Functional characterization of PBP1 gene in Helicoverpa armigera (Lepidoptera: Noctuidae) by using the CRISPR/Cas9 system

Pheromone binding proteins (PBPs) are thought to play crucial roles in perception of the sex pheromones particularly in noctuid moths, but this is rarely in vivo evidenced due to lacking an effective technique. Here, we reported an in vivo functional study of PBP1 in the important lepidopteran pest Helicoverpa armigera (HarmPBP1), by using the CRISPR/Cas9 system. Efficient and heritable mutagenesis was achieved by egg injection of mixture of Cas9-mRNA and HarmPBP1-sgRNA. The TA cloning and sequencing revealed various insertion and/or deletion (indel) mutations at the target site. Among those, one mutation resulted in a premature stop codon at the target site, which led to a highly truncated protein with only 10 amino acids. The HarmPBP1 with this mutation would completely loss its function, and thus was used to select the homozygous mutant insects for functional analysis. The electroantennogram recording showed that the mutant male adults displayed severely impaired responses to all three sex pheromone components (Z11-16:Ald, Z9-16:Ald and Z9-14:Ald). Our study provides the first in vivo evidence that HarmPBP1 plays important role in perception of female sex pheromones, and also an effective methodology for using CRISPR/Cas9 system in functional genetic study in H. armigera as well as other insects.

enzymes (PDEs) 37 . PBPs are small soluble proteins, highly concentrated in the sensillum lymph of lepidopteran antennae 38 . It is previously supposed to be simply pheromone carriers, by binding and transporting hydrophobic pheromone molecules across the aqueous lymph onto the ORs on dendritic membranes of the sensory neurons 39,40 . However, as multiple PBP genes have been found in H. armigera and other noctuid moths [41][42][43][44] , it is suggested that PBPs may function as recognizers, i.e. different PBPs selectively bind different component of the pheromone blend. Indeed, ligand-binding experiments showed that HarmPBP1 binds strongly to the two major components (Z11-16:Ald and Z9-16:Ald) 45,46 , but no binding to Z9-14:Ald 45 , while HarmPBP2 and HarmPBP3 showed only weak or no binding to all three components 45 . However, these results have not been validated by an in vivo functional study.
In the present study, by using the CRISPR/Cas9 system, we successfully obtained the high rate and heritable HarmPBP1-targeted mutagenesis and established a mutant H. armigera strain. With the mutant insects, we conducted the in vivo functional study of HarmPBP1 gene. Our study clearly demonstrates that HarmPBP1 plays important role in perception of all 3 sex pheromones, and has no obvious preference among the 3 pheromone components. In addition, our study provides an important reference for loss-of-function gene analysis in H. armigera.

Results
Targeted mutations in eggs after injection of Cas9/sgRNA. A mixture of 300 ng/μl Cas9-mRNA and 150 ng/μl sgRNA was injected into 308 newly laid eggs. To check the mutation efficiency, 20 eggs were randomly selected and pooled to extract the genomic DNA for PCR amplification, and then the PCR products were subjected to RED treatment and gel analysis (Fig. 1A). The result showed that portion of the PCR products from the treated eggs were uncleaved by RED treatment, indicating the induction of mutations in the target site of the HarmPBP1 gene. Based on the relative intensity, the mutation frequency of HarmPBP1 gene in the pooled eggs was calculated as 36.9%. In addition, the direct sequencing of the PCR products also showed occurrence of the mutations, indicated by the multiple peaks at the target site (Fig. 1B).
To investigate the exact mutations, the uncleaved band on the gel was cut, purified and subjected to the TA cloning. In total, 15 positive clones were successfully sequenced. The 13 clones showed 7 different types of mutations, including 3 deletions, 2 insertions, one mixed indel, and one with change in two nucleotides. Of the 13

