The false codling moth, Thaumatotibia leucotreta
(Meyrick) (Lepidoptera: Tortricidae), is one of the most important pests of citrus in Southern Africa [1
]. A range of products have been tested for its control on citrus since 1926 [3
]. These were reviewed by Moore [4
]. However, since that time increased effort has been poured into the development of new technologies. The existing control measures for T. leucotreta
are reviewed by Moore and Hattingh [5
]. One of these is the Cryptophlebia leucotreta
granulovirus (CrleGV) [4
CrleGV was first described by Angelini et al.
]. This isolate was obtained from infected T. leucotreta
larvae from the Ivory Coast. Thaumatotibia leucotreta
used to be known as Cryptophlebia leucotreta
, hence, the name of the virus, but the host genus was changed in the late 1990s [9
]. Angelini and Le Rumeur [10
] stated that CrleGV contamination, if not curtailed, was capable of causing a laboratory-reared T. leucotreta
culture to collapse. Incidentally, they also noted a cypovirus (CPV) infection in the laboratory culture. Another CrleGV isolate was obtained from diseased larvae, which were collected on the Cape Verde Islands [11
]. Whitlock [12
] was interested in the virus-like rods associated with CrleGV, which he isolated from a South African laboratory culture of the insect. A South African isolate was also obtained from larvae in a laboratory culture held by the Hoechst Corporation in Germany [13
]. The South African isolate, the Ivory Coast isolate and the Cape Verde isolate can be clearly distinguished by restriction analysis [13
]. Fritsch and Huber [14
] made reference to biological and biochemical characterization of the three different isolates mentioned. Fragment patterns were determined by restriction enzyme analysis with Eco
HI, and Hind
]. It was, thereby, demonstrated that all three isolates were distinct strains. Jehle et al.
] constructed a restriction fragment map covering almost the entire genome of the Cape Verde isolate of CrleGV. The position of the granulin
gene was identified by cross-hybridization with granulin coding fragments of Cydia pomonella
GV (CpGV) [16
]. The size of the viral genome was determined to be 112.4 kbp [13
]. Its granulin amino acid sequence was compared to that of Autographa californica
nucleopolyhedrovirus (AcNPV) polyhedrin, and other NPVs [17
]. Jehle et al.
] examined the genetic interaction between CrleGV and CpGV co-infecting larvae of T. leucotreta
. In so doing, the genetic interaction of unmodified GVs was examined in vivo
in order to assess possible risks of genetic exchange of modified baculoviruses. This work was based on the discovery that CpGV is cross-infectious for larvae of T. leucotreta
, but is about 1000 times less virulent than the specific GV [19
]. Subsequently, Lange and Jehle [20
] sequenced and analyzed the entire CrleGV genome. The genome contained 110,907 bp and potentially encoded 129 predicted open reading frames (ORFs), 124 of which were similar to other baculovirus ORFs. A baculovirus chitinase gene was identified, but Lange and Jehle [20
] concluded that it is most likely not functional, because its central coding region including the conserved chitinase active site signature was deleted. It was determined that CrleGV is indeed most closely related to CpGV, as revealed by genome order comparisons and phylogenetic analyses. However, the AT content of the CrleGV genome, which is 67.6% and the highest found so far in baculoviruses, differed by 12.8% from the AT content of CpGV. This resulted in a major difference in the codon usage of both viruses and may reflect adaptive selection constraints to their particular hosts.
Consequently, Reiser et al.
] considered T. leucotreta
as a suitable alternate host for mass production of CpGV for biological control purposes. This idea was apparently employed by Hoechst in Germany, but was unsuccessful, as CrleGV soon became the dominant virus in the culture [22
]. This possibility is again being tested by Chambers [23
], with renewed hope of success, due to improved techniques (based on qPCR) for rapid differentiation between the two viruses and, hence, establishment of virus purity [24
Unlike the closely related CpGV, which has been widely tested since 1966 [25
], culminating in the production of at least five commercial formulations [26
], CrleGV was not exploited for the biological control of T. leucotreta
on agricultural crops until 2004. Up to this time, only one record existed of a small-scale field trial with CrleGV, on citrus and Spanish pepper on the Cape Verde Islands [27
In the last 15 years, extensive work on CrleGV has been conducted in South Africa, initially in the laboratory, but subsequently in the field too. Moore [4
] described the discovery and development of a novel South African CrleGV isolate (CrleGV-SA) as a biological control agent for the management of T. leucotreta
in South Africa. The granulovirus was identified from Goedehoop citrus insectary at Citrusdal, Western Cape, South Africa [4
]. The CrleGV-SA isolate was subsequently characterized by Singh et al.
