Identification and Validation of Reference Genes for Expression Analysis Using RT-qPCR in Leptocybe invasa Fisher and La Salle (Hymenoptera: Eulophidae)

Simple Summary In gene expression investigations, the first crucial step is choosing appropriate housekeeping genes. However, the choice of reference genes is not absolute but relative and varies with different experimental conditions. It is vital to note that using unvalidated or unscreened internal reference genes can lead to erroneous inferences. This study was conducted on Leptocybe invasa to calculate the stability of eight housekeeping genes across various test conditions, such as sexes, somites, temperatures, diets, and pesticides. The relative expression of HSP90 at different temperature settings was evaluated to validate the results. This study aims to assist future gene expression research on this invasive species and lay the groundwork for further investigations into the gene function of this pest. Abstract Leptocybe invasa (Hymenoptera: Eulophidae) is a globally intrusive pest. Despite extensive research into the physiological responses of this pest, our understanding of the molecular mechanisms still needs to be improved. We want to accurately investigate the expression of L. invasa’s target genes, so it is imperative to select fitting reference genes. In this study, eight housekeeping genes’ stability (RPS30, ACTR, 18S rRNA, ACT, RPL18, GAPDH, 28S rRNA, and TUB) was tested under five different experimental conditions, including male or female adults, somites (head, thorax, and abdomen), temperatures (0 °C, 25 °C, and 40 °C), diets (starvation, clear water, 10% honey water, Eucalyptus sap), and pesticides (acetone was used as a control, imidacloprid, monosultap). Gene stability was calculated using RefFinder, which integrates four algorithms (the ∆Ct method, geNorm, NormFinder, and BestKeeper). The findings implied that ACT and ACTR were the most accurate when comparing sexes. For analyzing different somites, 28S rRNA and RPL18 were ideal; the 28S rRNA and RRS30 were perfect for analyzing at different temperatures. The combination of ACT and GAPDH helped to analyze gene expression in different diets, and GAPDH and 28S rRNA were suitable for various pesticide conditions. Overall, this research offers a complete list of reference genes from L. invasa for precise analysis of target gene expression, which can improve the trustworthiness of RT-qPCR and lay the foundation for further investigations into the gene function of this pest.


Introduction
The Eucalyptus tree is Australia's national tree and one of the southern hemisphere's most crucial fast-growing tree species for industries [1]. Leptocybe invasa Fisher and La Salle (Hymenoptera: Eulophidae) is a pest that infests Eucalyptus woodlands worldwide. Since its initial detection in the Middle East and the Mediterranean in 2000, L. invasa has caused significant damage to nurseries and young Eucalyptus forests. Moreover, it has quickly spread to many Eucalyptus-growing nations, such as Australia, China, India, and Male and female adult L. invasa newly emerged from DH201-2 of Eucalyptus grandis × Eucalyptus tereticornis were randomly collected, one replicate for every 60 male and female adults, for a total of three biological replicates. All insects were conserved in 1.5 mL RNAase-free centrifuge tubes with RNA preservation solution for L. invasa. All samples were left all night at 4 • C and then put at −20 • C pending the extraction of RNA. This approach was employed to collect and preserve samples without further specific descriptions in the following experimental treatments. Three biological replicates were set up during each of the subsequent experiments.

Adult Somite
First, place a sterile Petri dish on ice, cover the Petri dish with a layer of sterile filter paper, and cut off the head, thorax, and abdomen of the adult worms with a special scalpel. The adult head, thorax, and abdomen were placed into three RNAase-free centrifuge tubes containing an RNA preservation solution. A total of 500 adult worms were dissected.

Temperature Treatments
We collected newly emerged L. invasa adults from Eucalyptus grandis × Eucalyptus tereticornis DH201-2, and we kept all adults at 25 • C for 4 h. Afterwards to prevent the adults from dying of starvation, the adults were sited individually in 1.5 mL centrifuge tubes with 2 µL of 10% honey water in each tube, then 1 h at 0 • C, 25 • C, and 40 • C in MGZ light incubators, with 100 L. invasa per replicate.

Diet Treatments
Adults were starved for 4 h after emergence and divided into four treatment groups: (i) no food as a control, (ii) water, (iii) a 10% honey solution, and (iv) diluted Eucalyptus grandis × Eucalyptus tereticornis DH201-2 sap. Samples were collected after 6 h. Sixty adults were used as a replicate.

