Reverse-genetic analysis of the two biotin-containing subunit genes of the heteromeric acetyl-coenzyme A carboxylase in Arabidopsis indicates a unidirectional functional redundancy

Abstract

The heteromeric acetyl-coenzyme A carboxylase catalyzes the first and committed reaction of de novo fatty acid biosynthesis in plastids. This enzyme is composed of four subunits: biotin carboxyl-carrier protein (BCCP), biotin carboxylase, α-carboxyltransferase, and β-carboxyltransferase. With the exception of BCCP, single-copy genes encode these subunits in Arabidopsis (Arabidopsis thaliana). Reverse-genetic approaches were used to individually investigate the physiological significance of the two paralogous BCCP-coding genes, CAC1A (At5g16390, codes for BCCP1) and CAC1B (At5g15530, codes for BCCP2). Transfer DNA insertional alleles that completely eliminate the accumulation of BCCP2 have no perceptible effect on plant growth, development, and fatty acid accumulation. In contrast, transfer DNA insertional null allele of the CAC1A gene is embryo lethal and deleteriously affects pollen development and germination. During seed development the effect of the cac1a null allele first becomes apparent at 3-d after flowering, when the synchronous development of the endosperm and embryo is disrupted. Characterization of CAC1A antisense plants showed that reducing BCCP1 accumulation to 35% of wild-type levels, decreases fatty acid accumulation and severely affects normal vegetative plant growth. Detailed expression analysis by a suite of approaches including in situ RNA hybridization, promoter:reporter transgene expression, and quantitative western blotting reveal that the expression of CAC1B is limited to a subset of the CAC1A-expressing tissues, and CAC1B expression levels are only about one-fifth of CAC1A expression levels. Therefore, a likely explanation for the observed unidirectional redundancy between these two paralogous genes is that whereas the BCCP1 protein can compensate for the lack of BCCP2, the absence of BCCP1 cannot be tolerated as BCCP2 levels are not sufficient to support heteromeric acetyl-coenzyme A carboxylase activity at a level that is required for normal growth and development.

Figure 1.

Molecular structure of the cac1a-1, cac1b-1, and cac1b-2 mutant alleles. Exons are represented by black boxes, introns are represented by white boxes, and untranslated regions are represented by gray boxes. Arrows indicate the position of the PCR primers used to characterize each allele. A, In the cac1a-1 allele, the insertion is located in the first intron of the gene and consists of chimeric T-DNAs, in a head-to-head arrangement, so that both ends of the insert have outward-facing LB sequences. B, T-DNA insertions in the cac1b-1 (SALK_056228) and cac1b-2 (SALK_070569) alleles are located in the third intron and the first exon, respectively.

Figure 2.

Mutations in the CAC1A gene, but not in the CAC1B gene, show a seed abortion phenotype. Stereomicrographs (A, B, E–G) and scanning electron micrographs (C and D) of siliques at 9 DAF from sibling wild-type and mutant plants that were grown side by side. Bars = 350 μm (A, B, E–G); 100 μm (C and D). Seeds developing in siliques of wild-type (A and C) or sibling plants that are heterozygous for the cac1a-1 allele (B and D). White arrows identify aborted seeds, and yellow arrows identify nonfertilized ovules. Regardless of the allele carried at the CAC1B locus (homozygous wild-type [E], cac1b-1 [F], or cac1b-2 [G]), seeds develop normally.

Figure 3.

Effect of cac1b mutations on the accumulation of the htACCase subunits. Protein extracts from a mixture of flowers and flower buds were prepared from wild-type (WT), cac1b-1 homozygous, or cac1b-2 homozygous plants. Each lane was loaded with extracts containing 200 μg of total protein. Specific antisera were used to detect five htACCase subunits. In addition, streptavidin was used to detect biotinylated BCCP1 and biotinylated BCCP2 subunits. The mitochondrial MCCase subunit A was detected by its specific antibody and served as an internal control.

Figure 4.

Effect of the cac1a-1 mutation on embryo and endosperm development. Developing ovules and embryos were examined by CSLM (A–C, J–O, S–U) and SEM (D–I, P–R, V–X), at 3 DAF (A–L), 5 DAF (M–R), and 7 DAF (S–X). em, Embryo; en, endosperm; s, suspensor; oi, outer integument; ii, inner integument; ph, phragmoplasts; cw, cell wall; den, degenerated endosperm. As labeled, the development of the ovules and embryos were compared on sibling plants that were grown side by side, but carried either wild-type alleles at the CAC1A locus (WT) or were heterozygous for the cac1a-1 allele (cac1a). As described in the “Materials and Methods,” for CSLM, flowers and siliques were fixed, cleared, and the enclosed ovules examined in the intact organs. For SEM, collected flowers and siliques were fixed with glutaraldehyde/paraformaldehyde, post fixed in OsO₄, and following dehydration and drying, the ovules were fractured and examined. A to L, At 3 DAF, aberrations in ovule development are apparent in plants carrying the cac1a-1 allele. In wild-type plants, the embryos (em) are at the 32-cell stage, whereas in cac1a-1 plants 16-cell (B) and four-cell (C) embryos are present. Formation of cellular connections between sister and nonsister nuclei has begun to take place in the endosperm (en) of wild-type plants (D, G, J), but this process is delayed and less regular in aberrant ovules of cac1a-1 plants (E, F, H, I, K, L). Bars = 50 μm (A–C, J–L); or 10 μm (D–I). M to R, At 5 DAF, aberrations in ovule development become more apparent in plants carrying the cac1a-1 allele. Wild-type embryos are at the heart stage of development (M) and the endosperm has begun to cellularize (P). In cac1a-1 plants, the aberrantly developing ovules have globular embryos with highly vacuolated cells (N and O). In the most severely aberrant ovules, the embryo sac has collapsed (O). Endosperm is uncellularized or aberrantly cellularized and sometimes degenerated (den; Q and R). Bars = 50 μm (M–O); 10 μm (P–R). S to X, At 7 DAF, the aberrant ovules in the cac1a-1 plants have ceased development at the globular or heart stages (T and U), the endosperm is degenerated, and the embryo sacs are often collapsed (W and X). In contrast, in the wild-type siblings the embryos are at the torpedo stage (S) and the endosperm is fully cellularized (V). Bars = 50 μm (S–U); 10 μm (V–X).

