SEARCH

SEARCH BY CITATION

Keywords:

  • AICDε51;
  • Alzheimer's disease;
  • presenilin/γ-secretase;
  • βAPP;
  • γ-cleavage;
  • ε-cleavage

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

Background:  During intramembrane proteolysis of β-amyloid protein precursor (βAPP) by presenilin (PS)/γ-secretase, ε-cleavages at the membrane-cytoplasmic border precede γ-cleavages at the middle of the transmembrane domain. Generation ratios of Aβ42, a critical molecule for Alzheimer's disease (AD) pathogenesis, and the major Aβ40 species might be associated with ε48 and ε49 cleavages, respectively. Medicines to downregulate Aβ42 production have been investigated by many pharmaceutical companies. Therefore, the ε-cleavages, rather than the γ-cleavage, might be more effective upstream targets for decreasing the relative generation of Aβ42. Thus, one might evaluate compounds by analyzing the generation ratio of the βAPP intracellular domain (AICD) species (ε-cleavage-derived), instead of that of Aβ42.

Methods:  Cell-free γ-secretase assays were carried out to observe de novo AICD production. Immunoprecipitation/MALDI-TOF MS analysis was carried out to detect the N-termini of AICD species. Aβ and AICD species were measured by ELISA and immunoblotting techniques.

Results:  Effects on the ε-cleavage by AD-associated pathological mutations around the ε-cleavage sites (i.e., βAPP V642I, L648P and K649N) were analyzed. The V642I and L648P mutations caused an increase in the relative ratio of ε48 cleavage, as expected from previous reports. Cells expressing the K649N mutant, however, underwent a major ε-cleavage at the ε51 site. These results suggest that ε51, as well as ε48 cleavage, is associated with Aβ42 production. Only AICDε51, though, and not Aβ42 production, dramatically changed with modifications to the cell-free assay conditions. Interestingly, the increase in the relative ratio of the ε51 cleavage by the K649N mutation was not cancelled by these changes.

Conclusion:  Our current data show that the generation ratio of AICDε51 and Aβ42 do not always change in parallel. Thus, to identify compounds that decrease the relative ratio of Aβ42 generation, measurement of the relative level of Aβ42-related AICD species (i.e., AICDε48 and AICDε51) might not be useful. Further studies to reveal how the ε-cleavage precision is decided are necessary before it will be possible to develop drugs targeting ε-cleavage as a means for decreasing Aβ42 production.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

The transmembrane domain of β-amyloid protein precursor (βAPP) is proteolysed by presenilin (PS)/γ-secretase.1 Analysis of the resultant products has shown that the proteolysis proceeds by at least two distinct cleavages. The ‘ε-cleavage’ liberates its intracellular domain (i.e., βAPP intracellular domain (AICD)) into the cytoplasm, whereas the ‘γ-cleavage’ releases Alzheimer's disease (AD)-associated amyloid β-protein (Aβ).2–6

There are some variations in both the γ- and ε-cleavages of βAPP.6–8 The major N-termini of AICD species consist of leucine-49, valine-50 and leucine-52 (Aβ-numbering), whereas the major C-termini of Aβ species are comprised of valine-40 and alanine-42. (Fig. 1a).6 Among these, highly aggregatable Aβ42 is the major component of senile plaques in AD brains.9

