news and views

cation of the introduced DNA and its segre­ been available for many years, and they structure5• In either case, the replication of gation into daughter cells. At least two function when introduced into cells as MACs may not require careful provision for approaches are being explored towards this naked DNA. Mammalian geneticists, on the origins. goal. The first uses plasmid vectors carrying other hand, suffer from species envy, largely The main stumbling block for the con­ parts of the Epstein-Barr-virus genome that having had to be content with functional struction ofMACs, which Harrington et a/. 1 allow the retention and replication of DNA reintroduction of telomeres4• Moreover, have now overcome, has been the provision under the control of the host cell-cycle3• whereas origins of replication are defined of a centromere and the kinetochore that Another is to re-create the natural cellular sequence elements in yeast, the existence of assembles there. The centromere is not sim­ vehicle for DNA replication and segregation analogous mammalian origins remains ply a passive point at which the chromosome - the human chromosome. controversial. One suggestion is that poten­ attaches to the micro tubules of the mitotic The minimal requirements for an artifi­ tial sequence-specific origins are frequent in spindle - it actively participates in the seg­ cial chromosome include sequences to allow the genome. Another is that replication sim­ regation of chromosomes to daughter cells. for the maintenance of its ends (telomeres) , ply requires a threshold amount of mam­ The centromeric region contains at least one replication of its DNA, and mitotic segrega­ malian DNA, with the specific patterns of motor, which is involved in chromosome tion to daughter cells ( centromeres) (Fig. l ). replication initiation that are observed movement during cell division. The cen­ Yeast artificial chromosomes (YACs) have being created by overall chromosome tromere also mediates a critical checkpoint that prevents completion of cell division Ordering upaMPC until all chromosomes are properly attached to the mitotic spindle6• The centromere ( CEN) in Saccharomyces cerevisiae is a 125- base-pair DNA sequence; human cen­ tromeres, on the other hand, contain up to several million base-pairs of alpha satellite DNA, a large, tandemly repeating 171-bp DNA family7• Previous reintroduction stud­ ies had indicated that alpha satellite DNA could provide some aspects of centromere function8•9, but the microchromosomes constructed by Harrington et al. (see panel) are the first to have all the features necessary to designate them MACs. The successful creation ofmicrochromo­ somes with centromere function from intro­ duced naked DNA is a considerable step towards a MAC-based vector system for To make an artificial least two cases of de microchromosome gene therapy. Some problems remain, such chromosome, Harrington novo centromere (arrow) was detected as the incompletely defined DNA that is et a/.1 obtained formation seemed to using antibodies against used, and the variable and unpredictable megabase-sized alpha occur on the introduced CENP-G (green), and the structure, and relatively low yield, of satellite arrays, alpha satellite DNA. One chromosome-1 7 alpha the microchromosomes. This is further approaching the size of such microchromosome, satellite was detected by compounded by undesirable genomic those at natural created from transfected in situ hybridization (red). rearrangements, which are caused by the centromeres. Starting DNA including alpha b, A similar experiment introduced telomeric sequences. Ultimately, with a single satellite from human using antibodies against a practical MAC system for gene therapy characterized alpha chromosome 17, is CENP-E (green). Co- must use defined DNA and yield microchro­ satellite repeat unit. they shown. In most cells localization of green and mosomes in a predictable and efficient fash­ used a directional- only a single red signals leads to a ion. Even then, it still faces the problems of cloning method in a low- microchromosome is yellow signal. In both immune reactions against the therapeutic copy bacterial artificial seen, indicating faithful cases, the gene product, cell-specific targeting and so chromosome vector, maintenance of copy- chromosomes were forth. Manipulation and delivery of such followed by in vitro number control. These stained with the DNA large DNA molecules is a technological chal­ ligation, to make tandem microchromosomes are dye DAPI (blue). These lenge - these MACs are the whoppers of arrays up to 1 Mbp in stably inherited for r:na ny microchromosomes gene-therapy vectors - and it will be inter­ size. This DNA was then generations in the seem to be linear and esting to determine just how small a cen­ mixed with human absence of selection, they are 6-10 Mbp in tromere-containing human chromosome (TTAGGG lntel om eric with a loss rate of less size, but they are can be. MACs do not have the stringent size sequences and high than 0.5 per cent per cell composed of a complex limits ofm anyviral-delivery systems, so they molecular-weight division. The formation arrangement of may allow for natural patterns of gene genomic DNA, and of centromeres was sequences, expression by introducing large genomic introduced into cultured further shown by the characteristic of non- fragments carrying the therapeutic gene in human cells using ability of the homologous its natural genomic environment. cationic liposomes. microchromosomes to recombination of the Microchromosomes should also aid our Although most clones bind two known introd uced DNA cocktail. understanding of chromosome mainte­ represent integration or centromere-associated (Figure reproduced nance and transmission in both mitosis and chromosome- proteins, CENP-c and courtesy of Harrington meiosis. For example, by custom-building fragmentation events, at CENP-E. a, The et aC.) P.E.W.&D.K. arrays of mutated centromeric sequences from human and other species, the sequence

664 NATUREIYOL386.IlOAPR!L 1997