On a possible neutral charge state for the catalytic dyad in β-secretase when bound to hydroxyethylene transition state analogue inhibitors

Article abstract of DOI:10.1021/jm101568y

β-Secretase is one of the aspartic proteases involved in the formation of amyloid plaques in Alzheimer's disease patients. Our previous results using a combination of surface plasmon resonance experiments with molecular modeling calculations suggested that the Asp dyad in β-secretase bound to hydroxylethylene containing inhibitors adopts a neutral charged state. In this work, we show that the Asp dyad diprotonated state reproduced the binding ranking of a set of these inhibitors better than alternative protonation states.

Full text of DOI:10.1021/jm101568y

BRIEF ARTICLE
pubs.acs.org/jmc
On a Possible Neutral Charge State for the Catalytic Dyad in
β-Secretase When Bound to Hydroxyethylene Transition
State Analogue Inhibitors†
Fredy Sussman,*,‡ Josꢀe M. Otero,§ M. Carmen Villaverde,‡ Marian Castro,|| Josꢀe L. Domínguez,‡
Lucía Gonzꢀalez-Louro,‡,§ Ramꢀon J. Estꢀevez,§ and J. Carlos Estꢀevez*,§
‡Molecular Modeling Research Group, Departamento de Química Orgꢀanica, Universidad de Santiago de Compostela,
15782 Santiago de Compostela, Spain
§Carbohydrate Research Group, Departamento de Química Orgꢀanica, Universidad de Santiago de Compostela,
15782 Santiago de Compostela, Spain
Departamento de Farmacología, Instituto de Farmacia Industrial, Universidad de Santiago de Compostela,
15782 Santiago de Compostela, Spain
S Supporting Information
b
ABSTRACT: β-Secretase is one of the aspartic proteases involved in the formation of amyloid plaques in Alzheimer’s disease
patients. Our previous results using a combination of surface plasmon resonance experiments with molecular modeling calculations
suggested that the Asp dyad in β-secretase bound to hydroxylethylene containing inhibitors adopts a neutral charged state. In this
work, we show that the Asp dyad diprotonated state reproduced the binding ranking of a set of these inhibitors better than alternative
protonation states.
’ INTRODUCTION
we calculated the titration curves for a set of BACE-1 ligand
complexes.10 One of our most striking results of that work relates
to the protonation state of peptidic inhibitors with hydroxylethy-
lene (HE) based isosteres. For these inhibitors, our calculations
predicted that the Asp dyad will be diprotonated at low pH, as
opposed to all previous studies that pointed to a monoproto-
nated Asp dyad state.10
To further investigate this issue, in this work we have assembled
a set of transition state analogue peptidic inhibitors with HE
isosteres (see Table 1), some of which have been already synthe-
sized and tested by other authors, while others have been produced
and assayed by our group. Using molecular mechanics based
protocols we show that the diprotonated state reproduced the
binding ranking of our set of peptidic inhibitors better than the
ones based on either of the monoprotoned Asp dyad bearing
enzymes proposed by previous studies,6ꢀ9 a result that provides
further support to our enzyme’s charge assignment when bound to
this type of transition state analogue inhibitors. Finally, we discuss
the proposed Asp dyad protonation state in relation to the possible
existenceoflowbarrier hydrogenbonds(LBHBs) inthese systems.
