Preface The ability to express recombinant proteins in eukaryotic cells has greatly increased our ability to study protein function in a variety of settings. Initial work in this field focused on gaining an understanding of the cis-acting sequences necessary for mRNA expression, processing, and translation. As the field has matured, the more sophisticated task of finely regulating gene expression has taken center stage. Researchers want to be able to limit expression to particular tissues and to determine the timing and level of expression. In addition, exciting new technologies are being developed that use small RNA molecules to regulate expression of endogenous genes. In this volume of Methods in Enzymology, we have brought together a number of exciting new techniques for regulated gene expression in eukaryotic cells. We hope that this volume will prove to be a valuable resource for those who are looking for new methods to express recombinant proteins in eukaryotic systems. We realize that no single volume can be complete, and certainly there are methods for recombinant protein expression not represented here. However, we have tried to bring together many novel and creative methods currently being used in many top laboratories. We thank the authors for the time and effort they devoted to describing the current methods for recombinant protein expression in use in their labo- ratories. JOSEPH C. GLORIOSO MARTIN C. SCHMIDT xiii [ l ] MICROARRAY-BASED EXPRESSION MONITORING 3 [ 11 Fluorescence-Based Expression Monitoring Using Microarrays By ELIZABETH A. WINZELER, MARK SCHENA, and RONALD W. DAVIS Introduction The amount of DNA sequence available to researchers, along with the need to access this great wealth of information experimentally, has been increasing exponentially. Miniature nucleic acid arrays, often called "DNA chips" or "microarrays," offer opportunities to collect much of the same data that can be obtained with standard molecular biology hybridization methods but in a highly parallel fashion. These microarrays contain dense collections of nucleic acids [either polymerase chain reaction (PCR) prod- ucts or oligonucleotides] that are either synthesized or deposited at fixed spatial locations on specially prepared glass slides. When labeled DNA or RNA samples are hybridized to the microarrays, the abundance of specific target sequences in the sample can be estimated based on the observed signal intensity at the physical position of the complementary probe or probes. Pico- to femtomoles of thousands of different nucleic acid probes can be arranged at densities of 400 to 250,000 elements/cm 2. With this miniaturization and through the use of fluorescence the simultaneous analy- sis of entire genomes becomes possible. Although microarrays show great promise as tools for genotyping, map- ping, and resequencing, 1,2 an equally important application for the microar- ray is measuring transcript abundance. 3,4 Microarrays have been used to simultaneously measure the mRNA expression levels for every gene in Saccharomyces cerevisiae under several different growth conditions, 5-v to characterize the differences between normal and metastatic tissues, 8 and M. Chee, R. Yang, E. Hubbell, A. Berno, X. C. Huang, D. Stern, J. Winkler, D. J. Lockhart, M. S. Morris, and S. P. Fodor, Science 274, 610 (1996). 2 j. G. Hacia, L. C. Brody, M. S. Chee, S. P. Fodor, and F. S. Collins, Nature Gener 14, 441 (1996). 3 M. Schena, D. Shalon, R. W. Davis, and P. O. Brown, Science 270, 467 (1995). 4 D. J. Lockhart, H. Dong, M. C. Byrne, K. T. Follettie, M. V. Gallo, M. S. Chee, M. Mittmann, C. Wang, M. Kobayashi, H. Horton, and E. L. Brown, Nature Biotechnol. 14, 1675 (1996). 5 L. Wodicka, H. Dong, M. Mittmann, M H. Ho, and D. J. Lockhart, Nature Biotechnol. 15, 1359 (1997). 6 j. L. DeRisi, V. R. Iyer, and P. O. Brown, Science 278, 680 (1997). 7 R. J. Cho, M. J. Campbell, E. A. Winzeler, L. Steinmetz, A. Conway, L. Wodicka, T. G. Wolfsberg, A. E. Gabrielian, D. Landsman, D. J. Lockhart, and R. W. Davis, Mol. Cell (1998). 8 j. DeRisi, L. Penland, P. O. Brown, M. L. Bittner, P. S. Meltzer, M. Ray, Y. Chen, Y. A. Su, and J. M. Trent, Nature Genet. 14, 457 (1996). Copyright © 1999 by Academic Press All rights of reproduction in any form reserved. METHODS IN ENZYMOLOGY, VOL. 306 0076-6879/99 $30.00 4 ANALYSIS OF GENE EXPRESSION [ 1 ] as a vehicle for gene discovery. 9,1° They can also be used to probe genome content when DNA instead of RNA or cDNA is hybridized. 5,11 Global gene expression profiles will be used in diagnostics and to look for drug targets. 