Expression Profiling and Regulation of Genes Related to Silkworm

ARTICLE pubs.acs.org/jpr

Expression Profiling and Regulation of Genes Related to Silkworm Posterior Silk Gland Development and Fibroin Synthesis Jian-ying Li,†,‡ Hui-juan Yang,† Tian-yun Lan,† Hao Wei,† Hua-rong Zhang,§ Ming Chen,|| Wei Fan,† Ying-ying Ma,§ and Bo-xiong Zhong*,† College of Animal Sciences and §Zhejiang California International NanoSystems Institute (ZCNI), Zhejiang University, Hangzhou 310029, P.R. China ‡ Institute of Developmental and Regenerative Biology, Hangzhou Normal University, Hangzhou 310036, P.R. China College of Life Sciences, Zhejiang University, Hangzhou 310058, P.R. China

)

†

bS Supporting Information ABSTRACT: The posterior silk gland (PSG) is the most important suborgan responsible for the synthesis and secretion of silk core fibroin proteins in silkworm. Here, we performed genome-scale expression profiling analysis of silkworm PSG at the fourth molting (M4) and at day 1 (V1), day 3 (V3), day 5 (V5), and wandering stage (W) of the fifth instar by microarray analysis with 22 987 probes. We found that the five genes of silk proteins secreted from PSG including fibroin heavy (H) and light (L) chains, P25, seroin 1, and seroin 2 basically showed obvious up-regulation at V3 which lasted to V5, while slight down-regulation at W. The expression of translation-related genes including ribosomal proteins and translation initiation factors generally remained stable from M4 to V5, whereas it showed clear down-regulation at W. Clustering analysis of the 643 significantly differentially expressed transcripts revealed that 43 of the important genes including seroin 1 and sugar transporter protein had co-expression patterns which were consistent with the rate changes of fibroin synthesis and PSG growth. Pathway analysis disclosed that the genes in different clusters might have coregulations and direct interactions. These genes were supposed to be involved in the fibroin synthesis and secretion. The differential expression of several hormone-related genes also suggested their functions on the regulation of PSG development and fibroin synthesis. 2D gel-based proteomics and phosphoproteomics profiling revealed that the phosphorylated proteins accounted for no more than one-sixth of the total proteins at each stage, which was much lower than the level in normal eukaryotic cells. Changes in the phosphorylation status and levels of several proteins such as actin-depolymerizing factor 1 and enolase might be deeply involved in fibroin secretion and tissue development. Shotgun proteomic profiling combined with label-free quantification analysis on the PSG at V3, V5, and W revealed that many small heat shock proteins (sHSP) were specially expressed at W, which was substantially consistent with the results from 2-DE analysis, and implied the close correlations of sHSP with the physiological states of PSG at W. A majority of significantly up-regulated proteins at V5 were related to ribosome pathway, which was different from the microarray results, implying that the translation-level regulation of ribosomal proteins might be critical for fibroin synthesis. In contrast, the ubiquitin-proteasome pathway related proteins appeared obviously up-regulated at W, suggesting that the programmed cell death process of PSG cells might be started before cocooning. KEYWORDS: Bombyx mori, posterior silk gland, fibroin synthesis, development, microarray, proteomics, phosphoproteomics

’ INTRODUCTION The silkworm Bombyx mori is a holometabolous lepidopteran insect that has been domesticated for silk production for about 5000 years.1 It has now become the most well-studied lepidopteran model system for biochemical, molecular genetic, developmental, and genomic studies because of its detailed linkage maps and rich repertoire of well-characterized mutations affecting morphology, development, and behavior.1,2 The most attractive aspect of the silkworm, however, is the functional adaptation of its silk gland for silk protein synthesis and secretion. This gland is one of the most efficient protein synthesis systems among all r 2011 American Chemical Society

organisms. Its amazingly efficient protein synthesis makes the silk gland a desirable object for basic studies on gene expression and regulation and for biotechnological applications.3,4 The silkworm silk gland, which is homologous with the Drosophila salivary gland,5 can be anatomically and physiologically divided into three distinct compartments: posterior silk gland (PSG), middle silk gland (MSG), and anterior silk gland (ASG). The PSG is the longest suborgan and is responsible for Received: March 2, 2011 Published: June 09, 2011 3551

dx.doi.org/10.1021/pr200196x | J. Proteome Res. 2011, 10, 3551–3564

Journal of Proteome Research the synthesis of the silk core protein fibroin, which is composed of heavy (H) and light (L) chains plus P25, also named fibrohexamerin (fhx).6,7 The fibroin secreted exclusively from the PSG of silkworm larva moves to the MSG where heterogeneous sericin proteins are enveloped and then toward the ASG where the silk fiber is formed and spun. During silkworm larval life, the extensive growth of the silk gland and the increase in its secretory potential are achieved by about 500 PSG cells. At the early stages of the fifth (final) instar, the cellular structures necessary for the synthesis of fibroin are rapidly formed, and at the later stage, the synthesis of fibroin proceeds at a maximum rate.8 The third day of the fifth instar (V3) is the boundary for silkworm larval development because then rapid cell growth occurs in the silk gland and it begins to synthesize large amounts of silk proteins.9 The fifth day of the fifth instar (V5) is a specific stage when the silk gland size, as well as the amount of secreted fibroin into the lumen, reaches a climax, and two days later the wandering stage (W) is another special stage when the larvae stop feeding and begin wandering to search for a suitable place for cocooning. Interestingly, the expression of the fibroin genes is essentially controlled at the level of transcription, and changes periodically at the mRNA level; it is repressed at the molting stages and derepressed at the feeding stages.7,10 Several protein factors that bind the transcriptional regulatory elements of silk genes have been detected and characterized in silkworm, including FMBP1, Bm Fkh, POUM1,11 SGF-1,12 FF1,13,14 BMFA, and SGFB.15,16 Furthermore, the development, as well as the function, of silk glands has been shown to be under hormonal control.17 The effect of 20-hydroxyecdysone (20E) on the L-chain and P25 of fibroin is dose-dependent and it regulates the transcription of both genes.18 During the larval-pupal metamorphosis, programmed cell death (PCD) in the silk gland is also triggered by a steroid hormone.19,20 The silk gland provides a unique model system for gene expression and regulation that has been actively studied with multiple biotechnologies. The global gene expression profile of the silkworm PSG has been characterized by microarray analysis.21 The proteomic profiles of PSGs from silkworms fed on different diets were compared to determine the effects of differential nutrient conditions on silk protein synthesis.22 Zhang et al. found several phosphorylated forms of fibroin L-chain and P25 by analyzing the silk gland proteome.23 Chen et al. detected heavy phosphorylation of the fibroin H-chain in silk cocoons.24 In the present study, we characterized the gene expression profiles of the PSG at the fourth molt (M4), the first day of the fifth instar (V1), V3, V5, and W using DNA microarrays. The corresponding proteomic profiles were also characterized by twodimensional gel electrophoresis (2-DE) combined with MALDITOF/TOF analysis. Considering the importance of protein phosphorylation in the regulation of signaling pathways and various cellar processes,25 2D gel-based phosphoproteomic patterns were visualized with Pro-Q Diamond phosphoprotein dye technology.26 Furthermore, the total PSG proteomes at V3, V5, and W were analyzed with shotgun LCMS/MS. These results provide the first comprehensive view of the molecular basis of PSG development and the synthesis and secretion of fibroin at the transcriptional, translational, and post-translational modification (PTM) levels.

’ MATERIALS AND METHODS Insect Rearing and Dissection

Silkworm strain p50 was reared on fresh mulberry leaves under standard conditions. The developmental stages were synchronized

ARTICLE

after the fourth molting by collecting newly molted larvae. The PSG was dissected at five developmental stages: M4, V1, V3, V5, and W. Three groups of biological repeats were prepared at each time point. For proteomic analysis, the PSG had to be separated from fibroin proteins in the gland lumen at V3, V5, and W when large amounts of proteins were secreted. To do this, the PSG was immersed in prechilled 60% ethanol for 1 min to denature the fibroin, which was then drawn out from the PSG lumen with nippers. RNA Extraction and Microarray Hybridization

Total RNA was isolated from each sample using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The microarray hybridization procedure was performed according to a previous report with some modifications.21 Dual-dye experiments were performed for the PSG at M4, V1, V3, V5, and W; a total of 15 samples were tested including three biological replicates at each stage. Aliquots of each test sample were pooled as a common reference sample (CKA). The test samples and CKA were labeled with Cy3 and Cy5, respectively. The resulting labeled cDNAs were then hybridized with 22 987 previously designed 70-mer oligonucleotide probes (CapitalBio Corporation, Beijing, China) representing all the presently known silkworm genes and approximately 85% of the newly predicted ones (14 623 entries).2,21 Microarray Data Analysis