]. Ludewig [29
] attempted to induce a viral epizootic in larvae in a laboratory culture through stressing of the host, but concluded that this was not possible. He further concluded that this may be due to the T. leucotreta
culture being virus-free, as PCR analysis of DNA extracted from asymptomatic larvae, sensitive down to 60 fg (480 genome copies of CrleGV), was unable to detect any CrleGV. However, Opoku-Debrah et al.
] later succeeded in inducing outbreaks of CrleGV in five geographically distinct T. leucotreta
laboratory cultures through overcrowding of larvae.
An artificial diet for the larval host, a rearing technique and a virus production system were developed [4
]. Surface inoculation dose-response and time-response bioassays and detached fruit bioassays were conducted against T. leucotreta
neonate larvae (the only instar that would be exposed to virus in the field) [7
(the concentration required to kill 50% of the test insects) and LC90
(the concentration required to kill 90% of the test insects) values were estimated to be 4.095 × 103
occlusion bodies (OBs)/mL and 1.185 × 105
OBs/mL, respectively. LT50
(time to kill 50% of the test insects) and LT90
(time to kill 90% of the test insects) values were estimated to be 4 days 22 h and 7 days 8 h, respectively, categorising the virus as a fast or type 2 granulovirus [32
]. This was a clear indication that the virus was sufficiently virulent to warrant field trials. Consequently, extensive field trials were conducted [4
], leading to registration of the biopesticide Cryptogran (River Bioscience, South Africa) [6
]. Subsequently, a second CrleGV product, Cryptex (Andermatt Biocontrol, Switzerland) was registered for use against T. leucotreta
in South Africa [34
]. Registration of both products has been expanded to avocadoes and grapes [35
]. Recently a third CrleGV product, Gratham (also a product of Andermatt Biocontrol, Switzerland), with specifications identical to Cryptex, has also been registered in South Africa. Consequently, CrleGV has been used commercially in the field for 10 years.
] genetically and biologically characterized and compared the CrleGV isolates used in Cryptogran and Cryptex. Restriction analysis and partial amplifications of the granulin
genes, as amplicons of 690 bp and 1290 bp, revealed 99% and 98% nucleotide identities, respectively. The heterogeneity of the Cryptogran and Cryptex viral genotypes was further supported by significant differences in their biological activity determined by surface dose-response bioassays with neonate T. leucotreta
larvae. Cryptogran was shown to be significantly more virulent (specifically the LC90
) than Cryptex in dose-response bioassays. However, Opoku-Debrah et al.
] subsequently showed that although this was significantly so in one case (comparing LD50
values of the isolates against neonate larvae from a regionally specific laboratory culture according to a protocol described by Pereira-da-Conceicoa et al.
]), virulence is actually a very specific relationship between host and pathogen. Using seven CrleGV isolates and five T. leucotreta
host populations, it was demonstrated that certain isolates were significantly more or less virulent against certain regionally distinct host populations [37
The first reported case of insects developing resistance to a virus in the field was observed in C. pomonella
, where field populations in Europe developed resistance to a Mexican isolate of CpGV (CpGV-M), after repeated field applications in organic orchards had failed [39
]. In order to be prepared should a similar situation occur with T. leucotreta
in South Africa, Opoku-Debrah et al.
