© 2005 Nature Publishing Group
WUSCHEL controls meristem function by direct
regulation of cytokinin-inducible response
regulators
Andrea Leibfried1
, Jennifer P. C. To2
, Wolfgang Busch1
*, Sandra Stehling1
*, Andreas Kehle1
, Monika Demar1
,
Joseph J. Kieber2
& Jan U. Lohmann1
Plants continuously maintain pools of totipotent stem cells in
their apical meristems from which elaborate root and shoot
systems are produced. In Arabidopsis thaliana, stem cell fate in
the shoot apical meristem is controlled by a regulatory network
that includes the CLAVATA (CLV) ligand–receptor system and the
homeodomain protein WUSCHEL (WUS)1,2
. Phytohormones
such as auxin and cytokinin are also important for meristem
regulation3
. Here we show a mechanistic link between the
CLV/WUS network and hormonal control. WUS, a positive regulator
of stem cells, directly represses the transcription of several
two-component ARABIDOPSIS RESPONSE REGULATOR genes
(ARR5, ARR6, ARR7 and ARR15), which act in the negativefeedback
loop of cytokinin signalling4,5
. These data indicate that
ARR genes might negatively influence meristem size and that their
repression by WUS might be necessary for proper meristem
function. Consistent with this hypothesis is our observation
that a mutant ARR7 allele, which mimics the active, phosphorylated
form, causes the formation of aberrant shoot apical meristems.
Conversely, a loss-of-function mutation in a maize ARR
homologue was recently shown to cause enlarged meristems6
.
Genetic analyses have led to the discovery of several essential
regulators of stem cell fate in the shoot apical meristem of the reference
plant Arabidopsis thaliana. Among them, the homeodomain transcription
factors WUSCHEL (WUS) and SHOOTMERISTEM-LESS
(STM) have positive functions7,8
, whereas the CLAVATA (CLV) genes
negatively influence meristem size9–11
. WUS is expressed in the
organizing centre and induces stem cell fate in the overlaying cells12
that in turn express CLV3, a small secreted peptide13,14
that is thought
to act as ligand for the CLV1–CLV2 heteromeric receptor com-
plex15,16
. Activation of the CLV1–CLV2 receptor leads to the suppression
of WUS expression, creating a negative feedback loop that
controls the size of the stem cell pool1,2
.
Despite the central role of the WUS transcription factor in the
initiation and maintenance of stem cell fate, only a single direct
target, the floral homoeotic gene AGAMOUS (AG), which represses
the maintenance of stem cells in the flower, has been described17
. To
identify target genes of WUS and other meristem regulators, we
performed a comparative microarray screen using plants with
ethanol-inducible overexpression alleles18
of WUS as well as STM
and LEAFY (LFY), a floral regulator that interacts with WUS17,19
.
After 12 h of treatment with ethanol we harvested the shoot apex and
surrounding tissue (Fig. 1a) and subjected it to expression profiling
with Affymetrix Ath1 arrays. A combination of per-gene and common
variance20
filtering was used to identify 148 genes responsive to
LETTERS
Figure 1 | Expression profiles of ARR5, ARR6, ARR7
and ARR15. a, A 12-day-old seedling showing
ectopic AG::GUS reporter gene activation in
response to WUS induction. Tissue used for
expression profiling is indicated. b, Expression of
ARR5 (blue), ARR6 (purple), ARR7 (green) and
ARR15 (red) is specifically repressed by WUS as
detected by microarrays. c, Real-time qRT–PCR
confirms rapid repression of ARR genes by WUS.
Relative expression is normalized to induced
AlcA::GUS controls. Line colours are as in b. d, ARR
expression in response to downregulation of WUS
by induction of AlcA::CLV3 (grey bars). Black bars,
AlcA::GUS. Relative expression measured by realtime
qRT–PCR is normalized to TUBULIN. Error
bars indicate s.e.m.
1
Max Planck Institute for Developmental Biology, AG Lohmann, D-72076 Tu¨bingen, Germany. 2
Department of Biology, University of North Carolina, Chapel Hill, North Carolina
27599, USA.
*These authors contributed equally to this work.
