Background

With the improvement of people’s living standards, high-quality beef is becoming more popular. A crucial factor that affects the flavor, freshness, juiciness, tenderness and color in beef is the intramuscular fat (IMF) content. The latter plays an key role in improving taste and nutritional value, therefore it is important to shed light on the molecular mechanism of IMF deposition in cattle. Adipogenic differentiation is a complex process that includes mitotic clonal expansion (MCE) and terminal adipocyte differentiation [1,2,3]. Promotion of the cell cycle and maintenance of MCE, necessary for the terminal adipocyte differentiation, depend on cyclin-dependent kinases [4]. In turn, adipocyte terminal differentiation requires exiting from the mitotic cycle and entry into a permanent state of G1 arrest, together with expression of adipogenic phenotypic genes [1]. PPARγ C/EBPα and C/EBPβ are the key transcription factors in the regulatory network that functions during both MCE and terminal adipocyte differentiation [5]. During MCE, C/EBPβ can transactivate the expression of multiple cell cycle-related genes and also facilitate adipocyte hyperplasia by delaying the expression of C/EBPα and PPARγ [5]. When adipocytes accumulate above a threshold, terminal differentiation is initiated, and several genes are expressed to directly or indirectly regulate C/EBPα and PPARγ. Then activated PPARγ and C/EBPα regulate the expression of adipogenic phenotypic genes, e.g., FABP1, FABP4, LPL, aP2, CAP and Perilipin, leading to lipid droplet formation [5,6,7]. Thus, during adipogenic differentiation there is a variety of genes in families such as KLFs, SMADs, RARs, DGATs SREBPs, CDKs and others, that function synergistically [8, 9]. Herein, we focus on the regulatory mechanism of the CDK gene family in adipocyte differentiation.

CDKs are a large family of serine/threonine protein kinases. They were first discovered in the context of regulation of the cell cycle and have diverse functions in eukaryotes, like mRNA processing, regulation of transcription [10,11,12,13]. Recently, they have been shown to regulate adipocyte differentiation and lipid droplet formation by phosphorylating a series of associated transcription factors or adipocyte-specific genes. When the member of the CDK family genes CDK6 was targeted by miR107 in adipocytes, Notch and its downstream gene Hes1 were downregulated and this caused inhibition of glucose uptake and triglyceride synthesis in adipocytes [14]. MAPK and CDK2/cyclinA sequentially activate C/EBPβ by maintaining the phosphorylated state of Thr188 during MCE progression and adipocytes terminal differentiation [15]. Insulin activates the CCND3-CDK4 complex which in turn phosphorylates the insulin receptor IRS2 at Ser388; this maintains an active insulin signaling pathway in adipocytes, eventually promoting de novo lipid synthesis [16]. In addition, CDK4 can phosphorylate Rb to release E2F, leading to preadipocyte proliferation as well as phosphorylation of PPARγ which regulates the terminal differentiation of adipocytes [17]. CDK5 can reduce the insulin sensitivity of adipocytes by phosphorylating PPARγ at Ser273, and inhibition of this phosphorylation promotes browning and thermogenesis in white adipose tissue [18]. The CDK7 complex can inactivate PPARγ through PPARγ-S112 phosphorylation and inhibit adipogenesis [19]. CDK8 inhibits adipogenesis by phosphorylating the serine residues of SREBP-1c, leading to its ubiquitination and degradation [20]. These findings inspired our curiosity and guided our exploration of the impact of the CDK gene family on bovine adipocyte differentiation.

The expression patterns and regulatory mechanisms of CDK genes in bovine adipocytes have not been systematically studied and elucidated. Therefore, in the present study we aimed to first detect the CDK gene family in the bovine genome and to elucidate their physicochemical properties and structural features together with detailed classification, phylogenetic and functional analyses. In addition, to identify essential members of CDK family that affect adipogenic differentiation, we performed expression pattern analysis by RNA sequencing and qPCR. Our study provides deep insights into CDK genes that influence adipogenic differentiation and is beneficial for future studies to improve IMF in the context of bovine breeding.

