|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 METAL-ORGANIC MONOLAYERS AS SUBSTRATES FOR BIOMOLECULE MICROARRAYS By SARAH MARIE LANE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
2 2007 Sarah M. Lane
3 To my family (Papa, Mom, Margaret, Zac, Mellita, Pe ter and, of course, Jorge and Victoria)
4 ACKNOWLEDGMENTS I would like to thank my advisor, for mental and moral support; the University of Florida Chemistry Department for providing me with an outstanding education; Marcela Morado for her hard work on the lanthanide self-assembled monolayer project; and Eric Lambers and the members of MAIC for allowing me unbridled acc ess to the XPS instrument. Finally, I would like to thank the National Science Foundation for funding.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES................................................................................................................ .........8 LIST OF ABBREVIATIONS........................................................................................................11 ABSTRACT....................................................................................................................... ............12 CHAPTER 1 INTRODUCTION: XPS ANALYSIS OF DNA....................................................................14 Introduction................................................................................................................... ..........14 Qualitative and Semiquantitative XPS...................................................................................17 Quantitative................................................................................................................... ..........20 Conclusion..................................................................................................................... .........24 2 USE OF XPS TO STUDY OLIGONU CLEOTIDES AT A ZIRCONIUMPHOSPHONATE SURFACE................................................................................................26 Introduction................................................................................................................... ..........26 Experimental................................................................................................................... ........27 Materials...................................................................................................................... ....27 Zirconium-Phosphonate Substrates.................................................................................27 DNA Immobilization.......................................................................................................28 Results and Discussion......................................................................................................... ..30 The Zirconium-Phosphonate Surfaces............................................................................30 X-ray Photoelectron Spectroscopy..................................................................................30 Substrate-Overlayer Model.............................................................................................31 Model for the Zirconium-Phosphonate Surface..............................................................33 Conclusions.................................................................................................................... .........37 3 SPACER AND RINSING EFFECTS ON SSDNA SURFACE COVERAGE......................41 Introduction................................................................................................................... ..........41 Experimental................................................................................................................... ........41 Results and Discussion......................................................................................................... ..42 Conclusion..................................................................................................................... .........49 4 EFFECT OF PHOSPHATE LINKER PLACEMENT ON THE BINDING OF DOUBLE-STRANDED DNA TO ZIRCO NIUM-PHOSPHONATE SURFACES..............53
6 Introduction................................................................................................................... ..........53 Experimental................................................................................................................... ........55 Results and Discussion......................................................................................................... ..56 Linker and Spacer Placement..........................................................................................56 Single Versus. Double Stranded DNA............................................................................61 Conclusion..................................................................................................................... .........62 5 HYBRIDIZATION OF DNA AT A ZIRC ONIUM-PHOSPHONATE SURFACE AN XPS STUDY...................................................................................................................... .....68 Introduction................................................................................................................... ..........68 Experimental................................................................................................................... ........68 Confirmation of Hybridization with Fluorescence Confocal Microscopy......................69 Fluorescence Comparison of Hybridization Methods.....................................................70 Determination of Nitrogen Contamination from Hybridization Solution with XPS.......70 Hybridization Study with XPS........................................................................................71 Results and Discussion......................................................................................................... ..71 Conclusion..................................................................................................................... .........77 6 LANTHANIDE MONOLAYERS AS SUBSTRATES FOR PROTEIN MICROARRAYS PART 1: PREPAR ATION OF A ROBUST LANTHANIDE MONOLAYER...................................................................................................................... .80 Introduction................................................................................................................... ..........80 Experimental................................................................................................................... ........82 Materials...................................................................................................................... ....82 Lanthanide Phosphonate Langmuir-Blodgett Films........................................................82 Self-Assembled Lanthanide Films..................................................................................83 Analysis....................................................................................................................... ....84 Results and Discussion......................................................................................................... ..84 Langmuir-Blodgett Lanthanide Monolayers...................................................................84 Self-Assembled Lanthanide Films on Glass....................................................................86 Conclusion and Remarks........................................................................................................89 7 LANTHANIDE MONOLAYERS AS SUBSTRATES FOR PROTEIN MICROARRAYS PART 2: A MODEL STUDY...............................................................98 Introduction................................................................................................................... ..........98 Experimental................................................................................................................... ........99 Results and Discussion.........................................................................................................106 Future Work.................................................................................................................... ......108 8 CONCLUSIONS..................................................................................................................112 LIST OF REFERENCES.............................................................................................................114 BIOGRAPHICAL SKETCH.......................................................................................................120
7 LIST OF TABLES Table page 2-1 Intensities of the N 1s and Zr 3d peak s for a zirconium surface with and without DNA. The table also includes the surface cove rage of DNA that was calculated with the peak intensities........................................................................................................... ..40 5-1 Comparison of surface coverage of the DNA before hybridization, after hybridization with fewer water rinsings (A fter Hyb-1) and afte r hybridization with more water rinsings (After Hyb-2) for pr obe containing either a poly-A spacer or poly-G spacer.................................................................................................................. ...79
8 LIST OF FIGURES Figure page 1-1 Compounds used by Higashi et al. in their in binary SAM to immobilize DNA..............24 1-2 DNA with sulfur modified phosphate backbone used by Leavitt et al. .............................25 2-1 Illustration of the immobilization of the phosphorylated ssDNA and subsequent hybridization.................................................................................................................. ....38 2-2 Procedure for making the zirconium-phos phonate monolayers. First the ODPA is spread at the air-water interface and th en transferred onto a hydrophobic support. This followed by the addition of zirconyl ch loride which forms a network with the phosphonate headgroups....................................................................................................38 2-3 Graph of the analyzer transmission function. Ea is the analyzer energy (pass energy) and Ek is the kinetic energy of the photoelectr on. The diamonds are the scaled data points......................................................................................................................... .........39 2-4 Parameters that are used to calculate the DNA surface coverage. A) The parameters for the gold-DNA system. B) The parameters for the zirconium-phosphonate-DNA system......................................................................................................................... .......39 3-1 Fluorescence intensity af ter hybridization comparing th e four different spacers on three different probe sequences.........................................................................................50 3-2 Illustration of how the poly-dG quadr uplex might hold together the probe DNA............50 3-3 Poly-dG quadruplex. Illustrates th e hydrogen bonding between the bases of guanosine...................................................................................................................... .....51 3-4 Surface coverage of probe after each different rinsing condition......................................51 3-5 The bases and sugar deoxyadenosine and deoxyguanosine. From these structures, it is not apparent why a poly-dA strand would physisorb more strongly to the zirconium-phosphonate surface.........................................................................................52 4-1 Fluorescence intensities of the spots for the protein binding to the corresponding dsDNA. The dsDNA show the different spa cer and linker motifs that were used...........64 4-2 Steps used to test the binding the ArgR protein to dsDNA cont aining a 22 base-pair sequence....................................................................................................................... ......65 4-3 The dsDNA with the different phosphate-li nker/spacer motifs that were studied with XPS............................................................................................................................ ........65
9 4-4 Typical XPS spectra for the N 1s peak. This peak in particular is for 3G9PO4,3G9PO4............................................................................................................66 4-5 Comparison of the surface dens ities of the dsDNA strands..............................................66 4-6 Surface density comparison of dsDNA versus ssDNA......................................................67 5-1 Fluorescence confocal microscopy image of hybridized spots. The total spot (about 2mm in diameter) would not fit in the whole image..........................................................77 5-2 XPS spectra of blank hybridization take n to determine if nitrogen contamination would occur by following the N 1s peak. Only the significant region is shown, for this reason the percentages do not add up to 100...............................................................78 5-3 Fluorescence image of hybridized spot s comparing two hybridization methods. A) Fluorescence image comparing the two hybridization methods. B) Drawing indicating which part of the slide corr esponds to which method and which spots correspond to either the probe with poly-A spacer or poly-G spacer................................79 6-1 Two approaches explored here to prepar e a lanthanide substr ate for protein binding microarrays. A) The Langmuir-Blodgett met hod for preparing the substrate. B) The self-assembly method using a siloxane-c ontaining lanthanide-binding ligand.................90 6-2 The intramolecular energy transfer that can occur between a ligand and lanthanide........91 6-3 Procedure of the Langmuir-Blodgett me thod used to make the metal-phosphonate monolayers. Note that for clarity the Ca2+ in the subphase is not shown in the figure.....91 6-4 XPS survey scan of lanthanide phosphona te films. A) Samarium-phosphonate film. B) Terbium-phosphonate film............................................................................................92 6-5 AFM images of zirconium, samarium, and terbium phosphonate films............................93 6-6 XPS spectrum of self-assembled EDTA-siloxa ne lanthanide film. The insert is a magnification of the N 1s region.......................................................................................94 6-7 Molecular structure of dipico linic acid and picolinic acid.................................................94 6-8 XPS spectrum of self-assembled EDTA-siloxa ne lanthanide film after treatment with dipicolinic acid. The insert is a magnification of the N 1s region....................................95 6-9 XPS spectrum of self-assembled EDTA-siloxa ne lanthanide film after treatment with pH 9.5 picolinic acid. The insert is a magnification of the N 1s region...........................96 6-10 XPS spectrum of self-assembled EDTA-siloxa ne lanthanide film after treatment with pH 7 picolinic acid. The insert is a magnification of the N 1s region..............................97 7-1 Two common sensitizers of lanthanides are chelidamic acid a nd dipicolinic acid.........108
10 7-2 Scheme of different routes for a surf ace reactive chelidamic acid. The procedures used to prepare these compounds are s hown in the experimental section.......................109 7-3 Two possible routes of immobilizing the chelidamic acid/lanthanide complex onto a surface........................................................................................................................ ......110 7-4 Portion of an XPS spectrum where ligand 9 was allowed to react with Sm3+ and then the resulting complex was allowed to r eact with an epoxide surface. The XPS spectrum shows that 9 was contaminated with zinc, coming from the reduction process of the azide..........................................................................................................111
11 LIST OF ABBREVIATIONS AFM: Atomic Force Microscopy APTES: 3-Aminopropyltriethoxysilane DNA Deoxyribonucleic acid dsDNA: Double-stranded deoxyribonucleic acid EAL Effective Attenuation Length LB Langmuir-Blodgett NIST National Institute of Standards and Technology ODPA Octadecylphosphonic acid OTS Octadecyltrichlorosilane QCM Quartz crystal microbalance RNA Ribonucleic acid SAM Self-assembled monolayer SDS Sodium dodecylsulfate SPR Surface plasmon resonance SSC Saline sodium citrate buffer XPS X-ray photoelectron spectroscopy
12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy METAL-ORGANIC MONOLAYERS AS SUBSTRATES FOR BIOMOLECULE MICROARRAYS By Sarah Marie Lane August 2007 Chair: Daniel R. Talham Major: Chemistry It has previously been shown that a zi rconium-phosphonate surface can immobilize 5phosphorylated ssDNA and that this system can be used for DNA microarrays. These DNA microarrays had been studied with fluorescen ce imaging, but the probe surface density and hybridization efficiency was unknow n. X-ray photoelectron spectro scopy was used to study the surface density of DNA and dsDNA on a zirconi um-phosphonate substrate. A model was developed for the quantitative calculation of DNA surface density for the N 1s and Zr 3d peak intensities. It was found that a ssDNA probe wi th a poly-dG spacer gave a surface density of 2.8 1011 ssDNA molecules/cm2. While the same ssDNA probe with a poly-dA spacer gave a surface density of 1.4 1011 ssDNA molecules/cm2. It was also found that the surface density of the probe with the poly-dA spacer, unlike the probe with the poly-dG spacer, was highly dependent on the rinsing conditi ons. This indicated that the poly-dA spacer caused higher nonspecific adsorption. In additi on to the ssDNA studies, the use of the zirconium-phosphonate monolayer as a substrate for dsDNA microarrays was investigated, wh ere the dsDNA would act as a probe for proteins. To supplement this research, the surface c overage of dsDNA with phosphate linkers and poly-dG spacers on the 5 a nd 5ends; 3 and 3ends; and 3 and 5ends was calculated using the XPS data for the N 1s peak and Zr 3d peak. The surface coverage
13 results followed the same trend seen with fluor escence imaging of protein binding studies, where phosphate linkers on opposite ends of the dsDNA gave the highest surface coverages. A dsDNA sample with a random oligonucleotide spacer an d phosphate linkers on the 5 and 5 ends showed a lower surface density than the dsDNA with phosphate linkers and poly-dG spacers on the 5 and 5 ends, which also followed the trend seen in the fluorescence studies of the protein binding. Finally, this dissertation addre sses research aimed at making lanthanide monolayers, which could act as supports for protein microarrays. Monolayers of lantha nide phosphonates were prepared using the Langmuir-Bl odgett method, but it was found that these films were not stable in water. Using self-assembly of a siloxane m odified ethylenediamine triacetic acid/lanthanide complex a more robust monolayer was formed, wh ich was stable in water. Protein binding studies have not yet been performed on thes e more robust surfaces. Also, research was performed towards synthesizing a ligand wh ich would sensitize the luminescence and immobilize lanthanides on a surface by means of a chelidamic acid ligand containing an amine. This molecule was made, but the purification and use of this molecule was hindered by its poor solubility.
14 CHAPTER 1 INTRODUCTION: XPS ANALYSIS OF DNA Introduction As technology and science advance, resear chers have found many valuable reasons for immobilizing DNA on surfaces.1-5 One reason is to study and analyze DNA itself. For example, to code a genome, microarrays of DNA can be used.3 The general design of a DNA microarray is a surface covered with an array of many di fferent spots of known ss DNA sequences called the probe, which should be well attached to the surface. The surface is then exposed to a solution of unknown ssDNA sequences called the target, which ar e often fluorescently labeled. The target ssDNA binds the probes of complementary sequen ce, allowing the target sequences to be determined by following the probe-target interac tion with an analytical technique such as fluorescence imaging. One of the first examples of a DNA array used to study ssDNA interactions was given by Rease et al. This array was synthesized directly on a surface using photolithographic techniques and phosphoramidite ac tivated deoxynucleosides (protected at the 5'-hydroxyl with a photolabile group).3 The oligonucleotides were coupled together where the surface was exposed to light. Then selectiv e photodeprotection and coupling cycles were repeated until the desired sequences were obtai ned. Using this method they made a 1.28 x 1.28 cm array containing 256 spots. The spots were tested for hybrid ization specificity with target molecules. After this exampl e of in situ prepared oligonuc leotides, many other different methods for preparing ssDNA micr oarrays have been reported. Another use of immobilized DNA is for directed assembly. One of the first reports of using DNA for directed assemb ly came from Alivisatos et al. where they had derivatized 1.4 nm gold particles with a single DNA strand.1 Then using the specific in teractions of complementary DNA strands they created dimers and trimers of th e gold particles which ha d very specific inter-
15 particle distances based on the length of the DNA sequence. Similarly, Mucic et al. using complementary interactions be tween ssDNA on 31 nm Au nanopartic les and 8 nm Au particles, created a network of the larger partic les surrounded by the smaller particles.6 Another example of DNA directed assembly is seen in a paper by Becker et al. where ssDNA was spotted onto a substrate and then exposed to a solution of th e complementary strand modified with a protein.7 This second DNA strand bound to the complement creating a protein mi croarray which could then be used to probe protein-protei n interactions with mass spectrometry. These examples are just a drop in the bucket of the many uses scie ntists have found for surface immobilized DNA. With these many differe nt uses for immobilized DNA, the chemistry of the DNA on the great variety of surfaces plays an important role. Fo r instance, the surface density has been seen to a ffect hybridization efficiency8 and it has been observed that the different bases of DNA result in different non-specific binding.9 A number of methods have become more common for the quantitative surfac e analysis of DNA. These methods include surface plasmon resonance,8 radiometric quantification using 32P,10 quartz crystal microbalance11 and x-ray photoelect ron spectroscopy (XPS)12. Surface plasmon resona nce requires the DNA to be immobilized on either a gold or silver planar surface. Radiometric quan tification requires that the DNA be modified with a radioactive isotope. Quartz crystal microbalance again requires the DNA to be immobilized on a specific metal surfa ce like gold. Unlike th e other methods, with XPS, unmodified DNA on almost any surface can be analyzed. In addition, XPS can give both qualitative information about the chemical e nvironment of the different surface components through the shifts in the photoe lectron binding energies and al so quantitative information by looking at the peak areas of the photoelectrons.
