On-Nanoparticle Gating Units Render an Ordinary Catalyst Substrate- And Site-Selective

Substrate- and Site-Selective

Article

On-Nanoparticle Gating Units Render an Ordinary Catalyst

Minju Kim, Miroslaw Dygas, Yaroslav I. Sobolev, Wiktor Beker, Qiang Zhuang, Tomasz Klucznik,

Guillermo Ahumada, Juan Carlos Ahumada, and Bartosz A. Grzybowski*

Cite This: J. Am. Chem. Soc. 2021, 143, 1807 −1815

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: When an organometallic catalyst is tethered onto a nanoparticle and is

embedded in a monolayer of longer ligands terminated in “gating” end-groups, these groups

can control the access and orientation of the incoming substrates. In this way, a nonspeciﬁc

catalyst can become enzyme-like: it can select only certain substrates from substrate mixtures

and, quite remarkably, can also preorganize these substrates such that only some of their

otherwise equivalent sites react. For a simple, copper-based click reaction catalyst and for

gating ligands terminated in charged groups, both substrate- and site-selectivities are on the

order of 100, which is all the more notable given the relative simplicity of the on-particle

monolayers compared to the intricacy of enzymes’ active sites. The strategy of self-

assembling macromolecular, on-nanoparticle environments to enhance selectivities of

“ordinary” catalysts presented here is extendable to other types of catalysts and gating based

on electrostatics, hydrophobicity, and chirality, or the combinations of these eﬀects. Rational

design of such systems should be guided by theoretical models we also describe.

INTRODUCTION

research has focused on biomedical and sensing applications

capitalizing on the intrinsic properties of the nanoparticle

■

Enzymes are remarkable for their ability to perform reactions

selectively, only on the “ﬁtting” substrates and often only on a

speciﬁc group even if other, chemically equivalent groups are

cores, but there have also been some exciting developments

in which the organic ligand shell was engineered to endow

6,37

substrate speciﬁcity,

to template association of substrates

present in the same substrate. Such substrate- and site-

38−40

and their subsequent catalytic conversion,

dynamic layer for in-cell or photocontrolled catalysis (for

further examples, see recent reviews

or to act as a

selectivities derive from substrate recognition and preorganiza-

tion by the loci of the binding pockets that may be quite

distant from the catalytic center (Figure 1a). In the last several

decades, chemists have strived to build-in similar capabilities

33,43

). On the downside,

many of these systems achieve only moderate selectivities

(especially with respect to equivalent functionalities), are

tailored to very speciﬁc substrates (Figure 1b), and often

require elaborate synthesis.

In this work, we sought to “augment” an otherwise

nonselective organometallic catalyst (here, a popular Bipy-

Cu(I) for click reaction) and endow it with a high degree of

substrate and site-selectivity not by covalent modiﬁcation but

by self-assembly of a macromolecular environment controlling

access to the catalytic center. We hypothesized this objective

could be achieved by rather simple means, namely, by the

−26

into artiﬁcial systems.

Indeed, there have been numerous

−4

creative eﬀorts to prepare supramolecular enzyme mimics

presenting substrate-speciﬁc “pockets” based on molecular

6,7

9,10

sieves, cages,

ﬂasks, cucurbituril derivatives,

capsu-

bidentate ligands with

and several others. These

1,12

13,14

les,

porphyrin-based complexes,

5,16

built-in anion recognition sites,

systems can discriminate between substrates based on

,11,14,17

5,17

size,

presence of anionic groups

whose formation is disfavored in typical solution-phase

shape,

transition state conformation, or

5,16

and can result in products

,18,19

formation of a mixed, on-nanoparticle self-assembled mono-

reactions.

For instance, Davis and Jones demonstrated

44,45

layer, SAM,

in which the “gating” end groups of longer

substrate-selective catalysis (or separations) on various types of

organic-functionalized molecular sieves (OFMSs) or zeo-

ligands would control the approach and orientation of the

,20−25

26−28

lites.

and Reek

Groups of Breslow

(already in the 1990s)

9−31

Received: September 1, 2020

Published: January 20, 2021

and Bayley (more recently) synthesized

several elegant systems in which site-selectivity is enhanced by

supramolecular interactions mediating preorganization of

substrates. Simultaneously, there has been a signiﬁcant eﬀort

on the so-called nanozymes, that is, nanoparticles mimicking

3,34

the activity of enzymes.

A signiﬁcant portion of this

2021 American Chemical Society

J. Am. Chem. Soc. 2021, 143, 1807−1815

807

Article

by a theoretical model and simulations accompanying our

experiments.

Overall, the system we describe, and others that could be

designed based on similar principles for diﬀerent nonspeciﬁc

catalysts, can be construed as a hybrid between nanozyme and

macromolecular approaches to enzyme mimicry: (i) the

nanoparticle supports are important insofar as they present

high surface area and can also control the curvature of the

monolayer and the overall “enzyme” activity (Figure 6b,c),

whereas (ii) the substrate- and site-selectivities derive from the

details of the on-particle macromolecular environments.

RESULTS AND DISCUSSION

Figure 1c,d illustrates the architecture of our system whereby

■

gold nanoparticles (typically, 4.2 ± 0.4 nm in size but also 2.5

0.3 nm in some experiments discussed later) support a

mixed monolayer comprised of longer alkanethiols terminated

in positively charged, “gating” end-groups (N,N,N-trimethyl-

(

20-mercaptoicosanyl) ammonium bromide; TMA) and short-

er ligands terminated in bipyridine units (6-((5′-methyl-[2,2′-

bipyridine-5-yl)oxy)hexane-1-thiol; Bipy). Once made and

puriﬁed, the NPs were redispersed in 1:1 v/v water/MeOH

mixture at a concentration of ∼5.5 μM in terms of the NPs and

were then exposed to 10-fold excess CuI, allowing for the

coordination of copper to the Bipy units. These Bipy-Cu(I)

units were intended to act as catalysts of the so-called “click”

reaction, that is, an azide−alkyne cycloaddition leading to 1,4-

substituted 1,2,3-triazoles, which is a popular and powerful

Figure 1. Site-selectivity in enzymes and in enzyme mimics. (a)

Enzymes can recognize a speciﬁc place in a substrate by the use of

binding sites sometimes quite distant from the active site. For

example, trypsin cleaves only one (colored in red) out of multiple

peptide bonds present, and this recognition is mediated by a long,

positively charged side chain (in the ﬁgure, arginine) that ﬁts into a

−

distant pocket and interacts with a COO group present therein. (b)

Example of site-selectivity in an artiﬁcial system from Reek’s group

strategy for covalently joining molecules under a wide range of

III

(

ﬁgure adapted from ref. ). A Mn −porphyrin is functionalized with

46−48

conditions.