12 Data generated by array-based expression experiments will be essential for understanding genetic regulatory networks and integrated cellular responses. As the availability of fluorescence-based microarray technology grows, more applications for these tools will be discovered and new methods for generating microarrays will be developed. The different types, their differ- ent manufacturing processes, and different applications have been reviewed extensively elsewhere. 13-15 This article describes the two basic types of microarrays, protocols for fluorescently labeling messenger RNA from eu- karyotic cells for hybridization to microarrays, and considerations involved in experimental design and data analysis. Nomenclature The term DNA microarray (sometimes called cDNA microarray) will be used to describe ordered collections of plasmid clones or DNA fragments that have been attached to glass at densities of greater than 100 probe elements/cm 2. Oligonucleotide microarrays will be distinguished from DNA microarrays, not because their manufacture is necessarily different, but because hybridization conditions are generally different. Oligonucleotide or DNA arrays may also be fabricated on porous materials, such as nitrocel- lulose or nylon membranes, but the spot density is generally lower. Al- though these filter arrays can also be used to monitor expression using a large number of probes, the target is usually radiolabeled and discussion of such is outside the scope of this article. In addition, the word "target" will describe the molecule whose characteristics are unknown, in this case the messenger RNA or cDNA in solution, and the word "probe" will describe the molecules that are affixed to the microarray. Unlike conven- tional Southern or Northern methods, the target, not the probe, bears the label. 9 M. Schena, D. Shalon, R. Heller, A. Chai, P. O. Brown, and R. W. Davis, Proc. Natl. Acad, Sci. U.S.A. 93, 10614 (1996). 10 R. A. Heller, M. Schena, A. Chai, D. Shalon, T. Bedilion, J. Gilmore, D. E. Woolley, and R. W. Davis, Proc. Natl. Acad. Sci. U.S.A. 94, 2150 (1997). 11 D. Lashkari, J. DeRisi, J. McCusker, A. Namath, C. Gentile, S. Hwang, P. Brown, and R. Davis, Proc. Natl. Acad. Sci. U.S.A. 94, 13057 (1997). 12 L. H. Hartwell, P. Szankasi, C. J. Roberts, A. W. Murray, and S. H. Friend, Science 278, 1064 (1997). 13 A. Marshall and J. Hodgson, Nature Biotechnol. 16, 27 (1998). 14 M. Johnston, Curt. Biol. 8, R171 (1998). 15 G. Ramsay, Nature Biotechnol. 16, 40 (1998). [ 1 ] MICROARRAY-BASED EXPRESSION MONITORING 5 To construct DNA microarrays, nucleic acid probes (usually 500 to 2000 bases in size) are generated by PCR-amplifying plasmid library inserts (using primers complementary to the vector portion of the library) or portions of genomic DNA (using primers designed specifically for the open reading frames or genes of interest). The PCR products are then deposited, usually using a robotic microspotting device, at defined locations on a glass slide. 16 The robotic spotting device is equipped with a set of tips that pick up small amounts of PCR products (~0.2 tzl) from a microtiter plate by capillary action and then sequentially dispense the product at specific loca- tions on multiple slides by touching the tips to the slide surface in a way that is equivalent to the mechanism by which ink is released from a quill pen when the tip touches paper. When fine spotting pins are used, spots containing as little as 1 ng of nucleic acid can be arrayed at densities of up to 5000 elements/cm 2. Because each PCR reaction may produce a few micrograms of probe, thousands of microarrays may be produced from a single set of PCR reactions. Other robotic microfabrication technologies. such as ink jetting, in which the probe solution is placed in a piezoelectric controlled print head that can release microscopic droplets, are under devel- opment. ,7 After the PCR products are spotted on the microarray, they are fixed covalently to the surface and denatured. Fluorescently labeled mRNA, cDNA, or cRNA is hybridized to the microarray. The fluorescence intensity at each location on the slide is determined with a confocal microscope that has been modified to scan the slide surface or with a charge-coupled device (CCD) camera. 18 Detection by mass spectrometry or radioisotope is also possible, but will not be considered here. The scanned image is then ana- lyzed. An example of a scanned image of a DNA microarray is shown in Fig. 1 (see color insert). Messenger RNA was prepared from human Jurkat cells and hybridized to an array containing 248 human and Arabidopsis cDNAs using the procedures described here. As the amount of probe on the array exceeds the amount of target, the observed signal at any given position is a good estimate of the abundance of cognate target in the sample. Oligonucleotide Microarrays Oligonucleotide microarrays consist of shorter (less than 50 bases) nu- cleic acid fragments of known sequence, covalently attached to a derivatized glass slide at defined locations. The oligonucleotides can be either prefabri- cated and then attached using microdeposition technology (robotic printing 16 D. Shalon, S. J. Smith, and P. O. Brown, Genome Res. 6, 639 (1996). 