Several trypanosomatid cyclic nucleotide phosphodiesterases (PDEs) possess a unique, parasite-specific cavity near the ligand-binding region that is referred to as the P-pocket. One of these enzymes, Trypanosoma brucei PDE B1 (TbrPDEB1), is considered a drug target for the treatment of African sleeping sickness. Here, we elucidate the molecular determinants of inhibitor binding and reveal that the P-pocket is amenable to directed design. By iterative cycles of design, synthesis, and pharmacological evaluation and by elucidating the structures of inhibitor-bound TbrPDEB1, hPDE4B, and hPDE4D complexes, we have developed 4a,5,8,8a-tetrahydrophthalazinones as the first selective TbrPDEB1 inhibitor series. Two of these, 8 (NPD-008) and 9 (NPD-039), were potent ( Ki = 100 nM) TbrPDEB1 inhibitors with antitrypanosomal effects (IC50 = 5.5 and 6.7 μM, respectively). Treatment of parasites with 8 caused an increase in intracellular cyclic adenosine monophosphate (cAMP) levels and severe disruption of T. brucei cellular organization, chemically validating trypanosomal PDEs as therapeutic targets in trypanosomiasis.
Several trypanosomatid cyclic nucleotide phosphodiesterases (pan class="Gene">PDEs) possess a unique, parasite-specific cavity near the ligand-binding region that is referred to as the P-pocket. One of these enzymes, Trypanosoma brucei PDE B1 (TbrPDEB1), is considered a drug target for the treatment of African sleeping sickness. Here, we elucidate the molecular determinants of inhibitor binding and reveal that the P-pocket is amenable to directed design. By iterative cycles of design, synthesis, and pharmacological evaluation and by elucidating the structures of inhibitor-bound TbrPDEB1, hPDE4B, and hPDE4D complexes, we have developed 4a,5,8,8a-tetrahydrophthalazinones as the first selective TbrPDEB1 inhibitor series. Two of these, 8 (NPD-008) and 9 (NPD-039), were potent ( Ki = 100 nM) TbrPDEB1 inhibitors with antitrypanosomal effects (IC50 = 5.5 and 6.7 μM, respectively). Treatment of parasites with 8 caused an increase in intracellular cyclic adenosine monophosphate (cAMP) levels and severe disruption of T. brucei cellular organization, chemically validating trypanosomal PDEs as therapeutic targets in trypanosomiasis.
The family of trypanosomatids
is responsible for three major neglected
tropical diseases (NTDs) caused by protozoan parasites, namely, Chagas
disease (pan class="Species">Trypanosoma cruzi), leishmaniasis (Leishmania spp.), and human African trypanosomiasis, also
called African sleeping sickness (Trypanosoma brucei rhodesiense and T. brucei gambiense).[1] Although millions of people, mostly in underdeveloped countries,
are at risk, current options for drug treatment of these infectious
diseases remain limited.[2]
Recently,
the inhibition of 3′,5′-cyclic nucleotide
phosphodiesterases (PDEs) has emerged as a new approach to target
these kinetoplastid protozoans.[3−6] All three parasites, T. cruzi, Leishmania spp., and T. brucei spp., have the same set of
four different class I PDE families in their genomes.[7] Using an inducible siRNA approach, Seebeck and co-workers
demonstrated that simultaneous inhibition of T. bruceiPDE B1 (TbrPDEB1) and TbrPDEB2 blocks parasite proliferation and
eliminates parasitemia from infected mice,[8] thereby establishing the parasite PDEs as promising drug targets.Subsequent efforts to identify TbrPDEB1 inhibitors have resulted
in the repurposing of ligands that were originally developed as inhibitors
for pan class="Species">human PDE4 (hPDE4) as TbrPDEB1 inhibitors (Figure ).[9−12] However, until now, all compounds reported as TbrPDEB1
inhibitors remain more active on hPDE4. The most potent TbrPDEB1 inhibitor
reported to date, NPD-001 (1), is 10-fold more potent
on hPDE4 (hPDE4B1 IC50 = 0.6 nM) than on TbrPDEB1 (IC50 = 4 nM).[11] In a series of analogs
of 1 a close correlation between TbrPDEB1 inhibitory
potency and antiparasitic activity has been found.[11] Although these TbrPDEB1 inhibitors raise cAMP levels in T. brucei and show strong antiparasitic effects in vitro,[10−12] their potent inhibition of hPDE4 is undesirable.[13−16] Considerable screening and synthetic efforts have been reported,[9−12] but no TbrPDEB1-selective (over hPDEs) inhibitors have been identified
to date. Despite the high similarity of hPDE4 and TbrPDEB1 substrate
binding pockets,[17] tantalizing evidence
for discriminating selectivity is evident from the observation that
established hPDE4 inhibitors such as rolipram and etazolate do not
inhibit either the enzyme TbrPDEB1 or T. brucei cellular
PDE activity and do not kill the parasite.[9,10]
Figure 1
Representative
hPDE4 inhibitors and reported TbrPDEB1 inhibitors.[9,10,12]
Representative
pan class="Gene">hPDE4 inhibitors and reported pan class="Chemical">TbrPDEB1 inhibitors.[9,10,12]
In an effort to gain structural understanding to aid in the
design
of selective TbrPDEB1 inhibitors, we previously reported the crystal
structure of the apo-TbrPDEB1 catalytic domain as part of a structure-based
virtual screening effort.[17] The high-resolution
structure of the TbrPDEB1 catalytic domain revealed an open cavity
formed between helix 14 (H14), helix 15 (H15), and the M-loop (Figure ). This cavity is
also present in the apo structure of related parasite PDEs, including Leishmania majorPDE B1 (LmPDEB1)[18] and T. cruziPDE C (TcrPDEC),[19] but is not present in any of the 11 humanPDEs.[20,21] Being parasite-specific, this cavity has been named the P-pocket.
As the most prominent structural difference between parasite and humanPDE enzymes, the P-pocket has been considered a promising feature
for selective TbrPDEB1 inhibitor design.[17] Molecular docking studies suggested that the phenyltetrazole-containing
tail group of inhibitors 1 and 2 can interact
with residues deep inside the P-pocket.[11,12] However, the
preferential inhibition of hPDE4 over TbrPDEB1 by these compounds
is not consistent with this hypothesis, and therefore the usefulness
of targeting the P-pocket to achieve selectivity remains to be demonstrated.
Figure 2
Crystal
structure of apo-TbrPDEB1 (PDB code 4I15) active site highlighting
the parasite P-pocket. Helix 14 (H14), helix 15 (H15), and the M-loop
are labeled in red, the main binding site features are labeled in
black, water molecules have been omitted for clarity. The conserved
Gln874Q.50, the hydrophobic clamp residues Val840HC.32 and Phe877HC.52, and the distal aromatic residues Phe844S.35 and Phe880HC2.54 are shown as sticks. P-pocket
residues Ala837Q1.30, Thr841Q2.33, Tyr845Q2.36, Asn867Q2.43, Met868Q2.44, Glu869Q2.45, and Leu870Q2.46 are shown as lines. All binding
site residues have been named according to the PDEStrIAn nomenclature
convention.[21]
Crystal
structure of apo-TbrPDEB1 (PDB code 4I15) active site highlighting
the pan class="Chemical">parasite P-pocket. Helix 14 (H14), helix 15 (H15), and the M-loop
are labeled in red, the main binding site features are labeled in
black, water molecules have been omitted for clarity. The conserved
Gln874Q.50, the hydrophobic clamp residues Val840HC.32 and Phe877HC.52, and the distal aromatic residues Phe844S.35 and Phe880HC2.54 are shown as sticks. P-pocket
residues Ala837Q1.30, Thr841Q2.33, Tyr845Q2.36, Asn867Q2.43, Met868Q2.44, Glu869Q2.45, and Leu870Q2.46 are shown as lines. All binding
site residues have been named according to the PDEStrIAn nomenclature
convention.[21]
Here, we present detailed structural insights in TbrPDEB1
ligand
binding by reporting the first inhibitor-bound pan class="Chemical">TbrPDEB1 crystal structures,
including cocrystal structures with 1 and 2. These structures clearly explain the lack of selectivity of the
previously reported inhibitors and guide the design and synthesis
of novel P-pocket targeting inhibitors with, for the first time, selectivity
over hPDE4B. We show that these novel compounds inhibit cAMP degradation
in T. brucei with consequent trypanocidal activity.
