Nitrogen transformations in boreal forest soils—does composition of plant secondary compounds give any explanations?
- First Online:
- Cite this article as:
- Smolander, A., Kanerva, S., Adamczyk, B. et al. Plant Soil (2012) 350: 1. doi:10.1007/s11104-011-0895-7
- 769 Views
Two major groups of plant secondary compounds, phenolic compounds and terpenes, may according to current evidence mediate changes in soil C and N cycling, but their exact role and importance in boreal forest soils are largely unknown. In this review we discuss the occurrence of these compounds in forest plants and soils, the great challenges faced when their concentrations are measured, their possible effects in regulating soil C and N transformations and finally, we attempt to evaluate their role in connection with certain forest management practices. In laboratory experiments, volatile monoterpenes, in the concentrations found in the coniferous soil atmosphere, have been shown to inhibit net nitrogen mineralization and nitrification; they probably provide a C source to part of the soil microbial population but are toxic to another part. However, there is a large gap in our knowledge of the effects of higher terpenes on soil processes. According to results from laboratory experiments, an important group of phenolic compounds, condensed tannins, may also affect microbial processes related to soil C and N cycling; one mechanism is binding of proteins and certain other organic N-containing compounds. Field studies revealed interesting correlations between the occurrence of terpenes or phenolic compounds and C or net N mineralization in forest soils; in some cases these correlations point in the same direction as do the results from laboratory experiments, but not always. Different forest management practices may result in changes in both the quantity and quality of terpenes and phenolic compounds entering the soil. Possible effects of tree species composition, clear-cutting and removal of logging residue for bioenergy on plant secondary compound composition in soil are discussed in relation to changes observed in soil N transformations.
KeywordsCarbon mineralizationForest soilNitrogen cyclingPhenolic compoundsTanninsTerpenesTree species
Higher plants may allocate a significant proportion of assimilated carbon into the production of secondary metabolites, such as terpenes and phenolic compounds. Secondary metabolites are thought to have different roles in plants, particularly in defense against herbivores and pathogens, protection from UV radiation and excess light and by antioxidant activity also in response to other environmental stress factors (Close and McArthur 2002). Various hypotheses have been proposed to explain the type, distribution and abundance of secondary metabolites in different species and different biomes, but there is no clear consensus on this. There is evidence that some of these compounds may also mediate substantial changes in soil C and N cycling, but their exact role and importance in C and N transformations is largely unknown.
Production of secondary metabolites is particularly significant if plants are growing in nutrient-poor conditions, in low pH and short growing season or under other environmental stress (Thoss et al. 2004), conditions prevailing in a majority of boreal forests. In most boreal forest soils availability of N limits the rate of net primary production (Vitousek and Howarth 1991). Boreal forest soils contain large amounts of N, most of it in organic form. However, the rate of net mineralization of N, i.e. the release of ammonium-N from organic N-containing compounds in decomposition, is often very low (Persson and Wirén 1995; Merilä et al. 2002). The nitrogen cycle is relatively closed in undisturbed boreal forests, most of the nitrogen being recycled within the soil-microbe-plant system. Nitrification is the key process involved in opening the N cycle and may lead to losses of N from the ecosystem. Owing to its high mobility, nitrate is susceptible to leaching and is a potential source of contamination in groundwater. Another source of N losses from ecosystems is denitrification; the final product is N2, but a strong greenhouse gas N2O is also released. N2O can be also produced as a by-product in nitrification. An increased rate of nitrification may result in increased N2O emissions (Martikainen 1996). In undisturbed boreal forest soil net nitrification is usually negligible, although under certain conditions it can become significant (reviewed by Martikainen 1996; Smolander et al. 2000).
Although the great majority of N in boreal forest soils is organic N, we know little about the characteristics of it, mainly due to lack of proper methods of identification that do not destroy the structures. Soil organic N may include proteins, peptides and amino acids, amino sugars and nitrogen bases, as well as N in unidentified structures. Aspects of soil organic N have been the topic of several reviews (e.g. Schulten and Schnitzer 1998; Nannipieri and Eldor 2009; Knicker 2011). The most common form of N in soils is amino acid N; acid hydrolysis of soils often yields 20–50% of the total N as amino N, which is probably an underestimation (Senwo and Tabatabai 1998). According to Schulten and Schnitzer (1998), the major N components in soil are proteinaceous material (proteins, peptides and amino acids), comprising ca. 40%, and heterocyclic N compounds (including nucleic acids), comprising ca. 35%. Proteins and peptides do not usually occur in soil as pure compounds. Instead they are bound or complexed with soil minerals and with humic and other organic substances, including certain plant secondary compounds, and this protects organic N from degradation. The distribution of amino acid species in soil seems to be fairly uniform, with aspartate, glutamate, glycine and alanine being most common (Senwo and Tabatabai 1998; Lipson and Näsholm 2001). Reported values for extractable amino acids fall in the range of 0.04–24 mg N kg−1 soil (Lipson and Näsholm 2001).
To become available for plants, the complex organic N must be depolymerized first to organic N-containing monomers. Schimel and Bennett (2004) proposed that this depolymerization process is the key limiting process in N-poor ecosystems instead of the traditional view emphasizing the importance of N mineralization process. Accordingly, uptake of organic N by plants has recently received considerable attention. A Swedish research group showed that several plants, including forest trees like Scots pine (Pinus sylvestris L.) and Norway spruce (Picea abies (L.) Karst.), can take up significant amounts of amino acids in natural forest conditions (Persson and Näsholm 2001; Persson et al. 2003: Näsholm et al. 2009). The group showed intact amino acids, derived from soil, in the roots and shoots of plants and found that plants can also metabolize these amino acids. Moreover, for certain plants there is evidence that proteins can be their sole source of N (Adamczyk et al. 2008a, 2010), in addition to amino acids or small peptides. Adamczyk et al. (2009a) showed, using LC-MS techniques, that leek (Allium porrum L.J. Gay) seedlings exuded proteases able to degrade proteins. Whether these kinds of activities occur under soil conditions and are relevant for boreal forest-tree roots, which are usually mycorrhizal, is not known. We also have to understand that, although the existence of organic N uptake has been demonstrated in both laboratory and field studies, direct evidence in terms of how significantly this uptake contributes to plant N nutrition is still lacking (Näsholm et al. 2009). As summarized by Gärdenäs et al. (2011), until now no single experiment has been able to quantify the extent to which plants utilize organic N in a specific ecosystem. Moreover, since a large part of the organic N is not free, e.g. not as free proteins or amino acids, but is bound to other structures, the picture is very complicated.
As mentioned above, plant secondary compounds may affect soil C and N cycling and, particularly in boreal forest soil conditions these compounds may have a role in retaining N in soil. The aim of this review is to present information on the occurrence of two major groups, terpenes and phenolic compounds, in forest soils and how they affect soil processes. We also attempt to evaluate the importance of these compounds under boreal-forest soil conditions in connection with forest management practices such as tree-species composition, clear-cutting and removal of logging residue for bioenergy.
