Phosphinic Dehydrodipeptides: Diversification of the P1′ Residue with the Morita–Baylis–Hillman Acetates and Inhibition of Alanyl Aminopeptidases

Article abstract of DOI:10.1007/s10989-019-10004-7

Activated allyl acetates (the Morita–Baylis–Hillman acetates) were confirmed as convenient electrophilic precursors of the P1′ fragment of phosphinic dehydrodipeptides, and then, their selected saturated analogs were obtained in the subsequent reduction. This approach is particularly appropriate for structurally complex side chains and thus is an alternative to the addition of α-substituted acrylates with aminoalkylphosphinic acids. While phosphinic dehydrodipeptides were found to be good inhibitors of two mammalian alanyl metalloaminopeptidases, the saturated compounds showed higher activities than their structurally constrained counterparts. Docking calculations and molecular modeling showed that conformational freedom allowed for the favorable binding of distinct stereoisomers of phosphinates, while the presence of the double bond was more restrictive.

Full text of DOI:10.1007/s10989-019-10004-7

International Journal of Peptide Research and Therapeutics
https://doi.org/10.1007/s10989-019-10004-7
SHORT COMMUNICATION
Phosphinic Dehydrodipeptides: Diversifcation of the P1′ Residue
with the Morita–Baylis–Hillman Acetates and Inhibition of Alanyl
Aminopeptidases
Michał Talma1
Received: 4 November 2019 / Revised: 16 December 2019
© Springer Nature B.V. 2020
Abstract
Activated allyl acetates (the Morita–Baylis–Hillman acetates) were confrmed as convenient electrophilic precursors of the
P1′ fragment of phosphinic dehydrodipeptides, and then, their selected saturated analogs were obtained in the subsequent
reduction. This approach is particularly appropriate for structurally complex side chains and thus is an alternative to the
addition of α-substituted acrylates with aminoalkylphosphinic acids. While phosphinic dehydrodipeptides were found to
be good inhibitors of two mammalian alanyl metalloaminopeptidases, the saturated compounds showed higher activities
than their structurally constrained counterparts. Docking calculations and molecular modeling showed that conformational
freedom allowed for the favorable binding of distinct stereoisomers of phosphinates, while the presence of the double bond
was more restrictive.
Keywords Morita–Baylis–Hillman reaction · Nucleophilic substitution · Phosphinic peptide analogs · Enzyme inhibitors ·
Mammalian alanyl aminopeptidase
Introduction
side chains, was shown to provide highly potent dipeptide-
mimicking ligands (Mucha 2012). The most remarkable
examples involve action on zinc-containing aminopepti-
dases as demonstrated by the inhibition of the members of
M01 and M17 exopeptidase families that are multifunctional
broad-band specifcity enzymes involved in biological func-
tions in eukaryotic and prokaryotic cells; their deregulated
activity is associated with human diseases such as cancer,
and they are also etiological factors in protozoal and bacte-
rial infections (Chen et al. 2000; Skinner-Adams et al. 2010;
Suzuki et al. 2004; Vassliou et al. 2014; Wang et al. 2007).
Synthetic preparation of phosphinic dipeptides is achieved
either by a phospha-Michael addition of a silylated N-pro-
tected aminoalkyl-H-phosphinic acid to an α-substituted
acrylate (Thottathil et al. 1984) or by a three-component
amidoalkylation reaction (Chen and Coward 1996). The fun-
damental starting materials for these extensively explored
reactions are aldehydes that are precursors of the P1 residue
(Baylis et al. 1984) and acrylates that code the P1’ fragment
(Boyd et al. 1990; Stetter and Kuhlmann 1979). As the side
chains must ensure tight ligand-protein interactions, they fre-
quently demand modifcations with heteroatom-containing
functional groups. These additional modifcations require
the development of complex multistep synthetic pathways
Phosphinic acid peptide analogs are well-recognized transi-
tion-state analog inhibitors of metal-dependent hydrolases,
particularly the medically signifcant proteases (Collinsová
and Jirácek 2000; Georgiadis and Dive 2015; Mucha et al.
2011; Yiotakis et al. 2004). The competitive and reversible
mechanism of their action involves the coordination of the
tetrahedral phosphinic function with catalytic metal ion(s)
and specifc interactions of pseudopeptide side-chain resi-
dues with the corresponding binding pockets of an enzy-
matic target (Mucha et al. 2011; Yiotakis et al. 2004). The
structural complementarity is of particular importance for
the dipeptidyl analog inhibitors that basically explore the S1
and S1′ cavities that are the clefts closest to the reaction site.