Afterward, any excess CuI was removed from

cyclodextrin (CD) units linked to all meso-phenyl rings. When the

termini of an appropriately functionalized steroid derivative bind to

the CDs, the steroid’s C6 α position is placed precisely at the catalytic

center, resulting in hydroxylation of this CH function with

regioselectivity exceeding 90%. (c) General principle of on-nano-

particle, charge-based gating. In scenario “1”, the positively charged

end groups of the longer thiols interact repulsively with the positive

group on the incoming substrate (here, a dialkyne partner for the click

reaction). Consequently, substrate prefers to enter the ligand shell in

orientation illustrated in “2”. Only one alkyne group is expected to

react with the azide partner at the catalytic center (here, blue circles

denoting Cu-Bipy complexes at the end of shorter thiols). (d) A

snapshot of a Molecular Dynamics (MD) simulation showing the

nanoparticle and its ligand shell, with molecules’ contours traced by

van der Waals surfaces. Green = copper atoms coordinated to blue

Bipy thiols; orange = TMA ligands terminated in red positively

charged quaternary nitrogen groups; yellow Au = atoms of the NP

core. Not shown are substrate and product molecules. For details of

MD simulations, see the Supporting Information (SI), Section S10.3.

solution by multiple rounds of centrifugation and methanol

was removed by evaporation to yield ∼11 μM suspension of

NPs in water. Analyses by inductively coupled plasma atomic

emission spectroscopy (ICP-AES) showed that (i) concen-

trations of Cu in solution right after the synthesis and CuI

centrifugation, after 24 h, and also after using NPs for click

reactions were all below detection limit, and (ii) Cu was

localized on the NPs and the measured ratio of Au to Cu

matched the one expected for the ca. 50% content of Bipy-

Cu(I) units in the on-particle ligand shells. Additional control

experiments evidenced only residual catalysis for CuI-puriﬁed

solutions (nonselective conversions ∼1% at 24 h). Also,

AuNPs covered with only TMA, and no Bipy ligands showed

no catalytic conversion, excluding the possibility of Cu salts

49,50

adsorbing onto the NPs via the so-called salt adsorption.

For details of all these experiments, see the SI, Sections S7 and

S8 .

We emphasize six important aspects of the system’s design.

First, The TMA thiols were chosen as gating units because

they retain positive charge and the AuNPs they stabilize

remain soluble in water irrespective of pH. Second, with TMA

thiols shorter than Bipy thiols, the NP suspensions remain

stable prior to the addition of CuI but, after such addition, they

rapidly ﬂocculate, suggesting some form of “cross-linking” via

the Bipy-Cu(I) units (perhaps by the formation of binuclear

substrates with respect to the catalytic centers tethered at the

ends of shorter ligands (Figure 1c,d).

Although the on-particle monolayers are signiﬁcantly less

intricate than an enzymes’ active sites, they achieve

unexpectedly high selectivities. With charged gating groups,

these selectivities are as high as ∼100 for competitive selection

from mixtures of negatively and positively charged substrates

51−53

complexes, see

). Third, increasing the fraction of Bipy-

(

Figure 3), and tens to over 100 for the selection between

Cu(I) in the monolayer lowers the particles’ solubility in

waterthe ∼1:1 ratio we used here appears to be an optimal

trade-oﬀ between high catalyst loading and NP stability.

Fourth, because of electrostatic repulsions between the TMA

head-groups, the TMA thiols are unlikely to phase-separate

and form distinct patches on NP surfacesin fact, computer

simulations summarized in Figure S92 indicate that such

locally equivalent reaction sites (here, triple bonds) within the

same substrates (Figure 4). Even for the gating based on

weaker, van der Waals interactions, the selectivities are still

appreciable, 6−7 fold; in this case, the hydrophobic gating

units admit preferentially the more hydrophobic substrates

(Figure 6g−i). These experimental ﬁndings are substantiated

808

J. Am. Chem. Soc. 2021, 143, 1807−1815

Article

Figure 2. Selectivity for negatively- vs positively- charged substrates. (a) Scheme of the nanoparticle, the incoming substrates, and possible

products. Molecular details of the ligands are shown at the bottom. Red glow around the particle indicates electrostatic potential due to positively

charged TMA head-groups. (b) Speciﬁc substrates used and (c) conversions observed for diﬀerent pairs of substrates on nanoparticles (orange

bars) and for the copper catalysts dispersed in solution (green bars; same Cu concentration as for NPs). “0” means uncharged substrate; “+” or “−”

indicate, respectively, positively or negatively charged substrates. The ﬁrst symbol in the pair denotes azide and the second, alkyne. Error bars are

standard deviations from three independent experiments (whose individual outcomes are quantiﬁed by gray markers in the histograms). (d) Kinetic

plots tracing concentrations of products from panel (c) in the function of time. Lines are exponential ﬁts to experimental data.

remained similar to the initial particles (see SI, Section S8).

Sixth, after completion of a catalysis run, the NPs could be

reused for another round with similar results (see Figure S34).

