scientific correspondence
with other methods, such as T-DNA inser-
a R X b RNA
WT KO WT KO WT KO
6.0
4.1
3.9
2.8
1.3
Figure 2 Genetic analysis of agl5 mutant. a,
Genomic DNA blot with 6.0 kb Eco RI fragment as
probe. The 3.9-kb Xba I fragment was present in
wild-type and mutant plants, whereas the 1.3-kb
Xba I fragment was present only in the wild type.
This same probe hybridized to a 6.0-kb Eco RI frag-
ment in the wild type and to the expected 4.1 and
2.8-kb Eco RI fragments in the mutant. b, A probe
specific for the 3፱ end of the AGL5 complementary
DNA detected the AGL5 transcript in wild-type but
not in the knockout mutant plants. Details of
methodology available from the authors on request.
gous mutant plants. Recent improvements
in intact plant transformation technology2
have provided a more efficient method for
generating large numbers of transgenic
plants, so that it should be possible to iden-
tify a targeted insertion, even if the frequen-
cy of such an event is relatively low.
We used a polymerase chain reaction
(PCR)-based assay of transgenic plants to
identify targeted insertions into AGL5 3,4.
The targeting construct consisted of a
kanamycin-resistance cassette that was
inserted between a roughly 3-kilobase (kb)
segment from the 5፱ end of the AGL5 gene
and a 2-kb segment at the 3፱ end (Fig. 1).
We produced 750 kanamycin-resistant
transgenic lines by Agrobacterium-mediated
transformation, and analysed pools of
transformants using PCR to determine
whether any of these primary transfor-
mants contained the desired insertion into
AGL5. We found a single line that seemed to
contain the anticipated insertion, and this
line was allowed to self-pollinate to permit
further analyses in subsequent generations.
We analysed genomic DNA from the
homozygous mutant plants with five differ-
ent restriction enzymes and using several
distinct PCR amplifications. All data were
consistent with the desired targeting event
(Fig. 2a and data not shown). Additional
genomic DNA analyses and PCR amplifica-
tions confirmed that there were no
detectable deletions and/or rearrangements
outside the targeted region. As expected, we
detected the AGL5 transcript in wild-type
but not in agl5 mutant plants (Fig. 2b).
Targeted disruption by homologous
recombination provides a new tool along
NATURE | VOL 389 | 23 OCTOBER 1997
We studied a species of the deep-sea fish
tion and transposon tagging, for obtaining
loss-of-function (knockout) alleles of
cloned genes in Arabidopsis. Homologous
recombination has a number of advantages
over these methods, being a relatively rapid
procedure (roughly six weeks from the time
Arabidopsis plants are transformed to the
time that a targeted disruption can be iden-
tified) that allows precise construction of
both knockout and more subtle ‘knockin’
mutant alleles5.
Although we have not yet determined
the precise frequency of homologous
recombination, continued improvements in
plant transformation technology and the
use of negative selection strategies suggest
that targeted disruption will become a rou-
tine tool for molecular and genetic studies
of Arabidopsis. This is now the only higher
eukaryote in which both large-scale random
mutagenesis and targeted disruption by
homologous recombination are feasible.
Sherry A. Kempin, Sarah J. Liljegren
Laura M. Block, Steven D. Rounsley
Martin F. Yanofsky
Department of Biology,
Center for Molecular Genetics,
University of California at San Diego,
La Jolla, California 92093-0116, USA
e-mail: marty@ucsd.edu
Eric Lam
AgBiotech Center, Foran Hall, Dudley Road,
Rutgers the State University of New Jersey,
New Brunswick, New Jersey 08903, USA
1. Miao, Z.-H. & Lam, E. Plant J. 7, 359–365 (1995).
2. Bechtold, N., Ellis, J. & Pelletier, G. C. R. C. R. Acad. Sci. Paris,
Life Sci. 316, 1194–1199 (1993).
3. Ma, H., Yanofsky, M. F. & Meyerowitz, E. M. Genes Dev. 5,
484–495 (1991).
4. Savidge, B., Rounsley, S. D. & Yanofsky, M. F. Plant Cell 7,
721–733 (1995).
5. Hasty, P., Ramírez-Solis, R., Krumlauf, R. & Bradley, A. Nature
350, 243–246 (1991).
Speciation in the
open ocean
Species inhabiting the open oceans gen-
erally occupy broad distribution ranges,
extending from parts of ocean basins to
entire oceans1. Such oceanic species are
accepted as monospecific because of their
tendency to be dispersed across great dis-
tances by ocean currents. Indeed, there are
few known examples of cryptic diversity
among species in the oceanic environment2.
