|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 DNA APTAMERS HAVE AN ANALOG IN GENOMIC DNA By A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011
3 To Mom, daD, Carrie, Peter and the two great loves of my life: Josh and DNA
4 ACKNOWLEDGMENTS I would like to thank the scientists who prodded me over the years to think harder, dig deeper, and then reveled with me in the ecstasy of discovery, especially when the en, Lin Wang, Zeyu Xiao, Youngmi Kim, Prabodhika Mallikarachy, Maria Carmen Estevez, Karen Milians, Jennifer Martin, Zhi Zhu, Yanrong Wu, Ling Meng, Hui Wang, Suwussa Bamrungsap Jian Wang, and Dalia Lopez Colon; you taught me with style and grace how to w ork hard, cut loose, and earn a Ph.D. I am also grateful to my mentors in the scientific life, who have shaped my professional growth, and guided me down my path, including, David Ostrov, Richard Condit, Nancy Denslow, Bradley Fletcher, Joshua Socolar, Geo rgia Dadoly, Gregory Schultz, Chen Liu, Katherine Williams, and Weihong Tan. Much gratitude also goes to Kwame Sefah, Joseph Phillips, Dihua Shangguan, Haipeng Liu, Parag Parekh, Ibrahim Shukoor, and Yunpeng Cai who shared with me their expertise and scien tific rigor. They got me asking the right questions and provided me with the tools to get the answers I needed. Thanks to Xiaohong Fang, and her lab in Beijing, especially Xiaoli Shi and Zilong Zhao, for showing me immense hospitality, and helping me succe ed during the Olympic summer of 2008. Thanks to Stephen Larner for starting me in research and Josie Cooke for keeping me focused at the end of it. Finally, I would like to thank my dear friends who will remain, for now, in Florida, especially Weijun Chen, Xiangling Xiong, Guizhi Zhu, Tao Chen, Dimitri van Simaeys, Elizabeth Jimenez, Yunmin Chang, Sena Cansiz, Diane Turek, Jen Maier, Catherine Buchanan McGrath, and Jeff Boissoneault : may all of your pipettes be calibrated, and may all of your experiments wo rk the first time.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREV IATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 17 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 21 Aptamer Selection ................................ ................................ ................................ .. 23 The PTK7 Aptamer Sgc8c ................................ ................................ ...................... 25 Biology of PTK7 ................................ ................................ ................................ ...... 28 Wnt Signaling Pathway ................................ ................................ ..................... 28 Catenin/Wnt Pathway ................................ ................................ ................... 30 PCP/Wn t Pathway ................................ ................................ ............................ 31 PTK7 Structure ................................ ................................ ................................ 34 Pseudo Tyrosine Kinase Domain of PTK7 ................................ ....................... 35 Transmembrane Domain of PTK7 ................................ ................................ .... 36 Extracellular PTK7 and its MT MMP1 Cleavage Site ................................ ....... 37 PTK7 Gene Structure and Splicing ................................ ................................ ... 39 sgc8c Aptamer ................................ ................................ ................................ 40 DIXDC1 Biology ................................ ................................ ................................ ...... 40 DIXDC1 Increas catenin Signaling ................................ .................. 41 DIXDC1 Decreases Wnt/PCP Signaling ................................ ........................... 42 DIXDC1 Gene Structure ................................ ................................ ................... 43 2 PTK7 APTAMERS HAVE AN ANALOG IN GENOMIC DNA ................................ .. 46 Introduction ................................ ................................ ................................ ............. 46 Results and Discussion ................................ ................................ ........................... 47 38nt sgc8c Binds as w ell as 41nt sgc8c ................................ ........................... 47 Bioinformatic Analysis of cell SELEX Aptamers ................................ ............... 48 Statistical Analysis of Consensus Region ................................ ......................... 53 Consensus Region is Important to Aptamer Binding ................................ ........ 53 Mutational analysis ................................ ................................ .................... 54 Consensu s sequence blocking ................................ ................................ .. 55 Sgc8c pyrene mutants suggest sgc8c interacts intimately with PTK7 ........ 57 BLAST of the Consensus Sequence against the Human Genome .................. 60
6 DIXDC1b has the Consensus Sequence ................................ .......................... 60 KC2D4 has Identity with DIXDC1b Negative Strand ................................ ........ 63 Conclusions ................................ ................................ ................................ ............ 66 Materials and Methods ................................ ................................ ............................ 67 DNA Sequences ................................ ................................ ............................... 67 Cell Culture ................................ ................................ ................................ ....... 68 Bioinformatics ................................ ................................ ................................ .. 70 Significance Simulations ................................ ................................ ................. 70 Competition and Binding Experiments ................................ ............................. 71 Off Rate Experiments ................................ ................................ ....................... 72 Kd Measurements ................................ ................................ ............................ 72 Azobenzene Mutation Measurements ................................ .............................. 72 cDNA Blocking ................................ ................................ ................................ 73 Pyrene sgc8c B inding ................................ ................................ ...................... 73 3 POSSIBLE FUNCTIONAL ROLE FOR PTK7 DIXDC 1B INTERACTION ............... 74 Introduction ................................ ................................ ................................ ............. 74 Results and Discussion ................................ ................................ ........................... 75 PTK7 A ...................... 75 The Consensus Region is Conserved among Species ................................ .... 76 DNA is Double Stranded, but our Aptamers are Single Stranded .................... 77 PTK7 Structural Analysis ................................ ................................ .................. 79 PTK7 Has a Weak Predicted Nuclear Localization Sequence .......................... 82 Confocal Microscopy ................................ ................................ ........................ 83 Summary of PTK7 Known and Predicted Features ................................ .......... 84 Cleaved PTK7 Fragment is Higher in Cellular Nuclear Fraction ....................... 84 Sgc8c Prevents Wound Healing in HeLa Scratch Assay ................................ .. 85 Conclusions ................................ ................................ ................................ ............ 86 Materials and Methods ................................ ................................ ............................ 90 DNA Se quences ................................ ................................ ............................... 90 Cell Culture ................................ ................................ ................................ ....... 91 dsDNA Binding Experiments ................................ ................................ ............ 91 Western Blot ................................ ................................ ................................ ..... 92 Scratch Assay ................................ ................................ ................................ .. 92 4 ADDITIONAL SEQUENCE I DEN TITIES FOUND BETWEEN APTAMERS BY CELL SELEX BIOINFORMATIC A NALYSIS ................................ .......................... 93 Introduction ................................ ................................ ................................ ............. 93 Vaccinia Aptamers ................................ ................................ ................................ .. 93 BLAST of the Consensus Sequence agains t the Human Genome ......................... 95 Conclusions ................................ ................................ ................................ ............ 96 5 SGC8C APTAMER AND PTK7 ANTIBODY RUPTURE FORCES ARE COMPARABLE ON LIVE HELA CELLS ................................ ................................ 98
7 Introduction ................................ ................................ ................................ ............. 98 Results and Discussion ................................ ................................ ........................... 99 Blocking Controls ................................ ................................ ........................... 100 Fitting th e Collected Data ................................ ................................ ............... 101 Conclusions ................................ ................................ ................................ .......... 101 Materials and Methods ................................ ................................ .......................... 103 Preparation of AFM Tips and Substrate ................................ ....................... 103 Cell Culture ................................ ................................ ................................ ..... 104 AFM Measurements ................................ ................................ ...................... 104 6 MODIFYING CELLULAR PROPERTIES USING ARTIFICIAL LIPID APTAMER RECEPTORS ................................ ................................ ................................ ....... 106 Introduction ................................ ................................ ................................ ........... 106 Results and Di scussion ................................ ................................ ......................... 108 Streptavidin Aptamer Receptor Anchors on the Cell Membrane .................... 108 Streptavidin on Modified Cells Remains Functional ................................ ....... 112 SA ARs Can Be Used to Capture Cells for Analysis ................................ ...... 114 Thrombin Aptamer Receptor Captures Thrombin ................................ .......... 114 Thrombin Modified Cells Cause Clotting ................................ ........................ 116 Conclusion ................................ ................................ ................................ ............ 117 Materials and Methods ................................ ................................ .......................... 117 DNA Sy nthesis ................................ ................................ ............................... 117 Cell Culture ................................ ................................ ................................ ..... 118 Flow Cytometry ................................ ................................ .............................. 119 Fluorescence Microscopy ................................ ................................ ............... 119 MTS Assay ................................ ................................ ................................ ..... 120 Cell Capture with Streptavidin DynaBeads ................................ ..................... 120 Clotting Assay ................................ ................................ ................................ 121 7 FUTURE DIRECTIONS AND CONCLUSIONS ................................ .................... 123 Future Directions ................................ ................................ ................................ .. 123 Conclusions ................................ ................................ ................................ .......... 124 APPENDIX A COMPLEX TARGET SELEX ssDNA APTAMER DATABASE .............................. 128 B LIST OF SIMILARITY BETWEEN APTAMERS IN DATABASE ........................... 135 LIST OF REFERENCES ................................ ................................ ............................. 138 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 150
8 LIST OF TABLES Table page 1 1 The many uses of sgc8c ................................ ................................ ........................ 27 1 2 PTK7 orthologs in various species ................................ ................................ ......... 35 2 1 31nt sgc8c b inds as well as 41nt sgc8c ................................ ................................ .. 47 2 2 Aptamer sequen ces showing sequence similarity ................................ ................... 51 2 3 Azobenzene mutated sgc8c sequences ................................ ................................ 55 2 4 BLAST hits 14/15nt ident ity for GCTGCGCCGCCGGGA in genome ..................... 62 2 5 Aptamers share sequence similarity with DIXDC1b DNA sequence ...................... 63 2 6 KC2D4 has identity to the negative DIXDC1b strand ................................ ............. 65 2 7 All DNA sequences used in this ch apter ................................ ................................ 68 3 1 Consensus region is conserved among species ................................ .................... 77 3 2 Sequences used in this chapter ................................ ................................ ............. 91 4 1 Aligned aptamer sequences fr om three Vaccinia related selection ......................... 94 4 2 BLAST results of VV genome revealed ................................ ................................ .. 96 6 1 Sequences used for aptamer r eceptors ................................ ................................ 118
9 LIST OF FIGURES Figure page 1 1 Sgc8 structure and optimization ................................ ................................ .............. 26 1 2 Schematic Wnt si gnaling pathway ................................ ................................ ........... 29 1 3 Examples of planar cell polarity ................................ ................................ ............... 32 2 1 Hyperbolic fitting of data collected at dif ferent aptamer concentrations .................. 48 2 2 Competition studies ................................ ................................ ................................ 52 2 3 Azobenzene sgc8c mutants ................................ ................................ .................... 54 2 4 Bloc king consensus region with complementary DNA, blocks binding .................... 56 2 5 Pyrene sgc8c mutants ................................ ................................ ............................ 59 2 6 Pyrene sgc8c mutant data ................................ ................................ ...................... 61 3 ................................ .......... 75 3 2 NuPACK structures for DIXDC1b DNA ................................ ................................ ... 78 3 3 dsDIX DNA does not bind CEM cells ................................ ................................ ...... 78 3 4 Structural features of Ig like folds ................................ ................................ ............ 80 3 5 Model of PTK7 extracellular domain ................................ ................................ ....... 81 3 6 Annotated PTK7 primary amino acid sequence ................................ ...................... 84 3 7 Western blot of various cellular compartments of HEK293 cells ............................. 85 3 8 Scratch assay ................................ ................................ ................................ ......... 87 3 9 Scheme of possible PTK7 aptamer interactions ................................ ..................... 88 3 10 Model for PTK7 DIXDC1b interaction ................................ ................................ ... 89 4 1 Predicted secondary structures of consensus sequence ................................ ........ 95 4 2 Sequence identity between two aptamers selected in different labs ....................... 97 5 1 Scheme of AFM measurements on live cells ................................ .......................... 99 5 2 Binding probabilities of tip with ce lls ................................ ................................ ...... 100
10 5 3 Histograms of binding forces between tips and HeLa cells ................................ ... 101 5 4 Rupture forces from literature of different lengths of dsDNA and sgc8c ................ 102 5 5 Representati ve force distance curves for sgc8c AFM tip and HeLa ...................... 105 6 1 Characteristics of streptavidin artificial receptors ................................ .................. 109 6 2 Aptamer inserti on begins after 5min and saturates after an hour .......................... 111 6 3 SA AR does not inhibit cell growth ................................ ................................ ........ 112 6 4 Streptavidin modified cells bind biotin to make cell a ssemblies ............................ 113 6 5 SA AR modified cells are co llected with ma gnetic beads ................................ ...... 114 6 6 Cells modified with thrombin via a thrombin aptamer artificial receptor ................. 115 6 7 Thrombin Modified Cells Cause Clotting ................................ ............................... 116
11 LIST OF ABBREVIATION S PTK7 PTK7 antibody UTR untransl ated region aa Amino acids AFM Atomic force microscopy AFU Arbitrary Flourescence Units AKT Protein Kinase B AMA Ammonium hydroxide: methylamine 1:1 AML Acute myeloid leukemia ANOVA Analysis of variance APC Adenomatous polyposis coli ATCC American Type Culture Collection ATP Adenosine triphosphate BB Binding Buffer (WB, 1g/L BSA and 100mg/L tRNA) bp Base pair BLAST Basic local alignment search t ool BLASTn Nucleotide BLAST BTG4 BTG/TOb family protein BSA Bovine serum albumin Ccd1 Coiled coil domai n 1 protein cDNA Complementary DNA CEM Human T Cell Acute Lymphoblastic Leukemia cell line CH Calponin homology domain chz Chuzhoi mutant CLUSTAL Multiple sequence alignment computer program
12 CPG Controlled pore glass bead DAAM1/2 Disheveled associ ated activator of morphogenesis 1/2 DAPI 4',6 diamidino 2 phenylindole DIXDC1 DIX domain containing 1 protein DIXDC1a Long DIXDC1 isoform DIXDC1b Short DIXDC1 isoform DLAT Dihydrolipoamide S acetyltransferase DMEM DMSO Dimethyl sulfoxide DNA Deoxyribonucleic acid DOX Doxorubicin ds Double stranded dsDNA Double stranded DNA Dtrk PTK7 ortholog in zebrafi sh Dvl Disheveled ECM Extracellular matrix EDTA Ethylenediaminetetraacetic acid EGFR Epidermal growth fact or receptor ESPRIT Bioinformatics algorithm for sequence alignment FASTA Bioinformatics algorithm for sequence alignment FBS Fetal bovine serum FGFR Fibroblast growth factor receptor Flt 1 Vascular endothelial growth factor receptor 1 FRET Fluoresce nce resonance energy transfer Fz Frizzled receptor
13 GAPDH Glyceraldehyde 3 phosphate dehydrogenase GPCR G protein coupled receptor Glycogen synthase kinase 3 H23 Lung adenocarcinoma cell line HA Hemagglutinin HEK293 Human embryonic kidn ey cell line HeLa Henrietta Lacks's cervical cancer cell line Her3 Human epidermal growth factor receptor 3 HF Hydrofluoric acid HPLC High pressure liquid chromatography HRP Horseradish peroxidase Ig Immunoglobulin like domains JNK c Jun N terminal kinases K d Disassociation constant kDa KiloDalton KLG Chicken ortholog of PTK LRP5/6 Low density lipoprotein receptor related protein 5/6 MBTPS1 Membrane bound transcription factor peptidase MCOLN1 Mucolipin 1 MDCK Canine kidney cell line MEKK1 Mit ogen activated protein kinase kinase kinase 1/4 MINA Myc induced nuclear antigen miRNA MicroRNA MT1 MMP Membrane type 1 metalloprotease; MMP 14 MPTMS (3 mercaptopropyl) trimethoxysilane
14 MRI Magnetic resonance imaging mRNA Messenger ribonucleic acid MTH Central myosin tail homology domain NCBI National center for biotechnology information Nuclear factor kappa light chain enhancer of activated B cells NFAT/AP1 Nuclear factor of activated T cells/Fos/Jun NHS PEG MAL N hydroxysuccinimide ester poly(ethylene glycol) maleimide NLS Nuclear localization signal NLStradamus Nuclear localization prediction alogorithm NP Nanoparticle NPC Nuclear pore complex NPH II Vaccinia viral helicase nucleoside triphosphate phosphohydrolase II NR Nanorod nt Nucleotide NuPACK DNA and RNA secondary structure prediction program PAGE Poly acrylimide gel electrophoresis PBS Phosphate buffered saline PCP Planar cell polarity PCR Polymerase chain reaction PE Phycoerythrin PEG Polyethylene glycol Pk Prick le PP2A Protein phosphatase 2 PredictNLS Nuclear localization prediction algorithm pTK Tyrosine pseudokinase domain
15 PTK7 Protein tyrosine kinase 7 qRT PCR Quantitative reverse transcriptase PCR RACK1 Receptor of activated protein kinase C 1 Rho/ROCK Rho associated protein kinase RhoA GTPase Ras homolog gene family, member A RNA Ribonucleic acid Ror Receptor tyrosine kinase like orphan receptor RPMI 1640 Commonly used cell media Runx1 Runt related transcription factor 1 RYK Related to receptor tyrosine kinase SA Streptavidin SELEX Systematic Evolution of Ligands by EXponential enrichment SET Surface energy transfer SHANK1 SH3 and multiple ankyrin repeat domains protein 1 SIK2 Salt inducible serine/threonine kinase 2 siRNA Small interfering RNA sPTK7 Soluble PTK7; the cleavage fragment of PTK7 produced by MT1 MMP ssDNA Single stranded deoxyribonucleic acid STAT Signal Transducers and Activators of Transcription protein SV40 Simian virus 40 TBE Tris/Borate/EDTA buffer TCF/LEF T cell factor/lymphoid enhancer factor TEAA Triethylammonium acetate buffer TGF Transforming growth factor beta TK Tyrosine kinase
16 Transfac Transcription factor prediction program Tris HCl Tris(hydroxymethyl)aminomethane HCl TrkB TrkB tyrosine ki nase tRNA Transfer RNA UBE2D2 Ubiquitin conjugating Enzyme E2D2 UF University of Florida UV Ultra violet UV Vis UV Visible Vangl1/2 van Gogh protein VEGF A Vascular endothelial factor A VEGFR1 Vascular endothelial growth factor receptor 1 WB Washing buffer (PBS, 4.5g/L glucose, 1M MgCl2) Wnt Wingless Int VV Vacci nia virus
17 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DNA APTAMERS USED TO DETECT, TREAT, AND PROBE THE BASIC BIOLOGY OF CANCER By December 2011 Chair: Weihong Tan Major: Medical Sciences with a Concentration in Physiology and Pharmacology C ell SELEX generates artificial DNA and RNA molecules (aptamers) that bind to biological targets of interest with applications for research and therapy. Using this tool, our lab DNA aptamer that to breast, and colon cancer cell lines. extracellular portion of the receptor tyrosine kinase PTK7 which is important for non canoni cal Wnt signaling, and is mis regulated in numerous cancers. Bioinformatic ana lysis of aptamer sequence revealed significant seque nce identity between sgc8c and 3 GCC GCCGGGA) As predicted, all 4 aptamers competed with each other and bound to the cells with similar affinity, bound to the same
18 protein Furthermore, mutational analysis of sgc8c indicated that the consensus region, but not the surrounding nucleotides, is important to aptamer binding. repeat selection, as well as its importance in aptamer protein binding, in the human genome. in UTR) of human DIXDC1b DNA, which, along with PTK7, is a regulator of non canonical Wnt signaling. Surprisingly, aligning the aptamers with DIXDC1b DNA, we found 2 of the aptame rs, KMF9b and H01 had additional nucleotides in common with DIXDC1b DNA outside the consensus region. This 22nt region, shared by the PTK7 aptamers and DIXDC1b DNA, is unique in the human genome. We were also surprised that another aptamer, KC2D4, which also compet es with the PTK7 aptamers, but which shares no sequence similarity with them, has identity to the opposite strand of DIXDC1b DNA. In total, 5 aptamers compete for bind ing the same site on PTK7. Four of these aptamers share sequence identity to the positive strand of UTR. The fifth aptamer shares sequence identity 5 bases from this region, but on the opposite strand of DIXDC1b DNA. could be consistent with the protein PTK7 melting the genomic DNA and interacting with the resulting ssDNA hairpins, which share sequence identity to the aptamers formed by each melted strand. If this is indeed the case, it could affect DIXDC1b transcription. These sequences identities are conserved
19 in the putative DIXDC1b regulatory regions of Pan troglodytes Rattus norvegicus and Mus musculus PTK7 has a weak predicted nuclear localization sequence. By cell fractionat ion, followed by Western blot, we found an abundance of cleaved PTK7 in the nuclei of several different cell lines. This suggests that whole cell SELEX has not just found an aptamer for an extracellular protein, but identified a genomic DNA s equence, which the protein binds to naturally after it is cleaved, internalized, and transported to the nucleus. While we have only begun to address this receptor tyrosine kinase and this regulatory pathway, these findings suggest that whole cell SELEX mi ght be used more generally to identify novel extracellular transcription factors with highly specific binding motifs. Our bioinformatic analysis of other aptamer sequences selected against whole cells has yielded other examples of disparate selections yiel ding similar sequences, hinting that there may be other proteins that interact with DNA on the plasma membrane yet to be discovered. Our results are important for three main reasons: (1) The proteins PTK7 and DIXDC1b are murkily understood, yet key players in Wnt signaling, which is crucial to embryonic development and cancer progression. Understanding their relationship could be important for understanding Wnt signaling. For instance, there are two isoforms of DIXDC1, a and b; only DIXDC1b has the consensu s sequence. PTK7 binding this of PTK7 binding aptamers has sequence identity to both strands of DIXDC1b DNA, implying that, if melted, the DNA might be forming ssDN A hairpins that bind to PTK7 in
20 a manner similar to the aptamers. This would be a new type of transcription factor of interest to molecular biologists. (3) Finally, this finding is probably not a one time occurrence. Our bioinformatic analysis of 148 DNA s equences selected by whole cell SELEX identified other aptamers from disparate selections like those for Vaccinia infected cells and pure virus, which share significant sequence identity. Further comparison of existing DNA and RNA aptamers may yield other examples of SELEX identifying natural DNA or RNA sequences with functional roles, not initially envisioned. Future selections should also not be considered complete until the newly selected aptamers are compared with all other existing aptamers for sequen ce identity.
21 CHAPTER 1 INTRODUCTION G enetic al terations in cells cause cancer, which result in different cell behavior due to changes to the cell at the molecular level. There has been a big push in recent years to develop biomarkers for these changes to improve cancer diagnosis and treatment. To identify unique molecular features of target canc er cells, our lab has previously developed a cell SELEX ( cell systematic evolution of ligands by exponential enrichment) method for the selection of a panel of aptamers that specifically recognize leukemia, liver and lung cancer cells among others [ 1 2 3 4 5 6 7 8 9 ] These a ptamers are single stranded DNA molecules that bin d, through their unique secondary structures, to molecular targets on the cancer cell surface. A counter selection strategy is used to collect those DNA sequences that interact with the target cells but not with the control cells. Consequently, this process enriches aptamer candidates that exclusively bind to the target cells One such aptamer is sgc8c, which binds to PTK7 [ 6 ] a protein ty rosine pseudo kinase misregulated in numerous cancers including various colon cancer [ 10 ] leukemia [ 11 ] and melanoma [ 12 ] PTK7 has been found to play a role in development by regulating planar cell polarity in vertebrates through the Wnt pathway [ 13 ] and it is also believed to be important for invasiveness and metastasis in cancer [ 14 15 16 ] A ptamers including sgc8c are easily functionalized to ma ny different surfaces. T hroughout t he first part of my doctoral work I collaborated with several members of my lab to attach sgc8 c to var ious nanomaterials including dye doped silica nanoparticles (NPs) [ 17 ] liposomes [ 18 ] and gold NPs [ 19 ] to target PTK7 expressing cells These
22 sgc8c targeted materials could improve current diagnostic and therapeutic approaches to treat cancers that express PTK7. After several intense years of study about PTK7 biology and the selection process for sgc8 c conducted throughout my work on ap tamer functionalized nanomaterials I was perplexed by the observation that several different aptamers selected on different cell lines, by different people, years apart, competed with each other for binding to PTK7. So I created a database of 148 aptamers selected by whole cell SELEX against various cancer types and performed bioinformatic analysis on them. Curiously, I found that three other aptamers out of 148, shared sequence identity with sgc8 c I further found that these four aptamers competed with each other, and that the 14nt sequence that the four different aptamers shared was not only important for the untranslated UTR) of a p rotein also involved in Wnt signaling, DIX DC1b. A fifth aptamer, which also competed with the four other PTK7 aptamers, but which did not share sequence identity with them, shared identity instead with the opposite DNA strand of the UTR, adjacent to the region sharing sequence identity with the other aptamers. This suggested that whole cell SELEX has not just found an aptamer for an extracellular protein, but may also have identified a genomic DNA sequence the protein bi nds to naturally. Throughout this dissertation, I will try to develop a new concept, that whole cell SELEX might be used more generally to identify novel extracellular
23 transcription factors with highly specific binding motifs. Furthermore it may shed light on possible therapeutic pathways of current and future ssDNA aptamers. This Introduction chapter will review, in depth, aptamer selection, and what is currently known about PTK7 and another protein that is important to our story, DIXDC1b. These discussions consist of a comprehensive rev iew of all literature relating to these two proteins. For the casual reader, not interested in all of these details, the review of PTK7 and DIXDC1b in the Introduction can be skipped over and mainly used as a reference when reading the later chapters. Ch apter 2 will detail the discovery of the DIXDC1b DNA analog to the PTK7 aptamers in the genome. Chapter 3 will try to determine how PTK7 might be interacting with DIXDC1b DNA. Chapter 4 will explore other potential repeatedly selected aptamers identified t hrough bioinformatics. Chapter 5 will show how we used atomic force microscopy (AFM) to determine that sgc8c aptamer PTK7 protein interactions on live PTK7 antibody protein interactions. Chapter 6 will detail the modification of cell surfaces with aptamer that can act as artificial receptors. Finally, Chapter 7 will present some conclusions and detail further experiments that will help clarify and deepen our understanding of the relationship between PTK7 and DIXDC1b. Aptamer S election Molecular aptamers are s ingle stranded DNAs (ssDNAs) or RNAs 15 80nt in length, which can recognize target proteins, peptides and other small molecules. The dissociation constants of aptamers to targets can range from 10 12 M 10 8 M. Aptamers rec ognize their targets with high specificity, and can typically discriminate between protein targets that are highly homologous or differ by only a few amino acids [ 20 21 ] The secondary structures formed by the ssDNAs are the basis for targe t protein
24 recognition [ 22 23 ] These aptamers are selected by a process called SELEX (Systematic Evolution of Ligands by Exponential enrichment), where the aptamers are selected from libraries of random sequences of synthetic DNA or RNA by repetitive binding of these oligonucleotides to the target molecules [ 24 25 26 ] Through this iterative in vitro selection process, aptamers with high specificity and affinity to their targets can be obtained. M ost of the aptamers reported so far have been selected using pure molecules, such as purified proteins as the targets. Aptamer selection against complex targets (such as red blood cells and single protein on live trypanosomes) was also demonstrated [ 24 27 28 29 30 31 ] Aptamers have several key advantages over antibodies in molecular recognition and imaging, such as low molecular weight, easy and reproducible synthesis, simple modification, fast tissue penetration, low toxicity or immunogenicity, easy storage, high binding affinity and specificity that are very comparable with antibodies [ 24 29 ] Aptamers have shown great promise in molecular recognition, diagnosis and therapy. To produce probes for molecular analysis of tumor cells, our lab has developed a novel method, the cell based aptamer selection process (cell SELEX) for aptamer select ion [ 2 5 6 7 8 ] Instead of using a single type of molecule as a target the cell SELEX process uses whole cells as targets to select ssDNA aptamers that can distinguish target cells from control cells. In addition to the aptamer advantages me ntioned above, a big advantage of the cell SELEX technology is that there is no need of prior knowledge about potential biomarkers for cancer on these cells. A group of cell specific aptamers can be selected using a subtraction strategy in a relatively
25 sho rt period without knowing which target molecules are present on the cell surface Moreover, the selections may identify important biomarkers. Compared to 2 D gel electrophoresis and mass spectrometry u sed for proteomic studies aimed at identifying protein s, cell SELEX produce s molecular probes first, which can then be use d to identify the target proteins. Thus, not only can the selected aptamers be used as molecular probes for molecular analysis of cancer but also they can be used as tools for ide ntifying new biomarkers expressed by tumor cells or other cells in dicating disease status. The PTK7 Aptamer Sgc8c Sgc8, was among the first aptamers selected by whole cell SELEX in the Weihong Tan lab. It was discovered by Dihua Shangguan through selection for the T cell leukemia cell line, CEM CCRF but not Ramos, a B cell leukemia [ 7 ] The full sequence of sgc8 is: ataccagcttattcaatt AGT CAC ACT TAG AGT TCT AAC TGC TGC GCC GCC GGG AAA ATA CTG TAC GGT TA gatagtaagtgcaatct. The lower case, underlined regions, are the primers, and the upper case region is the variable part of the aptamer. This aptamer was interesting because it had a very low Kd of 0.8nM, and bound strongly to the target CEM cells, but not at all to the negative Ramos cells. Furthermore the aptamer showed a clear stem loop hairpin structure [Figure 1 1], wh ich allowed Shangguan, by progressive optimization, to shorten the 88nt aptamer to a 41nt version ATC TAA CTG CTG CGC CGC CGG GAA AAT ACT GTA CGG TTaga without altering the binding properties. Over time more was discovered about the aptamer: it binds at a physiological relevant temperature, 37C; it was swiftly internalized into cells [ 32 ] ; and most importantly it selectively binds a membrane protein tyrosine kinase called PTK7 [ 6 ]
26 While aptamers hold a lot of promise through their size and ease of functionalizati on, compared to antibodies they are relatively new and untested. Therefore, to show the full potential of aptamers for the diagnosis and treatment of cancer, a model aptamer can be used. Sgc8c is a great model aptamer: it is short; it binds at physiologica l temperature, with a low K d ; it has a known target, PTK7, which is upregulated in a lot of cancer types; it internalizes which is a major problem for drug delivery; and there are control cell lines, like Ramos, which do not express PTK7. As a result of th ese properties, sgc8c has been conjugated with different imaging and therapy modalities to specifically target a particular tumor type. Table 1 1 has an extensive list of different dyes and drugs sgc8c has been conjugated with and to what purpose. Figur e 1 1. A) Full length sgc8 structure. B) Shortened sgc8c structure. Both sequences were modeled by Nu PACK [ 33 ] A clear s tem loop hairpin is visible in both. The red dashed line shows where the full length sgc8 was cut to make sgc8c. A: adenine; C: cystosine; G: guanine; T: thymine.
