Friday, February 04, 2005
Invasion of the Brain Snatchers
Next week, my fair institution of higher learning will be visited by the world renowned mathematician and philosopher, Dr. William Dembski.
*mops up sarcasm that was dripping and is now flooding the floor*
See here for more information. It should of course be noted that it is not a department that is sponsoring him, but rather a group of various religious organizations (and the Philosophy Club). Rest assured, I am sure the media hacks who back ID will be spouting off, "But Dr. William Dembski gave a talk at Columbia University in New York, and that's one of them fancy Ivy League schools, blah blah, blah" as soon as the curtain calls on the session.
I'm contemplating attending the entire debacle (because I'm sure it will be one), but I wonder if I want to kill off my brain cells in that regard. It would be far more enjoyable to down a six pack to be honest.
|
*mops up sarcasm that was dripping and is now flooding the floor*
See here for more information. It should of course be noted that it is not a department that is sponsoring him, but rather a group of various religious organizations (and the Philosophy Club). Rest assured, I am sure the media hacks who back ID will be spouting off, "But Dr. William Dembski gave a talk at Columbia University in New York, and that's one of them fancy Ivy League schools, blah blah, blah" as soon as the curtain calls on the session.
I'm contemplating attending the entire debacle (because I'm sure it will be one), but I wonder if I want to kill off my brain cells in that regard. It would be far more enjoyable to down a six pack to be honest.
|
Wednesday, January 05, 2005
It would be nice....
In Praise of the Mad Scientist by Samantha Hunt (Village Voice)
There's a certain romance that I see in Hunt's essay - the idea of being able to pursue one's scientific interests without needing to slave over a lengthy grant proposal to the NSF or NIH; without requiring one to be independently wealthy; going after that which you think is important regardless of whether or not you can make a profit off of it. Sure, it would be nice to have a fully loaded lab all to myself. I would imagine that many lives would be greatly simplified if they could pursue important goals without needing lots of personal capital, or perhaps spared the effort of writing up grant proposals for submission to a government bureaucracy.
There is a strange disconnect I noticed right away in the argument - is Hunt after inventors who will revolutionize society or scientists who will add to society's knowledge of the world around us? While I am again going to be obscure about what it is that I do, my work has interesting relationships to issues in both environmental science and drug discovery. I personally don't have much interest in either field - I am far more intrigued by the fundamental physics and chemistry that underlies the problem. Scientists are not necessarily out to invent anything - it may come up that in the course of their work that something needs to be done that has never been done before, and therefore they invent. It also occurs that things come to be by luck - as I recall, Viagra was originally studied in relation to cardiovascular disease, not erectile dysfunction.
I may or may not come back to this in the days ahead (why, yes, I've left this blog in cobwebs for a few months now), but what really strikes me is that Hunt seems to have a very strong opinion of not only how science should be done but also what science should be done. It's actually kind of amusing when she's spent some effort in extolling the virtues of originality and independence.
P.S. - In response to a comment left about two months ago, I do not worship Darwin. That was a bad attempt at humor with the "honest-to-Darwin" comment.
P.P.S. - For everyone who's done a Google or Yahoo search linking to this page looking for information on solid state and magic angle spinning (MAS) NMR, Rob Schurko's solid state NMR pagehas a nice set of introductory notes. You may also want to check out this webpage by Klaus Eichele at the Universitat Tuebingen. For quadrupolar NMR, I'd suggest Pascal Man's very engrossing webpage .
|
There's a certain romance that I see in Hunt's essay - the idea of being able to pursue one's scientific interests without needing to slave over a lengthy grant proposal to the NSF or NIH; without requiring one to be independently wealthy; going after that which you think is important regardless of whether or not you can make a profit off of it. Sure, it would be nice to have a fully loaded lab all to myself. I would imagine that many lives would be greatly simplified if they could pursue important goals without needing lots of personal capital, or perhaps spared the effort of writing up grant proposals for submission to a government bureaucracy.
There is a strange disconnect I noticed right away in the argument - is Hunt after inventors who will revolutionize society or scientists who will add to society's knowledge of the world around us? While I am again going to be obscure about what it is that I do, my work has interesting relationships to issues in both environmental science and drug discovery. I personally don't have much interest in either field - I am far more intrigued by the fundamental physics and chemistry that underlies the problem. Scientists are not necessarily out to invent anything - it may come up that in the course of their work that something needs to be done that has never been done before, and therefore they invent. It also occurs that things come to be by luck - as I recall, Viagra was originally studied in relation to cardiovascular disease, not erectile dysfunction.
I may or may not come back to this in the days ahead (why, yes, I've left this blog in cobwebs for a few months now), but what really strikes me is that Hunt seems to have a very strong opinion of not only how science should be done but also what science should be done. It's actually kind of amusing when she's spent some effort in extolling the virtues of originality and independence.
P.S. - In response to a comment left about two months ago, I do not worship Darwin. That was a bad attempt at humor with the "honest-to-Darwin" comment.
