugar metabolism as an estimate of metabolic rate
Fluorescence bioenergetics assays (IQ-Tox™ Assays, Aqua Survey, Inc.) were used to investigate the response to the light stimulus providing information on metabolic rates (Hayes, Douglas, & Fisher, 1996). This test kit employed a fluorometric biomarker methylumbelliferyl galactoside. The concept behind the IQ-Tox™ Assay is that healthier D. magna will take up and metabolize more biomarker and fluoresce with more intensity under UV light, while less healthy individuals will metabolize less biomarker and therefore fluoresce less.
At termination of the experiments, 21-day-old D. magna were added to IQ exposure chambers, and each cell was filled with 10 ml of test water and three drops (≈60 µl) of the IQ substrate (a fluorogenically tagged sugar suspension). After 15 min, D. magna were removed and transferred onto the lid of a 24-well microplate (Product #3524, Corning Inc.) for fluorescence imaging, and excess water was removed. Fluorescence imaging was conducted using a ProteinSimple Red™ imaging system (ProteinSimple [formerly Cell Biosciences]).
The resulting images were analysed with image-Pro Plus software to quantify fluorescence intensity and density, which was used in the statistical analysis. The data generated herein used the fluorescence intensity as an indicator of sugar metabolism rate of the fluorogenically tagged sugar, thus providing a reliable estimate of D. magna metabolic rate to compare among the various photoperiod treatments. Data presented herein were generated from experiment 1, using 8-11 D. magna per treatment.

| Phototaxis
Phototaxis was assessed for each of the experimental photoperiods at week 1 and week 3. Short-term ex situ digital tracking of swimming behaviour was conducted using a real-time digital tracking system (Noldus DanioVision ® ). Horizontal movements in light and dark conditions were tracked to remove gravity as a confounding variable.
The DanioVision apparatus projected full spectrum light (measured to be 3,713 lux) from below the experimental chamber and allowed digital test organism tracking to be performed in the dark conditions by contrast to background using infrared light. Experiments were conducted in custom acrylic raceways (2.0 × 8.6 × 12.8 cm) consisting of eight separate 1.0 × 12.8 cm arenas (Figures S1 and S4).
Window tinting (Gila Window Film -20% visible light transmission, Eastman Performance Films, LLC) was placed on the bottom of the raceway so that half of each of the eight arenas was covered, thus providing D. magna a choice between light or dark conditions. Eight individuals were randomly selected from each replicate exposure chamber (total of 8 × 3 replicates = 24 individuals per photoperiod), and initially four were placed in the dark (tinted) zone, and four were placed in the light zone. A 5-min acclimation period was allowed, followed by 10-min trials during which D. magna movement was digitally tracked, and data were recorded at seven samples per second.
At the conclusion of the 1-week phototaxis trial, the D. magna that were examined were carefully returned to the primary experimental chambers. The measurement endpoints assessed included time (s) spent in each light/dark zones and velocity (mm 2 /s).

| Microarray experiments
The effects of photoperiod on global transcript expression in D. magna were investigated using microarray assays. An Agilent Technologies (Agilent Technologies) single colour custom 8 × 15K microarray format previously developed by Stanley et al. (2013) which was subsequently reannotated in 2015 was used for all investigations. The reannotated microarray design used in the present study is available at the Gene Expression Omnibus (GEO, http:// www.ncbi.nlm.nih.gov/) under Accession no GPL16579. The experimental design for the microarray experiments utilized animals collected directly from the photoperiod exposures at the termination of the 21-day assays and within 1 hr after initiation of the light phase of each photoperiod, for both experimental trials 1 and 2. This sampling approach ultimately included developing eggs/neonates within brood pouches; however, inclusion of such developmental processes is supportive of evaluating photoperiod influences on the molecular contributions to reproductive life history outcomes. Within each exposure replicate, two population subsamples consisting of single, unique D. magna were used for hybridization to microarrays. These population subsamples were selected at random (using a random number generator) from 2 to 6 D. magna sampled for the genomic investigation that met the RNA quality criteria described in the following paragraph. In summary, 2 duplicate photoperiod experimental trials were conducted where each included 5 photoperiods × 3 exposure replicates × 2 population subsamples for a total of 30 total microarray hybridizations per experiment and thus 60 microarray hybridizations overall. A completely randomized design was utilized for microarray hybridizations where the total set of 60 experimental samples were assigned to microarrays at random (using a random number generator).

