Crystallography for beginners – part 5 – monitoring and evaluating crystallization experiments results

So, you’ve set your first plate(s) as I have detailed in my previous post about setting up crystallization plates. That’s great, yet this is only the first half of a crystallization experiment. Now you need to monitor for crystal growth and interpret the result of the crystallization experiments so you know which future experiment to setup to obtain crystals (unless you’re lucky and already obtained a crystal in the initial experiment). It might sound like the fun part yet this step is crucial for obtaining crystals when those don’t come so easily (which is the case for most proteins).

What do you need?

Since protein crystals are pretty small, ranging from 0.001 mm to approximately 0.1 mm, you will need a stereomicroscope to observe them. While it might seem exciting to look for protein crystals, this is a tedious work, scanning hundreds of drops for crystal growth. Today, with the introduction of robotics to the life science applications, there are numerous robotics for imagine of protein crystal that can make the life of a crystallographer more comfortable. However, since many crystallography labs don’t have such robots, you will most probably do this manually. Take into consideration that this work is also  time consuming since you can quickly accumulate dozens of plate that need to be monitored on a regular basis just by using several crystallization screens with couple of concentrations and at different temperatures  You’ll need also a pen and a paper to mark yourself the name of the plate, position of the “hits” and their crystal/precipitate morphology.

When should you monitor your plates?

You’ll be most curious and eager to look at your plate right away and you SHOULD do just that. Looking at the plate at roughly time “0” will give you the assessment of the starting point of this dynamic and complex process. You should monitor the plate(s) each day for the first week and then check the plate once or twice per week up to two months, after which you should monitor your plates once or twice a month. The best way to monitor the changes over time is by taking photos and observing the change over time. This of course requires access to a stereomicroscope coupled to a digital camera and is time-demanding. Thus, unless you have access to an imaging robot (such as this one) you should use a scoring scheme such as based on numbers (see pictures below) coupled to an excel sheet 12×8 or 6×4 matrix (depending on the type of plate):

24-well crystallography plate scoring

The above plate is scored according to initials (C, clear drop; PS, phase separation; P, precipitate etc.). You can also see the exact details of every experimental factor (experimentation and monitoring date, protein and crystallization agent concentrations, temperature and even the origin of the specific conditions taken from a screen).

What’s in the drop and how should you proceed?

In protein crystallography our aim is to transform the liquid-state protein units into orderly-arrayed solid-state protein array. Since the process is very sensitive to numerous environmental as well as in-solution factors, we will expect to witness either no aggregation (if the saturation conditions are not met) or several types of aggregates. The type of aggregate will give us clues whether we are on the correct track toward protein crystal growth.

Following is a picture gallery (with a suggested number-scoring scheme) which represents most cases of drop appearance:

Clear drop (0)

The drop is clear without any signs of aggregations. This means that either protein or precipitant concentration is low (or both).

Contaminated drop #1 (Glass/fiber) (0)

These are non-biological contaminates that entered the drop while the plate was prepared. Glass (upper image) can be easily mistaken for crystals – monitoring these over time will prove their identity. Fibers might change their color under polarized light yet their curvature nature signal them as non-crystal species. Note that close to the fiber you can see translucent protein precipitates.

Granular Precipitate (1)

These are round-like precipitates that usually form small clusters. The brownish kind are usually dead-ends and even after several steps of optimization in rare cases these will eventually lead to crystal formation. Translucent precipitates, however, might be a better starting position for further optimization.

Full precipitate (2)

Full precipitates are usually brownish forms of aggregates which in most cases are dead ends. In many cases this phenotype is due to the presence of too high protein/precipitant or of both leading to the protein crashing onto large aggregates. Even so, keep track of these as some protein crystal can form out of these too (and by doing so, clearing the area surrounding the crystal).

Phase separation & oil drops (3)

These are highly concentrated protein aggregates that are packed together yet not in a crystal lattice, generating round shapes (upper image) or gelatinous shapes (lower image). This is a good place to optimize conditions, especially temperature, since the process of crystal packing might need to be slower/faster (and temperature controls that). If nothing helps, you might consider re-cloning and chop off possible tails from the protein.

Spherulites (5)

This is regarded as a certain form of microcystalline. Yes, this are usually perfectly round and they show the light befringing of crystals. Same as Phase separation, conduct an expansion grid.

Microcyrstalline precipitates (6)

This is a promising sign – such hits should be further optimized, in most cases requiring to lower the amount of nucleation events, usually by lowering the protein concentration/precipitant or lowering the incubation temperature.

1D crystals – needles (7)

Needles are one dimensional growing crystals, and as such these are on the one hand very promising, yet there is a need to improve the crystallization growth dynamics since only one dimensional is actively engaged in the growth process. Besides playing with protein concentration/precipitate, pH and temperature, you should also vary the buffer. If you find you just get more and more needles, with no improvement toward 2D or 3D crystals, then you should consider recloning and generating a new protein variant.

