Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic amines involves transamidation and carbon-carbon bond formation at the 2-position of the amine

Article abstract of DOI:10.1016/j.bbagen.2015.04.007

Background Tryptophan catabolism along the kynurenine pathway is associated with a number of pathologies including cataract formation and cancer. Whilst the chemical reactions of kynurenine are well studied, less is known about the reactivity of its precu

Full text of DOI:10.1016/j.bbagen.2015.04.007

BBAGEN-28194; No. of pages: 9; 4C: 5, 6, 7
Biochimica et Biophysica Acta xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbagen
1Q1 Formation of fluorophores from the kynurenine pathway metabolite
2
N-formylkynurenine and cyclic amines involves transamidation and
carbon–carbon bond formation at the 2-position of the amine
3
⁎
4Q2 Petr Tomek, Brian D. Palmer, Jackie D. Kendall, Jack U. Flanagan, Lai-Ming Ching
5
Auckland Cancer Society Research Centre, Faculty of Medical and Health Sciences, School of Medicine, University of Auckland, Auckland, New Zealand
OOF
6
a r t i c l e i n f o
a b s t r a c t
7
8
9
10
Article history:
Background: Tryptophan catabolism along the kynurenine pathway is associated with a number of pathologies 18
including cataract formation and cancer. Whilst the chemical reactions of kynurenine are well studied, less is 19
known about the reactivity of its precursor N-formylkynurenine (NFK). We previously reported the generation 20
of a strong fluorophore in an aqueous reaction of NFK with piperidine, and herein we describe its structure 21
Received 26 February 2015
Accepted 13 April 2015
Available online xxxx
and mechanism of formation.
Methods: Compounds were identified using NMR, mass and UV spectroscopic techniques. The products from the 23
reaction of amines with amino acids were quantified using HPLC-MS. 24
Results: The novel fluorophore was identified as a tetrahydroquinolone adduct (PIP-THQ), where piperidine 25
is N-formylated and attached at its 2-position to the quinolone. NFK is initially deaminated to generate an un- 26
saturated enone, which forms an adduct with piperidine and is subsequently converted into the fluorophore. 27
Testing of a variety of other secondary amines showed that only cyclic amines unsubstituted at both positions 28
adjacent to nitrogen could form fluorophores efficiently. The amino acids tryptophan and kynurenine, which 29
22
11
12
13
14
15
16
17
Keywords:
Indoleamine 2,3-dioxygenase
N-formylkynurenine
Tryptophan catabolism
Transamidation
Piperidine
Fluorophore
lack the formamide group do not form such fluorophores.
30
Conclusions: NFK forms fluorophores in a not previously published reaction with cyclic amines.
31
General significance: Our study is the first to provide evidence for concurrent transamidation and substitution at 32
the 2-position of a cyclic amine occurring under moderately-heated aqueous conditions with no added catalysts. 33
The high reactivity of NFK demonstrated here could result in formation of biologically relevant metabolites yet to 34
be characterised.
35
36
© 2015 Elsevier B.V. All rights reserved.
40
378
39
41
1. Introduction
1 (IDO1) and tryptophan 2,3-dioxygenases (TDO), [3] and then hy- 45
drolysed into kynurenine (KYN) by kynurenine formamidase [4]. 46
Further enzymatic conversion of KYN ultimately leads to production 47
of the cofactor nicotinamide adenine dinucleotide [1]. KYN can 48
also spontaneously deaminate at physiological pH to form a reactive 49
α,β-carboxyketoalkene (KYN-CKA) [5] which can either form 50
adducts with biological nucleophiles [6,7], undergo reduction or 51
cyclise to form kynurenic acid or kynurenine yellow (Scheme 1) 52
42
43
44
The kynurenine pathway is a major tryptophan (TRP) catabolic route
in humans [1,2]. In this pathway, TRP is firstly oxidised into N-
formylkynurenine (NFK) by haem enzymes indoleamine 2,3-dioxygenase
[8–10].
53
Abbreviations: 2-MePIP, (2-methylpiperidine); 3-MePIP, (3-methylpiperidine); 4-
The kynurenine pathway has recently attracted intense interest due 54
to its central role in causing tumour mediated immune suppression 55
[11]. IDO1 has been found in a broad range of human tumours causing 56
local depletion of TRP and accumulation of KYN, which result in accu- 57
mulation of T regulatory suppressor cells and inactivation of T effector 58
cells [12,13]. High IDO1 expression in tumours is associated with poor 59
prognosis for cancer patients [14–17]. Blocking IDO1 activity using 60
small-molecule inhibitors is now a recognised approach for cancer ther- 61
apy [18–20] and two IDO1 inhibitors are currently showing promise in 62
MePIP, (4-methylpiperidine); AZP, (azepane); DEA, (diethylamine); DMP, (cis-2,6-
UNCORRECTED PR
dimethylpiperidine); DPA, (dipropylamine); HRMS, (high-resolution mass spectrometry);
IDO1, (indoleamine 2,3-dioxygenase 1); KYN, (kynurenine); KYN-CKA, (kynurenine α,β-
carboxyketoalkene); NFK-CKA, (N-formylkynurenine α,β-carboxyketoalkene); MQW,
(Milli-Q water); MW, (molecular weight); NFK, (N-formylkynurenine); NMePIP, (N-
methylpiperidine); NMePYR, (N-methylpyrrolidine); PIP, (piperidine); PYR, (pyrrolidine);
Rt, (retention time); TDO, (tryptophan 2,3-dioxygenase); TFA, (trifluoroacetic acid); THQ,
(tetrahydroquinolone); TMP, (2,2,6,6-tetramethylpiperidine); TRP, (tryptophan).
⁎
Corresponding author at: Auckland Cancer Society Research Centre, Faculty of Medical
and Health Sciences, University of Auckland, Auckland, New Zealand. Tel.: +64 9 373
7599x86140.
human clinical trials [21,22].
63
E-mail address: l.ching@auckland.ac.nz (L.-M. Ching).
http://dx.doi.org/10.1016/j.bbagen.2015.04.007
0304-4165/© 2015 Elsevier B.V. All rights reserved.
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

2
P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
2. Materials and methods
86
87
2.1. Materials
DL-Kynurenine (N95%, cat. no. 69791) obtained from AK Scientific
88
(Union City, CA, USA) with only 62% purity by HPLC compared to crys- 89
talline L-Kynurenine (cat. no. K8625) from Sigma-Aldrich (St Louis, 90
MO, USA), was used with no further purification. 13C-labelled 91
formic acid (99%, cat. no. 14C-428) was purchased from Cambridge 92
Isotope Laboratories. All other chemicals were obtained from Merck or 93
Sigma-Aldrich.
94
2.2. General analytical methods
95
NMR spectra were obtained on a Bruker Avance 400 spectrometer at 96
400 MHz for 1H and 100 MHz for 13C spectra and were referenced to 97
tetramethylsilane. High resolution mass spectra were determined on a 98
Bruker micrOTOF-Q mass spectrometer (HRMS) operating under 99
electrospray ionisation conditions. LC-MS analyses were carried out on 100
Agilent 1100 HPLC/6150 single quadrupole electrospray mass spec- 101
trometer. Ionisation conditions were: drying gas temperature 270 °C, 102
gas flow 10 L/min, nebulizer pressure gauge 35 psi, capillary voltage 103
+3 kV and −3 kV, and fragmentor voltage 70 V. Spectra were acquired 104
over 107–1000 m/z range. Analytical separations were carried out on a 105
Luna C18 column (5 μm, 100 A, 150 × 2 mm; Phenomenex, Torrance, 106
CA, USA) eluted using 80% acetonitrile (MeCN) (A) and Milli-Q water 107
(MQW) containing 25 mM formic acid (pH ~2.7) (B) in a binary 108
gradient: 0 min (5% A), 11 min (58% A), 14–17 min (96% A), 18– 109
22 min (5% A) at flow rate 0.4 mL/min and 40 °C. UV/VIS chromato- 110
grams were acquired using diode-array detection at multiple wave- 111
lengths simultaneously ranging from 240 to 400 nm, and fluorescence 112
signals were acquired at emission and excitation wavelengths of 113
500 nm and 400 nm, respectively, using an in-line fluorescence detec- 114
tor. Data were analysed on Agilent Chemstation software.
115
The purity of isolated compounds was determined by HPLC 116
analysis (from 254 nm chromatogram) as a peak area of compound 117
of interest divided by the peak area of all analysed peaks in the 118
chromatogram. Peak selection was performed manually in Agilent 119
Chemstation software.
