Endogenous melatonin promotes rhythmic recruitment of neutrophils toward an injury in zebrafish

Neutrophil recruitment to injured tissue appears to be an evolutionarily conserved strategy for organisms to fight against exogenous insults. Recent studies have shown rhythmic migration of neutrophils and several factors, including melatonin, have been implicated in regulating this rhythmic migration. The mechanisms underlying how endogenous melatonin regulates rhythmic neutrophils migration, however, are unclear. Here we generated a zebrafish annat2 mutant that lacks endogenous melatonin and, subsequently, a Tg(lyz:EGFP);aanat2−/− transgenic line that allows for monitoring neutrophils migration visually in live zebrafish. We observed that migrating neutrophils are significantly reduced in aanat2−/− mutant zebrafish under a light/dark condition, and the disrupted migrating rhythmicity of neutrophils in aanat2−/− zebrafish is independent of the circadian clock. Further, we also found that endogenous melatonin enhances neutrophils migration likely by inducing the expression of cytokines such as interleukin-8 and interleukin-1β. Together, our findings provide evidence that endogenous melatonin promotes rhythmic migration of neutrophils through cytokines in zebrafish.

Zebrafish have emerged as a powerful model for studying innate immune functions, as their embryos are transparent, allowing for real-time visualization of fluorescent proteins at the single-cell level in vivo 3 . As a diurnal animal, zebrafish have been attractive for studying the circadian rhythm because of their conserved clock mechanisms shared with mammals 29 . In this study, we generated an annat2 mutant zebrafish line with CRISPR-Cas9, and also the Tg(lyz:EGFP);aanat2 −/− zebrafish line. Through visualizing rhythmic migration of neutrophils in vivo, we revealed that endogenous melatonin regulates this rhythmic neutrophils migration. Although normal oscillations of circadian clock genes are disrupted in aanat2 −/− fish, reduced neutrophils migration still persists in aanat2 −/− larvae. We also found that endogenous melatonin enhances neutrophils migration likely through inducing the expression of cytokines. These results demonstrated that endogenous melatonin participates in the regulation of neutrophil rhythmic migration in zebrafish.

Materials and Methods
Zebrafish lines and maintenance. Zebrafish embryos were harvested from natural matings of wild-type (AB), transgenic Tg(lyz:EGFP) labeled with neutrophils and aanat2 −/− lines. The Tg(lyz:EGFP);aanat2 −/− line was obtained by crossing aanat2 −/− and Tg(lyz:EGFP) for two consecutive generations. Embryos were maintained in 14/10 light/dark (LD) conditions at 28.5 °C. N-phenylthiourea (PTU, Sigma, USA) was used to prevent pigment formation. All animal manipulations were conducted in strict accordance with the guidelines and regulations set forth by the University of Science and Technology of China (USTC) Animal Resources Center and University Animal Care and Use Committee. The protocol was approved by the Committee on the Ethics of Animal Experiments of the USTC (Permit Number: USTCACUC1103013). All zebrafish surgeries were performed after anesthetization with Tricaine methane-sulfonate (MS-222, Sigma) treatment.
Design of the CRISPR-Cas9-targeted site and synthesis of Cas9 and gRNA. The gRNA was designed to target the first exon of zebrafish aanat2 by "seqbuilder" software (DNAStar, USA) according to the 5′-GGNNNNNNNNNNNNNNNNNNNG G-3′ form 30 . The targetingsequence started with GG, ended with NGG (PAM) and also contained a restrictive enzyme MspI near the PAM for genotyping. The Cas9 mRNA and gRNA were synthesized as described previously with modification 31 . Briefly, the Cas9 mRNA was synthesized using a T7 mMESSAGE mMACHINE Kit (Ambion, USA). The DNA fragment of the gRNA was amplified by PCR with a pair of primers (Table 1), and then purified by phenol and chloroform. The gRNA was in vitro transcribed with SP6 Riboprobe Systems (Promega, USA).

