Introduction

Horizontal gene transfer (HGT), a process in which the exchange of genes occurs between different species, is common among prokaryotes and recognized as one of the major forces in prokaryotic genome evolution (Ochman et al. 2000; Koonin et al. 2001; Pallen and Wren 2007; Polz et al. 2013). Compared with the frequency and level of HGT in prokaryotes, HGT events among eukaryotes are relatively rare (Boto 2014; Suzuki et al. 2015). An accumulation of evidence suggests that some horizontally transferred genes are truly expressed and have novel functions in eukaryotes (Jiggins and Hurst 2011; Acuna et al. 2012; Fan et al. 2020). Recently, reports of HGT have become increasingly frequent in insects and nematodes. HGT may play an important role in the adaptive evolution of these organisms (Acuna et al. 2012; Wybouw et al. 2016; Husnik and McCutcheon 2018).

Lepidoptera (butterflies and moths) is the second-largest order of insects. Increasingly, incidences of HGT from bacteria and fungi to Lepidoptera are being reported. The majority of HTGs to Lepidoptera have likely provided physiological functions related to nutritional metabolism and detoxification; these functions have in turn facilitated the adaptation of lepidopteran insects to specific plant hosts (Li et al. 2011; Sun et al. 2013). For example, chitinase genes acquired through HGT can degrade chitin during molting in lepidopteran insects (Daimon et al. 2005). In addition, Bombyx mori β-fructofuranosidase gained through horizontal transfer is related to tolerance to sugar-mimic alkaloids contained in mulberry latex (Daimon et al. 2008; Gan et al. 2018). Consistent with prokaryotes, transferred genes in eukaryotes, including plant-parasitic nematodes (Danchin et al. 2010), aphids (Nikoh et al. 2010; Novakova and Moran 2012), and lepidopterans (Li et al. 2011; Sun et al. 2013), are more frequently duplicated than endogenetic. For example, among 14 genes acquired by HGT in lepidopteran insects, more than half have undergone gene duplication (Sun et al. 2013). Paralogous genes can evolve in several ways following duplication events. The majority of duplicate gene copies become nonfunctionalized; however, both paralogs survive in some circumstances, presumably due to neofunctionalization or subfunctionalization (Lee and Irish 2011). Therefore, studying the subsequent fate of these duplicate gene pairs can contribute to understand the role of post-HGT gene duplication in the evolution of eukaryotes.

Cysteine synthase (CYS) is a crucial enzyme involved in sulfur amino acid biosynthesis and exhibit dual functions in plants, bacteria (Feldman-Salit et al. 2009; Lai et al. 2009; Bogicevic et al. 2012; Yi et al. 2012), and nematodes (Budde and Roth 2011). Cysteine synthase are responsible for cysteine synthase (CYS) activity that catalyzes the reaction between H2S and O-acetylserine (OAS) to generate cysteine, and β-cyanoalanine synthase (CAS) activity that catalyzes the conversion of hydrogen cyanide (HCN) into β-cyanoalanine. A recent study suggested that mites and lepidopteran insects possessed HGT-derived CYS genes and the recombinant mite CYS had dual functions (Wybouw et al. 2014). The recombinant CYS enzymes have CAS activity in eight lepidopteran species including pierids that feed on cyanogenic host plants and others species with varying diets; these enzymes are essential to lepidopterans being able to thrive on cyanogenic plants (Van Ohlen et al. 2016; Herfurth et al. 2017). However, previous studies were unable to clearly detect CYS activity in lepidopteran insects. Interestingly, CYS genes have also been identified in other lepidopteran insects that are not normally feed on cyanogenic plants, including B. mori and Manduca sexta (Li et al. 2011; Zhu et al. 2011; Sun et al. 2013; Wybouw et al. 2014). Moreover, the duplication of CYS genes has been observed in some lepidopteran species that feed on cyanogenic host plants such as Heliconius melpomene (Arias et al. 2016), Spodoptera litura, and Pieris rapae (Sun et al. 2013; Van Ohlen et al. 2016). Nevertheless, the diverse fates of CYS duplicates formed by post-HGT duplication remain to be elucidated.

In the present study, we examined the evolutionary fate of CYS genes by reconstructing their evolutionary history in lepidopteran insects. In addition, we characterized the functional divergence of duplicated CYS genes by examining the gene expression and enzymatic properties. Our results suggest that gene duplication and subsequent functional diversification of CYS genes could potentially facilitate the adaptation of lepidopteran insects to their plant hosts. Thus, our study provides valuable insights into the evolution and functional diversification of CYS genes in Lepidoptera; specifically, CYS genes acquired by HGT have apparent functional and ecological importance in lepidopteran insects.