120
Scheme 1. Overview of major chemical transformations of KYN and NFK. MW stands for
Mass fragmentation experiments were performed on an Agilent 121
1200 HPLC/6460 triple quadrupole mass spectrometer equipped with 122
Agilent JetStream electrospray ionisation interface (drying gas temper- 123
ature 250 °C, flow 10 L/min, nebulizer pressure 40 psi, capillary voltage 124
+2.75 kV and −3.5 kV, sheath gas temperature 250 °C, sheath gas flow 125
6 L/min, collision cell accelerator voltage 7 V, fragmentor voltage was 126
50 V and 135 V for MS2 scans and MS/MS fragmentation experiments, 127
respectively). MS2 scans were acquired over the 50–1000 m/z range. 128
Samples were injected into LC-MS with no column in a mobile phase 129
consisting of solution A (4.5 mM ammonium formate pH 3.5) and solu- 130
tion B (0.1% (v/v) formic acid in MeCN in ratio 6:4 (A:B) and at a flow 131
rate 0.3 mL/min). Data were analysed in Agilent MassHunter software. 132
Fluorescence titration of PIP-THQ (Fig. 2) was performed on an 133
EnSpire 2300 Multimode plate reader (Perkin-Elmer, Singapore) in a 134
black polypropylene 384-well plate (Cat. No. 781209, Greiner Bio-One, 135
Frickenhausen, Germany). Fluorescence emission spectra of the 136
fluorophore amine-THQ were acquired at 400 nm excitation during 137
HPLC analysis. Extinction coefficients were determined in potassium 138
phosphate buffer (0.1 M in MQW, pH 7) at 25 °C on an Agilent 8453 139
molecular weight in Daltons.
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
Testing of compounds for IDO1 inhibitory activity in the past
has relied on a colorimetric assay introduced in 1988 [23]. In this
assay, the NFK produced from IDO1-catalysed conversion of TRP is
firstly hydrolysed to KYN in trichloroacetic acid and then reacted with
p-dimethylaminobenzaldehyde to form a Schiff base and quantified at
480 nm. The IDO1 fluorescence assay introduced in 2006 measures
fluorescence of KYN produced from hydrolysis of NFK in sodium
hydroxide [24]. The sensitivity of this fluorescence assay is comparable
to that of the original absorbance assay, but has the advantage of
requiring only a single step incubation at 65 °C with sodium hydroxide.
We recently developed a fluorescence assay with 30-fold better
sensitivity than the previous assays for measuring IDO1 activity.
In this new assay, NFK is converted into a strong fluorophore in a
reaction at 65 °C with cyclic amine piperidine (PIP) [25]. To our
knowledge, the formation of fluorophore from NFK has not been report-
ed, and the only published transformation of NFK is photooxidation into
N-formylanthranilic acid and 4-hydroxyquinoline [26]. In this commu-
nication we report on the isolation and characterisation of the chemical
structure of the novel fluorophore which we have called piperidine-
tetrahydroquinolone (PIP-THQ) and provide experimental evidence
for a previously unreported chemical transformation pathway leading
to its formation.
UV–VIS spectrophotometer (Hewlett-Packard).
140
2.3. Small-scale reactions of amino acids and amines
141
Amino acids (2 mM) were incubated with amines (1 M) in 100– 142
200 μL MQW in 1.5 mL Eppendorf tubes heated in Thermomixer 143
Comfort (Eppendorf) at 65 °C for 20 min, then immediately cooled on 144
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
3
145 ice and analysed on LC-MS (single quadrupole). Peak areas (mV s) of
146 corresponding adducts were quantified from 254 nm (amine-KYN and
147 amine-NFK) and 400 nm (amine-THQ) chromatograms. Amine-amino
148 acid adducts were identified in chromatograms by their a) predicted
149 molecular ions in positive and negative mass spectra and b) virtually
150 identical absorption spectra to those of a parent amino acids KYN and
151 NFK. Amine-THQ adducts (fluorophores) were further validated from
152 fluorescence chromatograms. Since amine-NFK and amine-KYN were
153 incompletely separated, the double peak in chromatograms was split
154 at the trough as depicted on Fig. 3a (dashed line) to allow quantification.
155 Aliquots (10–20 μL) for each time point during reaction progress exper-
156 iments (Fig. 4) were taken from the same tube and immediately chilled
157 on ice prior to HPLC analysis.
vacuum manifold and fractions (10 mL) were analysed on HPLC. PIP- 206
THQ fractions were pooled, evaporated, resuspended in sodium hydrox- 207
ide (143 mM, 3.5 mL) and evaporated again. The sample was resus- 208
pended in MQW (0.5 mL), loaded onto a C18 cartridge (1 g/6 mL, 209
Varian) and eluted successively with MQW, 50% MeCN in MQW and 210
100% MeCN. MeCN eluates were pooled, evaporated and dried at high 211
vacuum to yield PIP-THQ as a yellow powder after freeze-drying 212
(17.6 mg, 0.96%). Purity was 94%. The product was a 1:1 mixture of di- 213
astereomers. 1H NMR δ (400 MHz, D2O, 298 K) 8.06 (s. 0.5H, CHO), 214
8.02 (s, 0.5H, CHO), 7.50–7.46 (m, 2H, H-5,7), 6.92 (dd, J = 8.2, 0.8 Hz, 215
1H, H-8), 6.80–6.74 (m, 1H, H-6), 4.45 (br d, J = 6.5 Hz, 0.5H, H-2′), 216
4.01 (dd, J = 14.6, 4.4 Hz, 0.5H, H-6′), 3.80 (dd, J = 6.8, 3.2 Hz, 0.5H, 217
H-2′), 3.50 (dd, J = 13.1, 4.6 Hz, 0.5H, H-6′), 3.40 (m, 0.5H, H-6′), 2.71 218
(m, 0.5H, H-6′), 2.52 (2xd, J = 15.1 Hz, 1H, H-3), 2.49 (2xd, J = 219
15.1 Hz, 1H, H-3), 1.59–1.20 (m, 6H, H-3′, 4′ 5′). 13C NMR spectrum-see 220
Table S.1. MS m/z 303.1 (90.94%, [M + H]+), 112.1 (100%, N-formyl- 221
2,3,4,5-tetrahydropyridin-1-ium), 627.1 (40.22%, [2 M + Na]+). HRMS 222
m/z calcd for C16H18N2NaO4 325.1159, found 325.1162 [M + Na]+. 223
158 2.4. Compound synthesis and characterisation
159
The isolation procedures described below are a result of optimisa-
160 tions. HPLC chromatograms of isolated compounds can be found in
161 Fig. S.1. The yields are calculated based on 62% purity of DL-KYN.
λ
max(H2O)/nm 400 (ε/M−1 cm−1 2136).
224
162 2.4.1. 2-Amino-4-(2-formamidophenyl)-4-oxobutanoic acid (NFK)
OOF
163
Acetic anhydride (0.24 mL, 2.54 mmol) was added to formic
2.4.4. [13C-formyl] labelled PIP-THQ
225
13C-formyl]-NFK (2.56 mg, 0.011 mmol) was incubated with 226
164 acid (0.48 mL, 12.7 mmol) and the solution was warmed at 50–55 °C
165 for 15 min, then cooled to room temperature. A solution of DL-
166 kynurenine (0.50 g, 2.40 mmol) in formic acid (14 mL) was added and
167 the mixture was stirred at room temperature for 2 h. Ether was added
168 to precipitate out the product, which was washed further with ether
169 and dried in vacuo to give NFK as a hygroscopic tan powder (0.38 g,
170 ~100%). Purity was 96%. 1H NMR δ (400 MHz, D2O, 298 K) (rotamers
171 about the formanilide group evident) 8.90 (s, 0.35H, CHO, rotamer A),
172 8.40 (s, 0.65H, CHO, rotamer B), 8.20 (d, J = 8.2 Hz, 0.65H, H-6′, rotamer
173 B), 8.09 (d, J = 8.2 Hz, 0.35H, H-6′, rotamer B), 8.05 (d, J = 8.0 Hz, 0.65H,
174 H-3′, rotamer B), 7.71 (dd, J = 8.2, 7.7 Hz, 1H, H-5′, rotamers A and B),
175 7.60 (d, J = 8.0 Hz, 0.35H, H-3′, rotamer A), 7.41 (dd, J = 8.0, 7.7 Hz, 1H,
176 H-4′, rotamers A and B), 4.20 (t, J = 4.9 Hz, 1H, H-2), 3.80 (d, J = 4.9 Hz,
177 2H, H-3). MS m/z 237.0 (100%, [M + H]+). HRMS m/z calcd
178 for C11H13N2O4 237.0870, found 237.0864 [M + H]+. λmax(H2O)/nm
179 261 and 322 (ε/M−1 cm−1 6289 and 1936), lit., [27] nm 260 and 321
[
PIP (0.3 mL, 3.04 mmol) in MQW (3 mL) at 65 °C for 20 min. The 227
reaction was cooled on ice, acidified to ~pH 1 with HCl and extracted 228
twice with ethyl acetate (EtOAc). The EtOAc extract was washed 229
twice with saturated sodium chloride solution containing 0.3% (v/v) 230
HCl, evaporated and resuspended in methanol (MeOH) prior to MS 231
analysis.