] bioprospected for new CrleGV isolates as possible alternatives to the existing ones used in the commercial formulations. This led to the isolation and genetic characterization of five novel CrleGV isolates. Single restriction endonuclease (REN) analysis of viral DNA and partial sequencing of granulin
genes and multiple alignments of nucleotide sequences were used to demonstrate these differences, leading to a proposal for two phylogenetic CrleGV-SA groups [30
To date, 13 years of field trials with CrleGV have been conducted on citrus in South Africa. This amounts to well over 50 distinct field trials. This period includes 10 years of commercial field usage of CrleGV products (initially on citrus but also avocadoes and grapes), hence the title of this paper. Differentiation has been made between the early developmental work (with unformulated CrleGV) and trials with commercial products, due to the formulated preparations of the latter (which are proprietary). Apart from internal reports, theses [4
] and one semi-popular paper [6
], these trials have not previously been published.
Consequently, our primary objective in this paper is to report on a comprehensive, large and representative sample of these trials conducted on citrus in South Africa. This is therefore the first published account of the field use and efficacy of CrleGV and should be of great value to scientists and biocontrol practitioners throughout the region of distribution of T. leucotreta. Additionally, we have provided a mini-review in this introduction of all known studies to date conducted on CrleGV.
is an important pest in the Southern African citrus industry [1
]. It is extremely important to control it effectively, particularly due to its endemism to Africa [4
] and the exporting of around 70% of South Africa’s fresh citrus to foreign markets [46
]. CrleGV-based biopesticides, such as Cryptogran and Cryptex, have proven to be effective tools for aiding in suppressing and controlling this cryptic pest.
Although a single-tree randomized block trial layout lends itself to more accurate and reliable comparison of treatment efficacy, as both the randomization and the use of a relatively small homogenous area manage for any possible variation very well (which is why the design was so often used in trials), it was shown in the trial conducted at Carden Farm in 2003, that CrleGV treatments applied to blocks of trees will provide far more effective control of T. leucotreta. This is because in a single tree layout, there is very little or no buffer against T. leucotreta pressure from outside of the trial area. Additionally, recolonization of T. leucotreta from adjacent or nearby untreated trees (or trees treated with less effective or ineffective treatments) will occur immediately on breakdown of a treatment. The semi-commercial block format used for testing CrleGV therefore provided a more accurate measurement of the true potential of CrleGV to control T. leucotreta in citrus under commercial conditions.
In four out of the 13 trials presented in this study, such a block format was used. Cryptogran succeeded in reducing T. leucotreta
infestation by between 58% and 92% (average of 72%) in these treatments (considering only those concentrations high enough to give optimal efficacy). One might argue that any level of control less than a percentage which is in the high 90s against a potentially phytosanitary pest is inadequate. However, Moore and Hattingh [5
] point out that it is essential that T. leucotreta
be controlled using an integrated suite of control options. Therefore, the efficacy of a single product application should not be judged in isolation but as part of a whole. For example, if a farmer uses five different control practices against T. leucotreta
(these could for argument’s sake be any of orchard sanitation, parasitoid conservation (or augmentation), CrleGV sprays, mating disruption and a chemical spray) and these very conservatively each provide around 50% control, the combined efficacy would be in the region of 97%. In reality, treatment efficacy would generally be expected to be well above this level for most products and technologies [5
Furthermore, CrleGV is completely harmless to beneficial insects. Grout et al.
] conducted a series of bioassays with Cryptogran field-weathered residues on citrus leaves against four key natural enemies of citrus pests: Chilocorus nigritus
(Fabricius) (Coleoptera: Coccinellidae), Aphytis lingnanensis
Compere (Hymenoptera: Chalcididae), Coccidoxenoides perminutus
(Timberlake) (Hymenoptera: Encyrtidae) and Trichogrammatoidea cryptophlebiae
(Nagaraja) (Hymenoptera: Trichogrammatoidea). These were conducted according to the protocol established and described by Hattingh et al.
]. They concluded that Cryptogran is probably the softest pesticide that they had tested with regard to its toxicity to natural enemies, as its overall impact ratings were below 10% for the natural enemies tested. Consequently, Cryptogran was categorised as “Harmless” against natural enemies considered important in the citrus ecosystem.
