Tip:
Highlight text to annotate it
X
Hi. I'm Nico Stuurman.
I work as a research specialist in the lab of Ron Vale, here at UCSF.
In the lab, one of my roles is to help everyone with microscopy and
set up new microscopes, and as such I am doing a lot of fluorescence microscopy.
Fluorescence microscopy is really a wonderful tool to let you look at the insides of cells,
either fixed or alive.
And just to give you a glimpse of the marvels that we see in lab regularly,
I actually was at the microscope and made a few movies that will
appear here during the talk.
And so one of these, here... what you see here
is a fruit fly cell that has been labeled with fluorescent proteins
that were coupled to tubulin, and you see that here in red.
So, this cell has some of those microtubules in red.
And they were also labeled with GFP coupled to a protein
called EB1. That EB1 are these green little arrows here.
EB1 sits on the tips of microtubules,
and you will see, when I start playing this movie,
that EB1 -- the tips of microtubules -- move very fast
throughout the cell and allow you to, here, see the
dynamics of this living cell.
So, this just gives you an idea of why
we so highly value fluorescence microscopy in the lab.
So, fluorescence -- we like the extreme high contrast
that you can get. So, you can label very specific components,
and you can then set up your experiment so that you have this very dark background
and a very high signal.
We're looking here at, again, a fruit fly S2 cell
that has been labeled with actin in red,
in green for microtubules, and in blue for DNA.
And that also goes to show that we can label these specific components
with very high specificity. We can use multiple colors simultaneously
to show these different components of the cell.
And we'll get back during this talk to how we exactly do that
and what kind of dyes we use.
Then, as you saw in the introduction, we can also
use fluorescence in live cell imaging
by virtue of fluorescence proteins, and these really give spectacular insights
into how proteins behave -- where they are localized over time -- in living cells.
And lastly, fluorescence is highly quantitative.
So, by virtue of CCD cameras that have a very linear range,
we can actually quantitate fluorescence very precisely, and in the end,
you can actually count how many molecules there are.
So, fluorescence is this great tool.
What is it, really?
And, it pays to go back and figure out what fluorescence means.
One thing that surprises me is that
fluorescence has been known to mankind only about 150 years or so.
And one of the first occurrences in the literature was this description
by Sir John Frederick William Herschel about 150 years ago.
He was a British scientist, and he was studying at the time extracts that he made
from the bark of a South American tree.
The bark was also used to treat malaria.
He made a watery extract of it, held it against the light in his study,
and noted that it gave this bluish color.
And he described that as, "Though perfectly transparent and colorless,
when held between the eye and the light, it yet exhibits in certain aspects
and under certain incidences of the light an extremely vivid and beautiful celestial blue color."
Now, we cannot fully recreate this experience of Sir John Frederick William Herschel,
because the bark of that South American tree is pretty hard to come by.
But it so happens that the active compound, which is quinine,
can actually be found in any bottle of tonic water.
So, the tonic water, if you would read carefully here,
it says, "Contains quinine."
And, so we will now demonstrate to you the fluorescence
that Sir John Frederick William Herschel experienced.
Ok, we've now got the UV illuminator up here.
So, we have the quinine in the tonic water.
This is a UV light source, so this will replace the sunlight
and will hopefully glow up the quinine in the tonic water.
I will wear a face mask, since I'm close to the UV light source,
and I don't want my face and eyes to get damaged by the UV.
So, we'll now pour the tonic water into the glass.
By the way, as a control, we have a beaker with water.
And we will switch on the UV illuminator.
And you see here, already the blue fluorescence coming from the quinine.
And, as you see, it's no trick -- the water is not fluorescent.
As in all fluorescence experiments, to see the fluorescence better,
you will need to reduce the background, and we just did that
by dimming the background light in the studio
to make more clear this blue fluorescence from the quinine.
You actually even see that it's more blue here in the beginning.
So, there's so much quinine in here that it absorbs the light
-- the UV light -- so that there's less excitation up here
than here in the beginning,
which is also what Herschel described in his initial description.
So, you now have seen fluorescence.
And what we have noted is that we are exciting
the quinine -- and this is the formula for the quinine molecule.
We were exciting it with this UV light.
The UV light has relatively high energy -- it's low wavelength --
and the quinine emitted light that was blue,
which was lower energy and higher wavelength.
So, there are certain wavelengths of light that excite the quinine,
and there are certain wavelengths of light emitted by the quinine.
The general rule is always that you go from higher energy
(lower wavelength) to lower energy (higher wavelength).
