2.2. Mapping and Differential Expression Gene Analysis
Using the Bowtie program, nearly 70% of filtered reads for all four experimental conditions were aligned to the WCR reference transcriptome. The output (read counts) from alignment were applied directly in DESeq, and MA-plots were generated to depict a general view of the distribution of the differentially expressed transcripts (DETs) (p
< 0.05) for all three treatments as compared to the control treatment (Figure 1
). In general, transcripts with altered expression varied considerably with the source of Cry toxin challenge (Figure 1
b,c). Among the three treatments, the Cry34/35Ab1 treatment produced almost 10,000 sequences defined as DETs (Figure 1
c). In contrast, less than 1000 DETs were detected in the Cry34Ab1 treatment (Figure 1
b) and none were detected in the treatment with Cry35 alone (Figure 1
The initial read count datasets were further processed by filtering transcripts with low read counts (cpm > 1) with edgeR, 26,218; 29,109 and 29,520 transcripts were remained in the Cry35Ab1, Cry34Ab1 and Cry34/35Ab1 compared to control comparisons, respectively, and further used as input in differential expression analysis.
As illustrated in Table 1
, in total, 116, 132 and 135 DETs were identified with DESeq, edgeR and limma, respectively, in the Cry34Ab1treatment as compared to the control.
Among them, 114 DETs were commonly identified by these three methods. In the combined treatment with Cry34/35Ab1, 2215, 1673 and 2336 genes were classified as DETs by DESeq, edgeR and limma, respectively, and 1300 DETs were commonly identified, which is at least 10-fold greater than the number of DETs caused by exposure to Cry34Ab1 alone.
These 1300 DETs were used for subsequent analyses. As previously described, no DETs were detected by any of the three methods in the treatment with Cry35Ab1 alone.
Among the DETs identified in Cry34Ab1 (114) and Cry34/35 Ab1 (1300) treatment (Table 1
), 92 DETs (31 up-regulated and 61 down-regulated) were found to be in common (Table 2
). The remainder (22 and 1108 DETs) were assigned only to Cry34Ab1 or Cry34/35Ab1 treatment, respectively. Among the DETs common to both treatments, the average fold change was generally higher in the Cry34/35Ab1 treatment (Table 2
For the unique DETs in each treatment, the average fold was similar (2.36 for Cry34Ab1 and 2.69 for Cry34/35Ab1 up-regulated genes and 3.16 and 3.3 for down-regulated genes). However, the range of fold change of DETs was greater in the Cry34/35Ab1 treatment (~6.46-fold) as compared to that in Cry34Ab treatment (3.57-fold) in down-regulated category.
Annotation of DETs
Nearly 70% of DETs in Cry34Ab1 treated WCR (114) and in Cry34/35Ab1 treated WCR (1300) treatment had no significant hits in the NCBI non-redundant (nr) database (Figure 2
a,b). Among the DETs with hits, 35 DETs in the Cry34Ab1 treatments and 385 DETs in Cry34/35Ab1 treatments were well annotated with most hits to other Coleopterans including Tribolium castaneum
and Dendroctonus ponderosae
Gene ontology classification of the DETs is provided in Figure 3
(Cry34Ab1 compared to control) and Figure 4
(Cry34/35Ab1 compared to control).
For the Cry34Ab1 treatment, the largest number of DETs was assigned to molecular function (17 transcripts) and cellular component (17 transcripts). Of those assigned to molecular function, catalytic activity (11 transcripts, 64.7%) and binding (6 transcripts, 35.3%) accounted for the largest number of genes assigned while all 7 transcripts represented in biological process were associated with metabolic process. In contrast, the largest group of DETs for the Cry34/35Ab1 treatment was associated with biological process (1007 transcripts), in which metabolic process (217 transcripts, 21.5%), cellular process (175 transcripts, 17.4%), and single-organism process (159 transcripts, 15.8%) accounted for the largest categories. The second largest group involved molecular function (531 transcripts), in which binding (196 transcripts, 10.7%) and catalytic activity (292 transcripts, 16%) were most abundant.