Results
and Discussion
The most potent TbrPDEB1 inhibitor reported
to date, pan class="Gene">NPD-001 (1), is a nanomolar inhibitor of TbrPDEB1
and a sub-nanomolar
inhibitor of hPDE4 subtypes (IC50 < 1 nM) but is over
100-fold less active against all other humanPDEs.[10,11] The structurally related pyrazolone analogs, such as NPD-340 (2), are also potent TbrPDEB1 inhibitors, but these too are
more potent at hPDE4.[12] Our earlier molecular
docking studies suggested that the central dialkoxyphenyl moiety of 1 and 2 forms a hydrogen bond to the PDE family
wide conserved Gln874Q.50 residue while occupying the hydrophobic
clamp, formed by Val840HC.32 and Phe877HC.52 in TbrPDEB1.[11,12] In addition to these hallmark
interactions of PDE inhibitor binding, the docking studies also suggested
that 1 and 2 bind in the P-pocket of TbrPDEB1
with their tetrazole-containing flexible tail group. The TbrPDEB1
activity of these compounds is promising, but their even higher potency
against hPDE4 seemed inconsistent with successful targeting of the
P-pocket as a means of obtaining selectivity.
In order to gain
a better understanding of the lack of selectivity
of these compounds and to guide our ongoing medicinal chemistry efforts,
we obtained the crystal structure of 1 in complex with
the catalytic domain of TbrPDEB1 at a resolution of 1.73 Å (Figure a). The overall fold
of the protein was highly similar (rms deviation of Cα = 0.38
Å over 630 atoms) to that observed for the apo structure of pan class="Chemical">TbrPDEB1
(PDB code 4I15). Two protein molecules pack in the crystal asymmetric unit of TbrPDEB1
where the binding site of one of the molecules lies close to a symmetry-related
molecule. We refer to the interactions of TbrPDEB1 with ligands at
the chain that is free from such crystal packing effects, unless noted
otherwise. Consistent with biochemical data, 1 was found
to bind in the substrate-binding pocket. The dialkoxyphenyl moiety
of 1 was situated in the hydrophobic clamp formed by
Val840HC.32 and Phe877HC.52, and its two ether
functionalities were within hydrogen bonding distance of the conserved
Gln874Q.50 residue. The (4aS,8aR)-enantiomer of 1 was observed in the crystal
structure. In contrast to the docking pose suggested by the earlier
computational studies,[11,12] the tail group of 1 is oriented away from the P-pocket, with the tetrazole moiety instead
forming a stacking interaction with Phe880HC2.54 in H15.
Figure 3
Crystal
structures of nonselective inhibitors with TbrPDEB1, hPDE4B,
and hPDE4D. (a) TbrPDEB1–1, including |Fo – Fc|αcalc electron density map contoured at 2.5σ. (b)
TbrPDEB1–2. (c) hPDE4B–1.
(d) hPDE4D–1. In all panels: key binding site
amino acid residues are shown as sticks, minor amino acid residues
are shown as lines. The P-pocket is shown as a gray surface with P-pocket
residues represented as lines, and key polar interactions are depicted
with black dashed lines. Water molecules are displayed as red spheres,
zinc cations as metallic blue spheres, and magnesium cations as green
spheres.
Crystal
structures of nonselective inhibitors with TbrPDEB1, pan class="Gene">hPDE4B,
and hPDE4D. (a) TbrPDEB1–1, including |Fo – Fc|αcalc electron density map contoured at 2.5σ. (b)
TbrPDEB1–2. (c) hPDE4B–1.
(d) hPDE4D–1. In all panels: key binding site
amino acid residues are shown as sticks, minor amino acid residues
are shown as lines. The P-pocket is shown as a gray surface with P-pocket
residues represented as lines, and key polar interactions are depicted
with black dashed lines. Water molecules are displayed as red spheres,
zinc cations as metallic blue spheres, and magnesium cations as green
spheres.
The crystal structure of the pan class="Chemical">TbrPDEB1
catalytic domain with 2 was determined at 2.25 Å
resolution (Figure b). Compounds 1 and 2 adopted an identical
binding mode with conservation
of the key interactions of the pan class="Chemical">catechol and phenyltetrazole tail with
TbrPDEB1.
We also determined the crystal structure of 1 in complex
with hPDE4B (Figure c) and pan class="Gene">hPDE4D (Figure d) at 2.4 and 2.25 Å resolution, respectively. As expected,
the hPDE4B structure displayed a similar overall fold to the TbrPDEB1
structure bound to 1 (rms deviation of Cα = 1.26
Å over 247 atoms). In line with its high potency as a hPDE4 inhibitor, 1 was observed in the cAMP-binding site and interacted with
hPDE4B through a bifurcated hydrogen bond to the invariant Gln615Q.50. As in the TbrPDEB1 structure, there are key interactions
with the hydrophobic clamp, formed by Ile582HC.32 and Phe618HC.52, in addition to a well-positioned stacking interaction
of the tetrazole moiety with Tyr621HC2.54. In the catalytic
domain of hPDE4D the interaction pattern exhibited by 1 was identical to that observed in the hPDE4B structure. In all structures,
only the single (4aS,8aR)-enantiomer
of 1 was observed. The similar binding mode observed
in these structures unequivocally establish that the tetrazole tail
present in 1 and 2 does not target the P-pocket
in TbrPDEB1, which potentially explains the lack of selectivity of
these compounds for TbrPDEB1 over hPDE4B.
Inspection of the
binding site characteristics of TbrPDEB1 and
the binding modes observed for 1 and 2 highlighted
that the hydrophobic clamp and the extended hydrophobic region (pan class="Gene">HC2)
(Figure S1) provide a favorable environment
for the flexible linker of these inhibitors to fold away from the
P-pocket and interact with the aromatic residue found in H15, being
Phe880HC2.54 in TbrPDEB1 and Tyr621HC2.54 in
hPDE4B. We hypothesized that modification of the flexible alkyl linker
could prevent such a hydrophobic collapse and utilized the Phosphodiesterase
Structure and Ligand Interaction Annotated (PDEStrIAn)[21] tool for the design of such ligands. More specifically,
it was observed that some cocrystallized humanPDE inhibitors (for
example, PDB codes 3IAD and 4AEL)
have a central aromatic moiety bound in the hydrophobic clamp while
having additional (hetero)aryl groups along the vector that would
point toward the P-pocket in TbrPDEB1. It was hypothesized that replacing
the flexible linker at the 3-position of the dialkoxyphenyl ring of 1 with a phenyl ring would provide a biphenyl system with
the desired rigidity and vector to target the P-pocket.
Accordingly,
we produced an initial series of biphenyl analogs 3–6 and screened them as inhibitors of
TbrPDEB1 and hPDE4B1 (Table ). Introduction of the biphenyl linker indeed led to some
selectivity for TbrPDEB1 over hPDE4B1 in carboxylic acids 3 and 4. The TbrPDEB1 selectivity was higher for the para-substituted 3 over its meta-substituted analog (4). The potency at TbrPDEB1 increased
10-fold upon introduction of a carboxamide group at the para-position in compound 5. In accordance with the overlapping
pharmacological profiles of hPDE4 subtypes, 5 showed
the same potency against all tested hPDE4s but was selective over
the other humanPDE subtypes (Table S5).
Table 1
Introduction of the Biphenyl Linkera
pKi values
are averages (n ≥ 2, with SD less than ±0.3
for all).
Compounds are
racemic mixtures of cis-isomers.
Selectivity is calculated as hPDE4B1
(Ki)/TbrPDEB1 (Ki).
pKi values
are averages (n ≥ 2, with SD less than ±0.3
for all).Compounds are
racemic mixtures of cis-isomers.Selectivity is calculated as pan class="Gene">hPDE4B1
(Ki)/pan class="Chemical">TbrPDEB1 (Ki).
