Transition state analogy of phosphonic acid peptide inhibitors of pepsin

Article abstract of DOI:10.1021/jo952074c

A series of 11 phosphonate peptide analogs, RO₂C-Xaa-Yaa-Phe-{PO₂^--O}-Phe O-(3-(4-pyridyl)propyl ester), were synthesized and evaluated as inhibitors of the aspartic peptidase pepsin. For the inhibitors with Gly or Ala at the P₂ position, the K(i) values correlate with the K(m)/k(cat) values of the corresponding substrates, demonstrating that these analogs mimic the transition state in the way the P₂-P₄ residues bind. Deviations from the correlation for the other inhibitor/substrate pairs imply a different binding orientation as a result of N-substitution at P₂ (Pro), contamination with the more potent diastereomer (D-Ala), or a change in rate-limiting step for turnover (Ile).

Full text of DOI:10.1021/jo952074c

J . Org. Chem. 1996, 61, 3433-3438
3433
Tr a n sition Sta te An a logy of P h osp h on ic Acid P ep tid e In h ibitor s of
P ep sin
Paul A. Bartlett* and Mark A. Giangiordano
Department of Chemistry, University of California, Berkeley, California 94720-1460
Received November 27, 1995X
A series of 11 phosphonate peptide analogs, RO2C-Xaa-Yaa-Phe-{PO2--O}-Phe O-(3-(4-pyridyl)-
propyl ester), were synthesized and evaluated as inhibitors of the aspartic peptidase pepsin. For
the inhibitors with Gly or Ala at the P2 position, the Ki values correlate with the Km/kcat values of
the corresponding substrates, demonstrating that these analogs mimic the transition state in the
way the P2-P4 residues bind. Deviations from the correlation for the other inhibitor/substrate
pairs imply a different binding orientation as a result of N-substitution at P2 (Pro), contamination
with the more potent diastereomer (D-Ala), or a change in rate-limiting step for turnover (Ile).
Sch em e 1. Asp a r tic P ep tid a se Mech a n ism
The aspartic peptidases constitute one of the primary
classes of proteolytic enzymes, utilizing two aspartic acid
residues in the active site to catalyze the direct addition
of water to the carbonyl of the scissile linkage.1-3 Early
investigations of pepsin and the fungal acid proteinases4,5
gave rise to a variety of mechanistic proposals which were
resolved in favor of the sequence shown in Scheme 1.6,7
As a result of attempts to develop renin inhibitors as
antihypertensive agents8-10 and continuing efforts to
exploit these and related inhibitor motifs for inactivation
of the HIV protease,11,12 there has been ongoing interest
in understanding how these enzymes work and how they
can be inhibited. Peptide analogs that incorporate
tetrahedral phosphorus units in place of the scissile
carbonyl group of a substrate are potent inhibitors of a
variety of aspartic peptidases.13-18 The activity of these
derivatives has been understood to result from “transition
state analogy”: their mimicry of the high-energy, tetra-
hedral intermediate that lies along the reaction pathway
(Scheme 1).15 It was on this basis that such analogs were
first studied, and the viewpoint has been supported by
structural analysis of the complexes of these inhibitors
with penicillopepsin19 and with HIV protease.20
However, potent inhibition does not of itself demon-
strate that a molecule is a transition state analog.21-25
Such evidence is obtained by showing that structural
alterations produce similar effects on inhibitor binding
(reflected in Ki) as they do on transition state stabiliza-
tion, as inferred from the inverse second order rate
X Abstract published in Advance ACS Abstracts, April 1, 1996.
(1) Fruton, J . S. Adv. Enzymol. Relat. Areas Mol. Biol. 1976, 44,
1-36.
22
constant Km/kcat
.
This approach, first proposed by
Wolfenden,26 and by Thompson,27 has been applied suc-
cessfully in assessing the transition state analogy of
phosphorus-containing inhibitors of the zinc peptidases
thermolysin28,29 and carboxypeptidase A,21,30 among other
enzymes.31,32 We have recently described a complemen-
(2) Dunn, B. M. Adv. Detailed React. Mech. 1992, 2, 213-241.
(3) Davies, D. R. Annu. Rev. Biophys. Biophys. Chem. 1990, 19, 189-
215.
(4) Wang, T.-T.; Hofmann, T. Biochem. J . 1976, 153, 691-699.
(5) J ames, M. N. G.; Sielecki, A. R. Biochemistry 1985, 24, 3701-
3713.
(6) Suguna, K.; Padlan, E. A.; Smith, C. W.; Carlson, W. D.; Davies,
D. R. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 7009-7013.
(7) J ames, M. N. G.; Sielecki, A. R.; Hayakawa, K.; Gelb, M. H.
Biochemistry 1992, 31, 3872-3886.
(19) Fraser, M. E.; Strynadka, N. C. J .; Bartlett, P. A.; Hanson, J .
E.; J ames, M. N. G. Biochemistry 1992, 31, 5201-5214.
(20) Abdel-Meguid, S. S.; Zhao, B.; Murthy, K. H. M.; Winborne,
E.; Choi, J .-K.; DesJ arlais, R. L.; Minnich, M.; Culp, J . S.; Debouck,
C.; Tomaszek, T. A., J r.; Meek, T. D.; Dreyer, G. B. Biochemistry 1993,
32, 7972-7980.
