Insights from Modeling the Tertiary Structure of Human BACE2 Kuo-Chen Chou†,‡ Gordon Life Science Institute, San Diego, California 92130 and Tianjin Institute of Bioinformatics and Drug Discovery (TIBDD), Tianjin, China Received June 3, 2004
BACE1, or β-secretase, is a putative prime therapeutic target for the treatment of Alzheimer’s disease. Mapping to the Down syndrome critical region (chromosome 21) and identified as a homologue of BACE1, BACE2 also cleaves amyloid precursor protein at the β-site. Thus, BACE2, named also as Asp1 or Memapsin1, represents a second β-secretase candidate. In this paper, the tertiary structure of the protease domain of BACE2 was developed. Although the overall structural topology between BACE1 and BACE2 protease domains is quite similar, the former contains 3 disulfide bonds but the latter only two. Particularly, a subtle structural difference around the DTG/DSG active site between the two structures has been observed that is useful for the in-depth selectivity study of BACE1 and BACE2 inhibitors, stimulating new therapeutic strategies for the treatment of Alzheimer’s disease and Down syndrome as well. Keywords: Alzheimer’s disease • Down syndrome • protease domain • DTG/DSG active site • hydrogen bond • disulfide bond
I. Introduction Alzheimer’s disease is the most common brain disorder among older people. The disease usually begins after age 60, and risk goes up with age. Once suffering from Alzheimer’s disease, the victims will eventually lose their entire memory. On the other hand, Down syndrome is one of the most common genetic birth defects, affecting approximately one in 800 to 1000 babies. The victim of Down syndrome usually has a combination of birth defects, such as some degree of mental retardation, characteristic facial features, heart defects, increased infections, trouble hearing, and problems with vision. The first chemical evidence of a relationship between Alzheimer’s disease and Down syndrome was discovered by Glenner and Wong,1 who found the presence of a unique cerebrovascular amyloid fibril protein (amyloid plaque) in both Down syndrome and Alzheimer’s disease. It is known that Down syndrome is generally caused by an extra chromosome. Normally, each egg and sperm cell contains 23 chromosomes. The union of the two creates 23 pair, or 46 chromosomes in total. Sometimes, an accident occurs when an egg or sperm cell is forming, causing it to have an extra chromosome number 21. All of the features and birth defects associated with Down syndrome result from having this extra chromosome 21 in each of the body’s cells. Therefore, Down syndrome is also called trisomy 21. But what is it that causes Alzheimer’s disease? Precious little is known about Alzheimer’s disease, which threatens to strike some 14 million Americans by 2050. Its precise cause is still * To whom correspondence should be addressed.
[email protected]. † Gordon Life Science Institute. ‡ Tianjin Institute of Bioinformatics and Drug Discovery. 10.1021/pr049905s CCC: $27.50
2004 American Chemical Society
E-mail:
largely mysterious, and effective treatments are still years away. However, it is known through biochemical studies that one of the major components of amyloid plaques is the amyloid β (Aβ) peptide, varying from 39 to 43 residues in length but primarily of 40 or 42. Also, studies in molecular pathology have indicated that a remarkable evidence observed in the brain (cerebral cortex) of Alzheimer’s disease patients is the accumulation of Aβ peptides, leading to Aβ fibril deposit or β-sheet (oligomer) amyloid deposit (see, e.g., Carter & Chou2-5). Accordingly, Aβ peptides have been considered one of the causal factors for the pathogenesis of Alzheimer’s disease, and the metabolism of Aβ peptides has attracted considerable interest. The source of Aβ peptide is a large membrane bound amyloid precursor protein (APP). Release of an Aβ peptide from APP requires sequential cleavage by two endopeptidases, βand γ-secretases, at the N-terminus and C-terminus of the Aβ peptide, respectively. As such, β-secretase is a prime therapeutic target for the treatment of Alzheimer’s disease. Until recently, the identity of both β- and γ-secretases had eluded researchers for over a decade. Recent reports suggest that γ-secretase is either a novel aspatic proteinase called presenilin6,7 or a large multimeric complex that includes presenilin. The highly elusive β-secretase was recently identified as a transmembrane aspartic proteinase.8 As a family member of BACE (β-site APP Cleaving Enzyme), β-secretase is often named BACE1 in the literature although it is sometimes called Asp2 or memapsin2 as well. Like many aspartyl proteases, BACE1 has a propeptide domain that is removed to form the mature enzyme. It is believed that the prodomain is removed by furin or furin-like proteases because BACE1 propeptide cleavage occurs at the sequence RLPRVE,9 a potential furin recognition motif. Although the prodomain is required for Journal of Proteome Research 2004, 3, 1069-1072
