A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iterative Carbon-Carbon Bond Formation

Article abstract of DOI:10.1016/j.chempr.2020.09.021

We report an artificial molecular machine that moves along a track, iteratively joining building blocks to form an oligomer of single sequence with a continuous backbone of carbon-carbon bonds. The rotaxane features a macrocycle bearing an aldehyde-terminated chain and an axle containing different phosphonium ylides separated by rigid spacers. Each ylide is large enough to block the passage of the macrocycle, trapping the ring between the stopper at the terminus of original threading and the next ylide along the track. Once a building block is reachable, it is removed from the track through a Wittig reaction that adds it to the terminus of the growing chain. Operation on a four-barrier tetra(phosphonium salt) track produces a tetra(diphenylpropane) of single sequence linked through alkene bonds. The prototype extends the principle for molecular machines that build polymers by moving along tracks to the synthesis of sequence-encoded chains with continuous carbon backbones. Sequence is crucial in the molecular world. Proteins are built from a common set of 20 amino acids, but different sequences afford materials as diverse as snake venom, muscle, and spider silk. However, the synthesis of artificial sequence polymers remains challenging. Biology uses molecular machines (e.g., ribosomes) for such tasks, inspiring the invention of artificial systems that move along tracks, picking off and joining building blocks in sequence. To date, such small-molecule machines have used amide formation to join building blocks, the same bonds the ribosome uses to make peptides. Here, we report on the design, synthesis, and operation of a track-based molecular machine that assembles a single-sequence oligomer with a continuous backbone of carbon-carbon bonds. This new class of de novo molecular synthesizer utilizes chemistry and reactivity patterns unavailable to biological machines. The long-term goal is for such molecular assemblers to ultimately be able to play significant roles in molecular construction. Molecular machines, such as ribosomes, are ubiquitous in biology. These natural systems are inspiring artificial systems that move along tracks, picking off and joining building blocks in sequence. To date, such small-molecule machines have used amide formation to connect building blocks, much like the ribosome. Here, the design, synthesis, and operation of a track-based molecular machine that iteratively forms a continuous backbone of carbon-carbon bonds is described. This new class of de novo molecular synthesizer utilizes chemistry and reactivity patterns unavailable to biological machines.

Full text of DOI:10.1016/j.chempr.2020.09.021

ll
OPEN ACCESS
Article
A Track-Based Molecular Synthesizer that
Builds a Single-Sequence Oligomer through
Iterative Carbon-Carbon Bond Formation
Charlie T. McTernan, Guillaume
De Bo, David A. Leigh
david.leigh@manchester.ac.uk
HIGHLIGHTS
A rotaxane ring on a
phosphonium ylide track joins
monomers through Wittig
reactions
Nanomachine makes single-
sequence oligomer with
continuous backbone of carbon
bonds
Utilizes chemistry and reactivity
patterns unavailable to
biomolecular machines
Future assemblers may provide
solutions to challenging problems
in synthesis
Molecular machines, such as ribosomes, are ubiquitous in biology. These natural
systems are inspiring artificial systems that move along tracks, picking off and
joining building blocks in sequence. To date, such small-molecule machines have
used amide formation to connect building blocks, much like the ribosome. Here,
the design, synthesis, and operation of a track-based molecular machine that
iteratively forms a continuous backbone of carbon-carbon bonds is described. This
new class of de novo molecular synthesizer utilizes chemistry and reactivity
patterns unavailable to biological machines.
McTernan et al., Chem 6, 1–10
November 5, 2020 ª 2020 The Author(s).
Published by Elsevier Inc.
https://doi.org/10.1016/j.chempr.2020.09.021

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
A Track-Based Molecular Synthesizer that Builds
a Single-Sequence Oligomer through Iterative
Carbon-Carbon Bond Formation
Charlie T. McTernan,1 Guillaume De Bo,1 and David A. Leigh1,2,3,
*
SUMMARY
The Bigger Picture
Sequence is crucial in the
We report an artificial molecular machine that moves along a track,
iteratively joining building blocks to form an oligomer of single
sequence with a continuous backbone of carbon-carbon bonds.
The rotaxane features a macrocycle bearing an aldehyde-termi-
nated chain and an axle containing different phosphonium ylides
separated by rigid spacers. Each ylide is large enough to block the
passage of the macrocycle, trapping the ring between the stopper
at the terminus of original threading and the next ylide along the
track. Once a building block is reachable, it is removed from
the track through a Wittig reaction that adds it to the terminus of
the growing chain. Operation on a four-barrier tetra(phosphonium
salt) track produces a tetra(diphenylpropane) of single sequence
linked through alkene bonds. The prototype extends the principle
for molecular machines that build polymers by moving along tracks
to the synthesis of sequence-encoded chains with continuous car-
bon backbones.