Analysis of mutagenesis frequencies and identification of aanat2 mutants. Cas9 mRNA and
aanat2 gRNA were co-microinjected into one-cell zebrafish embryos. Genomic DNAs of three groups (15 embryos each), including wild-type controls, were extracted at 24 hours post-fertilization (hpf), and used as templates for PCR. A 290-bp DNA fragment containing the aanat2 target fragment was PCR amplified, digested with MspI (New England Biolabs, UK) at 37 °C for 3 h, and electrophoresed on a 3% agarose gel. Intensities of cleaved and uncleaved bands were quantified with Image J software (NIH, USA). The uncleaved bands were recovered after gel electrophoresis and cloned into pMD-19T (Takara, Japan), and single clones were picked up for PCR and restriction enzyme digestion, and then sequenced by Sanger sequencing (GENEWIZ, Inc.). Primers used in the experiment are listed in Table 1.
The microinjected founder (F 0 ) embryos were raised to adulthood and then crossed with wild-type zebrafish to produce F 1 embryos. From each cross, 15 F 1 embryos were collected for genomic DNA extraction and enzymatic digestion. The F 1 embryos that carry heritable mutations were raised to adulthood, and then each individual  Table 1. Primers used in the experiment. F 1 fish was identified with PCR amplication of fin-clipped DNAs and enzymatic digestion. Homozygous aanat2 mutant fish were generated by crossing of the male and female fish carrying the same mutation.
Tail fin injury and live imaging. Zebrafish larval tail fin was transected at the end of the spinal cord by a sterile blade on a plastic petri dish after being anesthetized with Tricaine methane-sulfonate 22 (Sigma, USA). The injured larvae were recovered at 28.5 °C in embryo medium until live imaging. Three hours after wounding, transgenic larvae Tg(lyz:EGFP) labeled with neutrophils as well as Tg(lyz:EGFP);aanat2 −/− were embedded into low-melting agarose for visual monitoring under an Olympus microscope with a green fluorescent channel.
Melatonin measurement by ELISA. Melatonin concentrations were detected with an ELISA kit 18 (IBL international, Germany). Wild-type and aanat2 mutant larvae were raised under 14/10 LD for 5 days. At 12:00 and 24:00, fifty larvae were collected for ELISA evaluation. Melatonin samples were extracted with a column with methanol according to the kit instructions. The standard curve was generated with a series of concentrations of melatonin. Melatonin content was determined by measuring the optical density with a photometer at 405 nm within 60 min after pipetting of the stop solution.
RNA extraction and qRT-PCR. Total RNAs were extracted from larvae of wild-type (n = 50) and aanat2 −/− (n = 50) at 4-h intervals under LD and DD (dark-dark) conditions using Trizol (Takara, Japan) reagent. Larvae from the cloacal orifice to the incision end were also collected for RNA extraction to examine cytokine expression. Quantitative real-time PCR (qRT-PCR) was conducted with the SYBR green (Invitrogen, USA) system. The clock and cytokine genes were amplified using the profiles of 95 °C, 10 s, 60 °C, 30 s for 40 cycles. qRT-PCR was performed in triplicate with three individual biological samples (nine replicates) at corresponding time points, and the results were normalized to the expression level of the housekeeping gene β-actin and shown as a relative expression level calculated using the2 −ΔΔCt method 32 . P values were analyzed with one-way analysis of variance (ANOVA) test or Student's t test.

Rescue of melatonin content, neutrophils migration and clock gene expression by capped
wild-type aanat2 mRNAs. The zebrafish wild type cDNA and anti-sense cDNA of aanat2 were cloned into the pCS2+ plasmid with the restriction enzyme BamHI and EcoRI, and linearized with SacII. Capped aanat2 mRNAs and anti-sense mRNAs were transcribed from the linearized plasmids using the mMACHINE in-vitro transcription kit (SP6; Ambion, Austin, TX, USA) according to the manufacturer's instructions. To be specific, 2 μl 10X reaction buffer, 2 μl enzyme mix, 10 μl 2X NTP/CAP, 1 μl RNase Inhibitor and 1 μg plasmid DNA were mixed, and nuclease-free water was added to the mix solution up to 20 μl and incubate at 37 °C for 2 hr. mRNAs identification were using DNA gel electrophoresis and the content of mRNAs were measured using the spectrophotometer (Nanodrop2000, Thermo). 200 ng/μl aanat2 capped mRNAs or 200 ng/μl anti-sense capped mRNAs were microinjected into one-cell of zebrafish wild type, aanat2 −/− embryos or Tg(lyz:EGFP);aanat2 −/− embryos. None microinjected embryos of wild type, aanat2 −/− or Tg(lyz:EGFP);aanat2 −/− were as controls. Total RNAs were extracted from 50 larvae of ZT12 each sample.