image

Figure 1. Effect of familial Alzheimer's disease (AD)-associated β-amyloid protein precursor (βAPP) mutations around the ε-cleavage site. (a) Schematic diagram of intramembrane cleavage sites of βAPP and the familial AD mutations used in the present study. The amino acid sequence around the juxta membrane region of human βAPP is described (amyloid β-protein (Aβ) numbering). Filled inverted triangles indicate the cleavage sites. Substituted amino acids of the familial AD mutations are shown in open rectangles. The site of each mutant is also shown using APP695 numbering. (b) Mass spectra of de novoβAPP intracellular domain (AICD) species in the cell-free assay. Crude membrane fractions obtained from wild-type (wt) βAPP and the indicated βAPP mutant cells were used. (c) Relative secreted Aβ42 to Aβ40 ratio in the conditioned media of wt βAPP and the indicated βAPP mutant cells. The asterisks indicate statistical significance (*P < 0.05, **P < 0.001, one-way analysis of variance (anova) and Tukey–Kramer method). Error bars indicate standard error of the mean (SEM). (d) Hypothesis for explaining increased γ42 cleavage in each mutant βAPP (upper panels) and differential production of Aβ40 and Aβ42 (lower panels). FAD, familial Alzheimer's disease; m.w., molecular weight.

Download figure to PowerPoint

Are there any relationships between the ε- and γ-cleavages? How do these cleavages occur? Ihara et al. have tried to address these questions and recently showed that ε-cleavage precedes γ-cleavage in in vitroγ-secretase assays.10βAPP-CTF stubs, βAPP membrane-tethered remnants after β-cleavage, first undergo ε-cleavage.10 The ε-cleavage liberates AICD from the membrane and produces a membrane-bound 48/49 amino-acid-long Aβ species that undergoes further C-terminal truncation by PS/γ-secretase.11 Stepwise cleavages remove every three amino-acid residues from the C-terminus of the long Aβ species, which finally secretes Aβ40/42.12–14 For example, mutant PS causes increases in both ε48 and γ42 cleavages.8 Thus, the γ-cleavage seems to occur in an ε-cleavage-dependent manner.10 Furthermore, these results show that the production process for pathological Aβ42 is distinct from that of Aβ40.15 That is, the major ε49 cleavage causes the production of Aβ40, whereas a minor ε48 cleavage causes the production of pathological Aβ42.14

Modulation of γ-secretase function to specifically inhibit Aβ42 production is one of the promising strategies for developing drugs to modify the disease course of AD16. Given the possible correlation between the ε- and γ-cleavages, we think that targeting the upstream ε-cleavages will be a novel and more efficient method for developing Aβ42-lowering drugs. To test if precision of the ε-cleavage can be used as a novel target for drug development, we investigated the ε-cleavage pathway, particularly ε51 cleavage, which has previously not been well-characterized.7

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

Cell culture and cDNA constructs

cDNAs of βAPP V642I, L648P and K649N mutants were generated by PCR-based mutagenesis using a Quickchange mutagensis kit (Stratagene, La Jolla, CA, USA) or KOD plus (Toyobo, Osaka, Japan) with wt βAPP695 cDNA as a template. K293 cells were transfected and cultured as previously described.17

Membrane preparation

The crude membrane fraction was prepared as previously described with a slight modification.7,18 In the present study, the homogenization buffer contained 0.25 mol/L sucrose and 50 mmol/L HEPES (pH 7.4) containing a protease inhibitor cocktail (Roche Basel, Switzerland). To prepare the alkaline pretreated membrane, the membrane fraction was suspended in a 50 mmol/L bicarbonate buffer (pH 11.0) and incubated at 4°C for 1 h. The suspension was then centrifuged at 100 000 ×g for 1 h followed by washing once with a 50 mmol/L Mes buffer (pH 6.0).

Cell-free γ-secretase assay

The cell-free γ-secretase assay was carried out as previously described with a modification.7,18 The reaction buffer in the present study contained a 150 mmol/L citrate buffer (pH 6.0), 50 mmol/L MES (pH 6.0), 167 mmol/L NaCl and a protease inhibitor mixture comprised a 5x complete protease inhibitor cocktail (Roche), 0.5 mmol/L DIFP (WAKO, Osaka, Japan), 1 µg/ml TLCK (Sigma-Aldrich, St. Louis, MO, USA), 10 µg/ml antipain (Peptide Institute, Osaka, Japan), 10 µg/ml leupeptin (Peptide Institute), 5 mmol/L 1,5 phenanthroline (Sigma-Aldrich), 10 µmol/L amastatin (Peptide Institute), 10 µmol/L bestatin (WAKO), 1 µmol thiorphan (Sigma-Aldrich), 10 µmol/L phosphoramidon (Peptide Institute) and 1 µmol/L pepstatin A (Peptide Institute). To prepare the pH 7.4 buffer, 50 mmol/L HEPES (pH 7.4) was used instead of the citrate and MES buffers.

