Molecular Features of Phosphatase and Tensin Homolog (PTEN) Regulation by C-terminal Phosphorylation*

PTEN is a tumor suppressor that functions to negatively regulate the PI3K/AKT pathway as the lipid phosphatase for phosphatidylinositol 3,4,5-triphosphate. Phosphorylation of a cluster of Ser/Thr residues (amino acids 380–385) on the C-terminal tail serves to alter the conformational state of PTEN from an open active state to a closed inhibited state, resulting in a reduction of plasma membrane localization and inhibition of enzyme activity. The relative contribution of each phosphorylation site to PTEN autoinhibition and the structural basis for the conformational closure is still unclear. To further the structural understanding of PTEN regulation by C-terminal tail phosphorylation, we used protein semisynthesis to insert stoichiometric and site-specific phospho-Ser/Thr(s) in the C-terminal tail of PTEN. Additionally, we employed photo-cross-linking to map the intramolecular PTEN interactions of the phospho-tail. Systematic evaluation of the PTEN C-tail phospho-cluster showed autoinhibition, and conformational closure was influenced by the aggregate effect of multiple phospho-sites rather than dominated by a single phosphorylation site. Moreover, photo-cross-linking suggested a direct interaction between the PTEN C-tail and a segment in the N-terminal region of the catalytic domain. Mutagenesis experiments provided additional insights into how the PTEN phospho-tail interacts with both the C2 and catalytic domains.

PTEN is a ϳ45-kDa protein composed of an N-terminal PTP-type phosphatase domain, a midsection phospholipid membrane interacting C2 domain, and a 50-residue regulatory tail (1). A crystal structure of the core PTEN catalytic-C2 region has revealed intimate interactions between the catalytic and C2 domains but omitted the C-terminal tail, which is considered to be flexible and largely unstructured (11). One of the most intensively studied set of PTMs in PTEN is a cluster of four C-terminal phosphorylations at Ser 380 , Thr 382 , Thr 383 , and Ser 385 (12)(13)(14)(15). These phosphorylation events have been suggested to be catalyzed by protein kinases casein kinase 2 and/or glycogen synthase kinase 3␤ and demonstrated to drive a conformational change that inhibits membrane interactions and reduces catalytic activity (12)(13)(14)(15).
Several reports that shed light on the molecular basis for this phospho-dependent regulatory event suggest that the PTEN tail phosphorylation induces conformational closure involving intramolecular interactions between the tail and the CBR3 loop of the C2 domain of PTEN (14 -16). However, important aspects of this structural model remain undetermined. It is not known whether all four tail phosphorylation events are necessary for PTEN conformational closure or whether specific individual sites are critical for this tertiary structural change. Recently, a mutant form of PTEN, ePTEN, was identified through a genetic screen as a form of PTEN that constitutively binds the membrane, despite C-terminal phosphorylation (17). The five mutations in ePTEN, Q17R, R41G, E73D, N262Y, and N329H, are widely spaced throughout the PTEN domains (17). It was hypothesized, but not investigated in detail, that these mutations disrupt the phosphorylation-driven conformational change in PTEN (17). Other studies on PTEN including hydrogen/deuterium exchange experiments have suggested that the catalytic domain and/or the N terminus may also interact with the phosphorylated tail (18).
To help clarify these phospho-PTEN structural issues, we have applied expressed protein ligation (19) to generate several semisynthetic phosphorylated forms of PTEN, and we assessed their activity and tail interactions. We demonstrated that there is an approximately proportional relationship between the number of phosphates and degree of conformational closure. In contrast to expectations, tetraphosphorylated ePTEN (4p-ePTEN) showed a closed conformation similar to that of WT 4p-PTEN. However, mutations of a cluster of Lys and Arg residues in the C␣2 loop relaxed phospho-tail interactions with the PTEN body. The incorporation of a benzolyphenylalanine in the C-tail was employed in photo-cross-linking experiments and provided direct evidence for a tail-catalytic domain interaction. Here, we describe how this biochemical analysis of a new set of semisynthetic PTENs has led to significant insights into the structural basis for the effects of PTEN tail phosphorylation.