P.P.S. - For everyone who's done a Google or Yahoo search linking to this page looking for information on solid state and magic angle spinning (MAS) NMR, Rob Schurko's solid state NMR pagehas a nice set of introductory notes. You may also want to check out this webpage by Klaus Eichele at the Universitat Tuebingen. For quadrupolar NMR, I'd suggest Pascal Man's very engrossing webpage .
|
Friday, September 03, 2004
Adventures in Babysitting (Or, Watch Michael Try to Grow Protein Crystals)
Most of us probably have somewhat fond memories of growing simple salt crystals when younger, either in school or out of the classroom. It's fun, it's easy, and it's actually kind of pretty after a fashion. For those of us who went on to take more science classes at the university level, you may have fond (OK, probably not-so-fond) memories of trying to crystallize a compound you synthesized in a laboratory course. Crystallization is an important step for many investigations of chemical structure and function for at least two reasons - it is generally a sign that what you have is reasonably pure (although not necessarily what you're looking for), and is also suitable for analytical methods where a solid is required. This, of course, is what is needed for x-ray crystallography. This method depends upon the fact that x-rays will diffract in a comprehensible way from a regular, periodic arrangement of atoms and/or molecules in a crystal. The sorts of salts we used to play with as children and organic compounds that we may have played with as young adults are not that typically large - perhaps equivalent to a few hundred daltons (where one dalton is equivalent to the mass of a single hydrogen). They are typically pretty anhydrous (one can heat them up, driving away any stray water molecules, without doing much damage - although heating it up too much will melt the salt, and that's a different story) and, for the most part, quick to prepare.
Proteins, however, are a different story. And here things get interesting. Protein crystals can have lots of associated waters. It's been found that some protein crystals are, in fact, mostly water that's bound to the protein in one way or another. Some protein crystals can take up to a month to grow properly. As some of you are probably saying to yourselves right now, proteins are generally far larger than a few hundred daltons - tens to hundreds of thousands of daltons, particularly when one is talking about large respiratory complexes and other such monsters of biochemistry. Some proteins are notoriously difficult to crystallize - membrane proteins, which reside in a greasy envelope of lipids or are associated to one via a greasy anchor at one end of the protein, are one of the continuing challenges for protein crystallographers. Instead of trying to line up in three dimensions small, cute, cuddly atoms and molecules, one is attempting to line up large, bulky, recalcitrant proteins into crystals on the scale of a few hundred micrometers in each direction ideally.
As to how to do this....there is a curious blend of physical chemistry, one's intuition, and black magic. (OK, I'm kidding about the black magic. But if you happen to have any goats available for moonlight sacrifices.....) The idea is that one needs to precipitate the protein out of solution slowly and in an orderly fashion. Salts are often added to the buffer so as to take up water molecules in a solvation shell which would otherwise be surrounding the protein, as well as to affect the surface residues of the protein. Organic molecules such as polyethylene glycol are also added, which draw away water from the proteins in a similar fashion. Other chemicals may be added as needed - for example, dithiothreitol or 2-mercaptoethanol to prevent formation of disulfide bonds, or perhaps if one is working with an enzyme, the enzyme's preferred substrate so as to see where the substrate binds relative to the active site. One can also play with temperature, the amount of protein, and the geometry of the crystallization process (do you have your protein hanging above a reservoir of these compounds, sitting down surrounded by these compounds, or being injected into a microfluidics device?).
Even after all of this, you may still not end up with a protein crystal. Or, you may not end up with a large enough protein crystal - one might have a batch of so-called "microcrystals," so-called since they're best measured not in hundreds of micrometers but far fewer micrometers. What might look like messy precipitate might be nanocrystals, but it could just be unordered protein precipitate after all. My lab's specialty - solid state NMR - is able to deal with such cases, where crystals don't form large enough crystals for x-ray diffraction studies or fails to form crystals at all.
However, to my astonishment, I managed to grow a very few crystals back in the spring of x-ray diffraction size and so I've been following up on that as well as more expected pursuits. You see, there are some interesting questions that arise from thinking about the currently available crystal structures for the system I work with when considering temperature effects and under what conditions one finds my protein (to wit, crystals grown in a refrigerator don't necessarily correlate to an organism which does not grow at such temperatures). So, I figured, why not? However, as was once related to me by a protein crystallographer, "Sometimes it's just a matter of setting enough trays and wells before you start getting crystals." So, as to what I actually do when attempting to grow crystals? Let me show you.
First, I take one (or two, or three, or however many I need) crystallization tray(s) (we use the well known standard from Hampton Research, their VDX Plate), and apply a small bit of vacuum grease to each and every circular well ring. This makes sure that when I apply the cover slide (which will come later), I have a reliable seal and will not have contamination from the outside. This is done by gently heating some vacuum grease until it melts and dipping a small glass flask upside down into the grease, and then carefully applying the greased rim of the flask to the well ring. When I first started doing this, it would take almost an hour to do this for an entire tray (24 wells) - I was very concerned about dripping grease into the wells and, in general, making a mess of the entire process. Now it takes ten minutes on a good day. I'll usually make sure that my precipitation/crystallization solutions are prepared and and ready to be added, either directly or after a small bit of mixing if I'm using a high concentration stock solution of polyethylene glycol. The protein solution is in its appropriate buffer and at a high enough concentration, and has been mixed with the substrate if I'm doing such a trial. Now, I use pretty high concentrations of protein at times (over 200 milligrams per milliliter on occasion), so I often have to use centrifugal concentration methods - like these from Fisher - to get my protein to the desired protein concentration. If I'm cocrystallizing my protein with its substrate, I'll typically do this before concentrating - this way I can monitor the binding with UV/Vis spectra and then be sure that my substrate is in there before spending all that time concentrating my protein to a thin layer of protein sludge (well, not really sludge per se, just rather viscous solution). I then take some small plastic microscope cover slips, add a small bit of protein sludge, add a small bit of precipitation/crystallization solution, and then place the slide over a filled well. This is one of those places beginners can botch things up - if you take too long in flipping the slide over, it can run and you get something smeared out over more area than a neat small drop's worth. You then gently press down and end up with a sealed well, protein hanging above your solution. You do this 12, 24, 48, 72, ..... times and you wait. You need to check daily under a microscope (just a regular light microscope, nothing fancy is really needed for just monitoring growth or lack thereof of your protein crystals, and we have a digital camera to take pictures of anything interesting looking) and wait. Did I mention the waiting? This is the babysitting, and what usually takes the longest. You can make up your solutions, prepare your tray with the grease seals, concentrate your protein, and mix everything up in a day - the waiting is what really takes the time. One day of work - a week or longer (far, far longer in some cases) of watching and waiting.