This would obviously be in contrast to the chemical alternatives for T. leucotreta
, which would certainly have a far more adverse effect against natural enemies, including those which attack T. leucotreta
, than would CrleGV. In total at least 17 parasitoids of T. leucotreta
have been recorded [1
]. The most important of these is the egg parasitoid, T. cryptophlebiae
]. It can dramatically reduce T. leucotreta
levels in citrus orchards, either by inundative augmentation [50
] or conservation [51
]. If one couples this with the fact that the chemical alternatives tested did not perform better than did CrleGV, one can only conclude that CrleGV is a very attractive option for T. leucotreta
Only one previous record of a field trial with CrleGV against T. leucotreta
exists. This was a small-scale field trial on citrus and Spanish pepper on the Cape Verde Islands [27
]. Concentrations of 108
OBs/mL were used, and only skimmed milk powder and a wetting agent were added to the virus suspensions. Thaumatotibia leucotreta
damage was reduced by 77% in citrus and 65% in Spanish pepper [27
]. Although these concentrations used were extremely high compared to the registered concentrations with Cryptogran and Cryptex (5 × 106
and 6.6 × 105
OBs/mL, respectively), efficacy was not dissimilar to that reported in our studies. This may be because milk powder is not as effective as molasses at enhancing the efficacy of CrleGV in the field [33
In truth, it appears that from trials conducted in this study that compared different dose rates of CrleGV, it may be possible to further reduce the amount of virus applied, without loss of efficacy. For example, in the trial conducted on Bernol Farm in 2004, although the lowest Cryptogran rate (2 mL per 100 L water) was the least effective rate used, the difference in efficacy was not statistically significant. Nevertheless, this may be an indication that the application rate of CrleGV can be dropped to around 2 × 1013
OBs per ha without any immediate loss of efficacy. However, although this reduction in application rate from the registered Cryptogran rate may not reduce immediate efficacy, it may reduce residual efficacy, as breakdown (mainly due to ultraviolet (UV) irradiation) to below the critical minimum level of viable OB density on the tree for optimal efficacy, would then be reached sooner. It is not surprising that a dose-response was not observed in the trial on Bernol Farm, as dose-responses to baculoviruses are not easily observed in the field [52
]. Nevertheless, the large difference in concentration between the four highest rates and the lowest rate was large enough for there to be a discernible difference in efficacy.
It was noted that Cryptogran was consistently more effective than Cryptex. The reason for this consistent difference in efficacy must almost certainly be a result of the difference in concentration of OBs applied, being 7.6 times higher with Cryptogran (based on the registered concentrations of the two products). However, an additional explanation is a possible differential susceptibility of the local population of T. leucotreta
to the isolates of virus present in both commercial products. Most of the field trials were conducted in the Addo region of Sundays River Valley in the Eastern Cape Province. Opoku-Debrah et al.
] demonstrated in laboratory bioassays that the CrleGV isolate in Cryptogran was significantly more virulent to neonate T. leucotreta
larvae from this region than was the CrleGV isolate in Cryptex. Cryptex required an estimated mean of 2.58 OBs per larva to elicit 50% mortality (LD50
) in a given population as opposed to 1.02 OBs required for Cryptogran. LD90
for Cryptex and Cryptogran were 669 and 273 OBs per larva, respectively [37
Despite a number of general trends being observed in these field trials, it would be prudent to conduct a meta-analysis in order to confirm patterns [53
]. However, this would be superfluous with the relatively small number field trials reported here. A meta-analysis should be conducted on the full complement of more than 50 field trials and therefore warrants a separate study.
Considering the dearth of other CrleGV field trials, it is interesting to compare our results and experiences with those of the closely related system of the Cydia pomonella
granulovirus (CpGV) against the codling moth, Cydia pomonella
(Lepidoptera: Tortricidae), in apples (i.e.
, both cryptic tree fruit pests from the moth family, Tortricidae). Extensive field studies have been conducted with this system since 1966 [25
]. Lacey et al.
] listed numerous different field studies with CpGV, conducted on all continents of the world. There are therefore many examples that can be quoted.