The prevalent LOWESS normalization method was applied for the signal intensity data.27 The extracted signal data was filtered by removing faint spots, saturated spots (intensity >60 000 units), and signal intensities below 800 units after subtracting the background from both channels (Cy3 and Cy5). Biological replicate results for each stage were compared and combined to check that actively expressed genes had at most one missing value (because of discarded data that did not pass the intensity thresholds). Genes expressed differentially in two different stages were defined as those for which the average ratio had a 2-fold change with unpaired t-test significance P < 0.05. Hierarchical clustering (HC) analysis of the expressed genes at the five stages was performed using Cluster 3.0 software with the average linkage method and viewed with TreeView version 1.0.7. Differentially expressed genes with statistically significant changes (t test) were screened by using SAM (Significance Analysis of Microarrays) (version 2.1) with one class time course analysis.28 The differentially expressed genes were classed by SOM (Self-organizing Map Clustering) with GeneCluster 2.0 software. Real-Time Quantitative RT-PCR

Real-time qRT-PCR was carried out in 20-μL reactions including 1/20 volume of reverse-transcribed product, 2 units of Taq DNA polymerase, 0.6 μL of 20 EvaGreen dye (Biotium, Inc., Oakland, CA), and 0.3 μM each of the forward and reverse primers (Supplemental Table S2). The running program began at 95 °C for 10 min for activation, followed by 40 cycles of amplification with 95 °C for 15 s, 56 °C for 15 s, and 72 °C for 20 s. The gene expression level was calculated based on the delta Ct value normalized with the Rpl27a gene (accession no. NM_001044057). 2D Gel-Based Proteome and Phosphoproteome Analysis

The PSG protein extraction and separation with 2-DE followed by MALDI-TOF/TOF analysis were performed as we described previously.22 In brief, 250 μg of sample proteins was first loaded onto a 24-cm Immobiline DryStrip (pH 310, 3552

dx.doi.org/10.1021/pr200196x |J. Proteome Res. 2011, 10, 3551–3564

Journal of Proteome Research Linear) (Amersham Biosciences, Piscataway, NJ) for isoelectric focusing (IEF) and then transferred to a 12.5% polyacrylamide gel for SDS-PAGE separation. Phosphoproteins were detected by fluorescent staining with Pro-Q Diamond (Molecular Probes, Eugene, OR) according to the manufacturer’s specifications. The images were scanned with a Typhoon 9200 Imager (Amersham Biosciences) with excitation at 532 nm and emission at 580 nm. After washing, the total proteins were visualized by silver staining. Spot detection, matching, and quantitative intensity analysis were performed automatically with ImageMaster 2D software (version 6.0). Protein spots of interest were manually excised from the silverstained gels and subjected to in-gel digestion.22 The mass spectrometry (MS) spectra of digested peptides were acquired with a 4700 MALDI-TOF/TOF Proteomics Analyzer (Applied Biosystems, Foster City, CA) and recorded in reflector mode using 4700 Explore software (Applied Biosystems). Combined MS and MS/MS spectra were subjected to MASCOT (Version 2.2, Matrix Science, London, U.K.) by GPS Explorer software (version 3.6, Applied Biosystems) and searched against the NCBInr database (release date, Nov 20, 2008) with the following parameters: taxonomy of all entries, trypsin digest with one missing cleavage, MS and MS/MS tolerance of 100 ppm, fixed modifications of carbamidomethyl (C), and variable modifications of oxidation (M). The search yielded proteins with MASCOT scores more than 80 (P < 0.05), and individual MS/MS spectra with high confidence intervals (C.I. % >95%) were regarded as positive identifications. Shotgun Nano-LCMS/MS Analysis

PSG proteins at V3, V5, and W were extracted and separated followed by in-gel digestion according to our previously described methods.29 A total of 200 μg of protein was divided into 12 sections for each sample. Five microliters of digested peptide sample was subjected to nano-LCMS/MS using an Ettan MDLC nanoflow/capillary LC system (GE Healthcare, Pittsburgh, PA) coupled to a linear ion trap Orbitrap (LTQOrbitrap) mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). After preconcentration and washing of the sample on a C18 trap column (PepMap C18, 300-μm i.d. 5 mm, 3 μm, 100 Å (P/N 160454), Sunnyvale, CA), peptides were separated on an analytical column (PepMap C18, 75-μm i.d. 15 cm, 3 μm, 100 Å (P/N 160321), Sunnyvale, CA) using a 90 min gradient of buffer B (84% acetonitrile, 0.1% methanoic acid in water) at a 300 nL/min flow rate. The LTQ-Orbitrap with a nanospray configuration was operated in data dependent acquisition mode with XCalibur software version 2.0 (Thermo Electron, San Jose, CA). Collision-induced dissociation (CID) was conducted with an isolation width of 2 Da, normalized collision energy of 35%, and activation q of 0.25 for MS/MS acquisition. The five most intense ions were isolated for CID fragmentation and measured in the linear ion trap with the dynamic exclusion settings: repeat count 2, repeat duration 30 s, exclusion duration 180 s. Triplicate replicates were performed for each sample. Database Searching

An in-house database was constructed with protein sequences of B. mori downloaded from NCBI Refseq (release date: Jan 20, 2010; 1739 entries) in combination with the newly released genome coding sequences (14 623 entries).2 A composite targetdecoy database was made up with the forward and reverse sequences for the query of LCMS/MS data.30 MS/MS spectra

ARTICLE

were extracted from the raw data files by Mascot Deamon (Matrix Science, London, U.K.; 2.2) with the extract_msn program (version 2.07). The resultant .mgf files were searched against the in-house database on a local Mascot server (Matrix Science; 2.2). Parent and fragment ion mass tolerances were 50 ppm and 0.6 Da, respectively. Trypsin was used to cleave the peptides and two missing cleavage sites were allowed. A fixed (carbamidomethyl) modification of cysteine and variable modifications of oxidation (M), Phospho (ST), and Phospho (Y) were specified. The resultant .dat files were subjected to TransProteomic Pipeline (v4.0 JETSTREAM rev 2) for further process using PeptideProphet and ProteinProphet with the probability thresholds at 0.7 and 0.9, respectively.31 The protein lists were manually checked to reduce redundancy. For protein groups with shared peptides mapping to multiple proteins, a single protein from the protein group was selected unless more than one protein was also identified in the other samples. Label-Free Quantification

The relative expression levels of the proteins identified by LCMS/MS were evaluated with a modified spectral counting technique, Absolute Protein Expression (APEX).32 The APEX score of each protein was computed based on its Oi and the normalization factor (C), which was arbitrarily set to 100 000. The relative abundance of a protein was assessed on the basis of its APEX score. Bioinformatics Analysis

Gene Ontology (GO) terms for the proteins we identified were retrieved by searching against InterPro member databases using InterProScan software. Pathway Studio 7.0 software (Ariadne Genomics, Inc., Rockville, MD) was used to interpret, build, and visualize biological networks for our identifications based on published literature from the Drosophila and mammalian databases developed by Ariadne Genomics.

’ RESULTS Global Gene Expression Profiles

The normalized signal intensity ratios (Cy3/Cy5) of validated probes were used for the following analysis. After removing the low-confidence signals that did not pass the intensity threshold, we identified 4365, 4246, 3459, 3169, and 2772 active transcripts at M4, V1, V3, V5, and W, respectively (Supplemental Table S1). Collectively, 4880 transcripts were validated including 2471 (excluding the transcripts with saturated expression) expressed throughout the five investigated stages. Although there were many genes represented by multiple probes, the intensity ratios were perfectly consistent for the overwhelming majority of genes. To evaluate the quality of the microarray data, 17 genes were selected randomly for real-time qRT-PCR analysis. The correlation coefficient (R2) of the two data sets was 0.82, indicating the high reliability of our results (Supplemental Figure S1; Supplemental Table S2). We also compared our genes identified at V3 with previously published data,21 and 83.2% of the 3459 transcripts were supported regardless of different intensity thresholds. Out of the 4880 transcripts, 2746 with 80% intensity ratio values during the five stages were subjected to HC analysis and TreeView, which displayed the gene expression profiles across the developmental stages (Figure 1). The HC results showed that the PSG gene expression levels of the three replicates at each time point were well clustered. M4 and V1 showed the greatest differential compared to the other stages, whereas 3553


Journal of Proteome Research

ARTICLE

Figure 1. Clustering analysis of PSG genes detected during different developmental stages of silkworm larvae. Hierarchical clustering of the 2746 out of 4880 transcripts with 80% intensity ratio values during the five stages. The three repeats of genes at each stage were well clustered.

Figure 2. Expression patterns of genes clustered in different groups. (A) The profiles of characteristically expressed genes during the five stages. The standard for the stage-specific genes are defined in the text; (B) SOM analysis of the differentially expressed transcripts across time courses. The transcripts listed in Supplemental Table S5 were clustered into 9 centroids (c0c8) based on co-expression patterns by SOM using GeneCluster 2.0 software. The X-axis for each centroid consists of the five time points from M4 to W. The Y-axis indicates expression level (fold-change). The number in the top center of each centroid indicates the number of involved transcripts. The top line indicates the upper boundary of expression for all of these genes and the lower line indicates the lower boundary. The middle line is the mean value; (C) the expression patterns of the five silk genes; (D) the expression patterns of the ribosomal protein genes detected in this study; (E) the expression patterns of translation initiation factor genes during the five stages. The average values of expression ratios in panels CE were log2-transformed to represent the gene expression levels.

