No significant relationship exists between seedling relative growth rate

FEC_598.fm Page 122 Thursday, February 7, 2002 8:50 PM

Functional Ecology 2002 16, 122 –127

No significant relationship exists between seedling relative growth rate under nutrient limitation and potential tissue toxicity

Blackwell Science Ltd

J. S. ALMEIDA-CORTEZ* and W. SHIPLEY† Departement de Biologie, Université de Sherbrooke, Sherbrooke, Québec J1K 2K1 Canada

Summary 1. We asked two questions in this study: (i) is there is there a negative correlation between relative growth rate (RGR) among 19 species of herbaceous Asteraceae grown under nutrient limitation and potential tissue toxicity; and (ii) is the degree of toxicity of these species different than those measured in a previous study, in which the same species were grown under non-limiting conditions? 2. In the present study we grew 19 of the same species as previously in hydroponic sand culture for which the growth rate was limited by nutrient supply, and compared RGR and potential tissue toxicity due to secondary compounds. Under such conditions, RGR was reduced by 25%; specific leaf area was reduced by 32%; root : shoot ratios were increased by 320%; leaf nitrogen content was reduced by 69%; and potential toxicity was decreased by 46%. 3. Although nutrient limitation clearly decreased both RGR and potential toxicity, we found no interspecific correlation between RGR and toxicity. 4. We conclude that there was no detectable trade-off between RGR and potential toxicity due to secondary compounds. Key-words: Asteraceae, chemical defence, relative growth rate, RGR, secondary compounds Functional Ecology (2002) 16, 122 –127

Introduction Is there a trade-off in plants between the ability to grow rapidly and the amount of resources allocated to the production of secondary compounds (Herms & Mattson 1992)? Such a hypothesis is appealing, as a limiting resource that is allocated to the production of secondary compounds is necessarily diverted away from primary metabolism. This question has generated a large body of research, but is impossible to answer with current technology, especially when applied in multispecies comparisons. There are over 18 500 secondary plant compounds known or suspected of affecting plant–animal interactions (Harborne 1989), making it impossible in practice to quantitatively measure resource allocation to all secondary compounds that may occur in a collection of species. By concentrating on only a few secondary compounds with known anti-herbivore or anti-pathogen properties, one can quantitatively measure direct resource allocation, but this underestimates, perhaps severely, allocation to

© 2002 British Ecological Society

†Author to whom correspondence should be addressed. Email: [email protected] *Present address: Departamento de Botânica, Universidade Federal de Pernambuco, Rua Prof. Nelson Chaves s/n, Recife PE, CEP 50372-970, Brazil.

defence per se. Even concentrating on broader classes of compounds – for example, tannins or phenolics in general – underestimates allocation and has the added problem that such broad classes of compounds can only be measured quantitatively as rough indices, and undoubtedly include compounds not related to defence. These problems led us, in previous publications (Almeida-Cortez 1998; Almeida-Cortez, Shipley & Arnason 1999), to modify the underlying ecological question of interest. Rather than concentrating on the relationship between growth versus resource allocation to secondary compounds, we concentrated instead on the relationship between growth and the potential of constitutive secondary compounds to confer a defence, irrespective of the amount of resources allocated for such a defence. There are evolutionary reasons for suspecting a relationship between these variables, irrespective of competing resource substrates. When growth is limited by resource supplies, then any loss of tissues is also a loss of a limiting resource and, further, it will take longer for any tissues lost to herbivores to be replaced. One might expect plants to increase defence under such conditions. At an interspecific level, potential relative growth rate (RGR) is positively related to the average soil fertility of the habitats where the species occur (Grime & Hunt 1975; Poorter & Garnier 1999). 122