Careful structural refnement of the corresponding P1 and
P1′ substituents, which frequently leads to non-proteinogenic
* Michał Talma
michal.talma@pwr.edu.pl
1
Department of Bioorganic Chemistry, Faculty of Chemistry,
Wrocław University of Science and Technology, Wybrzeże
Wyspiańskiego 27, 50-370 Wrocław, Poland
Vol.:(0123456789)
1 3

International Journal of Peptide Research and Therapeutics
and an orthogonal protection strategy and are particularly
challenging for the diversifcation of α-substituted acrylates
(Mucha 2012).
bond that is conveniently performed in a stereoselective
manner, thus leading to phosphinic dipeptides with the
desired P1′ stereochemistry (Georgiadis et al. 2004; Par-
sons et al. 1988; Yamagishi et al. 2011). Accordingly,
d-Alaψ[P(O)(OH)CH2]-d-Ala (the phosphinic acid analog
of d-Ala-d-Ala dipeptide) was obtained and used to study
the inhibition of the d-Ala-d-Ala ligase, a bacterial enzyme
involved in cell wall biosynthesis (Parsons et al. 1988).
The α,β-conjugated system of P1’-dehydroalanine syn-
thons was also utilized as the Michael acceptors to add
nitrogen, carbon and sulfur nucleophiles (Matziari et al.
2001; Matziari et al. 2004). The reactions proceeded under
mild conditions and without the protection of the phosphi-
nate moiety. The diversifcation potential was implemented
to construct potent and selective inhibitors of matrix
metalloproteases, particularly MMP-11 (Matziari et al.
2004). Inhibition of renal dipeptidase is probably the only
reported example of direct biological application of phos-
phinic P1’-dehydrodipeptides (Gurulingappa et al. 2003;
Gurulingappa et al. 2004; Parsons et al. 1991). The com-
pound bearing hydrophobic side-chain residues showed a
high afnity against the enzyme (IC50 of a low nanomolar
value), with a preference for the Z confguration (stere-
oisomers separated chromatographically) (Gurulingappa
et al. 2004).
Activated allyl acetates (Morita–Baylis–Hillman acetates)
are synthetically feasible but rarely reported precursors of
the P1′ substituents in phosphinic pseudodipeptides. The
allyl acetates are products of acetylation of allyl alcohols
that are universally obtained in the Morita–Baylis–Hillman
reaction of aldehydes with alkyl acrylates (Morita et al.
1958). Such adducts are capable of an SN2′ displacement
reaction with phosphorus nucleophiles (Kalyva et al. 2015;
Talma and Mucha 2017), including silylated N-protected
aminoalkyl-H-phosphinic acid to provide P1′-dehydro-
modifed dipeptide analogs (Scheme 1, pathway A) (Mat-
ziari et al. 2001).
The synthesis and biological activity of phosphinic
dehydrodipeptides are much less recognized than their
saturated counterparts. Regarding the synthesis, in an
alternative to the abovementioned MBH adducts-based
approach (Kalyva et al. 2015; Matziari et al. 2001; Talma
and Mucha 2017), P1′-α,β-unsaturated phosphinic dipep-
tides may be obtained from active methylene compounds.
Additions of trialkyl phosphonoacrylates or dialkyl meth-
ylidenemalonates to H-phosphinic acids were reported to
provide phosphinate-substituted phosphonoacetates or
malonates, respectively, that could subsequently undergo
either Wittig-Horner olefnation (Parsons et al. 1988) or
Knoevenagel-type condensation (Gurulingappa et al. 2003;
Matziari et al. 2006) (Scheme 1, pathway B). The unsatu-
rated products have usually been considered as interme-
diates for further transformations. These transformations
mainly involved the catalytic hydrogenation of the double
In this work, we report the synthesis and activity of phos-
phinic dehydrodipeptides as inhibitors of mammalian alanyl
aminopeptidases. Bulky aromatic residues, including those
substituted with functionalities, were introduced as the rigid
P1’ portion of the molecules. The preparation protocol was
based on the Morita–Baylis–Hillman acetates as versatile
and diversifed precursors. The impact of the P1’ constraints
Scheme 1 General reaction scheme of the synthetic pathways leading to dehydrodipeptides.
1 3

International Journal of Peptide Research and Therapeutics
on the biological activity was assessed in relation to the
100%) → AcOEt/MeOH (0% → 100%) elution system to
yield 4.
potency of the selected saturated analogs.
Protective Group Removal—Representative
Examples 4e and 4h
Materials and Methods
General
The biphenyl derivative of amino- and carboxylate-pro-
tected phosphinic dehydrodipeptide 4e (0.3 g, 0.5 mmol)
was stirred with 33% solution of HBr in acetic acid (5 mL)
for 1–2 h, then, the volatiles were evaporated in vacuo. The
residue was dissolved in MeOH (20 mL) and propylene
oxide (2 mL) was added to quench the remaining HBr. The
following addition of cold diethyl ether (10 mL) precipitated
the crude product. The solid was separated by centrifugation
and solvent decantation. The crude peptide 5e was purifed
by fash chromatography on a C18 reversed phase column
in H2O/acetonitrile (0 → 100%) eluent.
