| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online September 16, 2004 as doi:10.1096/fj.04-1876fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Biochemistry,
Department of Anatomy, Faculty of Medicine, The National University of Singapore, Singapore;
Department of Cell Biology and Physiology and the Centre for Biologic Imaging, University of Pittsburgh Medical School, Pittsburgh, Pennsylvania, USA;
Department of Biochemistry, Hong Kong University of Science and Technology, Kowloon, Hong Kong; and
|| Institute of Molecular and Cell Biology, Singapore
1Correspondence: Department of Biochemistry, Faculty of Medicine, The National University of Singapore, 8 Medical Dr., Singapore, 117597, Singapore. E-mail: bchduanw{at}nus.edu.sg
SPECIFIC AIMS
PRK1/PKN, a cytosolic serine/threonine protein kinase, has previously been described as a RhoA effector. However the basis for PRK1 to signal downstream of activated RhoA at the plasma membrane remains elusive. In this study, biochemical and cell biological evidence is presented and suggests that PRK1/PKN contains a unique form that mediates signal transduction via activated RhoA at the plasma membrane. This has important implications in the context of neurobiology and cancer development
PRINCIPAL FINDINGS
1. Identification of an integral plasma membrane pool of PRK1
Our experimental evidence suggests that PRK1 exists in at least two distinct pools in vivo: a cytosolic/peripheral membrane pool and an integral membrane pool. The integral membrane form of PRK1 was defined by its partitioning into the detergent-rich phase in Triton X-114 phase partitioning and its resistance to extractions with chaotropic agents, alkali, or high concentration of salt (Fig. 1
A, B). It seems that membrane integration is an intrinsic property of PRK1 and the integral membrane pool is not derived from a PRK1 splice variant, as we obtained the same results in an analysis of membrane integration with native PRK1 from NIH 3T3, HEK 293, CHO-K1, and rat brain as with ectopically expressed PRK1. We showed that the integral membrane PRK1 was localized in the lipid rafts of erythrocyte plasma membrane (Fig. 1C-E
).
|
2. The integral plasma membrane PRK1 is functional
We found that the native form of PRK1 from the integral plasma membrane pool of erythrocytes could both autophosphorylate itself and transphosphorylate an S6 peptide substrate. The specific activity of the integral membrane PRK1 ectopically expressed in COS-1 cells was 
120% of that of PRK1 extracted from the aqueous phase of Triton X-114 phase partitioning that was mainly cytosolic/peripheral membrane PRK1.
3. PRK1 integrates with cellular membranes via novel mechanisms
PRK1 does not possess classic transmembrane domains nor the N-terminal methionine-glycine motif or a C-terminal CAAX motif required for N-myristoylation and prenylation, respectively. Biochemical data rule out involvement of lipid anchors in PRK1 membrane integration. By using RhoA binding sites deletion and kinase dead PRK1 mutants, we found that integration of PRK1 into cellular membranes does not require the catalytic competence of PRK1 or interaction with RhoA. Instead, phosphorylation of Ser377 of rat PRK1 (Ser374 in human PRK1) was required for the tight association of PRK1 with membranes.
4. The Ser377 to Ala377 mutation does not compromise known biochemical properties of PRK1
We further found that substitution of serine377 by alanine does not compromise the ability of PRK1-S377A to interact with and be activated by GTP-RhoA nor does it affect its catalytic activity when compared with wild-type protein in the detergent-soluble fraction.
5. Lysophosphatidic acid (LPA) does not activate the nonintegral membrane pool of PRK1
We examined which cellular pool of PRK1 is activated by RhoA upon activation by LPA in vivo. We found qualitative activation of the detergent-rich pool of PRK1 obtained by Triton X-114 extraction, which suggests that LPA activates only the integral membrane form of PRK1. There was no activation by LPA of PRK1 from the nonintegral membrane fraction (aqueous WT-PRK1 or PRK1-S377A). Finally, LPA activated membrane-bound PRK1 by at least 14-fold whereas the soluble PRK1 was minimally activated. Therefore, LPA preferentially activates a form of PRK1 that is tightly membrane bound in cells.
6. RhoA signals via the integral membrane pool of PRK1 in neuronal cells
We next determined whether enhancement of actin/myosin II contractility is mediated by the integral plasma membrane pool of PRK1. We transiently transfected expression plasmids for PRK1 or WT-RhoA into N1E-115 neuroblastoma cells. Once differentiated, these neuronal cells underwent RhoA-mediated neurite retraction and were rounded within 2–3 min of stimulation by physiological concentration of LPA. We noted that WT-PRK1 could mimic RhoA in augmenting LPA-stimulated neurite remodeling. The capacity of WT-PRK1 in modulating cytoskeletal contraction was in marked contrast to that of the PRK1-S377A mutant (P=0.001), which failed to enhance RhoA-stimulated neurite retraction and rounding (Fig. 2
A, B) even though it had a pattern of distribution in neurites and cell bodies hardly distinguishable from that of the WT-PRK1.
|
7. Only the integral membrane pool of PRK1 can initiate RhoA-mediated and ligand-dependent transcriptional superactivation of the androgen receptor in human epithelial cells
Using a cellular model of androgen receptor (AR) -dependent gene expression in prostate cells, we tested whether PRK1-S377A behaves differently from WT-PRK1 regarding transcriptional activation of the AR. HEK 293 cells were transfected with expression constructs for PRK1, AR, and MMTV-luciferase reporter, followed by activation of endogenous RhoA signaling pathways with LPA for 30 h. Only WT-PRK1 enhanced androgen agonist-stimulated AR gene expression (Fig. 2C
), indicating RhoA signals via the integral plasma membrane PRK1.
