Low aflatoxin levels and flowering delay in Aspergillus flavus-resistant maize lines are correlated with increased corn earworm damage and enhanced seed fumonisin content

Preharvest mycotoxin contamination of field-grown crops is influenced not only by the host genotype, but also inoculum load, insect pressure and their confounding interactions with seasonal weather. In two field trials, we observed a preferred natural infestation of specific maize (Zea mays L.) genotypes by corn earworm (Helicoverpa zea Boddie) and investigated this unexpected interaction. These studies involved four maize lines with contrasting levels of resistance to Aspergillus flavus. The resistant lines had 7 to 14-fold greater infested ears than the susceptible lines. However, seed aflatoxin B1 levels, in mock- or A. flavus-inoculated ears were consistent with maize genotype resistance to A. flavus. Further, the corn earworm-infested ears had greater levels of fumonisin content in seeds than uninfested ears, indicating that the insect may have vectored native Fusarium verticillioides inoculum. The two maize lines with heavy infestation showed delayed flowering. The availability of young silk for egg-laying could have been a factor in the pervasive corn earworm damage of these lines. At the same time, H. zea larvae reared on AF-infused diet showed decreasing mass with increasing AF and >30% lethality at 250 ppb. In contrast, corn earworm was tolerant to fumonisin with no significant loss in mass even at 100 ppm, implicating the low seed aflatoxin content as a predominant factor for the prevalence of corn earworm infestation and the associated fumonisin contamination in A. flavus resistant lines. These results highlight the need for integrated strategies targeting mycotoxigenic fungi and their insect vectors to enhance the safety of crop commodities. IMPORTANCE Aspergillus and Fusarium spp. not only cause ear rots in maize leading to crop loss, they can also contaminate the grain with carcinogenic mycotoxins. Incorporation of genetic resistance into breeding lines is an ideal solution for mycotoxin mitigation. However, the goal is fraught by a major problem. Resistance for AF or FUM accumulation is quantitative and contributed by several loci with small effects. Our work reveals that host phenology (flowering time) and insect vector-mycotoxin interactions can further confound breeding efforts. A host genotype even with demonstrable resistance can become vulnerable due to seasonal variation in flowering time or an outbreak of chewing insects. Incorporation of resistance to a single mycotoxin accumulation and not pairing it with insect resistance may not adequately ensure food safety. Diverse strategies including host-induced silencing of genes essential for fungal and insect pest colonization and broad-spectrum biocontrol systems need to be considered for robust mycotoxin mitigation.

These studies involved four maize lines with contrasting levels of resistance to 28 Aspergillus flavus. The resistant lines had 7 to 14-fold greater infested ears than 29 the susceptible lines. However, seed aflatoxin B 1 levels, in mock-or A. flavus-30 inoculated ears were consistent with maize genotype resistance to A. flavus. 31 Further, the corn earworm-infested ears had greater levels of fumonisin content 32 in seeds than uninfested ears, indicating that the insect may have vectored native 33 Fusarium verticillioides inoculum. The two maize lines with heavy infestation 34 showed delayed flowering. The availability of young silk for egg-laying could have 35 been a factor in the pervasive corn earworm damage of these lines. At the same 36 time, H. zea larvae reared on AF-infused diet showed decreasing mass with 37 increasing AF and >30% lethality at 250 ppb. In contrast, corn earworm was 38 tolerant to fumonisin with no significant loss in mass even at 100 ppm, 39 implicating the low seed aflatoxin content as a predominant factor for the 40 prevalence of corn earworm infestation and the associated fumonisin infection, but also vectors ear rot and stalk rot fungi, particularly F. verticillioides 89 and F. graminearum (Widstrom 1992). Extensive use of Bt (Bacillus thurigiensis 90 Crystal proteins-expressing) maize with its high efficacy against ECB, has and to Bt maize has also been documented (Capinera 2004;Dively et al. 2016;98 Kaur et al. 2019). Although CEW has multiple crop and weed hosts, maize is its 99 preferred host (Johnson et al. 1975

