Armillaria limonea

Armillaria limonea is a white rot wood decay fungus and root disease pathogen that has confirmed presence in New Zealand only, where it is presumed to be indigenous. It is closely related to A. luteobubalina in Australia and South America, and to South American A. montagnei. A. limonea occurs naturally on woody debris and as a cause of butt rot in living trees in podocarp-hardwood and southern beech (Nothofagus) forests where it contributes beneficially to carbon and nutrient recycling. It fruits prolifically in native forests, forming large clusters of “toadstool” fruitbodies during winter. Like many other Armillaria species it is recognized by characteristic white mycelial fans or ribbons produced beneath host bark, and by its bootlace-like rhizomorphs by which it spreads vegetatively from colonized buried woody material or stump root systems to infect living host plants.

Armillaria limonea (along with A. novae-zelandiae) was the cause of substantial disease losses in plantations of Pinus radiata planted on sites cleared of native forest in New Zealand during the 1970s and 1980s. Its importance has declined since planting on such sites is no longer practised, although it remains widely distributed. Unlike A. novae-zelandiae, it has not spread to or caused disease in orchards of kiwifruit (Actinidia deliciosa) or in P. radiata plantations established on sites not previously holding indigenous forest cover, and there is no evidence that A. limonea is invasive. Risk of unintended international spread appears to be negligible to nil. If intercepted, isolates of A. limonea may be identified by laboratory culture testing or more rapidly and precisely by molecular sequencing procedures. Armillaria limonea features in the United States Department of Agriculture Agricultural Research Service Fungal Databases. It is considered a risk organism in Hawaiʻi.

Armillaria limonea occurs in native forests in New Zealand from the north of North Island to the north of South Island, and there are three records from the southern South Island (New Zealand Fungarium, 2020). Armillaria limonea is reported but not confirmed from South America. It was reported in Tierra del Fuego, Argentina (Raithelhuber, 1987; 1991; 1992; 2004; Niveiro and Albertό, 2013) but needs molecular validation (Pildain et al., 2010). Coetzee et al. (2003) found an isolate from an unknown fruitbody resembling A. limonea in Chile was actually of A. luteobubalina.

The risk of A. limonea spreading outside New Zealand to other countries appears to be negligible to nil. A. limonea is largely confined to roots and root collars of infected trees and shrubs. In young plants mycelial fans may extend above soil level only if host resistance declines prior to or after death, or because of physiological stress from some additional external physical, chemical or biological cause. In older indigenous trees A. limonea may form a butt rot or basal canker immediately above ground level. However, the export of old growth native timber from New Zealand is restricted (Forestry New Zealand, 2020). A. limonea (probably) and A. novae-zelandiae are present causing non-lethal infection at and below soil level in a proportion of mature (up to ca. 25 years old) Pinus radiata trees growing on land cleared of native forest, indicating minimal risk of their being present in harvested logs cut at stump height. However, butt rot does not develop prior to harvest in these trees and such first rotation P. radiata stands are now becoming uncommon. In New Zealand much P. radiata is grown on land that was previously non-forested, but in these stands there is virtually no infection by A. limonea (in contrast to A. novae-zelandiae), yielding logs that pose no risk from this species.

Armillaria limonea requires a supporting wood- or organic-based substrate to survive, so its presence in nurseries is rare. Any infected plants would normally show readily detectable disease symptoms. It is possible that soil admixtures such as peat or other organic material might support A. limonea saprotrophically if the pathogen was somehow introduced, but the risk of movement in nursery stock would seem low.

Discussion on the risk of introducing Armillaria species to North America in logs of P. radiata and Douglas fir is included in White et al. (1992), who estimate it as low risk. A. limonea features in the United States Department of Agriculture Agricultural Research Service fungal databases (USDA-ARS, 2019), but is not included in the EPPO Global Database (EPPO, 2019). A. limonea is considered to be a risk organism in Hawaiʻi (DeNitto et al., 2015).

