Anatomical and chemical defenses of conifer bark against bark beetles and other pests

Summary

Conifers are long-lived organisms, and part of their success is due to their potent defense mechanisms. This review focuses on bark defenses, a front line against organisms trying to reach the nutrient-rich phloem. A major breach of the bark can lead to tree death, as evidenced by the millions of trees killed every year by specialized bark-invading insects. Different defense strategies have arisen in conifer lineages, but the general strategy is one of overlapping constitutive mechanical and chemical defenses overlaid with the capacity to up-regulate additional defenses. The defense strategy incorporates a graded response from ‘repel’, through ‘defend’ and ‘kill’, to ‘compartmentalize’, depending upon the advance of the invading organism. Using a combination of toxic and polymer chemistry, anatomical structures and their placement, and inducible defenses, conifers have evolved bark defense mechanisms that work against a variety of pests. However, these can be overcome by strategies including aggregation pheromones of bark beetles and introduction of virulent phytopathogens. The defense structures and chemicals in conifer bark are reviewed and questions about their coevolution with bark beetles are discussed.

	Contents
	Summary	353
I.	Introduction	354
II.	Integrating defense strategies	354
III.	Constitutive defense systems: first tier of defense	356
IV.	Inducible defense systems: second tier of defense	363
V.	Genetics and physiology of bark defense	365
VI.	Bark beetles: diminutive but deadly	367
VII.	The arms race: coevolution of conifer defense and bark beetle strategies	367
VIII.	Bark-beetle-vectored blue-stain fungi: multiple attacks against multiple defenses	369
IX.	Conclusions	370
	Acknowledgements	371
	References	371

I. Introduction

Conifers are a remarkably successful group of plants, and effective defense strategies are part of the key to their success and their colonization of diverse habitats, from alpine and arctic timberlines to subtropical swamps. During their long life span of decades to over 4000 yr (e.g. Pinus longaeva; Schulman, 1954), they are subject to numerous attacks from a wide range of organisms. This review focuses on the defensive capacity of the bark of the stem, or bole, consisting of the periderm and secondary phloem, which is of particular interest because of the evolution of bark-feeding beetles specialized in killing trees. The bark is rich in organic nutrients and is a target of many different organisms, including insects, vertebrates, fungi and bacteria. Between the bark and xylem is the cambium, which is responsible for secondary growth of the stem. Through partitioning of nutrients to defense in stems, a considerable investment is made in protecting the major pathways for transport of water and nutrients: from roots to shoots via the xylem and from mature leaves to growing parts via the phloem.

Underlying the bark defense strategies is the complex dynamics that are often present with respect to multiple organisms involved in an attack (e.g. mutualism or commensalism). A wound created by one organism provides an opportunity for other organisms that might not be able to breach the outer surface of the bark on their own. A prominent example is aggressive bark beetles that can kill healthy trees under certain conditions. These beetles enter the bark, breed, create tunnels (‘galleries’), and lay their eggs in the region of the secondary phloem/cambial zone. At the same time, they introduce pathogenic fungi that can rapidly destroy the phloem and cambium and can reduce the defenses of the stem, thus providing a suitable environment for growth of the larvae. Pathogens can also disrupt the xylem flow of water by causing embolism in the tracheids (Sperry & Tyree, 1988). If the infection is extensive around the bole, it will lead to death of the tree. Conifers have evolved defenses to inhibit or repel bark beetles and to kill or arrest the growth of the pathogens they vector (Lewisohn et al., 1991; Woodward, 1992; Oven & Torelli, 1994). These defenses are obviously quite effective, as can be seen by the dominance of conifer trees in many habitats in the presence of a multitude of enemies. However, it is not uncommon to have major losses of conifer trees due to outbreaks of beetles. A fundamental understanding of conifer defense mechanisms is thus important to elucidating the genetic and ecological limitations of these defenses.

Generating a defensive capacity that can deal with the diversity of potential attackers and their varying means of attack requires a wide range of mechanisms. As will be described here, the basic defense strategy of conifer stems involves multiple structures and chemicals that are overlapping in space and time. Because defense is a costly venture, not all defenses are expressed constitutively under normal growth. Investment in constitutive defenses provides for inhibition of an initial attack, whereas inducible defense mechanisms help to ensure that an initial invasion of tissues is both perceived and defended against actively and vigorously. Thus, two basic types of defense strategies can be discussed: constitutive defenses that are present in the tree without any challenge, and inducible defenses that are generated upon perception of a foreign challenge. A third strategy, acquired or systemic defense, can be considered to be a variation of inducible defenses but at some distance from the attack, temporally displaced with respect to the initial event, and with persistent properties. The ‘choice’ of strategy has been hypothesized to vary with the type of challenge (cf. Matson & Hain, 1985; Christiansen et al., 1987; Bonello et al., 2001). Here we review the nature of the basic bark defenses and how they are integrated. Questions concerning their evolution and coevolution with pests are discussed, focusing on scolytid bark beetles because of their dominating role in forest ecology and the large amount of research done with these organisms.

II. Integrating defense strategies

1. Basic defense strategies

The active phloem and cambium of stems represent a relatively small amount of tissue that can be easily damaged or destroyed, resulting in death of part, or all, of the tree. The basic function of bark defenses is to protect the nutrient- and energy-rich phloem, the vital meristematic region of the vascular cambium, and the transpiration stream in the sapwood. There are four basic steps or phases of defense systems in plants that are independent of the attacking organism. The first is an effective constitutive defense that can repel or inhibit invasion of tissues. If this is not effective, the next stage is to kill or compartmentalize the invading organism. A third phase of defense is to seal and repair damage incurred so that the plant can continue to function normally, and so that opportunistic infections are prevented. Finally, acquired or systemic resistance can be induced so that future attacks are more easily defended against. In addition, once an invading organism is identified, more specialized inducible defense responses can be elicited, such as gene-for-gene (R-gene) responses. The combination of constitutive and inducible systems provides a potent defense against attack.

2. Integrating constitutive and inducible defenses

Constitutive and inducible defenses are under different regulation, but must be coordinated to achieve optimal effect. Utilizing these defenses in a temporal sequence provides a multitiered system that can be regulated as a series of resistances. While each of the constitutive defenses are at a set resistance that is determined by genetics and prior history, the inducible defenses are of variable resistance, the value being determined by the nature of the attack as well as the genetic ability of the individual. This multitiered system has a spatial and a temporal component. The spatial component is determined by the disposition of constitutive defenses from the periderm surface to the cambial zone, and can be compared to the defenses of a medieval castle consisting of multiple rings of defenses like moats and walls, and inner battlements (Fig. 1a). Conifer bark also has concentric layers of mechanical and chemical defenses as indicated in Fig. 1(b,c). However, neither castles nor tree trunks can exist without some means of exchange with their surroundings. As a fortress has doors, windows, etc., the tree has lenticels, rays and cracks in the cortex, which may present potential weak points in case of an attack. The temporal component of the system consists of seasonal or continuous production of defenses in new tissues, and enhanced production, or production of different defenses, in attacked tissues. Inducible defenses are further differentiated into relatively rapid responses, such as a hypersensitive response and synthesis of pathogenesis-related (PR) proteins and toxic chemicals, and longer-term responses such as formation of traumatic resin ducts and wound periderms.

Figure 1

General pattern of constitutive defenses in conifer bark presents concentric layers of elements analogous to concentric castle defenses. (a) Diagram of a concentric castle, showing rings of defenses with moat, inner and outer walls and towers. (b) Diagram of concentric defenses of conifer bark with periderm, and layers of polyphenolic parenchyma (PP) cells and sclerenchyma, and resin-producing structures. (c) Drawings of two different basic types of conifer bark, one with fiber rows and one with stone cells, showing multiple concentric layers of defensive materials.