All used reagents, solvents, aminopeptidases (recombinant
human and microsomal porcine kidney APNs, the lyo-
philized forms) and the fuorogenic substrate (l-alanine-
4-methylcoumaryl-7-amide) were purchased from commer-
cial suppliers. Dry toluene was obtained by distillation over
P2O5. Dry THF was obtained by distillation with sodium
with benzophenone. Reactions were monitored by thin-layer
chromatography carried out on 0.25 mm silica gel plates
(silica gel 60F254) and components were visualized by
UV light absorbance. Purifcation of compounds by fash
chromatography was carried out on the CombiFlash® Rf+
Lumen chromatographic system with the use of prepacked
silica or C18 columns. 1H, 13C and 31P NMR spectra were
recorded on Bruker Avance DRX 600 MHz and JEOL JNM-
ECZ 400S Research FT NMR spectrometers. HRMS spectra
were recorded on a Waters ESI-Q-TOF Premier XE mass
spectrometer. All synthesized compounds gave satisfactory
NMR and HRMS spectra. Analytical HPLC was performed
on UltiMate 3000 LC System (Dionex) using H2O/ACN elu-
tion system (0%, 5 min, then, 0% → 100%, 20 min). Inhibi-
tion constants were calculated from the spectrofuorometric
measurements carried out on Spectra MAX Gemini EM
fuorimeter.
The mesyl derivative of amino- and carboxylate-protected
phosphinic dehydrodipeptide 4h (0.322 g, 0.5 mmol) was
stirred with trifuoroacetic acid (10 mL) in the presence of
thioanisole (1 mL) for 24 h at room temperature, then, the
volatiles were evaporated in vacuo. The following addition
of cold diethyl ether (10 mL) precipitated the crude prod-
uct. The solid was separated by centrifugation and solvent
decantation. The crude peptide 5h was purifed by fash
chromatography on a C18 reversed phase column in H2O/
acetonitrile (0 → 100%) eluent.
Double Bond Reduction (Kalyva et al. 2015;
Georgiadis et al. 2003)
Nickel(II) chloride (0.323 g, 2.50 mmol, 10 eq.) was added
to a protected phosphinic dehydrodipeptide 5e or 5h (0.25
mmol) dissolved in the mixture of dry THF (10 mL) and
absolute EtOH (5 mL), and cooled to - 30–40 °C. Sodium
borohydride was added (0.189 g, 5.00 mmol, 20 eq.) in
three portions and the reaction was stirred for 1 h, then,
evaporated in vacuo. The residue was partitioned between
ethyl acetate and 5% KHSO4 (25 mL:25 mL) for extraction.
The water phase was discarded and the organic phase was
washed again with 5% KHSO4 and brine (two times, 25 mL
of each). The organic phase was dried over Na2SO4, evapo-
rated in vacuo and the residue was purifed by fash chro-
matography on silica using AcOEt/MeOH (0% → 100%)
eluent. The removal of the protection groups was achieved
as described above for unsaturated derivatives.
Synthesis of Phosphinic Dehydrodipeptides
(Vassiliou et al 1999; Grembecka et al. 2003)
I n a n o ve n - d r i e d t h r e e - n e c k e d f l a s k ,
1-(N-benzyloxycarbonylamino)-3-phenylpropyl-H-phos-
phinic acid 3 (1.00 g, 3 mmol) (Baylis et al. 1984; Grem-
becka et al. 2003) and hexamethyldisilazane (10–15 mL)
were heated at 90–110 °C for 2–3 h under Ar. Then, the
reaction mixture was cooled to room temperature and the
appropriate MBH acetate 2 (Basavaiah and Pandiaraju 1996;
Yu et al. 2001) (4.20 mmol, 1.4 eq.) dissolved in toluene (5
mL) was added. The mixture was heated again at 90–110
°C for 2–3 h, then, cooled to room temperature. MeOH (5
mL) was added and the mixture was stirred overnight. The
volatiles were evaporated in vacuo and the residue was par-
titioned between ethyl acetate and 5% KHSO4 (50 mL:50
mL) for extraction. The water phase was discarded and the
organic phase was washed again with 5% KHSO4 and brine
(50 mL of each). The organic phase was dried over Na2SO4,
evaporated in vacuo and the residue was purifed by fash
chromatography on silica using n-hexane/AcOEt (0% →
Molecular Modeling
Molecular modeling studies were performed using the
Biovia Discovery Studio package and the GOLD package.
The crystal structure of the enzymes HsAPN (PDF:4FYT)
1 3

International Journal of Peptide Research and Therapeutics
(Wang et al. 2007) and SsAPN (PDB:6BV0) (Joshi et al.
2017) were used. The ligands and waters were removed and
appropriate protonation states in the relation to experimen-
tal studies were established: 7.0 pH for HsAPN and 7.5 pH
for SsAPN. Ligands were protonated as well and optimized
with the use of CHARMm force feld and Smart Minimizer
algorithm, up to 0.1 Å. The docking sites were determined
from the cavity option and the sphere representation limited
to the 11 Å (HsAPN) and 15 Å (SsAPN) radius from the
point of the metal ion in the active center. Molecular docking
studies were performed by consecutive application of GOLD
and In-Situ Minimization using Smart Minimize algorithms.
The GOLD program is implemented in BIOVIA DS. GA
parameters were changed to Automatic Very Flexible. Smart
Minimizer algorithm optimized the results with all the other
protein residues outside the docking sphere kept fxed. The
binding energy of all docking poses were calculated accord-
ing to the equation:
substrate is known to be favorably bound in the correspond-
ing S1 pocket of several M01 and M17 metalloaminopepti-
dases (Grembecka et al. 2003; Mucha 2005; Mucha et al.