| RNA extraction and microarray hybridizations
Frozen samples were homogenized using a mortar and pellet pestle (Kimble Kontes). Total RNA isolation was conducted using the Qiagen RNeasy Mini Kit (Qiagen) following manufacturer's recommendations. RNA quantity was estimated using a NanoDrop ND-1000 Spectrophotometer (NanoDrop technologies), and final quantity and quality were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). Only samples with a 28s/18s ratio ≥2.0 and RNA integrity number (RIN) ≥7.0 were used for downstream applications.
The Agilent Low Input Quick Amp Labeling Kit (one colour) and hybridization protocol (Agilent Technologies) were utilized for microarray hybridizations following manufacturer's recommendations using 80 ng of total RNA as starting material from each biological sample.

| Microarray analysis
An Agilent Technologies, High-Resolution Microarray Scanner (Model G2505C, Agilent Technologies) was used to scan microarray images at 2-μm resolution. Data were extracted from microarray images using agileNt Feature extractioN software, version 9.5.1 (Agilent Technologies). Microarray data were normalized to the 75th percentile within each array followed by median scaling among all exposures using geNesPriNg Software version GX12.5 (Agilent Technologies). Methods for the statistical analysis of microarray data are provided in Section 2.8.

| Microarray reannotation
The assembled sequence information used to develop all probes on the D. magna microarray was reannotated in December 2016 and used for functional interpretations of the microarray results.
Reannotation was conducted using Blast searches against the NCBI nt database filtered for only those GenBank entries in the Daphnia genus, where the best match was kept. The reannotation provided greatly expanded, expressed sequence tag matches which were predominated by matches to D. magna and Daphnia pulex putative protein coding sequences. Unfortunately, as of March 2019, these putative protein sequences are not supported by protein orthologs to enable functional enrichment analysis (i.e., pathway or gene ontology). Therefore, the original annotations were used for the functional analysis.

| Text mining of new protein annotations
To add value to the new protein annotations for D. magna, text mining was conducted for all transcripts having significant differential expression in response to the photoperiod treatment using a keyword bank (Table S2) developed by the authors of the present study to match key gene functions connected to the observed phenotypes.
Specifically, the keyword bank was developed to mine terms related to circadian rhythm, sex determination, moulting/cuticle processes, egg production and visual cues related to photoperiod or light stimulus to identify potential functional responses to the photoperiod treatments. Text matching was applied to the annotation list for the complete set of differentially expressed transcripts to produce the text mining results.

| Statistical analysis
Both life history data (survival, reproduction, size) and phototaxic swimming data were statistically analysed using sigmastat Software (V3.5, SPSS). The assumptions of normality (Kolmogorov-Smirnov test) and equality of variance (Levene's test) were tested. All life history data met the assumptions of normality and equality of variance.
Swimming data that did not meet normality or variance assumptions were rank transformed. The Holm-Sidak method was used for pairwise multiple comparisons to determine differences between treatments, which were determined at the α = 0.05. One-way ANOVAs compared photoperiod to life history data, and two-way ANOVA compared multiple factors (experimental trial × photoperiod). When the experimental trial (i.e., experimental trials 1 and 2 which were conducted separately in time, n = 3) was determined to be a nonsignificant factor, life history data were combined (n = 6). For photoperiod versus behavioural swim data analysis, one-way ANOVAs were performed within each experiment, while two-way ANOVAs compared two factors (photoperiod acclimation × duration in light/ dark zones) and three-way ANOVA compared three factors (experimental trial × photoperiod acclimation × duration in light/dark zones or organism age × photoperiod acclimation × duration in light/dark zones). When the experimental trial was determined to be a nonsignificant factor, swim behaviour data from replicates 1-3 and 4-6 were also combined (n = 6).
For the investigation of microarray data, two statistical approaches were conducted, one including data from both experiments 1 and 2, while the other investigated the data from experiments 1 and 2, separately. For the combined-experiment analysis, a threeway ANOVA (p = .01) was conducted to investigate three variables: to investigate the first two variables described above: (a) photoperiod and (b) variance from subpopulation sampling. In all of the analyses described above, effects of the primary variables and all interactions among variables were evaluated. As a note, in addition to the tests described above, ANOVA tests (p = .05) including Benjamini-Hochberg multiple-test corrections were conducted for each investigation which yielded very few differentially expressed transcripts. Therefore, the results described here-forward represent noncorrected tests with the low p-values (p = .01) to balance resolution (number of differentially expressed transcripts) against falsepositive discovery. In supplement to the ANOVAs described above, post hoc pairwise tests consisting of moderated t tests (p = .05) including ≥1.5 fold change cut-off were conducted to identify transcripts differentially expressed within each experimental photoperiod (4L:20D, 8L:16D, 12L:12D and 20L:4D) relative to the control (16L:8D). All microarray analyses described above were conducted using GeneSpring GX12.5 (Agilent Technologies). As a final note, the heatmaps presented in the Results include variance associated with two population subsamples; therefore, the averaged fold change for the control photoperiod (16L:8D) may deviate slightly from one (zero once log 2 transformed).

