INTRODUCTION

Glossy buckthorn (Frangula alnus Mill.) and common buckthorn (Rhamnus cathartica L.) are recognized as undesirable, invasive, woody plant species in North America (Gucker 2008; Zouhar 2011). Multiple factors lead to the dominance of invasive buckthorns in invaded ecosystems, including high fecundity and prolific growth rates, which enable them to form dense thickets. This, in combination with a generally longer growing season than some native plants, allows invasive buckthorns to displace native plant communities, including overstory tree regeneration (Boudreau and Wilson 1992; Knight et al. 2007). In addition to having negative effects on neighboring native plant species, buckthorn leaves and berries create nitrogen-rich soils where leaf litter falls, creating favorable conditions for earthworm invasion (Heneghan et al. 2007; Heimpel et al. 2010). In areas where invasive buckthorns and earthworms overlap, a positive feedback loop is created in which the presence of one helps to foster favorable conditions for the other (Heneghan et al. 2007; Knight et al. 2007). This co-facilitation further perpetuates their invasion while creating unsuitable conditions for native plant regeneration, and generally adversely impacting biodiversity and multiple trophic levels in forests (Klionsky et al. 2011; Roth et al. 2015; Schuh and Larsen 2015). The net effect of these buckthorn-driven ecosystem changes is decreased biodiversity and long-term, weakened ecosystem resilience.

Given the potentially deleterious and long-lasting effects of invasive buckthorns on forest ecosystems where present, their removal is often a critical component of forest restoration activities. The application of chemical herbicide is a common management practice because buckthorn prolifically resprout after cutting. However, chemical herbicide use on buckthorn often involves repeated applications. Additionally, using herbicide as a buckthorn management strategy is often costly, and could have environmental consequences (Delanoy and Archibold 2007; Annett et al. 2014; Florencia et al. 2017). Laborious repeated herbicide application is often required for removing well-established buckthorn colonies. Alternatively, manual removal of buckthorn, which involves removal of much of the root stock, can be a time- and labor-intensive process. Therefore, finding an alternative to chemical use and manual removal for buckthorn management is desired. A myco-biological control agent could meet these needs.

Chondrostereum purpureum (Fr.) Pouzar (Polyporales, Meruliaceae), a basidiomycete that is naturally occurring in temperate zones of North America, can act as a pathogen and saprotroph in sapwood of trees (Becker et al. 2005; Reina et al. 2019). It has broad-spectrum pathogenicity toward a variety of hardwood species (approximately 190 species on record), causing silverleaf disease primarily in plants in the Rosaceae family, although it is reported as low risk for nontarget infection associated with its application (Becker et al. 1999). C. purpureum usually enters its host through a fresh wound, cut stump, or stem lesion (Brooks and Moore 1925; Rayner and Boddy 1986; Spiers and Hopcroft 1988). The fungus grows through the xylem tissue of the tree, causing necrosis of the cambium, decay, and sometimes results in host death (Rayner 1977; Wall 1986; Wall 1991). Once entering the xylem, C. purpureum kills xylem parenchyma and uses the carbohydrates and nitrogenous compounds (Sinclair and Lyon 2005). The use of C. purpureum as a mycological, biological control agent has been tested on a variety of species, yet much is still unknown about its efficacy for invasive species management, including what ranges and climates it is most effective in, species effectiveness, and other practical applications of interest to potential managers.

Herein, we extend findings from previous C. purpureum studies to include both R. cathartica and F. alnus in northern Michigan mixed coniferous and deciduous forests. Our objective was to investigate the efficacy of C. purpureum as a biological control agent for treatment of both species of invasive buckthorns by comparing application treatments of C. purpureum and glyphosate to cut stump and girdled buckthorn stems. Based on earlier studies with other woody plants, we hypothesized that C. purpureum would colonize the stems of R. cathartica and F. alnus and inhibit regrowth through decay. Our specific hypotheses were that (1) control trees would be the most prolific stump sprouters, (2) C. purpureum-inoculated trees would have less than or equal to the number of basal resprouts compared to glyphosate-treated trees, and (3) girdled buckthorn trees inoculated with C. purpureum would have lower number of resprouts compared with cut stump buckthorn trees inoculated with C. purpureum after one growing season, based off findings from Au and Tuchscherer (2014).

