CRISPR-mediated knockout of cardinal and cinnabar eye pigmentation genes in the western tarnished plant bug

The western tarnished plant bug, Lygus hesperus, is a key hemipteran pest of numerous agricultural, horticultural, and industrial crops in the western United States and Mexico. A lack of genetic tools in L. hesperus hinders progress in functional genomics and in developing innovative pest control methods such as gene drive. Here, using RNA interference (RNAi) against cardinal (LhCd), cinnabar (LhCn), and white (LhW), we showed that knockdown of LhW was lethal to developing embryos, while knockdown of LhCd or LhCn produced bright red eye phenotypes, in contrast to wild-type brown eyes. We further used CRISPR/Cas9 (clustered regularly interspaced palindromic repeats/CRISPR-associated) genome editing to generate germline knockouts of both LhCd (Card) and LhCn (Cinn), producing separate strains of L. hesperus characterized by mutant eye phenotypes. Although the cardinal knockout strain Card exhibited a gradual darkening of the eyes to brown typical of the wild-type line later in nymphal development, we observed bright red eyes throughout all life stages in the cinnabar knockout strain Cinn, making it a viable marker for tracking gene editing in L. hesperus. These results provide evidence that CRISPR/Cas9 gene editing functions in L. hesperus and that eye pigmentation genes are useful for tracking the successful genetic manipulation of this insect.

The western tarnished plant bug, Lygus hesperus Knight (Hemiptera: Miridae) is a major pest of cotton and other crops throughout the western United States and other parts of North America [1][2][3] . Although an integrated pest management program has been implemented against L. hesperus in Arizona 3 , its success is dependent on the continued effectiveness of only a few insecticides which have been widely used for many years 4 . With the rapid evolution of insecticide resistance observed in the closely related Lygus lineolaris [5][6][7] , new tactics are needed to maintain control over members of this genus.
Among contemporary strategies for controlling arthropod pest species, homing-based gene drives are currently being developed for management of agricultural pests as well as those that vector human disease [8][9][10][11][12][13] . Clustered regularly interspaced short palindromic repeats and CRISPR-associated protein 9 (CRISPR/Cas9) gene drive systems have been developed and shown to effectively drive deleterious genes into laboratory insects, resulting in population crashes 11,13,14 . CRISPR-based gene drive systems developed in the laboratory include homing, split homing, translocation, X-shredder, killer-rescue, cleave-and-rescue, and TARE (reviewed in 15 ).
CRISPR/Cas9 gene editing is also widely used to ascertain gene function due to its efficiency and specificity in inducing mutations by cleavage and impairment of the genomic target sequences in model and non-model organisms 16,17 . In insects, genes that control eye pigmentation are frequently targeted because many of the induced mutations produce striking visible changes that facilitate screening for knockout efficacy. A frequent target is White, an ABC transporter that functions in transporting pigment precursor into pigment granules 18 . For example, CRISPR editing of white produced white-eyed adults in both Helicoverpa armigera 19 and Bactrocera dorsalis 20 , although homozygous mutations were lethal in the former and unexpectedly resulted in the loss of black head spots in the latter. In hemipterans, CRISPR-mediated null mutation of white results in white eyes in nymphs and bright red eyes in adults of Bemisia tabaci 21 and lighter red eyes and white ocelli in Nilaparvata lugens 22 .
Insect eye pigments are primarily from the guanine-derived pteridines and/or the tryptophan-derived ommochromes. Genes involved in the ommochrome pathway of pigment transport and formation are well studied in several model species, including Drosophila melanogaster, Bombyx mori and Tribolium castaneum [23][24][25] and typically follow the schematic shown in Fig. 1. Null mutations in the enzymes involved in the step-by-step process CRISPR/Cas9-mediated knockout of LhCd or LhCn. Given the phenotypes generated in the embryonic RNAi experiments, we next sought to assess the viability of embryonic CRISPR/Cas9 knockouts singly targeting LhCd and LhCn. For both transcripts, the sequences corresponding to the 20-nucleotide guide and the protospacer adjacent motif were searched via BLASTn against L. hesperus publicly available data (NCBI organism limit-L. hesperus; taxid 30085). No matches were found that would suggest potential off-target effects (Supplementary Table S1).