CONCLUSIONS AND SIGNIFICANCE
Rho GTPases play pivotal roles in the homeostasis of the central nervous system. Dysregulation of Rho GTPases is implicated in the pathogenesis of several neurological disorders. We report here that RhoA signals via an integral plasma membrane form of PRK1/PKN, although only cytosolic forms have been described to date. Our data are consistent with previous findings that RhoA-driven pathways in N1E-115 cells are only fully activated at the plasma membrane. The specificity of signaling between activated RhoA and the integral membrane pool of PRK1 demonstrated here suggests that RhoA may act primarily at specific compartments of the plasma membrane (Fig. 3
A). It is plausible that Rho effectors localize to discrete microdomains at the plasma membrane and suggests that proximity of the effectors for RhoA (and RhoA activators) ensures efficiency and specificity of Rho signaling.
|
Integration of PRK1 into cellular membranes is dependent, in part, on the phosphorylation status of Ser377 residue. Phosphorylation of Ser377 may create a binding motif for PRK1 to interact with cellular chaperone proteins that target PRK1 to the plasma membrane. Alternatively, the phosphoserine377 may allow PRK1 to assume a conformation favorable for its integration into membranes.
It was reported recently that a constitutively active form of PRK1 (i.e., the catalytic domain-only construct) stimulated sphingosine 1-phosphate (S1P)/RhoA-mediated and ligand-dependent activation of the androgen receptor. PRK1 activates AR transactivation in the presence of very low concentrations of adrenal androgens, or even in the presence of the AR antagonist cyproterone. However, it is the full-length PRK1 that is overexpressed in human prostate cancer. We have reconciled this apparent discrepancy by showing that the full-length integral membrane PRK1 and not the soluble counterpart, is able to mediate LPA/RhoA- and R1881-mediated AR transcription activation in cells, indicating that the signaling cascade downstream of Rho leading to activation of AR in cells can only be initiated by the integral membrane protein pool of Rho effectors at the plasma membrane. Our findings begin to shed new light on the molecular mechanisms of how physiological concentrations of bioactive lipids (LPA and S1P), as well as adrenal androgens, participate in the development of "androgen-independent" progression of prostate cancer cells via the novel RhoA-PRK1 signaling pathway (Fig. 3B
).
To our knowledge, this is the first report of an integral plasma membrane pool for mammalian intracellular serine/threonine kinases since the discovery of protein kinases some 50 years ago. The presence of an integral plasma membrane pool of PRK1 reported here is not unique or limited to PRK1. Rather, it seems to be a universal theme for at least the entire protein kinase C family, as we have found the existence of an integral membrane pool for all 10 protein kinase C isozymes. As in the case of PRK1, at least some well-characterized cellular signaling functions of PKC are only mediated by the integral membrane pool of PKC. The findings reported here define a novel paradigm for mammalian cell signaling in which a constitutive integral plasma membrane pool of "cytosolic" serine/threonine kinases are activated in situ at the plasma membrane and transduce signals into the cells, providing a cell biological basis for known millisecond kinetics and specificity of mammalian signal transduction processes at both cellular and organismal levels.
FOOTNOTES
To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.04-1876fje;
This article has been cited by other articles:
|
|
 |
|
  M. Betson and J. Settleman A Rho-Binding Protein Kinase C-Like Activity Is Required for the Function of Protein Kinase N in Drosophila Development Genetics, August 1, 2007; 176(4): 2201 - 2212. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Liu, E. V. Kostenko, G. M. Mahon, O. O. Olabisi, and I. P. Whitehead Transformation by the Rho-specific Guanine Nucleotide Exchange Factor Dbs Requires ROCK I-mediated Phosphorylation of Myosin Light Chain J. Biol. Chem., June 9, 2006; 281(23): 16043 - 16051. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Zhu, Q. Dong, B. J. Tan, W. G. Lim, S. Zhou, and W. Duan The PKC{alpha}-D294G Mutant Found in Pituitary and Thyroid Tumors Fails to Transduce Extracellular Signals Cancer Res., June 1, 2005; 65(11): 4520 - 4524. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
P > 0.05 (compared with empty vector control). C) HEK 293 cells were cotransfected with MMTV-Luc, androgen receptor, and other plasmids. LPA-stimulated activation of AR in the presence or absence of androgen agonist R1881 was assayed in a luciferase reporter gene assay. Luciferase activity was measured as relative light units and expressed as fold induction upon stimulation of 1 µM LPA with or without 10–10 M R1881 (mean±SE, n=6–8).
subunits 12 and 13; RhoGEF, Rho guanine nucleotide exchange factors; ARE, androgen receptor response element; ROCK, Rho GTPase-Rho kinase.