Recombinant Expression of ABCC2 Variants Confirms the Importance of Mutations in Extracellular Loop 4 for Cry1F Resistance in Fall Armyworm

Fall armyworm (FAW), Spodoptera frugiperda, is a highly destructive and invasive global noctuid pest. Its control is based on insecticide applications and Bacillus thuringiensis (Bt) insecticidal Cry toxins expressed in transgenic crops, such as Cry1F in Bt corn. Continuous selection pressure has resulted in populations that are resistant to Bt corn, particularly in Brazil. FAW resistance to Cry1F was recently shown to be conferred by mutations of ATP-binding cassette transporter C2 (ABCC2), but several mutations, particularly indels in extracellular loop 4 (ECL4), are not yet functionally validated. We addressed this knowledge gap by baculovirus-free insect cell expression of ABCC2 variants (and ABCC3) by electroporation technology and tested their response to Cry1F, Cry1A.105 and Cry1Ab. We employed a SYTOXTM orange cell viability test measuring ABCC2-mediated Bt toxin pore formation. In total, we tested seven different FAW ABCC2 variants mutated in ECL4, two mutants modified in nucleotide binding domain (NBD) 2, including a deletion mutant lacking NBD2, and S. frugiperda ABCC3. All tested ECL4 mutations conferred high resistance to Cry1F, but much less to Cry1A.105 and Cry1Ab, whereas mutations in NBD2 hardly affected Bt toxin activity. Our study confirms the importance of indels in ECL4 for Cry1F resistance in S. frugiperda ABCC2.


Introduction
Spodoptera frugiperda (J.E. Smith) (Lepidoptera: Noctuidae), fall armyworm (FAW), is a highly-destructive moth pest native to the tropical and subtropical Americas that feeds on a broad range of host plants [1,2], including important row crops such as corn, cotton and soybean [3,4]. FAW recently invaded the Eastern Hemisphere and the first outbreaks were described in West Africa in 2016 [5], from where it rapidly dispersed to many countries in Sub-Saharan Africa [6]. In 2018, FAW was detected in India [7], followed by other countries in the Asia-Pacific, including China and Australia [8][9][10].
FAW pest management strategies largely rely on the use of synthetic insecticides and transgenic crops, e.g., corn or cotton expressing Bacillus thuringiensis (Bt) insecticidal proteins such as Cry and Vip toxins [11][12][13]. Corn hybrids expressing Bt toxins remain a cornerstone in FAW control in North and South America [11,14]. Transgenic corn expressing Cry1F-particularly targeting FAW-was commercialized in North America and Brazil in 2003 and 2009, respectively [15,16]. Over the last 20 years, Bt crop technology was rapidly adopted for pest management on a large scale, thus augmenting the resistance risk due to increased cross-crop selection pressure on targeted pests, including FAW [17][18][19][20]. The first cases of FAW resistance compromising the field-efficacy of transgenic corn expressing 2. Results 2.1. Effect of SfABCC2 ECL4 Mutations on Cry1F, Cry1A.105 and Cry1Ab Toxicity In a previous study [44], we identified several mutations in the ABCC2 ECL4 of Cry1F-resistant FAW populations from Brazil ( Figure 1). To investigate the effect of these mutations on the ability of the FAW ABCC2 (SfABCC2) to serve as a Bt toxin receptor, SfABCC2 wildtype and variant proteins were functionally expressed in Sf9 cells transfected by electroporation technology, rather than commonly-used baculoviruses.
Toxins 2022, 14, x FOR PEER REVIEW 3 of 11 ABCC2 variants by electroporation technology and employing a highly efficient fluorescence-based cytotoxicity assay based on SYTOX TM orange in 384-well plates.