Materials and methods

Insects material

The N4 strain of B. mori were reared on fresh mulberry leaves at 25 °C under standard conditions (14-h light/10-h dark cycle and 70% relative humidity). Larvae of Spodoptera frugiperda were reared on fresh corn in a conditioned insect rearing room. Bombyx mori and S. frugiperda samples collected from different developmental stages, were dissected in phosphate buffered saline (PBS). All samples were stored at −80 °C until used for analysis.

Phylogenetic analysis and conserved synteny analysis of CYS genes

To identify more potential CYS paralogs, we downloaded the Hidden Markov Model (HMM) file for the PLP-dependent enzymes superfamily (PALP) from the Pfam database (http://pfam.xfam.org/). Using HMM of PALP domain (PF00291), hmmsearch program from the HMMER package was performed (e-value <10e − 5) to identify the putative CYS paralogs on 4 representative lepidopteran protein dataset (S. frugiperda, B. mori, P. rapae, and Amyelois transitella) (Eddy 2011). CYS paralogs were then used as queries to perform BLASTp and tBLASTn searches in Lepbase (http://lepbase.org/) and National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/gene). Phylogenetic tree was generated based on the protein sequence alignments in Molecular Evolutionary Genetics Analysis (MEGA, version 7.0) using neighbor-joining (NJ) method under JTT model (Edgar 2004; Kumar et al. 2016) (Fig. S1). To further investigate the evolutionary relationships of HGT-derived CYS genes, homologs of CYS were identified by BLASTp and tBLASTn searches in Lepbase and NCBI with BmorCYS (XP_004932948.1), PxylCYS (XP_011554718.1), MspCYS (EIZ83823.1), and TurtCYS (XP_015786551.1) used as queries (Wybouw et al. 2014). The species names, protein names, and accession numbers were presented in Table S1. All protein sequences were first aligned in the MEGA7 program using MUSCLE with the default settings (Edgar 2004; Kumar et al. 2016). We then inspected for regions of high-quality alignment and manually checked for misalignments. The best amino acid substitution model selected by ProtTest was LG + G (Darriba et al. 2011). Phylogenetic trees were generated based on the protein sequence alignments in MEGA7 using Maximum Likelihood method (ML) under LG + G protein model and NJ method under JTT model. The final trees were visualized using the Interactive Tree of Life (iTOL) online tool (http://itol.embl.de).

The syntenic relationships of CYS genes and their flanking genes in lepidopteran and mite genomes were manually analyzed in NCBI genome database using PrapCYS1 as the query gene. We analyzed genes flanking CYS gene and manually checked the results by performing reciprocal BLAST searches.

PCR amplification to exclude the possibility of bacterial contamination

Genomic DNA was extracted using DNA Isolation Mini Kit (Vazyme, Nanjing, China). Primer pairs were designed to amplify a 9400 bp-long genomic fragment of BmorCYS and its downstream esterase FE4-like (FE4) gene as well as their intergenic region on the B. mori genome. Briefly, the 20 µl reaction volume comprised 10 µl PCR Super Mix (Vazyme, Nanjing, China), 0.8 µl of each 10 µmol/L primer, and 1 µl template DNA. Amplification was performed using the following conditions: pre-denaturation at 94 °C for 3 min, followed by 35 cycles each with 94 °C for 10 s, 68 °C for 10 min. The PCR run ended with a post-extension step at 68 °C for 7 min.

Expression analysis by quantitative real-time PCR (qRT-PCR)

To obtain more detailed expression data of CYS genes in B. mori and S. frugiperda, we analyzed various tissues from different development stages: the third day of the last larval instar (L5D3), the eighth day after pupation (P8), and the first day of moth (A0). RNA was isolated with TRIzol reagent (Takara, Dalian, China) and quantified using a Nanodrop spectrophotometer (Thermo Fisher, USA). cDNA was synthesized using Primer Script RT Reagent Kit (Takara, Dalian, China).

qRT-PCRs were performed with specific primers as described previously (Gao et al. 2020). The cycling parameters were used as follows: 95 °C for 1 min, followed by 35 cycles at 95 °C for 15 s and 60 °C for 30 s. We analyzed three independent biological samples and calculated relative CYS expression levels against B. mori ribosomal protein L49 gene (BmorRpl49) or S. frugiperda glyceraldehyde 3-phosphate dehydrogenase gene (SfruGAPDH) using the 2−ΔΔCT method. The data were presented as means ± standard error of the mean. Statistical analysis between groups were performed using the one-way analysis of variance (ANOVA) on SPSS 20.0 software. Different letters over the bars showed that the expression levels between groups are statistically significant (P ≤ 0.05).