These findings
led to further design and synthesis of a series
of para-substituted pan class="Chemical">carboxamide analogs, using 5 as a template to further improve TbrPDEB1 potency and selectivity
over hPDE4B (Table ). Extending the carboxamide of 5 with a methoxyethyl
substituent (7) was found to increase the selectivity
for TbrPDEB1 7-fold, while the potency was comparable to that of its
parent compound. Gratifyingly, glycinamide 8, and its
isopropyl-substituted counterpart 9, were even more potent
TbrPDEB1 inhibitors (Ki = 100 nM and 99
nM, respectively). The cycloheptyl-substituted compound 8 displayed 10-fold selectivity over hPDE4B1 and was also selective
over the other hPDE4 subtypes (Table S5). The potency of 8 was fully explained by the biological
activity of a single enantiomer; (8a) was found to be
100-fold less potent against TbrPDEB1 than its optical isomer 8b. Both enantiomers of 8 were selective TbrPDEB1
inhibitors, with 8b showing 16-fold selectivity. Furthermore,
isopropyl analog 9 showed a 19-fold preference for TbrPDEB1
over hPDE4B1. These glycinamide compounds are the most selective TbrPDEB1
inhibitors reported to date. The introduction of N-methyl, N,N-dimethyl, or N-isopropyl substituents on different positions of the P-pocket
tail (compounds 10–13) led to decreases
in potency as well as selectivity. Alkylation was least tolerated
at the terminal nitrogen (12, 13), indicating
that changes in hydrophobicity or hydrogen bonding properties are
unfavorable. Increasing the chain length of glycinamide 8 with an additional carbon, to give 14, resulted in
a substantial decrease in potency at TbrPDEB1 but not at hPDE4B1,
decreasing selectivity for TbrPDEB1 to only 2-fold. Inversion of the
carboxamide of 14 affected the hPDE4B1 potency and restored
the TbrPDEB1 selectivity back to 5-fold in compound 15. Finally, further extension of the glycinamide of 8 with a hydroxyethyl (16) or methoxyethyl (17) substituent was tolerated and resulted in ∼8 fold selective
TbrPDEB1 inhibitors. These findings highlight that selectivity for
TbrPDEB1 over hPDE4B1 is very sensitive to small changes in the inhibitor
tail group.
Table 2
Exploration of SAR in the P-Pocketa
pKi values
are averages (n ≥ 2, with SD less than ±0.3
for all).
Compounds are
racemic mixtures of cis-isomers unless otherwise
noted.
Selectivity is calculated
as hPDE4B1
(Ki)/TbrPDEB1 (Ki).
pKi values
are averages (n ≥ 2, with SD less than ±0.3
for all).Compounds are
racemic mixtures of cis-isomers unless otherwise
noted.Selectivity is calculated
as pan class="Gene">hPDE4B1
(Ki)/pan class="Chemical">TbrPDEB1 (Ki).
In order to validate
the P-pocket targeting approach to selectivity,
we obtained the crystal structures of 8 (Figure a) and 9 (Figure b) with TbrPDEB1
at a resolution of 1.8 and 2.0 Å, respectively. The overall fold
of these structures was highly similar to the apo-pan class="Chemical">TbrPDEB1 (PDB code 4I15) and the TbrPDEB1–1 crystal structures (rms deviations of Cα < 0.3
Å). Both ligands bonded to the invariant Gln874Q.50 and occupied the hydrophobic clamp with the methoxyphenyl ring of
their biphenyl linker. The (4aR,8aS)-enantiomer was observed in the crystal structures of both inhibitors.
In the case of 8, we therefore propose the 100-fold more
potent enantiomer 8b to be (4aR,8aS)-configured. Compounds 8 and 9 successfully targeted the P-pocket. The glycinamide tail of 8 formed a direct hydrogen bond to Tyr845Q2.36 and
three water-mediated hydrogen bonds to the following residues: (1)
Thr841Q2.33 and Met861S.40; (2) Leu870Q2.46 and Gln874Q.50; and (3) Gly873Q2.49. Alignment
of the crystal structure of 8 with the nonselective inhibitor 1 clearly highlights the distinctive binding mode of the tail
groups (Figure ).
However, the 4a,5,8,8a-tetrahydrophthalazinone and methoxy-substituted
phenyl rings of 1 and 8 closely aligned
in the active site of TbrPDEB1 and therefore will be competitive inhibitors
of cAMP breakdown. The tail group of 9 was observed in
a slightly different conformation, with direct hydrogen bonds formed
to Tyr845Q.2.36 and Glu869Q2.45 and two water-mediated
hydrogen bonds to (1) Thr841Q2.33 and Met861S.40 and (2) Leu870Q2.46 and Gln874Q.50. Average
temperature factors (B factors) indicated greater flexibility in the
P-pocket region of TbrPDEB1 compared to the rest of the protein. The
observed electron density for the glycinamide tail of these inhibitors
was somewhat weaker than that for the rest of the ligand, indicative
of greater ligand flexibility in the P-pocket region. Both findings
have been supported by observations from molecular dynamics (MD) simulations
(data not shown).
Figure 4
TbrPDEB1 crystal structures of selective inhibitors. (a)
TbrPDEB1–8, including |Fo – Fc|αcalc electron density map
contoured at 2.5σ; (b) TbrPDEB1–9.
Figure 5
Alignment of the TbrPDEB1 crystal structures
of 1 (cyan)
and 8 (yellow). Coordinates of Gln874Q.50,
P-pocket residues, and metal cations were derived from TbrPDEB1–8 and are shown for clarity.
pan class="Chemical">TbrPDEB1 crystal structures of selective inhibitors. (a)
pan class="Chemical">TbrPDEB1–8, including |Fo – Fc|αcalc electron density map
contoured at 2.5σ; (b) TbrPDEB1–9.
Alignment of the TbrPDEB1 crystal structures
of 1 (cyan)
and 8 (yellow). Coordinates of pan class="Chemical">Gln874Q.50,
P-pocket residues, and metal cations were derived from TbrPDEB1–8 and are shown for clarity.
Attempts to elongate the glycinamide tail of compound 8 with a 2-hydroxyethyl substituent to further probe the P-pocket
led to compound 16. In the 2.1 Å crystal structure
of 16 with pan class="Chemical">TbrPDEB1, the elongated tail was found to
penetrate the P-pocket (Figure ). The overall binding mode for the single enantiomer (4aR,8aS)-16 was similar to that
observed for glycinamide analogs 8 and 9. This modification was not accompanied by an increase in TbrPDEB1
activity.
Figure 6
X-ray crystal structure of TbrPDEB1–16.
X-ray crystal structure of pan class="Chemical">TbrPDEB1–16.
The main residues lining the P-pocket
in TbrPDEB1 are pan class="Chemical">Ala837Q1.30, Thr841Q2.33, Tyr845Q2.36, Asn867Q2.43, Met868Q2.44, Glu869Q2.45, and
Leu870Q2.46, with residues from the adjacent M-loop potentially
playing a minor role. The character of these residues and the flexibility
of the P-pocket prompted us to direct chemistry efforts toward a series
of aliphatic heterocyclic tail substituents in order to introduce
rigidity into the tail region to further interrogate the P-pocket
(Table ).
Table 3
SAR of Aliphatic Heterocycles in the
P-Pocketa
pKi values
are averages (n ≥ 2, with SD less than ±0.3
for all).
Compounds are
racemic mixtures of cis-isomers.
Selectivity is calculated as hPDE4B1
(Ki)/TbrPDEB1 (Ki).
pKi values
are averages (n ≥ 2, with SD less than ±0.3
for all).Compounds are
racemic mixtures of cis-isomers.Selectivity is calculated as pan class="Gene">hPDE4B1
(Ki)/pan class="Chemical">TbrPDEB1 (Ki).
Compound 18 and its N-methylated
counterpart 19 were designed as constrained analogs of 8. However, cyclization of the pan class="Chemical">glycinamide led to a loss of
selectivity for TbrPDEB1 over hPDE4B1 and reduced potency. As seen
for the noncyclized analogs, N-methylation (19) led to a further reduction in potency. The pyrrolidin-3-one
analog (20) was comparable to 18 and 19 in TbrPDEB1 potency but was ∼3-fold selective over
hPDE4B1, indicating that a nitrogen in this ring is not necessary
for selectivity. Going from a pyrrolidin-3-one to a (R)-pyrrolidin-3-ol substituent (21) increased the TbrPDEB1
selectivity to 9-fold and led to a more potent TbrPDEB1 inhibitor.
Interestingly, its isopropyl-substituted counterpart (22) was even less potent at hPDE4B1 and displayed a 15-fold selectivity
for TbrPDEB1. Finally, we explored two (R)-pyrrolidine-2-carboxamide
analogs, 23 and 24, and both were more than
10-fold selective for TbrPDEB1, but overall less potent.
