J. Enzyme Inhibition, 1988, Vol. 2, pp. 129-142
`Reprints available directly from the publisher
`Photocopying permitted by license only
`
`0 1988 Harwood Academic Publishers GmbH
`Printed in Great Britain
`
`DIPEPTIDYLPEPTIDASE IV - INACTIVATION
`WITH N-PEPTIDYL-0-AROYL HYDROXYLAMINES
`
`H.U. DEMUTH2*, R. BAUMGRASS3, C . SCHAPERI, G. FISCHERI
`and A. BARTH’
`Departments of Biosciences’ and Biotechnology’, Martin Luther University, Halle,
`Academy of Sciences of the GDR, Central Institute of Nutrition,
`Potsdam-Rehbriicke3, German Democratic Republic
`
`(Received 10 August 1987)
`
`Eleven N-peptidyl-0-aroyl hydroxylamines have been synthesized and their hydrolytic stability, acidity and
`properties during reaction with dipeptidyl peptidase IV (E.C. 3.4.14.5) investigated. N-peptidyl-0-(4-
`nitrobenzoyl) hydroxylamines act as irreversible inhibitors of serine proteases’ . The serine enzyme, dipep-
`tidy1 peptidase IV (DP IV), is inactivated by substrate analog derivatives of this class by a suicide
`inactivation mechanism. During the enzymic reaction of DP IV with the suicide substrates most molecules
`are hydrolyzed but some irreversibly inactivate the target enzyme. In contrast to porcine pancreatic elastase
`and thermitase, DP IV exhibits a high ratio for hydrolysis of the compounds versus inhibition during their
`interaction with the enzyme. Variation of the leaving aroyl residue lowers this ratio. Variation of the
`substrate analog peptide moieties of the DP IV-inhibitors increases their ability to inhibit the enzyme to
`a remarkable extent. Possible reaction pathways are discussed.
`
`KEY WORDS: Dipeptidyl peptidase IV, diacyl hydroxylamines, nitrenes, mechanism-based inactiva-
`tion, Lossen reaction.
`
`INTRODUCTION
`
`Dipeptidyl peptidase IV, a serine peptidase with pronounced specificity for proline
`residues, is believed to be involved in modulation of proline containing peptide
`hormones and activation of proteins by limited proteolysis24.
`The enzyme is a membrane-bound aminopeptidase anchored by a hydrophobic
`peptide chain into the bilayer and is widely distributed in organs of mammals.
`Dipeptidyl peptidase IV has two identical but catalytically independent subunits and
`removes dipeptide units of the structure Xaa-Pro from the N-terminal end of polypep-
`tides and proteins (Xaa-Pro-Yaa). Substrates containing amino-acids other than
`proline in the PI-position are also accepted by the enzyme but the turnover rate is
`reduced. The unprotected and protonated aminofunction of the P,-amino acid is
`essential for catalysis (for a review see Walter’).
`The enzyme activity in plasma is useful for diagnostic purposes since it deviates
`significantly from normal values in several pathological circumstances, e.g. blood
`diseases and different cancersG8. However, the enzyme’s physiological role and the
`cause of the deviation of DP IV-activity in pathological processes are not yet fully
`understood.
`Enzyme inhibitors can be useful tools to help clarify the biological functions of
`
`*Correspondence to: Dr. Hans-Ulrich Demuth, Department of Biotechnology, Martin-Luther Univer-
`sity Halle-Wittenberg, Domplatz 1, Halle (Saale), DDR-4020, German Democratic Republic.
`
`129
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`AstraZeneca Exhibit 2075
`Mylan v. AstraZeneca
`IPR2015-01340
`
`Page 1 of 14
`
`

