A Journal of

Accepted Article

Title: Selection and Screening of DNA Aptamers for Inorganic

Nanomaterials

Authors: Yibo Zhou, Zhicheng Huang, Ronghua Yang, and Juewen Liu

This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.

To be cited as: Chem. Eur. J. 10.1002/chem.201704600

Link to VoR: http://dx.doi.org/10.1002/chem.201704600

Supported by
Chemistry - A European Journal 10.1002/chem.201704600

Selection and Screening of DNA Aptamers for Inorganic Nanomaterials

Yibo Zhou1‖, Zhicheng Huang2‖, Ronghua Yang1 and Juewen Liu1,2*

1. School of Chemistry and Biological Engineering, Changsha University of Science and Technology,

Accepted Manuscript
Changsha 410114, P. R. China

2. Department of Chemistry, Waterloo Institute for Nanotechnology, University of Waterloo, Waterloo,

Ontario, N2L 3G1, Canada

Corresponding Author Email: liujw@uwaterloo.ca

‖
Y. Zhou and Z. Huang contributed equally to this manuscript.

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

Abstract

Searching for DNA sequences that can strongly and selectively bind to inorganic surfaces is a long-
standing topic in bionanotechnology, analytical chemistry and biointerface research. This can be
achieved either by aptamer selection starting with a very large library of ~1014 random DNA sequences,
or by careful screening of a much smaller library (usually from a few to a few hundred) with rationally
designed sequences. Unlike typical molecular targets, inorganic surfaces often have quite strong DNA

Accepted Manuscript
adsorption affinities due to polyvalent binding and even chemical interactions. This leads to a very high
background binding making aptamer selection difficult. Screening, on the other hand, can be designed
to compare relative binding affinities of different DNA sequences and could be more appropriate for
inorganic surfaces. The resulting sequences have been used for DNA-directed assembly, sorting of
carbon nanotubes, and DNA-controlled growth of inorganic nanomaterials. It was recently discovered
that poly-cytosine (C) DNA can strongly bind to a diverse range of nanomaterials including nanocarbons
(graphene oxide and carbon nanotubes), various metal oxides and transition metal dichalcogenides. In
this Concept article, we articulate the need for screening and potential artifacts associated with
traditional aptamer selection methods for inorganic surfaces. Representative examples of application
are discussed, and a few future research opportunities are proposed towards the end of this article.

Keywords: Aptamers; Adsorption; Nanomaterials; DNA; Graphene

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

Introduction

Attaching biopolymers such as DNA and proteins to surfaces is highly useful for biosensor development,
directed materials assembly, and drug and DNA delivery.[1-8] To achieve optimal performance,
biopolymers need to be stably attached at a desired density, while still retaining their biological
activities and functions. Although chemically synthesized DNA can be modified with a wide range of
bioconjugation groups, covalent attachment remains difficult for many surfaces such as metal oxides

Accepted Manuscript
and dichalcogenides. Even for graphene oxide (GO) with abundant carboxyl groups (e.g. can be linked to
amino-modified DNA),[9-10] physisorption is still much more popular for its simplicity.[11-13] Between the
highly controllable covalent conjugation and random physisorption, an alternative approach is to design
a diblock DNA with a high surface affinity sequence serving as an anchoring block, and the other block
being the functional sequence oriented away from the surface.[14-18] This is different from random
physisorption, since the functional part of the molecule does not interact with the surface, and thus the
biological activity of the molecule is better retained.

Then, a key question is how to obtain surface binding sequences? The assumption is that this
sequence can displace other sequences on surface to achieve the intended directional adsorption.
Phage display has been used for isolating such peptides.[19-20] For DNA, this often means aptamer
selection. Aptamers refer to single-stranded nucleic acids that can selectively bind to target
molecules.[21-22] While most of work aimed at molecular targets, aptamers against surfaces have also
been reported in a few cases such as cellulose,[23-24] ZnO,[25] and calcium phosphate.[26] Compared to
hundreds of aptamers for various small molecules and proteins, the number of surface binding aptamers
is still quite low.

Beyond DNA immobilization, surface binding aptamers are also potentially useful for other
applications such as dispersing nanoparticles,[27] stabilization and sorting of nanomaterials,[28] DNA-
directed mineralization,[25, 29] and for understanding fundamental interactions between surfaces and
DNA. However, such aptamers have not caught much attention in application so far. The problem seems
to be the lack of high quality aptamers from the typical aptamer selection process.

Different from small molecules, inorganic surfaces have unique properties that make their
interaction with DNA a very interesting topic.[30-32] Fundamental studies on the surface features of
inorganic materials have led to other ways of obtaining affinity sequences. For example, in addition to
selection, screening has also been performed for surface binding DNA sequences. Although the library

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

size and throughput are much lower, screening has its own merit in this particular case. The best known
example was probably the DNA sequences binding to carbon nanotubes obtained by Zheng and co-
workers.[27-28, 33] In this Concept paper, we articulate the unique properties of inorganic surfaces in
comparison to typical small molecules, and its implication for finding related DNA aptamers. We
emphasize on a recent discovery that poly-cytosine (C) DNA can be used as a general high affinity ligand
for many surfaces.[17]

Accepted Manuscript
Inorganic surfaces versus molecules as an aptamer target.