Role of terpenes in N cycling in forest soil
Occurrence of terpenes in plants and soil
Terpenes in plant leaves often make up 1–2% of the dry mass (d.m.), but higher levels, up to 15–20%, can also be reached (Langenheim 1994). Of all terpenes, triterpenes are the most widely distributed in nature whereas the distribution of most diterpenes is more restricted to certain plant families (Dev 1989). Kanerva et al. (2008) determined the concentrations of sesqui-, di- and triterpenes in the common boreal tree species, silver birch (Betula pendula Roth.), Norway spruce (Picea abies (L.) Karst) and Scots pine (Pinus sylvestris L.). The sum concentrations of sesqui-, di- and triterpenes in silver birch leaves (1 g kg−1 d.m.) were lower than those in Norway spruce (2.7 g kg−1 d.m.) or Scots pine (2.3 g kg−1 d.m.) needles; birch leaves contained no diterpenes and only a small amount of sesquiterpenes, but the concentration of triterpenes was more than twice as high as that in conifer needles (Kanerva et al. 2008). The concentrations of both sesqui- and diterpenes were under the limit of detection in the understory vegetation species common in boreal forests, feather moss (Pleurozium schreberi (Brid.) Mitt.), blueberry (Vaccinium myrtillus L.), lingonberry (V. vitis-idaea L.) and wavy hair-grass (Deschampsia flexuosa (L.) Trin.); but they all contained triterpenes (Kanerva et al. 2008). Both conifers and deciduous trees may synthesize resin, a heterogeneous mixture of fats and fatty acids, steryl esters and sterols, terpenes and waxes; but conifers usually produce much larger amounts (Back and Ekman 2000). Common resin acids in Scots pine and Norway spruce are e.g. pimaric, isopimaric, palustric and abietic and dehydroabietic acids (Ekman and Holmbom 2000).
Not many reports on forest litter or soil terpene concentrations are available and in most of these, only monoterpenes have been measured (White 1994; Dijkstra et al. 1998; Smolander et al. 2005, 2006; Asensio et al. 2008; Kanerva et al. 2008; Leff and Fierer 2008; Maurer et al. 2008; Smolander et al. 2008, 2010a; Ketola et al. 2010). In the humus layer of boreal forests, depending on tree species, the sum concentrations of sesqui-, di- and triterpenes vary in the range of about 0.5–5 g/kg o.m. (organic matter), i.e. about 1–10 g/kg C, the concentrations being highest under pine and lowest under birch (Smolander et al. 2005, 2008, 2010a; Kanerva et al. 2008). Fewer data are available in which monoterpenes have also been measured from the same soils as the higher terpenes; however, including monoterpenes in the sum makes pine, with its high total concentration of terpenes, differ even more from spruce and birch (Smolander et al., unpublished). The concentration of terpenes decreases from the litter to the humus layer, but not as much for triterpenes as for the lower terpenes (Kanerva et al. 2008).
In the above-mentioned studies, the most abundant sesquiterpene has been junipene. Most of the diterpenes are resin acids, with a minority being manool-related compounds. The dominant resin acid is dehydroabietic acid; other common resin acids include pimaric, isopimaric, abietic and pinifolic acids. Dijkstra et al. (1998) also reported that similar types of resin acids occur in large concentrations in Scots pine needles and in the top of the organic layer under Scots pine. Accordingly, after 19 months of decomposition dehydroabietic acid was the main resin acid in Scots pine needle litter, degrading at a lower rate than other resin acids (Kainulainen and Holopainen 2002). β-sitosterol was the most abundant triterpene in leaf/needle material of both the trees and understory vegetation studied, as well as in the humus layer (Smolander et al. 2005, 2008, 2010a; Kanerva et al. 2008).
The concentrations of monoterpenes in aboveground litter of different conifers were 1–5 g/kg but decreased from the litter to the organic layer and mineral soil (White 1991; Wilt et al. 1993a; Asensio et al. 2008; Maurer et al. 2008; Ludley et al. 2009a). In the humus layer under Norway spruce and Scots pine the concentrations were 0.1–0.2 g/kg o.m., whereas under silver birch they were very low (Smolander et al. unpublished). The predominant monoterpenes were α-pinene, β-pinene, ∆-3-carene, camphene, myrcene and limonene (White 1991; Wilt et al. 1993a; Asensio et al. 2008; Ludley et al. 2009a; Smolander et al., unpublished); but their proportions varied depending on tree species and soil layer. In the organic layer under both Norway spruce and Scots pine, α-pinene made up the highest proportion (Smolander et al., unpublished).
Both spatial and temporal variations in monoterpene concentrations in the soil atmosphere are large, which certainly influences their biological effects. Loss of monoterpenes from litter probably occurs through gradual diffusion into the surrounding atmosphere (Wilt et al. 1993b; Van Roon et al. 2005). The total content of monoterpenes in the litter samples was shown to vary substantially, which affects the concentrations of different monoterpenes in the air phase (Wilt et al. 1988; White 1991; Strömvall and Peterson 2000). In addition, vapor pressure and the boiling point of the particular monoterpene as well as the temperature, water saturation and organic carbon content of the soil may also have an effect on monoterpene concentrations (Wilt et al. 1993b; Van Roon et al. 2005). In freshly fallen needle litter from Scots pine and two spruce species the majority of the monoterpene content of coniferous needle litter was lost in half a year (Ludley et al. 2009a). However, White (1994) emphasized that retention of monoterpenes in litter is related to the types of structures in which they are sequestered: short retention times in easily disrupted structures, longer retention time in resistant structures.
Due to differences in the volatility of different monoterpenes, the concentrations of different monoterpenes in the soil atmosphere do not follow exactly the pattern in the soil (Smolander et al., unpublished). The main monoterpene present in Norway spruce and Scots pine is α-pinene (Manninen et al. 2002), which also has the lowest boiling point as well as the highest vapor pressure at 38°C (Wilt et al. 1993b). Therefore, it was not surprising that α-pinene was the most abundant monoterpene in the forest soil atmosphere (Fig. 2, Smolander et al. 2006). Both litter and roots are important sources of terpenes in soil, but the relationship between the contribution of aboveground litter and roots to soil monoterpenes seems to vary depending on the compound in question and on the tree species (White 1991; Lin et al. 2007; Asensio et al. 2008; Maurer et al. 2008; Ludley et al. 2009a).
In addition to plants, soil microbes produce a wide variety of volatile organic compounds, including terpenes (Stahl and Parkin 1996; Leff and Fierer 2008; Bäck et al. 2010; Insam and Seewald 2010). However, their contribution is probably much smaller than that of plants, at least in soil under conifers (Paavolainen et al. 1998). Mycorrhizal infection may also affect volatile terpenes in soil. They consisted of the same compounds in roots of mycorrhizal and nonmycorrhizal Scots pine seedlings but mycorrhizal fungi showed different effects on the concentrations of individual terpenes (Napierała-Filipiak et al. 2002; Werner et al. 2004).
