Effects of toxic Microcystis aeruginosa on the expression of Hox genes in Daphnia similoides sinensis

Abstract Lake eutrophication and cyanobacterial blooms have become worldwide environmental issues. Under cyanobacterial blooms (especially Microcystis), Daphnia spp. can transfer beneficial information to their offspring in order to improve adaptability. Hox genes are important regulatory factors of transcription in metazoans, and are involved in the growth and development of organisms. However, the mechanisms of Microcystis on the expression of Hox genes in Daphnia are unclear. In this study, the effects of Microcystis aeruginosa on Hox gene expression in the mothers and offspring (F1) of two Daphnia similoides sinensis clones were investigated using a mixed diet of M. aeruginosa and Scenedesmus obliquus. Compared with the 100%S food treatment, the survival rates at the end of the experiment of clone 1‐F1 in the food treatments containing M. aeruginosa were significantly lower, but it was significantly higher for clone 2‐F1 in the 20%M + 80%S food treatment. Moreover, the survival rates at the end of the experiment of clone 1‐F1 in the food treatments containing M. aeruginosa were significantly higher than those of their mother. Based on previous transcriptome data, 14 Hox genes of D. similoides sinensis were identified, including Abd‐B, CDX‐1, Dll, HOX‐1, HOX‐2, HOXA1, HOXA2, HOXB3, HOXB3‐2, HOXB7, HOXC4, HOXC7, HOXC8, and HOXD10. The expressions of Abd‐B, HOX‐2, HOXA1, HOXC7, and HOXD10 of clone 2‐mothers in the 40%M + 60%S food treatment were 2.9–22.5 times as high as in the 100%S food treatment, whereas the expressions of CDX‐1, HOX‐1, HOXB3, and HOXD10 of clone 1‐mothers were 4.8–13.1 times at same food level. The expression of HOXA2, HOXC7, HOXC8, and HOXD10 of clone 1‐F1 in the 40%M + 60%S food treatment was 8.2–21.1 times as high as in the 100%S food treatment. However, compared with the 100%S food treatment, the expressions of CDX‐1 in the mothers and F1 of clone 2 and HOXB7 in the mothers of clone 1 in the food treatments containing M. aeruginosa were significantly lower (p < .05). Our results suggest that the offspring (F1) produced by D. similoides sinensis mother pre‐exposed to toxic M. aeruginosa had stronger adaptability to M. aeruginosa than their mothers. Moreover, Hox gene expressions of D. similoides sinensis had obvious differences between clones under stress of toxic M. aeruginosa.


