Drosha-DGCR8complexandDicerrecognizetheirRNAsubstratesindifferentmannersduringmiRNAprocessingFENGYong,CHENLang,HOUWei【推薦論文】.doc

精品論文drosha-dgcr8 complex and dicer recognize their rna substrates in different manners during mirna processing feng yong, chen lang, hou wei5(wuhan university school of basic medical sciences, wuhan 430070)abstract: drosha/dgcr8 complex cleaves primary micrornas (pri-mirnas) to yield 65 nucleotide (nt) precursor micrornas (pre-mirnas). dicer cleaves pre-mirnas to yield 22 nt mirna/mirna* duplexes. interactions of drosha/dgcr8 and dicer with their substrates are incompletely understood. by in vitro drosha and dicer assays, we proposed models that human10drosha/dgcr8 and dicer recognize their substrates in different manners. the flanking sequences and the terminal loop structure of a pri-mirna are critical for drosha/dgcr8 processing pri-mirnas.the terminal loop of a pre-mirna is not essential for hdicer processing. the end side and the stem of the pre-mirna contribute to the recognizing and binding with dicer. this study improved our understanding on rnase iii functional mechanism and mirna processing.15keywords: microrna; drosha; dicer; rna processing0introductionmicrornas (mirnas) are a group of 22 nucleotide (nt) rnas that modules gene expression post-transcriptionally. in mammalian cells, most of the mirnas genes are transcribed20by rna polymerase ii as primary transcripts (pri-mirna). the pri-mirnas are processed into 65 nt stem-loop structured precursor mirna (pre-mirna) by rnase iii endonuclease drosha and its partner protein dgcr8, and then delivered to cytoplasm, where another rnase iii enzyme, dicer, cleavages the 65 nt hairpin into 22 nt dupex of mirna and mirna* 1, 2.mirna processing includes two steps of cleavage by two rnase iii, drosha and dicer,25which work in two different manners to determine two ends of the matured mirna molecule. the drosha processing step, in which dgcr8 is involved together, is critical, as it defines the sequence of mirna by cropping the stem-loop hairpin structure embedding in the pri-mirna transcript 3. recent study also demonstrated that drosha processing controls the specificity and efficiency of global mirna expressions 4. substrate structures, which include single stranded30rna extensions from the pre-mirna hairpin, are required for drosha processing. for drosha processing, it does not matter whats the sequence of these rna extensions, however, a strong secondary structure within the extension or a blunt-ended pri-mirna hairpin blocks drosha cleavage 5.dicer defines the other end of the mirna sequence by cleaving the pre-mirna hairpin into35a 22 nt small rna duplex. human dicer protein forms an “l(fā)” shaped architecture with the paz domain in the head 6. the 3 2 nt end of the pre-mirna hairpin anchors at the paz domain and the dicer protein measures 22 nt distance from the 3 end to release the rna duplex. structure requirements of substrates for dicer processing are commonly believed to compromise the 3 end overhang, the double-stranded stem and a flexible single-stranded loop. previous study40demonstrated that human dicer has a loose substrate specificity, as human dicer tolerates remarkable structural variations in its pre-mirna substrates 7.rnase iii enzymes cleave double-stranded rna and leave a 2 nt overhang on the 3 end of its products 8. as enzymes from rnase iii family, both drosha and dicer form anintramolecular dimmer of rnase iii domain and share a basic action mechanism. however,foundations: ph.d. program foundation of chinese ministry of education (no. 20090141120048) and nationalnatural science foundation of china (no. 30972754)brief author introduction:feng yong, (1976-), male, lecturer, medical virology. e-mail:- 9 -45substrates for these two rnase are quite different, especially in their secondary structures. most importantly, drosha-dgcr8 complex and dicer define each end of the mature mirna. therefore, the mechanisms by which drosha-dgcr8 and dicer recognize and bind their substrates become an essential subject, which is incompletely understood. in the current study, we presented data of different length of rna