共查询到20条相似文献,搜索用时 46 毫秒
1.
The maT clade of transposons is a group of transposable elements intermediate in sequence and predicted protein structure to mariner and Tc transposons, with a distribution thus far limited to a few invertebrate species. We present evidence, based on searches of publicly available databases, that the nematode Caenorhabditis briggsae has several maT-like transposons, which we have designated as CbmaT elements, dispersed throughout its genome. We also describe two additional transposon sequences that probably share their evolutionary history with the CbmaT transposons. One resembles a fold back variant of a CbmaT element, with long (380-bp) inverted terminal repeats (ITRs) that show a high degree (71%) of identity to CbmaT1. The other, which shares only the 26-bp ITR sequences with one of the CbmaT variants, is present in eight nearly identical copies, but does not have a transposase gene and may therefore be cross mobilised by a CbmaT transposase. Using PCR-based mobility assays, we show that CbmaT1 transposons are capable of excising from the C. briggsae genome. CbmaT1 excised approximately 500 times less frequently than Tcb1 in the reference strain AF16, but both CbmaT1 and Tcb1 excised at extremely high frequencies in the HK105 strain. The HK105 strain also exhibited a high frequency of spontaneous induction of unc-22 mutants, suggesting that it may be a mutator strain of C. briggsae. 相似文献
2.
Halaimia-Toumi N Casse N Demattei MV Renault S Pradier E Bigot Y Laulier M 《Journal of molecular evolution》2004,59(6):747-760
Mariner-like elements (MLEs) are classII transposons with highly conserved sequence properties and are widespread in the genome of animal species living in continental environments. We describe here the first full-length MLE found in the genome of a marine crustacean species, the deep-sea hydrothermal crab Bythograea thermydron (Crustacea), named Bytmar1. A comparison of its sequence features with those of the MLEs contained in the genomes of continental species reveals several distinctive characteristics. First, Bytmar1 elements contains an ORF that may encode three transposase isoforms 349, 379, and 398 amino acids (aa) in long. The two biggest proteins are due to the presence of a 30- and 49-aa flag, respectively, at the N-terminal end of the 349-aa cardinal MLE transposase. Their GC contents are also significantly higher than those found in continental MLEs. This feature is mainly due to codon usage in the transposase ORF and directly interferes with the curvature propensities of the Bytmar1 nucleic acid sequence. Such an elevated GC content may interfere with the ability of Bytmar 1 to form an excision complex and, in consequence, with its efficiency to transpose. Finally, the origin of these characteristics and their possible consequences on transposition efficiency are discussed.Reviewing Editor: Dr. Nicolas Galtier 相似文献
3.
Mariner-like elements (MLEs) are the most diverse and widespread transposable elements, with members of the MLE superfamily found
in fungi, plants, ciliates and animals. In a previous study, we characterized 82 MLE transposase gene fragments (average length
383 bp) in 44 bamboo species, indicating that MLEs are widespread, abundant and diverse in the Bambusoideae subfamily. In
this study, we isolated 79 full-length MLE transposase genes from 63 bamboo species representing 38 genera in six subtribes
mainly found in China. The transposases were highly conserved, mostly uniform in length and contained intact DNA-binding motifs
and DD39D catalytic domains with few notable frameshift, indel and nonsense mutations. This suggested the MLEs are probably
still mobile, not yet affected by vertical inactivation. A phylogenetic tree of the Bambusoideae subfamily established using
ribosomal DNA internal transcribed spacer sequences was incongruent with a second tree based on the MLE transposase genes.
This evidence, together with the presence of near-identical MLEs in distantly related species and diverse MLEs in closely
related species, indicates that MLEs have evolved in a distinct manner, probably independently of speciation events in the
subfamily. The evolution and diversity of MLE transposase genes in the Bambusoideae subfamily is discussed. 相似文献
4.