(8) Lien, E. J .; Gao, H.; Lien, L. L. Prog. Drug Res. 1994, 43, 43-
86.
(9) Raddatz, P. Expert Opin. Ther. Pat. 1994, 4, 1347-1359.
(10) Raddatz, P. Expert Opin. Ther. Pat. 1994, 4, 489-504.
(11) Thaisrivongs, S. Annu. Rep. Med. Chem. 1994, 29, 133-144.
(12) Huff, J . R. J . Med. Chem. 1991, 34, 2305-2314.
(13) Allen, M. C.; Fuhrer, W.; Tuck, B.; Wade, R.; Wood, J . M. J .
Med. Chem. 1989, 32, 1652-1661.
(14) Dreyer, G. B.; Metcalf, B. W.; Tomaszek, T. A., J r.; Carr, T. J .;
Chandler, A. C., III; Hyland, L.; Fakhoury, S. A.; Magaard, V. W.;
Moore, M. L.; Strickler, J . E.; Debouck, C.; Meek, T. D. Proc. Nat. Acad.
Sci. U.S.A. 1989, 86, 9752-9756.
(15) Bartlett, P. A.; Hanson, J . E.; Giannousis, P. P. J . Org. Chem.
1990, 55, 6268-6274.
(16) Grobelny, D.; Wondrak, E. M.; Galardy, R. E.; Oroszlan, S.
Biochem. Biophys. Res. Commun. 1990, 169, 1111-1116.
(17) Peyman, A.; Budt, K.-H.; Spanig, J .; Stowasser, B.; Ruppert,
D. Tetrahedron Lett. 1992, 33, 4549-4552.
(18) Raddatz, P.; Minck, K.-O.; Rippmann, F.; Schmitges, C. J . J .
Med. Chem. 1994, 37, 486-497.
(21) Hanson, J . E.; Kaplan, A. P.; Bartlett, P. A. Biochemistry 1989,
28, 6294-6305.
(22) Wolfenden, R. Annu. Rev. Biophys. Bioeng. 1976, 5, 271-306.
(23) Wolfenden, R.; Radzicka, A. Curr. Opin. Struct. Biol. 1991, 1,
780-787.
(24) Morrison, J . F.; Walsh, C. T. Adv. Enzymol. Relat. Areas Mol.
Biol. 1987, 61, 201-301.
(25) Schloss, J . V. Acc. Chem. Res. 1988, 21, 348-353.
(26) Westerik, J . O.; Wofenden, R. J . Biol. Chem. 1971, 247, 8195-
8197.
(27) Thompson, R. C. Biochemistry 1973, 12, 47-51.
(28) Bartlett, P. A.; Marlowe, C. K. Biochemistry 1983, 22, 4618-
4624.
(29) Morgan, B. P.; Scholtz, J . M.; Ballinger, M.; Zipkin, I.; Bartlett,
P. A. J . Am. Chem. Soc. 1991, 113, 297-307.
(30) Phillips, M. A.; Kaplan, A. P.; Rutter, W. J .; Bartlett, P. A.
Biochemistry 1992, 31, 959-962.
S0022-3263(95)02074-3 CCC: $12.00 © 1996 American Chemical Society

3434 J . Org. Chem., Vol. 61, No. 10, 1996
Bartlett and Giangiordano
Ta ble 1. Kin etic P a r a m eter s for Su bstr a tes, 1, a n d
In h ibition Con sta n ts for P h osp h on a te In h ibitor s, 2, of
P or cin e P ep sin a
Sch em e 2. Syn th esis of P h osp h on a te In h ibitor s
RO2C-Xaa-Yaab Km (µM) kcat (s-1
)
kcat/Km (M-1 s-1
)
Ki (nM)
ZGP
140
350
510
170
750
420
70
110
250
110
40
0.056
0.19
2.1
3.5
19
72
12.6
25
145
410
280
400
560
4 000
21 000
25 000
260
84
570
200
67
17
0.31
12
ZG-D-A
Z-D-AG
ZSarG
MocGG
ZGG
ZGI
ZFG
ZAG
ZGA
170 000
180 000
225 000
580 000
3 700 000
7 000 000
4.2
0.73
0.34
ZAA
a
Kinetic data from Sachdev and Fruton at pH 3.5, 37 °C.34
Inhibition constants were determined under the same conditions
using Lys-Pro-Ala-Glu-Phe-(4-NO2)Phe-Arg-Leu as substrate.35
b
One-letter amino acid codes used; Z ) Cbz, Moc ) methoxycar-
bonyl, Sar ) sarcosyl.
tary correlation of inhibitor Ki versus substrate Km values
as evidence of “ground state analogy”.33 In this report,
we provide evidence for the transition state analogy of
phosphonate inhibitors of porcine pepsin.