16 The concept of XPS is fairly simple. A sa mple under ultra-high vacuum is bombarded with X-rays, usually from an aluminum or ma gnesium source, kicking out photoelectrons from the different energy levels of the atoms. The phot oelectrons are ejected in all directions, with some traveling towards the analyzer. They then pass through the analyzer which is adjusted to allow only photoelectrons with a specific kinetic energy to reach the detector. Thus the kinetic energies of the photoelectrons are known and w ith the known energy of th e X-rays the initial binding energy of the electron can be calculated. As the electrons travel towards the analyzer, th ey pass through the rest of the sample. As they travel through the solid, there is a certain probability they will collide with another atom, resulting in an elastic collision, where their trajectory is changed, or an inelastic collision, where they lose kinetic energy. The further the electr on must travel through the sample to reach the analyzer the higher the probabi lity it will undergo a collisi on. Thus the photoelectrons originating from deep within the sample are more likely to arrive at the analyzer with a kinetic energy less than its original kinetic energy. The photoelectrons that have lost kinetic energy do not contribute to their respective peak, but to th e background. Elastic colli sions increase the path length the photoelectron must travel to get to the analyzer, increasing the probability they will undergo a collision. Each electronic energy level of each element has an inherent efficiency with which it interacts with the X-rays called the photoelectri c cross-section; this also affects the peak intensity. The peak intensity is also dependent on the concentr ation of the material in the sample, which allows quantitative or semi-quantita tive information to be obtained from XPS. In addition to quantitative data, information about the chemical environment of the element can be drawn from slight shifts in the binding ener gy. This is because the electron binding energy
17 depends on the energy level, which can be affected by its chemical environment. For example, a photoelectron leaving an Fe3+ atom will have a higher binding energy than one leaving an Fe2+. As mentioned before, XPS is an analytical met hod that can be used on most solid surfaces. In 2004, our group published a paper showing that a zirconium-phosphonate surface can selectively immobilize 5-phosphorylated ss DNA over nonphosphorylated ssDNA and that this system can be used for DNA microarrays. Litt le was known about the nature of the DNA on the surface, such as the probe surface density or the hybridization efficiency. With the help of XPS analysis, we attempted to answer some of these questions. This PhD disse rtation discusses these XPS studies. This chapter introduces the topic of using XPS and outlines what other scientists have done in the way of studying DNA with XPS. In addition to studying DNA on a zirconium-phos phonate surface, research was performed in order to make lanthanide monolayers, which would act as substrates for protein microarrays and also which would provide a system to study th e luminescence of lantha nides in monolayers. More background information for these projects can be found in their respective chapters. Qualitative and Semiquantitative XPS XPS has been used a number of different ways to analyze DNA on a surface. One of the common uses is to simply look at a binding energy spectrum for the one of the elemental components of DNA, such as phosphorus or nitr ogen, to confirm that the DNA is present on the surface being studied.13-16 The binding energy of the N 1s photoelectron is around 400 eV while the P 2p peak has a binding energy around 133 eV. With a sufficient presence of DNA, both the N 1s and P 2p peak can be observed; however nitrogen is slightly more sensitive than phosphorus. Carbon and oxygen are not often us ed as an unconditional indicator for the presence of DNA due to the fact that they are often present on the surface either as contamination or a modification. This qualitativ e study often involves minimal analysis of the
18 peak shape or exact binding energy. For example, Wang et al. looked at the desorption of thiolated DNA on gold after applying a potential across the metal surface.15 They observed a loss of the N 1s peak and an increase in the Au 4f peak. The increase in the Au 4f peak means that there is less attenuation of this peak by the DNA overlayer. Another interesting example, involving the use of dsDNA to self assemble lysi ne-capped gold colloidal nanoparticles, comes from Kumar et al They used the P 2p peak as a positiv e indicator for the presence of DNA. According to the exact binding energy of the P 2p peak, they claimed no degradation of the DNA.14 XPS also easily offers semiquantitative results in the form of atomic percentages. The atomic percentages are calculated from either the peak intensity or peak area using a sensitivity factor for each energy level of each element. However, care need s to be taken when looking at the atomic percentages calculated this way b ecause attenuation of the peaks by elastic and inelastic collisions has not been taken into account. There are quite a few examples of scientists using semiquantitative XPS to study DNA. This is often combined with other studies such as an examination of the binding energies measured with high resolution.17-23 Frequently, when chemical information needs to be obtained from XPS by studying the binding energies, the element under analysis is present in more then one chemical environment. The multiple species give rise to more than one peak in XPS, which may be resolvable, but more often overlap. When the peaks overlap, peak fitting is often used to deconvolute the diffe rent species in the peak. An early example looked at the binding energy shif ts studied the interact ion of DNA with cisdichlorodiamine platinum(II).23 In this case, the DNA was not act ually on a surface. One of the earliest examples of using semiquantitative XPS to study DNA on a surface was given by Herne et al. who looked at thiolated and nonthiolated DNA self-assembling on gold.17 They observed
19 that post-spotting treatment w ith mercaptohexanol dramatica lly decreased the non-thiolated DNA N 1s peak, but only slightly decreased the th iolated DNA N 1s peak. They also explored different potassium phosphate buffer concentr ations. They saw, by following the N 1s normalized peak intensity, that increasing buffe r concentrations up to 1 M increased the DNA immobilization. Very recently, Lee et al. also looked at posttreatment with a mercapto-alcohol, mercaptoundecanol, on thiolated DNA adsorbed onto gold.19 They studied both elemental percentages and binding energies. They observed that posttreatme nt resulted in, not surprisingly, an increase in the elements associated with me rcaptoundecanol, such as carbon and sulfur. More interestingly, after treatment with the mercap toundecanol, the binding energies of most of the elements associated with the DNA shifted to lo wer binding energies espe cially the N 1s peak, which they thought indicated the DNA was standi ng upright and no longer interacting with the gold surface. Mohaddes et al saw similar behavior in their wor k. They reported shifts to lower binding energies while studying the influence of mercaptohexanol on thiol-modified DNA immobilized onto GaAs.21 On the other hand, Sastry et al. focused solely on studying the binding energies while looking at DNA immobili zed in thermally evaporated films of octadecylamine (ODA).22 The DNA was incorporated into the ODA by simply dipping the film into a DNA solution. The DNA was held in the film by the often-used elec trostatic interactions between the positively charged protonated amine and negatively charged backbone of the DNA. They looked at the C 1s, P 2p, and N 1s peaks with X PS. With peak fitting of the C 1s peak they found two components: one at 285.5 eV and anot her at 287.5 eV. They attributed the low binding energy component to the hydrocarbon peak of ODA and the DNA sugars and bases. The higher binding energy component was attributed to carbons coordinated to the oxygen in the backbone. They also looked at the P 2p peak which arises from the phosphate backbone. Its
20 binding energy at 132.8 eV, they report, is similar to others of immobilized DNA, which they say indicates there is no degradation of the DNA. An interesting study performed by Lee et al. which also used semiquantitative XPS and binding energies, in addition to time-of-flight secondary-ion mass spectrometry (TOF-SIMS), studied the variation in the purity of thiolated ssDNA from three different vendors: Alpha, TriLink, and Synthegen.18 They looked at the atomic percentage of phosphorus, carbon, nitrogen, oxygen, sulfur, and gold in the imm obilized DNA. In the films from Alpha and Synthegen, they obtained less intense P 2p and N 1s peaks and more intense Au 4f peaks, which indicated that the DNA from those two vendors self assembled with a lower density. In addition, the films made from the DNA from Synthegen a nd Alpha contained excess sulfur. The C 1s percentages were the same for the films from all three vendors However, upon studying the C 1s peak of the Synthegen and Alpha films with hi gh resolution XPS, they saw a larger portion of the peaks with a lower binding energy. The lo wer binding energy portion indicated an excess amount of C-C and C-H species. They proposed th at excess C-C and C-H plus the excess sulfur mentioned before was due to contamination with dithiothreitol, a reductant used to cleave disulfide precursors. With the help of TOF-SIMS they were able to confirm the contamination with dithiothreitol. They also looked at th e atomic percentages while varying the adsorption time and studied high resolution XPS scans of th e sulfur region while varying the adsorption time to gain insight into how the contaminati on affected the kinetics of immobilization. Quantitative The semiquantitative results offered by XPS are useful, but valuable more quantitative information can also be obtained with XPS. One important piece of information that can be calculated from XPS data is the thickness of an overlayer on a substrate. This is based on the
21 simple equation which relates the intensity I0 of a peak from an element in the bare substrate to the intensity I of the peak after it has been attenuated by an overlayer: ) sin exp(0 d I I (1-1) Here is the photoelectron mean free path, d is the thickness of the overlayer, and is the photoelectron take-off angle with respect to the su rface. The photoelectron mean free path is the average distance a photoelectron wi ll travel before it undergoes an inelastic collision. The mean free path depends on the kinetic energy of the phot oelectron and on the mate rial it is traveling through. Instead of the mean free path, an effec tive attenuation length ca n be used which takes into account elastic collisions as well. Equa tion 1-1 has been used in several instances to calculate the thickness of a DNA overlayer on a substrate.12, 24-26 Higashi immobilized dsDNA on a gold surface, which had been modified w ith a binary SAM with one of the compounds containing an acridine group (Figure 1-1).24 The acridine group can intercalate into the dsDNA leading to the dsDNA adsorption. Using Equation 1-1 and the Au 4f peak, they were able to calculate the thickness of the bi nary SAM (1.9 nm), and then the binary SAM plus the dsDNA layer (4.7 nm). From these thicknesses they estimated the dsDNA film thickness at 2.8 nm, which they say indicates that the dsDNA can be no more than a bilayer structure since the thickness of a dsDNA helix is 2 nm. Another way to take advantage of the relations hip in Equation 1-1 is to analyze the sample at multiple photoelectron take off angles. Leavitt et al. performed a similar study of ssDNA, which contained a sulfur atom replacing one of the oxygens in the phos phate backbone (Figure 1-2).27 However, they sought a depth profile of the atomic percentages which requires the following equation which relates the intensity Ii of a specific element i with the different parameters of XPS:
22 z z z c F Ii i i i id ) sin exp( ) (0 (1-2) In Equation 1-2, Fi is the spectrometer transmission function, i the photoionization cross section, and ci( z ) the concentration at depth z from the sample surface. The other variables have the same significance as in Equation 1-1. The exponential portion of th e equation takes into account the attenuation of photoele ctrons due to inelastic collisi ons. As is customary, Leavitt et al. used Equation 1-2 as a ratio of two elemental peaks. However, when using angle dependent data, Equation 1-2 cannot be solved direc tly. To get around this problem, Leavitt et al. treated the DNA layer as a series of flat homogeneous pa rallel layers and used an iterative method to find the atomic percentages of nitrogen, oxyge n, phosphorus, carbon, sulfur and gold at different depths. In their depth profile, they saw that in the top 5 carbon, ni trogen, and oxygen reached their maximum atomic percentages. Then at a depth between 2.5 to 10 sulfur and phosphorus reached their maximum atomic percentages, while Au increased steadily in this region. This depth profile indicates that the sulfur and phosphorus are oriented towards the gold as would be expected if the sulfur is bound to the gold. One of the most important works in quant itative XPS of DNA on gold was published by Petrovykh et al .12, 26 They looked at a thiolated 25-mer of thymine binding to gold. They first used Equation 1-1 to get the thickness of the film Then, as mentioned before, Equation 1-2 can be used to describe the ratio of concentrati on of gold to an element in the DNA, nitrogen or phosphorus: i Au Q i i i Q Au Au Au Au i Au iL z L z L F L F I I c c exp 1 exp (1-3) In this case, they used either quant itative effective attenuation lengths Q iL or average effective attenuations lengths AuL instead of photoelectron mean free paths. To calculate the
23 effective attenuation lengths, they used a progr am called Electron Effective-Attenuation-Lengths Database Version 1.0 written by Powell and Jablonski at NIST.28 In Equation 1-3, the iL z exp 1 portion of the equation accounts for the attenuation of the DNA film by itself. They could then use the density of gold to ca lculate the density of nitrogen and finally the surface coverage of the DNA, which they found to be 3.7 1013 DNA molecules per cm2 for a DNA immobilization time of 20 hours. Since Petrovykhs publication of this method for calculating the su rface coverage of DNA other reports have emerged with similar calculations. Saprigin et al. immobilized ssDNA homo 20-mers through a covalent linkage on amine-coated oxidized aluminum.25 To attach the DNA to the surface, they reacted phosphorylated DNA with carbodiimide to form a phosphoramidate linkage on the amine surface. They used the ratio of the Al 2p peak to P 2p peak and then the density of the aluminum to find the surface cove rage of the DNA. Their calculation, similar to Equation 1-3, must also take into account the attenuation of the Al 2p peak by the aluminum oxide layer, the amine modification as well as the DNA layer. The surface density they found for a guanine oligomer is 2.8 1013 DNA molecules per cm2. In addition to Saprigins paper, Shen et al. also used a similar method to quantify thiolated DNA immobilized on maleimidemodified fused silica. One difference in th eir method is that they sputter the DNA and maleimide modification off the silica support in situ to get a reading of the bare silica surface. They use the ratio of P 2p to Si 2p and Si 2s to perform the surface coverage calculation, but in their equation they use the intensity of the bare silica support negating the need to correct for the attenuation of Si signal by the DNA layer. The surface density they find under the optimum conditions is 9.2 1012 DNA molecules per cm2.
24 A final example involves the calcula tion of DNA surface coverage on silica nanoparticles.29 The nanoparticles were first coated w ith an amine and then with 6-hexanedioic acid to create a carboxylic acid terminated surface which could then bind an amine modified oligonucleotide. They followed each step of the silica nanoparticle modification and DNA immobilization using XPS. To determine the D NA surface coverage, first the surface coverage of amine on the nanoparticles was determined us ing a ninhydrin fluorescence test. Then when the DNA was bound, they assumed that any increas e in the N 1s signal was from the DNA and they could calculate the surf ace coverage of the DNA as 4.7 107 DNA molecules per bead. Conclusion This chapter offers an overview of the possi bilities for using XPS to investigate DNA. However, XPS has its limitations and the data sh own in most papers is often supported by other independent analytical techniques. Furthermor e, some authors reporting quantitative results caution the readers that the data should be ta ken to be only semiquantitative because of the assumptions that must be made when performing the calculations for overlayer thickness and surface coverage.27 O O N H N H S N S S O O Br S S Figure 1-1. Compounds used by Higashi et al. in their in binary SAM to immobilize DNA.
25 O B A S E O O P OO S Figure 1-2. DNA with sulfur modifi ed phosphate backbone used by Leavitt et al.
26 CHAPTER 2 USE OF XPS TO STUDY OLIGONUCLEO TIDES AT A ZIRCONIUM-PHOSPHONATE SURFACE Introduction Oligonucleotide microarrays, which allow hi gh throughput, highly para llel investigation, have been used to probe a number of different biological phenomena in cluding the analysis of gene expression.30 There are many platforms on the market which are used for DNA and RNA microarrays and with each new platform there are often different protocols. The oligonucleotide probes can be prepared either in situ for which the DNA or RNA sequences are synthesized directly on the surface,31 or ex situ where complete sequences are spotted on the surface.30 Our group recently reported that glass slid es modified with a zirconium-phosphonate surface layer are effective substrates for cova lently immobilizing oligonucleotides for array applications.32 This concept, illustrated in Figure 1, uses specific metal/ligand interactions to covalently attach the oligonucleotide to the surface. Zirconium-phosphonate surfaces are known to strongly immobilize phosphate an d phosphonate functionalized molecules.33, 34 For example, these surfaces can be used to prepare self-assembled m onolayers of densely-packed alkylphosphonates and other f unctional organophosphonates.35 In a related fashion, phosphate terminated oligonucleotides can be immobili zed and we demonstrated the zirconiumphosphonate surface selectively binds phosphorylat ed oligonucleotides over nonphosphorylated oligonucleotides. Some of the surface chemistry parameters a ffecting the performance of oligonucleotide arrays include the density of the probe, th e non-specific adsorption of DNA or RNA to the surface, and the distance of the probe from the surface. For example, it is now well documented that a high probe surface density can lead to low hybridization efficiencies and slow hybridization kinetics.8, 36 Also the use of a spacer between the linker and the probe, which is
27 surmised to lift the DNA off the surface and preven t steric hindrance caused by the surface, can increase hybridization efficiencies.37 To study the surface chemis try of immobilized probes several different quantitative t echniques have been used including surface plasmon resonance (SPR), quartz crystal microbalance(QCM) and X-ray photoelectron spectroscopy (XPS).26 XPS, traditionally a materials science analy tical technique, has been shown to be an efficient, label-free method for studying DNA on a surface. Leavitt et al. used angle-resolved XPS to study DNA with a thiol-modified backbone and determine its or ientation on the surface of gold.27 More recently, Petrovykh et al. used single-angle XPS w ith a substrate-overlayer model to determine the density of thiol-modified DNA on gold.12 It has also been demonstrated that XPS can be used to distinguish between the different DNA and RNA bases.25 In this thesis, we report ou r initial results quantifying DNA attached to a zirconiumphosphonate support using XPS. Using the probe spotting conditions optimized for microarray applications we immobilized phosphorylated ssDNA on the surface and quantified the surface density using XPS and a modified substrate-overlayer method. Experimental Materials Glass substrates were purchased from Gold Seal Products. Oligonucleotides were ordered lyophilized with HPLC purity from Invitrogen (Carlsbad, California). The ssDNA sequence used, which includes a spacer consisting of nine guanines, was 5-H2O3PO-(G)9CCGCCGGTAACCGGAGGTTAAGATCGAGATCCA ( PO4G9O33). Reagents were of analytical grade and used as received from co mmercial sources, unless indicated otherwise. Zirconium-Phosphonate Substrates Hydrophobic glass slides were made using oc tadecyltrichlorosila ne (OTS) following a method by Sagiv.38 The glass slides were clea ned with piranha etch (3:1 H2O2:H2SO4, boiling 20
28 minutes) and followed by the RCA method (4:1:1 H2O2:NH4OH:H2O, boiling 20 minutes and then 5:1:1 H2O2:HCl:H2O, boiling 20 minutes). Then they were made hydrophobic by treating with a 5mM solution of octadecy ltrichlorosilane (OTS) in bicy clohexyl for 2 minutes, rinsing with toluene for 30 seconds, and then drying with N2. The slides were treated again with the OTS, rinsed with toluene, and finally dried with N2.22 A KSV 2000 LB double-barrier Teflon tr ough, supplied through KSV Instruments (Stratford, CT), was used to form the ODPA monola yers at the air-water interface. A filter-paper Wilhelmy balance attached to a KSV microbalance was used to measure surface pressure. The zirconium-phosphonate films were made by first spreading a 0.3mg/mL solution of ODPA in chloroform on a 2.6 mM aqueous CaCl2 subphase, which had a pH adjusted to 7.8 using a KOH solution. The ODPA was compressed at the rate of 10 mm/min to a pressure of 20 mN/mm. Once the target pressure was reache d, a hydrophobic glass slide was dipped 50mm into the subphase at a rate of 8 mm/min. The slide wa s then lowered into a gl ass vial in the trough. The vial with the slide was removed from th e trough and an amount of a zirconyl chloride (ZrOCl2) solution was added to reach a concentration of 3 mM of Zr4+ in the vial with the slide. The slides sat for 7 days in the Zr4+ solution. If the slides were not used right away, they were stored in water until later use. DNA Immobilization To immobilize the oligonucle otides to the zirconium-phos phonate surface, first the appropriate amount of water was added to make a 100 M solution and then th e oligonucleotides were aliquoted into smaller one-time-use volumes The aliquoted oligonucleotides were stored in a freezer at -20 C. The ssDNA were prepared in a 1 x SSC (saline sodium citrate) buffer, pH 6 at a concentration of 40 M. The saline sodium citrate buffer solutions are prepared from a
29 stock solution of 20 x SSC, which corresponds to a 3.0 M NaCl and 0.30 M sodium citrate solution. Thus for a 1 x SSC solution, the concentrations are 0.15 M NaCl and 0.015 M sodium citrate, creating a solution which has an ionic strength of 0.225 M. To create a spot large enough for XPS analysis, 30 L of the DNA was pipetted onto th e rinsed and dried zirconiumphosphonate surface. Once the DNA had been spotte d, the slides were incubated overnight in Petri dishes at room temperature. Then the s lides were submerged for 1 hour in 3.5 x SSC, 0.3% SDS at 42 C, followed by rinsing 5 times with nanopure water and spin drying. The slides then underwent a mock hybridization, which should give th e true probe concentra tion if they were to undergo hybridization, by treatment with 25 L per spot of 3 x SSC, 0.1% SDS overnight at 42 C. Finally, the slides were rinsed in 2 x SSC, 0.1% (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min), followed by rinsing in water 5 times. XPS was performed using a UHV XPS/ESCA PHI 5100 system. Survey scans and multiplex scans (Zr 3d, P 2p, and N 1s) were taken with a Mg K X-ray source using a power setting of 300 W and a take off angle of 45 with respect to the surface. Survey scans were taken for all samples with a pass energy of 89.4 eV a nd multiplex scans were taken with a pass energy of 22.36 eV. Using commercial XPS analysis so ftware and Shirley background subtraction, the peak areas were determined. Four different spot s were analyzed to determine the scatter of the data. The analyzer transmission function, which is necessary for the surface coverage calculations, was determined using a method by Weng.39 To find the analyzer transmission function with this method, the XPS intensities of the C 1s, O 1s, Zr 3d, and Zr 3p3 peaks of a zirconium-phosphonate substrate were analyzed at nine different pass energies: 179.0, 143.0, 89.5, 71.5, 44.7, 35.7, 22.4, 18.0 and 11.2 eV.
30 Results and Discussion The Zirconium-Phosphonate Surfaces Zirconium-phosphonate surfaces can be made a number of ways, including adsorption of Zr4+ onto phosphorylated groups covalently attached to silica or gold, or also by adsorption of Zr4+ onto organicphosphonic acid monolayer s prepared by Langmuir-Blodgett (LB) deposition.40-43 We have found that zirconium monolayer s made using the LB technique provide a reproducible, highly-active surface for the binding of phosphate and phosphonate containing molecules.43 As shown in Figure 2-2, to prepar e the zirconium-phosphonate monolayers using the LB technique, octadecylphospho nic acid (ODPA) is spread at the air-water interface, is compressed into a monolayer, and then transf erred onto a hydrophobic substrate. The ODPAcoated substrate is then e xposed to a solution of ZrOCl2. The Zr4+ binds strongly to more than one phosphonate group creating a robust surface that is stable in water for months. X-ray Photoelectron Spectroscopy In a typical instrument setup for XPS, X-rays impinge a sample causing photoelectrons to be released from the core orbitals. The photoel ectrons are then retarded to a certain kinetic energy called the pass energy and then they enter th e analyzer and finally th e detector. Retarding the electrons does not change the absolute energy spread, but does increase resolution. Differences in the pass energy and also the photoele ctron kinetic energies can causes variations in the analyzer and detector efficienci es. An analyzer transmission function, T can be used to correct these variations. In a paper by Weng et. al., several methods were laid out for determining the analyzer transmission function.39 The most suitable method for our instrument was chosen. This involved using XPS p eaks at different ki netic energies ( Ek) taken with different pass energies. For this the C 1s, O 1s, Zr 3d, and Zr 3p3 peak intensities for a zirconium-phosphonate slide were obtained at nine different pass energies ( Ea): 179.0, 143.0,
31 89.5, 71.5, 44.7, 35.7, 22.4, 18.0 and 11.2 eV. The intensity data, I for a given kinetic energy was scaled with a multiplicative factor to take in to account the different inherent sensitivities for the different photoelectrons of the elements. The data was then used to make a log(I/Ea) vs log (Ea/Ek) plot, shown in Figure 2-3. The following third order polynomial equation was used to describe the trend of the data: C E E E E E E E Ik a k a k a a log 9387 0 log 4708 0 log 0222 0 log2 3 (2-1) When Equation 2-1 is solved fo r the intensity, the analyzer transmission function is obtained: n k a aE E E T (2-2) where 2log 0222 0 log 4708 0 987 0k a k aE E E E n (2-3) Using Equations 2-2 and 2-3; the kinetic energy of the Zr 3d phot oelectron, 1067 eV; and the pass energy of 22.36 eV, the analyzer transmi ssion function for Zr was found to be 31.2 eV. Similarly with the kinetic energy of the N1s photoelectron, 851.6 eV, and again an analyzer energy of 22.36 eV, the analyzer transmission fu nction for the N 1s photoelectron was found to be 37.1 eV. The analyzer transmission function for Zr and N are used in the calculations for the surface density of the DNA, as shown later. Substrate-Overlayer Model XPS is often used to determine the thic kness and surface coverage of thin films by quantifying the attenuation of the substrate photoelectrons as they pass through the thin film on their way to the analyzer. This technique, calle d the substrate-overlayer model (shown in Figure 2-4 A), has been used several times to quan tify the surface coverage of DNA on a substrate.25, 26 The calculation generally uses the intensity of an infinitely thick substrate and a finitely thick overlayer. The intensity of an elemen tal peak coming from the substrate, Is, can be written as
32 sin exp sineff S S S S S SL t L T FN I (2-4) or sin expeff S S SL t I I (2-5) where F is the X-ray photon flux; NS is the atomic density of the element; S is the effective cross section of the photoelectron for the specific energy level of the element; TS is the analyzer transmission function of the spectrom eter for the given kinetic energy; LS is the effective attenuation length (EAL) of the subs trate photoelectron as it travels through the substrate; t is the overlayer film thickness; eff sLis the EAL of the substrate photoe lectron as it trav els through the overlayer; is the angle of the photoele ctron detection with respect to the sample surface; and SI is the intensity of the bare substrate. Scof ield coefficients are usually used for the cross section of the photoelectron.44 Equation 2-4 can be simplifie d to Equation 2-5, and if the intensity of the bare substrate, SI is known or measured, the thic kness of the overlayer can be determined. Similarly, the intensity of a peak coming from an element in the overlayer, IX, can be written as sin exp 1 sineff X X X X X XL t L T FN I (2-6) where, NX, X, TX, LX, and eff XL signify the same parameters as in Equation 2-4 except with respect to an element from the overlayer. E quations 2-4 and 2-5 can be combined, cancelling some of the instrumental parameters, to give th e atomic density ratios of an element from the substrate and from the overlayer: sin exp 1 sin expeff X X X X S eff S S S S X S XL t L T I L t L T I N N (2-7)
33 As shown by Petrovykh et al. ,12 Equation 2-7 can be modified using the relationship of the atomic density, NX, to the atomic surface coverage, nX, and the overlayer thickness: t n NX X (2-8) When this is combined with Equation 2-7, the ra tio of overlayer surface coverage to substrate density is obtained: sin exp 1 sin expeff X X X X S eff S S S S X S XL t L T I L t t L T I N n (2-9) In the quantitative determination of DNA on a substrate such as gold, X could be the N 1s peak, or P 2p, but probably not C 1s or O 1s due to possible contaminants on the surface also containing these elements. While S can be any element from the substrate, it is best to use a peak specific to the substrate. Using the known atom ic density of the chosen substrate element and the number of nitrogens in the DNA strand, the surface coverage of the DNA can be calculated. Model for the Zirconium-Phosphonate Surface Our system, as shown in Figure 2-4B, requires a slightly differe nt approach due to the fact that we use the zirconium-phosphon ate layer as our modifying subs trate. We choose to use the Zr 3d peak as the substrate signal instead of the silicon XPS peaks, because these are strongly attenuated by the zirconium-phosphonate modifyi ng layer. The zirconium coverage corresponds to a monolayer and is therefore assumed not to attenuate itself.43 Thus the intensity of the Zr 3d peak attenuated by the DNA overlayer can be written as sin exp sineff Zr Zr Zr Zr ZrL t T Fn I (2-10) Equation 2-10 can then be used with Equations 2-7 and 2-8 to give sin exp 1 sin expeff N N N N Zr eff Zr Zr Zr N Zr NL t L T I L t t T I n n (2-11)
34 Using the number of nitrogens in the DNA sequence and the known surface coverage of zirconium, the DNA coverage can be calculated. The surface coverage of zirconium, 4.2 1014 atoms/cm2, was previously determined from the 24 2 cross sectional area of the zirconiumphosphonate sites in the LB monolayer.43 To calculate the EALs needed for Equation 2-11, the NIST SRD-82 software was used.28 This program requires the input of several pa rameters including the kinetic energy of the photoelectrons, the stoichiometry of the DNA, the band-gap energy of the DNA,45 the film density of DNA, 46 and the instrumental geometry. The software28 average practical EAL output number was used for eff NL and eff ZrL and the output number of the EAL for quantitative analysis was used for NL The EAL values were calculated as 25.5, 31, and 27 for eff NL eff ZrL and NL respectively. After spotting the probe on the zirconium-phos phonate modified slides and performing the washing procedures, survey scans and multiplex scans of the Zr 3d binding energy (BE) peak region (194 180 eV), P 2p3 peak region (143 131 eV) and N 1s (143 131 eV) were obtained on the DNA spot areas. The average peak areas after Shirley baseline subtraction are shown in Table 1. Nitrogen, which is more sensitive in XPS analysis than phosphorus and also more abundant in DNA than phosphorus, is a good choice as the element from the DNA overlayer. Furthermore, phosphorus is pres ent in the zirconium-phosphonat e films and would complicate the calculations if it were used. For these reasons, the N 1s and Zr 3d peak areas were used in Equation 2-11 to calculate the nitrogen to zirconi um surface coverage ratio. This ratio was then used with the number of nitrog ens per DNA strand and the known surface coverage of zirconium to calculate the surface coverage of the probe.