Vol 438|22/29 December 2005|doi:10.1038/nature04270
1172
© 2005 Nature Publishing Group
WUS but not to STM or LFY induction (Supplementary Table 1). Of
these 148 genes, 44 were repressed, including ARR5, ARR6, ARR7 and
ARR15, which belong to the 10-member type-A ARABIDOPSIS
RESPONSE REGULATOR gene family21
(Fig. 1b). Type-A ARR
proteins contain a phosphate-accepting receiver domain similar to
bacterial two-component response regulators, but in contrast to
type-B ARR proteins they lack a DNA-binding motif in their output
domain22
. Their expression is rapidly induced by cytokinin21
, which
has been shown to be a potent inductor of cell proliferation when
applied exogenously together with auxin and to induce shoot
development when acting alone23
. Type-A ARR proteins have been
implicated in the negative feedback regulation of cytokinin signalling
on the basis of the observation of decreased hormone sensitivity
in plants overexpressing type-A ARR genes4,24
. Furthermore, in
Arabidopsis, type-A arr multiple mutants have increased cytokinin
sensitivity. However, even in sextuple type-A arr mutants (arr3 arr4
arr5 arr6 arr8 arr9) morphological changes are minimal, indicating
strong redundancy within the gene family5
. ARR5 and ARR6, as well
as ARR7 and ARR15, constitute closely related pairs within the gene
family21
, and inspection of the AtGenExpress expression atlas25
revealed co-expression of each pair, marked by widespread transcription
with highest levels in meristematic tissue for ARR7 and ARR15,
and in roots for ARR5 and ARR6.
By using quantitative real-time reverse transcriptase-mediated
polymerase chain reaction (qRT–PCR), we found that 4 h after
WUS induction by ethanol, RNA levels of ARR5, ARR6, ARR7 and
ARR15 were already decreased, and after 12 h they reached a
minimum at about 10% of control levels. Expression levels remained
low for at least 48 h after treatment with ethanol (Fig. 1c). To test
whether WUS is not only sufficient but also necessary for the
repression of ARR5, ARR6, ARR7 and ARR15 in wild-type meristems,
we used inducible CLV3 to transiently repress WUS, because the
morphology of wus mutants deviates strongly from the wild type
even at very early stages of development7
. Besides a strong reduction
of WUS expression, we observed by qRT–PCR a moderate increase in
expression of the ARR genes after 24 h of CLV3 induction (Fig. 1d),
which is consistent with the idea that ARR expression extended into
the small WUS domain in these plants.
In situ hybridization on sections of inflorescence meristems
demonstrated that ARR7 RNA accumulates in a subdomain of the
meristem consistent with a potential function in this tissue (Fig. 2a,
and Supplementary Fig. 1). Reporter gene analysis confirmed this
pattern and showed in addition that ARR5, ARR6 and ARR15
promoters are also active in the meristem (Supplementary Fig. 2).
Figure 2 | Expression patterns of ARR7 and WUS in response to
meristematic signals. Upper panels show in situ hybridizations of ARR7;
lower panels show WUS. a, g, Wild type. b, h, Induced 35S::AlcR AlcA::WUS.
WUS is moderately expressed in all cells with hot spots in more mature
tissue (arrowheads). c, i, clv3-7 mutant. d, j, Induced 35S::AlcR AlcA::CLV3.
e, 6-Benzylaminopurine-treated wild type. f, CLV3 and WUS expression in
inflorescence apices of wild-type and arr3 arr4 arr5 arr6 arr7 arr8 arr9
septuple mutants as measured by qRT–PCR. Dark colours represent
wild-type, light colours indicate mutant. k, 35S::ARR7. l, arr3 arr4 arr5 arr6
arr7 arr8 arr9 septuple mutant.
Figure 3 | Direct interaction of WUS with regulatory sequences of ARR7.
a, Real-time qRT–PCR on 35S::WUS:GR plants. Dark grey bars represent
mock treatment, light grey bars induction with dexamethasone, crosshatched
bars mock treatment in the presence of cycloheximide, and hatched
bars induction with dexamethasone in the presence of cycloheximide.
