determining peak transcriptional response to drug in an organism
This is a stub for a longer post about how to finish the last experiment for my thesis.
(1) when strains grow in culture, they usually lag (period of no growth), and then they enter exponential or log-phase growth. Finally, nutrients may be depleted, and the culture will plateau, or reach stationary phase. If left to go too long, they will die. The problem here is determining when they are in log. Further, if you're going to do an experiment, you want to make sure that your strains are all in the same phase. That means you want to pulse them with drug, as well as sample, all while the cells are in the same physiological state growth-wise.
To determine when this occurs, you need to take a growth curve.
I did this by inoculating a large culture with a starting concentrations of 2 * 10^6 cells / mL of RPMI 1640 (a media meant to mimic blood). I then took an OD, or biomass estimation as determined by optical density, every hour until I had 3 readings at the same OD. This was initially a failed experiment, as my cells need to be grown overnight in minimal media, instead of rich media. Otherwise they become VERY sticky and stick together, and then sediment. This produced wonky growth curves.
I determined that all of my strains are safely in log-phase growth by the 5 hour mark.
(2) I then needed to determine at what concentration of drug my strains would be inhibited by. Inhibition is indicated by growth saturation that is 50% of non-drugged saturation for the same strain. To do this, I used 96-well plates, and tested 12 concentrations of drug against 8 strains at a time. I replicated the experiment by using overlapping concentration ranges (though not the same) for each strain I tested (24), and doing 3 biological replicates. The plates were inoculated to an OD (600 nanometers) of 0.01 in accordance with the CLSI / NCCLS lab standards for MIC-susceptibility testing.
(3) Knowing how much drug I needed for reach strain, and how long I had to grow each strain for, I needed to determine when the drug actually acts on my strains. I did this by doing a pilot experiment.
I grew 4 strains to mid-log, and then I spiked each strain with the MIC-value of drug I determined in step 2. Right before I spike in the drug, I took out some cells. I then pulled cells at 15 minutes, 30 minutes, 60 minutes, 90 minutes, and 120 minutes. I then chopped up the cells, took the RNA and made cDNA using RT, and I labeled the cDNA with one of two dyes. The Time-0 for each strain was labeled with a dye we use as a reference, and then each experimental time point was labeled with the other dye.
In parallel to this, I grew 2 of those 4 strains up and did the same experiment with DMSO, which is the chemical I suspend my drug in. The 2 strains I tested were the most and least drug resistant strains in the pilot.
I then hybridized each experimental time point to the mid-log time-0 control via a microarray.
I determined from this experimental that peak transcriptional response is between 60-90 minutes.
Now, the question is, would you do this pilot (step (3)) for ALL the strains you'd want to test for response, or is the pilot (1/6 of all strains to be tested) sufficient evidence that you can complete the experiment without testing response time in all strains.
I think it's sufficient to say that response is in the 60-90 minute window, and it is cheaper, faster, and no less accurate to profile gene transcription for all samples at that time point.
(1) when strains grow in culture, they usually lag (period of no growth), and then they enter exponential or log-phase growth. Finally, nutrients may be depleted, and the culture will plateau, or reach stationary phase. If left to go too long, they will die. The problem here is determining when they are in log. Further, if you're going to do an experiment, you want to make sure that your strains are all in the same phase. That means you want to pulse them with drug, as well as sample, all while the cells are in the same physiological state growth-wise.
To determine when this occurs, you need to take a growth curve.
I did this by inoculating a large culture with a starting concentrations of 2 * 10^6 cells / mL of RPMI 1640 (a media meant to mimic blood). I then took an OD, or biomass estimation as determined by optical density, every hour until I had 3 readings at the same OD. This was initially a failed experiment, as my cells need to be grown overnight in minimal media, instead of rich media. Otherwise they become VERY sticky and stick together, and then sediment. This produced wonky growth curves.
I determined that all of my strains are safely in log-phase growth by the 5 hour mark.
(2) I then needed to determine at what concentration of drug my strains would be inhibited by. Inhibition is indicated by growth saturation that is 50% of non-drugged saturation for the same strain. To do this, I used 96-well plates, and tested 12 concentrations of drug against 8 strains at a time. I replicated the experiment by using overlapping concentration ranges (though not the same) for each strain I tested (24), and doing 3 biological replicates. The plates were inoculated to an OD (600 nanometers) of 0.01 in accordance with the CLSI / NCCLS lab standards for MIC-susceptibility testing.
(3) Knowing how much drug I needed for reach strain, and how long I had to grow each strain for, I needed to determine when the drug actually acts on my strains. I did this by doing a pilot experiment.
I grew 4 strains to mid-log, and then I spiked each strain with the MIC-value of drug I determined in step 2. Right before I spike in the drug, I took out some cells. I then pulled cells at 15 minutes, 30 minutes, 60 minutes, 90 minutes, and 120 minutes. I then chopped up the cells, took the RNA and made cDNA using RT, and I labeled the cDNA with one of two dyes. The Time-0 for each strain was labeled with a dye we use as a reference, and then each experimental time point was labeled with the other dye.
In parallel to this, I grew 2 of those 4 strains up and did the same experiment with DMSO, which is the chemical I suspend my drug in. The 2 strains I tested were the most and least drug resistant strains in the pilot.
I then hybridized each experimental time point to the mid-log time-0 control via a microarray.
I determined from this experimental that peak transcriptional response is between 60-90 minutes.
Now, the question is, would you do this pilot (step (3)) for ALL the strains you'd want to test for response, or is the pilot (1/6 of all strains to be tested) sufficient evidence that you can complete the experiment without testing response time in all strains.
I think it's sufficient to say that response is in the 60-90 minute window, and it is cheaper, faster, and no less accurate to profile gene transcription for all samples at that time point.