Novel exon combinations generated by alternative splicing of gene fragments mobilized by a CACTA transposon in Glycine max

Abstract

Background: The recent discoveries of transposable elements carrying host gene fragments such as the Pack-MULEs (Mutator-like transposable elements) of maize (Zea mays), rice (Oryza sativa) and Arabidopsis thaliana, the Helitrons of maize and the Tgm-Express of soybeans, revealed a widespread genetic mechanism with the potential to rearrange genomes and create novel chimeric genes affecting genomic and proteomic diversity. Not much is known with regard to the mechanisms of gene fragment capture by those transposon elements or the expression of the captured host gene fragments. There is some evidence that chimeric transcripts can be assembled and exist in EST collections.

Results: We report results obtained from analysis of RT-PCR derived cDNAs of the Glycine max mutant flower color gene, wp, that contains a 5.7-kb transposon (Tgm-Express1) in Intron 2 of the flavanone 3-hydroxylase gene (F3H) and is composed of five unrelated host gene fragments. The collection of cDNAs derived from the wp allele represents a multiplicity of processed RNAs varying in length and sequence that includes some identical to the correctly processed mature F3H transcript with three properly spliced exons. Surprisingly, the five gene fragments carried by the Tgm-Express1 were processed through complex alternative splicing as additional exons of the wp transcript.

Conclusion: The gene fragments carried by the Tgm inverted repeat ends appear to be retained as functional exons/introns within the element. The spliceosomes then select indiscriminately the canonical intron splice sites from a pre-mRNA to assemble diverse chimeric transcripts from the exons contained in the wp allele. The multiplicity and randomness of these events provide some insights into the origin and mechanism of alternatively spliced genes.

Figure 1

Illustration of the effect of wp on flower and seed coat phenotypes. (A) Stable purple flower of plants with Wp genotype (left panel) or stable pink flower of plants with wp genotype (right panel) in lines LN89-5320-6 (iⁱRtW1Wp) and LN89-5322-2 (iⁱRtW1wp) both of which have yellow seed coats. In soybean I (CHS), R and T (F3'H) are three independent loci that control pigmentation in seed coats and W1 (F5'3'H) and Wp (F3H) were described as flower color markers, but all five loci seem to be encoding genes of the anthocyanin and proanthocyanidin pathways. Mutant alleles of those loci (i, iⁱ, r, t, w1 and wp) affect flower, seed coat, hypocotyle or pubescence coloration [1, 28, 30, 31]. (B) Imperfect black color of seed coats of plants with iRtW1Wp genotype (left panel) as contrasted with the lighter shaded seed coats of plants with iRtW1wp genotype (right panel). Effect on the seed coat phenotype was revealed by crossing the wp allele into lines having the recessive i allele that allows spatial pigmentation of the entire seed coat [24]. The cracks on both seed coat types result from an epistatic effect of t [31]. (C) Black seed coats of plants with iRTW1Wp genotype where the T allele drives the synthesis of cyanidins (left panel) contrasted with the lighter seed coats of plants with iRTW1wp genotype (right panel). (D) Abbreviated schematic representation of the three branches leading to the synthesis of the three anthocyanin classes and the genes encoding the enzymes relevant to the present study.

Figure 2

Variant flavanone 3-hydroxylase cDNAs from isolines containing mutant wp alleles. (A) Ethidium bromide-stained gel showing an array of cDNA bands between 5 and 1.4 kb in size that were amplified from RNAs of seed coats of the wp mutant line LN89-5322-2 through RT-PCR reactions. The (+) and (-) at top indicate reactions with and without Superscript RTII. The bright broad bands obtained from mutant RNA samples in the (+) reactions were resolved into a group of discreet bands with shorter photographic exposure of the same gel (far left lane). (B) Ethidium bromide-stained gel showing cDNAs amplified via RT-PCR with RNA from cotyledons of the LN89-5322-2 (wp) mutant line.

