J. Am. Chem. SOC. 1993,115, 12637-12638
`
`12637
`
`Immunosuppressive Boronic Acid Dipeptides:
`Correlation between Conformation and Activity
`
`Terence A. Kelly,'vt Julian Adams,? William W. Bachovchin,*
`Randall W. Barton,$ Scot J. Campbel1,ll Simon J. Coutts?
`Charles A. Kennedy,% and Roger J. Snowt
`Departments of Medicinal Chemistry, Pharmacology, and
`Analytical Sciences
`Boehringer Ingelheim Pharmaceuticals Inc.
`175 Briar Ridge Road, Ridgefield, Connecticut 06877
`Department of Biochemistry
`Tufts University School of Medicine
`Boston, Massachusetts 021 11
`Received September 27, 1993
`Dipeptidyl peptidase IV (DPP4; CD26) is a serine protease
`present on the surface of CD4+ cells.' Inhibition of this enzyme
`has been shown to suppress IL-2 production2 and antigen-induced
`T-cell proliferation3 in uitro. Recently, potentially therapeutic
`effects of inhibitors of DPP4 have been demonstrated using in
`uiuo models of immunos~ppression.~ Furthermore, CD26 has
`been identified as the adenosine deaminase binding protein,S a
`molecule involved in the severe combined immunodeficiency
`disease (SCID), and has also been proposed to be a coreceptor
`on the surface of CD4+ cells for the human immunodeficiency
`virus (HIV)? the causative agent of the acquired immune
`deficiency syndrome (AIDS).
`The DPP4 enzyme cleaves a two amino acid unit from the
`amino terminus of polypeptides which possess proline at the PI
`positiona7 Although its native substrate remains unknown,
`mutagenesis studies8 have demonstrated that the enzymatic
`activity of DPP4 is important in its role in the activation of T-cells.
`In order to explore more fully the role of this enzyme in
`immunosuppression we decided to examine the properties of the
`prolineboronic acid-containing dipeptides 1 ( Xu-boroPro) -9 These
`compounds have been shown by Bachovchin to be potent inhibitors
`of DPP4.10 The literature on boronate ester polypeptides indicates
`that they gain activity upon exposure to aqueous buffer,
`presumably through hydrolysis to the active boronic acid (e.g.,
`2)." In contrast, dipeptides 1 have been reported to lose activity
`in a time-dependent manner upon exposure to an aqueous
`medium,IO suggesting a more complicated scenario.
`t Department of Medicinal Chemistry, Boehringer Ingelheim.
`t Tufts University School of Medicine.
`I Department of Pharmacology, Boehringer Ingelheim.
`Department of Analytical Sciences, Boehringer Ingelheim.
`(1) (a) Dipeptidyl Peptidase IV-General and Applied Aspects; Barth,
`A,, Schowen, R. L., Eds.; Instituet fuer Pharmakologische Forschung: Berlin,
`1990; Vol. 38. (b) Lolda, Z . Histochemistry 1977, 54, 299. (c) Feller, A.
`C.; Heijnen, C. J.; Ballieux, R. E.; Parwaresch, M. R. Br. J. Haematol. 1982,
`51, 227. (d) Mentlein, R.; Heymann, E.; Scholz, W.; Feller, A. C.; Flad, H.
`D. Cell. Immunol. 1984, 89, 11.
`(2) Schoen, E.; Demuth, H.-U.; Eichmann, E.; Horst, H.-J.; Korner, LJ.;
`Kopp,J.;Mattern,T.;Neubert,F.;Noll,F.;Ulmer,A.
`J.;Barth,A.;Ansorge,
`S. Scand. J. Immunol. 1989, 29, 127.
`(3) Schoen, E.; Eichmann, E.; Grunow, R.; Jahn, S.; Kiessig, S. T.; Volk,
`H.-D.; Ansorge, S . Biomed. Biochim. Acta 1986, 11, 1523.
`(4) Kubota, T.; Flentke, G. R.; Bachovchin, W. W.; Stollar, B. D. Clin.
`Exp. Immunol. 1992, 89, 192.
`( 5 ) (a) Kameoka, J.;Tanaka,T.;Nojimi,Y.;Schlossman,S.
`F.; Morimoto,
`C.Science 1993,262,466. (b) Morrison, M. E.; Vijayasaradhi, S.;Engelstein,
`D.; Albino, A. P.; Houghton, A. N. J. Exp. Med. 1993, 177, 1135.
`(6) Hovancssian, A.; Callebaut, C.; Krust, B.; Jacotot, E. Manuscript
`submitted for publication (see: Chem. Eng. News, Nov 1, 1993, p 6).
`(7) Heins, J.; Welker, P.; Schoenlein, C.; Born, I.; Hartrodt, B.; Neubert,
`K.; Tsuru, D.; Barth, A. Biochim. Biophys. Acta 1988, 954, 161.
`(8) Tanaka, T.; Kameoka, J.; Yaron. A.; Schlossman, S. F.; Morimoto, C.
`Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 4586.