2011; Vassiliou et al. 2014). The displacement reaction pro-
ceeded gently under the activation of the H-phosphinic acid
to its tervalent ester with hexamethyldisilazane (Scheme 2).
The crude products 4a–h were separated after methanolysis
and were purifed chromatographically to provide either an
E/Z diastereomeric mixture (not separated) or pure Z stere-
oisomers. Deprotection in acidic conditions gave the fnal
compounds 5a–h. The cleavage of the mesyl group from the
phenolic -OH group of 5h was achieved using ammonium
hydroxide to yield 5i (Ohgiya and Nishiyama 2004).
The phosphinic dehydrodipeptides were tested for inhi-
bition of two mammalian M01 metalloaminopeptidases,
namely, porcine and human alanyl aminopeptidases (APNs).
The dipeptidyl analogs appeared to be potent, competitive
and fast-binding compounds with inhibitory constants, Ki,
in the low micromolar range for porcine kidney (SsAPN)
and with submicromolar values for the human ortholog
(HsAPN, Table 1). Expanded, hydrophobic and heteroatom-
free substituents (benzyl, styryl and biphenyl) of the P1’
double bond (compounds 2b, 2d and 2e, respectively) gave
rise to the most signifcant inhibition (Ki = 0.85–2.88 µM
for SsAPN and Ki = 0.134–0.151 µM for HsAPN). The
high afnity of the biphenyl-based P1’ residue to the APNs
was previously reported for phosphinic tripeptide analogs
(Chen et al. 2000). Dehydropeptides 5g and 5i containing
the phenyl group functionalized with nitro and hydroxy
groups also displayed very good afnity, particularly for the
human aminopeptidase (Ki = 0.192 µM and Ki = 0.18 µM,
respectively).
ΔGBinding_energy = ΔGComplex_energy − ΔGProtein_energy − ΔGLigand_energy
Enzyme Inhibition Assays (Vassiliou et al. 2014)
A 96-well plate format was used in spectrofuorometric
measurements. The excitation wavelength was set on 355
nm while the emission wavelength at 460 nm. Enzymes were
preincubated for 30 min at 37 °C. The release of the AMC
fuorophore was measured continuously for 45 minutes. The
velocity was calculated from the linear part of the progress
curve. The competitive version of Cheng–Prusof equation
allowed to compute the Ki values: Ki = IC50/[1 + (S/Km)].
Concentration of the inhibitor equal to 50% inhibition was
shown as IC50 value calculated from relation of the hydroly-
sis velocity to the logarithm of the inhibitor concentration
[I]. The SoftMax Pro, GraphPad Prism and Microsoft Excel
programs were used for kinetic parameters calculations.
Although the anti-aminopeptidase activity of phos-
phinic dehydrodipeptides was found to be satisfactory,
it did not achieve the excellent nanomolar level reported
previously for saturated analogs (Grembecka et al. 2003;
Vassiliou et al. 2014). For example, hPheψ[P(O)(OH)
CH2]-Phe inhibited HsAPN with Ki = 0.002 µM (Vassiliou
et al. 2014) (Ki = 0.62 µM for the corresponding dehydro
derivative 5a reported here) while hPheψ[P(O)(OH)CH2]-
Tyr inhibited SsAPN with Ki = 0.036 µM (Grembecka
et al. 2003). (Ki = 3.82 µM for the corresponding 5i).
Apparently, the double bond constraints led to a decrease
in the potency by more than two orders of magnitude. To
further illustrate this result, we reduced two phosphinic
dehydrodipeptides (5e and 5 h) to their saturated analogs
(6e and 6h) and compared their activities. The reduction
proceeded readily with the use of sodium borohydride
in the presence of nickel chloride at low temperature as
described elsewhere (Kalyva et al. 2015; Georgiadis et al.
2004). Indeed, the saturated phosphinic peptides inhibited
the tested metalloaminopeptidases much more efectively
than the dehydropeptides (Table 2). Only the potency of 6e
Results and Discussion
Activated allyl acetates 2a–h, the electrophilic starting
materials for obtaining the target dehydrodipeptides, were
synthesized in a typical two-step procedure that involved
the Morita–Baylis–Hillman reaction of aldehydes with an
acrylate, followed by acetylation (Basavaiah and Pandiaraju
1996; Yu et al. 2001) (Scheme 2). Phenylalkyl aldehydes and
p-substituted benzaldehydes 1a–h were selected to deliver
bulky aromatic P1’ side chains. The hydroxy group of 1h
was mesylated. N-Cbz protected α-aminoalkyl-H-phosphinic
acid 3, the analog of homophenylalanine (Baylis et al.
