Early accretion of water in the inner solar system from a

carbonaceous chondrite−like source

Adam R. Sarafian et al.
Science 346, 623 (2014);
DOI: 10.1126/science.1256717

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.
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

Downloaded from www.sciencemag.org on November 5, 2014

www.sciencemag.org (this information is current as of November 5, 2014 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/346/6209/623.full.html
Supporting Online Material can be found at:
http://www.sciencemag.org/content/suppl/2014/10/29/346.6209.623.DC1.html
This article cites 75 articles, 18 of which can be accessed free:
http://www.sciencemag.org/content/346/6209/623.full.html#ref-list-1
This article appears in the following subject collections:
Planetary Science
http://www.sciencemag.org/cgi/collection/planet_sci

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
2014 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
 RE S EAR CH | R E P O R T S

(529.1 eV) has been ascribed to O-Ce, the com-
ponent at 530 eV to O-metals (Rh and/or Pd),
and the component at 531.8 eV to –OH groups
present on the surface of the catalyst (Fig. 4, D to
F, O 1s) (28, 29). The ratio between the area of
the O-Ce peak and the –OH peak increases as the
photon energy increases; i.e., as the sampling
depth increases, which highlights the sensitivity
to different depths of the chosen photon energies
and the presence of the –OH groups on the sur-
face of the Rh0.5Pd0.5/CeO2 catalyst. The same
trend is observed for the unsupported NPs, where
the ratio between the O-metal peak and the –OH
peak also increases when the photon energy is
increased (Fig. 4, A to C, O 1s). In this case, the O 1s
high-binding-energy component could be related
to hydroxyl groups interacting with water mole-
cules (30), because they almost do not participate
in the ESR reaction (fig. S1), and the –OH groups
do not affect the oxidation state of the metals
(Fig. 2 C), because the inner shells are the most
oxidized ones.
Finally, the oxidation state of both metals in
the Rh0.5Pd0.5/CeO2 catalyst very nearly recov-
ered the values of the first reducing treatment at
lower temperature (Fig. 3A) during the reduction
with pure H2 at the reaction temperature (823 K,
Fig. 3C and fig. S5), unlike the model NPs, and
again this points to the stabilizing effect of the
ceria support. Analysis of the oxidation state of
cerium (performed only with hu = 1150 eV, fig.
S3) reveals variations depending on the gaseous
environment and temperature (22). Under H2
and 573 K, the Ce3+/Ce4+ atomic ratio was 0.63.
The ratio decreased to 0.58 during the ESR reac-
tion at 823 K, in accordance with the oxidation
experienced by the noble metals, and it strongly
increased up to 0.83 under H2 at 823 K.
In the absence of a support, the model NPs
were more strongly reduced for all the environ-
ments tested, and the reduction temperature af-
fected the amount of metallic phase found in Rh
and Pd. Compared with the unsupported NPs,
the capability of CeO2 as a support to activate wa-
ter and to donate oxygen atoms determined the
oxidation states of the noble metals under the
same reaction conditions. In addition, the inter-
action between the ceria support and the metal
nanoparticles prevented reorganization of the
Rh and Pd atoms. We have thus demonstrated
the pivotal role of the metal-support interaction
in the reorganization of metal atoms in supported
J. A. Rodriguez, J. C. Hanson, P. J. Chupas, Eds. (Wiley, 20. X. Peng, Q. Pan, G. L. Rempel, S. Wu, Catal. Commun. 11,