27 Each of these strategies has taken a non selective dye or drug and, by attaching the sgc8c ap tamer to its surface, made it able to selectively target those cells that express PTK7 on their surface. Throughout this dissertation we will return again and again to why and how this aptamer binds PTK7. To begin, the next section will be an in depth look at our current understanding of PTK7; its structure, functions, and role in development and cancer. Table 1 1. The many uses of sgc8c. Purpose sgc8c Conjugated Material Modality Targeted Detection Dye [ 34 ] Profiling Patient Blood Fluorophore doped and magnetic silica NPs [ 35 ] Extracti on then Fluorescent Au NPs [ 36 ] Colorimetric Ultrasound microbubbles [ 37 ] Ult rasound Different sized AU NPs Molecular ruler SET Microfluidic device [ 38 39 40 ] Enrichment and capture Multi dye loaded silica NPs [ 41 ] Fluorescence Dye doped silica NPs [ 42 ] Flow Cytometer Square capillary channels [ 43 ] Capture, microscopy Dye with a dynamic sgc8 structure [ 44 ] In vivo fluorescence Targeted Treatment Doxorubicin [ 45 ] Chemotherapy Doxorubicin l oad ed l iposomes or micelles [ 18 46 ] Danunorubicin [ 47 ] Acrylamide polymer [ 48 ] Au Ag NR [ 49 ] Photo thermal therapy DNA G Quadru plex loaded with TMPyP4 [ 50 ] Photodynamic therapy Virus particle [ 51 ] Gene delivery Multi modal Hollow magnetic silica NPs loaded with DOX [ 52 ] In vivo MRI imaging and chemotherapy Au Ag NRs coated in DNA nanogel with DOX [ 53 ] Photo thermal and Chemotherapy sgc8 : anti PTK7 aptamer; NP: nanoparticle; MRI: magnetic resonance imager; DOX: doxorubicin; Au : gold; Ag : silver; TMPyP4 : 5,10,15,20 tetrakis (1 methyl 4 pyridyl) 21 H ,23 H porphine (TMPyP4)
28 Biology of PTK7 Protein tyrosine kinase 7 (PTK7) is an ancient protein found as far back in the evolutionary tree of life as the fresh water animal Hydra While conserved and clearly important PTK7 mutant s lacking the first 114 amino acids ( aa ) or the cytoplasmic domain are embryonic lethal in mice [ 13 ] its exa ct purpose and signaling details are still unclear. Since its characterization in 1995 [ 54 55 ] various knockout models and mutation experiments have identified a central role for PTK7 in a process called planar cell polarity (PCP) [ 56 57 58 ] PCP causes changes in the underlying architecture of cells by rearranging the cells' actin filaments. Such rearrangements are crucial to early, dynamic developmental events such as gastrulation, neural tube closure, and convergent extension. Animal exper iments show PTK7 mutation produces stunted embryos with open spinal cords and other defects [ 13 ] As with many proteins important for development, dys regulation of PTK7 expression is also a factor in many different cancers especially colon cancer [ 10 ] leukemia [ 11 ] and melanoma [ 12 ] Wnt Signaling Pathway To re quires a basic understanding of the Wnt signal ing pathway Wnt signaling is crucial during embryonic development, and is one of eight major pathways mutated in cancer the others are AKT, Hedgehog, Death pathway is actually not just one pathway, but three that share common proteins but which have vastly different signaling outcomes. In Wnt signaling, one of nineteen different extracellular Wnt ligands binds to a Frizzled receptor (Fz) on the cell membra ne which in turn interacts with one of several different types of co receptor membrane proteins, such as low density lipoprotein
29 related receptor proteins 5/6 (LRB5/6) or PTK7, to bind a cytoplasmic protein called Disheveled (Dvl ). Depending on which of th e nineteen Wnts binds Fz, and what co receptor Fz is associated with, one of three different signaling cascades is activated: the catenin/Wnt pathway, or one of two non canonical Wnt pathways, either the PCP/Wnt pathway, or the Ca 2+ /Wnt path way Please refer to Figure 1 2 for a pictorial description of the components of Wnt signaling important for this discussion. Figure 1 2 Schematic of the canonical and non canonical Wnt signaling pathway. Please see accompanying text for figure details.
30 Catenin/Wnt Pathway A mutation consid ered the initiating event in 85% of colorectal cancers occurs in catenin/Wnt signaling [ 59 60 ] This is not a surprise bec ause the end catenin arrives in the nucleus and activates a powerful set of genes called the TCF/LEF genes that drive among other things, cell replication. When no Wnt ligands are present, a complex made up o catenin and three other proteins GSK 3, APC, Axin are found bound tightly together throughout the cell. If these protein complexes remain intact, they are continually destroyed, causing no signaling [Figure 1 2a] However, when a Wnt ligand binds to a Fz receptor that is interacting with LRP5/6 co receptor, several changes take place and Dvl which is bound to Fz, and LRP5/6 catenin complex which was held together by Axin becomes unstable catenin is released, fr ee to travel to the nucleus and activate the powerful TCF/LEF genes [Figure 1 2b] Until recently catenin/Wnt canonical pathway ; however, over the last year, two papers have challenge d this assumption. The catenin through its intracellular domain, potentially stabilizes it, and allows it to signal [ 61 ] Knocking down PTK7 expression prevents TCF/LEF gene transcription when cells are stimulated with Wnt3a a ligan catenin/Wnt signaling. The second paper asserts that the extracellular domain of PTK7 can interact directly with Wnts to inhibit catenin/Wnt signaling [ 62 ] In both studies PTK7, when bound to Fz, was shown to interact with catenin signaling The papers suggest that becaus e Wnt/ PTK7 /Fz complexes bind Dvl but not Axin like the
31 Wnt/ LRP5/6 catenin destruction complex. This leaves catenin free to enter the nucleus and activate TCF/LEF genes. The Wnt/PTK7/Fz/ D vl complexes are free, however, to signal through non canonical PCP/Wnt signaling. It is this PCP/Wnt signaling that PTK7 has been most closely associated with. The papers catenin/Wnt pathway to the PCP/Wnt pathway. PCP/Wnt Pathway catenin/Wnt signaling, Dvl and Fz proteins are also major players in PCP/Wnt signaling [Figure 1 2c]. The signaling cascade is very different with respect to PCP/Wnt signaling. For proper development and healthy functioning, cells align in relation to other cells on an xy plane. For instance, if you look down at the hairs on your arm right now, you will see they are, for the most part, tidily aligned and pointing in the same dire ction. The same is true for mesenchymal cells growing along the neural crest during embryo development ; these cells align their mitotic spindles in the same direction [Figure 1 3b] This causes dividing cells to pull toward the newly formed neural tube, he lping close it. Or more easily seen by researchers, the small hairs in animals' inner ears or the spiny projections on Drosophila larva called denticles [Figure 1 3c], are carefully aligned along their respective xy planes. PCP signaling is a major driver of this exquisite order and its main method of action is remodeling actin. Actin is a major component of a cellular shape. PCP acts in a left versus right (or distal versus proximal) direction, perp endicular to the more commonly described apical/basal polarity seen in a typical epithelial cell of the gut, where the top or apical side of the cell faces out into the lumen
32 and catches nutrients using cilia and the bottom or basolateral side of the cell is smooth and faces into the cell. Figure 1 3. Examples of planar cell p olarity. A) UF marching band members aligned along an xy plane. B) Neural tube closure aided by aligned mitotic spindles. C) Drosophila denticles are the white projections on each s egment of a 22h old larva, mutations of proteins in the PCP pathway causes perturbations in these denticle patterns. A number of proteins are involved in an complex, intricate dance to determine right relative to where that cell is positioned compared to others. Five proteins are especially important to development of the protein gradients important to forming this PCP: Flamingo catenin /Wnt signaling mentioned above), Dvl (same), van Gogh (Vangl1/2), and Prickle (Pk). Flamingo is found on both sides of the cell, while the other proteins migrate in pairs comprised of one transmembrane protein and one cytosolic protein with Fz and Dvl on one side and Vangl1/ 2and Pk on the other. Before an xy axis is patterned in the cell, these proteins are found mixed throughout the cytoplasm. Over time, in response to input from receptors on other cells, Fz and Dvl localize to one side of the cell and V angl1/2 and Pk localize to the other, creating a gradient in the cell. The absolute level of Fz on a cell surface is not what det ermines polarity. Instead, it is differences in the level of Fz between cells that
33 determines polarity. In an elegant, early ex periment on PCP, researchers poured hot wax on Drosophila pupae expressing an Fz transgene The Fz transgene was driven by a temperature sensitive heat shock promoter, which only made protein in areas where the hot wax fell. Regardle ss of where they would normally grow, wing hairs invariably grew away from where the wax had turned on Fz activation Please read A. Jennifer or [ 63 64 ] In addition to these 5 main proteins, there are a number of accessory proteins that appear to modulate PCP signaling. These include Ror, RYK and PTK7. When particular Wnt ligands espe cially Wnt5a, Wnt5b, or Wnt11, bind to Fz with a PTK7 co receptor, PTK7 recruits Dvl, which activates disheveled associated activator of morphogenesis 1 and 2 ( DAAM1/2 ), which in turn activat es RhoA GTPase, a key regulator of actin filaments through Rho/ROCK, among numerous other proteins (Figure 1.2c). DAAM1/2 and RhoA GTPase also activate longer term signaling, including c Jun N terminal kinases (JNK) [ 56 ] and/or protein kinase C [ 65 ] each of which cause changes in cell polarity, migration, and adherence. cate nin, and Wnt3a, PTK7 has been reported to interact with other proteins, including, Plexin A1 [ 66 ] Vangl2 [ 13 ] Flt 1 (VEGFR1; vascular endothelial growth factor receptor 1) [ 67 ] and a receptor of activated protein kinase C 1 (RACK1) [ 65 ] Recently, Flt 1/PTK7 interactions have been implicated in angiogenesis. In a mouse corneal model, PTK7 overexpression produced more angiogenesis, while siRNA knockdown of PTK7 produced less angiogenesis. The addition of VEGF A (vascular endothelial factor A) serves to strengthen the interaction between Flt 1 and
34 PTK7 [ 67 ] RACK1 also interacts with PTK7, and this interaction appears to be important for neural convergent extension in Xenopus [ 65 ] PTK7 expression is also linked to apoptosis. When PTK7 is knocked down in HCT116, a colon cancer cell line, apoptosis is increased through the caspase 10 pathway [ 68 ] Likewise increasing PTK7 appears to rescue patient derived acute myeloid leukemia (AML) cells from apoptosis, while de creasing PTK7 led to apoptosis through the effector caspases, 3 and 7 [ 69 ] PTK7 plays important roles in many different cell types from epi thelial to endothelial [ 70 ] and T cells [ 71 ] and it appears that in different cell types, depending of what sort of proteins PTK7 interacts with, different sign aling occurs. There is still a long way to go before the varied web of PTK7's protein interactions and behavior in various cell types is untangled and fully understood. PTK7 generally acts as a switch that turns the Wnt signaling machinery Fz and Dvl away catenin signaling toward Wnt/PCP signaling. PTK7 Structure Structurally, PTK7 has 7 immunoglobulin like domains (Ig) that hang outside of the cell, a single transmembrane domain, and a tyrosine pseudokinase domain (pTK) inside the cell This protein is structurally unique among tyrosine kinases (TKs) for a number of reasons: its extracellular domain is composed of only Ig domains, which in humans and zebrafish contain a matrix metalloprotease (MMP) cleavage site; its transmembrane domai n is more conserved than those of any other receptor TK; and its catalytically dead pTK domain, is also inert in all known orthologs [ 72 ] These characteristics indicate a protein with conserved biological function involving all three
35 protein domains, which arose long ago, in the metazoan radiation, during the Cambrian explosion, 530 million years ago. Pseudo Tyrosine Kinase Domain of PTK7 The pTK domain has structural similarities with other tyrosine kinases, but two crucial areas important for phosphorylating tyrosines have been mutated: The GXGXXG motif which acts as a clamp to help anchor transferable phosphates of ATP is mutated to 803 GXSXXG; and the DFG motif, which chelates a Mg 2+ phosphate for transfer, is mutated to 948 ALG [ 54 ] These PTK7 kinase mutations are highly conserved and whi le apparently not active, the pTK domain does have other biologically conserved fu nctions [Table 1 2]. Of the 58 TK proteins, 5 PTK7; RYK, another co receptor for Fz mentioned above ; Her3, an EGF receptor family member; and two ephrin receptor members, EphA10 and EphB6 are not functional tyrosine kinases [ 73 ] Table 1 2. PTK7 orthologs in various species. The active kinase motif, DFG, is mutated in all of them. Organism PTK7 O rtholog Mutated DFG Motif Hydra Lemon FLD Drosophila Dtrk YPA Chicken KLG ALS Human, mouse PTK7/CCK4 ALG A major binding partner of PTK7 pTK is Dvl the cytoplasmic phosphoprotein mentioned previously which is important in all three Wnt signaling pathways. The first Dvl was originally isolated from a Drosophila mutant with unruly wing hair (a symptom of poor PCP) hence the name [ 74 ] Dvl has three domains DIX, PDZ, and DEP. Dvl is catenin/Wnt signaling as this is where Axin binds [ 75 ] DIX domain later in this
36 work. The DEP domain is important for all Wnt signaling. The PDZ domain is important mainly for PCP, as PTK7 and Dvl interact directly through the PDZ when in a complex with Fz [ 56 ] In order for PTK7 to cause signaling, Dvl must first move from the cytoplasmic PDZ domain. In a Xenopus model transfected with mutant Dvl and wild type PTK7, o the cell membrane, but removal of PDZ did [ 56 ] Similarly when the pTK domain of PTK7 was removed, Dvl movement to the membrane was prevented. PTK7 and Dvl cannot, however, interact on their own; they need another protein, like Fz, to aid their interaction [ 56 ] Transmembrane Domain of PTK7 Many transmembrane proteins thread through the membrane several times, and have conserved transmembrane domains. By contrast, TKs thread through the membrane only once, and rarely have a h ighly conserved transmembrane domain. PTK7 has the most conserved transmembrane domain among TKs. It has a predicted 22aa helical structure, which is identical to chicken KLG, and it shares 55% homology with Hydra Lemon and Drosophila Dtrk, other PTK7 orth ologs. The PTK7 transmembrane domain includes a GxxxG motif, which in some proteins promotes interactions between transmembrane helices [ 72 ] T hese types of interactions, especially in TKs, promote signaling. However, studies where the transmembrane domain was made into a bacterial fusion protein found no increase in helix helix interactions indicating some other, yet unknown, method of action allows PTK7 to form homodimers and heterodimers with Fz [ 76 ]
37 Extracellular PTK7 and its MT MMP1 Cleavage Site The extracellular portion of PTK7 is composed of seven Ig like domains, held sheets interact by a single disulfide cysteine bond. The Ig fold is most commonly associated with antibodies, but is also found i n many proteins that encourage cell cell interactions, as well a number of critical [ 77 ] Statistical analysis of Ig folds across receptor TKs, found that PTK7 evolved very early in receptor TK development [ 78 ] To date, no ligand has been found that binds to PTK7, but the extracellular portion of the protein is clearly important for signaling. The ST Lee group from the University of Korea, one of the grou ps who first cloned PTK7 mRNA, found that addition of a soluble fragment of PTK7 made of just the 6 th and 7 th Ig domains (~200aa) had profound effects on cancer related signaling as seen by reduced VEGF induced tube formation, wound healing, migration, in vasion, and angiogenesis in HeLa [ 70 ] This showed the fragment was competing with normal PTK7 for signaling. This result was rep using the full length extracellular PTK7 which also caused an increase in the apoptosis [ 69 ] S everal years after this finding, the AY Strongin group from Stanford University might have discovered an explanation. They found an MT1 MMP cleavage site in the 7 th Ig fold, above the transmembrane domain [ 79 ] MT1 MMP, which is also known as MMP 14, is a membrane bound endopeptidase that cleaves extracellular matrix (ECM) proteins; other soluble MMPs, activating them; and, other membrane bound signaling receptors, including, PTK7. It is known to be overexpress ed in cancer, with a higher level of expression yielding more invasive and malignant tumors [ 14 15 16 ] Previous
38 studies had linked Wnt/PCP signaling and PTK7 in particular with MT1 MMP expression [ 79 80 ] They confirmed that MT1 MMP cleaves PTK7 PKP 621 LI into two pieces: a 50kDa C Terminal membrane bound pTK fragment and a 70k Da soluble PTK7 (sPTK7) fragment made up of the first 6 Ig like domains. They found these sPTK7 fragments in the cell media and bound to full length PTK7 on the surface of the cells. Furthermore, their mutant, L622D, which is resistant to MT1 MMP cleavage, could not cause downstream actin remodeling events. Finally, overexpression of sPTK7 in HT1080 primate fibrosarcoma cells did not affect actin remodeling, but did cause a large spike in RhoA GTPase expression, which is the upstream signaling molecule of a ctin remodeling that is activated by PTK7. Furthermore, although not explicitly stated in their conclusions, they found that MT1 MMP treatment of MDCK canine kidney cells caused PTK7 to move from the cell membrane and accumulate around the cell nucleus. Th e AY Strongin group followed up on this work by characterizing a mutant PTK7 protein with not one, but two MT1 MMP cleavage sites. This mutant PTK7 protein was made in a mouse strain, chuzhoi ( chz ), created by exposing the mice to the mutagen N ethyl N nit rosourea [ 57 ] The mutant mice showed signs of cl assic PCP signaling problems, including neural tube defects and disrupted hairs in the inner ear. In addition to the normal MT1 MMP cleavage site in the 7 th Ig contained a second cleavage site in the linker between the 5 th and 6 th Ig like domain where QVLEKLK 503 has been mutated to QVL ANP 503 by the addition of ANP In experiments comparing the mutant to wild type PTK7 in HT1080 cells, when MT1 MMP1 was added to wild type cells, PTK7 migrated from the cell surface to su rrounding the nucleus, and when sPTK7 was ectopically added to the cells, RhoA
39 GTPase expression rose. By contrast, when MT1 MMP1 was added to the chz mutant PTK7 was not able to localize to the membrane and the addition of sPTK7 caused no change in RhoA G TPase expression. However, it did cause an increase in the cells' invasion ability. PTK7 Gene Structure and Splicing PTK7 is located on chromosome 6 at 6p21.1, and has an 883 base pair (bp) s promoter lacks both a TATA box and CAAT box, and, instead, has a CpG island with several GC box/SP1 binding motifs [ 81 ] This promoter resembles those of housekeeping genes, which are constitutively expressed in most cells, and which are essential for general cell maintenance. It contains TCF/LEF SP1 and 5 bHLH binding sites [ 10 ] The TCF/LEF catenin pathway activation, dysregulated in c olon cancer. There is one main PTK7 variant, which has been the focus of our discussion; it has 1070aa stitched together from 20 exons. There are, however, four other PTK7 splice variants expressed at lower levels shown to have different expression levels in cancer cell lines. These variants lack various portions of the protein. PTK7 2 is missing the N terminal half of the 6 th Ig (Exon 10). PTK7 3 is missing the 5 th and N terminal half of the 6 th Ig (Exons 8 10) and has the highest mRNA expression levels n ext to the major PTK7. PTK7 4 is missing the 7 th Ig (portion of Exon 12, and Exon 13) this variant retains the MT1 MMP cleavage site, but removes a 55aa piece 5aa toward the C terminal of that site. PTK7 4 is truncated due to a frame shift mutation, which keeps the transmembrane domain, but which removes most of the pTK domain, and is only highly expressed in the testis. These splice variants may be interesting because in other
40 receptor TKs, splice variants exhibit different ligand binding affinities, as wi th FGFR2(IIIb) and FGFR2(IIIc), or form dominant negative heterodimers between major and minor splice variants, as with TrkB. sgc8c Aptamer In their paper [ 82 ] the A.D. Ellington group at the University of Texas was surprised that the anti PTK7 aptamer, sgc8c, bound to so many different types of cancer cell lines, including colon, leukemia, cervical, and breast cell lines. They conclude that this might be becau se PTK7 is a marker for cell adherence to culture dishes and not relevant. However, based on our review of the literature and PTK7's role in Wnt signaling, we contest the A.D. Ellington group's conclusion that sgc8c binds to many different types of cancer cells lines because PTK7 is a marker for cell adherence. Instead, we argue that sgc8c binds to many cancer cell lines because PTK7 is overexpressed in many different cancer types. DIXDC1 Biology Although PTK7 acts as a switch that shunts Wnt signaling toward PCP pathways, there is another protein that has the opposite effect and will become importa nt to our discussion : DIX Domai n Containing 1 protein (DIXDC1) is also known as coiled coil domain 1 protein (Ccd1) in zebrafish and mice, or KIAA1735. There are three known proteins with a DIX domain : Dvl and Axin, which we have previously mentioned ; and DIXDC1, which we will detail here. First identified in humans in 2000 [ 83 ] DIXDC1 is the least studied of the three DIX domai n containing proteins. It is becoming clear, catenin activity while inhibiting Wnt/PCP activity. For instance, DIXDC1 has been implicated in cancers with dysregulated Wnt signaling. Overexpression of DIXDC1 leads to survival of human
41 colorectal and lung adenocarcinoma s through catenin crosstalk w ith the PI3K/AKT pathway. By contrast, downregulation of DIXDC1, as in squamous cell carcinoma of the lung, leads to aberrant upregulation of Wnt/PCP signaling [ 84 ] DIXDC1 I catenin Signaling DIXDC1 is spatially coexpressed with other Wnt signaling proteins in developing zebrafish and mouse embryos [ 85 86 ] In embryos, overexpression of Ccd1, the homolog of DIXDC1 in zebrafish le a ds to a reduction in eyes and forebrain ; a similar phenotype to W nt8 overexpression catenin signaling. Addition of a dominant negative Ccd1, has the opposite phenotype more similar to Axin catenin signaling. Furthermore, Wnt8 overexpression rescues dominant negative DI XDC1 treatment. These in vivo results imply that DIXDC1 acts upstream of Axin but downstream of Wnt ligands [ 87 ] At the molecular level al so DIXDC1 is a positive regulator of canonical and a downregulator of non canonical Wnt signaling. The major downstream effect of catenin is expression of the TCF/LEF genes. siRNA knockdown of DIXDC1 suppresses Wnt3a TCF/LEF signaling, while upregulating DIXDC1 increases Wnt activation by the same amount as Dvl upregulation alone. Coexpression of both Dvl and DIXDC1 act synergistically to greatly increase TCF/LEF [ 88 ] Furthermore, D IXDC1 levels were increased after Wnt 3a stimulation, but it was less phosphorylated. While no change in DIXDC1 mRNA was found, a ctivation of canonical Wnt signaling decreased ubiquitin dependent degradation of ectopic and endogenous DIXDC1 cateni n signaling might upregulate DIXDC1 through a post translation mechanism by inhib iting degradation of the DIXDC 1 protein [ 89 ]
42 One major way DIXDC1 affects Wnt signaling is by controlling the amount of bioavailable Dvl protein found in cells. On its own, endogenous Dvl forms large found in discrete aggr egates, called puncta, throughout the cell. In these large aggregates, Dvl is unavailable for signaling, because they need to be broken up into smaller assemblies that can travel to the cell membrane and bind with receptors such as Fz or PTK7. By contrast Dvl DIXDC1 hetero assemblies also occur through DIX domain interactions, but the hetero assemblies are much smaller than homo assemblies of Dvl Dvl. Instead of having high molecular weight aggregates, Dvl DIXDC1 hetero assemblies typically have 3 Dvl and 1 DIXDC1 protein. This difference in oligomer size by a hydrophilic histidine residue in DIXDC1 [ 88 ] In fact domain alone is able to break up D vl puncta in cells. Upon Wnt activation, increases of DIXDC1 occur in the in cytosol where they are recruited to Dvl aggregates, depolymerizing them into the active trimer. Thus one way DIXDC1 controls Wnt activity is by making Dvl available for signaling [ 88 ] DIXDC1 Decreases Wnt/PCP Signaling DIXDC1 increases canonical Wnt signaling, but decreases non canonical Wnt signaling by modulating JNK signaling in several different ways [ 90 ] As mentioned in the above discussion of PTK7, JNK activation through PTK7 Dvl interactions is important for chang es in PCP, including cell polarity, migration, and adherence. DIXDC1 inhibits JNK activation through independent interactions with both Axin and Dvl.