Sinorhizobium meliloti natural populations show a high level of genetic polymorphism possibly due to the presence of mobile genetic elements such as insertion sequences (IS), transposons, and bacterial mobile introns. The analysis of the DNA sequence polymorphism of the nod region of S. meliloti pSymA megaplasmid in an Italian isolate led to the discovery of a new insertion sequence, ISRm31. ISRm31 is 2,803 bp long and has 22-bp-long terminal inverted repeat sequences, 8-bp direct repeat sequences generated by transposition, and three ORFs (A, B, C) coding for proteins of 124, 115, and 541 amino acids, respectively. ORF A and ORF C are significantly similar to members of the transposase family. Amino acid and nucleotide sequences indicate that ISRm31 is a member of the IS66 family. ISRm31 sequences were found in 30.5% of the Italian strains analyzed, and were also present in several collection strains of the Rhizobiaceae family, including S. meliloti strain 1021. Alignment of targets sites in the genome of strains carrying ISRm31 suggested that ISRm31 inserts randomly into S. meliloti genomes. Moreover, analysis of ISRm31 insertion sites revealed DNA sequences not present in the recently sequenced S. meliloti strain 1021 genome. In fact, ISRm31 was in some cases linked to DNA fragments homologous to sequences found in other rhizobia species. 相似文献
5.
In this review, we focus on the assembly of DNA/protein complexes that trigger transposition in eukaryotic members of the
IS630–Tc1–mariner (ITm) super-family, the Tc1- and mariner-like elements (TLEs and MLEs). Elements belonging to this super-family encode transposases with DNA binding domains of different
origins, and recent data indicate that the chimerization of functional domains has been an important evolutionary aspect in
the generation of new transposons within the ITm super-family. These data also reveal that the inverted terminal repeats (ITRs) at the ends of transposons contain three kinds
of motif within their sequences. The first two are well known and correspond to the cleavage site on the outer ITR extremities,
and the transposase DNA binding site. The organization of ITRs and of the transposase DNA binding domains implies that differing
pathways are used by MLEs and TLEs to regulate transposition initiation. These differences imply that the ways ITRs are recognized
also differ leading to the formation of differently organized synaptic complexes. The third kind of motif is the transposition
enhancers, which have been found in almost all the functional MLEs and TLEs analyzed to date. Finally, in vitro and in vivo
assays of various elements all suggest that the transposition initiation complex is not formed randomly, but involves a mechanism
of oriented transposon scanning.
Electronic Supplementary Material Supplementary material is available to authorised users in the online version of this article at .
An erratum to this article can be found at 相似文献
6.
The left (5) inverted terminal repeat (ITR) of the Mos1 mariner transposable element was altered by site-directed mutagenesis so that it exactly matched the nucleotide sequence of the right (3) ITR. The effects on the transposition frequency resulting from the use of two 3 ITRs, as well as those caused by the deletion of internal portions of the Mos1 element, were evaluated using plasmid-based transposition assays in Escherichia coli and Aedes aegypti. Donor constructs that utilized two 3 ITRs transposed with greater frequency in E. coli than did donor constructs with the wild-type ITR configuration. The lack of all but 10 bp of the internal sequence of Mos1 did not significantly affect the transposition frequency of a wild-type ITR donor. However, the lack of these internal sequences in a donor construct that utilized two 3 ITRs resulted in a further increase in transposition frequency. Conversely, the use of a donor construct with two 3 ITRs did not result in a significant increase in transposition in Ae. aegypti. Furthermore, deletion of a large portion of the internal Mos1 sequence resulted in the loss of transposition activity in the mosquito. The results of this study indicate the possible presence of a negative regulator of transposition located within the internal sequence, and suggest that the putative negative regulatory element may act to inhibit binding of the transposase to the left ITR. The results also indicate that host factors which are absent in E. coli, influence Mos1 transposition in Ae. aegypti.Communicated by G. P. Georgiev 相似文献
7.
Stefanie Kimbacher Ingrid Gerstl Branko Velimirov Sylvia Hagemann 《Molecular genetics and genomics : MGG》2009,282(2):165-172
P transposons belong to the eukaryotic DNA transposons, which are transposed by a cut and paste mechanism using a P-element-coded transposase. They have been detected in Drosophila, and reside as single copies and stable homologous sequences in many vertebrate species. We present the P elements Pcin1, Pcin2 and Pcin3 from Ciona intestinalis, a species of the most primitive chordates, and compare them with those from Ciona savignyi. They showed typical DNA transposon structures, namely terminal inverted repeats and target site duplications. The coding
region of Pcin1 consisted of 13 small exons that could be translated into a P-transposon-homologous protein. C. intestinalis and C. savignyi displayed nearly the same phenotype. However, their P elements were highly divergent and the assumed P transposase from C. intestinalis was more closely related to the transposase from Drosophila melanogaster than to the transposase of C. savignyi. The present study showed that P elements with typical features of transposable DNA elements may be found already at the base of the chordate lineage.
Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users. 相似文献
8.
Abstract
Numerous miniature inverted repeat transposable elements (MITEs) are present in the rice genome but their transposition mechanisms
are unknown. In this report, we present evidence that two novel MITE families may have arisen from Mutator-related transposable elements and thus may use a transposition mechanism similar to that of Mutator elements. Two families of novel MITEs, namely, MDM-1 and MDM-2, were identified by searching for MITEs nested with Kiddo, a previously identified MITE family. MDM-1 and MDM-2 bear hallmarks of Mutator elements, such as long terminal inverted repeats (LTIRs), 9-bp target-site duplications (TSDs), and putative transposase
binding sites. Strikingly, the MDM-1 family has a 9-bp terminus identical to that of a rice Mutator-like element (MULE-9) and the MDM-2 family has an 8-bp terminus identical to that of the maize autonomous Mutator element MuDR. A putative transposase homologous to MURA protein is identified for the MDM-2 family. Thus, these two novel MITE families, with a total copy number of several hundred in rice, are designated Mutator-derived MITEs (MDMs). Interestingly, sequence decay analysis of MDM families revealed a number of insertion site duplications (ISDs) in the alignment gaps, and widespread historical nesting
events are proposed to account for the existence of these ISDs. In addition to its value for discovering new MITEs, the nesting
analysis approach used in this study simultaneously identifies MITE insertion polymorphisms. 相似文献
9.
A family of repetitive DNA elements of approximately 350 bp—Sat350—that are members of Toxoplasma gondii satellite DNA was further analyzed. Sequence analysis identified at least three distinct repeat types within this family, called types A, B, and C. B repeats were divided into the subtypes B1 and B2. A search for internal repetitions within this family permitted the identification of conserved regions and the design of PCR primers that amplify almost all these repetitive elements. These primers amplified the expected 350-bp repeats and a novel 680-bp repetitive element (Sat680) related to this family. Two additional tandemly repeated high-order structures corresponding to this satellite DNA family were found by searching the Toxoplasma genome database with these sequences. These studies were confirmed by sequence analysis and identified: (1) an arrangement of AB1CB2 350-bp repeats and (2) an arrangement of two 350-bp-like repeats, resulting in a 680-bp monomer. Sequence comparison and phylogenetic analysis indicated that both high-order structures may have originated from the same ancestral 350-bp repeat. PCR amplification, sequence analysis and Southern blot showed that similar high-order structures were also found in the Toxoplasma-sister taxon Neospora caninum. The Toxoplasma genome database (
) permitted the assembly of a contig harboring Sat350 elements at one end and a long nonrepetitive DNA sequence flanking this satellite DNA. The region bordering the Sat350 repeats contained two differentially expressed sequence-related regions and interstitial telomeric sequences. 相似文献
10.
Jee SH Kim GE Hong SH Seo SB Shim JK Park SC Choo JK 《Molecular genetics and genomics : MGG》2007,278(4):479-486
The maT family is a unique clade within the Tc1-mariner superfamily, and their distribution is to date known as being limited to invertebrates. A novel transposon named EamaT1 is described from the genome of the earthworm Eisenia andrei. The full sized EamaT1 was obtained by degenerate and inverse PCR-based amplification. Sequence analysis of multiple copies of the EamaT1, which consisted of 0.9 and 1.4 kb elements, showed that the consensual EamaT1 with inverted terminal repeats (ITRs) of 69 bp was 1,422 bp long and flanked by a duplicated TA dinucleotide. The EamaT1 is present in approximately 120–250 copies per diploid genome but undergoes an inactivation process as a result of accumulating
multiple mutations and is nonfunctional. The open reading frame (ORF) of the EamaT1 consensus encoding 356 amino acid sequences of transposase contained a DD37D signature and a conserved paired-like DNA binding
motif for the transposition mechanism. The result of ITRs comparison confirmed their consensus terminal sequences (5′-CAGGGTG-3′)
and AT-rich region on the internal bases for ITRs-transposase interaction. 相似文献
11.
