20.1 n&v 203 MH

14/1/05

5:24 pm

Page 204

news and views

James A. Dias is at the Wadsworth Center, New York
State Department of Health, and in the Department
of Biomedical Sciences, State University of New York
at Albany, David Axelrod Institute, 120 New
Scotland Avenue, Albany, New York 12208, USA.
e-mail: james.dias@wadsworth.org
1. Wright, V. C., Schieve, L. A., Reynolds, M. A., Jeng, G.
& Kissin, D. MMWR Surveill. Summ. 53, 1–20 (2004).
2. Fan, Q. R. & Hendrickson, W. A. Nature 433, 269–277 (2005).
3. Kobe, B. & Kajava, A. V. Curr. Opin. Struct. Biol. 11, 725–732
(2001).

8. Lapthorn, A. J. et al. Nature 369, 455–461 (1994).
9. Wu, H., Lustbader, J. W., Liu, Y., Canfield, R. E. & Hendrickson,
W. A. Structure 2, 545–558 (1994).
10. Fox, K. M., Dias, J. A. & Van Roey, P. Mol. Endocrinol. 15,
378–389 (2001).
11. Palczewski, K. et al. Science 289, 739–745 (2000).

4. Hopkins, A. L. & Groom, C. R. Nature Rev. Drug Discov. 1,
727–730 (2002).
5. Dias, J. A. et al. in Vitamins and Hormones (ed. Litwack, G.)
249–322 (Academic, New York, 2002).
6. Smits, G. et al. N. Engl. J. Med. 349, 760–766 (2003).
7. Vasseur, C. et al. N. Engl. J. Med. 349, 753–759 (2003).

Climatology

Will soil amplify climate change?
David Powlson
It had been thought by some that rising atmospheric temperatures would
have no effect on the rate at which carbon is released from the soil. A
study that revisits the data behind this theory now finds otherwise.

K

a
250
Turnover time (yr)

norr et al.1 in this issue (page 298)
claim that rising temperatures
brought about by climate change will
cause microorganisms in the world’s soils
to decompose organic matter more rapidly,
releasing extra carbon dioxide (CO2) and
accelerating climate change. This may seem
unsurprising — a basic property of most
biological processes is that they proceed
faster with rising temperature, provided
that other factors do not become limiting.
However, this basic biological tenet was
challenged in 2000 by Giardina and Ryan2,
who sugges; ted that organic carbon decomposition in soil is not sensitive to temperature. If correct, that would mean that our
understanding of the process is in serious
error and predictions from current models
of soil carbon turnover cannot be trusted.
But Knorr et al. have re-examined the data
used in Giardina and Ryan’s work, and come
to the opposite conclusion. The new analysis
resolves the seeming paradox and suggests
that the positive feedback from soil to
climate might be even greater than is
currently thought.
The world’s soils hold about 1,500109
tonnes (1,500 Gt) of organic carbon; the
atmosphere contains about half this amount
as CO2 (720 Gt), and there is a further 600 Gt
in vegetation3. Thus, relatively small changes
in the flow of carbon into or out of soils
could be significant on a global scale, and
various studies have attempted to quantify
these flows using models of soil carbon
turnover4–6.
Increasing CO2 in the atmosphere can
enhance plant growth7 through ‘CO2 fertilization’, thus removing some of the excess
CO2. For long-lived plants such as trees, part
of this carbon is sequestered for decades or
centuries. With all plants, some organic
carbon is transferred by roots and litter-fall
into soil organic matter, some fractions of
which are so strongly stabilized that they
turn over on timescales of centuries or even
millennia8. This carbon sequestration would
tend to slow climate change9, so management practices such as re-forestation are
considered as mitigation measures under the

200
150
100
50
0
5

10
15
20
25
Mean annual temperature (ºC)

b
1.00
Relative turnover time

revealed that the 
-subunit’s carboxy-terminal tail is essential for receptor binding5. For
instance, single mutations of six amino acids
in this tail resulted in a 90% loss of binding
activity. But it was not obvious how these
amino acids contributed to binding. The
structure now reveals that the 
-subunit
plays a neat trick: it rotates its tail by 180 so
that these amino acids become fully conformationally constrained by making extensive
contacts with the receptor at the hormone–
receptor interface.
The structure of the FSH–FSHRHB complex also provides the first structural evidence that GPCRs form dimers,which might
be necessary for their full activity. In the case
of FSHR, the dimer contacts are not made
through the hormone. Rather, two receptors
interact directly with each other, with each
receptor binding one molecule of FSH. Yet
the structure of the dimer presents a conundrum as well. The transmembrane regions of
the two receptors in a dimer are too far apart
to associate,so what drives dimer formation?
One possibility is that other molecules link
receptors through cytoplasmic interactions.
Of course, the next goal is to determine
the structure of the full-length receptor. So
far, the full-length structure of only one
GPCR, rhodopsin, has been determined11,
and this molecule is composed solely of
transmembrane domains. The role of carbohydrate in activating FSHR also needs to be
defined — although receptor binding does
not require carbohydrate, full activity following binding does. And it will be essential
to determine the structure of FSHR without
its ligand, to find out whether rearrangements occur in the receptor upon hormone
binding.
Although it is in the nature of science that
each advance raises new questions, the new
structure2 does further our understanding
of the key–lock mechanism by which glycoprotein hormones work. This advance
defines the specific amino acids that make
contact between hormone and receptor, as
well as those that repulse hormone binding
and thereby provide specificity (Fig. 1).
Moreover, the orientation of FSH with
respect to the main extracellular regions of
the receptor is now resolved, and the structure allows a guess as to how the FSH–
FSHRHB complex is oriented relative to the
extracellular loops of the receptor’s transmembrane domains — goals of researchers
in the field for at least two decades.
■