New York, 2013), pp. 315–343. 62–66 (2009).
4. D. E. Starr, Z. Liu, M. Hävecker, A. Knop-Gericke, H. Bluhm, 21. J. C. Garcia-Martinez, R. M. Crooks, J. Am. Chem. Soc. 126,
Chem. Soc. Rev. 42, 5833–5857 (2013). 16170–16178 (2004).
5. C. Escudero, M. Salmeron, Surf. Sci. 607, 2–9 (2013). 22. See the supplementary materials.
6. F. Tao et al., Science 322, 932–934 (2008). 23. Y. Qi et al., Nanoscale 6, 7012–7018 (2014).
7. H. Idriss et al., ChemSusChem 1, 905–910 (2008). 24. C. Wheeler, J. Catal. 223, 191–199 (2004).
8. N. J. Divins, E. López, Á. Rodríguez, D. Vega, J. Llorca, 25. V. Pérez-Dieste et al., J. Phys. Conf. Ser. 425, 072023 (2013).
Chem. Eng. Process. Process Intensif. 64, 31–37 (2013). 26. C. J. Powell, A. Jablonski, NIST Electron Inelastic-Mean-Free-
9. G. A. Deluga, J. R. Salge, L. D. Schmidt, X. E. Verykios, Science Path Database - Version 1.2 (National Institute of Standards
303, 993–997 (2004). and Technology, Gaithersburg, MD, 2010).
10. J. Llorca, V. Cortés Corberán, N. J. Divins, R. O. Fraile, 27. Y. Zhu et al., ACS Catal. 3, 2627–2639 (2013).
E. Taboada, in Renewable Hydrogen Technologies, L. M. Gandía, 28. M. M. Natile, A. Glisenti, Surf. Sci. Spectra 13, 17–30 (2006).
G. Arzamendi, P. M. Diéguez, Eds. (Elsevier, Amsterdam, 2013), 29. C. Force, E. Roman, J. M. Guil, J. Sanz, Langmuir ACS J. Surf.
pp. 135–169. Colloids 23, 4569–4574 (2007).
11. F. Aksoy et al., Nucl. Instruments Methods Phys. Res. Sect. 30. H. Sánchez Casalongue et al., Nat. Commun. 4, 2817 (2013).
A Accel. Spectrometers Detect. Assoc. Equip. 645, 260–265
(2011). AC KNOWLED GME NTS
12. K. Mudiyanselage et al., Angew. Chem. Int. Ed. Engl. 52, This work has been funded through grant MINECO
5101–5105 (2013). ENE2012-36368. J.L. is grateful to the ICREA Academia program.
13. C. Wen et al., ACS Nano 6, 9305–9313 (2012).
14. A. Caballero et al., Chem. Commun. (Camb.) 46, 1097–1099
(2010). SUPPLEMENTARY MATERIALS
15. C. Mattevi et al., J. Phys. Chem. C 112, 12207–12213 www.sciencemag.org/content/346/6209/620/suppl/DC1
(2008). Materials and Methods
16. A. Bruix et al., J. Am. Chem. Soc. 134, 8968–8974 (2012). Figs. S1 to S5
17. A. Trovarelli, Catal. Rev. 38, 439–520 (1996). References (31–33)
18. H. Song, L. Zhang, R. Watson, D. Braden, U. Ozkan,
Catal. Today 129, 346–354 (2007). 30 June 2014; accepted 2 October 2014
19. H. Idriss, Platin. Met. Rev. 48, 105–115 (2004). 10.1126/science.1258106
EARLY SOLAR SYSTEM
Early accretion of water in the inner
solar system from a carbonaceous
chondrite–like source
Adam R. Sarafian,1* Sune G. Nielsen,1 Horst R. Marschall,1,2,3
Francis M. McCubbin,4 Brian D. Monteleone1
Determining the origin of water and the timing of its accretion within the inner solar system
is important for understanding the dynamics of planet formation. The timing of water
accretion to the inner solar system also has implications for how and when life emerged
on Earth. We report in situ measurements of the hydrogen isotopic composition of the
mineral apatite in eucrite meteorites, whose parent body is the main-belt asteroid 4 Vesta.
These measurements sample one of the oldest hydrogen reservoirs in the solar system
and show that Vesta contains the same hydrogen isotopic composition as that of
carbonaceous chondrites. Taking into account the old ages of eucrite meteorites and their
similarity to Earth’s isotopic ratios of hydrogen, carbon, and nitrogen, we demonstrate
that these volatiles could have been added early to Earth, rather than gained during a late
accretion event.