1984; Grembecka et al. 2003), was used as the conserved
nucleophilic component. The phenylethyl side chain of this
1 3

International Journal of Peptide Research and Therapeutics
Scheme 2 Synthesis of phosphinic dehydrodipeptides from the
Morita–Baylis–Hillman acetates as the P1’ precursors. Reagents and
conditions: a DABCO, DMSO/H2O, rt, 7 days; b AcCl, Et3N, dry
CH2H2, 0 °C, 1 h, rt, 16–24 h; c HMDS, 95–110 °C, Ar, 2 h; d 2a–h
in dry toluene, 90–115 °C, Ar, 2–3 h, then MeOH, rt, 16–24 h; e 33%
HBr in AcOH, rt, 2 h or TFA, thioanisole, rt, 24 h (for 4 h); f 25%
Me4N+OH− in H2O, MeOH, 10 Eq. rt, 24 h.
against SsAPN remained at the same level as the potency
of 5e. In the other cases, reduced compounds improved
the afnity toward APNs by a factor of 7–28. Ki values for
human APN were below 0.035 µM, with the biphenyl-sub-
stituted analog being the most active found in this study
(Ki = 0.021 µM).
of the corresponding binding pockets with the P1 and P1’
side chains.
Clearly, the binding is not optimal for each stereoisomer,
as indicated by the detailed correlations of the energies with
the confgurations. Among the unsaturated compounds, with
a single exception, Z geometrical isomers appeared disad-
vantageous and lost the opportunity for canonical binding
with the zinc ion (Table 2). As Z stereochemistry is domi-
nant for the synthesized samples, such an unfavorable com-
position could be the reason for their lower than expected
potency. The exception was the biphenyl derivative 5e bound
to HsAPN, and all four diastereoisomers were characterized
by high binding energy values. Among these, only the R-
(E) form was an exception to the standard mode of action.
Accordingly, 5e was one of the most potent compounds
against HsAPN among the dehydrodipeptides. In this case,
the advantageous overall conformations of the stereoisomers
are driven by the complementarity of the biphenyl to the
large hydrophobic pocket.
To obtain an additional insight into the structure-activ-
ity relationship, we performed docking simulations for
P1’-unsaturated and saturated dipeptides using the GOLD
docking protocols with in situ minimization and calculated
the free binding energies. Although the compounds were
assayed for kinetics with APNs as the mixtures, the calcula-
tion protocols were performed for individual stereoisomers,
taking into account the R/S and E/Z confgurations (Table 2).
Considering the best poses, compounds 6e and 6h were ener-
getically more favorable to some extent than 5e and 5h. Due
to their rotational freedom, the latter-mentioned inhibitors
could more easily balance the positioning of the phosphinic
group against the zinc ion, with the simultaneous penetration
1 3

International Journal of Peptide Research and Therapeutics
Table 1 Inhibition of porcine (Sus scrofa, Ss) and human (Homo
sapiens, Hs) alanyl aminopeptidase (APN) by phosphinic dehydrodi-
peptides 5a–i bearing aromatic substitution of the P1’ double bond
modeled binding mode of stereoisomer R,S-6e (correspond-
ing to the natural l,l confguration) with HsAPN revealed the
prototypical bidentate coordination of the negatively charged
phosphinate moiety to the zinc metal ion (Fig. 1b). In addi-
tion, one of the phosphinate oxygen atoms forms a hydrogen
bond network with the carboxylate of Glu411 and imidazole
nitrogen of His388. Another set of hydrogen bonds involve
interactions of the inhibitor amino group with the carboxy-
late Glu411 and the hydroxylate of Tyr477. The C-terminal
functionality has contacts with the guanidine of Arg381 and
the N-H of Ala353. The side chains ft well in the spacious
and hydrophobic binding pockets as the biphenyl ring has an
increased surface area involved in lipophilic interactions with
the residues of the S1’ cavity, e.g., His388 and Glu418. Quite
unexpectedly, its enantiomer (S,R-6e, corresponding to d,d
relative confguration) was capable of an even more favorable
set of polar contacts: (1)>PO2− with Zn2+ and with HO– of
Tyr477, (2) –NH3+ with COO− of Glu355, Glu389 and with
C=O of Ala353, and (3) COO– with guanidine of Arg381,
accompanied by the penetration of the hydrophobic enzyme
pockets with the side-chain residues (Fig. 1c). This comple-
mentarity gave rise to an increased free energy of binding.
Although bound to the central metal ion via phosphinate,
the best poses of the two remaining stereoisomers were
somewhat less conveniently accommodated by HsAPN and
avoided the canonical interaction mode. In fact, the overall
arrangements implied the reverse binding of the side-chain
residues, namely, the P1 to the S1’ pocket and the P1’ to the
S1 pocket. The bent backbone conformations oriented the
polar groups in front of each other on the same side of the
molecule. For R,R-6e, the acidic groups, namely, phosphi-
nate and carboxylate, were positioned in parallel and both
complexed with Zn2+ via one oxygen atom (Fig. 1a). The
remaining oxygen atoms formed hydrogen bonds with the
neighboring residues of Gln213, Ala351 and Glu355. The
amino group conserved an opportunity to form only a single
electrostatic contact with Glu389. For S,S-6e, the phosphi-
nate interacted in a bidentate manner with the zinc atom, and
with the hydroxyl group of Tyr477 (Fig. 1d). The inhibitor
carboxylate was positioned in proximity to the amino func-
tionality to form an intramolecular salt bridge that stabilized
the overall conformation. The C- and N-termini were also
involved in a set of hydrogen bonds with the heteroatoms of
Gln211, Glu355 and Met 354. The binding modes of these
two stereoisomers were less favorable and characterized by
a somewhat higher calculated free energy.
