Synthesis and Kinetic Characterisation of Water-Soluble Fluorogenic Acyl Donors for Transglutaminase 2

DOI: 10.1002/cbic.201600048

Full Papers

Synthesis and Kinetic Characterisation of Water-Soluble

Fluorogenic Acyl Donors for Transglutaminase 2

Robert Wodtke,^{[a, c]}Georg Schramm,^{[a, d]}Jens Pietzsch,^{[a, c]}Markus Pietsch,*^[b]and

Reik Lçser*^{[a, c]}

Small glutamate-containing peptides bearing coumarin deriva-

tives as fluorescent leaving groups attached to the g-carboxylic

acid group of the Glu residue were synthesised and investigat-

ed with regard to their potential to act as substrates for trans-

glutaminase 2 (TGase 2). Their synthesis was accomplished by

an efficient solid-phase approach. The excellent water solubility

of the compounds enabled their extensive kinetic characterisa-

tion in the context of TGase 2-catalysed hydrolysis and aminol-

ysis. The influence of the coumarin skeleton’s substitution pat-

tern on the kinetic properties was studied. Derivatives contain-

ing 7-hydroxy-4-methylcoumarin (HMC) revealed properties su-

perior to those of their 7-hydroxycoumarin counterparts; anal-

ogous amides are not accepted as substrates. Z-Glu(HMC)-Gly-

OH, which exhibited the best substrate properties out of the

investigated derivatives, was selected for representative kinetic

characterisation of acyl acceptor substrates and irreversible in-

hibitors.

Introduction

The quantitative determination of enzyme activities is of out-

standing importance in biochemical research. Enzymatic assays

provide the basis for the functional characterisation of en-

zymes in terms of kinetic parameters and are crucial for study-

ing these biocatalysts in their biological contexts. In the field

of drug discovery, the availability of methods that allow assay-

ing of enzymes is a prerequisite for the identification and char-

acterisation of inhibitors. In addition, the development of sub-

strate-based probes that can efficiently monitor enzymes for

diagnostic or imaging purposes requires their evaluation on

the basis of kinetic properties. To facilitate sensitive and facile

quantification, enzyme substrates should contain moieties that

are released upon conversion and can be reliably detected by

spectroscopic methods/analytical instruments. Enzymes that

catalyse the transfer of acyl moieties to water or to alternative

nucleophiles—such as hydrolases or acyl transferases—often

accept chromophores or fluorophores as leaving groups within

the acyl donor substrate. When these chromophores or fluoro-

phores contain electron-donating groups, such as OH or NH₂,

attachment of acyl residues to these groups usually results in

attenuation of the absorption and emission intensities. Upon

enzyme-catalysed cleavage, the spectroscopic properties are

restored, and this thus couples enzymatic activity to signal in-

crease.^[1]

Transglutaminases are acyl transferases of particular impor-

tance. Among them, transglutaminase 2 (TGase 2, tissue trans-

glutaminase) seems to be most obviously involved in human

disease progression. TGase 2 is

a ubiquitously occurring

enzyme in mammals, exhibiting several important biological

functions.^[2]The eponymous and best-characterised role is its

function as a Ca²⁺-dependent catalyst of acyl transfer between

protein-bound glutamine residues and primary amines, partic-

ularly protein-bound lysine residues and low-molecular-weight

polyamines.^[3]In addition to catalysing such transamidation re-

actions, TGase 2 is able to hydrolyse the g-carboxamide group

of a glutamine residue, thus generating a glutamate residue.^[4]

The acyl transferase activity of TGase 2 is considered to be

latent under physiological conditions, but rapid activation

occurs in stress situations, such as apoptosis, in which cross-

linking of proteins stabilises dying cells.^[5]In the course of

TGase 2 research, further diverse enzymatic and non-enzymatic

functions that might be more relevant under physiological

conditions for maintaining the integrity of the cell have been

ascribed to this enzyme.^[6]

[a] R. Wodtke, Dr. G. Schramm, Prof. Dr. J. Pietzsch, Dr. R. Lçser

Helmholtz-Zentrum Dresden–Rossendorf

Institut fꢀr Radiopharmazeutische Krebsforschung

Bautzner Landstrasse 400, 01328 Dresden (Germany)

E-mail: r.loeser@hzdr.de

[b] Dr. M. Pietsch

Zentrum fꢀr Pharmakologie, Medizinische Fakultꢁt, Universitꢁt zu Kçln

Gleueler Strasse 24, 50931 Kçln (Germany)

E-mail: markus.pietsch@uk-koeln.de

[c] R. Wodtke, Prof. Dr. J. Pietzsch, Dr. R. Lçser

Fachrichtung Chemie und Lebensmittelchemie

Technische Universitꢁt Dresden

Mommsenstrasse 4, 01062 Dresden (Germany)

TGase 2, particularly its acyl transferase activity, contributes

to different pathophysiological conditions, such as celiac dis-

ease,^[7]diseases related to fibrotic processes,^[8]neurodegenera-

tive disorders^[9]and cancer.^[10]Therefore, this enzyme appears

to be an attractive target for pharmacological inhibition to po-

[d] Dr. G. Schramm

Present address: Department of Imaging and Pathology, KU Leuven

UZ Herestraat 49, 3000 Leuven (Belgium)

Supporting information for this article can be found under htt p ://

dx.doi.org/10.1002/cbic.201600048.

ChemBioChem 2016, 17, 1 – 20

1

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

tentially treat these diseases. In addition, the development of

imaging agents targeted against TGase 2 should provide fur-

ther insight into its pathological functions and should support

clinical translation of inhibitors.

Here we describe the synthesis and kinetic characterisation

of six new glutamate-derived fluorogenic acyl donors. A solid-

phase synthesis strategy was developed for this purpose, and

the compounds were evaluated in terms of their enzymatic hy-

drolysis and aminolysis, both by standard regression analysis

and by numerical integration. Together with investigations into

different peptidic scaffolds for recognition by TGase 2, the in-

fluence of substitutions on the fluorophore was explored. Be-

cause all compounds exhibit distinct rates of spontaneous re-

actions, this feature was in turn used to assess their solubilities.

In addition to the coumarinyl esters, two coumarinyl amides

were synthesised in order to determine their potential to act

as fluorogenic acyl donors. Furthermore, the applicability of

the new acyl donors for kinetic characterisation of amine-

based acyl acceptor substrates [aminoacetonitrile and N-(bioti-

nyl)cadaverine] and irreversible inhibitors [iodoacetamide and

the recently described N^a-phenylacetyl-N^e-acryloyl-lysine-4-(6-

methylpyridine-2-yl)piperazide^[14]] was verified.

One of the prerequisites for the development of TGase 2 in-

hibitors is the availability of reliable assay methods for precise

measurement of the enzymatic activity, which should enable

the evaluation of the compounds’ inhibitory potencies. Various

continuous and discontinuous assay methods for TGase 2 have

been described so far and were recently discussed in detail.^[11]

Among these assay methods, the continuous fluorimetric activ-

ity assay described by Gillet et al.^[12]seems to be highly suita-

ble for inhibitor characterisation because it allows for the sen-

sitive detection of TGase 2 activity by measurement of increas-

es in fluorescence (Scheme 1).

Synthetic access to the required fluorogenic acyl donors,

each consisting of a Z-protected amino acid [Z: benzyloxycar-

bonyl, also referred to as carbobenzoxy (Cbz) group] connect-

ed to umbelliferone (7-hydroxycoumarin, HC) through a g-ami-

nobutyroyl (g-Abu) linker acting as mimetic of the glutamine

side chain, is fairly easy. In addition, the assay allows for the Results and Discussion

kinetic characterisation both of inhibitors and of acyl acceptor

Synthesis and kinetic evaluation of the g-Abu-derived acyl

substrates. Therefore, we set out to establish this fluorimetric

assay in our laboratories by using the fluorogenic acyl donor

Z-Phe-g-Abu-HC. However, as outlined later in this article, both

Z-Phe-g-Abu-HC and its 7-hydroxy-4-methylcoumarin-based

(HMC-based) derivative proved to have the disadvantage of

low solubility in aqueous media. Depending on the experi-

mental setup, this might limit the applicability of these com-

pounds for detailed kinetic studies, because such investiga-

tions require activity measurements over a broad range of sub-

strate concentrations. In particular, simultaneous measure-

ments of various substrate concentrations in a microplate

format can be more affected by solubility limits than single

measurements in a cuvette,^[12]due to the different time delays

caused by preparation.

donors

Initially, it was envisaged that a fluorimetric activity assay for

TGase 2 could be established by using the acyl donors de-

scribed by Gillet et al.^[12]Because the reported kinetic data

were more favourable for Z-Phe-g-Abu-HC (2a) than for its gly-

cine derivative Z-Gly-g-Abu-HC, we focused on the former

compound. Moreover, other coumarin derivatives—that is,

HMC and 7-amino-4-methylcoumarin (AMC)—were employed

as fluorogenic leaving groups. The syntheses of these three

compounds were each accomplished in two steps by the pub-

lished procedure with some modifications, starting from com-

mercially available Z-Phe-OH (Scheme 2 and Discussion S1 in

the Supporting Information for detailed experimental proce-

dures).^[12]After purification by column chromatography and re-

crystallisation, the fluorogenic acyl donors 2a–2c were ob-

tained in satisfactory yields (55–68%) and high purities.

We therefore set out to develop analogues with improved

solubility. A general approach to making compounds more

soluble in water is the attachment of ionisable functionalities

such as carboxylic acid groups.^[13]Applied to the TGase 2 sub-

strates discussed above, such derivatisation seems to be most

easily achievable by replacing the g-aminobutyroyl moiety

with glutamate.

Initial investigations were focussed on the enzymatic aminol-

ysis of the acyl donors 2a and 2b at pH 8.0 with aminoaceto-

nitrile as acyl acceptor in the presence of (ꢀ)-threo-dithiothrei-

tol (DTT) as antioxidant. As in the reported assay procedure,

Scheme 1. Principle of the fluorimetric TGase 2 activity assay developed by Gillet et al.^[12]Release of 7-hydroxycoumarin (HC) by TGase 2-catalysed hydrolysis

or aminolysis of Z-Phe-g-Abu-HC results in a measureable increase in fluorescence.

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

2

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

ing V_max, “since the concentration of substrate which saturates

an aqueous solution is insufficient to saturate the enzyme”.^[15]

To investigate the solubility behaviour of the acyl donors in

more detail, we decided to take advantage of their spontane-

ous decay that takes place in the absence of TGase 2 and ami-

noacetonitrile. Because these reactions each follow pseudo-

first-order kinetics, plots of initial rates (v_0control) versus concen-

tration should display a linear dependence in the case of com-

plete solubility.^[16]The results of this investigation confirmed

the initial observations on the low solubilities of the g-Abu-de-

rived acyl donors, because the curves deviated from linearity

even at low concentrations (<15 mm, Figure S2) before precipi-

tation could be observed visually, thus indicating aggregation

prior to precipitate formation. Moreover, as shown in Figure S2,

the solubilities of the two g-Abu derivatives are highly time

dependent and strongly influenced both by the organic co-sol-

vent and by the antioxidant present.

Scheme 2. Synthesis of the g-aminobutyric acid-derived fluorogenic acyl

donors 2a–2c. Reagents and conditions: a) N-methylmorpholine (NMM),

isobutyl chloroformate (IBCF), g-Abu-OH, aqueous NaOH, THF, ꢂ30!108C,

2.5 h; b) HC/HMC, N,N-dicyclohexylcarbodiimide (DCC), 4-(dimethylamino)-

pyridine (DMAP), CH₂Cl₂/THF (4:1), 25 h; c) AMC, NMM, IBCF, THF/DMF (12:1),

ꢂ258C!108C, 5 h.

However, to demonstrate their potential as substrates, com-

pounds 2a and 2b were evaluated in terms of gpTGase 2-cata-

lysed hydrolysis within their solubility limits (Figure S3). Non-

linear regression according to the Michaelis–Menten equation

[Experimental Section, Eq. (4)] revealed K_mand k_catvalues of

1.47 mm and 0.30 s^ꢂ1, respectively, for acyl donor 2a and

1.14 mm and 0.52 s^ꢂ1, respectively, for acyl donor 2b, resulting

a final concentration of 5% DMF or DMSO was used to achieve

sufficient solubility of the acyl donors in aqueous media.^[12]The

increase in fluorescence was monitored in 96-well microplates

at different concentrations of 2a (5–40 mm) and 2b (1–20 mm),

and initial rates (v_0total) were calculated as outlined in the Ex-

perimental Section. Unexpectedly, plots of v_0totalversus the acyl

donor concentration for both compounds each exhibited a

maximum at ꢁ5 mm, after which the rates suddenly dropped

and further increased with rising substrate concentrations (Fig-

ure S1). Concomitantly, visual inspection of plate wells revealed

the formation of precipitates even at acyl donor concentra-

tions below 20 mm, independently of the organic co-solvent

used.

in performance constants of 204000 and 456000m^ꢂ1^ꢂ1, re-

s

spectively (Table 1). These results clearly confirm the favourable

substrate properties of the g-Abu derivatives.^[12]Interestingly,

despite similar K_mvalues, substitution of HC by HMC resulted

in a k_catvalue for acyl donor 2b almost twice that for 2a. Be-

cause of their low K_mvalues, both coumarinyl esters can be ap-

plied for investigations into the hydrolytic activity of TGase 2

up to concentrations of 3–6ꢁK_m, depending on the buffer con-

ditions used.

In contrast to the coumarinyl esters 2a and 2b, neither

spontaneous nor enzymatic release of AMC was observed for

coumarinyl amide 2c (data not shown). This is discussed later

in this article.

This observation suggested that the anomalous Michaelis–

Menten plots are caused by limited solubility of 2a and 2b.

Generally, a poorly soluble substrate will follow the Michaelis–

Menten hyperbola only up to the solubility limit without reach-

Table 1. Kinetic parameters for the TGase 2-catalysed hydrolyses of the acyl donors 2–6 at pH 8.0 and 308C.

