Potential of two omnivorous iolinid mites as predators of the tomato russet mite, Aculops lycopersici

Two iolinid predatory mites were studied in the laboratory as potential biological control agents of the tomato russet mite, Aculops lycopersici (Acari: Eriophyidae). The development, reproduction and predation capacity of Pronematus ubiquitus and Homeopronematus anconai (Acari: Iolinidae) on A. lycopersici were investigated. Developmental time from egg to adult at 25 °C averaged 9.59, 9.31 and 9.52 days for P. ubiquitus on A. lycopersici, A. lycopersici and Typha angustifolia pollen, and pollen only, respectively. H. anconai required 11.18, 10.39 and 11.90 days to complete development on the respective diets. Survival of the immature stages exceeded 83% on all diets for both iolinids. In experimental arenas, both predators caused a substantial reduction of the immature population development of A. lycopersici. The number of tomato russet mite offspring was reduced by 78% and 57% by P. ubiquitus and H. anconai, respectively. The addition of pollen to the arena lowered this population reduction to 35% and 27% for the respective predators. However, supplementing a diet of tomato russet mites with pollen significantly increased the fecundity of both predatory mites from 14–15 to 24–25 eggs over a five-day period. The results suggest that both P. ubiquitus and H. anconai have good potential to suppress A. lycopersici populations and that T. angustifolia pollen can support population establishment of the predators. Overall, P. ubiquitus developed faster on the three diets tested and suppressed A. lycopersici stronger as compared to H. anconai, with or without pollen as a supplementary food source.


Introduction
Tomato, Solanum lycopersicum L., is globally the second most important solanaceous vegetable crop after potato (Solanum tuberosum L.) (Quinet et al. 2019). The annual worldwide production of tomatoes steadily increases every year and in 2020, approximately 187 million tonnes of tomatoes were produced on a ca. 5 million ha surface area (FAOSTAT 2022). The tomato russet mite, Aculops lycopersici (Tryon) (Acari: Eriophyidae), is an economically important pest of tomatoes in many tomato growing areas for which no commercial biological control solution exists (Vervaet et al. 2021). Feeding of this mite causes massive destruction of the epidermal cell layer, resulting in the characteristic bronzing of the damaged tissue and substantial yield losses (Al-Azzazy et al. 2018;Bailey et al. 1943).
In Europe, management practices are generally limited to pesticide application (e.g. abamectin and sulfur) which is often perceived as ineffective and it moreover interferes with biological control of other greenhouse tomato pests (e.g., whiteflies, spider mites and aphids) (Bailey et al. 1943;Brodeur et al. 1997;Duso et al. 2010). Additionally, due to concerns regarding the adverse effects of chemical pesticides on human health and the environment, there is an increasing demand for pesticide-free vegetables (Gerson et al. 2007;Van Lenteren et al. 2020). Consequently, it is imperative to find effective biological control agents against A. lycopersici.
Although A. lycopersici is a suitable food source for various natural enemies in the laboratory, especially for predatory mites of the Phytoseiidae, they are unable to control the tomato russet mite on tomato crops (van Houten et al. 2013;Vervaet et al. 2021). Tomato plants are covered with trichomes. These protect the plants from a range of herbivores but also hamper or kill natural enemies, including phytoseiid predators (Kant et al. 2015;Paspati et al. 2021Paspati et al. , 2022van Houten et al. 2013). The minute size of the tomato russet mite (150-200 µm) allows it to move between the trichomes that create a shelter against most of its competitors and predators (van Houten et al. 2013).
The iolinid predatory mites Homeopronematus anconai (Baker) and Pronematus ubiquitus (McGregor) (Acari: Iolinidae), however, are considerably smaller (225-280 µm) than phytoseiid mites (Baker 1943;Knop et al. 1983;McGregor 1932). This allows them to navigate under and between the tomato trichomes without being hindered and may thus enable them to contribute to the biological control of the tomato russet mite. Both iolinids are distributed throughout the world and have been observed on citrus (Ueckermann et al. 2007), apple (Bayan 1986), grape (Knop et al. 1983), fig (Abou-Awad et al. 1999) and different solanaceous crops, including tomato (Abou-Awad 1979;Brodeur et al. 1997;Hessein et al. 1986;Kawai et al. 2004). Moreover, they are predators of different eriophyids, including A. lycopersici (Abou-Awad 1979;Hessein et al. 1986;Kawai et al. 2004;Pijnakker et al. 2021;Van Houten et al. 2020). Besides predation on other arthropods, these iolinids display different feeding habits and have been reported to feed on pollen and plant tissue (without causing noticeable damage), fungi and dead arthropods (Flaherty et al. 1971;Hessein et al. 1986;Jeppson et al. 1975;Knop et al. 1983;Pijnakker et al. 2021;Walter et al. 2009). The pollinivory of both iolinids offers potential to pre-establish populations in tomato crops with pollen supplementation and enables them to sustain their populations in a tomato crop when prey is scarce or absent . In a recent greenhouse study, P. ubiquitus was shown to simultaneously control a tomato russet mite infestation and powdery mildew infection (Oidium neolycopersici L.) on potted tomato plants . The simultaneous protection of this mite against pest and pathogen was thought to show potential to solve the interference of fungicides and sulfur applications with biological control programs.
The objective of the present laboratory study was to compare the potential of P. ubiquitus and H. anconai as biological control agents of A. lycopersici. In particular, the developmental and reproductive performance of the two predatory mites was assessed using A. lycopersici, Typha angustifolia pollen, or a combination of both as food.

