Cloning, Characterization, and Expression Features of Chicken CDS2 Splicing Variants

Background: The CDS gene encodes the CDP-diacylglycerol synthase enzyme that catalyzes the formation of CDP-diacylglycerol (CDP-DAG) from phosphatidic acid. At present, there are no reports of CDS2 in birds. Here, we identied chicken CDS2 transcripts by combining conventional RT- PCR amplication, 5' RACE (Fig. 1A), and 3' RACE, explored the spatio-temporal expression proles of total CDS2 and the longest transcript variant CDS2-4, and investigated the effect of exogenous insulin on total the mRNA level of CDS2 by quantitative real-time PCR. Results: Four transcripts of chicken CDS2 (CDS2-1, -2, -3, and -4) were identied, which were alternatively spliced at the 3′-untranslated region (UTR). CDS2 was widely expressed in all tissues examined and the longest variant CDS2-4 was the major transcript. Both total CDS2 and CDS2-4 were prominently expressed in adipose tissue and the heart, and exhibited low expression in the liver and pectoralis of 49 day-old chickens. Quantitative real-time PCR revealed that total CDS2 and CDS2-4 had different spatio-temporal expression patterns in chicken. Total CDS2 exhibited a similar temporal expression tendency with a high level in the later period of incubation (embryonic day 19 [E19] or 1-day-old) in the brain, liver, and pectoralis. While CDS2-4 presented a distinct temporal expression pattern in these tissues, CDS2-4 levels peaked at 21 days in the brain and pectoralis, while liver CDS2-4 mRNA levels were highest at the early stage of hatching (E10). Total CDS2 (P < 0.001) and CDS2-4 (P = 0.0090) mRNA levels in the liver were differentially regulated throughout development of the chicken. Exogenous insulin signicantly downregulated the level of total CDS2 at 240 min in the pectoralis of Silky chickens (P < 0.01). Total CDS2 levels in the liver of Silky chickens were higher than that of the broiler in the basal state and after

The expression of mammalian CDS2 is ubiquitous [10][11][12]. Recent research revealed that CDS is involved in numerous cellular functions. Studies from Drosophila highlight the importance of CDS (CdsA) in the visual photo-transduction system [4]. Mutations in the eye-speci c CDS gene in Drosophila resulted in a defect in photo-transduction and retinal degeneration [4]. CDS also plays an important role in the regulation of vascular endothelial growth factor-A (VEGFA) signaling and angiogenesis [18]. Loss of CDS2 has been shown to cause a defect in VEGF signaling activity and angiogenic capacity in zebra sh, primarily by decreasing the level of PIP2 regeneration [19]. Genetic ablation of CDS2 switches the output of VEGFA signaling from promoting angiogenesis to inducing vessel regression and tumor inhibition [18].
Recent reports have revealed that the CDS gene is involved in lipid metabolism [1,2,15,20,21], and mammalian CDS1 and CDS2 regulate lipid droplets through distinct mechanisms [21]. Human CDS1 and CDS2 can create different CDP-DAG pools; CDS2 is selective for acyl chains at the sn-1 and sn-2 positions [22]. The importance of CDS enzymes stems from the need for cells to maintain their phosphoinositide levels, in particular, those of PIP2 [7,16].
Although Tam41 also catalyzes the synthesis of CDP-DAG using phosphatidic acid and cytidine 5′triphosphate as substrates, it shares no sequence or structural homology with the CDS enzyme [17,23]. In addition, CDS is an integral membrane protein, whilst Tam41 is a peripheral membrane protein [14,17,23].
Until now, there have been no reports of a CDS gene in birds. In this study, we identi ed chicken CDS2 transcripts, analyzed the genomic structure, chromosomal synteny, spatio-temporal expression pattern, and the effect of exogenous insulin on chicken CDS2, revealing the basic characteristics of chicken CDS2 and further identifying the function of CDS2 in the regulation of chicken development and growth.
