3-Phosphoinositide-Dependent Kinase 1 (PDK1)
The 3-phosphoinositide-dependent kinase-1 (PDK1) is a well-studied member of the AGC kinase subfamily which is implicated in many physiological functions and when dysregulated can play a role in cancer, among other pathological conditions. The AGC kinases can act on tyrosine (Tyr), serine (Ser), or threonine (Thr) residues of their substrates and comprise of 60 members, for most of which there are a number of splice variants isoforms. For the majority of the AGC kinases, phosphorylation of the T-loop in the catalytic domain and the hydrophobic motif in the noncatalytic region leads to activation of the kinase. Apart from these two highly conserved motifs, the turn motif can also be phosphorylated in several members of the ACG kinases family. In regard to interaction with their substrates, the specificity of the kinases is determined by the residues in the peptide binding site. PDK1 specifically contains a 100 amino acids long pleckstrin homology (PH) domain which binds to phosphatidylinositol lipids. Generally, AGC kinases can be activated by various extracellular stimuli (Calleja et al. 2014).
PDK1 Structure and Function
The origin of PDK1 is estimated 2.3 billion years ago and its maintenance and conservation throughout the eukaryotic species suggest that it is crucial for survival. Indeed, PDK1 knock-out mice are not viable and die during the embryonic development. The actual PDK1 protein is 556 amino acids long and is comprised of two main domains: an N-terminal kinase domain and a C-terminal pleckstrin homology (PH) domain. The kinase domain itself can be subdivided into an N-terminal and C-terminal region comprising the PDK1-interacting fragment (PIF)-binding pocket and the T-loop (Gagliardi et al. 2015). PDK1 had initially been known for its role in glucose metabolism since, following the binding of insulin to its receptor and subsequent phosphoinositide 3-kinase (PI3K) activation, it would localize to the plasma membrane together with Akt and activate it. Akt would then inhibit via phosphorylation the glycogen synthase kinase 3 (GSK3), and glucose storage would be facilitated (Cohen et al. 1997).
PDK1 Subcellular Localization and Activation and Deactivation Sites
PDK1 is characterized as a master regulator of numerous AGC kinases and understanding how the kinase itself is regulated, as well as how it interacts and modulates its substrates both spatially and temporally, is of great interest. PDK1 is one of the major components through which the PI3K/phosphatase and tensin homologue deleted on chromosome 10 (PTEN) signaling pathway regulates various cellular processes. PDK1 is found to be constitutively active within the cell due to a Ser241 trans-autophosphorylation in the activation segment residue, and it has also been found in the nucleus, with phosphorylation on a noncatalytic site (Ser396) inhibiting the nuclear export regardless of growth factors stimulation. Different mechanisms contribute to the distribution of PDK1 within the cells, including anchoring of the protein to the cytosol by binding to soluble inositol phosphates, and over time various cases of acute PDK1 regulation have emerged, such as homodimerization and subcellular localization. For this kinase, phosphorylation on Ser/Ther as well as Tyr sites has been reported. Among major PDK1 phosphorylation sites are Ser 25, 241, 393, 396, and 410, but only Ser241 is essential for PDK1 to exert its activity. Interestingly, this particular site is also not accessible by phosphatases and thus it is resistant to protein phosphatase 2A (PP2A) activity. Notably, although Ser241 autophosphorylation is essential for PDK1 activation, it does not suffice; full activation is accomplished upon Thr513 trans–autophosphorylation within the PH domain, in the simultaneous presence of phosphatidylinositol 3,4,5 trisphosphate (Ptdlns(3,4,5)P3). This highlights the regulatory role of the PH domain – Ptdlns(3,4,5)P3 interaction in the activation of PDK1. Other PDK1 worth-mentioned phosphorylation sites are Ser160, which is PI3K-dependent and stabilizes the protein’s active conformation, Ser501, which is PKCθ-dependent and abolishes the proteins’ activity upon platelet-derived growth factor or insulin stimulation, and, last but not least, Thr354 and Ser398 and 394 which in cooperation decrease PDK1 activity. The proximity of these negative-regulating phosphorylation sites to formerly described activation sites leads to the notion that there is a mutual exclusion governing positive and negative regulative mechanisms.
Apart from phosphorylation, another way of enhancing and stabilizing PDK1 activity is binding to regulators or other substrates. For instance, the interaction between Src protein and PDK1 gets stabilized by the binding of heat-shock protein 90 (HSP90) to the latter, which furthermore increases the kinase’s activation. Notably, both HSP90 and PDK1 are found in elevated levels in various cancers, leading to the conclusion that stabilizing PDK1 is very important in the cancer context (Calleja et al. 2014).
Apart from anchoring to the plasma membrane, PDK1 can be found predominantly in the cytoplasm but also in the nucleus, due to the existence of a nuclear export signal (NES). Retainment of PDK1 in the nucleus can be the result of Leu380/Phe383 mutation, and it is critical for tumorigenesis since, once in there, PDK1 inhibits the forkhead box O3A (FOXO3A) and the c-Jun N-terminal kinase (JNK) – thus promoting proliferation and survival, respectively, and increases intranuclear phospho-Akt. All these events are correlated with solid tumor progression and anchorage-independent growth in in vivo mouse models (Kikani et al. 2012).
