Linearized and truncated microcystin analogues as inhibitors of protein phosphatases 1 and 2A

Article abstract of DOI:10.1016/S0960-894X(03)00589-4

A series of acyclic, truncated microcystin analogues, comprised of the dienic β-amino acid (Adda) and up to four additional amino acids characteristic of the parent toxin, was synthesized and screened for activity as inhibitors of PP1 and PP2A. Despite a recent report to the contrary for a microcystin-derived tetrapeptide degradation product, none approaches the potency of microcystin itself.

Full text of DOI:10.1016/S0960-894X(03)00589-4

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2903–2906
Linearized and Truncated Microcystin Analogues as Inhibitors of
Protein Phosphatases 1 and 2A
Brian M. Gulledge, James B. Aggen and A. Richard Chamberlin*
Department of Chemistry, University of California at Irvine, Irvine, CA 92697, USA
Received 6 December 2002; accepted 2 May 2003
Abstract—A series of acyclic, truncated microcystin analogues, comprised of the dienic b-amino acid (Adda) and up to four addi-
tional amino acids characteristic of the parent toxin, was synthesized and screened for activity as inhibitors of PP1 and PP2A.
Despite a recent report to the contrary for a microcystin-derived tetrapeptide degradation product, none approaches the potency of
microcystin itself.
# 2003 Elsevier Ltd. All rights reserved.
The protein phosphatases 1 and 2A, (PP1 and PP2A,
respectively), are crucial elements in the regulation of
many cell signaling pathways.1 The inhibition of these
enzymes can disruptnumerous physiological processes,
ranging from glycogen synthesis,2 cell division,3 to
memory and learning.4 Only recently have the respective
roles of PP1 and PP2A begun to be delineated, due in
large part to the discovery of selective inhibitors—gen-
erally naturally occurring toxins—for cell biology stud-
ies.5
for PP1. While these initial results were gratifying, the
rather laborious synthesis of the macrocyclic heptapep-
tides suggested that it would be prudent to investigate
whether simplified, acyclic analogues might achieve the
same goal.
Recently, Jones et al. reported studies on the natural
enzymatic degradation pathway of 1,8,9 including the
Among the most potent inhibitors of PP1 and PP2A are
the microcystins (e.g., microcystin-LR, 1, Fig. 1), a large
family of natural toxins characterized by the presence of
an otherwise unusual b-amino acid (2S,3S,8S,9S,
4E,6E)-3-amino-9-methoxy-2,6,8-trimethyl-10-phenyl-4,6-
decadienoic acid (generally referred to by its acronym
‘Adda’). While these cyclic heptapeptides are quite
potent and selective for PP1 and PP2A over other ser-
ine-threonine phosphatases (e.g., PP2B, PP2C, etc.),
they do not distinguish at all between the two
(IC50=0.3 nM for PP1 and PP2A).6 To initiate a search
for more selective inhibitors, our group synthesized one
member of this family, microcystin-LA7 as well as a series
of rationally-designed ‘full-sized’ variants.6 Assays of
PP1/PP2A inhibition by these analogues demonstrated
that modification of the dehydroalanine, d-alanine, or
l-leucine residues gives modest increases in selectivity
*Corresponding author. Tel.: +1-949-824-7089; fax: +1-949-824-
8571; e-mail: archambe@uci.edu
Figure 1. Enzymatic degradation of MC-LR.
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00589-4