1069
Published on Web 08/10/2004
research articles proper folding of BACE1, the presence of the prodomain does not substantially suppress its protease activity,10,11 indicating BACE1 zymogen does not seem to confer strict zymogen-like properties. Such a unique phenomenon was recently elucidated based on the computed 3D (dimensional) structure of BACE1 zymogen.12 Meanwhile, it has been found by mapping to the Down syndrome critical region13 that BACE2 resides in the obligate Down syndrome regions of chromosome 21, and that BACE2, identified as a homologue of BACE1, also cleaves APP at the β-site and is expressed in brain and several tissues and cell lines. Thus, BACE2, also named Asp1 or Memapsin1, represents a second β-secretase candidate. Although both BACE1 and BACE2 are generated as proenzymes, the prodomain-removing process of BACE1 and BACE2 is very different: BACE1 requires the activity of a separate enzyme (furin or a furin-like enzyme) to cleave its prodomain at the site between Arg-45 and Glu46,9 while BASE2 is capable of autocatalytic at the cleavage site between Leu-62 and Ala-63 to remove its prodomain.14 As mentioned above, β-amyloid (Aβ) deposition in the brain is the hallmark of Alzheimer’s disease. To initiate Aβ formation, β-secretase cleaves APP at the N-terminus of Aβ to release APPsb (∼100 kDa soluble NT-fragment), and C99, a 12-kDa CT membrane fragment. Both C99 and C83 can be further cleaved by γ-secretase releasing Aβ and a p3 peptide, respectively. Alternatively, the secretase cleaves within the Aβ will prevent the formation of Aβ (see, e.g., Figure 1 of a previous paper).12 The 3D structure of the protease domain of BACE1 has been determined,15 and its molecular cleaving mechanism discussed.12 For further understanding the amyloidogenic cascade so as to discover an effective approach to inhibit the β-site APPcleaving process, it is also important to reveal the function and mechanism of BACE2 and its difference with BACE1 at a deeper level. To realize this, it is essential to find the 3D structure of BACE2 as well. This is the goal of the current study.
II. Materials and Methods According to SWISS-PROT databank,16 human BACE1 (locus BACE_HUMAN, accession P56817) contains 501 residues: residues 1-21 is the signal peptide, residues 22-45 the propeptide, residues 46-501 the mature chain, of which residues 22-457 the extracellular cellular region, residues 93/289 the active site aspartases, residues 458-478 the transmembrane region, and residues 479-501 the cytoplasmic domain. Human BACE2 (locus BAE2_HUMAN, accession Q9Y5Z0) contains 518 residues: residues 1-20 is the signal peptide, residues 21-62 the propeptide, residues 63-518 the mature chain, of which residues 21-473 the extracellular cellular region, residues 110/ 303 the active site aspartases, residues 474-494 the transmembrane region, and residues 495-518 the cytoplasmic domain. It should be pointed out that the end position of the propeptide and the beginning position of the mature chain for BACE2 (Table 1) are not from the original data in SWISS-PROT databank16 but derived from a recent report by Hussain et al.14 Since the current study is focused on the protease domain which is directly involved with the cleavage mechanism, the sequence of BACE2 that we need to consider is within the region of residues 63-473. This is because:1 residues 1-20 belong to the signal peptide (Table 1) that will be cleaved off by signal peptidase during the secretory process (see, e.g., refs 17, 18), and hence, there is no need to include;2 residues 2162 belong to the propeptide that will be removed to form the mature chain by the autocatalytic mechanism14 as mentioned 1070
Journal of Proteome Research • Vol. 3, No. 5, 2004
Chou
Figure 1. Sequence alignment between BACE2 (bac2) and BACE1 (1fkn), where the residues identical to each other between the two sequences are colored red, the DTG active sites for BACE1 and BACE2 are residues 93-95 and residues 110-112, respectively, and the corresponding DSG active sites are residues 289291 and residues 303-305. The codes connected by a solid line indicate the Cys pair involved in forming disulfide bond. The alignment was conducted by the PILEUP program in the GCG package.19 See the relevant text for further explanation. Table 1. Comparison of Residue Regions between Human BACE1 and BACE2 region
BACE1
BACE2
entire signal peptide propeptide extra-cellular mature chain active site trans-membrane cytoplasmic
1-501 1-21 22-45 22-457 46-501 93/289 458-478 479-501
1-518 1-20 20-62a 21-473 63a-518 110/303 474-494 495-518
a Unknown in the original data deposited in the SWISS-PROT databank16 and was determined according to the observations of ref 14.