PH Domain and PIF Pocket in PDK1 Function
Through its C-terminal PH domain which interacts mostly with Ptdlns(3,4,5)P3, PDK1 is able to localize at the plasma membrane together with Akt, in this way phosphorylating the latter at Thr308 in its activation segment, after it has changed conformation following interaction with Ptdlns(3,4,5)P3 or phosphatidylinositol 3,4-bisphosphate (Ptdlns(3,4)P2). In in vivo models there have been cases where the activation segment of Akt was still phosphorylated even when PDK1 was mutated in a way that it could not interact with phosphoinositides, leading to the notion that there might be additional components facilitating PDK1-Akt communication, such as the five repressor element under dual repression-binding protein 1 (FREUD1) or the growth factor receptor-bound protein 14 (GRB14) (Pearce et al. 2010). GRB14 has been reported to modulate PDK1 compartmentalization together with the insulin receptor which is localized in the plasma membrane (King and Newton 2004). Via a docking site in the kinase domain, the PIF pocket, PDK1 is also able to interact with substrates lacking a PH domain, such as p70 ribosomal S6 kinase (S6K) and serum-and glucocorticoid-induced protein kinase (SGK) – once their hydrophobic motifs are phosphorylated, and also members of the protein kinase C (PKC) family. Thanks to these distinct mechanisms, PDK1 is able to coordinate the regulation of its downstream targets and activate them independently (Pearce et al. 2010). Interestingly, a novel PDK1 target that did not belong to the AGC kinases family was recently revealed. More specifically, a study demonstrated that phospholipase Cγ1 (PLCγ1) activation is controlled by PDK1, and experiments involving chemical inhibition of PDK1 as well as genetic silencing of both PDK1 and PLCγ1 demonstrated that the two molecules act on the same pathway and are a prerequisite for cancer cell invasion. Therefore, PDK1 can create a bridge between PI3K signaling pathway and PLCγ1, in which both PDK1 and PLCγ1 would translocate at the membrane following PI3K activation and form a protein complex (Arteaga et al. 1991; Sala et al. 2008; Maurer et al. 2009; Raimondi et al. 2012). Notably, both PLCγ and PDK1 are found to be overexpressed in breast cancer especially in case of metastatic disease.
PDK1 and Cancer Invasion
As aforementioned, dysregulation of the AGC kinases leads to pathologic conditions, and studies have reported PDK1 alterations in both genetic and protein levels in human malignancies. For instance, whole gene or specific locus amplification have been reported in breast (Maurer et al. 2009) and prostate carcinoma, respectively, and protein overexpression has been reported in melanoma, esophageal squamous cell carcinoma, and acute myeloid leukemia. Moreover, this protein is involved in different hallmarks of tumor invasion (Gagliardi et al. 2015). To begin with, PDK1 is involved in the formation of protrusive structures in cancer cells called invadopodia. It is the PI3K pathway that triggers the formation of these structures and more specifically the p110α subunit which activates the cascade through PDK1 and Akt. In fact, this pathway is not only pivotal for invadopodia formation but also for their functionality and ability to facilitate cancer dissemination by degrading the basal membrane and the extracellular matrix (Murphy and Courtneidge 2011; Yamaguchi et al. 2011; Kung et al. 2012). Secondly, it was shown to activate ROCK1 kinase and in that way modulating amoeboid invasion, in which case when cancer cells invade a healthy tissue they simply incorporate between the fibers instead of degrading the extracellular matrix (Sahai and Marshall 2003; Pinner and Sahai 2008). Finally, PDK1 is implicated in the so-called mesenchymal and collective type of cancer invasion through activating MRCKα. In this case, tumor cells degrade the extracellular matrix and adhere to its components (Gagliardi et al. 2014).
To conclude, PDK1 is a very old and highly preserved kinase, implicated in physiological and pathological conditions and with a crucial role in multiple signal transduction pathways. Recent findings suggest that PDK1 loss-of-function or impairment of its activity can restrain cancer progression and induce apoptosis, and interestingly, in vivo work on pancreatic ductal adenocarcinoma driven by an oncogenic K-Ras mutation showed that PDK1 conditional ablation could confer normal life expectancy. This emphasizes the importance of the K-Ras/PI3K/PDK1 axis in this specific type of cancer and makes PDK1 an intriguing target for therapeutic intervention (Eser et al. 2013). Since K-Ras is considered to be an “undruggable” target, alternative ways of targeting this specific pathway have been explored, such as PDK1 inhibition. Another alternative way for impairing this axis is to target microRNAs. For instance, miR-375, which is upstream of many important oncogenes, has been found to be downregulated in many cancers including pancreatic cancer. Overexpression of this microRNA is shown to suppress the levels of transcription of pdk1, as well as its protein levels (reviewed in Ferro and Falasca 2014). Despite the research carried out up to date, there are still many aspects that need to be further elucidated, such as the precise spatial and temporal PDK1 regulation, as well as the development of effective and specific inhibitors.
- Cohen P, Alessi DR, Cross DAE. PDK1, one of the missing links in insulin signal transduction? FEBS Lett. 1997;410:3–10. 1 The previous Datta Lectures were given by: F. Melchers 1997(1st, 1986), N. Sharon (2nd, FEBS Lett. 217 (1987) 145–157)C.Google Scholar
- Gagliardi PA, Di Blasio L, Primo L. PDK1: A signaling hub for cell migration and tumor invasion. BBA – Rev Cancer. 2015;1856(2):178–88.Google Scholar