Volatile terpenes have considerable effects on the chemical composition and physical characteristics of the atmosphere (reviewed by Laothawornkitkul et al. 2009). They can, for example, form secondary organic aerosols that can scatter solar radiation leading to cooling of the Earth Surface.
Biological effects of monoterpenes
Because of their volatility, monoterpenes are a special group of compounds with regard to the soil microbial population. Monoterpenes are poorly soluble in water; however, oxygenated forms are more soluble, which may broaden their distribution in soil (Weidenhamer et al. 1993).
Nitrification is the key process in the loss of N from the ecosystem. Monoterpenes have been shown to inhibit net nitrification (White 1986, 1991, 1994; Paavolainen et al. 1998; Uusitalo et al. 2008). In laboratory incubation experiments, if the soils were otherwise nitrifying, addition of monoterpenes or exposure to monoterpene vapors inhibited net nitrification; mixtures of several monoterpenes like limonene, myrcene, α- and β-pinene and β-phellandrene, camphene, 3-carene as well as individual compounds like α- and β-pinene have been studied (White 1986; Paavolainen et al. 1998; Uusitalo et al. 2008). Even with small additions of monoterpenes, net nitrification was inhibited with little change in net N mineralization, which pointed to a direct effect on nitrification rather than only via ammonium formation, i.e. availability of the substrate (White 1991, 1994). In soil-suspension experiments with a continuous excess of NH4-N, net nitrification was also inhibited by α- and β-pinene (Paavolainen et al. 1998) These experiments confirmed that the decrease in nitrification is not only due to a decrease in ammonification but that there also exists some kind of direct inhibition of nitrifying bacteria. Accordingly, monoterpenes were shown to inhibit the nitrification rate and growth of Nitrosomonas europaea in whole-cell pure culture experiments (Ward et al. 1997). We must keep in mind that results also point to indirect inhibition due to enhanced immobilization of NH4-N (Bremner and McCarty 1988; Paavolainen et al. 1998). Monoterpenes probably affect net nitrification in forest soils both directly and indirectly.
The responses of soil microbial biomass, net N mineralization and nitrification as well as growth of nitrifiers in pure culture have been shown to vary with different monoterpenes (White 1988, 1994; Ward et al. 1997; Smolander et al. 2006), indicating that chemical structure of terpenes is important. Concentration is also important. With regard to nitrification, there could be two kinds of inhibiting mechanisms, at low concentrations a specific inhibition of ammonium monooxygenase by competitive or noncompetitive inhibition, and at high concentrations a general toxicity (Ward et al. 1997).
We know little about the response of denitrification, another source of N losses from ecosystems, to monoterpenes. The response of denitrification activity of environmental bacterial isolates to monoterpenes has been found to be variable (Amaral et al. 1998). In a laboratory incubation experiment, exposure of soil to a mixture of monoterpenes did not affect denitrification (Paavolainen et al. 1998).
Monoterpenes may affect a range of other types of microbes. They may inhibit methanotrophic bacteria and methane oxidation in soil (Amaral and Knowles 1998; Maurer et al. 2008). A microbial group in forest soils that has also been reported to be affected by monoterpenes is certain saprophytic and mycorrhizal fungi. Volatile organics from Scots pine or monoterpene vapors (α- and β-pinene and 3-carene) have been shown to inhibit growth of several ectomycorrhizal isolates, in addition to some saprotrophic isolates (Melin and Krupa 1971; Ludley et al. 2008); but the effect again depended on the compound. Monoterpenes also decreased the respiration rate of two litter-degrading fungal species (Ludley et al. 2008). In further studies, Ludley et al. (2009b) showed, however, that monoterpene vapors stimulated ectomycorrhizal colonization of Norway spruce seedlings.
Effects of higher terpenes and resin on soil processes
There is surprisingly little information available on how higher terpenes affect soil processes, but some studies indicate that there may be antibacterial and antifungal effects. Triterpenes from a wood-decay fungus, Fomitopsis rosea, were shown to inhibit a bacterium, Streptococcus aureus (Popova et al. 2009). β-sitosterol and other triterpenes have been shown to inhibit some species of fungi (Aderiye et al. 1989; Smania et al. 2003). Resin from a medicinal plant Pseudognaphalium viravira (Mol.) A. Anderb. and its diterpene component, kaurenoic acid, caused changes in the soil bacterial community (Gil et al. 2006). According to recent results, addition of diterpenes colophony (a mixture of several resin acids) and abietic acid and a triterpene β-sitosterol seemed to stimulate C mineralization and to increase microbial biomass in forest soil, but inhibit net N mineralization (Adamczyk et al. 2011b). Results of these experiments revealed no toxic effects on the soil microbial population, contrary to the case with monoterpenes and the decreased net N mineralization was probably due to increased immobilization of N in the soil microbial biomass due to these compounds acting as C sources. Still, toxic kinds of effects could not be excluded; and indeed, further studies pointed to selective inhibition of part of the soil microbial population (S. Adamczyk, O. Kiikkilä, A Smolander, unpublished.).
Secretory resin (canal resin) generally contains only terpenes, terpenoids and in certain plant genera polyisoprenes (Back and Ekman 2000). Exposure of the humus layer to coniferous resin (Lenoir et al. 1999) or resin volatiles (Uusitalo et al. 2008) increased C mineralization and decreased net N mineralization and net nitrification. Fungal biomass was increased by addition of resin (Lenoir et al. 1999), but resin volatiles did not affect microbial biomass C or N consistently (Uusitalo et al. 2008). In plant defense, mixtures of terpenes, such as conifer resin, may act synergistically to provide greater toxicity or deterrence than the equivalent amount of a single substance (Gershenzon and Dudareva 2007). The effects of resin on soil processes probably are also more complicated than the effects of its individual compounds. In addition, there may be physical interactions. The lower-molecular-weight monoterpenes are believed to act as solvents, enabling rapid transport of the higher-molecular-weight diterpene acids from resin ducts; on the other hand, the volatilization of monoterpenes may be retarded in the presence of the less volatile diterpenes (Gershenzon and Dudareva 2007). Therefore, highly volatile monoterpenes from canal resins, a point-like source of emission, probably survive in soil longer, allowing them to have a longer influence on soil microbes.
Have any relationships between terpenes and soil processes been found in field studies?
Phenolic compounds as N-retaining agents in forest soils
Phenolic compounds are a widespread and diverse group of plant secondary compounds. Their occurrence, chemistry, biological effects and role in allelopathy have been discussed in several excellent reviews (Harborne 1997; Northup et al. 1998; Schimel et al. 1998; Hättenschwiler and Vitousek 2000; Schofield et al. 2001; Kraus et al. 2003; Schimel and Bennett 2004). Here only certain selected interesting aspects and recent findings are discussed.
Chemistry and problems in analytics of phenolic compounds
In quantitative analysis of total phenolic compounds, a spectrophotometric Folin-Ciocalteu method is used (Box 1981) which is based on formation of a coloured complex between phenolic compounds and alkaline Folin-Ciocalteu reagent. Spectrophotometric methods are also used for measuring quantities of both condensed and hydrolyzable tannins. Acid-butanol-assay for measurements of condensed tannins involves HCl-catalyzed depolymerization of condended tannins in butanol to yield a pink-red anthocyanidin product (Waterman and Mole 1994). Hydrolyzable tannins can be measured using potassium-iodate-assay (Hartzfeld et al. 2002), in which hydrolyzable tannins are converted to methyl gallate via acid methanolysis and this methyl gallate reacts with KIO3 giving a red-colored product.
Part of the quantification methods is based on interaction of phenolic compounds with proteins. The Folin-Ciocalteu method (Box 1981) can, based on casein precipitation, further divide phenolic compounds into low- and high-molecular-weight phenolic compounds (< or >0.5 kDa, Haslam 1989). Methods which measure the amount of total tannins precipitated by protein have also been developed. Tannins precipitated by excess of protein are measured spectrophotometrically after reaction with FeCl3 and formation of colored iron-phenolate complex (Hagerman and Butler 1978). In radial diffusion assay tannin-containing solution is placed in a well on protein-containing agar and a visible precipitation develops (Hagerman 1987).
There are several methodological problems in measuring complex phenolic compounds in plant material and particularly in soil; and we must be aware of these limitations (Appel et al. 2001; Kraus et al. 2003; Adamczyk et al. 2008b). With regard to tannins, selection of a proper standard is one of the major problems since the tannin structure may affect its reactivity. Analysis is best standardized against purified tannins from the same plant species (Kraus et al. 2003; Preston et al. 2006), which is not always possible; as an extreme example, soil measurements in a mixed-tree-species forest with a diverse understory. Comparisons between different studies are also difficult due to differences in pretreatment methods, for example, fresh sample vs dry, ground sample, the latter may also contain differences due to the mesh size used. In addition, the efficiency of extraction may differ between the materials studied (Preston et al. 2006); and condensed tannins bound to soil organic matter are not necessarily extractable with 70% acetone (Nierop and Verstraten 2006). In spite of these problems, results within a study give valuable information for the comparisons in question.
Structural studies using more sophisticated methods, e.g. nuclear magnetic resonance (NMR) or different mass spectrometric (MS) methods (e.g. in Lorenz et al. 2000, Maie et al. 2003 and Nierop et al. 2005), have given extremely important information about tannin structure; however, these methods are not suitable for routine analysis in ecological studies. Both NMR and MS have also been useful for establishing the purity and structure of tannins used as standards for further analysis.
Occurrence of phenolic compounds in plants and soil
Condensed tannins are widely distributed, occurring in the leaves of all ferns and gymnosperms and in about half of the families (the woody members) of angiosperms (Harborne 1997; Hernes and Hedges 2004). On the contrary, hydrolyzable tannins have a more restricted occurrence, being found only in part of the dicotyledons. The most constant occurrence of tannins is in leaves, but they can occur throughout the plant, including the flower, fruit, seed, stem and root (Harborne 1997).
In general, high concentrations of phenolic compounds, including tannins, can be found in plants growing in conditions of low soil fertility and pH (Northup et al. 1998). In their review Kraus et al. (2003) stated that in woody species levels of foliar tannins commonly range from 15 to 20% dry mass; and in roots, values of 1–35% have been reported. Among boreal forest trees, Norway spruce needles had higher concentrations of total water-soluble phenolic compounds and total tannins than Scots pine needles or silver birch leaves, and also protein precipitation capacity was highest with spruce (Adamczyk et al. 2008b; Kanerva et al. 2008). Concentrations of condensed tannins were also more than twice as high in Norway spruce needles as in silver birch leaves (A. Smolander, unpublished). Pine and spruce can produce only condensed tannins, but birch contains both hydrolyzable and condensed tannins (Kraus et al. 2003). In a northern tree species experiment Kanerva et al. (2008) analyzed the total water-soluble phenolic compounds in the most common species of ground vegetation. Both the leaves and stems of blueberry and lingonberry contained large amounts of phenolic compounds, whereas in wavy hair-grass and in feather moss concentrations were low (Kanerva et al. 2008). Similarly, both blueberry and lingonberry leaves and stem contained large amounts of total tannins, whereas the amounts in wavy hair-grass and feather moss were very low (Adamczyk et al. 2008b).
Hernes and Hedges (2004) analyzed 77 plant species for condensed tannins and, in general, conifer needles were distinguished by high amounts of prodelphinidin-type condensed tannin overall and relative to procyanidin-type condensed tannin. Condensed tannins of spruce needles contain more procyanidin units while pine needles contain more prodelphinidin units (Maie et al. 2003; Kraus et al. 2004a; Kanerva et al. 2006; Norris et al. 2011).
When we consider concentrations of phenolic compounds in soil, both the input of above- and belowground litter and the concentration of phenolic compounds in it are important, as are soil processes and reactions. Being aware of the limitations posed by determination methods, in the humus layer of different forest sites concentrations of water-soluble phenolic compounds have generally been 1–3 TAE (tannic acid equivalents) g/kg o.m. in ground samples and 0.1–1 TAE g/kg o.m. in unground samples (Gallet and Lebreton 1995; Suominen et al. 2003; Smolander et al. 2005, 2008, 2010a; Kanerva et al. 2008). For the litter layer the values have been 2–5 times higher (Gallet and Lebreton 1995; Kanerva et al. 2008). The proportion of low-molecular weight (LMW) and high-molecular weight (HMW) phenolics varies, but in most cases the ratio is about 50:50 (Suominen et al. 2003; Smolander et al. 2005, 2008, 2010a; Kanerva et al. 2008). In particular, the proportion of HMW phenolic compounds seems to decrease with depth (Kanerva et al. 2008). For acetone-water extractable condensed tannins, the concentrations for organic layers of boreal forest soils have usually varied from about 1 to 5 g/kg o.m. in ground samples, with values up to 25 g/kg o.m. in black spruce (Picea mariana (Mill.) B.S.P) stands in Canada (Kranabetter and Banner 2000; Lorenz et al. 2000; Smolander et al. 2005, 2008, 2010a; Preston et al. 2006; Kanerva et al. 2008).