above; and3 residues 474-518 are within the transmembrane and cytoplasmic regions (Table 1), and hence are outside the scope of the current consideration. The sequence alignment performed by using the PILEUP program in the GCG package19 for BACE2 (bac2) and the protease domain of BACE1 (1fkn) is given in Figure 1, where the residues identical to each other between the two enzymes are colored red, the DTG active sites for BACE1 and BACE2 are residues 93-95 and residues 110-112, respectively, and the corresponding DSG active sites are residues 289-291 and residues 303-305. The codes connected by a solid line indicate the Cys pair involved in forming the disulfide bonds. As shown
3D Structure of BACE2
from the figure, the similarity and identity between BACE2 and BACE1 are 63.35% and 54.71, respectively. On the basis of the sequence alignment of Figure 1 and the template, i.e., the X-ray coordinates of 1fkn,15 the atomic coordinates for BACE2 was derived by the segment matching approach.20-24 The operation consisted of the following procedures: (1) The target chain was first broken into short segments of sequence. (2) The database (formed by more than 5200 high-resolution X-ray protein structures) was searched for matching segments according to the sequence alignment and the shape of the template protein chain. (3) These segments coordinates were fitted into the growing target structure under the monitor to avoid any van der Waal overlap until all atomic coordinates of the target structure were obtained. (4) The process was repeated 10 times and an average model was generated, followed by energy minimization to create the final 3D structure. The segment matching approach was previously used to model the structure of the protease domain of caspase8, at a time before the X-ray coordinates were released for caspase-3.25 In that particular study, the atomic coordinates of the catalytic domain of caspase-3 were predicted based on the X-ray structure of caspase-1, and then the caspase-3 structure thus obtained served as a template to model the protease domain of caspase-8. After the X-ray coordinates of caspase-3 protease domain were finally released and the X-ray structure of the caspase-8 protease domain was determined,26 it turned out that the RMSD (root-mean-square-deviation) for all the backbone atoms of the caspase-3 protease domain between the X-ray and predicted structures was 2.7 Å, whereas the corresponding RMSD was 3.1 Å for caspase-8, and only 1.2 Å for its core structure. This indicates that the computed structures of caspase-3 and -8 were quite close to the corresponding X-ray structures. Later on, this method was successively applied to model the CARDs (caspase recruitment domains) of Apaf-1, Ced-4, and Ced-3, based on the NMR structure of the RAIDD CARD,27 and to model the Cdk5-Nck5a* complex28 as well as the protease domain of caspase-9.29 It is intriguing to mention that, two years after the computed Cdk5Nck5a* complex structure was published,28 the corresponding crystal structure was determined.30 It has been found that the predicted Cdk5 and the crystal Cdk5 are almost the same. Furthermore, according to the report from the crystal structure, upon the binding of Cdk5 and Nck5a* (or p25), the buried surface area is 3400 Å2,30 which is quite close to 3461 Å2 derived from the computed structure.28 Also, based on the computed Cdk5-Nck5a*-ATP structure, the molecular truncation experiments were conducted and it has been found that the experimental results “confirm and extend specific aspects of the original predicted computer model”.31 Recently, it was also used to model the 3D structures of extracellular domains for the subtypes 1, 2, 3, and 5 of GABA-A receptors,32 clarifying the ambiguity about the directionality of the subunit arrangement in the heteropentamers and providing useful insights for understanding the molecular operation mechanism of the receptors.