V3 and V5, when the larvae are growing rapidly, shared the most similar patterns. The greatest changes of gene expression occurred at V3, which seems consistent with the variation of the larval growth rate. These differences can also be seen directly in the Volcano Plot of differentially expressed genes across the stages (Supplemental Figure S2). Compared with the expression levels of PSG genes at M4 (using unpaired t-tests), 105, 178, 190, and 272 transcripts were

up-regulated (ratio g 2, P < 0.05), while 146, 255, 327, and 543 were down-regulated (ratio e 0.5, P < 0.05) at V1, V3, V5, and W, respectively (Supplemental Table S3). Stage-Specific Gene Expression

From M4 to W, the silkworm experiences a serial of physiological changes, especially at M4 and W when the larvae are in 3554


Journal of Proteome Research specialized physiological states. To uncover their internal relationships with gene expression, we examined the expression patterns of stage-specific genes. In the present study, we defined these as genes for which the intensity ratios of three repeats at one stage were all over 2.0 and their average value was at least 2 times higher than that of the other stages. This arbitrary cutoff would unavoidably lose some specific genes for which one or two of the three repeats did not pass the threshold due to the variations of experiment. Nevertheless, it would not change the distribution of stage-specific genes, even those missing genes were considered. According to our thresholds, we found 82, 9, 2, 6, and 137 specific transcripts at the five respective stages (Figure 2A; Supplemental Table S4). It was obvious that there were many more stage-specific genes at M4 and W than at the other stages. This might be closely associated with the specific physiological states of molting and spinning, whereas few specific genes were expressed at active-eating stages. Indeed, we found many genes related to the formation of cuticle at M4, including chitin synthase, chitin binding peritrophin-A, and six cuticular protein coding genes. Of these cuticular proteins, glycinerich 9, glycine-rich 14, and tweedle motif 2 are rich in glycine (16.1%/14.8%/26.2%), serine (16.6%/6.2%/8.7%), and alanine (7.5%/5.3%/6.8%), which are the primary materials for fibroin protein synthesis.33 Co-expression Patterns

To identify genes with similar expression patterns, 2471 transcripts with no more than 2 missing ratio values in the 15 arrays were subjected to SAM by the slope and signed area methods. A total of 643 transcripts with time-dependent patterns, SAM scores more than 2, and estimated q-values lower than 0.05% were selected. These transcripts were clustered into 9 co-expression categories (c0c8) (Figure 2B; Supplemental Table S5). The expression of genes in c0 and c1 was obviously higher at M4 and decreased through the other stages. Apart from the M4-specific genes, most of them were cuticle-related. By contrast, the genes in c5 and c8 increased their expression gradually or showed dramatic up-regulation at W. They included the ribosomal protein P1, arginine kinase, and MybMuvB complex subunit Lin-52. Overall, more genes decreased their expression levels during the development of silk gland cells. We also found that the general expression pattern of 43 transcripts in c2 was consistent with the rate changes of PSG growth and fibroin synthesis. Seroin 1, sugar transporter protein, and mitochondrial cytochrome c are included in this group. Interestingly, the expression of 40 genes in c7 showed a roughly opposite trend from the genes in c2; they were gradually down-regulated until V5 and then up-regulated at W. Fibroin Synthesis and Protein Translation

Silk protein secreted from the PSG is composed of H-chain, L-chain, and P25 together with seroin 1 and seroin 2, which are also expressed in the MSG.34,35 These genes showed generally increasing expression levels during PSG development, especially at V3 when four genes showed dramatic increase (Figure 2C). It has been proposed that the expression of fibroin genes is regulated at the transcriptional level. Among the known regulators of fibroin gene expression, SGF-1 (sw16198) was highly expressed only at M4 and V1 (the missing values at M4 were retrieved from raw data), whereas the expression of FMBP1 (sw10503) failed to reach the intensity threshold. These factors regulate the expression of fibroin genes. However, unexpectedly, we also found that fibroinase was expressed throughout the five

ARTICLE

Figure 3. Expression patterns of hormone-related genes detected in the investigated stages. The mean fold-change value of each transcription was used to represent its expression level.

stages, whereas silk proteinase inhibitor was expressed more highly at M4 and V1. Fibroin is rich in glycine (43%), alanine (28%), and serine (12%). Synthesizing large quantities of fibroin proteins requires a functional adaptation of the PSG in which the tRNA species become highly enriched for the predominant amino acids in fibroin. We found that glycyl-tRNA synthetase (sw03059) was highly expressed at V3 and V5 (which is the exact time when large amounts fibroin proteins are synthesized), but this decreased at W. Ribosomal proteins make up the ribosomal subunits that participate in the cellular process of translation. We detected 103 transcripts representing 95 ribosomal protein-coding genes expressed throughout the five stages at high levels, including 77 transcripts with signal intensities higher than 8000 units which are 10 times of the threshold (800 units). Their expression basically remained stable from M4 to V5, but it became disorderly at W (Figure 2D). There were 71 (68.9%) transcripts lower at W than at V5, including 36 that were down-regulated more than 2-fold. By contrast, the predicted mitochondrial ribosome recycling factor showed a slight increase. Similar patterns were also observed for translation initiation factors. Of the 21 transcripts representing 17 translation initiation factors that were expressed throughout the five stages, 6 genes were significantly down-regulated at W (P < 0.05) (Figure 2E). Developmental Regulation

It is well-known that the growth and development of the silkworm are regulated by juvenile hormone and ecdysteroid hormones. The development and function of silk glands has also been shown to be under hormonal control.17 We detected 14 hormone-related proteincoding genes including five expressed throughout the investigated stages. They are esr20 (ecdysteroid regulated protein), AKH1 (adipokinetic hormone 1 precursor), E75 (nuclear hormone receptor E75 isoform B), Jheh2 (juvenile hormone epoxide hydrolase), and LOC100166921 (PREDICTED: similar to juvenile hormone esterase) (Figure 3). Of these genes, esr20 was specifically expressed at M4, whereas the expression levels of Jheh2 and AKH1 were significantly higher at V1 and W, respectively. Dynamics of Protein Expression and Phosphorylation Levels

The 2-DE patterns of total proteins and phosphoproteins of the PSG from M4 to W are shown in Figure 4. On the whole, phosphorylated proteins accounted for no more than one-sixth of the total proteins at each stage. Most of the phosphoproteins were distributed between pI (isoelectric point) 36.5, and more were detected at the molting stage than the other stages. A total of 62 interesting protein spots were analyzed by MALDI 3555



ARTICLE

Figure 4. 2D gel-based proteomic and phosphoproteomic profiles of the PSG during the investigated stages. (A, C, E, G, and I) Total protein profiles visualized by silver staining at M4, V1, V3, V5, and W, respectively; (B, D, F, H, and J) phosphoprotein profiles as stained by Pro-Q Diamond dye at each of the five stages, respectively. The proteins successfully identified by MALDI-TOF/TOF are marked with serial ID numbers. IDs beginning with “P” represent the phosphoproteins. K, representative proteins with differential expression levels. L, comparison of representative protein spots visualized by both fluorescent dye and silver staining, as shown in the upper and lower rows, respectively.

TOF/TOF. Of these spots, 40 were tentatively identified, including 14 phosphorylated and 26 nonphosphorylated protein spots (Table 1). The fluorescence intensities of these spots (which indicate the phosphorylation levels of the proteins) were sometimes consistent with the protein expression pattern and sometimes converse to it (Figure 4L). For example, the fluorescent signal intensity of spot P1 gradually increased together with an increase in the amount of protein; by contrast, the protein abundance of spot P2 increased accompanied by a decrease in its phosphorylation level. MS analysis revealed that spots P1 and P2 represent the same protein, ribosomal protein P2. The different pI and molecular weight (MW) values distributed in the gel revealed that ribosomal protein P2 has different levels of PTMs. The actin-depolymerizing factor 1 (spot P3) was phosphorylated throughout the late larval stages except for M4. Another interesting protein is enolase (spot P11), the expression of which was much higher at the feeding and wandering stages than the molting stage; in contrast, its phosphorylation intensity was very strong at M4 and disappeared from V1 to W. Moreover, we found several highly expressed stage-specific proteins. For example, cathepsin B (spots 7 and 8) and ATP synthase (spot 12) were highly expressed at M4, whereas they

basically disappeared at other stages (Figure 4K). Two small heat shock proteins (sHSPs), hsp20.4 (spot 18) and hsp20.8 (spot 26), were expressed at very low levels from M4 to V5 but increased dramatically at W. In addition, another sHSP (hsp21.4) gradually increased its expression from M4 to V5 and reached a maximum at W. However, in contrast to the molting and wandering stages, at feeding stages the PSG expressed few specific proteins, which is consistent with the microarray results. Whole Proteome Expression Profiles

