Home Introduction Methods Results References
The purpose of the present experiment was to utilize microarray analysis to characterize the differential expression pattern between a strain of wildtype S. cerevisiae, with a mutant strain. The mutant strain was a knockout for all hexose genes, which included transporter genes and also two receptor genes. The glucose receptor genes encoded for two receptors called RGT3 and SMF. These receptor genes were knocked out as well because they shared common primary transcripts with the hexose transporter genes. The media that both of the strains were grown on was 2% maltose. The choice of using the same maltose media complicated our microarray results. There was no glucose present for either the wildtype or mutant strains. The glucose transporters which were knocked out in the mutant strains had no glucose present to transport anyway. Therefore the alternative energetic pathways should have already been up regulated to compensate for the glucose depleted environment. If this is true, the changes in gene expression seen in our slides would only have been caused by chance. However, it is possible that the deleted delta hexose transporters might have other functions in signaling cascades. For example, the wildtype yeast receptors created a negative signal that there was no glucose in the environment where as the mutants, whose receptors were knocked out, were not able to signal at all. The difference in these molecular signals could have caused a change in gene expression between the two strains.
Samples with relatively pure amounts of RNA are usually measured at both 260nm and 280nm wavelengths. The ratio of absorbance readings of pure RNA should be between 19-2.2. The RNA separated from our wildtype yeast gave an absorbance ratio of 2.1 while the RNA isolated from out mutants yielded a 1.56. While the RNA from the wildtype stain was quite pure, the RNA from the mutant stain was contaminated. DNA is a possible contaminant for the mutant RNA. The phosphate backbone of the double helix acts as a shield and blocks the absorption of ultraviolet light. RNA's single stranded structure exposes more nucleotides which absorb much more light in the ultra violet range. Therefore, DNA contamination would cause the decrease in the 260/280 absorbance ratio that we observed in our results.
As we conducted our microarray procedure, inaccuracies were introduced through experimental error. In order to dry our slide after subsequent washings, we used a Dust-off compressed gas duster. Unfortunately the canister was dispersed at an inverted angle which caused liquid CO2 to come into contact with the slide. Upon contact, the water vapor on the slide immediately froze. When hybridizing the dye to the probes in complete darkness, an unknowing research assistant turned on the lights and pre-exposed the dye. Finally, when we conducted the second wash before shipment, the slides moisture created a decrease in the constant of static friction which caused the slide to slip away from the investigator and rapidly descend to the floor. Because of these events, the slide was unable to be read by the card reader and was not sent.
In order to increase changes in gene expression, we propose a different media for growing the two strains of yeast. By using a combination of carbohydrates instead of only maltose, the differences in gene expression should be much more pronounced. We suggest a mixture of 1% glucose with 1% maltose in the media. The wildtype yeast will have their glucose pathways more active while the mutants will have their maltose pathways up regulated. These changes will create on/off type variations in gene expression that can be more easily seen than the less noticeable changes caused by second messenger cascades. Also we would need to repeat the microarray experiments at least 8 times to give credible results. With any given repetition, biological background and noise can skew the data. Once numerous attempts have been recorded, statistical manipulations can be used to normalize the results. Only then can the results be trusted to be accurate.
Hierarchical clustering is an incredibly useful way to analyze data because it offers a simple way to compare and contrast your data. We not only used the clustering method to find similarities in expression, but we also were able to find out if these genes had related functions. By comparing their function, we were able to make predictions as to why they were expressed in such similar ratios. We were only able to cluster and analyze 5000 genes due to time and computer constraints. The computer program called MagicTool created a hierarchical tree which visually showed the most related genes, which were YMR137C, YMR106C, and YNL136W (Fig. 6). All of these genes seemed to have parallel functions in either DNA repair or packaging. The first two genes YMR137C and YMR106C both encoded for proteins involved in repairing double stranded DNA. When DNA is damaged there are many proteins and processes used by the cell to preserve its genetic material. This gene expression is quite interesting because we know that the mutant strains lost certain genes causing problems in their DNA. Therefore we would expect expression of DNA repair genes, which was what we saw in the YMR137C and YMR106C yeast genes. The third gene, YNL136W, is important in DNA packaging because it uses histone acetytransferases to unpackage DNA and allow for transcription to occur. We would expect that both the mutant and wildtype yeast cells would be regulating transcription of specific genes through DNA packaging. The gene expression in YNL136W was also significant in the clustering which indicates that there was some form of DNA packaging happening in the yeast cells. Since we were only able to analyze and cluster 5000 of genes, the information we obtained might be bias toward the limited data used. However, we found that the hierarchical clustering is extremely useful and we would like to study more genes in future trials.
