Index
I. Backround Information
II. Crystal Growth
III. The Vapor Diffusion Method
IV. Crystal Analysis
V. Filling in the Gaps
 

Background Information [back to top] We are trying to find the optimal crystallization conditions for horse skeletal muscle myoglobin. Myoglobin is a relatively small molecule whose structure is well-known from previous x-ray diffraction studies. This, coupled with the fact that it contains iron, makes it particularly useful to x-ray crystallographers. One of its common applications is as a calibration standard at the beamlines at SSRL. So why are we attempting to crystallize a molecule with such a familiar structure? In crystallography, higher symmetry of the sample leads to better and more easily-interpreted diffraction patterns. Also, a single type of protein may crystallize in several different forms (categorized by space group designations), depending upon the source of the protein and the experimental conditions. Sperm whale myoglobin crystallizes with high symmetry but has become nearly impossible to obtain, whereas myoglobin from horse and cow skeletal and heart muscle is much cheaper but less compliant in terms of forming desirable crystals. Furthermore, when protein crystals are placed directly in the path of an x-ray beam, they will undergo cracking and blistering from the intensity of the x-rays if left unprotected. To prevent this from occurring and destroying samples which have taken weeks or months to grow, crystals are frozen in a cryoprotectant (cryo) solution and placed in a stream of liquid nitrogen at 100K during analysis. Some protein crystals may be grown directly in a cryoprotectant by the vapor diffusion method, as described below, but it is often difficult to find a cryo solution whose pH and precipitant concentration will promote crystallization of a certain molecule.

Crystal Growth [back to top] The myoglobin (Mb) used in these experiments is a fluffy, reddish-brown solid. It is dissolved in deionized water to form a solution of 20mg/mL which is then sterilized by filtration and centrifugation.  With the horse myoglobin, we tested two sets of 48 prepared cryo solutions from Emerald BioStructures. Each vial contains a different amount of precipitant such as polyethylene glycol (PEG), various salt solutions in which the protein is insoluble, and often a buffer such as acetate or phosphate to stabilize the solution at a certain pH. For example, one of the non-cryo solutions we tested included 35%(v/v) 2-methyl-2,4-pentanediol (MPD), 100mM imidazole at pH 8.0, and 200mM magnesium chloride.

The photo above shows myoglobin being weighed on an analytical balance. One of our primary goals is maintaining the reproducibility of the experimental setup, so all volumes are converted to masses and combined on the analytical balance. This is especially important in the creation of screens, sets of solutions with slight variations of known crystallization conditions. Typically, one might come very close to the proper precipitant and pH levels with a prepared well solution and then screen around those parameters to produce crystals good enough for analysis by x-ray diffraction.

To the left is a set of twenty-four cryo solutions. These vials are refrigerated when not in use, and most of the experimental setup, aside from the creation of protein solutions and screens, is done in a cold room kept at 4ºC to prevent evaporation of small volumes and promote crystal growth.

The Vapor Diffusion Method [back to top]

In the vapor diffusion method, the bottom of a small cylindrical well is covered with a cryo solution, and a drop of protein plus cryo solution is suspended over the well on a glass coverslip. Because the concentration of the cryo solution is greater in the well than in the coverslip droplet, water diffuses out of the drop and into the well.

It is necessary to seal the wells with a small amount of oil or gel on their rims to prevent water from evaporating into the atmosphere.  Each tray has twenty-four wells; only alternating wells are greased so that the coverslips lie flat.

Obviously, the task of crystallization is an extremely delicate one. Typical volumes are approximately 650mL of solution in the well, with 2mL of protein solution and 2mL of the same well solution in the hanging droplet.

A micropipet is used to deposit the required amounts on the coverslip, as seen on the left, and tweezers are then used to transfer the coverslip to and invert it over the appropriate well. Each completed tray is covered, labelled, and left undisturbed in the cold room until it is reexamined for signs of crystal growth after a week or so.

If the rate of diffusion is ideal, the protein molecules which were initially soluble will align in a highly-ordered fashion, resulting in the growth of a crystal. If water diffuses too quickly, one might see the formation of an amorphous precipitate or a gel; on the other hand, the protein may remain soluble for an indeterminate amount of time.

Crystal Analysis [back to top] Crystallization can occur after a few hours have elapsed, or it may take weeks or months to note any changes in the hanging drops. Many crystals are visible to the naked eye, but the normal procedure is to observe each coverslip under a microscope at 40x or 50x magnification to determine whether or not any crystals are of sufficient size and quality for analysis.  Below are images of some common results seen under the microscope; even an inexperienced observer would be able to distinguish between the different cases in less than a second.

Here the protein has remained soluble in the well solution, as evidenced by the absence of any solids in the drop. The protein very well may crystallize in the future, though, so drops like these are reexamined at regular intervals (i.e., every other week). Sometimes the protein will denature before it precipitates, resulting in the loss of color for some molecules such as myoglobin.

This is an example of a precipitate, where the concentration of salts may have been too high, causing the protein to separate from the solvent in a random fashion. One of the most commonly observed results, this basically proves a certain well solution to be useless in promoting crystal growth since it is highly unlikely that the protein molecules will align themselves after precipitation.

Gel precipitates are also seen in the hanging drops. Here the protein is not quite soluble, nor has it precipitated completely. Gel formation does not preclude the growth of crystals, which may form on top of the gel.

A crystalline precipitate is seen here. There are no individual crystals, but the precipitate is obviously more crystalline than a gel due to the presence of incomplete edges in the drop. This drop would probably shimmer under the microscope with a change in the angle of incident light.

This is what crystallographers would like to see in coverslip droplets. These are obviously well-defined crystals, though they might not be suitable for x-ray diffraction analysis if they are inseparable or extremely hard (evidence of salt, rather than protein crystal growth). Desirable characteristics are an average crystal size of 200mm, the presence of single crystals, and, in the case of myoglobin, an opaque brown coloring.

Once a large, single crystal is selected for analysis in the beam line, it must be mounted on a nylon or glass fibre loop. The loop and crystal are then frozen, or flash-cooled, in liquid nitrogen to avoid damaging the sample during exposure to high-intensity x-rays.

Finally, the sample is placed in the x-ray beam so that diffraction patterns can be collected and a determination of the structure of the crystal itself or its contents can presumably be made. The picture on the right shows the experimental hutch at beam line 9-2 at the Stanford Synchrotron. At the lower center of the image is the charged-couple device (CCD) detector (white rectangular box), akin to a regular television camera with a phosphor screen, that captures and records diffraction patterns. Capturing data on X-ray film is a relatively tedious method which is rarely used. Current methods of data collection and analysis are highly dependent upon computer technology; a large amount of autoindexing and mathematical approximations must be done before one can determine the quality of the results. Good results are those that provide a high resolution map (less than 2Å) of the structure at hand, though it may take a number of different methods of diffraction analysis before one can actually solve a structure, especially for macromolecular crystals.

Filling in the Gaps [back to top]

Lawrence Livermore National Laboratory maintains an excellent introduction to and reference for crystallography, Crystallography 101. Here is a PowerPoint slide show on the basics of our summer work. Want the same information in more detail? Here is a written report of our work.

TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT

Back to home.