Statistical analysis.
All experiments were independently repeated three times. The data were analyzed with an unpaired, two-tailed t-test, one-way ANOVA using GraphPad Prism version 5.0 (Prism, USA). The results are shown as the mean ± SEM. The level of significance was set to P < 0.05. *, **, and *** represent P < 0.05, P < 0.01, and P < 0.001, respectively.

Results
Generation of aanat2 −/− zebrafish using CRISPR-Cas9. Our previous study showed that migrating neutrophils display a robust daily rhythm 17 . To further investigate the role of endogenous melatonin in rhythmic neutrophils migration, we generated aanat2 mutant lines. Aannat2 is the rate-limiting enzyme for melatonin synthesis in vertebrates 33,34 . Using "Seqbuilder" software, we designed a CRISPR-Cas9-targeted site in the first exon of zebrafish aanat2, which also contains an MspI restriction site for evaluating and screening mutants (Fig. 1A). In vitro synthesized capped Cas9 mRNAs and gRNA were then microinjected simultaneously into one-cell embryos. To evaluate the mutant efficiency, a 290-bp targeted DNA fragment was PCR amplified and digested with the MspI restriction enzyme. Results showed that the mutant efficiency was approximately 32-51% in F 0 larvae (Fig. 1B). Representative sequencing results of the two mutated fish lines showed that one had a 9-bp insertion and 1-bp deletion, the other had a 14-bp deletion, and both lines had frame-shift mutations (Fig. 1C). The mutant lines with the 9-bp insertion and 1-bp deletion were used in the study.
Establishment of a visual model monitoring neutrophils migration in aanat2 mutant zebrafish. To monitor neutrophils migration in live zebrafish in a real-time manner, we crossed transgenic line Tg(lyz:EGFP) labeling neutrophils with aanat2 mutant zebrafish ( Fig. 2A). Then we screened the homozygous aanat2 mutant whose neutrophils are also labeled with green fluorescent protein, called Tg(lyz:EGFP);aanat2 −/− fish, using enzymatic digestion of fin-clipped DNAs (Fig. 2B). We also determined the melatonin content in Tg(lyz:EGFP);aanat2 −/− larvae with ELISA. Results showed that melatonin is at a very low concentration in aanat2 −/− larvae during day and night, while melatonin is dramatically increased in wild-type larvae during the night. The melatonin content could be partly rescued by aanat2 capped mRNA treatment (Fig. 2C). Moreover, the aanat2 mutation did not cause a change in zebrafish body weight and length ( Supplementary Fig. S1A,B). Hence, neutrophils migration can be visualized in endogenous melatonin-deficient zebrafish.
Neutrophil recruitment is significantly reduced in aanat2 −/− larvae. Using the injury-induced inflammation model 6, 35 , we examined the effects of endogenous melatonin on rhythmic neutrophil recruitment.  Results showed that neutrophil recruitment towards the wound site is significantly reduced in aanat2 −/− larvae (Fig. 3A,B). Given the possibility that aanat2 −/− may affect the total number of circulating neutrophils, we evaluated the fluorescent neutrophils in Tg(lyz:EGFP) zebrafish. Considering the difficulty in counting all neutrophils, neutrophils located in the 800-μm region from the spinal cord to the anterior were counted and regarded as the total number, as in a previous study 22 . Results showed that there was no significant difference in the total number of neutrophils between wild-type and aanat2 −/− larvae (Fig. 3C). Results also showed that the neutrophil distribution in the control and mutant groups had no significant difference ( Supplementary Fig. S1C). We also found that the rhythmic patterns of neutrophils migration were abolished in aanat2 −/− mutants (Fig. 3D), consistent with our previous prediction that the rhythmic mode of neutrophils migration is likely disrupted in aanat2 mutants 17 .
The migrating neutrophils could be partly rescued by injection of aanat2 capped mRNA ( Supplementary Fig. S2). Further, the average number of migrating neutrophil was lower in aanat2 −/− larvae than in wild types during day and night (Fig. 3E). These results clearly indicated that loss of endogenous melatonin results in reduction of neutrophil recruitment and alters rhythmic neutrophils migration in zebrafish.