43 Axin activates JNK through the Mammalian Mitogen activated protein Kinase Kinase K inase 1 and 4 (MEKK1 an d MEKK4), which are serine/threonine kinases that to a region of the Axin protein, called PP2A, located 30aa from the N DIX domain. This is the same regi on through which Dvl and Axin interact [ 91 ] When DIXDC1 binds to Axin, the DIXDC1 Axin complex cannot bind to MEKK1, and thus this occurs, MEKK4 is physically seques tered by DIXDC1 and cannot bind to Axin, preventing JNK signaling [ 90 ] While Axin signals through MEKK1/4, as mentioned above, Dvl non canonical Wnt signaling occurs when Fz interacts with its co receptor PTK7, which in turn recruits Dvl, causing activation of DAAM1/2 and signaling through RhoA GTPase. While the details are not clear, when Dvl and DIXDC1 are bound together they can inhibit the downstream activation of JNK. This inhibition requires the coiled coil domain, which is called the central myosin tail homology domain (MTH) of DIXDC1. DIX DC1 mutants missing MTH domains do not inhibit JNK [ 90 ] Coiled coil domains common in centrosomal proteins and DIXDC1 have been shown to c o localize with gamma tubulin during interphase and mitotic phase in HEK293 suggesting a further role for DIXDC1 in chromosome segregation [ 92 ] DIXDC1 Gene Structure Human DIXDC1 is found at 11q23.1, located between microsatellite markers D11S927 and D11S1347 in a region commonly deleted in sporadic breast cancer Other genes in this region include KI AA1391, BTG/TOb family protein BTG4 and SIK2 (salt induc ible serine/threonine kinase 2) [ 93 ]
44 Human DIXDC1 has two different transcription initiation sites yielding two isoforms, a long form and short form: DIXDC1a and DIXDC1b respectively. DIXDC1b is predominantly expressed during development and in the brain [ 83 ] During adulthood, DIXDC1a is the more p revalent isoform [ 94 ] Both isoforms have an MTH and a C terminal DIX domain. In humans t he MTH region contains 2 tyrosine phosphorylation sites and a leucine zipper motif. The longer form DIXDC1a is 683aa coded by a 6.3kb mRNA. This isoform has an N terminal extension containing a calponin homology (CH) domain. The long form, DIXDC1a, but not the short form, DIXDC1b, directly binds to actin, with one DIXDC1a binding every 200 actin monomers localizing it to focal adhesions at the tips of the cell. DIXDC1a binding to actin requires amino acids 127 to 300 of DIXDC1a ; ho wever, it does not require the DIX or MTH domains [ 94 ] The shorter DIXDC1b is 472aa coded by a 5.1kb mRNA, and has a separate regulatory domain from DIXDC1a. Unlike DIXDC1a, DIXDC1b does not bind to actin and is found diffusely throughout the cytoplasm. In humans, the DIXDC1b g ene consists of 16 exons and is about 45kb. The gene is linked to dihydrolipoamide S acetyltransferase (DLAT) in a tail to head manner, with less than 4kb between them [ 95 ] DIXDC1b is highly conserved in vertebrates. Zebrafish only have the shorter DIXDC1b ortholog, called Ccd1 which is 54% conserved with humans [ 87 ] The mouse gene, by contrast, has 3 main isoforms of DIXDC1, and the isoform most similar to DIXDC1b (Ccd1B) is 90.9% similar to the human gene [ 96 ] Mouse DIXDC1 consists of 25 exons over 77kb in chromosome 8. All three mouse isoforms of DIXDC1 Ccd1A, Ccd1B, and Ccd1C combine with Dvl, and all act along the same pathway to cause
45 TCF/LEF gene transcription. However, only independent transfection of Ccd1B, and not of Ccd1A or of Ccd1C, caused activation of TCF/LEF genes. In the following chapters, I will endeavor to uncover more about PTK7 and DIXDC1b through the study of aptamers.
46 CHAPTER 2 PTK7 APTAMERS HAVE AN ANALOG IN GENOMIC DNA Introduction Cell SELEX generates artificial DNA and RNA molecules (aptamers) tha t bind to biological targets of interest with applications for research and therapy. Using this tool, our lab DNA aptamer that to breast [ 98 ] and colon [ 4 ] cancer cell lines In the process of cataloging colon cancer aptamers he selected, Kwame Sefah noticed several aptamers had the same binding profile as sgc8c. These three aptamers KC2D8, KC2D4, and KDED19 all bound the same cell types with similar affinities, and competed wi th each other for binding [ 4 ] I was intrigued by Kwame sgc8c. To my surprise CLUSTALx alignment showed sgc8c and KC2D8 share 38 contiguous nucleotides (nt). As previously discussed, sgc8c was selected by whole cell SELEX for a target T Cell leukemia cell line, CEM CCRF, and against a B Cell leukemia cell line, Ramos [ 7 ] The second aptamer, KC2D8, was selected by a Kwame Sefah several years later using a colon cancer cell line, DLD1, as the target, and no negative cell line [ 4 ] These tw o selections used two different primer pairs to PCR amplify the portion of the library that binds to cells after washing [Table 2 1]. The sgc8 aptamer cannot prime the KC2D8 library, making the likelihood of contamination by sgc8c during KC2D8
47 selection re mote. Starting with this initial observation, I was curious how and investigated why these two selections identified a nearly identical aptamer. Results and Discussion 38nt sgc8c Binds as well as 41nt sgc8c Sgc8c is the most commonly used aptamer in the We ihong Tan lab. It was one of the first aptamers selected by the lab. It started as an 88nt aptamer and was optimized by Dihua Shangguan to a shorter 41nt aptamer. As KC2D8 and sgc8c shared only 38nt of 41nt in common, I predicted these extra bases would be superfluous, and that an aptamer consisting of only shared bases would have the same binding ability as the 41nt aptamer. I synthesized different aptamers that incorporated different mixtures of the 3 bases not found in common between the two sequences, and found the different sequences derived from sgc8c and varying in length from 38 41nt all had comparable K d s [Table 2 1, Figure 2 1], indicating the three bases not shared between KC2D8 and sgc8c were not necessary for binding. In the future, all labs sh ould save resources, and use the 38nt sgc8c. Shortening the aptamer to 38nt would use 7% less resources to synthesize but wou ld not sacrifice aptamer quality. Table 2 1. 31nt sgc8c binds as well as 41nt sgc8c. Values averaged from two experiments with th ree replicates each. Bold, underlined bases are variable, depending on the sequence. K sgc8 41 AT CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG A 0.8 0.1 sgc8 38 CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG 0.9 0.9 sgc8x CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG A 1.4 0.4 xsgc8x T CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG A 0.5 0.1 xxsgc8 AT CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG 0.5 0.1
48 Figure 2 1. Hyperbolic fitting of data collected at different aptamer concentrations. Data is based on triplicate measurements of CEM cells. For each fitting R > 0.96. Bioinformatic Analysis of cell SELEX Aptamers Looking beyond saving resources by shortening the sgc8 sequence, the finding that two aptamers, selected by different researchers, on different cell lines, years apart, shared 38nt puzzled us. The selections that produced sgc8c and KC2D8 used initial libraries with different primers. Also, each selection started from a large pool of potential targets because, neglecting lipids and peripheral proteins and considering integral proteins alone, 20 30% of human proteins have a transmembrane domain [ 99 ] and a significant portion of these proteins are exported to the cell surface. This means upwards of 10,000 different proteins are expressed on each cell's surface. Moreover, each of these proteins has more than one unique domain serving as a potential ap tamer target, much as there are typically several antibodies or aptamers that recognize different regions called epitopes on the same protein. These proteins can be further modified by carbohydrates, creating many more sites for possible aptamer interactio n
49 and subsequent selection. An ovarian cancer selection that produced aptamers insensitive to proteases has been postulated to bind these glycoproteins [ 9 ] Furthermore, aptamers a re routinely selected against a wide array of single targets from small ions, to peptides, and purified proteins. A search of the literature did not uncover a single target that evaded selection. Building on this, we assumed that each potential target on t he cell surface had an equal probability of binding one of the sequences in the library, such that a selection would agnostically find targets [ 100 ] If this was the case, whole cell SELEX would select for a panel of aptamers that profile the entire range of proteins, carbohydrates, and lipids on a cell surface. Thus, the odds for repeat selection are remote because: (1) a large aptamer library is used for each selection, calculated to include 10 15 unique sequences; (2) there are a large pool of potential targets on each cell membrane surface; (3) and we assumed each potential target on the cell surface has an equal probability of binding one of the sequences in the library. The similarity between sgc8 and KC2D8 indicated one of these reasons did not hold. The first two were not in question; leading us to postulate the last assumption could be faulty. Instead of agnostically profiling the cell surface, perhaps cell SELEX could be preferentially sel ecting for aptamers with a biological function; for instance, those with an underlying affinity for DNA binding. Next I wondered if any other complex target selections had also found the same target and shared this common sgc8c KC2D8 motif. I compiled a d ataset of all ssDNA aptamers selected by complex selection or against membrane bound targets (Appendix A). This yielded 148 unique aptamer sequences from 33 different selections. The
50 ent, and the identity between them was determined by alignment with ESPRIT [ 101 ] Using this model, we made a set of identities with corresponding p va lues. From the results, two further aptamers, H01 and KMF9b, which had sequence identity to each other and to sgc8, were identified (Appendix B). When these sequences were aligned with sgc8c, their similarities were clustered around a core 15nt GC rich region: GCTGCGCCGCCGGGA [Table 2 2]. As predicted, all four aptamers -KC2D8, HO1, KMF9b, and sgc8c -competed with each other but not with a control, scrambled sgc8c sequence (Scr sgc8c) [Figure 2 2].
51 Table 2 2 Aptamer sequences showing sequence similarity. Black with white lettering shows consensus region. Note sequences have been shortened for clarity. Please refer to experimental section for full sequences. Aptamer Sequence Cell Target Primers Forward (F), Reverse (R) sgc8c [ 7 ] G GCGCAGC CTGC GCCGCCGGGA G CCTCCCTCCCAGTGGGAGAT G CEM (+) Ramos ( ) F ATACCAGCTTATTCAATT R GATAGTAAGTGCAATCT KC2D8 [ 4 ] AT CTAACTG CTGCGCCGCCGGGA AAATACTGTACGGTTAG A DLD1 (+) F ATCGTCCGCCACCACTACTC R GTGAGACTGCCTGCCGATGT KMF9b [ 102 ] A GCGCAGC AG CTG T GCC A CCGGGAG AATTTACGTACGGCTGAGCG A MCF7 10 AT 1 (+) F AGGCGGCAGTGTCAGAGT R CTGAGCGACGAAGACCCC H01 [ 103 ] AA GCAGC A G CTG T GCC AT CGGG T TCGGATTTTCTTCCTACGAC T CEM (+) F ATCGTCTGCTCCGTCCAATA T R TTTGGTGTGAGGTCGTGC
52 Figure 2 2. A) Competition studies between sgc8c and sgc8c, H01, KMF9b, and KC2D8. Scr sgc8c: Scrambled sgc8c, a negative control. Apparent K for each From the competition experiments [Figure 2 2a], it was clear that H01 was competed off of the cell surface faster than the other aptamers; because addition of sgc8c first, then H01 B (purple line) reduced the binding to background levels, but addition of H 01, then sgc8 B (dark blue line) or H01 then H01 B (light blue line) reduced binding, but at a much lower level. This means, if unlabeled sgc8c binds PTK7 first, H01 is not able to bind the PTK7. By contrast, H01 is readily replaced by both biotinylated sg c8c and 10x unlabeled H01. To investigate the off rate, a competition experiment where biotinylated aptamer was bound, washed off, and replaced by 10x unlabeled aptamer for varying periods of time was performed. After 120min the signal for H01
53 decreased b y 75% compared to less than 15% for each of the other aptamers, showing H01 has a faster off rate than the other PTK7 aptamers. Statistical Analysis of Consensus Region To determine whether the sequence identity between the aptamers could occur by chance, Yunpeng Cai, a bioinformatician, and I developed a bioinformatic simulation to determine the probability that four aptamer sequences, selected at random, would have the same level of identity. We performed 1 million random simulations to determine the probability that a random pair of sequences would have the same probability value, p, as the alignment between sgc8c and KC2D8. The overall probability of coincidence is P = n ( n 1)/2* p, where n is 1 million. With these simulations, we confirmed that the pair sgc8c KC2D8 is a confidently non coincident match, with significance level P sgc8 KC2D8 <0.01. Next, we did a random simulation with the third sequence, KMF9b. The overall probability between these three sequences is P=2*(1 P sgc8 KC2D8 )*( n 2)* p KMF9b We found KMF9b is matched to the pair sgc8c KC2D8, with significance level P KMF9b <0.03. Similarly, sequence H01 is matched to the above triplet at significance level P H01 <0. 05. Thus, we concluded that each of the 4 sequences matching with the others cannot be explained by coincidence. Consensus Region is Important to Aptamer Binding Although we found a statistically significant, conserved region between these four aptamers, we did not know if this region was relevant to how the aptamer bound PTK7. To determine if the consensus sequence had a larger influence on binding than sequences outside this region, we performed mutational, blocking, and structural probing of one of these PTK7 aptamers, sgc8c.
54 Mutational analysis Azobenzene phosphoramidites are DNA nucleoside analogs that have an azobenzene in place of the standard A, T, G, or C base [Figure 2 3]. Mutating aptamers with these bases perturbs the aptamers' secondary structure immediately surrounding the mutation, potentially having an effect on binding. I hypothesized that aptamer mutants with affected binding, determined by K d wou ld have mutations in key aptamer protein interacting nucleotides. Figure 2 3. Azobenzene nucleotide base analog. A) Phosphodiester backbone of the nucleoside. B) Azobenzene conjugated in place of a standard base. Under the aegis of Joseph Phillips and Ha ipeng Liu, we synthesized a number of sgc8c sequences with azobenzene mutations at different locations, and determined their K d with CEM cells. Sgc8c mutants with azobenzene bases inserted within the consensus region, the region shared by the aptamers, but not those with azobenzene mutations inserted outside that region, had significantly decreased aptamer binding ability, as measured by their K d [ 104 ] [Table 2 3]. This indicated that the consensus region is more important than other aptamer regions, including the stem and loop of sgc8c, for binding to CEM cells.
55 Table 2 3 Azobenzene mutated sgc8c sequences Black with white lettering shows consensus region. Z indicates an azobenzene insertion. One way ANOVA on K group gave a p value < 0.05. sgc8c Mutations Sequence with Azobenzene Mutations ( Z ) K (nM) No Azobenzene AT CTAACTG CTGCGCCGCCGGGA AAATACTGTACGGTT AG A 2 A zobenzene( Z )outside consensus sequence A Z T CTAACTG CTGCGCCGCCGGGA AAATACTGTACGGTT AG A 4 AT CTA Z ACTG CTGCGCCGCCGGGA AAATACTGTACGGTT AG A 5 AT CTAACTG CTGCGCCGCCGGGA A Z AATACTGTACGGTT AG A 2 AT CTAACTG CTGCGCCGCCGGGA AAATACTGTAC Z GGTT AG A 132 A zobenzene ( Z ) within consensus sequence AT CTAACTG CTG Z CGCCGCCGGGA AAATACTGTACGGTT AG A 1496 AT CTAACTG CTGCG Z CCGCCGGGA AAATACTGTACGGTT AG A 2429 ATCTAACTG CTGCGCCGCC Z GGGA AAATACTGTACGGTT AGA 523 Consensus sequence blocking To determine the region of sgc8c most important for binding in another way, a visiting scholar, Jian Wang, and I, blocked the sgc8c consensus sequence with different lengths of complementary DNA (cDNA). Aptamers are ssDNA; addition of a complementary seque nce to the aptamer sequence causes the aptamer and its complementary sequence to hybridize. This hybridized DNA does not have the same secondary structure as the single stranded aptamer. In order to bind a target, the aptamer binding confirmation must be m ore energetically favorable than the hybridized confirmation. The aptamer must be able bind in spite of having a piece of it hybridized, altering its secondary structure. If an important region for aptamer binding is blocked by cDNA, then it cannot form t he proper secondary structure and bind its target. If, however, the cDNA blocks a region unimportant for binding, then the remaining unblocked aptamer can form enough proper secondary structure to act as a toe hold and displace the blocking cDNA from the a ptamer [ 105 ] Therefore, we made different lengths of cDNA, complementary to the
56 Figure 2 4. Blocking consensus region with complementary DNA, blocks binding. A) Sequences used in these experiments. All complementary sequences are aligned with the bold faced sgc8 sequence they hybridize to. sequence is its lengt h. The putative consensus sequence in sgc8c is in red. B: biotin. C 41: complementary to the entire sgc8c sequence, a positive control. Scr sgc8c: scrambled sgc8c, a negative control. B) C 41, shifts binding all the way back. C) Complementary sequences th at block the consensus sequence shift sgc8c binding 14, 19, and 22). D) Complementary sequences that do not block the consensus sequence do not shift 19 which blocks one nucleotide of the consensus sequence.
57 cytometry [Figure 2 4]. The positive control was a strand entirely complementary to sgc8c, C 41. The sgc8 that was hybridized with C 41 shifted aptamer binding all the way back to the level of the negative control, scr sgc8c [Figure 2 14nt, 19nt, or 22nt cDNA resulted in a 10 fold decrease in aptamer binding [Figure 2 4c]. Th 19 showed a small shift backward [Figure 2 19 cDNA blocks 1nt of the consensus s sequence, was the most important region for binding. Sgc8c pyrene mutants suggest sgc8c interacts i ntimately with PTK7 To more specifically probe the interactions between the protein and the aptamer, 3 pyrene phosophoramidites sgc8c mutants were synthesized [Figure 2 5]. When UV light is shone on these sequences, the pyrene dye is excited, and electron transfer occurs. This electron transfer has enough energy to break disulfide bridges and am ide bonds that are within a 2nm radius of the pyrene [ 106 ] Mingxu You and I hypothesized that pyrene added to aptamers within the consensus sequence would be in close enough physical proximity to amino acids of the PTK7 target that UV irradiation could break important bonds in the protein. If this occurred the PTK7 aptamer binding pocket would be disrupted, leading to decreased binding of tho se pyrene sgc8c mutants to PTK7.
58 Our previous mutational analysis of sgc8c with azobenzene, discussed above, showed that insertion of a non standard base directly into the consensus region caused decreased binding. To avoid this, we inserted pyrene mutati ons either at the cusp of the consensus sequence (red star in Figure 2 5c) or 1 base away from the consensus region (green star, same figure). Thinking perhaps multiple pyrene mutations in the stem could yield a more intense effect, even though they are no t in the consensus region, we added 5 As expected, the pyrene mutations alone did not affect the binding of the probes, which all bound as well as unmodified sgc8c to CEM cells [Figure 2 6a]. Furthermore, only the mutant with pyrene inserted closest to the consensus sequence showed a significant decrease in aptamer binding to the target cells [Figure 2 6 A.IV and B]. The other two mutants, with pyrene inserted in the loop or stem, had no significant change [Figu re 2 6 A.III, A.V, and B.]. To ensure that UV alone was not damaging the cells or the DNA probes, all cells and aptamers received the same 15min dose of UV, regardless of their treatment group. Also, to ensure that internalization of the probes, which woul d decrease their abundance of the probes on the surface of the cells, and that heating from the UV lamp, did not interfere with our measurements, the entire protocol was performed at 4C. The cells appeared healthy under the microscope with trypan blue sta ining, and when Scr sgc8 was used as a negative control, no non specific binding was seen [Figure 2 6 A.I]. When pyrene is excited by the UV lamp, electron transfer occurs in a localized environment, within ~2nm. Amide and disulfide bonds within this radi us might possibly be disrupted. Disruption of the protein aptamer binding site could weaken protein
59 aptamer interactions, and decrease the signal seen by the flow cytometer. According to our results, there was a small but significant change in the binding ability of the pyrene sgc8c mutant adjacent to the consensus region. The most likely reason for this change is that, in this region, the consensus sequence is in close physical proximity to the protein it binds. Figure 2 5. Pyrene sgc8c mutants. A) Stru cture of pyrene nucleoside DNA base analog. I. Phosphodiester backbone. II. Pyrene base conjugated in place of standard sgc8c structure, above right, Figure 2 5c. Pyrene bases are sh own in red. Consensus sequence is bold underlined. Dashes are inserted for clarity, they do not represent bases. C) NuPACK predicted structure of sgc8c with colored stars representing pyrene insertion locations. Dashed line inside loop indicates consensus region.
60 Bioinformatic and competition analysis of aptamer sequences identified 4 aptamer sequences that shared a consensus region: GCTGCGCCGCCGGGA. Mutational analysis of sgc8c with azobenzene, blocking with sgc8c cDNA, and UV mediated pyrene degradation of the protein aptamer interaction, all showed that the region of sequence identity shared between sgc8c, KC2D8, KMF9b, and H01 aptamers was also the region most important for binding. Thus, I concluded the 15nt consensus region of these PTK7 aptamers is crucial for binding. BLAST of the Consensus Sequence against the Human Genome I hypothesized that these 4 independently selected, yet competing, aptamers might be mimics for a naturally occurring interaction between the PTK7 protein and natural DNA. If th is was the case, the consensus sequence should be found in a target sequence in the DNA. To investigate this possibility, I blasted the consensus sequence against the human genome using the NCBI nucleotide BLASTn algorithm, adjusted for short sequences. Th is returned 8 hits with 14/15nt identity, detailed in Table 2 4. Six of DIXDC1b has the Consensus Sequence Looking over this list, one of thes e results, DIXDC1b, immediately peaked my interest because, like its name suggests, it has a DIX domain. DIX domains are very rare. Aside from DIXDC1, only two other proteins have them: Axin and Dvl. As is extensively discussed in the Introduction, these p roteins are central in Wnt signaling, as
61 Figure 2 6. A) Representative data from one flow cytometry experiment. Scr sgc8: scrambled sgc8c, a negative control. B: biotin. Sgc8c: another negative control. All probes and cells received 15min of UV irradiation at 4C. B) Bar graph showing the ratio of fluorescence obtained by flow cytometry of UV irradiated probe only versus UV irradiated probe and cells together. p value test. Data is aver aged from 4 separate experiments with 3 replicates each.
62 Table 2 4. BLAST hits 14/15nt identity for GCTGCGCCGCCGGGA in human genome. Protein Name, Abbreviation Locus Location in Gene DIX Domain Containing 1, isoform b, DIXDC1b NM_033425.3 Mucolipin 1, MCOLN1 NM_020533.2 Coding region Membrane bound transcription factor peptidase, MBTPS1 NM_003791.2 Intron Ubiquitin conjugating Enzyme E2D2, UBE2D2 NM_003339.2 Intron Myc Induced Nuclear Antigen, MINA NM_032778.4 Intron SH3 and multiple ankyrin repeat domains protein 1, SHANK1 NT_011109.16 Intron Nearest protein KIAA1875 NR_024207.1 Unknown Nearest miRNA MIR302F NW_001838467.2 Unknown Recall from the Introduction that Dvl interacts with PTK7, the target of sgc8c, catenin pathway toward the Wnt/PCP pathway. PTK7 interacts with Dvl through its PDZ domain. DIXDC1 and Dvl interact through th eir respective DIX domains, and DIXDC1 causes Dvl to depolymerize from large aggregates into smaller, 4 protein heteromers, which are biologically active. This depolymerization makes Dvl available for signaling. Furthermore, Dvl DIXDC1 heteromers serve to inhibit JNK signaling, genes that are turned on by the Wnt/PCP pathway. Accordingly, DIXDC1 has the opposite effect of PTK7 on cell signaling. signal, driving the expression of T catenin pathway, while simultaneously inhibiting Dvl expression of the pro Wnt/PCP JNK genes. Therefore, PTK7 and DIXDC1 have opposite effects PTK7 switches Wnt catenin dominant to being PCP dominant, whereas DIXDC1 catenin dominant. These two proteins are intimately involved in the same signaling pathway; therefore, as with
63 many other signaling biofeedback loops within Wnt and othe r pathways, it would not be [ 107 ] Aligning the PTK7 aptamers with DIXDC1b DN A revealed a 1nt mismatch with sgc8 and KC2D8, 3nt mismatches with KMF9b, and 5nt mismatches with H01. Although H01 and KMF9b had more mismatches within the consensus region, they surprisingly both had additional bases in common with the DIXDC1b DNA adjace nt to the consensus sequence: 5 additional nts for H01, and 8 additional nts for KMF9b [purple in Table 2 5]. Also of interest, H01 had a 10 fold larger K d than KMF9b, indicating decreased binding affinity, and H01 also has 3 fewer bases in common with DIXDC1b DNA than KMF9b. The combined 22nt region in common between the PTK7 aptamers and DIXDC1b DNA is unique in the human genome. KC2D4 has Identity with DI XDC1b Negative Strand In addition to the 4 sgc8 type aptamers previously mentioned, there are 2 other aptamers, KDED19 and KC2D4, that had been shown to compete with sgc8 [ 4 ] Neither of these two sequences shared significant identity wit h each other, the other Table 2 5 Aptamers share sequence similarity with DIXDC1b DNA sequence. The sequence over the red line is the consensus sequence used for BLAST. Pink with white lettering denotes aptamer sequence similarity to DIXDC1b DNA within the consensus region. Purple with white lettering denotes sequence identity outside of the consensus region. Aptamer Sequence K d (nM) DIXDC1b G GCGCAGC CTGC GCCGCCGGGA G CCTCCCTCCCAGTGGGAGAT G n/a sgc8c AT CTAACTG CTGCGCCGCCGGGA AAATACTGTACGGTT AG A 0.8 0.1 KC2D8 TA CT AACTG CTGCGCCGCCGGGA AAATACTGTACGGTTAGTT 1.1 0.1 KMF9b A GCGCAGC AG CTG T GCC A CCGGGA G AATTTACGTACGGCTGAGCG A 0.3 0.1 H01 AA GCAGC A G CTG T GCC AT CGGG T TCGGATTTTCTTCCTACGAC TGC 4.0 0.3 PTK7 aptamers, the 146 other sequences in our SELEX database, or with the positive
64 identity with the negative strand of DIXDC1b DNA, five bases downstream from the sgc8 consensus sequence [Table 2 6]. Yunpeng Cai and I p erformed a simulation to determine the probability that this level of sequence identity 12/13nt between DIXDC1b DNA and KC2D4 was the result of chance. For this simulation, we took 1 million 39nt long random sequences (the same length as KC2D4), and aligne UTR of DIXDC1b DNA. Our simulation showed that 2,554 sequences out of 1 million sequences, or their reverse complement, have at least 1 13bp window that shares 12nt identity with the extracted DNA segment, which is the sa me similarity as the KC2D4 aptamer. Hence, the chance that a random sequence has the same similarity to the target region is p =0.0025.