Screening of homozygous mutant strains induced by CRISPR/Cas9. A single pair mating strategy
was used for the screening from generation G0 to G3. The number of single pairs made, number of single pairs that laid fertilized eggs, and the single pairs used for further screening were listed in Table 1. For each generation, genotypes of the single pairs that laid fertilized eggs were determined individually by the RED assay (Fig. 2), and accordingly the single pairs were selected for further screening (Table 1).
In G1 generation, the moths that laid fertilized eggs were also determined for their DNA sequence, displaying different mutant genotype (Fig. S1). In particular, an insertion of GT was occurred in target site of HarmPBP1 from parents of single pair #8, which results in the premature termination in protein translation (Fig. 3). This premature termination would lead to a truncated protein of only 10 amino acids, compared with 143 amino acids for the wild type protein. Therefore, G2 offspring of single pair #8 was used for further selection of the homozygous mutants by inbred crossing. In G2 generation, both parents of single pair #7 and the female parent of single pair #9 were homozygous, while other G2 parents were still heterozygous. The G3 male offspring of these 2 single pairs (#7 and #9) were expected to consist of homozygotes and heterozygotes, which were used for functional analyses by the electroantennogram (EAG) measurements (Fig. 2). EAG measurements of mutant males. The electroantennogram (EAG) experiment was conducted with the G3 males, to clarify the function of HarmPBP1 in perception of the sex pheromones. All males were checked for the genotype (homozygous or heterozygous) by RED assay (Fig. 2), after the EAG measurement. Compared  Fig. 2, for those adults from the single pairs that laid fertilized eggs.  Figure S3.
with the no injection wild type males, both homozygous and heterozygous mutants showed significantly lower EAG responses to all the 3 sex pheromone components (Z11-16:Ald, Z9-16:Ald and Z9-14:Ald) at the tested dosage (500 ng). In addition, homozygous mutants showed significantly lower EAG responses than heterozygous mutants to the first two pheromone components (Fig. 4).
Off target effect of the mutagenesis by CRISPR/Cas9 system. Without the available genome data of H. armigera, a transcriptomic data was used to search for the potential off target sequences using CasOT tool with the default settings. Total of 306 fragment sequences were found, and the top 4 sequences (Table S1) were determined in 4 G0 chimaera adults, by PCR amplification and DNA sequencing. The results showed that there were no multiple peak observed in the sequence chromatograms (Fig. S2), suggesting that the sgRNA used in our study was target sequence specific.  . The EAG responses of wide type (n = 6), heterozygous (n = 11) and homozygous (n = 8) male adults to the three sex pheromone components. The 3-day old virgin males were tested against pheromone dose of 500 ng/filter. Error bars represent standard error (SEM). EAG values among the 3 types of males were compaired by one-way ANOVA followed with least-significant difference test (LSD). Different letters indicate significant differences at p < 0.05.

