Rates of Molecular Evolution Are Linked to Life History in Flowering

Plants
Stephen A. Smith and Michael J. Donoghue
Science 322, 86 (2008);
DOI: 10.1126/science.1163197

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.

Downloaded from www.sciencemag.org on October 25, 2012

Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.

The following resources related to this article are available online at

www.sciencemag.org (this information is current as of October 25, 2012 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/322/5898/86.full.html
Supporting Online Material can be found at:
http://www.sciencemag.org/content/suppl/2008/10/02/322.5898.86.DC1.html
This article cites 34 articles, 15 of which can be accessed free:
http://www.sciencemag.org/content/322/5898/86.full.html#ref-list-1
This article has been cited by 50 article(s) on the ISI Web of Science
This article has been cited by 65 articles hosted by HighWire Press; see:
http://www.sciencemag.org/content/322/5898/86.full.html#related-urls
This article appears in the following subject collections:
Evolution
http://www.sciencemag.org/cgi/collection/evolution

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
2008 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
REPORTS
each phylogeny, we calculated the number of sub-
Rates of Molecular Evolution Are Linked stitutions per nucleotide site per million years using
branch lengths estimated from the dated molecu-
to Life History in Flowering Plants lar trees. Branch calculations were binned on the
basis of inferred life history to produce box plots
Stephen A. Smith* and Michael J. Donoghue for each clade (Fig. 1). Outliers (values >1.5 times
beyond the first and third quartiles) were excluded
Variable rates of molecular evolution have been documented across the tree of life, but the as artifacts of divergence-time estimation (e.g.,
cause of this observed variation within and among clades remains uncertain. In plants, it has been those with zero or near-zero branch lengths). Within
suggested that life history traits are correlated with the rate of molecular evolution, but each major clade, we noted that trees/shrubs were
previous studies have yielded conflicting results. Exceptionally large phylogenies of five major consistently evolving more slowly than related
angiosperm clades demonstrate that rates of molecular evolution are consistently low in trees herbaceous plants. Median rates of nucleotide
and shrubs, with relatively long generation times, as compared with related herbaceous plants, divergence were 2.7 to 10 times as high in herbs
which generally have shorter generation times. Herbs show much higher rates of molecular change as in trees/shrubs; herbs also showed higher
but also much higher variance in rates. Correlates of life history attributes have long been of ranges and variances (Fig. 1). None of the tree/
interest to biologists, and our results demonstrate how changes in the rate of molecular evolution shrub lineages examined here showed high rates
that are linked to life history traits can affect measurements of the tempo of evolution as of molecular evolution, but some herbaceous
well as our ability to identify and conserve biodiversity. lineages were inferred to have low evolutionary

Downloaded from www.sciencemag.org on October 25, 2012

rates, in the range characteristic of trees/shrubs.
ariation in the rate of molecular evolution those produced using other strategies (14). Spe- This asymmetry in variance may reflect the fact

