<%BANNER%>

Development and Applications of Miniaturized Protein Expression Systems

Permanent Link: http://ufdc.ufl.edu/UFE0021587/00001

Material Information

Title: Development and Applications of Miniaturized Protein Expression Systems
Physical Description: 1 online resource (135 p.)
Language: english
Creator: Mei, Qian
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: array, detection, microfluidic, miniaturization, protein, synthesis, toxin
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Recent development in miniaturization provides an opportunity for performing chemical and biomedical research in a novel, fast and unique way. This dissertation presents miniaturized plastic devices for in vitro protein expressions (synthesis), as well as their downstream applications. Biological synthesis of a protein includes the steps of gene transcription and protein translation, which are typically carried out in host cells. However, some proteins are difficult to be synthesized in cells due to their insolubility, degradation and cytotoxicity. To address some of these challenges, cell-free protein synthesis, a process called in vitro transcription and translation (IVT), has been developed. In this dissertation, IVT is first demonstrated in a microwell array for expression of proteins, including green fluorescent protein (GFP), chloramphenicol acetyl-transferase (CAT) and luciferase. The array device is then demonstrated for detecting toxin simulants, tetracycline (TC) and cycloheximide (CH), based on their inhibitory effects on protein synthesis. The differential response patterns from the array device for two toxin simulants suggest the feasibility of the detection concept. One drawback of the microwell array is that it does not consist of any fluid manipulation. To address this, a nested-well array is then developed. The nested-well device possesses a mechanism to supply nutrients continuously and remove byproducts, leading to higher protein expression yields. The productions of green fluorescent protein, chloramphenicol acetyl-transferase and luciferase are demonstrated in the device. In addition, the device is demonstrated for ricin detection. The calibration curve has been obtained between the luciferase expression yield and the ricin concentration. Finally, a microfluidic device consisting of a straight channel with herringbone ridges at the bottom wall and a spiral channel is designed and fabricated to demonstrate enzymatic reactions. The device with the ridge feature has shown better mixing than one with a smooth wall when luminescent enzymatic reaction is implemented in the device. Morever, the device can be used to detect luciferse synthesized off chip.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Qian Mei.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Fan, Zhonghui H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021587:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021587/00001

Material Information

Title: Development and Applications of Miniaturized Protein Expression Systems
Physical Description: 1 online resource (135 p.)
Language: english
Creator: Mei, Qian
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: array, detection, microfluidic, miniaturization, protein, synthesis, toxin
Mechanical and Aerospace Engineering -- Dissertations, Academic -- UF
Genre: Mechanical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Recent development in miniaturization provides an opportunity for performing chemical and biomedical research in a novel, fast and unique way. This dissertation presents miniaturized plastic devices for in vitro protein expressions (synthesis), as well as their downstream applications. Biological synthesis of a protein includes the steps of gene transcription and protein translation, which are typically carried out in host cells. However, some proteins are difficult to be synthesized in cells due to their insolubility, degradation and cytotoxicity. To address some of these challenges, cell-free protein synthesis, a process called in vitro transcription and translation (IVT), has been developed. In this dissertation, IVT is first demonstrated in a microwell array for expression of proteins, including green fluorescent protein (GFP), chloramphenicol acetyl-transferase (CAT) and luciferase. The array device is then demonstrated for detecting toxin simulants, tetracycline (TC) and cycloheximide (CH), based on their inhibitory effects on protein synthesis. The differential response patterns from the array device for two toxin simulants suggest the feasibility of the detection concept. One drawback of the microwell array is that it does not consist of any fluid manipulation. To address this, a nested-well array is then developed. The nested-well device possesses a mechanism to supply nutrients continuously and remove byproducts, leading to higher protein expression yields. The productions of green fluorescent protein, chloramphenicol acetyl-transferase and luciferase are demonstrated in the device. In addition, the device is demonstrated for ricin detection. The calibration curve has been obtained between the luciferase expression yield and the ricin concentration. Finally, a microfluidic device consisting of a straight channel with herringbone ridges at the bottom wall and a spiral channel is designed and fabricated to demonstrate enzymatic reactions. The device with the ridge feature has shown better mixing than one with a smooth wall when luminescent enzymatic reaction is implemented in the device. Morever, the device can be used to detect luciferse synthesized off chip.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Qian Mei.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Fan, Zhonghui H.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021587:00001


This item has the following downloads:


Full Text

PAGE 1

1 DEVELOPMENT AND APPLICATIONS OF MINIATURIZED PROTEIN EXPRESSION SYSTEMS By QIAN MEI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

2 2007 Qian Mei

PAGE 3

3 To my grandfather, Chongjia Mei, who passe d away in 2007. He will be always missed and I cherish the loving memories I had with him. To my beloved parents, Yuzhu Zhou and Guifang Mei, who have unconditionally been supporting me through my entire life. They are my model for responsibility, strength, persistence and personal sacrifices. Without their encouragem ent and guidance, I would not have the goals I have achieved. I am so proud that I am their daughter. To my husband and best friend, Hong Wang, whos e constant love, encouragement, endless patience and passion enable me to complete this work. To my dear sister, Xin Zhou, who has been always there for me, whenever good or bad time.

PAGE 4

4 ACKNOWLEDGMENTS It is my pleasure to express my gratitude to all those people without whom this dissertation would not be possible. First and foremost I gratefully acknowledge my supervisor, Dr. Z. Hugh Fan, for his inspiration, guidance, mentoring, and support throughout my thesis. Dr. Fan has been cultivating an interdisciplinary environment by populating th e laboratory with people from all different backgrounds, working on a number of different pr ojects and introducing all sorts of new ideas from various fields. I also thank him for providi ng me an opportunity to grow as a student and engineer in the unique research environment he creates. My next acknowledgement is to Dr. Steven S oper in Chemistry Department at Louisiana State University for providing me an opportunity to visit his lab and access the state-of-the-art fabrication facilities during the la st year of my Ph.D program. Ch apter 4 of this dissertation has been finished in his research lab. I appreciate Dr. Sopers generous supp ort and supervision, and all scientific discussion and creative suggestions from his research group. I thank Drs. Shouguang Jin, James Klausner, Malisa Sarntinoranont, and Roger Tran-SonTay for their time to serve on my supervisory committee and provide valuable comments and advice. I thank Dr. Brian Cain for the access of a luminometer and Dr. Weihong Tan for the access of a fluorescence spectrometer. I acknowledge Drs. Nancy Denslow, Quangfan Chen, Philip Laipis, Wei Lian, and Xinchun Shen for useful discussion. My deep appreciation also goes to Carl Fredrickson and Andrew Simon for their contributions to the development and fabricat ion of devices used in this dissertation.

PAGE 5

5 I thank Dr. Fans entire Research Group, bot h former and current members, for their support and friendship. I sincerel y cherish the moments we worked together as colleagues and friends. Lastly, and most importantly, I wish to th ank my parents, friends and entire extended family for providing a loving environment during my life.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............13 CHAPTER 1 INTRODUCTION..................................................................................................................15 1.1 Miniaturized Device and Its Applications.....................................................................15 1.1.1 Micro-Electro-Mechanical Systems (MEMS)..................................................16 1.1.2 BioMEMS, Lab-on-a-chip and Micro Total Analysis Systems (TAS).............17 1.2 Protein Expression........................................................................................................21 1.2.1 DNA, RNA, Protein and Central Dogma..........................................................21 1.2.2 In Vivo and In Vitro (Cell Free) Protein Expression.........................................22 1.3 In Vitro Protein Expression on a Microchip.................................................................24 1.4 Literature Review and Current Challenges...................................................................26 1.5 Objective and Organizati on of This Dissertation..........................................................28 2 IN VITRO PROTEIN EXPRESSION IN A MI CROWELL DEVICE A ND ITS USE FOR TOXIN DETECTION P ............................................................................................................150H34 15H2.1 Introduction............................................................................................................... ....151H34 16H2.2 Background................................................................................................................. ..152H35 17H2.2.1 Prokaryotic In Vitro Transcription/Translation (IVT) System..........................153H35 18H2.2.2 Eukaryotic IVT System.....................................................................................154H36 19H2.2.3 Choice of IVT Systems.....................................................................................155H36 20H2.3 Experimental............................................................................................................... ..156H37 21H2.3.1 Reagents and Materials.....................................................................................157H37 22H2.3.2 Device Fabrication............................................................................................158H38 23H2.3.3 DNA Template for Cell-Free Protein Synthesis...............................................159H38 24H2.3.4 Cell-Free Protein Synthesis...............................................................................160H39 25H2.3.5 Toxin Inhibition Assay......................................................................................161H40 26H2.3.6 Detection...........................................................................................................162H41 27H2.4 Results and Discussion..................................................................................................163H42 28H2.4.1 Toxin Detection Scheme...................................................................................164H42 29H2.4.2 Plasmid-based Protein Expression....................................................................165H43 30H2.4.3 Linear DNA Template-based Protein Expression.............................................166H44 31H2.4.4 Reaction Temperature.......................................................................................167H46 32H2.4.5 Inhibitory Effects of To xins on Protein Synthesis............................................168H46 33H2.4.6 Miniaturized IVT Array....................................................................................169H48 34H2.5 Conclusions................................................................................................................ ...170H50

PAGE 7

7 3 35HNESTED-WELL DEVICES FOR IN VITR O PROTEIN SYNTHESIS AND RICIN DETECTION...................................................................................................................... ....171H59 36H3.1 Introduction............................................................................................................... ....172H59 37H3.2 Background................................................................................................................. ..173H61 38H3.2.1 Ricin..................................................................................................................174H61 39H3.2.2 Continuous-Flow Cell-Free Pr otein Expression Systems.................................175H62 40H3.2.3 Continuous-Exchange Cell-Free System..........................................................176H63 41H3.3 Experimental............................................................................................................... ..177H64 42H3.3.1 Materials and Reagents.....................................................................................178H64 43H3.3.2 Device Fabrication............................................................................................179H65 44H3.3.3 In vitro Protein Expression...............................................................................180H66 45H3.3.4 Ricin Detection.................................................................................................181H67 46H3.4 Results and Discussion..................................................................................................182H68 47H3.4.1 Device Design...................................................................................................183H68 48H3.4.2 Protein Synthesis...............................................................................................184H70 49H3.4.3 Hydrostatic Flow...............................................................................................185H72 50H3.4.4 Effects of Membrane Pore Size.........................................................................186H75 51H3.4.5 Biological Signal Amplification.......................................................................187H76 52H3.4.6 Detection of Ricin by In Vitro Protein Synthesis..............................................188H77 53H3.4.7 Toxicity Level of Ricin.....................................................................................189H78 54H3.5 Conclusions................................................................................................................ ...190H79 4 55HFABRICATION OF A MICROFLUIDIC DE VICE AND ITS APPLICATION TO BIOLUMINESCENCE DETECTION...................................................................................191H90 56H4.1 Introduction............................................................................................................... ....192H90 57H4.2 Experimental............................................................................................................... ..193H92 58H4.2.1 Materials and Reagents.....................................................................................194H92 59H4.2.2 Device Design and Fabrication.........................................................................195H92 60H4.2.3 Mixing of Phenolphthalein and NaOH.............................................................196H93 61H4.2.4 Simulation.........................................................................................................197H94 62H4.2.5 Bioluminescence Reaction................................................................................198H95 63H4.3 Results and Discussion.....................................................................................................199H96 64H4.3.1 Device Fabrication...........................................................................................200H96 65H4.3.2 Embossing Fidelity...........................................................................................201H97 66H4.3.3 Holes Drilling....................................................................................................202H97 67H4.3.4 Reaction of Phenolphthalein and NaOH...........................................................203H98 68H4.3.5 Effect of Herringbone Ridges...........................................................................204H99 69H4.3.6 Simulation Results..........................................................................................205H101 70H4.3.7 Detection of Limit...........................................................................................206H102 71H4.3.8 Effect of Flow Rate.........................................................................................207H102 72H4.3.9 Detection of Synthesized Luciferase...............................................................208H103 73H4.4 Conclusion................................................................................................................. .209H104 5 74HSUMMARY AND FUTURE DIRECTION.........................................................................210H116

PAGE 8

8 75H5.1 Summary.................................................................................................................... .211H116 76H5.2 Future Work................................................................................................................212H118 77H5.2.1 Device Design and Fabrication.......................................................................213H118 78H5.2.2 Protein Expression System..............................................................................214H119 79H5.2.3 Protein Detection.............................................................................................215H120 80H5.2.4 Applications....................................................................................................216H120 81HLIST OF REFERENCES.............................................................................................................217H122 82HBIOGRAPHICAL SKETCH.......................................................................................................218H135

PAGE 9

9 LIST OF TABLES Table page 83H1-1 RNA codon table............................................................................................................ ....219H31 84H2-1 Principal composition of prokaryotic IVT systems...........................................................220H51 85H2-2 Pricinpal composition of eukaryotic IVT systems.............................................................221H51 86H2-3 Prokaryotic versus eukaryotic IVT systems......................................................................222H52 87H2-4 Oligonucleotide primers used for mutant GFP..................................................................223H52 88H3-1 Summary of Methods fo r detection of ricin.......................................................................224H80 89H3-2 Summary of IVT system....................................................................................................225H81

PAGE 10

10 LIST OF FIGURES Figure page 90H1-1 Materials, fabrication and a pplications of BioMEMS devices..........................................226H32 91H1-2 Structure of DNA and RNA...............................................................................................227H32 92H1-3 The central dogma of molecular biology...........................................................................228H33 93H1-4 Schematic diagram of the continuous exchange of cell-free system, adapted from Kim and Choi................................................................................................................... ..229H33 94H1-5 Schematic of in vitro red-shift GFP synthesis using emulsion droplets in microcrofluidic PDMS device by Dittrich et al.................................................................230H33 95H2-1 A picture of two devices form ing an array (3x4) of 12 wells............................................231H53 96H2-2 Vector maps. (a) pIVEX GFP vector map. (b) pIVEX CAT vector map. (c) Luciferase control DNA map. (d) phMGFP vector map...................................................232H53 97H2-3 A device with an array (12x8) of wells fo r in vitro expression of a group of proteins......233H54 98H2-4 In vitro protein expression in microcentrifuge tubes.........................................................234H55 99H2-5 PCR analysis by agarose gel electrophores is. Lane 1: PCR product after purification; Lane 2: DNA marker; Lane 3: raw PCR product..............................................................235H55 100H2-6 Excitation and emission spectra of the Monster PTMP GFP....................................................236H56 101H2-7 Comparison of in vitro GFP synthesis between at room temperature and at 30 oC...........237H56 102H2-8 Inhibitory effects of TC and CH on cel l-free protein expression in microcentrifuge tubes.......................................................................................................................... .........238H57 103H2-9 The response pattern of the 3x4 IVT se nsor array for two toxin simulants, tetracycline (TC, a) and cycloheximide (CH, b)................................................................239H58 104H3-1 Schematic representations of current formats for IVT system. (a) batch (a microcentrifuge tube); (b) CFCF system; (c) CECF system.............................................240H82 105H3-2 Illustrations of an array device I for pr otein synthesis. (a) Three dimensional view; (b) Picture of both parts; (c) Cross-sectional view of a tray nested in a well....................241H82 106H3-3 Comparison of in vitro GFP synthesis be tween a device and a microcentrifuge tube......242H83 107H3-4 (a) Comparison of fluorescence emission spectrum between GFP synthesized and GFP purchased. (b) The fluorescent signal as the function of the folding time................243H83

PAGE 11

11 108H3-5 Comparison of in vitro CAT synthesis between a device and a microcentrifuge tube......244H84 109H3-6 Production yield of luciferase as a function of the expression time in the device or in a microcentrifuge tube.......................................................................................................245H84 110H3-7 The effects of the fluid level di fference on the protein expression yield. .........................246H85 111H3-8 SEM images of polycarbonate membrane with 100 nm pore size. (a) room temperature. (b) 170 C......................................................................................................247H86 112H3-9 SEM images of polycarbonate membra ne with 10 nm pore size. (a) room temperature. (b) 170 C .....................................................................................................248H86 113H3-10 Production yield of luciferase as a f unction of time in the nanopore membrane devices, in the dialysis membrane or in a microcentrifuge tube........................................249H87 114H3-11 Signal amplification as a result of the i nhibitory effects of ricin on the production of every copy of protein.........................................................................................................250H87 115H3-12 The inhibitory effects of ricin on the production yield of luciferase.................................251H88 116H3-13 Synthesis of luciferase and dete ction of ricin in a few minutes.........................................252H88 117H3-14 Comparison among ricin A chain, B chain, whole ricin, and ricin treated with 2mercaptoethanol................................................................................................................ .253H89 118H4-1 (a) Schematic diagrams of two types of mi crofluidic devices; (b) Picture of a plastic microfluidic device for luciferase detection....................................................................254H105 119H4-2 (a) The picture of pre-assembly devi ce for thermal bonding. (b) A convection oven.....255H106 120H4-3 Photograph of experimental se tup for bioluminescent reaction......................................256H106 121H4-4 SEM pictures of a portion of the channe l with the herringbone structure. (a) The brass mold. (b) A representative PMMA substrate..........................................................257H107 122H4-5 Typical 3-D profiler results. (a) The brass mold. (b) A representative PMMA substrate...................................................................................................................... .....258H107 123H4-6 The picture of the in tegrated plastic device.....................................................................259H108 124H4-7 The images of air bubbles produced inside the channel. (a) Device I. (b) Device II......260H108 125H4-8 The images of a capillary with OD 360 m inserting into a hole (a) A mechanically drilled 1 mm-diameter hole. (a ) A laser drilled hole of 400 m......................................261H108 126H4-9 Mixing of phenolphthalein and NaOH at the flow rate of 1 L/min at the indicated locations. (a) Microfluidic device I and (b) Microfluidic device II.................................262H109

PAGE 12

12 127H4-10 CCD camera image of the luminescence at the indicated location in the microfluidic devices. (a) Microfluidic device I; (b) Microfuidic device II..........................................263H110 128H4 11 Normalized luminescence intensity cha nges along the channe l direction within microfluidic device I (solid circle) and micr ofluidic device II (ope n circle) with flow rate of 1 L/min...............................................................................................................264H111 129H4-12 CFD results of concentrati on distributions in the ridge re gion for different number of the mesh elements............................................................................................................265H112 130H4-13 The computational residue plot........................................................................................266H112 131H4-14 CFD results of concentration distributions and the interface betw een two solutions at the flow rate of 1 L/min.................................................................................................267H113 132H4-15 Calibration curves for bioluminescent a ssay using microfluidic device I (solid circles) and microfluidic device II (open circles)............................................................268H114 133H4-16 CCD image of the luminescence at the spiral channel at the flow rate of 1 L/min, 2 L/min, 4 L/min and 8 L/min from left to right..........................................................269H114 134H4-17 Effect of the flow rate on the lumine scent intensity along the channel direction in microfluidic device I (open) and microfluidic device II (solid)......................................270H115 135H4-18 Detection of synthesized luciferase using device II.........................................................271H115

PAGE 13

13 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DEVELOPMENT AND APPLICATIONS OF MINIATURIZED PROTEIN EXPRESSION SYSTEMS By Qian Mei December 2007 Chair: Z. Hugh Fan Major: Mechanical Engineering Recent development in miniaturization provi des an opportunity for performing chemical and biomedical research in a novel, fast and uniq ue way. This dissertation presents miniaturized plastic devices for in vitro protein expressions (synthesis) as well as their downstream applications. Biological synthesis of a protein includes th e steps of gene transcription and protein translation, which are typi cally carried out in host cells. However, some proteins are difficult to be synthesized in cells due to their insolubilit y, degradation and cytotoxicity. To address some of these challenges, cell-free protein synthesis, a process called in vitro transcription and translation (IVT), has been developed. In this dissertation, IVT is fi rst demonstrated in a microwell array for expression of proteins including green fluor escent protein (GFP), chloramphenicol acetyl-transferase (CAT) and luci ferase. The array device is then demonstrated for detecting toxin simulants, tetracycline (TC) and cycloheximide (CH), based on their inhibitory effects on protein synthesis. The diffe rential response patterns from the array device for two toxin simulants suggest the f easibility of the de tection concept. One drawback of the microwell array is that it does not consist of any fluid manipulation. To address this, a nested-well array is then developed. The nested-well device possesses a

PAGE 14

14 mechanism to supply nutrients continuously a nd remove byproducts, leading to higher protein expression yields. The productions of green fluorescent protein, chloramphenicol acetyltransferase and luciferase are dem onstrated in the device. In a ddition, the device is demonstrated for ricin detection. The calibration curve has b een obtained between the luciferase expression yield and the ricin concentration. Finally, a microfluidic device c onsisting of a straight channe l with herringbone ridges at the bottom wall and a spiral channel is designe d and fabricated to demonstrate enzymatic reactions. The device with the ridge feature has shown better mixing th an one with a smooth wall when luminescent enzymatic reaction is implemented in the device. Morever, the device can be used to detect lucife rse synthesized off chip.

PAGE 15

15 CHAPTER 1 INTRODUCTION In this chapter, miniaturization and the potenti al applications of miniaturized devices will be discussed first. It is follo wed by a brief introduction of basic principles and terminologies in molecular biology that are related to this interdisciplinary research. 1.1 Miniaturized Device and Its Applications Richard Feynman presented two famous talks about the miniaturized tools and devices: one is titled There is plenty of room at th e bottom an invitation to enter a new field of physics at the annual meeting of the American P hysical Society at the California Institute of Technology on December 29th, 1959;1 the other is Infinitesimal MachineryRevisiting there is plenty of room at the bottom at the Jet Pr opulsion Laboratory in California on February 23th, 1983.2 These talks invited scientists and engineers to explore the re alm of ultra-small structures and systems. Though Feynman never coined any fashion word, like nano or MEMS, he is now widely credited as the first visionary in th e field of miniaturized devices, and more recently, nanotechnology. Eleven months after his first talk, $1000 prize offered in Feynmans talk for the first guy who makes an operating electric motor in 1/ 64 inch cube was claimed by William McLellan.1 Since then, the engineering community has made exponential progress in miniaturization in microelectronics, micro-electro-mechanical syst ems and nanotechnologies. The miniaturized computers were partially realized wh en he gave his second talk in 1983. But he must be very disappointed by the slow progress in miniaturized biological systems, which he speculated in his talk: 1 (p. 62) The marvelous biological system. The biological example of writing information on a small scale has inspired me to think of some thing that should be possible. Biology is not simply writing information; it is doing something about it. A biological system can be exceedingly small. Many of the cells are very tiny, but they are very active; they

PAGE 16

16 manufacture various substances; they walk ar ound; they wiggle; and they do all kinds of marvelous things---all on a very small scale. Also, they store information. Consider the possibility that we too can make a thing very small which does what we want---that we can manufacture an object that maneuvers at that level! Chemists and biologists are beginning to unde rstand the potential of miniaturization and starting to work closely with engine ers to develop mini aturized devices.3-5 1.1.1 MEMS Micro-Electro-Mechanical Systems (MEMS) are the miniaturized systems or devices that integrate electrical and mechanical com ponents on a common silicon substrate through microfabrication technology.6-10 The sizes of these components are from the submicrometer level to the millimeter level; and there can be any number, from a few to millions, in a particular system to complete a specific function.11-16 Traditionally, MEMS devices are achieved by repeating the depos ition (thin film), patterning (photolithography) and etching steps on two major materials, silicon and glass.8, 17, 18 Besides the fabrication technologi es directly ported from the integrated circuit industry, new protocols have been developed in last three d ecades to manufacture mechanical elements, like gears, beams, springs, etc. in MEMS devices.19 During the last two decades, the MEMS devices sh ift rapidly from resear ch lab to industry. Nexus market analysis predicted the market for microsystems will grow from $12 billion in 2004 to $25 billion in 2009 at th e rate of 16% per year.20 Examples of applications of MEMS devices are read/write head of hard drive, inkjet printer cartridge, automotive accelerometers and pressure sensors. The topic of this dissertation is about Bi oMEMS, the most recent and dramatically expanding biological applications of MEMS.

