Location, location, location: in situ phosphorylation screen reveals novel cell cycle kinase substrate landscape

From the Clurman Laboratory, Human Biology and Clinical Research Divisions

Protein kinases are enzymes that mediate the addition of a phosphate group to the side chain of one of three amino acids: serine, threonine, or tyrosine –a fundamental mechanism for protein regulation. Cells contain hundreds of different protein kinases, each responsible for phosphorylating a different protein or set of proteins. Phosphorylation can have profound effects on protein conformation and function that regulate many cellular processes, including the complex cascade of events that control cell cycle progression.

The family of cyclin-dependent kinases (CDKs) regulate the transitions through the different phases of the cell cycle by phosphorylating complex substrate networks. Dysregulated cell cycle control –a fundamental aspect of cancer– is often the result of overexpressed or mutated CDKs. Thus, the identification of CDK phosphorylation targets (substrates) is an important goal in the search for cancer therapeutics. However, despite many advances, identifying protein kinase substrates remains challenging, partially because current methodologies rely on non-physiological conditions that disrupt regular cellular interactions. To address this problem, the Clurman lab in the Human Biology and Clinical Research Divisions sought to develop an in situ phosphorylation methodology to identify CDK2 substrates in near-physiological conditions. Principal investigator, Dr. Bruce Clurman, highlighted the importance of the study: “Cyclin-dependent kinases play critical roles in cell division and cancer, and the key to understanding their functions is to identify the proteins they regulate by phosphorylation. People have thus spent a lot of time trying to identify CDK targets. CDK2 is particularly important for many aspects of cell cycle control and it is central to oncogenic signaling”.  

The screen identified many known CDK2 substrates as well as ~70 previously undiscovered candidate substrates. Some of the newly identified candidates participate in the well-studied cell cycle control and DNA metabolism pathways, which are regulated by CDK2. Notably, other candidates are involved in histone modification, chromatin remodeling, and transcriptional regulation –a very different set of candidate substrates from those found previously in vitro. “We think that we were able to find all of these new targets by studying CDK2 within its normal cellular context within the nucleus.”, Dr. Clurman adds. Other substrates are found in multiprotein complexes associated with epigenetic regulation, which suggests possible links between cell division and other cellular processes mediated by CDK2. The group recently published their findings in the journal Science Advances.

To screen for CDK2 substrates in situ, the authors utilized an analog-sensitive kinase version of CDK2 (AS-CDK2) that codes for a mutation in a structurally conserved “gatekeeper” residue at the kinase active site. This gatekeeper mutation enlarges the size of the binding pocket and allows binding of bulky ATP analogs. The bulky ATP analog –in this case, a thiophosphate– functions as a tag that can distinguish the AS-CDK2 substrates from other phosphorylated proteins through mass spectrometry. But there is a caveat: CDK2 is a nuclear kinase, and bulky ATP cannot enter the nucleus; therefore, the use of AS-CDK2 requires cell lysis, which removes nuclear context and architecture. To solve this problem, the researchers isolated the nuclei of cells that express ectopic AS-CDK2 or WT-CDK2 but not the endogenous CDK2 and then incubated the isolated nuclei with the thiophosphate for labeling. To ensure that CDK2 substrates were available for labeling by the thiophosphate, the researchers pretreated the cells with a CDK inhibitor to promote substrate dephosphorylation before labeling in situ. The labeled nuclei were then lysed through sonication, and the resulting nuclear extract was treated with trypsin to generate a mixture of peptides. This peptide mixture was treated with disulfide beads that specifically capture thiophosphopeptides, which can later be eluted from the beads while retaining the phosphate tag for identification through mass spectrometry, a new elution method developed by the researchers that confers a higher accuracy for direct detection CDK2 substrates.

Schematic for in situ phosphorylation screen and pie chart indicating proportion of each gene category.
Schematic for identification of CDK2 substrates using an in situ nuclear labeling assay. Figure provided by Dr. Bruce Clurman.

The screen identified hundreds of unique thiophosphopeptides. To designate potential candidates, the researchers selected thiophosphopeptides that were exclusively found in the AS-CDK2 samples (but not in the WT-CDK2 samples) and contained the canonical CDK phosphorylation motif, a serine or threonine followed by proline. These filters resulted in 117 candidate substrates from which a high proportion (43%) were known CDK2 substrates, which highlights the utility of in situ phosphorylation to identify CDK2 substrates comprehensively. Some of the new CDK2 candidate substrates, especially the set of chromatin-associated proteins, are a surprising finding and likely the result of  in situ conditions that preserve relationships between CDK2 and its nuclear substrates. “By developing methods that allowed us to isolate the activity of CDK2 in a highly physiologic context, we were able to discover many new CDK2 substrates, including the unexpected finding that many of these new substrates are chromatin-associated proteins involved in regulating histone modification and other aspects of gene expression. We also found a number of new substrates with functions in DNA metabolism and repair.”, Dr. Clurman added.

The researchers selected nine substrate candidates that could be readily immunoprecipitated to test the ability of recombinant cyclin-CDK2 in vitro. They found that all nine proteins were directly phosphorylated in vitro, demonstrating that these proteins are direct CDK2 substrates. Finally, the researchers evaluated three candidate substrates in vivo using phospho-specific antibodies that discriminate between the phosphorylated or dephosphorylated forms of a specific protein, or using phosphosite-specific antibodies that can detect a particular phosphate isotope. For all three candidates, the researchers were able to detect phosphorylation signatures that suggest that these substrates are in vivo targets of CDK2.

Dr. Clurman expands on the implications of the study and future directions: “The initial questions involve how these new CDK2 substrates participate in the crucial cellular processes orchestrated by CDK2, and whether we have discovered fundamental new connections between cell division and gene expression?  We are also very interested in determining the roles that these substrates may play in cancer development and maintenance. CDK2 is a major target for cancer chemotherapy and our findings also raise the possibility that new CDK2 substrates may turn out to be novel targets for cancer therapeutics.”

This research was supported by the National Institutes of Health and the National Institute of General Medical Sciences.

UW/Fred Hutch Cancer Consortium member Dr. Bruce Clurman contributed to this work.

Chi, Y., Carter, J. H., Swanger, J., Mazin, A. V, Moritz, R. L., & Clurman, B. E. 2020. A novel landscape of nuclear human CDK2 substrates revealed by in situ phosphorylation. Science Advances, 6(16), eaaz9899. http://doi.org/10.1126/sciadv.aaz9899