Novel<Emphasis Type="Italic"> eceriferum</Emphasis> mutants in<Emphasis Type="Italic"> Arabidopsis thaliana</Emphasis> 总被引:3,自引:0,他引:3
We conducted a novel non-visual screen for cuticular wax mutants in Arabidopsis thaliana (L.) Heynh. Using gas chromatography we screened over 1,200 ethyl methane sulfonate (EMS)-mutagenized lines for alterations in the major A. thaliana wild-type stem cuticular chemicals. Five lines showed distinct differences from the wild type and were further analyzed by gas chromatography and scanning electron microscopy. The five mutants were mapped to specific chromosome locations and tested for allelism with other wax mutant loci mapping to the same region. Toward this end, the mapping of the cuticular wax (cer) mutants cer10 to cer20 was conducted to allow more efficient allelism tests with newly identified lines. From these five lines, we have identified three mutants defining novel genes that have been designated CER22, CER23, and CER24. Detailed stem and leaf chemistry has allowed us to place these novel mutants in specific steps of the cuticular wax biosynthetic pathway and to make hypotheses about the function of their gene products.Abbreviations EMS Ethyl methane sulfonate - SEM Scanning electron microscopy - SSLP Simple sequence length polymorphism - WT Wild type 相似文献
12.
Irene C Maciariello C Micheli G Theis JF Newlon CS Fabiani L 《Molecular genetics and genomics : MGG》2007,277(3):287-299
Eukaryotic chromosomal DNA replication is initiated by a highly conserved set of proteins that interact with cis-acting elements on chromosomes called replicators. Despite the conservation of replication initiation proteins, replicator
sequences show little similarity from species to species in the small number of organisms that have been examined. Examination
of replicators in other species is likely to reveal common features of replicators. We have examined a Kluyeromyces lactis replicator, KARS12, that functions as origin of DNA replication on plasmids and in the chromosome. It contains a 50-bp region with similarity
to two other K. lactis replicators, KARS101 and the pKD1 replication origin. Replacement of the 50-bp sequence with an EcoRI site completely abrogated the ability of KARS12 to support plasmid and chromosomal DNA replication origin activity, demonstrating this sequence is a common feature of K. lactis replicators and is essential for function, possibly as the initiator protein binding site. Additional sequences up to 1 kb
in length are required for efficient KARS12 function. Within these sequences are a binding site for a global regulator, Abf1p, and a region of bent DNA, both of which
contribute to the activity of KARS12. These elements may facilitate protein binding, protein/protein interaction and/or nucleosome positioning as has been proposed
for other eukaryotic origins of DNA replication. 相似文献
13.
14.
Sophie Casteret Najat Chbab Jeanne Cambefort Corinne Augé-Gouillou Yves Bigot Florence Rouleux-Bonnin 《Molecular genetics and genomics : MGG》2009,282(5):531-546
Previous studies have shown that the transposase and the inverted terminal repeat (ITR) of the Mos1
mariner elements are suboptimal for transposition; and that hyperactive transposases and transposon with more efficient ITR configurations
can be obtained by rational molecular engineering. In an attempt to determine the extent to which this element is suboptimal
for transposition, we investigate here the impact of the three main DNA components on its transposition efficiency in bacteria
and in vitro. We found that combinations of natural and synthetic ITRs obtained by systematic evolution of ligands by exponential
enrichment did increase the transposition rate. We observed that when untranslated terminal regions were associated with their
respective natural ITRs, they acted as transposition enhancers, probably via the early transposition steps. Finally, we demonstrated
that the integrity of the Mos1 inner region was essential for transposition. These findings allowed us to propose prototypes of optimized Mos1 vectors, and to define the best sequence features of their associated marker cassettes. These vector prototypes were assayed
in HeLa cells, in which Mos1 vectors had so far been found to be inactive. The results obtained revealed that using these prototypes does not circumvent
this problem. However, such vectors can be expected to provide new tools for the use in genome engineering in systems such
as Caenorhabditis elegans in which Mos1 is very active. 相似文献
15.