`(9) Prolineboronic acid is the analog of proline where the carboxylic acid
`is replaced by a boronic acid. It is abbreviated as boroPro.
`(10) Flentke, G. R.; Munoz, E.; Huber, B. T.; Plaut, A. G.; Kettner, C.
`A.; Bachovchin, W. W. Proc. NatI. Acad. Sci. U S A . 1991, 88, 1556.
`(11) (a) Kettner, C. A.; Shenvi, A. B. J. Biol. Chem. 1984,259, 15106.
`(b) Shenvi, A. B. Biochemistry 1986, 25, 1286.
`0002-7863/93/1515-12637$04.OO/0
`
`1
`
`2
`
`3
`
`To simplify the anlaysis of this apparent contradiction, we
`elected to synthesize the dipeptide in the unprotected boronic
`acid form (Scheme I). A single stereoisomer of (lS,2S,3R,SS)-
`pinanediol pyrrolidine-2(R)-boronate (4)Iz was coupled to Boc-
`valine to produce 5." Next, the pinanediol group was cleaved
`by treatment of the dipeptide with basic sodium ~eri0date.l~
`Finally, removal of the Boc group from 6 with HCl-Et20 afforded
`the HCl salt of compound 3 (val-boroPro).ls
`Val-boroPro (3) is a remarkably potent inhibitor of DPP4 (ICs0
`= 16 nM).16J7 However, the inhibitory activity of the material
`decreases rapidly upon standing in pH = 7.8 buffer, thus pointing
`to an intrinsic instability in the dipeptide motif.
`We hypothesized that the loss of inhibition was due to a
`cyclization taking place to generate a compound of structure 7
`(eq 1). This compound is a boron analog of a diketo piperazine,
`often a side product in peptide chemistry.
`
`3
`Active
`
`7
`Inactive
`
`Compound 7 could be isolated by treatment of 3 with aqueous
`base followed by purification via ion exchange chromatography.
`The cyclic structure of 7 was confirmed by "B-NMR. Compound
`3 shows a peak in the IlB spectrum at 6 28.0 ppm (relative to
`BFsmEtzO 6 = 0 ppm) while compound 7 shows a peak at 6 3.5,
`indicative of a tetracoordinated boron atom. This structure is
`supported further by X-ray data obtained for a protected analog.18
`A IH-NMR experiment was designed to follow the course of
`the rea~ti0n.l~ Compound 3 was dissolved in an aqueous buffer
`adjusted to pH = 7.8. Over time, increasing amounts of 7 were
`observed. The relative proportion of each compound was assessed
`at different time points by integrating the peaks at 6 3.09 and
`4.05 (3) and at 6 2.46 and 2.65 (7) (Figure 1, curve 1). Kinetic
`
`(12) Kelly, T. A,; Fuchs, V. U.; Perry, C. W.; Snow, R. J. Tetrahedron
`1993,49, 1009.
`(13) (a) Klausner,Y.S.;Bodansky,M.Synthesis 1972,453. (b) Bodansky,
`M.; Martinez, J. Synthesis 1981, 333.
`(14) Coutts, S.; Adams, J. A.; Snow, R. J.; Krolikowski, D. Manuscript
`in preparation.
`(15) All compounds were characterized by IH-, IlB-, and IT-NMR, CI-
`MS, and melting point. Compound 3 was further characterized by mi-
`croanalysis for C, H, N, and B.
`(16) The IClonumbcrrefers totheamountofcompoundrquiredtoinhibit
`50% of enzymatic activity for the duration of the assay.'' This number greatly
`underestimates the true activity of these compounds due to the cyclization
`process that is occurring over the 1-h course of the assay. Ki values for this
`series of compound have been estimated at <lo0 pM.22
`(17) DPP4 enzyme assay: This assay (based on the following: Smith, R.
`E.; Van Frank, R. M. Frontiers of Biology Volume 43: Lysosomes in Biology
`and Pathology; Neuberger, A., Tatum, E. L., Eds.; Amsterdam: North
`Holland, 1975; pp 193-249) relies on the ability of the enzyme to cleave the
`substrate ~-alanyl-~-prolinyl-2-(4-methoxy)naphthylamide
`( Ala-Pro-MNA).
`(18) Snow,R. J.;Coutts,S. J.;Kelly,T. A.;Krolikowski,D. A.;Campbcll,
`S.; Leonard, S.; Barton, R. W.; Kennedy, C. A.; Adams, J. A. Manuscript
`in preparation.
`(19) The 'H-NMR experiment was performed on a Bruker Instruments
`AF-270 spectrometer. A homonuclear presaturation routine using a spatially
`selective composite observe pulse (based on the following: Bax, A. J. Magn.
`Reson. 1985,65,142) was employed for solvent suppression. The sample was
`nonspinning. The compound was dissolved in a 0.5 M phosphate buffer (9:l
`H20-D20) adjusted to pH = 7.83.