`

`H.U. DEMUTH ct al.
`130
`enzymes and accordingly we have studied N-peptidyl-0-aroyl hydroxylamines as
`mechanism-based inactivators of serine proteases'. N-alanyl-prolyl-O-(4-nitroben-
`zoyl) hydroxylamine has been used successfully as a specific inhibitor in the investiga-
`tion of DP IV in human lymphocytes9.
`Two reaction pathways between serine peptidases and the substrate analog diacyl
`hydroxylamines were proposed (see Scheme 1).
`
`L F . + p e p t x d e - G O - N
`+ p e p t x d e - N = C = O 3
`
`8 - 1 1 -
`
`E L - C
`'< J - - p e p t x d e
`+ H U O C - R
`
`\
`
`p a t h 1
`
`t : + 1
`
`LE, p e p t ~ d e - C 0 - - N H - - O - C O - R 1
`
`p a t h 2
`
`L !.-.:t~e:,<-.r~t..;
`
`t h e f r e e e l i z y m r , H may h e a l k y l 01
`
`p h e n y l . E , - L ' : I
`
`I 1 s t .
`
`: L . - , L u i - ~ (ti L h e i n a c t i v e cornpie..: 1 s unkriuwri
`#,,, , : , c
`SCHEME 1 Pathways of reaction of peptidases with N-peptidyl-0-aroyl hydroxylamines.
`
`Pathway 1: During the formation of noncovalent or covalent complexes between
`suicide-substrate and target enzyme, N-O bond fission occurs leading to reactive
`intermediates (carbonyl nitrenes or isocyanates) which irreversibly modify the pro-
`tein. Final products are the inactivated enzyme and the 0-acyl residue.
`Pathway 2: The target enzyme hydrolyzes the compound as a substrate. The release
`of 0-acyl hydroxylamine during the catalytic process prevents N-0 bond fission
`leading to a simple acyl enzyme. Final products are active enzyme, peptide and the
`0-acyl hydroxylamine. The partition ratio (r) between both processes (substrate
`hydro1ysis:enzyme inactivation) is a characteristic measure of mechanism-based inac-
`tivation".
`In this paper work is reported on the mechanism of reaction of N-peptidyl-0-aroyl
`hydroxylamines with porcine kidney DP IV.
`
`MATERIALS AND METHODS
`S,~xfhesis sf' Diacyl Hydroxylarnines
`All N-peptidyl-0-benzoyl hydroxylamines used in this work were synthesized as
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 2 of 14
`
`

`

`INACTIVATION OF DIPEPTIDYLPEPTIDASE IV
`131
`described previously' by hydroxylaminolysis of the corresponding Boc-dipeptidyl-
`methylesters" . The resulting peptide hydroxamic acids (Boc-dipeptide-NHOH) were
`acylated to give diacyl hydroxylamines by treatment with the appropriate benzoyl
`chlorides in Schotten-Baumann-reactions.
`The Boc-dipeptidyl-methyl esters were synthesized according to standard meth-
`odsi2 starting with C-terminal amino acid methyl esters.
`Amino acids were purchased from Reanal, Budapest. Di-tert. butylcarbonate was
`from Serva, Heidelberg and tert. butylchloroformate, 3-chlorobenzoyl chloride, 4-nit-
`robenzoyl chloride and 3,5 dinitrobenzoyl chloride were obtained from Merck, Darm-
`stadt. All other chemicals were research grade from Laborchemie, Apolda. Organic
`solvents were dried before use by standard procedures.
`Melting points are uncorrected. TLC in product anlysis was performed on silicagel
`plates (Silufol, Kavalier, Czechoslovakia). Intermediate and final products were
`further characterized by 'H-NMR, elemental analysis and uv-spectrometry.
`Peptide hydroxamic acids were obtained amorphous in all cases with yields between
`90 and 95 YO. Their acylation to N-Boc-dipeptidyl-0-benzoyl hydroxylamines using
`substituted benzoyl chlorides gave yields of 6@80Y0 after crystallization from ethyl
`acetate/petrol ether. Finally, the Boc-groups were removed by HCl/CH, COOH to
`give the hydrochlorides of N-dipeptidyl-0-benzoyl hydroxylamines listed in Table I.
`All N-terminal deprotected compounds were highly hygroscopic, thus results of
`elemental analysis were influenced by moisture and these are not shown for com-
`pounds 1-5. Their purity was checked by TLC, by comparison of the free dipeptides
`(Xaa-Pro) with the products of DP IV-catalyzed total hydrolysis of compounds 1-1 1
`to the appropriate dipeptides and 0-benzoyl hydroxylamine~'~.
`Comparison of the absorption spectra of compounds 1-1 1 after complete degrada-
`tion to dipeptidyl hydroxamic acids and substituted benzoic acids with the spectra of
`solutions of the corresponding recrystallized commercial benzoic acids gave purity
`higher than 95.2% in all cases.
`
`TABLE I
`Hydrochlorides of N-peptidyl-0-benzoyl hydroxylamines; analytical parameters
`Compound*
`M,
`Formula
`Mp. ["C]
`C% H% N%
`141-142
`1 Gly-Pro-NHO-Bz(4-OCH,)
`357.81 C,,HmN30,CI
`133-134
`34 1.8 1 C,, H, N, 0,Cl
`2 Gly-Pro-NHO-Bz(4-CH3)
`157-1 58
`3 Gly-Pro-NHO-Bz
`327.78 C 14 H I8 N, 0, C1
`146148
`4 Gly-Pro-NHO-Bz(3-C1)
`362.23 C,,H,,N,04CI,
`see text
`5 Gly-Pro-NHO-Bz(3,5-(NO,),
`41 7.79 C,,H,,N5O,C1
`6 Gly-Pro-NHO-Bz(4-NO2)
`148-150
`372.80 C,4H,7N404C1
`
`found: 44.20 4.58 14.22
`requ.: 45.10 4.60 15.01
`found: 46.70 4.73 14.49
`requ.: 46.53 4.91 14.48
`114117 found: 48.52 6.21 11.72
`requ.: 50.41 5.88 13.07
`100-105 found: 52.81 5.07 11.90
`requ.: 54.48 5.01 12.10
`93- 95 found: 51.55 5.59 10.68
`requ.: 54.02 5.58 12.12
`found: 49.20 4.97 13.70
`requ.: 50.12 5.02 15.74
`*Nomenclature of peptide residues according to Schechter and Berger", -NHO- is the hydroxylamine
`function, N-acylated by petidyl residues, 0-acylated by various benzoic acids (Bz), substituent in brackets,
`Bz(X).
`
`578.01 C2,H3,N,O8 C1
`10 Lys(Z)-Pro-NHO-Bz(4-NO2)
`11 Lys(Z-4-N0,)-Pro-NHO-Bz(4-N02)
`623.01 C,,H,,N,O,,C1
`
`7 Ala-Pro-NHO-Bz(4-NO2)
`
`386.77 Cl5HIgN,O6C1
`
`148-149
`
`8 Leu-Pro-NHO-Bz(4-N02)
`
`428.87 C,,H,,N406 C1
`
`9 Phe-Pro-NHO-Bz(4-NO2)
`
`462.92 C2,H,,N40,C1
`
`101-103
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 3 of 14
`
`