I've had middling to fair success thus far - one of the proteins I worked with grew crystals pretty easily (then again, the conditions were already known) and the other one, my main interest, is coming along reasonably well. I seem to be able to get microcrystalline material without too much difficulty, and am playing around with conditions to see if I can get larger crystals. I'm also going to try such interesting techniques as microseeding - taking a small microcrystal and dropping it into a freshly set well to act as a growth initiator for a larger crystal - and magnetic field crystallization - where one basically sets up a magnetic field surrounding your trial in order to grow larger crystals. (As for an actual explanation of how magnetic fields affect protein crystallization, your guess is as good as mine - there are lots of ideas, but nothing firm or even reliable. One can wave one's hands about, but in the end.....remember the comment about black magic before?) Something that I need to be careful about is how conditions for crystal growth may not be optimal for doing NMR - too much salt in microcrystals or precipitate can cause unwanted heated during an NMR experiment, considering that we're pumping in high power radiofrequency waves; too much precipitant (polyethylene glycol) might cause unwanted signals in the resulting spectra of the protein.
Ultimately, though, it's well worth the time - if I can devise better ways to precipitate out protein (forming large, x-ray grade crystals or not), the easier my adventures in preparing future NMR samples should be, as well as the tantalizing idea of having comparative crystallographic data to work with when trying to synthesize what is already known with my original research.
Sites To Look At:
Crystallography 101 - An introduction to x-ray crystallography at Lawrence Livermore.
Crystals for Fun and Education - Hampton Research has some fun and cool crystallization projects for kids (and those who are still kids at heart).
|
Proteins, however, are a different story. And here things get interesting. Protein crystals can have lots of associated waters. It's been found that some protein crystals are, in fact, mostly water that's bound to the protein in one way or another. Some protein crystals can take up to a month to grow properly. As some of you are probably saying to yourselves right now, proteins are generally far larger than a few hundred daltons - tens to hundreds of thousands of daltons, particularly when one is talking about large respiratory complexes and other such monsters of biochemistry. Some proteins are notoriously difficult to crystallize - membrane proteins, which reside in a greasy envelope of lipids or are associated to one via a greasy anchor at one end of the protein, are one of the continuing challenges for protein crystallographers. Instead of trying to line up in three dimensions small, cute, cuddly atoms and molecules, one is attempting to line up large, bulky, recalcitrant proteins into crystals on the scale of a few hundred micrometers in each direction ideally.
As to how to do this....there is a curious blend of physical chemistry, one's intuition, and black magic. (OK, I'm kidding about the black magic. But if you happen to have any goats available for moonlight sacrifices.....) The idea is that one needs to precipitate the protein out of solution slowly and in an orderly fashion. Salts are often added to the buffer so as to take up water molecules in a solvation shell which would otherwise be surrounding the protein, as well as to affect the surface residues of the protein. Organic molecules such as polyethylene glycol are also added, which draw away water from the proteins in a similar fashion. Other chemicals may be added as needed - for example, dithiothreitol or 2-mercaptoethanol to prevent formation of disulfide bonds, or perhaps if one is working with an enzyme, the enzyme's preferred substrate so as to see where the substrate binds relative to the active site. One can also play with temperature, the amount of protein, and the geometry of the crystallization process (do you have your protein hanging above a reservoir of these compounds, sitting down surrounded by these compounds, or being injected into a microfluidics device?).
Even after all of this, you may still not end up with a protein crystal. Or, you may not end up with a large enough protein crystal - one might have a batch of so-called "microcrystals," so-called since they're best measured not in hundreds of micrometers but far fewer micrometers. What might look like messy precipitate might be nanocrystals, but it could just be unordered protein precipitate after all. My lab's specialty - solid state NMR - is able to deal with such cases, where crystals don't form large enough crystals for x-ray diffraction studies or fails to form crystals at all.
However, to my astonishment, I managed to grow a very few crystals back in the spring of x-ray diffraction size and so I've been following up on that as well as more expected pursuits. You see, there are some interesting questions that arise from thinking about the currently available crystal structures for the system I work with when considering temperature effects and under what conditions one finds my protein (to wit, crystals grown in a refrigerator don't necessarily correlate to an organism which does not grow at such temperatures). So, I figured, why not? However, as was once related to me by a protein crystallographer, "Sometimes it's just a matter of setting enough trays and wells before you start getting crystals." So, as to what I actually do when attempting to grow crystals? Let me show you.
