<%BANNER%>

Using fluorescence anisotropy for sensitive platelet-derived growth factor detection based on molecular aptamers

University of Florida Institutional Repository

PAGE 1

USING FLUORESCENCE ANISOTROPY FOR SENSITIVE PLATELET-DERIVED GROWTH FACTOR DETECTION BASE D ON MOLECULAR APTAMERS By ZEHUI CAO A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2002

PAGE 2

Copyright 2002 by Zehui Cao

PAGE 3

iii ACKNOWLEDGMENTS I would like to thank my parents for wh at they have gone through to bring me up to where I am now, and my wife, Qian, for her support and patience all the time. I would also like to thank my advisor, Dr. Wei hong Tan, for his confidence in me and the encouragement he gave me to continue my study, Dr. Xiaohong Fang for her guidance in the early PDGF work, Dr. James Winefordne r for his kind help and understanding during my difficult times, the members of my advisory committee for their helpful guidance and suggestions, and the Tan research group.

PAGE 4

iv TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii ABSTRACT....................................................................................................................... vi CHAPTER 1 INTRODUCTION........................................................................................................1 Introduction to PDGF Protein.........................................................................................2 Aptamer as a Probe for Protein.......................................................................................3 Fluorescence Anisotropy................................................................................................5 2 MOLECULAR APTAMER FOR REAL-TIME PDGF DETECTION USING FLUORESCENCE ANISOTROPY...................................................................................8 Experimental Section......................................................................................................8 Materials.....................................................................................................................8 Instrumentation...........................................................................................................9 Anisotropy Measurements..........................................................................................9 Results and Discussions................................................................................................10 Design of an Fluorescent Aptamer Probe.................................................................10 Binding of the Aptamer to PDGF.............................................................................11 Real-Time Binding Detection Usin g Fluorescence Anisotropy...............................12 Effects of Mg2+ on the Binding Assay......................................................................14 Effect of Temperature on the Binding Assay...........................................................16 Detection of PDGF-BB in Homogeneous Solution..................................................17 Selectivity of the Aptamer Probe..............................................................................21 Conclusions...................................................................................................................2 3 3 DEVELOPMENT OF FLUORESCENCE ANISOTROPY IMAGING SYSTEM FOR PROTEIN ARRAYS................................................................................................25 Protein Arrays...............................................................................................................25 Fluorescence Anisotropy Imaging................................................................................28 Experimental Section....................................................................................................30 Materials...................................................................................................................30 Anisotropy Imaging Setup........................................................................................31 Acquisitions of Anisotropy Images..........................................................................32 Results and Discussions................................................................................................35

PAGE 5

v Experimental Considerations....................................................................................35 Anisotropy Imaging of TAMR A-Glycerol Solutions...............................................36 Anisotropy Imaging for Aptamer-Protein Binding Assay........................................38 Selectivity of Aptamer/PDGF Binding on the Anisotropy Imaging System............41 Detection of Protein Mixture....................................................................................43 Conclusions...................................................................................................................4 5 4 SUMMARY AND FUTURE WORK........................................................................47 Summary.......................................................................................................................4 7 Future Work..................................................................................................................47 Improving and Refining the Techniques...................................................................47 Applications of the Anisotropy Imaging System in Cancer Diagnosis....................48 Anisotropy Imaging Technique for Real-Time Cell Imaging...................................49 LIST OF REFERENCES..................................................................................................50 BIOGRAPHICAL SKETCH............................................................................................53

PAGE 6

vi Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science USING FLUORESCENCE ANISOTROPY FOR SENSITIVE PLATELET-DERIVED GROWTH FACTOR DETECTION BASE D ON MOLECULAR APTAMERS By Zehui Cao December 2002 Chair: Dr. Weihong Tan Major Department: Chemistry Proteins play very important roles in al most all functions of life. Detection of proteins has drawn great attention from scien tists for many years. However, analytical methods for real-time protein detection in homogeneous solutions are scarce. The recent development of molecular aptamers, comb ined with fluorescence techniques, may provide an easy and efficient approach to se nsitive protein analysis. Aptamers are small oligonucleotides which have high affinity and high selectivity to their target molecules. They are isolated by the systematic evol ution of ligands by exponential enrichment (SELEX) process. We have designed a fluores cein-labeled aptamer for the detection of platelet-derived growth factor (PDGF) pr otein which has the potential as a cancer indicator. Fluorescence anisotropy was used as the detection method. Fluorescent molecules, when excited by plane-polarized li ght, will give an emission that has been depolarized to a certain extent. This can be described in terms of anisotropy. Factors that can change the rotational diffusion of the fluorescent molecules will affect their

PAGE 7

vii anisotropy, including their molecular weight s and the properties of the surrounding environment. Fluorescence anisotropy techni que is ideal for aptamer/protein binding assays where there is a large molecular weight change. By combining aptamers and fluorescence anisotropy, we are able to ach ieve highly sensitive P DGF detection with high selectivity. The detection limit was about 0.22 nM. We then extended this technique into a two dimensional format by building a fluorescence anisotropy imaging system. This format enables the simultaneous anisotropy measurements of multiple samples. We demonstrated the ability of this system to detect multiple proteins in one sample. With some modifications and improvements, the anisotropy imaging tec hnique may have the potential to be a very eff ective and easily implemented a pproach to high throughput and multiplex protein analysis in an array format.

PAGE 8

1 CHAPTER 1 INTRODUCTION Proteins are macromolecules that consis t of one or more unbranched chains of amino acids. Although typical pr oteins contain 200-300 amino ac ids, they can be much smaller or much larger. Proteins are the build ing blocks of life. Almost every function in living cells depends on proteins. Those incl ude catalysis of all biochemical reactions, construction of cells, motion of cells and orga nisms, transportation of materials in body fluid and many more. Since proteins realize their functions through interactions with other molecules, it is highly important to unde rstand those interactions in order to find out how proteins work in liv ing cells. Many techniques have been developed to detect and analyze proteins. Some of them, such as electrophoresis and affinity chromatography, have provided a good way to separate and detect proteins. However, they lack the ability of monitoring protein interactions in real time and in homogeneous solutions. Fluorescence techniques, on the other hand, have shown great capability in detecting and studying protein functions in their native envi ronments. This ensures a more direct and precise understanding of protein interactions Many protein probes have been used for protein detection using fluorescence techniqu es. Some of which are extracted from animals that have the inherent ability of binding to certain proteins, such as antibodies. They have great affinity and selectivity towa rds proteins, but they have more restricted requirements for their surrounding environmen ts in order to f unction properly. Other probes have been synthesized or selected by scientists, such as aptamers. Those synthetic protein probe molecules can also bind to prot eins with high affinity and high selectivity,

PAGE 9

2 but they are more robust and much easier to obtain. In our work, we were trying to develop a fluorescence detecti on technique, particularly a fluorescence anisotropy technique, for proteins using aptamers as th e probe. We later extended and modified this technique to an array format as we tried to examine the possibility of using fluorescence anisotropy for simple and accurate high th roughput protein screening and multiplex protein detection. We specifically developed our technique for platelet-derived growth factor (PDGF) detection because of the great interest in this protein for cancer studies. Introduction to PDGF Protein Understanding disease-related proteins coul d be the very first and most important step in disease studies and drug discovery. As increasing attention has been focused on cancer diagnosis, it is of great interest to fi nd out more about those proteins that may be related to cancers. One of which is platelet -derived growth factor (PDGF) found in many human cell types. Its biological function is to stimulate the division and proliferation of the cells through binding of its receptors on cell membranes. It is also believed to play a role in intercellula r signaling [1]. PDGF has severa l isoforms, among which are PDGFAA, PDGF-BB and PDGF-AB. Those isoforms consist of the two subunits, PDGF-A chain and PDGF-B chain. The three isoforms act differently in specific situations. PDGF-BB has been shown to be actively involved in cell transf ormation process and in tumor growth and progression [2-6]. PDGF-BB, when bound to its receptor on the cell membrane, activates phosphatidylinositol 3-kinase (P I3-K) inside the cell, whic h eventually leads to cell growth. PDGF-BB is expressed at unde tectable low levels in normal cell [7] while it is often found over-expressed or mu tated in a malignant tumor [8]. Because of the potential

PAGE 10

3 as a cancer indicator, its dete ction has been attempted using the traditional antibody based radioisotropic methods and ELISA techniques [3-6]. Aptamer as a Probe for Protein For the last several decades, antibodies ha ve been the most important probes for a variety of molecular recogn ition applications. Many diagnos tic tests based on antibodies are routinely conducted in laborat ories and clinics. The great success in these applications is the result of their very high sensitivity a nd high selectivity to the target molecules. However, antibody technology also has some majo r disadvantages. It relies on the animal host to produce antibodies, which means high cost, low efficiency and the very limited number of target molecules that can be dete cted. Antibodies are very sensitive to their surrounding environments and often easily undergo irreversible denaturation. Regeneration of antibodies is usually not easy, which also contributes to the high cost.. In some cases, ligand density is very limited by the antibodies are large molecules. It is also relatively more difficult to directionally im mobilize antibodies onto a surface, which may be critical in some applications. Recent development of the systematic evolution of ligands by exponential enrichment (SELEX) process, however, may provi de solutions to some of the problems associated with antibodies. This process has the ability to isolate the oligonucleotide sequences that recognize virtua lly any class of target mol ecules with high sensitivity and high selectivity [9]. The resulting oli gonucleotide ligands were given the name “aptamers,” which comes from the Latin word “aptus,” meaning “to fit.” The SELEX process begins with a library of synthe sized oligonucleotides usually containing 1014 to 1015 random sequences. This library is then incuba ted with the target molecule of interest under certain conditions. The sequences that interact with and bind to the target

PAGE 11

4 molecules are isolated for next round of in cubation. This process is repeated until a sequence that binds to the targ et with highest selectivity and affinity is determined. SELEX enables the discovery of ligands to vi rtually any target molecules of interest. Some aptamers have been well studied, such as the aptamer that specifically interacts with human -thrombin. New aptamers are also being discovered for a variety of different target molecules [10]. Although most aptamers are exploited to study proteins, there are some applications that use aptamers for detecting smaller molecules such as cocaine [11]. Compared to antibodies, aptamers have si milar high affinity and selectivity for proteins [12,13]. What makes aptamers so usef ul is that they have several important advantages over antibodies. First, their pr oduction is easier, ch eaper and not limited by the animal hosts. Because oligonucleotides have more stable structures than proteins, aptamers can withstand harsher experiment al conditions than an tibodies and can be stored and reused without causing much degr adation. Aptamers can be easily labeled or modified in different ways for different mol ecular recognition applications. They can also be easily immobilized onto solid surfaces w ithout much change in the binding affinities to proteins. In one of the examples, the aptamer for human -thrombin was labeled with a fluorophore and then immobilized onto glass surface for high sensitivity thrombin detection [14]. In another example, the -thrombin aptamer was labeled with a fluorophore and a fluorescence quench er at its two ends to form an aptamer beacon [15]. The change in the aptamer’s conformation upon binding to -thrombin brought the fluorophore and quencher far away from each other, thus causing the restoration of fluorescence. Highly sensitive thrombin detection was also achieved based on this

PAGE 12

5 scheme. Unlike antibodies, aptamers can also be used to inhibit ta rget protein’s normal functions by occupying the activ e binding sites of the protein. This gives aptamers the potential to be drug candidates for numerous diseases. Traditiona l protein inhibitors either are too toxic or lack good specificity. Aptamers, on the other hand, are just small DNAs or RNAs with high affinity and high sele ctivity for proteins. Thus, they should be very safe and effective. Some work has been done to investigate pos sibility of aptamers inhibiting the activity of HIV virus [10,16]. High-affinity aptamers for PDGF-B chai n have been reported using SELEX [17]. Several single-stranded DNA sequences were found to bind to PDGF-AB and PDGF-BB with high affinity (Kd 10-10 M) while to PDGF-AA with lower affinity (Kd > 10-8 M). Most of the ligands were found to have a st ructure of a three-way helix junction with a three-nucleotide loop at the branch point. Fluorescence Anisotropy Fluorescence anisotropy has b een effectively used previ ously to study interactions between macromolucules. Upon excitation with polarized light, some samples will have polarized emission. The extent of polarization of the emissi on is described in terms of anisotropy ( r ) [18]. The theory behind the polari zed emission can be explained as following: when excited by a po larized light, the sample molecules that have absorption transition moments oriented along the electric vector of the incident light are preferentially excited. Those excited molecule s may rotate to other directions before returning from the excited state to the ground state and emitting light, thus causing a depolarized emission. This depo larization is dependent on th e extent of the rotational diffusion of the excited molecule s. Two factors are believed to affect the diffusive motion of a molecule. One is the viscosity of the solvent surrounding the mo lecule and the other

PAGE 13

6 is the size of the molecule itself. A sma ll fluorophore molecule in a solvent with low viscosity will be almost fully depolarized on the excited state and exhibit a non-polarized emission. The anisotropy in this case is close to zero. To measure the anisotropy of a sample, th e sample is excited with a vertically polarized light. The inte nsity of the vertically polarized component ( IVV) and horizontally polarized component ( IVH) of the emission are measured with a polarizer on the emission side. The anisotropy r is then calculated using the following equation: r = ( IVVIVH) / ( IVV+2 IVH) where the subscripts V and H refer to the or ientation (vertical or horizontal) of the polarizers for the intensity measurements, with the first subscript indicating the position of the excitation polarizer and th e second for the emission polarizer. Note that even though from this equation, th e range of anisotropy value is from 0 to 1, in a real-world sample, with the molecules evenly distributed in all directions, the maximal anisotropy one can observe with one-photon excitation is 0.4. Fluorescence anisotropy is a simple signaling method for binding assays. In a typical binding assay, a small dye-labeled probe molecule, such as an aptamer, binds to a large target molecule, such as a protein. Th e increased molecular size will slow down the rotational movement of the dye molecule li nked to the probe, t hus causing a more polarized emission. This polarization, or th e binding event, can be reflected by an increase in the measured anisotropy value. Traditional binding assays like ELISA require multiple steps and complex procedures. With fluorescence anisotropy, the mixing of the binding probe and the target is the only step needed. Anisotropy measurement can then be performed directly on the sample mixtur e and the result will clearly show whether

PAGE 14

7 there is binding between the probe and the target. Compared to techniques such as surface plasmon resonance (SPR) and total internal reflection fluorescence (TIRF), fluorescence anisotropy is insensitive to cha nges of the refractive index of the sample solution. Compared to other fluorescence base d techniques for biomolecular interaction study, fluorescence anisotropy requires only one fluorescent dye molecule on the probe. Because all that is needed to give a signal ch ange in anisotropy is simply the change in the size of the molecule linked to the dye, th ere is no need to worry about whether the conformational change of the probe after the binding can gi ve a signal, such as in molecular beacon techniques. Furthermore, because fluorescence anisotropy is a rationing technique, some problems associated with fluorescence intensity techniques, such as photobleaching, nonuniform illuminati on and unstable light source, are not of major concern. Applications of fluorescence anisotropy technique are not limited to protein detection. Fluorescence anisotropy has also been used to study biom olecular interactions in order to understand some biological pr ocesses [19]. Although in a lot of cases, fluorescence anisotropy is used to study protein-DNA interactions it is also suitable for protein-protein [20], D NA-DNA and other type of interact ions. Theoretically, as long as there is a molecular weight change after a bi nding process, there should be an anisotropy change for the fluorescent molecule. This is not limited to small fluorescent molecules binding to large ones. Some work was done using fluorescent protei n molecules to study protein-DNA interaction [21].

PAGE 15

8 CHAPTER 2 MOLECULAR APTAMER FOR REAL-TIME PDGF DETECTION USING FLUORESCENCE ANISOTROPY Despite the presence of current techniques for protein detection, there is still few nonisotropic and sensitive methods for real time protein analysis in homogeneous solutions. In this part of our work, we ha ve developed an easy and effective way to specifically analyze PDGF due to its pot ential significance in cancer research. Experimental Section Materials The fluorescein-labeled PDGF aptame r was customer-designed and then synthesized by Trilink Biotechno logies (San Diego, CA). The sequence of the aptamer is 5’-fluorescein-CAGGC TACGG CA CGT AGAGC ATCAC CATGA TCCTG. Recombinant human PDGF-BB, PDGF-AB, and PDGF-AA were pur chased from R&D Systems (Minneapolis, MN). They were disso lved in 4 mM HCl and then diluted in a Tris buffer before use. Other recombinant human growth factors, epidermal growth factor (EGF), and insulin-like growth factor 1 (IGF-1), were bought from Roche (Indianapolis, IN). Human bovine serum albumin (BSA), human hemoglobin (HEM), porcine lactic dehydrogenase (LDH), horse myoglobin (MYO), chicken lysozyme (LYS), and human thrombin (THR) and other chemicals were from Sigma (St. Louis, MO). The buffer we used consisted of 20 mM Tris-HCl (pH 7.1), 140 mM NaCl, 5 mM KCl, 1 mM CaCl2, and 1 mM MgCl2 to simulate the ionic strength under physiological conditions. Superpurified water was used to prepare all of the solutions.

PAGE 16

9 Instrumentation Fluorescence measurements were performed on a Fluorolog-Tau-3 spectrofluorometer (Jobin Yvon, Inc., Edison, NJ) equipped with a thermostat accurate to 0.1 C. All experiments were carried out at 37 C unless otherwise specified. The sample cell was a 100-L cuvette. The fluorescence in tensity of the aptamer was monitored by exciting the sample (fluorescein label) at 470 nm and measuring the emission at 520 nm. Slits for both the excitation and emission were set to 10 nm. Corrections were also made for potential effects on sample concentrations caused by dilutions in the titration experiments. To achieve a better analytical sensitivity of PDGF by the anisotropy measurement, the emission monochromator box in the spectrofluorometer was removed. A 515-nm long-pass filter (Oriel, Stratfor d, CT) and a 525-nm bandpass filter (Chroma, Brattleboro, VT) were put in front of the PMT to select the desired fluorescence signal from the polarizer. This increased the optical signal collection efficiency, as compared with the spectrometer’s original optical path. Anisotropy Measurements Fluorescence anisotropy was measured us ing the L-format c onfiguration (Figure 2-1). The anisotropy ( r ) is calculated according to the following equation: r = ( IVV-G IVH) / ( IVV+2G IVH) where the subscripts V and H refer to the orientation (vertical or horizontal) of the polarizers for the intensity measurements, with the first subscript indicating the position of the excitation polarizer and the second for the emission polarizer. G is the G-factor of the spectrofluorometer, which is calculated as G = IHV/ IHH. G-factor represents the ratio of the sensitivities of the detection system for vertically and horiz ontally polarized light.

