S.R. DeSouza, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127342
of PLP with LMW-PTP IFB via 1H NMR. It forms several key inter-
tyrosines 131 and 132. However, when examining the alkene phos-
phonates, there is only a minimum increase in binding preference when
a methyl group is introduced at the R1 position. Again, this is suggestive
of a potential alternative binding within the active site for these alkene
phosphonates. Modeling studies suggest that the alkene phosphonates
without a methyl group in the R1 position bind within the active site but
also possess an unfavorable donor-donor interaction between the
phosphonate group and Gly14 and Leu13. The entropic decrease by
having an alkene substituent might be offset by unfavorable ligand
protein interactions and may lead to these mixed results.
molecular interactions that lead to the tight binding of PLP and LMW-
11
PTP IFB.
The pyridine nitrogen of PLP is protonated to give the
pyridinium form of the ring which interacts with the deprotonated form
of Asp-129 via ion-ion interactions. The phosphate group on PLP in-
teracts with the P-loop through a series of hydrogen bonds. The hy-
droxyl and aldehyde groups on PLP may interact with residues close to
the active site such as Asn-50 and Trp-49. Zhou and Van Etten proposed
that these two groups act to increase the basicity of the pyridine ni-
trogen which favors the protonated form.
In the current study, several non-hydrolysable analogs of PLP were
synthesized to probe which groups on PLP are essential to its binding
and to determine if these requirements differ between the two isoforms
of LMW-PTP. Results from these studies should be helpful in designing
new selective isoform inhibitors.
effect on binding. For the conformationally unrestrained phosphonates
an increase in binding bias for the hydroxyl group ranged from 36-fold
(comparing compounds 4 and 7) to 42-fold (comparing compounds 2
and 6). Modeling studies suggest that the hydroxyl group is potentially
interacting with either tyrosine 132 (compound 6) or glutamic acid 50
(compound 7). The hydroxyl group in the restricted rotation alkene
containing phosphonates gave mixed results. There was still a pre-
ference for the hydroxyl group when comparing compounds 10 and 14
(30-fold preference) but, very little when comparing compounds 12 and
15 (1.3-fold difference).
Phosphonate derivatives of PLP were synthesized as previously de-
scribed by Knobloch et al.15 Synthetic schemes and experimentals are
available in Supporting Information.
In addition, molecular modeling studies were performed using Auto
Dock Vina (version 1.1.2). Protein Databank files for the crystal struc-
tures of human LWM-PTP isoform A (5PNT) and isoform B (1XWW)
were used in docking studies with all 17 compounds plus PLP. Potential
ligand-protein interactions were explored using Discovery Studio
Visualizer (version 20.1.0.19295). Docking scores and 2-D ligand-pro-
tein interaction diagrams are available in Supporting Information.
Looking at the structural requirements for binding to IFA as pre-
sented in Tables 1 and 2, the pyridine nitrogen appears to be critical to
an increase in selectivity from 5- to 70-fold, respectively, when the
pyridine nitrogen is introduced as opposed to just a benzene ring.
Modeling studies did not suggest any additional intermolecular inter-
action between the pyridine nitrogen and the protein. Modest results
are seen when looking at compounds 10 and 11 (2-fold increase in
binding preference). Interestingly, however, when compounds 12 and
13 are compared there is an approximate 3-fold bias for the benzene
ring over the pyridine ring. This might suggest that the alkene phos-
phonates might bind differently in the active site or that the entropic
gain from the restricted rotation assists in binding or that the pyridine
nitrogen is not as critical for binding of the phosphonates as was sug-
gested for PLP or potentially some combination of all of these factors.11
Docking studies also suggest that there would be these slight pre-
ferences. Please see Supporting Information for docking scores.
Introduction of a methyl group at R1 position increases the se-
lectivity from 4-fold (comparing compounds 2 and 4) to 56-fold
(comparing compounds 3 and 5). According to modeling studies, this is
probably due to hydrophobic interactions of the methyl group with
Introduction of a hydrogen bond acceptor or hydrogen bond ac-
ceptor/donor at the R3 position increased the inhibitory ability of the
compounds whether the phosphonates had free rotation or restricted
rotation. Interestingly, there is a slight preference for a hydrogen bond
acceptor when the phosphonate has restricted rotation and a hydrogen
bond donor when the phosphonate has free rotation about the carbon-
carbon bond. Modeling studies suggest that when there is free rotation
about the carbon-carbon bond the hydrogen bond donor group in the R3
position is able to form a conventional hydrogen bond with aspartic
acid 129, whereas placing a hydrogen bond acceptor in this position
results in simply a diploe/dipole interaction with aspartic acid 129. In
the phosphonates with restricted rotation, aspartic acid 129 is still able
to interact using dipole/diploe interactions with the hydrogen bond
acceptor in R3 position but, is not able to hydrogen bond with a hy-
drogen bond donor group in that position as is seen with those phos-
phonates with free rotation.
Those inhibitors that bind well to IFA are shown below in Fig. 1.
They all essentially take advantage of similar interactions with the
enzyme. The charged phosphonate head interacts with arginine 18
through ion/ion interactions. The rest of the phosphonate is held in
place by a network of hydrogen bonding with residues Cys12, Cys17,
Gly14, Ile 16 and Asn 15. The aromatic ring on the inhibitors interacts
with tyrosine 131 via pi-stacking interactions. And finally, side chains
in the R1, R2 and R3 positions interact with residues Tyr132, Asp129,
Table 1
In vitro activity of phosphonates against LMW-PTP isoforms
Compound
X
R1
R2
R3
R4
Kis IFA (μM)
Kis IFB (μM)
X Fold Selectivity for IFA
1
2
3
4
5
6
7
8
9
CH
CH
N
H
H
H
F
359
552
103
130
1.84
13.0
3.59
10.4
74.9
82
N.I.
N.I.
> 28
> 18
10.4
2.5
8.5
4
H
H
H
H
H
H
H
H
H
H
H
278
36
H
H
H
1075
323
293
75
CH
N
CH3
CH3
H
H
H
60
H
H
0.47
1.7
0.84
2.1
26.1
15.6
52.2
39.9
552
8.9
CH
CH
N
OH
OH
OH
OH
H
4.8
4.6
CH3
CH3
CH3
H
11
CH2OH
CHO
134
124
53
N
218
2.9
N.I. No inhibition where Kis > 10 mM.
2