| Life history endpoints (survival, growth, reproduction, male production)
The life history endpoint results were summarized for each of the two individual experimental trials (n = 3) and as combined data (n = 6) that included both experimental trials. Combined survival (n = 6) was relatively high (80%-92%). Carapace length ( Figure   S5) from both experiments ranged from 4.7 to 6.0 mm where in experimental trial 1, Daphnia magna were slightly but significantly larger than in experimental trial 2. Generally, there were no significant differences determined by the Holm-Sidak method (p > .05) in carapace length relative to the 16L:8D control, with the exception of significantly smaller D. magna in the 4L:20D photoperiod in experimental trial 1 ( Figure S5). Neonate production among experimental trials 1 and 2 was not significantly different (p = .37); thus, the combined data sets were analysed (n = 6) where photoperiod caused a slight, but significant decrease (p < .01) in total neonate production in the 4L:20D photoperiod relative to the 16L:8D control photoperiod (Figure 1a). Perhaps the most interesting effect of short-day and equinox photoperiods on life history traits was a significant increase in male neonate production relative to the long-day (16L:8D) control photoperiod (Figure 1b) which was observed in each experimental trial (n = 3) and within the combined data set (p < .01, n = 6).

| Sugar metabolism as an estimate of metabolic rate
Analysis of the data from experimental trial 1 (n = 8-11) identified a statistically significant increase in pixel intensity (fluorescence output) using the Holm-Sidak method in the exposed D. magna in the short-day length photoperiods, 4L:20D and 8L:16D, relative to the control 16L:8D photoperiod ( Figure 2) which is indicative of increased sugar metabolism, suggesting an increased overall metabolic rate.

| Phototaxis -digital tracking of swimming behaviour (ex situ)
Photoperiod entrainment showed significant effects on light avoid- F I G U R E 2 Effect of photoperiod on metabolic rate in Daphnia magna (from experimental trial 1, n = 8-11), as estimated by metabolism of fluorescently labelled sugar molecules and quantified as pixel intensity from whole-animal images. Error bars represent standard deviations, and asterisks denote a statistically significant differences (p = .05) relative to the control photoperiod (16L:8D) D. magna relative to 21-day-olds. Photoperiod significantly affected swim speed, which again was significantly modulated by animal age where the independent variables (photoperiod and age) displayed a significant interaction. Overall, the extreme short-day photoperiod (4L:20D) showed the most consistently significant photonegative effect on the swim velocity response across animal ages ( Figure 3).

| Microarray analysis overview
Combining transcriptomic data for experimental trials 1 and 2 in a 3-way ANOVA indicated expression differences were dominated by the variance among experimental trials which was also confirmed by PCA ( Figure S8). Given these results, the effects of photoperiod for each experimental trial were evaluated separately. , 143 false-positive tests were expected by chance; thus, the variance associated with "within-population subsampling" and "photoperiod × within-population subsampling" was near background.
Correspondingly, "within-population" subsamples tended to cluster together within PCA plots, while photoperiod-specific clustering tended to show separation from the 16L:8D control ( Figure S8). In

