This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Bi1X – Cellular Decision Making Measuring Gene Expression 2010 Introduction: If you had to choose between broccoli and chocolate, would you eat both simultaneously or one at a time? Which one would you eat first? How and what do you think a mouse or a fly would choose? What about a bacterium? According to the famous Monod’s experiment, E. coli can choose to eat a certain sugar even when multiple types of sugar are present in the growth media. Today, we are going to focus on one of the most well known examples—the lac operon. The lac operon controls the production of three different genes associated with the metabolism of lactose. One of these genes is lacZ, which encodes for the enzyme beta‐galactosidase which metabolizes lactose. When the growth media has glucose and lactose E. coli will not produce any beta‐galactosidase. It will only produce it once it has consumed all the glucose. This means that the cell has a way of determining if there is no glucose and if there is lactose in the media. In this experiment we will explore the part of the circuit that takes care of detecting the presence or absence of lactose in the media. Instead of lactose, which would get eventually eaten by the cells, we use the inducer IPTG. This small molecule interacts with Lac repressor in much the same way that the real sugar does except that it cannot be cleaved by the enzyme beta‐galactosidase, making it a useful experimental substrate. In the absence of lactose or IPTG, Lac repressor binds to an operator that overlaps the lac promoter. Because of this overlap, RNA polymerase cannot bind to DNA to initiate transcription, resulting in repression. B.
A. Fig 1. Repression in the lac operon. (A) A RNA polymerase molecule binds to its promoter on the DNA to initiate transcription. (B) When a repressor molecule binds to its operator, it blocks part of the promoter and leads to repression. However, no response in biology is really digital. There is no such thing as something being “on” or “off”. You can instead ask an analog question: to what extent is something off? The Construct: In this experiment, we will use a synthetic version of the lac operon. This DNA construct has a single binding site (operator O1). The lacZ gene was replaced by a YFP gene controlled by a promoter called lacUV5 (a mutant of the wild type lac promoter). Fig 2. The construct. Sequence with a single operator O1. The promoter lacUV5 controls the expression of the YFP gene. Measuring gene expression: We will measure the fluorescence expressed by cells containing the construct described above for different concentrations of IPTG. We want to define a fold‐change in gene expression, a measure of the relative change in gene expression upon the addition of inducer. One problem with measuring fluorescence in cells is that the cells are fluorescent themselves! This means that if you measure the fluorescence of a cell expressing YFP you will actually be measuring the fluorescence of the YFP molecules on top of the autofluorescence of the cell. In order to account for this you can measure the autofluorescence of a cell that lacks any YFP. We will define the fold‐change in gene expression as /#0"([+,-. ]) 1 203"/#0"
!"#$‐%&'()*([+,-. ]) = /#0"([4'%+] = 0) 1 203"/#0" Here, AutoFluo is the autofluorescence and Fluo([IPTG]) is the fluorescence in cells grown in the presence of different amount of inducer and Fluo([LacI]=0) is the fluorescence of the cells that don’t have the repressor. Note that Fluo([IPTG]) refers to the total integrated fluorescence in a cell. This is because we are interested in the total gene expression in a cell. How will you correct for autofluorescence, given that cells have different areas (i.e., how should you calculate AutoFluo)? In order to quantify the level of gene expression you will need to write code in MatLab or other language of your choice. You will have to find the cells automatically and use the mask in order to integrate over the fluorescence of each cell. Experimental protocol: What happened behind the scenes: Five hours before you started your experiment we diluted cells from an LB overnight culture into minimal media in the presence of different concentrations of IPTG. This particular minimal media has salts, a carbon source and some amino acids. We also prepared agar pads now made with PBS (phosphate buffer). Cells cannot grow in this medium. The idea is that we will be taking snapshots of their state when we took them out of the culture. Due to time limitation, we had mounted the samples on the Wilco dishes for you. For each construct, there are 10 samples: 5uM, 10uM, 20uM, 50uM, 100uM, 200uM, 1mM, 2mM, NoLacI, and NoFluo. NOTE: Remember to keep track of what sample is on which pad! This is extremely important so that you can convert your images into a curve which reflects the level of gene expression as a function of the inducer. Your mission: 1) Set up the plate on the scope. We will take the images at room temperature. 2) Set up Micro‐Manager to take a picture of Brightfield (with phase contrast) and FITC or YFP (green/yellow fluorescence) at the same time (Ask the TAs about Multi‐D Acquisition if you are not sure how to use it). Decide on what exposure to use for your snapshots by looking at the brightest sample, the one with NoLacI. Make sure there is no saturation in the fluorescence channel! 3) Take 5 snapshots of the NoFluo control and of the NoLacI sample. You want to shoot for having more than 100 cells per sample. These two are the most important samples as they’re fundamental to calculating the fold‐change. That is why we want to make sure right from the beginning that everything went OK. 4) Now, take 5 snapshots on each of the remaining strains. Make sure to save all your data! Assignment: In your report answer all the following questions. Remember that independently of the particular questions being asked here a report should be enough for somebody to reproduce your experiment. This means that you need to include an introduction with a motivation, a methods section and a conclusion outlining what you learned. Ask the TAs for help with this if you need it! 1) Write a code to find the cells in the phase contrast image. Use the function regionprops to explore properties such as MajorAxisLength, MinorAxisLength, Area, etc. It is also a good idea to play with the ratio of these two magnitudes, as they set some bounds on the shape of the object. Explain the reasoning behind any of the thresholds that you draw using the regionprops properties. 2) Make and show an overlay of one of your favorite phase contrast pictures and the mask (Hint: explore the function cat and bwperim). 3) For each cell, overlay its corresponding mask with the fluorescence image and calculate the total fluorescence per cell. For example, if you want to choose cell number 4 in a labeled image called ImLabel (the one created by the function bwlabel) you can do >> ImCellFluo = immultiply(ImLabel==4,ImFluo); where ImFluo is your fluorescence image. The total (uncorrected) fluorescence of the cell is >> sum(sum(ImCellFluo)); and the area can be found using regionprops. Note: When calculating the total fluorescence per cell, make sure to subtract the autofluorescence and background (these steps are not shown above). Ask for help if it is not clear to you how to do this. 4) Finally, calculate the fold‐change for each one of your samples for both constructs. Use the function mean to calculate the mean level of fluorescence corresponding to each strain. Don’t forget to include error bars! You can obtain a standard deviation by using the function stdev and a standard error by combining your standard deviation with the length of the vector (using the function length). 5) Plot fold‐change vs. IPTG concentration. Do it on a linear and a log‐log scale. Try to fit these plots to the following formula (Hint: explore the functions fittype and fit): [+,-. ] (
5×6 7 8
!"#$‐%&'()* = [+,-. ] (
78 where ! is the maximal expression level of the promoter, K is a dissociation constant, and n is called the Hill coefficient. Be sure to include error bars in your plots and explain possible sources of the errors. What are the parameters corresponding to the best fit? What features of the curve change as you change the value of n or the value of K? ...
View Full Document
This document was uploaded on 01/03/2012.
- Fall '09