V has been attributed to a number of fac-

tors, including differences in body size,
metabolic rate, DNA repair, and generation time
cifically, we identified alignable clusters of homol-
ogous gene regions, which were then concatenated
with profile alignment (13). To minimize missing
that, although most trees/shrubs are not able to
reproduce within the first few years (23, 24), as
most herbs can, some herbs take as long as trees
(e.g., 1–4). In plants, differences in rates of mo- data, only phylogenetically informative clusters to flower. Consistent with the view that genera-
lecular evolution have been noted between an- (with at least four taxa) were used. The gene re- tion time influences the rate of molecular
nuals and perennials (5) and between woody and gions varied among the five matrices but in each evolution within the Commelinidae (Fig. 1), the
herbaceous species (6, 7). These differences have case included markers from the chloroplast, nu- longer-lived bromeliads [which take up to 18
been presumed to reflect differences in genera- clear, and mitochondrial genomes (figs. S1 and years to reproduce (25)] have remarkably short
tion time (the time from seed germination to the S2 and table S2). The average gene region in our branches, with even fewer substitutions per site
production of fruits/seeds). However, in plants the analyses contained 305 species; the smallest con- per million years than palms (0.00059 and
relationship between life history and the average tained 10 species. This process resulted in an 0.0014, respectively). Other factors, such as
length of time before a nucleotide is copied either Apiales matrix of 1593 species by 9522 sites population size, breeding system, and seed-
through replication or repair [nucleotide genera- (>15 megabases); for Dipsacales, it was 366 by banking, may also relate to the observed
tion time (1)] is complicated by the fact that so- 11374 (>4 megabases); for Primulales, 529 by asymmetry; for example, the rate of fixation of
matic mutations can accumulate during growth 11505 (>6 megabases); for Moraceae/Urticaceae, mutations by selection increases in large popula-
and can be transmitted through gametes (8, 9). 457 by 7820 (>3.5 megabases); and for Comme- tions. Although we do not dismiss these variables
Variation in breeding system and/or seed-banking linidae, 4657 by 22391 sites (>104 megabases). in explaining the observed variance, they are less
by annual plants (9) may also affect the ability to Phylogenetic trees (Fig. 1) were inferred clearly correlated with the life history distinction
detect a correlation between molecular rate and under maximum-likelihood (ML) with RAxML than is generation time [e.g., (26)].
generation time. (vers.7.0.0) (15), with gene regions treated as To explore whether the difference in rates of
Previous studies have been inconclusive with separate partitions (13). We conducted 100 rapid molecular evolution has remained constant over
respect to the extent and the correlates of rate bootstrap analyses, using every 10th bootstrap time, we compared substitutions per site per mil-
heterogeneity in plants (5, 7, 9, 10). Studies fo- tree as a starting tree for a full ML search, and lion years through 10-million-year segments for
cused on individual smaller clades, or on single chose the tree with the highest likelihood score; each dated phylogeny (Fig. 2) (13). We found that
gene regions, have yielded results of uncertain owing to the size of the Commelinidae matrix, the trend in rate heterogeneity holds through time,
generality (7), whereas broader phylogenetic studies only a single ML search was conducted. For all with some noteworthy exceptions in the earliest
have suffered from limited taxon sampling and, clades but Commelinidae, we used nonparame- time periods. For example, woody Dipsacales are
hence, comparisons among very distant relatives tric rate smoothing (16) to set branch lengths estimated to have a high rate of evolution before
(11). Some tests have failed to account for phy- proportional to time; we used the PATHd8 meth- the herbaceous habit is inferred to have evolved
logenetic relatedness (9). od (17) for the exceptionally large commelinid in this lineage (Fig. 2B). Fossil data might help to
We assembled molecular sequence data for analysis. Published studies were used to cali- distinguish whether these results are best explained
five major branches within the flowering plants: brate each phylogeny, using multiple calibra- by incorrect reconstructions (i.e., perhaps the first
three clades of asterids (Apiales, Dipsacales, and tion points to limit the impact of clade-specific Dipsacales were herbaceous), by faster evolution
Primulales), one clade of rosids (Moraceae/Urticaceae), rate heterogeneity (13, 18–21). For Apiales and of woody lineages during earlier times (e.g., due
and one of monocotyledons (Commelinidae). We Primulales, we separately calibrated the major to warmer climate in the early Tertiary), or by the
used group-to-group profile alignments (12) that subclades identified in previous analyses, which extinction of early woody lineages.
take advantage of previously recognized clades also accommodated the fact that our analyses Because these comparisons do not directly
within the groups analyzed (13) and yield denser included some taxa not represented in previous take into account phylogenetic relationships or
data matrices (containing less missing data) than studies. examine the effects of evolutionary change from
Ancestral states of the life history trait “trees/ one life history state to the other, we calculated
shrubs” versus “herbs” (a proxy for generation branch length contrasts (27) around each inferred
Department of Ecology and Evolutionary Biology, 21 Sachem
Street, Post Office Box 208105, Yale University, New Haven, CT time) (6, 7, 22) were inferred with ML methods evolutionary shift in life history (Fig. 3) (13).
06520–8105, USA. (Fig. 1) (13); palms (Arecaceae, Commelinidae), Specifically, we calculated the average accumu-
*To whom correspondence should be addressed. E-mail: which do not produce true wood (secondary xylem), lation of molecular changes from each branch tip
stephen.smith@yale.edu were scored as trees/shrubs. For each branch on to the shared ancestor of a tree/shrub clade and

86 3 OCTOBER 2008 VOL 322 SCIENCE www.sciencemag.org

 REPORTS
compared this to the average accumulation in its en the impact of incorrectly estimating singleton 2.5 times as fast as trees/shrubs. A maximum
herbaceous sister clade. We started from the most branches (branch lengths were averaged in clades rate difference of 4.75 times was found between
nested clades and worked toward the root, ex- with two or more species). Dorstenia (Moraceae) and its tree/shrub sister clade.
cising any nested contrasts from the more in- Of the 13 contrasts identified using these crite- The single exception occurred within Sambucus
clusive calculations to avoid measuring any node ria (Table 1 and Fig. 3), 12 showed a slower rate (Dipsacales), where the tree/shrub species showed
more than once. We omitted contrasts containing of molecular evolution in trees/shrubs than in herbs a slightly higher rate than the herbs (0.0075 and
only one tree/shrub or one herb branch to less- (sign test, P = 0.00342). On average, herbs evolve 0.0061, respectively). This case involved the smallest

