A Critical Cross-Catalytic Relationship Determines the Outcome of Competition in a Replicator Network

Article abstract of DOI:10.1021/jacs.7b06270

A network of two synthetic replicators exhibits a critical unidirectional cross-catalytic relationship that directs competing replication processes. In this network, nitrone N bearing a 6-methylamidopyridine recognition site can participate in 1,3-dipolar cycloaddition reactions with two maleimides that differ in the relative position of their carboxylic acid recognition site: either para (M^p) or meta (M^m) relative to the maleimide ring. These cycloaddition reactions create replicators trans-T^p and trans-T^m. In isolation, trans-T^p templates its own formation with an efficiency that is markedly greater than that of trans-T^m. Kinetic fitting and simulations reveal that this efficiency arises from a higher template-mediated rate constant for the cycloaddition and lower stability of the trans-T^p template duplex, compared to trans-T^m. By contrast, in a situation where M^p and M^m compete for a limited quantity of N, the normally less efficient trans-T^m outcompetes trans-T^p. Through a series of comprehensive kinetic ¹⁹F{¹H} NMR spectroscopy experiments, this system-level outcome is traced to a critical cross-catalytic pathway, whereby the presence of trans-T^p templates the formation of trans-T^m, but not vice versa. Replicator trans-T^m also reduces the efficiency of its competitor trans-T^p by sequestering trans-T^p in a heteroduplex that is more stable than homoduplex [T^p·T^p]. The addition of different templates as instructions reveals that, while the outcome of competition between replicators can be altered selectively, it is limited by the reaction environment employed. These results represent a conceptual and practical framework for the examination of selectivity in replication networks operating outside well-stirred batch reactor conditions.

Full text of DOI:10.1021/jacs.7b06270

Subscriber access provided by Warwick University Library
Article
A critical crosscatalytic relationship determines
the outcome of competition in a replicator network
Tamara Kosikova, and Douglas Philp
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b06270 • Publication Date (Web): 04 Aug 2017
Downloaded from http://pubs.acs.org on August 4, 2017
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.
Journal of the American Chemical Society 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 31
Journal of the American Chemical Society
1
2
3
Journal of the American Chemical Society
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
A critical crosscatalytic relationship determines the
outcome of competition in a replicator network
Tamara Kosikova and Douglas Philp*
School of Chemistry and EaStCHEM, University of St Andrews, North Haugh,
St Andrews, Fife KY16 9ST, United Kingdom
*Corresponding author eꢀmail: d.philp@stꢀandrews.ac.uk
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
ORCID ID: TK: orcid.org/0000ꢀ0001ꢀ7886ꢀ9660
DP: orcid.org/0000ꢀ0002ꢀ9198ꢀ4302
1
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 31
1
2
3
Abstract
4
5
6
7
8
A network of two synthetic replicators exhibits a critical unidirectional crosscatalytic relationship that
directs competing replication processes. In this network, nitrone N bearing a 6ꢀmethylamidopyridine
recognition site can participate in two 1,3ꢀdipolar cycloaddition reactions with two maleimides that
differ in the relative position of their carboxylic acid recognition site — either para (Mp) or meta (Mm)
relative to the maleimide ring. These two cycloaddition reactions create replicators transꢀTp and trans-
Tm. In isolation, transꢀTp templates its own formation with an efficiency that is markedly greater than
that of transꢀTm. Kinetic fitting and simulations reveal that this efficiency arises from a higher
templateꢀmediated rate constant for the cycloaddition and lower stability of the transꢀTp template
duplex, compared to trans-Tm. By contrast, in a situation where Mp and Mm compete for a limited
quantity of N, the normally less efficient trans-Tm outcompetes trans-Tp. Through a series of
comprehensive kinetic 19F{1H} NMR spectroscopy experiments, this systemꢀlevel outcome is traced to
a critical crosscatalytic pathway, whereby the presence of trans-Tp templates the formation of trans-
Tm, but not vice versa. Replicator trans-Tm also reduces the efficiency of its competitor trans-Tp by
sequestering trans-Tp in a heteroduplex that is more stable than homoduplex [Tp•Tp]. The addition of
different templates as instructions reveals that, while the outcome of competition between replicators
can be altered selectively, it is limited by the reaction environment employed. These results represent a
conceptual and practical framework for the examination of selectivity in replication networks operating
outside wellꢀstirred batch reactor conditions.
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
2
ACS Paragon Plus Environment

Page 3 of 31
Journal of the American Chemical Society
1
2
3
Introduction
Networks are the building blocks of the world around us, and their inherent interconnectedness1 often
exerts an influence over our everyday lives that can be challenging to predict. The allꢀpervasive
presence of networks in nature2 is reflected in the number of disciplines currently engaged1c,3 in the
study of complex systems3a,4 and the properties that emerge through the interactions of their
components. Chemistry experienced a paradigm shift around twenty years ago as it moved away from
the study of molecular matter in isolation and, instead, began to embrace the notion of complexity and
networks. By embracing the nascent field5 of systems chemistry, bottomꢀup approaches for the design,
development, and investigation of synthetic chemical models for complex systems have started to
appear. A phenomenon of particular interest within the field of systems chemistry is replication. In the
context of complex systems, synthetic replicators present a unique opportunity to study these processes
using networks constructed from molecules with wellꢀdefined structures and with catalytic and
recognition properties that can be probed and characterized in detail experimentally. The developments
in the field of systems chemistry to date have produced a remarkable variety of replicating systems
based on oligonucleotides,6 peptides,7 and small synthetic molecules,8 highlighting and demonstrating
that the ability to replicate is not exclusive to biological systems based on nucleic acids and equipped
with extensive enzymatic machinery. In addition to selfꢀreplication, these systems were shown to
express a number of systemꢀlevel properties such as errorꢀcorrection,9 stereoꢀspecific replication,10 and
Boolean logic.11 Peptideꢀbased replicating systems have, in particular, achieved a significant level of
sophistication, examining networks9,10,11c,12 comprising more than a single replicator—a feature
significantly less well explored in replicating systems based on oligonucleotides13 and smallꢀorganic
molecules.14 Although the requirements for the operation of minimal selfꢀreplicating systems in
isolation have been established15 firmly, the processes in complex networks in the real world do not
operate in isolation. For this reason, in this work, we build on our experience in developing selfꢀ
replicating systems based on small organic molecules8a,8c,8d,8f,8g,14a–c to design and analyze a network of
two replicators, connected directly through a shared building block. This network presents a model
system that allows us to study how reaction and recognitionꢀmediated processes impact and govern the
preference, i.e., the selectivity, of the network for one replicator over another.
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
3
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 31
1
2
3
Design and methodology
In order to create a minimal selfꢀreplicating system, two components, bearing complementary
recognition sites, must be connected together by one or more covalent bonds to afford a template that
can act as a specific autocatalyst for its own formation. Once formed, this autocatalytic template can
preꢀorganize the building blocks required for its formation in a ternary complex coꢀconformation that
renders their reaction pseudointramolecular. The bondꢀforming step produces a template dimer, or
duplex, which exist in equilibrium with the catalytically active monomeric template.
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
Here, a network of two replicators is constructed from two maleimides, Mp and Mm, and a
single nitrone N[Ref. 14c] (Figure 1a, Building blocks). The main distinguishing feature of these two
maleimides and the replicators16 Tp[Ref. 14c] and Tm (Figure 1a, Replicators) arising from their 1,3ꢀ
dipolar cycloaddition reactions with N, is in the location of the carboxylic acid recognition site on the
maleimide component (Figure 1). In particular, Mp integrates8a,8c,14a–c a phenylacetic acid in position
para relative to the maleimide ring, while Mm incorporates a benzoic acid functionality, in position
meta with respect to the maleimide ring. For this reason, we use superscript labels p and m to represent
the para and meta positions of the carboxylic acid recognition site in maleimides (M) and templates
(T) throughout this work. Initially, these two templates are formed via slow, templateꢀindependent
bimolecular pathways (Figure 1c), typically furnishing two diastereoisomeric products, labeled16 trans
and cis, in a 3:1 ratio in the absence effects originating from molecular recognition. Once the quantity
of template formed is sufficient to allow its association with the unreacted components in a
catalytically active ternary complex, [N•Mp•Tp] or [N•Mm•Tm] (Figure 1b), the template formation
proceeds with significant acceleration via the autocatalytic cycle. The ability of a template to initiate an
autocatalytic cycle depends on the strength of the nonꢀcovalent recognitionꢀmediated processes (Figure
1d) that govern the recognition between the network components. For example, a replicator with a
stronger17 recognition process would typically require a lower template concentration to enable selfꢀ
replication to proceed.
Replicators Tp and Tm possess a recognition unit complementary to the carboxylic acid,
namely, a 6ꢀmethylꢀ2ꢀamidopyridine ring that is originally associated with nitrone N (Figure 1a). The
relative positions and identities of recognition sites are conserved in the replicator templates. We
envisaged that the slight differences in the type and location of the carboxylic acid recognition
elements in each system would produce a reaction network of two replicators, referred to as the Tp–Tm
network, with the potential for crosscatalytic behavior. Hence, reaction of the three individual
components together allows the formation of a network where both templates can form simultaneously
(Figure 1b) through two autoꢀ and two crosscatalytic pathways. The ability of the templates Tp and Tm
4
ACS Paragon Plus Environment