Compound

Regression analysis

k_cat[s^ꢂ1

Numerical integration

k_cat[s^ꢂ1

K_m[mm]

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

K_m[mm]

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

gpTGase 2

1.47 (0.02)

1.14 (0.16)

12.1 (0.5)

17.7 (0.5)

2.83 (0.17)

2.53 (0.14)

11.0 (0.6)

2.66 (0.11)

hTGase 2

12.7 (0.9)

8.62 (0.09)

13.6 (1.3)

2a

2b

3

0.30 (0.03)

0.52 (0.03)

0.50 (0.01)

0.29 (0.03)

0.71 (0.02)

0.76 (0.02)

0.64 (0.03)

0.55 (0.02)

204000

456000

41300

1.24 (0.06)

1.20 (0.27)

10.9 (0.9)

0.25 (0.01)

0.46 (0.04)

0.43 (0.01)

0.23 (0.02)

0.62 (0.02)

0.71 (0.02)

0.55 (0.01)

0.54 (0.02)

202000

383000

39500

4

16400

12.4 (0.6)

18500

5a

5b

6a

6b

251000

300000

58200

2.57 (0.46)

2.49 (0.23)

8.87 (0.64)

2.14 (0.13)

241000

285000

62000

207000

252000

5a

5b

6b

0.84 (0.05)

0.68 (0.02)

0.50 (0.03)

66100

78900

36800

9.72 (0.62)

6.06 (0.16)

7.73 (0.54)

0.74 (0.04)

0.64 (0.02)

0.48 (0.03)

76100

106000

62100

For details on calculation of the kinetic parameters see the Experimental Section. Data shown are mean values (ꢀSEMs) of three separate experiments,

each performed in duplicate. Active concentrations of TGase 2 from guinea pig liver (E_T=31.6–35.3 nm, various lots) and human (E_T=30.8 nm) were calcu-

lated from Zedira’s activity data or determined by active site titration as recently described,^[30]respectively.

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

3

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

Synthesis and spontaneous reactivity of the glutamate-

derived acyl donors

the influence of methyl substitution in the coumarin fluoro-

phore on the substrate properties towards TGase 2, HC deriva-

tives of the most favourable peptidic structures Z-Glu-Gly-OH

and Z-Phe-Glu-Gly-OH (see Kinetics Section) were also synthes-

ised. Finally, Z-Glu(AMC)-Gly-OH (5c) was prepared to ascertain

whether or not coumarinyl amides can be used as fluorogenic

acyl donors for TGase 2.

To overcome the low solubilities, derivatisation of fluorogenic

acyl donors with a carboxylic acid group as a solubilising

moiety was envisaged.^[13]From a synthetic point of view, the

most straightforward way to achieve this seemed to be by re-

placement of the g-aminobutyroyl linker that connects Z-Phe

and the fluorophore by a glutamoyl moiety, thus resulting in

Z-Phe-Glu(fluorophore)-OH as the basic dipeptidic structure.

Moreover, this scaffold allows for the C-terminal extension of

the scaffold structure by an additional amino acid such as gly-

cine, which would place the solubility-mediating carboxylic

acid group further away from the reaction centre.

A major challenge to the synthesis of the glutamate-derived

acyl donors involves the regioselective installation of the fluo-

rogenic leaving group in the side chain of the glutamate resi-

due. Therefore, a strategy based on solid-phase synthesis with

use of orthogonal protecting groups was developed to provide

the different glutamate-derived acyl donors (Scheme 3). In ad-

dition, this modular approach should allow efficient access to

a broad spectrum of potential acyl donors. For this purpose,

the 2-chlorotrityl chloride (2-ClTrtCl) resin was considered to

be the optimal polymeric support because the anchoring

moiety readily reacts with carboxylic acid groups, and the cor-

responding highly acid-labile 2-chlorotrityl esters can be easily

cleaved under mildly acidic conditions (see below).^[17]Final

esterification at the glutamate side chain required careful opti-

misation with regard to the selection of the coupling reagents

(see Discussion S2). It was accordingly found that the use of

In order to define the minimal structural motif for efficient

TGase 2-catalysed conversion, the effect of deleting the N-ter-

minal Phe moiety had to be studied. Therefore, Z-Glu(fluoro-

phore)-OH, Z-Glu(fluorophore)-Gly-OH and Z-Phe-Glu(fluoro-

phore)-Gly-OH were examined as amino-acid-based and pepti-

dic scaffolds for substrate synthesis in addition to Z-Phe-Glu-

(fluorophore)-OH. Because g-Abu derivative 2b exhibits more

favourable substrate properties than 2a, HMC was selected as

fluorophore for all peptidic scaffolds. However, to investigate

Scheme 3. Solid-phase synthesis of the glutamate-derived fluorogenic acyl donors 3–6. Reagents and conditions: a) 1. Fmoc-AA, DIPEA, CH₂Cl₂, 4 h, 2. CH₂Cl₂/

CH₃OH/DIPEA (17:1:2); b) 20% piperidine/DMF; c) Z-OSu, DIPEA, CH₂Cl₂, 17 h; d) Z-Phe-OH, HBTU, DIPEA, DMF; e) 1. Fmoc-Glu(OAll)-OH, HBTU, DIPEA, DMF, 2.

20% piperidine/DMF; f) Pd(PPh₃)₄, CH₂Cl₂/NMM/HOAc (8:2:1), 4 h, Ar; g) HATU, DIPEA, HMC or HC or AMC, DMF, 5 h; h) HFIP/CH₂Cl₂(1:4), 3ꢁ10 min.

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

4

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

the uronium-based coupling agent (1-[bis(dimethylamino)-

methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxid hexafluor-

ophosphate (HATU) and N,N-diisopropylethylamine (DIPEA) as

base gave complete conversion of the carboxylic acids into the

corresponding coumarinyl esters (Discussion S3).^[18]Subsequent

release of the acyl donors from the resin was achieved under

mild conditions by treatment with a 1:4 mixture of hexafluoro-

isopropanol (HFIP) and dichloromethane.^[19]In this context, the

use of HFIP instead of trifluoroacetic acid (TFA) ensures the in-

tegrity of the coumarinyl esters, because they are potentially

susceptible to acid-mediated hydrolysis. All compounds were

obtained in sufficient yields (26–36%) and high purities after

purification by preparative RP-HPLC (see Discussion S4 for de-

tails of the experimental procedures).

Figure 1. Spontaneous reactions of the fluorogenic acyl donor 5b. Plots of

v_0control=f([5b]) at pH 8.0 and 308C in the presence of 5% DMSO and

~

!

*

500 mm DTT ( ), 500 mm TCEP ( ), no additive ( ) and 400 mm aminoaceto-

^

nitrile ( ). Analysis by linear regression (c) gave rate constants k_obsof

(24.3ꢀ1.1)ꢁ10^ꢂ3min^ꢂ1(DTT), (17.1ꢀ0.4)ꢁ10^ꢂ3min^ꢂ1(TCEP),

(17.5ꢀ0.2)ꢁ10^ꢂ3min^ꢂ1(no additive) and (15.3ꢀ0.2)ꢁ10^ꢂ3min^ꢂ1(aminoace-

tonitrile). Data shown are mean values ꢀSEMs of two separate experiments,

each performed in duplicate. When not apparent, error bars are smaller than

the symbols.

The established modular solid-phase synthesis approach

allows facile access to a variety of peptidic coumarinyl esters in

small amounts. In an alternative approach, Chung et al. pre-

pared the analogous chromogenic acyl donor Z-Glu(OpNp)-

Gly-OH in solution by starting from Z-protected glutamic anhy-

dride, which offers the advantage that the final product can be

obtained at a larger scale.^[20]

Table 2. Pseudo-first-order rate constants (k_obs) for the spontaneous reac-

tions of the different acyl donors at pH 8.0 and 308C.

The transfer of the conditions for on-resin esterification to

the coupling of AMC to the glutamate side chain carboxylic

acid group of resin-linked Z-Glu-Gly-OH yielded only minor

amounts of the desired coumarinyl amide 5c (yield 3%, Dis-

cussion S5) along with different side products. This might be

the result of the less nucleophilic character of the aromatic

amine AMC under these conditions, relative to the aromatic al-

cohols HMC and HC. To increase the yield for this amidation,

different coupling procedures were tested with limited success

(for details and further information see Discussion S5). Howev-

er, the amounts of 5c obtained by the various synthetic meth-

ods were sufficient for initial investigations into its behaviour

towards TGase 2.

k_obs(ꢁ10^ꢂ3min^ꢂ1

)

Acyl donor

DTT

TCEP

k_obs(TCEP)/k_obs(DTT)

2a

2b

3

15.8 (0.6)

14.3 (0.5)

–

12.2 (0.2)

25.1 (2.8)

24.3 (1.1)

29.2 (4.0)

32.7 (1.1)

2.11 (0.05)

1.86 (0.08)

4.60 (0.36)

4.75 (0.13)

16.9 (0.6)

17.1 (0.3)

18.1 (0.5)

25.3 (0.4)

0.13

–

0.39

0.67

0.70

0.62

0.77

4

5a

5b

6a

6b

Data shown are mean values ꢀstandard errors of the means (ꢀSEMs) of

two to three separate experiments, each performed in duplicate. The cor-

responding plots of v_0control=f([acyl donor]) are given in the Supporting

Information (Figure S5).

To determine the solubilities of the new glutamate-derived

fluorogenic acyl donors 3-6, the compounds were investigated

with regard to the rates of their spontaneous reactions at

pH 8.0 in the presence of 5% DMSO, as previously done for

2a and 2b. The resulting plots showed linearity up to

concentrations of 250 mm (see Figure 1 for the acyl donor 5b

and Figure S4); this highlights that the presence of the

carboxylic acid group significantly increases the solubility rela-

tive to the g-Abu derivatives. Whereas linearity predominates

over the entire concentration range for compound 4, a slight

bend between 100 and 200 mm is visible for compound 5b;

this indicates a diminished solubility at 200 and 250 mm. This

behaviour can also be observed for the other glutamate-

derived acyl donors 3, 5a, 6a and 6b (Figure S4). Neverthe-

less, enzyme-kinetics investigations up to substrate concentra-

tions of 200–250 mm should also be possible for these acyl

donors, because they still exhibit sufficient solubility in this

concentration range. Notably, none of the glutamate-derived

acyl donors forms any visible precipitate up to concentrations

as high as 500 mm. Furthermore, as indicated in Figure S4, the

extents of the spontaneous reactions vary within the different

acyl donors. To quantify these characteristics, pseudo-first-

order rate constants (k_obs) for the spontaneous reactions, sum-

marised in Table 2, were determined for all ester-based acyl

donors.

Because the assay of TGase 2 requires the presence of an an-

tioxidant to prevent oxidation of the active-site cysteine resi-

due, both DTT and tris(2-carboxyethyl)phosphine (TCEP) were

employed for investigations into the spontaneous reactivity of

the compounds at a concentration of 500 mm. In the presence

of TCEP, the obtained k_obsvalues reveal distinct spontaneous

reactions for all compounds; these seem to depend on the

kind of the peptidic scaffold and thus allow for a classification

into three groups. The g-aminobutyroyl derivatives 2a and 2b

display the lowest k_obsvalues. Compounds 3 and 4, each bear-

ing a free carboxylic acid group directly on the glutamate resi-

due, exhibited pseudo-first-order rate constants approximately

twice those of their descarboxy analogues. Interestingly, C-ter-

minal extension with glycine resulted in k_obsvalues that were

increased by a factor of approximately four, as observed for

the acyl donors 5a, 5b, 6a and 6b.

These results suggest that the carboxylate group might act

as an intramolecular general base to facilitate the nucleophilic

attack of water on the coumarinyl ester bond. General intramo-

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

5

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

Scheme 4. Structures of identified products from spontaneous reactions of 5b.

lecular base catalysis is obviously more effective for the glycine

derivatives, because a spatial orientation of the glycine carbox-

ylic acid group in proximity to the coumarinyl ester bond is

probably sterically less strained than that of the glutamate car-

boxylic acid group (Figure S5).

(Scheme 4 and Figures S6 and S7). The 5b-derived DTT-thioest-

er and O-ester, as well as the corresponding glutarimide 5d,

were also detectable under the assay conditions (pH 8.0,

500 mm DTT) in the absence of TGase 2 (Figure S8). Notably,

the rate of spontaneous reaction in the absence of DTT and

TCEP is nearly identical to that observed in the presence of

TCEP, which illustrates that the presence of TCEP does not con-

tribute to the spontaneous disintegration of the acyl donors

(Figure 1). Glutarimide formation from 5b under assay condi-

tions was unexpected; however, this cyclisation does not sub-

stantially contribute to the spontaneous disintegration of 5b

(Figure S8).

Monitoring of the spontaneous reactions in the presence of

DTT resulted in k_obsvalues up to 1.6 times greater than those

determined in the presence of TCEP for the glutamate-derived

acyl donors and up to 7.6 times greater for the g-Abu deriva-

tives 2a and 2b (Figure 1 and Table 2). Because of the nucleo-

philic character of its two thiol groups, DTT might contribute

to the spontaneous reaction of the ester substrates. This path-

way of spontaneous substrate disintegration is very likely to

occur; reactions between activated aryl esters and thiols to

afford thioesters having previously been reported.^[21]In gener-

al, thioesters are distinctly more stable towards hydrolysis than

the corresponding O-esters,^[22]so thiolysis of the coumarinyl-

ester-based substrates by DTT should be considered an inde-

pendent pathway for spontaneous reaction. To validate this as-

sumption and to identify the thiolysis products, compound 5b

was exposed to a tenfold excess of DTT in acetonitrile as sol-

vent in the presence of one equivalent of triethylamine. After

5 h at room temperature, analysis of the reaction mixture by

RP-HPLC indicated the formation of three major components

in addition to remaining reactant, the released 4-methylumbel-

liferone and the corresponding g-carboxylic acid. These were

identified as DTT-thioester, DTT-O-ester and glutarimide 5d

Another type of spontaneous reaction that can occur during

investigations into TGase 2-catalysed aminolysis is spontaneous

aminolysis. However, as shown for compound 5b, for example,

in Figure 1, no increased rate constant for spontaneous disinte-

gration was observed in the presence of aminoacetonitrile and

concomitant absence of any reducing agent.

In conclusion, because all coumarinyl esters exhibit distinct

rates of spontaneous reactions at pH 8.0, these nonenzymatic

conversions must be considered for analyses of TGase 2-medi-

ated hydrolyses and aminolyses to ensure correct data evalua-

tion. For this purpose, control measurements in the absence of

TGase 2 were carried out for the characterisation of all sub-

strates to determine the proportions of the enzyme-catalysed

reactions.

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

6

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

Kinetic analysis of TGase 2-catalysed conversions of the

glutamate-derived acyl donors

a TGase 2 concentration of 3 mgmL^ꢂ1provides the best com-

promise between signal/background ratio and speed of enzy-

matic reaction. If the concentration of the released fluorophore

exceeds 20 mm during the enzymatic conversion of the acyl

donor, fluorescence quenching might contribute to the bend

of the curves. Hence, progress curves for the enzymatic hydrol-

yses were analysed over the first 300 s by nonlinear regression

with use of Equation (1),^[24]which maintained the limits for the

fluorescence coefficient discussed above.

