Volatile Signaling in Plant-Plant Interactions:

"Talking Trees" in the Genomics Era

Ian T. Baldwin, et al.
Science 311, 812 (2006);
DOI: 10.1126/science.1118446

The following resources related to this article are available online at

www.sciencemag.org (this information is current as of September 7, 2007 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/cgi/content/full/311/5762/812
A list of selected additional articles on the Science Web sites related to this article can be

Downloaded from www.sciencemag.org on September 7, 2007

found at:
http://www.sciencemag.org/cgi/content/full/311/5762/812#related-content
This article cites 36 articles, 8 of which can be accessed for free:
http://www.sciencemag.org/cgi/content/full/311/5762/812#otherarticles

This article has been cited by 17 article(s) on the ISI Web of Science.

This article has been cited by 4 articles hosted by HighWire Press; see:
http://www.sciencemag.org/cgi/content/full/311/5762/812#otherarticles

This article appears in the following subject collections:

Botany
http://www.sciencemag.org/cgi/collection/botany

Information about obtaining reprints of this article or about obtaining permission to reproduce
this article in whole or in part can be found at:
http://www.sciencemag.org/about/permissions.dtl

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2006 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
 PLANT VOLATILES

Hence, plants that avoided investing fitness-

REVIEW limiting resources in the production of costly
defenses before an herbivore arrives, but were

Volatile Signaling in Plant-Plant able to prime defense metabolism to launch

defense responses when attacked, could realize
a fitness benefit over plants that ‘‘ignored’’ the
Interactions: ‘‘Talking Trees’’ information coded in the VOCs emanating
from their damaged neighbors.
in the Genomics Era The use of microarrays that monitor a
large fraction of the plant’s transcriptome can
free analysis from observer bias about plants’
Ian T. Baldwin,1* Rayko Halitschke,2 Anja Paschold,1 Caroline C. von Dahl,1
conversation topics and identify selective pres-
Catherine A. Preston3
sures other than impending attack from mobile
herbivores, which volatile signaling could be
Plants may ‘‘eavesdrop’’ on volatile organic compounds (VOCs) released by herbivore-attacked
used to anticipate. Herbivores frequently trans-

Downloaded from www.sciencemag.org on September 7, 2007

neighbors to activate defenses before being attacked themselves. Transcriptome and signal cascade
mit pathogens, and the elicited responses may
analyses of VOC-exposed plants suggest that plants eavesdrop to prime direct and indirect defenses
concern attack by impending pathogens more
and to hone competitive abilities. Advances in research on VOC biosynthesis and perception have
than attack by herbivores (18). The relentless
facilitated the production of plants that are genetically ‘‘deaf’’ to particular VOCs or ‘‘mute’’ in
competition with other plants for resources that
elements of their volatile vocabulary. Such plants, together with advances in VOC analytical
cannot be readily hoarded (such as light and
instrumentation, will allow researchers to determine whether fluency enhances the fitness of plants
nutrients) is likely the most important selective
in natural communities.
force for plants. Plants are able to anticipate im-
pending competition through far red (FR) light
lants excel at gas exchange: They can terpreted accordingly. Measures of herbivore signals and changes in the photon flux of blue