First, I take one (or two, or three, or however many I need) crystallization tray(s) (we use the well known standard from Hampton Research, their VDX Plate), and apply a small bit of vacuum grease to each and every circular well ring. This makes sure that when I apply the cover slide (which will come later), I have a reliable seal and will not have contamination from the outside. This is done by gently heating some vacuum grease until it melts and dipping a small glass flask upside down into the grease, and then carefully applying the greased rim of the flask to the well ring. When I first started doing this, it would take almost an hour to do this for an entire tray (24 wells) - I was very concerned about dripping grease into the wells and, in general, making a mess of the entire process. Now it takes ten minutes on a good day. I'll usually make sure that my precipitation/crystallization solutions are prepared and and ready to be added, either directly or after a small bit of mixing if I'm using a high concentration stock solution of polyethylene glycol. The protein solution is in its appropriate buffer and at a high enough concentration, and has been mixed with the substrate if I'm doing such a trial. Now, I use pretty high concentrations of protein at times (over 200 milligrams per milliliter on occasion), so I often have to use centrifugal concentration methods - like these from Fisher - to get my protein to the desired protein concentration. If I'm cocrystallizing my protein with its substrate, I'll typically do this before concentrating - this way I can monitor the binding with UV/Vis spectra and then be sure that my substrate is in there before spending all that time concentrating my protein to a thin layer of protein sludge (well, not really sludge per se, just rather viscous solution). I then take some small plastic microscope cover slips, add a small bit of protein sludge, add a small bit of precipitation/crystallization solution, and then place the slide over a filled well. This is one of those places beginners can botch things up - if you take too long in flipping the slide over, it can run and you get something smeared out over more area than a neat small drop's worth. You then gently press down and end up with a sealed well, protein hanging above your solution. You do this 12, 24, 48, 72, ..... times and you wait. You need to check daily under a microscope (just a regular light microscope, nothing fancy is really needed for just monitoring growth or lack thereof of your protein crystals, and we have a digital camera to take pictures of anything interesting looking) and wait. Did I mention the waiting? This is the babysitting, and what usually takes the longest. You can make up your solutions, prepare your tray with the grease seals, concentrate your protein, and mix everything up in a day - the waiting is what really takes the time. One day of work - a week or longer (far, far longer in some cases) of watching and waiting.
I've had middling to fair success thus far - one of the proteins I worked with grew crystals pretty easily (then again, the conditions were already known) and the other one, my main interest, is coming along reasonably well. I seem to be able to get microcrystalline material without too much difficulty, and am playing around with conditions to see if I can get larger crystals. I'm also going to try such interesting techniques as microseeding - taking a small microcrystal and dropping it into a freshly set well to act as a growth initiator for a larger crystal - and magnetic field crystallization - where one basically sets up a magnetic field surrounding your trial in order to grow larger crystals. (As for an actual explanation of how magnetic fields affect protein crystallization, your guess is as good as mine - there are lots of ideas, but nothing firm or even reliable. One can wave one's hands about, but in the end.....remember the comment about black magic before?) Something that I need to be careful about is how conditions for crystal growth may not be optimal for doing NMR - too much salt in microcrystals or precipitate can cause unwanted heated during an NMR experiment, considering that we're pumping in high power radiofrequency waves; too much precipitant (polyethylene glycol) might cause unwanted signals in the resulting spectra of the protein.
Ultimately, though, it's well worth the time - if I can devise better ways to precipitate out protein (forming large, x-ray grade crystals or not), the easier my adventures in preparing future NMR samples should be, as well as the tantalizing idea of having comparative crystallographic data to work with when trying to synthesize what is already known with my original research.
Sites To Look At:
Crystallography 101 - An introduction to x-ray crystallography at Lawrence Livermore.
Crystals for Fun and Education - Hampton Research has some fun and cool crystallization projects for kids (and those who are still kids at heart).
|
Sunday, July 04, 2004
I like bacteria.
This may not be totally surprising - after all, bacteria are exceptionally useful production factories for chemists and biophysicists interested in macromolecular structure and function and enzymology. They are a cheap, effective method of overexpressing proteins, nucleic acids, and other molecules of current interest. They are also of course targets in and of themselves, being the reason for a number of diseases and medical maladies.