232
2.4.5. 2-(1-formyl-5-methylpiperidin-2-yl)-4-oxo-1,2,3,4-tetrahydro- 233
quinoline-2-carboxylic acid (3-MePIP-THQ) 234
NFK (80.24 mg, 7.8 mmol) was incubated with 3-methylpiperidine 235
(3-MePIP; 2.0 mL, 17.04 mmol) in MQW (17 mL) at 65 °C for 30 min. 236
The reaction was evaporated and subsequently redissolved in 15 mL 237
MQW, acidified to ~pH 1 with HCl and extracted with EtOAc. The 238
dried extract was dissolved in 3 mL MeOH:MeCN (5:1) and diluted 239
with 30 mL MQW. The extract was fractionated on a Strata C18-E 240
cartridge (5 g/20 mL, Phenomenex) using mixtures of MeCN/MQW 241
containing 0.04% (v/v) TFA. Fractions containing 3-MePIP-THQ were 242
pooled, evaporated and dissolved in MQW, adjusted to ~pH 8 by 243
sodium hydroxide and mixed with ammonium bicarbonate buffer 244
(0.02 M final concentration, pH 8). 3-MePIP-THQ was subsequently 245
purified on a Strata C18-E cartridge (2 g/12 mL, Phenomenex) and 246
eluted by mixtures of MeCN/MQW containing ammonium bicarbonate 247
(0.02 M, pH 8). Eluted 3-MePIP-THQ was acidified to ~pH 1 and 248
extracted with EtOAc, washed with saturated sodium chloride 249
solution and dried over sodium sulphate. Subsequent evaporation and 250
drying at high vacuum afforded 3-MePIP-THQ as a yellow powder 251
(1.47 mg, 1.83%). Purity was 94%. The product was a mixture of 252
diastereomers. 1H NMR δ (400 MHz, CDCl3, 298 K) 8.21 (s. 0.5H, 253
CHO), 8.03 (s, 0.5H, CHO), 7.64 (dd, J = 8.4, 8.2 Hz, 1H, ArH), 7.53 254
(ddd, J = 8.2, 8.2, 1.2 Hz, 0.5H, ArH), 7.48 (ddd, J = 8.2, 8.2, 1.2 Hz, 255
0.5H, ArH), 7.02 (d, J = 8.2 Hz, 0.5H, ArH), 6.96–6.84 (m, 1.5H, ArH), 256
6.48 (br s, 0.5H, NH), 6.24 (br s, 0.5H, NH), 4.88 (br d, J = 4.1 Hz, 257
0.5H, CHN), 4.24 (dd, J = 13.4, 3.4 Hz, 0.5H, CHHN), 3.97 (d, J = 258
6.6 Hz, 0.5H, CHN), 3.23 (dd, J = 13.8, 4.1 Hz, 0.5H, CHHN), 2.97 (d, 259
J = 16.3Hz, 0.5H, CHHCO), 2.87 (dd, J = 11.3, 11.3Hz, 0.5H, CHHN), 260
2.69 (d, J = 16.0Hz, 0.5H, CHHCO), 2.61 (dd, J = 12.3, 12.3 Hz, 0.5H, 261
CHHN), 2.48 (d, J = 16.3 Hz, 0.5H, CHHCO), 2.23 (d, J = 16.0 Hz, 0.5H, 262
CHHCO), 1.77–1.28 (m, CH2), 0.94 (d, J = 6.4 Hz, 1.5H, CH3), 0.90 263
(d, J = 6.0 Hz, 1.5H, CH3). MS m/z 317.1 (98.18%, [M + H]+), 126.1 264
(100%, N-formyl-3-methyl-2,3,4,5-tetrahydropyridin-1-ium), 339.0 265
(60.60%, [M + Na]+), 655.2 (32.11%, [2 M + Na]+). λmax(H2O)/nm 266
180
181
182
183
184
(ε/M−1 cm−1 10,980 and 3750). Other authors reported λmax(H2O)/nm
321 (ε/M−1 cm−1 3152) from commercially available product [28]. We
have previously isolated a small quantity of pure NFK from enzymatically
catalysed oxidative cleavage of TRP [25] and obtained λmax(H2O)/nm 323
(ε/M−1 cm−1 3066).
185 2.4.2. [13C-formyl] labelled NFK
186
This material was prepared from DL-kynurenine as described above
187 for NFK, except that the acetic anhydride was added to [13C]-HCOOH
188 in the first step. The kynurenine was added to the resulting solution, dis-
189 solved in normal formic acid. The 1H NMR spectrum of the product was
190 the same as that described above for NFK, except that there were two
191 additional doublets present from the rotamers of the 13C-labelled
192 formyl group, due to 1H–13C coupling, at δ 8.86 (d, JH-C = 234.3 Hz,
193 CHO, rotamer A) and 8.40 (d, JH-C = 203.1 Hz, CHO, rotamer B). 1H
194
195
196
NMR analysis indicated the sample to consist of an approximately 3:2
mixture of 13C:12C-labelled NFK. MS m/z 237.0 (64.88%, [12C M + H]+),
238.0 (100%, [13C M + H]+).
UNCORRECTED PR
197 2.4.3. 2-(1-formylpiperidin-2-yl)-4-oxo-1,2,3,4-tetrahydroquinoline-2-
198 carboxylic acid (PIP-THQ)
199
NFK (1.84 g, 7.8 mmol) was incubated with PIP (45.5 mL, 460 mmol)
200 in 368 mL MQW at 60 °C for 20 min. The reaction was cooled on ice,
201 evaporated to dryness and subsequently redissolved in 0.04% (v/v)
202 trifluoroacetic acid (TFA) in MQW (15 mL). This solution was loaded
203 onto a Strata C18-E cartridge (5 g/20 mL, Phenomenex) preconditioned
204 with 70 mL MeCN and equilibrated with 70 mL of 0.04% (v/v) TFA in
205 MQW. The column was eluted with mixtures of MeCN/0.04% TFA on a
402 (ε/M−1 cm−1 2431).