| Functional annotation of differentially expressed transcripts
The functional annotation results indicated that the majority of differentially expressed transcripts had associations with the primary KO categories: (a) metabolism, (b) genetic information processing and (c) environmental information processing ( Figure S9). Over half of the differentially expressed transcripts had KO associations related to metabolism where genetic information processing represented 12%-28% of transcript associates across photoperiods. Although the directional fold change of transcriptional expression was mixed for metabolism and environmental information KO categories, expression for transcripts within the genetic information processing KO category was predominantly decreased relative the 16L:8D control photoperiod ( Figure S10).
Investigation of transcriptional expression within specific KEGG pathways again showed mixed expression for the metabolism and environmental signalling KOs, where expression tended to be decreased for a variety pathways related to genetic information processing including spliceosome, ribosomal biogenesis, RNA transport, proteasome, RNA degradation and DNA replication pathways ( Figure S10). Hierarchical clustering within these primary KOs indicated that transcript expression tended to be more similar within the short-day length and equinox photoperiods (4L:20D, 8L:16D and 12L:12D) compared to long-day length photoperiods (16L:8D and 20L:4D, Figure S10).

| Text mining of new annotations
Text mining for gene transcripts significantly affected in the photoperiod treatments provided putative matches to the whole-organism observations (Table 1). Transcript expression for genes matched to circadian rhythm, sex determination and visual cues tended to have decreased expression in short-day length and equinox photoperiods compared to the 16L:8D control and the 20L:4D photoperiods. A similar response to short-day length was observed for transcripts matched putatively to "egg" processes; however, the transcript "vitellogenin fused with superoxide dismutase" had increased expression in the short-day length photoperiods (Table 1) (Table 1). Functional associations of these transcriptional observations to phenotypes are provided in the discussion.

| D ISCUSS I ON
The experimental results provide strong evidence that differences in photoperiod alone, that is without the presence other cues (e.g., temperature, crowing, low nutrition), caused substantial changes in life history strategies, including (a) a decreasing trend in total fecundity ( Figure 1a); (b) significantly increased male production  (Table 1). These responses were putatively connected to observed life history phenotypes as well as phototaxis behaviours (Table 2).
In the following discussion, we provide context for the various responses to photoperiod with regard to the sensing of environmental cues, life history strategy and the basic ecology of D. magna.

| Photoperiod effects on life history traits
Photoperiod had no statistically significant effects on D. magna survival and caused only a minor, yet statistically significant, decreasing trend in fecundity with decreasing day length (Figure 1) over the course 21-day experiments. Little impact on basic survival from a relatively minor stressor (but important environmental cue) such as photoperiod is supported in previous studies (Korpelainen, 1986). The trend of a decreased number of total neonates produced in short-day photoperiods observed in the present study (Figure 1a) is also supported by the literature for Daphnia species (Parker, 1966). The most dramatic effect of photoperiod on life history traits was the interruption of cyclic parthenogenesis resulting in significantly greater percentage of male offspring produced (Figure 1b) in short-day length and equinox photoperiods (4L:20D, 8L:20D, 12L:12D). Korpelainen (1986) also observed that photoperiod affected male production in D. magna where, converse to our observations, their results showed increased male production at the 16L:8D photoperiod relative to a short-day length photoperiod (10L:14D). It should be noted that this observation occurred predominantly at low water temperatures (14 and 19°C) relative to those used herein (25°C). In the present study, decreased reproductive output may have been influenced by increased male production in the shortest day length photoperiod where Ginjupalli and Baldwin (2013) observed that D. magna exposed to a hormone analog that induced male production also decreased overall fecundity. Circadian rhythm Period circadian protein (PER) and a protein coding F-box/ LRR-repeats (possibly JET), each of which are critical in establishing Circadian clock, were affected by photoperiod (Koh et al., 2006;Zerr et al., 1990). Photoperiod influences to circadian rhythm are critical to establishing periodic activity and metabolism and likely influence the majority of responses observed herein Metabolism Differential expression of various pathways involved in carbohydrate, amino acid/ protein and fatty acid metabolism which are critical in establishing system-level energy metabolism Growth Decreased carapace length (4L:20D) Cuticle and chitin (development, molting, reproduction and more) Increased expression of cuticular protein analogous to peritrophins 1-G (CPAP1G) in the short-day length photoperiod exposures. Peritrophin 1 (CPAP1) family is essential for development, molting, cuticle integrity, locomotion and fecundity in arthropods (Jasrapuria et al., 2012) Reproduction Decreased neonates (4L:20D)