In
order to investigate how these heterocycles interact with the
P-pocket in TbrPDEB1, we determined the crystal structure of 22 (Figure ) bound to pan class="Chemical">TbrPDEB1 at a resolution of 1.8 Å. Only the (4aS,8aR)-enantiomer was observed in the crystal
structure. The (R)-3-hydroxypyrrolidine tail group
targeted the P-pocket through direct hydrogen bonds to the backbone
carbonyl of Met861S.40 and to the side chain of Asn867Q2.43. A water molecule was also observed making a polar interaction
with the hydroxyl substituent. Furthermore, the carbonyl oxygen of
the tail was within hydrogen bonding distance of a water molecule
coordinated by Glu869Q2.45, Lys872Q2.48, and
Gly873Q2.49. It should be noted that in this crystal structure, 22 was only present at the chain with a symmetry-related molecule
close to the ligand binding site. However, the pose of 22 was in line with the overall binding modes observed for the other
TbrPDEB1 inhibitors in this series.
Figure 7
X-ray crystal structure of TbrPDEB1–22.
X-ray crystal structure of pan class="Chemical">TbrPDEB1–22.
We have previously shown
that inhibition of TbrPDEB1 and TbrPDEB2
by 1 in T. b. brucei leads to a rapid
and dose-dependent increase in intracellular cAMP levels.[10,11,22] Incubation of bloodstream form
trypanosomes with 8 increased intracellular cAMP levels
in these parasites in a dose-dependent fashion (Figure ). Compound 8 showed a small
but significant effect at 100 nM (P < 0.05) and
a strong increase of cAMP level at 10 μM (P < 0.001), a concentration that was equally effective as 0.3 μM
of 1, used as positive control. The effect on the intracellular
cAMP levels of the antitrypanosomal compound pentamidine, used as
a negative control, was not statistically different from untreated
parasites.
Figure 8
Intracellular cAMP levels in T. brucei after treatment
with 8 (NPD-008), with 0.3 μM of compound 1 (NPD-001) used as a positive control, and pentamidine as
a negative control.
Intracellular cAMP levels in pan class="Species">T. brucei after treatment
with 8 (NPD-008), with 0.3 μM of compound 1 (NPD-001) used as a positive control, and pentamidine as
a negative control.
Having identified and
characterized 8 as a selective
TbrPDEB1 inhibitor, we initiated further phenotypic profiling. It
has been reported that inhibition of pan class="Chemical">cAMP metabolism causes severe
cellular defects in T. brucei, particularly the ability
of the cells to complete cell division, as cytokinesis and abscission
are impaired, while nuclear and kinetoplast division remain apparently
unaffected.[6,8,10] This leads
to misshapen and nonviable cells with multiple nuclei and kinetoplasts.
Incubation with 8 at a concentration of 10 μM had
the same effect: after incubation of healthy trypanosomes (Figure a) for 6 h (roughly
equivalent to one doubling time under standard culture conditions)
with 10 μM compound 8, cells could be seen to have
undergone at least one round of nuclear and kinetoplast division but
without completing cell division into daughter cells (Figure b), indicating a defect in
cytokinesis. At 12 h, this had resulted in rounded cells with multiple
flagella, some of them detached, and at least 4 nuclei (Figure c). At 24 h, only a few live
cells could be detected, all of which were severely misshapen (Figure d). These results
were consistent with the observed antitrypanosomal activity of 8 (IC50 = 5.5 ± 3.3 μM). The isopropyl
analog 9 displayed similar anti-T. brucei effects (IC50 = 6.7 ± 1.6 μM) in the same
cytokinesis assay. While exhibiting a potency in the sub-micromolar
range against TbrPDEB1, the observed lower antitrypanosomal activity
might be due to reduced cell penetration properties. At these concentrations,
off-target effects can also not be ruled out. However, the cytotoxic
effects of 8 and 9 against humanMRC-5 cells
were comparable for both compounds (CC50 = 36 and 35 μM,
respectively) and clearly lower than the antiparasitic effects. Yet,
because of the relatively narrow selectivity, these compounds do not
meet the published criteria of drug candidates for the treatment of
human African trypanosomiasis,[23] and their
trypanocidal activity, in particular, requires further phenotypic
optimization.
Figure 9
Fluorescence microscopy of wild-type T. brucei incubated with 8 (10 μM) for the following periods:
(a) 0 h (control); (b) 6 h; (c) 12 h; (d) 24 h. Left panels are brightfield
images, middle panels are fluorescence images after DAPI staining,
and the right-hand panels are the merged images. Scale bars, 5 μm.
Fluorescence microscopy of wild-type pan class="Species">T. brucei incubated with 8 (10 μM) for the following periods:
(a) 0 h (control); (b) 6 h; (c) 12 h; (d) 24 h. Left panels are brightfield
images, middle panels are fluorescence images after pan class="Chemical">DAPI staining,
and the right-hand panels are the merged images. Scale bars, 5 μm.
Chemistry
The chemical synthesis
of compounds 3–6 and intermediate 28 proceeded as shown in Scheme . The key pan class="Chemical">4a,5,8,8a-tetrahydrophthalazin-1(2H)-one building block (25)[24−26] was N-alkylated with the appropriate pan class="Chemical">alkyl halides, and a Suzuki
cross-coupling reaction afforded the biaryl systems.
Reagents
and conditions: (a)
alkyl halide, pan class="Chemical">NaH, DMF, RT, 4 h, 65–85%; (b) arylboronic acid,
Pd(dppf)Cl2·CH2Cl2, Na2CO3, DME, H2O, 100 °C, 16 h, 48–81%.
Compounds 7–10, 14–18, 21–24,
and 29 were prepared using carbodiimide-mediated pan class="Chemical">amide
coupling methods using 4 or 28 as starting
material (Scheme ).
Amide coupling was performed without the addition of DIPEA in the
synthesis of 15, 21, 23, and 24 or without Et3N in the case of analog 22.
Scheme 2
Reagents and conditions: (a)
amine, EDC·HCl, HOAt, DIPEA, CH2Cl2, RT,
3 h, 38–88% (compounds 7, 10, 14, 15, 17, 18, 21, 23, 24, and 29),
or amine, EDC·HCl, HOBt hydrate, Et3N, CH2Cl2, RT, 18 h, 54–65% (compounds 8, 9, 16, and 22), or separation
of the enantiomers of 4 by preparative chiral HPLC, then 4a or 4b, 2-aminoacetamide·HCl, EDC·HCl,
HOAt, DIPEA, CH2Cl2, RT, 24 h, 50–55%
(8a, 8b).
Reagents and conditions: (a)
amine, pan class="Chemical">EDC·HCl, HOAt, DIPEA, CH2Cl2, RT,
3 h, 38–88% (compounds 7, 10, 14, 15, 17, 18, 21, 23, 24, and 29),
or amine, EDC·HCl, HOBt hydrate, Et3N, CH2Cl2, RT, 18 h, 54–65% (compounds 8, 9, 16, and 22), or separation
of the enantiomers of 4 by preparative chiral HPLC, then 4a or 4b, 2-aminoacetamide·HCl, EDC·HCl,
HOAt, DIPEA, CH2Cl2, RT, 24 h, 50–55%
(8a, 8b).
To enable
access to the enantiomers of 8, racemic 4 was separated into its enantiomers 4a and 4b using prepan class="Chemical">parative chiral HPLC. Each of the enantiomers
of 4 was reacted with 2-aminoacetamide, as depicted in Scheme . Enantiomer 8a was obtained with 99.6% ee, while its optical isomer 8b was obtained with 96.6% ee. The X-ray crystal structures
indicate that TbrPDEB1 exhibits chiral discrimination, allowing the
tentative assignment of the absolute configuration of (4aR,8aS) to 8b, as this enantiomer possesses
a 100-fold higher potency than 8a.
The synthesis
of 11–13 is outlined
in Scheme . Methyl
ester 29 was treated with aqueous base to give intermediate 30, and subsequent carbodiimide-mediated pan class="Chemical">amide coupling yielded 11–13. Compound 13 was prepared
without the addition of DIPEA.