Protein purification and determination of recombinant CYS proteins

BmorCYS, SfruCYS1, and SfruCYS2 were respectively cloned into the corresponding site of a pCold vector (Takara, Dalian, China). Primers were listed in Table S2. Escherichia coli carrying pCold-BmorCYS, pCold-SfruCYS1, and pCold-SfruCYS2 were cultured in Luria–Bertani medium at 37 °C. The recombinant proteins were induced by isopropyl-β-D-thiogalactoside until the absorbance at 600 nm reaches approximately 0.6. After incubation for 15 h at 18 °C, 220 rpm, the cells were harvested by centrifugation at 6000 × g, 4 °C for 5 min. The pellets were suspended in cold PBS, followed by sonication. The lysates were centrifuged at 10,000 × g for 30 min. After centrifugation, the supernatant was collected and purified as previously described (Zhou et al. 2019). Briefly, the supernatant was loaded onto a Ni-nitrilotriacetic acid (NTA) column. The column was washed with washing buffer consisting of 0.5 M NaCl, 20 mM sodium phosphate, and 20 mM imidazole, pH 7.4. Then, the CYS proteins were eluted with a stepwise gradient of imidazole (up to 500 mM) in washing buffer. The eluate was dialyzed against 20 mM sodium phosphate buffer (pH 7.4). Subsequently, recombinant BmorCYS, SfruCYS1, and SfruCYS2 were stored at 4 °C or −80 °C until use.

Preparation of antiserum and western blotting analysis

Purified proteins were used to generate rabbit polyclonal antibody according to previous study (Lee et al. 2016). Western blotting analysis was performed as previously described (Dai et al. 2019). Proteins were extracted from seven different tissues of L5D3 using RIPA lysis buffer (FDbio Science, China) and then separated by 10% SDS-PAGE before being transferred to polyvinylidene fluoride membranes (Millipore, USA). After blocking with 5% nonfat milk for 2 h, the membranes were incubated overnight with polyclonal rabbit anti-CYS immunoglobulin G. Subsequently, the membranes were incubated with a florescence-labeled secondary antibody (DingGuo ChangSheng Biotechnology Co., Ltd., China) for 2 h. The protein level of tubulin was used as a loading control. Western blotting signals were detected using an ECL Plus Kit (FDbio Science, China).

Identifying the subcellular localization of CYS proteins

A subcellular localization assay was performed as previously described (Zhou et al. 2019). In brief, BmorCYS, SfruCYS1, and SfruCYS2 ORFs were cloned into a pIZ/V5-mCherry vector using specific primers (Table S2). Positive plasmids were transfected into Sf9 cells using Lipo8000™ Transfection Reagent (Beyotime, China). After 24 h of incubation, the nuclei were stained with Hochest 33258. The cells were then washed with PBS and observed using a Zeiss LSM780 laser scanning confocal microscope.

Enzyme assays

Recombinant CYS proteins were incubated at 30 °C for 10 min in the reaction buffer containing 500 µM pyridoxal-5′-phosphate before measuring enzyme activity. The CYS activity assays were performed based on the method of Lunn (Lunn et al. 1990), with 1 µg of purified recombinant BmorCYS, SfruCYS1, or SfruCYS2 protein used per respective reaction. Standard substrate concentrations were 10 mM and 5 mM for OAS and sodium sulfide, respectively. The reaction product cysteine was quantified using the method of Gaitonde (Gaitonde 1967; Wybouw et al. 2014). The CAS activity assay were performed based on the method of Hendrickson (Hendrickson and Conn 1969). The standard reaction was performed with 500 µl of 12 mM cysteine, 500 µl of 12 mM potassium cyanide (KCN), and 1 µg recombinant CYS protein. These reactions were performed in a 1.5 mL microcentrifuge tubes with sealed caps. Cyanide solution was prepared by adding solid KCN immediately before use. After incubation at 30 °C for 10 min, CAS activity was quantified by measuring products H2S with adding 250 µl 30 mM FeCl3, and 250 µl 20 mM N, N-dimethyl-p-phenylendiamindihydrochloride. The absorbance was determined by spectrophotometrically at 650 nm.

To determine the optimum pH for CYS activity, 20 µl of reaction solution (containing purified CYS protein, 10 mM OAS, 5 mM sodium sulfide, and 20 mM Britton–Robinson’s wide range buffer of pH 3–13) was incubated at 30 °C for 15 min. The product cysteine was determined as described above (Lunn et al. 1990). Various concentrations of OAS, sodium sulfide, and KCN were used to determine the kinetic parameters of the recombinant CYS proteins. Substrate affinity constants were determined by one-site-specific binding with the Hill slope using GraphPad Prism7 software. Each kinetic parameter, such as Vmax and Km, for OAS, sodium sulfide and KCN were derived using the Michaelis–Menten equation. GraphPad Prism7 was used to conduct statistical analysis. All experiments were repeated three times.

Results

Duplication of CYS genes occurred independently after the divergence of many lepidopteran species

To identify more potential CYS paralogs, we performed HMM and BLAST searches. The phylogenetic tree was then constructed using MEGA7. The phylogenetic trees indicated that CYS paralogs were clustered into three distinct, highly supported groups: Group I Cystathionine β-synthase, Group II Cysteine synthase, and Group III L-threonine ammonia lyase (Fig. S1). These are all members of pyridoxal-5′-phosphate (PLP) dependent enzyme family. 16 lepidopteran CYS homologs found within Group II tended to cluster with mite and bacterial sequences, suggesting that these CYS homologs within Group II might have been acquired from bacteria via horizontal gene transfer.

To further identify HGT-derived CYS homologs, we performed BLASTp and tBLASTn searches in the Lepbase and NCBI databases. In total, 57 genes encoding putative CYS were identified in lepidopteran, mite, and bacterial genomes. We found multiple putative CYS copies in several lepidopteran genomes from which 42 lepidopteran CYS proteins contained typical CBS-like domain and one lepidopteran CYS contained CysK domain, suggesting that they were members of the cysteine synthase family (Fig. S2). In addition, with the exception of M. sexta, all CYS genes lacked introns among 26 lepidopteran species (Fig. S2). Moreover, lepidopteran CYS genes had similar exon–intron structures to previously characterized members of the CYS gene family in mites and bacteria. To exclude the possibility that lepidopteran CYS genes were derived from contaminating bacterial sequences, we examined its position in the B. mori genome. The sequencing results showed that CYS genes were located on B. mori genome at downstream of FE4 gene, suggesting the CYS genes were not bacterial contamination (Figs. S3 and S4).

To further investigate the evolutionary relationships of HGT-derived CYS, we conducted an amino acid-based alignment of representative lepidopteran, mite, and bacterial CYS homologs and then constructed phylogenetic trees using the ML and NJ methods. ML and NJ methods yielded similar topologies and the phylogenetic tree reconstructed by ML was shown in Fig. 1. Phylogenetic analysis showed that arthropod CYS sequences form a highly supported cluster (bootstrap = 100/99) with a group of bacteria as their closest relatives; this suggests that the ancestors of the mite and lepidopteran species included in the phylogenetic tree acquired a CYS gene from bacteria via HGT. Despite some internal nodes being weakly supported within the lepidopteran CYS clade, most nodes of duplication events were supported with high confidence in our analysis (>80%; Fig. 1); thus, multiple homologous copies of CYS apparently exist in the studied lepidopteran species. In addition, the CYS homologs of several lepidopterans were clustered together by species, suggesting that duplication of CYS genes occurred independently within lepidopteran species. For instance, CYS1 and CYS2 paralogs of Heliothis virescens and A. transitella were reciprocally monophyletic. Moreover, CYS1 and CYS2 of Spodoptera clustered together within the species, suggesting that these proteins were derived from duplication before Spodoptera speciation. In D. plexippus and Papilio polytes, CYS1 and CYS2 paralogs were reciprocally monophyletic with high confidence. CYS paralogs of Pierid species were clustered into three groups, showing that gene duplication may have occurred in a common ancestor of these species. We also found that CYS duplication occurred in some polyphagous insects (such as P. rapae, Spodoptera littoralis, and S. frugiperda) that feed on cyanogenic plants (Fig. 2). Taken together, these results indicate that HGT-derived CYS genes are widespread and have undergone further duplications in many lepidopteran insects.