2904
B. M. Gulledge et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2903–2906
identities and inhibitory potencies of two ‘linearized,’
acyclic Adda-containing peptides. These experiments
address the important question of whether fragments of
the parent toxin might retain significant activity and
thus pose a toxicological risk in their own right. The
first degradation product in the pathway is the acyclic
heptapeptide resulting from scission of the Adda-argi-
nine amide bond to open the macrocyclic ring. This lin-
earized product was reported to be less active than its
macrocyclic precursor, 1, but it appears to retain sub-
stantial potency (PP1 IC50=90 nM). More surprisingly,
another degradation product, the linearized and trun-
cated tetrapeptide 2, was reported to be a low nano-
molar inhibitor (PP1 IC50=12 nM). This admittedly is
nearly a 100-fold less potent than the macrocyclic par-
ent but nonetheless amazingly potent considering it
lacks both the macrocyclic ring and three of the seven
amino acid residues presentin 1. Before proceeding with
simplified microcystin analogue design based on this
promising precedent, we decided to prepare the reported
tetrapeptide degradation product and several higher and
lower homologues (e.g., the corresponding penta- and
dipeptides) for comparison. We relate in this Letter that
all of these synthetic linearized and truncated peptides,
including the reportedly potent tetrapeptide degrada-
tion product, are only moderately active inhibitors of
PP2A and even less potent towards PP1.
isoglutamyl residue capped with a dimethyl amide. Pre-
vious studies had shown that Adda alone is not suffi-
cientfor inhibition, 10 so this dipeptide was chosen as the
simplest member of the linearized and truncated series
to be investigated. The synthesis was performed by
coupling 3, synthesized by a route disclosed recently by
Rinehart,11 with a suitably protected glutamate, fol-
lowed by standard protecting group manipulations to
yield 4 (Fig. 2).
The degradation tetrapeptide, 2, was synthesized using
the methodology developed previously for the synthesis
of the parent toxin 1.7 The dipeptide 5 (Fig. 3), was
deprotected by hydrogenolysis, followed by coupling of
the resultant free amine 6 to selectively protected gluta-
mate and in situ olefination to give the iso-glutamyl tri-
peptide 7. Next, 7 was coupled to 3 using HATU to
yield the protected tetrapeptide 8, which upon saponifi-
cation of the methyl esters and deprotection of the
amine afforded the degradation tetrapeptide 2.12 Unlike
the saponification of the methyl esters in our synthesis
of microcystin-LA, where significant epimerization was
observed, there was only a trace of the unwanted dia-
stereomer formed in this acyclic case.
The degradation peptide 2 differs from the parent toxin
not only in lacking the macrocyclic ring and two amino
acid residues, butalso in having a basic amino group at
the N-terminus of Adda rather than an amide group. In
order to address the issue of whether reducing the basi-
city of this nitrogen might have an effect on potency, we
prepared the corresponding N-acetylated tetrapeptide 9,
which mimics the amide group found at the corre-
sponding position in the parent toxin. This analogue
was synthesized (Fig. 4) from the intermediate 8, as
shown.
Design and Synthesis
The simplest analogue of the series to be synthesized
was the dipeptide 4, which consists of Adda and an
Figure 2. Synthesis of dipeptide 4.
Figure 4. Synthesis of acylated tetrapeptide 9.
Figure 3. Synthesis of the degradation tetrapeptide.

B. M. Gulledge et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2903–2906
2905
The olefin of the N-methyl-dehydroalanine residue is
another element of interest that may or may not play a
significant role in controlling the activity in this series of
linearized and truncated analogues. It has been shown
for the parent microcystins that neither borohydride
reduction of this residue to the corresponding N-methyl-
alanine diastereomers,13 nor replacementwith sarcosine