To gain further insight into the molecular basis of silk gland development and fibroin synthesis, the whole proteomes of PSG at V3, V5, and W were analyzed by shotgun LCMS/MS. Triplicate replicates were performed for each sample, and the results were combined together for the following quantitative analysis. After stringent filtering and manual selection, 1366, 1466, and 1280 proteins from V3, V5, and W, respectively, were successfully identified with a false discovery rate (FDR) lower than 0.3% (Supplemental Table S6). By reciprocal comparison, we found 825 proteins that were expressed throughout the three stages and 306, 205, and 126 stage-specific proteins. 3556


silk fibroin light chain

heat shock protein hsp21.4

thiol peroxiredoxin beta-tubulin

beta-tubulin

cathepsin B

cathepsin B

signal sequence receptor beta subunit

calreticulin

transgelin

ATP synthase receptor for activated protein kinase C RACK 1 isoform 1

2

3

4 5

6

7

8

9

10

11

12 13

gi|153792659 gi|112982743 gi|162952033 gi|112982743

gi|112984336 gi|153792659 gi|112982743

acetoacetyl-CoA thiolase

fructose 1,6-bisphosphate aldolase


short-chain dehydrogenease/reductase

adenosine kinase CDK-activating kinase assembly factor

actin-depolymerizing factor 1

elongation factor 1 beta0

ribosome-associated protein P40



ribosomal protein P2

ribosomal protein P2 actin-depolymerizing factor 1


ribosomal P0 protein

PREDICTED: similar to Actin-5C isoform 2

beta-tubulin

H+ transporting ATP synthase beta subunit isoform 1

15

16

3557

17

18

19

20 21

22

23

24

25

26

P1

P2 P3

P4

P5

P6

P7

P8

gi|114052072

gi|112983503

gi|91088367

gi|37359627

gi|112984336

gi|148298693

gi|114051596 gi|70942482

gi|114050771

gi|112983152

gi|148298685

gi|153791621

gi|114053313

26S protease regulatory subunit 8 (1856 protein)

GTP binding protein

14

gi|1709799

gi|114052278 gi|115345341

gi|114051357

gi|28804517

gi|114052941

gi|112983908

gi|112983908

gi|112983503

gi|112982996 gi|112983503

gi|154091278

gi|112982880 gi|112984494

translationally controlled tumor protein

1

no.

no.

protein name

accession

spot

5.26/55.0

4.83/50.3

5.36/37.9

5.69/34.1

4.49/24.5

4.68/11.5 6.17/17.0

4.68/11.5

5.98/20.8

4.49/24.5

4.87/33.4

4.49/24.5

6.17/17.0

5.14/38.1 8.65/29.5

8.62/27.1

6.54/20.4

8.38/39.7

7.56/39.6

7.08/44.4

8.51/45.3

9.21/59.6 8.07/36.0

8.42/20.9

4.49/45.7

6.9/20.9

5.95/37.5

5.95/37.5

4.83/50.3

6.09/21.9 4.83/50.3

5.8/21.4

5.23/27.7

4.66/19.8

(kDa)

pI/MW

cal.

447

360

379

296

178

187 200

171

178

105

96

107

243

409 94

83

85

132

156

94

84

188 155

190

184

387

132

105

114

208 142

293

368

191

score

protein

100

100

100

100

100

100 100

100

100

100

99.9

100

100

100 99.8

97

98.2

100

100

99.7

97.7

100 100

100

100

100

100

100

100

100 100

100

100

100

(%)

C.I.

21

21

13

8

7

3 5

4

6

5

4

4

9

10 6

4

3

6

7

5

7

7 10

10

10

5

6

4

7

6 7

7

6

3

count

pep.

Table 1. The Differentially Expressed Proteins and Phosphoproteins Identified by MALDI-TOF/TOFa

82.6

75.4

63.3

44.3

54.5

36.6 45.9

50

46.8

44.1

27.1

37.4

93.9

55.9 31.9

24

34.8

31.9

35.9

21.4

24.9

19 50.5

72.3

44.2

54.7

36.2

22.8

27.1

40 27.5

58.8

34.4

36

(%)

pep. cov.

0.55 (0.81)

0.21 (0.74)

0.92 (0.61)

0.34 (0.94)

0.18 (1.53)

0.21 (0.83) 0.10 (0.00)

0.28 (1.19)

0.13

0.04

0.17

0.12

0.24

0.08 0.15

0.10

0.29

0.19

0.14

0.11

0.13

0.41 0.21

0.06

0.71

0.55

0.46

0.43

0.16

0.22 0.16

0.22

0.27

0.38

M4

0.43 (0.71)

0.14 (1.15)

0.83 (0.58)

0.56 (1.57)

0.09 (1.75)

0.10 (0.27) 0.28 (0.61)

0.35 (1.77)

0.06

0.06

0.30

0.21

0.53

0.23 0.29

0.11

0.14

0.19

0.34

0.17

0.15

0.07 0.49

0.21

0.60

0.35

0.05

0.19

0.05

0.39 0.03

0.18

0.20

0.25

V1

0.36 (0.73)

0.02 (0.36)

1.12 (1.28)

0.88 (2.83)

0.18 (2.53)

0.03 (0.22) 0.38 (1.29)

0.42 (1.95)

0.10

0.04

0.19

0.17

0.52

0.00 0.31

0.10

0.12

0.00

0.00

0.07

0.16

0.02 0.35

0.20

0.60

0.58

0.04

0.00

0.10

0.30 0.02

0.20

1.61

0.41

V3

V5

0.53 (1.24)

0.11 (0.25)

1.02 (0.53)

0.95 (3.42)

0.29 (3.91)

0.02 (0.35) 0.13 (0.76)

0.48 (3.74)

0.12

0.00

0.18

0.41

0.35

0.02 0.22

0.04

0.09

0.00

0.00

0.00

0.22

0.00 0.34

0.04

0.47

0.96

0.09

0.00

0.27

0.16 0.07

0.15

0.60

0.69

intensity ratio (%)

0.89

0.80 (2.67)

0.10 (0.68)

1.02 (0.86)

1.00 (3.51)

0.64 (4.58)

0.08 (0.42) 0.16 (1.22)

0.68 (6.43)

1.02

0.22

0.46

0.19

0.25

0.03 0.19

0.07

0.63

0.02

0.00

0.00

0.05

0.00 0.25

0.00

0.49

0.85

0.00

0.00

0.32

0.18 0.00

0.00

0.93

W

Journal of Proteome Research ARTICLE


0.06 (0.00)

0.00 (0.00)

heterogeneous nuclear ribonucleoprotein A1 P14

The intensity ratio of the protein spots detected by 2-DE represent the relative protein expression levels or the phosphorylation levels. Spot IDs beginning with “P” represent the phosphoproteins whose protein intensity ratio values are shown in parentheses. a

0.07 (0.00) 0.13 (0.00) 0.40 (0.00) 0.35 (0.22) 26.9 5 182

glyceraldehyde-3-phosphate dehydrogenase P13

gi|153792009

6.55/35.0

100

0.00 (0.00) 0.00 (0.00)

0.00 (0.00) 0.00 (0.00)

0.22 (0.26) 0.69 (0.82)

0.08 (0.16) 0.28 (0.30)

0.33 (0.64) 36.7

19.3 5

9 99

8.31/35.4

100

88

146

9.31/38.5 lark P12

gi|112983816

heat shock cognate protein enolase P10 P11

gi|112983834

0.00 (0.49) 0.56 (0.00) 0.00 (0.47) 0.23 (0.00) 0.00 (0.60) 0.63 (0.00) 0.07 (0.60) 0.24 (0.00) 0.00 (0.71) 0.08 (1.88) 24 21.2 10 5 83 149 5.33/71.1 5.62/47.0

vacuolar ATP synthase subunit B

gi|112982828 gi|148298800

270

ARTICLE

P9

gi|148298717

5.25/54.4

97.1 100

W V5

0.06 (0.68) 0.26 (1.89)

V3 V1

0.13 (0.73) 0.07 (0.28)

M4 (%)

pep. cov.