Fig. 1: Phylogenetic analysis of CYSs.
figure 1

The CYS sequences from bacteria, mites, and Lepidoptera are marked with different colors. Bootstrap values over 60% are indicated above the nodes, with the number on the right indicating NJ and that on the left indicating ML. Moreover, the bootstrap value at the branching point of arthropod and bacterial CYS homologs is 100/99 as indicated by the black arrow.

Full size image
Fig. 2: Identification of CYS copy number in mites and Lepidoptera.
figure 2

The presence (black box) or absence (white box) of each CYS paralogs is detected by BLASTp and tBLASTn. This analysis indicated that there are multiple copies of CYS in many lepidopteran insects.

Full size image

Genomic organization and synteny of CYS genes

To understand the evolutionary history of the CYS gene family, we investigated CYS genes and their adjacent genes in the genome of P. rapae, which had three CYS paralogs in distinct genomic loci (Fig. 3B). PrapCYS1 was located downstream of the FE4 gene, whereas PrapCYS2 and PrapCYS3 were found in a distinct syntenic location.

Fig. 3: Synteny of genes flanking CYS in mite and 11 lepidopteran chromosomes.
figure 3

Synteny was analyzed using previously published genome sequences from the NCBI database in (A) 10 lepidopteran insects, (B) P. rapae, and (C) mite. Genes and their positions are not drawn to scale and shown as the series of colored arrows. Homologs are represented by the same color, and genes with no homolog were indicated by white arrows. Arrows point in the sense direction of the gene. Insect CYSs are placed in the middle and colored in red.

Full size image

We also extended our syntenic analyses to other lepidopteran CYS genes. In the examined species, CYS2 genes were in the same syntenic location and shared the FE4 gene as their 5′ neighbor gene (Fig. 3A). In addition, we also identified two CYS paralogs located tandemly on the same scaffold in several lepidopteran insects. The presence of many lepidopteran CYS genes in a shared syntenic location across species indicates that they were likely orthologous. Finally, we examined CYS genes in mites; they were not adjacent to FE4, histidine-rich glycoprotein-like (HRG), or protein extra-macrochaetae genes (Emc) (Fig. 3C), suggesting that the CYS locus of mites was not syntenic with that of lepidopteran insects.

Functionally important amino acid residues in cysteine synthase are conserved in lepidopteran orthologs

Enzymes that catalyze the formation of cysteine are highly conserved in plants and bacteria; they have a conserved lysine residue and two domains: a pyridoxal-5′-phosphate (PLP)-binding domain and an OAS-binding domain (Burkhard et al. 1998; Bonner et al. 2005). To evaluate the nature of lepidopteran CYS orthologs, we identified functionally important amino acid residues in lepidopteran CYS proteins based on previous studies (Bonner et al. 2005; Wybouw et al. 2014). Such residues, which were responsible for the molecular interaction between the PLP cofactor (H182, G206, T207, T210, T296) and substrate OAS (T91, S92, G93, N94, Q95, Q172), were conserved among lepidopteran CYS proteins; the exception was A/S/G208, which had different residues in Eumeta japonica, Plutella xylostella, Pieridae, and Spodoptera species (Figs. S5 and S6). Of note, the functionally important amino acids of lepidopteran CYS proteins were conserved.

Expression pattern and subcellular localizations of CYS proteins in B. mori and S. frugiperda

The spatial and temporal expression data for a corresponding gene should be useful for determining its biological or physiological role (Doxey et al. 2007). Bombyx mori was the first lepidopteran insect for which a genome was fully sequenced and has been extensively studied (Goldsmith et al. 2005). To explore the expression profiles of CYS, we investigated the expression of BmorCYS from SilkDB 3.0 (http://silkworm.swu.edu.cn/silkdb/) in various tissues and at different development stages (Fig. 4A). BmorCYS was expressed at relatively high levels in larval Malpighian tubule and testis. Transcriptional levels of BmorCYS were highest in the Malpighian tubule during feeding stages, whereas levels dropped dramatically during molting, wandering, and prepupal stages. In contrast, BmorCYS was expressed in the testis at all larval stages. To verify our findings and obtain more detailed expression data, we dissected larva, pupae, and adult tissues from B. mori and investigated the expression of BmorCYS (Fig. 4B). In general, qRT-PCR analysis suggested that BmorCYS was expressed at relatively high levels in the larval Malpighian tubule. Moreover, BmorCYS was expressed in pupal testis, wing, and Malpighian tubule, as well as adult testis, Malpighian tubule, and legs. These results suggest that BmorCYS shows diverse and dynamic expression patterns in multiple tissues.