Molecular targets have well-defined structures, and each molecule is exactly the same. Depending on
the shape and chemical groups, a molecule may interact with DNA via hydrogen bonding, electrostatic
interactions, hydrophobic interactions, and - stacking.[34-35] Usually, larger molecules have more
potential points of interaction and its aptamers can often bind with a higher affinity (e.g. typically µM
affinity with small molecules, while nM or pM affinity with proteins). Inorganic surfaces, however, are
very different at least in the following aspects, which strongly affect their interaction with DNA. We
focus our discussion on nanoparticles dispersed in water instead of bulk planar surfaces, since
nanoparticles are more popular for bio-related applications, and the chemistry of these two types of
surfaces is often quite similar.

1) Polyvalent binding. The surface area of most nanoparticles (e.g. often larger than 5 nm in
diameter) is much larger than small molecules and even proteins. The length of a 10-mer duplex DNA is
only 3.4 nm with a diameter of 2 nm. Therefore, even a very small nanoparticle can still interact with
many nucleotides in a DNA chain. The implication is that even relatively weak intermolecular forces that
may not be important for binding small molecules can be amplified via the polyvalent effect.[36] Stronger
interactions can also be promoted by simply increasing the length of DNA. A typical library for aptamer
selection is shown in Figure 1A, with two primer binding regions flanking the random library. With such
a long DNA library (often more than 80 nucleotides), non-specific interactions are likely to be very strong
for surface binding, which can make aptamer selection difficult. In other words, strong non-specific
binding can produce a very high background noise masking aptamer sequences.

2). Specific versus non-specific binding. Biological interactions are always quite specific, where
changing even a single atom can sometimes strongly disrupt binding. For most DNA aptamers, specific
binding refers to the folding of aptamers in 3D structures forming a binding pocket, implying also a very
strong DNA sequence requirement. Figure 1B shows the secondary structure of a DNA aptamer for

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

adenosine or ATP that can simultaneously bind two target molecules.[37] With a single nucleotide change
at an important location, binding can be fully inhibited. Another often cited example is the theophylline
aptamer that cannot bind caffeine (over 10,000-fold decreased affinity),[38] although caffeine and
theophylline differ only by a methyl group.

Such exquisite specificity is rarely seen for aptamers against inorganic surfaces. Most of the
nucleotides can contribute to surface binding, but it is difficult to identify a critical one. By changing any

Accepted Manuscript
single nucleotide, the change in affinity is often quite small. For example, DNA bind to gold
nanoparticles (AuNPs) via its bases, and it is known that poly-adenine (A) DNA binds to gold surface
more strongly than poly-thymine (T) DNA.[39-40] Figure 1C shows a poly-A sequence, and by replacing one
of the adenines by a thymine, the effect is unlikely to be noticeable. Only when all the adenines are
replaced by thymines, some difference can be detected.[39, 41-42]
DNA can also use its phosphate
backbone to bind and representative examples are metal oxides.[43-47] In such cases, the effect of DNA
base sequence is even smaller. Overall, with strong polyvalent interactions between DNA and surfaces,
specific sequences might be difficult to obtain or even to define.

When it comes down to binding energy, typical aptamer binding is not too far from KBT, the
thermal energy, such that small changes in temperature, pH or ionic strength can have a large influence
on aptamer binding. On the other hand, surface binding can be much stronger. For example, when a
piece of DNA is adsorbed on a gold surface, it is quite difficult to desorb it under ambient conditions by
changing buffer conditions, and individual DNA base adsorption on gold can reach over 100 kJ/mol.[31, 39-
40, 48-49]

Figure 1. The concept of specific versus non-specific binding. (A) A typical DNA library design for aptamer
selection with around 80 nucleotides. Only the middle 40 are the library while the rest are for PCR

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

purposes. Such a long library can have strong non-specific interactions with many inorganic surfaces. (B)
In the adenosine binding aptamer, changing a critical base can fully inhibit binding. (C) For a poly-A DNA
binding to AuNPs, changing an A to a T has almost no effect.

Classification of surfaces for DNA binding

Accepted Manuscript
Evolved as a genetic material, DNA has been demonstrated to be remarkable not only in binding various
molecules but also to inorganic surfaces. The structure of a 4-mer DNA is shown in Figure 2A. It has a
negatively charged phosphate backbone that can interact with positively charged surfaces or surfaces
with borderline and hard metals such as Ca(II), Ti(IV), and lanthanides. The DNA bases can use -
stacking and hydrogen bonding. The overall DNA chain as a polymer can interact with surfaces via van
der Waals force. Based on chemical composition and surface properties, we classify inorganic surfaces
as follows to interact with DNA.

