Introduction

Several herbivorous mites (Acari) damage citrus in Israel. Foremost amongst them is the citrus rust mite, Phyllocoptruta oleivora (Ashmead) (Eriophyidae), whose feeding causes much damage due to the extensive brown, rusty blemishes that develop on the fruit (Argov et al. 2002). The economic importance of this pest has greatly increased during the past decade in Israel, probably as a result of acaricide misuse in citrus groves (Palevsky et al. 2000). Other citrus mites, all assigned to the family Tetranychidae (spider mites), include two major pests, the citrus red mite, Panonychus citri (McGregor) and the oriental spider mite, Eutetranychus orientalis (Klein), and two minor citrus pests, the two-spotted spider mite, Tetranychus urticae (Koch) and the carmine spider mite, Tetranychus cinnabarinus (Boisduval) (Gerson 2003). Interest in the biological control of mites with acaropathogenic fungi has recently increased (Sztejnberg et al. 1997; Chandler et al. 2000; Roy and Pell 2000) and the present essay is intended to contribute in that area.

We recently discovered several indigenous fungi that were assigned to the Exobasidiomycetidae of the class Ustilaginomycetes (Basidiomycota), and were described as new taxa by Boekhout et al. (2003). They were obtained from the cadavers of pestiferous mites, and one of them, Meira geulakonigii Boekhout, Gerson, Scorzetti & Sztejnberg, was assayed against P. citri (75.3% corrected mortality), P. oleivora (95.4%) and T. cinnabarinus (79.4% corrected mortality) (Sztejnberg et al., 2004). The aims of the present study were (1) to test two other fungal species, namely Meira argovae Boekhout, Gerson, Scorzetti & Sztejnberg and Acaromyces ingoldii Boekhout, Gerson, Scorzetti & Sztejnberg, as antagonists of mites. We aimed at comparing their killing rates with those of M. geulakonigii when assayed against the same citrus mites in the laboratory, and (2) to examine the option of applying one of these fungi (M. geulakonigii) in a citrus orchard against P. oleivora.

Material and methods

Sources of mites

All mites were collected in the coastal plain of Israel. Phyllocoptruta oleivora was obtained from red grapefruit (Citrus paradisi) at Kibbutz Hatzor. Panonychus citri was collected from silverberry (Elaeagnus pungens), and E. orientalis from frangipani (Plumeria rubra), both in Rehovot. Tetranychus cinnabarinus was also from Rehovot, off castor bean (Ricinus communis), and T. urticae was obtained from magic flower (Achimenes sp.) in a greenhouse at Misgav, in the north. All five mite species were maintained on two month-old (two-leafed) sour orange (Citrus aurantium) seedlings in the laboratory, at 25°C and under a 12L:12D photoperiodic regime. Light was supplied by two fluorescent lamps (‘Day light’, 11W), providing 50–60 μA.

Source of fungi

The origins of all fungi are detailed in Boekhout et al. (2003). In this study, A. ingoldii (isolate AS001), M. geulakonigii (isolate AS004) and M. argovae (isolate AS005) were used. All isolates were routinely grown in the dark on 3.9% w/v potato dextrose agar [(PDA) (Difco, Detroit)] at 25°C, in Petri dishes (9 cm diameter).

Blastoconidia production and harvest

Durum wheat seeds were used as a substrate for production of blastoconidia for all three fungi (A. ingoldii does not sporulate in liquid media). The seeds were boiled in deionized water for 60 min and after removal of excess water, were placed in batches of 50 g within 250 ml Erlenmeyer flasks and then autoclaved. About 25 ml of sterilized deionized water were then added to each flask, along with three discs (6 mm diameter) of PDA on which each fungus was growing. The wheat and the agar discs were mixed with a sterilized spatula and rotated. The flasks were then kept at 25°C for 1 week, during which they were daily hand-shaken for four consecutive days. To remove blastoconidia, deionized water was added to the flasks and shaken in a vortex. The resultant suspension was forced through four layers of gauze and used. The concentration of the blastoconidia was calibrated using a haematocytometer (Sztejnberg et al. 2004).

