insectpeststoparental(Vip3AaandVip3Ca)andchimericproteins.

*Toxins* **2020** , *12*, 99

not statistically different based on the based on the overlap of standard deviation of the mean.

Regarding the chimeric proteins, the exchange of the Nt domain in Vip3Aa (Vip3\_ch1 chimera) decreased the insecticidal activity (compared to Vip3Aa) against all the insect species tested (detected when testing at lower concentrations 0.5 and 0.3 ug/cm2, 0.3 ug/mL and 5 ug/g) except for *S. exigua.* In the case of *A. gemmatalis* and *O. furnacalis*, this chimera completely lost toxicity (Table 1). In the case of the Vip3Ca protein, the exchange of the Nt domain (Vip3\_ch2 chimera) led to different outcomes depending on the insect species considered. Insecticidal activity did not significantly differ from that of Vip3Ca in *S. exigua*, *S. littoralis*, *H. armigera.* The insecticidal activity of the Vip3\_ch2 chimera decreased in *A. gemmatalis*, *M. brassicae*, and *O. furnacalis.* Most interestingly, in the case of *S. frugiperda* the Vip3\_ch2 chimera showed a strong gain of function compared to the Vip3Ca parental protein with mortality values similar to the ones of the most active parental protein, Vip3Aa (Table 1). Chimera Vip3\_ch4 (with the central domain from Vip3Aa and the flanking ones from Vip3Ca) was nontoxic for any of the insect species tested, except for *O. furnacalis* (Table 1). In the case of Vip3\_ch5, the chimeric protein did not cause any damage to any of the insects tested, most likely due to the instability of this protein or problems in the production and purification.

The toxicity of the three proteins active against *O. furnacalis* was confirmed by determining the LC50 values (Table 2 and Table S1). The results indicated that, though similarly toxic, Vip3Ca was the most toxic protein (LC50 = 1.2 μg/g) followed by Vip3\_ch2 (LC50 = 2.3 μg/g) and Vip3\_ch4 (LC50 = 3.9 μg/g). In the case of the chimera with gain of function for *S. frugiperda*, the LC50 value was determined and compared to the most toxic parental protein, Vip3Aa. The results indicated that the toxicity of Vip3\_ch2 (LC50 = 133 ng/cm2) did not significantly differ from that of Vip3Aa (LC50 = 162 ng/cm2), but was significantly increased compared to the Vip3Ca protein (LC50 > 7000 ng/cm2) (Table 1, Table 2 and Table S1).


**Table 2.** Determination of the lethal concentration (LC50) of the parental and selected chimeric Vip3 proteins in *Ostrinia furnacalis* and *Spodoptera frugiperda*.

\* Standard error of the slope and lethal dose concentration, respectively † Confidence interval at 95% for the slope and lethal dose concentration, respectively. ‡ Quantification of the standard deviation of the residuals (vertical distance of the point from the fit line or curve) expressed as % of mortality. At higher value, the data shows a greater variance and lower goodness of fit (R2). ¥ Degree of freedom. <sup>ζ</sup> For each insect species, the LC values followed by the same letter were not statistically different based on of the statistical analysis extra-sum-square F test analysis (α 0.05) (Table S1).

#### **3. Discussion**

Insecticidal proteins in the Vip3A family have been incorporated in commercial transgenic crop varieties [23] due their potent and broad spectrum activity against lepidopteran pests [4]. In contrast, members of the Vip3B and Vip3C protein families have a narrow insecticidal spectrum and a moderate activity [8,21,24–28]. In the case of Vip3Ca, among the 25 species of insects tested [21,25–28], only for *O. furnacalis* and *Mythimna separata* its toxicity was comparable to the most toxic Cry or Vip3 proteins (Cry1Ab for *O. furncacalis* or Vip3Aa for *M. separata*) [26,27]. The present work focused on testing the compatibility of domains exchanged between a member of the Vip3A family and one of the Vip3C family, with the possibility of generating novel proteins with increased insecticidal activity.

The results show that the exchange of the Nt domain does not affect the solubility and trypsin stability of the resulting chimeric Vip3 proteins (Vip3\_ch1 and Vip3\_ch2) (Figure 2). Similar results were obtained in another study testing the exchange of the Nt domain between Vip3Ab and Vip3Bc [8]. This is not a surprising result since the Nt domain is extremely conserved among Vip3 proteins, suggesting a structural role or a possible role in post binding events, such as pore formation. Wang et al. [20] generated a Vip3A protein mutated at the Nt domain (S175C and L177C) which was able to compete with binding of the wild type protein but not cause mortality, thus supporting the previous hypothesis. With regards to the Ct domain, of the four chimeras produced only one was soluble and stable to treatment with trypsin (Figure 2). These results suggest that the interaction of the Ct domain with the other domains in the 3D structure of Vip3 is more specific and critical to the physicochemical properties of the molecule. Furthermore, results from thermofluor assays showed that the chimeric proteins had a thermal stability intermediate between that of the two parental proteins and that the Tm values and the presence/intensity of one or two thermal transitions depended on the interaction between the respective Vip3 domains (Figure S1). Specifically, the denaturation curve profile for the Vip3Aa, Vip3\_ch1, Vip3\_ch2, and Vip3\_ch4 proteins indicates that these proteins have two motifs with different thermal stability, while the Vip3Ca protein would be more stable because of its single denaturation curve. Further understanding of the structure in this family of proteins would shed light on this aspect.

The results from the insecticidal spectrum of the chimeric proteins indicated that, in general (and considering that we only tested one species in the Crambidae family), the activity follows taxonomical relationships at the family level. Thus, species in Noctuidae presented a closer susceptibility profile to both parental and chimeric proteins when compared to the tested species in Crambidae (Table 1). This observation is in agreement with the results of Zack et al. [8], where the Noctuid insects (*Helicoverpa zea*, *S. frugiperda*, and *Pseudoplusia includens*) showed more similar susceptibility profiles for the parental (Vip3Ab and Vip3Bc) proteins and their chimeric proteins (generated by exchange of the Nt domain), compared to the Crambidae insects (*Ostrinia nubilalis*). That study also showed that the chimeras were less toxic than the parental proteins to *H. zea*, *S. frugiperda*, *O. nubilalis*, and *P. includens* [8]. Similarly, our results with the chimeras indicate that, with the exception of Vip3\_ch2, the chimera proteins are similarly or less toxic than the parental proteins (Table 1). The Vip3\_ch2 chimera, a Vip3C protein with the Nt domain from Vip3Aa, displayed gain of function only with *S. frugiperda* but not with other closely related species of the same genus (Table 1). A similar result was recently reported in which a "modified Vip3C protein" (i.e., ARP150v2, 98% similarity to the Vip3\_ch2) had higher toxicity against *S. frugiperda* than the parental Vip3Ca protein [29]. Sequence analysis indicates that ARP150v2 is a chimera in which the Nt domain of Vip3Ca has been replaced by that of Vip3Af1. We do not have a clear explanation for this unique increase in toxicity, but due to the specificity of the phenomenon, the reason has to be more efficient interaction with the receptors and/or facilitated post-binding events, such as membrane insertion or pore formation. Further research testing the mode of action of this family of proteins should clarify this particular phenomenon.

#### **4. Conclusions**

In summary, we present evidence for the relative importance of different Vip3 protein domains in stability and toxicity, and an example of how the design of chimeric Vip3 proteins may lead to novel proteins with improved and expanded insecticidal activity. Specifically, the Vip3\_ch2 protein, a Vip3C protein with the Nt domain from Vip3Aa, showed a gain of function for *S. frugiperda*. In addition, the Vip3\_ch4 protein showed that for the toxicity of the Vip3C protein in *O. furnacalis*, the specificity is provided by the Ct domain.

#### **5. Materials and Methods**

#### *5.1. Design and Construction of Chimeric Vip3 Proteins*

An overlap PCR method was used to generate the chimeric proteins from the parental Vip3Aa45 (JF710269.1) and Vip3Ca2 (JF916462.1) proteins [21,30]. To construct the Vip3A and Vip3C chimeric proteins, amino acids (aa) stretches at positions 188 (188FATET) and 509 (509SRLIT) of the Vip3Aa protein were used to define the protein fragments to exchange: fragment I (aa 0 - 188), fragment II (aa 189 - 508), fragment III (aa 509 - 788) (Figure 1). Six chimeric proteins were generated and classified as "single" (Vip3\_ch1, 2, 5, and 6) or "double" (Vip3\_ch3 and 4), depending on whether they were amplified from the parental or the Vip3\_ch5 and 6 proteins, respectively (Table S2).

To generate the chimeric genes, first a PCR was performed to amplify the necessary fragments separately with the annealing primers (Tables S2 and S3). The PCR reaction contained, in a final volume of 50 μL, 50 ng of the DNA template, 0.25 U of Kapa Hifi DNA polymerase, 5 μL of five-fold reaction buffer, 0.6 mM of each dNTPs, and 0.3 μM of the respective primers. PCR amplifications were carried out as follows: 5 min denaturation at 95 ◦C, 35 cycles of amplification ((20 s of denaturation at 98 ◦C, 15 s of annealing at 60 ◦C, and 30 s of extension at 72 ◦C), and an extra extension step of 5 min at 72◦C). The amplicons were purified form the agarose gel and a second PCR (overlap step + "amplification step") was performed with the respective DNA fragments (Table S2). First, the "overlap step" was conducted in a final volume of 50 μL with 100 ng (total amount) of the respective DNA fragments (Table S2) in an equimolecular ratio, 0.25 U of Kapa Hifi DNA polymerase, 5 μL of five-fold reaction buffer, 0.6 mM of each dNTPs. PCR amplifications were carried out as follows: 5 min denaturation at 95 ◦C, 15 cycles of amplification ((20 s of denaturation at 98 ◦C, 30 s (Vip3\_ch1, 2, 4, 5, and 6)/1 min (Vip3\_ch3) of annealing at 55 ◦C (Vip3\_ch1, 2, 4, 5, and 6) or 50 ◦C (Vip3\_ch3), 2 min of extension at 72 ◦C) and an extra extension step of 5 min at 72 ◦C). Second, the "amplification step" was performed with the respective end primers (Table S2), adding 0.3 μM of each to the PCR reactions. The PCR reactions were carried out in the conditions described for the "overlap step". In addition, for the Vip3\_ch3 protein a nested-PCR with the DNA amplified in the second PCR was carried out (PCR reaction: final volume 50 μL, 7 ng of the Vip3 chimera 3, 0.25 U of Kapa Hifi DNA polymerase, 5 μL of five-fold reaction buffer, 0.6 mM of each dNTPs, and 0.3 μM of the respective primers (Table S2). Conditions for this nested-PCR amplification were 5 min denaturation at 95 ◦C, 35 cycles of amplification (20 s of denaturation at 98 ◦C, 60 s of annealing at 50 ◦C, and 2 min of extension at 72 ◦C), and an extra extension step of 5 min at 72 ◦C). Amplicons were purified from an agarose gel, ligated into the pGEM®-T Easy plasmid or pCR®2.1-TOPO®, cloned in *E. coli* DH10β, and sequenced with the sequencing primers (Table S3).

For expression, the full length genes were amplified from the pGEM®-T Easy or pCR®2.1-TOPO® with the end primers (Table S3). The PCR reactions contained, in a final volume of 50 μL, 50 ng of the respective Vip3 chimeric genes, 0.25 U of Kapa Hifi DNA polymerase, 5 μL of five-fold reaction buffer, 0.6 mM of each dNTPs, and 0.3 μM of the respective end primers (Table S2). Conditions for PCR amplifications were as follows: 5 min denaturation at 95 ◦C, 35 cycles of amplification ((20 s of denaturation at 98◦C, 60 s of annealing at 55 ◦C (Vip3\_ch1, 2, 4, 5, and 6) 50 ◦C (Vip3\_ch3), 2 min of extension at 72 ◦C), and an extra extension step of 5 min at 72 ◦C). Amplicons were purified and together with the expression vector (pET30a (+)) were digested with BamHI and NotI for 2 h at 37 ◦C. The pET30a (+) plasmid was dephosphorylated for 2 h at 37 ◦C with alkaline phosphatase. The linearized/dephosphorylated pET30a (+) and the digested chimeric genes were purified prior to ligation using T4 DNA Ligase overnight at 4 ◦C. Ligation reactions were transformed in *E. coli* BL21 (D3) and transformants confirmed by sequencing with the sequencing primers (Table S3).

#### *5.2. Expression and Purification of Vip3Aa, Vip3Ca, and Chimeric Proteins*

#### 5.2.1. Expression of the Parental and Chimeric Vip3 Proteins

The Vip3Ca protein was expressed following the conditions described by Gomis-Cebolla et al. (2017) [16]. For expression of Vip3Aa and the chimeric proteins, a single colony was inoculated in 7 mL of LB-K medium (LB medium containing 50 μg/mL kanamycin) and grown overnight at 37 ◦C and 180 rpm. A 1/100 dilution of the culture in 700 mL LB-K medium was further incubated at 37 ◦C and 180 rpm. When the OD was 0.7–0.8, 1 mM IPTG (final concentration) was added for induction. Induced cultures were grown overnight at 37 ◦C and 180 rpm, and the cells were collected by centrifugation at 8800× *g* for 30 min at 4 ◦C. Cell pellets for the Vip3Aa and Vip3Ca proteins were lysed by chemical lysis. Briefly, three milliliters of lysis buffer-I (50 mM sodium phosphate buffer, 0.5 M NaCl, pH 8.0,

containing 3 mg/mL lysozyme, 10 μg/mL DNase, 10 mM DTT, and 100 μM PMSF) per gram of pellet were added to the samples. The pellets were resuspended with an Ultra Turrax T25 digital homogenizer (IKA, Janke & Kunkel-Str. 10 Staufen, DE) at 16,000× *g* and incubated at 37 ◦C for 60 min with strong shaking (200 rpm). Then, the lysate was sonicated on ice applying five cycles (1 min pulse at 70 W, 10 s off, 1 min pulse at 70 W). Insoluble materials were separated by centrifugation at 31,000× *g* for 15 min and 4 ◦C. The soluble cellular fractions were filtered through sterile 0.45 μm cellulose acetate filters. In the case of chimeric proteins, three milliliters of lysis buffer-II (50 mM sodium phosphate buffer, 0.5 M NaCl, pH 8.0, containing 3 mg/mL lysozyme, 10 mM DTT and 100 μM PMSF) per gram were added to the pellets and the samples were resuspended as described above, and then incubated at 37 ◦C for 60 min with strong shaking (200 rpm). After incubation, 8 mg of deoxycholic acid sodium salt per gram of pellet was added and incubated 30 min at 37 ◦C with gentle shaking (100 rpm), after which 40 μg/mL of DNase was added to eliminate the viscosity of the lysates and further incubated for 30 min at 37 ◦C with gentle shaking. The lysates of the chimeric proteins were then sonicated on ice applying five cycles (1 min pulse at 70 W, 10 s off, 1 min pulse at 70 W), centrifuged at 31,000× *g* for 15 min at 4 ◦C and the soluble cellular fraction was filtered through sterile 0.45 μm cellulose acetate filters. In the case of the Vip3 chimeras 3 and 6, they formed inclusion bodies that were not possible to dissolve in the conditions used in the present study (Figure S2).

#### 5.2.2. Purification of Vip3Aa, Vip3Ca, and Chimeric Vip3 Proteins by Isoelectric Point Precipitation

For the insect toxicity assays, two independent batches of the Vip3Aa, Vip3Ca and the chimeric proteins (Vip3\_ch1, Vip3\_ch2, Vip3\_ch4, and Vip3\_ch5), were purified by isoelectric point precipitation (IPP) in three steps (Figure 4A) [31]. First, the soluble cellular fractions of the Vip3 proteins were diluted three-fold with 50 mM sodium phosphate buffer pH 8.0, dialyzed overnight against the dialysis buffer (20 mM sodium phosphate buffer, 150 mM NaCl, pH 8), centrifuged at 14,000× *g* for 15 min at 4 ◦C, and then filtered through 0.45 μm cellulose acetate filters. Second, the pH of the respective lysates was lowered with acetic acid to pH 5.5 for Vip3Aa, pH 5.9 for Vip3Ca, pH 5.0 for Vip3\_ch1, pH 5.5 for the Vip3\_ch2, pH 5.2 for the Vip3\_ch4, pH 5.2 for the Vip3\_ch5. The Vip3 proteins were recovered by centrifugation (14,000× *g* for 15 min at 4 ◦C). Third, the pellets were resuspended in storage buffer (20 mM Tris buffer, 150 mM NaCl pH 8.6) for 1 h with shaking at 4 ◦C, and then centrifuged (14,000× *g* for 15 min at 4 ◦C) and filtered through 0.45 μm cellulose acetate filters. The Vip3 proteins were quantified by densitometry and the ratio of Vip3 protein/total protein (w:w) was calculated. The samples were stored at −80 ◦C and lyophilized prior to their use or shipping at room temperature to other laboratories.