Cultures and plant materials
All cultures and experimental arenas were maintained in climatically controlled chambers (PHCBI MLR-352H-PE, Japan). Plants of tomato (Solanum lycopersicum., cv. 'Moneymaker') and common bean (Phaseolus vulgaris L. cv. 'Prelude') were grown in an experimental greenhouse at the Department of Plants and Crops, Ghent University. Fresh cattail pollen (Typha angustifolia L., Nutrimite™) was supplied by Biobest N.V. (Westerlo, Belgium) and stored at -18 °C. Before use in the experiments, a small amount of pollen was thawed and kept at 4 °C for one week.
The laboratory colony of P. ubiquitus was initiated with mites collected on angel's trumpet (Brugmansia sp.) in Ternat, Belgium (50.866798 N, 4.159654 E). The colony of H. anconai was established from individuals collected on Viburnum burkwoodii Burkwood and Skipwith and Stephanandra incisa L. at the Research Station for Vegetable Production in Sint-Katelijne-Waver, Belgium (51.078727 N,4.529423 E). The colonies were reared on bean leaves which were placed with the lower side down on a 2 cm layer of water-saturated cotton in a plastic tray. The plastic trays (17.4 × 12.4 × 4.2 cm) were covered with a plastic lid (17.8 × 12.4 × 4.2 cm) having a central hole of 5 cm diameter. Tissue paper strips were placed half on the leaf margins and half on the cotton to provide water to the mites and prevent them from escaping. Once a week, the mites were supplied ad libitum with cattail pollen. The colonies of both iolinids were maintained at 25 ± 0.5 °C, 65 ± 5% relative humidity (RH) and a photoperiod of 16:8 h (L:D).
Tomato russet mites, A. lycopersici, were collected from a natural infestation on tomatoes (S. lycopersicum cv. 'Merlice') grown in a greenhouse at the Research Centre Hoogstraten in Belgium (51.471169 N, 4.798792 E) and were maintained on young tomato plants (Solanum lycopersicum, cv. 'Moneymaker') in an environmental chamber at 27 ± 0.5 °C, 50 ± 5% RH and a photoperiod of 16:8 h (L:D).

Experimental set-up
All experiments were conducted in a controlled environmental chamber at 25 ± 0.5 °C, 65 ± 5% RH and a photoperiod of 16:8 h (L:D). Freshly excised tomato leaf disks (Solanum lycopersicum, cv. 'Moneymaker', 2.1 by 2.1 cm) were used as arenas. The leaves were placed with the upper surface facing up on water-saturated cotton in plastic Petri dishes (ø 5.1 cm) with a mesh-covered hole (ø 1.4 cm) in the lid. The edges were covered with moist pieces of tissue paper to provide free water and prevent the mites from escaping and drowning.

