Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 61
4
Fusarium oxysporum formae
speciales as Candidate Biological
Control Agents for Cannabis and Coca
The genus Fusarium is one of the largest genera of fungi and includes
species that reproduce clonally and by recombination and species that reproduce
only clonally (Leslie and Summerell 2006). F. oxysporum is a species complex
(O’Donnell et al. 2004) that is pathogenic to a wide variety of plant species,
including several economically important vegetable and ornamental crops (Nel-
son et al. 1981; Michielse and Rep 2009). Sexual reproduction by F. oxysporum
has not been observed. However, genetic variation is significant in populations
of nonpathogenic strains, and strains of some form species (forma specialis, ab-
breviated f.sp.; plural formae speciales, abbreviated f.spp.) have few clonal
genotypes and large amounts of genetic variation (Leslie and Summerell 2006).
As Baayen et al. (2000) state, “although a teleomorph has not yet been found,
the sexual cycle may still be active in FOC [F. oxysporum complex].” F. ox-
ysporum occurs in all types of soils worldwide and causes severe vascular wilts,
damping-off, and crown and root rot in its hosts (Jarvis and Shoemaker 1978;
Nelson et al. 1981; Summerell et al. 2001; Di Pietro et al. 2003). In the absence
of a plant host, it can exist as a saprophyte (an organism that lives on dead or-
ganic matter) in soil for extended periods (Burgess 1981).
F. oxysporum has at least three and probably more species-level clades
(O’Donnell et al. 2004). Isolates are placed in form species on the basis of the
host-plant species that is attacked (Armstrong and Armstrong 1981; Kistler
1997). More than 150 form species of F. oxysporum have been characterized
(Armstrong and Armstrong 1981; Baayen et al. 2000). Nonpathogenic strains of
F. oxysporum also exist in soil and are distinguished from pathogenic strains
through pathogenicity testing (Edel et al. 2001). In some instances, a F. ox-
ysporum form species can have a wide host range and a single genotype is capa-
ble of parasitizing several different plants. For example, F. oxysporum f.sp. va-
sinfectum can be isolated from cotton, alfalfa, and tobacco (Assigbetse et al.
61
OCR for page 62
62 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
1994). Other forms show evidence of convergent evolution: several genotypes
have independently evolved the ability to parasitize the same plant species, for
example, F. oxysporum f.sp. cubense (O’Donnell et al. 1998).
BACKGROUND ON THE SPECIFIC FUNGI
Biological Control of Cannabis with Fusarium oxysporum f.sp. cannabis
Plants in the genus Cannabis, including C. sativa (fiber hemp and canna-
bis), are attacked by a number of pathogens, some of which may cause serious
damage or plant death. A review of scientific literature on cannabis pathogens
by McPartland (1992) identified at least 88 fungal species that attack cannabis.
Among those fungi are several Fusarium species that cause damping off, stem
canker, root rot, and wilt (McPartland 1996). Fusarium wilt of hemp, caused by
F. oxysporum f.sp. cannabis and F. oxysporum f.sp. vasinfectum, was first noted
in eastern Europe (Russia, the Czech Republic, Poland, and Romania) more than
50 years ago; it also reportedly occurs in western Europe (Italy and France),
central Asia (Kazakhstan and Pakistan), and North America (Canada and the
United States) (McPartland and Hillig 2004).
F. oxysporum f.sp. cannabis and F. oxysporum f.sp. vasinfectum are mor-
phologically similar in culture, but, according to McPartland and Hillig (2004),
the two can be differentiated on the basis of their host range. F. oxysporum f.sp.
cannabis is reported to infect only cannabis, whereas the host range of F. ox-
ysporum f.sp. vasinfectum includes (in addition to cannabis) cotton, mungbean,
pigeon pea, rubber tree, alfalfa, soybean, coffee, tobacco, and other plants
(McPartland and Hillig 2004). The study of genetic variability with DNA poly-
morphisms may provide an alternative route for identification of these two for-
mae speciales.
F. oxysporum f.sp. cannabis was considered a potential control agent for
cannabis as early as the 1970s on the basis of its purported specificity to mem-
bers of the genus Cannabis, its ability to survive in the soil for extended periods,
and the likelihood of infecting new plantings of the crop. Hildebrand and
McCain (1978) conducted laboratory experiments to develop a suitable tech-
nique for the production of F. oxysporum f.sp. cannabis inoculum consisting of
chlamydospores, which are suitable for soil application. Later, McCain and No-
viello (1985) explored the feasibility of using F. oxysporum f.sp. cannabis (iso-
lated in Italy) as a biological control agent against C. sativa and claimed that the
fungus caused disease only on C. sativa and that it was able to survive in the soil
for at least one growing season.
In the late 1990s, Tiourebaev et al. (2001) conducted experiments to test
the pathogenicity of F. oxysporum f.sp. cannabis isolates obtained from diseased
cannabis plants collected in various regions of Kazakhstan and to determine
their virulence, their host range, and the formulation best suited for field applica-
tion. The findings led them to conclude that the disease caused by F. oxysporum
OCR for page 63
63
Candidate Biological Control Agents against Cannabis and Coca
f.sp. cannabis was not severe enough to cause “permanent and lasting control”
of cannabis plants and that there was a need for an improved formulation and
improved delivery systems to enhance the pathogen’s efficacy as a mycoherbi-
cide (Tiourebaev et al. 2001). The results of those studies are reviewed below in
the context of efficacy, inoculum production and delivery, and persistence in the
environment. No studies of F. oxysporum f.sp. cannabis for the control of can-
nabis have been published since 2001.
Biological Control of Coca with Fusarium oxysporum f.sp. erythroxyli
Sands et al. (1997) isolated F. oxysporum from severely diseased coca
plants (Erythroxylum coca and E. novogranatense) grown by the U.S. Depart-
ment of Agriculture (USDA) on a secure experimental site on the island of
Kauai, Hawaii. This site was originally a coca plantation of a U.S. beverage
(soda) manufacturer in the 1960s and is now maintained by USDA for its herbi-
cide studies on coca. Coca plants grown on the site (grown from seeds imported
mainly from Peru) in the 1960s and 1970s exhibited damping off and wilt symp-
toms (Darlington 1996). Pathogenicity and host-range experiments led the inves-
tigators to identify the F. oxysporum as a unique forma specialis that attacks
members of the family Erythroxylaceae and to name it F. oxysporum f.sp.
erythroxyli (Sands et al. 1997).
In Peru, a similar disease of coca plants was observed in the 1930s (El
Comercio 1995), in the 1980s (Stevenson 1991), and in the 1990s (Arévalo et al.
1994). F. oxysporum was the causal agent of an epidemic in coca in the Hual-
laga Valley of Peru, and the disease was considered a threat to coca production
in that region (OTA 1993; Arévalo et al. 1994). In 1997, with the aid of random
amplified polymorphic DNA (RAPD) analysis, Nelson et al. identified two sub-
populations of F. oxysporum f.sp. erythroxyli in the Huallaga Valley (Nelson et
al. 1997). Later, Gracia-Garza et al. (1999) used RAPD and vegetative compati-
bility group (VCG) analyses to determine that F. oxysporum f.sp. erythroxyli
isolates from Peru and Hawaii were genetically similar. They speculated that the
pathogen might have been introduced into Hawaii in plant material from Peru
(Gracia-Garza et al. 1999). The known distribution of F. oxysporum f.sp.
erythroxyli today includes Peru, Hawaii, and possibly Colombia (El Comercio
1995; Nelson et al. 1997). On the basis of its ability to cause the observed natu-
ral wilt disease in coca plantings in Hawaii and Peru and its purported narrow
host range, it is considered a potential mycoherbicide for coca (Bailey et al.