The validity of the Ki versus Km/kcat correlation depends
on several key factors:28 (1) the chemical step whose
transition state is mimicked by the inhibitors is rate
limiting for the enzymatic reaction, (2) the structural
variation among the substrates does not alter their
intrinsic reactivity (i.e., knoncat is constant for the series),
and (3) the invariant regions of the ligands bind to the
enzyme in the same fashion across the series. The
importance of the latter point was demonstrated in a
recent study with phosphonate inhibitors of several
carboxypeptidase A mutants, where parallel (rather than
superimposed) correlation lines result from different
binding modes for the transition states for acyl-tripep-
tides and acyl-Gly-Xaa dipeptides.
acid chloride with oxalyl chloride36 and condensation with
the alcohol. The phosphinic acid analog of Cbz-Phe was
synthesized and resolved as described by Baylis et al.,37
converted to the methyl ester 7,38 and oxidized with CCl4
in the presence of water and triethylamine39 to give the
phosphonate monoester 8. This procedure proved to be
an operationally simpler route to this intermediate than
an alternative sequence via the diphenyl phosphonate
analog of Cbz-Phe.40-42 Activation of 8 with thionyl
chloride43 and condensation with the hydroxy ester 6 in
the presence of triethylamine gave phosphonate diester
9 in 87% yield. Alternatively, the phosphinate ester 7
could be oxidized with carbon tetrachloride in the pres-
We selected 11 substrates of the form RO2C-Xaa-Yaa-
Phe-Phe 3-(4-pyridyl)propyl ester, 1, which are cleaved
between the phenylalanine residues with a broad range
of Km and kcat values (Table 1).34 These substrates differ
in the S2-S4 positions, sufficiently remote from the Phe-
Phe linkage to justify the assumption that knoncat is
invariant. In addition to a wide range in kcat/Km values,
we selected this set of substrates to minimize any
correlation between Km and Km/kcat to allow a clear
distinction between ground state and transition state
analogy.
(31) Mookthiar, K. A.; Grobelny, D.; Galardy, R. D.; Van Wart, H.
E. Biochemistry 1988, 27, 4299-4304.
(32) Rahil, J .; Pratt, R. F. Biochemistry 1994, 33, 116-125.
(33) Bartlett, P. A.; Otake, A. J . Org. Chem. 1995, 60, 3107-3111.
(34) Sachdev, G. P.; Fruton, J . S. Biochemistry 1970, 9, 4465-4470.
(35) Dunn, B. M.; Valler, M. J .; Rolph, C. E.; Foundling, S. I.;
J imenez, M.; Kay, J . Biochim. Biophys. Acta 1987, 913, 122-130.
(36) Wissner, A.; Grudzinskas, C. V. J . Org. Chem. 1978, 43, 3972-
3974.
(37) Baylis, E. K.; Campbell, C. D.; Dingwall, J . G. J . Chem. Soc.,
Perkin Trans. 1 1984, 2845-2852.
(38) Karanewsky, D. S.; Badia, M. C. Tetrahedron Lett. 1986, 27,
1751-1754.
(39) Huber, J . W., III; Gilmore, W. F. J . Med. Chem. 1975, 18, 106-
108.
(40) Oleksyszyn, J .; Subotkowska, L.; Mastarlerz, P. Synth. Com-
mun. 1979, 985-986.
The analogous phosphonates, 2, were synthesized as
shown in Scheme 2. The 3-(4-pyridyl)propyl ester of L-â-
phenyllactate, 6, was prepared from the bis(tert-bu-
tyldimethylsilyl) derivative 3 by direct conversion to the
(41) Kafarski, P.; Leijczak, B.; Szewczyk, J . Can. J . Chem. 1983,
61, 2425-2430.
(42) Bartlett, P. A.; Lamden, L. A. Bioorg. Chem. 1986, 14, 356-
377.

Phosphonic Acid Peptide Inhibitors of Pepsin
J . Org. Chem., Vol. 61, No. 10, 1996 3435
F igu r e 1. Correlation of substrate kinetic parameters and phosphonate inhibition constants.
ence of the hydroxy ester 6 and triethylamine to give
diester 9 directly.44 The Cbz protecting group was
removed from 9 by hydrogenolysis, and the amine was
acylated with the appropriate dipeptide activated with
isobutyl chloroformate. Finally, the phosphonate methyl
ester was cleaved nucleophilically in refluxing tert-
butylamine,45 and the inhibitor 2 was purified by chro-
matography on DEAE Sephadex and isolated as the
lithium salt after ion exchange with Dowex-Li+ and
lyophilization.
Inhibition constants for the 11 analogs of 2 were
measured with the chromogenic substrate Lys-Pro-Ala-
Glu-Phe-(4-NO2)Phe-Arg-Leu35 under the same condi-
tions utilized previously for determination of the sub-
strate kinetic constants, pH 3.5 at 37 °C.34 Simple Dixon
analyses (eq 1)46,47 were used to determine the inhibition
constants for the weaker inhibitors (Ki > 10 nM), while
the method of Henderson (eq 2)48 was employed for the
the corresponding substrates in Figure 1A and B, re-
spectively. There is no meaningful correlation between
the Ki values (which span more than 3 orders of magni-
tude) and the Km values (which are within a factor of 20).