There are certain exceptions to this rule that we will not get into now.
So, when you now plot this as a spectrum,
where we have the wavelength of the light on the X axis
and the excitation or absorption on the Y axis,
as well as the intensity of the emitted fluorescence,
then you will see that, in these spectra, there are certain maxima,
which we call the excitation maximum
and the emission maximum.
The difference between that excitation and emission maximum
is called the Stokes shift, again named after a famous British scientist from the last century.
So, we talk about dyes that have a small Stokes shift,
and we talk about dyes that have a large Stokes shift.
So, this one looks relatively large.
There's a significant difference between the wavelength of the light
for excitation and for emission.
Then, to think a little bit further about what is happening -- what is going on --
in these dyes during fluorescence,
the so-called Jablonski diagrams are very useful.
Jablonski was a Polish professor from the beginning of the 20th century,
and so he draws these energy states.
So what you can think of is that there are electrons in these complex organic molecules
that can be excited -- they can absorb light --
and thereby appear in a higher energy state.
This absorption happens at a very small timescale.
So it happens on the order of 10^-15 seconds.
Now, times like that are kind of meaningless to me.
Something below milliseconds... maybe microsecond...
I don't have any natural feeling of what this means.
So, I like to express this in terms of the distance travelled by light during that time.
I do know that light goes very, very fast
3x10^8 meters per second.
So, when you realize that during the time that this photon is absorbed,
light travels only 0.3 microns, you get a bit of an impression of
how short that timescale is.
Now, once the molecule is in this excited state,
they fall back to the lowest level of the excited state
through a process called internal conversion.
That happens on the order of the timescale of 10^-12 seconds
or light travels about 0.3 mm in that time.
So, it's a much longer process.
But then, the electron will wait here for a considerable amount of time,
something on the order of 10 nanoseconds
(so light would actually travel about 3 meters in that time)
and then it can emit a photon.
So, in this case, you get now an impression of why
that emitted photon is actually of lower energy (higher wavelength)
simply because there is energy lost in all of these conversions.
Not always will the dye emit light once it's in the excited state.
So, energy can also be transferred to other molecules,
or it can lose energy through non-radiative decay.
So, that means that not for every photon we get an emitted photon out.
And therefore, that ratio can be expressed as the quantum efficiency.
The quantum efficiency is simply the ratio of absorbed and emitted photons.
And in general, we like dyes to have a very high quantum efficiency,
so that photons that you pump in also come back out again.
Even more general, we can say something about the brightness of a dye,
and of course, we'd like dyes to be bright because that makes them more useful
and easier to work with.
And the brightness is determined both by how well
the dye absorbs photons (how well the excitation works)
and by the QE. So in general, we like dyes that have
a very high absorption coefficient and that have a high quantum efficiency.
So that now brings us to the fluorescence microscope.
And this is a very schematic layout of the essential components in a microscope.
First of all, we need a light source that is capable of delivering
light needed for excitation.
So, we can use things like arc lamps, lasers are used for certain applications,
nowadays LEDs are becoming more and more available.
In general, you need something that's pretty bright.
Then, we hit these filters.
So, we need, first of all, a filter that only lets through the excitation light.
The excitation light then bounces off a dichroic mirror.
So, this mirror reflects the excitation light.
It goes through this objective to the sample.
Then, the emitted fluorescence passes back through the same objective lens.
It now hits again the dichroic mirror,
but the dichroic mirror is made such that the emitted fluorescence actually passes through.
And then, there's an emission or blocking filter
that will block all the excitation light that happens to make it this far,
but does let through all the emitted fluorescence
so that you can see it with your eyes or on the camera.
These three filters -- excitation filter, dichroic mirror, and emission filter
are often combined in a cube.
So, we have here, these fluorescence cubes that in one
unit contain the dichroic mirror, excitation filter,
and the emission or barrier filter.
We'll now to go the lab so that I can show you a real microscope,
and we'll show you a real fluorescence cube.
So we just now saw the schematic overview of a microscope,
and here's the real thing.
So, the light source that I talked about, in this microscope,
is sitting here. This is an arc lamp.
It's actually a mercury arc lamp, and I have one of those bulbs here.
This one has been used. So, these mercury arcs give an incredible amount of light.
Very, very bright.
Inside the microscope, the light then travels through here,
hits here the dichroic mirrors
that are sitting inside these fluorescent cubes,
and I'll show those in a second.
Then, there's a shutter here that blocks the light,
and I can open the shutter with a button on this microscope.
And so here we now see blue light coming out.
I'll move this away to make it a little bit more visible.