Page 5 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
to act as either autoꢀ or crosscatalytic templates allows the network to be instructed to form one
replicator over another by the addition of preformed Tp or Tm. Ultimately, the ability of our replicators
to compete for the shared nitrone building block will depend on their ability to take part in these autoꢀ
and crosscatalytic cycles and for this reason, we wished to undertake a comprehensive kinetic analysis
of this small network constructed from two selfꢀreplicators to develop an understanding of how
structural changes alter the selectivity for one replicator over another and their role in driving systemꢀ
level behavior.
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
Figure 1.
(a) A network of two replicators, Tp and Tm, can be constructed using the 1,3ꢀdipolar cycloaddition
reactions of two maleimides, Mp and Mm, and nitrone N. Superscript labels p and m represent the para and meta location of
the carboxylic acid recognition site. (b) Maleimides Mp and Mm can react with nitrone N to form two replicators, Tp and Tm
(central square). Each of these templates can theoretically take part in autoꢀ and crosscatalytic pathways. (c) Maleimides Mp
and Mm can react with nitrone N also via templateꢀindependent pathways, resulting in the formation of catalytically active
trans diastereoisomers and catalytically inert cis diastereoisomers of Tp and Tm. (d) The recognition between the network
components is governed by specific association constants.
5
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 31
1
2
3
4
Results and discussion
The first step in the analysis of the Tp–Tm network is the examination of individual bimolecular
recognitionꢀindependent pathways leading to Tp and Tm. Accordingly, we examined (Figure 2) the
reaction of nitrone N with two control, recognitionꢀdisabled maleimides, MpC and MmC. Details of the
synthetic procedures utilized in the syntheses of all reaction components are provided in the Supporting
Information. The kinetic analyses of all reaction pathways, facilitated by the presence of the aryl
fluorine atom present in nitrone N, and all Nꢀderived cycloadducts, were performed using 470.3 MHz
19F{1H} NMR spectroscopy. In all kinetic experiments reported in this study, one or both of the
maleimides and the nitrone ([M] = [N] = 5 mM) were allowed to react in CDCl3 at 5 °C, and the
reaction progress was monitored by recording 470.3 MHz 19F{1H} NMR spectra every 30 minutes over
a period of 16 h. The concentrations of the reaction components were determined at each time point
relative to 1ꢀbromoꢀ2ꢀfluoroꢀ4ꢀnitrobenzene, which was present as an internal standard.
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
Figure 2.
Kinetic experiments examining the reaction of N with recognitionꢀdisabled maleimides (a) MpC, and (b)
MmC, and recognitionꢀenabled maleimides (c) Mp, and (d) Mm, as determined by 470.3 MHz 19F{1H} NMR spectroscopy
relative to 1ꢀbromoꢀ2ꢀfluoroꢀ4ꢀnitrobenzene as an internal standard. Formation of recognitionꢀdisabled cycloadducts transꢀ
TpC (empty red diamonds), transꢀTmC (empty blue diamonds) and the corresponding cis cycloadducts (filled black
diamonds) in (a) and (b) proceeds slowly and unselectively. (c) Formation of replicator transꢀTp (no cis product was
detected) was examined in the absence of template (empty red circles), and in the presence of transꢀTp (filled red circles)
and transꢀTm (filled blue circles). (d) Formation of replicator transꢀTm (cisꢀTm formed <0.5 mM in each case, data omitted
for clarity) was examined in the absence of template (empty blue circles) and in the presence of transꢀTm (filled blue
circles) and transꢀTp (filled red circles). Graphs in (e) and (f) show the percentage enhancement factors (%EFs) determined
in templateꢀinstructed ([T]i) experiments (filled symbols in (c) and (d)) relative to the corresponding uninstructed
experiments ([T]u) (empty symbols in (c) and (d)). Reaction conditions: [all reactants] = 5 mM, 5 °C, CDCl3, 16 h. In each
case, instructing template was added at 20 mol% (1 mM). Note that the vertical scale in (a) to (d) is identical in order to
facilitate direct comparison of reaction conversions.
6
ACS Paragon Plus Environment

Page 7 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The control maleimides do not possess the carboxylic acid recognition site required for successful
recognition of the amidopyridine unit on nitrone component N and the absence of recognitionꢀmediated
reactivity is clearly evident in the reaction profiles determined for the reaction of MpC with N (Figure
2a), and MmC with N (Figure 2b). The formation of both recognitionꢀdisabled trans and cis
diastereoisomers of TpC and TmC proceeded very slowly. After 4 h, transꢀTpC reached 0.03 mM
(Figure 3a), while cisꢀTpC was below the limit of detection at this time. After 16 h, the concentration
of these two diastereoisomers increased to 0.14 mM and 0.04 mM for trans and cis, respectively
([trans]/[cis] = 3.8). By contrast, the formation of TmC proceeded slightly more efficiently, and the
trans and cis cycloadducts reached concentrations of 0.07 mM and 0.02 mM, respectively, after 4 h. At
t = 16 h, their concentration increased further to 0.24 mM and 0.10 mM ([trans]/[cis] = 2.5)—almost
twice as high as those determined for TpC. These results demonstrated that (i) in the absence of the
carboxylic acid recognition site, the formation of templates can proceed only via the slow templateꢀ
independent pathways, and (ii) formation of TmC proceeds noticeably faster than TpC, suggesting that
replicator Tm might be able to exploit the templateꢀmediated cycle at an earlier time point in the
reaction that Tp.
Next, we examined the formation of replicators Tp (Figure 2c, empty red symbols) and Tm
(Figure 2d, empty blue symbols), from their constituent components. Both reaction profiles are
sigmoidal, demonstrating the striking effect of recognitionꢀmediated replication processes on product
formation. At t = 4 h, replicator Tp was formed at a concentration of 0.53 mM, with complete
diastereoselectivity for the trans cycloadduct (i.e., cis < limit of detection by 19F{1H} NMR
spectroscopy). This concentration increased further to 3.34 mM after 16 h. The reaction that forms Tm
was shown to be less diastereoselective, affording transꢀTm and cisꢀTm at 0.30 mM and 0.13 mM after
4 h, respectively. After 16 h, the concentration of these cycloadducts increased to 2.10 mM and 0.40
mM. The ratio of the two diastereoisomer products ([trans]/[cis]) increased from 2.3 at 4 h to 5.2 after
16 h.
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
In the next stage of our analysis of this system, the formation of each replicator was examined
in the presence of 20 mol% (1 mM) of the respective trans template (e.g., Tp added to Mp and N) in
order to test whether Tp and Tm possess the capacity to template their own formation. The reaction
profile for the formation of Tp in the presence of Tp (Figure 2c, filled red symbols) and Tm in the
presence of Tm (Figure 2d, filled blue symbols) revealed disappearance of the lag period in each case,
thereby confirming the abilities of Tp and Tm to selfꢀreplicate. In the presence of the autocatalytic
template Tp, formation of Tp proceeded with significantly increased efficiency, reaching a
7
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 31
1
2
3
4
5
6
7
8
9
concentration of 1.26 mM after 4 h, and 3.69 mM after 16 h. As expected, the addition of preformed
template induces the most striking enhancement in the concentration of replicator at early reaction
times. The enhancement in the rate of formation of transꢀTm was, in contrast, less pronounced. The
concentration of transꢀTm reached 0.86 mM and 2.82 mM after 4 and 16 h, respectively. Catalyticallyꢀ
inert cisꢀTm reached a concentration of 0.09 mM and 0.32 mM after 4 and 16 h, giving [trans]/[cis]
ratios of 10.1 and 8.1, respectively.
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
Thus far, the kinetic analyses confirmed that both Tp and Tm are capable of selfꢀreplication.
Next, we examined the two crosscatalytic pathways—Tm acting as a template for the formation of Tp
(Figure 2c, filled blue symbols) and vice versa (Figure 2d, filled red symbols). To this end, each set of
reactants was instructed with 20 mol% of the corresponding crosscatalytic template. Interestingly, the
addition of Tm to the reactants required for the formation of Tp resulted in no observable enhancement
in the concentration of Tp. In fact, no change in the lag period for Tp formation was observed and the
reaction profile determined in the presence of Tm was essentially identical to that obtained in the
absence of instruction. In terms of concentration, cycloadduct transꢀTp reached 0.43 mM after 4 h—
similar to that (0.53 mM) reached in the absence of any instructing template. After 16 h, this
concentration increased to 3.15 mM. Strikingly, however, instructing the building blocks required for
the formation of Tm with template Tp resulted in the unambiguous disappearance of the lag period,
demonstrating that Tp can template the formation of transꢀTm. After 4 h, transꢀTm and cisꢀTm were
formed at concentrations of 0.87 mM and 0.08 mM, respectively. These concentrations increased
further after 16 h, reaching 2.52 mM and 0.22 mM, respectively ([trans]/[cis] ratios of 10.3 and 11.4).
The slight increase in diastereoselectivity for the trans product in the presence of preformed Tp
compared to Tm suggests that template Tp, which locates the two recognition sites at a slightly different
separation and geometry with respect to each other than Tm, can promote the formation of the trans
diastereoisomer more effectively than Tm.
In order to quantify the changes in replication efficiency observed in the templateꢀinstructed
experiments relative to those lacking any instruction, we calculated the percentage enhancement
factor18 (%EF) for each trans product determined after both 4 h and 16 h (Figure 2). Examination of
these results for Tp (Figure 2e) and Tm (Figure 2f) reveals a number of trends. Firstly, the %EFs are
significantly higher after 4 h than after 16 h, reflecting the fact that the presence of preformed template
exerts a stronger effect at earlier reaction times when the overall concentration of catalytically active
template within the uninstructed reaction mixture is typically low. Further, Figure 2e shows that the
addition of Tp to the building blocks required for its own formation enhanced its formation
significantly: 138 ± 7% at 4 h, and 10 ± 1% at 16 h. In contrast, the addition of Tm to the same mixture
8
ACS Paragon Plus Environment