| INTRODUC TI ON
Hox genes are important regulatory factors of transcription in metazoan animals and comprise a large family of highly conserved DNA transcription factors (Affolter et al., 1990). In vertebrates, the Hox gene family is often displayed in multiple cluster form, and participates in the regulation of embryonic development and morphological diversity (Krumlauf, 1994;McGinnis & Krumlauf, 1992). In metazoans, the target sites of the Hox gene homology domain are connected with specific DNA sequences (Affolter et al., 1990), which can regulate cell fates (Batas, 1993) and affect cell recognition via genetic address (Lawrence, 1992;Lawrence & Morata, 1983).
Hox genes were first identified in Drosophila melanogaster (McGinnis et al., 1984;Scott & Weiner, 1984), and Papillon and Telford (2007) studied the expression and evolution models of Hox3 and ftz genes in Daphnia pulex.
Animal mothers can transfer environmental information to their offspring so that their offspring can produce adaptive responses to environmental heterogeneity in terms of phenotype, physiology, behavior, and reproduction (Agrawal et al., 1999;Frost et al., 2010;Mousseau & Fox, 1998). In birds, lizards, insects, and crustaceans, maternal effects play an important role in their population adaptation to the environment (Badyaev et al., 2002;Mousseau & Dingle, 1991;Schwarzenberger & Elert, 2013;Uller, 2004). Boersma et al. (2000) observed that large-sized Daphnia magna could produce larger offspring as well as produce larger ephippia in order to improve their hatching rates. D. magna can improve net reproduction efficiency and fitness of their offspring after short-term exposure to the pesticide fenvalerate (Pieters & Liess, 2006). Furthermore, Badyaev (2008) found that the adaptability of a passerine bird to the environment obtained through maternal effects could be preserved for a long time before genetic evolution took place.
In recent decades, cyanobacterial blooms by species such as M. aeruginosa have become more frequent and severe in lakes due to eutrophication, leading to suppressed population dynamics of various Daphnia species (Deng et al., 2008;Hansson et al., 2007;Liess & Hillebrand, 2004;Przytulska et al., 2015). Cyanobacteria often release toxins such as microcystin (MC) which inhibits protein phosphorylation, affects physiological metabolism, and changes chromosomal structure, resulting in genotoxicity (Lankoff et al., 2004;Peng et al., 2018;Zegura et al., 2003). Microcystin (MC) can be accumulated in consumers through the food chain and can even affect human health (Christoffersen, 1996;Gilroy et al., 2000;Jorgensen, 1999;Reynolds, 1994). Usually, M. aeruginosa has an inhibitory effect on the life-history traits of Daphnia species (Gustafsson & Hansson, 2004;Jiang et al., 2013;Li et al., 2014;Lyu, Meng, et al., 2016;Yang et al., 2011). However, some studies have indicated that single-cell or small-colony Microcystis spp. can be fed by Daphnia spp. to favor their growth and reproduction (Chen & Xie, 2003;Hanazato, 1991;Li et al., 2014). Other studies have even shown that the offspring of Daphnia species can obtain more adaptability to toxic M. aeruginosa via maternal effect (Lyu, Guan, et al., 2016;Lyu et al., 2017). In Daphnia carinata, the offspring of the mothers pre-exposed to M. aeruginosa had quicker defensive responses than did their mothers previously unexposed to M. aeruginosa (Jiang et al., 2013). Gustafsson et al. (2005) found that the offspring of D. magna pre-exposed to M. aeruginosa had shorter time to maturation and a greater number of offspring. Schwarzenberger et al. (2009) observed that the offspring produced by the mothers pre-exposed to M. aeruginosa up-regulated the expression of target genes in D. magna, and suggested that the maternal effect was a short-term adjustment strategy to the environment.
In summary, M. aeruginosa could affect life-history traits and expression levels of some genes in Daphnia, but it was unknown how toxic M. aeruginosa affected the expression levels of Hox genes in Daphnia species and whether these genes of their offspring from the mother pre-exposed by M. aeruginosa had the adaptability to toxic M. aeruginosa. 14 Hox genes have been identified in D. similoides sinensis based on previous transcriptome data , including Abd-B, Dll,HOXA1,HOXA2,HOXB3,HOXB7,HOXC4,HOXC7,HOXC8,and HOXD10. In this paper, our goal is to compare the influences of M. aeruginosa on Hox genes of mothers and F1 in two D. similoides sinensis clones, and to examine the adaptability of F1 from pre-exposed mothers to toxic M. aeruginosa and the differences between two clones.

| Collection, identification, and culture of D. similoides sinensis
Lake sediment from the 0-to 1-cm layer was collected from Lake Junshan in Jiangxi province (28°9′41″-28°46′13″N, 116°1′15″-116°33′38″E) in August 2015 using an 8.4-cm-diameter columnar gravity corer (Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences). The sediment was washed using 200 mesh (0.074 mm) in the laboratory, and the residue was examined using a microscope (Olympus, Japan) in order to identify the ephippia of D. similoides sinensis according to the methods of Benzie (2005) and Gu et al. (2013). Ephippia containing resting eggs of D. similoides sinensis were incubated at 25 ± 1 °C in K E Y W O R D S Daphnia similoides sinensis, Hox genes, maternal effects, Microcystis aeruginosa

T A X O N O M Y C L A S S I F I C A T I O N
Ecotoxicology aerated tap water in an intelligent light incubator (Saifu, Ningbo, China). S. obliquus, a nontoxic microalgae species, was used as a food source.

| Culture of M. aeruginosa and S. obliquus
Microcystis aeruginosa was obtained from Lake Junshan in August 2015. A single colony of M. aeruginosa was chosen in the laboratory, and then cultured in BG-11 medium in an intelligent light shaker incubator (QZB-98B, China) at (28 ± 1) °C with illumination of a 12:12 h light/dark cycle. M. aeruginosa which were single or two cells in morphology were collected at the exponential phase of population growth and stored at 4°C.

