Fine Mapping Of An Extra-crescents And Degenerated Abdominal Legs Mutant(E^dl) And The Identification Of Relative Regulatory Elements

In order to survive and reproduce,insects have evolved a variety of appendages,which are mainly the result of changes in gene expression pattern and function during the evolution process.The development of appendages is a complex physiological process,which is regulated by multiple gene networks.Homeobox genes?Homeotic genes,Hox?are first discovered in Drosophila,and are principal regulators of the body plan development in insect.Hox genes are aligned on the genome in the order of the segments that they specify along the antero-posterior axis of the body pattern,and Hox genes specifying the location and morphology characteristics of appendages are important in the development and differentiation of appendages.The functions of Hox genes have been well studied in insect,and it is found that the functional evolution of Hox genes in different species results in varied appendages.The expression of Hox gene is regulated precisely,considerable cis elements and non encoding RNA?non-coding RNA,nc RNA?in Drosophila BX-C gene cluster are identified,which regulate the precise expression of Hox gene.However,the studies on the regulation of Hox gene expression in other insects,especially the regulation of Hox gene by functional elements in Hox gene clusters are still scarce.Owning complete genomic data and detailed maps of genetic variation,silkworm is not only an important economic insect but also an important mode of lepidoptera.In silkworm,there are abundant mutants in Hox loci such as E pseudoallele group?E group?,which are good materials for studying the function and regulation relationships of lepidopteran Hox gene.E group contains more than 30 mutants but the region of this group only occupies three Hox genes.In this study,we explored a mutant in the E group called extra-crescents and degenerated abdominal legs mutant(E^dl).We investigated the role of Hox gene in the formation of E^dl mutation by positional cloning,expression analysis and immunohistochemistry;morever,we also identified the regulatory elements in E^dl locus,and then analyzed the potential relationship between the regulatory elements and Hox genes by bioinformatics,expression association analysis and RNAi.The main results of this paper are as follows: 1.Morphological characteristics and genetic analysis of the E^dl mutant in silkwormE^dl mutant in E group is located at 21.1 c M on the 6th linkage of the silkworm genetic linkage map.Morphological characteristics of the E^dl mutant are previously described to be a dominant mutation with extra crescents and degenerated abdominal legs on the third abdominal segment as well as the loss of star spots on the fifth abdominal segment.The morphological characteristics of E^dl mutant observed by us also display extra crescents and degenerated abdominal legs on the third abdominal segment but display normal star spots on the fifth abdominal segment;we also observed that the distal limb structures are lost in the degenerated prolegs and few individuals of the E^dl mutant exist segmental malformation in the dorsal side of the second and third abdomen segments.Genetic analysis showed that F1 offsprings from the hybridization between Dazao?wildtype?and E^dl mutant showed the normal phenotype;self-corss of the F1 offsprings showed two separated phenotypes,the normal phenotype and E^dl phenotype,and the separation ratio is 3:1;reciprocal testcorsses between F1 offsprings and Dazao individuals showed normal phenotype;reciprocal testcorsses between F1 offsprings and E^dl mutant showed a 1:1 separation of the normal phenotype and E^dl phenotype.Above results of hybridizations and reciprocal testcorsses indicate that E^dl is a recessive mutant.Based on the characteristics of complete linkage of female individuals in silkworm.We performed the hybridization and backcorsses between E^dl mutant and Dazao for linkage analysis.According to studies in the ECs-l and EKp-1 mutants in E group,markers were designed to narrow down the E^dl locus.Using 386 BC1 M individuals,the E^dl locus was narrowed between SD02 and SD04 containing Bmabd-A,and marker SD03 was tightly linked with the E^dl locus.Collectively,E^dl may be the only recessive mutant reported in E group.2.Fine mapping and molecular analysis of the E^dl mutantFor fine mapping,we narrowed down the E^dl locus to approximately a 211 Kb region between D4 and D6 using 1205 BC1 M individuals and newly designed markers.The E^dl locus was in the intergenic region between Bmabd-A and Bmabd-B,which was approximately 10 Kb upstream of the Bmabd-A and approximately 100 Kb downstream of the Bmabd-B.As a result,no protein coding gene was predicted in this region?silk DB?.Considering the phenotype of E^dl mutant,we speculated