`

`H.U. DEMUTH ef al
`131
`No melting point is reported for 5 but the analytical data of the corresponding
`Boc-protected compound is as follows:
`N-BOC-G~~-P~~-NHO-B~(~,~-(NO~)~).
`M.p.: 131-1 3 2 T , (Found: C, 47.69; H,
`5.06; N, 13.99. C,,H,3N,0,0 requires C, 47.40; H, 4.82; N, 14.54%).
`
`Kinetic Methods and Inactivation Experiments
`All enzyme activity assays and inactivation experiments were performed and all
`spectra were recorded using a Unicam SP 800 and the microprocessor controlled
`Specord M 40.
`Estimation of pK-values was performed by measurement of spectra between 250
`and 450 nm for 0.1 mM solutions of diacyl hydroxylamines in silica cells containing
`2.0 ml McIlvain-buffer (pH 2.2-pH 8.0) and 0.1 ml acetonitrile or dimethyl forma-
`mide. Graphic analysis of differences in absorption at 300nm was used for pK-
`calculation.
`Spontaneous degradation of diacyl hydroxylamines was followed spectrometrically
`in the range of 225 to 400nm at 3 0 T .
`Solutions contained 40 mM sodium phosphate buffer, ionic strength 0.125, pH 7.6
`and 0.1 mM concentrations of diacyl hydroxylamines.
`Data at several wavelengths were collected as functions of time and the pseudo-
`first-order rate constants calculated using nonlinear regression programs provided
`with the Specord M 40 by Carl-Zeiss-Jena and using a Hewlett-Packard desktop
`computer HP 9825 A.
`The activity of DP IV was determined with Gly-Pro-4-nitroanilide and Ala-Pro
`4-nitroanilide in 40mM sodium phosphate buffer with an ionic strength of 0.125,
`maintained by potassium chloride as described previously‘.
`Specific activity of DP IV was in the range 3 5 4 5 U/mg. The kc,,-values given in
`Table 3 have been standardized and calculated assuming a maximal activity of
`55 Uimg and a molecular weight of 115 000 per subunit of the enzyme.
`DP IV-catalyzed hydrolysis of substrate analog diacyl hydroxylamines has been
`analyzed following the absorption change due to the release of 0-benzoyl hydroxyla-
`mines between 260 and 360 nm (for wavelengths and absorption coefficients see Table
`3).
`Activity was estimated in lOmm silica cells containing 2.5ml of 40mM sodium
`phosphate buffer, pH 7.6, ionic strength 0.125 at 30°C. The pseudosubstrate con-
`centration was varied between 1 .O pM and 0.1 mM. Final DP IV concentration was
`in all cases 50nM after addition of 50pl aliquots to initiate reaction.
`Initial rates were analyzed using the software cassette “reaction kinetics” built into
`the spectrophotometer M 40. The parameters k,,, and K, were calculated from the
`initial rates using nonlinear regression programs to fit a hyperbola using a Sinclair
`Spectrum Plus computer.
`Residual activity of DP IV after preincubation with several substrate analog diacyl
`hydroxylamines was estimated as follows: DP IV was incubated with suicide sub-
`strates in concentrations of 20 pM to 1.04 mM in 2.5 ml (40 mM) sodium phosphate
`buffer, pH 7.6, ionic strength 0.125 at 30°C. The reaction was initiated by adding
`enzyme to give 0.15 nM DP IV in the mixture. Decrease of activity was followed by
`fiithdrawing 0.1 ml aliquots of the incubation mixture and estimating its residual DP
`IV activity against 1.23 mM alanyl-prolyl-4-nitroanilide under the same conditions
`described above. The inactivation reaction was monitored until a completion but
`usually not longer than two hours.
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 4 of 14
`
`