PAGE 17

17 1.1.2 BioMEMS, Lab-on-a-chip and MicroTAS BioMEMS (biological or biomedical micro-electro-Mechanical system) gets its name from MEMS, but BioMEMS devices rarely use silicon or involve a movi ng part (like beams or gears), which are hallmarks of MEMS devices. It may be one of least precise terms to describe the application of MEMS and/or mi cro/nano technologies in biologi cal system or bioanalytical chemistry. BioMEMS is now much more than a subset or division of MEMS, it is becoming a field into itself, uniquely identified by its ma terials, fabrication technologies, chemistry and physics. While most researchers from engineering like to use BioMEMS to emphasize the microfabrication techniques used for manufacturi ng these miniaturized devices, chemists prefer to use Lab-on-a-chip (LOC) or micro total analysis system (MicroTAS or TAS) to indicate that more than one processing steps are integrated on to a single platform to perform a well-defined analysis task or a specific function.21 More and more researchers are now using these terms interchangeably. They sometimes also categorize devices without mechanical and/or electronic components, such as DNA or protei n microarrays, under BioMEMS. In the following part of this se ction, the material, fabrication, applications (Figure 1-1) and advantages of BioMEMS or LOC de vices will be briefly reviewed. BioMEMS devices are mainly fabricated with th ree categories of materials. They are (1) traditional MEMS materials, like silicon and gl ass; (2) plastic materials and polymers; (3) biomaterials, such as DNA, protein, cell and hydrogel. At the inception of BioMEMS, silicon wafers and glass slides were used extensively because the MEMS fabrication technologies, su ch as lithography, bulk micromachining, surface machining, could be directly port ed to fabrication such devices.17, 22, 23 However, the cost of

PAGE 18

18 such a device is very high. As reported by Manz et al. in 1998, the cost for each etched glass chip for continuous PCR amplificatio n was $500 and for the apparatus was $4000.24 In last couple of years, more and more resear ch groups started to use plastic materials and polymers. The main reasons of using plastic mate rials include: (1) low cost. The material itself and the compatible fabrication process are much less inexpensive than the silicon or glass and they can be used in disposable devices, reduci ng the possibility of crosscontamination of the reused devices;25-28 (2) multi-functionalities. There are di fferent plastic materials and polymers with different chemical, physical and biological properties. A ha rd and glassy or a soft and rubbery chip with different chemically functi onalized surface can be eas ily chosen from a doze of frequently used materials; (3) super bioc ompatibility. Its easie r to interface polymer materials to the biological tissues. The lowe r toxicity, less immuno-inflammatory and biodegradability permit in vitro applications of such devices The fabrication methods for manufacturing plastic devices include soft-lithography, mold injection, hot embossing, laser ablation and micromilling.29, 30 Most of these technologies still require a stamp, a mold, or a master fabricated with traditional MEMS technol ogies. Once the stamp is fabricated, the pattern on it can be mechanically transferred to hundreds to thousands of plastic chips. It is very costeffective. The miniaturized devices used in this dissertation were made from PMMA (poly methyl methacrylate) and polycabonate by direct micromilling or hot embossing. The third category is biomaterials, incl uding biomolecules (DNA, protein, etc.) and hydrogel. Hydrogel is widely used in scaffold fabrication for artifi cial organ and other implantable medical devices. Lots of well-de signed hydrogel may swell in responding to the environment, such as pH, temperature, etc., wh ich makes it as a perfect component in closedloop smart devices.31, 32 While the above mentioned two categories of materials are normally

PAGE 19

19 fabricated with top-down methods (from large sc ale to small scale, like photolithography), the bottom-up approaches (from atoms or molecules to a specific pattern), such as self assembly, are widely used for biomaterials. Nature is the best engineer.33 Combining the top-down and bottom-up approaches, much more complex systems can be manufactured. There are three categories of applications of BioMEMS devices: micro analysis, on chip synthesis and hybrid devices (Figure 1-1). The micro analysis devices are typically classi fied as either microarrays or microfluidic separation devices. A DNA or protei n microarray usually refers to a collection of miniaturized test sites, which are composed of DNA, protein, small molecules, probes, self-assembled or in situ synthesized on a solid substrate, permitting t housands or more tests to be performed at the same time. Developed in early 1990s, electrophore sis microchip is a typical microfluidic device for micro analysis.27, 34-50 It is still one of the most act ive research topics in LOC and/or BioMEMS, and lots of efforts have been devot ed to the integration and parallelization of microchip by different research groups. Burns et al. fabricated an integrated nanoliter DNA analysis device with microfluidic channels, heaters, temperature sensors and fluorescence detectors that could perform sample loading, mixing, PCR, electrophoresis and detection in a single chip.51 Mathies group pioneered in the de velopment of 16, 48 and 96 microchannel electrophoresis microchips for genotyping and DNA sequencing.50, 52, 53 A recent success of parallelization is the 768 lane DNA sequencing sy stem reported by Ehrlich et al. Within this system, two sets of 384 lanes could be altern atively cycled between electrophoresis and regeneration and the pr oductivity was 4M bases.54 While majority research reports are about the micro analysis system, considerable efforts have been devoted to the on chip synthesis or mi croreactors since later 1 990s. Take advantages

PAGE 20

20 from the fast mixing,55 improved mass and heat transfer from the microstructures in the microreactors,56-58 and on chip (bio)chemical synthesis studies59, 60 showed great success as predicted from the very beginni ng. One typical system is continuous-flow PCR microchip reported by Manz et al in 1998. Recent research focuses on the integration of multiple reaction steps, on line analysis feedback, the solutio n-based combinational synthesis which has the potential for oligonucleotide, and peptide library construction and gene assembly.61-70 There are full of challenges and opportunitie s to investigate the hybrid BioMEMS devices, such as implantable therapeutic devices and artif icial organs. An ideal implantable therapeutic device composed of a sensor for monitoring level of biomolecules and cont rolling an actuator for drug or other therapeutic reagents delivery. A model system is glucose biosensor and insulin delivery system to relieve diabetes patient s from repetitively injection of insulin.71 Although there are many successful cases in either in vivo glucose detection or in vivo insulin delivery system, but no integrated system has been repo rted by now. This category of BioMEMS devices is not within the scope of this dissertation. Fu rther information can be found in recent reviews.72, 73 After brief review of the mate rials, fabrication and applicat ions of BioMEMS devices, we can answer this question: why we need mini aturized BioMEMS devices ? Because there are several advantages of such devices, includi ng: (1) small dimension of the components will reduce the weight, size and energy consumption of the device, leadi ng to the high throughput, portable, and automatic complex system; (2) small amount of reagents may reduce the cost of the assay and small amount of analyte may save the precious sample consumption; (3) the performance (ultra-fast reaction rate, high resoluti on and sensitivity, etc.) of the device is greatly improved because of the fast mixing, heat exchange and mass transfer; (4) the reliability of the

PAGE 21

21 assay is improved because the batch fabrication pr ocesses make the low price, disposable device possible, and the confined miniat urized reactors almost eliminate the cross-contamination from different samples. These advantages and the po tential commercial interest pull the miniaturized BioMEMS devices from emerging to burgeoning in last 15 years. 1.2 Protein Expression In this section, the basic principles and termi nologies of molecule biology related to this interdisciplinary work will be introduced. 1.2.1 DNA, RNA, Protein and Central Dogma DNA or deoxyribonucleic acid, usually in the fo rm of double strand helix is the carrier of genetic information of all cellular forms of life and most viruses.74 Each monomer of DNA is made from simple units, called nucleotides, wh ich consist of a sugar (deoxyribose) with a phosphate group attached to it, and a base, whic h may be either adenine (A), guanine (G), cytosine (C), or thymine (T). The chemical bond between one phosphate to the next sugar form a backbone that links one nucleotide to the next The bases protruding from the backbone can pair with bases from another strand of DNA to form double strand helix, as shown in Figure 1-2. RNA or ribonucleic acid is slightly different from DNA. The sugar in the backbone is ribose and one of the four bases is different uracil (U) replaces t hymine (T). Unlike DNA, RNA is almost always a single-stranded molecule and has a much shorter chain of nucleotides. The hydroxyl group attached to the ribose is pron e to hydrolysis, which makes RNA less stable than DNA. But RNA always contains seconda ry structure to prompt the stability. Protein is a complex, high-molecular-mass and organic compound that is made from a long chain of amino acids, each linked to its nei ghbor through a peptide bond. Proteins constitute most of the dry mass of a cell a nd they are involved in nearly all biological processes of the

PAGE 22

22 living cell, including regulation of cellular functions such as signa l transduction and metabolism. Life, chemically speaking, is nothing but the functional activ ity of proteins. Figure 1-3 shows the central dogma of molecular biology fi rstly enunciated by Francis Crick in 1958 and re-stated in a Nature paper published in 1970.75 It summarized internal relations between DNA, RNA and protein and can be expressed in one sentence: DNA is transcribed to RNA, which is then translated to protein; but protein is never back-translated to RNA or DNA. The central dogma can be represented by th ree major steps: (1) DNA replication: it transmits the genetic information from the parent cell to each daughter cell. It is carried out by a complex group of proteins that unwind the doubl e stranded helix and using DNA polymerase and its associated proteins to faithfully replicate the original template; (2) transcription: the genetic information contained in DNA is transferred to mRNA. mRNA is synthe sized and catalyzed by the RNA polymerase and four ribonucleoside tr iphosphates using DNA as a template with a polymerase specific promoter sequence; (3) translation: after transcription, mRNA will be transported out of nucleus and bound to ribosom e. The sequence information of mRNA is translated to an amino acid chain (polypeptid e) by tRNAs (transfer RNAs) and associated enzymes according to the three letter codons (Tab le 1-1). After translation, the peptide chain may require additional process or directly folds to a mature protein with a specific conformation. Since Cricks paper published, there are a num ber of new facts have been discovered.76-79 For example, the retroviruses can transcribe RNA to DNA and some viruses only have RNA genome.76 But central dogma is still a very usef ul theory for guiding most experiments. 1.2.2 In Vivo and In Vitro (Cell Free) Protein Expression In vivo protein expression process refers to protein synthesis in cells, which is already discussed in above section. A wealth of genetic and biochemical knowledge has been

PAGE 23

23 accumulated both in prokaryotic cells (like E. coli ) and in eukaryotic cells (like insect cells and mammalian cells). Several companies provide commercial cloning and protein expression kit with improved protein expression level and simplifie d protocols. However, there are still some drawbacks, including tedious labo r in transfection, ce ll culturing and cell ly sing, instability of mRNA and expressed protein, improper folding, and the aggregation or degradation of the protein products in the cell. In addition, it is impossible to synthesize proteins which are unnatural or toxic for the host cells. In vitro or cell free protein synthesis is an attractive alternation of in vivo protein expression. As early as in 1950s, several groups independe ntly demonstrated protein biosynthesis machinery can work in absence of cell integrity. The limitation of the first generation cell-free expression system based on crude cell extract is the short lift time, and as a consequence, low productivity. The first cell-free protein synthesis system was prepared from extracts of rat liver cells.80 Later on, several E. coli extract systems were developed for translation.81, 82 In 1970s, cell-free systems for transl ation of mRNA were made using wheat germ extract 83and rabbit reticulocyte lysate.84 The final yields of these original systems are relative poor, though the produced proteins ha ve good biological activit ies. The limitation comes from two facts: one is the rapid deple tion of high energy phosphate pool; the other is the accumulation of the byproducts during the protein synthesis that inhibit the reaction. In 1988, Spirin et al overcame the bottleneck of the in vitro protein synthesis system by demonstrating a continuous flow system to continuously supply the consumable substrates and remove the byproducts.85 Following this work, several groups used different reagents exchange scheme to produce hundreds to thousands of microgram of proteins per milliliter of expression solution.85-92 After testing the production of various function proteins (enzyme, an tibody, hormone, etc) in

PAGE 24

24 more than 40 laboratories, Roche introduced a Rapid Translation System (RTS) in 2001.93 Theoretically, a cell-free protein synthesis coul d be prepared from any cell type. However, commercialized expression systems are based on E. coli wheat germ and rabbit reticulocyte. These systems will be summarized in chapter 2. Compared with in vivo protein expression system, the main advantages and potential application of cell free protei n synthesis systems include: 1 Absence of the cell membrane eliminates steps associated with introducing DNA into cells, lysing cells and clearing lysate. More purified protein can be easily obtained. 2 Being compatible with the miniaturized and automated instrument, which can fulfill the high-throughput protein expressi on requirement in protein engi neering and proteomic area. 3 The composition of the reaction mi xture can be easily modified to favor post translational modification and protein folding. 4 The non-natural and chemically modified ami no acids can be coopera ted into the protein, which provides the possibility of protein engineering, protein f unction research and pharmaceutical studies. 5 The cytotoxic proteins and unstable or low expressed protei ns can be obtained. 6 The procedure is extreme faster, making it hi ghly suited for screeni ng expression templates and conditions. 1.3 In Vitro Protein Expression on a Microchip Cell-free mRNA translation in a microfabricate d reactor made from a silicon wafer was first reported in 2000 by Hosokawa et al.94 Polyphenylalanine, coded by polyuridylic acid, is synthesized by pumping two reacta nts into two inlets and mixing them through a T-shaped or Hshaped microchannel. The resulting product was collected from the outle t using a filter paper and then analyzed off the device by radioisotope assay. This work demonstrated the feasibility to implement cell-free transc ription in a microfluidic device, though it showed only a polypeptide elongation step. Soon later, Nojima et al. showed the natural protein synthesis using the same device by cell-free transl ation of bacteriophage MS2 RNA.95 The key drawback of this

PAGE 25

25 device is the use of excessive ac cessories including external pump s and valves, and the lack of integration, making it difficult to be im plemented in a high-throughput format. Later on, an integrated PDMS-glass hybrid microreactor array was developed for in vitro protein synthesis by Yamamoto et al.96 The microreactor array consisted of a reaction chip and a temperature control chip. The reaction chip comp osed of reaction chambers, flow channels and inlet/outlet reservoirs, is fabri cated by casting PDMS replica on a silicon mold master, and then PDMS is bonded with a thin layer of PDMS to fo rm closed chambers and channels. While the temperature control sensors ar e patterned on a glass chip th rough conventional photolithography. The microreactor array was aligned and bonded by placing the temperature control chip on bottom of the reaction chip. In our study, the performance of the microreactor array for in vitro protein synthesis was demonstrat ed by using DNA templates of Gr een Fluorescent Protein (GFP) and Blue Fluorescent Protein (BFP) in E. coli extract. The functional activity of GFP and BFP collected in the outlet reservoirs was successful ly measured by the UV trans-illuminator. One limitation of this device is that the device was fabricated on a silicon wafer or a glass using traditional MEMS technology. Each device has to be fabricated individually, costing time and resources. In vitro protein synthesis has also been demonstrat ed in the microplate format. Tabuchi et al. first reported a PMMA microchip was de veloped for cell-free pr otein synthesis in 2002.97 The microchip was made by injection molding, w ith wells of 3 mm in diameter and 1 mm in depth. The synthesis of adipose-type fatty aci d binding protein (A-FABP) was investigated on this microchip and the product was detected usin g Western blotting. Two years later, Angenendt et al. accomplished protein synthesis in microfab ricated glass nanowell chip with submicroliter volumes, demonstrating enhanced throughput of protein production with reduced reagents.68

PAGE 26

26 Two Proteins were synthesized on a chip with the coding genes of wt-GFP and -galactosidase using E. coli extract. The biological activity of GFP and -galactosidase was confirmed by a highly sensitive laser scanner with an excitation of 488 nm and an emission of 535 nm. Meanwhile, a microchamber array was fabricated using PDMS for in vitro protein synthesis by Kinpara et al.98 The dimension of the chambers is about 100 (L) X 100(W) X 15 (D) m, and the volume is about 150 pL. Polymer beads with immobilized GFP gene coding sequence were added to microchamber and mixed with E. coli extract. The expressi on of GFP was detected by fluorescence using an optical microscope. In recent years, many research groups have attempted to perform in vitro protein synthesis on a microchip, but they only car ry it out using either E. coli extract or rabbit reticulocyte lysate on one chip. In our study, we developed e xpression of three proteins in both E. coli extract and rabbit reticulocyte lysate on a plastic microwells chip by direct micromilling, and used the response pattern of an array for id entification of two toxin simulants.99 Miniaturized array format of in vitro protein synthesis suggests that it is possible to have high-throughput protein synthesis in microsacle from a D NA library. It has advantage of less reaction solution and lowcost. However, cell-free protein synthesis on a mi crochip usually stopped within 2 h, due to lack of substrate or inhibitory by accumulated by prod ucts. Therefore, the productivity of protein synthesis is poor on such a microchip. 1.4 Literature Review and Current Challenges The challenge of BioMEMS is to integrate different components such as mciropumps and valves for mixing, purification and separation, and on-line detection on a single chip. In a miniaturized protein synthesis and analysis device, more challenges exist in performing fast, high-throughput and low cost synthesis, and sens itive assay. In addition, the poor productivity of in vitro protein synthesis limits the downstream app lications because large amount of proteins

PAGE 27

27 are required for traditional protei n analysis. In 1988, large scal e cell-free protein translation system using continuous flow mechanis m was first reported by Spirin et al..85 This system improved the production yield due to the extended duration of protein synt hesis. It has been demonstrated that the duration time for pr otein synthesis can be 20-40 hours in E. coli and wheat germ extract, resulting in a significant increase in the final amount of product. However, the setup is bulky because it require s a peristaltic pump or HPLC pu mp, and the system is very complicated. Later, simplified continuous exchange syst ems were reported by independent research groups using a dialysis bag or a flat dialysis membrane. For ex ample, Kim and Choi developed the continuous exchange system that employed a dialysis membrane for large scale cell-free protein synthesis.88 As shown in Figure 1-4, the reaction chamber was attached with a dialysis membrane at the bottom, and the chamber was i mmersed into a feeding solution containing all small molecular-weight substrate required for en ergy regeneration. Although the details of the membrane and the size of the chambers were no t given, authors reported the synthesis of CAT protein can continue for up to 14 hours in this reactor. Continuous-exchange cell-free (CECF) systems proved to be so attractive that CECF instrument have developed Roche Diagnostic.93 However, there are very limited numbers of repor ts about protein synthesis in miniaturized devices using flow mechanism. As mentioned above, Hosokawa et al. first fabricated a simple T-shape and H-shape glass-silicon mixers to test ce ll-free translation of poly (U), which is not a functional protein.94 Recently, a compartmentalized cell -free protein expression device was fabricated from PDMS to generate water-in -oil emulsion droplet as shown in Figure 1-5.65 In this device, cell-free translation/transcription reaction solutions and red-shift GFP (rs-GFP) DNA template were pumped into two inlets in the Yshape microchannel. At the cross-section of Y-

PAGE 28

28 shaped inlets, a side channel was used to provide a carrier liquid. Discre te liquid droplets were compartmentalized to mimic an artificial cell by mineral oil as a carrier liquid and in vitro protein synthesis was carried out inside the liquid droplets. In the microfluidic channel, the formation of emulsion droplets may prevent the re agents from contacting with the microchannel wall and decrease dispersion duri ng the transportation. However, the main problem for this experiment is that a heater is necessary to pr ovide a temperature control for each droplet to obtain the detectable amount of synthesized red shift-GFP. S econdly, the microchannels must maintain hydrophobic surface to ensure the formation of emulsion droplets. Rough channel surface could also induce the fusion of individual emulsion droplet. Recently, Hahn et al. published a short report and investigated a continuous-exchange cellfree protein synthesis syst em on a PDMS-based chip.100 Expression of bacterial chloramphenicol acetyltransferase (CAT) was demo strated in this chip and the duration time for reaction is up to 16 h. But no detail informati on was further discussed, such as the size of reaction chamber, the effects of te mperature and flow mechanism. In our study, we considered easy and simplicity of device fabricati on. Plastics are cheap and the devices can be disposable. Hence we chos e PMMA and PC for this research work. Two different types of microfluidic devices were designed for in vitro protein synthesis. One is designed in a well-in-well arra y format employing a porous memb rane to facilitate passive diffusion. The other is designed to contain three-dimension ridge structures in the bottom of microchannels to enhance mixing and reaction yield. 1.5 Objective and Organization of This Dissertation Detection and identification of toxic agen ts are important for medical diagnostics, food/water safety testing, and biological warfare defense. One way to detect them is to use recombinant technology to engineer cytoplasmic or cell surface receptors for regulated or

PAGE 29

29 specific changes in an organism after binding with an agent. An example is ricin, a plant toxin that is listed as one of poten tial warfare agents and known to be lethal by blocking protein synthesis in many cells of the human body. The objective of this work is to fabricate a miniaturized high-yi eld protein expression system for detecting toxins. The detection met hod we proposed to explor e is toxins inhibition of protein expression (synthesis), which is one of action mechanisms toxin use to cause toxin effects. Biological synthesis of proteins typically take places in cells but it is challenging to use it directly as a sensor due to its slow process and maintenance requirement such as continuous supply of nutrients. To address them, we develo ped miniaturized devices that are composed of an array of units; each unit consists of a reac tion chamber for protein expression and a feeding chamber for nutrient supply and byproduct removal. Different in vitro protein expression systems are evaluated in the device. This device has also demonstrated for screening anti-biotic inhibitors and ultra-fa st ricin detection. The significance of this work mainly lies in th ree aspects. First, the approach will provide a unique method to detect both known and new toxins With the increasing ability to modify and engineer potential toxins, the ability to detect t oxic agents that have not been identified or fingerprinted has become more important. To our knowledge, this new sensing principle using protein inhibition has not been studied. Second, the multiplexed IVT system offers a means for high-throughput protein expression that is needed by life sciences. As more and more new genes are being identified, there is a considerable need to determine the function and properties of the proteins these genes encode. Third, this study will also advance the field of microfluidics, bring forward sensor technology, and assist in unde rstanding nanoliter-scale cellular reactions. The rest of this dissertation is outlined as follows.

PAGE 30

30 In chapter 2, different in vitro protein expression systems are ev aluated in the miniaturized device consisting of an array of reaction chambers. As a proof of the concept for toxin detection, the expression yields of different proteins are measured in pr esence of antibiotic inhibitors. In chapter 3, the performances of protein synthesis in the static and the continuous exchange reaction chamber are compared. Ultra fast and sensitive detection of a toxin, ricin, is employed as an example of the application of th e continuous exchange pr otein synthesis device. In chapter 4, microfluidic r eactors are designed and fabricat ed to demonstrate enzymatic reactions. Bioluminescence produced by reaction of lucifera se-luciferin was de tected by a cooled CCD camera. Diffusion-based mixing and chaotic i nduced mixing in the microfluidic device was investigated and compared. The possibility of us ing such a microfluidic device for continuous flow protein expression will be discussed. In chapter 5, the conclusion and future work will be described.

PAGE 31

31 Table 1-1. RNA codon table Second Position Firs t Pos itio n U C A G Thir d Pors ition U phenylalanin e tyrosine cysteine C stop stop A U leucine serine stop tryptopha n G U histidine C A C leucine proline glutamin e arginine G U asparagin e serine C isoleucine A A methionine threonin e lysine arginine G U Aspartic acid C A G valine alanine Glutanic acid glycine G Start codon.

PAGE 32

32 Figure 1-1. Materials, fa brication and applications of BioMEMS devices. Figure 1-2. Structure of DNA and RNA.

PAGE 33

33 Figure 1-3. The central dog ma of molecular biology. Figure 1-4. Schematic diagram of the continuou s exchange of cell-free system, adapted from Kim and Choi.88 Figure 1-5. Schematic of in vitro red-shift GFP synthesis using emulsion droplets in microcrofluidic PDMS device by Dittrich et al.65

PAGE 34

34 CHAPTER 2 IN VITRO PROTEIN EXPRESSION IN A MI CROWELL DEVICE A ND ITS USE FOR TOXIN DETECTION0F1 2.1 Introduction Biological synthesis of a protein includes th e steps of gene transcription and protein translation, which are ty pically carried out in host cells such as Escherichia coli. Although it is prevalent among academic and industrial labs, E. coli -based recombinant protein production has inherent limitations including formation of inso luble protein aggregates (inclusion bodies), degradation of proteins by intracellular prot eases, low or no expression for the genes whose products are toxic to the host cell, and lack of post-translational modification.101-103 To address some of these challenges, a few eukaryot ic systems have been developed using Saccharomyces cerevisiae insect or mammalian cells. In addition, cel l-free protein synthesis, a process called in vitro transcription and translati on (IVT), has been developed.104-112 In IVT systems, a DNA template consisting of a coding se quence is transcribed into messenge r RNA, either eukaryotic or prokaryotic lysate is then expl oited to provide ribosome and a dditional components necessary for protein translation. The expression system may al so be reconstituted from purified recombinant elements.113, 114 IVT has been demonstrated for various applications, including in situ immobilization of expressed proteins onto solid surfaces,66, 115 synthesis of drug transporters 116, polypeptide display,117 gene expression,118 and high-throughput screening.119 High-throughput protein producti on in an array format is desirable for genomics and proteomics. Completion of mapping the human genome has prompted strong interest in identifying the functions of newly discovered gene s and the proteins encoded therein. To match high-throughput gene discovery, methods to produ ce a large number of pr oteins in parallel are 1 TA part of this chapter has been published in (99) Mei, Q.; Fredrickson, C. K.; Jin, S.; Fan, Z. H. Anal Chem 2005 77 5494-5500.