Disrupted expression of circadian clock genes in aanat2 mutant larvae. Previous studies have
shown that exogenous melatonin treatment could alter rhythmic expression of circadian clock genes [36][37][38][39] . Here, we examined whether endogenous melatonin deficiency in aanat2 −/− could affect the expression of circadian clock genes. qRT-PCR analyses show that clock1a, bmal1a, bmal1b and per1b were significantly down-regulated in aanat2 −/− larvae under the LD condition ( Fig. 4A-C). While per2 was up-regulated in aanat2 −/− larvae at ZT4 and ZT12.and cry1bb exhibited a phase advance of expression in aanat2 −/− larvae (Fig. 4E,F). Further, we synthesized wild-type aanat2 capped mRNAs and microinjected into the mutants to rescue the gene expression at ZT12 as most genes expression was altered at this time point. The rescue assay revealed relative mRNA expression of clock1a and bmal1b could be rescued by injection of aanat2 capped mRNA, and relative mRNA expression of cry1bb could be partly rescued ( Supplementary Fig. S3). Together, these results showed that expression of key circadian clock genes is disrupted in aanat2 mutant zebrafish.
Disrupted migrating rhythmicity of neutrophils in aanat2 −/− zebrafish is independent of the circadian clock. Because of altered expression of key circadian clock genes in aanat2 −/− zebrafish, we wondered whether the regulatory role of endogenous melatonin in rhythmic neutrophils migration is a direct effect, or is mediated through a disturbed circadian clock. We raised aanat2 −/− embryos at the onset of fertilization under constant darkness (DD), and established arrhythmic zebrafish lacking molecular circadian rhythms 40 . Results showed that oscillating patterns of clock genes per1b, per2, cry1bb, bmal1a, bmal1b and clock1a were abolished and almost all genes were expressed similarly at all circadian time points in these arrhythmic zebrafish larvae (Fig. 5A-F). These results showed that the circadian clock machinery did not function appropriately in these larvae raised continuously from the onset of fertilization under DD. Interestingly, we also observed that, under the same condition, neutrophil recruitment is still significantly reduced in aanat2 −/− larvae in a 24-h period (Fig. 5G), and the number of migrating neutrophil averages was lower in aanat2 −/− larvae than wild-type controls during day and night under the DD condition (Fig. 5H). Together with our previous study that showed treatment with melatonin at a physiological concentration promotes neutrophils migration in zebrafish 17 , our findings suggest that melatonin promotes neutrophils recruitment rather than through the circadian clock in zebrafish.
Down-regulation of cytokines in aanat2 −/− larvae. Exogenous melatonin also was shown to alter expression of cytokines 24 . We hypothesized that reduced neutrophil recruitment in aanat2 mutants may be meditated through cytokines. Cytokines were induced using the injury model in darkness as previously described 4 . qRT-PCR results showed that both il-8 and il-1β were significantly down-regulated in aanat2 mutants (Fig. 6A,B), which could be rescued by aanat2 capped mRNA injection. The results also showed that il-8 and il-1β exhibited obvious increase in the night period compared with the day period ( Supplementary Fig. S4). Both tnf-α and il-6 are barely changed in aanat2 −/− larvae (Fig. 6C,D). Similar results also appeared when embryos were raised in DD (data not shown). Previous studies have visually demonstrated that IL-8 directly induces neutrophil recruitment with the eye and optic vesicle model in zebrafish larvae 4,17 . Taken together, our results implied that endogenous melatonin might promote neutrophils migration by regulating IL-8 expression.