Alzheimer’s disease (AD) is a widespread, neurodegenerative,
dementia-inducing disorder, characterized by the formation of
amyloid plaques in the brain.1 The genesis of this construct is
catalyzed by a tandem of two proteases identified as β- and γ-
secretases.1,2 It is known that the former enzyme (also referred to
as BACE-1 and Memapsin 2) participates in the rate-limiting step
of the hydrolytic process that leads to the APP fragments,1ꢀ3 a
fact that has converted β-secretase into a major target for drugs
against Alzheimer’s disease.2,4,5
There are several issues related to the computer-aided design
of novel ligands that could enhance the success of in silico high
throughput screening protocols but that have not been fully
addressed yet. One of the most vital ones is the protonation state
of the many buried acidic residues found in this protein, including
the active site Asp dyad, which lends to this enzyme an optimal
catalytic activity at low pH.3
There have been many attempts to predict the charged state of
the active site Asp dyad by computational methods.6ꢀ9 The results
of those calculations support the hypothesis that the most favored
Asp dyad charge state is the one that has only one of the Asp
residues protonated. Nevertheless, quantum mechanical (QM)
based calculations have left open the possibility that the Asp dyad
might be neutral at the acidic optimal pH of the protein.8
Recently, we have determined the effect of the pH on the
affinities of a set of inhibitors with a variety of chemical motifs to
the ectodomain BACE-1 region by a surface plasmon resonance
(SPR) biosensor based assay.10 To understand the molecular
interactions that underlie the diverse optimum pH for the
binding of the various inhibitors as observed experimentally,
’ RESULTS AND DISCUSSION
Chemistry. The δ-amino acid isosters of inhibitors 3 and 4
were prepared as shown in Scheme 1, through a methodology
that allows the introduction of a variety of side chain groups at
P1/P10 using as a key intermediate the oxiranelactone 9, easily
Received: October 18, 2010
Published: March 31, 2011
r
2011 American Chemical Society
3081
dx.doi.org/10.1021/jm101568y J. Med. Chem. 2011, 54, 3081–3085
|

Journal of Medicinal Chemistry
BRIEF ARTICLE
Table 1. Inhibitors Studied in This Work and Their Experi-
mental BACE-1 Inhibitory Activity
Scheme 1. Total Synthesis of δ-Amino Acid Isosters 14a and
14ba
a Reagents and conditions: (a) as in ref 11; (b) K2CO3, acetone, rt, 20 h;
(c) iPrMgCl, Li2CuCl4, THF, ꢀ78 °C, 30 min; (d) MsCl, pyridine, 0 °C,
5 h; (e) NaN3, DMF, 80 °C, 12 h; (f) LDA, THF, ꢀ78 °C, 1 h; (g)
acetone for 12a and MeI for 12b; (h) H2 (1 atm), PdꢀC (10%),
(Boc)2O, AcOEt, rt, 5 h; (i) ClCOCO2Me, DMAP, Et3N, THF, 0 °C, 1
h and then HPh3Sn, AIBN, toluene, 110 °C, 3 h; (j) LiOH (1 M), DME,
rt, 12 h; (k) ClTBDMS, imidazole, DMF, rt, 15 h; (l) MeOH, rt, 2 h.
intermediate lactone 11 was alkylated at C-2 position by treating
it with LDA followed by methyl iodide to give 12b in 70% yield.
Azide 12b was then reduced to amine with hydrogen, using
palladium/carbon as catalyst, and the amine protected in the
reaction medium with tert-butoxycarbonyl anhydride to give 13b
in 95% yield. Finally compound 13b was transformed into the
isoster 14b in 82% yield following a described procedure.12
The above synthetic procedure for preparing the δ-amino acid
isosters presents two advantages over previous approaches
described in the literature. In the first place, the protocols
presented here improve previous synthetic routes for the prepara-
tion of the oxirane intermediate 9 starting from a sugar.13 Second,
our synthetic strategy affords the insertion of any alkyl group at
position C-5 in the lactone, a feature that will enable the synthesis
ofisosteres with a varietyof side chain fragments beyondthose that
could be obtained from commercially available R-aminoacids.14
Binding Affinity Prediction. Table 1 lists the peptidic in-
hibitors studied here and their BACE-1 inhibitory activity
expressed as IC50 and/or Ki values whenever available. This
table also displays the partition into fragments for inhibitor 1
(OM99-2) following the rules of Schechter and Berger.15 Having
as a reference inhibitor 1, compound 4 introduces an Ala f Val
replacement in P10. Studies on the substrate subsite specificity of
BACE-1 by the group of J. Tang16 indicate that the S10 site is ill
suited for bulky chains like those in Trp, Phe, or Val, the one that
is found in the inhibitor 4. Perusal of the IC50 values indicates
that its presence increases the IC50 value (as compared to 1) by
9-fold, in agreement with the predictions of the substrate
fragment specificity.16
a See SI for details on IC50 evaluation. b Data from ref 17. c Data from ref 16.
prepared from D-glucono-1,5-lactone (7). Thus D-glucono-1,5-
lactone (7) was transformed into lactone 8 by following a
procedure described in the literature,11 and this lactone, 8, when
treated in basic media with potassium carbonate in acetone, gave
quantitatively the expected oxiranelactone 9.