Turning to the actual click reactions, we ﬁrst tested whether

the NPs can selectively catalyze click reactions depending on

substrate charge (Figure 2a). For these comparisons, we chose

pairs of alkynes/azides (Figure 2b) in which one was neutral,

“

0”, and one charged “+” or “−”; in this way, the electrostatic

interactions were strictly between one of the substrates and the

NP, and not between the substrates (still, for completeness,

data for “++”, “− −”, “+−” and “−+” combinations are

included in Figure S40). All reactions were monitored up to 24

h at r.t. Analyses of the reaction mixtures by H NMR

summarized by the orange bars in Figure 2c and also kinetic

plots in Figure 2d evidence that the conversions are high as

long as one of the substrates bears negative charge (conversion

8.8 ± 0.52% for “0−” and 88.3 ± 2.78% for “−0”). In sharp

contrast, when the azide or the alkyne are positively charged

“0+” and “+0”), the conversion drops to ca. 10%. This result

Figure 3. Experiments with “competing” substrates. Substrates of

both polarities (alkynes in “entry 1” and azides in “entry 2”) compete

for reaction with a neutral azide/alkyne. Histograms on the right

(

echoes previous studies by us and others on the ability of

TMA nanoparticles to admit negatively charged nitrophenol

substratesin those studies, however, nitrophenol was

reduced at AuNP surfaces to a neutral aniline whereas, in

our current system, charge is still present on the click reaction’s

products. The fact that conversions for “0−” and “−0” pairs are

so high evidence not only that the positively charged NP

attracts negatively charged substrates but also that negatively

charged substrates and/or reaction products are not stuck in

the TMA monolayer and are not “poisoning” the NP catalyst.

Importantly, control experiments using equal amounts of

evidence that while CuSO /NaAs catalyst (green bars) does not show

pronounced preference for negatively charged substrates, the NP

catalystcarrying the same amount of copperchooses such

substrates with selectivities [azide(0)alkyne(−)]/[azide(0)alkyne(+)]

35 and [alkyne(0)azide(−)]/[alkyne(0)azide(+)] = 110. Error bars

are standard deviations from three independent experiments (whose

individual outcomes are quantiﬁed by gray markers in the histo-

grams).

repulsions are as high as ∼5 kT per thiol such that formation of

pure TMA patches (and, consequently, of pure Bipy patches

over remaining areas) would cost several hundred kT’s per

nanoparticle. Fifth, the particles were stable to ca. 80 °C − at

this temperature, some gradual coalescence/ripening of the

NPs was observed though their zeta potentials as well as

compositions of ligands shells determined by ICP-AES

copper free in water solution (as 0.46 mM CuSO /NaAs,

where As = ascorbate) rather than on NPs demonstrate no

similar selectivity. For the CuSO /NaAs system, all con-

versions are above 80% and as high as 97.4% for “−0” (green

bars in Figure 2c and green curves in Figure 2d).

809

J. Am. Chem. Soc. 2021, 143, 1807−1815

Article

Figure 4. Site-selective, on-nanoparticle catalysis. (a) Substrates 1−3 and their possible products (singly reacted, A and B, and doubly reacted, C).

Histograms quantify the distributions of these products at 24 h for CuSO /NaAs catalysts (green bars) and for the NP-based catalysts (orange bars;

same Cu concentration as for CuSO /NaAs). NP-based catalysts exhibit strong preference for products A (cf. Figure 1c). Gray markers = results of

individual experiments; error bars = standard deviations.

To further explore the selectivity of our NP catalysts, we

performed a series of experiments analogous to enzymatic

competitive-binding assays. In one example, we used a mixture

of approximately equal amounts of a neutral azide(0), a

negatively charged alkyne(−), and a positively charged

alkyne(+) (entry 1 in Figure 3) and analyzed the reaction

between the incoming substrate and gating/TMA charged

groups could preorient the substrate such that the reactive

group more distant from the charge center would enter the

monolayer ﬁrst and undergo the click reaction preferentially

(see Figure 1c). To test this hypothesis, we synthesized

substrates marked 1 through 3 bearing two terminal alkynes

(Figure 4; for synthetic details, see SI, Section S5). For each of

these substrates, reaction with an azide could lead to three

possible products − “A” singly reacted at an alkyne further

from the positively charged center, “B” singly reacted at an

alkyne closer to the positively charged center, and “C” doubly

reacted at both alkynes. As before, comparisons were made

products obtained in the presence of either CuSO /NaAs or

our NP catalysts (both at 24 h and 80 °C to maximize

conversions). In the ﬁrst case, CuSO /NaAs catalyzed

indiscriminately both of the possible reactions, and the

conversions for cycloadducts formed from azide(0)/alkyne(−)

and azide(0)/alkyne(+) were similar (respectively, 46.9% and

1.7%). However, when Cu(I)-loaded AuNPs were used, the

between CuSO /NaAs catalysis (green bars) and on-nano-

azide(0)/alkyne(−) product was formed in ca. 35-fold excess

over the azide(0)/alkyne(+) one (93.1% yield vs 2.6%).

Analogous results were observed when azide(−) and azide (+)

competed for reaction with a neutral alkyne(0) (entry 2 in

particle catalysis (orange bars) at 24 h (for details, see SI,

Section S6). As seen, for 1 through 3, the use of NPs always

increased the yield of product A and suppressed formation of B

and C − this eﬀect was observed irrespective of the native

Figure 3) − for CuSO /NaAs, there was only a slight

reactivity in the CuSO /NaAs system (approximately equal

preference for the azide(−)/alkyne(0) outcome but for NP

catalysts, this preference was much more pronounced (110-

fold; 95.5% vs 0.86%).

distribution of products for 1, favoring formation of B for 2, or

favoring formation of A for 3).