35 The average of the measured peak intensities are given in Table 1. With these intensities and Equation 2-11, the probe surface density can be calculated. In Equation 2-11, the Scofield coefficients for N 1s,N, is 1.77 and for Zr 3d, Z, is 7.3. As explained previously, Equation 2-5 can be used to calculate the thickness. To pe rform this calculation, the intensity of the Zr 3d peak from a bare zirconiumphosphonate substrate is needed, which was found to be 11800. Finally, using the average N 1s peak intensity of 691 and Zr 3d peak intensity, 11618, coming from the substrate with DNA, the following calculation can be performed: 45 sin 5 25 25 0 exp 1 27 eV 1 37 77 1 11618 5 4 sin 1 31 25 0 exp 25 0 eV 2 31 3 7 691 Zr Nn n 0.0875 Zr Nn n 11 1 1410 4 2 coverage surface DNA molecule DNA per atoms N 177 cm atoms Zr 10 4.2 0.0875 coverage surface DNA The measured peak intensities and corresponding surface coverage are given in Table 1. The calculated coverage from these measur ements of ssDNA on the surface is 2.8 1011 DNA molecules/cm2. This surface coverage is lower than wh at other researchers have measured for DNA immobilized on gold. For example, th e average maximum ssDNA coverage found by Georgiadis et al. on gold was 12 1012 molecules/cm2 and Petrovykh et al. found a maximum coverage on gold using XPS as 37 1012 molecules/cm2.8, 12 Despite some differences in immobilization procedures, such as the salt concentration in the probe buffer solution,47 there seem to be other factors, i nherent to the zirconi um-phosphonate surface, affecting the probe surface coverage. On the other hand, others have reported similar coverages. Zammatteo et al. studied the immobilization of DN A through a number of covalent coupling reactings including
36 the immobilization of amine-modified DNA on an aldehyde surface.48 They used scintillation counting of 32P-labeled ssDNA strands and found a maximum surface coverage of 3.6 1011 DNA molecules/cm2. They were, however, using much l onger strands, 250 bp, compared to our 42 bp, which might cause them to have lower surface coverages. From fluorescence microarray studies comp aring hybridization using nonphosphorylated and phosphorylated probes, it is seen that the phosphodiester bac kbone of the oligonucleotides does not covalently bind to the zircon ium surface for permanent immobilization.32 Nevertheless, once the oligonucleotides are covalently bound through the 5 phosphate, the phosphodiester backbone can physisorb to the su rface causing the oli gonucleotide to lie do wn on the surface, kinetically hindering further attach ment of oligonucleotides. This observation is consistent with film thickness measurements that show far less then complete m onolayer coverage. Therefore, although other molecular systems such as alkylp hosphonate molecules self-assemble into closepacked monolayers on the zirconium-phosphonat e surface, the phosphorylated DNA does not. Although the measured DNA coverage appears lo w, these conditions have been shown to be useful for array applications. The XPS qua ntification of the DNA surface coverage gives a reasonable answer. There are, however, several possible sources of random and systematic error in these experiments. The sources of random erro r include slight differen ces in the preparation of the sample and also error coming from the inst rument, such as change s in the X-ray flux and slight differences in the placement of the sample in the XPS chamber. The random error can be seen in the standard deviati on of the surface coverage, which is around 20%. This error can perhaps be improved by improving the c onsistency of the rinsing steps. The systematic error can come from a number of factors, including the surface coverage of zirconium ions and the choice of EALs. The intensity of the bare zirconium peak is taken from
37 an XPS of a rinsed unmodified zirconium-phosphona te surface. The zirconium-modified surface most likely has a monolayer of oxides and hydrox ides on the surface, which would be displaced by the DNA. The oxides and hydroxides could atte nuate the zirconium peak thus the measured peak area of the Zr 3d peak would be systema tically lower than the true value of the bare substrate. If this were true, it would give rise to a systematically lower DNA film thickness than the true value. Because in Equation 2-11, the EALs and overlayer thickness are always taken as a ratio to each other, the equation is not very sensitive to the valu e of the film thickness or EALs, and would only affect the surface coverage slightly. Conclusions The surface coverage of ssDNA on a zircon ium phosphonate monolayer was calculated from XPS N 1s and Zr 3d peak intensities. The surface coverage of ssDNA is lower than reported values for thiol-modified DNA on gol d. The DNA immobilization conditions were those optimized using fluorescence imaging a nd that indicated maximum DNA adsorption. Based on these findings, it indicates that, alt hough the phosphodiester ba ckbone cannot hold the DNA on the zirconium throughout the rinsings, an interaction between th e zirconium surface and the backbone may cause the DNA to lay down on th e surface, thus limiting the covalent binding of more DNA during spotting.
38 Figure 2-1. Illustration of the immobilization of the phosphorylated ssDNA and subsequent hybridization. Figure 2-2. Procedure for making the zirconi um-phosphonate monolayers. First the ODPA is spread at the air-water interface and th en transferred onto a hydrophobic support. This followed by the addition of zirconyl chloride which forms a network with the phosphonate headgroups. O P O O O P O O ZrZrZr 5 3 Zr Zr ODPA monolayer with Zr4+ bound Probe TargetO P O O O P O O O P O O O P O O O P O O O P O O ZrZrZr 5 3 Zr Zr ODPA monolayer with Zr4+ bound Probe TargetO P O O O P O O O P O O O P O O Zr4+( ) STEP 1: Transfer template from water surface STEP 2: Add Zr4+ CH3(CH2)17PO3H2 CH3(CH2)17PO3H2 Zr4+( ) Zr4+( ) STEP 1: Transfer template from water surface STEP 2: Add Zr4+ CH3(CH2)17PO3H2 CH3(CH2)17PO3H2 CH3(CH2)17PO3H2 CH3(CH2)17PO3H2
39 Analyzer Transmission Functiony = 0.0222x3 0.4708x2 0.9387x + 2.41622 2.2 2.4 2.6 2.8 3 -2.5-2-1.5-1-0.50 log(Ea/Ek)log(I/Ea) Figure 2-3. Graph of the anal yzer transmission function. Ea is the analyzer energy (pass energy) and Ek is the kinetic energy of the photoelectr on. The diamonds are the scaled data points. Figure 2-4. Parameters that are used to calcu late the DNA surface coverage. A) The parameters for the gold-DNA system. B) The para meters for the zirconium-phosphonate-DNA system. t NLeff NLeff ZrL Intensity ratio: ) sin ( exp ) sin ( exp 1 eff Zr Zr Zr Zr eff N N N N N Zr NL t T n L t L T N I I DNA Zr ODPA OTS Glass t NLeff NLeff AuL Intensity ratio: DNA Au AuL ) sin ( exp ) sin ( exp 1 eff Au Au Au Au Au eff X X X X X Au XL t L T N L t L T N I I AB t NLeff NLeff ZrL Intensity ratio: ) sin ( exp ) sin ( exp 1 eff Zr Zr Zr Zr eff N N N N N Zr NL t T n L t L T N I I DNA Zr ODPA OTS Glass t NLeff NLeff AuL Intensity ratio: DNA Au AuL ) sin ( exp ) sin ( exp 1 eff Au Au Au Au Au eff X X X X X Au XL t L T N L t L T N I I AB
40 Table 2-1. Intensities of the N 1s and Zr 3d peaks for a zirconium surface with and without DNA. The table also includes the surface cove rage of DNA that was calculated with the peak intensities. 11800 11618 Zr3d Area Without ssDNA 2.8 1011 0.5 1011691 With ssDNA DNA Surface Coverage (ssDNA/cm2) N 1s Area 11800 11618 Zr3d Area Without ssDNA 2.8 1011 0.5 1011691 With ssDNA DNA Surface Coverage (ssDNA/cm2) N 1s Area
41 CHAPTER 3 SPACER AND RINSING EFFECTS ON SSDNA SURFACE COVERAGE Introduction One of the many parameters changed to impr ove the performance of DNA microarrays is the use of a spacer between the prob e sequence and the surface-linking group.32, 37, 49, 50 The spacer, which creates distance between the linking group and the probe sequence, is used with many different DNA microarray platforms. It is thought to lift the probe off the surface to bring it more in contact with the solution phase. With our zirconium-phosphon ate surface, we have shown that when a stretch of at least 5 guanines are used as th e spacer, compared to the other nucleotides, increased hybridization is seen.32 Beyond 7 guanines there is no further increase in hybridization. We call this in crease in hybridization when usi ng the poly-guanine spacer the poly-dG effect. This chapter discusses the surface coverage measurements made with XPS of a probe sequence with either a poly-guanine or poly-adenine spacer. This da ta allows us to determine if the poly-dG effect is due to a difference in th e surface coverage or an increase in hybridization efficiency. Also in this chap ter are details on rinsing experime nts which give insight into nonspecific binding of these two sequences. Experimental Zirconium-phosphonate substrates we re prepared as stated in the experimental section of Chapter 1 of this dissertation. The general ssDNA sequence used was 5-H2O3PO-(X)9,12CCGCCGGTAACCGGAGGTTAAGATCGAGATCCA. In the sequence, X represents either guanine or adenine, which was e ither 11 or 9 nucleotides long. The ssDNA were prepared in a 1 x SSC (saline sodium citrate) buffe r, pH 6 at a concentration of 40 M. To create a spot large enough for XPS analysis, 30 L of the DNA was pipetted onto the rinsed and dried zirconium-
42 phosphonate surface. Once the DNA had been spotted, the slides were incubated overnight in Petri dishes at room temperature. Several diffe rent rinsing conditions we re explored before XPS was taken. In one set, which we are calling Rins ing 1, the slide was submerged successively in 2 x SSC, 0.1% SDS (sodium dodecylsulfate) (2 min) 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min), followed by dipping twice in water. Rinsing 1 experiments were done with a spacer 11 nucleotides long. Another set of rinsing conditions, Rinsing 2, be gan with the spotted slide being immersed for 1 hour in 3.5 x SSC, 0.3% SDS at 42 C, followed by rinsing 5 times with nanopure water and spin drying. A final set of rinsing c onditions was explored, Rinsing 3, which used the 1 hour submersion in 3.5 x SSC, 0.3% SDS at 42 C, followed by rinsing 5 times with nanopure water and spin drying. These slides then underw ent a mock hybridization, which should give the true probe concentration if they were to undergo hybridi zation, by trea tment with 25 L per spot of 3 x SSC, 0.1% SDS overnight at 42 C. Finally, the slides were rinsed in 2 x SSC, 0.1% (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min), followed by rinsing in water 5 times. At least four spots were examined at each different rinsing condition. Streaming surface potential measurements were performed using an asymmetric clamping cell on a Paar Physica Kinetic Analyzer. An excel lent description of this system is given by Walker et al .51 A 1 mM KCl solution with a pH of 6 was used as the electrolyte solution. A zirconium phosphonate slide was caref ully clamped (the film can be easily distur bed during this process) onto the cell and 16 measuremen ts were performed per slide. Results and Discussion There are a several different types of surf aces employed for DNA microarrays that use exsitu prepared DNA. These surfaces can use th e non-specific interactions between the DNA and the surface, such as in the case of the negatively-charged DNA phosphodiester backbone binding
43 to the positively charged amine-coated surface.30 On the other hand, the surface can use a specific interaction between a linker on the DNA a nd the surface; for example, a thiol-modified DNA molecule binding to a gold surface.17 With each different surface, there are different protocols for immobilization, ri nsing, and hybridization, the c onditions of which are usually suggested by the manufacturer. The different surfaces using specific interactions require different linkers. We use a zirconium-phosphonate surface, which uses a coordinate covalent bond between a 5-phosphorylated DNA and the zirconium-phosphonate su rface to immobilize the DNA. The two parameters explored here, th e spacer identity and the rinsing conditions, both make a considerable difference in the amount of DNA on the surface. The two seemingly discrete parameters are discusse d together in the same chapter because one affects the other. The use of a spacer on the probe DNA began ar ound the same time as the advent of the DNA microarray.52 The spacer is a moiety added between the linker and probe sequence, which is found to give increased hybridization. It is believed that the spacer lif ts the probe sequence off the surface so that it has more c ontact with the solution phase. There are a number of different spacers used depending on the type of surface. For DNA microarr ays on gold, a simple alkyl chain, often 6-carbons long, is frequently used.8 On the other hand, a stretch of different nucleotides can also be used. Guo et al studied the immobilizati on and hybridization of 5amine-functionalized probes on an isocyanate-func tionalized surface.52 They used poly-dT spacers of 0, 3, 6, 9, 12, or 15 nucleotides long. When no spacer was used, they observed no hybridization, but starting with 9 nucleotides an increase in hybridization was seen, which continued up to 15 nucleotides. However, they did not study spacers longer than 15 nucleotides. On the other hand, Shchepinov et al. built spacers from a number of different monomers using phosphoramidite chemistry, by condensation ont o an amine-functionalized polypropylene
44 support.37 They found that using a spacer of abou t 40 atoms gave a 150-fold increase in the hybridization. For our surface, fluorescence imaging was used to study hybridization when a stretch of adenine, guanine, thymine, or cytosine was adde d as spacers between the 5 phosphate linker and the probe sequence.32 It was observed that the use of a poly-dG spacer over any other nucleotide or no spacer gave increased hybridiz ation, as shown in Figure 3-1. This demonstrated the ability of a poly-dG spacer to give increased hybridi zation. The length of the spacer was also investigated, and it was found that beyond 7 gua nines, no further significant increase in hybridization was seen. It rema ined unknown whether the poly-dG sp acer caused an increase of probe on the surface or an increase in hybridizatio n efficiency. However, it was suspected that this was related to the ability of oligomers of gua nines to form quadruplex structures as shown in Figure 3-2, which is seen in the DNA of telomeres.53, 54 In this structure, which is seen in the DNA of telomeres, the guanine bases form a hydr ogen bonding scheme, as seen in Figure 3-3, where N(1) and N(2) act as hydr ogen donors to O(6) and N(7). To determine if the poly-dG effect was caused by an increase in probe density, probes with e ither a poly-dG or poly-dA were studied with XPS. The other parameter investigated here, rinsi ng, is a critical step in the DNA microarray process. After the DNA is spotted onto a su rface, the surface is gene rally rinsed before hybridization. The rinsing solutio ns often contain a buffer, such as SSC, and a detergent such as SDS. A rinsing solution that contains less salt and more surfactant is considered more stringent. The slide may be heated or not during the rinsin g process. There are only a few papers that investigate the importance of rinsing.55, 56 One paper demonstrated that automated rinsing compared to manual decreases vari ability across microarray slides.55 Another study showed that
45 more stringent washing cond itions improved signal to ba ckground ratios; however they concluded that the hybridization step played a greater role in improving signal to background ratios.56 A passivation step can also be used in be tween the rinsing and h ybridization. This employs a molecule, such as bovine serum album in (BSA), which binds to the surface where there is no probe. The BSA prevents the target molecule from physisorbing to the surface, ultimately improving the signal to noise. After a rinsing step, the passivation step and the hybridization can rinse o ff more probe. The experiments disc ussed in this ch apter attempt to take into account the rinsing from passivation and hybridization. Three different rinsings were investigated in this study. Rinsing 1, involves successively dipping the slide in 2 x SSC, 0.1% SDS (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min), successively, followed by di pping twice in water. In Rins ing 2, which is actually a modified passivation step, the slide is submer ged in a 3.5 x SSC, 0.3 % SDS solution at 42 C for 1 hour. Normally, this solution would contain BSA, but it was found that the BSA physisorbed onto the surface, causing nitrogen contamination. This contamination would hinder efforts to determine the DNA coverage based on the nitrogen signal in XPS. Rinsing 2 is more stringent than Rinsing 1 because of the increased surfactan t concentration, the heating and the increased time the slide stays in the solution. Also the in creased number of water rinsings at the end of Rinsing 2 raises the stringency. In Rinsing 3, again the slide is submerged in a 3.5 x SSC, 0.3% SDS 42 C for 1 hour, then the spot area is subjected to a step which is supposed to mimic a hybridization step. In this blank hybridizat ion step, the spot area is treated with 25 L of 3 x SSC, 0.1% SDS overnight at 42 C. Often the hybridization so lution contains formamide and other additives which are supposed to reduce non -specific interactions, but these molecules contain nitrogen and again it was found that th ey physisorbed onto the surface contributing to
46 nitrogen contamination. Due to the increased length of the rinsings and heating, Rinsing 3 should be more stringent than Rinsing 1 and 2. To investigate how these two different parame ters, rinsing and spacer identity, affected the surface coverage, XPS was taken af ter each of the different rinsing conditions while using a 5phosphorylated probe molecule containing a po ly-dA spacer and one containing a poly-dG spacer. To improve the statistics of the data, at least 4 different spots were analyzed for each different condition. The N 1s and Zr 3d peaks we re used to calculate th e surface coverage using the model shown in Chapter 2. A comparison of the average surface coverage s of the poly-dA cont aining probe and polydG containing probe is shown in Figure 3-4. As the graph demonstrates, while the surface coverage of the poly-dG containing probe stayed relatively constant with the different rinsing procedures, the surface coverage of the probe containing the pol y-dA spacer decreased with increasing rinsing stringency. With the less stringent SSC rins ing conditions (Rinsing 1) the calculated surface coverage of the poly-dA, 6.5 1011 6.2 1010 ssDNA molecules/cm2, is twice that of the polydG containing probe, 3.2 1011 3.5 1010 ssDNA molecules/cm2. After Rinsing 2, the surface coverages are almost equal, with the surface coverage of the poly-dA containing probe at 2.6 1011 4.8 1010 ssDNA molecules/cm2 and that of the poly-dG containing probe at 2.6 1011 5.9 1010 ssDNA molecules/cm2. Finally after rinsing with Rinsing 3, the surface of coverage of the poly-dG containing probe, 2.8 1011 5.1 1010 ssDNA molecules/cm2, is twice that of the su rface coverage of the polydA containing probe, 1.4 1011 3.4 1010 ssDNA molecules/cm2. The error shown with th e surface coverages is the standard deviation of the mean.
47 The final probe coverage after Rinsing 3, s hows that the poly-dG containing probe has a higher surface coverage than the poly-dA containi ng probe after the mock hybridization. From this we can conclude that the reason an increas e in hybridization is seen when a poly-dG spacer is used is due to a higher probe surface density. Although this is the first report of a poly-dG spacer giving increased probe surface density, Saprigin did an XPS study of 3-phosphorylated homo-oligonucleotides which were r eacted to a carbodiimide to form an O -phosphoryl isourea intermediate that reacts with surface-bound amine to produce a phosphoramidate linkage.25 They found that homopolymers of guanines gave a higher surface coverage than the other oligonucleotides. They surmised that this be havior was due to the formation of non-Watson Crick base-pairing. How these quadruplexes of DNA might increase the probe surface density is not yet known. The poly-dG quadruplex most like ly forms in solution, before the probes are spotted onto the surface. This more rigid stru cture likely takes up less space on the surface. The quadruplex may require the poly-dG quadruplex to stand erect on the surface, preventing some non-specific binding of the ssDNA. Another possibility is that the quadraplex requires the DNA to be more closely packed. A third possibility is that the quadraplex, with its four phosphate groups fairly close to each other, may have increased avidity towards the zirconium surface. The fact that the poly-dA sequence starts at a higher surface coverage after Rinsing 1, but rinses off the surface with more stringent rins ing conditions, while the poly-dG sequence remains fairly constant, indicates that the sequence with th e poly-dA exhibits st ronger non-specific binding onto the surface than the se quence containing the poly-dG. It is assumed that the bond between the 5-phosphate on the DNA and the zi rconium surface is stable under the rinsing conditions. The conclusion that poly-dA exhibits stronger nonspecific binding is supported by studies that looked at th e specific and non-specific bi nding of homo-oligomers of
48 oligonucleotides. One study done by Wolf et al. compared the binding of thiolated and nonthiolated 25-mers of homo-oligomers on gold.57 They did not look at pol y-dG because they said the non-Watson Crick base-pairing makes synt hesis of long homo-oligomers of guanines inefficient. They found that on gold there was cons iderable non-specific binding, which was sequence dependent. They found that non-thio lated poly-dA exhibite d more non-specific binding than non-thiolated poly-dC and poly-dT. When they studied the binding of thiolated poly-dA,-dT, and -dC, they found that the strand s with the most non-specific binding (poly-dA) bound at a slower rate. The non-specific bindi ng slowed down the bind ing kinetics of the thiolated poly-dA and that during rinsing, more of the poly-dA rinsed off. These studies help support the belief that the poly-dA exhibits hi gher non-specific binding, limiting the specific binding and thus perhaps limiting the final surfac e coverage after all the rinsing steps. Wolf et al. did not attempt to explain why the di fferent bases had different non-specific binding. Why the poly-dA would give rise to more non-specific than poly-dG binding is not apparent in the structure of the base (Figure 3-5). It was thought th at the charge of the zirconium-phosphonate surface may help to explain th is behavior. Streaming surface potential also known as zeta ( ) potential measurements were performed on the zirconium-phosphonate surface. The surface potentia l is the charge that develops at the interface between a solid surface and liquid medium. Two zirconium-phosphonate f ilms were analyzed giving an average surface potential of +19.22 mV. Apparently, the oxide s and hydroxides, which are assumed to cap the zirconium surface, do not completely mask the po sitive charge of the zirconium. What this could mean is that some of the non-specific binding is coming from the negatively charged phosphodiester backbone sticking to the somewhat positively charged surface. On the other hand, at pH 6 gold is reported to have a ne gative surface potential of approximately -20 mV.58
49 This indicates the reason that poly-dA appears to physisorb st rongly both on gold and zirconiumphosphonate is not due to similar surface poten tials. The final reason for the apparent physisorption of poly-dA is unclear. This data indicates that there may be two different reasons why the poly-dG probe after sufficient rinsings has a higher surface density. One, is likely the formation of non-Watson Crick base pairs in solution, resulting in the quadruplex structure. Th e other is increased nonspecific binding of poly-dAs increased non-specific binding which can limit the specific binding. While this non-specific binding is no t capable of holding th e oligonucleotide on the surface during the rinsings it is most likely inte rfering with the specific binding of the phosphate linker with the surface. Conclusion From the XPS quantitation, it is seen that afte r rinsings, the surface coverage of the probe with the poly-dA spacer decrea ses with each rinsing step, wh ile that of poly-dG remains constant. The final surface coverage of the probe with the poly-dG spacer after a mock hybridization is about twice that of the probe wi th the poly-dA spacer. From this data it is concluded there could be two factors contributing to the higher surface coverage for the poly-dG containing probe. The first one is the formation of a quadruplex between the guanine bases. However, it is not completely clear how this structure leads to higher surface coverage. The second is non-specific binding of the poly-dA sequence with the surface limiting the specific binding. Finally, these results show that th e poly-dG spacer leads to a higher probe surface coverage which is most likely the cause of the increased hybridization.