Page 9 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
of building blocks resulted in a suppression of Tp formation (4 h = −19 ± 5%, 16 h = −6 ± 1%) relative
to the uninstructed reaction. This observation suggests that Tp might be sequestered within the [Tp•Tm]
heteroduplex, which reduces the quantity of free template available for reaction. The same analysis of
the enhancements determined for Tm reveals that the addition of autocatalytic template Tm (4 h = 185 ±
15%, 16 h = 36 ± 2%) affords a marginally higher enhancement than the addition of crosscatalytic
template Tp (4 h = 187 ± 16%, 16 h = 22 ± 1%). Overall, a comparison of the EFs showed that the
addition of preformed template produces a more significant change in the formation of Tm than in the
formation of Tp. Most likely, this difference can be attributed to the fact that replicator Tm is less
efficient at making itself than Tp in the absence of template and is therefore affected less by the
decreasing availability of the reagents required for its formation at longer reaction times. In summary,
the comprehensive evaluation of all autoꢀ and crosscatalytic pathways available to Tp and Tm confirmed
that only three catalytic channels out of the four possible operate efficiently.
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
Kinetic fitting and simulations
In order to develop a better understanding of the differences in the kinetic behaviors of Tp and Tm in
terms of the key kinetic and thermodynamic parameters that characterize the autoꢀ and crosscatalytic
cycles, we fitted the reaction time courses shown in Figure 3 to appropriate kinetic models (see
Supporting Information). These kinetic models describe the interactions and reactions leading to the
formation of Tp or Tm and using established protocols,8a,13b,14a we were able to determine the
bimolecular reaction rate constants (kbi) for the formation of both trans and cis cycloadducts, the
unimolecular rate constants (kauto, kcross), and duplex association constants (KaDuplex) pertaining to all
autoꢀ and crosscatalytic pathways (Table 1).
Table 1.
Overview of rate and duplex association constants determined using SimFit13b for the bimolecular and
recognitionꢀmediated autoꢀ and crosscatalytic pathways of replicators trans-Tp and trans-Tm. The rate constants for the
formation of catalytically inactive cis diastereoisomers of Tp and Tm through the corresponding bimolecular templateꢀ
independent pathways were determined to be 0.431 and 0.704 10–4 M–1 s–1, respectively.
Reactions
N + Mp → trans-Tp
k
EMkinetic / M
1.12
1.79
71.6
17.0
4.67
20.3
—
Bimolecular
kbi / 10−4 M−1 s−1
N + Mm → trans-Tm
—
[N•Mp•trans-Tp] → [trans-Tp•trans-Tp]
[N•Mm•trans-Tm] → [trans-Tm•trans-Tm]
[N•Mp•trans-Tm] → [trans-Tp•trans-Tm]
[N•Mm•trans-Tp] → [trans-Tp•trans-Tm]
64.0
9.47
4.21
18.3
Autocatalytic
kauto / 10−4 s−1
Crosscatalytic
kcross / 10−4 s−1
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 31
1
2
Equilibria
N + Mp ↔ [N•Mp]
Ka
EMthermo / M
3
4
5
6
7
8
9
830
—
Binary
KaInd / M−1
N + Mm ↔ [N•Mm]
3320
17.9
68.6
35.0
—
trans-Tp + trans-Tp ↔ [trans-Tp•trans-Tp]
trans-Tm + trans-Tm ↔ [trans-Tm•trans-Tm]
trans-Tp + trans-Tm ↔ [trans-Tp•trans-Tm]
26.0
6.22
12.7
Template
KaDuplex / 106 M−1
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
These parameters, together with association values for the formation of [M•N] complexes (KaInd) were
subsequently used to calculate the kinetic effective molarity19 (EMkientic) and thermodynamic effective
molarity20 (EMthermo) for each replication pathway. Examination of the parameters shown in Table 1
reveals that both trans and cis diastereoisomeric products of Tm are formed with a higher bimolecular
rate constant their Tp counterparts. The initial advantage provided to Tm replicator as a result of the
higher bimolecular rate (Tm should reach the concentration required for initiating autocatalysis earlier
than Tp) is countered by the lower kauto for the formation of Tm. Comparison of the kinetic profiles and
values of EMkinetic determined for Tp and Tm confirmed that, in isolation, Tp is significantly more
efficient at directing its own formation than Tm. The fitting procedure allowed the determination of the
association constants21 for replicator homoduplexes for transꢀTp (17.9 × 106 M−1) and transꢀTm (68.6
× 106 M−1), respectively. The higher stability of the [Tm•Tm] duplex relative to the [Tp•Tp] duplex
indicates that Tm is hindered more significantly by product inhibition than Tp. Interestingly, while Tm
is associated with a higher duplex stability, it possesses lower EMthermo than Tp. This observation
suggests a lower degree of complementarity and, hence, cooperativity within the Tm duplex.
The fitting procedure employed in the determination of kinetic parameters pertaining to the two
crosscatalytic pathways required the introduction of an additional parameter—i.e., the association
constant for the formation of heteroduplex [Tp•Tm]. This heteroduplex contains one more rotatable
bond than [Tm•Tm] and one rotatable bond less than [Tp•Tp] and therefore, its stability should be
bracketed by the stabilities of the two homoduplexes. Applying rotor increments, as described18a by
Page and Jencks, the stability of [Tp•Tm] was estimated as 35.0 × 106 M−1, and this value was
employed in the determination of kcross for the formation of each trans product on the corresponding
crosscatalytic template (Table 1).
The parameters obtained through the kinetic fitting of crosscatalytic kinetic data allowed us to
assess the differences observed in the corresponding time course profiles of both replicators.
Specifically, Tm was found to be significantly worse at catalyzing the formation of Tp than Tp was at
catalyzing its own formation. In fact, the EMkinetic for this crosscatalytic pathway was ca. 15 × smaller
than that determined for the autocatalytic pathway leading to Tp. By contrast, the EMkinetic for the
10
ACS Paragon Plus Environment