Key intermediate lactone 9 was first alkylated at C-6 when
treated with isopropylmagnesium chloride to give compound 10
in 50% yield. The hydroxyl group in 10 was then transformed
into the corresponding mesylate by treatment with methanesul-
fonyl chloride, followed by reaction with sodium azide to give
azidolactone 11 in 89% yield. Alkylation of lactone 11 at C-2
position with LDA and subsequent addition of acetone afforded,
in a 67% yield, compound 12a, with all the functionality needed
for the final isoster. To transform compound 12a into the δ-
amino acid isoster 14a, compound 12a was reduced at the azido
position with hydrogen, using a mixture of palladium/carbon as
catalyst, and the resulting amine treated in the reaction mixture
with tert-butoxycarbonyl anhydride to give compound 13a in
95% yield. Elimination of the alcohol group in 13a when
submitted to reaction with methyl chlorooxoacetate and triphe-
nyltin hydride, using AIBN as starting radical reaction reactive,
gave the expected compound 13c in 70% yield. Finally com-
pound 13c was transformed into the isoster 14a in a 76% yield by
following a previously described procedure.12
On the other hand, the ligand 3 introduces three aminoacid
substitutions in the N-terminal segment of 1: the Asn f Asp at
P2, the Val f Ile at P3, and the Glu f Gln at P4. The substrate
specificity studies predict that while Val f Ile replacement at P3
should increase the enzyme affinity, the Asn f Asp at P2 and the
Glu f Gln at P4 modifications will decrease it.16 In accordance
with this line of thought, inhibitor 3 has a lower BACE-1
inhibitory activity (higher IC50 value) than 1, indicating that
the substitution at P3 does not compensate for the detrimental
effect of the modifications at P2 and P4.
The isoster 14b was prepared by an easier version of the
protocol that afforded compound 14a because alkylation at
position C-2 of the lactone can be done directly. Thus, the
3082
dx.doi.org/10.1021/jm101568y |J. Med. Chem. 2011, 54, 3081–3085

Journal of Medicinal Chemistry
BRIEF ARTICLE
Finally, inhibitors 5 and 6 were originally the result of a second
iteration drug design cycle aimed at reducing the size and
molecular weight of BACE-1 inhibitors like 1 and 2 (OM00-3),
based on the realization that in spite of the large size of this
protein, the active site of BACE-1 is smaller than that of other
aspartic protease enzymes.17 While the parent compounds
extended from fragment P4 to P40, the analogues 5 and 6 cover
a shorter span from either P4 to P30 in 6, or from P3 to P30 in 5. It
was found that in agreement with the design premises, some of
the shorter compounds were very potent BACE-1 inhibitors.17
As seen in Table 1, the experimental data obtained in this work
using an in vitro peptide cleavage FRET assay, together with the
data extracted from the literature,16,17 point toward the following
BACE-1 inhibitory activity:
Table 2. Binding Affinities Scoring Function Values Calcu-
lated for the Docked Poses with the Catalytic BACE-1 Asp
Dyad in Three Different Protonation Statesa
AspH32/
AspH228b
AspH32/
Asp(ꢀ)228b
Asp(ꢀ)32/
AspH32/
Asp(ꢀ)228c
inhibitor
AspH228c
1 (OM99-2) ꢀ95.5 ( 2.1 ꢀ89.2 ( 1.7 ꢀ91.8 ( 2.8
2 (OM00-3) ꢀ99.1 ( 1.5 ꢀ91.6 ( 2.0 ꢀ84.8 ( 1.7
ꢀ80.9 ( 1.0
ꢀ77.7 ( 1.5
ꢀ98.4 ( 2.6
3
4
5
6
ꢀ95.7 ( 1.4 ꢀ98.0 ( 2.0 ꢀ90.1 ( 2.3
ꢀ95.2 ( 1.2 ꢀ97.0 ( 2.6 ꢀ101.9 ( 2.4 ꢀ93.9 ( 2.8
ꢀ75.0 ( 1.4 ꢀ76.6 ( 2.0 ꢀ76.8 ( 2.0
ꢀ96.2 ( 2.2 ꢀ94.6 ( 1.7 ꢀ95.7 ( 1.1
ꢀ77.0 ( 2.1
ꢀ91.4 ( 2.0
a Energies are in kcal/mol. b The ionizable residues charge states were
calculated according to the protocol described in the SI. c The ionizable
residues, other than those in the Asp dyad, were assigned following their
pKa values in solution at pH 7.