TGase 2-catalysed hydrolyses: Prior to the characterisation of

the different acyl donors, fluorescence coefficients for the fluo-

rophores HC and HMC were determined. As can be seen in

Figure S17, sufficient linearity between the concentration of

fluorophore and the associated fluorescence signal is predomi-

nant up to a concentration of 20 mm. Example kinetic runs for

the enzymatic hydrolysis and the spontaneous reaction at

pH 8.0 are depicted for compound 5b in Figure 2. The enzy-

Michaelis–Menten plots for the enzymatic hydrolysis of com-

pound 5b by gpTGase 2 and by hTGase 2 at pH 8.0 are depict-

ed in Figure 3 (for the other compounds, see Figures S11 and

Figure 3. gpTGase 2- and hTGase 2-catalysed hydrolysis of the acyl donor

5b at pH 8.0. Plots of v_0corr=f([acyl donor]) with the nonlinear regressions

(c) with use of Equation (4) (Michaelis–Menten equation, Experimental

Section). Data shown are mean values ꢀSEMs of three separate experi-

ments, each performed in duplicate. When not apparent, error bars are

smaller than the symbols. Conditions: pH 8.0, 308C, 5% DMSO, 500 mm

DTT (for gpTGase 2) or 500 mm TCEP (for hTGase 2), 3 mgmL^ꢂ1of either

gpTGase 2 or hTGase 2.

S12). The obtained kinetic parameters for all compounds are

summarised in Table 1. The ratio k_cat/K_m, which we decided to

denote as performance constant according to Koshland’s sug-

gestion, can be regarded as the most appropriate parameter

for comparing the synthesised fluorogenic acyl donors with

regard to their TGase 2-catalysed conversion.^[25]Relative to 2b,

changing g-Abu to glutamate resulted in a significantly de-

creased k_cat/K_mvalue for compound 4 (456000m^ꢂ1s^ꢂ1and

16400m^ꢂ1s^ꢂ1for 2b and 4, respectively), mainly due to an in-

crease in K_m. This result indicates a detrimental effect on the af-

finity towards gpTGase 2 when the carboxylic acid group is di-

rectly attached at the glutamate residue; this is in agreement

with previous results observed with substrates that contain

glutamine^[26]or with inhibitors based on reactive glutamine an-

alogues.^[27]Therefore, a shift of the carboxylic acid group fur-

ther away from the reaction centre might be beneficial for the

recognition by gpTGase 2.

Figure 2. Progress curves for A) TGase 2-catalysed hydrolysis, and B) sponta-

neous reaction of acyl donor 5b. Conditions: pH 8.0, 308C, 5% DMSO,

500 mm DTT, 3 mgmL^ꢂ1gpTGase 2.

matic measurements for compound 5b generated nonlinear

progress curves over 900 s; these have also been observed for

the other compounds, including the g-aminobutyric acid deriv-

atives 2a and 2b. To ensure that inactivation of TGase 2 does

not occur due to protein instability under the experimental

conditions (which could also account for nonlinear progress

curves), the Selwyn test was performed with acyl donor 5b.^[23]

The results clearly indicate the absence of enzyme inactivation

over the entire measurement period of 900 s: plots of RFU=

f([E]*t) are almost congruent for three different enzyme con-

centrations (Figure S9). Therefore, nonlinear progress curves

result from rapid enzymatic conversion of the acyl donors,

which gradually leads to substrate depletion for TGase 2. Typi-

cally, this effect can be simply reduced by using a lower

enzyme concentration, as demonstrated by the Selwyn test,

which was performed at 1, 2 and 3 mgmL^ꢂ1of TGase 2. Howev-

er, with the reduced amount of enzyme the relative proportion

of the spontaneous reaction increases (Figure S10). Despite the

fact that the portion of the enzymatic rate even at the lowest

enzyme concentration is greater than that of the spontaneous

rate (over the applied range of substrate concentration),

Indeed, compound 6b, with Z-Phe-Glu-Gly-OH as peptidic

scaffold, shows a performance constant more than one order

of magnitude higher (k_cat/K_m=207000m^ꢂ1s^ꢂ1) than that of 4.

In order to define the minimal peptidic motif for efficient

TGase 2-mediated conversion, N-terminal deletion of the Phe

residue was envisaged. For this purpose, we synthesised com-

pounds 3 and 5b, derived from Z-Glu-OH and Z-Glu-Gly-OH,

respectively. From the kinetic data, it can be concluded that

the presence of the N-terminal phenylalanine residue does not

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

7

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

seem to be crucial for efficient conversion by gpTGase 2 be-

cause the performance constants even increase upon its dele-

tion (compare 3 and 4 as well as 5b and 6b). Again, the pres-

ence of a C-terminal glycine residue positively influenced the

recognition by gpTGase 2 (k_cat/K_mvalues of 41300 and

300000m^ꢂ1s^ꢂ1for 3 and 5b, respectively). In order to evaluate

the influence of the fluorophore on the substrate properties,

HC analogues of the best HMC-bearing substrates 5b and 6b

were synthesised. Surprisingly, whereas 5a displays almost un-

altered kinetic parameters, characterisation of tripeptide 6a re-

vealed a considerably lower performance constant in compari-

son with its methyl-substituted counterpart 6b. This result sug-

gests that the methyl group in the 4-position of the coumarin

system might be beneficial but in detail its influence seems to

depend on the peptidic scaffold. Finally, comparison of 5b

(k_cat/K_m=300000m^ꢂ1s^ꢂ1) with acyl donors reported in the liter-

ature showed that the compound displays not only the most

favourable substrate properties of all readily soluble fluorogen-

ic substrates investigated in this study, but also belongs to the

kinetically most favourable acyl donor substrates for gpTGase 2

described so far.^[11]However, it should be pointed out that, out

of all investigated fluorogenic acyl donors, compound 2b (k_cat/

envisaged in order to test whether or not separate recording

of spontaneous and enzymatic reactions is of sufficient accura-

cy to allow calculation of the kinetic parameters for the

enzyme-catalysed reactions. For this purpose, the differential

equations for the time dependence of [E], [SL], [ESL] and [P] of

the classical two-step Michaelis–Menten model extended by

the pathway for spontaneous reaction of the acyl donor sub-

strates were formulated (Scheme 5) and implemented in the

Scheme 5. Kinetic model for substrate hydrolysis by TGase 2 and the corre-

sponding differential rate laws derived from it. The kinetic model was pre-

pared according to ref. [32]. For all calculations, rate constants k_acyland k_deacyl

were combined to rate constant k_cataccording to the equation: k_cat

=

k_acylꢁk_deacyl/(k_acyl+k_deacyl). Abbreviations used in the scheme. E: enzyme. SL:

substrate (fluorogenic acyl donor). ESL: initial enzyme·substrate complex. L:

coumarin derivative. E–S: acylated enzyme intermediate. P: product.

K_m=456000m^ꢂ1^ꢂ1) clearly exhibits the best kinetic properties,

s

even though it is the least soluble compound of the series.

In accordance with this result, the corresponding non-fluo-

rescent substrate Z-Gln-Gly-OH exhibits favourable kinetic pa-

rameters towards gpTGase 2^[26]and a microbial transglutami-

nase.^[28]In addition to the evaluation of the acyl donors to-

wards gpTGase 2, compounds 5a, 5b and 6b, which display

favourable substrate properties towards that enzyme, were

also characterised with regard to their hydrolysis catalysed by

human TGase 2. The two orthologues share 83% sequence

identity, which in turn makes TGase 2 from guinea pig liver

a cost-efficient model for the human enzyme (Figure S13).^[29]

Notably, all three compounds are also suitable substrates for

hTGase 2, even though performance constants are diminished

relative to those of the guinea pig enzyme (Table 1).

freely available statistical software R (www.r-project.org/; see

the Experimental Section for further details). The kinetic pa-

rameters K_mand k_catobtained from nonlinear regressions were

then used as initial values for numerical integration of the cor-

responding progress curves (same data sets as for analysis by

nonlinear regression) over the first 300 s. Global fits to the data

from one example experiment for compound 5b are shown in

Figure 4. As is obvious, a global fit to the experimental data

provided the best-fit traces of sufficient quality for all concen-

Their K_mvalues are four to five times higher than for the

gpTGase 2. Trends are similar to those for gpTGase 2, and com-

pound 5b turned out to be the acyl donor with the most

favourable substrate properties towards the human enzyme

(k_cat/K_m=78900m^ꢂ1s^ꢂ1). A Michaelis constant similar to that of

5b has been reported for the chromogenic analogue Z-

Glu(OpNp)-Gly-OH towards hTGase 2, with a similar tendency

being found on comparison with the guinea pig enzyme.^[31]

An experimental setup with 96-well plates was used to

record spontaneous and enzymatic hydrolyses simultaneously,

so there is a time delay between the start of the reaction and

the data acquisition. In consequence, those substrate concen-

trations pipetted at earlier time points might be lower than

intended at the start of the measurement. Although it was at-

tempted to minimise this problem by pipetting in the right

order of substrate concentrations—that is, from low to high—

it cannot be completely avoided. The resulting variation of the

substrate concentrations is not considered by the classical re-

gression analysis but can be quantified by analysis with numer-

ical integration. Furthermore, application of this method was

Figure 4. Global fit of progress curves for gpTGase 2-catalysed hydrolysis of

acyl donor 5b at pH 8.0. Concentrations of 5b at the start of the reaction

~

!

&

*

^

were 0.5 ( ), 1 ( ), 2 (+), 3 (ꢁ), 5 ( ), 7.5 ( ), 10 ( ) and 15 mm (*). Data

shown [meanꢀstandard deviation (ꢀSD)] are from one experiment that

was performed in duplicate. The calculated concentrations at the time point

of data acquisition are given in the box in the top-left corner. Fitted values

for K_mand k_catare given in the box in the bottom-right corner; values calcu-

lated from three independent experiments can be found in Table 1. Condi-

tions: pH 8.0, 308C, 5% DMSO, 500 mm DTT, 3 mgmL^ꢂ1gpTGase 2.

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

8

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

trations. In this context, deviations of fit traces from experi-

mental data are a common problem of analysis by global fit-

ting because “it is not permissive to even small deviations

from ideal progress curves, that is, data should not contain

errors and should perfectly adhere to the model”.^[33]With

regard to the calculation of mean values and errors for K_mand

k_cat, the experiments for each acyl donor were separately ana-

lysed by global fitting followed by statistical analysis of the

obtained fitted parameters (Table 1). Notably, both methods

for data analysis gave reasonably comparable results for the

values of K_mand k_cat. Thus, the approach used to determine

the proportion of enzyme-catalysed reaction within the regres-

sion analysis by simple subtraction of the rates for spontane-

ous reaction from those of the overall reaction seems to be of

sufficient accuracy. However, the systematic deviation to lower

values for the parameters derived from numerical integration

in relation to those obtained by nonlinear regression might

occur due to the corrected substrate concentrations, which

highlights the advantage of the former method.

To obtain further information on the kinetic behaviour of

compound 5c, the gpTGase 2-catalysed hydrolysis of coumar-

inyl ester 5b in the presence of increasing concentrations of

5c was studied (Discussion S6). Judged from these results, 5c

has an inhibitory effect on the enzymatic hydrolysis of 5b. A

detailed kinetic analysis of the recorded data can be found in

Discussion S6. In summary, the obtained results indicate that

the coumarinyl amide seems most likely to act as a linear (b=

0) mixed, predominantly specific (1<a<1) inhibitor accord-

ing to the systematic classification of enzyme-modifier interac-

tions suggested by Baici¹.^[36]In view of the high structural simi-

larity between the coumarinyl ester 5b and the coumarinyl

amide 5c, the observed pronounced competitive component

for inhibition of gpTGase 2 by 5c is comprehensible. On the

other hand, binding of 5c to sites distinct from the active site

also appears to be reasonable if one considers the multi-

domain structure of the TGase 2 protein, which is known to

harbour several binding sites for various ligands apart from the

catalytic centre. As a representative example, Yi et al. recently

reported on an acylidenoxindole derivative that also exhibits

inhibitory potency of noncompetitive character on the transa-

midase activity.^[37]

Kinetic behaviour of coumarinyl amide 5c: The distinct

spontaneous reaction of the coumarinyl esters requires careful

handling of these compounds and more complex data evalua-

tion. Therefore, a substrate analogue with substrate proper-

ties similar to those of the coumarinyl esters but that is devoid

of spontaneous disintegration would be advantageous. A

promising way to obtain such a substrate seemed to be the

exchange of the ester bond between the fluorophore and the

peptidic scaffold for an amide bond. From the investigations

done in this study with regard to the substrate properties of

different peptidic scaffolds, the dipeptide Z-Glu-Gly-OH

seemed to be most suitable for performing such a sub-

stitution. Therefore, compound 5c, bearing AMC attached at

the carboxylic acid group of the glutamate side chain, was syn-

thesised.

TGase 2-catalysed aminolysis: Transglutaminases are special-

ised in the catalysis of transamidation reactions with primary-

amine-derived acyl acceptors, so the kinetics of TGase 2-cata-

lysed aminolysis of the glutamate-derived acyl donors—with

aminoacetonitrile as amine substrate—were investigated in ad-

dition to hydrolysis.

The corresponding kinetic data are shown in Table 3, with

the parameters K_mand k_catbeing of apparent character (see

Discussion S7 for detailed explanation). Nevertheless, K^a_m^ppand

app

cat

k

can be used to evaluate the different substrates because

they were determined under identical conditions at an aminoa-

cetonitrile concentration of 400 mm, thus guaranteeing excess

Indeed, no spontaneous release of AMC was observed in the

absence of TGase 2 (Discussion S6). Surprisingly, though, initial

investigations of 5c with regard to its enzymatic hydrolysis

and aminolysis also revealed no TGase 2-mediated release of

AMC (Discussion S6), and the same was also observed for the

g-Abu-derived coumarinyl amide 2c, as mentioned above. This

was not expected for the coumarinyl amides, because TGase 2

is actually specialised in the interconversion of amide bonds.

Moreover, aryl amides should be easier to cleave than aliphatic

amides from a purely electronic point of view, due to the less

pronounced double bond character of an amide bond adja-

cent to an aromatic ring system. This situation can also result

in faster reactions catalysed by enzymes involving acyl-enzyme

intermediates such as the serine protease trypsin, for which ni-

troanilide substrates exhibited greater performance constants

than the corresponding primary amides.^[34]However, results

comparable to those reported here were also obtained from

studies on chromogenic and fluorogenic acyl donor substrates

for factor XIIIa. Rapid cleavage of the aryl amide bond by factor

XIIIa to release p-nitroaniline was observed, whereas the con-

version of the corresponding AMC derivatives was considerably

slower. In contrast, the conversion of both aryl-amide-based

substrates by TGase 2 was negligible.^[35]

over the corresponding acyl donor. However, the calculated

app

ratios of k to K^a_m^pprepresent the true performance constants

cat

for aminolysis, because V_max/K_mis independent of the concen-

tration of either substrate, if the enzyme-catalysed reaction

adheres to a ping-pong mechanism,^[38]which is the case for

TGase 2.^[39,32b]The dependence of the kinetic parameters K^a_m^pp

app

cat

and k on substrate structure followed the trends observed

for hydrolysis; however, the values are considerably greater

(Table 3 and Figure 5). This result is reasonable because the de-

acylation step with aminoacetonitrile as acyl acceptor substrate

proceeds considerably more rapidly than in water.^[20b]Because

the Michaelis constant is the ratio of the sum of the rate con-

stants that describe the steps of disintegration of the Michaelis

complex (k_ꢂ1+k_cat) divided by the second-order rate constant

for the association of free enzyme and substrate (k₁), the

values for K_mhave to increase as deacylation becomes faster.

