I study Biochemistry with a Year in Industry/Research at Imperial College, and this year I'm doing my year in Research. Since September, I've been working in James Briscoe's lab in the NIMR. The lab's main focus is the development of the vertebrate neural tube, specifically, the patterning of the ventral area of the neural tube. We study the signaling mechanisms by which each region of the neural tube develops into different types of cells. One of the main proteins involved is my favourite protein: Sonic Hedgehog protein. I discovered it by accident during my first year at Imperial. I was flicking through the index of Alberts, looking for something or other during revision for my first year exams, when I came across an entry called Sonic Hedgehog protein. I thought this was quite funny, so I decided to check out what it was. I didn't understand a thing and soon dropped it and went back to revision (or so I tell myself). During our second year we studied the signaling pathway triggered by Sonic Hedgehog protein in Molecular Cell Biology II, and this is when I actually became interested in it seriously.
I'm going to try to explain a bit about the Sonic Hedgehog (Shh) pathway in the neural tube and a bit of what I'm trying to do in the lab in this placement year. First, let's start with a bit of anatomy and nomenclature. An embryo has three axis: the anterior-posterior axis, which goes from the head (anterior) to the tail (posterior); the dorso-ventral axis, which goes from the back (dorsal) to the stomach/chest (ventral); and the lateral or left-right axis (pretty self explanatory, I would I assume). I will generally refer to the dorso-ventral axis, since this is the direction in which Shh effects are most obvious. Shh is produced initially in the notochord (see figure below), which is ventral to the neural tube. From there, it reaches the cells in what is known as the floorplate of the neural tube, the most ventral region of the neural tube. These cells then start to produce Shh too. The rest of the cells in the neural tube don't produce Shh, so a concentration gradient of Shh is established, with higher concentrations on the ventral area of the neural tube, and lower concentrations in the dorsal side. Shh doesn't seem to have an effect on cells that are dorsal to the midline of the neural tube.
The Shh pathway is quite complex, but I can simplify it to essentially four proteins: Shh, the signaling protein; Patched (Ptc), the receptor protein, to which Shh binds; Smoothened (Smo), the first effector protein; and Gli, the final effector protein.
In the absence of Shh, Ptc inhibits Smo. Gli is part of a multiprotein complex that targets it for cleavage in the proteasome, giving place to a smaller protein, GliR, which diffuses into the cell nucleus and binds DNA, stopping the expression of genes. In the presence of Shh, Shh binds Ptc, so Ptc can't inhibit Smo. Smo acts to disassemble the complex that targets Gli for cleavage into GliR, so Gli is processed in a different way and gives place to GliA. GliA also diffuses into the nucleus, but instead of stopping gene expression, it promotes it. Generalizing, the presence of Shh leads to the expression of a set of genes. Different concentrations of Shh lead to different genes being activated due to the different concentrations of GliA or GliR produced (although the proportion of GliA to GliR doesn't seem to be as important as the different binding strength of each version of Gli). This is all further complicated by the fact that there are 3 Gli proteins, and by the fact that we don't fully understand how the Shh gradient is established.
Now that I've established a bit of anatomy and the basics of the Shh pathway I can continue and explain what I'm trying to do. Currently, a lot is understood about the gene regulation involved in the Shh pathway, most of the genes activated by Shh are known and some of the regulatory networks that link these genes are understood (the genes activated and repressed by Shh go on to repress and activate other genes, and each other, which leads to networks like the one described here), but not much is known about cell growth and behaviour in response to the Shh gradient, so that's what I'm trying to study.
It started off with me learning how to electroporate DNA into chick embryo neural tubes. Electroporation is a fairly tricky technique to master, but it's entirely doable with practice (if I can do it, anyone can, trust me). Basically, you incubate eggs for as long as you need (in my case it's usually 45 hours) and once they're incubated you make a hole in the shell to take out some of the albumen (egg white for the layman) out of the eggs. Once this is done, you take a pair of scissors and cut off a piece of the shell.
Once you have the egg open you can put it under the microscope and find the embryo. A very thin needle made from a glass capillary and a mouth pipette are needed at this point. The needle fits into the mouth pipette and you now have a system to suck liquid into the needle and then blow it into anywhere you stick the needle into. You suck DNA into the needle, and then stick it into the chick embryo's neural tube, which (to the trained eye) is pretty visible. Then you blow the DNA in. Once this is done, you apply electrodes to both sides of the neural tube and apply a charge so that the negatively charged DNA moves to the positive electrode and the cells on that side of the neural tube take up the DNA. At this point, you cover the hole on the egg with either transparent tape or parafilm and let it grow for a few hours until the cells that took up the DNA are expressing the protein coded in that DNA.