`

`INACTIVATION OF DIPEPTIDYLPEPTIDASE IV
`133
`The effectiveness” of suicide inactivation is characterized by the partition ratio
`r = k,,, /k,,,,, . Deviations from linearity in semilogarithmic plots (log residual activity
`versus time) occur caused by enzyme catalyzed hydrolysis ((Figure 3A) and an exact
`estimation of k,,,,, is not possible. Using the following approximation here a value
`representing “r” could be obtained:
`Since the partition ratio, r, is equal to the number of molecules processed as
`“substrate” divided by the number of molecules processed as “inhibitor” (giving
`inactivated enzyme), r can be set equal to [S]/[E,] in our case. The final concentration
`of processed substratemolecules is approximately equal to its initial concentration, if
`after the reaction has finished all suicide substrate was consumed and only catalytic
`quantities of enzyme were used.
`If in the experiment an enzyme concentration was chosen that after the reaction is
`complete still gives an active enzyme then the molar concentration of inactivated
`enzyme and therefore the concentration of molecules leading to inactivation can be
`calculated.
`Dipeptidyl peptidase IV was purified according to reference 24 using a slightly
`modified procedure.
`
`RESULTS
`
`The structure of diacyl hydroxylamines (R, -CO-NH-0-CO-R,) ideally permits varia-
`tion of affinity and reactivity of the molecule towards a target enzyme simply by
`selection of appropriate N-acyl- and 0-acyl residues.
`Bearing in mind the substrate specificity of DP IV two sets of modified diacyl
`hydroxylamines were synthesized and used in this study: (1) Compounds with the
`same substrate analog peptide moiety but different substituted benzoyl residues as
`leaving groups (R,-CO- = Gly-Pro, R,-CO- = substituted benzoic acid) and, (2)
`compounds with same leaving groups but different amino acids in the P,-position of
`the N-peptidyl residue of the molecule (R,-CO- = Xaa-Pro, R,-CO- = 4-nitroben-
`zoyl) (see Table 11).
`
`Stability and Ionization of Diacyl Hydroxylamines in Aqueous Solution
`Degradation studies with diacyl hydroxylamines in buffer solutions using TLC in
`silica gel plates showed peptide hydroxamic acids and benzoic acids as products of the
`spontaneous degradation of N-peptidyl-0-benzoyl hydr~xylamines’~.
`Taking advantage of the difference in absorption between solutions of free sub-
`stituted benzoic acids and solutions of N-peptidyl-0-benzoyl hydroxylamines the
`pseudo-first-order rate constants of this degradation have been estimated spectrome-
`trically (Table 11).
`The uv-spectra of 0.1 mM buffered solutions (pH 2.2-8.0) of N-peptidyl-0-(4-
`nitrobenzoyl) hydroxylamines exhibited a bathochromic shift of absorption maxima
`(263nm to 268nm) due to the known acidity of diacyl hydr~xylamines’~. The pH-
`dependent absorption change at 300 nm was used for estimation of pK-values of some
`N-peptidyl-0-(4-nitrobenzoyl) hydroxylamines (Figure 1, Table 11).
`With typical pK-values of 4.8 the compounds exist at neutral pH as anions as a
`result of the acidity of the -CO-NH-0- linkage. Investigation of the stability of
`Xaa-Pro-NHO-Bz(4-N02) derivatives (Xaa = Gly, Ala, Leu, Phe, Lys(Z)) in the
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 5 of 14
`
`

`

`I34
`
`H.U. DEMUTH er al.
`
`TABLE I1
`Rate of spontaneous degradation and pK value for N-peptidyl-0-benzoyl hydroxylamines in buffer
`solutions
`
`PK
`Pseudo-first order
`Suicide substrate
`rate constant+, kobs [min-'1
`value#
`-
`n.e.
`0.82. lo-'
`GIy-Pro-NHO-Bz(4-CH3 )
`me.
`Gly-Pro-NHO-Bz(3-CI)
`1.13. lo-'
`4.77
`2.20- 10-3
`Gly-Pro-NHO-Bz(CN0, )
`n.e.
`6.30. lo-'
`Gly-Pro-NHO-Bz(3.5-(NO, )2 )
`4.9
`Ala-Pro-Bz(4-N02 )
`1.30- 10-j
`4.8
`2.20.10-3
`Leu-Pro-NHO-Bz(4-N02 )
`4.5
`2.00- 10-
`Phe-Pro-NHO-Bz(4-N02 )
`2.80- lo-'
`Lys(Z)-Pro-NHO-Bz(4-NOZ)
`4.85
`2.30. lo-'
`4.80
`Lys(Z-4-NO, )-Pro-NHO-Bz(4-NOz )
`Release of benzoic acid followed uv-spectrometrically (see MATERIAL AND METHODS) in 0.04 M
`sodium phosphate buffer. pH 7.6. ionic strength was adjusted with potassium chloride to 0.125, tem-
`perature 30 C.
`# pK-values of ionisation of the -CO-NH-0- linkage were estimated in McIlvaine-buffer (pH 2.2-8.0) at
`25 c
`me. = not estimated
`
`I
`
`I
`
`I
`
`L.-.!-.i. I __-
`6.
`4.
`
`PH
`FIGURE I
`pHdependence of absorption of Lys(Z)-Pro-NHO-Bz(4-N02) 0.1 mM in McIlvaine buffer,
`pH 2 2~ 8.0. at 25.C.
`
`I
`
`0.
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 6 of 14
`
`