But no, you see, I really like bacteria. I think they're fascinating in and of themselves, without appeals to medical relevance or to biotechnological advantages. I used to work with cyanobacteria and green sulfur bacteria, both of which are photosynthetic organisms. I could make up a batch of solution containing not much more than inorganic salts and a few organic acids, inoculate with a colony or frozen culture, shine some light on them while in a warm water bath, and I'd end up with a dark green broth of dense photosynthetic bacteria. Out of a sulfur-laden bottle of smelly media I would grow some of the most fascinating bacteria out there. Did you know green sulfur bacteria are able to grow meters - up to ~70 meters - under the surface of water, where the transmitted light is only a slight percentage of that available at the surface? Their light harvesting apparatus are contained in what are known as chlorosomes - a monolayer of lipids and proteins - where long rods of aggregated chlorophyll are able to efficiently harvest light. The vast diversity of metabolism amongst bacteria is also endlessly fascinating and could keep a person busy for a lifetime. Modified glycolytic pathways where adenosine diphosphate (ADP) - not adenosine triphosphate (ATP) - serves as a phosphate donor, metalloproteins that utilize such "unconventional" metals like tungsten and vanadium....the list goes on. I cannot describe what a thrill it was to finally start learning these sorts of things when I reached university - while my experiences in high school biology were overwhelmingly positive, I left with the impression that bacteria weren't all that interesting. Then when you find out that there is spatial organization in bacteria - after having been inundated with "eukaryotes have internal organization and organelles, prokaryotes have neither of them" - you suddenly feel like the wool has been lifted from your eyes. When I first began to read about Carl Woese and his work in classifying microorganisms - and his proposal of the Archaea - it was wonderful to see how the main division that I had always been taught (prokaryotes and eukaryotes) was being rethought.
I may yet end up an honest-to-Darwin biologist after all.
|
This may not be totally surprising - after all, bacteria are exceptionally useful production factories for chemists and biophysicists interested in macromolecular structure and function and enzymology. They are a cheap, effective method of overexpressing proteins, nucleic acids, and other molecules of current interest. They are also of course targets in and of themselves, being the reason for a number of diseases and medical maladies.
But no, you see, I really like bacteria. I think they're fascinating in and of themselves, without appeals to medical relevance or to biotechnological advantages. I used to work with cyanobacteria and green sulfur bacteria, both of which are photosynthetic organisms. I could make up a batch of solution containing not much more than inorganic salts and a few organic acids, inoculate with a colony or frozen culture, shine some light on them while in a warm water bath, and I'd end up with a dark green broth of dense photosynthetic bacteria. Out of a sulfur-laden bottle of smelly media I would grow some of the most fascinating bacteria out there. Did you know green sulfur bacteria are able to grow meters - up to ~70 meters - under the surface of water, where the transmitted light is only a slight percentage of that available at the surface? Their light harvesting apparatus are contained in what are known as chlorosomes - a monolayer of lipids and proteins - where long rods of aggregated chlorophyll are able to efficiently harvest light. The vast diversity of metabolism amongst bacteria is also endlessly fascinating and could keep a person busy for a lifetime. Modified glycolytic pathways where adenosine diphosphate (ADP) - not adenosine triphosphate (ATP) - serves as a phosphate donor, metalloproteins that utilize such "unconventional" metals like tungsten and vanadium....the list goes on. I cannot describe what a thrill it was to finally start learning these sorts of things when I reached university - while my experiences in high school biology were overwhelmingly positive, I left with the impression that bacteria weren't all that interesting. Then when you find out that there is spatial organization in bacteria - after having been inundated with "eukaryotes have internal organization and organelles, prokaryotes have neither of them" - you suddenly feel like the wool has been lifted from your eyes. When I first began to read about Carl Woese and his work in classifying microorganisms - and his proposal of the Archaea - it was wonderful to see how the main division that I had always been taught (prokaryotes and eukaryotes) was being rethought.
I may yet end up an honest-to-Darwin biologist after all.
|
Monday, June 28, 2004
I don't normally speak about politics here, mainly since I really don't think anyone comes here for my politics (admit it, you come here for the occasional bit of science, dark grad student humor, and waiting for me to make some bad science jokes), but I figure I have to say this much, at least.
I went to see Fahrenheit 9/11 Friday evening with a few of my friends. Overall, I thought the movie could have been a bit more tightly edited and Moore could have strengthened the connections between the points he makes in the movie. However, if it takes a film to get people talking about public policy and the absolutely cracked out excesses of the Bush Administration and the direction of the U.S., then it is well worth the controversy and ensuing drama that has arisen. As I mentioned to a few people elsewhere, I don't think this movie is best served by praising it or attacking it, because everyone else is doing that. If it starts to inspire more honest and forthright discussion and action about what direction this country has been going and how to change it for the better, that's perhaps the best response ever.
I also saw Michael Moore heading out of the theater complex after one showing as I was waiting in line waiting to go into another theater for my showing. I would have ran after him and asked him something marginally intelligent, but I figured I should at least see his movie. Besides, the theater was packed, a good seat in NYC is hard to find at times. Heh.
|
I went to see Fahrenheit 9/11 Friday evening with a few of my friends. Overall, I thought the movie could have been a bit more tightly edited and Moore could have strengthened the connections between the points he makes in the movie. However, if it takes a film to get people talking about public policy and the absolutely cracked out excesses of the Bush Administration and the direction of the U.S., then it is well worth the controversy and ensuing drama that has arisen. As I mentioned to a few people elsewhere, I don't think this movie is best served by praising it or attacking it, because everyone else is doing that. If it starts to inspire more honest and forthright discussion and action about what direction this country has been going and how to change it for the better, that's perhaps the best response ever.
I also saw Michael Moore heading out of the theater complex after one showing as I was waiting in line waiting to go into another theater for my showing. I would have ran after him and asked him something marginally intelligent, but I figured I should at least see his movie. Besides, the theater was packed, a good seat in NYC is hard to find at times. Heh.
|
Sunday, June 27, 2004
In the June 30th issue of the Journal of the American Chemical Society, there's a paper by a MIT research group on an application of two-dimensional (2-D) infrared (IR) spectroscopy to protein structure and dynamics, namely antiparallel (AP) beta-sheets (see here for a convenient visualization of beta sheets). The potential usefulness of this work may or may not be immediately clear, if it's not, I urge you to keep reading I think it's quite fascinating.