Page 11 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
crosscatalytic pathway that involves Tm being made on Tp was actually somewhat higher (18.3 M) than
the corresponding autocatalytic process (9.47 M). These differences in the relative efficiencies of the
autoꢀ and crosscatalytic pathways directed by Tp and Tm can explain the absence of a template effect
for Tp being synthesized on a Tm template.
9
An important consequence of the order of stabilities determined for the duplexes in the full
network ([Tm•Tm] > [Tp•Tm] > [Tp•Tp]), is the preferential incorporation of Tp within the more stable
heteroduplex [Tp•Tm] in situations where Tm is present at significant concentrations. The kinetic data
derived from NMR spectroscopy reveals that the exchange between the various template species
present in solution is fast on the NMR chemical shift timescale (i.e., only a single resonance arising
from the trans isoxazolidine proton is observed for each replicator). For this reason, we exploited
kinetic simulations and the relevant reaction parameters obtained through fitting to demonstrate how
the presence of Tm serves to decrease the amount of catalytically active Tp in solution. Specifically,
early in the reaction where Tp is being synthesized in the presence of preformed Tm (Figure 3a), the
higher stability of the [Tp•Tm] heteroduplex relative to that of the [Tp•Tp] homoduplex leads to the
template Tp formed within the system being sequestered into the more stable heteroduplex. The higher
proportion of Tp in the [Tp•Tm] heteroduplex than in the [Tp•Tp] homoduplex reduces the catalytic
availability of this template and, also, the overall rate of production of Tp because the [Tp•Tm]
heteroduplex has a lower Kd. At a certain point in the reaction, however, the concentration of Tp
reaches a critical threshold and the fraction of Tp bound in the [Tp•Tp] homoduplex becomes higher
than the fraction present within the [Tp•Tm] heteroduplex.
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
By contrast, this order of stabilities exerts the opposite effect on the crosscatalytic pathway
where the formation of Tm is simulated in the presence of Tp (Figure 3b). In this situation, transꢀTp is
the dominant product early in the reaction and therefore, the molecules of Tm formed through the
crosscatalytic pathway are present predominantly in the form of the heteroduplex [Tp•Tm]. This
heteroduplex is less stable than the [Tm•Tm] duplex that is present exclusively in the absence of Tp,
resulting in higher concentrations of monomeric of Tm at early time points. Nevertheless, the affinity of
Tp for Mm and N is lower than the affinity of Tm for these components, and as the concentration of Tm
continues to increase, homoduplex [Tm•Tm] becomes the dominant species in solution, and the
efficiency of the crosscatalytic pathway (Tp → Tm) decreases.
11
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 31
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
Figure 3.
Simulations predict the distribution of templates Tm and Tp in homoꢀ (full lines) and heteroduplexes
(dashed lines) in experiments where the reagents required for the formation of (a) Tp are instructed with 20 mol% of transꢀ
Tm (1 mM), and (b) the reagents for the formation of Tm are instructed with 20 mol% of transꢀTp (1 mM). Templates Tp
and Tm present within the various categories are represented in red and blue, respectively. Dotted lines represent the %
content of Tp and Tm in their monomeric form or in other binary complexes. Red and blue arrows highlight the time point in
the reaction at which the % content of the homoduplex of the template being formed is higher than its content in the
corresponding heteroduplex. (c) Schematic overview of the catalytic relationship between replicators Tp and Tm. Dashed
line signifies absence of effective catalysis.
Kinetic analysis of competition between replicators
Kinetic analyses of our two replicators in isolation established clearly that Tp possesses the capacity to
template both its own formation and to crosscatalyze the formation of Tm, while Tm is capable of
autocatalytic activity only. Building on these results, we examined these replicators under conditions
where Mp and Mm compete for the nitrone N building block. In order to explore the possibility of
directing the reaction network towards selective formation of either Tp or Tm, we employed a series of
templateꢀinstructed experiments. In each of these experiments, an equimolar solution (5 mM)
containing all three reaction components, N, Mp, and Mm, was prepared in CDCl3 and the formation of
the replicators was monitored at 5 °C in the absence of any added template using 470.3 MHz F{1H}
19
12
ACS Paragon Plus Environment

Page 13 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
NMR spectroscopy (see Supporting Information, Figure S8). The instructed experiments examined the
formation of Tp and Tm in the presence of 20 mol% of one or both preformed replicators. A number of
scenarios were examined — (i) only transꢀTp added, (ii) only transꢀTm added and (iii) both transꢀTp
and transꢀTm added simultaneously (see Supporting Information, Figure S8). The results of these
experiments are summarized in Figure 4. Interestingly, the uninstructed competition experiment
revealed that, as a result of the interplay between the recognition and reaction processes taking place
simultaneously in the network, the less efficient template transꢀTm reached a higher concentration than
transꢀTp, the replicator found to be more efficient in isolation ([transꢀTm]/[transꢀTp] = 2.0 after 4 h,
Figure 4a).
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
(a)
(b)
2.5
2.0
1.5
1.0
0.5
2.5
2.0
1.5
1.0
0.5
4 h
16 h
Tm
dominates
Tm
dominates
Tp
Tp
dominates
dominates
Increase in time
m
Experiments with no T and T : selectivity
p
m
T
T
p
p
m
Experiments with no T and both T and T : selectivity unchanged
Figure 4.
The effect of instructing the Tp–Tm network with preformed template (added at t = 0), on the ratio of
[Tm]/[Tp] determined after (a) 4 h (black circles) and (b) 16 h (black squares). Dashed lines represent the [Tm]/[Tp] ratio
determined in the uninstructed kinetic experiments, for comparison. Concentrations were determined using 470.3 MHz
19F{1H} NMR spectroscopy ([Mp] = [Mm] = [N] = 5 mM, if present, [template] = 1 mM, CDCl3, 5 °C).
In the presence of preformed transꢀTp, added to the reaction mixture at the start of the reaction, the
observed outcome was altered significantly—the instructing replicator, transꢀTp, outperformed transꢀ
Tm in the competition for N and the two products were formed in a ratio of 0.80 ([transꢀTm]/[transꢀTp]
after 4 h, Figure 5a). When preformed transꢀTm was added at the start of the reaction, the network
displayed the opposite behavior—i.e., formation of replicator transꢀTm was enhanced further ([transꢀ
Tm]/[transꢀTp] = 2.38 after 4 h, Figure 5a). Interestingly, when both transꢀTp and transꢀTm were used
to instruct the replication processes simultaneously, the two replicators were formed at a ratio of 0.99
([transꢀTm]/[transꢀTp]) after 4 h (Figure 5a). The simultaneous addition of both Tp and Tm enhances
the formation of transꢀTp relative to the uninstructed scenario, nevertheless transꢀTm remained the
dominant species in solution in this experiment after 4 h. This observation can be attributed to the fact
13
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 31
1
2
3
4
that the formation of replicator transꢀTm proceeds through two templateꢀmediated channels, whilst
replicator transꢀTp is formed through a single autocatalytic pathway only.
5
The ratios of the two replicators, [transꢀTm]/[transꢀTp], were reꢀexamined after 16 h, revealing
a trend similar to that observed for the %EFs discussed earlier for the autoꢀ and crosscatalytic pathways
in isolation. Specifically, the values of [transꢀTm]/[transꢀTp] at t = 16 h were determined to be 1.13,
0.80, 1.58, and 0.93 in the presence of no template, transꢀTp, transꢀTm, and both transꢀTp and transꢀ
Tm, respectively (Figure 5b). Comparison of the ratios determined after 4 and 16 h showed that the
selectivity in experiments instructed with Tp only changed less over time (4 h → 16 h) than the
selectivity in experiments employing no template or Tm. These differences in product ratio as a
function of time can be attributed to the fact that only a single crosscatalytic pathway is active in this
network and thus, while the presence of Tp affects the formation of both templates, addition of Tm
enhances the production of itself exclusively. Overall, these results demonstrate that the [transꢀ
Tm]/[transꢀTp] ratio is strongly dependent on both the reaction time and the identity of the instructing
template or templates.
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
Examining the role of the [Tp•Tm] heteroduplex
The kinetic analyses revealed that the two replicators examined in this study exhibit markedly different
replication efficiencies as a result of small structural changes engineered in their molecular
frameworks. Although the kinetics of the competition experiments utilized data from 19F NMR
1
spectroscopy, H NMR spectroscopic data was also acquired at the same time during each of the
1
competition experiments. Examination of these H NMR spectra revealed interesting changes in the
chemical shifts of the resonances arising from the trans protons (Figure 5a) on the isoxazolidine rings
present in Tp and Tm. These changes are visible most clearly in the competition kinetic experiment
where the reaction mixture is instructed with 20 mol% of Tp (Figure 5a).
At the beginning of this reaction, replicator Tp, added at a concentration of 1 mM, is the
dominant template in solution and no Tm is present at the start of the reaction. In the presence of
preformed Tp template, Mp, Mm, and N, are transformed quickly into both Tp and Tm. As the
concentration of Tm increases in the reaction mixture over time, a significant change in the chemical
shift of the resonance associated with the trans proton present in Tp is observed. However, the
resonance arising from replicator Tm itself is apparently unperturbed. These observations are
instructive, since they demonstrate strikingly the ability of replicators Tp and Tm to assemble into both
homoduplexes ([Tp•Tp] and [Tm•Tm]) and the corresponding heteroduplex ([Tp•Tm]). The presence of
a single resonance for each template indicates that the exchange between these species is fast on the 1H
14
ACS Paragon Plus Environment