bimetallic catalysts under operating conditions
and shown that this effect has a strong influence
on the catalytic performance. Keeping this in
mind, it should be noted that AP-XPS studies (and
other studies) carried out on unsupported model
systems may not provide reliable information on
supported real catalysts. For this reason, it is very
important to also take into account the influence
of the support in spite of the experimental diffi-
culties often encountered.
RE FE RENCES AND N OT ES
1. F. F. Tao, M. Salmeron, Science 331, 171–174 (2011).
2. S. Zafeiratos et al., J. Catal. 269, 309–317 (2010).
3. D. E. Starr, H. Bluhm, Z. Liu, A. Knop-Gericke, M. Hävecker,
in In-situ Characterization of Heterogeneous Catalysts,
H
ydrogen is vitally important in cosmochem-
ical and geochemical processes. For ex-
ample, water (H2O) plays a critical role in
plate tectonics on Earth (1) and has likely
shaped the surface of Mars (2). Despite the
abundance of water on Earth and evidence of
1
Department of Geology and Geophysics, Woods Hole
Oceanographic Institution, Woods Hole, MA 02543, USA.
2
Department of Earth and Planetary Sciences, American
Museum of Natural History, New York, NY 10024, USA.
3
School of Earth Sciences, University of Bristol, Bristol BS8
1RJ, UK. 4Institute of Meteoritics, University of New Mexico,
Albuquerque, NM 87131, USA.
*Corresponding author. E-mail: asarafian@whoi.edu
water on the Moon, Mars, and the asteroid 4
Vesta (2–4), these planetary bodies are often
thought to have accreted dry (5–8). This prompts
two key questions: Where did the water come
from? And when was it present in the inner solar
system? The answers may reveal information
about accretion processes in terrestrial planets.
Additionally, the extent and importance of lat-
eral water transport throughout the history of
the solar system are debatable (9, 10).
The source of water in planetary bodies can be
investigated by measuring the ratio between the
isotopes of hydrogen (deuterium, D or 2H, and
hydrogen, 1H) because different regions of the