1) Classification of surfaces based on chemical composition. a) Metal nanoparticles (common

ones include gold, silver, platinum and palladium). They resist oxidation and thus can exist as metallic
nanoparticles in ambient conditions. Among them, gold nanoparticles have been the most extensively
studied.[50-52] DNA binds to gold surfaces via strong base coordination. While the affinity for different
bases are quite different, each base can be adsorbed very strongly approaching chemical adsorption
region in vacuum.[30-31, 39, 53] b) Metal oxide surfaces bind to DNA via its phosphate backbone, and the
effect of DNA base is relatively small.[43-47] Free inorganic phosphate ions can displace adsorbed DNA
from metal oxides in most cases. c) Carbon-based surfaces such as graphene oxide and carbon
nanotubes can bind DNA via - stacking and hydrogen bonding.[12, 54]
d) Transition metal
dichalcogenides interact with DNA via van der Waals force.[55] Therefore, each type of surface has a
different interaction mechanism with DNA, and DNA is extremely versatile for binding to surfaces. Note
we only discussed non-specific binding based solely on the chemical property of nucleotides and DNA,
without considering the sequence and structure of DNA.

Some other nanoparticles, such as semiconductor quantum dots, need to be capped with strong
hydrophobic ligands during synthesis and the surface chemistry is strongly dependent on the ligand shell.
Although water solution quantum dots have also been prepared, their interaction with DNA has not
been systematically studied and they are excluded from our discussion here.

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

2) Classification of surfaces based on crystallinity. Another type of classification is based on the

crystallinity of surfaces. Crystalline surfaces have simple repeating structures (Figure 2A). The dimension
of crystalline lattices is often comparable with that of DNA (e.g. in a duplex DNA the distance between
two neighboring bases is just 0.34 nm). Amorphous surfaces are also quite common for metal oxides,
hydroxides and phosphates, especially when they are freshly prepared by simple precipitation in water
(Figure 2B). Most common materials have a relatively simple composition. Even without a periodic
structure, they may still have simple interactions with DNA.

Accepted Manuscript
Graphene oxide is a unique example with heterogeneous surface structures (Figure 2C). The
domain size is comparable with the size of DNA oligonucleotides.[56-57] Depending on the spot of DNA
binding, it is possible to have a continuous distribution of interaction forces. For all the surfaces, we
need to realize that DNA is not adsorbed in a single conformation, and the final conformation of DNA
might be influenced by the kinetics of adsorption.

Figure 2. (A) Scheme of a crystalline surface with periodic structures together with a scanning tunneling
microscope (STM) micrograph of Au(111) surface. Adapted from Ref.[58] with permission. Copyright
Royal Society of Chemistry 2012. (B) Scheme of an amorphous inorganic surfaces together with a STM
micrograph of a silicon surface. Adapted from Ref.[59] with permission. Copyright Elsevier 2004. (C)

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

Scheme of a surface with nanoscale domains and the size of the domains are comparable with the size
of biopolymers together with a transmission electron microscopy (TEM) micrograph of graphene oxide.
Adapted from Ref.[60] with permission. An example of (C) is graphene oxide with both crystalline carbon
domains and highly oxidized amorphous domains in purple. (D) The structure of a 4-mer DNA with the
sequence of ACGT.

Accepted Manuscript
Surface binding aptamers from selection. With the above understanding of the properties of surfaces
and their non-specific interactions with DNA, we now discuss aptamer selection. Most aptamers were
obtained by a combinatorial biology method called systematic evolution of ligands by exponential
enrichment (SELEX), where a random DNA library is mixed with the target molecule. The binding DNA
sequences are amplified by polymerase chain reaction (PCR) to feed the next round of selection under
more stringent conditions. High quality and specific aptamers are enriched usually after more than ten
selection cycles.

For inorganic surfaces, the SELEX process is quite simple in most cases since the target is already
a surface, and there is no need for further immobilization (typically required for small molecule targets).
The large size of inorganic surfaces allows simple centrifugation to wash away non-binding or weakly
bound sequences, and the remaining sequences are eluted under denaturing conditions and then
amplified by PCR. When the binding activity is saturated, the resulting library is analyzed by sequencing.
Using this method, a number of aptamers have been selected against different surfaces. Gerdon and
coworkers used a precipitation SELEX method to isolate DNA aptamers for calcium phosphate.[26] A
single round of aptamer selection was performed by Morse and coworkers on ZnO, and a simple T30 DNA
sequence was concluded to be its aptamer.[25] By a screening method, however, poly-C DNA was found
to adsorb more strongly than poly-T DNA.[17, 61] Breaker and coworkers selected both DNA and RNA
aptamers for cellulose.[23-24]

However, these aptamers have not caught much attention for practical applications. We
attribute this to their poor binding properties. For example, the difference between such aptamers and
random sequences could be very small. Going through an aptamer selection protocol does not
necessarily yield desired aptamers. If non-specific surface interactions are too strong, it is unlikely to get
the desired sequence using the typical aptamer selection method. Therefore, the selection results need
to be looked more critically for surface based targets.

