Mechanistic Evaluation of Bioorthogonal Decaging with trans-Cyclooctene: The Effect of Fluorine Substituents on Aryl Azide Reactivity and Decaging from the 1,2,3-Triazoline

Article abstract of DOI:10.1021/acs.bioconjchem.7b00665

Bioorthogonal prodrug activation/decaging strategies need to be selective, rapid and release the drug from the masking group upon activation. The rates of the 1,3-dipolar cycloaddition between a trans-cyclooctene (TCO) and a series of fluorine-substituted

Full text of DOI:10.1021/acs.bioconjchem.7b00665

Subscriber access provided by READING UNIV
Article
Mechanistic Evaluation of Bioorthogonal Decaging with
trans-Cyclooctene: The Effect of Fluorine Substituents on
Aryl Azide Reactivity and Decaging from the 1,2,3-Triazoline
Siddharth Sai Matikonda, Jessica M Fairhall, Franziska Fiedler, Suchaya
Sanhajariya, Robert A. J. Tucker, Sarah Hook, Anna L. Garden, and Allan B. Gamble
Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.7b00665 • Publication Date (Web): 12 Jan 2018
Downloaded from http://pubs.acs.org on January 14, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Bioconjugate Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 13
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
Mechanistic Evaluation of Bioorthogonal Decaging with trans-
Cyclooctene: The Effect of Fluorine Substituents on Aryl Az-
ide Reactivity and Decaging from the 1,2,3-Triazoline
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Siddharth S. Matikonda,† Jessica M. Fairhall,† Franziska Fiedler,† Suchaya Sanhajariya,† Robert A. J.
Tucker,‡ Sarah Hook,† Anna L. Garden,‡ Allan B. Gamble*,†
†School of Pharmacy, University of Otago, Dunedin, 9054, New Zealand
‡Department of Chemistry, University of Otago, Dunedin, 9054, New Zealand
ABSTRACT: Bioorthogonal prodrug activation/decaging strategies need to be selective, rapid and release the drug from the maskꢀ
ing group upon activation. The rates of the 1,3ꢀdipolar cycloaddition between a transꢀcyclooctene (TCO) and a series of fluorineꢀ
substituted azidoꢀPABC selfꢀimmolative spacers caging two model drugs, and subsequent release from the 1,2,3ꢀtriazoline are reꢀ
ported. As the number of fluorine substituents on the PABC linker increases from one to four, the rate of cycloaddition increases by
1
almost oneꢀorder of magnitude. Using a combination of fluorescence, H/19F NMR, and computational experiments we have been
able to determine how substituents on the PABC ring can influence the degradation rates and also the product distribution of the
1,2,3ꢀtriazoline. We have also been able to determine how these substituents influence the rate of imine hydrolysis and 1,6ꢀselfꢀ
immolation decaging rates of the generated anilines. The NMR and computational studies demonstrate that fluorine substituents on
the aromatic ring lower the transition state energy required for converting the triazoline to the imine or aziridine intermediates via
extrusion of diatomic nitrogen, and that in the case of a tetrafluoro substituted aromatic ring, it is the imine hydrolysis and 1,6ꢀselfꢀ
immolation that is rateꢀlimiting. This knowledge further enhances the understanding of factors which influence the stability of triaꢀ
zolines, and enables potential applications of fluorinated aromatics, in particular perfluorinated aromatics, in synthetic chemistry
and
sustainedꢀrelease
drug
delivery
systems.
the Staudinger reaction and early examples of strainꢀpromoted
azideꢀalkyne cycloadditions (SPAAC).18 However, these rates
are not sufficiently rapid to accomplish prodrug activation at
the low ꢀM concentration of TCO and prodrug that is exꢀ
pected at the tumor site during in vivo experiments. To achieve
better prodrug and proꢀprobe activation we require rates to be
at least oneꢀtoꢀtwo orders of magnitude faster than our current
system.19 The rate determining step for our strategy is the 1,3ꢀ
dipolar cycloaddition, so in designing faster reaction kinetics
there is scope to modify the TCO or the azidoꢀPABCꢀcaging
linker. Addition of fluorine substituents to the PABC linker,
for example as perfluoroaryl azides20 and tetrafluoroaryl azꢀ
ides21 results in faster click reactions. Therefore, we chose to
investigate the effect of fluorine substituents on our azidoꢀ
PABC caging linker (Figure 1). As proofꢀofꢀconcept, 7ꢀ
hydroxycoumarin 1 and doxorubicin 2 were selected as the
model probe and drug for our proꢀprobe 1aꢀ1f and prodrug 2a,
2b, 2d, and 2e series, enabling us to examine the rates for the
1,3ꢀdipolar cycloaddition and subsequent decaging from the
generated triazoline (five total steps); all of which contribute
to prodrug activation and decaging (Scheme 1).8
ꢀ INTRODUCTION
The last fifteen years has seen the rapidly emerging field
of bioorthogonal ligations, a class of chemical reactions that
can proceed unhindered in the physiological environment, take
a prominent role in chemical biology.1ꢀ6 Key to the success of
these strategies is a rapid reaction between the bioorthogonal
reagents. However, there must be a balance between reactivity
and biological stability (bioorthogonality) so that the ligation
can be selective and occur before the reaction components are
cleared from the biological system.7 While most research in
the field has focused on chemical biology applications; i.e.
tracking biomolecules and cells in living systems, more reꢀ
cently, these bioorthogonal chemical reactions have been reꢀ
ported to clickꢀandꢀrelease drugs and probes; i.e. to selectively
target and activate prodrugs and proꢀprobes.8ꢀ16 In this emergꢀ
ing subꢀfield, one of the challenges for in vivo applications is
the need for sufficiently deactivated prodrugs and proꢀprobes
that are rapidly ligated to the target and subsequently release
the active drug/probe on a biologically relevant timescale.17
We recently reported a bioorthogonal reaction to decage
chemicallyꢀmasked drugs via a strainꢀpromoted 1,3ꢀdipolar
cycloaddition between pꢀazidobenzyloxycarbonyl (PABC)ꢀ
functionalized prodrug/probe and transꢀcyclooctenol (TCO).8
The click reaction resulted in secondꢀorder rates of 0.017 Mꢀ1sꢀ
1 and 0.027 Mꢀ1sꢀ1 for the equatorial and axial isomers of TCO,
respectively, being at least oneꢀorder of magnitude faster than
Further mechanistic insight for drug release from the triaꢀ
zoline, determined experimentally and supported by DFT calꢀ
culations, are reported and build on previous mechanistic studꢀ
ies of 1,3ꢀdipolar cycloadditions between perfluorinated arꢀ
ylazides and enamines or strained dipolarophiles.20
Scheme 1. Proposed mechanistic pathway for bioorthogonal activation and degradation/decaging of the PABC linker. Rates
1-5 contribute to the overall decaging process, and the potential rate-determining steps of 1-4 are investigated in this study.
ACS Paragon Plus Environment