For those who never had the pleasure of taking organic chemistry or certain other classes (or for those who have, you've since pushed some of them into your subconscious), infrared spectroscopy measures vibrational transitions in molecules. In analytical chemistry, particularly the type that is done in organic chemistry courses, this yields information regarding the presence or absence of various functional groups (e.g. a carbonyl or amide moiety), which typically relates to the bond strength and the masses of the atoms involved. IR data is normally presented in units of wavenumbers (which equal 1/wavelength of radiation in cm) and in terms of transmittance (so absorption of light by a certain vibrational transition shows up as a low amount of transmittance at that specific wavenumber). This sort of analysis is not particularly quantitative in nature IR spectroscopy's ability to serve as a general probe for functional groups means there are tradeoffs with being able to extract useful and local structural and dynamic information. There is also the aspect that repetitive motifs in a molecule's structure can add up to give a broadened, featureless resonance. For an example, the fundamental backbone structure of a protein repeats regularly (see this Birkbeck College page for instance), so any resonances that originate from the backbone would likely overlap. One transition in proteins, the so-called amide I transition, is quite overlapped in IR spectra, being primarily the stretching of the C=O bond coupled with some peptide bond (CN) stretching and CCN bending.
This amide I transition is in fact the window through which the group investigated secondary structure motifs in their 2-D IR experiments. If you look at the previously linked structure of antiparallel beta sheets, you'll notice that the short range interactions (hydrogen bonding) that give the beta sheet the appearance of being a sheet and not just a collection of protein strands are from the N-H of one strand to the C=O of another strand just the components which contribute to the amide I transition. Hydrogen bonding is a short-range interaction, sensitive to conformation and dynamics of the hydrogen donor and acceptor this makes it quite useful for probing protein dynamics (such as folding or conformational changes), as well as interactions of the protein with other proteins, nucleic acids, and small molecules. When one considers that infrared spectroscopy has an instrinsic timescale in the picosecond range, measuring the fast kinetic processes of protein folding becomes possible. The experimental setup is based upon a four wave mixing experiment three pulses of light are directed into the sample, and based upon the optical properties of the sample (as well as the parameters of the light pulses), a fourth signal is generated and then emitted from the sample which contains the information of interest. The samples used included a model polypeptide (poly-L-lysine, with a molecular weight range from 70,000 daltons to 150,000 daltons) and the following proteins concanavalin A (Con A), ribonuclease A, lysozyme, and myoglobin. Con A is exclusively composed of antiparallel beta sheets excluding any unstructured loops and strands, and the amount of AP beta-sheets decreases until you reach the exclusively helical myoglobin, with the same qualifer. The two-dimensional descriptive is seen in how the data from these experiments is presented and analyzed with the initially excited transition plotted against the resultant signal. This information coupled with the the model derived in the paper for 2-D IR spectroscopy of beta-sheets can provide insight into conformational dynamics, given frequencies, lineshapes, splitting patterns and crosspeaks in the two-dimensional plots. I was thinking of explaining more about the physical basis of this work, but I didn't want to overshadow the biological relevance in this post. Perhaps I will keep it in mind for future posts.
To wrap it up, the MIT group was able to identify patterns and data with 2D spectroscopy not extracted if at all observed in one dimensional IR spectroscopy with a fairly reasonable model underlying it all. Obviously, much more work must be done before its full biological utility is realized, especially in extending this idea to more complicated (e.g., mixed secondary structure) proteins, as well as to proteins interacting with other biopolymers and relevant molecules.
(For those interested in looking at visualizing biological macromolecules for themselves after reading this, I highly recommend Swiss PDB Viewer available at the ExPASy Proteomics Server - it's free and available for Mac, Windows, Linux, and SGI. I personally have used this for a while now, including for my fabled second year defense earlier this year as well as for giving talks and for other sundry tasks. You can make very nice pictures using POV-Ray, a ray-tracing program which is very easy to use with Swiss PDB Viewer. While the PDB is available on my side bar as a useful link, I'll pass along the PDB IDs for the proteins used in this study just so you can see for yourself - Concanavalin A :1NLS; ribonuclease A: 1FS3; lysozyme: 132L; and myoglobin: 1WLA.)
|
For those who never had the pleasure of taking organic chemistry or certain other classes (or for those who have, you've since pushed some of them into your subconscious), infrared spectroscopy measures vibrational transitions in molecules. In analytical chemistry, particularly the type that is done in organic chemistry courses, this yields information regarding the presence or absence of various functional groups (e.g. a carbonyl or amide moiety), which typically relates to the bond strength and the masses of the atoms involved. IR data is normally presented in units of wavenumbers (which equal 1/wavelength of radiation in cm) and in terms of transmittance (so absorption of light by a certain vibrational transition shows up as a low amount of transmittance at that specific wavenumber). This sort of analysis is not particularly quantitative in nature IR spectroscopy's ability to serve as a general probe for functional groups means there are tradeoffs with being able to extract useful and local structural and dynamic information. There is also the aspect that repetitive motifs in a molecule's structure can add up to give a broadened, featureless resonance. For an example, the fundamental backbone structure of a protein repeats regularly (see this Birkbeck College page for instance), so any resonances that originate from the backbone would likely overlap. One transition in proteins, the so-called amide I transition, is quite overlapped in IR spectra, being primarily the stretching of the C=O bond coupled with some peptide bond (CN) stretching and CCN bending.