17 A. M. Castellino, Genome Res. 7, 943 (1997). 18 M. Schena and R. Davis, in "PCR Methods Manual" (M. Innis, D. Gelfand, and J. Sninsky, eds.), p. 445. Academic Press, San Diego, 1999. 6 ANALYSIS OF GENE EXPRESSION [ 1 ] or ink jetting) or synthesized in situ using techniques such as photolithogra- phy or ink jetting of the individual dA, dC, dG, or dT phosphoramidite monomersJ 7 Affymetrix (Santa Clara, CA) manufactures high-density oli- gonucleotide microarrays using a light-directed combinatorial method. 19,2° With this method, photoreactive (versus chemically reactive) phosphor- amidite monomers are used in oligonucleotide synthesis. A series of physical photomasks that protect portions of the glass surface is designed based on the sequences of the oligonucleotides that will be synthesized. By shining a mercury light through the different photomasks [no more than 100 (4 x 25) masks are need to make 25-mers], the synthesis can be controlled in a spatially addressable fashion, allowing the production of microarrays whose sequence diversity is limited only by the resolution of the mask (up to 400,000 different 25-mer probes in a 1.6-cm 2 area). Each element on the array contains ~107 molecules of a particular probe sequence in a 50-/xm 2 area. One of the chief advantages of in situ methods over amplification/ microdeposition methods is that the sequences are selected from databases using software, and no physical handling of the probes is required, reducing the probability of probe cross-contamination or mix-up significantly. As with DNA microarrays, the mRNA (or cDNA) is labeled and hybrid- ized to the microarray. Hybridization occurs at lower temperatures, and protocols usually call for the fragmentation of the labeled material before hybridization to reduce the amount of secondary structure that forms at the lower temperatures. The microarrays are washed and scanned, and hybridization is detected by fluorescence. Analysis of Gene Expression Using Microarrays Preparation of Microarrays Microarrays designed to use fluorescence detection and containing probe sequences from different organisms can be purchased from suppliers such as Affymetrix (Santa Clara, CA) or Synteni (Fremont, CA) and many new producers are entering the market. The manufacture of certain types of microarrays may be technically inaccessible to most laboratories. How- ever, if a robotic spotting device is available, DNA microarrays can be fabricated in the laboratory. These instruments can be built using commer- cially available parts (see http://cmgm.stanford.edu/pbrown/) or can be pur- chased from companies such as Molecular Dynamics (Sunnyvale, CA) or 19 S. P. A. Fodor, J. L. Read, M. C. Pirrung, L. Stryer, A. T. Lu, and D. Solas, Science 251, 767 (1991). 20 A. C. Pease, D. Solas, E. J. Sullivan, M. T. Cronin, C. P. Holmes, and S. P. Fodor, Proc. Natl. Acad. Sci. U.S.A. 91, 5022 (1994). FIo. 1. Scanned image of a gene expression microarray. A fluorescent scan of a microarray prepared by mechanical microspotting is shown. A total of 240 human blood cDNAs and 8 Arabidopsis controls were amplified by PCR, purified, and arrayed on silylated microscope slides in duplicate at 200-t~m spacing. The 496 element microarray was then hybridized with a fluorescent probe prepared by enzymatic incorporation of Cy3-dCTP into cDNA by single- round reverse transcription of poly(A) + mRNA from cultured Jurkat cells, essentially as described in Table III. The microarray was scanned at 10-~m resolution with a confocal laser detection system. Fluorescent data are depicted in a rainbow pseudocolor palette for ease of visualization. Elements complementary to abundant messages are red, whereas black/dark blue elements correspond to genes having low or undetectable transcript levels. [ 1] MICROARRAY-BASED EXPRESSION MONITORING 7 Telechem International (Sunnyvale, CA). Generating arrays with a spotting device, although repetitive, is not prohibitively difficult (Tables I and II). An example of a scanned image of an array generated with a robotic spotting device is shown in Fig. 1. Hundreds or thousands of PCR reactions need to be performed using either a large number of different templates or a large number of different primers. Clones from any plasmid library (ordered or unordered cDNA, genomic DNA, etc.) can be used as tem- plates. In principle unamplified plasmid DNA could also be spotted, but the presence of vector sequence present in all