ABSTRACT 
 
 
 
 
Title of Dissertation: THE ROLE OF SEQUENCE IN THE 
STRUCTURE OF SELF-ASSEMBLING 3D 
DNA CRYSTALS 
  
 Maithili Saoji, Doctor of Philosophy, 2015. 
  
Dissertation Directed By: Dr. Paul Paukstelis, Assistant Professor,  
Department of Chemistry and Biochemistry. 
 
 
DNA is a widely used biopolymer for the construction of nanoscale objects due to its 
programmability and structural predictability. DNA oligonucleotides can, however, 
exhibit a great deal of local structural diversity. DNA conformation is strongly linked 
to both environmental conditions and the nucleobase identities inherent in the 
oligonucleotide sequence, but the exact relationship between sequence and local 
structure is not completely understood. We previously determined the X-ray crystal 
structure of a DNA 13-mer that forms a continuously hydrogen bonded three-
dimensional lattice through Watson-Crick and non-canonical base pairs. In the 
current work I examined how the sequence of the Watson-Crick duplex region 
influenced crystallization of this 13-mer. I screened all possible self-complementary 
sequences in the hexameric duplex region and found 21 oligonucleotides that 
crystallized. Sequence analysis showed that one specific Watson-Crick base pair 
influenced the crystallization propensity and the speed of crystal self-assembly. I 
  
determined X-ray crystal structures for 13 of these oligonucleotides and found 
sequence-specific structural changes suggesting that this base pair may serve as a 
structural anchor during crystal assembly. I explored the crystal self-assembly and 
nucleation process and demonstrated that crystals grown from mixtures of two 
different oligonucleotide sequences contained both the oligonucleotides. These results 
suggested that crystal self-assembly is nucleated by the formation of Watson-Crick 
duplexes. Finally, I also examined how a single nucleotide addition to the DNA 13-
mer leads to a significantly different overall structure under identical crystallization 
conditions. The 14-mer crystal structures described here showed that all of the 
predicted Watson-Crick base pairs were present, but the major difference as 
compared to the parent 13-mer structure was a significant rearrangement of non-
canonical base pairs. This included the formation of a sheared A-G base pair, a 
junction of strands formed from base triple interactions, and tertiary interactions that 
generated structural features similar to tandem sheared G-A base pairs. The adoption 
of this alternate non-canonical structure was dependent in part on the sequence of the 
Watson-Crick duplex region. These results provided important new insights into the 
sequence/structure relationship of short DNA oligonucleotides and demonstrated a 
unique interplay between Watson-Crick and non-canonical base pairs that are 
responsible for crystallization fate. 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
THE ROLE OF SEQUENCE IN THE STRUCTURE OF SELF-ASSEMBLING 3D 
DNA CRYSTALS 
 
 
 
by 
 
 
Maithili Saoji 
 
 
 
 
Dissertation submitted to the Faculty of the Graduate School of the  
University of Maryland, College Park, in partial fulfillment 
of the requirements for the degree of 
Doctor of Philosophy 
2015 
 
 
 
 
 
 
 
 
 
 
Advisory Committee: 
Professor Paul Paukstelis, Chair 
Professor Jason Kahn 
Professor David Fushman 
Professor Nichole LaRonde-LeBlanc 
Professor Ian White, Dean’s representative 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
© Copyright by 
Maithili Saoji 
2015 
  ii 
 
Dedication 
I dedicate this thesis to my loving husband Pushkar, for his relentless support and 
encouragement and my little angel Arya. I also dedicate it to my mom, dad, aajee, 
Shon and Purva for their constant faith in me and for helping me change my dream 
into reality, and to aai, baba and Neha for the never-ending love and support. Last but 
not the least, I dedicate this thesis to my loving friend Sudeshna for being with me 
through thick and thin. 
  iii 
 
Acknowledgements 
I want to thank Dr. Paul Paukstelis for his constant mentoring and support throughout 
my journey in his lab. My experience would not have been the same without him. I 
want to thank Dr. Jason Kahn, Dr. David Fushman and Dr. Nicole LaRonde for their 
support and guidance through all the committee meetings, Dr. Ian White for his time 
as the Dean’s representative and Dr. Amy Beaven, Dr. Silvia Muro and Dr. George 
Lorimer for opening their labs to me for my various experiments. I would like to 
extend a special thanks to my past and present lab mates, Mello, Stephanie, Diana, 
Ron, Alessandra, Ryan, Orion, Logan and Flo for their help from time to time and 
also for making the lab a fun place to work. 
The work presented in this document is adopted from the following research articles: 
 Probing the role of sequence in the assembly of three-dimensional DNA 
crystals 
 Maithili Saoji, Daoning Zhang & Paul J. Paukstelis 
 Biopolymers. 2015 May 26. doi: 10.1002/bip.22688. 
 Sequence-dependent structural changes in a self-assembling DNA 
oligonucleotide. 
Maithili Saoji, Paul J. Paukstelis 
Acta Crystallographica Section D (Accepted). 
 
  iv 
 
Table of Contents 
 
 
Dedication ..................................................................................................................... ii  
Acknowledgements ...................................................................................................... iii 
Table of Contents ......................................................................................................... iv 
List of Tables ............................................................................................................... vi  
List of Figures ............................................................................................................. vii  
Chapter 1: Introduction ................................................................................................. 1 
Chapter 1.1 DNA: A Structural Building Block ....................................................... 1 
Chapter 1.2 Structural DNA Nanotechnology .......................................................... 3 
Chapter 1.3 Structural Diversity of DNA ................................................................. 7 
Chapter 1.4 Motivation for Research ........................................................................ 8 
Chapter 1.5 Scope of Research ............................................................................... 10 
Chapter 2: Probing the Role of Sequence in the Self-Assembly Process of 13-mer 
DNA Crystals .............................................................................................................. 12 
Chapter 2.1 Introduction ......................................................................................... 12 
Chapter 2.2 G-A Pairs in the Duplex Not Required for Crystallization ................. 17 
Chapter 2.3 Sequence Requirements for Crystallization ........................................ 20 
Chapter 2.4 Structural Implications of the Duplex Sequences ............................... 25 
Chapter 2.5 B7 Does Not Form Base Pairs in Absence of Cations ........................ 31 
Chapter 2.6 Watson-Crick or Non-canonical Base Pairs as Drivers for 
Crystallization ......................................................................................................... 35 
Chapter 2.7 Influence of Sequence on Crystallizability, Structure and Self-
Assembly................................................................................................................. 42 
Chapter 2.8 Materials and Methods ........................................................................ 44 
1. Crystallization Screen ................................................................................. 44 
2. DNA Purification and Crystallization ......................................................... 44 
3. Data Collection and Refinement ................................................................. 45 
4. Sequence and Structure Comparison and Analysis .................................... 45 
5. Nuclear Magnetic Resonance ..................................................................... 46 
6. Heterogeneous Crystallization .................................................................... 47 
Chapter 3: Other Methods Used to Study the Cation Driven Assembly Process ....... 48 
Chapter 3.1 Dynamic Light Scattering ................................................................... 50 
Chapter 3.2 Ultra-Violet (UV) Spectroscopy ......................................................... 51 
Chapter 3.3 Fluorescence Spectroscopy ................................................................. 53 
Chapter 3.4 S1 Nuclease Digestion ........................................................................ 56 
Chapter 3.5 DNase I Digestion ............................................................................... 58 
Chapter 3.6 Confocal Microscopy .......................................................................... 60 
Chapter 3.7 Materials and Methods ........................................................................ 63 
1. Dynamic Light Scattering ........................................................................... 63 
2. UV Spectroscopy ........................................................................................ 63 
3. Fluorescence Spectroscopy ......................................................................... 63 
4. S1 and DNase I Digestion ........................................................................... 64 
5. Confocal Microscopy .................................................................................. 65 
  v 
 
Chapter 4: Sequence-Dependent Structural Changes in 14-mer DNA Oligonucleotide
..................................................................................................................................... 66  
Chapter 4.1 Introduction ......................................................................................... 66 
Chapter 4.2 Overview of the 14-mer Structures. .................................................... 69 
Chapter 4.3 B-Form Duplex Region Capped by Sheared A-G Pairs. ..................... 74 
Chapter 4.4 Tertiary Interactions Fulfil a Structural Role to Generate Tandem G-A 
Base Pairs. ............................................................................................................... 76 
Chapter 4.5 Triplex Junction................................................................................... 81 
Chapter 4.6 Sequence Requirements for Alternate Crystal Form........................... 83 
Chapter 4.7 Materials and Methods ........................................................................ 87 
1. DNA Synthesis and Purification ................................................................. 87 
2. Crystallization ............................................................................................. 87 
3. Data Collection and Structure Determination ............................................. 88 
Chapter 5: Conclusion and Future Prospects .............................................................. 89 
Bibliography ............................................................................................................... 91 
 
 
 
 
This Table of Contents is automatically generated by MS Word, linked to the 
Heading formats used within the Chapter text.   
 
 
 
 
 
 
  vi 
 
List of Tables 
 
Table 1. Data collection and refinement for B7. ......................................................... 18 
 
Table 2. Crystallized oligonucleotide sequences. ....................................................... 21 
 
Table 3. Data collection and refinement for 12 sequence variants. ............................ 26 
 
Table 4. Transverse relaxation experiments on B7 in presence of Mg
2+
. ................... 34 
 
Table 5. Heterogeneous crystals for pairwise combinations of 13-mer DNA 
oligonucleotides .......................................................................................................... 41  
 
Table 6. Data collection and refinement for 14-mer structures. ................................. 72 
 
Table 7. Sequence-dependent crystallization. ............................................................. 86 
 
 
  vii 
 
List of Figures 
Figure 1. B-form DNA helix
10
 (PBD ID: 1BNA). ........................................................ 2 
Figure 2. Representative 2D and 3D DNA-based architectures. .................................. 4 
Figure 3. The lattice formed by the tensegrity triangle
23
. ............................................. 6 
Figure 4 The crystal structure of d(GGACAGATGGGAG) (PDB ID 1P1Y)
59
. ........ 14 
Figure 5. Solvent channels in the 13-mer lattice......................................................... 15 
Figure 6. The 5’ GGA- motif in the structural context of the DNA 13-mer............... 16 
Figure 7. Structure of B7 variant compared to the original 13-mer............................ 19 
Figure 8. Representative crystal morphologies for DNA oligonucleotides identified in 
the screen. ................................................................................................................... 22 
Figure 9. Nucleobase frequency distribution of crystallizing DNA oligonucleotides. 23 
Figure 10. Superimposition of the 13-mer structures with and without the A5-T8 base 
pair. ............................................................................................................................. 29 
Figure 11. Average RMSD comparison for the duplex region of A5-T8 and the non-
A5-T8 structures.  . ..................................................................................................... 30  
Figure 12. 1D-
1
H NMR spectra of B7 variant. ........................................................... 33 
Figure 13. NMR signal broadening upon Mg
2+ 
addition. ........................................... 34 
Figure 14. Models for crystal pre-nucleation. ............................................................. 36 
Figure 15. Heterogeneous crystal from DNA mixtures. ............................................. 38 
Figure 16. Heterogeneous crystal from ratios of DNA mixtures. ............................... 39 
Figure 17. Monitoring the assembly of B7 DNA using UV Spectroscopy.  .............. 52 
Figure 18. Fluorescence intensity measurements on B7 DNA using nucleic acid 
specific fluorescent dyes. ............................................................................................ 55 
  viii 
 
Figure 19. Denaturing PAGE gel for samples digested with S1 nuclease. ................. 57 
Figure 20. Denaturing PAGE gel for samples digested with DNase I nuclease.. ....... 59 
Figure 21. Confocal images for unlabeled B7 core and fluorescent shells. ................ 61 
Figure 22. Confocal images to observe the B7 crystal assembly. .............................. 62 
Figure 23. 13-mer and 14-mer crystals. ...................................................................... 68 
Figure 24. Overview of 14-mer crystal structure ........................................................ 70 
Figure 25. Crystal packing. ......................................................................................... 71 
Figure 26. Structural overlap for the 14-mer structures. ............................................. 73 
Figure 27. Structural comparison of A3-13 and A3-14 monomers. ........................... 75 
Figure 28. Comparison of GA/AG motifs. ................................................................. 78 
Figure 29. Potential hydrogen bond between A3 and T9. .......................................... 79 
Figure 30. The inter- and intrastrand stacking at the sheared G-A base pair.............. 80 
Figure 31. Purine base triples...................................................................................... 82 
  1 
 
Chapter 1: Introduction 
Chapter 1.1 DNA: A Structural Building Block 
DNA is the genetic material for all living organisms. Through the complementary 
Watson-Crick base pairing (adenine-thymine and guanine-cytosine), it can 
successfully store and transmit hereditary information
1
. Remarkably, the properties 
that make DNA a powerful genetic material, also make it an attractive synthetic 
building block for nano and micrometer scale constructions
2
. Through the cooperative 
interplay of extremely precise Watson-Crick base pairing, base-stacking, electrostatic, 
and hydrophobic interactions, complementary DNA oligonucleotides self-assemble in 
solution to form an antiparallel B-form double helical structure
3
. The stability of the 
DNA double helix and its predictable structural and physical properties, like the 50 
nm persistence length, ∼2 nm diameter, ∼3.4 nm helical pitch, makes it desirable for 
creating nanostructures with predictable and precise geometries (Figure 1). Cations 
play an important role in the structure and folding of DNA
4
. Divalent cations, most 
commonly Mg
2+
, are used in the self-assembly of DNA based nanostructures to 
overcome electrostatic repulsion between DNA duplexes by screening the negative 
charge on the backbone phosphates
5–8
. Additionally, the success of using DNA for 
synthetic constructions is related to vast size of DNA sequence libraries that can be 
created through the permutation of the bases at each position of the oligonucleotides 
and the ease of chemical synthesis of any desired sequences with high yields and 
purity at relatively low costs using phosphoramidite chemistry
9
. By exploiting these 
unique properties of DNA the field of structural DNA nanotechnology has advanced 
  2 
 
to create highly sophisticated nanoscale architectures and to use DNA as a template 
for precise positioning of materials and molecules.  
 
 
Figure 1. B-form DNA helix
10
 (PBD ID: 1BNA). A. The right handed B-form DNA 
double helix highlighting the physical parameters. B. Top view of the B-form double 
helix with the diameter. 
 
̴ 
̴ 
  3 
 
Chapter 1.2 Structural DNA Nanotechnology 
DNA nanotechnology is a ‘bottom up’ construction process where dispersed 
DNA oligonucleotides self-assemble in solution to form complex architectures
11
. The 
first artificially constructed DNA nanostructure was the immobile crossover Holliday 
junction
12
 which was followed by a wide variety of more sophisticated architectures 
like the double-crossover (DX) DNA tiles
13
, triple-crossover (TX) tiles
14
, 4X4 tiles
7
,  
three-point-star tiles
15
 etc. Through an interplay of DNA helices and cross-over 
junction motifs, discrete higher order structures like the DNA cube
16
, polyhedra
15,17–
19
, nanotubes
20
 and repetitive two-dimensional (2D) DNA lattices
5,7,8,21,22
 and three-
dimensional (3D) tensegrity triangle based DNA crystals
23
 were constructed (Figure 2 
& 3).  
 Several non-periodic, spatially addressable and finite sized 2D and 3D DNA 
structures have been constructed using DNA origami technique
6,24–27
 where a long, 
circular, single-stranded genomic DNA, referred to as the scaffold-strand, is folded 
into various geometric shapes that are held together with short DNA oligonucleotides 
called the staple-strands. More recently the single-stranded DNA tile technique, in 
which single-stranded DNA tiles are interlocked with each other through local 
molecular connections,  has been successfully used to by-pass the use of staple-strand 
in the construction of discrete DNA objects
28–30 
(Figure 2). DNA has also been 
successfully used as a building block for the construction of dynamic nanoscale 
devices and in molecular electronics and computing
31,32
. 
 
 
  4 
 
Chen et. al. Nature. 
Goodman et. al. Science. 
He et. al. Nature. 
Aldaye et. al. J. Am Chem. 
Dietz et. al. Science. Han et. al. Science. 
Wei et. al. Nature. 
Douglas et. al. Nature. 
 
 
 
 
 
 
 
Figure 2. Representative 2D and 3D DNA-based architectures. The top panel 
shows the molecular model of a DNA cube
16
, DNA tetrahedron
17
, DNA 
dodecahedron
15
 and DNA biprism
18
 (left to right). The lower panel represents the 
sophisticated, discreet two and three dimensional architectures created using DNA 
origami
25-27
 (first three tiles) and single stranded tile technique
28
 (last tile). 
 