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

Surface binding DNAs from screening. With strong non-specific binding, aptamer selection is quite
difficult to perform on many surfaces. Only those surfaces with sufficiently weak binding, and those
interacting mainly with DNA bases (instead of phosphate backbone), can aptamer selection be
potentially useful. An alternative method is screening. The goal of screening is not just for binding (any
sequence can bind), but for stronger binders or for other properties.

Aptamer selection starts with a very large DNA library of around 1014 random sequences. The

Accepted Manuscript
main property searched by the selection is binding affinity. Screening involves a much smaller library (a
few to a few hundred), and each sequence is assayed individually. Since the surface of crystalline
materials often matches the spacing between nucleotides, optimal sequences might be simple repeating
nucleotides. 2D surfaces may not need complicated 3D DNA folding either, and simple repeating
sequences could be sufficient. Therefore, the limited library size used for screening might not be a major
disadvantage when the library is properly designed. Second, selection only selects for binding. For strong
binding surfaces, it is difficult to select for the strongest one (often kinetically determined). Screening
allows for a closer look at each sequence and other assay methods can be developed, for example,
displacement assays, the ability to disperse nanoparticles, and the sorting of a particular type of carbon
nanotubes. These are difficult to be performed by selection. Finally, no primer binding regions are
needed are screening (see Figure 1A for typical selection library design), and the whole sequence can all
be used for binding with little potential artifacts. Very short sequences (e.g. 5-mer DNA) difficult for
typical aptamer selections can be easily used for screening.

There are also a few disadvantages for screening. Since each sequence is individually made, it is
more costly. If each sequence is assayed in serial, the time to finish the whole screen can also be very
long. With high throughput screening, the time issue might be less important. With a very limited
sequence pool, the real optimal sequence might be hard to find, and it is difficult to be certain if the
screened sequence is the best possible candidate or not.

Examples of screened surface binding DNA sequences.

1) Screening by displacement. For strong binding surfaces, measuring direct binding affinity
could be difficult. Instead, competitive assay can be used. For example, with very strong binding,
essentially any sequence can bind to gold quite stably. Therefore, the goal is to obtain the most stable
binding sequence. Tarlov and coworkers screened the four 5-mer homo-DNA sequences for binding to
gold surface.[39] All the four DNA can adsorb, but they have identified that the poly-adenine sequence to

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

be the strongest. This knowledge has been used to functionalize gold surfaces with controllable DNA
densities.[14-15] The Liu lab also did small scale screening experiments based on the displacement of
fluorescently labeled DNA oligonucleotides on many inorganic surfaces.[17]

2) Screening of other properties. Zheng et al performed a series of experiments to screen for

DNA sequences that interact with carbon nanotubes (CNTs). First, they looked at the dispersion and
separation property. Among the 60-mer DNA homopolymers, they identified T60 being optimal for

Accepted Manuscript
dispersion CNTs, and later they varied the DNA length and found that T30 was optimal.[27] Then they
performed a systematic screening to obtain DNA sequences that can sort CNTs using not only
homopolymers but also sequences containing two different nucleotides. They found that poly(GT)
sequences are the most optimal.[28] A few years later, a larger scale screening was performed with a final
of ~350 DNAs screened, and 20 new sequences were obtained to effectively separate CNTs with
different chirality.[33]

Unique binding properties of Poly-C DNA. Liu and coworkers performed a simple screen of eight
different surfaces, and discovered that poly-C DNA is a general aptamer for all the inorganic surfaces.[17,
62]
It has very strong anchoring power on GO, MoS2, WS2 and various metal oxides. Figure 3 shows a
representative data for screening on GO using fluorescently labeled oligonucleotides. A FAM-labeled
DNA was first adsorbed, resulting in quenched fluorescence. Competing DNA sequences were then
individually added, and the best competitor resulted in the most fluorescence enhancement. Compared
to other DNA homopolymers, poly-C DNA showed most effectively displacement, while the adsorbed
poly-C was barely displaced by other sequences. Therefore, poly-C DNA provides a solution for
anchoring DNA on these surfaces. In addition, poly-C DNA can serve as a benchmark for aptamer
selection. To qualify as a high quality aptamer, it needs to be better than poly-C DNA of the same length.

It was found that for surfaces adsorbing DNA more weakly, the relatively adsorption affinity of
poly-C DNA compared to other DNA is larger. As discussed above, each type of material binds DNA with
a different mechanism. Yet poly-C DNA has the highest affinity in each case. This suggests the unique
property about poly-C DNA itself. Detailed mechanistic work is currently underway in the lab.