Experimental Procedures
Reagents-Soluble diC6-PIP 3 was purchased from Avanti Polar Lipids (Alabaster, AL). Anti-phospho-PTEN antibody was from Abcam (ab131107), and anti-total PTEN antibody was from Santa Cruz (sc-7974). Calf intestinal alkaline phosphatase was purchased from New England Biolabs (M0290). MESNA was supplied from Sigma. All canonical Fmoc amino acids were from EMD (Billerica, MA), whereas Fmoc-L-p-benzoylphenylalanine and Fmoc-L-Lys(biotin)-Wang resin were purchased from Novabiochem and Iris Biotech, respectively. All other reagents were of the highest quality and were purchased from either Sigma or Fisher.
Peptide Synthesis-Peptides were synthesized using the PS3 peptide synthesizer or the Prelude X peptide synthesizer from Protein Technologies (Tuscon, AZ) using conventional Fmoc solid phase peptide synthesis strategies. Synthetic peptides were estimated to be more than 90% pure after HPLC, and their correct structures were confirmed by MALDI mass spectrometry. Peptide samples were spotted using the sandwich method with ␣-cyano-4-hydroxycinnamic acid as the matrix (10 mg/ml ␣-cyano-4-hydroxycinnamic acid in 50% acetonitrile, 5 mM ammonium citrate). Mass spectra were acquired using the Voyager-DE STR biospectrometry work station (Applied Biosystems) by averaging 100 shots using a linear detector in either positive(unphosphorylatedpeptides)ornegativeion(phosphorylated peptides) mode.
Generation and Purification of Semisynthetic PTEN-Expressed protein ligation was used to generate different semisynthetic phosphorylated PTENs as previously reported by Bolduc et al. (14). In brief, a C-terminal truncated PTEN (amino acids 1-378) fused to Mycobacterium xenopi GyrA intein, and chitin-binding domain (truncated PTEN-intein-CBD) was subcloned into the pFastBac-1 vector and transformed into DH10Bac competent cells, to produce a bacmid containing the appropriate PTEN sequence. Site-directed mutants were generated by standard QuikChange methods using the primers shown in Table 1. The bacmid containing t-PTEN-intein-CBD was then transfected into Sf-21 insect cells to generate the corresponding baculovirus needed for protein expression. High Five insect cells were infected with the baculovirus at a cell density of 1 million/ml and cultured for 48 h at 27°C to express truncated PTEN-intein-CBD. The final culture was then harvested by centrifugation, and the cell pellet was lysed with a Dounce homogenizer by 40 strokes in 40 ml of 50 mM HEPES, pH 7.5, 250 mM NaCl, 1 mM EDTA, 10% glycerol, and one tablet of complete EDTA-free protease inhibitor mixture (Roche Diagnostics). The soluble lysate was then incubated with fibrous cellulose (Sigma) for 30 min at 4°C to remove chitinase. Next, t-PTEN-intein-CBD was immobilized by passing it over chitin resin, followed by rigorous washing with 25 mM HEPES, pH 7.5, 250 mM NaCl, 1 mM EDTA, 0.1% Triton, and 50 mM Enzyme Activity Assay-Lipid phosphatase activity of PTEN toward the aqueous soluble substrate diC6-PIP3 was measured by monitoring phosphate release using the Malachite Green assay kit (R&D Biosystems) at 620 nm. Initial velocities were measured for a 25-l reaction containing 50 mM Tris, pH 8.0, 10 mM BME, 0.05 mg/ml ovalbumin, and varying concentrations of diC6-PIP3 (20 -160 M). The reactions were initiated with 0.5-20 g of the corresponding PTEN, and reaction rates were measured for 5-10 min at 30°C. The reactions were then quenched, and inorganic phosphate was measured using the Malachite Green assay kit. Background corrections were performed with quenched enzyme. All measurements were performed in duplicate on at least two separate occasions, and replicates typically agreed within 20%. The k cat /K m measurements derived from these data were obtained using the Michaelis-Menten equation or a linear plot when no evidence of saturation was apparent by visual inspection.