Huber and Dickler [54
] tested CpGV in a commercial apple orchard for two years and compared it to organophosphate insecticides. They were able to achieve a 44%–85% reduction in injury to apples as opposed to a 72%–98% reduction with the use of chemical applications. Later studies by Jaques et al.
] showed that the use of CpGV could reduce C. pomonella
deep-entry damage to apples by 40%–83% compared to the respective control plots. In some of their trial data the protection of fruit by CpGV unexpectedly exceeded that of an organophosphate insecticide. Sheppard and Stairs [56
] tested a range of doses from 107
OBs/tree. All the doses tested had a similar effect on the reduction of infestation but it was found that with the higher dosages there was a larger reduction in larval population as they entered the fruit. Falcon et al.
] reported a 90% reduction in shallow entries. Stará and Kocourek [57
] tested various concentrations of CpGV ranging from 0.5 to 1 × 1013
OBs/ha, as well as varying numbers of applications per season. They succeeded in reducing the C. pomonella
population by 75%–96% compared to 91%–97% achieved with teflubenzuron. Arthurs et al.
] tested three concentrations of CpGV against high C. pomonella
populations, resulting in 81%–99% larval mortality in fruit and a reduced number of mature larvae collected in tree bands by 54%–98%. However, these studies showed that CpGV was more effective at reducing the C. pomonella
population density than reducing fruit injury. Glen and Clark [59
] found that different treatments of CpGV did not significantly affect the survival of the neonate larvae before they entered the fruit. In their first trials, 49% of larvae survived long enough to cause recognizable damage to the fruit. In a subsequent experiment 69% of larvae produced damage to the fruit irrespective of the treatment applied. However, it was noted that the neonate larvae usually died shortly after entering treated fruit. This highlighted a potential shortcoming of CpGV, namely its speed of kill.
Efficacy recorded in our trials with CrleGV against T. leucotreta
, fell within the range reported for trials with CpGV against C. pomonella
. However, unlike CpGV (against C. pomonella
on apples), speed of kill does not appear to be a shortcoming with CrleGV (against T. leucotreta
on citrus). Negligibly few dead (virus infected) larvae were found in fruit that had been treated with CrleGV. A first instar larva takes approximately four days to penetrate through the rind and albedo of a citrus fruit [7
]. If the larva dies or if its behaviour changes (and it reverses out of the fruit) before it manages to penetrate through the albedo into the flesh of the fruit, the damage to the fruit may be insignificant, meaning that the fruit will not decay and the minute blemish on the rind will not downgrade the fruit for export [7
], unlike an apple. This behaviour is typical of symptomatically baculovirus-infected lepidopteran larvae [32
Another drawback with CpGV in the field appears to be its rapid breakdown due to UV degradation. Half-life of CpGV in the field is generally estimated to be between two and three days [60
]. Glen and Payne [66
] showed that although CpGV infectivity was reduced by half in three days, some activity persisted for as long as four to eight weeks after spraying. Arthurs and Lacey [52
] reported that early season applications of label rates of three CpGV products remained highly effective for the first 24 h (producing 94% larval mortality) and moderately effective after 72 h (71% mortality), declining to 50% of its original value after eight days (early summer) during dry sunny conditions. However, some activity remained for up to 14 days, suggesting prolonged survival of the virus in UV-protected locations, such as the calyx of fruit. The decline to 50% activity was more rapid (four days) in mid-summer. Consequently, the recommended application intervals for CpGV against C. pomonella
range from 7 to 14 days [52
As with the slow speed of kill, so too does it appear that rapid breakdown of virus is not a problem with CrleGV on citrus, as it is with CpGV on apples. Particularly the trial conducted at Carden Farm in 2003 demonstrated efficacy of almost 70% recorded at 17 weeks after application. This was a minor decline in efficacy from the 81% recorded at three weeks after application. Fritsch and Huber [69
] estimated the half-life of CrleGV in the field to be two to three days, therefore similar to CpGV. However, Moore [4
] demonstrated that although CrleGV appeared to break down to less than 50% of its original activity within 3–6 days on the northern (sunny) sides of citrus trees, at 21 days after application, efficacy had not yet dropped to this level on the southern (shady) sides of trees. More recently, Mwanza [70
] confirmed this phenomenon, in an attempt to determine CrleGV reapplication frequency required in the field. He established that at 21 days after application to citrus trees in the field, LD50
of CrleGV (against neonate T. leucotreata
larvae) recovered from the northern sides of trees was 15 times higher than from the southern sides of trees. By 28 days after application, virulence of CrleGV on the northern sides of trees was indeterminable, whereas on the southern sides of trees, there was still a clear dose response.