123 Seedling RGR and tissue toxicity

© 2002 British Ecological Society, Functional Ecology, 16, 122 –127

If species typical of resource-poor environments have evolved greater degrees of chemical defence in order to protect captured resources that are more difficult to replace, then this too would generate a positive interspecific relationship between RGR and the defensive ability of secondary compounds. In a previous study (Almeida-Cortez, Shipley & Arnason 1999) we grew 31 species of herbaceous Asteraceae in non-limiting hydroponic solution. Despite substantial interspecific variation in RGR and tissue toxicity, we found no relationship between seedling RGR and potential tissue toxicity. To compare our species with other publications, we also measured total phenolic concentrations and found a significant, but positive, correlation with RGR. Almeida-Cortez (1998) grew six species of herbaceous Asteraceae hydroponically with differing nutrient availabilities and irradiances. Nutrients and irradiance had significant affects on tissue toxicity, although the responses were speciesspecific. This suggests that a general interspecific tradeoff might exist between growth and the potential of secondary compounds to confer a defence, but only when growth is resource-limited. This is the hypothesis that is addressed in the present paper. Specifically, we ask (i) is there is there a negative correlation between RGR in a collection of 19 species of herbaceous Asteraceae grown under nutrient limitation and potential tissue toxicity; and (ii) is the degree of toxicity of these species different from those measured in the previous study (Almeida-Cortez et al. 1999) in which the same species were grown in non-limiting conditions? It is important to circumscribe this biological hypothesis. Defence can be conferred through morphological and chemical means. We concentrate only on chemical defence. Similarly, secondary compounds can confer defence by changing tissue palatability (through bitter taste or by making proteins less digestible) or by actually interfering with animal metabolism and causing toxicity. We concentrate only on toxicity of secondary compounds and not on deterrence. Finally, even toxic chemical compounds can be constitutive or produced only after induced by some stimulus; we concentrate on constitutive compounds. Secondary compounds in the Asteraceae are relatively well studied and are primarily carbon-based (Malbry & Bohlmann 1977). These plants generally contain sesquiterpene lactones and tri-terpenes, derivatives of caffeic acid, acetylenic compounds (except for those in the tribes Senecioneae and Cichorieae), and essential oils (except in the latex-bearing Cichorieae). They have no true tannins and few alkaloids (none in the species used in this study). The notion of ‘toxicity’ also requires more careful definition. Clearly, toxicity is a property both of the plant and of the animal; compounds, which are lethal to one animal species may have no effect at all on another species. We therefore define ‘potential tissue toxicity’ as the toxicity conferred by a plant tissue on a naive organism that has not evolved any specific

defence; in this sense we are measuring a maximum possible toxicity that can be reduced by herbivore adaptations. Nonetheless, one would want an assay of potential tissue toxicity to indicate the ability of a plant to defend against many generalist herbivores. For this reason we measure potential tissue toxicity using the brine shrimp bioassay (Alkofahi et al. 1989; Arnason, Marles & Aucoin 1991; Meyer et al. 1982). This bioassay is widely used as an initial screening technique in pharmacology, is sensitive to many secondary compounds (Abel-Kader, Omar & Stermitz 1997; He et al. 1997; Lopez et al. 1997; Meyer et al. 1982; Spatafora & Tringali 1997), and is a good predictor of extract toxicity in insects such as southern armyworm, melon aphid, Mexican bean beetle and mosquito larvae (Alkofahi et al. 1989).

Materials and methods     We used 19 herbaceous species from the Asteraceae from six different tribes (Table 1). Taxonomy follows Gleason & Cronquist (1991). These are the same species as used in the study by Almeida-Cortez et al. (1999), and the seed source was the same in 15 of them. This study is restricted to herbaceous species that inhabit open, sunny habitats but with differing soil fertility, such as agricultural fields, meadows, waste places, roadsides and riverbanks. Seeds were collected from wild populations across south-western Quebec during the summer of 1994 and 1996 and stored in paper bags at 4 °C prior to germination. We estimated the germination rates and percentages for each species prior to the experiment so as to better synchronize germination within a 1-week period.

