Link to home

Plant Pathogens at Work: Progress and Possibilities for Weed Biocontrol

Part 1: Classical vs. Bioherbicidal Approach

C. B. Yandoc-Ables
E. N. Rosskopf
USDA-ARS, Fort Pierce, FL

R. Charudattan
Plant Pathology Department,
University of Florida, Gainesville, FL

(Corresponding author:
Yandoc-Ables, C.B., Rosskopf, E.N., and Charudattan, R. 2006. Plant Pathogens at Work: Progress and Possibilities for Weed Biocontrol. Part 1: Classical vs. Bioherbicidal Approach. doi:10.1094/APSFeature-2006-08.

See also Part 2

The Importance of Weeds

Weeds are a perpetual menace to agricultural productivity, causing significant reductions in the quantity and quality of crop yields. It is estimated that weeds can reduce crop yields by as much as 12%, which translates to $32 billion in losses, based on the potential value of all U.S. crops of approximately $267 billion/year (107). Growers also incur extra costs from herbicide use, employment of tillage and mowing, and cultural and biological inputs for weed management. In the 1980s, growers spent over $3 billion annually for chemical weed control and approximately $2.6 billion for cultural, ecological, and biological control methods (92). In 2001, approximately $6.4 billion was spent in the United States on herbicides, which represents 58% of the total pesticide expenditures in the country (69).

Fig. 1. Typical monoculture of Australian melaleuca (Meleleuca quinquenervia) trees in Florida prior to control efforts. Photo courtesy of Min Rayamajhi.


Weeds also pose serious ecological problems; invasive weeds in natural areas are capable of altering ecosystem processes and displacing native plant and animal species. They may also support populations of non-native animals and microbes and hybridize with native species, subsequently altering gene pools (89). Babbitt (7) estimated that non-native weed species are spreading and invading United States wildlife habitat at the rate of 700,000 ha (1,729,700 acres) annually. These non-native weeds include the European purple loosestrife (Lythrum salicaria), which is spreading at the rate of 115,000 ha/year (284,165 acres/year) and is changing the basic structure of wetlands (102), the European cheatgrass (Bromus tectorum), which now occurs on 5 million ha (12 million acres) in Idaho and Utah (111), and the Australian melaleuca (Melaleuca quinquenervia), which invades wetland ecosystems in Florida at the rate of 7.8 ha/day (19.3 acres/day) (Fig. 1) (31). In Florida alone, approximately $102.8 million were spent in 2003-2004 for the prevention and control of invasive species (49).

In addition, weeds serve as reservoirs for plant pathogens that impact crops (Fig. 2). This source can serve as the initial inoculum for epidemics in crops and can exacerbate existing problems by providing multiple hosts for reproduction. The survival of Phytophthora capsici on weed hosts is an example of a difficult-to-control pathogen that can utilize a weed host in the absence of a suitable crop, complicating attempts to utilize cultural control measures for this serious vegetable pathogen (50).


Fig. 2. American black nightshade (Solanum americanum), a serious weed problem in several solanaceous crops, is a host for multiple pathogens that impact these crops. The weed, infected with Tomato spotted wilt virus, is shown here. Photo courtesy of Scott Adkins.


Adding to the economic and ecological impacts of weeds are the environmental impacts associated with weed management practices, such as chemical herbicides and cultivation. Herbicides can cause non-target injury and contamination of ground and surface water. Herbicide resistance is now a widespread occurrence. At present, there are 307 herbicide-resistant weed biotypes worldwide, 113 of these biotypes occur in the United States (58). Cultivation predisposes the soil to wind and water erosion, and possibly increased herbicide deposition in lakes and streams due to adsorption of herbicides by mineral and soil fractions that are carried from the fields (9).

Numerous challenges for researchers, growers and land managers have resulted from: the negative effects of existing weed control practices; the loss of registration for some herbicides; the recent ban on methyl bromide; lack of herbicides labeled for weeds of minor crops, non-agricultural weeds and non-native invasive weeds; the cost of control methods in natural and other low-maintenance areas; and the need for control methods that are appropriate for use in organic production systems. Many of these challenges create niches where the use of plant pathogens as biological control agents for weeds could be a viable alternative.


