Comparative transcriptomic and metabolomics analysis of modified atmosphere responses in Tribolium castaneum (Coleoptera: Tenebrionidae)

Abstract

Modified atmosphere is effective in controlling Tribolium castaneum Herbst, but it has adaptations. Comprehending the potential mechanism of resistance to T. castaneum in a modified atmosphere will help advance related management methods. This study conducted a comparative transcriptomic and metabolomic analysis to understand the physiological mechanism of T. castaneum in adapting to CO₂ stress. Results showed that there were a large number of differentially expressed genes (DEGs) in T. castaneum treated with different concentrations of CO₂. Gene ontology (GO) analysis revealed significant enrichment of DEGs mainly in binding, catalytic activity, cell, membrane, membrane part, protein-containing complex, biological regulation, and cellular and metabolic process. Kyoto Encyclopedia of Genes and Genomes analysis showed that different treatments had different effects on the metabolic pathways of T. castaneum. DEGs induced by 25% CO₂ were involved in arginine and proline metabolism, and 50% air + 50% CO₂ treatment affected most kinds of metabolic pathways, mainly the signal transduction pathway, including PI3K-Akt signaling pathway, AMPK signaling pathway, neurotrophin signaling pathway, insulin signaling pathway, and thyroid hormone signaling. Ribosome and DNA replication were enriched under high CO₂ stress (75% and 95%). The metabolomics revealed that different concentrations of CO₂ treatments might inhibit the growth of T. castaneum through acidosis, or they may adapt to anoxic conditions through histamine and N-acetylhistamine. Multiple analyses have shown significant changes in histamine and N-acetylhistamine levels, as well as their associated genes, with increasing CO₂ concentration. In conclusion, this study comprehensively revealed the molecular mechanism of T. castaneum responding to CO₂ stress and provided the basis for an effectively modified atmosphere in the T. castaneum.

Graphical Abstract

Graphical Abstract.

Introduction

Carbon dioxide (CO₂) is a nonflammable, odorless gas that effectively affects insect respiration and does not remain in food (Sadeghi et al. 2021). Therefore, elevated CO₂ (hypercapnia)/low O₂ (hypoxia) at biologically achievable levels have been demonstrated to be cost-effective and environmentally friendly in the control of stored-grain pest control, which belongs to a modified atmosphere (Mehmood et al. 2018). Many studies have assessed the toxic effects of hypercapnia or hypoxia on warehouse pests. For example, elevated CO₂ levels within 24 h increased the mortality rate of Tribolium castaneum or Rhyzopertha dominica in dried apricots without affecting the quality of the product (Sadeghi et al. 2021). Exposure for 12–37 h to >99% CO₂ caused almost complete mortality of aphids, mites, thrips, whiteflies, Cadra cautella, and T. castaneum (Held et al. 2001; Husain et al., 2017). In addition to mortality, exposure to high levels of CO₂ or low levels of O₂ has reduced insect reproduction (Azzam et al. 2010). For example, CO₂ exposure decreased sperm storage by T. castaneum in females (Fedina and Lewis 2004), and 2% and 5% oxygen negatively affected the total egg production of Callosobruchus maculatus (Yan et al., 2016). The reasons for these toxic effects are multilayered, including a series of physiological and potential pathological effects on insects, such as immune dysfunction, metabolic disorder, and apoptosis signal transduction disorders (Cui et al. 2017).

Insects, in turn, have mechanisms to adapt to hypercapnia/hypoxia, which affected negatively the efficiency of modified atmosphere to control pests (Cao et al. 2019). To cope with the hypoxia stress, insects have developed a variety of tools, both physiological and metabolic.

Several studies have shown how insects respond to hypoxic stress through physiological changes. Insects are observed to compress their bodies in order to generate convection currents, increasing ventilation and facilitating gas exchange to maintain aerobic metabolism under low-oxygen conditions (Hoback and Stanley 2001, Matthews and White 2011, Nika et al. 2022). Furthermore, research on Tenebrio molitor has revealed a significant reduction in the duration of stomatal closure in low-oxygen environments, along with an increase in CO₂ emissions during agitation (Hetz and Bradley 2005). Notably, the tracheal system of insect larvae exhibits plasticity, as observed by the sprouting of tracheal branches toward oxygen-deprived tissues under hypoxic conditions (Centanin et al. 2010). Various physiological changes are aimed at improving gas exchange efficiency and increasing CO₂ emission to cope with hypoxia stress.

Besides physiological adaptation, metabolic adaptation is also crucial for insect hypercapnia/hypoxia adaptation. This includes intermediate metabolites related to key metabolic pathways such as autophagy, energy regulation, and signaling transduction pathways. (Wang and Zhao 2003, Zhou et al. 2021, Valko et al. 2022). Both autophagy and energy metabolism-related changes are considered to be key ways for Drosophila to cope with hypoxia (Feala et al. 2007, Valko et al. 2022). Notably, hypoxia-inducible factor (HIF) is a major transcription factor that responds to hypoxia (Du et al. 2021). Stimulating the expression of downstream genes related to apoptosis and glucose metabolism, HIF-1 plays a crucial role in hypoxia adaptation (Majmundar et al. 2010, Valko et al. 2022). Consequently, gaining a comprehensive understanding of the molecular mechanisms governing insects’ response to high CO₂/low O₂ environments is essential for harnessing the potential benefits of modified atmospheres.