45.3 13

count

pep. C.I.

score

(%)

protein

(kDa) no. no.

cal.

protein name

accession spot

pI/MW

100

intensity ratio (%)

Table 1. Continued

0.07 (0.71)


Interestingly, we found that two proteins involved in hormone regulation, the cytosolic juvenile hormone binding protein 36-kDa subunit and the ecdysone receptor (EcR), were specifically expressed at V3 and W, respectively. Furthermore, three sHSPs (hsp19.9, hsp20.1, and hsp20.8) were specifically expressed at W. The sHSPs highly expressed at W were also confirmed by the 2-DE profiles. The abundances of the identified proteins were evaluated using a label-free quantitation method, APEX,32 which is based on spectra counting. To identify differentially expressed proteins with significant 2-fold changes (P < 0.05), pairwise comparisons were performed among the three stages. In comparison with V3, there were 57 up-regulated and 154 down-regulated proteins at V5 (P < 0.05) (Figure 5; Supplemental Table S7). Compared with V5, there were 75 up-regulated and 40 down-regulated proteins at W (P < 0.05). Of the up-regulated proteins at V5, 22 (38.6%) were ribosomal proteins. Moreover, there were five transmembrane transport related proteins including tetraspanin E, transmembrane emp24 protein transport domain containing 9 (BGIBMGA010959-PA), translocon-associated protein gamma, H+ transporting ATP synthase O subunit, and transport protein Sec61 beta subunit. Most of the up-regulated proteins at V5 are engaged in protein synthesis and transportation. Although the five silk proteins as well as fibroinase were expressed throughout the three stages, their patterns were different than in the microarray data. Of the 100 ribosomal proteins, 71 were also detected by microarray analysis. A majority of these were highly expressed, and most of them were up-regulated at V5. GO category analysis showed that the up-regulated proteins at V5 with structural molecular activity, primarily the structural constituents of ribosomes (GO: 0003735), were significantly more abundant than that of down-regulated (P < 0.05) (Figure 6A). Transcriptional regulators were up-regulated while translation regulators and proteasome regulators were downregulated at V5. By contrast, translation regulators were upregulated at W (Figure 6B). Expression Regulatory Networks and Pathways

Co-expressed genes with similar patterns may have correlated pathways and share common functional tasks and regulatory mechanisms.36 We used Pathway Studio software with different algorithms to further uncover the connections of genes with co-expression patterns, and this revealed that they had complex immanent relationships. Many of them had direct interactions and several genes in different clusters had the same regulator (Supplemental Figure S3). For example, mirr in c2 and CG2263 in c4 are both under the regulation of epidermal growth factor receptor (egfr). Enrichment analysis of differential gene expression using expression subnetworks was performed by adding downstream neighbors to uncover regulatory networks.37 Thyroid hormone receptor (T3R) and klu inhibit the expression of egfr,38,39 whereas Jra can induce its expression (Figure 7).40 T3R and egfr both play crucial roles in the growth and development of organs, implying their importance in the development of the silkworm PSG. Moreover, the expression patters of T3R and Jra were similar. Jra regulates the expression of dm, and dm and luciferase have reciprocal regulation. Furthermore, multiple genes we identified including mtf-1, fk506-bp1 and thymidine kinase might regulate the expression of luciferase. Subnetwork enrichment analysis also helped to identify key upstream regulators. For example, the dynamic changes of general RNA polymerase II transcription factor might 3558



ARTICLE

Figure 5. Comparison of APEX values of the identified PSG proteins at V3, V5 and W. (A) V3 vs V5; (B) V5 vs W; (C) V3 vs W. The differential expressions of proteins between each pair were compared based on their log10-transformed APEX scores. The Pearson correlation coefficients for the three pairs are 0.87, 0.92, and 0.85, respectively. The blue squares and red spots represent proteins that are nonsignificantly and significantly differentially expressed (P < 0.05), respectively. Of the significantly differentially expressed proteins, the ones with changes lower than 2-folds were discarded in the following analysis.

Figure 6. The GO categories of the differentially expressed proteins at V3, V5, and W. The differentially expressed proteins at the three stages were classified into cellular component, molecular function, and biological process by WEGO according to the GO terms. The number of genes is the number of times the GO term is used to annotate genes in the cluster. The left-hand of each figure shows its proportion in total up- or down-regulated genes with GO terms. (A) Categories of the differentially expressed proteins of V5 compared with V3; (B) categories of the differentially expressed proteins of W compared with V5.

affect the expression of a series of downstream genes (Figure 7). The multifarious interactions of genes constitute complex expression regulatory networks. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of the up-regulated proteins revealed that the most-enriched proteins at V5 and W are involved in the ribosome and proteasome pathways, respectively. Conversely, the most-enriched proteins among the down-regulated proteins at V5 and W are involved in the proteasome and ribosome pathways, respectively (Table 2).

’ DISCUSSION This study is the first to systematically analyze the expression profiles and regulation of genes involved in the development of the silkworm PSG, fibroin synthesis and fibroin secretion during the late larval stages. Profiles of the transcriptome, proteome, and phosphoproteome pave the way for deep insights into the mechanisms underlying these biological processes of the PSG.

The expression analysis of stage-specific genes revealed that ecdysteroid-regulated protein, takeout-like protein (which is a homologue of juvenile hormone binding protein), and many cuticle-related proteins were specifically expressed at M4 when the final shedding of the cuticle occurs. It is known that ecdysones and juvenile hormone are required for the formation of the cuticle. Given that insect silk glands are ectodermal organs and are evolutionarily homologous with epidermal cells, the expression of ecdysis-related genes suggests that the surface layer of the PSG is renewed along with the shedding of epidermal cells, which is under the control of hormones. Furthermore, many cuticular proteins are rich in glycine, alanine, and serine, which are the primary materials for fibroin protein synthesis. It is an interesting question whether the shed cuticle is recycled as a source of amino acids for fibroin synthesis. W-specific genes were found to be involved in various processes such as development, immunity, heat shock, and lipid metabolism. Many of these genes have been characterized as having important regulatory functions. For example, trachealess 3559



ARTICLE

Figure 7. Gene expression regulatory networks built by Pathway Studio software with enrichment analysis. The genes with expression graphs are those detected by microarray, while the others are their downstream neighbors (gray) added from the database based on the literature. Multiple curves in one graph indicate the expression of a gene with one-to-many probes. By unpaired t-tests with M4 and enrichment analysis (P < 0.05), the genes with significant up-regulation (red) and down-regulation (green) are marked with different colors. The color deepness indicates the significance of the differential expression. The added genes (gray) without curves in the graph are detected genes that are not differentially expressed.

is a transcription factor playing an important role in salivary duct determination in Drosophila,41 and it has been proposed to be essential for the formation of the trachea and the ASG in the silkworm.5 However, compared with the other two particular stages, there were far fewer genes specifically expressed during the feeding stages. These results imply that the growth and development of PSG cells during the feeding stages are accompanied by an overall increase in transcription rather than by enhanced expression of certain genes. To comprehensively characterize the regulation of the genes involved in the development of the PSG and the process of fibroin synthesis, we examined the expression of genes for fibroin components, transcription factors, tRNA synthetases, ribosomal proteins, translation initiation factors, and hormone-related proteins. The efficient synthesis of fibroin is triggered at V3 and lasts until V5, at which point it gradually decreases. Although the expression of five silk genes increased markedly at V3 and V5, their patterns were inconsistent, implying that they were under control of different regulators. Additional control mechanisms, such as mRNA stability, translational regulation, and protein degradation, may also be important factors. Indeed, the fibroinase and silk proteinase inhibitors found to exhibit different expression

patterns might be involved in the regulation of fibroin abundance. Moreover, the activity of proteins can be altered through a variety of PTMs and through proteolytic cleavage. The P25, fibrion L-, and H-chain have different levels of PTMs.23,24 Glycyl-tRNA synthetase, which is responsible for synthesizing the transporter for the primary amino acid of fibroin, showed a pattern similar to that of the efficiency of fibroin synthesis. However, the expression levels of ribosomal protein genes and translation initiation factor genes remained relatively stable from M4 to V5 and most of them then decreased at W. In contrast, numerous ribosomal proteins were significantly up-regulated at V5 and down-regulated at W. These results suggest that regulation of ribosomal protein expression at the translational level is vital for the massive synthesis of fibroin proteins. However, the down-regulated ribosomal protein expression at mRNA and protein levels might lead to the decrease of translational capacity. Consequently, it reduced the synthesis of fibroin and of other proteins, and finally resulting in disintegration of the PSG. The silkworm glands begin to disintegrate at the prepupal stage, which is two days after the W stage. PCD of the silkworm silk glands is triggered by the insect steroid hormone 20E, and it proceeds sequentially through cell shrinkage, nuclear condensation, DNA fragmentation, nuclear fragmentation, and apoptotic 3560



ARTICLE

Table 2. The Differentially Expressed Proteins Associated KEGG Pathways protein KEGG pathway

count

protein p-value

q-value

KEGG pathway

V5-up Regulation Proteins

count

p-value

q-value

Glycerolipid metabolism

1

1.12 101 4.41 102

Tyrosine metabolism

1

1.12 101 4.41 102

Glutathione metabolism

1

1.12 101 4.41 102

Fructose and mannose

1

1.16 101 4.49 102

1 1

1.25 101 4.76 102 1.29 101 4.89 102

Associated Pathways Ribosome Oxidative phosphorylation

20 5

5.74 1037 6

4.61 10

5

1.66 1035 1.34 10

5 4

Proteasome

3

9.03 10

1.87 10

Aminoacyl-tRNA biosynthesis One carbon pool by folate

2 1

2.07 103 1.73 102

2.40 103 1.57 102

Folate biosynthesis Benzoate degradation

Protein export

1

2.16 102

1.79 102

Glycolysis/Gluconeogenesis

1

1.75 101 6.50 102

2

2

Limonene and pinene

1

3.56 101 1.10 101

1

3.59 101 1.10 101

metabolism

via CoA ligation Glycine, serine and

1

4.06 10

3.02 10

threonine metabolism

degradation

Glutamate metabolism

1

4.89 102

3.46 102

Purine metabolism

Glutathione metabolism

1

5.10 102

3.48 102

W-up Regulation Proteins

Valine, leucine and

1

6.54 102

4.21 102

Associated Pathways Proteasome

8

2.34 1014 1.12 1012

Oxidative phosphorylation

4

1.16 104 9.28 104

Glycine, serine and

2

8.09 104 3.92 103

Ribosome

3

8.98 104 3.92 103

SNARE interactions in vesicular transport

2

8.98 104 3.92 103

2

1.77 103 5.30 103

isoleucine degradation V5-down Regulation Proteins Associated Pathways Proteasome