    Within 3 days of germination, seedlings were transplanted individually into 1 l pots filled with silica sand (40 mesh) and moistened immediately with distilled water. The 400 1 l containers were placed randomly in a larger main reservoir of the growth chamber. Plants were supplied with a photosynthetic photon flux density (PPFD) of 450 µmol m–2 s–1 PPFD from fluorescent lamps and incandescent bulbs for 16 h a day for a daily integrated photon flux of 25·9 mol m–2 day–1. The temperature was maintained at 24 °C day and 20 °C night, and the relative humidity was 80%. These conditions were quite similar, but not identical, to those used by Almeida-Cortez et al. (1999). In that experiment, 500 µmol m–2 s–1 PFFD was provided for 16 h each day, providing a daily integrated photon flux of 28·8 mol m–2, while the temperature was maintained at 25 °C day and 20 °C night and the relative humidity was 80%. Five randomly chosen plants per species were harvested after 40 days for use in the bioassay of tissue toxicity. As this required 1 g fresh mass, measurements


124 J. S. AlmeidaCortez & B. Shipley

Table 1. Means of relative growth rate between days 20 and 40 (RGR), potential toxicity of the plant tissues at day 40 of 19 species grown in nutrient-limited conditions Tribe (subtribe)

Species

RGR (g g–1 day–1)

Toxicity (mg–1)

Anthemideae

Achillea millefolium Artemisia vulgaris Chrysanthemum leucanthemum Matricaria matricarioides Tanacetum vulgare Erigeron canadensis Solidago canadensis Cichorium intybus Hieracium aurantiacum Hieracium vulgatum Lactuca canadensis Lapsana communis Leontodon autumnalis Taraxacum officinale Tragopogon pratensis Arctium lappa Arctium minus Bidens cernua Rudbeckia hirta

0·03 0·12 0·17 0·11 0·18 0·14 0·17 0·16 0·07 0·13 0·08 0·12 0·16 0·09 0·08 0·08 0·08 0·11 0·15

0·05 0·03 0·01 0·01 0·05 0·01 0·02 0·01 0·01 0·01 0·01 0·03 0·01 0·01 0·01 0·01 0·01 0·02 0·01

Astereae Cichorieae

Cynareae (Carduinae) Heliantheae (Coreopsidinare) Heliantheae (Helianthinae)

Note that a toxicity value of 0·01 (i.e. LC50 of 100 mg fresh tissues) is the lowest detectable limit.

on individual plants were possible, but tissues from more than one individual were used for Erigeron canadensis and Rudbeckia hirta because of their small size. Another five to 12 randomly chosen plants per species were harvested after 20 days, and a further three to eight randomly chosen plants per species were harvested after 40 days to be used in the growth analysis and leaf nitrogen measurements. Leaf areas were measured with an image analyser, tissues were divided into leaf blades, roots and stems, and dry mass was obtained after drying tissues at 70 °C for at least 48 h.

  

© 2002 British Ecological Society, Functional Ecology, 16, 122–127

The nutrient delivery system consisted of a 200 l external nutrient holding tank filled with a 1/8 full-strength modified Hoagland’s solution. Three times a day the solution was pumped into a main reservoir in the growth chamber in which the pots were housed. The solution was allowed to saturate the silica sand via perforations at the base of each pot. Once saturated, the solution drained out of the pots and the reservoir by gravity into a holding tank. These pots held approximately 300 ml of solution at field capacity. The returning solution was filtered though cheesecloth and activated charcoal (to remove organic molecules) and recirculated into the external holding tank. A nutrient solution was prepared from distilled water and stock standards. Its macronutrient concentration (m) was 0·33 KNO3, 0·25 Ca(NO3)2·4H2O, 0·34 MgSO4·7H2O, 0·17 KH2PO4 and 0·08 (NH4)2SO4; micronutrient concentration (µ) was 1·51 MnSO4. H2O, 0·12 ZnSO4·7H2O, 7·7 H3BO3, 0·01 Na2MoO4·2H2O, 0·06 CuSO4 and 5FeSO4·7H2O with EDTA. Daily nitrate and pH measurements were recorded and pH

was adjusted to 5·5 as needed. The 200 l nutrient solution in the holding tank was completely replaced each week from distilled water and stock standards. Compared to Almeida-Cortez et al. (1999), this nutrient solution was eight times less concentrated.