I can then move the fluorescent cubes.
Green light with a different filter set in there. Red. This is UV, which you don't see very well.
And, back to blue light.
So now, let's take a second to look at these fluorescent cubes.
In this microscope, they are seated inside this turret.
And this is actually very nicely motorized so that I can press a button to move it around.
I can open up the turret, and here you see the cube.
So, the cube is sitting in the microscope in this way. The light comes in here.
The excitation filter hits the mirror, goes up, comes back,
and here in the bottom is the emission filter.
And you can see that better when I take out one of these cubes.
So, we have here, the excitation filter. There's the mirror inside,
and here on the bottom is the emission filter.
I have another type of cube that's a little easier to take apart.
So, here you can get a view of those filters themselves.
So, this would be an emission filter, and here you should see that it has a very clear color.
So, this is really the heart of fluorescence microscopy -- filters that separate very well
the excitation light from the light emitted by the fluorophore.
We've now seen what this microscope looks like, and we've seen some of these actual filters.
And I want to explain to you how these filters actually work.
So, the trick is that, on a glass substrate,
several layers are deposited, some of which are reflective and others that are transparent.
The distance between these layers is something on the order
of a half or a quarter of the wavelength of the light that you want to pass through.
So what now happens within this filter,
the incident light will bounce back and forth.
When the wavelength of the light matches the distance between these layers,
you will get constructive interference, and the light will pass through.
Versus when the wavelength of the light does not match the distance of these layers,
you will get destructive interference, and the light will reflect.
And that is how all these filters are made.
Enormous advances have been made in construction of these filters
over the last twenty or thirty years.
You can now get filters that have very impressive spectra that let through
only a very narrow band of light with very high efficiency.
So, thanks to these great filters, we are actually capable
of reducing the background fluorescence
to almost absolutely zero, and we can literally see single dye molecules
even with our naked eye.
So, this is what the spectra of these filters look like.
So, we have, for instance, an excitation filter from this filter cube,
where there are certain wavelengths of light that are passed through.
The dichroic mirror will the reflect the excitation light
and will transmit the emitted light, and in this case, we have a bandpass filter
that only lets through the emitted fluorescence.
So the combination of these three elements actually lets you see the fluorescent light.
Now, when you set up a microscope,
you will need to take care that you match the spectra of the filters
to the spectra of the fluorophores that you're working with.
So, we have here a spectrum of a dye -- this would be the excitation,
and here is the emission.
And this is probably something like Cy3, which emits in the red.
And we see here that when we put on this band filter here
that is centered around 650 nm
that we do get fluorescence coming through, but you also see from this
that we're only picking up a relatively small fraction of the emitted fluorescence.
So, all this fluorescence here, we will simply not see in our microscope,
and it's all photons lost,
which is a total waste.
So, it will be much better in this case to select a filter
that is more centered around something like 620 nm.
And this, as you see, will pick up 75% or so of the emitted fluorescence,
and so will therefore give you much higher throughput
and more signal.
So when you set up your experiment, you will have to think about the spectra of your dye --
of the dyes that you use and the filters that you use.
There are a couple of online resources that let you look at this,
and those will be available in the supplemental materials to this talk.
There are some other ways of setting up your experiment.
So, instead of putting all the filters in these filter cubes,
we can bring out the excitation filter, and the same can actually be done for the emission filter.
We can bring that out into a filter wheel.
The advantage of a filter wheel is that it can switch, usually, much faster
than the filter cube.
Also, when we put another dichroic mirror in place, under the microscope,
you will actually see that it's very difficult to get that mirror
at exactly the same angle in the microscope
and that there will often be slight shifts in the image.
So, using a filter wheel both allows you to switch faster
between wavelengths and also will keep the image much better in place.
The way this now works, over here...
We have an excitation filter in place -- in this case, the blue filter
is sitting in front of the light source,
so we're getting blue light out.
Green fluorescence, the green fluorescence passes through this dichroic mirror.
Then we switch the excitation filter.
We put the yellow filter in place, we get red fluorescence,
and that red fluorescence also passes through the dichroic mirror.
And in this case, through this multi-bandpass emission filter.
And then, in the computer, we can combine these different colored images
to then give you these multi-colored images.
There are certain requirements of the dichroic mirror.
So, it needs to be a multi-bandpass filter.
It will need to reflect the blue light with which we were exciting the green.
It will need to reflect the yellow light with which we were exciting the red.
But it will need to transmit the green fluorescence and the red fluorescence.