65 Table 2 6. KC2D4 has identity to the negative DIXDC1b strand. Blue with white lettering is KC2D4 identity with th e negative strand of DIXDC1b DNA. As in Table 2 3, Pink with white lettering denotes aptamer sequence similarity to DIXDC1b DNA within the consensus region. Purple with white lettering denotes sequence identity outside of the consensus region. p value for KC2D4 alignment with DIXDC1b DNA is p = 0.0025. Aptamer Sequence K (nM) DIXDC1b DNA + G GCGCAGC CTGCGCCGCCGGGA G CCTCCCTCCCAGTGGGAGAT GGGTTGAGA CCGCGTCG GACGCGGCGGCCCTCGGAGG GAGGGT C ACC C TC TACCCAACTCT n/a KC2D4 GAGGGT G ACC A TC GGTAAGGCGGGAATTGGCCCGGTAGC 54.3 7.9
66 Conclusions The 14nt identity among the 15nt consensus sequence is quite unusual, with a probability of 1.5 x 10 8 or one occurrence in 33.6M bps. This estimate predicts we should find 94 other instances of our 14/15nt sequence in the human 3.1647 x 10 9 bp genome. We only found this level of sequence identity for our consensus region 8 times. This discrepancy could be explained by the high number of repeated sequences throughout the human genome or because the BLASTn search only looks for contiguous bases. Unfortunately, the size of the human genome is so large that we are not able to enumerate all of the possible 14nt permutations. Hence, the coincidence cannot be judged by statistical inference, unless only a small part of the human genome is involved in whole cell SELEX procedure, which we cannot assume. Nevertheless, our simulations discourage the idea that these 5 aptamers share sequence similarity with DIXDC1b DNA through coincidence alone, because, if the 22nt sequence in common between the PTK7 aptamers and DIXDC1b DNA is a coincidence, then the KC2D4 DIXDC1b DNA match should also be a coincidence, as they would have happened independently. It is statistically improbable that both of these two coincidences would occur at the same time. Hence, if the PTK7 DIXDC1b DNA match is not a coincidence, then the KC2D4 DIXDC1b DNA match is also unlikely a coincidence. These simulations do not take into account that all 5 of these aptamers compete wit h each other for binding to the same site on the cell, or that the consensus region is the most important aptamer region for binding. If we take these experimental results into account, the likelihood that this confluence of data is just a random coinciden ce becomes very, very remote. Therefore, we concluded that if our observations were not a product of chance, there must be some underlying biologically relevant reason for a
67 DNA aptamer having an analog in genomic DNA. Using this as the basis for our reaso ning, we began to look for possible scenarios where PTK7 protein could interact directly with DIXDC1b DNA. These explorations will be detailed in the next chapter. Materials and Methods DNA Sequences All aptamer sequences, shown in Table 2 7, were synthesi zed in house at 1 scale synthesis using an ABI 3400 DNA/RNA synthesizer (Applied Biosystems). Special bases such as the azobenzene and pyrene were dissolved in dry acetonitrile before coupling for twice the time of standard bases on the synthesizer. Sequences were deprotected from CPG beads with AMA (ammonium hydroxide: methylamine 1:1) at 50C for either 20min or 12h for the azobenzene modified CPG (Glen Research) modifications, were purified using Glen Pak columns (Glen Research). Aptamers that included special bases such as azobenzene or pyrene, or that were shorter than 30nt in length were ethanol precipitated, resuspended in 1M TEAA and purified on a reverse pha se Prostar HPLC (Varian) using a C18 column (Econosil, 5U 250 x 4.6 mm from Alltech Associates) with a linear elution. These sequences were vacuum dried or ethanol precipitated, detritylated with 20% acetic acid, and stored at minus 20C for future use. T he absorbance at 260nm for the purified aptamers was quantified using a UV Vis spectrophotometer (Bio Rad). Aptamer concentration was determined using the Lambert cuvette pass length, and c is the DNA concentration. All sequences were tested for
68 (89mM Tris HCl, 89mM boric acid, 2mM EDTA, pH 8.0) for 40min at 120V and imaged under UV with Ethid ium Bromide (EtBr). Cell Culture CCRF CEM (CCL 119, a T cell line, Acute Lymphoplastic Leukemia) and Ramos (CRL 1596, a B Table 2 7. All DNA sequences used in this chapter. B: Biotin; UL: unlabeled; Z :Azobenzene; (T) 10 B: TTTTTTTTTT P : Pyrene. Experiment Name sgc8c Optimization sgc8c 41 ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B sgc8c 38 CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG B sgc8cx CTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B xsgc8cx TCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B xxsgc8 ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG B Competition and K d Experiments sgc8c B ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B sgc8 no ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA KC2D8 B ATCGTCCGCCACCACTACTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTT AGTTTGAGACTGCCTGCCGATGT B KC2D8 UL ATCGTCCGCCACCACTACTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTT AGTTTGAGACTGCCTGCCGATGT KMF9b B AGCGCAGCAGCTGTGCCACCGGGAGAATTTACGTACGGCTGAGCGA B KMF9b no AGCGCAGCAGCTGTGCCACCGGGAGAATTTACGTACGGCTGAGCGA H01 B AAGCAGCAGCTGTGCCATCGGGTTCGGATTTTCTTCCTACGACTGC B H01 no AAGCAGCAGCTGTGCCATCGGGTTCGGATTTTCTTCCTACGACTGC Scr sgc8 no ACATGGAGCGTCCAAATCGAGTCGGATATCACGTGCTCGAT B Scr sgc8 no ACATGGAGCGTCCAAATCGAGTCGGATATCACGTGCTCGAT TD05 B AACACCGTGGAGGATAGTTCGGTGGCTGTTCAGGGTCTCCTCCGGTG B TD05 no AACACCGTGGAGGATAGTTCGGTGGCTGTTCAGGGTCTCCTCCGGTG sgc8c Azobenzene Mutational Analysis Stem1A 10T B A Z TCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA (T) 10 B Stem1B 10T B ATCTA Z ACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA (T) 10 B Loop1C 10T B ATCTAACTGCTGCGCCGCCGGGAA Z AATACTGTACGGTTAGA (T) 10 B Stem1D 10T B ATCTAACTGCTGCGCCGCCGGGAAAATACTGTAC Z GGTTAGA (T) 10 B Stem1C 10T B ATCTAACTGCTG Z CGCCGCCGGGAAAATACTGTACGGTTAGA (T) 10 B Loop1A 10T B ATCTAACTGCTGCG Z CCGCCGGGAAAATACTGTACGGTTAGA (T) 10 B TD05 10T B AACACCGTGGAGGATAGTTCGGTGGCTGTTCAGGGTCTCCTCCGGTG (T) 10 B
69 Table 2 7. Continued Experiment Name cDNA Blocking Experiments sgc8c B ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGATTT B 14 TAGATTGACGACGC 19 TAGATTGACGACGCGGCGG 22 TAGATTGACGACGCGGCGGCCC 11 CATGCCAATCT 15 ATGACATGCCAATCT 19 TTTTATGACATGCCAATCT C 41 TAGATTGACGACGCGGCGGCCCTTTTATGACATGCCAATCT Scr sgc8 B AACACCGTGGAGGATAGTTCGGTGGCTGTTCAGGGTCTCCTCCGGTG B Pyrene UV disruption sgc8c B ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B sgc8c stem B A P T P C P T P A P ACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B sgc8c cons B ATCTAACT P GCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B sgc8c loop B ATCTAACTGCTGCGCCGCCGGGAA P AATACTGTACGGTTAGA B Type Culture Collection (ATCC). Both cell lines were cultured in RPMI 1640 medium (GIBCO) with 10% FBS (GIBCO) at 37C under 5% CO 2 100 units/mL penicillin streptomycin (Cellgro) were added to the cultures, occasionally, if bacterial contamination was suspected in the lab. Cultures were routinely monitored for Mycoplasma infection. Immediately before experiments, cells were collected by centrifugation at 1,260 RPM for 3min and washed with 2mL ice cold Washing Buffer (WB: PBS, 4.5g/L glucose, 1M MgCl 2 ). The cells were resuspended in Binding Buffer (BB: WB with 1g/L bovine serum albumin, and 100mg/L tRNA) at a concentration of 20 x 10 6 cell s/mL and placed on ice. For each experiment 1 x 10 6 cells were used. These conditions remained constant for all experiments unless otherwise noted.
70 Bioinformatics We compiled a dataset of 148 unique aptamer sequences from 33 different selections, removed t heir primers, and aligned them with ESPRIT (Appendix A). Using the convergent property of SELEX [ 108 ] the p value p obtained in our simulation i ndicates that, given a target sequence S1 a reference sequence R, which has a known similarity value t to S1 and a random sequence S2, with the same length as R the probability that S2 is at least as similar to S1 as R is p Under this non deterministic model, we supposed we have n (here n ~10,000) aptamers in the system. The probability that a random aptamer can be more similar than R is 1 (1 p ) n If we examine k targets (here k =148), the probability of one random hit will be 1 (1 p ) nk which is approxi mately nk*p when p is small. Hence, if nk*p << 1 we claim that R is not likely a random match. This means that R should be identical, or at least correlated to S1 In this case, we had k *( k 1)/2 random pairs in total where k =148. For a given pair, the chance of one random hit is 1 (1 p ) k ( k 1)/2 If k(k 1)p /2 << 1, we claim that the pair is not a random match, but either identical or correlated. From the results, sgc8c and KC2D8 had the greatest alignment. Two other aptamers, H01 and KMF9b, which had se quence identity to each other and to sgc8, were identified. Based on this result, a GenBank BLASTn [ 109 ] search was performed on the following consensus region: GCTGCGCCGCCGGGA. Significance Simulations We performed random simulations to find the probability that a random pair of sequences a and b resembles the similarity is P If we have n sequences, then we have n ( n 1)/2 pairs. Hence, the overall probability of coincidence is n ( n 1)/2* P With this
71 criterion, we confirmed the pair, sgc8c KC2D8, is a confidently non coincident match, with si gnificance level P 0 <0.01. Suppose we have found a matched pair a b with probability 1 P 0 that the match is not by chance, and we have found a third sequence c which shares some degree of similarity with a or b while random simulation showed the probabili ty of resembling this similarity is P 1 Then, the overall probability of coincidence is 2*(1 P )*( n 2)* P With this criterion, we find that sequence KMF9b is matched to the pair sgc8c KC2D8, with significance level P < 0.03. Similarly, sequence H01 is matched to the above triplet at significance level P <0.05. We can conclude that each of the 4 sequences matching with the others cannot be explained by coincidence. Furthermore, the 22nt common region between the 4 PTK7 ap tamers and DIXDC1b DNA is unique in the human genome. We examined the probability that another random sequence can match the same DNA in adjacent regions. We carried out a simulation by generating 1 million random sequences 39nt long, the same as KC2D4, an UTR of DIXDC1b DNA. Because local alignment is required to measure the similarity between the sequences and the DNA, we used FASTA with the following command: fasta35 a z 1 n q H b 1 d 1 O
72 aptamer treatments were as follows: no aptamer; 80nM scrambled sgc8 biotin; 80nM aptamer biotin; 800nM unlabeled aptamer then 80nM aptamer biotin; 800nM unlabeled sgc8c then 80nM aptamer biotin; and 800nM unl abeled aptamer then 80nM sgc8c biotin. For each experiment Ramos cells, which do not bind PTK7 aptamers, were used as a negative control. Off Rate Experiments CEM cells were incubated with biotinyl ated aptamers (125nM) for 20min. The cells were washed and then incubated with 10x of the same, unlabeled aptamer 120 m 60 m 30 m 20 m 10 m 6 m, 1m Cells were then washed and probed with 1:400 PE streptavidin and measured by flow cytometry. The experiment was carried out at 4 C to prevent aptamer internalization. Kd Measurements The apparent equilibrium dissociation constant (K ) for the aptamer cell interactions was determined by adding varying concentrations of biotin aptamer to 1x10 CEM CCRF cells in BB at 4C. Cells we re washed and incubated with 1:400 streptavidin PE (Invitrogen), washed again, and analyzed by flow cytometry, counting 20k cells. The mean fluorescence intensity of background binding with scr sgc8c aptamer or TDO5 aptamer (for the Azobenzene measurements ) was subtracted for each aptamer concentration. Data from 2 or 3 separate experiments with 3 replicates was averaged. Intensity and aptamer concentration were fit to the hyperbolic equation Y= B max X/( K +X), using Origin8.5 to measure K Azobenzene Mutation Measurements Azobenzene phosphoramidite was synthesized according the published protocol [ 104 ] K
73 For each aptamer tested, Ramos cells, which do not bind sgc8 c, were used as a negative control. cDNA Blocking Sgc8c B aptamers were hybridized by mixing equal molar equivalents of the various cDNA in H 2 0, which was heated to 95C for 5min and allowed to slowly cool to room temperature. Hybridization was confirmed by running a small aliquot on a 3% agarose gel in TBE for 40min at 120V and noting a shift in molecular weight. The sequences were then tested for binding with CEM cells similar to the competition experiments described above, using 250nM concentrations of probe. For each experiment Ramos cells, which do not bind sgc8, was used as a negative control. Pyrene sgc8c Binding For the photo regulation of the pyrene sgc8c, cells (1 x 10 6 ) and aptamer d at 302 nm (SANKYO DENKI, Japan) with a 352 nm optical filter (3 nm half bandwidth; Oriel Instruments, Stratford, CT, Newport). Aptamers and cells were separated into 3 groups: aptamer with UV irradiated cells; UV irradiated aptamer with UV irradiated ce lls; and UV irradiated cells pre bound to aptamer. The aptamers and cells (except for the aptamer that received no UV), either separately or together, all received a total of 15min UV irradiation. UV irradiation and aptamer binding was all conducted at 4C Cells were stained afterward with trypan blue and checked under the microscope to ensure the UV irradiation did not cause cell death during the experiment. For each experiment Ramos cells, which do not bind sgc8, was used as a negative control.
74 CHAPTER 3 POSSIBLE FUNCTIONAL ROLE FOR PTK7 DIXDC1B INTERACTION Introduction Our lab focuses on whole cell SELEX of DNA libraries against cancer cell lines, identifying aptamers to target specific tumor types. After many of these selections, we noticed a curious phenomenon: aptamers from different selections competed against each o ther for the same target, a membrane protein tyrosine kinase, PTK7. These selections involved different cell lines, DNA pools, and primers. They were performed years apart by different researchers. The chances of selecting aptamers against the same target multiple times were remote. Why was this happening? We hypothesized there might be a natural, physiological role for these sequences. We whittled down the smallest region of nucleotide consensus between the aptamers, and BLASTed it against the human genom e. Surprisingly, we found the consensus sequence appears in the regulatory region for the protein, DIXDC1b, which Moreover, aligning the aptamer sequences with this regula tory region, we found more bases in common than just the predicted consensus sequence. This family of PTK7 binding aptamers has significant sequence identity to both the positive and negative strands of the DIXDC1b DNA, implying PTK7 may interact naturally with this regulatory region, and explaining why we repeatedly selected aptamers for this protein.
75 Results and Discussion So if what we have found is not a coincidence, what is happening? Why have our DNA aptamers been found in the genome? To help answer these questions we needed to determine where in the DIXDC1b gene the aptamer sequence identity was found. DIXDC1 protein i n humans is coded by two main isoforms with separate regulatory regions. There is a deeper discussion of DIXDC1 in the Introduction. One isoform, DIXDC1a, is longer. It has an extra 200aa N terminal region with a CH domain and another domain that lets it b ind to actin. DIXDC1a is found in many adult tissues, and it localizes with actin to focal adhesions at the edge of the cell. The second isoform, DIXDC1b, lacks the CH and actin binding domains, does not bind actin, and is found diffusely throughout the c ytoplasm of cells, especially in neurons, during development. The region of PTK7 1]. The known site for translational regulation, but it can also be important for transcripti onal regulation. Therefore, it can act as a control region for mRNA expression. 1 ATCCGGAAGG TGGCACG GAG TGGGATCGCC GCTGG GGACT CGAG GCGCAG CCTGCGCCGC 61 CGGGAG CCTC C CTCCCA G TG G G AG ATGGGT TGAGATGCCC CC GCCAGGGG GGATGCCCGG 121 CACCGTGCGT CCGCG GAGGC CAAGATGCAG CGG CCAGGGG CCGGCAGCCT G CGAGGGGAG 181 GCAGCTTCCG CCGGG GCCGG GCTGCTGCAC AGT CTGAGCG GCCGGGACTG C GCGCTTCAG 241 AGCCTGGAGC ATCCC AGTCG CTGGGGCCGA GAC GCCG CCG CCGCCGCCGT TCCCG CTTTC 301 TCCCGCGAGC CGGGC CAGTA GCTTTGCTAG CTG GCCTTCC CGTGGAGGCG T TTTCCAGCC 361 CCAGCGCGGG GAGAC ATGCC TGAATTTGGG AGC GATGTGA CTCTCAGCCT C CCACTTCAC 421 CCGGGGACGC AGGCT TGCTG AAGCCCGAGA C AGGAGG GGG ACC ATG G GAG GGACGCAAGT Figure 3 DIXDC1b mRNA locus: NM_033425. Black with white lettering is sgc8c KMF9b type aptamer ATG methionine start site is in red. A rather strong Kozak sequence surrounds the start site: GGGACC ATG G [ 110 ]
76 The Consensus Region is Conserved among Species As both the DIXDC1 and PTK7 proteins are highly conserved across vertebrates, if there was an interaction between the protein PTK7 and DIXDC1 we might hypot hesize that it would be conserved. By BLASTn searches of various published genomes, this is indeed what we find. In the DIXDC1b putative regulatory regions (as determined by Transfac analysis) of human ( Homo sapiens ), chimpanzee ( Pan troglodytes ), mouse ( Mus musculus ), and rat ( Rattus norvegicus ), we find sequence identity with the aptamers [Table 3 1]. Zebrafish ( Danio rerio) also has PTK7 and DIXDC1 orthologs, but, unlike the mammals mentioned above, D. rerio only has one isoform of DIXDC1, which corresponds to DIXDC1b and is called Ccd1. There is no sequence identity between the putative regulatory region of the D. rerio gene Ccd1 and the aptamers. However, a 23nt sequence within a gene for a hypothetical pro tein called loc797129 was found in chromosome 12. Ccd1 is on chromosome 21. Nothing is known about this hypothetical protein, according the PubMed citation, it was found through automated computational analysis using GNOMON. D. rerio has ~331M base pairs in its non repetitive genome [ 111 ] ; the chance that a random sequence has 23bp in common is 1 .4 x 10 14. This means that, randomly, only 1 genome of ~1 million different genomes the same size of D. rerio would have this exact sequence. This result in D. rerio is intriguing, but it remains highly speculative that the hypothetical protein in D. rer io has anything to do with PTK7, as nothing is known about the potential protein. Of more importance to this discussion are the mammalian sequences that have retained some of the bases found in the regulatory regions, but not C1b orthologs. This indicates these sequences are
77 conserved, and lends credence to the supposition that the sequences play a functional role. Table 3 1. Consensus region is conserved among species. Blue with white lettering is KC2D4 identity with the nega tive strand of DIXDC1b. As in Table 2 3, Pink with white lettering denotes aptamer sequence similarity to DIXDC1b DNA within the consensus region. Purple with white lettering denotes sequence identity outside of the consensus region. P. troglodytes, M. mu sculus, and R. norvegicus sequences are all in the regulatory regions for DIXDC1b. D. rerio sequence is not. Species Sequence H. sapiens CGCAGC CTGCGCCGCCGGGA G CCT CC ---CTCCCA G TGG G AG ATGGG P. troglodytes CGCAGC CTGCGCCGCCGGGA G CCT CC ---CTCCCA G TGG G AG ATGGG M. CGGTGT CTG A GC TA CCGGGA CAGTACGCTT C A CCC TAA GG G AG CATCC R. norvegicus TGAATGAG GCGCCGCCGGGA GCCT GGCCTACCTTGCACTGGTCGAGTG D. rerio CGCAGC CTGCGCCGCCGGGA GCC CAGCGGGGCCTGCAGCGGGACGGGC DNA is Double Stranded, but our Aptamers are Single Stranded These aptamers are ssDNA, and their specific recognition is contingent on their secondary structure. If sgc8c and KC2D4 type aptamers are indeed a mimic for natural interactions between DIXDC1b DNA and PTK7, they need to interact with dsDNA. We modeled a 70nt region surrounding the aptamer consensus region with NuPACK DNA secondary structure prediction software. As expected, the region had apparent secondary structure with two prominent stem loop hairpins [Figure 3 2]. However, various attempts t o see if double stranded regions of the DIXDC1b DNA could bind with cells or compete with sgc8c showed no binding or competition [Figure 3 3]. We know there is a large amount of basal transcription machinery in place on the DNA during transcription that ca n melt small regions of dsDNA, allowing for a secondary structure to form. It is possible that some other factors are needed to allow the protein to bind to the DNA in situ Or, perhaps we have not identified the correct dsDNA region.
78 Figure 3 2. NuPACK structures for 70nt surrounding the aptamer consensus region in DIXDC1b DNA. A) DIXDC1b forward, positive strand. B) DIXDC1b reverse, negative strand. Pink line denotes 22nt region of aptamer identity to sgc8c KMF9b regions. Blue line denotes 15nt region of KC2D4 sequence identity. Bar on right shows strength of predicted structure. Red being most strong. Figure 3 3. dsDIX DNA does not bind CEM cells. Scr sgc8c: Scrambled sgc8c; B: biotin; DIXcs: DIX consensus sequence, 22nt sequence in common between th e KC2D8 and sgc8c aptamers and DIXDC1b DNA. dsDIX 1: Double stranded oligo of 22nt bases. dsDIX 2: Double stranded oligo of 40nt surrounding the consensus sequence.
79 PTK7 Structural Analysis The extracellular portion of PTK7 interacts with the PTK7 aptamer s. Thus, the region of PTK7 that interacts with the aptamers must have a protein structure that would be conducive to such protein DNA interactions. To determine where on the PTK7 extracellular portion the aptamers would be most likely to bind, we used Swi ss Modeler to model the extracellular region of PTK7, and analyzed that model for areas of likely DNA protein interaction. As detailed in the Introduction, the extracellular portion of PTK7 is made up of 7 Ig like domains. There are many important human transcription factor families that use an Ig Brachyhury T box proteins [ 112 ] sheets are held together by a disulfide cysteine bridge. Each Ig like fold has 8 strands, labeled A G in Figure 3 4a. While the individual structures of the 6 families of tran scription factors are unique to the families, there are several aspects that they all share. DNA is positively charged; consequently it has an attraction for the basic amino acids: lysine, arginine, and, under the right conditions, histidine. Each member o f the Ig like fold families of transcription factors interacts with DNA through basic residues on the A B loop, the E F loop, and the C Tail. An example of such an interaction is shown in Figure 3 4b. Here, the crystal structure of Runx1 (PDB: 1h9d [ 113 ] ), in complex with its dsDNA target, is shown with the important basic residues (shown as red stars) that determine its interaction with DNA.
80 Figure 3 4. Structural features of Ig like folds. A) Scheme showing an Ig like fold with each strand sheet and the blue letters (A, B, E, D) correspond to the other. The A B loop and E F loop are indicated by black arrows. A disulfide bridge is typically formed between the B and F strands. B) Crystal structure of Runx1 transcription factor, green, bound to DNA, brown backbone with blue green DNA bases). The A B loop, E F loop, and C tail are indicated by arrows. The red stars indicate basic residues important for binding. We modeled the 3 dimensional structure of the 7 extracellular Ig like domains of the PTK7 protein using titin (PDB 2nzi for Ig 6 7 and 3B43 for Ig 1 5) as a template [Figure 3 5a]. Then, using the classic Ig like fold architecture, and the common binding pattern of transcription fact ors with Ig like folds as a starting point, we tried to predict where PTK7 would most likely bind the aptamers. Of all the Ig like domains only 2 had a preponderance of basic residues: Ig7 and Ig3 [Figure 3 5b, c]. The Ig7 domain of PTK7 contains the MT1 MMP cleavage site discussed in the Introduction. The PTK7 receptor is cleaved at this site, releasing a 70kDa protein fragment that affects PTK7 signaling. As proteolytic cleavage at the Ig7 domain would leave the basic residue area
81 behind, still attached to the membrane, we do not predict that the PTK7 aptamers would bind here. Figure 3 like extracellular domains. The transmembrane domain and cytosolic pTK domain are not shown, but would be b elow the Ig7 domain. This protein fragment is ~70kDa. The purple X denotes the natural MT1 MMP cleavage site; the black X is the additional chuzhoi mutant MT1 MMP cleavage site. B) Detail of the Ig3 domain. C) Detail of the Ig7 domain. Basic residues are shown by red sheet is labeled. A list of basic residues and their location on the protein are shown to the right side of each model. The Ig3 domain also has many basic resides, including 4 arginine residues in the A B and E F loops that are important for other Ig like transcription factor DNA interactions mentioned above. Furthermore, when PTK7 is cleaved by MT1 MMP, the
82 Ig3 domain woul d remain with the cleaved PTK7 fragment. Therefore, we predict the Ig3 domain would be the most likely domain to bind the PTK7 aptamers. This prediction is tenuous, but may serve as a guide for future PTK7 mutation experiments that can pin point the actual site of PTK7 aptamer binding. PTK7 Has a Weak Predicted Nuclear Localization Sequence If PTK7 does indeed interact with the regulatory region of DIXDC1b DNA in the nucleus, it needs to first get into the nucleus. Although smaller proteins can diffuse thr ough the nuclear pore complex (NPC) relatively quickly, transport of proteins larger than 60kDa through the NPC by diffusion alone is very slow. Larger proteins therefore require a nuclear localization signal (NLS) to facilitate their nuclear transport. NL S can be specific amino acid sequences, typically a short string of basic residues, such as PPKKKRKV for the SV40 virus T structure. These NLSs interact with proteins found in the NPC and/or with special ized nuclear import receptors, which shuttle proteins from the cytoplasm into the nucleus. There are several algorithms developed in recent years that attempt to predict from a n NLS. Using two of these tools, PredictNLS [ 114 ] and NLStradamus [ 115 ] I searched the PTK7 protein for a predicted NLS. When searching for these sites I used another membrane tyrosine kinase, the protein epidermal growth factor receptor (EGFR), as a positive control. EGFR is transported to the nucleus using the NLS: RRRHIVRKRTLRR [ 116 ] If the alg PTK7. The PredictNLS did not find an NLS in PTK7, but it also did not identify the validated NLS in EGFR. NLStradamus did a better job identifying the EGFR NLS, and made a weak prediction for a NLS in the PTK7 protein (RPPHLRR). This predicted
83 I predicted for aptamer binding in the preceding section. Strangely, while this RPPHLRR motif is conserved in m ouse, chimpanzee, rat, and cow, it is deleted in its entirety from chicken PTK7, called KLG. KLG and PTK7 share 70% homology over the extracellular region. This is the only region deleted between PTK7 and KLG extracellular domains, and it is the longest co ntiguous stretch of difference between the two proteins. From this prediction analysis, PTK7 does not contain a validated NLS. However, not all NLSs have been characterized, and the absence of an NLS does not necessarily eliminate the possibility that PTK 7 can be transported to the nucleus. More experiments Confocal Microscopy The most direct way to see if PTK7 is found in the nucleus is through confocal microscopy. By staining the nucl eus with DAPI and performing slice by slice confocal analysis of cells stained with either an aptamer or antibodies against PTK7, colocalization of PTK7 in the nucleus can be determined. I tried numerous times to co stain cells for PTK7 and nuclei using ma ny different parameters. Unfortunately, while internalized sgc8c labeled aptamers are clearly seen by flow cytometry, under the microscope, the staining was too weak to determine where in the cells PTK7 was located. Due to the low quality of these images, I have not included them here. More optimization of the immunocytochemistry protocol for PTK7 needs to be done to produce pictures that can conclusively determine if PTK7 is localized to the nucleus. However, the AY Strongin group has succeeded in fluoresc ent microscopy of PTK7 in cells. They found addition of the PTK7 protease MT1 MMP to PTK7 expressing cell
84 lines caused PTK7 to be removed from the cell membrane and accumulate around the nucleus [ 117 ] Summary of PTK7 Known and Predicted Features In our discussion of PTK7 protein we have mentioned many structural and predicted features. Figure 3 6 summarizes and ma rks the features of PTK7 primary sequence mentioned in the text for clarity. MGAARGSPARPRRLPLLSVLLLPLLGGTQT AIVFIKQ PSSQDALQGRRALLR [1. CEVEAPGPVHVYWLLD GAPVQDTERRFAQGSSLSFAAVDRPQDSGTFQCVARDDVTGEEARSANASFNIKWIEAGPVVLKH ] PASEA EIQPQTQVTLR [2. CHIDGHPRPTYQWFRDGTPLSDGQSNHTVSSKERNLTLRPAGPEHSGLYSC ] CAHSA FGQACSSQNFTLSIA DESFARVVLAPQDVVVA RYEEAMF [ 3.HCQFSAQPPPSLQWLFEDETPITNRS RPP HLRR ATVFANGSLLLTQVRPR NAGIYRC ] IGQGQRGPPIILEATLHLAEIEDMPLFEPRVFTAGSEERVT [4. CLPPKGLPEPSVWWEHAGVRLPTHGRVYQKGHELVLANIAESDAGVYTC ] HA ANLAGQRRQDVNITVA TVPSWLKKPQDSQLEEGKPGYLD [5. CLTQATPKPTVVWYRNQMLISEDSRFEVFKNGTLRINSVEVYDGT WYRC ] MSSTPAGSIEAQARVQVLE K LK FTPPPQPQQCMEFDKEATVP [6. CSATGREKPTIKWERADGSS LPEWVTDNAGTLHFARVTRDDAGNYTC ] IASNGPQGQIRAHVQLTVAVFITFKVEPERTTVYQGHTA [7. L LQCEAQ G D QW 4 KGKDRILDPTKLGPRMHIFQNGSLVIHDVAPEDSGRYTC ] IAGNSCNIKHTEAP LYVVDK 4 PVPEESEGPGSPPPYK MIQ TIGLSVGAAVAYIIAVLGLMFY CKKRCKAKRLQKQPEGEEPEMEC LNGGPLQNGQPSAEIQEEVALTSLGSGPAATNKRHSTSDKMHFPRSSLQPITTL GKSEFG EVFLAKAQGLE EGVAETLVLVKSLQSKDE 5 QQQ LDFRRELEMFGKLNHANVVRLLGLCREAEPHYMVLEYVDLGDLKQFLRI SKSKDEKLKSQPLSTKQKVALCTQVALGMEHLSNNRFVHKDLAARNCLVSAQRQVKVS ALG LSKDVYNSEY YHFRQAWVPLRWMSPEAILEGDFSTKSDVWAFGVLMWEVFTHGEMPHGGQADDEVLADLQAGKARLPQPEG CPSKLYRLMQRCWALSPKDRPSFSEIASALGDSTVDSKP Figure 3 6. PTK7 primary amino acid sequence with identified features marked [GenBank AAH71557.1]. Protein is 1,070aa with 118kDa full form and 68kDa cleaved fragment. Single underlined numbers and brackets: denotes each of 7 Ig like fold domains; Grey hig PTK7 M02 antibody; Pink highlight and white lettering: region I predicted could bind PTK7; MGAARGSPARPRRLPLLSVLLLPLLGGTQT : signal peptide; RPPHLRR predicted NLS; K LK : chuzhoi mutant cleavage site; G : Site of : MT1 MMP cleavage site; GKSEFG and ALG : mutated regions of tyrosine kinase domain that inactivated the kinase, active kinase have GXGXXG and DFG respectively instead; TIGLSVGAAVAYIIAVLGLMFY : transmembrane doma in; 4 : Region between these two 4 s is deleted in PTK7 4 isoform. 5 : Truncation site for PTK7 5 isoform. Cleaved PTK7 Fragment is Higher in Cellular Nuclear Fraction Another way to determine if PTK7 is in the nucleus is to fractionate cells and probe them by western blot. Using this technique, we saw full length PTK7 in the
85 nuclear and pellet fractions of human embryonic kidney cells (HEK293). There was a larger level of full length PTK7 in the pellet fraction. Full length PTK7 was expected in the pellet fr action, which would contain the membrane; however, the presence of full length PTK7 in the nuclear fraction is unexplained. We saw cleaved PTK7 in all three fractions, and at much higher level in the nuclear fraction. We would expect to find cleaved PTK7 in the pellet fraction, as cleaved PTK7 has been found in complex with full length PTK7 on the membrane [ 79 ] The high level of cleaved PTK7 in the nuclear fraction may indicate that cleaved PTK7 is indeed present in the nucleus; however, this needs to be validated by confocal microscopy. These results were similar for the lung adenocarcinoma cell line H23, and in the cervical cancer cell line HeLa (Data not shown). Figure 3 7. Western blot of various cellular compartments of HEK293 cells. Full length PTK7 is ~ 118kDa cleaved ~ 68kDa. GAPDH is a cytoplasmic marker. Lamin B1 is a nuclear membrane marker. This blot is representative of three experiments. Sgc8c Prevents Wound Healing in HeLa Scratch Assay As discussed in the Introduction, PTK7 has been implicated in tumorigensis, invasive ness, and metastasis. The ST Lee group found that in vitro addition of a soluble
86 fragment of PTK7 made of just the 6th and 7th Ig domains (~200aa) had profound effects on cancer related signaling as seen by reduced wound healing in HeLa [ 70 ] Following their work, we hypothesized that, if sgc8c binds to a region of PTK7 important for downstream signaling, then addition of PTK7 aptam ers could potentially block the region of the protein and affect binding. One way to assay this would be to healing assay in the presence of sgc8c. In this assay, a monolayer of cells is serum starved and then the surface of them is s cratched. Cells migrate into the wound and heal the scratch. The speed and efficacy by which the cells heal the wound can be considered an indicator of their invasiveness. If sgc8c blocked a region important to PTK7 function, then the ability of the cells to migrate would be diminished. This is what we saw. Treatment of cells with sgc8c for 24h significantly reduced their ability to heal a wound [Figure 3 8]. This indicates that the addition of sgc8c might block a region of PTK7 important for downstream sig naling events. In the future, more experiments, using transwell plates or other assays of invasiveness, could help to strengthen this conclusion. Conclusions Bioinformatic, competition, and mutational analysis of PTK7 aptamers revealed a consensus sequence important for aptamer binding. An analog to the PTK7 aptamers untranslated region of DIXDC1b. This region of DIXDC1b was also conserved in the regulatory regions of chimpanzee, mouse, and rat DIXDC1b homologs. These result s indicated there might be a natural interaction between PTK7 protein and DIXDC1b DNA. Structural modeling of PTK7 identified a region of potential aptamer binding at PTK7 Ig3. A possible scheme for this interaction is shown in Figure 3 9.