Fig. 3. Virtual weed monoculture of purple nutsedge (Cyperus rotundus) in a squash production system that utilizes polyethylene mulch for fumigant retention and weed management.

Biological control has been defined as a selective process in which a pathogen or an insect is used against targeted weeds without damaging non-target plant species, such as crops and native plants (52). It clearly represents an "off-beat application" of plant pathology (8). The use of plant pathogens to suppress weeds is considered as one of the alternative weed control options for areas or production systems where the use of chemical herbicides is not permitted or feasible or where existing production practices lead to uncontrolled weed monoculture (Fig. 3). This may also be employed when the herbicide selection or usage must be rotated with other control methods in order to prevent the development of resistant weed biotypes or lessen the impact of herbicides on the environment. However, biological weed control agents are utilized only to mitigate weed infestations and not to eradicate entire weed populations. The inability of these agents to damage weeds as severely as chemical herbicides, and so serve as stand-alone replacements for chemical herbicides, has probably deterred the commercialization of many biological control agents. Biological control will be most effective when the strategy employed is tailored to the specific weed management needs of the situation.

Weed biocontrol with plant pathogens has been studied using two basic approaches: the classical and the bioherbicide approaches.

Classical Approach

The biocontrol approach using a pathogen imported from a foreign location to control a native or naturalized weed with minimal technological manipulations has been termed the classical or the inoculative biocontrol method. Introduced weeds, especially those that are clonally reproducing, genetically homogeneous, and reproductively conserved, can be easy targets for control by the introduction of coevolved pathogens from which they have been physically separated for a period of time. In classical biocontrol programs, the introduced pathogen is released or inoculated into small weed infestations relative to the total infestation (i.e., inoculated into the population as opposed to inundation of the population). If conditions are favorable, the pathogen multiplies on the weed host and spreads, causing a high level of disease (an epidemic) that may kill or severely limit the growth and reproduction of the weed. The weed population may then begin to decline. Since the disease increase will typically be gradual rather than instantaneous, several months or years may pass before a significant level of weed control is seen.

Two caveats concerning the classical approach: it is impossible to predict the success of an introduced pathogen, and it is not realistic to be prepared for a total recall should a released agent be subsequently deemed undesirable. Therefore, a careful evaluation of efficacy and safety must precede a pathogen’s introduction. Valid protocols based on conceptual frameworks as well as empirical examples exist for selection of safe and effective agents (14,43). As Barton (11) has found through her extensive analysis of projects during the past several decades, the safety record of worldwide pathogen introductions for weed control has been impeccable with not a single instance of unexpected, undesirable effects.

The overall success rate of classical weed biocontrol projects using imported pathogens has been estimated at 57% and for all pathogen-based weed control projects at 21%. These success rates are calculated from the number of projects for which success can be verified from published accounts or reliable anecdotes compared to the number of known projects (39). For a comparison, the record of success for insect-based weed biocontrol projects, at around 30-35% (see discussion in 39), is not considerably higher. Given the potential for additive or synergistic effects of different pathogens as well as the possible beneficial interactions between pathogens and insect biocontrol agents, it would be prudent to consider the integration of pathogens and insects in future weed control projects (29). Since the early 1970s, seven pathogens have been imported into the United States: Puccinia chondrillina, P. carduorum, P. jaceae (all in the mainland) and Entyloma ageratinae (also described as E. compositarum), Colletotrichum gloeosporioides f.sp. clidemiae, Septoria passiflorae, and a Septoria sp. on lantana (all in Hawaii). Worldwide, this number is 25 (11).

Generally, pathogens intended as classical biocontrol agents are subjected to rigorous safety and host-specificity testing under the oversight of the Technical Advisory Group (TAG) on Biological Control of Weeds. TAG, a voluntary committee under the aegis of the U.S. Department of Agriculture-Animal and Plant Health Inspection Service-Plant Protection and Quarantine (APHIS-PPQ), consists of representatives from several federal agencies (45,71,80). TAG’s role is to help coordinate research aimed at assuring safety in biological weed control introductions. TAG plays an oversight role in reviewing proposals to initiate classical weed biocontrol projects and in making recommendations with respect to the need, scope, and adequacy of the proposed research. Based on TAG’s recommendation, APHIS-PPQ grants permits to introduce foreign pathogens into approved quarantine facilities as well as their eventual release and field establishment. In the discovery phase, surveys are made in the native range of the weed to collect, test, and rank pathogens with biocontrol potential. Prospective pathogens, often rust fungi, are studied in their native range to determine their host specificity and virulence towards the target weeds. Those considered safe to non-target plants, including major and minor crops and ecologically important plants, are reviewed by TAG and approved by APHIS-PPQ for introduction into a quarantine facility in the United States. Once in quarantine, additional testing is done on plants of economic and ecological importance to North America. Release of the agents and assessment of their effectiveness are typically coordinated with appropriate federal (APHIS-PPQ), state (Department of Agriculture), and regional agencies (e.g., state and county agencies).

