Potato dextrose agar was used for stock culture of F. verticillioides originating from a single microconidium (storage: 10 °C) [8]. Laboratory stock colonies of O. nubilalis and H. armigera were established. For a stock colony of O. nubilalis, adults were collected during 2004 in Kéty (Hungary). Larvae were fed on a semi-synthetic diet [9]. For breeding of H. armigera, larvae were collected during 2008 in Zsámbék (Hungary). Larvae were fed on a different semi-synthetic diet [10].

In a preliminary examination, fecal pellets of H. armigera were collected in a maize field (Páty, August 8, 2009). The fecal samples contained F. verticillioides (∼80%) and F. proliferatum (∼20%) microconidia, thus further experiments were focused on F. verticillioides. RbGUM selective media was used for Fusarium species isolation (incubation: 25 °C, 7–10 days). Insect and plant samples under field conditions were collected and examined using RbGUM media [8].

Lepidopteran larvae reared for the field experiments in maize ear isolators were fed on sweet maize (Jubilee) discs with or without F. verticillioides mycelia. To prepare a F. verticillioides infected disc, sweet maize cobs were cut to 3 cm thick discs. Discs were dipped into an aqueous solution (200 mL) including a stock culture of F. verticillioides and 0.1% Tween 20. The final concentration was ∼80,000 microconidia/mm³. After four days (incubation: 22–25 °C), when the distribution of mycelia was at least ∼50% on the surface of a single disc, second or third instar larvae were transferred onto them. Larvae were fed on sweet maize discs with or without F. verticillioides mycelia until the next molt.

Field work was done at the Ecological Research Station, Plant Protection Institute at Julianna-major (the north-western outskirts of Budapest). Maize ear isolators were placed on MON 810 or near isogenic plants at the beginning of silking time (August 13, 2009). Isolators were placed on artificially infected cobs using third or fourth larval stadia. The stem was cut above the maize ear at the height of maize silk, an appropriately treated larva was placed on it, and the fine mesh maize ear isolator (diameter = 10 cm, length = 35 cm) was placed on the entire maize ear and was tightly closed at the bottom. Treatments were: (i) H. armigera larvae having consumed F. verticillioides; (ii) H. armigera larvae not having consumed F. verticillioides; (iii) O. nubilalis larvae having consumed F. verticillioides; (iv) O. nubilalis larvae not having consumed F. verticillioides mycelia; and (v) untreated control (Table 1). Repetitions were 10–14 for H. armigera and 18–20 for O. nubilalis. A treated control (vi; 20 repetitions) consisting of F. verticillioides mycelia on RbGUM media disc (5 mm diameter) fixed on maize silk was also used. After two weeks, maize ear isolators were collected and their content was carefully checked.

Cry1Ab toxin content in the main parts of the maize ears was determined by a commercial sandwich immunoassay, Abraxis Bt-Cry1Ab/Ac ELISA kit (#PN 51001, Warminster, PA, USA) carried out in 96-well microplates according to manufacturer-provided protocol. ELISA signals were detected on an iEMS microtiter plate reader (Labsystems, Helsinki, Finland). Cry1Ab protoxin calibrators (Abraxis) were put on every microplate at concentrations between 0.25 and 4 ng/mL, assays were used for determination of analyte concentration by linear regression. Activated Cry1Ab toxin concentrations were calculated from detected protoxin concentration values with Cry1Ab activated toxin/protoxin cross-reactivity, 56.4%, previously determined for the Abraxis Bt-Cry1Ab/Ac ELISA kit [11,12].

A MON 810 (DK-440 BTY) and its near isogenic line (DK-440) were investigated during October 5–8, 2009, in Julianna-major. Depending on the plot size, some hundreds (ca. 100–400) of maize ears were collected from every assortment. Upon removal of husk leaves, cobs were carefully investigated for symptoms and severity of larval and fungal damage (proportion of the damaged part in the cob, location of the damage, occurrence of pink ear rot as a sign of fungal infection). Four replicates were used resulting in 457 to 1,339 cobs investigated per a single plot. Data without transformation were analyzed using Statistica ver. 5.5 program (ANOVA and Tukey test). Insect and plant samples were also collected for Fusarium spp. identification in the laboratory.

