Introduction

After wheat, rice (Oryza spp.) is the most used cereal in human nutrition worldwide. The best-known species of this crop is Oryza sativa L., 1753 (^{Degiovanni et al., 2010}). Among the most important pests of rice is the Steneotarsonemus spinki^{Smiley, 1967} mite, which is commonly known as white mite or panicle rice mite. The reductions in yield attributed to this species have been estimated from 30 to 90% in China, and from 20 to 60% in Taiwan (^{Chen et al., 1980}; ^{Zhang et al., 1995}). In the states of Campeche, Tabasco and Veracruz, Mexico, the surface of cultivated rice suffered a gradual reduction in the last ten years, down to 45%, attributed, at least in part, to the infestation of S. spinki (^{Salazar-Santiago, 2017}).

This mite feeds by piercing the epidermal cells of the host plant, using its stylets to suck out the cell contents, resulting in the formation of brown spots with necrotic regions on the surface of leaf sheaths, panicles, and grains, as well as empty grains. It causes mechanical damage to the adaxial surface of the leaf sheaths, directly affects the rachis of the panicles, and influences the nutrient circulation mechanisms (^{Santos et al., 2004}).

Although S. spinki was originally described with specimens collected in Louisiana, USA (^{Smiley, 1967}), it appeared for the first time as an important pest in the Americas in Cuba in 1997 (^{Ramos & Rodríguez, 1998}), and in the following years it spread throughout several Latin American countries (^{Navia et al., 2010}). In Mexico, this mite was first detected in 2007 (^{Hummel et al., 2009}) and has spread to wide areas along the coast of the Gulf of Mexico (^{Salazar-Santiago et al., 2019}). Lack of proper management could cause this mite to spread to other rice production areas if it has not already spread to them.

In Cuba, this mite was observed to carry the spores of the Sarocladium oryzae (Sawada, 1976) fungus (^{Cabrera et al., 2005}). As a result of infestation by S. spinki, infection by S. oryzae, or the combined attack by both species, the panicles present malformed, stained grains with a significant percentage of empty grains (^{Santos et al., 2004}). However, ^{Salazar-Santiago et al. (2019)} did not find S. oryzae in rice fields in the state of Tabasco, Mexico, but they did observe symptoms such as those described above. They also did not find significant infections of other pathogenic fungal species in rice. With the previous results, these authors concluded that the emptying and staining of the grains are caused, at least in the study area, by S. spinki and not by its association with any pathogenic fungus.

Steneotarsonemus spinki is an extremely small mite; adult females measure approximately 274 x 108 μm, while adult males measure approximately 217 x 121 μm (^{Smiley, 1967}), making them difficult to detect. Apart from the above, this mite hides deeply under the leaf sheaths and panicles, which makes its control difficult (^{Jaimez-Ruiz et al., 2015}). Nevertheless, natural enemies have been detected that could be useful for its biological control, including mites from the families Ascidae, Laelapidae, and Phytoseiidae (^{Lo & Ho, 1984}; ^{Hummel et al., 2009}; ^{Quirós-McIntire & Rodríguez, 2010}). Likewise, various species of entomopathogenic (more properly, acaropathogenic) fungi have been detected infecting S. spinki and creating important epizootics (^{Hummel et al., 2009}). In rice fields on the coast of the Gulf of Mexico, ^{Salazar-Santiago (2017)} collected numerous isolates of S. spinki -pathogenic fungi. The species identified were all in the genus Hirsutella.

On the other hand, the Colegio de Postgraduados (Campus Campeche), in Mexico, has a collection of the generalist entomopathogenic fungi of the Metarhizium anisopliae (Metschnikoff, 1976) and Isaria fumosorosea (Wize, 1957). The M. anisopliae strain is widely used against rice pest insects under the trade name Ma005®; however, its effectiveness against S. spinki has not been evaluated.

The present work was developed with the objectives of identifying the predatory and pathogenic organisms of S. spinki in the most important rice-growing areas of the Coast of the Gulf of Mexico, which are in the states of Campeche, Tabasco, and Veracruz, and to evaluate the potential of select species of predators and pathogenic fungi to control S. spinki.