PAGE 17

10 Polarizer Light source IV IVV Sample Detector PolarizerSample emission IVH Detector Polarizer IV Sample PolarizerSample emissionG-factor is dependent on the emission wa velength and can be determined using a standard sample with known anisotropy. Figure 2-1 Schematic diagram for L-format meas urements of fluorescence anisotropy In all of our experiments, each anisotropy data point was the average of six measurements with an integration time of 1 second. The relative standard deviation was <2% for all measurements. Results and Discussions Design of an Fluorescent Aptamer Probe High-affinity aptamers for PDGF-B chain have been developed to inhibit binding of PDGF-BB to its receptor [ 17]. The consensus secondary structure motif of the PDGF aptamers is a three-way helix junction w ith a conserved singlestranded loop at the branch point. The helix junction domain of the aptamer represents the core of the structural motif required for high-affinity binding. The binding of the aptamers has been studied by radiolabeling an aptamer using a nitrocellu lose filter-binding method. To construct a

PAGE 18

11 fluorescent probe for PDGF-BB based on the aptamers, we adopted a 35-base singlestranded DNA sequence similar to the reported 36-base aptamer [17]. The 5’ end of the sequence was covalently linked to a fluorescei n molecule as a fluorescence reporter. The structure of the aptamer is shown in Figure 2-2. Figure 2-2 Structure of the fluorescein-labeled aptamer probe. The fluorescein label is far away from the helix junction, which is the binding center for PDGF-B [17]. So it will not affect the binding affinity of the aptamer. Our gel electrophoresis result clearly showed the eff ective binding of the dye -labeled aptamer to PDGF-BB. Binding of the Aptamer to PDGF The binding between the ap tamer and the PDGF-BB molecule was confirmed by gel electrophoresis (Figure 23). The experiment was done on a 4-20% precast gradient polyacrylamide gel (Bio-Rad, Hercules, CA). A running buffer of 0.5 TB (Bio-Rad) was used. The gel was prerun at 60 V for 10 min, and then samples were introduced and run

PAGE 19

12 at 150 V for 70 min. Fluorescence images we re taken with a Kodak DC290 digital camera (Eastman Kodak, Rochester, NY) a nd a UV illuminator. On the gel image, the right lane contained the fluorescein-labeled aptamer and resulted in only one bright band. The left lane, which had dye-labeled apta mer and unlabeled PDGF-BB, showed two bands. The weaker band appeared at the sa me position as the bright band on the right lane, indicating the unbound aptamer. The brig hter band on the left lane moved much slower than the aptamer band, which reveals the binding complex of the aptamer to the protein. With the ratio of the aptamer:PDGF being 1:10, only a very small fraction of the aptamer was unbound. Figure 2-3 Gel electrophoresis of the apta mer and its binding complex with PDGF. The sample injected to the right lane is 0.1 M aptamer, and that in the left lane was the mixture of 0.1 M aptamer and 1.0 M PDGF. Real-Time Binding Detection Us ing Fluorescence Anisotropy The fluorescein-labeled aptamer is a relativ ely small molecule compared to PDGF-BB molecule (M.W.=25000). When the aptamer binds to the protein, the increase in overall size and molecular weight will greatly sl ow down the rotation of the fluorescein

PAGE 20

13 molecule. The slower rotational diffusion will result in a lower ability of the fluorescein molecule to depolarize the incident polarized excitation. Consequent ly, the anisotropy of the emission will also increase to reflect th e slower motion of the dye itself. Based on this, we believe the fluorescence anisotropy should be a simple and reliable method for studying the interaction between aptamer a nd PDGF molecules, and for detecting the protein as well. Our anisotropy measurement s howed a significant increase in anisotropy with the addition of PDGF-BB to the aptamer solution (Figure 2-4). Figure 2-4 Anisotropy change upon the bi nding of PDGF to the aptamer. The concentration of the aptamer is 0.1 M. A 1.0 M PDGF solution was added at time 0 s. The time resolution for the data collecting was 3.3 s. As shown in this figure, with the add ition of PDGF-BB, the anisotropy increased more than 2-fold. In a time period of about 3 seconds, anisotropy increase was observed and the anisotropy value remained stable afte r that, meaning that the binding between the aptamer and the PDGF was fast and stable. C ontrol experiments were conducted with just fluorescein dye solutions and PDGF under the same conditions as above. No anisotropy change was observed, confirming that the anis otropy increase in the aptamer experiment 0 0.05 0.1 0.15 0.2 -20-10010203040 t(s)Anisotropy

PAGE 21

14 was really due to the interaction between th e aptamer and the protein. All measurements were done in real-time; no pre-separation wa s needed, and no need to label the target protein. It should be mentione d that after the binding, the ove rall fluorescence intensity decreased about 30%. Since there was no intens ity decrease in the fluorescein dye and PDGF control experiment, we concluded that this intensity decrease was due to the formation of the binding complex, which brough t the dye molecule and protein close to each other. The oligo bases may also have an effect if they were closer to the dye after the binding. Effects of Mg2+ on the Binding Assay Metal ions in the solution often have a significant impact on the binding of a single-stranded DNA (ssDNA) to a protein. A series of experi ments designed to study the effects of Mg2+ ions on the binding be tween the aptamer and PDGF molecules were carried out. The concentration of Mg2+ was varied in a solution with only the fluoresceinlabeled aptamer and also in a solution with same amount of aptamer and excess PDGFBB molecules. The anisotropies of th e solutions were measured at each Mg2+ concentration and results are shown in Figure 2-5. The results showed that the anisotr opy of the fluorescein-labeled aptamer increased significantly as Mg2+ concentration increased. This is because the divalent metal ions stabilize the three-way helix struct ure of the free aptamer [17], thus hindering the rotational rate of the labele d fluorophore linked to the aptamer. In the case of aptamer/PDGF binding complex, the Mg2+ had a very small effect on the anisotropy until its concentr ation reached 2 mM. With higher Mg2+ concentrations, the anisotropy of the solution began to decrease, which might be the result of the effects of high ionic strength on the conformation of the protein.

PAGE 22

15 Figure 2-5 Effects of Mg2+ concentration on sample’s an isotropy. (A) 0.1 M aptamer in buffer; (B) 0.1 M aptamer and 1.0 M PDGF-BB in buffer. The combination of the two effects of Mg2+ is shown in Figure 2-6. It shows that at different Mg2+ concentration, the anisotropy ch ange between the aptamer and the aptamer/PDGF binding complex is different. Higher Mg2+ concentration results in a smaller anisotropy change. A la rger anisotropy change is desirable when detecting protein. However, in order to study the ab ility of our method to detect PDGF-BB under physiological conditions, we still chose to use a buffer with 1 mM Mg2+. A 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 024681012[Mg2+] (mM)Anisotropy B 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 024681012[Mg2+] (mM)AnisotropyA B CMg (mM) CMg (mM)

PAGE 23

16 Figure 2-6 Effect of MgCl2 on the aptamer/PDGF binding. The concentrations of the aptamer and PDGF were 0.1 M and 1.0 M respectively. Effect of Temperature on the Binding Assay Temperature also has its effect on a fluorescent molecule’s anisotropy. As temperature increases, the rotation ability of a molecule will increase and the viscosity of the surrounding solvent will de crease. Both factors contribu te to a lower fluorescence anisotropy. Since temperature will have an effe ct on the anisotropies of both aptamer and the aptamer/PDGF binding complex, we did e xperiments to find out what the overall effect would be on this binding process. Experiments were carried out in a similar manner as that in the study of Mg2+ effect. Anisotropies of both an aptamer solution and an aptamer/PDGF binding complex solution were measured at different temperatures, as shown in Figure 2-7. Even though higher temper ature decreased the an isotropies of both aptamer and the aptamer/PDGF binding comple x (data not shown), a gradual increase in the anisotropy difference betw een the two species was obser ved. However, at 40 C, 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 024681012[Mg2+] mMRelative Anisotropy Change CMg (mM)

PAGE 24

17 there was a significant decr ease in the anisotropy of the binding complex (data not shown), which contributed to the sh arp decrease at 40 C in Figure 2-7. Figure 2-7 Effect of temperat ure on the anisotropy change during the aptamer/PDGF-BB binding process. Solutions contained 0.1 M aptamer and 1.0 M PDGF-BB in buffer. This may indicate that at this temperature, the three-way helix junction structure of the aptamer, which is essential to the ap tamer/ PDGF binding, is much less stable. As a result, more aptamer molecules became unboun d, and the anisotropy decreased. In fact, 40 C is close to the reported melting temper ature of the aptamer (about 44 C), at which the aptamer changes from the folded structure to an unfolded one [17] This result also shows that fluorescence anisotr opy is not just a detection te chnique. It can also be a useful tool for studying and understanding th e conformational changes of molecules in some biological processes. Detection of PDGF-BB in Homogeneous Solution To test if the fluorescence anisotropy me thod is a practical wa y to quantitatively detect PDGF-BB in homogeneous solution, we carried out a series of titration experiments to construct a calibration curve. The experiments were performed by adding 0 0.02 0.04 0.06 0.08 0.1 152025303540 Temperature (C) Anisotropy change

PAGE 25

18 aliquots of PDGF-BB stock solu tion to an aptamer solution. Potential dilution effects on sample concentrations during the titrati on experiments were corrected. A control experiment was also conducted using exactly the same procedures and conditions as in the PDGF titration experiments, with the only difference being that we used the buffer alone to do the titration inst ead of the PDGF-BB stock solu tion. The results are shown in Figure 2-8. Figure 2-8 Titration of the aptamer with ( ) PDGF solutions in the concentration range of 0-1.0 M; ( ) blank buffer. The aptamer concentration is 0.1 M. The control experiment showed a stab le anisotropy value and no increase, meaning that volume increase or aptamer concentration decrease does not affect the anisotropy of the aptamer. With the addition of PDGF-BB, the anisotropy of the aptamer solution greatly increased, indicating the bind ing between the aptamer and the PDGF-BB. This increase, however, is not linear to the PDGF concentration. This is because the measured anisotropy of a sample at a certain wavelength is actually the average anisotropy of all the components in this samp le that have emission at this wavelength. 0.03 0.04 0.05 0.06 0.07 0.08 0.09 00.20.40.60.81[PDGF-BB] MAnisotropy CPDGF-BB (M)

PAGE 26

19 For example, for a solution containing a dye-labeled aptamer and its target protein, if the binding is a 1:1 ratio reaction, there will be only two species that may have fluorescence anisotropy: the aptamer itself and the aptamer/protein bindi ng complex. And the anisotropy of this solution ( r ) can be calculated usi ng the following equation: r = (1) raptamer + rcomplex (Equation 1) where raptamer and rcomplex are the anisotropies of pure aptamer and pure binding complex respectively, and is the fraction of the total aptamer that is bound to the protein to form the binding complex. It is clear that the anis otropy of such a system is related to the aptamer/protein concentration ratio and the constant of the bindi ng reaction. In a more complex system, more components need to be considered in the above equation. For PDGF-BB and its aptamer, the titration curve we obtained (Figure 2-8) indicated that the interaction between aptame r and PDGF-BB might not be monophasic, but rather biphasic, as there seems to be two plateaus in the curve. This is in agreement with the previous report that studied the aptamer/ PDGF binding using radioisotropic technique [17]. The reason for this biphasic binding is proposed to be the coexistence of two noninterconverting components of the aptamer that bind to the PDGF with different affinities. Despite the biphasic curve in a range of PDGF-BB concentration up to 1 M, at lower PDGF-BB concentration (0-100 nM) th ere is a good linear rela tionship between the anisotropy and the amount of PDGF-BB. Clinic al studies showed th at the concentration of PDGF is about 0.4-0.7 nM in huma n serum and 0.008-0.04 nM in human plasma [4-6]. However, the PDGF concentration in the local tumor area should be higher than that in the blood as the PDGF has not diffused into blood. To demonstrate that our method is

PAGE 27

20 capable of detecting PDGF in that concentration range, we made some modifications to our experiments. Before modification, our de tection limit was ~2 nM of PDGF-BB. We first removed the monochromator box on the em ission side of the spectrofluorometer and put a band-pass filter directly between the sample and the PMT detector. This resulted in a broader emission wavelength range, a shorter optical path, and a more efficient optical signal collection. The second thing we ch anged was using an initial aptamer concentration of 2 nM, instead of 0.1 M we used before. According to Equation 1 and the principles behind it, the anisotropy cha nge caused by the addition of PDGF is not related solely to the PDGF concentration. It also de pends on the initial aptamer concentration. The lower aptamer concentratio n, the less PDGF is needed to achieve a similar level of anisotropy change. Because of this reason, as long as our detection system is sensitive enough to give sufficient signal for 2 nM aptamer, we should be able to detect a PDGF-BB concentration much lowe r than 2 nM. We carried out a series of experiments with the modifications menti oned above, and obtained a linear curve in a lower PDGF-BB concentration range (Figure 2-9). The sensitivity of our method was greatly improved as indicated in the figure. The detection limit was calculated to be ~0.22 nM of PDGF-BB when 2 nM fluorescein-labele d aptamer was used in the binding assay. We believe this detection limit will be f easible for PDGF detection in blood serum samples and in local tumor fluid samples. But in order to be useful for PDGF detection in clinical blood plasma sample, our method stil l needs to be improved. Possible approaches could be optimization of system design and optical detection, using a dye with higher fluorescence intensity and using a better light source, such as a laser.

PAGE 28

21 Figure 2-9 Titration of the aptamer with PDGF solutions in the concentration range of 01.25 nM. The aptamer concentration is 0.2 nM. Selectivity of the Aptamer Probe It is essential for a protein probe to ha ve the ability to di stinguish between its target protein and other proteins in order to be of practical use. While the aptamer sequence has been reported to be highly select ive for PDGF by other t echniques, it is still important for us to test the selectivity of the probe using fluorescence anisotropy. We need to show that the fluores cent label and the anis otropy method do not affect in anyway the selectivity of the aptamer probe for PDGF We carried out a se ries of experiments where excess of several common extracellula r proteins, such as albumin, hemoglobin, myoglobin and lysozyme, etc., were added to aptamer solutions and the anisotropy changes were recorded. These changes were then compared to the anisotropy change caused by PDGF. The result is shown in Figure 2-10. It is clear that compared to PDGF, at similar concentrations, other proteins show ed no or little anisotr opy increase due to the aptamer’s specificity for PDGF. C

PAGE 29

22 Figure 2-10 Binding selectivity of the aptame r. Different extracellular proteins were compared with PDGF-B chain (PDGF) in th eir capability to change the aptamer’s anisotropy. The 5 fold proteins (moles) are added into 0.02 M aptamer solution at 25 C. Other experiments were done to compare th e affinities of the aptamer to PDGFBB and to other growth factors which may coex ist with PDGF-BB in clinical samples. As shown in Figure 2-11, the aptamer did not bind to epidermal growth factor (EGF), or insulin-like growth factor-I (IG F1), indicating that the aptamer is highly selective for PDGF-BB. The other two isoforms of PDGF, PDGF-AA and PDGF-AB, however, showed some anisotropy increase. But their a ffinities are much lower than that of PDGFBB. The PDGF A chain and B chain have 60% similarity in their amino acid sequence [22]. This may explain the anisotropy ch ange caused by PDGF-AA and PAGF-AB. The fact that the A chain is more acidic than B chain, however, may be the reason for its lower affinity to the aptamer. -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 PDGFBSALYSMYOHEMLDHTHRRelative Anisotropy Change

PAGE 30

23 Figure 2-11 Comparing the bindi ng capability of the aptamer to PDGF-BB and to other growth factor such as PDGF-AA, PDGF-A B, epidermal growth factor (EGF), and insulin-like growth factor-I (IGF1 ). The molar ratio of the pr otein to the aptamer is 1:1. Conclusions In this part of our work, we have de veloped a molecular aptamer for real-time detection of the oncoprotein PDGF in homogeneous solution using a fluorescence anisotropy method. A fluorescein-labeled PDGF aptamer was designed and used as a probe to observe its an isotropy increase upon binding to it s target protein. The significant increase in anisotropy was attributed to the large difference in molecular size between the free fluorescein-labeled aptamer and its PDGF binding complex. This difference is significant enough to allow the mo lecular binding to be quantified for protein detection in real-time. The assay is highly selective a nd can detect PDGF down to 0.22 nM. The assay is quick and can detect PDGF without sepa ration. Anisotropy measur ements are ideally suited for measuring the binding of small ap tamer probes with protein macromolecules. This work demonstrates the potential ap plications of dye-labeled aptamers for oncoprotein and disease-related protein detectio n in clinical studies. This assay can be used in a noncompetitive homogeneous assay format. The same assay concept can also be -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 PDGF-BBPDGF-AAPDGF-ABEGFIGF1Relative Anisotropy Change

PAGE 31

24 used for biosensors by immobilization of the ap tamer onto a solid surface for in vivo or in vitro protein monitoring [23]. R ecently, the in vivo instability of aptamers in a biological fluid containing nucleases has been circumvent ed by chemical modification of the bases, particularly by substitutions at the 2’ position of the sugar [24], allowing aptamers to function adequately in biologica l fluids. This will ensure th e application of aptamer-based analytical methods in re al biological samples.

PAGE 32

25 CHAPTER 3 DEVELOPMENT OF FLUORESCENCE ANISOTROPY IMAGING SYSTEM FOR PROTEIN ARRAYS As databases for many sequenced genomes have been built, people begin to realize it is time to move to the next leve l: to understand more about proteins and their functions in life. Whereas gene s contain the information for li fe, the encoded proteins and RNAs fulfill nearly all the functions, from rep lication to regulation. It is well recognized that the complexity of the human proteome far exceeds that of the genome. The number of different molecular protei n species in the human body is likely to be at least 500,000. To be able to deal with such a large amount of proteins, it is very important to have new techniques that can realize protein analysis with high throughput. Not long ago, the term “proteomics” was proposed to define the larg e-scale study of the proteins expressed by a genome. Although microarray technology has enab led rapid development of genomics, it is not as easy, if not much mo re difficult, to apply similar techniques to proteomics. In this part of our work, we try to de velop a novel technique based on anisotropy measurements using imaging technique, which has the potential to realize high throughput and multiplex protein analysis. Protein Arrays Protein analysis has been done using a vari ety of techniques in small scale. They include techniques based on probe-prote in binding, such as enzyme-linked immunosorbent assay (ELISA), and techniques based on protein separations, such as gel electrophoresis. Those techniques have showed good sensitivity and selectivity in protein

PAGE 33

26 analysis, but in order to do large-scale protein studies, they still need to be modified or improved. A recent approach is the combinati on of 2-dimentional gel electrophoresis and mass spectrometry. This enables the analysis of multiple proteins in a short period of time. However, it also has some major drawba cks. First, it is a destructive technique, meaning you will lose some of your protein samp les. Second, it may not be able to collect and analyze all the protein sp ecies if they are expressed at low abundance. It can not provide reliable quantitative results. Lastly, while it is effective in separating and isolating individual proteins, it does not yiel d much information about the interactions between proteins and other biomolecules in real-time. Protein arrays, on the other hand, have been rapidly developed to address some of the problems faced by other techniques. Prot ein arrays are solid-phase ligand binding assay systems using immobilized proteins on surfaces which include glass, membranes, microtiter wells, mass spectrometer plates, and beads or other particles. The assays are highly parallel and often miniaturized. Th eir advantages include being rapid and automatable, capable of high sensitivity, ec onomical on reagents, and yielding a lot of data for a single experiment. Protein array is expected to be a very important technology in proteomics thanks to its ability to make possible high th roughput protein detections as well as parallel multiplex screening of intera ctions between multiple proteins and other biomolecules. Some work has demonstrated extraordinary power of protein chips to analyze thousands of proteins at the same ti me [25]. The capture ligands used in protein arrays are often antibodies, but may also be proteins, enzyme-substrates, receptor-ligands and aptamers [26]. Compared to antibodies, ap tamers have the advantages of ease of

PAGE 34

27 production by automated oligonucleotide synthesi s, robust nature of the nucleic acids, and easy modification for expanded applications. Even though it has a great potential and so me successful applications, the protein array technique is still facing some challeng es. Traditional protein arrays, or protein chips, when compared to DNA arrays, have some major drawbacks due to protein’s inherent properties. While DNA molecules are stable even in some harsh experimental conditions, proteins are much more fragile and very sensitive to their surrounding environment and will be easily damaged if not treated carefully. DNA molecules can be readily immobilized onto several kinds of surfaces without much loss of biological functions; proteins are much easier to be denatured when close to a solid surface, which will result in the loss of their abilities to r eact with their ligands. In traditional protein arrays where the detection system is based on fluorescent intensity or other intensity measurements, the immobilization of the capture ligand is inevitable. People usually need to immobilize the ligands on a chip, and then incubate the chip with the analytes. After the target protein in the sample binds to the ligands on the chip surface, all the unbound molecules need to be washed away. The protein on the chip will then be stained with a stain reagent. Finally the stain generates in tensity signals on the detection system, which indicates the presence of the target protein. All these procedures ar e the result of the detection methods adopted by the array system In our work, a new detection scheme has been developed for protein arrays base d on fluorescence anisotropy which does not require the ligand immobilization an d is very simple to operate. This will also allow us to circumvent the problems associated with denatu ration of proteins, either the ligand or the target, when immobilized to the solid surface.