**Figure 4.** SDS-PAGE analysis of the purified parental (Vip3Aa and Vip3Ca) and chimeric (Vip3\_ch1, Vip3\_ch2, Vip3\_ch4, and Vip3\_ch5) proteins. (**A**) Parental proteins and chimeric Vip3 proteins (5 μg) purified by isoelectric point precipitation (IPP). (**B**) Parental proteins and chimeric Vip3 proteins (5 μg) purified by ion metal affinity chromatography (IMAC) on a Hi-Trap chelating HP column (GE Healthcare). (**C**) Vip3\_ch5 protein (2 μg) purified by IPP and IMAC on a Hi-Trap chelating HP column (GE Healthcare). The arrowheads indicate the protein band corresponding to the chimeric Vip3 proteins. M1: Molecular Mass Marker.

5.2.3. Purification of Vip3Aa, Vip3Ca, and Chimeric Vip3 Proteins by Ion Metal Affinity Chromatography

For the proteolysis and thermal shift assays, the parental proteins (Vip3Aa and Vip3Ca) and the chimeric Vip3 proteins (Vip3\_ch1, Vip3\_ch2, Vip3\_ch4, and Vip3\_ch5) were purified by ion metal affinity chromatography (IMAC) on a His-Trap FF crude lysate column (GE Healthcare) (Figure 4B). The soluble cellular fractions of the Vip3 proteins (Vip3Aa, Vip3Ca, Vip3\_ch1, 2, 4, and 5) were diluted three-fold with 50 mM sodium phosphate buffer pH 8.0, dialyzed against the dialysis buffer to eliminate the presence of DTT and deoxycholic acid. Samples were dialyzed for 10-16 h at 4 ◦C, and the dialysis buffer exchanged twice. The unclarified lysates were filtered through 0.45 μm cellulose acetate filters to discard protein aggregates. The soluble Vip3 protein fractions were loaded into a His-Trap FF crude lysate column equilibrated in binding buffer (20 mM phosphate buffer, 150 mM NaCl, 10 mM imidazole, pH 8). After washing the column with binding buffer to eliminate unbound molecules, Vip3 proteins were eluted using elution buffer (20 mM phosphate buffer, 150 mM NaCl, 150 mM imidazole, pH 8) into 2 mL tubes containing 0.1 mM EDTA (pH 8.0).

Since the Vip3\_ch5 was expressed in low quantities (data not shown), first the protein was partially purified by IPP as described above, and then filtered through 0.45 μm cellulose acetate filter, prior to loading into the His-Trap FF crude lysate column (Figure 4C).

To avoid protein precipitation, buffer exchange was performed against storage buffer (20 mM Tris 500 mM NaCl, pH 8.6) by dialysis overnight at 4 ◦C. The concentration and quality of the purified proteins were estimated with the Bradford assay [32] using BSA as standard and by SDS-PAGE, respectively. The Vip3 proteins were snap frozen in liquid nitrogen and stored at −80 ◦C until used.

#### *5.3. Thermal and Protease Stability of the Parental and the Chimeric Vip3 Proteins*

The parental (5 μg) and chimeric (5 μg of Vip3\_ch1, Vip3\_ch2, Vip3\_ch4, and 2 μg of Vip3\_ch5) proteins were subjected to proteolysis with 1% (w/w) bovine trypsin (SIGMA T8003, Sigma-Aldrich, Madrid, Spain) for different time intervals (0, 0.5, 1, and 2 h) at 37 ◦C. The proteolytic reactions were stopped with 1 mM of AEBSF protease inhibitor for 10 min at room temperature, and then the samples were resolved by SDS-12%PAGE and stained with Coomassie brilliant blue R-250 (Sigma 1125530025, Sigma-Aldrich, Madrid, Spain). The size of the protoxin and trypsin-activated fragments were analyzed using the TotalLab 1D v 13.01 software.

The Tm of the parental proteins and the chimeric proteins resistant to trypsin treatment were determined using the environmentally sensitive extrinsic dye SYPRO-Orange [33]. The thermal shift reactions were prepared at room temperature (RT) and contained, in a final volume of 180 μL, 1 μM of the respective Vip3 proteins (filtered through 0.45 μm cellulose acetate filter), 15× of SPYRO-Orange (diluted in storage buffer-I) and storage buffer-I up to 180 μL. Eight replicates (20 μL) of the parental and chimeric proteins plus a negative control (15× of SPYRO-Orange and storage buffer-I) were incubated for 5 min at RT and centrifuged at 141× *g* for 1 min prior to analysis with the StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA). Thermal shift assays program was carried out as follows: Reporter: ROX; Passive Reference: None; Run Method: Mode Continuous, Program 2 min at 25 ◦C, Temperature Ramp 4% (4 ◦C/min), 2 min at 99 ◦C. The data were exported to an Excel file to determine the Tm of the respective Vip3 proteins by plotting the negative of the first derivative of the fluorescence as a function of temperature-dFv/dT, where Fv and T at (t+1)-t represent the increment of fluorescence and temperature between each measurement.

$$-\mathbf{d}\mathbf{F}\mathbf{v}/\mathbf{d}\mathbf{T} = (\mathbf{F}\mathbf{v}(\mathbf{t}+\mathbf{1}) - \mathbf{F}\mathbf{v}(\mathbf{t})) / (\mathbf{T}(\mathbf{t}+\mathbf{1}) - \mathbf{T}(\mathbf{t})) \tag{1}$$

The Tm values of the respective Vip3 proteins were compared with one-way ANOVA analysis and datasets statistically significant (α < 0.05), were analyzed by the multiple comparison Tukey post hoc test (α < 0.05).

#### *5.4. Insect Colonies and Toxicity Assays*

Insects were reared and bioassays performed at the insectaries of the University of Valencia (for *S. exigua* and *S. littoralis*, Spain), Public University of Navarra (for *H. armigera, M. brassicae*, Spain), University of Tennessee (for *S. frugiperda* and *A. gemmatalis*, Knoxville, TN, USA), and Chinese Academy of Agricultural Sciences (for *O. furnacalis*, Haidan district, Beijing, China) at 25 ◦C, 70% RH, 16:8 L/D photoperiod (*S. exigua*, *S. frugiperda*, *S. littoralis*, *M. brassicae*, *H. armigera*, and *A. gemmatalis*) and 27 ◦C, 80% RH, 16:8 h L/D photoperiod (*O. furnacalis*), respectively. The insect colonies of *S. exigua, S. littoralis*, *H. armigera*, *M. brassicae*, and *O. furnacalis* had been reared for several generations in laboratory conditions without exposure to any insecticide. In the case of *S. frugiperda* and *A. gemmatalis* the insects were purchased from Benzon Research Inc. (Carlisle, PA, USA). The laboratory insect colonies of *S. exigua*, *S. littoralis*, *M. brassicae*, and *H. armigera* were reared in a growth wheat germ-based semi-synthetic diet [34], while *O. furnacalis* had been reared using standard rearing techniques without exposure to any insecticide [35]. In the case of *S. frugiperda* and *A. gemmatalis*, they were reared with meridic diet (#F9772, Frontier Agricultural Sciences, Newark, DE, USA). The same diets and rearing conditions were used in the bioassays with the parental proteins and chimeric Vip3 proteins.

Different methodologies were used in the bioassays depending on the insect species tested. For *S. exigua*, *S. littoralis*, *S. frugiperda*, *H. armigera*, and *A. gemmatalis*, bioassays were performed on neonates using surface contamination. Briefly, two pairs of different concentrations were dispensed on the diet surface. Prior to the sample application, the surface of the diet was sterilized under UV light for 10 min. A volume of 50–75 μL of each concentration was applied on the surface of solidified diet (2 cm2 multiwell plates, Bio-Cv-16, C-D International) and let dry in a flow hood. Once dried, one larva was transferred to each well using a brush. In the case of *O. furnacalis*, bioassays were performed on neonates using diet incorporation assays [36], while for *M. brassicae* bioassays were performed on L2 instar larvae using a droplet feeding method [37]. To determine the effect of the domain exchange on toxicity, bioassays with Vip3 proteins were performed with two different concentrations (chosen as to give a discriminant mortality, range of mortality for each insect species between 1% and 99%) in at least two different insect generations (Table 1). Thirty-two neonates were used for each protein concentration for *S. exigua*, *S. littoralis*, *S. frugiperda*, and *A. gemmatalis*; 28 in the case of *H. armigera* and

*M. brassicae*; for *O. furnacalis* 42 neonates were tested. Mortality (number of dead larvae) was scored after 7 days for *S. exigua*, *S. littoralis*, *S. frugiperda*, *H. armigera*, *M. brassicae*, and *O. furnacalis*; while for *A. gemmatalis* mortality was scored after 5 days. Only data from bioassays with <10% control mortality were considered.

Determination of the LC50 (concentration of protein killing 50% of tested individuals) for the toxic parental and chimeric proteins was done for *O. furnacalis* (concentration range 0.04–50 μg/g) and *S. frugiperda* (concentration range 0.01–3 μg/cm2). For *S. frugiperda* a set of 16–32 neonates per concentration (7–8 concentrations of the respective Vip3 proteins) were tested under the same conditions as described above for bioassays, and bioassays replicated twice. The number of dead larvae was recorded after 7 days of exposure. In the case of *O. furnacalis*, neonates were introduced to individual wells of 48-well trays containing 9–11 concentrations of purified toxin, which were tested with a total of 96 larvae per concentration. Trays were incubated as per the rearing conditions above and mortality and survivor weight were recorded after 7 days of exposure. Bioassays were repeated with two insect generations. The storage buffer was used to dilute the parental and chimeric Vip3 proteins and as negative control. Bioassay data were subjected to nonlinear regression using the software GraphPad Prism7 to obtain the LC50 of the parental proteins and chimeric Vip3 proteins, which were compared the parental proteins vs. the chimeric proteins with the statistical analysis extra-sum-square *F* test (α 0.05) (Table S1).

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6651/12/2/99/s1, Figure S1. Thermal shift assays and multiple comparison of the thermal transitions of the parental and chimeric proteins. The dashed vertical lines in the thermal shift assays curves indicate the Tm (measured in ◦C) of respective thermal transitions. C- indicates the fluorescence intensity due to the SPYRO-Orange 15× in 20 mM Tris 500 mM NaCl pH 8.6. The thick line indicates the comparison of the Tm by one-way ANOVA (α 0.05). The dashed line indicates the multiple comparison analyzed by Tukey's range test (α 0.05). "\*\*\*\*" indicates a p value less than 0.0001 and "ns" indicates a p value greater than 0.05. Figure S2. Expression of the chimeric Vip3 proteins (Vip3\_ch3 and Vip3\_ch6). (A) SDS-PAGE of different dilutions of the pellet and supernatant of the Vip3\_ch3 and Vip3\_ch6 proteins. (B) Western blot analysis of different dilutions of the pellet and supernatant of the respective chimeric Vip3 proteins. The dilutions of the lysates were made with 50 mM phosphate buffer, 500 mM NaCl (pH 8.0), while the pellets were dissolved in the same volume of the supernatant and the dilutions were made with 50 mM phosphate buffer, 500 mM NaCl (pH 8.0). The arrowhead indicates the protein band corresponding to the chimeric Vip3 proteins. M1: Molecular Mass Marker "PINK Plus Prestained Protein Ladder" (Genedirex). M2: Molecular Mass Marker "Precision Plus Protein™ Dual Color Standards" (Biorad) developed with "Precision Protein™ Strep Tactin-HRP conjugate. Availability of data and material: Sequences encoding the Vip3 chimeras have been deposited in GenBanK with the following accession numbers: Vip3 chimera 1 (MH363727), Vip3 chimera 2 (MH363728), Vip3 chimera 3 (MH363729), Vip3 chimera 4 (MH363730), Vip3 chimera 5 (MH363731), and Vip3 chimera 6 (MH363732). Table S1. Comparison analyses of the respective dose-response assays (LC50 values) of the parental and chimeric Vip3 proteins in S. frugiperda and O. furnacalis. Table S2. Construction of the chimeric Vip3 proteins from the Vip3Aa and Vip3Ca proteins. Table S3. Primers used in construction and sequencing of the genes encoding the chimeric Vip3 proteins.

**Author Contributions:** J.F. and J.G.-C. conceived and designed the experiments. J.G.-C., R.F.d.S., Y.W. and J.C. performed the experiments. J.G.-C., and J.F. analyzed the data. J.G.-C., J.L.J.-F., P.C., K.H. and J.F. wrote the paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported to JF by the Spanish Ministry of Economy and Competitivity (Grants Ref. AGL2015-70584-C02-1-R and RTI2018-095204-B-C21), by the Generalitat Valenciana (GVPROMETEOII-2015-001), and by European FEDER funds. JGC was recipient of a PhD grant from the Spanish Ministry of Economy and Competitivity (grant ref. BES-2013-065848 and EEBB-I-17-12367). Support was also provided to JLJ-F by an Agriculture and Food Research Initiative Foundational Program competitive grant (No. 2018-67013-27820) from the USDA National Institute of Food and Agriculture, and to KH by a grant "Key Project for Breeding Genetically Modified Organisms" from China (grant ref. 2016ZX08003-001).

**Acknowledgments:** We thank R. González-Martínez for their help in rearing the insect colonies.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

1. Palma, L.; Muñoz, D.; Berry, C.; Murillo, J.; Caballero, P. *Bacillus thuringiensis* toxins: An overview of their biocidal activity. *Toxins* **2014**, *6*, 3296–3325. [CrossRef]


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## **Reduced Membrane-Bound Alkaline Phosphatase Does Not A**ff**ect Binding of Vip3Aa in a** *Heliothis virescens* **Resistant Colony**

#### **Daniel Pinos 1, Maissa Chakroun 1,**†**, Anabel Millán-Leiva 1, Juan Luis Jurat-Fuentes 2, Denis J. Wright 3, Patricia Hernández-Martínez <sup>1</sup> and Juan Ferré 1,\***


Received: 7 May 2020; Accepted: 17 June 2020; Published: 19 June 2020

**Abstract:** The Vip3Aa insecticidal protein from *Bacillus thuringiensis* (Bt) is produced by specific transgenic corn and cotton varieties for efficient control of target lepidopteran pests. The main threat to this technology is the evolution of resistance in targeted insect pests and understanding the mechanistic basis of resistance is crucial to deploy the most appropriate strategies for resistance management. In this work, we tested whether alteration of membrane receptors in the insect midgut might explain the >2000-fold Vip3Aa resistance phenotype in a laboratory-selected colony of *Heliothis virescens* (Vip-Sel). Binding of 125I-labeled Vip3Aa to brush border membrane vesicles (BBMV) from 3rd instar larvae from Vip-Sel was not significantly different from binding in the reference susceptible colony. Interestingly, BBMV from Vip-Sel larvae showed dramatically reduced levels of membrane-bound alkaline phosphatase (mALP) activity, which was further confirmed by a strong downregulation of the membrane-bound alkaline phosphatase 1 (*HvmALP1*) gene. However, the involvement of HvmALP1 as a receptor for the Vip3Aa protein was not supported by results from ligand blotting and viability assays with insect cells expressing *HvmALP1*.

**Keywords:** *Bacillus thuringiensis*; insecticidal proteins; insect resistance; tobacco budworm

**Key Contribution:** The biochemical characterization of a Vip3Aa-resistant colony of *H. virescens* shows that binding to receptors in the midgut is not affected and discards the role of mALP as a Vip3Aa receptor. This study suggests that Vip3A resistance may occur through mechanisms other than those commonly found for Cry proteins.

#### **1. Introduction**

The polyphagous pest *Heliothis virescens*(L.) (Lepidoptera: Noctuidae) is well known for producing substantial economic losses, particularly in cotton production, due to its ability to evolve resistance to different synthetic control products such as methyl parathion or pyrethroids [1,2]. As an alternative approach, genetically modified crops expressing Cry and Vip3A insecticidal protein genes from *Bacillus thuringiensis* (Bt crops) were introduced in 1996 for the control of this and other pests. However, extensive use threatens their effectiveness and cases of field-evolved practical resistance have already been reported for some lepidopteran and coleopteran pests [3].

Gene pyramiding has been proposed as an effective strategy for insect resistance management in Bt crops [4]. This approach consists of combined production of distinct insecticidal Bt proteins in the same plant, and its success heavily relies on the expressed insecticidal proteins having distinct mode of action, commonly defined as not sharing binding sites in target tissues [5,6].