Fig. 1. Phylogenies of five

angiosperm clades with
branch lengths proportion-
al to substitutions per site.
Branch colors represent in-

0.015
0.01
ferred life history states
(brown for trees/shrubs;
0.000

0.000
green for herbs). Box plots
show substitutions per site Apiales
per million years for the Commelinids

Downloaded from www.sciencemag.org on October 25, 2012

inferred life history cate-
gories; centerline represents
the median, hinges mark
the first and third quar-

0.01
tiles, whiskers extend to Bromeliaceae
0.010

the lowest and highest non-

0.000
outlier. Outliers (not shown)
have values >1.5 times
0.000

0.005
beyond the first or third Dipsacales
quartiles.

0.000
Urticaceae

Primulales

Moraceae

Fig. 2. Dated phylog-

enies for Apiales and

Valerianaceae
Dipsacales with substi- Apiales Dipsacales
tutions per site per mil-
lion years plotted for
10-million-year intervals
through the life of the
Apiaceae

clade. Branch colors rep-

Dips
resent inferred life his-
tory states (brown for M L
trees/shrubs; green for
herbs). Box plots as in
Fig. 1. PM, Pittospora-
Caprifolieae

ceae and Myodocarpa-

ceae; Dips, Dipsacaceae;
PM

M, Morinaceae; L, Linna-
ceae.
Araliaceae

Adoxaceae

70 60 50 40 30 20 10 0 100 80 60 40 20 0
0.0015
0.002
0.0000

0.0000

www.sciencemag.org SCIENCE VOL 322 3 OCTOBER 2008 87

REPORTS
numbers of species (three shrubby species versus ing all Sambucus species as trees/shrubs. These condition in Apiales, Dipsacales, and Primulales,
three herbs) and also presented the greatest diffi- alternatives (Table 1 and Fig. 3) yielded a simi- and with less certainly in Moraceae/Urticaceae.
culty in assigning life history states (the herbaceous larly strong historical correlation (P = 0.00049), The palms (Arecaceae) within the Commelinidae
species are subshrubby and the woody species ma- as did the exclusion of these contrasts altogether present the one clear instance in our sample of the
ture rapidly). As such uncertainties are inherent (P = 0.00195). evolution of trees/shrubs from herbaceous ances-
in large comparative analyses, we explored wheth- On the basis of our trees and broader phylo- tors (30). From our comparisons and a broader
er alternative phylogenetic hypotheses (13) af- genetic studies of angiosperms [reviewed in (28); analysis of monocotyledons (11), the shift to the
fected the results for the smallest clade examined see also (29)], the likely direction of evolution of tree/shrub habit in palms was associated with a
here, the Dipsacales, as well as the effect of scor- plant habit was from trees/shrubs to the herbaceous marked decrease in the rate of molecular evo-
lution (palms evolve 2.7 times as slow as their
sister commelinids), as predicted by the hypoth-
A 4 B esis that generation time drives the rate of mo-
0.4

lecular evolution.

5
5
Differences in rates of evolution associated

difference herb vs trees/shrubs

15
14 with generation time may be reflected most clear-

4
ly in synonymous substitutions within coding se-
0.3

1
molecular change

3 2
4 quences (31). We analyzed 1208 commelinid rbcL
11 sequences, pruning species lacking an rbcL se-

3
1

Downloaded from www.sciencemag.org on October 25, 2012

quence in GenBank from our Commelinidae phy-
0.2

13
9 7

6
8 logeny and using RAxML to estimate branch
9

2
8
6 lengths for several partitions of the data (Table 2)
3
5
(13). As expected, estimated amino acid branch
0.1

13 10
7
lengths showed the least difference in rate be-

1
14
2 12
11 tween life history classes (2.1 times as fast as in
15
10 herbs), with first and second nucleotide positions
0.0