`0 1993 American Chemical Society
`
`MYLAN - EXHIBIT 1072 Mylan et al. v. AstraZeneca IPR2015-01340
`
`

`
`12638 J. Am. Chem. SOC., Vol. 115, No. 26, 1993
`
`Communications to the Editor
`
`Scheme 1
`
`BocHN
`
`4
`6
`5
`a 2 equiv of Boc-valine, DCC, THF, 2 h, then filter and add to 4 and
`N-methylmorpholine in CHC13 (96%).* NaI04, acetone, H20 (61%).c
`HCI, EtOAc (90%).
`t l
`
`100
`
`75
`
`C .- "
`$? 50
`C -
`3
`
`25
`
`0
`
`
`
`T T T T
`
`$ $
`
`
`I I
`T
`I
`I
`
`7
`
`
`
`Y Y
`
`
`I I
`T
`T
`n I
`I
`
`[71
`
`% Compound 7
`% of original IC50
`
`o
`
`0
`
`0 c)
`
`0
`m
`
`0 2
`
`0 W
`time (min)
`Figure 1. Curves showing that the cyclization of 3 to 7 (curve 1, squares,
`determined by 'H-NMR) corresponds to the loss of ability to inhibit the
`DPP4-catalyzed hydrolysis of substrate (curve 2, diamonds). Both
`experiments were performed at pH = 7.8. Kinetic data for curve 2 are
`fitted up to f = 75 min since at high [7] the cumulative inhibition
`demonstrated in Figure 2 becomes significant over the time required to
`run the assay (ref 16).
`analysis of the data gave an observed first-order rate constant20
`of 3.9 X 10-4 f 0.7 X 10-4 s-l for the cyclization of 3 to 7. A
`corresponding biochemical experiment was run measuring the
`time-dependent ability of the compound to inhibit DPP4 after
`also standing in an aqueous buffer at pH = 7.8. As expected,
`the compound became less active over time and the inactivation
`proceeded with a similar observed rate constant20 (6.6 X 10-4 f
`0.6 X 10-4 s-l, Figure 1, curve 2). The correlation demonstrated
`between the structural and enzymatic experiments establishes
`that the cyclization is indeed responsible for the loss in activity.
`The cyclization is also reversible due to the dative nature of
`the B-N bond in 7. We have observed that although activity
`against the enzyme diminishes by 3 orders of magnitude, the
`final ICs0 is still submicromolar, suggesting that at pH = 7.8
`(20) The rate of the cyclization is dependent on the position of the
`equilibrium between protonated and unprotonated 3 (see eq l), so for the
`Since
`forward (i-e., cyclization) reaction d[7]/dt is equal to k&[3]/[H+].
`the solution is buffered, [H+] is constant and the rate equation reduces to
`d[7]/dt = k,b[3] where kat. = k&/[H+].
`
`Figure 2. Recovery of inhibitory activity of 7 (via 3) with time.
`compound 7 is in equilibrium with a small amount of the open
`from 3. We sought to exploit this process to see if the activity
`of compound 7 could be restored by driving the equilibrium back
`toward 3 using an acidic buffer. This proved possible. At pH
`= 3.0 the material reverted to its open form and completely
`regained its activity against the enzyme.2'
`Because boronic acid peptides are known to display tight-
`binding inhibition with a very slow off-rate,22 the equilibrium has
`profound consequences for this series of immunosuppressants.
`When the enzyme is incubated in the presence of cyclized (i.e.,
`inactive) inhibitor under standard (pH = 7.8 buffer) conditions
`and then is assayed for the ability to hydrolyze substrate, a
`cumulative, time-dependent inhibition of the enzyme is observed
`(Figure 2). These data are consistent with the formation of the
`equilibrium between 3 and 7 being reestablished after the active
`open form 3 binds (and remains bound)22 to the enzyme (eq 2).
`Thus it is the presence of the enzyme itself that drives the
`equilibrium toward the active form.
`
`3 + Enz - EnZ-3
`
`7 --
`
`(2)
`The results of these experiments show a system in which the
`ability to inhibit a serine protease relies on the position of an
`equilibrium between two isolable and interchangeable confor-
`mations. They also show that specific binding to an enzyme can
`drive an equilibrium if the off-rate is such that the enzyme can
`entrain one of the components from the system. The in vivo
`consequences of this phenomenon are currently under study.
`Acknowledgment. We are grateful to the Analytical Science
`Department at Boehringer Ingelheim for their fine efforts in
`support of this work. In particular we thank Dr. Phil Pitner for
`many helpful discussions.
`(21) The kinetic rationale for this phenomenon can be derived from the
`rate equation for the ring opening: d[3]/dt = k-&[7][H+]. As [H+] is
`increased, theobservedrateconstant oftheopening increases while theobserved
`rate constant of the closing (which is inversely related to [H+])M decreases.
`(22) Gutheil, W. G.; Bachovchin, W. W. Biochemistry 1993, 32, 8723.
`This work reports km, k.~, and Ki for Pro-boroPro as 5.02 X 106 M-1 s-1.77.8
`X 1V s-l, and 15.5 pM, respectively.