The CRISPR/Cas9 knockouts were conducted via injections in two independent experiments. Overall, we achieved a hatch rate between 8.8 and 27.5% for injected individuals compared to 86.3-100% for the noninjected control group (Supplementary Table S2). The survival of the instar nymphs to adults ranged from 22.7 to 71.4%   Table S2). Of these RNP-injected adults, 40-100% showed mutant eye phenotypes (Supplementary Table S2). To generate L. hesperus strains with the mutant eye phenotypes, we crossed surviving G 0 adults from both the LhCd-and LhCn-injected lines from experiment 1 according to the scheme depicted in Fig. 3. We estimate that the efficiency of CRISPR/Cas9 gene-editing of the germline was 86% (± 9.6%), as determined by averaging the efficiency of all samples from the Card and Cinn strains (Supplementary Methods and Table S3). Individuals from the Card strain showed a gradual darkening of the eye beginning at the 3rd instar that continued with subsequent molts and during adult maturation (Fig. 5a). Eye coloration in fully mature adult Card females closely resembled those from wild-type, while Card males displayed red eye phenotypes that were brighter than those found in the wild-type strain (Fig. 5b). In contrast, the eye phenotype across all stages of Cinn strain development was characterized by complete bright red pigmentation regardless of sex and stage (Fig. 5).
Target site mutations in LhCd and LhCn mutant strains. Both gDNA and cDNA corresponding to the gRNA target sites in G 0 -G 3 individuals from the Card and Cinn strains showed mutations (Table 1). For LhCd, we found a total of 43 mutations with 20 corresponding to LhCd1 sgRNA1 and 23 for LhCd2 sgRNA2 ( Table 1, Fig. 4). For LhCn, we found 12 and 13 mutations corresponding to LhCn1 sgRNA1 and LhCn2 sgRNA2, respectively ( Table 1, Fig. 4). Of these mutations, 25 unique combinations were found for LhCd and 14 for LhCn. Names of the mutations are based on the unique combination of mutations that were found within each individual (Table 1). With the only exception of mutation Cd1.7, in which LhCd showed a wild-type allele at LhCd1 target site, all targeted sites displayed more than one mutation. Four LhCd mutations (Cd1.1-Cd1.4) and two LhCn mutations (Cn1.1 and Cn1.2) were found to occur in gDNA and/or cDNA across generations (Table 1). Cd1.1 was the most common LhCd mutation, as it was found in a single G 0 individual and in two G 3 individuals from the Card strain (Table 1, Fig. 4). For the Cinn strain, two individuals, including one G 0 and one G 2, both harbored the Cn1.1 mutation ( Table 1, Fig. 4). There were 10 mutations in Card and 5 mutations in Cinn that were in-frame, whereas 14 and 9 mutations in Card and Cinn, respectively generated premature stop codons. Overall, the combination of mutations at the sgRNA1 and 2 target sites resulted in mutant eye color phenotypes regardless of the in-frame effect.

Discussion
Successful use of CRISPR/Cas9 in L. hesperus provides a first step towards the use of contemporary molecular control strategies against this pest species, as well as in other related species. The results here show that CRISPR editing of two L. hesperus genes (LhCd and LhCn) resulted in heritable mutations that affected eye pigmentation across nymphal and adult development. RNAi knockdown of L. hesperus genes involved in eye pigment transport in late 5th instars produced adults exhibiting primarily wild-type eyes apart from a red line extending from the rostrum to the antenna along the medial margins 28 . The extent of the change varied by individual and was often difficult to distinguish without close examination. It is evident that eye pigmentation in L. hesperus is a continuous process throughout the entire course of development (Fig. 5), which can only be marginally impacted by transient knockdown from RNAi. This attribute led us to attempt CRISPR/Cas9 knockout of the same genes previously targeted to produce a more pronounced and lasting change in phenotype. Our strongest result was with LhCn; CRISPR/Cas9-mediated knockout of this gene yielded persistent bright red eyes that were strikingly different from the typical brown coloration in wild-type (Fig. 5). Sequencing of the sgRNA target sites showed multiallelic mutations that confirmed LhCn was indeed knocked out. Although the final eye color appears to be species dependent, the bright red eye phenotype in the Cinn strain is consistent with Cinnabar functioning in ommochrome transport/biosynthesis, as has been found in knockdown or knockout studies in N. lugens 22 , Nasonia vitripennis 33,34 , Aedes aegypti 35 , T. castaneum 36 , and B. mori 37 . It is, however, possible that LhCn has other biological roles in L. hesperus. The homologous Cn gene in D. melanogaster modulates post-translational regulation of the mitochondrial fission gene Drp1, such that disruption of Cn activity negatively impacts mitochondrial morphology and function 38 . However, we have not observed any changes in L. hesperus beyond eye color in this or our previous study of LhCn 28 .