Red flour beetle (T. castaneum) is a worldwide pest that mainly affects the storage of grains, legumes, and oils. They destroy the grain by feeding the germ and endosperm and polluting the grain by producing feces that contain benzoquinones (Kharel et al. 2019). Insecticide is the main strategy for the prevention and control of T. castaneum at present, including slow-release grain fumigants such as phostoxin and essential oils (Sileem et al. 2020). However, frequent and high doses of insecticides increase costs, pollution, and risks to human health (Feroz 2020). Therefore, it is necessary to use green technology, such as modified or controlled atmospheres, for grain storage (Damcevski et al. 2010). However, although there have been experimental data showing increased mortality from elevated CO₂ levels, adults of T. castaneum have some adaptation to low O₂ and increased CO₂ (Krishnamurthy et al. 1986, Sadeghi et al. 2021). In addition, there have been some studies on the mechanism of insect tolerance to hypoxia and high CO₂ in recent years, but it is still not completely clear. The transcriptomics and metabolomics widely contribute to understand the comprehensive physiological and metabolic changes of organisms induced by environmental factors (Wang et al. 2016). In our previous work, we investigated the effects of different concentrations of CO₂ treatment on the growth and development of T. castaneum (Zhou et al., 2022). Currently, comprehensive exploration and annotation of the key genes involved in the CO₂ resistance of red flour beetle have not been conducted. In this study, continuous concentrations of CO₂ were established as the basis for investigating the responsive signaling pathways and metabolites of red flour beetle under CO₂ stress. The use of transcriptomics and metabolomics enabled a multiomics analysis, which, in turn, revealed the key pathways and metabolites involved in the red flour beetle’s response to high CO₂ concentration. Consequently, these findings offer valuable insights into the mechanisms governing insect responses to hypoxic stress.

Materials and Methods

Insects Rearing

All the T. castaneum used in this study came from the long-term breeding population of our laboratory. Their diet was 5% whole wheat flour containing yeast, and populations were cultured in an artificial climate incubator with a temperature of 29 ± 1°C, relative humidity of 65%± 5%, and photoperiod of 0L:24D.

Modified Atmosphere and Sampling

The 8th instar larvae (22th days after hatching) were put into a glass bottle, and an appropriate amount of whole meal flour containing 5% yeast was added as feed. The culture bottle was capped and sealed, and the CO₂ gas with different concentrations was passed through a tube from the CO₂ concentration control system (Changsha, Hunan, China). The ventilation was stopped after the CO₂ concentration ratio reached the requirements. In this experiment, the setting of the treatment group was the same as in the previous study (Wang et al. 2022), and the details are as follows: 100% air, 75% air + 25%CO₂, 50% air + 50%CO₂, 25% air + 75%CO₂, and 5% air + 95%CO₂. Treatment with 100% air was used as a control, and all experiments were repeated 3 times. After modified atmosphere treatment, the culture bottles were put into the artificial climate chamber for cultivation and observation. After 48 h, 90 nymphs were used for transcriptomic and metabolomics sequencing. The transcriptomic sample contained 3 biological replicates, while the metabolomics sample contained 6 biological replicates.

RNA Isolation and Transcriptomic Sequencing

Total RNA was extracted from 10 nymphs at modified atmosphere treatment for 48 h. In this study, total RNA was isolated by Trizol reagent (Takara, Kyoto, Japan). The purity and concentration of each RNA sample (1 μL) were measured with Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and the integrity was assessed using a 1% agarose gel. Subsequently, the qualified RNA was sent to the company (BGI Genomics Institute, Shenzhen, China) for transcriptomic sequencing.

Transcriptome Sequencing Analysis

RNA purification, reverse transcription, library construction, and sequencing were performed at BGI Genomics Institute. The raw read data obtained from sequencing were analyzed for standardized information. First, reads with low quality, contaminated joints, and high content of unknown base (N) were filtered out using SOAPnuke (V1.4.0, BGI), and the filtered data were called clean reads. All clean reads were then aligned to the reference genome (GCF_000002335.3) using HISAT (http://www.ccb.jhu.edu/software/hisat) (Kim et al. 2015), followed by new transcript prediction, SNP & InDel, and differential splicing gene detection. After obtaining the new transcript, the new transcript with coding potential was added to the reference gene sequence to form a complete reference sequence, and then the unigene expression level was calculated using RSEM (http://deweylab.biostat.wisc.edu/rsem/rsem-calculate-expression.html) (Langmead and Salzberg 2012; Li and Dewey 2011). Differentially expressed genes (DEGs) were identified with DEseq2 software, and the parameter is set as q-value (adjusted P-value) < 0.05 (Anders and Huber 2010, Love et al. 2014). In addition, functional-enrichment analyses, including GO (http://www.geneontology.org) and Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg/), were performed to identify which DEGs were significantly enriched in GO terms and metabolic pathways. Generally, a function with q value < 0.05 is significantly enriched.