Figure 3

Schematic representation of the wp recessive allele and the novel exon combinations generated in its transcribed RNAs. (A) Represents the genomic sequence of the mutant wp allele obtained from the line LN89-5322-2 (iⁱRtW1wp). The introns are indicated and their length given in bp. The 5,725 bp Tgm-Express insertion in Intron 2 is drawn at top with the arrow heads representing inverted repeats and the five captured gene fragments color coded. The full length of the mutant gene is 9,251 bp. The three Exons in purple represent the cDNA of the proper spliced wild type gene 1,422 bp in size. The 7F and 1428R primers used in the PCR reactions that generated the chimeric cDNA clones shown in Figure 3B and C map at the 5' end of Exon 1 (7F) and the 3' end of Exon 3 (1428R) respectively. (B) Graphic representation of six chimeric, multi-exon cDNA clones (wp-25s, -22s, -28s, -9s, 12s, -4s) derived from seed coat RNAs of the wp mutant line via RT-PCR. These clones contained besides the F3H three Exons (1, 2, 3) varying numbers of alternatively spliced exons (solid color boxes) and introns (dashed narrower boxes) from 3 or 5 of the Tgm-Express1 captured gene fragments (UP, CDC2, FPK, M and CS). A seventh cDNA clone, wp-15s, also derived from the mutant wp line is composed only of the wild type gene (Wp) Exons 1, 2 and 3. (C) Six chimeric cDNA clones (wp-9c, -8c, -2c, -13c, -12c, -6c) derived from cotyledon RNAs of the wp mutant line via RT-PCR. All clones contained the F3H Exons (1, 2,3) with varying numbers of alternatively spliced exonic and intronic regions from the Tgm-Express1 acquired host-gene fragments separating the Exon 2-Exon 3 junction. Abbreviations: UP, unknown protein; CDC2, cell division cycle 2; FPK, fructose-6-phosphate 2-kinase/fructose-2-6-biphosphatase; M, malate dehydrogenase; CS, cysteine synthase. Two CDC2 intronic regions captured by the transposon element and sandwished between the three exonic regions (C, D and C2, Figure 2A) were spliced out to form the CDC2 exon in the chimeric transcripts (Figure 2B and C). One FPK intronic fragment captured by the transposon between two flanking exons (F and PK, Figure 2A) was also spliced out to form the FPK exon in the chimeric transcripts (Figure 2B and C). A smaller FPK intron flanked by 15 bp exon fragment (narrow orange block not named) at the 5'end (Figure 2A) is not always spliced out (Figure 2B and C).

Figure 4

Schematic of relevant chimeric and non-chimeric orfs generated by the wp allele. In order of decreasing aa length the chimeric orfs from several of the chimeric mRNAs isolated were: A (418 aa) containing F3H Exons 1 and 2, the UP and CDC2 Exons; C (293 aa) with F3H Exon 1, 3'end 16 aa, and Exon 2 plus UP and CDC2 Exons; D (143 aa) had 9 aa of the UP intron plus the UP and CDC2 Exons. E (105 aa) with 34 aa of the FPK/MDH Intron plus 71 aa of the FPK Exon; F (105 aa) had 19 aa of FPK Intron, 77 aa of FPK Exon and 9 aa of MDH Exon. The non chimeric orf B (394 aa) had the three F3H Exons identical to the ones translated from the Wp allele the only cDNA clone with this orf was wp-15s. The chimeric orfs were generated from several of the cDNAs sequenced and they are listed underneath each orf class and also the frame in each one of the clones. * The orf from the wp-6c clone was 5 aa shorter. ** The orf from the wp-6c was 2 aa longer.

Figure 5

Expression of the Tgm-Express1 captured gene fragments and related host genes in the Wp-flower color isolines. (A) RNA gel blot with total RNA samples purified from flower buds, seed coats of three developmental stages and cotyledons of two developmental stages of the flower color isolines: LN89-5320-8-53 (wp^m wp^m), LN89-5320-6 (WpWp) and LN89-5322-2 (wpwp). Seed fresh weight of each seed coat and cotyledon sample in mg is shown at bottom. The chimeric DNA probe containing UP, CDC2, FPK, M and CS sequences was an amplification product from the seed coat RT-PCR derived wp-12 cDNA clone (Figure 3). (B) Ethidium bromide-stained gel prior to membrane transfer. The 25 S rRNA is shown to compare RNA sample loading.