Although the mechanism of action and receptors for Cry proteins have been widely studied [7], little is known about the biochemical mechanisms that underlie the action of Vip3A proteins. Several studies have shown that Vip3A proteins do not share binding sites with Cry1 or Cry2 proteins, yet their damage to the midgut epithelium resembles Cry action [8–11]. Supported by the lack of shared binding sites, transgenic corn and cotton varieties pyramided with Cry1, Cry2, and Vip3A genes are currently commercialized in several countries.

Knowledge of the biochemical and genetic factors involved in resistance is crucial to design management practices that delay the appearance of resistance and allow its rapid detection and ways to overcome it. The genetic potential to evolve resistance to Vip3A has already been shown in some laboratory-selected insect species such as *H. virescens* [12], *Spodoptera litura* [13], *Helicoverpa armigera* and *Helicoverpa punctigera* [14], *Spodoptera frugiperda* [15,16], and *Helicoverpa zea* [17]. However, the biochemical basis of resistance to Vip3A has only been studied in a laboratory-selected colony of *H. armigera*, for which alteration of binding sites was not the cause of resistance [18].

In the present study, we aimed to determine the biochemical basis of >2000-fold resistance to Vip3A in a *H. virescens* colony (Vip-Sel). In a previous study with this colony, resistance was shown to be polygenic, conferring little cross-resistance to Cry1Ab and no cross-resistance to Cry1Ac [19]. A transcriptomic analysis detected significant differences in gene expression compared to a susceptible strain, with 420 over-expressed and 1569 under-expressed genes in Vip-Sel [20]. Results herein support that Vip3Aa binding is not significantly altered in Vip-Sel compared to susceptible *H. virescens* and that membrane bound alkaline phosphatase (mALP) is not involved in Vip3Aa binding.

#### **2. Results**

#### *2.1. Vip3Aa Binding to Midgut Brush Border Membrane Vesicles (BBMV)*

In testing whether binding of Vip3Aa was altered in larvae from the Vip3A-resistant (Vip-Sel) compared to the reference susceptible (Vip-Unsel) colony, we measured binding of radiolabeled Vip3Aa to BBMV from the two colonies. Binding analyses showed specific Vip3Aa binding for BBMV from both colonies, with similar homologous competition curves (Figure 1a). A high percentage (35–40% of the input labeled toxin) of non-specific binding, i.e., not blocked by high concentrations of unlabeled Vip3Aa competitor, was detected, as previously reported [11,18]. The *K*<sup>d</sup> and *R*<sup>t</sup> values estimated from the competition curves (Table 1) indicated that Vip3Aa binds with low affinity to a high number of binding sites in BBMV from *H. virescens*. No major differences were found for these equilibrium binding parameters between the two *H. virescens* colonies, suggesting that binding alteration is not mechanistically related to Vip3Aa resistance in Vip-Sel.

**Figure 1.** Analysis of 125I-Vip3Aa binding to BBMV from susceptible (Vip-Unsel) and resistant (Vip-Sel) colonies of *H. virescens*. (**a**) Homologous competition binding assays of BBMV from the two colonies with 125I-Vip3Aa, using increasing concentrations of unlabeled Vip3Aa as a competitor. Each data point represents the mean of two replicates performed in technical duplicates (±SEM). (**b**) Ligand blot of BBMV proteins from Vip-Unsel and Vip-Sel colonies probed with Vip3Aa. Lane M, protein molecular weight marker (in kDa) (Precision Plus Protein ™ Dual Color Standards, Bio-Rad, St. Louis, MO, USA). The black arrow indicates expected molecular weight of mALP (*ca*. 66 kDa).

**Table 1.** Equilibrium *K*<sup>d</sup> (dissociation constant) and *R*<sup>t</sup> (concentration of binding sites) binding parameters estimated from Vip3Aa homologous competition assays with BBMV from resistant (Vip-Sel) and susceptible (Vip-Unsel) *H. virescens* insects.


<sup>1</sup> Values are the mean of two replicates. <sup>2</sup> Values are expressed in picomoles per milligram of BBMV protein.

#### *2.2. Reduced ALP Levels in the Vip3Aa-Resistant H. virescens Colony*

During the evaluation of BBMV quality, we determined and compared the specific activities of alkaline phosphatase (ALP) and aminopeptidase-N (APN) as brush border membrane marker enzymes in midgut homogenates and BBMV preparations from Vip-Unsel and Vip-Sel colonies (Figure 2). The specific APN activity in midgut homogenates from both colonies was around 12 mU/mg, while in the BBMV preparations it was around 70 mU/mg, indicating an enrichment of APN activity of around 5.8 folds. Importantly, no significant differences (Student's *t*-test, *p* > 0.05) in APN activity were observed between the midgut homogenates or BBMV from Vip-Unsel and Vip-Sel colonies. In agreement with the 5.8-fold enrichment value from APN activity comparisons, specific ALP activity was 7.44 mU/mg in midgut homogenates and 42.5 mU/mg in the BBMV from the Vip-Unsel colony. In contrast, dramatically reduced ALP activity was detected in both midgut homogenate (1.15 mU/mg) and BBMV (1.88 mU/mg) samples from the Vip-Sel colony. While unexpected, this observation is in line with reports of reduced ALP levels in Cry1-resistant lepidopteran species, including *H. virescens* [21–24]. Consequently, we further explored the extremely reduced ALP activity in Vip-Sel to determine whether it was due to a loss of enzymatic function or reduced gene expression.

**Figure 2.** Enzymatic activities in homogenates and BBMV from the two colonies of *H. virescens* (dashed-grey bars: Vip-Unsel; grey bars: Vip-Sel). Each bar represents the mean of three replicates (±SEM). Asterisks represent significant difference (Student's *t-*test, \*\*\*\* *p* < 0.0001).

Electrophoretic comparison of BBMV proteins from the two *H. virescens* colonies showed a protein band of ~66 kDa for the Vip-Unsel colony that was almost imperceptible in the BBMV from the Vip-Sel colony (Figure 3a). Western blotting indicated the presence of ALP in the ~66-kDa protein band, and confirmed the highly reduced levels of this protein in the Vip-Sel colony (Figure 3b). The composition of the ~66-kDa protein band and its relative abundance in the two *H. virescens* colonies were determined by liquid chromatography coupled to mass spectrometry (LC-MS) analysis. The spectra for the most abundant protein detected and identified in the ~66-kDa band matched to membrane-bound alkaline phosphatase (mALP) from *H. virescens* (Genbank Accession No. ABR88230). According to the exponentially modified protein abundance index (emPAI) expressing the proportional protein content in a protein mixture, the abundance ratio of mALP between Vip-Unsel and Vip-Sel was 22.7 folds.

**Figure 3.** Analysis of membrane ALP levels in the susceptible (Vip-Unsel) and resistant (Vip-Sel) colonies of *H. virescens*. (**a**) Protein gel electrophoresis (SDS–PAGE) of BBMV from the two colonies. (**b**) Western blot performed with anti-ALP antibody against BBMV from the two colonies. The black arrow indicates mALP (ca. 66 kDa). Lane M, protein marker (molecular weight in kDa). (**c**) Membrane ALP expression levels in Vip-Sel colony using transcript levels in Vip-Unsel colony as a reference. Fold changes calculated by REST-MCS Software. Bars represent the mean of three independent experiments (±SD, \* *p* < 0.05).

To test if the reduced mALP protein levels in Vip-Sel were controlled at the transcriptional level, we performed real-time quantitative PCR (RT-qPCR) with mRNA extracted from total RNA from the two colonies. Transcript levels for two *H. virescens* mALP genes, *HvmALP1* (Accession No. FJ416470.1) and *HvmALP2* (Accession No. FJ416471.1), were analyzed. Compared to insects from the Vip-Unsel colony, larvae from the Vip-Sel colony had significant (*p*-value < 0.05) nine-fold downregulation of the *HvmALP1* gene, while transcript levels for *HvmALP2* were not different between colonies (Figure 3c). These results support that reduced ALP enzyme activity in BBMV from Vip-Sel compared to Vip-Unsel is due to reduced expression of *HvmALP1* in the Vip-Sel colony.

#### *2.3. Functional Role of HvmALP1 in Vip3Aa Binding*

Since *H. virescens* ALP was proposed to play a role in binding of Cry1 proteins to the midgut membrane [25], we used ligand blotting to test whether mALP was involved in Vip3Aa binding. Binding of Vip3Aa to blots of resolved BBMV proteins was detected with anti-Vip3Aa antisera. No differences in the Vip3Aa-binding band pattern were detected between both colonies, in agreement with the binding results with radiolabeled Vip3Aa. However, no Vip3Aa binding was observed at the mALP position (~66 kDa) (Figure 1b).

To further discard mALP as a functional Vip3Aa receptor, we cloned and transiently expressed the *HvmALP1* gene in cultured (Sf21) insect cells and performed cell viability tests after challenge with Vip3Aa. Transfection was successful, as transfected cells showed ~5-fold increased specific ALP activity compared with non-transfected cells or cells transfected with the empty plasmid (Figure 4a). However, after a challenge with Vip3Aa, the viability of transfected cells was not significantly different (Student's *t*-test; *p* > 0.05) from that of the control cells (Figure 4b), confirming that mALP does not serve as a functional receptor for Vip3Aa during the toxicity process.

**Figure 4.** Specific ALP enzymatic activity and viability assays of Sf21 cells producing the HvmALP1 isoform. (**a**) Alkaline phosphatase enzymatic activity on non-transfected cells (empty bars), cells transfected with empty plasmid (grey bars, and plasmid with *HvmALP1* (dashed-grey bars). (**b**) Cell viability after 24 h of Vip3Aa intoxication (300 μg/mL final concentration) on the same three cell types. Each value represents the mean (±SEM). Means were compared by Student's *t*-test (\*\* *p* < 0.01).

#### **3. Discussion**

The use of resistant insect strains isolated from the field or selected in the laboratory has been a powerful tool to understand the biochemical and genetic bases of resistance to Bt insecticidal proteins. Many studies have found that the alteration of membrane receptors is a common mechanism conferring high levels of resistance to Cry proteins [26–28]. In the case of Cry1 proteins, an important body of literature identifies aminopeptidase N, ABC transporters, cadherins and membrane alkaline phosphatases as main receptors, and identifies their alterations in association with resistance [29,30]. In contrast, three candidate receptors have been proposed for Vip3A proteins, including the *Spodoptera spp.* ribosomal protein S2 [31], the fibroblast growth factor receptor-like

protein [32] and the scavenger receptor class C-like protein [33], yet their role in resistance has not been established.

In the present work, we aimed to determine whether alteration of membrane receptors was the basis for the observed 2040-fold resistance to Vip3Aa in the Vip-Sel colony of *H. virescens*. Results from binding assays with BBMV and radiolabeled Vip3Aa did not detect significant differences between the susceptible and resistant colonies, suggesting no involvement of binding site alteration in resistance. This conclusion was further supported by results from ligand blotting, where no differences between the binding patterns of Vip3Aa to BBMV proteins from the two colonies were observed. Similar results were reported for a laboratory-selected Vip3A-resistant colony of *H. armigera* [18], suggesting that high levels and narrow spectrum of Vip3A resistance may develop by mechanisms other than alteration of Vip3Aa binding sites.

Even though differences in binding were not found, a dramatic reduction in the ALP enzymatic activity was detected in midgut samples from the resistant compared to susceptible colony. Western blotting and RT-qPCR analyses showed that the decreased activity was due to a reduction in the amount of mALP protein, which was controlled at the transcriptional level, in agreement with a previous study [20]. Downregulation or reduced levels of mALP in the midgut membrane have been observed as a common phenomenon in resistance to Cry1Ac in *H. virescens* [25], *Helicoverpa zea* [21], *Plutella xylostella* [22], and *Helicoverpa armigera* [24]; to Cry1F in *S. frugiperda* [23]; to Cry1C in *Spodoptera litura* [34]; and even in *Aedes aegypti* resistant to Bt subsp. *israeliensis* (Bti) [35]. The fact that Cry1Ac and Cry1C do not share binding sites [36] suggests that the role of ALP downregulation in resistance may not be related to reduced Cry binding, but may represent a physiological response to resistance. In agreement with this hypothesis, susceptibility of Sf21 cells expressing HvmALP1 was not significantly different to Vip3Aa, supporting that ALP is not a functional receptor for Vip3Aa in *H. virescens*. In addition, in a Cry1Ac-resistant strain of *P. xylostella*, altered expression of different genes (including the *PxmALP*) was reported to be *trans*-regulated by upregulation of a mitogen-activated protein kinase, which was linked to resistance [22]. Similar *trans*-regulation of genes involved in resistance to Bt has also been observed for APN in *Trichoplusia ni* resistant to Cry1Ac [37] and *Ostrinia nubilalis* resistant to Cry1Ab [38], and for both APN and an ABCC transporter in *Bombyx mori* resistant to Cry1Ab [39]. Further research should test the involvement of this control mechanism in downregulation of *mALP* in Vip-Sel and other Bt-resistant colonies.

The two studies so far focused on the underlying mechanism of resistance to Vip3Aa proteins share a similar feature in that in vitro binding is not reduced [18] (and the present work). According to these results, mechanisms other than binding site alteration seem to be responsible for conferring specific and high-level resistance to Vip3A. This contrasts with the fact that the alteration of membrane receptors is a common mechanism conferring high levels of resistance to Cry proteins. Better knowledge of the mode of action of Vip3A proteins will help shed light on the biochemical basis of resistance to these proteins.

#### **4. Conclusions**

The results herein show lack of significant Vip3Aa binding alterations in a resistant colony of *H. virescens*. These observations are in contrast to most cases of high levels of resistance to Cry proteins for which decreased binding is commonly detected. In addition, this study provides evidence of downregulation of membrane bound alkaline phosphatase (mALP) in the Vip3Aa-resistant colony, although results do not support involvement of mALP as a receptor for the Vip3A protein.

#### **5. Materials and Methods**

#### *5.1. Insects*

Two colonies of *H. virescens* originating from the same field population collected in Arkansas (USA) were used in this study: Vip-Sel (Vip3Aa-resistant) and Vip-Unsel (Vip3Aa susceptible). The process of

selection of the Vip-Sel colony with Vip3Aa has been previously described [12,19]. After 13 generations of selection, the LC50 of the Vip-Sel colony was 2300 μg/mL, representing a 2040-fold resistance ratio relative to the control Vip-Unsel colony. Both colonies were reared at the Imperial College London, Silwood Park campus (UK), and frozen larvae were sent for analysis to the Universitat de València (Spain).

#### *5.2. BBMV Preparation and Enzyme Activity Assays*

Brush border membrane vesicles (BBMV) from 3rd instar *H. virescens* larval midguts from Vip-Sel and Vip-Unsel colonies were prepared according to the differential magnesium precipitation method [40]. Isolated BBMV were flash frozen in liquid nitrogen and kept at −80 ◦C until used. The protein concentration of the BBMV preparations was determined by the method of Bradford using bovine serum albumin (BSA) as a standard [41].

Alkaline phosphatase (ALP) and leucine aminopeptidase (APN) activities were used as brush border membrane enzymatic markers to determine the quality of the BBMV preparations [42]. Specific ALP activity was determined by chromogenic detection of *p*-nitrophenyl phosphate (*p*NPP) substrate hydrolysis into *p*-nitrophenol, and specific APN activity was detected by hydrolysis of L-leu-*p*-nitroanilide substrate into *p*-nitroanilide. In both cases, chromogenic variation was measured on 1 μg of either BBMV or midgut homogenate at 405 nm (Infinite m200, Tecan, Mannedorf, Switzerland). Two different batches of BBMV were used and all enzymatic activity assays were performed in triplicate. Means values for enzyme activities from Vip-Unsel and Vip-Sel were compared by Student's *t*-test at a 5% level of significance.

For measuring specific ALP enzymatic activity in cultured Sf21 cells, a 1.6-mL suspension of each cell type (non-transfected, transfected with empty plasmid, and transfected with plasmid with *HvmALP1*) was used. Culture cells were centrifuged, washed twice with 300 μL of phosphate buffered saline (PBS) and then resuspended in 50 μL of PBS. Protein concentration was determined by the method of Bradford and specific ALP activity measured as above.

#### *5.3. Vip3Aa Protein Expression and Purification*

The Vip3Aa16 (Vip3Aa) protein (NCBI Accession No. AAW65132) was overexpressed in recombinant *Escherichia coli* BL21 carrying the *vip3Aa16* gene. Protein expression and lysis was carried out following the conditions described elsewhere [43]. Soluble Vip3Aa in the cell lysate was purified by two different methodologies. For binding and cell viability assays, Vip3Aa was partially purified by isoelectric point precipitation (IPP), activated with trypsin treatment and further purified by anion-exchange chromatography, as previously described [11]. For ligand assays, affinity chromatography purification was carried out using a HiTrap chelating HP column (GE Healthcare, Uppsala, Sweden) and then activated with trypsin, as described [11].