Egg
Transcriptional expression in short-day length included: (a) Decreased expression of two vitelline membrane proteins which serve as a source chorion proteins important in early development (Pascucci et al., 1996), (b) Decreased expression of chorion peroxidase activity facilitates cross-linking of chorionic proteins during the "hardening" phase of egg development (Mindrinos et al., 1980), (c) Increased expression of vitellogenin fused with the superoxide dismutase where vitellogenin is a critical marker of Daphnia fecundity (Ebert, 1993), and (d) Decreased transcriptional expression of the female sterile 2, and inhibitor of oogenesis (Schüpbach & Wieschaus, 1991) Cuticle and chitin (development, molting, reproduction and more) Cuticular protein analogous to peritrophins 1-G (CPAP1G) had increased expression in the short day-length photoperiod where (CPAP1) family is essential for fecundity in arthropods (Jasrapuria et al., 2012) Male production Increased (12L:12D, 8L:16D and 4L:20D) Environmental signalling Decreased expression of glutamate receptor ionotropic, kainate. This target is an inhibitor of glutamate signaling in mammals (Contractor et al., 2000), which if this function is conserved in D. magna, suggests a putative increase in glutamate signaling. Glutamate signaling induces male production in the Daphnia pulex WTN6 strain (Toyota, Miyakawa, Hiruta, et al., 2015) and is a photoperiod-sensitive mechanism (Toyota, Miyakawa, Hiruta, et al., 2015) Sex determination Decreased transcriptional expression for the transformer-2 sex-determining protein in the short-day photoperiod. Differential exon splicing for transformer in Drosophila melanogaster leads to female-specific protein products in females and not males (Lopez, 1998). Involvement of transformer in D. magna sex determination is inconclusive Genetic information processing Decreased expression for multiple small nuclear ribonucleoprotein particles (snRNPs) and non-snNRP splicing factors in short day exposures. Arthropod sex determination has also been observed to be controlled by alternative splicing, although this mechanism remains to be tested in D. magna for sex determination genes including transformer (dmagtra, Kato et al., 2010) (Continues) Previous investigations combining temperature and photoperiod stimuli concluded temperature was the driving factor in affecting life history traits, such as survival, reproduction, growth brood times and male production (Armitage & Landau, 1982). The present study demonstrates that photoperiod entrainment alone can affect life history strategy in D. magna through the course of a 21-day exposure.

| Metabolic rate
To our knowledge, this is the first study to investigate the effect of photoperiod alone on an indicator of Daphnia metabolic/ bioenergetics rate. Based on literature searches, previous work integrated temperature and photoperiod as a combinational cue. Interestingly, our results indicate slight, yet statistically significant, increases in sugar metabolism as an indicator of overall metabolic rate in the two short-day length exposures (Figure 2) when ex situ measurements were taken during entrained daylight hours. We hypothesize that this is related to a circadian-like response where behavioural urgency might be increased in short-day length photoperiods (indicative of autumn/ senescence seasonal cues), though additional data generation is needed to test this hypothesis. In concert with this hypothesis, these short-day length photoperiod treatments produced more male progeny, which may be more energetically expensive to produce, necessitating increased metabolic rates. Again, this needs to be tested in dedicated experiments.

| Photoperiod effects on phototaxis
It is well documented that Daphnia species are generally photonegative (Dodson et al., 1997;Van Gool & Ringelberg, 1995), although there are some reports of photopositive clones (Michels et al., 1999;Ringelberg, 1964). Prior to this study, little was known about the effect of longterm ( where juveniles tended to be more photonegative than adults, even in clones exhibiting photopositive responses as adults. According to previous studies (e.g., Van Gool & Ringelberg, 1995), negative phototaxis responses lead to typical Daphnia diurnal water-column migration patterns which are hypothesized to optimize predatory avoidance during light conditions. The 7-day-old D. magna in the present study exposed to all photoperiods except for the extreme long-day length (20L:4D) swam significantly faster when in light conditions (approximately 7-9 mm/s), relative to dark conditions (approximately 5-6 mm/s; Figure 3). Faster swimming in the light is supported in the literature by Dodson et al. (1997) and is presumably a direct negative phototaxis response to avoid predators in light conditions and/or speed the process of moving out of the light. Overall, photoperiod was shown to affect phototaxis in D. magna with the most pronounced photonegative responses occurring at extreme photoperiods and in young life stages.

