Nuclear Structure and Function Research Group

The DNA molecules that run the length of each human chromosome are arguably the longest and most important biomolecules known. But although we know their DNA sequences, we still know little about how that DNA is folded in 3-D space within the nucleus of a living cell. Common sense suggests there must be some underlying order within the apparent tangle, and our current research is concerned with elucidating what that order might be. This research is based on the belief that we can only understand function if we understand the details of the architecture, and vice versa. Our ultimate goal, then, is to generate structural and functional maps of nuclei using a multi-disciplinary approach (e.g., combining molecular analysis, high-resolution imaging, modelling).

We try to preserve native structure during our studies in vitro; we also take great pains to preserve polymerizing activity. Unusually, then, we conduct our experiments in vitro in 'physiological' buffers. [Chromatin and polymerases are almost invariably studied using hypotonic salt concentrations.]

Functional sites within nuclei have traditionally been labelled using short pulses of radioactive precursors and autoradiography. However, such sites can now be marked with higher resolution. Living or permeabilized cells are allowed to make nucleic acids or proteins in the presence of non-radioactive analogues of the natural precursors; then, incorporated analogues are immunolabelled with antibodies conjugated with fluorochromes or gold particles. The following sites can now be labelled (we developed approaches 2-5):
(1) Replication sites (e.g., with BrdU, Br-dUTP, bio-dUTP or fluorescein-dUTP).
(2) Repair sites (e.g., with biotin-dUTP).
(3) Transcription sites (e.g., with Br-U, Br-UTP or biotin-CTP).
(4) Transcripts in transit to the cytoplasm (e.g., with Br-U).
(5) Translation sites (e.g., with biotin-lys-tRNA, BODIPY-lys-tRNA).
Generally, incorporated tags are not diffusely spread throughout euchromatin, but concentrated in a limited number of discrete sites. For example, nascent transcripts generated by RNA polymerase II are found in several thousand 'foci' or 'transcription factories' with diameters of ~90 nm. In various higher eukaryotes (human HeLa cells, mouse embryonic stem cells, newt cells) a typical nucleoplasmic transcription factory contains ~8 molecules of RNA polymerase II each transcribing a different template. Many of these factories are also actively engaged in translation – a process that was believed to occur only in the cytoplasm.

Our studies have led us to different views about the way the vital processes of replication and transcription occur, and how genomes are organized (model, illustrative movie).  We are currently analyzing the molecular content of these factories (particularly transcription factories). We are also determining their size, shape and numbers as cells progress around the cell cycle and differentiate. Increasingly, we are imaging sites in living cells using molecules like histones and polymerases tagged with the green fluorescent protein. Our results led us to a model for genome organization in which polymerases engaged on chromosomes cluster together to loop the intervening DNA; therefore, we are studying the size and distribution of these loops, and the molecular ties that maintain them.

We are also collaborating with physicists/mathematicians in an ambitious project – our ultimate goal is to write equations that will enable us to predict both the way genomes are folded in 3-D space, and how folding determines function.  We began by asking what forces might drive clustering of active polymerases (and so factory formation), and we have stumbled on what we believe is a general organizing principle.  Active polymerases in a crowded cell are modelled as beads threaded on a string.  Paradoxically, their aggregation increases the entropy of the system by increasing the entropy of the many small crowding molecules; this occurs despite the costs of looping.  Entropic forces (acting through what is known as the 'depletion attraction') inevitably drive active DNA and RNA polymerases into replication and transcription factories and the genome into loops!  We have also analyzed how the position of a gene in a loop affects the probability that it might access binding sites in a factory.  Loops (with varying lengths, thicknesses, and rigidity) are modelled as semi-flexible (self-avoiding) tubes attached to spheres.  Results provide insights into the packing problem (how long genomes are packed into small nuclei), action-at-a-distance (how firing of an origin/gene prevents firing of an adjacent one), and nuclear context (why genes are active in certain positions but not others). We are now taking the first steps towards modelling individual chromosomes packed into the nucleus, and – as we gain experience – we will try to model more and more chromosomes until we build up a complete picture.

A review:
Cook, P.R. (2010). A model for all genomes; the role of transcription factories. J. Mol. Biol. 395, 1–10. [PubMed]

Top | Home | Maintained by Peter Cook |