#### *5.4. Vip3Aa Labeling and Binding Experiments*

Purified Vip3Aa activated protein (25 μg) was labeled with 0.5 mCi of 125I using the chloramine T method [11]. The labeled protein was separated from the excess of free 125I in a PD10 desalting column (GE Healthcare, Uppsala, Sweden) and the purity of the 125I-labeled Vip3Aa was checked by autoradiography. The specific activity of the labeled protein was 2.2 mCi/mg.

Binding assays were performed as described elsewhere [11]. Prior to being used, BBMV were centrifuged and resuspended in binding buffer (20 mM Tris, 150 mM NaCl, 1 mM MnCl2, pH 7.4, 0,04% Blocking reagent from Sigma Aldrich, St. Louis, MO, USA). Competition binding experiments were conducted by incubating 1.4 μg of BBMV protein with 0.65 nM 125I-Vip3Aa in a final volume of 0.1 mL of binding buffer for 90 min at 25 ◦C in the presence of increasing amounts of unlabeled Vip3Aa. After incubation, samples were centrifuged at 16,000× *g* for 10 min and the pellet was washed once with 500 μL of ice-cold binding buffer. Radioactivity retained in the pellet was measured in a model 2480 WIZARD2 gamma counter. Data from the competition experiments were analyzed to determine

equilibrium binding parameters, dissociation constant (*K*d), and concentration of binding sites (*R*t) using the LIGAND software [44].

#### *5.5. Western and Ligand Blotting*

For the detection of ALP proteins in BBMV by Western blotting, BBMV (20 μg) were suspended in ice-cold PBS and heat denatured before separation on a SDS–10% PAGE gel. The resolved BBMV proteins were transferred to a nitrocellulose filter (Protran 0.45 μm NC, GE Healthcare, Uppsala, Sweden) using a BioRad Mini Trans-Blot system (Bio-Rad, Hercules, CA, USA) at 4 ◦C in blotting buffer (39 mM Glycine, 48 mM Tris-HCl, 0.037% SDS, 10% methanol, pH 8.5) for 1 h at constant voltage (100 V). After transfer, the nitrocellulose filter was blocked in blocking buffer (PBS, 0.1% Tween 20, 5% skimmed milk powder) overnight at 4 ◦C. After blocking and washing with PBST (PBS, 0.1% Tween 20) three times (5 min each), incubation with primary antibody against the membrane-bound form of ALP from *Anopheles gambiae* (generously provided by M. Adang, University of Georgia, USA) was performed for 90 min at a 1:5000 dilution at room temperature (RT). The membrane was then washed with PBST three times for 5 min each and then incubated with secondary antibody (goat anti-rabbit conjugated to horseradish peroxidase (HRP) at a 1:10,000 dilution) for 1 h at RT. After being washed with PBST three times for 5 min each, the membrane was developed using enhanced chemiluminescence (ECL Prime Western Blotting detection reagent, GE Healthcare, Uppsala, Sweden) in an ImageQuant LAS 4000 (GE Healthcare, Uppsala, Sweden), according to the manufacturer's instructions.

Ligand blotting for the detection of BBMV proteins binding Vip3Aa protein was performed with BBMV proteins resolved and immobilized as described above for Western blotting. The nitrocellulose membrane was blocked for 1 h at 4 ◦C in blocking buffer (5% skimmed milk), and after three washes for 5 min each with PBST buffer, it was incubated overnight at 4 ◦C with blocking buffer (1% skimmed milk) supplemented with affinity chromatography-purified Vip3Aa at a final concentration of 4 μg/mL. After washing with PBST three times for 5 min each, the membrane was incubated with primary antibody against Vip3Aa at a 1:5000 dilution for 1 h at RT. After three washing steps with PBST (5 min each), membranes were incubated with secondary antibody (goat anti-rabbit conjugated to HRP) for 1 h at RT. To visualize the marker, Precision Protein™ Streptactin-HRP conjugate (Bio-Rad, St. Louis, MO, USA) was used following the manufacturer's instructions. Upon washing three times (5 min each) with PBST, the membrane was developed as described for Western blotting.

#### *5.6. Proteomic Analysis*

After resolving BBMV proteins from Vip-Sel and Vip-Unsel colonies by SDS–10% PAGE, the gel was stained with Coomassie blue (Thermo Scientific™, Waltham, MA, USA). The band corresponding to the expected molecular weight of ALP (~66 kDa) was cut out and subjected to analysis by nano-electron spray ionization (nano-ESI) followed by tandem mass spectrometry (qQTOF) in a 5600 TripleTOF (AB Sciex, Madrid, Spain) system. Results were analyzed with ProteinPilot v5.0 software and the relative amount of the proteins detected was estimated using the exponentially modified protein abundance index (emPAI) as described elsewhere [45].

#### *5.7. RT-qPCR*

Relative expression levels for *HvmALP1* and *HvmALP2* isoforms (accession numbers FJ416470 and FJ416471, respectively) were determined by reverse transcription quantitative polymerase-chain reaction (RT-qPCR). For this purpose, total RNA of dissected midguts from both colonies (Vip-Unsel and Vip-Sel) was isolated using RNAzol (MRC Inc., Cincinnati, OH, USA) according to the manufacturer's protocol. Each RNA (1 μg) was reverse-transcribed to cDNA using random hexamers and oligo (dT) by following the instructions provided in the Prime-Script RT Reagent Kit (Perfect Real Time from TaKaRa Bio Inc., Otsu Shiga, Japan). RT-qPCR was carried out in a StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). Reactions were performed using 5× HOT FIREpol EVAGreen qPCR Mix Plus (ROX) from Solis BioDyne (Tartu, Estonia) in a total reaction volume of 25 μL. Specific primers for *HvmALP1*, *HvmALP2* and *Rps18* (endogenous control) genes were as described elsewhere [24]. The REST MCS software was used for gene expression analysis [46].

#### *5.8. Expression Vector Construction*

The full-length *HvmALP1* transcript was amplified from cDNA of *H. virescens* larvae and cloned into pET30a as described elsewhere [47]. Purified plasmid DNA was digested with *Eco*RI and *Not*I to excise the full-length sequence and ligate it in frame into *Eco*RI-*Not*I sites of the pIZT/V5His vector (Thermo Scientific™, Waltham, MA, USA), to generate the pIZT/V5His/*HvmALP1* construct. Ligation products were transformed into *E. coli* strain DH5α and transformants checked for correct insertion by sequencing (University of Tennessee Sequencing Facility, Knoxville, TN, USA). Purified plasmid was used to transform *E. coli* strain DH10β and liquid cultures of LB medium supplemented with Zeocin (25 μg/mL) were used to amplify the vector. To purify the plasmids for transfection, the NucleoSpin® Plasmid kit (Macherey-Nagel, Düren, Germany) was used. Double digestion with *Eco*RI and *Not*I (which cleaved the full-length *HvmALP1* insert) and 1% agarose gel electrophoresis were performed to check plasmid and/or insert integrity. The concentration of plasmid DNA was measured with a Thermo Scientific™ Nanodrop™ 2000 Spectrophotometer.

#### *5.9. Transient Expression of HvmALP1 in Sf21 Cells*

Cultured Sf21 insect cells, originally derived from *S. frugiperda*, were maintained in 25 cm<sup>2</sup> tissue culture flasks (Nunc T25 flasks, Thermo Scientific™, Waltham, MA, USA) at 25 ◦C with 4 mL of Gibco® Grace's Medium (1×) (Life Technologies™, Paisley, UK) supplemented with 10% heat-inactivated fetal bovine serum (FBS).

For transient expression, cells were seeded on 12-well plates with the same medium without FBS at ca. 70% confluency and transfected with 0.5 μg of the pIZT/V5His/*HvmALP1* or pIZT/V5His plasmid using Cellfectin® II Reagent (Thermo Scientific™, Waltham, MA, USA), following manufacturer's instructions. Five hours post-transfection, the medium was replaced with fresh medium containing 10% FBS. After 24 h, cells were examined using a confocal microscope (Olympus, FV1000MPE, Tokyo, Japan) equipped with the appropriate filter for green fluorescent protein (GFP) detection as transfection marker. The enzymatic activity of alkaline phosphatase was then measured as explained above.

#### *5.10. Cell Viability Assays*

Viability of transfected Sf21 cells exposed for 24 h to Vip3Aa was measured using the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay. Preliminary assays were performed to determine a final Vip3A concentration of 300 μg/mL as resulting in ~50% loss of viability in the control cell line (data not shown). Briefly, cells (100 μL per well) were transferred to 96-well ELISA plates (flat bottom) and incubated at 25 ◦C for at least 45 min. Then, 10 μL of trypsin-activated Vip3Aa toxin was added to each well (300 μg/mL final concentration). As negative and positive controls, 10 μL of Tris buffer (Tris 20 mM, NaCl 150 mM, pH 9) and 10 μL of 2% Triton X-100 were used, respectively. After 24 h of incubation at 25 ◦C, cell viability was assessed by applying 20 μL of CellTiter 96® AQueous One Solution Reagent (Promega, Madison WI, USA) to each well and incubating for 2 h at 25 ◦C. Absorbance was measured at 490 nm (Infinite m200, Tecan, Mannedorf, Switzerland). The percentage of viable cells was obtained as described elsewhere [48]. Mean values in the transfected cells against the non-transfected cells were compared by Student's *t*-test at 5% level of significance.

**Author Contributions:** Conceptualization, P.H.-M. and J.F.; Formal analysis, D.P., M.C., and A.M.-L.; Funding acquisition, J.L.J.-F. and J.F.; Methodology, D.P. and M.C.; Resources, J.L.J.-F. and D.J.W.; Supervision, P.H.-M. and J.F.; Writing—original draft, D.P.; and Writing—review and editing, J.L.J.-F., D.J.W., P.H.-M., and J.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the Spanish Ministry of Science, Innovation and Universities (grant no. RTI2018-095204-B-C21). D.P. is recipient of a PhD grant from the Spanish Ministry of Science, Innovation and Universities (grant ref. FPU15/05652). Support for JLJ-F was provided by the NC246 Multistate Hatch project from the USDA National Institute for Food and Agriculture (NIFA). The proteomic analysis was performed at the Proteomics Facility of the Servei Central de Suport a la Investigació Experimental (SCSIE) at the Universitat de València (Valencia, Spain) that belongs to ProteoRed, PRB2-ISCII, supported by grant PT13/0001, of the PE I+D+I 2013-2016, funded by ISCII and FEDERPT13/0001.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Oligomer Formation and Insecticidal Activity of** *Bacillus thuringiensis* **Vip3Aa Toxin**

**Ensi Shao 1,**†**, Aishan Zhang 1,**†**, Yaqi Yan 1, Yaomin Wang 1, Xinyi Jia 1, Li Sha 1, Xiong Guan 2, Ping Wang 3,\* and Zhipeng Huang 1,\***


Received: 19 March 2020; Accepted: 21 April 2020; Published: 23 April 2020

**Abstract:** *Bacillus thuringiensis* (Bt) Vip3A proteins are important insecticidal proteins used for control of lepidopteran insects. However, the mode of action of Vip3A toxin is still unclear. In this study, the amino acid residue S164 in Vip3Aa was identified to be critical for the toxicity in *Spodoptera litura*. Results from substitution mutations of the S164 indicate that the insecticidal activity of Vip3Aa correlated with the formation of a >240 kDa complex of the toxin upon proteolytic activation. The >240 kDa complex was found to be composed of the 19 kDa and the 65 kDa fragments of Vip3Aa. Substitution of the S164 in Vip3Aa protein with Ala or Pro resulted in loss of the >240 kDa complex and loss of toxicity in *Spodoptera litura*. In contrast, substitution of S164 with Thr did not affect the >240 kDa complex formation, and the toxicity of the mutant was only reduced by 35%. Therefore, the results from this study indicated that formation of the >240 kDa complex correlates with the toxicity of Vip3Aa in insects and the residue S164 is important for the formation of the complex.

**Keywords:** *Bacillus thuringiensis*; Vip3A; *Spodoptera litura*; site-directed mutagenesis

**Key Contribution:** Our results correlated the formation of a >240 kDa protein complex with the insecticidal activity of Vip3Aa toxin. The residue S164 in Vip3Aa protein was identified to be important for the formation of the >240 kDa protein complex.

#### **1. Introduction**

The vegetative insecticidal proteins (VIPs) from *Bacillus thuringiensis* (Bt) have been used as important insecticidal proteins for control of insect pests [1–3]. Vip toxins are divided into four families, including Vip1, Vip2, Vip3 and Vip4 [3]. Vip1 and Vip2 proteins act as binary toxins against some species of coleopteran and hemipteran insect [4,5]. Only 1 Vip4 protein has so far been identified but shows no activity in insects [6]. Vip3 proteins contain approximately 787 amino acid residues, showing no sequence homology with Vip1, Vip2 and Vip4 proteins [3]. Vip3 proteins have a high insecticidal activity against a wide variety of lepidopteran pests [7]. Vip3A proteins do not share the binding sites with the Bt Cry proteins [8–11], so pyramiding Vip3A proteins and Cry proteins has been widely adopted in Bt-crops [12].

Although Vip3A toxins have already been applied in transgenic crops for the control of lepidopteran pests, current understanding of the mode of action of Vip3A proteins remains limited. It is commonly assumed that Vip3A toxins exert their insecticidal activity by going through a similar sequence of events as Bt Cry1A toxins [13]. So far, the structure of Vip3A toxin has not been solved. Its structural information has been derived only by *in-silico* modeling [14,15], though the structure of the Vip3B was recently reported [16]. Studies on proteolytic activation of Vip3A proteins have shown that by a proteolytical process Vip3A protoxins are cleaved to become several major fragments, generally including fragments of 62–66 kDa, 45 kDa, 33 kDa and 19–22 kDa [17–20]. The 62–66 kDa fragment from the C-terminus of Vip3A toxins has been determined to be the main product from proteolytic processing. The 45 kDa and 33 kDa fragments are products from further processing of the 62–66 kDa fragment [18]. The 19–22 kDa fragment contains the first 199 amino acids at the N-terminus of Vip3A [21]. It has been suggested that the 62–66 kDa fragment at the C-terminus in Vip3A toxin is the activated core of the toxin [22–24]. However, recent studies have indicated that both of the 19–22 kDa and the 62–66 kDa fragments are required for the stability and specificity of Vip3A toxins [20]. More recently, a ~340 kDa homo-tetramer, constituted by the 19–22 kDa and the 62–66 kDa fragments, has been identified from Vip3A after treatment with trypsin or insect midgut proteases [18]. However, whether the formation of this ~340 kDa homo-tetramer is essential for the insecticidal activity of Vip3A in insects remains unknown.

A recent study of Vip3Af by Ala scanning to cover 558 out of the 788 residues showed that the most Ala substitutions in Vip3Af significantly decreased the insecticidal activity, and the proteolytically processed fragments of the Vip3Af substitution mutants displayed six different patterns by SDS-PAGE analysis [14]. Further analysis indicated that Vip3Af mutants with different proteolytic patterns could form a variety of oligomeric products [21]. The substitution of the residue T167 or G168 by Ala in the predicted 19 kDa N-terminal fragment of Vip3Af did not change the proteolytic proccessing, but both substitutions significantly decreased the insecticidal activity [14]. Sequence alignments indicated that the amino acid residues from K152 to P171 are highly conserved among the Vip3A toxins [3].

*Spodoptera litura* is a polyphagous species and a major pest of many crops worldwide due to its vigorous defoliation [25]. *S. litura* is not susceptible to Bt Cry1A toxins but highly susceptible to Vip3A toxins [26,27]. In this study, we constructed Vip3Aa mutants by site directed mutagenesis and investigated the insecticidal activity of the mutants in *S. litura*. The amino acid residue S164 in Vip3Aa protein was identified to be critical for the toxicity of Vip3Aa. Investigation of the toxicity of Vip3Aa in *S. litura* by substitutions of S164 with different amino acid residues indicated that a protein oligomer formed with the 19 kDa and the 65 kDa fragments of Vip3Aa is the toxin core necessary for the insecticidal toxicity.