| Transcriptomic response to photoperiod on generalized functions
Kyoto Encyclopedia for Genes and Genomes orthology annotations provided insights into generalized functions affected by photoperiod

Behavior Photoperiod effects Functions Functional response(s)
Phototaxis experimentslight avoidance (duration) 7 days old Daphnia -Increased, all but 8L:16D Visual cues Decreased expression of blue wavelength opsin (KZS21495) and class a rhodopsin G-protein coupled receptor 2, which are not typically involved in circadian rhythms (Zerr et al., 1990) but may be involved in short-term light responses 21 days old Daphnia -Increased at 4L:20D and 20L:4D Circadian rhythm Period circadian protein (PER) and a protein coding F-box/LRRrepeats (possibly JET), each of which are critical in establishing Circadian clock, were affected by photoperiod (Koh et al., 2006;Zerr et al., 1990) Phototaxis experiments -light avoidance (swim velocity) 7 days old Daphnia -Increased, all but 20L:4D Environmental signalling Decreased expression of glutamate receptor ionotropic, kainate (an inhibitor or glutamate signaling) in response to short-day photoperiods. In Drosophila, balanced glutamate signaling dynamics among lateral and dorsal clock neurons mediate light-avoidance behaviours and establish entrainment of light-responsive circadian rhythms (Collins et al., 2012) 21 days old Daphnia -Increased at 4L:20D and 12L:12D Genetic information processing Decreased expression of nearly every transcript associated with the spliceosome. Alternative mRNA splicing of the Down syndrome cell adhesion molecule (Dscam) homolog in Daphnia can affect in nervous system structure/function (Brites et al., 2008) TA B L E 2 (Continued) ( Figure S9). Below we discuss observations from the three major KO classes: environmental signalling, genetic information processing and metabolism.

| Environmental signalling
The pathways included within the KEGG environmental information processing ontology ( (Toyota, Miyakawa, Yamaguchi, et al., 2015), a response comparatively attributable to methyl farnesoate (putative juvenile hormone in daphnids) exposure (Toyota, Miyakawa, Hiruta, et al., 2015). Given the transcriptional evidence for potential increased glutamate signalling and the increased male production observed in the D. magna strain tested in present study, similarity in photosensitive mechanisms of male production is plausible between D. magna and the D. pulex

| Genetic information processing
The short-day length and equinox photoperiod exposures caused predominantly decreased expression for gene transcripts involved in multiple facets of genetic information processing ( Figure S10). A striking observation was the decreased expression of nearly every transcript associated with the spliceosome ( Figure S11). Spliceosomes are large complex ribonucleoproteins (RNPs) present in eukaryotes that facilitate splicing of introns from pre-mRNA as well as alternative splicing of exons to express multiple forms of an mRNA from a single gene (Will & Luhrmann, 2011). The transcriptomics for D. magna indicated decreased expression, predominantly in the short-day and equinox photoperiods, for multiple small nuclear ribonucleoprotein particles (snRNPs) and non-snNRP splicing factors ( Figure S11) that make up the spliceosome indicating potentially reduced capacity for spliceosome-based pre-mRNA processing (Black, Chabot, & Steitz, 1985;Blencowe, Bowman, McCracken, & Rosonina, 1999).
In the human genome, alternative exon splicing has been identified to be a major contributor to proteome diversity where estimates range from 40% to 80% of genes having alternative splicing potential (Matlin, Clark, & Smith, 2005). Extensive alternative exon splicing has been observed in Daphnia for specific genes, such as the Down syndrome cell adhesion molecule (Dscam) homolog, where 13,000 unique transcripts are formed (Brites et al., 2008).
Similar to observations in insects, the alternative exon splicing of Dscam in Daphnia has been implicated in playing a critical role both in immune responses and in nervous system structure/ function (Brites et al., 2008). In certain species, arthropod sex determination has also been observed to be controlled by alternative splicing, although this mechanism remains to be tested in D. magna for sex determination genes including transformer (dmagtra, Kato et al., 2010). Presently, possible outcomes of the overall decreased expression include overall differential exon splicing which may affect sex-specific phenotypes, determination of neural state/ function and potential effects on immunity.