PAGE 35

35 needed. In vitro protein expression has al so been implemented in miniaturized devices68, 95, 97, 98, 118, 120, 121 because miniaturization of an alytical instruments has been one of major developments in the past decades. In this chapter, we describe a device consis ting of an array of miniaturized wells. The device was used for in vitro expression of three proteins, in cluding green fluor escent protein (GFP), chloramphenicol acetyl-transferase (CAT), and luciferase. Two expression systems were used: a prokaryotic and a eukary otic system. Differential inhibitory effects of two toxin simulants, tetracycline (TC) and cycloheximide (CH), on the protein expression yield were observed, providing a unique response pattern of th e array device for each toxin. In addition, the quantitative relationship between the yield of expressed protei ns and the amount of a toxin provide a calibration curves, leading to true analysis. 2.2 Background In vitro protein synthesis systems are based on th e cellular protein synthesis machinery to perform protein synthesis outside intact cells, and are divided in to two types: prokaryotic and eukaryotic expression system. In vitro protein synthesis systems can be carried out either in a coupled manner, which requires an exogenous DNA as a template, or as an uncoupled system, in which messenger RNA (mRNA) is used as a template. Here we will focus on the coupled transcription-translation system (IVT). 2.2.1 Prokaryotic IVT System The most commonly used prokaryo tic system is derived from Escherichia coli lysate, including all the necessary enzyme s and machinery for translation.122 After two groups (Zubay; Gold and Scheiger) independently achieved some major improvements,106-112 IVT has been widely used. The principal components of Zuba y and Gold-Schweiger system are listed in Table 2-1. The major difference between them is Zubays system contains 30,000 g centrifugation

PAGE 36

36 supernatant (S30) E. coli lysate depleted of endogenous R NA and DNA, while Gold-Schweiger system is based on the ribosome-free and nucleic acid deprived of S100 fraction. In both systems, they contain an endogenous or phage T7 RNA polymerase, which transcribes mRNA from an exogenous DNA template. This RNA molecule is then used as template for translation. The DNA template may be either a gene cloned into a plasmid vector or a PCR amplified template. A ribosome binding site (RBS) is requ ired for template translated in prokaryotic systems. During transcription, the 5 end of the mRNA becomes available for ribosome bind and translation initiation, allowing transcription and translati on to occur simultaneously. 2.2.2 Eukaryotic IVT System Eukaryotic IVT systems, such as rabbit reti culocyte lysate and wheat germ extract, are prepared from the cytoplasmic fraction. Ho wever, they lack endogenous RNA polymerase activity because the process of transcription only takes place in the nuclei of the cell, while protein synthesis is carried out in the cytopl asm of the cell, which is the major difference between the prokaryotic and eukaryotic cell. Th is limitation can be addressed by addition of an exogenous RNA polymerase. Purified E. coli RNA polymerase was first used to couple transcription and translation in eukaryotic IVT systems.83, 123-126 Bacteriophage T7, T3 and SP6 RNA polymerases were then isolated to use in IVT systems and they were introduced into eukaryotic system to produce mRNA, using DNA template with corresponding promoters92, 127, 128 (Table 2-2). 2.2.3 Choice of IVT Systems Although any organism could potentially be used as a source to prepare an in vitro protein synthesis system, E. coli lysate, rabbit reticulocyte lysate a nd wheat germ extract are widely used and commercially available.64, 84, 93, 129-131 The choice of a sy nthesis system is based on the origin and biochemical nature of synthesized proteins and their downstream applications. Generally,

PAGE 37

37 prokaryotic IVT systems can provi de higher yields and are suitab le for structural analysis. However, eukaryotic IVT systems can allow a bette r platform for functiona l studies, especially for post-translational proteins. The comparison of prokaryotic and eukaryotic systems is summarized in Table 2-3. 2.3 Experimental 2.3.1 Reagents and Materials The RTS 100 E. coli HY kit, two expression vectors containing the genes encoding green fluorescent protein (GFP) and chloramphenicol acet yl-transferase (CAT), and anti-6xHis were obtained form Roche Diagnostics GmbH (M annheim, Germany). TNT Quick Coupled Transcription/Translation system T7 luciferase DNA vector, luci ferase assay reagent, phMGFP vector, PCR Master mix, PCR clean-up kit, DNA marker and nuclease-free water were from Promega Corporation (Madison, WI). RNA tran scription kit was bought from Stratagene (La Jolla, CA) and RNA quantitation kit was from Molecular Probes (Eugene, OR). Primers were synthesized from Integrated DNA technologies (Coralville, IA). Acrylamide-bisacrylamide (electrophoretic grade, 5% C) tetramethylethylenediamine, sodium dodecyl sulphate (SDS), ammonium persulfate, tris(hydroxymethyl)aminomet hane (Tris), glycine, sodium chloride, glycerol, bromophenol blue, -mercaptoethanol, Tween-20, tetr acycline, and cycloheximide were purchased from Fisher Scientific (A tlanta, GA). Polyvinylidene difluoride (PVDF) membranes (0.2 m), and filter papers were from Bi o-Rad Laboratories (Hercule, CA). Molecular weight standards, biotinylated secondary antibody and streptavidin-alkaline phosphatase were from Amersham Biosciences (Piscataway, NJ), while recombinant Green Fluorescent Protein (rGFP) and ra bbit anti-GFP polyclonal antibody were from BD Biosciences (Palo Alto, CA). The phosphatase staining solu tion (Bromo-chloro-indoryl phosphate/Nitro Blue Tetrazolium, BCIP/NBT) was obtaine d from KPL (Gaithersburg, MD).

PAGE 38

38 2.3.2 Device Fabrication Polymethyl methacrylate (PMMA) is widely used in bioMEMS device fabrication. Its excellent optical properties, ve ry low fluorescence background, a nd compatibility with chemical or biological reagents make it as the first choi ce of the materials for microfluidic device. Moreover, the fabrication steps are straightforw ard and cheap. A miniaturized device with an array of 2x3 wells was designed a nd fabricated for demonstrating the toxin detection concept. Two of such devices in Figure 2-1 form an array of 3x4 wells, which is the format for demonstrating toxin detection as discussed belo w. The device was made from PMMA (Lucite International, Cordova, TN) and the wells were created by a milling machine (Flashcut CNC, Menlo Park, CA). The distances between wells (center to center) are 9 mm according to the standards for microplates defined by the Societ y for Biomolecular Screen ing and accepted by the American National Standards Institute. This a rrangement conforms the device to the alignment of 96-well plates and ensures compatibility wi th a variety of commer cial fluid dispensing systems and plate readers. The diameter and depth of each well are 2.7 mm and 2.3 mm, respectively, providing the total we ll volume of ~13 L. This is about 25 times smaller than the wells in conventional 96-well microplates. Th e decrease in the well size will significantly reduce reagent consumption for high-throughput assays. After fabrication, the device was sterilized by exposing to UV light for 30 minutes that ensured the consistency of the protein expression. 2.3.3 DNA Template for Cell-F ree Protein Synthesis GFP vector (Figure 2-2a), CAT vector (Figur e 2-2b) and luciferase vector (Figure 2-2c) were used as the templates for the plasmid-based cell-free protein synthesis. For the experiments using linear DNA template produce d by PCR for cell-free protein sy nthesis, the coding region of mutant GFP was amplified from the plasmid phMGFP (Figure 2-2d) using the forward and

PAGE 39

39 backward primers given in Table 2-4. To generate linear DNA template, PCR Master Mix containing Taq DNA polymerase, was reconstitute d per the manufacturers recommendations. A PCR tube with 50 l of the prepared soluti on was placed in a PTC-100 programmable thermal controller (MJ Research, Waltham, MA). PCR amplification consisted of 25 cycles, in which denaturation (94 C) and annealing (57 C) steps we re 45 s, and the extension step (72 C) was 1 min. The PCR products were purified using PCRclean up kit (as manual instruction) prior to their use for protein expression. The amount of PCR products was determined by agarose gel electrophoresis and spectrophotometry. To prepare ge l, agarose was dissolved in tris-boric acidEDTA (TBE) buffer solution and cast into a gel slab. When solidified, it was immersed in TBE buffer containing ethidium bromide dye. The PCR products were added to the sample wells in the slab, which was placed in the separation cell and voltage was then applied. When separation was completed, the gel was removed from the cel l, de-stained and rinsed to remove excess ethidium bromide, and imaged using a UV transilluminator. 2.3.4 Cell-Free Protein Synthesis For the prokaryotic expression system, 50 L of RTS 100 reaction solution was composed of 12 L E. coli lysate, 10 L reaction mix (proprietary comp osition, supplied in the kit by the manufacturer), 12 L amino acids without methionine, 1 L methionine, 5 L reconstitution buffer (proprietary composition, supplied in the kit by the manufacturer), and 10 L nucleasefree water containing 1 g GFP or CAT vector. The reaction solution was then incubated in a microcentrifuge tube at 30 C for 4 hours. For G FP, the reaction solution was stored at 4 C for additional 24 hours for the maturation of GFP. For the eukaryotic expression system, rabbit re ticulocyte lysate or wheat germ extract was used. The reaction mix of 50 L for luciferase synthesis wa s prepared by combining 40 L TNT Quick master mix (proprietary composition, supplied in the kit by the manufacturer), 1 L

PAGE 40

40 methionine, and 9 L nuclease-free water containing 1 g luciferase vector. Wheat germ reaction mixture for monster GFP synthesis consisted of the following components: 25 L wheat germ extract, 2 L reaction buffer, 1 L T7 RNA polymerase, 1 L of 1 mM amino acid, 1 L Ribonulease inhibitor and 19 L nuclease-free water containing 2 g of linear MonsterPTMP GFP template or 1 g of phMGFP vector. Incubation was performed in a microcentrifuge tube at 30 C for 1.5 hours. In certain experiments, th e transcription and tran slation reactions were implemented sequentially. The plasmid temp late or PCR-product was incubated in a transcription mixture and then supplemented with wheat germ extract to initiate the translation reaction. 2.3.5 Toxin Inhibition Assay For the toxin inhibition assay in a microcentrifuge tube, a stoc k solution of tetracycline and that of cycloheximide were prepared at 15 g/ L and 10 g/ L, respectively. A series of amounts of tetracycline or cycloheximide were added into protein expression mixture. The concentrations of toxins used ar e listed in the figures or text. To save reagents and match with miniaturized devices, 8 L of prokaryotic or eukaryotic expression solution was used, making the total volume of each inhibition assay at 10 L. For each set of experiments, a positive control (without inhibitor) and a ne gative control (without the expres sion vector) have also been included. When the protein expression and toxin inhi bition assays were implemented in the miniaturized device, the volume of th e reaction mixture was reduced to 6.5 L. Films with pressure-sensitive adhesive, so-called PCR tape (3M, Minneapolis, MN), were used to seal the wells to prevent evaporation.

PAGE 41

41 2.3.6 Detection We used Western blot, fluorescence or lumine scence for measuring the yield of protein expression, depending on the property of the prot eins expressed. Detection of expressed WTGFP and CAT was achieved using Western blot. A reaction product solution of 1 L was mixed with 15 L of gel-loading sample buffer, wh ich contains 50 mM Tris-HCl, pH 6.8, 2% W/V SDS, 0.01% W/V bromophenol blue 10% V/V glycerol, and 5% V/V 2-mercaptoethanol. The mixture was then separated in a 15% SDS-poly acrylamide gel in the Mini-Protean III Cell system (BioRad). After electrophoresis, the gel was removed from the glass plates and then equilibrated in the transfer bu ffer, which is comprised of 48 mM Tris, 39 mM glycine, and 20% V/V methanol. PVDF membrane was pre-soaked with methanol, followed by soaking in the transfer buffer for 1 hour. The Mini Trans-Blot system (BioRad) was set up with pre-wetted fiber pad, filter paper, gel, and PVDF memb rane according to the instruction from the manufacturer. The cassette and ic e-cooling unit were placed in th e tank that was filled with the transfer buffer. After blotting, the PVDF membrane was re moved from the trans-blot apparatus and blocked with 5% w/v non-fat dried milk in Tr is-buffered saline (TBS) solution (with 0.05% Tween-20) for 1 hour at room temperature. Af ter being washed three times (5 minutes each time) with TBS solution, the membrane was incu bated for 1 hour at room temperature with 1 g/mL anti-GFP polyclonal antibody for GFP product or 0.3 g/mL anti-6xHisB Bmonoclonal antibody for CAT product, respectively. At the end of conjugation, the membrane was washed three times and then incubated with 1.5 g/mL biotinylated secondary antibody at room temperature for 1 hour. Upon comp letion of the incubation, the memb rane was rinsed again with TBS solution for three times, followed by incubati on at ambient temperature for 30 minutes with a solution of streptavidin-alka line phosphatase (1:2000 dilution from the stoc k solution). After

PAGE 42

42 being washed, the membrane was immersed in chromogenic substrate (BCIP/NBT) for 3 minutes, followed by rinsing with water (to st op reaction). Images of protein bands were acquired with a color laser scanner (Canon); prot eins bands were quantified using ImageJ from the National Institute of Health.132 Detection of synthesi zed WT-GFP and MonsterPTMP GFP was also carried out by Spex Fluorolog spectrofluorometer (Jobin Yvon Inc, Ed ison, NJ). The excitation spectrum of WTGFP was scanned from 350 nm to 500 nm w ith emission wavelength at 550 nm; and the emission spectrum was scanned from 450 nm to 580 nm with excitation wavelength at 470 nm. The emission spectrum of MonsterPTMP GFP was collected from 450 nm to 580 nm with excitation wavelength at 480 nm. Detection of luciferase expressed by IVT was ach ieved by SIRIUS luminometer from Berthold (Pforzheim, Germany). The lumino meter was programmed to have a two-second delay, followed by a five-second measurement of lu ciferase activity. The expression product of 2 L was added to a luminometer tube containing 40 L of luciferase assa y reagent and mixed evenly. The tube was then placed in the luminometer; and the data were acquired. 2.4 Results and Discussion 2.4.1 Toxin Detection Scheme Multiplexed sensor array is a unique approach to obtain the fingerprint of a new agent. The concept of the sensor array for detecting toxins us ing IVT is illustrated in Figure 2-3. The device consists of an array of IVT wells; each well is designed to express one pr otein and thus functions as a sensor. The number of IVT sensors can be as high as 96 or its integral multiples, in the format of traditional microplates and can be easily adapted to commercial plate readers. Multiple wells (e.g., 3x4 wells circumscribed w ithin the dashed lines and shadowed with diagonal lines) form one set, in which the top ro w is for the positive controls to express each of

PAGE 43

43 three proteins, the second row for the negative c ontrols, and the third and fourth rows are for the sample, allowing one repeat to e nhance the precision. Use of the positive and negative controls and comparison of the signal from the sample wells with those in the c ontrol wells will reduce false positives and negatives. The set will express a group (three in this case) of precharacterized proteins in different expression sy stems; the proteins and expression systems will be judiciously selected so that protein synthesis in each well is inhibited or affected differentially by different type of toxi ns. Therefore the unique response patte rn (or signature) of a toxin due to different inhibitory effects will be registered a nd used as a tool for detection and identification. New agents will be identified by comparing the re sponse pattern with signatures of known agents in a pre-acquired database. In the particular form at illustrated, the rest of wells in the 96-well array can be designed to detect additional seven types of toxins if 12-well set is proved to be enough for the identification. 2.4.2 Plasmid-based Protein Expression To demonstrate the concept of toxin detection by protein e xpression, we first synthesized three proteins in two types of expression syst ems. The first protein is wide type green fluorescent protein (WT-GFP), a widely-used fluorescent molecule with known DNA sequence and crystal structure.133 Protein expression was carried out by using an expression vector as a DNA template, which consists of GFP coding se quence and the necessary regulatory elements including T7-RNA polymerase prom oter, ribosome binding site start codon, stop codon, and T7 terminator. The expression vector was mixed with E. coli lysate and a reaction mix consisting of T7-RNA polymerase, nucleotides, amino aci ds, and other reagents. WT-GFP product was confirmed by Western blotting and fluorescence spect rometry. The result of Western blotting is shown in Figure 2-4a. A clear band in lane 4 indicates the presence of WT-GFP in the expression product. According to pre-stained protein markers, th e molecular weight of WT-GFP

PAGE 44

44 expressed is estimated ~31 KD. Expressed WT-GFP contains a stretch of additional six histidines (6xHis) at its C-terminal, causing its mo lecular weight slightly larger than recombinant GFP (rGFP) purchased commercially. The negati ve control in the expe riment contains all reagents except for the expression vector. The WT-GFP expression also can be quantified by fluorescence spectrometry. When it was excited with 488 nm light, the excitation and emission spectra of synthesized WT-GFP were shown in Figure 2-4b, indicating the synthesized WT-GFP folds properly and has biological activity. The second protein is chloramphenicol acetyltransferase (CAT), an enzyme responsible for bacterial resistance to an antibiotic drug, ch loramphenicol. CAT was expressed in the same E. coli expression system; success of the protein e xpression was also confirmed using Western blot as shown in Figure 2-4c. According to prestained protein markers, the molecular weight of CAT expressed is estimated ~26 KD, which agrees w ith the value reported in the literature. The third protein is luciferase, an enzyme from fire fly tails that catalyzes the production of light in the presence of luciferin adenosine tr iphosphate (ATP), molecular oxygen and MgP2+P. Synthesis of luciferase was carried out using rabbit reticu locyte expression system as described in the experimental section. Detection of the e xpression product was achieved by monitoring the intensity of luminescence after mixing 2 L of the product with luciferin ATP, OB2B and MgP2+P. As shown in Figure 2-4d, the luminescence signal of the product is 5 orders of magnitude higher than that of the negative control. 2.4.3 Linear DNA Template-based Protein Expression The WT-GFP gene originally cloned from Montastrea cavernosa can be expressed into the protein which tends to photobleach and is very dim. In addition, the correct folding of WT-GFP requires three distinct physical processes: the attainment of the correct three-dimensional structure, cyclization of the chromophore, and oxidation of the cyclized intermediate.134 So the

PAGE 45

45 folding of WT-GFP can not complete during transc ription/translation in a cell-free system and it takes addition time to have biological activit y. The mutants of GFP which have brighter fluorescence and fold more efficiently at 37 C were developed, such as MonsterPTMP GFP, enhanced GFP, blue fluorescent protein and yellow fluorescent protein. MonsterPTMP GFP can produce fluoresces at least 20% br ighter than other commercially available GFPs and the spectral properties of the MonsterPTMP GFP are also slightly red-shifted. Peak excitation occurs at 505 nm, with a shoulder at 480 nm, and the peak emission occurs at 515 nm. MonsterPTMP GFP coding sequence was first amplif ied from phMGFP vector by PCR with corresponding primer. PCR products were purifie d by a PCR clean-up system and confirmed by the agarose gel electrophoresis (Figure 2-5). The results indi cated that the length of PCR products is about 1190 bp, as same as we designed. The synthesized MonsterPTMP GFP from PCRamplified DNA or the vector using coupled transc ription and translation synthesis was confirmed by fluorescent spectrometry (Figur e 2-6). The excitation and emi ssion spectra of the synthesized MonsterPTMP GFP are consistent with those reported by Pr omega. The results also showed that the protein yield expressed from plas mid is higher than that from PCR amplified DNA, because the linear DNA are easily degrad ed by the endogenous nucleases thus, the amount of mRNA molecules is limited. Unfortuna tely, the expression of MonsterPTMP GFP can not be reproduced. Then PCR products were first transcripted to mRNA and then mRNA molecules were translated to protein but protein can not be detected by fl uorescent spectrometry. The reasons may include: (1) the amount of expressed MonsterPTMP GFP is too low to be de tected using fluorescent spectrometry; (2) no MonsterPTMP GFP was expressed due to rapi d degradation of mRNA, thus terminates the translational reaction quickly.

PAGE 46

46 2.4.4 Reaction Temperature All of the protein expression experiments di scussed above take place at the ambient temperature. For commercial RTS instrume nts from Roche, the recommended operation temperature is 30 oC. As a result, we studied the effect of the reaction temperature on protein synthesis in microcentrifuge tubes. Temperatur e control is achieved by placing the tubes in a PCR machine. Figure 2-7a show s the Western blot image of GFP synthesized at room temperature and at 30 oC. The GFP expression yield at room temperature is about 97% of that at 30 oC, as shown in the quantitative illustration in Figure 2-7b. This resu lt suggests that the GFP expression at room temperature is comparable to that at 30 oC and it is not necessary to integrate a heater in the array device. 2.4.5 Inhibitory Effects of T oxins on Protein Synthesis To illustrate the detection of toxins, we used tetracycline (TC) and cycloheximide (CH) as toxin simulants to study their i nhibitory effects on protein ex pression. TC is an antibiotic substance produced by Streptomyces species.135 It acts only on prokar yotic cells and it blocks binding of aminoaceyl-transfer RNA to A-site of ribosomes.136 CH acts specifically on eukaryotic cells and it inhibits the activity of peptidyl transferase, an enzyme needed in the translocation reaction on ribosomes. Figure 2-8 shows the effects of a series of concentrations of TC or CH on the expression yields of GFP, CA T, and luciferase synthesized in two protein expression systems. As illustrated in Figure 2-8a GFP synthesis was completely inhibited when 3000 ng/L of TC was used. Partial inhibi tion was observed when a series of lower concentrations (300 ng/L to 0.3 ng/L) of TC was added. The experiment also included the positive control, in which no inhibitor (TC) was added. The negative control contained no expression vector, thus repres enting the background signal. These results suggest that a

PAGE 47

47 qualitative and quantitative relationship exists between the expression yield and toxin amount, a critical figure of merit for a sensor. The comparison between TC and CH for GFP production in the E. coli expression system is shown in Figure 2-8b and 2-8c. The expression yield for each protein was normalized against the expression yield of the positiv e control (without toxin), so that it is easier to compare the toxic effects. The results clearly indicate that TC has inhibitory effect on GFP production and the degree of inhibition is propor tional to the amount of TC in the sample, whereas CH has a negligible effect on the yield of GFP production and the level of minor inhibition remained the same in the range of amount of TC we used. Thes e results suggest that th e differential inhibitory effects of a toxin on the expressi on of different proteins are possi bly used for toxin detection. Figure 2-8d and 2-8e exhibit similar disparity between TC and CH for CAT production in the E. coli expression system. Again, TC has inhi bitory effect on CAT production and the degree of inhibition is proportional to the amoun t of TC in the sample, whereas CH has a negligible effect on the yield of CAT producti on and the level of minor inhibition remained essentially same in the range of amount of TC we used. Furthermore, comparison of Figure 2-8b and 2-8d indicates that although TC has inhib itory effect on both GF P and CAT production, the degree of inhibition per unit amount of TC differs between these two proteins, evident from the difference in the slopes of respective linear regr ession lines. These results further suggest that each toxins differential inhibitory effects on expr ession of different protein can be used as a signature for toxin detect ion and identification. The comparison between TC and CH for luci ferase production in rabbit reticulocyte expression system is shown in Figure 2-8f and 2-8g. An opposite effect was observed; TC has a negligible effect on the luciferase production in the eukaryotic expression system, whereas CH

PAGE 48

48 has a significant inhibitory eff ect and the degree of inhibition is proportional to the amount of CH present in the sample. The result indicates that different expression systems can be used to expand the variability, so that a unique respons e pattern can be obtained for toxin detection by using a set of proteins produced in different expression systems. These results are significant because they indi cate not only the feasibi lity of the concept of toxin detection presented in this work, but al so the possibility of using IVT assay for highthroughput screening of drug candidates. We c onfirmed differential i nhibitory effects of antibiotic inhibitors such as TC and CH on protein expression in vitro in a way very similar to their effects on protein expression in vivo .137 Therefore, an IVT array device may provide a great platform for searching for the best antibiotic drug candidates. 2.4.6 Miniaturized IVT Array After the feasibility of toxin detection us ing protein expression is demonstrated in microcentrifuge tubes, we attempted IVT and toxin detection in a miniaturized well device. The design of the experiments was the same as in th e shadowed area of the 96-well array in Figure 23, in which a set of 3x4 wells is assigned for det ecting one toxin. Two of the 2x3 well devices in Figure 2-1 were combined to form a 3x4 well device. As illustrated in Figure 2-3, the first row of 3 wells was used as the positive control, expressing GFP and CAT vector in the E. coli expression system and luciferase vector in the ra bbit reticulocyte expression system. These wells were free of toxins. The second row of 3 wells was used as the negative control without DNA vectors added. The third and forth rows of 3 we lls were added with a certain amount of a toxin stimulant, either CH or TC, into the protein expression system. Figure 2-9a shows the response pattern of the IVT sensor array when 25 ng of TC was used whereas the response pattern of the same IVT array for 17 ng of CH is illustrated in Figure 2-9b. Although there is slight difference between two sample repeats for rows 3 and 4 fo r each toxin, the response pattern is reproducible

PAGE 49

49 as expected. The significant difference in the response patterns between CH and TC clearly indicates that it is feasible to use IVT sens or array to detect and identify toxins. Although we used Western blot and lumines cence detection to monitor protein production in this concept demonstration, it should be feas ible to use a common met hod for toxin detection. One of such methods is to use GFP as an indicat or for the detection of protein expression due to its green fluorescence. GFP has been used for visualization, tracking, an d quantification of a variety of proteins in cells after they are fused together.133 An increase of fluorescence signal in an IVT well will indicate the production of GFP or GFP-fused proteins. Quantitative information may be obtained by comparing the fluorescence signals of sample wells and of reference wells, which include both positive and ne gative controls in the array device. Any variation or adverse effects wi ll be cancelled out between co ntrol and sample wells. The magnitude of the signal can be correlated to th e amount of proteins prod uced in the device. Indeed, we used a fluorescence spectrometer to confirm the production of GFP expressed as mentioned above. However, we did not use it fo r CAT and luciferase because we have not yet expressed them by fusion with GFP. Rather than using GFP, we may also desi gn the expression vector containing a coding sequence for expressing an additiona l stretch of six histidines (6xHis) at the C-terminal of the protein of interest. Many proteins produced by r ecombinant techniques are designed to contain a 6xHis tag, so that they can be purified thr ough interactions between 6xHis tags and Ninitrilotriacetate chromatographic columns.138, 139 Both GFP and CAT proteins produced in this work contain 6xHis tag, even though we did not need the purification step. A variety of biological assays are available fo r detecting the amount of 6xHis ta g fused with a protein. There are many other tags that may be fused with proteins as reviewed in the literature.138 In addition,

PAGE 50

50 luminescence detection can also be used by fusi ng luciferase with prot eins of interest, as demonstrated in the present report. 2.5 Conclusions A novel concept for toxin detection is presen ted based on toxins inhibition of biological protein synthesisin the step of either DNA transcription or protein translation. We demonstrated the feasibility of the concept by (1) in vitro expression of four proteins either from plasmid or linear template, including wide t ype green fluorescent protein, chloramphenicol acetyl-transferase, luciferase, and MonsterPTMP GFP; (2) confirming differential inhibitory effects of two toxin simulants, tetracycli ne and cycloheximide, on the expres sion yields of these proteins in either prokaryotic or eukaryotic expression system; (3) obtaining unique response pattern (or signature) of the 3x4 IVT array device for each toxin simulant.

