|
|
||
|
Discussion |
|
|
|
Oxidative Stress has become an important area of study due to its role in many diseases and contribution to aging. The means of protection against oxidative stress are anti-oxidants. Two particular potential yeast anti-oxidants of interest are ZMS1 and ZMS2 which act as suppressors for a mutation in the yeast ZWF1 gene which exhibits oxygen sensitivity in the mutant phenotype. To better characterize and understand the role they may play in oxidative stress protection, a microarray analysis was performed. RNA was isolated during mid-log phase of yeast strains zms1Δ, zms2Δ, zms1Δ zms2Δ, ZMS1++, and ZMS2++. Enough RNA was isolated from each strain to provide appropriate concentrations for further analysis with relatively little protein contamination. However, the agarose gel revealed a significant amount of DNA contamination in some samples which could have potentially made RNA concentration appear greater than it actually was. The RNA was reversed transcribed into cDNA and used as probes for the microarrays. The resulting data for all strains except zms2Δ was not usable as there was low signal and several streaks across the slide. In addition, many of the slides fluoresced green but fluoresced very little red. The low signal may have resulted from poor RNA quality. Though the RNA appeared usable after initial isolation, it may have degraded over time in the freezer due to RNases or the DNA contamination was significant enough to affect the RNA concentration estimation and not enough cDNA was generated from the RNA to hybridize to the slide. There could have also been problems during washing or hybridization which may have led to the streaking as well. The streaks may have been caused by unwashed hybridization solution that was allowed to dry to the slide. The lack of red fluorescing may have been a result of the red dye fading due to its easy degradation in the presence of ozone and light which occurred a few times during washing and hybridization. If this experiment is to be repeated, a higher concentration of RNA may be used in labeling and careful note of washing will be used. Furthermore, all procedures involving dye exposure will be done in the dark and receive very little light As a result of poor quality data from all strains, but zms2Δ, analysis was continued using data from slides 104, zms1Δ, and 106, zms1Δ zms2Δ, in the year 2004. Data was analyzed and normalized using scanalyze. This data was used to gain information in three different areas: 1) to develop a better normalization protocol for more accurate data analysis, 2) to determine the most highly expressed or repressed genes for strains zms1Δ and zms1Δ zms2Δ, and to discover the possible relationship and function between these genes, 3) lastly to determine the expression of the genes involved in the pentose phosphate pathway for all three strains. Normalization of Microarrays Normalization is the technique that adjusts for variation in the microarray results that have been caused by microarray technology rather than biological differences between mutant and wild-type yeast (Smyth, G.K., Speed, T.P.). There are many different sources of error that have caused these variations. Variability could be caused due to manufacturing process of the probe DNA, the amount of oligos spotted on the slide, the ability for cDNAs to bind to the array. Dye bias could arise from the physical properties of the dye due to decay or ability to hybridize with cDNAs. Hybridization of the dyes to the cDNAs could be affected by humidity, dust, salts or other molecules. Different scanning settings could cause imbalances between the red and green dyes. For example, higher scanning intensities improve the quality of the signal but increase the risk of saturation (Jahan, Nusrat). All of these factors can affect the results for intensities of expression, therefore normalization must be considered when analyzing the data. By transforming the data into logarithmic scale, it makes the data easier for the experimenter to analyze. By transforming the intensity values into log base 2, a scale between –16 to 16 is formed, where high/low values before the transformation are the high/low values in the new scale. Using mean and median normalization lessens the effects of outliers, which are usually caused by sources of error listed earlier. It results in a more comparable data that still preserves each gene’s overall differences. Before the red and green intensities were transformed into their log base 2 values, they were graphed separately in a histogram for their frequencies. As seen in Figures 3 and 4, the data is right skewed. This was anticipated, since most genes do not have do not have high expression levels in microarrays. After transforming the intensities into log base 2 values, the graphs have a more symmetric distribution, which is shown in Figures 5-6. After mean and median normalization for both red and green intensities, the data is slightly more symmetrical. Mean normalizations may be a little more helpful when analyzing the data since the curve is centered on the value “0”, although median normalization is suppose to give better, symmetrical results. Better results could have been reached if more data or better methods during the experiment were performed. The data, especially for CH2 intensities, were not the best data to work with and did not give great results. Mean and median normalization