Uchino K Imamura M Shimizu K Kanda T Tamura T 《Molecular genetics and genomics : MGG》2007,277(3):213-220
We investigated the use of Minos as a vector for transgenesis in the silkworm, Bombyx mori. We first constructed a vector plasmid with the green fluorescent protein (GFP) gene fused with the silkworm cytoplasmic
actin gene (A3) promoter, and a helper plasmid with the Minos transposase gene controlled by the same A3 promoter. Injection of the vector and helper plasmid DNA into silkworm eggs produced
transgenic animals in the following generation. The efficiency of transgenic silkworm production using this method was much
lower than that obtained using piggyBac-mediated germ line transformation. However, >40-fold increase in the efficiency of producing transgenic silkworms was obtained
using an in vitro synthesized source of Minos transposase mRNA. We conclude that the Minos transposon is a useful vector for construction of transgenic silkworms, particularly when in vitro synthesized mRNA is used.
This is the first report showing that Minos can be used as a vector for germ-line transformation in lepidopteran insects. 相似文献
16.
A miniature inverted-repeat transposable element (MITE), designated as Hikkoshi, was previously identified in the null Wx-A1 allele of Turkish bread wheat lines. This MITE is 165 bp in size and has 12-bp terminal inverted repeats (TIRs) flanked by 8-bp target site duplications (TSDs). Southern and PCR analyses demonstrated the presence of multiple copies of Hikkoshi in the wheat genome. Database searches indicated that Hikkoshi MITEs are also present in barley, rice and maize. A 3.4-kb element that has Hikkoshi-like TIRs flanked by 8-bp TSDs has now been identified in the rice genome. This element shows high similarity to the 5 subterminal region of the wheat Hikkoshi MITE and contains a transposase (TPase) coding region. The TPase has two conserved domains, ZnF_TTF and hATC, and its amino acid sequence shows a high degree of homology to TPases encoded by Tip100 transposable elements belonging to the hAT superfamily. We designated the 3.4-kb element as OsHikkoshi. Several wheat clones deposited in EST databases showed sequence similarity to the TPase ORF of OsHikkoshi. The sequence information from the TPase of OsHikkoshi will thus be useful in isolating the autonomous element of the Hikkoshi system from wheat. 相似文献
17.
The evolutionary history of mariner-like elements (MLEs) in 49 mainly Neotropical drosophilid species is described. So far, the investigations about the distribution
of MLEs were performed mainly using hybridization assays with the Mos1 element (the first mariner active element described) in a widely range of drosophilid species and these sequences were found principally in species
that arose in Afrotropical and Sino-Indian regions. Our analysis in mainly Neotropical drosophilid species shows that twenty-three
species presented MLEs from three different subfamilies in their genomes: eighteen species had MLEs from subfamily mellifera, fifteen from subfamily mauritiana and three from subfamily irritans. Eleven of these species exhibited elements from more than one subfamily in their genome. In two subfamilies, the analyzed
coding region was uninterrupted and contained conserved catalytic motifs. This suggests that these sequences were probably
derived from active elements. The species with these putative active elements are Drosophila mediopunctata and D. busckii for the mauritiana subfamily, and D. paramediostriata for the mellifera subfamily. The phylogenetic analysis of MLE, shows a complex evolutionary pattern, exhibiting vertical transfer, stochastic
loss and putative events of horizontal transmission occurring between different Drosophilidae species, and even those belonging
to more distantly related taxa such as Bactrocera tryoni (Tephritidae family), Sphyracephala europaea (Diopsoidea superfamily) and Buenoa sp. (Hemiptera order). Moreover, our data show that the distribution of MLEs is not restricted to Afrotropical and Sino-Indian
species. Conversely, these TEs are also widely distributed in drosophilid species arisen in the Neotropical region. 相似文献
18.
Arabidopsis ACT2 represents an ancient class of vegetative plant actins and is strongly and constitutively expressed in almost all Arabidopsis sporophyte vegetative tissues. Using the beta glucuronidase report system, the studies showed that ACT2 5′ regulatory region was significantly more active than CaMV 35S promoter in Arabidopsis seedlings and gametophyte vegetative tissues of Physcomitrella patens. Its activity was also observed in rice and maize seedlings. Thus, the ACT2 5′ regulatory region could potentially serve as a strong regulator to express a transgene in divergent plant species. ACT2 5′ regulatory region contained 15 conserved sequence elements, an ancient intron in its 5′ un-translated region (5′ UTR),
and a purine-rich stretch followed by a pyrimidine-rich stretch (PuPy). Mutagenesis and deletion analysis illustrated that
some of the conserved sequence elements and the region containing PuPy sequences played regulatory roles in Arabidopsis. Interestingly, mutation of the conserved elements did not lead a dramatic change in the activity of ACT2 5′ regulatory region. The ancient intron in ACT2 5′ UTR was required for its strong expression in both Arabidopsis and P. patens, but did not fully function as a canonical intron. Thus, it was likely that some of the conserved sequence elements and gene
structures had been preserved in ACT2 5′ regulatory region over the course of land plant evolution partly due to their functional importance. The studies provided
additional evidences that identification of evolutionarily conserved features in non-coding region might be used as an efficient
strategy to predict gene regulatory elements. 相似文献
19.
