Many thanks to those who responded to my query. A summary of the
responses follows. Not every example is included as this is not a
review article.
Clusters of physiologically related genes are widespread, found in most
eucaryotes where sufficiently detailed studies have been made. My
question was directed toward eucaryotes, although I did not specify
that in my initial query, and several responses reminded me of the
operon clustering common to procaryotes. Interesting in this regard was
Jason Cummings response pointing out the paper by Zorio et al. (1994
Nature 372:270-272) providing evidence for polycistronic transcription
of gene clusters of unknown function in the roundworm Caenorhabditis
elegans. Nancy Keller pointed out the speculation that the gene cluster
for the penicillin biosynthetic pathway may have been transfered
horizontally between Streptomyces and Penicillium/Aspergillus ca 370
MYA (Miller, J. R. and T. D. Ingolia. 1989. Cloning §-lactam genes
from Streptomyces spp and fungi. In C. L. Hershberger, S. W. Queener
and G. Hegeman (Eds.), Genetics and Molecular Biology of Industrial
Microorganisms. pp. 246-255. Amer. Soc. Microbiology). This is an
interesting speculation that would require changing the cluster from
polycistronic with one promoter to a monocystronic cluster with
individual gene promoters. Perhaps the C. elegans example would allow
the assumption that promoter reorganization could occur secondarily.
However, chain termination occurs at stop codons causing eucaryotic
ribosomes to fall off the mRNA, unlike procaryotic ribosomes that
terminate but remain attached. How to initate the next cistron? Maybe
the transfer was from fungus to bacterium?
Clusters may be put into several classes. Metabolic gene clusters
encoding different enzymes in a pathway in which the genes are
unrelated (I think) in sequence homology, as the genes encoding
penicillin biosythesis of fungi (see above), catecholamine synthesis in
Drosophila (Wright 1987 Results Prob Cell Differ 14:95-120), and purine
nucleotide synthesis in humans (Brayton et al. 1994 J Biol Chem
269:5313-5321). Another kind of cluster is the multigene family, as the
globins genes of vertebrates, and histone genes (textbook examples).
The globins are clearly sequence related. Another kind of cluster are
the developmental regulatory genes that may or may not be sequence
related. Homeobox containing genes popped up here, the HOX cluster of
metazoans (Kenyon 1994 Cell 78:175-180) and the mating-type genes of
fungi (Casselton and KŸes 1994 Mating-type genes in Homobasidiomycetes.
In J. G. H. Wessels and F. Meinhardt (Eds.), Growth, differentiation
and sexuality. Vol. 1, pp. 307-321. Springer-Verlag). Also noted here
are the genes of the Enhancer split complex of Drosophila (Knust et al.
1992 Genetics 132:505-518).
Functional ideas about clustering were less strong. Tandem gene
duplication as with the ribosomal RNA genes and the globin genes fits
only a few examples of sequence related genes. Position effects in gene
regulation was another idea; though such effects have been demonstrated
in other contexts, they have not held up as a raison d'etre for gene
clusters. What then could be the evolutionary processes that generates
such clusters and maintains them in some cases but not in others?
David H. Griffin
Department of Environmental & Forest Biology
College of Environmental Science and Forestry
350 Illick Hall
One Forestry Drive
Syracuse NY 13210-2788
e-mail: griffin at mailbox.syr.edu