3. Cost–benefit analysis of defense

Plants must balance the cost of generalized defenses and pest-specific defenses vs growth and reproduction for an effective strategy against invading organisms. Assimilates used for defensive structures and chemicals cannot be used to support growth and reproduction, thus partitioning of resources to defense must be cost-effective (Wright et al., 1979; Christiansen & Ericsson, 1986; Miller & Berryman, 1986; Christiansen et al., 1987; Kozlowski, 1992; Christiansen & Fjone, 1993). During an initial phase of attack, a plant will have little or no information on the specific organism involved, and at this stage generalized broad-based defenses are important. Although investment into these defenses can be significant, the cost to the plant is warranted by the potential destruction that can occur in the absence of such defenses, particularly for long-lived species that do not reproduce for many years, such as many of the conifers.

Costs of defense are mitigated by the two-tier system of constitutive defense followed by inducible defenses. Constitutive defenses are a ‘fixed cost’ to the plant – insurance against inevitable attack. This cost can be scaled up or down based on recent history; if the plant was attacked early in the season and the attack was defeated, systemic or acquired resistance mechanisms may have been employed which enhance the subsequent resting levels of defensive materials. Inducible defenses are costly, adaptive responses (Karban & Baldwin, 1997) but can be further fine-tuned based on the severity of the attack, and allow for timing of the defense cost to conditions when they are specifically needed (Baldwin, 1998). However, elegant work by Baldwin and coworkers has demonstrated that there is a fitness cost associated with inducible defenses when the plant escapes attack (Baldwin et al., 1990; Baldwin, 1998). This work was done with herbaceous annuals, and it is more difficult to assess if such fitness costs of inducible defenses apply to long-lived conifer trees that may have large nutrient reserves to draw upon and a very different reproductive phenology.

The cost of constitutive defenses can be further mitigated by utilizing mechanisms that ‘add value’ to the overall structure of the plant. For instance, extensive formation of fibers can strengthen the structure of the bark, provide resistance to fire and insulate the cambium and phloem from severe environmental conditions, as well as provide a barrier to insects and microorganisms. Another strategy is to use compounds that might otherwise be waste; for example, excess calcium can be precipitated as calcium oxalate crystals (Volk et al., 2002; Mazen et al., 2004). If of adequate size or number, such crystals can provide resistance to chewing insects or animals attacking the bark (Hudgins et al., 2003b).

III. Constitutive defense systems: first tier of defense

1. General strategy

The constitutive defenses of conifers are ‘generalized’ defenses against a range of organisms that might be expected to try to penetrate the bark during the history of a tree. Constitutive production of a suite of stable defense products of varying properties provides a generalized capacity for resistance to a broad range of organisms. In addition, opportunistic secondary invasions are common, and because they are likely to be a different organism than the original invader it is useful to have multiple and overlapping constitutive defenses present (Fig. 1b,c).

Constitutive defenses are of two basic types: mechanical and chemical. Mechanical defenses are made of structural elements that provide a toughness or thickness to the tissues and inhibit chewing or pulling apart of the bark. They may also include ‘spines’ that can be effective against larger animals – for example, tough, pointed cauline leaves as in monkey puzzle trees. Impregnation of tissues with polymers such as lignins and suberin can add to the mechanical properties and enhance resistance to penetration, degradation and ingestion/chewing, particularly by smaller organisms. Constitutive chemical defenses often occur as pools of stored chemicals (e.g. phenolics, terpenoids and alkaloids) that can be released upon attack. Chemical defenses include antifeedants, toxins, defensive proteins and enzymes, and reservoirs of chemicals such as resins that can flush away, repel or physically entrap small organisms. These defenses are dispersed throughout the various tissues of the bark, which include the periderm, cortex (if the stem is young) and secondary phloem. The secondary xylem can also contain some of these defenses as constitutive products (cf. Pearce, 1996; Fäldt et al., 2003).

A description of the physical and chemical nature of the various constitutive defense structures and their distribution within the bark and among species is important to developing an understanding of defense strategies of conifers. A summary of the major defensive structures of the bark is presented, although this is neither comprehensive nor does it point out all the nuances inherent in the biological diversity of such structures. This description provides the basis for the conceptual framework of integrated defense strategies in bark of conifers.

2. Periderm, the outer defenses

The periderm provides a permeability barrier for control of gas exchange in the stem, and also is the first line of defense from adverse biotic and abiotic factors. It consists of the lateral meristem of the periderm (the phellogen or cork cambium) and its products (phellem or cork tissue outwards and phelloderm inwards). Periderms are quite variable among species but have some common components. They are generally characterized by the presence of multiple layers of structurally and chemically distinct cells (Fig. 2a,b). The phellem has layers of cells, mostly dead, that have lignified or suberized walls. In addition, the cells may contain large amounts of phenolic materials, and it is also common to have one or more layers of cells that are encrusted with calcium oxalate crystals (Fig. 2c). The combination of the mechanical properties of tough lignified walls and crystals, the suberized walls that provide a hydrophobic barrier and the chemical properties of phenolics, presents a multifunctional barrier to the external environment. The periderm is not a continuous barrier because there are structures present called lenticels that allow for gas exchange at the surface. The lenticels, although not a wide-open system, can be viewed as potential entry points for invasive organisms. Small insects such as the six-toothed bark beetle Pityogenes chalcographus have been shown to enter the bark through lenticels of Picea abies, and it is possible that larger species such as the spruce bark beetle Ips typographus also utilize the lenticels to facilitate their entry (Rosner & Führer, 2002). It seems likely that the bark beetles sense chemical signals emanating through the lenticels, although this remains to be determined.

Figure 2

Transverse sections showing basic anatomy of the periderm, cortex and secondary phloem of conifer stems. (a) Bright field image of Picea abies periderm. Liginified (L), suberized (S) and phenolic (P) cells indicated. (b) Autofluorescence of lignin, suberin and phenolics of a similar periderm section. (c) Same section viewed with partially crossed polarizing filters, showing calcium oxalate crystals in the suberized and lignified cell layers. (d) Cortex and secondary phloem and xylem in a young Abies balsamea stem. The cortex has an axial resin duct and many of the parenchyma cells have accumulated phenolics. Two layers of polyphenolic parenchyma (PP) cells can be seen in the phloem, separated by sieve cells. Bars, 100 µm.

3. Cortex

The cortex is produced during the primary development of the stem, and may stay alive for a number of years of secondary growth. It commonly contains large amounts of phenolics within vacuoles of cortical parenchyma, and in many of the Pinaceae axial resin ducts are present (Fig. 2d). Sclerenchyma and calcium oxalate crystals may also be formed within the cortex. Eventually, the cortex is either crushed, or shed by the activity of deep periderms. Although the cortex is an important defensive barrier during early stages of growth of the stem, the secondary phloem takes over this role once it develops more extensive layers over a number of years.

4. Secondary phloem, the inner defenses

The secondary phloem is a major site of constitutive defense mechanisms. Most conifers share three common constitutive defense components in their secondary phloem: sclerenchyma, calcium oxalate crystals and phenolic bodies, although the relative amount of each varies considerably among species (Hudgins & Franceschi, 2004). A fourth defense that is common in certain taxa is the presence of resin-producing structures such as radial resin ducts (extending into the xylem), axial resin ducts, resin blisters and resin cells. These will be discussed separately, because some conifers do not produce constitutive resin structures in the secondary phloem. The amount and combination of each of these components defines distinct defense strategies and likely different evolutionary pressures.