Results

Identification of CDK family members

To identify the CDK family members, 59 verified CDK amino acid sequences were used as the query. These belonged to cattle (Bos taurus, 8), human (Homo sapiens, 26) and mouse (Mus musculus, 25). Query sequences were used for genome-wide detection of homologous sequences in Bos taurus, Bos indicus, Bos grunniens, Hybrid-Bos indicus, Hybrid-bos taurus, Bos mutus, Bison bison bison, and Bubalus bubalis. In Bos taurus, 25 non-redundant CDK protein sequences were identified that included CDK1–10, CDK11B, CDK12–20 and CDKL1–5 (Table 1). CDK family proteins were also recognized in Bos grunniens (22), Hybrid-Bos indicus (21), Hybrid-bos taurus (22), Bos mutus (24), Bison bison bison (22), Bos indicus (25) and Bubalus bubalis (24) (Additional file 1). Protein sequences of all CDKs are provided in Additional file 2. A newly identified member of the CDK family was found in Hybrid-bos taurus, Transcript ENSBIXT00000049337, and was named as CDK20 following sequence similarity and collinearity.

Table 1 Characteristics of genome-wide identified CDK family members in Bos taurus
Full size table

The amino acid lengths of the 25 cattle CDK proteins ranged from 292 (CDK5) to 1512 (CDK13), whereas molecular weight (Mw) ranged from 33,288.47 to 164,717.14 Da, consistent with protein length. The isoelectric point (pI) of most CDK family proteins was higher than 8.0, as they contained more basic amino acids than acidic amino acids. However, CDK5 and CDK16 were neutral, with pIs 7.57 and 7.23, respectively. Also, five proteins were acidic (CDK4, CDK6, CDK11B, CDK15 and CDK20) with pIs ranging from 5.34 to 6.68. All 25 CDK proteins contained the Serine/Threonine Kinase (STK) conserved domain (Additional file 3).

Structural features of bovine CDK family members

To explore the structural characteristics of bovine CDK proteins and genes, the conserved motifs and gene structures were projected based on their phylogenetic relationships (Fig. 1). CDKs of cattle clustered into three main subfamilies according to the evolutionary clades. One subfamily contains 6 members CDKL1–5 and CDK20, another contains CDK10 and CDK11B, whereas the rest belongs to the third subfamily. All the CDK family proteins shared six conserved domains termed motifs 1, 3, 5, 6, 7, and 9, formed by 29, 21, 21, 21, 21 and 21 amino acids, respectively (Additional file 4). A small branch in the third subfamily formed by CDK16, CDK17 and CDK18, has all ten motifs. CDK4, CDK15 and CDK20 all have eight motifs, since CDK4 lacks 4 and 10, CDK15 lacks 2 and 10, and CDK20 lacks 4 and 10. The remaining CDK proteins comprise nine motifs lacking CDK10, which indicates they all have the same conserved patterns.

Fig. 1
figure 1

Characterization of the identified CDK proteins and genes in Bos taurus. The phylogenetic tree (left) was constructed by the Neighbor Joining method. Amino acid sequences (middle), where conserved motifs are indicated as rectangles with different colors. Gene structure map (right), where green rectangle, black line and red rectangle represent CDS, intron and UTR, respectively

Full size image

Introns, coding sequences (CDS) and untranslated regions (UTR) were variable among the CDK gene family. For instance, the length of CDK genes ranged from 3599 nt (CDK4) to 678,562 nt (CDK14), mainly due to intron variation. The number of CDS varied from 7 to 17, and the length and layout of 3’UTR and 5’UTR were also variable in the noncoding areas. Despite this variability in CDS, introns and UTRs, we discovered that CDK family members in the same evolutionary branch tend to show similar gene structures and conserved patterns in motifs.

Phylogenetic relationship of CDK proteins in different organisms

To assess the CDK proteins evolutionary relationships between cattle and other organisms, we conducted a phylogenetic analysis that included eight Bovinae species (Bos taurus, Bos indicus, Bos grunniens, Hybrid-Bos indicus, Hybrid-bos taurus, Bos mutus, Bison bison bison, and Bubalus bubalis) and Homo sapiens and Mus musculus. CDK proteins in human and mouse were also included since they have been studied extensively as model organisms. From these ten organisms, 236 amino acid sequences were aligned to generate a nonrooted Neighbor-Joining (NJ) tree (Fig. 2). Phylogenetic analyses revealed a clustering of CDK family proteins into eight major clades.