Immunoprecipitation/MALDI MS analysis

Immunoprecipitation/MALDI MS (IP-MS) analysis followed by cell-free incubation was carried out as previously described.7,18,19 The heights of the MS peaks and molecular weights were calibrated using angiotensin and bovine insulin β-chain as standards (Sigma-Aldrich).

ELISA analysis for Aβ

Aβ40 and Aβ42 levels in conditioned media were quantified by ELISA (WAKO).

Immunoblotting of Aβ

SDS-solubilized proteins were separated by SDS-PAGE using an 8 mol/L urea gel17,18,19 and transferred to a nitrocellulose membrane. Immunoblotting of Aβ species using 82E1 (IBL) was carried out as previously described.20

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

The βAPP K649N Belgian mutant increased both the relative ratio of AICDε51 and Aβ42 production in a cell-free γ-secretase assay

To test if the ε51 cleavage precedes the γ42 cleavage, we analyzed the effects of three βAPP mutants (V642I,21 L648P,22 and K649N23) around the ε-site. The L648P and K649N mutants (βAPP695 numbering) are located downstream of the ε51 site, and the V642I mutant is located upstream of the ε48 site (Fig. 1a). Each of the three mutants is familial AD-associated and therefore increases the relative ratio of Aβ42 production. We raised K293 cells stably expressing each of the mutants, prepared the crude membrane fractions24 and carried out the cell-free γ-secretase assays.7,18

As shown in Figure 1b, the K649N βAPP mutant caused a marked increase in the relative ratio of AICDε51 production (see also Table 1). However, the other two mutants caused completely different effects on the cleavage. The L648P mutant produced a barely detectable level of AICDε51, whereas in the V642I mutant cells, the ratio of AICDε51 production was comparable to that of wild-type (wt) expressing cells. It is of note that, instead of increased AICDε51 production, these V642I and L648P mutants substitutively increased the relative ratio of AICDε48 production. Next we measured Aβ species secretion by the stable cells in conditioned media using ELISA (Fig. 1c). As expected, we observed a significant increase in the ratio of Aβ42 to total Aβ secretion in the conditioned medium of the mutant cells. This data shows that the K649N mutant increased the ratio of Aβ42 production through upregulation of the ε51 cleavage, whereas the V642I and L648P mutants increased Aβ42 production through the ε48 cleavage. Based on these results, we suggest that not only the ε48, but also the ε51 cleavage precedes Aβ42 production, possibly by sequential three amino-acid C-terminal truncation14 (Fig. 1d).

Table 1.  Molecular species of β-amyloid protein precursor intracellular domain generated in the cell-free assay
AICD speciesm/z
Calculated (M + H)Observed (M + H)
MeanSD
  1. M + H, protonated molecular mass.

AICDε51 (52–99)Wild-type5677.795678.380.64
V642I5677.795678.300.70
K649N5663.745663.960.23
AICDε49 (50–99)Wild-type5907.95908.350.29
V642I5907.95908.490.21
L648P5891.875892.480.20
K649N5893.845894.100.27
AICDε48 (49–99)Wild-type6020.986021.360.40
V642I6020.986021.590.42
L648P6004.966005.590.33
K649N6006.936007.510.17
AICDε52 (53–99)K649N5550.655551.010.27