8

1.77 1011

3.31 1010

threonine metabolism Ribosome

5

6.31 105

2.14 104

4

4

Glutamate metabolism

3

1.91 10

4.87 10

Arginine and proline metabolism

3

3.47 104

6.83 104

Benzoate degradation

Lysine degradation

3

3.47 104

6.83 104

Butanoate metabolism

2

1.89 103 5.35 103

4

7.83 10

4

Vitamin B6 metabolism

1

6.66 103 1.31 102

7.83 10

4

Bisphenol A degradation

1

8.87 103 1.64 102

7.83 10

4

Ethylbenzene degradation

1

1.33 102 2.09 102

4

Caprolactam degradation

1

1.55 102 2.09 102

Limonene and pinene degradation

2

1.68 102 2.09 102

Methane metabolism

1

1.77 102 2.09 102

One carbon pool by folate

1

1.77 102 2.09 102

Phenylalanine, tyrosine and

1

2.20 102 2.25 102

via CoA ligation Aminoacyl-tRNA biosynthesis Tryptophan metabolism Fatty acid metabolism

3 3 3

4.72 10

4

4.72 10

4

4.72 10

4

Valine, leucine and isoleucine degradation

3

4.72 10

7.83 10


2

4.96 104

7.83 104

4

7.83 10

4

1.34 10

3 3

Fatty acid elongation in mitochondria Starch and sucrose metabolism

2 3

4.96 10

3

1.00 10

3

2

1.05 10

1.37 10


4

2.37 103

2.86 103

1- and 2-Methylnaphthalene

1

2.86 102 2.80 102

Selenoamino acid metabolism

2

2.42 103

2.86 103

degradation Phenylalanine metabolism

1

3.07 102 2.89 102

3

4.19 10

3

beta-Alanine metabolism

1

3.72 102 3.30 102

5.39 10

3

Alkaloid biosynthesis II

1

3.93 102 3.43 102

7.76 10

3

Propanoate metabolism

1

4.79 102 3.99 102

8.53 10

3

Lysine degradation

1

6.05 102 4.43 102

1.07 10

2

Fatty acid metabolism

1

6.68 102 4.47 102

2

Valine, leucine and isoleucine

1

6.68 102 4.47 102

Phenylalanine, tyrosine and tryptophan biosynthesis

Alanine and aspartate metabolism Propanoate metabolism Butanoate metabolism Citrate cycle (TCA cycle) Cyanoamino acid metabolism

tryptophan biosynthesis

2 2 2 2 1

3.89 10

3

5.20 10

3

8.94 10

2

1.08 10

2

1.47 10

2

Novobiocin biosynthesis

1

1.47 10

1.07 10

Benzoate degradation

1

1.96 102

1.34 102

degradation Tryptophan metabolism

1

6.68 102 4.47 102

1

2.92 102

1.85 102

Aminoacyl-tRNA biosynthesis

1

6.68 102 4.47 102

2

2.05 10

2

Pyruvate metabolism

1

6.68 102 4.47 102

2.18 10

2

Citrate cycle (TCA cycle)

1

6.89 102 4.47 102

2.18 10

2

Glycolysis/Gluconeogenesis

1

8.33 102 5.19 102

via hydroxylation Alkaloid biosynthesis I Pyrimidine metabolism One carbon pool by folate Methane metabolism

2 1 1

3.32 10

2

3.87 10

2

3.87 10

3561



ARTICLE

Table 2. Continued protein KEGG pathway Methionine metabolism

count 1

protein p-value 2

3.87 10

2

q-value 2.18 10

2 2

1

4.35 10

2.34 10

Cysteine metabolism

1

4.82 102

Glycan structures - biosynthesis 2

1

Arachidonic acid metabolism Sphingolipid metabolism Reductive carboxylate cycle

KEGG pathway Phenylpropanoid biosynthesis

count 1

p-value

q-value

1

8.23 102

1

8.48 102

1.65 10

Purine metabolism

1

1.82 10

2.50 102

gamma-Hexachlorocyclohexane

1

1.93 101 8.82 102

5.29 102

2.67 102

W-down Regulation Proteins

1 1

5.29 102 6.22 102

2.67 102 2.99 102

Ribosome Fatty acid metabolism

7 2

2.51 1011 1.67 1011 8.84 104 1.68 104

1

6.22 102

2.99 102

Valine, leucine and isoleucine

2

8.84 104 1.68 104

Bile acid biosynthesis

1

6.68 102

3.14 102

Fatty acid elongation

1

9.90 103 1.13 103

Phenylalanine metabolism

1

6.68 102

3.14 102


1

9.90 103 1.13 103

2

2


2

1.12 102 1.18 103

Valine, leucine and isoleucine iosynthesis

degradation Associated Pathways

(CO2 fixation)

degradation in mitochondria

beta-Alanine metabolism Galactose metabolism Glycine, serine and threonine

1

8.05 10

3.54 10

1 1

2

2

8.96 10 8.96 102

3.74 10 3.74 102

beta-Alanine metabolism Propanoate metabolism

1 1

2.39 102 2.12 103 3.08 102 2.57 103

1

8.96 102

3.74 102

Lysine degradation

1

3.90 102 2.89 103

2

2

Butanoate metabolism

1

4.04 102 2.89 103

Tryptophan metabolism

1

4.31 102 2.97 103

metabolism Carbon fixation Porphyrin and chlorophyll

1

9.40 10

3.87 10

1

1.03 101

4.21 102

metabolism Dorso-ventral axis formation

body formation.19,42 Jia et al. using comparative proteomic methods revealed hsp20.1 and elongation factor 1-beta0 might participate in the silk gland PCD process.43 In our study, several stage-specific proteins identified by LCMS/MS in the PSG at W might be involved in the process of PCD, such as EcR, hsp19.9, hsp20.1, hsp20.8, and elongation factor 1-alpha. 2-DE analysis detected three sHSPs (hsp20.4, hsp20.8, and hsp21.4) that were dramatically up-regulated at W. These results imply that these sHSPs have close association with the physiological changes and might be related to the forthcoming PCD. Given that the degradation of mRNAs is regulated by the heat shockubiquitin-proteasome pathway,44 and that PCD-related proteins, such as the cathepsins and proteasomes, were significantly upregulated at W, the sHSPs might play important roles in the PCD process of the PSG. Furthermore, the vast majority of the regulated proteolysis in eukaryotic cells is performed through the ubiquitin-proteasome pathway (UPP).45 Protein degradation via the UPP initiates with the conjugation of multiple ubiquitin moieties to the substrate which is followed by degradation of the tagged protein by the downstream 26S proteasome complex.45 The up-regulation of proteasome pathway proteins such as ubiquitin-conjugating enzyme E2L and 26S proteasome nonATPase regulatory subunit 13 occurred at W, implying an important role of the UPP in the PCD of the silk gland. The requirements for PCD seem to be all ready at the W stage and waiting to be triggered by 20E. The gene expression can be regulated at transcription- and translation-level, which may always lead to the inconsistency of levels between genes and proteins.46,47 It may also be an important reason for the differences between proteome analysis data and microarray results in the present study. For example, the actin-depolymerizing factor 1 (protein spot, P3; microarray probe, sw04646) constantly

increased its protein expression from M4 to V3, while clearly downregulated at V5 and then kept stable (Table 1). Although its variation tendency of gene expression was perfectly consistent with that of protein level, their changing magnitudes were different (Supplemental Table S1). Another example is the beta-tubulin (sw17349) of which the gene expression was much lower at M4 and W than at feeding stages. By contrast, its protein expression (spot P7) was much lower at V3 (Table 1). These inconsistencies implied the existence of possible PTM events such as phosphorylation. The phosphorylation and dephosphorylation of proteins and peptides regulate a number of key biological processes within the cell, including signal transduction, cell division, differentiation, and apoptosis in eukaryotic organisms.25 So far, there have been no phosphoproteomic reports on the silkworm, let alone the silk gland. Zhang et al. found several phosphorylated forms of fibroin L-chain and P25 in the silkworm PSG at V5.23 However, we did not detect the phosphorylated forms of fibroin L-chain by fluorescent dye staining, although it is expressed in high amounts through the late larval stages. This suggests that the fibroin L-chain we identified was not phosphorylated or was phosphorylated at an extremely low level. The global phosphoproteins of the PSG accounted for about one-sixth of the proteome, which is much lower than the phosphorylation level of 30% in eukaryotic cells.48,49 We presume that this low level of phosphorylation might be related to its reduced roles in larval PSG cells, whose division and differentiation were completed during the embryonic stage.5 However, phosphorylation events might still be very important at special physiological states.50 For example, the mRNA levels of the casein kinases gradually decreased from M4 to W. The shotgun proteomic data confirmed these results to some extent. Because the translation-level regulation of ribosomal proteins discussed above might be critical for fibroin synthesis, we presume that the 3562


