Endocrine Disrupters and Flavonoid Signalling Are Ducks Impressed By

brief communications

4 Sexual selection when fully unwound to 42.5 cm. between these organisms .In S. meliloti, the Like other stifftail ducks, lake ducks are constitutively expressed product of the nodD Are ducks impressed promiscuous and boisterous in their sexual genes acts as a positive regulator which, activity4. One explanation for the evolution upon recognizing agonistic phytochemicals by drakes’ display? of such a long penis might be selection for produced by alfalfa, activates transcription urprisingly few birds have penises, but its competitive advantage — for instance, of the common nodulation (nod) genes among those that do, the Argentine groups of drakes might display their everted needed to initiate symbiosis4,5. Whereas Slake duck (Oxyura vittata) tops the bill penises to attract females. Success in sperm some flavonoids, such as luteolin and — the penis of this small stifftail duck from competition5 may also be a factor. The base apigenin, stimulate nodD activation, others South America is shaped like a corkscrew of the lake-duck penis is covered with (such as chrysin) can antagonize the induc- and, at almost half a metre long, is the spines, yet the tip is soft and brush-like. tion of nod genes5–7. Chemicals in the en- largest of any bird measured so far. Factors Before ejaculation, drakes probably use their vironment that mimic or interfere with this responsible for the evolution of this remark- penises like bottle-brushes to remove sperm phytochemical signalling to S. meliloti may able organ could include runaway selection, stored in the oviduct by the female’s previ- confound recognition crucial to symbiosis. whereby drakes with longer penises gain ous consort. The larger the bottle-brush, the As luteolin and other flavonoids have dominance and copulate with more females, more effective it will be, clearing the way for been proposed to mimic oestrogen sig- or preference by females for drakes with the drake to father the ducklings. nalling in mammalian cells2 and a possible longer and more decorated penises. Many questions remain unanswered. homology has been identified between This small duck (which typically weighs How much of his penis does the drake NodD and the oestrogen receptor ER- around 640 g) was previously reported to actually insert, and does the anatomy of the (ref. 8), we investigated whether EDCs that have a 20-cm penis1,2, which is about half of female’s oviducts make them unusually diffi- disrupt oestrogen signalling could also inter- the drake’s total body length and is compa- cult to inseminate? The Argentine lake duck fere with symbiotic signalling3. We tested rable to the penis of an ostrich (Struthio offers a sizeable opportunity to study sexual the effects in S. meliloti of different com- camelus)3. A new voucher specimen from selection and sperm competition in birds. pounds, added in the presence of luteolin, Argentina’s Córdoba province (Fig. 1) Kevin G. McCracken*†, Robert E. Wilson*, on activation and transcription of a nod indicates, however, that the penis size of this Pamela J. McCracken†, Kevin P. Johnson‡ reporter gene using a -galactosidase assay9. species was underestimated. We measured *Institute of Arctic Biology and Department of The inhibitory flavonoids chrysin and the length of the unwound, everted penis of Biology and Wildlife, University of Alaska genistein7,10 were among the most potent an aroused O. vittata drake and found that Fairbanks, Fairbanks, Alaska 99775, USA phytochemical–NodD signalling inhibitors, it was 32.5 cm long when hanging under e-mail: [email protected] as were the organochlorine pesticides the force of gravity, but that it could stretch †University of Alaska Museum, Fairbanks, pentachlorophenol and methyl parathion, Alaska 99775, USA which each reduced nod expression by up ‡Illinois Natural History Survey, Champaign, to 90% even in the presence of luteolin Illinois 61820, USA (Fig. 1a) or apigenin (results not shown). 1. McCracken, K. G. Auk 117, 820–825 (2000). Half-maximal inhibitory concentrations 2. Mead, C. Nature 407, 461 (2000). (IC50) were comparable for the inhibitory 3. Briskie, J. V. & Montgomerie, R. J. Avian Biol. 28, 73–86 (1997). 