113
In addition to facilitating fungal colonization, insect infestation can also 114 enhance mycotoxin production in host tissues (Döll et al. 2013;Drakulic et al. 115 2015Drakulic et al. 115 , 2016. In turn, mycotoxigenic fungi can affect insect vector infestation by 116 inducing volatile production in host tissues. This is particularly well documented 117 in the case of Fusarium species (Schulthess et al. 2002;Piesik et al. 2011;Drakulic 118 et al. 2016 unseasonal and steep warming after protracted cold seems to have favored an 142 explosion of CEW population as indicated by a heavy infestation of ears in both 143 of our experimental plots. CEW incidence was also reported from maize fields in 144 other states in southern (Porter and Bynum 2018) as well as northern US (e.g.,

145
Handley 2018). In spite of two applications of a strong broad-spectrum 146 insecticide before and after silking, the insecticide seems to have failed to reach 147 silks covered by the husks. Further, all ears were bagged immediately after 9 inoculation/pollination, which concealed earworm damage until developing ears 149 were sampled for analysis. harvested and utilized all ears in the plots to obtain robust AF data. The insect 163 infestation was 8-fold greater in CML322 than observed in B73 ears in the mock-164 inoculated set. Inoculation with the highly toxigenic Tox4 strain resulted in a 165 significant (p<0.01) and nearly 4-fold decrease in the infestation of CML322, but 166 still 2-fold greater than infestation in B73. This is inversely correlated with >3-167 fold increase in seed AF content in Tox4-inoculated CML322 ears. As expected 168 from its susceptibility to A. flavus colonization, B73 seeds accumulated >100 ppb 169 of AF even in mock-inoculated (Control) ears and >500 ppb in Tox4-inoculated ears. These AF levels are >12-19 fold higher than those measured in CML322 171 seeds (Fig. 1B, right panel). CEW infestation was also greater in the resistant 172 hybrid (Mp313E x Mp717) than in the susceptible hybrid by >30-fold in the 173 control set and by 7-fold in the inoculated set (Fig. 1A, left panel). Infestation 174 was inversely correlated with seed AF levels in hybrids as well. The susceptible 175 hybrid (GA209×T173) had 100 ppb in control seeds and >400 ppb of AF in the 176 inoculated set (i.e., 3 and 24-fold greater than in the resistant hybrid). Unlike the 177 resistant inbred CML322, the resistant hybrid showed no difference in either AF 178 content or CEW infestation between the control and CA14-inoculated ears.

184
Not surprisingly, ANOVA of AF content revealed that the host genotype and 185 inoculation with toxigenic A. flavus strains showed highly significant 186 independent (or direct) as well as interaction effects on seed AF content. As 187 indicated by the data presented in Fig. 1, infestation was also significantly related 188 to AF content, although the interaction effect of genotype with infestation on AF combined for control and inoculated sets in each genotype (Fig. 2) or when all 197 data is combined (Fig. S2). It is of interest to note that the uninfected controls  analyzed in the same seed samples used for AF determination (Fig. 3 A) and 213 compared between uninfested and CEW-infested samples (Fig. 3B).
Both maize hybrids used in this study have been previously shown to be 215 resistant to FUM accumulation. In particular, Mp313ExMp717 (A. flavus resistant 216 hybrid) was shown to be more robustly resistant than GA209xT173 across 217 studies (Williams 2006;Henry et al. 2009;Williams and Windham 2009). In the 218 current study, however, the Mp313ExMp717 accumulated >7-fold FUM in its 219 seeds than GA209xT173 (Fig. 3A). Although CML322 accumulated a considerable 220 amount of FUM, it was >4-fold less than that in B73, which is known to be among The preferential infestation of A. flavus resistant lines by CEW and a negative 232 correlation between AF and infestation rate, taken together with greater FUM 233 levels in infested ears, suggested that AF may be more toxic to H. zea than FUM. 234 We tested this hypothesis by feeding experiments where CEW neonates were 235 reared on artificial diet containing graded levels of AF or FB1. Results shown in 236 Fig. 4 and 5 clearly demonstrate that the pest is more susceptible to AF than to FB1. As reported previously (Zeng et al. 2006), AF retarded CEW larval growth 238 even at the lowest concentration tested, although the effect was not significant 239 ( Fig. 5) and was toxic above 200 ppb (Fig. 4). On the other hand, FB1 was non-240 toxic to CEW even at the highest concentration tested. In fact, at lower 241 concentrations (below 30 ppm) the toxin seems to marginally enhance the growth silks may be an important factor for the increased H. zea infestation of these late 258 flowering genotypes. However, in an adjacent plot where B73 was planted two weeks later (unrelated to the current study), silk emergence coincided with that 260 of CML322 plants used in the present study. Nonetheless, B73 ears had highly 261 elevated levels of seed AF (400 ppb in controls and 800 ppb in inoculated plants) 262 and low levels of CEW infestation in this plot as well, suggesting that high seed 263 AF levels may act as a deterrent for CEW infestation because of its toxicity.