For the citrus grove (field) experiment, M. geulakonigii was cultured in 30 ml of 2.4% w/v potato dextrose broth [(PDB) (Difco, Detroit)] for 2 weeks in 100 ml Erlenmeyer flasks at 25°C. The fungus forms only negligent amounts of mycelia in liquid media, which were thus ignored. The broth was centrifuged at 2430g for 15 min using a Sorvall RC 5C centrifuge (Du Pont, Wilmington, DE, USA) at 25°C. The blastoconidia were washed twice, until no PDB residues remained in the suspension. They were then re-suspended in sterile deionized water and used, after a haematocytometer calibration.

Experimental design

For the laboratory bioassays, batches of approximately 30 adults were transferred to new seedlings, using five seedlings for each mite species. After 3–6 days later all mites were removed, leaving only their eggs on the leaves. After 10 days later, the deutonymphs and adults that had meanwhile developed were sprayed with the fungal suspensions. The mite-infested seedling were sprayed with a gas sprayer (Preval® Sprayer, NY, USA) to run off with a 1 × 108 ml−1 suspension of blastoconidia (of each fungal species) in deionized water, without any detergents. Seedlings were kept at 25°C at 12L:12D, RH conditions >95%, either for seven or for 14 days. Each experiment was similarly conduced twice.

Field experiment

The field experiment was conducted in Kibbutz Dan, in the north of Israel, at a 21 years-old commercial organic red grapefruit orchard (75,000 m2), which was naturally infested by P. oleivora. The fungal suspensions were sprayed with a 5 l garden sprayer (Holder, Model forma 5, Metzingen, Germany). Average temperatures and RH were recorded at Kfar Blum station, ca 8 km away. The experiment was a randomized block design and comprised of three treatments, each replicated five times (one per block). The treatments were: a single application of M. geulakonigii at the beginning of each season (July 1st, 2003 and July 15th 2004), monthly sprays (applied every month for the duration of the experiment), an early-solitary treatment in May 15th (only in 2004) and an untreated control. The latter was sprayed only with the same tap water used to dilute the treatment suspensions. Each group of five plants (which made up each treatment) was sprayed with 1 l of a 1 × 108 ml−1 suspension of unformulated blastoconidia in tap water. The pest populations were monitored monthly in 2003, and every 2 weeks in 2004, by taking at random five fruit samples before the application. Two randomly-chosen areas of 1 cm2 each were used to count the pest population on each fruit. The mean population in 1 cm2 was calculated per tree, and is expressed as mites/cm2 fruit. These data were used to calculate the adult cumulative mite-days (ACMD), mite days being defined as one mite present per fruit per day, and calculated as the mean of two successive counts multiplied by the number of intervening days (Palevsky et al. 2004). This value may serve as a parameter for the evaluation and influence of mite density on fruit damage. At the end of the season (after October 7th) fruit from all treatments were collected and the damage assessed in four categories: (1) Clean fruits (non-damaged); (2) 0.1%–25% of the fruit skin russeted (light); (3) Between 25% and 50% of the fruit russeted (medium); (4) More than 50% of the fruit skin damaged (severe). In addition, all fruit were weighed and measured at harvest. This experiment was repeated during two running seasons, in 2003 and in 2004.

Statistical analysis

Every mite species was tested separately in the laboratory trials, and mortality caused by each fungus was thus evaluated on its own. In addition, the effect of each fungus on the mites was compared in two trials, one lasting seven and the other 14 days after the fungal application. All resultant data were arc-sin transformed and analyzed by a two-way analysis of variance (ANOVA), and means were compared with Tukey’s pairwise comparisons test (P = 0.05). For the field trial, differences between means of mite populations were evaluated by Fisher`s least significant differences (LSD) test for each date separately. Damage distribution was compared by a two-way analysis of variance (ANOVA). Each category was compared between treatments, and differences of means were examined with Tukey’s pairwise comparisons test (P = 0.05). ACMD values were compared by a Tukey’s pairwise comparisons test (P = 0.05) between treatments at the same season. The analysis was conducted using JMP. 5.0.1 software (SAS, Drive Cary, NC).