III. Results and Discussion The overall structure thus obtained for the BACE2 is given in Figure 2, where the DTG/DSG active site is shown in red, and the residues involved in forming disulfide bonds in yellow. As expected, the overall topology is similar to that of BACE1. To provide an overall illustration of the general similarity and subtle differences between the 3D structures of BACE1 and
research articles
Figure 2. Computed structure for the BACE2 protease domain, where the DTG/DSG active site is shown in red, and the residues involved in forming disulfide bonds in yellow.
Figure 3. Superimposition of BACE1 and BACE2 protease domains, where BACE1 is colored in yellow with the DTG/DSG active site in blue and BACE2 green with the DTG/DSG active site in red.
BACE2, a superimposition of the two is given in Figure 3, where BACE1 is colored in yellow with DTG/DSG active site in blue and BACE2 green with DTG/DSG active site in red. As shown in Figure 3, the two structures are very similar to each other except the remarkable difference that BACE1 has three disulfide bonds but BACE2 has only two. Actually, the RMSD (rootmean-square-deviation) values between the two structures for all the CR atoms, all the backbone atoms, and all the heavy atoms are 0.66 Å, 0.68 Å, and 0.96 Å, respectively. Nevertheless, Journal of Proteome Research • Vol. 3, No. 5, 2004 1071
research articles
Figure 4. Hydrogen bond patterns around the DTG/DSG active site for (a) BACE1, and (b) BACE2. The hydrogen bond is shown by blue dotted line, and the DTG/DSG active site colored by red.
some subtle structural difference has been observed around the DTG/DSG active site between BACE1 and BACE2. As shown in Figure 4a, BACE1 has six hydrogen bonds (blue dotted line) formed between DTG/DSG active site (red) and Leu-180, Leu-182, Thr-292, Val-288, and Met-394 (yellow), while the number of corresponding hydrogen bonds for BACE2 has only five, and residues involved are Ser-113, Leu-199, Val-302, Thr306, and Ile-407 (Figure 4b). This kind of subtle difference may provide useful structural information for the selectivity study of BACE1 and BACE2 inhibitors, encouraging new therapeutic approaches for the treatment of Alzheimer’s disease and Down syndrome. Abbreviations: 3D, three-dimensional; Aβ, amyloid β; APP, amyloid precursor protein; BACE, β-site APP cleaving enzyme The tertiary structure of the protease domain of BACE2 is developed based on the structure of BACE1. Although the overall structural topology between the BACE1 and BACE2 protease domains is quite similar, some subtle structural difference around the DTG/DSG active site between the two structures has been observed that is useful for the in-depth selectivity study of BACE1 and BACE2 inhibitors.
References (1) Glenner, G. G.; Wong, C. W. Biochem. Biophys. Res. Commun. 1984, 122, 1131-1135.
1072
Journal of Proteome Research • Vol. 3, No. 5, 2004
