Casein kinase I isoform delta also known as CKI-delta or CK1δ is an enzyme that in humans is encoded by the gene CSNK1D, which is located on chromosome 17 (17q25.3). It is a member of the CK1 (formerly named casein kinase 1) family of serine/threonine specific eukaryotic protein kinases encompassing seven distinct isoforms (CK1α, γ1-3, δ, ε) as well as various post-transcriptionally processed splice variants (transcription variants, TVs) in mammalians. Meanwhile, CK1δ homologous proteins have been isolated from organisms like yeast, basidiomycetes, plants, algae, and protozoa.
Genetic coding
In 1993, the gene sequence of CK1δ was initially described by Graves et al. who isolated the cDNA from testicles of rats. After sequencing and characterization of the gene, the construct was described as a 1284 nucleotide sequence resulting in a protein consisting of 428 amino acids after transcription. The molecular weight of the according protein was published as 49 kDa. Three years later, the same gene was identified in humans. The human CSNK1D contains 1245 nucleotides and is transcribed into a protein consisting of 415 amino acids.
Ever since, CK1δ was investigated and described in various animals, plants, as well as parasites (Caenorhabditis elegans, 1998; Drosophila melanogaster, 1998; Mus musculus, 2002; Xenopus laevis, 2002.)
Transcriptional variants
So far, three different transcription variants (TVs) have been described for CK1δ in humans (Homo sapiens), mice (Mus musculus), and rats (Rattus norvegicus), which are highly homologous. The alignment of all CK1δ sequences of all organisms shows a high homology in the first 399 amino acids, except for position 381. While the human transcription variants are using isoleucine, the mouse and rat sequences incorporate a valine instead. The only exception is rat TV3, which is also transcribing its nucleotide sequence into an isoleucine.
After position 399, three different general structures can be observed. The first variant consists of 415 amino acids across all three organisms and is called TV1 in human and rat, while the murine counterpart is named CRAa. The shortest group of sequences consists of 409 amino acids: TV2 in humans and rats, CRAc in mice. The longest variant consists of 428 amino acids in rat (TV3) and mice (CRAb), while the human (TV3) variant is missing the second to last amino acid (threonine), resulting in a protein of a length of 427 amino acids.
The various transcription variants are based on a different usage of the exons that are encoding for CSNK1D. The whole gene consists of eleven different exons and is located in humans on chromosome 17 at position 17q25.3. CSNK1D has a length of 35kb and is overlapping with the gene Slc16a3. The intersecting part is exon 11, which is located downstream of exon 10. However, it does not interfere with Slc16a3 since it is located in a non-coding area.
TV1 and TV2 were postulated during an early analysis of human and murine genes in 2002. Both transcription variants share the first 399 amino acids, but differ at the following 16 amino acids for TV1 and ten amino acids for TV2, respectively. This is linked to the exon usage. While they share the first eight exons, TV1 is using exon 10and TV2 exon 9 to finish their respective sequence. The third transcription variant was postulated after a data bank analysis in the year 2014. The proposed sequence is sharing the first 399 amino acids with TV1 and TV2, but differs in the upcoming 28 amino acids. The exon usage of TV3 consists of exon 1 to 8, which is followed by exon 11 to finish the sequence.
Besides various sequences of the three different transcription variants, the variants also show differences in Michaelis-Menten kinetic parameters (Km and Vmax) in regard to their potential to phosphorylate canonical (α-casein) as well as non-canonical (GST-β-catenin1-181) substrates (Xu et al., 2019). TV3 shows an increase of phosphorylation of both substrates compared to TV1 and TV2, which is statistically significant. These differences can be explained by various degrees of autophosphorylation of the transcription variants.
Polyadenylation
Based on software analysis of the mRNA sequences various polyadenylation patterns could be identified for the transcription variants . TV1 and TV2 share the same pattern located on exon 10 starting at position 1246 resulting in a 32 nucleotide motif (AGUAGAGUCUGCGCUGUGACCUUCUGUUGGGC). TV3 uses a motif on exon 11 at the position 320. The motif is also 32 nucleotides long, but differs from the sequence used by TV1/2 (AGUGGCUUGUUCCACCUCAGCUCCCAUCUAAC). The difference in the polyadenylation sequence results in a variance of the minimum free energy values of the predicted RNA folding structures (-28.70 kcal/mol, TV1 and TV2 and -16.03 kcal/mol, TV3), which could result in different length of the Poly-A tail. Based on the observation that stable secondary structures result in decreased polyadenylation of the specific site, this might indicate that TV1 and TV2 are less polyadenylated in comparison to TV3.
Structure
Like eukaryotic protein kinases (ePKs) the different isoforms of the CK1 family consist of a N-terminal and a C-terminal lobe (N- and C-lobe, respectively), which are connected via a hinge region. While the N-lobe is mainly composed by β-sheet strands, the larger C-lobe predominantly consists of α-helical and loop structures. Between both lobes a catalytic cleft is formed, accommodating substrates and ATP for the kinase reaction.