qRT-PCR Verification

The qualified RNA mentioned in 2.3 was used as a template, and the 1st-strand complementary DNA (cDNA) was synthesized with the PrimeScript RT reagent kit with gDNA Eraser (Takara, Kyoto, Japan) following the manufacturer’s instructions. The verification of transcriptome sequencing was verified by qRT-PCR of 12 genes using the cDNA template diluted 3 times. The 12 genes were all related to trehalose metabolism pathway, including 655737 (facilitated trehalose transporter), 103312735 (MFS transporter), 103313476 (MFS transporter), 107397410 (MFS transporter), 655214 (MFS transporter), 655300 (MFS transporter), 657931 (MFS transporter), 658455 (MFS transporter), 660844 (MFS transporter), 662580 (MFS transporter), 664153 (MFS transporter), and 641504 (chitin synthase). Primer Premier 5 software was used for primer design, and rp13a was used as the internal reference (Table 1). The melting curve was used to confirm the specificity of primers. The qRT-PCR program was set as follows: predenaturation at 95 °C for 5 min, 39 cycles at 95 °C for 5 s, 60 °C for 20 s, followed by a final extension at 72 °C for 10 min. Three technical repeats were conducted. Finally, the 2^−ΔΔCT was used to calculate the relative expression of genes (Livak and Schmittgenb 2001).

Table 1.

Primers of transcriptome validation sequences used for qRT-PCR

Gene ID	Forward primer (5ʹ–3ʹ)	Reverse primer (5ʹ–3ʹ)
655737	CAAAACCGACCAGTCACGC	CATCAATCCAAGCACAATCAGTAG
103312735	TCAATGGGTATCCACGAAGC	CCAGAAGCGTACCAAAGAAGA
103313476	CTGGAGATAAGAGGGGTAGCG	TCGGACAACTTCTTTTGGTGA
107397410	CGTGCGTACCGTTGCTTT	CCATACTTCGGCTCCTTCCT
655214	TCCCGAAAATCGTAGCCG	CCCAAGCCTCCAATGAACC
655300	AGTCCTTCGGAATCTCCATAGT	ACAACCAAAACGTCCCCTC
657931	CGGTCGGGTTAGTGTCCTT	ATCGTGCTTCCAGCGTCA
658455	GAACAACCCACAACCTATCCA	CAAAACGCACCCAAAGTCA
660844	CCCGAGACTCAACGGAAGT	GGCAAAGCCACGATTAGG
662580	GGAAACTTTGATGAGTATCTGGC	CAACACCGCTGATTATGGC
664153	AGGATTTAGGCGTGGAGGG	CAGTGCTGTGATCTGATTTGGTAT
641504	GCCCTCCTGATATTCTACATTCT	CTGCGATTCCTTGTTGCTG

Untargeted Metabolomics

Sample collection methods are described in “Modified Atmosphere and Sampling” section. The thawed samples were processed by adding the extract solution (methanol: acetonitrile: water = 2:2:1, v: v: v, precooled at −20 °C) and steel balls, and then ground in a tissue grinder (JXFSTPRP, Shanghai Jingxin, China) at 50 Hz for 5min. After another 10 min of sonication in a 4 °C water bath, the grinding solution was allowed to stand for 1 h in −20 °C refrigerator. The next step involved centrifugation of the samples at 25,000 rpm for 15 min at 4 °C. After centrifugation, the supernatant was drained in a cryo-vacuum concentrator (MaxiVacbeta, GENE COMPANY), then dissolved by adding the solution (methanol: H₂O = 1:9, v: v). Liquid of dissolution was vortexed for 1 min, sonicated in a water bath at 4 °C for 10 min, and centrifuged at 25,000 rpm for 15 min at 4 °C. Finally, the supernatant was removed and placed in a loading flask for liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis.

The untargeted metabolomics analysis was performed using LC-MS/MS technique. A high-resolution mass spectrometer, Q Exactive HF (Thermo Fisher Scientific, USA), was used to collect positive and negative ion data, respectively, to improve the metabolite coverage. Compound Discoverer 3.1 (Thermo Fisher Scientific, USA) software was used for LC-MS/MS data processing, mainly for peak extraction, peak alignment, and compound identification. The identified metabolites were classified and annotated using KEGG Database and Human Metabolome Database (HMDB, https://hmdb.ca/). Functional annotation of the identified metabolites was carried out through KEGG database to understand the functional properties of different metabolites and determine the main biochemical metabolic pathways and signal transduction pathways involved in metabolites. In this experiment, principal component analysis (PCA) was used to detect the overall distribution of single samples in each group and the degree of dispersion between each group. Then, the partial least squares method-discriminant analysis (PLS-DA) model was used to calculate the variable important for the projection (VIP) values of the first 2 principal components. Finally, the fold change was obtained by the variation multiple analysis, and the P-value was obtained by t-test and corrected by false discovery rate to calculate q-value. The screening conditions of differential metabolites were VIP ≥ 1 and fold change ≥ 1.2 or ≤0.8 as well as q-value < 0.05.

Cluster analysis of differential metabolites was performed with a hierarchical cluster. The metabolic pathway enrichment analysis of differential metabolites was conducted based on KEGG database. The metabolic pathway with P-value < 0.05 was significantly enriched.

Combined Analysis of Transcriptomics and Metabolomics

Using regularized canonical correlation analysis (rCCA) (Rohart et al. 2017), we conducted a correlation analysis to measure the degree of association between differential genes and differential metabolites and then performed transcriptome and metabolome clustering. The differential genes and metabolites were analyzed using the block.splsda function in the mixOmics package and the results were visualized using the plotVar and circosPlot functions (González et al. 2012). Genes and metabolites with absolute correlation coefficients >0.99 and a P-value < 0.01 were selected for network analysis. The results were visualized using Cytoscape (version 3.10.1) software.

Data Analysis and Plotting

The data of qRT-PCR were shown as mean + standard errors and analyzed using Student’s t-test. The R package and SigmaPlot 14.0, as well as Chiplot (https://www.chiplot.online/) were used to plot.

Results

Identification of Differentially Expressed mRNAs by RNA-Seq

The total raw reads of 15 samples were all 43.82 M, and the total readings after filtration were at least 42.43 M and at most 43.25 M, respectively. The total base number after filtration was at least 6.37 Gb and at most 6.49 Gb, respectively, and the Q20 values were all above 96%, and even the Q30 values were all above 91% (Supplementary Table S1). In addition, the proportion of clean reads of reference genome in each sample group was above 80% (Supplementary Table S2).

Compared with the control group, 944, 4120, 3452, and 4560 genes were differentially expressed after 25% CO₂, 50% CO₂, 75% CO₂, and 95% CO₂ treatment, among which 579, 2,110, 1,939, and 2,415 genes were upregulated, respectively, and 365, 2,010, 1,513, and 2,145 genes were downregulated (Fig. 1A–D). There were 558 common DEGs between the 4 comparative groups (Fig. 1E). Meanwhile, there were 54 unique DEGs in 25% CO₂ group, 1,053 unique DEGs in 50% CO₂ group, 325 unique DEGs in 75% CO₂ group, and 1,535 unique DEGs in 95% CO₂ treatment group (Fig. 1E).

Fig. 1.

Identification of DEGs under different concentrations of CO₂ treatment. CK, 100% air; Treat A, 75% air + 25% CO₂; Treat B, 50% air + 50% CO₂; Treat C, 25% air + 75% CO₂; Treat D, 5% air + 95% CO₂. A) volcano plot for the DEGs between 100% air group and 75% air + 25% CO₂ group. B) volcano plot for the DEGs between 100% air group and 50% air + 50% CO₂ group. C) volcano plot for the DEGs between 100% air group and 25% air + 75% CO₂ group. D) volcano plot for the DEGs between 100% air group and 5% air + 95% CO₂ group. E) Venn plot for the DEGs.

GO Annotation of DEGs

The results showed that there were 830, 3,866, 3,296, and 4,220 DEGs belonging to molecular function genes in 25% CO₂, 50% CO₂, 75% CO₂, and 95% CO₂ groups, respectively. Meanwhile, there were 980, 5,169, 4,433, and 5,710 belonging to biological process, and 1,303, 6,888, 5,988, and 7,773 belonging to cellular components, respectively (Fig. 2). However, there were similarities in the functions of DEGs among the 4 groups, all of which focus on binding, catalytic activity, cell, membrane, membrane part, protein-containing complex, biological regulation, cellular process, and metabolic process.

Fig. 2.

GO enrichment analysis of DEGs in Tribolium castaneum treated with different concentrations of CO₂. The number above the black line represents the gene counts.

KEGG Pathway of DEGs

KEGG pathway enrichment results showed that arginine and proline metabolism were significantly enriched after 75% air + 25% CO₂ treatment (q < 0.05) (Fig. 3A); Ribosome, cholesterol metabolism, PI3K-Akt signaling pathway, ECM-receptor interaction, focal adhesion, AMPK signaling pathway, neurotrophin signaling pathway, axon guidance, osteoclast differentiation, insulin signaling pathway, and thyroid hormone signaling pathway were significantly enriched after 50% air + 50% CO₂ treatment (q < 0.05) (Fig. 3B); ribosome and DNA replication were significantly enriched after 25% air + 75% CO₂ treatment (q < 0.05) (Fig. 3C); ribosome was significantly enriched after 5% air + 95% CO₂ treatment (Fig. 3D). Therefore, different concentrations of CO₂ had different effects on the metabolic pathways of T. castaneum, and 50% air + 50% CO₂ treatment affected the most kinds of metabolic pathways.

Fig. 3.

The differentially expressed genes in the top 20 KEGG metabolic pathways. CK, 100% air; Treat A, 75% air + 25% CO₂; Treat B, 50% air + 50% CO₂; Treat C, 25% air + 75% CO₂; Treat D, 5% air + 95% CO₂. The KEGG pathway with q-value < 0.05 was considered as enriched significantly. A-D indicated the results of treatA-D vs CK.

RT-qPCR Verification of DEGs Identified With mRNA‑Seq

To confirm the reliability of our RNA-Seq results, we conducted RT-qPCR on selected DEGs and compared the results with those obtained from RNA-Seq. Twelve DEGs were selected for RT-qPCR detection, including trehalose transporter (655737), other sugar transporter (103312735, 103313476, 107397410, 655214, 655300, 657931, 658455, 660844, 662580, 664153) as well as a chitin synthase (641504). The results showed that the trend of relative expression of these genes determined by RT-qPCR was consistent with the DEGs determined by RNA-Seq, which proved that the data obtained by RNA-Seq was reliable (Fig. 4).

Fig. 4.

Comparison of the transcriptome analysis and qRT-qPCR validation. CK, 100% air; TA, 75% air + 25% CO₂; TB, 50% air + 50% CO₂; TC, 25% air + 75% CO₂; TD, 5% air + 95% CO₂. The data of qRT-PCR were shown as mean + standard errors and analyzed using Student’s t-test.

Analysis of Differential Metabolites

In negative ion mode, there were 254 significantly different metabolites between 75% air + 25% CO₂ treatment group and control group, among which 138 were upregulated and 116 were downregulated; there were 60 significantly different metabolites between the 50% air + 50% CO₂ treatment group and control group, among which 56 were upregulated, and 4 were downregulated; there were 344 significant different metabolites between 25% air + 75% CO₂ treatment group and control group, among which 190 were upregulated, and 154 were downregulated; there were 333 significant different metabolites between the 5% air + 95% CO₂ treatment group and the control group, among which 169 were upregulated and 164 were downregulated (Supplementary Fig. S1; Supplementary Table S3). In positive ion mode, there were 521 significantly different metabolites between 75% air + 25% CO₂ treatment group and control group, among which 180 were upregulated, and 341 were downregulated; there were 110 significantly different metabolites between the 50% air + 50% CO₂ treatment group and control group, among which 95 were upregulated, and 15 were downregulated; there were 992 significant different metabolites between 25% air + 75% CO₂ treatment group and control group, among which 436 were upregulated, and 556 were downregulated; there were 776 significant different metabolites between the 5% air + 95% CO₂ treatment group and the control group, among which 429 were upregulated, and 347 were downregulated (Supplementary Fig. S1; Supplementary Table S3).

The results indicated that CO₂ stress had significant effects on the physiological metabolism of T. castaneum, and the PCA analysis verified this difference (Supplementary Fig. S3). Overall, we detected 110 common differential metabolites in the 4 groups (Fig. 5A), among which 64 metabolites had identified information (Fig. 5B). Results showed that 5 metabolites, including 3-o-beta-d-galactosyl-sn-glycerol, ammeline, (r)-3-hydroxybutyrylcarnitine, etifenin, and cinnarizine were downregulated, while most of the metabolites were upregulated under CO₂ treatment, especially histamine, N-acetylhistamine, capryloylglycine, 3-[2-(hydroxymethyl)-4-methoxyphenyl]-6-methoxy-4-oxo-3,4-dihydro-1(2h)-quinazolinecarbaldehyde, procainamide, oxolamine, pentahomomethionine, capryloylglycine, miltefosine as well as acids such xanthurenic acid (Fig. 5B).

Fig. 5.

Impact of different concentrations of CO₂ on the metabolic profile of Tribolium castaneum. CK, 100% air; Treat A, 75% air + 25% CO₂; Treat B, 50% air + 50% CO₂; Treat C, 25% air + 75% CO₂; Treat D, 5% air + 95% CO₂. A) Venn plot for the differentially accumulated metabolites. B) The common metabolites with identified information among the CK vs. Treat A, CK vs. Treat B, CK vs. Treat C, and CK vs. Treat D group. C) The enriched KEGG metabolic pathways in CK vs. Treat A group are in negative mode. D) The KEGG metabolic pathways in CK vs. Treat B group are in negative mode. E) The KEGG metabolic pathways in CK vs. Treat C group are in negative mode. F) The KEGG metabolic pathways in CK vs. Treat D group are in negative mode. G) The enriched KEGG metabolic pathways in CK vs. Treat A group are in positive mode. H) The KEGG metabolic pathways in CK vs. Treat B group are in positive mode. I) The KEGG metabolic pathways in CK vs. Treat C group are in positive mode. J) The KEGG metabolic pathways in CK vs. Treat D group are in positive mode.

However, the different effects on the physiological metabolism of T. castaneum were caused by different concentrations of CO₂. We investigated the functions of differentially accumulated metabolites in T. castaneum using KEGG pathway annotation. Compared with the 100% air treatment group, amino metabolism (tryptophan, histidine, arginine, and tyrosine), galactose metabolism, and purine metabolism were significantly enriched after 25% CO₂ treatment (Fig. 5C and G); amino metabolism (tryptophan, histidine, phenylalanine, and lysine) and fatty acid biosynthesis were significantly enriched after 50% CO₂ treatment (Fig. 5D and H). The high concentration of CO₂ stress (75% and 95%) caused the changes of more metabolic pathways of T. castaneum. Amino metabolism (tyrosine, tryptophan, phenylalanine, histidine, alanine, aspartate, glutamate, arginine, and proline), nucleotide metabolism (purine and pyrimidine), taurine and hypotaurine metabolism, propanoate and butanoate metabolism, pantothenate and CoA biosynthesis, and energy metabolism (pyruvate metabolism and TCA cycle) were enriched after 75% CO₂ treatment (Fig. 5E and I). The differential metabolic pathways affected by 95% CO₂ stress were similar to those affected by 75% CO₂ stress, but other than those, the pentose phosphate pathway and glycolysis/gluconeogenesis pathway were significantly enriched (Fig. 5F and J).