2 > 1 ∼ 6 > 3 > 4 > 5
One of the main aims of this work is to compare the ligand
binding affinities generated by a diprotonated BACE-1 Asp dyad
proposed in our recent work10 with that of monoprotonated
states proposed by other groups.6ꢀ9 For this sake, we carried out
a very careful and involved docking by homology, having as a
template the X-ray structure of the complex between BACE-1
and inhibitor 1 (pdb code 1FKN).18 Once the starting docked
structures were obtained, we proceeded to calculate the pKa
values for all titratable residues by the protocol of Spassov and
Yan19 at pH 4.5, a value close to the optimal pH of the enzyme
(see Supporting Information (SI) for details). These calculations
produce titration curves that depict the degree of protonation for
each ionogenic residue at every pH value. Our pKa calculations
indicate that the catalytic Asp dyad of all BACE-1/inhibitor
complexes studied here will be neutral. To illustrate this issue, we
display the titration curves for the Asp dyad residues (see Figure
S1, SI) elicited by the presence of 1 and the shortest inhibitors 5
and 6. As seen from this figure, both Asp residues have an almost
constant protonation fraction close to 1, even at pH values
around 7, indicating that the Asp dyad will remain neutral in the
acidic medium optimum for the enzymatic activity. Moreover,
the protonation fraction pH dependence graphs indicate that the
Asp 32 has a higher pKa value than Asp 228 and hence will remain
protonated at higher pH.
To include the alternative monoprotonated states, we selected
additional starting structures with single charged Asp dyads,
predicted by previous studies.7,8 To assess the effect of the
monoprotonated Asp dyad on the charge state of the remainder
titratable residues, we fixed the Asp dyad in a monoprotonated
state (with the Asp 228 charged) and proceeded to recalculate
the pKa values of the remainder titratable residues, using the
protocol described above, and in detail in the SI. The specific Asp
dyad charge assignment for this calculation was based on the
ionization proclivity (see Figure S1, SI). Our results indicate that
the pKa values of the other residues do not vary, implying that in
this protocol the Asp dyad charge has a marginal effect on the
titration of the other ionogenic residues (results not shown). The
resulting structures with the protonation states that included
monoprotonated and diprotonated catalytic dyads were used in
our molecular mechanics (MM)/molecular dynamics (MD)
protocol, aimed at generating binding poses for each protonation
state. The final production stage of the MD protocol was used to
calculate the binding affinity scoring function for each of the
inhibitors listed in Table 1 (see SI for details).
contains the scoring function values for the charge distributions
calculated in this work, which include a neutral Asp dyad. The
other columns list the scoring function for the enzyme with the
Asp dyad in the monoprotonated states supported by previous
calculations.6ꢀ9 To assess the effect of the rest of the ionogenic
residues, we carried out two simulations for the case where the
Asp 32 is protonated. In the first one, we used the predicted pKa
values of the titratable residues other than the Asp dyad, and in
the second one, we left those residues with their individual charge
assignment in solution at pH 7.
Perusal of Table 2 indicates that the predictions based on the
Asp dyad for the two possible monoprotonated states are
unsuccessful in correctly ranking these ligands. For instance, the
ranking prediction based on the protonated Asp 228 points to 4 as
thetop binder, while2 isestimated tobe one ofthe worst, contrary
to the experimental observations. The binding affinities calculated
with protonated Asp 32, predict that 1 is one of the worst binders,
while 4 one of the best, at variance with the experimental results,
regardless of the charge state of the rest of the titrable residues. As
seen from this table, only the calculations carried out with an Asp
dyad in the doubly protonated state came closer to rank properly
the inhibitors studied here. For instance, calculations carried with
a neutral Asp dyad are the only ones to identify 2 as the inhibitor
with the highest affinity and 4 as next to worst binder, in
agreement with the experimental results (see Table 1).
