The CDK8 cyclin-dependent kinase and its paralog, CDK19, are not directly involved in regulation of the cell cycle phase transitions and belong to the so-called transcription-associated kinases that regulate gene transcription [1]. However, the mechanisms underlying such regulation are diverse. On the one hand CDK8/19 as part of a complex with cyclin C, MED12, and MED13 directly phosphorylate some transcription factors, such as intracellular domain NOTCH [2] or STAT1 [3]. On the other hand, together these four proteins form the kinase module of the Mediator complex that regulates gene expression via binding to the promoter and enhancer regions [1]. Despite the fact that in vitro studies have shown that, just like the CDK7 and CDK9 transcription-associated kinases, CDK8/19 can phosphorylate the RNA polymerase II C-terminal region, which represents an important event during transition to the elongation stage. This mechanism is likely to play no significant role in the cell [4]. At the same time it has been shown that CDK8/19 play an important role in expression of certain genes, especially in activation of expression of the earlier inactivated genes [5– 7] and the major oncogenes, such as c-Myc [8, 9] and genes of the Wnt/β-catenin signaling pathway [10]. In some cases the expression levels correlate with the presence of CDK8/19 in the enhancers and super-enhancers of the corresponding genes [11]. However, despite the important fundamental role played by CDK8/19 and their potential as therapeutic targets, specific molecular mechanism underlying the relationship between the expression of certain genes and CDK8/19 are unknown.

Recently there are papers reporting the comparison of chemical inhibition and genetic inactivation that show that genetic inactivation provides much larger effects in some models [7, 12, 13]. It must follow from this fact that CDK8/19 has some kinase-independent mechanisms of action. However, such comparison is not always precise, even in in vitro studies, due to limited selectivity and efficiency of inhibitors and is almost impossible in in vivo studies due to the features of biodistibution and metabolism of chemical inhibitors and the difficulties associated with crossing the blood-brain and blood–testis barriers by the substances.

The study was aimed to generate transgenic mice capable of tissue-specific induced expression of the CDK8 (D173A) mutant kinase-negative form for further identification of possible CDK8/19 kinase-independent mechanisms of action.

METHODS

The construct

Transgenesis involved the pKB2 vector that, in contrast to the previously used vector pKB1 [14, 15], was characterized by the absence of the reporter gene. The open reading frame was amplified using the KapaHiFi polymerase (Kapa Biosystems; USA) and primers P1 and P2 (hereinafter all olygonucleotides, the sequences of which are provided in tab. 1, have been synthesized by Evrogen, Russia) with cDNA. cDNAs were obtained by reverse transcription using the RevertAid reverse transcriptase (Thermo Scientific; USA) based on the RNA extracted from the murine brain using the ExtractRNA reagent (Evrogen; Russia). When performing amplification, the AgeI and MluI restriction sites were added to the 5' and 3' ends of the reading frame for further transfer to the final vector, along with the Kozak consensus sequence. The amplified reading frame was cloned into the CloneJet vector (Thermo Scientific; USA) and sequenced. The c.A518C mutation was created by site-directed mutagenesis during the polymerase chain reaction with the previously phosphorylated primers P3 and P4. The presence of target mutation and the absence of additional mutations were confirmed by sequencing. Then the ORF of the obtained mCdk8kd gene variant (kinase-dead) was re-cloned by the AgeI and MluI sites (all the restriction endonucleases were manufactured by Thermo Scientific, USA) into the pKB2 vector. The construct was linearized by sites SalI and NotI, separated by electrophoresis, extracted from the gel with the Cleanup Mini kit (Evrogen; Russia), further purified using the spin column with the Corning Costar Spin-X 0.22 μm nylon filter (Corning; USA), and diluted in the microinjection buffer (10 mmol Tris, 1 mmol EDTA) to the DNA concentration of 2 ng/μL.

Mice housing conditions

Embryos were obtained from 30 immature (weight 12–13 g) females (F1 hybrids CBA × C57BL/6) and similar males aged 6–8 weeks (Stolbovaya breeding station; Russia). The CD1 outbred mice (Stolbovaya breeding station; Russia) were used as recipients and foster mothers. The mice that were kept in the vivarium of the Institute of Gene Biology RAS had free access to water and food. The 22–24 °C ambient temperature was maintained, and the light/dark cycle (day/night) was 14/10 h.