PAGE 35

28 Fluorescence Anisotropy Imaging As described in Chapter 1, fluorescence an isotropy is a unique technique that has many distinctive features. Unlike intensity based fluorescence techniques, the excitation in fluorescence anisotropy is a plane-polar ized light. Depending on the size of the fluorescent molecules in the sample or the viscosity of the sample solution, the emission will have different intensities at two polarizat ion planes, one parallel to the excitation and the other perpendicular to it. The difference is measured and calculated using following equation: r = ( IVVIVH) / ( IVV+2 IVH) where the subscripts V and H refer to the or ientation (vertical or horizontal) of the polarizers for the intensity measurements, with the first subscript indicating the position of the excitation polarizer and the second for the emission polarizer. It is important to know that in a real-world measurement, the detection system may have different sensitivity to polarized light at different directions. So the actual equation used for anisotropy calculation is: r = ( IVV-G IVH) / ( IVV+2G IVH) where G is the G-factor of the detecti on system, which is calculated as G = IHV/ IHH. Gfactor represents the ratio of the sensitivitie s of the detection system for vertically and horizontally polarized light. Gfactor is dependent on the em ission wavelength and can be determined using a standard sample with known anisotropy. For a fluorophore-labeled probe, after it binds to its target prot ein, the increase of the overall weight of the binding complex w ill greatly decrease the fluorophore’s ability to rotate and result in a higher fluorescence anisotropy value. Intensity-based techniques can not tell the difference after the binding because only the total emission of the

PAGE 36

29 fluorophore is measure. However, an anisot ropy system can easily tell if there is a binding process even if the total emi ssions of the bound and unbound fluorophores are about the same. Most of th e current protein arrays require a long process of immobilization, incubation, washing and staining, in order to see the binding between the ligand and protein. By using fluorescence an isotropy, on the other hand, we can detect the binding event as soon as the ligand and the protein are mixed together. This method also has the potential to detect not only th e DNA-protein interacti on but also proteinprotein and RNA-protein interactions. Fluorescence anisotropy measurements ar e mostly done in fluorometers, where the type of the sample container is fixed and simultaneous multiple sample measurements can not be easily implemented. The idea presente d here is that if we could use a plate with wells on it that have different samples in them and then construct an anisotropy image of the plate, we should be able to look at samples with different anisotropies within one image. It has been reported that an anis otropy imaging method ha s been developed to study single molecules [27]. However, it was done with a very limited scale of samples, and should not be suitable for applications like protein arra ys. Other work has been done to study protein interactions and detect proteins with high throughput using fluorescence polarization (FP) [28]. It was based on a si milar method to fluorescence anisotropy and utilized a plate reader to measure samples one by one, instead of using one-time imaging system to measure all samples at the same time. In this part of our work, we have developed a system that can ta ke anisotropy images of fluor escent samples and make it a potentially better detection method for protein arrays.

PAGE 37

30 In order to obtain anisotropy images, two polarizers need to be coupled to an imaging system. One is the excitation polarizer that provides a polarized light source. Another is a polarizer on the emission side. While the excitation polarizer should be fixed in position, the emission polarizer need to pl aced in a way so that we can change its position to get two polarization planes, one para llel to the polarization plane of excitation light and the other perpendicula r to it. By changing the positi on of the emission polarizer, two images can be obtained, on which each pixe l does not show the real emission of the sample at that pos ition, but rather IVV or IVH of the total emission. An image processing software will be used to calculate the two images at pixel level to get an anisotropy image. The value of each pixel on this anis otropy image represents the anisotropy of the sample at that spot. Experimental Section Materials The carboxytetramethylrhodamine (T AMRA)-labeled PDGF aptamer was customer-designed and then synthesized by Integrated DNA Technologies, Inc. (Coralville, IA). The sequence of the aptamer is 5’CAGGC TACGG CACGT AGAGC ATCAC CATGA TCCTG -3’TAMRA. A TMARA-labeled human -thrombin aptamer, with a sequence of 5’TTTGG TTGGT GT GGT TGGT -3’TAMRA, and another TAMRA-labeled oligonucleotide with a random sequence of 5’CGGTA GTACC AAGTC CAGGT -3’TAMRA as a control DNA were also synthesized by Integrated DNA Technologies, Inc. Recombinant human PDGF-BB, PDGF-AB, and PDGF-AA were purchased from R&D Systems (Minneap olis, MN) and they were dissolved in 4 mM HCl and then diluted in a Tris buffe r before use. Human bovine serum albumin (BSA) and epidermal growth factor (EGF) were bought from Roche (Indianapolis, IN).

PAGE 38

31 Other chemicals were from Sigma (St. Loui s, MO). Buffer used in all experiments consisted of 10 mM Tris-HCl (pH 7.5), 75 mM NaCl, and 2.5 mM KCl. Superpurified water was used to prepare all of the solu tions. All experiments were done at room temperature. Anisotropy Imaging Setup A schematic diagram of the anisotropy im aging system is shown in Figure 3-1. An intensified charge-coupled device (I CCD) (Roper Scientific, EEV 512 1024 FT) was used to capture all images. A home-mad e cylindrical polarizer holder was placed under the ICCD. There was a 90 slit on the outer surface of the holder around the axis of the cylinder through which we coul d rotate a polarizer fixed in side the holder Rotation of the polarizer from on end of the slit to th e other could change between two emission polarization planes perpendicula r to each other. A TV zoom lens from Edmund Industrial Optics (Barrington, NJ) was also placed beneath the ICCD to help lo ok at a relatively large area. A mercury lamp by Olympus Ameri ca Inc. (Melville, NY) was used as the light source. The light coming out of the lamp was guided by a fiber bundle to the sample plate. A polarizer was placed just before the outlet of the fiber bundle as the excitation polarizer. All polarizers we re from Edmund Industrial Optics. A 520-550 nm bandpass filter from Olympus America Inc. was used for excitation. A 590 nm longpass filter (from Olympus America Inc.) and a 600 nm shortpass filter from Oriel Instruments (Stratford, CT) were used for emission. The sample plate was made of black plastic (DELRIN acetal resin) with dimensions of 20 mm20 mm2 mm. A 66 arra y of small wells was mechanically made on the plate. Each well is 1.5 mm in diam eter and 0.75 mm deep, and it holds about 1.5 L of liquid sample.

PAGE 39

32 A Fluorolog-Tau-3 spectrofluorometer (Jobin Yvon, Inc., Edison, NJ) was also used to perform some fluorescence anisotr opy measurements. In those cases, a 50 L quartz cuvette was used to hold the sample solution. Excitation and emission wavelength were set to be 545 nm and 595 nm respec tively for TAMRA. Bandpasses for both the excitation and emission were set to 10 nm. Figure 3-1 Schematic diagram of the anisotr opy imaging system. An ICCD with a zoom lens is the detector. A fibe r bundle coupled to a mercury lamp provides the excitation light. Acquisitions of Anisotropy Images Alignment of polarizers is done as follo wing. A sample of very high anisotropy, such as a TAMRA dye solution with 80% (V/V ) glycerol is used. First, the emission polarizer is fixed at one end of the 90 slit on the polarizer holder. Then the position of the excitation polarizer is adjusted until the lowest fluorescent signal is observed. The current position of the emission polarization should be perpendicular to that of the ICCD Emission p olarizer holde r TV zoom lens Sam p le p late Mercur y lamp Fiber bundle Filters and p olarizer

PAGE 40

33 excitation polarization, which will only allow the IVH signal of the sample to reach the detector. With rotation of the emission polarizer for 90, the IVV component of the emission will then be collected by the dete ctor. Same approach was also reported by other group [29]. By using ICCD as the detect or, two images can be obtained, with one containing the IVV part of the emission and the other containing the IVH part. We may call them Image IVV and Image IVH, or IIVV and IIVH respectively. 1.5 L of sample solution is added to the well on the sample plate. With the excitation on, we obtain two images of the sample, IIVV and IIVH, by rotating the emission polarizer. The exposure time is the same for the two polarization images of the same sample, but may vary from 0.05 sec to 3 sec from sample to sample depending on the sample intensity. The image capturing so ftware is Winview (Roper Scientific). Data processing is done us ing a Java-based computer program called ImageJ ( http://rsb.info.nih.gov/ij/index.html ). It was developed by the Research Services Branch at National Institute of Mental Health and National Institute of Neurological Disorders and Stroke. It has many image processing f unctions and allows you to develop plugins for your own image processing needs. We developed two plugins for the anisotropy imaging experiments. One plugin is used to clear the background area of the image IIVV and IIVH to zero. This is needed because the background noise on the two images is low and highly fluctuating. If it were not set to zero, on the final anisotropy image, the background would show very random anisotropies from very low to very high. The real anisotropy signal from the wells would be bur ied in this background. Another plugin we developed is used to directly calculate an anisotropy image Ir with image IIVV and IIVH and a user-specified G-factor using the following equation:

PAGE 41

34 Ir = ( IIVV-G IIVH) / ( IIVV+2G IIVH) In the anisotropy image, the value of each pixel represents the fluorescence anisotropy of the sample at that position. To get quantitative result s from the anisotropy image, the value of the pixels within the ar ea of one sample is averaged and the average value is used to represent the actual anisot ropy of the sample. The final anisotropy image might be rescaled or cut off of some trivial areas that are not related to the real signal for the sake of better presentation of data. In order to get reliable anisotropy values from our anisotropy imaging system, it is important to know the G-factor of the system. However, it is not easy to measure the Gfactor directly using the anisotropy imaging system. We solved this problem by correlating the anisotropy values measured by the imaging system to the anisotropies we get from the same sample using a spec trofluorometer that has the anisotropy measurement capability. This spectrofluorometer is able to determine the G-factor by measuring IHV and IHH, and thus can get relatively accu rate anisotropy values for a fluorescent sample. We set the excitation and emission wavelengths on the spectrofluorometer to be similar to those on the anisotropy imaging system so that the two systems should yield similar anisotropy va lues for the same sample. We then used the spectrofluorometer to determine the fluorescence anisotropies of the free proteinbinding ligand solutions. These ligand solutions with known anisotropies were used as references in all experiments conducted on the anisotropy imaging system. G-factors of the system were chosen to make sure th e calculated anisotropies for the reference solutions were the same as what we obtained from the spectrofluorometer. In this way,

PAGE 42

35 the results obtained from the anisotropy imag ing system should be parallel to those from the spectrofluorometer. Results and Discussions Experimental Considerations The carboxytetramethylrhodamine (TAM RA) dye was chosen to be the fluorescent label for all the ligands for differe nt proteins. TAMRA is much more stable than some of the other dyes such as fluorescein, which means we will have less severe photobleaching problem. This enables us to us e relatively longer expos ure time to obtain higher signals. Another advantage of TAMRA is that it can be readily excited by the intense 546 nm spectral line from mercury-ar c microscope lamps [30] and often shows brighter intensity than fluorescein. We use two TAMRA-labeled aptamers as the ligands for two proteins. One of them is the PDGF-aptamer that binds select ively to PDGF-BB protei n. It has a 35 base sequence and has a three-way he lix junction structure as sh own in Figure 2-2. Another aptamer is the aptamer for human -thrombin. It has a 15 base sequence and is believed to fold into a chair-form quadruplex with th e 5’ and 3’ ends in the corners of the quadruplex and two stacked G-quartets linked by TT and TGT loops [31,32]. The aptamer was found to bind selectively to human -thrombin with high affinity [33]. Aptamers are ideal for protein binding assay beca use of its high selectivity and affinity to their target protein. Compared to antibodies, aptamers have the advantages of low cost, easy handling and stable struct ure. We here use aptamers as the binding ligands for proteins and use fluorescence anisotropy im aging as the detection method since the binding of the aptamer to target protein result s in a larger molecular weight and greater anisotropy.

PAGE 43

36 Anisotropy Imaging of TAMRA-Glycerol Solutions To test the ability of our system to detect anisotropies of multiple samples simultaneously, we used TAMRA dye solutions with different glycerol concentrations (V/V). Viscosity of glycerol is much highe r than that of pure water. With glycerol concentration increases, the viscosity of the solutions also increases. As the rotational diffusion of the TAMRA dye molecules is re stricted by the higher viscosity of the solution, the measured anisotropies of the dye solutions should show an increase. Figure 3-2 shows an anisotropy image of eight TA MRA solutions in eight wells on the sample plate with glycerol concentra tion ranging from 10% to 80%. We can see clearly that the difference in anisotropy between tw o solutions can be seen directly from the image. By compar ing the colors of two samples on the same anisotropy image, we can tell the anisotropy difference and the information associated with the difference, such as differences in molecular size and solu tion viscosity. In the case of TAMRA-glycerol solutions, the anis otropy difference is caused by the difference in the viscosity of the solutions. On the anisotropy image, pi xels within one sample showed a distribution of anisotropy values. It is most likely due to signal fluctuation of each small photosensor on the ICCD during the data acqui sition process. Howe ver, this distribution is not totally random but rather centered around a certain an isotropy value. To get the quantitative results from the anisotropy image, we averaged the anisotropy values of each pixel within one sample to get the overall anisotropy of the sample. The re sults are shown in Figure 33.

PAGE 44

37 Figure 3-2 Anisotropy image of 200 nM TAMRA solutions with different glycerol concentrations. The table shows concentra tions of glycerol (V/V) and the relative positions of the eight solutions. Exposure time was 0.05 sec. Figure 3-3 Anisotropy vs. glyc erol concentration (V/V) in 200 nM TAMRA solutions. Data was averaged from four anisotropy images. The anisotropy values obtained from the anisotropy imaging system showed quite small error bars, partially due to the hi gh fluorescence of the TAMRA solutions. The anisotropy increases as the glycerol concentr ation increases. Similar curves were also obtained as a confirmation from the spectr ofluorometer, which wa s supposed to give rather accurate anisotropy measurements. It is worth noting that from the quantitative 40% 50% 30% 60% 20% 70% 10% 80% 0.00 0.30 0.00 0.05 0.10 0.15 0.20 0.25 020406080100 [Glycerol] (V/V)%Anisotropy Cglycerol (V/V)%

PAGE 45

38 curve and the anisotropy image, an anisotropy difference of as little as 0.015 can be seen with ease from the anisotropy image by just comparing colors. The imaging system provides a qualitatively very straightforwar d and quantitatively rather accurate way to examine multiple samples simultaneously. Anisotropy Imaging for Aptamer-Protein Binding Assay In our previously work, we used the sp ectrofluorometer to do all the anisotropy measurements for PDGF and aptamer binding study. Sample solutions of at least 50 L were used in order to fit in a cuvette or to get enough signals. With the anisotropy imaging system’s ability to conduct multiple anisotropy meas urements with very small amount of samples, we should be able to pl ace more than one solution with different aptamer/PDGF ratios in the plate wells and look at the extents of the binding reaction reflected by the anisotropy values in all wells at the same time. This will be quite useful and convenient when it comes to studyi ng protein binding process under different conditions. We carried out some experiments where the anisotropies of a series of solutions with the same aptamer concentrati on but different PDGF concentrations were determined at one time using anisotropy imaging. An anisotropy image is shown in Figure 3-4. A binding curve was also constructed based on the image, as shown in Figure 3-5.

PAGE 46

39 Figure 3-4 Anisotropy image of eight 50 nM PDGF-aptamer solutions with different PDGF-BB concentrations. The table on the right shows the concentration of the sample in nM at the corresponding position on the anisotropy image. Figure 3-5 Anisotropy vs. PDGF-BB concen tration in 50 nM TAMRA-labeled PDGFaptamer solutions. Data was averaged from 5 anisotropy images. Exposure time was 3 sec. Compare the shape of the curve in Figure 3-5 to the curve we obtained previously using the spectrofluorometer (Figure 2-8), we can see that in the PDGF-BB concentration 0.00 0.22 100 70 150 50 200 20 300 0 0.10 0.12 0.14 0.16 0.18 0.20 050100150200250300350[PDGF-BB] nMAnisotropy CPDGF-BB (nM)

PAGE 47

40 range of from 0 to 6 CPDGF-aptamer (0 to 0.6 M in Figure 2-8 and 0 to 300 nM in Figure 3-5), these two curves are very much alike. We did not go to higher PDGF concentration here since we just wanted to demonstrate the capability of the anisotropy imaging system in protein binding assay. This curve was al so repeatable on the spectrofluorometer. Compared to the anisotropy image we got from the TAMRA-glycerol solutions, the anisotropy image for aptamer/PDGF binding solutions showed less distinctive colors between samples. This is probably because at a much lower fluorescence intensity level and with a much longer exposure time, the sign al fluctuation of the pixels on the ICCD became more severe. Photobleaching of the TA MRA dye at low concentration might also have played a role in this problem. The result of this problem is a much larger anisotropy distribution within one sample on the anisotropy image. This is very much like broadened peaks in chromatography and there is colo r overlapping between two adjacent samples on the anisotropy image. This is why we see si milar colors between samples. However, the color difference is still discernable and the aver aged anisotropy values of the samples still showed significant change as PDGF-BB concen tration increased as shown in Figure 3-5. Similar reasons can be used to explain the larg er error bars in Figure 3-5 than in Figure 33. In a system where we use multiple sample wells and compare signals from different wells for analyte de tection, the well-to-w ell signal varia tion should be the indicator of the noise of this assay system [28]. The well-to -well anisotropy variation of our system was tested to be <0.002 by measur ing the anisotropy of the same sample in multiple wells. To qualify for a real signal, the anisotropy should be 3 times of this deviation, which accounts for an anisotropy difference of 0.006. Applying this to the