SCIENCE sciencemag.org 31 OCTOBER 2014 • VOL 346 ISSUE 6209 623

R ES E A RC H | R E PO R TS

solar system vary widely in measured D/H ratios Table 1. Hydrogen isotope and water analyses of apatite and epoxy.
(11). Thus, determining D/H ratios in meteorites
with well-known ages and planetary origins can Sample dD (‰) T 2s H2O (mg/g) T 2s
help constrain the timing of water delivery to PCA 91078 ap1 –141 T 40 804 T 60
accreting planetary bodies. Such analysis pro- PCA 91078 ap2 –109 T 53 668 T 52
vides an ideal opportunity to study the prove- Pasamonte ap1 –53 T 29 2207 T 153
nance of water in bodies throughout the solar Pasamonte ap2 –216 T 25 1630 T 584
system, including Earth, the Moon, Mars, and Pasamonte ap3 –200 T 34 1545 T 295
the asteroid belt (12–14).
Juvinas ap1 –223 T 21 1781 T 152
Eucrites, a class of asteroidal basaltic meteor-
Juvinas ap2 –179 T 29 1594 T 107
ites, can provide unique information about the
Stannern ap1 –107 T 38 2624 T 182
timing of water delivery to planets because most
of these rocks crystallized 4559 to 4547 million Stannern ap2 –37 T 68 2390 T 157
years ago (Ma), or ~8 to 20 million years after Stannern ap3 –75 T 36 1595 T 107
the first solids in the solar system, calcium- and Cachari ap1 –158 T 24 932 T 62
aluminum-rich inclusions (CAIs) (15, 16). Eucrites PCA 91078 epoxy –174 T 9 140,000 T 9000
belong to the howardite-eucrite-diogenite group Weighted mean apatite –162 T 127 1748 T 1689
of meteorites (HEDs) that are derived from the
asteroid belt, dominantly from the asteroid
Vesta (17–19). Determining the source of water
Fig. 1. Hydrogen isotopes
in eucrites can constrain the time at which water
versus water content of
existed in the inner solar system, because these
eucrite apatites. Note the
ancient rocks are some of the oldest igneous rocks
large variation in water con-
in the solar system. However, until recently it
tent that contrasts with a
was believed that eucrites were completely anhy-
small range in the hydrogen
drous and therefore could provide no insight
isotopic compositions. Error
into the D/H ratio of Vesta (4, 6).
bars are 2s; where error bars
Here, we report the concentration and isotopic
are not present, they are
composition of structurally bound water in the
smaller than the symbol.
mineral apatite [Ca5(PO4)3(OH,F,Cl)] from five differ-
ent basaltic eucrite samples: Juvinas, Pasamonte,
Cachari, Stannern, and Pecora Escarpment (PCA)
91078 (figs. S1 and S2). We measured the D/H
and water contents simultaneously in situ using
a Cameca 1280 ion microprobe. The water con-
centration in these apatites ranges from 668 to
2624 mg/g, which agrees with calculated water
contents from electron probe measurements (table
S1) (4, 20). We measured dD values ranging from
–231 per mil (‰) to –37‰ (Table 1) (21). No
systematic variation in D/H exists between sam-
ples or between different apatite grains in the
same meteorite (Fig. 1). The weighted mean
hydrogen isotopic composition for eucritic apa-
tite from the five different investigated samples
is dD = –162 T 127‰ (2s; n = 11).
There are three processes that could have mod-
ified the hydrogen isotopic composition of the
apatite we analyzed from that of the parental
magma: (i) hydrous degassing before or during
apatite crystallization, (ii) equilibrium stable iso-
tope fractionation between hydrous phases and
melt, and (iii) contamination or assimilation of
exogenous H or D, presumably from the regolith
of Vesta. Degassing is unlikely due to the small
observed spread in dD with the relatively large
spread of water contents of the analyzed apatite
(22, 23). Apatite is the only hydrous mineral
phase in eucrites, and stable isotope fractiona-
tion between apatite and melt is expected to be
small (~20‰) at high temperature; thus, apatite- Fig. 2. Nitrogen isotopes versus hydrogen isotopic compositions of objects in the Solar System.
melt fractionation should be minor relative to Note that the fields for the Moon and Mars are large; these represent all D/H data because of a lack of
the variation in dD observed (24). Contamination consensus about the hydrogen isotopic compositions for the bulk Moon and Mars. For carbonaceous
by exogenic H from solar wind and/or by D pro- chondrites, ice in equilibrium with bulk clays has dD values of ~ –400‰ to 100‰ (12). Early car-
duced during spallation processes is unlikely, be- bonaceous chondrite ice could be another source of water for the inner solar system.The degassing model
cause apatites have young exposure ages, high line originating at protosolar values does not intersect Vesta, Earth, or the Moon. See (20) for details and
water contents, and a small range of dD across data. [Figure modified from (3, 28)]