87 Figure 3 8. S cratch assay. A) Representative images of scratch in monolayer of HeLa cells at 0h, 12h, 24h: 10x resolution. Scr sgc8: scrambled sgc8. B). Quantitation of scratch assay data. The ratio of wound area at 24h and 0h were determined by ImageJ software. Result s are from 1 experiment with 4 replicates. Several other independent experiments had similar results. P values determined by ANOVA. In order for this interaction to occur, cleaved PTK7 must be transported to the nucleus. PTK7 has a weak predicted nuclear localization sequence. By cell fractionation, followed by Western blot, we found an abundance of cleaved PTK7 in the nuclei of several different cell lines Furthermore, addition of sgc8c to HeLa cells reduced their ability to migrate in a scratch assay, indicating sgc8c may block a region important for PTK7 signaling.
88 Figure 3 9. A) Scheme of possible PTK7 aptamer interactions on the surface of the cell. B) Scheme of possible PTK7 protein interaction with aptamer similar regions in DIXDC1b DNA. Pink s tem loop structure in dashed box: KC2D4 consensus sequence; Blue stem loop structure in dashed box: sgc8c consensus sequence. Green protein structure: model of PTK7 extracellular region. The purple X denotes the natural MT1 MMP cleavage site; the black X i s the additional chuzhoi mutant MT1 MMP cleavage site. Using the sum of this information, we developed a model for the possible functional role for PTK7 DIXDC1b DNA interaction [Figure 3 10]. In this model, PTK7 extracellular porti on is cleaved by MT1 MMP outside of the cell [1 ]. The soluble PTK7 fragment (sPTK7) is transported to the nucleus . In the nucleus, sPTK7 binds the region of DIXDC1b at the sites of aptamer sequence identity [3.], and causes the modulation of DIXDC1b transcription [4.]. The altered DIXDC1b affects the amount of active Dvl and the level of JNK gene expression [5.]. These changes affect Wnt
89 signaling [6.]. Much work still needs to be done to interrogate each one of the steps in the model below; it is presented only a s a starting point for discovery. Figure 3 10. Model for PTK7 DIXDC1b interaction and its effect on Wnt signaling. Please see text for description. Another possibility is that PTK7 interacts with the mRNA of DIXDC1b instead of the DNA and could modulate DIXDC1b by translational regulation instead of transcriptional regulation. This is not expected for two main reasons. Although DNA and RNA share simil ar structures, RNA has a hydroxyl group attached to every ribose. This modification has several effects; namely, it makes RNA more structurally unstable,
90 alters the secondary structure, and prevents the robust double stranded conformation seen with DNA. As aptamer binding is intimately connected to ssDNA secondary structure, the perturbations caused by replacement of RNA would affect binding. In the last chapter, we saw how introduction of a single azobenzene within the consensus region decreased the aptame r binding affinity 1,000 fold. Our review of the literature has not revealed one example of an ssDNA aptamer binding a target when it is translated into RNA, or vice versa, however, this does not mean it is not possible. The second main reason why it is unlikely PTK7 would interact with DIXDC1b mRNA instead of DIXDC1b DNA is that, when RNA is transcribed from DNA, mRNA is only made from 1 strand of the DNA. Our PTK7 aptamers have sequence identity to both strands of DIXDC1b DNA, and would presumably inter act with both strands of the DNA. Another option, if PTK7 does indeed bind mRNA, could be that an miRNA interacts with DIXDC1b mRNA at the region on the negative strand of DIXDC1b RNA that has sequence identity to the KC2D4 aptamer. The PTK7 protein would then bind the DIXDC1b mRNA and its miRNA. The question of whether PTK7 binds DIXDC1b mRNA or DNA can only be conclusively answered by synthesizing the sgc8c aptamer in RNA and seeing if it binds. Materials and Methods DNA Sequences DNA aptamers used in thi s chapter [Table 3 2] were synthesized, purified, and validated as described in Chapter 2 Materials and Methods.
91 Table 3 2 Sequences used in this chapter. B: Biotin; DIXcs: aptamer consensus region with DIXDC1b DNA; dsDIX 40 80: nucleotides 40 80 of the DIXDC1b mRNA sequence, positive direction; c dsDIX 40 80: sequence complementary to dsDIX 40 80.; Scr sgc8c: scrambled sgc8c. Experiment Name DIXcs GCGCAGCCTGCGCCGCCGGGA B dsDIX 40 80 CACTGGGAGGGAGGCTCCCGGCGGCGCAGGCTGCGCCTCG B c dsDIX 40 80 GCTGGGGACTCGAGGCGCAGCCTGCGCCGCCGGGAGCCTC B dsDNA Binding + Scratch Assay sgc8c TCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAGA B Scr sgc8 c ATCTAACTGCTGCGCCGCCGGGAAAATACTGTACGGTTAG B Cell Culture CEM CCRF, HeLa, and HEK cells lines were obtained from American Type Culture Collection (ATCC). CEM cells were cultured in RPMI 1640; HeLa and HEK cell lines were cultured in DMEM 1640 medium (GIBCO). All media was supplemented with 10% FBS (GIBCO) and cell were cultured at 37C under 5% CO 2 100 units/mL penicillin streptomycin (Cellgro) were added to the cultures, occasionally, if bacterial contamination was suspected in the lab. Cultures were routinely monitored for Mycoplasma infection. dsDNA Binding Experiments One million CEM cells were wa shed with 2mL washing buffer (WB: PBS, 4.5g/L glucose, 1M MgCl sgc8c, DIXcs B, hybridized dsDIX and c dsDIX 40 80; or hybridized dsDIX 40 80 and c dsDIX 40 buffer (BB: WB with 1g/L bovine serum albumin, and 100mg/L tRNA) for 20min at 4, washed, stained with 1:400 streptavidin PE, washed again, and measured using a FACScan flow cytometer (Becton Dickenson), counting 20k cells. The data was analyzed with FCS Express.
92 Aptamers were hybridized by mixing e qual molar equivalents of the various DNA in H 2 0, which was heated to 95C for 5min and allowed to slowly cool to room temperature. Hybridization was confirmed by running a small aliquot on a 3% agarose gel in TBE for 40min at 120V and noting a shift in mo lecular weight. Western Blot HEK293 cells were cultured in DMEM with 10% FBS under 5% CO One million cells were collected and fractionated using a NE PER kit (Pierce) following all 20% Tris Glycine gel (Invitrogen) and mouse horseradish peroxidase (HRP) 1:5000 (Pierce). The membranes was stripped and re probed for GAPDH (ABcam), a cytosolic marker, and Lamin B1 1:1000 (ABcam), a nuclear marker, foll Rabbit HRP 1:5000 (Pierce) and imaged on Kodak film. Scratch Assay The night before the experiment, a 24 well plate, fully confluent with HeLa cells (100k cells per well) sgc8c, or sgc8c The next day each well was s cratched with a 200uL pipet tip. Cell were washed 2x in PBS, and 100L DMEM with 2% FBS with no treatment, sgc8c, sgc8c was added to each well. Images were taken at 10x magnification under bright field using a Ze iss microscope at 0h, 12h, and 24h. To ensure the same region was photographed each time the plates were pre marked with 2 parallel lines black on top, blue on bottom. P ower P oint was used to align the images from different time points using the black and blue line guides. ImageJ was used to measure the area of the wound. Every well had the same size region analyzed for wound area.
93 CHAPTER 4 ADDITIONAL SEQUENCE IDENTITIES FOUND BET WEEN APTAMERS BY CELL SELEX BIOINFORMATIC ANALYSIS Introduction Thus far our discussion has centered around the PTK7 aptamers, which were found to share a common consensus region that had an analog in genomic DNA. The bioinformatic dataset that uncovered the commonality between these PTK7 aptamers included 145 other se quences from 28 independent selections (Appendix A). Among these other aptamers, there were also pairs from different selections that shared common sequence regions greater than would be expected from chance. This chapter will highlight another set of apta mers from one of these selections that share sequence identity sequences selected against Vaccina virus and Vaccina virus infected cells. After a discussion of the Vaccina aptamers, we will briefly explore general trends surrounding the GC/AT content of th e aptamers, and conclude with a brief catalog of some other aptamer pairs we have found that share suspiciously high sequence identity. Vaccinia Aptamers Vaccinia Virus (VV) is responsible for saving many millions of lives. It is the active constituent of the smallpox vaccine which eradicated smallpox in 1980. The virus is a roughly 350x250nm football shaped particle enclosed in a lipid membrane stolen from the host as it exits the cell. This purloined cell membrane is studded with many viral proteins, i ncluding, hemagglutinin (HA). HA is a nonessential gene for VV replication, and is known to interact with other VV encoded proteins such as VP37K and SPI 3 late in VV infection. HA interactions with these proteins prevent cell fusion. HA is
94 also expressed early in VV infection, but its function at early infection time points is unclear. Due to the need for rapid detection methods for viruses in the pox family, two independent aptamer selections were conducted in the Weihong Tan lab using VV infected cells as the target [ 118 119 ] These selections used different virus infected cell lines, different primer pairs and libraries, and were carried out several years apart by different people. Parag Parekh determined HA was the aptamer target for his selection of PP5 aptamer. In addition to these selections several years previously, the Andreas Kage group in Germany performed a ssDNA selection directly against VV particles themselves, in a one step selection using a VV particle affinity column, in a procedure they termed MonoLEX [ 120 ] Surprisingly, bioinformatic analysis of our ssDNA aptamer dataset revealed a common region of sequence identity in aptamers selected by these th ree independent aptamer selections [Table 4 1]. Structural prediction of aptamer secondary structure by NuPACK software showed all the sequences formed a strong stem loop structure with the consensus region found in the loop of each structure [Figure 4 1]. Based on these results, I predict all four sequences will bind to the HA, and these aptamers will compete against each other. Table 4 1. Aligned aptamer sequences from three Vaccinia related selection. Black highlighted sequence is shared between all the aptamers. SELEX target either VV infected cells: HeLa or A549, or target VV particles. Note: TOV2, PP5, and PP3 only show the aptamer variable region for clarity. Please refer to Materials and Methods for full sequences. Aptamer Sequence SELEX Target TOV 2 CACT TGCATATACA C TTTGCAT TATAGGG HeLa PP5 CC TGCATATACA C TTTGCAT GTGG A549 PP3 CGAGCCAGACATCTCACACCTGT TGCATATACA T TTTGCAT A549 A38 TACGACTCACTATAGGGATCCTG TATA T A T TTTGCA ACTAATTGAATTCCCTTTAGTGAG VV
95 Figure 4 1. Predicted secondary structures of consensus sequence and each of the full aptamer sequences. The bar on the right indicates Bar on right shows strength of predicted structure. Red being most strong. The selections that produced these aptamers for VV invo lved two different infected cell lines and a third selection on the virus particle alone. They were the product of three entirely different SELEX procedures carried out in two labs, on two different continents years apart. As with selection of the PTK7 apt amer, the chances of selecting aptamers against the same target multiple times were remote; even more so because HA is only a minor protein in the VV particle. It is not by any means the most highly expressed protein. Why was this happening? BLAST of the Consensus Sequence against the Human Genome VV is a dsDNA virus. Furthermore its DNA replication and particle maturation takes place entirely in the cytosol of the host cell. As a result, VV proteins, including HA which is expressed early, intermediately, and late in VV maturation, could have an opportunities to come in contact with the VV genome. Therefore we hypothesized there might be a natural, physiological role for these sequences and possibly HA interacts directly with the viral genome. To determine if this was indeed the case, we performed a
96 BLAST search for the 19nt consensus region of the aptamers: TGC ATA TAC ACT TTG CAT. The loop part of the consensus sequence was found twice in the VV genome in the coding regions these two proteins: D3R, GenBank and NPH II, GenBank: AAA48064. Neither of these proteins has any known direct link to HA. Table 4 2. BLAST results of VV genome revealed two proteins D3R and I8R share DNA sequences in common with the VV aptamers. Black highlighted sequences are shared b etween all the aptamers and VV DNA. Purple highlighted sequence is additional sequence identity found between D3R and I8R. Aptamer Sequence TOV2 CACT TGCATATACA C TTTGCAT TATAGGG PP5 CC TGCATATACA C TTTGCAT GTGG PP3 CGAGCCAGACATCTCACACCTGT TGCATATACA T TTTGCAT A38 TACGACTCACTATAGGGATCCTG TATA T A T TTTGCA ACTAATTGAATTCCCTTTAGTGAG VV DNA D3R TTGTCATTT ATGATA ---TATACA C TTTTGA CGCTTTCAAGAA I8R GATG ATGATA CGTA TATACA C TTTTG TAAAATATTATTCG Conclusions While this story about VV aptamers is incomplete, there is much intriguing evidence that more analysis might uncover a functional role for the aptamer sequence identity. Perhaps, HA is a naturally DNA binding protein. PTK7 and VV aptamers is not the end of sequence identities produced from SELEX bioinformatic analysis. Below are the aligned sequence from two selections performed in different labs against a neuronal cancer cell line (PC12 [ 121 ] ) and a mesenchymal stem cell line (20MSC [ 122 ] ). These two aptamers share an 18/20nt contiguous sequence. Unfortunately the protein targets for these aptamers have not yet been discovered, but when they are, I would bet these two aptamers bind the same target; perhaps for a functional reason. There are 3 other pairs of aptamers from our dataset that share similar levels of sequence identity.
97 Figure 4 2. Sequence identity between two aptamers with unknown targets selected in different labs on different cell lines.
98 CHAPTER 5 SGC8C APTAMER AND PTK7 ANTIBODY RUPTURE FOR CES ARE COMPARABLE ON LIVE HELA CELLS Introduction Aptamers and antibodies rely on a similar set of non covalent interactions to bind their targets hydrogen bonding, electrostatic interactions, and van der Waals forces, except antibodies can use hyrophopic interactions as well. Even though aptamers are much smaller that antibodies, i.e., 8 15kDa vers us 150kDa antibodies, and even though they are made of only 4 different nucleic acid bases compared to the 20 amino acids of antibodies, surprisingly, both antibodies and aptamers can have very strong non covalent interactions, producing binding affinities in the low nanomolar K d range. Due to their similar ability to target a wide range of proteins, yet their very different structural natures, there has been some concern that aptamers may not bind as robustly as antibodies. To address this issue, we used single molecule atomic force micr oscopy (AFM) to measure the rupture force between a protein and its respective aptamer and antibody. As we have mentioned before, the pseudokinase protein tyrosine kinase 7 (PTK7) is important in development [ 13 32 123 ] and cancer [ 67 68 70 ] Both a PTK7, and a DNA aptamer, sgc8c have been identified that exhibit strong binding to receptor PTK7 with a K d in the subnanomolar range [ 6 ] The specific interactions of protein receptor PTK7 with its two ligands have been confirmed by siRNA silencing and PTK7 plasmid insertion [ 68 ]
99 Figure 5 1. Scheme of AFM measurements on live cells. With the goal of using sgc8c for targeted delivery of chemotherapy to tumors, several questions surrounding sgc 8c PTK7 interactions needed to be answered: What is the force of interaction between the protein and the aptamer? And how does that force compare with the force for the antibody? A straightforward and novel way to answer these questions is via examination of aptamer target adducts using microscopes capable of single molecule tracking on live cells. In this study we used the single molecule AFM technique developed by XH Fang et al. [ 124 ] to comp are the rup PTK7. The experimental scheme is shown in Figure 5 1. Results and Discussion The rupture force was measured using AFM tips functionalized with either sgc8c PTK7 antibody, or th rombin antibody as a control (there is no thrombin on
100 the HeLa cell surface) to scan cell membrane surfaces for changes in the force curve that indicate a binding event. Three typical force curves are shown in Figure 5 5. Blocking Controls To ensure that actual binding was observed, several binding probability controls were performed [Figure 5 2] Blocking by addition of excess sgc8c aptamer significantly diminished the number of rupture peaks observed with HeLa cells from 13.21.3% of the total to only 4. PTK7 antibody was used to block the anti PTK7 tip, the binding probability decreased from 11.52.0% to 4.20.7%. The probability after blocking was only slightly higher than the background controls when a thrombin modified t ip was used (2.30.6%). This result indicates that there is specific PTK7 modified AFM tips and HeLa cells. Figure 5 2. Binding probabilities of tip with cells. Obtained using different blocking, cell, and tip modification conditions. 1,000 data points were recorded per cell.
101 Fitting the Collected Data At least 100 force curves were recorded to construct force histograms, which were then fitted using Gaussian peak functions [Figure 5 3]. The single molecule ru pture force between sgc8c and HeLa PTK7 was found to be 4626pN, while the rupture force between anti Figure 5 3. Histograms of binding forces between tips and HeLa cells. Bars: experimenta l data; solid line: Gaussian fit which reveals the most likely rupture force, and its respective theoretical Gaussian distribution curve. A) Binding force PTK7 antibody and HeLa cells. Conclusions For the first time the single molecule rupture force between a protein, PTK7, and both an aptamer and an antibody was measured using single molecule AFM on live cells. The single PTK7 was found to be 4626pN, PTK7 and HeLa PTK7 was PTK7 antibody,
102 significantly diminished the number of rupture peaks observed, indicating that PTK 7 modified AFM tips and HeLa cells was being observed. The measured force values between the two are very similar, indicating that, despite the differences in size and component diversity, they have similar binding profiles, and that DNA aptamer interactio ns with proteins can be as robust as those of antibodies in terms of rupture force. Interestingly, when we compare these results to previous AFM studies that have looked at rupture forces between hybridized dsDNA [Figure 5 4], we find the mean value for s gc8c PTK7 interaction live cells equates to 20 30bp dsDNA. Thus, even aptamer rupture force could be commensurate with 20 30bp interactions or 45 60 hydrogen bonds within the error [ 125 ] This indicates there might be unconventional, non hydrogen binding interactions occurring between the PTK7 and sgc8c. Figure 5 4 Rupture forces from literature of different lengths of dsDNA and sgc8c. Black squares [ 126 ] Blue square [ 127 ] Purple square [ 128 ] Pink square [ 129 ] Red star: Sgc8c rupture force.
103 Materials and Methods Preparation of AFM Tips and Substrate The natural oxide layer on the AFM silicon nitride (Si3N4) tips (type: NP, from Veeco, Santa Barbara, CA, USA) was removed by etching the substrates in HF for 30s at room temperature in a hood with proper protective eq uipment. The tips were then dipped in alkaline solutions of NH4OH:H2O2:H2O, 1:1:5 v/v for 30min, followed by oxidation in 90C piranha solution with 98% H2SO4:H2O2, 7:3 v/v for 30min. Between each of these steps the tips were washed well with water. To fun ctionalize the tips, they were transferred into a solution of (3 mercaptopropyl) trimethoxysilane (MPTMS) 1% v/v in toluene and incubated for 2h at room temperature, followed by thorough rinsing with toluene to remove any unbound silane. The tips were acti vated by incubating in 1mg/mL N hydroxysuccinimide ester poly(ethylene glycol) maleimide (NHS PEG MAL; Nektar Therapeutics) in DMSO for 3h at room temperature, followed by extensive rinsing with DMSO to remove unbound NHS PEG MAL. Finally, the tips we re immersed Thrombin antibody as a control (Haematologic PTK7 (Miltenyi Biotec, Germany); or in 100nM aptamer in Tris TTT TTT TTT ATC TAA house on an IDT DNA synthesizer and purified using a reverse phase Prostar HPLC (Varian) on a C18 column (Econosil, 5U 250 x 4.6 mm from Alltech Associates) with a linear elution. These sequences were vacuum dried or ethanol precipitated, detritylated with 20% acetic acid, and stored at minus 20C for future use. The absorbance at 260nm for the purified aptamers was quantified using a UV Vis spectrophotometer (Bio Rad). Aptamer concentr ation was determined using the
104 L ambert Beer equation: A bs 260nm where is the extinction coefficient, b is the cuvette pass length, and c is the DNA concentration. The sequence was tested for quality control by running 0.5 L of purified aptamer on a 3% agarose gel in TBE buffer (89mM Tris HCl, 89mM boric acid, 2mM EDTA, pH 8.0) for 40min at 120V and imaged under UV with Ethidium Bromide (EtBr). The density of protein on the AFM tips was sufficiently low to ensure that only one protein or DNA molecul e was measured at a given time. After rinsing in buffer, the tips were stored in buffer at 4C until use. The tips were calibrated using the thermal fluctuation method for the range of 0.040 0.075 N/m. Cell Culture All cells used in our experiments were c ultured in incubators at 37C with 5% CO 2 penicillin/streptomyocin (Sigma) and 5% Fetal Bovine S erum (Gibco). Just before use, the cells were washed three times with 2mL phosphate buffered saline (PBS; Sigma), and then 2mL DMEM was added. For blocking experiments, 1M sgc8c aptamer or free DMEM, followed by incubation for 10min prior to imaging. AFM Measurements All AFM force measurements were performed with a Nano Scope III AFM (Veeco, Santa Barbara, CA) with live cells in plates filled with freshly prepared DMEM buffer. The force curves [Figure 4 5] we re recorded and analyzed by Nanoscope 5.30b4 software (Veeco, Santa Barbara, CA), and statistics were calculated using Origin 8.5 software. Seven AFM tips were measured in total: 3 modified with Sgc8c, 3 modified
105 with PTK7 antibody, and 1 modified with thr ombin protein as a control. Data from 4 6 cells were collected for each tip with and without blocking. For each cell, 1,000 data points were collected from at least 50 different locations on the cell membrane. Each measurement took no longer than 1h. Fi gure 5 5. Representative force distance curves for sgc8c AFM tip and HeLa cells. Red curve is tip approach; blue curve is tip retraction. Top curve pair (red and blue) shows no binding, while black arrows on the bottom two pairs point to the measured bindin g. Expanded inset shows how measurements were F is directly proportional to spring constant, k and spring displacement, x
106 CHAPTER 6 MODIFYING CELLULAR PROPERTIES USING ARTIFICIAL LIPID APTAMER RECEPTORS Introduction Cell membranes are a central hub for receiving signals from outside the cell and membrane surface area, keeping its integrity intact. These phospholipids are studded wi th numerous proteins acting as recognition moieties, transporters, receptors, or structural connectors. Membrane proteins are vital to many cell functions and can control cell binding, fate, and signaling. Thus, the ability to choose which proteins are pr A number of methods have been used to alter which proteins are found on the cell surface. These include the use of recombinant proteins [ 130 ] the attachment of proteins to lipids which then insert themselves into the membrane, such as glycosylphosphatidylinositols anchored proteins [ 131 132 133 ] NHS functionalized poly (ethylene glycol) oleyl derivatives [ 134 ] and palmitated protein A complexes [ 135 ] Covalently coupling proteins to the surface by chemically modifying them with azide [ 136 ] or another chemistry has also been reported. However, these approaches all have shortcomings. Recombinant strategies are time consumi ng and require optimization to produce proteins that are reliably trafficked to the membrane. Insertion of lipid functionalized proteins, or covalently linked proteins, cannot be modulated, nor can they be strategically placed on the cell membrane. Moreove r, covalently bound proteins can In addition to proteins, groups have also linked cDNA to the surface of cells via lipid attachment [ 137 ] or broad chemical modification of the cell surface [ 138 ] These
107 modifications were made in order to attach cells to surfaces or to build artificial cell architectures for tissue engineering. Building on these attempts, we sought to create artificial aptamer aptamer receptors that capture proteins onto the cell surface in a rapid, reversible, and dose controllable manner. Aptamers are short DNA or RNA sequences, typically 1 5 to 80 nt in length. Selected from large libraries of many different sequences, aptamers bind to a particular target through a process known as SELEX. dissociation constant s (K d ) in the low nanomolar to micromolar range, aptamers have many advantages over antibodies, including increased stability, small size, simple cDNA, and flexible tar get recognition. Previously, our lab synthesized a diacyllipid phosphoramidite [ 46 ] a nucleoside building block that has two long saturated fatty acid chains held together with a glycer ol. This lipid end can form micelles which, when added to cells, can insert into cell membranes. This diacyllipid phosphoramidite can easily be can be easily functional ized with the lipid, and when added to cells, the lipid will anchor the aptamer in the membrane where it will protrude from the cell, thereby providing target binding capability. Our lab has previously made TD05 aptamer micelles, where the aptamer TD05 re cognized the B cell leukemia Ramos cell line [ 139 ] A dye, which was loaded into the micelle interior, was specifically delivered to Ramos cells by these micelles. Streptavidin (SA) and throm bin are two proteins having unique functions which, when attached to the cell surface, can dramatically modify cell behavior. SA is a
108 tetravalent protein that binds the small molecule biotin with a very high affinity K d > 10 14 making SA a useful tool in cell biology, especially when a rapid, specific interaction is needed. Many different fluorescently labeled SAs are available, as well as several aptamers that bind SA with high affinity. Thrombin is an enzyme important in the clotting of blood. In blood bound to cell membranes, it exists in an inactive, prothrombin, state. When activated by Factor Xa, it cleaves soluble fibrinogen protein into insoluble fibrin fragments. These fibrin fragments agglomerate, are covalently linke d through their lysine and glutamine side chains by Factor XIII, and promote clot stabilization Thrombin is a complex protein with several different active sites; in addition to cleaving fibrinogen, it also activates protein C, platelets, cell thrombin re ceptors, and Factors V, VII, and VIII, which regulate thrombin production. Thrombin has two very well studied aptamers: a 15nt aptamer (T 15), which 27), which binds exosite 2, which is impor tant for heparin binding. In this work, we will make aptamer receptor s for streptavidin and thrombin. We will show that cells modified with the receptors are able to capture proteins on their surface in a manner that is rapid and vity Results and Discussion Streptavidin Aptamer Receptor Anchors on the Cell Membrane In the first set of experiments, cells were modifi ed with streptavidin aptamer receptors (SA ARs), enabling them to capture fluorescently labeled streptavidin. The cap tured SA stained the cell membrane in a dose controllable manner. Streptavidin (SA) is a tetravalent protein known for its high affinity toward the small molecule biotin. SA ARs were made by attaching a lipid tail to a 29nt aptamer that binds streptavidin (40nM K d )
109 [ 140 ] To confirm that SA ARs retained their binding ability to SA, FITC labeled SA aptamers were competed off of SA coated magnetic beads by SA AR (data not shown). All cell lines tested were able to capture Alexa 488 labeled SA (SA 488) on their cell membranes after insertion of SA ARs [ Figure 6 1a ]. Figure 6 1. Characteristics of streptavidin artificial receptors (A) Artificial receptors can be inserted in many different cell types. Cells were incubated with 2M SA ARs for 2h at 37C, followed by addition of 1:400 SA 488, and imaged using bright field and fluorescence (488ex) microscopy. (B) CEM cells were incubated with different concentrations from 31nM to 5M SA ARs for 2h at 37C, followed by addition of 1:400 SA 488. SA 488 on the cell surface was analyzed by flow cytometry using channel 1 for SA 488. (C) CEM cells were incubated with 2 M SA AR for 2h at 37C, followed by washes and incubation at various times from 0 to 30h before addition of 1:400 SA 488. Following treatment, the cells were analyzed by flow cytometry. All histograms are ungated. Scale bar is 10M.