0
12 being next smallest (3.2 times as fast). The rate
trees/shrubs herb difference in the full Commelinidae data set (all
Fig. 3. Branch-length contrasts for trees/shrubs versus herbs. (A) Lines are drawn between the accumulated species, all genes) fell between these two values
average molecular branch lengths for each tree/shrub clade and its sister herbaceous clade (numbers (2.7 times as fast in herbs). The third positions
correspond to those in Table 1). All evolutionary shifts were inferred to be from trees/shrubs to herbs except showed the greatest difference in rate (4.98 times
for the evolution of palms within monocotyledons (arrowhead in contrast 4). Contrasts 1 to 13 were used in as fast in herbs). These findings are similar to
an initial sign test (P = 0.00342). Alternative contrasts within the Dipsacales (14 and 15) are marked by those based on a much smaller sample of rbcL
dotted lines and were substituted for 11 to 13 in one test (P = 0.00049); contrasts 11 to 15 were omitted in a sequences from grasses and palms (11).
third test (P = 0.00195). (B) Magnitude of change between each tree/shrub clade and its herbaceous sister Our findings highlight the need for the meth-
clades; values above 1 show higher rates of molecular evolution in herbs than in trees/shrubs. ods used to date phylogenies to address the form
of clade-dependent heterogeneity documented here.
A rate of nucleotide substitution obtained from
Table 1. Branch length contrasts 1 to 13 derive from the trees in Fig. 1 [see (13) for more exact locations an herbaceous group cannot be used to calibrate a
of the nodes in question]. Plants in the first taxon in each pair of representative taxa are trees/shrubs;
clade of trees/shrubs, or vice versa, without con-
plants in the second are herbs. Within Dipsacales, we explored alternative contrasts, substituting contrasts
founding age estimates. Likewise, relaxed clock
14 and 15 for 11 to 13 in one test and omitting contrasts 11 to 15 in another.
methods [e.g., (32)] are likely to estimate that slow-
Major clade Representative taxa Trees/shrubs Herbs Difference ly evolving groups are younger, and that rapidly
evolving groups are older, than their true ages. It
Apiales 1 Astrotricha–Hydrocotyle 0.0564 0.2173 3.8538
may be possible to avoid mixing clades with very
2 Aralia–Panax 0.0198 0.0679 3.4224
different life histories in designing dating studies.
3 Pittosporaceae–Apiaceae 0.1692 0.2724 1.6097
Otherwise, as we have attempted here, the use of
Commelinidae 4 Arecales–remaining 0.1363 0.4350 3.1915
multiple calibration points spanning clades that
Commelinidae
differ in life history may help alleviate this prob-
Moraceae– 5 Brosimum–Dorstenia 0.0213 0.1013 4.7527
lem. Also, as shown here for Commelinidae, the
Urticaceae 6 Moraceae–Urticaceae 0.1002 0.1800 1.7967
use of amino acid sequences (or the removal of
7 Cecropia/Coussapoa– 0.0361 0.0873 2.4169
third sites) may be useful. Bayesian models that
Boehmeria
do not assume an autocorrelated rate of molecular
Primulales 8 Ardisia–sister Myrsinaceae 0.0433 0.1011 2.3330
evolution [e.g., (33)] are promising, but current
9 Theophrastaceae– 0.0953 0.1824 1.9138
methods are incapable of analyzing large data sets.
Myrsinaceae/Primulaceae
We hope that our results will also focus new
Dipsacales 10 Symphoricarpos–Triosteum 0.0142 0.0210 1.4747
attention on the extent to which molecular and
11 Linnaeeae–Morinaceae 0.0217 0.0626 2.8912
morphological evolution are coupled [see (34, 35)].
12 Woody Sambucus– 0.0075 0.0061 0.8130
Are rates of morphological evolution also slower
herbaceous Sambucus
in trees/shrubs than in herbs [e.g., (36)]? Until
13 Viburnum–Adoxa 0.0352 0.0856 2.4304
this question is addressed, we urge caution in as-
14 Linnaeeae– 0.0219 0.0863 3.9432
suming that morphological change scales with mo-
Morinaceae/Dipsacaceae/
lecular change and in using molecular branch
Valerianaceae
lengths alone to assess “feature diversity” and de-
15 “Woody” Sambucus– 0.0133 0.0554 4.1783
sign conservation strategies [e.g., (37)]. A related
Adoxa
issue is the likely success of “barcoding” methods