Bioconjugate Chemistry
Page 2 of 13
INACTIVE
Caged Group/Drug
Unstable 1,2,3-triazoline
1
2
3
H
R2
O
PABC
N
PABC
N
N
N
Rate 2
Rate 1
R1
N3
O
X
+
H
Triazoline
degradation
Bioorthogonal
click reaction
R3
OH
4
5
6
7
8
9
R4
Rate 3
Imine
hydrolysis
trans-Cyclooctenol
(TCO)
Caging PABC Linker
OH
OH
X= O, NH; R1-4 = H, F
ACTIVE
Decaged Group/Drug
R2
O
R1
O
X
Rate 5
Rate 4
O
X
X
Decarboxylation
H2N
R3
1,6-Immolation
(self-elimination)
-O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R4
X= O, NH; R1-4 = H, F
X= OH, NH2
O
OH
O
obtained when 8f was first converted to the corresponding
methyl ester 9f and reduced. 4ꢀaminoꢀ3ꢀfluorobenzyl alcohol
3b and 4ꢀaminoꢀ2,3,5,6ꢀtetrafluorobenzyl alcohol 3d were
synthsised by LiAlH4 reduction of the benzoic acids 8b and
8d, while 4ꢀaminoꢀ2ꢀfluorobenzyl alcohol 3c was synthesized
via reduction of the corresponding methyl ester 9c (Scheme 3,
Scheme S1).
O
Probe/Drug
R
OH
O
X
OH
O
N3
H
O
O
OH
O
HO
O
O
OH
NH2
X= NH; Doxorubicin 2
Prodrugs = 2a, 2b, 2d, 2e
X= O; 7-hydroxycoumarin 1
Pro-probes = 1a-1f
F
F
Scheme 2. Synthesis of pro-probes 1a-1f and prodrugs 2a,
2b, 2d, and 2e from the corresponding 4-aminobenzyl al-
cohols 3a-3f.
F
F
N3
N3
N3
F
N3
F
N3
N3
F
F
F
F
NO2
O
R
R
R
(1a, 2a)
(1b, 2b)
(1c)
(1d, 2d)
(1e, 2e)
(1f)
(ii)
(i)
OH
OH
O
O
Figure 1. Suite of the fluorinated PABC caging linkers.
H2N
N3
N3
According to Sustmann theory,22,
our ambiphilic arꢀ
23
3a-3f
4a-4f
5a-5f
ylazide is a type II 1,3ꢀdipole, therefore the presence of elecꢀ
tronꢀwithdrawing substituents on the 1,3ꢀdipole can reduce the
energy gap between the LUMO of the dipole and HOMO of
the dipolarophile resulting in a faster click reaction.21, 23, 24 The
negative inductive (ꢀI) effect of fluoroꢀsubstitution on the azꢀ
idoꢀPABC aromatic ring is hypothesized to increase the rate of
the 1,3ꢀdipolar cycloaddition. We also envisaged that the elecꢀ
tronꢀwithdrawing substituents on the aromatic ring would reꢀ
duce the nucleophilicity of the nitrogen atom, funnelling the
intermediate product towards the imine rather than the aziriꢀ
dine which requires nucleophilic attack of the nitrogen atom at
the carbon that extrudes diatomic nitrogen.25, 26 In addition, the
use of fluorine in medicinal chemistry has added benefits27, 28
of imparting lipophilicity and increasing the binding to serum
proteins,29 and decreasing metabolism of the aromatic ring,
both of which aid in improving in vivo stability and increasing
halfꢀlife.30
(iii)
(iv)
O
OH
OH
O
O
OH
OH
R
OH
O
O
O
O
O
H
O
N3
1a-1f
O
O
R
O
O
N3
NH
2a, 2b, 2d and 2e
Reagents and conditions: i) NaNO2, 5M HCl and NaN3, 0°C, 18ꢀ
98%; ii) pꢀnitrophenyl chloroformate, pyridine, DCM, 25°C, 18ꢀ
93%; iii) 7ꢀhydroxycoumarin 1, Et3N, DMF, 25°C, 63ꢀ82%; (iv)
Doxorubicin.HCl 2, Et3N, 4Å M.S., DMF, 25°C, 49ꢀ66%.
Aziridation of 3a-3f was performed using one of three alꢀ
ternate aziridation conditions (Scheme 2; see Section S2 for
alternate procedures). The activated carbonate esters 5a-5f
were synthesized by reacting 4ꢀnitrophenyl chloroformate with
the corresponding 4ꢀazidobenzyl alcohols 4a-4f. Finally, the
acyl substitution of 5a-5f with 7ꢀhydroxycoumarin 1 or the
HCl salt of doxorubicin 2 afforded the prodrug analogues 1a-
1f and 2a, 2b, 2d, and 2e, respectively.
ꢀ RESULTS AND DISCUSSION
Synthesis. Proꢀprobe analogues of 7ꢀhydroxycoumarin 1a-
1f and doxorubicin prodrugs 2a, 2b, 2d, and 2e were syntheꢀ
sized from their corresponding 4ꢀaminobenzyl alcohol 3a-3f
(Scheme 2, Scheme S1). The synthesis of 1a/2a was achieved
following the protocols previously reported.8 4ꢀAminobenzyl
alcohol 3a was commercially available, while the other benzyl
alcohols 3b-3f were synthesised from various commercially
available starting materials (Scheme 3, Scheme S1). 4ꢀAminoꢀ
3,5ꢀdifluorobenzyl alcohol 3e and 4ꢀaminoꢀ2,5ꢀdifluorobenzyl
alcohol 3f were synthesized from their respective benzyl broꢀ
mides 6e and 6f via palladiumꢀcatalysed cyanation31 to afford
aryl cyanides 7e and 7f, which were further subjected to hyꢀ
drolysis under alkaline conditions to provide benzoic acids 8e
and 8f. Benzoic acid 8e could be directly reduced with LiAlH4
to the benzyl alcohol 3e, while better overall yields of 3f were
ACS Paragon Plus Environment

Page 3 of 13
1
Bioconjugate Chemistry
(1:1, 37 °C) and the disappearance of the proꢀprobe 1a-1f at
254 nm was analysed on RPꢀHPLC. As hypothesized, the
presence of fluoroꢀgroups on the aryl ring resulted in an apꢀ
proximate 2ꢀfold, 4ꢀfold, and 6.5ꢀfold acceleration of the 1,3ꢀ
dipolar cycloaddition (relative to 1a) for the monoꢀ (1b, 1c),
diꢀ (1e, 1f), and tetrafluoroꢀsubstituted analogues (1d), respecꢀ
tively. The location of the fluoroꢀsubstituents (orthoꢀ or metaꢀ
to the azide) did not greatly affect the rate of 1,3ꢀdipolar cyꢀ
cloaddition.
2
Scheme 3. Synthesis of benzylalcohols 3b-3f.
3
4
5
6
7
8
Next, the effect of the fluoroꢀsubstituent pattern on the rate
of drug release (Rates 2-5, Scheme 1) was examined (Table 1,
right column, Figure S4 and S5). We opted for a similar apꢀ
proach to our previous work8 in which the cycloaddition step
was stopped at the triazoline intermediate (in MeCNꢀd3) beꢀ
fore release of 7ꢀaminoꢀ4ꢀmethylcoumarin was initiated
through a 1000ꢀfold dilution into PBS buffer (measured by
spectrofluorometry). However, a slight modification to the
experiments was made, in that we examined the release of 7ꢀ
hydroxycoumarin 1 from the carbonate analogues 1a-1f, and
opted for 100% MeCNꢀd3 as solvent in the initial cycloaddiꢀ
tion step (c.f. 15% DMSOꢀd6:MeCNꢀd3). The absence of
DMSO ensured that anhydrous conditions could be maintained
in the initial NMR experiment, while the use of a phenolic
leaving group simplified the release kinetics measured in the
subsequent dilution experiment. The coumarin phenol 1 is
expected to undergo a spontaneous decarboxylation (Rate 5 =
secs in Scheme 1), unlike 7ꢀaminoꢀ4ꢀmethylcoumarin which
was recently reported to have a relatively slow decarboxylaꢀ
tion (~5 min due to low pKa of protonated aniline).34
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Reagents and conditions: i) Pd2(dba)3, XPhos, K4[Fe(CN)6],
NaOAc, dioxane:H2O (1:1), 25°C, 31 and 95%; ii) KOH, H2O,
100 °C, 80 and 92%; (iii) LiAlH4, THF, 0ꢀ25 °C, 35ꢀ55%; (iv)
MeOH, H2SO4, 65 °C, 54 and 78%; (v) LiAlH4, THF, 0ꢀ25 °C, 82
and 84%.
The azidoꢀbenzyl ether linked proꢀprobes 11a and 11d
were accessed via the mesylate of the benzyl alcohol (10a,
10d) and subsequent nucleophilic substitution by 7ꢀ
hydroxycoumarin 1 (Scheme 4). A modified procedure was
33
used for the synthesis of transꢀcyclooctenol (TCO) 12,32,
which resulted in yields higher than that reported in our previꢀ
ous work.
Scheme 4. Synthesis of benzyl ether analogues 11a and 11d
Table 1. Kinetics of the pro-probes 1a-1f
aRate of 1,3-dipolar
cycloaddition
(Rate 1)
bTime (min) for
50% release of 1
(Rates 2-5)
Reagents and conditions: (i) MesCl, Et3N, DCM (ii) 7ꢀ
hydroxycoumarin 1, K2CO3, MeCN, 41% (11a) and 72% (11d).
Entry
1a
1b
1c
1d
1e
0.017 Mꢀ1sꢀ1 ± 0.003c
0.035 Mꢀ1sꢀ1 ± 0.004
0.036 Mꢀ1sꢀ1 ± 0.006
0.110 Mꢀ1sꢀ1 ± 0.036
0.067 Mꢀ1sꢀ1 ± 0.001
1.2 ± 0.8
4.0 ± 2.2
2.4 ± 1.5
7.1 ± 1.8
5.9 ± 0.4
Release kinetics of prodrug 1a-1f activation. Initial exꢀ
periments examined the release of 7ꢀhydroxycoumarin 1 from
1aꢀ1f under the 1:1 mixture of PBS:MeCN conditions we preꢀ
viously reported.8 The fluorescence of 7ꢀhydroxycoumarin 1
that had been released upon addition of TCO 12 was measured
(Figure S1a). Six hours after addition of TCO, 66ꢀ100% of 1
was found to be released from 1a-1f (Figure S1a). In the abꢀ
sence of the trigger (TCO 12), only a low level of hydrolysis
to the carbonate group in 1bꢀ1f was observed (9ꢀ16% at 24 h,
Figure S1b). The tetrafluorinated analogue 1d, that was exꢀ
pected to react fastest in the 1,3ꢀdipolar cycloaddition (Step 1
in Scheme 1), released the lowest amount of 1 at 2 h (20%),
indicating that the effect of tetrafluorination may hinder/slow
one or all of the subsequent steps in which the linker is still
attached to the model drug (Rates 2ꢀ4). Also, rates 2ꢀ4 could
be dependant on solvent, and it may be that a 1:1 mixture of
organic and aqueous solvent is not polar enough to promote
sufficient triazoline, imine and aniline degradation. Therefore,
we investigated the mechanisms along this degradation pathꢀ
way in greater detail (vide infra).
1f
0.055 Mꢀ1sꢀ1 ± 0.002
6.9 ± 1.3
aCalculated from pseudoꢀfirst order rate of 1,3ꢀdipolar cycloaddiꢀ
b
tion reaction between 1aꢀ1f and TCO 12. The cumulative rate of
1,2,3ꢀtriazoline degradation, imine hydrolysis, 1,6ꢀimmolation,
c
and decarboxylation shown as steps 2ꢀ5 in Scheme 1. Data taken
from our previous work.8 Error represented as ± SD (n = 3).
1
Briefly, the H NMR spectra of the reaction between proꢀ
probes 1aꢀ1f and TCO 12 were obtained in MeCNꢀd3 (Figure
2; Figure S3a to S3f). Following a 20 h reaction, the complete
disappearance of the proꢀprobes 1aꢀ1f was confirmed and the
intermediate triazoline or imine could be observed. Complete
conversion to the triazoline was confirmed in all cases (fused
ring protons at δ 4.3ꢀ4.5 and 3.6ꢀ3.9 ppm) except the tetrafluoꢀ
roꢀsubstituted analogue 1d (Figure 2; Figure S3d) for which
transient triazoline peaks (δ 3.5ꢀ4.5 ppm) and strong exocyclic
imine proton peaks (two isomers) could be observed (δ 8.0
ppm). Table 1 (right column) shows the time required to reꢀ
lease 50% of 1 from the triazoline and imine (Scheme 1) when
an aliquot from the NMR mixture in MeCNꢀd3 was diluted
into PBS (1000ꢀfold dilution). Pseudo firstꢀorder kinetics
could be observed over the initial time points measured (Figꢀ
ure S4 and S5), but the complexity of having multiple degraꢀ
Following the preliminary release experiments, studies on
each of the individual steps involved in the activaꢀ
tion/decaging process were investigated. The pseudo firstꢀ
order rates (Rate 1 in Scheme 1) for the 1,3ꢀdipolar cycloaddiꢀ
tion between the modified azidoꢀPABC proꢀprobes 1a-1f and
TCO 12 were determined (Table 1, left column; Figure S2a to
Figure S2e). The experiments were conducted in PBS:MeCN
3
ACS Paragon Plus Environment