Fast pool
0.10

Very slow pool
0.01
10

20

30

40

50

Temperature (ºC)

Figure 1 One question but two answers. Is the
rate of decomposition of organic carbon in soil
influenced by temperature? a, Giardina and
Ryan2 used a model with a single carbon pool,
and calculated soil organic carbon turnover
times that were independent of temperature.
b, Knorr et al.1 used a model containing three
carbon pools (fast, intermediate and very slow).
Turnover time for each pool at different
temperatures is shown relative to that at 10 C.
The turnover times all decrease as temperature
increases; that is, decomposition speeds up as
temperature rises.

Kyoto Protocol on climate change. However,
current soil models predict that,in the longer
term, rising temperatures will speed up the
decomposition of organic carbon in soil,
releasing CO2 into the atmosphere in excess
of any carbon sequestered in the soil, and
adding to climate change.
Giardina and Ryan2 challenged this view
with an analysis of published data from
82 experiments across five continents. They

NATURE | VOL 433 | 20 JANUARY 2005 | www.nature.com/nature

204
©2005 Nature Publishing Group

20.1 n&v 203 MH

14/1/05

5:24 pm

Page 205

news and views
used field measurements and laboratory
incubations to calculate the turnover time of
soil organic carbon. Turnover time is the
inverse of the first-order rate constant for the
decomposition process. They calculated that
turnover times were almost independent of
mean annual temperature over the range
5–35 C (Fig. 1a).
Crucially, their analysis used the simplifying assumption that all soil organic carbon
behaves in the same way and can be regarded
as a single pool. Most current models, however, divide soil organic carbon into at least
three pools that differ in their turnover time,
and these models are fairly successful in
simulating long-term changes in the organic
carbon in soil10. The dangers of treating soil
organic carbon as a single pool were highlighted by Davidson et al.11, who showed
that the activity of small, very active pools
can mask the effects of a large, slow pool,
especially in short-term experiments.
Knorr et al.1 have now re-analysed Giardina and Ryan’s results using a soil model
with more than one carbon pool. To create
their model, they used published data in
which soil from a tropical rainforest in Brazil
was incubated for 24 weeks at temperatures
between 15 and 45 C, and the CO2 release
was measured. They then tested how well the
results were predicted by simple models of
decomposition using different numbers of
carbon pools, each with a temperaturesensitive rate constant. To explain the results,
they found it necessary to have two active
pools (one with a much faster turnover time
than the other), and a very slow pool that
was effectively inert during the 24 weeks
of the experiment. The very slow pool was
large, comprising about 95% of total soil
organic carbon.
By applying this model to the incubation
experiments considered by Giardina and
Ryan2, Knorr et al. demonstrate a fascinating
paradox that is particularly marked if a
single-pool model is assumed: because of
the dominance of the large, very slow pool,
short-term incubations of one to two years
can seem to show no sensitivity of decomposition to temperature, even when each pool
does in fact have a built-in temperature
dependence. This is because, as the temperature rises, fast pools turn over even faster
and their effect on the overall result is very
short-lived.
By applying their model to other data
sets, Knorr et al. reveal the perils of drawing
conclusions about the impacts of climate
change from short-term studies such as soilwarming experiments. Such experiments
rarely continue for more than a few years,
and so never provide information on the
response of the large, slower pools that will
dominate feedbacks from soil to atmosphere
over timescales of decades or more.
Several experimental studies12,13 have
seemed to show that decomposition is

