Back to Title

Discussion

    Microarray technology was used as an exploratory analysis tool to examine gene expression variation between ∆zms2 and wild type strains of Saccharomyces cerevisiae. The goal of this experiment was to identify effects and responses to oxidative stress in the mutant strain Δzms2, a strain of S. cervisiae known to be sensitive to oxidative stress.

Observed Gene Expression Trends

    From our analysis of a set of over-expressed and under-expressed genes, several interesting trends were observed. YNL130C, a gene known to be involved in phosphatidylcholine biosynthesis, and YMR307W, which functions in cell wall organization and biogenesis, were both found to be down-regulated in the ∆zms2 strain. In light of these observations we determined it is possible that in ∆zms2 mutations can have a negative effect on cell wall components and enzymes needed for its organization and synthesis.  Phosphatidylcholines are in a class of phospholipids that incorporate a choline into the head group.  Phosphatidylcholines are a major component of biological membranes as well.  It is possible that the down-regulation of cell wall organization, biogenesis, and the enzymes involved in these processes also affect the synthesis of phosphatidylcholine to be used in these cell walls. As to exactly how oxidative effects or mutations in Δzms2 cause this we cannot be certain without further experimentation, but it’s possible that energy in the cell is being diverted away from these processes in order to cope with stress caused by oxidizing agents.

    YNL021W was also found to be down-regulated. This gene encodes for a histone deacetylase activator. The function of histone deacetylase is to remove acetyl groups from DNA allowing histones which are positively to bind to the negatively charged DNA backbone. In the presence of oxidative stress it is possible that transcription for certain genes would be up-regulated to compensate for any damage caused by such stress. So it would make sense that a gene encoding an activator for repression of transcription would be down-regulated in the mutant Δzms2 strain. There may also be a possible connection between the down-regulation of YNL021W and YNL130C.  It is possible that because of the decreased activity of the histone deacetylases, the enzymes needed to synthesize phosphatidylcholine in the cell are not as readily available.  This may be the cause for the decreased phosphatidylcholine biosynthesis that is found in ∆zms2 mutants.

    YOL063C and YOL023 were both found to be upregulated on the arrays. YOL063C gene is of unknown function, however it lies on the same chromosome within 70 kb of YOL023, a known translation initiation factor. This could imply that YOL063C encodes for a response to oxidative stress and that YOL023 is responsible for initiating translation of this response.

    YDR510 a gene involved in protein sumoylation or more specifically tagging of proteins for sumoylation was found to be up-regulated strongly in two of the arrays. YLL039C a gene encoding protein tagging as a stress response was also found to be up-regulated. Protein sumoylation is often found to function in transcription regulation, stress response, and cytosolic transport.  It is possible that in response to the stresses put on the cell in the ∆zms2 mutants, both of these genes are up-regulated in order to compensate for the down-regulation of other proteins.  It is also notable that both of these genes are up-regulated, have uses in response to stress, and are both involved in protein tagging.

    It is important to note that these trends are observed from analysis of just four microarray slides and so further experimentation must be conducted in order to support any of these conclusions. For example, conducting a PCR of wild type and ∆zms2 strains to confirm over-expression of these genes and then possibly performing a knockout mutation of the translation initiation factor YOL023 to observe if any effects on the expression of YOL063C occur.

Analytical Results

    Due to the inherent variation in microarray it is necessary to remember that effects such as dye bias could have influenced expression in the conclusions noted above. This is especially true for this experiment, where two arrays using reversed dye scheme were used in the comparison. The standardization performed in this experiment normalized the data some but, future experiments could include a Lowess  normalization to decrease any dye bias further as well as further quantile normalization to even out the means between the arrays so that a t-test could be conducted to show statistical significance of any findings. This could not be performed on the data used for the analysis above because of the difference in the mean of array 2 (Δzms2 green bottom) which kept the data from satisfying conditions of normality needed for analysis in a t-test.

Experimental Issues

    When conducting this analysis data was used from analysis conducted by other groups with little to no collaboration. This leaves room for significant error because different groups used varied criteria for determining whether data should be flagged, as well as differences in molecular technique. Incorrect expression values could also be due to errors when constructing probe since RNA samples were found to be slightly contaminated with DNA and tRNA. It is also important to note that all arrays were shipped cross country to be scanned leaving a lot of time for dehydration of arrays which increase background and possible contamination by dust or air-borne particles.