Bioconjugate Chemistry
dation steps led to a deviation from firstꢀorder kinetics, and
Page 4 of 13
1
2
3
4
5
6
7
8
was particularly evident for analogues 1aꢀ1c. While the first
order rate could not be calculated and directly compared
across the series, the maximum observed release, indicated by
the plateau of 7ꢀhydroxycoumarin 1 release, occurred within
approx. 1 h of dilution (Figure S4 and S5). Interestingly, we
observed that sequential addition of fluorine groups on the aryl
ring, resulted in a decreased rate of coumarin 1 release as eviꢀ
denced by the right shift in the release profile for 1bꢀ1f c.f. 1a
(Figure S4). In particular, there was a marked difference obꢀ
served for the tetrafluoro analogue 1d, taking approx. 1 h for
the release of 1 to reach its plateau; i.e. 1d with four fluorine
groups was the slowest across steps 2ꢀ5.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Mechanistic investigation of fluorine substituent effect
on triazoline and imine degradation. The results from Table
1 suggest that while fluorine accelerates the 1,3ꢀdipolar cyꢀ
cloaddition, it slows the decaging process. For prodrug activaꢀ
tion, the 6.5ꢀfold faster click reaction of the tetrafluoroꢀ
substituted azide bodes well for increased localization of the
azide prodrug at the desired biological target. However, since
we require a clickꢀandꢀrelease event, it is critical to understand
the mechanism by which substituents on the PABC linker
influence the decaging process (Steps 2ꢀ4). Not only is this
important for mechanistic understanding, but moving forward
to in vivo models it could provide us with knowledge on how
to gain temporal control over the release process, and maybe
even lead to sustainedꢀrelease drug delivery systems.
Electron deficient 1,2,3ꢀtriazolines are known to be therꢀ
mally unstable.20, 25, 35 Since the triazolines for our analogues
1bꢀ1f are also electronꢀpoor we hypothesize that the degradaꢀ
tion (Scheme 1, Step 2) will not be rateꢀlimiting. Rather the
triazolines are expected to be even less stable as the number of
fluorine substituents increases. Therefore, we hypothesize that
the imine hydrolysis would be the rateꢀlimiting step (Scheme
1, Step 3), especially as the pKa for the nitrogen atom on an
aryl imine is lowered in the presence of electronꢀwithdrawing
groups,36 slowing the acidꢀcatalyzed hydrolysis step. The 1,6ꢀ
selfꢀimmolation step could also contribute to the rate, but our
expectation was that the elimination would be fast. However,
this did not prove to be the case for the analogue containing
the tetrafluorinated linker (vide infra). Interestingly, recent
reports on the Staudinger reaction between tetrafluoroaryl
azide and triphenylphosphine resulted in a stable iminophosꢀ
phorane, making the reaction unsuitable for clickꢀtoꢀrelease
applications.21 In comparision, the iminophosphorane formed
by the reaction of nonꢀsubstituted aryl azide results in unstable
iminophosphoranes.9,16 This highlights that the hydrolysis of
the imine generated from tetrafluoroaryl analogue 1d could
play a significant role in the rate of release.
Figure 2. Product distribution for reaction of 1a-1f and TCO 12
in 100% MeCN-d3 recorded at 20 h.
The identity of the observed peak at 8.02 was confirmed
by ESI+ mass spectral (MS) analysis, with a peak at m/z 530.1
assigned as the sodiated molecular ion [M+Na] of the imine
(C25H21F4NO6Na). In the spectrum of 1d, there was a relativeꢀ
ly low abundance of the bridging triazoline proton peaks.
From this we can conclude that the presence of four fluorine
groups on the PABC ring leads to a relatively unstable 1,2,3ꢀ
triazoline, even under pure organic solvent conditions (MeCNꢀ
d3), and that the imine hydrolysis could be rateꢀlimiting, espeꢀ
cially as fluorine substituents on the ring increase to four.
Further evidence for the proposed stabilization of the
imine bond by the four fluorine substituents on the aromatic
ring was provided by RPꢀHPLC analysis. Upon reaction of 1d
with TCO 12, only
a trace amount of released 7ꢀ
hydroxycoumarin 1 (RT = 4.9 min) was observed in the HPLC
chromatogram (Figure 3). Instead, a major peak at RT = 7.5
min was visible. In an attempt to confirm the identity of the
new peak, the NMR mixture of 1d containing the imine (from
the product distribution experiment in figure 2) was injected
onto the HPLC. The NMR mixture showed a retention time
matching that of the previously unknown peak at 7.5 min
(Figure 3), suggesting the presence of a stable imine. Howevꢀ
er, when the eluant for the expected imine (RT = 7.5 min) was
collected and reꢀanalysed by MS, the imine molecular ion or
its sodiated form was not present, and no conclusive product
peak could be identified. Based on all of the experimental
evidence (vide supra), the presence of a highly stabilized aniꢀ
line that was present under the HPLC assay conditions and yet
to undergo a 1,6ꢀselfꢀelimination, could not be ruled out. The
aniline itself could be generated via hydrolysis of the imine
upon dilution of the NMR sample in PBS. Of note, release of
the coumarin from the imine or the aniline could be influenced
by the solvent in a way that an increase in the amount of couꢀ
marin released would be expected as the percentage of waꢀ
ter/PBS increases. That is, the intermediate imine or aniline is
stabilised when there is a relatively high prodrug:proton ratio
To test our hypotheses; i.e. the triazoline is destabilised as
fluorine substitution increases, and the imine hydrolysis beꢀ
comes rateꢀlimiting; 1H NMR was used to examine the product
distribution for the reaction between proꢀprobes 1a-1f and
TCO 12 (Figure 2). Analogues 1a-1c, 1e and 1f did not exhibit
a peak for the imine proton and only 1,2,3ꢀtriazoline was eviꢀ
dent, but for the tetrafluoroꢀsubstituted PABC analogue 1d, a
strong peak corresponding to an imine proton at δ 8.02 ppm
was observed (Figure 2, Figures S3aꢀS3f).
4
ACS Paragon Plus Environment