267
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

4
P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
268 2.4.6. 4-(2-aminophenyl)-4-oxo-2-(piperidin-1-yl)butanoic acid
269 (PIP-KYN)
under aqueous conditions, and the fluorophore was purified using re- 329
versed phase preparative chromatography (see Section 2.4.3). The 330
product displayed a molecular ion adduct [M + H]+ at m/z 303.1 and so- 331
dium adduct [M + Na]+at 325.1159 amu on low resolution triple quad- 332
rupole mass spectrometry and high resolution time-off-flight mass 333
spectrometry, respectively, consistent with a molecular formula 334
C16H18N2O4 and molecular weight 302 Da. It was clear from both the 335
1H and 13C NMR spectra that the sample was a 1:1 mixture of isomers, 336
with most of the resonances being doubled up (Fig. S.2a,b). The four ar- 337
omatic resonances of the starting NFK were still present, and resonances 338
characteristic of a PIP group were clearly evident at δ 4.4–2.7 ppm 339
(CH\\N) and δ 1.8–1.3 ppm (CH2CH2). Singlets at δ 8.06 and 8.02 ppm 340
were consistent with the presence of a formyl group. Particularly diag- 341
nostic were resonances resulting from two overlapping AB spin systems 342
centred at around 2.5 ppm, each with a large coupling constant of ap- 343
proximately 15 Hz. This indicates the presence of an isolated CH2 344
group, probably in a ring system. Two-dimensional NMR studies 345
(Fig. S.3–S.7) were used to identify coupled proton systems and to cor- 346
relate proton and carbon resonances (Table S.1). From these it was evi- 347
dent that there were only three protons adjacent to the nitrogen atom of 348
the PIP, with two of these displaying geminal coupling to each other. 349
The remaining proton did not display geminal coupling, and appeared 350
as a much narrower resonance. This indicates that the PIP has reacted 351
at one of the carbon atoms adjacent to the nitrogen of the PIP, with 352
the loss of one of its hydrogen atoms. Finally, a nitrogen NMR spectrum 353
was obtained and indicated the presence of two nitrogen atoms in the 354
molecule (δ 88 and 134 ppm) (Fig. S.8). Based on considerations of all 355
this data we propose the structure shown in Scheme 2 for PIP-THQ. 356
The product is a mixture of two racemic diastereomers at the 2 and 2′ 357
positions (Table S.1). The placement of the formyl group on the PIP ni- 358
trogen and not the nitrogen of the quinolone ring was based on the 359
NOESY spectrum, which suggested that it was close to the PIP ring pro- 360
270
KYN (16.64 mg, 0.08 mmol) was incubated with PIP (0.395 mL,
271 4 mmol) in MQW (4 mL) at 80 °C for 30 min. The reaction was evaporat-
272 ed, redissolved in MQW and acidified to pH ~4 with TFA. Purification
273 was carried out on a C18-E cartridge (2 g/12 mL, Phenomenex) using
274 MeCN/MQW mixtures containing 0.016% (v/v) TFA. Freeze-drying of
275 purified PIP-KYN yielded a brown powder (4.7 mg, 45%). Purity was
276 95%. 1H NMR δ (400 MHz, D2O, 303 K) 7.99 (dd, J = 8.1, 1.3 Hz, 1H, H-
277 6′), 7.57 (ddd, J = 7.2, 7.1, 1.2 Hz, 1H, H-4′), 7.25 (ddd, J = 8.1, 7.2,
278 1.2 Hz, 1H, H-5′), 7.19 (dd, J = 7.1, 1.2 Hz, H-3′), 4.35 (dd, J = 7.0,
279 4.4 Hz, 1H, H-2), 3.85 (dd, J = 18.2, 7.0 Hz, 1H, H-3), 3.72 (dd, J =
280 18.2, 4.4 Hz, 1H, H-3), 3.51 (br d, J = 11.7 Hz, 1H, CHHN), 3.40 (br d,
281 J = 11.7 Hz, 1H, CHHN), 3.13–2.95 (m, 2H, CH2N), 1.93–1.68 (m, 5H,
282 CH2), 1.47–1.34 (m, 1H, CH). MS m/z 277.2 (100%, [M + H]+). HRMS
283 m/z calcd for C15H21N2O3 277.1547, found 277.1538 [M + H]+
284
max(H2O)/nm 259 and 364 (ε/M−1 cm−1 7213 and 4010).
.
λ
285 2.4.7. 4-(2-aminophenyl)-4-oxo-2-butenoic acid (KYN-CKA)
286
KYN (58.24 mg, 0.28 mmol) was incubated with sodium hydroxide
287 (7 mg, 0.175 mmol) in MQW (8.75 mL, pH ~12) at 80 °C for 9 min,
288 immediately cooled on ice and brought to pH ~2 using HCl. This solution
289 was extracted three times with EtOAc, and the organic layer was
290 pooled and washed four times with saturated NaCl. Subsequent
291 evaporation and drying at high vacuum afforded KYN-CKA as a bright
292 orange powder (9 mg, 24%). Purity was 96%. 1H NMR δ (400 MHz,
293 DMSO-d6, 298 K) 7.74 (dd, J = 8.2, 1.5 Hz, 1H, H-6′), 7.53 (d, J =
294 15.7 Hz, 1H, H-3), 7.32 (ddd, J = 7.1, 7.2, 1.5 Hz, 1H, H-4′), 6.78 (dd,
295 J = 8.7, 0.9 Hz, 1H, H-3′), 6.70 (ddd, J = 8.2, 7.1, 0.9 Hz, 1H, H-5′),
296 6.62 (d, J = 15.7 Hz, 1H, H-2). MS m/z 192.0 (100%, [M + H]+), 213.9
297 (85.37%, [M + Na]+), 189.9 (100%, [M-H]-). HRMS m/z calcd for
298 C10H9NNaO3 214.0475, found 214.0466 [M + Na]+. λmax(H2O)/nm
299 392 (ε/M−1 cm−1 4205). λmax(EtOH)/nm 411 (ε/M−1 cm−1 4493).
tons, although the correlation peaks were not strong (Fig. S.7).
361
300 2.4.8. 4-(2-Formamidophenyl)-4-oxo-2-butenoic acid (NFK-CKA)
3.1.2. MS fragmentation studies (confirming position of the formyl group on 362
PIP-THQ) 363
301
302
303
NFK (65.7 mg, 0.316 mmol) was heated with PIP (1.78 mL, 18 mmol)
in 81% (v/v) 2-propanol in MQW (20 mL) at 52 °C for 55 min. Subsequent-
ly, the reaction was cooled on ice and rapidly neutralised to ~pH 7
To confirm the position of the formyl group in the structure of PIP- 364
THQ, mass fragmentation studies of the compound were conducted 365
(Fig. 1). The fragmentation spectrum of [12C-formyl]-PIP-THQ (m/z 366
303.1) shows a prominent ion at m/z 112.2 likely corresponding to the 367
N-formyl-2,3,4,5-tetrahydropyridin-1-ium fragment (Fig. 1a). Further 368
fragmentation of this ion provided m/z 84.1, indicating a loss of the 369
formyl group (Δ m/z 28), and m/z 56.1, 42.1 and 28.1 (Δ m/z 14) typical 370
for PIP fragmentation (Fig. 1b) [29]. To confirm that m/z 112.2 is indeed 371
N-formyl-2,3,4,5-tetrahydropyridin-1-ium, we synthesised NFK using 372
13C labelled formic acid and reacted it with PIP. This yielded a mixture 373
of [13C-formyl]-PIP-THQ and [12C-formyl]-PIP-THQ in a ratio of 1.73:1 374
according to MS. Fragmentation of [13C-formyl]-PIP-THQ is expected 375
to yield M + 1 ion (m/z 113), and indeed the fragmentation spectrum 376
of [13C-formyl]-PIP-THQ (m/z 304.3) shows a dominant ion at m/z 377
113.1 (Fig. 1c), consistent with the formyl group in PIP-THQ being at- 378
tached to PIP. Other ions produced were the same as those observed 379
in the fragmentation spectrum of [12C-formyl]-PIP-THQ, consistent 380
with their formation following loss of the labelled formyl group 381
304 by HCl and evaporated to dryness. The residue was dissolved in
305 ammonium bicarbonate (0.05 M, pH 7), loaded onto a Strata C18-E car-
306 tridge (5 g/20 mL) and eluted using MeCN/MQW mixtures containing
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
ammonium bicarbonate (0.03 M, pH 7). Fractions containing NFK-CKA
adduct were put into a freezer (−20 °C). Next day, isolated NFK-CKA
was left at room temperature for N3 h, acidified with HCl to pH 2 and ex-
tracted with 2 volumes of EtOAc. The EtOAc extract was washed with sat-
urated sodium chloride solution and evaporated. Dry material was
resuspended in MeOH, sonicated, and centrifuged for 5 min at 14,000 g.
The supernatant, which contained impurities and a small amount of
NFK-CKA was removed and the pellet was dried at high vacuum to afford
NFK-CKA as a pale yellow powder (12.55 mg, 19.1%). Compound was
unstable in MS but appeared pure from NMR and HPLC. Purity was 98%.
1H NMR δ (400 MHz, DMSO-d6,298 K) 13.20 (br, 1H, COOH), 10.82
(s, 1H, NHCHO), 8.38 (s, 1H, CHO), 8.14 (d, J = 7.8 Hz, 1H, H-6′), 7.87
(d, J = 7.4 Hz, 1H, H-3′), 7.64 (dd, J = 7.4, 7.3 Hz, 1H, H-4′), 7.62
(d, J = 15.6 Hz, 1H, H-3), 7.29 (dd, J = 7.8, 7.3Hz, 1H, H-5′), 6.56
(d, J = 15.6 Hz, 1H, H-2). HRMS m/z calcd for C11H9NNaO4 242.0424,
found 242.0425 [M + Na]+. λmax(H2O)/nm 322 (ε/M−1 cm−1 1584).
(Fig. 1b,d).
382
3.1.3. Fluorescence titration and absorption spectrum of PIP-THQ
383
λ
max(THF)/nm 285 and 350 (ε/M−1 cm−1 4355 and 2596).