PAGE 51

51 Table 2-1. Principal compositi on of prokaryotic IVT systems.122 Zubays system Gold-Schweiger S30 of E. coli containing RNA polymerase, ribosomes, tRNAs, ARSes, and translation factors S100 of E. coli containing RNA polymerase, ARSes and translation factor Amino acids ribosome ATP, GTP, UTP and CTP ATP, GTP, UTP and CTP Energy regenerating system: PEP+PK, or CP+CK Energy regenerating system: PEP+PK, or CP+CK Sulfhydryl compounds: ME or DTT Amino acids cAMP Total tRNA Salts: MgP2+P, CaP2+P, KP+P and NHB4PB+P Sulfhydryl compounds: ME or DTT DNA template Salts: MgP2+P, CaP2+P, KP+P and NHB4PB+P DNA template Abbreviation: ARSes, aminoacyl-tRNA synthe tases; PEP, phophoenal pyruvate; PK, pyruvate kinase; CP, creatine phosphate; CK, creatine kinase; ME, mercaptoethanol; DTT, dithiotreitol Table 2-2. Pricinpal compositi on of eukaryotic IVT systems.122 Rabbit reticulocyte lysate Wheat germ extract Lysate containing ribosomes, tRNAs, ARSes, and translation factors Extract containing ribosomes, tRNAs, ARSes, and translation factors SP6,T3 or T7 RNA polymerase SP6,T3 and T7 RNA polymerase Amino acids Amino acids ATP, GTP, UTP and CTP ATP, GTP, UTP and CTP Energy regenerating system: CP+CK Energy regenerating system: CP+CK Sulfhydryl compounds: DTT Sulfhydryl compounds: DTT MgP2+P, KP+P and polyamine MgP2+P, CaP2+P, KP+P and polyamine Hemin, glutathione and cAMP DNA template with SP6, T3, or T7 promoter DNA template with SP6, T3, or T7 promoter

PAGE 52

52 Table 2-3. Prokaryotic vers us eukaryotic IVT systems.102 Prokaryotic Eukaryotic Ribosomes 70S 80S Initiation factors IF: 1, 2 and 3 eIFs 1, 1A, 2, 2A, 2B, 3, 4A, 4B, 4E, 4F,4G, 5, 5A,6 Elongation factors EF-Tu, EF-Ts, EF-G eEF1A, eEF1B, eEF2 Termination factors RF1, RF 2, RF3 eRF1, eRF3 Energy-regeneration system PEP+PK, or CP+CK, or pyrucate+NAD+CoA CP+CK Sulfhydryl compounds ME or DTT DTT Advantage (1)Simplicity and university of DNA template design; (2)High translation rate. (1)High stability; (2) Better compatibility with the synthesis of eukaryotic proteins Disadvantage (1)shorter reaction time; (2)High rate of degradation of DNA, mRNA and protein; (3)Aggregation of protein products (1)Lower translation rate; (2)Complexity of design of DNA template Abbreviation: IF, initiation factor; EF, elongation factor; eEF, eukaryotic elongation factor; RF, release factor; eRF, euka ryotic release factor. Table 2-4. Oligonucleotide primers used for mutant GFP. Orientation Sequence (5-3) Forward CAA CAG TCT CGA ACT TAA GC Backward AAA ACC TCC CAC ATC TCC

PAGE 53

53 Figure 2-1. A picture of two de vices forming an array (3x4) of 12 wells. A US penny is also pictured for comparison of the size. The we lls in each device were laid out according to the standards of 96-well plates, but w ith ~25 time reduction in the well volume. Figure 2-2. Vector maps. (a) pIVEX GFP v ector map. (b) pIVEX CAT vector map. (c) Luciferase control DNA map. (d) phMGFP vect or map. Additional description: ori, origin of plasmid replication; AmpR, b-la ctamase gene (resistant to ampicillin).

PAGE 54

54 protein DNA toxin RNA toxin transcription translation Figure 2-3. A device with an arra y (12x8) of wells for in vitro e xpression of a group of proteins. The area shadowed with dia gonal lines indicates a subs et (3x4) of 12 wells, among which the top row of 3 wells are for expre ssion of three different proteins as the positive controls, the second row for the nega tive controls, and other two rows for the samples with one repeat for enhanced precisi on. Total of 8 series of such a 12-well set may be made for detecting different toxins The insert in the expanded view shows transcription and translation steps in each well; and the processes may be inhibited by toxins.

PAGE 55

55 (a) (b) (c) (d) Figure 2-4. In vitro protein expression in microcentrifuge tubes. (a) WT-GFP expression confirmed by Western blot. Lane 1: pre-st ained protein markers; 2: negative control; 3: recombinant GFP (rGFP) purchased; 4: GFP expressed. (b) WT-GFP confirmed by fluorescence spectrometry. Top: excitation sp ectrum; Bottom: emission spectrum. (c) CAT expression confirmed by We stern blotting. Lane 1: prot ein markers; 2: negative control; 3: CAT expressed. (d) Lucifera se expression confirmed by luminescence detection. Lane 1: negative control; 2: luciferase expressed. The intensity of luminescence in the y axis is in log scale. Figure 2-5. PCR analysis by agarose gel electrop horesis. Lane 1: PCR product after purification; Lane 2: DNA marker; Lane 3: raw PCR product.

PAGE 56

56 Figure 2-6. Excitation and emi ssion spectra of the MonsterPTMP GFP. (a) The excitation spectra of expressed MonsterPTMP GFP. (b) The emission spectra of expressed MonsterPTMP GFP, excited at 480 nm. Top curve: MonsterPTMP GFP synthesized from phMGFP vector; Middle curve: MonsterPTMP GFP synthesized from PCR-amplified DNA; Bottom curve: negative control. 0 2000 4000 6000 8000 10000Negative controlroom temperature30 oC (b) ControlExpression Yield (A.U.) (a) Room Temperature 30 oC Figure 2-7. Comparison of in vitro GFP synthesis between at room temperature and at 30 oC. (a) Western blot analysis. Lane 1: pre-stained protein markers; 2: ne gative control; 3: GFP synthesized at room temperat ure. 4: GFP synthesized at 30 C. 5: recombinant GFP (rGFP) purchased. (b) Quantitative co mparison of the GFP expression yield between room temperature and 30 oC. (a) (b)

PAGE 57

57 0.0 4.0 8.0 12.0 16.0 -4048 0 0.2 0.4 0.6 0.8 1 -4048 0 0.2 0.4 0.6 0.8 1 -4048 0 4 8 12 16 -30369 0 0.2 0.4 0.6 0.8 1 -30369 0 0.2 0.4 0.6 0.8 1 -30369(c) CH effect on GFP expression (d) TC effect on CAT expression (e) CH effect on CAT expression (f) TC effect on luciferaseexpression (g) CH effect on luciferaseexpression123456789(a) (b) TC effect on GFP expression Figure 2-8. Inhibitory effects of TC and CH on cell-free protein expression in microcentrifuge tubes. (a) Western blot analysis confir ms the inhibitory effects of TC on the expression yield of GFP in E. coli expre ssion system. Lane 1: pre-stained protein markers; lanes 2-6: expression of GFP with 3000, 300, 30, 3, 0.3 ng/L of TC, respectively; lane 7: positive control; la ne 8: negative control; lane 9: rGFP purchased. (b-g) Inhibitory effects of TC (b, d, and f) or CH (c, e, and g) on the expression yield of GFP (b, c) in E. coli ex pression system, of CAT (d, e) in E. coli expression system, and of luci ferase (f, g) in rabbit retic ulocyte expression system. The expression yields of GFP and CAT we re quantified by Western blotting while that of luciferase was measured by lumine scence. All x axes are the concentration (ng/L) of toxin, in log scale. Y axes are the amount of expressed protein either normalized to the positive cont rol (b-e) or in log scale of luminescence signal (f, g).

PAGE 58

58 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1p o s i t i v en e g a t iv eT C T CGFP CAT luciferasesignal (normalized)(a)p o s i t iv e n e g a t iv eC H C HGFP CAT luciferasesignal (normalized)(b) Figure 2-9. The response pattern of the 3x4 IV T sensor array for two toxin simulants, tetracycline (TC, a) and cycloheximide (CH, b). The experiments were carried out in two of the 2x3 well devices in Figure 1. The signals for the positive control were from the first row of 3 wells in the device, in which GFP, CAT, and luciferase were expressed in their respective expression syst ems. These wells were free of toxins. The signals for the negative control were from the second row of 3 wells in the device, in which the expression vector wa s not added. The signals for the samples were from the remaining two rows of 3 wells in the device, in which either 17 ng of CH or 25 ng of TC was added into the protein expression system.

PAGE 59

59 CHAPTER 3 NESTED-WELL DEVICES FOR IN VITR O PROTEIN SYNTHESIS AND RICIN DETECTION1F1 3.1 Introduction Detection and identification of toxic agen ts are important for medical diagnostics, food/water safety testing, and biological warfar e defense. Methods to detect them include immunoassay,142, 143 sensors,144, 145 mass spectrometry,146 and genetic analysis.147-150 Nucleic acid-based genetic analysis, e.g., polymerase ch ain reaction (PCR), invol ves DNA amplification that offers high sensitivity and unambiguous iden tification. However, it is not applicable to toxins that contain no nucleic acids One example of such toxins is ricin, which is listed as a Category B bioterrorism agent according to th e Centers for Disease Control and Prevention151 and was used as an agent in the letter sent to US congress in Feb. 2004. The structure, toxicity, mechanism of action, and current detection methods of ricin will be discussed in section 3.2. Immunoassay is advantageous over other methods due to its high sensitivity. However, it requires an antibody that is specific to the agent of inte rest and it often involves labor-intensive sample preparation. Therefore, it can not be used for detecting unknown or a new agent because an antibody is simply not available. With the in creasing ability to modify and engineer potential warfare agents, a simple, fast, and highly sensitiv e method to detect agents is essential. To address this challenge, we describe an approach that exploits the mechanismby which toxins cause toxic eff ectsas the sensing scheme.137, 152 For example, ricin acts on the 28S ribosomal subunit and prevents the binding of elongation factor-2, a critical component in the process of protein translation. That is, ricin kills people by bl ocking protein synthesis in cells 1 A part of this chapter has been published in (140) Mei, Q.; Fredrickson, C. K.; Lian, W.; Jin, S.; Fan, Z. H. Ibid. 2006 78 7659-7664., and in (141) Mei, Q.; Fredrickson, C. K.; Simon, A.; Knouf, R.; Fan, Z. H. Biotechnology Progress 2007 23

PAGE 60

60 of human body. Similarly, a num ber of biological toxins exhi bit their toxin effects through inhibition of protein synthesis, includi ng Shiga toxin, diphtheria, and exotoxin A.137 As discussed in chapter 2, in vitro protein synthesis has been inplemented in microwell IVT array. One of the major problems of IVT pr eformed in this microwell device is its short reaction time, thus significan tly reduces protein expression yield. The problem has been addressed by the invention of continuous flow cell-free and continuous exchange cell-free systems, in which nutrients can be refurbis hed and inhibitory byproducts can be removed (summarized in section 3.2.2). The nested-well device in this chapter consists of a mechanism for fluid manipulation. The performance of the device is demonstrated by synthesizing green fluorescent protein, chloramphenicol acetyltransferae, and lucifera se. The production of IVT between the nestedwell device and a microcentrifuge tube was comp ared. The effects of hydrostatic flow and connection depth on the protein pr oduct yield are also investigat ed. Higher protein expression yields can be obtained in the nested-well device, leading to larger detection signals (lower detection limit) when ricin is present. More importantly, this detection approach accomplishes signal amplification. For each copy of DNA, thousa nds of copies of proteins can be produced. The inhibitory effects of ricin on the production of each copy of protein are accumulated, leading to a significantly enhanced dete ction signal. In addition, rici n detection using the one-step operation of protein expression simplifies the de tection procedure compared to ELISA that consists of many steps of a pplying reagents and washing.153 The detection can be achieved in as short as 5 minutes.

PAGE 61

61 3.2 Background 3.2.1 Ricin Ricin, extracted from the seeds of the castor bean plant ( Ricinus communis ), is one of the most toxic and easily produced plant toxins. Rici n consists of two chains, A and B, connected by a disulfide bond. The B chain binds to galactose or N-acetylgalactosamine of glycoproteins and glycolipids on the surface of a e ukaryotic cell, stimulating the e ndocytic uptake of the ricin. The A chain is an enzyme that binds and depurinat es a specific adenine of the 28S ribosome RNA (rRNA). The adenine ring of the ribosome beco mes sandwiched between two tyrosine rings in the catalytic cleft of the enzyme and is hydrolyzed by the enzyme N-glycosidase action. As a consequence, aminoacyl-tRNA can no longer bind to trancated 28S rRNA and protein synthesis is blocked in the cell, which lead to cell death and tissue image.137, 154 The target adenine is a specific RNA sequence that contains the unusua l tetranucleotide loop, GAGA. The B chain has two binding sites for galactose, and about 106 ricin molecules may bind per cell. However, just a single ricin molecule that en ters the cytosol can inactivate over 1,500 ribosomes per minute and kill the cell.155, 156 Although ricin has been shown to inhibit tumor growth and it has been applied for cancer therapy,157, 158 it is considered for biological wea ponization and is listed as a Category B bioterrorism agent due to its extreme ease of pr oduction and high toxicity. The lethal dosage of ricin, which is lethal to 50% of the exposed population, is approxima tely 3-5 g per kilogram of body weight.159 Since ricin has no selectivity for specifi c cells, the clinical signs and symptoms of ricin intoxicity depend on th e dose and the route of exposure. For example, inhalation causes cough, fever, nausea and pulmonary oedema; or al injection results in vomiting, diarrhea, hydration, low blood pressure, or necros is of spleen, liver and kidneys.160

PAGE 62

62 The methods to detect ricin at low concentration levels typically include enzyme-linked immunosorbent assay (ELISA)161 and immunoassay using radioactive labeling.162 Although offering high sensitivity, they involve many step s that are labor-intensi ve and time-consuming. For radio-immunoassay, the handling and disposal of radioisotopes are always environmental challenges. Recently, Liglers group reported a fluorescence-based multianalyte immunosensor that has a detection lim it of 25 pg/L of ricin.145, 163 Stine and Pishko exploited specific interaction between ricin and glycosphingolipid s and developed a quartz crystal microbalance sensor with the detection limit of 5000 pg/L of ricin.164 Shankar et al. depicted an immunoassay-based magnetoelastic sensor that shows a detection limit of 5 pg/L of ricin.165 The detection limit and time required for detectio n of ricin are summarized in Table 3-1. 3.2.2 Continuous-Flow Cell-Free Protein Expression Systems In an IVT system performed in the fixed volum e of a microcentrifuge tube (Figure 3-1a), the main limitation is its short lifetime because reaction terminates as soon as any essential substrate is exhausted, or by-products accumulate to the level of inhibition, usually within 60 minutes. As a result, this leads to poor yiel d of protein. The major breakthrough was achieved by Spirin and coworkers with the invention of the continuous-flow cell-free (CFCF) system, in which the reaction products were continuously removed and the nutrients were continuous supplied.85 This can also be achieved by a semi-per meable membrane, which maintains the high molecular weight components in the reaction ch amber and separates it from feeding chamber containing the small molecular components. In th is case, the nutrients are continuously supplied into and all the products includ ing synthesized proteins are re moved from the chamber by the forced flow of the feeding solution through an ultrafiltration membrane (Figure 3-1b). The first CFCF system developed by Sp irin was a modified Amicon 8MC mircoultrafiltration chamber.85 Typically, the volume of the r eaction mixture was 1 mL and the

PAGE 63

63 feeding solution was pumped in with a constant flow rate of 1 mL/hour. The ultrafiltration membranes used in their report were Amicon PM-10, PM-30, XM-50, with respective molecular weight cut-offs of 10 KDa, 30KDa and 50KDa. Both E. coli lysate and wheat germ extract were tested in this reactor for translation of various mRNA, resulting in extended reaction time (20-40 hr) of synthesis and a significantly high yield of products. The products are relatively small proteins including MS2 coat prot eins (16 KDa), brome mosaic vi rus coat protein (19kDa) and calcitomin polypeptide (3.4KDa).85 Since then, several other groups reproduced and optimized the CFCE systems and realized coupled transcription-translati on in the CFCF format. For examples, Endo et al. reported dihydrofolate reductase (DHFR) w ith enzymatic function was synthesized up to 20 hr using the same reactor as Spirins.86 Kigawa improved Spirins design by using a HPLC pump instead of peristaltic pump to make better control of a constant flow rate.87 In addition, Kigawa used a coupled E. coli IVT and produced active chloramphenico l acetyltransferase (CAT). Other similar reports are summarized in Table 3-2. 3.2.3 Continuous-Exchange Cell-Free System Another simple way to allow continuous supply of nutrients and removal of products is to employ a dialysis bag or a flat dialysis membra ne for the passive exchange of low molecular components, named as continuous-exchanged cell-free (CECF) systems166 (Figure 3-1c). Kim reported 1.2 mg of CAT was synthesized in 1 mL dialysis bag and the r eaction can extend up to 14 hr.88 Kigawa developed a high-productive E. coli CECF system and the yield of CAT and RAS protein increased to 6 mg /mL over 21 hours of incubation.101 More recently, Madin presented a highly productive whea t germ CECF system which can synthesize the amount of 0.5 to 1.5 mg/ml of several proteins, such as DHFR, GFP and luciferase in 24 hr.167 As an alternative, a hollow fiber reactor with incr eased filtration area was proposed by Yamamoto.168

PAGE 64

64 This modified configuration also enabled in situ concentration of extract, thus speeds the reaction rate. The commercial CECF instrument with a programmable control ling temperature, supplemented with well-developed reagent k its, was launched on to the market by Roche Applied Science; this is called Rapid Tr anslation Systems which use an optimized E. coli lysate or wheat germ extract. The production of various proteins can be synthe sized up to 5 mg/mL in 24 hours.93 Table 3-2 summarized nearly all publishe d data so far reported on CFCF and CECF protein synthesis. Although CFCF systems have several obvious adva ntages (purity and hi gh yield of protein products), the operational complex ities (pump and valve) limit the potential for practical use. Due to the simplicity and low cost, CECF systems became attractive to scientists during the past decade. In addition, CECF systems could be achieved in a miniaturized format and become a powerful tool for high-throughput expression of pr oteins, as discussed in following sections. 3.3 Experimental 3.3.1 Materials and Reagents The RTS 100 wheat germ CECF kit, RTS 500 E. coli kit, and the expression vectors containing the gene of green fluorescent protei n (GFP) and chlorampheni col acetyl-transferase (CAT) were obtained from Roche Diagnostics GmbH (Mannheim, Germany). T7 luciferase DNA vector, luciferase assay reagent, and nuc lease-free water were acquired from Promega Corporation (Madison, WI). Ricin (MW 60 KD a) and ricin B chain (MW 32 KDa) were purchased from Vector Labs (Burlingame, CA) while ricin A chain (MW 29 KDa) and 2mecaptoethanol were from Sigma (St. Louis, MO). The values of their molecular weight (MW) are provided by the respective ma nufacturers. Acrylic sheets with thickness of 0.25 (6.3 mm) and 0.22 (5.5 mm) was from Lucite International Inc.(Cordova, TN ). The dialysis membrane

PAGE 65

65 with the molecular weight cutoff of 8 KDa was obtained from Spectrum Labs (Rancho Dominguez, CA). 10 nm pore size of polycar bonate membrane was purchased from GE Osmonics Labstore (Minnetonka, MN) and pol ycarbonate membranes with 15 nm and 100 nm pore size were from Whatman Inc (Florham Pa rk, NJ). A biocompatible epoxy (353ND-T) was bought from Epoxy Technologi es (Billerica, MA). 3.3.2 Device Fabrication The design of the nested-well devi ce is shown in Figure 3-2a; it consists of two parts. The top part, tray, was fabricated by drilling through -holes and a common fla nge in a piece of 5.5 mm thick acrylic sheet using a CNC-mill (Flashcut CNC, Menlo Park, CA). The inside diameter of the hole is 3 mm, surrounded by a 1 mm thick wall, creating a structure with an outside diameter of 5 mm. The distance between the hole centers is 9 mm, following the microplate standards defined by the Society for Biomol ecular Screening (SBS) and accepted by the American National Standards Institute. The botto m part, well, was created by milling an array of 4 mm deep wells into a piece of a 6.3 mm-thick acr ylic sheet. The diameter of the wells is 7 mm; each well is concentric with the correspond ing tray chamber when they are assembled. A picture of both parts in a 2 x 3 array is in Figure 3-2b. A dialysis membrane is incorporated in th e device to connect the two chambers as illustrated in Figure 3-2c, providing a means to supply nutrients continuously and remove the reaction byproducts. The membrane is glued to the bottom of each tray chamber using adhesives. Epoxy 353ND-T is cured at 80 C in an oven for 20 minutes The assembled devices are rinsed with nuclear water and then sterilized by exposing to a UV light for 30 minutes to eliminate possible inhibiti on of enzymatic reactions.