After entering the soil, phenolic compounds may act in soil processes, react with other compounds, e.g. form tannin-protein complexes, or be degraded or leached. In litter bag experiments with both spruce litter (Lorenz et al. 2000) and birch litter (Stark et al. 2007), tannins were lost rapidly, in some cases more than 80%, during the first year. As discussed above, in the field the concentration of total water-soluble phenolic compounds, condensed tannins and total tannins generally decreased from the litter layer to the humus layer but not as drastically as was the case with several terpenes, and the decrease was dependent on tree species (Adamczyk et al. 2008b; Kanerva et al. 2008). However, NMR studies have shown that the extractability of tannins may decrease during decomposition and therefore the values for more decomposed materials can be underestimations (Lorenz et al. 2000; Preston et al. 2006). Roots can also contain high concentrations of tannins (Gallet and Lebreton 1995; Kraus et al. 2004b; Preston et al. 2006); and their contribution, plus leaching from the above layers, control the tannin concentrations in deeper soil layers.
Tannins affect soil microbial processes in several ways and by different mechanisms
Studies on the effects of condensed tannins or mixtures containing both condensed and hydrolyzed tannins on microbial processes of C and N cycling in litter and soil
Soil or litter where added
Populus balsamifera L.
Alder and poplar forest floor
Kalmia angustifolia L.
Black spruce humus
Bradley et al. (2000)
Abies balsamea L. Mill.
Pinus muricata D. Don
Bishop pine mineral soil
Kraus et al. (2004a)
Vaccinium ovatum Pursh
Gaultheria shallon Pursh
Arctostaphylos nummularia Gray
HT + CT
Rhododendron macrophyllum D. Don ex G. Don
HT + CT
Abies balsamea L. Mill.
Corsican pine forest floor
Nierop et al. (2006a)
Thuja plicata Donn ex D. Don
Kalmia angustifolia L.
Picea mariana (Mill.) Britton, Sterns & Poggenburg
Pinus nigra J.F. Arnold
Litter from Corsican pine forest
Picea abies (L.) Karst.
Norway spruce, Scots pine and silver birch humus layers (F + H)
High PC, HMW
High PC, LMW
Pinus sylvestris L.
Low PC, HMW
Low PC, LMW
Agathis australis (D. Don) Loudon
Kauri organic layer
Verkaik et al. (2006)
Dacrydium cupressinum Sol. ex Lamb.
Knightia excelsa R. Br. (1810)
Populus tremuloides Michx.
Air-dried silt loam from an aspen forest
Madricht et al. (2007)
Commercial guebracho (Aspidosperma quebracho Schltr.)
Soils from tropical agricultural and natural systems
Mutabaruka et al. (2007)
Acer saccharum Marschall
Soils from hardwood-conifer forests
HT + CT
Talbot and Finzi (2008)
Quercus rubra L.
HT + CT
Tsuga canadensis (L.) Carriere
HT + CT
Abies balsamea L. Mill.
Douglas fir humus layer (L and F layers removed)
Norris et al. (2011)
Pinus nigra J.F. Arnold
Vaccinium ovalifolium Sm.
Pinus banksiana Lamb.
Thuja plicata Donn ex D. Don
Kalmia angustifolia L.
A general phenomenon seems to be that addition of HMW condensed tannins decreases N mineralization in soil or litter. This has happened with the purified and characterized condensed tannin fractions originating from several tree or shrub species like Populus balsamifera L. (Schimel et al. 1996, 1998; Fierer et al. 2001), Kalmia angustifolia L. and Abies balsamea (L.) Mill (Bradley et al. 2000; Nierop et al. 2006a; Norris et al. 2011), Pinus muricata D. Don (Kraus et al. 2004a), P. sylvestris L. (Kanerva et al. 2006; Kanerva and Smolander 2008), P. nigra J.F. Arnold (Nierop et al. 2006b; Norris et al. 2011), Picea abies (L.) Karst (Kanerva et al. 2006; Kanerva and Smolander 2008), and P. mariana (Mill) (Nierop et al. 2006a). Even in cases where addition of condensed tannins have increased net N mineralization (Fierer et al. 2001; Kanerva et al. 2006; Kanerva and Smolander 2008; Kraal et al. 2009), gross N mineralization measurements have shown that increased net N mineralization was due to decreased immobilization, not to increased mineralization. The effects on C mineralization and microbial biomass are more variable, but part of the condensed tannins also show an inhibitory effects on both C mineralization and microbial biomass C and N (Fierer et al. 2001; Kanerva et al. 2006; Nierop et al. 2006a; Kanerva and Smolander 2008; Kraal et al. 2009). When all the results are combined, they often point to inhibitory effects of these HMW condensed tannins, caused by toxicity and/or protein-binding effects.
In some studies LMW condensed tannin fractions, also obtained in the tannin extraction procedure, were added to soils (Fierer et al. 2001; Kanerva et al. 2006; Kanerva and Smolander 2008). They seemed to enhance respiration and decrease net N mineralization. However, as emphasized by Kanerva et al. (2006), LMW fractions also contained compounds other than tannins, which may to some extent affect the reactions observed. Therefore, conclusions about their effects are on a weaker basis than those for HMW condensed tannins. This may also be a problem for some other studies; the plant extract may also contain other compounds, some of which are not even phenolic compounds.
Addition of a hydrolyzable tannin, commercially-available tannic acid, into soil resulted in a relatively rapid increase in CO2 production and a decrease in net N mineralization (Kraus et al. 2004a; Kanerva et al. 2006; Kanerva and Smolander 2008; Nierop et al. 2006b; Kraal et al. 2009). Tannic acid seemed to act as an ordinary carbon source, like cellulose, without detectable toxic effects; but due to its protein-precipitating capacity, it may cause an extra decrease in N availability in the soil. Studies with plant extracts containing a mixture of both hydrolyzable and condensed tannins have also shown a decrease in net N mineralization and an increase in CO2 production (Kraus et al. 2004a; Talbot and Finzi 2008). As discussed above, highly polymerized condensed tannins from several plant species clearly decreased soil respiration; the other extreme may be tannic acid, which seems to act as a readily-available C source for soil microbial population.
The picture of the effects of tannins on nitrification, as reviewed by Kraus et al. (2003), is not clear and we do not know so far whether tannins have any clear direct effect on nitrifying bacteria. De Boer and Kester (1996) found no indications that the relatively high content of polyphenolics in dwarf shrubs was important as a regulator of the nitrification process. In some laboratory experiments, polyphenols or condensed tannins inhibited nitrification in soil (Lodhi and Killingbeck 1980; Baldwin et al. 1983; Kraal et al. 2009); but in other experiments there was no effect (McCarty and Bremner 1986). In the study of Nierop et al. (2006a), condensed tannins extracted from several plant species affected net nitrification only slightly, although net N mineralization was clearly reduced. As with the N mineralization process, the effect of tannins on nitrification is probably dependent on soil properties and on tannin concentration and structure. It must also be determined when the effect is only due to a decrease in net N mineralization and a subsequent decrease in substrate supply, and when the effect is direct.