`

`INACTIVATION OF DIPEF'TIDYLPEPTIDASE IV
`I35
`range pH 5-pH 8 showed that they did not exhibit significant pH-dependence of the
`degradation reaction. This result does not fit a model involving hydroxyl ion catalyzed
`hydrolysis of either amide or ester bonds of the compounds (pathway la, ib in
`Scheme 2).
`
`H e 0
`
`R:* .c 9 * * N @ - O - C O - R n
`
`J
`1
`
`R I - C O - N H - O - C O - - K e
`
`\
`
`H I
`- C O - Q
`
`+ 0 0 - C O - R e
`
`2a
`R , - N = C = O + O O - C O - R n
`
`+ H e 0 1
`
`2 b
`
`R I - C O N H O H + OU-CCJ-R,
`
`SCHEME 2 Degradation pathways for diacyl hydroxylamines.
`
`Due to their relative stability i.e. half lives of 2-5 h under conditions used, investiga-
`tion of their reactions with DP IV were possible.
`
`DP IV-catalyzed Hydrolysis of Diacyl Hydroxylamines
`By definition, mechanism-based inactivation of enzymes includes inactivation of the
`enzyme by the suicide substrate as well as the enzyme-catalyzed turnover of the suicide
`substrate (Scheme 1).
`Indeed, incubation of N-Xaa-Pro-0-benzoyl hydroxylamines in buffered solutions
`with substantial amounts of DP IV results, in contrast to the nonenzymic degrada-
`tion, in complete hydrolysis to 0-benzoyl hydroxylamines and appropriate dipeptides
`as proved by TLCi4 (see pathway 2 in Scheme 1). The reaction leads to an absorption
`decrease in the uv-spectra between 250-400nm due to the release of 0-benzoyl
`hydroxylamines which permits estimation of the catalytic constants K, and k,,, (Table
`111). Figure 2 illustrates the dependence of the rate of DP IV-catalyzed hydrolysis with
`different concentrations of Leu-Pro-NHO-Bz(4-NOJ.
`
`Inactivation of DP ZV by Substrate Analog Diacyl Hydroxylamines
`Incubation of DP IV with suicide substrates (shown in Table IV) and measurement
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 7 of 14
`
`

`

`I36
`
`H.U. DEMUTH e t a /
`
`TABLE I11
`Kinetic parameters for dipeptidyl peptidase IV-catalyzed hydrolysis of N-peptidyl-U-benzoyl hydroxyl-
`amines
`
`Suicide substrate
`
`k,, *
`Km x lo5)
`k,,,/K,
`Wavelength+/
`w1
`absorb. coeff.
`( x 10')
`[see-']
`286 :3600
`117
`3.90
`Giy-Pro-NHO-Bz(4-OCH,)
`3.0
`263.5 : 4800
`2.77
`108
`3.9
`GI~-Pro-NHO-Bz(4-CH,)
`263.513372
`2.73
`3.7
`101
`Giy-Pro-NHO-Bz
`263.5:3285
`2.73
`101
`3.7
`Gly-Pro-NHO-Bz( 3-CI)
`270
`:1400
`108
`4.5
`2.4
`GIy-Pro-NHO-Bz(3.5-(NO2), )
`303.5: 2000
`2.9
`116
`4.0
`Giy-Pro-NHO-Bz(4-N02 )
`4.1
`325 :I230
`Ala-Pro-NHO-Bz(4-NO, )
`1.8
`74.5
`I .38
`55
`Leu-Pro-NHO-Bz(4-NO,)
`4.0
`300 :1847
`0.63
`43
`6.8
`300 :I864
`Phe-Pro-NHO-Bz(4-NO, )
`0.74
`28
`300 :I540
`Lys(Z)-Pro-NHO-Bz(4-NO2 )
`3.8
`1.05
`Lys(Z-4-N02 )-Pro-NHO-Bz(4-N02)
`2.1
`22
`300 :I513
`* DP IV-catalyzed release of U-(4-nitrobenzoyl) hydroxylamine followed spectrometrically in 0.04 M
`sodium phosphate buffer. pH 7.6. at 30'C with ionic strength adjusted with potassium chloride to 0.125.
`Wavelength in nm.
`*The k,,, -values have been calculated here assuming a molecular weight of I 15000 per subunit of DP IV
`and a specific activity of 55U;mg for the enzyme used.
`
`1/S * 10-5, M-1
`
`FIGURE Z Lineweaver-Burk plot of kinetic results in the DP IV-catalyzed hydrolysis of N-Leu-Pro-0-
`(4-nitrobenzoyl) hydroxylamine at pH = 7.6 in 0.04M sodium phosphate buffer at 30°C with ionic
`strtngth of 0.125 maintained with potassium chloride.
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 8 of 14
`
`