Results

All fungi significantly increased the death rate of all mites under laboratory conditions, except M. argovae against T. urticae (Table 1). All fungal applications against P. oleivora resulted in 82.1%–90.1% mortality after 7 days (significantly different from control), a value similar to that obtained after 14 days. The mortality of P. citri was significantly higher only after 14 days than within 1 week, when M. argovae (58.7%–72.7%), M. geulakonigii (63.5%–86.7%) and A. ingoldii (57.7%–87.5%) were applied. The exposure of E. orientalis to M. argovae, M. geulakonigii or A. ingoldii resulted in similar death rates after 1 week (79.7%, 67.8% and 83.6%, respectively) or 14 days (83.9%, 87.8% and 79.2%). The mortality of T. urticae, exposed to M. argovae, was different from control after 1 week (52.4%) but not significantly different after 14 days (46.1%) whereas M. geulakonigii and A. ingoldii caused a significant death rate 7 and 14 days post-application (81.9%–90.0% respectively). The mortality rates of T. cinnabarinus were significantly higher than in the controls only 14 days after each fungal application, being 61.1% due to M. argovae, 46.8% for M. geulakonigii and 57.8% when exposed to A. ingoldii.

Table 1 Percentage mortalities of Phyllocoptruta oleivora, Panonychus citri, Eutetranychus orientalis, Tetranychus urticae and Tetranychus cinnabarinus, sprayed either with Meira argovae, Meira geulakonigii or with Acaromyces ingoldii (all at 1 × 108 ml− 1 blastoconidia/ml).
Full size table

Only few mites (0.007 /cm2) were found on the grapefruit skin on July 1st 2003, when monitoring of P. oleivora in the field experiment was initiated (Fig. 1A). By the end of that month differences (albeit not significant) began to emerge between the control fruits and the two other treatments (0.4 mites/cm2 vs. 17.1 mites/cm2 on the control fruit). After 4 weeks later the differences became significant (137.6 mites/cm2 on the control fruit vs. 64.35 mites/cm2 on the seasonal-treatment fruit and 56 mites/cm2 on the monthly-treatment fruit); there was no significant difference between the two treatments (Fig. 1A). From September 16th to October 7th, the pest population underwent its natural decline. The average (day and night) RH values recorded in July to September were ∼90%, and the mean temperature was 28°C during July and August, 25°C in September. A similar pattern in mite mortality was observed in the second year (2004), when 0.7 mites/cm2 were monitored at the initiation of this experiment. After 14 days later the mite population increased to 7.8 mites/cm2 vs. 9.4 mites/cm2 on the control fruits. The highest mite population occurred on August 17th, when 198.5 mites/cm2 were counted on the controls, 105.4 mites/cm2 on the seasonal treatment, 76.5 mites/cm2 on the monthly-treatment fruit, and 132.9 mites/cm2 on the early treatment (Fig. 1B). No differences in mite numbers were observed between the early treatment application and the solitary or repetitive treatments in 2004. No differences in ACMD values were obtained amongst all three applications when treatment results were compared within each year (data for the early 2004 application are not shown in Fig. 2).