(N-methyl-glycine) has any effect on PP1/PP2A inhibi-
tion.6 This insensitivity towards removal of the olefin
group is particularly interesting in view of the fact that
microcystin-LR is known to undergo conjugate addition
to this residue, upon binding to PP1, by a free cysteine
in the active site. Nonetheless, the olefin still may be an
important structural feature for the linearized ana-
logues, particularly in view of additional potential con-
formational effects not present in the macrocyclic parent
toxin. Thus, two tetrapeptides substituting sarcosine for
dehydroalanine, 11 and 12, were synthesized from the
dipeptide 10 (Fig. 5) using chemistry similar to that
employed in the preparation of 2 and 9, as shown.
Additionally, two other compounds were screened as
controls: microcystin-LR (1), which gave IC50’s in good
agreement with literature values,6 and the methyl ester
of 4, which served as a negative control.
Discussion
The impetus for preparing this series of linearized ana-
logues was the surprisingly potent activity of the degra-
dation tetrapeptide 2, which was reported to have an
IC50 of 12 nM (PP1). Unfortunately, we were unable to
confirm this result with our synthetic tetrapeptide,
measuring instead a value more than three orders of
magnitude higher: IC50=45 mM for PP1 (slightly better
for PP2A, IC50=25 mM). A part—but probably not
all—of this discrepancy may be ascribed to the use of a
differentassay ( pNPP) than in the original report
(phosphorylase a), but the results obtained in the two
assays should differ only by an order of magnitude or so.
For example, in the assays reported here, microcystin has
an IC50 of 3 nM, only a factor of 10 higher than the 0.3
nM value we have determined previously in phos-
phorylase a assays. This discrepancy may be due to a
trace contaminant in the isolated degradation sample that
was responsible for mostof hte observed acitviyt. This
type of contamination is always a concern when testing
by-products isolated from mixtures containing very
active constituents. In this particular case, the presence
of only ꢀ1% of the parent toxin (or some other equally
active inhibitor) could have resulted in the reported IC50
of 12 nM reported for the degradation product.
Because our previous studies6 had shown that modifi-
cation of the leucine residue of microcystin-LA can
increase selectivity towards PP1, it was of interest to
prepare the next higher linearized homologue, the C-
terminal residue of which is that corresponding to leu-
cine. Two pentapeptide analogues were thus synthesized
(14 and 15, Fig. 6) with and without the olefin as dis-
cussed above, from the dipeptide 13 employing the
phosphosarcosine and sarcosine routes described for
other members of the series.
Despite the relatively weak activity of 2, itstill serves as
a benchmark for comparison with other linearized ana-
logues that may serve as lead compounds in the search
for structurally-simplified selective inhibitors. The tetra-
peptide 9 more closely mimics the parent compound 1,
with an amide terminus instead of the basic amine of 2,
but this acetylated derivative shows a 3-fold decrease in
PP1 inhibition, with no difference in PP2A inhibition.
Further comparison of 2 with the corresponding sarco-
sine analogues 11 (free amine) and 12 (acetylated amine)
reveals a further 5- to 10-fold erosion of potency
towards both PP1 and PP2A. This reduced activity is in
contrast to the parent toxin, which does not lose
potency when the N-methyl-dehydroalanine is replaced
Biological Assays
All of the analogues were screened in a pNPP assay14,15
against PP1 and PP2A to determine IC50 values (Table 1).
Figure 5. Sarcosine analogues 11 and 12.
Table 1. Inhibitory activity (IC50’s) againstPP1 and PP2A for
analogues
Analogue
IC50 (mM)
PP1
IC50 (mM)
PP2A
Ratio
PP1:PP2A
1
2
4
9
11
0.003
45
>100
140
280
0.003
25
20
1:1
1:2
>1:5
1:6
1:1
25
270
135
16
12
14
400
75
1:3
1:5
15
Methyl ester of 4
230
>100
17
>100
1:14
—
Figure 6. Pentapeptide analogues 14 and 15.