| Metabolism
Gene transcripts involved in a diverse suite of metabolic pathways were differentially expressed in response to different photoperiods, including effects on various carbohydrates, amino acids/ proteins, nucleic acids, fatty acids and etcetera ( Figure S10). Photoperiod proved to be an influential environmental stimulus affecting reproductive output, male production, energy expenditure and phototaxis (Figures 1-4). Such responses are underpinned by diverse suborganismal metabolic responses orchestrating the allocation of resources and energy towards the overall suite of life history traits (Ananthasubramaniam et al., 2015). Similar to the results of the present study, Toyota et al. (2016) demonstrated that the shortday length photoperiod caused differential expression of gene transcripts within various KEGG metabolomic pathways including alanine, aspartate and glutamate metabolism, starch and sucrose metabolism, amino sugar and nucleotide sugar metabolism, and galactose metabolism. Many of the common responses between our study and the Toyota et al. (2016) investigation were related to system-level energy metabolism. Effects of equinox and/or shortday photoperiod on circadian rhythms (i.e., increased phototaxis; Figure 3) and life history strategies (i.e., increased male production; Figure 1) undoubtedly influenced energy utilization in D. magna as evidenced by increased energy expenditure in the short-day length photoperiods (Figure 2). Overall, the transcriptional evidence suggests changes in multiple metabolic pathways that putatively drive the functional molecular mechanisms underlying the organismal response in D. magna that are described above, systemically initiated through environmental sensing/signalling and subsequent genetic information processing.

| Finding functions -connecting omics to organism-level responses
Similar to the use of KEGG orthology annotations, we leveraged the annotation text mining results to connect gene functions to observed organism-level responses (

| Sex determination
As described in the Environmental Sensing and Genetic Information Processing sections above, glutamate signalling pathways and alternative mRNA transcript splicing mechanisms, respectively, potentially contribute to the male production observed in the present experiment. Additionally, decreased transcriptional expression for the transformer-2 sex-determining protein (JAN91947; Table 1) was observed in concordance with increased male production (Figure 1) in D. magna exposed to short-day length photoperiods. Differential exon splicing of the transformer (tra) and Sex-lethal (Sxl) genes in Drosophila melanogaster led to production of female-specific protein products in females, but not males (Kato et al., 2010). Kato, Kobayashi, Watanabe, and Iguchi (2011) conducted a genomic study of the transformer (tra) gene in D. magna (dmagtra) which indicated no differential expression or exon splicing among males and females, although a potential splice variant of tra was observed in Daphnia galeata that did show some degree of female bias (Huylmans et al., 2016). Finally, the doublesex (Dsx) gene (DapmaDsx1) has been observed to control D. magna sex determination, where increased expression was associated with male formation (Kato et al., 2011). In the present study, Dsx probes were included in the microarray, but were not differentially expressed in response to the short-day length photoperiods.

| Cuticle and chitin (development, moulting, reproduction and more)
Chitin is an extracellular polysaccharide matrix that represents the predominant structural component of arthropod exoskeletons (Tetreau et al., 2015). The direction and magnitude of expression for multiple components of chitin metabolism and cuticle structure (chitin plus proteins) in D. magna varied in response to photoperiod (

| Egg
Given that neonate production was decreased by roughly 12% in the most extreme short-day length photoperiod, we investigated transcriptional expression relevant in egg productions to provide potential mechanistic insights. Transcript expression for genes coding vitelline membrane proteins (JAN90603, KZS10607) was decreased in the short-day length photoperiods (Table 1). The vitelline membrane has been demonstrated to act as the reservoir for chorion proteins during early developmental stages of Drosophila egg development (Pascucci, Perrino, Mahowald, & Waring, 1996). Chorion peroxidase precursors (JAN29198, JAN477987) also had decreased expression in short-day length photoperiods (Table 1). Mindrinos, Petri, Galanopoulos, Lombard, and Margaritis (1980) demonstrated that chorion peroxidase activity facilitates cross-linking of chorionic proteins during the "hardening" phase of egg development in Drosophila. In contrast to the decreased expression of transcripts related to the egg "shell," the expression of vitellogenin fused with the superoxide dismutase (BAE94324) was increased in the short-day length photoperiods (Table 1). This gene complex (vitellogenin fused to the superoxide dismutase-like domain) in D. magna represents a major component of yolk proteins in eggs (Kato, Tokishita, Ohta, & Yamagata, 2004) where vitellogenesis is recognized as a critical determinant of Daphnia egg production rate and quality (Ebert, 1993). Additionally, decreased transcriptional expression of the female sterile 2 (KZS06734) gene was observed in D. magna in the short-day length photoperiods (Table 1)