In New Zealand, A. limonea occurs naturally (along with A. novae-zelandiae) as a very common saprotrophic fungus causing decay in coarse woody debris and as a butt rot species in living trees in native forests, both podocarp-hardwood and southern beech (Nothofagus) (Colenso, 1890; Birch, 1937a; Gilmour, 1954; 1966; Hood and Sandberg, 1987; Hood et al., 2004b; 2019). Rhizomorphs are prolific in the soil beneath podocarp-hardwood forests, growing epiphytically across healthy roots without infecting them. Roots of older trees can be found with Armillaria-caused decay, but it is not clear whether the fungus is the cause of death or if the roots were killed or weakened from some other cause. A. limonea is not generally seen as a pathogen of native trees (Rawlings, 1952) but Armillaria-caused mortality has occurred e.g., in densely stocked Nothofagusmenziesii saplings regenerating in gaps after storm damage (Birch, 1937b) and in the same host planted as an ornamental specimen tree in urban settings (Hood, 1989). Nevertheless, Armillaria species provide an indispensable service as important members of the indigenous wood decay community. Along with other decomposer fungi, A. limonea contributes significantly to carbon and nutrient recycling in the native forest ecosystem (Hood et al., 2019).

Armillaria limonea occurs as a parasite on a wide range of host species, both angiosperms and gymnosperms. An online global list of hosts is provided by the United States Department of Agriculture Agricultural Research Service, Fungal Database (USDA-ARS, 2019). Host lists are given by Gilmour (1966), Dingley (1969), Laundon (1973), Shaw et al. (1976), Pennycook (1989), McKenzie et al. (1999; 2000), Gadgil (2005) and van der Pas et al. (2008). A. limonea and A. novae-zelandiae are not always distinguished in these lists. In New Zealand Armillaria species (A. limonea and A. novae-zelandiae) are or have been important in fruit orchards (Atkinson, 1971) and in forest plantations, particularly of Pinus radiata. Pathogenicity by A. limonea to P. radiata has been demonstrated by inoculation experiments with seedlings (Shaw et al., 1981; Benjamin, 1983; Kile, 1984; Hood and Sandberg, 1993b).

Genetics

Like many other Armillaria species, A. limonea is assumed to have a diploid vegetative mycelium, undergoing meiosis to form haploid basidiospores.

Virulence

Inherently, A. limonea appears to show an intermediate level of virulence when compared with other saprotrophic and pathogenic species of Armillaria. In New Zealand, A. limonea has not shown secondary spread between adjacent root systems of young infected Pinus radiata trees as occurs, for instance, with the more virulent Armillariaostoyae in Pinus pinaster stands in the Landes de Gascogne forest region in southwest France (Roth et al., 1979; van der Pas, 1981b; Lung-Escarmant and Guyon, 2004). Inoculation studies have demonstrated that different A. limonea isolates vary innately in their level of virulence to P. radiata (Shaw et al., 1981; Benjamin, 1983; Hood and Sandberg, 1993b).

Reproductive Biology

Fruiting occurs over several weeks between April and July (Hood et al., 2004a; Hood and Gardner, 2005) (Southern Hemisphere autumn-winter). Small and often very large clusters of fruitbodies appear in natural forests (along with those of A. novae-zelandiae) on woody debris in indigenous forests, producing copious quantities of basidiospores. Fruiting is rare on infected introduced host trees and shrubs. There is no asexual stage.

Physiology and Phenology

MacKenzie (1987) found that plantation P. radiata trees older than ca. 10 years (near mid-rotation) are not killed but interact dynamically with Armillaria while remaining green and apparently healthy. Over time, infection at the root collar decreases in some trees while increasing in others. This behavior has been confirmed for A. novae-zelandiae (Hood et al., 2002) and is also probably true of A. limonea.

Armillaria limonea demonstrates bioluminescence in both the fruitbody gills (Cooper, 2016) and mycelium, and Armillaria-decayed wood may be seen faintly glowing in native forests on dark nights. Cultures are also bioluminescent, especially when young and actively growing.