Western Blotting and Alkaline Phosphatase Sensitivity Assay-Semisynthetic phospho-PTEN forms were subjected to alkaline phosphatase treatment over a period of up to 300 min in 50 mM Tris, pH 8.0, 10 mM BME, 1 M of the corresponding phospho-PTEN, and 1 M calf intestinal alkaline phosphatase (NEB). The reaction samples were quenched in SDS loading dye at 95°C, and the fraction of phospho-PTEN remaining was measured by Western blotting analysis using an antibody to the PTEN phospho-C-tail cluster (Abcam) in combination with a HRP-conjugated anti-rabbit secondary antibody. Western positive bands were detected using Amersham Biosciences ECL Western blotting detection reagent with Syngene PXi and quantified by ImageJ image quantification software. Phosphotail half-lives were calculated using a single exponential decay curve. All measurements were performed at least twice on separate occasions, and half-lives from replicates typically agreed within 30%.
UV-induced C-terminal Tail Cross-linking-An 18-mer tetraphosphorylated C-terminal tail peptide was synthesized with a photoactivatable p-benzoylphenylalanine (Bpa) in the place of Phe 392 and an artificial C-terminal lysine that was N ⑀ -biotinylated using conventional Fmoc peptide synthesis strategies and ligated to immobilized t-PTEN-intein-CBD using the methods described above. UV-induced cross-linking was performed with 50 g of PTEN diluted with 50 mM Tris, pH 8.0, 10 mM DTT to ϳ150 g/ml. In a quartz reaction chamber with circulating water at 4°C and constant stirring, 4p-PTEN containing Bpa was UV-irradiated at 365 nm with a Blak-Ray C50 UV lamp for 3 h. Next, the UV-irradiated 4p-PTEN was treated with bovine alkaline phosphatase (Sigma) at a final concentration of 1 M for 2 h at room temperature. The alkaline phosphatase-treated cross-linked PTEN samples were subjected to buffer exchange with 50 mM NH 4 HCO 3 followed by reduction with 5 mM tris(2-carboxyethyl)phosphine for 30 min at 60°C and alkylation with 10 mM 2-chloroacetamide for 30 min at room temperature in the dark. The samples were desalted using Protein Desalting Spin Columns and overnight digestion was conducted using trypsin (2% w/w) for 16 h at 37°C. The digests were acidified by the addition of trifluoroacetic acid to 0.1% final concentration prior to desalting using C18 solid phase extraction. The eluate was dried in a SpeedVac and reconstituted in avidin column loading buffer for enrichment of the biotinylated peptides. Peptides were eluted from the avidin column with 2 mM biotin buffer (100 mM Na 2 PO 4 , pH 7.0, 150 mM NaCl, 2 mM biotin). The peptide concentration was approximated by UV absorption using a NanoDrop (A 280 nm ), and the samples were reconstituted in a sufficient volume of 0.2% formic acid in water to yield a concentration of 200 fmol/l prior to LC-MS/MS analysis.
LC-MS/MS-Chromatographic separation was performed using a Dionex Ultimate 3000 RSLCnano system (Thermo Scientific) with a 75 m x 15 cm Acclaim PepMap100 separating column (Thermo Scientific) protected by a 2 cm guard column (Thermo Scientific). The mobile phase flow rate was 300 nL/min and consisted of 0.1% formic acid in water (A) and 0.1% formic acid, 95% acetonitrile (B). MS analysis was performed using an LTQ Orbitrap Velos Pro mass spectrometer (Thermo Scientific). The spray voltage was set at 2.2 kV. Orbitrap spectra were collected from 400 -1800 m/z at a resolution of 30,000 followed by data-dependent HCD MS/MS (at a resolution of 7500, collision energy 35%, activation time 0.1 ms) of the 10 most abundant ions using an isolation width of 2.0 Da. Charge state screening was enabled to reject the generation of MS/MS spectra for unassigned and singly charged precursor ions. A dynamic exclusion time of 40 s was used to discriminate against previously selected ions.