Materials and methods

Collection and identification of predatory mites associated with S. spinki. Collections were carried out in rice fields in the states of Campeche, Tabasco, and Veracruz, Mexico, during the autumn 2019 crop cycle. The data obtained in that crop cycle were complemented with collections made between 2014 and 2022. The study sites are shown in Table 1. The collected mites were from families that include species recognized as predators of other mites. The collections were always carried out where the mentioned mites were associated with S. spinki or at least that they were found in the same microhabitat. The mites were captured with a fine brush, using a stereoscopic microscope (10 to 40x).

The mites were preserved in 70% alcohol and mounted between slides and coverslips in Hoyer's liquid (^{Krantz & Walter, 2009}) for subsequent observation in the phase contrast microscope and identification with the support of the relevant literature, appropriate for each cluster. The bibliographic sources used to identify each species are listed in Table 2. With additional specimens of each predatory species found, an attempt was made to found colonies to test their potential as natural enemies of S. spinki, for which the activities described below in detail were developed.

Collection and identification of pathogenic fungi of S. spinki in rice crops in the study area. During visits to rice fields aimed at collecting predators of S. spinki, mites of this species were carefully observed in search of specimens infected by fungi, which could look like mummies, with legs in an almost natural position and changes in the coloration, or mycelium that sprang from their bodies and spread radially. Mites suspected to be infected were collected along with a small portion of the plant tissue on which they were perched, placed in Eppendorf tubes, and stored in a cooler for transport to the laboratory.

Once in the laboratory, a small portion of mycelium or the entire mite was taken and seeded in H medium, developed by ^{McCoy et al. (1972)} and modified by ^{Cabrera et al. (2006)}, and the seeded specimens were incubated at 25°C, in the dark. They were checked daily to see if any opportunistic fungi were growing. When the growing fungi had the expected appearance for an acaropathogen, such as Hirsutella sp., they were reseeded in Petri dishes, 10 cm diameter, with H medium.

Pathogenic fungi of the genus Hirsutella were incorporated into the present study from a collection previously formed by M.A. Salazar-Santiago, collaborator in this study, with isolates taken from S. spinki in the state of Tabasco in 2015-2016. The identification of those isolates was carried out macroscopically and microscopically, as well as by amplification of the internal transcribed spacer of ribosomal DNA; they were preserved in 15% glycerin, at -70°C (^{Salazar-Santiago, 2017}).

Assessment of predatory mites and acaropathogenic fungi for the biological control of S. spinki. A method was developed to establish colonies of each species of predatory mite, aimed at having abundant specimens to test their effectiveness for the biological control of S. spinki. To do this, colonies of Tyrophagus sp. mites were initially bred to use them as a food source for the predators, as follows. Approximately 500 g of moist, organic matter-rich soil was collected from under palm trees (Phoenix canariensis H. Wildpret, 1882) in Texcoco, Mexico (19.465390, -99.909071). A search for mites was carried out with a stereoscopic microscope, aimed at locating species of the Acaridae family. Some specimens of these mites were mounted between slides and coverslips with Hoyer's liquid (^{Krantz & Walter, 2009}) and were identified as Tyrophagus sp. by coincidence with the genus diagnosis (^{Fan & Zhang, 2007}).

To provide food for the Tyrophagus sp. mites, pet food kibbles (Ganador Pedigree Purina®, Mexico) were moistened and left in a plastic container with an airtight seal to encourage fungal growth. Incubation temperature was estimated between 22 and 26°C; inside the containers, HR was 100%.

Observation arenas were made with plastic containers with lids, 5 cm in diameter by 2 cm high, with three 1 cm diameter holes covered with mesh (25 μm opening) to allow air entry but prevent mites from escaping. Four kibbles were placed in each arena, two that were in their original presentation and two that were invaded by saprophytic fungi. Next, Tyrophagus sp. mites were transferred to the arenas using a fine brush. Each of these arenas was checked daily to see if the mites were still alive and if they were feeding. Likewise, two drops of water were placed on one of the kibbles to keep them moist, and they were kept at room temperature (between 22 and 26 °C). When the colonies began to grow, they were moved to a new container, which had the bottom completely covered with kibbles to allow the mite populations to continue to thrive. The mites were checked every third day and incubated at 25 ±1 °C and relative humidity of 50-55%. Kibbles were added to the arenas once a week to promote fungal growth.