 
  5 
 
One of the long-standing goals of the DNA nanotechnology field has been the 
rational design and construction of periodic 3D DNA structures, or crystals, for use as 
molecular scaffolds. DNA based scaffolds as envisioned by Seeman could be used for 
ordering biological and non-biological guest molecules
12
. This could provide a 
method for determining crystal structures of otherwise difficult to crystallize proteins. 
Despite the advances in the DNA nanotechnology field, this goal has not been 
realized. This is mainly because only one truly periodic 3D DNA array composed 
entirely of Watson-Crick base pairs has been described to date (Figure 3)
23
. The 
Watson-Crick duplexes formed by complementary DNA oligonucleotides in solution 
are linear molecules, making the construction of periodic 3D array composed entirely 
of Watson-Crick duplexes challenging.  To create truly periodic 3D lattices we need 
to branch away from linearity. The DNA crossover junctions provide branching in the 
construction of the lattices however, these motifs are flexible making it difficult to 
achieve precise molecular associations. In our lab we use ‘non-Watson-Crick’ base 
pairing motifs to circumvent the issue of flexibility of the crossover junctions. Non-
canonical base pairs are rigid, could provide branching and controlled intermolecular 
interaction in multiple dimensions, for the construction of periodic 3D lattices. 
However the occurrence of these motifs is less predictable as compared to the 
Watson-Crick base pairs.  
  6 
 
 
 
Figure 3. The lattice formed by the tensegrity triangle
23
. The figure shows the X-
ray crystal structure of the lattice formed by the tensegrity triangle based tile, 
determined by the Seeman group (PDB ID: 3GBI). It is a 3D periodic lattice 
composed entirely of Watson- Crick base pairing interactions.  
 
  7 
 
Chapter 1.3 Structural Diversity of DNA 
  It has been known from some of the earliest structural studies that DNA can 
be both conformationally and structurally diverse
33
. Depending on the environmental 
conditions, B-form DNA can undergo conformational transitions to the A- and the Z-
forms
33–36
. Additionally, a variety of non-B-form DNA motifs have been 
characterized in vivo including DNA cruciform, hairpin structures, triplexes and 
quadruplexes
37–43
. There has been a growing appreciation for the use of these non-
Watson-Crick structural motifs
44
, including i-motifs
45
, A-motifs
46,47
, G-
quadruplexes
48,49
 and parallel-stranded motifs
50,51
, in the design of DNA nanoscale 
architectures and devices, as a way to provide structural diversity to nanoscale 
constructions. Watson-Crick duplexes are linear molecules and the use of non-
Watson-Crick motifs provide rigid branch points that enable precise and controlled 
intramolecular interactions in 3D space. One of the major areas of DNA structural 
biology over the course of several decades was in understanding how non-Watson-
Crick base pairs, or mismatches, could be accommodated in otherwise normal DNA 
helices, or are responsible for forming alternate DNA structures
52–54
. Non-canonical 
base pairs may be thermodynamically less stable in the context of a Watson-Crick 
helix
52
, but in certain sequences or other structural contexts they can be extremely 
stable
55
 making them invaluable for DNA structural nanotechnology. However, the 
non-canonical motifs are not as predictable as the Watson-Crick base pairing. 
  8 
 
Chapter 1.4 Motivation for Research 
To date several techniques like UV and fluorescent spectroscopy
56
, 
temperature controlled AFM assembly
57
, single molecule FRET
58
, have been 
employed to study the kinetics and thermodynamics of controlled complementary 
DNA self-assembly. These studies have enhanced the understanding of the DNA 
nanostructure assembly process, however, little is currently known about how 
periodic DNA arrays self-assemble into macroscopic crystals. A more thorough 
understanding of the assembly process would prove useful in the rational designing of 
DNA structures and help in optimizing the conditions for assembly, manipulation, and 
functionalization. This in turn will benefit both upstream design and downstream 
applications of DNA structural nanotechnology
56
. The goal of this study was to 
understand the effect of nucleobase identities on DNA crystallization and structure 
while examining the assembly process of 3D DNA crystals.  
Our lab previously determined a X-ray crystal structure of a continuously 
hydrogen bonded, self-assembling periodic 3D DNA crystal composed of both 
Watson-Crick and non-canonical base pairing interactions
59
. The 13-mer DNA lattice 
has been successfully used as a scaffold for the incorporation of small guest 
molecules, as a model biomaterial solid for terahertz spectroscopy studies
60,61
 and 
also as a template for the design and development of 3D crystal lattices with 
expanded solvent channels
62,63
. Here, I used this 13-mer DNA as a model to 
understand the relationship between sequence and structure. Specifically, I studied 
how the sequence in the duplex region of the 13-mer DNA influenced the 
crystallizability and the structure for these oligonucleotides and subsequently 
  9 
 
performed the first experiments to examine the self-assembly process for these 
crystals. 
  10 
 
Chapter 1.5 Scope of Research 
I identified 21 crystallizing sequence variants of the 13-mer DNA. Sequence 
comparison of the crystallizing DNA oligonucleotides suggested Watson-Crick base 
pair preferences at specific positions in the hexameric sequence, some of which also 
correlated with crystallization speed. Structure determination of 13 of these variant 
crystals revealed that the sequences are structurally isomorphous, with relatively 
minor sequence-related differences restricted to certain nucleotide positions.  Solution 
studies indicated that no base pairs are formed in the absence of additional divalent 
cations, and that assembly began immediately after the addition of Mg
2+
. Finally, I 
demonstrated that I could generate heterogeneous crystals from mixtures of 
oligonucleotides that differed in the sequence of the duplex region, suggesting that the 
formation of Watson-Crick duplexes may be the initial step enabling crystal 
nucleation. 
 I also examined how a single nucleotide change in the DNA sequence was 
capable of promoting a significant change in the oligonucleotide interactions and the 
corresponding crystal structure, under identical crystallization conditions. Here, I 
described the crystal structures of four 14-mer oligonucleotides obtained by adding an 
adenosine at the 3’ end of the DNA 13-mers. The 14-mer structures showed identical 
interactions in the Watson-Crick duplex region but a rearrangement of the non-
canonical base pairs. Additionally, I also examined the role sequence played in the 
adoption of the alternate crystal form. The analysis suggested that the sequence in the 
14-mer duplex region and the identity of the added nucleotide were necessary to 
promote this alternate structure. Remarkably, the added A14 residue from another 
  11 
 
strand made tertiary contacts to the guanosine adjacent to a single sheared A-G pair, 
resulting in a conformation similar to tandem sheared G-A pairs. Together with a 
series of purine base triples, these interactions were responsible for the formation of 
the alternate crystal form. This study is a step forward in the understanding of the 
complex sequence/structure relationship of DNA oligonucleotides. 
 
  
  12 
 
Chapter 2: Probing the Role of Sequence in the Self-Assembly 
Process of 13-mer DNA Crystals 
Chapter 2.1 Introduction 
Construction of a periodic 3D array composed entirely of Watson-Crick 
duplexes is challenging as the Watson-Crick duplexes are linear molecules. Repetitive 
periodic structures using immobile crossover junctions have been constructed 
however these junctions are flexible in nature
11
. Sophisticated 3D constructions 
require more precise junctions. However, DNA crystal assemblies are not limited to 
Watson-Crick base pairing interactions. Non-canonical base pairs can provide 
structural diversity for the creation of branched DNA structures. 
Hydrogen bonded non-Watson-Crick base pair combinations or Watson-Crick 
base pairs bonded through ‘non-Watson-Crick’ hydrogen bond donor-acceptor pairs 
are referred to as non-canonical base pairs. They occur naturally in RNA, the 
telomeric regions of DNA, and are associated with the structural and functional 
diversity of these molecules
55,64
. Even though non-canonical base pairs are 
thermodynamically less stable in the context of a Watson-Crick helix
52
, they can be 
both extremely predictable and stable
55
 in other structural contexts. There has been a 
growing appreciation for the use of non-Watson-Crick structural motifs in the design 
of DNA nanoscale architectures and devices.  
One such parallel-stranded homopurine 5'-GGA motif was first identified and 
reported in a 13-mer crystal structure from our lab
59
 (Figure 6A, B). In this crystal 
structure, the 13 nucleotide long DNA 5’ GGACAGATGGGAG 3’ is continuously 
hydrogen bonded through both Watson-Crick and non-canonical base pairs
59
 and self-
  13 
 
assembles in the presence of Mg
2+
. The crystal structure shows one DNA 13-mer base 
paired with three identical neighbors to form two distinct regions of base pairing 
(Figure 4A). The antiparallel Watson-Crick helical region is made up of a six base 
pair long self-complementary duplex composed of two sets of C4-G9 and A5-T8 
Watson-Crick base pairs and two Type II
65
 G6-A7 non-canonical base pairs about the 
central dyad axis (Figure 4B-D). Residues G1-A3 form parallel-stranded non-
canonical homopurine base pairs with G10-A12 of another monomer (Figure 4E). 
Two sets of these non-canonical base pairs stack on a crystallographic symmetry axis 
to serve as a junction that links different helical layers into a continuously hydrogen 
bonded 3D lattice structure (Figure 4F). These interactions lead to the formation of 
solvent-occupied channels running throughout the length of the crystal. Two kinds of 
solvent channels with 300 Å
2
 and 360 Å
2 
cross-section area, run parallel and 
perpendicular to the six-fold symmetry axis respectively (Figure 5A, B). These 
solvent channels allow incorporation of guest molecules within the lattice.  
The 13-mer crystals have also found application as a model biomaterial solid 
for terahertz spectroscopy studies
60,61
. The GGA motif identified from the 13-mer 
structure has been successfully used in different sequence contexts to rationally 
design DNA crystals with expanded solvent channels to allow incorporation of 
protein guest molecules
62,63
. My work was to understand how the nucleobase 
identities of the B-form duplex region influenced crystallization and in the process 
understand the assembly process for the model 13-mer DNA crystals, to help guide 
rational designing of 3D DNA crystals better suited for downstream applications as 
molecular scaffolds.  
  14 
 
 
 
Figure 4 The crystal structure of d(GGACAGATGGGAG) (PDB ID 1P1Y)
59
. A. 
Secondary structure showing the interaction between neighboring 13-mers (each 
strand colored differently). B-F. Show the generation of the continuously hydrogen 
bonded 3D lattice for the DNA 13-mer, through the formation Watson-Crick and non-
canonical interactions. 
 
  15 
 
                              
 
 
 
                                                     
 
 
                                
 
 
Figure 5. Solvent channels in the 13-mer lattice. The organization of the 13-mer in 
the 3D crystal lattice results in solvent channels of 300 Å
2 
and 360 Å
2
 running parallel 
(A) and perpendicular (B) to the six-fold symmetry axis respectively. 
90º 
  16 
 
 
 
 
Figure 6. The 5’ GGA- motif in the structural context of the DNA 13-mer. A. 
Two sets of GGA motifs (red and black) form parallel homopurine base pairs. B. 
Individual non-canonical base pairs of the GGA motif. In the first pair, the Watson-
Crick edge of G1 (N1, N2) hydrogen bonds with the Hoogsteen edge of G11 (N7, 
N6). In the second pair both guanines, G2 (N2, N3) and G11 (N3, N2) are hydrogen 
bonded through the sugar edge. In the third pair, the Hoogsteen edges of both 
adenosines, A3 and A12, are base paired to each other through symmetrical N6-N7 
hydrogen bonds. The hydrogen bonds are shown as dotted lines. 
 
 
  17 
 
Chapter 2.2 G-A Pairs in the Duplex Not Required for Crystallization 
 In the original 13-mer structure, the six base pair duplex region contains 
central G6-A7 non-canonical base pairs flanked by two Watson-Crick base pairs 
(Figure 3A). The presence of these non-canonical base pairs did not significantly 
distort the overall B-form helical geometry. However, the large propeller angle of the 
G6-A7 non-canonical base pair results in an additional cross-strand hydrogen bond 
between N2 of G6 and O2 of T8. A similar hydrogen bond was previously reported in 
the d(CCAAGATTGG) duplex structure
66
.  
 To determine if the duplex region could be composed entirely of Watson-
Crick pairs, I tested an A7C substitution that converted this region to a self-
complementary duplex. The A7C (also referred to as B7 in this document) 
oligonucleotide crystallized under the same conditions and required only the presence 
of Mg
2+
 cations to assemble. The B7 crystals were morphologically identical to the 
original 13-mer, crystallized in the same space group with one molecule in the 
asymmetric unit and had nearly identical unit cell parameters (Table 1). The crystal 
structure of B7 was highly similar to the original 13-mer (Figure 7) with the expected 
interactions leading to the formation of a Watson-Crick anti-parallel helical region 
and the parallel-stranded non-canonical junction. As in the original structure, the 
electron density of the G13 phosphate was observed, while the remainder of the 
nucleotide, which was located within the solvent channel, was disordered. 
  18 
 
      Table 1. Data collection and refinement for B7. 
 
B7  
Data collection 
 
Space group P6
4
 
Cell dimensions  
a, b, c (Å) 
40.641 40.641 51.983 
 () 90, 90, 120 
I/σI 17.2 (2.8) 
Number of reflections 2549 (386) 
Completeness (%) 97.0 (98.1) 
Redundancy 3.4 (2.9) 
Refinement 
 
Resolution (Å) 2.16-35.20 (2.16-2.21) 
Number of reflection 2287 (170) 
R factor 0.213 (0.492) 
R
free
 0.248 (0.408) 
Number of atoms  
DNA 254 
Ion 1 
Water 3 
Bond lengths (Å) 0.01 
Bond angles (˚) 1.57 
PDB ID 4ROK 
         *Values in parentheses are for the highest-resolution shell. 
  19 
 
 
 
Figure 7. Structure of B7 variant compared to the original 13-mer. The original 
1P1Y structure (green), superimposed on the B7 structure (blue). The sigma A-
weighted electron density map (2F
0
-F
c
) of B7 variant is contoured at 1σ. Refined 
solvent molecules are shown as points with associated density. Only the phosphate of 
G13 (G13P) was present in the electron density for both structures. 
 
 
  20 
 
Chapter 2.3 Sequence Requirements for Crystallization 
 To determine the sequence requirements within the duplex region necessary 
for crystal self-assembly, I screened all 64 possible self-complementary hexameric 
sequences. I maintained the GGA and GGAG sequences at the 5’ and the 3’ ends, 
respectively, to promote the formation of predictable non-canonical interactions 
required for the formation of the inter-layer junction that would hold the Watson-
Crick duplexes in a crystal framework similar to the original 13-mer. This also helped 
me to correlate any observed structural changes to the sequence of the duplex region. 
Using just a single crystallization condition, I identified 20 new crystallizing DNA 
oligonucleotides along with the B7 variant as a positive control (Table 2). The size 
and morphology of the crystals ranged from large hexagonal pyramids that were 
identical to the B7 variant, to birefringent spherulites, clusters, and microcrystals 
(Figure 8). Crystals also appeared at different times and were classified into two 
groups. The fast group crystallized within 16 hours, and the slow group took more 
than 48 hours to crystallize. Out of the 21 crystallizing DNA oligonucleotides in the 
screen (including the B7), 10 belonged to the fast group and the remaining 11 
belonged to the slower group (Table 2).  
  21 
 
 
Table 2. Crystallized oligonucleotide sequences. 
Designation Sequence Crystal 
morphology 
(Unpurified 
DNA) 
Crystallization 
speed 
Crystal 
morphology 
(purified DNA) 
1p1y 
GGA CAGATG GGAG 
- Fast Hexagonal Pyramids 
A1 
GGA AAATTT GGAG 
Microcrystals Fast Non-faceted hexagonal 
A2 
GGA AACGTT GGAG 
Microcrystals Fast Hexagonal Pyramids 
A3 
GGA AAGCTT GGAG 
Microcrystals Fast Hexagonal Pyramids 
A4 
GGA AATATT GGAG 
Spherulites Fast No crystals 
A7 
GGA ACGCGT GGAG 
Hexagonal Pyramids Slow No crystals 
A12 
GGA AGTACT GGAG 
Spherulites Slow No crystals 
B2 
GGA ATCGAT GGAG 
Hexagonal Pyramids Slow Hexagonal Pyramids 
B6 
GGA CACGTG GGAG 
Microcrystals Fast Hexagonal Pyramids 
B7 
GGA CAGCTG GGAG 
Hexagonal Pyramids Fast Hexagonal Pyramids 
B8 
GGA CATATG GGAG 
Spherulites Slow Spherulites 
B9 
GGA CCATGG GGAG 
Microcrystals Slow Hexagonal Pyramids 
B11 
GGA CCGCGG GGAG 
Hexagonal Pyramids Slow Hexagonal Pyramids 
C1 
GGA CGATCG GGAG 
Microcrystals Slow Non-faceted hexagonal 
C2 
GGA CGCGCG GGAG 
Spherulites Slow Spherulites 
C3 
GGA CGGCCG GGAG 
Intermediate 
morphology 
Fast Hexagonal Pyramids 
D5 
GGA GGATCC GGAG 
Spherulites Slow Spherulites 
E1 
GGA TAATTA GGAG 
Microcrystals Fast Hexagonal Pyramids 
E2 
GGA TACGTA GGAG 
Hexagonal Pyramids Fast Hexagonal 
Clusters 
E3 
GGA TAGCTA GGAG 
Hexagonal Pyramids Fast Hexagonal Pyramids 
E4 
GGA TATATA GGAG 
Spherulites Slow No crystals 
F1 
GGA TTATAA GGAG 
Cubic Slow No crystals 
 
Bold: DNA oligomers with crystal structures determined 
  22 
 
 
Figure 8. Representative crystal morphologies for DNA oligonucleotides 
identified in the screen (Table 2). A. Hexagonal pyramid. B. Microcrystals. C. Non-
Faceted hexagon. D. Spherulite. E. Hexagonal clusters. F. Cubic 
  
  
  23 
 
Sequence analysis of the crystallizing oligonucleotides revealed that all possible 
Watson-Crick base pairs were represented at least once at each of the six positions in 
the duplex region (Figure 9A). The strongest preference was for the A5-T8 base pair 
(11 of 21) with C4-G9 being the second most prevalent pair (8 of 21). Because of 
readily observable differences in the crystallization speed, I also compared sequences 
within the fast group (Figure 9B). Remarkably, 9 out of 10 oligonucleotides in the fast 
group had an A5-T8 base pair, while only 2 out of 11 from the slow group had the A5-
T8 base pair. The most uncommon base pair was the G4-C9 base pair, with only one 
sequence containing this pairing. Notably, this belonged to the slow crystallization 
group and formed only birefringent spherulites, even after purification. 
 