Pesticide Treatments
L. invasa adults were collected on the day of fledged, then the adults were sited in 1.5 mL centrifuge tubes, one head per tube, and each tube was filled with 2 µL of 10% honey water on the cap and fed for 3 h to avoid starvation. Then two commonly used pesticides for the control of L. invasa were selected, imidacloprid and monosultap. About 200 mg/mL film tubes were made by dissolving the drug in acetone. We introduced L. invasa into the drug film tubes for 1 h. One hundred L. invasa were used as one replicate. Acetone film tubes were controls.

RNA Extraction and cDNA Synthesis
TRIzol (Tiosbio, Beijing, China) and the instructions from the RNeasy Plus Mini Kit were used to quickly extract RNA from L. invasa (No. 74134; Qiagen, Hilden, Germany). Our analysis of the abstraction RNA was carried out using 1% agarose gel to verify its integrity. The concentration and pureness of RNA were determined by a Nan-oDrop 8000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). A range of absorbance ratios was observed for the RNA samples at A 260/280 and A 260/230 . Both are around 2.0, indicating that they are appropriate for future research. Based on instructions from TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix (TranGen Biotech, Guangzhou, China), the first strand of cDNA was generated from each sample set. The resulting cDNA was then diluted by 20 µL for RT-qPCR. For the RT-qPCR, we maintained total RNA at −80 • C while keeping the complete cDNA at −20 • C.

Reference Gene Selection and Primer Design
In this study, based on the transcriptome data of L. invasa, many candidate reference genes were initially screened based on functional annotations. Then the candidate reference genes were further screened based on the FPKM value (FPKM > 50, medium expression is optimal and has similar expression levels in different samples), CV (CV < 0.15), and log2 fold value (absolute value less than 0.2) between samples. The corresponding gene sequences were found and then Blast compared on NCBI to homologous genes of other insects with 90% sequence similarity, which were used as candidate reference genes. We designed 18 primer pairs (product length 90-300) based on the CDS sequences of the corresponding genes and subsequently verified the stability through semi-quantitative RT-PCR and RT-qPCR. A total of eight primer pairs were selected for subsequent experiments. Eight internal reference genes were ribosomal protein S30 (RPS30), actin-related protein (ACTR), 18S ribosomal RNA (18S rRNA), actin (ACT), ribosomal protein L18 (RPL18), glyceraldehyde 3-phosphate dehydrogenase (GAPDH), 28S ribosomal RNA (28S rRNA), and B-tubulin (TUB). Using the web software Primer 3.0, primer pairs for amplification were created carefully by the RT-qPCR primer design guidelines with primer lengths of 20-22 bases, annealing temperature between 54 and 56 • C, and amplification product length greater than 90 bp and less than 300 bp [23,24]. DynaScience Biotechnology generated each primer in Table 1 (Beijing, China). Electrophoresis was performed on a 1% agarose gel to confirm the correctness of each primer. The sequences, lengths, and amplification efficiencies (E) of the eight benchmark genes' primers are provided in Table 1.

RT-qPCR
RT-qPCR was performed using a LightCycler ® 480II Real-Time PCR System in 96well plates (Roche Molecular Systems, Germany). Using Genious 2X SYBR Green Quick qPCR Mix (No ROX), the cDNA was amplified (ABclonal Technology, Woburn, MA, USA). The total cDNA template was subjected to a 5-fold gradient dilution to obtain cDNA templates at 5 0 , 5 −1 , 5 −2 , 5 −3 , and 5 −4 ng.µL −1 concentrations for gradient concentration standard curve plotting. A 20 µL reaction system was used: Genious 2X SYBR Green Fast qPCR Mix 10 µL, forward and reverse primers 0.4 µL, cDNA template 1 µL, and ddH 2 O supplemented. We performed the RT-qPCR reaction in a 3-step standard reaction mode: 3 min pre-denaturation at 95 • C, 5 s denaturation at 95 • C, 30 s annealing, and extension at 60 • C, 40 cycles; 15 s at 95 • C, 60 s at 60 • C, and 15 s at 95 • C to form a melting curve. Each cDNA sample was subjected to three technical replicates, three biological replicates, and parallel inclusion of template-free controls. The relationship between Ct values and logarithmic cDNA template concentrations was analyzed using SPSS20 software, with the latter taken as the horizontal coordinate and the former as the vertical coordinate. To quantify the linearity of this relationship, we estimated the linear equation's slope and regression coefficient (R 2 ). According to the formula, the amplification efficiency (E) values were obtained [25,26].