Syntenic analysis revealed that chicken CDS2 was located between PCNA and ARHGAP25/BMP10 on chr.22, close to the chromosomal breakage/fusion point of the mammalian/bird evolutionary event. The upstream chromosomal region of CDS2 in birds containing the CDS2-PCNA-TMEM230-SVCT gene is homologous to chr.20 in human and chr.2 in mouse while the downstream chromosomal region containing ARHGAP25-BMP10-GKN2-GKN1 was homologous to chr.2 in human and chr.6 in mouse (

Tissue expression pro le of chicken CDS2
We conducted semi-quantitative reverse transcription PCR (RT-PCR) and qPCR to investigate the tissue expression pattern of chicken CDS2 in 49 day-old Silky chickens with CP2, AS1, and QP2 primer sets (Table S2).
Ampli cation with the CP2 primer set located in the coding region of the CDS2 gene re ects the total expression level of CDS2 (referred to as total CDS2). Total CDS2 was extensively expressed in all detected tissues (Fig. 3A). The qPCR revealed that total CDS2 was relatively highly expressed in adipose tissues including sebum (at the tail root), abdominal fat, and neck fat, followed by heart and leg muscle, and weakly expressed in the liver and pectoralis (P = 0.0528, Fig. 3B).
Primer pair AS1, which can simultaneously amplify four splicing forms, was used to observe the tissue expression pattern of multiple transcripts by semi-quantitative RT-PCR (Figs. 4A and S4). The ampli ed lengths of the CDS2-1, CDS2-2, CDS2-3, and CDS2-4 transcripts should be 245, 379, 504, and 1142 bp, respectively. QP2 is located in the speci c region of the longest transcript CDS2-4, which could speci cally detect expression of the CDS2-4 variant. CDS2-4 was extensively expressed, and predominantly present in nearly all detected tissues (Figs. 4 and S4). Expression of CDS2-3 was detected in almost all tissues, while the short transcripts CDS2-1 and CDS2-2 were undetected with the AS1 primer set in two separate experiments (Figs. 4A and S4).
The qPCR with the QP2 primer revealed that CDS2-4 showed a similar expression pattern as total CDS2, with a relative high mRNA level in adipose tissues including sebum, abdominal fat, and neck fat, followed by heart, testis, brain, and leg muscle, and low levels in the liver and pectoralis (Fig. 5). Compared with the tissue expression pattern of total CDS2 detected with the CP2 primer set (P = 0.0528, Fig. 3B), the level of CDS2-4 showed greater tissue uctuation (P < 0.0001, Fig. 5). CDS2-4 exhibited similar mRNA levels in heart and adipose tissues including sebum, abdominal fat, and neck fat, while CDS2-4 levels in abdominal fat were signi cantly higher than those of other tissues (P < 0.01). The CDS2-4 level in adipose tissue was approximately four-fold that in the testis and 120-fold that in the liver (Fig. 5).
The spatio-temporal expression patterns of total CDS2 and CDS2-4 We further investigated the spatio-temporal expression patterns of total CDS2 (with the CP2 primer set) and CDS2-4 (with the QP2 primer set) by qPCR. One-way ANOVA was used to analyze the effect of age and tissue on the expression of total CDS2 and CDS2-4 separately. Total CDS2 exhibited similar temporal expression patterns (Fig. 6A) in the brain (P = 0.1623), liver (P < 0.0001), and pectoralis (P = 0.1712). In general, the total CDS2 mRNA level was the highest in the later stage of embryogenesis (embryonic day 19 [E19]/ one-day-old [D1]), and the E19 level in the liver was signi cantly higher than that of the other time points (P < 0.05). In addition, the relative expression of total CDS2 in the brain was signi cantly (P < 0.05) higher than that in the liver (9-fold) and pectoralis (3-fold) at 49 days (Fig. 6A).