   Details of the method are given by Arnason et al. (1991) and Almeida-Cortez (1998); only the logic of the test is presented here. Ethanol was used because its intermediate polarity allows most biologically active secondary compounds to be extracted, including phenolics, alkaloids, acetylenes, terpenes and other less common secondary compounds (Liskens & Jackson 1992). However, the degree of extraction will naturally differ for different types of compound. After a general extraction of the fresh tissues in 100% ethanol followed by evaporation, the extractant is diluted again in ethanol at a ratio of 1 ml g–1 fresh mass of plant tissue. One µl of this solution therefore corresponds to extracts from 1 mg fresh tissue. A combination of this solution and ethanol (total volume 100 µl) is then added to test tubes containing 4 ml of a brine solution and 5-dayold brine shrimp in logarithmically increasing doses (0 : 100, 1 : 99, 10 : 90, 100 : 0 µl : µl of solution : ethanol). The proportional mortality after 24 h is regressed against the initial dosage using logistic regression, and the extract concentration (µl ~ mg fresh tissues) needed to obtain 50% mortality (LC50; mg) was calculated. Thus low LC50 values indicate that extracts from less tissue mass were required for 50% mortality, indicating higher toxicity. We therefore use the inverse of this value (1/LC50; mg–1) to measure the potential phytochemical toxicity of the extract, with

(a)

0·15

(b)

0·05

0·10

Toxicity high nutrients

0·15 0·10

0·0

0·0

0·05

RGR high nutrients

0·20

0·25


0·20


0·0

0·05

0·10

0·15

0·20

0·25

0·0

RGR low nutrients

0·05

0·10

0·15

0·20

Toxicity low nutrients

Fig. 1. Values of relative growth rate (RGR, g g–1 day–1) (a) and potential toxicity of tissue extracts (mg–1) (b) of 19 species of herbaceous Asteraceae grown with nutrient limitation (this study) and grown as an eightfold more concentrated nutrient solution. Species belonging to the same tribe (Table 1) have the same plotting symbol. Diagonal = 1 : 1 relation.

Table 2. Means (SD) of relative growth rate (RGR, g g–1 day–1), specific leaf area (SLA, cm–2 g–1), root : shoot ratios (g g–1), leaf nitrogen (% dry weight) and potential toxicity (mg–1) in this study and a study using the same species but cultured with eight times more concentrated nutrient solutions

Variable

Non-limiting nutrients (Almeida-Cortez et al. 1999)

Limiting nutrients (this study)

RGR** SLA* Root : shoot* Leaf nitrogen* Toxicity**

0·159 (0·036) 274·0 (87·4) 0·436 (0·103) 5·3 (0·830) 0·032 (0·033)

0·121 (0·037) 188·0 (39·9) 1·836 (0·687) 1·642 (0·659) 0·017 (0·013)

**P < 0·0001; *P < 0·002.

increasing values indicating increasing potential toxicity of the extract.

  The organic nitrogen (micro-Kjeldahl N) content of dried and ground samples was determined by digesting plant material in sulphuric acid and a mixture of potassium sulphate and selenium oxychloride as a catalyst (Lang 1962), followed by Nesslerization (Middleton 1960). Units are percentage nitrogen content relative to tissue dry weight. When plant material was sufficient, three replicates of each sample were analysed. Some dilutions were done if necessary. Leaves, stems and roots were analysed separately. © 2002 British Ecological Society, Functional Ecology, 16, 122 –127

Results Table 1 gives the mean relative growth rates and mean measurable toxicity (1/LC50; mg–1) in the brine shrimp

test for each of the 19 study species for each harvest date. It is clear that these plants experienced nutrient stress, as all but three of the species (Chrysanthemum leucanthemum, Cichorium intybus and Leontodon autumnalis) had reduced RGR relative to the values found by Almeida-Cortez et al. (1999) (Fig. 1a). Table 2 gives the mean values of each measured variable over all 19 species common to both this experiment and that of Almeida-Cortez et al. (1999). Paired t-tests showed that each variable in Table 2 differed significantly between the two experiments (P < 0·0002). Compared to the values obtained under non-limiting nutrients, RGR was reduced by 25%, SLA was reduced by 32%, root : shoot ratios increased by 320%, leaf nitrogen content decreased by 69%, and potential toxicity decreased by 46% (Fig. 1b). As secondary compounds often have a phylogenetic component, we classified species by tribe (Table 1) and then tested for significant differences in each measured variable between tribes using . Neither RGR nor