The site-selectivities of the “gated” NPs for A, deﬁned as a

Arguably, the most important results are for the situation

when the same substrate has two potentially reactive groups

but one of them is closer to the charge-bearing groupwe

expected that in this situation, the electrostatic interaction

ratio of singly reacted products, S_A_B= [A]/[B], are 74.8 for 1,

51.3 for 2, and 179.7 for 3. This measure, however, neglects

the doubly reacted product formed through A and B

intermediates. Accordingly, using solutions to kinetic equations

810

J. Am. Chem. Soc. 2021, 143, 1807−1815

Article

describing the kinetic system detailed in the SI Section S10,

selectivity can be more accurately deﬁned as S_A_B_C= ln(1 + C]/

[

B])/ ln(1 + [C]/[A]), which reduces to [A]/[B] for small

C]. For NP catalysts, S values are 44.2 for 1, 26.3 for 2, and

ABC

1 for 3. Yet another way of expressing these selectivities is

rel

AB Cu

NP AB

relative to the “native,” CuSO /NaAs values, S = S /S

rel

ABC

and S = S /S , such as to quantify the degree to which

ABC ABC

the NPs alone aﬀect preferential product formation. These

rel

values are S = 95.8 and S

= 52.3 for 1, S = 175.3 and

ABC

rel

= 73.1 for 2, and S = 22.4 and S

= 7.4 for 3.

ABC

While we do not attempt here to justify the native product

distributions in the CuSO /NaAs system (as they may reﬂect a

variety of factors, including details of coordination of the

substrate to the catalytic center), a simple theoretical model

can rationalize how the NPs impart the observed site-

preference. In this model, we consider a reaction L + M →

Z, in which a charged dialkyne L can have two orientations

with respect to the charged, catalytic surface of the

nanoparticle (in Figure 5a,c,d denoted L and L , respectively;

M stands for the azide). Depending on the orientation of L,

two reaction “pathways” with pseudo-ﬁrst-order reaction rates

k [L ] and k [L ], and two distinct, singly reacted products

are then possible, L + M → Z and L + M → Z (these

products can further react into disubstituted product C). Given

certain probabilities P(L ) = [L ]/[L] and P(L ) = [L ]/[L]

of molecule L to be oriented in two diﬀerent ways near the

catalytic surface of the nanoparticle, it can be shown (SI,

Section S10 ) that reactions forming Z and Z are pseudo-ﬁrst-

order in [L] with rate constants k and k whose ratio is k /k

k P(L )/(k P(L )) = W·P(L )/P(L ), where W = k /k is

A A B B A B A B

the “native” kinetic asymmetry between two reaction pathways.

Making a reasonable assumption that these orientations in

solution are thermally equilibrated, we can use Boltzmann

distribution to obtain k /k = W exp(−q(ϕ(r ) − ϕ(r ))/kT),

in which the diﬀerence in the energies of L and L is due to

the electric potential ϕ(r) of the charged nanoparticle and r =

r in orientation L , or r = r in orientation L . The electric

potential ﬁeld ϕ(r) is calculated by numerically solving full

nonlinear Poisson−Boltzmann (PB) equation for our nano-

particles, whereas the values of r and r are estimated from

Molecular Dynamics simulations (see SI, Sections S10.2 and

S10.3 ). In relation to experiments, k /k is equal to

experimental selectivity S_A_B_C, and approaches simple product

ratio S_A_B= [Z ]/[Z ] when disubstituted product is neglected.

Despite the relative simplicity of this model, it accurately

predicts the diﬀerences in kinetic rates in noncompetitive

reactions from Figure 2d (see SI, Section S10.2 ) as well as the

trends in selectivities S_A_B_Cobserved for particles harboring

diﬀerent ligands and/or having diﬀerent core sizes (Figures 5f,

Figure 5. Theoretical model. (a, c, d) Charged nanoparticle with

catalytic ligands (sky-blue) on its surface can be approached by a

substrate molecule (gray pins) oriented in two diﬀerent ways, L or

L . (b, e) Black curves trace potentials due to longer (C ) or shorter

C ) gating TMA ligands (dark green). MD calculations (Figure

S107) indicate that the eﬀective radius at which the potential becomes

approximately constant is shorter by ca. 0.6 nm than the average

radius at which TMA’s quaternary nitrogen groups reside. This

correction is incorporated into the PB model. Colored points on the

potential curves correspond to the potentials experienced by the

incoming ligands reaching the Bipy centers in two possible

orientations. Bipy centers are attached to NPs via (a, d) longer, C₆

a−c).

Speciﬁcally, an important parameter controlling selectivity is

ligand length. Figure 6a summarizes experiments in which the

lengths of the alkane chains were varied (C and C for Bipy

and C and C for TMA thiols). As seen, the C −Bipy/C −

alkane, chains or (c) shorter, C , chains. Δ’s are potential diﬀerences

TMA combination is more selective than either C −Bipy/

between L and L . (f) Comparison of theoretical and experimental

C −TMA (larger length diﬀerence) or C −Bipy/C −TMA

selectivities, S_A_B_Cfor substrate 1 and various nanoparticles (indicated

are lengths of Bipy and TMA ligands, as well as NP core size). Black

crosses are data from individual experiments, blue dots are mean

values, and blue error bars are standard deviations. Boxplots are

theoretical predictions computed based on the real experimental

distributions of NP sizes (see the SI, Section S10 for details). Boxplot

elements are median (orange line), 25%, and 75% quartiles (box

edges), maximum and minimum values (whiskers).