 
Figure 9. Nucleobase frequency distribution of crystallizing DNA 
oligonucleotides. Letter height reflects the frequency for each nucleotide position. A. 
Frequency distribution for the 21 DNA oligonucleotides identified in the initial 
crystal screen (including B7). B. Frequency distribution for the 10 DNA 
oligonucleotides belonging to the fast crystallizing sub-group (including B7). 
 
  24 
 
 The Ho laboratory took a similar screening approach to examine how 
sequence influences DNA structure
67
. There, self-complementary duplexes containing 
d(CCnnnNNNGG) sequences showed sequence and environmental preferences for A-
form, B-form, and 4-way junction structures. We anticipated that positive 
crystallization results from our study should correlate with sequences that 
preferentially take the B-form under crystallization conditions from their study. 
Interestingly, however, we found that only 3 of the 21 hexameric duplex sequences 
(A2, A12 and E1) were previously identified as preferential B-form sequences. Four 
of the sequences were identified as having A-form preferences (B9, B11, C1 and C3), 
while the remaining 14 had not been characterized. This indicates that the flanking 
regions, in our case the critical non-canonical base pairs, may play a significant role 
in how self-complementary sequences influence macromolecular conformation. 
  25 
 
Chapter 2.4 Structural Implications of the Duplex Sequences 
 I solved the X-ray crystal structures of 12 different sequence variants to 
determine if the different duplex sequences had any observable structural 
implications. 14 out of the 20 new DNA oligonucleotides identified in the screen 
crystallized after gel purification (Table 2; Materials & Methods). From these, 12 
gave morphologically similar hexagonal pyramidal crystals, while 2 yielded crystals 
that did not diffract x-rays. These 12 variants diffracted x-rays to high resolution 
limits of 2.03-2.39 Å, crystallized in the same space group, and had a narrow 
distribution of unit cell parameters (Table 3). Structurally, all variants were highly 
similar with the expected base pairing interactions. The average RMSD for the 
invariant 5' and 3' residues of all structures when compared to B7 was 0.31 Å, while 
the average RMSD for all backbone atoms compared to B7 was 0.49 Å. The G13 
sugar and nucleobase was disordered in all the structures. 
 
 
 
 
 
  26 
 
Table 3. Data collection and refinement for 12 sequence variants. 
 
A1 A2 A3 B2 B6 B9 
Data collection 
      
Space group P6
4
 P6
4
 P6
4
 P6
4
 P6
4
 P6
4
 
Unit cell (a,b,c) 
(Å) 
39.729 
39.729  
55.989 
40.399 
40.399 
54.627 
40.317 
40.317 
54.958 
40.232 
40.232 
52.910 
40.562 
40.562 
51.411 
40.719 
40.719 
51.214 
Resolution (Å)* 
2.08-55.99 
(2.08-2.19) 
2.04-34.99 
(2.04-2.15) 
2.03-54.96 
(2.03-2.14) 
2.08-52.91 
(2.08-2.19) 
2.08-35.13 
(2.08-2.20) 
2.39-51.21 
(2.39-2.52) 
R
merge
 
0.051 
(0.390) 
0.034 
(0.911) 
0.030 
(2.171) 
0.021 
(0.408) 
0.025 
(0.395) 
0.030 
(0.747) 
I/σI 24.8 (3.5) 24.3 (1.3) 26.3 (0.9) 26.2 (1.7) 23.0 (1.4) 27.1 (1.8) 
Number of 
reflections 
3045 (426) 3261 (483) 3297 (482) 2873 (347) 2578 (321) 1953 (289) 
Completeness 
(%) 
99.2 (95.7) 99.6 (99.5) 99.3 (99.5) 96.0 (79.2) 88.5 (75.3) 99.6 (100) 
Redundancy 5.2 (2.5) 3.8 (3.9) 5.0 (5.1) 3.1 (1.8) 2.8 (1.6) 5.0 (5.1) 
Refinement 
      
Resolution (Å) 
2.08-34.41 
(2.07-2.13) 
2.04-34.99 
(2.04-2.09) 
2.03-34.92 
(2.03-2.08) 
2.08-34.84 
(2.07-2.13) 
2.08-35.13 
(2.08-2.13) 
2.39-35.26 
(2.39-2.45) 
Number of 
reflections 
2738 (182) 2952 (227) 2952 (191) 2574 (149) 2325 (137) 1743 (128) 
R/R
free
 
0.200 
0.239 
0.226 
0.259 
0.221 
0.251 
0.225 
0.265 
0.216 
0.219 
0.210 
0.252 
Number of atoms       
DNA 254 254 254 254 254 254 
Ion 2 1 1 1 1 1 
Water 9 8 7 7 2 2 
Bond lengths (Å) 0.01 0.01 0.01 0.01 0.01 0.01 
Bond angles (˚) 1.8 1.06 1.17 1.23 1.13 1.33 
PDB ID 4RNK 4RO4 4RO7 4RO8 4ROG 4RON 
*Values in parentheses are for the highest-resolution shell. 
  27 
 
 
Table 3. Data collection and refinement (continued) 
 
B11 C1 C3 E1 E2 E3 
Data collection 
      
Space group P6
4
 P6
4
 P6
4
 P6
4
 P6
4
 P6
4
 
Unit cell (a,b,c) 
(Å) 
40.789 
40.789 
52.007 
40.442 
40.442 
52.514 
40.755 
40.755 
51.819 
40.177 
40.177 
53.015 
40.805 
40.805 
51.668 
40.780 
40.780 
51.813 
Resolution (Å)* 
2.37-52.01 
(2.37-2.50) 
2.09-35.02 
(2.09-2.20) 
2.08-35.29 
(2.08-2.19) 
2.19-34.79 
(2.19-2.31) 
2.27-35.34 
(2.27-2.40) 
2.32-35.32 
(2.32-2.45) 
R
merge
 
0.025 
(1.257) 
0.029 
(0.595) 
0.030 
(0.456) 
0.024 
(1.369) 
0.037 
(1.273) 
0.030 
(0.582) 
I/σI 23.0 (0.9) 33.6 (1.8) 19.5 (1.2) 31.2 (1.1) 16.4 (1.0) 21.1 (1.6) 
Number of 
reflections 
1999 (291) 2944 (421) 2663 (322) 2540 (375) 2219 (321) 2079 (317) 
Completeness 
(%) 
97.9 (99.3) 99.7 (98.4) 89.1 (74.8) 99.7 (99.2) 97.7 (97.4) 96.3 (97.4) 
Redundancy 3.8 (3.8) 6.1 (2.9) 2.8 (1.5) 5.5 (5.3) 2.8 (2.9) 2.9 (2.9) 
Refinement 
      
Resolution (Å) 
2.37-35.32 
(2.37-2.43) 
2.09-35.02 
(2.08-2.13) 
2.08-35.29 
(2.07-2.13) 
2.19-34.79 
(2.19-2.24) 
2.27-35.34 
(2.27-2.33) 
2.32-35.32 
(2.32-2.38) 
Number of 
reflections 
1772 (128) 2655 (204) 2397 (143) 2285 (148) 2003 (140) 1878 (147) 
R/R
free
 
0.237 
0.301 
0.221 
0.250 
0.220 
0.244 
0.222 
0.231 
0.202 
0.243 
0.202 
0.231 
Number of atoms       
DNA 254 254 254 254 254 254 
Ion 1 1 1 1 1 1 
Water 0 4 1 2 1 3 
Bond lengths (Å) 0.01 0.01 0.01 0.01 0.01 0.01 
Bond angles (˚) 1.44 1.01 0.98 1.08 1.29 1.16 
PDB ID 4ROO 4ROY 4ROZ 4RP0 4RP1 4RP2 
*Values in parentheses are for the highest-resolution shell. 
 
  
  28 
 
 To examine if the sequence preference for the A5-T8 base pair observed in the 
initial screen translated into local structure differences, I aligned the structures 
containing the A5-T8 pair and those without and compared the sugar-phosphate 
backbone atoms at the six duplex positions (Figure 10 & 11). The average backbone 
RMSDs for those structures containing the A5-T8 base pair was 0.65 Å, while the 
average RMSDs for the group without the A5-T8 pair was 0.93 Å. Analysis of the 
backbone by nucleotide position suggested that the lower RMSD values for the A5-
T8 group was not simply due to those structures having uniform nucleobase identities 
at these positions. The A5 and T8 backbone atoms had average RMSDs of 0.61 and 
0.67 Å, respectively, while positions 4 and 6 with variable nucleobases had lower or 
similar average RMSD values (0.51 and 0.64 Å). In contrast, the non-A5-T8 
structures had higher average backbone RMSDs for all but position 9, with positions 
5 and 6 having the highest variability (1.25 and 1.10 Å, respectively). The lower 
average RMSDs around position 5 when these residues are adenosines indicated that 
the backbone conformation in these structures is much less variable suggesting that 
this nucleobase is important for establishing a local structure that fits the crystal 
framework. This is also supported by the observation that A5-T8 is present in 9 out of 
10 fast crystallizing oligonucleotides. 
  29 
 
 
 
Figure 10. Superimposition of the 13-mer structures with and without the A5-T8 
base pair. The constant G1-A3 and G10-G12 residues were used as a reference for 
alignment. A. The overlap between 8 structures (A1, A2, A3, B6, B7, E1, E2 and E3) 
having the A5-T8 base pair. B. The duplex region of the overlapped A5-T8 structures. 
C. 5 superimposed structures (B2, C1, C3, B9 and B11) lacking the A5-T8 base pair. 
D. The duplex region of the non-A5-T8 structures. 
 
 
  30 
 
 
 
 
Figure 11. Average RMSD comparison for the duplex region of A5-T8 and the 
non-A5-T8 structures.  The average room mean square deviation (RMSD) values, 
calculated by comparing all structures to each other in a pairwise manner, for 
backbone atoms for positions 4 through 9 of the duplex regions are grouped into A5-
T8 (black) and non-A5-T8 (gray) groups. 
  
 
  31 
 
Chapter 2.5 B7 Does Not Form Base Pairs in Absence of Cations 
 To begin to understand how these DNA 13-mers assemble into periodic 
crystal lattices, we examined the behavior of the oligonucleotides in solution. Using 
the B7 variant as a representative sequence, we examined the oligonucleotide by 1D- 
1
H NMR with the assistance of Dr. Daoning Zhang. The spectrum showed sharp 
proton signals (Figure 12A) at conditions analogous to those used to store the 
oligonucleotide prior to crystallization (200 µM in water with 10% D
2
O). 
Importantly, the lack of imino proton signals in the 9-14 ppm range suggested that the 
oligonucleotides did not form stable Watson-Crick or non-canonical base pairs in 
absence of cations. These results indicated that B7 is primarily a dispersed monomer 
under these conditions.  
 Even though the initial crystal growth was performed in a buffer containing 
120 mM magnesium formate, 50 mM lithium chloride, 10% 2-methyl-2,4-
pentanediol, B7 DNA crystallized in just a 120 mM solution of  Mg
2+
 solution, 
demonstrating that divalent cations were the only minimal requirement for 
crystallization. To examine the effect of cation on the NMR spectra, we performed a 
Mg
2+
 titration on the B7 sample. We observed that addition of 5 mM Mg
2+
 led to 
significant signal broadening that became more severe with increasing Mg
2+
 
concentration (Figure 13). This suggested the formation of high molecular weight 
assemblies. Some areas of the spectrum showed chemical shift` changes upon 
addition of the cation. This was particularly true for the nucleobase proton region 
(Figure 12B) and the deoxyribose H1' region (Figure 12C). These chemical shift 
changes are likely the result of the altered environment upon base pairing 
  32 
 
interactions, while the signal broadening is a result of the assembly into high 
molecular weight complexes. However, only two low intensity peaks in the imino 
proton region appeared after the addition of Mg
2+ 
(Figure 12D). These peaks were 
most prominent at low Mg
2+
 and diminished at higher concentrations. The formation 
of discrete stable species through Watson-Crick or the non-canonical base pairs 
should likely have resulted in a larger number of imino signals with greater intensity. 
The limited number of imino peaks likely reflected the rapid base pairing and 
assembly of these oligonucleotides into high molecular weight structures. 
 To further investigate cation-induced assembly, we determined transverse 
relaxation times (T
2
) for the oligonucleotide both above and below stoichiometric 
Mg
2+
 concentrations, again in collaboration with Dr. Daoning Zhang (Table 4). T
2
 is 
inversely proportional to the tumbling time of the molecule in solution, resulting in 
lower T
2
 values as molecular weight increases.  Titration of sub-stoichiometric Mg
2+
 
(1:2) resulted in a small but measurable (3 ms) decrease in T
2
.  A similar decrease (3 
ms) was observed as Mg
2+
 was increased to equimolarity. Interestingly, a significantly 
greater decrease in T
2
 (19 ms) was observed as Mg
2+
 exceeded the DNA 
concentration (2.5:1), but did not change as Mg
2+
 was increased again (4:1). We could 
not accurately determine T
2
 at higher Mg
2+
 ratios due to the significant line 
broadening that resulted in a loss of peaks at the longer decay times necessary for 
fitting. However, these results show that even low concentrations of divalent cations 
can induce formation of higher molecular weight structures. 
  33 
 
 
 
Figure 12. 1D-
1
H NMR spectra of B7 variant. A. The complete spectrum for the 
B7 (A7C) variant in 0-200mM Mg
2+ 
ions obtained in collaboration with Dr. Daoning 
Zhang. The base proton region (B), and the deoxyribose H1’ region (C) shows 
broadening of peaks and chemical shift changes upon addition of increasing amounts 
of Mg
2+
 ions. D. The imino proton region lacks proton signals in the absence of Mg
2+
. 
Two low intensity signals appear at low Mg
2+
 concentrations, and disappear as Mg
2+
 
concentration was increased. 
  34 
 
 
Figure 13. NMR signal broadening upon Mg
2+ 
addition. Base proton resonance 
peaks were measured at different Mg
2+
 concentrations (right) in collaboration with 
Dr. Daoning Zhang. The width of the peak at half its total height is shown plotted 
against Mg
2+
 titration concentrations (left). 
 
 
Table 4. Transverse relaxation experiments on B7 in presence of Mg
2+
. 
 
DNA (mM) MgCl
2 
(mM) Mg
2+
/DNA 1/T
2
 T
2
 (ms) Fit (R
2
) 
0.2 0 0 16.1 62 0.9986
 
0.2 0.1 0.5 16.87 59 0.9932 
0.2 0.2 1 17.85 56 0.9997 
0.2 0.5 2.5 26.85 37 0.9979 
0.2 1 4 27.31 37 0.9776 
 
  35 
 
Chapter 2.6 Watson-Crick or Non-canonical Base Pairs as Drivers for 
Crystallization 
 Though there are many routes for crystal assembly and nucleation
68
, one 
interesting possibility in this system is that the addition of divalent cations initially 
promotes intermolecular interactions that form pre-nucleation assemblies.  In the 
absence of observable hydrogen bonding interactions in solution NMR experiments, 
we considered two simple scenarios for how base pairing could lead to pre-nucleation 
species and ultimately to crystal nucleation. The “duplex first” model would require 
the formation of base paired dimers through the specificity of the hexameric anti-
parallel Watson-Crick duplex (Figure 14A). The dimers could then form non-
canonical interactions through the free purine motifs at both ends of the duplexes, 
leading to crystal nucleation (Figure 14B). In the “non-canonical first” model the 
stability provided by the extensive base stacking interactions of the non-canonical 
region could initiate strand interactions, followed by the Watson-Crick interactions 
leading to the formation of a crystal nucleation site (Figure 14C). Analyzing these 
models from the perspective of mixed sequences, our analysis favored the duplex first 
model, as this process would always generate productive assemblies that could 
continue to propagate during crystal growth without base pair clashes. In contrast, 
depending on the associations formed during the assembly process, the non-canonical 
first model could lead to unproductive base pair clashes in the duplex regions that 
could potentially impact both crystal nucleation and growth (Figure 14D). Though we 
recognize that these models are not necessarily mutually exclusive and that the 
relatively long incubation times during crystal growth could allow for equilibrium 
  36 
 
partitioning, the identification of multiple DNA sequences capable of forming 
isomorphous structures allowed us to test these models. 
 