Page 5 of 13
Bioconjugate Chemistry
(purely organic or organic/aqueous mixture), but destabilized
under relatively low prodrug:proton ratio (purely aqueous;
expected conditions in vivo).
minutes) instead of the benzyl carbonate (release rate in the
order of seconds),38 enabling a better understanding of the 1,6ꢀ
selfꢀimmolation and its influence on the overall release rate.
To quantify the products formed in the reaction of 1d with 12,
an internal standard (1ꢀFluoroꢀ2,4ꢀdinitrobenzene) that could
be detected by 1H and 19F NMR spectroscopy was added.
1
2
3
4
1000
900
5
6
7
8
9
R
800
R
Peak at 8.6 min
R
R
O
O
O
H
without activation
R
N
700
O
O
O
N
R
with TCO 12
+
N
H
R
N3
R
(13a)
R = H
(11a)
R = H
R = F (11d)
OH
600
500
R = F (13d)
(Triazoline)
12
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OH
400
R
R
Peak at 7.5 min 20 min after
R
R
activation with TCO 12
R
O
O
O
300
O
R
O
O
Peak at 8.6 min
20 min after activation
N
R
N
R = H (14a)
R = F (14d)
with TCO 12
NMR mixture shows a
200
R
R = H (15a)
(15d)
Peak at 7.5 min
(Ketimine)
R = F
(Aldimine)
100
HO
OH
0
R
R
R
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
R
N
R
Minutes
O
O
O
O
O
O
Figure 3. HPLC traces of unidentified peak formed in the reaction
of 1d with 12. The NMR reaction mixture of 1d (with confirmed
imine), had the same peak at RT = 7.5 min.
R
H2N
R
R
HO
O
O
R = H (16a)
R = F (16d)
R = H (17a)
R = F (17d)
(Aniline)
1
HO
(7-hydroxycoumarin)
(Aziridine)
Based on the above NMR and HPLCꢀUV data, it was eviꢀ
dent that the triazoline for the tetrafluorinated analogue 1d
was destabilised by the additional fluorine substituents. Howꢀ
ever, there was still no direct evidence suggesting which step;
imine hydrolysis or 1,6ꢀselfꢀimmolation of the aniline, was
rateꢀlimiting. To acertain if fluorine substitution could influꢀ
ence the protonation of the imine; i.e. slowing the hydrolysis
and release of coumarin, the molecular electrostatic potential
(MEP) at the nucleus of the nitrogen atom for the imines of
1a-1f were computed and correlated to a relative pKa. This
quantity has previously been shown to correlate strongly with
the experimentally determined pKa such that a more negative
MEP corresponds to a larger pKa value.37 The MEP at the
nuclei was calculated via the “Prop” keyword in Gaussian09.
Table S3 (Section S6) summarises the values of predicted pKa
values of the imines corresponding to 1a-1f. The data demonꢀ
strates that the predicted pKa of the imine corresponding to the
tetrafluoroꢀsubstituted linker 1d (pKa ~5) is approx. twoꢀ
orders of magnitude below physiological pH and threeꢀorders
of magnitude lower than the imine of the nonꢀsubstitued anaꢀ
logue 1a (pKa ~8). The computed data supports the notion
that, as the number of fluorine substituents increase the pKa of
the imine nitrogen decreases. This means that the tetrafluoriꢀ
nated imine of 1d is least susceptible to protonation under
physiological conditions (pH 7.4), and that acidꢀcatalyzed
imine hydrolysis could be, at least in part, responsible for the
slow release of coumarin 1 from 1d.
Figure 4. Possible product distribution for benzyl ethers 11a and
11d after reaction with TCO 12. Products of the reaction were
monitored by H NMR (for 11a and 11d) and 19F NMR (for 11d
1
only).
Initially, the product distribution after addition of TCO 12
to 11a or 11d in MeCNꢀd3 was monitored by 1H and 19F NMR
spectroscopy for 48 h (11a) or 72 h (11d). It was evident from
the H and 19F NMR (for 11d only) spectra, that all of azide
1
11a (Figure S9a) and 11d (Figure 5, 6 and S9b) had been conꢀ
verted to the triazoline 13a or imine 15d (confirmed by HRMS
analysis with a peak at m/z 486.1307; [M+Na]), respectively.
1
While not conclusively identified in the H NMR spectrum,
aziridine 16d has the same molecular weight as imine 15d,
thus could also be present in the reaction mixture. In the reacꢀ
tion of 11d with 12, the relative integration of imine protons at
δ 8.02 ppm did not exhibit the expected 1:1 ratio with the
coumarin protons (measured at 72 h; Figure S9b). Instead
there was a 1:1.5 relative integration. Closer inspection of the
coumarin core proton peaks and the presence of a peak at δ
5.25 ppm, indicated that a structurally related product with
similar chemical shifts to the imine was present. The 19F NMR
spectrum further supported the presence of other minor prodꢀ
ucts (multiplets at δ ꢀ155.8, ꢀ154.8, ꢀ147.2, ꢀ142.6 ppm) in
addition to the imine (multiplets at δ ꢀ146.1 and ꢀ156.8 ppm)
and a small amount of remaining triazoline (multiplets at δ ꢀ
148.8 and ꢀ144.6 ppm). We hypothesized that these additional
peaks could correspond to the ketimine 14d or aziridine 16d.
After 72 h (reaction time in NMR tube), a 1000ꢀfold dilution
of the reaction mixture containing triazoline 13a in PBS and
monitoring by spectrofluorometry, showed 50% release of 1 in
2.40 ± 0.53 min (Figure S9d), comparable to other 1,6ꢀselfꢀ
To further examine the product distribution, imine hydrolꢀ
ysis and 1,6ꢀselfꢀimmolation steps experimentally, a series of
NMR and fluorescence experiments were conducted using
benzyl ether analogues 11a and 11d (Figure 4). The benzyl
ether linker was selected in order to avoid any background
hydrolysis that could occur to the carbonate bond in the correꢀ
sponding analogues of 1a and 1d. In addition, the experiment
was designed so that the 1,6ꢀselfꢀimmolative step would proꢀ
ceed via a benzyl ether (expected release rate in the order of
immolative processes for benzyl ethers releasing low pKa
39
phenols.38,
Strikingly, the release from the tetrafluoroꢀ
5
ACS Paragon Plus Environment

Bioconjugate Chemistry
Page 6 of 13
aniline 17d, 10% of aziridine 16d (unchanged), and 35% of
substituted imine 15d was ~125× slower, taking ~5 h for 50%
of 1 to be released after dilution into PBS (Figure S9d).
1
2
3
4
5
6
7
8
what is suspected to be the enamine tautomer of aldimine 15d.
1
This result was supported by the H NMR spectrum acquired
To ascertain which step was rateꢀlimiting in the case of
15d (imine hydrolysis or 1,6ꢀselfꢀimmolation) and determine
if the additional products were ketimine 14d or azidirine 16d,
D2O (10% by volume) was added to the NMR solution (at 72
h). For comparision, D2O (10% by volume) was also added to
the triazoline 13a (at 72 h). Addition of D2O to the mixture
containing triazoline 13a resulted in significant release of 1
within 24 h (Figure S9a, 7ꢀhydroxycoumarin peaks illustratꢀ
ed). The aldehyde peak at δ 9.56 ppm indicated that 13a had,
at least in part, released 1 via imine hydrolysis and 1,6ꢀselfꢀ
immolation via 17a. Addition of D2O to the mixture containꢀ
ing 15d did not result in the release of 1 (Figure 5, S9b), howꢀ
ever, an aldehyde peak was present indicating that imine hyꢀ
drolysis had occurred. Using HRMS analysis, the major prodꢀ
96 h post D2O addition (Figure S9b); i.e. 50% of imine hyꢀ
drolysis evident via presence of aldehyde (δ 9.56 ppm).
The results suggest that the imine hydrolysis does occur
quite rapidly under relatively low abundance of water (10%
D2O) for the nonꢀsubstituted imine (Figure S9a; no imine peak
observed and aldehyde peak evident within 5 mins of D2O
addition). As the number of electronꢀwithdrawing fluorine
substituents increase, the hydrolysis is slowed (Figure S9b;
trace amounts of imine 15d remain at 24 h post D2O addition),
as predicted by the calculated relative pKa values of the imine
nitrogen. The presence of four fluorine atoms on the aryl
group would also be expected to increase the electrophilicity
of the carbon atom of the imine bond (increasing the hydrolyꢀ
sis rate), but the results suggest that the reduced pKa of the
nitrogen dominates (decreasing the hydrolysis rate). However,
the overall rateꢀlimiting step in the degradation process of the
tetrafluoroꢀsubstituted triazoline 13d and imine 15d, is likely
to be the 1,6ꢀselfꢀimmolation of aniline 17d. One possible
explanation for the high stability of 17d vs the low stability of
17a could be that shortening of the benzylic carbonꢀoxygen
bond by the electronegative fluorine atoms slows down 1,6ꢀ
selfꢀimmolation, an opposite effect to that reported for addiꢀ
tion of electronꢀdonating methoxy aryl ring substituents (bond
elongation leading to rapid 1,6ꢀselfꢀimmolation).39
In summary, the NMR and spectrofluorometry data for 1aꢀ
1f, 11a, and 11d suggest that, after the 1,3ꢀdipolar cycloaddiꢀ
tion and addition of a protic solvent, the 1,2,3ꢀtriazoline 13a/d
is rapidly converted to an aldimine 15a/d (and enamine), a
ketimine 14a/d and an aziridine 16a/d. The aldimine/enamine
15a/d are converted to the aniline 17a/d via hydrolysis of the
aldimine. The rate of imine hydrolysis and the contribution of
the 1,6ꢀselfꢀimmolation to the overall release rate of coumarin
1 increases as the number of fluorine substituents on the
PABC linker increases. For tetrafluoroꢀsubstitution (1d and
11d), imine hydrolysis and/or 1,6ꢀselfꢀimmolation (especially
for benzyl ether 11d) dominate the observed rate.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
uct identified in the H NMR spectrum was confirmed as aniꢀ
line 17d (m/z 362.0402; M+Na ion). Closer inspection of the
19F NMR spectrum (Figure 6 and S9c) confirmed one major
product with peaks at δ ꢀ148.3 and ꢀ163.5 ppm (~60% of total
19F integration), which combined with the 1H NMR and
HRMS data could be assigned to aniline 17d. There were
some minor products present (~40% of total integration). The
second most notable product after D2O addition, was noted at
δ ꢀ146.9, ꢀ147.2, ꢀ158.2, and ꢀ158.6. This product starts formꢀ
ing immediately after addition of D2O, and then slowly deꢀ
creased over the 96 h period resulting in further formation of
aniline 17d. Therefore, these peaks likely correspond to the
enamine tautomer (two possible isomers) of 15d, which exists
in equilibrium with the imine. There was a slight doubling in
1
the peaks of the aromatic region of the H NMR spectrum
(Figure S9b) at the corresponding timeꢀpoints, indicating that
the enamine has very similar chemical shift values to aniline
17d. The minor product with peaks centred at δ ꢀ147.0 and ꢀ
155.5 (~15% of total integration) were present in the sample
prior to addition of D2O (note the slight change in chemical
shift observed due to change in solvent). Since they were staꢀ
ble under both NMR solvent compositions, the peaks likely
correspond to aziridine 16d, a compound that does not underꢀ
go hydrolysis, and not ketimine 14d or the enamine tautomer
of 15d. Upon addition of D2O, the multiplets at δ ꢀ154.8 and ꢀ
142.6 ppm disappeared, indicative of ketimine 14d. The low
abundance of these peaks in the original experiment (preꢀD2O)
also points towards ketimine 14d.25
Quantification of the products formed from 11d using the
internal standard (1ꢀFluoroꢀ2,4ꢀdinitrobenzene) was also perꢀ
formed, and supported the initial relative 19F integrations (vide
supra). Due to chemical shifts of the products/intermediates
being very similar in the 1H NMR spectrum, more quantitative
information could be gained from the 19F NMR study. After
incubating at room temperature for 48 h, the product distribuꢀ
tion was; 60% aldimine 15d, 10% aziridine 16d (see above for
identification), 15% of 1,2,3ꢀtriazoline 13d, and trace amounts
14d. Prior to addition of D2O (at 72 h), the product distribution
was; 65% aldimine 15d, 10% aziridine 16d, 10% of 1,2,3ꢀ
triazoline 13d. After D2O addition (10% by volume) and incuꢀ
bation for 96 h (Figure 6), the product distribution was; 55%
Computational analysis of the 1,2,3-triazoline degrada-
tion. Triazolines are an important class of compounds used in
synthetic chemistry, and there is a significant amount of effort
involved in understanding their formation and subsequent
25, 26, 40ꢀ47
degradation mechanisms.20,
Therefore, the results
obtained in our NMR and HPLCꢀUV experiments provide an
interesting addition to the current mechanistic understanding
of triazoline synthesis, stability, and subsequent degradation
products. To understand the reason for the instability of the
triazoline formed in the case of 1d/11d relative to the triaꢀ
zolines resulting from the cycloaddition of the nonꢀsubstituted
1a/11a and other fluoroꢀsubstituted analogues 1b, 1c, 1e, and
1f, a mechanistic study was conducted to relate the fluorine
substitution to the triazoline degradation and imine hydrolysis.
Herein, we discuss the outcome of the computational evaluaꢀ
tion and compare this to the experimental observations reportꢀ
ed above.
6
ACS Paragon Plus Environment

