Insights from Biological Stoichiometry and Nutritional Geometry

bioRxiv preprint doi: https://doi.org/10.1101/2021.03.11.434999; this version posted March 12, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Title: Understanding the Evolution of Nutritive Taste in Animals: Insights from Biological

2 Stoichiometry and Nutritional Geometry

3 Running title: Biological Stoichiometry, Nutrition and Taste

4 Authors: Lee M. Demi1*, Brad W. Taylor1, Benjamin J. Reading1, Michael G. Tordoff2, Robert

5 R. Dunn1,3

6 Affiliations:

7 1Department of Applied Ecology, North Carolina State University, Raleigh, NC 27695, USA

8 2 Monell Chemical Senses Center, Philadelphia, PA 19104, USA

9 3 Center for Evolutionary Hologenomics, University of Copenhagen, Copenhagen, DK

10 *Corresponding author: [email protected]

12 Abstract:

13 A major conceptual gap in taste biology is the lack of a general framework for understanding the

14 evolution of different taste modalities among animal species. We turn to two complementary

15 nutritional frameworks, biological stoichiometry theory and nutritional geometry, to develop

16 hypotheses for the evolution of different taste modalities in animals. We describe how the

17 attractive tastes of Na, Ca, P, N and C containing compounds are consistent with principles of

18 both frameworks based on their shared focus on nutritional imbalances and consumer

19 homeostasis. Specifically, we suggest that the evolution of multiple nutritive taste modalities can

20 be predicted by identifying individual elements that are typically more concentrated in the tissues

21 of animals than plants. Additionally, we discuss how consumer homeostasis can inform our

22 understanding of why some taste compounds (i.e., Na, Ca and P salts) can be either attractive or

23 aversive depending on concentration. We also discuss how these complementary frameworks can

24 help to explain the phylogenetic distribution of different taste modalities and improve our

25 understanding of the mechanisms that lead to loss of taste capabilities in some animal lineages.

26 The ideas presented here will stimulate research that bridges the fields of evolutionary biology,

27 sensory biology and ecology.

29 Keywords (5): gustation, nutritional ecology, chemoreception, homeostasis, optimal foraging

30 1. Introducing a general framework for consumptive taste evolution

31 All organisms are faced with the challenge of procuring the essential elements and biochemicals

32 of life for growth, maintenance and reproduction, often in proportions that deviate substantially

33 from their environmental availability. For animals, nutrients (i.e., elements and/or biochemicals)

34 are acquired through their diet by consuming foods that may vary considerably in their chemical

35 composition and that frequently provide an inadequate supply of one or more essential nutrient

36 (1). Thus, many animals suffer periodic nutritional imbalances whereby the nutrient content of

37 available foods does not match their nutritional requirements. Such imbalances (i.e., over- or

38 undersupply of essential nutrients) can affect consumer metabolism and performance, ultimately

39 resulting in reduced growth and fitness (2, 3). Because nutritional imbalances between

40 consumers and their foods are both pervasive and consequential animals have evolved a range of

41 adaptations, from behavioral to physiological and chemosensory, that allow them to modulate

42 nutrient intake, as well as post-ingestion nutrient processing (i.e., assimilation and allocation), to

43 minimize potential imbalances (3).

44 Selective foraging is a key behavioral adaptation that allows animals to regulate nutrient

45 intake in order to achieve balanced nutrient supply (3). Choosing exactly which foods to eat,

46 however, requires that animals differentiate among potential food sources of differing chemical

47 composition to select those that are most nutritionally advantageous. One of the primary

48 mechanisms that animals use to assess the nutritional quality of foods is gustation, or taste,

49 which provides animals the ability to evaluate the chemical composition of potential foods prior

50 to ingestion via complex chemosensory systems (4, 5, 6). Indeed, there is broad recognition that

51 gustatory systems function primarily as a screening mechanism that drives consumption of

52 nutrient- and energy-dense foods, as well as the rejection of potentially harmful ones based upon

53 the presence, and concentrations, of a variety of different chemicals. These various chemical

54 stimuli interact with specialized receptor cells to produce signals that are transduced and

55 interpreted as unique taste modalities, including the salty, sweet, bitter, sour and umami tastes

56 familiar to humans (4, 6).

57 The variety of different taste modalities provide organisms with a way to evaluate the

58 quality of food based on multiple aspects of its chemical composition and can be broadly

59 grouped by whether they taste “good”, and therefore promote consumption, or “bad”, thus

60 producing an aversive response. Consumptive responses, such as those elicited by sweet and

61 umami tastes, promote the ingestion of foods that supply compounds essential for growth and

62 metabolism. Conversely, aversive responses, such as those typically generated by bitter and sour

63 tastes, encourage animals to reject, rather than consume specific foods. Aversive tastes are

64 generally thought to inform consumers that food may contain toxic or harmful chemicals, such as

65 allelochemicals, though not all chemicals that produce innate aversive responses are necessarily

66 harmful (5, 6). Additionally, some elements, such as Na, Ca and P which are essential but can be

67 toxic when oversupplied, may elicit either consumptive or aversive responses depending on

68 concentration, thereby allowing consumers to regulate intake within a relatively narrow window.

69 All animals possess gustatory sensing capabilities (3, 7), reinforcing the notion that

70 regulation of nutrient intake is highly consequential for consumer fitness. Interestingly, gustatory

71 sensing appears to have evolved largely independently among prominent animal lineages (i.e.,

72 mammals and insects) (5), yet exhibits remarkable convergence around taste capabilities for a

73 small suite of compounds (e.g., amino acids, carbohydrates, Na salts). This suggests that animals

74 have faced similar nutritional constraints throughout their evolutionary history and indicates

75 potential for developing a nutritional framework for understanding the evolution of taste that has

76 broad applicability. Nevertheless, we currently lack a predictive understanding of when and why

77 particular taste modalities might be evolutionarily favored (3), thereby leading to differences in

78 the breadth of gustatory capabilities among species.

79 In this paper, we turn to two complementary frameworks, biological stoichiometry theory

80 (2, 8, 9) and nutritional geometry (1, 3, 10), that have been developed by ecologists to

81 characterize nutritional imbalances (and their consequences) in trophic interactions to draw

82 insights into the evolutionary contexts under which different taste modalities have evolved in

83 animals. In the following sections, we summarize the basic tenets of biological stoichiometry

84 theory and nutritional geometry, including the unifying concept of consumer homeostasis, and

85 draw upon a review of the taste biology literature to show how these complementary frameworks

86 advance our understanding of the evolution of multiple taste modalities in animals. We use the

87 principles of biological stoichiometry to demonstrate that the evolution of multiple nutritive taste

88 modalities can be predicted by identifying individual elements that are typically more

89 concentrated in the tissues of animals than plants. Moreover, we describe how consumer

90 homeostasis can inform our understanding of why some taste compounds (i.e., Na, Ca and P

91 salts) can be either attractive or aversive depending on concentration. Additionally, we

92 demonstrate how these complementary frameworks can help to explain the phylogenetic

93 distribution of different taste modalities, including why the taste of Ca (at least at some

94 concentrations) is attractive to some vertebrate species but appears to be strictly aversive to

95 insects. We also discuss how biological stoichiometry and nutritional geometry can improve our

96 understanding of the mechanisms that lead to loss of taste capabilities, as well as the

97 pseudogenization of associated taste receptor genes, in some animal lineages (11, 12, 13). We

98 also discuss some caveats to our application of these complementary frameworks and finally,

99 propose several ideas that we hope will stimulate future research, including into the evolutionary

100 history and phylogenetic distribution of different taste modalities, and incorporation of taste into

101 functional trait-based ecological analyses.

102

103 Nutritional frameworks in ecology

104 Ecologists have long recognized that the quantity and nutritional quality of foods affect the

105 growth and performance of animals (1). Given that nutritional requirements vary among animal

106 species, shifts in the nutrient composition of available foods can scale beyond individual and

107 population effects to alter community and ecosystem dynamics, including fluxes of energy and

108 nutrients to and from the environment and among trophic levels (14, 15). Increasing recognition

109 of the role of animal nutrition in driving ecological dynamics, from foraging behavior to

110 consumer driven nutrient recycling, has led to the development of multiple conceptual

111 frameworks that seek to characterize the nature and consequences of nutritional imbalances in

112 trophic interactions (1). Biological stoichiometry theory and nutritional geometry are perhaps the

113 most prominent of these frameworks and have been developed in parallel over the last several

114 decades (14, 15). Though these frameworks focus on different nutritional currencies (i.e.,

115 elements versus macro- and micronutrients) there is substantial conceptual overlap, including an

116 emphasis on homeostasis, that suggest promise in their ability to inform our understanding of

117 how nutritional imbalances have shaped the evolutionary history of gustation in animals (14).

118 There are large bodies of literature devoted to the development and application of these

119 frameworks, including several recent papers highlighting their connections and encouraging their

120 synthesis (14, 15, 16). As such, we provide only brief summaries of their basic principles to

121 establish their usefulness in describing the ecological contexts under which different taste

122 modalities have evolved.