88 3 OCTOBER 2008 VOL 322 SCIENCE www.sciencemag.org

 REPORTS
Table 2. Branch length contrast estimates for different partitions of rbcL sequences from Commelinidae. 20. N. J. C. Zerega, W. L. Clement, S. L. Datwyler, G. D. Weiblen,
Mol. Phylogenet. Evol. 37, 402 (2005).
Estimate data Palms Rest of Commelinidae Difference 21. T. Janssen, K. Bremer, Bot. J. Linn. Soc. 146, 385 (2004).
22. J. Bousquet, S. H. Strauss, A. H. Doerksen, R. A. Price,
Third sites 0.0369 0.1842 4.9872 Proc. Natl. Acad. Sci. U.S.A. 89, 7844 (1992).
First, second, and third sites 0.0331 0.1284 3.8808 23. R. J. Petit, A. Hampe, Annu. Rev. Ecol. Evol. Syst. 37, 187
(2006).
First and second sites 0.0313 0.1007 3.2180 24. M. Verdú, Evol. Int. J. Org. Evol. 56, 1352 (2002).
Amino acids 0.0464 0.0993 2.1379 25. P. Hietz, J. Ausserer, G. Schindler, J. Trop. Ecol. 18, 177
All genes/species 0.1363 0.4350 3.1915 (2002).
26. S. S. Renner, R. E. Ricklefs, Evol. Int. J. Org. Evol. 82,
596 (1995).
27. T. J. Davies, V. Savolainen, Evol. Int. J. Org. Evol. 60, 476 (2006).
for identifying plant species from short DNA se- References and Notes
28. D. E. Soltis, P. S. Soltis, P. K. Endress, M. W. Chase,
quences [reviewed in (33)]. We predict that the 1. A. P. Martin, S. R. Palumbi, Proc. Natl. Acad. Sci. U.S.A.
Phylogeny and Evolution of Angiosperms (Sinaur
90, 4087 (1993).
chloroplast genes proposed as universal barcode 2. A. O. Mooers, P. H. Harvey, Mol. Phylogenet. Evol. 3, 344
Associates, Sunderland, MA, 2005).
loci will be most successful in resolving herba- 29. P. F. Stevens, Angiosperm Phylogeny Website,
(1994).
www.mobot.org/MOBOT/research/APweb (2008).
ceous species and may be incapable of confident- 3. L. Bromham, A. Rambaut, P. H. Harvey, J. Mol. Evol. 43,
30. M. W. Chase et al., Aliso 22, 63 (2006).
ly distinguishing closely related woody species. 610 (1996).
31. T. Ohta, Proc. Natl. Acad. Sci. U.S.A. 90, 10676 (1993).
4. J. F. Gillooly, A. P. Allen, G. B. West, J. H. Brown, Proc.
Finally, our studies underscore the need for Natl. Acad. Sci. U.S.A. 102, 140 (2005).
32. M. J. Sanderson, Mol. Biol. Evol. 19, 101 (2002).
better and more accessible information on the 33. A. J. Drummond, S. Y. W. Ho, M. J. Phillips, A. Rambaut,
5. K. Andreasen, B. G. Baldwin, Mol. Biol. Evol. 18, 936 (2001).