In contrast, for eight of the 11 inhibitor/substrate pairs,
there is a strong correlation between Ki and Km/kcat; linear
regression gives a slope of 1.00 and a correlation coef-
ficient R of 0.995. For these eight inhibitors, the P2-P4
residues bind to the enzyme active site the same way they
do in the transition state complexes and distinctly
differently than they do in the ground state, Michaelis
complexes. This correlation is logically the result of the
tetrahedral geometry of the phosphonate linkage, which
conveys a different orientation to the remote side chains
in the active site than the trigonal amide does. The
potent binding of the Cbz-Ala-Ala inhibitor is therefore
a manifestation of the remote substituent effects that are
known to be important for activity of the aspartic
peptidases.49
v0
vi
[I]
[S]
Km
) 1 +
1 +
(1)
(2)
A surprising aspect to the correlation is the broad
range over which it is observed: from substrates with
turnover rates of 4 × 103 M-1 s-1 (Cbz-D-Ala-Gly-) to 7 ×
106 M-1 s-1 (Cbz-Ala-Ala). As the second-order rate
constant approaches 107 M-1 s-1, the diffusion-limited
association step becomes increasingly rate-limiting for a
peptidase, and the “transition state” is no longer the
tetrahedral species whose geometry is mimicked by the
phosphonates. Tighter-than-expected binding of phos-
phonate analogs of very good substrates (kcat/Km ≈ 106
M-1 s-1) has indeed been observed for the zinc peptidases
thermolysin50 and carboxypeptidase A.51 The absence of
such a deviation for the Cbz-Ala-Ala and Cbz-Gly-Ala
pairs in this study suggests that the chemical step
remains rate-limiting for these substrates. This conclu-
sion differs from that of Sachdev and Fruton, who found
that nonchemical events may become kinetically signifi-
(
)
Ki
It
v
[S]
Km
) Et + 0Ki 1 +
(
)
vi
vi
1 -
(
)
v0
more potent analogs where inhibitor concentrations on
the order of [E]t had to be used. A full kinetic analysis
of the Cbz-D-Ala-Gly analog demonstrated that it binds
as a competitive inhibitor.46 The more potent inhibitors
were incubated with the enzyme for 5 min prior to
addition of the substrate to overcome the slow binding
step;24,25 under these circumstances, the [S]/Km term was
dropped from the analysis, since the E‚I complex was not
in equilibrium with the substrate during the timescale
of the assay.
The inhibition constants are given in Table 1 and are
compared graphically with the Km and Km/kcat values of
cant for substrates with kcat values greater than 20 s-1 52
.
Moreover, such behavior contrasts sharply with that
(43) Hirschmann, R.; Yager, K. M.; Taylor, C. M.; Moore, W.;
Sprengeler, P. A.; Witherington, J .; Phillips, B. W.; Smith, A. B., III.
J . Am. Chem. Soc. 1995, 117, 6370-6371.
(44) Sampson, N. S.; Bartlett, P. A. J . Org. Chem. 1988, 53, 4500-
4503.
(45) Gray, M. D. M.; Smith, D. J . H. Tetrahedron Lett. 1980, 21,
859-862.
(46) Segel, I. H. Enzyme Kinetics; J ohn Wiley and Sons: New York,
1975.
(49) Medzihradszky, K.; Voynick, I. M.; Medzihradszky-Schweiger,
H.; Fruton, J . S. Biochemistry 1970, 9, 1154-1162.
(50) Bartlett, P. A.; Marlowe, C. K.; Giannousis, P. P.; Hanson, J .
E. In Cold Spring Harbor Symposia on Quantitative Biology; Cold
Spring Harbor Laboratory: Cold Spring Harbor 1987; Vol. LII; pp 83-
90.
(51) Kaplan, A. P.; Bartlett, P. A. Biochemistry 1991, 30, 8165-
8170.
(47) Dixon, M. Biochem. J . 1972, 129, 197-202.
(48) Henderson, P. J . F. Biochem. J . 1972, 127, 321-333.
(52) Sachdev, G. P.; Fruton, J . S. Proc. Natl. Acad. Sci. U.S.A. 1975,
72, 3424-3427.

3436 J . Org. Chem., Vol. 61, No. 10, 1996
Bartlett and Giangiordano
observed with smaller peptides, whose anomalous kinetic
properties and cleavage and recombination patterns have
been explained by slow binding of substrates and release
of products.1,53,54 However, Pohl and Dunn have pre-
sented evidence that Km represents the true dissociation
the rate-limiting step for turnover of this substrate is not
formation of the tetrahedral intermediate. Hydrophobic
peptides and products are known to stick to pepsin, and
the anomalously low turnover rate of the Cbz-Gly-Ile
peptide may simply be the result of slow dissociation of
product after the cleavage step.
constant even for substrates with kcat > 100 s-1 55
.
Con clu sion . The demonstration of transition state
analogy for a closely related series of phosphonate pepsin
inhibitors indicates that the binding orientation of these
derivatives can serve as a model for that of the transition
state. As such, they may help to elucidate the structural
basis for remote substituent effects on catalysis. An
understanding of these effects would be especially sig-
nificant because of the structural and mechanistic simi-
larities between pepsin, renin, and the fungal and viral
aspartic peptidases.
