A bifunctional locus (BIO3-BIO1) required for biotin biosynthesis in Arabidopsis

Abstract

We identify here the Arabidopsis (Arabidopsis thaliana) gene encoding the third enzyme in the biotin biosynthetic pathway, dethiobiotin synthetase (BIO3; At5g57600). This gene is positioned immediately upstream of BIO1, which is known to be associated with the second reaction in the pathway. Reverse genetic analysis demonstrates that bio3 insertion mutants have a similar phenotype to the bio1 and bio2 auxotrophs identified using forward genetic screens for arrested embryos rescued on enriched nutrient medium. Unexpectedly, bio3 and bio1 mutants define a single genetic complementation group. Reverse transcription-polymerase chain reaction analysis demonstrates that separate BIO3 and BIO1 transcripts and two different types of chimeric BIO3-BIO1 transcripts are produced. Consistent with genetic data, one of the fused transcripts is monocistronic and encodes a bifunctional fusion protein. A splice variant is bicistronic, with distinct but overlapping reading frames. The dual functionality of the monocistronic transcript was confirmed by complementing the orthologous auxotrophs of Escherichia coli (bioD and bioA). BIO3-BIO1 transcripts from other plants provide further evidence for differential splicing, existence of a fusion protein, and localization of both enzymatic reactions to mitochondria. In contrast to most biosynthetic enzymes in eukaryotes, which are encoded by genes dispersed throughout the genome, biotin biosynthesis in Arabidopsis provides an intriguing example of a bifunctional locus that catalyzes two sequential reactions in the same metabolic pathway. This complex locus exhibits several unusual features that distinguish it from biotin operons in bacteria and from other genes known to encode bifunctional enzymes in plants.

Figure 1.

BIO1 gene identification and nature of the point mutation in bio1-1. The 3′ end of the last intron of At5g57590 in wild-type plants (TTTCAG) is modified in bio1-1 homozygotes (TTTCAA). The 5′ end of the last exon (GTAT) remains unchanged. Refer to Supplemental Figure S1 for additional details on the location of this sequence polymorphism. A, Sequencing of genomic DNA from wild-type plants reveals a G nucleotide at the mutation site. B, Genomic DNA from heterozygotes yields a doublet peak that results from the expected mixture of A and G nucleotides at the mutation site. C, Rescued homozygotes exhibit a single peak, consistent with the G to A substitution. D, RT-PCR products obtained from this region demonstrate that transcripts from rescued homozygotes (lane 3) are longer than normal because they include the final intron (confirmed by sequencing) not found in transcripts from wild-type plants (lane 2). The five smallest bands in the DNA ladder (lane 1) range from 100 to 500 nucleotides.

Figure 2.

Genome annotation and mutation sites for the BIO3-BIO1 region. Two separate genes (BIO3 short and BIO1 long) are predicted at TAIR (www.arabidopsis.org). Bicistronic cDNA has the potential to encode two distinct proteins (BIO3 long and BIO1 short). Monocistronic cDNA contains a single ORF that encodes a bifunctional fusion protein. Monocistronic (−10) and bicistronic (+10) transcripts differ with respect to the presence or absence of 10 nucleotides (arrow) at the end of intron 4. White triangles represent insertion sites for T-DNA mutant alleles. Horizontal arrows designate the locations of flanking sequences obtained. The asterisk marks a single nucleotide substitution that disrupts splicing of the last intron in the bio1-1 allele. The final 33 nucleotides of BIO1 exon 7 in the TAIR model (white rectangle) are differentially spliced and are not present in the full-length cDNAs.

Figure 3.

Responses of mutant embryos in culture. Immature embryos were removed from heterozygous siliques, plated on agar medium containing the supplements noted, and observed after the specified number of days in culture. Top, Dual images of three embryos at two different time points. DAPA, a biotin intermediate, was expected to rescue bio1 embryos but not bio3 embryos. DTB, a later intermediate, was expected to rescue bio3 embryos. See Table V for additional details. Scale bar = 1 mm.

Figure 4.