Bmabd-A as the candidate gene.The 20 stage of embryo is critical for proleg development,we detected the gene expression in this stage and found that Bmabd-A involved in proleg development was down-regulated.Neighboring genes in Hox gene cluster can interact with each other,and posterior Hox genes on the genome can inhibit the expression of the anterior gene.We investigated the expression of Hox genes adjacent to Bmabd-A,and found that the expression of Bm Ubx?anterior gene?and Bmabd-B?posterior gene?were both significantly increased.The development of distal limb structures was controlled by the expression of Bm Dll,we performed whole-mount Bm Dll antibody staining to detect whether its expression pattern was changed in the same stage.Bm Dll staining was not detected in the distal part of the degenerated prolegs but was normally detected in all normally developed prolegs,which was consistent with the mutant phenotype of degenerated prolegs in E^dl.We suppose that down-regulation of Bmabd-A expression may lead to the change of Bm Dll expression,and then generates the degenerated proleg phenotype in E^dl.3.Analysis of the mi RNAs and cis-elements in the E^dl mutantThe sequences in E locus are homologous to those in Drosophila BX-C,and there are mi RNAs and a large number of regulatory elements in the homologous region of Drosophila which regulate the expression of relative Hox genes.In this part,we analyzed the expression pattern of mi RNAs in E^dl locus and their relationships with the Hox genes.Morever,we also identifed the CTCF binding sites in E^dl locus and then analyzed their sequence differences and conservation.There are two mi RNA loci in E^dl locus,mi R-iab-4 and mi R-2835;and three mi RNAs,mi R-iab-4-3p,mi R-iab-4-5p and mi R-iab-8,were transcribed in mi R-iab-4 loucs.Firstly we compared the genomic sequences of pre-mi R-iab-4 and pre-mi R-2835 and found that those sequences in E^dl were identical to Dazao.We then predicted the target sites of mi RNA in Bm Ubx and Bmabd-A sequences by different softwares.The predictions by RNAhybrid software showed that mi R-iab-4-3p had 4 and 5 target sites on Bm Ubx and Bmabd-A sequences,respectively;mir-iab-4-5p had 4 target sites on Bm Ubx sequence but no target site in Bmabd-A sequence;mi R-iab-8 had only 1 target sites both on Bm Ubx and Bmabd-A sequences,and mi R-2835 had 4 and 1 target sites on Bm Ubx and Bmabd-A sequences.The predictions by mi Rnada software showed that all mi RNAs had no sites on Bm Ubx,while mi R-iab-4-3p,mi R-iab-8 and mi R-2835 had 1,2,and 1 target sites on Bmabd-A,respectively.The predictions by PITA software showed that only mi R-iab-8 had 1 target site while other mi RNAs had no targets on Bm Ubx;mi R-iab-4-3p,mi R-iab-4-5p,mi R-iab-8 and mi R-2835 had 2,1,4,and 3 target sites on Bmabd-A,respectively.The target sites of mi R-iab-8 and mi R-2835 on Bmabd-A predicted by mi Rnada software were covered by those predicted by PITA software,however no target site was predicted by all the three softwares.The silkworm prolegs develop in the embryonic stage,we detected the expression patterns of mi R-iab-4-3p,mi R-iab-4-5p,mi R-2835,Bm Ubx and Bmabd-A in Dazao embryos and calculated the correlation between the three mi RNAs and Bm Ubx and Bmabd-A.Resultly,we found that no correlations between the three mi RNAs and two Hox genes in the expression pattern of embryo stage were detected.CTCF binding sites are important cis elements in the Hox cluster,playing a critical role in regulating Hox gene expression.We predicted the CTCF binding sites in E^dl locus and found 8 CTCF binding sites.Agarose gel electrophoresis showed that only the second site was polymorphic but not specific in multi-strains.Conservative analysis of the 15 bp CTCF binding sites and ±50 bp sequences along the binding sites in multi-strains containing both domestic and wild silkworm,showed that 2nd,6th,7th and 8th sites were conserved while others were unconserved.Sequence alignment analysis of the CTCF binding sites in the E^dl mutant found that only the 1st and 3rd loci had Polymorphism compared with Dazao but were found to be nonspecific in multi-strains.Sequence alignment analysis of the CTCF binding sites in multi-strains found that 2nd,6th,7th and 8th sites were the most conserved with no sequence difference while the 1st and 5th sites had nonspecific sequence differences and the 3rd as well as 4th sites had specific sequence differences.Then,we analyzed the sequences in the 1st,3rd,4th,5th sites which were unconseved,and found that only the 13 th base of the 4th site changed from G to A would lead to the 4th site inactivated,other variations in all the four sites did not lead to the site inactivated.The 4th and 5th sites were in the adjacent genomic loci,and the variations in the 4th and 5th sites did not exist simultaneously in