Data Analysis-Mass spectrometry data from the crosslinked PTEN samples were analyzed using Crossfinder version 1.0 (20,21) with the default parameters. Prior to analysis with Crossfinder, MS/MS data were converted to mzXML files using ProteoWizard.

Individual Tail Phosphorylation Effects on PTEN
Conformation-To assess the specific effects of the individual Ser/Thr phosphorylations among Ser 380 , Thr 382 , Thr 383 , and Ser 385 , we employed expressed protein ligation to generate individual semisynthetic phospho-forms of PTEN. In this way, amino acids 1-378 were produced as an intein fusion protein using a baculovirus expression system and then reacted with MESNA to generate the PTEN recombinant fragment C-terminal thioester (Fig. 1, A and B). In addition, we synthesized a series of eight new 25-mer N-Cys containing mono-, di-, and tri-phosphorylated C-terminal peptides, along with the unphosphorylated and tetraphosphorylated peptides prepared previously. These peptides were then chemoselectively ligated individually to the recombinant PTEN thioester with high efficiency (Ͼ80%) and then the semisynthetic full-length PTEN proteins purified using anion exchange chromatography to Ͼ90% purity (Fig. 1C). As shown previously (14), the introduction of Cys 379 (in place of Tyr 379 ) and the conditions of ligation are well tolerated by PTEN regarding its catalytic activity.
We then measured the PIP3 phosphatase activity of these semisynthetic phospho-PTENs using soluble diC6-PIP3 substrate and monitoring inorganic phosphate release using malachite green. As shown in Fig. 2 and Table 2, the monophosphorylated forms p380-PTEN, p382-PTEN, p383-PTEN, and p385-PTEN all showed a modest reduction in catalytic efficiency, with each showing a k cat /K m ϳ3-fold below that of unphosphorylated PTEN (WT n-PTEN). The three di-phos-phorylated PTENs 2p-380/382-PTEN, 2p-382/383-PTEN, and 2p-380/385-PTEN each hydrolyze diC6-PIP3 ϳ6-fold below that of WT n-PTEN. The triphosphorylated PTEN form 3p-380/382/383-PTEN showed approximately a 12-fold reduction in catalytic efficiency compared with WT n-PTEN, and its catalytic efficiency was very similar to that of tetraphosphorylated (WT 4p-PTEN). These data suggest that no single phospho-modification of the Ser/Thr C-terminal cluster is dominant and that each is partially additive in antagonizing catalysis.
To assess the effects of the particular phosphorylation events on PTEN conformation, the mono-, di-, and tri-phosphorylated semisynthetic PTENs were treated with alkaline phosphatase to gauge tail accessibility. As reported previously, natively  folded 4p-PTEN versus denatured 4p-PTEN shows resistance to tail phosphate removal catalyzed by the nonspecific hydrolase alkaline phosphatase (14). The resistance of natively folded 4p-PTEN to alkaline phosphatase is understood to be caused by the masking of the phospho-tail through its intramolecular interactions with the PTEN body. We first determined which of the mono-,di-, and tri-phosphorylated semisynthetic PTENs are recognized by a commercially available phospho-PTEN Ab used in Western blotting. We found that all forms of p380containing semisynthetic PTEN gave a strong Western blot signal with the anti-phospho-PTEN Ab, but those PTEN forms lacking p380 were not reliably detected (Fig. 3A). Thus, we proceeded to analyze the kinetics of alkaline phosphatase-catalyzed dephosphorylation of p380-PTEN, 2p-380/382-PTEN, 2p-380/385-PTEN, and 3p-380/382/385-PTEN benchmarked to WT 4p-PTEN standard (Fig. 3, B and C, and Table 2). These measurements showed that p380-PTEN was the most rapidly dephosphorylated by alkaline phosphatase (t1 ⁄ 2 ϭ 3 min), 30-fold faster than WT 4p-PTEN (t1 ⁄ 2 ϭ 104 min), and was followed closely by 2p-380/382-PTEN (t1 ⁄ 2 ϭ 6 min), which was ϳ20-fold faster than WT 4p-PTEN. 2p-380/385-PTEN (t1 ⁄ 2 ϭ 20 min) was dephosphorylated by alkaline phosphatase ϳ5-fold faster than WT 4p-PTEN. The sensitivity of 3p-380/382/385-PTEN (t1 ⁄ 2 ϭ 96 min) to alkaline phosphatase was nearly identical to that of WT 4p-PTEN. Taken together, the pattern of alkaline phosphatase sensitivity correlates fairly well with the levels of semisynthetic PTEN catalytic activity, indicating that the individual phosphorylation events show partial additivity in driving the tail interaction with the PTEN body.