. Female sterile mutations in
Drosophila have been observed to impair reproduction by inhibiting egg production by blocking oogenesis or by impairing egg morphology (Schüpbach & Wieschaus, 1991). Finally, notch (DM03056) expression was increased in short-day length photoperiod exposures (Table 1), where notch is critical in Daphnia segmentation and overall organismal development (Eriksson, Ungerer, & Stollewerk, 2013). Overall, a variety of functions critical in egg production and development were differentially expressed at the transcriptional level in response to short-day length photoperiod. Given changes in energy expenditure (Figure 2), male production ( Figure 1) and overall neonate production ( Figure 1) in short-day length and equinox photoperiods, shifts in the expression of genes underlying reproductive physiology provide insights into processes influenced/affected by photoperiod.

| Visual cues
There were significant, albeit low relative decreases in transcriptional expression of blue wavelength opsin (KZS21495) and class a rhodopsin G-protein coupled receptor 2 (JAN36420) in the short-day length photoperiods. Rhodopsin dynamics in Anopheles gambiae mosquitoes were observed in response to light/dark exposures, where the "mature" rhodopsin protein tended to increase in concentration within the major R1-6 photoreceptor class of the mosquito eye in response to dark exposure (Moon, Metoxen, Leming, Whaley, & O'Tousa, 2014).
In the present study, it is plausible that dynamics in rhodopsin accumulation in light/dark cycles could have implications related to phototaxis, including the observation of intensified light aversion (Figure 3).

| CON CLUS IONS
In Daphnia magna, photoperiod as a sole environmental cue had powerful influence over phototaxis behaviours and multiple life history outcomes where functional genomic analysis identified a variety of logical connections to potential underlying molecular mechanisms (a synthesis of connections is provided in Table 2).
Entrainment to short-day photoperiods caused D. magna to have significantly increased light-avoidance behaviours relative to conspecifics reared in the control (long-day photoperiod -16L:8D) where young Daphnia (7 days old) displayed the most pronounced light-avoidance responses. Functional transcriptomic analysis identified photoperiod-induced differential expression of gene targets involved in glutamate signalling, which is critical in arthropod light-avoidance responses, as well as period circadian protein and proteins coding F-box/LRR-repeat domains, all of which contribute to establishing circadian rhythms. Short-day photoperiods also produced indications of increased metabolic rate during light hours which corresponded with broadscale changes in expression across multiple pathways that establish system-level energy metabolism. A striking life history response to photoperiod was increased male production across short-day and equinox photoperiods (4L:20D, 8L:16D and 12L:12D). Correspondingly, transcriptional expression consistent with multiple putative mechanisms of male production was observed including photoperiod-dependent transcriptional expression of transformer-2 sex-determining protein (Lopez, 1998) and associated small nuclear ribonucleoprotein particles (snRNPs) involved in splice variant expression for genes such as transformer. Additionally, changes in transcriptional expression that were suggestive of increased glutamate signalling show parallelism to responses observed to induce male production in Daphnia pulex (Toyota, Miyakawa, Yamaguchi, et al., 2015) via photoperiod-sensitive mechanisms (Camp & Haeba 2019;Toyota, Miyakawa, Hiruta, et al., 2015). Overall, the results demonstrate the importance of photoperiod on behaviour and life history trajectories in D. magna where we have now established multiple putative mechanistic pathways underlying these critical responses.

ACK N OWLED G EM ENTS
Opinions, interpretations, conclusions and recommendations are those of the author(s) and are not necessarily endorsed by the U.S.
Army. This work was supported by the US Army's 6.1 Basic Research Program, Environmental Quality and Installations, Project 14-44, under Technical Director, Dr. Elizabeth Ferguson.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest or relationship, financial or otherwise that might be perceived as influencing the authors' objectivity in the reporting of the study results or interpretations.

AUTH O R CO NTR I B UTI O N S
The study idea was conceived by K.G., A.K., M.K. and T.S. The

DATA AVA I L A B I L I T Y S TAT E M E N T
All microarray data/results are publicly available at the Gene