1260·06 J. S. AlmeidaCortez & B. Shipley 0·05

Toxicity

0·04

0·03

0·02

0·01

0·00 0·00

0·05

0·10

0·15

0·20

0·25

RGR

Fig. 2. Relationship between relative growth rate (RGR, g g–1 day–1) and potential toxicity of tissue extracts (1/ LC50, mg–1) in 19 species of herbaceous Asteraceae grown under nutrient limitation. Species belonging to the same tribe (Table 1) have the same plotting symbol.

potential toxicity varied significantly between tribes (F5,14 = 0·90 and 1·91, respectively). We asked two general questions in this paper. First, is there a negative correlation between relative growth rate in a collection of 19 species of herbaceous Asteraceae grown under nutrient limitation and potential tissue toxicity? We found no significant correlation between RGR and potential toxicity in this experiment (Spearman’s r = 0·12, P > 0·05) as seen in Fig. 2(a). Second, we asked if the degree of toxicity of these species was different from those measured in AlmeidaCortez et al. (1999) in which these same species were grown in non-limiting conditions. There was a significant decrease in toxicity when the plants were grown in nutrient-limited conditions (Fig. 1b). The mean potential toxicity was 0·017 (i.e. an average LC50 of 59 mg fresh tissue) in this experiment, as compared to a mean of 0·032 (LC50 of 31 mg–1 fresh tissue) in Almeida-Cortez et al. (1999). Although all species had detectable toxicities when grown under non-limiting conditions, 12 of the 19 species (Arctium lappa, A. minus, Chrysanthemum leucanthemum, Cichorium intybus, Hieracium aurantiacum, H. vulgatum, Lactuca canadensis, Leontodon autumnalis, Matricaria matricarioides, Rudbeckia hirta, Taraxacum officinale and Tragopogon pratensis) had potential toxicities below the detectable limit (0·01 mg–1, i.e. >100 mg fresh tissue) in the present experiment.

Discussion


We failed to observe any significant relationship between potential toxicity and seedling RGR. This is the same result as obtained by Almeida-Cortez et al. (1999) with 31 species growing in non-limiting conditions. It is also the same result as obtained by McCanny et al. (1990);

these authors measured potential toxicity by adding the tissue extracts of 32 species of wetland plants, collected in the field, to a standard diet fed to a generalist herbivore (European corn borer) and compared the growth reduction of the larvae to the RGR values of these species measured in controlled fertile growth conditions in the greenhouse. There is therefore little empirical evidence to support the hypothesis that there is an interspecific trade-off between the ability to grow rapidly and the ability to defend tissues based on toxicity of secondary compounds; if it exists, then it is quite weak. In each of these studies there was substantial interspecific variation in both RGR and toxicity, but no clear relationship between the two variables. It is interesting to compare our results with those of Almeida-Cortez (1998), consisting of six species (all included in the present experiment) grown under three levels of irradiance crossed with three levels of nutrient availability. The resource levels most relevant to the present study were those obtained at 500 µmol m–2 s–1 PFFD comparing the full strength versus 1/10 dilution of the hydroponic solution. Under such conditions, potential toxicity decreased in three of the six species. We conclude that potential toxicity in these herbaceous Asteraceae decreases with decreasing nutrient supplies (Fig. 1b), but in a manner that is species-specific. This is a plastic response to changing resource availability. On the other hand, there is no (or a very weak) interspecific trade-off between growth and potential toxicity when comparing species growing in the same light and nutrient environment (Fig. 2a). If such results are also found in the field, then this would suggest that species are less able to defend themselves through toxic secondary compounds when experiencing nutrient stress. This possibility will be the subject of a future study.