123 Biological, or commonly, ecological stoichiometry is the study of the balance of energy

124 and multiple elements in ecological interactions (2, 8, 9). For the purposes of this paper, we

125 prefer the term “biological stoichiometry”, as its broader scope explicitly integrates cellular and

126 genetic mechanisms (9). Similar to stoichiometry of chemical reactions, biological stoichiometry

127 applies thermodynamic principles to the mass balance of multiple elements in biological

128 interactions (8). It is rooted in the observation that individual organisms are composed of

129 elements in relatively fixed ratios (stoichiometric homeostasis) and that available resources

130 frequently supply elements in ratios that deviate from those of the organism. This is particularly

131 true of animals, which tend to be far less plastic in their elemental composition and exhibit far

132 narrower interspecific variation in elemental composition than plants (and many autotrophic

133 microorganisms; Fig. 1), thereby resulting in notable stoichiometric imbalances between plants

134 and animals (17; 18). Core principles of biological stoichiometry include the observations that

135 (1) organisms have limited capacity to alter their elemental composition even when the supply of

136 elements is variable and that (2) large imbalances in elemental composition between consumers

137 and their foods (e.g., herbivores and plants, Fig. 1) are common (2). A central thesis of biological

138 stoichiometry theory is that reconciling these imbalances (i.e., via selective foraging behavior or

139 post-ingestive regulation) while maintaining elemental homeostasis has consequences for the

140 growth and fitness of animals, as well as the cycling of elements between animals and the

141 environment. Hence, a key strength of biological stoichiometry is the ability to scale and make

142 predictions across different levels of organization (e.g., from consumer taste receptors to

143 ecosystem nutrient supply).

144 Biological stoichiometry has historically focused primarily on imbalances between the

145 supply and demand of C, N and P in trophic interactions. This bias reflects how commonly these

146 elements limit biological productivity across trophic levels in many different ecosystems (19, 20)

147 and a tendency to view limitation through the prism of single nutrient limitation rooted in

148 Liebig’s law of the minimum. However, biological stoichiometry is not constrained to these

149 three elements and incorporation of a wider suite of biologically essential elements into

150 stoichiometric analysis is an emerging frontier (21, 22). Additionally, there is increasing

151 recognition that key molecular and cellular level processes may be co-limited by multiple

152 chemical elements (23) and that studies of nutrient limitation should consider this possibility and

153 expand the elements considered in stoichiometric analyses (24). A notable limitation of

154 biological stoichiometry theory with respect to its ability to characterize nutritional imbalances is

155 its focus on individual elements (25, 14). Given that animal nutrition is largely biochemical in

156 nature, focusing solely on the ratios of elements in foods may obscure potential imbalances in the

157 supply of certain essential macro- and micronutrients that affect consumer performance and

158 reduce fitness (25).

159 Similar to biological stoichiometry, nutritional geometry, or the “geometric framework”

160 for animal nutrition, focuses on how the availability of multiple nutrients affect consumer growth

161 and fitness and relies on homeostatic regulation as a conceptual cornerstone (1, 3, 26). However,

162 nutritional geometry differs by focusing on the availability of micro- and macronutrients (i.e.,

163 amino acids, carbohydrates, lipids), rather than chemical elements, making it a more nutritionally

164 explicit framework (1, 14). This framework recognizes that foods are complex mixtures of

165 nutrients and other non-nutritive compounds and that any single food item is likely to provide an

166 insufficient supply of at least some essential nutrients relative to consumer requirements at a

167 given point in time. Nutritional geometry models relationships among various nutrients (or non-

168 nutrients, such as allelochemicals in some applications) using a multi-dimensional state-space

169 approach to characterize the mechanisms, such as foraging behavior or nutrient assimilation, that

170 animals use to achieve nutritional homeostasis while foraging among foods of varying nutrient

171 composition (26). This geometric approach was designed to provide quantitative predictions

172 regarding how consumers utilize multiple foods that vary in their respective amounts of different

173 nutrients (i.e., proteins and carbohydrates) in the service of homeostatic regulation of nutrient

174 intake (1, 26).

175 Much of the early formulation and application of nutritional geometry focused on how

176 animals forage among various resources to achieve a specific target ratio of two or more

177 nutrients in their diet (14). Proponents of nutritional geometry have recognized that taste and

178 other chemosensory systems (i.e., olfaction), as well as learned nutritional associations for

179 various foods, play an important role in nutrient-based foraging decisions, allowing consumers to

180 achieve balanced supply of multiple nutrients (3). In contrast, most applications of biological

181 stoichiometry have focused on how organisms reconcile imbalances in the supply of essential

182 elements via post-ingestive mechanisms, such as altering element use efficiencies (i.e.,

183 assimilation) and differential excretion of individual elements, rather than through selective

184 foraging. However, the potential for biological stoichiometry theory to inform our understanding

185 of how animals forage among nutritionally heterogenous food resources, particularly through the

186 evolution and employment of chemosensory systems, has been previously suggested (17), but

187 remains largely unexplored.

188 In the following sections we describe how these two complementary frameworks provide

189 a platform, rooted in the concept of consumer homeostasis, to describe the evolutionary context

190 from which multiple different taste modalities have evolved. We rely primarily on the more

191 reductionist approach of biological stoichiometry theory, which focuses on nutritional

192 imbalances at the elemental level, for many examples, including hypotheses related to

193 evolutionary drivers of taste associations for N, P, Na and Ca containing molecules. We

194 acknowledge that our use of biological stoichiometry at times treats individual elements as

195 surrogates for complex biochemicals, and that gustatory receptors may be responding to those

196 biochemicals rather than to individual elements contained within them, as in the case of N and

197 amino acids. However, this approach remains useful in most cases given that a standard

198 assumption of biological stoichiometry is that tissue concentrations of individual elements can

199 often be related to a small number of specific biochemicals, such as protein being the primary N

200 pool in most organisms (2, 14). There are, however, notable cases where more biochemically

201 focused nutritional geometry provides greater insight into potential evolutionary drivers of taste,

202 such as for carbohydrate and lipid taste. As such, we rely on both frameworks in our discussion

203 of the evolutionary contexts for taste modalities, with the common thread being that consumers

204 have relatively specific nutritional requirements, determined largely by their own chemical

205 composition (which is relatively homeostatic), and that potential foods (at least individually)

206 often do not satisfy those requirements.

207

208 3. Consumer homeostasis and nutritional imbalance: an evolutionary context for nutritive taste

209 The concept of elemental homeostasis, meaning that organisms have limited capacity to alter

210 their chemical composition even when the supply of chemical elements is variable (2), is

211 foundational to both biological stoichiometry theory and nutritional geometry.

212 A second shared concept is that for many animals, the chemical composition of potential foods

213 can be very different than that of the consumer. These differences result in nutritional

214 imbalances, where certain foods may provide either an under- or oversupply of certain essential

215 nutrients (i.e., elements or biochemicals), that can constrain individual performance and reduce

216 reproductive potential (2, 3). Because consumers have limited capability to modify their

217 chemical composition to more closely match that of their foods, they must reconcile imbalanced

218 nutrient supply by using pre- and/or post-ingestive mechanisms (2, 3, 27), such as selective

219 feeding or differential release of individual elements as waste products to achieve nutritional, or

220 maintain tissue, homeostasis. However, post-ingestion nutrient processing, particularly of

221 nutrients which are oversupplied, may have metabolic costs that affect consumer performance,

222 highlighting the importance of selective foraging to reduce nutritional imbalances at the front

223 end (28, 29). Gustation is commonly viewed as having evolved to function as a sensory tool that

224 allows animals to assess the chemical composition of their food. Expanding further, we suggest

225 that the evolution of many individual taste modalities should be explained, at least in part, by

226 common nutritional imbalances in the diets of animals. Moreover, we contend that these

227 common imbalances are the drivers of the remarkable convergence in gustatory evolution among

228 major animal lineages (5). We believe that biological stoichiometry and nutritional geometry

229 provide useful frameworks for identifying and characterizing those imbalances and therefore

230 have great potential to inform our understanding of the evolutionary contexts for many

231 individual taste modalities in animals.

232 Biological stoichiometry in particular provides a simple, yet useful, framework for

233 characterizing common nutritional imbalances between consumers and their foods by comparing

234 their respective elemental composition (though see carbon, section 4). We can make some useful

235 generalizations about consumer nutritional imbalances by identifying elements that are

236 commonly less concentrated in foods than in the tissues of consumers. The common convention

237 in biological stoichiometry is to express the relative concentrations of different elements of

238 consumers and their food as ratios, particularly as C:nutrient (i.e., N or P) ratios. For example,

239 Fig. 1 displays biomass C:N and C:P ratios of autotrophs (i.e., plants) and animals across

240 multiple trophic levels. The median C:N and C:P of autotroph biomass is considerably greater

241 than that of animals (at any trophic level; Fig. 1), indicating that plant tissue has a greater amount

242 of C per unit N, and C per unit P, than does animal tissue. As such, an animal of typical chemical

243 composition (i.e., around the median) that is feeding on a plant of typical chemical composition

244 will experience undersupply of both N and P, relative to C. We would expect that such elements,

245 here N and P, would have a greater likelihood of limiting consumer growth and metabolism and

246 therefore predict that animals should find them (or the primary chemicals that contain them, as in

247 the case of N and amino acids) pleasing to taste ─ the pleasing oral sensation would reward

248 animals for consuming foods that are relatively high in those potentially limiting nutrients.