Incubation in higher pH does not cancel the K649N βAPP mutant effects

We previously found that the precision of ε-cleavage changes depending on the buffer pH.7,18 The relative ratio of AICDε51 production is the most sensitive to such changes. Therefore, we next determined whether the relative ratio of AICDε51 and/or Aβ42 production by the K649N mutant is affected by changing the buffer pH during the cell-free assay. As expected, incubation in the higher pH (pH 7.4 vs pH 6.0) buffer decreased the relative ratio of AICDε51 generation in both the K649N mutant and wt βAPP membrane fraction. However, the pH effect was not so strong as to cancel the AICDε51 upregulation effect by the K649N mutant (Fig. 2a). We further analyzed the pH effects on the increase in the relative ratio of Aβ42 production by the mutant (Fig. 2b). Surprisingly, the assay pH elevation did not cause any changes in the relative ratio of Aβ42 generation. Therefore, unlike the effects of the K649N mutant on the ε51- and γ42-cleavages, the elevation of the buffer pH causes a decrease in the relative ratio of AICDε51 production, but does not cause any changes in Aβ42 production. The data suggest that two distinct mechanisms might contribute to the determination of the relative ratio of AICDε51 production.

image

Figure 2. Effect of cell-free incubation pH levels on the precision of ε/γ-cleavages. (a) Mass spectra of β-amyloid protein precursor intracellular domain (AICD) generated in the cell-free assay carried out at the indicated pH (upper and middle panels). Peak heights of AICDε49 and ε51 were measured and the ratios of AICDε49 to ε51 were calculated (lower panel). The asterisks indicate statistical significance (*P < 0.05, **P < 0.001, one-way anova and Tukey–Kramer method). Error bars indicate SEM. (b) Levels of amyloid β-protein (Aβ) generated at the indicated pH. Levels of Aβ40 and 42 were measured by western blotting and the Aβ42 to 40 ratios calculated. The asterisks indicate statistical significance. Error bars show SEM. m.w., molecular weight.

Download figure to PowerPoint

Alkali pretreatment of the crude membrane fraction cancels the effect of higher pH cell-free incubation on ε-cleavage

Because the ε51 cleavage occurs at the membrane-cytosol interface, we considered that membrane-bound substances might induce the pH-dependent effects on AICDε51 production. Many substances detach from the membrane on treatment with alkali solution.25 To test this theory, we washed the wt βAPP membrane fraction in a pH 11 solution (see ‘Materials and Methods’), then we carried out the cell-free assay at pH 6.0. The relative ratio of AICDε51 production markedly decreased (Fig. 3a), whereas that of the Aβ42 did not (Fig. 3b). The phenomena are reminiscent of the effects of raising the pH of the incubation buffer (see Fig. 2). Thus, we further considered that the decrease in the AICDε51 production resulting from the use of a higher incubation buffer pH might also be a result of detachment of substances from the membrane. When the membrane fraction was incubated in a pH 7.4 buffer after alkali treatment, we could no longer observe the pH-dependent incubation buffer effects on the AICDε51 ratio (Fig. 3c). Collectively, although incubation at lower pH buffer increased in the AICDε51 ratio (Fig. 2a), the effects were cancelled by the alkali pretreatment (Fig. 3a). These results suggest that substances removed by the alkali treatment might induce the changes in the relative ratio of AICDε51 production.

image

Figure 3. Effect of alkali pretreatment on the precision of ε/γ-cleavages of wild-type (wt) β-amyloid protein precursor. (a) Mass spectra of β-amyloid protein precursor intracellular domain (AICD) generated in the cell-free assay with and without alkali pretreatment. Peak heights of AICDε49 and ε51 were measured and the AICDε49 to ε51 ratios calculated. The asterisk indicates statistical significance (*P < 0.05, paired t-test). Error bars indicate SEM. (b) Levels of β-amyloid (Aβ) generated in the cell-free assay after alkali pretreatment. Levels of Aβ40 and 42 were measured by western blotting with and without alkali pretreatment and the Aβ42 to 40 ratios calculated. (c) Mass spectra of AICD generated in the cell-free assay at the indicated pH after alkali pretreatment. MW, molecular weight; WT, wild type.