Discussions
As a newly developed genome editing technique, the CRISPR/Cas9 system has been widely used in studies of various organisms, but very limited in non-model insects especially the agriculturally important pests 47,48 . Our study successfully expanded this technique to a globe agricultural pest, H. armigera. By direct egg injection with the mixture of Cas9 and the sgRNA, the HarmPBP1 gene has been efficiently and specifically knocked out, allowing us to conduct the in vivo functional analyses of genes involved in the sex pheromone detection in lepidopteran species. It is noted that, to make results of the study more convincing, we intriguingly selected the target site at the far 5′ end of the ORF (50 bp-73 bp), and among the various mutations, we selected a mutation that resulted in a stop codon at the target site, leaving only 10 amino acids compared to 143 amino acids in the wild type HarmPBP1. Therefore, the homozygous mutant insects screened from this mutation is completely lost of HarmPBP1 function.
Multiple PBP genes have been reported in moth insects [49][50][51][52][53][54] , but their functional differentiation and relative importance in sex pheromone perception is mostly unknown, due to lack of an effective in vivo functional study technique. Some indirect or in vitro functional studies have been conducted to address this topic. For example, studies on the relative expression showed that, among 3 PBP genes, PBP1 displayed the highest antennal expression in males, and also the highest male bias in antennal expression 43,50,52,55 . Further binding assays using the recombinant proteins showed that PBP1 has higher binding affinity for sex pheromones than other PBPs in H. armigera 45 and other investigated moth species 50,52 . These results all suggests that PBP1 plays a role much more important than other PBPs in the olfaction of female sex pheromone, i.e. the majority part regarding the sex pheromone binding and transportation is contributed by PBP1. In the present study, however, the PBP1 knocked out male of H. armigera shows ~40% reduction in EAG responses to each of the three pheromone components. It seems that the relative importance of HarmPBP1 in the sex pheromone perception is somewhat overestimated in the in vitro binding assays, or that the relative importance could not be simply estimated on the percent reduction in EAG response. Other assays on behavioral responses and mating success might need to be incorporated into the evaluation.
There is no significant difference between sgRNA and RNase-free water injected eggs, in terms of survival and development (Table 2) at the G0 generation, however, we encountered a lower mating success and egg hatching rates (data not shown), in selection of the HarmPBP1 homozygous mutant strain by pairing the heterozygous moths. This is in accordance with the EAG assay, in which both homo-and heterozygous mutant moths displays a significant reduction in EAG values, compared to that of wild type moths. This reduction in mating success suggests that, sex pheromone take roles not only in female finding but also in mating with the female. Similar phenomenon is also found in a study on knocking out of an olfactory co-receptor gene (Orco) in Ostrinia furnacalis 56 . Therefore, the sex pheromone perception is physiologically related to or has crosstalk with the mating. The off target effect may also be a reason for the difficulty in maintaining the homozygous mutants. We did not observed any mutation in the top 4 potential sequences in the G0 chimaeras, but the potential off target sequences were identified using a H. armigera transcriptomic data, as the genome data is not available. It is possible that sequences of even higher off target potentiality might exist, and that the resulting off target induces the mating failure in the homozygous mutants.
In conclusion, our study provides the first in vivo evidence that HarmPBP1 plays important role in perception of female sex pheromones, using insects with HarmPBP1 knockout by CRISPR/Cas9 system. In addition, the study provides an important methodological reference for genome editing in other genes and in other lepidopteran pest species.

Materials and Methods
Insect and rearing. The wild strain of H. armigera was introduced from Prof. Wu's lab at Nanjing agricultural university, China. This strain originally started with insects from the Côte D'Ivoire (Ivory Coast), Africa over 30 years ago and has been maintained in the laboratory with no outcrossing 57 . We reared larvae on an artificial diet at 27 ± 1 °C with a 14: 10 (Light: Dark) photoperiod and 65 ± 5% relative humidity (RH). Pupae were sexed and placed individually in cages for eclosion. Adults were held under the same lab condition and supplied with a 10% honey solution as food.

Reagents. Restriction enzyme Msl1 was purchased from New England Biolabs (Beijing, China).
Oligonucleotides were custom synthesized by GENEWIZ Biotechnology (Suzhou, China). All three sex pheromones (Z11-16:Ald, Z9-16:Ald and Z9-14:Ald) were purchased from Sigma-Aldrich (http://www.sigmaaldrich. com/china-mainland) with more than 95% purity. Target site and potential off-target site searching. The target site of the CRISPR/Cas9 system was determined at exon 1 of HarmPBP1 sequence (GenBank accession no. HQ436362.1) according to the criteria of 5′-GG-(N)-18-NGG-3′ 11 . Based on the target site sequence, the DNA template for sgRNA synthesis was designed to contain the T7 promoter, the target site and the guideRNA (gRNA) sequences (Fig. 5). The software CasOT 58 was used to search potential off target sites of the CRISPR/Cas9 system, against a transcriptomic database of H. armigera (GenBank accession NO. GBDM01000816-GBDM01029784). The top 4 potential off target sequences determined were listed in Table S1. To check the possible off target mutations, 4 G0 moths were randomly selected and analyzed for these specific sequences.