249 To illustrate, we compared the whole-body elemental composition of insects, marine

250 fishes, and mammals with the elemental composition of foliar plant tissue (Fig. 2, Table 1) in

251 order to identify which elements are typically less concentrated in plants than in animals. We use

252 plants and animals in this example because of the typically greater difference in the elemental

253 composition of herbivores and plants than carnivores and their animal prey (Fig. 1) and with the

254 assumption that this is representative of conditions faced by herbivorous animals (particularly

255 terrestrial) throughout their evolutionary history. Based on principles of biological stoichiometry,

256 we might expect there to be consumptive taste associations for elements that are typically more

257 concentrated in the tissues of animals than in plants, given that these elements are likely to limit

258 performance and reproductive potential. By extension, we would not necessarily expect positive

259 taste associations for elements that are more concentrated in plants than animals. Comparison of

260 animal and plant elemental composition reveals that Na, P, N and sometimes Ca (i.e., in

261 mammals but not insects) are considerably more concentrated (between ~3.5 - 40×) in animal

262 than in plant biomass (Fig. 2, Table 1). A comprehensive review of taste research reveals that

263 each of these four elements is indeed associated with at least one chemical, or biochemical,

264 known to have a taste that elicits a consumptive response in some animal species. Conversely,

265 we were unable to find evidence of nutritive taste modalities that were obviously associated with

266 specific elements that are closely balanced between plants and animals (except C, see below), or

267 that are more concentrated in plants than in the tissues of the three animal groups in our analysis.

268 For many animals some of these elements (i.e., Ca, Na, P) or associated biochemicals

269 may taste good at some concentrations, but also may produce aversive taste responses when

270 consumed at other, especially high, concentrations. Here again we can turn to the concept of

271 consumer homeostasis to provide a potential explanation for this pattern. Because oversupply of

272 some essential nutrients, such as Na and Ca, is toxic to consumers and because consumers have

273 limited ability to alter their constitution (homeostasis) consumers may be required to regulate

274 homeostasis by managing nutrient intake within a relatively narrow window. The combination of

275 both consumptive and aversive responses to some essential nutrients, especially those that can

276 have toxic or other negative effects when oversupplied, when supplied at different concentrations

277 can be viewed as helping consumers achieve nutritional homeostasis by regulating nutrient

278 intake (3). In the following subsections we discuss taste associations for Na, P, N, and Ca and

279 summarize the evidence that supports these taste capabilities. We provide examples of taste

280 associations, or in some cases preferential feeding behaviors, for these elements within both

281 major vertebrate and invertebrate lineages. Though in some cases direct links between

282 preferential feeding behaviors and associated taste capabilities have not been explicitly made, we

283 believe this anecdotal evidence is at least suggestive of a role for taste in driving these behaviors

284 and warrants further research. We do, however, acknowledge that such behaviors are not

285 necessarily indicative of taste associations, per se, and may be driven by other characteristics of

286 foods such as texture, or by learned nutritional associations (3). Nevertheless, consumptive taste

287 associations for each of these four elements have been described for at least some taxa and we

288 argue that the large number of additional species that display preferential feeding behavior

289 suggests that these taste modalities may be more widespread among animals than what is

290 currently known from research on a relatively limited number of model organisms.

291

292 3.a. Na imbalances and salty taste

293 Na comprises ~0.3% of the body dry mass of animals but only trace amounts of the mass of terrestrial

294 plants, making it on average ~40× more concentrated in animals than in foliar plant tissue (Fig. 1). Na is

295 typically attractive to both vertebrates and at least some invertebrates at particular concentrations (5, 6,

296 27, 30, 31, 32). Na salts (i.e., NaCl) are the primary taste molecules associated with gustatory sensing of

297 Na and are responsible for salty taste in humans. The greater Na imbalances between plants and animals

298 than among major animal groups (Fig. 2) suggests that selection for Na taste should be stronger for

299 herbivores than predators (Table 1). However, Na taste remains intact and putatively functional in most

300 mammalian species (12, 13, 33). Nevertheless, herbivorous mammals appear to have a greater affinity

301 for Na than do carnivorous mammals (33), and there is considerable evidence of Na-related feeding

302 behaviors among herbivores (30). For example, many terrestrial herbivores seek out Na-rich mineral

303 deposits on the landscape. Examples include puddling behavior by insects (34), use of dry or wet licks

304 by vertebrates, drinking of sea water by reindeer, soil eating by mountain gorillas (30, 35), and

305 preference for Na rich aquatic plants over terrestrial plants by moose and other ungulates (36). Though

306 Na appears to be attractive to most animals at low to moderate concentrations, aversive responses to

307 high Na concentrations have been reported for both mammals and insects (6, 27, 31). Excess levels of

308 dietary Na salts are detrimental to animals and aversive taste responses to high concentrations can be

309 seen as a mechanism to protect against oversupply. Thus, the evolution of both consumptive and

310 aversive taste responses across Na supply gradients is consistent with biological stoichiometry, as it

311 provides animals sensory cues to guide feeding behavior that maintains relatively tight homeostatic

312 regulation of body Na content, a critical component of osmoregulation. Thus, we suggest that dietary Na

313 imbalances (Fig. 1) combined with the negative consequences of deviation from strict Na homeostasis

314 have driven selection for both consumptive and aversive Na taste responses.

315

316 3.b. P imbalances and P taste

317 P imbalances are pervasive at the base of food webs, with P being on average 13× more

318 concentrated in animal than plant biomass (Fig. 2). P availability is often rate-limiting for many

319 organisms and its relationship to growth has garnered considerable interest from a biological

320 stoichiometry perspective (i.e., growth rate hypothesis, 37). Though there has historically been

321 substantial interest among ecologists in the role of P as a limiting element in trophic interactions,

322 taste associations for P-containing compounds have received relatively little attention compared

323 to other basic taste senses (i.e., sweet, salty, bitter, etc.). Nevertheless, P taste has been confirmed

324 in some animal taxa, including cattle (30, 38, 39) and rodents (40), and appears to be elicited

325 most strongly by phosphate, which is the primary form of P in many P-rich biomolecules (i.e.,

326 nucleic acids, ATP, phospholipids). Moreover, preferential feeding on P-rich foods has been

327 observed among numerous invertebrate (i.e., grasshoppers, 41) and vertebrate (i.e., rats, 42) taxa

328 and may be fairly widespread. Because generalist herbivores are more likely to suffer greater

329 dietary P deficiencies than carnivores (Fig. 1), and thus stronger selective pressure for P-

330 associated taste, we propose that P taste should be more common among herbivores (and

331 omnivores) than strict carnivores. Additionally, the higher P-requirements imposed by building

332 and maintaining skeletal bone suggest that vertebrates likely experience stronger selection for P

333 taste than do invertebrates. Lastly, excess dietary P supply has been associated with reduced

334 individual growth rates in a variety of organisms (29), reflecting the importance of

335 stoichiometrically balanced P supply for consumers. As such, animals may be likely to find P-

336 associated taste to be aversive at high concentrations, similar to Na and Ca, which may explain

337 why some species reduce consumption rates when reared on high-P diets (crustaceans [43],

338 chickens [44] and mice [40]).

339

340 3.c. N imbalances and umami taste

341 N is, on average, ~4× more concentrated in animal than foliar plant biomass (Fig. 2). The

342 frequently limiting role of N makes it a strong candidate for consumptive taste associations.

343 Indeed, such taste associations for various N-rich amino acids appear to be widespread, having

344 been described in numerous vertebrate and invertebrate taxa. In humans, umami taste is

345 stimulated by the N-rich amino acids L-glutamate and L-aspartate whereas mice are attracted to

346 many L-amino acids (45). Insects also make consumptive responses to a variety of amino acids,

347 although the relative sensitivity to different amino acids varies among insect taxa (46). We posit

348 that selective pressure for N-associated taste should generally be stronger in herbivores than

349 predators, given the greater N-content of animal than plant biomass and thus greater likelihood

350 of N-limitation among herbivores (Fig. 1). However, retention of umami taste in obligate

351 carnivores, may be an indication of the potential for carnivore N-limitation. Despite the

352 appearance of relatively stoichiometrically similar prey, carnivores may experience N-limitation

353 due to their reliance on amino acids as a source of glucose for cellular respiration and the low

354 carbohydrate content of their food (animal prey, 47).

355

356 3.d. Ca imbalances and Ca taste

357 Ca is, on average, ~3.5× more concentrated in the tissues of animals than in plants, although

358 there is major taxonomic variation in the Ca content of animals due to the production of Ca-rich

359 skeletal bone in vertebrates (Fig. 2). Consumptive taste associations for several Ca salts have

360 been observed among diverse groups of animals including amphibians, birds, and mammals (48,

361 49, 50, 51). Calcium taste is likely to be under stronger selection in herbivores than predators,

362 due to the generally lower Ca content of plant biomass (Fig. 2, Table 1). Moreover, selection for

363 Ca taste should vary considerably between invertebrates and vertebrates due to large differences

364 in Ca-content associated with skeletal bone (and eggshell formation) in the latter. Indeed, Ca

365 appears to be strictly aversive in insects, which are depleted in Ca relative to plants (Fig. 1, [32,

366 52]), whereas Ca is attractive at low and moderate concentrations to a diverse suite of

367 vertebrates. For example, the Ca content of plant material has been linked to consumptive

368 preferences in rats and mice in laboratory experiments (53), as well as the feeding behaviors of

369 wild gorilla populations, which use forest clearings to access their preferred Ca and Na rich plant

370 species (54). Additionally, dietary Ca deficiencies in herbivorous desert tortoises have been

371 proposed to explain their occasional consumption of numerous Ca-rich resources such as bones,

372 bone-rich vulture feces, reptile skin castings, feathers, mammalian hair (55), and Ca carbonate

373 stones (56). Similar to Na taste, Ca taste is aversive at high concentrations that may result in

374 hypercalcemia (57, 52). Thus, the presence of both consumptive and aversive responses to

375 different levels of dietary Ca supply indicate a role for taste in regulating Ca homeostasis in

376 animals and suggest homeostatic maintenance as a likely driver behind the evolution of Ca-

377 associated taste.