Microinjections and embryo transfer

Microinjections and embryo transfer were performed as described previously [16]. On day 19 after the embryo transfer, recipients underwent caesarean section, and the newborn mouse pups were housed together with foster mothers.

Genotyping and copy number measurement

Animal genotyping was performed in accordance with the previously used protocol [14]. A STOP cassette (primers P5 and P6) and a terminator (primers P7 and P8) being parts of the pKB2 vector were simultaneously detected in transgenic animals by multiplex polymerase chain reaction (PCR). The copy number of the inserted construct was defined by comparison with genes with the known varying copy numbers (HPRT, HbA, H3C7) based on the results of the real-time PCR [14].

RESULTS

To generate transgenic mice capable of induced and tissue-specific expression of the kinase-negative CDK8 variant, we created a genetic construct containing the open reading frame of the mCdk8 gene with the c.A518C substitution in the DNA (and, consequently, with the D173A substitution in the protein) based on the pKB2 vector [17]. Along with the insulator and terminators which are protection against the position effect during insertion, the vector used contains the CAG promoter and the STOP cassette flanked by LoxP sites that separates promoter from the ORF. The STOP cassette inserted in such a way significantly reduces thetransgene transcription and makes further translation of the transcript impossible [14].

The linearized construct was microinjected into the zygotes, among which 487 zygotes survived and were transferred to recipients. As a results, 14 mouse pups were born, among which six pups were transgenic. The presence of the transgene was confirmed by PCR in all newborn animals (fig. 1). A numerical representation of the results of activities on generating transgenic animals is provided in tab. 2.

Among six primary transgenic animals, two died before puberty and four produced offspring when bred with the C57BL/6J mice. To date, four independent substrains have been generated: 369, 372, 375, 376. When random integration takes place in the genome, multimers of the construct can be formed and then inserted in chromosomes. In turn, the expression levels can be a fuction of the number of monomers. Furthermore, since insertion in distinct blastomeres may occur independently, F0 animals may be mosaic due to the varying copy numbers of the construct in various cells. We have defined the average copy number of insertion for F0 and the copy number for the substrains obtained (fig. 2). The resulting values can be non-integers because of using exponential approximation when assessing the copy numbers and averaging the values of various cells. However, in fact this is due to the measurement error. Starting from the F1 generation, the copy numbers are constant integer values. Two substrains of F1 hybrids with the copy numbers of about 2, which correspond to the copy numbers of the majority of genes in diploid genomes, were selected for further breeding.

DISCUSSION

We have generated the mice line capable of the induced and tissue-specific expression of the kinase-negative mCDK8 form: six F0 animals, among which four animals have produced offspring and two have become the founders of substrains by now. Furthermore, the efficiency of obtaining transgenes was 50% (6 out of 12), and further segregation compared to F1 that was associated with mosaicism in F0 together with the wide range of copy numbers of the inserted construct (1–17) was found in the animals generated. This allowed us to select the lines with the copy numbers most close to natural for further studies.

A series of studies consider the consequences of the CDK8 and/or CDK19 genetic knockout [18–20], however, the models used in these studies make it impossible to to draw a final conclusion about the role played by the lack of CDK8/19 proteins and impossibility of the Mediator complex kinase module assembly in the observed phenotypes, and about the role of appropriate phosphorylation. Moreover, the use of the CDK8/19 inhibitors in animals has resulted in the effects that are different from that reported in knockout animals [21]. The animal line we have generated is a right tool for distinguishing between kinase-dependent and kinase independent CDK8/19 functions.

CONCLUSIONS

During the study transgenic mice were generated providing the possibility of the tissue-specific and induced expression of the kinase-negative form of the CDK8 cyclin-dependent kinase. In the future, after breeding with the CDK19^–/–, CDK8^fl/fl mice and various activators, animals will be generated, in which the wild type CDK8 and CDK19 could be replaced by the kinase-negative variant in the ubiquitous or tissue-specific manner. The experiments involving such animals can possibly reveal the CDK8/19 kinase-independent mechanisms of action.