1997; Sands et al. 1997).
Several studies of F. oxysporum f.sp. erythroxyli were conducted in the
United States in the middle to late 1990s to develop methods for inoculum pro-
duction and formulation for the purpose of evaluating its efficacy against coca
(Bailey et al. 1997; Hebbar et al. 1997; Connick et al. 1998) and to study its dis-
persal in the field (Bailey et al. 1997; Bailey et al. 1998; Gracia-Garza et al.
1998, 1999). The results of the studies are reviewed below in the context of effi-
OCR for page 64
64 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
cacy, inoculum production and delivery, and persistence in the environment. No
studies of F. oxysporum f.sp. erythroxyli for the control of coca have been pub-
lished since 1999.
EFFICACY AND IMPLEMENTATION
The overarching consideration in determining the feasibility of the pro-
posed mycoherbicides is the ability of F. oxysporum f.sp. cannabis and F. ox-
ysporum f.sp. erythroxyli to inflict severe damage or death on their target plants.
The committee reviewed the available data for evidence of such high levels of
efficacy as a prerequisite for the use of the identified strains as mycoherbicides.
In the case of registered mycoherbicides—such as Chontrol, Collego (Lock-
down), DeVine, and Sarritor—the most commonly used measure of efficacy is
reduction in weed numbers due to complete and fairly rapid killing of the target
weed after application. Accordingly, in the following analysis, efficacy is
viewed as the ability to yield an acceptable level of control of the target crops
within a short period after application, namely, a few days to a few weeks for
the annual crops of cannabis and a few months in the case of coca, a long-lived
tree crop. As mentioned in Chapter 1, eradication of the crops is not a realistic
goal, because biological agents rarely, if ever, kill 100% of their hosts. There-
fore, the committee looked for data on different steps in efficacy measurements,
such as greenhouse and field evaluations, with emphasis on the latter; appraisal
at relevant growth stages of the target crops; and suitable measures of control,
such as plant death and reduction in plant population density, plant growth rate,
or crop yield (as harvestable biomass of cannabis and coca). Disease assess-
ments reported in the literature, including visual ratings and direct measurement
of disease on individual plants and populations, help to quantify the destructive-
ness of the pathogens in controlled experiments, but they provide only an indi-
rect measure of the potential efficacy of the proposed mycoherbicides under
field conditions.
Fusarium oxysporum f.sp. cannabis
Only two publications shed light on the efficacy of F. oxysporum f.sp.
cannabis as a mycoherbicide for cannabis. McCain and Noviello (1985) in con-
ference proceedings reported on the effectiveness of F. oxysporum f.sp. canna-
bis against cannabis (industrial hemp) plants in the greenhouse, growth chamber,
and fields in Italy. Tiourebaev et al. (2001) in a short communication described
greenhouse and field studies to evaluate the pathogenicity and virulence of F.
oxysporum f.sp. cannabis isolates obtained from diseased cannabis plants col-
lected in various regions of Kazakhstan. The details of the experimental methods
used in those studies are presented in Tables 4-1 and 4-2.
OCR for page 65
65
Candidate Biological Control Agents against Cannabis and Coca
TABLE 4-1 Greenhouse and Field Studies in Italy by McCain and
Noviello (1985)
Greenhouse Study
Inoculum F. oxysporum f.sp. cannabis isolate recovered from industrial
hemp plants collected from fields in Italy in 1972
Inoculum (composed of mycelium, conidia, and chlamydospores)
produced on wheat straw and soybean meal
Five inoculum levels tested: 7, 70, 700, 1,400, and 7,000
propagules/g of soil
Test plants C. sativa cultivar Iran (susceptible variety)
Assessment method Efficacy based on percentage of plants killed
Field Studies
Location Field experiments conducted in Vitulazio, Alvignano, and
Portici, Italy, in 1974
Inoculum Air-dried straw-soybean inoculum used; applied by hand and
mixed with top 10 cm of soil
Vitulazio field trial: 10 g/m2; inoculated plots seeded with C.
Application rate and
cultivars used sativa cultivars CS and SF
Alvignano field trial: 10 g/m2; inoculated plots seeded with C.
sativa cultivars Iran and SF
Portici field trial: 1, 10, and 30 g/m2; inoculated plots seeded with
C. sativa cultivars Iran, CS, and SF; trial repeated with field soil in
large ceramic pots
Assessment method Efficacy based on plant mortality
In their greenhouse study, McCain and Noviello noted that the time re-
quired to kill all the seedlings was proportional to the inoculum level. At the
highest inoculum level (7,000 propagules/g of soil), 100% of the plants were
killed within 9 days after planting, whereas it took 47 days for all the plants to
die at the inoculum level of 70 propagules/g of soil. Only 50% of the plants were
killed at the lowest inoculum level (7 propagules/g of soil) at 47 days after plant-
ing.
McCain and Noviello reported that in field plots at Vitulazio, no Fusa-
rium-infected CS or SF hemp plants were observed during the study, but they
did not provide an explanation for the apparent failure of this trial. At the
Alvignano field trials, 4 months after the cultivars were planted in the fungus-
infested soil, 71% of the Iran plants had died and the ones that survived were
shorter than the plants in the noninoculated control plots. Only two SF plants
became infected—an indication that this cultivar is resistant to the disease. In
the Portici trial, 50%, 94%, and 94% of the Iran plants had died 4 months after
the cultivars were planted in the soil treated at 1, 10, and 30 g/m2, respectively.
The Iran plants that survived were stunted by 26 and 45% at the 1-g/m2 and 10-
g/m2 inoculum levels, respectively, compared with the Iran plants in the non-
inoculated control plots. At the highest inoculum level, 30 g/m2, only 4% and
14% of the cultivars CS and SF died.
OCR for page 66
66 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
TABLE 4-2 Greenhouse and Field Studies in Kazakhstan by Tiourebaev
et al. (2001)
Greenhouse Study
Inoculum 125 isolates of F. oxysporum obtained from diseased C.
sativa plants collected in various regions of Kazakhstan
125 isolates of F. oxysporum “evaluated for pathogenicity”
to C. sativa
All isolates formulated with wheat or oat seeds or birch
sawdust and applied at 0.5 g per 2-cm2 pot of soil
Test plants C. sativa (cultivar not stated)
Environmental conditions 28°C; 19-h photoperiod
Assessment method Disease severity rating: 1 = healthy plant; 2 = wilting of
lower leaves; 3 = wilting of 25-50% of leaves; 4 = over 50%
of leaves wilted; 5 = dead plant
Field Studies
Location Kazakhstan field naturally infested with C. sativa plants
(cultivar not stated)
Inoculum 25 F. oxysporum isolates “that showed virulence and host
specificity towards C. sativa” in the greenhouse
Trial 1: 25 isolates formulated with wheat seed, oat seed,
and birch sawdust
Trial 2: 25 isolates formulated in birch sawdust
12.5 g/m2 (plot size, 2 m2); inocula of the 25 isolates applied twice:
Application rate
when C. sativa seedlings were 2-3 wk old and when the plants
were 6-7 wk old
Assessment method Estimation of “percentage of infected plants within treated plots”
done on 2-m transect laid across plot; disease severity calculated
with the equation D = a/b(100), where D = disease severity,
a = number of C. sativa plants with disease severity rating of
3-5 (see rating scale in greenhouse studies), and b = total number
of C. sativa along the transect with 1-5 disease severity rating
There are a few important shortcomings in the experiments and paper by
McCain and Noviello (1985): the greenhouse trial of the inoculum rate study
was done only once, and the number of replications per treatment was not stated;
the temperature and humidity conditions during the greenhouse and field trials
were not stated; and the isolates used were not specifically identified. Thus, the
experiments are not readily repeatable, and the results cannot be accepted as
critical evidence.