50 Figure 3-1. Fluorescence intens ity after hybridization compari ng the four different spacers on three different probe sequences. Figure 3-2. Illustration of how the poly-dG quadruplex might hold together the probe DNA. poly-guanine spacer oligonucleotide phosphate end 3 5 G G G G G G G G G G G G G G G G OPO3 2-OPO3 22-O3PO2-O3PO poly-guanine spacer oligonucleotide phosphate end 3 5 G G G G G G G G G G G G G G G G G G G G G G G G G G G G OPO3 2-OPO3 22-O3PO2-O3PO5'-H2O3PO-(B)11-O33 0 10000 20000 30000 40000 50000---Fluorescence IntensityGATC OligoXOligoYOligoZ 10 M50 M5 Mno spacer no spacer no spacerIdentity of Spacer B GATCGATC 5'-H2O3PO-(B)11-O33 0 10000 20000 30000 40000 50000---Fluorescence IntensityGATCGATC OligoXOligoYOligoZ 10 M50 M5 Mno spacer no spacer no spacerIdentity of Spacer B GATCGATC GATC GATC
51 Figure 3-3. Poly-dG quadrupl ex. Illustrates the hydrogen bonding between the bases of guanosine. Rinsing 1Rinsing 2Rinsing 3 0.00E+000 2.00E+011 4.00E+011 6.00E+011 8.00E+011 poly-dG poly-dAssDNA molecules/cm2 Figure 3-4. Surface coverage of probe after each different rinsing condition. N N N N N H H H O R N N N N H H H O R N N N N N H H H O R N N N N N N H H H O R 2 1 3 6 7 8 9
52 Figure 3-5. The bases and sugar deoxyadenosine and deoxyguanosine. From these structures, it is not apparent why a poly-dA strand would physisorb more strongly to the zirconium-phosphonate surface. O N N N N O O NH2 O N N N NH O O O NH2 DeoxyadenosineDeoxyguanosine O N N N N O O NH2 O N N N NH O O O NH2 DeoxyadenosineDeoxyguanosine
53 CHAPTER 4 EFFECT OF PHOSPHATE LINKER PL ACEMENT ON THE BINDING OF DOUBLESTRANDED DNA TO ZIRCONI UM-PHOSPHONATE SURFACES Introduction Soon after the advent of the ssDNA microarray, came the dsDNA array.59-66 The main use of the dsDNA microarray is to follow the intera ctions of DNA and proteins, like transcription factors.63 Transcription factor prot eins, like the name suggests, regulate transcription by regulating the binding of RNA polymerase. They are generally modular proteins that bind to dsDNA at a specific region called the promoter sequence, or they can also bind to other transcription factor proteins. By doing this, they can initiate or inhibit transcription. Their influential role in the translati on of the genetic code makes them an important area of research. The more commonly used met hods of studying transcription factors and determining the promoter sequence includ es gel shift analysis,67 DNA footprinting,68 and fluorescence polarization.69 While these methods are reliable, they are not convenient for multiplexed studies of DNA-protein interactions, which is why dsDNA microarra ys are becoming a valuable analytical technique. Just as for ssDNA arrays, there are a number of different proposed methods for making dsDNA arrays. The dsDNA array was first reported by Bulyk et al. who used primer extension of an ssDNA array, which had been made us ing a combination of photolithography and solidphase chemistry.63 The ssDNA array contained, between the linker and the variable sequence, a sequence constant across the area which was complementary to the primer they used. They could then use a polymerase to synthesize the complement to the ssDNA on the surface, creating the dsDNA probe. With a fluorescent tag on the dsDNA, digestion of the DNA by a restriction enzyme could be followed using fluorescence im aging. Another method to create a dsDNA array, which was demonstrated by OBrien et al. used the self-assembly of 5,3 disulfide-
54 modified dsDNA onto bare gold sp ots made through photolithography.64 In this study, after the dsDNA was attached to the surface, they includ ed a further step to remove some of the immobilized dsDNA by exposing the su rface to an alkane thiol They stated that decreasing the density of the duplex was a key factor in incr easing the enzyme accessibility. With AFM, they looked at the change in height of the spots after cleavage with a restriction enzyme. Smith et al. also used a gold surface but with a slightly mo re complex immobilization procedure, ultimately immobilizing thiol-modified ssDNA to maleim ide groups on the surface and then hybridizing with the complement to get the dsDNA.66 Using SPR imaging, they then studied the interactions of different dsDNA sequences with two transcripti on factor proteins and the behavior of these proteins in the presence of an inhibitor. Yet another example of an immobilization procedure on gold, given by Shumaker-Parry et al. first uses a step self-assembling thiol-modified oligo ethyleneglycol and biotin onto gold.65 They then either spotted streptavidin on the surface or self-assembled a film of streptavidin on the surf ace. Finally they spotted biotinylated-dsDNA onto the strepatividin spots or film. In their study, they followed the surface density of the dsDNA with different spotting condi tions, but did not look at protein bi nding to the surface. It is worth mentioning the use of dsDNAs arrays for investigation of protein-protein interactions with mass spect rometry, reported by Becker et al.59 They used the DNA for DNAdirected immobilization, first by spotting 5-amino-modified ssDNA to a succinimide functionalized surface. They could then use the complementary strand with an attached protein to specifically bind different proteins to diffe rent spots. The surface was then exposed to a second protein, which interacted with the protein immobilized on the surface. After a matrix was applied to the surface, they could use MALDI (m atrix assisted laser desorption ionization) mass spectrometry to study the interaction of the two proteins.
55 As demonstrated in the examples given above, there are many uses for and ways to make dsDNA microarrays. Some of the key require ments of a substrate for dsDNA microarrays include that there is little non-specific binding of th e protein and that the dsDNA is bound to the surface in such a way that does not inhibit the bi nding of the protein. Current research is underway to investigate the usefulness of our zirconium-phosphonate monol ayer as a substrate for dsDNA microarrays. Our collaborators, Dr. Br uno Bujoli and the members of the Laboratoire de Synthse Organique, in Nantes France, performed studies analyz ing the ability of a protein, ArgR, to bind to dsDNA on the zirconium-phosphona te surface. With fluorescence imaging, they looked at protein binding to dsDNA microarrays, which containe d the binding site of ArgR. While they used the same probe sequence, th ey did, however, vary length, identity and placement of the spacer and also the placement of the phosphate linker. They found that the dsDNA sequence with a poly-G spacer and phosphat e linker on both 5 ends gave the highest protein binding. In this chapter are the XPS studies of dsDNA on our zirconium-phosphonate surface, which help to explain why more protein bound to the spot s of the 5-5 phosphorylated dsDNA. Experimental Zirconium-phosphonate substrates we re prepared as stated in the experimental section of Chapter 2 of this dissertation. To prepare the dsDNA, ssDNA was prepared in a 1 x SSC (saline sodium citrate) buffer, pH 6 at double the desired concentration. The saline sodium citrate buffer solutions are prepared from a stock solution of 20 x SSC, which corres ponds to a 3.0 M NaCl and 0.30 M sodium citrate solution. Thus for a 1 x SSC solution, the c oncentrations are 0.15 M NaCl and 0.015 M sodium citrate, creating a solution which has an ionic strength of 0.225 M. The ssDNA was then mixed with its complement in equal volume to give the final desired concentration in a 1 x SSC buffer. The DNA was hybridized using a thermocyler, first by
56 heating at 98 C for 2 minutes, 65 C for 2 minutes, and finally cooling to 4 C for 4 minutes. The dsDNA was then spotted onto the slide; to cr eate a spot large enough for XPS analysis, 30 L of the DNA was pipetted onto the rinsed and dr ied zirconium-phosphonate surface. Once the DNA had been spotted, it was allowed to react with the slides overni ght in Petri dishes at room temperature. Two different rins ing conditions were used. One set of conditions was performed to allow the probe coverage to be compared to the protein binding expe riments, the other to allow the probe coverage to be compared to th e ssDNA experiments. Th e rinsing conditions set to mimic those used for the protein binding experi ments involved first submersion of the slide in 3.5 x SSC, 0.3% SDS (sodium dodecylsulfate) for 45 minutes with gentle rocking. Then the slide was submerged in 1 x SSC for 5 minutes with gentle rocking. Finally, the slide was rinsed in water by dipping 10 times in 5 di fferent vials of water. The other set of rinsing conditions, which are the same as the ssDNA experiments, began with the spotted slide being immersed for 1 hour in 3.5 x SSC, 0.3% SDS at 42 C, followed by rinsing 5 tim es with nanopure water and spin drying. The slides then underwent a mock hybridization, which shoul d give the true probe concentration if they were to underg o hybridization, by treatment with 25 L per spot of 3 x SSC, 0.1% SDS overnight at 42 C. Finally, the slides were rinsed in 2 x SSC, 0.1% (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min), followed by rinsing in water 5 times. At least four spots were examined at each di fferent rinsing condition. Results and Discussion Linker and Spacer Placement The employment of dsDNA microarrays to probe different DNA-protein interactions is being seen more and more in the literature. To study how well a zirc onium-phosphonate surface worked as a substrate, our co llaborators, Dr. Bruno Bujoli an d his group, investigated the
57 interaction of dsDNA, immobilized on the zirconium-phosphonate surf ace, with a protein. They chose to look the protein ArgR, which plays a major role in the control of certain biosynthetic and catabolic arginine genes.70 They also looked at the influen ce of the identity of the spacer and the placement of the phosphate linker and spac er on the dsDNA. Relative to ssDNA the number of possible ways for attaching dsDNA to the surface increases. For ssDNA with a covalent linker, the two possibilities are to use the linker on the 5 end or th e 3 end. In earlier chapters, ssDNA phosphorylated on the 5 end was used, based on earlier work which showed this effectively immobilized the ssDNA on our zirc onium-phosphonate surface. For dsDNA, a number of different linker and spacer motifs were investigated by Bujoli et al. The corresponding fluorescent data fo r the protein binding experiments are shown in Figure 4-1. To look at the protein binding to the dsDNA on the surface, first Bujoli et al. spotted 10 M ds-DNA in a 1x SSC solution. The dsDNA contained a 22-bp sequence located in the middle, which is recognized by the protein. They then exposed the slide to a 0.3% -casein, 3.5 x SSC, 0.3% SDS solution for 45 min. Casein, a pr otein that contains phos phate groups, binds to the area of the slide that does not contain any ds DNA. This prevents non-specific binding of the target protein, lessening the back ground signal. After rinsing the slide, they exposed it to the target protein, ArgR. The prot ein did not contai n a fluorophore, but it did have what is called a histag, a sequence of 6 histidines. After rinsing again, they exposed the surface to an anti-histag antibody solution, which binds to a hi stag on the ArgR protein. Afte r another rinsing, the slide is exposed to a secondary antibody, Anti-Mouse Ig G, containing a fluorophore. The secondary antibody binds to the anti-histag antibody. The s lide was studied with fluorescence imaging after a final rinsing. Figure 4-2 illustrates the main steps in the experiment.
58 Through fluorescence imaging it was seen that the dsDNA with 5,5 phosphates gave the most intense fluorescence and therefore the highes t amount of protein bindi ng. It was not clear why more protein bound to the 5-5 phosphorylated dsDNA spot s. It was surmised that it could be due to increased surface density of dsDNA, or possibly from the DNA binding on both ends using both phosphate groups creating a loop of dsDNA on the surface perhaps better exposing the segment where the protein would bind. To determine if surface density played a role in the binding, quantit ative XPS experiments were performed. Four different phosphate li nker/spacer motifs, shown in Figure 4-3, were investigated. These four motifs consisted of a dsDNA sample with phosp horylation and a 9-mer guanine spacer on both 5 ends (5G9PO4,5G9PO4); a dsDNA sample with phosphorylation and a 9-mer guanine spacer on both 3 ends (3G9PO4,3G9PO4); a dsDNA sample with phosphorylation and a 9-mer guanine spacer on the 5 end of one strand and the same on the 3 end of the other strand (5G9P O4,3G9PO4); and finally a dsDNA sample with phosphorylation and a 9-mer random spacer on both 5 ends (5X9PO4 ,5X9PO4). The rinsing procedures used for the XPS experiments were set to mimic thos e used in the fluorescence experiments, before the ArgR protein binding step wa s performed. However, the ri nsing had to be modified to prevent nitrogen contamination. For example, in the passivation step with the casein, the casein was not used for the XPS experiments. To immobilize the dsDNA, first it was hybridized in a thermocyler, and then spotted onto the surface. After the ds DNA had reacted with the substrate overnight, it was rinsed with gen tle rocking in a 3.5 x SSC/0.3% SDS solution, followed by rinsing in a 1 x SSC solution, and fina lly the slide was rinsed in water and dried. XPS was then taken on the area where the dsDNA had been spotted.
59 The surface coverage can be calculated using a similar method to that outlined in Chapter 2. Again the N 1s peak was used with the Zr 3d peak in Equation 2-8 from Chapter 2: sin exp 1 sin expeff N N N N Zr eff Zr Zr Zr N Zr NL t L T I L t t T I n n (2-8) Equation 2-8 calculates the ratio of the nitrogen surface dens ity to the zirconium surface density. As shown in Chapter 2, to get the DNA density from the nitrogen to zirconium ratio, the nitrogen stoichiometry is used along with known surface density of zirconium. A typical N 1s peak for the dsDNA is shown in Figure 4-4. Wh en calculating the surface coverage of dsDNA, there are two different ways to consider the nitrogen stoichiometr y. The surface coverage can be calculated as the number of DNA strands/cm2 or the number of dsDNA strands/cm2. To use the former, the average number of nitrogens per strand is taken for the dsDNA complex. For the latter, the total number of nitr ogens in the dsDNA complex can be used. To compare the surface coverage among the different phosphate-linker an d spacer placement experiments, the number of dsDNA strands/cm2 is used. The comparison of the surface coverages is show n in Figure 4-5. The graph shows that the 3G9PO4,3G9PO4 DNA gave the hi ghest surface coverage at 2.8 1011 4.8 1010 dsDNA molecules/cm2. Then the 5G9PO4,5G9PO4 DNA gave th e next highest surface coverage at 2.1 1011 2.2 1010 dsDNA molecules/cm2, followed by the 5X9PO4,5X9PO4 DNA at 1.7 1011 1.3 1010 dsDNA molecules/cm2. Finally the 5G9PO4,3G9PO4 DNA had the lowest surface coverage at 1.2 1011 1.0 1010 dsDNA molecules/cm2. The error reported here is the standard deviation of the mean. This follows the general trend seen w ith fluorescence indicating that the surface coverage does play a role in the differences seen with the protein immobilization.
60 Based on the XPS data, it appe ars that having two linking gr oups on opposite ends of the dsDNA increases the surface coverage. The increased dsDNA density seen with 3G9PO4,3G9PO4 and 5G9PO4,5 G9PO4 DNA could be due to an increased probability of binding with a phosphate group on both ends compar ed to two phosphate groups on a single end. This stems from the fact that with a phosphate group on both ends, binding can occur with the zirconium monolayer when either end approaches the surfa ce. The data does not, however, indicate the orientation in how the dsDNA binds to the surface. A surface coverage of 2.85 1011 corresponds to 357 nm2/dsDNA molecule. This is ample distance for the 3G9PO4,3G9PO4 and 5G9PO4, 5G9PO4 dsDNA to lie down and bind with both phosphate groups. If the dsDNA binds with both phospha tes, the dsDNA may be more strongly bound to the surface, which could also be a reason for higher surface coverage. The XPS data also demonstrate, based on the higher surf ace coverage of 5G9PO4,5G9PO4 over 5X9PO4,5X9PO4, that the poly-dG effect di scussed in Chapter 3 also applies to dsDNA. It should be noted that, alt hough the XPS data follows a tren d that corresponds slightly with the fluorescence data, the magnitudes of the differences seen in the XPS data are smaller than those seen in the fluorescence data. For example, in the fluorescence data the 3G9PO4,3G9PO4 DNA was 4 times more intens e than the 5X9PO4,5 X9PO4 DNA, where as the XPS data shows the 3G9PO4,3G9PO4 DNA ha s a surface coverage of only 1.7 times more than the 5X9PO4,5X9PO4 DNA. This may just be due to the variability of the data. On the other hand, there has also been so me suggestion in the literature that larger spot sizes (2 mm spots compared to 150 m spots) can lead to lower surface coverages.71 Bujoli et al. our collaborators performed fluores cence experiments comparing 2 mm and 8 mm spot sizes and no difference in intensity was seen between these tw o spots, but perhaps the 2 mm spot needs to be
61 smaller to see a difference. The spots necessary for XPS analysis are approximately, 1.5 cm x 1 cm. If the larger spot sizes lead to lower su rface coverage, then the X PS data may not represent the surface coverages of the DNA microarrays. Single Versus. Double Stranded DNA Immobilizing ssDNA and dsDNA on a surface for microarrays each serve different analytical purposes. It has been shown by Peterson et al. that 5-thiolated ssDNA binds at a higher surface density on gol d than 5-thiolated dsDNA.36 A comparison of the surface coverage of ssDNA and dsDNA would help us to determine the most efficient method for making dsDNA microarrays. Would a higher dsDNA probe dens ity be achieved for ssDNA first immobilized and then hybridized or dsDNA directly immob ilized on the surface? A comparison of ssDNA and dsDNA would also help us understand further the dynamics of the DNA immobilization. In this study, there were th ree different strands compar ed. A 5 phosphorylated ssDNA (5PO4G9O33), a 5 phosphorylated ds DNA (5PO4G9O33O33), and a doubly 5 phosphorylated dsDNA (5PO4G9O37,5PO 4G9O37). The sequence of the 5PO4G9O37,5PO4G9O37 is different than that of the other two. In order to do this study, we wanted to use the same sequence, but for some reason when Invitrogen, the manufacturer of the DNA, tried to synthesize a 5 phosphorylat ed complement to 5PO4G9O33 the sample repeatedly failed quality control. After trying the synthesis four times, they stopped trying. For this reason, the probe sequence of the doubl y phosphorylated dsDNA had a different sequence than the samples. As in the other studies, the DNA was spotted on th e surface with a large enough surface area to be sampled by XPS. Th e washing procedure is outlined in the experimental section. In brief, it begins by submerging the slide in a SSC/SDS solution, followed by water rinsing, and then treating the s lide with SSC/SDS overnight. Then the slide is rinsed with water a final time and dried, after which XPS was taken.
62 The surface coverage results are shown in Fi gure 4-6. For ease of comparing surface coverage with the ssDNA, the dsDNA is reported as number of DNA strands/cm2 instead of the number of dsDNA strands/cm2. The calculated surface c overages for 5PO4G9O33, 2.8 1011 5.1 1010 DNA strands/cm2, and 5PO4G9O37,5PO4G9O37, 2.6 1011 4.0 1010 DNA strands/cm2, are approximately the equal. While, 5PO4G9O33O33, the singly phosphorylated dsDNA, has a lower surface coverage: 1.1 1011 1.4 1010 DNA strands/cm2. This data shows that the bulkier dsDNA, if it has only a si ngle phosphate like the ssDNA, binds at a lower density than the ssDNA. This is similar to what Peterson observed on gold.36 Peterson claims in his paper that because the duplex has double the anionic charge on the backbone compared to the ssDNA, it is expected the dsDNA would bind at half the density of the ssDNA due to repulsion of the dsDNA molecules. He goes on to say that the density is even lower than half the ssDNA density indicating ther e are other factors coming into play such as conformation or flexibility of the DNA strands. It does not seem likely that one of the main causes of lower dsDNA density is increased electrostatic repulsion of the dsDNA strands. The charge of the dsDNA strands should be shielded by the ions in solution, minimizing repulsive forces. The most reasonable cause for the de creased surface coverage of the dsDNA is the different dynamics of the dsDNA such as decreased flexibility. Conclusion The surface densities of dsDNA with four different phosphate -linker/spacer motifs were calculated from the N 1s and Zr 3d XPS data. Th e highest surface density was achieved with the 3G9PO4,3G9PO4 dsDNA at 2.8 1011 4.8 1010 dsDNA molecules/cm2 followed by the 5G9PO4,5G9PO4 DNA, which gave the next highest surface coverage at 2.1 1011 2.2 1010. These are also the two strands that gave the highest fluorescence in the protein binding
63 experiments done by our colleagues. This data indicates that the reas on the protein exhibited more binding for the 5G9PO4,5G9PO4 and 3G9PO4,3G9PO4 was due to higher dsDNA surface density. In addition to a comparison of the different phosphate-linker/spacer motifs, a comparison of dsDNA vs ssDNA was made. It was seen that dsDNA with a single phosphate group on the 5 had a lower surface coverage at 1.1 1011 1.4 1010 DNA strands/cm2 than an ssDNA strand with a single phosphate, wh ich had a surface coverage of 2.8 1011 5.1 1010 DNA strands/cm2. A dsDNA strand with a phosphate group on both 5 ends gave a surface coverage of 2.6 1011 4.0 1010 DNA strands/cm2.