624 31 OCTOBER 2014 • VOL 346 ISSUE 6209 sciencemag.org SCIENCE


Fig. 3. Carbon versus hydrogen isotopes. Note that Earth, the Moon, and Vesta plot within the
carbonaceous chondrite field. Because carbon isotopes vary from –250‰ to 200‰ in comets, C
isotopes are not useful for excluding comets as a source of C for inner solar system bodies.The degassing
line originating from the protosolar point assumes minimal C isotope fractionation and maximal H
fractionation. See (20) for details and data.
exposure ages and metamorphic grades (20). We
conclude that none of these processes have sub-
stantially affected the H isotopic compositions
measured in the present study (20).
On the basis of the weighted mean dD of the
apatites we analyzed, the dD value of eucrite
magmatic water is –162‰, which is the value we
adopt for bulk Vesta (20). This value is within the
combined uncertainties of the value for the bulk
Earth (~ –90‰) (22) and possibly the Moon
(~ +90‰; Fig. 2) (3, 23). The fact that Earth and
Vesta have indistinguishable dD values suggests
that they have the same source of water. Our new
D/H data, in combination with published bulk
nitrogen (N) isotope data for eucrites, shows that
the most likely common source of volatiles for
Earth and eucrites (Vesta) is carbonaceous chon-
drites (13, 25). Note that N isotopes exclude the
Jupiter family comets as a potential source of
water for Earth and Vesta. Nitrogen isotopes also
exclude Oort cloud comets as a source of water
for the inner solar system planetary bodies as
well. Kinetic isotope fractionation during both
accretion and magmatic processes would not al-
low the H and N isotope systems as currently
measured on Earth and Vesta to have evolved
from the material implied by the signature of the
Jupiter family comets or of the Oort cloud com-
ets (Fig. 2). Although we propose that carbona-
ceous chondrites provided water for the inner
solar system, there is evidence to show that at
least Earth accreted from a heterogeneous mix-
ture of hydrous and anhydrous materials, which
rules out carbonaceous chondrites as the only
terrestrial building blocks (8).
The hypothesis that inner solar system water
came from a carbonaceous chondrite source is
further supported by bulk carbon (C) isotope data.
The C isotopic compositions of Earth and Vesta
SCIENCE sciencemag.org
are very similar and overlap with that of car-
bonaceous chondrites but not with the solar com-
position (Fig. 3). Thus, the isotopic compositions
of three different volatile elements as well as their
concentration ratios for Earth and Vesta are
consistent with the composition of carbonaceous
chondrites.
Earth and Vesta have indistinguishable H, C,
and N isotopic compositions, and thus likely
have the same volatile source. This source must
have had an H isotopic composition similar to
or lower than the values measured in our sam-
ples, because possible magmatic fractionation pro-
cesses could only lead to a relative enrichment
and not a depletion in deuterium (20). One im-
portant reservoir that has low D/H and low
15
N/14N is the Sun (Fig. 2), which could theoret-
ically be an alternative source of volatiles in the
inner solar system, assuming that such a solar-
derived reservoir would have been later altered
by a secondary degassing process. Our model for
degassing uses the most conservative input val-
ues, which lead to the largest isotope fractiona-
tions for N and H isotopes and reflect degassing
of a 1:1 mixture of NH4 and H2. It is evident that
degassing alone cannot realistically evolve solar
H and N isotope ratios to values measured in
Vesta (Fig. 2). Degrees of degassing in excess of
99.99% are required to evolve N and H isotope
ratios of the Sun to ratios close to those of the
inner solar system. Additionally, if this extreme
degassing also involved volatile loss of C, then C
isotopes would significantly disagree with a solar
source of volatiles because the solar C isotopic
composition would become heavier than that of
Earth, the Moon, or Vesta (Fig. 3). It is highly
unlikely that accretion and degassing would
strongly fractionate H and N isotopes yet would
allow C isotopes on Vesta to remain isotopically
RE S EAR CH | R E P O R T S
light. Consequently, we rule out a solar source of
volatiles for the bodies in the inner solar system.
The H, C, and N isotopic similarities between
eucrites, Earth, and potentially the Moon allow
us to place important limits on the timing of
water delivery to the inner solar system. Earth
cannot provide timing of water delivery because
it is currently geologically active. The Moon like-
ly accreted its water at or before ~200 million
years after CAIs, or around 4367 Ma (3, 23), but
such a constraint is not very rigorous, given that
all the planets in the inner solar system are
thought to have fully accreted by this time. Eu-
crites provide a substantially earlier data point,
which suggests that the source of Earth’s water
was present in the inner solar system very early,
~8 to 20 million years after CAIs (15, 16). This
evidence moves back the time at which the
terrestrial water reservoir is thought to exist and
have been available for accretion. Additionally,
this reservoir was present between 1 and 2.4 AU
and perhaps throughout the inner solar system.
Late-stage addition of water to planets from outer
parts of the solar system is therefore unlikely to
have affected the water budgets of inner solar
system bodies. Thus, the bulk of the highly vola-
tile elements H, C, and N now present in Earth
and the asteroid belt most likely arrived from a
local source (i.e., carbonaceous chondrite–like
material) very early in solar system history. The
limited variation in dD over a large range of helio-
centric distances (1 to 2.4 AU) supports the notion
of a uniform source of water in the inner solar
system. Our findings cannot preclude a late addi-
tion of water for Earth with a carbonaceous
chondrite–like D/H, but the observation indicates
that a late addition of water is not necessary.
Our geochemical results are in agreement with
several dynamic solar system models for plane-
tary water delivery. These models indicate that
carbonaceous chondritic planetesimals delivered
water during primary accretion of Earth (9, 26, 27).
In addition, it is believed that Earth’s water con-
densed in the outer asteroid belt and giant plan-
et regions and was then transported to Earth by
dynamical processes related to giant planet migra-
tion (27). The implications are that (i) the migra-
tion of water in the inner solar system must have
started by 8 to 20 million years after CAIs, or (ii)
water was always present in the inner solar sys-
tem, in which case no water migration is needed
to satisfy the H isotopic composition of terres-
trial planets (10).
REFERENCES AND NOTES
1. I. Campbell, S. Taylor, Geophys. Res. Lett. 10, 1061–1064
(1983).
2. M. H. Carr, J. W. Head III, Earth Planet. Sci. Lett. 294, 185–203
(2010).
3. A. E. Saal, E. H. Hauri, J. A. Van Orman, M. J. Rutherford,
Science 340, 1317–1320 (2013).
4. A. R. Sarafian, M. F. Roden, A. E. Patiño Douce, Meteorit. Planet.
Sci. 48, 2135–2154 (2013).
5. F. Albarède, Nature 461, 1227–1233 (2009).
6. J. J. Papike, Rev. Mineral. 43, 7-01–7-10 (1998).
7. P. Lucey et al., Rev. Mineral. Geochem. 60, 83–219 (2006).
8. M. Schönbächler, R. W. Carlson, M. F. Horan, T. D. Mock,
E. H. Hauri, Science 328, 884–887 (2010).
9. A. Morbidelli, J. I. Lunine, D. P. Obrien, S. N. Raymond,
K. J. Walsh, Annu. Rev. Earth Planet. Sci. 40, 251–275 (2012).
31 OCTOBER 2014 • VOL 346 ISSUE 6209 625
R ES E A RC H | R E PO R TS

