Because of anthropogenic increases in atmospheric CO2 content, there is a need to understand how organisms sense and respond to CO2 variation. An important distinction is whether CO2 responses result from direct effects of CO2 on signal-transduction pathways, enzyme catalysis, or regulatory processes, as opposed to indirect, secondary responses that are a consequence of the direct effects. In plants, direct effects occur because rising CO2 A) increases the activity of Ribulose-1,5-bisphopshate carboxylase/oxygenase (Rubisco) via its role as a substrate for RuBP carboxylation and its inhibition of RuBP oxygenation; B) reduces stomatal aperture; C) alters mitochondrial respiration; and D) possibly reduces transcription of genes for Rubisco activase and carbonic anhydrase. Because of these direct effects, the carbon and water balance of plants is altered leading to secondary effects on growth, resource partitioning and defense compound synthesis. Reduced investment in photosynthetic protein is one of the characteristic acclimation responses of plants to high CO2. This is modulated by increased carbohydrate levels, probably in concert with hormone signals from the roots. Roots are hypothesized to be the main control points for CO2 acclimation because they are well situated to integrate the carbohydrate status of the plant. In higher fungi, development of the mushroom fruiting body is inhibited at high CO2, but the mechanism is poorly known. Fungal CO2 sensing may serve to position the spore-bearing tissue above the soil boundary layer to ensure effective spore dispersal. The animals that are most sensitive to anthropogenic CO2 enrichment are insects. Many insects have a well-developed ability to sense CO2 variation as a means of locating food. Unlike plants, insects have CO2 receptors that can detect variation in CO2 as low as 0.5 ppm. However, the sensitivity of these receptors is reduced in atmospheres with double or triple current levels of CO2, indicating some insect species may be threatened by rising atmospheric CO2.
Because of fossil fuel use and deforestation, humans have increased the CO2 level in the atmosphere from 270 ppm two hundred years ago to 370 ppm today. In the coming centuries, atmospheric CO2 levels are expected to double, and may triple before peaking. The human effect on atmospheric CO2 content is superimposed upon natural CO2 variations controlled by long-term changes in the global carbon cycle. Over the past 400,000 yr, atmospheric CO2 has oscillated between 180 and 300 ppm in concert with glacial/interglacial cycles (Petit et al., 1999). During interglacial periods, CO2 levels peaked near 300 ppm for about 10,000 to 15,000 yr. During glacial phases, low CO2 levels persisted for much longer. CO2 levels were below 240 ppm for two-thirds of the past 400,000 yr and over 80,000 yr of this period corresponded to CO2 levels below 200 ppm (Petit et al., 1999; Sage and Coleman, 2001). Because of these generally low CO2 levels of the past half-million years, plants are probably adapted to much lower CO2 than exists today (Sage and Coleman, 2001).
In addition to the natural cycles of recent geological time, a long-term decline in atmospheric CO2 has occurred over the past 100 million years, from levels estimated to be 3 to 10 times above current levels (Berner, 1994). Associated with this decline has been a general cooling and drying of the global climate, and an extensive radiation of plant and animal life forms (Prothero, 1994). For example, most angiosperm families appeared during the time when CO2 declined from the high levels that existed during the time of the dinosaurs (Wolfe, 1997). This reduction in CO2 has been implicated as a contributing agent for numerous evolutionary trends, notably the rise of plants that use the C4 photosynthetic pathway, and evolutionary specialization of numerous faunal lines (Ehleringer et al., 1991, 1997; MacFadden, 1997; Sage, 1999). In addition, numerous water conservation features in plants may have arisen as CO2 levels fell, because of the reduced water use efficiency that accompanies reduction in atmospheric CO2 (Sage and Cowling, 1999).
Over the years, the focus of CO2-related studies has been the response of plants to rising CO2, largely because of their ability to acquire CO2 through photosynthesis. It is now recognized that a wide variety of organisms directly respond to CO2 variation, and plants are not necessarily the most responsive, nor are they the most threatened, by atmospheric CO2 enrichment. Atmospheric CO2 variation affects organisms through direct interactions with enzymes, sensory molecules, or regulatory systems, and indirectly as a result of secondary responses to these direct effects. To understand the range of effects of CO2, it is necessary to identify the responses that are direct, versus those that are secondary in nature, or indirect. The purpose of this paper is to review our understanding of the mechanisms by which organisms directly respond to variation in CO2, and how these responses cascade to affect higher order processes within an organism.
EFFECTS OF ATMOSPHERIC CO2 VARIATION ON PLANTS
The direct effects of rising atmospheric CO2 on plants are due to three major processes: a) direct modulation of the activity of Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubico), the primary CO2 fixing enzyme in photosynthesis, b) sensing of CO2 levels by stomatal guard cells, and c) modulation of mitochondrial respiration (Fig. 1; Sage and Reid, 1994; Drake et al., 1997). The consequence of each of these direct effects is to alter water and carboydrate levels in plants, which in turn modulate a diverse set of secondary responses ranging from stimulation of growth, alteration of biomass allocation, increased chemical defense, and greater density of mycorrhizal and nitrogen fixing associations (Sage, 1995; Lindroth, 1996; Drake et al., 1997; Rogers et al., 1999). Direct sensing of CO2 has been suggested to control expression of photosynthetic genes coding for carbonic anhydrase and Rubisco activase (Fig. 1; Coleman, 2000; von Caemmerer and Quick, 2000), but this control is relatively minor and does not have a major effect on photosynthetic performance at CO2 levels near the current ambient and above. Instead, the principal mechanism by which photosynthesis responds to CO2 is via the direct stimulation of Rubisco activity.
The initial reaction in photosynthetic CO2 fixation is the carboxylation of RuBP; however, Rubisco also can oxygenate RuBP in the first step of a process termed photorespiration (Fig. 2; Jordan and Ogren, 1984). Oxygenation of RuBP produces phosphoglycolate, a useless two-carbon compound. Metabolism of phosphoglycolate to recover its carbon requires ATP and reducing power, and results in the loss of previously fixed CO2 at a ratio of 1 C lost for every two oxygenation events (Sharkey, 1988). The uptake of oxygen via RuBP oxygenation, and the loss of CO2 during phosphoglygolate metabolism is termed photorespiration because of its superficial similarity with mitochondrial respiration. CO2 and O2 compete for Rubisco active sites, and thus the ratio of photorespiration to photosynthesis is highly dependent on the ratio of CO2 to O2 in the chloroplast (Fig. 3; Sage, 1999). In terrestrial plants, Rubisco has a much higher specificity for CO2 relative to O2, but atmospheric O2, levels are 580 times higher than CO2 in the current atmosphere, and over 1,000 times higher in the atmosphere of the late-Pleistocene. This abundance of O2 relative to CO2 compensates for the greater CO2 specificity of Rubisco such that at warmer temperatures, photorespiration can be significant in atmospheres of today and recent geological time. Rising temperature reduces both the specificity of Rubisco for CO2, and the CO2/O2 ratio in the chloroplast solution (Jordan and Ogren, 1984). As a result, photorespiration rises with increasing temperature, except at very high CO2 levels (>1,000 ppm) where oxygenase activity is largely suppressed at all temperatures (Fig. 3). In large part because of the temperature dependence of photorespiration, the proportional stimulation of photosynthesis by rising CO2 increases with temperature (Fig. 4). As a result, the CO2 effect on the earth's flora is predicted to be greatest in warmer climates of low latitude, and during the warmest peaks of the growing season at temperate latitude (Long, 1991).
At the molecular level, CO2 has not been shown to directly interact with any receptor or step in the signal-transduction pathway of photosynthetic genes. Limited evidence has been presented that elevated CO2 inhibits the transcription of genes coding for carbonic anhydrase and Rubisco activase (Majeau and Coleman, 1996; Mate et al., 1996; Eckard et al., 1997). It is not clear whether these responses are directly mediated by CO2, or indirectly controlled by other factors, for example, carbohydrate status in the leaf (Coleman, 2000). Both enzymes enhance photosynthesis at low CO2 levels but are less important at elevated CO2. Carbonic anhydrase reversibly catalyzes the conversion of CO2 to bicarbonate and in doing so can speed the diffusion of inorganic carbon into the chloroplast (Coleman, 2000). Rubisco activase enhances the activation state of Rubisco by removing inhibitors from the catalytic site (Portis, 1992). At low CO2, plants are unable to maintain a high activation state of Rubisco in the absence of an abundant amount of Rubisco activase, while at elevated CO2, less activase is required for complete Rubisco activation (Mate et al., 1996; Eckard et al., 1997).
SECONDARY RESPONSES OF THE PHOTOSYNTHETIC APPARATUS
In C3 plants, photosynthesis is limited by one of three general processes: the capacity of Rubisco to consume RuBP, the capacity of electron transport and the Calvin cycle reactions to regenerate RuBP, and the capacity of starch and sucrose synthesis to regenerate phosphate for photophosphorylation (Sharkey, 1985; Sage, 1994). The degree to which these processes limit photosynthesis varies with CO2 level. At low CO2 (370 ppm ambient CO2 and below), Rubisco capacity exerts greater control over photosynthesis; at elevated CO2, processes contributing to either RuBP or Pi regeneration are limiting (Harley and Sharkey, 1991). As a result, a shift in CO2 level produces a metabolic imbalance between the capacities of the different components of photosynthesis. This leads to a regulatory response that modulates activation states of key enzymes, but does not initially alter levels of existing enzyme (Sage, 1990). For example, tripling of atmospheric CO2 causes Rubisco capacity to become excessive, and the response within the leaf is to reduce the activation state of Rubisco (Sage et al., 1988, 1989). Similarly, CO2 reduction leads to a limitation of Rubisco capacity, and an excess of light harvesting and sucrose synthesis capacity (Sharkey et al., 1988; Sage, 1990; Sage et al., 1990). Regulation within the leaf reduces the capacity of light harvesting in CO2-depleted atmospheres by activating carotenoid systems that quench excess light energy, while excess sucrose synthesis capacity is reduced by deactivation of sucrose-phosphate synthase and cytosolic fructose-bisphosphatase (Sharkey et al., 1988; Sage and Reid, 1994). These regulatory responses can be profound; for example, half of the Rubisco active sites can be switched off by large CO2 increases (Sage et al., 1988, 1989). They are, however, not directly related to CO2 acting as a regulatory molecule. Instead, the regulatory system responds to imbalances in the supply and demand for energy which CO2 variation perturbs through its effect on the carboxylation capacity of Rubisco (Sage, 1990; von Caemmerer and Quick, 2000).
Long-term regulation (photosynthetic CO2 acclimation)
The photosynthetic stimulation that initially follows high CO2 exposure generally leads to a significant enhancement of leaf carbohydrate levels and a burst of growth (Jitla et al., 1997; Centritto et al., 1999). Within a few days to weeks of CO2 enhancement, leaf protein levels may begin to decline, photosynthetic capacity at a given CO2 concentration drops, and the growth enhancement is reduced if not eliminated altogether (Stitt, 1991; Moore et al., 1997, 1999). These changes are also coupled to increases in root growth relative to shoot growth such that root to shoot ratios increase (Rogers et al., 1999). Together, these responses are termed the acclimation response to rising CO2. Much work has recently focused on understanding the signal-transduction system that controls the acclimation response.
The strength of the acclimation response is dependent upon nutrient status, plant age, photoperiod and source to sink ratio (Arp, 1991; Stitt, 1991; Roitsch, 1999; Stitt and Krapp, 1999). Low nutrients, advanced age, extended photoperiod, and sink loss all enhance the degree to which plants acclimate to CO2 enrichment. For example, older plants with low nutrient supply and few sinks often acclimate to the point where photosynthesis and growth enhancements disappear (Arp, 1991; Sage, 1994; Sims et al., 1998a). By contrast, young, rapidly growing plants with high nutrition and abundant sinks typically show little if any acclimation to rising CO2 (Sage, 1994; Drake et al., 1997; Sage and Coleman, 2001). Notably, many environmental factors that perturb nutrient supply and source to sink ratios in the absence of CO2 variation promote the same acclimation responses that occurs in elevated CO2 (Stitt, 1991). A common element in each of these responses is an enhancement of leaf carbohydrate content, and it is now recognized that hyper-accumulation of carbohydrates is the major signal controlling plant acclimation to elevated CO2, as well as acclimation to a variety of treatments that alter source to sink ratios (Stitt, 1991; Smeekens and Rook, 1997; Moore et al., 1999).
Carbohydrate signaling in photosynthetic acclimation
In leaves, most of the major photosynthetic genes respond to carbohydrate status. When carbohydrate levels are high, transcription of genes coding for Rubisco, chlorophyll-protein complexes of the light harvesting antennae, and chloroplast electron transport proteins is repressed (Moore et al., 1997, 1999). In sink tissues, multiple genes coding for the utilization and storage of carbohydrate are expressed, leading to enhanced activity of growth and storage tissues (Koch, 1996). The means by which tissues sense variation in carbohydrate status remains unclear, although multiple sensory pathways are postulated. The most discussed model of sugar sensing is hexokinase signaling. Hexoses (glucose and fructose) appear to be major activators of regulatory hexokinases in a manner that has not yet been identified (Jang et al., 1997; Smeekens, 2000). Once activated, hexokinase is hypothesized to initiate a signal-transduction sequence involving a number of intermediate complexes (termed SNF complexes for sucrose non fermenting after the mutant yeast in which they were first identified), that modulate transcription of specific genes. Hexokinase signaling in plants is similar to hexokinase sugar-sensing in animal and fungi, because regulatory hexokinases in each have high sequence similarity and the intermediate SNF complexes appear homologous (Jang et al., 1997; Halford and Hardie, 1998; Smeekens, 2000). Key differences are present, however. In animal and fungi, the primary purpose of hexokinase signaling is to coordinate sugar metabolism with availability of carbohydrates in the diet, and thus to avoid wasteful expression of enzymes if carbohydrates are deficient. In plants, the situation is more complex in that carbohydrate signals are required to coordinate protein investment within photosynthetic cells with metabolism in remotely located sinks, as well as transport between those sinks and local storage pools (Moore and Sheen, 1999). To accomplish this, plants have multiple sugar sensing mechanisms in addition to the hexokinase system. Sucrose, for example, directly modulates transcription of sucrose transporters (Smeekens and Rook, 1997; Chiou and Bush, 1998). Membrane-bound receptors for specific carbohydrates are postulated, and acetate levels may have an important signaling function (Koch et al., 2000). Invertase activity is correlated with high CO2 acclimation (Moore et al., 1998), indicating an important role for this enzyme in sugar signaling. Moore et al. (1999) postulate that invertases hydrolyze excess sucrose to glucose and fructose, which are then rephosphorylated by hexokinase, leading in turn to hexokinase activation and signal production. In this manner, sucrose levels could be linked to the hexokinase regulatory system.
Interaction with phytohormones, nutrients and light
While much of the research on CO2 acclimation has focused upon the role of carbohydrates, it is increasingly clear that in plants, carbohydrates status is but one component of a complex interaction of controls that include phytohormones, light and nutrients (Roitsch, 1999; Smeekens, 2000). Carbohydrates alone cannot explain the range of acclimation responses to elevated CO2 as has been demonstrated by Sims et al. (1998b) using soybean plants grown in either normal or CO2-enriched atmospheres (Table 1). In their experiments, plants were grown in growth chambers in atmospheres of either 250 or 1,000 ppm CO2. From each plant, a single, attached leaf was placed in a smaller chamber and grown for ten days at either 250 or 1,000 ppm CO2. Single leaves grown at 1,000 ppm showed little sign of photosynthetic acclimation when attached to plants growing at 250 ppm, because they had similar Rubisco levels as leaves exposed to 250 ppm on plants grown at 250 ppm (Table 1). By contrast, leaves in 250 ppm CO2, but attached to plants at 1,000 ppm, showed marked acclimation, with 60% reduction in Rubisco levels, similar to what the high-CO2-exposed leaves on high-CO2-grown plants exhibited. Notably, carbohydrate levels did not correlate with the degree of acclimation, and tended to be lower in leaves of the high-CO2 grown plants. As demonstrated by Sims et al. (1998b), entire plants, not simply leaves, acclimate to high CO2.
Where acclimation within a leaf is mediated by whole-plant factors, long-distance signals molecules are likely involved. In plants, the major long-distance signals include the phytohormones (ABA, ethylene, giberellins, auxin and cytokinins), xylem pH, and the nutrients nitrogen and calcium (Jackson, 1993; Davies et al., 1994; Davies and Gowing, 1999). High ABA in leaves may act in concert with ethylene and sucrose to stimulate the reduction in leaf protein levels observed during the acclimation response (Roitsch, 1999; Smeekens, 2000). ABA in particular shares common sensing elements as sugars, as shown by Arabidopsis mutants that are insensitive to both ABA and sucrose (Huisjer et al., 2000; Laby et al., 2000). Cytokinins may reduce the effect of sugars on acclimation processes in leaves, as elevated cytokinins enhance greening and leaf protein levels, and antagonize senescence responses (Wingler et al., 1998).
Long-distance signals influencing CO2 acclimation probably originate in the sinks in the plant, particularly reproductive and root sinks. Conceptually, the roots are better situated than the source leaves to monitor the carbohydrate status of the plant, because they can integrate the activity of all source organs. By contrast, photosynthetic acclimation in response to leaf carbohydrate levels alone could inhibit photosynthesis in individual leaves that are well positioned within the canopy for high photosynthetic performance. For example, a leaf at the top of the canopy in ideal conditions could accumulate high levels of carbohydrate, perhaps much more than the rest of the canopy. Plants rely heavily upon these leaves for much of their carbon supply during the year, and should, if anything, invest more resources into these well-positioned leaves. (They do, as indicated by canopy patterns of nitrogen distribution, Anten et al., 2000). Repressing photosynthesis of a few high-performance leaves would be counterproductive, as it would represent a waste of light and would be a lost opportunity for photosynthetic carbon gain. Alternatively, if all leaves photosynthesize at a high rate and whole plant carbohydrate status becomes excessive, then it would be necessary to reduce photosynthetic investment in order to free internal resources for the acquisition of other limiting resources, such as mineral nutrients (Bloom et al., 1985). How might roots act as the master control of carbohydrate status and CO2 acclimation within plants?
Root meristems are well known to modulate shoot growth and responsiveness to a wide range of environmental factors including water, nutrients and temperature (Jackson, 1993; Wilkinson and Davies, 2002). The response to drying soil is a well documented example of long-distance signaling mediated by roots (Davies and Gowing, 1999). Upon sensing a mild water deficit, roots release ABA to the xylem stream as well as altering xylem pH and calcium levels. Enhanced levels of ABA induce stomatal closure, slow cell division and expansion, accelerate senescence, and reduce export of carbon from leaves (Trewavas and Jones, 1991; Smeekens, 2000). Similarly, roots can increase the flux of auxins, cytokinins, and GAs under more favorable conditions (Jackson, 1993). Cytokinins modulate nutrient levels in leaves and antagonize senescence, and thus may counteract the tendency of elevated carbohydrates to accelerate leaf senescence (Wingler et al., 1998). From these considerations, a simple conceptual model is proposed to explain how roots poise leaves for high or low carbohydrate sensitivity during CO2 acclimation (Fig. 5).
STOMATAL RESPONSE TO CO2 VARIATION
The second major direct effect of CO2 is upon stomatal aperture. As atmospheric CO2 levels increase, stomatal conductance (an index of stomatal aperture) declines, reducing transpiration and increasing the water use efficiency of photosynthesis (Morison, 1987; Mott, 1990). Drought stress and low atmospheric humidity enhance the degree of stomatal closure that is caused by CO2 enrichment, so that the CO2 effect on stomatal aperture is greatest in arid conditions (Schulze et al., 1987). Stomata respond to intercellular rather than ambient CO2 levels, so the effect of rising atmospheric CO2 is because intercellular CO2 increases (Mott, 1990).
Despite this long-standing knowledge of stomatal responses to CO2, the means by which CO2 is sensed by guard cells remains uncertain. Guard cells do have functional chloroplasts, indicating internal photosynthetic activity could be a means by which guard cells sense variation in CO2. Significantly, however, stomata respond to CO2 enrichment in the dark when photosynthesis is not operating, indicating there is a direct sensing capability (Morison, 1987). Four models have been proposed to explain CO2 sensing by stomata (Assmann, 1999). Two involve malate formation as a result of increased PEP carboxylation by PEP carboxylase. Rashke (1979) proposed elevated malate levels reduce cellular pH and the permeability of the tonoplast and plasmalemma of guard cells. In support of this model, reduction in cytosolic pH of guard cells by elevated levels of organic acids activates K+-efflux (Rashke, 1979). Hedrich and co-workers (Hedrich et al., 1994) argue that increased malate production enhances extracellular malate levels, which in turn activate anion efflux channels at the plasmalemma. As a result, osmotically-active anions leak out of the cell, causing the cell's turgor pressure to decline, closing the stomata. Instead of malate, elevated zeaxanthin levels within the chloroplast may modulate guard cell CO2-sensitivity in the light. Zeaxanthin content in guard cells is correlated with CO2 level, but how this shift is transduced to the plasma-membrane is not known (Zhu et al., 1998). The fourth model involves CO2 induced changes in cytosolic calcium levels, possibly in a synergistic manner with ABA. This final model is the only model that may involve a specific CO2 receptor, which upon interacting with CO2 would send a signal that could feed into signal-transduction pathways mediated by intracellular calcium (Assmann, 1999). Finally, multiple modes of CO2 perception may exist, raising the possibility that more than one of these models are correct. Zeaxanthin signaling, for example, may be an additional means of CO2-sensing as it only is present in the light, and thus could not account for dark CO2 responses (Zhu et al., 1998).
As with the photosynthetic apparatus, stomata can acclimate to long-term variation in CO2 supply. CO2 acclimation of stomata involves either a relative change in the aperture at the growth CO2 level, or a change in the sensitivity to a given level of CO2 variation (Santrucek and Sage, 1996). In Chenopodium album, growth at 750 ppm CO2 resulted in a weakening of the stomatal response when measurement CO2 levels were experimentally varied between 300 and 800 ppm (Fig. 6). The consequence of CO2 acclimation by the stomata was a recovery of the ability of the guard cells to respond to CO2 variation above growth CO2 levels. In C. album plants grown at current atmospheric CO2 levels, stomata were insensitive to increasing CO2 levels above 750 ppm, but stomata of plants grown at 750 ppm could respond to CO2 up to at least 1,200 ppm. In general, however, consistent responses of stomatal conductance to rising CO2 are not apparent, and strong interactions with water status, leaf age, and humidity are present (Sage, 1994; Assmann, 1999). Despite this variation, it is clear that in most species, stomatal acclimation does not override the closing response, and transpiration rates remain depressed in high CO2 environments (Johnson et al., 1993; Sage and Santrucek, 1996; Anderson et al., 2001). The improvement in water status of the plant and soil associated with reduced stomatal aperture leads to a wide range of higher order responses that can have profound effects on overall performance, soil properties and competitive ability (Sage, 1995; Owensby et al., 1996).
The third direct effect of CO2 on plant performance is a modulation of mitochondrial respiration. Doubling of atmospheric CO2 is reported to cause respiration to vary between a 20% increase and a 60% decrease, with a 15 to 20% reduction being the average (Drake et al., 1999; Gonzalez-Meler and Siedow, 1999). Even within a species such as soybean, the range of responses is between +10% and −30% (Thomas and Griffen, 1994; Bunce, 1995). Respiratory inhibition improves the carbon status of plants and stimulates yield, but the exact mechanism causing the respiratory shift is unclear (Drake et al., 1999; Amthor, 2000). A doubling of atmospheric CO2 is reported to directly inhibit cytochrome C oxidase and succinate dehydrogenase, although the magnitudes of these effects are too small to account for the larger respiratory shifts (Gonzalez-Meler and Siedow, 1999). Instead, CO2 effects on the demand for respiratory ATP may account for a greater proportion of the responses that have been observed. Indirectly, increases in carbohydrate supply can stimulate respiration, but this is more of a short-term response that tends to weaken with long-term exposure to high CO2 (Azcon-Bieto, 1983; Amthor, 2000).
FUNGAL CO2 SENSING
Because they largely live in soil or the remains of other organisms where CO2 levels are often very high (>1,000 ppm), fungal mycelia are not very sensitive to rising CO2, except indirectly through alteration in carbohydrate supply provided by higher plants in parasitic or mutualistic associations (Cooke and Whipps, 1993; Rogers et al., 1999). By contrast, pileus (fruiting body, or the cap of the mushroom) development in numerous basidiomycete species is directly sensitive to moderate CO2 enrichment. CO2 levels above 1,000 ppm repress expansion of the pileus and stimulate stipe growth in a wide variety of commercially-grown species, leading to spindly mushrooms in high CO2 conditions (Table 2; Turner, 1977; Elliot, 1994). The mechanism for the CO2 effect is unclear but has been associated with reduced synthesis of cell wall constituents (Elliot, 1994). In oyster mushrooms (Pleurotus spp.), increasing CO2 two to three times above ambient reduces expansion of the spore-bearing cap, and causes lengthening of the stalk of the mushroom cap (Fig. 7). CO2 increases above 5,000 ppm produce spindly oyster mushrooms with elongated stalks and little pileus expansion, and in some species of Pleurotus, CO2 increases to 10,000 ppm largely arrest pileus formation (Fig. 7 and Sage, unpublished). Apparently, certain mushroom species use CO2 as a cue to position their fruiting bodies in the boundary layer above the soil (Turner, 1977). Where the boundary layer is thick and air movement low, CO2 generated by soil respiration accumulates. Under these conditions, spore dispersal is also poor, and the mushroom may elongate its stalk to elevate the gills into a more favorable dispersal location, which would be indicated by reduced CO2 levels. Because oyster mushrooms respond to CO2 increases predicted for the future, there may be some loss of fecundity if elevated CO2 weakens their ability to optimally position their spore-bearing tissue.
Because they lack photosynthesis, there has been far less research on animal responses to CO2 variation than plant responses. Most global change research addressing animals and elevated CO2 has addressed secondary responses of herbivores to the changes in forage quality caused by CO2 enrichment (Lindroth, 1996; Coviella and Trumble, 1999). Animals are directly affected by high CO2, but for most, the effects are not significant until CO2 levels well above 1,000 ppm (Table 3; Nicolas and Sillans, 1989). Respiratory poisoning by high CO2 affects all animals, but does not generally happen until CO2 levels rise above 50,000 ppm. Social insects such as ants, bees and termites use high CO2 concentration to sense location and activity within the hive (Stange, 1996). For example, if CO2 levels in a beehive become too high (>5,000 ppm), poor ventilation is indicated and bees begin fanning their wings. Many soil insects locate food sources by detecting CO2 emitted by roots and other food items, but the CO2 levels detected are well above atmospheric values (Nicolas and Silens, 1989; Coviella and Trumble, 1999). CO2 also regulates the aperture of the insect spiracle (a breathing pore on the thorax and abdomen), but significant effects of atmospheric CO2 on spiracle aperture occur at higher CO2 levels than predicted for the future (Miller, 1974). From a global change perspective, the most important response of animals to rising CO2 is the olfactory response of insects, particularly that used by haematophagous (blood-sucking) and phytophagous (leaf eating) insects to locate prey.
Olfactory sensing of CO2 in insects
Large numbers of insect species sense CO2 variation by olfactory sensilla on their mouth parts, head capsules, antennae and legs (Bogner et al., 1986; Nicolas and Sillans, 1989; Stange, 1996). A specific CO2 receptor is present in the insect sensilla, and this initiates a change in the membrane potential of the surrounding tissue, which in turn propagates an action potential in the neural axons, sending a signal to the central nervous system (Fig. 8). High CO2 sensitivity occurs when the numbers of CO2 sensilla increase, and when they are flattened to enhance the surface area exposed to the CO2 source (Stange, 1996). CO2 sensing is generally coupled with other chemical cues to allow discrimination between food sources and false signals (Stange, 1996). In the absence of complementary chemical and visual signals, CO2 signals are ignored.
Insects feeding on the blood of higher animals locate their prey by following the high CO2 gradient produced by its breath. CO2 levels in the breath of warm-blooded animals approaches 50,000 ppm, and can be easily resolved at 3 to 10 m by the olfactory sensors on the insect's antennae and labial palps. Mosquitoes, biting flies, and ticks are some of the well-known blood parasites that use CO2 as a directional cue (Bogner, 1992; Takken and Knols, 1999). In blood parasites, a typical pattern is for relatively low levels of CO2 (10 to 300 ppm) to activate flight or searching behavior, while a relatively high concentration (>1,000 ppm) is required to initiate directional locomotion. For instance, in ticks, perception of CO2 changes as small as 10 ppm initiate questing behavior (the lifting of legs to sample the air—CO2 sensilla on ticks are located on leg tarsi). Directed locomotion in ticks does not follow until CO2 levels rise above 3,000 ppm (Stange, 1996).
The organisms with the most acute sense of CO2 perception are the lepidopteran herbivores (moths and butterflies) who can discriminate small changes (<1 ppm) in CO2 brought about by photosynthetic activity of their plant foods (Stange, 1992, 1996, 1997). Caterpillars of the moth Helicoverpa armigea discriminate between CO2 levels of 160 and 800 ppm, indicating an ability to distinguish between photosynthetic and respiring plant parts (Rasch and Rembold, 1994). This aids in locating fruit as opposed to leaf food supplies. In adult Helicoverpa armigea, CO2 sensilla on labial palps are able to discriminate 0.5 ppm on a background of 350 ppm (Stange, 1992). The best-studied CO2 sensing system in lepidopterans is that of the prickly-pear cactus moth, Cactoblastis cactorum (Stange et al., 1995; Stange, 1996, 1997). Cactoblastis cactorum locates its food supply by sensing nighttime reductions in CO2 caused by the cladodes (leafy stems) of prickly pear cacti (Opuntia spp). Opuntia use CAM photosynthesis, which involves nighttime opening of the stomates and CO2 fixation into a temporary pool of four-carbon acids. Cactoblastis species are important herbivores of Opuntia spp. and are well-known because they were introduced to control a severe Opuntia invasion in Australia early in the 20th century. CO2 is sensed by Cactoblastis for two purposes. First, CO2 gradients above the soil are used to determine the flight location with respect to the ground, apparently to position the moth at the height where the cladodes will be easily detected. Second, CO2 gradients of just a few ppm on a background of 350 ppm are detected in the vicinity of the cladodes. The ability of Cactoblastis to resolve these gradients is so acute that it can distinguish the more robust cladodes that have higher concentrations of photosynthetic enzymes and hence are more nutritious. This allows the insect to select the best locations on which to lay its eggs.
Olfactory CO2 perception in a high-CO2 world
In most insect species capable of using CO2 as an olfactory signal, anthropogenic increases in atmospheric CO2 are not likely to have major impacts because the CO2 levels being sensed are higher than what is predicted to occur. The survival of many lepidopteran species, however, may be threatened because their ability to resolve a given CO2 signal is weakened as the background CO2 level increases (Stange, 1997). In addition, the CO2-receptor neurons of highly responsive species such as Cactoblastis cactorum are adapted for high sensitivity at low CO2 but approach signal saturation at double current atmospheric CO2 levels (Stange, 1996, 1997). CO2-receptor neurons are temperature compensated, and this capacity is also weakened as the background CO2 level increases (Stange and Wong, 1993). Blood-sucking insects may also be adversely affected, mainly in their ability to resolve small signals in the activation response. As a result, they may be more sluggish in initiating search behavior. However, because host animals produce high CO2 concentrations in the breadth, the ability to find the host once the activation response has been initiated will likely be little affected (Stange, 1996).
CONCLUSIONS—GLOBAL CHANGE AND THESENSING OF CO2
The CO2 rise of the past century, and what is predicted for the coming century, represent a return to atmospheric CO2 levels not seen for 10 to 20 million years, if not longer. Most of the research associated with rising CO2 has focused on plants because of the central role of photosynthesis. In plants, responses to high CO2 are relatively benign or improve performance by improving the carbon and water balance of the plant. By contrast, the ability of certain insects to detect food by variation in atmospheric CO2 is compromised because their ability to resolve small gradients declines as the background level of CO2 increases. With large enough increases in atmospheric CO2, the ability of these creatures to sense CO2 could be reduced enough to threaten their survival, unless they can adapt and recover an ability to detect subtle changes in CO2. It is not known whether this will be possible in an atmosphere with a high background level of CO2. Because a large number of insect species potentially use CO2 as a cue, the consequences of their lost CO2 sensitivity could be great. Even in remote, relatively pristine areas of the globe, substantial stress could be imposed on insect populations that lose their ability to use CO2 to identify food sources. Global insect biodiversity could be further eroded, with unpredictable consequences as the loss of insect populations cascade to other trophic levels.
Ironically, however, it may be the insect sensory system that may provide key insights into how plants sense CO2. In plants, no CO2 receptors have been identified, although they may be central to the ability of the stomata to respond to CO2. An intriguing possibility is that CO2 receptors in insect sensilla may be homologous with plant CO2 receptors, should they exist. Identification of the genes for the insect CO2 receptors could allow for sequence comparisons with homologues from plant genomes, possibly leading to identification of plant CO2 receptors. With the completion of the genome for Drosophila and Arabidopsis, such possibilities may soon become reality.
From the symposium Plant/Animal Physiology presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 3–7 January 2001, at Chicago, Illinois.