I started my experiments off by electroporating H2BGFP DNA into the cells. H2BGFP is a protein that is a fusion of the histone protein H2B (histones are proteins that bind DNA and keep it organised) and green fluorescent protein. This protein localises to the nucleus of the cell, so that if you electroporate the neural tube cells with it, their nuclei (approximately in the centre of the cell) will emit green fluorescence. Once the embryos had fluorescent neural tubes, I cultured them so they would grow on plates, and then I took them to the multiphoton microscope. A multiphoton microscope uses two low-energy photons to excite fluorophores. Usually, to excite a fluorophore you need to illuminate it with photons of shorter wavelength (higher energy) than the fluorescence emission you are trying to produce, but multiphoton fluorescence allows two long wavelength (low energy) photons to arrive at the sample at the same time, having the same effect as one shorter wavelength photon. Multiphoton microscopy allows imaging of sections of a sample (if you had a column, with a confocal microscope or a multiphoton microscope you could image different planes of that column), and because the photons used have relatively low energy, they won't damage samples as much as normal fluorescence microscopy would. This makes multiphoton microscopy especially good for in vivo imaging. This allowed me to make movies of the cells electroporated with H2BGFP, which, once processed, will hopefully allow me to track the cells in the neural tube and see how they behave as the neural tube grows. As time wore on, however, we realized that the culture method needed to make these movies didn't guarantee high survival of embryos, and it also produced a lot of drift (the embryo drifted outside the microscope's field of view while we were making the movies, which usually lasted a whole night), so we decided to change the experiments. Instead of using H2BGFP, we started to use Kaede protein. Now, I said before that Sonic Hedgehog protein is my favourite protein. Kaede is a close second. The name means "maple tree" in Japanese, and this i s why: Kaede emits green fluorescence, much like GFP, but unlike GFP when it is illuminated with UV light this fluorescence shifts from green to red. This change in fluorescence is called photoconversion. This property makes it a good protein to track cell behaviour, because if you manage to only photoconvert a small number of cells from green to red you can image a few hours later and see where these red cells are. The main problem with Kaede is that as long as the embryo is alive (and we want it to be alive so we can image the growth) it produces green Kaede protein, and the red Kaede protein is degraded, so if enough time has gone by, it will completely disappear and the marked cells will be lost.
At the moment, I electroporate Kaede, photoconvert a region of the neural tube to red using UV light, and take images right after photoconversion and 8 to 18 hours later. The images look quite good at the moment, and I'm happy with what I'm obtaining, but I'm not making movies at the moment. In any case, I should get back to my presentation...
Now that I've established a bit of anatomy and the basics of the Shh pathway I can continue and explain what I'm trying to do. Currently, a lot is understood about the gene regulation involved in the Shh pathway, most of the genes activated by Shh are known and some of the regulatory networks that link these genes are understood (the genes activated and repressed by Shh go on to repress and activate other genes, and each other, which leads to networks like the one described here), but not much is known about cell growth and behaviour in response to the Shh gradient, so that's what I'm trying to study.
It started off with me learning how to electroporate DNA into chick embryo neural tubes. Electroporation is a fairly tricky technique to master, but it's entirely doable with practice (if I can do it, anyone can, trust me). Basically, you incubate eggs for as long as you need (in my case it's usually 45 hours) and once they're incubated you make a hole in the shell to take out some of the albumen (egg white for the layman) out of the eggs. Once this is done, you take a pair of scissors and cut off a piece of the shell.
Once you have the egg open you can put it under the microscope and find the embryo. A very thin needle made from a glass capillary and a mouth pipette are needed at this point. The needle fits into the mouth pipette and you now have a system to suck liquid into the needle and then blow it into anywhere you stick the needle into. You suck DNA into the needle, and then stick it into the chick embryo's neural tube, which (to the trained eye) is pretty visible. Then you blow the DNA in. Once this is done, you apply electrodes to both sides of the neural tube and apply a charge so that the negatively charged DNA moves to the positive electrode and the cells on that side of the neural tube take up the DNA. At this point, you cover the hole on the egg with either transparent tape or parafilm and let it grow for a few hours until the cells that took up the DNA are expressing the protein coded in that DNA.
I started my experiments off by electroporating H2BGFP DNA into the cells. H2BGFP is a protein that is a fusion of the histone protein H2B (histones are proteins that bind DNA and keep it organised) and green fluorescent protein. This protein localises to the nucleus of the cell, so that if you electroporate the neural tube cells with it, their nuclei (approximately in the centre of the cell) will emit green fluorescence. Once the embryos had fluorescent neural tubes, I cultured them so they would grow on plates, and then I took them to the multiphoton microscope. A multiphoton microscope uses two low-energy photons to excite fluorophores. Usually, to excite a fluorophore you need to illuminate it with photons of shorter wavelength (higher energy) than the fluorescence emission you are trying to produce, but multiphoton fluorescence allows two long wavelength (low energy) photons to arrive at the sample at the same time, having the same effect as one shorter wavelength photon. Multiphoton microscopy allows imaging of sections of a sample (if you had a column, with a confocal microscope or a multiphoton microscope you could image different planes of that column), and because the photons used have relatively low energy, they won't damage samples as much as normal fluorescence microscopy would. This makes multiphoton microscopy especially good for in vivo imaging. This allowed me to make movies of the cells electroporated with H2BGFP, which, once processed, will hopefully allow me to track the cells in the neural tube and see how they behave as the neural tube grows. As time wore on, however, we realized that the culture method needed to make these movies didn't guarantee high survival of embryos, and it also produced a lot of drift (the embryo drifted outside the microscope's field of view while we were making the movies, which usually lasted a whole night), so we decided to change the experiments. Instead of using H2BGFP, we started to use Kaede protein. Now, I said before that Sonic Hedgehog protein is my favourite protein. Kaede is a close second. The name means "maple tree" in Japanese, and this i s why: Kaede emits green fluorescence, much like GFP, but unlike GFP when it is illuminated with UV light this fluorescence shifts from green to red. This change in fluorescence is called photoconversion. This property makes it a good protein to track cell behaviour, because if you manage to only photoconvert a small number of cells from green to red you can image a few hours later and see where these red cells are. The main problem with Kaede is that as long as the embryo is alive (and we want it to be alive so we can image the growth) it produces green Kaede protein, and the red Kaede protein is degraded, so if enough time has gone by, it will completely disappear and the marked cells will be lost.
At the moment, I electroporate Kaede, photoconvert a region of the neural tube to red using UV light, and take images right after photoconversion and 8 to 18 hours later. The images look quite good at the moment, and I'm happy with what I'm obtaining, but I'm not making movies at the moment. In any case, I should get back to my presentation...
i´ll try again on week end. too many trees to understand how it works por
ReplyDelete