This relationship results in apparent K_mvalues up to approxi-

mately 100 times higher than the Michaelis constants for hy-

drolysis. Consequently, the increased single parameters K_mand

1

In that system for classification of enzyme–modifier interactions, competitive

inhibition is referred to as specific inhibition whereas noncompetitive inhibi-

tion is referred to as catalytic inhibition.

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

9

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

Table 3. Kinetic parameters for the TGase 2-catalysed aminolysis of the acyl donors 3–6 at pH 8.0 and 308C with aminoacetonitrile (400 mm) as acyl accept-

or.

Cpd

Regression analysis

Numerical integration

K^a_m^pp[mm]

k

[s^ꢂ1

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

K^a_m^pp[mm]

k

[s^ꢂ1

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

app

cat

gpTGase 2

148 (52)

140 (24)

120 (12)

121 (8)

236

3

4

5a

5b

6a*

6b*

6.26 (1.26)

3.05 (0.39)

47.1 (3.6)

52.7 (1.7)

29.7

42300

21800

393000

436000

126000

296000

99.5 (13.9)

135 (19)

82.9 (7.6)

82.1 (3.6)

249 (104)

150 (21)

4.28 (0.17)

2.66 (0.29)

34.3 (1.1)

35.7 (0.5)

29.8 (9.5)

45.6 (4.2)

43000

19600

414000

435000

120000

304000

238

70.4

hTGase 2

252

5b*

22.7

90100

171 (51)

16.7 (4.6)

97800

For details on calculation of the kinetic parameters see the Experimental Section. Data shown are mean values (ꢀSEMs) of three separate experiments,

each performed in duplicate. Active concentrations of TGase 2 from guinea pig liver (E_T=3.16 nm for 5–6 and 21.0 nm for 3 and 4) and from human

TGase 2 (E_T=3.08 nm) were calculated from Zedira’s activity data or determined by active site titration as recently described,^[30]respectively. * Data ob-

tained by the method of Cornish-Bowden and Eisenthal.^[40]

ures S14 and S15).^[41,40]The results of data analysis by numeri-

cal integration are in good agreement with those obtained by

regression analysis (Table 3).

Exceptions can be observed for the K^a_m^ppvalues of 6b in the

presence of gpTGase 2 and of 5b in that of the human

enzyme obtained from the different methods of data analysis.

These discrepancies might result from the mentioned problem

that in those cases substrate concentrations were restricted to

values <K^a_m^pp. In this context, it might seem that the improved

Figure 5. gpTGase 2- and hTGase 2-catalysed aminolysis of acyl donor 5b at

pH 8.0. Plots of v_0corrr=f([5b]) with nonlinear regressions (c) with use of

the g-Abu derivatives is of minor importance because sub-

the Michaelis–Menten equation [Eq. (4), Experimental Section]. Data shown

solubility of the glutamate-derived acyl donors in relation to

strate concentrations <K^a_m^ppcould also be investigated for the

are mean values ꢀSEMs of three separate experiments, each performed in

duplicate. When not apparent, error bars are smaller than the symbols. Con-

ditions: pH 8.0, 308C, 5% DMSO, 500 mm TCEP, 400 mm aminoacetonitrile,

0.3 mgmL^ꢂ1of either gpTGase 2 or hTGase 2.

latter compounds. However, the observed increase in K_mupon

changing the acyl acceptor from water to aminoacetonitrile,

which was also demonstrated for compound 2a by Gillet

et al.,^[12]implicates that substrate concentrations for 2a and

k_catcompensate each other when the performance constants

are calculated, with the latter ones being of similar value to

those of the hydrolysis reaction. As in the case of hydrolysis,

5b was revealed to be the most efficient substrate, with a

performance constant of 436000m^ꢂ1s^ꢂ1towards TGase 2 from

guinea pig, whereas its reaction with aminoacetonitrile cata-

lysed by the human enzyme was much slower (k_cat/K_m=

2b would be restricted to even lower values in relation to K_m^app

.

Although we refrained from attempts to characterise 2a and

2b with regard to their enzymatic aminolyses, it can be con-

cluded that the increased solubility of the glutamate-derived

acyl donors might indeed be beneficial, because a broader

concentration range is available for these substrates.

To confirm the occurrence of TGase 2-catalysed transamida-

tion in the presence of competing spontaneous reaction path-

ways, the assay mixture containing 5b and aminoacetonitrile

was analysed by RP-HPLC and ESI-MS, which confirmed the for-

mation of the expected corresponding cyanomethyl amide

(Figure S16).

90100m^{ꢂ1 ꢂ1}).

s

Representatively for compound 5b, additional investigations

that allowed for estimation of its true kinetic parameters k_cat

and K_mwere performed (Discussion S7). Because of the high

values of K^a_m^pp, measurements at higher acyl donor concentra-

tions were necessary, and this in turn resulted in higher rates

of spontaneous reactions. For regression analysis, the initial

substrate concentrations were therefore corrected to allow for

the time delay between preparation of the reaction mixture

and the start of reaction by addition of the enzyme. Further-

more, fitting the Michaelis–Menten equation to the experimen-

tal data yielded unreliable results in the cases of compounds

6a and 6b, as well as for 5b, in the presence of hTGase 2, due

to the high K^a_m^ppvalues and limitations in substrate concentra-

tion. For this reason, evaluation of these data sets was per-

formed by the method of Cornish-Bowden and Eisenthal (Fig-

Application of acyl donor 5b for the kinetic characterisation

of amine substrates and inhibitors

Characterisation of amine substrates: One motivation behind

this study was to establish an activity assay for TGase 2 that

allows kinetic evaluation both of amine-based acyl acceptor

substrates and of irreversible inhibitors of this enzyme. After

identifying fluorogenic acyl donor substrates with suitable

properties, we therefore aimed to establish their applicability

for that purpose. Because 5b exhibited the best substrate

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

10

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

properties of the studied analogues, the following investiga-

tions were performed with this acyl donor substrate. N-(Bioti-

nyl)cadaverine is often applied as amine substrate to deter-

mine the activity of TGase 2,^[42]so this compound was selected

in addition to aminoacetonitrile in order to verify the applica-

bility of the new fluorogenic acyl donors for the kinetic charac-

terisation of primary amines as acyl acceptors (for synthesis

see Discussion S8).

Table 4. Kinetic parameters of different amines as acyl acceptors for

gpTGase 2 and hTGase

2

at pH 8.0 and 308C with compound 5b

(100 mm) as acyl donor.

app

cat

Amine

K^a_m^pp[mm]

k

[s^ꢂ1

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

gpTGase 2

aminoacetonitrile

0.037 (0.002) 27.0 (1.8)

hTGase 2

730000

0.015 (0.001)

7.39 (0.15) 493000

The two primary amines were evaluated with regard to their

kinetic parameters K_mand k_cat, which were determined at a con-

stant concentration of 5b (100 mm). This concentration of the

acyl donor substrate is not sufficient to fulfil the condition

K_m(5b)![5b], neither for the guinea pig nor for the human

N-(biotinyl)cadaverine (7)* 1.67

9.02

5400

For details of the calculation of the kinetic parameters see the Experimen-

tal Section. Data shown are mean values (ꢀSEMs) either of one (com-

pound 7) or of three (aminoacetonitrile) separate experiments, each per-

formed in duplicate. Active concentrations of TGase 2 from guinea pig

liver (E_T=3.16 nm) and human TGase 2 (E_T=6.16 nm) were calculated

from Zedira’s activity data or determined by active site titration as recent-

ly described,^[30]respectively. *K_i=8.04 mm.

TGase 2, so the parameters K_mand k_catare of apparent charac-

app

ter and therefore denoted as K^a_m^ppand k . The Michaelis–

cat

Menten plots are shown in Figure 6, and the calculated param-

eters are summarised in Table 4.

app

cat

For aminoacetonitrile, values for K^a_m^ppand k of 37 mm and

27 s^ꢂ1, respectively, were determined with the guinea pig

enzyme. With human TGase 2, the K^a_m^ppof aminoacetonitrile

Because N-(biotinyl)cadaverine (7) has frequently been used

as molecular tool to study TGase 2 at the cellular level, the util-

ity of 5b for the evaluation of amine-based acceptor substrates

was demonstrated for this compound with human TGase 2. In

app

cat

was slightly lower, whereas its k value decreased to a larger

extent, thus resulting in a somewhat reduced performance

constant. To gain insights into the true Michaelis constant of

aminoacetonitrile, K^a_m^ppvalues were determined at four differ-

ent concentrations of the fluorogenic acyl donor 5b, which

concomitantly provided insight into the true K_mvalue of 5b.

The detailed analysis of these kinetic data along with the asso-

ciated discussion is given in Discussion S7.

contrast to aminoacetonitrile, compound

7 displayed a

K^a_m^ppvalue in the millimolar range. This dramatic difference in

the K^a_m^ppvalues reflects the differing basicities of aminoacetoni-

trile and 7. Whereas the amino group of aminoacetonitrile ex-

hibits a pK_avalue of 5.6 for the conjugate ammonium ion, the

corresponding value for 7 is in the range of 10–11.^[20b]Because

the deacylation of acyl-donor-derived TGase 2 thioester inter-

mediates can only be effected by unprotonated amines,^[43]the

free substrate concentration of 7 is much lower than that of

aminoacetonitrile. Furthermore, Brønsted plots for a series of

primary amines as acyl acceptor substrates of varying basicity

indicated a correlation between the performance constants of

TGase 2-catalysed transamidation and their pK_avalues.^[20b]It

app

cat

should be noted that the k value of 7 with human TGase 2

is in the same range as that of aminoacetonitrile and that its

rate versus concentration plot indicates substrate inhibition

(Figure 6B). The latter finding is in accordance with the behav-

iour of the analogous fluorescent acyl acceptor N-(dansyl)cada-

verine, which displayed substrate inhibition for TGase 2-cata-

lysed transamidation in the cases both of the guinea pig

enzyme and of the human one. This phenomenon has been

attributed to the formation of an unproductive complex of the

acyl acceptor substrate and free TGase 2.^[32a]Moreover, sub-

strate inhibition in enzyme-catalysed group-transferring reac-

tions by the substrate that is devoid of the transferred group

(in this case the amine substrate) is very common, due to its

premature binding to the free enzyme.^[44]

Figure 6. gpTGase 2- and hTGase 2-catalysed incorporation of different

amines into compound 5b. Plots of v_0corr=f([amine]) with nonlinear regres-

sions (c) with use of A) Equation (4) (Michaelis–Menten equation, Experi-

mental Section) for aminoacetonitrile, and B) Equation (5) (substrate inhibi-

tion, Experimental Section) for N-(biotinyl)cadaverine (7). Data shown for

aminoacetonitrile are mean values ꢀSEMs of three separate experiments,

each performed in duplicate. Data shown for N-(biotinyl)cadaverine (7) are

mean value ꢀSD of one experiment, performed in duplicate. When not ap-

parent, error bars are smaller than the symbols. Conditions: pH 8.0, 308C,

5% DMSO, 100 mm acyl donor 5b, 500 mm TCEP, 0.3 mgmL^ꢂ1gpTGase 2 or

0.6 mgmL^ꢂ1hTGase 2.

Characterisation of irreversible inhibitors: Evaluation of

enzyme inhibitors for the purposes of drug discovery projects

is often based on kinetic assays. This is particularly valid for in-

hibitors that interact with enzymes in an irreversible manner.

Therefore, we aimed to validate the usefulness of the devel-

oped fluorogenic TGase 2 substrates for the characterisation of

inhibitors representatively for substrate 5b with the estab-

lished irreversible TGase 2 inhibitors iodoacetamide and the

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

11

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

Scheme 6. Structures of iodoacetamide and acrylamide 8.

Figure 7. gpTGase 2- and hTGase 2-catalysed hydrolysis of acyl donor 5b at

pH 6.5. Plots of v_0total=f([5b]) with the nonlinear regressions (c) by use of

Equation (4) (Michaelis–Menten equation, Experimental Section). Data shown

are meansꢀSEMs of three separate experiments, each performed in dupli-

cate. When not apparent, error bars are smaller than the symbols. Condi-

tions: pH 6.5, 308C, 5% DMSO, 500 mm TCEP, 3 mgmL^ꢂ1of either gpTGase 2

or hTGase 2.

recently reported selective N^e-acryloyllysine derivative

8

(Scheme 6).^[14]

Inhibition studies with iodoacetamide for the TGase 2-cata-

lysed hydrolysis of 5b at pH 8.0 were problematic because the

product-release progress plots were curved even in the ab-

sence of inhibitor, due to substrate depletion (data not

shown). To obtain steady-state conditions over a broad time

window, neither reducing the amount of enzyme nor increas-

ing the substrate concentration was possible, due to the sus-

ceptibility of 5b to spontaneous hydrolysis. Therefore, the con-

version of 5b was decelerated by reducing the pH from 8.0 to

6.5; this slowed down the rate of the spontaneous reaction to

less than 5%, whereas the enzymatic hydrolysis was reduced

to a much lesser extent. These modified conditions resulted in

linear progress curves for concentrations of 5b greater than

20 mm both for gpTGase 2 and for hTGase 2 over a time

window of 900 s. This linearity ensures the absence of sub-

strate depletion, which is of crucial importance for the correct

kinetic evaluation of time-dependent enzyme inhibition.^[45]The

fluorescence coefficients of 4-methylumbelliferone and the

Michaelis–Menten parameters were determined for pH 6.5 (Fig-

ures 7 and S17, Table 5).

kylation of the active-site thiol Cys277.^[47]Irreversible enzyme

inactivation usually results in steady-state rates that are equal

to zero independent of inhibitor concentration. However, be-

cause substrate 5b continues to react in the presence of com-

pletely inactivated enzyme, we use v_sas an open parameter in

Equation (6) to take account of the remaining spontaneous re-

action after completed enzyme inhibition. Replotting the ob-

tained pseudo-first-order rate constants k_obsagainst the inhibi-

tor concentration and linear regression of the resulting line

provided the ratio k_obs/[I] as slope. Correction of that value by

the factor 1+[S]/K_mgave a value of 29800m^ꢂ1s^ꢂ1for the

second-order inactivation constant k_inact/K_Itowards the guinea

pig enzyme, with a similar value being obtained for hTGase 2

(Table 6).^[46,48]In contrast to the k_obsvalues, which linearly in-

creased with rising inhibitor concentration, the initial velocities

did not systematically vary when the iodoacetamide concen-

tration was increased (Figure S18). This finding indicates a one-

step mechanism for inactivation of both guinea pig and

human TGase 2 by iodoacetamide and suggests that no stable

noncovalent enzyme·inhibitor complex is formed prior to the

inactivating alkylation step. This result is in agreement with ob-

servations for the inactivation of factor XIIIa by iodoacetate^[35a]

and can be explained by considering that this small-molecule

inhibitor does not provide many contact points for noncova-

lent interactions.