All previous calculations started from the premise that the Asp
dyad is monoprotonated, but they disagreed on which Asp
residue the proton resides.2 The earliest attempt to assign a
charge to the catalytic dyad was based on MD simulations of the
X-ray structure of the β-secretase complexed with inhibitor 1 in
solution.7 Their results indicate that a neutral Asp32 and ionized
Asp228 combination is the only option that maintains the
intricate network of hydrogen bonds around the Asp dyad
observed in the X-ray structure.7 This outcome is in agreement
with our recent results10 for the 1/BACE-1 complex around a
neutral pH, the one used for the determining its structure by
X-ray crystallography.
Nevertheless, as pointed out above, our calculations suggest
that the dyad is totally neutralized at the pH at which the FRET
based enzyme inhibition assays have been carried out and at
which the enzyme presents its highest activity (pH 4.5). Other
calculations carried out on the 1/BACE-1 complex, like the linear
scaling quantum approach,8 supported a monoprotonated state
Asp dyad with a neutralized Asp 228 and a charged Asp 32, also in
Table 2 lists the binding scoring function values for the three
Asp dyad protonation states studied here. The first column
3083
dx.doi.org/10.1021/jm101568y |J. Med. Chem. 2011, 54, 3081–3085

Journal of Medicinal Chemistry
BRIEF ARTICLE
disagreement with the conclusions based on our binding affinity
predictions. Nevertheless, the authors of this QM study found
out that a diprotonated state could be accessible at optimal
enzyme pH given the rather small energy difference found with
the monoprotonated state.8 The only other study that evaluated
the protonation state for the Asp dyad from the calculated
titration curves is the one carried out by Polgar and Keser€u.6
From these curves, they surmised that the Asp dyad should be
monoprotonated, with the proton located at Asp 32. Never-
theless, perusal of their titration curves indicate that the proton-
ation fraction at pH 4.5 (close to the enzyme’s optimal pH) is
well above 0.5 for both Asp residues, indicating that the Asp dyad
should be diprotonated at that relevant pH value.
There have been a large number of studies on the Asp dyad
protonation state of other aspartic proteases. The results are not
unambiguous, but they support the idea that the dyad charge will
depend on the nature of the inhibitor and of the aspartic protease
studied. For instance, NMR studies combined with Poissonꢀ
Boltzmann calculations carried out on the HIV-1 PR, an aspartic
protease that is the target for many AIDS drugs, support the
premise that there are some HIV-1 PR inhibitors (cyclic ureas)
that elicit a diprotonated Asp dyad, while others, like KNI-272,
produce a monoprotonated Asp dyad.20 Our recent results
performed on a chemically diverse set of BACE-1 inhibitors also
support the ligand dependence of the Asp dyad protonation
state.10
Figure 1. (A) and (C) represent two of the three possible protonation
states (mono- and diprotonated) for the Asp dyad, while (B) schema-
tically represents a LBHB, showing the possible proton positions
enclosed by red and blue circles.
(QM) or quantum mechanical/molecular mechanics (QM/
MM) studies have been applied to the unbound enzymes of
this family25 and enzymes bound to phosphinate and phospha-
nate containing inhibitors.26 Nevertheless, no calculations of this
kind have been applied to aspartic proteases bound to HE
peptidomimetic transition state analogue inhibitors. At present,
we are involved in the development of quantum mechanical
protocols that would overcome some of the shortcomings of this
kind of calculations, like the representation of the quantum
region. The results of our QM calculations will be presented
elsewhere.
’ CONCLUSIONS
To further validate the Asp dyad protonation state in BACE-1
bound to HE transition state analogues, found in our
laboratory,10 we have compared the binding ranking predictions
for a set of HE peptidomimetic inhibitors based on alternative
Asp dyad protonation states by a molecular mechanics method
developed in our group. Some of the inhibitors studied were
synthesized in our laboratory by a protocol that can afford many
isostere variants at P1/P10. Our calculations indicate that Asp
dyad neutral state reproduces better the binding ranking that the
Asp dyad charged alternatives. The results are discussed in the
light of possible LBHB between the Asp dyad components and
the HE isostere hydroxyl group.