Scenedesmus obliquus was obtained from the Freshwater Algae
Culture Collection (Institute of Hydrobiology, Chinese Academy of Sciences), and cultured in BG-11 medium in an intelligent light incubator (Saifu, Ningbo, China) at 25°C, with a 12:12 h light/dark cycle, then collected at the exponential phase of population growth and stored at 4°C.

| D. similoides sinensis mother experiment
Two D. similoides sinensis ephippia containing resting eggs were randomly selected, and then hatched in a 50-ml beaker in an intelligent light incubator at 25°C with a 12:12 light/dark cycle, respectively. The individual hatched from each ephippium containing resting eggs represented one clone, and each clone was respectively cultured through parthenogenesis. Two clones from different resting eggs were employed in the experiment. Third generation youngs (<12 h old) produced by each clone were used as experimental animals in the mother experiment. Three food treatments were designed based on biomass content: 100% S. obliquus (100% S), serving as a control, 20% M. aeruginosa + 80% S. obliquus (20% M + 80% S), and 40% M. aeruginosa + 60% S. obliquus (40% M + 60% S). The total biomass of each food treatment was 40 mg/L wet weight. There were three replicates in each food treatment, yielding a total of 18 experimental groups (2 clones × 3 food treatments ×3 replicates). At the beginning of the experiment, 20 young females (<12 h old) at third generation were randomly placed in each 250-ml beaker. The culture medium was 200 ml aerated tap water (over 48 h). Therefore, 180 youngs were employed for each clone in the mother experiment. The experiments were carried out in an intelligent light shaker incubator (QZB-98B, China) at (25 ± 1) °C and 12:12 light/dark cycle. All neonates produced by the mothers in each 250-ml beaker were promptly removed during the experiment. The survival rates of the mothers were calculated daily and lasted at the end of the 14-day experiment. The culture medium was replaced every two days before D. similoides sinensis mothers became pregnant, from which point on it was replaced daily. The cultural density (20 young females) of D. similoides sinensis and temperature (25°C) in this experiment are according to our previous experimental designs Xu et al., 2018).
On the fourteenth day, 12-h-old neonates produced by the mother in the 20% M + 80% S food treatment were removed and placed in new 250-ml beakers for an offspring (F1) experiment. At the end of the mother experiment, all D. similoides sinensis mothers in each food treatment were pooled into an EP tube and stored in liquid nitrogen for later measurement of Hox genes.

| D. similoides sinensis F1 experiment
In the mother experiment, owing to fewer offspring produced in the 40% M + 80% S food treatment, the offspring (<12 h old, F1) produced by the mothers of two D. similoides sinensis clones in only the 20% M + 80% S food treatment on the fourteenth day were collected and regarded as experimental animals in the F1 experiment, and 180 individuals (F1) in each clone were employed. The F1 experimental designs were the same as described in the mother experiment. After 14 days, all F 1 females in each food treatment were pooled into an EP tube and stored in liquid nitrogen for later measurement of Hox genes. Fluorescence was measured using a melting curve from 55°C to 95°C in order to detect single gene-specific peaks and primer-dimer peaks.
Three biological replicates were used for each sample.

| Gene identification and sequence analyses
The homologous genes were searched and compared in NCBI (https://www.ncbi.nlm.nih.gov/). Reading frames and functional domains based on the complete sequence information of these homologous genes were predicted using the ORF Finder (https://www. ncbi.nlm.nih.gov/orffi nder/) from the NCBI database. Sequence alignment, similarity, and homology analyses were performed using BLASTX (https://blast.ncbi.nlm.nih.gov/Blast.Cgi) and ClustalX.