substrates processed by drosha-dgcr8 and dicer, to show the50difference of interaction manners for these two important rnase iii enzymes with their substrates.1materials and methods1.1 preparation of rna substratessequences and secondary structures of human pri-mirnas and pre-mirnas were predicted based on mirbase 9 and mfold (10, 11. dna templates were amplified by pcr, with 555primers containing the t7 promoter sequence. the pcr products were in intro transcribed into rnas by t7 rna polymerase (promega, madison, us). rnas were labeled by - 32p ctp and the probes were gel-purified before use. primers used in this study and the predicted sizes of rnas are listed in table 1.60tab.1 primers used in the study and predicted rna sizes of pri-mirnas and pre-mirnas.forward primer (from 5 to 3)reverse primer (from 5 to 3)pri-mirnasize (nt)pre-mirnasize (nt)pri-let-7a (+25)taatacgactcactataggatgttctcttcactgtgcagacttttctatcacgttag11572pri-mir-7-2(+25)taatacgactcactatagagtggaccggctggcccccgctgcgatggctggcacca10562pri-mir-7-2(+50)taatacgactcactatagaacgccttgcagaactggccaggggggctgcccccatt16162pri-mir-325(+25)taatacgactcactatagcagaaccaatacagtgcttgcagagcctagcacagtgc10560pri-mir-325(+100)taatacgactcactatagacattcacaaaggaaagaaaaatcactctactgtaaccac25660pre-let-7ataatacgactcactataggaggtagtaggttgtatagtgaaagacagtagattgtata72pre-mir-877taatacgactcactatagtagaggagatggcgcagctgggaggagggagaagag86pre-mir-554taatacgactcactatagctagtcctgactcagccaatgccaaaatatgaaccc65pre-mir-554-m1taatacgactcactatagctagaatatcctgactcagccagtaatgccaaaatatgaaccc71pre-mir-554-m2taatacgactcactatagctagtcctgactcagccaatgccaatgaaccctgacccatcac61pre-mir-554-m3taatacgactcactatagctagaatatcctgactcagccagtaatgccaatgaaccctgacccatcac651.2 drosha cleavage assaysfor a drosha cleavage assay, a 32p-labeled pri-mirna was mixed with humandrosha/dgcr8 protein mix in a total volume of 6-10 l at 37oc as previously described 4, 12.65at time point of 60 min, an equal volume of 2 x sample buffer (98% formamide, 10mm edta and 0.1% bromophenol blue) was added. after running the samples on a 12% denaturing polyacrylamide gel and fixing the gel, data were analyzed using a phosphorimager (ge healthcare) or by autoradiography.1.3 dicer cleavage assays70for a dicer cleavage assay, a 32p-labeled pre-mirna was mixed with human dicer in a total volume of 6-10 l at 37oc as described previously 7. the reaction buffer contained 20 mm tris ph 7.4, 50 mm nacl, 5% glycerol, supplemented with 3 mm mgcl2, 10 mm dtt, and rnasin (promega). at specified time point, an equal volume of 2 x sample buffer was added. after running the samples on a 12% denaturing polyacrylamide gel and fixing the gel, data were75analyzed using a phosphorimager (ge healthcare) or by autoradiography. protein of human dicer was expressed and purified from insect cells via baculovirus expression system based on the plasmid pfast-hdicer which was kindly provided by dr. j. a. doudna 13.1.4 gel shift assayshuman dicer proteins were incubated with -32p ctp labeled pre-mirnas and pri-let-7a80859095100105in binding buffer (20mm triscl ph7.4, 50 mm kcl, 0.1% tween-20, 10% glycerol) on ice for30 min. then 1l dye was added to the reaction and electrophoresis in non-denaturing 7.5%polyacrylamide gels. the dicer:rna complex was visualized by phosphorimager.2results2.1 drosha cleaves pri-mirnas with variant rna extensions into 65 nt rnahairpinwe first performed drosha assays on pri-mirna transcripts with 25 nt extensions to the pre-mirna stem-loop hairpins (fig 1a). we applied in vitro transcription to prepare the rna substrates as described previously 4. the drosha/dgcr8 complex efficiently cut these pri-mirnas into a pre-mirna and extended rna fragments (fig 1a). the band presented in the gel reflected the size of pre-mirnas products as expected, for pre-let-7a was 72 nt (fig 1a lane5,6), pre-mir-7-2 was 62 nt (lane 3,4) and pre-mir-325 60 nt (lane 1,2)(table 1).to investigate the module of dorsha/dgcr8 responses on variation of pri-mirna, especially on variations within the extended sequences to the pre-mirna stem-loop structure, we transcribed pri-mir-7-2 with 50 nt extended to the pre-mirna hairpin (pre-mir-7-2(+50) and pre-mir-325(+100) for the dorsha assays (fig 1b). we noticed that pri-mir-7-2 folded as a long stem-loop structure, while the pri-mir-325 folded as triple-connected stem-loop structure (fig 1b, fold by rna folding software indicated in the method). drosha efficiently cut pri-mir-7-2 into a60 nt pre-mirna hairpin and 50 nt fragments (fig 1b, lane 1, 2. the extended sequences to thehairpin).interestingly, drosha/dgcr8 complex cut the pri-mir-325(+100) into a single 60 nt hairpin rna and 100 nt fragments (the extended sequences to the hairpin) instead of cutting two or three hairpins with much smaller fragments (fig 1b, lane 3, 4). this observation indicated that the drosha/dgcr8 cut the pri-mir-325 only once, and this cleavage reaction occurred on the longest hairpin where the mature mir-325 exists. notably, dorsha cut pri-mir-325(+100) much less efficiently than it did on pri-mir-7-2 (fig 1b).110115120125figure 1. schematics of pri-mirna are shown on left. predicted hairpin structures of the sequences harbor themirnas. the mirna and the mirna* are represented by character and other sequences includes extended rna and loop by lines, respectively. for drosha assays, - 32p ctp-labeled pri-mirna was mixed with human drosha/dgcr8 protein complex in a total volume of 6-10 l at 37oc. an equal volume of 2 x sample buffer was added at time point of 60 min. after running the samples on a 12% denaturing polyacrylamide gel and fixing the gel, data were analyzed using a phosphorimager. arrows indicate the pre-mirna hairpin products. the dna size markers did not accurately co-migrate with rnas of the same size. a: pri-mirnas with 25 nt extended rna to the pre-mirna hairpins. b: pri-mirnas with 50 nt and 100 nt extended to the hairpins. abbreviations: pri-7-2 for pri-mir-7-2 and pri-325 for pri-mir-325. “+”: add drosha/dgcr8 protein complex; “-”: no drosha/dgcr8 added.2.2 hdicer cleaves pre-mirna hairpins into 22nt rna duplexwe performed human dicer cleavage assays on selected pre-mirnas. as in other studies, pre-let-7a, pre-mir-16 and pre-mir-30a were cleaved by human dicer (hdicer) into 22 nt mirna/mirna* duplex (fig 2a, lane 1-6). we also noticed that an 86 nt long pre-mir-877 was efficiently processed by hdicer (fig 2a, lane 7, 8), though it had a much larger size than average pre-mirna has ( 65 nt).we then asked whats the responses of human dicer (hdicer) cleavage pre-mirnas with different length additional to the canonical 65 nt pre-mirna. we performed hdicer assays on two dicer substrates, pre-let-7a (total length is 72 nt) and pri-let-7a(+25 as described in the130135140145150above, total length is 115nt). pre-let-7a was cleaved by hdicer into 22 nt rna duplex and 29 nt fragments (fig 2b, lane 7-10). pri-let-7a(+25) is a good substrate for drosha cleavage to yield a 72 nt pre-let-7a (fig 1a), however, in this part of hdicer assays, we observed that it was processed by hdicer as well (fig 2b, lane 2-5).notably, there was a band of 75 nt immediate product (ip) in the gel at the time points of 1min and 5 min (fig 2b, lane 2, 3, dashed arrows), which indicated that hdicer first cut pri-mirna(+25) to release a 22 nt rna duplex and a 75 nt pre-mirna (ip). at the time points of 30 min and 60 min, the ip was cleaved into a second 22 nt rna duplex and fragments (lane 4,5), as the hdicer did to pre-let-7a (lane 7-10). this observation demonstrated that hdicer cut pri-let-7a(+25) twice to yield two different rna duplex, which indicated that hdicer could cleave this length of rna hairpin substrate (pri-let-7a(+25) in this study) gradually.figure 2. a: schematics of selected pre-mirna are shown on left. the mirna and the mirna* are represented by character and other sequences by lines, respectively. rna substrates were in vitro transcribed and labeled by - 32p ctp then incubated with hdicer protein at 37oc for 60 min. lane 1, 2: pre-let-7a; lane 3, 4: pre-mir-16;lane 5,6: pre-mir-30a; lane 7,8: pre-mir-877. after running the samples on a 12% denaturing polyacrylamide gel and fixing the gel, data were analyzed using a phosphorimager. “+”: add hdicer protein; “-”: no hdicer protein added. b: pri-let-7a(+25) and pre-let-7a were incubated with