Analysis of Mutant Semisynthetic 4p-PTENs-We next combined site-directed mutagenesis along with expressed protein ligation to prepare three semisynthetic 4p-PTEN proteins to investigate the biochemical properties of the ePTEN mutant residues and those of the PTEN C␣2 loop. As previously mentioned, cellular expression of ePTEN indicates that it localizes to the plasma membrane, suggesting that tail phosphorylation may be unable to drive a closed PTEN conformation (17). The C␣2 loop of PTEN has been implicated in membrane binding and in hydrogen/deuterium exchange experiments to participate in PTEN phospho-tail-mediated conformational change (11,18). The 4p-ePTEN mutant containing Q17R, R41G, E73D, N262Y, and N329H was prepared by semisynthesis as described above, as were two 4p-PTEN C␣2 mutants, C␣2D (K327D, K330D, K332D, and R335D) and C␣2A (K327A, K330A, K332A, and R335A) to Ͼ90% purity (Fig. 4). Interestingly both n-ePTEN and 4p-ePTEN had 2-and 5-fold higher diC6-PIP3 phosphatase catalytic efficiencies compared with their WT counterparts ( Fig. 4 and Table 2). However, the diC6-PIP3 phosphatase catalytic efficiency of 4p-ePTEN was a marked 6-fold lower than that of n-ePTEN ( Fig. 4B and Table 2), suggesting that tail phosphorylation could drive ePTEN into a closed state. Furthermore, 4p-ePTEN alkaline phosphatase sensitivity showed a slightly (ϳ40%) longer half-life of phos-  Fig. 3B (n ϭ 2). Error bars indicate Ϯ standard error.
pho-tail hydrolysis relative to that of WT 4p-PTEN (Fig. 5, A and B, and Table 2). These results suggest that the point mutants of ePTEN do not substantially weaken phospho-tail-PTEN body interactions, and there is an alternative, unidentified explanation for ePTEN membrane localization.
In contrast, analysis of the 4p-PTEN C␣2 mutants demonstrated that this region participates significantly in phosphorylation-mediated conformational change. Both the C␣2D and C␣2A 4p-PTEN mutants showed increased catalytic efficiencies of diC6-PIP3 hydrolysis relative to that of 4p-PTEN ( Fig.  4C and Table 2). Although still lower than the truncated PTEN C␣2 mutants that lack phospho-tails, the lipid phosphatase activities of these 4p-PTEN C␣2 mutants suggest that replacement of these Lys and Arg residues weakened phospho-tail interactions in phosphorylated PTEN. Furthermore, both C␣2D and C␣2A 4p-PTEN exhibited a marked 6-fold greater sensitivity to alkaline phosphatase-mediated tail dephosphorylation compared with WT 4p-PTEN. The fact that both the Asp and Ala replacements of the Lys/Arg residues in the C␣2 4p-PTEN mutants display similar behaviors is consistent with the possibility that one or more of the Lys/Arg side chains are making important electrostatic interactions with residues in the anionic PTEN tail.