 
 
Figure 14. Models for crystal pre-nucleation. The duplex first and the non-
canonical first models for crystal pre-nucleation explained by mixed sequence 
crystallization. DNA oligonucleotides differing in just the duplex sequence are shows 
in red and cyan. A. In the duplex first model, a mixture would partition first into 
respective complementary duplexes following the addition of Mg
2+
. B. These 
duplexes could then assemble into heterogeneous higher-order structures through 
non-canonical pairings (black) leading to crystal nucleation. C. In the non-canonical 
first model, DNAs would assemble randomly through the non-canonical interactions, 
resulting in both homogeneous and heterogeneous dimeric structures. D. These 
dimeric structures could then assemble into higher-order structures, however, this 
could result in duplex clashes, as indicated. 
  37 
 
 To test the two models for crystal pre-nucleation, I created heterogeneous 
crystals from mixtures of different DNA oligonucleotide sequences identified from 
our screen. To detect both DNAs present in the heterogeneous crystals by gel 
electrophoresis I first tested mixtures in which one oligonucleotide contained a single 
nucleotide 3' extension. I confirmed that the addition of a 14
th
 adenosine nucleotide 
did not prevent crystallization for 7 out of 10 extension variants I tested (A1, A2, A3, 
B6, B7, E1 and E3). However, for three of these (A1, A2 and A3), the addition of the 
3' residue resulted in morphologically distinct crystals with different space groups and 
unit cell parameters (Figure 23B-E). These structures will be described in detail in 
Chapter 3. The remaining three 14-mer variants, all belonging to the slow group in 
the initial screen of 13-mers (B2, B11 and C1), did not crystallize under the tested 
condition. 
 Washed and dissolved crystals grown from two oligonucleotides with identical 
duplex sequences, but different lengths showed that the relative ratios of the two 
oligonucleotides in single crystals were consistent with the ratio of the 
oligonucleotides in the input mixture (Figure 15A and Figure 16). This indicated there 
was no preference for the longer or shorter oligonucleotide during crystal assembly, 
consistent with the 13 and 14 positions being oriented into the crystal solvent 
channels. Heterogeneous crystals grown with different duplex sequences of different 
length mixed together showed similar results (Figure 15B, C & Figure 16).  Both 
oligonucleotides were present in all combinations of heterogeneous 13- and 14-mer 
mixtures that I tested, with only the E3-13/B7-14 combination showing some 
preference for incorporation of the 14-mer relative to the input mixture.  
  38 
 
 
 
  
 
 
 
  
Figure 15. Heterogeneous crystal from DNA mixtures. A. Denaturing polyacrylamide gel 
for heterogeneous single crystals grown from a mixture of 13 and 14 nucleotide long DNAs 
with the same sequence in the duplex region (B7-13 and B7-14). B. A denaturing 
polyacrylamide gel for heterogeneous single crystals grown from a mixture of 13 and 14 
nucleotide long DNAs with different sequence in the duplex region (indicated). For both gels, 
lanes designated 13 and 14 are the separate oligonucleotides; M, oligonucleotide mixture 
used for crystallization; X, single crystals grown from the mixture; 13X and 14X, single 
crystals grown from the individual oligonucleotides. C. Quantification of gels (A and B) 
showing the ratio of 13 nucleotide (light grey) and 14 nucleotide (dark grey) DNAs in the 
crystallization mixture (M) and averages of single crystals (X) obtained from the mixture. 
Black regions at the interface are standard deviations for four crystal (B7-13 + B7-14) or 
three crystals (remaining mixtures). 
  39 
 
 
 
Figure 16. Heterogeneous crystal from ratios of DNA mixtures. Denaturing 
polyacrylamide gel for heterogeneous single crystals grown from a mixture of 13 and 
14 nucleotide long DNAs with the same sequence in the duplex region (B7-13 and 
B7-14) and mixed in different ratios to each other. The ratio’s are 14 mer DNA : 13 
mer DNA based on the input mixture. Lanes designated 13 and 14 are the separate 
oligonucleotides; M, oligonucleotide mixture used for crystallization; X, single 
crystals grown from the mixture; 13X and 14X, single crystals grown from the 
individual oligonucleotides. 
 
 
 
 
  40 
 
 Finally, I screened 78 combinations through pair-wise mixing of the 13 DNA 
oligonucleotides for which the structures were determined. 74 out of 78 mixtures 
produced crystals of the typical size (Table 5) irrespective of the number of base pair 
differences in the mixed oligonucleotides. All the combinations gave me normal sized 
crystals. These results supported the model of crystal assembly by the formation of 
Watson-Crick pairs. The DNA mixtures did not inhibit crystallization in any 
appreciable way as expected from the non-canonical first model. While it is possible 
that some combinations could support crystallization through non-Watson-Crick 
pairings in the duplex region (similar to the original 13-mer structure), it is unlikely 
that all 74 of 78 pairwise mixtures could do so. Therefore, it is likely that both the 
initial crystal assembly steps and the propagation of crystal growth rely primarily on 
the formation of Watson-Crick base pairs. This also suggests that during crystal 
growth, the units incorporated into the crystals have already formed Watson-Crick 
duplexes. It is not clear if the incorporated unit is a duplex or some larger assembly. 
Further studies will be necessary to directly detect these base pairs either before or 
during the assembly process.  
  41 
 
Table 5. Heterogeneous crystals for pairwise combinations of 13-mer DNA 
oligonucleotides 
 
    A1 A2 A3 B2 B6 B7 B9 B11 C1 C3 E1 E2 E3 
    AAA AAC AAG ATC CAC CAG CCA CCG CGA CGG TAA TAC TAG 
E3 TAG 
                          
E2 TAC 
                        
E1 TAA 
                      
C3 CGG 
                    
C1 CGA 
                  
B11 CCG 
                
B9 CCA 
              
B7 CAG 
            
B6 CAC 
          
B2 ATC 
        
A3 AAG 
      
A2 AAC 
    
A1 AAA 
  
 
78 total combinations 
19 combinations with 3 base pair difference 
35 combinations with 2 base pair difference 
20 combinations with 1 base pair difference 
4 combinations with no crystals 
single nucleotide crystals 
  42 
 
Chapter 2.7 Influence of Sequence on Crystallizability, Structure and Self-Assembly 
 
 This work has revealed many new DNA oligonucleotide sequences capable of 
adopting isomorphous crystal structures. Through analysis of sequence and structure I 
have found a strong preference for only one base pair position, A5-T8, which also 
correlates with crystallization speed. The structural differences between those 
structures with this base pair and those without, suggests that the A5-T8 pair provides 
a more uniform structural framework for crystallization.  Based on these 
observations, we proposed that the A5-T8 base pair may function as a structural 
anchor that enables the hexameric duplex region to adopt an overall structure that is 
more amenable for crystallization. Interestingly, the overall magnitude of the 
differences between the A5-T8 and non-A5-T8 structures was relatively modest. This 
may be due in part to all the structures fitting into a common crystal framework, but it 
may also suggest that the significance of the A5-T8 pair may not be restricted to its 
contribution to the final structures. One interesting possibility is that this base pair 
identity is important during the crystal assembly process. Nucleobase identities at 
particular positions can have a significant influences on helix bending and 
breathing
69–72
. Analysis using X3DNA
73
 indicated there was little difference in the 
duplex helical bend angle in the final structures, however, it remains possible that the 
A-T base pairs might influence global helical parameters during the crystallization 
process that facilitates assembly. This may also explain the larger number of A5-T8 
sequences that crystallized in the fast group of the initial screen. 
 Having identified sequences with isomorphous structures also provides new 
tools for understanding the crystal self-assembly process, particularly the nucleation 
  43 
 
of these DNA oligonucleotides. Our mixed sequence crystallization results supported 
a model of nucleation through formation of Watson-Crick duplexes, followed by 
crystal growth through non-canonical pairs. Further, we established through NMR 
that the 13-mer oligonucleotide sequence does not form base pairs in solution without 
divalent cations, and that the addition of a divalent cation leads to the apparent rapid 
assembly into higher molecular weight complexes. Thus, these oligonucleotides may 
provide a unique system for further understanding how a dispersed biopolymer can 
self-assemble into macroscopic objects through a simple chemical trigger. 
  
 
  44 
 
Chapter 2.8 Materials and Methods 
1. Crystallization Screen 
 The 64 oligonucleotides representing all self-complementary sequences in the 
duplex region of the 13 nucleotide DNA oligomer were purchased from Integrated 
DNA Technologies (Coralville, IA), on the 100 nmol scale. Unpurified, desalted 
oligonucleotides were dissolved in water to 350 µM and screened for crystallization 
by sitting drop vapor diffusion. Crystallization drops were created using a Crystal 
Phoenix crystallization robot with 3-well Intelliplate crystallization trays (Art 
Robbins Instruments; Sunnyvale, CA). Each oligonucleotide was screened at three 
different ratios (1:1, 2:1, 3:1) of DNA to crystallization buffer (120 mM magnesium 
formate, 50 mM lithium chloride, 10% 2-methyl-2,4-pentanediol) to final drop 
volumes of 1 µl, 1.5 µl and 2 µl, respectively. The reservoir contained 100 µl of 
crystallization buffer. Trays were incubated at 22˚C. 
2. DNA Purification and Crystallization 
 The 20 molecules identified from the screen and the A7C variant were were 
purchased from Integrated DNA Technologies (Coralville, IA) on the 1 µmol scale 
and were purified by 20% (19:1) polyacrylamide gel electrophoresis, electroeluted, 
and ethanol precipitated as previously described
59
. The purified DNA samples were 
dialyzed against deionized water and the concentration was adjusted to 260 µM. The 
crystals were grown by sitting drop vapor diffusion in 24 well Cryschem M plate 
(Hampton Research, Aliso Viejo, CA).   Prior to crystallization, DNA samples, in 
water, were heated at 95˚ for 2 minutes and cooled to room temperature. DNAs were 
  45 
 
mixed (1:1) with crystallization buffer (120 mM magnesium formate, 50 mM lithium 
chloride, 10% 2-methyl-2,4-pentanediol) in a 4 µl drop. The reservoir contained 400 
µl of crystallization buffer. The crystal plates were incubated at 22˚C.  
3. Data Collection and Refinement 
 Crystals were harvested by nylon loop, washed sequentially in crystallization 
buffer containing 30% and 40% 2-methyl-2,4-pentanediol, and flash-cooled in liquid 
nitrogen. Data were collected at Advanced Photon Source, Argonne National Labs, 
Sector 24-ID-C with an X-ray wavelength 0.979200 Å. Data were indexed and 
integrated using XDS
74
, and scaled using Aimless
75
. The isomorphous cell and space 
group for all datasets allowed direct refinement using the original 13-mer structure 
(PDBID: 1P1Y) as a starting model with individual nucleotide identities modified 
using Coot
76
. Prior to refinement all atomic B-factors were reset and all sugar atoms 
were randomized to avoid biasing puckers. Refinement was performed using 
Phenix
77
. Water molecules and ions were added manually during refinement.   
Following converged refinement in Phenix, all of the structures were run through the 
PDB-REDO
78
 pipeline with single-group TLS (Translation/Liberation/Screw) and 10-
fold cross-validation due to the small number of reflections in these datasets. 
Coordinates and structure factors were deposited in the Protein Data Bank
79
 with 
accession codes provided in Table S2. 
4. Sequence and Structure Comparison and Analysis 
 Depending on the time required for the appearance of crystals, the 21 
crystallizing DNA oligonucleotides were classified into fast or slow crystallizing 
  46 
 
groups, and their sequences were compared to calculate the frequency of nucleobases 
at each position of the duplex. For structural analysis, the variant structures were 
grouped depending on the presence or absence of the A5-T8 base pair. Structures 
were initially superimposed using the invariant residues (1-3, 10-12) to account for 
the slight differences in unit cell parameters. PyMol
80
 was used to calculate RMSD 
values for backbone atoms (C1', C2', C3', O3', C4', O4', O5', P, O1P, and O2P). 
5. Nuclear Magnetic Resonance 
 NMR spectra were recorded on a 600 MHz Bruker Advance III NMR 
spectrometer equipped with a CPTCI cryoprobe at a sample temperature of 298K. 
The purified B7 sample (200 µM) was used in 500 µL water containing 10% D
2
O. 2 
M MgCl
2
 stock solution was added to NMR sample stepwise to reach target Mg
2+
 
concentrations. At each titration point, a 1-D proton spectrum was acquired to observe 
signal chemical shift changes and/or signal intensity changes. The pulse program 
zgesgp was used to suppress water signal using gradient excitation sculpting
81
.  
 NMR transverse relaxation measurement were performed using the pulse 
program cpmgpr1d
82
.  Each experiment consisted of a series of 1D measurement with 
at least 7 different decay times.  Peak intensities plotted against decay time decreased 
exponentially and were fit to the equation: 
   
2
/
0
Tt
xyxy
eM=tM

to determine values 
for T
2
. A deoxyribose H2’ signal at 1.7 ppm was chosen for peak intensity 
measurements because it was separated from other signals and had high intensity. 
MgCl
2
 was added to the 0.2 mM B7 sample step wise, and at each addition point, a T
2
 
experiment was performed.  
  47 
 
6. Heterogeneous Crystallization 
 For heterogeneous crystallization, two different DNAs were mixed and heated 
at 95˚ for 2', cooled to room temperature, and then mixed (1:1) with crystallization 
buffer in a 4 µl drop and incubated at 22˚C. For polyacrylamide gel analysis, single 
heterogeneous crystals were washed in crystallization buffer three times and 
dissolved in water. Dissolved products were labeled with T4 polynucleotide kinase 
(New England Biolabs, Ipswich, MA) using 5 µCi/crystal 
32
P ɣ-ATP (3000 Ci/mM; 
Perkin Elmer, Waltham, MA). The reaction was carried out at 37˚ for 30' and then 
terminated at 65˚. 2 µl samples were run on a 20% acrylamide (19:1). The gels were 
exposed to the phosphor screen overnight and imaged using a Storm 860 
phosphorimager (Molecular Dynamics, Sunnyvale, CA). 
 
  48 
 
Chapter 3: Other Methods Used to Study the Cation Driven 
Assembly Process 
 
 I employed various direct and indirect methods to study the formation of 
higher order assemblies for the 13-mer DNA and to gather clues about self-assembly 
process. However, most of the used methods were only partially successful largely 
because of the high salt concentrations required to bring about the self-assembly 
process. We reasoned that the crystal pre-nucleation assemblies would be marked by 
the formation of stable intermediate species which could be distinguished from 
dispersed single-stranded DNA oligonucleotides based on their size, differential UV 
absorbance and specific dye binding capabilities. Dynamic light scattering is an 
effective method for fast and accurate size measurements of DNA nanostructures. As 
the presence of cations induces the 13-mer assembly, I wanted to monitor the 
formation of pre-nucleation complexes by measuring the hydrodynamic radii of the 
particles formed as a function of Mg
2+ 
concentration. The challenge to this approach 
was to get an accurate size distribution of dispersed 13-mer DNA oligonucleotides in 
solution to compare with the particle sizes observed in presence of cations. 
Unfortunately, I was not able to get a consistent particle size distribution for the 13-
mer in water probably owing to the low scattering from these short oligonucleotides 
or due to non-specific aggregation.  
 The differential absorbance of UV by single and double-stranded DNA has 
been traditionally used to monitor the assembly of DNA. The formation of double-
stranded DNA duplex is associated with a hypochromic shift in the UV absorption 
due to the extensive base stacking interactions. I recorded the absorbance of the 13-
  49 
 
mer at 260 nm as a function of Mg
2+
 concentration. I observed a decrease in the 
absorbance with an increase in cation concentration but the results were unreliable 
due to an interplay of sample evaporation, aggregation and sample heating in the 
nanodrop during the course of observations. As a way to monitor the assembly 
process for the 13-mer crystals, I also explored the use of double-strand specific 
fluorescent dye, SYBR Green I. This method has been successfully used in 
quantitative PCR techniques to monitor the copy number of the amplified product. I 
established a baseline fluorescence using single-stranded 13-mer and then measured 
the fluorescence as a function of Mg
2+
 concentration. The biggest challenge to the use 
of SYBR Green I was the quenching of its fluorescence at the Mg
2+ 
concentration 
required for promoting the DNA self-assembly.  
 Additionally, I probed the assembly process of the 13-mer DNA using indirect 
methods like S1 and DNase I endonuclease digestion. S1 nuclease is a single-strand 
specific endonuclease whereas DNase I does not show such specificity. I anticipated 
to see a protection against these endonucleases upon formation of higher order 
assemblies promoted by the addition of cations to the DNA sample. However, the 
narrow range of Mg
2+
 concentrations over which the endonucleases were active made 
it difficult for me to accurately and reproducibly draw clues about the self-assembly 
process. Finally I also tried to use 3’ fluorescence labelled B7 sample in conjunction 
with an unlabeled sample to monitor the assembly process under a fluorescence 
microscope. This work has shown promising results and is being explored by other 
members in our lab. 
  50 
 