4. Johnsgard, P. A. & Carbonell, M. Ruddy Ducks and Other phytochemicals and the EDCs tested Stifftails: Their Behavior and Biology (Univ. Oklahoma Press, (Fig. 1a). Even nanomolar concentrations of Norman, 1996). some pesticides significantly inhibited phy- 5. Birkhead, T. R. & Møller, A. P. Sperm Competition in Birds: tochemical–NodD signalling (Fig. 1a). The Evolutionary Causes and Consequences (Academic, London, 1992). insecticide dichlorodiphenyltrichloroethane (DDT) and its metabolites inhibited nod expression by an average of 40% (see, for example, Fig. 1b); other EDCs3, including Nitrogen fixation the plasticizer bisphenol A (Fig. 1a) and the herbicides 2,4-D and 2,4,5-T (results Endocrine disrupters and not shown), reduced expression by 66%, flavonoid signalling 32% and 37%, respectively. Endosulfan, methoxychlor, aldrin, dieldrin, dursban, itrogen fixation is a symbiotic process vinclozolin and diazinon had no effect on initiated by chemical signals from symbiotic signalling (results not shown). Nlegumes that are recognized by soil The inhibitory EDCs are all planar bacteria. Here we show that some phenolic compounds, whereas those that endocrine-disrupting chemicals (EDCs)1–3, had no effect in our assay are non-planar and so called because of their effect on hor- have no free hydroxyl group. Inhibition of mone-signalling pathways in animal cells, transcription regulated by the luteolin–NodD also interfere with the symbiotic signalling pathway can be overcome by increasing the that leads to nitrogen fixation. Our results concentration of luteolin (Fig. 1b). raise the possibility that these phytochemi- We grew alfalfa seedlings inoculated with cally activated pathways may have features the bacterial reporter strain on medium con- in common with hormonal signalling in taining X-gal11,12; NodD-initiated gene tran- vertebrates, thereby extending the biologi- scription produces -galactosidase which, in cal and ecological impact of EDCs. the presence of X-gal, generates a blue prod- Initiation and maintenance of nitrogen- uct11. Figure 1c–f shows that -galactosidase Figure 1 Photograph of the 42.5-cm penis of a male Argentine fixing symbiosis between rhizobial bacteria is expressed along the entire root system in lake duck (Oxyura vittata) taken near Arroyo Chucul, Córdoba, (Sinorhizobium meliloti) and alfalfa (Medica- the alfalfa control, whereas blue product is Argentina on 30 April 2001. go sativa) requires an exchange of signals almost absent in roots treated with chrysin,

Japan during glacial periods, whereas a c higher sea levels during interglacial 100 periods isolated these regions and warmer 80 temperatures restricted temperate taxa to disjunct refuges. However, palaeovegeta- 60 tion data from east Asia2–6 show that 40 d temperate forests were considerably less induction (%) extensive than today during the Last nod 20 Glacial Maximum, calling into question the 0 coalescence of tree populations required by 10–10 10–9 10–8 10–7 10–6 10–5 10–4 the hypothesis of Qian and Ricklefs1. [Chemical] (log) + 1 µM luteolin Pollen data2–6 from 14C-dated sediment b 1,500 e cores, processed using a standard interpre- 7 1,250 tive technique , record the changes in vegetation between the Last Glacial Maximum 1,000 (Fig. 1a) and the present (Fig. 1b). These 750 data are derived from work carried out by Chinese and Japanese palynologists during Miller units 500 f the two decades since the reconstruction by 250 CLIMAP cited in ref. 1. The data indicate 0 that temperate forests were restricted in 10–7.0 10–6.5 10–6.0 10–5.5 10–5.0 10–4.5 10–4.0 distribution during the Last Glacial Maxi- [Luteolin] (log) + 50 µM chemical mum, a finding that is contrary to the inference by Qian and Ricklefs1. The forests Figure 1 Endocrine-disrupting chemicals (EDCs) alter symbiotic signalling in vitro and in vivo. a, Phytochemicals (chrysin, orange) or were displaced either by boreal and mixed EDCs (methyl parathion, green; pentachlorophenol, blue; bisphenol A, red) at 50 nM to 50 M were added to Sinorhizobium meliloti forests lacking most of the temperate forest 1021pRmM57 (ref. 9; provided by S. R. Long) treated with 1 M luteolin. Luteolin alone was set as 100% induction10; luteolin plus taxa, or by treeless vegetation (steppe and 7 7 chemical was compared to 100% induction. IC50 values (molar): methyl parathion, 4.3 10 ; pentachlorophenol, 9.9 10 ; chrysin, desert, with pollen assemblages dominated 7.0107; bisphenol A, 1.7105. b, 50 M o,p-DDT (purple) or pentachlorophenol (blue) caused 43% and 90% inhibition, by Artemisia, Chenopodiaceae and grass- respectively, of nod activation (measured as Miller units of -galactosidase activity). Increasing the luteolin concentration (100 nM to es), over much of their present range. 