The fitness effects of transgenic disease resistance in wild sunflowers-Scienc

1. BREVIA
These results also illustrate the importance of
Fitness Effects of Transgenic quantifying fitness directly, rather than using
a presumptive correlate such as disease inci-
Disease Resistance in Sunflowers dence; had we relied solely on infection rates,
our conclusions would have been quite dif-
ferent. Although the mechanism(s) responsi-
John M. Burke1* and Loren H. Rieseberg2 ble for the decoupling of infection rate and
disease severity are unclear, one possibility is
that conditions favoring white mold infection
Fears about transgene escape have focused then monitored them for symptoms of infection. differ from those favoring development of the
attention on the potential for hybridization Presence or absence of the OxOx transgene disease once it has been acquired. The hotter,
between crops and their wild relatives. Al- had no effect on seed output (P 0.25), indi- drier climate of California might, for exam-
though transgenes will often escape from cul- cating that there was no cost of resistance in the ple, make plants more susceptible to infection
tivation (1), their rate of spread will be main- absence of a pathogen challenge (Fig. 1A). In but hinder development of the disease. Re-
ly governed by their fitness effects, not the terms of infection rates, the OxOx transgene did gardless, it appears that, by giving the OxOx
migration rate (2). Thus, only highly advan- provide protection against white mold (P transgene to wild sunflower, we gave it some-
tageous transgenes will spread rapidly 0.002) (Fig. 1B). The transgene did not, how- thing that it already had: some degree of
enough to have a substantial ecological im- ever, have any effect on seed output after inoc- white mold resistance.
pact. Therefore, research on the risks associ- ulation (P 0.84) (Fig. 1C). Though the trans- One caveat is that this work was per-
ated with transgene escape should focus on gene provided protection against white mold formed within a single season and on a single
the fitness effects of the gene(s) in question. infection, it had no effect on reproductive out- genetic background. Therefore, our results
Here, we examined the fitness effects of a put. This result has a simple explanation: Vari- may not be generalizable. Another caveat is
transgene conferring resistance to white mold ation in the likelihood of infection was offset by that plants in this experiment were not sub-
(Sclerotinia sclerotiorum) in sunflower (He- variation in the severity of infection (compare jected to environmental stresses such as
lianthus annuus). Unfortunately, attempts to Fig. 1, B and D). In California, where the drought; plants grown under stress may be
breed white mold resistance have met with transgene provided the most protection against less able to protect themselves from disease,
little success, and chemical control methods infection, disease onset had no effect on seed and thus derive a greater benefit from trans-
are costly and often ineffective. Therefore, output. In contrast, white mold infection caused genic disease resistance. Of course, stressful
attention has turned to genetic modification. a severe decline in seed output in Indiana, but conditions could also reveal a cost of resis-
An oxalate oxidase (OxOx) transgene has infection rates were unaffected by the transgene. tance not otherwise observed.
now been used to enhance white mold resis- Thus, the transgene had a significant effect on the Future studies assessing the environmen-
tance in cultivated sunflower [supporting on- likelihood of infection, and infection had a nega- tal impact of transgenes should not only be
line material (SOM) text], presumably by tive effect on seed output (P 0.0001) (Fig. 1D), replicated over space and time, but should
degrading oxalic acid, which contributes to but the disease effect varied across locations (P also examine the effects of genetic back-
white mold pathogenicity (3). 0.001), nullifying any advantage of the transgene. ground and environmental stresses. Regard-
Most of the sunflower acreage in the United Our results suggest that the OxOx trans- less of the form of future research, an in-
States occurs within the range of wild sunflow- gene will do little more than diffuse neutrally formed judgment of the risks and benefits of
er, and many fields flower coincidentally with after its escape. This is especially true be- genetic modification on a case-by-case basis
neighboring wild populations (4). Where they cause our experiment simulated the worst- is preferable to either the dismissal of trans-
come into contact, cultivated and wild sun- case scenario, in which early generation hy- genic approaches entirely, or the introduction
flower often hybridize (5), making transgene brids faced a severe pathogen challenge. of transgenic crops in the absence of appro-
escape a virtual cer- priate scientific scrutiny.
tainty. Combined
References and Notes
with the efficacy of A 60 Wild type C 45 Wild type 1. N. C. Ellstrand, H. C. Prentice, J. F. Hancock, Annu.