110 Each cell type had a un ique staining profile relative to (1) the amount of SA 488 incorporated: CEM and HeLa were high, while A549 were lower; (2) the smoothness of staining: HeLa and A549 were punctate, while CEM, Ramos, and Ludlu 1 were smoother; and (3) the predominant locati on of staining: junctions were between cells as in MCF7 or were on the main body of the cell as in HeLa and A549. These differences plasma membrane and the ease of SA AR insertion While SA 488 was captured with ease, phycoerythrin (PE) labeled SA could not be captured on the cell surfaces by the SA ARs, possibly because PE is larger than Alexa 488, and PE sterically interferes with In cubation of CEM cells with different concentrations of SA AR resulted in the dose dependent capture of SA on the cell surface [ Figure 6 1b]. Specifically, incubation with as little as 31nM of SA AR, which is below the K d for the SA aptamer, was enough to d etect SA 488 on the cell surface with flow cytometry. Increasing the SA AR concentration increased the amount of SA 488 signal in a linear fashion until it plateaued at around 5M. SA ARs persisted on the cell membrane for an extended time, but the amount of aptamer slowly decreased over two days after incubation [Figure 6 1c]. After two days, fluorescence became undetectable, indicating SA AR modification is temporary, and cells return to normal after being cultured for two days. Furthermore the insertion of SA ARs is rapid, after addition to cells, d etectable levels of SA AR insertion wer e apparent within 5min and reached saturation levels within an hour [Figure 6 2].
111 Figure 6 2. Aptamer insertion begins after 5min and reaches saturation after an hour CEM cells were incubated with 1M SA AR for different times from 5 120min, after which cells were washed and stained with SA 488. To demonstrate that the SA aptamer was responsible for localizing SA on the membrane, we attached SA 488 to the cell surfac e at 4C. However, when the temperature was increased to 37C, the aptamer did not bind, and we lost all SA 488 signaling on the cells. The fluorescence measured by flow cytometry for treated cells went from 393 arbitrary fluorescent units (AFU) at 4C to 3 AFU after 30 minutes at 37C, which was the same as background, unmodified cells. These results demonstrate that the binding of SA AR to its target SA 488 can be modulated by altering such environmental conditions as temperature. Furthermore, SA ARs ins ertion had no effect on cell proliferation as measured by MTS [Figure 6 3]. These results indicate SA AR treatments do not negatively affect cell growth an important criteria for future applications.
112 Figure 6 3 SA AR does not inhibit cell growth 20,000 CEM cells were incubated with 2M SA AR in 100L media over 72h. After which proliferation was determined by MTS. Strepta vidin on Modified Cells Remains Functional Proteins captured by aptamer receptors need to remain functional to be useful. To confirm the active, i.e. biotin binding, state of SA captured by SA ARs, a sandwich assay was performed, as seen in Figure 6 4 a S A is tetravalent, with four distinct biotin binding sites. Therefore, even though one site is taken up by the aptamer, there a re theoretically three other sites on each SA 488 capable of binding. In this assay, two different suspension cell lines that do not naturally bind to each other, were induced to bind each other when a biotinylated aptamer that recognized the other cell type was added to SA modified cells. This aggregation did not occur when the control aptamer, which only recognized the modified cell type, was added. Sgc8 is an aptamer that binds a pr otein called PTK7, which is found on CEM cells, but not on Ramos cells. TD05 is an aptamer that binds IgM, which is found on Ramos cells, but not on CEM cells. In the first case, CEM cells were modified with SA AR and then coated with SA 488. Biotinylated TD05 aptamer was then added to the cells, and after washing, a 5x
113 together by the SA modified CEM cells were formed. However, for the control, when biotinylated Sgc8 aptamer, which only binds CEM cells, was added, instead of TD05, no aggregation was seen [Figure 6 4 b T OP ] In the second case with Ramos, the same phenomenon occurred, but in reverse [Figure 6 4b, B OTTOM ]. These experimental findings indicate that the SA 488 m odified on the surf ace of the cells remained functional, and able to form complex assemblies. Figure 6 4 Streptavidin modified cells bind biotin to make cell a ssemblies (A) Scheme showing the cellular assembly. (B) Top: CEM cells modified with strepta vidin 488 via SA AR incubated with TDO5 aptamer that binds Ramos cells and then either CEM (control, left) or Ramos (target, right). Bottom: Ramos cells modified with streptavidin 488 via SA AR incubated with sgc8 aptamer that binds CEM and then either Ram os (control, left) or CEM (target, right).
114 SA ARs Can Be Used to C ap ture Cells for Analysis In order to determine how robust the SA AR system was and to show we could isolate SA AR modified cells, we captured SA AR functionalized cells with streptavidin coated DynaBeads [Figure 6 5]. In this experiment CEM cells were modified with either the control, PDGF AR, o r the target, SA AR, aptamer receptors. When they were mixed with streptravidin coated magnetic beads in a buffer, the cells were enriched on the actin, the positive control s either sgc8 biotin bound cells (wel l 3) or pure cell lysate (well 7) actin. In addition SA AR modified cells (well 5) were able to enrich the cells in buffer and consequently actin. The wells with the untreated cells (well 1), TD05 biotin which does not bind CEM (well 2), or wit h PDGF AR, which does not bind streptavidin, had no actin. Thus, SA AR can be used to modify cells and then capture them again with SA coated beads. Figure 6 5. SA AR modified cells are collected with streptavidin coated magnetic beads. CEM cells modif ied with SA AR or a non specific PDGF AR were collected with streptavidin coated DynaBeads. Cells were lysed, and the lysates were Actin. 1) Unmodified cells. 2) TD05 biotin 3) sgc8 biotin. 4) PDGF aptamer lipid. 5) SA AR. 6) E mpty. 7) CEM lysate. Thrombin Aptamer Receptor Captures Thrombin Thrombin is an enzyme important in the clotting of bl ood. Thrombin has two very well studied aptamers: a 15nt (T 15), which binds the fibrinogen cleavage site on
115 thrombin with relatively low affinity ( K d = 450nM), and a 27nt aptamer (T 27), which binds exosite 2, important for heparin binding, but has a lower binding affinity ( K d = 0.7nM). We made our thrombin aptamer artificial receptor (TA AR) by synthesizing T 27 with a lipid tail. Conseque ntly, when thrombin binds the TA AR, the protein will be tethered to the cell membrane through T The fibrinogen cleaving active site is found on the other side of the thrombin protein, allowing it to remain f ree and active after interacting with the TA AR. To show that we were able to localize thrombin to the cell surface via the TA AR, we made an assembly on the cell surface where the TA AR captured the thrombin protein, which, in turn, captured a biotinylate d T 15. This was visualized by adding SA 488, which bound the biotin on T 15, and was analyzed using a flow cytometer [Figure 6 6]. Figure 6 6 Cells are modified with thrombin via a thrombin aptamer artificial receptor. 500k healthy CEM cells were washed and incubated for two hours at 37C in 50L media (black and blue traces), 5uM TA AR (green and orange), or 5M control PDGF AR (purple). After washing, all tubes, except black and green, were incubated with 500nM thrombin. After further washing, al l tubes were incubated with 500nM biotinylated T 15 followed by streptavidin 488. Cells were examined in FL 1 by flow cytometry for streptavidin 488 signal.
116 Thrombin Modified Cells Cause Clotting Since T 15 bound the fibrinogen cleavage site on thrombin, and, as a result, inhibited the ability to cleave fibrinogen, we asked whether the cells modified with thrombin would induce clotting through cl eavage of fibrinogen to fibrin To address this question, thrombin protein was modified on the surface of CEM ce lls using the TA AR, followed by adding fibrinogen to a clotting buffer. Cells that were pre modified with thrombin v ia the TA AR, and exposed to fibrinogen caused a clot to form within 15 sec [Figure 6 7, Tube 3]. The clots had a gel like consistency, whi ch could be visualized by staining the mixtures with trypan blue. However when a T 15 aptamer, which inhibits washing, there was no clotting [Tube 4]. If a scrambled T 15 wa s was added, clotting was recovered [Tube 5]. Likewise use of a control protein, BSA [Tube 6] or control aptamer receptor, PDGF AR, did not cause clotting. Figure 6 7. Thrombin Modified Cells Cause Clotting. 500k healthy CEM cells were washed and incub ated for two hours at 37C in 50L media (Tubes 1+2), 5uM TA AR (Tubes 3 6), or PDGF AR (Tube 7). After washing, cells were incubated with buffer (Tube 1), 500nM thrombin (Tube 1+7), 500nM thrombin pre incubated with 1M T 15 (Tube 4), 1M scrambled T 15 ( Tube 5), or 200nM BSA for 30min at RT. After five washes, fibrinogen was added to Tubes 2 7. Clots formed and were visualized by adding 50L trypan blue stain.
117 Conclusion This work demonstrates a simple, biomimetic, non toxic, reversible and dose controllable strategy for modifying cell membranes with any protein for which there is a known aptamer. This is an improvement over current methods to modify the surface of cells w ith new proteins because this method is rapid, one step, dosable, reversible and does not alter the captured protein. Using artificial aptamer receptors for streptavidin and thrombin, we further showed that proteins captured on the cell membrane surface re tained their activity, giving cells new properties, such as enzymatic ability, fluorescence, or a way to collect modified cells from a mixture. In the future, these aptamer receptors could be used for various bioanalytical applications or to modify cells b efore their use in therapeutic interventions such as targeting moieties for stem cell homing after autologous bone marrow transplants. Materials and Methods Unless otherwise noted, all chemicals and buffers came from Sigma Aldrich and were not further purified. All DNA bases, except for the lipid phosphoramidite, which was synthesized in house, were purchased from Glen Research. Oligonucleotides were synthesiz ed in house on an automated ABI 3400 DNA synthesizer from Applied Biosystems. HPLC of the DNA sequences was done with a Varian Prostar Instrument. UV/Vis measurements for purity and concentration determination were carried out on a Varian Cary 100 spectro photometer. DNA S ynthesis All DNA sequences were synthesized with an ABI 3400 synthesizer on a 1.0 micromolar scale. Biotinylated CPG Lipid phosphoramidite was dissolved in 0.4mL
118 dichloromethane for coupling. For the lipid DNA after synthesis, the DNA wa s cut from the CPG beads and deprotected in ammonia hydroxide at 55C for 14 hours. Next, the DNA was dissolved in 100mM triethylamine acetic acid buffer (TEAA, pH7.5) and purified by reverse phase HPLC using a C4 column with an acetonitrile gradient (0 30 min, 10 100%) as an eluent. For other DNA sequences with no biotin or dye modifications, the sequences were deprotected for 20min in AMA (1:1 ammonia hydroxide: 40% methylamine) and purified using Gel Pak Purification Columns (Glenn Research) followed by desalting on a Nap 5 column (GE Healthcare). The sequences used can be found in Table 4 1. Table 6 1. Sequences used for Aptamer Receptors Name Sequence TA AR Lipid TT TTT TTG TCC GTG GTA GGG CAG GTT GGG GTG A C SA AR (St 2 1) SA complement ATT GAG TGT TGC GTC ACA CAG CGG TCA AT T 15 FITC GGT TGG TGT GGT TGG FITC T 15 complement CCA ACC ACA CCA ACC T 15 scrambled CAC CAC CAA CAC CAC T 27 complement GTC ACC CCA ACC TGC CCT ACC ACG GAC VEGF complement CCC TGC ACT CTT GTC TGG AAG ACG GGA PDGF AR TTT TTC AGG CTA CGG CAC GTA GAG CAT CAC CAT GAT CCT G VEGF AR Lipid TT TTT TTC CCG TCT TCC AGA CAA GAG TGC AGG G SA: streptavidin; AR : aptamer receptor; T 15 : 15nt thrombin aptamer; T 27 : 27nt thrombin aptamer; VEGF: Vascular endothelial growth factor; PDGF: platlet derived growth factor. Cell C ulture CCRF CEM cells (T cell, human acute lymphoblastic leukemia), Ramos (B cell, human Burkitt's lymphoma), HeLa (human cervical adenocarcinoma), and A549 (human
119 lung adenocarcinoma) were obtained from ATCC (American Type Culture Association). Ludlu 1 cells were obtained from the European Collection of Cell Cultures (ECA CC). CEM, Ludlu 1, and Ramos cells were grown in RPMI 1640 media (GIBCO). HeLa cells grown in Minimal Essential Media (MEM, GIBCO), and A549 cells were grown in F 12k media (GI BCO). All media were supplemented with 10% fetal bovine serum (Invitrogen), Flow Cytometry 500k cells were washed with 2mL washing buffer (WB; 1x PBS, 5mM MgCl2, 4.5g/L glucose) and spun down at 1200g for 3min CEM cells were incubated with different concentrations of SA ARs, ranging from 31nM to 5M, for 2h at 37C and washed 3x. This was followed by addition of 1:400 streptavidin 488 (SA 488) (Invitrogen) or streptavidin phycoerythrin (SA PE) (Invitrogen) an d 30min incubation at 4C in WB. Cells were then washed once in WB and analyzed by flow cytometry on a FACScan (Becton Dickinson) using FL 1 for SA 488 and FL 2 for SA PE. Data were analyzed using either Win MDI or FCS Express 2.0. All data shown are ungat ed. To study aptamer receptor permanence on the cell surface, CEM cells were incubated with 2M of SA AR for 2h at 37C, followed by washes and incubation at various times from 0 to 30h before adding 1:400 SA 488, washing, and performing flow analysis. Fl uorescence Microscopy Cells were plated at low confluence on Lab Tek four chambered slides and allowed to grow for 24 hours before washing with WB and incubating with 2M of SA ARs for 2h at 37C. Afterwards, the cells were washed again 3x and incubated wi th 1:400 SA 488
120 for 30min at 4C. After washing, cells were imaged at 40x magnification using incubation and imaged using bright field and fluorescence (488ex) microscopy using a Leica DM6000B microscope. MTS Assay MTS works on the principle of a cell's ability to reduce the tetrazolium reagent (Owen's Reagent) via NADH or NADPH when it is alive. The reduced produce absorbs at 490nm and can be read at that wavelength. The higher the is the absorbance, the more viable are the cells. 250k CEM cells were washed and put in sterile flow tubes with 1mL of RPMI media. For each time point, each tube was treated in the same manner. 2mL of PBS was added, and cells were washed by spinning at 1300RPM for three minutes. The cell m edia was removed by pouring out the supernatant, and cells were incubated with either 50uL 500nM streptavidin Lipid #3 for 1 2hours at 37C. 2mL PBS was used to wash the cells. Then the cells in each tube were resuspended in 1mL RPMI. Finally cells were wa shed, resuspe n ded in 100uL RPMI split into three 96 wells to which 20uL MTS reagent CellTiter 96 AQ ueous Non Radioactive Cell Proliferation (Promega)was added. Cells were measured at 490nm after 4hours. The wells were all bright purple. Cell Capture with S treptavidin DynaBeads 12 x10 6 Healthy CEM CCRF cells were washed in 10mL washing buffer. Cells were resuspended in 6 flow tubes with the following treatments: (1) Just Cells. Incubated 1hour RT, washed 2x with 2mL WB. ( control). (2) TD05 biotin 500nM i n 200uL BB. Incubated 30min @ 4C, washed 1x with WB. ( control). (3) Sgc8 Biotin 500nM in 200uL BB. Incubated 30min @ 4C, washed 1x with WB. (+ control). (4) PDGF lipid aptamer 2uM in RPMI media. Incubated 1hour washed 2x with 2mL WB.
121 ( control). ( 5) Streptavidin lipid aptamer 2uM in RPMI media. Incubated 1hour RT, washed 2x with 2mL WB. (6) Just Cells no DynaBead extraction (+ control). Tubes 1 5 were resuspended in 500uL BB. 400uL of each were added to a fresh tube with 50uL DynaBeads (Invitrogen) Cells were placed on a rotator at 4C for 30min. 150L RIPA buffer with protease inhibitor was added to each tube. Cells were shaken at 4C for 30min and sonicated briefly. 150L 2x SDS Lammeli Sample Buffer was added, and boiled for 5 min. 20L sample was added to each lane of a 4 12% Bis Tris Nupage NOVEX Gel (Invitrogen) and run at 200V in MOPs running buffer for 1hour. Blot was transferred at 30V for 1hour on a PDVF membrane. Blots were blocked with 5% milk for one hour then incubated with 1:1000 1 anti rabbit actin (Cell Signaling #4967S) overnight at 4C. The next day blots were washed and probed with 2 goat anti rabbit HRP (Pierce #1858414, 1:2000). Blots were visualized with SuperSignal West Dura Extended Duration Substrate (Thermo Scientific) and imaged on Kodak film. Clotting Assay 500k healthy CEM cells were washed and incubated for two hours at 37C in 50L RPMI media (Tubes 1+2), 5uM TA AR (Tubes 3 6) or PDGF AR (Tube 7). After washing 3x, cells were incubated with 100L WB alone (Tube 1) 500nM thrombin thrombin; Haematologic Technologies, HCT 0020) (Tube 1+7), 500nM thrombin pre incubated with 1M T 15 (Tube 4), 1M scrambled T 15 (Tube 5), or 200nM bovine serum albumin (BSA, Invitrogen) for 30min at RT. After five washes, cells were resuspended in 200L clotting buffer (25mM Tris HCL, 150mM NaCl, 5mM KCl, 1mM MgCl2, 1mM CaCl2, 10% glycerol, pH 7.5). Then, 4L 20mg/mL fibrinogen (fraction I, type I from human plasma; Sigma F3879) was added to Tubes 2 7. Clots
122 formed and were visu alized by adding 50L trypan blue stain (GIBCO). Pictures were taken by placing the tubes on their sides and imaging against a white background with a digital camera.
123 CHAPTER 7 FUTURE DIRECTIONS AND CONCLUSIONS Future Directions The next step in this UTR. A tri nity of experiments would confirm the conclusions suggested by our prior results: (1) Performing a gel shift assay where primers specific for DIXDC1b are us ed to PCR amplify a dsDNA region surrounding the consensus region, and where the dsDNA is run on a gel. If the DIXDC1b region containing the consensus sequence does indeed interact with PTK7 protein, the addition of purified PTK7 extracellular domain to th e amplified dsDNA DIXDC1b region, but not to a control region lacking the consensus sequence, would cause the DNA to run more slowly on a PAGE gel, causing the DNA to shift. (2) Measuring any changes in DIXDC1b mRNA levels compared to control after the fol lowing modulations: a. PTK7 knockdown by siRNA; b. PTK7 overexpression by addition of a PTK7 plasmid; c. addition of sgc8c and KC2D4 aptamers; and d. addition of MT1 MMP, which would cause increased cleavage of PTK7 into sPTK7. (3) Confocal immunocytochemi stry of cells co stained for PTK7 (either with an antibody or an aptamer), a nuclear marker, and a membrane marker. Other useful experiments would determine exactly where the PTK7 aptamers bind the PTK7 protein. The specific binding region could be found b y making a series of PTK7 mutants, each with a progressively shorter N terminal extracellular domain. If mutants lacking a particular Ig like domain cannot bind the aptamers, the location of the missing Ig like domain would be in close proximity to the sit e of aptamer protein
124 interaction. When this is found, a mutant that does not bind the aptamers can be made, and the phenotype for this PTK7 protein can be studied. Lastly, there is one PTK7 aptamer, KDED19, which remains a puzzle. This aptamer competes wit h the other PTK7 aptamers, but does not share sequence identity and/or the rest of the PTK7 binding aptamer family will prove interesting and informative. Conclusion s Test t ube evolution (SELEX) of RNA and DNA aptamers has become common C ell SELEX generates artificial DNA and RNA molecules (aptamers) that bind to biological targets of interest with application s for research and therapy. Using this tool, we DNA aptamer that to breast, and colon cancer cell lines. extracellular portion of the receptor tyrosine kinase PTK7 which is im portant for non canoni cal Wnt signaling, and is mis regulated in numerous cancers. Bioinformatic analysis of aptamer sequence revealed significant seque nce identity between sgc8c and 3 s
125 GCC GCCGGGA) As predicted, all 4 aptamers competed with each other and bound to the cells with similar affinity, bound to the same protein Furthermore, mutational analysis of sgc8c indicated that the co nsensus region, but not the surrounding nucleotides, is important to aptamer binding. repeat selection, as well as its importance in aptamer protein binding, in the human genome. in UTR) of human DIXDC1b DNA, which, al ong with PTK7, is a regulator of non canonical Wnt signaling. Surprisingly, aligning the aptamers with DIXDC1b DNA, we found 2 of the aptamers, KMF9b and H01 had additional nucleotides in common with DIXDC1b DNA outside the consensus region. This 22nt reg ion, shared by the PTK7 aptamers and DIXDC1b DNA, is unique in the human genome. We were also surprised that another aptamer, KC2D4, which also competes with the PTK7 aptamers, but which shares no sequence similarity with them, has identity to the opposite strand of DIXDC1b DNA. In total, 5 aptamers compete for bind ing the same site on PTK7. Four of these aptamers share sequence identity to the positive strand of UTR. The fifth aptamer shares sequence ident ity 5 bases from this region, but on the opposite strand of DIXDC1b DNA. could be consistent with the protein PTK7 melting the genomic DNA and interacting with the res ulting ssDNA hairpins, which share sequence identity to the aptamers formed by each melted strand. If this is indeed the
126 case, it could affect DIXDC1b transcription. These sequences identities are conserved in the putative DIXDC1b regulatory regions of Pan troglodytes Rattus norvegicus and Mus musculus PTK7 has a weak predicted nuclear localization sequence. By cell fractionation, followed by Western blot, we found an abundance of cleaved PTK7 in the nuclei of several different cell lines. This suggests that whole cell SELEX has not just found an aptamer for an extracellular protein, but identified a genomic DNA sequence, which the protein binds to naturally after it is cleaved, internalized, and transported to the nucleus. Consistent with t his hypothesis, a matrix metalloprotease cleavage site that frees the extracellular portion of PTK7 has been recently reported. Adding this protease to PTK7 expressing cells caused PTK7 to be removed from the cell membrane and accumulate around the nucleus [ 117 ] While we have only begun to address this receptor tyrosine kinase and this regulatory pathway, these fi ndings suggest that whole cell SELEX might be used more generally to identify novel extracellular transcription factors with highly specific binding motifs. Our bioinformatic analysis of other aptamer sequences selected against whole cells has yielded othe r examples of disparate selections yielding similar sequences, hinting that there may be other proteins that interact with DNA on the plasma membrane yet to be discovered. Our results are important for three main reasons: (1) The proteins PTK7 and DIXDC1b are murkily understood, yet key players in Wnt signaling, which is crucial to embryonic development and cancer progression. Understanding their relationship could be important for understanding Wnt signaling. For instance, there are two isoforms of
127 DIXDC1, a and b; only DIXDC1b has the consensus sequence. PTK7 binding this of PTK7 binding aptamers has sequence identity to both strands of DI XDC1b DNA, implying that, if melted, the DNA might be forming ssDNA hairpins that bind to PTK7 in a manner similar to the aptamers. This would be a new type of transcription factor of interest to molecular biologists. (3) Finally, this finding is probably not a one time occurrence. Our bioinformatic analysis of 148 DNA sequences selected by whole cell SELEX identified other aptamers from disparate selections like those for Vaccinia infected cells and pure virus, which share significant sequence identity. F urther comparison of existing DNA and RNA aptamers may yield other examples of SELEX identifying natural DNA or RNA sequences with functional roles, not initially envisioned. Future selections should also not be considered complete until the newly selected aptamers are compared with all other existing aptamers for sequence identity.