Downloaded from www.sciencemag.org on October 25, 2012

underlying drivers of rates of molecular evolu- PLoS Biol. 4, e88 (2006).
6. J. Laroche, P. Li, L. Maggia, J. Bousquet, Proc. Natl. Acad.
34. R. Lahaye, Proc. Natl. Acad. Sci. U.S.A. 105, 2923 (2008).
tion. In addition to data on generation times, we Sci. U.S.A. 94, 5722 (1997).
35. T. G. Barraclough, V. Savolainen, Evol. Int. J. Org. Evol.
7. K. M. Kay, J. B. Whittall, S. A. Hodges, BMC Evol. Biol. 6,
need better knowledge of effective population 55, 677 (2001).
36 (2006).
sizes. Past analyses (e.g., in mammals) have as- 8. E. J. Klekowski, P. J. Godfrey, Nature 340, 389 (1989).
36. E. W. Sinnott, Am. Nat. 50, 466 (1916).
sumed that larger, longer-lived organisms have 37. F. Forest et al., Nature 445, 757 (2007).
9. C. A. Whittle, M. O. Johnston, J. Mol. Evol. 56, 223 (2003).
38. We thank D. Tank, J. Beaulieu, S. Stearns, D. Weinreich,
smaller population sizes, but this may be reversed 10. K. Andreasen, Conserv. Genet. 6, 399 (2005).
C. Dunn, and E. Edwards for helpful discussions.
in plants, where tropical trees often appear to have 11. B. S. Gaut, B. R. Morton, B. C. McCaig, M. T. Clegg, Proc.
Funded by the NSF through the Cyberinfrastructure for
Natl. Acad. Sci. U.S.A. 93, 10274 (1996).
large population sizes (31). Our analyses imply 12. R. C. Edgar, Nucleic Acids Res. 32, 1792 (2004).
Phylogenetic Research (EF-0331654) and the angiosperm
that somatic mutation has not counteracted the Assembling the Tree of Life (EF-0431258) projects.
13. Materials and methods are available as supporting
influence of generation time on rates of evolu- material on Science Online. Supporting Online Material
tion, but more data are needed on the rate and fate 14. M. M. McMahon, M. J. Sanderson, Syst. Biol. 55, 818 (2006). www.sciencemag.org/cgi/content/full/322/5898/86/DC1
15. A. Stamatakis, Bioinformatics 22, 2688 (2006). Materials and Methods
of such mutations (8). In any event, our analyses Fig. S1 to S8
16. M. J. Sanderson, Mol. Biol. Evol. 14, 1218 (1997).
demonstrate a general pattern that must now be Tables S1 and S2
17. T. Britton, C. L. Anderson, D. Jacquet, S. Lundqvist,
taken into account in evolutionary studies and K. Bremer, Syst. Biol. 56, 741 (2007). References
whose existence demands the elaboration of a 18. K. Bremer, E. M. Friis, B. Bremer, Syst. Biol. 53, 496 (2004). 14 July 2008; accepted 5 September 2008
cohesive causal explanation. 19. C. D. Bell, M. J. Donoghue, Am. J. Bot. 92, 284 (2005). 10.1126/science.1163197

involute at the gastrula margin (blastopore) and

Chemokine Signaling Controls move anteriorly toward the animal pole (future
head) while converging toward the midline (con-
Endodermal Migration During vergent extension).
The chemokine receptor CXCR4 controls
Zebrafish Gastrulation directional migration in many contexts and is
expressed in the endoderm. It is up-regulated by
the endodermal determinants Mixer and Sox17b
Sreelaja Nair and Thomas F. Schilling* (9–13) and is required for gastrointestinal
vascularization (14). Of the two closely related
Directed cell movements during gastrulation establish the germ layers of the vertebrate embryo zebrafish Cxcr4s, Cxcr4b regulates the migration
and coordinate their contributions to different tissues and organs. Anterior migration of the of many cell types (12, 15–18), but no roles have
mesoderm and endoderm has largely been interpreted to result from epiboly and convergent- been reported for Cxcr4a during embryogenesis.
extension movements that drive body elongation. We show that the chemokine Cxcl12b and its Zebrafish embryos deficient in Cxcr4a or
receptor Cxcr4a restrict anterior migration of the endoderm during zebrafish gastrulation, thereby Cxcl12b, generated by injection with antisense
coordinating its movements with those of the mesoderm. Depletion of either gene product causes morpholino oligonucleotides (MO), appeared
disruption of integrin-dependent cell adhesion, resulting in separation of the endoderm from the morphologically normal (Fig. 1, A to C, and
mesoderm; the endoderm then migrates farther anteriorly than it normally would, resulting in fig. S1). However, analysis of Tg(gutGFP)s854
bilateral duplication of endodermal organs. This process may have relevance to human transgenic embryos in which the entire gut
gastrointestinal bifurcations and other organ defects. fluoresces [(19); GFP, green fluorescent protein]
revealed duplications of endodermal organs at
crucial feature of vertebrate embryo- and blood), whereas mesodermal defects disrupt 56 hours post-fertilization (hpf) (Fig. 1, D to F,

A genesis is the coordinated morphogene-

sis of germ layers (endoderm, mesoderm,
and ectoderm) during gastrulation (1). Interac-
the locations of the liver and pancreas (2–5).
Morphogenesis is regulated by Wnt (6) and
Nodal signaling (7) when cells are intermingled
and fig. S2), including the pancreas (normally
on the right; fig. S3, A to F) and liver (normally

tions between the endoderm and mesoderm in a bipotential “mesendoderm” (8). However, Department of Developmental and Cell Biology, University
specify organ locations and symmetries (2). relatively little is known about germ layer– of California, Irvine, CA 92697, USA.
Defects in the endoderm alter the morphogene- specific pathways that establish organ rudiments. *To whom correspondence should be addressed. E-mail:
sis of mesodermal organs (e.g., heart, kidneys, In zebrafish, mesendodermal organ progenitors tschilli@uci.edu

www.sciencemag.org SCIENCE VOL 322 3 OCTOBER 2008 89