Abstract
The human parvulin Pin1 is a member of the peptidyl–prolyl cis–trans isomerase group of proteins, which modulate the assembly, folding, activity, and transport of essential cellular proteins. Pin1 is a mitotic regulator interacting with a range of proteins that are phosphorylated before cell division. In addition, an involvement of Pin1 in the tau-related neurodegenerative brain disorders has recently been shown. In this context, Pin1 becomes depleted from the nucleus in Alzheimer's disease (AD) neurons when it is redirected to the large amounts of hyperphosphorylated tau associated with the neurofibrillary tangles. This depletion from the nucleus may ultimately contribute to neuron cell death. Recently we have devised a novel methodology in which exogenous Pin1 is used as a TEM probe for its target proteins. Here we extend this methodology to provide further evidence that Pin1 binds at enhanced levels to mitotic nuclear proteins and to hyperphosphorylated tau in AD brain. We suggest that exogenous Pin1 labeling can be used to elucidate the phosphorylation status of its target proteins in general and could specifically provide important insights into the development of tau-related neurodegenerative brain disorders.
Keywords
P
Pin1 is a nuclear protein that can regulate entry of cells into mitosis and whose PPIase activity is required for normal progression through mitosis in yeast and mammalian cells (Lu and Hunter 1995; Lu et al. 1996; Fujimori et al. 1999). Deletion of Pin1 activity in HeLa cells causes mitotic arrest and apoptosis (Lu et al. 1996). In addition to binding to the targets of the mitotic phosphoprotein monoclonal-2 antibody (MPM-2), Pin1 specifically interacts with and regulates the activity of a subset of mitotic and nuclear proteins in a phosphorylation-dependent manner (Yaffe et al. 1997). These targets include the NIMA kinase, Cdc25, Wee1, topoisomerase IIa, Cdc27, RNA polymerase II, Plk1, Myt1, Rab4, and p70/p85S6K (Lu et al. 1996, 1999b; Yaffe et al. 1997; Crenshaw et al. 1998; Shen et al. 1998; Albert et al. 1999; Morris et al. 1999; Patra et al. 1999; Wells et al. 1999). In addition to having a role in pre-mRNA 3′ end formation (Morris et al. 1999), it appears that nucleus-localized Pin1 modulates the structure of target proteins by isomerizing phosphorylated Ser/Thr-Pro bonds and is thus intimately involved in regulating many aspects of cell cycle control.
Alzheimer's disease (AD), the most common cause of dementia in the elderly, is characterized by the presence of two histopathological hallmark brain lesions, extracellular deposits of β-amyloid in neuritic plaques and intracellular neurofibrillary tangles (NFTs). The latter is composed of bundles of paired helical filaments (PHFs), the major protein subunit of which is the microtubule-associated protein (MAP) tau (Iqbal et al. 1989; Lee et al. 1991). Tau in PHF is in a different form from that in normal neurons, being abnormally hyperphosphorylated and aggregated into filaments. This hyperphosphorylated tau is unable to bind to microtubules and therefore cannot promote or maintain microtubule assembly (Iqbal et al. 1998). Evidence from many studies indicates that hyperphos-phorylation of tau is responsible for its loss of biological activity, its resistance to proteolytic degradation, and probably plays a key role in neurofibrillary degeneration in AD patients (reviewed in Iqbal et al. 1998). It is still unclear how hyperphosphorylation of tau occurs in AD brain, but this probably reflects a combination of regulation at the level of both increased kinase and decreased phosphatase activity (Gong et al. 2000).
Recently, evidence for an involvement of Pin1 in AD has been shown (Lu et al. 1999a). Elevated levels of Pin1 binding to the NFT-rich cytoplasm of AD-affected compared with healthy neurons were reported (at the light microscopic level). These authors showed that Pin1 binds to only one phosphorylated threonine-proline motif in tau and that the protein co-purified with the PHFs. They suggested that the resultant depletion of available soluble Pin1 in the AD-affected neurons could contribute to cell death.
