Regulation of Fibroin Gene Expression in Bombyx mori

Abstract

The silkworm Bombyx mori has been exploited as a convenient model system for the study of eukaryotic gene expression and its regulation. The genes investigated include chorion genes, certain tRNA genes (alanyl and glycyl), actin genes, and genes encoding cocoon proteins-the silk fiber protein fibroin, the glue proteins sericin(s), and P25, the fibroin light chain component. The silk glands, which are modified salivary glands, are the site of synthesis of the silk proteins. Each larva possesses a pair of silk glands, physiologically and anatomically divided into three parts: the anterior silk gland (ASG), middle silk gland (MSG), and posterior silk gland (PSG). The silk fiber proteins are synthesized in the PSG, while the glue proteins (sericins) are synthesized in the MSG. Fibroin synthesized in the PSG is secreted into the lumen and transported to the MSG, where sericins are synthesized and secreted; the sericin coated fibroin is stored until secretion. The ASG serves as the duct, transporting the silk proteins. The two ASGs converge to form the sclerotized spinneret through which silk fibers are extruded. Thus, silk protein gene expression is regulated in a tissue specific manner. It is also developmentally regulated, occurring only during larval stages, with maximal expression during the 5th instar, just prior to cocoon formation. Previous work has shown that fibroin expression is regulated at the transcriptional level. Work from our laboratory has additionally shown that developmental regulation also occurs post transcriptionally, presumably at the translational level due to limited amounts of cognate tRNAs. The present study was undertaken to identify the cis acting DNA sequence elements and trans activating proteins involved in the transcriptional regulation of fibroin gene expression. The underlying premise was that specific protein-DNA interactions at regulatory regions govern expression. The study focused on the region of the fibroin gene encompassing ~800 nucleotides upstream from the transcription start site. The experimental approach involved constructing DNA segments containing specific upstream regions and assessing their transcription in vitro. To perform in vitro transcription, a method for preparing nuclear extracts from homologous silk gland tissue was developed. This method differed from earlier protocols for Bombyx mori cell free extracts. The nuclear extracts were competent for both RNA polymerase II and III transcription. Deletion subclones were generated from the full length genomic clone pFb29 covering various fibroin upstream regions. These subclones were used in transcription assays and DNA-protein interaction studies. Conditions for optimal in vitro transcription using exogenous fibroin templates were standardized by monitoring incorporation of radiolabeled NTPs. Incubation for 1 hour at 25°C with 1-2µg template DNA yielded best results. Extract activity was strictly template dependent. Constructs containing sequences between +66 and 233 were most efficiently transcribed, and sequences between 95 and 233 enhanced transcription. Sequences upstream of 233 appeared to contain a negative regulatory element. Heterologous extracts (from MSG and ovary, tissues that do not express fibroin in vivo) also supported transcription but at much lower levels. Sequences required for maximal transcription were similar across PSG, MSG, and ovarian extracts. Run off transcription confirmed that transcription initiated at the correct fibroin start site. S1 nuclease analysis with subclones pF94, pF233, and pF540 showed pF233 gave highest transcription, while inclusion of 233 to 540 sequences reduced transcription, again indicating a negative regulatory element in that region. These results confirmed that the homologous extracts could: • transcribe added fibroin DNA faithfully and specifically, and • distinguish functional upstream regulatory elements. Therefore, the study proceeded to characterize protein-DNA interactions using gel retardation assays (EMSA), Southwestern blotting, UV crosslinking, and footprinting. Optimal EMSA conditions were established. Binding reactions at 0°C and pre incubation with nonspecific competitor DNA produced best signal discrimination. Poly(dI:dC) was the most effective nonspecific competitor. Specificity was verified by self competition and non specific competitor DNA. Competition experiments showed that probes pF94 and pFEn95 shared common binding factors. pF233 also shared the same factors. Binding downstream of 94 was stronger than binding between 95 and 233. Probe pFU288 showed a unique complex that was not competed by sequences between 233 and +66, indicating a distinct protein likely responsible for the negative regulatory influence. Binding pattern comparison showed similar protein interactions in MSG and PSG extracts, but distinct patterns in ovary extracts. Southwestern analysis revealed: • a 70kDa PSG specific protein binding to pF233, • two proteins (75-80kDa and 53kDa) binding specifically to pFU288, • a 70kDa protein binding only to pFEn95, • ovarian specific proteins (~45-47kDa and 25kDa) binding to pFU288, and • a 50kDa protein binding to pF233 only in ovary extracts. UV crosslinking showed pF94 and pF233 bound a 47kDa protein, while pF233 additionally bound 70-75kDa and 97kDa proteins. Substituting TTP with BrdUTP inhibited protein-DNA binding, likely because the sequences (>70% AT rich) require intact thymine residues for binding. Footprinting using copper-phenanthroline nuclease revealed extensive protection between 94 and +66.