#### **2. Results**

#### *2.1. Insecticidal Activity of Residue Substituted Vip3Aa Mutants Against Neonates of S. litura*

The wild-type Vip3Aa protein and its mutants at K152, D154 and S164, respectively, were prepared through a glutathione S-transferase (GST) tagged protein purification system. Vip3Aa mutants with substitution of K152 or D154 with Ala were expressed as GST-Vip3Aa-K152A and GST-Vip3Aa-D154A fusion proteins. Vip3Aa mutants from substitution of S164 with Ala, Pro and Thr, respectively, were expressed as GST-Vip3Aa-S164A, GST-Vip3Aa-S164P and GST-Vip3Aa-S164T fusion proteins. The purified GST-Vip3Aa fusion proteins and the wild-type Vip3Aa (Vip3Aa-WT) were fed to neonates of *S. litura* to determine their insecticidal activity respectively. The bioassay results showed that the substitution mutations of K152A and D154A did not significantly change the toxicity of the toxin, in comparison with the wild-type Vip3Aa protein (Table 1). However, substitution of S164 with Ala or Pro completely abolished the toxicity of Vip3Aa. In contrast, substitution of S164 with the similar amino acid residue Thr only slightly reduced the insecticidal activity (35% reduction). Mortality of the

control groups fed with 100 μg/mL or 250 μg/mL of purified GST tag protein were below 5% after 96 h feeding (results not shown).


**Table 1.** Insecticidal activity of Vip3Aa toxins in neonates of *S. litura.*

CI: confident interval.

#### *2.2. Analysis Vip3Aa Fragments After Proteolytic Processing*

To examine the difference in the proteolytic processing among the Vip3Aa-WT and three S164 mutants, each Vip3Aa protein was processed by trypsin or midgut proteases of *S. litura* and analyzed by SDS-PAGE after heat denaturation. The tryptic fragments from the three S164 mutants and Vip3Aa-WT contained major bands at 65 kDa, 35 kDa and 19 kDa and multiple weak bands from 29 kDa to 66.4 kDa (Figure 1). Vip3Aa proteins processed by midgut proteases of *S. litura* present a different pattern to the tryptic proteins. Besides the major fragments at 65 kDa, a band at 38 kDa and another at 30 kDa were observed in the midgut proteases processed Vip3Aa-WT and three S164 mutants. The band of 19 kDa was weak or invisible after in vitro proteolytic processing of Vip3Aa by the midgut proteases of *S. litura* (Figure 1). It could be observed that after proteolytic treatment with trypsin or midgut proteases, Vip3Aa-WT and three S164 mutants showed the same protein patterns. All trypsin digested Vip3Aa proteins contained a strong band at 19 kDa. Several other protein fragments were observed with the molecular weight from 14.3 kDa to 19 kDa from Vip3Aa-WT and three S164 mutants, although some bands were weak (Figure 1).

**Figure 1.** Analysis of Vip3Aa proteins after treatment by trypsin or midgut proteases of *S. litura.* Purified Vip3Aa-WT, Vip3Aa-S164T, Vip3Aa-S164A and Vip3Aa-S164P were *in vitro* digested by commercial trypsin or midgut proteases of *S. litura*. Processed proteins were mixed with 5 × SDS-PAGE sample buffer followed by heat denaturation and analyzed by the electrophoretic analysis in an SDS-PAGE gel.

#### *2.3. Analysis of Vip3Aa Protein Complexes by Native PAGE After Proteolytic Processing*

The same or similar protein digestion patterns were observed by SDS-PAGE from Vip3Aa-WT and three S164 mutants after proteolytic processing by either trypsin or midgut proteases of *S. litura*. Protein fragments from trypsin- or gut proteases-processed Vip3Aa proteins were then analyzed by native PAGE to identify the protein complexes. Two similar major bands, representing the protein complex 1 and 2, were observed from Vip3Aa-WT, Vip3Aa-S164T, Vip3Aa-S164A and Vip3Aa-S164P. However, a band, representing the protein complex 3, with a higher molecular weight than the two bands were observed from the trypsin-processed Vip3Aa-WT, Vip3Aa-S164T and Vip3Aa-S164P but not from Vip3Aa-S164A (Figure 2a). For the midgut proteases-processed Vip3Aa proteins, the band of protein complex 3 was observed from Vip3Aa-WT and Vip3Aa-S164T but not from Vip3Aa-S164A and Vip3Aa-S164P (Figure 2b). To estimate the molecular weight of the three protein complexes, the trypsin-processed Vip3Aa was analyzed by native SDS-PAGE. Two clear bands at ~240 kDa and ~200 kDa were observed (Figure 2c). The two bands in Figure 2c are assumed to the relatively dominant protein complex 1 and 2 in Figure 2a, and the molecular weight of complex 3 is predicted to be >240 kDa.

**Figure 2.** Analysis of native Vip3Aa proteins after proteolytic processing. Protease treated Vip3Aa-WT, Vip3Aa-S164T, Vip3Aa-S164A and Vip3Aa-S164P by either commercial trypsin or midgut proteases of *S. litura* larvae were analyzed by the electrophoretic analysis without heat denaturation. (**a**) Vip3Aa proteins after tryptic processing were analyzed in a native gel; (**b**): Vip3Aa proteins after processing by midgut proteases were analyzed in a native gel; (**c**): Vip3Aa proteins after tryptic processing were mixed with 5 × SDS-PAGE sample buffer without β-mercaptoethanol and analyzed in an SDS-PAGE gel. Protein complex 1, protein complex 2 and protein complex 3 in panel (**a**) indicate gel bands sliced from each lane in the native gel.

#### *2.4. Composition of the Three Protein Complexes Formed from Vip3Aa Toxins after Tryptic Processing*

To analyze the compositions of the three protein complexes, the bands corresponding to protein complexes 1, 2 and 3 (Figure 2a) were excised from the native PAGE gel (Figure 2a). The gel slices were mixed with SDS-PAGE sample buffer, heat denatured and loaded on SDS-PAGE gel to separate the proteins. All protein complexes contained the 65 kDa major fragment and multiple weak bands from 29 kDa to 66.4 kDa (Figure 3). Difference in composition was observed between the wild type and the mutants in the fragments below 20 kDa. A 19 kDa fragment was observed from protein complex 3 of trypsin-processed Vip3Aa-WT and Vip3Aa-S164T (Figure 3a). In comparison with the 19 kDa fragment, a slightly smaller fragment (17 kDa) was observed from the protein complex 3 from Vip3Aa-S164P (Figure 3a). An even smaller 15 kDa fragment was observed from protein complex 1 of Vip3Aa-WT and three mutants (Figure 3b). No protein bands below 20 kDa were observed from the protein complex 2 (Figure 3b). In addition, a peptide showing molecular weight around 95 kDa was observed from the protein complex 1 and 3 but not from the protein complex 2 (Figure 3).

**Figure 3.** Separation of peptides from protein complexes of tryptic Vip3Aa proteins. Major protein bands representing different protein complexes in Figure 2a were sliced and separated in an SDS-PAGE gel. (**a**) peptides separated from the protein complex 3 in Figure 2a; (**b**) peptides separated from the protein complexes 1 and 2 respectively in Figure 2a. The yellow, red, black and white arrows indicate the bands of 95 kDa, 19 kDa, 17 kDa and 15 kDa fragments respectively.

#### *2.5. Identification of Tryptic Fragments from the 15, 17 and 19 kDa Protein Fragments by Peptide Fingerprinting*

The 15 and 19 kDa fragments, isolated from protein complexes 1 and 3 of Vip3Aa-WT and the 17 kDa fragment from protein complexes 3 of Vip3Aa-S164P were analyzed by nano LC-MS/MS to identify the protein fragments. The identified peptides derived from the 15, 17 and 19 kDa fragments were mapped to the amino acid sequence from D32 to K195 of Vip3Aa protein, located at the N terminal region (Figure 4).

**Figure 4.** Schematic representation of the 15 kDa, 17 kDa and 19 kDa fragments from Vip3Aa protein. The first 198 amino acids at the N terminus of Vip3Aa were presented. The red box indicates amino acids corresponding to the N terminal 19 kDa fragment of Vip3Af. The yellow, blue and green boxes represent LC-MS/MS identified peptides from the 15 kDa, 19 kDa and 17 kDa fragments respectively.

#### *2.6. Correlation of Toxicity of Vip3Aa Protein with the Formation of the Protein Complex 3 Composed of 19 kDa and 65 kDa Peptides*

Trypsin-processed Vip3Aa-WT, Vip3Aa-S164T, Vip3Aa-S164A and Vip3Aa-S164P were fed to the neonates of *S. litura,* respectively, to assay their insecticidal activity. After 96 h, 100% mortality was observed by feeding *S. litura* larvae with 5 μg/mL and 50 μg/mL of trypsin-processed Vip3Aa-WT or Vip3Aa-S164T. In contrast, neither trypsin-processed Vip3Aa-S164A nor Vip3Aa-S164P showed significant toxicity to the larvae of *S. litura* (Figure 5). The insecticidal activity of trypsin treated Vip3Aa proteins was consistent to that of Vip3Aa protoxins (Table 1) and correlated with the formation of the protein complex 3 composed of the 19 kDa and the 65 kDa peptides (Figures 2a and 3a)

**Figure 5.** Mortality of *S. litura* larvae fed with trypsin-processed Vip3Aa proteins. After tryptic processing, Vip3Aa-WT, Vip3Aa-S164T, Vip3Aa-S164A and Vip3Aa-S164P were respectively fed to neonates of *S. litura* for 96 h to test their insecticidal activity. Error bars indicate the standard error of mortality among five replications.

#### **3. Discussion**

Previous studies have indicated that proteolytic processing of Vip3A proteins in insect midgut is a key step to exert the insecticidal activity [3,13]. In insect midgut, Vip3A proteins are processed by midgut proteases to produce a 62–66 kDa protease resistant toxic core from the C-terminal part of Vip3A. However, a recent study indicated that deletion of the first 198 residues at the N-terminus outside the ~65 kDa fragment region could lead to a complete loss of insecticidal activity and the resulting Vip3Aa fragments became sensitive to trypsin degradation [28]. Current studies have also shown that with treatment of Vip3A by trypsin, a 19~20 kDa peptide from the N-terminal region could bind with the C-terminal 62~65 kDa peptide, leading to the formation of a ~360 kDa homo-tetramer, which was tolerant to degradation in the protease-rich environment [29]. This 19~20 kDa peptide was proposed to play a functional role in protecting the 62~65 kDa peptide from proteolytic degradation and is necessary for the toxicity of the toxin in insects [30]. The K152 to E168, included in the N terminus 19~20 kDa peptide of Vip3Aa, were predicted to be a loop structure by the three-dimensional structure modeling software LOMETS (https://zhanglab.ccmb.med.umich.edu/LOMETS/). Significant decrease of toxicity of Vip3Af in *Spodoptera frugiperda* and *Agrotis segetum* was observed after substitution of T167 or E168 by Ala [14,20]. S164 was considered to be a polar amino acid located at the C terminus of the K152-E168 loop. Both K152 (carrying a basic polar side chain) and D154 (carrying an acidic polar side chain) were predicted to be at the N terminus of the K152-E168 loop. Consequently, in this study, we chose K152, D154 and S164 as our targets to analyze potential effects on the toxicity of Vip3Aa after substitution of these three amino acids respectively by Ala. Results of bioassay showed that only substitutions at S164 affect the toxicity of Vip3Aa in *S. litura*. Three main protein complexes were observed in protease treated Vip3Aa-WT and its three S164 mutants by native PAGE (Figure 2a,b). The protein complexes were composed of protein fragments of 19 and 65 kDa, 17 and 65 kDa, 15 and 65 kDa, or a single 65 kDa peptide only, respectively (Figure 3). The 95 kDa protein band was observed in protein complexes 1 and 3 but not protein complex 2 from each Vip3Aa protein. We interpret

that the 95 kDa fragment is complexed with a 15 kDa peptide in protein complex 1 or a 19 kDa peptide in protein complex 3 with a 65 kDa peptide (Figure 3). LC-MS/MS analysis of the peptides between 15~19 KDa from the complexes indicated that the 15 kDa, 17 kDa and 19 kDa peptides were all from the N terminus of Vip3Aa protein, which corresponds to the previous reported domain I in Vip3Af protein [21]. Vip3Aa protoxin could be processed in vivo by the midgut proteases of *S. litura*. Toxicity of the Vip3Aa toxin, pretreated by midgut proteases, in *S. litura* should be similar to that of Vip3Aa protoxins. In order to build the relationship between the toxicity and the presence of the protein complex 3, mortality of *S. litura* fed with trypsin-processed Vip3Aa toxin was calculated and compared to the bioassay results by feeding *S. litura* with Vip3Aa protoxins (Table 1). Bioassay results showed that the trypsin-processed Vip3Aa-WT and Vip3Aa-S164T, in which the protein complex 3 was formed (composed of 19 and 65 kDa peptides), had significant toxicity in larvae of *S. litura* while the trypsin-processed Vip3Aa-S164A and Vip3Aa-S164P did not form the protein complex 3 with the 19 and 65 kDa peptides and completely lost the toxicity (Figure 5). These results are corresponding to the results of bioassay using Vip3Aa protoxins (Table 1). The toxicity of Vip3Af against *S. frugiperda* and *A. segetum* has been suggested to correlate with the transient formation of a tetramer composed of 20 kDa and 62 kDa peptides before the final processing to smaller fragments [14,21]. In this study, 100% mortality was observed in *S. litura* larvae fed with tryptic Vip3Aa-WT or Vip3Aa-S164T, while <10% mortality was observed in *S. litura* larvae fed with tryptic Vip3Aa-S164A or Vip3Aa-S164P (Figure 5). The protein complex 3 (composed of the 19 and 65 kDa fragments but not the 17 and 65 kDa fragments) could only be observed in tryptic Vip3Aa-WT or Vip3Aa-S164T (Figure 3). Consequently, formation of the protein complex 3 from Vip3Aa-WT and Vip3Aa-S164T was directly correlated to the toxicity of Vip3Aa in *S. litura*. A protein complex showing closed molecular weight to the protein complex 3 in Vip3Aa-WT and Vip3Aa-S164T was observed from tryptic Vip3Aa-S164P (Figure 2a) but not observed from midgut proteases-processed Vip3Aa-S164P (Figure 2b). Composition of this protein complex was identified to be the 17 kDa and the 65 kDa fragments (Figure 3a), different from the complex 3 from Vip3Aa-WT and Vip3Aa-S164T (Figure 4). This protein complex was degraded after treatment of Vip3Aa-S164P with midgut proteases of *S. litura*, while the protein complex 3 from Vip3Aa-WT and Vip3Aa-S164T remained stable (Figure 2b). Both the 17 kDa and the 19 kDa peptides were identified from the N terminal 199 amino acid residues of Vip3Aa proteins (Figure 4). These results indicated that the complete 19 kDa fragment is essential for the stability of the protein complex 3 which correlated to the insecticidal activity of Vip3Aa toxin.

The oligomer of Vip3A formed after proteolytic processing was observed through gel filtration chromatography analysis [14,21,29,31]. In this study, three protein complexes were observed from Vip3Aa-WT and its three mutants by native PAGE (Figure 2). Molecular weights of protein complexes 1 and 2 observed in the native SDS-PAGE gel were predicted to be ~240 kDa and ~200 kDa (Figure 2c). It is interesting that the protein complex 3 from tryptic Vip3Aa-WT could not be observed in the native SDS-PAGE. Previous studies identified a further degradation of Vip3A proteins due to the introduction of the secondary cleavage sites after treating with SDS contained SDS-PAGE sample buffer [17]. We speculate that protein complex 3 is easy to be disassembled in the native SDS-PAGE. Due to the disappearance of the less abundant protein complex 3 in the native SDS-PAGE gel, its molecular weight could only be predicted to be >240 kDa. From the toxin Vip3Af, a ~360 kDa homo-tetramer composed of 20 kDa and 62 kDa peptides has been proposed to correlate with the toxicity of Vip3Af toxin [14,21]. In this study, Vip3Aa-WT and Vip3Aa-S164T were observed to form the complex 3 composed of the 19 and 65 kDa peptides (Figure 3a) and have toxicity (Figure 5). It is possible that the protein complex 3 from Vip3Aa-WT and Vip3Aa-S164T corresponds to the previously reported 360 kDa homo-tetramer from Vip3Af, which was predicted to be formed by four 85–90 kDa protein complexes, each of them was composed of a 19 kDa peptide and a 65 kDa peptide [14,29]. SDS-PAGE analysis of trypsin- or midgut proteases-processed Vip3Aa proteins showed nearly the same protein fragment patterns below 20 kDa from Vip3Aa-WT and three S164 mutants (Figure 1). However, the protein complex 3 composed of the 19 kDa and 65 kDa fragments could only be observed from Vip3Aa-WT and Vip3Aa-S164T

(Figure 2a,b). This is because the 15 kDa, 17 kDa and 19 kDa fragments were all identified from the N terminus of Vip3Aa (Figure 4). The S164 is critical for the formation of the protein complex 3, composed of 19 kDa and 65 kDa peptides from Vip3Aa proteins.