Knocking down cardinal, another gene in the ommochrome pathway, also impacts eye color in L. hesperus. Manipulation of LhCd expression in late 5th instars had a moderate impact on eye color 28 , similar to that observed for N. lugens in which RNAi yielded eyes with a mixture of red and brown pigment 29 . Unlike the LhCn results, LhCd knockout mutant eyes appeared to gradually accumulate brown pigment with each successive developmental stage after the 3rd instar (Fig. 5). Although adult eyes in the LhCd mutants were redder than that of wild-type individuals, they were substantially darker than LhCn mutant eyes. Gradual changes to the eye color have also been found in adults with crimson mutations in Culex pipiens 39 and cardinal mutants in Plutella xylostella 27 . Cardinal functions as the final step in ommochrome biosynthesis to produce xanthommatin. In the absence of a functional cardinal, the oxidative 3-hydroxy kynurenine can auto-dimerize to xanthommatin over time 33,40,41 . This may explain the gradual darkening in eye pigmentation that was reminiscent of the wild-type phenotype that we observed in L. hesperus.
The white gene functions in both pteridine and ommochrome transport 18 , where White dimerizes with either Brown or Scarlet to transport pigments into pigment granules. It is frequently used as a marker of insect genetic manipulation 20,30,[42][43][44] . Although knockdown of LhW was frequently fatal in dsRNA-injected 5th instar L. hesperus nymphs, it did disrupt pigment accumulation in survivors 28 . Here, knockdown of LhW in embryos completely disrupted embryonic development prior to visible eye formation, we were thus unable to confirm if LhW is involved in L. hesperus eye pigmentation. Similar mortality induced by white knockout was observed in H. armigera 19 and Oncopeltus fasciatus 30  www.nature.com/scientificreports/ . Both mutant G 0 females were crossed with a mutant male (Card × 1 and × 2) and the remaining mutant males were crossed with four wild-type females (Card × 3-5). G 1 offspring of Card × 1 and × 2 were inbred to generate G 2 . G 1 heterozygous offspring from Card × 3-5 were inbred to generate a mix of wild-type and red-eye G 2 progeny. The selected G 2 (boxed) with red eyes from all Card crosses were pooled to establish the Card colony. Cross numbers are indicated by an "x" and the respective number. (b) Two G 0 female from Cinn were crossed with wild-type males generating heterozygous G 1 progeny that were subsequently inbred to produce G 2 . Red-eye mutants (boxed) in G 2 were crossed and perpetuated to establish the Cinn colony. L. hesperus models with black or red eyes represent wild type or mutant, respectively, in (a) and (b). www.nature.com/scientificreports/ produces retinal degeneration, shortened life span, and progressive loss of ability to climb 45 . The multitude of biological processes affected by white indicate that its substrate specificity is not limited to pigment precursors. White has been shown to be expressed in the nervous and excretory system of Drosophila and functions in the transport of important substrates like biogenic amines 46 and cyclic guanosine monophosphate 47 , respectively. The exact functions of LhW remain to be elucidated, but are clearly crucial for normal development.

Scientific Reports
As is evident from the changes to L. hesperus eye color, our methods produced efficient and on-target gene editing that was inherited among increasing proportions of progeny for each new generation. Both Card and Cinn colonies have been reared beyond 10 generations with no changes in the mutant eye phenotypes, indicating that respective mutant lines can be stably generated under established laboratory rearing conditions without any observable negative impacts on fitness. Functional genomic studies can now be conducted on key genes, such as those regulating development, reproduction, and insecticide resistance. Use of relatively new molecular approaches, such as Receptor Mediated Ovarian Transduction of Cargo (ReMOT Control) and/or branched amphiphilic peptide capsules (BAPC), can also be applied to facilitate and accelerate functional genomic studies in L. hesperus. ReMOT Control uses an ovary specific-ligand to target the RNP cargo to developing oocytes by injecting the abdomens of mature adult females 21,34,[48][49][50][51] . This technology has potential for generating mutant insects at a higher rate and reduced cost. In contrast, BAPC assisted-delivery of CRISPR/Cas9 into the developing oocyte by injecting the BAPC-RNP mixture near the ovaries may enable enhanced uptake and improve efficiency of editing 48 . The immediate next step in optimizing genetic engineering of L. hesperus is to use CRISPR/Cas9 to target genes that reduce fitness, alter sex differentiation, or induce mortality. The generation of conditional gene-drive systems in other model insects exemplifies the feasibility of driving traits with associated negative phenotypes into populations 8,11,52 . Future work in L. hesperus will continue to focus on expanding our understanding of basic functional genomics as it relates to pest biology, but also begin to develop new transgenic and/ or gene-drive practical approaches for potential pest management purposes.