molecular world. Proteins are built
from a common set of 20 amino
acids, but different sequences
afford materials as diverse as
snake venom, muscle, and spider
silk. However, the synthesis of
artificial sequence polymers
remains challenging. Biology uses
molecular machines (e.g.,
ribosomes) for such tasks,
inspiring the invention of artificial
systems that move along tracks,
picking off and joining building
blocks in sequence. To date, such
small-molecule machines have
used amide formation to join
building blocks, the same bonds
the ribosome uses to make
peptides. Here, we report on the
design, synthesis, and operation
of a track-based molecular
INTRODUCTION
Polymers with backbones of continuous carbon-carbon bonds (e.g., polystyrene,
polyethylene, polymethyl methacrylate, etc.) are among the most widely used man-
made materials.1–3 However, the synthesis of monodisperse versions of these mac-
romolecules remains challenging,4 and analogs with encoded sequence information
cannot be prepared using current synthetic methods.5,6 Biology uses molecular ma-
chines to build sequence polymers but is limited in constitutional diversity to pro-
teins, DNA, RNA, and carbohydrates.7 Polyketide and fatty acid biosynthesis (bio-
molecules with a continuous chain of carbon atoms built from iterative Claisen
condensations of malonyl-CoA or related building blocks) involves post-synthetic
modification of specific-length oligomers,7 but nature has not developed molecular
machines that build long-chain, sequence-specific polymers with an uninterrupted
carbon backbone. Artificial DNA8–10 and small-molecule machines11 have been de-
signed that move along tracks, iteratively detaching and connecting building blocks
to form monodisperse oligomers of defined sequence12–14 and polymers of tran-
scribed polydispersity.15 Systems able to modify polymeric chains16 or to thread
precise numbers of macrocycles onto polymeric threads,17–20 including controlling
their sequence,18 have been demonstrated. To date, small-molecule track-based
synthesizers have all connected building blocks through amide bond formation, a
feature reminiscent of the ribosome.21 Here, we explore extending this concept
beyond peptide synthesis in the form of an artificial molecular assembler22,23 that
moves along a track iteratively joining together building blocks to form an oligomer
of single sequence with a continuous backbone of carbon-carbon bonds.
machine that assembles a single-
sequence oligomer with a
continuous backbone of carbon-
carbon bonds. This new class of
de novo molecular synthesizer
utilizes chemistry and reactivity
patterns unavailable to biological
machines. The long-term goal is
for such molecular assemblers to
ultimately be able to play
significant roles in molecular
construction.
Chem 6, 1–10, November 5, 2020 ª 2020 The Author(s). Published by Elsevier Inc.
1
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
RESULTS
Molecular machine-track conjugate 1 (Figure 1) was designed to use iterative
Wittig reactions to form carbon-carbon double bonds between a macrocycle and
building blocks abstracted one at a time and in sequence from a track. The Wittig
reaction24–26 was chosen as it is robust and structurally tolerant, lending itself to
exploitation in a range of contexts, including dynamic DNA-template synthesis.9
Our machine is based on a rotaxane architecture, in which the macrocycle has a reac-
tive aldehyde attachment and the axle has the building-block sequence encoded as
phosphonium salts during its synthesis. The 2,2-diphenylpropane phosphonium
units act both to restrict the position of the ring on the track and, upon deprotona-
tion, as reactive ylide functionalities. Each ylide is large enough to block the passage
of the macrocycle, trapping the ring within a compartment defined by the bulky
stopper at the terminus of original threading and the next ylide along the track.
Once a reactive building block can be reached by the macrocycle-appended alde-
hyde, it can be removed from the track through a Wittig reaction that adds it to
the terminus of the growing chain. Each barrier also contains an aldehyde unit, so
that once the building block is added to the end of the chain, it is able to react
with the next barrier on the track that the macrocycle can access, enabling the
alkene-connected oligomer to grow through successive Wittig reactions.
The specific size and constitution of the 2,2-diphenylpropane motif of the building
blocks proved important for successful machine operation. Early track designs in
which the ylide and aldehyde were attached to the same aromatic ring or extended
conjugated system proved insufficiently reactive (see Section S7 for a brief discus-
sion of initial designs). Embedding the phosphorus atoms within the vector of the
track allowed synthetically accessible triaryl phosphines to be the basis of the track
design, expediting the synthesis (see Sections S2 and S3). The phenyl substituent at
each phosphorus center (e.g., 4a–4d) also proved important: when a tolyl (4-meth-
ylphenyl) linking group was investigated, it proved difficult to develop macrocycles
that could both thread during the rotaxane-forming reaction and, subsequently,
pass over the phosphine oxide in the track formed from the Wittig reaction.