Page 15 of 31
Journal of the American Chemical Society
1
2
3
4
5
NMR chemical shift timescale. For each replicator, the observed chemical shift depends on the mole
fraction of the replicator bound within its homoduplex and the heteroduplex.
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
1
Figure 5.
(a) Arrayed partial 499.9 MHz H NMR spectra showing the changes in the chemical shifts of the
resonances associated with the trans protons on the isoxazolidine ring of Tp (red) and Tm (blue) replicators formed within a
competition experiment instructed with Tp template ([Mp] = [Mm] = [N] = 5 mM, [transꢀTp] = 1 mM, CDCl3, 5 °C) over
time. (b) Overlay of partial 400.1 MHz H NMR spectra (RT) showing the trans resonances belonging to the Tp and Tm
1
cycloadducts. The chemical shift changes for these trans products were examined at five different conditions: starting with
Tp only and progressively increasing the content of Tm to 100% (combined concentration of Tp and Tm = 10 mM). (c)
Simulated changes in the chemical shifts of the isoxazolidine trans proton (see (a)) in transꢀTp (red triangles) and in transꢀ
Tm (blue triangles), arising as a result of the distribution of these products in homo and heteroduplexes. Data are derived
from simulations run using Gepasi 3 (Ref. 22), using the duplex association constants determined through fitting of kinetic
data and estimated/determined chemical shifts for each duplex in the system.
15
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 31
1
2
3
4
5
6
In order to develop a better understanding of these chemical shift changes, we examined the 1H
NMR spectra (Figure 5b) recorded on five solutions of Tp and Tm at a combined concentration of 10
mM in CDCl3, starting with a composition of 100% Tp and increasing the fraction of Tm progressively
7
8
1
from 0% to 25%, 50%, 75%, and, finally, 100%. Each mixture was analyzed by 500.1 MHz H NMR
spectroscopy (Figure 5b, for the corresponding 19F{1H} NMR spectra, see Figure S9). These five
experiments allowed the changes in the chemical shifts arising from the resonances of both replicators
to be examined under conditions where only homoduplex [Tp•Tp] or homoduplex [Tm•Tm] are present,
as well as the three intermediate conditions where both homoꢀ and heteroduplexes are present together.
Analysis of the chemical shift changes observed in the isoxazolidine ring protons colored red and blue
(Figure 5) as a function of the Tp and Tm composition revealed a markedly different behavior for
replicator Tp relative to Tm. While the chemical shift of the resonance arising from Tm varied very little
across the four different compositions examined (25% Tm → 100% Tm), the chemical shift of the
resonance arising from Tp exhibited a dramatic downfield shift (>0.2 ppm, Figure 5b) as the fraction
of Tm in the solution increases.
We modelled the steady state distributions of Tp and Tm within the various product duplexes
present in solution and, consequently, the expected chemical shift for the resonances arising from the
isoxazolidine protons present in Tp and Tm (colored red and blue in Figure 5c) by combining the data
in Figure 5b with the corresponding association constants for the homoꢀ and heteroduplexes
determined earlier through kinetic fitting. Figure 5c shows the calculated chemical shifts for the
appropriate proton resonances in Tp and Tm at nine different concentration ratios. Comparison of these
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
calculated chemical shifts to the experimentally determined H NMR data shows a clear qualitative
agreement. The high stability of [Tm•Tm] is reflected by the fact that the chemical shift of the
isoxazolidine proton associated with this product remains virtually unchanged throughout the
experiments. Since [Tp•Tp] is the least stable duplex, as the concentration of Tm is increased, the
distribution of Tp is shifted towards the more stable [Tp•Tm] duplex, evidenced by the 0.2 ppm
downfield shift in the resonance arising from transꢀTp. The analysis of the behavior of the [Tp•Tm]
heteroduplex demonstrated that the differences in duplex stabilities have a dramatic influence over not
only the autoꢀ and crosscatalytic efficiencies of the two replicators, but also the chemical shifts of the
resonances associated with these templates. In this case, we were able to use the thermodynamic
parameters derived though kinetic fitting to successfully relate the variation in duplex and heteroduplex
stabilities to the variation in the chemical shifts observed for templates Tp and Tm within the highly
interconnected network. Additionally, these differences in the strengths of recognition processes that
underlie the formation of replicators Tp and Tm and the associated complexes/duplexes also help to
16
ACS Paragon Plus Environment

Page 17 of 31
Journal of the American Chemical Society
1
2
3
4
5
determine the outcome of competition in this network of replicators—Tm is not only able to exploit Tp
for its formation, it is also capable of sequestering Tp in a duplex that decreases its availability in
solution and, as a result, the efficiency of the autocatalytic pathway leading to the formation of Tp.
6
7
8
9
What governs the outcome of competition in Tp–Tm network?
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
The comprehensive set of kinetic experiments described here allowed us to examine each catalytic
pathway in isolation and within a number of competition scenarios. The observed results demonstrate
clearly that the absence of one crosscatalytic relationship (Tm is not a catalyst for the formation of Tp)
has a significant influence on the behavior of the replicator network in a situation where Tm and Tp
have to compete for a shared building block. The kinetic fitting demonstrates that there are significant
differences between the two replicators in terms of the duplex association constants and unimolecular
rate constants for the reactions that take place on templates. The origin of the disparity in the
crosscatalytic efficiencies of Tm and Tp, identified by our kinetic simulations and fitting, must lie in
subtle differences in the stabilities and structures of the ternary complexes, [N•Mp•Tm] and
[N•Mm•Tp], and the transition states accessed from these complexes. In order to gain a better
understanding of these processes, we undertook a series of density functional theory calculations that
examined the structures of the ternary complexes [N•Mp•Tm] and [N•Mm•Tp], the heteroduplex
[Tp•Tm], and the transition states (TSs) that connect these complexes. All of these calculations were
performed at the ωB97X/def2ꢀSVP level of theory using a continuum solvation model (PCM) for
chloroform (see Supporting Information for details). In addition, we compared the transition state
structures calculated for the templateꢀdirected reactions with that computed for the prototypical 1,3ꢀ
dipolar cycloaddition of diphenylnitrone to Nꢀphenyl maleimide. Although the relative stereochemistry
of the forming cycloadduct is controlled very well in favor of the trans diastereoisomer, the absolute
stereochemical relationships between the templates are more complex. The templateꢀdirected reactions
can proceed by one of two pathways. The stereochemistry of the newly formed replicator can match
that of the template, for example, RSRꢀTp → RSRꢀTm — we will refer to this pathway as “matched”.
Alternatively, the stereochemistry of the newly formed replicator can be mismatched compared to that
of the template, for example, RSRꢀTp → SRSꢀTm — we will refer to this pathway as “mismatched”.
17
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 31
1
2
3
∆G‡
+95.2
Tm templates Tp
(a)
4
5
6
7
8
9
‡
[N•Mp•Tm]
[N•Mm•Tp]
[N•Mp•Tm]‡
[N•Mm•Tp]‡
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
B
[Tp•Tm]
Relative Energy
+31
‡
∆G‡
+88.9
Mismatched
Matched
Tp templates Tm
∆G‡
+96.5
Tm templates Tp
Relative Energy
0
[Tp•Tm]
‡
C
[N•Mp•Tm]
[N•Mp•Tm]‡
‡
A
[N•Mm•Tp]
∆G‡
[N•Mm•Tp]‡
Tp templates Tm
+98.0
d1
d2
d3
d4
d5
dtet
ɸ
(b)
θ
Comparison
N
Reagents
Parent TS
—
1.264 1.309
—
1.332 0.015
—
—
C
dtet
C
2.086 1.279 1.345 2.140 1.386 0.141 96.95 103.09
d1 d2 d3 d4 d5 dtet
O
ɸ
θ
Matched
d3
m
p ‡
[N•M •T ]
2.078 1.277 1.346 2.149 1.387 0.151 95.25 100.90
2.106 1.277 1.348 2.106 1.387 0.149 95.26 106.18
d2
d1
p
m ‡
d4
[N•M •T
]
ɸ
d1
d2
d3
d4
d5
dtet
ɸ
θ
Mismatched
m
p ‡
d5
[N•M •T ]
m ‡
2.002 1.289 1.341 2.187 1.393 0.166 100.35 102.87
2.096 1.277 1.346 2.128 1.386 0.148 96.95 104.37
θ
p
[N•M •T
]
Figure 6.
(a) Calculations reveal that the ternary complexes [N•Mm•Tp] and [N•Mp•Tm] can react to form the
template duplex [Tp•Tm] through two sets of transition states—in which, the stereochemistries of the forming cycloadducts
can be either matched or mismatched with respect to the replicator template—giving rise to four possible reaction pathways.
Calculations were performed at the ωB97X/def2ꢀSVP level of theory using a continuum solvation model for chloroform.
Energies are in kJ mol–1. Hydrogen bonds are represented by dashed lines. The colored shaded areas indicated by letters are
discussed in the main text. (b) Structural parameters describing the calculated structures (ωB97X/def2ꢀSVP) for the four
transition states accessed from either [N•Mm•Tp] or [N•Mp•Tm]. Data for the corresponding parameters for the transition
state located for the reaction of diphenylnitrone and Nꢀphenyl maleimide are provided for comparison. All distances are in
Å and angles are in degrees.
18
ACS Paragon Plus Environment