We have recently (Rulten et al. 1999) identified and characterized a novel human eukaryotic parvulin homologue [hEPVH, also known as hPar14 (Uchida et al. 1999)], so named because of its homology with bacterial parvulin (Rahfeld et al. 1994). Routine immunogold labeling transmission electron microscopy (TEM) revealed a preferentially mitochondrial location in HEK 293 cells. TEM analysis also demonstrated that both hEPVH and Pin1 protein can bind to tissue sections (Thorpe et al. 1999). However, the subcellular patterns of their binding were different, with Pin1 being preferentially nuclear and hEPVH associated with the mitochondria. Because Pin1 has nuclear targets, we suggested that it was binding to these proteins in the sections. On the evidence of its binding in a different pattern, we further suggested that hEPVH might be binding to its own predominantly mitochondrial target proteins. The possible use of PPIase proteins in general as probes to investigate specific target or substrate proteins was proposed.
In this latter work, sections were incubated with free unconjugated Pin1 or hEPVH protein, followed by standard immunogold labeling. In the work here, we have also prepared specific probes by the direct conjugation of Pin1 and hEPVH to gold particles and investigated their use in a simple one-step labeling procedure.
Using mitotic HEK 293 cells and AD brain tissue, we have analyzed binding of Pin1 to nuclear proteins in the former and to NFTs in the latter. These data show that Pin1 binds strongly to the Ser/Thr-Pro motif of its target proteins in the nucleus of mitotic cells but that Pin1 is redirected from the nucleus to NFTs in AD brain tissue.
Materials and Methods
HEK 293 Cell Cultures
HEK 293 cells were grown in DMEM supplemented with Glutamax (Life Technologies; Paisley, UK), 10% fetal calf serum (Labtech International; Lewes, UK), 100U/ml penicillin, and 0.1mg/ml streptomycin (Life Technologies).
A “mitotically enriched” HEK 293 cell population was achieved by growth to 50% confluency, then washing twice in DMEM before the addition of nocodazole (Sigma; Poole, UK) at 75, 150, or 375 ng/ml (in dimethyl sulfoxide; final concentration of 0.05% in the DMEM) for 22 hr. For cycling cultures, the nocodazole treatment was substituted with either DMEM alone or DMEM + 0.05% DMSO. Each culture was split; half was processed for TEM and the other half for FACS (fluorescence-activated cell sorter) analysis.
Determination of HEK 293 Cellular DNA Content by FACS Analysis
HEK 293 cells were stained for their DNA content with pro-pidium iodide (PI) and analyzed in a Beckman Coulter Epics Elite ESP FACS (Beckman Coulter; High Wycombe, UK), as follows. After overnight fixation at 4C in an absolute ethanol/PBS (70%/30%) mixture, the cells were rinsed in PBS and the volume then reduced to 50 μl. RNase A (Sigma; 20 μl of 1 mg/ml) and PI (Sigma; 500 μl of 30 μg/ml) were added and the cells were left to stain for 30 min before analysis. For FACS analysis, an argon laser was used to excite fluorescence, which was then collected through a 675-nm bandpass filter. Ten thousand cells were analyzed and the data were gated on the ‘Time of Flight’ vs PI fluorescence scattergram to select the singlet cell population (i.e., to discriminate against doublets and triplets).
Brain Tissue
Postmortem samples from the frontal lobe of an AD and a normal human brain were obtained with the kind assistance of Dr. Nigel Cairns from the Brain Bank at the Institute of Psychiatry, (King's College, University of London). The AD tissue was from a 64-year-old man at 10 hr postmortem and the normal tissue from a 57-year-old man at 45 hr postmortem.
Recombinant Pin1 and hEPVH Protein Production
The recombinant proteins were produced as described previously (Thorpe et al. 1999).
Antibodies and Gold Probes
Goat anti-Pin1 polyclonal antisera to the C-terminal [anti-(C)Pin1] and N-terminal [anti-(N)Pin1] domains were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit anti-tau polyclonal antiserum (anti-tau) was obtained from Sigma. Ten-nm rabbit anti-goat IgG-gold (RaG10) and goat anti-rabbit IgG-gold (GaR10) were obtained from British BioCell International (Cardiff, UK).
Preparation of HEK 293 Cells and Brain Tissue for TEM
HEK 293 cells were prepared as previously described (Thorpe et al. 1999). The frozen brain tissue was allowed to warm gradually to −20C. Then the gray matter was selected and sliced directly into a cold (4C) fixative consisting of 4% formaldehyde and 0.1% (vacuum-distilled) glutaraldehyde in PBS. All the following procedures were carried out at 4C. After fixation for 18 hr, the samples were rinsed thoroughly in PBS, then dehydrated in an ethanol series and embedded in Unicryl resin (British BioCell), as previously described (Thorpe et al. 1999).