Analyzing Reference Genes and Handling Data
The stability of eight benchmark genes was valued in diverse conditions using analysis and screening tools, including ∆Ct, geNorm, NormFinder, and BestKeeper. Furthermore, the online tool RefFinder (https://blooge.cn/RefFinder/?type=reference, accessed on 18 January 2023) was employed to comprehensively rank all housekeeping genes. While the original quantized cyclic values (Ct) can satisfy the criteria for the BestKeeper and comparative ∆Ct algorithms, for geNorm and NormFinder Analysis, the actual Ct values must be transformed to relative quantities. This part of the data was plotted using Origin 2021 (OriginLab, Northampton, MA, USA).

Verification of Reference Gene Stability
Heat shock proteins are a current research hotspot because they are widely distributed throughout most animals and are highly conserved. These heat shock proteins repair damaged proteins in response to heat or cold stimuli to sustain the organism's regular life activities [26]. The accuracy of our experimental findings was further corroborated by the expression of the heat shock protein (HSP90) gene in L. invasa standardized by the two most optimal (28S rRNA and RPS30) and least reliable (RPL18 and TUB) reference genes in different temperature, with 25 • C serving as the control. Using the 2 −∆∆Ct , the relative expression levels of HSP90 at various temperatures were calculated [27]. The expression levels of genes in diverse dealings were examined using one-way ANOVA, and the results were compared using Tukey's highly significant difference test (Tukey's HSD). This section uses GraphPad Prism 9 (GraphPad, San Diego, CA, USA) to process and plot the data.

RNA Quality and Amplification Efficiency
The putative internal reference genes 28S rRNA, TUB, RPS30, ACTR, 18S rRNA, ACT, RPL18, and GAPDH were chosen based on the transcriptome analysis findings. Sequencing matching showed a greater than 90% sequence similarity with the same genes from other insects. Additionally, for the eight benchmark genes of L. invasa, the match's expected value (E) was 0 (or nearly 0), indicating a perfect match for the genes, which also shows the highly conserved nature of these internal benchmark genes. The Ct values for the eight benchmark genes were significantly correlated with the cDNA values at various concentration gradients (p ≤ 0.001, 0.988 ≤ R 2 ≤ 0.998, Table 1). A distinct single peak on the RT-qPCR solubility plots confirmed the specificity of the primers. These genes' amplification effectiveness (E) values varied from 93% to 114%, with R 2 > 0.990.

Levels of Expression of Reference Genes
The violin plot combines a bar chart (with the median as a white dot in the center) with a kernel density plot to provide a visual representation of the probability distribution of the data. The size of the area in the plot corresponds to the likelihood of the data being distributed around a certain value. Unlike a box line plot, the violin plot can show and more accurately represent the data distribution. The cycle of quantification (Ct) represents the transcript level of the mRNA. The stability of Ct values plays a crucial role in housekeeping gene selection: the level of expression of a gene depends on its Ct value; the lower the Ct value, the higher the expression level, and vice versa. RT-qPCR was employed in the evaluation of the expression patterns of eight internal control genes under diverse conditions. Figure 1 shows that the Ct values of the eight housekeeping genes ranged from 17.93 (TUB) to 29.7 (ACTR), with most between 22 and 27. According to further studies on the distribution of Ct values, the Ct values of the eight housekeeping genes were different under different conditions. Under different sex conditions, GAPDH expression levels were higher, and the Ct values of the eight housekeeping genes were mostly concentrated between 24 and 28, but the Ct values of GAPDH genes were concentrated around 22. Under different somites conditions, GAPDH expression levels were higher, and the Ct values of the eight internal reference genes were mostly concentrated between 22 and 26, while the Ct values of GAPDH genes were concentrated around 21.5. The expression levels of TUB were higher under different temperatures, diet, and pesticide conditions, but the Ct values of TUB genes were concentrated in 19, 21, and 21, respectively. Specifically, RPL18 and TUB had mean Ct values in the sex of 27.06 and 23.23, respectively, but under temperature conditions, they had mean Ct values of 22.81 and 19.55, respectively. In the sex condition, the Ct mean value of the ACTR was 28.86, while it was 25.61 under situations involving various other conditions. Overall, TUB was the most abundant gene, and ACTR was the least expressed gene.