Multiomics Analysis Revealed the Mechanism of Histamine Metabolism Under Hypoxia Stress

The integrated analysis of transcriptomic and metabolomic data complements and verifies the relationship between genotypic and phenotypic data, jointly explaining the influence of gene function and environmental factors on biological phenotypes. To verify the relatedness between transcription and metabolism under hypoxic stress, a multiomics integrated analysis was conducted to gain a comprehensive understanding of the regulatory mechanisms and changes of T. castaneum under CO₂ stress.

In the correlation clustering analysis of transcriptomic and metabolomic data, each row represents a differential metabolite, while each column represents a differential gene. The clustering heatmap indicates a strong correlation between differential genes and metabolites, with blue color indicating a negative correlation and red color indicating a positive correlation (Fig. 6A). To further validate the strength of their correlation, a multiomics multivariate dimensionality reduction technique was employed, coupled with the integration of multiple datasets to reveal relationships between omics. The Circos plot visualizes the details of all differential metabolites and genes, highlighting the significantly correlated pairs (Fig. 6B). The results of the Circos plot show that as CO₂ concentration increases, the number of correlated genes and metabolites generally increases, with a majority of differentially correlated metabolites and genes showing a positive association. In the correlation circle plot, a more dispersed pair of differential metabolites and genes indicates a higher correlation (Fig. 6C). Analysis of the correlation circle plot reveals that with increasing CO₂ concentration, there is a higher number of differentially correlated metabolites and genes, indicating a stronger correlation, which is consistent with the results of the Circos plot. Notably, in the 50% concentration treatment group, while the overall correlation strength increased compared to the 25% concentration treatment group, the number of correlated differential metabolites decreased. These findings are potentially related to the maximum tolerance capacity of T. castaneum.

Fig. 6.

Results of transcriptome metabolome association analysis. In the figures, Treat A, Treat B, Treat C, and Treat D correspond to the 25%, 50%, 75%, and 95% CO₂ treatments, respectively. A) Transcriptome and metabolome correlation clustering heat maps. Each row represents one differential metabolite, and each column represents one differentially expressed gene. Blue represents a negative correlation, and red represents a positive correlation. B) Circos plots of correlation between differential genes and differential metabolites. The lines in the circles represent differential genes and differential metabolism. The correlation coefficient between objects is greater than or equal to 0.99. The blue and orange curves in the periphery represent differential metabolites and differential gene expression in the two groups of samples. C) Concentric circles plots of correlation between differential genes and differential metabolites. Each point in the circle represents a gene, and each square represents a metabolite. The relationship between differential genes and differential metabolites was determined by being on the first and second principal components. The projection is determined. A positive correlation was considered if the angle between the differential gene and the differential metabolite was acute (<90 °C). If the angle between the differential gene and the differential metabolite is a ton angle (>90 °C and <180 °C), the correlation is negative. The heart was used as the starting point, and the line was connected to the differential metabolites and differential genes; the longer the line length was, the stronger the relationship was, and vice versa. In general, variables far from the center of the circle are more strongly correlated.

We observed significant changes in the levels of histamine and N-acetylhistamine in response to differential metabolites (Fig. 7A), with nearly identical shifting patterns. N-acetylhistamine, an important histamine metabolite, serves as a key biomarker in allergic reactions in numerous animals. Based on the metabolomics results, under CO₂ stress, the levels of histamine and N-acetylhistamine in T. castaneum increased significantly with the intensification of stress. Hence, histamine and N-acetylhistamine can serve as biomarkers for T. castaneum responding to CO₂ stress. Although the exact catalyzing enzyme for N-acetylhistamine generation has not been precisely identified, its type (EC: 2.3.1.-) and reaction are relatively well established. We compiled and clustered the expression profiles of all relevant enzyme DEGs (Fig. 7B). The results revealed that the majority of the genes were significantly upregulated, with higher expression levels in the 95% CO₂ treatment group, while the 25% concentration treatment group exhibited the lowest level of DEGs changes. Using histamine and N-acetylhistamine as key metabolites for screening highly correlated DEGs, a large number of associated genes were discovered. Still, only a small subset of genes appeared repeatedly in multiple treatments (Supplementary Fig. S2). Six differential genes (661916, 660223, 660695, 663651, 662302, and 655676) were revealed through histamine-based screening that appeared in 3 treatments and displayed an upward trend with increasing CO₂ concentration (Fig. 7C). Similarly, when using N-acetylhistamine to screen for associated genes, 5 DEGs (658229, 656141, 657796, 103313165, and 659253) across the 3 treatments were identified. Like histamine-based genes, these genes also exhibited an increase in response to elevated CO₂ concentration (Fig. 7D). These genes may be associated with the response to CO₂ stress, but their enrichment cannot be traced back to a specific pathway or function, indicating their involvement in diverse changes rather than a specific function.

Fig. 7.

Histamine metabolism changes significantly under hypoxia stress. A) Histamine is catalyzed via EC: 2.3.1.- to produce N-acetylhistamine. The line graph shows the levels of these substances under various CO₂ concentration treatments, with histamine in left and N-acetylhistamine in rihgt. B) All gene expression profiles annotated as EC:2.3.1.-. C) and D) The expression profiles of highly associated genes were screened using histamine and N-acetylhistamine as key metabolites. Turkey’s HSD test was used for significance analysis. “*” represents statistical significance at P < 0.05, while “**” represents statistical significance at P < 0.01.

Discussion

Pests in storage have long been a challenge for agricultural production. Existing methods use phosphine to control pests, such as the red flour beetle. However, this approach brings concerns about toxic residue and environmental harm. Although some researchers have begun to promote plant essential oils for storage pest control, such methods seem to be environmentally friendly and safe, but there are problems of high cost and residue (Kavallieratos et al. 2021). Modified atmosphere stress is an emerging means of storage pest control in recent years. As a green and environmentally friendly pest control program, it has been a hot program to replace phosphine pest control. Red flour beetle is a major household pest against a wide variety of cereals, nuts, and grains (Athanassiou et al. 2010), and modified atmospheres such as CO₂ and N₂ are the most environment-friendly approaches (Held et al. 2001, Cheng et al. 2012, Qu et al. 2022). Previous studies have reported that the mortality rate of T. castaneum is 99% after exposure timings of 29.3 h, 153.9 h, and 78.4 h against adult, pupa, and larval stages, respectively, under 99.9% CO₂ treatment, and high concentration CO₂ resulted in a lower weight and shrunken body size of T. castaneum, which demonstrated that CO₂ treatment is an effective means to eliminate the T. castaneum (Husain et al., 2017, Zhou et al., 2022). However, with the increase in hypoxia/hypercapnia stress, the insects adopt different response strategies, such as tracheal dilation, behavioral change, increased respiratory metabolism, HIF transcription regulation system, and regulated energy pathways (Harrison and Haddad 2011, Zhou et al. 2022). However, only a few studies have focused on the physiological changes of T. castaneum under CO₂ stress. Transcriptomic sequencing is helpful for understanding the molecular mechanism that T. castaneum responds to CO₂ stress at the molecular level (Du et al. 2023). In this study, Q30 values of clean reads in each group were all above 91%, and the comparison efficiency of clean reads was also as high as above 80%, indicating high data quality of transcriptome sequencing. Results of qRT-PCR also demonstrated the accuracy of transcriptome sequencing.

A large amount of DEGs and metabolites were found. However, the DEGs and metabolite counts were different with the different concentrations of CO₂ treatment, which demonstrated that different concentrations of CO₂ had different effects on the physiological metabolism of T. castaneum. Previous studies also suggested that the upregulation and downregulation of the gene were not consistent under different concentrations of CO₂ (Zhou et al. 2022). KEGG pathway enrichment results showed that DEGs were significantly enriched in arginine and proline metabolism after 75% air + 25% CO₂ treatment, and the differential metabolites were significantly enriched in amino acid metabolism (tryptophan, histidine, arginine, and tyrosine), galactose metabolism, and purine metabolism. It is not clear what role amino acids play in insect tolerance to hypoxia stress, while previous studies also indicated that amino acid metabolism was the most significantly changed metabolic pathways in Callosobruchus chinensis larvae under hypoxia/hypercapnia treatment (Cui et al. 2017, 2019). It was worth noting that tryptophan metabolism in each group was all affected in this study, and it is an essential amino acid that plays an indisputable role in several physiological processes, such as neuronal function and immunity, which suggested that tryptophan metabolism might be a protective mechanism under CO₂ stress (Davidson et al. 2022). DEGs produced by 50% CO₂ treatment enriched the most abundant pathways. Specifically, transcriptome results showed that several signaling pathways were affected, including AMPK, PI3K-Atk, neurotrophin, and thyroid hormone signaling pathways. AMPK is a cell energy sensor, which is regulated by the ratio of AMP/ATP, while O₂ is the key factor for driving cellular metabolism in mitochondria to maintain cellular energy homeostasis (Garcia and Shaw 2017). The increase of reactive oxygen species caused by low-oxygen conditions may directly activate AMPK, which is consistent with our results (Hinchy et al. 2018). Similarly, PI3K-Akt and neurotrophin receptors could regulate the HIF-1, which maintains the internal environment stability of tissues and cells under low-oxygen conditions (Majmundar et al. 2010, Le Moan et al. 2012, Xie et al. 2019). Besides, it was reported that thyroid hormone metabolite also influenced the HIF-1 signaling pathway, reflecting the relationship between thyroid hormone signaling and hypoxia (Zhou et al. 2020). However, differential metabolites produced by 50% CO₂ treatment were enriched the least abundant pathways, as opposed to transcriptome, and most of them were related to amino acid metabolism.