In addition to the differences in number of transcripts that responded to Cry34/35Ab1 as compared to Cry34Ab1 alone, there were 18 more additional functional activities associated with these differences in the Cry34/35Ab1 treatment (Table S3
). Activities included antioxidant activity, molecular transducer activity, enzyme regulatory activity, receptor activity, transporter activity and membrane-enclosed lumen.
Transcripts coding for putative Bt toxin receptors identified from other insect species, such as cadherin, aminopeptidase N and ATP-binding cassette transporter (ABC) and metalloprotease were not detected with either Cry34Ab1 or Cry34/35Ab exposure. However, two different alkaline phosphatases (Dv_137932_c0_seq1 and Dv_149197_c0_seq1), which have been associated with Bt toxin binding in Lepidoptera, were identified as differentially expressed. The Dv_ 137932_c0_seq1 was down-regulated 7.46- and 2-fold in both Cry34/35 and Cry34Ab1 treatment, respectively. The Dv_149197_c0_seq1was up-regulated 1.65- and 1.2-fold in Cry34Ab1 and Cry34/35Ab1 treatment, respectively, based on DESeq results only.
2.3. GO-Term Enrichment and Pathway Analysis
As illustrated in Table S4
, five significantly overrepresented GO terms were associated with DETs in Cry34Ab1 challenge. Four of these were related to molecular function, in which two GO terms were up-regulated and correlated with zinc ion binding (GO:0008270) and transition metal ion binding (GO:0046914). The other two were down-regulated and were related to hydrolase activity (GO:0004553 and GO:0016798). The remaining GO term (GO:0005975) was down-regulated and associated with carbohydrate metabolism.
In the Cry34/35Ab1 treatment, a total of 168 GO terms were significantly enriched (Tables S5 and S6
). Among them, 152 GO terms (Table S5
) were up-regulated and 16 (Table S6
) were down-regulated. Of the up-regulated GO terms, 35 (23%) were associated with molecular function and mostly related to binding, such as ATP binding (GO:0005542), cation binding (GO:0043169), anion binding (GO:0043168), ion binding (GO:0043167) and carbohydrate derivative binding (GO:0097367). The remaining 117 GO terms were identified as biological processes and the GO terms associated with regulation accounted for the largest group (23 GO terms, 19.6%), including regulation of signaling (GO:0023051), regulation of Ras protein signal transduction (GO:0046578), regulation of hydrolase activity (GO:0051336) and regulation of lipid catabolic processes. Interestingly, only two GO terms were related to cellular process (GO:0009987) and cellular metabolic process (GO:0044237) (Table S6
) were found to be under-represented as down-regulated transcripts in the Cry34/35Ab1 treatment. Moreover, no cellular component-related GO term was found to be enriched in either the Cry34Ab1 or Cry34/35Ab1 treatment.
Due to the limited annotation of the reference transcriptome, KEGG analysis was also conducted to identify pathways associated with DETs to help identify higher-level functions. For the Cry34Ab1 treatment, 7 out of 114 (6%) DETs were assigned in two pathways: (1) amino sugar and nucleotide sugar metabolism and (2) starch and sucrose metabolism. In contrast, almost 323 out of 1300 (25%) DETs associated with the Cry34/35Ab1 treatment were associated with 42 different pathways (Table S7
). Among them, the majority (10 in up-regulated DETs and 22 in down-regulated DETs) of the identified were related to “metabolic pathway”. The top two pathways with most designated DETs were pyrimidine metabolism with 7 DETs and glycan degradation with 6 DETs. In addition, 6 pathways were with 5 DEGs, 9 pathways with 4 DETs, 4 pathways with 3 DETs, 9 pathways with 2 DETs assigned and the rest 28 pathways with 1 DET assigned only. In addition, among these 42 pathways identified, two pathways related to detoxification, drug metabolism-cytochrome P450 and glutathione metabolism assigned with a transcript (Dv_138610_c0_seq1) were included with a 1.61-fold change in the down-regulated category in Cry34/34Ab1 treatment.