50 M) restores nod activation to normal. c–f, Alfalfa seedlings were grown on buffered nodulation medium plus X-gal11,12 and Low levels of precipitation, compound- inoculated with dilute12 S. meliloti 1021pRmM57 treated with vehicle (dimethylsulphoxide, c), chrysin (d), methyl parathion (e) or ed by the low water-use efficiency of pentachlorophenol (f) at 50 M. NodC–lacZ expression is visualized as blue product. Further details are available from the authors. C3 photosynthesis at atmospheric CO2 concentrations during the Last Glacial pentachlorophenol or methyl parathion, 977–980 (1986). Maximum (less than 200 p.p.m.; refs 7, 8), except between the tip and the first emerging 6. Djordjevic, M. A. et al. EMBO J. 6, 1173–1179 (1987). confined temperate forests to medium 7. Firmin, J. L. et al. Nature 324, 90–93 (1986). root hairs, where luteolin concentration is 8. Gyorgypal, Z. & Kondorosi, A. Mol. Gen. Genet. 226, elevations in northern China, whereas low 6,10 highest . This effect is similar to that shown 337–340 (1991). temperatures caused mixed forests (mix- in Fig. 1b, where inhibition is overcome by 9. Mulligan, J. T. & Long, S. R. Proc. Natl Acad. Sci. USA 82, tures of boreal conifers with only the most increasing concentrations of luteolin. 6609–6613 (1985). cold-tolerant temperate taxa) to dominate 10.Peters, N. K. & Long, S. R. Plant Physiol. 88, 396–400 (1988). 2–6,8 Nitrogen fixation is controlled by both 11.Redmond, J. W. et al. Nature 323, 632–635 (1986). in much of Japan and southern China . agonistic and antagonistic phytochemical 12.Gage, D. J., Bobo, T. & Long, S. R. J. Bacteriol. 178, To provide a spatially continuous signalling to Rhizobium5,13. As with the 7159–7166 (1996). vegetation map (Fig. 1c) that is physically vertebrate endocrine system, plant–bacterial 13.Baker, M. E. Adv. Exp. Med. Biol. 439, 249–276 (1998). consistent with climate forcing for the Last symbiosis relies on recognition of specific Glacial Maximum, we used a model of agonists and antagonists for gene regula- climate in the last glacial maximum (based tion2,10,13. We have shown that some pesti- on known changes in insolation, atmos- cides and planar phenolic EDCs can interfere COMMUNICATIONS ARISING pheric composition and physiography)9 as with plant–Rhizobium signalling and nitro- Palaeovegetation input to an equilibrium biogeography gen-fixing symbiosis, thereby exposing a pre- model, BIOME4 (Fig. 1d)10. The large-scale viously unrecognized similarity to the effects Diversity of temperate features of this simulation are supported by of EDCs on vertebrate endocrine signalling. plants in east Asia pollen data (Fig. 1a). Temperate forests are Jennifer E. Fox*†, Marta Starcevic*, shown as confined to a relatively narrow Kelvin Y. Kow*, Matthew E. Burow*‡, he exceptionally broad species diversi- belt at low elevations, whereas non-forest John A. McLachlan*‡ ty of vascular plant genera in east vegetation is shown to be greatly extended, *Environmental Endocrinology Laboratory, Center TAsian temperate forests, compared as the pollen data indicate. On the other for Bioenvironmental Research at Tulane and Xavier with their sister taxa in North America, has hand, a band of temperate deciduous forest Universities, †Molecular and Cellular Biology been attributed to the greater climatic is shown extending across the East China Program, Tulane University, and ‡Department of diversity of east Asia, combined with Sea, consistent both with Qian and Rick- Pharmacology, Tulane University Medical School, opportunities for allopatric speciation lefs’ hypothesis1 and with pollen evidence New Orleans, Louisiana 70112, USA afforded by repeated fragmentation and for the existence of an offshore glacial e-mail: [email protected] coalescence of populations through Late refuge for temperate deciduous trees. 1 1. Tham, D. M., Gardner, C. D. & Haskell, W. L. J. Clin. Cenozoic ice-age cycles . According to Qian Quaternary palaeodata have repeatedly Endocrinol. Metab. 83, 2223–2235 (1998). and Ricklefs1, these opportunities occurred overturned historical biogeographical 2. Collins-Burow, B. M. et al. Nutr. Cancer 38, 229–244 (2000). 3. McLachlan, J. A. Endocrinol. Rev. 22, 319–341 (2001). in east Asia because temperate forests hypotheses that were based on present- 4. Schultze, M. & Kondorosi, A. Annu. Rev. Genet. 32, 33–57 (1998). extended across the continental shelf day taxon distributions. The palaeodata 5. Peters, N. K., Frost, J. W. & Long, S. R. Science 233, to link populations in China, Korea and indicate a more complex history in which