And so, the filter manufacturers are capable of making dichroic mirrors
that really can do these very complicated things, and that have very complicated spectra,
but still transmit with very high efficiencies.
Now, one of the main concerns when we are doing fluorescence imaging
is photobleaching. And, I illustrate this in this movie, here.
Again, this is green fluorescent protein coupled to EB1.
And you see now that while we're talking that the fluorescence is slowly disappearing.
And, so basically, any fluorescent molecule, when it is in the excited state,
there is a certain probability that the molecule will self-destruct
or that it will go into a dark state.
And so that means that the longer we look at something,
the more molecules will become invisible and will photobleach.
And, as you see here, that can go pretty fast.
So, what can we do about this? How do we deal with this problem?
There are a couple of approaches.
One is to select dyes that resist photobleaching
as much as possible.
As I'll explain in a little bit, there is an enormous number of dyes that you can choose from.
And so, you want to choose the ones that photobleach the least.
Then, you can try to just put as much label on your molecule of interest as possible
to make it brighter, so that you have more molecules there before they're all photobleached.
You can use anti-fade compounds.
So, often, molecular oxygen is involved in these photobleaching processes,
and so by depleting this molecular oxygen, you can reduce
photobleaching. Even something as simple as adding glycerol can decrease photobleaching.
And then very importantly, you want to really budget the photons
that you will get out of your dye.
So, you only want to have the shutter open while you are actually observing your sample.
So, most beginning microscopists don't really think about this,
and they just open the shutter, and they're exposing their sample
continuously with this bright light
All experienced microscopists will get very nervous about that
and will want to immediately close the shutter.
So, really, you only want to expose while you're taking data.
Otherwise, it's all photons gone to waste.
Likewise, you will want to play with your exposure time and excitation power
and keep that as low as possible so that you will get all the information
you need from that limited number of photons that will come out of your dye.
So now, we'll go a little bit through the dyes that are available,
what they are and how we use them.
So, the kind of classic dyes are this coumarin in blue, fluorescein in green,
and rhodamine, which is a red dye.
These dyes have been known -- fluorescein and rhodamine for sure --
for more than a century.
Fluorescein actually was used to probe whether the Rhone and Danube river in Europe
are connected. So, it just goes to show
with what high efficiency scientists have been able to detect these fluorescent molecules.
So what happened was that the sources of the river Rhine were spiked with
fluorescein, and then that fluorescein was then traced in the Danube river
demonstrating that the two river systems are connected.
Also, fluorescein is actually the compound that is used to dye the Chicago river green.
So, during Saint Patrick's day in Chicago...
During St. Patrick's day, everything in the US has to be green,
including the Chicago river, and it is this dye, fluorescein,
that is being used to stain the river green.
You can see here that fluorescein is actually pretty good.
It has a pretty high quantum efficiency.
The downside of this guy is that it bleaches extremely fast,
and that's kind of why we don't use it so much anymore.
One general rule to see is that these all have these connected ring structures.
And in general, the larger that system is, the longer the wavelength,
which you can see nicely here in rhodamine,
which also means the more hydrophobic these dyes become.
Now, dyes have been synthesized all over the place, and there's now a wide spectrum
of different dyes that have different excitation and emission maxima,
that have different photobleaching specificities,
and you can get these from several companies.
This company, Molecular Probes, now owned by Invitrogen,
makes lots of these different dyes.
So, study up on what you use before doing your experiments.
Having a dye is, of course, not enough.
We want to use that dye to detect certain aspects of our cells.
We're kind of lucky in some respects, because some dyes in themselves
already stain cellular components without any effort.
So, for instance, these DNA-staining dyes, like DAPI and these Hoechst components
bind themselves to DNA, and then increase their fluorescence many, many folds.
So, we can stain cells for DNA just by adding these compounds, and there you go.
Likewise, this compound called Mitotracker is actually taken up by the cells.
The compound enters the mitochondria and is then oxidized into a fluorescent component
that also cannot leave the mitochondria anymore.
It's a very easy way of labeling mitochondria.
So, often we are not so lucky, and we need to couple the dye
to the protein we're interested in.
So, we can then either do a direct labeling -- we couple the dye to the protein of interest
and then put it into the cell.
Or, what's often done is that the dye is coupled to an antibody,
and the antibody recognizes the protein of interest.
So, that still then requires every primary antibody that you get to be labeled with the dye.
That's a lot of work, and it was soon realized that it's much easier
to label just a few secondary antibodies that bind to, for instance,
these mouse or rabbit antibodies.
Have these labeled so that you have only a limited set of those in your refrigerator
that you can pull out whenever you make a new primary antibody.