Unlike total CDS2, the mRNA level of the CDS2-4 variant presented a distinct temporal change in pattern in the brain, liver, and pectoralis (Fig. 6B). CDS2-4 levels peaked at 21 days in the brain (P = 0.1336) and pectoralis (P = 0.0104). The level of CDS2-4 in the pectoralis at 21 days was signi cantly higher than that at the other ages tested, with the exception of 1 day (P < 0.05); whereas hepatic CDS2-4 levels presented a decreasing trend with development, with highest levels at the early stage of hatching (E10), a weak decrease during embryogenesis (E14-D1), and a clear decrease with age after hatching (D21 and D49, P = 0.009). Hepatic CDS2-4 levels at E10 were signi cantly higher than at 21 days (P < 0.05) and 49 days (P < 0.01) (Fig. 6B). In addition, the brain had the highest level of CDS2-4 at all time points among the three tissues; brain CDS2-4 levels at 21 days were signi cantly higher than those in the liver and pectoralis (P < 0.01, Fig. 6B).
Effect of exogenous insulin on chicken CDS2 expression Exogenous insulin resulted to the rapid drop of blood glucose (until 120min) in both Silky chickens and broilers, and Silky chickens presented a more rapid recovery of blood glucose than broilers after 120 min ( Fig. 7A and 7B). Exogenous insulin downregulated the expression of total CDS2 in the pectoralis of Silky fowl (p = 0.003, Fig. 7F), where the total CDS2 level at 240 min was signi cantly lower than at 0 and 120 min (Fig. 7F), and the total CDS2 level in the pectoralis of Silky chickens was lower than that of broiler chickens at 240 min after insulin injection (P < 0.05, Fig. 7H). The mRNA level of total CDS2 in the liver of Silky chickens was higher than that of broilers at the basal state and after insulin stimulation (P < 0.05, Fig. 7G), and exogenous insulin could weakly upregulate the mRNA level of total CDS2 at 240 min in the livers of broilers (P = 0.078, Fig. 7C).

Discussion
Four transcripts of chicken CDS2 (CDS2-1-4) were identi ed; the transcripts were alternatively spliced at the 3′-UTR. Consistent with the 3'-RACE results (Fig. 1B), the longest product CDS2-4 was the major transcript observed using RT-PCR (with AS1 primers), while only two long transcript variants (CDS2-3 and CDS2-4) were detected from multiple tissues in two separate experiments (Figs. 4A and S4). The differences in identi ed transcript types may be because RACE technology is more sensitive than conventional RT-PCR. We also observed that the expression level of CDS2-3 was similar to that of CDS2-4 in the small intestine, kidney, brain, and abdominal fat tissues of one Silky chicken (Fig. 3A), but was weak in nearly all detected tissues of another individual (Fig. S4), suggesting that the expression of CDS2-3 varied greatly among individuals/tissues and the function of CDS2-3 needs to be further clari ed. 3′-UTRs are major players in gene regulation that enable local function, compartmentalization, and cooperativity [24]. Birds and mammals share high sequence similarity for the deduced CDS2 proteins, while possessing their own conservative 3'-UTR, suggesting that the 3′-UTR of CDS2 may play an important role in regulating the phenotypic diversity of higher organisms [24]. The average length of 3′-UTRs in human are 1278 nt [25]; CDS2 exhibited a long 3′-UTR both in mammals (9000 nt in human) and birds (5210 nt in chicken), suggesting there may be an abundance of regulatory elements in the 3′-UTR of CDS2. The shortening/lengthening of the 3′-UTR may result in a change in miRNA binding sites [26]. Further studies are needed to con rm the relationship between these miRNAs and CDS2.
CDS has been linked with lipid metabolism [1,2,15,20,21], with a stronger effect on phosphatidic acid levels. CDS2 de ciency (while not CDS1) could impair the maturation of initial lipid droplets in cultured mammalian cells [21]. A transcriptome pro le revealed that CDS2 was differentially expressed in adipose tissues of fat-and short-tailed Chinese sheep [27]. Here, we observed that chicken CDS2 was predominantly expressed in multiple adipose tissues, which provide direct evidence for a function of CDS2 in fat metabolism.