Expression values are normalized to the respective mock treatment controls
(see Supplementary Fig. 3a for alternative normalization). b, Detection of
ARR7 regulatory sequences by real-time qRT–PCR after ChIP with antiWUS
antiserum (see Supplementary Fig. 3b). Enrichment of overlapping
genomic fragments upstream of the ARR7 start codon is shown after
normalization to unrelated control sequences (see also Supplementary
Fig. 3c for alternative normalization). ChIP was performed on induced
35S::WUS:GR tissue. Asterisk, promoter fragment used for gel shifts.
c, EMSA using ARR7 promoter sequences identified in b; 2, free probe; E,
control protein extract from yeast expressing LEAFY; W, protein extract
from yeast expressing WUSCHEL; dIdC, poly(dIdC) used as unspecific
competitor; pARR7, unlabelled probe used as specific competitor.
NATURE|Vol 438|22/29 December 2005 LETTERS
1173
© 2005 Nature Publishing Group
In plants with an inducible WUS transgene (Fig. 2h), ARR7 RNA
could no longer be detected 24 h after WUS induction (Fig. 2b),
which is similar to the situation in clv3 mutants (Fig. 2c), in which
WUS expression is expanded (Fig. 2g, i). Conversely, after suppression
of WUS by CLV3 induction (Fig. 2j), we observed an expansion
of the ARR7 expression domain (Fig. 2d), confirming the qRT–PCR
results. Activation of ARR7 in cells outside the WUS domain might
indicate a more direct effect of CLV3 on ARR7 expression in parallel
to its WUS dependent activity. Similarly to what has been observed
for the maize homologue ABPH1 (ref. 6), a 30-min treatment with
the synthetic cytokinin 6-benzylaminopurine caused an expansion of
ARR7 expression in the wild type (Fig. 2e).
An additional level of regulation is provided by negative feedback
of ARR7 on WUS, because plants that overexpress ARR7 from the
constitutive 35S promoter have lower WUS RNA levels (Fig. 2k).
However, residual WUS activity in 35S::ARR7 plants is sufficient for
correct function of the meristem, because 35S::ARR7 plants have no
obvious defects in the shoot apical meristem, similar to induced
AlcA::CLV3 plants, which show a wus mutant phenotype only in
flowers (data not shown).
Having established a regulatory interaction between WUS and
ARR7, we next asked whether this interaction is direct. To this end,
we first made use of an inducible form of WUS by means of a
translational fusion to the ligand-binding domain of the rat glucocorticoid
receptor (WUS:GR). Application of a steroid such as
dexamethasone causes translocation of the fusion protein from the
cytoplasm to the nucleus, allowing activation or repression of direct
targets in the absence of protein synthesis26,27
. After treatment of
35S::WUS:GR plants with dexamethasone for 4 h, we observed robust
repression of ARR5, ARR6, ARR7 and ARR15. Repression of the ARR
genes also occurred in the presence of the protein synthesis inhibitor
cycloheximide (Fig. 3a, and Supplementary Fig. 3a), which is
compatible with a direct interaction of WUS with the regulatory
elements of the ARR genes. We then confirmed in vivo binding of
WUS to ARR7 promoter sequences by chromatin immunoprecipitation
(ChIP) with a polyclonal anti-WUS antiserum (Fig. 3b, and
Supplementary Fig. 3b, c). We observed a twofold enrichment
of ARR7 promoter DNA in wild-type inflorescences in comparison
with leaves, in which WUS is not expressed, whereas in WUSoverexpressing
tissue ARR7 promoter DNA was enriched 68-fold.
The ChIP results indicated binding of WUS to sequences located
about 1,000 base pairs upstream of the start codon of ARR7 in a
region harbouring multiple TAATelements, which have been shown
to be the core binding sites for WUS (Fig. 3b)17
. Subsequently, we
were able to confirm sequence-specific binding of WUS protein to
this promoter element by electrophoretic mobility-shift assays
(EMSAs) (Fig. 3c).
It has recently been shown that maize mutants defective for
ABPH1, a type-A ARR homologue, have defects in phyllotaxis and
meristem size regulation6
. In contrast, neither Arabidopsis plants
lacking individual type-A ARR genes nor plants overexpressing
ARR5, ARR6, ARR7 or ARR15 have obvious phenotypes (data not
shown, and refs. 4, 5). We therefore constructed arr7 arr15 double
mutants, because they are closely related and both are expressed in
meristematic tissue. However, the double mutant combination
caused female gametophytic lethality, precluding analysis of the
progeny. To reduce redundancy outside the ARR7/ARR15 pair, we
then extended our analysis to arr3 arr4 arr5 arr6 arr7 arr8 arr9
septuple mutants. These plants were viable, although they had defects
in phyllotaxis and organ initiation (Fig. 4a, b), indicating that the
redundant function of ARR7 and ARR15 might be sufficient for
meristem maintenance. WUS expression in the inflorescence meristem
of septuple mutants was decreased (Fig. 2f, l), indicating that, in
addition to the negative regulatory activity of ARR7 on WUS, there
might be positive effects on WUS expression by other type-A ARR
genes.