PAGE 48

41 curve in Figure 3-5, we determined the detection limit of our anisotropy imaging system for PDGF-BB to be around 13 nM. Since we are detecting a sample of only 1.5 L in each well, the mole detection limit should be 13 nM 1.5 L 2.010-14 mole =20 fmole of PDGF-BB. This sensitivity can still be im proved by optimization of the optical path, changing to a near IR dye w ith higher fluorescence and less photobleaching, and using high power laser as the light source. Selectivity of Aptamer/PDGF Binding on the Anisotropy Imaging System We carried out a series of experiments to demonstrate that the selectivity of the aptamer/PDGF-BB binding can be preserved on the anisotropy system. First, 50 nM PDGF-aptamer solutions with 300 nM of differ ent proteins were added to six wells. The resulting anisotropy image is shown in Figure 3-6. Figure 3-6 Anisotropy image of 50 nM PDGF-aptamer soluti ons with 300 nM different proteins [epidermal growth factor (EGF ), bovine serum albumin (BSA), PDGF-AA, PDGF-BB and PDGF-AB]. The table on the right shows the relative positions of these aptamer/proteins solutions on the anisotropy image and their anisotropy values based on the image. The relative anisotropy changes over pure aptamer solution caused by other proteins compared to PDGF-BB are shown in Figure 3-7. We can see that on the PDGF-BB 0.187 BSA 0.099 PDGF-AA 0.128 EGF 0.109 PDGF-AB 0.139 No protein 0.105 0.00 0.22

PAGE 49

42 anisotropy imaging system, the PDGF-aptamer still showed very good selectivity for PDGF. At similar concentrations as PDGF-BB, EGF and BSA showed very little or no anisotropy increase while PDGF-AA and PD GF-AB showed some increase which is agreeable with previous result in Figure 2-11. Figure 3-7 Comparing the bindi ng capability of the PDGF-a ptamer to PDGF-BB and to other proteins such as PDGF-AA, PDGF-AB, epidermal growth factor (EGF), and bovine serum albumin (BSA). The aptamer and protei n concentrations are 50 nM and 300 nM respectively. Another experiment was designed to s how that only PDGF-aptamer binds to PDGF-BB while other dye-labeled oligonucleotides have little affinity to PDGF-BB. An anisotropy image is shown in Figure 3-8. We used a TAMRA-labeled control DNA with a random sequence and the TAMRA-labeled human -thrombin aptamer as ligands to bind to the same concentration of PDGFBB. As shown clearl y in Figure 3-8, no anisotropy increase was observed fo r control DNA/PDGF-BB or thrombinaptamer/PDGF-BB. Only PDGF-aptamer showed significant anisotropy increase with the addition of PDGF-BB. This shows that apta mers are ideal binding ligands for protein -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 PDGF-BBPDGF-AAPDGF-ABEGFBSARelative Anisotropy Change

PAGE 50

43 binding assay. By combining aptamers with fluorescence anisotropy techniques, we should be able to provide an easy and effective way for proteomics studies. Figure 3-8 Anisotropy image for comparison of binding ability of PDGF-BB to different TAMRA-labeled oligonucleotides [a c ontrol DNA sequence (ctrDNA), the human thrombin aptamer (tAPT) and PDGF-aptamer (pAPT)]. The first column is 50 nM pure oligonucleotide solutions. The second column is 50 nM oligonucleotides plus 50 nM PDGF-BB. The table on the ri ght shows the sample s’ relative locations on the anisotropy image and their anisotropy values. Detection of Protein Mixture In protein array or protein chip techniques, it is important to be able to not only detect proteins with high th roughput but also detect multi ple proteins in one sample simultaneously. In order to demonstrate that the anisotropy imaging system can easily realize multiplex protein detection, we conducte d a simplified example of the detection of a protein mixture. The anisotropy image of this experiment is shown in Figure 3-9. 0.00 0.15 pAPT+PDGF-BB 0.122 pAPT 0.099 tAPT+PDGF-BB 0.090 tAPT 0.088 ctrDNA+PDGF-BB 0.096 ctrDNA 0.098

PAGE 51

44 Figure 3-9 Anisotropy image of 50 nM different TAMRA-la beled oligonucleotides [a control DNA sequence (ctrDNA), the human -thrombin aptamer (tAPT) and PDGFaptamer (pAPT)] with different protei ns or protein mixture (50 nM human -thrombin, 100 nM PDGF-BB and the mixture of these two) The table shows all samples’ relative locations on the anisotropy imag e and their anisotropy values. In this experiment, human -thrombin, PDGF-BB and th eir mixture were added separately to three TAMRAlabeled oligonucleotide sequen ces (a control DNA sequence, the thrombin-aptamer and the PDGF-aptamer) and resulting anisotropy image was recorded. It is clear from Figure 3-9 that the control DNA showed little anisotropy tAPT+mixture 0.081 tAPT+PDGF-BB 0.052 tAPT+ thrombin 0.088 tAPT 0.057 pAPT+mixture 0.129 ctrDNA+mixture 0.092 pAPT+PDGF-BB 0.142 ctrDNA+PDGFBB 0101 pAPT+ thrombin 0.107 ctrDNA+thrombin 0.090 pAPT 0.109 ctrDNA 0.096 0.00 0.17

PAGE 52

45 increase when mixed with any of the protei ns, indicating there was minimal interaction between the control DNA and the proteins. The thrombin-aptamer showed anisotropy increase with addition of thrombin or thro mbin/PDGF-BB mixture, but no increase with addition of PDGF alone, meaning that it was -thrombin that caused the anisotropy increase. Similar results were observed fo r PDGF-aptamer, which showed anisotropy increase only with PDGF-BB and thromb in/PDGF-BB mixture. This experiment demonstrated that such an array of aptamers can be used to accurately detect proteins in a mixture sample when the concentrations of th e proteins are within comparable ranges. A sample that causes the thrombin-aptamer an anisotropy increase, but not other DNA sequences, would probably contain -thrombin but not other proteins. Similar things can be stated for PDGF-aptamer and PDGF-BB. On the other hand, a sample that increases the anisotropies of both thrombin-aptamer and PDGF-aptamer, would probably contain a mixture of both proteins. This is the simple st example of detecti ng protein mixtures. By using more aptamers or antibodies targeted for more proteins of interest and making a larger array, we can use anisotropy imaging te chnique to easily dete ct more proteins in one sample. Conclusions We have developed a novel technique that ha s the potential to be used to build a very simple but effective protein array. It uses an imaging system to measure fluorescence anisotropy in a 2-dimensi onal format, which enables anisotropy measurements of multiple samples in an array format. Due to fluorescence anisotropy technique’s unique ability of effectively detecting protein interactions with other molecules, anisotropy imaging technique has the possibility of ma king a protein chip much simpler than conventional protein chips. It does not require tedious procedures that

PAGE 53

46 are necessary in most of the current protei n chips, such as immobilization, washing and protein staining. At the same time, it has th e same ability to realize high throughput and multiplex protein detection which is essential in a wide range of applications in immunodiagnostics, protein function and inte raction screening, and drug discovery. Although we used aptamers as the liga nds for proteins, th e application of anisotropy imaging is not limited by aptamers Basically, anything that exhibits a significant molecular weight increase after bind ing to proteins can be used as a protein ligand in anisotropy imaging, which include s antibody, antigen, DNA, RNA, and protein. This ensures the versatility of anisotr opy imaging in a variety of applications.

PAGE 54

47 CHAPTER 4 SUMMARY AND FUTURE WORK Summary We have demonstrated the possibility of detecting platelet-derived growth factor (PDGF) using aptamer based fluorescence anis otropy technique. This method has been proved to be highly sensitive, highly selective and very simple. We then extended this method to a 2-dimentional array format for possible high throughput and multiplex protein analysis by developing a fluorescence anisotropy imaging system. This system allowed us to examine multiple aptamer/pr otein binding samples simultaneously. Very small amount of samples were needed and all measurements could be done in seconds. We also used this system to demonstrate de tection of protein mixt ures. We believe this anisotropy imaging technique has the potential to help build very simple yet highly efficient protein arrays for various applications. Future Work Improving and Refining the Techniques The fluorescence anisotropy techniques we used in our work can be improved in many aspects. First, we could use a better fl uorescent dye such as th e Cy5 dye for all our experiments. It has much longer excitati on and emission wavelengths than TAMRA. Longer excitation wavelength will cau se less photobleaching and sample autofluorescence problem. It also dramatical ly reduces light scattering in biological samples as light scattering is inversely propo rtional to the fourth power of wavelength. Since light scattering is the major source of e rror, this will greatly improve the accuracy

PAGE 55

48 and sensitivity of our anisotropy measuremen ts. We can also use a high power ion laser as the light source for those longer wavelengt h dyes. Since the power of ion lasers is much higher and more stable than that of the lamp we currently use, a much better sensitivity is expected. For the anisotropy imaging system, we may need to optimize it for better sensitivity and easier handling. Currently in our system, the fluorescence signals from the sample plate have to travel a quite long dist ance before reaching the ICCD detector. This contributes to the relatively low optical efficiency and sensitivity of the system. We should be able to improve the sensitivity by using a way to image large area with a much shorter optical path. We may also improve our image processing software to combine background subtraction and image calculation in to one simple step, instead of two steps we currently use. In order to be used in real world high throughput protein analysis, the automated sample handling capability and a mo re miniaturized sample plate should be included in our system. Applications of the Anisotropy Imaging System in Cancer Diagnosis A cancer cell is a cell that grows out of c ontrol. Since growth factors are believed to play a role in intercellular communicat ion and regulating cell growth and division, people are interested in studyi ng the relationship between cancer and growth factors. However, study showed that cancer is often re lated to the level of not just one growth factor, but rather leve ls of several growth factors and other related proteins in human body [34]. The ability of our an isotropy imaging system to detect multiple proteins in a mixture may make it useful as a simple and quick cancer diagnostic method. Body fluid samples of patients with no cancer, early stag e cancer and late stage cancer can be added to different probe solutions targeted to diffe rent cancer related proteins. Levels of those

PAGE 56

49 proteins can then be determined using the anisotropy imaging system. A database can be built in a short period of time for study of the specific cancer. Anisotropy Imaging Technique for Real-Time Cell Imaging It will be of great interest if one ca n obtain anisotropy im ages of fluorescent samples in real-time. Some optical devices ha ve been developed that are able to divide the emission from the sample into two imag es of different polarization state using a polarization beam splitter. They can cast the two images side by side on a CCD detector. By using new software that allows real-time image calculation within one image, we will be able to construct an isotropy images in real time. This technique may be very useful in monitoring binding processes inside or outsi de living cells. For example, real-time monitoring of PDGF secretion from a can cer cell may be very difficult for other techniques. Using real-time anisotropy imag ing technique, we may have dye-labeled PDGF aptamer in the cell cu lture medium and monitor any anisotropy change that indicates the release of PDGF from the cell. Further more, we might also be able to monitor PDGF transportation between cells. This might help improve our understanding of intercellular signaling. We believe that the real significant applications of the fluorescence anisotropy imaging t echnique may be in the field of real-time monitoring of biological processes at cell level. This is a field that is worth more attention and investigation. The experience we gained from our anisotropy imaging experiments might help us a lot in building a sy stem for real-time cell imaging.

PAGE 57

50 LIST OF REFERENCES [1] AnandApte, B.; Zetter, B. Stem Cells 1997 15 (4): 259-267. [2] Raines, E. W.; Bowen-Pope, D. D.; Ross, R. Handbook in Experimental Pharmacology ; Sporn, M. B., Roberts, A. B., Eds.; Springer: Heidelberg, 1990, pp 173-262. [3] Blaskovich, M. A.; Lin, Q.; Delarue, F. L.; Sun, J.; Park, H. S.; Coppola, D.; Hamilton, A. D.; Sebti, A. M. Nat. Biotechnol. 2000 18 (10), 1065-1070. [4] Leitzel, K.; Bryce, W.; Tomita, J.; Manderino, G.; Tribby, I.; Thomason, A.; Billingsley, M.; Podczaski, E.; Harvey, H.; Bartholomew, M.; Lipton A Cancer Res. 1991 51 (16), 4149-4154. [5] Singh, J. P.; Chaikin, M. A.; Stiles, C. D. J. Cell Biol. 1982 95 667-671. [6] Bowen-Pope, D. F.; Malpass, T. W.; Foster, D. M.; Ross, R. Blood 1984 64 458469. [7] Heldin, C.-H. EMBO J. 1992 11 4251-4259. [8] Pillai, R. Eur. J. Surg. Oncol. 1992 18 417-424. [9] Sumedha, D.; Jayasena, Clin. Chem. 1999 45 (9), 1628–1650. [10] Andreola, M. L.; Pileur, F.; Calmels, C.; Ve ntura, M.; Tarrago-L itvak, L.; Toulme, J. J.; Litvak, S. Biochemistry 2001, 40 (34), 10087-10094. [11] Stojanovic, M. N.; de Prada, P, Landry, D. W. J. Am. Chem. Soc. 2001 123 (21), 4928-4931. [12] Gold, L. J. Biol. Chem. 1995 270 13581-13584. [13] Xu, W.; Ellington, A. D. Proc. Natl. Acad. Sci. U. S. A. 1996 93 7475-7480. [14] Potyrailo, R. A.; Conrad, R. C.; Ellington, A. D.; Hieftje, G. M. Anal. Chem. 1998 70 3419-3425. [15] Li, J. J.; Fang, X.; Tan, W. Biochem. Biophys. Res. Commun. 2002 292 (1), 31-40.

PAGE 58

51 [16] Kensch, O.; Connolly, B. A.; Steinhoff, H. J.; McGregor, A.; Goody, R. S.; Restle, T. J. Biol. Chem. 2000 275 (24), 18271-18278. [17] Green, L. S.; Jellinek, D.; Jenison, R.; Ostman, A.; Heldin, C. H.; Janjic, N. Biochemistry 1996 35 14413-14424. [18] Lakowicz, J. R. Principles of Fluorescence Spectroscopy 2nd ed.; Kluwer Academic/Plenum Publishers: New York, 1999. [19] Hey, T.; Lipps, G.; Krauss, G. Biochemistry 2001 40 (9), 2901-2910. [20] Heyduk, T.; Ma, Y.; Tang, H.; Ebright, R. H. Methods Enzymol. 1996 274 492-503. [21] Takahashi, M.; Sakumi, K.; Sekiguchi, M. Biochemistry 1990 29 (14), 3431-3436. [22] Betsholtz, C.; Johnsson, A.; Heldin, C.-H.; Westermark, B.; Lind, P.; Urdea, M. S.; Eddy, R.; Shows, T. B.; Philpott, K.; Mello r, A. L.; Knott, T. J.; Scott, J. Nature 1986 320 695-699. [23] Cordek, J.; Wang, X.; Tan, W. Anal. Chem. 1999 71 1529-1533. [24] Floege, J.; Ostendorf, T.; Janssen, U.; Burg, M.; Radeke, H. H.; Vargeese, C.; Gill, S. C. Am. J. Pathol. 1999 154 169-179. [25] Zhu, H.; Bilgin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.; Lan, N.; Jansen, R.; Bidlingmaier S.; Houfek T.; Mitc hell T.; Miller P.; Dean R. A.; Gerstein M.; Snyder M. Science 2001 293 (5537), 2101-2105. [26] Templin, M. F.; Stoll, D.; Schrenk, M.; Tra ub, P. C.; Vohringer, C. F.; Joos, T. O. Trends Biotechnol. 2002 20 (4), 160-166. [27] Harms, G. S.; Sonnleitner, M.; Schutz, G. J.; Gruber, H. J.; Schmidt, T. Biophys. J. 1999 77 (5), 2864-2870. [28] Wu, P.; Brasseur, M.; Schindler, U. Anal. Biochem. 1997 249 29–36. [29] Gough, A. H.; Taylor, D. L. J. Cell Biol. 1993 121 (5), 1095-1107. [30] Haugland, R.P. Handbook of Fluorescent Probes and Research Products 9th edition, Molecular Probes, Eugene, OR, USA, 2002. [31] Macaya, R. F.; Schultze, P.; Smith, F. W.; Roe, J. A.; Feigon, J. Proc. Natl. Acad. Sci. U. S. A. 1993 90 (8), 3745-3749. [32] Wang, K. Y.; Mccurdy, S.; Shea, R. G.; Swaminathan, S.; Bolton, P. H. Biochemistry 1993 32 (8), 1899-1904.

PAGE 59

52 [33] Bock, L. C.; Griffin, L. C.; Latham, J. A.; Vermaas, E. H.; Toole, J. J. Nature 1992 355 (6360), 564-566. [34] Harman, S. M.; Metter, E. J.; Blackman, M. R.; Landis, P. K.; Carter, H. B. J. Clin. Endocrinol. Metab. 2000 85 (11), 4258-4265.

PAGE 60

53 BIOGRAPHICAL SKETCH Zehui Cao is from Zhuzhou, Hunan Province, China. He spent 18 years in that same province before he traveled almost 1000 miles to Nanjing, a city in eastern China, for college study. He got his B.S. degree in chemistry from Nanji ng University in 1998. He spent one more year at the graduate school of Nanjing Univ ersity studying organic chemistry before he finally realized he was more attracted to control panels than to flasks and beakers. So he traveled again over 10000 miles to University of Florida to study analytical chemistry. He likes what he is doing and would like to pursue a Ph.D. degree after graduation.


Permanent Link: http://ufdc.ufl.edu/UFE0000511/00001

Material Information

Title: Using fluorescence anisotropy for sensitive platelet-derived growth factor detection based on molecular aptamers
Physical Description: vii, 53 p.
Language: English
Creator: Cao, Zehui ( Dissertant )
Tan, Weihong ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2002
Copyright Date: 2002

Subjects

Subjects / Keywords: Anisotropy   ( lcsh )
Chemistry thesis, M.S
Fluorescence   ( lcsh )
Dissertations, Academic -- UF -- Chemistry
Proteins -- Analysis   ( lcsh )

Notes

Abstract: Proteins play very important roles in almost all functions of life. Detection of proteins has drawn great attention from scientists for many years. However, analytical methods for real-time protein detection in homogeneous solutions are scarce. The recent development of molecular aptamers, combined with fluorescence techniques, may provide an easy and efficient approach to sensitive protein analysis. Aptamers are small oligonucleotides which have high affinity and high selectivity to their target molecules. They are isolated by the systematic evolution of ligands by exponential enrichment (SELEX) process. We have designed a fluorescein-labeled aptamer for the detection of platelet-derived growth factor (PDGF) protein which has the potential as a cancer indicator. Fluorescence anisotropy was used as the detection method. Fluorescent molecules, when excited by plane-polarized light, will give an emission that has been depolarized to a certain extent. This can be described in terms of anisotropy. Factors that can change the rotational diffusion of the fluorescent molecules will affect their anisotropy, including their molecular weights and the properties of the surrounding environment. Fluorescence anisotropy technique is ideal for aptamer/protein binding assays where there is a large molecular weight change. By combining aptamers and fluorescence anisotropy, we are able to achieve highly sensitive PDGF detection with high selectivity. The detection limit was about 0.22 nM. We then extended this technique into a two dimensional format by building a fluorescence anisotropy imaging system. This format enables the simultaneous anisotropy measurements of multiple samples. We demonstrated the ability of this system to detect multiple proteins in one sample. With some modifications and improvements, the anisotropy imaging technique may have the potential to be a very effective and easily implemented approach to high throughput and multiplex protein analysis in an array format.
Subject: anisotropy, aptamers, arrays, flourescence, PDGF, protein
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 60 p.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2002.
Bibliography: Includes bibliographical references.
Original Version: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000511:00001

Permanent Link: http://ufdc.ufl.edu/UFE0000511/00001

Material Information

Title: Using fluorescence anisotropy for sensitive platelet-derived growth factor detection based on molecular aptamers
Physical Description: vii, 53 p.
Language: English
Creator: Cao, Zehui ( Dissertant )
Tan, Weihong ( Thesis advisor )
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2002
Copyright Date: 2002

Subjects

Subjects / Keywords: Anisotropy   ( lcsh )
Chemistry thesis, M.S
Fluorescence   ( lcsh )
Dissertations, Academic -- UF -- Chemistry
Proteins -- Analysis   ( lcsh )

Notes

Abstract: Proteins play very important roles in almost all functions of life. Detection of proteins has drawn great attention from scientists for many years. However, analytical methods for real-time protein detection in homogeneous solutions are scarce. The recent development of molecular aptamers, combined with fluorescence techniques, may provide an easy and efficient approach to sensitive protein analysis. Aptamers are small oligonucleotides which have high affinity and high selectivity to their target molecules. They are isolated by the systematic evolution of ligands by exponential enrichment (SELEX) process. We have designed a fluorescein-labeled aptamer for the detection of platelet-derived growth factor (PDGF) protein which has the potential as a cancer indicator. Fluorescence anisotropy was used as the detection method. Fluorescent molecules, when excited by plane-polarized light, will give an emission that has been depolarized to a certain extent. This can be described in terms of anisotropy. Factors that can change the rotational diffusion of the fluorescent molecules will affect their anisotropy, including their molecular weights and the properties of the surrounding environment. Fluorescence anisotropy technique is ideal for aptamer/protein binding assays where there is a large molecular weight change. By combining aptamers and fluorescence anisotropy, we are able to achieve highly sensitive PDGF detection with high selectivity. The detection limit was about 0.22 nM. We then extended this technique into a two dimensional format by building a fluorescence anisotropy imaging system. This format enables the simultaneous anisotropy measurements of multiple samples. We demonstrated the ability of this system to detect multiple proteins in one sample. With some modifications and improvements, the anisotropy imaging technique may have the potential to be a very effective and easily implemented approach to high throughput and multiplex protein analysis in an array format.
Subject: anisotropy, aptamers, arrays, flourescence, PDGF, protein
General Note: Title from title page of source document.
General Note: Document formatted into pages; contains 60 p.
General Note: Includes vita.
Thesis: Thesis (M.S.)--University of Florida, 2002.
Bibliography: Includes bibliographical references.
Original Version: Text (Electronic thesis) in PDF format.

