Regulation of protein prenyltransferase in central neurons

Abstract

Geranylgeranyltransferase I (GGT) is a protein prenyltransferase that mediates lipid modification of some proteins such as Rho family small GTPases. Since the activation of Rho GTPases mediates tumorgenesis and metastasis, GGT has become an attractive target for anti-tumor drug design. Although GGT is extensively expressed in the brain, the function of GGT in central nervous system (CNS) is totally unknown. We have previously shown that GGT was involved in neuromuscular synaptogenesis. In this study, we report that neuronal activity- and brain-derived neurotropic factor (BDNF)-dependent dendritic morphogenesis requires activation of GGT. Furthermore, GGT was activated by depolarization or BDNF in cultured neurons or in hippocampus of the mice under novelty exploration test, suggesting that neuronal activity activates GGT in vitro and in vivo. In this addendum, we further discuss the significance of this study and the possible implication to the field.

Geranylgeranyltransferase I (GGT) is a protein prenyltransferase, which catalyzes the attachment of 20 carbons long prenyl lipid anchors to the carboxyl terminal (C-terminal) of a variety of eukaryotic proteins.¹ It is composed of two subunits—a distinct β subunit (GGTβ) and a common α subunit (GGTα) that is shared with farnesyltransferase (FT). Substrate proteins of GGT and FT contain a C-terminal “CAAX” box, where C is an invariant cysteine residue fourth from the C terminus, A is an aliphatic amino acid, and X is any amino acid.^1,2 Substrate proteins undergo geranylgeranylation if X is Leu or farnesylation if X is Ser, Met, Ala or Gln. Known targets of GGT include the γ subunit of brain heterotrimeric G proteins and Rho small GTPases such as Rac, RhoA and Cdc42. Substrates for FT in mammalian cells include all known Ras proteins, nuclear lamins A and B, the γ subunit of the retinal trimeric G protein transducin, rhodopsin kinase.¹ Because the targets of GGT and FT are tightly related to tumorgenesis and metastasis,^3–5 GGT and FT have become very attractive targets for anti-tumor drug design.

Protein prenyltransferases are usually believed to function constitutively. However, recent studies show that FT and GGT can be regulated by multiple factors in non-neuronal cells, such as breast cancer cells or skeletal muscle cells.^6–8 We found that motor neuron-derived glycoprotein agrin activates GGT through the muscle specific receptor tyrosine kinase MuSK, leading to acetylcholine receptor clustering at the postsynaptic membrane.⁹ Although GGT is extensively expressed in the brain,^10,11 the function of GGT in central nervous system (CNS) and whether it is also regulated through receptor tyrosine kinase during development of central neurons is totally unknown.

We found that the expression of GGTα or GGTβ in cultured hippocampal neurons gradually increased and peaked at approximately 8 days in vitro (DIV), a pattern that coincides with the period of extensive dendrite growth. Overexpression GGTβ, but not FTβ, promoted dendritic branching significantly in cultured hippocampal neurons compared with vector control. Treatment with GGT specific inhibitor GGTi-2147 reduced the dendritic development. The role of GGT in dendrite development was further confirmed by downregulating GGT using the approach of small interference RNA. This result indicates that GGT is sufficient and necessary for dendrite development.¹²

Since neuronal activity and several secreted factors, including BDNF and insulin-like growth factor 1(IGF-1), are known to promote dendrite growth,^13–18 we investigated whether these factors regulate GGT activity. We found that GGT was activated by high KCl-induced depolarization or BDNF in cultured neurons. Although insulin and IGF-1 can activate GGT in non-neuronal cells,^7,19 neither insulin nor IGF-1 had any effect on GGT activity under our experimental conditions. In agreement with these results, high KCl treatment increased the membrane level and activity of Rac, the known substrate of GGT. Moreover, exploration of a novel environment that is known to increase neuronal activity caused activation of GGT in the mice hippocampus, suggesting that neural activity activates GGT in vivo.