265
There are few studies where CEW infestation patterns have been compared 266 in maize genotypes with varying resistance to A. flavus or AF accumulation. Nie Further, high AF contamination (≤100 ppb) in uninoculated as well as inoculated 278 plots of only susceptible lines allowed to make robust comparisons.

279
The premise for this study is an unprecedented or unreported observation, 280 in that two unrelated maize lines (Tuxpeño germplasm versus CML) with proven 281 resistance to A. flavus were heavily infested by CEW. Conversely, the two A. flavus 282 susceptible lines (stiff-stalk inbred B73 and non-stiff stalk hybrid GA209 x T173) 283 were spared from heavy CEW damage. Although late flowering maize is known 284 to be susceptible to CEW infestation by providing green silks, availability of silks 285 alone could not fully explain our observations. Late flowering is more often a problem in the northeastern US where it coincides with CEW migration from 287 southern states. Furthermore, late planted B73 in an adjacent plot had delayed 288 silk emergence but showed no CEW infestation. The other and more likely 289 explanation is that the susceptible lines had very high levels of AF that were toxic 290 to CEW. Even mock-inoculated controls had 100 ng of AF per gram of seed meal 291 prepared from entire ears with both moldy and non-moldy seeds. This argument 292 is supported by previous studies on AF toxicity to CEW in feeding experiments 293 (Zeng et al. 2006) as well as our current work ( Fig. 4 and 5). Zeng et al (2006) 294 showed that AF at 200 ppb strongly inhibited the growth and development of 295 first instar larvae, leading to >50% larval death after 9 d and 100% death after 15 296 d of feeding. Even lower concentrations (1-20 ppb; FDA-regulated levels) affected 297 larval development, delayed pupation rate and led to >40% mortality when the 298 exposure was longer than 7 d (Zeng et al. 2006). Although concentrations below 299 20 ppb were not tested in our study, we observed a steady decline in larval mass 300 as AF concentration increased with ≥30% mortality at or above 250 ppb during 301 10-15 d exposure (Fig. 5). We did not continue our observations beyond the larval 302 stage to assess the longer term developmental effects (e.g., pupation or 303 emergence of adults). An apparent exception to the correlation between low AF 304 and high CEW infestation was a significant decrease in CEW infestation observed 305 in TOX4-inoculated ears compared to uninoculated ears in the A. flavus resistant 306 inbred CML322, although average AF levels did not exceed 30 ppb. Given the 307 highly variable distribution of AF in individual kernels of a maize ear (e.g., Lee et 308 al. 1980), it is possible that AF content particularly in damaged kernels (close to the silk canal, the site of inoculation as well as CEW infestation) was much greater 310 than the average for the entire ear and high enough to be toxic to CEW survival.  It is not surprising that AF is toxic to insects, not merely to mammals. A. 318 flavus is predominantly a soil-living saprophyte, feeding on decaying organic 319 matter, including dead insects. It is also an opportunistic pathogen and can 320 colonize a wide variety of insects, e.g., moths, silkworms, bees, grasshoppers,   (Zeng et al. 2006;Opoku et al. 2019). Spatial correlation analysis of natural infestation by different pests and seed AF content in field-grown maize 333 plants indicated that AF content was correlated to the frequency of weevils and 334 stink bug-affected kernels, but not with CEW damage (Ni et al 2011).