And the added benefit of this is that you get an extra amplification
by building up this small little tree.
And, indirect immunofluorescence is used a lot to label cells
or other structures to visualize your protein of interest.
And we'll now go to the microscope and see what it actually looks like
to take such an image that's been labeled with different dyes.
Now, when we take images, we actually have a monochrome camera.
Those black and white monochrome cameras give us higher sensitivity in the end.
But, we will now see different dyes depending on which filter cube we put in place.
So, in a typical sequence of acquisition, you would first excite with UV light,
get blue fluorescence. In our example, that will be blue DNA.
You will then excite with blue light that gives green fluorescence.
Our example of that will be green actin.
And then you will excite with yellow/orange light that will give red fluorescence,
which in our example are mitochondria.
Here, you see now the end result.
We have, in green, actin labeled with this phalloidin compound.
We have in red mitochondria labeled with Mitotracker.
And in blue, we have nuclei labeled with this DAPI compound.
And this is a typical end result of an indirect immunofluorescent experiment.
Now, there are other types of dyes.
One that should be mentioned are these quantum dots.
Quantum dots are made of inorganic semiconductor material.
They are relatively large -- they are kind of on the scale of a protein.
So, they're much larger than the organic dyes that we talked about before.
They have certain advantages, though, and one of the major ones
is that they photobleach at a very, very slow rate,
meaning that you can get many, many more photons out of them than out of organic dyes
before they photobleach.
Another characteristic is that their emission wavelength depends on the size of
the quantum dot. So, the bigger the quantum dot, the higher the wavelength.
Whereas the absorption efficiency does not depend on the size,
but is the same for more or less all of the quantum dots,
meaning that at lower wavelength, you get higher absorption, which means
that you can excite different quantum dots all simultaneously
with one and the same wavelength.
So, as we've seen in the introduction,
where it really becomes interesting is when we can label proteins within living cells.
And that is so easy thanks to these fluorescent proteins.
Now, the discovery of fluorescent proteins is a very interesting story.
So this year, a Nobel prize was awarded for the discovery of fluorescent proteins.
What was realized was that certain jellyfish will exhibit green light.
They can do it just by themselves, but they can also exhibit green light
when you put them on a UV illuminator,
as we used earlier to show fluorescence of the tonic water.
Then Shimomura, he spent many, many years figuring out
where this fluorescence came from.
He actually isolated the green fluorescent protein from these jellyfish.
So, when you hold a jellyfish like this on the UV illuminator,
what you will see is this ring, which is this ring here of green fluorescence.
Shimomura isolated the protein and showed that the protein itself was fluorescent.
Doug Prasher, who at the time was in Woods Hole, Massachusetts,
had the idea that it was possible that when expressing this protein just by itself,
it might also be fluorescent, and therefore could be a great tool
to couple to other proteins of interest
and to express it in other organisms.
He made the first steps towards realizing that idea,
and he cloned the gene for green fluorescent protein.
However, once he had cloned that gene, he ran out of funding,
and he actually left science altogether.
And it was then Chalfie and Roger Tsien who picked up that work.
And Chalfie was the first to demonstrate that when you express the GFP in
an organism like E. coli or C. elegans, it also fluorescences all by itself.
And this was a bit unexpected, since all other fluorescent proteins
actually need cofactors. But, in this case, no cofactors are needed.
Simply expressing the protein itself gives you a fluorescent product inside other cells.
And that is why this has become such an amazing and great tool in cell biology.
Then, Roger Tsien was one of the leaders in further manipulating
this green fluorescent protein.
So, here we see the crystal structure of green fluorescent protein.
It has this beta barrel structure, and the chromophore is within the beta barrel.
So, Roger Tsien and his coworkers made many different varieties -- different mutants --
that have different spectral properties.
Also, there were other fluorescent proteins isolated from other organisms
which actually were the source for all these red fluorescent proteins.
And so, this is still an ongoing process.
GFP is probably the most engineered protein known to human kind at the moment.
We can get GFPs with all kinds of different characteristics.
New ones are coming out continuously, so that you really will have to study up
to decide which is the best one for your specific experiment.
So, I hope you now have a solid understanding of what fluorescence microscopy is,
and why we, as cell biologists, like fluorescence microscopy so much.
And just to end, I want to show you this movie of a dividing cell,
which is similarly labeled as the cell in the beginning.
So, it's an S2 cell with microtubules here in red, and this EB1 protein in green.
And you see again, how fluorescence lets us see these very specific processes within cells.
Thank you very much.