To evaluate the real yield loss, cobs (10 replicates) in different sizes (small, medium and big size of cob) infected by H. armigera larva were chosen. The infected part of the cobs was removed and masses were measured.

Fusarium verticillioides was identified in laboratory fungal rearing tests (using Fusarium-selective RbGUM media) from plant and fecal pellet samples collected by the maize cob isolators in the field experiments. Thus, F. verticillioides infection was verified in fecal pellets of H. armigera and O. nubilalis, husk leaves, ripening kernels and cobs from maize plants damaged by H. armigera and O. nubilalis, as well as stem tunnels caused by O. nubilalis. Insect-transmitted infestation of F. verticillioides (manifested in the occurrence of pink ear rot) appeared to be dependent on the choice of the plant part by the insect to feed on, and on the varying subsequent survival rates due to different Cry1Ab toxin exposures. In contrast, no infestation by F. verticillioides was observed in the untreated control, and none was produced by fixing F. verticillioides mycelia on corn silk, either (Table 1). All the H. armigera larvae used (third and fourth stadia) chose maize ears for feeding. The majority of O. nubilalis larvae (third and fourth stadia) preferred to feed on the fresh wound on the stem, where it had been cut prior to the placement of the maize ear isolators on isogenic maize. A small proportion (10–15%) of O. nubilalis larvae tended to feed on husk leaves on MON 810 maize (Table 1). Larvae of both species investigated, having been fed previously in laboratory rearing during one larval instar on F. verticilliodes mycelia, showed increased mortality (10–17%). F. verticillioides infection was transmitted only by larvae having been fed previously on its mycelia (Table 1). Nonetheless, not all larvae (H. armigera and O. nubilalis) having been fed on F. verticillioides mycelia could transmit Fusarium infestation; the rate was only 20–42%. All of the O. nubilalis larvae died on MON 810 maize, although only this species could transmit F. verticillioides infection feeding on the base of maize ears. Some H. armigera larvae (third and fourth stadia) could survive on MON 810 maize, feeding on husk leaves (Table 1). These larvae stop feeding from time to time, and starve—showing the symptoms of Cry1-toxicosis—consuming the minimum possible.

Similarly high levels of maize ear infestation (42–46%) were observed in all three, closely located maize fields investigated. Only 20–30% of cobs with larval damage were also infected by F. verticillioides. Larval damages were mostly attributed to H. armigera in two of the maize fields, while both H. armigera and O. nubilalis larvae occurred in nearly identical rates in the third case (Table 2). Independently from the damaging insect species, mostly apical cob infestation occurred. Fusarium verticillioides infestation was significantly higher in one case, but it did not correlate with the overall larval damages (Table 2).

In case of O. nubilalis, the tunnels were found mostly on stems, and only a quarter of the damage occurred on maize ears (Table 3).

The position where the larvae fed on the maize ears were different for the two species evaluated. In the case of H. armigera, ∼90% of the damage occurred on the apical region. A lower value (∼80%) of apical damage was related to O. nubilalis (Table 4). First instar larvae of both species usually try to reach the kernels through the maize silks, although a more significant portion of O. nubilalis larvae choose the base of maize ear. In case of microbial infection (like F. verticillioides), older H. armigera larvae might move (∼10%) toward the middle of the cob, seeking a drier environment, or come out from the maize ear at the top and choose another maize ear. In this latter case, older H. armigera larvae choose husk leaves to reach the middle of the cob. The most frequent damage type by H. armigera larvae occurs on the top of the cob by feeding on ripening seeds. An ample amount of fecal pellets may be found at places of earlier seeds. Fecal pellets of H. armigera are sometimes covered by mycelia of different fungi. In our case, plant pathogenic F. verticillioides was the most abundant (Table 5). The damage type caused by O. nubilalis larva is different. It makes a longer tunnel toward the base of the cob under the seed surface. Fecal pellets of O. nubilalis are usually not visible on the cob surface, thus damage is not so bulky.

The ∼Cry1Ab toxin content (corrected with toxin/protoxin cross-reactivity [11,12]) was found to be 826 ng (s.d. = 237); 1280 ng (s.d. = 150) and 2075 ng (s.d. = 1287) ∼CryAb toxin/g fresh mass in corn silk; husk leaves and young cob, respectively. The high standard deviation in the cob is most likely due to this plant part being a mixture of different tissues with variable amounts of ∼Cry1Ab toxin.