Page 7 of 13
Bioconjugate Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Figure 5. 1H NMR product distribution experiments for reaction of TCO 12 with 11d a) t = 5 min (MeCNꢀd3) b) t = 72 h (MeCNꢀd3) c) t =
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5 min after adding 10% D2O d) t = 24 h after adding 10% D2O e) t = 96 h after adding 10% D2O.
Figure 6. 19F NMR product distribution experiments for reaction of TCO 12 with 11d a) t = 5 min (MeCNꢀd3) b) t = 72 h (MeCNꢀd3) c) t =
5 min after adding 10% D2O d) t = 24 h after adding 10% D2O e) t = 96 h after adding 10% D2O.
R
The relative rates of 1,2,3ꢀtriazoline degradation were inꢀ
Me
R
R
Me
Me
vestigated using ab initio techniques (Section S6). Inclusion of
implicit solvation was found to have a minimal effect on the
thermodynamics of 1,2,3ꢀtriazoline (Table S1) therefore simꢀ
pler, solventꢀfree methods were used to investigate the kinetꢀ
ics. As the aryl azide is the variable portion of the proꢀprobes
N
N
N
N
N
N
N
N
N
OH
OH
OH
1a-1f,
a simplified model was used where the 7ꢀ
(r), Reactant
(t), Transition state
(p), Product
hydroxycoumarin and the carbonate group were replaced by a
methyl group in the paraꢀposition with respect to the azide on
the linker motif (18a-18f; Figure 7). The degradation process
was modelled by the elementary reaction step of the ring openꢀ
ing (Figure 7).
F
F
Me
Me
F
F
Me
F
Me
Me
F
Me
N3
N3
N3
N3
N3
N3
F
F
F
F
(18a)
(18b)
(18c)
(18d)
(18e)
(18f)
Figure 7. Elementary step chosen to model 1,2,3ꢀtriazoline degꢀ
radation.
Transition states of the triazoline degradation were identiꢀ
fied using a relaxed surface scan through the NNN angle, reꢀ
fined using the QST3 method and confirmed using frequency
and IRC calculations (Refer to Table S2 for the coordinates
and energies of the transition states of 1,2,3ꢀtriazoline degraꢀ
7
ACS Paragon Plus Environment

Bioconjugate Chemistry
Page 8 of 13
IC50a,b,c (ꢀM) on acti-
dation). Table 2 summarises the energetics and relative rates
IC50a,b,c (ꢀM)
Entry
1
2
3
4
5
6
7
8
of the triazoline degradation. Evidently the triazoline correꢀ
sponding to 18d is highly unstable with a 27000ꢀfold (fourꢀ
orders in magnitude) higher krel than 18a (Figure S8). Addiꢀ
tionally, 18e and 18f were computed to have a 60 to 600ꢀfold
(approximately one to twoꢀorders in magnitude) higher krel
than 18a. Hence, it can be concluded that introducing the fluoꢀ
rine group onto the aryl ring of the PABC linker destabilizes
the 1,2,3ꢀtriazoline and increases the rates of triazoline degraꢀ
dation. This supports the experimental data obtained in the
1H/19F NMR and fluorometry studies (vide supra).
vation with TCO
Doxorubicin
2a
0.84 (0.79 – 0.89)
29.40 (11.64 – 74.28)
37.00 (19.43 – 70.46)
>50
32.95 (19.06 – 56.97)
ꢀ
1.39 (1.25 – 1.53)
2b
2d
2e
1.06 (0.95 – 1.17)
5.97 (4.68 – 7.60)
0.98 (0.88 ꢀ 1.09)
aIC50 values for doxorubicin 2, prodrugs (2a, 2b, 2d, and 2e) and
activation of the prodrugs with TCO 12 (100 ꢀM), incubated for
b
48 h at 37 °C. IC50 = concentration required to kill 50% of cells.
9
c95% confidence interval (n ≥ 3) in parenthesis.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Energetics and computed krel of 1,2,3ꢀtriazoline degꢀ
radation
ꢀ CONCLUSIONS
The effect of fluorineꢀsubstitution on aromatic azides that
take part in a potentially bioorthogonal 1,3ꢀdipolar cycloaddiꢀ
tion with TCO 12 has been investigated. The role of the fluoꢀ
rine substituents on the kinetics of the reaction and subsequent
release/decaging of the PABC linker from a drug or probe
have been discussed, and key mechanistic insights have been
made. The rate of the cycloaddition increased upon sequential
addition of fluoroꢀsubstituents, with a maximum rate of 0.110
Mꢀ1sꢀ1 achieved in the tetrafluoroꢀsubstituted analogue 1d, an
approximate orderꢀofꢀmagnitude improvement over the nonꢀ
substituted linker 1a. However, the opposite trend was obꢀ
served for the subsequent release/decaging step; i.e. as fluoꢀ
rineꢀsubstitution increased, the rate of release/decaging
slowed. Spectrofluorometry, NMR and computational studies
demonstrated that the slower release/decaging was a result of
the electronegative fluorine atoms, which stabilize the interꢀ
mediate aldimine, and in the case of the tetrafluorinated aryl
azides, the aniline that is primed for 1,6ꢀselfꢀimmolation.
Doxorubicin prodrugs caged with the fluorinated aryl azides
2a, 2b, 2d, and 2e were relatively inactive against a melanoma
cell line, and upon reaction with TCO 12 were decaged, and
cytotoxicity of the parent doxorubicin was almost entirely
restored, except for the tetrafluorinated prodrug 2d. We are
currently investigating the use of these and other caging
groups in in vitro and in vivo preꢀtargeting strategies (TCOꢀ
antibody conjugates). Our lab is also investigating the use of
fluorinated PABC caging groups in sustainedꢀrelease drug
delivery. Lastly, as part of our investigation we were able to
examine the stability of 1,2,3ꢀtriazolines (experimentally and
computationally), and the subsequent product distribution
following degradation of the ring. This adds mechanistic inꢀ
sight into the degradation process, and complments the work
of others in this area.20, 25, 26, 40ꢀ47 Due to the importance of the
triazoline in organic chemistry, this provides valuable inforꢀ
mation for synthetic chemists, enabling them to design triaꢀ
zolines of varying stability.
krel
(298 K)
Entry
ꢁG / kJ molꢀ1
Ga / kJ molꢀ1
18a
18b
18c
126
123
124
199
194
194
1
9
9
18d
18e
18f
112
119
120
174
183
189
27268
585
62
ꢁG = Gproduct – Greactant, Ga = Gtransition state – Greactant
Cytotoxicity of TCO-activated doxorubicin-based pro-
drugs 2a, 2b, 2d, and 2e. Based on the kinetics reported in
Table 1, the number of fluorine groups on the PABC linker
determines the kinetics of drug/probe decaging irrespective of
its position on the aromatic ring. Hence, we analysed the doxꢀ
orubicin prodrugs 2a, 2b, 2d, and 2e to examine if activation
by 1,3ꢀdipolar cycloaddition restored the cytotoxicity of doxoꢀ
rubicin 2 against a B16ꢀOVA melanoma cell culture (Table
3). After 48 h incubation in the presence of TCO 12 (100 ꢀM),
varying degrees of activity were restored, indicating activation
of the prodrug and subsequent release of free doxorubicin 2. In
the absence of TCO 12, prodrugs 2a, 2b, 2d, and 2e were 10ꢀ
toꢀ35ꢀfold less toxic to the melanoma cells, indicating that the
azidoꢀPABC linkers deactivate doxorubicin and are relatively
stable in vitro over 48 h. Notably, the tetrafluoroꢀsubstituted
prodrug 2d had the highest IC50 (>50 ꢀM) of all prodrugs, but
upon activation full activity of doxorubicin 2 was not restored
(5.97 ꢀM vs. 0.84 ꢀM). This may be a result of slow aldimine
hydrolysis or 1,6ꢀselfꢀimmolation of the aniline derivative, as
previously demonstrated for coumarin analogues 1d and 11d.
Based on our fluorescence experimental results with 1d
and 11d, the release under a protic solvent should still be relaꢀ
tively rapid (minutes to hours), and the 48 h incubation should
have been sufficient time for complete release of doxorubicin
2. Therefore, the reason for not seeing complete conversion of
the prodrug 2d to doxorubicin 2 may lie in the fact that four
fluorine atoms impart a high lipophilicity on the prodrug 2d
and the intermediate aldimine and aniline. In this situation, the
imine and aniline may diffuse more readily into the lipid biꢀ
layer of the cells (favoured equilibrium) where the proton
source is low and hence imine protonation and 1,6ꢀselfꢀ
immolation is slowed.
ꢀ EXPERIMENTAL METHODS
Spectrofluorometry Release Experiments (Section S3).
In a typical experiment: Stock solutions of coumarin probe 1a-
1f (5 mM, Stock A), in MeCN and TCO 12 (50 mM, Stock
B), in MeCN were prepared. A solution of PBS:MeCN was
prepared to a volume of 800 ꢀL (500 ꢀL PBS and 300 ꢀL
MeCN), to which, 100 ꢀL of Stock A was added (resulting in
a volume of 900 ꢀL). In the case of control studies, 100 ꢀL of
MeCN was added to the result in a total volume of 1 mL (0.5
mM of the coumarin probe). To begin the release experiments,
Table 3. Cytotoxicity study of TCO 12 activated prodrugs 2a,
2b, 2d, and 2e against B16ꢀOVA melanoma cells.
8
ACS Paragon Plus Environment