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0000511:00001


This item has the following downloads:


Full Text











USING FLUORESCENCE ANISOTROPY FOR SENSITIVE PLATELET-DERIVED
GROWTH FACTOR DETECTION BASED ON MOLECULAR APTAMERS
















By

ZEHUI CAO


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2002































Copyright 2002



by


Zehui Cao















ACKNOWLEDGMENTS

I would like to thank my parents for what they have gone through to bring me up

to where I am now, and my wife, Qian, for her support and patience all the time. I would

also like to thank my advisor, Dr. Weihong Tan, for his confidence in me and the

encouragement he gave me to continue my study, Dr. Xiaohong Fang for her guidance in

the early PDGF work, Dr. James Winefordner for his kind help and understanding during

my difficult times, the members of my advisory committee for their helpful guidance and

suggestions, and the Tan research group.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ......... .................................................................................... iii

A B STR A C T ............... ............................................................................................ vi

CHAPTER

1 INTRODUCTION ...... ........ ..................... ...... ........... .............. 1

Introduction to PD G F Protein............. ................................ .................. .............. 2
A ptam er as a Probe for Protein........... ................................................. .............. 3
F lu orescen ce A n isotropy ................................................................................................ 5

2 MOLECULAR APTAMER FOR REAL-TIME PDGF DETECTION USING
FLUORESCEN CE ANISOTROPY ............................................ ........................... 8

E xperim mental Section .......................................................... ....... .... ...... ........ .. 8
M materials .............................................................. .......... ...... 8
Instrum entation ........................................ ................ ...................... ......... 9
A nisotropy M easurem ents .......................................................................................... 9
Results and Discussions ................ .................................. .... ........ .. ........ .. 10
Design of an Fluorescent Aptamer Probe....................... ........ ............. 10
Binding of the Aptamer to PDGF ............................ .................. 11
Real-Time Binding Detection Using Fluorescence Anisotropy ............................ 12
Effects of Mg2+ on the Binding Assay.................................. .............. 14
Effect of Temperature on the Binding Assay ................................................... 16
Detection of PDGF-BB in Homogeneous Solution............................................ 17
Selectivity of the Aptamer Probe.... .. ............................................ .. .............. 21
Conclusions...................... ............... ..... .............. 23

3 DEVELOPMENT OF FLUORESCENCE ANISOTROPY IMAGING SYSTEM
F O R PR O T E IN A R R A Y S ......................................................... ................................. 25

Protein Arrays .............. .. .................. .................................. ........ 25
Fluorescence Anisotropy Im aging ....................................................... .............. 28
Experim mental Section ............................................................................. .. 30
M materials ............................... .................... .......... 30
A nisotropy Im aging Setup .......................................................... .............. 31
A acquisitions of A nisotropy Im ages ........................................ ........ .............. 32
R results and D discussions ............................................... ........ .. .......... 35









Experim mental Considerations ........................................................ ........... .... ..... 35
Anisotropy Imaging of TAMRA-Glycerol Solutions............................................. 36
Anisotropy Imaging for Aptamer-Protein Binding Assay.................................... 38
Selectivity of Aptamer/PDGF Binding on the Anisotropy Imaging System.......... 41
D election of Protein M ixture ......................................................... ........ ...... 43
C onclusions............................... ........... .......... 45

4 SUMMARY AND FUTURE WORK ........................................................ 47

S u m m ary ............................................................................... 4 7
Future W ork ........................................ ................... ..... ..... ........ 47
Improving and Refining the Techniques................. ... ... .............. 47
Applications of the Anisotropy Imaging System in Cancer Diagnosis .................... 48
Anisotropy Imaging Technique for Real-Time Cell Imaging............................ 49

L IST O F R EFER EN CE S ......... ................................................................ .............. 50

B IO G R A PH ICA L SK ETCH ................................................ ..................... .............. 53















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

USING FLUORESCENCE ANISOTROPY FOR SENSITIVE PLATELET-DERIVED
GROWTH FACTOR DETECTION BASED ON MOLECULAR APTAMERS

By

Zehui Cao

December 2002


Chair: Dr. Weihong Tan
Major Department: Chemistry

Proteins play very important roles in almost all functions of life. Detection of

proteins has drawn great attention from scientists for many years. However, analytical

methods for real-time protein detection in homogeneous solutions are scarce. The recent

development of molecular aptamers, combined with fluorescence techniques, may

provide an easy and efficient approach to sensitive protein analysis. Aptamers are small

oligonucleotides which have high affinity and high selectivity to their target molecules.

They are isolated by the systematic evolution of ligands by exponential enrichment

(SELEX) process. We have designed a fluorescein-labeled aptamer for the detection of

platelet-derived growth factor (PDGF) protein which has the potential as a cancer

indicator. Fluorescence anisotropy was used as the detection method. Fluorescent

molecules, when excited by plane-polarized light, will give an emission that has been

depolarized to a certain extent. This can be described in terms of anisotropy. Factors that

can change the rotational diffusion of the fluorescent molecules will affect their









anisotropy, including their molecular weights and the properties of the surrounding

environment. Fluorescence anisotropy technique is ideal for aptamer/protein binding

assays where there is a large molecular weight change. By combining aptamers and

fluorescence anisotropy, we are able to achieve highly sensitive PDGF detection with

high selectivity. The detection limit was about 0.22 nM. We then extended this technique

into a two dimensional format by building a fluorescence anisotropy imaging system.

This format enables the simultaneous anisotropy measurements of multiple samples. We

demonstrated the ability of this system to detect multiple proteins in one sample. With

some modifications and improvements, the anisotropy imaging technique may have the

potential to be a very effective and easily implemented approach to high throughput and

multiplex protein analysis in an array format.














CHAPTER 1
INTRODUCTION

Proteins are macromolecules that consist of one or more unbranched chains of

amino acids. Although typical proteins contain 200-300 amino acids, they can be much

smaller or much larger. Proteins are the building blocks of life. Almost every function in

living cells depends on proteins. Those include catalysis of all biochemical reactions,

construction of cells, motion of cells and organisms, transportation of materials in body

fluid and many more. Since proteins realize their functions through interactions with

other molecules, it is highly important to understand those interactions in order to find

out how proteins work in living cells. Many techniques have been developed to detect

and analyze proteins. Some of them, such as electrophoresis and affinity chromatography,

have provided a good way to separate and detect proteins. However, they lack the ability

of monitoring protein interactions in real time and in homogeneous solutions.

Fluorescence techniques, on the other hand, have shown great capability in detecting and

studying protein functions in their native environments. This ensures a more direct and

precise understanding of protein interactions. Many protein probes have been used for

protein detection using fluorescence techniques. Some of which are extracted from

animals that have the inherent ability of binding to certain proteins, such as antibodies.

They have great affinity and selectivity towards proteins, but they have more restricted

requirements for their surrounding environments in order to function properly. Other

probes have been synthesized or selected by scientists, such as aptamers. Those synthetic

protein probe molecules can also bind to proteins with high affinity and high selectivity,









but they are more robust and much easier to obtain. In our work, we were trying to

develop a fluorescence detection technique, particularly a fluorescence anisotropy

technique, for proteins using aptamers as the probe. We later extended and modified this

technique to an array format as we tried to examine the possibility of using fluorescence

anisotropy for simple and accurate high throughput protein screening and multiplex

protein detection. We specifically developed our technique for platelet-derived growth

factor (PDGF) detection because of the great interest in this protein for cancer studies.

Introduction to PDGF Protein

Understanding disease-related proteins could be the very first and most important

step in disease studies and drug discovery. As increasing attention has been focused on

cancer diagnosis, it is of great interest to find out more about those proteins that may be

related to cancers. One of which is platelet-derived growth factor (PDGF) found in many

human cell types. Its biological function is to stimulate the division and proliferation of

the cells through binding of its receptors on cell membranes. It is also believed to play a

role in intercellular signaling [1]. PDGF has several isoforms, among which are PDGF-

AA, PDGF-BB and PDGF-AB. Those isoforms consist of the two subunits, PDGF-A

chain and PDGF-B chain.

The three isoforms act differently in specific situations. PDGF-BB has been

shown to be actively involved in cell transformation process and in tumor growth and

progression [2-6]. PDGF-BB, when bound to its receptor on the cell membrane, activates

phosphatidylinositol 3-kinase (PI3-K) inside the cell, which eventually leads to cell

growth. PDGF-BB is expressed at undetectable low levels in normal cell [7] while it is

often found over-expressed or mutated in a malignant tumor [8]. Because of the potential









as a cancer indicator, its detection has been attempted using the traditional antibody based

radioisotropic methods and ELISA techniques [3-6].

Aptamer as a Probe for Protein

For the last several decades, antibodies have been the most important probes for a

variety of molecular recognition applications. Many diagnostic tests based on antibodies

are routinely conducted in laboratories and clinics. The great success in these applications

is the result of their very high sensitivity and high selectivity to the target molecules.

However, antibody technology also has some major disadvantages. It relies on the animal

host to produce antibodies, which means high cost, low efficiency and the very limited

number of target molecules that can be detected. Antibodies are very sensitive to their

surrounding environments and often easily undergo irreversible denaturation.

Regeneration of antibodies is usually not easy, which also contributes to the high cost.. In

some cases, ligand density is very limited by the antibodies are large molecules. It is also

relatively more difficult to directionally immobilize antibodies onto a surface, which may

be critical in some applications.

Recent development of the systematic evolution of ligands by exponential

enrichment (SELEX) process, however, may provide solutions to some of the problems

associated with antibodies. This process has the ability to isolate the oligonucleotide

sequences that recognize virtually any class of target molecules with high sensitivity and

high selectivity [9]. The resulting oligonucleotide ligands were given the name

"aptamers," which comes from the Latin word "aptus," meaning "to fit." The SELEX

process begins with a library of synthesized oligonucleotides usually containing 1014 to

1015 random sequences. This library is then incubated with the target molecule of interest

under certain conditions. The sequences that interact with and bind to the target









molecules are isolated for next round of incubation. This process is repeated until a

sequence that binds to the target with highest selectivity and affinity is determined.

SELEX enables the discovery of ligands to virtually any target molecules of interest.

Some aptamers have been well studied, such as the aptamer that specifically interacts

with human a-thrombin. New aptamers are also being discovered for a variety of

different target molecules [10]. Although most aptamers are exploited to study proteins,

there are some applications that use aptamers for detecting smaller molecules such as

cocaine [11].

Compared to antibodies, aptamers have similar high affinity and selectivity for

proteins [12,13]. What makes aptamers so useful is that they have several important

advantages over antibodies. First, their production is easier, cheaper and not limited by

the animal hosts. Because oligonucleotides have more stable structures than proteins,

aptamers can withstand harsher experimental conditions than antibodies and can be

stored and reused without causing much degradation. Aptamers can be easily labeled or

modified in different ways for different molecular recognition applications. They can also

be easily immobilized onto solid surfaces without much change in the binding affinities

to proteins. In one of the examples, the aptamer for human a-thrombin was labeled with a

fluorophore and then immobilized onto glass surface for high sensitivity thrombin

detection [14]. In another example, the a-thrombin aptamer was labeled with a

fluorophore and a fluorescence quencher at its two ends to form an aptamer beacon [15].

The change in the aptamer's conformation upon binding to a-thrombin brought the

fluorophore and quencher far away from each other, thus causing the restoration of

fluorescence. Highly sensitive thrombin detection was also achieved based on this









scheme. Unlike antibodies, aptamers can also be used to inhibit target protein's normal

functions by occupying the active binding sites of the protein. This gives aptamers the

potential to be drug candidates for numerous diseases. Traditional protein inhibitors

either are too toxic or lack good specificity. Aptamers, on the other hand, are just small

DNAs or RNAs with high affinity and high selectivity for proteins. Thus, they should be

very safe and effective. Some work has been done to investigate possibility of aptamers

inhibiting the activity of HIV virus [10,16].

High-affinity aptamers for PDGF-B chain have been reported using SELEX [17].

Several single-stranded DNA sequences were found to bind to PDGF-AB and PDGF-BB

with high affinity (Kd Z 10-10 M) while to PDGF-AA with lower affinity (Kd > 10-8 M).

Most of the ligands were found to have a structure of a three-way helix junction with a

three-nucleotide loop at the branch point.

Fluorescence Anisotropy

Fluorescence anisotropy has been effectively used previously to study interactions

between macromolucules. Upon excitation with polarized light, some samples will have

polarized emission. The extent of polarization of the emission is described in terms of

anisotropy (r) [18]. The theory behind the polarized emission can be explained as

following: when excited by a polarized light, the sample molecules that have absorption

transition moments oriented along the electric vector of the incident light are

preferentially excited. Those excited molecules may rotate to other directions before

returning from the excited state to the ground state and emitting light, thus causing a

depolarized emission. This depolarization is dependent on the extent of the rotational

diffusion of the excited molecules. Two factors are believed to affect the diffusive motion

of a molecule. One is the viscosity of the solvent surrounding the molecule and the other









is the size of the molecule itself. A small fluorophore molecule in a solvent with low

viscosity will be almost fully depolarized on the excited state and exhibit a non-polarized

emission. The anisotropy in this case is close to zero.

To measure the anisotropy of a sample, the sample is excited with a vertically

polarized light. The intensity of the vertically polarized component (Ivv) and horizontally

polarized component (IVH) of the emission are measured with a polarizer on the emission

side. The anisotropy r is then calculated using the following equation:

r = (Ivv-IvH) / (Ivv+2IVH)

where the subscripts V and H refer to the orientation (vertical or horizontal) of the

polarizers for the intensity measurements, with the first subscript indicating the position

of the excitation polarizer and the second for the emission polarizer.

Note that even though from this equation, the range of anisotropy value is from 0

to 1, in a real-world sample, with the molecules evenly distributed in all directions, the

maximal anisotropy one can observe with one-photon excitation is 0.4.

Fluorescence anisotropy is a simple signaling method for binding assays. In a

typical binding assay, a small dye-labeled probe molecule, such as an aptamer, binds to a

large target molecule, such as a protein. The increased molecular size will slow down the

rotational movement of the dye molecule linked to the probe, thus causing a more

polarized emission. This polarization, or the binding event, can be reflected by an

increase in the measured anisotropy value. Traditional binding assays like ELISA require

multiple steps and complex procedures. With fluorescence anisotropy, the mixing of the

binding probe and the target is the only step needed. Anisotropy measurement can then

be performed directly on the sample mixture and the result will clearly show whether









there is binding between the probe and the target. Compared to techniques such as

surface plasmon resonance (SPR) and total internal reflection fluorescence (TIRF),

fluorescence anisotropy is insensitive to changes of the refractive index of the sample

solution. Compared to other fluorescence based techniques for biomolecular interaction

study, fluorescence anisotropy requires only one fluorescent dye molecule on the probe.

Because all that is needed to give a signal change in anisotropy is simply the change in

the size of the molecule linked to the dye, there is no need to worry about whether the

conformational change of the probe after the binding can give a signal, such as in

molecular beacon techniques. Furthermore, because fluorescence anisotropy is a

rationing technique, some problems associated with fluorescence intensity techniques,

such as photobleaching, nonuniform illumination and unstable light source, are not of

major concern.

Applications of fluorescence anisotropy technique are not limited to protein

detection. Fluorescence anisotropy has also been used to study biomolecular interactions

in order to understand some biological processes [19]. Although in a lot of cases,

fluorescence anisotropy is used to study protein-DNA interactions, it is also suitable for

protein-protein [20], DNA-DNA and other type of interactions. Theoretically, as long as

there is a molecular weight change after a binding process, there should be an anisotropy

change for the fluorescent molecule. This is not limited to small fluorescent molecules

binding to large ones. Some work was done using fluorescent protein molecules to study

protein-DNA interaction [21].














CHAPTER 2
MOLECULAR APTAMER FOR REAL-TIME PDGF DETECTION USING
FLUORESCENCE ANISOTROPY

Despite the presence of current techniques for protein detection, there is still few

nonisotropic and sensitive methods for real time protein analysis in homogeneous

solutions. In this part of our work, we have developed an easy and effective way to

specifically analyze PDGF due to its potential significance in cancer research.

Experimental Section

Materials

The fluorescein-labeled PDGF aptamer was customer-designed and then

synthesized by Trilink Biotechnologies (San Diego, CA). The sequence of the aptamer is

5'-fluorescein-CAGGC TACGG CACGT AGAGC ATCAC CATGA TCCTG.

Recombinant human PDGF-BB, PDGF-AB, and PDGF-AA were purchased from R&D

Systems (Minneapolis, MN). They were dissolved in 4 mM HC1 and then diluted in a

Tris buffer before use. Other recombinant human growth factors, epidermal growth factor

(EGF), and insulin-like growth factor 1 (IGF-1), were bought from Roche (Indianapolis,

IN). Human bovine serum albumin (BSA), human hemoglobin (HEM), porcine lactic

dehydrogenase (LDH), horse myoglobin (MYO), chicken lysozyme (LYS), and human y-

thrombin (THR) and other chemicals were from Sigma (St. Louis, MO). The buffer we

used consisted of 20 mM Tris-HCl (pH 7.1), 140 mM NaC1, 5 mM KC1, 1 mM CaC12,

and 1 mM MgC12 to simulate the ionic strength under physiological conditions.

Superpurified water was used to prepare all of the solutions.