The fluorescence titration of PIP-THQ showed that pH between 3 384
and 11.5 did not affect PIP-THQ fluorescence intensity but extremely 385
acidic (b1) or basic (N13) pH destroys the signal (Fig. 2). The decrease 386
in fluorescence of PIP-THQ is likely caused either by ionisation of the 387
quinolone nitrogen or carboxylic acid groups in PIP-THQ (pKa 1.78 388
and 12.56, respectively) or decomposition at pH extremes. The absorp- 389
tion spectrum of PIP-THQ shows distinct bands at 400 nm and 235 nm, 390
and a shoulder at 260 nm (see Fig. 3a—inset), and differs to those of 391
324 3. Results and discussion
325 3.1. Structure and properties of PIP-THQ fluorophore
326 3.1.1. NMR studies
327
NFK was synthesised by formylation of KYN using a modification of a
328 published method [27]. It was then reacted with an excess of PIP at 60 °C
previously reported KYN-related compounds.
392
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
5
and PIP (Fig. 3b). The product in the m/z 277.4 peak was isolated 408
(see Section 2.4.6) and confirmed to be PIP-KYN (Scheme 2). The results 409
suggest the formyl group is necessary for fluorophore formation, and 410
KYN or its adduct PIP-KYN appear not to be intermediates in this 411
process.
412
The unsaturated amino acid KYN-CKA has been shown to form 413
adducts with nucleophilic protein residues and glutathione [6,10], 414
therefore we investigated the reaction of 2 mM KYN-CKA or NFK-CKA 415
with 1 M PIP heated at 65 °C for 20 min. The reaction of KYN-CKA and 416
PIP showed only one dominant product (PIP-KYN) establishing that 417
KYN-related analogues are not involved in fluorophore formation, and 418
the hydrolysis of NFK into KYN on reaction with PIP is only a side reac- 419
tion. On the other hand, reaction of NFK-CKA and PIP gave a comparable 420
chromatogram as in Fig. 3a, consistent with deamination of NFK as a 421
first step in fluorophore formation. When an aliquot of NFK-CKA dis- 422
solved in DMSO was mixed with PIP (1 M) at room temperature, the 423
product corresponding to PIP-NFK (MW 304 Da, absorption spectrum 424
comparable to NFK) was formed rapidly at ~80% purity. This strongly in- 425
dicated that the formation of PIP-NFK adduct is the second step in the 426
fluorophore formation from PIP (Scheme 2). Unfortunately, PIP-NFK 427
could not be isolated for structural studies because it decomposed 428
OOF
back to NFK-CKA during purification, suggesting that the reaction is re- 429
versible favouring NFK-CKA formation in the absence of PIP. However, 430
PIP-NFK converted to PIP-KYN when heated at 60 °C for 10 min at 431
pH 1 (Scheme 2) lending support to our proposed structure of PIP-NFK. 432
No reaction products were observed when TRP was heated with PIP 433
at 65 °C for 20 min, indicating that the γ-keto group on the amino acid 434
chain of KYN and NFK is critical for the addition of the amine.
435
During our studies, we noticed that N80% 2-propanol or DMSO in the 436
reaction of NFK and PIP or NFK-CKA and PIP suppressed the formation of 437
fluorophore PIP-THQ and PIP-KYN but not the PIP-NFK intermediate. 438
This suggests that the conversion of PIP-NFK to the fluorophore PIP- 439
THQ requires a high content of water.
440
3.3. Rate of formation and consumption of PIP-NFK and PIP-KYN and 441
PIP-THQ 442
The kinetics of PIP-NFK formation is very fast compared to that of 443
PIP-KYN formation (Fig. 4). When NFK (2 mM) was incubated with 444
PIP (1 M) in MQW at 65 °C, the maximal amount of PIP-NFK was formed 445
after 3 min and then decreased exponentially (Fig. 4a). In contrast, a 446
30 min incubation of KYN (5 mM) and PIP (1 M) at 80 °C is necessary 447
to reach maximal formation of PIP-KYN (Fig. 4b). Rates of PIP-KYN 448
formation (kobs = 8.2 × 10−2 min−1) and PIP-NFK consumption 449
(kobs = 4.5 × 10−2 min−1) are comparable. As expected, KYN decom- 450
poses at about 10-fold slower rate (kobs = 2.9 × 10−2 min−1) than 451
NFK (kobs = 3.4 × 10−1 min−1) when incubated with PIP, consistent 452
with the higher reactivity of NFK to KYN. Whilst PIP-NFK decreases 453
Scheme 2. Proposed pathway of NFK and KYN reactions with PIP. Step d) is not based on
experimental observation. Note that PIP-KYN does not undergo any further modification
when incubated with PIP. Percent yield of purified solid is indicated in parentheses.
at
a nearly linear rate, PIP-THQ forms exponentially (kobs = 454
1.3 × 10−1 min−1) and reaches a plateau before PIP-NFK is depleted. 455
This suggests two things. Firstly, PIP-THQ is decomposing as the 456
reaction progresses. Secondly, PIP-THQ is not formed directly from 457
PIP-NFK. The missing step is likely a rearrangement and transamidation 458
393 3.2. Reaction pathway leading to the formation of PIP-THQ
394
Considering that PIP-THQ is produced in a moderately heated
395 aqueous reaction with no obvious catalysts, the N-formylation and 2-
(Scheme 2).
459
UNCORRECTED PR
396 substitution of PIP to form PIP-THQ raise an intriguing question as to
397 the reaction mechanism involved. The UV trace from the LC-MS chro-
398 matogram of the reaction products of NFK (2 mM) and PIP (1 M) at
399 pH 12.56 shows a large number of compounds and some were identi-
400 fied (Fig. 3a). The most intense peaks correspond to the PIP-THQ
401 fluorophore eluting at Rt 9.6 min. Molecular ions [M + H]+ at m/z
402 305.2 and m/z 277.4 coeluting at Rt 5.7 and 5.9 min (peak area ratio
403 1:3.4) observed in the spectrum were tentatively assigned to PIP-NFK
404 and PIP-KYN, respectively, whilst KYN eluted at Rt 2.8 min. KYN is a sub-
405 stantial product of the PIP and NFK reaction and could potentially be the
406 intermediate for fluorophore formation. However, only m/z 277.4 and
407 KYN were observed in the chromatogram from the reaction of KYN
Whilst the optimal reaction temperature and time for maximum for- 460
mation of PIP-THQ are 65 °C for 20 min, we also observed that PIP-THQ 461
could form at 55 °C, although at 3× slower rate [25] indicating that the 462
higher temperature (65 °C) is not essential for the PIP-THQ formation. 463
3.4. Effect of PIP and NFK concentration on the formation of fluorophore 464
PIP-THQ
465
The effect of varying concentrations of PIP on fluorophore formation 466
was investigated. When a fixed concentration of PIP (1 M) was incubat- 467
ed with different concentrations of NFK (0.5–9 mM), PIP-THQ formation 468
was initially nonlinear in relationship to NFK concentration and reached 469
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

6
P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
Fig. 1. Mass fragmentation of PIP-THQ isotopes in positive electrospray ionisation mode and triple quadrupole. Panel shows fragmentation spectra of ion a) m/z 304.3 ([13C-formyl] PIP-
THQ), b) m/z 113.1, c) m/z 303.1 ([12C-formyl] PIP-THQ) and d) m/z 112.1. Voltage applied to collision cell was 10 V (a,c) and 20 V (b,d), respectively. Asterisk marks the ion being
fragmented.
470 a plateau at 9 mM NFK (Fig. 5a). When the concentration of NFK was
471 kept constant (2 mM) and reacted with varying concentrations of PIP
472 (0.4, 1 and 3 M) (Fig. 5b), PIP-THQ formation at both 3 M and 0.4 M
473 PIP was approximately half of that formed at 1 M PIP. The large molar
474 excess (~500-fold) of PIP to NFK is necessary for efficient fluorophore
475 formation however, concentrations of PIP higher than 1 M appear to de-
476 crease PIP-THQ formation possibly due to decomposition caused by
477 high pH.
connected (Fig. S.9a). This showed that the fluorophore formation is 493
not significantly affected by the size of the amine ring.