In conclusion, the present study identified a > 240 kDa protein complex composed of the 19 kDa and 65 kDa fragments from Vip3Aa after proteolytic processing. The formation of this protein complex was determined to correlate with the toxicity of Vip3Aa in *S. litura* larvae. The S164 in Vip3Aa is critical for the formation of the >240 kDa protein complex and consequently the insecticidal activity. The results from this study provided new information on the insecticidal mechanism of Vip3Aa toxins.

#### **4. Materials and Methods**

#### *4.1. Site Directed Mutation on the vip3Aa Gene*

The *vip3Aa* gene (NCBI Accession No. AF500478.2) was cloned from plasmid of Bt WB7, a native strain isolated from soil collected in Wuyi mountain (Fujian, China) [32], by the use of primer pairs P-3Aa-F and P-3Aa-R (Table 2). The pGEX-KG vector [33] was used for the heterologous expression of the *vip3Aa* gene in *E. coli* BL21 (DE3). To substitute nucleotides coding for a single amino acid residue in Vip3Aa protein, primer pairs containing the site substitutions in *vip3Aa* gene was carried out by PCR using the pGEX-Vip3Aa as the template. Primers for the site directed mutation were listed in Table 2. To replace the S164 by an Ala in Vip3Aa, primers P-3Aa-F and P-164A-R were used as the primer pair for the 1st round PCR to obtain the N-terminal part of *vip3Aa* gene. P-164A-F and P-3Aa-R were used as the primer pair for the 2nd round PCR to obtain the C-terminal part of *vip3Aa* gene. Both PCR products were diluted 1000 folds in water and used as the template for the 3rd round PCR using P-3Aa-F and P-3Aa-R as the primer pair to obtain the full-length *vip3Aa* gene with the codons coding for S164 replaced by codons coding for A164. PCR reactions were performed using the iProofTM High-Fidelity Master Mix DNA polymerase (Bio-RAD, Hercules, CA, USA). Other site substituted mutants of *vip3Aa* gene were generated according to the same procedure by the use of corresponding primers (Table 2). The final products from the 3rd round PCR were purified and digested with restricted enzymes of *Nco* I and *Sac* I. Digested products were purified and ligated with pGEX-KG plasmid linearized with *Nco* I and *Sac* I. Plasmids carrying mutated *vip3Aa* gene were transformed into *E. coli* BL21 (DE3) cells for protein preparations.


**Table 2.** Primers for the site substitutions in *vip3Aa* gene.

Underlined sequences indicate the restricted enzyme sites of *Nco* I and *Sac* I. Lower case sequences indicate the mutant nucleotides in each primer.

#### *4.2. Expression and Purification of Vip3Aa Proteins*

To prepare Vip3Aa proteins, 250 μL of overnight culture of BL21 cells carrying a plasmid of pGEX-Vip3Aa were inoculated to 250 mL of LB ina1L flask. The bacterial cultures were incubated at 37 ◦C and shaken at 150 rpm until OD600 reached 0.5. Protein expression was induced by addition of 0.8 mM IPTG (Isopropyl β-D-thiogalactoside) to the cultures, followed by incubation at 16 ◦C for 24 h. The *E. coli* cells were then harvested by centrifugation at 14,000× *g* for 1 min and the cell pellets were resuspended and washed in GST binding buffer (PBS, pH 7.3). The cell suspension was sonicated with a sonicator (VC-50, Sonics & Materials Inc. Danbury, CT, USA), followed by centrifugation at 21,000× *g* for 10 min to pellet the cell debris. The supernatant containing soluble GST-Vip3Aa proteins was loaded onto a Glutathione Sepharose column. Purification of GST fusion proteins and removing of GST tag by thrombin followed the standard purification procedure described by manufacturer of Glutathione Sepharose 4B (GE healthcare, Chicago, IL, USA). All purification steps were conducted at 4 ◦C or on ice. The purified GST-Vip3Aa fusions or thrombin treated Vip3Aa were quantified by the Bradford method [34].

#### *4.3. Insects Rearing and Bioassays*

An inbred colony of *S. litura* reared in the laboratory for over 3 years (~30 generations) was used in this study. The *S. litura* colony was maintained on a soybean-based artificial diet at 27◦C with 50% humidity and a photoperiod 16 h of light and 8 h of darkness.

Bioassays were conducted by diet overlay method [35]. GST tag of Vip3Aa-WT was removed while other Vip3Aa mutants were prepared as GST-Vip3Aa fusions for bioassays. Briefly, the Vip3Aa-WT toxins or GST-Vip3Aa fusions in a series of dilutions were prepared in water. A 200 μL aliquot of diluted toxin was overlaid on the surface (~7 cm2) of diet in each cup (30 mL plastic cup containing ~5 mL diet). Each concentration included 5 replications. Ten neonates were placed into each cup. Cups were covered with lids and kept in the rearing room at 27 ◦C, 50% humidity and a photoperiod of 16:8 (light:dark) for at least 96 h. Mortality of larvae in each cup was recorded every 24 h. Cups contain diet overlaid by 100 μg/mL or 250 μg/mL of GST tag protein diluted in water were prepared as negative controls. Probit analysis of the bioassay data was carried out using the POLO program [36] to estimate the LC50 and 95% confidence limits (CL). For bioassays using tryptic Vip3Aa proteins, Glutathione Sepharose carrying GST-Vip3Aa fusions were digested with 10 mg/mL trypsin. Tryptic Vip3Aa proteins were quantified by the Bradford method and diluted to the concentration of 5 μg/mL and 50 μg/mL for the diet overlay bioassays described above. Mortality of *S. litura* larvae fed with tryptic Vip3Aa proteins were recorded in 96 h and analyzed by Prism (version 8.2.0, GraphPad Software, San Diego, CA, USA).

#### *4.4. In Vitro Proteolytic Processing of Vip3Aa Proteins*

To prepare midgut proteases of *S. litura* larvae, mid-fifth-instar larvae of *S. litura* were immobilized on ice for several minutes and dissected to isolate the complete midgut without loss of its contents. Midgut homogenates were prepared by thorough homogenization of 5 midguts in a 1.5 mL microcentrifuge tube. Grounded tissues were centrifugated at 16,000× *g* for 10 min at 4 ◦C. The supernatant was collected and distributed into small aliquots, snap frozen in liquid nitrogen and then stored at −80 ◦C until use. The protein concentration of midgut protease preparations was measured using the Bradford method.

Affinity-purified Vip3Aa proteins were subjected to in vitro proteolytic processing with trypsin (Sigma-Aldrich Inc. St. Louis, MO, USA) or midgut proteases of *S. litura*. Vip3Aa fusions were incubated with 10 mg/mL of trypsin or 400 μg/mL midgut protease preparation in Tris-HCl buffer (20 mM Tris-HCl, 0.15 M NaCl, 5 mM EDTA, pH 8.6) at the ratio of 120:100 (trypsin:Vip3Aa, *w*/*w*) for the trypsin treatment and 40:100 (midgut protease:Vip3Aa, *w*/*w*) for the midgut protease treatment. In vitro digestion was carried out at 30 ◦C for 6 h.

#### *4.5. Analysis of Vip3Aa Proteins by the Native Gel and SDS-PAGE Gel*

Protease treated Vip3Aa proteins were immediately analyzed by native PAGE or SDS-PAGE. To analyze protein complexes of Vip3Aa by native PAGE, trypsin- or midgut proteases-processed Vip3Aa proteins were mixed with the 2 × native PAGE sample buffer (0.5 M Tris-HCl pH 6.8, 5% bromophenol blue, 30% glycerol) and separated by the electrophoretic analysis on a 6% native PAGE gel. To analyze denatured Vip3Aa proteins by SDS-PAGE, the processed Vip3Aa proteins or protein complexes excised from the native PAGE gel were mixed with 5 × SDS-PAGE sample buffer (0.2 M Tris-HCl pH 6.8, 1 M sucrose, 5 mM EDTA, 0.1% bromophenol blue, 10% SDS and 5% β-mercaptoethanol), heated at 99 ◦C for 10 min and centrifugated at 15,000× *g* for 5 min. The supernatant was loaded to a 10% SDS-PAGE gel for the electrophoretic analysis. To estimate the molecular weight of protein complexes of Vip3Aa proteins by native SDS-PAGE, trypsin-processed Vip3Aa proteins were mixed with 5 × SDS-PAGE gel sample buffer with the absence of β-mercaptoethanol and loaded to a 6% SDS-PAGE gel for the electrophoretic analysis.

#### *4.6. Identification of Trypsin-Processed Fragments*

Protein bands of the 19 kDa fragment and the 15 kDa fragment contained by the protein complexes of trypsin treated wild-type Vip3Aa and a protein band of the 17 kDa fragment contained by the protein complex from trypsin treated Vip3Aa-S164P were excised from the SDS-PAGE gel after staining with Coomassie blue and detained with 30% acetonitrile and 100 mM NH4HCO3. The gel slices were dried in a vacuum centrifuge, and the proteins were reduced in-gel with dithiothreitol (10 mM DTT and 100 mM NH4HCO3) for 30 min at 56 ◦C, then alkylated with iodoacetamide (200 mM IAA and 100 mM NH4HCO3) in dark at room temperature for 30 min. Gel slices were briefly rinsed with 100 mM NH4HCO3 and acetonitrile, respectively, followed by digestion with 12.5 ng/μL trypsin in 25 mM NH4HCO3 overnight. The peptides were extracted three times with 60% acetonitrile and 0.1% trifluoroacetic acid. The extracts were pooled and dried completely in a vacuum centrifuge. The peptide mass and sequence were determined by Liquid Chromatography (LC)—Electrospray Ionization (ESI) Tandem mass spectrometry (MS/MS) in a Q Exactive mass spectrometer which was coupled to Easy nLC (Proxeon Biosystems, Thermo Fisher Scientific, Shanghai, China). The MS data were analyzed using Max Quant (version 1.6.4.0, Max Planck Institute of Biochemistry, Munich, Germany) by searching the data against the amino acid sequence of Vip3Aa, and the intensity of sequenced peptide in the target protein was calculated.

**Author Contributions:** E.S. and P.W. conceived and designed research. E.S., A.Z., Y.Y., Y.W. and X.J. conducted experiments. E.S., A.Z., L.S., and Z.H. analyzed data. E.S., P.W., X.G. and Z.H. wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by National Natural Science Foundation of China (31772539); National Key R&D Program of China (2017YFE0121700); FAFU Science Fund for Distinguished Young Scholars (XJQ201819); Science and Technology Innovation Fund of FAFU (CXZX2018040).

**Acknowledgments:** We thank Forestry college of Fujian Agriculture and Forestry University for providing us with an insect breeding room for the breeding of *S. litura* species.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Mitochondria and Lysosomes Participate in Vip3Aa-Induced** *Spodoptera frugiperda* **Sf9 Cell Apoptosis**

**Xiaoyue Hou 1, Lu Han 1, Baoju An 1, Yanli Zhang 1, Zhanglei Cao 1, Yunda Zhan 1, Xia Cai 1, Bing Yan <sup>1</sup> and Jun Cai 1,2,3,\***


Received: 20 December 2019; Accepted: 12 February 2020; Published: 13 February 2020

**Abstract:** Vip3Aa, a soluble protein produced by certain *Bacillus thuringiensis* strains, is capable of inducing apoptosis in Sf9 cells. However, the apoptosis mechanism triggered by Vip3Aa is unclear. In this study, we found that Vip3Aa induces mitochondrial dysfunction, as evidenced by signs of collapse of mitochondrial membrane potential, accumulation of reactive oxygen species, release of cytochrome c, and caspase-9 and -3 activation. Meanwhile, our results indicated that Vip3Aa reduces the ability of lysosomes in Sf9 cells to retain acridine orange. Moreover, pretreatment with Z-Phe-Tyr-CHO (a cathepsin L inhibitor) or pepstatin (a cathepsin D inhibitor) increased Sf9 cell viability, reduced cytochrome c release, and decreased caspase-9 and -3 activity. In conclusion, our findings suggested that Vip3Aa promotes Sf9 cell apoptosis by mitochondrial dysfunction, and lysosomes also play a vital role in the action of Vip3Aa.

**Keywords:** Vip3Aa; lysosome; mitochondria; apoptosis; Sf9 cells

**Key Contribution:** Vip3Aa-induced apoptosis involves mitochondria and lysosomes.

#### **1. Introduction**

Vip3Aa is a protein produced by *Bacillus thuringiensis* (*Bt*) during vegetative growth. It can bind to the brush border membrane vesicles (BBMV) specifically in susceptible and non-susceptible insects [1–3]. Moreover, the brush border membrane binding sites of Vip3Aa are different from those of insecticidal crystal proteins (ICPs), and Vip3Aa could extend its activity to pests non-susceptible to ICPs. Consequently, it is widely accepted that Vip3Aa can not only broaden the insecticidal spectrum, it may also delay the resistance development in insects [3–5]. Thus, Vip3Aa is considered a second-generation insecticidal toxin and has been used in genetically modified crops, such as *Bt* cotton and *Bt* corn products [6].

The pore-forming model is generally accepted to explain the virulence of ICPs [7] and Lee et al. [3] corroborated that the Vip3 proteins share a similar mode of action. In short, the Vip3 proteins (protoxins) are ingested by the insect and activated to the active form (act-Vip3A) by the midgut proteases. After that, the act-Vip3A binds to its receptor on the BBMV and exerts toxicity to the midgut cells, eventually leading to the death of the pests. Additionally, Kunthic et al. [8] found that the pH could regulate the properties of the tetramer made by the act-Vip3Aa, which further supported the pore-forming

model and suggested that the pH could regulate the post-binding events such as membrane insertion or pore formation. Regarding the binding sites of the Vip3Aa, recent research has found some proteins interacting with Vip3Aa that are closely related to cell toxicity in *Spodoptera frugiperda* cells, such as S2, SR-C, and FGFR [5,9,10]. Additionally, Jiang et al. [9] found that the toxicity of Vip3Aa to Sf9 cells correlated with its endocytosis mediated by Sf-SR-C and that internalization is essential for Vip3Aa to exert its toxic effects.

Bel et al. [11] showed Vip3Aa provoked a wide transcriptional response in *Spodoptera exigua* larvae. The upregulated genes were involved in innate immune response and pathogen response, while the downregulated ones were mainly related to metabolism. However, genes related to the action of ICPs were found to be slightly overexpressed. Crava et al. [12] further indicated that Vip3Aa upregulated genes coding for antimicrobial peptides and lysozymes in *S. exigua* midgut. Ayra-Pardo et al. [13] reported a transcriptomic study, showing that the decreased translation rate could be an important adaptation for Vip3Aa resistance in *Heliothis virescens*. Hernández-Martinez et al. [14] suggested Vip3Aa could activate different insect response pathways that trigger the regulation of some genes, APN shedding, and apoptotic cell death. These results suggest that there are other mechanisms that are participating in cell death apart from the pore-forming model. Jiang et al. [15] observed that the Vip3Aa-treated Sf9 cells had some apoptosis characteristics, such as DNA breakage, mitochondrial membrane potential (ΔΨ*m*) collapse, and Sf-caspase-1 activation. Hernández-Martinez et al. [14] confirmed that there was apoptosis occurrence in midgut epithelial cells when *S. exigua* larvae were treated with Vip3Aa and Vip3Ca. However, how Vip3Aa induces apoptosis is unclear and further experiments will be needed to determine the underlying mechanism.

Apoptosis is indispensable to the homeostasis and development of organisms [16]. Bcl-2 family proteins are crucial regulators of cell survival and cell death. They are divided into anti- and pro-apoptotic proteins. After apoptotic stimulation, Bax, a pro-apoptosis protein, can transfer to mitochondria, resulting in mitochondrial membrane permeability increase and cytochrome c release. The mitochondrion, a highly sensitive organelle, plays a critical role in apoptosis. Increased mitochondrial membrane permeabilization may represent the point of no return of the lethal stressors-induced signal [17]. Cytochrome c normally localizes in the inner mitochondrial membrane through weak electrostatic interactions with acidic phospholipids. When mitochondria permeability increases, it releases to the cytoplasm and subsequently activates the apoptotic cascades. Anti-apoptotic proteins, such as Bcl-2 and Bcl-XL, inhibit apoptosis by locally preventing ΔΨ*m* loss [17,18]. Environmental stimuli may contribute to mitochondrial injury, which causes ΔΨ*m* collapse, oxidative stress, resulting in increased cellular ROS, changed Bcl-2 family protein levels, and apoptosis factor release [19–21].