GGT is associated with muscle specific receptor tyrosine kinase MuSK at the neuromuscular junction.⁹ Here we found that GGT was physically associated with tropomyosin-related kinase B (TrkB), the receptor for BDNF, and this association was enhanced by depolarization. When GGT-TrkB interaction was interrupted by dominant-negative GGTα mutant K164A or Y200F, which abolished GGT kinase activity, high KCl or BDNF induced dendritic arborization was suppressed. This result indicates that GGT-TrkB association is required for high KCl or BDNF induced dendritic arborization. To investigate the role of TrkB in neuronal activity-induced GGT activation, we treated hippocampal neurons with K252a, a tyrosine kinase inhibitor, or a specific inhibitor for TrkB signaling—TrkB-Fc chimera protein (extracellular domain of human TrkB fused to the C-terminal histidine-tagged Fc region of human IgG₁). We found that high KCl induced GGT activation was blocked by pretreatment with K252a or TrkB-Fc (Fig. 1), suggesting that KCl activates GGT through TrkB.

Figure 1.

KCl depolarization-induced GGT activation was blocked by K252a or TrkB/Fc. Neurons at DIV6 were pretreated with K252a or TrkB/Fc respectively for 45 min, then treated with KCl for another 45 min. After treatment, neurons were lysed and GGT activity was assayed by using Dansyl-GCVLL as substrate. Data shown are mean ± SEM of three independent experiments, with each performed in triplicate. All values were normalized to control neurons. *p < 0.05; ANOVA with Tukey test.

Rac is important for dendritic development and its amino acid residue cystine in the C-terminal CAAX box is the key site for geranylgeranylation.^1,20 We generated two mutated forms of Rac1—Rac1^ΔC and Rac1^C189S—by deleting the C-terminal CAAX box or substituting the cystine with serine, respectively. We found that although Rac1^ΔC and Rac1^C189S were unable to associate with plasma membrane, they remained the ability to interact with Tiam1, a specific Rac GEF that has been shown to mediate TrkB-mediated Rac activation.²¹ We reasoned that these mutated forms of Rac1 might act as dominant-negatives by competing with endogenous Rac for GEFs such as Tiam1. In line with this notion, overexpression of these Rac mutants caused inhibition of high KCl-induced activation of endogenous Rac in cultured neurons. Furthermore, overexpression of these mutated forms of Rac attenuated dendrite arborization induced by GGT overexpression or treatments with KCl or BDNF. Therefore, the prenylation of Rac1 is required for depolarization- or BDNF-induced Rac activation and dendrite development.

In summary, we have identified GGT as important regulator for dendritic morphogenesis and found that GGT activity can be strictly regulated by neuronal activity or neurotrophin. Most studies on the regulatory mechanisms of Rho small GTPases focus on GTPase activating proteins (GAPs) or guanine nucleotide exchange factors (GEFs).^20,22 For example, neuronal activity or BDNF regulates Rho small GTPases through Rho-GEF Trio or Rac1-GEF Tiam1.^21,23 Here we identify an alternative mechanism by which neuronal activity or BDNF regulates Rac activity through activation of GGT. The prenyl moiety is derived from the mevalonate/cholesterol synthetic pathway and this pathway has been shown to affect neural plasticity or synaptogenesis.^24–26 Since GGT is implicated in the development of neuromuscular junction and its partner TrkB is important for synapse formation by interacting with PSD95, the main scaffold protein of the postsynaptic structure,^27,28 it shall be interesting to determine the role of protein prenylation in regulating other neuronal functions, such as synaptic plasticity.

Addendum to: Xiu-Ping Zhou, Kong-Yan Wu, Bin Liang, Xiu-Qing Fu, Zhen-Ge Luo. TrkB-mediated activation of geranylgeranyltransferase I promotes dendritic morphogenesis. Proc Natl Acad Sci USA. 2008;105:17181–17186. doi: 10.1073/pnas.0800846105.