Lygus hesperus rearing.
A laboratory colony of L. hesperus, collected in Maricopa, AZ, served as the source of insects. Adults were maintained in 0.03 m 3 screened plastic cages containing shredded paper that were housed in an environmental chamber set at 27 ± 1 °C, 40-60% RH, and a 14:10 (L:D) h photoperiod. Fresh green beans (Phaseolus vulgaris L.), an artificial diet pack 53 , and a bottle of water with a wick were used to nourish and hydrate the colony and were replaced as needed.
Embryonic RNAi. dsRNA targeting LhCd (MH806847), LhCn (MH806848), and LhW (MH806842) was produced as described in Brent and Hull (2019) to a concentration of 1 µg/µl. In brief, ~ 500-bp products were PCR amplified from validated plasmid DNAs harboring LhCd, LhCn, or LhW using T7 promoter containing primers (Supplementary Table S4). PCR products were purified then used as templates for in vitro transcription using a MEGAscript RNAi kit (Thermo Fisher Scientific). dsRNAs corresponding to the fluorescent protein gene venus, which was injected as a negative control, were similarly generated.
Gel packs, made of Parafilm M (Pechiney Plastic Packaging, Chicago, IL) and filled with carrageenan (1.25% w/v), were provided to L. hesperus as an oviposition substrate for one hour. The gel was removed from the packs and the egg-embedded parafilm sheet retained. The parafilm was stretched to release the eggs, which were then transferred to a moistened filter paper using a wet, fine-tip paintbrush. Eggs were aligned in 4 rows of 10 for a total of 40 eggs per treatment per experiment (LhCd, LhCn, LhW, and venus) and then gently covered with a No.1, 24 × 40 mm coverslip coated with permanent linerless double-sided Scotch tape (3 M, Maplewood, MN). The bare side of the coverslip was mounted on a glass slide with double-sided tape that adhered to the corners www.nature.com/scientificreports/ of the coverslip. The slide was placed on a Leica DMIL scope (Allendale, New Jersey). An IM 300 Microinjector (Narishige International USA, Amityville, NY) with a quartz needle loaded with dsRNA was used to inject the embryos at the posterior pole. Needles were produced by pulling capillary tubes with filament using a P-2000 needle puller (Sutter Instrument, Novato, CA) with the following two-line program: Line 1) Heat = 850, Filament = 5, Velocity = 25, Delay = 128; and line 2) Heat = 700, Filament = 5, Velocity = 50, Delay = 150. Needles were beveled using a Model EG-44 micropipette grinder (Narishige) at a 30° angle and an approximate rotor speed of 1800 rpm or 90% of the maximum speed. Needles were backfilled using a Microloader tip (Eppendorf, Enfield, CT). Following injection, coverslips with eggs were placed in a covered plastic petri dish containing 1% agarose (1 g agarose in 100 ml of distilled water), which was sealed with parafilm and placed in a growth chamber with the same settings as the laboratory colony. Images of eggs were taken 5 days post-injection using a Nikon SMZ18 microscope equipped with a Nikon D5-Ri2 camera (Nikon Instruments Inc., Melville, NY).
To confirm knockdown of targeted transcripts, expression of LhCd, LhCn, and LhW was measured by semiquantitative RT-PCR, using actin (GDHC01004191) as a loading control. Total RNA was isolated from four replicated groups of three eggs using a Quick-RNA Microprep kit (Zymo Research, Irvine, CA). RNA quality and quantity were assessed using the Take3  Design and synthesis of sgRNAs. sgRNAs were designed using LhCd and LhCn with a focus on identifying guide sites near the 5′-end of the gene using CRISPOR 54 . Both sets of gene-specific sgRNAs were designed in proximity to one another; LhCd1 and 2 are separated by 8 nucleotides while LhCn1 and 2 are separated by 5 nucleotides. sgRNAs were screened for potential off-target sites by BLASTn of the 20-bp target sequence and the PAM sequence against the L. hesperus taxid 30085 database. Potential off-target sites were determined by comparing the BLASTn hit sequences that exactly matched the 3′ end of each sgRNA and the PAM sequence.