In this paper, we try to further explore the mechanism of Vip3Aa-induced apoptosis and probe the signaling pathways and molecules involved in Vip3Aa-induced cell death.

#### **2. Results**

#### *2.1. The E*ff*ects of Vip3Aa on Sf9 Cell Viability and the Subcellular Localization of Vip3Aa in Sf9 Cells*

Sf9 cells were exposed to Vip3Aa (10, 20, 30, 40, or 50 μg/mL) for different times (24, 48, 60, and 72 h). Cell viability of Sf9 cells was assessed by the CCK-8 assay, by measuring the amount of orange–yellow formazan that is directly proportional to the number of living cells. As illustrated in Figure 1, when Sf9 cells were exposed to the same Vip3Aa concentration, the cell viability decreased as the time of treatment prolonged. If the Vip3Aa-treated time was the same, cell viability decreased with the increase of Vip3Aa concentration. Vip3Aa (final concentration, 40 μg/mL) treatment for 48 h reduced the cell viability of Sf9 cells to about 50%. Thus, the final concentration of Vip3Aa used in the following experiments was 40 μg/mL.

**Figure 1.** Viability impacts of Vip3Aa on Sf9 cells. The Sf9 cells were exposed to different concentrations of Vip3Aa for 24, 48, 60, and 72 h, respectively. Significant tests from the corresponding controls (without Vip3Aa treatment) are indexed via \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.001.

Jiang et al. [9] revealed that Vip3Aa could enter into the Sf9 cells via endocytosis. Since the lysosome is the endpoint of endocytosis, we further explored the subcellular localization of Vip3Aa in Sf9 cells. Sf9 cells were exposed to Vip3Aa-RFP for different times (2, 4, and 6 h). As shown in Figure 2, there was abundant co-localization of Vip3Aa and lysosomes from 4 h after Vip3Aa treatment. These results suggested that lysosomes might be involved in the action of Vip3Aa.

**Figure 2.** Co-localization of Vip3Aa and lysosomes in Sf9 cells. Cells were treated with Vip3Aa-RFP for 0, 2, 4, and 6 h, respectively, and were stained with fluorescent probe LysoSensor™ Green DND-189 at 28 °C for 45 min. Then, the cells were observed under a confocal laser scanning microscope. Scale bar, 20 μm.

#### *2.2. Vip3Aa Impairs Mitochondrial Function and Induces Cytochrome c Release*

Mitochondria are the meeting point of many apoptotic signals, so we examined the mitochondrial ultrastructure (red arrows) in Sf9 cells by transmission electron microscope (TEM) (Figure 3A). The Vip3Aa-untreated cells showed a normal ultrastructure with an intact cristae structure. However, the number of twisted and swollen mitochondria increased from 12 to 24 h after Vip3Aa treatment. After 36 h, the outer membranes of most mitochondria were intact, but the cristae structures were disrupted.

**Figure 3.** Effects of Vip3Aa on mitochondria in Sf9 cells. (**A**) Representative photographs of mitochondria ultrastructure in Sf9 cells after exposure to Vip3Aa, obtained by TEM. N, nucleus. Nm, nuclear membrane (blue arrows). M, mitochondria (red arrows). Magnification, 30000 ×. Scale bar, 1 μm. (**B**) Effects of Vip3Aa on ROS production and mitochondrial membrane potential (ΔΨ*m*) in Sf9 cells, which were determined by the fluorescent probe DCFH-DA and Rhodamine123, respectively. Scale bar, 20 μm.

Mitochondria are not only the main source of endogenous ROS but also the "absorption bank" of ROS. Mitochondria play an essential role in regulating ROS metabolism. In turn, ROS also impacts the function of mitochondria [22]. We then explored the impact of Vip3Aa on the ROS in Sf9 cells. For Vip3Aa-treated cells, ROS levels increased within 12 h, peaked at 24 h, and then decreased (Figure 3B).

Mitochondrial membrane potential (ΔΨ*m*) plays a key role in mitochondria function [23]. Rhodamine 123 was used to detect ΔΨ*m*. Results showed a significant decrease in ΔΨ*m* appeared firstly at 24 h after Vip3Aa treatment, and the fluorescence intensity reduced to a lower level at 48 h (Figure 3B).

To determine the influence of Vip3Aa on cytochrome c distribution, we evaluated the cytochrome c content in the cytosol and mitochondria via Western blotting. As indicated in Figure 4A and Figure S1D, the cytochrome c content in the cytoplasm increased, while that in the mitochondria decreased significantly. Subsequently, this phenomenon became more apparent.

**Figure 4.** Subcellular distribution of cytochrome c and the levels of mitochondria-associated proteins in Sf9 cells after Vip3Aa treatment. (**A**) Subcellular distribution of cytochrome c after Vip3Aa treatment. Cytochrome c distribution in cytoplasm and mitochondria was detected by Western blotting. (**B**) Cytochrome c distribution in Sf9 cells after different treatments. CsA (5 μM) or BKA (10 μM) pretreated the Sf9 cells for 2 h before Vip3Aa treatment. (**C**) The mitochondria-associated proteins levels in Sf9 cells after Vip3Aa treatment. The original pictures of Western blotting (with protein maker) are shown in Supplementary Materials (Figure S1).

To further explore the way of cytochrome c release, mitochondria permeability transition pore (mPTP) inhibitors, CsA and BKA, were used. Results showed that both inhibitors prevented cytochrome c release partly, while BKA exerted a stronger inhibitory effect than CsA did (Figure 4B). These results suggested that an mPTP-dependent mechanism was involved in cytochrome c release in Vip3Aa-treated Sf9 cells.

#### *2.3. E*ff*ects of Vip3Aa on the Mitochondria-Associated Proteins Levels*

The protein levels of the Bcl-2 family are related to mitochondrial function. As shown in Figure 4C and Figure S1E, Bax expression increased, while Bcl-2 and Bcl-XL decreased with the extension of Vip3Aa treatment time.

In the mitochondrial pathway, cytochrome c release induces the formation of apoptotic protein complexes, which convert pro-caspase-9 into active caspase 9. Subsequently, caspase-9 will further activate caspase-3 and leads to cell apoptosis. So, we investigated whether caspase activation was involved in the Vip3Aa-induced cell death. Results suggested that the levels of cleaved-caspase-9 and cleaved-caspase-3 increased in different degrees with the extension of Vip3Aa treatment time (Figure 4C and Figure S1F). Additionally, when cytochrome c in the cytosol increased significantly, the protein level of cleaved-caspase increased accordingly. These results indicated that Vip3Aa induced dysregulation of mitochondria-associated proteins and subsequently led to the activation of caspases.

#### *2.4. E*ff*ects of Vip3Aa on Lysosome Morphological and Physicochemical Property*

Several lines of evidence suggest that the lysosomal pathway contributes to apoptosis. To explore the impact of Vip3Aa on lysosomes, we observed the lysosomal ultrastructure in Sf9 cells by TEM (Figure 5A). The control cells showed a few lysosomes and the cytoplasm was homogeneous. However, the type of lysosomes in the Vip3Aa-treated cells became diverse, and some lysosomes increased distinctly in volume. Meanwhile, we measured the ability of lysosomes to retain acridine orange (AO). As shown in Figure 5B, the fluorescence signals for Sf9 cells were mostly kept in the Q2 region (normal cells), while the percentage of the Q3 region (cells with weak lysosomes) increased from 6.07% to 29.16% with the prolongation of Vip3Aa treatment time. These data showed that Vip3Aa increased the proportion of cells with abnormal lysosomes, and the ability of these lysosomes to keep AO was poor.

**Figure 5.** Effects of Vip3Aa on lysosomes in Sf9 cells. (**A**) Representative photographs of lysosomes ultrastructure in Sf9 cells after exposure to Vip3Aa, obtained by TEM. N, nucleus. Nm, nuclear membrane (blue arrows). L, lysosomes (yellow arrows). Magnification, 0 h, 24 h, 36 h, and 48 h, 10000 ×; 12 h, 5000 ×. (**B**) The physicochemical property of lysosomes was detected by acridine orange (AO) staining in Sf9 cells. (**C**) The lysosomal pH in Sf9 cells was detected using LysoSensor Yellow/Blue DND-160. Significant tests from the corresponding controls (without Vip3Aa treatment) are indicated by \* *p* < 0.05, \*\* *p* < 0.01, and \*\* \**p* < 0.001.

Additionally, we also measured lysosomal pH to further study the impact of Vip3Aa on lysosomes. The lysosome pH value of the Vip3Aa-untreated cells was estimated at 4.91, whereas the lysosomal pH increased at 5.48 after Sf9 cells were exposed to Vip3Aa for 36 h (Figure 5C).

#### *2.5. The Relationship between Sf9 Cell Cathepsins and Vip3Aa-Induced Apoptosis and Cytotoxicity*

Lysosomes could be involved in apoptosis via lysosomal proteases, especially cathepsins. Therefore, we measured the mRNA level of cathepsins B, L, and D, which are the significant proteins in the lysosome function. The results indicated that the expression levels of cathepsins (L and D) increased differently depending on the cathepsins analyzed (Figure S2). The mRNA level of cathepsin L and cathepsin D peaked at 36 and 6 h, respectively. However, the expression level of cathepsin B had little change after the cells exposed to Vip3Aa in all the times analyzed.

Meanwhile, we detected the effects of cathepsins (B, L, and D) on Vip3Aa-induced apoptosis and toxicity using the inhibitors CA-074me (a cathepsin B inhibitor), Z-Phe-Tyr-CHO (a cathepsin L inhibitor) and pepstatin (a cathepsin D inhibitor). As illustrated in Figure 6A, the percentage of late apoptotic cells was 0.1%, which rose to 12.56% after 48 h of Vip3Aa treatment. Z-Phe-Tyr-CHO and pepstatin reduced the percentage of late apoptotic cells from 12.56% to 1.7% and 1.44%, respectively. However, CA-074me, a cathepsin B inhibitor, had a little impact on the late apoptotic rate. Compared with Vip3Aa used alone, when Z-Phe-Tyr-CHO was used, there was a little effect on the proportion of early apoptotic cells, but the proportion of late apoptotic cells decreased significantly. However, when pepstatin was used, the proportion of early and late apoptotic cells all decreased significantly. These results suggested that cathepsin D contributes to Vip3Aa-induced apoptosis more than cathepsin L, and cathepsin D plays a more important role in apoptotic signal transduction and enhancement.

**Figure 6.** Effects of cathepsin inhibitors on Vip3Aa-treated Sf9 cells. Sf9 cells were pretreated with the inhibitors, CA-074me (10 μM), Z-Phe-Tyr-CHO (10 μM), or pepstatin (15 μM) 2 h before Vip3Aa was added. (**A**) The apoptotic rate of Vip3Aa-treated cells incubated with or without inhibitors. The apoptotic rate was evaluated by Annexin V-FITC/PI stains. (**B**) The cell viability of Vip3Aa-treated cells incubated with or without inhibitors. The cell viability was measured by a CCK-8 assay. Significant tests from the corresponding controls (without Vip3Aa treatment) are indicated by NS, not significant, \*\* *p* < 0.01.

We also detected the effect of inhibitors on cell livability. The results showed that Z-Phe-Tyr-CHO and pepstatin increased cell livability from 51.3% to 75.1% and 77.8%, respectively (Figure 6B). However, CA-074me had little effect on cell viability. These results suggested that cathepsins (L and D) play a critical role in Vip3Aa-induced cell death rather than cathepsin B.

#### *2.6. E*ff*ects of Inhibition of Cathepsins (L and D) on Cytochrome c Distribution and Caspase-9 and -3 Activity*

Cytochrome c plays a vital role in apoptosis when the mitochondrial pathway is the executor. Thus, we investigated whether the cathepsin (L and D) inhibitors, Z-Phe-Tyr-CHO and pepstatin, could impact the release of cytochrome c. As shown in Figure 7A, the cathepsin (L and D) inhibitors, especially pepstatin, could reduce cytochrome c release. These results indicated that the function of lysosomes affected Vip3Aa-induced mitochondrial dysfunction.

**Figure 7.** Effects of cathepsins inhibitors on caspases activity and cytochrome c release. (**A**) Impacts of cathepsin (L and D) inhibitors on cytochrome c distribution. The original pictures of Western blotting (with protein maker) are shown in Supplementary Materials (Figure S3). (**B**) Impacts of cathepsin inhibitors on caspase-9 activity. (**C**) Impacts of cathepsin inhibitors on caspase-3 activity. Significant tests from the corresponding controls (without Vip3Aa treatment) and densitometry of the protein bands are indicated by NS, not significant, \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.001.

Vip3Aa induced apoptosis in Sf9 cells in a caspase-dependent mode [15]. Mitochondrial membrane permeation and the release of cytochrome c from mitochondria activate caspase-9. Caspase-9 further activates caspase-3 and induces apoptosis. To further confirm the influence of lysosomes on Vip3Aa-induced mitochondrial pathway, we detected the caspase-9 and -3 activity without or with

the cathepsin inhibitors. As shown in Figure 7B,C, the activity of caspase-9 and -3 increased from 24 h, peaked at 48 h after Vip3Aa treatment without cathepsin inhibitor. However, the caspase-9 and -3 activity decreased significantly after Vip3Aa treatment for 36 h with Z-Phe-Tyr-CHO or pepstatin. As expected, CA-074me had a little impact on caspase activity, especially on caspase-3. These findings indicated that lysosomes are involved in Vip3Aa-induced apoptosis and that cathepsins (L and D) have a vital impact on the Vip3Aa-induced mitochondrial pathway.

#### **3. Discussion**

Vip3Aa is a potent toxin against lepidopteran pests, especially to some pests of Noctuidae which are insensitive to ICPs. Recently, studies have shown that Vip3Aa could exert cytotoxicity by triggering apoptosis of insect cells and tissues besides formatting pores [3,5,14,15]. However, the specific mechanism of apoptosis induced by Vip3Aa remains unclear. Hence, we dissected the mechanism of mitochondrial pathway in Vip3Aa-induced apoptosis and found the lysosomes play a crucial role in Vip3Aa-induced apoptosis. The action mechanism of Vip3Aa found in Sf9 cells may also occur in insect intestinal epithelial cells.

Apoptosis includes two important pathways: the extrinsic pathway and the intrinsic pathway, mediated by the death receptor and mitochondria, respectively [24]. Jiang et al. [15] found that mitochondrial membrane potential decreased in Vip3Aa-treated Sf9 cells. We confirmed that Vip3Aa reduced Sf9 cell viability and caused mitochondria morphological alterations, which included swelling and disrupted cristae structure. In our study, we also found that Vip3Aa-induced apoptosis was mediated by mitochondrial dysfunction caused by loss of ΔΨ*m*, which subsequently led to cytochrome c release and caspase-9 and -3 activation. Bcl-2 family proteins and caspases are involved in programmed cell death by regulating the protein levels. Moreover, the trends of capase-3 activity were consistent with those of caspase-9 in Vip3Aa-treated Sf9 cells. These results supported that the intrinsic mitochondrial pathway is involved in Vip3Aa-induced apoptosis in Sf9 cells.

Studies have indicated that the extrinsic pathway can be triggered by activating the death receptor on the cell membrane [25]. Additionally, the receptor-mediated pathway contains two types of mechanisms. In type I cells, the extrinsic apoptotic pathway leads to the activation of caspase-8, which directly activates effector caspases (caspase-3), causing apoptosis [26]. Nevertheless, in type II cells, the two apoptotic pathways, i.e., extrinsic pathway and intrinsic pathway, can be linked by caspase-8, which can cleave non-activated Bid protein into truncated Bid (tBid) [27]. tBid could activate Bax, resulting in cytochrome c release and caspase-3 activation [28]. To explore whether the death receptor pathway involves Vp3Aa-induced apoptosis, we also detected the activity of caspase-8 (Figure S4). The caspase-8 activity increased a little bit from 12 to 36 h, but it was lower than that of untreated cells from 48 h. This suggested that caspase-8 might not contribute to the activation of caspase-3. Moreover, the two Vip3Aa receptors SR-C and FGFR did not contain the death domain [9,10]. Jiang et al. [9] revealed the toxicity of Vip3Aa to Sf9 cell correlates with its endocytosis mediated by Sf-SR-C and the internalization is essential for Vip3Aa to exert its toxic effects. On this basis, we further showed that internalized Vip3Aa impacted the features of lysosomes. These results suggested that Vip3Aa-induced apoptosis might involve the internalization of Vip3Aa and the denaturation of lysosomes rather than the death receptor-mediated pathway.