McCain and Noviello screened several other unnamed C. sativa cultivars
besides those in Table 4-1 and seed collections and discovered that some indus-
trial hemp varieties were resistant to F. oxysporum f.sp. cannabis. That led them
to conclude that the presence of resistant varieties might limit the effectiveness
of the mycoherbicide.
OCR for page 67
67
Candidate Biological Control Agents against Cannabis and Coca
The results of the study by McCain and Noviello confirm that F. ox-
ysporum f.sp. cannabis is a pathogen of C. sativa, but the pathogen’s efficacy as
a mycoherbicide agent is low. By the authors’ own admission, the time required
for disease development and death of cannabis (at least 4 months) is a limitation
(McCain and Noviello 1985). The decline of populations of F. oxysporum f.sp.
cannabis in the soil in the absence of the host plant suggests that the fungus may
not persist unless the fields are continuously cropped with cannabis. The au-
thors’ claim that complete coverage of a field is unnecessary because natural
spread would occur is not backed up by the data that they presented. Therefore,
the hypothesis that F. oxysporum f.sp. cannabis will persist in treated soil and
can infect new crops of cannabis remains untested.
Tiourebaev et al. (2001) conducted greenhouse and field experiments with
a fairly large collection of F. oxysporum f.sp. cannabis isolates. In the green-
house study that was conducted only once, 25 of the 125 isolates tested were
found to cause disease in cannabis. Plant mortality for the wheat-seed, oat-seed,
and birch-sawdust formulations was 64%, 61%, and 59%, respectively. Tioure-
baev et al. noted that cannabis mortality varied according to the fungal isolate.
In the first field experiment, 12 of the 25 isolates that were previously
screened in the greenhouse study were found to be pathogenic to cannabis. Iso-
late CR-21 caused the most damage, namely, wilting of the plant within 2
weeks. Other symptoms on the treated plants were stunting, leaf curling, root
discoloration, and loss of structural integrity in the upper part of the plant. At the
end of the first field experiment (5 months after the initial application of the
formulations), disease incidence ranged from 12% to 67% for the isolates tested.
Disease severity ranged from 1.27 to 2.00 on a scale of 1-5 (see Table 4-2), in
which a rating of 2.00 corresponds to wilting of lower leaves. In the second field
trial, disease incidence ranged from 6.8% to 39% and disease severity from 1.4
to 2.7 for the isolates tested. Plot size was 2 m2, and the number of replicates
was not mentioned (Tiourebaev et al. 2001).
Like McCain and Noviello, Tiourebaev et al. (2001) confirmed F. ox-
ysporum f.sp. cannabis as a pathogen of cannabis. However, the levels of dis-
ease, based on disease-severity and disease-incidence estimates, were low to
moderate and led the authors to conclude that under field conditions “the infec-
tion rate was still too low to affect permanent and lasting control of the weed.”
Thus, the Kazakh isolates of F. oxysporum f.sp. cannabis tested by Tiourebaev
et al. (2001) do not appear to be efficacious or suitable for development as a
mycoherbicide.
According to Tiourebaev et al. (2001), the “pathogenicity of the isolates
varied greatly” between greenhouse and field trials. One explanation for the
variation could be temperature: the greenhouse temperature during the experi-
ment was 28°C with 19 hours of daylight, whereas the temperature in the field
ranged from 20°C to 30°C during May or June, the time when infection began to
appear. The disease levels were highest in late August and September and, ac-
cording to the authors, the “disease tapered off as cooler conditions prevailed”
OCR for page 68
68 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
(Tiourebaev et al. 2001). Another factor might be the genetic composition of the
cannabis plants in the field; inasmuch as the trial took place in a field with a
natural cannabis infestation, the hosts’ genetic variability could have affected
the isolates’ field performance.
The identification of the isolates as F. oxysporum was based on “micro-
scopic analysis and colony morphology” (Tiourebaev et al. 2001). The authors
assigned the name F. oxysporum f.sp. cannabis without completing Koch’s pos-
tulates to prove their pathogenicity toward cannabis. Without additional support
from molecular or VCG analysis, this identification is unlikely to be acceptable
for mycoherbicide-registration purposes. It cannot be ruled out that this large
collection of F. oxysporum isolates from the wild included several formae spe-
ciales that are cross-infective on cannabis and one or more additional hosts, with
cannabis being an alternative rather than a primary host.
Overall, the data on F. oxysporum f.sp. cannabis gleaned from the publi-
cations of McCain and Noviello (1985) and Tiourebaev et al. (2001) are not
sufficient to draw conclusions about the feasibility of F. oxysporum f.sp. canna-
bis as a mycoherbicide.
Fusarium oxysporum f.sp. erythroxyli
The committee reviewed the available data on the efficacy of F. ox-
ysporum f.sp. erythroxyli in three publications: Sands et al. (1997), Bailey et al.
(1997), and Bailey et al. (1998). The study by Sands et al. consisted of experi-
ments in the field and in a growth chamber to assess the virulence of F. ox-
ysporum f.sp. erythroxyli isolated from an Erythroxylum population growing in
Hawaii. Bailey et al. (1997) performed three field trials in Hawaii with a rice-
alginate prill formulation of strain EN-4 of F. oxysporum f.sp. erythroxyli,
which was one of the Hawaiian isolates studied by Sands et al. (1997), to test the
feasibility of enhancing the pathogen populations in the soil and to cause disease
in coca. Later, Bailey et al. (1998) examined six formulations, including the
rice-alginate prill, for their ability to enhance the pathogen’s populations and
cause disease on coca in the field. The experimental methods used in the studies
are summarized in Tables 4-3 and 4-4.
In the growth-chamber study by Sands et al. (1997), severe disease was
not observed with any of the F. oxysporum f.sp. erythroxyli strains tested when
the inoculum was placed in 10-cm holes around the plant. Better results were
obtained by first infesting the soil with the fungus and then transplanting coca
plants into the soil. Wilting was observed 3 weeks after the plants were trans-
planted, and the plants eventually died; however, no quantitative data were pro-
vided.