Population Size and Structure

In native forests, populations of A. limonea are composed of comparatively high colony densities per unit ground area. In a study in natural podocarp hardwood forest in New Zealand, cultural pairing of isolates demonstrated a density of colonies (vegetative compatibility groups) of A. limonea ranging from 15-56 colonies/ha (Hood and Sandberg, 1987; Hood, 1989). Similar studies in young P. radiata plantations that had replaced native forest yielded equivalent values of 3 colonies/ha (Benjamin, 1983; Shaw, 1984) and between 15 and 46 colonies/ha (Hood and Sandberg, 1993b). Although numerous, these were not as high as values for A. novae-zelandiae in the same stands.

Rhizomorphs are produced prolifically. Hood and Sandberg (1987) recorded an average aggregate rhizomorph length (comprising both A. limonea and A. novae-zelandiae) ranging from 2 to 9 m/m² of soil surface down to a depth of 22 cm beneath a podocarp-hardwood forest logged of the podocarp element three decades earlier.

Human food and beverage

• Food

To identify the species first requires laboratory isolation of fungal cultures from infected tissue, mycelial sheeting, rhizomorphs or, if present, fruitbody tissue. This is done by antiseptically transferring small pieces onto plates of an Armillaria-selective medium such as malt extract agar with suitable antibiotic &/or biocidal additives (e.g., o-phenylphenol or benomyl) to discourage contaminating bacteria and other unwanted fungi (Morrison et al., 1991; Worrall, 1991). One recipe prescribes the addition of from 2 to 10 ppm benomyl and (after autoclaving and cooling before pouring plates) 100 ppm streptomycin sulphate to a 2% malt extract agar base (although a substitute may need to be found if the fungicide Benlateä, active ingredient 50% benomyl, becomes unavailable). Cultures may be maintained in the laboratory by sub-culturing on potato dextrose or 2% malt extract agar. Armillaria isolates are readily recognized in culture from their reddish-purple crustose appearance and production of characteristic rhizomorphs (distinct in appearance, however, from those formed in the field), but additional techniques are necessary for distinguishing species.

An older procedure to identify an unknown diploid field isolate, first developed with Northern Hemisphere Armillaria species (Korhonen, 1978), was to pair it with a range of selected, white, fluffy, single spore, haploid, tester isolates of known species. The paired tester isolate that changed from white fluffy to a dark purple, crustose, diploid form identified the unknown as being of the same species (Kile and Watling, 1988; Hood and Sandberg, 1987). The method is mostly reliable but takes time and the result is not always clear cut. Isolates of A. limonea can also be distinguished from those of A. novae-zelandiae by the production of rhizomorphs only in the former after growing for ca. three weeks under a 24-hour photoperiod using specified lighting conditions (Shaw et al., 1981; Benjamin, 1983; Hood and Sandberg, 1987).

More recently, cultures of A. limonea have been distinguished from other Armillaria species using molecular techniques (Coetzee et al. 2001; 2003; Maphosa et al. 2006; Pildain et al. 2009; 2010). Molecular procedures have been developed for identifying A. limonea using the ribosomal internal transcriber spacer region (Dodd et al. 2006; 2010).

As with other Armillaria species, first evidence of disease in a crop will be symptoms of wilting or crown discoloration. Removal of soil to expose the root collar using a small trowel or grubber may reveal evidence of oozing resin or gum in conifers or hardwood species, respectively. This presence of resinosis or gummosis indicates damage to the living host, implying that the pathogen is acting parasitically and not simply feeding saprotrophically on already dead tissue. Black, branching, bootlace-like rhizomorphs consisting of a thin, outer, black rind and a soft, inner, pinkish, hyphal core, may be present. Another defining feature is characteristic, thick, white mycelial sheets, fans or ribbons which are revealed by using a knife to cut and prise away the bark. If the host is already dead and no longer resisting the pathogen, these sheets may be present more extensively running up into the stem. Rarely, fruitbodies may be present (see “Description”, above). Unlikely in crop plants, but in older, mature trees, a butt heartwood rot may be present with a distinctive moist, cheesy, yellowish, fibrous decay interspersed with thin, black zone lines (the edges of sclerotial sheeting forming large, 2-3 cm “pockets” within the wood) which give rise to a scalloped appearance on drying.