Preparation of Protein-Gold Probes
Initially, a titration series was carried out for Pin1 and hEPVH to find the amount of protein required to stabilize the colloidal gold (10 nm; British BioCell), with the pH of the latter adjusted to be above the pI of the protein being bound [9.5 for Pin1 (pI = 8.8) and 10.5 for hEPVH (pI = 9.8)]. The protein-gold probes were then prepared by mixing the colloidal gold with the required amount of protein for 5 min (with continuous agitation) and then adding 1% Carbowax 20M (0.45-μm filtered) to give a final concentration of 0.1%. After centrifugation at 30,000 × g for 25 min to remove aggregates of gold particles, the supernatant was decanted to fresh tubes and re-centrifuged at 30,000 × g for 90 min. The supernatant was carefully removed and discarded and the remaining probe volume adjusted with 0.1% Carbowax 20M to give an optical density at 530 nm (OD530) of 16.0. The protein–gold probes were stored at either 4C or −20C (after flash-freezing in liquid nitrogen). The 10-nm protein A-gold (pA-gold) probe (OD530 3.2; British BioCell) was centrifuged at 30,000 × g for 90 min and the supernatant removed to achieve an OD530 of 16.0.
Protein-Gold Labeling of HEK 293 Cells
Thin sections were cut and collected on formvar-coated grids. A modified PBS, pH 8.2, containing 1% BSA, 500 μl/liter Tween-20, 10 mM NaEDTA, and 0.2 g/liter NaN3 (henceforward termed PBS +) was used throughout all the following procedures for blocking, dilution of gold probes, and rinsing.
The sections were blocked in PBS+ for 30 min at room temperature (RT) and then incubated in 60-μl drops of either Pin1–or hEPVH-gold probe for 18 hr at 4C. The incubation pH was 8.2 (in PBS+) and the probes were routinely diluted 1:3 (OD530 of 5.33), except for the initial labelings of the confluent cells for which the probe OD530 was 3.74 and 2.26 for Pin1 and hEPVH, respectively. As a control “irrelevant” protein, sections were incubated in pA-gold at an OD530 of 5.33. In one experiment for Pin1–gold labeling, the incubation pH was varied from 7 to 10, and in another (again for Pin1), a titration of probe dilution was carried out from 1:3 to 1:30. Sections were rinsed in PBS+ (four times for 10 min) and then in distilled water (four times for 3 min).
Unconjugated Pin1 Labeling of Brain Tissue
Sections of normal and AD brain were blocked in normal rabbit serum (1:10 in PBS+) for 30 min at RT and then incubated in 40-μl drops of either PBS+ (to reveal endogenous Pin1 levels) or 20 μg/ml Pin1 for 24 hr at 4C. After three 2-min PBS+ rinses, the sections were immunogold-labelled for Pin1 as described below. Further serial sections were immunolabeled for tau (see below).
Immunogold Labeling TEM of Brain Tissue
For localization of endogenous Pin1 and tau, sections were blocked either in normal goat serum (for tau) or in normal rabbit serum (for Pin1; both 1:10 in PBS+) for 30 min and then incubated overnight at 4C in either specific antiserum [anti-tau 1:100 dilution (in PBS+; 0.48 mg/ml final protein concentration)] or a mixture of anti-(C)Pin1 and anti-(N)Pin1; 20 μg/ml final protein concentration or in the relevant non-immune serum at an identical protein concentration. After three 2-min PBS+ rinses, the sections were immunolabeled for 1 hr at RT in GaR10 (for tau) or in RaG10 (for Pin; both 1:10 in PBS+). The immunolabeled sections were then rinsed in PBS+ (three times for 10 min) and then in distilled water (four times for 3 min).
