CaM interaction and Ser181 phosphorylation as new K-Ras signaling modulators

Abstract

The small G-protein Ras was the first oncogene to be identified and has a very important contribution to human cancer development (20–23% prevalence). K-RasB, one of the members of the Ras family, is the one that is most mutated and plays a prominent role in pancreatic, colon and lung cancer development. Ras proteins are membrane bound GTPases that cycle between inactive, GDP-bound and active, GTP-bound, states. Most of the research into K-RasB activity regulation has focused on the analysis of how GTP-exchange factors (GEFs) and GTPase activating proteins (GAPs) are regulated by external and internal signals. In contrast, oncogenic K-RasB has a very low GTPase activity and furthermore is not deactivated by GAPs. Consequently, the consensus was that activity of oncogenic K-RasB was not modulated. In this extra view we recapitulate some recent data showing that calmodulin binding to K-RasB inhibits phosphorylation of K-RasB at Ser181, near to the membrane anchoring domain, modulating signaling of both non-oncogenic and oncogenic K-RasB. This may be relevant to normal cell physiology, but also opens new therapeutic perspectives for the inhibition of oncogenic K-RasB signaling in tumors.

Ras proteins are well-known small GTPases involved in the regulation of signal transduction pathways which control processes as varied as cell proliferation, differentiation, survival and apoptosis. Hyperactivation of these proteins is a crucial step in the development of the vast majority of cancers.^1,2

Three different genes code for a total of four different Ras isoforms named H-Ras, N-Ras, K-RasA and K-RasB. Although they have a highly conserved globular domain (from amino acid 1 to 165), the last C-terminal residues of Ras proteins, named the hypervariable region (HVR), are not conserved among the different isoforms; K-RasB (K-Ras) is the one that exhibits the greatest differences in this region. Like all small GTPases, Ras proteins cycle between an inactive GDP-bound state and an active GTP-bound state. In the GTP-bound form, Ras is able to interact with different effector proteins and consequently activates signal transduction pathways. Cycling from the inactive to the active state in the cells is tightly regulated by GTP-exchange factors (GEFs) and GTPase activating proteins (GAPs).^3,4 Moreover, as in many other proteins, the presence of an active protein does not have a single outcome; the outcome is regulated by further post-translational modifications that may regulate interaction with other proteins, intracellular localization, stability, etc. In fact, it is well known that the HVR region of the different Ras isoforms is modified by farnesylation. This modification, together with a second signal, determines their location and transport to specific membrane domains. The second signal in H- and N-Ras comprises palmitoylation of cystein residues, while in K-Ras comprises a polybasic stretch of six lysine residues near the farnesyl group. These basic residues associate with the phospholipid anionic heads at the inner plasmalemmal leaflet. Palmitoylation is not a permanent modification, and cycling between palmitoylated and depalmitoylated states is important for H- and N-Ras signaling regulation and spatial localization.^5,6 Phosphorylation of K-Ras by PKC at Ser181 within the HVR, first described by Ballester et al. more than two decades ago,⁷ now appears to be a post-translational modification with capacity to cause an electrostatic-switch by reducing the positive charge generated by the polybasic region of K-Ras and consequently regulating the second membrane targeting signal of K-Ras.

At present, the main questions being addressed by research groups are: Does K-Ras phosphorylation affect cellular physiology? Does it modify the functionality of both non-oncogenic and oncogenic K-Ras? How does phosphorylation regulate K-Ras activity? Does it modulate interaction with other proteins directly or alter its plasma membrane association? And finally, how is K-Ras phosphorylation regulated? In this extra view article we will try to offer responses to these questions and present speculative models on the basis of our preliminary unpublished data, our recent publication and other previously published results concerning this issue.