The importance of precise identification of the race, pathotype, or genotypes of pathogens and confirmation of their virulence toward the target genotype is important because the host-pathogen specificity can be governed by single-gene differences or a by a small number of genes, particularly at the subspecies level. A poor understanding of the taxonomy of the weed target, plant pathogen or the underlying host-pathogen relationship can cause temporary delays or a permanent halt to the biocontrol projects due to the agent’s unsatisfactory performance in assessment studies.

Formal attempts at classical biocontrol of weeds began in the late 1960s with a project to find and use a pathogen or pathogens for sorrels or docks (Rumex spp.) in the United States (65) and blackberries (Rubus spp.) in Chile (85). Since the 1970s, there have been several highly successful classical biocontrol programs (for reviews of this topic see: 12,25,39,110).

Among the most successful is the control of Acacia saligna by the rust fungus Uromycladium tepperianum introduced into South Africa from Australia (84). Acacia saligna is regarded as the most important invasive weed that threatens the Cape Fynbos Floristic Region of South Africa, a unique ecosystem. The fungus causes extensive gall formation on branches and twigs causing heavily infected branches to droop; the tree is eventually killed. The fungus was introduced into the Western Cape Province between 1987 and 1989. In about eight years after introduction the rust disease had become widespread in the province and tree density was decreased by 90-95%. The number of seeds in the soil seed bank had also stabilized in most sites and the process of tree decline was reported to be continuing (82). It has been determined that the benefits of this biocontrol program far outweigh the potential loss of social benefits, mainly as fire wood, to be derived from this invasive tree species (82,83).

Another widely acclaimed example of success is the use of Puccinia chondrillina to control Chondrilla juncea (rush skeleton) in Australia. The rust fungus was introduced from the Mediterranean. This biocontrol project has been estimated to have yielded a cost to benefit ratio of 1:100 to 1:200 (46). Puccinia chondrillina was also introduced into the western United States to control a skeletonweed biotype. However, unlike in Australia, the rust was only partially successful. Hence, under these conditions of less-than-expected efficacy, the rust has been utilized along with chemical herbicides, and the insect biocontrol agents, Cystiphora schmidti (a gall forming midge) and Aceria chondrillinae (a gall forming mite), in an integrated weed management program to maximize its benefits (75). As in Australia, the rust has been the most successful of the three introduced biocontrol agents in California and other western states (99).

Another successful weed biocontrol program has involved the use of a foliar smut fungus, Entyloma ageratinae from Jamaica, to control Hamakua pamakani (Ageratina riparia) in Hawaiian forests and rangelands (105) (Fig. 4). The fungus, originally misnamed as Cercosporella sp. and subsequently described as Entyloma ageratinae by Barreto and Evans (10) and E. compositarum by Trujillo et al. (105), was introduced into Hawaii in 1974. About two to three months after the pathogen was released in the field, devastating epidemics were recorded in dense stands of A. riparia in cool, high-rainfall sites in Oahu, Hawaii, and Maui. The weed populations were reduced 80% to less than 5% in a 9-month period. Similar reductions in weed populations were recorded 3 to 4 years after the pathogen was released at sites with adequate moisture. At sites with low temperatures and low rainfall there was greater than 50% reduction in the weed population in 8 years after the pathogen’s release. It is estimated that more than 50,000 ha (123,550 acres) of pasture land have been rehabilitated to their full potential due to this pathogen. No evidence of host resistance or the presence of mutant stains of the pathogen has been encountered (104,106). This pathogen has also been released against A. riparia in South Africa (81) and New Zealand (51), where it has established and is showing signs of success (57, 81).