Bark of all conifer families studied has axial phloem parenchyma cells that are specialized for synthesis and accumulation of phenolic compounds (Fig. 3; Hudgins et al., 2003a; Hudgins et al., 2004). These are referred to as polyphenolic parenchyma cells (PP cells; Franceschi et al., 1998; Krekling et al., 2000). Within their vacuoles, PP cells contain variable amounts of phenolic bodies (Fig. 3) that are thought to serve as antifeedants and as antifungal agents (Beckman, 2000). In addition to their phenolic contents, they also have thickened cell walls, albeit with abundant plasmodesmata that allow for axial and tangential exchange of information, and possibly for defense signaling (Krekling et al., 2000). PP cells are commonly produced as an annual tangential ring, or cylinder, of cells and so can be used to determine age of the region of secondary phloem (Fig. 3a). As the rows of PP cells are ‘displaced’ further away from the cambium through its activity in secondary growth, they become larger and may even divide to produce more PP cells (Krekling et al., 2000). It is significant that PP cells more than 70 yr old in 100-yr-old Norway spruce (Picea abies (L.) Karst.) bark were found to be still living (Krekling et al., 2000). This, along with their presence in all conifer families, attests to their importance to defense reactions in general.

Figure 3

The anatomy and distribution of polyphenolic parenchyma (PP) cells in Picea abies secondary phloem. (a) Transverse section showing annual layers of PP cells separated by sieve cell (S) layers. R, radial ray. Bar, 200 µm. (b) Enlargement of PP cells (radial section) showing the large vacuole (V) filled with phenolics and storage of starch (St) in these living cells. N, nucleus. Bar, 20 µm. (c) Tangential section through a layer of PP cells. Bar, 100 µm. (d) Radial section with files of cylindrical PP cells separated by layers of sieve cells. Bar, 100 µm.

It is unfortunate that the chemistry of the phenolic components found in PP cells has not been extensively studied. One might expect that different species produce different phenolic components, dependent upon the type of organisms commonly attacking them. There is also some evidence that relative resistance to pathogens might partly be a function of the type of phenolics produced (Brignolas et al., 1995; Bonello et al., 2003). The phenolics deposited are not static structures, as shown by studies of changes in PP cell contents on a seasonal basis (Krekling et al., 2000), and, as will be discussed, PP cells are also responsible for important inducible defense responses. Overall, PP cells represent a very dynamic component of the defense strategy of conifer bark, and because they are the most abundant living cell type in the secondary phloem, the argument can be made that they are the single most important component. The PP cells, along with the ray parenchyma, are also a major site for storage of starch and/or lipids (Fig. 3b; Krekling et al., 2000). In this capacity, they can be seen as a target for beetles and fungi, and constitutive phenolics can be hypothesized to protect the cells themselves as well as prevent fungal penetration towards the cambial zone. In any case, the multiple layers of PP cells between which are the crushed remnants of the sieve cells can be viewed as providing physical and chemical resistance to penetration of the bark (Franceschi et al., 2000).

Sclerenchyma is a mechanical tissue common to all conifer bark, although the amount and type produced varies dramatically among conifer taxa. Sclerenchyma is made of cells with lignified secondary wall thickenings, which can serve as structural elements as well as mechanical defense. They can occur as massive, often irregular-shaped stone cells (sclereids), such as can be seen in many of the Pinaceae (Fig. 4a–d), or as well organized rows of fibers, as in the Taxaceae (Fig. 4e; Hudgins et al., 2004). Stone cells can be present as individual cells or as clusters of cells of variable size (Fig. 4a–d). Their physical toughness can be a deterrent to grazing or bark-boring organisms, particularly if they are abundant (Wainhouse et al., 1990, 1997, 1998; Hudgins et al., 2004).

Figure 4

Sclerenchyma is common in conifer bark but has different forms and distribution patterns. All are transverse sections except (c), which is tangential. (a) Larix occidentalis has fibers (arrows) scattered throughout the secondary phloem. Bar, 100 µm. (b) Picea pungens has large patches of stone cells (SC). Bar, 100 µm. (c) Tangential section of Tsuga heterophylla gives a good representation of the amount of lignified tissue that can be present as stone cells. Bar, 50 µm. (d) Enlargement of a stone cell cluster in Picea abies. The stone cells are generally derived from polyphenolic parenchyma (PP) cells as the phloem ages. Bar, 100 µm. (e) Taxodium sp. showing fibers (arrows) arranged as tangential rows separated by two rows of sieve cells between which is a row of PP cells. Bar, 100 µm.

In contrast to the more or less scattered distribution of stone cells, fibers occur as distinct tangential rows of cells interspersed in a highly ordered organization with rows of PP cells and sieve cells (Fig. 4e). Unlike stone cells, fibers develop from a layer of precursor cells that are distinct from PP cell layers, and they constitute a complete tangential layer of lignified cells disrupted only by radial rays. Full lignification of fibers occurs only after 2–3 yr, so that the two layers of precursor cells closest to the cambium are unlignified or have only partially thickened walls during normal growth (Hudgins et al., 2004). In general, layers of fibers are separated by much fewer layers of other cells (3–5 nonfiber rows) as compared with species with stone cells (7–12 nonstone cell rows between), and appear visually to provide a much more formidable mechanical barrier. In mature secondary phloem, stone cells and fibers are not generally found together. This may be partly due to separate evolution of these mechanical defenses, because stone cells are most abundant in the Pinaceae, whereas fibers are found in other families, and these groups represent a major split in the evolution of conifers (Fig. 5).

Figure 5

The presence or absence of various bark defenses is mapped against a conifer phylogeny. Note that aggressive, tree killing bark beetles are primarily found in the Pinaceae. (a) Distribution of the mechanical defense components, calcium oxalate crystals and sclerenchyma. (b) Distribution of a chemical defense component, resin-producing structures.

The sieve cells themselves may provide some mechanical resistance against penetration. The layers of sieve cells older than 3 or 4 yr, which are dead and no longer participate in assimilate translocation, progressively collapse under the pressure of new cell layers interior to them. They thus form, at least in the Pinaceae where 6–12 layers are produced per year, thick layers of cell wall material.

Calcium oxalate crystals are another potential mechanical defense component that is common in the secondary phloem of conifers. They occur as intracellular deposits in the Pinaceae and extracellular wall deposits in the non-Pinaceae (Fig. 6; Hudgins et al., 2003b), as well as in the periderm of various conifer species (Fig. 2). Although little is known about their formation or function in conifers, the tough physical nature of the crystals and their relative abundance would implicate a role in deterrence of bark-boring or -chewing animals. Because they are chemically inert, they are unlikely to have any effect on fungal attack. Calcium oxalate crystals are common in plants (reviewed by Arnott & Pautard, 1970; Franceschi & Horner, 1980; Webb, 1999; Nakata, 2003; Franceschi & Nakata, 2005), and while a role in bulk calcium regulation has been demonstrated (Franceschi, 1989; Volk et al., 2002; Mazen et al., 2004), they clearly have also evolved secondary functions in defense (Franceschi, 2001; Franceschi & Nakata, 2005). The Pinaceae produce calcium oxalate crystals embedded within the phenolic bodies in vacuoles of modified PP cells, typically as scattered axial strands of crystalliferous cells (Fig. 6c). At maturity, the cell walls of these modified PP cells have a suberin layer and the cells are no longer living, unlike the normal PP cells. In contrast, crystals of the non-Pinaceae are deposited within or along the walls of various cell types of the secondary phloem, depending upon the species (Fig. 6d–f). They are particularly abundant in the Taxodiaceae, where almost all cell walls are heavily encrusted with small crystals, including the walls of fibers. The combination of multiple layers of fibers and heavy incrustation with crystals may provide a particularly potent defense against bark beetles (see Fig. 5a), as proposed by Hudgins et al. (2003b).

Figure 6

The distribution of calcium oxalate crystals in various conifer species. Panels (c–f) are scanning electron microscope back-scattered electron (BSE) images where crystals appear as bright particles. (a) Young Pinus sp. stem in transverse section. C, cortex; P, phloem; X, xylem. Bar, 250 µm, (b) The same Pinus section viewed with crossed polarizing filters, with crystals showing up as bright spots in the phloem. Bar, 250 µm. (c) A crystal parenchyma cell in Cedrus libani with four prismatic crystals (arrows). Bar, 5 µm. (d) Transverse section of Taxus brevifolia phloem. Crystals appear as white spots along the walls of the sieve elements and ray parenchyma. Bar, 50 µm. (e) Tangential section of Juniperus scopulonum phloem. Crystals are abundant along the walls of almost all cell types. Bar, 50 µm. (f) Radial section of Chamaecyparis nootkatensis. Crystals are abundant along the walls of most cell types. Bar, 50 µm.