Reagents and conditions:
(a)
NaOH, pan class="Chemical">EtOH, RT, 2 h, 89%; (b) amine, EDC·HCl, HOAt, DIPEA, CH2Cl2, RT, 36 h, 29–54%.
Scheme depicts
the synthesis of 19 and 20. Analog 19 was prepared by the N-alkylation of 18 using pan class="Chemical">iodomethane. Finally, compound 20 was
attained by the oxidation of 21 using Dess–Martin
periodinane.
Reagents and conditions: (a)
iodomethane, pan class="Chemical">NaH, DMF, RT, 1.5 h, 54%; (b) Dess–Martin periodinane,
CH2Cl2, RT, 4 h, 58%.
Conclusion
In summary, we have identified the first series of selective TbrPDEB1
inhibitors reported to date and explain their specificity for pan class="Chemical">TbrPDEB1
over hPDE4 isoforms on the basis of the cocrystal structures obtained.
We have also explained why previously reported inhibitors failed to
show any selectivity. Our novel trypanocidal TbrPDEB1 inhibitors feature
a tail group containing a rigid biphenyl system with polar substituents
for targeting the P-pocket. Structural data currently available on
the PDEs of L. major and T. cruzi indicate the presence of similar P-pockets in those enzymes,[18,19] while humanPDEs do not have such a pocket.[21] Therefore, this work identifies important possibilities for the
development of parasite-selective PDE inhibitors for a variety of
neglected tropical diseases.
Experimental Section
Chemistry
All reagents and solvents were obtained from
commercial suppliers and were used as received. All reactions were
magnetically stirred and carried out under an inert atmosphere. Reaction
progress was monitored using thin-layer chromatography (TLC) and LC-MS
analysis. Silica gel column chromatography was carried out manually
or with automatic purification systems using the indicated eluent.
Nuclear magnetic resonance (NMR) spectra were recorded on Bruker Avance
500 (500 MHz for 1H and 126 MHz for pan class="Chemical">13C) or
Bruker Avance 600 (600 MHz for 1H and 151 MHz for 13C) instruments equipped with a Bruker CryoPlatform, or a
Bruker DMX300 (300 MHz). Chemical shifts (δ in ppm) and coupling
constants (J in Hz) are reported with residual solvent
as internal standard (δ 1H NMR, CDCl3 7.26,
DMSO-d6 2.50; δ 13C NMR,
CDCl3 77.16, DMSO-d6 39.52).
Various compounds exhibited rotamers leading to more complicated 1H NMR spectra and less accurate integrations. LC-MS analysis
was performed on a Shimadzu LC-20AD liquid chromatograph pump system,
equipped with an Xbridge (C18) 5 μm column (50 mm, 4.6 mm),
connected to a Shimadzu SPD-M20A diode array detector, and MS detection
using a Shimadzu LC-MS-2010EV mass spectrometer. The LC-MS conditions
were as follows: solvent B (acetonitrile with 0.1% formic acid) and
solvent A (water with 0.1% formic acid), flow rate of 1.0 mL/min,
start 5% B, linear gradient to 90% B in 4.5 min, then 1.5 min at 90%
B, then linear gradient to 5% B in 0.5 min, then 1.5 min at 5% B;
total run time of 8 min. Analytical chiral HPLC was performed with
a Chiralpak AD-H column (250 mm × 4.6 mm, 5 μm) with the
following conditions: flow, 1 mL/min; column temperature, 35 °C;
detection, 270 nm; eluent, heptane/isopropanol 9:1; runtime, 30 min.
Exact mass measurement (HRMS) was performed on a Bruker micrOTOF-Q
instrument with electrospray ionization (ESI) in positive ion mode
and a capillary potential of 4500 V. Systematic names for molecules
were generated with ChemBioDraw Ultra 14.0.0.117 (PerkinElmer, Inc.).
The reported yields refer to isolated pure products; yields were not
optimized. The purity, reported as the peak area % at 254 nm, of all
final compounds was ≥95% based on LC-MS.
Synthetic Procedures
Building pan class="Gene">block 25,[24−26] pan class="Gene">NPD-001 (1),[11] and NPD-340
(2)[12] were prepared as described
elsewhere. All other compounds were prepared as described below.
EDC·pan class="Chemical">HCl (183 mg, 0.952
mmol) was added to a solution of 4 (300 mg, 0.635 mmol),
HOBt hydrate (97 mg, 0.63 mmol), 2-aminoacetamide·HCl
(105 mg, 0.952 mmol), and Et3N (0.265 mL, 1.90 mmol) in
CH2Cl2 (6 mL). The reaction mixture was stirred
at RT for 18 h. EtOAc (100 mL) was added, and the resulting solution
was washed with water (2 × 50 mL) and brine (50 mL). The organic
phase was dried over Na2SO4, filtered and concentrated
to obtain the crude product as a light brown oil. The crude oil was
purified on a silica gel column eluting with CH2Cl2/MeOH (gradient, 100:0 to 95:5), to give 8 as
a white solid in 65% yield. 1H NMR (500 MHz, CDCl3) δ 7.91 (d, J = 8.0 Hz, 2H), 7.81 (dd, J = 8.7, 2.3 Hz, 1H), 7.75 (d, J = 2.4
Hz, 1H), 7.62 (d, J = 8.1 Hz, 2H), 7.29–7.24
(m, 1H), 7.03 (d, J = 8.7 Hz, 1H), 6.54 (s, 1H),
5.82–5.61 (m, 3H), 4.85–4.74 (m, 1H), 4.23 (d, J = 4.7 Hz, 2H), 3.85 (s, 3H), 3.35–3.27 (m, 1H),
3.04–2.94 (m, 1H), 2.73 (t, J = 6.0 Hz, 1H),
2.24–2.12 (m, 2H), 2.10–1.94 (m, 2H), 1.93–1.83
(m, 1H), 1.82–1.68 (m, 4H), 1.67–1.42 (m, 6H). 13C NMR (126 MHz, CDCl3) δ 171.2, 167.6, 165.9,
157.6, 153.4, 141.9, 132.0, 129.8, 129.7, 128.2, 128.1, 127.1, 127.0,
126.1, 123.9, 111.3, 56.3, 55.8, 43.3, 34.7, 33.2, 33.0, 31.9, 31.1,
28.43, 28.36, 25.1, 25.0, 23.1, 22.4. LC-MS (ESI): tR = 4.69 min, area: >98%, m/z 529 [M + H]+. HRMS (ESI) m/z: [M + H]+ calcd for C31H37N4O4 529.2809, found 529.2811.
N-(2-Amino-2-oxoethyl)-5′-((4aS,8aR)-3-cycloheptyl-4-oxo-3,4,4a,5,8,8a-hexahydrophthalazin-1-yl)-2′-methoxy-[1,1′-biphenyl]-4-carboxamide
(8a, NPD-949) and N-(2-Amino-2-oxoethyl)-5′-((4aR,8aS)-3-cycloheptyl-4-oxo-3,4,4a,5,8,8a-hexahydrophthalazin-1-yl)-2′-methoxy-[1,1′-biphenyl]-4-carboxamide
(8b, NPD-1373)
The enantiomers of racemic 4 were separated into 4a (first eluting isomer) and 4b (second eluting
isomer) using prepan class="Chemical">parative chiral chromatography. Analytical chiral
HPLC indicated 4a and 4b to have the same
retention times as their respective counterparts in racemic mixture 4. The title compounds were synthesized from separate solutions
of enantiomerically pure 4 (124 mg, 0.262 mmol) and 2-aminoacetamideHCl (34.8 mg, 0.315 mmol) as described for 7, and were
isolated in 50–55% yield. Absolute configuration was tentatively
assigned to enantiomer 8b. Compound 8a,
Chiralpak AD-H 99.6% ee; compound 8b, Chiralpak AD-H tR = 19.71 min, 96.9% ee.