We have now found large, localized genetic
differences within a circumglobal, mono-
typic species of deep-sea fish. This shows
that cryptic allopatric lineages can and have
split without discernible barriers, suggest-
ing that there might have been a serious
underestimation of oceanic biological
diversity.
Nature © Macmillan Publishers Ltd 1997
genus Cyclothone (family Gonostomatidae)
that comprises 13 oceanic micronektonic
species3 (20–70 mm body length). They
are so ubiquitous and numerically domi-
nant in meso- and bathypelagic depths
(200–2,000 m) in the world’s oceans that
some authors have claimed them to be the
most abundant vertebrates on Earth4. Taxo-
nomic confidence in Cyclothone has been
achieved through several critical revisions
using worldwide collections3, and mito-
chondrial DNA sequence data have sup-
ported the recognition of the 13 species5.
We sequenced a portion of the 5፱ half of
the mitochondrial 16S ribosomal RNA and
adjacent transfer RNAval genes (366 base
pairs) from 41 specimens of Cyclothone
alba, which occurs circumglobally along
tropical–subtropical latitudes at depths of
300–600 m (Fig. 1), in addition to seven
specimens of C. signata, four of C. braueri
and one of C. acclinidens, and analysed the
unambiguously aligned sequences (350 base
pairs) using neighbour-joining and maxi-
mum parsimony methods (Fig. 1).
There is complete congruence between
genealogy and geography in this phylogeny.
Each individual belonged to one of five
monophyletic groups (bootstrap probabili-
ties 91–100%) that corresponded to the
region of its origin (Fig. 1). The extant sister
populations seem to have undergone rela-
tively long, independent lineage-sorting
processes and attained reciprocal mono-
phyletic status6 with low levels of gene flow7,
but no discernible morphological diver-
gence. Indeed, the maximum intraspecific
uncorrected pairwise sequence difference in
C. alba (11.1%) was very similar to the
minimum interspecific differences between
C. alba and C. signata (11.9%), the average
intra- and interpopulational sequence dif-
ferences of C. alba ranging from 0.1 to 0.7%
and from 3.0 to 10.3%, respectively.
The central and western North Pacific
populations are more closely related to the
western North Atlantic (bootstrap probabil-
ities, 96%/99%) and eastern equatorial
Indian Ocean populations (bootstrap prob-
ability, 93%/92%), respectively, than to the
other Pacific populations. The existence of
two such inter-oceanic ancestral popula-
tions can be explained by the fact that tropi-
cal Pacific and Atlantic waters had been
connected through the Panamanian seaway
until the final closure of the isthmus about
3.2–2.8 million years ago8 and that Pacific
and Indian deep waters have been narrowly
connected through the Strait of Timor in the
equatorial area. Subsequent lineage split-
tings may have resulted from the closure of
the isthmus and the Indonesian Archipelago
acting as effective geographical barriers.
There seem to have been earlier lineage
splittings of the western North Atlantic and
central North Pacific populations, the west-
803
scientific correspondence

Extraterrestrial
handedness
M. H. Engel and S. A. Macko1 tentatively
ascribe the enantiomeric excess of L-amino
acids in extraterrestrial sources to the
circular polarization of synchrotron radi-
ation, from a neutron star, incident on the
interstellar molecular cloud from which
the Solar System formed (see also the
accompanying News and Views article by
C. F. Chyba2). But the Kuhn–Condon zero-
sum rule3,4 for the rotational strengths
of a chiral molecule requires that broad-
band circularly polarized radiation cannot
discriminate between the enantiomers of
a racemic substance in photochemical
reactions.
Kuhn first demonstrated photochem-
ical optical resolution by showing that
monochromatic circularly polarized light
of a given handedness, tuned to the fre-
quency of a specific circular dichroism
absorption of an enantiomer, preferentially
photolysed that enantiomer in a racemic
mixture3. The mirror-image enantiomer
was favoured if the radiation was tuned to
a circular dichroism absorption of opposite
sign, under the same conditions. Kuhn
Figure 1 Neighbour-joining tree of the four species of Cyclothone. Tree topology is congruent with one of the used coupled-oscillator theory to account
two maximum parsimony trees (186 steps; consistency indexǃ0.742; retention indexǃ0.924). Pairwise dis- for the observation that the circular
tances were calculated using the Kimura two-parameter method, and neighbour joining analyses used dichroism bands of an enantiomer alter-
MEGA10. Maximum parsimony analyses used the heuristic algorithm in PAUP11, with no transition or transver- nate in sign along the wavelength ordinate,
sion weightings. Phylogenetically uninformative sites were excluded. Haplotypes are designated by the first and so cancel out over its electronic spec-
three letters of specific names and two digits. Number of individuals (if any) with identical sequences follows trum. Therefore the optical rotatory power
in parentheses. Numbers beside internal branches indicate bootstrap probabilities12 (>50% only) for 1,000 of a given enantiomer, and its susceptibility
replicates. Wedge-shapes denote possible population subdivision events. Shaded portions of maps show to differential photochemical change with
distribution ranges5. Specimens were taken from two stations in the Sargasso Sea (western North Atlantic, circular radiation, sum to zero over the
WNA), two in Hawaiian waters (central North Pacific, CNP), two in the equatorial eastern Indian Ocean (EEI), electromagnetic spectrum as a whole3.