2906
B. M. Gulledge et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2903–2906
with sarcosine, as discussed above. These results suggest
that for the acyclic tetrapeptides the olefin is important
for binding, perhaps because itbiases acyclic con-
formational preferences in favor of a pseudo-cyclic
arrangement favored for binding. Alternatively, the
aforementioned conjugate addition to the N-methyl-
dehydroalanine double bond that does not affect mea-
sured potencies in the parent toxins may come into play
for the acyclic analogues, slowing the off-rates for 2 and
9 relative to 11 and 12 and resulting in lower IC50 values
for the former pair.
Despite their reduced potencies, some of the simplified
analogues retain sufficient activity to serve as lead com-
pounds for the development of new, selective inhibitors.
The clear lesson is that linearized and truncated ana-
logues of microcystin do not benefit—at least in terms
of potency - from additional residues beyond those pre-
sent in the Adda-containing dipeptide. In fact, we have
found that further truncation of the iso-glutamyl-resi-
due in this dipeptide is well tolerated, resulting in even
simpler analogues that retain inhibitory activity against
PP1 and PP2A. Initial results on these further simplifi-
cations are described in an accompanying Letter.
The next higher homologues, the pentapeptides 14 and
15, incorporate an additional leucine residue at the C-
terminus and behave similarly to the tetrapeptides in
most respects, but with one interesting difference. Com-
parison of 9 and 14 (both with dehydroalanine) shows a
slight, 2-fold improvement in PP1 and PP2A inhibition
for the pentapeptide over the tetrapeptide. For 12 and
15 (with sarcosine), there was also a 2-fold improvement
in PP1 inhibition, but an 8-fold increase in PP2A
potency for the pentapeptide. The net effect on selectiv-
ity is that for 15, butnot 14, there is a significant
increase in PP2A selectivity in going from the tetrapep-
tide to the corresponding pentapeptide. Both pentapep-
tides (14 and 15) were slightly more potent than the
corresponding tetrapeptides (9 and 12, respectively),
and they exhibited a similar decrease in PP1 potency for
the sarcosine-derived derivatives compared to the cor-
responding dehydro-peptides. Interestingly, however,
the absence of the olefin in the pentapeptides does not
have the same deleterious effect for PP2A; 14 and 15
have essentially identical IC50 values of 16 and 17 nM.
Acknowledgements
We gratefully acknowledge the financial support of the
National Institutes of Health (GM-57550).
References and Notes
1. Cohen, P. T. W. J. Cell Sci. 2002, 115, 241.
2. Aggen, J. B.; Nairn, A. C.; Chamberlin, A. Chem. Bio.
2000, 7, R13.
3. McCluskey, A.; Sim, A. T. R.; Sakoff, J. A. J. Med. Chem.
2002, 45, 1151.
4. Genoux, D.; Haditsch, U.; Knobloch, M.; Michalon, A.;
Storm, D.; Mansuy, I. M. Nature 2002, 418, 970.
5. Sheppeck, J. E.; Gauss, C. M.; Chamberlin, A. R. Bioorg.
Med. Chem. 1997, 5, 1739.
6. Aggen, J. B.; Humphrey, J. M.; Gauss, C. M.; Huang, H.-
B.; Nairn, A. C.; Chamberlin, A. R. Bioorg. Med. Chem. 1999,
7, 543.
7. Humphrey, J. E.; Aggen, J. B.; Chamberlin, A. R. J. Am.
Chem. Soc. 1997, 118, 11759.
As the tetrapeptide was only slightly less potent than the
pentapeptide, we were not surprised to observe that
further simplification of the structure to the dipeptide 4
did not abolish activity. Indeed, for PP2A, the IC50 was
essentially the same as the tetrapeptide 9 and the pen-
tapeptide 14, while the potency towards PP1 was
reduced (only a lower limit could be determined due to
limited solubility of the compound).
8. Bourne, D. G.; Jones, G. J.; Blakeley, R. L.; Jones, A.;
Negri, A. P.; Riddles, P. Appl. Environ. Microbiol. 1996, 62,
4086.
9. Bourne, D. G.; Riddles, P.; Jones, G. J.; Smith, W.; Blake-
ley, R. L. Environ. Toxicol. 2001, 16, 523.
10. Gulledge, B. M.; Aggen, J. B.; Huang, H. B.; Nairn, A. C.;
Chamberlin, A. R. Curr. Med. Chem. 2002, 9, 1991.
11. Pearson, C.; Rinehart, K. L.; Sugano, M.; Costerison,
J. R. Org. Lett. 2000, 2, 2901.
1
12. 2: H NMR (500 MHz, MeOH-d4) d 1.08 (dd, J=1.2, 6.5
The Adda-containing analogues described herein, rang-
ing from di- to penta-peptides with the sequence found
in microcystins, exhibit remarkably similar activities.
Their inhibitory potencies towards PP1 and PP2A have
been found to be only moderate-to-mediocre, contra-
dicting a literature report of potent inhibition by the
acyclic tetrapeptide microcystin degradation product.
Whatis notknown is why there is no significantchange
in activity between the di-, tetra-, and pentapeptides.
The three variable residues could bind in a similar
manner to microcystin, with the beneficial binding con-
tacts from each additional residue offset by the entropic
cost of restraining the conformation for binding. Alter-
natively, the residues may not make any contacts, and
extend away from the enzyme. Further experiments are
being conducted in our laboratory to address this issue.
Hz, 3H), 1.32 (s, 3H), 1.50 (s, 3H), 1.71 (dd, J=1.1, 3.5 Hz,
3H), 2.01–2.06 (m, 1H), 2.19–2.24 (m, 1H), 2.31–2.36 (m, 1H),
2.49 (m, 1H), 2.67–2.78 (m, 5H), 2.85 (dd, J=4.6, 13.8 Hz,
1H), 3.10 (br s, 2H), 3.24–3.29 (m, 1H), 3.28 (s, 3H), 3.35 (s,
3H), 3.95 (dd, J=8.5, 14.6 Hz, 1H), 4.43–4.47 (m, 1H), 4.52
(app q, J=7.3 Hz, 1H), 5.54 (dd, J=8.8, 15.5 Hz, 1H), 5.63 (d,
J=10.1 Hz, 1H), 5.77 (br s, 1H), 6.23 (br s, 1H), 6.52 (dd,
J=1.8, 15.6 Hz, 1H), 7.18–7.24 (m, 4H), 7.30 (t, J=7.4 Hz,
3H).
13. Meriluoto, J. A. O.; Nygard, S. E.; Dahlem, A. M.;
Eriksson, J. E. Toxicon 1990, 28, 1439.
14. Takai, A.; Mieskes, G. Biochem. J. 1991, 275, 233.
15. The time dependence of the activity–concentration curves
shows no changes with inhibitor-enzyme incubation periods
ranging from 5 min to 1 h for microcystin-LR (data not
shown); the time dependence for the acyclic analogues is cur-
rently being determined.