Page 9 of 13
Bioconjugate Chemistry
spectrometer were plotted and the 50% release was calculated
100 ꢀL of stock B was added for a final volume of 1 mL (1:1
1
2
3
4
5
6
7
8
using GraphPad Prism 7 software. (Figure S4)
PBS:MeCN), a 0.5 mM solution of coumarin probe and a 5
mM solution of TCO 12. The solutions (including the control
runs) were incubated at 37 °C and at relevant time points, a 50
ꢀL aliquot was taken from the incubated solutions and diluted
in PBS (2950 ꢀL, 60× dilution). The PBS solution was vorꢀ
texed briefly, transferred to a plastic cuvette and the fluoresꢀ
cence measured at 455 nm (Ex 360 nm). The experiment was
repeated for a total of three runs (n = 3). The amount of 1 reꢀ
leased was calculated from a standard curve.
HPLC Kinetic Experiments ꢀ 1,3-Dipolar Cycloaddi-
tion (Section S4). A stock of TCO 12 in MeCN was prepared
at 100 mM or 200 mM, and a 7ꢀhydroxycoumarin probe 1a-1f
stock was prepared at 5mM in MeCN (Stock A). For the conꢀ
trol experiment (run in triplicate), 100 µL of coumarin probe
1a-1f stock (Stock A, in MeCN) was added to a solution conꢀ
taining 400 µL MeCN and 500 µL PBS (total volume of 1
mL). From this solution was taken a 50 ꢀL aliquot, which was
diluted into a 950 ꢀL solution of PBS:MeCN (1:1) in an
HPLC vial. The solution was injected onto the HPLC (inj. vol.
Computational Studies (More details in Section S6).
Initial structures were generated using a Monte Carlo conꢀ
former distribution search performed in Spartan ’1448 with a
Merck molecular force field (MMFF06).49 This was performed
only for 18e and other molecules were built from this one.
10,000 conformers were searched and the lowest energy conꢀ
former was retained for further calculation. This conformer
was found to account for either a very high proportion of the
population at 298 K (>96 %) or, in the case of other lowꢀlying
conformers, the structures were very similar near the triazoline
and varied only further out on the probe. Minima and transiꢀ
tion states were optimised using density functional theory
(DFT) with the B3LYP functional and the 6ꢀ31G+(d) basis set
with the Gaussian 09 program suite.50 Vibrational frequency
calculations were performed to confirm either minima (no
imaginary frequencies) or transition states (one imaginary
frequency) had been found. Intrinsic reaction coordinate (IRC)
calculations were performed to ensure the TS connected the
relevant minima.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
o
50 ꢀL) at t = 0 min. The control was then incubated at 37 C
Relative rates of 1,2,3ꢀtriazoline degradation were investiꢀ
gated by using a simplified model (Figure 7) where the probe
and carbonate group were replaced by a methyl group at the
paraꢀposition on the PABCꢀlinker. The degradation process
was modelled by the elementary reaction step of the ring openꢀ
ing, as shown in (Figure 7). The difference between the transiꢀ
tion state and reactant energies of this step was taken as the
activation energy and a constant prefactor was assumed for all
linkers. Transition states were identified using a relaxed surꢀ
face scan through the NNN angle, refined using the QST3
method and confirmed using frequency and IRC calculations.
To calculate the MEP at the nucleus, the “Prop” keyword
is added to the root section of a calculation. This keyword
directs Gaussian to compute electrostatic properties. By deꢀ
fault, the potential, electric field, and electric field gradient at
each nucleus are computed. The density used for the electroꢀ
static analysis is controlled by the “density” keyword. MEP
values are converted to pKa values based on an empirical relaꢀ
tionship established by Liu and Pedersen37 based on the measꢀ
ured pKa and computed MEP of 154 amines and anilines (see
SI). Note that the relationship only holds for N atoms and difꢀ
ferent relationships exist for other atom types.
for 240 mins before a further aliquot was taken and measured
by HPLCꢀUV at 254 nm (this was run in triplicate).
To begin the 1,3ꢀdipolar cycloaddition reaction, 100 µL of
the TCO 12 stock was added to a vial containing 100 µL of the
coumarin probe stock (Stock A), 300 µL MeCN and 500 µL
PBS. The reaction was incubated at 37 °C and at the time
points indicated, a 50 ꢀL aliquot was taken and diluted in a
950 ꢀL solution of PBS:MeCN (1:1). The solution was injectꢀ
ed onto the HPLC (inj. vol. 20 ꢀL or 50 ꢀL), and the disapꢀ
pearance in absorbance for the coumarin probe 1a-1f was
measured at 254 nm (triplicate experiments). From the pseudo
firstꢀorder plots, the secondꢀorder rates could be calculated
using the concentration of TCO 12 (5 mM, 10 mM or 20 mM).
Spectrofluorometry Kinetic Experiments - Triazoline
and Imine Degradation 1a-1f (Section S5). In a typical ex-
periment: Stock solutions of coumarin probe 1a-1f (40 mM,
Stock C), in MeCNꢀd3 and 12 (50 mM, Stock B), in MeCNꢀd3
were prepared. An aliquot was taken from the Stock C (130
ꢀL) and added to an NMR tube containing 195 ꢀL of MeCNꢀ
d3. To start the 1,3ꢀdipolar cycloaddition reaction, 325 ꢀL of
Stock B was added to the mixture (shaken well for 5 secs),
and analysed by NMR spectroscopy. The addition of the
stocks resulted in a final concentration of 8 mM for coumarin
probe 1a-1f and 25 mM of 12. The NMR sample was incubatꢀ
ed at 25 °C and the progress was measured at 0 h and 20 h.
After 20 h of incubation, the 1H NMR spectra indicated that all
of the azide had reacted and that conversion of the coumarin
probe to the corresponding triazoline or imine had occured
(See Figure S3a – S3f).
1H NMR studies of bioorthogonal activation of the pro-
probes 11a and 11d:
A 10 mM stock of the 7ꢀ
hydroxycoumarin proꢀprobes 11a and 11d in MeCNꢀd3, 250
mM solution of the internal standard (1ꢀfluoroꢀ2,4ꢀ
dinitrobenzene) and a 935 mM stock of TCO 12 in MeCNꢀd3
were prepared. 500 ꢂL of coumarin (11a and 11d) stock, 20
ꢂL of the interal standard was further diluted with 210 ꢂL of
MeCNꢀd3. To begin the 1,3ꢀdipolar cycloaddition, 20 ꢂL of
TCO 12 stock was added (resulting in a final concentation of
24.93 mM of 12 and 6.67 mM of the coumarin probe and the
internal standard; the mixture was incubated in the dark at
room temperature). For 48 h (for 11a) and 72 h (for 11d), the
An aliquot of the NMR sample containing triazoline (3 ꢂL)
was further diluted into PBS (2997 ꢂL, 1000ꢀfold dilution;
resulting in 8 ꢂM of the coumarin proꢀprobe), starting the triaꢀ
zoline/imine degradation. The PBS solution was vortexed
briefly, transferred to a plastic cuvette and the fluorescence
was scanned over 3600 seconds at 455 nm (ex. 360 nm). The
experiment was repeated for a total of three runs (n = 3). From
the standard curve of 7ꢀhydroxycoumarin 1 (Figure S6), the
maximum amount of 1 release from proꢀprobes was calculated
to be 1146.63 units (8 ꢂM). The plots from the timeꢀscan on
1
reaction progress was analysed via H NMR spectroscopy at
room temperature to ensure conversion to the corresponding
triazoline 13a (from 11a) or aldimine 15d (from 11d). And
upon complete conversion, 75 ꢂl of D2O was added to the
sample to initiate triazoline and imine degradation; NMR
samples were measured at indicated time points.
9
ACS Paragon Plus Environment