2D & 3D crystals (8 & 9)

These are the most promising and the sought for protein crystals.  Since protein units are packed in 2D and 3D dimensions  these can lead to high resolution diffraction data. As a rule, a crystallographer should aim at growing single large crystals. Looking at the upper image, two phenotypic elements an be observed: shape of the crystals and their number. Under suboptimal conditions the 3D crystals will look part round as not all crystals facets will be sharp, a strong indication that either the growth rate is too fast or that the protein concentration is too high. The large number of crystal units is a classic demonstration of over nucleation possibly due to high protein concentration in the drop. When the amount of protein is smaller, less nuclei form and thus leads to fewer yet larger protein crystals. Even so, when obtaining many small crystals (which has clear sharp facets) it is worth to challenge them under a strong radiation source such a synchrotron beam – sometimes that’s the best you get, and among these some might diffract beyond 2.5A.

One important note – a crystallography experiment is in progress until the drop dries up.

Salt crystals?! 

Yep, salts can quite easily crystallize and produce some very beautiful (and useless!) crystals in your drop. Since you are a protein crystallography (I presume), these will be a nuisance in your path to growing protein crystals. How can you differentiate the salt from the protein crystals? here are some tips:

  • Salt crystals are hard and dense, thus you will mostly see them at the bottom of the well. if you try to prick them with a needle you will see them bouncing around happily (that’s because salt crystals pack very well and thus doesn’t have high water content). Protein crystals, on the other hand, are characterized by a relatively lower packing density, which is filled by water molecules. This characteristics lead protein crystals to be found at the upper parts of the well (low density) and to be very fragile. A glancing blow with a needle will usually be sufficient to make the crystal crumble right in front of your eyes (that’s one good reason why you should not do that if this is your only crystal).
  • Salt crystals can withstand dehydration (fishing the crystal from the well and let it stay at the open air). protein crystals, on the other hand, will crumble, crack and disintegrate rapidly.
  • Salt crystals have a very strong bifinging properties and using a polarizer filter attached to the microscope you can see the strong colorful effect of the polarizing light. While proteins can also bifringe, it is of lesser strength.
  • UV absorbance – if your protein absorbs at UV wavelength (should contain the amino acids tryptophan and tyrosine) than under UV light protein crystals should shine very brightly while salt crystals will look like ghostly shapes. For this you will need an image robot that is also equipt with a UV lamp.
  • Dye – Special dyes (such as Izit) will penetrate the crystals (since it is loosely packed) and bind to the protein’s amino acids, lead to coloring of the crystals.
  • Diffraction – this is the ultimate test and if you have next door a home source with liquid nitrogen running this should be the only test required.

X-ray diffraction pattern of crystallized 3Clpro, a SARS protease. (2.1 Angstrom resolution) Jeff Dahl   Diffraction image of Brushite's crystal

The above image is a diffraction image of a SARS protease,  while the lower image represents a diffraction pattern of a phosphate mineral, Brushite. A (good) diffraction pattern from a protein crystal is commonly characterized by many spots across all resolution levels (low, mid and high) while that of a salt crystal is usually sporadic and mostly at the mid high resolution. Low resolution diffraction data is located close to the center of the diffraction image (close to the beam stop black circle) while the mid and high resolution extends to the outer rim of the image, the high resolution located at the edge of the diffraction image.

Time to celebrate?! Not so fast…

Even if you got a single, large protein crystal, don’t pop out the champagne just yet. The ultimate goal of a protein crystal is not its morphology but rather its packing and diffracting information. A large beautiful crystal can be poorly packed and thus both fragile and will diffract poorly or not diffract at all. In future post I will discuss the diffraction experiment, what should you expect and how to proceed from there.

Next post will discuss how one can rescue a difficult to crystallize protein – don’t miss it!

Drop photos & crystals are courtesy of the Zarivach laboratory and the Macromolecular crystallography research center, BGU.


10 comments on “Crystallography for beginners – part 5 – monitoring and evaluating crystallization experiments results

  1. beautifully ….written artticle …many thanks for your nice explainations….

  2. It’s a pretty good article, very helpful and informative. Thanks.


  3. Just starting some crystallography experiments, and very excited (and curious) about it (I’m a cellular biology with no knowledge of protein crystallography)..and found your explanations very clear! Hopefully the crystal we found today is not salt! 🙂 Thanks for sharing.

  4. Very informative indeed. thanks.

  5. Hello,

    Regarding the Salt and protein crystals I am a little confused.
    I learned from several different sources that it is exactly the opposite of what you say.

    Salt crystals are very fragile and easy to break, and you can even hear a distinct ”click” sound when you break it. (which is kind of surprising since its such a tiny tiny crystal but the sound is as loud as dropping your pen on the table)
    And protein crystals are very strong and hard to break.

    Please correct me if I’m wrong 🙂

    P.s. When will part six be continued? 🙂

    With Kind Regards

  6. Thank you so much Chen for such a nice explanation. I have done robotic screening just a week ago and looking forward for my crystals to grow. I was confused to observed my plates but this article makes easy for me. Thanks a lot for sharing this…..

  7. Hi,
    Nice notes and photos.
    Just wanted to mention that you may consider changing “that’s because salt crystals pack very well and thus doesn’t have low water content” to ” that’s because salt crystals pack very well and thus doesn’t have high water content”.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: