Source–sink balance and carbon allocation below ground in plants exposed to ozone

Authors

  • Christian P. Andersen

    Corresponding author
    1. Western Ecology Division, National Health and Environmental Effects Research Laboratory, Office of Research and Development, US Environmental Protection Agency, 200 SW 35th St, Corvallis, Oregon 97333, USA
    Search for more papers by this author

  • Note: The information in this article has been funded by the U.S. Environmental Protection Agency. It has been subjected to the Agency's peer and administrative review, and it has been approved for publication as an EPA document. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.

Author for correspondence: Christian P. Andersen Tel: +1541 7544791 Fax: +1541 7544799 Email: andersen.christian@epa.gov

Abstract

Contents

  • Summary 213

  • I. Introduction 213
  • II. Source–sink model: carbohydrate signaling 214
  • III. Effect of ozone on above-ground sources and sinks 216
  • IV. Decreased allocation below ground 218
  • V. Carbon flux to soils 220
  • VI. Soil food web 223
  • VII. Summary, conclusions and future research 223
  • Acknowledgements 223

  • References 223

Summary

The role of tropospheric ozone in altering plant growth and development has been the subject of thousands of publications over the last several decades. Still, there is limited understanding regarding the possible effects of ozone on soil processes. In this review, the effects of ozone are discussed using the flow of carbon from the atmosphere, through the plant to soils, and back to the atmosphere as a framework. A conceptual model based on carbohydrate signaling is used to illustrate physiological changes in response to ozone, and to discuss possible feedbacks that may occur. Despite past emphasis on above-ground effects, ozone has the potential to alter below-ground processes and hence ecosystem characteristics in ways that are not currently being considered.

I. Introduction

‘The most curious result obtained appears to me to be that relating to the effect of a highly ozonized atmosphere upon the roots of plants.’ – M. Carey Lea 1864.

Professor M. Carey Lea was perhaps the first to study the effects of a highly oxidized atmosphere on plant growth. A member of the chemistry faculty at the University of Pennsylvania, Professor Lea used 3-l bell jars as exposure vessels in studies of wheat and corn, and generated O3 by ‘the action of sulphuric acid upon chameleon mineral’ (potassium permanganate). The exposure was regenerated every 2–3 d as ‘the presence of vegetation would tend to destroy the O3 rapidly.’ Ozone reduced ‘moulding’ and caused roots to grow upward, away from the bathing nutrient and water solution.

There has been increased understanding of the effects of O3 on plants since 1864, particularly during the last half of the 20th century. Recognition that O3 was a problem in agricultural crops dates at least to 1944 when certain vegetables were observed to have leaf injury (Middleton et al., 1950; Richards et al., 1958). Although originally attributed to sulfur dioxide, it was later recognized that other components of smog were causing the damage. Soon oxidant damage was recognized as a potential stress in sensitive species throughout the US and in Europe (Heggestad & Middleton, 1959; Daines et al., 1960; Bell & Cox, 1975; Posthumus, 1976; Ro-Paulsen et al., 1981). Since that time, emphasis has often been on readily observable effects of O3, such as leaf injury, growth and yield, including factors leading to variation in response. This has led to a substantial database on the effects of O3 on plant growth and development.

Ozone is only one of many stresses present in natural systems that can lead to shifts in ecosystem structure. An example to illustrate this point is found in southern California, where unique topographical and climatic conditions have combined with air pollutants in the Los Angeles basin to result in elevated O3 exposure to forests in the San Bernardino mountains during the last half-century (Miller et al., 1982; Miller et al., 1989). Ponderosa pine (Pinus ponderosa) and Jeffery pine (Pinus jeffreyi) were most sensitive to O3, and therefore were weakened by chronic exposure. Weakened trees were more susceptible to bark beetle (Dendroctonus brevicomis) and root rot (Fomes annosus), and O3-tolerant species such as shrubs and oak were favored as pine mortality increased (Miller et al., 1982). It is important to note that O3 was a predisposing factor for stresses that were already present in this ecosystem. The interaction of stresses affecting San Bernardino forests is still under investigation, particularly the role of N deposition in conjunction with O3 (Fenn et al., 1996; Takemoto et al., 2001).

The relative inaccessibility of plant roots has hampered efforts to understand effects of O3 below ground. Current levels of O3 are capable of altering the timing and quantity of carbon flux to soils, and therefore are affecting interactions in the rhizosphere. We still have very limited understanding of how O3 affects interactions of roots with soil organisms, and no idea how these changes alter soil physical and biological properties in ecosystems. Ozone may cause greater disruption of processes below ground than above, and these changes may occur before changes are observed above ground (Hofstra et al., 1981). Past emphasis on shoot responses, for example, photosynthesis, foliar injury, and reduced yield, has diverted our attention from a discussion of below-ground effects, which may in fact be more critical than above-ground effects in determining the long-term consequences of O3 exposure to ecosystems.

Improved understanding of whole-plant response to O3 is fundamental to identifying changes that may be occurring below ground, and further, to predicting possible consequences of exposure on ecosystem structure and function. Soil ecosystems are dependent on carbon inputs from plants, and any stress that alters carbon flux to soil could change interactions among soil organisms. Understanding the role of O3 on below-ground processes is complicated by: the broad time scales of response, ranging from minutes to years; the trophic complexity associated with soil organismic groups; and our general lack of understanding of processes controlling resource acquisition and use in both individual plants and in ecosystems. These factors have made it difficult to identify the effects of O3 on soil, and hence ecosystem, processes.

The goal of this article is to discuss whole-plant processes leading to decreased carbon allocation below ground, and to discuss possible consequences of decreased allocation on soil processes and properties. Information is drawn from both tree and crop literature to identify common mechanisms of response to O3 that may lead to changes in natural ecosystems. Changes in the quality or quantity of carbon flux to soil are very likely to influence interactions among soil foodweb organisms, potentially altering carbon retention, mineralization and important soil properties. Whether or not these effects are transient or long-term remains to be determined, but there is sufficient evidence to indicate that O3 has the potential to alter soil properties and hence ecosystem structure and function in ways that are not currently being considered.

II. Source–sink model: carbohydrate signaling

Carbon allocation patterns are very complex in perennial plants owing to the multiple-age structure of leaves and roots. Farrar & Jones (2000) examined four hypotheses related to carbohydrate allocation and concluded that the ‘shared-control’ hypothesis was most consistent with empirical data from a number of studies. The hypothesis states that carbon allocation to various sinks is controlled both by sink demand (activity and size) and source control of photosynthate production.

Carbohydrate levels affect gene regulation and provide a mechanism for control of resource distribution among various sources and sinks (Koch, 1996; Sheen et al., 1999). Carbohydrate regulated genes have both direct and indirect roles in sugar metabolism, providing a way to adjust allocation and growth in response to stress and changes in environmental signals. Metabolic regulation involves monosaccharides, particularly hexoses and hexokinases (Koch, 1996; Moore et al., 1999; Sheen et al., 1999).

Carbohydrate signaling can provide a way to understand changes in allocation in response to O3 stress. In general, carbohydrate depletion up-regulates genes responsible for photosynthesis, mobilization, and export, while carbohydrate abundance up-regulates genes responsible for storage and use (Fig. 1) (Koch, 1996). Through modifications in gene expression therefore, source and sink activities can be effectively altered to allow plants to adjust their growth patterns in response to carbon resource availability and acquisition (Koch, 1996; Farrar & Jones, 2000). Applying this general pattern to activities in source and sink tissues suggests that depletion of carbohydrates stimulates photosynthetic enzymes in source tissues, while increased levels of carbohydrate stimulate storage and use in sink tissues.

Figure 1.

Processes in source and sink tissues that are under carbohydrate control in plants. A downward arrow indicates gene down regulation in response to high or low sugar abundance, while an upward arrow indicates up regulation in response to high or low sugar abundance. Information based on Koch (1996 ).

Carbohydrate signaling also provides a mechanism to respond to changes in available resources such as nitrogen (N). Increased N uptake and inorganic N availability in leaf tissue favors amino acid synthesis over gluconeogenesis, leading to carbohydrate retention in source tissue at the expense of allocation to heterotrophic tissues such as roots (Wingler et al., 1994; Hampp & Nehls, 2001). Decreased allocation to roots leads to down-regulation of ‘use’ genes and reduced root growth. In this fashion, shoot growth is favored over root growth, a common response observed in experiments where N levels are high. Similarly, decreased leaf inorganic N leads to decreased amino acid synthesis and increased carbohydrate availability for transport to heterotrophic tissues including roots. As carbon availability in roots increases, genes involved in storage and use are up-regulated (Fig. 2) (Koch, 1996), which leads to root growth and increased N acquisition. It appears that root proliferation in N-rich zones, which occurs primarily in N-stressed plants, also is controlled in part by long-range signaling pathways involving carbohydrates (Forde, 2002).

Figure 2.

A conceptual diagram of processes and storage pools in sources and sinks that are affected by ozone exposure. A plus (+) denotes an increase in process rate or pool size, a minus (–) denotes a decrease in process rate or pool size, and a plus-minus (+) denotes that both increases and decreases have been reported in response to O 3 . Primary effects in the shoots (1°) are distinguished from secondary effects in roots (2°) since the primary site of ozone action occurs in the leaves. See text for details.

Compounds such as cytokinins, ABA and auxin also influence gene regulation and growth of shoots and roots (Aiken & Smucker, 1996; Forde, 2002). It is possible that carbohydrate concentrations provide a means to control responses to light, nutrient concentrations, and stresses, while intrinsic developmental programs are directed by several plant hormones (Sheen et al., 1999).

Although growth-regulating signals can originate in both sources and sinks including roots, the site of O3 action is in the leaf and therefore the site where physiological changes are initiated (Fig. 2). Since roots are not exposed directly to O3 (Turner et al., 1973; Blum & Tingey, 1977), responses are secondary through changes in allocation and subsequent sink activity. Changes in root activity therefore are thought to be a result of carbohydrate availability rather than a direct effect of O3 on tissue metabolism per se. The distinction between primary and secondary effects is important since it determines the temporal dynamics of compensation and response mechanisms, which appear to occur first in shoots.

The discussion will follow carbon as it is fixed from the atmosphere in the leaves and transported to other plant tissues to meet growth and metabolic needs. Carbohydrate signaling provides a means to understand how O3 affects carbon allocation patterns, which will help to organize the discussion.

III. Effect of ozone on above-ground sources and sinks

Ozone causes physiological changes in leaves that affect source strength, that is, the amount of carbon available for allocation to sink tissues (Fig. 2). Changes include decreased carbon assimilation, increased metabolic costs, and possibly decreased phloem loading.

1. Decreased carbon assimilation

Ozone affects carbon assimilation by altering stomatal conductance, by decreasing the activity and concentration of Ribulose 1,5 bisphosphate carboxylase (Rubisco), and by reducing leaf longevity (Fig. 2) (Dan & Pell, 1989; Pell et al., 1994; Miller et al., 1999; Zheng et al., 2002). Stomatal responses alter the amount of O3 reaching the target site, and are therefore considered avoidance mechanisms.

Ozone decreases stomatal conductance in plants, decreasing O3 uptake but also reducing assimilation (Beckerson & Hofstra, 1979; Dan & Pell, 1989; Mulchi et al., 1992; Weber et al., 1993; Fiscus et al., 1997). Decreased stomatal conductance in response to O3 also has been demonstrated using isotopic analysis of δ13C ratios (Greitner & Winner, 1988; Matyssek et al., 1992), although analysis of water use in conjunction with δ13C ratios suggests shifts in 13C/12C ratios may result from other factors besides changes in stomatal conductance (Tingey et al., 1994).

Decreased stomatal conductance in response to O3 may be due to changes in photosynthesis and internal CO2 concentration (Ci) rather than a direct effect on guard cells (Winner et al., 1988; Fiscus et al., 1997; Torsethaugen et al., 1999; Noormets et al., 2001). Results with Plantago major indicate both direct effects on the stomatal complex as well as indirect effects through changes in Rubisco and internal CO2 (Zheng et al., 2002). Although decreased stomatal conductance reduces further O3 uptake and damage, CO2 fixation would decline, possibly leading to increased photorespiration and production of phosphoglycolate.

Prolonged exposure to O3 also can lead to increased stomatal conductance and loss of stomatal control (Keller & Hasler, 1987; Skarby et al., 1987). Ozone-sensitive (based on foliar injury) ponderosa pine were found to have higher stomatal conductance compared to O3 tolerant seedlings (Beyers et al., 1992; Temple et al., 1992). Sensitive (i.e. symptomatic) ponderosa pine in the field exhibited sluggish stomatal control in late summer (Grulke, 1999). Matyssek et al. (1991) and Maurer & Matyssek (1997) found that O3-exposed plants lost the ability to control water loss, leading to reduced water use efficiency. Delayed closure and loss of stomatal control may involve changes in lignin content in guard cells, altering the relationship between water potential and guard cell closure (Maier-Maercker, 1998). Loss of guard cell sensitivity to water stress compromises plants in two respects, that is, increased exposure to water stress and increased O3 uptake.

The second way O3 can affect assimilation (and hence source strength, Fig. 2) is through decreased protein synthesis associated with photosynthetic enzymes (Pell & Pearson, 1983; Dann & Pell, 1989). In the substomatal cavity, O3 disassociates in the leaf apoplast to form H2O2, organic radicals, and other chemically reactive compounds (Hippeli & Elstner, 1996; Pell et al., 1997; Wohlgemuth et al., 2002). These chemical intermediates, particularly the highly reactive free radicals, are thought to be short-lived since they readily react with proteins and membrane lipids in leaf mesophyll cells (Pell et al., 1997). Chloroplast degradation occurs through a complex series of events starting with damage to mesophyll plasmalemma, and leading to decreased photosynthesis (Sandermann, 1996; Matyssek & Innes, 1999). The effect may be due to increased degradation of Rubisco or decreased synthesis under prolonged exposure. In either case, carbon assimilation and source strength decreases, leading to decreased availability of carbon for export to sink tissues such as roots (Fig. 2).

Ozone reduces assimilation by decreasing leaf longevity and increasing senescence (Miller et al., 1982; Duriscoe & Stolte, 1989; Plocher et al., 1994; Karnosky et al., 1996; Miller et al., 1999) (Fig. 2). Photosynthetic compensation of newer foliage has been observed in some species exhibiting premature senescence, but overall carbon gain is generally reduced (Greitner et al., 1994). From a whole-plant perspective, decreased leaf longevity decreases the plants ‘return on investment’ since assimilation is a function of rate and duration of photosynthesis.

2. Increased metabolic costs

Several processes occur in O3-exposed leaves that lead to increased carbon demand and sink strength in leaves, and thus a reduction in carbon allocation below ground (Fig. 2). Factors that increase the sink strength of leaves for carbon include repair processes (tolerance), as well as synthesis of antioxidants and other secondary compounds (avoidance) (Fig. 3). Carbon costs associated with repair processes are thought to be less than for avoidance mechanisms associated with antioxidant synthesis (Tingey & Andersen, 1991). Nonetheless, tolerance and avoidance mechanisms both increase carbon demands and hence sink strength of shoots.

Figure 3.

Stress avoidance and tolerance processes in foliage of ozone-exposed plants leading to a change in translocation to heterotrophic sink tissues. Repair processes (tolerance) and antioxidant synthesis (avoidance) have an associated carbon cost. Stomatal closure does not have a metabolic cost, but results in reduced assimilation.

Increased Costs Associated with Maintenance and Repair (Tolerance) Ozone stimulates biosynthetic pathways that provide carbon intermediates in support of defense and repair processes (Guderian et al., 1985; Landolt et al., 1997; Dizengremel, 2001), and can lead to increased costs to construct tissues (Tingey et al., 1994). Phenolic production increases and the lignin biosynthetic pathway is stimulated as evidenced by increased enzyme activity of phenylalanine ammonia lyase, cinnamyl alcohol deydrogenase, and other enzymes (Dizengremel, 2001; Saleem et al., 2001). Activity of PEP-carboxylase also increases, suggesting a shift from sucrose production to processes in support of defense and repair (Landolt et al., 1997). Repair leads to increased respiration losses in foliage (Landolt et al., 1997). In P. vulgaris, maintenance respiration increased by as much as 25% in leaves exposed to O3 (Amthor & Cumming, 1988; Amthor, 1988).

Compensatory processes also may increase carbon retention in shoots. In the source–sink model, decreased carbon assimilation and carbohydrate concentrations may lead to up-regulation of Rubisco enzymes (Figs 1 and 2). Increased production and turnover of photosynthetic enzymes could lead to increased C and N demand in leaf chloroplasts. Ozone increased N concentration in shoots of Pinus halepensis (Kytoviita et al., 2001) and ponderosa pine (Andersen et al., 2001). Increased protein synthesis is also supported by observations of increased PEP-carboxylase activity in response to O3 (Dizengremel, 2001).

Increased Costs associated with Antioxidant Synthesis (Avoidance) Plants can avoid O3 stress by synthesizing antioxidant compounds capable of scavenging free radicals before they interact with membrane lipids (Melhorn et al., 1986; Alscher & Amthor, 1988; Tingey & Andersen, 1991). Some secondary metabolites are constitutive while others are inducible (Edwards & Wratten, 1983; Haukioja, 1990). Although inducible resistance is energetically less expensive than constitutive resistance, energy costs are thought to be relatively high compared to tolerance mechanisms, that is, repairing damage after it occurs (Herms & Mattson, 1991; Tingey & Andersen, 1991). Synthesis of antioxidant compounds results in diversion of carbohydrate from other metabolic processes (Dizengremel, 2001).

Recently, endogenous isoprene was identified as an antioxidant that protects against O3 at low concentrations (Loreto & Velikova, 2001). When isoprene synthesis was inhibited, leaf sensitivity to O3 increased. It appeared that H2O2 formation and lipid peroxidation were reduced when isoprene was allowed to form.

The amount of carbon that is available for defense and repair influences the degree of O3 response, and hence, a relationship should exist between response and carbon availability at the whole-plant level (Fig. 2). Tingey et al. (2002) found that P. vulgaris was more sensitive to O3 during pod filling than during vegetative growth. Although others have reported seasonal variation in O3 sensitivity of crops such as P. vulgaris (Bender et al., 1990; Vandermeiren et al., 1995), Tingey et al. (2002) removed flowers to eliminate the carbon sink associated with pod fill and found significantly less foliar injury and biomass loss during O3 treatment when pod fill was eliminated. Yield is often affected by O3, but in some cases the relative strength of sinks does not appear to change with exposure (Drogoudi & Ashmore, 2001).

3. Decreased phloem loading

It has been suggested that O3 impairs phloem loading, possibly due to plasmalemma or plasmodesmatal damage in the mesophyll cells (Rennenberg et al., 1996; Landolt et al., 1997; Grantz & Farrar, 2000) (Fig. 3). Several studies have shown increased soluble sugar concentrations and carbohydrate retention in leaves of O3-exposed plants (Tingey et al., 1976; McLaughlin et al., 1982; Friend & Tomlinson, 1992; Grantz & Farrar, 1999). Ozone increased foliar retention of carbohydrates in ponderosa pine (Tingey et al., 1976), white pine (Pinus strobus) (McLaughlin et al., 1982), and loblolly pine (Pinus taeda) (Friend & Tomlinson, 1992). Ozone inhibited carbon export from leaves of Pima cotton (Gossypium barbadense) to a greater extent than it inhibited photosynthesis (Grantz & Farrar, 1999; Grantz & Yang, 2000). Spence et al. (1990) similarly found an effect of O3 on the allocation of carbon to stems and roots of loblolly pine seedlings in short-term studies using 11C, although direct effects of O3 on phloem loading were not evident. Additional research will be required to evaluate the relationship between phloem loading and O3, and whether symplastic and apoplastic phloem-loading species vary in their sensitivity. This is particularly interesting in light of recent results showing the potential role of endogenous isoprene in phloem loading and its role as an antioxidant capable of protecting plants against O3 damage (Loreto & Velikovea, 2001).

In the source–sink model, decreased phloem loading and increased carbohydrate concentrations would lead to feedback inhibition of photosynthesis (Koch, 1996). Moore et al. (1999) suggest that decreased export of leaf sucrose and subsequent increase in leaf sucrose and sucrose cycling are responsible for photosynthetic down-regulation in plants exposed to elevated CO2. Although via a different mechanism, O3 may result in photosynthetic down-regulation in some species if phloem loading is impaired.

4. Summary: above-ground sources and sinks

The leaves are the direct site of O3 action, and determine source supply of carbohydrate for allocation to other plant parts (Fig. 2). Ozone can alter stomatal function, leading to changes in assimilation and water use efficiency. Lipid peroxidation and plasmalemma damage in mesophyll cells leads to increased repair costs, and initiates a sequence of events leading to chloroplast damage. Activity and concentration of photosynthetic enzymes decline, reducing CO2 fixation and potentially increasing the sink strength for carbohydrates and N in support of protein synthesis (Fig. 2). Inducible antioxidant systems can protect tissues from damage, but require carbon for their synthesis. Increased costs associated with membrane repair, protein/antioxidant synthesis, possible photosynthetic down-regulation, and decreased phloem loading all contribute to the reduction in allocation below ground (Fig. 3).

IV. Decreased allocation below ground

1. Roots

Ozone stress decreases carbon allocation to roots (Manning et al., 1971; McCool & Menge, 1983; McLaughlin & McConathy, 1983; Cooley & Manning, 1987; Gorissen & van Veen, 1988; Spence et al., 1990; Gorissen et al., 1994; Rennenberg et al., 1996; US Environmental Protection Agency, 1996). Since roots are often dependent on current photosynthate for their structural development (van den Driessche, 1978; Ritchie & Dunlap, 1980; Marshall & Waring, 1985; van den Driessche, 1991), carbon-limiting stresses such as O3 can have a rapid and significant effect on root growth. In the source–sink model, decreased carbon allocation affects root growth through down-regulation of ‘use’ genes (Figs 1 and 2) (Koch, 1996).

In many cases decreased allocation to roots in response to O3 occurs quickly (Gorissen & van Veen, 1988; Spence et al., 1990; Andersen & Rygiewicz, 1991, 1995; Gorissen et al., 1991), with subsequent reductions in root biomass occurring within one growing season (US Environmental Protection Agency, 1996). Decreased allocation below ground is often associated with decreased root : shoot ratio, but observed responses in root : shoot ratio are highly variable owing to several factors including intra- and interspecies variation, culture conditions, and ontogenetic drift (Reich, 2002). Root : shoot ratio is a point-in-time measurement that does not include carbon lost to exudation, respiration or turnover. Therefore, biomass and ratios of biomass such as root-shoot ratio do not necessarily reveal physiological changes in response to O3 stress.

Decreased carbon allocation leads to reduced carbohydrate levels and storage pools in O3-exposed plants (Tingey et al., 1976; Ito et al., 1985; Cooley & Manning, 1987; Rebbeck et al., 1988; Andersen et al., 1991; Gorissen et al., 1994; Andersen et al., 1997). Although difficult to quantify changes in the field, Grulke et al. (1998) found decreased coarse, medium and fine root biomass with increased pollutant load across a gradient in southern California. Coarse and fine root starch concentrations also were lowest in mature trees at the most polluted site (Grulke et al., 2001). The effects of O3 could not be completely separated from other known stresses across the pollutant gradient, but it appeared that O3 was an important factor in the patterns observed.

Decreased storage pools can lead to carry-over effects on root growth over time. Decreased carbohydrate storage pools were associated with decreased root growth during the spring following exposure to ozone, even in the absence of additional O3 exposure (Andersen et al., 1991, 1997). Decreased spring root growth was attributed to decreased stored reserves as well as premature loss of older foliage age classes the previous fall. Aside from the loss of photosynthetic surface area associated with premature senescence, early loss of foliage in the fall occurs when allocation to roots is at a maximum in many species (Kozlowski & Pallardy, 1997). Older needle age classes preferentially allocate photosynthate basipetally to stems and roots (Rangnekar et al., 1969; Gordon & Larson, 1970), and their absence in the fall during allocation to root growth and storage, and in the spring during periods of root growth, preferentially impacts roots and root processes.

There is still little understanding of the effects of O3 on root metabolism, although in the source–sink model decreased allocation will lead to down regulation of metabolic processes. As a measure of root metabolic activity, Edwards (1991) found decreased root and soil CO2 efflux during a 2-yr exposure to loblolly pine. Fine root respiration increased in mature red oak exposed to O3, while total soil CO2 efflux increased in the spring and decreased in the summer and fall (Kelting et al., 1995). The authors attributed increased root respiration to increased nutrient uptake in support of increased demands in the shoot. Ozone decreased root system respiration in aspen after 12 wk of exposure, but the decrease was closely associated with decreased root biomass (Coleman et al., 1996). Whether or not other metabolic shifts occur in roots of plants exposed to O3 needs to be examined.

Since the site of action of O3 is in the leaf, physiological changes in the root are considered 2° (Fig. 2). Applying the shared-control model (Farrar & Jones, 2000), O3 stress affects source control of allocation and not sink control since O3 does not penetrate the soil to affect roots directly. However, measurable effects on roots may occur before effects on shoots are observed since shoots have immediate access to carbon for repair and compensation. Mortensen (1998) found decreased root but not shoot growth in Betula pubescens at exposures of 42 nMol mol-1 (applied 12 h d−1), whereas both root and shoot growth were reduced at higher exposures. Chromosomal aberrations were found in root tips of Norway spruce exposed to O3, even in the absence of biochemical changes in needles (Wonisch, 1999). Using relatively high concentrations of O3 (0.15 ppm O3 6 h d−1), Hofstra et al. (1981) found metabolic changes in P. vulgaris root tips before the development of leaf injury. Morphological changes in root tips occurred within 2–3 d, and metabolism declined within 4–5 d of initiation of exposure.

Feedback signals from roots can influence the degree of O3 response. Stolzy et al. (1964) exposed tomato roots (Lycopersicon esculentum) to periods of anerobic conditions and followed a change in leaf susceptibility to O3. An exposure of roots to low oxygen conditions for 3 h did not alter photosynthesis, but it decreased foliar damage when subsequently exposed to O3. In this case, a signal originating in the root appeared to alter leaf sensitivity to O3, possibly hydraulic in nature leading to decreased O3 uptake.

2. Symbionts

Decreased allocation associated with ozone exposure alters N fixation in legumes and actinorrhizal species. Ozone exposure was found to decrease nodulation in a number of species (Manning et al., 1971; Tingey & Blum, 1973). In alder (Alnus serrulata), host root cells of nodules showed cytoplasmic breakdown and lacked organelles when seedlings were exposed to 27 d of O3 (Greitner & Winner, 1989). In N-fixing perennial plants, decreased N fixation and N nutrition may lead to feedbacks that stimulate root growth through increased carbon allocation, leading to additional N acquisition.

It is possible that increased sink strength of roots due to mycorrhizal colonization is sufficient to partially overcome the negative effects of O3 on allocation to roots and hence root–shoot balance (Mahoney et al., 1985). Mycorrhizal symbioses increase the sink strength of roots for carbon, primarily by increasing metabolic activity of the root system but also by increasing the size of the root-hyphal sink (Reid et al., 1983; Rygiewicz & Andersen, 1994; Smith & Read, 1997). Andersen & Rygiewicz (1995) measured carbon allocation to roots of ponderosa pine and found less carbon was allocated to extrametrical hyphae, but the presence of ECM did not offset the negative effects of O3 on allocation below ground.

The presence of mycorrhizal fungi may increase overall carbon demands on the plant, reducing available carbon for defense and repair and increasing O3 sensitivity. Colonization of ‘Troyer’ citrange by the AM fungus Glomus fasciculatus reduced plant mass in the presence of O3, possibly due to the increased sink demands placed on the host by the fungus in combination with O3 (McCool et al., 1979). A similar response was observed in subterranean clover, where the presence of an AM fungus suppressed plant growth even at low concentrations, presumably due to the extra carbon demands on the host by the fungus (Miller et al., 1997).

Stresses such as O3 that limit carbon allocation to roots would be expected to decrease mycorrhizal colonization and alter species-host compatibility (Ho & Trappe, 1981; McCool et al., 1982; Simmons & Kelly, 1989; Adams & O’Neill, 1991; Edwards & Kelly, 1992; Smith & Read, 1997). Even so, several short-term studies have found enhanced mycorrhizal short-root formation under O3 stress. White pine (Pinus strobus) (Stroo et al., 1988), Norway spruce (Rantanen et al., 1994), Northern red oak (Quercus rubra) (Reich et al., 1985), Douglas-fir (Pseudotsuga menziesii) (Gorissen et al., 1991), European silver-fir (Abies alba) (Wöllmer & Kottke, 1990), and Scots pine (P. sylvestris) (Kasurinen et al., 1999) all showed some increase in mycorrhizal presence when exposed to O3. Others have shown minimal or no effects of O3 on mycorrhizas (Mahoney et al., 1985; Meier et al., 1990; Kainulainen et al., 2000). Stroo et al. (1988) found that percent infection increased from 0.02 to 0.06 ppm O3, then decreased from 0.06 to 0.14 ppm; the total number of short roots were unaffected, however. In cases where stimulation was observed, the response was often noted shortly after initiation of exposure, often at relatively low concentrations. Good examples of this transitory response can be found in results with Norway spruce and Scots pine (Rantanen et al., 1994; Kasurinen et al., 1999). In these studies, O3 increased mycorrhizal short roots initially but differences were not evident at the end of the experiment.

One explanation for the short-term, stimulatory processes below ground in response to O3 is the availability and mobilization of stored reserves. It is generally accepted that colonization by ECM is positively correlated with root soluble sugars (Björkman, 1970; Smith & Read, 1997). Ozone decreases sugar transport to roots, and genes responsible for starch and lipid metabolism are up-regulated (Koch, 1996) (Fig. 1). In leaves, up-regulation of photosynthetic enzymes is related to hexose metabolism rather than sugar abundance per se (Moore et al., 1999). Similarly, down-regulation of ‘use’ genes in roots caused by decreased sucrose cycling may be unaffected by availability of sugar concentrations resulting from starch breakdown. Increased mobilization of reserves coupled with down-regulation of ‘use’ genes in roots could yield a transient increase in sugar availability to mycorrhizal fungi.

Increased soluble sugar concentrations from breakdown of stored reserves may up-regulate fungal genes responsible for sugar uptake (Nehls et al., 1998). Hampp & Nehls (2001) suggest that hexose concentration at the fungus/root interface is responsible for sugar uptake in the ectomycorrhizal (ECM) fungus Amanita muscaria. Up-regulation of hexose transporter genes in adjacent root cells did not occur in their studies, suggesting that the fungus rather than root cells control uptake of available sugars by the fungus. The inducible transporter system may enable the fungus to obtain carbohydrate at a sustained rate despite decreased total allocation to roots under O3 stress. This mechanism, if it occurs, could increase ‘root’ surface area with minimal carbon cost, and could be an adaptive strategy for nutrient and water uptake when insufficient carbohydrate is available to support root growth.

Currently, the above hypothesis has not been tested. In the only study where carbon allocation to extrametrical hyphae was measured, O3 significantly reduced allocation to the fungus (Andersen & Rygiewicz, 1995). Short-term, pulse-chase experiments using isotopes, such as this one, would not reveal a change in partitioning from current to stored reserves in support of the symbiosis. In at least one study, O3 was found to increase reducing and nonreducing sugar concentrations in roots over a 5-wk exposure period (Mahoney et al., 1985), but it was not clear whether increased levels were due to increased allocation or metabolism of reserves, or both.

A second explanation for the stimulatory effect of O3 on mycorrhizal symbioses could be related to membrane leakiness and release of carbohydrates and amino acids into the rhizosphere. Nutrient deficient roots become ‘leaky’ (Ratnayake et al., 1978), and it is possible that carbohydrate depletion in roots also results in loss of membrane integrity. Increased carbohydrate leakage may foster development of mycorrhizal fungi (Graham et al., 1981).

Decreased allocation below ground associated with O3 may influence host-symbiont compatability. Edwards & Kelly (1992) found a shift in fungal morphotypes present on loblolly pine roots, even though mycorrhizal short roots per gram fine root was not significantly affected by O3. Qiu et al. (1993) found increased numbers of morphotypes present on O3-sensitive loblolly pine seedlings exposed to O3. Roth & Fahey (1998) found an interaction between O3 and acid precipitation treatments on the composition of fungi forming ectomycorrhizae on red spruce saplings, possibly driven by nutrient availability. Carbohydrate requirements vary among fungal species (Bidartondo et al., 2001), and O3 may effect species composition by altering carbohydrate availability in roots. A shift in species dominance could lead to a change in successional patterns of mycorrhizal communities. Unlike shifts in mycorrhizal community structure associated with increased N deposition (Egerton-Warburton & Allen, 2000; Lilleskov et al., 2002), shifts in mycorrhizal diversity associated with O3 stress would be unrelated to soil resource availability and hence could result in decreased nutrient or water uptake.

Decreased allocation below ground will lead to smaller root systems and possibly less extensive hyphal networks, which would reduce the size and hence strength of the below-ground carbon sink (Fig. 2). Ozone decreased carbon allocation to extrametrical hyphae in ponderosa pine seedlings colonized with Hebeloma crustuliniforme (Andersen & Rygiewicz, 1995), but the extent of O3's impact on extrametrical hyphal growth has not been studied.

Compensation responses in roots and mycorrhizae may be temporally separated from the O3 stress. Compensatory responses in roots may be tertiary, resulting from nutrient or water stress that occurs as a result of decreased root system size or activity under O3 stress. Although decreased carbohydrate allocation below ground would result in down-regulation of ‘use’ genes (Koch, 1996), decreased shoot N would favor allocation of carbon to roots, possibly offsetting the effect of O3. Increased relative sink strength of roots compared to shoots may lead to reestablishment of a suitable root-shoot ratio even though the signal inducing the shift in allocation is not O3 but rather nutrient or water limitation. Clearly, improved understanding is needed of the interactions of stresses in combination with O3.

V. Carbon flux to soils

There is little doubt that O3 alters carbon flux to soil, although the extent of these changes is poorly characterized. Decreased allocation to roots can lead to a suite of changes in the root–soil interface (Fig. 4). Two primary ways O3 alters carbon flux to soil are through: altered rhizodeposition, including exudation, and root and hyphal turnover; and changes in leaf litter quality or quantity.

Figure 4.

Conceptual diagram showing where O 3 disrupts carbon flow in a tree-soil system, including transfer between biotic and abiotic components below ground that influence soil physical and chemical properties. Arrows denote carbon flux pathways that are affected by O 3 . Dashed lines indicate where the impact of O 3 is suspected but unknown.

1. Rhizodeposition

In the few studies conducted, O3 was found to alter the quantity and quality of root exudates. McCool & Menge (1983) found a significant decrease in exudation of amino acids in tomato (L. esculentum) exposed to 300 ppb O3. McCrady & Andersen (2000) observed increased root exudation in nonmycorrhizal wheat (Triticum aestivum). No apparent change in root exudation was found in labeling studies of ECM ponderosa pine (Andersen & Rygiewicz, 1995). Inconsistency in the literature resulting from species differences and experimental protocols might be expected; the important point is that carbon flux to the rhizosphere is altered, affecting interactions between roots and rhizosphere organisms.

Decreased allocation to roots of ozone-exposed plants may lead to reduced root longevity and accelerated root turnover (Fig. 4). Fine root turnover decreased in mature northern red oak exposed to elevated O3 (seasonal exposure ranging from 152 to 189 ppm h), whereas seedlings did not show any reduction in turnover (Kelting et al., 1995). King et al. (2001) found a trend toward decreased live root biomass and increased dead root biomass in aspen exposed to O3 in a free-air exposure study, suggesting possible changes in both production and longevity. Root longevity and turnover are critical determinants of N and C cycling rates in ecosystems (Zak et al., 2000), and full understanding of the effects of ozone on soil processes requires better characterization of turnover times. In addition, ozone may alter N or C status of roots, affecting foodweb interactions.

Indirect studies also suggest that ozone alters carbon flux to soils. As noted earlier, soil CO2 efflux decreased from loblolly pine pots exposed to ozone (Edwards, 1991). Soil CO2 efflux increased in response to O3 in ponderosa pine seedlings (Andersen & Scagel, 1997; Scagel & Andersen, 1997). In a three year study designed to follow the effects of O3 on competition, O3 exposure increased soil CO2 efflux at the end of the first growing season, but no increase was observed during years two and three of the exposure study. Subsequent measures of individual root respiration revealed no significant O3 response (unpublished data), indicating heterotrophic rather than autotrophic respiration was responsible for the increased flux.

No direct assessments of hyphal growth and turnover in response to ozone stress have been conducted. Decreased allocation of carbon to extrametrical hyphae (Andersen & Rygiewicz, 1995) might be expected to decrease hyphal extent and longevity.

2. Litter quality and quantity

Leaf litter contributes substantially to nutrient cycling and soil carbon (Fig. 4). The role of early senescence in reducing carbon assimilation and allocation to roots (as discussed above) may be less important than the quantity and quality of senesced foliage, once it reaches the litter layer.

Ozone can alter leaf nutrient contents that affect litter quality. Yellow-poplar (Liriodendron tulipifera) and black cherry (Prunus serotina) litter exposed to O3 showed greater N loss during decomposition than charcoal filtered controls, although mass loss did not vary among O3 treatments (Boerner & Rebbeck, 1995). Subsequent studies showed that O3 reduced foliar N concentrations in yellow-poplar leaves, which resulted in slower decomposition following senescence (Scherzer et al., 1998). Other studies have also shown a change in foliar N concentration in response to O3 treatment, affecting C : N ratio and possible litter quality (Berg & Staaf, 1980; Andersen et al., 2001).

In some cases it appears that N remobilization from foliage into the plant is not complete at the time of foliage abscission in O3-exposed plants (Findlay & Jones, 1990; Stow et al., 1992; Matyssek et al., 1993; Patterson & Rundel, 1995). Greater N content of senesced litter could increase rates of decomposition. However, prematurely senesced foliage from O3-exposed cottonwood (Populus deltoides) was found to decompose more slowly than controls despite higher N content (Findlay & Jones, 1990; Findlay et al., 1991). Higher N in senesced leaves appeared to be related to complexes formed by bound phenolics in O3-exposed leaves, which were responsible for decreased decomposition rates (Jones et al., 1994; Findlay et al., 1996). Increased phenolics also have been found in European silver birch (Betula pendula) exposed to O3 (Saleem et al., 2001).

Compositional changes in leaf structural characteristics, such as lignin content, would be expected to alter rates of litter decomposition (Fogel & Cromack, 1977; Meentemeyer, 1978; Kim et al., 1998). Blackberry (Rubus cuneifolus) litter exposed to elevated O3 had greater permanganate lignin than control treatments, which was inversely related to mass-loss rates in decomposition studies (Kim et al., 1998).

Nitrogen allocated to synthesis of secondary metabolites, or lost in litter due to incomplete remobilization, is not available for internal recycling, and therefore represents a significant cost to the plant. Internal recycling of N is important in perennial species, and accelerated loss of N in O3-exposed plants requires additional energy for uptake of nitrogen to maintain plant N balance. Any change in rate of litter decomposition will lead to a change in nutrient cycling rates (Fig. 4). Additional research is needed on the short and long-term effects of O3 on litter quantity and quality due to its potential to alter soil properties, which may lead to changes in soil properties in ecosystems.

Ozone may affect early stages of decomposition by altering populations of leaf surface organisms, before or after senescence. Magan et al. (1995) found a shift in phyllosphere fungi on Scots pine (Pinus sylvestris), Sitka spruce (Picea sitchensis), and Norway spruce (Picea abies) exposed to O3, but the potential effect of these changes on subsequent litter decomposition was uncertain. The slowest decomposition rates of pre-exposed blackberry leaves were found when senesced foliage was exposed to O3 during decomposition, suggesting a possible direct effect of O3 on microorganisms on decomposing litter (Kim et al., 1998). Whether or not O3 concentrations at the soil surface influence initial stages of litter decomposition remains to be addressed.

VI. Soil food web

Processing of plant-derived carbon compounds by organisms comprising the soil food web is a fundamental property of a functional and stable below-ground ecosystem (de Ruiter et al., 1998; Wolters, 1998). Processing of carbon residues leads to the formation of soil organic matter, which influences soil physical and chemical properties (Fig. 4). Soil foodweb organisms are responsible for recycling nutrients and for development of soil properties such as porosity, aggregate structure, water holding capacity and cation exchange capacity. It is these properties that are fundamental to the establishment and maintenance of ecosystems and which may be affected by altered allocation patterns in plants exposed to O3.

The high degree of interdependence among organisms comprising food webs suggests that a change in the quality or quantity of carbon movement from plants to soils resulting from O3 stress would be propagated through other levels of the food web (de Ruiter et al., 1998). Food webs are highly integrated systems that have a regular trophic structure (Pimm & Lawton, 1977; Pimm, 1982). Trophic structure is constrained even when species diversity is high, generally resulting in three or four transfers within most food chains (Pimm, 1982; Moore et al., 1993). Well developed food webs with long trophic ‘loops’ are thought to be stable as a result of weak links between preditor and prey organisms (Neutel et al., 2002). Despite apparent stability resulting from these weak links, it is not known which soil processes will be affected by altered energy flow through soil food webs.

Predicting the effect of O3 on the below ground ecosystem is especially difficult because it is difficult to understand the specific role of organisms in the trophic structure (de Ruiter et al., 1998). Naeem et al. (1994) manipulated species diversity while maintaining trophic structure in model ecosystems, and found changes in soil N, K and P availability, but the response was unpredictable based on levels of diversity. The results show that specific ecosystem processes may respond uniquely to loss of species diversity. In addition, alteration of certain soil organisms, such as mycorrhizal fungi, may have a greater impact on ecosystem productivity than others (Moore et al., 1993; van der Heijden et al., 1998). In Southern California, N deposition led to changes in AM community composition, which may be partially responsible for a shift in dominance from coastal sage to Mediterranean grass (Egerton-Warburton & Allen, 2000). Clearly, species diversity affects soil processes and hence ecosystems structure and function (Naeem et al., 1994), but the idiosyncratic nature of soil foodweb organisms makes it extremely difficult to predict which soil processes will be affected by the loss of a species or a shift in species composition.

There are no comprehensive studies on the effects of O3 on structural or functional components of soil food webs. In the few cases where soil microbial communities have been examined, O3 has led to changes in bacterial and fungal biomass, and in some cases changes in soil enzyme activity. Phillips et al. (2002) examined the effects of elevated CO2 and O3 on carbon flow through heterotrophic microbial communities in soils collected from a free-air exposure study in Wisconsin, USA. Ozone decreased abundance of fungal PLFA in aspen and birch-aspen plots but had few other direct effects on measured soil parameters. The greatest effect of O3 was to eliminate significant increases in microbial respiration resulting from elevated CO2, suggesting an important role for O3 in altering carbon flow through soils. Shafer (1988) found that O3 tended to increase the number of fungal propagules and bacteria exhibiting phosphatase activity in the rhizosphere of Sorghum. Ozone in combination with simulated acid rain stimulated soil arylsufatase activity (Reddy et al., 1991). The response was observed at low, but was reversed at high concentrations, suggesting a threshold level of O3, possibly involving different mechanisms. Ozone significantly decreased soil microbial biomass in the fall after one season of exposure in a wheat (Triticum aestivum) and soybean (Glycine max) system (Islam et al., 2000). Other studies have shown shifts in microbial and fungal biomass in response to ozone stress, but responses were variable (Scagel & Andersen, 1997; Yoshida et al., 2001). It is possible that O3 may shift microbial species composition without altering total microbial biomass, as found under elevated CO2 (Phillips et al., 2002). Collectively, these studies show the potential for O3 to alter functional characteristics of soil food webs, and illustrate the need for comprehensive approaches to study carbon flow through soils in ecosystems exposed to O3.

VII. Summary, conclusions, and future research

It is clear that O3 alters source–sink balance in plants, resulting initially in carbon retention in shoots and decreased carbon allocation below ground. Decreased carbon allocation to roots occurs relatively quickly, and in some cases root responses develop before shoot effects are observed. Decreased allocation below ground alters carbon flux to soil, and it is probable that soil processes are altered as a result.

The combination of stresses present in natural ecosystems makes predicting O3 effects particularly difficult. In plants, gene up-regulation in response to loss of roots could lead to a new balance between roots and shoots, and decreased allocation of carbon to roots may be partially offset. Changes in resource distribution in response to O3 may predispose plants to other stresses, particularly if the stresses are temporally separated. It has been proposed that elevated CO2 will partially offset the effects of O3 on plant growth and photosynthesis, and several short-term studies have found this to be the case (Fiscus et al., 1997; Volin et al., 1998; Grams et al., 1999; Cardoso-Vilhena & Barnes, 2001). However, there is still much debate regarding the effects of elevated CO2 on below-ground processes (New Phytologist, 2000, 147(1)), particularly over the long-term. Information on the combined effects of elevated CO2, O3, amount of nutrient capital, and elevated temperature on below-ground processes is not well understood, yet this information is critical for predicting how these stresses are affecting ecosystems. Additional work is required to better understand the interactions among stresses in natural ecosystems, and how O3 may cause predisposition to naturally occurring stresses.

Given large uncertainty, three hypotheses are presented that warrant further investigation in natural ecosystems:

Ozone stress alters soil carbon pools As noted above, studies have revealed both increased and decreased soil CO 2 efflux in response to O 3 . If episodic occurrences of elevated O 3 increase rhizodeposition by accelerating root turnover, increasing exudation, or increasing litter inputs as found in controlled studies, microbial communities may shift from older to newer, more labile forms of carbon. A similar response has been observed in studies with elevated CO 2 ( Islam et al., 2000 ; Cardon et al., 2001 ). In the case of O 3 , such a response would likely be episodic and transient, occurring in areas where seasonal variation in O 3 exposure is high. In areas that experience chronically high O 3 exposure, decreased allocation of carbon to roots would lead to smaller root systems, smaller plants and less carbon flux to soils. Since soil organisms are generally carbon limited ( Zak et al., 1994 ), long-term exposure to O 3 is hypothesized to lead to increased decomposition of older soil organic matter as labile pools shrink, decreasing sequestration of carbon in soils.

Ozone stress alters nutrient cycling in soils Ozone may alter nutrient cycling in soils by altering litter quality and quantity (root and leaf), and by altering rates of energy flow through soil food webs. Decreased rates of nitrogen fixation also will lead to reduced soil nitrogen. Changes in the quantity or quality of soil carbon will affect nutrient retention in soils. Plant response to O 3 is affected by nutrient availability ( Whitfield et al., 1998 ; Utriainen & Holopainen, 2001 ), so any change in nutrient cycling and plant availability may lead to feedback effects on plant sensitivity to O 3 .

Ozone stress alters plant species diversity indirectly through structural or functional changes in components of the soil food web There is evidence that feedbacks between roots and soil microbial populations affect plant species diversity and community structure ( Bever et al., 1997 ; van der Heijden et al., 1998 ). For example, soil pathogens may increase plant species diversity through negative feedbacks, while beneficial organisms such as mycorrhizas may decrease diversity through positive feedbacks ( Mills & Bever, 1998 ). Shifts in mycorrhizal species occurrence, which has been observed in response to O 3 , may reflect broader structural changes in soil fungal diversity. The relationship between soil biodiversity and plant productivity has been demonstrated in model ecosystems ( Naeem, 1995), and future studies need to determine if O 3 stress alters soil biotic communities sufficiently to result in a change in plant species diversity or system productivity.

In order to address these and other hypotheses, long-term studies need to be under taken. Coupling of field, laboratory, and modeling efforts will help to identify and predict long-term consequences of shifts in process rates. Future efforts should be directed at quantifying the extent of changes occurring below ground in response to O3, and how these changes may alter the long-term stability of ecosystems.

Acknowledgements

I thank the following individuals for extremely helpful comments on earlier versions of this manuscript: David Tingey, John Laurence, Paul Rygiewicz, William Hogsett and Ruth Yanai. I also thank Virginia Robinson and Karen Gundersen for their editorial assistance.

Ancillary