In the field study, the initial disease symptoms were leaf drop and death
of a “few lower stems”; plant death was observed 7 weeks after the plants
were transplanted into the fungus-infested soil. The time from the appearance
OCR for page 69
69
Candidate Biological Control Agents against Cannabis and Coca
TABLE 4-3 Growth-Chamber and Field Studies by Sands et al. (1997)
Growth-Chamber Study
Inoculum F. oxysporum strains tested: Ec1-3 and EN1-4 (from Hawaii)
and SA1 (from South America)
Soil in pots planted with test plants infested with millet seeds
colonized by F. oxysporum by placing seeds in three 10-cm
holes (10 seeds/hole)
Test plants 6- to 12-mo E. coca plants
Environmental conditions 30°C/28°C day/night temperature; 12-h light/dark periods
Test plants watered daily (80-100 mL)
Assessment method Test plants evaluated for disease weekly over 3.5-mo period
Field Study
Location Conducted in 1989 on field plots transplanted with E. coca
and E. novogranatense in the previous fall (1988)
Inoculum Field plots infested with millet seeds colonized by F. oxysporum
Seven Hawaiian isolates of F. oxysporum evaluated for
pathogenicity on coca
Environmental conditions All infested plots received trickle irrigation
Application method Subsurface method: inoculum (5.0 g /plot) deposited in
and rate 2-cm-deep V-shaped trench and covered with 2-3 cm of soil
Surface method: inoculum applied directly to the soil surface
and covered with mesh screen to protect from predation
Assessment method Plants in treated and nontreated soil evaluated for severity
and mortality after 7, 8, 9, 10, and 15 mo
Efficacy assessed with a disease-severity rating scale of
0-2: 0 = no disease, 1 = wilt, 2 = plant death
of initial disease symptoms to plant death ranged from days to months. The av-
erage disease severity for all seven isolates tested was 1.1 (on a scale of 0-2) in
the subsurface-applied plots and 1.03 in the surface-applied plots. Plants in the
control plots had an average disease severity rating of 0.71, that is, plants were
symptomatic but not dead. Some control plants also developed disease, pre-
sumably from naturally occurring inoculum of F. oxysporum f.sp. erythroxyli. E.
coca plants were more severely diseased (rating, 1.22) than E. novogranatense
plants (0.93).
Apparently on the basis of a combination of the data from all seven iso-
lates and two methods of inoculum application, Sands et al. (1997) reported that
94% of E. coca and 49% of E. novogranatense plants (42 per species) in the
fungus-infested soil were killed at 15 months after inoculation, whereas about
95% of E. coca and 43% of E. novogranatense plants (28 plants per species) in
the noninfested control soil (values read off the graph in Figure 1 of Sands et al.
[1997]) had died. F. oxysporum was isolated from the vascular tissues of symp-
tomatic and asymptomatic plants from the fungus-infested soil and from symp-
OCR for page 70
70 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
tomatic plants in the noninfested soil; this implied nonpathogenic colonization
of coca plants by F. oxysporum f.sp. erythroxyli and fairly high levels of natural
incidence of wilt at the test site.
TABLE 4-4 Field Studies in Hawaii by Bailey et al. (1997, 1998)
First Set of Field Trials
Location Three experiments conducted from 1995 to 1996 in
Kauai, Hawaii
Two trials in fields continuously planted with coca for at
least 7 years; one trial in a field not previously planted to coca
Environmental conditions Fields irrigated daily for 14 d after F. oxysporum f.sp.
erythroxyli strain EN-4 formulations were incorporated in soil
Temperature 17-21°C minimum and 23-27°C maximum;
regular afternoon rains (3.9-57.9 cm)
Inoculum Alginate prill formulation F. oxysporum f.sp. erythroxyli
strain EN-4
Application rate 33.6 kg/ha
Assessment method F. oxysporum f.sp. erythroxyli strain EN-4 efficacy assessed
with disease-severity rating scale of 0-2: 0 = asymptomatic,
1 = symptomatic (plant defoliating), 2 = dead
Second Set of Field Trials
Location Three experiments in 1995-1996 in Kauai, Hawaii
Two experiments in fields continuously planted with coca
for at least 7 years; one trial in a field not previously planted
to coca
Low percentage of diseased coca plants present in all field
plots before the experiment was conducted
Environmental conditions Fields irrigated daily for 14 d after the F. oxysporum f.sp.
erythroxyli strain EN-4 formulations were incorporated in soil
For experiments 2 and 3, average low air temperature: 19.5°C
and 21.3°C; average low soil temperature: 22.6°C and 24.2°C;
relative humidity for both trials from less than 80% to over
97% during the first 10 d after application of formulations
Inoculum Six formulations (rice-alginate, C6, Pesta, canola-alginate,
rice-alginate + canola oil, corn cob-alginate) and biomass
alone of F. oxysporum f.sp. erythroxyli strain EN-4
Application rate 33.6 kg/ha
Assessment method F. oxysporum f.sp. erythroxyli strain EN-4 efficacy assessed
with disease-severity rating scale of 0-2: 0 = asymptomatic,
1 = symptomatic (plant defoliating), 2 = dead
OCR for page 71
71
Candidate Biological Control Agents against Cannabis and Coca
These results confirm F. oxysporum f.sp. erythroxyli as a virulent patho-
gen of E. coca and E. novogranatense; with respect to efficacy, however, the
data are equivocal. No quantitative data from the growth-chamber study were
provided, so these data are of limited value in assessing efficacy. The data in
Figure 1 of the paper shows that the percentages of dead control and fungal-
treated E. coca plants were nearly equal at the end of the experiment. In the case
of E. novogranatense, nearly 60% (not 49% as stated in the publication) of the
inoculated plants and nearly 50% of the control plants appear to have died.
Without the benefit of statistical analysis of data in Figure 1, the small number
of plants tested (42 and 28 plants per species in the fungus-infested and fungus-
free soils, respectively) raises the question whether the naturally prevalent in-
oculum at the field site rather than the experimentally applied inoculum was
responsible for the observed results. Furthermore, the lack of information on
climatic conditions (average daily temperature, relative humidity, soil moisture)
during the study and the use of 6- to 12-month-old seedling transplants instead
of older and established coca plants are additional limitations.
In two later studies, Bailey et al. (1997, 1998) reported an increase in the
rate of disease development in plots treated with different formulations com-
pared with the rate in untreated plots, which had only background levels of
Fusarium wilt. However, the mortality of coca plants was variable, ranging from
35% to 85% among all seven formulations tested (Bailey et al. 1998).
The results obtained by Bailey et al. are consistent with the finding by
Sands et al. (1997) that F. oxysporum f.sp. erythroxyli is a virulent pathogen of
coca plants. However, as in the study by Sands et al., the results with respect to
efficacy are ambivalent. Bailey et al. (1997) claimed that “the 33.6 kg/ha rate of
the rice-alginate prill formulation enhanced the killing of coca plants in three
different field experiments over 7-9 months,” but the data in the paper do not
report plant mortality; the maximum disease rating for up to 350 days was under
1; that is, the plants were symptomatic, not dead. Bailey et al. (1997) also re-
ported that some plants “appear to seal off the infected areas and resume normal
growth patterns,” and this could indicate the plants’ ability to overcome the dis-
ease.
The incidence of natural wilt of about 58% at the experimental site in Ha-
waii where Bailey et al. (1997) performed the first set of field studies and the
17% mortality in the second set of field studies (Bailey et al. 1998) are problem-
atic. Against such background disease levels, the increase in disease incidence
resulting from the added inoculum cannot be ascertained on the basis of the col-
lective data of Bailey et al. (1997, 1998) and Sands et al. (1997).
Bailey et al. (1997) also state that by “maintaining soil surface moisture
with drip irrigation, overhead watering, and/or rain, the rice-alginate prill formu-
lation germinated relatively uniformly in two greenhouse experiments and at
least two field experiments when applied to the soil surface.” Thus, the rice-
alginate prill formulation appears to be a feasible and effective way to deliver
the inoculum if it is deployed under conditions where adequate moisture is pre-
sent.