Chou (2) Klunk, W. E.; Pettegrew, J. W.; Abraham, D. J. J. Histochem. Cytochem. 1989, 37, 1273-1281. (3) Klunk, W. E.; Xu, C. J.; Pettegrew, J. W. J. Neurochem. 1994, 62, 349-354. (4) Soto, C.; Kindy, M. S.; Baumann, M.; Frangione, B. Biochem. Biophys. Res. Commun. 1996, 226, 672-680. (5) Carter, D. B.; Chou, K. C. Neurobiol. Aging 1998, 19, 37-40. (6) Wolfe, M. S.; Xia, W.; Ostaszewski, B. L.; Diehl, T. S.; Kimberly, W. T.; Selkoe, D. J. Nature 1999, 398, 513-517. (7) Lin, X.; Koelsch, G.; Wu, S.; Downs, D.; Dashti, A.; Tang, J. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 1456-1460. (8) Vassar, R.; Bennett, B. D.; Babu-Khan, S.; Kahn, S.; Mendiaz, E. A.; Denis, P.; Teplow, D. B.; Ross, S.; Amarante, P.; Loeloff, R.; Luo, Y.; Fisher, S.; Fuller, J.; Edenson, S.; Lile, J.; Jarosinski, M. A.; Biere, A. L.; Curran, E.; Burgess, T.; Louis, J.-C.; Collins, F.; Treanor, J.; Rogers, G.; Citron, M. Science 1999, 286, 735-741. (9) Bennett, B. D.; Denis, P.; Haniu, M.; Teplow, D. B.; Kahn, S.; Louis, J.-C.; Citron, M.; Vassar, R. J. Biol. Chem. 2000, 275, 37 71237 717. (10) Shi, X. P.; Chen, E.; Yin, K. C.; Na, S.; Garsky, V. M.; Lai, M. T.; Li, Y. M.; Platchek, M.; Register, R. B.; Sardana, M. K.; Tam, M. J.; Thiebeau, J.; Wood, T.; Shafer, J. A.; Gardell, S. J. J. Biol. Chem. 2001, 276, 10 366-10 373. (11) Benjannet, S.; Elagoz, A.; Wickham, L.; Mamarbachi, M.; Munzer, J. S.; Basak, A.; Lazure, C.; Cromlish, J. A.; Sisodia, S.; Checler, F.; Chre´tien, M.; Seidah, N. G. J. Biol. Chem. 2001, 276, 10 87910 887. (12) Chou, K. C.; Howe, W. J. Biochem. Biophys. Res. Commun. 2002, 292, 702-708. (13) Acquati, F.; Accarino, M.; Nucci, C.; Fumagalli, P.; Jovine, L.; Ottolenghi, S.; Taramelli, R. FEBS Lett. 2000, 468, 59-64. (14) Hussain, I.; Christie, G.; Schneider, K.; Moore, S.; Dingwall, C. J. Biol. Chem. 2001, 276, 23 322-23 328. (15) Hong, L.; Koelsch, G.; Lin, X.; Wu, S.; Terzyan, S.; Ghosh, A. K.; Zhang, X. C.; Tang, J. Science 2000, 290, 150-153. (16) Bairoch, A.; Apweiler, R. Nucleic Acids Res. 2000, 25, 31-36. (17) Chou, K. C. Protein Eng. 2001, 14, 75-79. (18) Chou, K. C. Curr. Protein Peptide Sci. 2002, 3, 615-622. (19) Devereux, J. In Genetic Computer Group (GCG); Madison, Wisconsin, 1994. (20) Jones, T. A.; Thirup, S. EMBO J. 1986, 5, 819-822. (21) Blundell, T. L.; Sibanda, B. L.; Sternberg, M. J. E.; Thornton, J. M. Nature (London) 1987, 326, 347-352. (22) Finzel, B. C.; Kimatian, S.; Ohlendorf, D. H.; Wendoloski, J. J.; Levitt, M.; Salemme, F. R. In Crystallographic and Modeling Methods in Molecular Design; Bugg, C. E., Ealick, S. E., Eds.; Springer-Verlag: Berlin, 1989; pp 175-188. (23) Levitt, M. J. Mol. Biol. 1992, 226, 507-533. (24) Chou, K. C.; Nemethy, G.; Pottle, M.; Scheraga, H. A. J. Mol. Biol. 1989, 205, 241-249. (25) Chou, K. C.; Jones, D.; Heinrikson, R. L. FEBS Lett. 1997, 419, 49-54. (26) Watt, W.; Koeplinger, K. A.; Mildner, A. M.; Heinrikson, R. L.; Tomasselli, A. G.; Watenpaugh, K. D. Structure 1999, 7, 11351143. (27) Chou, J. J.; Matsuo, H.; Duan, H.; Wagner, G. Cell 1998, 94, 171180. (28) Chou, K. C.; Watenpaugh, K. D.; Heinrikson, R. L. Biochem. Biophys. Res. Commun. 1999, 259, 420-428. (29) Chou, K. C.; Tomasselli, A. G.; Heinrikson, R. L. FEBS Lett. 2000, 470, 249-256. (30) Tarricone, C.; Dhavan, R.; Peng, J.; Areces, L. B.; Tsai, L. H.; Musacchio, A. Mol. Cell 2001, 8, 657-669. (31) Zhang, J.; Luan, C. H.; Chou, K. C.; Johnson, G. V. W. PROTEINS: Structure, Function; Genetics 2002, 48, 447-453. (32) Chou, K. C. Biochem. Biophys. Res. Commun. 2004, 316, 636-642.
PR049905S