Since eight of the derivatives appear to be transition
state analogs, why aren’t the other three? The points
for the Cbz-Gly-Pro, Cbz-Gly-D-Ala, and Cbz-Gly-Ile
analogs deviate significantly from the correlation line;
these phosphonates are bound 20-50 times more tightly
than predicted by the turnover rate of the corresponding
substrates. Interestingly, these are the three analogs
with amino acids other than glycine or L-alanine at the
P2 position. The enhanced binding of the Cbz-Gly-D-Ala
analog (observed Ki ) 84 nM) is probably due to con-
tamination from the L-Ala diastereomer (Ki ) 0.73 nM)
formed during the coupling step (10 f 11); less than 1%
of this material would account for the observed deviation.
Exp er im en ta l Section 58
3-(4-P yr id yl)p r op yl (2S)-2-[[[(1R)-1[N-[N-[[(P h en yl-
m eth oxy)ca r bon yl]glycyl]-L-a la n yl]a m in o]-2-p h en yleth -
yl]h ydr oxyph osph in yl]oxy]-3-ph en ylpr opan oate, Lith iu m
Sa lt (2-GA, R2 ) Me, R3 ) H, R ) Bn ). A solution of 120
mg (0.16 mmol) of the methyl phosphonate 10-GA (see below)
in 10 mL of freshly-distilled tert-butylamine in the presence
of 4 Å molecular sieves was heated at reflux for 14 h. The
solvent was evaporated under a stream of N2 and under high
vacuum, and the residue was dissolved in 1:1 CH2Cl2/methanol
(1:1) and filtered through a pad of Celite. The filtrate was
concentrated to a minimum volume of solution and loaded onto
Less obvious are the implications from the binding
behavior of the P2-proline and the Cbz-Gly-Ile derivatives.
Peptide substrates and inhibitors with proline at the P2
position are likely to bind in a different orientation than
peptides with an unsubstituted NH. Structural studies
of a variety of aspartic peptidases, including pepsin,
reveal a favorable hydrogen-bonding interaction be-
tween this amide hydrogen and an active site carbonyl
group.3,19,56,57 Disruption of this interaction by introduc-
tion of an N-substituted amino acid at the P2 position
may alter the orientation of the Phe{X-Y}Phe-OR resi-
dues in the active site and thus explain why this point
does not fit the correlation. This deviation does not imply
that this phosphonate is not a mimic of the transition
state, simply that the transition state for the proline-
containing substrate is bound in a different orientation
than the others; similar behavior has been seen with
phosphonate inhibitors and substrates of carboxypepti-
dase A.30 In contrast, an N-methyl group on the P3
residue, in the Cbz-Sar-Gly analog, does not disturb the
binding orientation because the active-site contact in
pepsin is a more flexible threonine side chain at the
protein surface.57
-
a 20 × 1 cm column of DEAE-Sephadex (HCO3 form) and
eluted with a linear gradient of 0-0.25 M triethylammonium
bicarbonate buffer (pH 7.5; this eluting solvent contained up
to 50% methanol depending on the analog). Fractions were
monitored by UV absorbance, and those containing product
were pooled and lyophilized to give an off-white, hygroscopic
powder. This material was dissolved in water (with up to 50%
methanol, again depending on the analog), passed through a
1 × 5 cm column (50 mequiv) of Dowex in the Li+ form, and
lyophilized to give 22 mg (18% yield) of 2-GA, lithium salt, as
a white powder: 1H NMR (D2O) δ 0.84 (d, 3, J ) 7.1 Hz), 1.6
(m, 2), 2.32 (t, 2, J ) 7.6 Hz), 2.45 (m, 1), 2.9 (m, 3), 3.56 (s,
2), 3.82 (m, 2), 4.1 (m, 2), 4.7 (q, 1), 4.9 (s, 2), 6.90-7.20 (m,
17), 8.15 (d, 2, J ) 5.3 Hz); 31P NMR (D2O) δ 18.17; FAB MS
737 (MH+); HRMS (FAB) calcd for C38H42N4LiO9P + H m/z
737.2928, found 737.2938 (MH+). Anal. Calcd for C38H42N4-
LiO9P: P, 4.21. Found: P, 3.91 (weight purity ) 93%).
3-(4-P yr id yl)p r op yl O-(ter t-Bu tyld im eth ylsilyl)-(2S)-3-
p h en ylla cta te (5). To a solution of 2.6 g (16 mmol) of L-â-
phenyllactic acid and 10 g (67 mmol) of tert-butyldimethylsilyl
chloride in 40 mL of CH2Cl2 at 0 °C was added 10 mL (67
mmol) of DBU via syringe. After 20 min, TLC analysis
indicated that the diacid had been completely silylated; the
reaction mixture was washed twice with 70-mL portions of
saturated NH4Cl, once with water, and once with brine. The
organic layer was dried (MgSO4), two drops of DMF from a
disposable pipet were added, and the pale yellow solution was
added to a flame-dried flask containing 4 Å molecular sieves
and cooled in an ice bath. Freshly-distilled oxalyl chloride (5.0
mL, 57 mmol) was added carefully via syringe, with evolution
The deviation observed for the Cbz-Gly-Ile analog is
more difficult to understand: the inhibition constant of
31 nM suggests that pepsin should be capable of hydro-
lyzing the corresponding peptide substrate at a rate of 8
× 106 M-1 s-1, rather than the observed rate of 1.8 × 105
M-1 s-1. Indeed, the corresponding Cbz-Gly-Leu analog
is hydrolyzed at a rate of 4 × 106 M-1 s-1 34
.