The response of T. castaneum to high concentration CO₂ (75% and 95%) was different from low concentration CO₂ (25% and 50%). The ribosome, as the place for protein synthesis, was significantly changed after high-concentration CO₂ treatment, according to transcriptomics. This is consistent with 25% CO₂ treatment and similar to previous studies in cells, which found that hypoxia condition contributed to the regulation of ribosome biogenesis by TNF receptor‑associated protein 1 (Bruno et al. 2022). The change of ribosomes might be an effective mechanism to minimize hypoxia-induced specific protein changes and to enhance hypoxia tolerance of T. castaneum (Shah et al., 2011). In addition, the results of metabolomics showed that high concentrations of CO₂ affected many metabolic pathways. In addition to amino acid metabolism, nucleotide metabolism (purine and pyrimidine) and energy metabolism (pyruvate metabolism and TCA cycle) were also significantly changed. ATP degradation results in the accumulation of purine metabolites under hypoxic conditions in mammalian tissues, while the astrocytes are capable of normal metabolism during hypoxia in the presence of the pyrimidine nucleotide precursor (Sonnewald et al. 1998, Del Castillo Velasco-Martínez et al. 2016). Therefore, it is speculated that hypoxia resulted in the accumulation of purine, while pyrimidine might be a protective substance for hypoxia adaption. As for energy metabolism, its change is reasonable. Pyruvate metabolism and the TCA cycle are responsible for energy production, and the TCA cycle requires oxygen supply, as oxygen acts as a final electron acceptor in oxidative phosphorylation for energy production (Goda and Kanai 2012).

From a comprehensive analysis, there were 64 common differential metabolites with identified information in the 4 groups. Results showed that many acid substances such as xanthurenic acid, 3,4-dihydro-7-methoxy-2-methylene-3-oxo-2h-1,4-benzoxazine-5-carboxylic acid. It might be a hypercapnic acidosis caused by hypercapnia (Johnson 2017). Previous studies showed that high concentrations of CO₂ inhibited the growth and development of T. castaneum, and acidosis might be one of the potential mechanisms (Zhou et al. 2022).

In addition, it is evident from the heat map that the levels of histamine and N-acetylhistamine are significantly increased, and studies in mice have shown that histamine contributes to changes in metabolic rate during hypoxia to increase hypoxic ventilatory decline (Ishiguro et al. 2006, Ohshima et al. 2007). Hypoxic ventilatory decline is a physiologic survival mechanism that reflects reduced cellular activity in response to reduced cellular metabolism and energy requirements under hypoxia conditions (Ishiguro et al. 2006). Therefore, it is speculated that the accumulation of histamine and N-acetylhistamine was a protective mechanism of T. castaneum. It is worth noting that histamine and N-acetylhistamine, as important neurotransmitters in arthropods, are associated with the formation of various sensory systems (Aryal and Lee 2021). Some studies also suggest that histamine serves as a source of pheromones in some insects. The increase in histamine and its metabolites under CO₂ stress may signify potential severe damage or hindrance to the nervous system of the test animal, necessitating further research. Although the enzyme responsible for catalyzing the conversion of histamine to N-acetylhistamine remains unclear in this study, almost all relevant enzymes showed some degree of upregulation under CO₂ stress. This finding reinforces the possibility that histamine metabolism pathways may play an important role in responding to hypoxia stress.

In summary, these results have indicated that different concentrations of CO₂ have a wide range of effects on signal transduction and physiological metabolism in T. castaneum. Amino acid metabolism was the most significantly changed metabolic pathways, especially the tryptophan metabolism, while 50% CO₂ treatment significantly changed the signal transduction pathway related to energy metabolism. Besides, high-concentration CO₂ (75% and 95%) treatment affected the nucleotide metabolism and energy metabolism. It is worth noting that different concentrations of CO₂ treatments may inhibit the growth of T. castaneum through acidosis, or they may adapt to anoxic conditions through histamine and N-acetylhistamine. These results have provided potential molecular mechanisms of T. castaneum responding to CO₂ stress and were helpful for improving the effect of modified atmosphere combined with other methods, such as blocking the protective substance through RNA interference.

Funding

This research was financially supported by the National Natural Science Foundation of China (Grant No. 31960542), the Academician Workstation of Guiyang University, Guizhou Province (QKHRCP No. [2019] 5605), the Special Funding of Guiyang Science and Technology Bureau and Guiyang University (GYU-KY-[2021]), the Innovation Group Project of Education Department of Guizhou Province ((2021)013), the Guiyang Science and Technology Planning Project [2022] 5-20, Discipline and Master’s Site Construction Project of Guiyang University by Guiyang City Financial Support Guiyang University [2022-xk], and the Program for Natural Science Research in Guizhou Education Department (QJJ[2023]024), Special Project for Science and Technology Development of Local (Guizhou) under the Guidance of the Central Government (QKZYD[2022]4013).

Author Contributions

Min Zhou (Conceptualization [equal], Writing—original draft [equal]), Biying Pan (Conceptualization [equal], Data curation [equal], Investigation [equal], Visualization [equal]), Liwen Guan (Investigation [equal]), Yuanyuan Wang (Investigation [equal]), Kang-Kang Xu (Data curation [equal], Visualization [equal]), Shigui Wang (Data curation [equal], Visualization [equal]), Bin Tang (Writing—original draft [equal]), and Can Li (Writing—original draft [equal])