PAGE 66

66 3.3.3 In vitro Protein Expression For the prokaryotic expression system, GFP and CAT were synthesized using a RTS 500 E. coli kit. Reaction and feeding so lutions were prepared as r ecommended by the manufacturer. In brief, the reaction solution contained 0.525 mL E.coli lysate, 0.225 mL reaction mix, 0.27 mL amino acid without methionine and 30 L methioni ne. The feeding solution consists of 8.1 mL feeding mix, 2.65 mL amino acid w ithout methionine and 0.3 mL me thionine. To run the protein synthesis in the miniaturized membrane device, th e tray was filled with the 8 L of the reaction solution containing GFP vector or CAT vector while 80 L of the feeding solution was added to each well in the well plate. A biocompatible sealing tape was used to seal the tray to prevent evaporation. The device was placed on a shaker with a speed 100 rpm and incubated at room temperature for 2, 4, 6, 10, 14 and 20 hrs. As fo r comparison, 8 L of reaction solution was dispensed into microcentrifuge t ube and incubated the same time period as that of the membrane device. After GFP synthesis, th e reaction tubes were stored at 4 C overnight for the maturation of GFP. The synthesized GFP and CAT were analyzed using Western blotting.99 Gel images are scanned by a flatbed scanning and quantified us ing ImageJ from the National Institute of Health.132 GFP production is also conf irmed by a microplate reader (Tecan Safire, Research Triangle Park, NC). For the eukaryotic expression system, luci ferase was synthesized using RTS 100 wheat germ expression kit. The reaction solution for trays was prepared by mixing 15 L wheat germ extract, 15 L reaction mix (provided in the kit) 4 L amino acids, 1 L methionine and 15 L of nuclease-free water containing 1 g of luciferase vector. For each tray chamber, 8 L of the reaction solution was used. The feeding soluti on for wells was prepared by combining 900 L feeding mix (provided in the kit), 80 L amino acid and 20 L methionine. In each well, 80 L of the feeding solution was introduced. The tray and well plates were assembled and then placed

PAGE 67

67 on a shaker (at room temperature) for a period of time (e.g., 0.5 hour). The amount of luciferase synthesized was determined by mixing the expre ssion product with lucife rase assay reagents, followed by luminescence detection in a lumino meter (Berthold, Germany), as described in chapter 299. When luciferase was synthesized in a microcentrifuge tube, the same reaction solution (8 L) was used w ithout the feeding solution. 3.3.4 Ricin Detection Ricin is highly toxic to human. The material safety data sheet (MSDS) should be reviewed before handling this chemical. Ricin samples should be handled in de dicated laboratories and appropriate safety precautions are required when preparing rici n solutions. The personnel who perform the experiments must wear protective de vices including a lab co at, gloves, and a face mask. Contaminated labware such as microcentr ifuge tubes, pipet tips, and used devices are disposed into a biohazard container. A series of concentrations of ricin A chai n solutions, ranging from 0.035 to 0.69 nM, were prepared from a stock solution of 35 M. To demonstrat e ricin detection, 6 L of the reaction solution (discussed above) is pipe ted into the tray, followed by 2 L of ricin samples. The volume of the feeding solution remained at 80 L. For the positive controls in the same device, 2 L of water was added. The negative controls contain no luciferase vect or, providing with the background signal. To achieve lower detection limit, 4-hr protein expression was used, though ricin detection can be achieved in as short as 5 minutes. The same protocol was used in studying the toxicity level of various forms of ricin, including whole ricin, ricin A ch ain, ricin B chain, heat-denatured ricin A chain, and whole ricin treated with 2-mercaptoethanol. Denature of Ricin A chain was achieved by heating the samples at 95 oC for 5 minutes. Reduction of disulfide bond of two chains in ricin was carried out by

PAGE 68

68 mixing ricin with 50 mM 2-mercaptoethanol, fo llowed by the incubation of the mixture at 52 oC for 10 minutes. 3.4 Results and Discussion 3.4.1 Device Design As mentioned above, protein expression can be produced in a cell-free medium employing IVT. It is also well-recognized that high-yield protein expression of IVT can be attained in the flow of a feeding solution, but not un der static conditions in a fixed volume.85, 102 As a result, commercial bench-top instruments incorporate the principle of c ontinuous flow with a magnetic stirrer.93, 169 However, the bench-top instruments often employs milliliters of reagents93 and it is difficult to achieve the high-throughput form at as discussed by Angenendt et al.68 We fabricated a nested-well device consisti ng of a mechanism for fluid manipulation; it also has a potential to implement protein synthesis in a high-throughput format due to miniaturization. As illustrated in Figure 3-2, IVT was implemented in an array of units; each unit is for expression of one prot ein (e.g., GFP). Each unit in the de vice consists of a tray and a well (Figure 3-2c). The tray chamber is for th e IVT reaction; the well is concentric with the corresponding tray chamber and func tions as a nutrient reservoir. The well contains amino acids, adenosine triphosphate (ATP), and other reagents The tray contains the cell-free expression mixture extracted from E. coli or wheat germs, as well as the same reagents in the well. The solution array is designed to conform with 96-well microplates, ensuring compatibility with a variety of commercial fluid dispensing systems and commercial plate readers for detection. A dialysis membrane is glued to the bottom of the tray, connecting the tray and well and providing a means to supply nutrients and remove the reaction byproducts. The incorporation of membrane is critical 85, 129 because of two facts: (1) the flow of a nutrient-feeding solution will lead to higher expression yield compared to st atic conditions, because protein synthesis will not

PAGE 69

69 terminate earlier due to fast depletion of th e energy source (ATP); (2) removal of small molecular byproducts is also crit ical to high yield e xpression of proteins in a cell-free medium, because possible inhibition of protein synthesi s by the byproducts (e.g., hydrolysis products of triphosphates) will not take place. The connection between the tray and well cham bers is a hollow circ le, though it looks like a channel in the cross-sectional view in Figure 3-2c. The height of the connection ranges from 100 m to 2 mm. One of the mechanisms for su pplying nutrients from the well chambers to the tray chamber and for removing byproducts from the tray chamber to the well chambers is diffusion, the transport of species resulting from the difference in the c oncentration of solutes between two solutions separated by the membrane. The rate of diffusion is determined by the diffusion coefficient and the concentration gradient of solutes. In addition, the flow to supply fresh solution from the well to the tray is a ugmented by a hydrostatic pressure, which is caused by the difference in the solution level between th e tray and well. When the well has slightly higher solution level than the tr ay, the pressure difference resulting from the height difference will drive nutrients from the well into the tray. Incorporation of the membrane and fluidi c connection enables continuous supply of nutrients and selective removal of small molecule byproducts. Th e nutrient-feeding solution will avoid fast depletion of the energy source, a nd the accumulation of small molecular byproducts will not take place. This device design allows pr otein synthesis to continue for up to 20 h with higher protein yield as discussed later. Using the dimension discussed in 3.3.2, the volume of a tray chamber is 18 L. We typically fill in the tray with 8 L of an IVT so lution; this volume is more than 2 orders of magnitude less than 1 mL of a reaction solutio n in RTS 500. Similarly, the volume of a well

PAGE 70

70 chamber is 107 L after a tray is placed inside. We typically fill in th e well with 80 L of a feeding solution; this volume is again more than 2 orders of magnitude less than 10 mL of a feeding solution in RTS 500. The decrease in the volume of both reaction mix and feeding resolution will significantly reduce reagent consumption when using IVT for high-throughput assays 170 Although Figures 3-2 shows a 2 x 3 array device an array of a higher number (e.g., 8 x 12) can be implemented as is readily appreciated by th ose in the field. As explained above, such an nested-well device has the fo llowing advantages: (1) less reagent consumption due to miniaturization; (2) higher prot ein expression yield due to incor poration of a mechanism for fluid manipulation; and (3) a potential to implement protein synthe sis in a high-throughput format. 3.4.2 Protein Synthesis To demonstrate the function of the nested -well device, GFP was synthesized using commercial E. coli system with a GFP coding sequen ce. GFP product was analyzed using Western blotting. Figure 3-3a show the Western blot image, and the signal intensity is plotted in Figure 3-3b To show the effects of the flow manipul ation, we did the same reactions in a microcentrifuge tube and found that protein e xpression ceased after 4 h. This result clearly illustrates the importance of incorporation of the dialysis membrane and fluidic connection in the device. Continuous supply of nutrients and se lective removal of small molecule byproducts enable protein synthesis to continue for up to 20 h in the device, which is 5 times longer than the expression time in the tube. The longer expre ssion time results in higher expression yield. The amount of GFP synthesized in the device is about 13 times larger than in the tube. The quantified data as shown in Figure 3-3b indicat ed the device design is proper. Since GFP is naturally fluor escent, the expression produc t can also be quantified by fluorescence emission. Using a commercial mi croplate reader, the emission spectrum of

PAGE 71

71 synthesized GFP is shown in Figure 3-4a, as is the spectrum of the recombinant GFP we purchased. The same fluorescence spectrum also confirms our success of GFP expression in the device. It should be noted that the G FP synthesized must be incubated at 4 oC to let proteins fold correctly in order to have appr opriate three-dimensional structur es for fluorescence. Figure 3-4b shows the fluorescence intensity as the func tion of the incubation time. After 20 hours incubation, the fluorescence intensity reached the plateau. Fluorescence is a highly sensitive and widely used analytical method. But there are potential limitations when we try to use the fluorescence intensity to quantify the GFP expressi on in the microwell devi ce. First, the long folding time of GFP lead slow response time. Even though lots of mutagenesis studies have improved the folding properties, the slow matura tion time is still a big obstacle of practical application of GFP. Second, in tegrating the excitation light source into the microdevice will involve a lot of microfabrication time-consuming and costly steps. Moreover, it is very difficult to completely filter the excitation light in the chip, which may induce high background and greatly decrease the sensitivity of the system. Similar expression has been achieved for chlora mphenicol acetyl-trans ferase (CAT). CAT is expressed in the same E. coli expression system; success of the protein expression is confirmed by Western blotting in Figure 3-5a. The signal intensity as the function of the reaction time is shown in Figure 3-5b. These resu lts also suggest longer expression time can be achieved in the device than in th e tube. As a result, the produc tion yield increased more than 22 fold in the device. This confirms that we have achieved the desire d fluid manipulation, both continuous feeding of the nutrient solution and removal of the byproducts, in the device. Because of the limitation of fluorescence dete ction on the chip, we further investigated a low background bioluminence system. Synthesis of luciferase is carrie d out using wheat germ

PAGE 72

72 expression system. Detection of the expression product is achie ved by monitoring the intensity of luminescence after mixing th e product with luciferin and ATP as reported previously 140. Figure 3-6 indicates the production yi eld of luciferase increased more than 2.6 fold in the device than that in a microcentrifuge tube. In additi on, luciferase folding can be processed during the process of translation.171 These results suggest that we ach ieved the desired fluid manipulation in the device. 3.4.3 Hydrostatic Flow As mentioned above, one mechanism for supplying nutrients from the well to the tray is the hydrostatic flow due to the difference in the solution level between the tray and well. We studied the effects of the fluid level difference on the protein e xpression yield. Figure 3-7 shows the protein expression yield temporal profiles wh en the fluid level difference changes from -0.5 mm to 2 mm. The fluidic connection depth is fixe d at 500 m. When the he ight is -0.5 mm, i.e., a reverse flow from the tray to the well exist, protein expression is significantly hindered due to depletion of the expression compone nts in the tray. When the flui d level difference is near 0, the expression yield increases 183% a nd normal temporal profile is obs erved. In this condition, the fluid flow should be primarily due to diffusion and agitation. Wh en the fluid level difference increases from 0 to 2 mm, the expression yield increases 38%, presumably due to additional hydrostatic flow. This result indi cates that hydrostatic flow pl ays a role in the replenishing nutrients in the device. The flow of material across the membrane is believed to be driven by a three possible mechanisms: hydrostatic pressure, osmotic pressure, and diffusion. Hydrostatic pressure is due to the differe nce in solution levels on both sides of the membrane; any difference will result in a pressure gradient across the membrane. The magnitude of the pressure difference is given by

PAGE 73

73 P = gh where P is the hydrostatic pressure, is the density of fluid, g is gravitational acceleration, and h is the height difference between liquid in terfaces of the reaction and feeding chambers Osmosis is the spontaneous net movement of water across a dialysis membrane from a region of high solvent potential to an area of low solvent potential when a solute concentration gradient is present. Osmosis likely take pl ace because solutions possess different strengths, and water crossing the membrane from the dilute side to the concentrated side in an attempt to balance the concentrations. The magnitude of osmosis is given by Vant Hoffs equation: = cRT where is the osmotic pressure, c is concentratio n, R is the universal gas constant, and T is temperature in Kelvin. This equation predicts that the osmotic pressure will be higher when concentration is greater (Note that it doesnt depend on what th e solute is, just its overall concentration; if the solution has many solutes, c is the overall concentration regardless of species). Transport of materials across the semi-permeab le membrane should be governed by Ficks law of diffusion, which in one-dimensional form is expressed as: dy dC D NA A A In this equation, NA represents the molar flux of A with respect to a fixed reference frame (mol/t L2), DA is the mass diffusion coefficient, and CA is the concentration. By integrating this equation and applying the boundary conditions (concentrations on each side of the membrane and membrane thickness), we find that low A high A A AC C D N,

PAGE 74

74 This equation suggests that if the solutions on either side of the membrane have different concentrations, solute will travel from the re gion of high concentra tion to region of low concentration. Since solutions in the feeding and reacti on chambers are from Roche, the exact compositions are unknown. Our calculation below i ndicates that the concen tration of amino acid is same between the reaction and feeding chamber. In addition, Kim et al. mentioned that the concentration and the components in reaction solution and feeding solution are identical except for RNA polymerase, DNA template, and tRNA mixture.88 The following calculation is based on the concentration of com ponents provided by Kim et al:88 At the beginning of reactions: Camino acid, tray= 500 M Camino acid, well= 500 M CATP, tray= 1.2 mM CATP, well= 1.2 mM Cbyporduct, tray= 0 Cbyproduct, well=0 h=2 mm Hydrostatic pressure= g h=(998) (9.8) (0.002)=19.5608 Pa tray= well Ntray=Nwell In the middle of reactions, the concentra tions are unknown, making calculation difficult. However, we believe that the molecules of amin o acids and ATP diffuse from the well to tray due to higher concentration in the well whereas by -products diffuse from the tray to the well. Moreover, the total concentration of solutes is lik ely slightly higher in th e tray, thus the osmosis will drive water from the tray to well. However, this effect should be overcome by the hydrostatic difference.

PAGE 75

75 3.4.4 Effects of Membrane Pore Size Different pore sizes of membranes have an e ffect on transportation of molecules through the membrane. 10 nm, 15 nm and 100 nm pore size of polycarbonate membrane and polycarbonate membranes are studied for CECF prot ein synthesis. The field emission scanning electron microscopy (SEM) images of 100 nm and 10 nm pore size of PC membrane at room temperature and 170 C are shown in Figure 3-8 and Fi gure 3-9. The images indicated increasing temperature has negligible e ffects on the pore size of PC membrane. The chip with the connection depth of 0.5 mm was created and the effects of membranes on the yield of protein synthesis were investigated. Luciferase was synthesized using a RTS 100 wheat germ expression kit. For each inner ch amber, 6 L of the reaction solution was used; while in outer chamber, 120 L of the feeding solution. The amount of luciferase synthesized was determined by mixing the expression product with luciferase assa y reagents, followed by luminescence detection in a lumino meter, as described previously.99 The yield of luciferase production among the device with different membrane s and a microcentrifuge tube are compared and plotted in Figure 3-10. The results indicated the polycarbonate membranes with 10 nm and 15 nm pore size can improve the yield of protei n expression, while no sign ificant change using 100 nm pore size of polycarbonate membrane on protein expression compared with a microcentrifuge tube. The yield of luciferase synthesized in the de vice using polycarbonate membranes with 10 nm and 15 nm pore size ar e about 2.3 fold and 1.8 fold higher than in a microcentrifuge tube. However, the yield of luciferase production in the device using polycarbonate membranes is less than that using a dialysis membrane. The result indicated the pore size of membranes is critical for CECF prot ein synthesis. The dialysis membrane with the molecular weight cutoff 6 KDa-8 KDa contains tort uous pores of 2 nm in average. The pore is large enough to allow small molecular weight substrates (amino acids, ATP and PPi) pass

PAGE 76

76 through, but that are small enough that larger mo lecular weight compounds (e.g., proteins and ribosome) cannot get through. However, polycarbona te membranes contain uniform, cylindrical pores. With the increasing of pore size, the larg e molecules such as the synthesized protein, polymerase and ribosome can not be retained in the inner chamber. Therefore the leaking of synthesized protein will reduce the productivity of protein synthesis. In the following experiment, the dialysis was sele cted for the best results. 3.4.5 Biological Sign al Amplification As discussed above, ricin causes toxic effect s by inactivating ribosomes and inhibiting protein synthesis in biological cells and then leading to cell death and tissue damage.137, 154 We exploit its toxicity mechanism as the sensing sc heme to detect ricin. This detection method possesses inherent biological signa l amplification, as illu strated in Figure 311. For each copy of DNA, thousands of copies of proteins can be pr oduced. This is estimated by the amount of DNA vector used and the amount of the corresponding proteins produ ced in IVT. The inhibitory effects of ricin exist on the production of every copy of protein, as shown in Figure 3-11. As a result, the detection signal (i.e., the difference between the sample and th e positive control) is accumulated, leading to an amplified signal. This signal amplification is similar, to some degree, to the enzyme-enabled signal amplification in ELISA. Intrinsic to ELISA is the addition of reagents conjugated to enzymes; assays are then quantified by the build-up of colored products after the addition of the corresponding product. The signal am plification results from the enzyme that catalyzes a certain amount of substrates to dete ctable products. Two widely-used enzymes are horseradish peroxidase and alkaline phosphatase, which transfer o-phenylene diamine and pnitrophenylphosphate, respectively, and generate colored products. Therefore, we expect the detection method based on prot ein inhibition has comparable sensitivity with ELISA.

PAGE 77

77 3.4.6 Detection of Ricin by In Vitro Protein Synthesis To demonstrate the detection of ricin, we studied its inhibitory effects on luciferase expression by adding a series of concentrations of ricin A chain into the IVT reactions in the array device. As shown in Figure 3-2a, the units on the left of the array are for the positive controls (free of ricin), the units in the middl e for the negative controls (no DNA vectors), and the units on the right are for samples. Two ro ws of each unit are for repeat experiments to enhance the precision. The positiv e and negative controls are used for quantification as well as for reducing false positives and negatives for to xin detection. As shown in Figure 3-12a, the expression yield of luciferase indicated by the lu minescence decreased with the concentration of ricin A chain. However, the expression yield rema ined the same when the ricin A chain was heat denatured and its toxicity was deactivated (open circles). The error bar of each data point indicates the standard deviation that was obtained from three repe at experiments. The calibration curve is obtained by plotting the detection signal (i.e., the differe nce between the samples and the controls) as a function of the ri cin concentration (Figure 3-12b) A linear relationship exists from 0.035 to 0.69 nM. For a larger concentratio n range, a non-linear be havior of inhibition reactions is observed as expecte d. The detection limit is calcul ated to be 0.01 nM (0.3 ng/mL) by using the criterion that the si gnal-to-noise ratio is three times the standard deviation of the blank. Since the detection signal is dependent on the accumulation of inhibitory effects of ricin on protein production, longer protei n expression (within the ascending range of the curve in Figure 3-12) will lead to lower detection limit. The result in Figure 3-12 was obtained after 4 hours of protein expression, at which protein expre ssion yield just reaches a plateau as indicated in Figure 3-7. However, luciferase can be synt hesized in as short as 5 minutes, as shown in Figure 3-13a. Similar to the result in Figure 3-6, when a longe r expression time was used, we also observed the

PAGE 78

78 difference in the luciferase expression yi eld between the miniaturized device and a microcentrifuge tube when a short expression tim e (<30 minutes) was used. The result suggests that enough difference takes place in as short as 5-minute. A shor t IVT time is critical for those applications that need a quick res ponse. We confirmed that we were able to detect ricin within 5 minutes, as shown Figure 3-13b, though the dete ction limit is higher than when a longer expression time is used. 3.4.7 Toxicity Level of Ricin It is well known that ricin is a dimer, in which B chain binds to cell surface, allowing A chain to penetrate the cell to inhibit protein synthesis.154, 172-176 In other words, the toxicity of ricin comes from the A chain. Since IVT does not provoke dissociation of the A-B dimer, all experimental results discussed above are from ricin A chain. If whole ricin (with A and B chains) is used, it will be intere sting to see what the toxic eff ect is in the IVT device. The literature indicates that the A chain alone has a gr eater inhibitory effect on translation than does the complete ricin molecule.177 As shown in Figure 3-14, we obs erved the similar result in IVT device. The A chain has highest toxicity, B chain has no detect able toxicity, and whole ricin shows a toxicity level slightly le ss than the A chain. However, 2mercaptoethanol-treated ricin is as effective as A chain due to reduct ion of disulfide bond between two chains. We also studied the effects of heat-denatur e on the ricin. We observed that there was a negligible inhibitory effect on luciferase expression when ricin A chain was heat-denatured and its biological activity was disabled (Figures 312a and 3-13b). In other words, there was no significant difference between the positive controls (free of ricin) and the one with the denatured ricin A chain. We can infer from these results th at the IVT method can dete ct the toxicity level of ricin after physical and/or chemical treatments.

PAGE 79

79 3.5 Conclusions An array of nested wells with microfluidic connection has been developed for cell-free protein expression. The array device is fabric ated by milling a tray chamber from acrylic and polycarbonate, attaching a porous membrane, and th en nesting it in a well chamber. The device has been demonstrated for in vitro synthesis of GFP, CAT, and luciferase. The duration of protein expression is significan tly extended due to continuous fe eding of nutrients and removal of byproducts. The reactions in the wells with the feature of continuous-exchange last 5-10 times longer than the same reactions in a micr ocentrifuge tube. The production yield in the device is 13-22 time higher than that in the conve ntional reaction vessels (either microcentrifuge tubes or 96 well plates). We also demonstrated the capac ity of our device for ricin detection using the mechanism of protein expression pathway. Protein expression in our de vice can sustain a translation reaction as long as 20 h, thereby the expression efficiency is hi gh. Moreover, the miniaturized device has the further advantage of reducing the volume of reagents, thus the costs. The nestedwell device appears to perform better for ricin detection, offering bette r sensitivity, larger dynamic range and shorted detection time. The array device has potential to be used for highthroughput protein synthesis and pa rallel functional analysis.

PAGE 80

80 Table 3-1. Summary of met hods for detection of ricin. Research group Methods Labeling Limit of Detection Detection time Reference Ramakrishnan Immunoassay 125I 50 pg/tube 30 h 178 Leith ELISA Avidin/biotin immunoperoxidase 0.2 pg/L 30 h 179 Narang Immunoassay Cy5 0.1 pg/L 1 h 180 Dill Immunoassay Horseradish Peroxidase 0.3 pg/ L 4 h 181 Rubina Immunoassay Cy3 0.1 pg/ L 30 h 182 Yu Immunoassay No label N/A N/A 183 Shyu Immunoassay Gold particles 50 pg/ L 10 min 184 Shankar Immunoassay Alkaline phosphatase 5 pg/ L 30 h 165 Wadkins Immunoassay Cy5 25 pg/ L 1.5 h 163 Stine Quartz crystal microbalance sensor No label 5000 pg/ L 6 h 164

PAGE 81

81 Table 3-2. Summary of IVT system. Expression System Reaction System Protein synthesized Duration hours Yield (mg/mL) Reference E. coli CFCF MS2 coat protein 20 0.1 85 Wheat germ CFCF BMV coat protein 20 0.18 85 E. coli CFCF Calcitonin 40 0.06 85 E. coli CFCF CAT 25 0.1 185 Rabbit reticulocyte CECF CAT 34 0.1 92 Wheat germ CFCF DHFR 24 0.2 92 Wheat germ CFCF DHFR 20 0.917 86 E. coli CECF CAT 14 1.2 88 E. coli CECF CAT 15 1.5 87 E. coli CECF CAT 8 3.5 186 Wheat germ CECF DHFR 60 4 167 Wheat germ CECF Luciferase 60 1.1 167 Wheat germ CECF TMV RNA replicase 72 0.6 167 E. coli CECF GFP 21 4.4 187 Wheat germ CECF DHFR 30 0.8 188 E. coli RTS 500 CAT 24 0.9 169 E. coli RTS 500 DHFR 24 0.2 93 E. coli RTS 500 GFP 24 5.9 189 E. coli RTS 500 Erythropoietin24 0.3-2.2 190 Wheat germ RTS 100 GUS 24 0.05 191 Wheat germ RTS 100 SHR 24 0.794 191 Wheat germ RTS 100 SRC 24 0.95 191 Wheat germ RTS 500 SHR 24 1.07 191 Wheat germ RTS 500 SRC 24 1.34 191

PAGE 82

82 Figure 3-1. Schematic representations of cu rrent formats for IVT system. (a) batch (a microcentrifuge tube); (b) continuous-flo w cell-free (CFCF) system; (c) continuous exchange cell-free (CECF) system. Figure 3-2. Illustrations of an array device I for protein synt hesis. The units are laid out according to the standards of 96-well plates (i.e., 9 mm pitch). The tray chamber is for IVT reactions; the well chamber is concentric with the corresponding tray chamber and functions as a nutrients rese rvoir. (a) Three dimensional view; (b) Picture of both parts; (c) Cro ss-sectional view of a tray ne sted in a well, showing the dialysis membrane and fluidic connection.

PAGE 83

83 Figure 3-3. Comparison of in v itro GFP synthesis between a devi ce and a microcentrifuge tube. (a) Western blot analysis of GFP expressed. The first lane is protein markers (M) and the second lane is the negative control (N); the lanes on the left are for GFP expressed in the device and the numbers on the top indicate the reaction time (hours); and the lanes on the right are for GFP expressed in the tube and the reaction time is also indicated on the top. (b) The production yield of GFP as the function of the expression time. The data for the device are represented by the open circles while those for the tube are by the closed circles. 0 10000 20000 30000 40000 50000 0510152025 (b)Time (hr.) 0 10000 20000 30000 40000 50000 450500550600Wavelength (nm)Fluorescence Signal (A.U.) (a) GFP purchased GFP synthesized background Fluorescence Signal (A.U.) Figure 3-4. (a) Comparison of fluorescence emi ssion spectrum between GFP synthesized and GFP purchased. The excitation wavelength is 488 nm. 5 L of GFP synthesized is mixed with 35 L of pH 7.4 PBS buffer to fill in the 50 L of cuevette. The concentration of GFP purcha sed is 2 g/mL in this spectrum. (b) The fluorescent signal as the function of the folding time.

PAGE 84

84 Figure 3-5. Comparison of in vitro CAT synthesis between a devi ce and a microcentrifuge tube. (a) Western blot analysis of CAT expressed. The first lane is protein markers (M) and the second lane is the negative control (N); the lanes on the left are for CAT expressed in the device and the numbers on th e top indicate the re action time (hours); and the lanes on the right are for CAT expr essed in the tube and the reaction time (hours) is also indi cated on the top. (b) The production yield of CAT as a function of the expression time. The data for the devi ce are represented by the open circles while those for the tube are by the closed circles. Figure 3-6. Production yield of luciferase as a function of the expression time in the device or in a microcentrifuge tube.

PAGE 85

85 0 0.2 0.4 0.6 0.8 1 1.2 051015 Time (h)Signal intensity L=-0.5 mm L=0 L=0.5 mm L=1 mm L=1.5 mm L=2 mm (a) height 0.5 mm (b) Figure 3-7. The effects of the fluid level difference on the prot ein expression yield. (a) Device cartoon illustrating the fluid level difference between the tray and well chambers when the connection depth is fixed at 500 m (b) Luciferase expression yield as the function of the IVT time. The expression yi eld is normalized against the experiment showing the highest luminescence signal.

PAGE 86

86 Figure 3-8 SEM images of polycarbonate membrane with 100 nm pore size. (a) room temperature. (b) 170 C. Figure 3-9. SEM images of polycarbonate me mbrane with 10 nm pore size. (a) room temperature. (b) 170 C.

