Dr. Brown is also Assistant Professor of Pediat- 
rics and Biochemistry at the Stanford University 
School of Medicine. 
Articles 
Chow, S.A, Vincent, K.A., Ellison, V., and Brown, 
P.O. 1992. Reversal of integration and DNA splic- 
ing mediated by integrase of human immunodefi- 
ciency virus. Science 255:723-726. 
Tsuchihashi, Z., and Brown, P.O. 1992. Sequence 
requirements for efficient translational frame- 
shifting in the Escherichia co/? dnaX gene and the 
role of an unstable interaction between tRNA''''' 
and an AAG lysine codon. Genes Dev 6:511-519. 
TARGETED MUTATIONS IN EMBRYONIC STEM CELLS: 
A MOLECULAR GENETIC DISSECTION OF DEVELOPMENT 
Mario R. Capecchi, Ph.D., Investigator 
Gene targeting in mouse embryo-derived stem 
cells can be used to introduce designed modifica- 
tions into virtually any gene in the mouse. The ad- 
vantage of this technology is that the investigator 
chooses both the gene and the mutation. Because of 
the limitations imposed by the generation time of 
mice and the size of the available colonies, it is not 
practical to rely on random mutagenesis to generate 
mice with specific mutations. Thus gene targeting 
provides the only available technology for achieving 
this goal. 
Dr. Capecchi 's laboratory is using gene targeting 
to determine the role of two sets of genes in mouse 
development. The first set is involved in localized 
developmental decisions through cell-cell signal- 
ing, and the second comprises members of a tran- 
scriptional complex involved in specifying posi- 
tional value along the anterior-posterior axis of the 
mouse. In order to make better use of the gene tar- 
geting technology, parameters that influence the 
gene targeting frequency have also been analyzed. 
Parameters That Influence Gene Targeting 
The capacity to create predetermined genomic al- 
terations is dependent upon the precision of the 
recombination reaction. Recently it has been 
suggested that "replacement vectors" mediate ho- 
mologous recombination with low fidelity. In this 
particular study, fewer than 5% of the recombinants 
were the result of simple legitimate replacement. 
This novel lack of fidelity prompted scrutiny of po- 
tential difi'erences in experimental protocols. The 
important variable turned out to be the length of the 
genome-homologous DNA sequence flanking the 
mutation. If the replacement vector was designed 
such that the desired mutation was flanked on both 
sides by several kilobases of DNA homologous to the 
target locus, replacement proceeded with high fidel- 
ity. If, on the other hand, the mutation was flanked 
on one side by <1 kb of target-homologous DNA, 
abortive recombination occurred. 
A parameter that strongly influences the gene tar- 
geting frequency is the extent of homology between 
the targeting vector and the target locus. However, 
there were discrepancies in the literature on the im- 
portance of this parameter in determining the tar- 
geting frequency. Furthermore, there were claims 
that "replacement" vectors and "insertion" vectors 
targeted with very different frequencies. To clarify 
this issue, a systematic analysis of this parameter was 
undertaken, using both replacement and insertion 
vectors. The vectors behaved similarly with respect 
to their targeting efficiency and dependency on the 
extent of homology between the targeting vector 
and the target locus. Furthermore, when a targeting 
vector behaved anomalously, an unfavorable distri- 
bution of heterology between the vector and the tar- 
get locus was shown to be the culprit. 
The Mouse box Genes 
Mice and humans contain homologues of many 
genes that control early embryonic development in 
Drosophila, including 38 genes that directly corre- 
spond to the Drosophila homeotic genes of the 
Ultrabithorax and Antennapedia complexes. The 
evolutionary conservation of these genes may re- 
flect the inheritance of a whole transcriptional pro- 
gram for specifying positional information in the 
embryo. However, no known mutations of these 
genes have been identified in either mouse or hu- 
mans that would allow direct assignment of func- 
tion. For this reason. Dr. Capecchi is seeking to pro- 
vide a genetic definition for the function of these 
genes by creating mice with null alleles in each 
gene. From such a systematic analysis, the laboratory 
hopes not only to reveal the phenotypes associated 
with inactivating any individual gene but also to de- 
fine, through epistasis and molecular analysis of 
162 
