Functional analysis of the protein phosphatase activity of PTEN

Abstract

In vitro, the tumour suppressor PTEN (phosphatase and tensin homologue deleted on chromosome 10) displays intrinsic phosphatase activity towards both protein and lipid substrates. In vivo, the lipid phosphatase activity of PTEN, through which it dephosphorylates the 3 position in the inositol sugar of phosphatidylinositol derivatives, is important for its tumour suppressor function; however, the significance of its protein phosphatase activity remains unclear. Using two-photon laser-scanning microscopy and biolistic gene delivery of GFP (green fluorescent protein)-tagged constructs into organotypic hippocampal slice cultures, we have developed an assay of PTEN function in living tissue. Using this bioassay, we have demonstrated that overexpression of wild-type PTEN led to a decrease in spine density in neurons. Furthermore, it was the protein phosphatase activity, but not the lipid phosphatase activity, of PTEN that was essential for this effect. The ability of PTEN to decrease neuronal spine density depended upon the phosphorylation status of serine and threonine residues in its C-terminal segment and the integrity of the C-terminal PDZ-binding motif. The present study reveals a new aspect of the function of this important tumour suppressor and suggest that, in addition to dephosphorylating the 3 position in phosphatidylinositol phospholipids, the critical protein substrate of PTEN may be PTEN itself.

Figure 1. The domain structure of PTEN

Schematic representation to illustrate the major domains of PTEN, the catalytic and C2 domains and the C-terminal PDZ-binding motif (PDZB). The signature motif of the catalytic domain is shown, with the residues that were altered in the phosphatase activity mutants highlighted in bold. The acidic sequence motif in the C-terminal segment of the protein is also shown with the four phosphorylation sites highlighted in bold.

Figure 2. Expression of GFP–PTEN in hippocampal CA1 pyramidal neurons

(A) Top panel: a schematic representation of the position of dendrites imaged in the neurons. Examples of GFP–PTEN/DsRed fluorescence of a transfected hippocampal CA1 neuron (middle panels) and of apical dendrites in a transfected hippocampal CA1 neuron (bottom panels) are presented. White arrows highlight the spines. (B) Comparison of the levels of expression of the various GFP-tagged forms of PTEN, made by measurement of fluorescence intensity of single spines at a fixed region of interest, standardized according to the co-transfected cytoplasmic marker DsRed.

Figure 3. Effect of PTEN expression on spine density in CA1 pyramidal neurons requires its protein, but not lipid, phosphatase activity

Representative images of apical dendrites from biolistically transfected CA1 pyramidal neurons expressing the indicated forms of PTEN in organotypic slice cultures of rat hippocampus. Scale bar=10 μm. The histogram shows quantification of the effects of each plasmid on spine density and each bar is presented below the appropriate image (means±S.E.M., *P<0.0001). n=8 cells for each group.

Figure 4. PTEN displays potential for autodephosphorylation

(A) PTEN dephosphorylated the C-tail peptide in vitro. Purified wild-type and mutant PTEN was tested for protein phosphatase activity using [³²P]P_i-labelled phosphorylated PTEN C-tail peptide as substrate. Activity is expressed as pmol of [³²P]P_i released/min per mg of enzyme. The assay was performed in triplicate. Values are means±S.D. (B) The phosphorylation status of the C-tail segment of PTEN in cells coincided with protein phosphatase activity. GFP or GFP-tagged PTEN constructs were transfected into HEK-293 cells and the expressed proteins were immunoprecipitated (IP) with an antibody against GFP, followed by immunoblotting first with an antibody against phospho-PTEN (Ser³⁸⁰/Thr^382/383) (top panel) and secondly with an antibody against total PTEN (bottom panel).

Figure 5. Effect of mutating phosphorylation sites in the C-terminal segment of wild-type and C124S inactive mutant forms of PTEN on its ability to reduce spine density

Representative images of apical dendrites from biolistically transfected CA1 pyramidal neurons expressing the indicated forms of PTEN in organotypic slice cultures of rat hippocampus. Scale bar=10 μm. The histograms show quantification of the effects of each plasmid on spine density (means±S.E.M., *P<0.0001). n=8 cells for each group.

Figure 6. Expression of the PTEN C2 domain alone did not reduce the spine density in CA1 pyramidal neurons

Effects on spine density of wild-type and isolated PTEN C2 domain were compared. The top panels are representative images of apical dendrites of biolistically transfected CA1 pyramidal neurons expressing the indicated forms of PTEN in organotypic slice cultures of rat hippocampus. Scale bar=10 μm. The histogram shows quantification of the effects of each plasmid on spine density (means±S.E.M., *P<0.0001). n=8 cells for each group.

Figure 7. Effect of the PTEN PDZ domain-binding motif on spine density in neurons

(A) The PTEN PDZ-binding domain is required for the effects of PTEN on spine density in CA1 pyramidal neurons. The effects on spine density of wild-type and mutant PTEN-(1–398), which lacked the extreme C-terminal five residues of the protein, were compared. (B) Effects on spine density of a PTEN-(354–403) truncation mutant, which comprised only the C-terminal segment of the protein, were tested to determine the importance of the phosphorylation sites and the PDZ domain-binding motif in this construct. The top panels are representative images of apical dendrites of biolistically transfected CA1 pyramidal neurons expressing the indicated forms of PTEN in organotypic slice cultures of rat hippocampus. Scale bar=10 μm. The histograms show quantification of the effects of each plasmid on spine density (means±S.E.M., **P<0.0001). n=8 cells for each group.