multi-strains,thus maintaining the binding ability of CTCF protein to this genomic locus.In a summary,all the CTCF binding sites except 4th site are conserved and all loci for CTCF binding in the genome are conserved in E^dl locus in multi-strains.4.Analysis of the lnc RNA in E^dl locusWe also discovered a new transcript in the intergenic region between Bmabd-A and Bmabd-B,and got the full-length sequence by Race technology.Its ORF was less than 300 nucleotides with no conserved domains.Therefore we proposed the new transcript to be a lnc RNA?Long non-coding RNA,lnc RNA?and named it lnc RNA-iab1.lnc RNA-iab1 bore a variety of splice variants which could be divided into two categories type1 and type2,according to the position of 3' terminal sequence of the different splice variants in the genome.Then we used Lnc Tar software to predict the potential interaction between lnc RNA-iab1 and Bm Ubx,Bmabd-A as well as Bmabd-B.However,it was found that all the lnc RNA-iab1 isoforms could not directly interact with Bm Ubx,Bmabd-A and Bmabd-B.Three softwares,including RNAhybrid,mi Rnada and PITA,were used to predict the target sites of mi RNAs in the two lnc RNA-iab1 isoforms.Using mi Rnada software,we found that only mi R-iab-4-5p and mi R-iab-4-3p each had 1 target sites on the type2 splice isoform of the lnc RNA-iab1.Using RNAhybrid software,we found that mi R-iab-4-5p,mi R-iab-4-3p and mi R-2835 had 1,1 and 2 binding sites on lnc RNA-iab1,respectively;mi R-iab-8 had no binding sites on lnc RNA-iab1;except one site of mi R-2835 which was only existed in the type2 splice isoform,other sites existed in both splice isoforms.When Predicted with PITA software,we found that mi R-iab-4-3p and mi R-iab-8 both had 2 target sites on the type2 splice isoform of lnc RNA-iab1 transcript,respectively.Collectively,there may be potential binding sites of these mi RNAs on lnc RNA-iab1.We detected the expression pattern between lnc RNA-iab1 and mi R-iab-4-3p,mi R-iab-4-5p as well as mi R-2835 during the embryonic stage and found that no correlations existed between lnc RNA-iab1 and mi RNAs.The correlations between lnc RNA-iab1 and Bmabd-A as well as Bmabd-B in the expression pattern of embryonic and larval stages were very strong,and then the correlations decreased during metamorphosis.The expression patterns of lnc RNA-iab1 and Bm Ubx only had a strong correlation in the embryonic stage.Tissue specific expression analysis before the 4th moulting showed that lnc RNA-iab1 was highly expressed in the nerve and epidermis with the highest expression in the epidermis;and the results showed that lnc RNA-iab1 was highly expressed in the posterior part of the epidermis?7-10 abdominal segments?,while lower expressed in the anterior part.Treatment of si RNA targeting lnc RNA-iab1 led to the death of most si RNA individuals?the proportions of lethal in total inference individuals was 13/14,12/12?,which was significantly higher than control?the proportion of lethal in total injected individuals was 1/12?.We found that the expression of lnc RNA-iab1 was significantly down-regulated in the lnc RNA-iab1 si RNA individuals,while expressions of the relative Hox gene Bmabd-A and Bmabd-B were not significantly changed.We guess that the correlations between Hox genes and lnc RNA-iab1 in the expression pattern may be related to the production process of lnc RNA-iab1 but not the effect of lnc RNA-iab1 products.lnc RNA-iab1 may be involved in other physiological functions,as the down-regulation of lnc RNA-iab1 expression lead to individual death.Only the type2 splice isoform of lnc RNA-iab1 was detected in the E^dl mutant,and some of the type2 splice isoform in E^dl did not produce the same 3' end as in Dazao,and one of the type2 splice isoform which partly overlapped with Bmabd-A transcript was also not detected.There was no difference in the transcription level of lnc RNA-iab1 between Dazao and E^dl mutant at 20 stage,a critical period for proleg development.Because only the type2 splice isoform of lnc RNA-iab1 was detected in the E^dl mutant,we detected the transcription level of type2 splice isoform in Dazao and E^dl mutant and found that the transcription level of type2 splice isoform in E^dl was significantly higher than that in Dazao.The genomic region covered by the type2 splice isoform is much larger than that of the type1 splice isoform,and could influence more functional components including two mi RNA loci,probably having a greater impact on the expression of Hox genes.Only the 3' end of type1 splice isoform is located in the E^dl locus,we speculate that mutation of some functional components in the E^dl locus leads to the failure of transcription termination at the 3' end of type1 splice isoform and then only produces the type2 splice isoform.This may also be beneficial for explaining the formation of the E^dl mutation.