335
Our work also showed that FUM is not toxic to H. zea (Fig. 4). This may 336 have allowed CEW to vector F. verticillioides and other FUM-contaminating fungi, 337 as indicated by an increased seed FUM content in infested ears (Fig. 3). CEW 338 damage is also frequently associated with the colonization by another

371
In contrast to a mutual antagonism reported previously between A. flavus 372 and F. verticillioides (Zummo and Scott 1992; also see Fig. S3), we observed high 373 levels of AF and FUM co-contaminating our samples. B73, in particular with its 374 high susceptibility to both mycotoxigenic fungi, had very high levels of both AF 20 and FUM in many of its seed samples. Although CEW damage was very low in this 376 inbred ( Fig. 1B and 2), FUM levels were exacerbated in infested ears (Fig. 3B). 377 There is some evidence for an additive or even synergistic effect on 378 carcinogenicity from co-exposure to AF and FUM (World Health Organization 379 2018). Based on biomarker studies and food analyses, the co-occurrence of these 380 two mycotoxins has been widely documented in developing countries (Shirima et   used for AF analysis to have robust AF data. When the seed meal exceeded more than 100 g (in uninoculated controls), we took more than one sample to minimize 433 sampling error. AF from seed meal was extracted and measured as before          (mock-inoculated) and CA14-inoculated set. Seed AF content in CA14-inoculated 5 set and the control were also similar in the resistant hybrid (Mp313E x Mp717). 6 (B) Data shown is from inbreds. There was a similar negative relationship 7 between CEW infestation rate and seed AF content as was observed in hybrids.

8
Infestation was significantly dependent on the host genotype with very little 9 difference between control (mock-inoculated) and Tox4-inoculated plots except 10 in the case of CML322. The resistant inbred showed only 30% infestation in Tox4 11 inoculated set compared to the control. Seed AF levels were significantly higher 12 in B73 both in control and inoculated ears than those of CML322. Values shown 13 are average + SE. Significant differences (P value <0.05) between each data set 14 were tested using an ANOVA (Supplemental Table 1) followed by Tukey's 15 multiple-comparisons post hoc test (Supplemental Table 2) in R (version 3.6.2).

16
Means are significantly different if marked by a different letter.

FIG. 2. CEW damage is negatively correlated with seed AF content in maize 20
18 lines. The infestation and AF data from control and infected ears is combined in 19 each genotype. Significant differences (P value <0.05) between each data set were 20 tested using an ANOVA (Supplemental Table 3    The values marked with the same letter are not statistically significant. FB 1 had 41 no significant effect on larval growth at concentrations tested. (B) Data shown is from inbreds. There was a similar negative relationship between CEW infestation rate and seed AF content as was observed in hybrids.
Infestation was significantly dependent on the host genotype with very little difference between control (mock-inoculated) and Tox4-inoculated plots except in the case of CML322. The resistant inbred showed only 30% infestation in Tox4 A B inoculated set compared to the control. Seed AF levels were significantly higher in B73 both in control and inoculated ears than those of CML322. Values shown are average + SE. Significant differences (P value <0.05) between each data set were tested using an ANOVA (Supplemental Table 1) followed by Tukey's multiple-comparisons post hoc test (Supplemental Table 2) in R (version 3.6.2).
Means are significantly different if marked by a different letter.

Fig. 2. CEW damage is negatively correlated with seed AF content in maize
lines. The infestation and AF data from control and infected ears is combined in each genotype. Significant differences (P value <0.05) between each data set were tested using an ANOVA (Supplemental Table 1) followed by Tukey's multiplecomparisons post hoc test (Supplemental Table 2) in R (version 3.6.2). Average (+SE) infestation and AF values between A. flavus susceptible and resistant lines are highly significant (p<0.01).  well bioassay plate for 10 d. Each well had 1 g of feed and a single neonate at the start of the assay. A representative assay from 4 replicates is shown. In an additional assay, 100 ppm of FB 1 and 300 ppb of AF were tested. Results were not different, except for a greater larval mortality at 300 ppb of AF (data not shown). Scale Bar = 1 cm. The values marked with the same letter are not statistically significant. FB 1 had no significant effect on larval growth at concentrations tested.  Contents of the two mycotoxins from the same seed sample are poorly correlated in both sets as indicated by Pearson correlation coefficient values (r = -0.0983 for hybrids and 0.3344 for inbreds). This lack of correlation indicated that there was no mutual effect in the production of the two mycotoxins by the fungi infecting seeds from same ears.