Egg microinjection and mutation analysis.
Gauzes with freshly laid eggs were rinsed in diluted sodium hypochlorite solution for 3-5 minutes, then eggs were collected and washed with distilled water for several times. After the water was removed by a filter screen, the eggs were lined up and fixed on a microscope slide. About one nanoliter mix of Cas9-coding mRNA (300 ng/μL) and sgRNA (150 ng/μL) was injected into an egg using a FemtoJet and InjectMan NI 2 microinjection system (Eppendorf, Hamburg, Germany). A total of 496 eggs were injected in 2 hours. The pre-experiment of microinjection with RNase-free water showed no deleterious effect on growth and development for eggs. Injected eggs were incubated at 25 °C, 65 ± 5% RH for 3-4 days until hatching. Survival rates of larvae were calculated at the third instar.
About 20 injected eggs were collected to detect the mutation after 24 hours. The genomic DNA extracted (QIAamp DNA Mini Kit, Qiagen, Hilden, Germany) from these pooled eggs was used as template for PCR amplification. The fragment (767 bp) flanking the CRISPR target site was amplified using a pair of specific primers (forward: 5′-CTGACAGCCCAGCGATACC-3′, reverse: 5′-CGCCTTGCCGTGATGTAG-3′). The PCR products were purified using AxyPrep PCP Cleanup Kit (Axygen, Suzhou, China). A restriction enzyme cutting site (Msl1 for HarmPBP1) adjacent to the NGG PAM was selected to analyze the putative mutations by restriction enzyme digestion (RED) assay. After Msl1 digestion, the PCR products were separated by gel electrophoresis. The uncleaved band by Msl1 indicated the mutation occurred at the target site. The mutation frequency was calculated by dividing uncleaved band intensity to the total band intensity from a single digestion experiment 59 . Band intensity was measured using Quantity One Software (Bio-Rad, Hercules, CA, USA). To further determine the mutant sequences, the uncleaved band was recovered and sub-cloned. The positive clones were sequenced and aligned with the wild-type sequence. Sequencing chromatogram with multiple peaks around the target site also indicated the occurrence of mutations.

Screening of the HarmPBP1 homozygous mutant lines. The generation with the eggs injected
with Cas9/sgRNA was designated as G0. Apart from the 20 eggs used for checking the mutation efficiency, other G0 eggs were maintained properly until adults and were used to screen homozygous mutant lines, using the inbred and single pair mating strategy. The G0 moths were paired with each other to obtain G1 offspring, then G1 moths from selected G0 single pairs were paired with each other to obtain G2 offspring, and so forth. For each generation, after laying fertilized eggs, DNA typing of parent moths were determined by PCR amplification of the HarmPBP1 genome DNA followed with the RED assay. For RED assay-positive G0 and G1 moths, a further direct sequencing of the PCR product was also conducted to determine the mutant sequence. Among the various mutated HarmPBP1 sequences in G1 moths, a mutation with a premature stop codon that would lead to a highly truncated HarmPBP1 (of only 10 amino acids) was chosen to screen the homozygous mutants. The G3 homozygous and heterozygous mutant moths of this mutation type were used for in vivo functional analysis.
Electroantennogram (EAG) recordings. The solutions of 3 sex pheromone components (Z11-16:Ald, Z9-16:Ald, and Z9-14:Ald) were prepared in 50 ng/μL using hexane as solvent. EAG values were recorded by using our previous reported method 60,61 , and hexane solution was used as the control. Briefly, the antenna of unmated males (3 days after eclosion) was cut off at the base of the moth's head with a knife blade, and a short fragment in terminal of the antenna was excised to keep better contact with the electrode. The antenna with both ends removed was connected by gel (SPECTRA 360, Fairfield, NJ, USA) to the two recording electrodes, respectively. The filter paper strip (2.5 × 0.9 cm) containing 10 μL test solution was allowed to evaporate solvent for 5 minutes, then the paper strip was inserted into a Pasteur pipette placed perpendicularly through a hole in a metal line tube with an airflow of 4 ml/s. Stimulations were achieved by directing a puff of air (4 ml/s) through the pipette with a timer-controlled solenoid valve. For each antenna to each sex pheromone compound, the EAG response was tested in three repeats with an interval of 30 seconds. The EAG signals were recorded as voltage waveforms with EAG-adapted software (Syntech ® , The Netherlands).

Statistical analysis.
The Chi-square test was used for significance evaluation of survival rates between insects derived from HarmPBP1-sgRNA injected and RNase-free water injected eggs. The one-way ANOVA followed by least-significant difference test (LSD) was used to compare differences in EAG amplitude between HarmPBP1 mutant and wild type adult males. Data were analyzed using SPSS 20.0 software.