Microscopy!

Today (November 8th), we worked with cells from the blood-brain barrier using microscopy! By definition, microscopy is "investigation using a microscope." To refresh my knowledge of the microscopy processes we use, JP gave me this link to read. In our work, we used Alexa 488 and Hoechst 33342 (also known as DAPI). Before I arrived, JP fixed the cells and added the primary antibody. We had to decide whether or not to permeablize the cells before adding the DAPI. We decided to try imaging the cells without permeablization first. We added a DAPI solution to the sample, incubated the cells, and then removed the excess solution. We then washed the sample a couple times to remove excess dye.

Once the sample was ready to be imaged, we had to decide what wavelengths of lasers to use to image them. To do this, we used a program called SpectraViewer. The resulting graphs are shown below.

 The vertical line on the graphs represents the wavelength of the selected laser. The green curve represents the Hoechst 33342 and the red curve represents the Alexa 488. The shaded regions represent the region that will be recorded. We decided to use the 405 nm laser for the Hoechst 33342 and the 488 nm laser for the Alexa 488.

To image the cells, we used a Zeiss Spectral Confocal Microscope (seen in the image below). It was so cool to be able to see the cells magnified to such a high degree!

The nuclei were very clearly visible, but the staining from the Hoechst 33342 was not visible. After viewing the images, we decided that it would be necessary to permeabilize the cells and re-add the Hoechst.

I won't be able to go to RPI next week because I will be in Kansas City for my national championship horse show, but I can't wait to return on the 22nd to see the images from the cells and continue our work!

Analyzing Peptide Mapping Data

On Friday (November 1st), I did a lot of work with analyzing peptide mapping data using the Origin program. But before I get into that, JP said that the peptide formation that I prepped for last week came out perfect! Anyway, I worked with three different sets of data from the same experiment- one set of 20 peptides and two sets of 56. After copying the data from a previous Excel sheet, my first step was to change the settings of all of the odd-numbered columns to "Y-error" because it represented standard deviation data. After I had done that for all three sets of data, I adjusted the fit curves by using a one-site bind pharmacology non-linear cure fit. In that process, I had to change the k-values of every data set to 1 to standardize the fits.

 Our goal is to look at the data for each of the peptides and determine which peptides have low Kd (affinity towards the substrate) values and high R-squared values (good fit to the best-fit line). The area we are looking for is represented by the green area above. In looking at the plots, I narrowed the range to an R-squared value of 0.5 to 1 and Kd value of 3E-6 to 8E-6, and changed the data point labels to display their peptide numbers. Looking at the layout of peptides in this region, we determined that the most accurate set of data is the second set of 56, which makes sense because those peptides were made in 2012, while the peptides used in the other two sets were made in 2011.

After looking at those plots, I the plotted Peptide # vs. 1/Kd. This produced plots with many sporadic troughs and peaks. JP introduced me to the idea of curve-smoothing. Using the Kd values of the peptides, we are ideally looking for the Kd values of individual amino acids. To do this, we use averages of the Kd values of different peptides. For example, A would equal KD1, B would equal the average of KD1 and KD2, etc.

Using an Excel template that JP made for a previous experiment, I plotted averages of three up to fifteen peptides. While the fifteen-peptide averages obviously smoothed the curve the most, it is hard to determine the correct plot we should use because as the curve gets smoother, it hides certain important peaks.

After I finished my analysis work, JP showed me some of the brain cells that he is culturing! He just started growing them on Tuesday, so they had not yet grown into a full layer of cells. They will be used to run a second trial of an experiment we did last winter, which involves testing the permeability of the blood-brain barrier by passing different sizes of sugar molecules through a layer of cells. I can't wait to continue working with the data and see where that experiment goes!