A major constitutive defense of particular importance to the Pinaceae is resin-producing and -storing structures. These include radial resin ducts derived from radial rays, axial resin ducts or canals, resin blisters and resin cells (Fig. 7a–c). The ducts and blisters have a lining of plastid-enriched epithelial cells (Fig. 7d) that synthesize terpenoid resins and secrete them into an extracellular lumen where they accumulate under pressure. Resin cells accumulate resin internally under pressure and expand into fairly large structures (Fig. 7c). Upon damage by wounding or by an invading organism, the pressurized resin is released and can repel or flush the organism out of the bark, entrap the organism in sticky resin or otherwise kill the invader due to the toxic nature of the resin. Resin flow from radial phloem ducts can be enhanced by their connection to constitutive or induced axial resin ducts in the xylem (Christiansen et al., 1999a; Nagy et al., 2000). When the volatile components of the released resin evaporate, the nonvolatile components will crystallize to sterilize and seal the wounded region effectively. Whereas resin-producing structures are found in all Pinaceae, they do not occur constitutively in the secondary phloem of some other conifer taxa (Hudgins et al., 2003a, 2004). It is interesting to note that those species that lack constitutive phloem resin structures are the same species that have tangential fiber rows and high amounts of calcium oxalate crystals (see Fig. 5b). Although resin is considered to be a defense against bark-boring insects, and although a correlation between resin duct number and resistance to pine weevil has been shown (Alfaro et al., 1997), it is clear that other strategies have evolved that can work equally or more effectively (Hudgins et al., 2004). The chemistry and biochemistry of conifer resins has been extensively reviewed (Gershenzon & Croteau, 1991; Bohlmann et al., 1998; Phillips & Croteau, 1999, and others) and so is not covered here.

Figure 7

Resin-producing structures found in conifer secondary phloem. (a) Radial resin duct in Pseudotsuga menziesii. Bar, 100 µm. (b) Axial resin duct in Thuja plicata. Bar, 100 µm. (c) Resin cells in Abies grandis. Bar, 50 µm. (d) Epithelial cells (E) of a resin duct from Picea abies. Bar, 10 µm.

The role of radial ray cells in constitutive defense has not been determined. These cells in conifers tend to have thin walls, and generally lack accumulation of polyphenolics. However, in a few species that have been examined, they have an abundance of the enzyme phenylalanine ammonia lyase (PAL), a key enzyme in the phenylpropanoid pathway, and it has been speculated that they may be involved in production of soluble phenolics that can be secreted (Franceschi et al., 1998; Hudgins et al., 2004). Because they represent the main radial pathway for movement of information in the bark, ray parenchyma likely function in signal sensing and transduction at some level (Hudgins & Franceschi, 2004). Recent work has shown that they contain 1-aminocyclopropane-1-carboxycylic acid oxidase, a key enzyme in ethylene biosynthesis, and that they may be involved in ethylene-mediated responses to attack (Hudgins & Franceschi, 2004). However, the ray parenchyma also provides a nutrient-rich radial infection route for pathogenic fungi, and therefore rays may need their own defense mechanisms.

5. Secondary xylem

Although the focus of this review is on the secondary phloem, a quick note on constitutive defenses of xylem is warranted for reference, because defense strategies are overarching in the stem. Xylem parenchyma, both axial and radial systems, is present in the secondary xylem and can be involved in synthesis and storage of phenolics and resin, as well as other secondary products. Xylem parenchyma is also involved in synthesis and secretion of ‘extractives’, such as lignans, into the heartwood of conifers (Kwon et al., 2001), which provide a defense against wood rotting fungi and other organisms. The radial rays appear to be the only means of translocation of phenolics or their precursors to the border between sapwood and heartwood, to promote the development of phenol-impregnated heartwood, or reaction zones that form to stop the spreading of heart rot (Shain, 1967, 1971). Constitutive axial resin ducts are also present in the xylem of some conifers (Fig. 5b; Wu & Hu, 1997), and may contribute to resin flow when they are connected to radial resin ducts that traverse between xylem and phloem.

IV. Inducible defense systems: second tier of defense

Inducible defenses enhance the overall defense capacity of the plant. They include a combination of responses seemingly targeted at the specific organism, increased generalized responses, and processes aimed at limiting the extent of the damage inflicted by an organism, walling it off and sealing any wounded tissue. Acquired resistance is a long-term consequence of inducible defense. Inducible defenses are diverse and include structural changes and synthesis of chemical and biochemical agents.

1. Inducible structural defenses

Structural defenses occurring in the bark are important in containing and isolating an invasive organism, repairing damaged tissue, and limiting opportunistic or subsequent attack or invasions. The hypersensitive response can be thought of in part as a structural defense, although the endpoint is dead tissue. The hypersensitive response occurs locally at the site of infection or attack (Bleiker & Uzunovic, 2004) and results in production of reactive oxygen species, and rapid cell death that is intended to kill and contain organisms such as fungal pathogens, bacteria and viruses. This ‘scorched earth’ defense (Berryman, 1972) sacrifices a small volume of tissue in an attempt at rapid containment of invading organisms. Beyond this highly localized response, a more generalized response to wounding is the formation of callus tissue that can subsequently become lignified, suberized or impregnated with phenolics to provide a barrier, and may form part of the wound periderm. This reaction provides protection against further intrusion as well as walling off an organism such as a fungal pathogen. Wound callus also can repair damaged tissue so that continuity of function can be re-established – for instance, by reorganization of a region of damaged cambium.

Wound periderms are induced to form around an invaded or damaged region of bark, and serve to wall off the region as well as to re-establish a continuous surface barrier in the previously disrupted area. Wound periderms are produced by activation of PP cells in the secondary phloem, as well as callus tissues in severe wounds, which begin to divide to form a wound phellogen (Fig. 8). This meristem will then produce the tissues, phellem and phelloderm, typical of a normal periderm (Fig. 8b). The wound periderm essentially isolates the damaged area and effects a permanent repair of tissues. Wound periderms form at the boundaries of lesions induced by bark beetle (Christiansen & Kucera, 1999) or fungal attacks in the stems of conifers (Fig. 8a), but will form around any damaged tissue. Damaged or infected tissue is isolated from nutrient supplies by the wound periderm, and will eventually die if not already killed by an invasive organism.

Figure 8

Induction of wound periderm and polyphenolic parenchyma (PP) cell activation in Picea abies secondary phloem after inoculation with the blue-stain fungus, Ceratocystis polonica. (a) Wound periderm (WP) anatomy. Bar, 200 µm (b) Wound periderm showing the phellogen (PG) and its derivatives of phelloderm (PD) and suberized (SC) and lignified (L) cells of the phellem. Bar, 100 µm (c) PP cell anatomy of control bark. Three rows of PP cells can be seen (numbered) separated by layers of sieve cells (S). Bar, 100 µm (d) PP cell anatomy 3 week after inoculation. The PP cells have swollen, accumulated increased amounts of phenolics, and have crushed the surrounding sieve cells. Bar, 100 µm.

Another induced structural defense is early lignification of fibers (Hudgins et al., 2003a, 2004). As indicated in Section III.4, incipient fiber rows are produced yearly by the activity of the vascular cambium in many conifer species, but they do not become fully lignified until they are c. 3 yr old. Recent studies on induced defense responses demonstrated that these incipient fibers can be induced to become fully lignified after methyl jasmonate treatment. This compound generates a response similar to that of wounding, and is thought to be a signaling agent in normal wound responses of the bark of conifers (Franceschi et al., 2002), although the response may actually be mediated by ethylene induced by methyl jasmonate (Hudgins & Franceschi, 2004).