Each phosphorus center is attached to a methylene group bearing a diarylpropane
building block derivatized with
a different pair of substituents (H, Ph,
C6H4CH2CHMe2, or C6H4OMe). These provide different sidechains in the machine
product, the same role that different amino acids play in proteins. However, two
(identical) sidechains are present per monomer using this artificial molecular ma-
chine design compared with one sidechain per amino acid in proteins. This was cho-
sen partly to illustrate how artificial machines and their products are not subject to
the same constraints as biomolecular synthesizers but, conveniently, the symmetry
of the building blocks also makes their synthesis more straightforward. Each phos-
phonium moiety is separated from the next by rigid spacers that prevent folding
of the track and so ensure that the phosphonium salts can only react with the alde-
hyde group at the end of the chain attached to the macrocycle rather than others on
the track. Rotaxane systems with a sufficiently rigid track such that the axle cannot
fold are rare and making such a long rigid track presents particular challenges.27,28
The inter-residue spacers in the track of 1 contain methyl groups adjacent to the
biphenyl connections to break the planarity and conjugation of the track and
improve the solubility of the machine-track conjugate.
1Department of Chemistry, University of
Manchester, Oxford Road, Manchester M13 9PL,
UK
2School of Chemistry and Molecular Engineering,
East China Normal University, Shanghai 200062,
China
3Lead Contact
Macrocycle 2 features an endotopic pyridine group in order to direct an active
metal template29–31 copper-catalyzed azide-alkyne cycloaddition (CuAAC)32 of
*Correspondence:
david.leigh@manchester.ac.uk
https://doi.org/10.1016/j.chempr.2020.09.021
2
Chem 6, 1–10, November 5, 2020

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
Figure 1. Operation of Machine-Track Conjugate 1, Assembled by Iterative CuAAC Reactions, Generates a Tetrastilbene Oligomer of Single
Sequence, 7
4a R=H, 4b R=Ph, 4c R=C6H4CH2CHMe2, 4d R=C6H4OMe. Conditions: (i) CuPF6(CH3CN)4 (1 equiv), 3 (1 equiv), 4 (3 equiv), 2 (5 equiv), CH2Cl2, RT, 16 h,
23%. (ii–vii) 4,40’-Diazido-2,20’-dimethyl-1,10:40,10’-terphenyl (S9) (for ii), 4b (for iii), S9 (for iv), 4c (for v), S9 (for vi), or 4d then phenyl azide (for vii) (10 equiv;
see Section S3 for further details), CuPF6(CH3CN)4 (1 equiv), Tentagelª TBTA (Tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine, resin-bound, 1.5 equiv),
CH2Cl2, d6-DMSO, RT, 2 h, 35% overall yield in 6 steps. (viii) Resin-bound 2-tert-butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-
diazaphosphorine (44 equiv), CH2Cl2, ꢀ78ꢁC (15 min) to RT, 5 days. Isolated yield of 8, 69%; isolated yield of 7, 21%. TBDPS, tButyldiphenylsilyl. For
synthetic procedures and characterization data, see Section S3.
Chem 6, 1–10, November 5, 2020
3

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
Figure 2. 1H NMR Stackplot Showing Characteristic Proton Shifts upon Rotaxane Formation in 1 and 5
(A–C) Partial 1H NMR (600 MHz, 298 K, d6-DMSO) of (A) 6, (B) 5, and (C) 1. See Section S3 for full assignment.
azide stopper 3 and dialkyne building block 4 through the macrocycle cavity. This
was used to assemble a one-barrier rotaxane building block, 5, which was isolated
in modest yield (23%), in part because the rotaxane-forming step also involves de-
symmetrization of the phosphonium bis-alkyne 4 (Figure 1, step i) and also
because 5 was somewhat unstable with respect to purification by chromatography
on silica gel, as it is likely incompatible with silica OH groups. However, the short
route to 5 proved more effective than desymmetrization at later stages of the syn-
thesis, and 5 proved a convenient intermediate for constructing the rest of the
track. The active metal template strategy avoids the requirement for persistent
attractive interactions between the macrocycle and thread that could slow compo-
nent dynamics in the fully assembled machine. In this way, the macrocycle is
restricted in its movements along the track only by the bulky building blocks. Ro-
taxane 5 was elongated through a series of further CuAAC reactions (Figure 1,
steps ii–vii) to assemble the four-building-block molecular machine 1 in 35% over-
all yield from 5 (Figure 1; Section S3). Two shorter two- and three-barrier machines,
12 and 14, were elaborated from 5 in similar fashion in 54% and 39% overall yields,
respectively (Figure 3; Section S5). Distinctive upfield shifts of triazole proton t1
(from 8.90 ppm in 6 to 8.37 ppm in 5 and 1, see Section S3 for full assignment)
and A2 in the proximal alkyl chain (from 2.34 ppm in 6 to 1.93 ppm in 5 and 1)
1
in the H NMR spectrum of rotaxane 5, compared with thread 6, confirm the local-
ization of the macrocycle in the left-hand compartment of rotaxanes 1 (Figure 2),
12, and 14 (Section S5).
4
Chem 6, 1–10, November 5, 2020

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
Figure 3. Operation of One-, Two-, and Three-Building-Block Machines 11, 12, and 14 to Form Mono-, Di-, and Tri-Stilbenes 10, 13, and 15 and
Corresponding Non-interlocked Systems
(A) (i) Lithium diisopropylamine (8 equiv), THF, ꢀ78ꢁC (15 min) to 60ꢁC, 1 day.