Fig. 4: Expression profiles of CYS in Bombyx mori.
figure 4

A Heat map for expression patterns of CYS gene in different stages and tissues of B. mori. B qRT-PCR assays of BmorCYS. C Protein expression analysis of BmorCYS in different tissues of L5D3 larvae. D Protein expression analysis of BmorCYS in Malpighian tubule. E Protein expression analysis of BmorCYS in testis. L4D3 day 3 of fourth-instar larvae, L4 molting fourth larval molting, L5D0 newly emerged fifth-instar larvae, L5D3 day 3 of fifth-instar larvae, PP pre-pupa, Ep epidermis, Mt Malpighian tubule, mFb male fat body, fFb female fat body, Ov ovary, Te testis, Mg midgut, Sg silk gland, Asg anterior silk gland, Msg middle silk gland, Psg posterior silk gland, Wi wing disc, Hd head, Tr Trachea.

Full size image

We also investigated the expression of BmorCYS protein on the third day of fifth-instar larvae (Fig. 4C). Western blotting analysis revealed that BmorCYS was expressed in the Malpighian tubule and testis but not in the fat body, midgut, and silk gland, which was consistent with the mRNA expression level of BmorCYS. Moreover, BmorCYS protein was detected in the Malpighian tubule during molting, the newly emerged fifth-instar, day-2 and day-3 fifth-instar larval, and wandering stages (Fig. 4D). BmorCYS was expressed in the testis at all larval stages examined (Fig. 4E). Thus, dynamic expression analyses revealed that BmorCYS was stably expressed in the testis but variable in the Malpighian tubule.

Duplicated gene copies often differ in their expression patterns (Farre and Alba 2010). To test whether such differences exist for CYS paralogous genes, we extended our expression analyses to another lepidopteran species, S. frugiperda, containing duplicate CYS genes and compared gene expression patterns of duplicate gene pairs. We found that two SfruCYS genes were significantly differentially expressed in larval tissues (Fig. 5). SfruCYS1 was highly expressed in the Malpighian tubule, fat body, and testis. SfruCYS2 was highly expressed in the larval fat body and midgut, whereas low expression of SfruCYS2 was observed in the Malpighian tubule, suggesting that the expression pattern of CYS genes had diverged in S. frugiperda. SfruCYS1 and SfruCYS2 were similarly expressed in the adult stage: expression levels were high in the wing and testis. Moreover, the expression of SfruCYS1 was similar to that of BmorCYS, which had high expression levels in the Malpighian tubule during the feeding stage.

Fig. 5: Expression profiles of CYS in Spodoptera frugiperda.
figure 5

qRT-PCR assays of SfruCYS1 and SfruCYS2. Ep epidermis, Te testis, Ov ovary, Mt Malpighian tubule, Sg silk gland, Fb fat body, Wi wing disc, Mg midgut, Hd head.

Full size image

To identify the subcellular locations of lepidopteran CYS proteins, pIZ-mCherry-BmorCYS, pIZ-mCherry-SfruCYS1, and pIZ-mCherry-SfruCYS2 were transfected into Sf9 cells, which encode transiently expressed mCherry-BmorCYS, mCherry-SfruCYS1, and mCherry-SfruCYS2 fusion proteins, respectively (Fig. S7). Strong signals from the three proteins were detected in the cytoplasm of Sf9 cells, indicating that B. mori and S. frugiperda CYS proteins had similar subcellular localizations in the cytoplasm.

Biochemical divergence of lepidopteran CYS proteins in CYS and CAS activity

In order to identify whether lepidopteran CYS have CYS activity and CAS activity, we expressed and purified recombinant BmorCYS, SfruCYS1, and SfruCYS2 proteins (Fig. S8). Subsequent biochemical assays suggested that three purified recombinant proteins had CAS and CYS activity (Figs. 6 and 7). The CYS/CAS ratio of the specificity constants was 2.75, 0.3, and 0.03 for BmorCYS, SfruCYS1, and SfruCYS2, respectively (Tables S3 and S4). This suggested that CYS activity of BmorCYS was higher than the CAS activity, whereas CAS activity was higher for SfruCYS2. The optimum pH of BmorCYS, SfruCYS1, and SfruCYS2 for cysteine synthesis activity was 9, 10, and 9, respectively. Moreover, the pH curve was broad for all CYS enzymes, with ~50% of the maximum activity occurring over pH 7–11 (Fig. 7B). This finding suggests that lepidopteran CYS proteins might have adapted to a highly alkaline pH environment.