2. Inducible chemical defenses

Whereas constitutive chemical defenses are generally nonselective with respect to the pest species, inducible chemical defenses include both broad-spectrum and specific compounds. Chemical defenses are extremely diverse and can thus cover a range of pests. Nonprotein chemicals such as products of the phenylpropanoid pathway (phenolics) and isoprenoid pathway (terpenoid resins) as well as alkaloids can have potent effects on invasive organisms. They have the potential of being more quickly produced than protein-based defenses because the pathway often already exists in the tissues and just requires activation, as opposed to the need to produce mRNA and translate that to produce a protein-based defense. However, some of the biochemical pathways are created de novo in newly generated tissues or structures as well, a good example being resin produced in traumatic resin ducts (Thomson & Sifton, 1926; Bannan, 1936; Cheniclet, 1987; Christiansen et al., 1999a; Nagy et al., 2000; Martin et al., 2002; Fäldt et al., 2003; Hudgins et al., 2003a, 2004). Another advantage of these chemical defenses is that they are often effective against a broad range of organisms, and thus can slow down an attack while recognition mechanisms come into play to identify the organism and activate specific defenses against it. Most work with conifers has been done on phenolics, terpenoids and enzymes such as some of the PR proteins.

Phenolics are abundant in the bark of conifers (Pan & Lundgren, 1995, 1996; Viiri et al., 2001), occurring primarily in PP cells as discussed in Section III.4. Phenolics and tannins can act as antifungal agents and antifeedants, and can bind hydrolytic enzymes secreted by pests, thus inhibiting their advancement into tissues (Hunter, 1974; Nicholson & Hammerschmidt, 1992; Appel, 1993). By binding amino acids and proteins of the plant tissue when it is disrupted, phenolics and tannins reduce the nutrient value to the attacking organism, whereas binding to digestive enzymes in the gut further diminishes the ability to digest the plant tissue. Wounding or invasion of the bark has been shown to activate PP cells, which includes cell expansion and accumulation of increased amounts of phenolics (Fig. 8c,d; Klepzig et al., 1995; Franceschi et al., 2000; Kusumoto & Suzuki, 2003). Various chemical studies have shown that production of phenolics, or up-regulation of enzymes of the pathway (Richard et al., 2000), is rapidly induced in invaded bark and that the amount and type of phenolic compounds produced during invasion may be quite different from the constitutively produced phenolics (Brignolas et al., 1995; Lieutier et al., 1996; Bois & Lieutier, 1997; Viiri et al., 2001; Bonello & Blodget, 2003; Bonello et al., 2003). The implication is that the induced phenolics are more toxic or more specific to an invasive organism than are the constitutive phenolics. It is also possible that conversion of polyphenolics to soluble phenolics may occur, adding to the defense capacity, as indicated by the reduction of polyphenolics in vacuoles of intact PP cells close to the region of attack (Franceschi et al., 1998). The ability of bark PP cells to convert or change their vacuole polyphenolic contents is further demonstrated by studies of seasonal changes that occur (Krekling et al., 2000).

Terpenoid resin production can also be induced by attack. During and following an attack, resin flow from the wound can be quite extensive, especially in members of the Pinaceae. Part of this resin is from stored resin in existing resin-producing structures, and there is evidence that the constitutive ducts can be activated to produce more resin (Ruel et al., 1998; Lombardero et al., 2000; Hudgins & Franceschi, 2004). Within 2–3 wk after attack, new resin ducts can also be induced to form, referred to as traumatic resin ducts (Fig. 9; Alfaro, 1995; Alfaro et al., 1996; Tomlin et al., 1998; Byun McKay et al., 2003), and the resin formed by traumatic ducts can be different than constitutive resin (Nault & Alfaro, 2001; Martin et al., 2002; Miller et al., 2005). These ducts are formed above and below a damaged site or induced point on the stem (Ito, 1998; Franceschi et al., 2000; Nagy et al., 2000; Hudgins & Franceschi, 2004; Hudgins et al., 2004; Krekling et al., 2004). In the Pinaceae and some other conifer groups, the traumatic ducts are formed in the xylem by a reprogramming of the cambial zone xylem mother cells (Krekling et al., 2004), resulting in axial ducts being embedded in the new sapwood (Fig. 9; Hudgins et al., 2003a). These ducts are interconnected with the radial resin ducts of the phloem (Nagy et al., 2000). Some conifer species are induced to form traumatic ducts in the phloem rather than the xylem (Hudgins et al., 2004). Although the mechanisms that are signaling traumatic duct formation after an attack are not completely worked out, there is ample evidence that the octadecanoid pathway is involved as an early event. In many plants, tissue wounding results in the formation of jasmonates, which are involved in a host of defense responses (Farmer et al., 2003). Little is known about jasmonate signaling in conifers; however, recent studies have shown that exogenous application of methyl jasmonate to the surface of conifer stems in the absence of wounding induces formation of traumatic resin ducts (Franceschi et al., 2002; Martin et al., 2002; Hudgins et al., 2003a, 2004). The effect of exogenous jasmonate treatment shows a dose-dependent response (Hudgins & Franceschi, 2004), and other responses seen in wounding or bark beetle attack, such as PP cell activation, are also induced by methyl jasmonate application. Further evidence of involvement of jasmonates in traumatic duct induction is illustrated by the work of Miller et al. (2005), who demonstrated that transcripts for enzymes in the octadecanoid pathway are up-regulated in conifer stems by wounding. Recent evidence indicates that jasmonate-induced ethylene production is responsible for reprogramming of the cambial zone for traumatic duct formation, and ethylene may also induce activation of existing radial resin ducts to produce additional resin (Hudgins & Franceschi, 2004; Hudgins et al., 2004). The end result of traumatic resin duct development is increased resin formation and accumulation (Martin et al., 2002; Miller et al., 2005) and enhanced resin flow. Increased resin flow can help to kill or flush out invaders as well as seal the wound, and resin-soaked regions of bark and sapwood may also be more resistant to introduced microbial activity. In addition, there is evidence that traumatic ducts may impart acquired resistance to subsequent attack (Christiansen et al., 1999b; Krokene et al., 2003) and that the resin in traumatic ducts may be more toxic through changes in terpenoid components or addition of phenolics (Nagy et al., 2000).

Figure 9

Transverse section of Picea abies stem demonstrating the development of traumatic axial resin ducts in the new xylem after inoculation with the blue-stain fungus, Ceratocystis polonica. Bar, 200 µm.

Protein-based chemical defenses in trees include enzymes such as chitinases and glucanases that can degrade components of invasive organisms, toxic proteins such as porins and lectins, and inhibitors of enzymes such as proteinase and amylase inhibitors. The enzyme inhibitors interfere with the ability of the organism to utilize resources of the invaded tissue. Other inducible enzymes such as peroxidases and laccases can make cell walls tougher through crosslinking reactions or promotion of lignification, or can be involved in directly affecting the invasive organism. Protein-based defenses can be highly specific to a particular organism. For example, in Norway spruce, chitinases exist as a large family of proteins, but only a small subset of them may be up-regulated during attack by a specific fungal pathogen (Hietala et al., 2004; Nagy et al., 2004), and it is presumed that these are effective against cell walls of that particular organism.

Overall, inducible chemical defenses follow a similar mechanistic concept as inducible structural defenses; that is, multiple overlapping strategies. Production of a toxic cocktail of diverse chemical components will maximize the potential to halt or destroy an aggressive or virulent invading organism, in contrast to a more conservative production of one or a few more directed defenses.