(B) (ii) Resin-bound 2-tert-butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorine (72 equiv), CH2Cl2, ꢀ78ꢁC (15 min) to RT, 1 day.
(C) (iii) Resin-bound 2-tert-butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorine (44 equiv), CH2Cl2, ꢀ78ꢁC (15 min) to RT, 3 days.
TBDPS, tButyldiphenylsilyl. For operation conditions, synthetic procedures, and characterization data, see Sections S5 and S6.
Effective procedures for machine operation were first established for the two-build-
ing-block rotaxane, 12 (Figure 3; Section S6). Under optimized conditions, machine
12 was activated by deprotonation of the bis(phosphonium salt) using a polymer-
bound phosphazene base (BEMP resin) in CH2Cl2. A resin-bound base was found
to facilitate reaction monitoring by mass spectrometry. The resin was added to
the solution of 12 in CH2Cl2 at ꢀ78ꢁC, and the reaction mixture was allowed to
warm to room temperature, forming the corresponding bis(phosphonium ylide) of
the machine. The subsequent intra-machine reaction was allowed to proceed until
the product distribution no longer changed (as evidenced by mass spectrometry):
1 day for machine 12, 3 days for machine 14, and 5 days for machine 1. The machine
products, the sequence oligomer and the phosphine oxide track, were then isolated
after a short work-up procedure followed by extensive chromatography (see Sec-
tions S4 and S6). Phosphine oxides are often difficult to remove by chromatography
and, indeed, this proved to be the case with several of the molecular machine prod-
ucts. In addition to the fully elongated oligomeric product, small amounts of byprod-
ucts were detected by mass spectrometry that correspond to products missing one
diarylpropane monomer from the chain. As the pristine molecular machine-track
conjugates showed no evidence of missing barriers prior to operation (e.g., by
mass spectrometry), these byproducts apparently arise from hydrolysis of the build-
ing blocks from the track under the machine’s operating conditions. In the case of
the two-building-block machine operation, barrier loss corresponds to ꢂ5% of the
Chem 6, 1–10, November 5, 2020
5

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
isolated products, i.e., ꢂ2.5% loss per barrier. This rises to approximately 4% barrier
loss during the operation of four-building-block machine 1.
When non-interlocked thread 9 was reacted with macrocycle 2 (Figure 3A, for syn-
thesis and operation see Sections S4 and S5) the stilbene product (10) was formed
in a 58:42 E:Z ratio of alkene isomers.33 However, when the same components
were interlocked as rotaxane 11, the E:Z ratio of 10 formed was 88:12, indicating a
significant stereochemical effect of the restricted vector of the macrocycle-ap-
pended aldehyde’s approach to the phosphonium ylide in the rotaxane, favoring
the E-alkene-forming oxaphosphetane. The operation of the two-building-block
machine 12 formed stilbene product 13 and had an 87:13 E:Z ratio for the first
formed alkene, while the second formed in a 55:45 ratio, apparently as a conse-
quence of the greater flexibility of the extended oligomer chain (Figure 3B; Section
S6). Operation of three-building-block machine 14 under the same reaction condi-
tions afforded the anticipated tri-stilbene product 15 (Figure 3C; Section S6). The
sequence of 15 was confirmed by tandem MS/MS spectroscopy (Figure S22), con-
firming that the machine selectively extracts the building blocks embedded in the
track in sequence. Although we were unable to fully assign the stereochemistry of
all of the alkenes in 15 due to a lack of clearly resolved signals in the 1H NMR, we
were able to determine that the right-hand-terminus alkene of 15 is formed in a
73:27 E:Z ratio. If the stereochemistry of the first two alkenes mirrors that of the
one- and two-building-block machines 11 and 12, then the stereochemistry of the
machine product from 14 is approximately E(87%),E(55%),E(73%)-15.
Molecular machine 1 was activated by deprotonation of the tetra(phosphonium salt)
using the polymer-bound phosphazene base (BEMP resin) in CH2Cl2, forming the
corresponding tetra(phosphonium ylide) (Figure 1, step viii). Initially, macrocycle 2
is constrained to the left-most compartment by the first barrier. However, after a Wit-
tig reaction between the aldehyde of the macrocycle and the first phosphonium
ylide, the resulting phosphine oxide is not large enough to prevent the macrocycle
from passing. The macrocycle now accesses the second compartment as well as the
first, shuttling at random over the length of the increased amount of accessible
thread until it reacts with the second barrier, forming another carbon-carbon double
bond. This process is repeated twice more with barriers three and four. Once the
fourth alkene connection is generated, all of the phosphonium ylides have been
removed from the track, leaving the smaller phosphine oxides in their place, and
the macrocycle bearing the oligomeric product, 7, dethreads from the residual track,
8. After 5 days, mass spectrometry indicated complete consumption of the starting
material. Size-exclusion chromatography followed by preparative thin-layer chroma-
tography led to the isolation of two major products: the tetra(phosphine oxide)
thread 8 (69%) and oligomer 7 (21% isolated yield; the low yield is a reflection of
the material lost during chromatography), obtained as a mixture of stereoisomers
(E/Z for each of the four alkenes).