P literally build forests from CO2 taken

from the air at about 120 Pg C yearj1,
half of which is respired back to the atmo-
performance have been broadened to include
the elicitation of various direct plant defenses
(e.g., phenolics, alkaloids, terpenes, and de-
light transmitted through their neighbors’ can-
opies. These light signals are perceived by pho-
toreceptors (e.g., phytochrome B) and elicit a
sphere. Up to 36% of the assimilated carbon fense proteins). Indirect defenses have also at- complex of traits known as the shade-avoidance
is released as complex bouquets of VOCs (1). tracted attention, including food rewards that syndrome (SAS) (19). Experiments with tobac-
Although some of these VOCs may be mere increase predation pressure on herbivores (4) co plants transformed with a mutant ethylene
waste, others mediate various pollination and and VOCs that help predators or parasitoids receptor (etr1-1), which inhibits ethylene per-
defense mutualisms with animals. These VOC- locate feeding herbivores (5, 6). Moreover, the ception, have demonstrated that ethylene-
mediated interactions of plants with orga- signal cascades that elicit direct and indirect insensitive tobacco could not respond rapidly
nisms of higher trophic levels suggest that they defenses have been scrutinized (7, 8) as have to FR signals and consequently was outcom-
communicate similarly with each other (2). transcriptional responses (9–12) (Fig. 1). peted by wild-type plants (20). At concentra-
Two decades ago, researchers serendipitously VOC exposure alone, without actual herbi- tions apparently possible in dense plant canopies,
discovered changes in herbivore resistance and vore attack, may directly increase the produc- ethylene by itself elicits the SAS (21). Similarly,
secondary metabolites in plants (Breceivers[) tion of defenses. Alternatively, VOC exposure exposure to unidentified VOCs from barley cul-
growing adjacently to herbivore-attacked plants may allow nearby plants to ready their tivars changes the allocation of biomass between
(Bemitters[). Because in some experiments defenses for immediate use once the herbi- roots and shoots without influencing biomass
results were best explained by the aerial trans- vores move from the neighboring plant to production of receiver barley genotypes (22), a
fer of information (3), the phenomenon was attack the ‘‘listening’’ receiver. Exposure to re-allocation that may influence competitive
popularly dubbed Btalking trees.[ This phrase volatiles from damaged sagebrush primes the ability. Thus, responses to the most important
seems unfortunate, because selection most elicitation of defensive proteinase inhibitors environmental factors in a plant’s life may be
likely favors plants that Beavesdrop[ on VOCs (PIs) in wild tobacco, and exposed plants anticipated by signals from neighboring plants.
released from neighbors and respond by tai- subsequently receive less damage (13–15) Almost anything can be a signal as long as it can
loring their phenotypes to enhance their own (Fig. 2). Corn seedlings previously exposed to be perceived and provides reliable information.
fitness. either individual components or to the entire
blend of VOCs released from herbivore- What Does It Take to Be a Signal?
What Are Plants Talking About? attacked seedlings responded to simulated Four steps characterize the transfer of VOC
An obvious conversation topic concerns im- herbivory with increased VOC production signals between plants: the release of the signal
pending attack from mobile herbivores, and and higher jasmonate (JA) accumulations by the emitter plant and its transport, absorp-
most VOC-elicited responses have been in- compared with the responses of unexposed tion, and perception by the receiver plant (Fig.
plants (8). Whether these enhanced VOC 1). All are influenced by the signal’s properties
1
Department of Molecular Ecology, Max Planck Institute for emissions protect corn seedlings remains to and its biological context. Most research on
Chemical Ecology, Hans Knöll Strasse 8, Jena 07745, be determined. The priming of defense cas- signal release has focused on the activation of
Germany. 2Ecology and Evolutionary Biology, Cornell cades may benefit plants that would incur biosynthetic enzymes and their substrate sup-
University, E443 Corson Hall, Ithaca, NY 14853, USA. fitness costs by activating defense responses ply. The biochemical control mechanisms for
3
Biotechnology Regulatory Services, Animal and Plant
Health Inspection Service, U.S. Department of Agriculture, (16), particularly in the absence of herbivore the major VOC constituents are rapidly being
4700 River Road, Riverdale, MD 20737, USA. attack (17). If VOC exposure directly elicited clarified (Fig. 1). However, the release of foliar
*To whom correspondence should be addressed. E-mail: defense responses, receiver plants would incur VOCs is also controlled by their physico-
[email protected] similar fitness costs without being damaged. chemical properties (23): Volatility is deter-

812 10 FEBRUARY 2006 VOL 311 SCIENCE www.sciencemag.org

 SPECIALSECTION
mined by partitioning the compound between ceiver plants, define the last steps in the signal chamber increases VOC concentrations and also
the ‘‘liquid’’ phase of the leaf and the atmo- transfer process: adsorption at the plant surface reduces CO2 once the chambers are illuminated
sphere, whereas molecular size and stomatal and uptake into the leaf via stomatal openings because of photosynthetic carbon fixation. Under
aperture constrain diffusive transport from the or cuticle diffusion. The low concentration such conditions, plants increase the number of
leaf into the air surrounding the leaf, its gradient between atmosphere and leaf during open stomata, enhancing exposure of mesophyll
headspace. Once released into the headspace the adsorption step amplifies the effects of the cells to the VOCs. Therefore, sealed chambers
of the emitter, the potential signal has to be signal’s physicochemical properties. Transport are likely to influence the responsiveness of re-
transported to receivers. Direction and dy- into the receiver leaf is influenced by stomatal ceiver plants, and studies that use them are more
namics of this transport are dictated by tem- conductance. The limited air volume of a sealed likely to report ecologically insignificant results.
perature, convective transport,
and wind for above-ground
signaling or water for below-
ground signaling. Small high-
ly volatile compounds (e.g.,
ethylene, methanol, isoprene,