Fig. 2
figure 2

Phylogenetic Neighbor-Joining (NJ) tree of CDK proteins from ten organisms. Identified CDKs in Bos taurus (BOSTA), Bos grunniens (BOSGR), Hybrid-Bos Indicus (BOSIH), Hybrid-Bos taurus (BOSTH), Bos mutus (BOSMU), Bison bison bison (BISOM), Bos indicus (BOSIN) and Bubalus bubalis (BUBBU) together with verified CDKs from Homo sapiens (HUMAN) and Mus musculus (MOUSE). CDK proteins are grouped into eight clusters (I-VII) shown as different colors

Full size image

Chromosomal distribution and collinearity analysis of CDK genes

CDK genes were mapped on the chromosomes of six Bovinae species; 25 bovine CDK genes distribute on 16 chromosomes (Fig. 3). The CDK genes of cattle have a similar distribution as those in Bos indicus. However, the order of CDK14 (7.94–8.62 Mb), CDK6 (9.94–10.19 Mb), CDK13 (80.95–81.08 Mb) and CDK5 (113.630–113.634 Mb) in Bos taurus Chr 4 was opposite from that in Hybrid-Bos indicus, Hybrid-bos taurus and Bos grunniens. In Bos taurus, CDK4, CDK2 and CDK17 were three tandem genes located at 29.66–29.69 Mb, 29.66–29.69 Mb and 29.72–29.92 Mb on Chr 5, whereas in Bos grunniens, Hybrid-Bos indicus Chr 5 and Bubalus bubalis Chr 4, the arrangement of these three genes was reversed. In addition, compared with Bos taurus, several genes were missing in Bos grunniens (CDK7, CDK11B and CDK20), Hybrid-Bos indicus (CDK11B, CDK16, CDK20 and CDKL4), Hybrid-bos taurus (CDK11B, CDK20, CDKL4 and CDKL5), and Bubalus bubalis (CDK11B). Also, CDKL5 is located at chromosome X in Bos taurus, whereas it is in chromosome Y in Bos grunniens.

Fig. 3
figure 3

Chromosomal distribution of CDK genes. The black font on the left represents chromosome numbers and the red font on the right represents CDK genes

Full size image

Genome collinearity analysis resulted in the identification of pairs of collinear genes between Bos taurus and Bos indicus (31,691), Hybrid-Bos indicus (34,495), Hybrid-bos taurus (33,570), Bos grunniens (32,378) and Bubalus bubalis (33,327), respectively (Fig. 4). There is a one-to-one correspondence between chromosomes of Bos taurus and Hybrid-Bos indicus, Hybrid-bos taurus, Bos indicus and Bos grunniens. A large chromosome homology also existed between cattle (2 N = 60) and buffalo (2 N = 50), although the chromosome number is different in the two species. The syntenic blocks revealed syntenic relationships in the CDK gene family between cattle and the other five species in Bovinae (Table 2).

Fig. 4
figure 4

Collinearity analysis of CDK genes between cattle and other organisms. Syntenic genes pairs are linked by grey lines whereas syntenic CDK genes are shown as red lines

Full size image
Table 2 Syntenic relationships of CDK family genes between cattle and the other five species
Full size table

Expression analysis of CDK genes in different tissue

The expression patterns of genes could provide important references for their function. To explore the expression patterns of the CDK gene family during adipogenic differentiation, we investigated the relative expression levels in 163 samples of 60 tissue types. The results showed that CDK genes displayed differential expression patterns in diverse tissues (Fig. 5 a), which could be classified into five groups (A to E). PPARγ is a marker gene for adipocyte differentiation, and consistent with this, its expression was high in Group B which included omental, intramuscula, subcutaneous and mammary gland fats. According to their expression patterns, the 25 CDK genes could be grouped into four categories. They were all expressed in the 60 tissues, suggesting a broad regulatory role in life activities. Group I (CDK4, CDK9 and CDK11B) showed the highest expression levels, followed by Group III (CDK3, CDK5, CDK7, CDK8, CDK10, CDK18, and CDK20) and Group IV (CDK1, CDK2, CDK6, CDK12, CDK13, CDK14, CDK16, and CDK17). The rest of the members were in Group II and their expression was the lowest. Further analysis of the five different fat tissues revealed that CDK9 was highly expressed in all of them with an expression pattern similar to PPARγ (Fig. 5 b).