METHODS

Inoculation Preparation

An isolate of C. purpureum was collected on 6 November 1994 from Populus tremuloides in Houghton, Michigan, and has been maintained as a live culture on 2% malt media (2M) and stored at 4°C in between annual transfers as part of the culture collection in the Forest Microbiology Lab, Wood Protection Group, Michigan Technological University. Fungal mycelia from this sample were cultured on 2M. Inoculum was prepared by plating out hyphae on 2M for 10 d of growth at room temperature. For field application, C. purpureum and 2M inoculum was slurried (hand-stirred) together, and diluted 1:10 with DI water. The slurry was then transferred to sterilized plastic squeeze bottles and was stored at 4°C until needed for field trials. To verify that C. purpureum was active in the slurry, following Koch's postulates, an inoculation test was conducted using an Olympus BH-2 6V 20W microscope to verify anatomical features of C. purpureum mycelium grown from each test batch before use in the field. Distinguishing anatomical features of C. purpureum include evidence of clamp connections, and hyaline hyphae of 2–3 micrometers across.

The six treatment combinations in this study were girdled trees with either C. purpureum or glyphosate applications, and cut trees with either C. purpureum or glyphosate applications. Five girdled and five cut stems were also included to act as controls at each site.

Treatment with C. purpureum began in early summer 2020 (23–25 and 29 June 2020) when temperature ranges reached 15–25 °C, while treatment with herbicide AquaNeat (glyphosate) concentrated at 50% (1:1 herbicide and water) began in late summer 2020 (4 August 2020), based on findings from Au and Tuchscherer (2014). All buckthorn trees were selected from existing populations, and treatment trees were selected in the following manner: trees spaced greater than 1 m apart, and trees that received the same treatment were grouped near each other, while also taking into consideration >1 m distance between them, to reduce treatment interactions. Trees were girdled by removing a 5 cm wide band of bark to expose the sapwood, at 20 cm above ground level. Cut-stump trees were cut with a handsaw 10 cm above ground level. C. purpureum inoculations and glyphosate treatments were applied immediately after wounding by coating either the cut surface of the stump or the exposed sapwood after girdling. In October 2020 and August 2021, buckthorn trees were evaluated for regrowth, and the number of stump sprouts (basal resprouts extending from the root collar and trunk) was recorded. Following Koch's postulates (Agrios 2005) to verify presence of C. purpureum in inoculated trees, a small hatchet was used to remove a chip of vascular xylem tissue from the base of 39 C. purpureumtreated buckthorn trees in August 2021. Tissue samples were plated on 2M to incubate for 10 d, and mycelial growth was examined under an Olympus BH-2 6V 20W microscope for the presence of C. purpureum anatomical features such as clamp connections and hyaline hyphae spanning 2–3 micrometers across (Sinclair and Lyon 2005). C. purpureum was successfully reisolated from 26 of the 39 tissue samples, verifying successful inoculation in the majority of trees.

Figure 1.

Study sites located at Michigan Tech Trails and Recreational Forest, Houghton Greenspace, and Swedetown Trails in Houghton County, Michigan, USA; 2020 through 2021 field trials of fungal biocontrol and chemical herbicide treatment on invasive buckthorn trees. Triangle icons are sites infested with R. cathartica (n = 5) and diamond icons are F. alnus infested sites (n = 5).

Statistical Analysis

A generalized linear mixed model (GLMM) was employed using sprout count data from August 2021, with treatment as a fixed effect, and research site and buckthorn species as random effects. Basal diameter, tree height, percent canopy, whether a buckthorn tree was multi-stemmed (multistem), and treatment × percent canopy were examined as covariates. Pairwise comparisons were made using the Tukey method with an alpha = 0.05. Standardized residual plots were examined to confirm that model assumptions were met. A binary logistic regression was used to examine the likelihood of apparent mortality in 2021 (no sprouting) between treatments as a function of percent canopy. Pairwise odds ratios are presented for each treatment combination. Odds ratios greater than one indicate an increased likelihood and less than one a reduced likelihood of mortality. Ninety-five percent confidence intervals are provided, and intervals inclusive of values both less than and greater than one indicates a lack of evidence of a treatment effect. Regrowth comparisons were made using sprout count data recorded in October 2020 and August 2021 to assess how the amount of stump sprouting varies over time for a given treatment. Statistical analyses were performed using Minitab 20 software (State College, Pennsylvania, USA).

Table 1.

Full and reduced generalized linear models. Full model examines variables of basal diameter, multistem, tree height, percent canopy, buckthorn species, research site, treatment, and treatment × percent canopy. Variables that were statistically significant (P < 0.05) are in bold font.

RESULTS

A total of 283 buckthorn trees were resurveyed from the 300 original trees included in this study. Eight R. cathartica and nine F. alnus trees were lost before final evaluations could be made, due to them either being mistakenly cut during other recreational or invasive species management activities in the area, or their flagging being removed. Mean stem diameter at 10 cm was 3.52 cm (min: 1.8 cm, max: 10.90 cm, n = 283). Mean tree height before treatments was 4.32 m (min: 1.44 m, max: 6.1 m, n = 283).