geNorm Analysis
geNorm evaluated the stability of each of the eight possible internal benchmark genes using the M value. The smaller the value of M, the more reliable the expression of the gene [10,19]. All eight housekeeping genes had M values lower than 0.15 in each setting, as shown in Figure 2, and they all varied in their levels of stability between settings. The fittest genes for different sexes and diets are ACTR and ACT, which have the same M

geNorm Analysis
geNorm evaluated the stability of each of the eight possible internal benchmark genes using the M value. The smaller the value of M, the more reliable the expression of the gene [10,19]. All eight housekeeping genes had M values lower than 0.15 in each setting, as shown in Figure 2, and they all varied in their levels of stability between settings. The fittest genes for different sexes and diets are ACTR and ACT, which have the same M value. The housekeeping gene, 18S rRNA, demonstrated excellent stability under temperature and pesticide conditions. In sex and somite, GAPDH was the most unreliable housekeeping gene. geNorm software also gives data on the perfect amount of benchmark genes to be tested based on the pair-wise variance between ranking genes (V n/n + 1 ). Typically, V n/n + 1 is utilized to decide whether more housekeeping genes are required [10,19]. In V n/n + 1 > 0.15, case n + 1 housekeeping genes must be utilized. Conversely, just n housekeeping genes are necessary [10,19]. Figure 2 indicated, to properly normalize these treated samples, that only the two housekeeping genes were required, as evidenced by the V 2/3 values for the sex (0.042), somite (0.037), temperature (0.058), diet (0.079), and pesticide (0.045) samples being less than 0.15. The use of two benchmark genes is preferred in gene quantification research. As shown in Figure 3, geNorm analysis was utilized to identify the fittest benchmark gene pairs under different settings. Results revealed that ACT+ACTR exhibited the best stability under various sex conditions, while RPS30+28S rRNA demonstrated superior performance in somite-related analyses. The two genes with the most excellent stability under diverse for the sex (0.042), somite (0.037), temperature (0.058), diet (0.079), and pesticide (0.045) samples being less than 0.15. The use of two benchmark genes is preferred in gene quantification research. As shown in Figure 3, geNorm analysis was utilized to identify the fittest benchmark gene pairs under different settings. Results revealed that ACT+ACTR exhibited the best stability under various sex conditions, while RPS30+28S rRNA demonstrated superior performance in somite-related analyses. The two genes with the most excellent stability under diverse temperature settings were ACTR+18S rRNA. Meanwhile, ACT+ACTR was the most reliable housekeeping gene across different diet settings. Lastly, GAPDH+18S rRNA showed the most excellent stability under various pesticide conditions.

Comparative ∆Ct Analysis
In this approach, gene expression stability is evaluated by calculating each gene's mean and standard deviation (SD) value. Comparative ∆Ct analysis revealed that GAPDH had the most unstable expression across different sex and somite conditions. The ACT was the best housekeeping gene for gene normalization between sex and various dietary conditions. The most reliable housekeeping gene was 28S rRNA for various somites, temperatures, and pesticide circumstances ( Figure 4 and Table 2).

Comparative ∆Ct Analysis
In this approach, gene expression stability is evaluated by calculating each gene's mean and standard deviation (SD) value. Comparative ∆Ct analysis revealed that GAPDH had the most unstable expression across different sex and somite conditions. The ACT was the best housekeeping gene for gene normalization between sex and various dietary conditions. The most reliable housekeeping gene was 28S rRNA for various somites, temperatures, and pesticide circumstances (Figure 4 and Table 2).

NormFinder Analysis
NormFinder software directly assesses the reliability of internal benchmark genes based on intra-and inter-group differences, with lower values indicating more excellent stability [20]. Figure 4 and Table 2 display the steadiness of the E values of benchmark genes under each treatment. Results showed that RPL18 was the most trustworthy benchmark gene across different somites, while ACT exhibited superior performance in sex and diet-related analyses. Moreover, RPS30 was optimal under varying temperature conditions. Lastly, GAPDH was identified as the most applicable under multiple pesticide environments.