Chapter 3.1 Dynamic Light Scattering 
 To study the particle sizes in the pre-nucleation assembly process, as a 
function of Mg
2+ 
concentration, I recorded Dynamic Light Scattering (DLS) 
measurements on B7 in water and at 50 and 100 mM Mg
2+
 concentrations. The 
samples were preheated at 37ºC for 2 minutes to get rid of non-specific hydrogen 
bond interactions. As a control, particle sizes were recorded in presence of 100 mM 
K
+
 as it does not promote any crystal assembly. Similar readings were recorded on 
the original 13-mer (1P1Y), A7 sample which failed to crystallize in presence of 
Mg
2+ 
(Table 2) and on EET11 sample (5’ CAATCAGTCAG 3’) which does not form 
any potential self-complementary dimers in presence of Mg
2+
 ions.   
 The B7 sample in water showed multiple peaks with average particle sizes of 
2.3, 213 and 5416 nm. The observed sizes were larger than expected for a dispersed 
13 nucleotide monomer. The readings for the sample upon addition of 50 mM Mg
2+ 
and 100 mM Mg
2+
 were inconsistent in the number of observed peaks as well as the 
particle sizes. Similar inconsistencies were observed in the other three samples. As 
there were no clear differences in the sample and control particle sizes for the DLS 
experiment I concluded that it was not a useful technique to measure the self-
assembly process of these 13-mer crystals. The inconsistent data could be from the 
low scattering from these short DNA molecules or due to the presence of non-specific 
aggregation.  
  51 
 
Chapter 3.2 Ultra-Violet (UV) Spectroscopy 
 The most common real time measurement strategy to study the assembly 
process of complementary DNA strands include monitoring changes in UV 
absorbance based on the hypochromic effect of DNA. This phenomenon is attributed 
to the formation of extensive stacking interactions between the bases upon formation 
of the helical structure. To monitor the formation of higher order assemblies for the 
13-mer DNA oligonucleotides, I measured the hypochromic shift in the absorption at 
260 nm. I performed a time course experiment and recorded the absorption of B7 
oligonucleotide as a function of Mg
2+
 concentration (Figure 17). B7 in storage 
conditions (absence of Mg
2+
) was used as a negative control. The observation showed 
a significant drop in the absorption at 260 nm at the 3 hour time point for B7 samples 
in 100 mM and 200 mM Mg
2+
 with no further drop for the subsequent time points. 
This effect was more gradual for the B7 sample in 50 mM and 25 mM Mg
2+
. The B7 
samples in 5 mM and 10 mM Mg
2+
 showed a slight drop in the absorption 
immediately after mixing, but did not show any further change in the absorption in 
the course of the experiment. All the samples were measured again after 18 hours 
(overnight) with and without the application of heat and I observed an increase in the 
absorption for samples in Mg
2+ 
concentration ranging from 25 mM – 200 mM upon 
application of heat. This result suggested that I could monitor association between the 
DNA molecules as a function of time and Mg
2+
. This effect was most significant at 
higher Mg
2+
 concentrations (50 mM - 200 mM) and was reversed upon application of 
heat which would cause disruption of the DNA assemblies.   
  52 
 
 
Figure 17. Monitoring the assembly of B7 DNA using UV Spectroscopy.  250 µM 
B7 DNA sample was mixed with various concentrations of MgCl
2
 (Colored lines) and 
the absorption intensity is measured at 260 nm as a function of time. After the 
overnight time point, the samples were heated at 95º for 2 minutes and then the 
absorption was measured again. 
 
  53 
 
Chapter 3.3 Fluorescence Spectroscopy 
 One of the ways to monitor the assembly process of DNA oligonucleotides is 
to record the changes in emission signal from fluorescent dyes that are able to 
distinguish between dispersed single-stranded vs. double-stranded DNA molecules. I 
monitored the self-assembly process of B7 using SYBR Green I, a double-stranded 
DNA specific dye. It has an excitation maximum at 450 nm and an emission 
maximum at 520 nm. This dye is commonly used to monitor the copy number of the 
product in the quantitative polymerase chain reactions (qPCR). First, to determine the 
effect of SYBR Green I on the growth of B7 crystals, I performed post crystallization 
soaking along with co-crystallization and confirmed that SYBR Green I did not 
interfere in the self-assembly process for B7 and also did not have any visible effect 
on the crystal stability and geometry. I analyzed the assembly process by monitoring 
the fluorescence intensity as a function of Mg
2+
 concentration (Figure 18A). I used 
B7 oligonucleotide in absence of Mg
2+
 to establish a baseline and subsequently added 
increasing amounts of Mg
2+
 anticipating an increase in the fluorescence intensity with 
the formation of higher order assemblies. I used K
+
 ions (results not shown) and 
SYBR Gold with B7 oligonucleotides as negative controls (Figure 18B).  
 I observed an increase in the fluorescence intensity upon increasing the Mg
2+ 
concentration from 1 mM, to 2.5 mM and then to 5 mM.  However, a drop in the 
fluorescence intensity was seen as the Mg
2+
 concentration was further increased to 10 
mM, then to 100 mM and then again to 200 mM. Furthermore, through experiments 
conducted to measure the effect of Mg
2+
 on the SYBR green I fluorescence, it was 
found that the SYBR green I was active only over a small range of magnesium 
  54 
 
concentration (optimum fluorescence at 1-5 mM Mg
2+
), and higher Mg
2+
 
concentrations were inversely related to the fluorescent intensity
83 
(http://www.biofiredx.com/pdfs/LightCycler/LC_Exp_Design.pdf). To confirm this I 
used a self-complementary 13 nucleotide DNA as a control and performed 
fluorescence reading as a function of Mg
2+
 and observed the expected quenching of 
fluorescence intensity at 100-200 mM Mg
2+ 
concentrations (data not shown here). So, 
I concluded that SYBR green I fluorescence spectroscopy was not a reliable method 
for monitoring the 13-mer DNA assembly process.  
  55 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 18. Fluorescence intensity measurements on B7 DNA using nucleic acid 
specific fluorescent dyes. Fluorescent intensity measurements were carried out as a 
function of time (reading every 3s up to 200s) on B7 DNA sample in presence of 
SYBR Green I, a double-strand specific DNA dye (A) or SYBR Gold (B) which 
binds to both single and double-stranded DNA. The time course experiment was 
carried at various concentrations of MgCl
2
 
(Colored lines). 
B. 
A. 
  56 
 
Chapter 3.4 S1 Nuclease Digestion 
 S1 Nuclease is a single-strand specific endonuclease that hydrolyses the 
single-stranded regions in DNA duplexes or assemblies
84
. As increasing the Mg
2+
 
concentrations leads the free B7 oligonucleotides to assemble into duplexes and 
higher order assemblies, we believed that the process could be marked by an 
increased protection from S1 digestion which could then be monitored through gel 
electrophoresis. The heterogenous crystal experiments, described in Chapter 2.6, 
strongly suggest that the pre-nucleation event for the 13-mer crystals is driven by the 
formation of stable Watson-Crick duplexes, which would mean the S1 treatment 
would lead to the hydrolysis of the single-stranded non-canonical overhangs from one 
or both ends of the 13-mer. We anticipated that this would limit the assembly 
formation and propagation. For the S1 treatment, I incubated the oligonucleotide 
sample with Mg
2+
 for 4 hours and the samples were then digested with S1 nuclease. 
The samples were analyzed on a denaturing urea gel (Figure 19A). I performed 
similar experiment using EET (5’ CAATCAGTCAG 3’) as a control DNA 
oligonucleotide as it was unlikely to form any duplexes due to lack of any self-
complementary sequence (Figure 19B). 
 The results showed that the B7 oligonucleotide was protected from digestion 
by S1 nuclease as the concentrations of Mg
2+
 increased. However, the band intensities 
were inconsistent at higher Mg
2+
 concentrations. The EET control showed complete 
S1 digestion at 0 mM and 100 mM, but small amount of protection was observed in 
10 mM Mg
2+
 concentration. Due to these inconsistencies I did not use S1 nuclease for 
studying the assembly process of B7 oligonucleotides.
  57 
 
 
 
  
 
 
 
 
 
Figure 19. Denaturing PAGE gel for samples digested with S1 nuclease. The B7 
sample at 300 µM (A.) or EET sample at 413 µM (B.) were incubated at various 
concentrations of MgCl
2
 for four hours and digested with S1 at nuclease at room 
temperature for 10 minutes and visualized on a denaturing PAGE gel stained with 
SYBR Gold. 
1. B7 + no enzyme + No 
Mg
2+
 
2. B7 + No Mg
2+
 
3. B7 + 5 mM Mg
2+
 
4. B7 + 10 mM Mg
2+
 
5. B7 + 25 mM Mg
2+
 
6. B7 + 50 mM Mg
2+
 
7. B7 + 100 mM Mg
2+
 
1.  EET11 + no enzyme + No 
Mg
2+ 
2.  EET11 + No Mg
2+ 
3.  EET11 + 10 mM Mg
2+ 
4.  EET11 + 100 mM Mg
2+ 
A. 
B. 
  58 
 
Chapter 3.5 DNase I Digestion 
 DNase I is an endonuclease that cleaves both single and double-stranded DNA 
oligonucleotides. I used DNase I to monitor the self-assembly process of B7 
oligonucleotides, as a function of Mg
2+
 concentration. The formation of higher order 
assemblies for the DNA 13-mer in presence of Mg
2+
, should theoretically make it 
inaccessible for digestion by DNase I enzyme which could then be resolved on a 
PAGE gel. We incubated the B7 oligonucleotide sample with water or various 
concentrations of Mg
2+
 for 4 hours and then the samples were digested with the 
nuclease. The samples were analyzed on a denaturing urea gel (Figure 20A). I 
performed similar experiment using T7 primer (5’-TAATACGACTCACTATA-3’) as 
a control DNA oligonucleotide as it did not have any self-complementary sequence 
(Figure 20B, C). 
 The results showed that B7 oligonucleotide was protected from digestion by 
DNase I as the concentrations of Mg
2+
 increased, suggesting the formation of higher 
order assemblies.  However, these results were not reproducible. Also, the control T7 
DNA showed no digestion under identical reaction conditions and so we concluded 
that the DNase I experiments were not optimum for characterization of DNA 
assemblies. Reports show that DNase I is active over a relatively narrow range of 
Mg
2+
 concentrations and so the discrepancies in the results could be a result of partial 
or complete deactivation of the enzyme at high Mg
2+ 
concentrations used in our 
reaction conditions
85
. 
  59 
 
B. 
C.  
7 
 
 
 
 
 
 
 
 
Figure 20. Denaturing PAGE gel for samples digested with DNase I nuclease. A. The B7 
sample at 300 µM was incubated at various concentrations of MgCl
2
 for 4 hours and digested 
with DNase I nuclease for 10 minutes and visualized on a denaturing PAGE gel. B. The T7 
sample was incubated with either 0 or 100 mM MgCl
2 
for 1.5 hours and then digested at room 
temperature for up to 4 hours. Samples corresponding to 30 mins, 1 hour and 4 hour digestion 
time points are loaded onto the gel. C. The B7 sample at 300 µM was incubated at various 
concentrations of MgCl
2
 for 2 hours and digested with DNase I nuclease for 4 hours at room 
temperature and visualized on a denaturing PAGE gel. The gels were stained with SYBR 
Gold.
  
1. B7 + No enzyme + No Mg
2+
 
 
2. B7 + No Mg
2+
 
3. B7 + 5 mM Mg
2+
 
4. B7 + 10 mM Mg
2+
 
5. B7 + 25 mM Mg
2+
 
6. B7 + 50 mM Mg
2+
 
7. B7 + 100 mM Mg
2+
 
8. B7 + 200 mM Mg
2+
 
1. T7 + No enzyme + No Mg
2+
 
 
2. T7 + No Mg
2+
- 30 min 
3. T7 + 100 mM Mg
2+
- 30 min 
4. T7 + No Mg
2+
- 1 hour 
5. T7 + 100 mM Mg
2+
- 1hour 
6. T7 + No Mg
2+
- 4 hour 
7. T7 + 100mM Mg
2+
- 4 hour 
1. B7 + No enzyme + No Mg
2+
 
 
2. B7 + No Mg
2+
 
3. B7 + 5 mM Mg
2+
 
4. B7 + 10 mM Mg
2+
 
5. B7 + 25 mM Mg
2+
 
6. B7 + 50 mM Mg
2+
 
7. B7 + 100 mM Mg
2+
 
 
A. 
  60 
 
Chapter 3.6 Confocal Microscopy 
 I used a 3’ fluorescein (λ
exct
 = 494 nm and λ
emiss 
= 521 nm) labelled B7 to 
monitor the assembly process of the 13-mer crystals. It has been previously 
determined in our lab that a mixture of fluorescein labelled and an unlabeled B7 (1:9) 
can be used to grow a shell on a pre-grown unlabeled B7 crystal (core). We reasoned 
that we can use this mixture on the S1 nuclease treated or the DNase I treated crystals 
and gather clues about the assembly process. We also wanted to demonstrate the 
growth of fluorescently labelled shell around unlabeled microcrystals. We believed 
that this would give us a visual aid to observe the assembly process and help us 
understand the time scale at which the assembly took place. 
 I observed that both the S1 and the DNase I treated crystals were able to form 
a fluorescent shell within 2 hours when incubated in 120 mM Mg
2+
 (Figure 21A, B). 
No visual difference was observed in the rate of growth of the shell or its thickness 
between the untreated and the enzyme treated crystals (Figure 21C). In the seeding 
experiment, I observed appearance of the fluorescent shell around the non-fluorescent 
core within 2 hours (Figure 22A, C) and the appearance of fluorescent crystals in 
absence of the seed was seen after 5 hours of incubation (Figure 22B). However 
taking into account the previous S1/DNase I digestion experiments, these 
observations are not sufficient to unambiguously prove that they are a true 
representation of the assembly process and not the inhibition of the enzyme activity in 
high Mg
2+
 concentrations and so these experiments need to be explored further for 
studying the assembly process of 13-mer DNA crystals. 
  61 
 
  
  
 
Figure 21. Confocal images for unlabeled B7 core and fluorescent shells. The 
unlabeled B7 crystals were grown to a standard size and then treated with either S1 
nuclease (A) or DNase I (B) or no enzyme (C) for 1 hour at room temperature and 
then transferred to a mixture of unlabeled B7: fluorescein labelled B7 (9:1). After 
washing in buffer solution the crystals were imaged under a confocal fluorescence 
microscope. The images in panel A, B and C are the fluorescent image and the 
overlay of the fluorescent image with the bright field image.  
B. 
C. 
A. 
  62 
 
 
 
 
 
Figure 22. Confocal images to observe the B7 crystal assembly. A mixture of 
unlabeled B7: fluorescein labelled B7 (9:1) was used to observe the assembly process 
under confocal microscope. The assembly process was monitored in presence (A, B) 
or absence (C) of unlabeled B7 microcrystals as seeds for nucleation. Images were 
taken at various time points. In the presence of the microcrystal seeds the fluorescent 
shells started appearing as early as 2 hours (A) and were completely developed in the 
overnight sample (B) where as in the absence of the micro crystal seeds, the 
fluorescent assembly was visible at a 5 hour time point (C). The three panels in (B) 
indicate the fluorescence image, a bright-field image and an overlay of the two 
images going left to right. 
A. 
B. 
C. 
  63 
 
Chapter 3.7 Materials and Methods 
1. Dynamic Light Scattering 
 Unpurified B7, DET66, A7 and EET11 samples were dialyzed against 
deionized water over night and filtered through a 0.2 µm filter. 70 µl of samples were 
heated at 37º for 2 minutes and mixed with 7 µl of either deionized water, 500 mM or 
1 M Magnesium formate or 1 M Potassium chloride to get a final concentration of 50 
mM or 100 mM Mg
2+ 
or K
+
 ions. The samples were then read in Zetasizer Ver. 6.20 
(Malvern Instruments Ltd) in Dr. Silvia Muro’s lab. The particle diameters were 
recorded three times for each sample maintained at 25º, with a material refractive 
index of 1.45 and 3s recording time. 
2. UV Spectroscopy 
 Reactions were set up using 5 µl of 500 µM B7 DNA oligonucleotide with 5µl 
of either water or MgCl
2 
(10, 20, 50, 100, 200 and 400 mM). Samples were mixed 
and 1 µl of sample was used to measure the absorption at 260 nm using a 
NanoPhotometer (Denville Scientific Inc.) at 0, 3, 6, 9, 12 and 18 hour time points. 
Readings were also recorded by heating the sample at 95º for 1 minute after the 18 
hour time point. 
3. Fluorescence Spectroscopy 
 To monitor self-assembly process of 13-mer oligonucleotides using 
fluorescent spectroscopy, 90 µl of 100 µM B7 oligonucleotide sample in 1 mM EDTA 
was mixed with either 10 µl of water or 10X Magnesium formate solution (10, 25, 50, 
100, 1000 mM) to get final concentrations of 0, 1, 2.5, 5, 10, 100 and 200 mM Mg
2+
 
  64 
 
ions respectively. The stock solution of SYBR Green I, a double-stranded DNA 
specific dye, was diluted 20X using DMSO and 0.1 µl of the diluted solution was 
added to the DNA-Mg
2+
 mixture to get a 20000X dilution. Upon mixing the 
fluorescence intensity (λ
excit
 = 450 nm, λ
emis
 = 520 nm) was recorded every 3s for a 
total of 200s. Similar readings were obtained for a non-specific DNA dye, SYBR 
Gold (20000X dilution), and using 10 µl of  100 or 1000 mM K
+
, instead of Mg
2+
, 
with SYBR Green I (20000X dilution). To determine the effect of Mg
2+ 
on the 
fluorescence of SYBR Green I, fluorescent intensity studies were performed on a 100 
µM 13 nucleotide self-complementary duplex by mixing it with 10 µl of 100 or 1000 
mM Mg
2+
 and SYBR Green (20000X dilution). 
4. S1 and DNase I Digestion 
 For S1 and DNase I digestion experiments, 1.5 µl of 300 µM B7 
oligonucleotide sample was incubated with 1.5 µl of either water or Mg
2+
 (10, 20, 50, 
100, 200 mM) for 4 hours. The reaction mix was then digested with either S1 
nuclease or DNase I (Thermo Scientific) in their respective reaction buffers (S1 1X 
reaction buffer: 40 mM Sodium acetate, pH 4.5, 1.5 M Sodium chloride and 10 mM 
Zinc sulphate) (DNase I 1X Reaction Buffer: 10 mM Tris-HCl, 2.5 mM Magnesium 
chloride, 0.5 mM Calcium chloride
,
 pH 7.6) for 10 minutes at room temperature. The 
reaction was quenched using 5 µl of stop buffer (200 mM EDTA and 0.2% SDS) at 
65º for 10 minutes for S1 and at room temperature for DNase I. 12 µl of the reaction 
mix was analyzed on a 20% polyacrylamide urea gel and stained using SYBR Gold 
(1:10000X dilution) in TBE buffer.  
  65 
 
5. Confocal Microscopy 
 B7 crystals were grown in crystallization buffer (120 mM Magnesium 
formate, 50 mM Lithium chloride and 10% MPD) and were washed 2X in fresh 
buffer and transferred to a reaction mixture containing either 10 µl buffer + no 
enzyme or 8 µl buffer + 2 µl S1 reaction buffer + 0.2 µl S1 nuclease or 8 µl buffer + 1 
µl DNase I reaction buffer + 1 µl DNase I enzyme. The samples were incubated at 
room temperature for 1 hour and transferred to a fresh mixture of 2 µl buffer + 2 µl 
B7: fluorescein labelled B7 mixture (9:1) following a 2X washing step in buffer. The 
crystals were incubated for 2 hours and then imaged using the Leica SP5X confocal 
microscope in the imaging core facility of the cell biology and genetics department of 
University of Maryland. 
 