Fig. 1
figure 1

Population dynamics of Phyllocoptruta oleivora on grapefruit during two seasons: 2003 (A) and 2004 (B) at Kibbutz Dan, as affected by four treatments: Spraying one liter of a suspension of Meira geulakonigii (1 × 108 ml−1 blastoconidia per tree) every 15-30 days (monthly treatment—▲); a similar spray applied only once, in July (seasonal treatment—■); a similar spray applied in May 15th (early in the season) and conducted only in 2004—◆); and control (water only—●). Bar between treatments denotes LSD at P < 0.05

Full size image
Fig. 2
figure 2

Mean adult cumulative mite-days (ACMDs) on grapefruits treated either seasonally or monthly (see Fig. 1 for explanation) during 2003 and 2004. Different capital letters denote significant differences between treatments at P < 0.05 in 2003, and lower case letters, in 2004

Full size image

The extent of mite injury (i.e. russeting) was similar in the two application treatments and both differed significantly from the damage in the control. The percentages of “medium” and “severe” fruit injury were 3.3%-3.6% and 1.4%–2.0%, respectively, in the “one application” and “repetitive applications” treatments, as compared to 8.2% medium and 8.2% severe damage in the control (Fig. 3),

Discussion

The laboratory data reveal a variable pattern of fungal effects on the five target-mites. Each of the tested fungi affected all mites, with the exception of M. argovae being harmless to T. urticae (Table 1). These results imply that there is a degree of selectivity in the effects of these fungi, which may be exploited when using them in combination with other natural enemies.

No reports are at hand about non-parasitic fungi that are antagonistic to mites. Mite-specific fungi, e.g. Hirsutella thompsonii Fisher and Neozygites floridana Weiser and Muma, as well as mite and insect pathogenic fungi (e.g. Beauveria spp. and Metarhizium spp.) penetrate the body of the mite and then secrete toxic compounds (Chandler et al. 2000). As parasitism is the mode of action of neither Meira spp. nor A. ingoldii (Sztejnberg et al. 2004), we assume that mite death was due to toxic secretions. A related fungus, Pseudozyma flocculosa (Traquair, Shaw and Jarvis) Boekhout and Traquair is antagonistic to powdery mildew diseases and to other phytopathogenic fungi, and its mode of action is the secretion of toxic fatty acids (Avis et al. 2001). This suggests that the killing mechanism of Meira spp. may be more potent because these fungi are also antagonists of powdery mildew diseases (Sztejnberg et al. 2004) and other phytopathogenic fungi and bacteria (unpublished data).

Testing the effects of the fungi under field conditions may provide variable results, due to the effects of fluctuating temperatures, low RH and U.V. irradiation. Nevertheless, field data are essential as initial steps towards evaluating the commercial potential of these fungi as biocontrol agents.

Both field application schedules of M. geulakonigii had a comparable effect on the pest (Fig. 1) and on damage (Fig. 3), indicating that a single application, given early in the season, could reduce injury. One method to express the relationship between mite numbers and damage to fruits is to calculate the ACMD values for the whole period. These values incorporate the population size during each day for the entire period of the experiment. The ACMD values of the treated fruits were significantly lower than those of the controls, showing the reduction of pest numbers during the experimental period (Fig. 2).

Fig. 3
figure 3

Fruit damage categories as affected by the seasonal or monthly treatments (see Fig. 1 for explanation) during 2003. “Non damaged” are fruit without any visible russeting, “light” signifies less than 25% of the fruit area damaged, “medium” means that 25%–50% of the fruit face was russeted and “severe” denotes more than 50% of the fruit skin covered with `rust`. Different letters denote significant differences (P < 0.05) between damage categories as a result of the treatments

Full size image

No evidence was obtained that the discussed fungi damage the plants whence their acarine hosts were recovered, nor did the augmentative sprays, seasonal or monthly, reduce fruit weight or size distribution, and there were never any signs of phytotoxicity (data not shown). Thus M. geulakonigii probably lives on various parts of the host-plants throughout the year as a saprophyte, attacking mites as these become numerous enough to constitute pests.

In the laboratory all three fungi reduced the numbers of citrus damaging mites, and in the field M. geulakonigii significantly decreased the numbers of P. oleivora as well as the economic injury caused by this pest. These data clearly indicate the potential of M. geulakonigii to control P. oleivora and possibly other citrus mites. Understanding the ecological interactions between fungi, plant-feeding mites and host-plants could further the use of these antagonists for better pest control.