Entry
Inhibition, Ki [μM]
SsAPN
HsAPN
36.8 1.2
0.620 0.20
5a
5b
2.88 0.83
0.151 0.031
6.85 2.6
0.254 0.060
5c
1.92 0.15
5d
0.134 0.023
0.850 0.060
5e
0.151 0.028
13.0 5.1
9.70 0.28
19.5 4.9
0.491 0.086
0.192 0.018
0.984 0.34
5f
5g
5h
3.82 0.22
5i
0.180 0.001
The most signifcant activities are highlighted in bold
In the cases of 6e and 6h, the S absolute confguration (the
d relative confguration) of the P1 fragment was discriminating
for the porcine-originated aminopeptidase. The S,R and S,S
diastereoisomers interacted poorly in the active site and did
not utilize contacts with the catalytic metal cation. By con-
trast, HsAPN was less demanding and accepted distinct forms,
e.g., all four stereoisomers of 6e gave rise to high-energy
interactions (Table 2 and Fig. 1). Thus, the highest measured
potency of this phosphinic dipeptide could be explained by
the Zn-phosphinate coordination, and a range of tight con-
tacts between both the inhibitor backbone and the side chains
with the enzyme residues found for each diastereoisomer. The
Conclusion
To conclude, we suggested phosphinic dehydrodipeptides
as potential inhibitors of metalloaminopeptidases; however,
these compounds did not reach the high afnity of their
1 3

International Journal of Peptide Research and Therapeutics
Table 2 Inhibition
of mammalian alanyl
aminopeptidases by selected
phosphinic dipeptides (6e and
6h) compared to their dehydro
analogs (5e and 5h)
Configuration
R =
R-(E) R-(Z) S-(E) S-(Z) R,R
R,S
S,R
S,S
SsAPN Ki [μM]
0.850 0.060
1.28 0.26
ΔG [kcal/mol] -272
-38.2* -158
0.151 0.028
-49.2* -199
-259 -150* -258*
0.021 0.011
HsAPN Ki [μM]
ΔG [kcal/mol] -297* -349
-352
-406
-308
-367 -398
2.85 0.75
-315
R
SsAPN Ki [μM]
19.5 4.9
-169 -232
ΔG [kcal/mol] -204
HsAPN Ki [μM]
ΔG [kcal/mol] -373
-93.1* -241
-236 -162* -120*
0.035 0.0045
0.984 0.34
-290* -370
-284* -210* -375 -399
-390
The kinetic data are further illustrated by free binding energy values, calculated for the stereo-defned
ligand-protein complexes using the GOLD docking protocols with in situ minimization
saturated analogs. Because of structural constraints, the
porcine APN, but the rotational freedom gave rise to a lower
binding energy of the distinct diastereomeric forms. While
the detailed mode of interaction was apparently dependent
on the structure of the side-chain substituents, an advan-
tageous ft was found more frequently for the saturated
compounds than for their unsaturated counterparts. Thus,
to achieve active compounds, the convenient synthetic pro-
tocol adopted for phosphinic dehydrodipeptides and based
on the substitution of the Morita–Baylis–Hillman acetates
with aminoalkylphosphinic acids should be extended by the
subsequent reduction. The whole procedure is mild and can
be readily performed for diferent P1’ substituents.
anti-aminopeptidase activity of the double bond-containing
compounds appeared to be much more dependent on the
confguration of individual stereoisomers and composition
of the materials. Certain arrangements of dehydrodipeptides
were suggested to lead to weak nonclassical binding that
did not implicate coordination with the catalytic metal ion
in the active center. To achieve the desired ligand-enzyme
complementarity, these compounds must be synthesized in
an entirely stereoselective manner. For phosphinic dipep-
tides, the confguration appeared to be less discriminating.
Certain stereoisomers were clearly preferred, in particular by
1 3

International Journal of Peptide Research and Therapeutics
Fig. 1 Modeled binding modes of diastereoisomers of 6e to HsAPN
(PDB:4FYT) (Wong et al. 2012): a (R,R), b (R,S), c (S,R), d (S,S).
Inhibitors and enzyme amino acid residues are shown as sticks, while
the zinc ion is represented by a dark blue sphere. Hydrogen bonds
and interactions with metal ions are shown as thin green lines. The
surface of the enzyme is shown in gray.
Acknowledgements This work was financed by a grant from the
Polish National Science Centre to Artur Mucha (Grant 2013/09/B/
ST5/00090). The Biovia Discovery Studio package was used under a
Polish country-wide license. The use of the software resources (BIO-
VIA Discovery Studio program package) of the Wrocław Centre for
Networking and Supercomputing is kindly acknowledged.