494
3.5.2. Monomethyl-substituted PIP and acyclic amines
495
Fluorophore formation from the reaction of NFK with isomeric 496
methyl substituted PIPs was investigated (Table 1). When the methyl- 497
substituent was located on the nitrogen of PIP (NMePIP), the mass 498
and Rt of the formed fluorophore (NMePIP-THQ) were identical to 499
that of PIP-THQ (MW 302, Rt 9.5 min), but 2.5-fold lower yields 500
were obtained. A similar pattern was also observed comparing 501
fluorophore formation between PYR and NMePYR. This indicates that 502
the N-methyl group of the tertiary amine had to undergo demethyla- 503
tion. Much lower yields of fluorophore were obtained using 2- 504
methylpiperidine (2-MePIP), cis-2,6-dimethylpiperidine (DMP) and 505
2,2,6,6-tetramethylpiperidine (TMP) (4%, 3% and 0%, respectively, com- 506
pared to PIP). Furthermore, the mass of the fluorophore DMP-THQ (MW 507
316) formed from DMP was identical to fluorophores formed from 508
monomethylated amines (3-MePIP-THQ and 4-MePIP-THQ) (Table 1). 509
DMP forms both amine-KYN and amine-NFK adducts with slight de- 510
creases in efficiency compared to PIP (29% and 36% respectively). This 511
suggested that the commercial DMP used in the study (Sigma-Aldrich, 512
478
In contrast, the reaction of KYN and PIP does not require a high molar
479 excess of PIP. When KYN (20 mM) was heated at 80 °C for 10 min with
480 either a 50-fold (1 M) or 10-fold (0.2 M) excess of PIP, the yield of PIP-
481 KYN by HPLC was comparable.
482 3.5. Formation of fluorophores from amines other than PIP
483 3.5.1. Cyclic amines
484
We compared fluorophore formation when NFK was reacted with 5-
485 or 7-membered cyclic amines to that with 6-membered ring PIP.
486 Fluorophore formation with 5-membered ring pyrrolidine (PYR), or
487 with 7-membered ring azepane (AZP), was decreased 12% and 10% re-
488 spectively, compared to PIP (Table 1). The mass and Rt of the
489 fluorophores AZP-THQ (MW 316, Rt 10.7 min) and PYR-THQ (MW
490 288, Rt 8 min) were as expected in relation to the mass of the ring
491 (Table 1). The fragmentation pattern of PYR-THQ resembled that of
492 PIP-THQ indicating that the amine ring was similarly formylated and
Fig. 3. HPLC chromatogram of reaction of a) NFK (2 mM) or b) KYN (2 mM) with PIP (1 M)
incubated at 65 °C for 20 min, cooled on ice and analysed on HPLC-MS. Detection wave-
length was 254 nm. Compound name and retention time (Rt) are indicated. Absorption
spectrum of PIP-THQ is shown in the inset. Dashed line denotes the separation of PIP-
KYN and PIP-NFK peaks for quantification purposes.
Fig. 2. Fluorescence titration of 15 μM PIP-THQ measured in 0.1 M KCl at 25 °C. pH was
adjusted with KOH and HCl. Excitation and emission wavelengths were 400 and
500 nm, respectively. Data were fitted with bell shaped curve which afforded values at
inflection points.
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
7
Fig. 4. Reaction progress of a) 2 mM NFK and 1 M PIP at 65 °C and b) 5 mM KYN and 1 M PIP at 80 °C. Each time point aliquot was analysed by HPLC and peak areas of corresponding
compounds were quantified from 400 nm (PIP-THQ), 360 nm (PIP-KYN and KYN), and 325 nm (PIP-NFK and NFK) chromatograms. Data points were fitted with one-phase exponential
curves. Data point of PIP-KYN at 40 min in panel b) was not included in the fit.
513 98%) might have been contaminated with isomers of monomethyl PIP.
514 Indeed, comparing 13C NMR spectra of DMP, 2-Me, 3-Me and 4-Me
515 PIP showed that DMP contains about 1% of 3-MePIP but other impurities
516 were also present. In Section 3.4, we showed that formation of PIP-KYN
517 adducts does not require a high molar excess of PIP, in contrast to
518 formation of fluorophores. This strongly indicates that the adducts ob-
519 served in the reaction of DMP and NFK are 3-MePIP adducts and that
520 DMP itself is only marginally reactive. It also corroborates the require-
521 ment for a high molar excess of PIP (and most likely also other amines)
522 to NFK for the fluorophore formation.
2) Only cyclic amines without substitutions at both carbons adjacent to 543
nitrogen (2- and 6-position on PIP ring) appear to be able to form 544
fluorophores efficiently. One position next to nitrogen is required 545
for bond formation during the rearrangement step, but why both po- 546
OOF
sitions next to nitrogen need to be unsubstituted is not yet clear. As 547
suggested in point 1), a methyl group adjacent to nitrogen might 548
sterically hinder the interaction of the formyl group with the 549
amine nitrogen.
550
3) 3- and 4-methyl substituents on PIP enhance fluorophore formation. 551
523
Formation of the amine-THQ fluorophore from 3-MePIP and 4-
3.6. Reactions of amines with KYN
552
524 MePIP was increased 1.84-fold and 1.38-fold respectively compared to
525 PIP. Acyclic amines diethylamine (DEA) and dipropylamine (DPA)
526 formed THQ products in much lower quantities than with PIP.
In a final aspect of our investigations, we examined reactions of 553
amines with KYN, an NFK analogue without the formamide group. 554
Chromatograms of the reaction of KYN with amines showed only 555
unreacted KYN and amine-KYN adduct(s) (Table 1). NMePIP-KYN and 556
DMP-KYN adducts also showed a loss of the methyl substituent, similar 557
to their respective fluorophores NMePIP-THQ and DMP-THQ formed 558
from NFK (Table 1). 4-MePIP did not significantly increase the yield of 559
amine-KYN adduct relative to PIP. This contrasts with the increased 560
yields of amine-NFK and amine-THQ that were obtained in the reaction 561
of NFK with 4-MePIP and 3-MePIP compared to PIP. Thus, a 4-methyl 562
substituent on the PIP skeleton is favoured during fluorophore forma- 563
527
The fluorescence emission maxima of PIP-THQ and the other amine-
528 THQ fluorophores were comparable (Table 1), showing that the
529 structure of amine has negligible effect on the electronic properties of
530 the fluorophores. We purified 3-MePIP-THQ and confirmed using
531 mass fragmentation (Fig. S.9b) and NMR (Fig. S.10–S.12), that it was in-
532 deed an analogue of PIP-THQ, differing only in the attached amine.
533 Moreover, comparable extinction coefficients of 3-MePIP-THQ
534 (ε402 = 2431 M−1 cm−1) and PIP-THQ (ε400 = 2136 M−1 cm−1
)
535 strongly indicates the yields in Table 1 obtained from peak areas of
536 400 nm chromatogram are genuine, and not a result of different extinc-
537 tion coefficients.
tion but has a negligible effect on amine-KYN adduct formation.
564
Acyclic secondary amines DEA and DPA formed DEA-KYN and DPA- 565
KYN adducts with 86% and 39% efficiency, respectively, compared to 566
PIP-KYN. This again contrasts with NFK reactions where the acyclic 567
amines did not form adducts apart from small amounts of fluorophore 568
amine-THQ. Moreover, acyclic amines DPA and DEA produced two dif- 569
ferent KYN adducts; the more abundant adduct having apparently lost 570
one of the alkyl chains, whilst the lower yielding adduct did not. In con- 571
trast, reactions of cyclic amines produced only a single amine-KYN ad- 572
duct with exception of 2-MePIP which also formed two products but 573
with identical mass (MW 290, Table 1). The two 2-MePIP-KYN adducts 574
538 3.5.3. Summary of reactions of amines with NFK
539 1) Cyclic amines form fluorophores more efficiently than acyclic
540
541
542
amines of comparable size and basicity. This suggests that the rear-
rangement and transamidation step requires a finely tuned spatial
arrangement that does not favour acyclic structures.
are most likely diastereomers.
575
4. Concluding remarks
UNCORRECTED PR
576
The reaction of NFK and cyclic amines generates tetrahydroquinolone- 577
based fluorophores where the amine is N-formylated and attached 578
through its 2-position. Substitutions on the 2-position of piperidine 579
[30–34] and transamidations [35–39] in general require high tempera- 580
tures or catalysts. In contrast, our newly-described reaction does not re- 581
quire catalysts and occurs in aqueous solution under mild conditions. 582
Fluorophore formation is driven by a high molar excess of cyclic amine 583
relative to NFK which provides the basic environment for initial NFK de- 584
amination and the subsequent formation of a transient adduct, PIP-NFK. 585
This is then followed by amine-induced transamidation of the formyl 586
group from NFK to the nitrogen of the amine, adduct rearrangement 587
Fig. 5. Formation of PIP-THQ in reaction of a) variable NFK and 1 M PIP and b) 2 mM NFK
and variable PIP. Reactions were initiated by addition of PIP into aqueous solution of NFK,
incubated at 65 °C for 20 min, chilled on ice and analysed on HPLC. Peak area correspond-
ing to PIP-THQ was quantified from 400 nm chromatogram.