One of the most stringent experimental tests on the Asp dyad
protonation state proposals will be a neutron diffraction struc-
tural determination for some of the BACE-1 inhibitor complexes
studied here. Previously, neutron diffraction combined with
X-ray studies in endothiapepsin (a fungal aspartic protease)
bound to a gem-diol inhibitor implied the possible existence of
low barrier hydrogen bonds (LBHB), characterized by short
distances between one of the inhibitors hydroxyl groups and the
oxygens of the Asp dyad (∼2.50 Å).21 There has been some
debate about which kind of experimental observables or com-
puted variables should be used to unambiguously identify a
LBHB, and some authors have even disputed its very existence.22
Nevertheless, the compact hydrogen bonds (HBs), which sup-
port the existence of LBHBs, have been observed in several
X-ray structures of ligand bound aspartic proteases, besides
endothiapepsin, like in HIV-1 PR23 and BACE-1 bound to
peptidic inhibitors,18 suggesting that LBHBs may be a common
feature for some aspartic proteaseꢀinhibitor complexes. It has
been posited that in LBHBs, the proton that mediates the HB
could undergo tunnelling, from the energy well located next to
hydroxyl oxygen to the one next to the carboxylate oxygen.21 In
such a system, the proton could be found at both positions (see
Figure 1B), resulting in the dispersion of the negative charge of
the Asp residue. It may be argued that the tunneling in the LBHB
populates the Asp dyad diprotonated state, and hence it may be
claimed that among the three active site charge distributions
studied here (two monopronated and one diprotonated states),
the latter could be a compromise molecular mechanics repre-
sentation of the entities implied by LBHB (Figure 1). The
improved ranking provided by including a neutral Asp dyad over
the alternative monoprotonated states (shown in this work)
could be a result of a better representation of the LBHB in
molecular mechanics simulations. There has been an increased
body of work aimed at identifying and characterizing LBHBs in
enzymes since they were proposed among others by the group of
Frey.24 In the case of aspartic proteases, quantum mechanical
’ EXPERIMENTAL SECTION
Synthesis of Inhibitors. Inhibitors 4 and 3 were prepared from
the corresponding δ-amino acids isosters 14a and 14b at the Serveis
Cientificotꢁecnics Universitat de Barcelona (Spain) by a Boc/Bzl strategy
following standard protocols. The purity was greater than 95% as judged
by HPLC. See SI for details.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures for
b
the synthesis of isosters 14a and 14b and inhibitors 3 and 4 as
well as IC50 evaluation. Detailed docking and affinity prediction
protocols. This material is available free of charge via the Internet
at http://pubs.acs.org..
’ AUTHOR INFORMATION
Corresponding Author
*For F.S.: phone, 34-881-814402; fax, 34-981-595012; E-mail,
fredy.sussman@usc.es. For J.C.E.: phone, 34-881-815730; fax,
34-981-591014; E-mail, juancarlos.estevez@usc.es.
3084
dx.doi.org/10.1021/jm101568y |J. Med. Chem. 2011, 54, 3081–3085

Journal of Medicinal Chemistry
BRIEF ARTICLE
’ ACKNOWLEDGMENT
(12) Ghosh, A. K.; Shin, D.; Downs, D.; Koelsch, G.; Lin, X.;
Ermolieff, J.; Tang, J. Design of Potent Inhibitors for Human Brain
Memapsin 2 (β-Secretase). J. Am. Chem. Soc. 2000, 122, 3522–3523.
(13) Ghosh, A. K.; McKee, S. P.; Thomson, W. J. An Efficient
Synthesis of Hydroxyethylene Dipeptide Isosteres: The Core Unit of
Potent HIV-1 Protease Inhibitors. J. Org. Chem. 1991, 56, 6500–6503.