Downloaded from www.sciencemag.org on September 7, 2007

acrolein, methacrolein, and
some monoterpenes) diffuse
rapidly into the headspace
and are diluted in the atmo-
sphere (Fig. 2). For such
compounds, signaling func-
tion is likely limited to the
foliage of the emitter (as a
systemic within-plant signal)
and of neighbors with in-
tertwined canopies. Heavier
compounds with less vola-
tility, such as terpene alcohols,
methyl jasmonate (MeJA), ar-
omatic compounds including
methyl salicylate (MeSA), and
green-leaf volatiles (GLVs),
are more likely to function as
signals over longer distances,
because their comparatively
slower dispersal allows devel-
opment of plumes of higher
concentrations (24) that may
be carried farther as intact par-
cels by turbulent flow (Fig. 2).
During transport, some VOC
species are oxidized or other-
wise processed in the atmo-
sphere (1), possibly causing
dilution but also activation.
The concentration gradients,
which ultimately regulate the
receiver’s exposure, remain
largely uncharacterized. An ex- Fig. 1. Scheme of plant-plant interaction mediated by VOCs emphasizing the use of genetically manipulated plants to
ample of a characterized con- investigate the mechanisms underlying this process. Plants (e.g., wild tobacco) can be exposed to VOCs released from
centration gradient comes from either conspecifics or from emitters of different species (e.g., sagebrush). The VOC bouquet of stressed plants consists
of GLVs, terpenoids, MeJA, MeSA, methanol, ethylene, and other substances (32). Various biotic and abiotic stress
a study of corn seedlings that
factors modulate the chemical vocabulary emitted in quantity, quality, and timing. If the signal is recognized by the
release the volatile sesqui-
receiver plant, it may respond with changes in its signal transduction, transcriptome, proteome, and metabolome,
terpene (E)-b-caryophyllene which may or may not result in functionally significant changes in its fitness (Y). Comparing responses to wild-type
into the soil from their roots, (WT) emitter plants with responses to mute emitters (Y - - ) whose VOC bouquet is deficient in one or more VOCs allows
a below-ground plume used researchers to identify compounds mediating the interaction between emitters and receivers. In addition to insertional
by entomopathogenic nema- mutants [e.g., def-1 (33)], various transgenic lines are generated by the expression of endogenous genes in antisense
todes to locate root-attacking (as) orientations to silence enzymes necessary for eliciting or synthesizing VOCs, such as hydroperoxide lyase [HPL
beetle larvae (25). (34, 35)], lipoxygenase [LOX3 (36)], allene oxide synthase [AOS (35)], 1-aminocyclopropane-1-carboxylic acid synthase
Signal volatility and dif- [ACS (37)], or 1-aminocyclopropane-1-carboxylic acid oxidase [ACO (38)]. These lines represent possible mute emitters.
fusion rates, as well as the Deaf receiver plants, such as the etr1-1 line (39) impaired in functional VOC receptors for individual substances, could
stomatal conductance of re- be used to verify each individual VOC’s bioactivity.