Acknowledgements This research was funded by the Natural Sciences and Engineering Research Council of Canada.

References Abel-Kader, M., Omar, A.A. & Stermitz, F.R. (1997) Erythroxan diterpenes from Fagonia glutinosa. Planta Medica 63, 374 – 376. Alkofahi, A., Rupprecht, J.K., Anderson, J.E., McLaughlin, J.L., Mikolajczack, K.L. & Scott, B.A. (1989) Search for new pesticides from higher plants. American Chemical Society Symposium Series 387, 25 – 43. Almeida-Cortez, J. (1998) La relation entre l’allocation aux composés secondaires et le taux de croissance relative chez les Ateraceae. PhD Thesis, Université de Sherbrooke, Sherbrooke, Canada. Almeida-Cortez, J.S., Shipley, B. & Arnason, J.T. (1999) Do plant species with high relative growth rates have poorer chemical defenses? Functional Ecology 13, 819–827. Arnason, J.T., Marles, R.J. & Aucoin, R.R. (1991) Isolation and biological activity derived phototoxins. Photobiological Techniques (eds D. P. Valenzeno, R. H. Pottier, P. Mathis & R. H. Douglas), pp. 187–196. Plenum Press, New York, NY.




Gleason, H.A. & Cronquist, A. (1991) Manual of the Vascular Plants of Northeastern United States and Adjacent Canada, 2nd edn. NY Botanical Garden, NY. Grime, J.P. & Hunt, R. (1975) Relative growth rate: its range and adaptive significance in a local flora. Journal of Ecology 63, 393 – 422. Harborne, J.B. (1989) Recent advances in chemical ecology. Natural Products Reports 6, 85 –109. He, K., Shi, G., Zeng, L., Ye, Q. & McLaughlin, J.L. (1997) Konishiol, a new sesquiterpene, and bioactive components from Cunninghamia konishii. Planta Medica 63, 158 –160. Herms, D.A. & Mattson, W.J. (1992) The dilemma of plants: to grow or defend. Quarterly Review of Biology 67, 283 – 335. Lang, C.A. (1962) Simple microdetermination of kjeldahl nitrogen on biological materials. Annals of Chemistry 30, 1692 –1694. Liskens, H.F. & Jackson, J.F. (1992) Plant Toxin Analysis. Springer-Verlag, New York, NY. Lopez, J.A. et al. (1997) Granulosin, a new chromone from Galipea gradulosa. Journal of Natural Products 60, 24 – 26. Malbry, T.J. & Bohlmann, F. (1977) Summary of the chemistry of the compositae. The Biology and Chemistry

of the Compositae (eds V. H. Heywood & J. B. Harborne), pp. 1097 –1104. Academic Press, New York, NY. McCanny, J.S., Keddy, P.A., Arnason, T.J., Gaudet, C.L., Moore, D.R.J. & Shipley, B. (1990) Fertility and the food quality of wetland plants: a test of the resource availability hypotheses. Oikos 59, 373–381. Meyer, B.N., Ferrigni, N.R., Putnam, J.E., Jacobsen, L.B., Nichols, D.E. & McLaughlin, J.L. (1982) Brine shrimp: a convenient general bioassay for active plant constituents. Journal of Medicinal Plant Research 45, 31–34. Middleton, K.R. (1960) New Nessler reagent and its use in the direct Nesslerization of kjeldahl digests. Journal of Applied Chemistry 10, 281–286. Poorter, H. & Garnier, E. (1999) Ecological significance of inherent variation in relative growth rate and its components. Handbook of Functional Plant Ecology (eds F. I. Pugnaire & F. Valladores), pp. 82 –120. Marcel Dekker, New York, NY. Spatafora, C. & Tringali, C. (1997) Bioactive metabolites from the bark of Fangara. Phytochemical Analysis 8, 139–142. Received 20 April 2001; revised 26 July 2001; accepted 14 August 2001