64 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 SSC 1xP-9G-NS37 / CNS37 P-9G-37 / C37 P-9G-37 / C37-9G-P P-9G-379GP / C37 37-9G-P / C37-9G-P P-9G-37 / P9G-C37 P-9N-37 / P9N-C37 P-NN-37 / PNN-C37 (G9) PO4H2 (G9) PO4H25 5 (X9) PO4H2PO4H2 (X9) 5 5 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2H2O4P (G9) 53 (G9) PO4H25 (G9) PO4H25 3 (G9) PO4H2 (G9) PO4H25 PO4H25 5 PO4H2 (G9) PO4H2 (G9) PO4H25 5 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 5 (X9) PO4H2PO4H2 (X9) 5 5 (X9) PO4H2PO4H2 (X9) (X9) PO4H2 (X9) PO4H2PO4H2 (X9) PO4H2 (X9) 5 5 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2H2O4P (G9) 53 (G9) PO4H2H2O4P (G9) (G9) PO4H2 (G9) PO4H2H2O4P (G9) H2O4P (G9) 53 (G9) PO4H25 (G9) PO4H2 (G9) PO4H25 (G9) PO4H25 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 (G9) PO4H2 (G9) PO4H25 PO4H25 5 PO4H2 PO4H2 PO4H25 5 PO4H2 PO4H2FlourescenceIntensity 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 SSC 1xP-9G-NS37 / CNS37 P-9G-37 / C37 P-9G-37 / C37-9G-P P-9G-379GP / C37 37-9G-P / C37-9G-P P-9G-37 / P9G-C37 P-9N-37 / P9N-C37 P-NN-37 / PNN-C37 (G9) PO4H2 (G9) PO4H25 5 (X9) PO4H2PO4H2 (X9) 5 5 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2H2O4P (G9) 53 (G9) PO4H25 (G9) PO4H25 3 (G9) PO4H2 (G9) PO4H25 PO4H25 5 PO4H2 (G9) PO4H2 (G9) PO4H25 5 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 5 (X9) PO4H2PO4H2 (X9) 5 5 (X9) PO4H2PO4H2 (X9) (X9) PO4H2 (X9) PO4H2PO4H2 (X9) PO4H2 (X9) 5 5 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2H2O4P (G9) 53 (G9) PO4H2H2O4P (G9) (G9) PO4H2 (G9) PO4H2H2O4P (G9) H2O4P (G9) 53 (G9) PO4H25 (G9) PO4H2 (G9) PO4H25 (G9) PO4H25 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 (G9) PO4H2 (G9) PO4H25 PO4H25 5 PO4H2 PO4H2 PO4H25 5 PO4H2 PO4H2FlourescenceIntensity Figure 4-1. Fluorescence intensities of the spots for the protein binding to the corresponding dsDNA. The dsDNA show the different spacer and linker motifs that were used. The dsDNA with the pink band contains a random sequence which should not be recognized by the protein ArgR. The blue band is a dsDNA seque nce that should be recognized by the protein ArgR. Th e X spacer is a random sequence.
65 Figure 4-2. Steps used to test the binding the ArgR protein to dsDNA containing a 22 base-pair sequence. First, the dsDNA is spotted, which is followed by a surface passivation with casein. The slide is exposed to the Ar gR protein containing a histag and then to an anti-histag antibody. Fina lly, the slide is exposed to the Anti-Mouse IgG antibody, containing a fluorophore. (G9) PO4H2 (G9) PO4H25 5 (X9) PO4H2PO4H2 (X9) 5 5 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2H2O4P (G9) 535G9PO4,5G9PO4 3G9PO4,3G9PO4 5X9PO4,5X9PO4 5G9PO4,3G9PO4 (G9) PO4H2 (G9) PO4H25 5 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H25 5 (X9) PO4H2PO4H2 (X9) 5 5 (X9) PO4H2PO4H2 (X9) (X9) PO4H2 (X9) PO4H2PO4H2 (X9) PO4H2 (X9) 5 5 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H2 (G9) PO4H23 3 (G9) PO4H2H2O4P (G9) 53 (G9) PO4H2H2O4P (G9) (G9) PO4H2 (G9) PO4H2H2O4P (G9) H2O4P (G9) 535G9PO4,5G9PO4 3G9PO4,3G9PO4 5X9PO4,5X9PO4 5G9PO4,3G9PO4 Figure 4-3. The dsDNA with the different phosphate-linker/spacer mo tifs that were studied with XPS. PO3 2PO3 2PO3 2PO3 22) Incubation withthe protein(His6tag) Incubation withan antiHistag antibody Incubation witha fluorescent secondary antibody 1) Saturation witha-casein PO3 2PO3 2PO3 2PO3 2PO3 2PO3 2PO3 22) Incubation withthe protein(His6tag) Incubation withan antiHistag antibody Incubation witha fluorescent secondary antibody 1) Saturation witha-casein
66 Figure 4-4. Typical XPS spectra for the N 1s peak. This peak in particular is for 3G9PO4,3G9PO4. 5 G 9 P O 4 3 G 9 P O 4 5 G 9 P O 4 5 G 9 P O 4 3 G 9 P O 4 3 G 9 P O 4 5 X 9 P O 4 5 X 9 P O 40.00E+000 1.00E+011 2.00E+011 3.00E+011 dsDNA/cm2dsDNA Figure 4-5. Comparison of the surf ace densities of the dsDNA strands. Binding Energy (eV) N(E)Min: 2939Max: 3614 410408.6 407.2 405.8 404.4 403 401.6 400.2 398.8 397.4 396
67 5 P O 4 G 9 O 3 7 5 P O 4 G 9 O 3 7 5 P O 4 G 9 O 3 3 O 3 3 5 P O 4 G 9 O 3 30.00E+000 1.00E+011 2.00E+011 3.00E+011 DNA strands/cm2 Figure 4-6. Surface density comp arison of dsDNA versus ssDNA.
68 CHAPTER 5 HYBRIDIZATION OF DNA AT A ZIRCONI UM-PHOSPHONATE SURFACE AN XPS STUDY Introduction DNA microarrays can be used for a number of different purposes. More commonly they are used to follow gene expressi on. To make a microarray for th is purpose requires a series of complex steps including probe selection, probe spotting, target preparation (mRNA extraction, reverse transcription, and cDNA la beling), hybridization of target w ith probe, rinsing, scanning, and data acquisition.30 Hybridization is a key step in the process and can affect the noise (specificity of the probe-target interaction) and the signal to b ackground ratio. The hybridization efficiency and amount of target physisorbed to the surface where no probe is spotted determine the signal to background ratio. However, spec ificity and hybridization e fficiency work against each other; consequently, optimization of the hybrid ization conditions is used to give the best results. Optimization is facilitated by the use of fluores cence imaging with a microarray scanner. which is how the hybridization conditions we re optimized for our zirconium-phosphonate support.32 However, fluorescence imaging does not gi ve quantitative information on the probe surface density or the hybridizati on efficiency. It has been shown that XPS can be used to calculate these values,12 and in this chapter we discuss our attempt to determine the hybridization efficiency with XPS. Experimental The zirconium-phosphonate monolayers were prepar ed as stated in Chapter 2. Reagents were purchased from Aldrich and were used as received. DNA was purchased from Invitrogen, Biosource (which was bought by Invitrogen), or MWG biotech all with HPLC purification.
69 Confirmation of Hybridization with Fluorescence Confocal Microscopy This fluorescence study was done on a Nikon lase r confocal scanning microscope. For these experiments DNA was ordered from MWG Biotech with HPLC purity and used following sequence for the probe: 5 GGGGGGGGG GGGACCCCAGAGGTATACATACGTTGCAGT CAGGA and the following sequence for the target: 5 TCCTGACTGCAACGTATGTATACC TCTGGGGTC 3. The target was labeled with a 5 Cy3. A 40 M probe solution was prepared in a pH 6 1 x SSC solution. The saline sodium citrate buffer solutions ar e prepared from a stock solution of 20 x SSC, which corresponds to a 3. 0 M NaCl and 0.30 M sodium citrate solution. Thus for a 1 x SSC solution, the concentrations are 0.15 M NaCl and 0. 015 M sodium citrate, creating a solution which has an ionic strength of 0.225 M. The spots were made by dropping 5 L on the slide with a pipetter. The slide was inc ubated overnight in a Petri dish. Then the slide was rinsed in 2 x SSC (saline s odium citrate), 0.1% SDS (sodium dodecyl sulfate) (2 min), 1 x SSC (2 min), 0.2 x SSC (2 times, 2 min). Followi ng the rinsing, the slide is treated with a 1% BSA (bovine serum albumin), 3.5 x SSC, and 0.3% SDS for 1 hour at 42 C. The slide was then rinsed with water by dipping 10 times To perform the hybridization, 40 L of the hybridization solution (5 M of target, 0.3% SDS, 3.5 x SSC, 5 x Denharts, 1 x TE buffer, and 50% formamide) was dropped onto the ar ea containing the spots and th en a 1 x 2 coverslip was placed on top. The slide was placed in a Petri dish and kept at 42 C overnight. The hybridization was followed with rinsing in 2 x SS C, 0.1 % SDS (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min.). Finally, the slide was spun dry. After whic h the slide was analyzed under the microscope.
70 Fluorescence Comparison of Hybridization Methods This experiment was performed with an Agile nt 2-dye fluorescent microarray scanner. The probe sequence was 5-H2O3PO-(X)9-CCGCCGGTAACCGGAGGTTAAGATCGAGA-TCCA and the target sequence was 5-TGGATCT CGATCTTAACCTCCGGTTACCGGCGG with a Cy3 fluorophore on the 5 end. Both a poly-ad enine and guanine spacer on the probe was studied. A 40 M solution of the two probes were prepar ed in a 1 x SSC, pH 6 buffer. These were spotted on the sl ide by dropping 4, 2.5 L spots, which coalesced to form one large 10 L spot. The slide was then incubated overnight at room temperature. Following incubation, the slide was submerged in a 3.5 x SSC, and 0.3% SDS for 1 hour at 42 C, after which it was rinsed by dipping 10 times in water. Two different hybrid izations were performed on the slide. On one half of the slide 2 spots of the probe cont aining poly-A and 2 spots of the probe containing poly-G it was subjected to the normal hybridization, where 20 L of the hybridization solution (5 M target, 0.3% SDS, 3.5 x SSC, 5 x dernhard ts, and 50% formamide) is dropped on the spotted area and a 1 x 1 coverslip is placed on th e drop. On the other half of the slide, where there was also 2 spots of the poly-A contai ning and poly-G containi ng probe, was dropped 15 L of the nitrogen-free hy bridization solution (5 M target, 3 x SSC, 0.1% SDS) on each area where the probe was. After sitting ov ernight in a humid Petri dish at 42 C, the slide was rinsed 2 x SSC, 0.1 % SDS (2 min), 1 x SSC (2 min), and 0. 2 x SSC (2 times, 2 min.). Finally the slide was spun dry and the fluores cence imaging was performed. Determination of Nitrogen Contamination from Hybridization Solution with XPS Blank hybridization experiments were performe d by first spotting a mock probe solution of 1 x SSC, pH 6 (contained no probe) onto a zirconiu m phosphonate substrate. The slide was then submerged in a 3.5 x SSC, and 0.3% SDS for 1 hour at 42 C. After rinsing 5 times with water,
71 the slide was treated with one of two mock hybridization soluti ons (contained no target). One solution was composed of 0.3% SDS, 3.5 x SSC 5 x Dernhardts, and 50% formamide. The other solution was 0.3% SDS, 3.5 x SSC, and 50% formamide (the same as the first except without the Dernha rdts). The 20 L of the mock hybridization so lution was dropped on the slide and a 1 x 1 was placed over the drop. The slide was then placed in a humid chamber and left overnight at 42 C. After which, the slide was rinsed with 2 x SSC, 0.1 % SDS (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min.), followe d by rinsing 5 times with water. The slide was spun dry and XPS analysis was performed on th e area where the 1 x SSC had been spotted. Hybridization Study with XPS Real hybridization experiments were pe rformed using the probe sequence, 5-H2O3PO(X)9-CCGCCGGTAACCGGAGGTTAAG ATCGAGATCCA, and the target sequence: 5TGGATCTCGATCTTAACCTCCGGTTACCGGCGG A 40 M solution of the probe DNA was prepared in a 1 x SSC buffer, pH 6. Th e DNA was spotted on the slide to cover an area large enough for XPS analysis. The slide was then allowed to incubate overnight at room temperature, followed by submersion in a 3.5 x SSC, and 0.3% SDS solution for 1 hour at 42 C. A 5 M target solution was prepared in a 3 x SSC, 0.1% SDS buffer. 25 L of the target solution was spotted on the slide, covering the area wher e there was probe. The slide was placed in a humid Petri dish overnight at 42 C. Then the slide was rinsed with 2 x SSC, 0.1 % SDS (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min.), followed by either dipping 2 times in 5 vials of water or 10 times in 5 vials of water. After this, XPS was taken. Results and Discussion The basis of the DNA microarray is the specifi c hybridization of each probe to a labeled complementary target. There are many parameters th at affect the efficiency of hybridization of a
72 probe with its complementary target at a surface. Thes e parameters include the identity of the probe and target, the probe environment, the buffer conditions and the temperature.36, 37, 52, 72, 73 Some of these parameters can easily be changed and are adjusted to give the optimum result, which is generally high specificity and a high si gnal to background ratio. A high signal to background ratio is obtained with a high probe surface density, high hybridization efficiency, and low physisorption of the target to non-probe containing regions. The specificity of each probe is associated with its melting temperature. To help optimize the specificity there is an equation, which was first proposed by Howley et al. that relates the melting temperature of the DNA on a surface to the conditi ons of the hybridization:74 bp / 500 %form 61 0 % 41 0 Na log 6 16 5 81 GC Tm (5-1) where [Na+] is the concentrati on of sodium ions; %GC is th e percentage of guanine and cytosine in the probe and target; %form is th e percentage of formamide in the hybridization solution; and bp is the number of base pairs. The equation was developed for DNA 50 base pairs or longer. Nevertheless, the equation shows the important relationship of two components which are often included in a h ybridization solution: Na+ and formamide. The Na+ increases the melting temperature and the formamide decreases it. Thus, to increase specificity, the formamide can be increased, but this will also lower hybridization efficiency. On the other hand, to increase specificity [Na+] can be decreased, but again this can lead to lower hybridization efficiency. As mentioned above the probe environment can affect the hybr idization efficiency. For example, if the density of the probe is too hi gh this will lead to low hybridization efficiency.36 Of course, if the probe density is too low, then only a small amount of targ et will be hybridized onto the surface, which may result in low signal to background. It has also been shown that a
73 spacer placed between the linker and probe se quence can increase hybrid ization efficiency supposedly by bringing the probe sequence closer in contac t with the solution phase.37, 52 Furthermore, there is evidence that the ch arge on the surface can affect hybridization.73 For the ssDNA arrays made on our zircon ium-phosphonate surface, the spotting and hybridization conditions were optimized by our colleagues in France using fluorescence imaging.32 These optimized conditions begin with spotting the ssDNA (10 40 M) prepared in a pH 6, 1 x SSC buffer. The slide is then incu bated overnight in 50% humidity. Then, the slide is subjected to a passivation step, where it is submerged in 1% BSA (bovine serum albumin), 3.5 x SSC, and 0.3% SDS for 1 hour at 42 C, which is followed with rinsing 5 times with water. The next step is hybr idization where the 20 L of the hybridization solution (5 M target, 0.3% SDS, 3.5 x SSC, 5 x Denhardts, 1 x TE (tris( hydroxymethyl)aminomethane/EDTA) buffer, and 50% formamide) is dropped onto the surface and a coverslip is placed on this drop so that it spreads out over the array area. The slide is then place in a hu mid sealed chamber in an oven overnight at 42 C. Finally the slide is washed with 2 x SSC, 0.1 % SDS (2 min), 1 x SSC (2 min), and 0.2 x SSC (2 times, 2 min.). This experi ment was tried with larger spots in our lab to confirm we were using the correct techni ques. The spots, which were about 2 mm in diameter, were analyzed with confocal fluores cence microscopy (Figure 5-1). The spotting and hybridization in our lab appeared successful. To determine the hybridization efficiency, the XPS experiments simply looked at the increase in the N 1s signal afte r hybridization. Several modificati ons had to be made in order to make the experiments viable with XPS. One simp le modification made was that the slides had to be rinsed water to remove salts after the final SSC rinsings. This was because the salts could attenuate the signals from the DNA and zirconium layers. Also, the hybridization solution had to
74 be modified so that the nitrog en containing solution components, other than the target, would not contaminate the surface giving a false increase in nitrogen signal. The nitrogen containing constituents that might cause contamination we re formamide, Denhardts solution, and Tris E buffer. Denhardts is a mixture of ficoll (a sucrose polymer), polyvinylpy rrolidine, and BSA, of which the latter two contain nitroge n. Thus before the real hy bridization experiments took place blank experiments were performed to determ ine which components might cause nitrogen contamination. This was done using spotting so lutions and hybridization solutions containing no DNA and varying the nitrogen containing additiv es. The TE buffer was eliminated without testing simply because our French collaborators changed the hybridization solution such that the TE buffer was no longer used. The blank spotti ng solutions were spot ted onto the zirconiumphosphonate surface. Then the slides were rinsed as outlined in th e Experimental section. Then the slides were subjected to one of two blank hybridization solu tions overnight at 42 C. One contained SDS, SSC, Denhardts reagent, a nd formamide (Blank 1). The other solution contained SDS, SSC, and formamid e (Blank 2). Following the blank hybridization, the slide was rinsed with several SSC solutions and then wa ter. The XPS survey scans (Figure 5-2) both contain a considerably large N 1s peak, giving evidence that bot h solutions produced nitrogen contamination with Blank 1 (the Denhardtscontaining solution) gi ving more nitrogen contamination. From these results came the decision to use a hybridi zation solution containing only SSC, SDS, and the target DNA. Eliminating formamide and Denhardts solution may actually increase the ap parent hybridization si nce both are in the hybridization solution to prevent nonspecific interactions. The hybridization solution cont aining only SSC, SDS, and ta rget was compared using fluorescence detection along side the hybridization solution c ontaining SDS, SSC, Denhardts
75 reagent, and formamide and target. To do this the ssDNA was spotted onto a slide. After the slide was rinsed, on one half of the spots, the 20 L of the latter hybridization solution was dropped and then a coverslip was placed over it. On the other half, the former hybridization solution was dropped onto the probe areas just enough to cover the spot. The slide then underwent the routine SSC rinsings, which we re followed by water rinsings. Scanning fluorescence imaging was then used to analyze the spots (Figure 5-3). From these results it was seen that there was little difference in the in tensities from the spots using the two different hybridization methods. This validated the use of the nitrogen-free hybridization solution. After the nitrogen-free hybridization buffer was tested with fluor escence, hybridization using this buffer was followed with XPS. This was done by spotting the probe, rinsing the slide, and then spotting the target, prepared in the nitrog en-free buffer, in the area of the probe. The slide was incubated overnight and with the targ et then a final rinsi ng with SSC buffers was performed. As mentioned before it is necessary to rinse the slides with water before taking XPS. This removes excess salt from the buffers which can cause attenuation of the N 1s and Zr 3d XPS peaks. There was concern that the water rinsings may cause denaturation as seen in Equation 5-1, DNA needs salt to hybridize. For this reason, two water rinsing conditions were explored. One used 10 dips in 5 different vials of wate r and the other used 2 dips in 5 different vials of water. Using the N 1s and Zr 3d peak intensities, the surface coverage (reported in DNA strands/cm2 not dsDNA/cm2 molecules) were calculated usi ng the methods outlined in Chapter 2 and 4. The DNA surface coverage calculations are shown in Table 5-1. The surface coverage of the poly-G containing probe and poly-A containing probe are 2.8 1011 5.1 1010 ssDNA molecules/cm2 and 1.4 1011 3.3 1010 ssDNA molecules/cm2, respectively. These values are
76 also given in Chapter 3, and the procedure for imm obilizing the probe is also in Chapter 3. After Hyb-1, the surface coverages for the poly-G c ontaining probe and polyA containing probe are 2.1 1011 5.2 1010 DNA strands/cm2 and 1.2 1011 2.0 1010 DNA strands/cm2, respectively. After Hyb-2, th e surface coverages for the polyG containing pr obe and poly-A containing probe are 1.7 1011 1.8 1010 DNA strands/cm2 and 0.7 1011 2.8 1010 DNA strands/cm2, respectively. As the data show, no increase in DNA surface coverage could be seen with XPS after hybridization; conversely, there ap pears to be a decrease in th e surface coverage. However to determine the probe surface coverage, a mock hybr idization was included, so that the probe was subjected to the same conditions as hybridization except no target wa s present. For this reason, it makes the decrease in coverage seen very puzzli ng. There are several possible explanations for this. One is that the water rinsings, which are ne cessary to remove salts before XPS analysis, are causing denaturation of the dsDNA (and possibly more desorption of the probe). However, our French colleagues sometimes rinse their dsDNA slid es in water briefly, but do not experience denaturation. Unfortunately, there are no studi es that investigate water rinsings of dsDNA immobilized on a surface. If the hybridization is o ccurring, which seems to be the case based on the fluorescence data, unless th e hybridization efficiency wa s above around 20%, it would be difficult to detect with XPS based on the expe rimental error of the probe data. A low hybridization efficiency is lik ely related to the low probe surface coverage, where the ssDNA strands are able to lie down on the zirconi um-phosphonate surface. The positive potential (as shown in Chapter 3) of the surface attracts the negatively charged phosphodiester backbone. Others have reported low or zero hybridization if the probe was t oo close to the surface. For example Guo et al. reported that when they did not us e a spacer between the linking group and
77 the probe sequence no hybridization was seen.52 Perhaps with our surf ace, when the probe is close to the surface, the target is sterically hindered from hyb ridizing with probe. Overall, considering the XPS experiments and the fluores cence experiments, hybridi zation is occurring, but not at a high enough e fficiency to be detectable with XPS. However, why a decrease is seen in the DNA surface coverage is puzzling. Perhap s, after hybridization, probes that are not bound with sufficient strength, as a duplex the DNA rinses o ff more easily. Conclusion In conjunction with fluorescence imaging expe riments, XPS experime nts were performed in order to determine the hybridization effici ency. Hybridization was observable with the fluorescence, but not with XPS. Th is indicates that the extent of hybridization is too little to observe with XPS. Even though, when dete rmining the probe surface coverage, a mock hybridization was used, a decrease was seen in the overall DNA surface coverage during the real hybridization experiments, making the data more difficult to interpret. Figure 5-1. Fluorescence confocal microscopy image of hybridized spots. The total spot (about 2mm in diameter) would not fit in the whole image.