| Statistical analyses
All statistical analyses were performed using SPSS 20.0. Two-way ANOVA was employed to analyze the influences of food treatment, mother-F1 generation, and their combinations on the survival rates at the end of the experiment and each Hox gene expression of each D. similoides sinensis clone. For each clone, multiple comparisons (Tukey's HSD) were also used to test the differences of the survival rates at the end of the experiment and each Hox gene expression of both mothers and F1 among different food treatments, respectively.

| Survival rates of two D. similoides sinensis clones under different food treatments
The survival rates of the mothers and F1 in clone 1 showed a gradual dropping trend with the increasing of M. aeruginosa concentration.
However, it was an opposite pattern in clone 2 ( Figure 1).
For clone 1, both food treatment and mother-F1 generation af- For clone 2, food treatment affected significantly the survival rates at the end of the experiment (F = 7.600, p = .007), but both mother-F1 generation and their combinations of food treatment and mother-F1 generation had no significant effects (mother-F1 generation: F = 0.400, p = .539; their combinations: F = 0.400, p = .679).
Multiple comparisons (Tukey's HSD) showed that the survival rates at the end of the experiment of F1 in the 20%M + 80%S food treatment were significantly higher than those in the 100%S food treatment (p = .0128).

TA B L E 1
The qRT-PCR primer sequences of D. similoides sinensis in the experiment Name Sequence Name Sequence  Note: Bold values indicates p < .05 is significant; p < .01 is very significant.

TA B L E 3 (Continued)
HOXD10 were respectively clustered into different clades with orthologs in other species (Figure 2).

| Hox gene expression in the mothers and F1 of two D. similoides sinensis clones under different food treatments
For clone 1, food treatment and mother-F1 generation affected significantly the relative expression of CDX-1, HOX-1, HOXA2, HOXB3, HOXB3-2, HOXB7, HOXC8, HOXD10 genes as well as their combinations (Table 3). Moreover, both food treatment and mother-F1 generation affected significantly the relative expression of HOXC7 gene (Table 3). In clone 1-mothers, compared to that in the 100%S food HSD) showed that the expressions of HOXA2, HOXC7, HOXC8, and HOXD10 in the 40%M+60%S food treatment were significantly higher than those in the 100% S food treatment (p < .05). In addition, the expressions of HOXA2, HOXB7, HOXC7, HOXC8, and HOXD10 in the 40%M+60%S food treatment were significantly higher than those in the 20%M+80%S food treatment (p < .05).
For clone 2, food treatment and mother-F1 generation affected significantly the relative expressions of HOXA1, HOXB3, HOXB3-2, HOXC8, and HOXD10 genes as well as their combinations (Table 3). Moreover, both food treatment and mother-F1 generation affected significantly the relative expressions of Abd-B and CDX-1 genes (Table 3). In clone 2-mothers, the expressions of 10 Hox genes (Abd-B, Dll, HOX-2, HOXA1, HOXA2, HOXB3, HOXB3-2 was significantly lower than that in the 100% S food treatment (p < .05), whereas it was only significantly lower in the 20%M + 80%S food treatment for HOXB3.