hdicer protein at 37oc for 1, 3, 30 and 60 min. an equal volume of 2 x sample buffer was added at each time point. dashed arrows indicate the immediate products (ips) and the arrows indicate the rna duplexes. the dna size markers did not accurately co-migrate with rnas of the same size. lane 1 and lane 6, no hdicer were added.2.3 stem of pre-mirnas contributes to binding with hdicerto gain insight into hdicer and pre-mirna interaction, we performed gel shift assays to examine the hdicer:pre-mirna complex. as revealed by the gel shift assay, pri-let-7a(+25) formed a shifted complex as pre-let-7a did. in addition, pri-let-7a(+25) bound hdicer more155160165weakly than pre-let-7a did(fig 3a). in our previous work, we demonstrated that the stem of pre-mirna bound to the hdicer protein by rnase vi digestion experiments 7. here, we mutated pre-mir-554, which was found to bind hdicer very weakly 7. from the gel shift assays, we observed that mutants with regular stem had better binding affinity with hdicer (fig 3c, lane3-8), since pre-mir-554 (wild type) had a big asymmetric mismatch bugle located in the stem (fig3b). these results suggested that the stem structure contributed to the dicer-pre-mirna binding, and the binding was stronger with a perfect stem.figure 3. gel shift assays on hdicer:pre-mirna complex. 2 nm of hdicer was incubated with 1000 cpm of32p-labelled pre-mirnas in binding buffer on ice for 30 min, before fractionation on a 7.5% native gel. the arrows and stars mark the protein:rna complex and free rna, respectively. a: lane 1, 2: pri-let-7a(+25); lane 3,4: pre-let-7a. b: pre-mir-554 and its mutants. c: lane 1, 2: pre-mir-554; lane 3, 4: pre-mir-554-m1; lane 5, 6:pre-mir-554-m2; lane 7, 8: pre-mir-554-m3; lane 9, 10: pre-let-7a.1701751801851901952002052103discussiona mirna gene transcript undergoes two cleavage steps by two rnase iii enzymes, drosha and dicer, which defines each side of the mature mirna. therefore, the recognition of these two enzymes with their rna substrates and via which the two catalysis reactions take place play important roles in choosing the mirna sequence. to our knowledge, only a few works were reported about how drosha recognizes its substrate. a canonical animal pri-mirna compromises a double-stranded stem, a single-stranded loop and long flanking sequences. zeng et al proved that a large terminal loop was critical for drosha processing 12, 14 and the cleavage site might be determined 2 helical turns distance from this structure. however, han et al argued that the terminal loop was unessential to the processing and they demonstrated that dgcr8, not drosha, interacted with pri-mirna directly and specifically 3. from data we obtained in this study, we proposed our model how drosha recognizes pri-mirna (fig 4a and 4b). the dgcr8 protein is responsible for recognizing and binding to the pri-mirna, via interacting with both the terminal loop and the flanking sequences. drosha, as a small molecular weight protein, binds and locates in one side of dgcr8 where the cleavage catalysis reaction takes place. the terminal loop of pri-mirna has an interaction with dgcr8 (or drosha/dgcr8 complex) during the binding and cleavage process. this model explained why a larger terminal loop facilitated drosha/dgcr8 processing pri-mirna and why drosha/dgcr8 cut pri-mir-325(+100) only once in this study (fig 1b, lane 3, 4). its most possible that the other two stem-loop structures in pri-mir-325(+100) were too small, so that the strong secondary structure within the extension of the other two hairpins blocked drosha cleavage, therefore the terminal loop and the flanking arms could not interact with drosha/dgcr8 complex at both sides (fig 4b). in conclusion, our data showed that both the flanking sequences and the terminal loop structure of pri-mirnas are critical for drosha/dgcr8 processing.in the meanwhile, we explored the dicer processing model by dicer assays on over-sized pre-mirna. pre-mirna is typically 65 nt long and compromises a loop and a stem with a 2 nt 3 overhang structure. previously, we showed that human dicer tolerated structural variations in its pre-mirna substrates 7. it wa