(14) (a) D’Aniello, F.; Taddei, M. A Stereoselective Method for the
Preparation of HIV-1 Protease Inhibitors Based on the Lewis Acid
Mediated Reaction of Allylsilanes and N-Boc-R-amino Aldehydes. J. Org.
Chem. 1992, 57, 5247–5250. (b) Ghosh, A. K.; Cappiello, J.; Shin, D.
Ring-closing metathesis strategy to unsaturated γ- and δ-lactones:
synthesis of hydroxyethylene isostere for protease inhibitors. Tetrahe-
dron Lett. 1998, 39, 4651–4654.
This work was supported by financial aid from the Xunta de
Galicia (to F.S. and grant PGIDIT 08CSA002209PR to M.C.V.)
and Ministerio de Educaciꢀon y Ciencia (grants CTQ2005-
00555/BQU and CTQ2008-03105 to R.J.E. and fellowships to
J.M.O. and J.L.D.). The Supercomputing Center of Galicia
(CESGA) provided computer time.
DEDICATION
†Dedicated to the memory of Professor Rafael Suau, recently
deceased.
(15) Schechter, I.; Berger, A. On the size of the active site in
proteases. I. Papain. Biochem. Biophys. Res. Commun. 1967, 27, 157–162.
(16) Turner, R. T., III; Koelsch, G.; Hong, L.; Castenheira, P.;
Ghosh, A.; Tang, J. Subsite Specificity of Memapsin 2 (β-Secretase):
Implications for Inhibitor Design. Biochemistry 2001, 40, 10001–10006.
(17) Ghosh, A. K.; Bilcer, G.; Harwood, C.; Kawahama, R.; Shin, D.;
Hussain, K. A.; Hong, L.; Loy, J. A.; Nguyen, C.; Koelsch, G.; Ermolieff,
J.; Tang, J. Structure-Based Design: Potent Inhibitors of Human Brain
Memapsin 2 (β-Secretase). J. Med. Chem. 2001, 44, 2865–2868.
(18) Hong, L.; Koelsch, G.; Lin, X.; Wu, S.; Terzyan, S.; Ghosh,
A. K.; Zhang, X. C.; Tang, J. Structure of the Protease Domain of
Memapsin 2 (β-Secretase) Complexed with Inhibitor. Science 2000,
290, 150–153.
(19) Spassov, V. Z.; Yan, L. A fast and accurate computational
approach to protein ionization. Protein Sci. 2008, 17, 1955–1970.
(20) Trylska, J.; Antosiewicz, J.; Geller, M.; Hodge, C. N.; Klabe,
R. M.; Head, M. S.; Gilson, M. K. Thermodynamic linkage between the
binding of protons and inhibitors to HIV-1 protease. Protein Sci. 1999,
8, 180–195.
’ ABBREVIATIONS USED
AD, Alzheimer’s disease; APP, amyloid precursor protein;
BACE, β-site APP cleaving enzyme; QM, quantum mechanics;
SPR, surface plasmon resonance; HE, hydroxyethylene; LBHB,
low barrier hydrogen bond; FRET, fluorescence resonance
energy transfer; MM, molecular mechanics; MD, molecular
dynamics; HIV-1 PR, human immunodeficiency virus type 1
protease; HB, hydrogen bond
’ REFERENCES
(1) Selkoe, D. J. Translating cell biology into therapeutic advances in
Alzheimer’s disease. Nature 1999, 399, A23–A31.
(2) Villaverde, M. C.; Gonzꢀalez-Louro, L.; Sussman, F. The Search
for Drug Leads Targeted to the β-Secretase: An Example of the Roles of
Computer Assisted Approaches in Drug Discovery. Curr. Top. Med.
Chem. 2007, 7, 980–990.
(3) Lin, X.; Koelsch, G.; Wu, S.; Downs, D.; Dashti, A.; Tang, J.
Human aspartic protease memapsin 2 cleaves the β-secretase site of
β-amyloid precursor protein. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 1456–1460.
(21) Coates, L.; Tuan, H.-F.; Tomanicek, S.; Kovalevsky, A.; Mus-
tyakimov, M.; Erskine, P.; Cooper, J. The Catalytic Mechanism of an
Aspartic Proteinase Explored with Neutron and X-ray Diffraction. J. Am.