| Identification and phylogenies of D. similoides sinensis Hox genes
In this study, 14 Hox genes of D. similoides sinensis were identified based on previous transcriptomic data  Table 2).
In the shrimp L. vannamei, there were 13 Hox gene protein sequences at the transcriptomic level (Sun et al., 2015). However, 39 Hox gene sequences in Ichthyophis bannanicus were found based on genomic data (Wu et al., 2015). Therefore, the 14 Hox genes in D. similoides sinensis in this study might be underestimated based on the data of the transcriptome rather than the genome.
A phylogenetic tree constructed based on amino acid sequences from vertebrates and invertebrates showed that Hox genes had evolved into different functions after multiple genomic duplication or genomic doubling events. Abd-B, CDX-1, Dll, HOXA1, HOXA2, HOXB3, HOXB3-2, HOXB7, HOXC4, HOXC7, HOXC8, and HOXD10 of D. similoides sinensis were clustered into different clades with orthologs from other species. There was an orthologous relationship between HOXB3 from D. similoides sinensis and HsHOXB3 from H. sapiens (Sun et al., 2015), and HOXB3-2 had an orthologous correlation with LmHOXB3 from L. menadoensis (Koh et al., 2003). HOXC4 from both D. similoides sinensis and D. magna were clustered into a separate clade with Dfd from L. vannamei (Sun et al., 2015), suggesting that these three spe- food treatments were significantly lower, whereas it was significantly higher for clone 2-F1 in the 20%M + 80%S food treatment. Peng et al. (2018) observed also that the mother exposed to toxic M.
aeruginosa enhanced the fitness of D. similoides sinensis offspring to Microcystis and had the differences among clones. Similarly, different genotypes of D. galeata showed different tolerance to M. aeruginosa PCC7806 (Druga et al., 2016). However, Dao et al. (2018) found that the survival rates of Daphnia lumholtzi offspring from the mothers pre-exposed to toxic Microcystis evidently dropped in spite of transgenerational adaptability to cyanobacterial toxin. Therefore, M.
aeruginosa affecting Daphnia survival rates had the differences between species or clones. Moreover, it had potential limitations using only the survival rate to evaluate the adaptability of D. similoides sinensis offspring to M. aeruginosa in this study, and more the lifehistory parameters should be employed to study the mechanism.
Microcystis can affect related gene expression of Daphnia spp. (Druga et al., 2016;Lyu et al., 2015;Schwarzenberger et al., 2009;Schwarzenberger & Elert, 2013;Xu et al., 2018).  (Druga et al., 2016;Xu et al., 2018). In this study, in the 40%M + 60%S food treatment, the survival rates at the end of the experiment of clone 1-mothers were significantly lower than those of clone 2-mothers (p < .05), and the expression of Abd-B in clone 2-mothers was higher than in clone 1-mothers. In insects, Abd-B is able to regulate the development of the posterior nodules (Hou et al., 2004), affecting the ecdysis and survival. Moreover, in this study, Clone 2-mother and Clone 2-F1 had similar survival rates under 20%M+80%S food treatment, whereas their Hox gene expression patterns are different under the same condition. Therefore, the expression patterns of Hox genes may be related to the tolerance of D. similoides sinensis offspring to M. aeruginosa and have the differences between clones.
Daphnia spp. have an inductive defense mechanism against M.
aeruginosa, which can transfer environmental information and tolerance to M. aeruginosa to their offspring, and reduce the toxic effects of M. aeruginosa (Gustafsson et al., 2005;Jiang et al., 2013;Schwarzenberger & Elert, 2013). Compared with the mothers unexposed to M. aeruginosa, the offspring from mothers exposed to M. aeruginosa have a shorter time to maturation and produce much more offspring, and so had greater fitness for an adverse environment (Gustafsson et al., 2005). Schwarzenberger and Elert (2013) observed that cyanobacterial protease inhibitors could lead to an increase in protease gene expression of D. magna offspring. Arginine kinase transcript level of D. magna offspring whose mothers had been previously exposed to M. aeruginosa were significantly higher than those of mothers fed with pure S. obliquus (Lyu et al., 2015).
The Hox genes, as a family encoding transcriptional regulator, could regulate the growth and development of crustaceans as well as body formation (Hou et al., 2004). Dll is an important gene regulating the growth of arthropods (Hou et al., 2004), and could similarly regulate appendage development in insects (Hughes & Kaufman, 2002). Vachon et al. (1992) found also that the abdomen appendages in insects might not be developed if Dll was inhibited by other Hox genes.
In this study, compared to those in the 100%S food treatment, the expression of Dll of clone 1-mothers and clone 1-F1 in the 40%M+  Therefore, further studies need to be promoted in the future.

ACK N OWLEG M ENTS
The authors thank Qi Liu, Tingting Zhang, and Yuchen Sun for their assistance with sample collection and culturing. This work was supported by the National Natural Science Foundation (No. 31870451, 31370470) of China and National Youth Science foundation (No.

CO N FLI C T O F I NTE R E S T
All authors declare that they have no conflict of interest. Zhongze Zhou: Conceptualization (supporting); Project administration (equal); Supervision (equal).