378

379 4. Carbon, carbohydrate and lipid taste, a nutritional geometry perspective

380 We have demonstrated that a relatively simple analysis of elemental imbalances between plants

381 and animals, rooted in foundational concepts of biological stoichiometry, explains the context for

382 the evolution of multiple nutritive taste modalities, including those for Na, N (amino acids), Ca

383 and P or associated compounds. However, conspicuously absent from this list are nutritive taste

384 associations for certain C-rich molecules, including the pleasing taste of sugars experienced by

385 humans (sweet taste) and many other organisms. Indeed, biological stoichiometry appears

386 insufficient in providing an explanation for why many animals find certain C-rich biomolecules,

387 such as sugars, pleasing, and others (i.e., structural carbohydrates) not. Carbon is the dominant

388 element by mass in animals and plants and is similarly concentrated in both animal and plant

389 foliar tissue (Fig. 2). However, an organism’s total demand for C is not determined solely by its

390 body C composition but must also account for additional C for energy metabolism (58), given

391 that C compounds are relied upon as energy by nearly all life forms. Thus, focusing strictly on

392 biomass proportions of C in plants and animals suggests that the C concentration of food is

393 unlikely to limit consumer growth and metabolism. As such, dietary C imbalances for consumers

394 may be obscured by the similar C content of animals and plants (Fig. 2).

395 Assessing potential C-related nutritional imbalances based on bulk C concentration of

396 foods is also difficult given the wide variety of C-containing biochemicals, that vary in their

397 essentiality, digestibility and ability to be synthesized by consumers. For example, a large

398 proportion of the C in foliar tissue of plants occurs in the form of recalcitrant structural

399 carbohydrates (i.e., cellulose and lignin) that many animals cannot digest, rather than structurally

400 simpler carbohydrates, such as sugars and starches, that are more labile and energy dense.

401 Additionally, many essential C-containing and C-rich macronutrients, including some lipids and

402 amino acids, cannot be synthesized endogenously by animals and therefore must be acquired

403 through the diet. As such, it would be advantageous for animals to differentiate among foods not

404 by their total C-concentration, but by assessing the presence and relative abundances of different

405 C-containing macronutrients.

406 Here we turn to the framework of nutritional geometry to inform our understanding of

407 how the evolution of C-associated tastes, such as that for sugars, are shaped by the need to

408 regulate nutritional homeostasis (i.e., a target nutrient intake ratio) while foraging among foods

409 that vary in their relative concentrations of multiple essential macronutrients. Nutritional

410 geometry studies have often focused on the relative concentrations of C-containing

411 macronutrients, including carbohydrates, lipids and proteins, in multiple foods. Given that some

412 of these nutrients cannot be synthesized by animals, and that they are variably concentrated in

413 different foods, one would expect that animals would experience strong selective pressure on

414 mechanisms, including taste, that allow them to detect their presence and assess their

415 concentrations in potential foods (3). In the next several paragraphs we discuss the evidence for

416 taste associations for carbohydrates and lipids and how nutritional geometry can inform our

417 understanding of their evolutionary history. Although we previously discussed the evolution of

418 amino acid taste in context of dietary N deficiencies, we acknowledge here that nutritional

419 geometry may provide additional insight into the evolution and radiation of amino acid taste in

420 animals. For example, the more nutritionally explicit framework of nutritional geometry may

421 help us understand why different animal species taste different combinations of amino acid

422 molecules by considering the unique nutrient requirements and diets of different species.

423 Carbon is the primary energy currency of life and it is glucose that fuels the process of

424 cellular respiration in animals. Dietary carbohydrates, such as sugars and starches, are the

425 primary source of glucose for many animals, though it can also be derived from other sources

426 (i.e., lipids, amino acids). Given that foods can vary substantially in their total carbohydrate

427 content, and that different carbohydrates (i.e., sugars vs. fiber) vary in their digestibility, animals

428 should experience selective pressure on mechanisms that allow them to detect and differentiate

429 between types of dietary carbohydrates in foods. Indeed, many animals taste simple sugars and

430 find sugar taste attractive, while many mammal species are also able to taste, and enjoy, more

431 complex carbohydrates, including starches and starch derivatives (59). Animals, however, do not

432 appear to taste more complex structural carbohydrates, such as cellulose fiber. Thus, it appears

433 that carbohydrate taste is fine-tuned for labile, energy-rich carbohydrates, thereby enabling

434 consumers to regulate their intake of energetically important foods. The link between sugar taste

435 in animals and metabolic C demand that we suggest is supported by multiple lines of evidence.

436 For example, in frugivorous non-human primates, sucrose taste threshold increases with

437 decreasing body size (60), indicating that the sweet taste receptors of small-bodied species are

438 tuned for selecting foods with higher sucrose than those of large-bodied species. This pattern is

439 consistent with basic allometric scaling rules relating to organismal metabolic rates, where mass-

440 specific metabolic rates decrease with increasing body size (61) and is suggestive of a link

441 between sugar taste thresholds and C demands of at least some primate species. Moreover,

442 bumble bees and honey bees have gustatory receptor neurons that are triggered by especially

443 high sugar concentrations (62) providing a gustatory mechanism to meet their high metabolic C

444 demand. Likewise, hummingbird’s high metabolic demands for C are consistent with the

445 evolution of their sweet taste receptor from their ancestor’s umami taste receptor as their diet and

446 habits changed (63). These patterns are consistent with the principles of nutritional geometry

447 which recognizes animals use sensory systems, like taste, to inform foraging decisions that

448 balance the intake of multiple nutrients that vary in their concentration among different foods.

449 Lipids, including fatty acids, phospholipids and steroids, are another important and

450 diverse class of C-rich macromolecules that are essential in animal nutrition and can be an

451 important source of glucose for cellular respiration. Potential foods (i.e., nuts and seeds vs fruits

452 and foliage, plant vs animal prey) may vary substantially in the both the amount and types of

453 lipids that they contain and some nutritionally important lipids, including various essential fatty

454 acids, must be obtained directly from the diet. As such, animals should benefit from the ability to

455 evaluate the lipid composition of different foods to maintain a balanced intake of essential lipids

456 and other nutrients relative to nutritional demands. Indeed, animals possess multiple systems,

457 including taste, olfaction, and gut-nutrient sensing, that contribute to lipid detection in their foods

458 and lipids are increasingly recognized as having pleasing “flavor”, in addition to smell and

459 texture, for many animals (64). However, evidence for specific positive taste associations for

460 major lipid molecules, such as various fatty acids, is not widespread among animals, indicating

461 that other mechanisms such as olfaction and gut nutrient sensing may be more important for

462 detecting dietary lipids and driving lipid intake in some species. For example, taste for fatty acids

463 appears to be aversive in many species, including humans, and the attractive taste of lipids

464 experienced by other species, including mice, is apparently stimulated only by certain long chain

465 fatty acid molecules, such as linoleic acid (64, 65), an essential polyunsaturated omega-6 fatty

466 acid that must be acquired through the diet. Thus, fatty acid taste in some species may be seen

467 primarily as a mechanism for regulating consumption of specific essential fatty acids, which

468 animals cannot synthesize endogenously, a pattern that is consistent with concepts from

469 nutritional geometry. The pleasing sensation experienced when consuming dietary fats by some

470 species may be triggered primarily by sensory cues other than taste [i.e., olfaction, texture,

471 postingestive cues (66)], thereby highlighting the need to consider how other sensory systems

472 interact to influence foraging behaviors that play a role in regulating the intake of lipids (3).

473 Nutritional geometry implies that animals that consume a variety of foods of varying

474 nutrient composition will benefit from the ability to detect, whether by taste or another

475 mechanism, the presence of multiple nutrients in their food. Animals should experience

476 particularly strong pressure for nutrient sensing systems that detect essential biochemicals that

477 cannot be synthesized and must be consumed. Tastes for certain fatty acids, as discussed above,

478 are examples. Another example is the attractive taste of vitamin C (ascorbic acid) to guinea pigs

479 (67) and certain laboratory strains of rats (68), which are incapable of vitamin C synthesis.

480 Consumptive taste associations for vitamin C, which has a chemical formula of C6H8O6, would

481 not necessarily be predicted from an analysis of stoichiometric imbalances between guinea pigs

482 or rats and their food.