Effects of diet on development and survival of P. ubiquitus and H. anconai
For each predator and diet, 20 to 30 fresh eggs (less than 8 h old) from the stock colony were transferred individually to a tomato leaf arena. The predators were supplied ad libitum with one of the following food sources: (1) cattail pollen, (2) mixed stages of A. lycopersici, and (3) mixed stages of A. lycopersici + cattail pollen. The mixed stages of A. lycopersici were provided by placing a 1-cm section of infected stem from the stock colony on the tomato leaf disk for 24 h. Pollen was placed on the arenas using a fine brush, twice a week. The development and survival of the predators were monitored twice a day. Once the adult stage was reached, the sex of the predators was determined. Whenever the quality of a leaf substrate began to deteriorate, it was replaced with a fresh leaf. A small piece of the leaf disk containing the mite was cut out and transferred to the new leaf disk to avoid injury to a predatory mite during transfer. As soon as the predator had moved to the new disk, the old leaf section was removed.

Fecundity and impact of P. ubiquitus and H. anconai on the population development of A. lycopersici, in the presence or absence of pollen
This experiment assessed the predation and reproduction capacity of P. ubiquitus and H. anconai on A. lycopersici. In addition, the effect of pollen on the performance of the predatory mites was evaluated. Four treatments were set up: P. ubiquitus with or without pollen and H. anconai with or without pollen. Two-days-old female predatory mites were starved for 24 h, i.e., they were kept on a tomato leaf without any additional food, and then confined individually on tomato leaf disks inoculated with fifty tomato russet mite adults. In the treatments with pollen, fresh cattail pollen was provided at the start of the experiment and on the third day. Controls consisted of tomato leaf disks infested with fifty tomato russet mites, with or without pollen and without a predatory mite. For each treatment, 14 to 21 replicates were set up. Over the next five days, the number of tomato russet mite eggs, nymphs and adults was recorded daily, as well as the number of predatory mite eggs.

Statistical analysis
All data were analyzed using RStudio 4.0.2. Data from mites that drowned were not included in the analysis. Normality was visually assessed using normal QQ-plots, as well as using Shapiro-Wilk normality tests. The data for the developmental time were analyzed in hours using a generalized linear model (GLM). A quasi-Poisson distribution was used instead of a Poisson distribution to correct for overdispersion. The significance of interactions and factors was assessed by removing them from the model using the ANOVA function of R until a minimum adequate model was reached (Crawley 2007). In addition, contrasts among treatments were determined with general linear hypothesis testing (glht function from the multcomp package) to perform Tukey HSD tests for post-hoc pairwise comparisons (Hothorn et al. 2008). Survival rates and sex ratios had a binomial distribution and were analyzed using a GLM with a logit link and a binomial error function. Tukey's post hoc analysis was used to identify significant differences between treatments. The non-normally distributed predation data were analyzed using generalized linear mixed effect models (GLMMs). A negative binomial distribution was used instead of a Poisson distribution to correct for overdispersion. The number of predatory mite eggs and tomato russet mite adults killed after 5 days were compared using a GLM with a quasi-Poisson distribution due to the overdispersion of the data. In addition, contrasts among treatments were determined with general linear hypothesis testing. For all tests, the level of statistical significance was set at p = 0.05.