OCR for page 90
90 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
Transmission and Spread
The potential pathways by which spores and other propagules from a par-
ticular application site might move via physical transport mechanisms are illus-
trated in Chapter 2 (see Figure 2-1). In general, dispersal of the proposed myco-
herbicides after application will depend on the production and natural dispersal
of secondary inoculum. In the case of F. oxysporum, secondary spread was
originally thought to be limited to root-to-root contact and distribution of in-
fected plant tissues and seeds (Gambogi 1983; Gracia-Garza et al. 1999; Rekah
et al. 2001). However, additional evidence indicates that under humid condi-
tions, several formae speciales of F. oxysporum produce conidia in stroma (mass
of hyphae) on the lower stems of infected plants (Rowe et al. 1977; Timmer
1982; Woudt et al. 1995; Gamliel et al. 1996; Vakalounakis 1996; Katan et al.
1997; Rekah et al. 1999, 2000, 2001; El-Hamalawi and Stanghellini 2005).
These conidia can be dispersed by rain, wind, or insects (Gillespie and Menzies
1993; El-Hamalawi 2008). Inoculum could also be transferred from the applica-
tion site by several vectors, including humans (transporting plant material or
infested seeds), the hides or fur of mammals (such as pack animals), the drop-
pings of birds, or the surface of insects. Long-distance transport via waterways
is a potential route inasmuch as viable and infectious propagules of F. ox-
ysporum have been found in rivers and seas (Palmero et al. 2009), although the
distance over which such transport might occur is unknown. Finally, windstorms
could carry inoculum over long distances to areas far outside the target areas
(Maldonado-Ramirez et al. 2005). Thus, the potential exists for the proposed
mycoherbicides to move out of the geographic areas where they are applied.
Only a few studies have examined the dispersal of F. oxysporum f.sp.
cannabis and F. oxysporum f.sp. erythroxyli. In a field study in Kazakhstan,
consumption and dispersal by insects, lizards, rodents, and birds was reported
for a seed-based inoculum of F. oxysporum f.sp. cannabis (Tiourebaev et al.
2001). In a field study in Hawaii, the applied inoculum of F. oxysporum f.sp.
erythroxyli was foraged by ants and became clustered in their nests (Gracia-
Garza et al. 1998).
In addition, many Fusarium species, including F. oxysporum, survive on
and in seeds (Nelson et al. 1997; Garcia-Garza et al. 1999; Mbofung and Pryor
2007). Transmission by seeds and planting materials carries the risk of unin-
tended spread beyond target areas in that farmers and traders could carry in-
fected materials throughout a region or even into new areas, as has been pro-
posed for the transport of F. oxysporum f.sp. erythroxyli that appears to have
arrived in Hawaii from Peru (Fravel et al. 1996; Sands et al. 1997).
Other Host Plants
In general, Fusarium species are ubiquitous in the environment and can
survive on host plants that may or may not be susceptible to them (Blok and
OCR for page 91
91
Candidate Biological Control Agents against Cannabis and Coca
Bollen 1997). In a host-range study of F. oxysporum f.sp. erythroxyli, Sands et
al. (1997) reported that the pathogen was isolated from the crown tissue of
symptomless plants of 26 species grown in treated soil. Thus, this fungus could
colonize plants other than coca. However, no studies have evaluated the long-
term persistence of either F. oxysporum f.sp. erythroxyli or F. oxysporum f.sp.
cannabis through observations on their survival in soil after the host plant dies
or evaluated their potential to colonize later plantings of coca or cannabis.
The recent study of genetic variation at two loci among F. oxysporum in-
dividuals pathogenic on animals and plants (O’Donnell et al. 2009) showed that
F. oxysporum f.sp. erythroxyli has a unique genotype at both loci but that F.
oxysporum f.sp. cannabis shared one or both alleles with individuals that were
isolated from other plants. Although it certainly is possible that either Fusarium
mycoherbicide could attack nonhost plants, it also is true that the mycoherbicide
strains would not be expected to behave differently from members of the two
Fusarium species that are already present in the endemic areas.
Competition with Other Soil Microorganisms
Although it was demonstrated that F. oxyspsorum f.sp. erythroxyli can be
suppressed in some soil types (Fravel et al. 1996), no studies are available on the
interactions of F. oxysporum f.sp. cannabis or F. oxysporum f.sp. erythroxyli
with particular soil microorganisms or other organisms. Studies of other F. ox-
ysporum formae speciales and other fungi suggest that the presence of competi-
tor or antagonistic microorganisms could reduce persistence. For example,
springtails and mites feed on plant pathogenic fungi (Nakamura et al. 1992;
Okabe 1993), and several bacteria and fungi can suppress the pathogens respon-
sible for Fusarium wilt of melon (Suárez-Estrella et al. 2007), tomato (Larkin
and Fravel 1999), and chickpea (Landa et al. 2004). In fact, such antagonists
have been pursued as biological control agents against these wilt diseases. Thus,
antagonistic microorganisms in the soil could theoretically lessen the likelihood
that the proposed mycoherbicides would establish sufficiently high inoculum
levels in the rhizosphere to cause wilt disease of cannabis or coca. In contrast, it
is possible that the mycoherbicide strains could displace resident strains; but,
again, no data are available to evaluate this possibility.
Conclusions
It is clear that F. oxysporum f.sp. erythroxyli has the ability to colonize
and survive in soil and coca roots for several months, but there are insufficient
data to predict the long-term persistence of F. oxysporum f.sp. erythroxyli, in
that none of the studies lasted for more than 7 months. However, the natural
presence of this fungus and the recurrent incidence of the coca wilt disease in
the Huallaga Valley of Peru (see discussion later in this chapter) suggest that F.
oxysporum f.sp. erythroxyli is capable of long-term persistence. Inasmuch as F.
OCR for page 92
92 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
oxysporum f.sp. erythroxyli could survive on plants other than coca and on de-
caying organic material, it is likely that the mycoherbicide strain would persist
in the indigenous fungal populations at some level once it is introduced in coca
fields. In general, the environmental factors that affect the persistence of F. ox-
ysporum f.sp. erythroxyli may also affect F. oxysporum f.sp. cannabis, but too
little information is available on the latter forma specialis to draw any conclu-
sions beyond this generalization.
EFFECTS ON NONTARGET ORGANISMS
Microbial pesticides are regulated by the U.S. Environmental Protection
Agency, which requires a variety of testing of the environmental fate and safety
of pesticides before they are registered. Chapter 2 and Appendix B describe the
types of testing required, including product analysis, pesticide residue analysis,
toxicity testing, toxicity and pathogenicity testing on nontarget organisms, and
assessment of environmental fate. The aforementioned studies have not been
conducted with the goal of registering the tested strains of F. oxysporum f.sp.
cannabis or F. oxysporum f.sp. erythroxyli as mycoherbicides, so this section
reviews the existing data on them that are available in the open literature and are
pertinent to understanding potential adverse effects on nontarget plants and or-
ganisms. Consideration is given to the issues specified in the committee’s state-
ment of task, including potential effects on licit crops, other soil fungi, animals,
humans, biodiversity, and other relevant aspects of environmental health.
Effects on Nontarget Plants
Some research has been performed to evaluate the effects of F. oxysporum
f.sp. cannabis and F. oxysporum f.sp. erythroxyli on native plant species, but the
reported data are sparse. For example, Tiourebaev et al. (2000, 2001) tested F.
oxysporum f.sp. cannabis on several plant species in Kazakhstan, including food
crops and native plants (see Table 4-11). The first paper appeared in the pro-
ceedings of a conference and reported that none of 13 plants tested exhibited
symptoms of Fusarium wilt but provided no experimental data to support the
claim (Tiourebaev et al. 2000). The second paper reported that host-range tests
were conducted under greenhouse conditions on five crops that were commonly
cultivated in the test region of Kazakhstan (tomato, corn, wheat, bean, and po-
tato) but contained no data or discussion of the results obtained pertaining to
those crops (Tiourebaev et al. 2001); instead, the authors simply state that 12
nontarget plants “showed no evidence of infection by the pathogen” and seemed
to imply that the results were the results of field tests.