Journal of Proteome Research variation of the phosphorylation level of some of the identified proteins play important roles. The phosphorylation level of the proteome seemed to remain stable at all the examined stages except for a few changes. The phosphorylation and dephosphorylation of some proteins at different physiological stages hinted at internal correlations. For example, enolase was dephosphorylated after M4, implying that its phosphorylation is closely associated with the function of the PSG during the molting stage. By contrast, the phosphorylation of actin-depolymerizing factor 1 at feeding stages suggested that this change might be involved in the growth and development of PSG. Actin-depolymerizing factor is a member of the actin-binding protein family involved in actin remodeling.51 The cumulative expression of the actin A3 gene has been studied in relation to the silk production cycle in the two secretory regions of silkworm silk glands.52 Because the secretory channels of silk gland cells are formed by microtubules in association with actin-containing microfilament bundles extending around the circumference of the gland lumen, disruption of these microtubules and microfilaments results in an interruption of the secretion of silk proteins.53 We hypothesize that phosphorylated actin-depolymerizing factor 1 either protects these microtubules and microfilaments from disruption or is required for the growth of PSG cells.

’ CONCLUSIONS This large-scale characterization of transcriptome, proteome, and phosphoproteome profiles provides copious data relevant to a comprehensive understanding of the molecular basis and regulatory mechanisms of PSG development and fibroin synthesis. The results revealed notable expression attributes of the PSG during molting and wandering stages and that diverse factors have effects on the accumulation of fibroin. Especially, the significant upregulation of ribosomal proteins occurred at the very time of large amounts of fibroin synthesis, revealing one of critical factors for the high efficient protein synthesis system. Additionally, the vast amount transcriptomic and proteomic data we obtained will contribute greatly to the silkworm genome annotation. ’ ASSOCIATED CONTENT

bS

Supporting Information Tables of transcriptome profiles, primers for qRT-PCR, differential expression genes, stage-specific genes, proteome profiles, differentially expressed proteins; figures of correlation analysis of microarray and qRT-PCR data, Volcano Plots of differentially expressed genes, and correlations of differential expression genes with co-expression patterns. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Prof. Bo-xiong Zhong, College of Animal Sciences, Zhejiang University, Hangzhou 310029, P. R. China. Tel/Fax: +86-57186971302. E-mail: [email protected]..

’ ACKNOWLEDGMENT This work was supported by the National Basic Research Program of China (Grant No. 2005CB121003), the National Natural Science Foundation of China (Grant No. 30972142), the National Hi-Tech Research and Development Program of China (Grant No. 2006AA10A118), and the Doctoral Fund of the

ARTICLE

Ministry of Education of China (Grant No. 20070335148). We thank Ms. Xiaohong Chen and Dr. Liang Zhang in Corporation CapitalBio, Beijing, and Dr. Ye Feng in ZCNI, Zhejiang University, for their kind help in microarray experiment. We are grateful to the 985-Institute of Agrobiology and Environmental Sciences of Zhejiang University for providing access to their experimental equipment.

’ REFERENCES (1) International Silkworm Genome Consortium.. The genome of a lepidopteran model insect, the silkworm Bombyx mori. Insect Biochem. Mol. Biol. 2008, 38 (12), 1036–1045. (2) Goldsmith, M. R.; Shimada, T.; Abe, H. The genetics and genomics of the silkworm, Bombyx mori. Annu. Rev. Entomol. 2005, 50, 71–100. (3) Tomita, M.; Munetsuna, H.; Sato, T.; Adachi, T.; Hino, R.; Hayashi, M.; Shimizu, K.; Nakamura, N.; Tamura, T.; Yoshizato, K. Transgenic silkworms produce recombinant human type III procollagen in cocoons. Nat. Biotechnol. 2003, 21 (1), 52–56. (4) Iizuka, M.; Ogawa, S.; Takeuchi, A.; Nakakita, S.; Kubo, Y.; Miyawaki, Y.; Hirabayashi, J.; Tomita, M. Production of a recombinant mouse monoclonal antibody in transgenic silkworm cocoons. FEBS J. 2009, 276 (20), 5806–5820. (5) Matsunami, K.; Kokubo, H.; Ohno, K.; Xu, P.; Ueno, K.; Suzuki, Y. Embryonic silk gland development in Bombyx: molecular cloning and expression of the Bombyx trachealess gene. Dev. Genes Evol. 1999, 209 (9), 507–514. (6) Suzuki, Y.; Brown, D. D. Isolation and identification of the messenger RNA for silk fibroin from Bombyx mori. J. Mol. Biol. 1972, 63 (3), 409–429. (7) Couble, P.; Moine, A.; Garel, A.; Prudhomme, J. C. Developmental variations of a nonfibroin mRNA of Bombyx mori silkgland, encoding for a low-molecular-weight silk protein. Dev. Biol. 1983, 97 (2), 398–407. (8) Tashiro, Y.; Morimoto, T.; Matsuura, S.; Nagata, S. Studies on the posterior silk gland of the silkworm, Bombyx mori. I. Growth of posterior silk gland cells and biosynthesis of fibroin during the fifth larval instar. J. Cell Biol. 1968, 38 (3), 574–588. (9) Grzelak, K. Control of expression of silk protein genes. Comp. Biochem. Physiol., Part B: Biochem. Mol. Biol. 1995, 110 (4), 671–681. (10) Maekawa, H.; Suzuki, Y. Repeated turn-off and turn-on of fibroin gene transcription during silk gland development of Bombyx mori. Dev. Biol. 1980, 78 (2), 394–406. (11) Takiya, S.; Kokubo, H.; Suzuki, Y. Transcriptional regulatory elements in the upstream and intron of the fibroin gene bind three specific factors POU-M1, Bm Fkh and FMBP-1. Biochem. J. 1997, 321 (Pt 3), 645–653. (12) Hui, C. C.; Matsuno, K.; Suzuki, Y. Fibroin gene promoter contains a cluster of homeodomain binding sites that interact with three silk gland factors. J. Mol. Biol. 1990, 213 (4), 651–670. (13) Suzuki, T.; Matsuno, K.; Takiya, S.; Ohno, K.; Ueno, K.; Suzuki, Y. Purification and characterization of an enhancer-binding protein of the fibroin gene. I. Complete purification of fibroin factor 1. J. Biol. Chem. 1991, 266 (25), 16935–16941. (14) Suzuki, T.; Takiya, S.; Matsuno, K.; Ohno, K.; Ueno, K.; Suzuki, Y. Purification and characterization of an enhancer-binding protein of the fibroin gene. II. Functional analyses of fibroin factor 1. J. Biol. Chem. 1991, 266 (25), 16942–16947. (15) Durand, B.; Drevet, J.; Couble, P. P25 gene regulation in Bombyx mori silk gland: two promoter-binding factors have distinct tissue and developmental specificities. Mol. Cell. Biol. 1992, 12 (12), 5768–5777. (16) Horard, B.; Julien, E.; Nony, P.; Garel, A.; Couble, P. Differential binding of the Bombyx silk gland-specific factor SGFB to its target DNA sequence drives posterior-cell-restricted expression. Mol. Cell. Biol. 1997, 17 (3), 1572–1579. 3563