V. Genetics and physiology of bark defense

1. Genetic component and genes of defense

The defense capacity and relative resistance of conifers will depend on both the genetics and the physiological status of the individual. Defense mechanisms and resistance to pests have a heritable component that has been studied in a number of conifers and with respect to a variety of pests. Genetic resistance can work through mechanisms as diverse as R-gene interactions (Liu & Ekramoddoullah, 2004a), changes in chemical pathways giving rise to a qualitative or quantitative change in defense chemicals (Brignolas et al., 1995; Lieutier et al., 1996; Bois & Lieutier, 1997; Tomlin et al., 1998; Bonello et al., 2003; Miller et al., 2005), or constitutive defenses (Wainhouse et al., 1990; O’Neil et al., 2002; Rosner & Hannrup, 2004). This genetic component is a product of the pressures under which the various taxa evolved or have adapted to, although parallel evolution of similar strategies is also apparent. For example, initial studies suggest that resin ducts had multiple independent lines of evolution in different conifer lineages (Fig. 5b; Hudgins et al., 2003a; Hudgins et al., 2004). They occur as quite diverse structures (radial resin canals, axial resin ducts, resin cells and resin blisters) and in both phloem and xylem, and thus come from completely different cell lineages.

Combinations of genetically diverse defense mechanisms can give rise to similar defense results, even though the genetic components are quite different. For example, the Taxodiaceae have constitutive expression of layers of phloem fibers but no phloem resin ducts, whereas the Pinaceae constitutively express phloem resin-producing structures, and in place of fibers produce what are arguably less effective stone cells (Fig. 5a). These two conifer families have genetic components of constitutive defense that are quite different, but both seem to be effective against bark beetles. However, both families are capable of inducible expression of traumatic axial resin ducts, and although the nature of the ducts and their resin components are different in the two taxa, this inducible defense strategy is the same. Overall, genetic makeup strictly controls the type of defenses present in the bark, how extensive they are, and how quickly or strongly inducible defenses are up-regulated in response to a signal. Thus, the relative resistance of individuals of a species or of related species to a particular pest depends initially on genetic makeup, which can be thought of as a sort of ‘genetic capacitance’.

The increasing amount of genetic information available for conifer species has provided powerful new tools to explore and compare defense-related genes in conifer systems. For example, cDNA libraries and arrays have been made that can be applied to experiments to examine a host of changes in expression of selective genes during infection or experimental treatments of conifer plants, as illustrated by the recent studies of Morse et al. (2004) and Miller et al. (2005). Most studies have looked at a few selected genes or members of a gene family involved in defense, which are only briefly cited here. There is a considerable amount of information on constitutive and inducible expression of genes of the terpenoid resin pathway in conifers (cf. Steele et al., 1998; Bohlmann et al., 1999; Fäldt et al., 2003; Martin et al., 2004). Some studies have looked at induction of phenylpropanoid pathway genes as well (Chiron et al., 2000; Richard et al., 2000; Nagy et al., 2004), but much work remains to be done on this important chemical defense pathway. Expression patterns of various PR proteins have been examined, but these usually deal with one or a few PR proteins (Liu & Ekramoddoullah, 2004a; Nagy et al., 2004; Piggott et al., 2004; Zamani et al., 2004; Asiegbu et al., 2005). Inducible expression of genes as diverse as peroxidases (Fossdal et al., 2001) and antimicrobial peptide (AMP; Li & Asiegbu, 2004a) have also been reported. Recent studies have explored differential expression of genes of potential regulatory or signaling pathways such as 14-3-3 genes (Lapointe et al., 2001), R-gene analogues or leucine rich repeats (LRR) genes (Richard et al., 1999; Li & Asiegbu, 2004b; Liu & Ekramoddoullah, 2004b; Asiegbu et al., 2005) and the octadecanoid pathway members (Miller et al., 2005). Using the growing molecular genetic resources, important new information on regulation of defenses, the range of defense genes present in various conifer species and the evolution of these defenses will become available.

2. Physiological components of defense

Overlaying the genetic predisposition relative to defense is the physiological condition of the individual. Maintaining constitutive defenses and expressing inducible defenses requires resources in the form of carbon and nitrogen. Physiological condition is thus another capacitance of the tree, which will modulate the expression of the genetic capacitance. The cumulative effects of abiotic and biotic stresses on an individual tree will be reflected in its ability to maintain basal constitutive defenses and to mount effective inducible defenses.

Inducible defenses in the bark are influenced not only by local nutrient stores but also by translocation of new photosynthate from needles (Christiansen & Ericsson, 1986; Miller & Berryman, 1986). Various abiotic factors, such as water stress, air pollution and temperature stress, as well as attacks by biotic agents, may alter the resources available for defense responses to such a degree that even relatively resistant genotypes become susceptible. At first glance, a tree growing under ideal conditions might be thought to have plenty of resources to invest in constitutive defense and divert to inducible defenses, as opposed to a stressed tree. However, the connection seems to be more complicated: stressed trees do not seem to be favorable for all insect guilds, although bark-boring insects generally appear to perform better on stressed trees (Koricheva et al., 1998). Bark beetles/fungi may be favored by water stress (e.g. Miller & Keen, 1960; Kalkstein, 1976; Croise & Lieutier, 1993; see also references in Christiansen & Bakke, 1988), but the degree of water stress appears to be crucial. Mild drought does not necessarily reduce the resistance to bark beetles and their associated fungi; there are indications of the opposite (Dunn & Lorio, 1993; Reeve et al., 1995; Christiansen & Glosli, 1996), including increases of resin defenses (Turtola et al., 2003).

To the extent that physiological resistance depends on resource availability, a useful strategy might be to invest to the maximum capacity in constitutive defenses when conditions are good, because energy and nutrient availability cannot be ensured even in the short term (and particularly not from season to season). This may not always be the case, however. White pine weevil (Pissoides strobi) attack of spruce increased with fertilization intensity, whereas constitutive resin canal defenses were decreased (vanAkker et al., 2004). Southern pines are more susceptible to southern pine beetle (Dendroctonus frontalis) attacks in the spring, when water stress is absent, than after the onset of summer drought. This variation may be explained by changes in the growth differentiation balance of the trees, as less energy will be available for defense when growth processes are at a maximum (Lorio, 1986). However, whereas constitutive resin flow from Pinus taeda followed this pattern, induced increases in resin flow were greatest during the season of fastest growth, indicating that constitutive and induced defenses may be differently influenced by environmental factors (Lombardero et al., 2000).

VI. Bark beetles: diminutive but deadly

This review of bark defenses indicates that conifers have potent anatomical and chemical mechanisms to protect the phloem and cambium. The success of these defenses is seen in the ability of conifers to live for hundreds or thousands of years, to colonize large tracts of land and to dominate many ecosystems. As potent as these defenses appear to be, they are not impenetrable and it is interesting that the most successful attacker is also one of the smallest.

1. Bark beetles: the enemy at the gate

A major challenge to bark defenses are bark beetles (Scolytidae). Bark beetles occur in a variety of habitats, but most species breed subcortically in trees (Wood, 1982; Kirkendall, 1983; Beaver, 1989). All species breed in galleries excavated in plant material, and live inside their host for most of their life cycle. True bark beetles, which include all tree killing or ‘aggressive’ species, excavate breeding galleries in the inner bark of trees, feeding directly on the phloem and cambial area, although a few species also feed on associated fungi. The bark beetle life cycle can be divided into three main phases: dispersal, colonization and development (Wood, 1972; Borden, 1982). During the dispersal phase, the new generation emerges and flies to a new tree. The colonization phase includes host selection and boring into the bark. In many species, numerous additional colonizers of the same species are attracted to the host by pheromones produced by the initial attackers (secondary attraction; Wood, 1982; Borden, 1985; Vanderwel & Oehlschlager, 1987). The development occurs within the host, and includes mating, gallery construction, oviposition and brood development.

