equilibrium sugar concentrations due to treatment order in the groups with three bats [Fig. 4,
B, C, E, and F; lm in R (linear model function in
R language), t = 0.11, P = 0.92; overall model
F3, 82 = 13. 28, P < 0.001, adjusted R2 = 0.30].
Our studies of the dynamic interactions of
bats and flowers show that the evolution of
nectars with lower sugar concentration can be
driven by pollinators acting as economic decision-makers subject to the principles of psychophysics
and reinforcement. If pollinators simply maximized energetic gains through linear-value
encoding, they would not exert stabilizing selection pressure on the flowers, and an intermediate concentration would not be selected
for (Fig. 3).
Our approach of combining field experiments,
simulations, and confirmatory laboratory studies
allowed us to elucidate an intricate evolutionary
narrative. The dynamic interaction between nectar volume, sugar concentration, and psychophysics would have made predicting the direction
of selection from analyses based on only a single
reward dimension impossible. The competition
among bats that determines the supply/demand
ratio provided an additional layer of causal complexity (figs. S2 and S3). These results demonstrate the power of iterating between simulation
and experimentation and suggest a plausible
account of the transition to producing nectars
with low sugar concentrations in bat-pollinated
plants as well as in other species that undergo
evolutionary shifts between different pollinators
( 12, 17–20).
We replicated the natural situation most closely when we reduced the supply/demand ratio
by increasing the number of consumers exploiting limited resources. The increase in demand
reduced median food portions available at flowers
from 60 to 6 mL (Fig. 2B), but weakly affected
sugar concentration (Fig. 2A, 30% versus 24%).
When overall value is the product of multiple reward dimensions, proportional magnitudes attain importance. Since perceived differences are
stronger at smaller physical magnitudes, discrimination along the volume dimension could
take priority over the concentration dimension.
Therefore, as predicted by our model (Fig. 2C),
bats favored increases in volume instead of concentration, shifting the balance toward more dilute nectars. Similar processes are likely to
affect the behavior of invertebrate pollinators,
where the different shape of the psychometric
function for concentration (2) and the preference
for nectars with higher sugar concentrations
even at the cost of profitability, ( 21) presumably
cause selection for more concentrated nectars.
For vertebrate and invertebrate pollinators,
body size correlates positively with nectar production rates and negatively with sugar concentration ( 5).
In contrast to some psychological models of
economic choice that assume nonlinearity in
utility ( 15), our approach is based directly on
physiological processes underlying proportional
(Fechnerian) reward evaluation. When multi-
ple perceptual dimensions determine value, a
trade-off ( 22) situation may arise as does be-
tween nectar volume ( 13) and sugar concentration
(2, 10). These effects of proportional psycho-
physics on reward evaluation are of a general
nature and should be applicable to other choice
situations.
REFERENCES AND NOTES
1. F. Roces, Y. Winter, O. von Helversen, in Animal-Plant
Interactions in Tropical Environments, W. Barthlott,
C. M. Naumann, K. Schmidt-Loske, K.-L. Schuchmann, Eds.
(Zoologisches Forschungsinstitut und Museum Alexander
Koenig, Bonn, 1993), pp. 159–165.
2. V. Nachev, K. P. Stich, Y. Winter, PLOS ONE 8, e74144 (2013).
3. M. W. Roberts, Anim. Behav. 52, 361–370 (1996).
4. A. B. Bolten, P. Feinsinger, Biotropica 10, 307–309 (1978).
5. G. H. Pyke, N. M. Waser, Biotropica 13, 260–270 (1981).
6. S. W. Nicolson, R. W. Thornburg, in Nectaries and Nectar,
S. W. Nicolson, M. Nepi, E. Pacini, Eds. (Springer, Dordrecht,
Netherlands, 2007), pp. 215–263.
7. J. M. Ruhlmann, B. W. Kram, C. J. Carter, J. Exp. Bot. 61,
395–404 (2010).
8. I. W. Lin et al., Nature 508, 546–549 (2014).
9. M. Perret, A. Chautems, R. Spichiger, M. Peixoto, V. Savolainen,
Ann. Bot. (Lond.) 87, 267–273 (2001).
10. V. Nachev, Y. Winter, Anim. Cogn. 15, 393–400 (2012).
11. M. Tschapka, O. von Helversen, J. Trop. Ecol. 23, 385–395
(2007).
12. D. W. Schemske, H. D. Bradshaw Jr., Proc. Natl. Acad. Sci. U.S.A.
96, 11910–11915 (1999).
13. U. Toelch, Y. Winter, J. Comp. Physiol. A Neuroethol. Sens.
Neural Behav. Physiol. 193, 265–269 (2007).
