Nature Structural Biology
8, 824 - 826 (2001)
doi:10.1038/nsb1001-824
A cell death-promoting kinaseAdi KimchiAdi Kimchi is in the Department of Molecular Genetics, Weizmann Institute of Science, Rehovot 76100, Israel. adi.kimchi@weizmann.ac.il Death-associated protein kinase (DAP-kinase; DAPk) has been implicated in programmed cell death and tumor suppression. The recently solved crystal structure of the catalytic domain of human DAP-kinase reveals interesting 'fingerprint' regions that may be functionally important.Death-associated protein kinase (DAPk or DAP-kinase) is a multidomain Ser/Thr kinase regulated by Ca2+−calmodulin (CaM). It is localized to the cytoskeleton, specifically in association with the actin microfilaments1,
2. It was originally identified by applying a function-based genetic screen in mammalian cell cultures aimed at discovering novel genes that participate in programmed cell death3. Since its discovery in 1995, this large multidomain protein kinase has been implicated in a wide range of apoptotic systems and in tumor suppression4,
5,
6,
7. In parallel, loss of DAP-kinase expression was detected in a wide spectrum of human cancers, in some cases in association with more aggressive stages of disease8,
9,
10,
11,
12,
13,
14. Thus, it is clear that molecular studies of DAP-kinase are critical to the advancement of several areas of biomedical research and that the protein is a likely drug target.
Both the pro-apoptotic and tumor suppressor functions of DAP-kinase depend on its kinase activity2,
4,
7. This further implies that a better definition of the catalytic domain is necessary for deciphering important features concerning its regulation and mode of action. On page 899 of this issue of Nature Structure Biology, Tereshko et al.15 describe the crystal structure of the catalytic domain of human DAP-kinase. This structure highlights unique features of DAP-kinase in comparison to other Ser/Thr kinases with known structures. One of these features is a highly basic and structured loop that may possess a critical function, based on its specific location in the three-dimensional structure. The importance of this basic loop is further supported by its conservation in the recently identified members of DAP-kinase family of proteins16,
17.
DAP-kinase
The region encompassing the catalytic domain of DAP-kinase, which was utilized by Tereshko et al. for determining the crystal structure, occupies 285 out of the 1,431 amino acid residues of the full length protein. The catalytic domain resides at the N-terminus of the protein and precedes a segment comprising the calmodulin (CaM) binding and regulatory domains (Fig. 1). Ca2+−calmodulin binding to this segment relieves its inhibitory effect on the catalytic site and is necessary for activation of the catalytic domain2. In addition, the protein carries eight ankyrin repeats, two extra-catalytic nucleotide-binding P-loops, a cytoskeleton binding region and a conserved death domain. The death domain is followed by a short Ser-rich stretch of amino acids at the C-terminus of the protein (for a review, see ref. 18).
 | | |
It has been shown experimentally that various domains of the protein contribute to the final cellular action of the protein2,
4,
7. This includes modules that mediate protein−protein interactions, such as the death domain and the ankyrin repeats that may bind downstream or upstream effector proteins, possibly generating muti-protein complexes, or may influence the specificity and/or stability of kinase−substrate interactions. The fact that expression of these individual protein domains by themselves counteracts the effect of the full length DAP-kinase by operating in a dominant negative manner reflects their functional relevance in the context of the entire protein19. Yet, the intact catalytic activity is central for mediating cellular effects. All cellular functions of DAP-kinase examined so far are abrogated when a catalytically inactive mutant — that is, a single point mutation of a conserved Lys (K42A) essential for ATP binding1,
2 — is used experimentally. These functions include the subcellular changes induced by DAP-kinase during programmed cell death2,
4 — for example, membrane blebbing — as well as its tumor suppressing activities7. This implies that phosphorylation of specific substrate(s) is the most critical step in these cellular processes, further emphasizing the importance of resolving the function and regulation of the catalytic domain. The paper by Tereshko et al.15 is the first step in this direction.