Ecologically, bark beetles form a continuum from species that can colonize healthy trees to species that are limited to dead ones. Most belong to the latter group, but a few aggressive species cause major economic impacts due to their habit of killing healthy trees capable of mounting some resistance (Rudinsky, 1962; Beaver, 1989; Raffa, 1991). Of the more than 5800 described bark beetle species in the world (Wood & Bright, 1992), probably less than a dozen species, mostly in the genera Dendroctonus and Ips, are aggressive. When aggressive beetles attack a living tree, there are only two possible outcomes: either the tree successfully defends itself and the beetles are expelled or killed, or the beetles colonize and kill the tree or, in some cases, parts of it (e.g. Berryman, 1989). Thus, successful reproduction in aggressive bark beetles depends on killing host tissues, a requirement that places special selective pressures on both host plants and insects (Raffa & Berryman, 1987). Almost all bark beetles, irrespective of their aggressiveness, require dying, nonresistant bark for successful brood development (Raffa et al., 1993).

2. Mass destruction in a small package

Tree killing bark beetles are minuscule in size compared with a mature conifer tree, but are the most destructive agents in temperate conifer forests (Fig. 10). During outbreaks, certain bark beetles can kill virtually all host trees over extensive areas (Berryman, 1982). For example, in the north-western USA, the mountain pine beetle Dendroctonus ponderosae Hopkins killed 80 million pine trees from 1979 to 1983 (McGregor, 1985); during the same period, the southern pine beetle D. frontalis Zimmermann killed pines equivalent to 17.4 million m³ in the USA's southern states (Hoffard, 1985). In Europe, the spruce bark beetle Ips typographus L. has killed around 50 million m³ of Norway spruce in large-scale outbreaks since the 1940s (Worrell, 1983; Christiansen & Bakke, 1988; Führer, 1996). Aggressive bark beetles have clearly evolved strategies to overcome the potent defenses of even healthy trees. How can these tiny attackers kill a well-defended giant? Several factors explain this, and many of them are probably complementary rather than mutually exclusive. Three factors are usually considered most important: efficient population aggregating pheromones, tolerance towards host defensive chemicals and association with phytopathogenic fungi (e.g. Berryman, 1986; Christiansen et al., 1987; Berryman et al., 1989). Pheromone-coordinated mass attack is an essential element in the tree killing strategy, and hundreds or thousands of beetles are required to overwhelm host defenses (Rudinsky, 1962; Berryman, 1982; Raffa & Berryman, 1983a,b; Christiansen, 1985a,b).

Figure 10

Specialized insects, bark beetles, can breach the defenses of conifer bark under certain conditions, leading to dramatic consequences. (a) Aerial view showing extensive tree mortality due to an outbreak of the mountain pine beetle, Dendroctonus ponderosae. Dead trees are red. (Photo by Mark McGregor, USDA Forest Service.) (b) Ips typographus. Bar, 1 mm (c) Old lesions (arrows) on the surface of a spruce tree caused by bark beetle attack several years earlier. (d) When the outer bark is removed a few weeks after an Ips typographus attack, beetle galleries can be seen (arrows). Dark material is an indication of the blue-stain fungus, Ceratocystis polonica, which is vectored by the beetle. Resin has accumulated in and around the aborted galleries. (e) Stem section from a fungus-inoculated tree showing blue stain (arrows) extending into the sapwood. (f) Disc from a tree infected with the blue-stain fungus. The mycelium-filled, dark, occluded patches of sapwood no longer conduct water, in contrast to the orange, semitransparent water conducting areas. The darker area in the center of the disc is heartwood.

VII. The arms race: coevolution of conifer defense and bark beetle strategies

Conifers and bark beetles have coexisted since the early Mesosoic, and a special relationship has developed between bark beetles and members of the Pinaceae who produce an abundance of constitutive resin (Seybold et al., 2000). The coevolution of these species suggests an arms race that includes the usurpation of part of the constitutive defenses of the plant. It appears paradoxical that only the Pinaceae have serious aggressive bark beetle pests, being the conifer family with the most well-developed constitutive resin system. Conventional wisdom has been that resin is important in defending against bark beetle colonization by repelling, flushing out or entrapping and killing the insects. In comparison, conifers that produce no constitutive bark or wood resins, such as members of the Taxodiaceae, have no aggressive bark beetle pests. Two hypotheses can be drawn from these observations: completely different combinations of defense strategies can be effective against the same pest, and, through coevolution, a pest may be able to turn a defense into an opportunity. The latter possibility is discussed here.

1. When a defense becomes a weakness

The interesting relationship among aggregation pheromones, conifer resin defenses and bark beetle mass attacks probably reflects the coevolution of these insects and their host trees. Fast-acting and efficient population aggregation pheromones are important to achieve the rapid and massive attacks that are required to deplete the defenses of healthy trees (Borden, 1974). Most bark beetle pheromones are secondary alcohols or bicyclic ketals that are synthesized from host monoterpenes (Vanderwel & Oehlschlager, 1987) or by de novo biosynthesis (Lanne et al., 1989; Ivarsson et al., 1993; Seybold et al., 1995; Seybold & Tittiger, 2003). In an evolutionary sense, pheromones probably originated as detoxification products of host monoterpenes, which were later incorporated into an intraspecific communication system (Borden, 1982; Vanderwel & Oehlschlager, 1987). Thus, aggressive bark beetles have developed mechanisms to use one of the primary chemical defenses of conifer bark to their advantage. The widespread occurrence of aggregation pheromones among bark beetles suggests that this is a preadaptation for evolution of tree killing by mass attack. However, pheromones are not always essential for aggregation, because bark beetles without aggregation pheromones also can achieve high attack densities provided that their population is high and that primary (i.e. host-induced) attraction is strong (e.g. Tomicus piniperda L., Hylurgopinus rufipes (Eichhoff); Kirkendall et al., 1997). Host-induced attraction is often based on volatile terpenoids, again showing how a primary defense has been turned into a weakness.

2. Tolerance to host defense chemicals

Tolerance to monoterpenes, the most conspicuous and well-studied constituents of conifer resin, has been suggested to be characteristic for aggressive bark beetle species (Berryman, 1986; Berryman et al., 1989). However, the evidence for higher tolerance in aggressive species is weak or conflicting. Monoterpenes or oleoresin at concentrations found in healthy trees are toxic to adults, larvae or eggs of several aggressive and nonaggressive bark beetles, with considerable variation in effect among individual monoterpenes (Smith, 1961a,b, 1963, 1965; Berryman & Ashraf, 1970). Interspecific comparisons are made difficult by the different experimental methods employed, but the few studies that compared several species found aggressive species to be equally or less tolerant to monoterpenes than nonaggressive species (Cook & Hain, 1988, P. Krokene, Norwegian Forest Research Institute, unpublished data). The most tolerant bark beetle species are probably those that breed solitarily in living trees and those that colonize tree species with well-developed constitutive resin defenses (Everaerts et al., 1988; Raffa, 1991).

The efficient aggregation pheromones of aggressive species may reduce the need for resin tolerance, as exposure to host terpenoids is reduced when a large number of attackers quickly deplete resin defenses and kill the tree (Schwerdtfeger, 1955; Raffa & Berryman, 1987; Everaerts et al., 1988). Some degree of resin tolerance is probably common in bark beetles, because even dying and dead trees contain residual resin. A low level of tolerance to host terpenoid resins may thus be widespread among bark beetles, and may be regarded as another preadaptation for the evolution of aggressiveness in these insects.

VIII. Bark-beetle-vectored blue-stain fungi: multiple attack against multiple defenses

Coordinated mass attacks by bark beetles rapidly deplete the bark defenses of even healthy trees, and associated pathogenic fungi are suspected to assist in overcoming defenses and providing a suitable environment for beetle offspring. Association with fungi is common in bark beetles, and beetle–fungus relationships may be mutualistic, with clear benefits to both partners (e.g. ambrosia beetles and their fungi), antagonistic (e.g. competitors for resources or beetle pathogens), commensal or casual, or they may have unknown effects (Francke-Grosmann, 1967; Whitney, 1982; Beaver, 1989).