BestKeeper Analysis
By measuring the standard deviation (SD), coefficient of variation (CV), Pearson correlation coefficient (CC), and p (probability value) of the Ct values, BestKeeper evaluated the steadfastness of gene expression. Less SD and CV indicate a better level of gene expression. The gene was deemed unacceptable for the benchmark genes when SD > 1 or p > 0.05 [21]. The analyses' findings are presented in Table 3 and Figure 4. The top-ranked gene under different sex conditions was GAPDH, but its p value was higher than 0.05, disqualifying it from internally serving as a reference gene. RPS30, RPL18, 18S rRNA, and TUB were four other genes whose SD values were higher than 1, disqualifying them from being used as benchmark genes. Finally, an evaluation of stable internal benchmark genes under various sexes was ACTR > 28S rRNA > ACT. The eight genes under different somite's SD and p values complied with the reference genes' norms, and a stability ranking was GAPDH > ACT > RPL18 > 28S rRNA > TUB > RPS30 > ACTR > 18S rRNA. Even though ACT was the most precise internal benchmark gene, under different temperature states, its p value was higher than 0.05, which was unsuitable as a benchmark gene. Meanwhile, the SD value of RPL18 was more than 1, which was also problematic as an internal benchmark gene. The remaining six's stability order internal reference genes were GAPDH > 28S rRNA > RPS30 > TUB > ACTR > 18S rRNA. Under various diet conditions, only the SD and p values of RPS30 and TUB met the requirements. They were suitable for the internal reference genes, having a stability score of RPS30 > TUB. The p values of GAPDH, ACT, RPL18, 28S rRNA, ACTR, and 18S rRNA were all greater than 0.05 and did not meet the requirements of the benchmark genes. The SD value of ACT was more than 1, which did not fulfil the standards of the reference genes under various pesticide circumstances. The remaining seven reference genes' SD and p values were acceptable, and RPL18 > ACTR > 28S rRNA > RPS30 > GAPDH > 18S rRNA > TUB.

Comprehensive Ranking of Reference Genes
Using the online tool RefFinder (https://blooge.cn/RefFinder/type=reference, accessed on 18 January 2023), the combined stability ranking of the benchmark genes was determined to lessen the effects of a single algorithm's limitation. The geometric mean used to rank genes was calculated, and the stability increased as the geometric mean decreased [28]. Figure 5 displays that the two most trustworthy housekeeping genes for various sex conditions were ACT and ACTR; 28S rRNA and RPL18 were thought to be the best combination under different somites; 28S rRNA and RPS30 were the most appropriate housekeeping genes under various temperature circumstances; the two housekeeping genes with the highest levels of stability were ACT and GAPDH under diverse dietary circumstances; and GAPDH and 28S rRNA were the most trustworthy housekeeping genes under diverse pesticides. Under most circumstances, TUB is the housekeeping gene that is the most unreliable.

Verification of Reference Genes
The heat shock protein 90 of L. invasa was utilized as the objective gene to attest stability of benchmark genes. In this study, two of the most reliable housekeeping genes (28S rRNA and RPS30) and two of the least reliable reference genes (RPL18 and TUB) were chosen to be evaluated under diverse temperature states (25 °C was used as a control). The gene expression trends consistently used 28S rRNA and RRS30 as housekeeping genes in Figure 6. Specifically, the expression of HSP90 at 0 °C was lower than that at 25 °C, and at the same time, that at 40 °C was significantly higher than that at 25 °C and 0 °C. Overall, gene expression results were consistent when both 28S rRNA and RPS30 were used individually or combined as housekeeping genes. The expression of HSP90 at 0 °C was greater than that at 25 °C when we used RPL18 and TUB as housekeeping genes, in contradiction of the consequences obtained with 28S rRNA or RRS30 as housekeeping genes.

Verification of Reference Genes
The heat shock protein 90 of L. invasa was utilized as the objective gene to attest stability of benchmark genes. In this study, two of the most reliable housekeeping genes (28S rRNA and RPS30) and two of the least reliable reference genes (RPL18 and TUB) were chosen to be evaluated under diverse temperature states (25 • C was used as a control). The gene expression trends consistently used 28S rRNA and RRS30 as housekeeping genes in Figure 6. Specifically, the expression of HSP90 at 0 • C was lower than that at 25 • C, and at the same time, that at 40 • C was significantly higher than that at 25 • C and 0 • C. Overall, gene expression results were consistent when both 28S rRNA and RPS30 were used individually or combined as housekeeping genes. The expression of HSP90 at 0 • C was greater than that at 25 • C when we used RPL18 and TUB as housekeeping genes, in contradiction of the consequences obtained with 28S rRNA or RRS30 as housekeeping genes.