Apical infection by H. armigera larvae—the most frequent damage type (Table 4)—results in the loss of only 10–15% of the entire cob (Table 6). Infection and damage on the base and the middle of the cobs might cause more damage.

MON 810 maize shows a high efficacy against H. armigera and O. nubilalis larval infection. None of the O. nubilalis larvae survived in stems, but some in maize ears (Table 7). There are some survivors in the case of H. armigera in maize ears. MON 810 drastically reduced F. verticillioides infection as well.

It was apparent from the laboratory experiments that young larvae of H. armigera and O. nubilalis do not tolerate well F. verticillioides mycelia in their food (Table 1). In a laboratory experiment, H. armigera L1 could develop only until the third larval stadium (data not shown). O. nubilalis larvae could tolerate F. verticillioides mycelia much better. Nonetheless, H. armigera and O. nubilalis larvae feeding on F. verticillioides mycelia for a short period can spread fungal microconidia, which can survive the digestive channel of these insects. F. verticillioides was readily identified from the fecal pellets of both species. H. armigera larvae usually attempt to escape from moldy tunnels. During this period, the larva can transmit F. verticillioides. Evaluating co-occurrence of insect and fungal infection on the isogenic maize line, only 20–40% of the larva feeding on F. verticillioides infected seeds was found to transmit maize pink ear rot (Table 1). Similarly, some 20–30% of larval infection was found to be related to F. verticillioides infection in the field experiments (Table 2). In contrast, MON 810 maize, highly effective against the larvae, strongly reduced the F. verticillioides infection as well (Table 7). Incidental wounds on the top of maize ears [13] may play a more important role in this relationship than direct transmission. Insecticide (lambda-cyhalothrin) application against O. nubilalis, monoculture and effects of sowing date also indicate this effect [14–18]. In agreement with reported results [19,20], infestation through intact maize silk with F. verticillioides did not occur. Certain, but not all Fusarium mycotoxins were found to be related to O. nubilalis damage on Bt-plants (SYN-EV176 and MON 810 events) [21].

MON 810 maize is advertised to be used against O. nubilalis larvae, which usually feed on the base of the leaves after hatching and subsequently make a tunnel into the stem (Table 4). MON 810 maize leaves produce the largest concentration and amount of the truncated toxin (ca. 10,000 ng ∼CryAb toxin/g fresh mass), while only a moderate concentration (ca. 1,500 ng ∼CryAb toxin/g fresh mass) is produced in the stem [11,12]. Different parts of the maize ear produce moderate (husk leaves, cobs) or low (maize silk) levels of ∼Cry1Ab toxin, therefore, are more suitable plant parts for the survival of those larvae which can feed on it. This is the probable reason why first instar larvae usually survive in the upper part of the maize ear feeding first on the maize silk. Similarly to reported results [22], no survival of O. nubilalis larvae was seen in the stem of MON 810 (Tables 1 and 7), but several survived in the lower part of the cob creating tunnels in the base of the husk leaves [23,24]. Larvae surviving in ∼Cry1Ab toxin containing maize ear with variable toxin content may become the source of Cry1-resistant strains in the future [25,26]. This applies especially to spots near the rescue zone, where intraspecific hybrid seeds with only one transgenic parental line and, consequently, lower Cry1Ab toxin production levels than the parental GMO trait (data not shown), are frequent.

Although the most frequent apical damage of maize ear is very spectacular (Tables 4 and 5), the yield loss is rather moderate. High cob damage (40–50%) caused by H. armigera and/or O. nubilalis larvae is very rare in Hungary, and the yield loss is only 4–8% of the cob mass even in those infrequent cases (corresponding to 10–16% cob mass, Table 6). In reality, yield losses due to infected cob tops are even smaller, as seeds on the thinner top part of the cobs are often lost anyway during mechanic shelling of cobs. As the fungal infection occurs predominantly apically on the cobs, such loss of cob tops also reduces the average fusarotoxin content in the shelled kernels. This is the basic reason why Hungarian farmers practically do not use chemical insecticides against H. armigera and O. nubilalis, even though authorized preparations are available [27].