Instrumentation

Fluorescence measurements were performed on a Fluorolog-Tau-3

spectrofluorometer (Jobin Yvon, Inc., Edison, NJ) equipped with a thermostat accurate to

0.1 C. All experiments were carried out at 37 C unless otherwise specified. The sample

cell was a 100-pL cuvette. The fluorescence intensity of the aptamer was monitored by

exciting the sample fluoresceinn label) at 470 nm and measuring the emission at 520 nm.

Slits for both the excitation and emission were set to 10 nm. Corrections were also made

for potential effects on sample concentrations caused by dilutions in the titration

experiments. To achieve a better analytical sensitivity of PDGF by the anisotropy

measurement, the emission monochromator box in the spectrofluorometer was removed.

A 515-nm long-pass filter (Oriel, Stratford, CT) and a 525-nm bandpass filter (Chroma,

Brattleboro, VT) were put in front of the PMT to select the desired fluorescence signal

from the polarizer. This increased the optical signal collection efficiency, as compared

with the spectrometer's original optical path.

Anisotropy Measurements

Fluorescence anisotropy was measured using the L-format configuration (Figure

2-1). The anisotropy (r) is calculated according to the following equation:

r = (Ivv-GIvH) / (Ivv+2G'IvH)

where the subscripts V and H refer to the orientation (vertical or horizontal) of the

polarizers for the intensity measurements, with the first subscript indicating the position

of the excitation polarizer and the second for the emission polarizer. G is the G-factor of

the spectrofluorometer, which is calculated as G = IHV/IHH. G-factor represents the ratio

of the sensitivities of the detection system for vertically and horizontally polarized light.









G-factor is dependent on the emission wavelength and can be determined using a

standard sample with known anisotropy.


Polarizer Iv Sample Polarizer Iv Sample

Light .
source
+' Sample A Sample
emission emission

Polarizer Polarizer

T Ivv IVH



Detector Detector
Figure 2-1 Schematic diagram for L-format measurements of fluorescence anisotropy

In all of our experiments, each anisotropy data point was the average of six

measurements with an integration time of 1 second. The relative standard deviation was

<2% for all measurements.

Results and Discussions

Design of an Fluorescent Aptamer Probe

High-affinity aptamers for PDGF-B chain have been developed to inhibit binding

of PDGF-BB to its receptor [17]. The consensus secondary structure motif of the PDGF

aptamers is a three-way helix junction with a conserved single-stranded loop at the

branch point.

The helix junction domain of the aptamer represents the core of the structural

motif required for high-affinity binding. The binding of the aptamers has been studied by

radiolabeling an aptamer using a nitrocellulose filter-binding method. To construct a









fluorescent probe for PDGF-BB based on the aptamers, we adopted a 35-base single-

stranded DNA sequence similar to the reported 36-base aptamer [17]. The 5' end of the

sequence was covalently linked to a fluorescein molecule as a fluorescence reporter. The

structure of the aptamer is shown in Figure 2-2.




C CA

G \T C A
G-^ A AGC A





GOC

A*T

COG
I
F'
Figure 2-2 Structure of the fluorescein-labeled aptamer probe.

The fluorescein label is far away from the helix junction, which is the binding

center for PDGF-B [17]. So it will not affect the binding affinity of the aptamer. Our gel

electrophoresis result clearly showed the effective binding of the dye-labeled aptamer to

PDGF-BB.

Binding of the Aptamer to PDGF

The binding between the aptamer and the PDGF-BB molecule was confirmed by

gel electrophoresis (Figure 2-3). The experiment was done on a 4-20% precast gradient

polyacrylamide gel (Bio-Rad, Hercules, CA). A running buffer of 0.5 x TB (Bio-Rad) was

used. The gel was prerun at 60 V for 10 min, and then samples were introduced and run









at 150 V for 70 min. Fluorescence images were taken with a Kodak DC290 digital

camera (Eastman Kodak, Rochester, NY) and a UV illuminator. On the gel image, the

right lane contained the fluorescein-labeled aptamer and resulted in only one bright band.

The left lane, which had dye-labeled aptamer and unlabeled PDGF-BB, showed two

bands. The weaker band appeared at the same position as the bright band on the right

lane, indicating the unbound aptamer. The brighter band on the left lane moved much

slower than the aptamer band, which reveals the binding complex of the aptamer to the

protein. With the ratio of the aptamer:PDGF being 1:10, only a very small fraction of the

aptamer was unbound.



















Figure 2-3 Gel electrophoresis of the aptamer and its binding complex with PDGF. The
sample injected to the right lane is 0.1 kM aptamer, and that in the left lane was the
mixture of 0.1 kM aptamer and 1.0 kM PDGF.

Real-Time Binding Detection Using Fluorescence Anisotropy

The fluorescein-labeled aptamer is a relatively small molecule compared to PDGF-BB

molecule (M.W.=25000). When the aptamer binds to the protein, the increase in overall

size and molecular weight will greatly slow down the rotation of the fluorescein









molecule. The slower rotational diffusion will result in a lower ability of the fluorescein

molecule to depolarize the incident polarized excitation. Consequently, the anisotropy of

the emission will also increase to reflect the slower motion of the dye itself. Based on

this, we believe the fluorescence anisotropy should be a simple and reliable method for

studying the interaction between aptamer and PDGF molecules, and for detecting the

protein as well. Our anisotropy measurement showed a significant increase in anisotropy

with the addition of PDGF-BB to the aptamer solution (Figure 2-4).



0.2


0.15


0 0.1


0.05


0
-20 -10 0 10 20 30 40
t(s)

Figure 2-4 Anisotropy change upon the binding of PDGF to the aptamer. The
concentration of the aptamer is 0.1 [tM. A 1.0 [tM PDGF solution was added at time 0 s.
The time resolution for the data collecting was 3.3 s.

As shown in this figure, with the addition of PDGF-BB, the anisotropy increased

more than 2-fold. In a time period of about 3 seconds, anisotropy increase was observed

and the anisotropy value remained stable after that, meaning that the binding between the

aptamer and the PDGF was fast and stable. Control experiments were conducted with just

fluorescein dye solutions and PDGF under the same conditions as above. No anisotropy

change was observed, confirming that the anisotropy increase in the aptamer experiment









was really due to the interaction between the aptamer and the protein. All measurements

were done in real-time; no pre-separation was needed, and no need to label the target

protein. It should be mentioned that after the binding, the overall fluorescence intensity

decreased about 30%. Since there was no intensity decrease in the fluorescein dye and

PDGF control experiment, we concluded that this intensity decrease was due to the

formation of the binding complex, which brought the dye molecule and protein close to

each other. The oligo bases may also have an effect if they were closer to the dye after

the binding.

Effects of Mg2+ on the Binding Assay

Metal ions in the solution often have a significant impact on the binding of a

single-stranded DNA (ssDNA) to a protein. A series of experiments designed to study the

effects of Mg2+ ions on the binding between the aptamer and PDGF molecules were

carried out. The concentration of Mg2+ was varied in a solution with only the fluorescein-

labeled aptamer and also in a solution with same amount of aptamer and excess PDGF-

BB molecules. The anisotropies of the solutions were measured at each Mg2+

concentration and results are shown in Figure 2-5.

The results showed that the anisotropy of the fluorescein-labeled aptamer

increased significantly as Mg2+ concentration increased. This is because the divalent

metal ions stabilize the three-way helix structure of the free aptamer [17], thus hindering

the rotational rate of the labeled fluorophore linked to the aptamer.

In the case of aptamer/PDGF binding complex, the Mg2+ had a very small effect

on the anisotropy until its concentration reached 2 mM. With higher Mg2+ concentrations,

the anisotropy of the solution began to decrease, which might be the result of the effects

of high ionic strength on the conformation of the protein.












A 0.07
0.06
0.05
g 0.04
0
0.03
0.02
0.01
0
0 2 4 6 8 10 12
CMg (mM)
B
0.16
0.14
0.12
,. 0.1
o 0.08
0.06
0.04
0.02
0
0 2 4 6 8 10 12
CMg (mM)

Figure 2-5 Effects of Mg2+ concentration on sample's anisotropy. (A) 0.1 tM aptamer in
buffer; (B) 0.1 pM aptamer and 1.0 gM PDGF-BB in buffer.

The combination of the two effects of Mg2+ is shown in Figure 2-6. It shows that

at different Mg2+ concentration, the anisotropy change between the aptamer and the

aptamer/PDGF binding complex is different. Higher Mg2+ concentration results in a

smaller anisotropy change. A larger anisotropy change is desirable when detecting

protein. However, in order to study the ability of our method to detect PDGF-BB under

physiological conditions, we still chose to use a buffer with 1 mM Mg2+










0.14

& 0.12

| 0.10

| 0.08

0.06

S0.04

S0.02

0.00
0 2 4 6 8 10 12
CMg (mM)

Figure 2-6 Effect of MgCl2 on the aptamer/PDGF binding. The concentrations of the
aptamer and PDGF were 0.1 pLM and 1.0 pLM respectively.

Effect of Temperature on the Binding Assay

Temperature also has its effect on a fluorescent molecule's anisotropy. As

temperature increases, the rotation ability of a molecule will increase and the viscosity of

the surrounding solvent will decrease. Both factors contribute to a lower fluorescence

anisotropy. Since temperature will have an effect on the anisotropies of both aptamer and

the aptamer/PDGF binding complex, we did experiments to find out what the overall

effect would be on this binding process. Experiments were carried out in a similar

manner as that in the study of Mg2+ effect. Anisotropies of both an aptamer solution and

an aptamer/PDGF binding complex solution were measured at different temperatures, as

shown in Figure 2-7. Even though higher temperature decreased the anisotropies of both

aptamer and the aptamer/PDGF binding complex (data not shown), a gradual increase in

the anisotropy difference between the two species was observed. However, at 40 OC,









there was a significant decrease in the anisotropy of the binding complex (data not

shown), which contributed to the sharp decrease at 40 OC in Figure 2-7.



0.1

c 0.08

0.06
C.
o 0.04
o
'E 0.02

0 -
15 20 25 30 35 40
Temperature (OC)

Figure 2-7 Effect of temperature on the anisotropy change during the aptamer/PDGF-BB
binding process. Solutions contained 0.1 gM aptamer and 1.0 [M PDGF-BB in buffer.

This may indicate that at this temperature, the three-way helix junction structure

of the aptamer, which is essential to the aptamer/ PDGF binding, is much less stable. As a

result, more aptamer molecules became unbound, and the anisotropy decreased. In fact,

40 C is close to the reported melting temperature of the aptamer (about 44 C), at which

the aptamer changes from the folded structure to an unfolded one [17]. This result also

shows that fluorescence anisotropy is not just a detection technique. It can also be a

useful tool for studying and understanding the conformational changes of molecules in

some biological processes.

Detection of PDGF-BB in Homogeneous Solution

To test if the fluorescence anisotropy method is a practical way to quantitatively

detect PDGF-BB in homogeneous solution, we carried out a series of titration

experiments to construct a calibration curve. The experiments were performed by adding









aliquots of PDGF-BB stock solution to an aptamer solution. Potential dilution effects on

sample concentrations during the titration experiments were corrected. A control

experiment was also conducted using exactly the same procedures and conditions as in

the PDGF titration experiments, with the only difference being that we used the buffer

alone to do the titration instead of the PDGF-BB stock solution. The results are shown in

Figure 2-8.


0.09

0.08

S0.07

0.06 -

S0.05

0.04

0.03 .....
0 0.2 0.4 0.6 0.8 1

CPDGF-BB (pM)

Figure 2-8 Titration of the aptamer with (A) PDGF solutions in the concentration range
of 0-1.0 pM; (+) blank buffer. The aptamer concentration is 0.1 pM.

The control experiment showed a stable anisotropy value and no increase,

meaning that volume increase or aptamer concentration decrease does not affect the

anisotropy of the aptamer. With the addition of PDGF-BB, the anisotropy of the aptamer

solution greatly increased, indicating the binding between the aptamer and the PDGF-BB.

This increase, however, is not linear to the PDGF concentration. This is because

the measured anisotropy of a sample at a certain wavelength is actually the average

anisotropy of all the components in this sample that have emission at this wavelength.









For example, for a solution containing a dye-labeled aptamer and its target protein, if the

binding is a 1:1 ratio reaction, there will be only two species that may have fluorescence

anisotropy: the aptamer itself and the aptamer/protein binding complex. And the

anisotropy of this solution (r) can be calculated using the following equation:

r = (l-X) raptamer + X complex (Equation 1)

where raptamer and complex are the anisotropies of pure aptamer and pure binding complex

respectively, and x is the fraction of the total aptamer that is bound to the protein to form

the binding complex. It is clear that the anisotropy of such a system is related to the

aptamer/protein concentration ratio and the constant of the binding reaction. In a more

complex system, more components need to be considered in the above equation. For

PDGF-BB and its aptamer, the titration curve we obtained (Figure 2-8) indicated that the

interaction between aptamer and PDGF-BB might not be monophasic, but rather

biphasic, as there seems to be two plateaus in the curve. This is in agreement with the

previous report that studied the aptamer/PDGF binding using radioisotropic technique

[17]. The reason for this biphasic binding is proposed to be the coexistence of two

noninterconverting components of the aptamer that bind to the PDGF with different

affinities.

Despite the biphasic curve in a range of PDGF-BB concentration up to 1 atM, at

lower PDGF-BB concentration (0-100 nM) there is a good linear relationship between the

anisotropy and the amount of PDGF-BB. Clinical studies showed that the concentration

of PDGF is about 0.4-0.7 nM in human serum and 0.008-0.04 nM in human plasma [4-6].

However, the PDGF concentration in the local tumor area should be higher than that in

the blood as the PDGF has not diffused into blood. To demonstrate that our method is









capable of detecting PDGF in that concentration range, we made some modifications to

our experiments. Before modification, our detection limit was -2 nM of PDGF-BB. We

first removed the monochromator box on the emission side of the spectrofluorometer and

put a band-pass filter directly between the sample and the PMT detector. This resulted in

a broader emission wavelength range, a shorter optical path, and a more efficient optical

signal collection. The second thing we changed was using an initial aptamer

concentration of 2 nM, instead of 0.1 tM we used before. According to Equation 1 and

the principles behind it, the anisotropy change caused by the addition of PDGF is not

related solely to the PDGF concentration. It also depends on the initial aptamer

concentration. The lower aptamer concentration, the less PDGF is needed to achieve a

similar level of anisotropy change. Because of this reason, as long as our detection

system is sensitive enough to give sufficient signal for 2 nM aptamer, we should be able

to detect a PDGF-BB concentration much lower than 2 nM. We carried out a series of

experiments with the modifications mentioned above, and obtained a linear curve in a

lower PDGF-BB concentration range (Figure 2-9). The sensitivity of our method was

greatly improved as indicated in the figure. The detection limit was calculated to be -0.22

nM of PDGF-BB when 2 nM fluorescein-labeled aptamer was used in the binding assay.

We believe this detection limit will be feasible for PDGF detection in blood serum

samples and in local tumor fluid samples. But in order to be useful for PDGF detection in

clinical blood plasma sample, our method still needs to be improved. Possible approaches

could be optimization of system design and optical detection, using a dye with higher

fluorescence intensity and using a better light source, such as a laser.










0.056


S0.052


o 0.048


< 0.044


0.04 i --r*
0 3.25 0.5 0.75 1 1.25
C PDGF (nM)
Figure 2-9 Titration of the aptamer with PDGF solutions in the concentration range of 0-
1.25 nM. The aptamer concentration is 0.2 nM.

Selectivity of the Aptamer Probe

It is essential for a protein probe to have the ability to distinguish between its

target protein and other proteins in order to be of practical use. While the aptamer

sequence has been reported to be highly selective for PDGF by other techniques, it is still

important for us to test the selectivity of the probe using fluorescence anisotropy. We

need to show that the fluorescent label and the anisotropy method do not affect in anyway

the selectivity of the aptamer probe for PDGF. We carried out a series of experiments

where excess of several common extracellular proteins, such as albumin, hemoglobin,

myoglobin and lysozyme, etc., were added to aptamer solutions and the anisotropy

changes were recorded. These changes were then compared to the anisotropy change

caused by PDGF. The result is shown in Figure 2-10. It is clear that compared to PDGF,

at similar concentrations, other proteins showed no or little anisotropy increase due to the

aptamer's specificity for PDGF.










1.2

1.0











PDGF BSA LYS MYO HEM LDH THR
-0.2
0- 0.6
0


0.2

0.0
PDGF BSA LYS MYO HEM LDH THR
-0.2

Figure 2-10 Binding selectivity of the aptamer. Different extracellular proteins were
compared with PDGF-B chain (PDGF) in their capability to change the aptamer's
anisotropy. The 5 fold proteins (moles) are added into 0.02 [LM aptamer solution at 250C.

Other experiments were done to compare the affinities of the aptamer to PDGF-

BB and to other growth factors which may coexist with PDGF-BB in clinical samples. As

shown in Figure 2-11, the aptamer did not bind to epidermal growth factor (EGF), or

insulin-like growth factor-I (IGF1), indicating that the aptamer is highly selective for

PDGF-BB. The other two isoforms of PDGF, PDGF-AA and PDGF-AB, however,

showed some anisotropy increase. But their affinities are much lower than that of PDGF-

BB. The PDGF A chain and B chain have 60% similarity in their amino acid sequence

[22]. This may explain the anisotropy change caused by PDGF-AA and PAGF-AB. The

fact that the A chain is more acidic than B chain, however, may be the reason for its

lower affinity to the aptamer.










1.2

1.0 -

` 0.8

e 0.6

0.2
S0.0



-0.2

Figure 2-11 Comparing the binding capability of the aptamer to PDGF-BB and to other
growth factor such as PDGF-AA, PDGF-AB, epidermal growth factor (EGF), and
insulin-like growth factor-I (IGF 1). The molar ratio of the protein to the aptamer is 1:1.

Conclusions

In this part of our work, we have developed a molecular aptamer for real-time

detection of the oncoprotein PDGF in homogeneous solution using a fluorescence

anisotropy method. A fluorescein-labeled PDGF aptamer was designed and used as a

probe to observe its anisotropy increase upon binding to its target protein. The significant

increase in anisotropy was attributed to the large difference in molecular size between the

free fluorescein-labeled aptamer and its PDGF binding complex. This difference is

significant enough to allow the molecular binding to be quantified for protein detection in

real-time. The assay is highly selective and can detect PDGF down to 0.22 nM. The assay

is quick and can detect PDGF without separation. Anisotropy measurements are ideally

suited for measuring the binding of small aptamer probes with protein macromolecules.

This work demonstrates the potential applications of dye-labeled aptamers for

oncoprotein and disease-related protein detection in clinical studies. This assay can be

used in a noncompetitive homogeneous assay format. The same assay concept can also be






24


used for biosensors by immobilization of the aptamer onto a solid surface for in vivo or in

vitro protein monitoring [23]. Recently, the in vivo instability of aptamers in a biological

fluid containing nucleases has been circumvented by chemical modification of the bases,

particularly by substitutions at the 2' position of the sugar [24], allowing aptamers to

function adequately in biological fluids. This will ensure the application of aptamer-based

analytical methods in real biological samples.














