Problem-based Learning Experimental Approaches to Microarray Analysis of Tumor Samples Received for publication, October 3, 2007 Laura Lowe Furge‡¶, Michael B. Winter‡, Jacob I. Meyers‡, and Kyle A. Furge§ From the ‡Department of Chemistry, Kalamazoo College, Kalamazoo, Michigan 49006-3295, and the §Laboratory of Computational Biology, Van Andel Research Institute, Grand Rapids, Michigan 49503 Comprehensive measurement of gene expression using high-density nucleic acid arrays (i.e. microar- rays) has become an important tool for investigating the molecular differences in clinical and research samples. Consequently, inclusion of discussion in biochemistry, molecular biology, or other appropriate courses of microarray technologies has become essential in training the modern scientist. The following article offers preparatory and problem-solving questions for engaging students in the understanding of microarray technologies for use in conjunction with other course materials relating to discussions of microarray technology. Keywords: Microarray, cDNA, cancer. A central tenet of biochemistry is that the regions of DNA are transcribed to mRNA, and then the mRNA is translated into protein. As the high-throughput measure- ment of all proteins present within a cell or tissue is cur- rently difficult, nucleic acid-based microarray technology exploits this central dogma to relate mRNA levels to pro- tein expression [1]. The power of microarray technology comes from comparing the mRNA levels between control and experimental samples to determine differences in gene expression under different experimental conditions or to compare different cells/tissues [1–3]. Indeed, microarray technology allows for the simulta- neous comparison of expression for virtually all genes in the genome in a single experiment. Microarray technol- ogy has been applied to developmental biology (changes in gene expression over developmental time), drug analy- sis (identification of drug induced altered gene expres- sion, drug effects on pathways, identification of drug targets, etc.), analysis of disease states (identification of differences in gene expression between normal and diseased tissue, identification of genetic signatures of a certain disease, etc.), analysis of environmental changes in cells or tissues (gene expression associated with lack or abundance of nutrient, oxygen, etc.), and, more recently, mutational analysis (to identify a single nucleo- tide polymorphism—SNP or block of SNPs that contrib- ute to interindividual genetic differences at the DNA level) [1–3]. Of course, a disadvantage of gene expression microarrays is that while measurements of mRNA are taken, protein expression levels are inferred. Although significant correlations exist between mRNA levels and protein levels, these two measures are not always well correlated in some instances [1, 4]. Two critical pieces are required to perform a gene expres- sion microarray experiment: two populations of mRNA (called targets) that are to be compared and a microarray chip that contains nucleic-acid fragments (the probes) to which the derivatives of the targets will bind. Rather than measuring mRNA levels directly, the mRNA is usually con- verted to the more stable cDNA (complementary DNA) by the action of reverse transcriptase in the presence of fluo- rescently labeled nucleotide substrates [1]. Synthesis of cDNA in the presence of labeled nucleotides allows the detection of these species in later experimental steps. After labeling of the mRNA derivatives, nucleic acid- based microarray technology uses the hybridization prop- erties of complementary nucleic acids. On the microarray chip, oligonucleotides (probes) are immobilized on a solid surface, such as glass. The probe sequence placed on the solid surface can be constructed in a variety of ways. Ei- ther a single long probe (such as a 60-mer oligonucleotide) or multiple shorter probes (such as 10–20-mer oligonucleo- tides) are constructed for each gene in the genome. Each probe sequence is then placed within micrometer distan- ces of one another [5, 6]. This allows the presence of many probes in a small amount of space. Examples of microarrays using long oligo probes are those commer- cially produced by Agilent Technologies. On these arrays, each gene is represented by a 60-mer oligo probe. These 60-mer probes have been optimized to both represent unique regions in the gene and to have similar melting temperatures as other probes on the same array. Typical arrays that use short oligo probes are commercial arrays constructed by Affymetrix [6]. For these arrays, a set of 10–20-mer probes are synthesized that represent unique regions of each gene. Moreover, sets of control probes ¶To whom correspondence should be addressed. Tel.: 269- 337-7020; Fax: 269-337-7251. E-mail: lfurge@kzoo.edu. This paper is available on line at http://www.bambed.org DOI 10.1002/bmb.20161 149 Q 2008 by The International Union of Biochemistry and Molecular Biology BIOCHEMISTRY AND MOLECULAR BIOLOGY EDUCATION Vol. 36, No. 2, pp. 149–152, 2008