128 APPENDIX A COMPLEX TARGET SELEX ss DNA APTAMER DATABASE
129 SELEX GQ? Target Name N t Sequence 1 Yes hnRNP A1 BC 15 74 TGTGGCGAGGTAGGTGGGGT GTGTGTGTAT 2 No IgE IgE 21 TTTATCCGTTCCTCCTAGTG G 3 No small cell lung cancer 16 1 25 GAATCCTTCTTTGTCCCGGG CCCGT No small cell lung cancer 0 25 25 TACTCAATTACTCTCTTGTC CCTCT 4 Yes Shp2 Phosphatse HJ24 80 GGGGTTTTGGTGGGGGGGGC TGGGTTGTCTTGGGGGTGGG 5 No Tenascin C GB 10 34 CCCAGAGGGAAGACTTTAGG TTCGGTTCACGTCC 6 Yes Mucin S1.3/S2.2 72 GCAGTTGATCCTTTGGATAC CCTGG 7 No RET Kinase D24 50 GCGCGGGAATAGTATGGAAG GATACGTATACCGTGCAATC CAGGGCAACG 8 Yes Nucleolin AS1411 26 GGTGGTGGTGGTTGTGGTGG TGGTGG 9 Yes IL 17RA RA10 2 30 CTAAGGATCGGATCCACGGC CTACCAGGTC No IL 17RA RA10 6 30 CTTGGATCACCATAGTCGCT AGTCGAGGCT No IL 17RA RA10 7 30 ACGCGCTAGGATCAAAGCTG CACTGAAGTG No IL 17RA RA10 13 30 CCAGAAGAAGCCCACTAGCG TGCTTTTGTC No IL 17RA RA10 14 30 CCAGACGTGAGCACTAGATC AGTACGGAAG 10 Yes NSCLC: A549 v HLAMP S1 45 GGTTGCATGCCGTGGGGAGG GGGGTGGGTTTTATAGCGTA CTCAG Yes NSCLC: A549 v HLAMP S6 45 GTGGCCAGTCACTCAATTGG GTGTAGGGGTGGGGATTGTG GGTTG Yes NS CLC: A549 v HLAMP S11a 45 AGAGTGGGGGGGTGGGTGGA TTTGACAGGTGGCATGCTGG AGAGT Yes NSCLC: A549 v HLAMP S11b 45 TGGGGTTATTAATTTTGGGT GGGGGGGAAGATGTAGCATC CGACG Yes NSCLC: A549 v HLAMP S11c 45 AGCTTGAGGGTGGGCGGGTG GACGCGGTAGTGGTATATAG GTCGG Yes NSCLC: A549 v HLAMP S11d 45 GATCGGTGGGTGGGGGGGTT GGAGATCATCCTCAGGGATT ACGTC Yes NSCLC: A549 v HLAMP S11e 45 ATGCGAACAGGTGGGTGGGT TGGGTGGATTGTTCGGCTTC TTGAT Yes NSCLC: A549 v HLAMP S11f 45 GGTCGCAGATGGATTAAGTA TGTGGGTGGGGGGGTGGAAG TTAAT Yes NSCLC: A549 v HLAMP S15 45 GCTATCTTATGGAAATTTCG TGTAGGGTTTGGTGTGGCGG GGCTA 11 Yes PigPen III.1 96 AGGCGGTGCATTGTGGTTGG TAGTATACATGAGGTTTGGT TGAGACTAGTCGCA 12 No RBC Ghost: CD71 C56t 26 AACTCAGTAATGCCAAGGTA ACGGTT No RBC Ghost Motif 2a 33 CGAATCGCATTGCCCAACGT TGCCCAAGATTCG 13 Yes Differentiated PC12 1 25 TGGTTGGGGATAGAGGTGGG TGTTT Yes Differentiated PC12 2 25 TGAGGGTCTAGGGTGGTGGG GTGGA
130 Yes Differentiated PC12 3 25 TGATGGATGTGGGGATGCGG GGGCG Yes Differentiated PC12 4 25 TATGGGGGTGGGTCAGGTTT CGGTA Yes Differentiated PC12 5 24 GGGAGGTTGGGGTATCAGGG GGGG Yes Differentiated PC12 7 25 GGGTGTGGGAGGTGATGGGG TAGGT Yes Differentiated PC12 8 24 AGGGGGGTTCGGCGGAGGTA TCAG Yes Differentiated PC12 10 25 GCTGGGGTGTTGGGTGTGGG GGTGA No Differentiated PC12 12 25 GTGCGACATAGCTAAACCGG TTCGT Yes Differentiated PC12 13 25 GAGGAGGGAGAATAGGGGTG GGTGG No Differentiated PC12 14 24 AGTCAGACAGGGGGGAGGAT CCGT Yes Differentiated PC12 15 25 TGGGTAGGTTCGAGGGGTGG GTGTG Yes Differentiated PC12 16 25 AGAGTGGGGGGGATGTAGGT GGGTT Yes Differentiated PC12 17 25 GGTTGGATGTAAGGTTGGAG GGGGG No Differentiated PC12 18 25 GTGTCCGTGGACTAAACCGG CCTGT No Differentiated PC12 20 24 GTGGAAGCCTCCTAAGCGGT GTGT No Differentiated PC12 22 24 TGGGTGAGTTCAATGGGGGT ATGT Yes Differentiated PC12 23 25 GGGTGTGAGAGGTTGAGGGG GTTCG 14 No Vaccinia Virus A549 TVO1 25 GTGCATTGAAACTTCTGCAT CCTCG No Vaccinia Virus A550 TVO2 24 CCTGCATATACACTTTGCAT GTGG No Vaccinia Virus A551 TVO4 33 AACCTGCATAATTTATAAGT CTAGACTGCTGCA No Vaccinia Virus A552 TVO6 27 GGACCGATAGGAACCACGGA CTGCATG 15 No Vaccinia Virus Hela PP2 38 ACACCGTTTGTATTCTGCAT TGTTTTGCATTCTACATG (Primer) Vaccinia Virus Hela PP5 31 CACTTGCATATACACTTTGC ATTATAGGGTG 16 No Mucin MUC15TR1 25 GAAGTGAAAATGACAGAACA CAACA No Mucin MUC15TR2 25 GGCTATAGCACATGGGTAAA ACGAC No Mucin MUC15TR3 25 CAAACAATCAAACAGCAGTG GGGTG No Mucin MUC15TR4 25 TACTGCATGCACACCACTTC AACTA 17 No HL60/CEM KH1C12 42 TGCCCTAGTTACTACTACTC TTTTTAGCAAACGCCCTCGC TT No HL60/CEM KHG11 45 TGCTCATCCACGATTCTGGC GAATTTAGTGCCTGTCTTTT TCTCT No HL60 KH2B05 42 CACACAACCTGCTCATAAAC TTTACTCTGCTCGAACCATC TC
131 No Ramos KH1A02 44 GGCATAGATGTGCAGCTCCA AGGAGAAGAAGGAGTTCTGT GTAT No Ramos KK1B10 45 GATCAGTCTATCTTCTCCTG ATGGGTTCCTATTTATAGGT GAAGC No HL60, NB4, K562/CEM KH1B08 45 TTCAAATCACACGACGCATT GAAACACTCTACAATATCAC ATTTA No HL60, NB4, K562/CEM KH3H03 45 CTGGCGCCTTCTACTTCAAG GCAATAAGCTCAATCAATAT CATCG 18 No CEM H01 46 AAGCAGCAGCTGTGCCATCG GGTTCGGATTTTCTTCCTAC GACTGC No CEM H04 45 TATCAAAGGCGAATTTTGTC AAGGTGTTAAACGATAGTCC CTACC No CEM, Ramos, Toledo H11 44 TCGCCTGTACATAGACTGTT GCGTTAGGGTCTGCCTTTAT CTTG No CEM, Ramos, Toledo B07 44 CATAGAGACTTGGATGCAAC TTAGCTACTAACGCTAGCTC TATG 19 Yes Ramos TD05 47 AACACCGTGGAGGATAGTTC GGTGGCTGTTCAGGGTCTCC TCCGGTG No Ramos, CEM, Toledo TD08 85 TACTCTAATTGCCGTATAAG GTCAGGGGGTTGGTTGGTTC CTAGTGCTT No Ramos TE02 44 GCAGTGGTTTGACGTCCGCA TGTTGGGAATAGCCACGCCT CGGG No Ramos, CEM TE04 44 CACTCCTCGATGCACCAGTT CACCTTATTTGCTTCTTCTC TCTG No Ramos, CEM, Toledo TE13 42 GCCCCCAGGCTCGGTGGATG CAAACACATGACTATGGGCC CG No Ramos TE17 52 ACCTGCTTGACCGACCGATA CAGCTACGCAATACAAAACT CCGAACACCTGC 20 No CEM, Ramos, Toledo TC01 33 CCAAACACAGATGCAACCTG ACTTCTAACGTCA No CEM TC02 46 AGCATCAACAAGGTCATAAA AACACGTCAGCTCCTTCACA TTTGCC 21 No CEM, Jurkat sgd3 53 AGGGGGAGCTTGCGCGCATC AAGGTGCTAAACGAAAGCCT CATGGCTTCTATA No CEM, Jurkat, Ramos sgc4a 36 CGAGTGCGGATGCAAACGCC AGACAGGGGGACAGGA No CEM, Jurkat sgc5 45 ACCGACGACGAACTATCTAT CACTATCTTACACATCATAC CTCGA No CEM, Jurkat sgc7 52 ACCGCAGCGACTATCTCGAC TACATTACTAGCTTATACTC CGATCATCTCTA No CEM, Jurkat Sgd2 53 GAGTGAAGCAAGGATGCAAC CTCGGCTCCAACCCGTGAGA GTCGCGAAACTCA Yes CEM, Jurkat, Toledo Sgd5a 66 ACTTATTCAATTATCGTGGG TCACAGCAGCGGTTGTGAGG AAGAAAGGCGGATAACAGAT AATAAG No CEM, Jurkat sgc8c 41 ATCTAACTGCTGCGCCGCCG GGAAAATACTGTACGGTTAG A Yes CEM, Jurkat sgc3b 51 ACTTATTCAATTCCTGTGGG AAGGCTATAGAGGGGCCAGT CTATGAATAAG 22 Yes Liver Cancer mouse: IMEA TLS1c 55 ACAGGAGTGATGGTTGTTAT CTGGCCTCAGAGGTTCTCGG GTGTGGTCACTCCTG No Liver Cancer mouse: IMEA TLS3 45 TGGGAATATTAGTACCGTTA TTCGGACTCCGCCATGACAA TCTGG yes Liver Cancer mouse: IMEA TLS4 45 ACGGTGGTCGTACACGGCCA TTTTATTCCCGGAATATTTG TCAAC No Liver Cancer mouse: IMEA TLS7 45 TGCGCCCAAAGTTCCCATAT TGCTTCCCTGTTGGTGAGTG CCGAT No Liver Cancer mouse: IMEA TLS11a 63 ACAGCATCCCCATGTGAACA ATCGCATTGTGATTGTTACG GTTTCCGCCTCATGGACGTG CTG
132 No Small Cell Lung Cancer HCH07 35 GCCGATGTCAACTTTTTCTA ACTCACTGGTTTTGC 23 No Small Cell Lung Cancer HCA12 35 GTGGATTGTTGTGTTCTGTT GGTTTTTGTGTTGTC Yes Small Cell Lung Cancer HCC03 35 CCGGGGACCGGGGCACCGGG GGCCAGTGGCACGGA No Small Cell Lung Cancer HCH01 71 GTCAACCGAATGCGTCAGCT GGATCTTAAAGATTGCATGC GCTCACTATGGGACTGAGCA TCGCACTGGTA 24 Yes Colon Cancer KMF2 1a 42 GAATAGGGGATGTACAGGTC TGCACCCACTCGAGGAGTGA CT Yes Colon Cancer KMF3 42 AGGATAGCCATACCACCGGG GAGTTTATAACGGTACGGTC CT No Colon Cancer KMF9b 46 AGCGCAGCAGCTGTGCCACC GGGAGAATTTACGTACGGCT GAGCGA 25 No Colon Cancer KDED1 39 CTAAACAAAATACGAGCAGG GAGACTTCTATCCGATTGT Yes Colon Cancer KDED2 55 AACTGCTATTACGTGTGAGA GGAAAGATCACGCGGGTTCG TGGACACGGTGGCTT No Colon Cancer KDED3 39 GGGGTGGTTTTCAAAGAGTC TTGCCTGACTCCCCTGTGG No Colon Cancer KDED7 40 GCGGACGCACTTTTAGCAAG CAAGTCGACAATGGAGGTTT No Colon Cancer KDED9 40 GCAACTGAAGCTAGAACTGT GTGGGGTTTGGGGTATAATT Yes Colon Cancer KDED20 45 TAGGTTGGATAGGGATGGTA GAGCAGGCTAAGCACTTTTT TTTAT 26 Yes Colon Cancer KCHA10a 59 ACGCAGCAGGGGAGGCGAGA GCGCACAATAACGATGGTTG GGACCCAACTGTTTGGACA No Colon Cancer KCHB10 63 ATCCAGAGTGACGCAGCAGA TCTGTGTAGGATCGCAGTGT AGTTGACATTTGATACGACT GGC 27 Yes Colon Cancer KC2D3 37 CGGGAAAGGAACAAACTGCT ATTAGGTCGCAGGCCGG No Colon Cancer KC2D4 37 CCCACTGGTAGCCATTCCGC CCTTAACCGGGCCATCG No Colon Cancer KC2D8 38 AACTGCTGCGCCGCCGGGAA AATACTGTACGGTTAGTT 28 No Ovarian TOV Cells aptTOVl 45 GATCTGTGTAGGATCGCAGT GTAGTGGACATTTGATACGA CTGGC No Ovarian TOV Cells aptTOV2a 42 CAATCTCTACAGGCGCATGT AATATAATGGAGCCTATCCA CG No Ovarian TOV Cells aptTOV3 42 CTCACTCTGACCTTGGATCG TCACATTACATGGGATCATC AG No Ovarian TOV Cells aptTOV4 42 GGCACTCTTCACAACACGAC ATTTCACTACTCACAATCAC TC No Ovarian TOV Cells aptTOV5 42 CAACATCCACTCATAACTTC AATACATATCTGTCACTCTT TC No Ovarian TOV Cells aptTOV6 42 CGGCACTCACTCTTTGTTAA GTGGTCTGCTTCTTAACCTT CA No Ovarian TOV Cells aptTOV7 42 CCAACTCGTACATCCTTCAC TTAATCCGTCAATCTACCAC TC No Ovarian TOV Cells aptTOV8 42 CCAGTCCATCCCAAAATCTG TCGTCACATACCCTGCTGCG CC No Ovarian TOV Cells aptTOV9 42 GCAACACAAACCCAACTTCT TATCTTTTCGTTCACTCTTC TC 29 Yes Ovarian DOV Cells DOV3 37 ATGCAGAGGCTAGGATCTAT AGGTTCGGACGTCGATG Yes Ovarian DOV Cells DOV6 37 AATGTTGGGGTAGGTAGAAG GTGAAGGGGTTTCAGTT
133 30 No Adenocarcinoma: H23 EJD1 42 CCCTCACCACCAAACAACAA TATTAGAGACAATGAGTTCC CT No Adenocarcinoma : H23 EJD2 41 AGTGGTCGAACTACACATCC TTGAACTGCGGAATTATCTA C No Adenocarcinoma: H23 EJD4 41 GAAGACGAGCGGCGAGTGTT ATTACGCTTGGAAACAACCC C No Adenocarcinoma : H23 EJD5 41 TACGGGCTGGATCCACTGTT ACGGCGTGTATCCGCTATCA A No Adenocarcinoma : H23 EJD7 42 CAACTCTTAAGTAAATACCT TTTTCTGGCGTGTAAGAAAA TG No Adenocarcinoma: H23 ADE1 42 GGCAAAGCACGACGACATGG TATTACACGAACTACAATCC CT No Adenocarcinoma: H23 ADE2 42 GAGCCCTATCTCACACCGCA CCCGCAAACTATCATCCTAC AT 31 Yes Cancer Stem Cells: DU145 CSC01 40 AGGTGGTTTGCTGCGGTGGG CTCAAGAAGAAAGCGCAAAG No Cancer Stem Cells: DU146 CSC08 43 GCTCTGAGCCTAGCTTGACC ACTTTTCTTTATTCGCTCTG AGG No Cancer Stem Cells: DU147 CSC13 43 GGGGTGTCGTATCTTTCGTG TCTTATTATTTTCTAGGTGG AGG No Cancer Stem Cells: DU148 CSC17 42 CACCAGCTCCATAACGACAC GACCCTCATTCCAACACACA GG No Cancer Stem Cells: DU149 CSC22 43 GTGGGGCTGTGATACTTTAC ATCTTATTTCTCTAGTGACT AGG 32 Yes Activated Protein C HS02 52G 52 GCCTCCTAACTGAGCTGTAC TCGACTTATCCCGGATGGGG CTCTTAGGAGGC 33 Yes Mesenchymal Stem Cells 1MSC 40 CGACTTCGGTTATTACGTTG TTGGCCTCACAAGGACGCCC Yes Mesenchymal Stem Cells 2MSC 39 CACGATCCAGATGTCATAGT TTAGGCTCTCTCTACTACT Yes Mesenchymal Stem Cells 3MSC 40 GGCGGGAGGTCACGTTGAGA ATTTACGAGGCAGGGGGCAC No Mesenchymal Stem Cells 4MSC 39 GAGGGGCCGCCAAAGCTAGC TCAAGTGATATCCTGTACT No Mesenchymal Stem Cells 5MSC 41 CACCCGTATGCCAAGTCAGA TCCAGTGTAGATGCGCGCCC C No Mesenchymal Stem Cells 6MSC 41 CGACACGCGCACGGTTCTCA TCAATACTGCCTCGCCGGTA C No Mesenchymal Stem Cells 7MSC 38 CAGCATGCAGAGGCGTCAAA TAACGGGACCTCTCGGAC Yes Mesenchymal Stem Cells 8MSC 53 GGGGAGTGGTGGAGAAAGGC TTACAGGGTAGATAAGGTTC AGGTGCTTCGTTC Yes Mesenchymal Stem Cells 9MSC 50 GGGTCATTGCAGGGTAAGGT TGGATTTATTGATGCCTCGG AGTTGGGTGG No Mesenchymal Stem Cells 10MSC 50 GTAGGCGTTGCCTTAGTTAT TGTTTTGAGGTAGAGCAGAG TTTTACTCAG Yes Mesenchymal Stem Cells 11MSC 50 CGAGGTGGATGACAGGGTAT GTGGATTGGTAGTGTGTTTG GTGCTAACGC Yes Mesenchymal Stem Cells 12MSC 50 GGAGGAAGGGTTACGGAGGA AGAGTTAGGATCGGTGGGGA TGATGATGGG Yes Mesenchymal Stem Cells 13MSC 50 GGTTTAATGTGTGGGTAGTT GGGCGTGACGGGGTAGTCCT GGGGGTTAGG Yes Mesenchymal Stem Cells 14MSC 50 GTGGAGTGGCCGTAGTCTGG CCAGGTCCCGTTGGTGATGG GTAGAGTGGG No Mesenchymal Stem Cells 15MSC 50 TTTGCGCTGGATGCGATAAC GTGTTCGACATGAGGCCCGG ATCCACTCCC No Mesenchymal Stem Cells 16MSC 50 TGTGCTTATGCTCGAGATGG TGTTATCCGTGTTGCCACGA TGGGGGGACC
134 Yes Mesenchymal Stem Cells 17MSC 50 TGGATGGGTGGGCGTAGGTG AGGTGTTGTAAGAGCCTCTC CACAGGTGCG Yes Mesenchymal Stem Cells 18MSC 50 TGCTCCAAGGGACAGGGCAA GGGATCTATCCTGCCGCGGG GATGTAAGGC Yes Mesenchymal Stem Cells 19MSC 50 TGGGGGGAAGCGGACTGTTC GCACTTAGGGCGTATGATGG TAGTGGACCG Yes Mesenchymal Stem Cells 20MSC 50 GAGTAATGTAGGGTGAAGGG TGTGGGGGCTATGGGGATAG TGGCACGGCC
135 APPENDIX B LIST OF SIMILARITY BETWEEN APTAMERS IN DATABASE Aptamer 1 Aptamer 2 %Identity p value 95% CI 1 sgc8c KC2D8 97.3 0 3.69E 06 2 KCHB10 aptTOVl 97.78 0 3.69E 06 3 S11c S11f 73.33 2.70E 05 3.93E 05 4 S11a 16 83.33 4.00E 05 0.000102413 5 MUC1 5TR 3 TE02 81.82 6.00E 05 0.00013059 6 TVO2 PP5 84.62 8.00E 05 0.000157626 7 sgc8c KMF9b TMR 72.09 0.0001 0.000183896 8 H01 KMF9b TMR 72.73 0.00011 0.000196812 9 KH1B08 aptTOV4 71.05 0.00015 0.00024739 10 sgc8c 18MSC 71.87 0.00015 0.00024739 11 7 23 80 0.00018 0.000284463 12 KH1A02 KCHB10 68.97 0.00019 0.000296693 13 B07 KDED9 71.43 0.0002 0.000308867 14 AS1411 10 73.91 0.00022 0.000333064 15 S6 S11a 68.75 0.00024 0.00035708 16 KDED9 20MSC 68.29 0.00024 0.00035708 17 RA10 2 20MSC 72.73 0.00027 0.000392812 18 S11b 13MSC 67.44 0.00033 0.000463412 19 sgd3 KDED1 70 0.00034 0.000475084 20 DOV3 12MSC 68.57 0.00035 0.000486732 21 TC01 HCH07 74.07 0.00037 0.000509961 22 BC 15 15 76 0.0004 0.00102384 23 HJ24 TD08 60.76 0.0004 0.00102384 24 1 16 77.27 0.0004 0.00102384 25 3 10 77.78 0.0004 0.00102384 26 17 23 75 0.0004 0.00102384 27 7 20MSC 76 0.0004 0.000544647 28 7 DOV6 75 0.00043 0.000579164 29 5 8 76.19 0.0005 0.00116644 30 S6 7 76 0.00051 0.000670502 31 S1.3/S2.2 KCHA10a 64.15 0.0006 0.00130549 32 AS1411 3 71.43 0.0006 0.00130549 33 15 17 76.19 0.0006 0.00130549 34 RA10 2 15MSC 71.43 0.0007 0.00144173 35 III.1 PP2 65.91 0.0007 0.00144173 36 7 13MSC 73.08 0.0007 0.00144173 37 TVO4 TC01 71.43 0.0007 0.00144173 38 sgc8c 6MSC 69.57 0.0007 0.00144173
136 39 AS1411 S11a 73.33 0.0008 0.00157571 40 S11a MUC1 5TR 3 71.43 0.0008 0.00157571 41 S11c 2 72.73 0.0008 0.00157571 42 sgc5 sgc7 67.35 0.0008 0.00157571 43 D24 RA10 14 70 0.0009 0.00170779 44 S11c 13MSC 63.64 0.0009 0.00170779 45 S11e 11MSC 65.12 0.0009 0.00170779 46 2 7 73.91 0.0009 0.00170779 47 KMF3 TMR KMF9b TMR 66.67 0.0009 0.00170779 48 AS1411 S11e 73.08 0.001 0.00183826 49 S11a 13 72.73 0.001 0.00183826 50 S11f 17 73.33 0.001 0.00183826 51 2 10 75 0.001 0.00183826 52 CSC13 CSC22 66.67 0.001 0.00183826 53 GB 10 DOV3 70.59 0.0011 0.00196735 54 S1 15 73.91 0.0011 0.00196735 55 S1 S11d 65 0.0012 0.00209522 56 S11e 16 76 0.0012 0.00209522 57 10 20MSC 72 0.0012 0.00209522 58 23 20MSC 70.83 0.0012 0.00209522 59 TLS1c HCA12 66.67 0.0012 0.00209522 60 S1 17 72 0.0013 0.00222201 61 1 12MSC 70 0.0013 0.00222201 62 KH2B05 aptTOV9 65.71 0.0013 0.00222201 63 DOV6 20MSC 65.79 0.0013 0.00222201 64 BC 15 S1 63.83 0.0014 0.00234785 65 S11b S11f 64.86 0.0014 0.00234785 66 1 4 73.68 0.0014 0.00234785 67 2 8 75 0.0014 0.00234785 68 2 20MSC 71.43 0.0014 0.00234785 69 16 12MSC 70.37 0.0014 0.00234785 70 aptTOV5 aptTOV7 67.5 0.0014 0.00234785 71 aptTOV7 EJD2 67.57 0.0014 0.00234785 72 HJ24 S11f 61.11 0.0015 0.00247282 73 S11d S11f 64.86 0.0015 0.00247282 74 20 HS02 52G 72.22 0.0015 0.00247282 75 TC02 sgd3 65.12 0.0015 0.00247282 76 KC2D4 aptTOV7 66.67 0.0015 0.00247282 77 RA10 6 B07 69.7 0.0016 0.002597 78 1 DOV6 72.41 0.0016 0.002597 79 S11d S11e 65 0.0017 0.00272047 80 S1 KDED3 65.52 0.0018 0.00284329 81 2 15 72.73 0.0018 0.00284329
137 82 15 KDED20 72.73 0.0018 0.00284329 83 H04 sgd3 64.44 0.0018 0.00284329 84 TLS3 TLS11a 64.58 0.0018 0.00284329 85 HJ24 S6 60.87 0.002 0.00308716 86 2 19MSC 70.83 0.002 0.00308716 87 S1 S11a 63.16 0.0021 0.00320829 88 KH1B08 CSC17 64.71 0.0021 0.00320829 89 Sgd5a 8MSC 62.5 0.0021 0.00320829 90 TLS3 KC2D8 66.67 0.0021 0.00320829 91 KDED9 KCHB10 64.29 0.0021 0.00320829 92 AS1411 S11c 70.37 0.0022 0.00332894 93 S11d 17MSC 63.83 0.0022 0.00332894 94 S15 8 71.43 0.0022 0.00332894 95 CSC13 10MSC 64.1 0.0022 0.00332894 96 1 7 71.43 0.0023 0.00344914 97 5 17 72.73 0.0023 0.00344914 98 12 Sgd5a 70.83 0.0023 0.00344914 99 TVO2 H11 72 0.0023 0.00344914 100 TLS4 EJD2 65.12 0.0023 0.00344914
138 LIST OF REFERENCES 1. Bayrac AT, Sefah K, Parekh P, Bayrac C, Gulbakan B, et al. (2011) In vitro Selection of DNA Aptamers to Glioblastoma Multiforme. ACS Chem Neurosci 2: 175 181. 2. Chen HW, Medley CD, Sefah K, Shangguan D, Tang Z, et al. (2008) Molecular recognition of small cell lung cancer cells using aptamers. ChemMedChem 3: 991 1001. 3. Mallikaratchy P, Tang Z, Kwame S, Meng L, Shangguan D, et al. (2007) Aptamer Directly Evolved from Live Cells Recognizes Membrane Bound Immunoglobin Heavy Mu Chain in Burkitt's Lymphoma Cells. Mol Cell Proteomics 6: 2230 2238. 4 Sefah K, Meng L, Lopez Colon D, Jimenez E, Liu C, et al. (2010) DNA Aptamers as Molecular Probes for Colorectal Cancer Study. PLoS ONE 5: e14269. 5. Sefah K, Tang ZW, Shangguan DH, Chen H, Lopez Colon D, et al. (2009) Molecular recognition of acute myeloid leukemia using aptamers. Leukemia 23: 235 244. 6. Shangguan D, Cao Z, Meng L, Mallikaratchy P, Sefah K, et al. (2008) Cell specific aptamer probes for membrane protein elucidation in cancer cells. J Proteome Res 7: 2133 2139. 7. Shangguan D, Li Y, Tang Z, Cao ZC, Chen HW, et al. (2006) Aptamers evolved from live cells as effective molecular probes for cancer study. Proc Natl Acad Sci U S A 103: 11838 11843. 8. Tang Z, Shangguan D, Wang K, Shi H, Sefah K, et al. (2007) Selection of aptamers for molec ular recognition and characterization of cancer cells. Anal Chem 79: 4900 4907. 9. Van Simaeys D, Lpez Coln D, Sefah K, Sutphen R, Jimenez E, et al. (2010) Study of the Molecular Recognition of Aptamers Selected through Ovarian Cancer Cell SELEX. PLoS ON E 5: e13770. 10. Katoh M (2007) Comparative integromics on non canonical WNT or planar cell polarity signaling molecules: transcriptional mechanism of PTK7 in colorectal cancer and that of SEMA6A in undifferentiated ES cells. Int J Mol Med 20: 405 409. 11. Prebet T, Lhoumeau AC, Arnoulet C, Aulas A, Marchetto S, et al. (2010) The cell polarity PTK7 receptor acts as a modulator of the chemotherapeutic response in acute myeloid leukemia and impairs clinical outcome. Blood 116: 2315 2323. 12. Easty DJ, Mitchel l PJ, Patel K, Florenes VA, Spritz RA, et al. (1997) Loss of expression of receptor tyrosine kinase family genes PTK7 and SEK in metastatic melanoma. Int J Cancer 71: 1061 1065.