Thinking
that the reported kinetic parameters for the Cbz-Gly-Ile
analog may be incorrect, we resynthesized the peptide;
however, our kinetic analysis did not reveal any mean-
ingful errors in the reported Km or kcat values. Nor did
examination of a model of the enzyme-substrate complex
suggest any reason why the isoleucine side chain could
not be accommodated sterically in the S2 pocket.57 The
lack of transition state analogy in this case implies that
(58) Unless otherwise stated, all reactions were run in base-washed,
oven-dried round-bottom flasks containing Teflon-coated stir bars,
under house nitrogen atmosphere. Dichloromethane, THF, and tri-
ethylamine were distilled from CaH2 prior to use. Unless otherwise
indicated, NMR spectra were taken in CDCl3; phosphorus chemical
shifts are given with reference to an internal capillary containing
trimethyl phosphite (δ 3.086) in CDCl3. FAB mass spectra were
performed using a matrix of thioglycerol/glycerol unless otherwise
indicated. Silica gel column chromatography was performed according
to Still59 using the solvent system specified. Triethylammonium
bicarbonate buffer (“TBK”) was prepared by treating a solution of
triethylamine in H2O with CO2 gas until a pH of 7.5 was achieved.
(59) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.
(53) Silver, M. S.; J ames, S. L. T. Biochemistry 1981, 20, 3183-
3189.
(54) Silver, M. S.; J ames, S. L. T. Biochemistry 1981, 20, 3177-
3182.
(55) Pohl, J .; Dunn, B. Biochemistry 1988, 27, 4827-4834.
(56) Lunney, E. A.; Hamilton, H. W.; Hodges, J . C.; Kaltenbronn,
J . S.; Repine, J . T.; Badasso, M.; Cooper, J . B.; Dealwis, C.; Wallace,
B. A.; Lowther, W. T.; Dunn, B. M.; Humblet, C. J . Med. Chem. 1993,
36, 3809-3820.
(57) Fujinaga, M.; Chernaia, M. M.; Tarasova, N.; Mosimann, S. C.;
J ames, M. N. G. Protein Sci. 1995, 4, 960-972.

Phosphonic Acid Peptide Inhibitors of Pepsin
J . Org. Chem., Vol. 61, No. 10, 1996 3437
of gas. After 2 h, the clear yellow solution was evaporated in
vacuo, 20 mL of benzene was added and the solution was
lyophilized. The residue (4) was dissolved in 10 mL of CH2-
Cl2 and added via cannula to a flask containing 6.4 g (47 mmol)
of 3-(4-pyridyl)propanol and 190 mg (1.6 mmol) of DMAP in
30 mL of CH2Cl2 cooled in a dry ice/acetone bath. After 14 h,
the solvent was evaporated in vacuo, the residue was dissolved
in ether, and the solution was filtered through a 0.5-in. pad of
silica gel and evaporated in vacuo to give 6.0 g of 3 as a clear,
colorless oil which can be purified by flash chromatography
(70% ether/hexanes f ether). This material proved difficult
to separate from the silyl ether of 3-(4-pyridyl)-1-propanol, so
the crude product was carried on to the next step directly,
when 6 can be separated easily from the alcohol by column
chromatography: 1H NMR δ -0.21 (s, 3), -0.11 (s, 3), 0.80 (s,
9), 1.90-1.98 (m, 2), 2.59-2.62 (m, 2), 2.97 (ddd, 2, J ) 66.8,
13.4, 4.2 Hz), 4.10-4.13 (m, 2), 4.34 (dd, 1, J ) 8.6, 4.3 Hz),
7.07 (dapp, 2), 7.20-7.30 (m, 5), 8.49 (dapp, 2); 13C NMR δ -5.87,
-5.55, 17.88, 25.34, 28.76, 31.12, 41.38, 63.42, 73.41, 123.41,
126.30, 127.85, 129.45, 136.91, 149.49, 149.55, 172.60; FAB
MS 400 (MH+). Anal. Calcd for C23H33NO3Si: C 66.16; H 8.27,
N 3.51. Found: C 66.21, H 7.77, N 3.31.
66.43 and 66.57, 74.80 (J ) 7 Hz) and 75.27 (J ) 7.5 Hz),
123.71, 126.55-129.49 (17 peaks), 135.30 and 135.46, 136.38
and 136.56, 136.71 and 136.85, 136.94, 149.62 and 149.79,
155.80 and 155.91, 169.87, 170.34; 31P NMR (from a prep with
racemic 8) δ 25.5, 26.2, 26.5, 27.0; FAB MS Calcd for
C34H37N2O7P + H m/z ) 617, found m/z 617.1 (MH+). This
material proved to be unstable on storage and was therefore
used directly in subsequent reactions.