Fig. 4. (A-D) Biological control of Hamakua ‘Pa-makani,’ Ageratina riparia, with Entyloma compositarum, the white smut fungus introduced from Jamaica in 1974. (A) Nonseptate, hyaline, slender, arcuate conidia (30-40 × 3-3.5 µm) of E. compositarum. Scale bar = 12 µm. (B) Abaxial surface of a diseased ‘Pa-makani’ leaf showing the characteristic white sporodochia, containing masses of spores of E. compositarum. (C) Infestation of A. riparia at 900 m elevation at Palani ranch, North Kona, Hawaii, before inoculations in December 1975. (D) Striking biological control of the ‘Pa-makani’ weed at Palani ranch, a site at 900 m elevation, 8 years after inoculation with the biocontrol fungus.

(E-H) Biological control of banana poka Passiflora tarminiana with Septoria passiflorae introduced from Ipiales, Colombia in 1993. (E) Filiform, multiseptate, hyaline conidia of S. passiflorae (35-52 × 1.5-2 µm). Scale bar = 24 lm. (F) Symptoms of septoria leaf spot on banana poka leaves 30 days after inoculations with ‘lei inoculum.’ Notice necrotic dry leaves, remains of the original lei that was placed below the symptomatic leaves. (G) Piha-road bordering the US Department of the Interior-Fish and Wildlife Preserve Photopoint No. 2, 1850 m elevation, before the 1997 inoculation with S. passiflorae. Notice the banana poka vine climbing to the top of the koa forest. (H) Photopoint No. 2 showing 99% banana poka biomass reduction 6 years after inoculation. Reprinted from Biological Control, Volume 33, Issue 1, Eduardo E. Trujillo, History and success of plant pathogens for biological control of introduced weeds in Hawaii, pp. 113-122, Figure 2, 2005, with permission from Elsevier. Use of (G) and (H) by permission of the American Phytopathological Society.


A more recent example is that of Puccinia carduorum, imported from Turkey and released into the northeastern United States (Virginia and Maryland) in 1987 to control musk thistle, Carduus thoermeri. The rust has spread widely from its original introduction to the western states of Wyoming and California (13,26,77). Puccinia carduorum has been found to reduce musk thistle density by accelerating senescence of rust-infected musk thistle and reducing seed production by 20 to 57% (13). The effects of this fungus on insect biocontrol agents of this weed are negligible (72).

Two other rust fungi, Maravalia cryptostegiae and Puccinia evadens, introduced into Australia from Madagascar and Florida, USA, respectively, to control Cryptostegia grandiflora and Baccharis halimifolia, are beginning to have significant impacts on the densities of their respective weed hosts (Rachel McFadyen, Queensland Department of Lands, Australia, personal communication). Another introduction into Australia that appears to be producing successful results is Sphaerulina mimosae-pigrae (anamorph: Phloeospora mimosae-pigrae) on Mimosa pigra. The most recent introduction to the United States, Puccinia jaceae var. solstitialis imported from Bulgaria and Turkey, was released in California in 2003 (Figs. 5 and 6). The host range tests on this pathogen were extensive (23,24), and the effects of this introduction are currently being monitored.


Fig. 5. Infestation of yellow starthistle (Centaurea solstitialis) in California. The weed is the target of the classical biological control agent, Puccinia jaceae var. solstitialis. The host range of the rust was studied extensively prior to its release in California in 2003. The impact and movement of the fungus is currently being monitored (Bill Bruckart, personal communication). Photo courtesy of Bill Bruckart.



Fig. 6. (A) Initial field infection of yellow starthistle leaf in Tehama Co., California, by the biocontrol agent, Puccinia jaceae var. solstitialis (Photo credit Baldo Villegas, CDFA) and (B) prolific greenhouse production of urediniospores (Photo credit Bill Bruckart). Photos courtesy of Bill Bruckart.