Download figure to PowerPoint

Alkali pretreatment of the crude membrane fraction did not cancel the effects of the K649N mutant on the ε-cleavage

As shown in Figure 1, the K649N βAPP mutation causes upregulation of both the AICDε51 and Aβ42 ratio, whereas alkali pretreatment causes downregulation of only the AICDε51 ratio (Fig. 3). These data show that changes in the AICDε51 ratio caused by the mutation and by the treatment occur by two distinct processes. A further experiment was carried out to confirm whether the K649N mutation cause a change in the relative ratio of AICDε51 production through the effect of the alkali treatment (Fig. 4a). After treatment of the K649N mutant membrane fraction in the alkali solution, the cell-free assay was carried out at pH 6.0. As shown in Figure 4a, even after the alkali treatment, the K649N mutant membrane produced a relatively higher level of AICDε51 than that of the wt fraction (Fig. 3a). Furthermore, the elevated Aβ42 ratio was not changed by the pretreatment (Fig. 4b).

image

Figure 4. Effect of alkali pretreatment on the precision of ε/γ-cleavages of βAPP K649N Belgian mutant. (a) Mass spectra of β-amyloid protein precursor intracellular domain (AICD) generated in the cell-free assay with and without alkali pretreatment. (b) Levels of of β-amyloid generated in the cell-free assay following alkali pretreatment. N.S, not significant.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

In the present study, we determined that there are at least two factors that change the precision of ε-cleavage: (i) a process induced by a pathological βAPP mutation; and (ii) another process induced by possibly unidentified substances removed from the membrane fraction by alkali pretreatment. In the case of βAPP mutations, the relative ratio of ε51 and ε48 production increases in parallel with the ratio of AD-associated Aβ42.

It has been reported that ε-cleavage precedes γ-cleavage and γ-cleavage seems to occur in an ε-cleavage-dependent manner.10 Considering these reports and our own preliminary results, it seemed possible that measurement of the relative ratio of AICDε48/AICDε51 production might help develop Aβ42-lowering anti-AD drugs. Further study showed, however, that the relative level of AICDε51 production is drastically affected by the removal of unidentified substances from the membrane as a result of alkali pretreatment. Interestingly, the alkali pretreatment did not cause any changes in the relative ratio of Aβ42 generation. These results show that changes in the precision of ε-cleavage do not always cause parallel alterations in the precision of γ-cleavage, even though ε-cleavage occurs upstream of the γ-cleavage. Therefore, although measuring the levels of AICD species is a potentially attractive new target for developing Aβ42 lowering compounds, challenges must still be overcome before screening methods for such compounds can be established. For example, the paradoxical mechanism discussed previously must first be understood before an assay in which the ε-cleavage precision accurately reflects the γ-cleavage precision can be developed.

How does alkali pretreatment result in a decreased ratio of AICDε51 production? One might consider the presence of unknown substances that (i) transiently associate with the PS/γ-secretase and affect its intramembrane cleavage precision, or (ii) truncate a couple of N-terminal amino-acid residues of AICD produced by the ε-cleavage. The second possibility is reminiscent of angiotensin-converting enzyme activity to truncate the C-terminus of Aβ42.26 Of course, the possibility that alkali pretreatment might change the character of PS/γ-secretase itself also cannot be excluded.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

Our current data suggest that the precision of ε-cleavage does not always change in parallel with the precision of γ-cleavage, even though ε-cleavage occurs upstream of the γ-cleavage. Thus, to measure the levels of AICD species might be an attractive new target for developing Aβ42 lowering compounds, there still remain some challenges.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES

M.O and coworkers are funded by the National Institute of Biomedical Innovation (05–26), the Ministry of Education, Culture, Sports, Science and Technology, and the Ministry of Health, Labor and Welfare, Japan. The authors declare no competing financial interests.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. ACKNOWLEDGMENTS
  9. REFERENCES