1. Blue-stain fungi: deadly weapon or opportunist?

Because all aggressive bark beetles are more or less closely associated with relatively virulent fungi, an attacked tree will normally face a beetle–fungus complex that can act synergistically to overwhelm its defenses. The fungi that are suspected to be important in tree killing belong to a group of unrelated genera of ascomycetes known as blue-stain fungi (e.g. Ceratocystis, Ophiostoma, Leptographium; Whitney, 1982, Harrington, 1993; Wingfield et al., 1993). Most blue-stain fungi are associated with bark beetles, and most conifer bark beetles vector some species of these fungi (Harrington, 1988; Perry, 1991).

For more than a century, blue-stain fungi have been suggested to play an important role in tree killing by bark beetles (Hartig, 1878; Craighead, 1928), but the idea has also been challenged (Parmeter et al., 1992; Harrington, 1993; Hobson et al., 1994). Trees that are colonized by bark beetles may wilt and die in a matter of weeks, whereas trees that have been girdled can live, and even continue to grow, for 1–2 yr until the roots starve and die from lack of photosynthate (Vité, 1961; Raffa & Berryman, 1982). Thus, although girdling by bark beetles depletes constitutive resin reserves and destroys cambium and phloem translocation (Vité, 1961; Molnar, 1965; Solheim, 1988; Parmeter et al., 1989), this does probably not affect the water relations of the tree quickly enough to be the primary cause of tree death. Destruction of the phloem by the beetles has been suggested to lead to sapwood occlusion (Harrington, 1993; Hobson et al., 1994), conceivably by inducing death of ray parenchyma cells and subsequent cavitation of adjacent tracheids. However, because virulent blue-stain fungi are able to grow into healthy sapwood, they appear to be likely candidates as contributors to tree death (Fig. 10). The fact that all aggressive bark beetles are associated with relatively virulent blue-stain fungi, and that these fungi usually are relatively vector-specific, suggests that blue-stain fungi may be important in tree killing.

The use of artificial inoculation provides some of the most convincing experimental evidence for a role of virulent blue-stain fungi in overcoming bark defenses. Experiments using point inoculation of fungal mycelium into host trees have shown that some species can kill trees in the absence of their bark beetle vectors (Horntvedt et al., 1983; Krokene, 1996). An examination of tree responses to a few vs multiple inoculations gives some insights into the resistance mechanisms. Low-density inoculations do not kill mature trees, and the fungus is confined within discrete lesions produced by the induced defenses of the tree (Fig. 10; Reid et al., 1967; Berryman, 1972; Christiansen et al., 1987). The mass inoculation technique, on the other hand, simulates the depletion of bark defenses that takes place during a bark beetle mass attack, and healthy trees can be killed using this technique. Dose–response experiments comparing fungal inoculations and induced bark beetle attacks further implicate pathogenic fungi in tree killing by showing that a similar number of inoculations and attacks is required to kill a tree. For example, the number of Ceratocystis polonica inoculations needed to kill a Norway spruce tree roughly corresponds to the number of Ips typographus attacks required to kill a similar spruce tree (Christiansen, 1985a,b). In other words, the ‘threshold of successful attack’ (sensuThalenhorst, 1958) is roughly the same. In conclusion, it appears that the bark beetle–fungus complex enhances the overall strength of the assault and acts synergistically to overcome the potent bark defenses.

It is important to keep in mind that the relationship between aggressive bark beetles and pathogenic blue-stain fungi has both positive and negative effects for each partner (Jones, 1991). Fungi and beetles may share mutual benefits in overcoming host defenses when they colonize healthy trees, but they also compete for the same nutrients in the phloem. The benefits to blue-stain fungi from an association with bark beetles seem quite obvious – dispersal and introduction under the bark of suitable host trees (Dowding, 1969; Harrington, 1993; Malloch & Blackwell, 1993). The net benefit for aggressive bark beetles from an association with fungus depends on the need to overwhelm healthy trees, and may be restricted to epidemic periods with high beetle activity (Owen et al., 1987; Raffa, 1995). When aggressive bark beetles colonize dead and dying trees during nonepidemic periods, there are probably no positive effects of being associated with virulent fungi, and only the negative, competitive aspects of the interaction are visible (Raffa, 1995). The most important advantage aggressive bark beetles gain from pathogenic fungi may be that the fungi help them exhaust tree defenses and prevent the host from expressing its potential resistance. For example, multiple wounding or inoculations with Ophiostoma clavigerum or C. polonica reduce the quantity of resin that accumulates at each wound site (Schwerdtfeger, 1955; Raffa & Berryman, 1983; Christiansen, 1985a). Fungi may also prevent the host from recovering after the initial attack and killing the developing brood with secondary resinosis (Nelson, 1934; Barras, 1970; Fares et al., 1980). As pointed out by Paine et al. (1997), an attacked tree may have no external symptoms for weeks, but still be irreversibly stressed by a bark beetle–fungus attack in the phloem. Thus, tree death may occur after the critical interactions between the host tree and the beetle–fungus complex have been completed.

IX. Conclusions

During their long lives, conifer trees meet challenges from a wide variety of organisms, among the most serious being bark beetles and their associated phytopathogenic fungi. Conifer defenses against stem infesting insects and pathogens can be classified as constitutive or induced. The constitutive defense system includes resin accumulating cells and channels in the phloem and wood, cells in the phloem that store toxic substances (e.g. phenolics), and mechanical properties of the bark such as suberized and lignified cell layers, stone cells and calcium oxalate crystals. The induced defense system involves de novo synthesis or activation of a wide range of defense chemicals, including terpenoids, phenolics, PR proteins and enzymes. The induced defense system may act against a current infection (the hypersensitive response and local resistance) or against future infections or bark beetle attacks (acquired resistance). These multiple overlapping defense structures and systems provide a formidable defense against a wide range of possible attacking organisms. However, conifers are still susceptible to certain organisms that have evolved strategies to overcome the defenses or avoid them. A number of challenges remain in improving our understanding of these defenses and the interaction of invading organisms with conifer defense strategies, as well as our understanding of the coevolution of conifer defense mechanisms and bark infesting organisms. Defensive chemicals such as the terpenoid resins and phenolics involve complex pathways that can lead to a large range of different chemical components, and characterization of the pathways operating in diverse species (and their regulation) will help determine the importance of various compounds singly and in combinations. Another key to understanding these defenses is to identify the signaling mechanisms involved in establishing patterns of constitutive defenses and in each phase of the induced defense responses. Another perspective that needs to be integrated with our knowledge of defense systems is the nature of the tree–beetle interaction. The mechanisms by which bark beetles are attracted to host trees and decide on the suitability of a tree are not well established, but evidence of the interaction of otherwise defensive compounds from the bark with beetle attraction or deterrence is particularly intriguing and merits further intensive study.

The remarkable longevity of various conifer species is a testament to the success of their defense strategies, some of which have been reviewed here, and the persistence of minute beetles capable of overcoming these defenses indicates a fascinating coevolutionary history. Understanding bark defenses and these interactions is critical in predicting the effect of ecological disturbances on tree–pest dynamics as well as in developing management and genetic strategies for improving the health of conifer forests or conifer production.

Acknowledgements

We gratefully acknowledge the support of The Research Council of Norway, the Norwegian Forest Research Institute, and the WSU College of Sciences, and the use of the WSU Electron Microscopy Center and the NLH Electron Microscopy Facility. Discussions within the Conifer Defense group (CONDEF, an international interdisciplinary working group; http://www.skogforsk.no/condef/) are also appreciated. We dedicate this review to our colleague, Dr Alan A. Berryman, Washington State University, who catalyzed the formation of CONDEF, and whose creative and lively discussions inspired our enthusiasm for exploring conifer–bark beetle interactions.