483

484 5. Lack and loss of taste senses among animal species

485 Understanding why some animal species possess, or lack, certain taste senses has generated

486 considerable interest among taste researchers, particularly in cases where specific taste functions

487 have been lost within some animal lineages (12, 13). The loss of taste receptor function has been

488 linked, in part, to evolutionary shifts in feeding ecology, including dietary specialization, that

489 reduce the selective pressures on taste receptor genes (11, 12, 69, 70, 71, 72). For example, sweet

490 taste has been independently lost in multiple lineages of terrestrial carnivores, including the

491 strictly carnivorous felids, cetaceans, and vampire bats (12, 13, 71). This loss of sweet taste has

492 been linked to pseudogenization of the T1R2 (sweet) receptor gene following dietary shifts

493 towards obligate carnivory (13) as the relative lack of carbohydrates in animal tissue is thought

494 to reduce selective pressure for maintaining a functioning sweet taste receptor (12). Shifts in

495 feeding mode, meaning how an organism feeds (i.e., swallowing versus chewing prey, fluid

496 feeding, etc.), have also been linked to loss of taste functions in some animal lineages. For

497 example, the substantial loss of taste (sweet, umami, bitter and sour) in cetaceans has been linked

498 to pseudogenization of multiple genes following an ancestral dietary shift from herbivory to

499 carnivory (animals contain fewer carbohydrates and bitter tasting compounds than plants), as

500 well as a behavioral shift towards swallowing prey whole, which precludes the release of taste

501 compounds within the oral cavity through mastication (72). Additionally, in vampire bats, the

502 shift to blood feeding and use of infrared sensing to locate blood flows have presumably reduced

503 selection for functional carbohydrate and amino acid taste receptors (71) resulting in the loss of

504 sweet and umami tastes, respectively. Despite these notable examples, lineage-specific

505 pseudogenization events do not always align with shifts in feeding ecology or behavior,

506 suggesting gaps in our understanding of the physiological function of taste receptor genes and

507 the selective pressures that influence their maintenance (13, 73).

508 We suggest that biological stoichiometry can provide additional insight into the

509 ecological contexts under which natural selection is likely to favor, or possibly disfavor, the

510 evolution and maintenance of functional taste capabilities across diverse animal lineages. By

511 focusing on differences in the elemental composition of plants and animals, as well as major

512 animal lineages, we can use biological stoichiometry to understand how selective pressure for

513 certain taste capabilities might vary among species on the basis of their trophic position

514 (herbivore, omnivore, or carnivore) and phylogenetic relationships (i.e., mammals and insects).

515 For example, differences in organismal stoichiometry among trophic levels, such as those

516 depicted in Fig. 1, suggest that animals are likely to experience differential selective pressure for

517 certain taste capabilities based on their trophic position. The narrower variation in the elemental

518 composition of animals compared with plants (Fig. 1) indicates that stoichiometric imbalances

519 for N and P, and potentially Na and Ca (Fig. 2), are likely to be greater, and more common, for

520 animals that feed at lower trophic positions, such as generalist herbivores and omnivores, than

521 those that feed at higher trophic levels (i.e., obligate carnivores). Put another way, carnivores are

522 more likely to encounter food that has a chemical composition matching their own, thereby

523 reducing selective pressure on the genes that allow them to differentiate among foods of varying

524 composition. Conversely, generalist herbivores and omnivores consume foods that are much

525 more heterogeneous in their elemental composition, and less likely to be in balance with their

526 own chemical composition, meaning they should experience stronger selective pressure on their

527 gustatory systems than strict carnivores.

528 When taste receptor genes are no longer under strong selective pressure, they should be

529 more likely to accumulate deleterious mutations that result in their becoming broken

530 pseudogenes. To illustrate this concept, we turn to data compiled by Feng and Zhao (73) who

531 used available genomes of 48 mammals to demonstrate widespread pseudogenization of taste

532 receptor genes, including those for the T1R1 umami receptor. Based on principles of biological

533 stoichiometry, we would expect herbivores to experience much greater dietary N-deficiencies

534 than carnivores, because of the higher and more variable C:N ratio of plants than animals (Fig.

535 1), and therefore stronger selective pressure to maintain functioning umami (amino acid) taste

536 receptor genes. Though Feng and Zhao (73) concluded that there was no common dietary factor

537 that explained the pseudogenization of the T1R1 umami taste receptor across those 48 mammal

538 species, it is true that a higher proportion of predatory species (carnivores, insectivores,

539 piscivores) lacked a functioning umami receptor gene (60% defective or absent, n=10) than

540 omnivores (13.5% defective or absent, n=22), or herbivores (31% defective or absent, n=16) in

541 their dataset. Thus, there is some limited evidence that N-associated umami taste is more

542 common for consumers that experience large dietary imbalances in N-supply, suggesting that

543 biological stoichiometry may be useful for understanding general patterns regarding the loss of

544 some taste functions in animals. However, we caution that biological stoichiometry may not

545 accurately predict the presence or absence of specific tastes for individual species as those are

546 undoubtedly shaped by each species unique evolutionary history. For example, the loss of amino

547 acid taste in some herbivores, like the giant panda, would not be expected when viewed from the

548 perspective of biological stoichiometry. In the case of the giant panda, loss of umami taste is

549 likely a reflection of reduced selective pressure on taste receptors resulting from the reduced

550 dietary breadth of giant pandas, which almost exclusively consume bamboo, relative to ancestral

551 species (11). Additionally, retention of umami taste in some carnivores, such as felids, may be an

552 indication of predator N-limitation, despite the appearance of relatively stoichiometrically

553 balanced prey (47, 74). In such cases, N-limitation may be driven by consumer metabolism, as

554 many predators rely on gluconeogenesis to convert N-rich amino acids into glucose for cellular

555 respiration due to the lack of carbohydrates in animal prey (47).

556 Though interspecific variation in elemental composition tends to be lower among

557 different animal than plant species, there are several elements that vary considerably among

558 animal species based on broad scale phylogenetic relationships. For example, P and Ca tend to

559 be more concentrated in vertebrates than insects, a result of the presence P- and Ca-rich skeletal

560 bones in the former (Fig. 1). As such, vertebrates may be more likely to experience, or

561 experience greater magnitude, P and Ca imbalances than insects within any given trophic level,

562 making it more likely they should find P and Ca taste attractive. Indeed, Ca is attractive at some

563 concentrations to a diverse suite of vertebrates (birds, amphibians, mammals), but appears to be

564 strictly aversive in insects, which are depleted in Ca relative to plants (Fig. 1; 32, 52). Thus, we

565 believe that biological stoichiometry can inform our understanding of the phylogenetic

566 distribution of taste by revealing differential elemental imbalances between animals and their

567 food both among major animal lineages as well as consumer trophic levels, while accepting that

568 the full suite of taste capabilities possessed by any species will ultimately be influenced by that

569 species’ evolutionary history.

570

571 6. Limitations of biological stoichiometry and nutritional geometry

572 We do not intend to provide a comprehensive review of taste biology research, as several such

573 papers exist and remain useful (i.e., 5, 6), nor do we intend to provide a comprehensive synthesis

574 of either biological stoichiometry theory or nutritional geometry. Rather, our intent here is to

575 highlight linkages between taste and animal nutrition using two prominent nutritional

576 frameworks in ecology. We do, however, draw support for this effort from a comprehensive

577 review of the taste literature, which in some cases (i.e., for P- and Ca-related taste) is limited to a

578 relatively small number of species and model systems, including humans, mice and livestock. As

579 such, some of the evidence we present to support our general thesis, that biological stoichiometry

580 and nutritional geometry can inform our understanding of the evolutionary contexts that

581 produced multiple nutritive taste modalities, is anecdotal in nature. We believe that these

582 examples suggest a possible role for taste in driving the documented behaviors, especially in

583 species for which specific taste capabilities have not been previously investigated. It is important

584 that we acknowledge that animals employ other sensory systems, as well as learned nutritional

585 associations, to assess the nutritional content of their foods and that these systems may also

586 contribute to selective feeding behaviors (3, 66). Finally, we acknowledge that both biological

587 stoichiometry and nutritional geometry may not accurately predict the full suite of consumptive

588 taste capabilities for a given species, but nonetheless are useful for generating hypotheses to

589 further understand taste evolution.

590 We have previously discussed that dietary specialization has been linked to reductions in

591 diversity of taste capabilities in some species. In such cases, the presence or absence of taste

592 senses may deviate from expectations based solely on principles of biological stoichiometry and

593 nutritional geometry. In cases of extreme dietary specialization, especially those that result in

594 large nutritional imbalances between the consumer and its food, animals may rely on a diverse

595 suite of adaptations ranging from specialized physiology and digestive morphology, to

596 mutualisms with endosymbiotic microorganisms to reduce the severity of nutritional deficiencies

597 in their diets. For example, some termites that feed on nutritionally poor wood rely on microbial

598 endosymbionts to breakdown cellulose (75), and for microbial N-fixation within the gut (76).

599 Animals that possess such adaptations may circumvent the effects of nutritional limitation on

600 poor quality diets by means other than dietary selectivity informed by gustatory sensing, and

601 therefore may experience little selective pressure to maintain functioning gustatory systems.

602 We have primarily focused our discussion of taste senses on those that promote

603 consumptive, rather than aversive, behavioral responses, though we do discuss aversive

604 responses to high concentrations of compounds that produce consumptive responses under some

605 conditions (i.e., Na, P and Ca salts). This distinction is important given that some aversive tastes,

606 such as bitter and, perhaps to some extent sour, are generally thought to inform consumers that

607 foods may contain toxic compounds, or low pH, that may be harmful if ingested, even in low

608 quantities (5, 6). Thus, aversive tastes encourage consumers to reject foods, and do not

609 necessarily align with our applications of biological stoichiometry theory and nutritional

610 geometry, which we link primarily to consumptive tastes that drive nutrient acquisition. Indeed,

611 aversive tastes may alert consumers to the presence of potentially harmful compounds in foods

612 that may appear nutritious when considered from a strictly biological stoichiometry perspective.