10. H. King et al., Earth Planet. Sci. Lett. 300, 11–18 (2010).
11. F. Robert, D. Gautier, B. Dubrulle, Space Sci. Rev. 92, 201–224
(2000).
12. C. M. Alexander et al., Science 337, 721–723 (2012).
13. F. Robert, Science 293, 1056–1058 (2001).
14. L. Hallis, G. Taylor, K. Nagashima, G. Huss, Earth Planet. Sci. Lett.
359–360, 84–92 (2012).
15. Q. Zhou et al., Geochim. Cosmochim. Acta 110, 152–175
(2013).
16. K. Misawa, A. Yamaguchi, K. Hiroshi, Geochim. Cosmochim. Acta
69, 5847–5861 (2005).
17. R. N. Clayton, N. Onuma, T. K. Mayeda, Earth Planet. Sci. Lett.
30, 10–18 (1976).
18. T. B. McCord, J. B. Adams, T. V. Johnson, Science 168,
1445–1447 (1970).
19. H. Y. McSween Jr., D. W. Mittlefehldt, A. W. Beck, R. G. Mayne,
T. J. McCoy, in The Dawn Mission to Minor Planets 4 Vesta and
1 Ceres (Springer, New York, 2012), pp. 141–174.
20. See supplementary materials on Science Online.
21. dD = {[(D/Hunknown)/(D/HVSMOW)] – 1} × 1000, where
VSMOW = Vienna standard mean ocean water.
22. Z. D. Sharp, F. M. McCubbin, C. K. Shearer, Earth Planet. Sci. Lett.
380, 88–97 (2013).
23. R. Tartèse et al., Geology 42, 363–366 (2014).
24. T. Chacko, D. R. Cole, J. Horita, Rev. Mineral. Geochem. 43,
1–81 (2001).
25. Y. Miura, N. Sugiura, Antarctic Meteor. Res. 6, 338 (1993).
26. A. Morbidelli et al., Meteorit. Planet. Sci. 35, 1309–1320
(2000).
27. D. P. O’Brien, K. J. Walsh, A. Morbidelli, S. N. Raymond,
A. M. Mandell, Icarus 239, 74–84 (2014).
ACKN OWLED GMEN TS
We thank the anonymous reviewers; K. Righter, D. Ebel, C. Agee,
and T. McCoy for sample allocations; and N. Shimizu and
G. Gaetani for their immense help. Supported by NASA graduate
fellowship NNX13AR90H (A.R.S.), an Andrew W. Mellon
Foundation Award for Innovative Research (S.G.N.), and NASA
Cosmochemistry Program award NNX11AG76G (F.M.M.).
Secondary ion mass spectrometry analyses were made at the
Northeast National Ion Microprobe Facility (NENIMF) at the Woods
Hole Oceanographic Institution (WHOI). NENIMF acknowledges
support from the NSF Instrumentation and Facilities Program,
Division of Earth Sciences, and from WHOI. Data are presented in
the main text of the manuscript and the supplementary materials.
SUPPLEMENTARY MATERIALS
www.sciencemag.org/content/346/6209/623/suppl/DC1
Materials and Methods
Supplementary Text
Figs. S1 and S2
Table S1
References (28–81)
30 May 2014; accepted 29 September 2014
10.1126/science.1256717