Effect of SfABCC2 ECL4 Mutations on Cry1F, Cry1A.105 and Cry1Ab Toxicity
In a previous study [44], we identified several mutations in the ABCC2 ECL4 of Cry1F-resistant FAW populations from Brazil ( Figure 1). To investigate the effect of these mutations on the ability of the FAW ABCC2 (SfABCC2) to serve as a Bt toxin receptor, SfABCC2 wildtype and variant proteins were functionally expressed in Sf9 cells transfected by electroporation technology, rather than commonly-used baculoviruses.  [44] tested in this study. Red: amino acid substitution; blue: inserted amino acids; red dashed box: deleted amino acids.
The toxicity of the trypsin-activated Bt toxins Cry1F, Cry1A.105 and Cry1Ab to these ABCC2-expressing Sf9 cells was determined in a SYTOX TM orange cytotoxicity assay measuring the membrane permeabilization upon incubation with Bt toxins ( Table 1). The expression of the SfABCC2 wildtype protein in Sf9 cells conferred susceptibility to Cry1F, Cry1A.105 and Cry1Ab with EC50-values of 0.395, 0.0946 and 0.110 nM, respectively, while non-transfected Sf9 cells were not affected. Sf9 cells expressing the SfABCC2 variants 1 (GY-deletion) and 4-7 (see Table 1) showed no susceptibility to Cry1F at the highest tested concentration of 1000 mM, resulting in Cry1F resistance ratios (RR) of >2532-fold. These variants contain ECL4 indels of various length (see Figure 1B). Similarly, the amino acid substitutions P799K and P799R (variants 2 and 3, respectively) conferred high resistance to Cry1F, i.e., RR 630-and 1643-fold, respectively. In contrast, the toxicity of Cry1Ab and Cry1A.105 to the cells expressing the SfABCC2 ECL4 mutant variants was much less affected compared with Cry1F. For the variants 1-4, in which, at most, three amino acids in the ECL4 were altered (substituted or deleted), low RRs of 3.6-to 16.7-fold were observed.  [44] tested in this study. Red: amino acid substitution; blue: inserted amino acids; red dashed box: deleted amino acids.
The toxicity of the trypsin-activated Bt toxins Cry1F, Cry1A.105 and Cry1Ab to these ABCC2-expressing Sf9 cells was determined in a SYTOX TM orange cytotoxicity assay measuring the membrane permeabilization upon incubation with Bt toxins ( Table 1). The expression of the SfABCC2 wildtype protein in Sf9 cells conferred susceptibility to Cry1F, Cry1A.105 and Cry1Ab with EC 50 -values of 0.395, 0.0946 and 0.110 nM, respectively, while non-transfected Sf9 cells were not affected. Sf9 cells expressing the SfABCC2 variants 1 (GY-deletion) and 4-7 (see Table 1) showed no susceptibility to Cry1F at the highest tested concentration of 1000 mM, resulting in Cry1F resistance ratios (RR) of >2532-fold. These variants contain ECL4 indels of various length (see Figure 1B). Similarly, the amino acid substitutions P799K and P799R (variants 2 and 3, respectively) conferred high resistance to Cry1F, i.e., RR 630-and 1643-fold, respectively. In contrast, the toxicity of Cry1Ab and Cry1A.105 to the cells expressing the SfABCC2 ECL4 mutant variants was much less affected compared with Cry1F. For the variants 1-4, in which, at most, three amino acids in the ECL4 were altered (substituted or deleted), low RRs of 3.6-to 16.7-fold were observed. For the variants with larger indels (variants 5-7), RRs of 46.3 to 301.3 were determined.

SfABCC3 Is a Receptor for Cry1Ab but Not Cry1F
Next, we tested the toxicity of Cry1F and Cry1Ab to Sf9 cells expressing SfABCC3. Expression of the SfABCC3 conferred high susceptibility of the Sf9 cells to Cry1Ab (EC 50 3.32 nM), whereas no toxicity of Cry1F to SfABCC3-expressing Sf9 cells was observed at the highest tested concentration of 1000 nM ( Table 1), suggesting that the ABCC3 shows a functional redundancy to ABCC2 as a Cry1Ab receptor but does not function as a Cry1F receptor, mediating pore formation.