Specificity Tests of the Mixture of anti-(C)Pin1 and anti-(N)Pin1 Antisera
Solid-phase Homologous Preabsorption. A homologous pre-absorption of the mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera was carried out by solid-phase coupling of recombinant Pin1 to NHS-activated Sepharose. HiTrap NHS-activated affinity columns were obtained from Amersham Pharmacia Biotech (Little Chalfont, UK). Recombinant Pin1 was concentrated to 3.5 mg/ml using a Centricon-3 column (Millipore; Watford, UK) and diluted in coupling buffer (0.2 M NaHCO3, 0.5 M NaCl, pH 8.3) to give a final concentration of 0.5 mg/ml. One ml of this solution was injected into the activated column, which was then incubated at RT for 1 hr. Excess active groups were deactivated by serially washing the column with 0.5 M ethanolamine, 0.5 M NaCl, pH 8.3, and 0.1 M acetate, 0.5 M NaCl, pH 4. A blank Sepharose column was made by loading coupling buffer only onto another column, incubating for 1 hr, and deactivating the excess groups in the same way as for the Pin1 column. Both columns were equilibrated in PBS+ before loading 1 ml of an 80 μg/ml mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera and incubating at 4C for 2.5 hr. The preabsorbed antiserum was eluted from the column in 4 ml of PBS+ to obtain a 20 μg/ml antiserum, which was used directly for immunolabeling. Western Blotting. Western blotting was carried out to determine the specificity of the mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera. HeLa cell nuclear extracts were run on a 15% SDS-PAGE gel and transferred to PVDF membrane. One lane was probed with the antiserum eluted from the blank column from the solid-phase homologous preabsorption described above. Another lane was probed with the antiserum eluted from the Pin1-conjugated column.
Image Acquisition and Analysis
Thin sections were post-stained in filtered 2% aqueous uranyl acetate for 20 min and then examined in a Hitachi-7100 TEM at 100 kV. Random images at ×50,000 magnification were acquired digitally with a charge-coupled device camera (800 × 1200 pixel; Digital Pixel Co., Brighton, UK). The areas of the different cell compartments were computed (Kinetic Imaging; Liverpool, UK), gold particles counted, and the labeling results expressed as numbers of gold particles per μm2.
Results
Protein-Gold Labeling of Confluent HEK 293 Cells
Figure 1 shows the pooled data from two separate experiments using the directly conjugated protein-gold probes to label confluent HEK 293 cells. Pin1 binds in a pattern that is preferentially to the nucleus, whereas hEPVH binds more intensely to the mitochondria. Control labeling using a commercial pA-gold probe at identical OD exhibited low levels of binding to all cell compartments.
Subcellular distribution of Pin1-gold and hEPVH-gold probe labeling in confluent HEK 293 cells. Labeling densities for sections incubated in either protein A–(pA), hEPVH–(hEPVH), or Pin1–gold (Pin1) probes. Pooled data from three separate labelings. Nu, nucleus; Cyt, cytoplasm; Mito, mitochondria; Cell, Nu, Cyt, and Mito combined. Data, gold particles/μm2; n = 30 ± SEM (except for pA, where n = 60).
Effect of Incubation pH on Pin1-Gold Probe Binding
The effect of the pH of the incubation buffer (PBS+) on the efficiency of Pin1–gold binding was determined by labeling actively growing HEK 293 cells at pH 7, 8.2, 9, and 10. Labeling density at pH 8.2 was higher than that at pH 7, and there was a sharp drop in label density at the higher nonphysiological (with regard to the bulk of the cytosol) pHs of 9 and 10. This latter observation lends some weight to the apparently specific in vivo nature of the observed protein binding. These results therefore confirmed that the pH of 8.2 used throughout this work is at or near optimal for binding of Pin1 to target proteins (data not shown).
Titration of the Pin1-Gold Probe
Titration of the Pin1-gold probe (from 1:3 to 1:30) showed that there were still significant levels of labeling in the nuclei of actively growing HEK 293 cells at a dilution of 1:30 (data not shown).
Effects of Nocodazole Treatment on the Cell Cycle of HEK 293 Cells
Efforts to arrest HEK 293 cells in mitosis with the drug nocodazole were only partially successful. As revealed by FACS analysis (Figure 2; upper panel and tabulated data), at a concentration of 75 ng/ml nocodazole the proportion of cells in the G2/M-phase of the cell cycle was doubled compared with a control cell population. A control incubation of the cells in the drug vehicle (0.05% DMSO) had no effect on the cellular DNA content (data not shown). Increasing the nocodazole concentration two- and five-fold produced no further increase in this G2/M cell population but did increase the level of the apoptotic cell population (data not shown).