elements could increase back- ground. If the genomic DNA is to be used as template, different pairs of oligonu- cleotide primers need to be designed and synthesized for each element TABLE I PCR AMPLIFICATION OF CLONES FROM GENOM1C DNA OR cDNA LIBRARIES 1. Assemble PCR reactions in 96-well plate Reagent Volume (/xl) 10x PCR buffer (15 mM Mg 2+) 10.0 dNTP cocktail (2 mM each) 10.0 Primer 1 (100 pmol/tzl)" 1.0 Primer 2 (100 pmol//xl) 1.0 Genomic or plasmid DNA (10 ng//xl) b 1.0 H20 76.0 Taq DNA polymerase (5 U/tzl) 1.0 100/~1 2. Amplify targets in 96- or 384-well format using 30 rounds of PCR (94 °, 30 sec; 55 °, 30 sec; 72 °, 60 sec) 3. Purify using PCR product purification kit and elute products with 100 ~1 of 0.1× TE (pH 8.0) 4. Dry products to completion in Speed-Vac 5. Resuspend each PCR product in 7.5 txl 5× SSC (0.3-1.0 mg/ml DNA) 6. Transfer to flat bottom 384-well plate (Nunc) for arraying Suggested materials PCR primers modified with a 5'-amino modifier C6 (Glen Research) 96-well thermal cycler (PCR system 9600-Perkin Elmer, Norwalk, CT) 96-well PCR plates (MicroAmp 96-well Perkin Elmer) Taq DNA polymerase (Stratagene, La Jolla, CA) PCR product purification kit (Telechem International) Flat-bottom 384-well plates (Nunc, Naperville, IL) " Use of generic primer pairs (-21-mers) to vector sequences allows high- throughput processing. h Plasmid DNA can be prepared by alkaline lysis and purified. The 96- well REAL prep (Qiagen) facilitates rapid preparation. 8 ANALYSIS OF GENE EXPRESSION [ 1] TABLE II MICROARRAYING AND SLIDE PROCESSING 1. Obtain silylated (free aldehyde) microscope slides (CEL Associates, Houston, TX) 2. Print cDNAs using microspotting device according to manufacturer's instructions 3. Allow printed arrays to dry overnight in slide box a 4. Soak slides twice in 0.2% (w/v) sodium dodecyl sulfate (SDS) for 2 min at room tem- perature with vigorous agitation b 5. Soak slides twice in doubly distilled H20 for 2 min at room temperature with vigorous agitation 6. Transfer slides into doubly distilled H20 at 95-100 ° for 2 min to allow DNA denatur- ation 7. Allow slides to dry thoroughly at room temperature (~5 min) 8. Transfer slides into a sodium borohydride solution c for 5 min at room temperature to reduce free aldehydes 9. Rinse slides three times in 0.2% SDS for 1 min each at room temperature 10. Rinse slides once in doubly distilled H20 for l min at room temperature 11. Submerge slides in doubly distilled H20 at 95-100 ° for 2 sec d 12. Allow the slides to air dry and store in the dark at 25 ° (stable for >1 year) Suggested materials Microspotting device (Telechem International, Molecular Dynamics, Cartesian) Microscope slides (CEL Associates) a Drying increases cross-linking efficiency. Several days or more is acceptable. b This step removes salt and unbound DNA. c Dissolve 1.0 g NaBH4 in 300 ml phosphate-buffered saline. Add 100 ml 100% ethanol to reduce bubbling. Prepare just prior to use! d Heating the slides aids greatly in the drying process. on the microarray. The Whitehead Institute Primer program (http://www. genome.wi.mit.edu/genome software/), which uses a nearest neighbor algo- rithm to calculate melting temperatures, has been used successfully to pick primers for the amplification of open reading frames from the yeast genome. 1L21 Using this program, greater than 94% of PCR reactions gener- ated useable products, 11 as detected by analyzing a small amount of product on a gel. Ideally, primer pairs should be chosen so that they have similar melting temperatures and so that the resulting PCR products are of similar sizes. When possible, primers should be synthesized in the same format as used for the PCR reactions (microtiter plates, either 96 or 384 well). Com- puter scripts can be written that automate the process of primer selection for groups of sequences or for an entire genome. The success of PCR reactions can be checked by analyzing the products on an agarose gel or by examining the microarray elements once they have been spotted: fluorescein-labeled dNTPs can be included in the PCR 21 W. Rychlik, W. J. Spencer, and R. E. Rhoads, Nucleic Acids Res. 18, 6409 (1990). [ 1] MICROARRAY-BASED EXPRESSION MONITORING 9 reactions and the array can be scanned for fluorescein to determine the reaction success. PCR reactions that fail are noted and repeated, although failures are generally arrayed along with successes in order to ease handling and documentation (if space is not an issue). It is generally not necessary to quantitate the amount of product if two-color hybridization (described later) will be performed. In addition, it is assumed that even with inefficient reactions, the amount of product will exceed the amount of target. PCR reactions are spotted onto glass that is derivatized chemically by treatment with reactive aldehydes (Table II) or polycations, such as polylysine. 