Page 19 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The consequence of these relationships is that heteroduplex [Tp•Tm], and the transition states leading to
it, can exist either as a structure in which the stereochemistries of the templates are matched or as a
structure in which the stereochemistries of the templates are mismatched. We computed the four
possible pathways—matched and mismatched for both Tp → Tm and Tm → Tp. The results of these
calculations are summarized in Figure 6.
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
In the case of the template duplexes, the matched form of [Tp•Tm] is predicted to be around
30 kJ mol–1 more stable than the corresponding mismatched duplex. In part, this observation may be
explained by the presence of stabilizing C–H•••O interactions along the spine of the template (marked
A and highlighted in blue in Figure 6a) that are present in the matched duplex, but not in the
mismatched one. Additionally, there are some distortions in the conformations of the fused bicyclic
cycloadducts present in the mismatched [Tp•Tm] duplex that are required in order to maintain the four
hydrogen bonds connecting the two templates.
Comparison of the transition states that were located for the four available pathways reveals that
one structure—mismatched Tp → Tm (Figure 6a)—is significantly lower in energy than the other
three. This observation is consistent with the experimental results indicating that Tp can act as a
catalyst for the formation of Tm, but Tm is not a particularly efficient catalyst for the formation of Tp.
Comparison of the transition state structures with that of the parent cycloaddition between
diphenylnitrone and Nꢀphenyl maleimide (Figure 6b) reveals that the mismatched transition state
[N•Mm•Tp]‡ has a geometry that differs considerably from the other three and from the parent. In this
transition state, C–O bond formation is more advanced (C•••O = 2.002 Å vs. 2.086 Å in parent TS) and
C–C bond formation is much less advanced (C•••C = 2.187 Å vs. 2.146 Å in parent TS) than in any of
the other three transition states accessible from the ternary complexes. This distortion suggests a
transition state that may be slightly more polar23 that the other three and may be stabilized by a series
of C–H•••O interactions present along the spine of the template (marked B and highlighted in green in
Figure 6a).
Whilst the structures of the transition states and the duplexes have limited conformational
freedom as they adopt coꢀconformations that maintain four hydrogen bonds, the ternary complexes are
much less restricted in a conformational sense. Bruice and Lightstone have shown24 that the rates of
intramolecular reactions can be related to the fraction of low energy conformations that are present as
nearꢀattack conformations (NACs). Previously, we have applied25 this type of analysis successfully to a
recognitionꢀmediated pseudointramolecular cycloaddition reaction. Therefore, when considering these
structures, we restricted our examination to structures that could be categorized loosely as NACs, i.e.,
we examined the structures in which the termini of the 4π and 2π components of the cycloaddition were
19
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 20 of 31
1
2
3
4
5
6
7
8
9
in closest proximity. Interestingly, in the lowest energy structure located for each ternary complex,
there is a striking difference between the distances that separate the reactive termini of the 4π and 2π
components (C•••O and C•••C) depending on the identity of the replicator that is acting as the template.
In the ternary complex [N•Mp•Tm], where Tm is templating the formation of Tp, all of the
C•••O and C•••C distances are above 4 Å and, in the case of matched [N•Mp•Tm], over 5 Å (marked C
and highlighted in red in Figure 6a). These large distances suggest that the low energy coꢀ
conformations of [N•Mp•Tm] are unlikely to contain a high fraction of NACs. This observation is
consistent with lower efficiency of this ternary complex determined experimentally. By contrast, in
[N•Mm•Tp], where Tp is templating the formation of Tm, all of the C•••O and C•••C distances are
below 4 Å and in the case of the mismatched ternary complex, these distances are both below 3.40 Å.
These very short distances suggest that the low energy coꢀconformations of mismatched [N•Mp•Tm]
are likely to contain a high fraction of NACs and this observation is consistent with this ternary
complex accessing the lowest energy transition state.
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
These calculations also offer a window into a more complex world that is currently beyond the
reach of our experiments, in which the stereochemical relationships between replicators can play an
important role in determining the outcome of competition8b,8e,26 between the replicators. Going
forward, these results demand that, in order to maximize the efficiency of a templateꢀdirected process,
it will be necessary to consider the detailed relationship between individual molecular structures and
their coꢀconformations within complexes.
Conclusions
Small networks of minimal replicators may serve27 as simple models for primitive metabolic pathways
in that the population of an individual replicator within these networks can be manipulated by
exploiting the autocatalytic properties of the replicating template in order to specifically upregulate the
rate of formation of that template. In this work, we describe the experimental implementation of such a
network constructed from three simple building blocks—two maleimides (Mm and Mp) and a single
nitrone (N)—that react to create two synthetic replicators (Tm and Tp) that can template their own
formation. Within this network, the two replicating templates compete for a common reagent, namely
nitrone N. In order to understand the systemꢀlevel behavior of this network, it is important to
characterize each replicator in isolation. In this case, our kinetic studies on Tm and Tp in isolation
reveal that Tp is the more efficient replicator of the pair. In isolation, Tp benefits from a significantly
higher effective molarity for the key cycloaddition step than Tm. Additionally, the [Tp•Tp] duplex is
somewhat less stable than the [Tm•Tm] duplex, thereby ensuring that the concentration of catalytically
20
ACS Paragon Plus Environment