(

smaller length diﬀerence). This result means that selectivity is

not simply a function of the distance from the Bipy units to the

TMA gating groups. On the other hand, the results are well

explained by our electrostatic model. Speciﬁcally and with

reference to Figure 5, Gauss’ law necessitates that electric

potential inside a charged spherical shell is uniform in the PB

model (ﬂat segments of black curves in Figure 5b,e). In order

811

J. Am. Chem. Soc. 2021, 143, 1807−1815

Article

Figure 6. Diﬀerent eﬀects inﬂuencing gating and selectivity. (a) Experimentally measured conversions, into two singly reacted and one doubly

reacted products, for the on-particle click reactions between a positively charged dialkyne 1 and a negatively charged azide (same as in Figure 4, 6.9

and 2.3 mM, respectively). The particles all have 4.2 nm Au cores but diﬀer in the lengths of the Bipy and/or TMA ligands, as indicated in the

legend. Reactions were performed at 80 °C and conversions were measured at 24 h. (b) Scheme illustrating that the distance between the gating,

charged TMA groups decreases when the size of the nanoparticle increases. (c) Same reaction as in (a) but for nanoparticles diﬀering in the size of

Au cores. The concentrations of nanoparticles (10 μM for 2.5 nm NPs, 3.55 μM for 4.2 nm NPs) were such that the concentration of Bipy-Cu(I)

catalytic units was approximately the same in both cases (0.46 mM). (d) Singly and doubly reacted products from dialkyne substrate 4. The

distributions of these products at 24 h (80 °C) are quantiﬁed in the rightmost column for CuSO /NaAs catalysts (green bars) and for the NP-

based catalysts (orange bars). In panels b, c, and d, error bars are standard deviations from three independent experiments (whose individual

outcomes are quantiﬁed by gray markers in the histograms). (e) Scheme of pincer-like arrangement of 4 into the monolayer of NP catalysts. (f)

Corresponding snapshot from MD simulation. (g) A proof-of-the-principle example of hydrophobic/hydrophilic gating by n-alkane, C₂₂ligands

surrounding the catalytic centers on the particle. (h) Among the reaction substrates, the two alkynes were chosen such that they (1) were both

soluble in CHCl yet (2) diﬀered in the phobicity of their n-alkane vs tetra(ethylene glycol) “tails”. These substrates competed for the reaction with

the azide partner shown. The two possible products that can form are, correspondingly, labeled as A and B. (i) In the CuSO /NaAs system, there is

no selectivity, and A and B form in roughly equal proportions. In contrast, with Cu-Bipy/NP catalysts, the alkyne with n -alkane tail enters the

hydrophobic monolayer preferentially and gives product A in 7-fold excess. For all synthetic details, see the SI, Section S11 .

to reach the Cu atom of C −Bipy (Figure 5c), the dialkyne

and orange pairs of points in Figure 5b on a shared potential

curve (deﬁned by TMA ligand length and NP size). As a

consequence, the charged group of dialkyne in orientation L_A

must approach the Au core closer than in the case of C −Bipy

(Figure 5a); this corresponds to a horizontal shift between red

812

J. Am. Chem. Soc. 2021, 143, 1807−1815

Journal of the American Chemical Society

left red point in Figure 5b) must enter the region of constant

https://pubs.acs.org/doi/10.1021/jacs.0c09408.

Article

(

sites. We see this approach as a form of “enhancing” known

catalysts to endow them with new selectivity modalitiesin

addition to the electrostatic gating described here, we envision,

for instance, that hydrophilic/hydrophobic gating ligands could

diﬀerentiate substrate sites based on the polarity of nearby

groups (see proof-of-the concept experiments in Figure 6g−i),

while chiral ligands could control access of select enantiomers.

An unexplored but promising aspect of this work would be to

also harness the properties of the NP core, e.g., to retrieve from

reaction mixture catalysts tethered onto magnetic particles or

to couple Ir or Ru-based photoredox catalysts to the plasmonic

response of AuNPs.

potential in order to react, and therefore, the potential

diﬀerence Δϕ between L and L is reduced relative to the

C −Bipy case. Because selectivity depends exponentially on

Δϕ, it follows the same trend (Figure 5f). Along similar lines,

for a constant alkane chain length (C ) of Bipy ligand, a change

from C −TMA to a shorter C −TMA shifts the potential

curve toward the Au core (Figure 5a,b vs d,e). The dialkyne

substrate can now reach the Cu site at C −Bipy while being

largely outside the layer of C −TMA quaternary nitrogens

(

green in Figure 5d,e) where the potential is less steep. Thus,

Δϕ is reduced (Figure 5e, purple points) relative to the C −

Bipy/C −TMA case and selectivity drops accordingly (Figure

f, bottom).

ASSOCIATED CONTENT

Supporting Information

The Supporting Information is available free of charge at

■

The eﬀect of NP size also merits an additional comment. On

sı

one hand, when the NPs become larger and their curvature

decreases, the thiol monolayers fan out less and form a more

tightly packed monolayer in which the access to the catalytic

centers is sterically restricted (Figure 6b), possibly compro-

mising the total yield. Indeed, as seen in Figure 6c, the total

yield for 4.2 nm NPs is ∼10% lower than for 2.5 nm NPs (in

these two cases, the NP concentrations were adjusted such that

the concentration of Bipy−Cu(I) was constant, 0.46 mM and

Synthetic and experimental details and NMR data,

computational details, and simulation results (PDF)

AUTHOR INFORMATION

Corresponding Author

■

0% of substrates’ concentration). On the other hand, NP’s

Bartosz A. Grzybowski − Center for Soft and Living Matter,

Institute for Basic Science (IBS), Ulsan, Republic of Korea;

Department of Chemistry, Ulsan National Institute of Science

and Technology (UNIST), Ulsan, Republic of Korea;

Institute of Organic Chemistry, Polish Academy of Sciences,

Warsaw, Poland; orcid.org/0000-0001-6613-4261;

Email: nanogrzybowski@gmail.com

surface electric potential ϕ(r) increases with particle size, in

turn increasing the magnitude of the electric potential energy

diﬀerence (ϕ(r ) − ϕ(r )) between the two orientations of a

substrate and therefore improving the selectivity by ∼exp-

(

−q(ϕ(r ) − ϕ(r ))/kT); this eﬀect is, indeed, seen in the

distribution of products shown in Figure 6c.

Finally, we would like to emphasize that despite the apparent

simplicity of the NPs we used, the dynamics of the monolayer/

catalytic center/substrate system can be nuanced. In particular,

the selectivities observed for some substrates cannot be

predicted “by inspection” and, instead, careful modeling is

needed to understand the experimental outcomes, in some

sense strengthening the analogy to enzymes, for which

understanding substrate binding and reactivity also beneﬁts

from simulations. As a case in point, we synthesized the

substrate marked 4 in Figure 6d; in this molecule, the charged

group is only slightly closer to one of the two alkynes.