Treatment Comparisons

We compared six management treatments on two different buckthorn species across five sites per species, approximately one-year post-treatment. Treatment type was a highly significant fixed effect (Table 1, P < 0.001). Buckthorn species was not significant (Table 1, P = 0.424), thus, both species of buckthorn were combined for all remaining analyses. Wounding a tree by either cutting or girdling without additional application of either glyphosate or C. purpureum resulted in the most stump sprouts of all the treatments (Figure 2). C. purpureum application significantly reduced buckthorn stump sprouts compared with control treatments after just one growing season (Figure 2, P < 0.001). The C. purpureum treatment applied to girdled buckthorn trees and both glyphosate treatments were not significantly different from each other (Figure 2, P ≥ 0.823). Glyphosate applied to girdled stems showed similar success in reducing stump sprouts as C. purpureum applied to cut trees and to girdled trees (Figure 2, P ≥ 0.238). There was not a significant difference in stump sprouting between C. purpureum treatments applied to girdled and cut surfaces (Figure 2, P = 0.426).

One year after treatment applications, apparent tree mortality (i.e., zero stump sprouts) as proportion of total number of trees per treatment was highest for the glyphosate cut (94%) and girdle (81%) treatments and lowest for the controls (cut = 11% and girdle = 8%). Apparent mortality for C. purpureum treatments was 51% for cut surface and 67% for girdle applications. Binary logistic regression revealed a significant treatment effect (P < 0.001) and that the likelihood of apparent mortality increased with increasing canopy cover (P = 0.043). Treatment with C. purpureum increased the likelihood of apparent mortality for girdle and cut surface applications versus their respective controls (Table 2). The increase in the likelihood of apparent mortality, however, was greatest for herbicide treatments. Within C. purpureum and herbicide treatments, the likelihood of apparent mortality did not vary significantly between cut and girdle treatments (Table 2).

Figure 2.

August 2021 (1 year) mean number of stump sprouts present after one growing season for a given treatment (control cut, control girdle, C. purpureum [C.purp] cut, C. purpureum [C.purp] girdle, glyphosate [gly] cut, and glyphosate [gly] girdle). Standard error bars shown. Factors that are not significantly different at P < 0.05 from each other share the same lowercase letter(s) denoted above.

Table 2.

Logistic regression pairwise odds ratios assessing the influence of treatment combination on the likelihood of apparent mortality (no live sprouts in 2021). Odds ratios greater than one indicate an increased likelihood of apparent mortality and less than one a reduced likelihood. Ninety-five percent confidence intervals (95% CI) inclusive of values both less than and greater than one indicate that there was no evidence of a treatment effect.

Table 3.

GLM exploring treatment × percent canopy interaction to examine significance of percent canopy on treatment success for various treatment combinations on invasive buckthorns.

In cambium samples collected to verify Koch's postulates, verified samples showed fewer mean stump sprouts compared with samples where Koch's postulates were unverified for both C. purpureum cut stump and girdle treatment combination. The mean number of stump sprouts was six in samples where Koch's postulates were unverified. In samples where Koch's postulates were verified, the mean number of stump sprouts was two.

Percent canopy was a significant covariant (coeff: –0.0585, P = 0.013). Additionally, the interaction between treatment and percent canopy was significant (P < 0.001). The treatment combinations significantly influenced by percent canopy were C. purpureum cut, control cut, and weakly significant herbicide girdle and control girdle (Table 3). As percent of canopy increased (denser canopies, lower light availability), treatment efficacy increased in C. purpureum cut and control cut buckthorn trees (Figures 3A and 3E). Contrastingly, in herbicide girdle and control girdle buckthorn trees, treatment efficacy decreased as percent canopy increased (Figures 3B and 3D). The interactions between percent canopy and C. purpureum girdle and herbicide cut treatments were not significant (Table 3, Figures 3C and 3F).

Figure 3.

A–F: Buckthorn stump sprout count versus forest percent canopy for treatment combinations, trend lines shown for treatment combinations in which percent canopy had a significant effect on treatment success.

Stump sprouts were also assessed in October 2020 and compared to sprout counts from August 2021. The number of stump sprouts decreased by 46% for C. purpureum cut stump and 59% for C. purpureum girdling treatments between October 2020 and August 2021, showing a significant reduction in stump sprout counts for C. purpureum application to both cut and girdled trees between October 2020 and August 2021 (P = 0.037 and 0.001, respectively; Table 4). Sprout suppression did not increase over time for glyphosate application to cut or girdled trees, with the number of stump sprouts increasing by 100% in cut and 550% in girdling treatments between October 2020 and August 2021 (Table 4). Glyphosate application to girdled buckthorn trees showed a significantly higher amount of stump sprouting over the time period (Table 4, P = 0.0025), indicating less effectiveness over time.

