Mollier L. Kelly and Jonathan Chernoff; Molecular and Cellular Biology; doi:10.1128/MCB.01267-10
p21-activated kinases (PAK) are a family of Cdc42/Rac-activated serine/threonine kinases involved in a variety of cellular processes, such as motility, migration and cytoskeletal reorganization. Various members of the PAK family are known to influence brain development and cognitive function, but there remain many unanswered questions regarding PAK functions in neurons, the significance of individual PAK isoforms and their molecular mechanisms of action. A new genetic study by Huang and colleagues provides insight into how murine PAKs, specifically PAK1 and PAK3 are involved in brain development through morphogenesis of dendritic spines as well as development of synaptic networks. As PAK3 mutations in humans cause mental retardation, the findings in this new study may lead to new approaches to treat disorders of cognitive function.
On structural and biochemical grounds, the PAK family of kinases can be divided into two groups, group I(PAK1, –2 and –3) and group II (PAK4,-5 and –6). All PAKs contain a highly conserved N-terminal Cdc42/Rac-binding domain and C-terminal protein kinase domain but differ substantially between these two domains. Within the group I PAKs, PAK1 and PAK3 are nearly identical in amino acid sequences (81% over all and >95% in the Cdc42/Rac-binding and kinase domains). and both are highly expressed in the brain. These features suggest that PAK1 and PAK3 share redundant functions. However, despite these close similarities only PAK1 has a binding site for LC8, which is required for its nuclear entry. Since nuclear functions of PAK1 play an important role in development in zebrafish as well as in signaling in mammalian cells one might suppose on these grounds that PAK1 and PAK3 functions are nonredundant. Thus, both the similarities and the one known difference between PAK1 and PAK3 have made it difficult to determine the unique functions, if any. of each of these two isoforms.
Due to their high levels of expressions in the brain both PAK1 and PAK3 have been closely examined for their role in nerve cell function. A number of studies suggest a role for PAK1 in regulating dendritic spine morphology although the underlying mechanism remains unclear. PAK3 has also been implicated in having a role in neuron development and plasticity. The most powerful evidence in this regard has come from studies of human mental retardation which revealed a causal association between certain X-linked nonsyndromic forms of mental retardation and PAK3 loss-of-function mutations. To date, five distinct point mutations have been found in various kindreds:three in the kinase domain (abolishing kinase activity), one in the Cdc42/Rac- binding domain (abrogating binding to these GTPases) and one in an intron (causing a premature stop). Patients with these mutations display hyperactivity, excessive anxiety, restlessness and impaired memory. Such behavioral and cognitive impairments are often associated with abnormal neuron plasticity, suggesting a plausible link to PAK1/PAK3 cellular functions in neurons. However, the exact roles of these proteins in brain development and their levels of redundancy are unknown. Additionally, the role of these PAKs in the adult brain and their contributions to a stable neuron network have yet to be elucidated.
To understand the mechanism by which PAK1 and PAK3 contribute to brain function, single-knockout mouse models have been produced. However, mouse knockouts of PAK1 and PAK3 have been notable for their lack of dramatic neuronal phenotypes. For example, loss of PAK1 alone gives rise to modest defects in long-term potentiation in hippocampal CA1 synapses, as does loss of PAK3 alone but neither are associated with neuroanatomic defects or abnormalities in cellular actin structures. Not so the double knockout. When Huange et al. crossed PAK1 and PAK3 null mice, the resulting double knockout mice were normal at birth but soon showed major loss of brain volume compared to that of wild type mice, despite normal brain organization. PAK1/PAK3 double knockout mice also had severely impaired learning and memory and hyperactive behavior, a phenotype that echoes that seen in human patients with mutations in PAK3. For these reasons, the PAK1/PAK3 double knockout mouse represents a useful model to study the role of PAK function in postnatal brain development and a plausible platform for testing therapeutic agents.
What is the cellular basis for the brain defects in PAK1/PAK3 knockout mice? To answer this question, Huang et al. evaluated morphogenesis and maturity of the neurons. As PAK1 and PAK3 are both expressed in neurons after mitosis and in differentiated neurons, the authors hypothesized that neuron maturation is affected in the double knockout. The group found that the morphology of neurons was much less complex with reduced dendrite length and number of dendritic tips in the double knockout mice, showing that PAK1 and PAK3 are involved in branch formation. Surprisingly, however, the double knockout mice displayed enhanced synaptic transmission.This is likely related to the fact that the few synapses that were present on neurons in double knockout mice were functionally more potent, leading to the apparent enhanced synaptic transmission. This finding shows that PAK1 and PAK3 are essential for normal spine morphology and synaptic properties. Such defects in dendritic spine structure are likely to explain their further finding that long-term potentiation and depression were significantly reduced in the double- knockout mice. These are noteworthy observations, as long-term potentiation and depression are important in learning and memory and may explain the diminished cognitive function in the double knockout mice.
Given these cellular and physiological defects other questions remain: what molecular pathways do these kinases regulates and why is deletion of both required to obtain a phenotype? To assess the biochemical mechanisms by which PAKs exerts their effects in neurons, Huang et al. examined the activity of a number of well known PAK regulated pathways . Surprisingly, the combined loss of PAK1/PAK3 had no effect on extracellular signal regulated kinase (ERK) activity or on other well studied PAK substrates, such as myosin light chain kinase. In fact, of the seven substrates tested, phosphorylation was significantly reduced in only one, cofilin, which is a target of PAK substrate LIM kinase (LIMK). That cofilin should be involved makes perfect sense, as the major phenotype observed in the double-knockout neurons, namely, altered dendritic spine morphology, is consistent with the loss of F-actin and a natural consequence of cofilin activation, which is associated with the unphosphorylated state. An interesting corollary that emerges from these data is that the remaining group I PAK, PAK2, which is also highly expressed in neurons, must lack the ability to phosphorylate LIMK. Perhaps the lack of effect on other PAK-activated pathways in the double knockouts is due to redundancy among the three group I PAKs in most signaling circuits in neurons, with LIMK/cofilin representing an important exception.
The findings of Huang et al. could have implications beyond mental retardation, as synaptic dysfunction associated with underphosphorylated cofilin underlies the cognitive impairments accompanying a wide range of neurological disorders and normal aging. Thus, the PAK1/PAK3 double knockout model, in revealing what had been obscured by redundant group I PAK functions, enlightens us regarding broader issues of brain development and function.
No comments:
Post a Comment