The possible taxonomic and etiological relationships of F. oxysporum f.sp.
cannabis to other wilt-causing Fusarium species that attack plants related to
cannabis, such as the economically important hops (Humulus lupulus), were not
OCR for page 93
93
Candidate Biological Control Agents against Cannabis and Coca
adequately addressed either by McCain and Noviello (1985) or by Tiourebaev et
al. (2001). The claim by McCain and Noviello that hops “are not known to be
susceptible to a [Fusarium] wilt disease anywhere in the world” is erroneous
(McPartland and West 1999; Solarska 2001).
One host-range study of F. oxysporum f.sp. erythroxyli was identified.
Sands et al. (1997) tested strains of this fungus from Hawaii and South America
on 26 plant species (see Table 4-12) grown in growth chambers from seeds
planted in pathogen-infested soil. Wilt symptoms were not observed on any of
the native species. However, when crown tissues of the species grown in the
infested soil were plated on a selective medium at the conclusion of the experi-
ment, F. oxysporum was isolated from all species. Those results indicate that F.
oxysporum f.sp. erythroxyli was not pathogenic to the tested hosts, but because it
was present in the crowns the authors concluded that it could infect nontarget
host tissues.
TABLE 4-11 Plants Reportedly Tested in Host-Range Studies of Fusarium
oxysporum f.sp. cannabis
Tiourebaev et al. 2000 Tiourebaev et al. 2001
Agropyron spicatum (wheat grass) Crops:
Cucumus sativus (cucumber) Lycopersicon esculentum (tomato)
Daucus carota (wild carrot) Phaseolus vulgaris (kidney bean)
Festuca arundinaceae (tall fescue; grass) Solanum tuberosum (potato)
Gossypium hirsutum (cotton) Triticum aestivum (wheat)
Hordeum vulgare (barley) Zea mays (corn)
Lycopersicon esculentum (tomato)
Native plants:
Melilotus indica (sweet clover; herb)
Agropyrum pectiniforme (crested wheatgrass)
Phaseolus vulgaris (kidney bean)
Alhagi pseudalhagi (camelthorn; legume)
Pisum sativum (garden pea)
Artemisia vulgaris (common wormwood;
Raphanus sativus (radish) herbaceous perennial)
Triticum aestivum (wheat) Atriplex alba (lambsquarters goosefoot; shrub)
Zea mays (corn) Bromus inermis (awnless brome; grass)
Ceratocarpus arenarius (jiao gulo li; herb)
Chenopodium sp. (common lamb’s quarters;
herbaceous annual)
Dactylis glomerata (orchardgrass)
Glycyrrhiza glabra (cultivated licorice; legume)
Kochia prostrata (forage kochia; shrub)
Stipa dasyphylla (grass)
Tragopogon major (yellow salsify)
OCR for page 94
94 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
TABLE 4-12 Plants Reportedly Tested in Host-Range Studies of Fusarium
oxysporum f.sp. erythroxyli
Abelmoschus esculentus (okra) Helianthus annuus (sunflower)
Allium cepa (onion) Hordeum vulgare (barley)
Arachis hypogaea (peanut) Lactuca sativa (lettuce)
Beta vulgaris (sugar beet) Lycopersici esculentum (tomato)
Capsicum annuum (cayenne pepper) Oryza sativa (rice)
Carthamus tinctorius (safflower) Phaseolus vulgaris (kidney bean)
Citrullus lanatus (watermelon) Pisum sativum (garden pea)
Cucumis melo var. cantalupensis (cantaloupe) Raphanus sativus (radish)
Cucumis sativus (cucumber) Sorghum bicolor (sorghum)
Cucurbita moschata (pumpkin) Triticum aestivum (wheat)
Daucus carota subsp. sativus (carrot) Vigna unguiculata (cowpea)
Glycine max (soybean) Zea mays (corn)
Gossypium barbadense (annual long-fiber cotton)
Source: Adapted from Sands et al. 1997.
Overall, the data on the host range of F. oxysporum f.sp. cannabis and F.
oxysporum f.sp. erythroxyli are insufficient to draw any conclusions. Erythroxy-
lum is a large genus; it has as many as 250 species, of which about 200 have the
same native habitat as the coca plants (Plowman 1980). Thus, it is important to
determine whether F. oxysporum f.sp. erythroxyli could attack potential alterna-
tive hosts and whether any of them are of ecological significance, for example,
endangered, threatened, or of particular value.
Conducting such tests and interpreting the data from them may be difficult
because F. oxysporum is ubiquitous and “the host range tests required to prove
specificity [of the formae speciales] would be very large and likely prohibitive”
(Berner and Bruckart 2005, p. 227).
Effects on Legal Crop Production
The potential risk to legal production of cannabis and coca has not been
given much attention. Commercial cultivation of hemp is not legal in the United
States, but hemp is grown in Canada and many other countries (for example,
several European countries, China, and Russia) for fiber, for linen production,
and for seeds, which are a source of meal and oil used in food and other prod-
ucts (USDA 2000).
McCain and Noviello (1985) claimed that “because of the restricted host
range, no special precautions are necessary for controlled use of the fungus,”
even though the industrial hemp cultivars tested developed 10-100% disease.
They also claimed that cannabis cultivars or seed collections from Czechoslova-
kia, India, Iran, Mexico, Nepal, Pakistan, Poland, South Africa, Thailand, and
OCR for page 95
95
Candidate Biological Control Agents against Cannabis and Coca
Turkey were “all highly susceptible to the disease and in general there were no
survivors or escapes in [a] growth chamber study.” However, no data were pro-
vided on the extent of cultivar differences in susceptibility. On the basis of the
McCain and Noviello (1985) data, F. oxysporum f.sp. cannabis could pose a risk
to commercial hemp grown for fiber.
Toxicity to Wildlife, Domestic Animals, and Humans
This subsection focuses on the potential toxicity to nontarget organisms.
As discussed in Chapter 1, toxicity refers to the degree or extent of harm caused
by a chemical. In the case of the proposed mycoherbicides, the greatest risk to
animals and humans probably would come from the consumption of the patho-
gen in colonized plant material. Toxins might be produced in the materials used
to deliver the inoculum. Tiourebaev et al. (2001) found that when F. oxysporum
f.sp. cannabis was applied to wheat and oat seeds, insects, lizards, and rodents
consumed the inoculum. There are no studies of mycotoxin production by F.
oxysporum f.sp. cannabis or F. oxysporum f.sp. erythroxyli. As noted earlier in
this chapter, F. oxysporum is a large and diverse species complex, and the genus
Fusarium is even larger, so generalizations or assumptions about the production
of mycotoxins by the proposed mycoherbicide strains cannot be made, even
though there is a great deal of information about mycotoxins produced by mem-
bers of the genus Fusarium (Marasas et al. 1984; Desjardins 2006).
Filamentous fungi produce a variety of metabolites. Primary metabolites
are chemicals that are required for the survival of an organism, and secondary
metabolites are organic compounds whose production by an organism is not
required for growth, development, reproduction, or immediate survival. Biosyn-
thesis of secondary metabolites often begins with components derived from pri-
mary metabolism. Nearly all fungi produce at least one secondary metabolite.