With gpTGase 2, the performance constants were reduced

s

to 130000m^ꢂ1^ꢂ1: that is, to less than half of the value deter-

mined at pH 8.0 (Table 1). This is caused by a reduction in k_cat,

with K_mbeing unaffected by the change in pH. A similar trend

was observed for the human enzyme. Data analysis by numeri-

cal integration confirmed these results for the two TGase or-

thologues.

The reduction of pH in the assay medium enabled analysis

of inhibition by iodoacetamide under the formalism of slow-

binding inhibition [Eq. (6), Experimental Section].^[46]Iodoaceta-

mide has been shown to inactivate TGase 2 irreversibly by al-

To characterise a pharmaceutically more relevant inhibitor

with the established assay, compound 8 was investigated with

Table 5. Kinetic parameters for the TGase 2-catalysed hydrolyses of the acyl donor 5b at pH 6.5 and 308C.

Cpd

Regression analysis

k_cat[s^ꢂ1

Numerical integration

k_cat[s^ꢂ1

K_m[mm]

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

K_m[mm]

]

k_cat/K_m[m^ꢂ1s^ꢂ1

]

gpTGase 2

2.53 (0.18)

hTGase 2

6.60 (1.06)

5b

0.33 (0.01)

0.32 (0.02)

130000

48500

2.62 (0.20)

5.63 (1.00)

0.33 (0.01)

0.31 (0.02)

126000

55100

For details on calculation of the kinetic parameters see the Experimental Section. Data shown are mean values (ꢀSEMs) of three separate experiments,

each performed in duplicate. Active concentrations of TGase 2 from guinea pig liver (E_T=31.6 nm) and human (E_T=30.8 nm) were calculated from Zedira’s

activity data or determined by active site titration as recently described,^[30]respectively. The k_obsvalue for the spontaneous reaction of 5b at pH 6.5 and

308C in the presence of TCEP was determined to be (0.70ꢀ0.03)ꢁ10^ꢂ3min^ꢂ1

.

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

12

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

and the initial velocities (v_i) hyperbolically decreased with in-

creasing inhibitor concentration. This observation indicates the

formation of a rapid-equilibrium noncovalent complex be-

tween acrylamide 8 and hTGase 2 prior to the inactivation

step, for which a dissociation constant (K_i) of 5.73 mm was cal-

culated from the plot of v_iversus [I] (Figure S19). From the

double reciprocal plot of 1/k_obsversus 1/[I] (Figure 8B) it was

possible to obtain the isolated value for the first-order inactiva-

tion constant k_inact(0.35 s^ꢂ1) describing the transition of the ini-

tial noncovalent enzyme·inhibitor complex to the final covalent

complex. From this value, K_Iwas calculated to be 68.7 mm. The

meaning of this value is distinctly different from that of K_ibe-

cause it represents a merely kinetic parameter that signifies

the inhibitor concentration at which the inactivation proceeds

with half the maximum rate in the absence of substrate.^[46]

Usually, irreversible enzyme inhibition results in steady state

rates equal to zero, whereas in the present case these rates are

different from zero because substrate conversion continues

after completed enzyme inactivation, due to spontaneous hy-

drolysis. To take this factor into account for the kinetic analysis

of TGase 2 inhibition, the kinetic parameters were independ-

ently determined by numerical integration.^[33,23b]For this pur-

pose, differential equations were formulated on the basis illus-

trated in Scheme 7 and implemented into the statistical soft-

Table 6. Kinetic parameters for the irreversible TGase 2 inhibitors iodo-

acetamide and acrylamide 8 at pH 6.5 and 308C.

Inhibitor

Regression analysis

_inact/K_I[m^ꢂ1s^ꢂ1

Numerical integration

k_inact/K_I[m^ꢂ1s^ꢂ1

]

k

]

gpTGase 2

29800 (1280)

740 (78)

iodoacetamide

acrylamide 8

39600 (973)

821 (10)

hTGase 2

iodoacetamide

acrylamide 8*

50700 (3510)

5110 (68)

44400 (3030)

5290 (135)

*K_i=5.73 mm (0.77), K_I=68.7 mm (24.2), k_inact=0.35 s^ꢂ1(0.12); for details

on the calculation of the kinetic parameters see the Experimental Section.

Inhibition experiments on gpTGase 2 and hTGase 2 were performed in

the presence of 25 mm (ꢁ10ꢁK_m) and 30 mm (ꢁ4.5ꢁK_m), respectively, of

acyl donor 5b. Data shown are mean values (ꢀSEMs) of two separate

experiments, each performed in duplicate.

regard to its interaction with guinea pig and human TGase 2.

The N^e-acryloyllysine derivative 8 was prepared by the pub-

lished procedure with slight modifications;^[14]details on its syn-

thesis will be reported in the context of a follow-up study. The

results of the kinetic characterisation of 8 are included in

Table 6. With the guinea pig enzyme, compound 8 exhibited

a second-order inactivation constant of 740m^ꢂ1^ꢂ1. No system-

s

atic variation of the initial velocities with increasing inhibitor

concentration was discernible; this led to the conclusion that

inactivation of gpTGase 2 by 8 follows a one-step mechanism.

Notably, the situation is obviously different for the human

enzyme (Figure 8). Here, the value for k_inact/K_Iwas approximate-

ly seven times higher than that for the guinea pig orthologue,

Scheme 7. Kinetic model for the two-step irreversible inhibition of TGase 2

and the associated differential rate laws derived from it. For the case that

the ratio k_ꢂ2/k₂=K_iis high, the concentration of the initial noncovalent enzy-

me·inhibitor complex becomes negligible and, in consequence, the two-step

mechanism for irreversible inhibition simplifies to a one-step mechanism.

Abbreviations used in the scheme. E: enzyme. I: inhibitor. EI: initial enzyme·

inhibitor complex. E–I: covalent enzyme·inhibitor complex.

ware R. The series of progress curves in the absence and in the

presence of different inhibitor concentrations were globally

fitted, and this provided values for k_inact/K_Ithat are essentially

similar to those obtained by conventional regression analysis

(Table 6).

Thus, it can be concluded that the common analysis using

integrated rate equations and subsequent evaluation of the

obtained data yields reliable parameters to describe the inhibi-

tory potency. Taken together, the obtained results demonstrate

the suitability of the established assay for the kinetic character-

isation of TGase 2 inhibitors.

Figure 8. Kinetic characterisation of acrylamide 8 with hTGase 2 and acyl

donor 5b. A) Typical time courses of the hTGase 2-catalysed hydrolysis of

*

5b in the presence of different concentrations of acrylamide 8 [0 mm ( ),

~

!

^

3 mm ( ), 6 mm (+), 12 mm (ꢁ), 24 mm ( ) and 30 mm ( )] in the presence of

30 mm (ꢁ4.5ꢁK_m) of acyl donor 5b. B) Double reciprocal plot [1/k_obs=f(1/

[8])] with linear regression to the data for determination of k_inact, K_Iand k_inact

/

K_I. Conditions: pH 6.5, 308C, 5% DMSO, 500 mm TCEP, 3 mgmL^ꢂ1of hTGase 2.

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

13

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

might stimulate further developments directed towards fluoro-

genic substrates for this important enzyme.

Conclusions

The recently described acyl donor substrate 2a and com-

pounds derived from it that release fluorescent coumarins

upon enzymatic conversion appear to be attractive substrates

for the fluorimetric assay of TGase 2. However, both 2a and its

methyl-substituted analogue 2b proved to be prone to aggre-

gation in aqueous media; this can hamper detailed kinetic

investigations depending on the particular conditions and in-

tended application purposes. Therefore, analogues based on

small glutamate-containing peptides were synthesised by a

solid-phase approach and were demonstrated to be soluble in

the assay medium up to concentrations of 250 mm. This re-

markably improved water solubility allowed for the extensive

kinetic characterisation of the newly designed substrates in

terms of their TGase 2-catalysed hydrolysis and aminolysis,

both by conventional regression analysis and by numerical in-

tegration. In addition, their spontaneous reactivity under assay

conditions was studied in detail, and this has revealed a signifi-

cant contribution of spontaneous thiolysis in the presence of

DTT to the overall rate of spontaneous disintegration, amongst

other findings. This result suggests that TCEP should be used

as antioxidant instead of DTT.

Experimental Section

General: All commercial reagents and solvents were used without

further purification unless otherwise specified. Melting points were

determined with a Galen III Boetius apparatus from Cambridge

Instruments. Nuclear magnetic resonance spectra were recorded

with Varian Unity 400 MHz or Agilent Technologies 400 NMR spec-

trometers. Spectra were processed by using the programme Mes-

treNova (version 6.1.1-6384).^[49]NMR chemical shifts were refer-

enced to the residual solvent resonances relative to tetramethylsi-

lane (TMS). Mass spectra (ESI) were obtained with a Micromass

Quattro LC or a Waters Xevo TQ-S mass spectrometer, each driven

by Mass Lynx software. Elemental analysis was performed with

a LECO CHNS-932 apparatus. Determination of the resin loading

was performed with a Thermo Scientific Helios a UV/Vis spectro-

photometer.

Chromatography: Thin-layer chromatography (TLC) was performed

with Merck silica gel F-254 aluminium plates and visualisation

under UV (254 nm) and/or by staining with a ninhydrin solution

(0.1%, m/v) in ethanol. Preparative column chromatography was

carried out with Merck silica gel (mesh size 230–400 ASTM) and

solvent mixtures as specified for the particular compounds. Analyti-

cal and preparative HPLC of compounds 2–7 was performed with

a Varian Prepstar system equipped with a UV detector (Prostar,

Varian) and an automatic fraction collector (Foxy 200). Two Micro-

sorb C18 60-8 columns (Varian Dynamax 250ꢁ4.6 mm and 250ꢁ

21.4 mm) were used as the stationary phases for analytical and

preparative HPLC, respectively. A binary gradient system of 0.1%

CF₃COOH/water (solvent A) and 0.1% CF₃COOH/CH₃CN (solvent B)

at a flow rate of 1 mLmin^ꢂ1or 10 mLmin^ꢂ1served as the eluent.

For analytical HPLC, the programme for elution of compounds 2–6

was as follows: 0–3 min 90% A, 3–25 min gradient to 90% B, 25–

35 min 90% B, 35–36 min gradient back to 90% A, 36–40 min 90%

A. For compound 7 the following elution regime was used: 0–

5 min 95% A, 5–25 min gradient to 95% B, 25–30 min 95% B, 30–

31 min gradient back to 95% A, 31–35 min 95% A. For preparative

HPLC, the conditions for the gradient elution of compounds 2–6

were as follows: 0–3 min 90% A, 3–25 min gradient to 90% B, 25–

30 min 90% B, 30–31 min gradient back to 90% A, 31–35 min 90%

A. For compound 7 the following elution regime was used: 0–

7 min 90% A, 7–22 min gradient to 80% B, 22–30 min 80% B, 30–

31 min gradient back to 90% A, 31–35 min 90% A. HPLC for inves-

tigating the spontaneous reactivity of compound 5b was carried

out with a system consisting of a Merck Hitachi L7100 gradient

pump combined with a Jasco DG2080 four-line degasser with UV

detection with a Merck Hitachi L7450 diode array detector. The

system was operated with D-700 HSM software and use of a Merck

Hitachi D7000 interface. A Luna C18 5 mm column (Phenomenex,

250ꢁ4.6 mm) served as stationary phase. The following elution

programme (binary gradient system as detailed above, flow rate

1 mLmin^ꢂ1) was run to separate the components: 0–3 min 80% A,

3–25 min gradient to 70% B, 25–26 min gradient to 95% B, 26–

30 min 95% B, 30–35 min gradient back to 80% A, 35–40 min 80%

A.

The influence of the fluorogenic leaving group on the kinet-

ic properties was investigated. Substrates 5b and 6b, each

containing 7-hydroxy-4-methylcoumarin attached to the g-car-

boxylic acid group of the Glu residue, showed superior kinetic

properties to their unsubstituted counterparts 5a and 6a;

analogous amides 2c and 5c, containing 7-amino-4-methyl-

coumarin, are not accepted as substrates. Z-Glu(HMC)-Gly-OH

(5b) exhibited the best substrate properties out of the syn-

thesised compounds, with performance constants of 300000

and 436000m^ꢂ1s^ꢂ1determined for the gpTGase 2-catalysed

hydrolysis and aminolysis, respectively, at pH 8.0. Therefore,

5b was selected for representative kinetic characterisation

of amine-based acyl-acceptor substrates and irreversible in-

hibitors. The determination of performance constants at

pH 8.0 was demonstrated for aminoacetonitrile and N-

(biotinyl)cadaverine as representative primary amine substrates

on both guinea pig and human TGase 2. Inhibitor characterisa-

tion was based on monitoring and analysing the TGase 2-cata-

lysed hydrolysis of 5b at pH 6.5, which resulted in reliable

second-order inactivation constants for iodoacetamide and

compound 8 towards both the guinea pig and the human

enzyme.

Furthermore, this study confirmed compounds 2a and 2b

as potentially suitable substrates for fluorimetric activity deter-

mination of TGase 2. In view of their reduced susceptibility to

spontaneous reaction relative to that of the glutamate-derived

analogues introduced here, 2a and 2b might be more advan-

tageous for assaying cellular TGase 2 activities.^[12]

Taken together, the results of this study show compound 5b

to be a powerful fluorogenic substrate of guinea pig and

human TGase 2. Therefore, this compound allows for robust

assay methods to identify and to characterise molecules capa-

ble of targeting TGase 2 for therapeutic inhibition and molecu-

lar imaging, although its susceptibility to spontaneous decay

Analytical data for the glutamate-derived fluorogenic acyl

donors

Z-l-Glu(HMC)-OH (3): The synthesis with use of HMC yielded 3

(67 mg, 30%) as a yellow oil. Because 3 tends to decompose in the

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

14

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

isolated state, this compound should be stored below 08C. ¹H NMR

153.49, 152.95, 152.83, 136.90, 128.31, 127.78, 127.71, 126.37,

118.39, 117.50, 113.72, 110.09, 65.54 (CH₂O of Z), 53.50, 40.68,

30.02, 27.00, 18.17 ppm (CH₃); MS (ESI⁺): m/z calcd for C₂₅H₂₅N₂O₉:

497.16 [M+H]⁺; found: 497.1; elemental analysis calcd (%) for

C₂₅H₂₄N₂O₉: C 60.48, H 4.87, N 5.64; found: C 59.86, H 4.84, N 5.65.