Binding of substrates and co-substrates
Binding of phosphorylated substrates to distinct regions of the C-lobe has previously been detected by binding of a tungstate derivative (as a phosphate analog). Instead of phospho-primed substrate also the C-terminal regulatory domain of CK1δ is able to bind to this position for the purpose of autoregulatory function. Binding of ATP is mainly mediated via the glycine-rich P-loop (L-12, bridging strands β1 and β2), forming the top cover of the WTP binding site, and the so-called catalytic loop (L-67). Conformational changes affecting the activation loop (L-9D) are related to regulation of kinase activity. When the activation loop moves out of the catalytic site the catalytically relevant DFG motif (Asp-149, Phe-150, and Gly-151) shifts to an internal position. The aspartate residue chelates a Mg2+ ion allowing proper binding and orientation of ATP. Another residue, which is essentially involved in the regulation of kinase activity, but also in forming interactions with small molecule inhibitors, is Met-82, the so-called gatekeeper residue. Directly located within the ATP binding pocket this residue controls access of small molecules to certain binding pockets (selectivity pockets) located beyond the position of the gatekeeper.
Additional functional domains
Apart from domains directly involved in catalytic activity further functional domains are present in the CK1δ protein. A kinesin homology domain (KHD) as well as a putative dimerization domain (DD) can be found in the kinase domain. While the KHD allows CK1 isoforms to interact with components of the cytoskeleton. the DD is supposed to be involved in regulation of kinase activity (see below). In the C-lobe, furthermore, a nuclear localization signal (NLS) as well as a centrosome localization signal (CLS) can be found. However, the first one is not sufficient to locate CK1δ to the nucleus.
Regulation of expression and activity
Rigorous control of CK1δ expression and kinase activity is crucial due to its involvement in important cellular signal transduction pathways. Generally, basal expression levels of CK1δ differ between various tissues, cell types, and physiological circumstances. Increased expression levels of CK1δ mRNA can be detected after treatment of cells with DNA-damaging substances, like etoposide and camptothecin, or by γ-irradiation, while increased CK1-specific activity is observed after stimulation of cells with insulin or after viral transformation.
Subcellular sequestration
On protein level, CK1δ activity can be regulated by sequestration to particular subcellular compartments bringing the kinase together with distinct pools of substrates in order to guide its cellular function. This sequestration is usually facilitated by scaffolding proteins, which are also supposed to allosterically control the activity of the interacting kinase. For CK1δ subcellular sequestration has been described to be mediated by A-kinase anchor protein (AKAP) 450, the X-linked DEAD-box RNA helicase 3 (DDX3X), casein kinase-1 binding protein (CK1BP), and the regulatory and complex-building/-initiating molecule 14-3-3 ζ. AKAP450 recruits CK1δ and ε to the centrosome to exert centrosome-specific functions in the context of cell cycle regulation. DDX3X promotes CK1ε-mediated phosphorylation of Dishevelled (Dvl) in the canonical Wnt pathway but has also been demonstrated to stimulate CK1δ- and ε-specific kinase activity by up to five orders of magnitude. On the contrary, proteins being homologous to CK1BP (e.g. dysbindin or BLOC-1 [biogenesis of lysosome-related organelles complex-1]) are able to inhibit CK1δ kinase activity in a dose dependent manner.
Dimerization
Dimerization of CK1δ has also been described as a regulatory mechanism through the interaction interface contained by the DD of CK1δ. Following dimerization, Arg-13 inserts into the adenine binding pocket and prevents binding of ATP and perhaps also of large substrates. Although CK1δ in solution is always purified as monomers, biological relevance of dimerization could be demonstrated by showing that the binding of dominant-negative mutant CK1δ to wild type CK1δ resulted in the total reduction of CK1δ-specific kinase activity.
Site-specific phosphorylation
Posttranslational modifications, especially site-specific phosphorylation mediated either by upstream kinases or by intramolecular autophosphorylation, have been demonstrated to reversibly modulate CK1δ kinase activity. Several residues within the C-terminal regulatory domain of CK1δ were identified as targets for autophosphorylation, including Ser-318, Thr-323, Ser-328, Thr-329, Ser-331, and Thr-337. Upon autophosphorylation sequence motifs within the C-terminal domain are generated, which are able to block the catalytic center of the kinase by acting as a pseudosubstrate. Regulatory function of the C-terminal domain has furthermore been confirmed by the observation that kinase activity is increased after proteolytic cleavage of this domain.
Besides autophosphorylation, site-specific phosphorylation by other cellular kinases has been demonstrated to regulate kinase activity. So far, C-terminal phosphorylation of CK1δ by upstream kinases has been confirmed for protein kinase A (PKA), protein kinase B (Akt), cyclin-dependent kinase 2/cyclin E (CDK2/E) and cyclin-dependent kinase 5/p35 (CDK5/p35), CDC-like kinase 2 (CLK2), protein kinase C α (PKCα), and checkpoint kinase 1 (Chk1). For several phosphorylation events also effects on kinase function have been described. For residue Ser-370, which can be phosphorylated at least by PKA, Akt, CLK2, PKCα and Chk1, major regulatory function has been demonstrated. As a consequence of altered kinase activity of a CK1δ S370A mutant, subsequently affected Wnt/β-catenin signal transduction resulted in development of an ectopic dorsal axis in Xenopus laevis embryos. Further residues targeted by site-specific phosphorylation are depicted in Figure 2. Mutation of identified target sites to the non-posphorylatable amino acid alanine leads to significant effects on catalytic parameters of CK1δ in most cases, at least in vitro.