To breed the predatory mites, hermetically sealed translucent plastic containers (8 cm in diameter, 5 cm in height) were prepared with a hole 2.5 cm in diameter in the lid, to which a mesh was glued to prevent the mites from escaping and allow the aeration. A plastic capsule (1.5 cm in diameter, 1 cm in height) with water-saturated cotton was placed inside each container to maintain high humidity and offer it as a drink for the mites. Then, kibbles that were highly infested with the mites that would be provided as food (Tyrophagus sp.) were placed in the containers. From rice plants infested by S. spinki and depending on their availability, between 10 and 13 mites, preliminarily determined as predators, were transferred with a fine brush moistened with water into each of the containers, each mite was placed directly on the kibbles with abundant Tyrophagus mites. These containers were checked every 24 hours to see if these predators were still alive, in addition to moistening the cotton. They were regularly provided with kibbles infested with Tyrophagus sp. as food.

Acquisition of rice plants. Rice plants were collected from a field located in Piedras Negras, Veracruz (18.762508, -96.24183), which were in the initial stage of tillering, approximately 80 days old from sowing. These plants were transferred to pots with soil from the same plot and taken to an incubation chamber at a temperature of 25 ±2 °C with LED lighting and a 14:10 (light: dark) photoperiod. They were watered every third day and fertilized with Peters® (Tricel-20®, Cosmocel, S.A., Mexico) at a dose of 0.25g/L every week. These plants were infested with S. spinki specimens and were intended to obtain enough specimens of this species necessary to evaluate the impact of its natural enemies.

Predation tests in “sandwich” observation arenas. Arenas were made consisting of two 7.5 x 2.5 cm glass slides and an acrylic sheet of the same size with a central hole of 1 cm in diameter, a wet paper napkin, a layer of cotton, a portion of rice leaf sheath with high infestation by S. spinki, slightly larger than the hole in the acrylic sheet, and two paper clips. The sandwich arenas were assembled as illustrated by ^{Overmeer (1985}, in Fig. 2.1.4.1.5).

Once each arena or sandwich was assembled, the S. spinki specimens that had been confined in the hole were counted, separately by developmental stages. In the same hole, an adult female of the predatory mite was transferred with a fine brush, the second slide was quickly placed on the arena in preparation and pressed with the two clips, placed at both ends, to prevent the escape of the mites, both predator and prey. Ten arenas (replicates) were prepared with the only predatory mite that managed to establish a colony, Gaeolaelaps aculeifer (Canestrini, 1884); they were incubated at 25 °C in a plastic box with moist napkins to maintain relative humidity close to 100% and observed daily for 13 days. Ten other arenas were daily prepared to replace those under observation with new specimens of S. spinki as a food source.

To examine the arenas, each one was opened under the stereoscopic microscope, quickly to prevent the predator from escaping. With a wet brush, the mite was transferred to the new arena, in which the mites destined to be its prey had previously been counted. It was then immediately closed. Once the first arena was opened, its contents were observed: number of prey mites in each stage and whether there were eggs from the prey and the predator; the number of preyed mites was calculated as the difference between the number of mites of each stage, before and after the exposure to predators.

Characterization of Hirsutella sp. fungi. Hirsutella isolates that were stored in special tubes (cryovials) at -70°C and that had been obtained by M.A. Salazar-Santiago, collaborator of this study, were activated in Petri dishes with H medium. To do this, small portions of mycelium were sown under aseptic conditions in Petri dishes with H medium, as previously described. All the Petri dishes were incubated in a growth chamber at 25 °C to observe the mycelial growth of the colonies.

Small portions of mycelium were taken from the aforementioned isolates, and microscopic preparations were made between slides and coverslips with lactic acid as mounting medium. They were confirmed to belong to the Hirsutella genus by observing the characteristic phialides (^{Minter & Brady, 1980}).