`

`INACTIVATION OF DIPEF’TIDYLPEPTIDASE IV
`
`137
`
`A
`
`e houri prdnorrbatlon
`
`0
`
`.e
`
`.4
`
`lo mM
`
`.6
`
`0
`
`20.
`
`40.
`time (min)
`(A) - Inactivation of Dipeptidyl peptidase IV with N-Leu-Pro-O-(4-nitrobenzoyl) hyd-
`FIGURE 3
`roxylamine. Measurements carried out at 30°C, pH 7.6, in 0.04M sodium phosphate buffer with ionic
`strength 0.125, maintained with potassium chloride. DP IV = 0.165 nM, I, = 0.26mM. (B) Insert shows
`dependence of inactivation of DP IV with N-Leu-Pro-O-(4-nitrobenzoyl) hydroxylamine as function of
`inhibitor concentration at 2 hours incubation time. DP IV = 0.146nM.
`
`60.
`
`of the residual activity at different time intervals shows a time-dependent loss of
`enzyme activity (Figure 3). The irreversibility of the inactivation was proved by
`dilution experiments and by gel chromatographic separation of the inactivated pro-
`tein from excess of inhibitor. Incubation of the enzyme with diacyl hydroxylamines
`having other N-acyl-residues as Xaa-Pro- does not affect DP IV activity’.
`Diminution or suppression of the inactivation process by adding “normal” sub-
`strate or competitive inhibitors to the incubation mixture may indicate that a specific
`reaction within the active site of a target enzyme is occurring. Thus, a change of the
`half life for inactivation in the presence of Lys(Z-4-N02)-Pro, a competitive inhibitor
`of DP IVI4, was expected:
`To a preincubation mixture containing 2.8 mM N-Ala-Pro-NHO-Bz(4-N02) and
`57 pM Lys(Z4-N02)-Pro, DP IV was added to give a final concentration of 1.1 nM.
`The resulting inactivation rate constant (residual activity estimated as described in
`Material and Methods), kobs = 7.7 x 10-3min-’, is about one order of magnitude
`lower than the rate obtained in a control experiment in the absence of Lys(Z4-NO2)-
`Pro.
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 9 of 14
`
`