 
  66 
 
Chapter 4: Sequence-Dependent Structural Changes in 14-mer 
DNA Oligonucleotide 
Chapter 4.1 Introduction 
 DNA oligonucleotides are both conformationally and structurally diverse
33
. 
Depending on the environmental conditions, B-form DNA can undergo 
conformational transitions to the A- and the Z-forms
33–36
. Additionally, a variety of 
non-B-form DNA motifs have been characterized in vivo including DNA cruciform, 
hairpin structures, triplexes and quadruplexes
37–43
. One of the major areas of DNA 
structural biology over the course of several decades was in understanding how non-
Watson-Crick base pairs, or mismatches, could be accommodated in otherwise 
normal DNA helices, or are responsible for forming alternate DNA structures
52,53,86
. 
 G-A base pairs have been one of the most well-characterized non-Watson-
Crick base pairings that can be readily integrated into the B-form duplex
66,87–94
. 
Structural studies revealed that these G-A base pairs can adopt up to four different 
base pairing combinations depending on the local sequence and environment
95
. The 
two most prevalent types represented in the Nucleic Acid Databank
96
 include the 
Type I pair involving the Watson-Crick edges of the bases, and the Type IV sheared 
G-A pair involving the guanosine sugar edge and the adenosine Hoogsteen edge
65,94
. 
However, the type and stability of G-A base pair formed is highly dependent on the 
local sequence
97
. The Type I pairing is favored for d(AGAT)
2
 sequence due to an 
additional interstrand hydrogen bond between the N2 amino group of the paired G 
and O2 of the thymidine in the flanking A-T pair
66
. The sheared G-A base pairs are 
favored in d(YGAR)
2
 sequences. In nearly all cases the sheared G-A pairs are found 
  67 
 
in tandem (GA/AG) and are thermodynamically quite stable within a canonical 
duplex due to the interstrand stacking between the sheared base pairs and the 
extensive intrastrand stacking between the sheared pairs and the flanking base pairs
95
. 
 The 14-mer DNA oligonucleotide structures described in this section were 
obtained by adding an additional adenosine at the 3’ of the DNA 13-mers described in 
Table 2. The 14
th
 nucleotide was added for the easy resolution of the heterogenous 
crystals on a denaturing gel. Out of all the 14-mer DNA’s screened for crystallization, 
the addition of the 14
th
 nucleotide lead to a significantly different crystal habit under 
identical Mg
2+
 conditions for four DNA oligonucleotides (Figure 23B-E). 
Remarkably, the added A14 residue made tertiary contacts to the guanosine adjacent 
to a single sheared A-G pair from the neighbouring duplex, resulting in a 
conformation similar to tandem sheared G-A pairs. Together with a series of purine 
base triples, these interactions were responsible for the formation of the alternate 
crystal form. Here, I also examined the role sequence played in the adoption of the 
alternate crystal form. Our analysis suggested that the sequence in the 14-mer duplex 
region and the identity of the added nucleotide were necessary to promote this 
alternate structure. 
  68 
 
 
Figure 23. 13-mer and 14-mer crystals. A. The 13-mer DNAs crystallize with a 
hexagonal unipyrimidal crystal habit.  Under identical crystallization conditions the 
four 14-mer DNAs, A1-14, A2-14, A3-14 and A4-14 crystallize with habits shown in 
B-E, respectively. All of the 14-mers crystallized in the same space group with 
almost identical unit cell dimensions. 
 
  69 
 
Chapter 4.2 Overview of the 14-mer Structures. 
 We determined the X-ray crystal structures of four 14-mer DNA 
oligonucleotides differing by one base pair in the self-complementary duplex region 
(Figure 24A). The structures were highly isomorphous to each other with an average 
RMSD of 0.40 Å for all identical aligned atoms, and 0.58 Å for the backbone atoms. 
All of DNAs crystallized with one molecule in the asymmetric unit, with crystal 
symmetry generating interstrand hydrogen bonding and base stacking interactions. 
Each strand in the crystal formed hydrogen bonds with 5 other strands to form two 
distinct regions of nucleobase interactions (Figure 24A, B). The B-form duplex 
region is formed from residues A3 through G10 of two strand, and the triplex junction 
is formed from G1 and G2 of one strand, G11-G13 of two different strands, and A14 
of another strand. End-to-end stacking of the triplex junction regions lead to columns 
of pseudo-infinite coaxially stacked helices that interact only through the A14 
residues (Figure 25). For convenience I will restrict the structural description to the 
A3-14 structure, noting that the only substantial differences in the other structures 
were the central base pairs of the duplex region (Table 7 & Figure 26). 
  70 
 
 
Figure 24. Overview of 14-mer crystal structures. A. Secondary structure of 14-
mer crystals. The A3-14 sequence is diagrammed and sequence differences of the 
other oligonucleotides are shown. Each DNA 14-mer is hydrogen bonded to five 
identical molecules related by crystallographic symmetry indicated by different 
colors. Interactions between DNA molecules lead to the formation of two distinct 
regions of base pairing. The duplex region is formed from residues A3-G10 of partner 
strands, and the triplex junction is formed by residues G1-G2 of one duplex (black-
red) and the G11-G13 of the coaxially stacked duplex (green-blue) and the A14 
residue of a neighboring duplex (magenta). The 5’ nucleotide of each strand is 
denoted by bold. B. The overall 3D arrangement of the 14-mers is shown in A.
  71 
 
 
Figure 25. Crystal packing. A. The overall 3D arrangement of the 14-mer in the 
crystal lattice looking perpendicular to the three-fold symmetry axis, and B. looking 
down the three-fold symmetry axis, . Duplex regions are coaxially stacked, with 
adjacent duplexes interacting only through tertiary contacts via the added A14 
nucleotide. 
 
 
 
 
 
 
  72 
 
Table 6. Data collection and refinement for 14-mer structures. 
 
A1-14 A2-14 A3-14 A4-14 A3-14-Br 
Data collection 
     
Wavelength (Å) 0.97919 0.97919 0.97919 0.97919 0.91940 
Detector 
ADSC Quantum 
315 
ADSC Quantum 
315 
ADSC Quantum 
315 
ADSC Quantum 
315 
Pilatus 6M 
Space group P3
1
21 P3
1
21 P3
1
21 P3
1
21 P3
1
21 
Number of crystals 4 4 3 2 1 
Cell dimensions      
Avg. a, b, c (Å) 
26.01   26.01   
122.02 
25.99   25.99   
121.53 
25.83   25.83   
123.08 
25.96   25.96 
 121.53   
26.27   26.27      
123.30 
α, β, γ (º) 90, 90, 120 90, 90, 120 90, 90, 120 90, 90, 120 90, 90, 120 
Resolution 
2.03-22.52 (2.03-
2.09) 
2.10-22.51 (2.10-
2.17) 
2.15-41.02 (2.15-
2.23) 
2.40-22.48 (2.40-
2.53) 
1.99-122.34 (1.99-
2.09) 
I/σI 13.9 (1.0) 16.3 (3.0) 14.7 (2.5) 8.9 (2.6) 14.8 (1.0) 
CC
1/2
 0.997(0.883) 0.987(0.978) 0.978(0.959) 0.993(0.954) 0.999(0.866) 
R
pim
 0.036 (0.45) 0.047 (0.12) 0.061 (0.15) 0.055 (0.13) 0.023 (0.38) 
Number of reflections 3436 (268) 3035 (291) 2965 (281) 2131 (299) 3818 ( 507) 
Completeness (%) 99.4 (97.3) 97.8 (97.4) 99.9 (99.9) 99.1 (98.7) 99.1( 96.0) 
Multiplicity 10.3 (6.9) 10.6 (4.6) 6.8 (4.5) 3.8 (3.5) 5.0 (3.8) 
Anomalous 
Completeness (%) 
N/A N/A N/A N/A 95.2 (77.9) 
Anomalous multiplicity N/A N/A N/A N/A 2.7 (1.8) 
Refinement 
     
Resolution (Å) 
2.03-22.52 (2.03-
2.07) 
2.10-22.51 (2.10-
2.15) 
2.15-41.02 (2.15-
2.20) 
2.40-22.48 (2.40-
2.46) 
 
Number of reflections 3092 (206) 2716 (202) 2626 (178) 1898 (117)  
Average R
free 
a
 0.2937 0.313 0.265 0.312  
R factor 
b
 0.233 (0.502) 0.270 (0.454) 0.244 (0.421) 0.245 (0.523)  
R
free 
b
 0.293 (0.538) 0.311 (0.416) 0.262 (0.606) 0.312 (0.641)  
Number of atoms      
DNA 293 293 293 293  
Ion 2 2 2 2  
Water 13 13 13 6  
Bond lengths (Å) 0.007 0.006 0.008 0.005  
Bond angles (˚) 1.567 1.406 1.887 1.220  
PDB ID 5BZ7 5BZ9 5BXW 5BZY  
  *Values in parentheses are for the highest-resolution shell. 
 
  73 
 
 
Figure 26. Structural overlap for the 14-mer structures. The three 14-mer 
structures, A1-14 (yellow), A2-14 (orange) and A4-14 (pink), are superposed with 
A3-14 structure (blue). The RMSD for the overlap is 0.40 Å. The weighted electron 
density map (2F
o
-F
c
) of A3-14 contoured at 1.0σ is shown in grey. 
  74 
 
Chapter 4.3 B-Form Duplex Region Capped by Sheared A-G Pairs. 
 The B-form duplex region was formed through base pairing interactions 
between A3 and G10 of two DNA strands. The central six base pairs of the helix were 
composed of self-complementary base pairs A4-T9, A5-T8 and G6-C7. These six 
nucleotides were structurally isomorphous to the duplex region in the parent A3-13 
structure with a RMSD of 0.69 Å (Figure 27). The sugar-phosphate backbone of 
residue A4 showed the greatest variability between the two structures (RMSD 3.40 Å 
for backbone atoms) and was the result of a significantly different conformations 5' of 
the A4 nucleotide. In all of the 13-mer structures we determined, residues G1-A3 
were flipped out of the helical axis toward the major groove of the duplex where they 
were positioned to make non-canonical interactions with G10-A12 of another strand. 
In the 14-mer structures, A3 remained stacked with A4 and was base paired with G10 
in a Type IV sheared base pair. This resulted in a duplex region containing six self-
complementary base pairs flanked on either end by A-G pairs. 
  75 
 
 
Figure 27. Structural comparison of A3-13 and A3-14 monomers. Stereo view of 
the X-ray crystal structures of A3-13 (orange) and A3-14 (grey) superposed using 
residues A4-T9 of the duplex region. The Watson and Crick region (A4-T9) was 
isomorphous in both structures whereas the 5’ non-canonical region (G1-A3) and the 
3’ non-canonical region (G10-A14) shows a complete rearrangement in the 14-mer 
structure as compared to the parent 13-mer.  
 
  76 
 
Chapter 4.4 Tertiary Interactions Fulfil a Structural Role to Generate Tandem G-A 
Base Pairs. 
 The secondary and tertiary structural environment surrounding the sheared 
A3-G10 base pair established a local structure that was highly similar to the tandem 
sheared GA/AG steps that have been previously observed in B-form helices
97–101
, 
with the sheared A3-G10 base pair being structurally equivalent to the second base 
pair (Figure 28A, B). This base pair was formed through the Hoogsteen edge of A3 
(N6, N7) and the sugar edge of G10 (N2, N3) and displayed the characteristic base 
pair buckling (Figure 28A). The non-planarity of the base pair lead to the formation 
of a potential interstrand hydrogen bond between N6 of A3 and O2 of T9, though the 
geometry is not ideal (Figure 29). Like previous solution structures, inter- and 
intrastrand stacking interactions played an important part in stabilizing the A3-G10 
pairing. Despite the relatively large twist angle (60.3
o
) at the A3A4/T9G10 step, there 
was significant intrastrand stacking between A3 and A4 (4.85 Å
2
 overlap based on 
polygon projections using X3DNA
73
 (Figure 30B). Intrastrand stacking interactions 
were even more pronounced for the partner strand, with T9-G10 stacking having 8.10 
Å
2 
overlap. The overall 12.95 Å
2
 overlap at this base pair step was the single most in 
the entire structure, suggesting that the capping A-G pairs provide significant stability 
to the duplex ends. Interstrand stacking interactions, one of the hallmarks of 
structures having tandem sheared G-A base pairs, were also present (Figure 30A). G2 
of one strand stacks with G10 of the partner strand. Overall, the structural 
environment around the A3-G10 pair was remarkably similar to previous solution 
structures (Figure 28A), including the presence of several phosphate linkages in the 
  77 
 
B
II
 conformer (G2, A3, G10) which is a hallmark of tandem sheared G-A structures
98
. 
The major difference was the lack of the first G-A pair. Interestingly, tertiary contacts 
between A14 from a different column of coaxially stacked helices and G2 maintained 
base stacking interactions and lead to a similar overall structure.   
 The spacing between the G10 and G11 nucleobases allowed A14 to stack 
between the A3 and G11 nucleotides from the partner strands, while forming a base 
pair with G2 (Figure 28A, C). Unlike the tandem sheared G-A structures, this tertiary 
contact occurred between the Watson-Crick face of the A14 and the sugar edge of 
G2, with N6 of A14 in almost the identical position to the first G-A in the solution 
structures (Figure 28C). Along with these base pairing and base stacking interactions, 
A14 made additional 3' OH contacts with the A4 phosphate and was also involved in 
the base capping interactions with the A3 sugar. This is an example of a tertiary 
structure interaction providing a structural equivalence to a previously observed 
secondary structure motif. 
  78 
 
 
Figure 28. Comparison of GA/AG motifs. A. Stereo view of the tandem sheared 
GA/AG base pairs with the flanked Watson-Crick base pairs from a solution structure 
(yellow; PDB ID: 175D) superposed on  G1-A4 from one strand (grey) and T9-G11 
of the partner strand (red).  A14 from a neighboring duplex is shown in magenta. A 
hydrogen bond between the A14 3'-OH and the A4 phosphate group is shown as a 
dashed line. B. The A3-G10 base pair from the 14-mer structure is highly similar to 
the second base pair in the tandem sheared GA solution structure. Hydrogen bonds 
are shown as dashed lines. C. The tertiary contact between A14-G2 mimics the first 
sheared GA base pair. The hydrogen bonds between the Watson-Crick face of A14 
and the Hoogsteen face of G2 are shown as dotted lines.  
  79 
 