NEURODEVELOPMENT
apparently affected by sparse trkC removal
(figs. S2 and S3A), consistent with previous ob-
Dendrite morphogenesis servations (6, 7), we observed a distinctive Purkinje
cell dendrite phenotype (Fig. 1).
Both trkC+/+ and trkC+/– Purkinje cells ex-
depends on relative levels of tended complex dendritic arbors that spanned
the entire molecular layer of the cerebellum
NT-3/TrkC signaling (Fig. 1A, left and middle). In contrast, arbors
from trkC–/– cells failed to reach the pial surface
(65 out of 72 cells; Fig. 1A, right). These stunted
William Joo,1,2 Simon Hippenmeyer,1* Liqun Luo1,2† arbors exhibited normal dendritic spine density
(fig. S3B), but spanned only ~75% of the mo-
Neurotrophins regulate diverse aspects of neuronal development and plasticity, but their lecular layer (Fig. 1B). To determine when the
precise in vivo functions during neural circuit assembly in the central brain remain unclear. trkC–/– dendrite phenotype emerges, we com-
We show that the neurotrophin receptor tropomyosin-related kinase C (TrkC) is required pared trkC+/+ and trkC–/– cells between post-
for dendritic growth and branching of mouse cerebellar Purkinje cells. Sparse TrkC natal day 7 (P7) and P21, when Purkinje cells
knockout reduced dendrite complexity, but global Purkinje cell knockout had no effect. normally elaborate their dendrites and begin
Removal of the TrkC ligand neurotrophin-3 (NT-3) from cerebellar granule cells, which to form synapses (Fig. 1C and fig. S4). Although
provide major afferent input to developing Purkinje cell dendrites, rescued the dendrite indistinguishable from control cells at P7 and
defects caused by sparse TrkC disruption in Purkinje cells. Our data demonstrate that P10, trkC–/– cells exhibited reduced dendritic
NT-3 from presynaptic neurons (granule cells) is required for TrkC-dependent competitive arbor height, branch number, and total den-
dendrite morphogenesis in postsynaptic neurons (Purkinje cells)—a previously unknown drite length compared to trkC+/+ cells by day 14
mechanism of neural circuit development. (Fig. 1, C and D, and fig. S4). Growth and branch-