Effects of diet on development and survival of P. ubiquitus and H. anconai
Both P. ubiquitus and H. anconai have six life stages: egg, larva, three nymphal stages (protonymph, deutonymph, tritonymph) and adult; in between the active stages, a distinct quiescent phase (chrysalis) occurs. Feeding was observed in all active stages. The duration of the different immature stages and the total developmental times on the three tested diets are shown in Table 1; the results of the GLMs conducted are reported in Supplementary Table 1. Total developmental time (egg to adult) was affected by predator species (GLM: F 1,139 = 117.16; p < 0.001), diet (GLM: F 2,137 = 8.12; p < 0.001) and by their interaction (GLM: F 2,135 = 4.90; p = 0.009). Diet had a clear effect on the total developmental time of H. anconai but not on that of P. ubiquitus. The latter completed its development faster than H. anconai on all three diets. P. ubiquitus showed similar egg to adult developmental times of 9.31-9.59 days on the different diets while H. anconai required 10.39-11.90 days. Diet had no significant effect on the duration of the different developmental stages of P. ubiquitus. H. anconai, on the contrary, developed significantly faster on a mixed diet as compared to pollen alone (p < 0.001) but not as compared to A. lycopersici alone (p = 0.120). However, there was no significant difference between the two single diets (p = 0.205). In H. anconai, the developmental times of the deutonymph and the four chrysalis stages were not affected by diet. In contrast, in the larval, protonymph and tritonymph stages, diet significantly affected the developmental time of these stages. The larval stage developed faster on A. lycopersici alone than on pollen alone (p = 0.016) but not as compared to a mixed diet (p = 0.969). The developmental time of the protonymph stage on a mixed diet was shorter than that on pollen alone (p = 0.001) but was similar to that on A. lycopersici alone (p = 0.413). The tritonymph stage was significantly shorter on the mixed diet compared to either pollen (p = 0.045) or tomato russet mites alone (p = 0.001), whereas there was no difference between the single diets (p = 0.892).

Fecundity and impact of P. ubiquitus and H. anconai on the population development of A. lycopersici, in the presence or absence of pollen.
From the first day onwards, the population of tomato russet mites increased quickly in the controls, reaching up to 290 offspring individuals per tomato leaf arena (Fig. 2). Both P. ubiquitus and H. anconai successfully suppressed the population growth of A. lycopersici in the small-scale laboratory arenas, with or without pollen as a supplementary food source. In the absence of pollen (Fig. 2a), the number of nymphs and eggs of A. lycopersici on the fifth day averaged 290 ± 11 (mean ± SE) in the control compared to 64 ± 10 when P. ubiquitus was present. The population of immature tomato russet mites was thus reduced by 78% in the presence of P. ubiquitus. When pollen was available as a supplementary food source (Fig. 2b) the reduction of the immature pest population on day five averaged 35% (236 ± 21 offspring in the control versus 155 ± 21 in the treatment). A similar trend was found for H. anconai. With A. lycopersici as a sole food source (Fig. 2c), H. anconai reduced the prey population from 234 ± 10 in the control to 1 3 110 ± 17 in the treatment (− 57%) on day five. When pollen was supplemented (Fig. 2d), offspring numbers of A. lycopersici after five days averaged 219 ± 19 in the control versus 159 ± 10 in the treatment (− 27%). The addition of pollen significantly lowered the predation of adult tomato russet mites by both iolinids (Fig. 3; GLM, F 1,68 = 70.059; p < 0.001). There was a significant difference in predation of adult A. lycopersici between the two mite species (GLM, F 1,69 = 12.618; p < 0.001). On the fifth day of the experiment, P. ubiquitus showed the highest predation rates on tomato russet mite adults in the absence of pollen, with a total average of 13 ± 2.2 killed adults over a five-day period versus 8 ± 1.1 adults for H. anconai (p = 0.038). When pollen was added, the difference between the two predators disappeared (p = 0.578) and only 2.0 ± 0.3 and 1.0 ± 0.3 tomato russet mite adults were killed after five days by P. ubiquitus and H. anconai, respectively. P. ubiquitus and H. anconai showed similar fecundity on either diets (GLM, F 1,69 = 0.0717; p = 0.7897). However, in both predatory mites fecundity was significantly affected by diet ( Fig. 4; GLM, F 1,69 = 29.968; p < 0.001). On a tomato russet mite diet, total oviposition over five days averaged 14.5 ± 1.7 and 14.4 ± 1.4 eggs for P. ubiquitus and H. anconai, respectively (p > 0.999). A combination of pollen and A. lycopersici averaged oviposition of 24.5 ± 2.3 and 24.4 ± 2.0 eggs over five days for the respective predators (p > 0.999). The addition of T. angustifolia pollen to the diet of A. lycopersici and plant sap (obtained from the leaf disk) substantially increased the reproductive performance of both P. ubiquitus (p = 0.003) and H. anconai (p < 0.001).