Armillaria limonea is one of many Armillaria species broadly similar in appearance and behavior (Watling et al., 1991; Pegler, 2000). All have white basidiospores, an anulus or ring (partial veil) around the fruitbody stipe (toadstool stalk), and many produce rhizomorphs (species without a ring belong in the recently erected genus, Desarmillaria (Koch et al., 2017)). Species vary in their level of intrinsic pathogenicity against different hosts, but many are responsible for economic diseases in crops, plantations and managed natural forests around the world (Hood et al., 1991; Kile et al., 1991). All species also feed saprotrophically, colonizing and producing a white rot in wood, on which they fruit. Molecular analysis has shown that A. limonea belongs in a broad group of Southern Hemisphere Armillaria species that cluster separately from those in the Northern Hemisphere (Maphosa et al., 2006; Coetzee et al., 2011; Klopfenstein et al., 2017). Armillaria limonea is closest to A. luteobubalina, found in Australia and South America, and to South American A. montagnei (Coetzee et al., 2003; 2011; 2018; Maphosa et al., 2006; Pildain et al., 2009; 2010; Klopfenstein et al., 2017). Fruitbodies of these species are very similar morphologically, but tissues of A. limonea lack the persistent unpleasant taste of those of A. luteobubalina and A. montagnei (Podger et al., 1978; Coetzee et al., 2018). These species likely diverged after the breakup of Gondwana (Coetzee et al., 2011; 2018; Klopfenstein et al., 2017).

In New Zealand A limonea and A. novae-zelandiae are the two most common Armillaria species. The rhizomorphs of these species are difficult to distinguish in the field (Hood and Sandberg, 1987) but according to Benjamin (1983) those of A. novae-zelandiae are slightly thicker on average and produced in greater abundance than those of A. limonea. Armillaria limonea fruit bodies tend to be larger than those of A. novae-zelandiae, and unlike the latter the cap is not viscid (tacky) when first formed. Rhizomorphs are common, dichotomous in form (Benjamin, 1983; Morrison, 1989).

Unusually for Armillaria species, A. limonea was first recognized as a distinct taxon on morphological grounds (Stevenson, 1964). However, to distinguish it confidently from other species internationally requires cultural or molecular procedures (see “Diagnosis”).

Due to the variable regulations around (de)registration of pesticides, your national list of registered pesticides or relevant authority should be consulted to determine which products are legally allowed for use in your country when considering chemical control. Pesticides should always be used in a lawful manner, consistent with the product's label.

Cultural Control and Sanitary Measures

Field experiments have demonstrated the importance of the quality of the planting stock on sites known to be infested with A. novae-zelandiae and A. limonea. In a first rotation Pinusradiata stand on a site cleared of podocarp-hardwood forest in New Zealand mortality from the disease among newly planted seedlings was lower than among rooted cuttings (Klomp and Hong, 1985). However, with improved quality of P. radiata cuttings, the trend was reversed when a second rotation trial stand was established on the same site (Hood et al., 2006). In each case, better planting stock was at an advantage when exposed to Armillaria inoculum.

Physical/Mechanical Control

Removal of stumps and woody debris before planting is the most effective method of control of Armillaria root disease. However, this procedure has been applied only occasionally in forest plantations in New Zealand because benefits are not seen by forest managers to outweigh the initial expense, despite economic analyses to the contrary (Shaw and Calderon, 1977; MacKenzie, 1987; Self and MacKenzie, 1995). Unlike some other management options, stump extraction also benefits subsequent rotations by removing inoculum from the site, although its effects on soil quality (removal of topsoil, compaction) must also be considered. Soil cultivation without stump removal has a number of advantages e.g., improved planting efficiency and better establishment, but it is not clear if this procedure reduces disease levels.

In one study, removal of indigenous forest stumps reduced mortality among planted pine seedlings to 12-21% after 5 years, compared with 52% on the untreated site (Shaw and Calderon, 1977; van der Pas, 1981a). In another, similar trial, which included clearing debris into windrows after removing indigenous forest stumps, mortality after 4 years was 1% in treated and 23% in untreated plots (van der Pas and Hood, 1984). In a third trial, in a second rotation Pinus radiata pine plantation following Pinus ponderosa that had been planted on a cleared indigenous forest site and subsequently poison thinned, mortality was reduced after 5 years from 22% (untreated) to 5% (stumps removed) at one site and from 10% to less than 1% at another (Self and MacKenzie; 1995). After 8 years, non-lethal infection in the same trial was reduced from 67% to 31% of trees affected and from 85% to 10% of trees, in the respective sites. In all these studies no attempt was made to distinguish A. limonea from A. novae-zelandiae.