Although we expected that this smaller diﬀerence might result

in decreased selectivity between 4A and 4B, we were surprised

to ﬁnd that the main eﬀect was a dramatically increased

formation of the disubstituted product 4C (in a ∼5:2 ratio

with respect to the sum of singly reacted products; see orange

bars in Figure 6d). This eﬀect is clearly due to the NP since, in

Authors

Minju Kim − Center for Soft and Living Matter, Institute for

Basic Science (IBS), Ulsan, Republic of Korea; Department of

Chemistry, Ulsan National Institute of Science and

Technology (UNIST), Ulsan, Republic of Korea

Miroslaw Dygas − Center for Soft and Living Matter, Institute

for Basic Science (IBS), Ulsan, Republic of Korea; Institute of

Organic Chemistry, Polish Academy of Sciences, Warsaw,

Poland

Yaroslav I. Sobolev − Center for Soft and Living Matter,

Institute for Basic Science (IBS), Ulsan, Republic of Korea

Wiktor Beker − Institute of Organic Chemistry, Polish

Academy of Sciences, Warsaw, Poland

Qiang Zhuang − Center for Soft and Living Matter, Institute

for Basic Science (IBS), Ulsan, Republic of Korea; School of

Chemistry and Chemical Engineering, Northwestern

Polytechnical University, Xi’an, China

Tomasz Klucznik − Institute of Organic Chemistry, Polish

Academy of Sciences, Warsaw, Poland

the CuSO /NaAs system, singly reacted products 4A and 4B

are dominant. Also, in the NP system, the 5:2 product ratio is

established already within ∼1 h, suggesting that reactions on

both alkynes happen in rapid succession. These results can be

rationalized by MD simulations illustrated in Figure 6e,f (and

detailed in the SI, Section S10.3), namely, the substrate orients

Guillermo Ahumada − Center for Soft and Living Matter,

Institute for Basic Science (IBS), Ulsan, Republic of Korea;

orcid.org/0000-0002-1507-0816

Juan Carlos Ahumada − Center for Soft and Living Matter,

Institute for Basic Science (IBS), Ulsan, Republic of Korea

such that the charged NMe₃group on the phenyl ring is

pointing away from the NP whereas the two alkynes arrange in

a pincer-like fashion such that they can both enter the on-

particle monolayer and react on the catalytic sites within it.

Complete contact information is available at:

https://pubs.acs.org/10.1021/jacs.0c09408

CONCLUSIONS

■

Author Contributions

In summary, ligands self-assembling on the surface of

nanoparticles can create “active sites” controlling access to

catalytic units and preorganizing the incoming substrates for

selective reactions at only some of otherwise similar-reactivity

M.K. designed and performed most experiments and analyzed

the data. M.D., G.A., and J.C.A. synthesized ligand molecules

for nanoparticles. Y.S. developed theoretical models and

conducted data analysis. W.B. performed molecular dynamics

813

J. Am. Chem. Soc. 2021, 143, 1807−1815

Journal of the American Chemical Society

MD) simulation and related analysis. Q.Z. optimized

nanoparticle functionalization and performed some experi-

ments. M.K., M.D., Q.Z., T.K., G.A., and J.C.A. synthesized

substrates used for catalysis. G.A. and J.C.A. performed

experiments on hydrophobic gating. B.A.G. conceived and

supervised the research. B.A.G. wrote the manuscript with help

from other authors, mostly M.K., Y.S., and W.B. All authors

discussed the results and corrected the manuscript.

(17) Leung, D. H.; Bergman, R. G.; Raymond, K. N. Highly

Article

(

Selective Supramolecular Catalyzed Allylic Alcohol Isomerization. J.

Am. Chem. Soc. 2007, 129, 2746−2747.

18) Yoshizawa, M.; Tamura, M.; Fujita, M. Diels-Alder in Aqueous

Molecular Hosts: Unusual Regioselectivity and Efficient Catalysis.

Science 2006, 312, 251−254.

19) Pluth, M. D.; Bergman, R. G.; Raymond, K. N. Acid Catalysis in

Basic Solution: A Supramolecular Host Promotes Orthoformate

Hydrolysis. Science 2007 , 316 , 85 −88.

Notes

(20) Davis, M. E. New Vistas in Zeolite and Molecular Sieve

Catalysis. Acc. Chem. Res. 1993, 26 , 111 −115.

The authors declare no competing ﬁnancial interest.

21) Wight, A. P.; Davis, M. E. Design and Preparation of Organic-

Inorganic Hybrid Catalysts. Chem. Rev. 2002, 102, 3589 −3614.

ACKNOWLEDGMENTS

■

22) Margelefsky, E. L.; Zeidan, R. K.; Davis, M. E. Cooperative

This work was supported by the Institute for Basic Science,

Republic of Korea (IBS-R020-D1) and by the National Science

Center, Poland (Grant Maestro, No. 2018/30/A/ST5/00529).

Catalysis by Silica-Supported Organic Functional Groups. Chem. Soc.

Rev. 2008, 37, 1118−1126.

(23) Tsuji, K.; Jones, C. W.; Davis, M. E. Organic-Functionalized

Molecular Sieves (OFMSs): I. Synthesis and Characterization of

OFMSs with Polar Functional Groups. Microporous Mesoporous Mater.

1999, 29, 339−349.

(24) Jones, C. W.; Tsuji, K.; Davis, M. E. Organic-Functionalized

Molecular Sieves (OFMSs): II. Synthesis, Characterization and the

Transformation of OFMSs Containing Non-Polar Functional Groups

into Solid Acids. Microporous Mesoporous Mater. 1999, 33 , 223 −240.