Discussion
This laboratory study shows that A. lycopersici is an acceptable prey for both P. ubiquitus and H. anconai. The iolinids were able to attack and kill the tomato russet mites and effectively exploit them as food, allowing complete development. Both predatory mites were observed piercing and sucking the contents of all tomato russet mite stages. They were also able to successfully develop from egg to adult on fresh cattail pollen or on a mixed diet of pollen and tomato russet mites. The duration of the different life stages of P. ubiquitus on the diets tested in the present study is considerably shorter than those reported by Abou-Awad et al. (1999) for this predator when feeding on the fig leaf mite Rhyncaphytoptus ficifoliae (Keifer) or the fig bud mite Eriophyes ficus (Cotte) at 29 °C and 70-80% RH. These authors reported a mean total developmental time of P. ubiquitus females of 19 and 19.5 days on the respective prey species. Under similar climatic conditions (24-25 °C), the total developmental duration of H. anconai in this study was shorter than that reported by Knop et al. (1983), who offered the iolinid cattail pollen on blackberry leaves (13.6 days). In the present study, the developmental time of H. anconai was shorter on a mixed prey-pollen diet than on pollen alone. Hessein et al. (1988) noted a positive influence of a mixed diet of A. lycopersici and Typha latifolia L. pollen on the survival of H. anconai compared to either food alone. A beneficial effect of a mixed diet on life history parameters has been reported for other predatory mites as well (Messelink et al. 2008;Muñoz-Cárdenas et al. 2014;van Rijn et al. 2002). Whereas this beneficial effect of a mixed diet was not observed on immature survival in our experiment, survival of the immature stages exceeded 83% for both mites on all diets.
Both P. ubiquitus and H. anconai effectively suppressed the population increase of the tomato russet mite in the small-scale laboratory arenas, with or without pollen as a supplementary food source (Fig. 2). These results confirm earlier reports of effective control of A. lycopersici by H. anconai (Kawai et al. 2004) and P. ubiquitus ). However, the suppressive effect of P. ubiquitus on the pest's population growth was stronger than that of H. anconai, with or without pollen supplemented. Additionally, our results indicate that the presence of pollen decreases the overall predation of immature and adult tomato russet mites by both predatory mites. Without pollen, a significant reduction of the tomato russet mite population was found from the second day onwards for P. ubiquitus compared to the control treatment, while this effect was only significant from the third day for H. anconai. With pollen available, a significant reduction of the pest population was found only from the fourth day onwards for P. ubiquitus, whereas this effect was further delayed to the fifth day for H. anconai. In the absence of pollen, P. ubiquitus kills more adult tomato russet mites than H. anconai. When pollen is present, this difference between the iolinid species disappears, but only few adults are killed. These findings agree with those of Hessein et al. (1988), who reported that H. anconai consumed less tomato russet mite adults in the presence of Typha latifolia L. pollen.
Although the addition of pollen to the arena lowered the tomato russet mite predation by both iolinids, it substantially increased the fecundity of the predatory mites. The average number of 3-5 eggs/female/day in P. ubiquitus observed over a five-day period in the present study is higher compared to the 1.6-1.9 eggs/female/day reported by Abou-Awad et al. (1999)  angustifolia pollen supplied on tomato leaf arenas. Similarly, Hessein et al. (1988) showed that H. anconai laid a significantly higher number of eggs on A. lycopersici plus cattail pollen and on cattail pollen alone than on a sole diet of A. lycopersici. The numbers of eggs laid by H. anconai reported in the latter study (1-2 eggs/female on day 5) are lower than those in our study (3-5 eggs/female). This difference could be related to the plastic substrate used in Hessein et al. (1988) study. The availability of leaf tissue is key to the fitness of H. anconai (Flaherty et al. 1971;Hessein et al. 1988;Knop et al. 1983) and P. ubiquitus (Vervaet et al., unpublished). According to Knop et al. (1983), H. anconai fails to reproduce on an artificial substrate (plastic or cork) with or without cattail pollen and survival is low. Hessein et al. (1988) reported some reproduction of H. anconai on a plastic substrate when offered A. lycopersici alone or A. lycopersici with cattail pollen, but the availability of leaf tissue clearly enhanced population development. Duarte et al. (2021) reported that P. ubiquitus females taken from the mass-rearing units and allowed to lay eggs for 4 days on a tomato leaf-only diet produced less than 0.5 eggs/female/day. However, this reproductive output may result from their food uptake during the rearing. On the other hand, attempts to rear H. anconai on grape leaf alone (Flaherty et al. 1971;Hessein et al. 1988;Knop et al. 1983) or on tomato leaf alone (Similon 2018) failed. Similarly, P. ubiquitus is unable to develop to the protonymph stage on a tomato leaf alone (Vervaet et al., unpublished). Thus, whereas feeding on leaf tissue alone provides sufficient nutrients to support the survival of both iolinids to some extent, it is insufficient to sustain full development or allow reproduction. Whether leaf tissue provides one or more essential nutrients, moisture, specific microhabitat requirements, ovipositional cues, or a combination of these, remains unclear.