CHAPTER 3
DEVELOPMENT OF FLUORESCENCE ANISOTROPY IMAGING SYSTEM FOR
PROTEIN ARRAYS

As databases for many sequenced genomes have been built, people begin to

realize it is time to move to the next level: to understand more about proteins and their

functions in life. Whereas genes contain the information for life, the encoded proteins and

RNAs fulfill nearly all the functions, from replication to regulation. It is well recognized

that the complexity of the human proteome far exceeds that of the genome. The number

of different molecular protein species in the human body is likely to be at least 500,000.

To be able to deal with such a large amount of proteins, it is very important to have new

techniques that can realize protein analysis with high throughput. Not long ago, the term

"proteomics" was proposed to define the large-scale study of the proteins expressed by a

genome. Although microarray technology has enabled rapid development of genomics, it

is not as easy, if not much more difficult, to apply similar techniques to proteomics. In

this part of our work, we try to develop a novel technique based on anisotropy

measurements using imaging technique, which has the potential to realize high

throughput and multiplex protein analysis.

Protein Arrays

Protein analysis has been done using a variety of techniques in small scale. They

include techniques based on probe-protein binding, such as enzyme-linked

immunosorbent assay (ELISA), and techniques based on protein separations, such as gel

electrophoresis. Those techniques have showed good sensitivity and selectivity in protein









analysis, but in order to do large-scale protein studies, they still need to be modified or

improved. A recent approach is the combination of 2-dimentional gel electrophoresis and

mass spectrometry. This enables the analysis of multiple proteins in a short period of

time. However, it also has some major drawbacks. First, it is a destructive technique,

meaning you will lose some of your protein samples. Second, it may not be able to collect

and analyze all the protein species if they are expressed at low abundance. It can not

provide reliable quantitative results. Lastly, while it is effective in separating and

isolating individual proteins, it does not yield much information about the interactions

between proteins and other biomolecules in real-time.

Protein arrays, on the other hand, have been rapidly developed to address some of

the problems faced by other techniques. Protein arrays are solid-phase ligand binding

assay systems using immobilized proteins on surfaces which include glass, membranes,

microtiter wells, mass spectrometer plates, and beads or other particles. The assays are

highly parallel and often miniaturized. Their advantages include being rapid and

automatable, capable of high sensitivity, economical on reagents, and yielding a lot of

data for a single experiment. Protein array is expected to be a very important technology

in proteomics thanks to its ability to make possible high throughput protein detections as

well as parallel multiplex screening of interactions between multiple proteins and other

biomolecules. Some work has demonstrated extraordinary power of protein chips to

analyze thousands of proteins at the same time [25]. The capture ligands used in protein

arrays are often antibodies, but may also be proteins, enzyme-substrates, receptor-ligands

and aptamers [26]. Compared to antibodies, aptamers have the advantages of ease of









production by automated oligonucleotide synthesis, robust nature of the nucleic acids,

and easy modification for expanded applications.

Even though it has a great potential and some successful applications, the protein

array technique is still facing some challenges. Traditional protein arrays, or protein

chips, when compared to DNA arrays, have some major drawbacks due to protein's

inherent properties. While DNA molecules are stable even in some harsh experimental

conditions, proteins are much more fragile and very sensitive to their surrounding

environment and will be easily damaged if not treated carefully. DNA molecules can be

readily immobilized onto several kinds of surfaces without much loss of biological

functions; proteins are much easier to be denatured when close to a solid surface, which

will result in the loss of their abilities to react with their ligands. In traditional protein

arrays where the detection system is based on fluorescent intensity or other intensity

measurements, the immobilization of the capture ligand is inevitable. People usually need

to immobilize the ligands on a chip, and then incubate the chip with the analytes. After

the target protein in the sample binds to the ligands on the chip surface, all the unbound

molecules need to be washed away. The protein on the chip will then be stained with a

stain reagent. Finally the stain generates intensity signals on the detection system, which

indicates the presence of the target protein. All these procedures are the result of the

detection methods adopted by the array system. In our work, a new detection scheme has

been developed for protein arrays based on fluorescence anisotropy which does not

require the ligand immobilization and is very simple to operate. This will also allow us to

circumvent the problems associated with denaturation of proteins, either the ligand or the

target, when immobilized to the solid surface.









Fluorescence Anisotropy Imaging

As described in Chapter 1, fluorescence anisotropy is a unique technique that has

many distinctive features. Unlike intensity based fluorescence techniques, the excitation

in fluorescence anisotropy is a plane-polarized light. Depending on the size of the

fluorescent molecules in the sample or the viscosity of the sample solution, the emission

will have different intensities at two polarization planes, one parallel to the excitation and

the other perpendicular to it. The difference is measured and calculated using following

equation:

r = (Ivv-IvH) / (Ivv+2IVH)

where the subscripts V and H refer to the orientation (vertical or horizontal) of the

polarizers for the intensity measurements, with the first subscript indicating the position

of the excitation polarizer and the second for the emission polarizer. It is important to

know that in a real-world measurement, the detection system may have different

sensitivity to polarized light at different directions. So the actual equation used for

anisotropy calculation is:

r = (Ivv-GIVH) / (Ivv+2G-IvH)

where G is the G-factor of the detection system, which is calculated as G = IHV/IHH. G-

factor represents the ratio of the sensitivities of the detection system for vertically and

horizontally polarized light. G-factor is dependent on the emission wavelength and can be

determined using a standard sample with known anisotropy.

For a fluorophore-labeled probe, after it binds to its target protein, the increase of

the overall weight of the binding complex will greatly decrease the fluorophore's ability

to rotate and result in a higher fluorescence anisotropy value. Intensity-based techniques

can not tell the difference after the binding because only the total emission of the









fluorophore is measure. However, an anisotropy system can easily tell if there is a

binding process even if the total emissions of the bound and unbound fluorophores are

about the same. Most of the current protein arrays require a long process of

immobilization, incubation, washing and staining, in order to see the binding between the

ligand and protein. By using fluorescence anisotropy, on the other hand, we can detect

the binding event as soon as the ligand and the protein are mixed together. This method

also has the potential to detect not only the DNA-protein interaction but also protein-

protein and RNA-protein interactions.

Fluorescence anisotropy measurements are mostly done in fluorometers, where

the type of the sample container is fixed and simultaneous multiple sample measurements

can not be easily implemented. The idea presented here is that if we could use a plate

with wells on it that have different samples in them and then construct an anisotropy

image of the plate, we should be able to look at samples with different anisotropies within

one image. It has been reported that an anisotropy imaging method has been developed to

study single molecules [27]. However, it was done with a very limited scale of samples,

and should not be suitable for applications like protein arrays. Other work has been done

to study protein interactions and detect proteins with high throughput using fluorescence

polarization (FP) [28]. It was based on a similar method to fluorescence anisotropy and

utilized a plate reader to measure samples one by one, instead of using one-time imaging

system to measure all samples at the same time. In this part of our work, we have

developed a system that can take anisotropy images of fluorescent samples and make it a

potentially better detection method for protein arrays.









In order to obtain anisotropy images, two polarizers need to be coupled to an

imaging system. One is the excitation polarizer that provides a polarized light source.

Another is a polarizer on the emission side. While the excitation polarizer should be fixed

in position, the emission polarizer need to placed in a way so that we can change its

position to get two polarization planes, one parallel to the polarization plane of excitation

light and the other perpendicular to it. By changing the position of the emission polarizer,

two images can be obtained, on which each pixel does not show the real emission of the

sample at that position, but rather Ivy or IVH of the total emission. An image processing

software will be used to calculate the two images at pixel level to get an anisotropy

image. The value of each pixel on this anisotropy image represents the anisotropy of the

sample at that spot.

Experimental Section

Materials

The carboxytetramethylrhodamine (TAMRA)-labeled PDGF aptamer was

customer-designed and then synthesized by Integrated DNA Technologies, Inc.

(Coralville, IA). The sequence of the aptamer is 5'- CAGGC TACGG CACGT AGAGC

ATCAC CATGA TCCTG -3'- TAMRA. A TMARA-labeled human a-thrombin aptamer,

with a sequence of 5'- TTTGG TTGGT GTGGT TGGT -3'- TAMRA, and another

TAMRA-labeled oligonucleotide with a random sequence of 5'- CGGTA GTACC

AAGTC CAGGT -3'- TAMRA as a control DNA were also synthesized by Integrated

DNA Technologies, Inc. Recombinant human PDGF-BB, PDGF-AB, and PDGF-AA

were purchased from R&D Systems (Minneapolis, MN) and they were dissolved in 4

mM HC1 and then diluted in a Tris buffer before use. Human bovine serum albumin

(BSA) and epidermal growth factor (EGF) were bought from Roche (Indianapolis, IN).









Other chemicals were from Sigma (St. Louis, MO). Buffer used in all experiments

consisted of 10 mM Tris-HCl (pH 7.5), 75 mM NaC1, and 2.5 mM KC1. Superpurified

water was used to prepare all of the solutions. All experiments were done at room

temperature.

Anisotropy Imaging Setup

A schematic diagram of the anisotropy imaging system is shown in Figure 3-1.

An intensified charge-coupled device (ICCD) (Roper Scientific, EEV 512 x 1024 FT)

was used to capture all images. A home-made cylindrical polarizer holder was placed

under the ICCD. There was a 900 slit on the outer surface of the holder around the axis of

the cylinder through which we could rotate a polarizer fixed inside the holder. Rotation of

the polarizer from on end of the slit to the other could change between two emission

polarization planes perpendicular to each other. A TV zoom lens from Edmund Industrial

Optics (Barrington, NJ) was also placed beneath the ICCD to help look at a relatively

large area. A mercury lamp by Olympus America Inc. (Melville, NY) was used as the

light source. The light coming out of the lamp was guided by a fiber bundle to the sample

plate. A polarizer was placed just before the outlet of the fiber bundle as the excitation

polarizer. All polarizers were from Edmund Industrial Optics. A 520-550 nm bandpass

filter from Olympus America Inc. was used for excitation. A 590 nm longpass filter (from

Olympus America Inc.) and a 600 nm shortpass filter from Oriel Instruments (Stratford,

CT) were used for emission.

The sample plate was made of black plastic (DELRIN acetal resin) with

dimensions of 20 mmx20 mmx2 mm. A 6x6 array of small wells was mechanically made

on the plate. Each well is 1.5 mm in diameter and 0.75 mm deep, and it holds about 1.5

[tL of liquid sample.










A Fluorolog-Tau-3 spectrofluorometer (Jobin Yvon, Inc., Edison, NJ) was also

used to perform some fluorescence anisotropy measurements. In those cases, a 50 PL

quartz cuvette was used to hold the sample solution. Excitation and emission wavelength

were set to be 545 nm and 595 nm respectively for TAMRA. Bandpasses for both the

excitation and emission were set to 10 nm.





ICCD






Mercury lamp Emission
polarizer holder

Fiber bundle
STV zoom lens


Filters and
polarizer

S/ Sampplate



Figure 3-1 Schematic diagram of the anisotropy imaging system. An ICCD with a zoom
lens is the detector. A fiber bundle coupled to a mercury lamp provides the excitation
light.

Acquisitions of Anisotropy Images

Alignment of polarizers is done as following. A sample of very high anisotropy,

such as a TAMRA dye solution with 80% (V/V) glycerol is used. First, the emission

polarizer is fixed at one end of the 900 slit on the polarizer holder. Then the position of

the excitation polarizer is adjusted until the lowest fluorescent signal is observed. The

current position of the emission polarization should be perpendicular to that of the









excitation polarization, which will only allow the IVH signal of the sample to reach the

detector. With rotation of the emission polarizer for 900, the Ivy component of the

emission will then be collected by the detector. Same approach was also reported by

other group [29]. By using ICCD as the detector, two images can be obtained, with one

containing the Ivy part of the emission and the other containing the IVH part. We may call

them Image Ivy and Image IVH, or IIvv and IIVH respectively.

1.5 [tL of sample solution is added to the well on the sample plate. With the

excitation on, we obtain two images of the sample, IIvv and IIVH, by rotating the emission

polarizer. The exposure time is the same for the two polarization images of the same

sample, but may vary from 0.05 sec to 3 sec from sample to sample depending on the

sample intensity. The image capturing software is Winview (Roper Scientific).

Data processing is done using a Java-based computer program called ImageJ

(http://rsb.info.nih.gov/ij/index.html). It was developed by the Research Services Branch

at National Institute of Mental Health and National Institute of Neurological Disorders

and Stroke. It has many image processing functions and allows you to develop plugins

for your own image processing needs. We developed two plugins for the anisotropy

imaging experiments. One plugin is used to clear the background area of the image IIvv

and IIVH to zero. This is needed because the background noise on the two images is low

and highly fluctuating. If it were not set to zero, on the final anisotropy image, the

background would show very random anisotropies, from very low to very high. The real

anisotropy signal from the wells would be buried in this background. Another plugin we

developed is used to directly calculate an anisotropy image Ir with image IIvv and IIVH

and a user-specified G-factor using the following equation:









= (Ivv-GIIVH) / (Ilvv+2G'IIVH)

In the anisotropy image, the value of each pixel represents the fluorescence

anisotropy of the sample at that position. To get quantitative results from the anisotropy

image, the value of the pixels within the area of one sample is averaged and the average

value is used to represent the actual anisotropy of the sample. The final anisotropy image

might be rescaled or cut off of some trivial areas that are not related to the real signal for

the sake of better presentation of data.

In order to get reliable anisotropy values from our anisotropy imaging system, it is

important to know the G-factor of the system. However, it is not easy to measure the G-

factor directly using the anisotropy imaging system. We solved this problem by

correlating the anisotropy values measured by the imaging system to the anisotropies we

get from the same sample using a spectrofluorometer that has the anisotropy

measurement capability. This spectrofluorometer is able to determine the G-factor by

measuring IHV and IHH, and thus can get relatively accurate anisotropy values for a

fluorescent sample. We set the excitation and emission wavelengths on the

spectrofluorometer to be similar to those on the anisotropy imaging system so that the

two systems should yield similar anisotropy values for the same sample. We then used

the spectrofluorometer to determine the fluorescence anisotropies of the free protein-

binding ligand solutions. These ligand solutions with known anisotropies were used as

references in all experiments conducted on the anisotropy imaging system. G-factors of

the system were chosen to make sure the calculated anisotropies for the reference

solutions were the same as what we obtained from the spectrofluorometer. In this way,









the results obtained from the anisotropy imaging system should be parallel to those from

the spectrofluorometer.

Results and Discussions

Experimental Considerations

The carboxytetramethylrhodamine (TAMRA) dye was chosen to be the

fluorescent label for all the ligands for different proteins. TAMRA is much more stable

than some of the other dyes such as fluorescein, which means we will have less severe

photobleaching problem. This enables us to use relatively longer exposure time to obtain

higher signals. Another advantage of TAMRA is that it can be readily excited by the

intense 546 nm spectral line from mercury-arc microscope lamps [30] and often shows

brighter intensity than fluorescein.

We use two TAMRA-labeled aptamers as the ligands for two proteins. One of

them is the PDGF-aptamer that binds selectively to PDGF-BB protein. It has a 35 base

sequence and has a three-way helix junction structure as shown in Figure 2-2. Another

aptamer is the aptamer for human a-thrombin. It has a 15 base sequence and is believed

to fold into a chair-form quadruplex with the 5' and 3' ends in the corners of the

quadruplex and two stacked G-quartets linked by TT and TGT loops [31,32]. The

aptamer was found to bind selectively to human a-thrombin with high affinity [33].

Aptamers are ideal for protein binding assay because of its high selectivity and affinity to

their target protein. Compared to antibodies, aptamers have the advantages of low cost,

easy handling and stable structure. We here use aptamers as the binding ligands for

proteins and use fluorescence anisotropy imaging as the detection method since the

binding of the aptamer to target protein results in a larger molecular weight and greater

anisotropy.









Anisotropy Imaging of TAMRA-Glycerol Solutions

To test the ability of our system to detect anisotropies of multiple samples

simultaneously, we used TAMRA dye solutions with different glycerol concentrations

(V/V). Viscosity of glycerol is much higher than that of pure water. With glycerol

concentration increases, the viscosity of the solutions also increases. As the rotational

diffusion of the TAMRA dye molecules is restricted by the higher viscosity of the

solution, the measured anisotropies of the dye solutions should show an increase. Figure

3-2 shows an anisotropy image of eight TAMRA solutions in eight wells on the sample

plate with glycerol concentration ranging from 10% to 80%.

We can see clearly that the difference in anisotropy between two solutions can be

seen directly from the image. By comparing the colors of two samples on the same

anisotropy image, we can tell the anisotropy difference and the information associated

with the difference, such as differences in molecular size and solution viscosity. In the

case of TAMRA-glycerol solutions, the anisotropy difference is caused by the difference

in the viscosity of the solutions.

On the anisotropy image, pixels within one sample showed a distribution of

anisotropy values. It is most likely due to signal fluctuation of each small photosensor on

the ICCD during the data acquisition process. However, this distribution is not totally

random but rather centered around a certain anisotropy value. To get the quantitative

results from the anisotropy image, we averaged the anisotropy values of each pixel within

one sample to get the overall anisotropy of the sample. The results are shown in Figure 3-

3.
















0.3080% 10%

70% 20%

60% 30%

0.001 50% 40%


0.00

Figure 3-2 Anisotropy image of 200 nM TAMRA solutions with different glycerol
concentrations. The table shows concentrations of glycerol (V/V) and the relative
positions of the eight solutions. Exposure time was 0.05 sec.


0.25


0.20

0.15
0

0.10

0.05


0.00
0 20 40 60 80 100
Cglycerol (V/V)%

Figure 3-3 Anisotropy vs. glycerol concentration (V/V) in 200 nM TAMRA solutions.
Data was averaged from four anisotropy images.

The anisotropy values obtained from the anisotropy imaging system showed quite

small error bars, partially due to the high fluorescence of the TAMRA solutions. The

anisotropy increases as the glycerol concentration increases. Similar curves were also

obtained as a confirmation from the spectrofluorometer, which was supposed to give

rather accurate anisotropy measurements. It is worth noting that from the quantitative









curve and the anisotropy image, an anisotropy difference of as little as 0.015 can be seen

with ease from the anisotropy image by just comparing colors. The imaging system

provides a qualitatively very straightforward and quantitatively rather accurate way to

examine multiple samples simultaneously.

Anisotropy Imaging for Aptamer-Protein Binding Assay

In our previously work, we used the spectrofluorometer to do all the anisotropy

measurements for PDGF and aptamer binding study. Sample solutions of at least 50 tL

were used in order to fit in a cuvette or to get enough signals. With the anisotropy

imaging system's ability to conduct multiple anisotropy measurements with very small

amount of samples, we should be able to place more than one solution with different

aptamer/PDGF ratios in the plate wells and look at the extents of the binding reaction

reflected by the anisotropy values in all wells at the same time. This will be quite useful

and convenient when it comes to studying protein binding process under different

conditions. We carried out some experiments where the anisotropies of a series of

solutions with the same aptamer concentration but different PDGF concentrations were

determined at one time using anisotropy imaging. An anisotropy image is shown in

Figure 3-4. A binding curve was also constructed based on the image, as shown in Figure

3-5.












0.22


0 300

20 200

50 150

70 100


0.00


Figure 3-4 Anisotropy image of eight 50 nM PDGF-aptamer solutions with different
PDGF-BB concentrations. The table on the right shows the concentration of the sample in
nM at the corresponding position on the anisotropy image.