Discussion
In this study, the expression steadiness of eight benchmark genes was evaluated under different sexes, adult somites, temperature treatments, different dietary conditions, and pesticide treatments. According to this finding, no benchmark gene was suitable for each condition. Using a gene consistent across all experimental conditions as a control is much better than using a gene previously found to be highly consistent under only a very limited number of conditions, because there may be a nuisance variable that makes a gene that looks like a good fit in a given situation turn out to be a bad control. However, it is impossible to accurately predict any gene's expression under a specific set of conditions. Therefore, to select an appropriate reference gene for an RT-qPCR experiment, the recommended strategy is to choose several candidate genes and evaluate their expression levels in various experimental conditions and treatments. This approach will identify those genes that exhibit the most stable expression levels under different test conditions and thus serve as the most appropriate control genes in the experiment. However, this topic has yet to get much attention, and its significance does not seem adequately understood. Because the option of reference genes varies drastically, insects of the same species, in terms of insect morphology, developmental stage, temperature, sex, and diet conditions, suggest that there is also no absolute generality between benchmark genes for the same species [29][30][31][32]. For instance, PGK and RPL13 are acceptable internal housekeeping genes in Cnaphalocrocis medinalis under different sexes [31]. PGK and EF1α are stable housekeeping genes expressed in Cnaphalocrocis medinalis larvae under different temperature conditions [31]. Thus, for specific experimental treatments of L. invasa, it is indispensable to select appropriate housekeeping genes.
So far, many Hymenoptera species have found trustworthy housekeeping genes under different conditions, including Solenopsis invicta [32], Aphidius gifuensis [33], and Apis mellifera [34]. Nevertheless, reference genes for L. invasa have not been chosen or verified in earlier studies. L. invasa is an important pest of the genus Eucalyptus, mainly affecting seedlings and young forests. It forms galls on leaf veins, petioles, and current year branches, which in severe cases can lead to seedling mortality, up to 100% plant damage in young stands, and a significant reduction in yield in affected stands. Given future dispersal trends and changing environments, we should investigate molecular pathways for better management and control measures. The stability of eight regularly used internal