The sequence of the building blocks in 7 was unambiguously determined by tandem
mass spectrometry (MS/MS). Fragmentation along the chain of 7 shows sequential
loss of fragments from each monomer strictly in the anticipated order, demon-
strating the sequence integrity of the molecular machine operation (Figures 4A–4C).
The regulation of the building-block sequence in 7 corresponds to control over the
primary structure of a sequence polymer, a degree of order critical in proteins, where
a single monomer missing or positioned out of sequence can cause loss or change of
function.34 The next level of structural order in proteins, secondary structure,
6
Chem 6, 1–10, November 5, 2020

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
Figure 4. Sequence of Machine Operation Product 7 (Formed as a Mixture of Stereoisomers)
(A) Fragmentation of 7 that occurs during tandem mass spectrometry.
(B and C) Tandem mass spectrometry (ESI+) fragmentation of 7 at (B) low and (C) high m/z values.
(D) Ribbon representation of the helical structure of Z,Z,Z,Z-7, a stereoisomer of 7 formed with a fixed secondary structure.
(E) Energy-minimized (MMFF) conformation of truncated Z,Z,Z,Z-7.
corresponds to the 3D shape of local segments (e.g., motifs such as a-helices).34 The
E- or Z- configuration of the stilbene units offers at least the potential to control this
in carbon backbone polymers produced by molecular machines such as 1.35 For
example, molecular modeling indicates that the all-Z- stereoisomer of 7 should
adopt a helical conformation with the two side chains per residue directed outward
˚
with a pitch of ꢂ7 A and two monomers per turn (Figures 4D and 4E; in comparison, a
˚
peptide a-helix has one side chain per residue, a pitch of 5.4 A, and 3.6 monomers
per turn). However, although we were able to determine that the right-hand terminus
alkene of 7 forms in a 70:30 E:Z ratio (similar to the E:Z ratio obtained with the oper-
ation of the three-barrier machine), suggesting that the E:Z distribution is approxi-
mately E(87%),E(55%),E(73%),E(70%)-7 by extrapolating from the shorter machine
products,36 the complete distribution of the alkene stereochemistry in 7 could not
be established unambiguously. Furthermore, attempts to isomerize all of the al-
kenes in 7, 13, or 15 to a single stereoisomer (for example, by extended heating
or treatment with I2 to try to form the all-E isomer or photochemistry to try to favor
the all-Z isomer)37 were unsuccessful, with forcing conditions leading to decompo-
sition rather than uniform alkene stereochemistry. Control of chain secondary struc-
ture38 will be one of the next major challenges for continuous-carbon-backbone
oligomers9 and polymers built by artificial molecular machines.
DISCUSSION
Artificial molecular machines that can perform tasks in chemical synthesis12–20,39–45
strain the limits of rational design46 given current levels of understanding of chem-
istry. For a molecular machine to build a new class of sequence polymer, the machine
must use building blocks that are sufficiently robust, both initially and as intermedi-
ates, for them to be stable while attached to the track (in this case not undergoing
side reactions for up to 5 days under the reaction conditions used for machine oper-
ation) but sufficiently reactive to be abstracted by the machine-tethered chain. The
track must not be able to fold (to avoid reactions occurring out of sequence) yet not
be so rigid as to cause problems with the solubility of either the initial machine-track
conjugate or subsequent intermediates. The macrocycle cavity and building-block
size and shape need to be such that the macrocycle cannot pass a monomer on
Chem 6, 1–10, November 5, 2020
7

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
the track but can (as freely as possible) pass over the derivatized track once the build-
ing block is abstracted. The design must also cause the macrocycle to thread onto
the track, while not subsequently having attractive intercomponent interactions
that would slow the machine dynamics. During machine operation the reactive inter-
mediates on the chain and track must not be able to react with other functionalities
present in the machine, track, or other molecules (product, reagents, solvent, waste,
impurities, etc.). The chain tethered to the macrocycle must be able to adopt a
conformation that allows it to react with a building block on the track, using chem-
istry that simultaneously breaks a covalent bond (detaching the monomer from
the track) while forming another that adds the unit to the end of the growing chain.
Finally, the linking chemistry must be able to take place iteratively and repeatedly,
each time in high yield, without being so reactive that intermolecular reactions
occur.
Our approach12–15,17–19,39,41,43,47 to such challenges is to develop an initial concept
based on potentially appropriate chemistry, and apply multiple iterations of molec-
ular modeling, to give an idea of sizes, shapes, conformations, and the dynamics of
various components, together with experimental studies that probe various aspects
of the design and determine what works in practice with molecules of increasing
complexity.48 The lessons learned feed back into further iterations of (hopefully
improved) designs culminating in the assembly of a machine conjugate, such as 1,
that successfully performs the task. Such a process reveals effective chemistries, fea-
tures, and design principles that can be applied and adapted for use with other mo-
lecular machine systems, while also sometimes identifying issues and challenges for
the future. For track-based molecular synthesizers such future challenges include:
(1) The ability to reload building blocks (with sequence information), enabling a
single machine to ‘‘print’’ multiple copies of a product.