Alter n a tive P r ep a r a tion of P h osp h on a te 9: Oxid a tive
Cou p lin g of 6 a n d 7. Meth yl (1R)-1-[(Ben zyloxyca r bon -
yl)a m in o]-2-p h en yleth yl p h osp h in a te (7). A solution of 3.2
g (10 mmol) of the phosphinic acid,37 4.0 mL (100 mmol) of
methanol, and 2.1 g (11 mmol) of EDC in 50 mL of dichlo-
romethane was stirred for 16 h, diluted with ethyl acetate,
washed with saturated NaH2PO4 (pH 5) and saturated NaH-
CO3, dried (MgSO4), and evaporated to give 2.46 g (74%) of
methyl phosphinate 7 as a waxy solid: 1H NMR δ 2.95 (m, 1),
3.2 (m, 1), 3.73 and 3.76 (2d, Σ ) 3, J ) 5.0 Hz), 4.25 (m, 1),
5.0 (2br s, Σ ) 2), 5.27 and 5.42 (2br d, Σ ) 1, J ) 9 Hz), 7.02
and 7.04 (2 dd, Σ ) 1, J ) 560 Hz), 7.1-7.4 (m, 10); 31P NMR
δ 36.32, 36.53.
A suspension of 53 mg (160 µmol) of methyl phosphinate 7,
84 mg (300 µmol) of hydroxy ester 6, and 4 Å molecular sieves
in 1 mL of CCl4 under nitrogen was cooled to -10 °C, 50 µL
(0.59 mmol) of triethylamine was added, and the mixture was
stirred at -10 °C for 18 h. The mixture was filtered and
evaporated, and the residue was chromatographed to give 66
mg (72% yield) of phosphonate 9.
3-(4-P yr id yl)p r op yl (2S)-2-[[(1R)-1-[[N-[N-[(P h en yl-
m eth oxyca r bon yl)glycyl]-L-a la n yl]a m in o]-2-p h en yl-eth -
yl]m et h oxyp h osp h in yl]oxy]-3-p h en ylp r op a n oa t e (10-
GA). A solution of 600 mg (ca. 1 mmol) of phosphonate 9 in
30 mL of ethanol was prepared, the reaction vessel was purged
with N2, and 70 mg of 20 wt % Pd(OH)2 on carbon was added
under a stream of N2. A hydrogen balloon was then attached,
and the progress of the hydrogenolysis was followed by TLC.
After 1.5 h, the mixture was filtered through Celite, evapo-
rated, and placed under high vacuum for 1 h to remove any
residual ethanol. Half of this material was carried through
the following coupling procedure.
3-(4-P yr id yl)p r op yl (2S)-3-P h en ylla cta te (6). To a solu-
tion of 6.0 g (<15 mol) of the crude silyl ether 5 in 70 mL of
THF in an ice bath was added 15 mL of a 1.0-M solution of
tetra-n-butylammonium fluoride in THF. After 20 min, TLC
(5% methanol/ether) indicated that the reaction had gone to
completion, so 1.5 mL (25 mmol) of acetic acid was added,
followed by pyridine (1.5 mL), and the reaction mixture was
partitioned between ether and brine. The organic phase was
dried (MgSO4), and the solvent was evaporated in vacuo to
give 5.0 g of crude alcohol which was purified by column
chromatography (silica gel eluted with a gradient of ether f
5% methanol/ether) to give 2.4 g (54% overall from L-â-
phenyllactic acid) of the hydroxy ester 6: 1H NMR δ 1.93-
2.00 (m, 2), 2.60-2.64 (m, 2), 3.05 (ddd, 2, J ) 50.2, 13.9, 4.8
Hz), 4.11-4.22 (m, 2), 4.45 (dd, 1, J ) 6.7, 4.9 Hz), 7.00 (d, 2,
J ) 4.9 Hz), 7.13-7.35 (m, 5), 8.49 (d, 2, J ) 4.9 Hz); 13C NMR
δ 28.37, 30.84, 40.40, 49.56, 63.62, 71.23, 123.48, 126.26,
127.90, 128.73, 129.05, 136.47, 148.85, 150.05, 173.73; FAB
A solution of 140 mg (0.5 mmol) of Cbz-Gly-Ala-OH was
prepared in 3 mL of dry acetonitrile over 4 Å molecular sieves,
and the suspension was cooled to -10 °C. Triethylamine (70
µL, 0.5 mmol) and 65 µL (0.5 mmol) of isobutyl chloroformate
were added, followed after 3 min by a solution of the amine in
acetonitrile, and the suspension was stirred for 15 h. Flash
grade silica gel was added to the reaction mixture, and the
solvent was removed in vacuo to give a silica gel composite
which was loaded onto a column of 10 g of plate-grade silica
packed into a 30-mL fritted filter funnel. Elution was per-
formed using 40% THF/ethyl acetate to give 125 mg (34%) of
the coupled product (10-GA) as a mixture of two diastereo-
mers: 1H NMR δ 1.20-1.30 (m, 3), 1.89-2.22 (m, 2), 2.57 and
2.64 (2 t, Σ ) 2, J ) 7.6 Hz), 2.88-2.95 (m, 1), 3.04-3.32 (m,
5), 3.20 (d, J ) 11 Hz), 3.74-3.80 (m, 3), 3.78 (d, J ) 11 Hz),
4.04 and 4.24 (2t, Σ ) 2, J ) 6.4 Hz), 4.40-4.46 (m, 1), 5.24
(dt, 1, J ) 4, 9 Hz), 5.68-5.76 (m, 1), 6.90-7.46 (m, 19), 8.49-
8.51 (br, 2); 31P NMR δ 26.11 and 26.24; IR (CDCI3) 3420 (wk),
MS 286 (MH+); [R]25 ) -16.1° (c ) 0.012 in CHCl3); HRMS
D
(FAB) calcd for
C17H19NO3 + H m/z ) 286.1443, found
286.1442 (MH+).