OxOx 40 OxOx
the OxOx transgene, 50 Rev. Ecol. Syst. 30, 539 (1999).
SQRT (seeds)
SQRT (seeds)
35
this raises the possi- 40 30 2. L. H. Rieseberg, J. M. Burke, Taxon 50, 47 (2001).
25 3. R. D. Noyes, J. G. Hancock, Physiol. Plant Pathol. 18,
bility of hybridization 30
20 123 (1981).
giving rise to a more 20 15 4. J. M. Burke, K. A. Gardner, L. H. Rieseberg, Am. J. Bot.
10
invasive wild sun- 10 89, 1550 (2002).
5
5. C. R. Linder, I. Taha, G. J. Seiler, A. A. Snow, L. H.
flower. 0 0
Rieseberg, Theor. Appl. Genet. 96, 339 (1998).
California North Dakota Indiana California North Dakota Indiana
To simulate the 6. Materials and methods are available as supporting
early stages of es- B material on Science Online.
D 50 7. We thank C. Alarcon, G. Cole, J. Durphy, R. Essner, E.
Frequency of infection
cape, we back- 0.4 Wild type Healthy
OxOx 45 Infected Haro, E. Hoeft, J. Holte, C. Scelonge, and L. Wang. This
crossed the OxOx
SQRT (seeds)
0.3
40 work was supported by grants from Pioneer Hi-Bred,
35
transgene into wild 30 Intl. (to L.H.R.) and the United States Department of
sunflower and grew 0.2 25 Agriculture (to J.M.B.).
20
the resulting plants 15 Supporting Online Material
0.1
10 www.sciencemag.org/cgi/content/full/300/5623/1250/DC1
in containment cages 5 SOM Text
at field sites in Indi- 0 0
Materials and Methods
California North Dakota Indiana California North Dakota Indiana
ana, North Dakota,
and California (6). Fig. 1. Effects of the OxOx transgene and white mold infection on crop 1
Department of Biological Sciences, Vanderbilt University,
wild sunflower hybrids. (A) Seed output of control individuals. (B) Fre- Nashville, TN 37235–1634, USA. 2Department of Biology,
Just before flower- quency of infection after inoculation. (C) Seed output after inoculation.
Indiana University, Bloomington, IN 47405, USA.
ing, we inoculated (D) Effect of infection on seed output after controlling for presence or
half of all the plants absence of the transgene. All values are expressed as least-squares *To whom correspondence should be addressed. E-mail:
with white mold and means one standard error. john.m.burke@vanderbilt.edu
1250 23 MAY 2003 VOL 300 SCIENCE www.sciencemag.org

2. PERSPECTIVES
pathogen or by mass fruiting episodes). will remain a powerful tool for understanding References
In summary, the chances of capturing car- carbon fluxes in forests and for testing mod- 1. R. K. Dixon et al., Science 263, 185 (1994).
2. Y. Malhi et al., Plant Cell Environ. 22, 715 (1999).
bon release with similar accuracy as carbon els (4, 6, 10, 12). But it is important to re- 3. E. D. Schulze et al., Science 289, 2058 (2000).
uptake are minute, preventing meaningful member that they cannot produce a realistic 4. J. L. Ehman et al., Global Change Biol. 8, 575 (2002).
budgeting, irrespective of the precision of the picture of a landscape’s contribution to car- 5. I. C. Prentice et al., Ecol. Appl. 10, 1553 (2000).
equipment used. Reliable carbon budgeting bon sequestration. The same holds for growth 6. S. Frolking et al., Global Change Biol. 2, 343 (1996).