139 13. Lu X, Borchers AG, Jolicoeur C, Rayburn H, Baker JC, et al. (2004) PTK7/CC K 4 is a novel regulator of planar cell polarity in vertebrates. Nature 430: 93 98. 14. Uchibori M, Nishida Y, Nagasaka T, Yamada Y, Nakanishi K, et al. (2006) Increased expression of membrane type matrix metalloproteinase 1 is correlated with poor prognos is in patients with osteosarcoma. Int J Oncol 28: 33 42. 15. Kanazawa A, Oshima T, Yoshihara K, Tamura S, Yamada T, et al. (2010) Relation of MT1 MMP gene expression to outcomes in colorectal cancer. Journal of Surgical Oncology 102: 571 575. 16. Cavalheir o BG, Junqueira CR, Brando LG (2010) Expression of membrane type 1 matrix metalloproteinase in medullary thyroid carcinoma: Prognostic implications. Head & Neck 32: 58 67. 17. Estevez MC, Huang YF, Kang H, O'Donoghue MB, Bamrungsap S, et al. (2010) Nanopa rticle aptamer conjugates for cancer cell targeting and detection. Methods Mol Biol 624: 235 248. 18. Kang H, O'Donoghue MB, Liu H, Tan W (2010) A liposome based nanostructure for aptamer directed delivery. Chem Commun (Camb) 46: 249 251. 19. Chen Y, O'Don oghue MB, Huang YF, Kang H, Phillips JA, et al. (2010) A Surface Energy Transfer Nanoruler for Measuring Binding Site Distances on Live Cell Surfaces. J Am Chem Soc 132: 16559 16570. 20. Conrad R, Keranen LM, Ellington AD, Newton AC (1994) Isozyme specific inhibition of protein kinase C by RNA aptamers. J Biol Chem 269: 32051 32054. 21. Hirao I, Spingola M, Peabody D, Ellington AD (1998) The limits of specificity: an experimental analysis with RNA aptamers to MS2 coat protein variants. Mol Divers 4: 75 89. 22. Kelly JA, Feigon J, Yeates TO (1996) Reconciliation of the X ray and NMR structures of the thrombin binding aptamer d(GGTTGGTGTGGTTGG). J Mol Biol 256: 417 422. 23. Macaya RF, Schultze P, Smith FW, Roe JA, Feigon J (1993) Thrombin binding DNA aptamer forms a unimolecular quadruplex structure in solution. Proc Natl Acad Sci U S A 90: 3745 3749. 24. Daniels DA, Chen H, Hicke BJ, Swiderek KM, Gold L (2003) A tenascin C aptamer identified by tumor cell SELEX: systematic evolution of ligands by exponential enrichment. Proc Natl Acad Sci U S A 100: 15416 15421. 25. White R, Rusconi C, Scardino E, Wolberg A, Lawson J, et al. (2001) Generation of species cross reactive aptamers using "toggle" SELEX. Mol Ther 4: 567 573.
140 26. Bock LC, Griffin LC, Latham JA, Verma as EH, Toole JJ (1992) Selection of single stranded DNA molecules that bind and inhibit human thrombin. Nature 355: 564 566. 27. Hicke BJ, Marion C, Chang YF, Gould T, Lynott CK, et al. (2001) Tenascin C aptamers are generated using tumor cells and purifie d protein. J Biol Chem 276: 48644 48654. 28. Wang C, Zhang M, Yang G, Zhang D, Ding H, et al. (2003) Single stranded DNA aptamers that bind differentiated but not parental cells: subtractive systematic evolution of ligands by exponential enrichment. J Biot echnol 102: 15 22. 29. Morris KN, Jensen KB, Julin CM, Weil M, Gold L (1998) High affinity ligands from in vitro selection: complex targets. Proc Natl Acad Sci U S A 95: 2902 2907. 30. Blank M, Weinschenk T, Priemer M, Schluesener H (2001) Systematic evolu tion of a DNA aptamer binding to rat brain tumor microvessels. selective targeting of endothelial regulatory protein pigpen. J Biol Chem 276: 16464 16468. 31. Homann M, Goringer HU (1999) Combinatorial selection of high affinity RNA ligands to live African trypanosomes. Nucleic Acids Res 27: 2006 2014. 32. Xiao Z, Shangguan D, Cao Z, Fang X, Tan W (2008) Cell specific internalization study of an aptamer from whole cell selection. Chemistry 14: 1769 1775. 33. Zadeh JN, Steenberg CD, Bois JS, Wolfe BR, Pierce MB, et al. (2011) NUPACK: Analysis and design of nucleic acid systems. J Comput Chem 32: 170 173. 34. Shangguan D, Cao ZC, Li Y, Tan W (2007) Aptamers evolved from cultured cancer cells reveal molecular differences of cancer cells in patient samples. Clin Chem 53: 1153 1155. 35. Medley CD, Bamrungsap S, Tan W, Smith JE (2011) Aptamer Conjugated Nanoparticles for Cancer Cell Detection. Analytical Chemistry 83: 727 734. 36. Liu X, Dai Q, Austin L, Coutts J, Knowles G, et al. (2008) A One Step Homogeneous Imm unoassay for Cancer Biomarker Detection Using Gold Nanoparticle Probes Coupled with Dynamic Light Scattering. Journal of the American Chemical Society 130: 2780 2782. 37. Chung Hsin W, Chih Kuang Y. One step covalently conjugated aptamer microbubbles for u ltrasound targeted imaging; 2010 11 14 Oct. 2010. pp. 1960 1963. 38. Xu Y, Phillips JA, Yan J, Li Q, Fan ZH, et al. (2009) Aptamer Based Microfluidic Device for Enrichment, Sorting, and Detection of Multiple Cancer Cells. Analytical Chemistry 81: 7436 7442
141 39. Phillips JA, Xu Y, Xia Z, Fan ZH, Tan W (2008) Enrichment of Cancer Cells Using Aptamers Immobilized on a Microfluidic Channel. Analytical Chemistry 81: 1033 1039. 40. Liu Y, Bae SW, Wang K, Hong J I, Zhu Z, et al. (2010) The effects of flow type on aptamer capture in differential mobility cytometry cell separations. Analytica Chimica Acta 673: 95 100. F, Chen Y, et al. (2009) Using Aptamer Conjugated Fluorescence Resonance Energy Transfer Nanoparticles for Mult iplexed Cancer Cell Monitoring. Analytical Chemistry 81: 7009 7014. doped silica nanoparticles increase flow cytometry sensitivity for cancer cell monitoring. Nano Research 2: 448 461 43. Martin JA, Phillips JA, Parekh P, Sefah K, Tan W (2011) Capturing cancer cells using aptamer immobilized square capillary channels. Molecular BioSystems 7: 1720 1727. 44. Shi H, He X, Wang K, Wu X, Ye X, et al. (2011) Activatable aptamer probe for co ntrast enhanced in vivo cancer imaging based on cell membrane protein triggered conformation alteration. Proceedings of the National Academy of Sciences 108: 3900 3905. 45. Huang Y F, Shangguan D, Liu H, Phillips JA, Zhang X, et al. (2009) Molecular Assemb ly of an Aptamer Drug Conjugate for Targeted Drug Delivery to Tumor Cells. ChemBioChem 10: 862 868. 46. Liu H, Zhu Z, Kang H, Wu Y, Sefan K, et al. (2010) DNA Based Micelles: Synthesis, Micellar Properties and Size Dependent Cell Permeability. Chemistry A European Journal 16: 3791 3797. 47. Taghdisi SM, Abnous K, Mosaffa F, Behravan J (2009) Targeted delivery of daunorubicin to T cell acute lymphoblastic leukemia by aptamer. Journal of Drug Targeting 18: 277 281. 48. Yang L, Meng L, Zhang X, Chen Y, Zhu G et al. (2011) Engineering Polymeric Aptamers for Selective Cytotoxicity. Journal of the American Chemical Society 133: 13380 13386. 49. Huang Y F, Chang H T, Tan W (2008) Cancer Cell Targeting Using Multiple Aptamers Conjugated on Nanorods. Analytical Ch emistry 80: 567 572. 50. Wang K, You M, Chen Y, Han D, Zhu Z, et al. (2011) Self Assembly of a Bifunctional DNA Carrier for Drug Delivery. Angewandte Chemie International Edition 50: 6098 6101.
142 51. Tong GJ, Hsiao SC, Carrico ZM, Francis MB (2009) Viral Cap sid DNA Aptamer Conjugates as Multivalent Cell Targeting Vehicles. Journal of the American Chemical Society 131: 11174 11178. 52. Chen T, Shukoor MI, Wang R, Zhao Z, Yuan Q, et al. (2011) Smart Multifunctional Nanostructure for Targeted Cancer Chemotherapy and Magnetic Resonance Imaging. ACS Nano. 53. Kang H, Trondoli AC, Zhu G, Chen Y, Chang Y J, et al. (2011) Near Infrared Light Responsive Core Shell Nanogels for Targeted Drug Delivery. ACS Nano 5: 5094 5099. 54. Mossie K, Jallal B, Alves F, Sures I, Plow man GD, et al. (1995) Colon carcinoma kinase 4 defines a new subclass of the receptor tyrosine kinase family. Oncogene 11: 2179 2184. 55. Park SK, Lee HS, Lee ST (1996) Characterization of the human full length PTK7 cDNA encoding a receptor protein tyrosin e kinase like molecule closely related to chick KLG. J Biochem 119: 235 239. 56. Shnitsar I, Borchers A (2008) PTK7 recruits dsh to regulate neural crest migration. Development 135: 4015 4024. 57. Paudyal A, Damrau C, Patterson VL, Ermakov A, Formstone C, et al. (2010) The novel mouse mutant, chuzhoi, has disruption of Ptk7 protein and exhibits defects in neural tube, heart and lung development and abnormal planar cell polarity in the ear. BMC Dev Biol 10: 87. 58. Yen WW, Williams M, Periasamy A, Conaway M, Burdsal C, et al. (2009) PTK7 is essential for polarized cell motility and convergent extension during mouse gastrulation. Development 136: 2039 2048. 59. Korinek V, Barker N, Morin PJ, van Wichen D, de Weger R, et al. (1997) Constitutive transcriptional activation by a beta catenin Tcf complex in APC / colon carcinoma. Science 275: 1784 1787. 60. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, et al. (1997) Activation of beta catenin Tcf signaling in colon cancer by mutations in beta catenin or APC. Science 275: 1787 1790. 61. Puppo F, Thome V, Lhoumeau AC, Cibois M, Gangar A, et al. (2011) Protein tyrosine kinase 7 has a conserved role in Wnt/beta catenin canonical signalling. EMBO Rep 12: 43 49. 62. Peradziryi H, Kaplan NA, Podleschny M, Liu X, Wehner P, et al. (2011) PTK7/Otk interacts with Wnts and inhibits canonical Wnt signalling. EMBO J 30: 3729 3740.
143 63. Jennifer A Z (2007) Planar Polarity and Tissue Morphogenesis. Cell 129: 1051 1063. 64. Paul N A (2002) Planar Signaling and Morphogenesis in Drosophila. Developmental Cell 2: 525 535. 65. Wehner P, Shnitsar I, Urlaub H, Borchers A (2011) RACK1 is a novel interaction partner of PTK7 that is required for neural tube closure. Development 138: 1321 1327. 66. Toyofuku T, Zhang H, Kumanogoh A, Tak egahara N, Suto F, et al. (2004) Dual roles of Sema6D in cardiac morphogenesis through region specific association of its receptor, Plexin A1, with off track and vascular endothelial growth factor receptor type 2. Genes Dev 18: 435 447. 67. Lee HK, Chauhan SK, Kay E, Dana R (2011) Flt 1 regulates vascular endothelial cell migration via a protein tyrosine kinase 7 dependent pathway. Blood 117: 5762 5771. 68. Meng L, Sefah K, O'Donoghue MB, Zhu G, Shangguan D, et al. (2010) Silencing of PTK7 in Colon Cancer Cells: Caspase 10 Dependent Apoptosis via Mitochondrial Pathway. PLoS ONE 5: e14018. 69. Prebet T, Lhoumeau A C, Arnoulet C, Aulas A, Marchetto S, et al. (2010) The cell polarity PTK7 receptor acts as a modulator of the chemotherapeutic response in acute m yeloid leukemia and impairs clinical outcome. Blood 116: 2315 2323. 70. Shin W S, Maeng Y S, Jung J W, Min J K, Kwon Y G, et al. (2008) Soluble PTK7 inhibits tube formation, migration, and invasion of endothelial cells and angiogenesis. Biochemical and Bio physical Research Communications 371: 793 798. 71. Haines CJ, Giffon TD, Lu LS, Lu X, Tessier Lavigne M, et al. (2009) Human CD4+ T cell recent thymic emigrants are identified by protein tyrosine kinase 7 and have reduced immune function. J Exp Med 206: 27 5 285. 72. Miller MA, Steele RE (2000) Lemon encodes an unusual receptor protein tyrosine kinase expressed during gametogenesis in Hydra. Dev Biol 224: 286 298. 73. Robinson DR, Wu YM, Lin SF (2000) The protein tyrosine kinase family of the human genome. O ncogene 19: 5548 5557. 74. Wallingford JB, Habas R (2005) The developmental biology of Dishevelled: an enigmatic protein governing cell fate and cell polarity. Development 132: 4421 4436.
144 75. Tada M, Smith JC (2000) Xwnt11 is a target of Xenopus Brachyury: regulation of gastrulation movements via Dishevelled, but not through the canonical Wnt pathway. Development 127: 2227 2238. 76. Kobus FJ, Fleming KG (2005) The GxxxG containing transmembrane domain of the CCK4 oncogene does not encode preferential self i nteractions. Biochemistry 44: 1464 1470. 77. Rudolph MJ, Gergen JP (2001) DNA binding by Ig fold proteins. Nat Struct Biol 8: 384 386. 78. Grassot J, Gouy M, Perriere G, Mouchiroud G (2006) Origin and molecular evolution of receptor tyrosine kinases with i mmunoglobulin like domains. Mol Biol Evol 23: 1232 1241. 79. Golubkov VS, Chekanov AV, Cieplak P, Aleshin AE, Chernov AV, et al. (2010) The Wnt/planar cell polarity protein tyrosine kinase 7 (PTK7) is a highly efficient proteolytic target of membrane type 1 matrix metalloproteinase: implications in cancer and embryogenesis. J Biol Chem 285: 35740 35749. 80. Holmbeck K, Bianco P, Caterina J, Yamada S, Kromer M, et al. (1999) MT1 MMP deficient mice develop dwarfism, osteopenia, arthritis, and connective tissu e disease due to inadequate collagen turnover. Cell 99: 81 92. 81. Jung JW, Ji AR, Lee J, Kim UJ, Lee ST (2002) Organization of the human PTK7 gene encoding a receptor protein tyrosine kinase like molecule and alternative splicing of its mRNA. Biochim Biop hys Acta 1579: 153 163. 82. Li N, Ebright JN, Stovall GM, Chen X, Nguyen HH, et al. (2009) Technical and biological issues relevant to cell typing with aptamers. J Proteome Res 8: 2438 2448. 83. Nagase T, Kikuno R, Hattori A, Kondo Y, Okumura K, et al. (20 00) Prediction of the Coding Sequences of Unidentified Human Genes. XIX. The complete Sequences of 100 New cDNA Clones from Brain Which Code for Large Proteins in vitro. DNA Research 7: 347 355. 84. Hu Y, Galkin AV, Wu C, Reddy V, Su AI (2011) CAFET Algori thm Reveals Wnt/PCP Signature in Lung Squamous Cell Carcinoma. PLoS ONE 6: e25807. 85. Ikeuchi Y, Stegmller J, Netherton S, Huynh MA, Masu M, et al. (2009) A SnoN Ccd1 Pathway Promotes Axonal Morphogenesis in the Mammalian Brain. The Journal of Neuroscien ce 29: 4312 4321. 86. Soma K, Shiomi K, Keino Masu K, Masu M (2006) Expression of mouse Coiled coil DIX1 (Ccd1), a positive regulator of Wnt signaling, during embryonic development. Gene Expression Patterns 6: 325 330.
145 87. Shiomi K, Uchida H, Keino Masu K, Masu M (2003) Ccd1, a Novel Protein with a DIX Domain, Is a Positive Regulator in the Wnt Signaling during Zebrafish Neural Patterning. Current Biology 13: 73 77. 88. Liu Y T, Dan Q J, Wang J, Feng Y, Chen L, et al. (2011) Molecular Basis of Wnt Activatio n via the DIX Domain Protein Ccd1. Journal of Biological Chemistry 286: 8597 8608. 89. Wang L, Li H, Chen Q, Zhu T, Zhu H, et al. (2010) Wnt signaling stabilizes the DIXDC1 protein through decreased ubiquitin dependent degradation. Cancer Science 101: 700 706. 90. Wong CK, Luo W, Deng Y, Zou H, Ye Z, et al. (2004) The DIX Domain Protein Coiled coil DIX1 Inhibits c Jun N terminal Kinase Activation by Axin and Dishevelled through Distinct Mechanisms. Journal of Biological Chemistry 279: 39366 39373. 91. L uo W, Zou H, Jin L, Lin S, Li Q, et al. (2005) Axin Contains Three Separable Domains That Confer Intramolecular, Homodimeric, and Heterodimeric Interactions Involved in Distinct Functions. Journal of Biological Chemistry 280: 5054 5060. 92. Wu Y, Jing X, M a X, Wu Y, Ding X, et al. (2009) DIXDC1 co localizes and interacts tubulin in HEK293 cells. Cell Biology International 33: 697 701. 93. Katoh M (2003) Identification and characterization of human KIAA1391 and mouse Kiaa1391 genes encoding novel RhoG AP family proteins with RA domain and ANXL repeats. Int J Oncol 23: 1471 1476. 94. Wang X, Zheng L, Zeng Z, Zhou G, Chien J, et al. (2006) DIXDC1 isoform, l DIXDC1, is a novel filamentous actin binding protein. Biochemical and Biophysical Research Communic ations 347: 22 30. 95. Katoh M (2003) KIAA1735 gene on human chromosome 11q23.1 encodes a novel protein with myosine tail homologous domain and C terminal DIX domain. Int J Oncol 23: 145 150. 96. Shiomi K, Kanemoto M, Keino Masu K, Yoshida S, Soma K, et al (2005) Identification and differential expression of multiple isoforms of mouse Coiled coil DIX1 (Ccd1), a positive regulator of Wnt signaling. Molecular Brain Research 135: 169 180. 97. Wallin E, von Heijne G (1998) Genome wide analysis of integral memb rane proteins from eubacterial, archaean, and eukaryotic organisms. Protein science : a publication of the Protein Society 7: 1029 1038. 98. Zhang KJ, Tang LL, Sefah K, Zhao ZL, Zhu GZ, et al. (2011) Novel aptamer developed for breast cancer cell internali zation. ChemMedChem In press.
146 99. Wallin E, Heijne GV (1998) Genome wide analysis of integral membrane proteins from eubacterial, archaean, and eukaryotic organisms. Protein Science 7: 1029 1038. 100. Sefah K, Shangguan D, Xiong X, O'Donoghue MB, Tan W (20 10) Development of DNA aptamers using Cell SELEX. Nat Protocols 5: 1169 1185. 101. Cai Y, Sun Y (2011) ESPRIT Tree: hierarchical clustering analysis of millions of 16S rRNA pyrosequences in quasilinear computational time. Nucleic Acids Res 39: e95. 102. Se fah K (2011) KMF9b Selection against MCF7 10 AT 1 cells. Unpublished. 103. Tang Z (2004) HO1 Selection against CEM cells. Unpublished. 104. Phillips Ja, Liu H, O'Donoghue MB, Xiong X, Wang R, et al. (2011) Using azobenzene incorporated DNA aptamers to prob e molecular binding interactions. Bioconjugate chemistry 22: 282 288. 105. Xing Y, Yang Z, Liu D (2011) A Responsive Hidden Toehold To Enable Controllable DNA Strand Displacement Reactions. Angewandte Chemie International Edition: n/a n/a. 106. You M, Zhu Z, Liu H, Gulbakan B, Han D, et al. (2010) Pyrene assisted efficient photolysis of disulfide bonds in DNA based molecular engineering. ACS Appl Mater Interfaces 2: 3601 3605. 107. Lustig B, Jerchow B, Sachs M, Weiler S, Pietsch T, et al. (2002) Negative Fe edback Loop of Wnt Signaling through Upregulation of Conductin/Axin2 in Colorectal and Liver Tumors. Molecular and Cellular Biology 22: 1184 1193. 108. Seo Y J, Chen S, Nilsen Hamilton M, Levine Ha (2010) A mathematical analysis of multiple target SELEX. B ulletin of mathematical biology 72: 1623 1665. 109. Altschul SF, Madden TL, Schffer AA, Zhang J, Zhang Z, et al. (1997) Gapped BLAST and PSI BLAST: a new generation of protein database search programs. Nucleic Acids Research 25: 3389 3402. 110. Kozak M (1 989) The scanning model for translation: an update. The Journal of Cell Biology 108: 229 241. 111. Moss SP, Joyce DA, Humphries S, Tindall KJ, Lunt DH (2011) Comparative Analysis of Teleost Genome Sequences Reveals an Ancient Intron Size Expansion in the Z ebrafish Lineage. Genome Biology and Evolution 3: 1187 1196. 112. Rudolph MJ, Gergen JP (2001) DNA binding by Ig fold proteins. Nat Struct Mol Biol 8: 384 386.
147 113. Bravo J, Li Z, Speck NA, Warren AJ (2001) The leukemia associated AML1 (Runx1) CBF[beta] co mplex functions as a DNA induced molecular clamp. Nat Struct Mol Biol 8: 371 378. 114. Rost B, Yachdav G, Liu J (2004) The PredictProtein server. Nucleic Acids Research 32: W321 W326. 115. Nguyen Ba A, Pogoutse A, Provart N, Moses A (2009) NLStradamus: a simple Hidden Markov Model for nuclear localization signal prediction. BMC Bioinformatics 10: 202. 116. Hsu S C, Hung M C (2007) Characterization of a Novel Tripartite Nuclear Localization Sequence in the EGFR Family. Journal of Biological Chemistry 282: 1 0432 10440. 117. Golubkov VS, Chekanov AV, Cieplak P, Aleshin AE, Chernov AV, et al. (2010) The Wnt/Planar Cell Polarity Protein tyrosine Kinase 7 (PTK7) Is a Highly Efficient Proteolytic Target of Membrane Type 1 Matrix Metalloproteinase. Journal of Biological Chemistry 285: 35740 35749. 118. Parekh P, Tang Z, Turner PC, Moyer RW, Tan W (2010) Aptamers Recognizing Glycosylated Hemagglutinin Expressed on the Surface of Vaccinia Virus Infected Cells. Analytical Chemistry 82: 8642 8649. 119. Tang Z, Pare kh P, Turner P, Moyer RW, Tan W (2009) Generating Aptamers for Recognition of Virus Infected Cells. Clin Chem 55: 813 822. 120. Nitsche A, Kurth A, Dunkhorst A, Panke O, Sielaff H, et al. (2007) One step selection of Vaccinia virus binding DNA aptamers by MonoLEX. BMC Biotechnol 7: 48. 121. Wang C, Zhang M, Yang G, Zhang D, Ding H, et al. (2003) Single stranded DNA aptamers that bind differentiated but not parental cells: subtractive systematic evolution of ligands by exponential enrichment. Journal of Biot echnology 102: 15 22. 122. Schafer R, Wiskirchen J, Guo K, Neumann B, Kehlbach R, et al. (2007) Aptamer based isolation and subsequent imaging of mesenchymal stem cells in ischemic myocard by magnetic resonance imaging. Rofo 179: 1009 1015. 123. Lhoumeau A C, Puppo F, Prebet T, Kodjabachian L, Borg JP (2011) PTK7: a cell polarity receptor with multiple facets. Cell Cycle 10: 1233 1236. 124. Jiang Y, Zhu C, Ling L, Wan L, Fang X, et al. (2003) Specific aptamer protein interaction studied by atomic force micro scopy. Anal Chem 75: 2112 2116. 125. Basnar B, Elnathan R, Willner I (2006) Following aptamer thrombin binding by force measurements. Anal Chem 78: 3638 3642.
148 126. Strunz T, Oroszlan K, Schfer R, Gntherodt H J (1999) Dynamic force spectroscopy of single DNA molecules. Proceedings of the National Academy of Sciences 96: 11277 11282. 127. Pope LH, Davies MC, Laughton CA, Roberts CJ, Tendler SJ, et al. (2001) Force induced melting of a short DNA double helix. Eur Biophys J 30: 53 62. 128. Nguyen T H, Steinbo ck LJ, Butt H Jr, Helm M, Berger Rd (2011) Measuring Single Small Molecule Binding via Rupture Forces of a Split Aptamer. Journal of the American Chemical Society 133: 2025 2027. 129. Sattin BD, Pelling AE, Goh MC (2004) DNA base pair resolution by single molecule force spectroscopy. Nucleic Acids Research 32: 4876 4883. 130. Grisshammer R (2006) Understanding recombinant expression of membrane proteins. Current Opinion in Biotechnology 17: 337 340. 131. Legler DF, Doucey MA, Schneider P, Chapatte L, Bender FC, et al. (2005) Differential insertion of GPI anchored GFPs into lipid rafts of live cells. Faseb J 19: 73 75. 132. McELHINNY AS, Ginger E E, Carol M W (2000) Pa inting Qa 2 o nto Preimplantation Embryos Increases the Rate of Cleavage. Ame rican Journal of Reproductive Immunology 44: 52 58. 133. Daniel RDP, Yoshihiro F, Daniel S, Elaine B, Terrone LR, et al. (2001) Properties of exogenously added GPI anchored proteins following their incorporation into cells. Journal of Cellular Biochemistry 82: 234 245. 134. Koichi K, Chika I, Tohru Y, Teruyuki N (2004) Rapid Protein Anchoring into the Membranes of Mammalian Cells Using Oleyl Chain and Poly(ethylene glycol) Derivatives. Biotechnology Progress 20: 897 904. 135. Chen A, Zheng G, Tykocinski ML (2000) Hierarchical Costimulator Thresholds for Distinct Immune Responses: Application of a Novel Two Step Fc Fusion Protein Transfer Method. J Immunol 164: 705 711. 136. Saxon E, Bertozzi CR (2000) Cell Surface Engineering by a Modified Staudinger Reactio n. Science 287: 2007 2010. 137. Borisenko GG, Zaitseva MA, Chuvilin AN, Pozmogova GE (2009) DNA modification of live cell surface. Nucleic Acids Res 37: e28. 138. Hsiao SC, Shum BJ, Onoe H, Douglas ES, Gartner ZJ, et al. (2009) Direct cell surface modifica tion with DNA for the capture of primary cells and the investigation of myotube formation on defined patterns. Langmuir 25: 6985 6991.
149 139. Wu Y, Sefah K, Liu H, Wang R, Tan W (2009) DNA aptamer micelle as an efficient detection/delivery vehicle toward can cer cells. Proc Natl Acad Sci U S A 107: 5 10. 140. Bing T, Yang X, Mei H, Cao Z, Shangguan D (2010) Conservative secondary structure motif of streptavidin binding aptamers generated by different laboratories. Bioorganic & Medicinal Chemistry 18: 1798 1805
150 BIOGRAPHICAL SKETCH engineer. From the age of one, she grew up overseas with her family. For college she returned to the United States, studying at Duke University where she majored in botany and Chinese. After graduation, she sent herself down to do manual work in a hospital emergency room, a farm then a zoo. After three years, she returned to formal education, becoming fascinated by nanoparticles and aptamers while working toward her Ph.D. at the University of Florida. She is currently a post doc at the NIH studying virus cell interactions. After spending so much time trying to functionalize nanoparticles,