Protein-Gold Labeling of Nocodazole-treated HEK 293 Cells
The Pin1-gold probe was used to label cycling and G2/M-enriched (75 ng/ml nocodazole-treated) HEK 293 cell populations. Data from random images showed a 69% increase in labeling density over the nuclei (or chromosomal profiles in the case of mitotic cells, where the nuclear membrane has broken down) of the nocodazole-treated cells compared with the untreated (cycling) cells (Figure 2, lower panel). Significantly, data acquired from visually selected mitotic cells [Figure 2, lower panel: Chr (striped bar)] in the nocodazole-treated culture exhibited labeling densities over regions of chromosomal profiles that were 143% higher than those of the (randomly selected) cycling cell nuclei (Figure 2, lower panel; see also Figure 3). These findings are consistent with the reported role of Pin1 in binding to nuclear targets during mitosis (Yaffe et al. 1997).
Cycling and mitotically enriched HEK 293 cell cultures: FACS analysis and Pin1-gold labeling. (
Ultrastructure and Tau Immunolabeling of Brain Tissue
Thin sections of both normal and AD brain tissue revealed only moderate preservation of cellular ultra-structure, reflecting the forced constraint of using postmortem frozen tissue. However, neurons could be identified, with their nuclei being well-preserved and having electron-dense content and identifiable nucleoli. In the areas between neuron bodies, profiles of neuritic processes were abundant. These contained filaments with a range of appearances, some being rather loosely-packed, whereas others had an appearance of more tightly-packed regular arrays of filaments. In the normal brain sample, no such filamentous regions were observed. The loosely-packed filamentous processes were very specifically immunoreactive for tau (Figure 4). Immunolabeling with non-immune rabbit serum at an identical protein concentration showed low levels of nonspecific binding spread uniformly over the tissue (data not shown).
In the AD brain tissue, extensive filamentous regions were occasionally observed close to the nucleus in the neuron body. These filaments had an identical appearance to the loosely-packed filamentous processes and were also specifically immunolabeled for tau protein (Figure 5). Our measurements of the widths of these filaments (7–18 nm) fall within the range reported for the paired helical filaments (PHFs) characteristic of the hyperphosphorylated form of the tau protein (5–25 nm) (Wischik et al. 1985; Greenberg and Davies 1990; Ruben et al. 1992, 1993, 1995).
Endogenous Pin1 Immunolabeling and Exogenous Pin1 Labeling of Brain Tissue
Initially, we tested the specificity of the mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera by two approaches. A homologous (recombinant Pin1-bound Sepharose) column preabsorption of the mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera resulted in a 77% reduction in immunolabeling density over the nuclei of the AD tissue (compared with a blank column; n = 20). Incubation with non-immune goat serum at an identical protein concentration revealed only very low levels of nonspecific binding (data not shown). We next used the eluents from these two columns to probe nuclear extracts by Western blotting (Figure 6). The dominant band detected with eluent from the blank column corresponded to Pin1; the detection was virtually abolished with eluent from the Pin1 column.
Having confirmed the specificity of the antisera, we examined the distribution of endogenous Pin1 protein in cell compartments of normal and AD brain samples. Figures 7 and 8 (white bars) show data (pooled from three separate labeling experiments) for endogenous Pin1 immunolabeling of normal and AD brain, respectively. These data indicate that endogenous Pin1 is preferentially nuclear in normal brain tissue. However, in AD brain tissue, Pin1 immunolabeling is more uniformly spread throughout the neurons. Although the immunolabeling densities appear to suggest that there is more Pin1 in normal brain tissue, the absolute levels of immunolabeling in the normal and AD brain tissues cannot be compared directly because they are derived from two distinct postmortem samples.
Concurrent incubations of serial sections (of those immunolabeled for endogenous Pin1) in exogenous Pin1 (Figures 7 and 8, black bars) revealed a preferentially cytoplasmic binding pattern for Pin1 to its target proteins in the AD brain, with the filamentous, tau-immunoreactive regions of the cytoplasm exhibiting the highest levels of binding. It was also evident that there is a much higher level of Pin1 binding to the AD tissue sections in all cell compartments (relative to the endogenous protein levels and compared with the normal brain control), and especially to the tau regions [compare the “Exo” data (striped bars) for normal (Figure 7) and AD (Figure 8) brain].