([D₆]DMSO): d=7.81 (d, ³J_H,H=8.7 Hz, 1H; H-5 of coumarin), 7.70 (d,

4

³J_H,H=8.3 Hz, 1H; NH), 7.39–7.27 (m, 5H; H_phenyl), 7.25 (d, J_H,H

=

3

4

2.1 Hz, 1H; H-8 of coumarin), 7.17 (dd, J_H,H=8.6, J_H,H=2.2 Hz, 1H;

4

H-6 of coumarin), 6.39 (d, J_H,H=1.1 Hz, 1H; H-3 of coumarin), 5.10–

5.01 (m, 2H; CH₂O of Z), 4.15–4.07 (m, 1H; C_aH of Glu), 2.78–2.65

(m, 2H; C_gH₂of Glu), 2.44 (d, J_H,H=1.0 Hz, 3H; CH₃), 2.20–2.09 (m,

4

Z-l-Glu(AMC)-Gly-OH (5c): The synthesis with use of AMC yielded

1

1H; C_bHH of Glu), 2.01–1.89 ppm (m, 1H; C_bHH of Glu); ¹³C NMR

([D₆]DMSO): d=173.26, 170.75, 159.60, 156.19, 153.49, 152.95,

152.80, 136.92, 128.32, 127.81, 127.72, 126.40, 118.38, 117.53,

113.73, 110.07, 65.51 (CH₂O of Z), 52.84 (C_aof Glu), 30.24, 25.91,

18.17 ppm (CH₃); MS (ESI⁺): m/z calcd for C₂₃H₂₂NNaO₈: 462.12

[M+Na]⁺; found: 462.1; elemental analysis calcd (%) for C₂₃H₂₁NO₈:

C 62.87, H 4.82, N 3.19; found: C 60.45, H 5.14, N 3.14.

5c (7.3 mg, 3%) as a white solid. H NMR ([D₆]DMSO): d=10.34 (s,

3

1H; NH of coumarin), 8.22 (t, J=5.6 Hz, 1H; NH of Gly), 7.75 (d,

⁴J=1.9 Hz, 1H; H-8 of coumarin), 7.70 (d, ³J=8.7 Hz, 1H; H-5 of

coumarin), 7.49–7.44 (m, 2H; H-6 of coumarin, NH of Glu), 7.40–

7.25 (m, 5H; H_phenyl), 6.25 (d, ⁴J=1.0 Hz, 1H; H-3 of coumarin),

5.08–4.95 (m, 2H; CH₂O of Z), 4.14–3.99 (m, 1H; C_aH of Glu), 3.82

2

3

2

3

(dd, J=17.4, J=6.1 Hz, 1H; C_aHH of Gly), 3.73 (dd, J=17.8, J=

5.5 Hz, 1H; C_aHH of Gly), 2.54–2.43 (m, 2H; C_gH₂of Glu), 2.39 (d,

⁴J=0.7 Hz, 3H; CH₃), 2.11–1.95 (m, 1H; C_bHH of Glu), 1.92–

1.75 ppm (m, 1H; C_bHH of Glu); ¹³C NMR ([D₆]DMSO): d=171.84,

171.31, 171.08, 160.01, 159.74, 153.65, 153.09, 142.53, 128.30,

127.76, 127.69, 125.85, 115.04, 114.78, 112.08, 105.42, 65.49 (CH₂O

of Z), 27.36, 17.95 ppm (CH₃), 2ꢁC_aand C_gnot visible; MS (ESI⁺):

m/z calcd for C₂₅H₂₆N₃O₈: 496.17 [M+H]⁺; found: 496.2.

Z-l-Phe-l-Glu(HMC)-OH (4): The synthesis with use of HMC yielded

4 (99 mg, 36%) as a white solid. Because 4 tends to decompose in

the isolated state, this compound should be stored below 08C.

3

¹H NMR ([D₆]DMSO): d=8.41 (d, J=8.0 Hz, 1H; NH of Glu), 7.81 (d,

³J=8.6 Hz, 1H; H-5 of coumarin), 7.53 (d, ³J=8.6 Hz, 1H; NH of

Phe), 7.36–7.15 (m, 12H; H_phenyl, H-6,8 of coumarin), 6.39 (s, 1H; H-3

of coumarin), 4.95–4.87 (m, 2H; CH₂O of Z), 4.45–4.36 (m, 1H;

C_aH), 4.36–4.28 (m, 1H; C_a), 3.02 (dd, ²J=13.8, ³J=3.8 Hz, 1H;

C_bHH of Phe), 2.81–2.69 (m, 3H; C_bHH of Phe, C_gH₂of Glu), 2.44 (s,

3H; CH₃), 2.26–2.14 (m, 1H; C_bHH of Glu), 2.00–1.88 ppm (m, 1H;

C_bHH of Glu); ¹³C NMR ([D₆]DMSO): d=172.94, 171.91, 170.78,

159.60, 155.90, 153.50, 152.95, 152.87, 138.08, 136.89, 129.20,

128.21, 128.04, 127.63, 127.42, 126.38, 126.25, 118.41, 117.52,

113.72, 110.11, 65.21 (CH₂O of Z), 56.10, 50.69, 37.26, 29.86, 26.10,

18.17 ppm (CH₃); MS (ESI⁺): m/z calcd for C₃₂H₃₁N₂O₉: 587.20

[M+H]⁺; found: 587.3; elemental analysis calcd (%) for C₃₂H₃₀N₂O₉:

C 65.52, H 5.15, N 4.78; found: C 63.94, H 5.23, N 4.81.

Z-l-Phe-l-Glu(HC)-Gly-OH (6a): The synthesis with use of HC yielded

6a (44 mg, 35%) as a white solid. Because 6a tends to decompose

in the isolated state, this compound should be stored below 08C.

¹H NMR ([D₆]DMSO): d=12.59 (s, 1H; COOH), 8.29–8.22 (m, 2H; NH

3

of Glu and Gly), 8.07 (d, J=9.5 Hz, 1H; H-4 of coumarin), 7.77 (d,

³J=8.5 Hz, 1H; H-5 of coumarin), 7.55 (d, ³J=8.5 Hz, 1H; NH of

Phe), 7.35–7.10 (m, 12H; H_phenyl, H-6,8 of coumarin), 6.48 (d, ³J=

9.6 Hz, 1H; H-3 of coumarin), 4.96–4.87 (m, 2H; CH₂O of Z), 4.52–

2

3

4.43 (m, 1H; C_aH), 4.38–4.27 (m, 1H; C_aH), 3.84 (dd, J=17.5, J=

2

3

6.0 Hz, 1H; C_aHH of Gly), 3.74 (dd, J=17.5, J=5.8 Hz, 1H; C_aHH

2

3

of Gly), 3.04 (dd, J=13.7, J=3.8 Hz, 1H; C_bHH of Phe), 2.81–2.68

(m, 3H; C_bHH of Phe, C_gH₂of Glu), 2.16–2.05 (m, 1H; C_bHH of Glu),

1.98–1.86 ppm (m, 1H; C_bHH of Glu); ¹³C NMR ([D₆]DMSO): d=

171.67, 171.12, 171.01, 170.88, 159.68, 155.88, 154.05, 152.89,

143.82, 138.06, 136.88, 129.28, 129.19, 128.20, 128.01, 127.62,

127.42, 126.20, 118.64, 116.64, 115.50, 110.11, 65.23 (CH₂O of Z),

56.14, 51.22, 40.67, 37.25, 29.64, 27.16 ppm; MS (ESI⁺): m/z calcd

for C₃₃H₃₂N₃O₁₀: 630.21 [M+H]⁺; found: 630.1; elemental analysis

calcd (%) for C₃₃H₃₁N₃O₁₀: C 62.95, H 4.96, N 6.67; found: C 61.75, H

5.15, N 7.22.

Z-l-Glu(HC)-Gly-OH (5a): The synthesis with use of HC yielded 5a

(34 mg, 35%) as a white solid. Because 5a tends to decompose in

the isolated state, this compound should be stored below 08C.

3

¹H NMR ([D₆]DMSO): d=8.32 (t, J=5.9 Hz, 1H; NH of Gly), 8.07 (d,

³J=9.3 Hz, 1H; H-4 of coumarin), 7.76 (d, ³J=8.5 Hz, 1H; H-5 of

coumarin), 7.57 (d, ³J=8.3 Hz, 1H; NH of Glu), 7.39–7.25 (m, 6H;

H

_phenyl, H-8 of coumarin), 7.15 (dd, ³J=8.4, ⁴J=2.2 Hz, 1H; H-6 of

3

coumarin), 6.48 (d, J=9.6 Hz, 1H; H-3 of coumarin), 5.10–4.98 (m,

2H; CH₂O of Z), 4.22–4.13 (m, 1H; C_aof Glu), 3.83 (dd, ²J=17.5,

³J=5.9 Hz, 1H; C_aHH of Gly), 3.73 (dd, ²J=17.5, ³J=5.6 Hz, 1H;

C_aHH of Gly), 2.78–2.64 (m, 2H; C_gH₂of Glu), 2.12–2.01 (m, 1H;

C_bHH of Glu), 1.99–1.86 ppm (m, 1H; C_bHH of Glu); ¹³C NMR

([D₆]DMSO): d=171.54, 171.07, 170.85, 159.68, 155.94, 154.05,

152.82, 143.81, 136.89, 129.28, 128.30, 127.77, 127.69, 118.61,

116.64, 115.52, 110.08, 65.53 (CH₂O of Z) 53.49, 40.67, 30.01,

27.00 ppm; MS (ESI⁺): m/z calcd for C₂₄H₂₃N₂O₉: 483.14 [M+H]⁺;

found: 483.1; elemental analysis calcd (%) for C₂₄H₂₂N₂O₉: C 59.75,

H 4.60, N 5.81; found: C 58.28, H 4.58, N 5.75.

Z-l-Phe-l-Glu(HMC)-Gly-OH (6b): The synthesis with use of HMC

yielded 6b (88 mg, 27%) as a white solid. Because 6b tends to de-

compose in the isolated state, this compound should be stored

below 08C. ¹H NMR ([D₆]DMSO): d=8.29–8.24 (m, 2H; NH of Glu

and Gly), 7.81 (d, ³J=8.7 Hz, 1H; H-5 of coumarin), 7.55 (d, ³J=

8.4 Hz, 1H; NH of Phe), 7.34–7.16 (m, 12H; H_phenyl, H-6,8 of coumar-

in), 6.39 (d, ⁴J=1.2 Hz, 1H; H-3 of coumarin), 4.95–4.88 (s, 2H;

CH₂O of Z), 4.52–4.44 (m, 1H; C_aH), 4.37–4.29 (m, 1H; C_aH), 3.84

2

3

2

3

(dd, J=17.6, J=5.9 Hz, 1H; C_aHH of Gly), 3.74 (dd, J=17.5, J=

Z-l-Glu(HMC)-Gly-OH (5b): The synthesis with use of HMC yielded

5b (63 mg, 26%) as a white solid. Because 5b tends to decompose

in the isolated state, this compound should be stored below 08C.

¹H NMR ([D₆]DMSO): d=12.58 (s, 1H; COOH), 8.32 (t, ³J=5.8 Hz,

2

3

5.7 Hz, 1H; C_aHH of Gly), 3.04 (dd, J=13.8, J=3.7 Hz, 1H; C_bHH

of Phe), 2.80–2.68 (m, 3H; C_bHH of Phe and C_gH₂of Glu), 2.44 (d,

⁴J=0.9 Hz, 3H; CH₃), 2.17–2.05 (m, 1H; C_bHH of Glu), 1.98–

1.86 ppm (m, 1H; C_bHH of Glu); ¹³C NMR ([D₆]DMSO): d=171.70,

171.14, 171.04, 170.89, 159.61, 155.89, 153.50, 152.96, 152.90,

138.08, 136.89, 129.20, 128.21, 128.03, 127.63, 127.42, 126.37,

126.22, 118.41, 117.50, 113.71, 110.12, 65.23 (CH₂O of Z), 56.15,

51.23, 40.68, 37.25, 29.64, 27.17, 18.17 ppm (CH₃); MS (ESI⁺): m/z

calcd for C₃₄H₃₄N₃O₁₀: 644.22 [M+H]⁺; found: 644.3; elemental

analysis calcd (%) for C₃₄H₃₃N₃O₁₀: C 63.45, H 5.17, N 6.53; found: C

62.10, H 5.17, N 6.46.

3

1H; NH of Gly), 7.81 (d, J=8.7 Hz, 1H; H-5 of coumarin), 7.57 (d,

³J=8.3 Hz, 1H; NH of Glu), 7.39–7.25 (m, 6H; H_phenyl, H-8 of cou-

3

4

marin), 7.18 (dd, J=8.6, J=2.2 Hz, 1H; H-6 of coumarin), 6.39 (d,

⁴J=1.2 Hz, 1H; H-3 of coumarin), 5.10–4.99 (m, 2H; CH₂O of Z),

3

4.22–4.14 (m, 1H; C_aH of Glu), 3.83 (dd, ²J=17.5, J=5.9 Hz, 1H;

2

3

C_aHH of Gly), 3.73 (dd, J=17.5, J=5.7 Hz, 1H; C_aHH of Gly), 2.77–

4

2.65 (m, 2H; C_gH₂of Glu), 2.44 (d, J=1.1 Hz, 3H; CH₃), 2.13–2.01

(m, 1H; C_bHH of Glu), 1.98–1.86 ppm (m, 1H; C_bHH of Glu);

¹³C NMR ([D₆]DMSO): d=171.56, 171.09, 170.86, 159.60, 155.95,

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

15

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

Fluorimetric assay

RFU₀: RFU_t=0. Plateau: RFU_t!1. k: rate constant to reach the pla-

teau.