As an alternative to exploring ARR7 function, we constructed
alleles that either mimic the active, phosphorylated state or the
inactive non-phosphorylated state of ARR7 by mutating aspartate
85 to glutamate or asparagine, respectively28
. Whereas ubiquitous
overexpression of the dominant-negative form (Asp 85 ! Asn) did
not cause any morphological defects, the constitutively active form
(Asp 85 ! Glu) had severe effects on the function of the shoot apical
meristem. In some of the transgenic seedlings meristems were
arrested for several days after expansion of the cotyledons, resulting
in an almost complete block of organ formation, very similar to that
Figure 4 | Phenotypes of type-A ARR mutant plants. a, Wild type. b, arr3
arr4 arr5 arr6 arr7 arr8 arr9 septuple mutant; note irregular organ
positioning indicated by arrowheads. c–f, Activity of the shoot apical
meristem is arrested in 35S::ARR7 (Asp 85 ! Glu) plants similar to wus
mutants 5 days after sowing. Scale bars, 1 mm for seedlings and 100 mm for
meristem insets unless otherwise noted. c, 35S::ARR7 (Asp 85 ! Glu) plant
withwild-type morphology. Scale bar, 200 mm. d, 35S::ARR7 (Asp 85 ! Glu)
plant with intermediate phenotype. e, 35S::ARR7 (Asp 85 ! Glu) plant with
strong phenotype. f, wus mutant seedling. g, 35S::ARR7 (Asp 85 ! Glu)
seedling shortly after recovery of meristematic activity. Arrowheads indicate
duplicated meristems. h, Phenotype of an adult 35S::ARR7 (Asp 85 ! Glu)
plant after recovery. Note duplicated rosettes. Arrowheads indicate irregular
side-shoot positions.
LETTERS NATURE|Vol 438|22/29 December 2005
1174
© 2005 Nature Publishing Group
observed in wus mutants (Fig. 4c–f). Subsequently, shoot apical
meristems recovered proliferative activity, but often split into two or
three independent meristems (Fig. 4g), giving rise to multiple
primary shoots. Similarly to the abph1 mutant of maize6
, these
shoots had defects in phyllotaxis (Fig. 4h) and flower formation; in
addition they did not produce seeds.
Our results show that direct interaction between the CLV/WUS
network and the cytokinin signalling circuitry is required for proper
meristem function. Together with the recently uncovered role of the
type-A response regulator ABPH1 in maize6
, our findings are a first
step towards understanding how global hormonal signals are integrated
with local transcriptional inputs in the regulation of cell
behaviour at the shoot apical meristem.
METHODS
Plant material and treatments. Plants were of Columbia background and
grown at 23 8C in continuous light. Inductions with ethanol were performed
at 20 8C by watering with 1% ethanol. For inductions with dexamethasone, tissue
was incubated in 15 mM dexamethasone and 0.015% Silwet L-77. Cycloheximide
was used at 10 mM. For 6-benzylaminopurine treatments, tissue was incubated in
1 mM 6-benzylaminopurine and 0.1% DMSO. The Columbia wus allele corresponds
to wus-4 (provided by Martin Hobe and Ru¨diger Simon); details on the
arr3 arr4 arr5 arr6 arr7 arr8 arr9 septuple mutant are available in Supplementary
Information.
Microarray experiments. Affymetrix Ath1 microarrays were hybridized as
described29
in duplicates using RNA from pools of 20 plants for each replicate.
Expression estimates were calculated by gcRMA (ref. 30) and statistical testing
for differential expression was performed with LogitT (ref. 20).
Quantitative real-time RT–PCR. qRT–PCR was performed as described29
with
the use of either SYBR-green or Taq-Man probes (Fig. 1d). Experiments were
performed in triplicates from RNA of pooled tissue. Amplification of TUBULIN
served as control. Oligonucleotides are listed in Supplementary Table 2.