(2) The ability to ‘‘read’’ information, not just translate a sequence from one form
into another, perhaps using a molecular ‘‘tape’’ that enables the machine to
be programmed41,43 to produce different outcomes.
(3) Incorporating ratchet mechanisms,49 so that the molecular synthesizer is
located only in the region of the track where the next reaction needs to
take place.
(4) Machines that work together, and/or in tandem, to carry out more sophisti-
cated synthetic tasks than a single machine.
All these features, in some form or other, can be found in various biomolecular
synthesizers.
The ubiquity and function of sequence polymers in biology illustrates the tremen-
dous potential of sequence polymers assembled from unnatural building blocks
and linking chemistries.5 A carbon-carbon-bond-forming artificial small-molecule
machine, which moves along a track performing four iterative Wittig reactions to
generate a sequence oligomer with a continuous backbone of carbon-carbon
bonds, has been demonstrated to autonomously carry out a task in chemical synthe-
sis that biological machines cannot do. Excellent control of primary structure in the
machine product was established and there was significant influence of the machine
on the stereochemistry of some building-block connections that augurs well for
future systems being able to control product secondary structure. The design and
operation of de novo molecular synthesizer 1, which is able to position reactive
building blocks during a multistep reaction sequence and utilize chemistries and
reactivity patterns unavailable to biological machines, offers a tantalizing glimpse
8
Chem 6, 1–10, November 5, 2020

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
of how track-based molecular synthesizers,12–15 and other types of molecular assem-
blers43 may one day be able to provide novel solutions to significant problems in
chemical synthesis.
EXPERIMENTAL PROCEDURES
Resource Availability
Lead Contact
David A. Leigh, david.leigh@manchester.ac.uk
Materials Availability
This study did not generate new unique reagents.
Data and Code Availability
This study did not generate datasets or codes.
SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.chempr.
2020.09.021.
ACKNOWLEDGMENTS
We thank the Engineering and Physical Sciences Research Council (EPSRC; EP/
P027067/1), the European Research Council (ERC Advanced grant to D.A.L.,
786630) and East China Normal University for funding, and the University of Man-
chester Mass Spectrometry Service Centre for high-resolution mass spectrometry.
C.T.M. thanks the Leverhulme Trust and the Isaac Newton Trust, and Sidney Sussex
College, Cambridge, for Fellowship support during the preparation of the manu-
script. G.D.B. is a Royal Society University Research Fellow; D.A.L. is a Royal Society
Research Professor.
AUTHOR CONTRIBUTIONS
C.T.M. and G.D.B. planned the work. C.T.M. carried out the experiments and
analyzed the data. D.A.L. directed the research. All authors contributed to the
design of the project and the writing of the manuscript.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: July 7, 2020
Revised: August 3, 2020
Accepted: September 23, 2020
Published: October 15, 2020
REFERENCES
1. Geyer, R., Jambeck, J.R., and Law, K.L. (2017).
Production, use, and fate of all plastics ever
made. Sci. Adv. 3, e1700782.
4. Lawrence, J., Lee, S.H., Abdilla, A., Nothling,
M.D., Ren, J.M., Knight, A.S., Fleischmann, C.,
Li, Y., Abrams, A.S., Schmidt, B.V.K.J., et al.
(2016). A versatile and scalable strategy to
discrete oligomers. J. Am. Chem. Soc. 138,
6306–6310.
information into polymers. Nat. Rev. Chem. 2,
365–381.
7. Voet, D., and Voet, J.G. (2011). Biochemistry,
Fourth Edition (John Wiley & Sons).
2. Matyjaszewski, K. (2011). Chemistry.
Architecturally complex polymers with
controlled heterogeneity. Science 333, 1104–
1105.
8. He, Y., and Liu, D.R. (2010). Autonomous
multistep organic synthesis in a single
isothermal solution mediated by a DNA walker.
Nat. Nanotechnol. 5, 778–782.
5. Lutz, J.F., Ouchi, M., Liu, D.R., and Sawamoto,
M. (2013). Sequence-controlled polymers.
Science 341, 1238149.
3. Lutz, J.-F., Lehn, J.-M., Meijer, E.W., and
Matyjaszewski, K. (2016). From precision
polymers to complex materials and systems.
Nat. Rev. Mater. 1, 16024.
6. Rutten, M.G.T.A., Vaandrager, F.W., Elemans,
J.A.A.W., and Nolte, R.J.M. (2018). Encoding
9. Meng, W., Muscat, R.A., McKee, M.L., Milnes,
P.J., El-Sagheer, A.H., Bath, J., Davis, B.G.,
Chem 6, 1–10, November 5, 2020
9

Please cite this article in press as: McTernan et al., A Track-Based Molecular Synthesizer that Builds a Single-Sequence Oligomer through Iter-
ative Carbon-Carbon Bond Formation, Chem (2020), https://doi.org/10.1016/j.chempr.2020.09.021
ll
OPEN ACCESS
Article
¨
Brown, T., O’Reilly, R.K., and Turberfield, A.J.