14. A. Kacelnik, F. Brito e Abreu, J. Theor. Biol. 194, 289–298 (1998).
15. D. Kahneman, A. Tversky, Econometrica 47, 263–291 (1979).
16. S. Shafir, A. Bechar, E. U. Weber, Plant Syst. Evol. 238,
195–209 (2003).
17. R. J. Gegear, J. G. Burns, Am. Nat. 170, 551–566 (2007).
18. J. D. Thomson, P. Wilson, Int. J. Plant Sci. 169, 23–38 (2008).
19. S. De Witt Smith, New Phytol. 188, 354–363 (2010).
20. S. Martén-Rodríguez, C. B. Fenster, I. Agnarsson, L. E. Skog,
E. A. Zimmer, New Phytol. 188, 403–417 (2010).
21. J. Cnaani, J. D. Thomson, D. R. Papaj, Ethology 112, 278–285
(2006).
22. R. J. Mitchell, D. C. Paton, Oecologia 83, 238–246 (1990).
ACKNOWLEDGMENTS
We thank A. Schatz for programming support, K. Reinhold for
help with the selection algorithms, and S. Shafir and K. Akre for
comments on the manuscript. Financial support was provided
by the Volkswagen Foundation, Bat Conservation International,
and the Deutsche Forschungsgemeinschaft (NeuroCure EXC-
257, CITEC EXC-277). This work benefited greatly from the
support provided by the Organization for Tropical Studies
(OTS) and its La Selva Biological Field Station staff. Data and
code are available at doi.org/10.5281/zenodo.164615 and doi.
org/10.5281/zenodo.164617. Y. W., V.N., A.K., and A.B.
conceived the study. V.N., Y. W., A.K., and A.B. designed the
virtual pollination ecology algorithms. Y. W. developed the radio-frequency ID automated nectar-feeder system. V.N. performed
the field experiments. C. W., Y. W., and V.N. wrote the code for
the simulations. K.P.S. performed the laboratory experiments.
V.N. performed the simulations and analyzed all data. V.N.,
Y. W., and A.K. wrote the paper with help from A.B.
SUPPLEMENTARY MATERIALS
www.sciencemag.org/content/355/6320/75/suppl/DC1
Materials and Methods
Figs. S1 to S6
References ( 23–33)
1 July 2016; accepted 22 November 2016
10.1126/science.aah4219
CANCER THERAPY
Rb1 and Trp53 cooperate to suppress
prostate cancer lineage plasticity,
metastasis, and antiandrogen resistance
Sheng Yu Ku,1 Spencer Rosario,1 Yanqing Wang,1 Ping Mu,2 Mukund Seshadri,1
Zachary W. Goodrich,1 Maxwell M. Goodrich,1 David P. Labbé, 3, 4
Eduardo Cortes Gomez, 5 Jianmin Wang, 5 Henry W. Long, 3, 4 Bo Xu, 6 Myles Brown, 3, 4
Massimo Loda, 4, 7, 8, 9 Charles L. Sawyers,2, 10 Leigh Ellis,1† David W. Goodrich1†
Prostate cancer relapsing from antiandrogen therapies can exhibit variant histology with altered
lineage marker expression, suggesting that lineage plasticity facilitates therapeutic resistance.
The mechanisms underlying prostate cancer lineage plasticity are incompletely understood.
Studying mouse models, we demonstrate that Rb1 loss facilitates lineage plasticity and
metastasis of prostate adenocarcinoma initiated by Pten mutation. Additional loss of Trp53
causes resistance to antiandrogen therapy. Gene expression profiling indicates that mouse
tumors resemble human prostate cancer neuroendocrine variants; both mouse and human
tumors exhibit increased expression of epigenetic reprogramming factors such as Ezh2 and
Sox2. Clinically relevant Ezh2 inhibitors restore androgen receptor expression and sensitivity to
antiandrogen therapy. These findings uncover genetic mutations that enable prostate cancer
progression; identify mouse models for studying prostate cancer lineage plasticity; and suggest
an epigenetic approach for extending clinical responses to antiandrogen therapy.
As molecularly targeted cancer therapy im- proves, lineage plasticity is increasingly appreciated as a potential mechanism under- lying therapeutic resistance. Lineage plas- ticity facilitates conversion of a cancer cell that is dependent on the therapeutic target to ne that is indifferent to its function. For exam- ple, relapse of EGFR (epidermal growth factor eceptor) mutant lung adenocarcinomas after EGFR-targeted therapy is associated with the
78 6 JANUARY 2017 • VOL 355 ISSUE 6320 sciencemag.org SCIENCE
RESEARCH | REPORTS