Buffer systems and TGase 2 stock solutions: For the measurements

at pH 8.0 and pH 6.5, the following two assay buffers of the corre-

sponding pH value were used.

v_{0 total}¼ k ꢃ plateau

ð2Þ

buffer A: 3-(N-morpholino)propanesulfonic acid (MOPS, pH 8.0,

100 mm), CaCl₂(3 mm), EDTA (50 mm), adjusted to pH 8.0 with

NaOH (1m)

With regard to the measurements of compound 5b at pH 6.5, time

courses of type (RFUꢂRFU₀)=f(t) were analysed either by nonlinear

[Eq. (1)] or by linear regression over the first 300 s to the experi-

mental data, depending on the shape of the curve.

buffer B: 2-(N-morpholino)ethanesulfonic acid (MES, pH 6.5,

100 mm), CaCl₂(3 mm), EDTA (50 mm), adjusted to pH 6.5 with

NaOH (1m)

The rates for the spontaneous reactions (v_0control) in the absence of

enzyme were determined as described below. All fluorescence

rates (RFUs^ꢂ1) were converted into molar rates [mmmin^ꢂ1] by divid-

ing by the corresponding fluorescence coefficients. Subsequently,

Two different enzyme buffers were used for the preparation of the

TGase 2 stock solutions:

the two sets of initial rates {v_0total=f([acyl donor]) and v_0control

=

buffer I: MOPS (pH 8.0, 100 mm), CaCl₂(3 mm), DTT (10 mm), glyc-

erol (20%, v/v)

f([acyl donor])} were globally analysed by use of the model of total

and nonspecific binding as implemented in GraphPad Prism to

determine the kinetic parameters for the enzymatic conversion.

Accordingly, the following rule was defined [Eq. (3)]:

buffer II: MOPS (pH 8.0, 100 mm), CaCl₂(3 mm), TCEP (10 mm), glyc-

erol (20%, v/v)

For kinetic measurements at pH 6.5, TGase 2 stock solutions in buf-

fer II were used; these were diluted with buffer III.

v_{0 total}¼ v_{0 corrþ}v_{0 control}

ð3Þ

where v_0corrrepresents the rates for the enzymatic conversions.

buffer III: MES (pH 6.5, 100 mm), CaCl₂(3 mm), TCEP (10 mm), glyc-

erol (20%, v/v)

Within this model, the portion of _0corr=f([acyl donor]) and

v

v_0control=f([acyl donor]) were analysed by nonlinear regression by

All buffers were stored at 08C for periods of up to two weeks and

freshly prepared after that period. The concentrations of the

enzyme stock solutions were 0.5 mgmL^ꢂ1and 1 mgmL^ꢂ1for

gpTGase 2 and hTGase 2, respectively.

use of Equation (4) (Michaelis–Menten equation) and linear regres-

sion (v_0control=k_obs*[acyl donor]), respectively, with the data (see Fig-

ures S11 and S12 for global plots).

v_maxꢃ ½Sꢄ

ð4Þ

v_{0 corr}

¼

General assay procedure and analysis: All measurements were

conducted at 308C over 900 s (interval of 20 s) with a Synergy 4

Multi-Mode Microplate Reader (BioTek Instruments) and black 96-

well BRANDplates with transparent bottoms (BRAND). Fluorescence

was detected in bottom read mode. To detect released HMC or

HC, a combination of optical filters adjusted to 365/40 nm and

465/40 nm as ranges of wavelengths for excitation and emission,

respectively, were used. Measurements at pH 8.0 and pH 6.5 were

conducted with sensitivities of 35 and 45, respectively. The assay

mixture (200 mL) contained aqueous solution (190 mL) and DMSO

(5%, v/v, 10 mL). All regression analyses were accomplished with

GraphPad Prism (version 5.02, 17.12.2008). To provide values of

means and SEMs, the corresponding regression analyses were sep-

arately accomplished for each experiment, and the obtained fit

values were collected and statistically analysed.

K_mþ ½Sꢄ

Because of the negligible spontaneous reaction of compound 5b

at pH 6.5, plots of v_0total=f([acyl donor]) were directly analysed by

nonlinear regression to the data by use of Equation (4).

Analysis of the spontaneous reactions: The recorded time cours-

es of type (RFUꢂRFU₀)=f(t) for the spontaneous reactions were

analysed by linear regression to the experimental data over the

first 120 s. The respective slopes are equal to the values of v_0control

(units of RFUs^ꢂ1).

To determine the pseudo-first-order rate constants (k_obs), plots of

v

_0control=f([acyl donor]) were analysed by linear regression to the

data. The obtained slopes are identical to the k_obsvalues.

Assay procedure and kinetic analysis of the enzymatic aminoly-

sis: For investigations of enzyme-catalysed aminolysis reactions, six

different concentrations of the acyl donors were used (three inde-

pendent measurements, each concentration in duplicate). The ap-

propriate stock solutions were prepared in DMSO. Aminoacetoni-

trile was chosen as reference acyl acceptor. Aminoacetonitrile in

DMSO (16 mm, 5 mL) and the acyl donor stock solution (5 mL) were

added to assay buffer (180 mL). The reactions were initiated by ad-

dition of TGase 2 (10 mL, 6 mgmL^ꢂ1for 5a, 5b, 6a and 6b or

40 mgmL^ꢂ1for 3 and 4). For the measurement of the spontaneous

reactions, the TGase 2 solution was replaced by the appropriate

enzyme buffer. The recorded time courses of type (RFUꢂRFU₀)=

f(t) for the enzymatic conversions were analysed by nonlinear

regression with the experimental data over 180 s by use of Equa-

tion (1) (one-phase association). Values of v_0totaland v_0controlwere

obtained as described for the enzymatic hydrolysis. Because of the

high rates of spontaneous and enzymatic reactions, the calculated

substrate concentrations from the numerical integration were

used. Finally, the two sets of initial rates (v_0totaland v_0control) were

globally analysed with use of a model of total and nonspecific

Assay procedure and kinetic analysis of the enzymatic hydroly-

ses: For investigations on enzyme-catalysed hydrolysis reactions,

six or eight different concentrations of the acyl donors were used

(three independent measurements for each concentration in dupli-

cate). The corresponding stock solutions were prepared in DMSO.

DMSO (5 mL) and the acyl donor stock solution (5 mL) were added

to assay buffer (180 mL). The reactions were initiated by addition of

TGase 2 (10 mL, 60 mgmL^ꢂ1). For measurements of the spontaneous

reactions, the TGase 2 solution was replaced by the appropriate

enzyme buffer. The recorded time courses of type (RFUꢂRFU₀)=

f(t) for the enzymatic conversions were analysed by nonlinear

regression to the experimental data over the first 300 s by use of

Equation (1) (one-phase association). Subsequently, the first deriva-

tive of this function at t=0 [Eq. (2)] afforded the initial slopes,

which are equal to the values of v_0total(units of RFUs^ꢂ1).

RFUꢂRFU₀¼ plateau ꢃ ð1ꢂe^ꢂkt

Þ

ð1Þ

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

16

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

binding as described for the enzymatic hydrolysis (see Figures S14

and S15 for global plots). The evaluation of the data sets for com-

pounds 6a and 6b (gpTGase 2) and compound 5b (hTGase 2)

were performed by the method of Cornish-Bowden and Eisen-

thal.^[40]

v_s: steady-state velocity. v_i: initial velocity.

For iodoacetamide (gpTGase 2 and hTGase 2) and acrylamide 8

(gpTGase 2), the plot of k_obs=f([I]) was analysed by linear regres-

sion to the data. The corresponding slope (k_obs/I) was converted

into k_inact/K_Iby multiplication [Eq. (7)] by (1+[S]/K_m). For the inhibi-

tion of hTGase 2 by acrylamide 8, the double reciprocal plot 1/

Characterisation of acyl acceptors: To characterise the kinetic

properties of the chosen acyl acceptor substrates [aminoacetoni-

trile and N-(biotinyl)cadaverine (7)·TFA] six different concentrations

of the acyl acceptors were used. The appropriate stock solutions

were prepared in DMSO. Compound 5b was chosen as acyl donor.

Acyl acceptor stock solution (5 mL) and 5b in DMSO (4 mm, 5 mL)

were added to assay buffer (180 mL). The reactions were initiated

upon addition of TGase 2 (12 mgmL^ꢂ1, 10 mL). For measurement of

the spontaneous reactions, the TGase 2 solution was replaced by

the appropriate enzyme buffer. The two sets of initial rates (v_0total

and v_0control) were obtained as described for the enzymatic hydroly-

sis. Subsequently, the v_0controlvalues were subtracted from the

corresponding v_0totalvalues, and the resulting plot of v_0corr=f([S])

was analysed by use either of Equation (4) for aminoacetonitrile or

of Equation (5) for N-(biotinyl)cadaverine (7).

k

_obs=f([1/[I]) [Eq. (8)] was analysed by linear regression to the data

to determine the values for k_incat, K_Iand k_inact/K_I. The appropriate K_i

value was determined by analysis of v_i=f([I]) by use of Equation (9)

and the equation K_i=K_i’/(1+[S]/K_m), with K_m=6.60 mm.

k_inactꢃ ½Iꢄ

ð7Þ

ð8Þ

ð9Þ

k_obs

¼

K_Iþ ½Iꢄ

1

K_I

1

k_inact

ꢃ

þ

k_obsk_inact

½Iꢄ

v₀

v_i¼

þ v_0control

1 þ ½Iꢄ=K⁰_i

Numerical integration: For analysis of progress curves [RFU=f(t)]

by numerical integration the differential equations shown in

Scheme 5 and Scheme 7 were implemented in the freely available

software R (www.r-project.org/). A series of parameters were also

implemented: fluorescent coefficients of HMC or HC (coP), the acyl

donors (coSL) and the background fluorescence (bg), active con-

centrations of TGase 2, the rate constants k_obsfor spontaneous re-

actions of the acyl donors, time span between data points (20 s),

the entire time period for analysis (same as for the analysis by non-

linear regression) and starting values for the concentrations of the

acyl donors and inhibitors. These parameters had to be defined de-

pending on the data set prior to performing numerical integration.

Furthermore, the following relations were also implemented:

v_maxꢃ ½Sꢄ

ð5Þ

v_{0 corr}

¼

K_mþ ½Sꢄ ꢃ ð1 þ ½Sꢄ=K_iÞ

Because the measurements with N-(biotinyl)cadaverine (7) required

highly concentrated amine stock solutions (400 mm), the volumes

of the stock solutions were significantly greater than that of the

added DMSO for dissolving the compounds, due to the large

amount of amine, which finally resulted in lower concentrations

than intended. Therefore, the applied concentrations had to be

corrected. For this purpose, the density of the appropriate stock

solution was determined by weighing a defined volume, and the

correct concentration was calculated by using the obtained density

and the overall weight of the solution (DMSO+amine). The ratio

between the actual and the intended concentration of the stock

solution provided a factor (0.94) that was then used to correct the

concentrations of all other solutions made from the stock solution.

Enzymatic hydrolysis and aminolysis:

k

¼ k₁ꢃ K_mꢂk_cat

ꢂ1

where k₁was kept constant at the assumed value of 1 mm^ꢂ1s^ꢂ1

RFU ¼ coP ꢃ ½Pꢄ þ coSL ꢃ ½SLꢄ þ bg

Inhibition experiments: For the characterisation of coumarinyl

amide 5c with gpTGase 2 at pH 8.0, the enzymatic hydrolysis of

5b in the presence of four concentrations of 5c (0, 200, 300 and

500 mm) was recorded. The corresponding stock solutions of 5b

and 5c were prepared in DMSO. Stock solution of 5c (5 mL) and

stock solution of 5b (5 mL) were added to assay buffer (180 mL).

The reactions were initiated upon addition of gpTGase 2

(60 mgmL^ꢂ1, 10 mL). For the measurement of the spontaneous reac-

tions, the TGase 2 solution was replaced by the appropriate

enzyme buffer. Data evaluation was carried out as described in the

Supporting Information (Discussion S7).

For fitting of the progress curves, the kinetic parameters K_mand

k_catobtained by nonlinear regression or by the method of Cornish-

Bowden and Eisenthal were used as initial values.

Enzymatic hydrolysis in the presence of an irreversible inhibitor:

k

_ꢂ1þ k_cat

k₁¼

K_m

For the characterisation of iodoacetamide and acrylamide 8 with

gpTGase 2 and hTGase 2 at pH 6.5, eight or six concentrations of

the inhibitors were used. The appropriate stock solutions were pre-

pared in DMSO. Compound 5b was chosen as acyl donor. Iodoace-

tamide stock solution (5 mL) and either 5b (1 mm, for gpTGase 2)

or 5b (1.2 mm, 5 mL, for hTGase 2) in DMSO were added to assay

buffer (180 mL). The reactions were initiated upon addition of

TGase 2 (60 mgmL^ꢂ1, 10 mL). The recorded time courses of type

(RFUꢂRFU₀)=f(t) were analysed by nonlinear regression with the

experimental data over the entire measurement period (900 s) by

use of Equation (6):

where k_ꢂ1was kept constant at the assumed value of 1 s^ꢂ1

k

_ꢂ2þ k_inact

k₂¼

K_I

where k_ꢂ2was kept constant at the assumed value of 1 s^ꢂ1

k_inact¼ r ꢃ K_I

thus r=k_inact/K_I

_obst

ðv_iꢂv_sÞð1ꢂe^ꢂk

Þ

ð6Þ

RFUꢂRFU₀¼ v_sꢃ t þ

RFU ¼ coP ꢃ ½Pꢄ þ coSL ꢃ ½SLꢄ þ bg

k_obs

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

17

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

Full Papers

Giuffrida, O. Luinetti, E. Solcia, G. R. Corazza, Autoimmun. Rev. 2012 , 11 ,

74 6 –7 53; c) C. Klçck, T. R. DiRaimondo, C. Khosla, Semin. Immunopathol.

2012 , 34 , 51 3 –5 22.

The value of k_inact/K_Iobtained by regression analysis was used as in-

itial value for r, and an arbitrary value of K_I=15 mm was set. There-

fore, this procedure only allowed for the reliable calculation of the

second-order inactivation constant k_inact/K_I. Each experiment was

fitted separately, and final values of means ꢀSEMs shown in

Tables 2 and 3 were calculated by statistical analysis of all experi-

ments.

[8] a) L. Elli, C. M. Bergamini, M. T. Bardella, D. Schuppan, Dig. Liver Dis.

2009 , 41 , 54 1 –5 50; b) D. Caccamo, M. Curro, R. Ientile, Expert Opin.

Ther. Targets 2010 , 14 , 98 9 –1 003; c) K. C. Olsen, R. E. Sapinoro, R. M.

Kottmann, A. A. Kulkarni, S. E. Iismaa, G. V. Johnson, T. H. Thatcher, R. P.

Phipps, P. J. Sime, Am. J. Respir. Crit. Care Med. 2011 , 184 , 69 9 –7 07.

[9] a) H. Grosso, M. M. Mouradian, Pharmacol. Ther. 2012 , 133 , 39 2 –4 10;

b) A. C. Brett, T. R. Rosenstock, A. C. Rego, Curr. Drug Targets 2014, 15,

313–334; c) R. Ientile, M. Curro, D. Caccamo, Amino Acids 2015 , 47 , 1 9 –

26.