(25) Jones, C. W.; Hwang, S.-J.; Okubo, T.; Davis, M. E. Synthesis of

Hydrophobic Molecular Sieves by Hydrothermal Treatment with

Acetic Acid. Chem. Mater. 2001, 13, 1041−1050.

(26) Breslow, R.; Zhang, X.; Huang, Y. Selective Catalytic

Hydroxylation of a Steroid by an Artificial Cytochrome P-450

Enzyme. J. Am. Chem. Soc. 1997 , 119 , 4535 −4536.

(27) Yang, J.; Breslow, R. Selective Hydroxylation of a Steroid at C-9

by an Artificial Cytochrome P-450. Angew. Chem., Int. Ed. 2000, 39,

REFERENCES

■

(

1) Berg, J. M.; Tymoczko, J. L.; Stryer, L. Biochemistry, 6th ed.; W.

H. Freeman and Company: New York, 2006.

2) Davis, H. J.; Phipps, R. J. Harnessing non-Covalent Interactions

to Exert Control over Regioselectivity and Site-Selectivity in Catalytic

Reactions. Chem. Sci. 2017, 8, 864−877.

3) Meeuwissen, J.; Reek, J. H. H. Supramolecular Catalysis beyond

Enzyme Mimics. Nat. Chem. 2010, 2, 615−621.

4) Margelefsky, E. L.; Zeidan, R. K.; Davis, M. E. Heterogeneous

Catalyst Design by Multiple Functional Group Positioning in

Organic-Inorganic Materials: On the Route to Analogs of Multifunc-

tional Enzymes. In Model Systems in Catalysis: Single Crystals to

Supported Enzyme Mimics; Springer: New York, 2010; pp 495 −516.

5) Jones, C. W.; Tsuji, K.; Davis, M. E. Organic-Functionalized

Molecular Sieves as Shape-Selective Catalysts. Nature 1998, 393, 52−

6) Nishioka, Y.; Yamaguchi, T.; Yoshizawa, M.; Fujita, M. Unusual

2 + 4] and [2 + 2] Cycloadditions of Arenes in the Confined Cavity

of Self-Assembled Cages. J. Am. Chem. Soc. 2007, 129, 7000−7001.

7) Bhawe, Y.; Moliner-Marin, M.; Lunn, J. D.; Liu, Y.; Malek, A.;

Davis, M. Effect of Cage Size on the Selective Conversion of

Methanol to Light Olefins. ACS Catal. 2012 , 2 , 2490 −2495.

8) Murase, T.; Horiuchi, S.; Fujita, M. Naphthalene Diels-Alder in a

Self-Assembled Molecular Flask. J. Am. Chem. Soc. 2010, 132, 2866−

867.

9) Mock, W. L.; Irra, T. A.; Wepsiec, J. P.; Manimaran, T. L.

Cycloaddition Induced by Cucurbituril. A Case of Pauling Principle

Catalysis. J. Org. Chem. 1983, 48, 3619−3620.

10) Jon, S. Y.; Ko, Y. H.; Park, S. H.; Kim, H. J.; Kim, K. A Facile,

Stereoselective [2 + 2] Photoreaction Mediated by Cucurbit[8]uril.

Chem. Commun. 2001, 19 , 1938 −1939.

11) Chen, J.; Rebek, J. Selectivity in an Encapsulated Cycloaddition

Reaction. Org. Lett. 2002, 4, 327−329.

12) Kang, J. M.; Santamaria, J.; Hilmersson, G.; Rebek, J. Self-

Assembled Molecular Capsule Catalyzes a Diels-Alder Reaction. J.

Am. Chem. Soc. 1998, 120, 7389−7390.

13) Kuil, M.; Soltner, T.; van Leeuwen, P.; Reek, J. N. H. High-

Precision Catalysts: Regioselective Hydroformylation of Internal

Alkenes by Encapsulated Rhodium Complexes. J. Am. Chem. Soc.

006, 128, 11344−11345.

14) Merlau, M. L.; del Pilar Mejia, M.; Nguyen, S. T.; Hupp, J. T.

Artificial Enzymes Formed Through Directed Assembly of Molecular

Square Encapsulated Epoxidation Catalysts. Angew. Chem., Int. Ed.

001, 40, 4239−4242.

15) Slagt, V. F.; Roder, M.; Kamer, P. C. J.; van Leeuwen, P.; Reek,

28) Fang, Z.; Breslow, R. Metal Coordination-Directed Hydrox-

692−2694.

ylation of Steroids with a Novel Artificial P-450 Catalyst. Org. Lett.

006, 8, 251−254.

29) Dydio, P.; Ploeger, M.; Reek, J. N. H. Selective isomerization-

[

hydroformylation sequence: A Strategy to Valuable α -Methyl-

Branched Aldehydes from Terminal Olefins. ACS Catal. 2013, 3,

939−2942.

30) Dydio, P.; Reek, J. N. H. Supramolecular control of selectivity

in hydroformylation of vinyl arenes: Easy Access to Valuable β -

aldehyde Intermediates. Angew. Chem., Int. Ed. 2013 , 52 , 3878 −3882.

(31) Dydio, P.; Reek, J. N. H. Supramolecular Control of Selectivity

in Transition-Metal Catalysis through Substrate Preorganization.

Chem. Sci. 2014, 5, 2135−2145.

(32) Qing, Y.; Tamagaki-Asahina, H.; Ionescu, S. A.; Liu, M. D.;

Bayley, H. Catalytic Site-Selective Substrate Processing within a

Tubular Nanoreactor. Nat. Nanotechnol. 2019, 14, 1135−1142.

(33) Wu, J.; Wang, X.; Wang, Q.; Lou, Z.; Li, S.; Zhu, Y.; Qin, L.;

Wei, H. Nanomaterials with Enzyme-Like Characteristics (nano-

zymes): Next-Generation Artificial Enzymes (II). Chem. Soc. Rev.

34) Liang, M.; Yan, X. Nanozymes: From New Concepts,

019, 48, 1004−1076.