3
An effective biological agent against A. lycopersici should meet several requirements (Vervaet et al. 2021). First, tomato is an unsuitable host plant for many predators because of the abundant (glandular) trichomes. The results obtained in our study confirm earlier reports that both H. anconai and P. ubiquitus are not hindered by these trichomes and perform well on tomato leaf surfaces (Kawai et al. 2004;Pijnakker et al. 2021). Furthermore, a tomato russet mite infestation is difficult to detect and early symptoms are easily missed or misdiagnosed as a nutritional deficiency, plant disease or water stress (Vervaet et al. 2021). The preventive establishment of a predator to create a'standing army' against A. lycopersici offers a solution to this problem . The suitability of pollen as a diet may increase the potential of biological control agents by enabling them to pre-establish well in the crop before the pest is present and to sustain their population when pest densities are low, as well as by making mass-production easier and cheaper McGregor et al. 2020). Our results corroborate earlier reports that pollen is an excellent food source for H. anconai and P. ubiquitus Flaherty et al. 1971;Hessein et al. 1988;Knop et al. 1983). Duarte et al. (2021) reported that pollen needs to be supplemented at least every other week to allow a sufficient population build-up of P. ubiquitus in tomatoes. Whether pollen supplementation needs to be continued even when A. lycopersici is present in the crop remains to be investigated in greenhouse conditions. Based on the findings of the present study, the positive effect of provisioning pollen on the fecundity of both iolinids is expected to compensate the associated decrease in predation on A. lycopersici in the longer run, but this also remains to be investigated under practical field conditions. Finally, compatibility with other management strategies in the greenhouse is key to the success of a biological control agent. Pesticides can negatively interfere with the performance of natural enemies in the crop. Five compounds (dicofol, abamectin, sulfur, cyhexatin and thuringiensin) tested on A. lycopersici proved toxic to H. anconai, but selective doses of abamectin showed the best potential to control A. lycopersici without reducing predator numbers (Royalty et al. 1987). Sulfur, widely used in greenhouse production as a fungicide, was also noted to be highly detrimental to P. ubiquitus (Van Houten et al. 2020). Interestingly, P. ubiquitus was recently found to simultaneously control a tomato russet mite infestation and a powdery mildew infection (Oidium neolycopersici) on potted tomato plants, suggesting that augmentative releases of this iolinid might reduce the dependence on sulfur in tomato greenhouses ). In addition, intraguild predation with other beneficials in the greenhouse ecosystem should be investigated. Given the small size of both iolinids, the direct effects of most of the currently used beneficials in tomato greenhouses are expected to be limited. However, as both iolinids feed on plant sap as well, their performance might be influenced by indirect effects associated with induced plant defenses triggered by other omnivorous natural enemies (e.g. zoophytophagous predators (Pérez-Hedo et al. 2015.
In conclusion, the present laboratory study indicates that the omnivorous predatory mites P. ubiquitus and H. anconai can play a crucial role in the biological control of the tomato russet mite in protected tomato cultivation. Aculops lycopersici, T. angustifolia pollen or a combination of both as a diet are adequate food sources to sustain the development and reproduction of both iolinid mites. Overall, P. ubiquitus developed faster and suppressed the population development of A. lycopersici stronger than H. anconai. Survival and fecundity on the tested diets were similar for both iolinids. Further field studies are required to determine whether augmentative releases of P. ubiquitus or H. anconai can maintain A. lycopersici under economic threshold levels as well as to optimize pollen supplementation to help establish and sustain populations of the iolinids.