Presetting of the parameters above to arbitrarily assumed values

was necessary because the rate constants describing the reversible

enzyme·substrate and enzyme·inhibitor complexes cannot be cal-

culated independently.^[33]

[10] a) P. Kotsakis, M. Griffin, Amino Acids 2007 , 33 , 37 3 –3 8 4; b) K. Mehta, A.

Kumar, H. I. Kim, Biochem. Pharmacol. 2010 , 80 , 192 1 –1 92 9; c) S. Kumar,

K. Mehta, Amino Acids 2013 , 44 , 8 1 –8 8.

Acknowledgements

[11] M. Pietsch, R. Wodtke, J. Pietzsch, R. Lçser, Bioorg. Med. Chem. Lett.

2013 , 23 , 652 8 –6 543.

[12] S. M. Gillet, J. N. Pelletier, J. W. Keillor, Anal. Biochem. 2005 , 347 , 22 1 –

226.

[13] M. Walker, Top. Med. Chem. 2015, 9, 69–106.

[14] J. Wityak, M. E. Prime, F. A. Brookfield, S. M. Courtney, S. Erfan, S. John-

sen, P. D. Johnson, M. Li, R. W. Marston, L. Reed, D. Vaidya, S. Schaertl, A.

Pedret-Dunn, M. Beconi, D. Macdonald, I. MuÇoz-Sanjuan, C. Domi-

nguez, ACS Med. Chem. Lett. 2012 , 3 , 102 4 –1 028.

[15] M. Dixon, E. C. Webb, C. J. R. Thorne, K. F. Tipton, Enzymes, Academic

Press, New York, 1979, pp. 78–79.

[16] J. Polster, Reaktionskinetische Auswertung spektroskopischer Meßdaten,

Vieweg, Braunschweig, 1995, pp. 80–115.

[17] F. Garcꢂa-Martꢂn, N. Bayꢃ-Puxan, L. J. Cruz, J. C. Bohling, F. Albericio,

QSAR Comb. Sci. 2007 , 26 , 102 7 –1 035.

[18] J. D. K. Twibanire, T. B. Grindley, Org. Lett. 2011 , 13 , 298 8 –2 991.

[19] R. Bollhagen, M. Schmiedberger, K. Barlos, E. Grell, J. Chem. Soc. Chem.

Commun. 1994 , 255 9 –2 560.

[20] a) S. I. Chung, R. I. Shrager, J. E. Folk, J. Biol. Chem. 1970, 245, 6424–

6435; b) A. Leblanc, C. Gravel, J. Labelle, J. W. Keillor, Biochemistry 2001 ,

40 , 833 5 –8 342.

The authors thank Elisabeth Jꢁckel and Martin Lohse (Helmholtz-

Zentrum Dresden–Rossendorf, Institute of Radiopharmaceutical

Cancer Research) for assisting in the synthesis of compound 8

and 5c, respectively, and are grateful to Gçzde Sarican (Trakya

University, Edirne, Turkey, financially supported by a scholarship

within the Erasmus programme for residence at the University of

Cologne) for initial investigations into compounds 2a–2c. R.L.

appreciates continuing support by Prof. Dr. Jçrg Steinbach. Par-

tial financial support by the Helmholtz Portfolio Topic “Technolo-

gie und Medizin—Multimodale Bildgebung zur Aufklꢁrung des in

vivo-Verhaltens von polymeren Biomaterialien” (R.W., J.P. and

R.L.) and by the Fonds der Chemischen Industrie (R.L.) is gratefully

acknowledged. Furthermore, we wish to thank Prof. Dr. Jçrg van

den Hoff for helpful discussions concerning numerical integration.

[21] a) W. P. Jencks, J. Carriuolo, J. Am. Chem. Soc. 1960 , 82 , 177 8 –1 78 6;

b) O. S. Tee, O. J. Yazbeck, Can. J. Chem. 2000 , 78 , 110 0 –1 10 8; c) C.

Gravel, D. Lapierre, J. Labelle, J. W. Keillor, Can. J. Chem. 2007 , 85 , 16 4 –

174.

Keywords: enzyme assays · inhibitors · reactive peptides ·

solid-phase synthesis · transferases

[22] a) P. J. Bracher, P. W. Snyder, B. R. Bohall, G. M. Whitesides, Origin Life

Evol. Biosph. 2011 , 41 , 39 9 –4 12; b) Z. P. Gates, J. R. Stephan, D. J. Lee,

S. B. H. Kent, Chem. Commun. 2013 , 49 , 78 6 –7 88.

[23] a) M. J. Selwyn, Biochim. Biophys. Acta 1965 , 105 , 19 3 –1 95; b) A. Baici,

Kinetics of Enzyme–Modifier Interactions, Springer, Vienna, 2015, pp. 1–

64.

[24] R. A. Copeland, Enzymes: A Practical Introduction to Structure, Mecha-

nism, and Data Analysis, Wiley, Hoboken, 2000, pp. 109–145.

[25] a) D. E. Koshland, Bioorg. Chem. 2002 , 30 , 21 1 –2 13; b) R. Eisenthal, M. J.

Danson, D. W. Hough, Trends Biotechnol. 2007 , 25 , 24 7 –2 49.

[26] J. E. Folk, P. W. Cole, J. Biol. Chem. 1965, 240, 2951–2960.

[27] M. Griffin, A. Mongeot, R. Collighan, R. E. Saint, R. A. Jones, I. G. Coutts,

D. L. Rathbone, Bioorg. Med. Chem. Lett. 2008 , 18 , 555 9 –5 562.

[1] a) G. W. Jameson, D. V. Roberts, R. W. Adams, W. S. A. Kyle, D. T. Elmore,

Biochem. J. 1973 , 131 , 10 7 –1 17; b) S. P. Leytus, L. L. Melhado, W. F.

Mangel, Biochem. J. 1983 , 209 , 29 9 –3 07; c) J. L. Reymond, V. S. Fluxa, N.

Maillard, Chem. Commun. 2009, 34–46; d) S. S. van Berkel, B. van der

Lee, F. L. van Delft, R. Wagenvoord, H. C. Hemker, F. P. Rutjes, ChemMed-

Chem 2012, 7, 606–617; e) M. Sakabe, D. Asanuma, M. Kamiya, R. J. Iwa-

tate, K. Hanaoka, T. Terai, T. Nagano, Y. Urano, J. Am. Chem. Soc. 2013 ,

135 , 40 9 –4 14; f) M. Prost, L. Canaple, J. Samarut, J. Hasserodt, ChemBio-

Chem 2014 , 15 , 141 3 –1 417; g) K. R. Tallman, K. E. Beatty, ChemBioChem

2015 , 16 , 7 0 –7 5; h) A. Zadlo, D. Koszelewski, F. Borys, R. Ostaszewski,

ChemBioChem 2015 , 16 , 67 7 –6 82.

[2] a) D. Park, S. S. Choi, K. S. Ha, Amino Acids 2010 , 39 , 61 9 –6 31; b) S. Gun-

demir, G. Colak, J. Tucholski, G. V. Johnson, Biochim. Biophys. Acta Mol.

Cell Res. 2012 , 1823 , 40 6 –4 19; c) R. L. Eckert, M. T. Kaartinen, M. Nurmin-

skaya, A. M. Belkin, G. Colak, G. V. Johnson, K. Mehta, Physiol. Rev. 2014 ,

94 , 38 3 –4 17.

[3] a) N. K. Sarkar, D. D. Clarke, H. Waelsch, Biochim. Biophys. Acta 1957 , 25 ,

45 1 –4 52; b) M. J. Mycek, D. D. Clarke, A. Neidle, H. Waelsch, Arch. Bio-

chem. Biophys. 1959 , 84 , 52 8 –5 40; c) J. Schrode, J. E. Folk, J. Biol. Chem.

1978, 253, 4837–4840.

[4] M. J. Mycek, H. Waelsch, J. Biol. Chem. 1960, 235, 3513–3517.

[5] a) P. A. Smethurst, M. Griffin, Biochem. J. 1996 , 313 , 80 3 –8 08; b) N. S.

Caron, L. N. Munsie, J. W. Keillor, R. Truant, PLoS One 2012 , 7 , e44159;

c) M. S. Pavlyukov, N. V. Antipova, M. V. Balashova, M. I. Shakhparonov,

Biochem. Biophys. Res. Commun. 2012 , 421 , 77 3 –7 79.

ˇ

´

[28] M. Malesevic, A. Migge, T. C. Hertel, M. Pietzsch, ChemBioChem 2015,

16, 1169–1174.

[29] V. Gentile, M. Saydak, E. A. Chiocca, O. Akande, P. J. Birckbichler, K. N.

Lee, J. P. Stein, P. J. A. Davies, J. Biol. Chem. 1991, 266, 478–483.

[30] C. Hauser, R. Wodtke, R. Lçser, M. Pietsch, Amino Acids 2016, DOI:

10.1007/s00726-00016-02192-00725.

[31] I. Roy, O. Smith, C. M. Clouthier, J. W. Keillor, Protein Expression Purif.

2013 , 87 , 4 1 –4 6.

[32] a) A. Case, R. L. Stein, Biochemistry 2003 , 42 , 946 6 –9 481; b) J. W. Keillor,

C. M. Clouthier, K. Y. Apperley, A. Akbar, A. Mulani, Bioorg. Chem. 2014 ,

57 , 18 6 –1 97.

[33] A. Baici in Proteases: Structure and Function (Eds.: K. Brix, W. Stçcker),

Springer, Vienna, 2013, pp. 37–84.

[34] H. Tsunematsu, H. Nishimura, K. Mizusaki, S. Makisumi, J. Biochem. 1985,

97, 617–623.

[35] a) K. Hardes, G. L. Becker, M. Z. Hammamy, T. Steinmetzer, Anal. Biochem.

2012 , 428 , 7 3 –8 0; b) K. Hardes, M. Zouhir Hammamy, T. Steinmetzer,

Anal. Biochem. 2013 , 442 , 22 3 –2 30.

[6] a) H. Nakaoka, D. Perez, K. Baek, T. Das, A. Husain, K. Misono, M. Im, R.

Graham, Science 1994 , 264 , 159 3 –1 59 6; b) S. S. Akimov, D. Krylov, L. F.

Fleischman, A. M. Belkin, J. Cell Biol. 2000 , 148 , 82 5 –8 38; c) A. M. Belkin,

FEBS J. 2011 , 278 , 470 4 –4 716.

[7] a) W. Dieterich, T. Ehnis, M. Bauer, P. Donner, U. Volta, E. O. Riecken, D.

Schuppan, Nat. Med. 1997 , 3 , 79 7 –8 01; b) A. Di Sabatino, A. Vanoli, P.

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

18

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Full Papers

[36] A. Baici, Kinetics of Enzyme-Modifier Interactions, Springer, Vienna, 2015,

pp. 127–169.

[43] M. T. Gundersen, J. W. Keillor, J. N. Pelletier, Appl. Microbiol. Biotechnol.

2014 , 98 , 21 9 –2 30.

[44] A. Cornish-Bowden, Fundamentals of Enzyme Kinetics, Wiley-Blackwell,

Oxford, 2012, pp. 211–213.

[37] M. C. Yi, B. A. Palanski, S. A. Quintero, N. M. Plugis, C. Khosla, Bioorg.

Med. Chem. Lett. 2015 , 25 , 492 2 –4 926.

[45] R. L. Stein, Kinetics of Enzyme Action: Essential Principles for Drug Hunt-

ers, Wiley, Hoboken, 2011, pp. 122–137.

[38] R. L. Stein, Kinetics of Enzyme Action: Essential Principles for Drug Hunt-

ers, Wiley, Hoboken, 2011, pp. 141–168.

[46] R. A. Copeland, Evaluation of Enzyme Inhibitors in Drug Discovery, Wiley,

Hoboken, 2005, pp. 214–248.

[39] J. E. Folk, J. Biol. Chem. 1969, 244, 3707–3713.

[40] A. Cornish-Bowden, Fundamentals of Enzyme Kinetics, Wiley-Blackwell,

Oxford, 2012, pp. 45–51.

[47] J. E. Folk, P. W. Cole, J. Biol. Chem. 1966, 241, 3238–3240.

[48] A. G. McDonald, K. F. Tipton in eLS, Wiley, Chichester, 2012, pp. 1–17.

[49] C. Cobas, S. Dominguez, N. Larin, I. Iglesias, C. Geada, F. Seoane, M.

Sordo, P. Monje, S. Fraga, R. Cobas, C. Peng, J. A. Garcia, M. Goebel, E.

Vaz, MestReNova 6.1.1–6384, Mestrelab Research S. L., 2010.

[41] a) R. Eisenthal, A. Cornish-Bowden, Biochem. J. 1974 , 139 , 71 5 –7 20;

b) R. Eisenthal, A. Cornish-Bowden, Biochem. J. 1974, 139, 721–730;

c) A. Baici, Biochemist 2006, 28, 36–39.

[42] a) T. F. Slaughter, K. E. Achyuthan, T. S. Lai, C. S. Greenberg, Anal. Bio-

chem. 1992 , 205 , 16 6 –1 71; b) J.-H. Jeon, C.-W. Kim, D.-M. Shin, K.-i. Kim,

S.-Y. Cho, J.-C. Kwon, K.-H. Choi, H.-S. Kang, I.-G. Kim, FEBS Lett. 2003 ,

534 , 18 0 –1 84; c) X. Jin, J. Stamnaes, C. Klçck, T. R. DiRaimondo, L. M.

Sollid, C. Khosla, J. Biol. Chem. 2011 , 286 , 3786 6 –3 7873.

Manuscript received: January 25, 2016

Accepted article published: April 28, 2016

Final article published: && &&, 0000

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

19

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

&

These are not the final page numbers! ÞÞ

FULL PAPERS

R. Wodtke, G. Schramm, J. Pietzsch,

M. Pietsch,* R. Lçser*

Fluorimetric assay for Transglutami-

nase 2: A set of fluorogenic acyl donor

substrates for transglutaminase 2 was

synthesised by a modular solid-phase

strategy. Their good aqueous solubility

allowed for a detailed kinetic characteri-

sation with regard to their TGase 2-cata-

lysed hydrolysis and aminolysis. Their

applicability to characterising amine

substrates and inhibitors of TGase 2 was

also demonstrated.

&& – &&

Synthesis and Kinetic Characterisation

of Water-Soluble Fluorogenic Acyl

Donors for Transglutaminase 2

&

ChemBioChem 2016, 17, 1 – 20

www.chembiochem.org

20

ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ÝÝ These are not the final page numbers!

Article Doi

Synthesis and Kinetic Characterisation of Water-Soluble Fluorogenic Acyl Donors for Transglutaminase 2

DOI: 10.1002/cbic.201600048

Source and publish data:

Authors:

Article abstract of DOI:10.1002/cbic.201600048

Full text of DOI:10.1002/cbic.201600048

Products guided by the article

R&D Labs maybe for 873561-88-7

Relevant to this article

Hot Product