Mycoplasma infection significantly alters microarray gene expression profiles.

Mycoplasmas are the smallest and simplest self-replicating prokaryotes. Most are parasites, and their con-tamination of primary and continuous eukaryotic cell lines represents a sig-nificant problem in research, diagnosis, and biotechnological production (1). Unlike bacteria and fungi, they rarely produce turbid growth or cellular dam-age, and most are resistant to the antibi-otics commonly used in long-term cell cultures (1,2). These factors combine to make them a stubborn contaminant, and recent surveys have shown that they affect up to 87% of cell lines. Mycoplasmas have also been known to infect the respiratory, gastrointestinal, and urogenital tracts of many patients, often without any apparent illness (3). Their ability to cause chromosomal re-arrangements has led some to postulate them as a cause of cancer. Others have argued that they are cofactors for a va-riety of conditions including arthritis, Crohn’s disease, and acquired immuno-deficiency syndrome (AIDS) (4).Mycoplasma genomes are small and have limited options for metabolism and replication. Genomic analysis of Mycoplasma pneumonia and Myco-plasma genitalium, for example, show that they have no genes involved in amino acid biosynthesis and very few responsible for that of cofactors (2,5,6). Instead, they must depend on their host cells for the supply of essential amino acids and vitamins. Competition for nu-trients can disrupt host cell integrity and alter function. To colonize a host cell, mycoplasma must adhere to the cell’s surface, thereby interfering with its receptors, altering its transport mecha-nisms, and triggering specific signal cascades. Its lack of a cell wall can allow fusion between the mycoplasma and host, and when this occurs, a va-riety of components are delivered, in-cluding hydrolytic enzymes, cytokines, chemokines, prostaglandins, and active oxygen and nitrogen metabolites. My-coplasmas also show mitogenic activity. Mycoplasma penetrans, for instance, has components that display properties similar to leukocyte integrins (2). Thus, mycoplasma infection has diverse and far-reaching consequences, from unsafe biological products, to unreliable experimental results (1). Irrespective of their role as a benign colonist, a cofactor, or even a cause of disease, their ubiquity has profound implications for diagnosis, classifica-tion, or characterization on the basis of patient samples or cell lines. Studies of the effects of mycoplasmas are often dif-ficult to perform, because their presence is generally undesirable, and there is an understandable reluctance to introduce them into the laboratory intentionally.During the course of a microarray experiment, routine screening identified mycoplasma contamination within one replicate set of experimental controls (after RNA extraction, but before hy-bridization to the chips). Rather than simply dispose of the samples, we took the opportunity to explore the effects of mycoplasma infection by hybridizing the contaminated samples to arrays and comparing their expression profiles to the replacement, uncontaminated, controls. There was a certain amount of serendip-ity in this, as we chose to retrospectively analyze the results of a compromised experiment, rather than to deliberately infect a cell line with mycoplasma. The design of the original experi-ment, which was intended to test the response of MCF7 human breast ad-enocarcinoma cells to a pair of drug treatments, is shown in Figure 1. There were three separate control groups: (i) cells that were simply resuspended for 2 h (CSC); (ii) cells that were treated for 2 h with dimethyl sulfoxide (DMSO) (DC); and (iii) cells that were treated for 4 h with DMSO (2DC). Mycoplasma contamination resulted in one sample from each of these groups becoming infected, and an extra set of replicates was generated to replace the infected ones. In order to investigate the effects of contamination, the three contaminated samples were compared to the nine uncontaminated replicates (three from each group). MCF-7 cells were cultured in RPMI with 10% fetal calf serum (FCS) (both from PAA Laboratories, Yeovil, UK) and L-glutamine (Sigma, Gillingham, UK). No antibiotics or antifungal treatments were used. During routine mycoplasma screening using the Hoechst technique (7), the first batch of cells were shown to be positive. This identified the pres-ence of mycoplasma, but not the spe-cific strain. Retrospective analysis was not possible, because the contaminated samples were immediately destroyed. Subsequent batches of cells grown from fresh stocks were shown to be clear by repeated PCR screening both before and after the experimental procedure. All experiments were performed us-ing HgU133A oligonucleotide arrays (Affymetrix, High Wycome, UK) as described at (http://www.affymetrix.com/products/arrays/specific/hgu133.affx). Total RNA from each sample was used to prepare biotinylated target RNA, with minor modifications from the manufacturer’s recommendations at (http://www.affymetrix.com/support/technical/manual/expression_manual.affx). Briefly, 10 µg of mRNA were used to generate first-strand cDNA by using a T7-linked oligo(dT) primer. After second-strand synthesis, in vitro transcription was per-formed with biotinylated UTP and CTP (Enzo Diagnostics, Farmingdale, NY, USA), resulting in approximately 100-fold linear amplification of RNA. A complete description of biochemical procedures is available at (http://bioinf.picr.man.ac.uk/mbcf/downloads/GeneChip_Target_Prep_Protocol_CRUK_v_2.pdf). The target cDNA generated from each sample was processed as per the manufacturer’s recommendation using a GeneChip