PAGE 87

87 Figure 3-10. Production yield of luciferase as a function of time in the nanopore membrane devices, in the dialysis membrane or in a microcentrifuge tube. Figure 3-11. Signal amplification as a result of the inhibitory effects of ricin on the production of every copy of protein. Thousands of copies of a protein can be expressed from each copy of DNA when ricin is absent. While ricin is present, the damaged ribosome (indicated with a dent) is unable to produce the corresponding proteins. 0 0.2 0.4 0.6 0.8 1 1.2 024681012Time (hr.)Normalized Expression Yield dialysis membrane 10 nm PC membrane 15 nm PC membrane 100 nm PC membrane tube

PAGE 88

88 Figure 3-12. The inhibitory effects of ricin on the production yield of luciferase. (a) The expression yield of luciferase (indicated by the luminescence) is plotted as a function of the concentration of ricin A chain (solid circles) and heat-denatured ricin A chain (open circles). Experiments were carried out in the device an d the expression time was 4 hours. The error bars ar e the standard deviation ob tained from three repeat experiments. Lines are the best fit of lin ear regression of the e xperimental results. (b) Calibration curve for ricin detection. The signal is the difference between the samples and the controls. 0 4 8 12 01234 0 20 40 60 0102030Luminescence (Arb. Units)Time (min.)device tube Luminescence (Arb. Units)Ricin Concentration (nM)ricin heat-denatured ricin (a) (b) Figure 3-13. Synthesis of lucife rase and detection of ricin in a few minutes. (a) Luciferase production as a function of the expression ti me in a miniaturized device or in a microcentrifuge tube. The expression yi eld is indicated by the luminescence. A circular DNA vector was used. (b) Protein expression yield, indicated by luminescence, decreased with the concentration of ricin A chain when the expression time was fixed at 5 minutes (solid circles) However, the expression yield remained the same when the ricins toxicity was deact ivated by heat denature (open circles).

PAGE 89

89 0.0 5.0 10.0 15.0 0.010.11101001000Luminescence (Arb. Units)Ricin Concentration (nM) ricin B chain Ricin + SHC2H4OH A chain Figure 3-14. Comparison among ricin A chain, B chain, whole ri cin, and ricin treated with 2mercaptoethanol. Luciferase expression yiel d, indicated by luminescence, is plotted with the concentration of each reagent.

PAGE 90

90 CHAPTER 4 FABRICATION OF A MICROFLUIDIC DE VICE AND ITS APPLICATION TO BIOLUMINESCENCE DETECTION 4.1 Introduction As briefly discussed in the first chapter, the purpose of this project is to fabricate an integrated microfluidic system for protein expres sion that can be used for toxin detection. The system also has a potential to be used in appl ications like large scale protein engineering, drug screen, etc. So we are more interested in fa bricating a protein synthe sis and analysis device instead of a protein preparation device. Fo r the latter one, improving the total yield of synthesized protein is one of the most impor tant research focuses and there are several commercialized systems available, like RTS ProteoMaster commercialized by Roche. To fabricate a miniaturized protein synthesis and analysis device, challenges lie in performing ultra-fast, high sensit ive, high throughput and low cost synthesis and assay. There are very limited numbers of reports about protein synthesis in microfluidic devices. Endo et al. fabricated a simple T-shaped mixer to test cell-free tran slation of poly (U).95, 192 Fujii et al developed a PDMS-glass hybrid device with temper ature controller and successfully synthesize GFP and BFP.121 More recently, a compartmentalized cell-free protein expression device was fabricated to generate water-in -oil emulsion for protein synthesis.65 However, no protein with biofunctionality has be en expressed in microfluidic devices in all these papers, including our previous results. The reason is that effective mixing plays an important role for the success of most chemical and biochemical reactions. in the microscale.193201 It is well known that the flows inside the microchannels are predominantly laminar due to low Renolds number (Re) and mixing is typically dominated by diffusion.202, 203 When the mixing relies on molecular diffusion, the total mi xing time can be calculated by the following Equation 4-1:

PAGE 91

91 D d t22 (4-1) where t, d, and D are the time needed to obt ain mixing, the thickness of diffusion path and diffusion coefficient, D, respectively. In orde r to enhance the mixing pe rformance, a variety of mechanisms to effectively mix in microfluidic devices have been studie d and reported. These devices can be categorized as using either active204-210 or passive mixing194, 211-234 methods. Most active micromixers facilitate ra pid mixing by stirring flow, which require external sources, such as ultrasonic,205 magnetic force,208, 210 electro-kinetic field207, 209 and an acoustic wave.204, 206 Although these active micromixers are more effec tive in mixing performance, they are difficult to fabricate, operate and inte grate with microfluidic devices On the other hand, passive micromixers exert no energy input except for th e mechanism provided to drive flows at a constant rate. They usually use channel geomet ry to stir or divide fluids, increasing the interfacial area between fluids a nd shortening the diffusion length. Two methods of passive mixing include multilamination and chaotic advection. Multilayer or multi-stage micromixers are designed to tw ist and split fluids and then to speed fluid mixing in the channel.215, 220, 223, 226, 230, 235 A representative example is a micromixer with a three-dimensional serpentine ch annel developed by Liu et al.211 A serpentine flow channel keeps twisting and stretching the interface between the two fl uids, which is effectively at high flow rate (Re =70). This kind of micromixers is not suita ble for expensive biological samples or reagents because a large volume of liquid is required. In addition, the micros tructures in these mixers are complex, requiring complicated fabrication proce ss such as multi-step lithography or multi-layer stacking. The micromixers based on chaotic ad vection using simple microstructures in a microchannel have been reported by severa l research groups in recently years.212, 213, 218, 219, 228, 229, 236-240 One of most promising examples is the mi xer with the bas-relief structure initially

PAGE 92

92 proposed by Stroock et al.212 In the mixer, helical flows can be produced in a microchannel with patterned grooves over a wide flow rate range (0.01 Re 10) and efficient mixing can be obtained in a short distance. In this chapter, I discussed microfluidic devices consisting of a straight microchannel patterned with herringbone ridge s and a spiral channel. The de vices we fabricated relatively simple fabrication processes are designed for bi oluminescence detection of enzymatic reactions. The PMMA microfluidic devices are replicated by hot embossing with a brass mold, in which the microstructures are milled by a high-precis ion micromilling machine. The bioluminescent luciferase-luciferin reaction is undertaken in the microflow devices to study the effect of the channel structure on intensity of light emission. The detection of limit of bioluminescent assay in the device is determined. In addition, the factors affecting the enzymatic reaction on the intensity of luminescence are investigat ed as the function of the flow rate. 4.2 Experimental 4.2.1 Materials and Reagents PMMA polymer sheet (5 mm th ickness) and cover film (250 m thickness) were obtained from Plexiglas MC (GE Polymershapes, Ne w Orleans, LA). QuantiLum recombinant luciferase, T7 luciferase DNA vector, TNT wheat germ extract system and luciferase assay system were acquired from Promega (Madison, WI), while acetylated bovine serum albumin (BSA) was from Sigma-Aldrich (St. Louis, MO ). Phenolphthalein, sodium hydroxide and Alconox were from Fisher Sc ientific (Atlanta, GA). 4.2.2 Device Design and Fabrication The design of microfluidic de vice is shown in Figure 4-1a. Two devices, called device I and device II, are used in bioluminescent reactio n experiments. Each microfluidic deivce is composed of T-shaped inlets, a st raight channel connected to a sp iral channel, and one outlet.

PAGE 93

93 Device II possesses herringbone ridg es in the straight channel for better mixing whereas device I does not. The detail of herringbone ri dges is illustrated in an exploded view in the figure insert. All straight and spiral channels are 200 m wide and 100 m deep, except where specified otherwise. The herringbone ridge is 50 m wi de and 25 m high. The patterns of the devices are designed using AutoCAD and then milled in a 6.3 mm thick brass (353 engravers brass, McMaster-Carr) with a high-precision microm illing machine (Kern MMP 2522). Micromilling i carried out at 40,000 rpm using 500-, 200and 50 m-diameter milling bits with the corresponding feeding rate of 200 mm/min, 120 mm/min and 15 mm/min using procedures described previously.241 The negative relief of the patt ern is hot embossed into the PMMA sheets using a commercial hydraulic press (Precision Press TS-21-HC, City of Industry, CA). The brass mold and PMMA substrate are put in to a vacuum chamber to remove air at the pressure of 8 kPa. During hot embossing, the mo ld is heated at 150 C and then pressed to PMMA substrate with a force of 1000 lb for 4 min. Following embossing, the chamber is removed from the press and the embossed PMMA s ubstrate was cooled. After the holes are then drilled at the ends of each channel serving as so lution reservoirs for samples into and out of the device, the devices are thoroughl y cleaned with 0.5% Alconox solution and DI water in an ultrasonicated bath and dried with compress air. For the final assembly, the open face of microchannel on PMMA substrate is thermal bonded to a thin PMMA film. That is, the PMMA device and thin film are clamped between two gl ass plates and put in a convection oven at 107 C for 20 min. Figure 4-2 shows a picture of th e assembled plastic device and the convection oven used in our study. 4.2.3 Mixing of Phenolphthalein and NaOH Phenolphthalein is a pH indicator and change s its color from colorless to pink when the value of pH is larger than 8. It is easy to observe the colo r change by the naked eye so the

PAGE 94

94 reaction between phenolphthalein an d NaOH is used to simply demonstrate the performance of the device. 1% phenolphthalein and 0.05 mol/L NaOH are injected into microchannels by a syringe pump (PHD 2000, Harvard Apparatus, Holliston, MA) at the flow rate of 1 L/min. When phenolphthalein and NaOH m eets and reacts inside the channel, the color change are observed and captured by Nikon upright micr oscope equipped with a DXM 1200 color CCD camera. 4.2.4 Simulation Computational simulation of 3D flow field and two-fluid mixing through a microchannel based on Finite Volume Method was carried out in the computational fluid dynamics (CFD) package, CFD-ACE (ESI Group, Huntsville, AL). The depth and width of the channel are set at100 m and 200 m, while the width and height of the ridge are set at 50 m and 25 m. 1.2mm and 6-mm-long channels corresponding to one cy cle of herringbone ridges and five cycles of herringbone ridges are examined. The 3D geom etric models of microchannel and 8-node hexahedral elements are built by CFD-GEOM to achieve a stable solution. The density of mesh elements for the 1.2-mm-long channel is studied and optimized to speed up convergence and produce accurate simulation results. The flow module in CFD-ACE is used to solve the incompressible Navier-Strokes equations. The simulations are run with an assumption of a steady and laminar flow of a Newton ian fluid. Both inlets are assigned with a pressure of 10 Pa. Pure fluid 1 and fluid 2 are in the two inlets; their diffusion co efficient (D) is set at 1.0 10-6 cm2/s. No-slip condition is assumed at all channe l walls and the outlet is set at atmospheric pressure. Conjugated Gradient Squared (CGS) and Preconditioning (Pre) solver are used for velocity and species, and Algebr aic MultiGrid (AMG) solver is us ed for pressure correction. The central differencing scheme (with blending set to 0.1) and second-or der upwind scheme are

PAGE 95

95 adopted for the velocity and concentration calc ulations, respectively. The maximum iteration is 1000, while the convergen ce critical is 10-4 and minimum residual is 10-18. 4.2.5 Bioluminescence Reaction Luciferase, an enzyme found in the firefly, cata lyzes luciferin oxidation in the presence of adenosine triphosphate (ATP), magnesium ion as well as oxygen and generates the light with a peak emission at ~560 nm. A luciferase assay ki t contains optimal concentration of luciferin, magnesium, ATP and coenzyme; appropriate reconstitution of these components forms luciferase assay reagents (LAR). A series of dilution of recombinant luciferase from 14.6 ng/mL to 14.6 g/mL is obtained in 1 x cell culture lysis r eagent (provided by the lu ciferase assay kit) with 1 mg/mL acetylated BSA; they are used to i nvestigate the detection of limit in microfluidic devices. Addition of BSA prevents luciferase fr om nonspecific adsorption to the walls of the device, eliminating possible background signal. In vitro synthesis of luciferase is carried out in wheat germ extr act as previously reported. Briefly, The reaction mix of 50 L for luciferase synthesis was prepared by combining 25 L wheat germ extract, 2 L reaction buffer, 1 L T7 RNA polymerase, 1 L of 1 mM amino acid, 1 L Ribonulease inhibitor and 19 L nuclease-free water containing 1 g of luciferase DNA vector. Incubation was performed in a micro centrifuge tube at 30 C for 1.5 hours. Recombinant luciferase or synt hesized luciferase and LAR are injected into a device by a syringe pump with two 50L glass syringes (SGE Internationa l Pty. Ltd., Austin, Texas) at a constant flow rate. The syringe s are connected to the fused-si lica capillaries through low dead volume luer-to-microtight assembly (P-662, Upc hurch Scientific, Oak Harbor, WA), and the capillaries are glued into reservoirs on microchi p to deliver the reagents into microchannel. Various volumetric flow rates are tested to i nvestigate the effect of the flow rate on the enzymatic catalysis of bioluminescence. A th ermoelectrically cooled (-70C) CCD camera

PAGE 96

96 (Roper Scientific Instru mentation, Trenton, NJ) coupled with an optical lens subsystem (4X Zoom) is used to capture the image of lumine scence along the channel direction. The device is installed on a X-Y stage, so that different region in the device can be imaged as shown in Figure 4-3. The setup is in a black box to reduce the background signal. For each measurement, an image of 1024 x 1024 pixels was collected with an exposure time of 60 s. The quantitative data of each image are analyzed quantitativ ely using the software Winview 32. 4.3 Results and Discussion 4.3.1 Device Fabrication As mentioned above, we exploit a high-pre cision micromilling machine to fabricate a mold, which was then used as the master for rep licating microfluidic device. The approach is simple, with low cost and quick turn-around time. In addition, the mold can be used for a wide range of thermoplastics by hot embossing, compress molding or injection molding. The devices are made from polymethylmethacr ylate (PMMA). The straight channel in device II contains staggered herringbone mixer (S HM) as reported. Alternating cycles of asymmetric herringbone ridges are incorporated to improve th e mixing efficiency because interfacial area and the bulk advection will increase due to the three dimension ridge structures.212 The spiral channel is designed to ob tain enough length for possible long time reaction in the device. The layout of the device is shown in Figur e 4-1. The device consists of two inlet channels joined by a T-junction, a straight channel of 6.6 mm length, a 56-mm-long spiral channel, and an outlet. Se veral sets of T-junctions inlets are designed to enclose numbers of herringbone ridges in the cha nnel. Each set is comprised of 5 cycles of asymmetric herringbone structures, which are com posed of twenty single ridges. Only the first set of inlets numbered 1 and 2, and the fifth set of inlets numbered 9 and 10 are used in this work. The

PAGE 97

97 straight channel in device I is smooth without any feature, functio ning as the baseline control to evaluate the performance of the herringbone structure channel. 4.3.2 Embossing Fidelity As discussed before, hot embossing is an eas y and mass-productive method for replication of the micropatterns. The embossing is achie ved by pressing a brass mold into the PMMA substrate at desired temperature in a vacuum chamber and then the desired microstructures are transferred into the plastic substrate within a fe w minutes. The topology of the plastic device is the negative image of the brass mold so the device has the exactly same pattern as the microfabricated brass mold. The brass mold a nd the embossed PMMA device coated with a thin layer of Au are observed under scanning electr on microscopy (SEM) to examine the detail features. The pictures in Fi gure 4-4a and b show the SEM photography of the brass mold and plastic device with the portion of staggered herri ngbone structure. The plas tic substrate shown in Figure 4-4b is a negative image of Figure 4-4a, i. e., a ridge becomes a channel. The results indicate the microstructures are successfully tran sferred from the mold to the plastic device at a high degree of fidelity. The depth of groove (or the height of ridge) and width are determined and quantified by a non-contact profiler (Nanovea ST 400, Allentown, PA). Typical 3D profiler results are shown in Figure 4-5. The depth and wi dth of the channel are 100 m and 200 m, while the height of the ridge is about 25 m. The replication of th e overall layout in Figure 45a and b shows additional evidence of a high degree of fidelity during the tran sfer from the brass mold to plastic substrate. 4.3.3 Holes Drilling Before final assembly of PMMA substrate and thin film, access holes are drilled at the end of each channel to serve as reservoirs for samp les and wastes. Originally, 1 mm-diameter holes are drilled by a milling machine. The end of the capillaries (OD 360 m) is glued into a hole on

PAGE 98

98 the assembly device as shown in Figure 4-6. It is noted that when two solu tions are injected into the microchannels, the air bubbles are continuousl y produced inside the channel at the first 30 l solution pumping in both device I and II (Figure 4-7), which has an effect on mixing of the solutions. It also can be seen from Figure 4-7b th at a small bubble is always trapped by the ridge structure and it is difficult to re move the bubble from the channel. In some cases, the solution flows from one inlet to the other. This phe nomenon can be explained by the large dead volume formed in the reservoir during the fabrication proc ess. Figure 4-8a provides a clear view of the dead volume in the reservoir when the OD 360 m of capillary sits into a 1 mm-diameter hole. To shrink the dead volume, the small size of holes is created using laser ablation technique. 400 m-diameter of holes were drilled by a UV la ser with 264 nm wavelength. After inserting 360 m OD of capillary into the la ser drilled hole, no dead volume is observed under the microscope as shown in Figure 4-8b. 4.3.4 Reaction of Phenolphthalein and NaOH To evaluate the performance of device I and II, a simple experiment of mixing phenolphthalein and NaOH was first carried out. Two solutions of 1% phenolphthalein and 0.05 mol/L NaOH were pumped into the micr ochannels at the flow rate of 1 L/min. As discussed before, when phenolphthalein meets and reacts w ith NaOH, its color changes from colorless to pink. Therefore, the interface be tween two solutions can be easily observed by the pink color. Figure 4-9 shows the experimental results for the device I and II. In the case of device I, two solutions flow in parallel and the reaction only takes place only at the cen ter of the channel so that the interface is clearly obser ved. The fully mixing is not achieved at the spiral channel only by pure diffusion as indicated in Figure 4-9a. However, figure 4-9b shows the mixing has been completely achieved after the solu tions pass through a set of the he rringbone structure. We can

PAGE 99

99 obviously tell the advantage of device I over de vice II by naked eyes. It should be mentioned that the air bubbles are produced in both devices due to large dead volume of the reservoir. 4.3.5 Effect of Herringbone Ridges As discussed in Section 4.2.4, luciferase and LAR are pumped through inlets 1 and 2 into the microchannel simultaneously. When lucife rase meets with LAR at the T-intersection, catalytical reactions start and ge nerate the luminescence. Theref ore, the mixing efficiency can be inferred from the luminescence detected. Figure 4-10 shows the top view luminescence images of both devices I and II, when the flow rate is at 1 L/min and the concentration of luciferase is 14.6 g/mL. These images are obtained from a cooled CCD camera. Only a part of channel is recorded due to the field view of the camera. In microdevice I without herringbone structures, two solutions flow in parallel in a laminar flow as e xpected. Two solutions interact with each other only in the centerlin e of the entire channel, as sugge sted by the stable interface in Figure 4-10a. The thickness of the luminescent trace slowly increases along the channel. The mixing is not completed even at the end of the spir al channel. In contra st, the interfacial line between two fluids disappeared almost immediat ely after they enter into the region containing herringbone structures, as shown in Figure 4-10b. Two fluids were completely mixed before the end of the region with herringbone ridges. The mixing efficiency is analyzed by calcula ting the intensity of th e luminescence on a line perpendicular to the microchannel by the following formula:94 I = maxwI I yi i (4-2) where I y Ii, w and Imax represent the normalized average intensity, interval between sampling points, intensity at sampling point i, width of the channel and maximum intensity

PAGE 100

100 among the sampling points. The numerator in Equation 4-2, i iI y, defines the area of intensity profile. The normalized average intensity, based on equa tion (2), at different positions downstream for device I and II is plotted in Figure 4-11. Th e luminescence intensity in device I increases along the microchannel from 0.11 at the inlet to 0.3 at the distance of 30 mm. However, the signal of device II increases from 0.76 to 0.99 a nd reach the maximum at 24 mm, which consists of three sets of herringbone patt ern. The average intensity of the device II grows much faster that of the device I, indicati ng that alternate repe tition of the set of herringbone structure improves fluid mixing efficiently. It is shown that the normali zed average intensity I can be represented by the equation:218-220 I=1-exp(-x/ ) (4-3) where x is channel axis and is a required mixing length. It can be seen that I is 0 at the cross of inlets and reaches 1 at cert ain length of the channel. is considered as a required mixing length when the intensity I increase from 0 to 0.632.218-220 The required mixing length of device I and II are 88.18 mm and 5.41 mm, respectively. It is show n that the required mixing length of device II is 16 fold shorter than device I. These results indicate th at device II has better mixing performance and one set of herringbone ridge is enough to improve the mixing efficiency. Several research groups have reported the co mputational work concerning the staggered herringbone mixer, including the geometric effect s on the mixing performance and the velocity profiles in the mixer.213, 228, 236, 238 Here we utilize the CFD simulation results of the concentration distribution in th e microchannel to visualize the process of mixing. Figure 4-12 displays the concentration distri butions along the channe l of both devices, indicating the mixing processes of those two devices ar e totally different. From Figur e 4-12b, it can be seen that the

PAGE 101

101 distortion of the interface in th e channel with herri ngbone structure, where there is no distortion occurred in the straight channel. The simu lation results are highly consistent with the experiment results, as shown in Figure 4-11. 4.3.6 Simulation Results Several research groups have reported the co mputational work concerning the staggered herringbone mixer, including the geometric effect s on the mixing performance and the velocity profiles in the mixer.213, 228, 236, 238 Here we utilize the CFD simulation results of the concentration distribution in the microchannel to visualize the pro cess of mixing. The density of mesh elements for 1.2 mm of channel length is first studied. The typica l dimension of mesh elements are set at 0.833, 1.5625, 3.125, 6.25 and 12.5 m, corresponding to the number of mesh elements at 344926, 256418, 205842, 180554 and 167920. The c oncentration distributions in the cross-section of the ridge cha nnel with different numbers of the mesh element are shown in Figure 4-12. Figure 4-13 indicate s that the computational residue plot is converged for velocity, pressure, and sample concentration when the dimension of mesh el ement is set at 3.125 m. Simulations indicate higher number of the mesh elements could achieve more accurate concentration distribution in th e channel, although a ll the residue plots (d ata are not shown) show the solution is converged. It is known that large number of mesh elements requires longer time to achieve the results, though it may give more accurate solution. For 6 mm long channel, we choose the medium dimension size of 2.5 m, therefore the number of mesh element is about 1299780. Figure 4-14 displays the concentration dist ributions along the channel of both devices, indicating the mixing processes of those two devices are totally different. From Figure 4-14b, it can be seen that the distortion of the interface in the channel with herringbone structure, where there is no distortion occurred in the straight channel. The simulation results are highly consistent with the experiment results, as shown in Figure 4-11.

PAGE 102

102 4.3.7 Detection of Limit The luciferase protein is freque ntly used as a reporter gene for measuring promoter activity or transfection efficiency in biol ogical studies. As a result, we studied the detection limit of luciferase and dynamic range of detection using the current set up. To obtain these results, a series of concentrations of luciferase is pumped into Devices I or II. Both luciferase and LAR are pumped at the constant flow rate of 1 L/min, while LAR concentr ation maintains at the constant. Since the intensity of generated light is proportional to the amount of luciferase in the reaction when luciferin and ATP are in excess,242 we can get a calibratio n curves between the light intensity and luciferase concentration. Figure 4-15 shows the luminescent intensity in the region of spiral channel as a function of lucife rase concentration. The plots display a linear relationship in both devices as ex pected. The error bar of each da ta point indicates the standard deviation that was obtained from three repeat ex periments. The data shoe a dynamic detection range from 0.7 g/mL up to 14.6 g/mL and a correlation coefficient (R2) of 0.997 using microfluidic device 1. However, the detec tion limit for microfluidic device II is 0.0146 g/mL and the dynamic range is up to 14.6 g/mL with R2 of 0.994. The detection limit for luciferase was defined as the concentration of luciferase pr oducing an analytical sign al equal to three-times the standard deviation of the blank bioluminescen ce intensity. The sensitivity for luciferase in the mircodevice II was increased by 48 fold compar ed to that in device I. Therefore, the sensitivity for luciferase obtained in the microf luidic device with the herringbone structure can be used to detection of in vitro synthesized protein in microfluidic device. 4.3.8 Effect of Flow Rate The dependence of the luminescence intensity on the flow rate was investigated in device I and II. Reagents are introduced in inlets 9 a nd 10 and only one set of herringbone feature is contained in the channel. The c oncentration of luciferase is 14.6 g/mL. Figure 4-16 shows the

PAGE 103

103 CCD image of the spiral channel at the various flow rates. Calculated normalized intensity was plotted as the function of location for both devices as shown in Figure 4-17. For the range of flow rates from 1 to 8 L/min, the signal intensity of device II is much higher than that of device I, which confirms again the herringbone structure in the channel providing much more efficiency of mixing. Based on Equation 4-3, the requi red mixing lengths of device II are 5.75, 6.97, 9.67 and 12.79 mm, and those of device I are 77.94, 97.18, 116.78 and 176.40 mm, corresponding to the flow rates of 1, 2, 4 and 8 L/min. These results indicate that the required mixing length increases as flow rate increases. No surprising, mixing intensity in the device I decreases with increasing flow rate due to the reduction in the molecular diffusion of the react ants. However, luminescence intensity in the microreator II reduces with the increase of the flow rate, which could be explained the light generated from the luciferase assa y reaction is the rate-determining. In such type of enzymatic reaction in the micro scale, mixi ng and reaction occur simultaneous ly rather than consecutively so that reactions may be limited by insufficient mixing.243 DeLuca and MaElroy presented evidence indicating there are two distinct rate-l imiting steps that occur after rapid mixing of ATP, luciferin and luciferase befo re maximum light emission reaches.242 After mixing, there is a lag of 25 msec at room temperature before any light is emitted and then the rate of light emission rises slowly to the maximum in about 0.3 sec. Therefore, the decrease of the luminescence intensity can probably be attributab le to a reduction in the residence time of the reactants, since the increase in the flow rate co uld be responsible for the decrea se in the residence time of the reactants. 4.3.9 Detection of Synthesized Luciferase In vitro synthesis of luciferase is carried out using wheat germ extract in the microcentrifuge tube as reported previously. S ynthesized luciferase a nd LAR are pumped into

PAGE 104

104 the inlets 9 and 10 using both device s I and II at the flow rate of 1 L/min. Detection of the expression product is achieved by monitoring the intensity of luminescence after mixing the product with LAR. Figure 4-18 shows the e xperimental results of online detection of synthesized luciferase in device II. The amount of synthesized lu ciferase in a microcentrifuge tube is about 0.0146 ng/ L according to the detected signal an d the calibration curve in Figure 416. However, the sample can not be detect ed using device I. It is suggested in vitro luciferase synthesis could be carried out in the device II. 4.4 Conclusion We have developed a microfluidic device co nsisting of the straight channel with the herringbone ridge at the bottom wall and the spiral channel for luminescence detection. The PMMA microfluidic device was hot embossed from the brass mold, which was fabricated with a high-precision micromilling machine. Therefore, th e fabrication of PMMA devices can be easily implemented with low cost. The efficiency for the luminescence detection of luciferase in the microfluidic device was relied on the microstructu re of the channel within the devices, indicating the device with herringbone featur e has a better mixing performance. It may be noted that luminescence intensity decreased as the flow rate increased for both devices. In addition, the device II can be used to detect in vitro synthesized luciferase. A de vice with a similar feature may be used for cell-free protein synthesis and onlin e detection.