78 Figure 5-2. XPS spectra of bl ank hybridization taken to determ ine if nitrogen contamination would occur by following the N 1s peak. On ly the significant region is shown, for this reason the percentages do not add up to 100. A) XPS spectrum of slide prepared using Denhardts and formamide, which s hows a higher N 1s si gnal, an indication that both formamide and Denhardts physisor b. B) XPS spectrum of slide prepared using only formamide. Binding Energy (eV) N(E)Min: 1507Max: 70299 462415.8 369.6 323.4 277.2 231 184.8 138.6 92.4 46.2 0 Si 2p3 2.3 % P 2p3 2.1 % Si 2s Zr 3d 3.1 % C 1s 6 Zr 3p3 Zr 3p1 N 1s 2.3 % B Binding Energy (eV) N(E)Min: 1507Max: 70299 462415.8 369.6 323.4 277.2 231 184.8 138.6 92.4 46.2 0 Si 2p3 2.3 % P 2p3 2.1 % Si 2s Zr 3d 3.1 % C 1s 6 Zr 3p3 Zr 3p1 N 1s 2.3 % B Binding Energy (eV) N(E)Min: 1353Max: 85323 451405.9 360.8 315.7 270.6 225.5 180.4 135.3 90.2 45.1 0 Si 2s Si 2p3 1.9 % P 2p3 2.0 % Zr 3d 2.9 % C 1s 60.3 % Zr 3p3 Zr 3p1 N 1s 4.5 % A Binding Energy (eV) N(E)Min: 1353Max: 85323 451405.9 360.8 315.7 270.6 225.5 180.4 135.3 90.2 45.1 0 Si 2s Si 2p3 1.9 % P 2p3 2.0 % Zr 3d 2.9 % C 1s 60.3 % Zr 3p3 Zr 3p1 N 1s 4.5 % Binding Energy (eV) N(E)Min: 1353Max: 85323 451405.9 360.8 315.7 270.6 225.5 180.4 135.3 90.2 45.1 0 Si 2s Si 2p3 1.9 % P 2p3 2.0 % Zr 3d 2.9 % C 1s 60.3 % Zr 3p3 Zr 3p1 N 1s 4.5 % A
79 Figure 5-3. Fluorescence image of hybridized spots comparing two hybridization methods. A) Fluorescence image comparing the two hybridization methods. B) Drawing indicating which part of the slide corr esponds to which method and which spots correspond to either the probe with poly-A spacer or poly-G spacer. The 4 spots on left were hybridized with the original method, with the Denhardts solution and formamide and the 4 spots on the right are the nitrogen-free method. As the illustration shows, in the nitrogen free me thod, the target solution does not cover the whole slide, just the area where the spot is. Also there appears to be a bad area in the zirconium-phosphonate film, which led to strong adsorption of the target. Table 5-1. Comparison of surface coverage of the DNA before hybridizati on, after hybridization with fewer water rinsings (After Hyb-1) and after hybridization with more water rinsings (After Hyb-2) for probe contai ning either a poly-A spacer or poly-G spacer. 1.7 1011 .8 10102.1 1011 .2 10102.8 1011.1 1010Poly-G 7.0 1010 .8 10101.2 1011.0 10101.4 1011.3 1010Poly-A After Hyb-2 (DNA strands/cm2) After Hyb-1 (DNA strands/cm2) Before Hyb (DNA strands/cm2) 1.7 1011 .8 10102.1 1011 .2 10102.8 1011.1 1010Poly-G 7.0 1010 .8 10101.2 1011.0 10101.4 1011.3 1010Poly-A After Hyb-2 (DNA strands/cm2) After Hyb-1 (DNA strands/cm2) Before Hyb (DNA strands/cm2) A G G G G A A A Original method Nitrogenfree method Bad spot? Bad spot? A B A A G G G G G G G G A A A A A A Original method Nitrogenfree method Bad spot? Bad spot? A B
80 CHAPTER 6 LANTHANIDE MONOLAYERS AS SUBSTRATES FOR PROTEIN MICROARRAYS PART 1: PREPARATION OF A ROBUST LANTHANIDE MONOLAYER Introduction With the success of DNA microarrays, many scie ntists are trying to expand this technology to other biomolecules, such as carbohydrates or proteins. Protein microarrays work similarly as DNA microarrays, in that there is a probe mol ecule immobilized on a su rface and a fluorescently labeled target interacts with some of the probes.75 Thus, the probe targ et interaction can be followed using fluorescence imaging. The possibl e uses for protein arrays cover a wide area including antibody-based th erapeutics and general laboratory research investigating proteinbiomolecule interactions.76 While the use of DNA microarrays is well established in the laboratory, the science of prot ein microarrays is still being developed. There are several differences between DNA and proteins that hinde r the construction of a universal, protein microarray. Proteins are more complex and di verse than DNA; where one substrate may work with one type of protein, it may not work with another. In the literature, there are a wide variety of substrates proposed for protein microarrays. These include glass-, gold-, and polymer-based substrates; each has their advantages and disadvantages. Glass slides ar e generally modified with an alkylalkoxysilane (siloxane), which should provide binding sites and st ability from protein denaturati on. There are various methods of coupling the proteins on a gl ass microarray platform, for example, an amine coated slide, which electrostatically binds protei ns, or an epoxy coated slide, wh ich covalently binds with Lys, Gln, or Arg in the protein.76 These examples do not control the orientation of the protein on the surface, which has been shown to play a consid erable role in the proteins functionality.77 To get around this problem, a Ni-NTA (nitrilotriacetic acid) surface has been developed.78
81 The Ni-NTA support binds the pr otein through a 6-hist dine oligomer called histag. With this system, all proteins should be oriented in one known direction. Also, the Ni-NTA surface is very selective for the histagged proteins. Th ere are still some draw backs with the Ni-NTA surface. One being that the bindi ng of the histag to the nickel is easily reversible; with minor changes in pH or additives, bound protein can wash away.79 The chapter describes work towards a proposed new metal-organic s ubstrate and binding tag system to be used for protein microarrays. The substrate will consist of a lanthanide, Tb3+, Eu3+, or Sm3+, monolayer. The probe molecule will th en bind to the lanthanide through a sensitizing, binding tag consisti ng of amino acids. The optimum sequence of amino acids will be determined in the course of the experiments, but will contain tryptophan, which will act as the fluorescence sensitizer to the lanthanide, and either aspartic acid or glutamic acid, which will bind strongly to the lanthanide. Two different approaches were taken in the development of the lanthanide monolayer. One approach, the Langmuir-Blodgett method, used an alkyl-phosphonate surfactant, which creates a lanthanide-phosphonate network on the surface. The othe r approach investigated uses self-assembly on a glass surface and a covale nt linkage between a siloxane-containing lanthanide-binding ligand (N-trimethoxysilylpropyl)ethylenedia mine triacetic acid trisodium salt (EDTA-siloxane) and the glass surface. These con cepts are demonstrated in Figure 6-1. There is one report in the literature of using the EDTA-siloxane as a gadolinium chelator in silica nanoparticles.80 Lanthanides can be sensitized by a large number of different lig ands. In general the ligand is excited and energy transfer occurs from the tr iplet excited state of th e organic ligand to the lower level excited states of the lanthanide, as represented in Figur e 6-2. The sensitization of the
82 lanthanide fluorescence by polypeptides is based on r ecent literature of lanthanide binding tags. These tags are highly fluorescent an d bind strongly to lanthanides. This microarray system would have several a dvantages over those currently used. First, each probe molecule will contain this tag, which will give rise to fluorescence once it is bound to the lanthanide, confirming the orientation of the protein on the surface. Second, using the fluorescence of the lanthanide and that of the target fl uorophore a ratio can be taken for all spots of the microarray. This will give a ratiometric account of the probe-target interaction strength. Third, the polypeptide tag will be developed to bi nd strongly to the lanthanides, unlike the Nihistag system. Finally, because the tag will consist of natural amino acids, the tagged proteins can be synthesized using DNA containing the code for the tag sequence. Experimental Materials The reagents were purchased through commerc ial sources, except the octadecylphosphonic acid (ODPA) which was synthesized by our collaborators in Dr. Bujolis laboratory in Nantes, France. (N-trimethoxysilylpropyl) ethylenediamine triacetic acid trisodium salt was purchased from Gelest. To form samarium chloride, HCl was simply added to samarium oxide (Sm2O3) and any excess HCl was evaporated off. To form the terbium chloride, terbium oxide (Tb4O7) was refluxed in HCl, and any excess HCl was evaporated off. Lanthanide Phosphonate Langmuir-Blodgett Films Substrate preparation. Glass microscope slides were us ed as the supports for most of the experiments. The glass slides were cleaned with piranha etch and the RCA method, and made hydrophobic by treating with a 5mM so lution of octadecyltrichloros ilane (OTS) in bicyclohexyl for 2 minutes, rinsing with toluene fo r 30 seconds, and then drying with N2. The slides were treated again with the OTS, rinsed with toluene, and finally dried with N2.
83 LB film formation. A KSV 2000 LB double-barrier Te flon trough, purchased from KSV Instruments (Stratford, CT), was used to form the ODPA monolayers at th e air-water interface. A filter-paper Wilhelmy balance attached to a KSV microbalance was used to measure surface pressure. The lanthanide-phosphonate films were made by first spreading a 0.3 mg/mL solution of ODPA in chloroform on a 2.6 mM aqueous CaCl2 subphase, which had a pH adjusted to 7.8 using a KOH solution. The ODPA was compressed at the rate of 10 mm/min to a pressure of 20 mN/mm. Once the target pressure was reached, a hydrophobic glass slide was dipped 50mm into the subphase at a rate of 8 mm/min. The slide wa s then lowered into a gl ass vial in the trough. A volume of the lanthanide chloride solution was added to the vial with the OPDA coated slide to reach a concentra tion of 1.5 mM of Ln3+ in the vial with the slide. The slides sat for at least 7 days in the Ln3+ solution. If the slides were not used right away, they were stored in a 1.5 mM, pH 6 Ln3+ solution until later use. Before this previous procedure was tried, we also attempted to make the films by incorporating the lanthanide ion in the subphase, but we did not find much success with this method. Self-Assembled Lanthanide Films Glass supports were cleaned us ing the RCA cleaning method and piranha etch. They were then submerged in a solution cont aining 4% EDTA-siloxane, 0.086 M Tb3+, and 0.02% acetic acid. The slides were heated overnight at 90 C. The EDTA-siloxane so lution was poured out. The slides were rinsed 2 times with water and th en sonicated with more fresh water and finally rinsed 2 more times with water. The slides were stored in water until later use.
84 To test the stability of the f ilm, when exposed to different chelating ligands, the slides were submerged in either a 1 mM, pH 10 dipicoli nic acid solution; 1 mM, pH 7.5 picolinic acid; 1 mM, pH 9.5 picolinic acid. Analysis XPS was performed using a UHV XPS/ESCA PH I 5100 system. Survey scans were taken with either a Mg or Al K X-ray source using a power settin g of 300 W, a pass energy of 89.4 eV and a take off angle of 45 with respect to th e surface. Using commercial XPS analysis software and Shirley background su btraction, the peak areas we re determined. The elemental percentages were determined using the peak area s and sensitivity factors given by the maker of the XPS instrument. Tapping-mode AFM was carried out on air-dri ed samples using a Multimode AFM with a Nanoscope IIIa controller (Digital Instruments, Santa Barbara, CA) and commercially available silicon cantilever probes (Nanosensors, Phoenix, AZ). Results and Discussion Langmuir-Blodgett Lanthanide Monolayers The general procedure for making the lanthanide LB films is shown in Figure 6-3. The transfer ratios of ODPA onto the OTS coated slid es were always around unity. After one week of sitting in the Ln3+ solutions, the slides were taken out of the solution and were observed for hydrophilicity. If a slide did not appear hydrophi lic over the entire 5 cm where the film was supposed to be, it was considered bad and the sl ide was discarded. Both the samarium films and terbium films exhibited hydrophili city where the film should have been. This was the first indication that the terb ium and samarium had formed meta l-phosphonate monolayers. Only the films that were continuously hydrop hilic were analyzed further. The purpose of the Ca2+ in the subphase is to supply st ability by forming a weak network with the phosphonate groups, whic h aids the transfer of the ODPA onto the slide. The Ca2+
85 should be replaced by the other metal ion, i.e. samarium, or terbium, while the slide sits in the metal ion solution. In the an alysis of the samarium and te rbium-phosphonate monolayers, the XPS data provides information in to whether there is residual Ca2+ in the film and whether there is an equal elemental percentage of phosphorus and lanthanide. The XPS images in Figure 6-4 give the elemental percentages for the samari um-phosphonate and terbium-phosphonate films. First, the samarium-phosphonate film does not appear to have any calcium, which would give rise to a Ca 2p3 peak around 346-347 eV. Furthe rmore, the XPS results for the samariumphosphonate film give the ratio of samarium to p hosphorus at 1.2:1. These results are based on the Sm 4d peak and the P 2s peak. For the terbium-phosphonate films, first, ther e is no Ca 2p peak around 346 eV. On the other hand, the terbium to phosphorous ratios are about 43:1 calculated using the Tb 3d5 and P 2p3 peaks. The high percentage of terbium is an indication that there is more than just a monolayer of it. There are other indicators of this including the attenuation of the Si 2p3 peak from the glass support in the te rbium films as compared to th e Si 2p3 peak in the samariumphosphonate films. The peaks from silicon will be more attenuated as the terbium overlayer thickness increases. Overall, this indicates that the film does not consist only of a terbiumphosphonate monolayer. Most likely there is a precipitation of terbium oxide/hydroxide on the surface. Whether there is a terbium-phosphonate la yer underneath the terbium oxide/hydroxide is unclear from the XPS results. The AFM images (Figure 6-5) support the XPS results. An AFM image of a zirconiumphosphonate film is shown for reference. The film is smooth with some holes, which allow depth analysis. The holes in the zirconiumphosphonate film have a depth of 25 This corresponds to the height of the ODPA and zirconium layer. The samarium-phosphonate film is
86 not as smooth as the zirconium-phosphonate film but does have the same topography as seen with the zirconium-phosphonate films and does have a film height of approximately 25 The terbium film appears smooth, but it does not have the typical topography that the samarium and zirconium-phosphonate films have, which not only ma kes it difficult to measure the height of the film. The topography and XPS results indicate that a monolayer of terbium is not being achieved, but most likely a terbium oxide/hydroxide film is forming on the slide. Currently, samarium and terbium are the only la nthanides that have be en investigated for forming a metal-phosphonate monolayer. Because the lanthanides exhibit similar chemical properties, the binding behavior of the lanthanides should be sim ilar. For this reason, if a method to make a lanthanide-phos phonate monolayer works with one lanthanide, it should work with the other lanthanides. The zirconium-phosphonate films ar e stored in water after the formation of the film, but the samarium-phosphonate films do not appear to be stable in water. Although, there is no data to quantify this, it is obvious th at the samarium-phosphonate films loose their hydrophilicity after sitting in water for several hours. This sugge sts that the samarium is dissolving off the phosphonate headgroup. Without the network of metal ions at th e surface, the monolayer is no longer as robust and washes off the OTS coat ed support. The samarium-phosphonate films do appear stable, i.e. they do not loose their hydrophilicity, if th ey are sitting in a 3 mM, pH 6 Sm3+ solution. Self-Assembled Lanthanide Films on Glass Based on the lanthanide-phosphonate films, a mo re stable system was needed that would not dissolve in water. One way to resolve this problem was to use a ligand covalently linked to the glass support that could also bind the lanthanide. The commercially available EDTAsiloxane was a relatively inexpensive and conve nient answer to this problem. Also, the
87 preparation of the EDTA-l anthanide coated slide was a simple process. There were two possible methods for immobilizing the EDTA-lanthanide co mplex. The EDTA-siloxane could first be immobilized on the surface and then the surface could be exposed to a lanthanide solution. The other possibility is to first prepare the ED TA-lanthanide complex in solution and then immobilize this on the surface. The latter method was used, even though there might be a possibility of having more th an one EDTA-siloxane per lant hanide, because it was thought a higher lanthanide surface concentr ation would be obtained this way. Terbium was used in these experiments. The XPS survey spectrum of the lanthanide-E DTA siloxane complex immobilized on glass is shown in Figure 6-6. The XPS spectra are labeled with el emental percentages. In the lanthanide-EDTA film the elemental percentages that give information about the film are the nitrogen: 2.6%, carbon: 28.1%, sodium: 0.5%, terbium: 11.1%, and silicon: 10.7%. An XPS survey spectrum of bare glass, gives a silicon percentage of 25.5 % and carbon percentage of 8.6%. The attenuation of the si licon peak with the lanthanide-EDTA film is one of the first indications that film formation was successful. The carbon on bare glass is due to adventitious carbon contamination. If the carbon contamination is taken into account, the nitrogen to carbon works out to be approximately 1:7, the expect ed ratio of the EDTA-siloxane based on the molecular formula. However, the terbium to n itrogen ratio is higher than expected. We would expect the terbium to nitrogen rati o to be about 1:2, but ratio turns out to be around 4:1. It is possible this is due to a formation of a thin layer of terb ium oxide on surface. Three different solutions were investigated to determine robustness of the lanthanide film to the chelators dipicolinic and picolinic acid (F igure 6-7). These three solutions were 1 mM, pH 10 picolinic acid; 1 mM, pH 9.5 picolinic acid; and 1 mM, pH 7 picolinic acid. Based on the
88 literature, dipicolinic acid binds strongly to terbium in a wide pH range, but more strongly at a higher pH, around 10 or 11.81 Picolinic acid binds less strongly to terbium and in a narrower pH range, with a maximum around 7 or 8.81 After the lanthanide-EDTA surface is treated with pH 10, 1 mM picolinic acid, it can be seen that th e terbium peaks are no longer present in the XPS spectrum (Figure 6-8). It appears that the dipicolinic acid is pu lling the terbium off the surface. After the lanthanide-EDTA surface is treated w ith 1 mM, pH 9.5 picolinic acid, a good portion of the terbium remains on the surface, with the terbium to nitrogen ratio at 1:1 (Figure 6-9). Again, this ratio is still higher th an the expected ratio of 1:2. Wh en 1 mM, pH 7 picolinic acid is used, a larger decrease in terbium is seen, with the terbium to nitrogen ratio at 1:1.5 (Figure 610). However, the terbium to nitr ogen ratio is still higher than the expected 1:2. At pH 7, the literature shows that the terbium-picolinic acid interaction is much gr eater than at pH 9.5.81 The XPS data shows that when the lanthanide -EDTA slide is placed in a solution of an efficient chelator, such as dipicolinic acid, th e terbium is stripped off the surface. When a weaker chelator is used, picolini c acid, pH 9.5, more of the terb ium stays on the surface. Also, a slight decrease in the ni trogen signal is seen after treatment with the two more efficient chelators. The terbium-EDTA film has a nitrogen percentage of 2.6%, after treatment with pH 7 picolinic acid the nitrogen decreases to 1.8% and after treatment with the di picolinic acid the percentage decreases to 1.1%. This shows that with the te rbium some of the EDTA-siloxane is coming off, which may inidicate that the terbium is binding more than one EDTA-siloxane. If the terbium binds more than one EDTA-siloxa ne, it is unlikely that more than one EDTA-siloxane is bound to the glass. Therefore, the EDTA-siloxane whic h is not bound to the glass, comes off with the terbium.
89 These studies demonstrate that a terbium ED TA-siloxane surface was prepared on a glass support. The terbium EDTA monolayer is stable in water and some weakly chelating solutions. However, when the surface is placed in a solutio n of a ligand which binds efficiently to the terbium, the ligand strips the terbium from the EDTA. Most likely, the EDTA ligand is a stronger chelator than dipicolinic acid, but because there is so much more of the dipicolinic acid, the terbium is taken off the slide by the dipicolinic acid. It would be intere sting to try a chelating ligand, such as dipicolinic acid, at a much smalle r volume and concentration, similar to that used when spotting proteins for microarrays. It shou ld be mentioned that preliminary fluorescence spectroscopy was attempted with the picolinic and dipicolinic treated slides. However, no appreciable luminescence was seen. It is possible to try an even stronger chelating ligand to hold the terbium on the surface. Nevertheless, the lant hanide needs binding sites left available for the lanthanide binding tag on the protein. Conclusion and Remarks A method for making samarium phosphonate monolayers was developed, which with minimal changes should be applicable to other la nthanides. Preliminary experiments were tried with terbium, but it appears to be forming more than a monolayer. Most likely with the right conditions a terbium-phosphonate monolayer should form. The samarium films appear to have inferior stability, they dissolv e in water, compared to the zirconium-phosphonate films, but do appear stable in a Sm3+ solution. More stable lanthanide films were made using terbium and an EDTA-siloxane. These films proved stable in water, but when placed in a solution of an efficien t chelating ligand, the terbium was removed from the film. It would be interesting to investigat e the stability of the terbium, when a much smaller volume and concentration, similar to that used with the microarray spotting of proteins.