Journal of Proteome Research (17) Sehnal, F.; Akai, H. Insect silk glands: their types, development and function, and effects of environmental factors and morphogenetic hormones on them. Int. J. Insect Morphol. Embryol. 1990, 19 (2), 79–132. (18) Chaitanya, R. K.; Dutta-Gupta, A. Light chain fibroin and P25 genes of Corcyra cephalonica: Molecular cloning, characterization, tissuespecific expression, synchronous developmental and 20-hydroxyecdysone regulation during the last instar larval development. Gen. Comp. Endrocrinol. 2010, 167 (1), 113–121. (19) Terashima, J.; Yasuhara, N.; Iwami, M.; Sakurai, S. Programmed cell death triggered by insect steroid hormone, 20-hydroxyecdysone, in the anterior silk gland of the silkworm, Bombyx mori. Dev. Genes Evol. 2000, 210 (11), 545–558. (20) Tsuzuki, S.; Iwami, M.; Sakurai, S. Ecdysteroid-inducible genes in the programmed cell death during insect metamorphosis. Insect Biochem. Mol. Biol. 2001, 31 (45), 321–331. (21) Xia, Q.; Cheng, D.; Duan, J.; Wang, G.; Cheng, T.; Zha, X.; Liu, C.; Zhao, P.; Dai, F.; Zhang, Z.; He, N.; Zhang, L.; Xiang, Z. Microarraybased gene expression profiles in multiple tissues of the domesticated silkworm, Bombyx mori. Genome Biol. 2007, 8 (8), R162. (22) Zhou, Z. H.; Yang, H. J.; Chen, M.; Lou, C. F.; Zhang, Y. Z.; Chen, K. P.; Wang, Y.; Yu, M. L.; Yu, F.; Li, J. Y.; Zhong, B. X. Comparative proteomic analysis between the domesticated silkworm (Bombyx mori) reared on fresh mulberry leaves and on artificial diet. J. Proteome Res. 2008, 7 (12), 5103–5111. (23) Zhang, P. B.; As, Y. C.; Yamamoto, K.; Banno, Y.; Wang, Y. Q.; Tsuchida, K.; Kawaguchi, Y.; Fujii, H. Proteome analysis of silk gland proteins from the silkworm, Bombyx mori. Proteomics 2006, 6 (8), 2586–2599. (24) Chen, W. Q.; Priewalder, H.; John, J. P. P.; Lubec, G. Silk cocoon of Bombyx mori: Proteins and posttranslational modifications heavy phosphorylation and evidence for lysine-mediated cross links. Proteomics 2010, 10 (3), 369–379. (25) Gruhler, A.; Olsen, J. V.; Mohammed, S.; Mortensen, P.; Faergeman, N. J.; Mann, M.; Jensen, O. N. Quantitative phosphoproteomics applied to the yeast pheromone signaling pathway. Mol. Cell. Proteomics 2005, 4 (3), 310–327. (26) Steinberg, T. H.; Agnew, B. J.; Gee, K. R.; Leung, W. Y.; Goodman, T.; Schulenberg, B.; Hendrickson, J.; Beechem, J. M.; Haugland, R. P.; Patton, W. F. Global quantitative phosphoprotein analysis using multiplexed proteomics technology. Proteomics 2003, 3 (7), 1128–1144. (27) Yang, Y. H.; Dudoit, S.; Luu, P.; Lin, D. M.; Peng, V.; Ngai, J.; Speed, T. P. Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res. 2002, 30 (4), e15. (28) Tusher, V. G.; Tibshirani, R.; Chu, G. Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. U.S.A. 2001, 98 (9), 5116–5121. (29) Li, J. Y.; Chen, X.; Fan, W.; Moghaddam, S. H.; Chen, M.; Zhou, Z. H.; Yang, H. J.; Chen, J. E.; Zhong, B. X. Proteomic and bioinformatic analysis on endocrine organs of domesticated silkworm, Bombyx mori L. for a comprehensive understanding of their roles and relations. J. Proteome Res. 2009, 8 (6), 2620–2632. (30) Elias, J. E.; Gygi, S. P. Target-decoy search strategy for increased confidence in large-scale protein identifications by mass spectrometry. Nat. Methods 2007, 4 (3), 207–214. (31) Keller, A.; Eng, J.; Zhang, N.; Li, X. J.; Aebersold, R. A uniform proteomics MS/MS analysis platform utilizing open XML file formats. Mol. Syst. Biol. 2005, 1, 0017. (32) Lu, P.; Vogel, C.; Wang, R.; Yao, X.; Marcotte, E. M. Absolute protein expression profiling estimates the relative contributions of transcriptional and translational regulation. Nat. Biotechnol. 2007, 25 (1), 117–124. (33) Schroeder, W. A.; Kay, L. M.; Lewis, B.; Munger, N. The amino acid composition of Bombyx mori silk fibroin and of tussah silk fibroin. J. Am. Chem. Soc. 1955, 77 (14), 3908–3913. (34) Zurovec, M.; Yang, C.; Kodrik, D.; Sehnal, F. Identification of a novel type of silk protein and regulation of its expression. J. Biol. Chem. 1998, 273 (25), 15423–15428.

ARTICLE

(35) Nirmala, X.; Mita, K.; Vanisree, V.; Zurovec, M.; Sehnal, F. Identification of four small molecular mass proteins in the silk of Bombyx mori. Insect Mol. Biol. 2001, 10 (5), 437–445. (36) Priness, I.; Maimon, O.; Ben-Gal, I. Evaluation of gene-expression clustering via mutual information distance measure. Bmc Bioinformatics 2007, 8, 111. (37) Kotelnikova, E.; Ivanikova, N.; Kalinin, A.; Yuryev, A.; Daraselia, N. Atlas of signaling for interpretation of microarray experiments. PLoS One 2010, 5 (2), e9256. (38) Hudson, L. G.; Santon, J. B.; Glass, C. K.; Gill, G. N. Ligandactivated thyroid hormone and retinoic acid receptors inhibit growth factor receptor promoter expression. Cell 1990, 62 (6), 1165–1175. (39) Rusconi, J. C.; Fink, J. L.; Cagan, R. klumpfuss regulates cell death in the Drosophila retina. Mech. Dev. 2004, 121 (6), 537–546. (40) Li, G.; Gustafson-Brown, C.; Hanks, S. K.; Nason, K.; Arbeit, J. M.; Pogliano, K.; Wisdom, R. M.; Johnson, R. S. c-Jun is essential for organization of the epidermal leading edge. Dev. Cell 2003, 4 (6), 865–877. (41) Kuo, Y. M.; Jones, N.; Zhou, B.; Panzer, S.; Larson, V.; Beckendorf, S. K. Salivary duct determination in Drosophila: roles of the EGF receptor signalling pathway and the transcription factors fork head and trachealess. Development 1996, 122 (6), 1909–1917. (42) Iga, M.; Iwami, M.; Sakurai, S. Nongenomic action of an insect steroid hormone in steroid-induced programmed cell death. Mol. Cell. Endocrinol. 2007, 263 (12), 18–28. (43) Jia, S. H.; Li, M. W.; Zhou, B.; Liu, W. B.; Zhang, Y.; Miao, X. X.; Zeng, R.; Huang, Y. P. Proteomic analysis of silk gland programmed cell death during metamorphosis of the silkworm Bombyx mori. J. Proteome Res. 2007, 6 (8), 3003–3010. (44) Laroia, G.; Cuesta, R.; Brewer, G.; Schneider, R. J. Control of mRNA decay by heat shock ubiquitin proteasome pathway. Science 1999, 284 (5413), 499–502. (45) Glickman, M. H.; Ciechanover, A. The ubiquitin-proteasome proteolytic pathway: Destruction for the sake of construction. Physiol. Rev. 2002, 82 (2), 373–428. (46) Gygi, S.; Rochon, Y.; Franza, B. R.; Aebersold, R. Correlation between protein and mRNA abundance in yeast. Mol. Cell. Biol. 1999, 19 (3), 1720–1730. (47) Pandey, A.; Mann, M. Proteomics to study genes and genomes. Nature 2000, 405 (6788), 837–846. (48) Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.; Shabanowitz, J.; Hunt, D. F.; White, F. M. Phosphoproteome analysis by mass spectrometry and its application to Saccharomyces cerevisiae. Nat. Biotechnol. 2002, 20 (3), 301–305. (49) Ptacek, J.; Devgan, G.; Michaud, G.; Zhu, H.; Zhu, X. W.; Fasolo, J.; Guo, H.; Jona, G.; Breitkreutz, A.; Sopko, R.; McCartney, R. R.; Schmidt, M. C.; Rachidi, N.; Lee, S. J.; Mah, A. S.; Meng, L.; Stark, M. J. R.; Stern, D. F.; De Virgilio, C.; Tyers, M.; Andrews, B.; Gerstein, M.; Schweitzer, B.; Predki, P. F.; Snyder, M. Global analysis of protein phosphorylation in yeast. Nature 2005, 438 (7068), 679–684. (50) Chessa, T. A.; Anderson, K. E.; Hu, Y.; Xu, Q.; Rausch, O.; Stephens, L. R.; Hawkins, P. T. Phosphorylation of threonine 154 in p40phox is an important physiological signal for activation of the neutrophil NADPH oxidase. Blood 2010, 116 (26), 6027–6036. (51) Maciver, S. K.; Hussey, P. J. The ADF/cofilin family: actinremodeling proteins. Genome Biol. 2002, 3 (5), reviews3007.3001–3007.3012. (52) Mounier, N.; Coulon, M.; Prudhomme, J. C. Expression of a cytoplasmic actin gene in relation to the silk production cycle in the silk glands of Bombyx mori. Insect Biochem. 1991, 21 (3), 293–301. (53) Couble, P.; Blaes, N.; Prudhomme, J. C. Actin microfilaments and fibroin secretion in the silkgland cells of Bombyx mori. Effects of cytochalasin B. Exp. Cell Res. 1984, 151 (2), 322–331.

3564


Expression Profiling and Regulation of Genes Related to Silkworm

Recommend Documents