Page 21 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
active, monomeric Tp is higher than in the corresponding reaction involving Tm. The abilities of both
Tp and Tm to replicate are confirmed by their capacity to template their own formation as evidenced by
experiments in which preꢀformed template is added to a mixture of the nitrone and the appropriate
maleimide at the start of the reaction. The structural similarity between Tp and Tm opens the possibility
that a mutual crosscatalytic relationship may exist between these two templates. By exploring the
response of reactions that form either Tp or Tm to the addition of the other template as an instruction,
we established that, although Tp can template the formation of Tm, Tm does not catalyze the formation
of Tp effectively. The absence of reciprocity in the crosscatalytic relationship between Tp and Tm is
critical to understanding the outcome of experiments in which Tp and Tm are placed in a competitive
situation. Analysis of experiments in which Tp and Tm replicate in the same flask reveal the apparently
anomalous observation that the replicator that is less efficient in isolation (Tm) becomes the dominant
species in the complete network. This observation is a direct result of the presence of the nonꢀ
reciprocal crosscatalytic interaction between Tp and Tm, which allows Tm to exploit Tp for its
formation but not vice versa. The systemꢀlevel outcome is also driven by the subtle interplay of the
recognition and reaction processes. Replicator Tm forms more stable complexes and duplexes, which
allows it to reduce the replication efficiency of Tp by sequestering it within the [Tp•Tm] heteroduplex
that is more stable than the [Tp•Tp] homoduplex. However, our attempts to direct the competition
between the replicators using preformed templates under the wellꢀstirred batch reactor conditions
employed here always result in the erosion of the imbalance between the two replicators generated at
early time points. This effect is a direct result of the exhaustion of building blocks that limits the
efficiency of replication processes, thereby preventing the initially dominant replicator from enhancing
its advantage. This observation illustrates clearly that the ability of one replicator to dominate within
this network depends not only on its catalytic relationship with the other replicator in the network but
also their reaction environment. The current work represents a firm foundation for the expansion of the
study of replicator networks under conditions where dynamic processes or diffusion contribute to the
outcome of replicator competition—i.e., reaction conditions where replication processes can operate
with increased efficiency. These studies are currently underway in our laboratory.
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
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 22 of 31
1
2
3
Acknowledgements
The financial support for this work was provided by the University of St Andrews and the Engineering
and Physical Sciences Research Council (Grant EP/K503162/1).
4
5
6
7
8
9
Supporting information:
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
Supporting information is available free of charge at:
General experimental procedures; synthetic procedures and compound characterization; details of
kinetic analyses, fitting (including fitted kinetic profiles), and simulations, and details of DFT
calculations.
Notes and References
(1)
(a) Newmann, M. E. J. SIAM Rev. 2003, 45, 167. (b) Barabási, A.ꢀL. Rev. Mod. Phys. 2002, 74,
1. (c) Strogatz, S. H. Nature 2001, 410, 268.
(2)
(a) Rieckmann, J. C.; Geiger, R.; Hornurg, D.; Wolf, T.; Kveler, K.; Jarossay, D.; Sallusto, F.;
ShenꢀOrr, S. S.; Lanzavecchia, A.; Mann, M.; Meissner, F. Nat. Immunol. 2017, 18, 583. (b)
Boccaletti, S.; Latora, V.; Moreno, Y.; Chavez, M.; Hwang, D. U. Phys. Rep. 2006, 424 (4–5),
175. (c) Barabási, A.ꢀL.; Oltvai, Z. N. Nat. Rev. Genet. 2004, 5, 101. (d) Jeong, H.; Tombor, B.;
Albert, R.; Oltvai, Z. N.; Barabási, A.ꢀL. Nature 2000, 407, 651.
(3)
(a) Mitchell, M. Complexity: A Guided Tour; Oxford University Press: New York, 2009. (b)
Watts, D. J. Small Worlds: The Dynamics of Networks between Order and Randomness;
Princeton University Press: New Jersey, 2003. (c) Barabási, A.ꢀL. Linked: The New Science of
Networks; Perseus Book Group: New York, 2002.
(4)
(5)
Kauffman, S. A. At Home in the Universe: The Search for the Laws of Self-Organization and
Complexity; Oxford University Press Inc.: New York, 1995.
(a) Ashkenasy, G.; Hermans, T. M.; Otto, S.; Taylor, A. F. Chem. Soc. Rev. 2017, 46, 2543. (b)
De la Escosura, A.; Briones, C.; RuizꢀMirazo, K. J. Theor. Biol. 2015, 381, 11. (c) Mattia, E.;
Otto, S. Nat. Nanotechnol. 2015, 10, 111. (d) RuizꢀMirazo, K.; Briones, C.; de la Escosura, A.
Chem. Rev. 2014, 114, 285. (e) von Kiedrowski, G.; Otto, S.; Herdewijn, P. J. Syst. Chem. 2010,
1, 1. (f) Peyralans, J. J.; Otto, S. Curr. Opin. Chem. Biol. 2009, 13, 705. (g) Stankiewicz, J.;
Eckardt, L. H. Angew. Chem. Int. Ed. 2006, 45, 342.
(6)
(7)
(a) Plöger, T. A.; von Kiedrowski, G. Org. Biomol. Chem. 2014, 12, 6908. (b) Paul, N.; Joyce,
G. F. Proc. Natl. Acad. Sci. USA 2002, 99, 12733. (c) Luther, A.; Brandsch, R.; von Kiedrowski,
G. Nature 1998, 396, 245. (d) Sievers, D.; von Kiedrowski, G. Nature 1994, 369, 221. (e) von
Kiedrowski, G.; Wlotzka, B.; Helbing, J.; Matzen, M.; Jordan, S. Angew. Chem. Int. Ed. Engl.
1991, 30, 423. (f) von Kiedrowski, G. Angew. Chem. Int. Ed. Engl. 1986, 25, 932.
(a) ColombꢀDelsuc, M.; Mattia, E.; Sadownik, J. W.; Otto, S. Nat. Commun. 2015, 6, 7427. (b)
Carnall, J. M. A.; Waudby, C. A.; Belenguer, A. M.; Stuart, M. C. A.; Peyralans, J. J.ꢀP.; Otto,
S. Science 2010, 327, 1502. (c) Li, X.; Chmielewski, J. J. Am. Chem. Soc. 2003, 125, 11820. (d)
Issac, R.; Chmielewski, J. J. Am. Chem. Soc. 2002, 124, 6808. (e) Yao, S.; Ghosh, I.; Zutshi, R.;
Chmielewski, J. Angew. Chem. Int. Ed. Engl. 1998, 37, 478. (f) Severin, K.; Lee, D. H.; Kennan,
A. J.; Ghadiri, M. R. Nature 1997, 389, 706. (g) Yao, S.; Ghosh, I.; Zutshi, R.; Chmielewski, J.
J. Am. Chem. Soc. 1997, 119, 10559. (h) Lee, D. H.; Granja, J. R.; Martinez, J. A.; Severin, K.;
Ghadri, M. R. Nature 1996, 382, 525.
22
ACS Paragon Plus Environment