Discussion
In this study, the expression steadiness of eight benchmark genes was evaluated under different sexes, adult somites, temperature treatments, different dietary conditions, and pesticide treatments. According to this finding, no benchmark gene was suitable for each condition. Using a gene consistent across all experimental conditions as a control is much better than using a gene previously found to be highly consistent under only a very limited number of conditions, because there may be a nuisance variable that makes a gene that looks like a good fit in a given situation turn out to be a bad control. However, it is impossible to accurately predict any gene's expression under a specific set of conditions. Therefore, to select an appropriate reference gene for an RT-qPCR experiment, the recommended strategy is to choose several candidate genes and evaluate their expression levels in various experimental conditions and treatments. This approach will identify those genes that exhibit the most stable expression levels under different test conditions and thus serve as the most appropriate control genes in the experiment. However, this topic has yet to get much attention, and its significance does not seem adequately understood. Because the option of reference genes varies drastically, insects of the same species, in terms of insect morphology, developmental stage, temperature, sex, and diet conditions, suggest that there is also no absolute generality between benchmark genes for the same species [29][30][31][32] . For instance, PGK and RPL13 are acceptable internal housekeeping genes in Cnaphalocrocis medinalis under different sexes [31]. PGK and EF1α are stable housekeeping genes expressed in Cnaphalocrocis medinalis larvae under different temperature conditions [31]. Thus, for specific experimental treatments of L. invasa, it is indispensable to select appropriate housekeeping genes.
So far, many Hymenoptera species have found trustworthy housekeeping genes under different conditions, including Solenopsis invicta [32], Aphidius gifuensis [33], and Apis mellifera [34]. Nevertheless, reference genes for L. invasa have not been chosen or verified in earlier studies. L. invasa is an important pest of the genus Eucalyptus, mainly affecting seedlings and young forests. It forms galls on leaf veins, petioles, and current year branches, which in severe cases can lead to seedling mortality, up to 100% plant damage in young stands, and a significant reduction in yield in affected stands. Given future dispersal trends and changing environments, we should investigate molecular pathways for better management and control measures. The stability of eight regularly used internal housekeeping genes is explored in this study using various algorithms under five different experimental circumstances.
The commonly used statistical-analysis-based algorithms for evaluating the suitability of internal benchmark genes include geNorm, NormFinder, BestKeeper, and comparative ∆Ct. The rankings generated in this research by geNorm, NormFinder, and comparative ∆Ct are more like one another and different from the orders obtained by BestKeeper. For instance, ACT was the fittest in different sexes according to the results of geNorm, NormFinder, and comparing ∆Ct. However, according to BestKeeper analysis, ACT was the fourth trustworthy housekeeping gene in different sexes. In contrast, GAPDH took the top spot in the BestKeeper analysis. In Nitraria tangutorum [35], relative differences between BestKeeper and other programs have also been noted. These differences probably result from these systems' different algorithms [35]. RefFinder is a comprehensive evaluation tool that generates stability scores by estimating the geometric mean of internal control genes to lessen the effects of a single algorithm's limitations [36]. Many species, including Neoceratitis asiatica [18] and Anastatus japonicus [17], were studied using a similar strategy. The expression of HSP90 under different temperature treatments was evaluated to authenticate our findings. After normalization with 28S rRNA+RPS30 and RPL18+TUB, the expression results of HSP90 were different. This outcome indicates that it is essential to pick the appropriate benchmark gene to normalize the expression of the target gene. Some studies reported that accurate RT-qPCR results required two or more stable internal benchmark genes [37,38]. The geNorm algorithm can compute the perfect amount of internal control genes for standardization based on whether V n/n+1 is less than 0.15 [10,19]. Only two housekeeping genes were used as a benchmark for qRT-PCR in this research to increase the accuracy of the data.
Actin genes can encode the cytoskeleton and regulate the structural integrity of cells [38]. Actin genes include ACT8 and ACT11, which we have long used as housekeeping genes [39]. In fact, under various diapause states and in different tissues, Anastatus japonicus [17] displayed that ACTIN was the most reliable, consistent with earlier findings on Locusta migratoria [40] and Spodoptera litura [41]. In this study, ACT expresses steadily in most cases. Moreover, five algorithms assessed ACT as the perfect housekeeping gene under various diet circumstances. The validity of ACT, which was previously used in the study as a housekeeping gene for Hymenoptera, was further validated by this work. Still, our study found that the recommended amount of benchmark genes is two under diverse conditions. However, some studies on L. invasa used a single benchmark gene in a previous study. Therefore, we propose using two benchmark genes for normalization in future molecular experiments on L. invasa. Under varied sex and diet conditions, the most trustworthy housekeeping gene for L. invasa was ACT, but it was less reliable under different pesticides and somites. In the same species, a housekeeping gene may react differently to various conditions, like Kentucky bluegrass [42] and Klebsormidium nitens [43]. In conclusion, depending on the species, tissue, and treatment, it is frequently required to choose specific housekeeping genes.

Conclusions
As a result, the steadiness of the eight benchmark genes was tested using five trustworthy approaches in various experimental conditions. Regardless of which algorithm was used to assess the reference genes, ACT was most stable under different dietary conditions. After ranking the housekeeping genes' stability, geNorm was used to calculate V n/(n + 1). Two housekeeping genes were required as a benchmark for RT-qPCR to improve the trustworthiness of the qRT-PCR results. Five algorithms were combined to screen for the best combination of housekeeping genes under different conditions. These combinations included ACT and ACTR for different sexes, 28S rRNA and RPL18 for different somites, 28S rRNA and RPS30 for various temperature treatments, ACT and GAPDH for various diet treatments, GAPDH for various pesticide conditions, and 28S rRNA. This finding will improve the precision of target gene expression quantification and lay the foundation for the study of gene function and the molecular mechanisms involved in L. invasa resistance. Despite efforts to identify stable internal control genes for use in gene expression studies, it is important to note that there is no one-size-fits-all solution. The housekeeping genes recommended in this study demonstrate high stability and accuracy under specific experimental conditions, but their applicability cannot be assumed across all settings.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.