www.sciencemag.org SCIENCE VOL 311 10 FEBRUARY 2006 813

 PLANT VOLATILES

Once a VOC enters the leaf, a response will these highly volatile compounds increased the laboratory under experimental conditions (sealed
only occur if the compound is ‘‘active,’’ a production of endogenous phytohormones, or low air-flow chambers) that maximized the
poorly understood condition. Several pro- their activity was partially independent of the probability of detecting responses in receiver
posed between-plant signals have hormone or JA, SA, and ethylene signal cascades. A redox- plants by increasing exposure [reviewed in
hormone-like functions, including MeSA (26), based signal process, generated by the deple- (10)]. Although this work has shown that plants
GLVs (8, 9, 27), ethylene (28), and MeJA (29). tion of cellular reductants resulting from the respond to being fumigated, its ecological rel-
However, proof that any of these are released electrophile reactivity of these compounds, sug- evance will remain unclear until the responses
and transported to receiver plants in quantities gests a mechanism for their activity that resem- are verified in open-grown plants.
sufficient to elicit responses under natural bles the activation of the regulatory protein for One solution to the problems of ecological
conditions is either lacking or belies a signal pathogen defense, NPR1 (31). Similar processes realism in between-plant signaling studies is to
function (10, 13, 30). Although most studies of may provide the basis of a general chemical use mutants or transgenic plants whose ability
bioactivity have examined whether the presence ‘‘sense,’’ which may have predated the evolu- to either release or perceive particular compo-
of a VOC elicits a response, removing certain tion of receptors for particular volatiles. nents of the wild-type volatile blend is defi-
components from a volatile bouquet can also cient. The use of ‘‘mute’’ emitters (10) allows
elicit a response. The removal of GLVs from Ecological Realism: ‘‘Deaf’’ and ‘‘Mute’’ complex herbivore-induced VOC blends to be

Downloaded from www.sciencemag.org on September 7, 2007

the wound-induced volatile blend by silencing Plants to the Rescue dissected (Fig. 1). Complementation studies, in
hydroperoxide lyase strongly influenced the Constitutive and herbivore-induced VOC emis- which synthetic constituents supplement the vol-
regulation of gene expression in neighboring sions are influenced by a variety of abiotic atile blend to determine whether the receivers’
conspecific tobacco plants (10). In other words, factors [nutrient availability, temperature, wind, response is subsequently restored, confirm func-
plants may respond to the ‘‘sounds of silence.’’ ultraviolet (UV) radiation and photosynthetical- tion. The biosynthetic pathways contributing
A class of electrophilic a,b-unsaturated car- ly active radiation (PAR), and ozone exposure]. constituents to the herbivore-induced volatile
bonyl compounds represents potent regulators To lessen this variability, most studies of plant- bouquet and their regulatory cascades repre-
of gene expression (11). Although exposure to plant signaling have been performed in the sent possible genetic targets. Mutants whose

Fig. 2. Aerial interaction of the wild tobacco (Nicotiana attenuata) and in the cis epimer of MeJA, which is thermodynamically unstable but putatively
sagebrush (Artemisia tridentata tridentata) (40) is the best-documented more biologically active than the trans epimer (14, 30, 42). Hence, MeJA was
example of between-plant signaling via above-ground VOCs in nature the most obvious candidate for the volatile signal mediating the response;
(14, 15, 41). When transplanted to within 15 cm of clipped sagebrush, to- subsequent studies were unable to confirm that either epimer of MeJA
bacco plants suffered less herbivory and produced more seed capsules elicited known herbivore defenses when applied in quantities relevant to
than did plants transplanted adjacent to undamaged sagebrush. Damaged those released by damaged sagebrush (30, 42). Rather than directly elicit-
sagebrush releases a variety of VOCs, which are composed of highly volatile ing defenses, exposure to volatiles from excised sagebrush foliage (and two
substances that disperse by diffusion, namely, methacrolein (A) and less constituents of its aromatic headspace: trans-2-hexenal and methacrolein)
volatile compounds such as GLVs [e.g., cis-3-hexenal (B) and trans-2-hexenal primes defense responses, so that plants increase the production of their
(D)], oxygenated monoterpenes [e.g., cineole (E), thujone, and camphor] and defense protein, PI, faster when attacked (13). The progress in this system
the epimers of MeJA (C), which are likely transported by turbulent flow in highlights the difficulty of predicting how plant-plant signaling functions
fragmented plumes. The plume from damaged sagebrush is highly enriched from first principles.