3-(4-P yr id yl)p r op yl 2-[[(1R)-1-[N-[(P h en ylm et h oxy)-
car bon yl]am in o]-2-ph en yleth yl)m eth oxyph osph in yl]oxy]-
3-p h en ylp r op a n oa te (9). To a solution of 1.50 g (4.28 mmol)
of the resolved phosphonic acid 842 in 20 mL of dichlo-
romethane was added 570 mL (6.42 mmol) of thionyl chloride
via syringe, causing the cloudy solution to become clear. After
1 h, an aliquot was removed for 31P NMR analysis (CH2Cl2,
82 MHz), which revealed that complete conversion to the acid
chloride (δ 41.04, 42.07) had occurred; a trace of the dichloride
(δ 50.98) was also present. The solvent was evaporated under
a stream of nitrogen, 10 mL of benzene was added, and the
solution was lyophilized to give a white solid. A solution
containing 1.50 g (5.14 mmol) of hydroxy ester 6, 1.0 mL (7.2
mmol) of triethylamine, and 50 mg (0.43 mmol) of DMAP in
20 mL of dichloromethane dried over 4 Å molecular sieves was
added via cannula to the phosphonochloridate cooled in a dry
ice/acetone bath. After being stirred for 20 h, the reaction
mixture was loaded directly onto a chromatography column
containing 150 mL of silica gel, slurry-packed in ether, and
the column was eluted with an ether f 10% methanol/ether
gradient. Fractions containing product were diluted with
toluene and the solvent was removed in vacuo (in order to
remove any water as an azeotrope) to afford 2.3 g (87%) of the
diastereomeric phosphonates 9 as a clear, viscous oil: 1H NMR
δ 1.95 (quintet, 2), 2.6 (m, 2), 2.8-3.3 (m, 4), 3.15 and 3.75
(2d, Σ ) 3, J ) 11.4 Hz), 4.2 (m, 2), 4.3 and 4.5 (2m, Σ ) 1),
4.7 and 5.5 (2d, Σ ) 1), 4.9-5.0 (m, 2), 5.05 and 5.2 (2ddd, Σ
) 1, J ) 4, 4, 8 Hz), 7.0-7.4 (m, 17), 8.5 (2d, Σ ) 2); 13C NMR
δ 28.69 and 31.08 (J ) 8.5 Hz), 35.37 and 35.83, 38.95 (J ) 7
Hz) and 39.07 (J ) 7 Hz), 48.40 (J ) 14 Hz) and 49.96 (J ) 14
Hz), 52.09 (J ) 7 Hz) and 53.22 (J ) 7 Hz), 64.37 and 64.85,
3060, 2980, 1720, 1680, 1500, 1510 (wk), 1420 (wk), 1230 cm-1
FAB MS 745 (MH+).
;
En zym e Assa ys. Porcine pepsin was obtained from Sigma
Chemical Co. (P-6887, lot 10H8075) prepared as 100 mM stock
solution, aliquots of which were diluted to 100 nM working
solutions. The activity of the working solutions was deter-
mined prior to each assay, and fresh solution was prepared
every 3-5 h. The substrate Lys-Pro-Ala-Glu-Phe-(4-NO2)Phe-
Arg-Leu35 was prepared by standard solid phase methods and
purified by HPLC (ꢀ310 ) 3770). Kinetic analysis60 of this
material gave values of Km ) 19 µM and kcat ) 66 s-1 (reported
values: 50 µM and 100 s-1, respectively35). Assays were
initiated by addition of ca. 10 µL of working enzyme solution
to a preequilibrated cuvette containing substrate (25-30 µM),
(60) Leatherbarrow, R. J . ENZFITTER; Elsevier Science Publishers
BV: Amsterdam, 1987.

3438 J . Org. Chem., Vol. 61, No. 10, 1996
Bartlett and Giangiordano
inhibitor, and buffer (0.1 M sodium acetate, pH 3.5); the final
enzyme concentration was 0.5-2.0 nM. The decrease in
absorbance at 310 nM (∆ꢀ ) -800 M-1 cm-1) was followed over
6-30 min, as appropriate.
Su p p or tin g In for m a tion Ava ila ble: Characterization of
other phosphonate analogs and kinetic plots for Ki determina-
tions (7 pages). This material is contained in libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
Ack n ow led gm en t. This work was supported by a
grant from the National Institutes of Health (GM-
46627).
J O952074C