10

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

Accepted Manuscript
Figure 3. Kinetics of desorption of FAM-labeled (A) A15, (B) T15, (C) C15, and (D) G15 from GO by the four
15-mer non-labeled homo-DNA (5 µM) added at 10 min. Figure adapted from ref.[17] with permission.
Copyright ®Wiley 2017.

Application of surface binding DNA sequences. With such affinity DNA sequences, many interesting
applications are possible, and a few examples are presented here.

1) Surface hybridization. Directional immobilization of unmodified DNA for hybridization is of

significant practical interest for its cost-effectiveness. This hybrid material can be used for hybridizing
with its cDNA for analytical applications. Current DNA synthesis can be afforded by most researchers
while after adding thiol or other modifications, the synthetic cost can go up many folds. Excellent
example of poly-A containing DNA for anchoring on gold surfaces have been reported.[14-15] With the
discovery of poly-C DNA being a high affinity sequence for many other surfaces, the application of this
method can be potentially expanded. On GO, it was demonstrated that poly-C containing DNA can still
hybridize specifically even in the presence of high concentrations of competing ligands (Figure 4A). The
performance of such sequences was much better than the typically used poly-A anchoring sequences.[17]
It needs to be noted that such diblock DNA is likely to be a research tool since the current examples still
cannot compete with covalent conjugates for long-term stability. With their low cost, such DNA can be
used for proof-of-concept and optimization work. Once optimized, then covalent linkages can be made
for practical applications.

11

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

2) DNA-directed assembly. DNA has been used as designer bonds for building nanostructures.
This has been performed using AuNPs functionalized with thiolated DNA. With non-thiolated diblock
DNA containing a poly-A block, it is also possible to perform similar assembly of materials.[14, 63] This can
be cost-effective and can easily tune the density of DNA on AuNP surface (Figure 4B).[14] For some other
surfaces, such as GO,[9-10] MoS2,[64] and most metal oxides, covalent attachment takes even more steps,
and such DNA anchors could be more preferred by some researchers. Li and coworkers used a poly-A
anchoring sequence to assemble GO with DNA for ultrasensitive DNA detection.[16] The disadvantage is

Accepted Manuscript
still the relatively low binding affinity compared to covalently attached DNA, while the advantages are
adjustable affinity and density.

3) DNA-templated materials synthesis. With a high surface binding affinity, such DNA can also be
useful for controlling the growth of nanoparticles. For example, attempts have been made with
mineralization using ZnO.[25] The much more systematically studied examples are the growth of gold
nanostructures (Figure 4C).[65] In addition, gold nanorods,[66] and palladium nanoparticles[67] were also
used as seeds to grow various nanostructures and mechanistic studies have also been performed.[29]

4) Sorting of carbon nanotubes. The work pioneered by Zheng and coworkers for screening DNA
sequences for screening carbon nanotubes has been an interesting example, where very subtle
differences in chirality of the nanotubes can be obtained by recognized by DNA. The implication is that
DNA can ‘sense’ the periodic carbon structures with a very high resolution. It has suggested the
significant potential for DNA in interacting with surfaces.

12

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

Accepted Manuscript
Figure 4. Examples of applications of surface binding DNA sequences. All these sequences were from
screening or rational design. (A) DNA surface hybridization on GO using a diblock DNA design, where the
poly-C anchor works better.[17] Copyright ®Wiley 2017. (B) Controlling the density of DNA on AuNPs by
varying the length of the anchoring poly-A block.[14] Copyright ®ACS 2012. (C) DNA templated seeded
growth of AuNPs or other shapes.[65] Copyright ®Wiley 2012. (D) Sorting carbon nanotubes with various
DNA sequences. Figure adapted from ref.[33] with permission. Copyright ®Nature Publication Groups.

Summary and future perspectives

In summary, we articulated the difference between molecular targets and inorganic surfaces (e.g.
nanoparticles) for DNA aptamer selection and also for DNA binding in general. Inorganic surfaces often
have much stronger non-specific interactions, making typical aptamer selection difficult to perform. To
better illustrate this concept, we classified inorganic surfaces based on their chemical composition and
crystallinity. The concept of screening then arises for inorganic surfaces. In particular, we highlighted a
recent discovery of poly-C DNA being effective anchoring sequences for a diverse range of surfaces. On a
larger scale screening involving a few hundred sequences, Zheng and co-workers focused on carbon
nanotube sorting.[33] It needs to be noted that not all surfaces can strongly adsorb DNA and weakly

13

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

interacting ones still exist, for which it might be still possible for aptamer selection to gain subtle
differences in binding energy, and aptamers for cellulose are good examples.[23-24] Given the merits of
screening surface binding DNA sequences, we propose the following future research opportunities.