Aqueous pan class="Chemical">sodium hydroxide (6.2 mL, 25 mmol)
was added to a suspension of 29 (0.84 g, 1.5 mmol) in
pan class="Chemical">EtOH (25 mL) and stirred at RT for 2 h. The reaction mixture was acidified
with HCl (2 M) and extracted with EtOAc (2 × 30 mL). The combined
organic phases were washed with water, dried over Na2SO4, filtered, and concentrated to give the title compound as
a brown solid in 89% yield. 1H NMR (300 MHz, DMSO-d6) δ 8.87 (t, J = 5.8
Hz, 1H), 7.95 (d, J = 8.5, 2H), 7.91 (dd, J = 8.9, J = 2.2 Hz, 1H) 7.79 (d, J = 2.3 Hz, 1H), 7.64 (d, J = 8.5 Hz, 2H),
7.26 (d, J = 8.8 Hz, 1H), 5.73–5.61 (m, 2H),
4.74–4.65 (m, 1H), 3.97 (d, J = 5.8 Hz, 2H),
3.85 (s, 3H), 3.55–3.47 (m, 1H), 2.81–2.72 (m, 2H),
2.20–2.11 (m, 2H), 2.00–1.36 (m, 13H). LC-MS (ESI): tR = 4.93 min, m/z 530 [M + H]+
Interference Compounds
All final
compounds have been
expan class="Chemical">amined for the presence of substructures classified as Pan Assay
Interference Compounds (pan class="Disease">PAINS) using a KNIME workflow.[27]
Phosphodiesterase Activity Assay
To determine the effect
of test compounds on the enzymatic activity of full length TbrPDEB1
(Km = 7.97 ± 2.32 μM) and full
length recombinant pan class="Gene">hPDE4B1 (Km = 2.0 ±
0.7 μM), the standard scintillation proximity assay (SPA) was
used, as reported previously.[10,12] In this assay, the
cAMP substrate concentration was 0.5 μM, and the enzyme concentration
was adjusted so that <20% of substrate was consumed. The Ki values are represented as the mean of at least
two independent experiments with the associated standard deviation
(SD) as indicated.
Gene Constructs for Structural Studies
TbrPDEB1
Catalytic Domain
A gene segment coding for
TbrPDEB1 catalytic domain residues 565–918 (Uniprot entry Q8WQX9) was PCR
amplified with flanking NdeI and EcoRI restriction sites and cloned into a pET28a(+) expression vector
(Novagen) previously digested with the same set of restriction enzypan class="Gene">mes.
An N-terminal, thrombin cleavable 6×His tag was kept in-frame
with the gene to facilitate subsequent purification of the expressed
protein by metal affinity chromatography. The resultant recombinant
vector was named pET28a(+)-TbrPDEB1_CD.
Coding sequence for pan class="Gene">hPDE4B UCR2
and catalytic domain residues 241–659
(Uniprot entry Q07343) was synthesized and cloned into a pMA vector (GeneArt, Invitrogen,
Life Technologies). This vector was then used as a PCR template for
subcloning of the gene segment into a pFastBacHTA insect cell expression
vector (Invitrogen, Life Technologies) with a C-terminal 6×His
purification tag.
hPDE4D2 Catalytic Domain
A gene
segment coding for
residues 381–740 of hPDE4D2 (Uniprot entry Q08499) was PCR
amplified using a forward primer including an NdeI restriction site and a reverse primer including a XhoI restriction site. The PCR product was cloned into a pET15b pan class="Species">E. coli expression vector (Novagen) previously digested
with the same set of enzymes. The primer design was such to keep an
N-terminal 6×His tag from the vector in frame with the target
gene. The resultant recombinant vector was named pET15b-hPDE4D_CD.
Protein Expression and Purification for Structural Studies
pan class="Species">Escherichia coli pan class="CellLine">BL21 (DE3)
cells were transformed with pET28a(+)-TbrPDEB1_CD and
allowed to grow in 2 L of 2×YT medium at 37 °C until the
optical density at 600 nm reached 0.6–0.8. At this stage, the
culture was cooled down, induced with 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and further grown overnight at 16
°C. Cells were collected by centrifugation, resuspended in a
buffer containing 20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 10 mM imidazole,
5% glycerol, 2 mM β-mercaptoethanol (BME), and protease inhibitor
cocktail tablet (Roche), and lysed by passing through a cell disruptor
(20 kpsi/pass). Cleared cell lysate was loaded onto a 5 mL HisTrap
HP nickel affinity column (GE Healthcare Biosciences) and bound protein
was eluted with a linear gradient of 0–1 M imidazole. Fractions
containing the target protein, as assessed by SDS gel electrophoresis,
were pooled and desalted to remove imidazole. Removal of N-terminal
6×His tag was performed by overnight incubation at 4 °C
of the sample with humanthrombin (Abcam) at 5 NIH units/mL of the
pooled sample followed by a second nickel affinity purification step
to remove any remaining tagged fraction. The protein was then dialyzed
against ion exchange buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl,
5% glycerol, 2 mM BME), loaded onto a HiTrap Q HP column (GE Healthcare
Biosciences), and eluted with a linear gradient of 0–1 M NaCl.
A final size exclusion chromatography step was performed on the collected
sample using a Superdex 200 increase 10/300 GL column (GE Healthcare
Biosciences) pre-equilibrated with 20 mM Tris-HCl, pH 7.5, 50 mM NaCl,
5% glycerol, and 2 mM BME after which the protein was concentrated
using Amicon Ultra concentrators (Millipore) to 7 mg/mL and stored
at −80 °C prior to use. For crystallization trials, NaCl
was removed from the buffer while keeping other components intact.
hPDE4B
regulatory domain expression and purification was performed
according to the method described.[28] Recombinant
baculovirus generation, insect cell (Sf21) culture, and pan class="Disease">infection
were carried out following manufacturer’s instructions (Bac-to-Bac
expression system; Invitrogen, Life technologies). Infected cells
were grown for 48 h at 28 °C and harvested by centrifugation
at 500g for 15 min. Lysis was performed by resuspending
the cells in a hypotonic buffer of 10 mM HEPES, pH 7.5, 50 mM NaCl,
and 1 mM tris(2-carboxyethyl) phosphine (TCEP) followed by centrifugation
to remove cell debris. The obtained cleared lysate was loaded onto
a 5 mL HisTrap HP nickel affinity column (GE Healthcare Biosciences)
pre-equilibrated with 100 mM HEPES, pH 7.5, 150 mM NaCl, 50 mM arginine,
10 mM imidazole, and 1 mM TCEP and eluted with a linear gradient of
0–1 M imidazole. Eluted protein was then dialyzed against ion-exchange
buffer (100 mM HEPES, 50 mM NaCl, 1 mM dithiothreitol (DTT)) and loaded
onto a HiTrap Q HP ion exchange column (GE Healthcare Biosciences).
A linear gradient elution of 20–250 mM NaCl was performed followed
by dialysis of the collected sample into size exclusion buffer (10
mM HEPES, pH 7.5, 100 mM NaCl, 1 mM DTT). Protein was then passed
through a Superdex 200 increase 10/300 GL size exclusion column (GE
Healthcare Biosciences). The collected sample was concentrated to
10 mg/mL and stored at −80 °C prior to use in crystallization.
BL21 (DE3) Codon Plus cells
were transformed with pET15b-pan class="Gene">hPDE4D_CD and allowed to grow in 1 L
of 2×YT medium at 37 °C until the optical density at 600
nm reached 0.6–0.8. At this stage, culture temperature was
lowered, and expression was induced by addition of 0.5 mM IPTG followed
by further overnight growth at 22 °C. The same cell disruption
and protein purification procedures were followed as performed in
the case of TbrPDEB1 catalytic domain protein with the following changes
in the buffers used: cell resuspension and nickel affinity chromatography
buffer (50 mM Tris-HCl, pH 8, 150 mM NaCl, 5 mM BME, 20 mM imidazole),
ion exchange chromatography buffer (50 mM Tris-HCl, pH 8, 50 mM NaCl,
5 mM DTT), and size exclusion chromatography buffer (50 mM Bis-tris,
pH 6.8, 100 mM NaCl, 5 mM DTT). Purified protein was concentrated
to 9 mg/mL and stored at −80 °C prior to use in crystallization.