It remains to be determined which soil organisms are most sensitive to tannins, but Kraus et al. (2003) listed several fungi and bacteria that tannins have been reported to inhibit in pure culture. In some tropical soils, addition of condensed tannin (commercial quebracho)—protein (bovine serum albumin, BSA) complex increased the fungi-to-bacteria ratio (Mutabaruka et al. 2007). When the relative availability of HMW condensed tannin fractions to bacteria and fungi was studied, adopting 3H-thymidine and 14C-acetate incorporation methods, respectively, higher concentration seemed to be more inhibitory whereas lower concentration was stimulatory to bacteria; but the effects on fungi were not clear (Kanerva et al. 2006). As discussed by Talbot and Finzi (2008), concentration is important; they explained the decrease in net N mineralization at low tannin (extracts containing both condensed and hydrolyzable tannins) concentrations as being due to greater microbial immobilization, while at higher concentrations the decrease in mineralization was consistent with the formation of tannin-protein complexes.
Tannins as N-binding agents
As mentioned above, one mechanism by which tannins affect soil processes is their ability to precipitate proteins (Haslam 1989). Complexes with tannins are formed through hydrogen bonding and/or hydrophobic effects (Hagerman and Butler 1981; Hagerman et al. 1998). There is specificity among proteins; in competitive binding assays, Hagerman and Butler (1981) showed that one protein can be efficiently precipitated in a large excess of some other protein. For example, conformationally loose proteins had higher affinities for tannins than tightly coiled globular proteins (Hagerman and Butler 1981).
Another way to affect soil microbial processes, in addition to making protein N less available due to formation of tannin-protein complexes, is also based on the protein-precipitating capacity of tannins. Exoenzymes, being proteins, can be immobilized by tannins; and this process may reduce their activity (Schimel et al. 1998; Hättenschwiler and Vitousek 2000; Joanisse et al. 2007). Kraus et al. (2003) listed those enzymes that have been reported in the literature to be inhibited by tannins, but they emphasized that most studies were performed in solution culture and may not reflect how enzymes would react in soil. In an experiment without soil, both commercial tannic acid and condensed tannins from silver birch and Norway spruce inhibited commercial protease (Adamczyk et al. 2009b). The results of Joanisse et al. (2007) with condensed tannins from kalmia and black spruce showed inhibition of enzymes in water extracts from the forest floor. However, plant studies indicated that, although condensed tannins precipitated the enzymes studied, their activity did not necessarily decrease (Juntheikki and Julkunen-Tiitto 2000). Accordingly, addition of condensed tannin fractions from balsam poplar in alder and poplar soils did, in general, not affect enzyme rates (Fierer et al. 2001).
In legume tree litter, consisting of various mixtures of Gliricidia sepium (Jacq.) Steudd and Peltophorum dasyrrachis Kurz ex. Baker prunings, a strong negative correlation was found between net rate of N mineralization and capacity for protein (BSA) precipitation (Handayanto et al. 1997). Adamczyk et al. (2008b) studied protein (BSA) precipitation of plant, litter and humus-layer material to obtain a picture of their protein-binding capacity, and then compared the capacity to amounts of tannins. There were some correlations between amounts of tannins and protein-precipitating capacity, but no consistent overall pattern was observed. This points to differences in the astringency of different tannins. Gundale et al. (2010) also observed variation in protein complexation capacity among materials from different plant species expressed per dry mass of plant material but noted that the picture changed when the capacity was expressed per unit of soluble phenol.
Molecular size and structure of tannins may influence both protein-binding capacity and also the more direct biological effects, such as C mineralization (Maie et al. 2003; Kraus et al. 2003, 2004a; Nierop et al. 2006a; Mutabaruka et al. 2007; Kanerva and Smolander 2008; Talbot and Finzi 2008). However, linking structure to function is still a problem. Condensed tannins seemed to bind more protein than tannic acid, a hydrolyzable tannin did (Kraus et al. 2003; Mutabaruka et al. 2007; Adamczyk et al. 2011a), but this is not always the case (reviewed by Kraus et al. 2003). The protein-precipitation capacity of condensed tannins increased with an increase in degree of polymerization of the tannin (Kumar and Horigome 1986). Among condensed tannins, prodelphinidin is considered to possess higher reactivity than procyanidin (Maie et al. 2003; Kraus et al. 2003). Tannins with a high proportion of prodelphinidin (having predominantly three hydroxy groups at the B ring) inhibited net N mineralization more than procyanidin tannins (having two OH groups) (Nierop et al. 2006a). However, tannins in Norway spruce needles, which contain predominantly procyanidins, and tannins in Scots pine needles, which consist mostly of prodelphidines, did not have very different effects on microbial biomass or rates of C and net N mineralization (Kanerva and Smolander 2008). Moreover, Norris et al. (2011) who studied condensed tannins from six different plant species was not able to demonstrate any clear and consistent effects of tannin structure on C and N mineralization rates in soil. They stated that a problem in developing better structure-function relationships is that the tannin preparations are not typically single pure compounds and the derived structural parameters are averages.
An important question is whether tannins also bind other soil organic N-containing compounds in addition to proteins. Some reports suggest that they do: tannic acid forms complexes with arginine, an amino acid, and with choline, an amine (Kalina and Pease 1977; Mole and Waterman 1987), tannic acid adsorbs to chitosan (amino sugar) beads and chitosan-montmorillonite (Chang and Juang 2004; An and Dultz 2007), and condensed tannins precipitate amino acids arginine and histidine (Mole and Waterman 1987). Adamczyk et al. (2011a) estimated the capacity of both commercial tannic acid and condensed tannins extracted from Norway spruce needles to precipitate a wide range of organic N-containing compounds: the compounds studied were proteinaceous amino acids, proteins and peptides, polyamines, N bases and aminosugars. Evidently both tannic acid and condensed tannins could precipitate several but not all of the compounds studied. Of all amino acids, tannins precipitated only arginine, and of the peptides studied, they precipitated insulin (MW 5733 Da; 51AA). The proteins studied, polyamines, N bases, chitin and chitosan (chitosan studied only with tannic acid) were also all precipitated. As Hagerman and Butler (1981) previously reported for proteins, complex formation was dependent on pH and on the concentration of the compounds. BSA is the protein usually used in protein precipitation studies. Interestingly, both types of tannin also precipitated a large protein, Rubisco (D-ribulose 1,5-diphosphate carboxylase), the most abundant protein in plant leaves and needles (Parry et al. 2003), as also shown earlier (Juntheikki 1999; McAllister et al. 2005). Based on the above information, it seems that tannins have the ability to form complexes with several N-containing compounds other than proteins. More research is needed to determine whether similar complexation occurs under soil conditions, and whether these complexes are as stable as tannin-protein complexes are considered to be.