This amide I transition is in fact the window through which the group investigated secondary structure motifs in their 2-D IR experiments. If you look at the previously linked structure of antiparallel beta sheets, you'll notice that the short range interactions (hydrogen bonding) that give the beta sheet the appearance of being a sheet and not just a collection of protein strands are from the N-H of one strand to the C=O of another strand just the components which contribute to the amide I transition. Hydrogen bonding is a short-range interaction, sensitive to conformation and dynamics of the hydrogen donor and acceptor this makes it quite useful for probing protein dynamics (such as folding or conformational changes), as well as interactions of the protein with other proteins, nucleic acids, and small molecules. When one considers that infrared spectroscopy has an instrinsic timescale in the picosecond range, measuring the fast kinetic processes of protein folding becomes possible. The experimental setup is based upon a four wave mixing experiment three pulses of light are directed into the sample, and based upon the optical properties of the sample (as well as the parameters of the light pulses), a fourth signal is generated and then emitted from the sample which contains the information of interest. The samples used included a model polypeptide (poly-L-lysine, with a molecular weight range from 70,000 daltons to 150,000 daltons) and the following proteins concanavalin A (Con A), ribonuclease A, lysozyme, and myoglobin. Con A is exclusively composed of antiparallel beta sheets excluding any unstructured loops and strands, and the amount of AP beta-sheets decreases until you reach the exclusively helical myoglobin, with the same qualifer. The two-dimensional descriptive is seen in how the data from these experiments is presented and analyzed with the initially excited transition plotted against the resultant signal. This information coupled with the the model derived in the paper for 2-D IR spectroscopy of beta-sheets can provide insight into conformational dynamics, given frequencies, lineshapes, splitting patterns and crosspeaks in the two-dimensional plots. I was thinking of explaining more about the physical basis of this work, but I didn't want to overshadow the biological relevance in this post. Perhaps I will keep it in mind for future posts.
To wrap it up, the MIT group was able to identify patterns and data with 2D spectroscopy not extracted if at all observed in one dimensional IR spectroscopy with a fairly reasonable model underlying it all. Obviously, much more work must be done before its full biological utility is realized, especially in extending this idea to more complicated (e.g., mixed secondary structure) proteins, as well as to proteins interacting with other biopolymers and relevant molecules.
(For those interested in looking at visualizing biological macromolecules for themselves after reading this, I highly recommend Swiss PDB Viewer available at the ExPASy Proteomics Server - it's free and available for Mac, Windows, Linux, and SGI. I personally have used this for a while now, including for my fabled second year defense earlier this year as well as for giving talks and for other sundry tasks. You can make very nice pictures using POV-Ray, a ray-tracing program which is very easy to use with Swiss PDB Viewer. While the PDB is available on my side bar as a useful link, I'll pass along the PDB IDs for the proteins used in this study just so you can see for yourself - Concanavalin A :1NLS; ribonuclease A: 1FS3; lysozyme: 132L; and myoglobin: 1WLA.)
|
Wednesday, June 16, 2004
When reading a story or article about some aspect of modern biology, my interest always piques when there is mention of how recent structural studies of the constituent parts has enabled new insights into the function of said system. The main reason for this interest is twofold - one, I rather like hearing that structural biology does have something to contribute to interesting problems in biology and medicine; and two, I always wonder just how reliable the structural work actually is, and not in the sense of the resolution of the structure.
Kleywegt and Jones (1995;Structure, 3: 535-540) describe some of their always interesting work in trying to improve validation of macromolecular structures, namely, intentionally tracing a protein backwards and attempting to fit it to the obtained electron density to see what differences arise. The intentionally traced backwards model of the protein had an R-factor (to slightly oversimplify, a statistical measure to compare the experimentally observed diffraction pattern to the expected diffraction pattern given the determined structure) of 0.214 (21.4%), while the model that had been obtained with a more conservative protocol (without being traced backwards) had an R-factor of 0.251 (25.1%). As the authors mention in this paper, the threshold tossed about in the crystallography community for an R-factor whereupon structures could be problematic floated around 25% and then upwards. Thankfully, the authors showed that other measures (particularly an updated R-factor, the so-called R(free)-factor where one randomly excludes some of the reflections from the refinement, introduced by Axel Brunger) clarified which was the incorrect structure and protein trace. The rise in the usage of NMR spectroscopy in structural biology sets off an interesting train of thought - structures derived by NMR are typically presented as an ensemble of individual structures, with an RMSD value determined for ensemble of structures. The RMSD is typically for the protein backbone at the very least. The idea is that a low RMSD suggests a higher quality structure - which really can't be justified in my opinion. The wonderful thing about NMR is that you can probe protein dynamics (as well as the dynamics of whatever the protein is interacting with). What follows next is perhaps obvious to the astute visitor - if the RMSD is not within a few angstroms (or is glaringly enormous along certain stretches of the protein), there is likely something interesting going on there with protein motion, solvation, protein-protein interactions, protein-ligand interactions, or some other biologically relevant process. And that is what should really get you excited, not that you can say you have an RMSD of 0.87 angstroms.