We explored those genes which were significantly up-regulated and down regulated similarly in both the Δhxt and wt with the explore function in Magic Tool. We identified approximately four that were greatly under expressed and five that greatly over expressed. Of the high expressers sampled, two were involved in the glycolytic pathway though neither cell is importing glucose. In light of these results, we believe that maltose was broken down and sent through the glycolytic pathway. In exploring the expression pattern of low expressers, little was known about majority of the genes investigated. It would have been more interesting to look at those genes that were differentially expressed between the Δhxt and wt yeast. Furthermore, this process could have been ameliorated by transforming our values to the log base 2 as this would have allowed us to view our data on the same scale and present data in terms expression fold changes (i.e. a 4 fold up-regulation). Magic Tool unfortunately would not perform this operation for us and when we attempted to do so in excel not all of our genes would open in that program.
We looked at 12 genes of interest which included transports proteins, metabolic enzymes, and membrane receptors. To verify that the delta hexose mutant's transporters were in fact disabled, we analyzed the expression of the YJL219W gene which codes for hexose transports. As we expected, our results showed that there was significantly reduced amounts of expression as they were selectively knocked out. To explore if other metabolic pathways were being taken advantage of by the yeast instead glycolytic ones, we examined the YBR298C gene which codes for a maltose transporter. Because the yeast are grown in a 2% maltose media, we hypothesized that the yeast's disaccharide transporter would be highly active. Our suspicion was confirmed as the gene was highly up-regulated. The glucose membrane receptor, YDL194W, was slightly down regulated. This was surprising because the delta hexose mutants should have had their receptors knocked which would yield no expression at all. The primary structure of both hexose transporters and hexose receptors are strikingly similar and probably undergo post translational modification to gain their specific function. Therefore the small expression value exposes an unreliability in the data. These three genes represent the three categories of genes (metabolizers, transporters, and receptors), we thought would be affected by the delta hexose knock out. For more specific information about our other genes of interest examine the table below.
Genes of Interest:
YJL219W is a hexose transporter gene. The expression ratio confirms that the hexose transporters were knocked out, but our data did have some variation because it shows a very small amount of expression. This is most likely due to segmentation error or some stray dye in on the slide.
YDL194W encodes for a glucose sensor located in the plasma membrane. The data suggests that this gene was slightly down regulated, yet the ratio should be much lower because the sensors are knocked out in the mutants. This indicates some unreliability in the data.
YLR121C encodes for a plasma membrane protease. This gene was down regulated in the mutant yeast. This may be because the mutant cells have recruited more internal methods of energy production instead of attempting to gather energy from the environment.
YPL075W encodes for a transcription factor that activates genes involved in glycolysis. This gene was very slightly down regulated in the mutants. This was a surprise because it was expected that this gene would be much more down-regulated due to the lack of glucose in the cell.
YKL060C encodes for Fructose 1,6-bisphosphate aldolase, which is required for gluconeogenesis. This gene was highly up regulated in the delta hexose mutants. This is because no glucose from the medium is present in the cell, so it is favorable for the yeast to make its one energy. This gives evidence that the delta hexose yeast compromise for there lack of hexose transporters by making glucose instead of moving it in.
YBR298C encodes for a high affinity maltose transporter. As seen in the figure, this gene was also highly up regulated, yet it the data also had a very large standard deviation. It would be expected that this gene would be up regulated in order to transport more maltose into the cell to compromise for the absence of glucose.
YGR289C encodes for another high affinity maltose transporter. This gene was also up regulated, but to a much lesser degree than the YBR298C maltose transporter. Yet this gene had a much lower standard deviation than the previous maltose transporter data.
YOL086C encodes for alcohol dehydrogenase, which is required for the last step in the glycolytic pathway. Alcohol dehydrogenase metabolizes ethanol into the energy rich molecule acetaldyhe. This would help compromise for the absence of hexose sugars available for energy. As expected, the gene was highly up regulated in the mutant yeast.
YDL052C encodes for 1-acyl-sn-gylcerol-3-phosphate acyltransferase, which is a key enzyme in lipid metabolism. This gene was up regulated with a fairly small standard deviation for the data. This was as expected because producing more of this protein would enable the cell to metabolize more energy from a source other than hexose sugars.
YMR008C encodes for phospholipase B, which also plays a part in lipid metabolism. Surprisingly, this gene was down regulated. It is possible that this enzyme acts to replenish and increase the size of the plasma membrane, which may not be a high energy usage priority for the delta hexose mutants.
YKL091C encodes for a lipid transport enzyme localized in the nucleus. This gene was slightly down regulated, but considering the standard deviation for the data the expression probably did not change much in the delta hexose mutants. It is possible that the YKL091C encoded enzyme is not directly involved in metabolism.
YDR155C encodes for an isomerase involved in protein degradation in metabolism. This gene was up regulated. Again, this may be because the mutant yeast cell is trying to harvest energy from alternative sources other than hexose sugars.
In conclusion, many more repetitions of this experiment coupled with computer analysis that can assess the entire spectrum of results are necessary for any speculations to be purported. Of the genes investigated, the manner in which they were explored proved to be the most rewarding part as we gained familiarity with analyzing gene expression patterns but found few significant trends to mention. Mircoarray analysis proved to be a very functional and legitimate way to study gene expression.