Main page

Contents

Introduction

Sequencing Resources

Systems Biology

Analysis Results

Bioinformatics

Environmental Genomics

Technology Development

ELSI

Author Index

Institution Index

Pdf versions:
faster download (339 kb) or
higher resolution (610 kb)

Return to HGPI

Technology Development

57

Developing a Lox-Based Recombinatorial Cloning System for Ligand Libraries

Robert Siegel[1], Nileena Velappan[2], Peter Pavlik[2], Leslie Chasteen[2], and Andrew Bradbury[2] (amb@lanl.gov)

[1]Pacific Northwest National Laboratory and [2]Los Alamos National Laboratory

The selection of binding ligands (e.g. single chain Fvs - scFvs) against protein targets can be done using a number of different systems, including phage, phagemid, bacterial or yeast display vectors. Genetic selection methods have also been developed based on yeast two hybrid and enzyme complementation systems. In general, selection vectors are not suitable for subsequent production. Furthermore, once scFvs have been selected, they can be usefully modified by cloning into other destination vectors (e.g. by adding dimerization domains, detection domains, eukaryotic expression in eukaryotic vectors etc.). However, this is relatively time consuming, and requires checking of each individual construct after cloning. An alternative to cloning involves the use of recombination signals to shuttle scFvs from one vector to another. These have the advantage that DNA restriction and purification can be avoided. Such systems have been commercialized in two general systems: Gateway™, uses lambda att based recombination signals, while Echo™ uses a single lox based system to integrate a source plasmid completely into a host plasmid.

We have examined the potential for using heterologous lox sites and cre recombinase for this purpose. Five apparently heterologous lox sites (wild type, 511, 2372, 5171 and fas) have been described. A GFP/lacZ based assay to determine which of these were able to recombine with each other was designed and implemented. Of the five, three (2372, 511 and wt) were identified which recombined with one another at levels less than 2%.

To use recombination as a cloning system, it is important to be able to select against host vectors which do not contain the insert of interest. Two toxic genes were examined for this purpose. The tetracycline gene confers sensitivity to nickel, while the sacB gene confers sensitivity to sucrose. We confirmed these sensitivities, although found that some antibiotic resistances interfere with survival of bacteria hosting non-tetracycline containing plasmids.

In preliminary experiments we have demonstrated that recombination from one plasmid to another, using 2272 and wild type lox sites and sacB or tetracycline, can occur in vivo at very high efficiency. This opens the possibility of using this system to easily transfer scFvs after selection to other plasmids. However, the utility of this system is not limited to scFvs - any DNA fragment (gene, open reading frame, promoter etc.) can easily be shuttled from one plasmid to another using these lox based signals.



58

Towards High-Throughput Selection of Binding Ligands

Milan Ovecka, Nileena Velappan, Leslie Chasteen, Peter Pavlik, and Andrew Bradbury (amb@lanl.gov)

Los Alamos National Laboratory

Phage display libraries represent a relatively easy way to generate binding ligands against a vast number of different targets. Although in principle, phage display selection should be amenable to automation, this has not yet been described and present selection protocols are far from high throughput. We have examined the selection process in a systems approach and attempted to automate each individual step. Selection is carried out in the microtiter format using 24 targets as the individual selection lot size. Output is plated onto large assay trays, and a program to pick colonies in specific orders corresponding to the selection arrangement is in the process of being developed for the Qbot picking robot. High density dot blots (400 clones in the footprint of 4 "96 well" wells) as a first round clone testing method is in the process of being developed, while ELISA as final confirmation has been completely automated in the 384 well format for the Tecan Genesis workstation.