Region of the Arabidopsis genome spanning the BIO3-BIO1 junction. Exons are shown in capital letters (orange) and introns in lower case (blue). Boxed and underlined sequences identify potential start and stop codons for bicistronic and single gene transcripts. The long BIO1 start site is located upstream of the long BIO3 stop site. Alternative BIO3 stop and BIO1 start sites for translation are underlined. The large boxed (+10) region is differentially spliced from the full-length monocistronic transcript (GenBank accession no. EU089963). RT-PCR primers are noted beneath the sequence with black (BIO3, forward), green (BIO1, long), violet (BIO1, short), and red (BIO3, reverse) arrows. Flanking sequences from bio3-3 begin at green (Oklahoma State sequence) and red (Salk sequence) arrowheads located within the second exon.

Figure 5.

RT-PCR confirmation of (+10) and (−10) transcripts. A, One strategy used a reverse primer (underlined) that spanned the fourth and fifth exons and skipped the 10 nucleotides (red) that are alternatively spliced. Sequencing confirmed that the single product obtained was derived from the (−10) transcript. B, A second strategy used a reverse primer located in a downstream exon. As expected, two products that differed in length by 10 nucleotides were obtained from leaves (L), flowers (F), and siliques (S). Sequencing confirmed that these products differed with respect to the 10 nucleotides in question. A small amount of genomic DNA was also amplified. C, Semiquantitative RT-PCR analysis of the (+10; white rectangles) and (−10; black rectangles) products obtained using a reverse primer that spanned two downstream exons. Ubiquitin served as an internal standard. Band intensities were quantified using ImageJ (http://rsb.info.nih.gov/ij) and normalized relative to the maximal intensity in the flower sample.

Figure 6.

Genomic organization of biotin biosynthetic genes in microorganisms. Orthologs are depicted using the same color. Gene names (bio; BIO) and directions of transcription (pointed edge) are noted. Adjacent genes are abutted, genes that are close but not adjacent are joined by a thin line, and unlinked genes are separated by a hatched line. Green rectangles represent dissimilar genes involved in the biosynthesis of the initial biotin precursors. The BIO5 gene of yeast is involved in transport. A BIO3-BIO1 fusion protein is found in some fungi but not in yeasts.

Figure 7.

Heterologous expression (A and B) and functional characterization (C–H) of BIO3-BIO1 gene products in E. coli. A and B, Protein extracts from isopropylthio-β-galactoside-induced E. coli cultures harboring either the empty vector (pDEST17) or the BIO3-BIO1 bicistronic (+10) or monocistronic (−10) full-length cDNA were subjected to SDS-PAGE and stained with Coomassie Blue. Putative BIO3 (A) and fusion (B) proteins are marked with arrows. C and D, Expected responses of wild-type (WT) and mutant strains of E. coli in the presence and absence of biotin. E and F, Responses of a wild-type control strain and bioD (E) and bioA (F) mutants transformed with either the pDEST17 (empty) vector or recombinant vectors containing the (−10) or (+10) cDNA. G and H, Responses of wild-type and transformed bioD (G) and bioA (H) strains in liquid cultures. E to H, Strains were lysogenized with λDE3 and cultured on a kanamycin medium without biotin.

Figure 8.

Summary of expression data for BIO genes of Arabidopsis. Results of microarray experiments were obtained from https://www.genevestigator.ethz.ch. BIO3 expression levels are low in all tissues examined.

Figure 9.

Semiquantitative RT-PCR evidence of BIO1 and BIO3 single gene transcripts in extracts from wild-type plants. BIO1 forward primers were located either in the predicted 5′-UTR for the long transcript (lanes 1 and 2) or short transcript (lanes 4 and 5) or upstream (U) of the 5′-UTR for the short transcript (lane 3), but in all cases within an intron for the BIO3-BIO1 transcript. The BIO3 reverse primer was located in the predicted 3′-UTR for the long transcript but in an intron for the BIO3-BIO1 transcript. The lower band in lanes 6 and 7 was confirmed by sequencing to be the expected product. This band is less abundant than the BIO1 (long) product (lanes 1 and 2) when gels are run under equivalent conditions. The top band in lanes 6 and 7 represents contaminating genomic DNA. Plant extracts were prepared from leaves (lanes 1–6) and flowers (lane 7).