Although the functions of most secondary metabolites are unknown, many have
been shown to be biologically active. Many have antimicrobial activity and thus
might play a role in competition with other microorganisms. Others are toxic to
plants and function as pathogenicity factors.
Secondary metabolites that are toxic to animals and humans are called
mycotoxins. Mycotoxins are grouped into classes on the basis of their chemical
structure. All members of a class have a common core structure; each member
has its own unique chemical modifications of the common core. The toxicity of
members of a class is diverse; they can range from nontoxic to lethal. Sensitivity
to mycotoxins also varies: some animals exhibit acute reactions to a particular
mycotoxin and members of other species find it innocuous or require long expo-
sures to produce any observable symptoms.
Some members of the genus Fusarium produce toxic secondary metabo-
lites (Marasas et al. 1984; Desjardins 2006). Table 4-13 presents a summary of
the classes of mycotoxins reportedly produced by F. oxysporum. Most research
OCR for page 96
96 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
TABLE 4-13 Classes of Mycotoxins Produced by Fusarium oxysporum
Mycotoxin Class Source Reference
Beauvericin Maize culture Logrieco et al. 1998
Fumonisins Wheat culture Seo et al. 1996
Fusaric acid Maize culture Bacon et al. 1996
Fusarin C Defined medium Cantalejo et al. 1999
Moniliformin Rice culture Abbas et al. 1990
Trichothecenes Defined medium Cantalejo et al. 1999
Wortmannin Rice culture Abbas and Mirocha 1988
on Fusarium mycotoxins has focused on the ones that contaminate important
food and feed crops, such as corn, wheat, barley, and potato. Some highly toxic
metabolites—such as trichothecenes, moniliformin, fumonisins, fusarin C, and
zearalenone—have been found (Kim and Lee 1994; Cantalejo et al. 1999; Seo et
al. 1999; Miller 2002; Rheeder et al. 2002).
Trichothecene toxins include T-2 toxin and diacetoxyscirpenol, which are
on the U.S. select-agents list and pose severe threats to animal or plant health or
products. T-2 and diacetoxyscirpenol are most commonly produced by Fusa-
rium species other than F. oxysporum. The data on trichothecene production in
the report by Cantalejo et al. (1999) are questionable because the identification
of the one F. oxysporum isolate to produce trichothecene was based solely on
morphologic characteristics of cultures that were not grown under the standard
conditions (Leslie and Summerell 2006) used for the accurate identification of
these fungi. Given that background, it is unlikely that any F. oxysporum strains
proposed as mycoherbicides would mutate to produce T-2 or diacetoxyscirpenol,
although it would be straightforward to test mycoherbicide strains for these
molecules. Because the biosynthetic pathways for these toxins are known (e.g.,
Desjardins et al. 1993; Kimura et al. 2007), sequencing of appropriate portions
of the genomes of the proposed mycoherbicide strains could be used to deter-
mine whether the genes encoding the enzymes in these pathways are present. If
some of or all the necessary genes are missing, it is unlikely that the strains
could mutate to produce either of the toxins.
Fumonisins are a large family of related compounds, a few of which are
produced by some strains of F. oxysporum. The analogues most commonly as-
sociated with human health problems are those in the fumonisin B group. F.
oxysporum strains are known to produce members of the fumonisin C series
(Seo et al. 1996; Sewram et al. 2005), but these toxins are not produced at high
concentrations even under optimal conditions. Some isolates have also been
shown to produce small amounts of fumonisin B1/B2 in culture (Waskiewicz et
al. 2010). Strains of F. verticillioides have been identified as responsible for the
production of the fumonisin that caused neural-tube defects in newborns in the
Rio Grande valley of Texas and in the Transkei region of South Africa (Ncayi-
OCR for page 97
97
Candidate Biological Control Agents against Cannabis and Coca
yana 1986; Hendricks 1999; Gelineau-van Wase et al. 2009). A strain of F. ox-
ysporum used as a bioherbicide of striga was grown in culture medium known to
encourage mycotoxin production (Savard et al. 1997). Small amounts of fusaric
acid were detected but no other mycotoxins.
Pathogenicity to Animals and Humans
There are reports in the literature of about 150 cases of Fusarium infec-
tions in immunocompetent people and 330 cases in immunocompromised peo-
ple (references include Nelson et al. 1994; Guarro and Gene 1995; Baran 1997;
Gianni et al. 1997; Romano et al. 1998, 2010; Sander et al. 1998; Musa et al.
2000; Bodey et al. 2002; Nucci and Anaissie 2002; Albisetti et al. 2004; Naiker
and Odhav 2004; O’Donnell et al. 2004; Guilhermetti et al. 2007; Nucci and
Anaissie 2007; Nucci et al. 2010). Among the cases reported in immunocompe-
tent people, virtually all were nail infections (onychomycosis) with or without
accompanying localized cutaneous infection (paronychia). Cases of keratitis
were excluded from this review because Fusarium solani was reported as the
etiologic agent in the overwhelming majority of cases. Increased Fusarium in-
oculum in the soil and in the environment could potentially increase the number
of these types of cases in the population of the target areas. Although the infec-
tions lead to cosmetic disfigurement of affected nails, the infections are self-
limited in immunocompetent hosts.
In immunocompromised hosts, fusariosis is a much more serious infec-
tion. Although many of the infections in this population begin as nail and super-
ficial skin infections, the organism rapidly invades blood vessels and spreads
hematogenously (Nucci and Anaissie 2002, 2007). These infections are difficult
to diagnose and treat and lead to poor survival rates in infected patients. Al-
though increased Fusarium in the environment could lead to increased infections
in immunocompromised populations, the population at risk in developing na-
tions would be estimated to be somewhat smaller than in developed countries, in
part because of the much lower numbers of procedures performed in developing
countries that involve treatments for suppressing the immune system, such as
organ transplantation (Ota 2004). However, the committee is aware that the
variability of conditions that contribute to immunosuppression in members of a
population and the difficulty of estimation of the prevalence of such conditions
clearly hinder long-term estimation of risk. A final caveat in interpreting these
studies is that in many cases the species of Fusarium was not determined and in
others the accuracy of the identification, which was almost always based only on
structure, could be called into question. In addition, the phenotype of the poten-
tial mycoherbicide strains of Fusarium appears to be largely uncharacterized.
Although a few cases of infection with Fusarium have been reported in
animals (for example, see Mayayo et al. 1999; Conkova et al. 2003; Evans et al.
2004; Marangon et al. 2009), most of the reports deal with cases of proven or
presumed Fusarium toxicosis.
OCR for page 98
98 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
An additional factor that should be considered in an evaluation of potential
effects of Fusarium on humans is the paucity of effective antifungal drugs avail-
able to treat systemic fusariosis (Azor et al. 2009). In a recent review, Lortholary
et al. (2010) reported an overall rate of success of voriconazole treatment for
systemic fusariosis of 47%, with a range of 0-64%, depending on the site of in-
fection. Combination therapy with amphotericin B did not improve overall suc-
cess. The median survival rate of patients infected with F. oxysporum was 112
days. That is perhaps not surprising, in light of the poor in vitro response of F.
oxysporum isolates to antifungal drugs. Although the efficacy of voriconazole is
low, it compares favorably with the administration of amphotericin B and its
lipid derivatives (Pfaller et al. 2002). Voriconazole and the other mold-active
azoles posaconazole and ravuconazole are likely to be the mainstays of therapy
for systemic fusariosis for now (Stanzani et al. 2007).