90 Ln P P H N O O O O LnLnLnAB LanthanideBinding Tag Lanthanide Substrate N N O O O O O Si O O O O H N O O O O Ln Ln P P H N O O O O LnLnLn P P H N O O O O LnLnLnAB LanthanideBinding Tag Lanthanide Substrate N N O O O O O Si O O O O H N O O O O Ln N N O O O O O Si O O O O H N O O O O N N O O O O O Si O O O O H N O O O O N N O O O O O Si O O O O H N O O O O Ln Figure 6-1. Two approaches expl ored here to prepare a lanthani de substrate for protein binding microarrays. A) The Langmuir-Blodgett met hod for preparing the substrate. B) The self-assembly method using a siloxane-cont aining lanthanide-b inding ligand. The molecule binding to the lanthanide is a hypothetical illustration of a polypeptide binding to the lantha nide substrate.
91 Figure 6-2. The intramolecular energy transfer that can occur between a ligand and lanthanide. Figure 6-3. Procedure of th e Langmuir-Blodgett method used to make the metal-phosphonate monolayers. Note that for clarity the Ca2+ in the subphase is not shown in the figure. So* S1* S2* T1*LigandMetal Energy So* S1* S2* T1*LigandMetal So* S1* S2* T1*LigandMetal EnergyLn3+( ) STEP 1: Transfer template from water surface STEP 2: Add Ln3+ CH3(CH2)17PO3H2Ln3+( ) Ln3+( ) STEP 1: Transfer template from water surface STEP 2: Add Ln3+ CH3(CH2)17PO3H2
92 Binding Energy (eV) N(E)Min: 1330Max: 190370 1100990 880 770 660 550 440 330 220 110 0 Sm 4d 3.0 % Ca 2p 0.5 % Sm 4p3 P 2s 2.4 % Si 2p3 8.2 % Si 2s C 1s 57.3 % O 1s 28.6 % O KLL Sm 3d5 Binding Energy (eV) N(E)Min: 2140Max: 370633 13001170 1040 910 780 650 520 390 260 130 0 Tb 3d5 34.6 % P 2s P 2p3 0.6 % Si 2p3 2.5 % Si 2s C 1s 36.3 % O 1s 26.1 % O KLL Tb 3d3 Binding Energy (eV) N(E)Min: 1330Max: 190370 1100990 880 770 660 550 440 330 220 110 0 Sm 4d 3.0 % Ca 2p 0.5 % Sm 4p3 P 2s 2.4 % Si 2p3 8.2 % Si 2s C 1s 57.3 % O 1s 28.6 % O KLL Sm 3d5 Binding Energy (eV) N(E)Min: 2140Max: 370633 13001170 1040 910 780 650 520 390 260 130 0 Tb 3d5 34.6 % P 2s P 2p3 0.6 % Si 2p3 2.5 % Si 2s C 1s 36.3 % O 1s 26.1 % O KLL Tb 3d3 Figure 6-4. XPS survey scan of lanthanide phosphonate films. A) Samarium-phosphonate film. B) Terbium-phosphonate film. A B
93 Figure 6-5. AFM images of zirconium, samarium, and terbium phosphonate films Sm Tb Zr Sm Tb Sm Tb Zr Zr Zr Sm Tb
94 Figure 6-6. XPS spectrum of se lf-assembled EDTA-siloxane lanthanide film. The insert is a magnification of the N 1s region. Figure 6-7. Molecular structure of dipicolinic acid and picolinic acid. Dipicolinic Acid Picolinic Acid N OH HO O O N OH O
95 Figure 6-8. XPS spectrum of se lf-assembled EDTA-siloxane lanthanide film after treatment with dipicolinic acid. The insert is a magnification of the N 1s region.
96 Figure 6-9. XPS spectrum of se lf-assembled EDTA-siloxane lanthanide film after treatment with pH 9.5 picolinic acid. The insert is a magnification of the N 1s region.
97 Figure 6-10. XPS spectrum of self-assembled ED TA-siloxane lanthanide film after treatment with pH 7 picolinic acid. The insert is a magnification of the N 1s region.
98 CHAPTER 7 LANTHANIDE MONOLAYERS AS SUBSTRATES FOR PROTEIN MICROARRAYS PART 2: A MODEL STUDY Introduction It has been known for quite some time that certain ligands can bind to lanthanides and through an intramolecular energy transfer from th e ligand to the lanthanide, a dramatic increase in luminsence from the lanthanide can be seen.82 This energy transfer occurs from the triplet excited state of the organic li gand to the lower level excited states of the lanthanide, as represented in Figure 6-2. The po ssible use of lanthanides as lumi nescent indicators of protein binding to surfaces, as shown in Chapter 6, brou ght up the question of the luminescent behavior of lanthanides in a monolayer. A project was developed to stu dy the ligand-enhanced luminescent behavior of lanthanide s at the air solid interface. The aim of this project was to investigate the luminescent behavior of a monolayer of sensitized lanthanides on a gla ss surface. Specifically, we wanted to study the intermolecular energy transfer from one lanthanide to a different lanthanide, such as from terbium to europium. One of the first studies of lanthanide intermol ecular transfer was done in dimethyl sulfoxide between europium and terbium.83 It is thought that in solution it is necessary for the lanthanides to form polymeric structures, or in other words to share ligands, in order for the intermolecular lanthanide energy transfer to occur.81 However, there has been one group to study this intermolecular lanthanide ion energy transfer in multilayer LB (Langmuir-Blodgett) films, where the lanthanides are not a pparently sharing ligands.84 To perform this stu dy they used films of different molar percentages of Eu(TTA)3Phen (TTA=2-thenoyltrif luoroacetone; Phen=1,10phenanthroline) with Gd(TTA)3Phen coexisting with arachidic aci d. They developed a model of their system called active enhancem ent circle, which showed that there is must be a minimum distance between the donor and acceptor for the en ergy transfer to occur. In the case of
99 gadolinium and europium, they found this distance to be 1.2 nm. Fr om previous research it was already known that the energy was transfer from the donor to the accepto r ion through a triplet excited state of TTA.85 In addition to this energy transfer study, there have been a number of other studies that have looked at the luminescence of lanthani des in LB monolayers and thin films.86, 87 In this project, we are attempting a similar st udy, but using self assembly on a glass surface to make the lanthanide monolayer and also using a different ligand system. The first part of the project involved the developmen t of a ligand that would hold th e lanthanides on a surface and sensitize the luminescence through an intramolecu lar energy transfer from the ligand to the lanthanide ion. The second part of the projec t would be studying the lu minescence behavior of mixed lanthanide monolayers made various ways. The lanthanide ratios would be varied as well as the surface density. The surface modification and lanthanide ratios would be followed with XPS. The ligand that was chosen for this sy stem contains a chelidamic acid moiety (4hydroxypyridine-2,6-dicarboxylic acid) which clos ely resembles the well known sensitizer of lanthanides, dipicolinic acid (pyridin e-2,6-dicarboxylic acid) (Figure 7-1).81 Several different routes, as shown in Figure 7-2, were explored to find a suitab le ligand, which contained the chelidamic acid moiety and would react with an amine-reactive modified glass surface or directly with a glass surface. Unfortunately, this ligand in pure form has not yet been synthesized. This rest of this chapter outlines the procedures for the reactions that were attempted and a short explanation of the different synthetic routes. Experimental Reagents were ordered from Acros or Aldrich an d used without further purification, except when stated. Samarium oxide wa s converted into the chloride by treating with concentrated HCl.
100 Diethyl 2,4,6-trioxoheptanedioate (1). The synthesis for 1 and 2 was taken from Riegel et al.88 250 mL of ethanol was distilled over Na into a 3-neck round bottom flask under argon. A condenser under argon was added to the 3-neck flask and then 16.2 g (0.7 moles) of sodium was added slowly to the ethanol. The sodium a nd ethanol were allowed to react to completion with heating at the end. Usi ng a large graduated pipet, 100 mL of the sodium ethoxide solution was transferred to another 3-neck round botto m under argon. The remaining 150 mL was kept warm. To the 100-mL sodium ethoxide, 50 g (0. 34 moles) of diethyl oxalate and 19.7 g (0.34 moles) of acetone was added with stirring, which resulted in ra pid precipitation of a very thick yellow solid. Then, 50 g (0.34 moles) of diethyl oxalate was added with stirring to the remaining 150 mL of the sodium ethoxide solution and fina lly this solution was added to the sodium ethoxide/acetone/diethyloxalate flask. This th ick mixture was stirred for 1 hour and then the ethanol was taken off under vacuum until dry. To the dark greenish-yellow solid was added 350 g of ice and 100 mL of concentrated HCl. Th is yellow slurry was st irred so that all the clumps were broken up and then the yellow paste was filtered, washed several more times with ice water until a pale yellow paste was obtained and then dried under vacuum. These procedures gave 70 g of crude product ( 80% yield). 1H (300 MHz, CDCl3) H: 6.9 (s). 13C (75 MHz, CDCl3) C: 14.51, 63.24, 104.46. 162.34. 4-Oxo-4H-pyran-2,6-dicarboxylic acid (chelidonic acid) (2). In a 200mL round bottom flask, with an attached reflux conde nser, was added 70 g (0.27 moles) of 1 and 100 mL of concentrated HCl. This mixture was refluxed at 100 C for 20 hours. The reaction was allowed to cool to room temperature and 100 g of ice was added. The precipitate was filtered and washed several times with 30 mL of cold water. The pink solid was dried unde r vacuum. 29 g (56%
101 yield). 1H (300 MHz, DMSO-d6) H: 6.9784. 13C (75 MHz, DMSO-d6) C: 119.71, 154.71, 161.55, 180.11. 4-Hydroxypyridine-2,6-dicarboxylic acid (chelidamic acid) (3). The synthesis for 3 was taken from King et al.89 In a 250-mL round bottom flask wa s added 29 g of 2 and 155 mL of 10% NH4OH was added. This solution was allo wed to reflux for 4 hours and with each hour, 8.5 mL of concentrated NH4OH was added. The water was removed under vacuum, followed by the addition of 150 mL of cold wa ter and 25 mL of concentrated HC l. The beige precipitate was filtered and washed several times with cold water. 1H (300 MHz, DMSO-d6) H: 7.60 (s). 13C (75 MHz, DMSO-d6) C: 110.91, 145.33, 161.47, 162.67. Diethyl 4-hydroxypyridine-2,6-dicarboxylate (4). The synthesis for 4 and 5 were taken from Cooper et al.90 In a 100-mL flask with an attached reflux c ondenser was added 25mL 100% ethanol at 0 C under N2. To the same flask was slow ly added 6.2 mL (85 mmol) of thionyl chloride with stirring. Then 2.5 g (13.7 mmol) of 3 was added. This solution was stirred at room temperature for 18 hours and then refluxed for 2 hours. The solvent was removed under reduced pressure, followed by the addition of 20 mL cold water. The solution was neutralized with 5 mL of cold 10% aqueous Na2CO3, upon which a light beige precipitate formed, and then 5 mL of cold 50% aqueous ethano l. The solution was allowed to stir overnight, so that the precipitate was less clumpy. 3.24 g (96% yield). 1H (300 MHz, CDCl3) H: 1.42 (t, 6H), 4.47 (q, 4H), 7.34 (bs). Diethyl 4-(10-bromodecyloxy)py ridine-2,6-dicarboxylate (5). 100 mL of acetone was distilled over anhydrous K2CO3 into a 250-mL flask contai ning 2.5 g (10.4 mmol) of 4 and 6.3 mL of 1,10-dibromodecane. To this solution was added 2.32 g of K2CO3. This solution was allowed to reflux for 40 hours under argon while following with TLC, after which the solvent
102 was removed under reduced pressure. The oil wa s purified using a column chromatography with silica as the stationary phase. Hexane was us ed to elute the excess 1,10-dibromodecane, which was followed with methylene chloride to elute th e product. The solvent was removed with rotary evaporation to give a pale yello w oil. 2.5 g (53% yield). 1H (300 MHz, CDCl3) H: 1.39 (m, 18H), 1.85 (m, 4H), 3.41 (t, 2H), 4.13 (t, 2H), 4.47 (q, 4H). 13C (75 MHz, CDCl3) C: 14.40, 26.01, 28.36, 28.92, 29.38, 29.51, 29.56, 62.57, 69.20, 114.59, 150.23, 165.01, 167.23. Diethyl 4-(10-cyanodecyloxy)pyridine-2,6-dicarboxylate (6). The synthesis for 6 was adapted from information given by Vogel.91 Into a 3-neck flask, was distilled under reduced pressure 15 mL of DMF dried over anhydrous MgSO4. To the same flask was added 1.5 g (3.3 mmol) of 5, which was followed with the addition of 0. 7 g (10.75 mmol) of potassium cyanide. This solution was allowed to stir for 2 hours. To the DMF solution was added 10 mL of water and then this solution was extracted 3 times with 20 mL of a 1:1 mixtur e of diethyl ether and petroleum ether. The organic pha se was washed 2 times with 30 mL of brine and then dried with MgSO4. The organic phase was retained and the so lvent was removed by rotary evaporation. 0.25 g (18% yield). 1H (300 MHz, CDCl3) H: 1.40 (m, 18H), 1.65 (m, 2H), 1.83 (m, 2H), 2.34 (t, 2H), 4.13 (t, 2H), 4.47 (q, 4H), 7.77 (s, 2H). 13C (300 MHz, CDCl3) C: 14.43, 26.05, 28.34, 28.95, 29.40, 29.54, 29.59, 33.03, 34.24, 62.60, 69.21, 114.55, 150.27, 165.05, 167.26. Diethyl 4-(11-aminoundecyloxy)py ridine-2,6-dicarboxylate (7). The synthesis for 7 was adapted from a procedure shown by Borkowski et al.92 The reduction of the nitrile was attempted using a Co2+ and NaBH4 mixture. In 21 mL of 100% ethanol, 0.25 g (0.62 mmols) of 6 was dissolved. To the solution was added 0.24 g of CoCl2 6H2O. The cobalt was allowed to dissolve and then 0.21 g of NaBH4 was added to the solution. Af ter the reaction had stirred for 3 hours, 3 M HCl was added until all of the solid dissolved. The solu tion was extracted twice with
103 10 mL of diethyl ether. The aqueous pha se was basified with concentrated NH4OH and then extracted with 4 10-mL portions of diethyl ether. After extrac ting the organic phase with an equal volume of brine, it was dried with MgSO4. Finally, the solvent was removed under reduced pressure. Only a small amount of a residue was obtained as a product, and an NMR taken in CDCl3 showed significant degradati on of the parent molecule. Diethyl 4-(10-azidodecyloxy)py ridine-2,6-dicarboxylate (8). The synthesis for 8 was adapted from a procedure written by Roy et al.93 In a round bottom flask under argon, 1.5 g (3.3 mmols) of 5 was dissolved in 30 mL of dry DMF. To this was added 0.29 g (4.4 mmols) of sodium azide and the flask was heated at 90 C for 14 hours. After the reaction had cooled to about 50 C, the DMF was removed under vacuum. To the residue was added 15 mL of diethyl ether and 20 mL of water. Th e aqueous portion was extracted twi ce with 15 mL of diethyl ether and then the combined diethyl ether extractions were extracted once w ith 30 mL of water and once with 30 mL of brine. After whic h, the organic portion was dried with MgSO4, filtered, and rotary evaporated. This gave 1 g (75% yield) of a viscous oil. 1H (300 MHz, CDCl3) H: 1.37 (m, 18H), 1.58 (m, 2H), 1.82 (m, 2H), 3.24 (t, 2H ), 4.12 (t, 2H) 4.46 (q, 4H), 7.76 (s, 2H). 13C (300 MHz, CDCl3) C: 14.44, 26.06, 26.93, 28.96, 29.05, 29.33, 29.41, 29.58, 51.67, 62.59, 69.21, 114.83, 150.34, 164.81, 167.30. Diethyl 4-(10-aminodecyloxy)py ridine-2,6-dicarboxylate (9). Several methods were explored for the reduction of the azide to an amine. The first method employed H2 gas and 10% palladium on carbon catalyst, which was ad apted from a procedure written by Roy et al.93 To use this method, a 3 neck flask containing 0.06 g of 10% Pd/C with a glass tube to disperse the H2 was evacuated and purged with N2 several times. Then, it was evacuated and filled with H2, which was followed with the injec tion of 0.5 g (1.3 mmols) of 8 and 1 mole equivalent of HCl
104 dissolved in 20 mL of methanol. The reaction was stirred with H2 bubbling for 45 minutes. The material was filtered over celite and the solvent was removed using rotary evaporation. 1H NMR was taken in CDCl3, and although an -NH3 peak with a H of 2.94 ppm was seen, degradation of parent compound could be observed by the appearance of multiple aromatic groups. The second method used the Staudinger proc edure, which was adapted from a procedure written by Somfai et al.94 For this method, 0.50 g (1.3 mmols) of 8 was dissolved in 20 mL of 10:1 THF and water and then 382 mg (1.4 mmole s) of triphenylphosphine was added. The solution was allowed to stir overnight, after which the so lvent was removed under reduced pressure. Purification was attempted using a mini column made with a small glass frit funnel and silica gel. First 150 mL of ethyl acetate was run through the column, followed by a mixture of 10:0.1 methanol:NH4OH. A very sticky residue was obtained. By 1H NMR, taken in CDCl3, it was seen to contain a significan t amount of the triphenylphosphine. The third method, which used zinc powder, wa s adapted from a procedure written by Lin et al.95 First, 1.0 g (2.4 mmols) of 8 was dissolved in 3:1 ethanol:water solution. To this solution was added with stirring, 0.29 g of ammonium chloride, which was followed by the addition of 0.300 g of zinc powder. The solution was refluxed for 15 minutes and then filtered to remove the remaining zinc powder. Extraction w ith dichloromethane was attempted, but at this point a gooey precipitate had formed. This precipitate was wash ed with HCl and H2O extensively. The precipitate was dried under vacuum to form a white powder. 1H (300 MHz, D2O) H: 1.21 (m, 13H), 1.81 (m, 2H), 2.61 (t, 2H), 4.22 (t, 2H), 7.54 (s, 2H) Diethyl 4-(10-(3-triethoxysilyl)propylamino) decyloxy)pyridine-2,6-dicarboxylate (10). In 30 mL of dry dichlorometh ane, 0.5 g (1.2 mmoles) of 5 and 0.179 g (1.0 mmoles) of 3aminopropyltriethoxysilane (APTES) was added under N2. The solution was allowed to reflux
105 for 3 days, after which, 10 mL more dichloro methane was added and it was filtered 3 times over celite. The solvent was then remove d using rotary evaporation. An 1H NMR was taken which showed that there was lik ely the desired product mixed with a single by-product. Diethyl 4(3-(trimethoxysilyl)propylcarbam oyloxy)pyridine-2,6-dicarboxylate (11). To make this molecule two differe nt methods were explored. The first method, adapted from Lenaerts et al. ,86 used 0.25 g of 4 and 5 mL of triethoxy(3-isocyana topropyl)silane. The reaction was performed neat with heating at 85 C for 72 hours. Precipitation was attempted out of cold hexane, heptane, and petane, but none of th ese solvents resulted in precipitation. The other condition tried, wh ich was adapted from Cui et al. ,96 used 0.30 g (1.25 mmoles) of 4 and 0.34 mL (1.375 mmols) of triethoxy( 3-isocyanatopropyl)silane and 1 drop of triethylamine in THF. Th e solution was reluxed at 80 C for 48 hours. The material was filtered through celite and the solvent was rem oved with rotary evaporation. An 1H NMR in CDCl3 was taken which showed many by pr oducts and no clear product. Diethyl 4-(3-(trimethoxysilyl)propo xy)pyridine-2,6-dicarboxylate (12). For this reaction 1.5 g (5.2 mmols) of (3-iodopropyl)t rimethoxysilane and 0.5 g (2.1 mmols) of 4 was dissolved in 30 mL of dichlorome thane with 0.37 g of anhydrous K2CO3. This was refluxed for 2 days, after which, it was filtered through celite using dichloromethane. An 1H NMR was taken in CDCl3 which showed no real product. Slide modification. Epoxide coated slides were made using glass slides freshly cleaned with the RCA and then piranha cleaning solutions (See the Experimental section of Chapter 2). The glass slides were placed in a 10% v/v solution of glycidyloxypr opyl trimethoxysilane and toluene. The slides sat in this solution overn ight after which they were washed twice with toluene, sonicated in toluene fo r 15 minutes and then washed agai n two times more with toluene.