7. M. L. Goulden et al., Science 271, 1576 (1996).
beyond the plot level will thus mostly remain responses of forests to elevated CO2. When 8. E. H. DeLucia et al., Science 284, 1177 (1999).
the domain of forest inventories and remote forest dynamics are accounted for, a stimula- 9. R. Valentini et al., Nature 404, 861 (2000).
sensing. At larger (hemispheric or global) tion of growth rate and acceleration of devel- 10. J. G. Canadell et al., Ecosystems 3, 115 (2000).
11. Ch. Körner, Ecol. Appl. 10, 1590 (2000).
scales, we depend on atmospheric signals. opment may reduce the residence time of car- 12. D. Schimel et al., Nature 414, 169 (2001).
Plot-based carbon flux measurements in bon per unit land area, and thus may even di- 13. J. Granados, Ch. Körner, Global Change Biol. 8, 1109
connection with tree and soil studies are and minish long-term carbon storage (11, 13). (2002).
PLANT SCIENCES
to release genetically modified plants for
commercial purposes. This document in-
Super Sunflowers— cludes a discussion of the environmental
risk-assessment process, including consid-
Stopping the Rot? eration of the potential impact of gene flow
to unmodified plants and to wild relatives.
This document does not make direct refer-
John Heritage ence to an estimation of the “fitness” of the
inserted transgene in a genetically modi-
he creation of a “superweed” is an en- at least in Europe. This perhaps explains fied plant. It recognizes, however, that
T
during image used by environmental the high level of skepticism or hostility to- genes will escape from genetically modi-
lobbyists opposed to the introduction ward the introduction of genetically modi- fied plants and that this is a significant fac-
of genetically modified plants. On page fied plants in agriculture in Europe, and tor in a proper risk assessment. This recog-
1250 of this issue, Burke and Rieseberg (1) probably also contributes to Europe having nition is central to the argument made by
highlight an impor- the most thorough regulation of genetical- Burke and Rieseberg (1).
Enhanced online at tant point concerning ly modified crops. Farmers grow sunflowers commercially
www.sciencemag.org/cgi/ the potential escape The European Union has recently re- for their oil, and sunflower seed cake, a by-
content/full/300/5623/1243 of genes inserted into viewed legislation regarding the deliber- product of oil production, is an important
genetically modified ate release of genetically modified organ- feed commodity for ruminants. Sunflower
crops. They examine the transgene that con- isms into the environment. This resulted crops are particularly important in temper-
fers resistance to the white rot fungus in the publication of European Directive, ate climates, where the growing conditions
(Sclerotinia sclerotiorum) in sunflowers 2001/18/EC on the deliberate release into are too dry for other oilseed crops, although
(Helianthus annuus L). These authors the environment of genetically modified in Europe sunflowers are less important as a
demonstrate that it is not the transfer of organisms (2). This legislation recognizes crop than are soyabeans and canola (rape-
genes per se that is important, but rather their that “…the effects of such releases [of live seed) (4). Sunflower plants are, however,
fitness to persist in a new host plant (1). genetically modified organisms] on the en- prone to attack by the white rot fungus S.