Functional Significance of K-Ras Phosphorylation in Cell Physiology and Oncogenicity

The effect of K-Ras phosphorylation on its function is controversial. But considering that K-Ras can induce proliferation, differentiation or even apoptosis, depending on the cell context and the activation kinetics,^8,9 it is also possible to harmonize the results under the hypothesis that phosphorylation increases K-Ras signaling. Firstly, we demonstrated that K-Ras phosphorylation induces a more sustained K-Ras activation. This correlated with a higher ability of GAP to inactivate non-phosphorylable K-Ras, without any change in the affinity between K-Ras and GAP. Suggesting that localization of GAP near K-Ras was hampered by its phosphorylation.¹⁰ Secondly, by using mouse embryo fibroblasts lacking all Ras isoforms, we showed that, under strong stimulation (10% FCS), expression of non-phosphorylable K-Ras could recover the ability of the cells to proliferate to the same extent as a wild-type K-Ras, but, when cells were subjected to lower growth factors concentration (5% FCS) K-Ras phosphorylation was essential for driving normal cell proliferation. This was concomitant with a reduced ability of non-phosphorylable K-Ras to activate PKB.¹⁰

Interestingly, phosphorylation also increases signaling of oncogenic K-Ras. Plowman et al.¹¹ showed that phosphorylation of oncogenic K-Ras increased its capacity to induce PC12 cell differentiation in correlation with increased levels of phospho-ERK1,2 output. It is also well established that expression of oncogenic K-Ras in cultured fibroblasts induces cell transformation. We showed that when oncogenic K-Ras could not be phosphorylated, focus formation capacity, in vitro cell mobility and resistance to apoptosis inducing agents were decreased. Here again, non-phosphorylable oncogenic K-Ras showed an impaired ability to activate PKB and ERK1,2 at very low serum conditions (0.1% FCS).¹⁰ Thus, although the effects related to cellular transformation should be corroborated in vivo, they suggest that, if K-Ras phosphorylation is avoided, its ability to induce cell transformation would be reduced. M. Phillip's group.¹² showed that expression of pseudo-phosphorylated K-Ras induced apoptosis in established cell lines. Induction of apoptosis and senescence by oncogene overexpression has been observed in primary cultures and animals,¹³ and this seems to be due to checkpoint activation in response to oncogenic insults or to replicational stress. The existence of this response in the fibroblast used in those experiments could explain the fact that phosphorylation increases apoptosis under our hypothesis that K-Ras phosphorylation increases K-Ras signaling. Moreover, apoptosis induction by pseudo-phosphorylated K-Ras was mediated by Bcl-x_L, and alteration in the level of expression of this pro-apoptotic protein between different cell types could also modify the response induced by K-Ras phosphorylation.

Creation of knock-in mice expressing non-phosphorylable non-oncogenic K-Ras is essential to determine the role of this phosphorylation in different cell types during development and in adult pluricellular organisms, but no data regarding this issue have been published to date.

Phosphorylation and K-Ras Activity Regulation

As mentioned above, the farnesyl group and the polybasic sequence of the HVR of K-Ras allow its interaction with the inner negatively-charged leaflet of the plasma membrane. It is reasonable to think that introduction of a negative charge in the HVR might modify the plasma membrane affinity of K-Ras and alter its localization. Certain groups have reported that pseudo-phosphorylated K-Ras is internalized to intracellular membranes such as mitochondria, Golgi apparatus, endoplasmic reticulum or endosomes.^12,14 But this issue remains controversial. In fact, our group and others have shown that pseudo-phosphorylated K-Ras is still primarily found in the plasma membrane.^11,15 It has been shown that to avoid interaction of K-Ras with the membrane, at least 3 lysines among the ones found in the polybasic domain should be substituted for non-charged amino acids.^16,17 Considering that phosphorylation of a Ser will reduce the net charge of the region by 2, it would not be surprising that an important percentage of K-Ras is still associated to the plasma membrane. In fact a K-Ras with three lysines substituted by glutamates resulted only in a partial dissociation of the protein from the plasma membrane.¹⁸ Finally, we have also shown that non-phosphorylable K-Ras can be internalized to endosomes¹⁹ and Golgi.¹⁵ We propose that phosphorylation changes K-Ras affinity for the plasma membrane negative domains, allowing its relocalization both to other plasma membrane domains and to internal membranes. Relocalization to either the internal or the plasma membrane will dependent on the cell status. For instance, it has recently been shown that after different apoptotic stimuli anion-rich lipids are exposed in the outer mitochondrial membrane and, consequently, affinity for proteins with polybasic domains (such as K-Ras) is increased.²⁰ Thus, phosphorylated K-Ras localization at the mitochondria might be an effect of an apoptotic state of the cells and may potentially contribute to the execution of programmed cell death.