Fig. 6: The CAS activity analysis of BmorCYS/SfruCYS1/SfruCYS2.
figure 6

Michaelis–Menten analysis of recombinant BmorCYS/SfruCYS1/SfruCYS2. Various concentrations of KCN was used to determine the kinetic parameters of the recombinant CYSs.

Full size image
Fig. 7: The CYS activity analysis of BmorCYS/SfruCYS1/SfruCYS2.
figure 7

A Michaelis–Menten analysis of recombinant BmorCYS/SfruCYS1/SfruCYS2. Various concentrations of OAS and sodium sulfide were used to determine the kinetic parameters of the recombinant CYSs. B pH profile of recombinant BmorCYS/SfruCYS1/SfruCYS2. The pH–activity relationship was determined using 10 mM OAS and 5 mM sodium sulfide as the substrate over a pH range of 3–13. Recombinant CYSs (1 µg) were incubated for 15 min at 30 °C. Cysteine was quantified by measuring the absorbance at 560 nm. These experiments were repeated three independent times.

Full size image

To determine whether lepidopteran CYS enzymes are biochemically diverged, we compared previously published data (Van Ohlen et al. 2016; Herfurth et al. 2017) and the enzyme activity of BmorCYS and SfruCYSs. For CAS activity, Km and Vmax values of CYS3 for KCN were one to two orders of magnitude higher than CYS2 and CYS1 among homologs from the same species in Pieridae. Km and Vmax values of CYS2 for KCN were one order of magnitude higher than CYS1 among homologs from the same species in Noctuidae (Fig. S9). In addition, the CAS activity of BmorCYS was low. For CYS activity, we found no significant difference in Km values between SfruCYS1 and SfruCYS2. However, SfruCYS2 had a higher Vmax than SfruCYS1 (Table S3). Taken together, our results suggested lepidopteran CYS enzymes have undergone divergence in CAS and CYS activities.

Discussion

In this study, we provided a comprehensive analysis of the evolution of the CYS family in Lepidoptera. We showed that CYS genes underwent further duplications in many lepidopteran insects, and duplicated CYS genes showed marked divergence in gene expression patterns and enzymatic properties. Lepidopteran CYSs not only have β-cyanoalanine synthase activity but also possess cysteine synthase activity; these processes may facilitate the adaptation of lepidopteran insects to various diets and diverse habitats. Thus, our findings provide valuable insights into the function and fate of lepidopteran CYS genes.

CYS genes may have horizontally transferred to common ancestor of Lepidoptera (Wybouw et al. 2014). Interestingly, CYS gene is widespread and has been maintained as a functional gene in several lepidopteran insects. Recombinant CYS enzymes apparently have β-cyanoalanine synthase activity in several lepidopteran insects, which is essential to thrive on cyanogenic plants (Stauber et al. 2012; Van Ohlen et al. 2016). In addition, CYS genes were observed in other lepidopteran insects that do not feed on cyanogenic host plants. It is unclear why these CYS genes have been retained in these organisms. We showed that lepidopteran CYS proteins not only have β-cyanoalanine synthase activities but also possess cysteine synthase activities. Both of these activities have also been reported in CYS proteins of plants, bacteria, and mites (Yamaguchi et al. 2000; Wada et al. 2004; Bogicevic et al. 2012; Wybouw et al. 2014). Moreover, CYS activity of BmorCYS was higher than the CAS activity, suggesting that lepidopteran CYSs are not only related to detoxification but also possess cysteine synthase activity that is involved in sulfur amino acid biosynthesis. Cysteine is the least abundant amino acid in plant leaves, which may be growth limiting for herbivorous insects (Barbehenn et al. 2013a; Barbehenn et al. 2013b). CYS activity may increase the levels of cysteine to promote insect’s growth. Moreover, the synthesis of glutathione (GSH) is tightly regulated and highly dependent on the availability of cysteine (Aoyama et al. 2008). A large fraction (20%) of total cysteine content is used to produce GSH (Jeschke et al. 2016), which may exert an anti-oxidant effect as a metabolic precursor of GSH and play roles in detoxification of specific plant secondary metabolites to facilitate organisms adaptation to host plants (Wadleigh and Yu 1988; Schramm et al. 2012). These might be one of the reasons why these CYS genes acquired from bacteria have been retained in these organisms that are not feeding on cyanogenic host plants. In addition, previous studies suggested that insects either obtain cysteine directly from their diet or synthesize it from methionine (Jeschke et al. 2016). Our results provide evidence for an alternative cysteine biosynthesis route in Lepidoptera: they can synthesize cysteine independently from methionine by CYS activity (Fig. 8).