Bioherbicide Approach

The bioherbicide approach utilizes indigenous plant pathogens that are isolated from weeds and are cultured to produce large numbers of infective propagules (such as spores). Infective propagules are applied at rates that will cause high levels of infection, leading to suppression of the target weed before economic losses are incurred (101). Annual applications are required since the pathogen does not generally survive in sufficient numbers between growing seasons and does not produce the level of inoculum needed to initiate a new epidemic on new weed infestations (37,98,100). In the United States, the bioherbicide approach to weed control is often exemplified by the discovery, development, and commercialization of DeVine®, a bioherbicide composed of Phytophthora palmivora for the control of stranglervine (Morrenia odorata) in citrus, and Collego®, a bioherbicide composed of Colletotrichum gloeosporioides f. sp. aeschynomene for the control of northern jointvetch (Aeshynomene virginica) in rice and soybeans.

It is estimated that there are over 200 plant pathogens that have been or are under evaluation for their potential as bioherbicides; these include fungi and bacteria that cause foliar diseases, soil-borne fungal and bacterial pathogens and deleterious rhizobacteria (DRB) (16,38,93). However, out of these, there are few that have been registered, commercially produced, and available for use. A survey conducted by Barton (12) indicated that from the 1960s to 2005, there have been numerous bioherbicidal plant pathogens that have been registered, including the following: in the United States, Acremonium diospyri, Phytophthora palmivora [DeVine®], Colletotrichum gloeosporioides f. sp. aeschynomene [Collego®], Alternaria cassiae [CASST™], Puccinia canaliculata [Dr. BioSedge], P. thlaspeos [Woad Warrior], Chondrostereum purpureum [Myco-Tech™ Paste and Chontrol™ Paste] and Alternaria destruens [Smolder]; in Canada, C. gloeosporioides f. sp. malvae [BioMal®], Chondrostereum purpureum [Myco–Tech™ paste and Chontrol™/Ecoclear™]; in China, C. gloeosporiodes f. sp. cuscutae [Lubao]; in South Africa, Cylindrobasidium leave [Stumpout] and C. gloeosporioides [Hakatak®]; in the Netherlands, C. purpureum [Biochon™]; and in Japan, Xanthomonas campestris pv. poae [Camperico™]. Of these bioherbicides, only a few are currently available for purchase. The others are unavailable due to the lack of continued commercial backing, high cost of mass production, introduction of newer herbicidal chemistries, resistant weed biotypes (e.g., Dr. BioSedge), or limited markets. One bioherbicide agent, C. gloeosporioides f.sp. aeschynomene (previously Collego), has been re-registered as of March 2006 under the commercial name LockDown for use in rice in Arkansas, Lousianna and Mississippi (David O. TeBeest, personal communication).

Bioherbicidal plant pathogens are generally evaluated for their efficacy (virulence toward the target weed), their potential for commercialization (performance under field conditions, specificity and host range, ease of inoculum production), and their compatibility with other pesticides that are used in crop production (herbicides to control other weeds, insecticides, fungicides).

The ability of a pathogen to cause damage on its target weed is influenced by several factors and their interactions, such as inoculum concentration and application rate, environmental conditions (temperature and relative humidity or moisture), formulation, spray parameters (droplet size, deposition, and distribution), target weed plant age or growth stage, non-target plant species, micro and macroorganisms in the phyllosphere or rhizosphere, and pesticides applied in the same area. Numerous studies have been done to find ways for improving bioherbicidal efficacy and achieving acceptable levels of weed control under conventional, non-conventional (organic, sustainable) and natural systems.

In Part 2 of this article, research that has been aimed at enhancing bioherbicidal activity, including formulation and application technology, will be discussed. Applications that have improved weed control efficacy for both classical biological control agents and bioherbicides, such as the combination of pathogens and insects, also will be examined. Lastly, we will discuss the future of pathogens for biological weed control.


Mention of a trademark, warranty, proprietary product or vendor does not constitute a guarantee by the United States Department of Agriculture and does not imply its approval to the exclusion of other products or vendors that may also be suitable.

Additional Resources

Weeds Newsletters
from Manaaki Whenua - Landcare Research

Classical Biological Control of Weeds
from Lethbridge Research Centre, Agriculture & Agri-Food Canada

TEAM Leafy Spurge
from USDA-ARS, Northern Plains Agricultural Research Laboratory

Invasive Plant Research Laboratory

Agricultural Permits: Weed Biocontrol

International Survey of Herbicide-Resistant Weeds

Multistate Research Project S-1001: Development of Plant Pathogens as Bioherbicides for Weed Control from Botany & Plant Pathology, Purdue University

Literature Cited

See Literature Cited page.