Among the first generation of trans- vironment may be irreversible.” The sclerotiorum (see the figure). This fungus is
genic plants introduced into agriculture European Commission has recently pub- a highly successful plant pathogen that at-
were crops rendered resistant to herbicides lished a guidance document (3) to aid risk tacks a range of economically important
or made tolerant to insect attack. These assessors intending to apply for permission plants, including those involved in oilseed
traits presented clear benefits to farmers production. It has a broad eco-
who wished to minimize crop loss and to logical and geographical distri-
simplify agricultural practices. It is easy to bution, often being found in
envisage the genes that encode these traits temperate climes. Fungal mycelia
escaping from the crop plant and entering may persist in soil to cause in-
conventionally bred plants or related fection, but aerial spores are also
species in the wild. In Europe, a series of responsible for dissemination of
major food and feed scares have sensitized this plant pathogen. Head rot
the public to “green” issues, perhaps to a disease of sunflowers results
much greater extent than elsewhere. The from aerial spread of the fungal
largest such problem has been the emer- spores. Disease in sunflowers
gence of bovine spongiform encephalopa- may also manifest as wilt and
thy and its link to variant Creutzfeldt-Jakob middle stem rot. S. sclerotiorum
disease in humans. The recent epidemic of is one of the most economically
CREDIT: D. LEBOURGEOIS/CETIOM
foot-and-mouth disease among farm ani- significant pathogens of sun-
mals in the United Kingdom has refocused Sunflowers afflicted with white rot. Stem rot in a sun- flowers. Despite its widespread
public attention on agricultural practices, flower plant caused by the white rot fungus S. sclerotiorum. geographical distribution and its
The sunflower plant has an extensive infection that has broad host range, its hosts ap-
The author is in the Division of Microbiology, School caused it to wilt as evidenced by the leaves, which are droop- pear not to be naturally resistant
of Biochemistry and Molecular Biology, University of ing down the stem. The fungal mycelia and fruiting bodies to attack by S. sclerotiorum.
Leeds, LS2 9JT, UK. E-mail: j.heritage@leeds.ac.uk can be seen erupting from the stem of the plant. Oxalic acid is an important viru-
www.sciencemag.org SCIENCE VOL 300 23 MAY 2003 1243

3. PERSPECTIVES
lence determinant for this fungus (5). their study, the transgene provides protec- ment on the risks, and benefits, of geneti-
Certain cereal plants are resistant to attack tion against white rot infection. Despite cally modified plants.
by S. sclerotiorum through the production of this finding, reproductive output did not
oxalate oxidase (OxOx), an enzyme that improve in hybrid plants containing the References and Notes
1. J. M. Burke, L. H. Rieseberg, Science 300, 1250 (2003).
breaks down oxalate, producing carbon OxOx transgene compared with wild-type 2. European Commission, Directive 2001/18/EC of the
dioxide and hydrogen peroxide. The intro- plants. Antagonists of the use of genetical- European Parliament and of the Council of 12 March
duction into sunflowers of a gene that codes ly modified plants in agriculture may argue 2001 on the deliberate release into the environment of
genetically modified organisms and repealing Council
for oxalate oxidase results in plants that re- that this is a single case and that results de- Directive 90/220/EEC–Commission Declaration, Official
sist white rot attack (6). rive from a single growing season. It would Journal of the European Communities L106, 1 (2001).
Burke and Rieseberg measure the im- be unwise to extrapolate from this particu- 3. European Commission Scientific Steering Committee,
Guidance Document for the Risk Assessment of
pact on the relative fitness of sunflowers of lar set of observations to other constructs in Genetically Modified Plants and Derived Food and
the transgene conferring resistance to the different conditions, but the authors them- Feed; http://europa.eu.int/comm/food/fs/sc/ssc/
white rot fungus. Even in the presence of a selves highlight this limitation. The out327_en.pdf (23 April 2003).
4. European Commission Directorate-General for
direct challenge from the pathogen, the strength of this paper does not lie in its Agriculture, Prospects for Agricultural Markets 2000-
transgene has no direct effect on sunflower demonstration that not all genetically mod- 2007; http://europa.eu.int/comm/agriculture/publi/
fitness. With this demonstration, the au- ified plants will form “superweeds.” caprep/prospects2000/fullrep.pdf (23 April 2003).
5. S. G. Cessna, V. E. Sears, M. B. Dickman, P. S. Low, Plant
thors show that the effect of transgene Rather, it shows that, when considering the Cell 12, 2191 (2000).
spread is more important than considering persistence of transgenic material in the en- 6. R. D. Noyes, J. G. Hancock, Physiol. Plant Pathol. 18,
the likelihood that the spread may happen. vironment, it is important to ask the right 123 (1981).