106 Immobilization of 9 complexed with Sm3+ was performed on an epoxi de modified slide. To do this a 6.0 10-4 M and 2.0 10-4 M solution of 9 and Sm3+, respectively was prepared at pH 11. The epoxide slide was then subm erged in this solution overnight at 42 C. The slide was then washed repeatedly with water, however, a so lid could still be see on the surface of the glass. Results and Discussion The first area of synthesis focused on making an amine modified chelidamic acid, which would chelate lanthanides and sens itize their fluorescence. As s hown in Figure 7-3, this could then react with a glass surface wh ich has been modified with an amine-reactive molecule, such as an epoxide. Several different synthetic routes were tried. One route investig ated a nitrile modified chelidamic acid which could then be redu ced to the amine. The synthesis of the nitrile gave very low yield. When the reduction wa s attempted using a coba lt borohydride mixture, which is supposed to be a mild procedure, de gradation of the parent molecule occurred. In a second route, an azide modified che lidamic acid was made, which could then be reduced to the amine. The facile synthesis of the azide gave a good yiel d and thus our efforts shifted from the nitrile to the azide. Several different met hods were investigated for the reduction of the azide. These methods included the Staudinger procedur e, where the azide is reduced using tryphenylphosphi ne; a classic reduction in the presence of H2 gas and a palladium catalyst; and a lesser known me thod of using zinc powder. The Staudinger procedure, considered a mild reduction route, resulted in some degradation of the parent compound. Saponification of the ester protectin g the carboxylic acid groups of the chelidamic acid occurred with this procedure as well as with the other azide-reduction pr ocedures. The saponification of the ester greatly decreased the solubility of the mo lecule making purification difficult in general. This further complicated the removal of th e triphenylphosphine oxide by-products of the
107 reaction. When using the H2 gas and a palladium catalyst, there seemed to be a tradeoff which could not be perfected. When the hydrogenation was performed, degradation of the parent compound occurred. The degradation could be lessened by decreasing th e reaction time, but then this led to only partial reduction of the azi de. Again saponification of the ester was seen which complicated the purification. In the zinc reduction, full reduction of the azide could be achieved with little to no degrad ation of the parent compound, th ere was however saponification of the ester. Saponification of th e ester left the molecule spari ngly soluble in basic water making it difficult to wash. However, based on the NMR, th e material appeared to be fairly pure and it was decided to use this product for the next pha se of the project: imm obilization on a surface and subsequent luminescence studies. The product from the zinc reduction was mixed w ith either samarium or terbium and then reacted with epoxy coated glass slides. However when XPS was taken of th is slide, it showed there was still contamination with zinc (Figure 7-4). Preliminar y fluorescence spectroscopy of a terbium slide did show a typical emission picks of terbium. However, because of the poor solubility of the material, it coul d not be certain if all the materi al that was not bound to the glass had washed off. Because of the difficulty in obtaining 9 in pure form and the apparent poor solubility after ester saponificat ion, it was decided that a different approach to forming these luminescent lanthanide films should be explored. The next possibility explored was the preparation of a chelidamic acid containing a siloxane so that it could be immobilized directly onto bare gl ass. Three different synthetic approaches were explored, wh ich tried to make compounds 10, 11, or 12. One difficulty in preparation of silanes is that they cannot be purified through column chromatography. Many of the papers that report the preparation of siloxanes use eith er precipitation or distillation for
108 purification, but neither of th ese methods proved successful in these reactions. NMR of the material from the reactions for 11 and 12 after it had been filtere d through celite showed many side products. The reaction that showed the most promise was that for 10. In the 1H NMR of the material, it showed mostly the product. However, the integration of the NMR indicated that there is a mixture of APTES with two of 5 added onto the amine and APTES with a single 5 added onto the amine, which is to be expected, bu t because the siloxane pu rification with column chromatography is not feasible, separa tion of the two would be difficult. Future Work This project is not yet finished. The target molecule, 9, was apparently synthesized, but was contaminated with zinc. Also, the molecule exhibited poor solubility which would make its immobilization as a monolayer on glass difficult. A different approach was taken, which focused on synthesizing chelidamic acid modi fied silanes, but little progr ess was made. There are still many routes that can be tried for making a siloxa ne modified chelidamic acid. There are also different coupling schemes besides the amine-e poxide method shown in Figure 7-3. Once the chelidamic-lanthanide m onolayer is constructed, the fluores cence experiments would still need to be conducted. Figure 7-1. Two common sensit izers of lanthanides are chelid amic acid and dipicolinic acid. ChelidamicacidDipicolinicacidN O H OH HO O O N OH HO O O ChelidamicacidDipicolinicacidN O H OH HO O O N OH HO O O
109 Figure 7-2. Scheme of different routes for a surface reactive chelidamic acid. The procedures used to prepare these co mpounds are shown in the e xperimental section. 4 5 6 7 8 9 10 11 12 N O O OH O O O Br 10N O O O O O H N 10 Si O O O N O O O O O CN 10N O O O O O 10NH2 N O O O O O N 10N N N O O O O O NH2 10N O O O O O H N O Si O O O N O O O O O SiO O O N O O O O 4 5 6 7 8 9 10 11 12 N O O O H O O O Br 10N O O O O O H N 10 Si O O O N O O O O O CN 10N O O O O O 10NH2 N O O O O O N 10N N N O O O O O NH2 10N O O O O O H N O Si O O O N O O O O O SiO O O N O O O O
110 O O O O O O O O + O O O O O O O O N O H O O O OH (CH2)10 NH2+ O O O O O O O O N O H O O O OH (CH2)10 NH2+ A B O O O O O O O O + O O O O O O O O N O H O O O OH (CH2)10 NH2+ O O O O O O O O N O H O O O OH (CH2)10 NH2+ A B Figure 7-3. Two possible routes of immobilizin g the chelidamic acid/lanthanide complex onto a surface. A) Method involves the reaction of the chelidamic-containing ligand with the surface and then this su rface with the lanthanide solution. B) The route other involves first the reac tion of the lanthanide with the ligand and then immobilization of this complex on the surface.
111 Figure 7-4. Portion of an XPS spectrum where ligand 9 was allowed to react with Sm3+ and then the resulting complex was allowed to r eact with an epoxide surface. The XPS spectrum shows that 9 was contaminated with zinc, coming from the reduction process of the azide. Binding Energy (eV) N(E)Min: 63548Max: 174027 12001176 1152 1128 1104 1080 1056 1032 1008 984 960 Sm 3d5 3.7 % Na 1s 5.5 % O KLL Zn 2p1 Zn 2p3 0.4 % Sm 3d3
112 CHAPTER 8 CONCLUSIONS XPS has proven to be a powerful tool to study DNA on zirconium-phosphonate surfaces. A model was developed, based on th e substrate-overlayer model, for the quantitative calculation of the DNA surface coverage on zirconium-phosphonate monolayers. This model could also be used for the calculation of su rface density of other substances immobilized on a zirconiumphosphonate monolayer. With this model the surface coverage of ssDNA was found to be around 2.8 1011 DNA molecules/cm2 for a probe molecule cont aining a phosphate linker and poly-dG spacer on the 5 end. For a probe with a poly-dA spacer, the surface coverage was calculated as 1.4 1011 DNA molecules/cm2; however, this surface coverage was found to be highly dependent on the rinsi ng conditions. The surface coverage of dsDNA was also investigated. The highest surface coverage, 2.8 1011 dsDNA molecules/cm2, was obtained for dsDNA which had a poly-dG spacer and phosphate linker on both 3 ends. A lower surface coverage, 2.1 1011 dsDNA molecules/cm2, was found for dsDNA which had a poly-dG spacer and phosphate linker on both 5 ends. The lowe st surface coverages we re obtained for dsDNA with a random spacer and phosphate linker on both 5 ends (1.7 1011 dsDNA molecules/cm2) and for dsDNA with a poly-dG spacer and phosph ate linker on the 3 end and 5 end (1.2 1011dsDNA molecules/cm2). XPS was also used to follo w the hybridization efficiency of immobilized ssDNA probes with a target of comp lementary sequence, but no increase in the N 1s signal could be seen after hybr idization, indicating that the amount of target hybridized was too low to be seen with XPS. Finally, the synthesis and immobilization of several lanthanide-chelating ligands was explored. A robust lanthanide monolayer was deve loped using an ethylenediamine triacetic acid ligand with a siloxane moiety. It would be inte resting to test this su rface as a substrate for
113 protein microarrays. The synthesis of a liga nd which would sensitize the luminescence of lanthanides as well as immobili ze the lanthanide on a surface wa s also explored. A molecule containing a chelidamic acid moiety and an am ine group was synthesized but the purification and use of this molecule was hindered by its poor solubility. Again, XPS played a key role in developing and studying the lanthanide monolayers.
114 LIST OF REFERENCES (1) Alivisatos, A. P.; Johnsonn, K. P.; Peng, X.; Wilson, T. E.; Loweth, C. J.; Bruchez, M. P.; Schultz, P. G. Nature 1996, 382 609-611. (2) Braun, E.; Eichen, Y.; Sivan, U.; Ben-Yoseph, G. Nature 1998, 391 775-778. (3) Pease, A. C.; Solas, D.; Sullivan, E. J.; Cronin, M. T.; Holmes, C. P.; Fodor, S. P. A. Proceedings of the National Academy of Sc iences of the United States of America 1994, 91 5022-5026. (4) Smith, S. B.; Finzi, L.; Bustamantet, C. Science 1992, 258 1122-1126. (5) Storhoff, J. J.; Mirkin, C. A. Chemical Reviews 1999, 99 1849-1862. (6) Mucic, R. C.; Storhoff, J. J.; Mirkin, C. A.; Letsinger, R. L. Journal of the American Chemical Society 1998, 120 12674-12675. (7) Becker, C. F. W.; Wacker, R.; Bouschen, W.; Seidel, R.; Kolaric, B.; Lang, P.; Schroeder, H.; Miller, O.; Niemeyer, C. M.; Spengl er, B.; Goody, R. S.; Engelhard, M. Angewandte Chemie International Edition 2005, 44 7635 (8) Georgiadis, R.; Peterlinz, K. P.; Peterson, A. W. Journal of the Americ an Chemical Society 2000, 122 3166-3173. (9) Kimura-Suda, H.; Petrovykh, D. Y.; Tarlov, M. J.; Whitman, L. J. Journal of the American Chemical Society 2003, 125 9014-9015. (10) Gong, P.; Lee, C.-Y.; Gamble, L. J.; Castner, D. G.; Grainger, D. W. Analytical Chemistry 2006, 78 3326-3334. (11) Casero, E.; Darder, M.; Daz, D. J.; Pariente, F.; Martn-Gago, J. A.; Abrua, H.; Lorenzo, E. Langmuir 2003, 19 6230-6235. (12) Petrovykh, D. Y.; Kimura-Suda, H.; Tarlov, M. J.; Whitman, L. J. Langmuir 2004, 20 429-440. (13) Dai, S.; Zhang, X.; Du, Z.; Dang, H. Materials Letters 2004, 59 423 429. (14) Kumar, A.; Pattarkine, M.; Bhadbhade, M.; Mandale, A. B.; Ganesh, K. N.; Datar, S. S.; Dharmadhikari, C. V.; Sastry, M. Advanced Materials 2001, 13 341-344. (15) Wang, J.; Rivas, G.; Jiang, M.; Zhang, X. Langmuir 1999, 15 6541-6545. (16) Zhang, D.; Chen, Y.; Chen, H.-Y.; Xia, X. H. Analytical and Bioanalytical Chemistry 2004, 379 1025.
115 (17) Herne, T. M.; Tarlov, M. J. Journal of the Americ an Chemical Society 1997, 119 89168920. (18) Lee, C.-Y.; Canavan, H. E.; Gamble, L. J.; Castner, D. G. Langmuir 2005, 21 5134-5141. (19) Lee, C.-Y.; Gong, P.; Harbers, G. M. ; Grainger, D. W.; Castner, D. G.; Gamble, L. J. Analytical Chemistry 2006, 78 3316-3325. (20) May, C. J.; Cana van, H. E.; Castner, D. G. Analytical Chemistry 2004, 76 1114-1122. (21) Mohaddes-Ardabili, L.; Martnez-Miranda, L. J.; Silverman, J.; Christou, A.; SalamancaRiba, L. G.; Al-Sheikhlya, M.; Bentley, W. E.; Ohuchi, F. Applied Physics Letters 2003, 83 192194. (22) Sastry, M.; Ramakrishnan, V.; Pattarkine, M.; Ganesh, K. N. Journal of Physical Chemistry B. 2001, 105 4409-4414. (23) Millard, M. M.; M acquet, J. P.; Theophanides, T. Biochimica et Biophy sica Acta, Nucleic Acids and Protein Synthesis 1975, 402 166-70. (24) Higashi, N.; Takahashi, M.; Niwa, M. Langmuir 1999, 15 111-115. (25) Saprigin, A. V.; Thomas, C. W.; Dulcey, C. S.; Patterson, C. H.; Spector, M. S. Surface and Interface Analysis 2004, 36 24-32. (26) Petrovykh, D. Y.; Kimura-Suda, H.; Whitman, L. J.; Tarlov, M. J. Journal of the American Chemical Society 2003, 125 5219-5226. (27) Leavitt, A. J.; Wenzler, L. A.; Williams, J. M.; Thomas P. Beebe, J. Journal of Physical Chemistry 1994, 98 8742-8746. (28) Powell, C. J.; Jablonski, A. NIST Electron Effective-A ttenuation-Length Database, Version 1.0 (NIST SRD-82) 2001. (29) Corrie, S. R.; Lawrie, G. A.; Trau, M. Langmuir 2006, 22 2731-2737. (30) Schena, M.; Shalon, D. ; Davis, R. W.; Brown, P. O. Science 1995, 270 467-470. (31) Lockhart, D. J.; Dong, H. L.; Byrne, M. C.; Follettie, M. T.; Gallo, M. V.; Chee, M. S.; Mittmann, M.; Wang, C. W.; Kobayashi, M.; Horton, H.; Brown, E. L. Nature Biotechnology 1996, 14 1675-1680. (32) Nonglaton, G.; Benitez, I. O.; Guisle, I.; Pipelier, M.; L ger, J.; Dubreuil, D.; Tellier, C.; Talham, D. R.; Bujoli, B. Journal of the Americ an Chemical Society 2004, 126 1497-1502.
116 (33) Benitez, I. O.; Bujoli, B.; Camus, L. J.; Lee, C. M.; Odobel, F.; Talham, D. R. Journal of the American Chemical Society 2002, 124 4363-4370. (34) Wu, A. P.; Talham, D. R. Langmuir 2000, 16 7449-7456. (35) Byrd, H.; Whipps, S. ; Pike, J. K.; Talham, D. R. Thin Solid Films 1994, 244 768-771. (36) Peterson, A. W.; H eaton, R. J.; Georgiadis, R. M. Nucleic Acids Research 2001, 29 5163-5168. (37) Shchepinov, M. S.; Ca se-Green, S. C.; Southern, E. M. Nucleic Acids Research 1997, 25 1155-1161. (38) Frydman, E.; Cohen, H.; Maoz, R.; Sagiv, J. Langmuir 1997, 13 5089-5106. (39) Weng, L. T.; Vereecke, G.; Genet, M. J.; Bertrand, P.; Stone, W. E. E. Surface and Interface Analysis 1993, 20 179-192. (40) Lee, H.; Kepley, L. J.; Hong, H. G.; Mallouk, T. E. Journal of the American Chemical Society 1988, 110 618-620. (41) Lee, H.; Kepley, L. J.; H ong, H. G.; Akhter, S.; Mallouk, T. E. Journal of Physical Chemistry 1988, 92 2597-2601. (42) Thompson, M. E. Chemistry of Materials 1994, 6 1168-1175. (43) Byrd, H.; Pike, J. K.; Talham, D. R. Chemistry of Materials 1993, 5 709-715. (44) Scofield, J. H. Journal of Electron Spectro scopy and Related Phenomena 1976, 8 129137. (45) The band-gap energy determined fr om the average UV absorption maxima of the nucleotides, which is 4.8 eV. (46) An ideal film density of 0.893 g/cm3 was used. (47) In the referenced pa per by Georgiadis and colleagues, th ey demonstrated that higher salt concentrations, which they used their experiment s, can lead to higher probe surface coverages. (48) Zammatteo, N.; Jeanmart, L.; Hamels, S.; Courtois, S.; Louette, P.; Hevesi, L.; Remacle, J. Analytical Biochemistry 2000, 280 143-150. (49) Halperin, A.; Buhot, A.; Zhulina, E. B. Langmuir 2006, 22 11290-11304. (50) Hong, B. J.; Oh, S. J.; Youn, T. O.; Kwon, S. H.; Park, J. W. Langmuir 2005, 21 42574261.
117 (51) Walker, S. L.; Bhattacharj ee, S.; Hoek, E. M. V.; Elimelech, M. Langmuir 2002, 18 2193-2198. (52) Guo, Z.; Guilfoyle, R. A.; Thiel, A. J.; Wang, R.; Smith, L. M. Nucleic Acids Research 1994, 22 5456-5465. (53) Pinnavaia, T. J.; Miles, H. T.; Becker, E. D. Journal of the American Chemical Society 1975, 97 7198-7200. (54) Blackburn, E. H. Nature 1991, 350 569-573. (55) Yauk, C.; Berndt, L.; Williams, A.; Douglas, G. R. Journal of Biochemical and Biophysical Methods 2005, 64 69. (56) Han, T.; Melvin, C. D.; Shi, L.; Branham, W. S.; Moland, C. L.; Pine, P. S.; Thompson, K. L.; Fuscoe, J. C. BMC Bioinformatics 2006, 7(Suppl 2) S17. (57) Wolf, L. K.; Gao, Y.; Georgiadis, R. M. Langmuir 2004, 20 3357-3361. (58) Giesbers, M.; J. Mi eke Kleijn; Stuart, M. A. C. Journal of Colloid and Interface Science 2002, 248 88. (59) Becker, C. F. W.; Wacker, R.; Bouschen W.; Seidel, R.; Kolaric, B.; Lang, P.; Schroeder, H.; MuIler, O.; Niemeyer, C. M.; Spengl er, B.; Goody, R. S.; Engelhard, M. Angewandte Chemie International Edition 2005, 44 7635. (60) Berger, M. F.; Philippakis, A. A.; Qureshi, A. M.; He, F. S.; Estep, P. W.; Bulyk, M. L. Nature Biotechnology 2006, 24 1429-1435. (61) Brockman, J. M.; Frutos, A. G.; Corn, R. M. Journal of the American Chemical Society 1999, 121 8044-8051. (62) Bulyk, M. L. Current Opinion in Biotechnology 2006, 17 422. (63) Bulyk, M. L.; Gentalen, E.; Lockhart, D. J.; Church, G. M. Nature Biotechnology 1999, 17 573-577. (64) OBrien, J. C.; St ickney, J. T.; Porter, M. D. Journal of the American Chemical Society 2000, 122 5004-5005. (65) Shumaker-Parry, J. S.; Zareie, M. H.; Aebersold, R.; Campbell, C. T. Analytical Chemistry 2004, 76 918-929. (66) Smith, E. A.; Erickson, M. G.; Ulijasz, A. T.; Weisblum, B.; Corn, R. M. Langmuir 2003, 19 1486-1492.
118 (67) Sato, N.; Ohta, N. Nucleic Acids Research 2001, 29 2244-2250. (68) Shoeman, R. L.; Hartig, R.; Traub, P. Biochemistry 1999, 38 16802-16809. (69) Tsung, K.; Brissette, R. E.; Inouye, M. The Journal of Biological Chemistry 1989, 264 10104 -10109. (70) Park, S.-M.; Lu, C.-D.; Abdelal, A. T. Journal of Bacteriology 1997, 179 5309. (71) Gong, P.; Harbers, G. M.; Grainger, D. W. Analytical Chemistry 2006, 78 2342-2351. (72) Meinkoth, J.; Wahl, G. Analytical Biochemistry 1984, 138 267-284. (73) Vainrub, A.; Pettitt, B. M. Biopolymers 2002, 68 265. (74) Howley, P. M.; Israel, M. A.; Law, M.-F.; Martin, M. A. The Journal of Biological Chemistry 1979, 254 4876-4883. (75) Zhu, H.; Bilgin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.; Lan, N.; Jansen, R.; Bidlingmaier, S.; Houfek, T.; Mitchell, T.; M iller, P.; Dean, R. A.; Mark.Gerstein; Snyder, M. Science 2001, 293 2101-2105. (76) Kambhampti, D. Protein Microarray Technology 2004, Wiley-VCH Verlag, Weinham. (77) Cha, T.; Guo, A.; Zhu, X.-Y. Proteomics 2005, 5 416-419. (78) Keller, T. A.; Duschl, C.; Kroger, D.; Anne-Francoise; Sevi n-Landais; Vogel, H.; Cervigni, S. E.; Dumy, P. Supramolecular Science 1995, 2 155-160. (79) Nieba, L.; Nieba-Axmann, S. E.; Persson, A.; Hamalainen, M.; Edebratt, F.; Hansson, A.; Lidholm, J.; Magnusson, K.; Karl sson, A. F.; Pluckthun, A. Analystical Biochemistry 1997, 252 217. (80) Santra, S.; Bagwe, R. P.; Dutta, D.; St anley, J. T.; Walter, G. A.; Tan, W.; Moudgil, B. M.; Mericle, R. A. Advanced Materials 2005, 17 2165-2169. (81) Brittain, H. G. Inorganic Chemistry 1978, 17 2762-2766. (82) Crosby, G. A.; Whan, R. E.; Freeman, J. J. Journal of Physical Chemistry 1962, 66 2493-2499. (83) Chrysochoos, J.; Evers, A. Chemistry Physics Letters 1973, 20 174. (84) Zhong, G.-L.; Wang, Y.-H.; Wang, C.-K.; Yang, K.-Z. Luminescence 2001, 385 234238.
119 (85) Zhong, G.; Yang, K. Langmuir 1998, 14 5502. (86) Lenaerts, P.; Storms, A.; Mullens, J.; DHaen, J.; Gorller-Walrand, C.; Binnemans, K.; Driesen, K. Chemistry of Materials 2005, 17 5194-5201. (87) Rodriguez-Mendez, M. L.; Aroca, R.; DeSaja, J. A. Chemistry of Materials 1992, 4 1017-1020. (88) Riegel, E. R.; Zwilgmeyer, F. Organic Syntheses 1937, 17 40. (89) King, H.; Ware, L. L. Journal of the Chemical Society 1939, 873. (90) Cooper, C. G. F.; MacDonald J. C.; Soto, E.; McGimpsey, W. G. Journal of the American Chemical Society 2004, 126 1032-1033. (91) Vogel. Vogel's Textbook of Prac tical Organic Chemistry 1978, Longman, London ; New York. (92) Borkowski, P. R.; Horn, J. S.; Rapoport, H. Journal of the American Chemical Society 1978, 100 276-281. (93) Roy, B. C.; Santos, M.; Sanku Mallik; Campiglia, A. D. Journal of Organic Chemistry 2002, 68 3999-4007. (94) Somfai, P.; Marchand, P.; Torsell, S.; Lindstrom, U. M. Tetrahedron 2003, 59 1293 1299. (95) Lin, W.; Zhang, X.; He, Z.; Jin, Y.; Gong, L.; Mi, A. Synthetic Communications 2002, 32 3279. (96) Cui, Y.; Chen, L.; Qian, G.; Wang, M. Dyes and Pigments 2006, 70 232-237.
120 BIOGRAPHICAL SKETCH Sarah Lane was born in Pensacola, FL the da y after Christmas way back in 1978. She graduated from Escambia High School, home of the fightin gators, in 1997. Subsequently, she began attending the University of Florida as a psychology major, but her major quickly switched to environmental science. Lack ing direction, she dropped out for a semester to wait tables at a diner in Pensacola. After whic h, she settled on a major in ch emistry. In December 2001, she graduated with a BS in chemistry and continued at the University of Florida with graduate studies in surface chemistry. She defended her PhD dissertation in July 2007.