613 Interestingly, animals can, in some cases, learn to tolerate, or even prefer, bitter and sour tastes in

614 their food and drink, particularly when the toxicity of taste compounds is negligible relative to

615 potential nutritional and pharmacological benefits (77). Though aversive responses to bitter and

616 sour tastes are not necessarily consistent with principles of biological stoichiometry, we have

617 documented several examples where aversive tastes for high concentrations of certain elements,

618 such as Na, P and Ca, do align. The aversive responses to highly concentrated supply of Na, P

619 and Ca contrast to the consumptive response that these same elements induce at low and

620 moderate concentrations. These differential responses to low and high levels of dietary Na, P and

621 Ca can be viewed as adaptations that reduce the potential for physiological consequences

622 associated with deviation from strict homeostatic Na, P and Ca regulation, and thus provide

623 examples of aversive taste responses that are consistent with the core principles of biological

624 stoichiometry theory.

625

626 7. Future directions

627 We have provided numerous hypotheses for understanding the evolution of consumptive taste in

628 animals based on the core principles of biological stoichiometry theory and the complementary

629 framework of nutritional geometry. We believe that our review of the taste biology literature,

630 albeit brief and incomplete, has produced compelling evidence that supports the usefulness of

631 these frameworks for informing our understanding of the evolution of taste. Here, we highlight

632 several observations and questions that have emerged from this analysis that point to exciting

633 avenues for future work that we hope will bridge the fields of taste biology, physiology,

634 biological stoichiometry, ecology, and evolution.

635 We have previously acknowledged that support for some of the ideas presented in this

636 manuscript come from relatively small bodies of literature. For example, both P and Ca-

637 associated tastes have been described for a small number of taxa, reflecting the lack of attention

638 for these particular taste capabilities relative to the basic taste modalities recognized in humans

639 (e.g., salty, bitter, sour, sweet, umami). Nevertheless, our analysis suggests that P- and Ca-

640 associated taste may be widespread given the large number of species that display preferential

641 feeding behaviors related to these elements, and provides some guidance on where to look for

642 candidate species that might possess these specific taste capabilities (i.e., herbivores and

643 omnivores for P, vertebrates for Ca). Thus, we believe that both P- and Ca-associated tastes are

644 deserving of increased attention in a wider array of taxa, including invertebrates.

645 In addition to the elements surveyed here (Na, P, N, Ca and C), several trace elements

646 such as Bo, Fl, Fe and Zn are present in much greater concentrations in animal tissues than the

647 foliar tissue of plants, suggesting the potential for consumptive taste associations for those

648 elements. For example, it is worth noting that Zn is essential for the proper functioning of taste

649 receptors themselves (78). Future research may assess the potential for gustatory sensing of those

650 elements in animals, followed by efforts to characterize the genetic and molecular nature of those

651 gustatory systems. Further research is also warranted to better understand potential taste

652 associations for essential biochemicals that might not be predicted by biological stoichiometry.

653 Gustatory sensing of essential biochemicals, particularly those that are not endogenously

654 synthesized by animals, should be highly favored leading to the evolution of taste associations.

655 Rapid advances in molecular and genetic techniques have enabled great insight into the

656 chemical and genetic underpinnings of taste in both vertebrate and invertebrate models. We now

657 know the specific receptor genes involved in producing many different taste modalities. This

658 opens the possibility of analyzing the distribution of different taste modalities throughout the

659 animal kingdom and thereby facilitating the extension of taste research to a larger number of

660 species, including a greater survey of invertebrate fauna. This might prove especially useful for

661 ecologists seeking to understand how animals employ behavioral or life history strategies to

662 minimize nutritional imbalances in natural settings, thereby influencing interspecific predator-

663 prey interactions, food web and ecosystem dynamics. For example, a species ability to taste a

664 particular nutrient is likely an important trait that influences the role of that species in nutrient

665 cycling. We encourage continued efforts to understand the genetic and molecular bases of

666 different taste capabilities, particularly those exhibited by invertebrates. Identification of specific

667 taste receptors and their associated genes will facilitate broad phylogenetic analyses to reveal the

668 taxonomic distribution and evolutionary history of individual taste capabilities.

669 Data Accessibility

670 Data sharing not applicable to this article as no datasets were generated and all data presented

671 (Table 1, Fig.1, Fig.2) can be accessed from the original sources.

672 References

673 1. D. Raubenheimer, S. J. Simpson, D. Mayntz, Nutrition, ecology and nutritional ecology:

674 Toward an integrated framework. Funct. Ecol. 23, 4–16 (2009).

675 2. R. W. Sterner, J. J. Elser, Ecological Stoichiometry: The Biology of Elements from Molecules

676 to the Biosphere (Princeton University Press, New Jersey USA, 2002).

677 3. S. J. Simpson, D. Raubenheimer, The nature of nutrition: A unifying framework from animal

678 adaptation to human obesity (Princeton University Press, New Jersey USA, 2012).

679 4. B. Lindemann, Receptors and transduction in taste. Nature. 413, 219–225 (2001).

680 5. D. A. Yarmolinsky, C. S. Zuker, N. J. P. Ryba, Common Sense about Taste: From Mammals

681 to Insects. Cell. 139, 234–244 (2009).

682 6. P. A. S. Breslin, An evolutionary perspective on food and human taste. Curr. Biol. 23, R409–

683 R418 (2013).

684 7. M. Kirino, J. Parnes, A. Hansen, S. Kiyohara, T. E. Finger, Evolutionary origins of taste buds:

685 Phylogenetic analysis of purinergic neurotransmission in epithelial chemosensors. Open

686 Biol. 3, 1-12 (2013).

687 8. J. J. Elser, D. R. Dobberfuhl, N. a Mackay, J. H. Schampel, Size , and Life Stoichiomet

688 Toward a unified view of cellular and ecosystem processes. Bioscience. 46, 674–684

689 (1996).

690 9. J. J. Elser, W. F. Fagan, R. F. Denno, D. R. Dobberfuhl, A. Folarin, A. Huberty, S. Interlandi,

691 S. S. Kilham, E. McCauley, K. L. Schulz, E. H. Siemann, R. W. Sterner, Nutritional

692 constraints in terrestrial and freshwater food webs. Nature 408, 578–580 (2000).

693 10. D. Raubenheimer, S. J. Simpson, The geometry of compensatory feeding in the locust. Anim.

694 Behav. 45, 953-964 (1993).

695 11. H. Zhao, J. R. Yang, H. Xu, J. Zhang, Pseudogenization of the umami taste receptor gene

696 Tas1r1 in the giant panda coincided with its dietary switch to bamboo. Mol. Biol. Evol.

697 27, 2669-2673 (2010).

698 12. P. Jiang, J. Josue, X. Li, D. Glaser, W. Li, J. G. Brand, R. F. Margolskee, D. R. Reed, G. K.

699 Beauchamp, Major taste loss in carnivorous mammals. Proc. Natl. Acad. Sci. U. S. A.

700 109, 4956–4961 (2012).

701 13. M. Antinucci, D. Risso, A matter of taste: Lineage-specific loss of function of taste receptor

702 genes in vertebrates. Front. Mol. Biosci. 4, 1–8 (2017).

703 14. E. Sperfeld, N. D. Wagner, H. M. Halvorson, M. Malishev, D. Raubenheimer, Bridging

704 Ecological Stoichiometry and Nutritional Geometry with homeostasis concepts and

705 integrative models of organism nutrition. Funct. Ecol. 31, 286-296 (2017).

706 15. T. R. Anderson, D. Raubenheimer, D. O. Hessen, K. Jensen, W. C. Gentleman, D. J. Mayor,

707 Geometric Stoichiometry: Unifying Concepts of Animal Nutrition to Understand How

708 Protein-Rich Diets Can Be “Too Much of a Good Thing.” Front. Ecol. Evol. 8, 1-12

709 (2020).

710 16. S. M. Wilder, P. D. Jeyasingh, Merging elemental and macronutrient approaches for a

711 comprehensive study of energy and nutrient flows. J. Anim. Ecol. 85, 1427–1430 (2016).

712 17. P. C. Frost, M. A. Evans-White, Z. V. Finkel, T. C. Jensen, V. Matzek, Are you what you

713 eat? Physiological constraints on organismal stoichiometry in an elementally imbalanced

714 world. Oikos 109, 18-28 (2005).

715 18. J. Persson, P. Fink, A. Goto, J. M. Hood, J. Jonas, S. Kato, To be or not to be what you eat:

716 Regulation of stoichiometric homeostasis among autotrophs and heterotrophs. Oikos.

717 119, 741–751 (2010).

718 19. J. J. Elser, M. E. S. Bracken, E. E. Cleland, D. S. Gruner, W. S. Harpole, H. Hillebrand, J. T.

719 Ngai, E. W. Seabloom, J. B. Shurin, J. E. Smith, Global analysis of nitrogen and

720 phosphorus limitation of primary producers in freshwater, marine and terrestrial

721 ecosystems. Ecol. Lett. 10, 1135–1142 (2007).

722 20. 1. K. LeBauer, D; Treseder, Nitrogen Limitation of Net Primary Productivity. Ecology. 89,

723 371–379 (2008).

724 21. 1. M. Filipiak, K. Kuszewska, M. Asselman, B. Denisow, E. Stawiarz, M. Woyciechowski,

725 J. Weiner, Ecological stoichiometry of the honeybee: Pollen diversity and adequate

726 species composition are needed to mitigate limitations imposed on the growth and

727 development of bees by pollen quality. PloS ONE. 12, 1-31 (2017).