Moore et al.
] surmised that there are four reasons for the protracted CrleGV persistence recorded on citrus. Firstly, a citrus tree provides substantial shading and therefore protection of virus against UV inactivation–more so than probably any other crops on which viruses have been tested for pest control. Secondly, it has been observed that during most of the growing season, the vast majority of T. leucotreta
larvae penetrate a Navel orange through its navel end. It is precisely here that CrleGV could be well protected against sunlight and possibly even rainfall. Thirdly, T. leucotreta
takes a long time to recolonise an area, even once the efficacy of a spray might have expired. This slow migration is confirmed by Timm et al.
] and Stotter et al.
]. Lastly, as CrleGV would have little, if any, detrimental impact on the highly effective and naturally occurring egg parasitoid, T. cryptophlebiae
, this biocontrol agent could aid in maintaining control of T. leucotreta
once virus was no longer effective.
Despite all of these positives, there are a number of challenges that may occur and should be addressed in future research. The risk of development of resistance by the target pest to CrleGV has been mentioned. This concern is based on the experiences with CpGV and C. pomonella
in Europe [39
]. However, as CpGV is recommended to be applied every 7 to 14 days [45
] and CrleGV is applied far less frequently, the risk of resistance development must surely be less. Nevertheless, the study initiated on identification of novel isolates [30
] should be continued and expanded. The potential for resistance can be tested in the laboratory by inducing resistance under selection pressure in subsequent generations, such as was achieved with Phthorimaea operculella
(Zeller) to PhopGV [72
] and Anticarsia gemmatalis
(Hubner) to AgMNPV [75
]. The ability of novel CrleGV isolates to overcome resistance can then be tested against these resistant individuals in laboratory assays. The genetic basis for this ability to overcome resistance should then be determined. For example, it has been ascertained that the viral gene pe38 is not only essential for the infectivity of CpGV but it is also the key factor in overcoming CpGV resistance in codling moth [76
As Opoku-Debrah et al.
] has already determined that certain CrleGV isolates are significantly more virulent than others against laboratory cultures of certain regionally distinct T. leucotreta
populations, this study should be extended to the field (using isolates at equivalent dose rates) to determine if these differences do indeed translate into practice–something which we may already have observed with the differences in efficacy between the two main commercial preparations in the Eastern Cape Province. This may lead to the development of regionally appropriate commercial preparations of CrleGV. This is a possibility that should also be investigated for other baculovirus-host systems. For example, similar differences have been recorded in the laboratory for both virulence of different CpGV genomes [78
] and susceptibility of different C. pomonella
Another challenge that warrants attention is that of UV protection. Although it has been stated that protection from UV by the architecture of a citrus tree is superior to that of an apple tree, there must be exceptions. For example, a young small tree will be far sparser than a mature tree and will thus provide less shading. Additionally, cultivars other than Navel oranges do not possess a navel end in which OBs can be protected against direct sunlight and where T. leucotreta
larvae will preferentially penetrate. Although numerous studies have demonstrated significant protection of baculoviruses under laboratory conditions (e.g., [80
]), there is as yet insufficient evidence that this makes a substantial difference in the field under commercial practices (e.g., [85
]). Consequently, examination of these published formulations with CrleGV, all the way up to full field trials is justified. Additionally, it can be assumed that effective commercial formulations will be kept proprietary. Consequently, sophisticated research on novel and effective formulations should be conducted outside of the commercial sector in order that this information can be made available to scientists and practitioners in the field.