Page 23 of 31
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
(8)
(a) Bottero, I.; Huck, J.; Kosikova, T.; Philp, D. J. Am. Chem. Soc. 2016, 138, 6723. (b)
Dieckmann, A.; Beniken, S.; Lorenz, C. D.; Doltsinis, N. D.; von Kiedrowski, G. Chem. Eur. J.
2011, 17, 468. (c) Kassianidis, E.; Philp, D. Angew. Chem. Int. Ed. 2006, 45, 6344. (d) Pearson,
R. J.; Kassianidis, E.; Slawin, A. M. Z.; Philp, D. Chem. Eur. J. 2006, 12, 6829. (e) Kindermann,
M.; Stahl, I.; Reimold, M.; Pankau, W. M.; von Kiedrowski, G. Angew. Chem. Int. Ed. 2005, 44,
6750. (f) Pearson, R. J.; Kassianidis, E.; Slawin, A. M. Z.; Philp, D. Org. Biomol. Chem. 2004,
2, 3434. (g) Quayle, J. M.; Slawin, A. M. .; Philp, D. Tetrahedron Lett. 2002, 43, 7229. (h)
Wang, B.; Sutherland, I. O. Chem. Commun. 1997, 1495. (i) Rotello, V.; Hong, J.ꢀI.; Rebek, Jr.,
J. J. Am. Chem. Soc. 1991, 113, 9422. (j) Tjivikua, T.; Ballester, P.; Rebek, Jr., J. J. Am. Chem.
Soc. 1990, 112, 1249.
(9)
Severin, K.; Lee, D. H.; Martinez, J. A.; Vieth, M.; Ghadiri, M. R. Angew. Chem. Int. Ed. Engl.
1998, 37, 126.
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
(10) Saghatelian, A.; Yokobayashi, Y.; Soltani, K.; Ghadiri, M. R. Nature 2001, 409, 797.
(11) (a) Ashkenasy, G.; Dadon, Z.; Alesebi, S.; Wagner, N.; Ashkenasy, N. Isr. J. Chem. 2011, 51,
106. (b) Allen, V. C.; Robertson, C. C.; Turega, S. M.; Philp, D. Org. Lett. 2010, 12, 1920. (c)
Ashkenasy, G.; Ghadiri, M. R. J. Am. Chem. Soc. 2004, 126, 11140.
(12) (a) Sadownik, J. W.; Mattia, E.; Nowak, P.; Otto, S. Nat. Chem. 2016, 8, 264. (b) Dadon, Z.;
Wagner, N.; Ashkenasy, G. Angew. Chem. Int. Ed. 2008, 47, 6128. (c) Ashkenasy, G.; Jagasia,
R.; Yadav, M.; Ghadiri, M. R. Proc. Natl. Acad. Sci. USA 2004, 101, 10872. (d) Yao, S.; Ghosh,
I.; Zutshi, R.; Chmielewski, J. Nature 1998, 396, 447. (e) Lee, D. H.; Severin, K.; Yokobayashi,
Y.; Ghadiri, M. R. Nature 1997, 390, 591.
(13) (a) Kim, D.ꢀE.; Joyce, G. F. Chem. Biol. 2004, 11, 1505. (b) Sievers, D.; Von Kiedrowski, G.
Chem. Eur. J. 1998, 4, 629.
(14) (a) Kosikova, T.; Hassan, N. I.; Cordes, D. B.; Slawin, A. M. Z.; Philp, D. J. Am. Chem. Soc.
2015, 137, 16074. (b) Kassianidis, E.; Pearson, R. J.; Wood, E. A.; Philp, D. Faraday Discuss.
2010, 145, 235. (c) Sadownik, J. W.; Philp, D. Angew. Chem. Int. Ed. 2008, 47, 9965. (d) Xu, S.;
Giuseppone, N. J. Am. Chem. Soc. 2008, 130, 1826.
(15) (a) Bissette, A. J.; Fletcher, S. P. Angew. Chem. Int. Ed. 2013, 52, 12800. (b) Philp, D.; Huck, J.
Supramolecular Chemistry: From Molecules to Nanomaterials; John Wiley & Sons, Ltd.: New
York, 2012; Vol. 4, pp 1415. (c) Vidonne, A.; Philp, D. Eur. J. Org. Chem 2009, 2009, 593. (d)
Patzke, V.; von Kiedrowski, G. ARKIVOC 2007, 46, 293. (e) von Kiedrowski, G. Bioorganic
Chem. Front. 1993, 3, 113.
(16) The 1,3ꢀdipolar cycloaddition reaction between a nitrone and a maleimide can give rise to two
diastereoisomers, trans and cis. Only the trans diastereoisomers are capable of taking part in
templateꢀdirected replication processes in the network of Tp and Tm replicators. For this reason,
the notation for the trans cycloadducts capable of replication will generally omit the trans
throughout this work (for example, transꢀTp will be generally referred to as Tp, unless emphasis
on the identity of the diastereoisomer is necessary). All preformed templates used in instructed
experiments consisted exclusively of the trans diastereoisomer of the relevant replicator.
(17) This relationship presents a slight generalization, as whilst generally true, the increase in the
strength of the recognition between two network components might also result in the formation
of [M•N] type of complexes, which can reduce the efficiency of the autocatalytic pathway.
(18) The error values for the calculated %EFs were determined using standard methods for
calculation of normally distributed errors. These calculations employed the estimated value of
concentration error of ±0.02 mM determined for all of the kinetic experiments described in this
work (for details, see Supporting Information).
(19) Kinetic effective molarity provides a measure of the enhancement in the templateꢀmediated
pathway relative to the bimolecular pathway. EMkinetic also tells us about the concentration at
which the reaction would have to be performed in order for the bimolecular pathway to perform
23
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 24 of 31
1
2
3
4
5
6
7
at the efficiency of the recognitionꢀmediated pathway. For this reason, rate acceleration takes
place even if EMkinetic < 1, as long as the concentration at which the reaction is conducted is
below the value of EMkinetic
.
(20) Thermodynamic effective molarity provides information about the stability of the template
duplex relative to the strength of the ternary complex.
8
9
(21) The values of template duplex association constants determined through kinetic fitting in this
work were in the range of ~ 17.9 to 68.6 × 106 M–1, i.e., a situation where the concentration of
the catalytically active ternary complex will be in the ꢁM range (see Ref. 8c for detailed
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
analysis) under the conditions employed experimentally. Ideally, selfꢀreplicating systems would
be examined at a concentration (C) that promotes the dissociation of the template duplex
([C]initial = 1/KaDuplex). At such a concentration, however, the formation of the catalytically active
ternary complex, which is driven by the considerably smaller KaInd would be strongly
disfavoured. The optimum conditions, therefore, need to balance these two competing
influences.
(22) (a) Mendes, P. Comput. Applic. Biosci. 1993, 9, 563. (b) Mendes, P. Tends Biochem. Sci. 1997,
22, 361.
(23) Cowie, R. M.; Turega, S. M.; Philp, D. Org. Lett. 2006, 8, 5179.
(24) (a) Lightstone, F. C.; Bruice, T. C. J. Am. Chem. Soc. 1996, 118, 2595. (b) Bruice, T. C.;
Lightstone, F. C. Acc. Chem. Res. 1999, 32, 127.
(25) Sinclair, A. J.; del Amo, V.; Philp, D. Org. Biomol. Chem. 2009, 7, 3308.
(26) (a) Ribó, J. M.; Crusats, J.; ElꢀHachemi, Z.; Moyano, A.; Hochberg, D. Chem. Sci. 2017, 8, 763.
(b) Ribóꢂ, J. M.; Blancoꢂ, C.; Crusatsꢂ, J.; ElꢀHachemiꢂ, Z.; Hochberg, D.; Moyano, A. Chem.
Eur. J. 2014, 20, 17250. (c) Dieckmann, A.; Houk, K. N. Chem. Sci. 2013, 4, 3591.
(27) Shapiro, R. Q. Rev. Biol. 2006, 81, 105.
24
ACS Paragon Plus Environment

Page 25 of 31
Journal of the American Chemical Society
1
2
3
For Table of Contents Only:
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
25
ACS Paragon Plus Environment

(a)
O
F
F
Journal of the American Chemical Society Page 26 of 31
N
N
O–
N+
O
O
H
N
N
N
N
H
H
H
H
N
N
F
O
O
H
H
1
2
3
4
N
H
O
H
O
O
N
N
O
C8H17
CO2H
O
O
Mp
N
Mm
N
HO2C
CO2H
C8H17
O
O
HO2C
5
6
Replicator Tp
Replicator Tm
Building blocks
7
(b) Two-replicator interconnected network
[Tp•Tm]
8
9
kcross (Tp → Tm)
kcross (Tm → Tp)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Mp + N with Tm
Crosscatalytic
cycle
Mm + N with Tp
Crosscatalytic
cycle
Ka [Tp•Tm]
Ind
Ind
Ka
Ka
Tp
[Mm•N•Tp]
[Mp•N•Tm]
Mp
Mm
N
Ind
Ind
Ka
Ka
Tm
Tp Autocatalytic
Tm Autocatalytic
[Mm•N•Tm]
[Mp•N•Tp]
cycle
cycle
kauto (Tm → Tm)
p
kauto (Tp → T )
Ka [Tp•Tp]
Ka [Tm•Tm]
[Tp•Tp]
[Tm•Tm]
(d) K values
(c) Bimolecular pathways
a
Mp
R1
R1
N
KaInd / M–1
R2
kbi
kbi
R2
N
N
H
H
O
O
cis
trans
830
H
H
H
H
O
O
ACS Paragon Plus EnvOironmeOnt
Mm
N
N
R3
N
3320
kbi
kbi
R3
cis
trans
32
cis-Tp or cis-Tm
trans-Tp or trans-Tm
33

p
(a)
(c)
(e)
Formation of T
5
+Tp
5
200
150
100
50
p
Page 27 of O31
Journal of the American Chemic+Tal Society
+Tm
4
pC
4
N
M
No T
CO2Me
O
3
2
13
22
3
0
m
+T
1
0
-50
41
50
6
4 h
16 h
0
4
8
12
16
0
4
8
12
16
Time / h
Time / h
]
]
u
[T − [T
i
7
8
%EF = 100 ⋅
[T]u
m
(b)
(d)
Formation of T
9
5
5
4
3
2
1
O
N
O
10
(f)
mC
+Tp
+Tm
4
200
150
100
50
M
11
m
+T
CO2Me
12
3
13
2
14
15
1
0
ACS Paragon Plus Enpvironment
No T
+T
16
-50
0
0
17
0
4
8
12
16
0
4
8
12
16
4 h
16 h
18
Time / h
Time / h

(a)
Formation of Tp in
Journal of the American ChemPicaagleSo28cioetfy31
the presence of Tm
100
Tp in [Tp•Tp]
80
1
2
3
4
5
6
7
8
9
60
Tm in [Tp•Tm]
Tm in [Tm•Tm]
40
20
0
Tp in [Tp•Tm]
0
4
8
12
16
Time / h
10
(b)
Formation of Tm in
the presence of Tp
11
12
100
13
Tm in [Tm•Tm]
80
60
40
20
14
15
16
17
18
19
20
Tp in [Tp•Tm]
Tp in [Tp•Tp]
Tm in [Tp•Tm]
21 0
22
23
0
4
8
12
16
Time / h
24
(c)
25
26
27
28
29
ACS Paragon Plus Environment
Tp
Tm

(a)
(b)
2.5
2.5
Page 29 of 31 Journal of the American Chemical Society
4 h
16 h
2.0
1.5
1.0
0.5
2.0
1.5
1.0
0.5
1
2
3
4
5
6
7
m
m
T
T
dominates
dominates
p
p
T
T
dominates
dominates
8
9
10
11
Increase in time
ACS Paragon Plus Environment
m
Experiments with no T and T : selectivity
p
m
T
T
p
p
m
12
Experiments with no T and both T and T : selectivity unchanged
13