STRUCTURE AND DYNAMICS OY DROSOPHILA DIPLOID NUCLEI 
John W. Sedat, Ph.D., Investigator 
The extremely rapid progress in the elucidation 
of DNA sequence information in eukaryotes has 
highlighted the quest for an architecture for chro- 
mosomes, from interphase to mitotic. Such an ar- 
chitecture will lead to a unity of structure and func- 
tion. Dr. Sedat and his colleagues, in collaboration 
with Dr. David Agard (HHMI, University of Califor- 
nia at San Francisco), have focused on structural 
questions probing nuclear and chromosomal orga- 
nization. They seek answers to three questions. 1) 
How do interphase chromosomes fold up within 
the intact diploid nucleus, and how, in detail, does 
an interphase chromosome change as a function of 
progression through the cell cycle or development? 
2) What is the defined interphase chromosome ar- 
chitecture of a specific gene? 3) Is molecular infor- 
mation reflected in structural attributes? Because of 
the extensive knowledge of its genetics, develop- 
ment, and biochemistry, Drosophila melanogaster 
is an ideal biological system for these integrated 
structural and functional studies. 
The development of generalized computational- 
based tools for three-dimensional optical micros- 
copy (OM) and electron microscopy (EM) of large 
noncrystalline subcellular structures has been de- 
scribed. In a resolution-overlapping fashion, data 
are collected with charge-coupled devices (CCD), 
processed to remove observational problems, and 
analyzed quantitatively. 
The microscopy tools that are required for the 
structural questions are being developed, both at 
the OM and EM levels. The CCD for the EM430 
(the HHMI state-of-the-art intermediate voltage EM) 
was brought on line and very high-quality data have 
resulted. This past year, four-dimensional OM imag- 
ing (three-dimensional images as a function of 
time) in living embryos was perfected and became 
routine. In addition, computer-controlled wave- 
length selection filter wheels with double-wave- 
length dichroic mirrors working in tandem were in- 
corporated into the OM, so that rapid wavelength 
changes could be made. This allows Dr. Sedat and 
his colleagues to record fluorescent images from 
several cellular structures as a function of time. 
Modifications to this instrumentation, together 
with new software, have greatly speeded up the 
rate that three-dimensional data are collected. Soft- 
ware is being written to correct for systematic 
image acquisition problems, to display these data 
in a variety of formats, and to model and analyze 
(in many cases quantitatively) the intricate three- 
dimensional data. 
Dr. Sedat and his colleagues continue to study 
the structure of the diploid nucleus in Drosophila, 
using the prophase and anaphase cell cycle time 
points as examples. Several diploid nuclear struc- 
tures have been modeled, and specific chromo- 
somes have been assigned to three-dimensional 
chromosomal paths. Several exciting structural fea- 
tures, including mirror-symmetric homologues at 
prophase, were described. A major effort is under 
way to model many examples, so that statistical 
comparison (one nucleus with another) can be 
made. Software has been written to model the 
chromosome paths within nuclear structures rap- 
idly and computationally, so that large numbers of 
nuclear examples can be determined. 
I. Real-Time Three-dimensional Optical Microscopy 
of Diploid Nuclear Structures. 
Chromosome and nuclear structures were visual- 
ized in living embryos, using a procedure devel- 
oped by Dr. Jon Minden in the laboratory of Dr. 
Bruce Alberts. Rhodamine-labeled histones were 
microinjected into embryos, where they were stably 
incorporated into nuclei. Three-dimensional images 
(e.g., 80 sets of three-dimensional data taken every 
25 s) were collected. These three-dimensional im- 
ages were projected into stereo pairs, followed by 
computational processing to remove out-of-focus 
information. The ability to study three-dimensional 
nuclear structures as a function of time in living 
embryos made it possible to identify functional cor- 
relations quickly. These structures showed that the 
four-dimensional nuclear images were excellent 
controls for fixation artifacts in the previous static 
diploid nuclear studies. Careful comparisons of the 
real-time and fixed-images data have indicated that 
the fixed samples were minimally perturbed. So far 
no structural discrepancy has been noted. The real- 
time study reveals that there were discrete chromo- 
some condensation sites (—2-3 per nucleus) at- 
tached to the nuclear envelope that were the last to 
decondense during telophase. At the subsequent 
interphase-to-prophase transition, bright chromo- 
some condensation spots were again observed at 
discrete sites on the nuclear envelope in a similar 
position to that of the late decondensation sites. At 
present it is not known if these are the identical 
Continued 
597 