0.20


0.18


g 0.16


S0.14


0.12


0.10


0 50 100 150 200
CPDGF-BB (M)


250 300 350


Figure 3-5 Anisotropy vs. PDGF-BB concentration in 50 nM TAMRA-labeled PDGF-
aptamer solutions. Data was averaged from 5 anisotropy images. Exposure time was 3
sec.

Compare the shape of the curve in Figure 3-5 to the curve we obtained previously

using the spectrofluorometer (Figure 2-8), we can see that in the PDGF-BB concentration









range of from 0 to 6x CPDGF-aptamer (0 to 0.6 [LM in Figure 2-8 and 0 to 300 nM in Figure

3-5), these two curves are very much alike. We did not go to higher PDGF concentration

here since we just wanted to demonstrate the capability of the anisotropy imaging system

in protein binding assay. This curve was also repeatable on the spectrofluorometer.

Compared to the anisotropy image we got from the TAMRA-glycerol solutions,

the anisotropy image for aptamer/PDGF binding solutions showed less distinctive colors

between samples. This is probably because at a much lower fluorescence intensity level

and with a much longer exposure time, the signal fluctuation of the pixels on the ICCD

became more severe. Photobleaching of the TAMRA dye at low concentration might also

have played a role in this problem. The result of this problem is a much larger anisotropy

distribution within one sample on the anisotropy image. This is very much like broadened

peaks in chromatography and there is color overlapping between two adjacent samples on

the anisotropy image. This is why we see similar colors between samples. However, the

color difference is still discernable and the averaged anisotropy values of the samples still

showed significant change as PDGF-BB concentration increased as shown in Figure 3-5.

Similar reasons can be used to explain the larger error bars in Figure 3-5 than in Figure 3-

3.

In a system where we use multiple sample wells and compare signals from

different wells for analyte detection, the well-to-well signal variation should be the

indicator of the noise of this assay system [28]. The well-to-well anisotropy variation of

our system was tested to be <0.002 by measuring the anisotropy of the same sample in

multiple wells. To qualify for a real signal, the anisotropy should be 3 times of this

deviation, which accounts for an anisotropy difference of 0.006. Applying this to the










curve in Figure 3-5, we determined the detection limit of our anisotropy imaging system

for PDGF-BB to be around 13 nM. Since we are detecting a sample of only 1.5 aL in

each well, the mole detection limit should be 13 nM x 1.5 aL z 2.0x10-14 mole =20 fmole

of PDGF-BB. This sensitivity can still be improved by optimization of the optical path,

changing to a near IR dye with higher fluorescence and less photobleaching, and using

high power laser as the light source.

Selectivity of Aptamer/PDGF Binding on the Anisotropy Imaging System

We carried out a series of experiments to demonstrate that the selectivity of the

aptamer/PDGF-BB binding can be preserved on the anisotropy system. First, 50 nM

PDGF-aptamer solutions with 300 nM of different proteins were added to six wells. The

resulting anisotropy image is shown in Figure 3-6.


0.22


0.00 -
Figure 3-6 Anisotropy image of 50 nM PDGF-aptamer solutions with 300 nM different
proteins [epidermal growth factor (EGF), bovine serum albumin (BSA), PDGF-AA,
PDGF-BB and PDGF-AB]. The table on the right shows the relative positions of these
aptamer/proteins solutions on the anisotropy image and their anisotropy values based on
the image.

The relative anisotropy changes over pure aptamer solution caused by other

proteins compared to PDGF-BB are shown in Figure 3-7. We can see that on the


No protein PDGF-AB
0.105 0.139

EGF PDGF-AA
0.109 0.128

BSA PDGF-BB
0.099 0.187


'''




r,


~:










anisotropy imaging system, the PDGF-aptamer still showed very good selectivity for

PDGF. At similar concentrations as PDGF-BB, EGF and BSA showed very little or no

anisotropy increase while PDGF-AA and PDGF-AB showed some increase which is

agreeable with previous result in Figure 2-11.


1.2

1.0 -

0.8

S0.6
0
S0.4

0.2 -

0.0 -
PDGF-BB PDGF-AA PDGF-AB EGF
-0.2

Figure 3-7 Comparing the binding capability of the PDGF-aptamer to PDGF-BB and to
other proteins such as PDGF-AA, PDGF-AB, epidermal growth factor (EGF), and bovine
serum albumin (BSA). The aptamer and protein concentrations are 50 nM and 300 nM
respectively.

Another experiment was designed to show that only PDGF-aptamer binds to

PDGF-BB while other dye-labeled oligonucleotides have little affinity to PDGF-BB. An

anisotropy image is shown in Figure 3-8. We used a TAMRA-labeled control DNA with

a random sequence and the TAMRA-labeled human a-thrombin aptamer as ligands to

bind to the same concentration of PDGF-BB. As shown clearly in Figure 3-8, no

anisotropy increase was observed for control DNA/PDGF-BB or thrombin-

aptamer/PDGF-BB. Only PDGF-aptamer showed significant anisotropy increase with the

addition of PDGF-BB. This shows that aptamers are ideal binding ligands for protein









binding assay. By combining aptamers with fluorescence anisotropy techniques, we

should be able to provide an easy and effective way for proteomics studies.



0.15


0.00w
Figure 3-8 Anisotropy image for comparison of binding ability of PDGF-BB to different
TAMRA-labeled oligonucleotides [a control DNA sequence (ctrDNA), the human a-
thrombin aptamer (tAPT) and PDGF-aptamer (pAPT)]. The first column is 50 nM pure
oligonucleotide solutions. The second column is 50 nM oligonucleotides plus 50 nM
PDGF-BB. The table on the right shows the samples' relative locations on the anisotropy
image and their anisotropy values.

Detection of Protein Mixture

In protein array or protein chip techniques, it is important to be able to not only

detect proteins with high throughput but also detect multiple proteins in one sample

simultaneously. In order to demonstrate that the anisotropy imaging system can easily

realize multiplex protein detection, we conducted a simplified example of the detection of

a protein mixture. The anisotropy image of this experiment is shown in Figure 3-9.


ctrDNA ctrDNA+PDGF-BB
0.098 0.096

tAPT tAPT+PDGF-BB
0.088 0.090

pAPT pAPT+PDGF-BB
0.099 0.122











0.17


0.00


Figure 3-9 Anisotropy image of 50 nM different TAMRA-labeled oligonucleotides [a
control DNA sequence (ctrDNA), the human a-thrombin aptamer (tAPT) and PDGF-
aptamer (pAPT)] with different proteins or protein mixture (50 nM human a-thrombin,
100 nM PDGF-BB and the mixture of these two). The table shows all samples' relative
locations on the anisotropy image and their anisotropy values.

In this experiment, human a-thrombin, PDGF-BB and their mixture were added

separately to three TAMRA-labeled oligonucleotide sequences (a control DNA sequence,

the thrombin-aptamer and the PDGF-aptamer) and resulting anisotropy image was

recorded. It is clear from Figure 3-9 that the control DNA showed little anisotropy


ctrDNA pAPT tAPT
0.096 0.109 0.057

ctrDNA+thrombin pAPT+ thrombin tAPT+ thrombin
0.090 0.107 0.088

ctrDNA+PDGF- pAPT+PDGF-BB tAPT+PDGF-BB
BB 0.142 0.052

ctrDNA+mixture pAPT+mixture tAPT+mixture
0.092 0.129 0.081









increase when mixed with any of the proteins, indicating there was minimal interaction

between the control DNA and the proteins. The thrombin-aptamer showed anisotropy

increase with addition of thrombin or thrombin/PDGF-BB mixture, but no increase with

addition of PDGF alone, meaning that it was a-thrombin that caused the anisotropy

increase. Similar results were observed for PDGF-aptamer, which showed anisotropy

increase only with PDGF-BB and thrombin/PDGF-BB mixture. This experiment

demonstrated that such an array of aptamers can be used to accurately detect proteins in a

mixture sample when the concentrations of the proteins are within comparable ranges. A

sample that causes the thrombin-aptamer an anisotropy increase, but not other DNA

sequences, would probably contain a-thrombin but not other proteins. Similar things can

be stated for PDGF-aptamer and PDGF-BB. On the other hand, a sample that increases

the anisotropies of both thrombin-aptamer and PDGF-aptamer, would probably contain a

mixture of both proteins. This is the simplest example of detecting protein mixtures. By

using more aptamers or antibodies targeted for more proteins of interest and making a

larger array, we can use anisotropy imaging technique to easily detect more proteins in

one sample.

Conclusions

We have developed a novel technique that has the potential to be used to build a

very simple but effective protein array. It uses an imaging system to measure

fluorescence anisotropy in a 2-dimensional format, which enables anisotropy

measurements of multiple samples in an array format. Due to fluorescence anisotropy

technique's unique ability of effectively detecting protein interactions with other

molecules, anisotropy imaging technique has the possibility of making a protein chip

much simpler than conventional protein chips. It does not require tedious procedures that









are necessary in most of the current protein chips, such as immobilization, washing and

protein staining. At the same time, it has the same ability to realize high throughput and

multiplex protein detection which is essential in a wide range of applications in

immunodiagnostics, protein function and interaction screening, and drug discovery.

Although we used aptamers as the ligands for proteins, the application of

anisotropy imaging is not limited by aptamers. Basically, anything that exhibits a

significant molecular weight increase after binding to proteins can be used as a protein

ligand in anisotropy imaging, which includes antibody, antigen, DNA, RNA, and protein.

This ensures the versatility of anisotropy imaging in a variety of applications.














CHAPTER 4
SUMMARY AND FUTURE WORK

Summary

We have demonstrated the possibility of detecting platelet-derived growth factor

(PDGF) using aptamer based fluorescence anisotropy technique. This method has been

proved to be highly sensitive, highly selective and very simple. We then extended this

method to a 2-dimentional array format for possible high throughput and multiplex

protein analysis by developing a fluorescence anisotropy imaging system. This system

allowed us to examine multiple aptamer/protein binding samples simultaneously. Very

small amount of samples were needed and all measurements could be done in seconds.

We also used this system to demonstrate detection of protein mixtures. We believe this

anisotropy imaging technique has the potential to help build very simple yet highly

efficient protein arrays for various applications.

Future Work

Improving and Refining the Techniques

The fluorescence anisotropy techniques we used in our work can be improved in

many aspects. First, we could use a better fluorescent dye such as the Cy5 dye for all our

experiments. It has much longer excitation and emission wavelengths than TAMRA.

Longer excitation wavelength will cause less photobleaching and sample

autofluorescence problem. It also dramatically reduces light scattering in biological

samples as light scattering is inversely proportional to the fourth power of wavelength.

Since light scattering is the major source of error, this will greatly improve the accuracy









and sensitivity of our anisotropy measurements. We can also use a high power ion laser

as the light source for those longer wavelength dyes. Since the power of ion lasers is

much higher and more stable than that of the lamp we currently use, a much better

sensitivity is expected.

For the anisotropy imaging system, we may need to optimize it for better

sensitivity and easier handling. Currently in our system, the fluorescence signals from the

sample plate have to travel a quite long distance before reaching the ICCD detector. This

contributes to the relatively low optical efficiency and sensitivity of the system. We

should be able to improve the sensitivity by using a way to image large area with a much

shorter optical path. We may also improve our image processing software to combine

background subtraction and image calculation into one simple step, instead of two steps

we currently use. In order to be used in real world high throughput protein analysis, the

automated sample handling capability and a more miniaturized sample plate should be

included in our system.

Applications of the Anisotropy Imaging System in Cancer Diagnosis

A cancer cell is a cell that grows out of control. Since growth factors are believed

to play a role in intercellular communication and regulating cell growth and division,

people are interested in studying the relationship between cancer and growth factors.

However, study showed that cancer is often related to the level of not just one growth

factor, but rather levels of several growth factors and other related proteins in human

body [34]. The ability of our anisotropy imaging system to detect multiple proteins in a

mixture may make it useful as a simple and quick cancer diagnostic method. Body fluid

samples of patients with no cancer, early stage cancer and late stage cancer can be added

to different probe solutions targeted to different cancer related proteins. Levels of those









proteins can then be determined using the anisotropy imaging system. A database can be

built in a short period of time for study of the specific cancer.

Anisotropy Imaging Technique for Real-Time Cell Imaging

It will be of great interest if one can obtain anisotropy images of fluorescent

samples in real-time. Some optical devices have been developed that are able to divide

the emission from the sample into two images of different polarization state using a

polarization beam splitter. They can cast the two images side by side on a CCD detector.

By using new software that allows real-time image calculation within one image, we will

be able to construct anisotropy images in real time. This technique may be very useful in

monitoring binding processes inside or outside living cells. For example, real-time

monitoring of PDGF secretion from a cancer cell may be very difficult for other

techniques. Using real-time anisotropy imaging technique, we may have dye-labeled

PDGF aptamer in the cell culture medium and monitor any anisotropy change that

indicates the release of PDGF from the cell. Further more, we might also be able to

monitor PDGF transportation between cells. This might help improve our understanding

of intercellular signaling. We believe that the real significant applications of the

fluorescence anisotropy imaging technique may be in the field of real-time monitoring of

biological processes at cell level. This is a field that is worth more attention and

investigation. The experience we gained from our anisotropy imaging experiments might

help us a lot in building a system for real-time cell imaging.
















LIST OF REFERENCES

[1] AnandApte, B.; Zetter, B. Stem Cells 1997, 15 (4): 259-267.

[2] Raines, E. W.; Bowen-Pope, D. D.; Ross, R. Handbook in Experimental
Pharmacology; Sporn, M. B., Roberts, A. B., Eds.; Springer: Heidelberg, 1990, pp
173-262.

[3] Blaskovich, M. A.; Lin, Q.; Delarue, F. L.; Sun, J.; Park, H. S.; Coppola, D.;
Hamilton, A. D.; Sebti, A. M. Nat. Biotechnol. 2000, 18 (10), 1065-1070.

[4] Leitzel, K.; Bryce, W.; Tomita, J.; Manderino, G.; Tribby, I.; Thomason, A.;
Billingsley, M.; Podczaski, E.; Harvey, H.; Bartholomew, M.; Lipton A Cancer Res.
1991, 51 (16), 4149-4154.

[5] Singh, J. P.; Chaikin, M. A.; Stiles, C. D. J. CellBiol. 1982, 95, 667-671.

[6] Bowen-Pope, D. F.; Malpass, T. W.; Foster, D. M.; Ross, R. Blood 1984, 64, 458-
469.

[7] Heldin, C.-H. EAMBO J. 1992, 11, 4251-4259.

[8] Pillai, R. Eur. J. Surg. Oncol. 1992, 18, 417-424.

[9] Sumedha, D.; Jayasena, Clin. Chem. 1999, 45 (9), 1628-1650.

[10] Andreola, M. L.; Pileur, F.; Calmels, C.; Ventura, M.; Tarrago-Litvak, L.; Toulme, J.
J.; Litvak, S. Biochemistry 2001, 40 (34), 10087-10094.

[11] Stojanovic, M. N.; de Prada, P, Landry, D. W. J. Am. Chem. Soc. 2001, 123 (21),
4928-4931.

[12] Gold, L. J. Biol. Chem. 1995, 270, 13581-13584.

[13] Xu, W.; Ellington, A. D. Proc. Natl. Acad. Sci. U. S. A. 1996, 93, 7475-7480.

[14] Potyrailo, R. A.; Conrad, R. C.; Ellington, A. D.; Hieftje, G. M. Anal. Chem. 1998,
70, 3419-3425.

[15] Li, J. J.; Fang, X.; Tan, W. Biochem. Biophys. Res. Commun. 2002, 292 (1), 31-40.









[16] Kensch, O.; Connolly, B. A.; Steinhoff, H. J.; McGregor, A.; Goody, R. S.; Restle,
T. J. Biol. Chem. 2000, 275 (24), 18271-18278.

[17] Green, L. S.; Jellinek, D.; Jenison, R.; Ostman, A.; Heldin, C. H.; Janjic, N.
Biochemistry 1996, 35, 14413-14424.

[18] Lakowicz, J. R. Principles ofFluorescence Spectroscopy, 2nd ed.; Kluwer
Academic/Plenum Publishers: New York, 1999.

[19] Hey, T.; Lipps, G.; Krauss, G. Biochemistry 2001, 40 (9), 2901-2910.

[20] Heyduk, T.; Ma, Y.; Tang, H.; Ebright, R. H. Methods Enzymol. 1996, 274, 492-503.

[21] Takahashi, M.; Sakumi, K.; Sekiguchi, M. Biochemistry 1990, 29 (14), 3431-3436.

[22] Betsholtz, C.; Johnsson, A.; Heldin, C.-H.; Westermark, B.; Lind, P.; Urdea, M. S.;
Eddy, R.; Shows, T. B.; Philpott, K.; Mellor, A. L.; Knott, T. J.; Scott, J. Nature
1986, 320, 695-699.

[23] Cordek, J.; Wang, X.; Tan, W. Anal. Chem. 1999, 71, 1529-1533.

[24] Floege, J.; Ostendorf, T.; Janssen, U.; Burg, M.; Radeke, H. H.; Vargeese, C.; Gill,
S. C. Am. J. Pathol. 1999, 154, 169-179.

[25] Zhu, H.; Bilgin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.; Lan, N.;
Jansen, R.; Bidlingmaier S.; Houfek T.; Mitchell T.; Miller P.; Dean R. A.; Gerstein
M.; Snyder M. Science 2001, 293 (5537), 2101-2105.

[26] Templin, M. F.; Stoll, D.; Schrenk, M.; Traub, P. C.; Vohringer, C. F.; Joos, T. O.
Trends Biotechnol. 2002, 20 (4), 160-166.

[27] Harms, G. S.; Sonnleitner, M.; Schutz, G. J.; Gruber, H. J.; Schmidt, T. Biophys. J.
1999, 77 (5), 2864-2870.

[28] Wu, P.; Brasseur, M.; Schindler, U. Anal. Biochem. 1997, 249, 29-36.

[29] Gough, A. H.; Taylor, D. L. J. CellBiol. 1993, 121 (5), 1095-1107.

[30] Haugland, R.P. Handbook ofFluorescent Probes and Research Products 9th edition,
Molecular Probes, Eugene, OR, USA, 2002.

[31] Macaya, R. F.; Schultze, P.; Smith, F. W.; Roe, J. A.; Feigon, J. Proc. Natl. Acad
Sci. U. S. A. 1993, 90 (8), 3745-3749.

[32] Wang, K. Y.; Mccurdy, S.; Shea, R. G.; Swaminathan, S.; Bolton, P. H.
Biochemistry 1993, 32 (8), 1899-1904.






52



[33] Bock, L. C.; Griffin, L. C.; Latham, J. A.; Vermaas, E. H.; Toole, J. J. Nature 1992,
355 (6360), 564-566.

[34] Harman, S. M.; Metter, E. J.; Blackman, M. R.; Landis, P. K.; Carter, H. B. J. Clin.
Endocrinol. Metab. 2000, 85 (11), 4258-4265.















BIOGRAPHICAL SKETCH

Zehui Cao is from Zhuzhou, Hunan Province, China. He spent 18 years in that

same province before he traveled almost 1000 miles to Nanjing, a city in eastern China,

for college study. He got his B.S. degree in chemistry from Nanjing University in 1998.

He spent one more year at the graduate school of Nanjing University studying organic

chemistry before he finally realized he was more attracted to control panels than to flasks

and beakers. So he traveled again over 10000 miles to University of Florida to study

analytical chemistry. He likes what he is doing and would like to pursue a Ph.D. degree

after graduation.