The identification of the fungi was complemented molecularly by sequencing of the internal transcribed spacer gene (ITS). For this, total DNA was extracted using the method described by ^{Freeman et al. (1993)}, and polymerase chain reaction (PCR) was performed to amplify a segment of the ITS gene using the primers ITS1 and ITS4 (^{White et al., 1990}). The PCR mixture was as follows: the dyNAzyme TM II (Thermo Fisher Scientific, Madrid, Spain) reaction mixture, 10 µL; molecular biology grade water, 4 µL; sense and antisense primers, 0.5 µL of each; and template DNA, 5 µL, for a reaction volume of 20 µL. The PCR was performed in a MyCycler™ (BIORAD Laboratories Inc., Hercules, CA, USA) thermocycler with the following program: 94 °C for 5 min; 30 cycles of 94 °C for 30 s, 56 °C for 30 s, and 72 °C for 1 min; 72 °C for 10 min; 4 °C indefinitely. The PCR products were visualized on a 1.5% agarose gel (IBI Scientific, Dubuque, Iowa, USA) in 1X TAE buffer plus 1 µL of ethidium bromide, subjected to electrophoresis at 90 volts for 30 min. The amplicons were sequenced by Macrogen Inc. The sequences were edited and assembled in the Mega 10.0.5 program and compared with sequences deposited in GenBank using the BLAST program. Multiple alignment was performed with the ClustalW program.

Pathogenicity test. The observation arenas called sandwiches were the same as those used to evaluate predation effectiveness. Entomopathogenic fungi of the species M. anisopliae, I. fumosorosea, and Hirsutella thompsonii Fisher, 1950, were applied to them; the first two were provided by J. Lara-Reyna; the third was isolate M-2 (B-1), selected from those obtained by M.A. Salazar-Santiago during a previous research work (^{Salazar Santiago, 2017}), since it had the most profuse growth of all the isolates.

Suspensions of conidia of M. anisopliae and I. fumosorosea were prepared at a concentration of 10⁶ conidia/mL in a 0.5% Tween® solution; the concentration was estimated using a Neubauer chamber. In the case of H. thompsonii, the culture obtained did not allow the same concentration to be reached, so it was applied at 10³ conidia/mL. The application was done by spraying on a piece of rice leaf sheath infested with variable numbers of S. spinki specimens previously placed in the sandwich arena. An attempt was made to simulate a drip point application, and this arena was immediately closed with a slide, as previously described. The experiment was carried out with 10 replicates (sandwich arenas) per fungal isolate, including a control that was treated with a 0.5% Tween solution, in a completely randomized design.

The sandwich arenas were placed in a plastic container with moist napkins, incubated at 25 °C in the dark, and checked one and two weeks after fungi application. In each of these checks, an attempt was made to recover half of the S. spinki mites present and they were mounted between slides and coverslips with Hoyer's liquid to estimate the proportion of those that were infected by fungi, by observation under a phase contrast microscope, 400 and 1000x.

The sandwich arenas functioned as wet chambers, so it was expected that they would stimulate the development of pathogenic fungi and their infection in mites, but they surely also stimulated the development of saprophytic fungi already present in the leaf sheaths. A mite was determined as infected if the mycelium was growing profusely within its body, or diagnostic structures of the applied fungi, such as the phialides of Hirsutella, were clearly seen. Percent infestations among treatments were compared by the Kruskal Wallis test (P = 0.05), using the software SAS 9.4 (^{SAS Institute, 2010}).

Results and discusion

Identification of mites associated with rice, including those determined to be predators of S. spinki. The presence of S. spinki was confirmed in all sampled locations (Table 1), which confirms that this mite has spread widely in rice-growing areas after approximately 12 years of its establishment in Mexico. Additionally, the close species Steneotarsonemus furcatus de Leon, 1956, was identified in the same habitat in the town of La Tranca, Tlalixcoyan, Veracruz (18.7656433, -96.251406). Steneotarsonemus furcatus had already been reported as inhabiting rice in Cuba and Brazil and apparently coexisting with S. spinki (^{Navia et al., 2006}). The damage that S. furcatus causes to rice has not been determined. The list of predatory mites identified in this study, real or potential natural enemies of S. spinki, with their respective collection data, is shown in Table 2.