(2016). An autonomous molecular assembler
for programmable chemical synthesis. Nat.
Chem. 8, 542–548.
24. Wittig, G., and Schollkopf, U. (1954). Uber
¨
38. Hill, D.J., Mio, M.J., Prince, R.B., Hughes, T.S.,
and Moore, J.S. (2001). A field guide to
foldamers. Chem. Rev. 101, 3893–4012.
triphenyl-phosphin-methylene als
olefinbildende reagenzien (I. Mitteil). Chem.
Ber. 87, 1318–1330.
39. Hannam, J.S., Lacy, S.M., Leigh, D.A., Saiz,
C.G., Slawin, A.M.Z., and Stitchell, S.G. (2004).
Controlled submolecular translational motion
in synthesis: a mechanically interlocking
auxiliary. Angew. Chem. Int. Ed. Engl. 43, 3260–
3264.
10. Gu, H., Chao, J., Xiao, S.J., and Seeman, N.C.
(2010). A proximity-based programmable DNA
nanoscale assembly line. Nature 465, 202–205.
25. McDonald, R.N., and Campbell, T.W. (1960).
The Wittig reaction as a polymerization
method 1a. J. Am. Chem. Soc. 82, 4669–4671.
11. Erbas-Cakmak, S., Leigh, D.A., McTernan, C.T.,
and Nussbaumer, A.L. (2015). Artificial
molecular machines. Chem. Rev. 115, 10081–
10206.
26. Anuragudom, P., Newaz, S.S., Phanichphant,
S., Lee, T.R., and Horner-Emmons, F. (2006).
Facile HornerꢀEmmons synthesis of defect-
free poly(9,9-dialkylfluorenyl-2,7-vinylene).
Macromolecules 39, 3494–3499.
40. Chao, S., Romuald, C., Fournel-Marotte, K.,
Clavel, C., and Coutrot, F. (2014). A strategy
utilizing a recyclable macrocycle transporter for
the efficient synthesis of a triazolium-based [2]
rotaxane. Angew. Chem. Int. Ed. Engl. 53,
6914–6919.
12. Lewandowski, B., De Bo, G., Ward, J.W.,
Papmeyer, M., Kuschel, S., Aldegunde, M.J.,
Gramlich, P.M.E., Heckmann, D., Goldup, S.M.,
D’Souza, D.M., et al. (2013). Sequence-specific
peptide synthesis by an artificial small-
27. De Juan, A., Pouillon, Y., Ruiz-Gonza´lez, L.,
Torres-Pardo, A., Casado, S., Martı´n, N., Rubio,
´
A., and Pe´ rez, E.M. (2014). Mechanically
interlocked single-wall carbon nanotubes.
Angew. Chem. Int. Ed. Engl. 53, 5394–5400.
molecule machine. Science 339, 189–193.
41. Beswick, J., Blanco, V., De Bo, G., Leigh, D.A.,
Lewandowska, U., Lewandowski, B., and
Mishiro, K. (2015). Selecting reactions and
reactants using a switchable rotaxane
organocatalyst with two different active sites.
Chem. Sci. 6, 140–143.
13. De Bo, G., Kuschel, S., Leigh, D.A.,
Lewandowski, B., Papmeyer, M., and Ward,
J.W. (2014). Efficient assembly of threaded
molecular machines for sequence-specific
synthesis. J. Am. Chem. Soc. 136, 5811–5814.
28. Balakrishna, B., Menon, A., Cao, K., Gsanger,
¨
S., Beil, S.B., Villalva, J., Shyshov, O., Martin, O.,
Hirsch, A., Meyer, B., et al. (2020). Dynamic
covalent formation of concave disulfide
macrocycles mechanically interlocked with
single-walled carbon nanotubes. Angew.
Chem. Int. Ed. Engl. https://doi.org/10.1002/
anie.202005081.
42. Martinez-Cuezva, A., Lopez-Leonardo, C.,
Bautista, D., Alajarin, M., and Berna, J. (2016).
Stereocontrolled synthesis of b-lactams within
[2]rotaxanes: showcasing the chemical
consequences of the mechanical bond. J. Am.
Chem. Soc. 138, 8726–8729.
14. De Bo, G., Gall, M.A.Y., Kitching, M.O.,
Kuschel, S., Leigh, D.A., Tetlow, D.J., and
Ward, J.W. (2017). Sequence-specific
b-peptide synthesis by a rotaxane-based
molecular machine. J. Am. Chem. Soc. 139,
10875–10879.
29. Aucagne, V., Hanni, K.D., Leigh, D.A., Lusby,
¨
P.J., and Walker, D.B. (2006). Catalytic ‘‘click’’
rotaxanes: a substoichiometric metal-template
pathway to mechanically interlocked
architectures. J. Am. Chem. Soc. 128, 2186–
2187.