Wu S Yu Z Wang F Li W Ye C Li J Tang J Ding J Zhao J Wang B 《Molecular biotechnology》2007,36(2):102-112
N-methylation of phosphoethanolamine, the committing step in choline (Cho) biosynthesis in plants, is catalyzed by S-adenosyl-l-methionine: phosphoethanolamine N-methyltransferase (PEAMT, EC 2.1.1.103). Herein we report the cloning and characterization of the novel maize phosphoethanolamine
N-methyltransferase gene (ZmPEAMT1) using a combination of bioinformatics and a PCR-based allele mining strategy. The cDNA sequence of ZmPEAMT1 gene is 1,806 bp in length and translates a 495 amino acids peptide. The upstream promoter sequence of ZmPEAMT1 were obtained by TAIL-PCR, and contained four kinds of putative cis-acting regulatory elements, including stress-responsive elements, phytohormone-responsive elements, pollen developmental
special activation elements, and light-induced signal transduction elements, as well as several other structural features
in common with the promoter of rice and Arabidopsis homologies. RT-PCR analysis showed that expression of ZmPEAMT1 was induced by salt stress and suppressed by high temperature. Over-expression of ZmPEAMT1 enhanced the salt tolerance, root length, and silique number in transgenic Arabidopsis. These data indicated that ZmPEAMT1 maybe involved in maize root development and stress resistance, and maybe having a potential application in maize genetic
engineering.
Note: Nucleotide sequence data are available in GenBank under the following accession numbers: maize (Zea mays, ZmPEAMT1, AY626156; ZmPEAMT2, AY103779); rice (Oryza sativa, OsPEAMT1/Os01g50030, NM_192178; OsPEAMT2/Os05g47540, XM_475841); wheat (Triticum aestivum, TaPEAMT, AY065971); Arabidopsis (Arabidopsis thaliana, AtNMT1/At3g18000, AY091683; AtNMT2/At1g48600, NM_202264; AtNMT3/At1g73600, NM_106018); oilseed rape (Brassica napus, BnPEAMT, AY319479), tomato (Lycopersicon esculentum, AF328858), spinach (Spinacia oleracea, AF237633). 相似文献
20.
Sarkar A Sim C Hong YS Hogan JR Fraser MJ Robertson HM Collins FH 《Molecular genetics and genomics : MGG》2003,270(2):173-180
piggyBac is a short inverted-repeat-type DNA transposable element originally isolated from the genome of the moth Trichoplusia ni. It is currently the gene vector of choice for the transformation of various insect species. A few sequences with similarity to piggyBac have previously been identified from organisms such as humans ( Looper), the pufferfish Takifugu rubripes (Pigibaku), Xenopus (Tx), Daphnia (Pokey), and the Oriental fruit fly Bactrocera dorsalis. We have now identified 50 piggyBac-like sequences from publicly available genome sequences and expressed sequence tags (ESTs). This survey allows the first comparative examination of the distinctive piggyBac transposase, suggesting that it might contain a highly divergent DDD domain, comparable to the widespread DDE domain found in many DNA transposases and retroviral integrases which consists of two absolutely conserved aspartic acids separated by about 70 amino acids with a highly conserved glutamic acid about 35 amino acids further away. Many piggyBac-like sequences were found in the genomes of a phylogenetically diverse range of organisms including fungi, plants, insects, crustaceans, urochordates, amphibians, fishes and mammals. Also, several instances of "domestication" of the piggyBac transposase sequence by the host genome for cellular functions were identified. Novel members of the piggyBac family may be useful in genetic engineering of many organisms.Electronic Supplementary Material Supplementary material is available in the online version of this article at 相似文献