`

`138
`
`H.U. DEMUTH er (11.
`
`TABLE IV
`Inactivation of dipeptidyl peptidase IV by N-peptidyl-0-benzoyl hydroxylamines
`
`Suicide substrate
`
`Turnover ratio
`“r”
`2.03.106
`6.97.105
`6.67- 105
`3.47. los
`2.13 * lo5
`3.01 * 10’
`6.73.104
`1.50. 104
`1.40-104
`8.00*10’
`1.20. 104
`
`0.58
`1.54
`1.52
`2.92
`5.08
`3.83
`11.07
`36.69
`30.07
`35.10
`18.31
`
`5.0
`5.0
`5.0
`5.0
`5.0
`5.0
`4.4
`2.6
`2.6
`2.6
`2.6
`
`~
`
`~~~~
`
`~~
`
`~
`
`GIy-Pro-NHO-Bz(4-OCH, )
`Gly-Pro-NHO-Bz(4-CHl
`Gly-Pro-NHO-Bz
`Gly-Pro-NHO-Bz( 343)
`Gly-Pro-NHO-Bz(3.5-(N02 ): )
`Gly-Pro-NHO-(4-N02)
`Ala-Pro-NHO-Bz(4-NO,)
`Leu-Pro-NHO-Bz(4-N02 )
`Phe-Pro-NHO-Bz(4-NO2 )
`Lys(Z)-Pro-NHO-Bz(4-N02)
`Lys(Z-rl-NO, )-Pro-NHO-Bz(4-N02 )
`Initial inhibitor concentration.
`*DP IV incubation with inhibitors in 0.04M sodium phosphate buffer, pH 7.6, at 30°C with ionic
`strength adjusted with potassium chloride to 0.125. Preincubation time: all Gly-Pro-derivatives one hour,
`all others two hours. Estimation of “r”-values see “Material and Methods”.
`‘The k,,,,,-values in this paper have been calculated assuming r = kcat/k,,,,l and r = [hydrolyzed
`inhibitor].’[inactivated enzyme] using kc,,-values in Table 2.
`
`The results described meet the typical criteria established for suicide inactivators’6
`which also include enzyme-catalyzed turnover of inhibitor during the incubation
`period. In Figure 3 the deviation of the estimated residual activities from a straight
`line is due to this turnover. Although some suicide substrate molecules inhibit the
`target enzyme others are hydrolyzed by it simultaneously (compare Scheme 1).
`The higher the enzyme concentration or the less effective the inactivation process
`a more curved dependence of residual activity from incubation time in a semilogarith-
`mic presentation may be observed. Alternatively the lower the enzyme concentration
`or the more effective the inactivation compared to catalysis is, a closer approach to
`a semilogarithmic presentation of residual activity versus time will be obtained.
`Ideally, if no inhibitor molecule is destroyed by the enzyme or the inhibitor
`concentration does not change significantly during incubation, kinetics follow
`pseudo-first order reactions. Accordingly suicide inactivation is best characterized by
`i.e. values of “r” (ratio of kCat/kinact).
`ratios of catalysis versus ina~tivation’~~’’.’~
`Values of “r” for the reaction of DP IV with the substrate analog diacyl hydroxyl-
`amines estimated as described previously”, are summarized in Table IV.
`
`DISCUSSION
`
`The concept of introducing substrate analog diacyl hydroxylamines as specific in-
`hibitors of DP IV was aimed at overcoming the undesirable features of N-terminal
`unprotected proline containing dipeptide derivatives which as amides or esters tend
`to form diketopiperazines during intramolecular reactions. Earlier attempts to use
`Xaa-Pro-chloromethylketones gave compounds with half lives less than ten minutes
`and complete inactivation of the target enzyme was achieved only with substantial
`amounts of inhibitor. Compounds were sought where the chemical reactivity is
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 10 of 14
`
`

`

`INACTIVATION OF DIPEPTIDYLPEPTIDASE IV
`139
`masked and are only activated by the catalytic attack of the target enzyme which so
`generates its own inactivation.
`Falling in this category of masked compounds that deserves special mention, are
`N-halogenamides, carbonylazides and diacyl hydroxylamines i.e. compounds prone
`to rearrangement (Hofmann-, Curtius- and Lossen reactions). It was considered that
`should the compounds generate reactive intermediates during degradation, the mech-
`anism could be used to inactivate proteolytic enzymes.
`Azides and diacyl hydroxylamines may share this feature since possible inter-
`mediates in the Curtius and Lossen reactions are carbonyl nitrenes and/or iso-
`cyanate~'~. As expected the peptide azides, Boc-Ala-Ala-N, and Ala-Ala-N,, were
`found to be inhibitors of chymotrypsin and DP IV, respectively but there was still
`substantial hydrolytic degradation of these compounds in buffer solution due to the
`good leaving azide ionz0.
`Since diacyl hydroxylamines exist as anions at neutral pH-valuesJ5 a higher stability
`to hydrolysis of the scissile carbonyl linkage in N-peptidyl-0-benzoyl hydroxylamines
`was expected. Incubation of chymotrypsin, elastase, thermitase and dipeptidyl pep-
`tidase IV with specific substrate analog N-peptidyl-0-benzoyl hydroxylamines resul-
`ted in irreversible enzyme inhibition'. In most cases enzyme-catalyzed hydrolysis of
`the substrates was negligible. Only during reaction between substrate analog diacyl
`hydroxylamines with DP IV have we found substantial enzyme-catalyzed hydrolysis
`of the suicide substrates.
`Therefore a systematic study of the factors influencing the effectiveness of inactiva-
`tion of DP IV was perforemd. The compounds used are stable and suitable in the
`enzymological experiments under the standard conditions used. Their degradation in
`solution may be caused by hydrolysis of the ester or the amide bonds of the hy-
`droxylamine function or by N-0 bond fission (see pathways la, lb and 2 in Scheme
`2).
`Variation of the peptidyl residue of the compounds does not influence the degrada-
`tion rate remarkably (Table 111). Obviously, the distance between the P,-amino acid
`and reactive center of the molecule excludes noticeable effects of side chain structure
`in the reaction. In contrast, rate constants increase proportionally to the electron
`withdrawing effects of substituents in the leaving benzoic acid in compounds 1-4
`(Table 11).
`These kinetic findings are in good agreement with studies on the degradation of
`diacyl hydroxylamines in aqueous ammonia described as Lossen degradation of the
`substrates2'.22.
`Hammett correlation of the pseudo-first-order rate constants of the degradation
`give for N-Gly-Pro-0-benzoyl hydroxylamine derivatives a correlation coefficient
`e = 0.62 (X = 4-CH3-, 3-C1-, 4-N02-, 3,5-(N0,)2, r = 0.97). The value resembles
`the value 4 = 0.86 found for the degradation of dibenzoyl hydroxylamines, with
`variation of the 0-benzoyl substituents, in 0.1 M ammonia".
`However, here no rearrangement products of a complete Lossen reaction are
`detectable (2a, Scheme 2), thus excluding the occurrence of isocyanates as inter-
`mediate~'~. The decomposition products are peptide hydroxamic acids and 0-benzoyl
`hydroxylamines, so that only distinction between hydrolysis of the ester bond (path-
`way la), and insertion of carbonyl nitrene in water (pathway 2b), is necessary.
`Recently we obtained evidence of a nitrene-generating a-elimination step followed
`by insertion in H-0 bond of a water molecule (2b, Scheme 2)22.
`In contrast to the nonenzymic degradation of the compounds, the electronic nature
`
`Downloaded by [bianca hamilton] at 11:16 20 January 2016
`
`Page 11 of 14
`
`