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

8
P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
t1:1
t1:2
t1:3
t1:4
Table 1
We have been using the formation of PIP-THQ fluorophore in 598
our most current and most sensitive IDO1 enzymatic assay [25], but 599
the studies here suggest that using 3-MePIP instead of PIP might in- 600
crease the sensitivity of this assay even more. Isolation of remaining 601
fluorophore intermediates and determination of reaction kinetics and 602
reaction orders in future experiments would aid in identifying the 603
Molecular weights (MW), retention times (Rt) and amount of amine-KYN, amine-NFK,
amine-THQ adducts formed in reaction of amines with NFK (red font), and the amine-
KYN adduct formed in the reaction of amines with KYN (blue font).
final steps of this reaction mechanism.
604
Conflicts of interest
605
606
amine–KYN
Peak area (%)
amine–NFK
Peak
amine–THQ
b
Peak
MW
R
MW
R
MW
R
λ
em
(nm)
t
t
t
area
area
d
NFK
c
KYN
c
Authors do not have any conflicts of interest to disclose.
(Da) (min)
(Da) (min)
b
b
(Da) (min)
(%)
(%)
rxn
rxn
100*
(100%)
276
276
5.9
5.9
100*
16
100*
70
304
304
5.7
5.7
100
102
0
302
302
316
316
9.6
9.6
486
Transparency Document
607
PIP
The Transparency document associated with this article can be 608
found, in the version.
39
4
486
N/O
485
609
NMePIP
2–MePIP
290
290
6.3
6.8
0
39
17
N/O
10.6
10.9
48
Acknowledgements
610
184*
(190%)
290
290
290
7.1
7.1
7.1
94
114
71
N/T
113
82
318
318
318
6.8
6.8
6.8
272
This work was supported by grants from the Auckland Cancer 6Q131
Society and the University of Auckland doctoral scholarship to P.T.
3–MePIP
612
275
64
316
316
10.9
10.9
138
3
484
484
Appendix A. Supplementary data
613
4–MePIP
DMPe
Supplementary data to this article can be found online at http://dx. 614
doi.org/10.1016/j.bbagen.2015.04.007.
615
N/O
0
N/T
N/O
0
N/O
0
N/O
TMP (40%
References
616
2–propanol)
a
a
[1] Y. Chen, G.J. Guillemin, Kynurenine pathway metabolites in humans: disease and 617
healthy states, Int. J. Tryptophan Res. 2 (2009) 1–19. 618
[2] G. Costantino, New promises for manipulation of kynurenine pathway in cancer and 619
neurological diseases, Expert Opin. Ther. Targets 13 (2009) 247–258. 620
[3] M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Heme-containing oxygenases, Chem. 621
Rev. 96 (1996) 2841–2888. 622
[4] A.H. Mehler, W.E. Knox, The conversion of tryptophan to kynurenine in liver: II. The 623
enzymatic hydrolysis of formylkynurenine, J. Biol. Chem. 187 (1950) 431–438. 624
[5] L.M. Taylor, J. Andrew Aquilina, J.F. Jamie, R.J.W. Truscott, UV filter instability: conse- 625
quences for the human lens, Exp. Eye Res. 75 (2002) 165–175. 626
262
4.9
190
N/T
N/T
N/T
290
N/O
190
288
288
8
8
88
35
488
486
PYR
N/O
0
0
0
0
NMePYR
N/O
236
N/O
N/O
316
290
10.7
9
90
10
483
485
AZP
DEA
4.1
5.5
0
0
59
27
0
0
264
250
[6] B. Garner, D.C. Shaw, R.A. Lindner, J.A. Carver, R.J.W. Truscott, Non-oxidative modifi- 627
cation of lens crystallins by kynurenine: a novel post-translational protein modifica- 628
tion with possible relevance to ageing and cataract, Biochim. Biophys. Acta Protein 629
5.2
7.8
0
0
30
9
0
0
N/O
318
11
1
483
DPA (40%
2–propanol)
292
Struct. Mol. Enzymol. 1476 (2000) 265–278.
[7] N.R. Parker, A. Korlimbinis, J.F. Jamie, M.J. Davies, R.J.W. Truscott, Reversible binding 631
of kynurenine to lens proteins: potential protection by glutathione in young lenses, 632
Invest. Ophthalmol. Vis. Sci. 48 (2007) 3705–3713.
[8] Y.P. Tsentalovich, O.A. Snytnikova, M.D.E. Forbes, E.I. Chernyak, S.V. Morozov, 634
Photochemical and thermal reactivity of kynurenine, Exp. Eye Res. 83 (2006) 635
630
N/O = not observed; N/T = not tested. Molecular weights were determined by presence
of [M + H]+ and [M − H]− ions in mass spectra of respective compounds at given reten-
tion time (Rt). Reactions of TMP and DPA were performed in 40% 2-propanol due to their
poor water solubility.
t1:6
t1:7
t1:8
633
t1:10
a PYR-KYN and PYR-NFK adducts coeluted in a single peak, and the value represents their
combined peak areas.
t1:9
1439–1445.
636
t1:11
t1:12
t1:13
t1:14
t1:15
t1:16
t1:17
t1:18
t1:19
t1:20
[9] L.V. Kopylova, O.A. Snytnikova, E.I. Chernyak, S.V. Morozov, M.D.E. Forbes, Y.P. 637
Tsentalovich, Kinetics and mechanism of thermal decomposition of kynurenines 638
and biomolecular conjugates: ramifications for the modification of mammalian 639
eye lens proteins, Org. Biomol. Chem. 7 (2009) 2958–2966.
[10] P.T. Yu, O.A. Snytnikova, R.Z. Sagdeev, Photochemical and thermal reactions of 641
kynurenines, Russ. Chem. Rev. 77 (2008) 789. 642
[11] A.L. Mellor, D.H. Munn, IDO expression by dendritic cells: tolerance and tryptophan 643
catabolism, Nat. Rev. Immunol. 4 (2004) 762–774. 644
b Peak areas expressed relative to PIP reactions (100%) in each respective column. The per-
cent value in parentheses denotes the yield of purified solid relative to PIP-THQ.
c These two columns show peak areas of amine-KYN adducts observed in reactions of NFK
(NFK rxn) and KYN (KYN rxn) with amines.
640
d
The fluorescence emission maximum at 400 nm excitation, acquired during HPLC
analysis.
e The products in reaction of DMP with amino acids are probably formed from impurities
in the DMP reagent.
[12] G.K. Lee, H.J. Park, M. Macleod, P. Chandler, D.H. Munn, A.L. Mellor, Tryptophan dep- 645
rivation sensitizes activated T cells to apoptosis prior to cell division, Immunology 646
⁎
Asterisk denotes that the compound was purified and isolated.
107 (2002) 452–460.
647
[13] F. Fallarino, U. Grohmann, S. You, B.C. McGrath, D.R. Cavener, C. Vacca, C. Orabona, R. 648
Bianchi, M.L. Belladonna, C. Volpi, P. Santamaria, M.C. Fioretti, P. Puccetti, The com- 649
bined effects of tryptophan starvation and tryptophan catabolites down-regulate T 650
cell receptor ζ-chain and induce a regulatory phenotype in naive T cells, J. Immunol. 651
588
589
590
591
592
593
594
595
and cyclisation which requires a high content of water. The sequence of
the final steps is not clear, but cyclisation is likely to be a rate-limiting
step as thermal decomposition of KYN to form KNY (Scheme 1) proceeds
at a slow rate [8]. The formamide group of NFK is essential for fluorophore
formation, and efficient fluorophore formation is limited to cyclic amines
that are unsubstituted on both positions adjacent to the nitrogen of the
amine.
176 (2006) 6752–6761.
652
[14] Y. Suzuki, T. Suda, K. Furuhashi, M. Suzuki, M. Fujie, D. Hahimoto, Y. Nakamura, N. 653
Inui, H. Nakamura, K. Chida, Increased serum kynurenine/tryptophan ratio corre- 654
lates with disease progression in lung cancer, Lung Cancer 67 (2010) 361–365.