References
Basavaiah D, Pandiaraju S (1996) Nucleophilic addition of triethyl
phosphite to acetates of the Baylis–Hillman adducts: stereose-
lective synthesis of (E)- and (Z)-allylphosphonates. Tetrahedron
52:2261–2268. https://doi.org/10.1016/0040-4020(95)01055-6
Baylis EK, Campbell CD, Dingwall JG (1984) 1-Aminoalkylphos-
phonous acids. Part 1. Isosteres of the protein amino acids. J
Chem Soc Perkin Trans 1:2845–2853. https://doi.org/10.1039/
P19840002845
Compliance with Ethical Standards
Ethical Approval This article does not contain any studies with human
participants or animals performed by any of the authors.
Boyd EA, Corless M, James K, Regan AC (1990) A versatile route to
substituted phosphinic acids. Tetrahedron Lett 31:2933–2936.
https://doi.org/10.1016/0040-4039(90)80188-R
Conflict of interest Author declares that he has no confict of interest.
Chen H, Noble F, Mothé A, Hervé M, Coric P, Danascimento S,
Roques BP, George P, Fournié-Zaluski MC (2000) Phosphinic
derivatives as new dual enkephalin-degrading enzyme inhibi-
tors: synthesis, biological properties, and antinociceptive
1 3

International Journal of Peptide Research and Therapeutics
Parsons WH, Patchett AA, Bull HG, Schoen WR, Taub D, Davidson
C, Combs P, Springer JP, Gadebusch H, Weissberger B, Valiant
ME, Mellin TN, Busch RD (1988) Phosphinic acid inhibitors of
d-alanyl-d-alanine ligase. J Med Chem 31:1772–1778. https://doi.
org/10.1021/jm00117a017
activities. J Med Chem 43:1398–1408. https://doi.org/10.1021/
jm990483l
Chen S, Coward JK (1996) A general method for the synthesis of
N-protected α-aminoalkylphosphinic acids. Tetrahedron Lett
37:4335–4338. https://doi.org/10.1016/0040-4039(96)00839-8
Collinsová M, Jirácek J (2000) Phosphinic acid compounds in bio-
chemistry, biology and medicine. Curr Med Chem 7:629–647.
https://doi.org/10.2174/0929867003374831
Skinner-Adams TS, Stack CM, Trenholme KR, Brown CL, Grembecka
J, Lowther J, Mucha A, Drag M, Kafarski P, McGowan S, Whiss-
tock JC, Gardiner DL, Dalton JP (2010) Plasmodium falciparum
neutral aminopeptidases: new targets for anti-malarials. Trends
Biochem Sci 35:53–61. https://doi.org/10.1016/j.tibs.2009.08.004
Stetter H, Kuhlmann H (1979) Eine einfache herstellung von
α-alkylacrylsäure-estern. Synthesis 1979:29–30. https://doi.
org/10.1055/s-1979-28537
Georgiadis D, Cuniasse P, Cotton J, Yiotakis A, Dive V (2004) Struc-
tural determinants of RXPA380, a potent and highly selective
inhibitor of the angiotensin-converting enzyme C-domain. Bio-
chemistry 43:8048–8054. https://doi.org/10.1021/bi049504q
Georgiadis D, Dive V (2015) Phosphinic peptides as potent inhibitors
of zinc-metalloproteases. Top Curr Chem 360:1–38. https://doi.
org/10.1007/128_2014_571
Suzuki K, Muto Y, Fushihara K, Kanemoto K, Iida H, Sato E, Kikuchi
C, Matsushima T, Kato E, Nomoto M, Yoshioka S, Ishii H (2004)
Enhancement of fbrinolysis by EF6265 [(S)-7-amino-2-[[[(R)-
2-methyl-1-(3-phenylpropanoylamino)propyl]hydroxyphosphi-
noyl]methyl]heptanoic acid], a specifc inhibitor of plasma car-
boxypeptidase B. J Pharmacol Exp Ther 309:607–615. https://doi.
org/10.1124/jpet.103.062729
Grembecka J, Mucha A, Cierpicki T, Kafarski P (2003) The most
potent organophosphorus inhibitors of leucine aminopeptidase.
Structure-based design, chemistry, and activity. J Med Chem
46:2641–2655. https://doi.org/10.1021/jm030795v
Gurulingappa H, Buckhalts P, Kinzler KW, Vogelstein B, Khan SR
(2004) Synthesis and evaluation of aminophosphinic acid deriva-
tives as inhibitors of renal dipeptidase. Bioorg Med Chem Lett
14:3531–3533. https://doi.org/10.1016/j.bmcl.2004.04.057
Gurulingappa H, Buckhaults P, Kumar SK, Kinzler KW, Vogelstein
B, Khan SR (2003) Design, synthesis and evaluation of new
RDP inhibitors. Tetrahedron Lett 44:1871–1873. https://doi.