59

Fluorobodies: Fluorescent Binding Ligands for Genomic Studies

Ahmet Zeytun, Geoff Waldo, and Andrew Bradbury (amb@lanl.gov)

Los Alamos National Laboratory

Antibodies are the most widely used binding ligands in research, and recent molecular diversity techniques (e.g. phage display) permit the generation of antibody fragments, comprising the binding domain alone, without the use of animals. As a result, much hope has been placed in the idea of creating genome-wide panels of antibodies selected with high throughput procedures. Such antibodies could be used in the high throughput genome wide study of gene products by immunofluorescence, immunoprecipitation and western blotting, as well as novel applications such as antibody chips and intracellular inhibition studies. While we are in the process of automating the selection of antibodies against different targets, and feel this is feasible, we have noted that antibody fragments suffer from a number of problems, foremost among these is the inability to detect binding without the use of secondary enzymatically, or fluorescently, labeled reagents. Other problems include relatively poor expression levels in bacteria and poor stability.

The use of GFP as a binding scaffold, rather than antibodies, would resolve many of these problems. However, due to destabilization of GFP folding upon the insertion of extraneous sequences, attempts to engineer standard GFP have been unsuccessful to date. We have overcome these problems with two essential modifications. The first involves the use of a novel form of GFP (superfolder GFP) which is far more stable than traditional GFP. The second modification involves the use of CDRs from antibodies as diversity elements, rather than random peptides encoded by oligonucleotides. We have created a small library of 5e6 fluorobody clones and displayed them on phage. From this small library we selected specific binders for a number of different targets. These fluorobodies bind their targets specifically as shown by band shift assays, immunofluorescence, ELISAs and crude microarrays. The affinities are similar to those expected from antibody libraries of similar sizes, and they can be expressed at very high levels (100mg/L) compared to antibodies. We have also shown that loss of fluorescence is associated with loss of binding function, permiting easy monitoring of their use. They hold tremendous potential in genomics, proteomics, diagnostics and drug screening.



60

Microbioreactor Arrays with Parametric Control for High-Throughput

Michel Maharbiz, William Holtz, Afshan Shaik, Roger Howe, and Jay D. Keasling (keasling@socrates.berkeley.edu)

Departments of Electical Engineering and Computer Sciences and of Chemical Engineering, University of California, Berkeley, CA 94720

We present a scalable array technology for parametric control of high-throughput cell cultivations. The technology makes use of commercial printed circuit board (PCB) technology, integrated circuit sensors, and an electrochemical gas generation system. We present results and for an array of eight 250 µl microbioreactors. Each bioreactor contains an independently addressable suite that provides closed-loop temperature control, generates feed gas electrochemically, and continuously monitors optical density. The PCB technology allows for the assembly of additional off-the-shelf components into the microbioreactor array; we demonstrate the use of a commercial ISFET chip to continuously monitor culture pH. The electrochemical dosing system provides a powerful paradigm for gas delivery to high-density arrays of microreactors. We show growth data for Escherichia coli cultured in the array with varying microaerobic conditions using electrochemically generated oxygen. Additionally, we present data on carbon dioxide generation and pH control.



61

Selective Genotyping of Individual Cells by Capillary Polymerase Chain Reaction

Hanlin Li and Edward S. Yeung (yeung@ameslab.gov)

Ames Laboratory, Iowa State University, Ames, IA 50011

On-line capillary polymerase chain reaction (PCR) coupled with laser-induced fluorescence detection was successfully demonstrated for individual human cells. A single 50-mm i.d. fused-silica capillary served both as the reaction vessel and for isolating single cells. SYBR Green I dye was added into the reaction mixture for dynamic fluorescence labeling. Because of the small i.d. of the capillary, PCR-amplified DNA fragments from single cells were localized in the capillary, providing discrete product zones with concentrations at readily detectable levels. With selective primer design, only cells containing the DNA of interest were amplified. By counting the number of peaks in the capillary via electromigration past a detection window, the number of targeted cell templates could be determined. Identification of the 295-bp fragment beta-actin gene from individual human lymphoblast cell was demonstrated. Independent on-column cell counting provided positive correlation between the starting cell templates and the final PCR products. This opens up the possibility of highly selective and sensitive disease diagnosis at an early stage, when only a few cells in the population are defective.

[an error occurred while processing this directive]