 
Figure 29. Potential hydrogen bond between A3 and T9. The sheared A3-G10 
base pair leads to the formation of an interstrand hydrogen bond between N6 of A3 
and O2 of T9, shown in dotted lines. 
  80 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 30. The inter- and intrastrand stacking at the sheared G-A base pair. A. 
The interstrand stacking between the G2 nucleotide of one strand (grey) and the G10 
nucleotide of the partner strand (red) in a duplex and the A3 nucleotide of one duplex 
(grey) and the A14 nucleotide from a different duplex (magenta) involved in a tertiary 
mimic of tandem sheared G-A base pairs. B, C. Intrastrand stacking between the 
sheared A3-G10 base pair and the flanking T9-A4 base pair and G2-A14 base pair 
  81 
 
and the flanking G1-G11 base pair respectively. The overlap values are a sum of both 
the base pairs, based on polygon projections obtained from X3DNA. 
Chapter 4.5 Triplex Junction. 
 The triplex junction connected two duplex segments into pseudo-infinite 
coaxially stacked helices. Two distinct triple interactions were present within the 
junction. First, G2 was involved in a sugar-edge contact with A14 as described, but it 
also made a single hydrogen bond through O6 with G13 N2 from the next coaxially 
stacked duplex (Figure 31A). This interaction was mediated in part by the hydrogen 
bonding of a solvent ion that was within hydrogen bonding distance of G2 N1, G13 
N1, G1 O6, and the G11 phosphate of the partner duplex. Next, G1 was base paired to 
G11 of the partner duplex through their Watson-Crick and Hoogsteen faces, 
respectively. A12 from a coaxially stacked duplex made a single hydrogen bond with 
G1 (Figure 31B).  To our knowledge these kind of all purine triple interactions have 
not been previously observed in DNA. Additionally, they provide an atypical 
example of coaxial helical stacking. In this case, only the 5'-most residues (G1) are 
directly stacked, while the 3'-most residues (A14) are involved in the tertiary contacts 
that allow the parallel arrangement of adjacent duplex stacks. The purine base triples 
effectively “stitch” the four strands together at the major groove, without significant 
stacking interactions between the duplexes. 
  82 
 
 
Figure 31. Purine base triples. The two purine base triple interactions A. The first 
base triple is mediated by the hydrogen bonding between the sugar edge of G2 and 
the Watson-Crick edge of A14 and with an additional hydrogen bond between O6 of 
G2 and N2 of G13 from the adjacent stacked duplex. Additional hydrogen bonding is 
mediated by the interaction of the water molecule with G2 N1, G13 N1, G1 O6, and 
the G11 phosphate of the partner duplex. B. The second base triple is formed between 
the Watson-Crick edge of G1 and the Hoogsteen edge of G11 from its duplex partner, 
along with the single hydrogen bond between G1 and A12 from a coaxially stacked 
duplex. In both A. and B. the hydrogen bonds are denoted by dashed lines and sigma 
A-weighted electron density map (2F
o
-F
c
) contoured at 1.0σ is shown in blue. 
  83 
 
Chapter 4.6 Sequence Requirements for Alternate Crystal Form. 
 To understand why the addition of a single 3' adenosine could result in a 
significantly different structure under identical crystallization conditions, I set out to 
understand the sequence requirements for the alternate crystal form in the context of 
the determined crystal structures. I screened 30 variants of the 14-mer 
oligonucleotides by altering the nucleobase identities at positions mediating key 
interactions in the structures. I probed these interactions in three different groups. In 
the first group, I screened all four of the oligonucleotides described here, but with the 
different nucleobase identities at the added 14
th
 residue. Out of the 12 DNA 
oligonucleotides screened, 7 crystallized with a hexagonal unipyramidal habit and the 
other 5 failed to show any crystals (Table 8).  Notably, 3 of these sequences were the 
variants of A4-14, which did not crystallize as the 13-mer (Table 2).  These results 
suggested adenosine at 14
th
 position is a requirement for the alternate crystal form. 
This is consistent with our structural observations of the G2-A14 tertiary base pair. 
Simple modelling with different nucleotide identities at position 14 indicated that 
pyrimidines would be unable to pair with G2 without significant backbone clashes, 
while a guanosine at this position would present incompatible hydrogen bonding 
partners.  
 In the second group I examined six sequences with different self-
complementary base pairs formed by positions 4 and 9, adjacent to the A3-G10 
pairing (Table 8). I observed crystals in all cases, with sequences having the Y4-R9 
base pair exhibiting the hexagonal crystal habit, while the other two sequences having 
a G4-C9 base pair forming microcrystals or irregular crystals. Though I could not 
  84 
 
ascertain if the G4-C9 containing crystals belonged to one of the two crystal forms, 
these results indicated that the sequence rules observed for the formation of tandem 
sheared G-A base pairs
97
 also applied to these crystals. Only sequences with a 
thymidine 5' of G10 adopted the alternate crystal form described here, though it is 
possible that a cytosine at this location could promote the alternate crystal form. The 
strong stacking interactions between T9 and G10, along with geometric constraints, 
were previously suggested as reasons for the presence of the pyrimidine 5' of 
guanosines in the tandem sheared pairs
97,98,101
. Our structural and crystal screen 
results support this analysis, but also indicate that a potential A3-T9 interstrand 
hydrogen bond (Figure S2) may help stabilize these structures. Notably, the hydrogen 
bond acceptor at O2 would be present in either pyrimidine at position 9. 
 Finally, I screened several sequences variants at the A5-T8 base pair. Based 
on the local sequence rules for the formation of the sheared G-A pair, I anticipated 
that this position should have little impact on the interactions necessary to form the 
alternate crystal form. Interestingly, 10 out of the 12 sequences screened in this group 
failed to crystallize, while the remaining two sequences formed only poor crystals 
(Table 8). This somewhat surprising result may be explained in several ways. First, 
this may indicate that the significant stacking interactions between T8 and T9 (7.56 
Å
2
) are required to adopt the location conformation necessary to form the alternate 
crystal form, though this does not appear to be the case for solution structures 
containing tandem sheared GA pairs. Second, in previous chapters we have 
established that the A5-T8 base pair was an important determinant for crystallization 
and crystallization speed in the context of 13-mers. It is possible that this base pair 
  85 
 
may have a more fundamental role in the formation of the short self-complementary 
duplex that is a common feature of the 13-mer and 14-mer structures.  
 Altogether the sequence study and the structural observations strongly suggest 
that presence of A14-G2, A4-T8 and A5-T9 are all critical to the formation and 
stabilization of the alternate crystal form. 
 
 
 
 
 
 
 
 
 
 
 
 
  86 
 
Table 7. Sequence-dependent crystallization. 
Designation Sequence* Crystal Habit 
A1-14-T GGAAAATTTGGAGT Hexagonal 
A1-14-G GGAAAATTTGGAGG Hexagonal 
A1-14-C GGAAAATTTGGAGC Hexagonal 
A2-14-T GGAAACGTTGGAGT None 
A2-14-G GGAAACGTTGGAGG Hexagonal 
A2-14-C GGAAACGTTGGAGC Hexagonal 
A3-14-T GGAAAGCTTGGAGT None 
A3-14-G GGAAAGCTTGGAGG Hexagonal 
A3-14-C GGAAAGCTTGGAGC Hexagonal 
A4-14-T GGAAATATTGGAGT None 
A4-14-G GGAAATATTGGAGG None 
A4-14-C GGAAATATTGGAGC None 
B6-14-A GGACACGTGGGAGA Hexagonal 
B7-14-A GGACAGCTGGGAGA Hexagonal 
E1-14-A GGATAATTAGGAGA Hexagonal 
E3-14-A GGATAGCTAGGAGA Hexagonal 
C9-14-A GGAGAATTCGGAGA Microcrystals 
C11-14-A GGAGAGCTCGGAGA Clusters 
B1-14-A GGAATATATGGAGA None 
B2-14-A GGAATCGATGGAGA None 
B3-14-A GGAATGCATGGAGA None 
B4-14-A GGAATTAATGGAGA None 
A5-14-A GGAACATGTGGAGA None 
A6-14-A GGAACCGGTGGAGA None 
A7-14-A GGAACGCGTGGAGA None 
A8-14-A GGAACTAGTGGAGA Clusters 
A9-14-A GGAAGATCTGGAGA Hexagonal 
A10-14-A GGAAGCGCTGGAGA None 
A11-14-A GGAAGGCCTGGAGA None 
A12-14-A GGAAGTACTGGAGA None 
*Red indicate the position of sequence variability. 
 
  87 
 
Chapter 4.7 Materials and Methods 
1. DNA Synthesis and Purification  
 The four DNA 14-mers, designated A1-14: 5’-d(GGAAAATTTGGAGA); 
A2-14: 5’-d(GGAAACGTTGGAGA); A3-14: 5’-d(GGAAAGCTTGGAGA); A4-14: 
5’- d(GGAAATATTGGAGA), were synthesized on the 1 µmol scale (Integrated 
DNA Technologies, Coralville, IA) and were purified by 20% (19:1) polyacrylamide 
gel electrophoresis, electroeluted, and ethanol precipitated as previously described
59
. 
The A3-14 (BrU9) oligonucleotide was synthesized using standard phosphoramidite 
chemistry on an Expedite 8909 DNA synthesizer (PerSeptive BioLabs) with reagents 
from Glen Research (Sterling, VA). The purified DNA samples were dialyzed against 
deionized water and the concentration was adjusted to 260 µM. Oligonucleotides 
used to examine sequence effects on crystallization (Table 2) were synthesized on the 
100 nmol scale, dissolved in deionized water, and used without purification.  
2. Crystallization 
 The DNA oligonucleotides were crystallized by sitting drop vapour diffusion. 
Prior to crystallization, DNA samples (260 µM) were heated at 95˚ for 2' and cooled 
to room temperature. Samples were mixed (1:1) with crystallization buffer (120 mM 
magnesium formate, 50 mM lithium chloride and 10% 2-methyl-2,4-pentanediol) in a 
4 µl drop. The reservoir contained 400 µl of crystallization buffer. The crystal plates 
were incubated at 22˚C. Crystals appeared in 16-20 hours and grew to an average size 
of 250x75x100 µm. 
  88 
 
3. Data Collection and Structure Determination 
 Crystals were harvested by nylon loop, washed sequentially in crystallization 
buffer containing 30% and 40% 2-methyl-2,4-pentanediol, and flash-cooled in liquid 
nitrogen. Native data sets were collected at Advanced Photon Source, Argonne 
National Labs, Sector 24-ID-E. Data were indexed and integrated using XDS
74
, and 
scaled using Aimless
75
. Several data sets had relatively low completeness due to 
crystal orientation. However, each crystal type was highly isomorphous with respect 
to the unit cell dimensions (RMSD of unit cell dimensions ≤ 0.1 Å) allowing the 
merging of observations from multiple crystals to improve completeness. Phases were 
initially determined using an A3-14 (BrU9) derivative with data collected at 
Advanced Photon Source beamline 24-ID-C. Phases were determined by single 
wavelength anomalous dispersion with the substructure sites identified by HySS in 
the Phenix crystallography package
77,102
. Models were built in Coot
76
. The other three 
14-mer structures were solved by molecular replacement using the completed A3-14 
structure as a search model. Refinement was performed with Phenix
77
. Water 
molecules and ions were added manually during the refinement process. Following 
converged refinement in Phenix, all of the structures were run through the PDB-
REDO pipeline
78
 with 10-fold cross-validation applied due to the small number of 
reflections in these datasets. Average R
free
 for these 10 different test sets are reported 
in Table 1. Coordinates and structure factors were deposited in the Protein Data 
Bank
79
. 
  89 
 
Chapter 5: Conclusion and Future Prospects 
 
 Through my work, I have identified 12 new DNA oligonucleotide sequences, 
capable of adopting isomorphous crystal structures, starting from a single 
oligonucleotide sequence. This study opens up the sequence space and provides 
diversity in the design and construction of 3D DNA crystals. Importantly, the 13-mer 
crystals described in this study contains solvent channels that run throughout the 
length of the crystals, making them ideal candidates for use as molecular scaffolds. 
The identification of new sequences that form isomorphous structures will provide 
additional diversity for attaching and characterizing guest molecules for this and other 
applications.  
 Having identified sequences with isomorphous structures, I demonstrated that 
hybrid crystals could be grown from a mixture of two different DNA 
oligonucleotides. This provides a new tool for understanding the crystal self-assembly 
process, particularly the nucleation of these 13-mer DNA oligonucleotides. These 
oligonucleotides provides a unique system for further understanding how a dispersed 
biopolymer can self-assemble into macroscopic objects through a simple chemical 
trigger. 
 In this study I demonstrated how a single base pair (A5-T8) in the 13-mer 
structure is related to both crystallizability and the crystallization speed. Similarly, the 
14-mer structure described here demonstrates how a single nucleotide change can 
lead to significantly different secondary and tertiary interactions and subsequently the 
overall structure. Additionally, through the sequence dependent crystallization study I 
  90 
 
determined that adoption of the 14-mer structure is in part influenced by the 
neighboring sequence.  
 In our lab we are exploring the use of 13-mer oligonucleotides for developing 
“core-shell” crystals as drug delivery vehicles and for solid state catalysis. The added 
diversity provided by the newly identified isomorphous sequences, will help in 
attaching multiple guest molecules to the DNA to have a ‘multi-functional’ crystal. 
My work also proved useful in the crystal crosslinking project in our lab. The various 
sequences with isomorphous sequences proved to be a good system for studying the 
crosslinking efficiencies for various 13-mer sequences. The work on heterogenous 
crystals could be explored further to use these 13-mer crystals as ‘crystallization-aid’, 
where a small amount of crystallizing oligonucleotide promotes the crystallization of 
other difficult to crystallize oligonucleotides. 
 I believe that my work will have important implications for improving the 
characteristics of periodic 3D DNA crystals, which have been one of the most highly 
sought after DNA architectures. I also believe that my study is a step forward in 
understanding the assembly process for DNA oligonucleotides and provide insights in 
the sequence-structure relationship for 3D DNA crystals. This would prove useful in 
the rational designing of DNA structures and help in optimizing the conditions for 
assembly, manipulation, and functionalization. This in turn will benefit both upstream 
design and downstream applications of DNA structural nanotechnology.  
 
  91 
 
Bibliography 
 
1. Dahm, R. Discovering DNA: Friedrich Miescher and the early years of nucleic 
acid research. Hum. Genet. 122, 565–581 (2008). 
2. Jones, M. R., Seeman, N. C. & Mirkin, C. A. Programmable materials and the 
nature of the DNA bond. Science 347, 1260901 (2015). 
3. Aldaye, F. A., Palmer, A. L. & Sleiman, H. F. Assembling Materials with DNA 
as the Guide. Science 321, 1795–1799 (2008). 
4. Lipfert, J., Doniach, S., Das, R. & Herschlag, D. Understanding Nucleic Acid–
Ion Interactions. Annu. Rev. Biochem. 83, 813–841 (2014). 
5. Winfree, E., Liu, F., Wenzler, L. A. & Seeman, N. C. Design and self-assembly 
of two-dimensional DNA crystals. Nature 394, 539–44 (1998). 
6. Rothemund, P. W. K. Folding DNA to create nanoscale shapes and patterns. 
Nature 440, 297–302 (2006). 
7. Yan, H., Park, S. H., Finkelstein, G., Reif, J. H. & LaBean, T. H. DNA-templated 
self-assembly of protein arrays and highly conductive nanowires. Science 301, 
1882–4 (2003). 
8. He, Y., Chen, Y., Liu, H., Ribbe, A. E. & Mao, C. Self-assembly of hexagonal 
DNA two-dimensional (2D) arrays. J. Am. Chem. Soc. 127, 12202–3 (2005). 
9. Roy, S. & Caruthers, M. Synthesis of DNA/RNA and Their Analogs via 
Phosphoramidite and H-Phosphonate Chemistries. Molecules 18, 14268–14284 
(2013). 
10. Drew, H. R. et al. Structure of a B-DNA dodecamer: conformation and dynamics. 
Proc. Natl. Acad. Sci. U. S. A. 78, 2179–2183 (1981). 
  92 
 
11. Seeman, N. C. DNA in a material world. Nature 421, 427–431 (2003). 
12. Seeman, N. C. Nucleic acid junctions and lattices. J. Theor. Biol. 99, 237–47 
(1982). 
13. Li, X., Yang, X., Qi, J. & Seeman, N. C. Antiparallel DNA Double Crossover 
Molecules As Components for Nanoconstruction. J. Am. Chem. Soc. 118, 6131–
6140 (1996). 
14. LaBean, T. H., Yan H., Kopatsch J., Liu F., Winfree E., Reif J. H., Seeman N. C. 
Construction, Analysis, Ligation, and Self-Assembly of DNA Triple Crossover 
Complexes. J. Am. Chem. Soc. 122, 1848–1860 (2000). 
15. He, Y., Ye T., Su M., Zhang C., Ribbe A. E., Jiang W., Mao C. Hierarchical self-
assembly of DNA into symmetric supramolecular polyhedra. Nature 452, 198–
201 (2008). 
16. Chen, J. H. & Seeman, N. C. Synthesis from DNA of a molecule with the 
connectivity of a cube. Nature 350, 631–3 (1991). 
17. Goodman, R. P., Schaap I. A. T., Tardin C. F., Erben C. M., Berry R. M., 
Schmidt C. F., Tuberfield A.J. Rapid chiral assembly of rigid DNA building 
blocks for molecular nanofabrication. Science 310, 1661–5 (2005). 
18. Aldaye, F. A. & Sleiman, H. F. Modular Access to Structurally Switchable 3D 
Discrete DNA Assemblies. J. Am. Chem. Soc. 129, 13376–13377 (2007). 
19. Erben, C. M., Goodman, R. P. & Turberfield, A. J. A self-assembled DNA 
bipyramid. J. Am. Chem. Soc. 129, 6992–3 (2007). 
20. Mitchell, J. C., Harris, J. R., Malo, J., Bath, J. & Turberfield, A. J. Self-assembly 
of chiral DNA nanotubes. J. Am. Chem. Soc. 126, 16342–3 (2004). 
  93 
 