1) Tool development. A key challenge is to increase the throughput of screening. With 1

nucleotide, there are only 4 sequences. With 2 different nucleotides, there are additional six (if we
ignore the difference between TA and AT). With 3, the number of combinations goes up to over 30. So

Accepted Manuscript
far, only the carbon nanotube work has been screened at a relatively large scale of over 100 DNA
sequences. Most other work used only some DNA homopolymers sequences or simple nucleotide
repeats. Luckily, with the assumption that inorganic surfaces have short-ranged ordering, we expect the
optimal binding sequences to be simple DNA repeats. Developing more efficient DNA screening tool will
be important. DNA microarray could be a good idea since the technology of synthesis and screening
have already been developed for this purpose. This idea has been used for screening of DNA for making
fluorescent silver nanoclusters.[68] Other screening methods such as those based on microdroplets are
also highly desirable. They have now been used for aptamer screening for molecular targets,[69-72] and
their application for surfaces has yet to be demonstrated. Most previous work focused on gold and
carbon nanotubes for screening, and even for these two, the number of tested DNA sequences is still
very limited. For most other surfaces, they were screened only for simple homo DNA polymers. With the
development of better synthetic and analytical tools, this is likely to be a natural outcome.

2) Mechanistic studies. Fundamental surface force studies are important but difficult to measure.
For example, it remains unclear why poly-C DNA is so unique that it can tightly bind to all types of
surfaces, although the mechanism of binding is different in each surface. Model surfaces with well-
defined chemistry might be needed to provide deeper structural and spectroscopic data on this front.
Similar work has been performed to select surface-binding peptides, but peptides have 20 different side
chains with very different properties. This makes DNA particularly useful for systematic studies and
fundamental understanding.

3) Using modified DNA. Most of the current work used only natural DNA bases and backbones.
We reason that interesting DNA binders might also be produced by using modified DNA, especially those
modifications can potentially improve the binding of specific surfaces. For example, phosphorothioate
modification might be useful for surfaces with thiophilic metals.[73-75]

14

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

4) Further demonstration of application. The eventual value of such DNA has to be

demonstrated through application. Most of the current work has focused on proof-of-concept studies
and the stability of physisorbed DNA is still weaker than those with covalent conjugation. Therefore, the
applications need to be less relied on adsorption stability but more geared towards specific DNA/surface
interactions.

Accepted Manuscript
Acknowledgement

The data from the Liu lab was mainly from work supported by the Natural Sciences and Engineering
Research Council of Canada (NSERC).

References

[1] M. R. Jones, N. C. Seeman, C. A. Mirkin, Science 2015, 347, 1260901.

[2] K. E. Sapsford, W. R. Algar, L. Berti, K. B. Gemmill, B. J. Casey, E. Oh, M. H. Stewart, I. L. Medintz,

Chem. Rev. 2013, 113, 1904-2074.

[3] L. Yang, X. B. Zhang, M. Ye, J. H. Jiang, R. H. Yang, T. Fu, Y. Chen, K. M. Wang, C. Liu, W. H. Tan,

Adv. Drug Deliver. Rev. 2011, 63, 1361-1370.

[4] J. Liu, Z. Cao, Y. Lu, Chem. Rev. 2009, 109, 1948–1998.

[5] S. P. Song, Y. Qin, Y. He, Q. Huang, C. H. Fan, H. Y. Chen, Chem. Soc. Rev. 2010, 39, 4234-4243.

[6] O. I. Wilner, I. Willner, Chem. Rev. 2012, 112, 2528-2556.

[7] H. Wang, R. H. Yang, L. Yang, W. H. Tan, ACS Nano 2009, 3, 2451-2460.

[8] H. Sun, J. Ren, X. Qu, Acc. Chem. Res. 2016, 49, 461-470.

[9] N. Mohanty, V. Berry, Nano Lett. 2008, 8, 4469-4476.

[10] P.-J. J. Huang, J. Liu, Anal. Chem. 2012, 84, 4192-4198.

[11] C. H. Lu, H. H. Yang, C. L. Zhu, X. Chen, G. N. Chen, Angew. Chem. Int. Ed. 2009, 48, 4785-4787.

[12] B. Liu, S. Salgado, V. Maheshwari, J. Liu, Curr. Opin. Colloid Interface Sci. 2016, 26, 41-49.

15

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

[13] F. Li, H. Pei, L. Wang, J. Lu, J. Gao, B. Jiang, X. Zhao, C. Fan, Adv. Funct. Mater. 2013,

10.1002/adfm.201203816.

[14] H. Pei, F. Li, Y. Wan, M. Wei, H. Liu, Y. Su, N. Chen, Q. Huang, C. Fan, J. Am. Chem. Soc. 2012, 134,

11876-11879.

[15] A. Opdahl, D. Y. Petrovykh, H. Kimura-Suda, M. J. Tarlov, L. J. Whitman, Proc. Natl. Acad. Sci.

Accepted Manuscript
U.S.A. 2007, 104, 9-14.