Localization of K-Ras in nanoscale protein clusters (nanoclusters) which operate as temporary signaling platforms at the plasma membrane is emerging as an important mechanism for regulating Ras signaling.^21,22 These nanoclusters contain certain mixtures of kinases, phosphatases and other signaling proteins that are anchored directly to the membrane or by lipid-anchored regulatory proteins or subunits. Targeting of proteins to these nanoscale domains would increase its effective concentration, hence accelerating reaction rates of signaling events that require specific protein-protein, protein-lipid interactions. These nanoclusters are critical for functional activity and essential for high fidelity signaling through the MAPK pathway.²² Although other scenarios cannot be excluded, our data^10,15 fit with a model in which phosphorylated K-Ras regulates localization of K-Ras in different membrane nanoclusters, and consequently regulates K-Ras signaling. It is also possible that K-Ras phosphorylation status may actively induce a local change in membrane lipids, hence changing nanocluster composition. Similar to the proposal for MARCKS (myristylated alanine-rich C-kinase substrate) protein, the basic cluster of K-Ras could laterally sequester, concentrate, and also hide phosphatidylserine, phospatidylinositol-bis-phosphates (PIP2) and tri-phosphate (PIP3).^23,24 After K-Ras phosphorylation, interaction may decrease and those lipids may be released, being more accessible to different enzymes such as phospholipases or PI3K. To date, the major scaffold protein shown to associate with K-Ras on the plasma membrane is Galectin-3. This interaction is restricted to the K-Ras-GTP nanocluster, but using immunoprecipitation techniques we could not see differences in the interaction depending on K-Ras phosphorylation status (our unpublished results). Furthermore galectin-3 overexpression increases K-Ras nanoclustering regardless of its phosphorylation status.^11,25 Other scaffold proteins, regulators of K-Ras nanoclustering, have been identified, but in these cases the differences depending on K-Ras phosphorylation have not been investigated.²⁶ The Ca²⁺-binding protein calmodulin (CaM), which acts as a second messenger in cell signal transduction pathways and regulates cell proliferation,^27,28 is one of the molecules involved in the modulation of Ras activity.^9,29–31 We have shown that CaM directly binds to K-Ras-GTP and among the important K-Ras regions that mediate this interaction are the farnesyl group and the polybasic domain. It may also be possible that, as in the case of galectin-3, CaM regulates K-Ras nanoclustering at the plasma membrane. But while galectin-3 increases affinity of K-Ras with the plasma membrane, CaM seems to decrease it.³² Nevertheless, we have shown colocalization between K-Ras and CaM at the plasma membrane.¹⁵ Interestingly, K-Ras phosphorylation inhibits CaM interaction, and consequently one might propose the existence of nanoclusters with phosphorylated K-Ras and without CaM and nanoclusters with non-phosphorylated K-Ras and CaM (Fig. 1). At the same time, CaM may attract other CaM-binding proteins to these membrane domains that would modulate the final output of K-Ras signaling. In this way CaM may work as a scaffolding protein. MARCKS, phosphodiesterase and EGF receptor are examples of CaM binding proteins located at the plasma membrane that may modulate K-Ras signaling output.^24,33

Figure 1.