21. Lund, K., Liu, Y., Lindsay, S. & Yan, H. Self-assembling a molecular pegboard. 
J. Am. Chem. Soc. 127, 17606–7 (2005). 
22. Liu, Y., Ke, Y. & Yan, H. Self-assembly of symmetric finite-size DNA 
nanoarrays. J. Am. Chem. Soc. 127, 17140–1 (2005). 
23. Zheng, J., Birktoft J. J., Chen Y., Wang T., Sha R., Constantinou P.E., Ginell 
S.L., Mao C., Seeman N. C. From molecular to macroscopic via the rational 
design of a self-assembled 3D DNA crystal. Nature 461, 74–77 (2009). 
24. Andersen, E. S. et al. Self-assembly of a nanoscale DNA box with a controllable 
lid. Nature 459, 73–76 (2009). 
25. Douglas, S. M. et al. Self-assembly of DNA into nanoscale three-dimensional 
shapes. Nature 459, 414–418 (2009). 
26. Dietz, H., Douglas, S. M. & Shih, W. M. Folding DNA into Twisted and Curved 
Nanoscale Shapes. Science 325, 725–730 (2009). 
27. Han, D. et al. DNA origami with complex curvatures in three-dimensional space. 
Science 332, 342–346 (2011). 
28. Wei, B., Dai, M. & Yin, P. Complex shapes self–assembled from single–stranded 
DNA tiles. Nature 485, 623–626 (2012). 
29. Ke, Y., Ong, L. L., Shih, W. M. & Yin, P. Three-Dimensional Structures Self-
Assembled from DNA Bricks. Science 338, 1177–1183 (2012). 
30. Ke, Y. et al. DNA brick crystals with prescribed depths. Nat. Chem. 6, 994–1002 
(2014). 
  94 
 
31. Mao, C., LaBean, T. H., Relf, J. H. & Seeman, N. C. Logical computation using 
algorithmic self-assembly of DNA triple-crossover molecules. Nature 407, 493–6 
(2000). 
32. Zhang, F., Nangreave, J., Liu, Y. & Yan, H. Structural DNA nanotechnology: 
state of the art and future perspective. J. Am. Chem. Soc. 136, 11198–11211 
(2014). 
33. Drew, H. R., McCall, M. J. & Calladine, C. R. Recent Studies of DNA in the 
Crystal. Annu. Rev. Cell Biol. 4, 1–20 (1988). 
34. Drew, H., Takano, T., Tanaka, S., Itakura, K. & Dickerson, R. E. High-salt 
d(CpGpCpG), a left-handed Z[prime] DNA double helix. Nature 286, 567–573 
(1980). 
35. Dickerson, R. E. & Drew, H. R. Structure of a B-DNA dodecamer. J. Mol. Biol. 
149, 761–786 (1981). 
36. Rich, A. & Zhang, S. Z-DNA: the long road to biological function. Nat Rev Genet 
4, 566–572 (2003). 
37. Lee, J. S., Johnson, D. A. & Morgan, A. R. Complexes formed by (pyrimidine)
n
 . 
(purine)
n
 DNAs on lowering the pH are three-stranded. Nucleic Acids Res. 6, 
3073–3091 (1979). 
38. Lilley, D. M. Hairpin-loop formation by inverted repeats in supercoiled DNA is a 
local and transmissible property. Nucleic Acids Res. 9, 1271–1289 (1981). 
39. Panayotatos, N. & Wells, R. D. Cruciform structures in supercoiled DNA. Nature 
289, 466–470 (1981). 
  95 
 
40. Sundquist, W. I. & Klug, A. Telomeric DNA dimerizes by formation of guanine 
tetrads between hairpin loops. Nature 342, 825–9 (1989). 
41. Bacolla, A. & Wells, R. D. Non-B DNA Conformations, Genomic 
Rearrangements, and Human Disease. J. Biol. Chem. 279, 47411–47414 (2004). 
42. Burge, S., Parkinson, G. N., Hazel, P., Todd, A. K. & Neidle, S. Quadruplex 
DNA: sequence, topology and structure. Nucleic Acids Res. 34, 5402–5415 
(2006). 
43. Zhao, J., Bacolla, A., Wang, G. & Vasquez, K. M. Non-B DNA structure-induced 
genetic instability and evolution. Cell. Mol. Life Sci. 67, 43–62 (2009). 
44. Yatsunyk, L. A., Mendoza, O. & Mergny, J.-L. ‘Nano-oddities’: unusual nucleic 
acid assemblies for DNA-based nanostructures and nanodevices. Acc. Chem. Res. 
47, 1836–1844 (2014). 
45. Dong, Y., Yang, Z. & Liu, D. DNA nanotechnology based on i-motif structures. 
Acc. Chem. Res. 47, 1853–1860 (2014). 
46. Chakraborty, S., Sharma, S., Maiti, P. K. & Krishnan, Y. The poly dA helix: a 
new structural motif for high performance DNA-based molecular switches. 
Nucleic Acids Res. 37, 2810–2817 (2009). 
47. Saha, S., Chakraborty, K. & Krishnan, Y. Tunable, colorimetric DNA-based pH 
sensors mediated by A-motif formation. Chem. Commun. 48, 2513–2515 (2012). 
48. Yatsunyk, L. A. et al. Guided assembly of tetramolecular G-quadruplexes. ACS 
Nano 7, 5701–5710 (2013). 
  96 
 
49. Zhang, D., Huang, T., Lukeman, P. S. & Paukstelis, P. J. Crystal structure of a 
DNA/Ba
2+
 G-quadruplex containing a water-mediated C-tetrad. Nucleic Acids 
Res. 42, 13422–13429 (2014). 
50. Muser, S. E. & Paukstelis, P. J. Three-dimensional DNA crystals with pH-
responsive noncanonical junctions. J. Am. Chem. Soc. 134, 12557–12564 (2012). 
51. Tripathi, S., Zhang, D. & Paukstelis, P. J. An intercalation-locked parallel-
stranded DNA tetraplex. Nucleic Acids Res. 43, 1937–1944 (2015). 
52. Peyret, N., Seneviratne, P. A., Allawi, H. T. & SantaLucia, J. Nearest-neighbor 
thermodynamics and NMR of DNA sequences with internal A.A, C.C, G.G, and 
T.T mismatches. Biochemistry 38, 3468–77 (1999). 
53. Tikhomirova, A., Beletskaya, I. V. & Chalikian, T. V. Stability of DNA duplexes 
containing GG, CC, AA, and TT mismatches. Biochemistry 45, 10563–10571 
(2006). 
54. Rossetti, G. et al. The structural impact of DNA mismatches. Nucleic Acids Res. 
43, 4309-4321 (2015).  
55. Bochman, M. L., Paeschke, K. & Zakian, V. A. DNA secondary structures: 
stability and function of G-quadruplex structures. Nat. Rev. Genet. 13, 770–780 
(2012). 
56. Wei, X., Nangreave, J. & Liu, Y. Uncovering the Self-Assembly of DNA 
Nanostructures by Thermodynamics and Kinetics. Acc. Chem. Res. 47, 1861–
1870 (2014). 
57. Song, J. et al. Isothermal hybridization kinetics of DNA assembly of two-
dimensional DNA origami. Small  9, 2954–2959 (2013). 
  97 
 
58. Johnson-Buck, A., Nangreave, J., Jiang, S., Yan, H. & Walter, N. G. 
Multifactorial Modulation of Binding and Dissociation Kinetics on Two-
Dimensional DNA Nanostructures. Nano Lett. 13, 2754–2759 (2013). 
59. Paukstelis, P. J., Nowakowski, J., Birktoft, J. J. & Seeman, N. C. Crystal 
Structure of a Continuous Three-Dimensional DNA Lattice. Chem. Biol. 11, 
1119–1126 (2004). 
60. Sizov, I., Rahman, M., Gelmont, B., Norton, M. L. & Globus, T. Sub-THz 
spectroscopic characterization of vibrational modes in artificially designed DNA 
monocrystal. Chem. Phys. 425, 121–125 (2013). 
61. Zhang, W., Brown, E. R., Rahman, M. & Norton, M. L. Observation of terahertz 
absorption signatures in microliter DNA solutions. Appl. Phys. Lett. 102, 023701 
(2013). 
62. Paukstelis, P. J. Three-dimensional DNA crystals as molecular sieves. J. Am. 
Chem. Soc. 128, 6794–5 (2006). 
63. Geng, C. & Paukstelis, P. J. DNA Crystals as Vehicles for Biocatalysis. J. Am. 
Chem. Soc. 136, 7817–7820 (2014). 
64. Leontis, N. B. & Westhof, E. The Annotation of RNA Motifs. Comp. Funct. 
Genomics 3, 518–524 (2002). 
65. Greene, K. L., Jones R.L., Li Y., Robinson H., Wang A.H.J., Zon G., Wilson W. 
D. Solution Structure of a GA Mismatch DNA Sequence, d(CCATGAATGG)2, 
Determined by 2D NMR and Structural Refinement Methods. Biochemistry 33, 
1053–1062 (1994). 
  98 
 
66. Prive, G. G., Helnemann U., Chandrasegran S., Kan L. S., KopkaM. L., 
Dickerson R.E. Helix geometry, hydration, and G.A mismatch in a B-DNA 
decamer. Science 238, 498–504 (1987). 
67. Hays, F. A. Teegarden A., Jones Z. J. R., Harms M., Raup D., Watson J., 
Cavaliere E., Ho P. S. How sequence defines structure: A crystallographic map of 
DNA structure and conformation. Proc. Natl. Acad. Sci. U. S. A. 102, 7157–7162 
(2004). 
68. Baumgartner, J. et al. Nucleation and growth of magnetite from solution. Nat. 
Mater. 12, 310–314 (2013). 
69. Koo, H. S. & Crothers, D. M. Calibration of DNA curvature and a unified 
description of sequence-directed bending. Proc. Natl. Acad. Sci. U. S. A. 85, 
1763–1767 (1988). 
70. Hagerman, P. J. Flexibility of DNA. Annu. Rev. Biophys. Biophys. Chem. 17, 
265–86 (1988). 
71. Haran, T. E., Kahn, J. D. & Crothers, D. M. Sequence Elements Responsible for 
DNA Curvature. J. Mol. Biol. 244, 135–143 (1994). 
72. Kahn, J. D., Yun, E. & Crothers, D. M. Detection of localized DNA flexibility. 
Nature 368, 163–166 (1994). 
73. Lu, X.-J. & Olson, W. K. 3DNA: a versatile, integrated software system for the 
analysis, rebuilding and visualization of three-dimensional nucleic-acid 
structures. Nat. Protoc. 3, 1213–1227 (2008). 
74. Kabsch, W. XDS. Acta Crystallogr. D Biol. Crystallogr. 66, 125–132 (2010). 
  99 
 
75. Evans, P. R. & Murshudov, G. N. How good are my data and what is the 
resolution? Acta Crystallogr. D Biol. Crystallogr. 69, 1204–1214 (2013). 
76. Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development 
of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010). 
77. Afonine, P. V. et al. Towards automated crystallographic structure refinement 
with phenix.refine. Acta Crystallogr. D Biol. Crystallogr. 68, 352–367 (2012). 
78. Joosten, R. P., Long, F., Murshudov, G. N. & Perrakis, A. The PDB_REDO 
server for macromolecular structure model optimization. International Union Of 
Crystalography (IUCr) J 1, 213–220 (2014). 
79. Berman, H. M. et al. The Protein Data Bank. Nucleic Acids Res. 28, 235–242 
(2000). 
80. Delano, W. The PyMOL Molecular Graphics System, Schrödinger, LLC. 
81. Hwang, T. L. & Shaka, A. J. Water Suppression That Works. Excitation 
Sculpting Using Arbitrary Wave-Forms and Pulsed-Field Gradients. J. Magn. 
Reson. A 112, 275–279 (1995). 
82. Carr, H. Y. & Purcell, E. M. Effects of Diffusion on Free Precession in Nuclear 
Magnetic Resonance Experiments. Phys. Rev. 94, 630–638 (1954). 
83. Nath, K., Sarosy, J. W., Hahn, J. & Di Como, C. J. Effects of ethidium bromide 
and SYBR® Green I on different polymerase chain reaction systems. J. Biochem. 
Biophys. Methods 42, 15–29 (2000). 
84. Roberts, T. M., Kacich, R. & Ptashne, M. A general method for maximizing the 
expression of a cloned gene. Proc. Natl. Acad. Sci. U. S. A. 76, 760–764 (1979). 
  100 
 
85. Price, P. A. The essential role of Ca
2+
 in the activity of bovine pancreatic 
deoxyribonuclease. J. Biol. Chem. 250, 1981–1986 (1975). 
86. Rossetti, G. et al. The structural impact of DNA mismatches. Nucleic Acids Res. 
43, 4309–4321 (2015). 
87. Kan, L. S., Chandrasegaran, S., Pulford, S. M. & Miller, P. S. Detection of a 
guanine-adenine base pair in a decadeoxyribonucleotide by proton magnetic 
resonance spectroscopy. Proc. Natl. Acad. Sci. U. S. A. 80, 4263–4265 (1983). 
88. Patel, D. J., Kozlowski, S. A., Ikuta, S. & Itakura, K. Deoxyguanosine-
deoxyadenosine pairing in the d(C-G-A-G-A-A-T-T-C-G-C-G) duplex: 
conformation and dynamics at and adjacent to the dG-dA mismatch site. 
Biochemistry 23, 3207–3217 (1984). 
89. Brown, T., Hunter, W. N., Kneale, G. & Kennard, O. Molecular structure of the 
G-A base pair in DNA and its implications for the mechanism of transversion 
mutations. Proc. Natl. Acad. Sci. U. S. A. 83, 2402–2406 (1986). 
90. Brown, T., Leonard, G. A., Booth, E. D. & Chambers, J. Crystal structure and 
stability of a DNA duplex containing A(anti) · G(syn) base-pairs. J. Mol. Biol. 
207, 455–457 (1989). 
91. Hunter, W. N., Brown, T. & Kennard, O. Structural Features and Hydration of 
d(C-G-C-G-A-A-T-T-A-G-C-G); a Double Helix Containing Two G.A Mispairs. 
J. Biomol. Struct. Dyn. 4, 173–191 (1986). 
92. Leonard, G. A., Booth, E. D. & Brown, T. Structural and thermodynamic studies 
on the adenine.guanine mismatch in B-DNA. Nucleic Acids Res. 18, 5617–5623 
(1990). 
  101 
 
93. Nikonowicz, E. P. & Gorenstein, D. G. Two-dimensional proton and phosphorus-
31 NMR spectra and restrained molecular dynamics structure of a mismatched 
GA decamer oligodeoxyribonucleotide duplex. Biochemistry  29, 8845–8858 
(1990). 
94. Li, Y., Zon, G. & Wilson, W. D. NMR and molecular modeling evidence for a 
G.A mismatch base pair in a purine-rich DNA duplex. Proc. Natl. Acad. Sci. U. S. 
A. 88, 26–30 (1991). 
95. Li, Y., Zon, G. & Wilson, W. D. Thermodynamics of DNA duplexes with 
adjacent G-A mismatches. Biochemistry (Mosc.) 30, 7566–7572 (1991). 
96. Coimbatore Narayanan, B. et al. The Nucleic Acid Database: new features and 
capabilities. Nucleic Acids Res. 42, D114–D122 (2014). 
97. Cheng, J.-W., Chou, S.-H. & Reid, B. R. Base pairing geometry in GA 
mismatches depends entirely on the neighboring sequence. J. Mol. Biol. 228, 
1037–1041 (1992). 
98. Chou, S.-H., Cheng, J.-W. & Reid, B. R. Solution structure of 
[d(ATGAGCGAATA)]
2
: Adjacent G-A mismatches stabilized by cross-strand 
base-stacking and BII phosphate groups. J. Mol. Biol. 228, 138–155 (1992). 
99. Chou, S.-H., Zhu, L. & Reid, B. R. The Unusual Structure of the Human 
Centromere (GGA)
2
 Motif: Unpaired Guanosine Residues Stacked Between 
Sheared G·A Pairs. J. Mol. Biol. 244, 259–268 (1994). 
100. Chou, S.-H., Cheng, J.-W., Fedoroff, O. & Reid, B. R. DNA Sequence 
GCGAATGAGC Containing the Human Centromere Core Sequence GAAT 
  102 
 
Forms a Self-complementary Duplex with Sheared G·A Pairs in Solution. J. Mol. 
Biol. 241, 467–479 (1994). 
101. Chou, S.-H., Chin, K.-H. & Wang, A. H.-J. Unusual DNA duplex and hairpin 
motifs. Nucl Acids Res 31, 2461–2474 (2003). 
102. Grosse-Kunstleve, R. W. & Adams, P. D. Substructure search procedures for 
macromolecular structures. Acta Crystallogr. D Biol. Crystallogr. 59, 1966–1973 
(2003). 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  103 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  104