MUTATION
The potential of fungi used as mycoherbicides to mutate is similar to that
of other fungi in general. Fungal genetic variation can be affected by many
processes, including nucleotide substitution (Kasuga et al. 2002), gene loss
(Sharpton et al. 2009), gene gain by duplication (Sharpton et al. 2009), and gene
gain by horizontal gene transfer (that is, not from parent to offspring) from
closely related (Neafsey et al. 2010) or distantly related (Friesen et al. 2006)
fungi. The gain of genetic material can be as small as a few genes or as large as
complete chromosomes (Ma et al. 2010). Genetic variation can become estab-
lished, or fixed, in fungal populations by natural selection or by chance. Natural
selection leads to adaptation to changing environments, and the adaptation could
include the ability to attack new cultivars of existing host species or new host
species. Adaptation may occur in fungi that reproduce sexually or asexually, but
genetic recombination speeds the process by enabling the rapid synthesis of nu-
merous genotypes (Goddard et al. 2005). Recombination is accomplished by
mating and meiosis but also could occur by nonsexual anastomosis and nonmei-
otic recombination via the parasexual cycle (even though many fungi have
mechanisms to limit nonsexual hyphal anastomosis to partners with almost iden-
tical genotypes). All those evolutionary processes apply to fungi used as myco-
herbicides.
Strains of Fusarium have a reputation for being highly mutable in culture,
especially if the strains are maintained for a long period on a medium high in
simple sugars (Puhalla 1981; Leslie and Summerell 2006). Mass production of
Fusarium would involve such long-term maintenance of the strains. Mutations
that occur under laboratory conditions usually are of a loss-of-function nature,
such as the inability to synthesize a toxin, the inability to produce one or more
types of spores, abnormal hyphal structure, or loss of plant pathogenic capabili-
ties. Such mutations are probably due to the inactivation of one or more func-
tional genes that are already in the strain’s genome. Under field conditions, mu-
OCR for page 99
99
Candidate Biological Control Agents against Cannabis and Coca
tations would certainly occur, although there is no reason to think that they
would be more likely in these fungi than in indigenous strains or any other
fungi.
Although no sexual stage has been reported for F. oxysporum, it is not
possible to rule out its existence. For example, there are two mating types in
sexual Fusarium, MAT1-1 and MAT1-2, and both are found in the strains of F.
oxysporum that have been evaluated (Enya et al. 2008; Lievens et al. 2009).
Therefore, it is possible that adaptation in F. oxysporum could be aided by re-
combination if an as yet undetected sexual stage is occurring.
It is possible that mutation could appear to be driven in F. oxysporum by
selective conditions. For example, some conditions might favor a particular
strain among the very large pre-existing populations of F. oxysporum in the soil.
F. oxysporum is ubiquitous in soils where crops are grown, and its populations
would harbor substantial genetic variation. As a consequence, the sudden ap-
pearance of such a strain could be attributed to mutation even though the strain
had existed at low levels for an extended period. Alternatively, phenotypic
changes could result from a mutagenic mechanism, such as the one responsible
for the spontaneous generation of nit mutants, which are rendered insensitive to
chlorate toxicity when the strains are cultured in the presence of high concentra-
tions of chlorate (Klittich and Leslie 1988). The mechanism underlying this
mutagenic process has not been characterized, but it appears to have a multi-
genic base (Klittich et al 1988). In both those cases, mutations appear to be di-
rected by the selective conditions under which they occur.
Mutations could make mycoherbicide strains more virulent and thus make
a host plant more susceptible to them. It is equally likely that the host plant
would produce offspring that could be more resistant to the mycoherbicide.
Changes in susceptibility and pathogenicity are expected as the host-pathogen
system evolves. Selection of plants resistant to mycoherbicide strains should be
relatively straightforward; the more virulent the mycoherbicide strain is and the
fewer genetically susceptible plants escape it by chance, the more rapidly selec-
tion should occur.
Mutations could affect mycotoxin production, provided that a mycoherbi-
cide has the genes for toxin production. Mutations also could alter the synthe-
sized toxin by blocking its biosynthetic pathway; this would lead to the accumu-
lation of an intermediate in the pathway rather than of the end product. No on
toxin production in the proposed F. oxysporum mycoherbicides are available.
WHAT WE CAN LEARN FROM A NATURAL
EPIDEMIC OF FUSARIUM OXYSPORUM F.SP.
ERYTHROXYLI WILT OF COCA IN PERU
The natural occurrence of diseases is common in plant populations, and
plants that are intensively grown as crops are particularly vulnerable to periodic
devastating epidemics. Recent history tells us that coca populations are no ex-
OCR for page 100
100 Feasibility of Using Mycoherbicides for Controlling Illicit Drug Crops
ceptions. For example, an epidemic on coca in Peru in the late 1980s was linked
to F. oxysporum f.sp. erythroxyli (Arévalo et al. 2000). That natural epidemic
can be viewed to some extent as a large-scale field trial that is informative in
considering several questions: Did the epidemic inflict sufficient damage to re-
duce crop production? Did it cause recurrent or permanent effects on coca in the
affected areas? If the effects of this natural epidemic are an indication, how ef-
fective would it be if the same pathogen were applied as a mycoherbicide to
curtail the production of illicit-drug crops?
A wilt of coca (E. coca) caused by F. oxysporum resulted in extensive
losses to the coca crop in the Huallaga Valley, the main coca-producing region
in Peru (Arévalo et al. 2000). Although definitive reports of the presence of the
disease go back at least as far as 1932, its incidence increased sharply in the
1980s, at the time of the increase in coca production, the subsequent increase in
use of agricultural chemicals, and the reduction in cultivation time (Nelson et al.
1997). The pathogen was reported as F. oxysporum f.sp. erythroxyli (Nelson et
al. 1997; Arévalo et al. 2000) and one of its strains was found to be identical to a
strain of the pathogen studied in Hawaii by Sands et al. (1997) and Nelson et al.
(1997).
The most recent epidemic of F. oxysporum f.sp. erythroxyli wilt of coca in
the Huallaga Valley of Peru began in 1987. It was estimated that 52% of the
coca crops were affected during 1992-1994. To describe the impacts of the dis-
ease, Arévalo et al. (2000) quantified its incidence and development in 11 coca
fields in five regions in the valley. The coca plants sampled were 14-93 months
old. Overall, the coca leaf yields were reduced by 74%. Disease incidence (the
proportion of sampled plants that were diseased) ranged from 54% of 32-month-
old plants to 79% of 93-month-old plants. The level of disease, measured as the
area under the disease-progress curve, ranged from 5.67 on 39-month-old plants
to 28.2 on 93-month-old plants (that is, the older plants were more diseased).
Thus, the Peruvian F. oxysporum f.sp. erythroxyli epidemics of the 1980s and
1990s were highly devastating to coca production in the Huallaga Valley.
No formal followup studies of the disease have been performed, but anec-
dotal observations suggest that the coca wilt disease persists and affects coca
leaf production in some portions of the valley. In some areas, however, cultiva-
tion of coca has been moved to new fields to escape the disease (personal com-
munication, E. Arévalo, Instituto de Cultivos Tropicales, November 16, 2010, to
B. Bailey, USDA). Hence, the epidemic, even while inflicting severe losses, has
not deterred coca production, which has continued even in the presence of spo-
radic disease outbreaks.