ABCC2 Gating Activity has no Effect on Cry1 Activity
In addition to the mutations in the ECL4, we previously observed an amino acid substitution (G1088D) close to the ABCC2 NBD2 of Cry1F-and Cry1Ab-resistant FAW strains [44]. We expressed the SfABCC2 G1088D variant as well as a variant in which NBD2 was completely deleted in Sf9 cells to investigate the effect of these mutations and the necessity of the ABCC2 gating activity to function as a Bt toxin receptor (Figure 2). The mutation G1088D had no significant effect on the toxicity of Cry1F (EC 50 (Figure 2A,B; Table 1). Additionally, the expression of the NBD2-deletion variant conferred nanomolar susceptibility of Sf9 cells to Cry1F and Cry1Ab. The deletion of the NBD2 had no significant effect on the toxicity of Cry1Ab (EC 50 0.183 nM, CI95% 0.154-0.218), but resulted in a 15fold decreased Cry1F toxicity (EC 50 5.77 nM, CI 95% 5.26-6.34) compared to the wild-type ABCC2 ( Figure 2C; Table 1). However, based on EC 95 values there was no significant difference in Cry1F efficacy between wild-type ABCC2 (47.0 nM, CI95% 25.7-85.8) and the NBD2-deletion mutant (45.9 nM, CI95% 35.4-59.5).
Toxins 2022, 14, x FOR PEER REVIEW 5 of 11 [44]. We expressed the SfABCC2 G1088D variant as well as a variant in which NBD2 was completely deleted in Sf9 cells to investigate the effect of these mutations and the necessity of the ABCC2 gating activity to function as a Bt toxin receptor (Figure 2 (Figure 2A,B; Table  1). Additionally, the expression of the NBD2-deletion variant conferred nanomolar susceptibility of Sf9 cells to Cry1F and Cry1Ab. The deletion of the NBD2 had no significant effect on the toxicity of Cry1Ab (EC50 0.183 nM, CI95% 0.154-0.218), but resulted in a 15fold decreased Cry1F toxicity (EC50 5.77 nM, CI 95% 5.26-6.34) compared to the wild-type ABCC2 ( Figure 2C; Table 1). However, based on EC95 values there was no significant difference in Cry1F efficacy between wild-type ABCC2 (47.0 nM, CI95% 25.7-85.8) and the NBD2-deletion mutant (45.9 nM, CI95% 35.4-59.5).  ) were treated with activated Cry1F and Cry1Ab, respectively, and the toxicity was determined in a SYTOX TM orange cytotoxicity assay measuring membrane permeabilization. Data are mean values ± CI95% (n = 4). Blue star: G1088D substitution; orange dashed box: deleted NBD2.

Discussion
There is an emerging body of evidence that ABC-transporters are critical determinants of three-domain (3D) Cry toxin-mediated toxicity and resistance in several lepidopteran species [29,32,33,[49][50][51]. This is also supported by results obtained with ABCC2 knockout lines of S. frugiperda expressing high levels of resistance against Cry1F and Cry1Ab [41]. Studies with genome-edited S. exigua confirmed a crucial role for ABCC2, but not ABCC3 in Cry1F and Cry1Ac toxicity [50]. In B. mori only the double knockout of ABCC2 and ABCC3 conferred Cry1F resistance [51], however, in contrast, the double knockout in FAW was lethal [49].
Recently, a number of studies have linked mutations in the ABCC2 gene to Cry1F resistance in FAW populations [39,40,44]. High levels of field-relevant Cry1F resistance in FAW have been shown to be caused by truncated non-functional ABCC2 receptors in strains from Puerto Rico [39,40], whereas a GY-deletion and P799K in ECL4 has been shown to result in Cry1F resistance in Brazilian FAW populations [44]. This result is

