Assessing Differential Gene Expression

As the complete human genome sequence emerges, research shifts from questions of genomics to those of proteomics--determining the function of individual gene products and mapping global gene expression patterns. Gene expression patterns change continually during the course of tissue development and differentiation. The expression of different gene products at any given time within a particular cell defines the cell's characteristics and helps determine how it will react to external stimuli. Alte

Nov 26, 2001
Barbara Cunningham
As the complete human genome sequence emerges, research shifts from questions of genomics to those of proteomics--determining the function of individual gene products and mapping global gene expression patterns. Gene expression patterns change continually during the course of tissue development and differentiation. The expression of different gene products at any given time within a particular cell defines the cell's characteristics and helps determine how it will react to external stimuli. Alterations in these expression patterns often accompany various disease states. Thus, understanding how the expression of certain genes is toggled on and off during normal development and in disease, and how environmental factors affect these patterns, helps researchers focus drug development efforts.

Researchers have many tools at their disposal to measure differential gene expression.1 Certainly microarrays are a viable option,2 but relatively few researchers have access to this big ticket technology. Instead scientists often rely on relatively "low tech" protocols including Northern blotting, RNase protection assays, differential plaque hybridization, subtractive hybridization, differential display, representational difference analysis, serial analysis of gene expression (SAGE), and rapid analysis of gene expression (RAGE).3,4


For many expression analysis protocols, reverse transcriptase (RT)-PCR is just a preliminary step. However, recent improvements to this method have rendered it sufficiently sensitive and reliable to be an effective differential gene expression analysis technique in its own right. RT-PCR is based on the theory that when researchers subject equal amounts of RNA from different samples to identical amplification conditions, the product yield will reflect the expression levels in the starting material. In practice, however, technical problems can complicate the issue. Scientists have overcome many of these pitfalls using improved reverse transcriptases, specialized RNA-isolation kits, and RT-PCR kits that are designed for improved expression quantification.5

Many companies now offer one-step RT-PCR kits, in which reverse transcription and amplification occur within the same reaction tube and cycle. Furthermore, scientists can run multiple reactions in the same tube, a process called multiplexed PCR, to control for sample-to-sample variation. This technique involves the use of primers that amplify an internal standard--typically a housekeeping gene--within each experimental reaction. Analogous to the interpretation of Northern blot signals, the quantification of the experimental signal relative to the control reflects expression changes between samples. Alternatively, some kits make use of a competitive standard that contains primer-binding sites identical to those of the sequence being studied. Comparison of differences in the competitor concentration needed between samples reflects expression differences in the gene of interest.

A relatively new advance developed approximately four years ago is real-time RT-PCR, which enables "online" monitoring of expression changes. Real-time RT-PCR methods usually require the use of a labeled, sequence-specific probe that imparts a level of fluorescence proportional to the amount of product. Some companies have found ways to work around this requirement. For example, Valencia, Calif.-based QIAGEN has eliminated the need for sequence-specific probes with its QuantiTect™ SYBR® Green PCR and RT-PCR kits. Introduced this past summer, both kits contain the fluorescent dye SYBR Green I in a master-mix format. This dye, manufactured by Eugene, Ore.-based Molecular Probes, indiscriminately binds to any double-stranded DNA molecule in the reaction, including nonspecific products and primer dimers. To reduce the problem of inaccurate quantitation due to low specificity sometimes associated with SYBR Green, the kits use HotStarTaq™ DNA polymerase, which prevents formation of primer dimers and nonspecific amplification products during reaction setup and the first denaturation step. As a general rule, the SYBR Green I detection method is a good starting system to use when studying several different targets that may otherwise require the prohibitively expensive construction of several different sequence-specific probes.

Differential Display

Peng Liang and Arthur Pardee invented mRNA differential display in 1992.6 That same year, Liang and Pardee founded GenHunter Corp. of Nashville, Tenn., to market the technology. By 1998, the technique had had such an impact that the German Society of Molecular Biology and Biological Chemistry gave its Molecular Bioanalytics Award to Liang and Pardee for their contribution to genomic analysis. Indeed, differential display has become the most widely used technique for studying differential gene expression. According to Jonathan Meade, product manager at GenHunter, the number of publications per year using differential display exceeds those published using all other techniques combined.

The procedure involves the systematic amplification of the 3' terminal portions of mRNAs followed by resolution of the resulting fragments on a sequencing gel. Oligo-dT primers containing a single 3' anchor nucleotide (A, C, or G)--called "one-base anchored primers"--bind to the junction between the mRNA's 5' poly-A tail and its coding sequence for reverse transcription. The scientist then amplifies the resulting cDNAs using both the anchor and additional, arbitrary-sequence primers, and finally resolves the different mRNA subpopulations by denaturing polyacrylamide gel electrophoresis.