Photo-cross-linking of Benzoylphenylalanine Containing 4p-PTEN-To further investigate the structural basis of phospho-tail-body interactions within 4p-PTEN, we pursued a photo-cross-linking strategy. In this regard, Phe 392 was replaced with the photo-activatable Bpa and a C-terminal biotin-modified Lys. In this semisynthetic strategy, a 18-mer N-Cys peptide ( Fig. 6A) was used rather than the full-length 25-mer tail for synthetic ease and because prior studies have shown the 4p-17mer tail is sufficient to promote high affinity interaction with the core of PTEN (14). This Bpa/biotin-modified semisynthetic 4p-PTEN (BB-4p-PTEN) was efficiently produced and appeared greater than 90% pure after anion exchange chromatography (Fig. 4A). Irradiation of BB-4p-PTEN was followed by alkaline phosphatase treatment (to remove the mass spectrometry electrospray ionization suppressive effects of the phosphates) and trypsinization and avidin treatment to enrich for biotin-containing peptides. These peptides were then subjected to LC-MS/MS, and then the data were analyzed using Crossfinder (20,21). This led to the identification of a cross-linked peptide that contained a C-terminal peptide ( 379 CSDTTDSD-PENEP(B)DEDK bio 396 ) attached to an N-terminal peptide ( 42 LEGVYRNNIDDVVR 55 ) (Fig. 6, B and C). The assignment of this cross-linked peptide was based on a combination of its intact molecular mass (4090.7 Da), as well as its fragmentation pattern using tandem mass spectrometry, although the site of conjugation to the N-terminal peptide (between Gly 45 and Ile 50 ) could not be precisely defined (Fig. 6C). This cross-linked peptide was not detected in the absence of UV irradiation. These data thus suggest that the C-tail can access a conformation that places the C-terminal peptide in proximity to the N-terminal catalytic domain segment containing amino acids 42-55.
To further probe the importance of this interaction, we generated semisynthetic 4p-PTEN containing three point mutations in the vicinity of the three-dimensional surface of the cross-linked N-terminal segment, R41D, R47D, and R74D (3R/ D). Even unligated 3R/D PTEN mutant showed significantly impaired catalytic activity (Fig. 7A), as expected from prior studies that have investigated the phosphatase effects of mutating Arg 47 (22,23). Thus, it was not feasible to use diC6-PIP3 hydrolytic activity to gain insight into the phosphorylation effects of 3R/D-4p-PTEN. To assess the conformation of the 3R/D-4p-PTEN, we relied on alkaline phosphatase sensitivity.
In these experiments, we found that 3R/D-4p-PTEN (t1 ⁄ 2 ϭ 63 min) showed approximately a 2-fold rate increase in alkaline phosphatase-mediated dephosphorylation compared with 4p-PTEN (Fig. 7, B and C, and Table 2). These data support at least a modest contribution of the N-terminal catalytic segment in stabilizing the closed conformation of 4p-PTEN.

Discussion
The application of protein semisynthesis, mutagenesis, and photo-cross-linking has been integrated to refine our understanding of how phosphorylation of a cluster of four C-terminal Ser/Thr residues drives a conformational change in PTEN. Prior to this work, it was unclear whether any single phosphorylation site in PTEN would be predominantly or wholly responsible for inducing the PTEN conformational change, and the influence of the three other sites would be negligible. Our data shown here support the idea that each of the four sites can contribute incrementally to stabilizing the closed PTEN conformation, and at least three sites are needed to match the full effects of tetraphosphorylated PTEN. These results imply that the surface of interaction between the phospho-tail and the PTEN body may be rather broad to encompass multiple contact sites with these phosphates. These findings also suggest the concept that a dynamic stepwise degree of PTEN conformational closure may occur by modifying only a subset of the tail Ser/Thr residues, which in turn could give rise to a sliding scale of cellular signaling effects. Our results also highlight that the antibody used here cannot readily distinguish between monophosphorylated p380 and the multiply modified forms that contain p380. This suggests the need for caution in interpreting Western blots and immunocytochemistry data that rely on this reagent for signaling analysis.