15. De Bo, G., Gall, M.A.Y., Kuschel, S., De Winter,
J., Gerbaux, P., and Leigh, D.A. (2018). An
artificial molecular machine that builds an
asymmetric catalyst. Nat. Nanotechnol. 13,
381–385.
43. Kassem, S., Lee, A.T.L., Leigh, D.A., Marcos, V.,
Palmer, L.I., and Pisano, S. (2017).
Stereodivergent synthesis with a
programmable molecular machine. Nature
549, 374–378.
30. Crowley, J.D., Goldup, S.M., Lee, A.L., Leigh,
D.A., and McBurney, R.T. (2009). Active metal
template synthesis of rotaxanes, catenanes and
molecular shuttles. Chem. Soc. Rev. 38, 1530–
1541.
16. Thordarson, P., Bijsterveld, E.J.A., Rowan, A.E.,
and Nolte, R.J.M. (2003). Epoxidation of
polybutadiene by a topologically linked
catalyst. Nature 424, 915–918.
44. Riss-Yaw, B., Clavel, C., Laurent, P., Waele` s, P.,
and Coutrot, F. (2018). The importance of
length and flexibility of macrocycle-containing
molecular translocators for the synthesis of
improbable [2]rotaxanes. Chemistry 24, 13659–
13666.
31. Denis, M., and Goldup, S.M. (2017). The active
template approach to interlocked molecules.
Nat. Rev. Chem. 1, 61.
17. Fuller, A.M., Leigh, D.A., and Lusby, P.J. (2007).
One template, multiple rings: controlled
iterative addition of macrocycles onto a single
binding site rotaxane thread. Angew. Chem.
Int. Ed. Engl. 46, 5015–5019.
32. Meldal, M., and Tornøe, C.W. (2008). Cu-
catalyzed azide-alkyne cycloaddition. Chem.
Rev. 108, 2952–3015.
45. Martinez-Cuezva, A., Saura-Sanmartin, A.,
Alajarin, M., and Berna, J. (2020). Mechanically
interlocked molecules for asymmetric catalysis.
ACS Catal. 10, 7719–7733.
18. Fuller, A.M., Leigh, D.A., and Lusby, P.J. (2010).
Sequence isomerism in [3]rotaxanes. J. Am.
Chem. Soc. 132, 4954–4959.
33. Likhtenshtein, G. (2010). Stilbenes.
Applications in Chemistry, Life Sciences and
Materials Science (Wiley-VCH Verlag GmbH &
Co. KGaA).
46. Heard, A.W., and Goldup, S.M. (2020).
Simplicity in the design, operation, and
applications of mechanically interlocked
molecular machines. ACS Cent. Sci. 6, 117–128.
19. Erbas-Cakmak, S., Fielden, S.D.P., Karaca, U.,
Leigh, D.A., McTernan, C.T., Tetlow, D.J., and
Wilson, M.R. (2017). Rotary and linear
34. Fersht, A. (1998). Structure and mechanism in
protein science (Freeman).
molecular motors driven by pulses of a
chemical fuel. Science 358, 340–343.
47. Zhang, L., Marcos, V., and Leigh, D.A. (2018).
Molecular machines with bio-inspired
mechanisms. Proc. Natl. Acad. Sci. USA 115,
9397–9404.
35. Nakano, T., and Okamoto, Y. (2001). Synthetic
helical polymers: conformation and function.
Chem. Rev. 101, 4013–4038.
20. Qiu, Y., Song, B., Pezzato, C., Shen, D., Liu, W.,
Zhang, L., Feng, Y., Guo, Q.H., Cai, K., Li, W.,
et al. (2020). A precise polyrotaxane
36. It is not clear whether the geometry formed in
the first Wittig reaction affects the geometries
of subsequent Wittig reactions. The
stereochemistry indicated are averaged values
based on the outcome of the shorter machine
outputs.
48. Our general approach is often to first use CPK
modelling to assess size and fit, followed by
simple molecular mechanics calculations to
assess conformations (along with DFT
synthesizer. Science 368, 1247–1253.
21. Steitz, T.A. (2008). A structural understanding
of the dynamic ribosome machine. Nat. Rev.
Mol. Cell Biol. 9, 242–253.
calculations in certain cases), before moving to
the synthesis of model systems to inform
machine design (see Section S7).
22. Drexler, K.E. (1986). Engines of Creation: The
Coming Era of Nanotechnology (Anchor).
37. Yu, Z., and Hecht, S. (2013). Control over
unfolding pathways by localizing
49. Kay, E.R., and Leigh, D.A. (2015). Rise of the
molecular machines. Angew. Chem. Int. Ed.
Engl. 54, 10080–10088.
23. Kelly, T.R., and Snapper, M.L. (2017).
Nanotechnology: a molecular assembler.
Nature 549, 336–337.
photoisomerization events within
heterosequence oligoazobenzene foldamers.
Angew. Chem. Int. Ed. Engl. 52, 13740–13744.
10
Chem 6, 1–10, November 5, 2020