Mechanisms, and Standards to Applications. Acc. Chem. Res. 2019,

2, 2190−2200.

35) Wang, H.; Wan, K.; Shi, X. Recent Advances in Nanozyme

Research. Adv. Mater. 2019, 31, 1805368.

36) Zhang, Z. J.; Zhang, X. H.; Liu, B. W.; Liu, J. W. Molecular

Imprinting on Inorganic Nanozymes for Hundred-Fold Enzyme

Specificity. J. Am. Chem. Soc. 2017, 139, 5412−5419.

(

(37) Phan, N. T. S.; Van Der Sluys, M.; Jones, C. W. On the Nature

J. N. H. Supraphos: A Supramolecular Strategy to Prepare Bidentate

of the Active Species in Palladium Catalyzed Mizoroki-Heck and

Suzuki-Miyaura Couplings - Homogeneous or Heterogeneous

Catalysis, A Critical Review. Adv. Synth. Catal. 2006, 348, 609−679.

(38) Fillon, Y.; Verma, A.; Ghosh, P.; Ernenwein, D.; Rotello, V. M.;

Chmielewski, J. Peptide Ligation Catalyzed by Functionalized Gold

Nanoparticles. J. Am. Chem. Soc. 2007, 129, 6676−6677.

Ligands. J. Am. Chem. Soc. 2004, 126, 4056−4057.

(16) Dydio, P.; Dzik, W. I.; Lutz, M.; de Bruin, B.; Reek, J. N. H.

Remote Supramolecular Control of Catalyst Selectivity in the

Hydroformylation of Alkenes. Angew. Chem., Int. Ed. 2011, 50,

396−400.

814

J. Am. Chem. Soc. 2021, 143, 1807−1815

Journal of the American Chemical Society

(39) Zaramella, D.; Scrimin, P.; Prins, L. J. Self-Assembly of a

Article

Catalytic Multivalent Peptide-Nanoparticle Complex. J. Am. Chem.

Soc. 2012, 134, 8396−8399.

(40) Jin, L.; Liu, B.; Duay, S. S.; He, J. Engineering Surface Ligands

of Noble Metal Nanocatalysts in Tuning the Product Selectivity.

Catalysts 2017, 7, 44.

(41) Tonga, G. Y.; Jeong, Y.; Duncan, B.; Mizuhara, T.; Mout, R.;

Das, R.; Kim, S. T.; Yeh, Y.-C.; Yan, B.; Hou, S.; Rotello, V. M.

Supramolecular Regulation of Bioorthogonal Catalysis in Cells using

Nanoparticle-Embedded Transition Metal Catalysts. Nat. Chem.

42) Neri, S.; Garcia Martin, S.; Pezzato, C.; Prins, L. J.

015, 7, 597−603.

Photoswitchable Catalysis by a Nanozyme Mediated by a Light-

Sensitive Cofactor. J. Am. Chem. Soc. 2017 , 139 , 1794 −1797.

(43) Prins, L. J. Emergence of Complex Chemistry on an Organic

Monolayer. Acc. Chem. Res. 2015, 48, 1920−1928.

(44) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.;

Whitesides, G. M. Self-Assembled Monolayers of Thiolates on Metals

as a Form of Nanotechnology. Chem. Rev. 2005, 105, 1103−1169.

(46) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Click Chemistry:

45) Yan, Y.; Warren, S. C.; Fuller, P.; Grzybowski, B. A.

Chemoelectronic Circuits Based on Metal Nanoparticles. Nat.

Nanotechnol. 2016, 11, 603−608.

Diverse Chemical Function from a Few Good Reactions. Angew.

Chem., Int. Ed. 2001, 40, 2004−2021.

(47) Lewis, W. G.; Magallon, F. G.; Fokin, V. V.; Finn, M. G.

Discovery and Characterization of Catalysts for Azide-Alkyne

Cycloaddition by Fluorescence Quenching. J. Am. Chem. Soc. 2004,

48) Hein, J. E.; Fokin, V. V. Copper-Catalyzed Azide-Alkyne

26, 9152−9153.

Cycloaddition (CuAAC) and Beyond: New Reactivity of Copper(I)

Acetylides. Chem. Soc. Rev. 2010, 39, 1302−1315.

(49) Borukhov, I.; Andelman, D.; Orland, H. Steric Effects in

Electrolytes: A Modified Poisson-Boltzmann Equation. Phys. Rev. Lett.

997, 79, 435−438.

50) Kilic, M. S.; Bazant, M. Z.; Ajdari, A. Steric Effects in the

Dynamics of Electrolytes at Large Applied Voltages. I. Double-Layer

Charging. Phys. Rev. E 2007, 75, 021502.

(51) Rodionov, V. O.; Presolski, S. I.; Diaz Diaz, D.; Fokin, V. V.;

Finn, M. G. Ligand-Accelerated Cu-Catalyzed Azide-Alkyne Cyclo-

addition: A Mechanistic Report. J. Am. Chem. Soc. 2007, 129, 12705−

2712.

52) Ahlquist, M.; Fokin, V. V. Enhanced Reactivity of Dinuclear

Dopper(I) Acetylides in Dipolar Cycloadditions. Organometallics

007, 26, 4389−4391.

53) Straub, B. F. μ -Acetylide and μ -Alkenylidene Ligands in “click ”

Triazole Syntheses. Chem. Commun. 2007, 37, 3868−3870.

54) Zhuang, Q.; Yang, Z.; Sobolev, Y. I.; Beker, W.; Kong, J.;

Grzybowski, B. A. Control and Switching of Charge-Selective

Catalysis on Nanoparticles by Counterions. ACS Catal. 2018, 8,

55) Roy, S.; Rao, A.; Devatha, G.; Pillai, P. P. Revealing the Role of

469−7474.

Electrostatics in Gold-Nanoparticle-Catalyzed Reduction of Charged

Substrates. ACS Catal. 2017, 7, 7141−7145.

815