Enhanced Pin1-go d binding to chromosomal profiles of mitotic HEK 293 cells. Digital TEM images of Pin1-gold labeled HEK 293 cells showing (
Figure 9 shows an example of an AD neuron containing an NFT used to derive the data shown in Figure 8. Figures 9A–9C are from a section incubated in PBS+ alone (to reveal endogenous Pin1 immunolabeling). Figures 9E and 9F are of the same neuron from a serial section incubated in 20 μg/ml Pin1. Figure 9A shows the neuron, with its associated NFTs, at low magnification. Note the increased levels of labeling in the nucleus (compare Figure 9E with Figure 9B) and especially over the tau-immunoreactive (NFT) regions (compare Figure 9F with Figure 9C) after incubation in exogenous Pin1. Tau positivity of the NFT region of this (and the other) neuron(s) was confirmed by examination of a serial section immunolabeled with its specific antibody (not shown; but see similar tau-immunoreactive neurons in Figure 5).
Immunogo d abe ing of tau protein in AD brain tissue. Digital TEM images showing examples of filamentous processes in AD brain tissue immunogold-labeled for tau protein (using anti-tau antibody and GaR10 gold probe).
The characteristic PHF structure of the NFT region was usually hard to discern, presumably because of the nature of the sample and the necessarily minimal fixation (reduced aldehyde concentrations and no osmication) required to maintain the protein–protein binding interactions. However, in certain regions of the filaments, some PHF-like structures could be seen (e.g., Figure 9D, arrows).
Discussion
The subcellular distribution and specificity of binding of the Pin1– and hEPVH–gold probes were initially tested by labeling the confluent HEK 293 cells used in our previous studies (Thorpe et al. 1999). We confirmed the preferentially nuclear and mitochondrial binding patterns of Pin1 and hEPVH, respectively, with the low levels of nonspecific protein A–gold binding to sections, reinforcing the specificity of the Pin1–gold binding to target proteins.
Neurof ibri ary tang es and associated tau immunoreactivity. Digital TEM images showing examples of neurons in AD brain tissue containing NFT (∗) and tau protein [sections immunogold-labeled for tau protein (using anti-tau antibody and GaR10 gold probe)].
Subsequently, we prepared cycling and mitotically enriched cultures of these cells to further test our Pin1–gold probe. If the Pin1 target protein binding of this probe (in sections) is really similar to that of in vivo binding events, then the presence of enhanced levels of its many phosphorylated targets in mitotic cells would be expected to increase the observed labeling density. Initial collection of data from randomly acquired images of the mitotically enriched culture confirmed a significant increase in Pin1–gold labeling density in the nucleus (c. 70% above that of the standard cycling cell culture). Furthermore, when mitotic cells were visually selected, heavy labeling was associated with chromosomal profiles. Quantification of the labeling density in these regions revealed an increase of around 140% above the mean labeling density of randomly selected cycling cell nuclei. We propose that this is very firm evidence that the Pin1–gold probe is recognizing its specific target proteins, presumably via a phosphorylated Ser/Thr-Pro motif (e.g. Yaffe et al. 1997).
Western blot showing the specificity of the mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera. Western blot of nuclear extracts run on a 15% SDS-PAGE gel and transferred to PVDF membrane. Lanes A, B, probed with anti-(C)Pin1 and anti-(N)Pin1 antiserum eluted from a blank and a Pin1-conjugated Sepharose column, repectively (as described in Materials and Methods) Size markers are shown at left.
Normal brain tissue: endogenous Pin1 immunolabeling and exogenous Pin1 labeling. Labeling densities in the neuronal nuclei and cytoplasm of normal brain incubated in either PBS + alone [to reveal endogenous Pin1 levels (End, white bars)] or in 20 μg/ml Pin1 protein [to reveal endogenous Pin1 levels plus exogenous Pin1 binding (End/Exo; black bars)]. Striped bars represent levels of exogenous Pin1 binding (Exo). Latter data are derived by subtracting endogenous (PBS+ incubation) Pin1 immunolabeling density from the endogenous plus exogenous Pin1 immunolabel density (20 μg/ml Pin1 incubation). All sections subsequently immunolabeled for Pin1 with a mixture of the anti-(C)Pin1 and anti-(N)Pin1 antisera followed by RaG10. Pooled results from three separate labelings. Data = gold particles/μm2; n = 30 ± SEM.