The catalytic domain
The X-ray crystal structure reported by Tereshko et al.15 was determined at 1.5 Å — the highest resolution of the 26 protein kinase structures determined to date. The authors have determined several structures of the catalytic domain, both in the free form and in complexes with AMPPNP and with divalent metal ion-bound AMPPNP.
Like every kinase, DAP-kinase displays a classical bilobal conformation, with a smaller and mostly  -stranded N-terminal domain and a larger helix-rich C-terminal domain. In addition to these common conformational features, the crystal structure reveals several important findings whose uniqueness became apparent upon a detailed comparison with the three-dimensional structures of other kinases. First, unlike many other kinases, which display auto-inhibited forms, the catalytic domain of DAP-kinase is in the active 'closed' conformation in the absence of a peptide substrate. Second, and most importantly, a highly ordered basic loop is located on the surface of the N-terminal domain (residues 46−56). The loop extends from the mouth of the ATP-binding cleft, which lies between the two domains, and protrudes above the putative peptide-binding region of the large lobe. This positively charged patch on the molecular surface of DAP-kinase provides a unique 'fingerprint' region that may contribute to its specific functions (see below). Last but not least, the structure reveals that a loop within the putative peptide-binding region, mapped to residues 165−181, has unique flexible features manifested by large conformational variations.
Functional implications
How do all these structural characteristics promote our understanding of regulation, function and mode of action of this interesting death-promoting protein? The most obvious benefit in solving the three-dimensional structure of an enzyme is to provide new tools to identify specific peptide substrates. One initial approach in this direction was recently reported by Velentza et al.20. In this work the authors (who were also involved in the crystallographic work) superimposed the structures of DAP-kinase and phosphorylase kinase, which contains a bound peptide substrate. This provided insight into which regions and, more specifically, which specific sets of amino acids in DAP-kinase are potentially important for recognizing peptide substrates. This information was used as a starting point to construct a positional scanning substrate library and has resulted in the discovery of a tentative synthetic peptide substrate. The search for substrates and specific peptide inhibitors has important clinical implications especially in light of recent work that suggests DAP-kinase is involved in ischemia-induced neuronal cell death21.
Another critical question relates to the functional role of the above-mentioned basic loop present on the surface of the N-terminal domain. Interestingly, this unique basic loop, which is absent from all other kinases with known structures, represents a 'fingerprint' of the new members of the death-associated protein kinase family (Box 1). In principle, this basic loop could be involved in regulating the kinase or in recognizing substrates. In a recent work it was found that mutations that perturbed this loop in DAP-kinase had no significant effect on the Km values of a synthetic peptide substrate20. The lack of additional experimental data concerning other potential effects of these mutations leaves the present status of our knowledge on the functional role of the basic loop in a stage of mere speculation. The basic loop could be a site of interaction with another regulatory protein, or a site of interaction with another domain in DAP-kinase protein itself. In any case it should regulate an event that has a common basic role in all death kinases belonging to the DAP-kinase family. Thus, the report of Tereshko et al.15 provides improtant leads for designing future experiments that otherwise could not be conducted due to lack of critical data.
Finally, the finding that the DAP-kinase catalytic domain is found in its active closed conformation, combined with the flexibility of the peptide binding loop discussed above, is consistent with the lack of activating phosphorylation in this region15. This feature distinguishes members in the DAP-kinase family from many other kinases that are activated by phosphorylation. This further suggests the existence of additional tightly controlled mechanisms that would keep the active kinase in a silent harmless state in viable cells. One obvious mechanism is the auto-inhibition by the calmodulin regulatory segment (see above) that is specifically released once calcium-activated calmodulin binds to it2. This mechanism may be further reinforced by additional layers of regulation that restrain the DAP-kinase function in growing cells to prevent its harmful death effects — as long as the cells are not exposed to external death signals. An example of such regulation was recently identified, involving autophosphorylation of the calmodulin regulatory segment that inhibits the activity of the protein. The theme of auto-inhibition extends to other functional domains of DAP-kinase, as the Ser-rich C-terminal tail also displays auto-inhibitory effects on the death-promoting functions of this protein19.