Fig. 8: Summary of the synthesis and metabolism of cysteine in lepidopteran larvae.
figure 8

Insects can obtain cysteine directly from their diet or via synthesis from methionine. Our results provide an alternative cysteine biosynthesis route in Lepidoptera: synthesis of cysteine from OAS and sodium sulfide by CYS activity (marked by a red box). Other reactions involving the lepidopteran CYS gene are marked by a black box. CAS detoxifies cyanide by incorporating it into cysteine to form β-cyanoalanine, which is further metabolized into asparagine and aspartate. In addition, cysteine can be used to produce GSH, which plays a role in detoxifying specific plant secondary metabolites.

Full size image

CYS gene has undergone further duplications in many lepidopteran insects. 2-3 CYS paralogs were found in the butterflies D. plexippus, H. melpomene, and P. rapae (Sun et al. 2013). However, the number of CYS paralogs has yet to be identified in other lepidopteran genomes. Of 27 lepidopteran species examined, we found that 14 species have 2-3 copies of CYS genes. Interestingly, we found that CYS duplication occurred in some polyphagous insects (such as P. rapae, S. littoralis, and S. frugiperda) that feed on cyanogenic plants, which may be due to challenges in the dietary conditions required to metabolize plant secondary metabolites. The large number genes found in many insect lineages perhaps reflect their enormous ecological success and diversity (Helmkampf et al. 2015; Rane et al. 2016). For instance, P. rapae is a specialist herbivore that feeds on cyanogenic plants, whereas Spodoptera species, such as S. littoralis and S. frugiperda, are major polyphagous pests that infest >80 economically important plant species from 40 different families (Brown and Dewhurst 2009; Baloch et al. 2020). These insect species are exposed to very different plant secondary metabolites, and the expansion of CYS genes may allow them to handle the broad range of secondary metabolites (Schramm et al. 2012).

Over the course of genome evolution, functional divergence is the most likely explanation for the retention of duplicate genes. Divergence in expression patterns has been reported for various functional duplicated genes among multicellular species (Gu et al. 2002; Wagner 2002; Gu et al. 2005; Ganko et al. 2007; Farre and Alba 2010; Liu et al. 2015; Leite et al. 2018; Zhou et al. 2019). In the present study, pronounced divergence between the spatial expression patterns of S. frugiperda CYS paralogs was observed, suggesting that the CYS paralogs have undergone subfunctionalization or neofunctionalization. Previous studies demonstrated that the transcript levels of PrapCYS1 were highest in gut tissue, which was likely the major site of cyanide liberation upon ingestion of cyanogenic plants. PrapCYS1 has β-cyanoalanine synthase activity and plays an important role in cyanide detoxification (Van Ohlen et al. 2016). Our results show that SfruCYS2 was highly expressed in the larval midgut and the recombinant SfruCYS2 also showed higher CAS enzyme activity. In contrast, SfruCYS1 and BmorCYS were highly expressed in the Malpighian tubule, but showed relatively low expression in the midgut. These results suggest that BmorCYS and SfruCYS1 were likely involved in diuresis, detoxification, and metabolism in larval Malpighian tubule. Early expression divergence between duplicated genes may be a phenomenon that reduces the chances that one paralog becomes a pseudogene (Force et al. 1999; Lynch and Force 2000). In addition, lepidopteran CYS duplicate genes showed different enzymatic activities. For example, the kinetic values of Pieridae CYS3 and Noctuidae CYS2 for KCN was higher than that of other CYSs from the same species in Pieridae and Noctuidae, indicating divergence in their biochemical properties.

The evolutionary mechanisms responsible for the retention and subsequent functional divergence of duplicate genes after HGT events are poorly understood. In this study, we reconstructed the evolutionary history of HGT-derived CYS genes in lepidopteran genomes, supporting hypothesis that HGT events might have occurred in mite ancestor and lepidopteran ancestor, respectively (Wybouw et al. 2014). By examining the gene expression and the enzymatic properties, we revealed the evolutionary and functional dynamics of CYS genes. Lepidopteran CYS genes were differentially expressed in a tissue-specific manner, which presumably accelerates functional diversification to optimally regulate cysteine biosynthesis and cyanide detoxification. Our investigation provides valuable insights into the function and fate of lepidopteran CYS genes. This study also provides an example of how gene transfer and gene duplication can contribute to the adaptation and ecological success of insects.