In this study, we found that the abundant colocalization of Vip3Aa and lysosome in Sf9 cells and Vip3Aa had a distinct effect on lysosome morphological and physicochemical properties. Thus, we thought the lysosomes contain the Vip3Aa, and the deformed lysosomes might be the consequence of Vip3Aa action. Duve et al. [29] put forward that lysosomes play a role in apoptosis in 1966. A new death theory, the lysosome–mitochondria axis, mainly emphasized that hydrolytic enzymes were released to the cytosol from the lysosome when lysosomal membranes were permeabilized, resulting in mitochondrial dysfunction, cytochrome c release, and caspase activation. Many studies had reported that cathepsins could be involved in the signaling of apoptosis. Cathepsin D can activate Bax and the active form of Bax translocates to the mitochondria, leading to the opening of transition pores on

the mitochondrial membrane, which cause apoptosis factors such as cytochrome c to be released [16]. Another study indicated that cathepsin L acts as a death signal integrator and cytosolic cathepsin L regulated the cytochrome c release and caspase-3 activity in cervical cancer cells [30]. In this study, the results (Figures 6 and 7) showed that cathepsin (L and D) inhibitors could protect Sf9 cells from Vip3Aa and suppress cytochrome c release and inhibit the caspase-9 and -3 activity, suggesting that cathepsins (L and D) played a significant role in Vip3Aa-induced cell death. Moreover, some studies found that cathepsin B associated with programmed cell death of the fat body cells in the process of silkworm metamorphosis [31]. However, there was a little effect of cathepsin B on the Vip3Aa-treated Sf9 cells. As for the role of cathepsins (L and D), they may be released to the cytoplasm and could cleave Bid to tBid, and the latter triggers the mitochondrial outer membrane permeabilization, resulting in mitochondria dysfunction. On the other hand, cathepsins (L and D) may contribute to activating Vip3Aa in lysosomes. Many studies indicated that cathepsins were involved in the physiological reaction of insects, but the exact mechanism is unclear. In this study, lysosomes were firstly found to be involved in the process of Vip3Aa-induced apoptosis. However, the mechanism needs further investigation.

We found the caspase inhibitor (Z-VAD-FMK) could not protect all the Sf9 cells from Vip3Aa (Figure S5). This result suggested that some other apoptosis-independent cell death mechanisms, such as pore-forming, might be involved in cell death caused by Vip3Aa [3,8]. Some microbial toxins, such as aerolysin produced by *Aeromonas hydrophila* and α-toxin generated by *Staphylococcus aureus*, could contribute to pore-forming and apoptosis in their target cells [32]. Similarly, Vip3Aa may cause insect cell death through two mechanisms at the same time.

In conclusion, in Sf9 cells, we showed that the mitochondria pathway serves as the executor in Vip3Aa-induced apoptosis, while lysosomes are involved in Vip3Aa-induced mitochondrial dysfunction and apoptosis. Our findings can provide a venue for promoting the knowledge of Vip3Aa action.

#### **4. Materials and Methods**

#### *4.1. Cell Culture and Reagents*

The Sf9 cells were cultured in Sf-900 II SFM medium (Gibco, 10902088, Grand Island, NY, USA) supplemented with 6% FBS (GIBCO, Grand Island, NY, USA), at 28 °C. RIPA buffer (#9806S), and antibodies against Bax (#2772), Bcl-2 (#15071), caspase-9 (#9508), cytochrome c (#11940), and β-actin (#8457) were obtained from Cell Signaling Technology (Beverly, MA, USA). Acridine orange (#A8120) was purchased from Solarbio Life Science (Beijing, China). Antibody against Bcl-XL (#abs131907) was purchased from Absin Bioscience (Shanghai, China). Antibody against Caspase-3 (#A2156) was purchased from ABclonal (Wuhan, China). Goat anti-mouse IgG-HRP conjugate (#sc-2005) or anti-rabbit (#sc-2357) was purchased from Santa Cruz (Santa Cruz, TX, USA). DCFH-DA (#S0033) and Rhodamine123 (#C2007) were purchased from Beyotime Biotechnology (Nanjing, China). LysoSensor™ Green DND-189 (#40767ES50) was purchased from Yeasen Biotechnology (Shanghai, China).

#### *4.2. Vip3Aa Purification*

pET-28a (+) vector was used to construct a recombinant expression plasmid. The BL21 (DE3) strains transferred with pET28a-Vip3Aa were cultured to OD600 0.8–1.0, and IPTG (0.5 mM) were used to induce the protein expression at 16 ◦C for 12–16 h. Then, the cells were collected, broken by ultrasonication, and purified using Ni SepharoseTM affinity column. The Vip3Aa was dialyzed in a buffer containing 25 mM Tris-Hcl (pH 7.4) and 150 mM NaCl at 4 ◦C. The result of purified Vip3Aa was shown in Figure S6. The concentration of Vip3Aa was measured via the protein-dye method of Bradford. BSA was used as a standard protein. The full-length Vip3Aa was used directly in Sf9 cells.

#### *4.3. Cell Viability Assay*

The cell viability was detected using the CCK-8 Counting Kit (Dojindo, Kumamoto, Japan). Cell suspensions (100 <sup>μ</sup>L, 2.5 <sup>×</sup> 105 cells/mL) were pipetted into a 96-well plate and incubated overnight at 28 °C. Then, Vip3Aa was added into the suspensions. The cells were exposed to Vip3Aa for 24, 48, 60, and 72 h. The final concentration of Vip3Aa was 10, 20, 30, 40, and 50 μg/mL. Sf-900 II SFM medium and cell suspensions without Vip3Aa were used as blank group and control group, respectively. Then, CCK-8 reagent (10 μL) was added and incubated in darkness for 2–4 h at 28 °C. The results were monitored at 450 nm using a microplate reader (PerkinElmer, Boston, MA, USA). The experiments were performed six times. Cell viability was the ratio of absorbance of Vip3Aa-treated group/control group.

#### *4.4. Vip3Aa Subcellular Localization in Sf9 Cells*

The cells were exposed to Vip3Aa-RFP for 0, 2, 4, and 6 h. Then, the cells were incubated with Sf-900 II SFM medium containing 1 μM LysoSensor™ Green DND-189 at 28 °C in the darkness for 45 min. The cells were washed with phosphate-buffered saline (PBS) (pH 7.4) three times and imaged with a Zeiss LSM710 fluorescence microscope.

#### *4.5. Transmission Electron Microscopy (TEM)*

Cell suspensions (5 <sup>×</sup> <sup>10</sup>5–1 <sup>×</sup> <sup>10</sup><sup>6</sup> cells/mL) were incubated overnight in 25 cm<sup>2</sup> flasks. The cultures were exposed to Vip3Aa (final concentration, 40 μg/mL) for 12, 24, 36, and 48 h, respectively. Transmission electron microscope (TEM) was used to observe and record the ultrastructure of Sf9 cells. The cells were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde in phosphate-buffered saline (PBS, pH 7.4) for 2 h after washing with PBS three times. Then, the fixed cells were treated with 1% osmic acid (OsO4) at 25 °C for 1 h after washing with PBS. The cell samples were dehydrated in different concentration ethanol solutions, soaked, and embedded in EPON812. Ultrathin (60 nm) sections were cut and counterstained with lead citrate and uranyl acetate. The sections were observed with TEM (JEOL-1200EX).

#### *4.6. Measurement of Intracellular ROS and Mitochondrial Membrane Potential (*ΔΨ*m)*

DCFH-DA and Rhodamine 123 were utilized to measure intracellular ROS [33] and ΔΨ*m*, respectively. The cells were exposed to Vip3Aa (final concentration, 40 μg/mL) for different times. Then, the cells were incubated with Sf-900 II SFM medium containing 10 mM of DCFH-DA or 50 nM Rhodamine 123 at 28 °C in the darkness for 30 min. The cells were washed with PBS (pH 7.4) three times and imaged with a Zeiss LSM710 fluorescence microscope.

#### *4.7. Total Protein and Cytosolic Protein Extraction*

The cells were exposed to Vip3Aa (final concentration, 40 μg/mL) for different times. The cells were lysed in 350 μL RIPA buffer with 1 mM PMSF and incubated on ice for 15 min after washing with PBS (pH 7.4) three times. The suspension was centrifuged at 12,000× *g* for 15 min. Then, the supernatant, which was the total protein extraction, was collected carefully.

Cells were washed and collected by centrifugation at 200 × *g* for 5 min. The cells were resuspended with 500 μL isotonic buffer (IB, 10 mM HEPES, 200 mM mannitol, 1 mM EGTA, 70 mM sucrose). The cell suspension was centrifuged at 3000× *g* for 5 min. The collected cells were resuspended in 500 μL IB with 20 mM NaF, 20 mM Na3VO4 and 1 mM PMSF. Then, 26-G needles were used to homogenize the cell suspension, which was passed through 14 times and stood on ice for 5 min. The suspension was centrifuged at 4 °C, 10,000× *g* for 15 min. Centrifugation sediment contains lysosomes and mitochondria. Then, the supernatant was diverted to a fresh cold centrifuge tube and centrifuged at 4 °C, 14,000× *g* for 30 min in an ultracentrifuge. The supernatant, which was the cytosolic protein extraction, was collected carefully.

#### *4.8. Quantitative Real-Time PCR*

Trizol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract total RNA. Chloroform and isopropanol were used to isolate RNA. The primers used in this study are listed in Table 1. A PrimescriptTM RT reagent kit with gDNA Eraser (TakaRa, Dalian, China) was utilized to reverse-transcribe RNA. The quantitative real-time PCR was performed with SYBR® Premix Ex Taq™ (TakaRa, Dalian, China) in an ABI Prism 7900HT Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA). *GAPDH* was used as the control for normalization by the 2-ΔΔCt method [34].


**Table 1.** Primers used in this study.

#### *4.9. Western Blotting Analysis*

A BCA Protein Assay kit (TIANGEN, Beijing, China) was used to test the concentrations of protein samples. Next, 12% SDS-PAGE electrophoresis was utilized to separate the target proteins, which were transferred onto the PVDF membrane. Primary antibodies were anti-caspase-3 (1:1000), anti-caspase-9 (1:500), anti-Bcl-XL (1:500), anti-Bcl-2 (1:500), anti-Bax (1:500), anti-cytochrome c (1:500), anti-cathepsin L (1:500), and anti-β-actin (1:500). Mouse anti-rabbit IgG-HRP (1:1000) and goat anti-mouse IgG-HRP (1:1000) were the secondary antibodies. Finally, the PVDF membranes were visualized using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Milan, Italy).

#### *4.10. Acridine Orange (AO) Staining Analysis*

The AO staining analysis was performed by flow cytometry (Becton Dickinson, USA). The Sf9 cells were treated with Vip3Aa (final concentration, 40 μg/mL) for different times. Then, the cells were incubated with Sf-900 II SFM medium containing 5 μg/mL AO at 28 °C in the darkness for 10 min. The stained cells were used to analyze the fluorescence distribution (FL1-H/FL3-H) after washing with PBS (pH 7.4) three times. After adjusting the fluorescence compensation of the channels, the number of recorded cells was 10,000.

#### *4.11. Lysosomal pH Assay*

The LysoSensor Yellow/Blue DND-160 (Life Technologies, Carlsbad, CA), a lysosomal pH indicator, was used to measure the Sf9 cells lysosomal pH. Cell suspensions (100 <sup>μ</sup>L, 2.5 <sup>×</sup> 105 cells/mL) were pipetted into a 96-well black plate. All the cells were incubated with Sf-900 II SFM medium containing 5 μM fluorescent probe at 28 °C in the darkness for 5 min. Then, the cells were washed and cultured in an MES calibration buffer (1.2 mM MgSO4, 115 mM KCl, 5 mM NaCl, 25 mM MES, pH 3.5–6.0) containing 10 μM monensin and 10 μM nigericin. The fluorescence value (Ex340 nm/Em540 nm and Ex380 nm/Em540 nm) was monitored by a microplate reader (PerkinElmer, Boston, MA, USA). The pH calibration curve was generated using ratios of the two light emission intensities and the corresponding pH value. To find the effect of Vip3Aa on lysosomal pH, the Vip3Aa-treated cells were incubated with Sf-900 II SFM medium containing 5 μM fluorescent probe at 28 °C in the darkness for 5 min, washed, and resuspended in MES buffer (pH 7.0) and detected using a microplate reader (PerkinElmer, Boston, MA, USA). The lysosomal pH was estimated using the ratios and the pH calibration curve. Sf-900 II SFM medium and cell suspensions without Vip3Aa were used as blank group and control group, respectively.

#### *4.12. Apoptosis Assay*

Sf9 cells were treated with Vip3Aa (final concentration, 40 μg/mL) for 48 h with/without cathepsins inhibitor for 2 h. We evaluated the proportion of apoptotic cells using the FITC annexin V apoptosis detection kit I (BD Biosciences, USA). Cells without Vip3Aa-treated were used as a control group. After washing twice with PBS (100 × g, 5 min), cells incubated with 1 × binding buffer containing FITC annexin V at 28 °C in the darkness for 30 min. Then, 1 × binding buffer containing propidium iodide was added to each sample. After incubating at 28 °C in the darkness for 5 min, the cells were monitored with a flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).

#### *4.13. Caspase Activity Analysis*

Sf9 cells were treated with Vip3Aa (final concentration, 40 μg/mL) for different time with/without cathepsins inhibitor for 2 h. Caspase-Glo® assay kit (Promega, Madison, WI, USA) was utilized to determine caspase activity. Cell suspensions (100 <sup>μ</sup>L, 2.5 <sup>×</sup> 105 cells/mL) were pipetted into a 96-well white plate and incubated overnight at 28 °C. Then, Vip3Aa was added into the suspensions. The cells were exposed to Vip3Aa for 12, 24, 36, 48, 60, and 72 h. Sf-900 II SFM medium and cell suspensions without Vip3Aa were used as blank group and control group, respectively. Caspase-Glo® Reagent was prepared according to the protocol and all the operations should be performed in the darkness. Equilibrate the reagent and plates to room temperature. Caspase-Glo® Reagent (100 μL) was added to the plates containing cells in Sf-900 II SFM medium. A plate shaker was used to mix the plates containing cells and reagent at 300–500 rpm for 0.5–2 min. Then, the plates were incubated at room temperature for 2 h. Finally, the luminescence of each plate was detected using a microplate reader (PerkinElmer, Boston, MA, USA).

#### *4.14. Statistical Analysis*

The results were obtained from at least three independent experiments. The densitometry values were evaluated by the software Image J. Origin 8.0 (OriginLab, Northampton, MA, USA) was used to draw the graphs. The significance was tested by one-way analysis of variance utilizing Student t test. If *p*-value ≤0.05, the results were considered significant.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2072-6651/12/2/116/s1, Figure S1: Impacts of Vip3Aa in Sf9 cells on the cytochrome c distribution and the level of mitochondria-associated proteins. (A) Cytochrome c distribution in mitochondria and cytosol was detected by Western blotting. (B) Cytochrome c distribution in Sf9 cells after various treatments. CsA (5 μM) or BKA (10 μM) was pretreated the Sf9 cells for 2 h before adding Vip3Aa. (C) Mitochondria-associated protein levels in Sf9 cells were tested using Western blotting. (D, E, and F) Densitometry analysis of (A) and (C). Figure S2: Effect of Vip3Aa on cathepsins mRNA level in Sf9 cells at different time. Figure S3: Effects of cathepsin (L and D) inhibitors on cytochrome c distribution. (A) cytochrome c distribution in cytosol was detected by Western blotting. (B) Densitometry analysis of (A). Figure S4: Effects of Vip3Aa on caspase-8 activity. Significant tests from the corresponding controls (without Vip3Aa treatment) are indicated by NS, not significant, \**p* < 0.05, \*\**p* < 0.01, \*\*\**p* < 0.001. Figure S5: Effects of caspase-3 inhibitor on Sf9 cell viability. Significant tests from the corresponding controls (without Vip3Aa treatment) are indicated by \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001. Figure S6: Result of purified Vip3Aa detected by SDS-PAGE electrophoresis analysis. M, protein maker 26619.

**Author Contributions:** Data curation, X.H.; funding acquisition, J.C.; investigation, X.H., L.H., B.A., Y.Z. (Yanli Zhang), Z.C., Y.Z. (Yunda Zhan), X.C., and B.Y.; methodology, X.H. and L.H.; project administration, J.C.; supervision, J.C.; writing—original draft, X.H.; writing—review and editing, J.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by grants from the National Key R&D Program of China (No. 2017YFD0200400), and the National Natural Science Foundation of China (No. 31670081 and 31371979).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Toxins* Editorial Office E-mail: toxins@mdpi.com www.mdpi.com/journal/toxins

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com

ISBN 978-3-0365-2050-6