`

`H.U. DEMUTH e t a /
`I40
`of the leaving group does not influence the k,,, values for DP IV-catalyzed hydrolysis
`of Glycyl-Prolyl- derivatives (Table 111). Departure of 0-benzoyl hydroxylamine as
`the first product of catalysis occurs prior to the step analyzed as k,,, so that k,,, must
`reflect the rate determining deacylation of Glycyl-Prolyl-DP IV (see items 1-6, Table
`111).
`This observation is in complete agreement with similar results obtained using a
`series of substituted alanyl-prolyl-anilides as substrate^^^. In control measurements
`using Gly-Pro-4-nitroanilide as a substrate a k,,, value of 1 11 sec-' has been estimated
`supporting the above view (Table 111: mean k,,, value 109sec-I & 6%).
`Analysis of DP IV-catalyzed hydrolysis of suicide substrates with different peptide
`residues but a constant leaving group shows that: a slight variation of the deacylation
`rate constant k,,, occurs (Table 111, items 6 - 1 1). The rate constants obtained resemble
`similar values obtained" for DP IV substrate specificity, which was interpreted as
`being due to structural differences of Xaa-Pro-DP IV (acyl enzymes).
`As in nonenzymic degradation, but in contrast to the independence of k,, values
`for enzyme-catalyzed hydrolysis, N-Gly-Pro-0-benzoyl hydroxylamines exhibit a
`remarkable influence of the nature of benzoyl substituents on the inactivation of DP
`IV. A Hammett correlation of the k,,,,,-values for compounds 1-6, (Table IV) gives
`a p-value of 0.49 (r = 0.9). This result indicates electronic control by the leaving
`0-benzoyl group in the transition state of the enzyme inactivation pathway. Electron
`withdrawing substituents promote rate determining N-0 bond fission by stabilization
`of the leaving negative charged benzoyl residue.
`A comparison of the turnover values, r, of suicide substrates with a constant leaving
`group but different Xaa-residues in the P,-position demonstrates that the more
`hydrophobic or longer the side chains in substrate analog molecule are, the more
`successfully DP IV is inactivated. The enhancement of the rate of inactivation
`between the Gly-Pro- and the Lys(2)-Pro-derivatives is two orders of magnitude.
`With a turnover ratio of 8000 and an apparent pseudo-first-order inactivation rate
`constant of 0.21 min- ' N-Lys(Z)-Pro-0-(4-nitrobenzoyl) hydroxylamine is the most
`potent of the inhibitors tested and may be used in further biological studies. However
`reasons for this dramatic change in activity are unclear. Possibly, the mechanism of
`reaction of DP IV with substrate analog diacyl hydroxylamines includes a ther-
`modynamic partitioning between processes leading to hydrolysis as substrate or
`enzyme inactivation (Scheme 3, A). If so, both processes, enzyme catalyzed hydrolysis
`and inactivation, are influenced by the nature of the peptide part of the molecules
`since k,,, values increase as the size of side chain decreases (Table III), consistent with
`decreasing k,,,,, values (Table IV).
`During the intermediate steps of catalysis of suicide substrate molecule, tighter
`binding caused by bulky or hydrophobic P,-side chains may increase the lifetime of
`noncovalent or covalent complexes formed prior to the acyl enzyme so that more time
`exists for N-0 bond fission as the initial step of the inactivation; the ratio between
`catalysis and inactivation events would decrease. The lower k,,, values for DP IV
`catalyzed hydrolysis of the compounds with Ala, Leu, Phe, Lys(Z) and Lys(Z-4-N02)
`as P2- amino acids compared with the Gly-derivative may support this hypothesis
`(Table 111).
`Possible reactions occurring during interact