Bioconjugate Chemistry
Page 10 of 13
ꢀ
ACKNOWLEDGMENT
Similar to spectrofluorometry kinetic studies, a 3 ꢂL aliꢀ
quot of the NMR sample (before the addition of D2O) was
diluted into 2997 ꢂL of PBS to initiate triazoline degradation,
imine hydrolysis and 1,6ꢀelimination which was measured by
spectrofluorometry timeꢀscan over 3600 seconds for 11a and
various time points over 2 days for 11d at excitation 360 nm
and emission 455 nm. The time for 50% release was calculatꢀ
ed to be 2.40 ± 0.53 minutes (n=3). The time for 50% release
from tetrafluoroꢀsubstituted linker was calculated to be ~4.83
hours (n=2).
Cell Culture and Proliferation Assay: The murine B16ꢀ
OVA melanoma cells were maintained in a humidified CO2
(5%) incubator (Heraeus) at 37 °C in RPMIꢀ1640 media (Gibꢀ
co), supplemented with 10% fetal bovine serum (FBS) (Gibꢀ
co), 1% GlutaMAX (Gibco), 1% Pen Strep (Gibco) and 0.1%
2ꢀMercaptoethanol (55 mM) (Gibco). Cells were plated out in
flat bottom 96ꢀwell plates (BD Falcon) at a density of 7500
cells/well and allowed to attach for 24 h. The doxorubicin
prodrugs 2a, 2b, 2d, and 2e were dissolved in DMSO at 10
mM and subsequently serialꢀdiluted in preꢀwarmed culture
media. TCO 12 was dissolved in DMSO at 100 mM and dilutꢀ
ed to 200 µM in preꢀwarmed culture media. After 24 h cell
attachment, cell culture medium was replaced by 50 µL preꢀ
warmed media containing the compounds in different concenꢀ
trations and 50 µL preꢀwarmed media containing 12 at 200
µM (final concentration in each well was 100 ꢀM). After 48 h
incubation, cell proliferation was assessed by means of a MTT
assay.
1
2
3
4
5
6
7
8
The authors would like to thank the NMR and MS facility at Department
of Chemistry (Otago) for allowing us to use their facilities, Dr Pummy
Kritaphol for technical support with HPLC experiments and the Health
Research Council (HRC) for a doctoral scholarship (J.M.F).
ꢀ
REFERENCES
(1) Sletten, E. M. and Bertozzi, C. R. (2009) Bioorthogonal
chemistry: Fishing for selectivity in a sea of functionality. Angew.
Chem. Int. Ed. 48, 6974ꢀ6998.
(2) Saxon, E. and Bertozzi, C. R. (2000) Cell surface engineering
by a modified staudinger reaction. Science 287, 2007ꢀ2010.
(3) King, M. and Wagner, A. (2014) Developments in the field of
bioorthogonal bond forming reactions − Past and present trends.
Bioconjugate Chem. 25, 825ꢀ839.
(4) Prescher, J. A. and Bertozzi, C. R. (2005) Chemistry in living
systems. Nat. Chem. Biol. 1, 13ꢀ21.
(5) Bertozzi, C. R. (2011) A decade of bioorthogonal chemistry.
Acc. Chem. Res. 44, 651ꢀ653.
(6) Ramil, C. P. and Lin, Q. (2013) Bioorthogonal chemistry:
strategies and recent developments. Chem. Commum. 49, 11007ꢀ
11022.
(7) Sletten, E. M. and Bertozzi, C. R. (2011) From mechanism to
mouse: a tale of two bioorthogonal reactions. Acc. Chem. Res. 44,
666ꢀ676.
(8) Matikonda, S. S., Orsi, D. L., Staudacher, V., Jenkins, I. A.,
Fiedler, F., Chen, J., and Gamble, A. B. (2015) Bioorthogonal
prodrug activation driven by a strainꢀpromoted 1,3ꢀdipolar
cycloaddition. Chem. Sci. 6, 1212ꢀ1218.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(9) Brakel, R. V., Vulders, R. C. M., Bokdam, R. J., Grüll, H.,
and Robillard, M. S. (2008) A doxorubicin prodrug activated by
the Staudinger reaction. Bioconjugate Chem. 19, 714ꢀ718.
(10) Xu, M., Tu, J., and Franzini, R. M. (2017) Rapid and
efficient tetrazineꢀinduced drug release from highly stable
benzonorbornadiene derivatives. Chem. Commun. 53, 6271ꢀ6274.
(11) Versteegen, R. M., Rossin, R., Ten Hoeve, W., Janssen, H.
M., and Robillard, M. S. (2013) Click to release: Instantaneous
doxorubicin elimination upon tetrazine ligation. Angew. Chem.
Int. Ed. 52, 14112ꢀ14116.
(12) Weiss, J. T., Dawson, J. C., Fraser, C., Rybski, W., Torresꢀ
Sánchez, C., Bradley, M., Patton, E. E., Carragher, N. O., and
UncitiꢀBroceta, A. (2014) Development and bioorthogonal
activation of palladiumꢀlabile prodrugs of gemcitabine. J. Med.
Chem. 57, 5395ꢀ5404.
(13) Weiss, J. T., Dawson, J. C., Macleod, K. G., Rybski, W.,
Fraser, C., TorresꢀSánchez, C., Patton, E. E., Bradley, M.,
Carragher, N. O., and UncitiꢀBroceta, A. (2014) Extracellular
palladiumꢀcatalysed dealkylation of 5ꢀfluoroꢀ1ꢀpropargylꢀuracil as
a bioorthogonally activated prodrug approach. Nat. Commun. 5,
3277.
(14) Ritter, C., Nett, N., AcevedoꢀRocha, C. G., Lonsdale, R.,
Kraling, K., Dempwolff, F., Hoebenreich, S., Graumann, P. L.,
Reetz, M. T., and Meggers, E. (2015) Bioorthogonal enzymatic
activation of caged compounds. Angew. Chem. Int. Ed. 54, 13440ꢀ
13443.
Thiazolyl Blue Tetrazolium Bromide (MTT) (Sigma Alꢀ
drich) was freshly dissolved in PBS at 5 mg/mL, passed
through a 0.22 µm filter, and 10 µL were added to each well.
After an incubation period of 4 h, the medium was gently reꢀ
moved. The formed formazan crystals were dissolved in 100
µL DMSO (Sigma Aldrich), subsequently the absorbance was
measured with a plate reader (PolarStar Omega, BMG Labꢀ
tech)) at 570 nm. As a reference wavelength (background),
690 nm was chosen and subtracted from the absorbance at 570
nm. The cytotoxicity assay was performed in at least three
independent experiments, with duplicate experiments on each
plate (i.e. 2 x 3 wells per drug/prodrug concentration in each
of the independent experiments for a total of n ≥ 3. IC50 values
were calculated using the normalized (variable) doseꢀresponse
parameters in GraphPad Prism 7 software (Figure S10).
ꢀ ASSOCIATED CONTENT
Supporting Information. Synthetic procedures and characterizaꢀ
tion of compounds 1bꢀ1f, 2b, 2d, 2e, 11a, 11d and all intermediꢀ
ates, 1H and 13C NMR spectra of final compounds, supplementary
figures and tables, coordinates and energies of transition states.
This material is available free of charge via the Internet at
http://pubs.acs.org.
ꢀ AUTHOR INFORMATION
(15) Van de Bittner, G. C., Bertozzi, C. R., and Chang, C. J.
(2013) Strategy for dualꢀanalyte luciferin imaging: In vivo
bioluminescence detection of hydrogen peroxide and caspase
activity in a murine model of acute inflammation. J. Am. Chem.
Soc. 135, 1783ꢀ1795.
(16) Gorska, K., Manicardi, A., Barluenga, S., and Winssinger,
N. (2011) DNAꢀtemplated release of functional molecules with an
azideꢀreductionꢀtriggered immolative linker. Chem. Commun. 47,
4364ꢀ4366.
Corresponding Author
*Email: allan.gamble@otago.ac.nz
Notes
The authors declare no competing financial interest.
Funding Sources
This research was supported by a contract from the Health Research
Council of New Zealand (A.B.G).
10
ACS Paragon Plus Environment