Several models have been proposed regarding where the phospho-tail in PTEN contacts the core of the protein (14 -16). In the current study, we found additional evidence for key interactions between the PTEN C2 domain and the phospho-tail. Prior studies have shown that the CBR3 loop of the PTEN C2 domain is involved in the phospho-tail interactions (14), with an 8-fold enhanced sensitivity to alkaline phosphatase-catalyzed tail dephosphorylation compared with WT 4p-PTEN. However, although the penta-Lys (260, 263, 266, 267, and 269) to penta-Asp CBR3 loop 4p-PTEN mutant appears to have a relatively open structure, mutation of the penta-Lys motif to penta-Ala is quite similar to WT 4p-PTEN (14). Thus, it appears that none of the CBR3 loop Lys residues are making key electrostatic interactions with the phospho-tail, and only when replaced with the negatively charged Asp residues do they prevent conformational closure, arguing for a more indirect role in stabilizing the closed conformation. In contrast, as shown here the C␣2 Lys/Arg side chains cannot be substituted functionally by Ala, suggesting that these basic residues are more directly involved in tail phospho-interactions.
Some prior experiments including hydrogen/deuterium exchange, mutagenesis, and protease sensitivity have suggested that the catalytic domain of PTEN might also somehow participate in phospho-tail interactions (14,15,18). Through the use of photo-cross-linking, we find that at least part of the phospho-tail binds to an N-terminal segment in the PTEN catalytic domain. Mutagenesis of this segment is consistent with this region playing a partial role in mediating the conformational closure of phosphorylated PTEN. If we consider the energetic sum of the alkaline phosphatase sensitivity parameters of the N-terminal segment (2-fold), the C␣2 loop (6-fold), and the CBR3 loop (8-fold) mutants reported here and previously, we can more than account for the 25-fold protection measured when comparing folded and unfolded 4p-PTEN. Thus, it is plausible that these three PTEN surfaces are fully responsible for locking the phospho-tail to the PTEN body (14). It is perhaps noteworthy that this N-terminal segment as well as the CBR3 loop and the C␣2 loop all have been implicated in PTEN membrane recruitment (1,18,22,23). It is therefore reasonable to postulate that the PTEN phosphate tail acts as a molecular mimic of the phospholipid membrane. These results under- score the very large interaction surface that may govern the 4p-PTEN conformational closure.
The observation that 4p-ePTEN behaves conformationally similarly to WT 4p-PTEN was unexpected, given the dramatic cellular phenotype associated with ePTEN (17). We speculate that these ePTEN point mutations may strengthen a specific protein-protein or protein lipid interaction between PTEN and a molecule not yet identified. Future work will be needed to address these possibilities that could uncover new targets to enhance cellular PTEN functions.
This study also illustrates the power of expressed protein ligation in elucidating how a complex series of PTMs can alter the structure of a protein (24). Because mass spectrometry has identified hundreds of thousands of PTMs in proteins, it has become daunting to cope with illuminating the structural and functional effects of these PTMs at an individual protein level. The incorporation of four phosphorylations, a benzoylpheny-lalanine, and a biotin tag site-specifically into a single folded protein of this size would have been difficult to achieve without expressed protein ligation. Expressed protein ligation when coupled with other biochemical, biophysical, and analytical approaches thus offers atomic precision in working through the 21st century challenges of protein science.
Author Contributions-Z. C., D. R. D., and P. A. C. designed the experiments, and Z. C., D. R. D., D. H., and D. M. B. performed the biochemical experiments in this manuscript. S. N. T. designed and performed the protein mass spectrometry experiments and analyzed the data along with D. R. D. and P. A. C. All authors contributed to the analysis of the data. The manuscript was drafted by Z. C., D. R. D., and P. A. C., and all authors contributed to writing and editing the manuscript.