DISCUSSION

In a study conducted by Wall (1990), fewer stump sprouts were reported in nine hardwood species treated with C. purpureum than in their uninoculated counterparts (control trees). Our study is the first known study that has investigated the efficacy of C. purpureum on both R. cathartica and F. alnus as a myco-biological control agent. We hypothesized that untreated control trees would be the most prolific stump sprouters, which was the case here. Overall, all C. purpureum and herbicide treatment combinations showed fewer stump sprouts than control trees and higher probabilities of apparent mortality. C. purpureum treatment when applied to girdled trees performed as well as both herbicide treatments in terms of number of sprouts, but less well in terms of apparent mortality as of 2021. Additionally, the C. purpureum cut treatment combination performed as well as the herbicide girdle treatment in terms of number of sprouts, but the herbicide girdle treatment was four times more likely to result in apparent mortality as of 2021. In buckthorn trees treated with C. purpureum, Au and Tuchscherer (2014) saw fewer stump sprouts in girdled R. cathartica than in cut stump R. cathartica trees. We did not find a difference between girdling and cut stem application of either C. purpureum.

Table 4.

Comparisons of buckthorn stump sprout data from October 2020 and August 2021 for a given treatment combination. Treatment combinations include cut or girdled buckthorn with Chondrostereum purpureum or glyphosate applications. Percent change was calculated by (Aug 2021 count – Oct 2020 count)/|Oct 2020 count|. Statistical significance between treatments (P < 0.05) is shown in bold font.

In our study, buckthorn trees that were inoculated with C. purpureum did not show a significant difference in treatment success whether they were cut or girdled. Consequently, our hypothesis, that girdled buckthorn trees inoculated with C. purpureum will have fewer resprouts compared with cut stump buckthorn trees inoculated with C. purpureum, is not supported. Overall, our results showed that both C. purpureum treatment combinations were not significantly different from girdle herbicide treatment in their efficacy in reducing stump sprouting in invasive buckthorns after just one season of monitoring, which would give land managers a degree of flexibility when using C. purpureum for buckthorn management.

Fewer average number of stump sprouts were found in samples where the fungus was verified in the stumps following Koch's postulates, suggesting that C. purpureum did not successfully establish in all treated trees. These trees stump sprouted in similar numbers to that of control treatments, further supporting the need to ensure successful inoculation conditions are met. We purposefully did not inoculate trees when rain was forecast within the following two days, however we noted a large presence of slugs and invertebrates on the slurry paste in the field on some trees soon after C. purpureum was applied: it is possible they consumed too much of the media and the fungus dried out before successfully penetrating the wood.

It has been shown that for many hardwood species a decrease in percent canopy results in an increase in stump sprouting (e.g., Gardner and Helmig 1997; Dey et al. 2007; Atwood et al. 2009). In smaller-diameter hardwoods (10–20 cm dbh), stump sprout growth can be prolific in high light environments, especially in clearcut areas (Dey et al. 2007). Increasing amounts of residual overstory trees suppress growth, displaying a negative relationship between overstory shade and hardwood sprout growth in mechanically wounded hardwoods (Dey et al. 2007). Gardner and Helmig (1997) examined the effects of thinning on stump sprouts of water oak (Quercus nigra L.), and showed that there was greater stump sprout survival after heavy thinning (60% reduction in basal area) compared to light thinning (40% reduction in basal area). Partial harvesting systems have also been shown to result in lower rates of sprouting than clearcutting in nine hardwood species (Atwood et al. 2009). Almost complete mortality of invasive buckthorns has been reported in canopies that permit ≤3% transmission of incoming light, with surviving buckthorn strongly tied to light availability (Schuster et al. 2020). This indicates a tight linkage between buckthorn performance and light availability, where decreased light availability reduces survival in invasive buckthorns (Schuster et al. 2020).