The novelty of differential display stems from its use of multiple primer combinations to statistically represent all genes expressed in a eukaryotic cell in an ordered, sequence-dependent fashion. The method offers ready access to sequence information, facilitates the development of probes for the isolation of cDNA and genomic DNA for subsequent studies, and is highly sensitive. According to GenHunter, five micrograms of total RNA is enough to process all three anchored primers in combination with 80 arbitrary 13-mers, statistically representing the majority of eukaryotic cellular mRNAs. Unlike other techniques whose success is biased by relative mRNA abundances, differential display detects both abundant and rare mRNAs, as long as the arbitrary primer sequences match the target mRNA. Furthermore, the technique requires no previous knowledge of the genes within the study.

GenHunter offers two lines of differential display kits, each containing 10 separate kits. The company's RNAimage kits enable radioactive detection of the PCR products, whereas the RNAspectra Green and RNAspectra Red kits enable fluorescent detection. Each kit contains a unique set of eight different arbitrary 13-mers, with a total of 80 arbitrary 13-mers in all 10 sets, which, in combination with each of the three one-base anchor primers, provides up to 240 separate reaction conditions. Each kit provides approximately 28 percent coverage of expressed genes; incorporating increasing numbers of kits into a study increases the representation of expressed genes, with 98 percent coverage obtained by using all 10 kits.7

GenHunter also offers two different gene-profiling services that are fully automated and use fluorescent differential display (FDD). The FDD Preview Service uses four primer combinations, giving approximately five percent gene expression coverage. Researchers provide a sample containing 20-50 micrograms of total RNA, and the company removes contaminating DNA, confirms RNA integrity and quantity, and reverse transcribes mRNA with a one-base anchored oligo-dT primer. An automated FDD reaction is then run in duplicate using the single one-based anchored primer and a combination of four arbitrary 13-mers. Finally, the fragments are separated on a six percent denaturing gel and visualized using a fluorescent laser scanner. The researcher receives FDD images and is offered advice regarding which bands should be selected for excision and further characterization. For additional fees, GenHunter can excise, re-amplify, clone, and sequence these interesting bands on-site.

The Comprehensive FDD Screening Service includes all the same steps as the Preview Service, but the company will use all three one-base anchor primers for reverse transcription, and as many arbitrary primers as the researcher wants, in order to increase gene expression coverage. Four different packages are available, offering 24, 72, 144, or 216 primer combinations per RNA, yielding estimated gene expression coverages of 28, 62, 86, and 95 percent, respectively. Prospective clients must pre-apply for both services, and only selected projects will be accepted.

Other companies featuring differential display kits include Palo Alto, Calif.-based BD Biosciences-CLONTECH, Fullerton, Calif.-based Beckman Coulter, and Carlsbad, Calif.-based Qbiogene. CLONTECH's Delta Differential Display Kit requires two micrograms of total RNA and begins with an oligo-dT-primed synthesis step. The resultant cDNAs are then amplified by long-distance PCR using pairwise combinations of arbitrary "P" primers and oligo-dT "T" primers; a total of 90 primer combinations are possible when all 10 arbitrary "P" and nine "T" primers are used.

CLONTECH's "P" primers correspond to commonly-found sequence motifs within mRNA coding regions that were identified by computer analysis of over 200 mRNAs. This kit's use of longer primers allows researchers to employ higher annealing temperatures to increase stringency, thereby decreasing background. In addition, long-distance PCR yields fragments up to two kilobases in length, increasing the likelihood of detecting differentially expressed bands with a given primer. These modifications allow the researcher to reduce the number of PCR cycles from 40 to 25.

Beckman Coulter offers five different HIEROGLYPH mRNA Profile Kits, which come with 12 distinct oligo-dT anchoring primers and a unique set of four arbitrary primers.8 The anchored primers are 31-mers with a T7 RNA polymerase promoter sequence at the 5' end. Each 26-nucleotide-long arbitrary primer has a core of 10 nucleotides with an M13 reverse sequencing primer at the 5' end. Thus, after re-amplifying potential positives with M13 reverse and T7 promoter primers, fragments can immediately enter a sequencing reaction cycle, bypassing the cloning step. The company also offers a Fluorescent Differential Display Adaptor Kit (fluoroDD Kit), which contains fluorescently labeled primers that users can incorporate into the HIEROGLYPH kits for laboratories with fluorescent detection capabilities.

Finally, Qbiogene's displayPROFILE™ expression profiling kits overcome one of DD's fundamental limitations--the inapplicability of the technique to prokaryotic messages due to the method's reliance on oligo-dT- containing primers. The displayPROFILE system employs a differential display variant, called restriction fragment differential display (RFDD-PCR). Researchers use the restriction enzyme TaqI, which preferentially digests in coding sequences, to digest double-stranded cDNA. They then ligate adapter molecules, one of which contains any of 64 possible trinucleotide anchors, to the fragments and use these to prime amplification reactions. After resolving these fragments on sequencing gels, the scientists can determine which genes could have produced any given fragment using Qbiogene's displayFIT™ Web site ( Qbiogene also offers a displayPROFILE service, which includes the option of excising interesting genes and arraying them on custom DNA microarrays.