728 22. P. D. Jeyasingh, J. M. Goos, S. K. Thompson, C. M. Godwin, J. B. Cotner, Ecological

729 stoichiometry beyond redfield: An ionomic perspective on elemental homeostasis. Front.

730 Microbiol. 8 (2017).

731 23. E. Sperfeld, D. Martin-Creuzburg, A. Wacker, Multiple resource limitation theory applied to

732 herbivorous consumers: Liebig’s minimum rule vs. interactive co-limitation. Ecol. Lett.

733 15, 142–150 (2012).

734 24. M. Kaspari, J. S. Powers, Biogeochemistry and Geographical Ecology: Embracing All

735 Twenty-Five Elements Required to Build Organisms. Am. Nat. 188, S62–S73 (2016).

736 25. T. R. Anderson, M. Boersma, D. Raubenheimer, Stoichiometry: Linking elements to

737 biochemicals. Ecology. 85, 1193–1202 (2004).

738 26. S. J. Simpson, D. Raubenheimer, The nature of nutrition: A unifying framework. Aust. J.

739 Zool. 59, 350–368 (2011).

740 27. C. P. Richter, Total self-regulatory functions in animals and human beings. Harvey Lectures.

741 38, 63-103 (1942-1943).

742 28. S. J. Simpson, R. M. Sibly, K. P. Lee, S. T. Behmer, D. Raubenheimer, Optimal foraging

743 when regulating intake of multiple nutrients. Anim. Behav. 68, 1299-1311 (2004),

744 29. M. Boersma, J. J. Elser, Too much of a good thing: On stoichiometrically balanced diets and

745 maximal growth. Ecology, 87, 1325-1330 (2006).

746 30. D.A. Denton, The hunger for salt: An anthropological, physiological and medical analysis.

747 (Springer-Verlag, New York, 1982)

748 31. E. A. Hallem, A. Dahanukar, J. R. Carlson, Insect odor and taste receptors. Annu. Rev.

749 Entomol. 51, 113–135 (2006).

750 32. Y. C. D. Chen, A. Dahanukar, Recent advances in the genetic basis of taste detection in

751 Drosophila. Cell. Mol. Life Sci. 77, 1087–1101 (2020).

752 33. V. G. Dethier, The taste of salt. Am. Sci. 65, 744–751 (1977).

753 34. S. R. Smedley, T. Eisner, Sodium uptake by puddling in a moth. Science, 270, 1816-1818

754 (1995).

755 35. J. R. Blair-West, J. P. Coghlan, D. A. Denton, J. F. Nelson, E. Orchard, B. A. Scoggins, R.

756 D. Wright, K. Myers, C. L. Junqueira, Physiological, morphological and behavioural

757 adaptation to a sodium deficient environment by wild native Australian and introduced

758 species of animals. Nature. 217, 922–928 (1968).

759 36. G. E. Belovsky, P. A. Jordan, Sodium Dynamics and Adaptations of a Moose Population. J.

760 Mammal. 62, 613–621 (1981).

761 37. J. J. Elser, K. Acharya, M. Kyle, J. Cotner, W. Makino, T. Markow, T. Watts, S. Hobbie, W.

762 Fagan, J. Schade, J. Hood, R. W. Sterner, Growth rate-stoichiometry couplings in diverse

763 biota. Ecol. Lett. 6, 936–943 (2003).

764 38. K. Theiler, H. H. Green, P. J. Du Toit, 1924. Phosphorus in the livestock industry. Journal of

765 the Department of Agriculture, South Africa, 8, 460-504 (1924).

766 39. H. H. Green, Perverted appetites. Physiol. Rev. 5, 336-348 (1925).

767 40. M. G. Tordoff, Phosphorus taste involves T1R2 and T1R3. Chem. Senses. 42, 425–433

768 (2017).

769 41 S. Ibanez, A. Millery, M. D’ottavio, R. Guilhot, E. Vesin, Phosphorus-rich grasshoppers

770 consume plants high in nitrogen and phosphorus. Ecol. Entomol. 42, 610–616 (2017).

771 42. J. M. Sweeny, H. E. Seibert, C. Woda, J. Schulkin, A. Haramati, S. E. Mulroney, Evidence

772 for induction of a phosphate appetite in juvenile rats. Am. J. Physiol. - Regul. Integr.

773 Comp. Physiol. 275, 1358–1365 (1998).

774 43. K. Plath, M. Boersma, Mineral limitation of zooplankton: Stoichiometric constraints and

775 optimal foraging. Ecology. 82, 1260–1269 (2001).

776 44. D. J. Holcombe, D. A. Roland, R. H. Harms, The ability of hens to regulate phosphorus

777 intake when offered diets containing different levels of phosphorus. Poult. Sci. 55, 308–

778 317 (1976).

779 45. G. Nelson, M. A. Hoon, J. Chandrashekar, Y. Zhang, N. J. P. Ryba, C. S. Zuker, Mammalian

780 sweet taste receptors. Cell. 106, 381–390 (2001).

781 46. R. F. Chapman, Contact Chemoreception in Feeding by Phytophagous Insects. Annu. Rev.

782 Entomol. 48, 455–484 (2003).

783 47. K. D. Kohl, S. C. P. Coogan, D. Raubenheimer, Do wild carnivores forage for prey or for

784 nutrients?: Evidence for nutrient-specific foraging in vertebrate predators. BioEssays. 37,

785 701–709 (2015).

786 48. M. G. Tordoff, Calcium: Taste, intake, and appetite. Physiol Rev. 81, 1567–1597 (2001).

787 49. M. G. Tordoff, H. Shao, L. K. Alarcón, R. F. Margolskee, B. Mosinger, A. A. Bachmanov,

788 D. R. Reed, S. McCaughey, Involvement of T1R3 in calcium-magnesium taste. Physiol.

789 Genomics. 34, 338–348 (2008).

790 50. A. S. Gabriel, H. Uneyama, T. Maekawa, K. Torii, The calcium-sensing receptor in taste

791 tissue. Biochem. Biophys. Res. Commun. 378, 414–418 (2009).

792 51. S. Niknafs, E. Roura, Nutrient sensing, taste and feed intake in avian species. Nutr. Res. Rev.

793 31, 256–266 (2018).

794 52. Y. Lee, S. Poudel, Y. Kim, D. Thakur, C. Montell, Calcium Taste Avoidance in Drosophila.

795 Neuron. 97, 67-74 (2018).

796 53. M. G. Tordoff, M. A. Sandell, Vegetable bitterness is related to calcium content. Appetite.

797 52, 498–504 (2009).

798 54. F. Magliocca, A. Gautier-Hion, Mineral content as a basis for food selection by western

799 lowland gorillas in a forest clearing. Am. J. Primatol. 57, 67–77 (2002).

800 55. A. D. Walde, D. K. Delaney, M. L. Harless, L. L. Pater, Osteophagy by the desert tortoise

801 (Gopherus agassizii). Southwest. Nat. 52, 147–149 (2007).

802 56. T. C. Esque, E. L. Peters, Ingestion of bones, stones, and soil by desert tortoises. Fish Wildl.

803 Res. 13, 105–112 (1994).

804 57. M. G. Tordoff, L. K. Alarcón, S. Valmeki, P. Jiang, T1R3: A human calcium taste receptor.

805 Sci. Rep. 2, 3–6 (2012).

806 58. P. C. Frost , J. P. Benstead, W. F. Cross, H. Hillebrand, J. H. Larson, M. A. Xenopoulos, T.

807 Yoshida T, Threshold elemental ratios of carbon and phosphorus in aquatic

808 consumers. Ecol. Lett. 9, 774–779 (2006).

809 59. A. Sclafani, The sixth taste? Appetite. 43, 1–3 (2004).

810 60. 1. C. M. Hladik, B. Simmen, Taste perception and feeding behavior in nonhuman

811 primates and human populations. Evol. Anthropol. 5, 58–71 (1996).

812 61. J. H. Brown, J. F. Gillooly, A. P. Allen, V. M. Savage, G. B. West, Toward a metabolic

813 theory of ecology. Ecology. 85, 1771–1789 (2004).

814 62. A. Miriyala, S. Kessler, F. C. Rind, G. A. Wright, Burst Firing in Bee Gustatory Neurons

815 Prevents Adaptation. Curr. Biol. 28, 1585-1594 (2018).

816 63. M. W. Baldwin, Y. Toda, T. Nakagita, M. J. O’Connell, K. C. Klasing, T. Misaka, S. V.

817 Edwards, S. D. Liberles, Evolution of sweet taste perception in hummingbirds by

818 transformation of the ancestral umami receptor. Science. 345, 929–933 (2014).

819 64. P. Besnard, P. Passilly-Degrace, N. A. Khan, Taste of fat: A sixth taste modality? Physiol.

820 Rev. 96, 151–176 (2015).

821 65. D. Gaillard, F. Laugerette, N. Darcel, A. ElYassimi, P. PassillyDegrace, A. Hichami, N.

822 A. Khan, J. Montmayeur, P. Besnard, The gustatory pathway is involved in

823 CD36mediated orosensory perception of longchain fatty acids in the mouse. FASEB

824 J. 22, 1458–1468 (2008).

825 66. A. Sclafani, K. Ackroff, Role of gut nutrient sensing in stimulating appetite and conditioning

826 food preferences. Am. J. Physiol. Regul. Integr. Comp. Physiol. 302, R1119-R1131

827 (2012).