[16] L. H. Tang, Y. Wang, Y. Liu, J. H. Li, ACS Nano 2011, 5, 3817-3822.

[17] C. Lu, Z. Huang, B. Liu, Y. Liu, Y. Ying, J. Liu, Angew. Chem., Int. Ed. 2017, 56, 6208-6212.

[18] S.-W. Lee, C. Mao, C. E. Flynn, A. M. Belcher, Science 2002, 296, 892-895.

[19] R. R. Naik, S. J. Stringer, G. Agarwal, S. E. Jones, M. O. Stone, Nat. Mater. 2002, 1, 169-172.

[20] S. R. Whaley, D. S. English, E. L. Hu, P. F. Barbara, A. M. Belcher, Nature 2000, 405, 665-668.

[21] A. D. Ellington, J. W. Szostak, Nature 1990, 346, 818-822.

[22] C. Tuerk, L. Gold, Science 1990, 249, 505-510.

[23] B. J. Boese, K. Corbino, R. R. Breaker, Nucleos. Nucleot. Nucl. 2008, 27, 949-966.

[24] B. J. Boese, R. R. Breaker, Nucleic Acids Res. 2007, 35, 6378-6388.

[25] L. A. Bawazer, A. M. Newman, Q. Gu, A. Ibish, M. Arcila, J. B. Cooper, F. C. Meldrum, D. E. Morse,

ACS Nano 2014, 8, 387-395.

[26] K. R. Baillargeon, K. Meserve, S. Faulkner, S. Watson, H. Butts, P. Deighan, A. E. Gerdon, Chem.

Commun. 2016.

[27] M. Zheng, A. Jagota, E. D. Semke, B. A. Diner, R. S. McLean, S. R. Lustig, R. E. Richardson, N. G.

Tassi, Nat. Mater. 2003, 2, 338-342.

[28] M. Zheng, A. Jagota, M. S. Strano, A. P. Santos, P. Barone, S. G. Chou, B. A. Diner, M. S.

Dresselhaus, R. S. McLean, G. B. Onoa, G. G. Samsonidze, E. D. Semke, M. Usrey, D. J. Walls,

Science 2003, 302, 1545-1548.

16

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

[29] L. H. Tan, Y. Yue, N. S. R. Satyavolu, A. S. Ali, Z. Wang, Y. Wu, Y. Lu, J. Am. Chem. Soc. 2015, 137,

14456-14464.

[30] J. Liu, Phys. Chem. Chem. Phys. 2012, 14, 10485-10496.

[31] T. M. Herne, M. J. Tarlov, J. Am. Chem. Soc. 1997, 119, 8916-8920.

[32] J. M. Carnerero, A. Jimenez-Ruiz, P. M. Castillo, R. Prado-Gotor, ChemPhysChem 2017, 18, 17-33.

Accepted Manuscript
[33] X. Tu, S. Manohar, A. Jagota, M. Zheng, Nature 2009, 460, 250-253.

[34] T. Hermann, D. J. Patel, Science 2000, 287, 820-825.

[35] W. Zhou, R. Saran, J. Liu, Chem. Rev. 2017, 117, 8272–8325.

[36] M. Mammen, S. K. Choi, G. M. Whitesides, Angew. Chem. Int. Ed. 1998, 37, 2755-2794.

[37] D. E. Huizenga, J. W. Szostak, Biochemistry 1995, 34, 656-665.

[38] R. D. Jenison, S. C. Gill, A. Pardi, B. Polisky, Science 1994, 263, 1425-1429.

[39] H. Kimura-Suda, D. Y. Petrovykh, M. J. Tarlov, L. J. Whitman, J. Am. Chem. Soc. 2003, 125, 9014-

9015.

[40] B. Liu, J. Liu, Anal. Methods 2017, 9, 2633-2643.

[41] J. J. Storhoff, R. Elghanian, C. A. Mirkin, R. L. Letsinger, Langmuir 2002, 18, 6666-6670.

[42] M. Ostblom, B. Liedberg, L. M. Demers, C. A. Mirkin, J. Phys. Chem. B 2005, 109, 15150-15160.

[43] F. Wang, X. Zhang, Y. Liu, Z. Y. Lin, B. Liu, J. Liu, Angew. Chem., Int. Ed. 2016, 55, 12063–12067.

[44] B. Liu, Z. Sun, P.-J. J. Huang, J. Liu, J. Am. Chem. Soc. 2015, 137, 1290-1295.

[45] B. Liu, J. Liu, ACS Appl. Mater. Inter. 2015, 7, 24833-24838.

[46] X. Zhang, F. Wang, B. Liu, E. Y. Kelly, M. R. Servos, J. Liu, Langmuir 2014, 30, 839–845.

[47] B. Liu, J. Liu, Chem. Commun. 2014, 50, 8568-8570.

[48] W. Zhao, M. A. Brook, Y. Li, ChemBioChem 2008, 9, 2363-2371.