6 Preparation of Polyadenylated mRNA Small differences in the environmental conditions to which a cell is exposed can have a profound impact on the global pattern of transcription. 5 If two growth conditions are to be compared, extreme caution should be taken to ensure that the cells from which the mRNA is isolated are treated equivalently. For example, cells should ideally be grown in the same batch of media, in tissue culture-treated plasticware, and should be harvested at similar densities if possible. Various methods can be used to rapidly isolate total RNA from cells, depending on the organism. Following isolation, the polyadenylated RNA is usually purified from total RNA on oligo(dT) resin. The amount of polyadenylated RNA needed will depend on the particular microarray, the labeling method, and the size of the hybridization chamber, but is generally between 0.5 and 10/zg. With 10/zg of labeled material, transcripts that are present at one copy per mammalian cell or one copy per every 20 yeast cells can be detected. If mRNA quantities are limited (0.5 tzg or less), the mRNA sample can be amplified by an in vitro transcrip- tion step as described later. Two-Color Analysis Because microarrays have typically been manufactured using relatively simple instruments, ensuring that identical amounts of DNA are deposited at different locations has been difficult. This theoretical limitation can be overcome by using a two-color labeling strategy. 16 Here target derived from mRNA from one condition is labeled with one fluor, whereas target from a different condition is labeled with a second fluor. Similar amounts of labeled material (usually eDNA) from the two samples are cohybridized to the microarray and the fluorescence intensity at the two appropriate emission wavelengths is determined. A good estimate of the relative differ- ences in abundance of a target in the two samples can be obtained by comparing the ratio of the fluorescence intensities at the two wavelengths. By always using the same reference sample, microarrays produced using 10 ANALYSIS OF GENE EXPRESSION [1] different sets of PCR products and by different individuals can be compared. Two-color strategies have been employed for mutation screening with oligo- nucleotide microarrays 2 and could theoretically be used in expression analy- sis, reducing the number of hybridizations that would need to be performed. Array Design The design of arrays for gene expression experiments should include appropriate controls for signal linearity and specificity. Probes to targets whose abundance is well characterized and probes to genes from different organisms should be included on the microarray, for example, human probes on an Arabidopsis microarray or bacterial genes on a yeast microar- ray. Target for these control probes can be generated by cloning the probe into a vector containing a phage promoter, and a poly(A) tract, and then generating polyadenylated mRNA by runoff in vitro transcription using the phage polymerase. This target can be added at various concentrations or at various times during the mRNA isolation and hybridization. Probes to ribosomal genes and to nontranscribed regions of the genome may also be included. In addition, controls that allow different scans to be normalized (as described later) should be arrayed. These may include multiple probes to genes whose expression is not expected to vary in the experiment (in yeast, probes complementary to the genes encoding actin and the TATA- binding protein have been used) or spots of total genomic DNA (for mi- croarray experiments)) 1 Direct Labeling of Messenger RNA for Hybridization to Oligonucleotide Microarrays One of the simplest methods for generating target is to label mRNA directly. At least two different methods have been reported: the mRNA is first fragmented to 30-50 base sizes by precipitating the RNA, resus- pending the sample in a buffer containing magnesium ions [40 mM Tris- acetate (pH 8.1), 100 mM potassium acetate, 30 mM magnesium acetate] and then heating the sample to 94 ° for 35 min. The RNA fragments are then kinased by diluting the fragmentation reaction 2-fold and adding ATP to 8/xM, dithiothreitol (DTr, to 3 mM), bovine serum albumin (BSA, 10 tzg/ml), and polynucleotide kinase. The reaction is incubated for 2 hr at 37 °. Then a biotinylated oligoribonucleotide (5'-biotin-AAAAAA 3') is ligated directly to fragmented, kinased mRNA using T4 RNA ligase in a buffer [50 mM Tris-HCl (pH 7.6), 10 mM DTT, and 1 mM ATP] containing a 10-fold molar excess of 5'-biotin-AAAAAA 3'. 4 Heat-denatured mRNA (or total RNA) can also be incubated with a biotinylated psoralen derivative (35 tzM, from Schleicher and Schuell, [...]