Discussion
There is an emerging body of evidence that ABC-transporters are critical determinants of three-domain (3D) Cry toxin-mediated toxicity and resistance in several lepidopteran species [29,32,33,[49][50][51]. This is also supported by results obtained with ABCC2 knockout lines of S. frugiperda expressing high levels of resistance against Cry1F and Cry1Ab [41]. Studies with genome-edited S. exigua confirmed a crucial role for ABCC2, but not ABCC3 in Cry1F and Cry1Ac toxicity [50]. In B. mori only the double knockout of ABCC2 and ABCC3 conferred Cry1F resistance [51], however, in contrast, the double knockout in FAW was lethal [49].
Recently, a number of studies have linked mutations in the ABCC2 gene to Cry1F resistance in FAW populations [39,40,44]. High levels of field-relevant Cry1F resistance in FAW have been shown to be caused by truncated non-functional ABCC2 receptors in strains from Puerto Rico [39,40], whereas a GY-deletion and P799K in ECL4 has been shown to result in Cry1F resistance in Brazilian FAW populations [44]. This result is supported by another study systematically investigating the crucial role of ECL4 and other ECLs for Cry1F toxicity [45]. However, recently, a whole genome sequencing approach identified a novel mutation in ABCC2 in two Brazilian FAW individuals, which were heterozygous for a 12 bp insertion, leading to a premature stop codon [47]. The data obtained in the present study confirmed the key role of FAW ABCC2 and ABCC3 as Cry1Ab receptors, but did not suggest functional redundancy of ABCC2 and ABCC3 as Cry1F receptors, as for example shown for Cry1F in B. mori [51], or for Cry1A toxins in H. armigera [52]. We observed no membrane permeabilization in Sf9 cells expressing ABCC3 when incubated with Cry1F, whereas Cry1Ab and Cry1A.105 interacted with both ABCC2 and ABCC3 at low nanomolar concentrations.
Cry toxin induced cytotoxicity is often quantified in cell-based assays by the measurement of lactate dehydrogenase (LDH) release indicating necrotic cell death [53]. This is a commonly-used assay to measure membrane permeabilization, along with the microscopic assessment of cell swelling mediated by water influx via aquaporins [53][54][55]. We employed a slightly different approach to determine the effect of ABCC2 mutations on Cry toxinfacilitated pore formation by using SYTOX TM orange, a highly sensitive fluorescent DNA stain, allowing us to quantitatively measure Sf9 cell membrane permeabilization across a wide range of toxin concentrations in 384-well plates.
Our work supplements findings made in a previous study, where pooled population sequencing identified field-derived allelic variations of FAW ABCC2, particularly indels in ECL4, most likely arisen independently as soft selective sweeps [44]. By functional expression of mutant FAW ABCC2 receptors in Sf9 cells, we confirmed the important role of ECL4 for Cry1F interaction and resistance.
A GY-deletion was previously detected at high frequency and was shown to be linked to high levels of Cry1F resistance in a resistant Brazilian field-collected population, Sf_Des [44]. Two more mutations in ABCC2, P799K/R and G1088D, were found in Sf_Des and another Cry1F resistant strain, Sf_Cor [44]. The impact of GY-deletion and P799K on ABCC2-mediated pore formation by Cry1F was confirmed by the functional expression of mutant receptors in Sf9 cells and subsequent cytotoxicity measurements after Cry1F treatment [44], whereas P799R (detected in 14 Brazilian populations), in addition to other ECL4 indels and G1088D (close to NBD2), remained untested. Here, we provided substantial evidence that those ECL4 mutations in FAW ABCC2 not yet tested are highly relevant, resulting in Cry1F resistance levels between 600-and >2500-fold in cellular assays, whereas G1088D (close to NBD2) did not affect Cry1F-induced pore formation in Sf9 cells. In fact, we demonstrated that the deletion of the entire NBD2 domain did not substantially alter the activity of Cry1F and Cry1Ab, a finding in line with earlier results showing that B. mori ABCC2 devoid of NBD2 retained its Cry1Aa receptor activity [56]. In addition, these authors demonstrated, by investigating 29 ABCC2 mutants, that a conserved "DYWL" motif in ECL4 is a determinant for Cry1Aa binding. In a more recent study, S. frugiperda and Mythimna separata ABCC2 ECL4 was divided into three domains, M1-M3, which were shown to differentially contribute to Cry1F selectivity [45]. The authors replaced the different ECL's between species and demonstrated that M. separata ABCC2 expressing the entire coding sequence of FAW ECL4 showed a remarkable increase in Cry1F sensitivity compared to M. separata wildtype ABCC2 [45]. All allelic ABCC2 variants tested here have indels downstream from the conserved "DYWL" motif ( Figure 1), particularly at the interface of the M1 ("TTTDYWLSFWTNQVDGYIQTL") and M2 ("PEGESPNPELDT") regions, supporting the importance for Cry1F interaction of these sequence stretches, as defined by Liu et al. [45]. We also noticed a single polymorphism at position 804 (in M2 of ECL4) where our FAW wildtype ABCC2 sequence [GenBank OL955491] harbours an asparagine in contrast to aspartic acid, reported by Liu et al. [45].
Interestingly, neither the GY-deletion nor the P799K/R amino acid substitution showed a major effect on Cry1Ab and Cry1A.105 toxicity towards Sf9 cells expressing these ABCC2 variants when compared to wildtype ABCC2. Cry toxin proteins have three principle domains (DI-III) and the critical role of domain II for B. mori ABCC2 receptor binding has been recently shown for Cry1Aa [57]. Domain II of Cry1Ab and chimeric Cry1A.105 is identical, but different from Cry1F [48], possibly explaining the observed difference in specificity towards functionally expressed FAW ABCC2 variants mutated in ECL4. Indeed, the EC 50 -values measured for Cry1Ab and Cry1A.105 on different ABCC2 variants significantly coincided, supporting the role of domain II for ABCC2 interaction and/or binding. Additional studies are warranted to identify and better understand the structural determinants of Cry1F selectivity in comparison to other Cry1 toxins, e.g., by mutagenesis of domain II loop regions, resulting in disabled proteins lacking functional features essential for toxicity, such as recently demonstrated with Cry proteins modified in domain I involved in pore formation [58].
Our results indicated that the >400-fold Cry1Ab resistance described in the Brazilian strain Sf_Des [59] is not linked to the presence of GY-deletion and P799K in ABCC2 ECL4. Further research is necessary to investigate if other extracellular loop structures such as ECL2 (described for B. mori ABCC2 [53]), or alternative receptors contribute to Cry1Ab resistance in FAW; except cadherin which was shown to be not involved in Cry1F and Cry1Ab susceptibility in FAW [60]. Indeed, it has been demonstrated that extracellular loops other than ECL4 contribute to Cry1 toxin specificity, such as ECL1 to Cry1Ac, binding to different Spodoptera ssp. ABCC2 transporters [61]. Here, Cry1Ab and Cry1A.105 were much less affected than Cry1F by the investigated indels in ECL4 and still induced pore formation at nanomolar concentrations in Sf9 cells, irrespective of the ABCC2 mutant variant expressed. It is currently unknown if the low-resistance ratios in vitro for Cry1A.105 and Cry1Ab would translate into significant resistance in vivo. Further work is necessary to explore the molecular determinants of Cry1A toxin specificity to FAW ABCC2 and to shed light on potential cross resistance issues between Cry1F and other Cry1 toxins, including Cry1Ab and Cry1A.105.