PAGE 105

105 Figure 4-1. (a) Schematic diagrams of two type s of microfluidic devi ces; Top: Microfluidic device I; Bottom: Microfluidic device II. Both devices consist of multiple T-shaped inlets, a straight channel, a spiral channel for reactions and one outlet. Device II possesses herringbone ridges in the straight channel whereas devi ce I does not. The depth and width of all channels are 100 m and 200 m, except where specified otherwise. The pattern and geometry of the staggered herringbone groove. The groove is 50 m wide and 30 m deep. (b) Picture of a plastic microfluidic device for luciferase detection. Channels and wells are in the plastic substrate, which is thermal bonded with a thin film.

PAGE 106

106 Figure 4-2. (a) The picture of pre-assembly device for thermal bonding. (b) A convection oven. Figure 4-3. Photograph of experiment al setup for bioluminescent reaction. CCD device stage lens pump syringes

PAGE 107

107 Figure 4-4. SEM pictures of a portion of the ch annel with the herringbon e structure. (a) The brass mold. (b) A representative PMMA substr ate. The length of the scaling bars is 1 mm in pictures and 200 m in the exploded view. Figure 4-5. Typical 3-D profile r results. (a) The brass mold. (b) A representative PMMA substrate.

PAGE 108

108 Figure 4-6. The picture of th e integrated plastic device. Figure 4-7. The images of air bubbles produced insi de the channel. (a) Device I. (b) Device II Figure 4-8. The images of a capillary with OD 360 m inserting into a ho le. (a) A mechanically drilled 1 mm-diameter hole. (a ) A laser drilled hole of 400 m.

PAGE 109

109 Figure 4-9. Mixing of phenolphthalein and NaOH at the flow rate of 1 L/min at the indicated locations. (a) Microfluidic device I and (b) Microfluidic device II.

PAGE 110

110 Figure 4-10. CCD camera image of the lumi nescence at the indicated location in the microfluidic devices. (a) Microfluidic device I; (b) Microfuidic device II.

PAGE 111

111 0 20 40 60 80 100 0 0.2 0.4 0.6 0.8 1 downchannel length (mm)normailized intensity Figure 4 11. Normalized luminescence intensity ch anges along the channel direction within microfluidic device I (solid circle) and micr ofluidic device II (ope n circle) with flow rate of 1 L/min. Symbols: experimental data and trend line w ith I=1-exp(-x/ ).

PAGE 112

112 Figure 4-12. CFD results of concentration distribu tions in the ridge region for different number of the mesh elements: (a) 167910; (b) 180554; (c) 205842; (d) 256418; (e) 344826. Figure 4-13. The computational residue plot. (b) (a) (c) (d) (e)

PAGE 113

113 Figure 4-14. CFD results of concentration distri butions and the interface between two solutions at the flow rate of 1 L/min. (a) Microfluidic device I; (b) Microfluidic device.

PAGE 114

114 0 500 1000 1500 2000 2500 3000 0246810121416 Concentration ( g/mL)Signal Intensity Figure 4-15. Calibration curves for bioluminesce nt assay using microflu idic device I (solid circles) and microfluidic device II (open circles). Figure 4-16. CCD image of the luminescence at the spiral channel at the flow rate of 1 L/min, 2 L/min, 4 L/min and 8 L/min from left to right. (a ) Microfluidic device I; (b) Microfuidic device II. (a) (b)

PAGE 115

115 0 10 20 30 40 50 60 0 0.2 0.4 0.6 0.8 1 Channel length (mm)Normalized intensity Figure 4-17. Effect of the flow rate on the lu minescent intensity along the channel direction in microfluidic device I (open) and microfluid ic device II (solid) fo r flow rate of 1 L/min (circle), 2 L/min(rectangle), 4 L/min (triangle) and 8 L/min(pentagram). Figure 4-18. Detection of synthesi zed luciferase using device II.

PAGE 116

116 CHAPTER 5 SUMMARY AND FUTURE DIRECTION 5.1 Summary The objective of this research is to fabricat e and develop the miniat urized devices for high yield and high-throughput protein production and its downstream applications. The combination of microfabricat ion technology and in vitro protein synthesis provides a state-of-the-art platform for producing a large number of proteins at low co st. Moreover, such a platform can be used for simple, fast and sensitive toxin detection. The background information about fabrication an d applications of mini aturized devices is first reviewed in Chapter 1. The methods for large scale protein synthesis from in vivo to in vitro are described. The development and challenges of in vitro protein synthesis in microchips are discussed. Chapter 2 presents a microwell array device ma de from acrylic by micromilling. The array device has been demonstrated for in vitro synthesis of GFP, CAT and lu ciferase in prokaryotic or eukaryotic expression system. The device is then used to de monstrate the concept of toxin detection based on the mechanism that toxin i nhibits biological protein synthesis, while tetracycline (TC) and cycloheximide (CH) are select ed as toxin simulants. It is confirmed that TC has an inhibitory effect on the expression of GFP and CAT in the prokaryotic expression system, whereas CH has a negligible effect. In contrast, TC has a negligible effect on the production of luciferase in the e ukaryotic expression system, while CH has an inhibitory effect. In addition, the response pattern of TC and CH in a 3 X 4 array microwell device shows the potential of toxin detection based on the mechanisms of toxin actions. The major shortcoming of the microwell arra y device is the limited reaction time, thus leading to a low yield of synthe sized proteins. For example, expression of CAT and GFP ceased

PAGE 117

117 after 4 h and 1.5 h, respectively. To addresses th is challenge, a nested-well device is designed and it incorporates continuous ex change cell free (CECF) technol ogy as discussed in chapter 3. The nested-well device consists of a tray and a well separated by a dialysis membrane. In this device, the reaction byproducts (Pi and NMPs) are removed, and the consumable substrates (amino acids, NTPs, and energy-regenerating co mpounds) are continuously supplied through the porous membranes. As a result, the nested-well device allows protein expression to continue for up to 20 h with a high protein yield, because th e accumulation of inhibitory byproducts and the exhaustion of substrates are avoi ded. The performances of prot ein synthesis in a tube and the device are compared. In addition, the use of th e device for ricin detection is demonstrated. Compared with traditional la bor-intensive and time-consumi ng ELISA method, this detection method is simpler and faster. The detection limit of 1 pg/L of ricin is at least comparable to the reported ELISA method and at least one order of magnitude more sensitive than most other reported detection methods. Chapter 4 focuses on development of the microfluidic devices based on hot embossing technique and its application for enzymatic reacti ons. The microfludic device is composed of the straight channel with herringbone ridges at the bottom wall and a spiral channel for luminescence detection. The advantages of using herringbon e ridge mixer for biological reactions are emphasized. Bioluminescence produced by the enzymatic reaction between luciferase and luciferin is detected by a cooled CCD camera, wh ich allows us to visualize fluid behavior. Diffusion-based mixing and chaotic mixing in th e microfluidic device are investigated and compared. In addition, the device is demonstrated to detect luci ferase synthesized off chip so that the device may be uesd for online in vitro protein synthesis and detection.

PAGE 118

118 5.2 Future Work As summarized in the previous section, miniaturized protein expression systems have been successfully established and their potential applica tions have also been demonstrated in this dissertation. In this section, the futu re work is proposed and discussed. 5.2.1 Device Design and Fabrication A two-part, nested-well array is described in the dissertation for protein synthesis. The upper part, or tray, functions as the reaction chamber for protein synthesis; each well has a dialysis membrane on the bottom to hold DNA te mplate, cell extract or lysate, RNA, and synthesized proteins. The lower part, well, contains the feeding solution, including amino acid and energy source. As described in Chapter 3, the lower parts are milled in the lab. Alternatively, a commercially available 96-well plat e can be used as a substitute. This will allow the use of commercial well-plate readers to m easure the protein expres sion yield directly. Although the current design is designed in agreem ent with 96-well plate format, the device with a smaller number of wells is not compatible with commercial plate readers. Well size and alignments corresponding to 384-well plates coul d also theoretically be developed, though the issues related to injection of tray must be addressed. The nested-well device consists of two parts, which are separated by a dialysis membrane. Dialysis membrane is glued to tray, which leads to several limitations. First, the glue needs to be carefully chosen and tested for biocompatibility. Second, it is difficult to ensure the uniformity of glue thickness since it depends on the size of a microwell and contact area of the membrane. Third, leakage is always possible. Finally, it is di fficult to automate fabrication of device. These limitations may be addressed by laminating a na noporous membrane with the tray. Moreover, the membranes must meet the following criteria: (1) the pore size is large enough for nutrient supplies and byproduct removal and small enough to keep the synthesized proteins; (2) the

PAGE 119

119 membrane is compatible with lamination proce ss. The pore structures must be maintained during the sealing. Nanoporous PC membranes w ith the pore size of 10 nm and 15 nm are under testing. 5.2.2 Protein Expression System In this dissertation, we employ E. coli extract and wheat germ extract to demonstrate prokaryotic and eukaryotic protei n expression system. Theoretica lly, any organism that makes a large amount of protei ns can be used for in vitro protein expression, but the discussion on the future work is still focused on commercial systems above. Many different DNA templates may be designed and tested. Successful in vitro protein expression depends on a correct starting DNA temp late, which consists of a gene encoding a target protein and the necessary regulatory elements including promoter, terminator, ribosome binding site and tags. Although GFP, CAT a nd luciferase have been expressed from corresponding vectors provided by kit, a variety of different T7 driv en plasmids, such as pIVEX, pET, and pEXP for in vitro protein expression are availabl e and DNA coding sequence can be inserted into these plasmids by standard cloning procedures. An alternative method is to use PCR to ge nerate linear DNA template to avoid timeconsuming and labor-intensive steps of plasmid preparation. The advantages of this method include: (1) easy deletion and inse rtion of sequences; (2 ) easy introduction of point mutations and random mutations; (3) easily and fast fuse of different proteins. Th e disadvantage of this method is that the yield of prot ein synthesis is lower than usi ng plasmid. The low yield may be caused either by the fast decay of linear DNA temp lates by the exonuclease presented in the cellfree extracts or by the limited acces sibility of transcription and translation factors due to the secondary structure in DNA or RNA. In chapter 2, we tried to use a PCR amplified template for protein expression, but failed to obtain detectable amount of pr otein. We suspect that the

PAGE 120

120 sequence of mRNA preceding the translation f actor binding site may form the secondary structure and the bindin g sites are embedded. 5.2.3 Protein Detection In this work, three proteins, CAT, GFP and luci ferase, were successfu lly synthesized in the miniaturized devices. After expression, the synt hesized protein was detected and quantified by one or more methods. CAT and GFP were qua ntified by Western blotting. GFP was also detected by fluorescence whereas lu ciferase is detected by lumine scence. Western blotting is the most commonly-used method for protein detect ion. However, it takes long time and it is difficult to be integrated with the miniaturized device. Fluorescence dete ction based on GFP is a good approach, since GFP is a relatively small pr otein and many proteins have been successfully fused with GFP. Compared with fluorescence, luminescence based on luciferase requires no excitation light source, simplifying the design an d fabrication of the microfluidic device. Imaging setup discussed in chapter 5 can be used to detect synthesized luciferase or extended for other targeted proteins, if the prot ein can be fused with luciferase. 5.2.4 Applications In this dissertation, the feasibility of using in vitro protein synthesis sy stem for detecting inhibitory effects of different toxin simulants/drugs has been demonstrated. Other toxin simulants can be tested and the response pattern (or signature) database of the array device can then be constructed. The database may be used as a tool for detection and identification of known and unknown toxins. Other potential applications of in vitro pr otein expression array may include (1) high throughput cloning and screening of mutant and engineered prot eins; (2) construction of the protein library from cDNA library for func tional genomic and proteomic research.

PAGE 121

121 Inspired by Feynmans talk There is plenty of room at the bottom an invitation to enter a new field of physics, I have acco mplished this research. I would lik e to end this thesis with an invitation, there is plenty of chemistry and biology in miniatur ized devices. The door is open; scientists and engineers from di fferent disciplinary are welcome to explore such a marvelous small world.

PAGE 122

122 LIST OF REFERENCES (1) Feynman, R. Journal of Microelect romechanical systems 1992 1, 60-66. (2) Feynman, R. Journal of Microelect romechanical systems 1993 2, 4-14. (3) Dittrich, P. S.; Manz, A. Nat Rev Drug Discov 2006 5, 210-218. (4) Auroux, P. A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal Chem 2002 74, 2637-2652. (5) Vilkner, T.; Janasek, D.; Manz, A. Anal Chem 2004 76, 3373-3385. (6) Gad-el-Hak, M. The MEMS Handbook, 2nd edition ed.; CRC Press: Boca Raton, 2002. (7) Hsu, T.-R. MEMS and Microsystems: Design and Manufacture; McGraw Hill: Boston, 2002. (8) Rai-Choudhury, P. Handbook of Microlithography, Micromachining, and Microfabrication., 1st edition ed.; SPIE: Bellingham, 1997. (9) Terry, S. C.; Jerman, J. H.; Angell, J. B. IEEE Transactions on Electron Devices 1979 ED-26, 1880-1886. (10) Manz, A.; Miyahara, Y.; Miura, J. ; Watanabe, Y.; Miyagi, H.; Sato, K. Sensors and Actuators, B: Chemical 1990 B1, 249-255. (11) van Lintel, H. T. G. Sensors and Actuators 1988 15, 153-167. (12) Smits, J. G. Sensors and Actuators A: Physical 1990 21, 203-206. (13) Esashi, M.; Shoji, S.; Nakano, A. Sensors and Actuators 1989 20, 163-169. (14) Shoji, S.; Nakagawa, S.; Esashi, M. Sensors and Actuators A: Physical 1990 21, 189-192. (15) van de Pol, F. C. M.; van Lintel, H. T. G.; Elwenspoek, M.; Fluitman, J. H. J.; van Lintel, H. T. G.; Elwenspoek, M.; Fluitman, J. H. J. Sensors and Actuators A: Physical 1990 21, 198-202. (16) van de Pol, F. C. M.; Wonnink, D. G. J.; Elwenspoek, M.; Fluitman, J. H. J. Sensors and Actuators 1990 17, 139-143. (17) Madou, M. J. Fundamentals of Microfabrication: The Science of Miniaturization, 2nd edition ed.; CRC Press: Boca Raton, 2002. (18) Rai-Choudhury, P. Handbook of Microlithography, Micromachining, and Microfabrication, 1st edition ed.; SP IE: Bellingham, 1997. (19) Sadia National Laboratories. Microsystems, http:// www.sadia.gov/mstc/index.html, November 1st, 2007

PAGE 123

123 (20) NUXUS III Report. Market Analysis fo r MEMS and Microsystems. http:// www.utcconsult.de/english/r_n3_e.html, April 8th, 2006 (21) Manz, A.; Graber, N.; Widmer, H. M. Sensors and Actuators, B: Chemical 1990 B1, 244248. (22) Kovacs, G. T. Micromachined Transducers Sourcebook, 1st edition ed.; McGraw-Hill Science: Boston, 1998. (23) Wise, K. D.; Najafi, K. Science 1991 254, 1335-1342. (24) Kopp, M. U.; Mello, A. J.; Manz, A. Science 1998 280, 1046-1048. (25) Quake, S. R.; Scherer, A. Science 2000 290, 1536-1540. (26) Madou, M.; Florkey, J. Chem Rev 2000 100, 2679-2692. (27) Boone, T. D.; Fan, Z. H.; Hooper, H. H.; Ricco, A. J.; Tan, H.; Williams, S. J. Anal Chem 2002 74, 78A-86A. (28) Soper, S. A.; Ford, S. M.; Qi, S.; McCarley, R. L.; Kelly, K.; Murphy, M. C. Anal Chem 2000 72, 642A-651A. (29) Xia, Y.; Whitesides, G. M. Annual Review of Materials Science 1998 28, 153-184. (30) Whitesides, G. M.; Ostuni, E.; Ta kayama, S.; Jiang, X.; Ingber, D. E. Annu Rev Biomed Eng 2001 3, 335-373. (31) Peppas, N. A. Hydrogels in Medicine and Pharmacy, 1st edition ed.; CRC Press: Boca Raton, 1986. (32) Beebe, D. J.; Moore, J. S.; Bauer, J. M.; Yu, Q.; Liu, R. H.; Devadoss, C.; Jo, B. H. Nature 2000 404, 588-590. (33) Vogel, S. D., Kathryn K. Cats' Paws and Catapults: Mechanical Worlds of Nature and People; W. W. Norton & Company: Norton, 2000. (34) Fan, Z. H.; Harrison, D. J. Analytical Chemistry 1994 66, 177-184. (35) Sia, S. K.; Whitesides, G. M. Electrophoresis 2003 24, 3563-3576. (36) Stoyanov, A. V.; Das, C.; Fredrickson, C. K.; Fan, Z. H. Electrophoresis 2005 26, 473479. (37) Vreeland, W. N.; Locascio, L. E. Anal Chem 2003 75, 6906-6911. (38) Lai, S.; Wang, S.; Luo, J.; Lee, L. J.; Yang, S. T.; Madou, M. J. Anal Chem 2004 76, 1832-1837.

PAGE 124

124 (39) Tan, W.; Fan, Z. H.; Qiu, C. X.; Ricco, A. J.; Gibbons, I. Electrophoresis 2002 23, 36383645. (40) Holden, M. A.; Jung, S. Y.; Cremer, P. S. Anal Chem 2004 76, 1838-1843. (41) Paegel, B. M.; Blazej, R. G.; Mathies, R. A. Curr Opin Biotechnol 2003 14, 42-50. (42) Song, S.; Singh, A. K.; Shepodd, T. J.; Kirby, B. J. Anal Chem 2004 76, 2367-2373. (43) Huber, D. L.; Manginell, R. P.; Samara, M. A.; Kim, B. I.; Bunker, B. C. Science 2003 301, 352-354. (44) Hong, J. W.; Studer, V.; Hang, G.; Anderson, W. F.; Quake, S. R. Nat Biotechnol 2004 22, 435-439. (45) Schmalzing, D.; Belenky, A.; Novotny, M. A.; Koutny, L.; Salas-Solano, O.; El-Difrawy, S.; Adourian, A.; Matsudaira, P.; Ehrlich, D. Nucleic Acids Res 2000 28, E43. (46) Harrison, D. J.; Fluri, K.; Seiler, K. ; Fan, Z. H.; Effenhauser, C. S.; Manz, A. Science 1993 261, 895-897. (47) Harrison, D. J.; Manz, A.; Fan, Z. H.; Ludi, H.; Widmer, H. M. Analytical Chemistry 1992 64, 1926-1932. (48) Fan, Z. H.; Mangru, S.; Granzow, R.; Heaney, P.; Ho, W.; Dong, Q.; Kumar, R. Anal Chem 1999 71, 4851-4859. (49) Li, Y.; Buch, J. S.; Rosenberger, F.; DeVoe, D. L.; Lee, C. S. Anal Chem 2004 76, 742748. (50) Woolley, A. T.; Mathies, R. A. Proc Natl Acad Sci U S A 1994 91, 11348-11352. (51) Burns, M. A.; Mastrangelo, C. H.; Sammarc o, T. S.; Man, F. P.; Webster, J. R.; Johnsons, B. N.; Foerster, B.; Jones, D.; Fields, Y.; Kaiser, A. R.; Burke, D. T. Proc Natl Acad Sci U S A 1996 93, 5556-5561. (52) Woolley, A. T.; Hadley, D.; Landre, P.; deMe llo, A. J.; Mathies, R. A.; Northrup, M. A. Anal Chem 1996 68, 4081-4086. (53) Woolley, A. T.; Mathies, R. A. Analytical Chemistry 1995 67, 3676-3670. (54) Aborn, J. H.; El-Difrawy, S. A.; Novotny, M.; Gismondi, E. A.; Lam, R.; Matsudaira, P.; Mckenna, B. K.; O'Neil, T.; Streechon, P.; Ehrlich, D. J. Lab on a Chip 2005 5, 669-674. (55) Kakuta, M.; Bessoth, F. G.; Manz, A. The Chemical Record 2001 1, 395-405. (56) Janickea, M. T.; Kestenbauma, H.; Hagendor fa, U.; Schutha, F.; Fichtnerb, M.; Schubertb K. Journal of Catalysis 2000 191, 282-293.

PAGE 125

125 (57) Losey, M. W.; Schmidt, M. A.; Jensen, K. F. Industrial & Engineering Chemistry Research 2001 40, 2555-2562. (58) Chambers, R. D.; Spink, R. C. H. Chemical Communications 1999 883-884. (59) De Bellefon, C.; Tanchoux, N.; Caraviei lhes, S.; Grenoullet, P.; Hessel, V. Angewandte Chemie, International Edition 2000 39, 3442-3445. (60) Hisamoto, H.; Saito, T.; Tokesh i, M.; Hibara, A.; Kitamori, T. Chemical Communications 2001 2662-2663. (61) Ismagilov, R. F.; Ng, J. M.; Kenis, P. J.; Whitesides, G. M. Anal Chem 2001 73, 52075213. (62) Kikutani, Y.; Horiuchi, T.; Uchiyama, K.; Hisamoto, H.; Tokeshi, M.; Kitamori, T. Lab Chip 2002 2, 188-192. (63) Neils, C.; Tyree, Z.; Finlayson, B.; Folch, A. Lab Chip 2004 4, 342-350. (64) DiTursi, M. K.; Cha, J.; Newman, M. R.; Dordick, J. S. Biotechnol Prog 2004 20, 17051709. (65) Dittrich, P. S.; Jahnz, M.; Schwille, P. Chembiochem 2005 6, 811-814. (66) He, M.; Taussig, M. J. Nucleic Acids Res 2001 29, E73-73. (67) Ramachandran, N.; Hainsworth, E.; Bhulla r, B.; Eisenstein, S.; Rosen, B.; Lau, A. Y.; Walter, J. C.; LaBaer, J. Science 2004 305, 86-90. (68) Angenendt, P.; Nyarsik, L.; Szaflarski, W. ; Glokler, J.; Nierhaus, K. H.; Lehrach, H.; Cahill, D. J.; Lueking, A. Anal Chem 2004 76, 1844-1849. (69) Steffen, J.; von Nickisch -Rosenegk, M.; Bier, F. F. Lab Chip 2005 5, 665-668. (70) de Figueiredo, P.; Roberts, R. L.; Nester, E. W. Proteomics 2004 4, 3128-3140. (71) Zahn, J. D.; Hsieh, Y. C.; Yang, M. Diabetes Technol Ther 2005 7, 536-545. (72) Grayson, A. C. R.; Shawgo, R. S.; Johnson, A. M.; Flynn, N. T.; Li, Y.; Cima, M. J.; Langer, R. Proceedings of The IEEE 2004 92, 6-21. (73) Ziaie, B.; Baldi, A.; Lei, M.; Gu, Y.; Siegel, R. A. Adv Drug Deliv Rev 2004 56, 145-172. (74) Alberts, B.; Johnson, A.; Lewis, J.; Raff, M.; Roberts, K.; Walter, P. Molecular Biology of the Cell, 4 Edition ed.; Garland Science: New York, 2002. (75) Crick, F. Nature 1970 227, 561-563. (76) Temin, H. M.; Baltimore, D. Adv Virus Res 1972 17, 129-186.