Possible ecological consequences
Protein-tannin complexation is considered to be a phenomenon that conserves litter N within the forest ecosystem (Northup et al. 1995a, b, 1998; Kraus et al. 2003). High levels of polyphenols and tannins in litter have been suggested to shift N cycling from mineral- to organic-dominated pathways (Northup et al. 1995a, b, 1998; Schimel and Bennett 2004). In their review Kraus et al. (2003) stated that both complex formation and resistance to degradation are greater at lower pH—most boreal forest soils are acidic, the pH often being around 4 or even less (Tamminen 2000). It was further suggested by Northup et al. (1995b, 1998) that certain mycorrhizal symbionts are able to recover N from protein–tannin complexes and thus give the host plant a competitive advantage. However, the picture is not clear since at least ectomycorrhizal fungi of pine have been shown to very weakly break down these N complexes; to make its N available to ectomycorrhizal fungi, pretreatment of the complex by saprotrophs seems to be necessary (Bending and Read 1996; Wu et al. 2003).
Northup et al. (1995b, 1998) further suggested that a high level of polyphenols may not only inhibit N mineralization but also correlate positively with the release of dissolved organic N (DON) from pine leaf litter. They explained this correlation to mean that in strongly N-limited ecosystems certain plants may benefit from increased DON/mineral N ratio. Joanisse et al. (2008), however, questioned the relationship between tannins and the DON-to-mineral N ratio and suggested that this ratio is controlled more by other confounding factors. In any case, in N-poor forest ecosystems, DON is the dominant form of N in soil solution and percolation water (Näsholm et al. 1998; Smolander et al. 2001). Although soluble tannin-protein complexes can be formed, particularly with an excess of protein (Hagerman and Robbins 1987), protein-tannin complexes are generally considered to be insoluble. Accordingly, tannic acid was shown to reduce the solubility of soil organic N (Halvorson and Gonzales 2008; Halvorson et al. 2009), and addition of polymerized condensed tannins from spruce and pine into soil decreased soil concentration of DON (Kanerva et al. 2006). We may conclude that protein-tannin complexes probably do not appear in large amounts in DON.
For net N mineralization the picture is not clear. Northup et al. (1995b, 1998) observed a negative correlation between phenolic content in litter and rate of mineral N release. Net N mineralization and total phenolic compounds correlated negatively also when samples of birch and spruce litter and F layer were studied (Kanerva et al. 2008). However, there was no clear relationship between total phenolics or condensed tannins and net N mineralization in the humus layer with data from several coniferous and birch stands (Fig. 8).
Phenolic secondary compounds in plants and soil also include simple phenolic acids. Some old studies indicate that these compounds may also play a role in N transformations. For example ferulic acids may inhibit nitrification in soil (Rice and Pancholy 1974) and both ferulic and coumaric acids have been reported to inhibit bacterial and fungal activity (Blum and Shafer 1988).
Can composition of plant secondary compounds explain changes in N cycling caused by forest management?
Forestry practice may affect both C and N cycling in soil by changing physical conditions, such as microclimate, or chemical conditions, like C-to-N ratio or pH. These changes in chemical or physical conditions may be caused by differences in, e.g. ground vegetation cover, the amount or chemical composition of above- or belowground litter and root activities. Different forest management practices may result in a different input of litter; thus there are changes in both the quantity and quality of plant secondary compounds entering the soil.
Tree species composition
Several studies have shown that the dominant tree species may affect the biological characteristics of soil (e.g. Menyailo et al. 2002; Grayston and Prescott 2005; Prescott and Vesterdal 2005). This is also the case for silver birch, Norway spruce and Scots pine (Priha and Smolander 1999; Priha et al. 2001; Smolander and Kitunen 2002; Smolander et al. 2005; Kanerva and Smolander 2007). In a recent study results from six tree species experiments of different ages and growing on different site types were compared in different parts of Finland. In all older stands and also in part of younger stands, birch, compared to spruce or pine, increased soil pH, NH4-N concentration and amounts of C and N in the microbial biomass and decreased the C-to-N ratio in the humus layer (Smolander and Kitunen 2011). Birch also increased the rates of both C and net N mineralization compared to spruce and pine, but only on two sites. The DON-to-mineral N ratio was always higher under Scots pine and Norway spruce than under silver birch (Smolander and Kitunen 2011). In the litter layer the degradability of dissolved organic C and N was highest under birch but in the humus layer the results varied (Kiikkilä et al. 2005, 2006). General explanations given for all these tree-species-specific differences have included differences in microclimatic conditions, ground vegetation cover, number of roots and amount and composition of root exudates as well as in the chemical composition of the litter.
Despite great variation within plant species, tree species and the ground vegetation species developed under them have their characteristic patterns of secondary compounds. This may reflect soil concentrations of these compounds, which are also affected by degradation, volatilization and leaching. As discussed above, microbial activities in both C and N cycling are often higher in soil under birch than under spruce or pine, and N is rather mineralized all the way to mineral N as compared to coniferous soils where it stays more as organic N. This is what would be expected based on the plant secondary compound composition in litter and soil. The concentration of condensed tannins was much higher in spruce needles than in birch leaves (Smolander et al. unpublished), protein precipitation was more effective with spruce needle tannins than with birch leaf tannins (Adamczyk et al. 2008b; Gundale et al. 2010), and birch leaves contained no diterpenes (Kanerva et al. 2008). Accordingly, in the humus layer, condensed tannins were most abundant under spruce (Smolander et al. 2005; Kanerva et al. 2008), hydrolyzable tannins under birch (Adamczyk et al. 2009b) and diterpenes under conifers (Kanerva et al. 2008). Perhaps the most striking difference is in volatile monoterpenes; under spruce and in particular pine they were abundant in the soil atmosphere, whereas under birch the concentration was negligible (Fig. 2, Smolander et al. 2006).
Clear-cutting has been shown to increase net N mineralization and to either initiate net nitrification or increase it in several boreal forest ecosystems (Tamm et al. 1974; Vitousek and Matson 1984; Dahlgren and Driscoll 1994; Smolander et al. 1998, 2001), which may lead to increased risk of nitrate leaching. An increase in the production of DON, in particular, its low-molecular weight fraction (<1 kDa), and a simultaneous decrease in DON-to-mineral N ratio was also observed after clear-cutting a Norway spruce stand (Smolander et al. 2001).
In clear-cutting, the soil first receives a large pulse of plant secondary compounds in dead roots and logging residues, if the latter are not harvested. Later the developing understory produces fresh litter, but the composition of secondary compounds is different. For example, in a Norway spruce stand growing on relatively fertile soil, the change in ground vegetation cover can be from a thick layer of needle litter with mosses to herbs and grasses. There is little information available on how clear-cutting affects composition of plant secondary compounds in the soil. Paavolainen et al. (1998) compared a mature Norway spruce stand and an adjacent plot that had been clear-cut 3 years earlier and found a dramatic decrease in the concentration of volatile monoterpenes in the soil atmosphere due to clear-cutting, indicating that the tree stand was the main source of monoterpenes. Accordingly, terpene concentration in soil decreased with increasing distance from the trunks of Pinus spp., thus from pine to grasses and herbs (Lin et al. 2007; Asensio et al. 2008). Increased net nitrification after clear-cutting (Smolander et al. 1998) can be partly explained by the increase in pH (Paavolainen and Smolander 1998) and NH4-N concentration (Smolander et al. 1998), but lack of volatile monoterpenes in the soil of a clear-cut area may also play a role (Paavolainen et al. 1998). Whether changes in the amounts and composition of higher terpenes or phenolic compounds play an important role in regulation of N cycling after clear-cutting is not known. Studies are needed on immediate changes in plant secondary compound composition in soil after clear-cutting but subsequent monitoring for several years is important as well.