Discussion
Melatonin has been widely regarded as a regulator of inflammation and circadian rhythms [41][42][43][44] . Its role in regulating inflammation and circadian rhythms, however, is controversial 24 , and particularly the function of endogenous melatonin is largely uncertain. Here we generated a melatonin-deficient model in diurnal zebrafish (Fig. 1) and we raised the mutant zebrafish for several generations before conducting all experiments to reduce the off-target effect. In vivo imaging of transgenic lyz:EGFP;aanat2 −/− larvae using the injury model showed that neutrophils migration is reduced and its migration rhythmicity is abolished in aanat2 −/− larvae (Fig. 3). Because rhythmic expression of key circadian clock genes is disrupted in aanat2 −/− larvae (Fig. 4), the enhancing effect of melatonin on neutrophils migration may be independent of circadian regulation. To test this hypothesis, we generated arrhythmic larvae by raising them immediately at the onset of fertilization under constant darkness (DD), wherein all key circadian clock genes lost their rhythmic expression (Fig. 5). Intriguingly, under this treatment, neutrophils migration is still reduced in these arrhythmic aanat2 −/− larvae from day to night (Fig. 5), suggesting that endogenous melatonin directly regulates neutrophils migration, rather than through modulation of the circadian clock.
Cytokines have been shown to be rhythmically expressed 8,45 and exogenous melatonin can modulate their expression 42,44,46 . In this study, we found that both il-1β and il-8 are significantly down-regulated in aanat2 −/− larvae (Fig. 6), both of which have been shown to be able to attract neutrophils migration in zebrafish 4, 47 . These qRT-PCR analysis showed that key circadian clock genes clock1a, bmal1a, bmal1b, per1a, per1b, and cry1bb were all expressed at the same level without rhythmicity. Each sample contained 50 embryos. The data were analyzed from three independent samples in both the control and mutant groups. The experiment was repeated three times. (G) Neutrophil migration towards the injury is reduced in aanat2 mutants under the same condition (control, n = 40; mutant, n = 40; mutant + mRNA, n = 40). The experiment was repeated three times. (H) The average number of migrating neutrophils was lower in aanat2 −/− larvae than wild types during day and night (control, n = 120; mutant, n = 120) (unpaired t-test). The data was analyzed from Fig. 5G. The "day" and "night" in constantly-dark are consistent with the time-cycle in nomal light cycle. (**P < 0.01, ***P < 0.001, unpaired t-test and ANOVA analysis). results implied that the promoting effect of endogenous melatonin on neutrophils migration may be mediated, at least in part, by cytokine signaling, although this hypothesis should be investigated in detail in the future.
Melatonin is known to play roles in the circadian clock and peripheral immune system 39,48 . While a large number of studies have employed nocturnal animals such as mice or rats to investigate effects of melatonin on immune functions, only a few have used a diurnal model, such as zebrafish, to explore roles of endogenous melatonin on immune functions. Here we showed that endogenous melatonin modulates rhythmic neutrophils migration in diurnal zebrafish. Reduced melatonin levels often occur in the elderly and related patients 49 , implying that reduced endogenous melatonin may contribute partially to the subdued immune functions in the elderly and related patients. A deep understanding of how endogenous melatonin interacts with the immune system and other physiological functions using the aanat2 −/− zebrafish may set the stage for developing novel therapies for the elderly and related patients. were collected for RNA extraction at 1.5 h after injury in the darkness. qRT-PCR analysis showed significantly down-regulation of il-1β and il-8, which could be partly rescued by injection of aanat2 capped mRNA (one way ANOVA analysis). However, tnf-α and il-6 expressions showed no significant difference between the aanat2 −/− and control groups (unpaired t-test). Each independent sample contained fifty embryos. The data were analyzed using three samples in both the control and mutant groups. The experiment was repeated three times. (*P < 0.05, **P < 0.01).