insensitive to temperature. The reasons for
this are unclear, but as Davidson et al.11 point
out, the effects of climate change are much
more complex than just an increase in temperature: changes in soil water and nutrient
availability will occur, and these can influence decomposition. The observation of
apparent insensitivity of soil decomposition
to temperature is sometimes termed
acclimatization to suggest a gradual biological adaptation of soil organisms to a higher
temperature. Knorr and colleagues’ analysis1
shows that biological adaptation does not
have to be invoked — such results can simply
be a consequence of using a model that
ignores the dominance of large, slowly
changing pools of organic matter.
As a final twist, Knorr et al. predict that,
over a timescale of decades to centuries, the
dominant slow pools will be more sensitive to
temperature than the faster pools (Fig. 1b),
causing a larger positive feedback in response
to global warming than previously thought.
The feedback suggested by current models
is already very significant; for instance, one
forecast6 is that organic carbon will accumulate in soil as a result of CO2 fertilization until
the middle of the twenty-first century and
then decline as the effect of rising temperature

on decomposition becomes dominant.
Although Giardina and Ryan’s analysis2
now appears incorrect, it was useful in provoking a deeper examination of the temperature sensitivity of the decomposition of soil
organic carbon1,11,14, and has led to a greater
appreciation of the importance of stable
organic carbon fractions in influencing
climate change.
■
David Powlson is in the Agriculture and
Environment Division, Rothamsted Research,
Harpenden, Hertfordshire AL5 2JQ, UK.
e-mail: david.powlson@bbsrc.ac.uk
1. Knorr, W., Prentice, I. C., House, J. I. & Holland, E. A. Nature
433, 298–301 (2005).
2. Giardina, C. P. & Ryan, M. G. Nature 404, 858–861 (2000).
3. Batjes, N. H. Eur. J. Soil Sci. 47, 151–163 (1996).
4. Jenkinson, D. S., Adams, D. E. & Wild, A. Nature 351, 304–306
(1991).
5. Cox, P. M., Betts, R. A., Jones, C. D., Spall, S. A. & Totterdell, I. J.
Nature 408, 184–187 (2000).
6. Jones, C. et al. Glob. Change Biol. 11, 154–166 (2005).
7. Grace, J. & Rayment, M. Nature 404, 819–820 (2000).
8. Jenkinson, D. S. Phil. Trans. R. Soc. Lond. B 329, 361–369 (1990).
9. The Role of Land Carbon Sinks in Mitigating Global Climate
Change Policy doc. 10/01 (Royal Society, London, 2001).
10. Smith, P. et al. Geoderma 81, 153–225 (1997).
11. Davidson, E. A., Trumbore, S. E. & Amundson, R. Nature 408,
789–790 (2000).
12. Oechel, W. C. et al. Nature 406, 978–981 (2000).
13. Luo, Y., Wan, S., Hui, D. & Wallace, L. L. Nature 413, 622–625
(2001).
14. Ågren, G. I. & Bosatta, E. Soil Biol. Biochem. 34, 129–132 (2002).

Evolution

A taste for mimicry
Graeme D. Ruxton and Michael P. Speed
Looking inedible is a great way to deter predators, but the warning
signs must be learnt first. It seems that unpalatable species employ
some unexpected strategies to make the education a quick one.
arwin saw mimicry — strong visual
resemblances between unrelated
species — as an excellent test case for
his theories of natural selection1. The phenomenon continues to exercise evolutionary
biologists today, with the latest salvo coming
from Skelhorn and Rowe2. Writing in Proceedings of the Royal Society, they find that
mimicry can work in an unexpected way to
provide safeguards against predators.
Mimicry generally occurs in two forms,
Batesian and Müllerian. Batesian mimicry is
essentially parasitic: a prey species evolves
to look like a species that is unattractive to
predators (because it is poisonous, for example), and in so doing degrades the effectiveness of the signals used by the inedible species
to warn off predators. By contrast, Müllerian
mimicry involves two unpalatable species,
and is thought to be mutualistic because the
two species share the mortality costs incurred
when naive predators sample them before
learning to avoid the warning signal they both
use3. Butterflies are among the commonest
examples of Müllerian mimicry; a possible
example is shown in Figure 1 (overleaf).

D

NATURE | VOL 433 | 20 JANUARY 2005 | www.nature.com/nature

Müllerian mimicry has become highly
contentious, with a body of mainly theoretical work arguing that systems identified
previously as Müllerian might in fact be
parasitic rather than mutualistic4,5. But this
conclusion will need to be re-examined in
light of the ingenious experiments reported
by Skelhorn and Rowe2. An (often implicit)
assumption of previous work on Müllerian
mimicry has been that the two species are
unappealing to predators because they have
the same defence — the same toxin, say.
However, there is no logical or observational
foundation for this supposition6. Skelhorn
and Rowe now demonstrate that Müllerian
mimicry can provide highly effective protection from predation when the two species
concerned have different defences.
The authors’experiments involved giving
domestic chicks coloured food crumbs
flavoured with aversive chemicals. All the
crumbs were coloured identically, but some
were flavoured with quinine, some with an
equally bitter solution of Bitrex (a preparation designed to discourage people from
biting their nails) and some with a blend of
205

©2005 Nature Publishing Group