In situ hybridization. In situ hybridization was performed in accordance with
standard protocols, with the addition of 10% poly(vinyl alcohol) (molecular
mass 70–100 kDa) to the staining solution.
ChIP. Genomic fragments were analysed by real-time qRT–PCR in triplicates.
Unrelated sequences in the experimental tissue and ARR7 sequences in leaves,
where WUS should not be present, served as controls. A detailed protocol is
available as Supplementary Information.
EMSA. EMSA was performed as described in ref. 17.
Transgenes. Complementary DNAs flanked by the AlcA promoter and the OCS
terminator were inserted into a pMLBART-derived binary vector, which
harbours a 35S::AlcR cassette18
. Constitutive overexpression constructs were
made in pMLBARTor pART27 binary vectors using a 35S promoter and an OCS
terminator.
Received 4 May; accepted 30 September 2005.
1. Schoof, H. et al. The stem cell population of Arabidopsis shoot meristems in
maintained by a regulatory loop between the CLAVATA and WUSCHEL genes.
Cell 100, 635–-644 (2000).
2. Brand, U., Fletcher, J. C., Hobe, M., Meyerowitz, E. M. & Simon, R. Dependence
of stem cell fate in Arabidopsis on a feedback loop regulated by CLV3 activity.
Science 289, 617–-619 (2000).
3. Leyser, O. Regulation of shoot branching by auxin. Trends Plant Sci. 8, 541–-545
(2003).
4. Kiba, T. et al. The type-A response regulator, ARR15, acts as a negative
regulator in the cytokinin-mediated signal transduction in Arabidopsis thaliana.
Plant Cell Physiol. 44, 868–-874 (2003).
5. To, J. P. et al. Type-A Arabidopsis response regulators are partially redundant
negative regulators of cytokinin signalling. Plant Cell 16, 658–-671 (2004).
6. Giulini, A., Wang, J. & Jackson, D. Control of phyllotaxy by the cytokinininducible
response regulator homologue ABPHYL1. Nature 430, 1031–-1034
(2004).
7. Laux, T., Mayer, K. F., Berger, J. & Jurgens, G. The WUSCHEL gene is required
for shoot and floral meristem integrity in Arabidopsis. Development 122, 87–-96
(1996).
8. Long, J. A., Moan, E. I., Medford, J. I. & Barton, M. K. A member of the
KNOTTED class of homeodomain proteins encoded by the STM gene
Arabidopsis. Nature 379, 66–-69 (1996).
9. Clark, S. E., Running, M. P. & Meyerowitz, E. M. CLAVATA1, a regulator of
meristem and flower development in Arabidopsis. Development 119, 397–-418
(1993).
10. Clark, S. E., Running, M. P. & Meyerowitz, E. M. CLAVATA3 is a specific
regulator of shoot and floral meristem development affecting the same
processes a CLAVATA1. Development 121, 2057–-2067 (1995).
11. Kayes, J. M. & Clark, S. E. CLAVATA2, a regulator of meristem and organ
development in Arabidopsis. Development 125, 3843–-3851 (1998).
12. Mayer, K. F. X. et al. Role of WUSCHEL in regulating stem cell fate in the
Arabidopsis shoot meristem. Cell 95, 805–-815 (1998).
13. Fletcher, J. C., Brand, U., Running, M. P., Simon, R. & Meyerowitz, E. M.
Signaling of cell fate decisions by CLAVATA3 in Arabidopsis shoot meristems.
Science 283, 1911–-1914 (1999).
14. Rojo, E., Sharma, V. K., Kovaleva, V., Raikhel, N. V. & Fletcher, J. C. CLV3 is
localized to the extracellular space, where it activates the Arabidopsis
CLAVATA stem cell signalling pathway. Plant Cell 14, 969–-977 (2002).
15. Clark, S. E., Williams, R. W. & Meyerowitz, E. M. The CLAVATA1 gene encodes
a putative receptor kinase that controls shoot and floral meristem size in
Arabidopsis. Cell 89, 575–-585 (1997).
16. Jeong, S., Trotochaud, A. E. & Clark, S. E. The Arabidopsis CLAVATA2 gene
encodes a receptor-like protein required for the stability of the CLAVATA1
receptor-like kinase. Plant Cell 11, 1925–-1934 (1999).