N
eurotrophins regulate the survival, differ-
entiation, and plasticity of peripheral and
central neurons (1–3). The mammalian neu-
rotrophin family signals through three
tropomyosin-related kinase (Trk) recep-
tors, as well as the p75 neurotrophin receptor
(p75NTR). Whereas brain-derived neurotrophic
factor (BDNF) has been intensely studied, much
less is known about neurotrophin-3 (NT-3) and
its receptor, TrkC, despite their widespread ex-
pression in the developing and adult brain (4, 5).
The lack of central brain cell death in mice lack-
ing NT-3 and TrkC contrasts starkly with the
severe reductions in sensory and sympathetic neu-
rons and suggests survival-independent functions
(6, 7). NT-3 functions in dendrite morphogenesis
in brain slice culture and in proprioceptive axon
1
Howard Hughes Medical Institute and Department of
Biology, Stanford University, Stanford, CA 94305, USA.
2
Neurosciences Program, Stanford University, Stanford, CA
94305, USA.
*Present address: Institute of Science and Technology Austria,
3400 Klosterneuburg, Austria. †Corresponding author. E-mail:
lluo@stanford.edu
patterning (8–10). However, evaluating the roles
of NT-3/TrkC signaling in the central brain in
vivo has been hindered by the early postnatal
lethality of NT-3 or TrkC knockout mice and the
limited cellular resolution of phenotypic analyses
(7, 11).
To study the cell-autonomous function of
TrkC in mouse neural development, we used
mosaic analysis with double markers (MADM)
(12, 13) with a null allele of trkC that removes all
isoforms of the receptor (14), and a pan-neural
Nestin-Cre line (15) to drive recombination through-
out the brain. Thus, in an otherwise heterozy-
gous animal (trkC+/–), Cre/loxP-mediated mitotic
recombination between homologous chromo-
somes sparsely labels wild-type (trkC+/+) and
homozygous mutant (trkC–/–) cells in distinct
colors (figs. S1A and S2). In these animals,
sparse trkC–/– cells were labeled with green flu-
orescent protein (GFP) (green), trkC+/+ cells with
tdTomato (red), and trkC+/– cells with both GFP
and tdTomato (yellow). Cells without recombina-
tion, which remained colorless, were all trkC+/–.
Although survival of central brain cells was not
ing phenotypes persisted in 3-month-old animals
(Fig. 1, C and D, and fig. S4), indicating that loss
of TrkC caused a morphogenesis defect rather
than a developmental delay. Furthermore, the
most distal dendritic branches were preferen-
tially lost in trkC–/– cells (fig. S3C). Uniparental
disomies (13) (fig. S5) and fluorescent markers
(fig. S6) did not affect Purkinje cell dendrite phe-
notypes. Thus, our mosaic analysis suggested that
TrkC is required cell-autonomously for proper
Purkinje cell dendrite growth and branching.
Although Trk signaling normally requires the
tyrosine kinase domain (16), TrkC also has a
kinase-independent role in synaptogenesis (17).
To determine whether dendritic arborization re-
lies on kinase activity, we examined MADM mice
harboring a conditional allele in which loxP
sites flank an exon encoding part of the TrkC
kinase domain (18). Here, Nestin-Cre mediated
interchromosomal recombination within the
MADM cassettes, and also excised this exon to
generate a “kinase-dead” allele (trkCKD; fig. S1B).
At P21, Purkinje cells homozygous for this allele
(trkCKD/KD) exhibited dendrite height, branch
number, and total dendritic length phenotypes

626 31 OCTOBER 2014 • VOL 346 ISSUE 6209 sciencemag.org SCIENCE