Chem. Soc. 2008, 130, 7235–7237.
(22) Schutz, C. N.; Warshel, A. The Low Barrier Hydrogen Bond
(4) Silvestri, R. Boom in the Development of Non-peptidic β-
Secretase (BACE1) Inhibitors for the Treatment of Alzheimer’s Disease.
Med. Res. Rev. 2009, 29, 295–338.
(LBHB) Proposal Revisited: The Case of the Asp His Pair in Serine
Proteases. Proteins 2004, 44, 711–723.
3 3 3
(23) Jaskꢀolski, M.; Tomaselli, A. G.; Sawyer, T. K.; Staples, D. G.;
Heinrikson, R. L.; Schneider, J.; Kent, S. B. H.; Wlodawer, A. Structure at
2.5 Å Resolution of Chemically Synthesized Human Immunodeficiency
Virus Type 1 Protease Complexed with a Hydroxyethylene-Based
Inhibitor. Biochemistry 1991, 30, 1600–1609.
(24) Frey, P. A.; Whitt, S. A.; Tobin, J. B. A low-barrier hydrogen
bond in the catalytic triad of serine proteases. Science 1994,
264, 1927–1930.
(25) Porter, M. A.; Molina, P. A. The Low-Barrier Double-Well
Potential of the Oδ1ꢀHꢀOδ1 Hydrogen Bond in Unbound HIV
Protease: A QM/MM Characterization. J. Chem. Theory Comput.
2006, 2, 1675–1684.
(26) Vidossich., P.; Carloni, P. Binding of Phosphinate and Phos-
phonate Inhibitors to Aspartic Proteases: A First-Principles Study
J. Phys. Chem. B 2006, 110, 1437–1442.
(5) (a) Ghosh, A. K.; Gemma, S.; Tang, J. β-Secretase as a
Therapeutic Target for Alzheimer’s Disease. Neurotherapeutics 2008,
5, 399–408. (b) Ghosh, A. K. Harnessing Nature’s Insight: Design of
Aspartyl Protease Inhibitors from Treatment of Drug-Resistant HIV to
Alzheimer’s Disease. J. Med. Chem. 2009, 52, 2163–2176.
(6) Polgꢀar, T.; Keser€u, G. M. Virtual Screening for β-Secretase
(BACE1) Inhibitors Reveals the Importance of Protonation States at
Asp32 and Asp228. J. Med. Chem. 2005, 48, 3749–3755.
(7) Park, H.; Lee, S. Determination of the Active Site Protonation
State of β-Secretase from Molecular Dynamics Simulation and Docking
Experiment: Implications for Structure-Based Inhibitor Design. J. Am.
Chem. Soc. 2003, 125, 16416–16422.
(8) Rajamani, R.; Reynolds, C. H. Modeling the Protonation States
of the Catalytic Aspartates in β-Secretase. J. Med. Chem. 2004,
47, 5159–5166.
(9) Yu, N.; Hayik, S. A.; Wang, B.; Liao, N.; Reynolds, C. H.; Merz,
K. M., Jr. Assigning the Protonation States of Key Aspartates in β-
Secretase Using QM/MM X-ray Structure Refinement. J. Chem. Theory
Comput. 2006, 2, 1057–1069.
(10) Domínguez, J. L.; Christopeit, T.; Villaverde, M. C.; Gossas, T.;
Otero, J. M.; Nystr€om, S.; Baraznenok, V.; Lindstr€om, E.; Danielson,
U. H.; Sussman, F. Effect of the Protonation State of the Titratable
Residues on the Inhibitor Affinity to BACE-1. Biochemistry 2010,
49, 7255–7263.
(11) (a) Bock, K.; Lundt, I.; Pedersen, C. Preparation of some
bromodeoxyaldonic acids. Carbohydr. Res. 1979, 68, 313–319. (b)
Lundt, I.; Pedersen, C. Preparation of some 2,3-Dideoxylactones by
an Unusual Catalytic Hydrogenolysis. Synthesis 1986, 1052–1054.
3085
dx.doi.org/10.1021/jm101568y |J. Med. Chem. 2011, 54, 3081–3085