Page 11 of 13
Bioconjugate Chemistry
(17) Zawilska, J. B., Wojcieszak, J., and Olejniczak, A. B. (2013)
Prodrugs: a challenge for the drug development. Pharmacol. Rep.
65, 1ꢀ14.
(18) Lang, K. and Chin, J. W. (2014) Bioorthogonal reactions for
labeling proteins. ACS Chem. Biol. 9, 16ꢀ20.
(19) Selvaraj, R. and Fox, J. M. (2013) transꢀCyclooctene — a
stable, voracious dienophile for bioorthogonal labeling. Curr.
Opin. Chem. Biol. 17, 753ꢀ760.
(20) Xie, S., Lopez, S. A., Ramström, O., Yan, M., and Houk, K.
N. (2015) 1,3ꢀDipolar cycloaddition reactivities of perfluorinated
aryl azides with enamines and strained dipolarophiles. J. Am.
Chem. Soc. 137, 2958ꢀ2966.
(35) Chow, A. H., Chiu, F. C., Damani, L. A., and Kadaba, P. K.
(1996) Triazolines. 34. Structure and stability relationships of
novel ꢁ2ꢀ1,2,3ꢀtriazoline anticonvulsants. Pharm. Res. 13, 179ꢀ
182.
(36) Morgenthaler, M., Schweizer, E., HoffmannꢀRöder, A.,
Benini, F., Martin, R. E., Jaeschke, G., Wagner, B., Fischer, H.,
Bendels, S., Zimmerli, D., et al. (2007) Predicting and tuning
physicochemical properties in lead optimization: Amine
basicities. ChemMedChem 2, 1100ꢀ1115.
1
2
3
4
5
6
7
8
(37) Liu, S., Schauer, C. K., and Pedersen, L. G. (2009)
Molecular acidity: A quantitative conceptual density functional
theory description. J. Chem. Phys. 131, 3648.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(21) Sundhoro, M., Jeon, S., Park, J., Ramström, O., and Yan, M.
(2017) Perfluoroaryl azide Staudinger reaction: A fast and
bioorthogonal reaction. Angew. Chem. Int. Ed. 56, 12117ꢀ12121.
(22) Sustmann, R. and Trill, H. (1972) Substituent effects in 1,3ꢀ
dipolar cycloadditions of phenyl azide. Angew. Chem. Int. Ed. 11,
838ꢀ840.
(23) Sustmann, R. (1971) A simple model for substituent effects
in cycloaddition reactions. I. 1,3ꢀdipolar cycloadditions.
Tetrahedron Lett. 12, 2717ꢀ2720.
(24) Jewett, J. C. and Bertozzi, C. R. (2010) Cuꢀfree click
cycloaddition reactions in chemical biology. Chem. Soc. Rev. 39,
1272ꢀ1279.
(25) Shea, K. J. and Kim, J. S. (1992) Influence of strain on
chemical reactivity. Relative reactivity of torsionally strained
double bonds in 1, 3ꢀdipolar cycloadditions. J. Am. Chem. Soc.
114, 4846ꢀ4855.
(26) Cases, M., Duran, M., Mestres, J., Martín, N., and Solà, M.
(2001) Mechanism of the addition reaction of alkyl azides to
[60]fullerene and the subsequent N2 Extrusion to form
monoiminoꢀ[60]fullerenes. J. Org. Chem. 66, 433ꢀ442.
(27) Gillis, E. P., Eastman, K. J., Hill, M. D., Donnelly, D. J., and
Meanwell, N. A. (2015) Applications of fluorine in medicinal
chemistry. J. Med. Chem. 58, 8315ꢀ8359.
(28) Zhou, Y., Wang, J., Gu, Z., Wang, S., Zhu, W., Acena, J. L.,
Soloshonok, V. A., Izawa, K., and Liu, H. (2016) Next generation
of fluorineꢀcontaining pharmaceuticals, compounds currently in
Phase IIꢀIII clinical trials of major pharmaceutical companies:
New structural trends and therapeutic areas. Chem. Rev. 116, 422ꢀ
518.
(29) Meanwell, N. A. (2011) Synopsis of some recent tactical
application of bioisosteres in drug design. J. Med. Chem. 54,
2529ꢀ2591.
(30) Böhm, H. J., Banner, D., Bendels, S., Kansy, M., Kuhn, B.,
Müller, K., ObstꢀSander, U., and Stahl, M. (2004) Fluorine in
medicinal chemistry. ChemBioChem 5, 637ꢀ643.
(38) Matikonda, S. S., Fairhall, J. M., Tyndall, J. D. A., Hook, S.,
and Gamble, A. B. (2017) Stability, kinetic, and mechanistic
investigation of 1,8ꢀselfꢀimmolative cinnamyl ether spacers for
controlled release of phenols and generation of resonance and
inductively stabilized methides. Org. Lett. 19, 528ꢀ531.
(39) Schmid, K. M., Jensen, L., and Phillips, S. T. (2012) A selfꢀ
immolative spacer that enables tunable controlled release of
phenols under neutral conditions. J. Org. Chem. 77, 4363ꢀ4374.
(40) Hamadouche, M., GaudelꢀSiri, A., Pons, J. M., and Abed, D.
E. (2010) Relative stability of a series of 1,2,3ꢀtriazolines.
Theoretical study of substituent effects. J. Mol. Struc.:
THEOCHEM 956, 33ꢀ37.
(41) Bodnar, B. S. and Miller, M. J. (2007) Reactions of nitroso
heteroꢀDiels−Alder cycloadducts with azides:ꢃ Stereoselective
formation of triazolines and aziridines. J. Org. Chem. 72, 3929ꢀ
3932.
(42) Contini, A. and Erba, E. (2012) Clickꢀchemistry approach to
azacycloalkene
monosulfonyl
diamines: synthesis and
computational analysis of the reaction mechanism. RSC Adv. 2,
10652ꢀ10660.
(43) Wladkowski, B. D., Smith, R. H., and Michejda, C. J. (1991)
Theoretical investigation of the protonꢀinduced decomposition of
4,5ꢀdihydroꢀ1,2,3ꢀtriazole to form the aziridinium ion: instability
of the (2ꢀaminoethyl)diazonium ion. J. Am. Chem. Soc. 113,
7893ꢀ7897.
(44) Audran, G., Adiche, C., Brémond, P., El Abed, D.,
Hamadouche, M., Siri, D., and Santelli, M. (2017) Cycloaddition
of sulfonyl azides and cyanogen azide to enamines. Quantumꢀ
chemical calculations concerning the spontaneous rearrangement
of the adduct into ringꢀcontracted amidines. Tetrahedron Lett. 58,
945ꢀ948.
(45) Chung, T. S., Xue, Y., Carranza, A., and GarciaꢀGaribay, M.
A. (2017) Stereospecific photochemistry of ꢁ2ꢀ1,2,3ꢀtriazolines in
solution and in the solid state: scope and mechanistic studies.
Photochem. Photobiol. Sci. 16, 1458ꢀ1463.
(46) Chen, Y. J., Hung, H. C., Sha, C. K., and Chung, W. S.
(2010) Photochemistry of benzene and quinoxaline fused
ꢁ2−1,2,3−triazolines and their trapping products. Tetrahedron 66,
176ꢀ182.
(47) Bailey, A. S. and Wedgwood, J. J. (1968) The reaction of
picryl azide with olefins. J. Chem. Soc. (C), 682ꢀ684.
(48) Shao, Y., Molnar, L. F., Jung, Y., Kussmann, J., Ochsenfeld,
C., Brown, S. T., Gilbert, A. T. B., Slipchenko, L. V., Levchenko,
S. V., O'Neill, D. P., et al. (2006) Advances in methods and
algorithms in a modern quantum chemistry program package.
Phys. Chem. Chem. Phys. 8, 3172ꢀ3191.
(49) Halgren, T. A. (1996) Merck molecular force field. I. Basis,
form, scope, parameterization, and performance of MMFF94. J.
Comput. Chem. 17, 490ꢀ519.
(31) Senecal, T. D., Shu, W., and Buchwald, S. L. (2013) A
general, practical palladiumꢀcatalyzed cyanation of (hetero)aryl
chlorides and bromides. Angew. Chem. Int. Ed. 52, 10035ꢀ10039.
(32) Inoue, T., Matsuyama, K., and Inoue, Y. (1999)
Diastereodifferentiating
ZꢀE
photoisomerization
of
3ꢀ
benzoyloxycyclooctene:ꢃ Diastereoselectivity switching controlled
by substrate concentration through competitive intraꢀ vs
intermolecular photosensitization processes. J. Am. Chem. Soc.
121, 9877ꢀ9878.
(33) Royzen, M., Yap, G. P., and Fox, J. M. (2008) A
photochemical synthesis of functionalized transꢀcyclooctenes
driven by metal complexation. J. Am. Chem. Soc. 130, 3760ꢀ3761.
(34) Alouane, A., Labruère, R., Le Saux, T., Aujard, I.,
Dubruille, S., Schmidt, F., and Jullien, L. (2013) Light activation
for the versatile and accurate kinetic analysis of disassembly of
selfꢀimmolative spacers. Chem. Eur. J. 19, 11717ꢀ11724.
(50) Frisch, M. J. T., G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.;
11
ACS Paragon Plus Environment

Bioconjugate Chemistry
Page 12 of 13
Mennucci, B.; Petersson, G. A., et al. (2009), Gaussian 09,
revision D.01, Gaussian, Inc., Wallingford CT.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12
ACS Paragon Plus Environment

Page 13 of 13
Bioconjugate Chemistry
TOC Graphic:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
ACS Paragon Plus Environment