PAGE 126

126 (77) Mulder, C.; Arrand, J. R.; De lius, H.; Keller, W.; Pettersson, U. ; Roberts, R. J.; Sharp, P. A. Cold Spring Harb Symp Quant Biol 1975 39 Pt 1, 397-400. (78) Cech, T. R. Nature 1989 339, 507-508. (79) Cech, T. R. Biochem Int 1989 18, 7-14. (80) Littlefield, J. W.; Keller, E. B.; Gross, J.; Zamecnik, P. C. J Biol Chem 1955 217, 111123. (81) Lamborg, M. R.; Zamecnik, P. C. Biochim Biophys Acta 1960 42, 206-211. (82) Nirenberg, M. W.; Matthaei, J. H. Proc Natl Acad Sci U S A 1961 47, 1588-1602. (83) Roberts, B. E.; Paterson, B. M. Proc Natl Acad Sci U S A 1973 70, 2330-2334. (84) Pelham, H. R.; Jackson, R. J. Eur J Biochem 1976 67, 247-256. (85) Spirin, A. S.; Baranov, V. I.; Rya bova, L. A.; Ovodov, S. Y.; Alakhov, Y. B. Science 1988 242, 1162-1164. (86) Endo, Y.; Otsuzuki, S.; Ito, K.; Miura, K. J Biotechnol 1992 25, 221-230. (87) Kigawa, T.; Yokoyama, S. J Biochem (Tokyo) 1991 110, 166-168. (88) Kim, D. M.; Choi, C. Y. Biotechnol Prog 1996 12, 645-649. (89) Kim, D. M.; Kigawa, T.; Choi, C. Y.; Yokoyama, S. Eur J Biochem 1996 239, 881-886. (90) Yamane, T.; Kawarasaki, Y.; Nakano, H. Ann New York Acad Sci 1995 750, 146-157. (91) Alakhov, Y. B.; Baranov, V. I.; Ovodov, S. J.; Ryabova, L. A.; Spirin, A. S.; Morozov, I. Y.; United States Patent: USA, 1995. (92) Baranov, V. I.; Spirin, A. S. Methods Enzymol 1993 217, 123-142. (93) Betton, J. M. Curr Protein Pept Sci 2003 4, 73-80. (94) Hosokawa, K.; Fujii, T.; Nojima, T.; Shoji, S.; Yotsumoto, A.; Endo, I. Journal of Micromechatronics 2000 1, 85-98. (95) Nojima, T.; Fujii, T.; Hosokawa, K.; Yotsumoto, A.; Shoji, S.; Endo, I. Bioprocess Engineering 2000 22, 13-17. (96) Yamamoto, T.; Nojima, T.; Fujii, T. Lab on a Chip 2002 2, 197-202. (97) Tabuchi, M.; Hino, M.; Shinohara, Y.; Baba, Y. Proteomics 2002 2, 430-435.

PAGE 127

127 (98) Kinpara, T.; Mizuno, R.; Murakami, Y.; Koba yashi, M.; Yamaura, S.; Hasan, Q.; Morita, Y.; Nakano, H.; Yamane, T.; Tamiya, E. J Biochem (Tokyo) 2004 136, 149-154. (99) Mei, Q.; Fredrickson, C. K.; Jin, S.; Fan, Z. H. Anal Chem 2005 77, 5494-5500. (100) Hahn, G. H.; Asthana, A. ; Kim, D. M.; Kim, D. P. Anal Biochem 2007 365, 280-282. (101) Kigawa, T.; Yabuki, T.; Yoshida, Y.; Tsut sui, M.; Ito, Y.; Shibata, T.; Yokoyama, S. FEBS Lett 1999 442, 15-19. (102) Spirin, A. S. Trends Biotechnol 2004 22, 538-545. (103) Gilbert, M.; Albala, J. S. Current Opinion in Chemical Biology 2002 6, 102-105. (104) Jewett, M. C.; Swartz, J. R. Biotechnol Bioeng 2004 87, 465-472. (105) Katzen, F.; Chang, G.; Kudlicki, W. Trends Biotechnol 2005 23, 150-156. (106) Gold, L. M.; Schweiger, M. Proc Natl Acad Sci U S A 1969 62, 892-898. (107) Gold, L. M.; Schweiger, M. Methods Enzymol 1971 20, 537-542. (108) Lederman, M.; Zubay, G. Biochim Biophys Acta 1967 149, 253-258. (109) Schweiger, M.; Gold, L. M. Proc Natl Acad Sci U S A 1969 63, 1351-1358. (110) Zubay, G. Annu Rev Genet 1973 7, 267-287. (111) Zubay, G.; Lederman, M.; DeVries, J. K. Proc Natl Acad Sci U S A 1967 58, 1669-1675. (112) DeVries, J. K.; Zubay, G. Proc Natl Acad Sci U S A 1967 57, 1010-1012. (113) Shimizu, Y.; Inoue, A.; Tomari, Y.; Suzuki T.; Yokogawa, T.; Nishikawa, K.; Ueda, T. Nat Biotechnol 2001 19, 751-755. (114) Ohashi, H.; Shimizu, Y.; Ying, B. W.; Ueda, T. Biochem Biophys Res Commun 2007 352, 270-276. (115) Oleinikov, A. V.; Gray, M. D.; Zhao, J.; M ontgomery, D. D.; Ghindills, A. L.; Dill, K. Journal of Proteome Research 2003 2, 313-319. (116) Elbaz, Y.; Steiner-Mordoch, S.; Danieli, T.; Schuldiner, S. Proc. Natl. Acad. Sci. USA 2004 101, 1519-1524. (117) Tao, S. C.; Zhu, H. Nat Biotechnol 2006 24, 1253-1254. (118) Tian, J.; Gong, H.; Sheng, N.; Zhou, X.; Gulari, E.; Gao, X.; Church, G. Nature 2004 432, 1050-1054.

PAGE 128

128 (119) Sawasaki, T.; Hasegawa, Y.; Tsuchimochi, M.; Kamura, N.; Ogasawara, T.; Kuroita, T.; Endo, Y. FEBS Lett 2002 514, 102-105. (120) Jung, G. Y.; Stephanopoulos, G. Science 2004 304, 428-431. (121) Yamamoto, T.; Fujii, T.; Nojima, T. Lab Chip 2002 2, 197-202. (122) Spirin, A. S. In In cell-free translation systems.; Spirin, A. S., Ed.; Springer: Berlin, 2002, pp 3-22. (123) Bujard, H.; Gentz, R.; Lanzer, M.; Stueber, D.; Mueller, M.; Ibrahimi, I.; Haeuptle, M. T.; Dobberstein, B. Methods Enzymol 1987 155, 416-433. (124) Coen, D. M.; Bedbrook, J. R.; Bogorad, L.; Rich, A. Proc Natl Acad Sci U S A 1977 74, 5487-5491. (125) Lewis, J. B.; Anderson, C. W.; Atkins, J. F.; Gesteland, R. F. Cold Spring Harb Symp Quant Biol 1975 39 Pt 1, 581-590. (126) Stueber, D.; Ibrahimi, I.; Cutle r, D.; Dobberstein, B.; Bujard, H. Embo J 1984 3, 31433148. (127) Spirin, A. S. In Frontiers in Bioprocessiong II; Todd, P., K., S. S., Beer, M., Eds.; Washington,DC: ACS, 1991, pp 31-43. (128) Craig, D.; Howell, M. T.; Gibbs, C. L.; Hunt, T.; Jackson, R. J. Nucleic Acids Res 1992 20, 4987-4995. (129) Jermutus, L.; Ryabova, L. A.; Pluckthun, A. Curr Opin Biotechnol 1998 9, 534-548. (130) Endo, Y.; Sawasaki, T. J Struct Funct Genomics 2004 5, 45-57. (131) Jackson, A. M.; Boutell, J.; Cooley, N.; He, M. Brief Funct Genomic Proteomic 2004 2, 308-319. (132) ImageJ. http://rsb.info.nih.gov/ij, July 18th, 2007 (133) Tsien, R. Y. Annu Rev Biochem 1998 67, 509-544. (134) Prendergast, F. G. In Green fluorescent proteins; Sullivan, K. F., Kay, S. A., Eds.; Academic press: San Diego, 1999, pp 1-18. (135) Bucdavari, S. The Merck Index; Merck &Co.: Whitehouse Station,NJ, 1996. (136) Watson, J. D.; Baker, T. A.; Bell, S. P.; Gann, A.; Kevine, M.; Losich, R. Molecular biology of the gene, 12th edition ed.; CSHL press: Cold Spring Harbor,NJ, 2004. (137) Salyers, A. A.; Whitt, D. D. Bacterial pathogenesis: a molecular approach, 2nd edition ed.; ASM press: Washington, DC, 2002.

PAGE 129

129 (138) Terpe, K. Appl Microbiol Biotechnol 2003 60, 523-533. (139) Buchel, C.; Morris, E.; Orlova, E.; Barber, J. J Mol Biol 2001 312, 371-379. (140) Mei, Q.; Fredrickson, C. K.; Lian, W.; Jin, S.; Fan, Z. H. Anal Chem 2006 78, 7659-7664. (141) Mei, Q.; Fredrickson, C. K. ; Simon, A.; Knouf, R.; Fan, Z. H. Biotechnology Progress 2007 23. (142) Moreno-Bondi, M. C.; Alarie, J. P.; Vo-Dinh, T. Anal Bioanal Chem 2003 375, 120-124. (143) Phillips, K. S.; Cheng, Q. Anal Chem 2005 77, 327-334. (144) Liu, W.; Montana, V.; Chapman, E. R.; Mohideen, U.; Parpura, V. Proc Natl Acad Sci U S A 2003 100, 13621-13625. (145) Taitt, C. R.; Anderson, G. P.; Lingerfelt, B. M.; Feldstein, M. J.; Ligler, F. S. Anal Chem 2002 74, 6114-6120. (146) Tao, L.; Yu, X.; Snyder, A. P.; Li, L. Anal Chem 2004 76, 6609-6617. (147) Mothershed, E. A.; Cassiday, P. K.; Pierson, K.; Mayer, L. W.; Popovic, T. J Clin Microbiol 2002 40, 4713-4719. (148) Lagally, E. T.; Emrich, C. A.; Mathies, R. A. Lab Chip 2001 1, 102-107. (149) Koh, C. G.; Tan, W.; Zhao, M. Q.; Ricco, A. J.; Fan, Z. H. Anal Chem 2003 75, 45914598. (150) Belgrader, P.; Young, S.; Yuan, B.; Prim eau, M.; Christel, L. A.; Pourahmadi, F.; Northrup, M. A. Anal Chem 2001 73, 286-289. (151) CDC, http://www.bt.cdc.gov/agen t/agentlist-category.asp, March 16th, 2005 (152) Teter, K.; Holmes, R. K. Infect Immun 2002 70, 6172-6179. (153) Crowther, J. R. ELISA : theory and practice; Humana Press: Totowa, N.J., 1995. (154) Olsnes, S.; Refsnes, K.; Pihl, A. Nature 1974 249, 627-631. (155) Lord, J. M.; Roberts, L. M.; Robertus, J. D. Faseb J 1994 8, 201-208. (156) Lord, M. J.; Jolliffe, N. A.; Marsden, C. J.; Pateman, C. S.; Smith, D. C.; Spooner, R. A.; Watson, P. D.; Roberts, L. M. Toxicol Rev 2003 22, 53-64. (157) Vitetta, E. S.; Thorpe, P. E.; Uhr, J. W. Trends in Pharmacological Sciences 1993 14, 148-154. (158) Uckun, F. M.; Frankel, A. Leukemia 1993 7, 341-348.

PAGE 130

130 (159) Franz, D. R.; Jaax, N. K. In Medical Aspects of Chemic al and Biological Warfare; Sidell, F. R., Takafuji, E. T., Franz, D. R., Eds.; Borden Institute, Walter Reed Army Medical Center: Washington, D.C., 1997, pp 632. (160) Bigalke, H.; Rummel, A. Toxicology 2005 214, 210-220. (161) Poli, M. A.; Rivera, V. R.; Hewetson, J. F.; Merrill, G. A. Toxicon 1994 32, 1371-1377. (162) Godal, A.; Olsnes, S.; Pihl, A. Journal of Toxicology and Environmental Health 1981 8, 409-417. (163) Wadkins, R. M.; Golden, J. P.; Pritsiolas, L. M.; Ligler, F. S. Biosensors & Bioelectronics 1998 13, 407-415. (164) Stine, R.; Pishko, M. V. Analytical Chemistry 2005 77, 2882-2888. (165) Shankar, K.; Zeng, K.; Ruan, C.; Grimes, C. A. Sensors and Actuators B 2005 107, 640648. (166) Alakhov, Y. B. B., V. I.; Ovodov, S. J.; Ryabova, L. A.; Spirin, A.S.; Morozov,I. Y.; United States Patent: USA, 1995. (167) Madin, K.; Sawasaki, T.; Ogasawara, T.; Endo, Y. Proc Natl Acad Sci U S A 2000 97, 559-564. (168) Yamamoto, Y.; Sugimoto, S.; Shen, X. C.; Nagamune, T.; Yao, S. L.; Suzuki, E. Biochemical Engineering Journal 1999 3, 151-155. (169) Martin, G. A.; Kawaguchi, R.; Lam, Y. ; DeGiovanni, A.; Fukushima, M.; Mutter, W. Biotechniques 2001 31, 948-950, 952-943. (170) Angenendt, P.; Nyarsik, L.; Szaflarski, W. ; Glokler, J.; Nierhaus, K. H.; Lehrach, H.; Cahill, D. J.; Lueking, A. Anal. Chem. 2004 76, 1844-1849. (171) Kolb, V. A.; Makeyev, E. V.; Spirin, A. S. Embo J 1994 13, 3631-3637. (172) Salyers, A. A.; Whitt, D. D. Bacterial pathogenesis: a molecular approach, 2nd ed.; ASM Press: Washington, D.C., 2002. (173) Endo, Y.; Mitsui, K.; Motizuki, M.; Tsurugi, K. J Biol Chem 1987 262, 5908-5912. (174) Endo, Y.; Tsurugi, K. J Biol Chem 1987 262, 8128-8130. (175) Harley, S. M.; Beevers, H. Proceedings of the National Acad emy of Sciences of the United States of America-Biological Sciences 1982 79, 5935-5938. (176) Harley, S. M.; Beevers, H. Plant Science Letters 1984 36, 1-5.

PAGE 131

131 (177) Cawley, D. B.; Hedblom, M. L.; Houston, L. L. Arch Biochem Biophys 1978 190, 744755. (178) Ramakrishnan, S.; Eagle, M. R.; Houston, L. L. Biochimica Et Biophysica Acta 1982 719, 341-348. (179) Leith, A. G.; Griffiths, G. D.; Green, M. A. Journal of the Forensic Science Society 1988 28, 227-236. (180) Narang, U.; Anderson, G. P.; Ligler, F. S.; Burans, J. Biosensors and Bioelectronics 1997 12, 937-945. (181) Dill, K.; Montgomery, D. D.; Ghindilis, A. L.; Schwarzkopf, K. R.; Ragsdale, S. R.; Oleinikov, A. V. Biosensors and bioelctronics 2004 20, 736-742. (182) Rubina, A. Y.; Dyukova, V. I.; Dementieva E. I.; Stomakhin, A. A.; Nesmeyanov, V. A.; Grishin, E. V.; Zasedatelev, A. S. Anal Biochem 2005 340, 317-329. (183) Yu, X.; Zhao, L.; Jiang, H.; Wang, H.; Yin, C.; Zhu, S. Sensors and Actuators B 2001 76, 199-202. (184) Shyu, R. H.; Shyu, H. F.; Liu, H. W.; Tang, S. S. Toxicon 2002 40, 255-258. (185) Kudlicki, W.; Kramer, G.; Hardesty, B. Anal Biochem 1992 206, 389-393. (186) Kigawa, T.; Yabuki, T.; Yokoyama, S. Tanpakushitsu Kakusan Koso 1999 44, 598-605. (187) Klammt, C.; Lohr, F.; Schafer, B.; Haase, W.; Dotsch, V.; Ruterjans, H.; Glaubitz, C.; Bernhard, F. Eur J Biochem 2004 271, 568-580. (188) Shirokov, V. A. S., P. N.; Biryukov, S. V.; Spirin, A. S. In In cell-free translation; Spirin, A. S., Ed.; Spiringer: Berlin, 2002, pp 91-107. (189) Budisa, N. In In cell-free protei expression; Swartz, J. R., Ed.; Springer: Berlin, 2003, pp 89-98. (190) Nemetz, C. W., S.; Krupka, S.; Watzele, M; Mutter, W. In In cell-free translation systems; Spirin, A. S., Ed.; Springer: Berlin, 2000, pp 157-163. (191) Madin, K. W., Manfred; Hoffmann, Michae l; Metzler, Thomas; Rode, Hans-Jrgen; Schmid, Markus; Mayr, Dora; Nikolaus, Thomas; Schlossman, Thilo; Obermeier, Wolfgang; Reichhuber, Rolf ; Buchberger, Bernd Biochemica 2005 33-36. (192) HosoKawa, K.; Fujii, T.; Nojima, T.; Shoji, S.; Yotsumoto, A.; Endo, I. Journal of Micromechatronics 2000 1, 85-98. (193) Service, R. F. Science 1998 282, 400-400.

PAGE 132

132 (194) Schwesinger, N.; Frank, T.; Wurmus, H. Journal of Micromechanics and Microengineering 1996 6, 99-102. (195) Liu, R. H.; Lenigk, R.; Grodzinski, P. Journal of Microlithography Microfabrication and Microsystems 2003 2, 178-184. (196) Burns, M. A.; Johnson, B. N.; Brahmasandra, S. N.; Handique, K.; Webster, J. R.; Krishnan, M.; Sammarco, T. S.; Man, P. M.; J ones, D.; Heldsinger, D.; Mastrangelo, C. H.; Burke, D. T. Science 1998 282, 484-487. (197) Kricka, L. J.; Wilding, P. Anal Bioanal Chem 2003 377, 820-825. (198) Bilitewski, U.; Genrich, M.; Kadow, S.; Mersal, G. Anal Bioanal Chem 2003 377, 556569. (199) Golbig, K.; Kursawe, A.; Hohmann, M.; Taghavi-Moghadam, S.; Schwalbe, T. Chemical Engineering Communications 2005 192, 620-629. (200) Liu, R. H.; Lenigk, R.; Druyor-Sanchez, R. L.; Yang, J. N.; Grodzinski, P. Analytical Chemistry 2003 75, 1911-1917. (201) Schwalbe, T.; Autze, V.; Wille, G. Chimia 2002 56, 636-646. (202) Ehrfeld, W.; Golbig, K.; Hessel, V.; Lowe, H.; Richter, T. Industrial & Engineering Chemistry Research 1999 38, 1075-1082. (203) Benz, K.; Jackel, K. P.; Regenauer, K. J.; Schiewe, J.; Drese, K.; Ehrfeld, W.; Hessel, V.; Lowe, H. Chemical Engineering & Technology 2001 24, 11-17. (204) Yang, Z.; Goto, H.; Matsumoto, M.; Maeda, R. Electrophoresis 2000 21, 116-119. (205) Yang, Z.; Matsumoto, S.; Goto, H.; Matsumoto, M.; Maeda, R. Sensors and Actuators aPhysical 2001 93, 266-272. (206) Zhu, X.; Kim, E. S. Sensors and Actuators a-Physical 1998 66, 355-360. (207) Wu, H. Y.; Liu, C. H. Sensors and Actuators a-Physical 2005 118, 107-115. (208) Lu, L. H.; Ryu, K. S.; Liu, C. Journal of Microelectromechanical Systems 2002 11, 462469. (209) Oddy, M. H.; Santiago, J. G.; Mikkelsen, J. C. Anal Chem 2001 73, 5822-5832. (210) Lemoff, A. V.; Lee, A. P. Sensors and Actuators B-Chemical 2000 63, 178-185. (211) Liu, R. H.; Stremler, M. A.; Sharp, K. V. ; Olsen, M. G.; Santiago, J. G.; Adrian, R. J.; Aref, H.; Beebe, D. J. Journal of Microelect romechanical Systems 2000 9, 190-197.

PAGE 133

133 (212) Stroock, A. D.; Dertinger, S. K. W.; Ajdari A.; Mezic, I.; Stone, H. A.; Whitesides, G. M. Science 2002 295, 647-651. (213) Stroock, A. D.; Dertinger, S. K.; Whitesides, G. M.; Ajdari, A. Analytical Chemistry 2002 74, 5306-5312. (214) Park, S. J.; Kim, J. K.; Park, J.; Chung, S.; Chung, C.; Chang, J. K. Journal of Micromechanics and Microengineering 2004 14, 6-14. (215) Bessoth, F. G.; deMello, A. J.; Manz, A. Analytical Communications 1999 36, 213-215. (216) Jen, C. P.; Wu, C. Y.; Lin, Y. C.; Wu, C. Y. Lab on a Chip 2003 3, 77-81. (217) Jeon, M. K.; Kim, J. H.; Noh, J.; Kim, S. H.; Park, H. G.; Woo, S. I. Journal of Micromechanics and Microengineering 2005 15, 346-350. (218) Kim, D. S.; Lee, I. H.; Kwon, T. H.; Cho, D. W. Journal of Micromechanics and Microengineering 2004 14, 1294-1301. (219) Kim, D. S.; Lee, S. W.; Kwon, T. H.; Lee, S. S. Journal of Micromechanics and Microengineering 2004 14, 798-805. (220) Kim, D. S.; Lee, S. H.; Kwon, T. H.; Ahn, C. H. Lab on a Chip 2005 5, 739-747. (221) Kim, D. J.; Oh, H. J.; Park, T. H.; Choo, J. B.; Lee, S. H. Analyst 2005 130, 293-298. (222) Kirner, T.; Albert, J.; Gunther, M.; Mayer, G.; Rei nhackel, K.; Kohler, J. M. Chemical Engineering Journal 2004 101, 65-74. (223) Lee, N. Y.; Yamada, M.; Seki, M. Analytical and Bioanalytical Chemistry 2005 383, 776782. (224) Lee, Y. K.; Shih, C.; Tabeling, P.; Ho, C. M. Journal of Fluid Mechanics 2007 575, 425448. (225) Mae, K.; Maki, T.; Hasegawa, I.; Eto, U.; Mizutani, Y.; Honda, N. Chemical Engineering Journal 2004 101, 31-38. (226) Melin, J.; Gimenez, G.; Roxhed, N. ; van der Wijngaart, W.; Stemme, G. Lab on a Chip 2004 4, 214-219. (227) Chung, Y. C.; Hsu, Y. L.; Jen, C. P.; Lu, M. C.; Lin, Y. C. Lab on a Chip 2004 4, 70-77. (228) Hassell, D. G.; Zimmerman, W. B. Chemical Engineering Science 2006 61, 2977-2985. (229) Howell, P. B.; Mott, D. R.; Fertig, S.; Kapla n, C. R.; Golden, J. P.; Oran, E. S.; Ligler, F. S. Lab on a Chip 2005 5, 524-530. (230) Hong, C. C.; Choi, J. W.; Ahn, C. H. Lab on a Chip 2004 4, 109-113.

PAGE 134

134 (231) Johnson, T. J.; Ross, D.; Locascio, L. E. Anal Chem 2002 74, 45-51. (232) Chew, Y. T.; Xia, H. M.; Shu, C.; Wan, S. Y. M. Modern Physics Letters B 2005 19, 1567-1570. (233) Xia, H. M.; Wan, S. Y. M.; Shu, C.; Chew, Y. T. Lab on a Chip 2005 5, 748-755. (234) Xia, H. M.; Shu, C.; Wan, S. Y. M.; Chew, Y. T. Journal of Micromechanics and Microengineering 2006 16, 53-61. (235) Vijayendran, R. A.; Kathleen, M. M.; Beebe, D. J.; Leckband, D. E. Langmuir 2003 19, 1824-1828. (236) Yang, J. T.; Huang, K. J.; Lin, Y. C. Lab on a Chip 2005 5, 1140-1147. (237) Fu, X.; Liu, S. F.; Ruan, X. D.; Yang, H. Y. Sensors and Actuators B-Chemical 2006 114, 618-624. (238) Aubin, J.; Fletcher, D. F.; Bertrand, J.; Xuereb, C. Chemical Engineering & Technology 2003 26, 1262-1270. (239) Wang, H. Z.; Iovenitti, P.; Harvey, E.; Masood, S. Journal of Micromechanics and Microengineering 2003 13, 801-808. (240) Lynn, N. S.; Dandy, D. S. Lab on a Chip 2007 7, 580-587. (241) Hupert, M. L.; Guy, W. J. ; Llopis, S. D.; Shadpour, H.; Rani, S.; Nikitopoulos, D. E.; Soper, S. A. Microfluidics and Nanofluidics 2007 3, 1-11. (242) DeLuca, M.; McElroy, W. D. Biochemistry 1974 13, 921-925. (243) David, R.; Lintz, H. G. Chemie Ingenieur Technik 1984 56, 104-110.

PAGE 135

135 BIOGRAPHICAL SKETCH Qian Mei was born in 1977 in China. She was ra ised in Luoyang, a beautif ul city in central China, known as the city of peonies. She ente red Southeast University in Nanjing, China, in 1996, and earned her Bachelor of Science in biomed ical engineering four years later. She then proceeded to her graduate studies in the same uni versity and she received a Master of Science in biomedical engineering in 2003. In the same year, she enrolled in biomedical engineering at the Health Science Center of theUniversity of Te nnessee at Memphis. In the spring of 2004, she relocated to Florida and started to pursue her Ph.D. in mechanical engineering in University of Florida.