Fig. 5
figure 5

Expression analysis of the CDK gene family in different bovine tissue types. a. Expression analysis of theCDK gene family in 60 bovine tissues. The tissues were classified into 5 groups (A to E) and the 27 genes were classified into 4 groups (I-IV) according to their expression pattern. b. Expression analysis of CDK gene family in 5 bovine fat tissues. The horizontal axis represents 25 CDKs and 2 marker genes (PPARG and CEBPA) of adipocyte differentiation. The vertical axis represents different bovine tissues

Full size image

Expression analysis of CDK genes in preadipocytes and differentiated adipocytes by RNA-seq

As revealed by transcriptome analysis, 25 CDK genes showed an up-regulation trend in preadipocytes compared with differentiated adipocytes, except for CDK1, CDK3, CDK6, CDK19, CDKL1 and CDKL4 (Fig. 6). CDK7 displayed high expression, whereas CDK1 showed low expression in preadipocytes (within the 95% confidence interval). CDK4, CDK8, CDK9 and CDK14 displayed a high expression in preadipocytes (within the 99% confidence interval).

Fig. 6
figure 6

Expression analysis of CDK genes in preadipocytes and differentiated adipocytes by RNA-seq. Error bars were obtained from two measurements. Symbols * and ** above the bars indicate significant differences at 0.05 and 0.01 level between preadipocytes and differentiated adipocytes

Full size image

Expression analysis of CDK genes during adipocyte differentiation by qPCR

To further explore the expression patterns of the CDK gene family, preadipocytes collected from perirenal adipose tissue of premature calves were induced to differentiate. The results of Oil Red O staining showed that lipid droplet accumulation increased in adipocytes induced for 10 days compared to preadipocytes (Additional file 5), indicating that induction and differentiation was successful. A qPCR analysis was conducted to detect the expression of CDK genes at 0, 2, 4, 6, and 10 days during adipocytes differentiation (Fig. 7). CDK genes showed a relatively high expression in preadipocytes and then decreased as differentiation progressed in addition to CDK1, CDK15, CDK18, CDKL3 and CDKL5. CDK1, CDKL3 and CDKL5 showed the highest expression on the second day of differentiation whereas the lowest expression points were on the 6th, 8th and 6th day, respectively. CDK15 and CDK18 increased with adipocyte differentiation and reached a peak on the fourth day, then decreased.

Fig. 7
figure 7

qPCR expression analysis of CDK genes during adipocyte differentiation. Error bars were obtained from three measurements. Symbols * and ** above the bars indicate significant differences at 0.05 and 0.01 level compared with preadipocytes (0 day). These symbols above the lines represent the significance of differences between two labeled pillars

Full size image

Discussion

Cattle is known as an important species for supplying meat. The IMF content directly affects the taste and flavor of beef and it is of great scientific significance to reveal the molecular regulation mechanism of IMF deposition. The CDK gene family encoding functional proteins have been well studied in the regulation of transcription, metabolism and cell differentiation [10,11,12]. However, research on the role of CDK genes in adipocyte differentiation is limited, especially in Bovidae. Since cattle and several species of Bovidae were sequenced, these resources may serve as reference to explore the evolution and function of CDK genes and in advancing genome science in Bovidae.

Structural features of bovine CDK family proteins and genes

The activity of proteins depends on their functional motifs and domains [21]. Six conserved amino acid sequences (Motifs 1, 3, 5, 6, 7 and 9) were conserved among all CDK family members in cattle. Highly conserved motifs are usually located in active sites of enzymes and are essential in maintaining their structure, binding to substrate or in catalysis [22, 23]. CDK16–18, a small branch in a subfamily, have all of the ten motifs, pointing to a common function in these three members. It is likely that Motif 10 is not located at the core of the catalytic domain since the other 22 members where it is absent still show kinase activities. In addition to Motif 10, other motifs are absent in CDK4 (Motif 4), CDK15 (Motif 2) and CDK20 (Motif 8) indicating some sequence losses during evolution. CDK20 is a newly identified member in Hybrid-bos taurus, and has a conserved STKc domain and nine motifs (Additional file 6). Domain and motif analysis revealed its similarity with the other CDK proteins.