To our knowledge, canopy cover has not been examined in previous studies that examined the impacts of C. purpureum as a control on invasive buckthorns. Percent canopy was a significant variable in our statistical analyses; however, the influence of percent canopy varies by treatment combination. Treatment success in C. purpureum cut buckthorn was significantly related to percent canopy (Table 3). As percent canopy decreased (increasing available light on the forest floor), the efficacy of C. purpureum in cut stump trees also decreased. A higher percent canopy may be more beneficial for the C. purpureum cut stump treatment, as it provides a more humid, cooler microenvironment for fungal establishment and growth. It may be that the exposure of a cut stump to more direct sunlight and higher temperatures creates unfavorable conditions for C. purpureum growth in areas where daytime temperatures surpass C. purpureum limits (Au and Tuchscherer 2014). Temperatures exceeding 30°C may be too hot for C. purpureum establishment, as the mean maximum temperature for successful field application of C. purpureum is 27°C and the optimum temperature range for spore germination and fungal growth is 24–26°C (Sinclair and Lyon 2005). Additionally, greater light availability to the root crown and axillary buds of the cut stump may stimulate sprouting. This is the likely explanation for the negative relationship between percent canopy and stump sprouting in the control cut surface treatment. This interpretation is supported by the lack of relationship between percent canopy and sprouting of girdled trees inoculated with C. purpureum because their immediate canopies remain intact.

In contrast to the negative relationship between percent canopy and stump sprouting described above, there was a positive relationship between percent canopy and stump sprouting for the glyphosate with girdling treatments. In trees with higher percent canopy above, presumably from the trees being treated, net photosynthesis and carbon capture may be higher. A more productive buckthorn tree may overcome wounding better because it has more carbohydrates sequestered; therefore, girdling a more productive buckthorn may have a higher capacity of producing more stump sprouts compared with that of a less productive tree. However, more research would be beneficial to determine if this logic can be fully supported.

The mortality of other hardwood trees inoculated with C. purpureum has been reported as increasing from 92% to 100% a year following inoculation (Becker et al. 2005). Similarly, treatment efficacy of C. purpureum improved from October 2020 to August 2021 in our study. It is likely that this decrease in the number of buckthorn stump sprouts coincides with C. purpureum continuing to colonize and necrotize the buckthorn sapwood, further exploiting its resources and hindering vascular transport abilities as more time passed (Rayner 1977; Wall 1986, 1991). This suggests that future sprout count measurements will continue to decrease until host death occurs. Following these treatments for another 2–3 y will be important to determining lasting effects of either treatment combination with C. purpureum.

C. purpureum as a biocontrol mechanism for this purpose was not commercially available to the public in the United States or Canada at the time of this writing. We cultured our own isolate and prepared C. purpureum inoculum to carry out field trials. Without commercialization of this fungus, its application as a myco-biological agent is limited, as it would be inaccessible to most land managers. A patent for the use of C. purpureum as a biological control agent for weeds and invasive species is in effect (United States Patent 5587158), which further limits the accessibility and development of potential new commercial sources of C. purpureum at this time. While multiple companies in North America and Europe have offered it for sale in the past, it appears that none is currently on the market or being regularly produced for this purpose. Trends in pest management suggest a public preference for fewer chemical control options when possible (Fravel 2005), thus there is a likely potential market for this product.

Wounding a tree by either cutting it to a stump or girdling the stem, and subsequently applying fungal inoculum, was a time-consuming process, one that might not save land managers much time in comparison to using a chemical herbicide. However, the use of C. purpureum can alleviate concerns of chemical herbicide off-target species and water quality impacts, which may be especially important at sensitive sites. It may also be that C. purpureum can act as a longer-lasting control mechanism than the chemical herbicides, as it continues to live in the wood and grow through any belowground root tissue over multiple seasons. Testing different methods for wounding buckthorn trees and applying fungal inoculum, such as brush cutting and broadcast spraying on smaller stems, could be beneficial in potentially developing more time-efficient protocols for using C. purpureum. Future testing could expand to examine the efficacy of C. purpureum on other woody invasive species in similar temperate zones, such as invasive honeysuckles (Lonicera spp.), autumn olive (Elaeagnus umbellata), or Japanese barberry (Berberis thunbergii), which often co-invade with buckthorn.

CONCLUSION

C. purpureum inoculation treatments via girdling showed similar reduction of stump sprouting as that of glyphosate treatments via girdling or cut stump, in both F. alnus and R. cathartica species. Additionally, C. purpureum inoculation via cut stump showed similar reduction of stump sprout as that of C. purpureum and glyphosate treatments via girdling. Reduction of stump sprouts in C. purpureum-inoculated buckthorn was shown regardless of the species of buckthorn, the height and diameter of the tree, and whether it was multi-stemmed. Our results indicate that C. purpureum could provide a chemical-free alternative for buckthorn management when applied during the appropriate air temperature range. This potential myco-biological control alternative would act as a step forward to creating a more resilient, biodiverse ecosystem in places where invasive buckthorn species are currently present. Ideally, these findings could provide landowners and land managers with additional options for managing invasive buckthorns.