The level of CDS2 was low in chicken livers at 49 days, but mRNA levels of CDS2 showed dramatic changes during development. Livers in birds play a central role in lipid metabolism, serving as the center for lipoprotein uptake, formation, and export to the circulation [28,29], while adipose tissue functions primarily as a storage tissue [30]. Zhao et al. reported that hepatic triglyceride levels increased during embryonic development [31,32]. While post-hatching, the liver lipid content correlated negatively with broiler age; it was the highest on the rst day then decreased sharply at day 7 [29,33]. Here, we observed that total CDS2 mRNA exhibited a similar temporal change pattern as hepatic lipids during development. The changes in lipid and gene expression of the liver were indicative of the rapid alteration of the role of the liver in lipid metabolism during development. In addition, it has been reported that the serum lipid level (total cholesterol and triglyceride) of Silky chickens was higher than that of broilers [34]. Here, we found that total CDS2 levels in livers of Silky chickens were twice that of broilers at 47 days, which suggested that CDS2 contributes to lipid metabolism.
In past work, our group found a 30% energy restriction in broiler chickens signi cantly reduced the abdominal fat ratio and subcutaneous fat thickness [35], and downregulated an unknown differentially expressed fragment C7-2 (mapped to chicken EST CR406381.1) [36] in broiler's liver tissue. With the update of the chicken genomic database in NCBI, this fragment has now been mapped to the 3′-UTR of chicken CDS2 (XM_004947552). The downregulation of CDS2 in chicken liver by energy restriction demonstrated the association of chicken CDS2 with fat metabolism.
It has been reported that Silky chickens exhibit a stronger ability to regulate glucose homeostasis than broilers under exogenous insulin stimulation. Silky chickens showed a greater blood glucose recovery than broilers at 240 min regardless of whether food was served at 120 min after insulin stimulation [34,37]. Skeletal muscles have been considered a major regulator of systemic glucose homeostasis [38]. Here, exogenous insulin signi cantly downregulated the level of total CDS2 at 240 min in the pectoralis of Silky chickens, which re ects the involvement of CDS2 in the insulin signaling pathway in a breed-and tissue-speci c manner. Loss of CDS in Drosophila has been reported to cause a decrease in phosphatidylinositol and insulin signaling [1].

Conclusion
In summary, four alternatively spliced forms of CDS2 with variations at the 3′-UTR were identi ed in chicken, and the longest form, CDS2-4, was the main transcript. Both total CDS2 and CDS2-4 were extensively expressed in all detected tissues, and predominantly present in adipose tissues including sebum, abdominal fat, and neck fat. Total CDS2 and CDS2-4 presented distinct spatio-temporal expression patterns in the brain, liver, and pectoralis during chicken development. Total hepatic CDS2 and CDS2-4 were dramatically regulated with age in a distinctly different manner. Total CDS2 in the liver of Silky chickens was signi cantly higher than that of broilers at the basal state and after insulin stimulation. Total CDS2 in the pectoralis of Silky chickens was insulin sensitive; this level was signi cantly downregulated at 240 min after exogenous insulin stimulation.

Sample collection
Chickens and fertilized eggs were obtained from the poultry germplasm resources farm of Henan Agricultural University. The one-day-old Arbor Acres (AA) broilers and Silky chickens were cage raised routinely until 49 days with free access to food and water as described previously [34]. The diet was prepared based on nutritional standards for broilers recommended by the NRC (Nutrient Requirements for Poultry, 1994). The fertilized eggs from the Silky population were collected and hatched under conventional conditions. The research protocol was approved by the Animal Care and Use Committee of Henan Agricultural University (Zhengzhou, China).
To identify the spatio-temporal characteristics of chicken CDS2 transcripts, brain, liver, and pectoralis tissues were collected from Silky chickens at E10, E15, E19, one-day-old (newly hatched birds, D1), D21, and D49 (n = 3 for each time point). The samples were snap-frozen immediately in liquid nitrogen and stored at -80℃ for RNA isolation.