org/10.1016/S0040-4039(03)00082-0
Talma M, Mucha A (2017) P-C bond formation in reactions of Morita-
Baylis-Hillman adducts with phosphorus nucleophiles. Arkivoc
324–344. https://doi.org/10.3998/ark.5550190.p009.787
Thottathil JK, Ryono DE, Przybyla CA, Moniot JL, Neubeck R (1984)
Preparation of phosphinic acids: Michael additions of phos-
phonous acids/esters to conjugated systems. Tetrahedron Lett
25:4741–4744. https://doi.org/10.1016/S0040-4039(01)81507-0
Vassiliou S, Mucha A, Cuniasse P, Georgiadis D, Lucet-Levannier
K, Beau F, Kannan R, Murphy G, Knauper V, Rio M-C, Basset
P, Yiotakis A, Dive V (1999) Phosphinic pseudo-tripeptides as
potent inhibitors of matrix metalloproteinases: a structure—activ-
ity study. J Med Chem 42:2610–2620. https://doi.org/10.1021/
jm9900164
Joshi S, Chen L, Winter MB, Lin YL, Yang Y, Shapovalova M,
Smith PM, Liu C, Li F, LeBeau (2017) The rational design of
therapeutic peptides for aminopeptidase N using a substrate-
based approach. Sci Rep 7:1424–1434. https://doi.org.10.1038/
s41598-017-01542-5
Kalyva M, Zografos AL, Kapourani E, Giambazolia E, Devel L, Papa-
kyriakou A, Dive V, Lazarou Y, Georgiadis D (2015) Probing
the mechanism of allylic substitution of Morita–Baylis–Hillman
acetates (MBHAs) by using the silyl phosphonite paradigm:
scope and applications of a versatile transformation. Chem Eur J
21:3278–3289. https://doi.org/10.1002/chem.201405626
Matziari M, Beau F, Cuniasse P, Dive V, Yiotakis A (2004) Evaluation
of P1’-diversifed phosphinic peptides leads to the development of
highly selective inhibitors of MMP-11. J Med Chem 47:325–336.
https://doi.org/10.1021/jm0308491
Vassiliou S, Węglarz-Tomczak E, Berlicki Ł, Pawełczak M, Nocek B,
Mulligan R, Joachimiak A, Mucha A (2014) Structure-guided,
single-point modifcations in the phosphinic dipeptide structure
yield highly potent and selective inhibitors of neutral aminopepti-
dases. J Med Chem 57:8140–8151. https://doi.org/10.1021/jm501
071f
Wang YX, Zhao L, Nagashima M, Vincelette J, Sukovich D, Li W,
Subramanyam B, Yuan S, Emayan K, Islam I, hrvatin P, Bryant
J, Light DR, Vergona R, Morser J, Buckman BO (2007) A novel
inhibitor of activated thrombin-activatable fbrinolysis inhibitor
(TAFIa)—part I: pharmacological characterization. Thromb Hae-
most 97:45–53. https://doi.org/10.1160/TH06-09-0551
Wong AHM, Zhou D, Rini JM (2012) The X-ray crystal structure of
human aminopeptidase N reveals a novel dimer and the basis for
peptide processing. J Biol Chem 287:36804–36813. https://doi.
org/10.1074/jbc.M112.398842
Matziari M, Georgiadis D, Dive V, Yiotakis A (2001) Convenient syn-
thesis and diversifcation of dehydroalaninyl phosphinic peptide
analogues. Org Lett 3:659–662. https://doi.org/10.1021/ol006
9103
Matziari M, Nasopoulou M, Yiotakis A (2006) Active methylene
phosphinic peptides: a new diversifcation approach. Org Lett
8:2317–2319. https://doi.org/10.1021/ol060575m
Morita K, Suzuki Z, Hirose H (1958) A tertiary phosphine-catalyzed
reaction of acrylic compounds with aldehydes. Bull Chem Soc Jpn
41:2815–2815. https://doi.org/10.1246/bcsj.41.2815
Yamagishi T, Tashiro N, Yokomatsu T (2011) Diastereoselective syn-
thesis of the Leu-Pro type phosphinyl dipeptide isostere. J Org
Chem 76:5472–5476. https://doi.org/10.1021/jo200669r
Yiotakis A, Georgiadis D, Matziari M, Makaritis A, Dive V (2004)
Phosphinic peptides: synthetic approaches and biochemical evalu-
ation as Zn-metalloprotease inhibitors. Curr Org Chem 8:1135–
1158. https://doi.org/10.2174/1385272043370177
Mucha A (2012) Synthesis and modifcations of phosphinic dipeptide
analogues. Molecules 17:13530–13568. https://doi.org/10.3390/
molecules171113530
Mucha A, Kafarski P, Berlicki Ł (2011) Remarkable potential of the
α-aminophosphonate/ phosphinate structural motif in medicinal
chemistry. J Med Chem 54:5955–5980. https://doi.org/10.1021/
jm200587f
Yu C, Liu B, Hu L (2001) Efcient Baylis–Hillman reaction using
stoichiometric base catalyst and an aqueous medium. J Org Chem
66:5413–5418. https://doi.org/10.1021/jo015628m
Ohgiya T, Nishiyama S (2004) A simple deprotection of trifate esters
of phenol derivatives. Tetrahedron Lett 45:6317–6320. https://doi.
org/10.1016/j.tetlet.2004.06.104
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
Parsons WH, Hajdu R, Schoen WR, Combs PL, Sundelof J, Patchett
AA (1991) A new class of potent, slowly reversible dehydropepti-
dase inhibitors. Biochem Int 23:1107–1115
1 3