One of the other things is that you needn't abandon your basic chemical/biological intuition. If there look to be ridiculous conformations in the determined structure, they may very well be mistakes or artifacts. If the Ramachandran plot (where one plots the backbone torsion angles phi and psi against one another, see here at Birkbeck College for a more detailed explanation) has points where you really shouldn't expect to see them, there may be something wrong. There is an enormous database (the Cambridge Structural Database) where an enormous number of high resolution small molecule structures can be found - comparing the statistical trends in structures from there to what you see in proteins can be an enlightening exercise. Given that a number of the small molecule structures were determined by neutron diffraction methods, this can also be a way to critically examine hydrogen bonding as well. And in an even more basic fashion, if it simply doesn't make sense, then nothing compels you to take that structure as a gold standard. If the positioning of the electron transfer cofactors seems odd given what you know about the physiological rates, then it may very well be an artifact of crystallization. (Go read a couple of structure papers, with an emphasis on the materials and methods section. See how many crystallize proteins at non-physiological - even non-ambient laboratory - temperatures.) Does it look like there may be some electrostatic repulsion between the two surfaces of the interacting macromolecules? It may be that there needs to be a bound water to form some hydrogen bonds and minimize that repulsion.
The main reason I find this topic important is because researchers outside the structural biology community utilize this sort of information and expect that they are looking at something which reflects what the macromolecule really looks like, for lack of a better phrase. There needs to be care taken that if you're working with a protein or other macromolecule that does something interesting, you should be able to assay for that activity so you're not spending all that effort on examining an inactive assembly of amino acids (or what have you). It's crucial that we not lose sight of the fact that these beautiful structures should be able to be connected to the system's physical properties and biological structure.
|
Kleywegt and Jones (1995;Structure, 3: 535-540) describe some of their always interesting work in trying to improve validation of macromolecular structures, namely, intentionally tracing a protein backwards and attempting to fit it to the obtained electron density to see what differences arise. The intentionally traced backwards model of the protein had an R-factor (to slightly oversimplify, a statistical measure to compare the experimentally observed diffraction pattern to the expected diffraction pattern given the determined structure) of 0.214 (21.4%), while the model that had been obtained with a more conservative protocol (without being traced backwards) had an R-factor of 0.251 (25.1%). As the authors mention in this paper, the threshold tossed about in the crystallography community for an R-factor whereupon structures could be problematic floated around 25% and then upwards. Thankfully, the authors showed that other measures (particularly an updated R-factor, the so-called R(free)-factor where one randomly excludes some of the reflections from the refinement, introduced by Axel Brunger) clarified which was the incorrect structure and protein trace. The rise in the usage of NMR spectroscopy in structural biology sets off an interesting train of thought - structures derived by NMR are typically presented as an ensemble of individual structures, with an RMSD value determined for ensemble of structures. The RMSD is typically for the protein backbone at the very least. The idea is that a low RMSD suggests a higher quality structure - which really can't be justified in my opinion. The wonderful thing about NMR is that you can probe protein dynamics (as well as the dynamics of whatever the protein is interacting with). What follows next is perhaps obvious to the astute visitor - if the RMSD is not within a few angstroms (or is glaringly enormous along certain stretches of the protein), there is likely something interesting going on there with protein motion, solvation, protein-protein interactions, protein-ligand interactions, or some other biologically relevant process. And that is what should really get you excited, not that you can say you have an RMSD of 0.87 angstroms.
One of the other things is that you needn't abandon your basic chemical/biological intuition. If there look to be ridiculous conformations in the determined structure, they may very well be mistakes or artifacts. If the Ramachandran plot (where one plots the backbone torsion angles phi and psi against one another, see here at Birkbeck College for a more detailed explanation) has points where you really shouldn't expect to see them, there may be something wrong. There is an enormous database (the Cambridge Structural Database) where an enormous number of high resolution small molecule structures can be found - comparing the statistical trends in structures from there to what you see in proteins can be an enlightening exercise. Given that a number of the small molecule structures were determined by neutron diffraction methods, this can also be a way to critically examine hydrogen bonding as well. And in an even more basic fashion, if it simply doesn't make sense, then nothing compels you to take that structure as a gold standard. If the positioning of the electron transfer cofactors seems odd given what you know about the physiological rates, then it may very well be an artifact of crystallization. (Go read a couple of structure papers, with an emphasis on the materials and methods section. See how many crystallize proteins at non-physiological - even non-ambient laboratory - temperatures.) Does it look like there may be some electrostatic repulsion between the two surfaces of the interacting macromolecules? It may be that there needs to be a bound water to form some hydrogen bonds and minimize that repulsion.
The main reason I find this topic important is because researchers outside the structural biology community utilize this sort of information and expect that they are looking at something which reflects what the macromolecule really looks like, for lack of a better phrase. There needs to be care taken that if you're working with a protein or other macromolecule that does something interesting, you should be able to assay for that activity so you're not spending all that effort on examining an inactive assembly of amino acids (or what have you). It's crucial that we not lose sight of the fact that these beautiful structures should be able to be connected to the system's physical properties and biological structure.
|