[49] K. M. Koo, A. A. I. Sina, L. G. Carrascosa, M. J. A. Shiddiky, M. Trau, Anal. Methods 2015, 7, 7042-

7054.

17

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

[50] K. Saha, S. S. Agasti, C. Kim, X. Li, V. M. Rotello, Chem. Rev. 2012, 112, 2739−2779.

[51] J. I. Cutler, E. Auyeung, C. A. Mirkin, J. Am. Chem. Soc. 2012, 134, 1376-1391.

[52] X. Zhang, M. R. Servos, J. Liu, Langmuir 2012, 28, 3896-3902.

[53] D. Y. Petrovykh, H. Kimura-Suda, L. J. Whitman, M. J. Tarlov, J. Am. Chem. Soc. 2003, 125, 5219-

5226.

Accepted Manuscript
[54] J. Lee, Y. Yim, S. Kim, M.-H. Choi, B.-S. Choi, Y. Lee, D.-H. Min, Carbon 2016, 97, 92-98.

[55] C. Lu, Y. Liu, Y. Ying, J. Liu, Langmuir 2017, 33, 630–637.

[56] C. Gomez-Navarro, J. C. Meyer, R. S. Sundaram, A. Chuvilin, S. Kurasch, M. Burghard, K. Kern, U.

Kaiser, Nano Lett. 2010, 10, 1144-1148.

[57] M. Wu, R. Kempaiah, P.-J. J. Huang, V. Maheshwari, J. Liu, Langmuir 2011, 27, 2731–2738.

[58] S. Yoshimoto, Y.-G. Kim, K. Sato, J. Inukai, K. Itaya, Phys. Chem. Chem. Phys. 2012, 14, 2286-2291.

[59] K. Arima, H. Kakiuchi, M. Ikeda, K. Endo, M. Morita, Y. Mori, Surf. Sci. 2004, 572, 449-458.

[60] K. Erickson, R. Erni, Z. Lee, N. Alem, W. Gannett, A. Zettl, Adv. Mater. 2010, 22, 4467-4472.

[61] L. Ma, B. Liu, J. Huang Po-Jung, X. Zhang, J. Liu, Langmuir 2016, 32, 5672–5680.

[62] Z. Huang, J. Liu, Langmuir 2017, 10.1021/acs.langmuir.7b02812.

[63] X. Zhang, B. Liu, N. Dave, M. R. Servos, J. Liu, Langmuir 2012, 28, 17053-17060.

[64] B. L. Li, M. I. Setyawati, L. Chen, J. Xie, K. Ariga, C.-T. Lim, S. Garaj, D. T. Leong, ACS Appl. Mater.

Inter. 2017, 9, 15286-15296.

[65] Z. Wang, L. Tang, L. H. Tan, J. Li, Y. Lu, Angew. Chem., Int. Ed. 2012, 51, 9078-9082.

[66] T. J. Song, L. H. Tang, L. H. Tan, X. J. Wang, N. S. R. Satyavolu, H. Xing, Z. D. Wang, J. H. Li, H. J.

Liang, Y. Lu, Angew. Chem. Int. Ed. 2015, 54, 8114-8118.

[67] N. S. R. Satyavolu, L. H. Tan, Y. Lu, J. Am. Chem. Soc. 2016, 138, 16542-16548.

[68] C. I. Richards, S. Choi, J.-C. Hsiang, Y. Antoku, T. Vosch, A. Bongiorno, Y.-L. Tzeng, R. M. Dickson, J.

Am. Chem. Soc. 2008, 130, 5038-5039.

18

This article is protected by copyright. All rights reserved.

Chemistry - A European Journal 10.1002/chem.201704600

[69] H. Qu, A. T. Csordas, J. Wang, S. S. Oh, M. S. Eisenstein, H. T. Soh, ACS Nano 2016, 10, 7558-7565.

[70] S. S. Oh, K. M. Ahmad, M. Cho, S. Kim, Y. Xiao, H. T. Soh, Anal. Chem. 2011, 83, 6883-6889.

[71] A. Autour, E. Westhof, M. Ryckelynck, Nucleic Acids Res. 2016, 44, 2491-2500.

[72] W. Y. Zhang, W. H. Zhang, Z. Y. Liu, C. Li, Z. Zhu, C. J. Yang, Anal. Chem. 2012, 84, 350-355.

[73] N. Ma, E. H. Sargent, S. O. Kelley, Nat. Nanotechnol. 2009, 4, 121-125.

Accepted Manuscript
[74] J. Farlow, D. Seo, K. E. Broaders, M. J. Taylor, Z. J. Gartner, Y.-w. Jun, Nat. Methods 2013, 10,

1203-1205.

[75] W. Zhou, F. Wang, J. Ding, J. Liu, ACS Appl. Mater. Inter. 2014, 6, 14795-14800.

19

This article is protected by copyright. All rights reserved.