Creation of CRISPR/Cas9 eye pigmentation mutant strains. The experimental design included two independent injection groups; the first injections used a Cas9 protein with a nuclear localization signal (PNA Bio, Newbury Park, CA), whereas the second set used the Alt-R Streptococcus pyogenes HiFi Cas9 nuclease V3 (Integrated DNA Technology, Coralville, Iowa). The injection mixture consisted of the RNP complex of Cas9 (300 ng/µl) with two sgRNAs each at 150 ng/µl or a total of 300 ng/µl. Each sgRNA was preincubated with Cas9 at room temperature for 15 min and both solutions of RNP were combined to make the injection mixture. Negative controls include "no injection" and water only. Embryos were prepared and injected as previously described for the RNAi experiments. A total of 80 eggs per treatment (LhCd1 + 2, LhCn1 + 2, non-inject) were injected in the first experiment and the second experiment included 160, 80, 80, and 20 eggs for LhCd1 + 2, LhCn1 + 2, non-inject, and water, respectively.
Six to nine days post-injection, 1st instar nymphs that hatched were collected into a 355 ml mesh lidded paper cup and reared to adulthood under rearing conditions identical to those outlined above. Pairs from each subsequent generation were mated using the crossing schemes shown in Fig. 3. Card cross 1 (× 1) between a mutant male and a mutant female produced 19 G 1 males and 15 G 1 females. Card × 2, which also crossed mutants of both sexes, generated 9 G 1 males and 12 G 1 females. All G 1 males from Card × 1 were group-mated with females from Card × 2 and vice versa to generate G 2 . To perpetuate the mutant line, G 2 with red eyes were selected and transferred to new cages. In addition, Card × 3-5 each consisted of one mutant male and four wild type females. The G 1 progeny with wild-type eye color from Card × 3-5 were combined and group-mated to generate a mix of wild-type and mutant eye color G 2 progeny. From the resulting G 2 progeny, 19 males with the mutant eye phenotype were crossed with 38 G 2 mutant females from the Card × 1 and × 2 lines. Mutant progeny arising from this group-mating, as well as individuals from the ongoing mutant lines of Card × 1 and × 2 formed the Card colony.
To generate the Cinn colony, two mutant G 0 females were crossed with two wild-type males to produce G 1 progeny with wild-type eyes. Of these, three females and four males from Cinn × 1 were backcrossed with three males and 13 females from Cinn × 2, respectively. G 2 progeny with the mutant eye phenotype were selected to perpetuate the Cinn colony. Mutant colonies were reared in 355 ml paper cups covered with mesh lids with up to 50 individuals. To prevent overcrowding in larger groups of > 50 individuals, 1.89 L paper cups were used 55 . Fresh green beans and sunflower seeds were provided twice a week. Diet and oviposition carrageenan packs were placed into rearing cups one week post-adult emergence. www.nature.com/scientificreports/ Sequencing LhCd and LhCn from mutant eye pigment strains. Representative insects from CRISPR strains that displayed altered eye pigmentation were collected at G 0 , G 1 , G 2 , and/or G 3 and stored at -80 °C in RNALater (Invitrogen, Carlsbad, CA). gDNA was extracted using a DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany). Total RNA was extracted using TRI Reagent following the manufacturer's protocol. Total RNA was treated with DNase I (Thermo Fisher Scientific) and cDNA was synthesized using a SuperScript IV First-Strand Synthesis kit (Invitrogen). LhCd and LhCn were PCR amplified from gDNA and cDNA using a Phusion High-Fidelity PCR kit (Thermo Fisher Scientific) with gene-specific primer pairs (Supplementary Table S4) and thermocycler conditions of 1 cycle at 98 °C for 30 s; 35 cycles at 98 °C for 5 s, 60 °C for 10 s, 72 °C for 5 s; and 1 cycle at 72 °C for 5 min. PCR products were cloned into pJET1.32/blunt vector (Thermo Fisher Scientific) and transformed into One Shot OmniMAX 2 T1 Chemically Competent E. coli (Thermo Fisher Scientific). Multiple clones (n = 4-10) from each transformation reaction were Sanger sequenced (Retrogen Inc., San Diego, CA).