of gene expression was also performed for each individual microarray. In Figures 9-16, symmetry is seen slightly better after both normalizations and the data is centered on the value “0” for both normalizations. There wasn’t much difference between the two normalization techniques when normalizing the individual microarrays. Again, more data could have been needed in order to gain significant results and better techniques could have been used in order to obtain the best results in the microarrays. Clustering Analysis The top and bottom sections from slide 104 do not show the same genes being up or downregulated. Slide 106 does not show any similarities as far as specific genes on slide 104. However, a common theme is that genes involved in amino acid synthesis are affected by these mutations. Serine and histadine were downregulated on slide 104. Slide 106 showed tryptophan synthesis downregulated. A tentative proposal is that genes involved in amino acid synthesis are affected by ZMS1 single knockout and ZMS2 + ZMS1 double knockout, but since these experiments have yet to be reproduced, no conclusions can be made. Genes that are related to meiosis, conjugation and growth are also downregulated, which makes sense since RNA samples were collected at log phase of yeast growth in optimal growing conditions. Meiosis in yeast only occurs during bad growing conditions, which increases chance of diversity. As a result of diversity, some yeast will have traits enabling them to survive and produce offspring. Slide 104 top & bottom (Table 1 & 2 respectively) and slide 106 (Table 3) follow the idea that these yeast were reproducing asexually. Genes functioning in metabolism were also upregulated or downregulated in Table 1 and 3, but did not show a distinguishable pattern of expression. Expression Determination The pentose phosphate pathway generates electrons, via NADPH production, to be used by anti-oxidants to remove harmful oxidizing agents. ZWF1 is one of the key enzymes in this pathway because it reduces glucose-6-phosphate and produces a molecule of NADPH. When ZWF1 is mutated, the yeast becomes sensitive to oxidative stress since there are not enough electrons available for the antioxidants to dispose of the oxygen radicals. However, overexpression of ZMS1 and ZMS2 genes has been shown to suppress this mutation (Slekar laboratory, unpublished work). Yet, it is still not known how these genes are able to do this, therefore, genes from the pentose pathway were analyzed to reveal if ZMS1 or ZMS2 have any effect on this pathway as ZMS1 and ZMS2 may be involved with an alternative NADPH synthesis pathway. The activation or inactivation of this alternate pathway may cause the pentose pathway genes to be repressed or induced since it may be wasteful for the cell to turn on both pathways at the same time. Results from the data analysis show that no genes were induced in either strain except ZMS1 in zms1Δ and zms2Δ and ZMS2 in zms1Δ zms2Δ. Yet, ZMS1 was knocked out in zms1Δ and ZMS2 was knocked out in zms1Δ zms2Δ, thus, this result is most likely due to the poor quality of some regions in the data. In zms1Δ, gene TKL1, which catalyzes conversion of xylulose-5-phosphate into sedoheptulose-7-phosphate and ribose-5-phosphate into glyceraldehyde-3-phosphate, was repressed, all other genes exhibited normal expression. In zms2Δ RPE1, which catalyzes the conversion of ribulose-5-phosphate into ribose-5-phosphate, was also repressed. In zms1Δ zms2Δ three genes were repressed; SOL4 which converts D-6-phospho-glucono-δ-lactone into 6-phospho-gluconate, GND1, which catalyzes 6-phospho-gluconate into ribulose-5-phosphate, and TKL1 again. All other genes showed normal expression. It is possible that ZMS1 and ZMS2 may play a role in these genes’ expression as they control a wide range of genes. It is likely they do not directly influence these genes since the pentose pathway takes place in the cytoplasm and ZMS1 acts as a transcription factor mainly for nuclear and mitochondrial proteins (Lu et al 2005). Yet, one of the proteins it activates may in turn activate the genes in the pathway. More replications and more reliable data must be done in order to confirm this conclusion. In addition, genes that ZMS1 and ZMS2 activate can be used to determine if the activated genes have any affect on altered pentose pathway genes through tests such as blotting and reverse transcription PCR. In addition, expression of ALD6, an aldeyhde dehydrogenase, was also observed since ZMS1 is a known transcription factor for this gene (Grabowska, D., and Chelstowska, A 2003). Overexpression of ALD6 has been shown to suppress a ZWF1 mutation. Overexpression of ZMS1 is also shown to cause an overexpression of ALD6 which maybe the way ZMS1 works to suppress a ZWF1 mutation. ZMS2 may also overexpress ALD6 when overexpressed and, thus, the two genes ZMS1 and ZMS2 may have redundant roles. Therefore, ZMS1, ZMS2, and ALD6 may be involved in an alternative NADPH synthesis pathway. Data from the microarray supports this theory because ALD6 was repressed in both zms1Δ and zms1Δ zms2Δ, however there was no data for zms2Δ on ALD6 expression. Thus another replication of this experiment is needed to produce zms2Δ and confirm ALD6 repression in all three strains. If this is the case then further studies using Reverse Transcription PCR or blotting test can be used to specifically identify ZMS1’s and ZMS2’s affect on ALD6.
|
|