828 67. D. F. Smith, S. Balagura, Taste and physiological need in vitamin C intake by guinea

829 pigs. Physiol. Behav. 14, 545–549 (1975).

830 68. T. Yasuo, T. Suwabe, N. Sako, Behavioral and Neural Responses to Vitamin C Solution in

831 Vitamin C-deficient Osteogenic Disorder Shionogi/Shi Jcl-od/od Rats. Chem. Senses. 44,

832 389–397 (2019).

833 69. P. Jiang, J. Josue, X. Li, D. Glaser, W. Li, J. G. Brand, R. F. Margolskee, D. R. Reed, G. K.

834 Beauchamp, Reply to Zhao and Zhang: Loss of taste receptor function in mammals is

835 directly related to feeding specializations. Proc. Natl. Acad. Sci. U. S. A. 109, 19104

836 (2012).

837 70. Z. Liu, G. Liu, F. Hailer, P. Orozco-terWengel, X. Tan, J. Tian, Z. Yan, B. Zhang, M. Li,

838 Dietary specialization drives multiple independent losses and gains in the bitter taste gene

839 repertoire of Laurasiatherian Mammals. Front. Zool. 13 (2016.

840 71. H. Zhao, D. Xu, S. Zhang, J. Zhang, Genomic and genetic evidence for the loss of umami

841 taste in bats. Genome Biol. Evol. 4, 73–79 (2012).

842 72. P. Feng, J. Zheng, S. J. Rossiter, D. Wang, H. Zhao, Massive losses of taste receptor genes in

843 toothed and baleen whales. Genome Biol. Evol. 6, 1254–1265 (2014).

844 73. P. Feng, H. Bin Zhao, Complex evolutionary history of the vertebrate sweet/umami taste

845 receptor genes. Chinese Sci. Bull. 58, 2198–2204 (2013).

846 75. J. A. Breznak, A. Brune, Role of microorganisms in the digestion of lignocellulose by

847 termites. Annu. Rev. Entomol. 39, 453–487 (1994).

848 76. C. J. Potrikus, J. A. Breznak, Nitrogen-Fixing Enterobacter agglomerans Isolated from Guts

849 of Wood-Eating Termites1. 33, 392–399 (1977).

850 77. E. D. Capaldi, G. J. Privitera, Decreasing dislike for sour and bitter in children and adults.

851 Appetite. 50, 139–145 (2008).

852 78. R. I. Henkin, Zinc in taste function. Biol. Trace Elem. Res. 6, 263–280 (1984).

853 79. M. J. Vanni, P. B. McIntyre, D. Allen, D. L. Arnott, J. P. Benstead, D. J. Berg, Å. Brabrand,

854 S. Brosse, P. A. Bukaveckas, A. Caliman, K. A. Capps, L. S. Carneiro, N. E. Chadwick,

855 A. D. Christian, A. Clarke, J. D. Conroy, W. F. Cross, D. A. Culver, C. M. Dalton, J. A.

856 Devine, L. M. Domine, M. A. Evans-White, B. A. Faafeng, A. S. Flecker, K. B. Gido, C.

857 Godinot, R. D. Guariento, S. Haertel-Borer, R. O. Hall, R. Henry, B. R. Herwig, B. J.

858 Hicks, K. A. Higgins, J. M. Hood, M. E. Hopton, T. Ikeda, W. F. James, H. M. Jansen, C.

859 R. Johnson, B. J. Koch, G. A. Lamberti, S. Lessard-Pilon, J. C. Maerz, M. E. Mather, R.

860 A. McManamay, J. R. Milanovich, D. K. J. Morgan, J. M. Moslemi, R. Naddafi, J. P.

861 Nilssen, M. Pagano, A. Pilati, D. M. Post, M. Roopin, A. T. Rugenski, M. H. Schaus, J.

862 Shostell, G. E. Small, C. T. Solomon, S. C. Sterrett, Ø. Strand, M. Tarvainen, J. M.

863 Taylor, L. E. Torres-Gerald, C. B. Turner, J. Urabe, S. I. Uye, A. M. Ventelä, S. Villeger,

864 M. R. Whiles, F. M. Wilhelm, H. F. Wilson, M. A. Xenopoulos, K. D. Zimmer, A global

865 database of nitrogen and phosphorus excretion rates of aquatic animals. Ecology. 98,

866 1475 (2017).

867 80. H. J. M. Bowen, Trace elements in biochemistry (Academic Press, New York, USA, 1966).

868 81. B. Markert, Establishing of “Reference Plant” for inorganic characterization of different

869 plant species by chemical fingerprinting. Water, Air, Soil Pollut. 64, 533–538 (1992).

870 Competing Interests

871 The authors have no competing interests.

872 Author contributions

873 LMD, BWT, BJR and RRD were primarily responsible for the conceptual development of the

874 manuscript and MGT provided guidance and expertise on discussion of taste biology. LMD

875 compiled relevant data and created figures. All authors made writing contributions during

876 manuscript preparation and were involved in editing and revision of various drafts.

877 Acknowledgements

878 This manuscript was improved by comments from P Jiang, JA Balik, R Irwin, and S Jordt. This

879 work was supported by the U.S. National Science Foundation [grant number 1556914].

880

881 Figure legends

882 Fig 1. Carbon:nitrogen and carbon:phosphorus ratios of biomass across multiple trophic levels.

883 Circles represent median values, error bars indicate ranges of C:N and C:P observed for

884 organisms within each trophic level. Autotroph data are from Elser et al. (9), where Autotroph

885 (Plants) includes measurements of foliar chemistry of terrestrial autotrophs (for C:N, n = 406; for

886 C:P, n = 413) and Autotroph (Plankton*) represents nutrient chemistry of seston from freshwater

887 lakes (for C:N, n = 267; for C:P, n = 273). Seston contains some detritus and heterotrophic

888 biomass (i.e., bacteria, protozoa) but is typically dominated by phytoplankton (9). Consumer

889 stoichiometry data are from Vanni et al. (79) and represent primarily aquatic animals. Data

890 include 190 families from 9 animal phyla (Herbivore and Detritivore, n = 168; Omnivore, n =

891 77; Predator, n = 163).

892 Fig 2. The ratio of biomass concentrations of biologically “essential” elements in animals and

893 plants for all elements which are at least 0.1% of dry mass in animals. Mass ratio is calculated as:

894 XA/ XP; where XA and XP represent the % of dry mass for element X in animals and plants,

895 respectively. Thus, values >1 indicate elements that are more concentrated in animal than plant

896 tissues, 1 indicates equal concentrations and values <1 indicate greater concentrations in plant

897 tissue. Circles containing elemental symbols represent the average of data from mammals

898 (orange), insects (black) and fish (blue) compiled by Bowen (80). Data for plants are from

899 Markert (81) and represent the elemental composition of foliar material.

900

901

902

903

904 Fig. 1

905

906 Fig. 2

907 Table 1. The tissue concentrations (as % of dry mass) of essential elements in plants, animals,

908 insects, fish, and mammals. Values for animals represent the average of elemental concentrations

909 in insects, fish and mammals. Animal data are from Bowen (1980). Plant data are from Markert

910 (81) and are representative of foliar tissue.

Percent of total dry mass Atomic Element Mass Plants Animals Insect Fish Mammals C 12.11 44.50 46.83 44.60 47.50 48.40 O 15.99 42.50 26.63 32.30 29.00 18.60 N 14.01 2.50 10.80 12.30 11.40 8.70 H 1.01 6.50 6.90 7.30 6.80 6.60 Ca 40.08 1.00 3.52 0.30 2.00 8.50 P 30.97 0.20 2.60 1.70 1.80 4.30 K 39.09 1.90 1.02 1.10 1.20 0.75 Na 22.99 0.02 0.61 0.30 0.80 0.73 S 32.97 0.30 0.56 0.44 0.70 0.54 Cl 35.45 0.20 0.35 0.12 0.60 0.32 Si 28.09 0.10 0.21 0.60 7.00E-03 1.20E-02 Mg 24.31 0.20 9.8E-02 0.08 0.12 0.10 F 18.99 0.0002 9.5E-02 0.14 0.05 Zn 65.41 0.01 2.1E-02 0.04 8.00E-03 1.60E-02 Fe 55.85 0.02 1.3E-02 0.02 3.00E-03 1.60E-02 Cu 63.55 1.00E-03 2.0E-03 5.00E-03 8.00E-04 2.40E-04 B 10.81 4.00E-05 1.1E-03 2.00E-03 2.00E-04 Mn 54.94 0.02 3.7E-04 1.00E-03 8.00E-05 2.00E-05 Ni 58.69 1.50E-04 3.7E-04 9.00E-04 1.00E-04 1.00E-04 Se 78.96 2.00E-06 1.7E-04 1.70E-04 Mo 95.94 5.00E-05 8.7E-05 6.00E-05 1.00E-04 1.00E-04 I 126.9 3.00E-04 7.8E-05 9.00E-05 1.00E-04 4.30E-05 Co 58.93 2.00E-05 5.0E-05 7.00E-05 5.00E-05 3.00E-05 Cr 51.99 1.00E-05 2.5E-05 3.00E-05 2.00E-05 V 50.94 5.00E-05 2.3E-05 1.50E-05 1.40E-05 4.00E-05