Biological Control

Isolates of Trichoderma fungi species have been tested for controlling Armillaria root disease in New Zealand. In one trial Pinusradiata seedling root systems were immersed in a selected Trichoderma slurry treatment before planting on an Armillaria-infested site. Mortality after 2 years was 6% of treated plants compared with 22% in the untreated controls (Cutler and Hill, 1994). Surviving treated plants were healthier. Certain basidiomycete fungi showing promise against A. novae-zelandiae and A. limonea in laboratory testing were not effective when inoculated into P. radiata thinning stumps in a field trial (Li and Hood, 1992; Hood et al., 2015). This work has not continued.

Chemical Control

There has been some research to test the effectiveness of chemical treatments in reducing the incidence of Armillaria root disease in forest plantations (without distinguishing the causal species) although none have been used operationally. Mortality 4 years after planting Pinus radiata on an indigenous forest cutover site was 9% in plots with stumps treated with a commercial hydrocarbon mixture containing methyl isothiocyanate, compared with 23% in untreated plots, despite there being no difference in the frequencies and quantities of soil rhizomorphs between sites (van der Pas and Hood, 1984). A similar result was obtained after applying agricultural lime to the soil surface. Sodium pentachlorphenate and pentachlorphenol introduced to the soil gave no protection to container grown P. radiata against artificial inoculum of A. limonea (Shaw et al., 1980).

Host Resistance (Incl. Vaccination)

There is evidence that host species vary in resistance or tolerance to A. limonea. In inoculation studies, Benjamin and Newhook (in Kile, 1984) found more seedlings of Pinus radiata became infected and died than did those of Eucalyptus species, among which variation also occurred. Field observations of the effects of Armillaria root disease (A. novae-zelandiae and A. limonea) on tree species planted on sites cleared of indigenous forest have been published (Weston, 1957, Gilmour, 1966). Most susceptible were P. radiata, species of Larix and Chamaecyparislawsoniana, while least affected were Cryptomeria japonica and Thuja plicata (Hocking and Mayfield, 1939; Jolliffe, 1940; Lysaght, 1942). Douglas fir (Pseudotsuga menziesii) was as susceptible as P. radiata but showed lower rates of mortality. Losses were low among eucalypts (Newhook, 1964). Although apparently less vulnerable, C. japonica was found to develop butt heart rot caused by Armillaria. A list of species recorded as susceptible to the disease in New Zealand (without distinguishing between A. novae-zelandiae and A. limonea) was published by van der Pas et al. (2008).

Monitoring and Surveillance (Incl. Remote Sensing)

Monitoring of mortality is normally undertaken on the ground by means of transects or other appropriate field design (Hood and Kimberley, 2002). In Pinusradiata plantations, non-lethal infection is estimated in the same way, except that in the absence of crown symptoms it is also necessary to expose the root collar to reveal the presence and extent of girdling infection (Shaw and Toes, 1977). Firth and Brownlie (2002) found that by using high resolution, colour, stereo aerial photography it was possible to detect most trees greater than 2 m tall that had been killed by Armillaria (without distinguishing species), reducing the time and investment necessary for ground monitoring. Mapping of infested sites in second rotation stands indicated that the distribution of non-lethally infected trees showed a broad approximation to that of trees visibly killed by the disease (e.g., areas with an average pre-thinning mortality greater than 3% of trees killed were reasonable indicators of stand areas with non-lethal infection exceeding 25%). This implied that it might be possible to use the distribution of visible mortality (e.g., from aerial photography) to produce contour maps that predict and delineate areas requiring treatment, either during or at the end of the rotation (Hood and Kimberley, 2002; Hood et al., 2006). This work has not proceeded further.

16/10/19 Original text by:

Ian Hood, New Zealand Forest Research Institute (Scion), Rotorua, New Zealand

Charles Shaw, Consultant, USA