7. I thank CETIOM (Paris) for permission to use the pho-
They conclude that, for this particular con- questions. Only when we ask the correct tograph and A. Hardy (DEFRA Central Science
struct and for the three locations chosen for questions will we reach an informed judg- Laboratory, York) for help in obtaining the picture.
M AT E R I A L S S C I E N C E
tain grain size. This has also been found by
several atomistic simulations (7, 8).
Nanocrystals Get Twins Individual dislocation may interact with
other dislocations and other lattice defects,
Jørgen B. Bilde-Sørensen and Jakob Schiøtz such as grain boundaries. The generation,
propagation, and annihilation of disloca-
n the messy world of materials science, nor expected in coarse-grained aluminum tions during deformation are therefore
I experimental observations often precede owing to the high energy of the planar faults. rather complex processes. Atomistic simu-
theoretical attempts to find a suitable ex- Nanocrystalline materials have attract- lation is an obvious tool to apply to these
planation. But on page 1275 of this issue, ed much interest in recent years because of processes. However, even with modern
Chen et al. (1) report the reverse scenario: their extraordinary properties. The strength computers this is a challenging task, be-
an experiment that corroborates recent mo- and hardness of materials increase with de- cause a very large number of atoms has to
lecular dynamics simulations (2–4). creasing grain size until, below a certain be handled during the computation, and the
The authors study the plastic deformation grain size, the material starts to become geometry of the simulated volume has to
of nanocrystalline aluminum. Textbooks tell softer again (5, 6). As the grain size de- be chosen carefully.
us that deformation twinning has never been creases to very small values, the fraction of The first indication that deformation
observed in coarse-grained aluminum, but in the atoms that resides in or near a grain twinning could occur in aluminum under
the simulations, deformation twinning oc- boundary increases drastically, and it is certain circumstances was reported by
curred in aluminum with a sufficiently small therefore to be expected that grain-bound- Tadmor et al. (2). The authors had used the
grain size. And that is exactly what Chen et ary effects will start to dominate at a cer- so-called quasi-continuum method to sim-
al. now show experimentally. ulate what happens when a
Plastic deformation of crystalline materi- nanometer-scale, rigid object
als has long been known to occur by the is pressed into the surface of
movement of line defects called dislocations. a thin aluminum film.
A perfect dislocation moves one part of the 5 However, the geometry was
crystal by a complete lattice vector with re- unusual, creating essentially
spect to the other, such that it leaves a perfect 4 a two-dimensional state of
lattice. A partial dislocation, however, moves 3 deformation. Furthermore,
one part by less than a complete lattice vec- 2 the deformation twins creat-
tor, leaving a faulted layer in the crystal. If ed during the indentation
1
this occurs on consecutive planes in a face- disappeared again on un-
centered cubic (fcc) crystal, it results in the loading.
formation of a twin, which has a lattice that Further support was pro-
is a mirror of the host lattice. Deformation vided by Yamakov et al. (3,
twinning is well known from some fcc sys- 4), who observed several de-
tems, such as brass, but is neither observed 6 formation twins in molecular
dynamics simulations of
CREDIT: ADAPTED FROM (3)
J. B. Bilde-Sørensen is in the Materials Research nanocrystalline aluminum
Department, Risø National Laboratory, DK-4000 Ros- under tensile deformation
kilde, Denmark. E-mail: j.bilde@risoe.dk J. Schiøtz is
in the Department of Physics, Technical University Emergence of twins in a simulation of nanocrystalline alu- (see the figure). These simu-
of Denmark, DK-2800 Lyngby, Denmark. E-mail: minum. (1) to (5) mark one emerging twin; (6) marks another lations still have some obvi-
schiotz@fysik.dtu.dk twin formed by a slightly different mechanism. ous limitations. For example,
1244 23 MAY 2003 VOL 300 SCIENCE www.sciencemag.org