Four of the species mentioned in Table 2 have a history of their role in the biological control of pest mites. Neoseiulus paspalivorus (^{De Leon, 1957}) has a wide global distribution, especially in tropical and subtropical areas (^{Demite et al., 2020}) and has been tested in Brazil for the biological control of the Aceria guerreronis Keifer, 1965 mite, which causes coconut scab (^{Lawson-Balagbo et al., 2008a}), but it has also been found in Cuba in rice, where its role in the control of S. spinki is unknown (^{Ramos & Rodríguez, 2004}). Gaeolaelaps aculeifer (Canestrini, 1884) has a wide global distribution in temperate climates (^{Mahjoori et al., 2014}) but has also been found in Colombia (^{Rueda-Ramírez et al., 2018}). This mite is an efficient predator of insects in the soil that is produced commercially as a biological control agent (^{Gerson et al., 2003}). Lasioseius youcefi Athias-Henriot, 1959 has a wide distribution in Africa, Asia, Europe, and North America (^{Negm, 2014}). ^{Lo and Ho (1984)} observed that it preys on S. spinki, and ^{Walter and Lindquist (1989)} define this mite as omnivorous, feeding on both small animals and fungi. Finally, Lasioseius subterraneus^{Chant, 1963} was found in Baja California, Mexico (^{Chant, 1963}), but it was also found in Brazil, inhabiting coconut palm (Cocos nucifera L., 1753) and associated with the mite A. guerreronis (^{Lawson-Balagbo et al., 2008a}).

For its part, P. curtipilis was collected in sorghum imported from Mexico (^{Chant, 1958}) and in Santa Cruz, California (^{Chant, 1963}). There is no data about its diet or its potential as a natural enemy of any pest. The identification of N. paspalivorus, G. aculeifer, and L. youcefi are the first records in Mexico of the presence of these species.

Pathogenic fungi of S. spinki in rice crops in the study area, identified by traditional means and molecularly characterized. Five isolates of acaropathogenic fungi of S. spinki obtained by M.A. Salazar-Santiago during 2015-2016 were reactivated, no additional isolates were obtained in 2019. Their observation under a microscope confirmed them as Hirsutella. The Ma 034 and Isaria isolates of M. anisopliae and I. fumosorosea, respectively, massively cultured on wet rice grains, and provided by J. Lara-Reyna, were added.

The identification results by comparison with sequences deposited in GenBank (BLAST) are shown in Table 3, which allow us to confirm the identity of all the fungi that were part of the present study. In the case of H. thompsonii, these are close haplotypes and only isolates M-1 and M-3 are identical in the amplified sequence. Due to the presence of synnemata in isolate M-2 (B-1) grown in H medium, it is classified as the synnematous variety of H. thompsoni.

Hirsutella nodulosa Petch, 1926, has been identified infecting S. spinki in Cuba (^{Cabrera et al., 2005}), Costa Rica (^{León González & Avilés Chávez, 2011}), and Sri Lanka (^{Cabrera et al., 2002}), so the finding of H. thompsonii as the only species that infects S. spinki in the study area represents novel data and suggests that native strains acquired the ability to infect this mite, recently established in Mexico.

Predation test. A thriving colony of mites identified as Tyrophagus sp. was established. With these mites used as food for predators, only one colony could be established with abundant specimens of G. aculeifer. Daily observations of consumption, survival, and reproduction of this predator appear in Figs. 1 to 3. Fig. 1 shows that G. aculeifer maintained a daily consumption of between four and 11 specimens of S. spinki. Given the above, it can be said that the predator was able to feed on the prey that was offered to it. However, the specimens of this predator were dying throughout the days of confinement, so that after 13 days there was only one survivor per arena on average (Fig. 2). Most of the eggs were laid in the first days of confinement, suggesting that the females were already in a gravid condition before receiving S. spinki as food. The maximum number of eggs per female per day was two (Fig. 3), and six females from each replicate did not lay any eggs, so it is postulated that S. spinki is not a suitable diet for G. aculeifer. Alternatively, it is possible that the confinement conditions, especially the high humidity inside the sandwich arenas, were not favorable for the development of G. aculeifer. However, this species has been observed mainly in soil, where it is effective in controlling small arthropods (^{Gerson et al., 2003}). Presumably, high relative humidity should not be a limiting factor for said mite.