655
[15] D.H. Munn, M.D. Sharma, D. Hou, B. Baban, J.R. Lee, S.J. Antonia, J.L. Messina, P. 656
Chandler, P.A. Koni, A.L. Mellor, Expression of indoleamine 2,3-dioxygenase by 657
plasmacytoid dendritic cells in tumor-draining lymph nodes, J. Clin. Invest. 114 658
(2004) 280–290.
659
[16] K. Ino, N. Yoshida, H. Kajiyama, K. Shibata, E. Yamamoto, K. Kidokoro, N. Takahashi, 660
M. Terauchi, A. Nawa, S. Nomura, T. Nagasaka, O. Takikawa, F. Kikkawa, Indoleamine 661
2,3-dioxygenase is a novel prognostic indicator for endometrial cancer, Br. J. Cancer 662
We have demonstrated that NFK is a highly reactive TRP catabolite
596 indicating that it could form yet to be identified biologically relevant
597 metabolites warranting further study.
95 (2006) 1555–1561.
663
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007

P. Tomek et al. / Biochimica et Biophysica Acta xxx (2015) xxx–xxx
9
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
[17] M. Hoshi, H. Ito, H. Fujigaki, M. Takemura, T. Takahashi, E. Tomita, M. Ohyama, R.
Tanaka, K. Saito, M. Seishima, Indoleamine 2,3-dioxygenase is highly expressed in
human adult T-cell leukemia/lymphoma and chemotherapy changes tryptophan ca-
tabolism in serum and reduced activity, Leuk. Res. 33 (2009) 39–45.
[18] C. Uyttenhove, L. Pilotte, I. Théate, V. Stroobant, D. Colau, N. Parmentier, T. Boon, B.J.
Van den Eynde, Evidence for a tumoral immune resistance mechanism based on
tryptophan degradation by indoleamine 2,3-dioxygenase, Nat. Med. 9 (2003)
1269–1274.
[27] C.E. Dalgliesh, The synthesis of N′-formyl-DL-kynurenine, Nα-acetyl- DL- 701
kynurenine and belated compounds, and observations on the synthesis of 702
kynurenine, J. Chem. Soc. (1952) 137–141 (Resumed).
703
[28] R. Dick, B.P. Murray, M.J. Reid, M.A. Correia, Structure–function relationships of rat 704
hepatic tryptophan 2,3-dioxygenase: identification of the putative heme-ligating 705
histidine residues, Arch. Biochem. Biophys. 392 (2001) 71–78.
[29] L.W. Daasch, High resolution mass spectrum of piperidine, J. Phys. Chem. 69 (1965) 707
3196-3196. 708
706
[19] E. Dolušić, R. Frédérick, Indoleamine 2,3-dioxygenase inhibitors: a patent review
[30] O. Calvez, A. Chiaroni, N. Langlois, Enantioselective synthesis of 2,3-disubstituted pi- 709
peridines from (S)-methylpyroglutamate, Tetrahedron Lett. 39 (1998) 9447–9450. 710
[31] J. Lee, T. Hoang, S. Lewis, S.A. Weissman, D. Askin, R.P. Volante, P.J. Reider, Asymmetric 711
synthesis of (2S,3S)-3-hydroxy-2-phenylpiperidine via ring expansion, Tetrahedron 712
(2008–2012), Expert Opin. Ther. Pat. 23 (2013) 1367–1381.
[20] X. Zheng, J. Koropatnick, M. Li, X. Zhang, F. Ling, X. Ren, X. Hao, H. Sun, C. Vladau, J.A.
Franek, B. Feng, B.L. Urquhart, R. Zhong, D.J. Freeman, B. Garcia, W.-P. Min,
Reinstalling antitumor immunity by inhibiting tumor-derived immunosuppressive
molecule IDO through RNA interference, J. Immunol. 177 (2006) 5639–5646.
[21] X. Liu, N. Shin, H.K. Koblish, G. Yang, Q. Wang, K. Wang, L. Leffet, M.J. Hansbury, B.
Thomas, M. Rupar, P. Waeltz, K.J. Bowman, P. Polam, R.B. Sparks, E.W. Yue, Y. Li, R.
Wynn, J.S. Fridman, T.C. Burn, A.P. Combs, R.C. Newton, P.A. Scherle, Selective inhi-
bition of IDO1 effectively regulates mediators of antitumor immunity, Blood 115
(2010) 3520–3530.
[22] M.R. Mautino, F.A. Jaipuri, J. Waldo, S. Kumar, J. Adams, C. Van Allen, A. Marcinowicz-
Flick, D.H. Munn, N. Vahanian, C.J. Link, NLG919, a novel indoleamine-2,3-
dioxygenase (IDO)-pathway inhibitor drug candidate for cancer therapy. [abstract],
Proceedings of the 104th Annual Meeting of the American Association for Cancer
Research; 2013 Apr 6–10, Cancer Res, 73(8 Suppl), AACR, Washington, DC. Philadel-
phia (PA), 2013http://dx.doi.org/10.1158/1538-7445.AM2013-491 (Abstract nr
491, (2013)).
[23] O. Takikawa, T. Kuroiwa, F. Yamazaki, R. Kido, Mechanism of interferon-gamma
action—characterization of indoleamine 2,3-dioxygenase in cultured human-cells
induced by interferon-gamma and evaluation of the enzyme-mediated tryptophan
degradation in its anticellular activity, J. Biol. Chem. 263 (1988) 2041–2048.
[24] A. Matin, I.M. Streete, I.M. Jamie, R.J.W. Truscott, J.F. Jamie, A fluorescence-based
assay for indoleamine 2,3-dioxygenase, Anal. Biochem. 349 (2006) 96–102.
[25] P. Tomek, B. Palmer, J. Flanagan, S.-P. Fung, D.A. Bridewell, J. Jamie, L.-M. Ching,
Formation of an N-formylkynurenine-derived fluorophore and its use for measuring
indoleamine 2,3-dioxygenase 1 activity, Anal. Bioanal. Chem. 405 (2013) 2515–2524.
[26] A. Pirie, Formation of N′-formylkynurenine in proteins from lens and other sources
by exposure to sunlight, Biochem. J. 125 (1971) 203–208.
Lett. 42 (2001) 6223–6225.
713
[32] Y. Matsumura, D. Minato, O. Onomura, Highly enantioselective introduction of 714
bis(alkoxycarbonyl)methyl group into the 2-position of piperidine skeleton, 715
J. Organomet. Chem. 692 (2007) 654–663.
716
[33] G. Pelletier, L. Constantineau-Forget, A.B. Charette, Directed functionalization of 1,2- 717
dihydropyridines: stereoselective synthesis of 2,6-disubstituted piperidines, Chem. 718
Commun. 50 (2014) 6883–6885.
719
[34] L. Zhou, D.W. Tay, J. Chen, G.Y.C. Leung, Y.-Y. Yeung, Enantioselective synthesis of 2- 720
substituted and 3-substituted piperidines through a bromoaminocyclization pro- 721
cess, Chem. Commun. 49 (2013) 4412–4414.
722
[35] N.A. Stephenson, J. Zhu, S.H. Gellman, S.S. Stahl, Catalytic transamidation reactions 723
compatible with tertiary amide metathesis under ambient conditions, J. Am. 724
Chem. Soc. 131 (2009) 10003–10008.
725
[36] C.L. Allen, B.N. Atkinson, J.M.J. Williams, Transamidation of primary amides with 726
amines using hydroxylamine hydrochloride as an inorganic catalyst, Angew. 727
Chem. Int. Ed. 51 (2012) 1383–1386.
[37] T.B. Nguyen, J. Sorres, M.Q. Tran, L. Ermolenko, A. Al-Mourabit, Boric acid: a highly 729
efficient catalyst for transamidation of carboxamides with amines, Org. Lett. 14 730
728
OOF
(2012) 3202–3205.
731
[38] S.P. Pathare, A.K.H. Jain, K.G. Akamanchi, Sulfated tungstate: a highly efficient cata- 732
lyst for transamidation of carboxamides with amines, RSC Adv. 3 (2013) 7697–7703. 733
[39] L. Becerra-Figueroa, A. Ojeda-Porras, D. Gamba-Sánchez, Transamidation of
carboxamides catalyzed by Fe(III) and water, J. Org. Chem. 79 (2014) 4544–4552.
UNCORRECTED PR
Please cite this article as: P. Tomek, et al., Formation of fluorophores from the kynurenine pathway metabolite N-formylkynurenine and cyclic
amines involves transamidation and carbon–carbon bond..., Biochim. Biophys. Acta (2015), http://dx.doi.org/10.1016/j.bbagen.2015.04.007