AD brain tissue: endogenous Pin1 immunolabeling and exogenous Pin1 labeling. Labeling densities in the neuronal nuclei, cytoplasm, and tau-immunoreactive filamentous regions of AD brain incubated in either PBS+ alone [to reveal endogenous Pin1 levels (End; white bars)] or in 20 μg/ml Pin1 protein [to reveal endogenous Pin1 levels plus exogenous Pin1 binding (End/Exo; black bars)]. Striped bars represent levels of exogenous Pin1 binding (Exo). Latter data were derived by subtracting endogenous (PBS + incubation) Pin1 immunolabeling density from the endogenous plus exogenous Pin1 immunolabeling density (20 μg/ml Pin1 incubation). All sections were subsequently immunolabeled for Pin1 with a mixture of anti-(C)Pin1 and anti-(N)Pin1 antisera, followed by RaG10. Pooled results from three separate labelings. Data = gold particles/μm2; n = 30 ± SEM.
We have also used unconjugated Pin1 to investigate the localization of its specific target proteins in human AD and normal brain tissue. The involvement of Pin1 in this tau-related neurodegenerative brain disorder has been established recently (Lu et al. 1999a) and, indeed, it was proposed that there might be a possible therapeutic use for Pin1 in this disease. Using light microscopy, these authors have shown that Pin1 binds tightly to the NFTs. They also confirmed that Pin1 binds to phosphorylated tau. Therefore, it is possible that sequestration of Pin1 to these tangles could result in a concomitant reduction in its availability to fulfill its normal functions in the cell. These authors suggested that their observed high levels of Pin1 binding to these features were indicative of this shortfall of available, soluble Pin1 in the affected neurons. Such a shortfall was confirmed by their biochemical results, and this lack of available Pin1, they proposed, could ultimately lead to cell death.
To elucidate possible shortfalls of Pin1 in the different subcellular compartments of neurons in AD brain tissue at the EM level, we used unconjugated Pin1 protein as a label followed by standard immunolabeling. A concurrent incubation in buffer alone was carried out to reveal endogenous Pin1 protein levels. Because both of these incubations (buffer and Pin1) were concurrently immunolabeled in an identical fashion, the data directly reflect the relative amounts of the endogenous and bound exogenously applied protein, with the levels of the latter reflecting the amount of unbound Pin1 targets. Our data agree with the findings of Lu et al. (1999a) in that, relative to the endogenous protein levels and compared with the normal brain control, there were elevated levels of Pin1 binding to the cytoplasm of the AD neurons. Because the levels of Pin1 binding were highest in the filamentous tau-immunoreactive regions, this provides more firm evidence that the Pin1 is binding specifically to phosphorylated Ser/Thr-Pro motifs. The data thus confirm a shortfall of available Pin1 within the neurones of AD-affected brain. We suggest that using Pin1 in this way as a probe on an animal model system (e.g., Masliah et al. 1996; Ishihara et al. 1999), in which tissue can be readily obtained and optimally preserved, could provide further important insights into the development of this and related neurodegenerative diseases.
AD brain tissue: TEM images of endogenous Pin1 immunolabeling and exogenous Pin1 labeling. Examples of the immunolabeling densities of the same neuron cell body observed after incubation of serial AD brain sections in PBS+ [to reveal endogenous Pin1 (
In summary, we have presented firm evidence that Pin1 binds strongly to its phosphorylated target proteins in mitotic nuclei and to hyperphosphorylated tau in the neurofibrillary tangles of AD-affected neurons. We have also prepared a novel Pin1–gold probe that is extremely simple to use in a one-step labeling procedure for TEM (a procedure that obviates the need for antibody). The methods outlined could be of use in the key research areas of the cell cycle and the tau-related neurodegenerative disorders.
Footnotes
Acknowledgments
Stuart Rulten is in receipt of a BBSRC CASE Award, sponsored by Pfizer Central Research (Sandwich, UK). Simon Morley is a Senior Research fellow of the Wellcome Trust.
We would like to thank Colin Robinson, Richard Bazin, and Vikash Malde (Pfizer Central Research, UK), for provision of the Pin1 expression construct and for assistance with hEPVH and Pin1 expression and purification. JRT would like to thank Richard Killick for his help in arranging tissue provision and an introduction to Nigel Cairns; the latter for his provision of the brain samples and many useful discussions about the AD portion of the work; Dan Pullen for provision of some very useful data. Finally and respectfully, we thank the anonymous tissue donors and their next of kin.