17. Lohmann, J. U. et al. A molecular link between stem cell regulation and floral
patterning in Arabidopsis. Cell 105, 793–-803 (2001).
18. Roslan, H. A. et al. Characterization of the ethanol-inducible alc geneexpression
system in Arabidopsis thaliana. Plant J. 28, 225–-235 (2001).
19. Lenhard, M., Bohnert, A., Ju¨rgens, G. & Laux, T. Termination of stem cell
maintenance in Arabidopsis floral meristems by interactions between
WUSCHEL and AGAMOUS. Cell 105, 805–-814 (2001).
20. Lemon, W. J., Liyanarachchi, S. & You, M. A high performance test of differential
gene expression for oligonucleotide arrays. Genome Biol. 4, R67 (2003).
21. D’Agostino, I. B., Deruere, J. & Kieber, J. J. Characterization of the response of
the Arabidopsis response regulator gene family to cytokinin. Plant Physiol. 124,
1706–-1717 (2000).
22. D’Agostino, I. B. & Kieber, J. J. Phosphorelay signal transduction: the emerging
family of plant response regulators. Trends Biochem. Sci. 24, 452–-456 (1999).
23. Skoog, F. & Miller, C. O. Chemical regulation of growth and organ formation in
plant tissues cultured in vitro. Symp. Soc. Exp. Biol. 54, 118–-130 (1957).
24. Hwang, I. & Sheen, J. Two-component circuitry in Arabidopsis cytokinin signal
transduction. Nature 413, 383–-389 (2001).
25. Schmid, M. et al. A gene expression map of Arabidopsis thaliana development.
Nature Genet. 37, 501–-506 (2005).
26. Brand, U., Grunewald, M., Hobe, M. & Simon, R. Regulation of CLV3 expression
by two homeobox genes in Arabidopsis. Plant Physiol. 129, 565–-575 (2002).
27. Lenhard, M., Jurgens, G. & Laux, T. The WUSCHEL and
SHOOTMERISTEMLESS genes fulfil complementary roles in Arabidopsis shoot
meristem regulation. Development 129, 3195–-3206 (2002).
28. Hass, C. et al. The response regulator 2 mediates ethylene signalling and
hormone signal integration in Arabidopsis. EMBO J. 23, 3290–-3302 (2004).
29. Schmid, M. et al. Dissection of floral induction pathways using global
expression analysis. Development 130, 6001–-6012 (2003).
30. Wu, Z., Irizarry, R. A., Gentleman, R., Murillo, F. M. & Spencer, F. A. A model
based background adjustment for oligonucleotide expression arrays. Working
Paper 1 (Dept of Biostatistics Working Papers, Johns Hopkins Univ., Baltimore,
Maryland, 2004).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements We thank A. Greenland for providing the AlcA system,
J. Palatnik for sharing unpublished results, R. Schwab for establishing the in situ
protocol, R. Chen for preparing the WUS antiserum, K. Harter and D. Weigel for
discussion, and K. Bomblies, I. Lohmann, M. Schmid, J. Palatnik and D. Weigel
for reading the manuscript critically. This work was supported by a Career
Development Award of the International Human Frontier Science Program
(HFSP) Organization (J.U.L.), a Ph.D. fellowship of the Cusanuswerk (W.B.),
grants from the NSF and the NIH (J.J.K.) and the Max Planck Society (J.U.L).
Author Contributions A.L. performed in situ hybridizations and qRT–PCRs,
constructed reporter genes, the mutated alleles of ARR7 and performed electron
microscopy; J.P.C.T. and J.J.K. generated and analysed the arr double and
septuple mutants; W.B. performed the ChIP experiments; S.S. generated
constructs and transgenic lines of ARR genes; A.K. generated AlcA::CLV3 plants;
M.D. performed qRT–PCRs; and J.U.L. carried out the microarray experiment
and analysis, performed gel-shifts and wrote the paper. All authors discussed
the results and commented on the manuscript.
Author Information Microarray data have been deposited at ArrayExpress
(http://www.ebi.ac.uk/arrayexpress/) under accession number E-MEXP-432.
Reprints and permissions information is available at npg.nature.com/
reprintsandpermissions. The authors declare no competing financial interests.
Correspondence and requests for materials should be addressed to J.U.L.
(jlohmann@tuebingen.mpg.de).
NATURE|Vol 438|22/29 December 2005 LETTERS
1175