814 10 FEBRUARY 2006 VOL 311 SCIENCE www.sciencemag.org

 SPECIALSECTION
herbivore- or wound-induced vocabularies deaf plants that are available to complement the 22. V. Ninkovic, J. Exp. Bot. 54, 1931 (2003).
have been modified by silencing genes involved growing list of available mute plants, the more 23. U. Niinemets, F. Loreto, M. Reichstein, Trends Plant Sci. 9,
180 (2004).
in either the biosynthesis of particular volatiles tools researchers will have to fully evaluate the 24. H. W. Thistle et al., Forest Sci. 50, 610 (2004).
or the oxylipin signal cascade represent poten- significance of volatile signaling among plants 25. S. Rasmann et al., Nature 434, 732 (2005).
tial mute emitter plants (Fig. 1). in natural settings. These experiments will de- 26. V. Shulaev, P. Silverman, I. Raskin, Nature 385, 718
Mutants whose perception of specific VOCs termine whether being a native speaker en- (1997).
27. G. Arimura, R. Ozawa, J. Horiuchi, T. Nishioka, J. Takabayashi,
is impaired (‘‘deaf’’ plants) represent another hances a plant’s fitness in its community. Biochem. Syst. Ecol. 29, 1049 (2001).
tool for analyzing the consequences of VOC sig- 28. T. Tscharntke, S. Thiessen, R. Dolch, W. Boland, Biochem.
naling as illustrated by the ethylene-insensitive References and Notes Syst. Ecol. 29, 1025 (2001).
tobacco plants, etr1-1. The produce industry 1. J. Kesselmeier et al., Global Biogeochem. Cycles 16, 29. E. E. Farmer, C. A. Ryan, Proc. Natl. Acad. Sci. U.S.A. 87,
1126 (2002). 7713 (1990).
long ago developed a sophisticated ethylene 2. M. Dicke, A. A. Agrawal, J. Bruin, Trends Plant Sci. 8, 403 30. C. A. Preston, G. Laue, I. T. Baldwin, Biochem. Syst. Ecol.
trapping and releasing technology, but the first (2003). 29, 1007 (2001).
clear demonstration of the functional sig- 3. I. T. Baldwin, A. Kessler, R. Halitschke, Curr. Opin. Plant 31. X. Dong, Curr. Opin. Plant Biol. 7, 547 (2004).
nificance of ethylene signaling in competitive Biol. 5, 351 (2002). 32. P. W. Paré, J. H. Tumlinson, Plant Physiol. 114, 1161
4. C. Kost, M. Heil, Basic Appl. Ecol. 6, 237 (2005). (1997).
interactions required plants that were ‘‘deaf’’