one in the Coral Sea (western South Pacific, WSP) and three off southern Japan (western North Pacific, WNP). Kuhn’s classical result was confirmed by
Condon4, who demonstrated quantum-
ern South Pacific population, and the west- Masaki Miya mechanically that an enantiomer’s rota-
ern North Pacific and eastern equatorial Natural History Museum and Institute, Chiba, tional strengths (measured by the circular
Indian Ocean populations, which required Aoba-cho, Chuo-ku, Chiba 260, Japan dichroism band areas) sum to zero over the
three within-ocean fragmentations of the e-mail: mxe01346@niftyserve.or.jp spectrum.
ancestral population in the Pacific (Fig. 1). Mutsumi Nishida The zero-sum rule does not exclude a
It is evident that these three lineages have Department of Marine Bioscience, photochemical origin for biomolecular
not been ephemeral, as there has been a lack Fukui Prefectural University, homochirality under severely restrictive
of within-ocean coalescence between popu- Obama, Fukui 917, Japan initial conditions, covering time, place,
lations during the past few million years, 1. van der Spoel, S. & Heyman, R. P. A Comparative Atlas of radiation filters, and so on, for example
which probably extends back beyond the Zooplankton: Biological Patterns in the Oceans (Springer, Berlin, by a prebiotic pool of racemic amino
1983).
closure of the Panamanian seaway. 2. Knowlton, N. Annu. Rev. Ecol. Syst. 24, 189–216 (1993).
acids on an east-facing slope exposed to
Our results show that there are geneti- 3. Miya, M. Copeia 1994, 191–204 (1994). solar radiation only at dawn on the early
cally structured, isolated populations in 4. Nelson, J. S. Fishes of the World 3rd edn (Wiley, New York, 1994). Earth5. But such suggestions are essentially
5. Miya, M. & Nishida, M. Ichthyol. Res. 43, 375–398 (1996).
open oceans. Given the long periods where ad hoc.
6. Avise, J. C. Molecular Markers, Natural History and Evolution
there were no physical barriers, such as con- (Chapman & Hall, New York, 1994). Stephen F. Mason
tinents, it seems surprising that the three 7. Slatkin, M. Genetics 121, 609–612 (1989). 12 Hills Avenue,
populations in the Pacific have not coa- 8. Coates, A. G. et al. Geol. Soc. Am. Bull. 77, 804–828 (1992). Cambridge CB1 4XA, UK
9. Gibbs, R. H. Jr in Pelagic Biogeography (eds Pierrot-Bults, A. A.
lesced. Numerous questions remain unan- et al.) 98–103 (Unesco, Paris, 1986).
e-mail: jm148@cam.ac.uk
swered, such as what factors might structure 10. Kumar, S., Tamura, K. & Nei, M. MEGA: Molecular 1. Engel, M. H. & Macko, S. A. Nature 389, 265–268
genetic diversity in such a homogeneous Evolutionary Genetic Analysis Ver. 1.0 (Pennsylvania State Univ., (1997).
environment. As stated by Gibbs9, we are a University Park, 1993). 2. Chyba, C. F. Nature 389, 234–235 (1997).
11. Swofford, D. L. PAUP: Phylogenetic Analysis Using Parsimony 3. Kuhn, W. Trans. Faraday Soc. 26, 293–310 (1930).
long way from knowing what species really Ver. 3.1.1 (Illinois Natural History Survey, Champaign, 1993). 4. Condon, E. U. Rev. Mod. Phys. 9, 432–457 (1937).
exist in the oceanic pelagic realm. 12. Felsenstein, J. Evolution 39, 783–791 (1985). 5. Mason, S. F. Chem. Soc. Rev. 17, 347–359 (1988).

804 NATURE | VOL 389 | 23 OCTOBER 1997

Nature © Macmillan Publishers Ltd 1997