Removal of logging residue in harvest
Increasing demand for biomass for bioenergy production has led to a situation where removal of logging residue from both thinning stands and clear-cuttings is becoming more common. Repeated removal of logging residue in harvesting has tended to decrease both C mineralization and net N mineralization in thinned Norway spruce stands 10–19 years after the last harvest, compared to a situation where logging residue was retained on the site (Smolander et al. 2008, 2010a). In a study performed in a Scots pine stand 4 years after thinning, rates of both net N- and C mineralization were higher with larger amounts of logging residue retained on the soil surface (Smolander et al. 2010b). On some other sites, decreased net mineralization of N over the long term has also been observed after removal of logging residue (Piatek and Lee Allen 1999). In several Nordic Norway spruce and Scots pine stands, removal of logging residue resulted in a long-term decrease in tree volume growth (Jacobson et al. 2000). This decrease can probably be partly explained as being due to removal of nutrients with the logging residue. Whether it is also explained by long-term lowered net N mineralization is not known.
Logging residue, which consists of tree tops, branches and needles, contains large amounts of many plant secondary compounds (Obst 1998). Therefore whole-tree harvest leads to a decreased input of tree secondary compounds compared to stem-only harvest. There is, for example, much resin, containing different terpenes, in the residue. In a study performed after repeated removal of logging residue, 10–19 years after the last treatment, concentrations of sesqui- and diterpenes were lower in whole-tree harvest plots where logging residue had been removed than in stem-only harvest plots where residue had been left on the site, whereas there were no differences in total phenolic compounds and condensed tannins (Smolander et al. 2010a). The situation was different in a Scots pine stand 4 years after treatment: the larger the amount of logging residue, the higher was the concentration of condensed tannins but with terpenes there were no major differences (Smolander et al. 2010b). It should be noted that in both studies, soil C and N transformations reacted to the removal of logging residue in the same soils as did the plant secondary compounds.
We can only speculate as to how long-term changes in C and N transformations due to logging residue removal are related to smaller input of different types of plant secondary compounds. However, it was suggested that in the long-term higher terpenes may offer C sources for part of the soil microbial population and that phenolic compounds may have a quicker effect on soil C and N transformations (Smolander et al. 2010a, b). This is based simply on the different timing with which these groups of compounds may enter the humus layer. At least part of phenolic compounds may enter the humus layer by leaching from the residue above, but terpenes enter it after the decaying residue starts to be part of the humus layer, which may take years. With monoterpenes it is not known how much they are volatilized before entering the humus layer. Both short- and long-term studies are necessary to compare the effects of traditional stem-only harvest and whole-tree harvest on soil secondary compound composition.
Considerable research has been done on plant secondary compounds, especially on tannins; but we still are at the beginning in terms of considering their importance for regulation C and N cycling in forest soil. Plant secondary compounds are extremely challenging to study; the structure of condensed tannins varies and may affect their reactivity, so there are no appropriate commercial standards available; some terpenes are volatile, etc. So far, in laboratory experiments the ability of several compounds to regulate C and N transformations in forest soils has been demonstrated; and in theory, we know that this can happen in soil. We also know something about the mechanisms behind these phenomena: binding organic N compounds, acting as a C source for the soil microbial population or being toxic. However, there are two problems in interpretation of laboratory experiments. The first is methodological and connected particularly with tannins extracted from plants: It is not always very clear what exactly has been added to soil, i.e. to which compounds or compound groups the soil is exposed, or whether other compounds that are included, some of which are not even phenolics, can cause some of the observed effects. Therefore it is important to make a 100% chemical characterization, and as carefully as possible. Another problem is more general and is always connected to this kind of experiments: How can we extrapolate results from laboratory experiments to the field where the input of plant secondary compounds to the soil occurs together with hundreds of other compounds and many other factors differ (i.e. the presence of living roots and their associated mycorrhizae, to mention only one). In any case, some correlation studies in the field point to relationships similar to those we have seen in the laboratory, while others are more contradictory.
In addition to questioning the importance of the effects of plant secondary compounds on soil C and N transformations, there are also many gaps in our knowledge concerning more specific points. A good example is the fact that surprisingly little is known about the occurrence of higher terpenes in soils and about their effects on C and N transformations. On monoterpenes some studies are available but more detailed information and, in particular, studies performed in different forest ecosystems are needed. Coniferous resin may form a point-like source of monoterpene emissions. Therefore, with volatile monoterpenes, one relevant research problem is their effect on spatial variation in soil microbial activities. With regard to phenolic compounds, particularly tannins, attempts to link structure more definitely to function are needed. This is a difficult task since even if plant tannins can be purified from extraneous compounds, they still may not be single pure compounds but mixtures of structurally different tannin polymers. Therefore we are often making conclusions based on averages of different structural parameters. Finally, a more general problem is that typically different research groups concentrate on research on terpenes or phenolic compounds, terpenes being studied much less than phenolic compounds. However, we need also studies where the importance of both compound groups as N-retaining agents in forest soils are evaluated in the same forest ecosystems together.
Results obtained so far indicate that in certain cases plant secondary compounds may play a role of utmost importance, for example, the ability of monoterpenes to decrease N mineralization and to inhibit net nitrification completely. If more N remains in organic form due to plant secondary compounds, what does this mean? From the environmental standpoint, it probably means less harmful losses of N, although dissolved organic N may also leach. From the standpoint of plant nutrition, this question is more difficult to answer, but in the short-term these secondary compounds seem to decrease the availability of N to plants.
How relevant is regulation of soil C and N cycling by plant secondary compounds, and what are the impacts of this possible regulation in a forest ecosystem? These are questions we can only speculate about, although changes in plant secondary compound composition well explain certain changes in N transformations occurring after forest management, like tree species change or clear-cutting. However, the importance of plant secondary compounds in relation to other soil characteristics probably depends to a great extent on the ecosystem in question.
We are grateful to Dr. Oili Kiikkilä and Prof. Heljä-Sisko Helmisaari for thoughtful comments, Dr. Joann von Weissenberg for checking the English language of this paper, to Anne Siika for making the figures and to previous members of the research group, particularly Dr. Laura Höijer and Dr. Outi Priha, for their contribution. The Finnish Forest Research Institute, Academy of Finland and Maj and Tor Nessling Foundation have funded several studies discussed in this paper.