To investigate the effect of exogenous insulin on CDS2 mRNA expression, at 47 days (after 12 h fasting), female AA broilers (average weight, 1.86 kg, n = 12) and Silky fowls (average weight, 0.75 kg, n = 12) were randomly selected and administered an insulin tolerance test (with hypodermic injection of 80 μg/kg body weight insulin or equal PBS solution as the control) as described previously [34], and birds were given feed at 120 min. The fowls were sacri ced at 0 min (before insulin injection), 120 min, and 240 min after insulin injection (5-6 birds/breed at each time point). All chickens were euthanized by cervical dislocation prior to tissue isolation. The pectoralis and liver were collected and snap-frozen immediately in liquid nitrogen and stored at -80℃ for RNA isolation.
RNA extraction and rst-strand complementary DNA (cDNA) synthesis Total RNA was extracted from tissues according to the manufacturer's instructions (RNAiso Plus, TaKaRa, Dalian, China). The quality and concentration of the extracted RNA was determined by agarose gel electrophoresis and a spectrophotometer (Thermo NanoDrop One, DE, USA) respectively. The RNA was then reverse-transcribed using a PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa) in a 20 µl reaction containing 1 μg of total RNA and random primers. The synthesized cDNA samples were stored at -20°C.
Identi cation of CDS2 transcripts CDS2 transcripts were identi ed by combining conventional RT-PCR ampli cation, and 5′ and 3′ rapid ampli cation of cDNA ends (RACE). The primers for amplifying the CDS2 gene were designed based on the sequence of EST CR406381.1 and the predicted chicken CDS2 sequence (accession no. XM_417669.5), and listed in Table S2. CDS2 transcripts were identi ed from the ovary tissue of Silky chickens.
5′ and 3′ RACE were used to identify CDS2 transcripts using the SMART TM RACE cDNA Ampli cation Kit (Clontech Laboratories Inc., CA, USA) according to the manufacturer's instructions. PCR was conducted in a total volume of 50 μl, containing 2.5 μl 5' or 3'-ready cDNA, 1.0 μl GSP1 or GSP2 primer (10 μM), 2.5 μl UPM (10 μM), and 41.5 μl Master Mix and was performed according to a touchdown PCR protocol: ve cycles of ampli cation (94℃ for 30 s and 72℃ for 3 min) followed by ve cycles of 94℃ for 30 s, 70℃ for 30 s, and 72℃ for 3 min, and 25 cycles of 94℃ for 30 s, 68℃ for 30 s, and 72℃ for 3 min. The ampli ed products for 5'-and 3'-RACE were puri ed and cloned into the pMD-18T vector (TaKaRa), and transformed into E. coli DH5α. In addition, 6-10 positive clones were con rmed by Sanger sequencing.
Quantitative real-time PCR (qPCR) qPCR was carried out with a CFX96™ Real-Time PCR Detection Systems (Bio-Rad Laboratories, Hercules, CA, USA) using 2× M5 HiPer Dual Real-time PCR Super Mix (SYBRgreen with anti-Taq, Mei5bio). The βactin gene was used as an internal control. The primers for qPCR (Table S2) were optimized as previously described [39]. PCR ampli cation was performed in a total volume of 10 µl containing 5 μl Real-time PCR Super Mix (SYBRgreen with anti-Taq, Mei5bio), 1.0 μl cDNA, and 0.25 μmol/L of forward and reverse primer as follows: 95℃ for 5 min, followed by 40 cycles of 95℃ for 10 s, 60℃ for 15 s, and 72℃ for 30 s, and a nal incubation at 72℃ for 10 min, followed by a nal 5 min extension at 72°C. Three technical replicates were performed for each sample. Relative gene expression was calculated using the 2 −ΔΔCt method [7].

Statistical analysis
Data were analyzed by one-way analysis of variance using IBM SPSS Statistics 22.0 software and expressed as mean ± standard error. Multiple comparisons were conducted using the Bonferroni method. GraphPad Prism 5.0 software was used to prepare graphs. P < 0.05 was considered signi cant and P < 0.01 was considered extremely signi cant.