Protein Crystallization, Ligand Soaking and Data Collection
All crystallization trials were performed by vapor diffusion hanging
drop technique, typically with 500 μL reservoir volume and 2
μL droplets with a protein to crystallization solution ratio
of 1:1. Crystals of the TbrPDEB1 catalytic domain were grown in 20%
pan class="Chemical">PEG 3350, 400 mM sodium formate, 300 mM guanidine, and 100 mM MES,
pH 6.5, at 4 °C. Soaking of these crystals with various compounds
was performed maintaining a final compound concentration of 5–15
mM and for varying duration of overnight to 48 h. Soaked crystals
were then briefly dipped in growth solution supplemented with 20%
(v/v) glycerol or ethylene glycol and were mounted using CryoLoop
(Hampton Research) or LithoLoops (Molecular Dimensions) and vitrified
in liquid nitrogen for data collection. In the case of hPDE4D2, thick
plate-like crystals appeared within 5–6 days in a condition
containing 24% PEG 3350, 30% ethylene glycol, and 100 mM HEPES, pH
7.5, at 19 °C. For hPDE4B, a condition containing 20% PEG 400,
50 mM calcium acetate, and 100 mM sodium acetate, pH 4.6, resulted
in well-diffracting crystals. Compound soaking and crystal harvesting
were performed in the same way as for TbrPDEB1 crystals except that
in case of hPDE4D2 no cryopreservative was added prior to vitrification.
X-ray diffraction data sets were collected at Diamond Light Source
(DLS; Didcot, Oxfordshire, UK) beamlines I03 and I04 at 100 K. The
data sets were processed by xia2[29] or autoPROC,[30] which incorporates XDS[31] and AIMLESS,[32] or were integrated using
iMOSFLM[33] and reduced using POINTLESS,
SCALA, and TRUNCATE,[34] all of which are
part of CCP4.[35]
X-ray Crystal Structure
Determination, Refinement, and Analysis
The crystal structure
of TbrPDEB1 bound to pan class="Gene">NPD-008 (8) was solved by molecular
replacement (MR) using PHASER,[36] taking
the apo structure (PDB code 4I15) as search model.
Reflections for calculating Rfree were
selected randomly, and the same set was used in all other ligand-bound
TbrPDEB1 data sets except for NPD-038 (22) where the
crystal was nonisomorphous. All isomorphous crystal forms were solved
by Fourier synthesis using the partially refined, ligand-free 8 model, whereas MR was applied for the 22 data
set. hPDE4B and hPDE4D2 structures with NPD-001 (1) were
solved by MR using their respective apo models (PDB codes 3G45 and 3SL3, respectively).
Ligand descriptions were generated by ACEDRG available within the
CCP4 package[35] or with the grade Web Server
(http://grade.globalphasing.org/). Adjustment of the models and ligand fitting were performed with
COOT[37] and refinement with REFMAC5.[38] The final structures had good geometry and could
be refined to low R-factors (Tables S1 and S3). All refined models were validated with
MOLPROBITY.[39] Data collection and refinement
statistics are given in Supplementary Tables S1–S4. Root-mean-square (rms) deviation values were calculated from a
sequence alignment, structural superposition, and refinement cycle
on Cα carbons with the align function as implemented in PyMOL
1.7.4.4 (The PyMOL Molecular Graphics System, Schrödinger,
LLC). All binding site residues have been named according to the PDEStrIAn
nomenclature system (http://pdestrian.vu-compmedchem.nl/).[21] Structural figures were prepared with PyMOL 1.7.4.4. For clarity,
selected residues from the helix capping the substrate binding pocket
(i.e., D784, M785, A786, K787, H788, G789, S790, A791, L792, and E793
in TbrPDEB1; D518, M519, S520, K521, H522, M523, S524, and L525 in
hPDE4B; D272, M273, S274, K275, H276, M277, N278, and L279 in hPDE4D)
have been omitted in the rendering of the figures. Coordinates of
the structures have been deposited to the RCSB Protein Data Bank with
following accession codes: 5G57 (TbrPDEB1–NPD-001); 5LAQ (hPDE4B–NPD-001); 5LBO (hPDE4D–NPD-001); 5L9H (TbrPDEB1–NPD-340); 5G2B (TbrPDEB1–NPD-008); 5L8C (TbrPDEB1–NPD-039); 5G5V (TbrPDEB1–NPD-038); 5L8Y (TbrPDEB1–NPD-937).
Molecular Dynamics
Molecular dynamics simulations of
the TbrPDEB1–pan class="Gene">NPD-008 (8) crystal structure (PDB
code 5G2B) were
performed with GROMACS 5.1,[40,41] using the Amber ff99SB-ILDN
force field[42] and a TIP3P water model,[43] running with MPI parallelization. Ligand input
files, based on 8 extracted from chain A, were prepared
with Antechamber as implemented in AmberTools14 (AMBER 14, University
of California, San Francisco, 2014) by the calculation of AM1-BCC
partial charges, and generation of a GAFF topology. Subsequent preparation
of the parameters and ligand topology was performed with LEaP. ACPYPE[44] was used to convert the ligand files to GROMACS
input format. The protein–ligand complex of 8 and
chain A from TbrPDEB1, together with the Mg2+ and Zn2+ ions and the crystal water network directly coordinating
the binding site cations consisting of six crystal waters, was centered
at least 1.0 nm from the edges of a cubic box, that was fully solvated
and charge neutralized by the addition of Na+ ions. The
system was energy minimized with a steepest descent algorithm for
5000 steps with an initial step size of 0.01 nm. Bonds were constrained
with the P-LINCS algorithm,[45] an integration
time step of 2 fs was used, and the smooth particle mesh Ewald (PME)
method was used for the calculation of long-range electrostatics.
The Verlet cutoff scheme was used for neighbor searching; cutoff distances
of 1.0 nm were set for the short-range neighbor list as well as for
Coulomb and van der Waals interactions. Initial equilibration was
performed in three 250 ps steps by gradually increasing the temperature
from 100 to 298 K while decreasing the position restraint force constant
from 1000 kJ mol–1 nm–1 to 500
kJ mol–1 nm–1 and finally to 50
kJ mol–1 nm–1. In the canonical
(NVT) ensemble, a velocity rescaling thermostat was
used. The Parrinello–Rahman barostat and a position restraint
force constant of 25 kJ mol–1 nm–1 was used in the isothermal–isobaric (NPT) ensemble equilibration for 250 ps at 298 K and 1.0 bar. A final
unrestrained equilibration step was executed for 500 ps at 298 K and
1.0 bar, followed by a production run of 100 ns. The first 10 ns of
the simulation was omitted from analysis. The stability of the system
and the results of the simulations were validated and analyzed with
gmx rms, energy, gyrate, rmsf, distance, and hbond tools as implemented
in GROMACS 5.1.
Parasite Culturing for cAMP Measurements
and Microscopy
Bloodstream forms of pan class="Species">T. brucei Lister 427 were cultured
in pan class="Chemical">Hirumi-9 (HMI-9) medium (Invitrogen), supplemented with 10% heat
inactivated fetal bovine serum (FBS; Gibco) in vented culture flasks
(Corning), at 37 °C, in a 5% CO2 atmosphere, as described
previously.[46]
Intracellular cAMP Measurements
Intracellular cAMP
was measured as described previously,[10] with minor changes. Briefly, log-phase bloodstream form tryne">panosopan class="Gene">mes
were inoculated into HMI-9/FBS media and incubated at 37 °C overnight.
The cells were counted and suspended in HMI-9/FBS at 2 × 106 cells/mL, which was divided into 8 subcultures of 6 mL each.
To each culture flask, a small volume of HMI-9 medium containing either
NPD-001 (1) (positive control; final concentration 0.3
μM), pentamidine (a known trypanocide not acting on cAMP signaling;
final concentration 0.025 μM), NPD-008 (8) (at
0.1, 0.33, 1, 3.3, and 10 μM), or no drug (negative control)
was added. The cultures were incubated under standard conditions (37
°C, 5% CO2) for 5 h, after which cell densities were
determined in each culture using a hemocytometer; 5 × 106 cells were transferred into new tubes and collected by centrifugation
at 1500g for 10 min at 4 °C. Cell pellets were
resuspended in 100 μL of 0.1 M HCl and left on ice for 20 min
to complete cell lysis. The samples were centrifuged in a microfuge
at 12 000g for 10 min at 4 °C, and the
supernatants (cell extracts) were stored at −80 °C for
the determination of intracellular cAMP. Each cAMP determination was
performed in duplicate, data are represented as the mean of four independent
cAMP determinations with the standard error of the mean (SEM) and
were analyzed with Student’s t-test, differences
were considered significant at P < 0.05, with P values as indicated.