Figure 1 mean number of Steneotarsonemus spinki specimens consumed by adult females of Gaeolaelaps aculeifer, separated by stages, as well as total consumption, over 13 days in sandwich arenas. Vertical lines represent the standard error of each reading. 

Figure 2 Survival curve of Gaeolaelaps aculeifer fed with Steneotarsonemus spinki in sandwich arenas. 

Figure 3 Average fecundity (number of eggs laid) over successive days of confinement of Gaeolaelaps aculeifer females in sandwich arenas, fed with Steneotarsonemus spinki. Vertical lines represent the standard error of each reading. 

It was not possible to establish colonies of predatory mites of other species, which was particularly unfortunate for N. paspalivorus because it is a small and flattened mite, which would make it easier for it to enter the small spaces where S. spinki lives, as was observed when testing its effectiveness to prey A. guerreronis (^{Lawson-Balagbo et al., 2008b}). According to ^{McMurtry et al. (2013)}, N. paspalivorus belongs to subtype III-d in the lifestyle classification of phytoseiid mites. This group is characterized for being generalist predators that live in tight sites in monocotyledons, precisely the microhabitat where it was found, so theoretically it should have the potential to control S. spinki. A closely related species, Neoseiulus cucumeris (Oudemans, 1930), is commercially raised using mites from the Acaridae family as food but prey on a wide variety of mites (^{McMurtry et al., 2015}).

Similarly, and according to ^{McMurtry et al. (2013)}, L. youcefi, L. subterraneus, G. aculeifer, and P. curtipilis are considered generalist predators, so it was expected that they would have the capacity to use Tyrophagus sp. as food.

Fungal pathogenicity test. The percentage of mites infected by pathogenic fungi in the experiment in sandwich arenas is shown in Table 4. The so-called sandwiches are arenas in which the mites, the substrate, and the applied pathogens are confined in a narrow space, where the humidity is concentrated, so the conditions are theoretically favorable for the development of pathogenic fungi, but also for saprophytic fungi. Under these conditions, the three fungal species were able to infect S. spinki mites in variable proportions. A greater proportion of mites infected by fungi was observed in the first reading (seven days after application), which is tentatively attributed to the fact that as the days passed, the saprophytic fungi inhibited the action of the pathogens. It should be noted that M. anisopliae and I. fumorosea were applied at a concentration of 10⁶ conidia/mL, while H. thompsoni was applied at a concentration 1000 times lower and, despite this, the latter fungus infected a higher percentage of mites, significantly different from the control one week after the inoculation. The first two species are considered pathogenic fungi for insects (^{Skinner et al., 2014}), while H. thompsoni, like other species of the same genus, are considered specialists in mite infection (^{Wekesa et al., 2015}). The isolate tested was obtained from S. spinki, which shows it to have greater potential to be used in the biological control of said mite.

It is notable that during the observations carried out in 2019, no infections were observed in S. spinki caused by Hirsutella spp., contrary to the observations of ^{Salazar-Santiago (2017)}, who obtained several isolates of species of this genus in natural infection. This suggests that infections, and possibly epizootics, may occur erratically and may not be a natural control mechanism for S. spinki. However, this does not exclude that activities can be carried out to promote the action of said fungus, such as its mass production and application in the field. The data from ^{Cabrera et al. (2005)}, who associate rice emptying with the S. oryzae fungus in Cuba, would motivate producers to use fungicides to control this problem. Apparently, the situation in Mexico is different, since ^{Salazar-Santiago et al. (2019)} showed that emptying is caused by the S. spinki mite alone, so stopping the application of fungicides can be a resource to facilitate the development of epizootics caused by Hirsutella spp.

In summary and by way of conclusion, H. thompsonii, or some species of this genus, among those identified, is the one that shows the greatest potential to be tested in the biological control of the panicle rice mite. It is not an easy species to cultivate (^{Wekesa et al., 2015}), so research on its mass production and application methods is a challenge to face. For its part, it is suggested to continue research on the predatory capacity of N. paspalivorus, due to its wide distribution and close association with S. spinki.