Bombyx mori Posterior and Median Silk Gland SAGE transcriptome

To identify functions that distinguish the posterior and median cells producing fibroin and sericin in the silk gland of Bombyx mori, serial analysis of gene expression (SAGE) profiles from both silk gland regions were analyzed and compared. The construction of a B. mori reference tag collection extracted from a set of 38000 Bombyx EST sequenced from the 3’ side, helped annotating the SAGE libraries. Most of the tags appeared at similar relative concentration in the two libraries except for those corresponding to silk proteins that were found region-specific and highly abundant. Strikingly, besides tags from silk protein mRNAs, 19 tags were found in the class of high abundance in the median cell library, which were absent in the posterior cell tag collection. Except tags from SP1 mRNA, no PSG specific tags were found in the same class of abundance. The analysis of MSG-specific different transcripts led to suggest that middle silk gland cell realizes more diversified functions as those already known, of synthesis and secretion of the silk sericins. EST libraries from 11 silkworm tissues were 3’-sequenced to ensure the identification of the most terminal tag. 37,920 sequences were analyzed on ABI 3700 or 3730XL sequencers. Electrophoregrams were processed with KB Basecaller (3730XL traces) or with PHRED (3700 traces) to obtain the .phd files from which were extracted text sequences and their corresponding quality files in Fasta format (Phd2Fasta, Green and Ewing, 1995; 2002). Vector sequences and bad quality regions were removed with an home made software after identification by Lucy (Chou and Holmes, 2001). Chimera were removed by an home made software and retrotransposon sequences were masked by RepeatMasker (http://repeatmasker.org). The cleaned sequences were clusterized with TGICL package (TIGR) and assembled into contigs with CAP3 (Huang and Madan, 1999). Contigs were identified with Blastn and Blastx on GenBank and SwissProt/TREMBL, respectively. Some clusters, splitted by CAP3 procedure, have been regrouped on the basis of Blastx identity. Identitag (Keime et al., 2004) was used to extract all possible tags (forward and reverse) to create a reference database. Moreover, a quality index is attached to each tag depending to the presense of poly-A, polyadenylation signal and its proximity to 3-prime extremity in the original mRNA sequence. This database was supplemented with the tags extracted from public B. mori sequences (GenBank, Silkbase) with the same software. Tags were extracted from MSG and PSG libraries (from 2304 and 3072 sequenced inserts respectively) with the Sagenhaft software (Beissbarth et al., 2004) then identified and compared with Identitag. The assessment of significant differences among the two libraries was performed by using the Z-test used for comparison of SAGE libraries of different size (Kal et al., 1999). For graphic purpose and to avoid division by zero we used a tag value of 1 for tags that were not detected in MSG or PSG libraries. Since the two SAGE libraries showed up different TAG number, we used the Z-test for comparing the two mRNA populations.

为了鉴定区分家蚕（Bombyx mori）丝腺中分泌丝素蛋白与丝胶蛋白的后部与中部细胞功能，本研究对两处丝腺区域的基因表达系列分析（Serial Analysis of Gene Expression, SAGE）谱进行了分析与比较。本研究基于从38000条家蚕3'端测序的表达序列标签（Expressed Sequence Tag, EST）中提取的序列，构建了家蚕参考标签集合，用于注释SAGE文库。两个文库中多数标签的相对丰度较为接近，仅区域特异性且高丰度的丝蛋白相关标签除外。值得注意的是，除丝蛋白mRNA对应的标签外，中部细胞文库中存在19个高丰度标签，而这些标签在后部细胞标签集合中完全缺失；除SP1 mRNA对应的标签外，该丰度等级中未发现后部丝腺（Posterior Silk Gland, PSG）特异性标签。对中部丝腺（Median Silk Gland, MSG）特异性转录本的分析表明，中部丝腺细胞具备比现有认知更为多样化的功能，包括丝胶蛋白的合成与分泌。为确保能够识别最末端标签，本研究对11个家蚕组织的表达序列标签文库进行了3'端测序，共计37920条序列在ABI 3700或3730XL测序仪上完成分析。使用KB Basecaller（处理3730XL测序数据）或PHRED（处理3700测序数据）对电泳图谱进行处理，得到.phd格式文件；随后通过Phd2Fasta工具（Green与Ewing, 1995; 2002）从该文件中提取序列文本及其对应Fasta格式的质量文件。通过Lucy软件（Chou与Holmes, 2001）识别载体序列与低质量区域后，使用自研软件将其移除；使用自研软件移除嵌合序列，并通过RepeatMasker（http://repeatmasker.org）屏蔽反转录转座子序列。清理后的序列使用TGICL工具包（TIGR）进行聚类，并通过CAP3软件（Huang与Madan, 1999）拼接为重叠群。分别通过GenBank数据库的Blastn比对与SwissProt/TREMBL数据库的Blastx比对完成重叠群的注释。部分被CAP3拆分的聚类簇可基于Blastx比对相似度重新合并。使用Identitag工具（Keime等人, 2004）提取所有可能的标签（正向与反向）以构建参考数据库。此外，每个标签都会附带一个质量指数，该指数基于原始mRNA序列中poly-A尾、多聚腺苷酸化信号的存在情况，以及标签与3'端的相对距离计算得到。本数据库还补充了使用相同软件从公共家蚕序列（GenBank、Silkbase）中提取的标签。使用Sagenhaft软件（Beissbarth等人, 2004）从中部丝腺与后部丝腺文库（分别对应2304与3072条测序插入片段）中提取标签，随后通过Identitag工具完成标签的识别与比对。为评估两个文库间的显著差异，本研究采用适用于不同规模SAGE文库比较的Z检验（Kal等人, 1999）。为便于可视化并避免除零错误，对于未在中部丝腺或后部丝腺文库中检测到的标签，将其标签值设为1。由于两个SAGE文库的标签总数存在差异，本研究采用Z检验完成两个mRNA群体的比较。