Q is for QB3

Biochemistry professor David Agard has long appreciated UCSF's collaborative culture and the depth and influence of its biological research — a record his own work has advanced. Yet for all of the university's prowess, there was something missing in the equation: a focus on cutting-edge quantitative biology. Until now.

Thanks to Governor Gray Davis' California Institutes for Science Innovation initiative, which created a collaborative research enterprise among UCSF, UC Berkeley and UC Santa Cruz, Agard has not only a dream, but also a plan and a building, which will take physical shape on UCSF's new Mission Bay campus. The UCSF portion of this enterprise, called the Institute for Quantitative Biological Research, has been dubbed QB3. "The idea of three major universities coming together to pool their resources and tackle a wide class of fundamental problems in biomedical sciences is extraordinarily exciting," says Agard, now the Scientific Director of QB3.

The whole mission of QB3, Agard explains, is to focus on a broad range of quantitative biology, especially areas that are just undergoing rapid development. "These are the things that are setting the stage for the future," he says. The QB3 building, which should be complete by mid-2004, will function as an enticement for recruiting faculty working in these emerging fields, as well as a foundation for a whole new way of collaborating between academia and industry. "We're trying to see if we can foster a much closer partnering than was possible before," he says, "while still retaining intellectual rights and the academic freedom of the UCSF community."

Making Biology Visible

Not only is Agard spearheading the idea of a new type of collaborative, multidisciplinary biology with QB3, his own research is also making biology itself more visible.

Agard has been collaborating with genetics professor John Sedat at UCSF for 20 years using high-resolution imaging to look at very complex problems in biology, such as the structure of chromosomes and centrosomes, also called the "microtubule organizing center."

While the centrosome has long been known to play a role in organizing the microtubules — which provide the framework for cell division, the transport of materials within cells and general cell structure and shape — the underlying machinery remains a mystery. For this reason, it is necessary to have a range of imaging techniques to take in the whole picture — from the atomic to the whole cell.

There have been several key gaps in the technology to view this entire range, and Agard and Sedat have been filling these gaps. The highest resolution that can be achieved currently is by X-ray crystallography and NMR, which are very good for determining the structure of molecules, but don't work well on very large complex assemblies.

Agard and Sedat have pioneered new microscopic techniques that will go where X-ray crystallography leaves off, essentially up to the level of the whole cell.

Called electron microscope tomography, which is the subcellular equivalent of CAT scans on people — it uses exactly the same mathematical concepts, taking many different views from different angles. From the whole set of hundreds of views, a picture can be reassembled to provide the internal structure of the sample. This seemingly straightforward technique actually dates back to the late '60s, but it turns out that there are "loads and loads of critical but annoying little technical problems that have to be solved," says Agard.

The concept is very simple, but making it work in practice is a significant challenge, so challenging that less than a dozen labs in the world use this technique.

"The utility is such that every cell biologist would love to put a sample in and do this automatically," he says, so they have chiseled away at the problems and have successfully automated many of the features to make tomography a usable reality.

The two UCSF scientists are closing in on the automation part, aided by the manufacture of a new generation of electron microscopes designed to be run by computers. Agard's lab has just taken delivery of one of the most powerful of this new generation of microscopes, soon to be operational.

The new microscope will give them a two- to threefold improvement in resolution, which comes at a critical point in their studies.

"What we've been doing is almost good enough, but we've been frustrated and held back by needing just a little bit more resolution. This factor of two or so should push us well over the edge."

The problem is that in a structure such as the centrosome or the chromosome, they need to make molecular interpretations of what's going on, and the 50 to 60 angstrom resolution that they have now is just not enough. Individual proteins look like blobs, Agard says.

When they can get down to 20-25 angstrom resolution, then they could see subunit structure and tell the shape of one protein from another. "This will really be enabling for bridging the gap between the molecular and cellular structure levels.

We're just thrilled with what we think we'll be able to do," says Agard. "Our hope is that it will give us pictures of such incredible clarity of what's going on inside of cells that it will be revolutionary."

A Light Squeeze

Equally revolutionary work involves the far end of the visualization spectrum, the realm of whole, live cells. The main problem has always been that the resolution was limited by the wavelength of light itself. A postdoc in Sedat's lab, Mats Gustafsson, has invented two new classes of light microscopes that get much, much higher resolution than was ever thought possible.

Instead of directly viewing a sample through an eyepiece, Gustafsson instead focused on what information could be squeezed through the lens.

Then, as Agard says, "you can play tricks, but it takes a computer to sort out those tricks and generate an image that makes sense. It's astoundingly clever in terms of the underlying concept, as well as wonderful engineering."

Page 2 - Cassman assumes QB3 post