In conclusion, the structural information provided by Tereshko et al.15 has revealed important new features concerning the function of the catalytic domain in its isolated form. It has also paved the way for future structure-assisted discovery of new active site-directed inhibitors of DAP-kinase as therapeutics. To get a closer physiological insight in the mode of action of DAP-kinase, one must analyze the catalytic properties of the kinase domain in the context of the full length protein and, perhaps, in the context of complexes with other interacting proteins. This study of Tereshko et al.15 is at the beginning of a long and complicated road.
REFERENCES
- Deiss, L.P., Feinstein, E., Berissi, H., Cohen, O. & Kimchi, A. Genes Dev. 9, 15−30 (1995). | PubMed | ISI | ChemPort |
- Cohen, O., Feinstein, E. & Kimchi, A. EMBO J. 16, 998−1008 (1997). | Article | PubMed | ChemPort |
- Deiss, L.P. & Kimchi, A. Science 252, 117−20 (1991). | PubMed | ISI | ChemPort |
- Cohen, O. et al., J Cell Biol. 146, 141−148 (1999). | PubMed | ISI | ChemPort |
- Inbal, B. et al. Nature 390, 180−184 (1997). | Article | PubMed | ISI | ChemPort |
- Kissil, J.L. et al. Oncogene 15, 403−407 (1997). | Article | PubMed | ChemPort |
- Raveh, T., Droguett, G., Horwitz, M.S., DePinho, R.A. & Kimchi, A. Nature Cell Biol. 3, 1−7 (2001). | Article | PubMed | ISI | ChemPort |
- Sanchez-Cespedes, M. et al. Cancer Res. 60, 892−895 (2000). | PubMed | ISI | ChemPort |
- Katzenellenbogen, R.A., Baylin S.B. & Herman, J.G. Blood. 93, 4347−53 (1999). | PubMed | ISI | ChemPort |
- Esteller, M. et al. Cancer Res. 59, 67−70 (1999). | PubMed | ISI | ChemPort |
- Tang, X. et al. J Natl Cancer Inst. 92, 1511−1516 (2000). | Article | PubMed | ChemPort |
- Nakatsuka, S. et al. Laboratory Investigation 80, 1651−1655 (2000). | PubMed | ISI | ChemPort |
- Ng, M.H. et al. Clin. Cancer Res. 7, 1724−1729 (2001). | PubMed | ISI | ChemPort |
- Kim, D.H. et al. Oncogene. 29, 1765−1770 (2001). | Article | ChemPort |
- Tereshko, V., Teplova, M., Brunzelle, J., Watterson, M.D. & Egli, M. Nature Struct. Biol. 8, 899−907 (2001). | Article | PubMed | ISI | ChemPort |
- Inbal, B., Shani, G., Cohen, O., Kissil, J.L. & Kimchi, A. Mol. Cell. Biol. 20, 1044−1054 (2000). | Article | PubMed | ISI | ChemPort |
- Kogel, D. Prehn, J.H. & Scheidtmann, K.H. Bioessays. 23, 352−358 (2001). | Article | PubMed | ISI | ChemPort |
- Cohen. O. & Kimchi, A. Cell Death and Differentiation 8, 6−15 (2001). | Article | PubMed | ISI | ChemPort |
- Raveh, T., Berissi, H., Eisenstein, M., Spivak, T. & Kimchi, A. Proc. Natl. Acad. Sci. USA 97, 1572−1577 (2000). | Article | PubMed | ChemPort |
- Velentza, A.V., Schumacher, A.M., Weiss C., Egli, M. & Watterson, D. J. Biol. Chem. In the press. (published online August 1, 2001). | Article | PubMed |
- Yamamoto, M. et al. J. Neurosci. Res. 58, 674−683 (1999). | Article | PubMed | ISI | ChemPort |
- Shani G. et al., EMBO J. 20, 1−15 (2001). | Article | PubMed |
Acknowledgments
I thank B. Inbal, S. Bialik, G. Shohat and G. Shani for their helpful comments, discussions and advice.
|