Illustration of the speculative model proposed for the regulation of K-Ras functionality and clustering by PKC phosphorylation. When PKC activity is low and [Ca²⁺] high: K-Ras is cycling between GDP-bound and GTP-bound states depending on the extracellular signalling; K-Ras is found in in membrane microdomains rich in GAPs and consequently K-Ras molecules are immediately inactivated; in the GTP-bound state, K-Ras interacts with CaM which prevents its phosphorylation by PKC; the polybasic domains would be interacting with (and hiding) negatively charged phospholipids and consequently PI3K interaction could be precluded. When PKC activity is high and [Ca²⁺] low: K-Ras becomes phosphorylated, its interaction with the membrane changes and it is allowed to move to other plasma membrane microdomains (maybe due to its affinity for specific scaffold proteins), or to endomembranes; in such domains GAPs concentration would be lower, and consequently the GTP-bound state is maintained for a longer period; phosphorylation of the Ser181 in the polybasic domain of K-Ras displaces this region from the membrane and the phospatidylinositol-bis-phosphates would be accessible to PI3K.

Regulation of K-Ras Phosphorylation

Regardless of how phosphorylation regulates K-Ras activity, since phosphorylation of both oncogenic and non-oncogenic K-Ras has important consequences in cellular physiology, it is extremely important to understand how this phosphorylation is regulated. We recently showed that CaM inhibits K-Ras phosphorylation both in vitro and in vivo.¹⁰ Since CaM interacts directly with K-Ras³⁴ and the polybasic domain of the HVR is important for this interaction,¹⁵ CaM is most probably directly hiding the PKC phosphorylation site. This mechanism of action of CaM is common for other CaM-binding proteins, such as neurogranin, neuromodulin, MARCKS²⁴ and p21.^35,36 Thus, an increase in Ca²⁺ or increased accumulation of CaM at the plasma membrane would prevent K-Ras phosphorylation by PKC (Fig. 1). Our previous data showing that CaM inhibition specifically enhances K-Ras activation only if PKC is active support the notion that phosphorylation increases K-Ras activity.³⁷ Another important issue to be addressed is which PKC isoform is responsible for K-Ras phosphorylation, or even whether PKC is the only kinase able to phosphorylate K-Ras at Ser181. In fact, although K-Ras is phosphorylated by PKC in vitro, and although PKC activity is essential to achieve K-Ras phosphorylation in vivo, other kinases activated by PKC may be directly responsible for K-Ras phosphorylation. Nevertheless, CaM, by directly binding to the HVR region of K-Ras, may block phosphorylation independently of the kinase involved. Thus, modulating CaM binding to K-Ras may be a wise strategy for regulating K-Ras phosphorylation and consequently the activity of oncogenic and non-oncogenic K-Ras in cells.

Since we have shown that K-Ras phosphorylation at Ser181 favours resistance of K-RasG12V expressing cells to adryamicin treatment and that CaM prevents this phosphorylation,¹⁰ inhibiting K-Ras phosphorylation or mimicking CaM binding may be a good strategy for improving anti-tumoral therapy in cancers such as pancreas, lung or colon, where K-Ras has an important role.

We still have much to learn about K-Ras regulation, one of the first oncogenes described, depending on its posttranslational modifications and the effect in protein and lipid interactions and localization in specific membrane microdomains. The coming years are likely to yield exciting findings in relation to all these issues.

Abbreviations

CaM

calmodulin

ERK

extracellular-signal-regulated kinase

GDP

guanine diphosphate

GTP

guanine triphosphate

GEF

GTP-exchange factor

GAP

GTPase activating protein

HVR

hypervariable region

K-Ras

K-RasB

MAPK

mitogen activated protein kinase

MARCKS

myristoylated alanine-rich C-kinase substrate

PI3K

phosphoinositide 3-kinase

PKB

protein kinase B

PKC

protein kinase C

Extra View to: Alvarez-Moya B, López-Alcalá C, Drosten M, Bachs O, Agell N. K-Ras4B phosphorylation at Ser181 is inhibited by calmodulin and modulates K-Ras activity and function. Oncogene. 2010;29:5911–5922. doi: 10.1038/onc.2010.298.