The distribution of CDSs, introns, and UTRs was variable in CDK genes, mainly due to the length and layout of introns and UTRs, whereas gene coding sequences and amino acid sequences were similar. These results suggest that amino acid sequence similarity, especially in conserved motifs, may play essential roles in kee** the kinase function in CDK proteins.

Phylogenetic relationship of CDK family proteins

Phylogenetic analysis of proteins can provide an in-deep insight for their evolutionary relationships [24]. CDK family proteins were classified into eight major clades. The same member from each species first clustered in one branch, indicating that they were conserved in sequences among the ten species. Initially, Clade I separated, whereas Clade II, III and IV clustered into a subfamily and the remaining clustered into another subfamily. This shows that they have evolved asymmetrically and that these subfamilies are distant. Notably, Clade VII includes all the cyclin-dependent kinases like proteins (CDKL1, CDKL2, CDKL3, CDKL4 and CDKL5), consistent with human study that divided the CDK gene family into CDK and CDKL [12]. As expected, members that are closer evolutionarily clustered together first. For example, Bos indicus CDK1 first clustered with that of Bubalus bubalis, and then get together with Hybrid-bos taurus, Bos taurus, Bos mutus, Hybrid-Bos indicus, Bos grunniens, Bison, human and mouse.

Collinearity analysis of CDKs in Bovidae

Members of a gene family may distribute in the same or in different chromosomes. These two cases are due to tandem or segmental duplication events, respectively [50]. Quality control of raw sequence data, including removal of the adapter sequences and low-quality sequences was performed using the Trim_galore [51]. Clean reads were then mapped to the Bos taurus genome (ARS-UCD1.2.101) using STAR [52]. RSEM [53] and FeatureCounts [54] were used to calculate the expression of the transcripts. Data was normalized by calculating the RPKM for each gene. The limma package [55] in R was used to analyze the differential expression under the threshold of fold change ≥1.2 and p-value < 0.05. The Sequence Read Archive Run (SRR) number and adjusted RPKM values of 163 tissue samples were downloaded from the Ruminant Genome Database (http://animal.nwsuaf.edu.cn/code/index.php/Ruminantia) [56] (Additional file 8). The heat map was constructed using the pheatmap package in R software [57].

Isolation, culture and induction differentiation of bovine primary adipocytes

Primary adipocytes were isolated and cultured from perirenal adipose tissue of premature calf in the Zerui ecological breeding farm. Type I collagenase digestion method was used for the isolation and cultivation of calf preadipocytes. Induction of preadipocytes differentiation [58] and Oil Red O staining [59] were performed as described. Images were captured using a 10× objective lenses without filter model on an OLYMPUS IX50 microscope (Olympus Corporation, Tokyo, Japan) equipped with a charge-coupled device (CCD) camera (Diagnostic Institute, Inc., Sterling Height, MI, USA). CellSens V3.2 acquisition software was used and the images with a resolution at 96 dpi were acquired.

RNA extraction and quantitative RT-PCR (qPCR)

The primers of the CDK genes were designed using Primer Premier 5.0 software (Additional file 9). Total RNA was extracted by the phenol-chloroform method using TRIzol (9109, Takara), and bovine adipocytes (n = 3) during the differentiation of 0, 2, 4, 6 and 10 days were used. Absorbance was measured at 260 nm and 280 nm and samples with an OD260/OD280 ratio between 1.8 and 2.0 were used in subsequent experiments. Then, 1000 ng total RNA was reverse transcribed using random primers with Moloney murine leukemia virus reverse transcriptase (Takara Bio, Kyoto, Japan). Realtime PCR was carried out in a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) with SYBR Green Master Mix (Takara Bio, Kyoto, Japan). The qPCR reaction procedure was 40 cycles of pre-denaturation for 3 min (95 °C), denaturation for 10 s (95 °C), annealing for 20 s (59–61 °C), extension for 30 s (72 °C). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was selected as the reference gene.

Statistical analysis

All qRT-PCR results were determined using the 2−∆∆Ct method [60, 61]. Three independent technical repetitions were performed for each test. Statistical significance determined using Graphpad Prism 7.0 software.