Microscopy and DAPI Staining
For the monitoring of
the cell cycle (division of nucleus, kinetoplast, and cells) and cellular
morphology, trypanosomes were stained with 4′,6-diamidino-2-phenylindole
(pan class="Chemical">DAPI) and observed by fluorescence microscopy as described.[47] Briefly, a culture of T. brucei bloodstream forms was inoculated at 2 × 105 cell/mL
in the presence of 10 μM NPD-008 (8). Samples for
microscopy were taken at 0, 6, 12, and 24 h and centrifuged at 2600
rpm for 10 min at 4 °C in a Heraeus Biofuge centrifuge. Supernatant
was decanted, and the pellet was washed with phosphate-buffered saline
(PBS), pH 7.4, resuspended in 20 μL of PBS and spread on a microscope
slide. The slides were air-dried and then fixed in 4% formaldehyde/PBS
for 15 min. Slides were rinsed 3 times with PBS after which the samples
were mounted in VectaShield (Vector Laboratories Inc., USA) mounting
medium with DAPI. Samples were imaged on an Axioskop II microscope
(Zeiss, Inc.) and a DeltaVision Core (AppliedPrecision).
Phenotypic
Cellular Assays
For the cellular assays,
the following reference drugs were used as positive controls: suramin
(Sigma-Aldrich, Germany) for pan class="Species">T. brucei (pIC50 = 7.4 ± 0.2, n = 5), and tamoxifen (Sigma-Aldrich,
Germany) for MRC-5 cells (pIC50 = 5.0 ± 0.1, n = 5). All compounds were tested at five concentrations
(64, 16, 4, 1, and 0.25 μM) to establish a full dose-titration
and determination of the IC50 and CC50; data
are represented as the mean of triplicate experiments ± SD. The
final concentration of DMSO did not exceed 0.5% in the assays.
Antitrypanosomal
Cellular Assay with T. brucei
Squib-427
strain (suramin-sensitive) was cultured at 37
°C and 5% pan class="Chemical">CO2 in HMI-9 medium, supplemented with 10%
fetal calf serum (FCS). About 1.5 × 104 trypomastigotes
were added to each well, and parasite growth was assessed after 72
h at 37 °C by adding resazurin. The color reaction was read at
540 nm after 4 h, and absorbance values were expressed as a percentage
of the blank controls.
MRC-5 Cytotoxicity Cellular Assay
MRC-5 SV2 cells,
originally from a pan class="Species">human diploid lung cell line, were cultivated in
MEM, supplemented with l-glutamine (20 mM), 16.5 mM sodium
hydrogencarbonate, and 5% FCS. For the assay, 104 MRC-5
cells/well were seeded onto the test plates containing the prediluted
sample and incubated at 37 °C and 5% CO2 for 72 h.
Cell viability was assessed fluorimetrically 4 h after the addition
of resazurin. Fluorescence was measured (excitation 550 nm, emission
590 nm), and the results were expressed as percentage reduction in
cell viability compared to control.
Statistical Analysis
Details of the applied statistical
analyses are provided with each experiment. No statistical methods
were used to predetermine the size of samples. The experiments were
not randomized, and the investigators were not blinded to allocation
during experiments or outcome assessment.
Authors: Nicholas D Bland; Cuihua Wang; Craig Tallman; Alden E Gustafson; Zhouxi Wang; Trent D Ashton; Stefan O Ochiana; Gregory McAllister; Kristina Cotter; Anna P Fang; Lara Gechijian; Norman Garceau; Rajiv Gangurde; Ron Ortenberg; Mary Jo Ondrechen; Robert K Campbell; Michael P Pollastri Journal: J Med Chem Date: 2011-11-08 Impact factor: 7.446
Authors: Sander Pronk; Szilárd Páll; Roland Schulz; Per Larsson; Pär Bjelkmar; Rossen Apostolov; Michael R Shirts; Jeremy C Smith; Peter M Kasson; David van der Spoel; Berk Hess; Erik Lindahl Journal: Bioinformatics Date: 2013-02-13 Impact factor: 6.937
Authors: Harry P de Koning; Matthew K Gould; Geert Jan Sterk; Hermann Tenor; Stefan Kunz; Edith Luginbuehl; Thomas Seebeck Journal: J Infect Dis Date: 2012-01-30 Impact factor: 5.226
Authors: Garib N Murshudov; Pavol Skubák; Andrey A Lebedev; Navraj S Pannu; Roberto A Steiner; Robert A Nicholls; Martyn D Winn; Fei Long; Alexei A Vagin Journal: Acta Crystallogr D Biol Crystallogr Date: 2011-03-18
Authors: Martyn D Winn; Charles C Ballard; Kevin D Cowtan; Eleanor J Dodson; Paul Emsley; Phil R Evans; Ronan M Keegan; Eugene B Krissinel; Andrew G W Leslie; Airlie McCoy; Stuart J McNicholas; Garib N Murshudov; Navraj S Pannu; Elizabeth A Potterton; Harold R Powell; Randy J Read; Alexei Vagin; Keith S Wilson Journal: Acta Crystallogr D Biol Crystallogr Date: 2011-03-18
Authors: Vincent B Chen; W Bryan Arendall; Jeffrey J Headd; Daniel A Keedy; Robert M Immormino; Gary J Kapral; Laura W Murray; Jane S Richardson; David C Richardson Journal: Acta Crystallogr D Biol Crystallogr Date: 2009-12-21
Authors: Julianna Siciliano de Araújo; Raiza Brandão Peres; Patrícia Bernardino da Silva; Marcos Meuser Batista; Geert Jan Sterk; Louis Maes; Guy Caljon; Rob Leurs; Harry P de Koning; Titilola D Kalejaiye; Maria de Nazaré Correia Soeiro Journal: Antimicrob Agents Chemother Date: 2021-02-17 Impact factor: 5.191
Authors: Julianna Siciliano de Araújo; Alfonso García-Rubia; Victor Sebastián-Pérez; Titilola D Kalejaiye; Patrícia Bernardino da Silva; Cristina Rosa Fonseca-Berzal; Louis Maes; Harry P De Koning; Maria de Nazaré Correia Soeiro; Carmen Gil Journal: Antimicrob Agents Chemother Date: 2019-03-27 Impact factor: 5.191
Authors: Lorena Zara; Francesca Moraca; Jacqueline E Van Muijlwijk-Koezen; Barbara Zarzycka; Robert Abel; Iwan J P de Esch Journal: ACS Med Chem Lett Date: 2022-05-23 Impact factor: 4.632
Authors: Erik de Heuvel; Albert J Kooistra; Ewald Edink; Sjors van Klaveren; Jeffrey Stuijt; Tiffany van der Meer; Payman Sadek; Dorien Mabille; Guy Caljon; Louis Maes; Marco Siderius; Iwan J P de Esch; Geert Jan Sterk; Rob Leurs Journal: Front Chem Date: 2021-01-21 Impact factor: 5.221
Authors: Alberto Venturelli; Lorenzo Tagliazucchi; Clara Lima; Federica Venuti; Giulia Malpezzi; George E Magoulas; Nuno Santarem; Theodora Calogeropoulou; Anabela Cordeiro-da-Silva; Maria Paola Costi Journal: Microorganisms Date: 2022-06-27
Authors: Jane C Munday; Stefan Kunz; Titilola D Kalejaiye; Marco Siderius; Susanne Schroeder; Daniel Paape; Ali H Alghamdi; Zainab Abbasi; Sheng Xiang Huang; Anne-Marie Donachie; Samia William; Abdel Nasser Sabra; Geert Jan Sterk; Sanaa S Botros; David G Brown; Charles S Hoffman; Rob Leurs; Harry P de Koning Journal: PLoS Negl Trop Dis Date: 2020-07-30
Authors: Sanaa S Botros; Samia William; Abdel-Nasser A Sabra; Naglaa M El-Lakkany; Sayed H Seif El-Din; Alfonso García-Rubia; Victor Sebastián-Pérez; Antoni R Blaazer; Erik de Heuvel; Maarten Sijm; Yang Zheng; Irene G Salado; Jane C Munday; Louis Maes; Iwan J P de Esch; Geert J Sterk; Koen Augustyns; Rob Leurs; Carmen Gil; Harry P De Koning Journal: Int J Parasitol Drugs Drug Resist Date: 2019-01-14 Impact factor: 4.077
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Scheme 1
Scheme 2
Scheme 3
Scheme 4