Conclusions
Our study, employing a baculovirus-free expression system of ABCC2 mutant receptors in Sf9 cells, combined with a SYTOX TM orange cell viability stain, underpinned the importance of ABCC2, particularly ECL4, for Cry1F mediated pore formation and toxicity in FAW. The failure of Cry1F to facilitate cytotoxicity in Sf9 cells expressing FAW ABCC3 suggests the absence of functional redundancy between ABCC2 and ABCC3, which contrasts to the results obtained for Cry1Ab and Cry1A.105. Lack of the NBD2 in FAW ABCC2 hardly affected Cry1F and Cry1Ab toxicity, supporting earlier findings with Cry1A toxins and beet armyworm ABCC2 [62], as well as silkworm ABCC2 [56].

Insect Cell Culture
Sf9 cells were maintained in suspension culture at 27 • C, 120 rpm in Sf900 II culture medium supplemented with 1% foetal bovine serum (FBS) (all from Thermo Fisher Scientific, Waltham, MA, USA), hereafter referred to as cell medium. Cells were passaged twice a week at a density of 4 × 10 5 cells/mL. One day before the transfection, a 4-day culture was split into a density of 1.5 × 10 6 cells/mL.

Expression of ABC-Transporter Variants in Sf9 Cells
For the functional expression of the ABC-transporter variants, Sf9 cells were transiently transfected by electroporation technology using a MaxCyte STx transfection system (MaxCyte, Gaithersburg, MD, USA). The ABCC2 and ABCC3 sequences with a C-terminal 3xFLAG-tag (GenBank accession numbers are stated in Table 1), cloned into pIB/V5-His vector, were purchased from Thermo Fisher Scientific (Waltham, MA, USA). For the transfection, Sf9 cells were harvested by centrifugation (120× g, 5 min, 22 • C) and resuspended in 50:50 electroporation buffer (MaxCyte, Gaithersburg, MD, USA): Sf900 II medium to achieve a density of 1 × 10 8 cells/mL. Cells were mixed with plasmid DNA (100 µg pDNA (in H 2 O)/mL cell suspension) and transferred into the electroporation processing assemblies. Cells were electroporated using the MaxCyte Sf9 protocol. After the electroporation, the cells were transferred into a 6-well plate, mixed with one volume of Sf900 II medium, as well as DNase I (final concentration 0.05 U/µL, Thermo Fisher Scientific, Waltham, MA, USA) and incubated for 30 min at 27 • C, 85% RH. Afterwards, the cells were resuspended in 10 volumes cell medium, and the cell density was adjusted to 3 × 10 5 cells/mL. The cells were seeded into black µCLEAR 384-well plates (Greiner Bio-One, Frickenhausen, Germany) at 15,000 cells/well and incubated for 48 h at 27 • C, 85% RH.

SYTOX TM Orange Cytotoxicity Assay
To determine the toxicity of activated Bt toxins to Sf9 cells expressing ABC-transporter variants a SYTOX TM orange stain was employed [63]. SYTOX TM orange is a highly sensitive fluorescent dye that stains DNA in cells with permeabilized membranes. Trypsin-activated and purified proteins were kindly provided by Kristina Berman (Bayer Crop Science, Chesterfield, MO, USA) and were prepared as previously described [48]. For the cytotoxicity assay, the medium was removed from the cells and replaced with a 25 µL buffer (20 mM HEPES, 5 mM NaHCO 3 , 130 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , pH 7.4 at room temperature) containing activated Bt toxins Cry1F, Cry1Ab or Cry1A.105 in the range of 0.000003 to 1000 nM (Cry1F) and 100 nM (Cry1Ab and Cry1A.105), respectively (four replicates per concentration). The buffer was used as the negative control and 1:

Data Availability Statement:
We will provide all data generated in this study upon request.