Downloaded from www.sciencemag.org on September 7, 2007

5. J. Ruther, S. Kleier, J. Chem. Ecol. 31, 2217 (2005). 33. J. S. Thaler, M. A. Farag, P. W. Pare, M. Dicke, Ecol. Lett. 5,
to this VOC (20). Receptors for most of the 6. Y. Choh, T. Shimoda, R. Ozawa, M. Dicke, J. Takabayashi, 764 (2002).
herbivore-induced VOCs remain to be dis- J. Chem. Ecol. 30, 1305 (2004). 34. G. Vancanneyt et al., Proc. Natl. Acad. Sci. U.S.A. 98,
covered, but transcriptional responses to VOC 7. G. Arimura et al., Plant J. 29, 87 (2002). 8139 (2001).
8. J. Engelberth, H. T. Alborn, E. A. Schmelz, J. H. Tumlinson, 35. R. Halitschke, J. Ziegler, M. Keinanen, I. T. Baldwin, Plant
exposure can be used in mutant screens to iden- Proc. Natl. Acad. Sci. U.S.A. 101, 1781 (2004). J. 40, 35 (2004).
tify new VOC receptors. Identification of these 9. N. J. Bate, S. J. Rothstein, Plant J. 16, 561 (1998). 36. R. Halitschke, I. T. Baldwin, Plant J. 36, 794 (2003).
genetic elements and the creation of VOC- 10. A. Paschold, R. Halitschke, I. T. Baldwin, Plant J., in press. 37. P. W. Oeller, L. M. Wong, L. P. Taylor, D. A. Pike,
reporter plants [with b-glucuronidase (GUS) or 11. E. Alméras et al., Plant J. 34, 205 (2003). A. Theologis, Science 254, 437 (1991).
12. G. Arimura et al., Nature 406, 512 (2000). 38. A. J. Hamilton, G. W. Lycett, D. Grierson, Nature 346, 284
green fluorescent protein] will allow research-
13. A. Kessler, R. Halitschke, C. Diezel, I. T. Baldwin, (1990).
ers to readily determine the quantity of signals Oecologia, in press. 39. J. Hua, C. Chang, Q. Sun, E. M. Meyerowitz, Science 269,
that are perceived by receivers at different dis- 14. R. Karban, I. T. Baldwin, K. J. Baxter, G. Laue, G. W. Felton, 1712 (1995).
tances from an emitter. Combining deaf and Oecologia 125, 66 (2000). 40. B. Wignall, in Shrubs and Trees of the Southwest Deserts,
mute plants with wild-type plants in natural 15. R. Karban, J. Maron, Ecology 83, 1209 (2002). J. E. Bowers, Ed. (Western National Parks Association,
16. M. Heil, I. T. Baldwin, Trends Plant Sci. 7, 61 (2002). Tucson, AZ, 1993), p. 19.
settings will clarify the relevance of VOC sig- 17. I. T. Baldwin, Proc. Natl. Acad. Sci. U.S.A. 95, 8113 (1998). 41. R. Karban, M. Huntzinger, A. C. McCall, Ecology 85, 1846
naling for a plant’s performance and/or fit- 18. E. S. Jimenéz-Martinéz et al., Environ. Entomol. 33, 1207 (2004).
ness in the real world. Because differences in (2004). 42. C. A. Preston, G. Laue, I. T. Baldwin, J. Chem. Ecol. 30,
performance among plants that are unable to 19. C. L. Ballaré, Trends Plant Sci. 4, 97 (1999). 2193 (2004).
20. R. Pierik, E. J. W. Visser, H. De Kroon, L. Voesenek, Plant 43. Funded by the Max Planck Society. The views expressed do
produce or perceive certain volatiles are likely Cell Environ. 26, 1229 (2003). not necessarily represent U.S. government views.
to be subtle, the analysis will likely require 21. R. Pierik, G. C. Whitelam, L. Voesenek, H. de Kroon,
long-term studies in natural settings. The more E. J. W. Visser, Plant J. 38, 310 (2004). 10.1126/science.1118446

during flowering, ripening, or maturation. Al-

REVIEW though a single fruit or vegetable synthesizes
several hundred volatiles, only a small subset
Plant Volatile Compounds: Sensory generates the Bflavor fingerprint[ that helps ani-
mals and humans recognize appropriate foods
and avoid poor or dangerous food choices.
Cues for Health and Nutritional Value? Although perception of flavor is often de-
scribed as a combination of taste and smell (5),
Stephen A. Goff1* and Harry J. Klee2 appearance, texture, temperature, mouth feel,
and past experience also play major roles in
Plants produce many volatile metabolites. A small subset of these compounds is sensed by flavor perception, indicating that multiple dis-
animals and humans, and the volatile profiles are defining elements of the distinct flavors of tinct sensory inputs are processed to generate the
individual foods. Flavor volatiles are derived from an array of nutrients, including amino acids, overall sensation (Fig. 1). Integration of this
fatty acids, and carotenoids. In tomato, almost all of the important flavor-related volatiles are sensory information in the brain ultimately re-
derived from essential nutrients. The predominance of volatiles derived from essential nutrients sults in a flavor preference or aversion with a
and health-promoting compounds suggests that these volatiles provide important information strong influence on subsequent perception and
about the nutritional makeup of foods. Evidence supporting a relation between volatile perception behavior. Studies of flavor preferences and
and nutrient or health value will be reviewed. aversions suggest that flavor perception may
be linked to the nutritional or health value as-
lants are capable of synthesizing tens to ary metabolites are generally low molecular sociated with the perceived foods (6–11). For

P hundreds of thousands of primary and

secondary metabolites with diverse bio-
logical properties and functions. Plant volatile
weight lipophilic compounds (1, 2). More than
7000 flavor volatiles have been identified and
cataloged from foods and beverages (3, 4).
example, fatty acids that stimulate taste responses
are essential long-chain cis-polyunsaturated fatty
acids rather than nonessential saturated fatty
organic compounds (defined hereafter as vola- Many volatiles are produced in plant tissues acids (11). Flavor preferences begin to develop
tiles) generated from both primary and second- at specific developmental stages—for example, before birth and develop rapidly in the newborn

www.sciencemag.org SCIENCE VOL 311 10 FEBRUARY 2006 815