Subtractive Hybridization

cDNA expression libraries are collections of bacteria, each harboring a plasmid that contains a single cDNA under the control of a bacterial promoter. Thus, the collection, in theory, represents the set of genes expressed by the cells, tissue, or organism from which the library was derived. As different cell types produce different expression patterns, a different cDNA library can be made from each one.

Comparison of cDNA libraries representing different developmental time points can expose changes in expression pattern that accompany the differentiation process. One method for carrying out this comparison is subtractive hybridization, in which genes expressed under both developmental conditions are "subtracted out" of the analysis. For example, researchers can mix an mRNA sample from hormone-treated cells with a sample from unstimulated cells. Sequences that are expressed in both samples, regardless of the presence of hormone, form complementary pairs, whereas differentially expressed sequences remain single stranded. A cDNA library is then constructed from the unpaired sequences, every one of which, in theory, was up- or down-regulated by the experimental treatment.

Subtractive hybridization is biased toward abundant messages. However, one advantage of this technique is that there are a wide variety of commercially available, affordable libraries on the market. Those who either want or need to make a highly specialized library can find detailed protocols and user-friendly kits to ease library construction. For those without the time or expertise required, many companies offer special-order library construction services, although this option is more expensive than using off-the-shelf libraries available through a catalog.

Representational Difference Analysis (RDA) combines subtractive hybridization and PCR to provide a unidirectional analysis that is easy, efficient, and affordable even for small laboratories. The method begins with two distinct mRNA samples from control and experimentally manipulated cells or tissues. The samples are subjected to reverse transcription to produce a pool of cDNAs, which are then digested with a frequently cutting restriction enzyme. Specific linkers are ligated to both ends of each cDNA fragment, the fragments are amplified by PCR, the linkers are removed from both samples, and a new linker is added only to cDNAs of the experimental sample. The experimental cDNA sample is mixed with a 100-fold excess of control cDNA, and hybridization between complementary single strands is allowed to proceed.1 PCR is again employed to amplify only experimental, differentially expressed cDNA sequences using the available sites on both sides of the double-stranded cDNA. This technique selectively enriches for differentially expressed sequences, even those that are not highly abundant.

Other Techniques

Rockville, Md.-based OriGene Technologies Inc. has developed the Rapid-Scan™ Gene Expression Panel, a PCR-based method for gene expression profile analysis for genes or identified expressed sequence tags (ESTs) from human or mouse. The Rapid-Scan system uses first-strand cDNAs from different tissues or developmental stages that have been tested to ensure the representation of low-abundance transcripts and that have been normalized using an internal standard. cDNAs are serially diluted over a four log range (to ensure amplification within the linear range, thereby allowing semi-quantitation of relative mRNA expression) and are arrayed onto 96- or 48-well plates. After PCR with gene specific primers, investigators analyze the products by agarose gel electrophoresis. This technique produces a complete expression profile in three hours and eliminates the need to use radioactive probes.

Biofrontera Pharmaceuticals AG of Leverkusen, Germany, recently developed Digital Expression Profiling Display (DEPD®), a sequence-independent, PCR-based alternative to differential display technology that, according to company literature, offers greater sensitivity for very low abundance transcripts and a low rate of false positives. Biofrontera uses the technology to study diseases of the nervous system, including pain, stroke, schizophrenia, and Parkinson's disease. The method offers much higher resolution power than does differential display for the number of transcripts in a given sample--it can identify more than 100,000 individual DNA fragments--making it an excellent choice for the analysis of very complex tissues like the brain. Although DEPD technology is not commercially available, Biofrontera collaborates with various companies and universities.

A number of additional options exist for laboratories seeking out the changes in gene expression patterns that define developmental and pathogenic pathways. Differential display, subtractive hybridization, and RT-PCR must share space in a laboratory playbook that also includes microarray analysis, SAGE, RAGE, RNAse protection assays, and others. What is clear is that with the variety of technologies available, a laboratory can pinpoint a technique best suited for its individual needs, goals, and budget.

Barbara A. Cunningham ( is a freelance writer in York, Maine.
1. D.H. Kozian, B.J. Kirschbaum, "Comparative gene-expression analysis," Trends in Biotechnology, 17:73-8, 1999.

2. J. Cortese, "Array of options," The Scientist,14[11]:26, May 29, 2000.

3. A. Constans, "SAGE advice," The Scientist, 15[11]:21, May 28, 2001.

4. A. Wang et al., "Rapid analysis of gene expression (RAGE) facilitates universal expression profiling," Nucleic Acids Research, 27:4609-18, 1999.

5. A. Constans, "Reverse psychology," The Scientist, 14[17]:29, Sept. 4, 2000.

6. P. Liang, A.B. Pardee, "Differential display of eukaryotic messenger RNA by means of the Polymerase Chain Reaction," Science, 257:967-71, 1992.

7. P. Liang et al., "Analysis of altered gene expression by differential display," Methods in Enzymology, 254:304-20, 1995.

8. "Making light work of differential display," The Scientist, 12[1]:15, Jan. 5, 1998.

Supplemental Materials