... replication of plasmids carrying an SV40 origin of replication 38 On transfection of an expression plasmid featuring the gene of interest under the control of a strong, mostly viral, promotor and an SV40 ori, this construct replicates rapidly to high copy numbers, allowing the recovery of protein from cell supernatants or recombinant cells within 48-72 hr 39 However, as the protein-processing machinery of the... MOIs Analysis of recombinant protein production in these samples will indicate how much virus is needed to obtain maximal yield after a given time of infection Only this type of experiment, in combination with the careful titration of virus stocks, will guarantee the reproducibility of experiments during repeated production runs 5 Production in Insect Cells Finally, a working virus stock of sufficient... identification of successful recombination events at a frequency of 80-100% 61 Instead of cotransfection of viral D N A and recombinant transfer vector into insect cells, recombination can also be performed in bacteria with subsequent isolation and transfection of the recombinant bacmid 62 Verification of recombination events is done by polymerase chain reaction ( P C R ) 63 as well as on the protein level,... used myeloma cell lines for recombinant protein production are J558L [secreting immunoglobulin (Ig) Zl light chains19], nonsecreting Sp2/0 Agl42°,21 and NSO cells.22-24 For expression of engineered chimeric or humanized antibodies, these cell lines proved to be extremely useful Other types of proteins were also expressed successfully in myeloma cells 19,25 Following cotransfection of dominant antibiotic... to achieve recombinant protein production Table I summarizes the most frequently employed expression systems for this purpose Prominent examples of the different types of systems will be described in theoretical and practical detail later A few general remarks should precede the overview, however One of the key factors for success in recombinant protein expression is optimal cell culture maintenance... Accordingly, early expression protocols using the baculovirus system were based on the identification of virus populations exhibiting an occlusion body-negative phenotype 59'6° As insertion of the gene of interest under control of the polyhedrin or pl0 promotor was achieved by homologous recombination, occurring with a frequency of 1-3%, the search for recombinant viruses was a tedious and painstaking procedure... batch procedure" in the case of secreted proteins, albeit not in every instance Comparative Protein Production in Different Systems: Expression of hu-LIF In order to exemplify the overview on expression strategies discussed earlier, the following section is used as an example of the practical applica74A Day,T Wright,A Sewall,M Price-Laface,N Srivastava,and M Finlayson ,in "Methods in MolecularBiology"(C... Results of the protein analysis are shown in Fig 1 All purified protein preparations were also analyzed by N-terminal amino acid sequence determination and were found to be homogeneous, except for the CHO-derived material, where the first N-terminal amino acid was missing in 25% of the analyzed protein sample A summary of data on yields and time frames for each expression system is compiled in Table III Interpretation... via infection: Several systems based on recombinant D N A or RNA viruses (including retroviruses) are available, which allow expression of transgenes again on a stable or transient basis The transfer of genetic material into suitable recipient cells is used for many purposes, ranging from expression cloning and mutational analysis to gene or cell therapy In this article the focus of discussion will be... 22, 41 (1992) 26 ANALYSIS OF GENE EXPRESSION [2] Expression in Murine Erythroleukemic (MEL) Cells One of the main contributors to the successful establishment of hightiter production cell lines is the so-called "position effect," i.e., integration of the expression plasmid into a transcriptionally highly active site of the genome Unless gene targeting via homologous recombination is performed, this . express recombinant proteins in eukaryotic cells has greatly increased our ability to study protein function in a variety of settings. Initial work in this field focused on gaining an understanding. looking for new methods to express recombinant proteins in eukaryotic systems. We realize that no single volume can be complete, and certainly there are methods for recombinant protein expression. to genes whose expression is not expected to vary in the experiment (in yeast, probes complementary to the genes encoding actin and the TATA- binding protein have been used) or spots of total