Figure 4.

Effect of the cac1a-1 mutation on embryo and endosperm development. Developing ovules and embryos were examined by CSLM (A–C, J–O, S–U) and SEM (D–I, P–R, V–X), at 3 DAF (A–L), 5 DAF (M–R), and 7 DAF (S–X). em, Embryo; en, endosperm; s, suspensor; oi, outer integument; ii, inner integument; ph, phragmoplasts; cw, cell wall; den, degenerated endosperm. As labeled, the development of the ovules and embryos were compared on sibling plants that were grown side by side, but carried either wild-type alleles at the CAC1A locus (WT) or were heterozygous for the cac1a-1 allele (cac1a). As described in the “Materials and Methods,” for CSLM, flowers and siliques were fixed, cleared, and the enclosed ovules examined in the intact organs. For SEM, collected flowers and siliques were fixed with glutaraldehyde/paraformaldehyde, post fixed in OsO₄, and following dehydration and drying, the ovules were fractured and examined. A to L, At 3 DAF, aberrations in ovule development are apparent in plants carrying the cac1a-1 allele. In wild-type plants, the embryos (em) are at the 32-cell stage, whereas in cac1a-1 plants 16-cell (B) and four-cell (C) embryos are present. Formation of cellular connections between sister and nonsister nuclei has begun to take place in the endosperm (en) of wild-type plants (D, G, J), but this process is delayed and less regular in aberrant ovules of cac1a-1 plants (E, F, H, I, K, L). Bars = 50 μm (A–C, J–L); or 10 μm (D–I). M to R, At 5 DAF, aberrations in ovule development become more apparent in plants carrying the cac1a-1 allele. Wild-type embryos are at the heart stage of development (M) and the endosperm has begun to cellularize (P). In cac1a-1 plants, the aberrantly developing ovules have globular embryos with highly vacuolated cells (N and O). In the most severely aberrant ovules, the embryo sac has collapsed (O). Endosperm is uncellularized or aberrantly cellularized and sometimes degenerated (den; Q and R). Bars = 50 μm (M–O); 10 μm (P–R). S to X, At 7 DAF, the aberrant ovules in the cac1a-1 plants have ceased development at the globular or heart stages (T and U), the endosperm is degenerated, and the embryo sacs are often collapsed (W and X). In contrast, in the wild-type siblings the embryos are at the torpedo stage (S) and the endosperm is fully cellularized (V). Bars = 50 μm (S–U); 10 μm (V–X).

Figure 5.

The CAC1A function is required for normal pollen development and germination. A and B, Scanning electron micrographs of pollen grains collected from sibling plants that are either wild type (A) or carry the cac1a-1 allele (B). Arrows in B point to aberrantly shaped pollen grains that have an altered exine layer; these are recovered from plants carrying the cac1a-1 allele, but not from wild-type siblings. Insets show magnified views of individual pollen grains from wild-type plants and the aberrant pollen grains that were recovered from mutant plants. C to E, Scanning electron micrographs of pollen grains after 20 h on germination medium. Pollen collected from wild-type plants germinate and elongate a pollen tube (C), but those that are collected from plants that carry the cac1a-1 allele have an elevated number of grains that do not germinate or germinate aberrantly (D) or upon germination the elongating pollen tube appears to burst (E). F, The proportion of pollen grains that are classified as germinated, not germinated, or burst after 20 h on germination medium; dark bars are for pollen grains collected from wild-type plants and the white bars are for pollen grains collected from siblings that carry the cac1a-1 allele. Under each category are micrographs of Nile Blue stained pollen grains representative of each category. Bars = 10 μm. [See online article for color version of this figure.]

Figure 6.

The CAC1A function is required for normal in planta pollen germination and pollen tube elongation. Larger number of pollen grains germinate on the stigma (arrow), and pollen tubes more fully elongated in the style of flowers of wild-type (WT) plants (A and C) than on plants carrying the cac1a-1 allele (B and D). In addition, during processing for microscopic examination, more pollen grains have germinated in the anthers of wild-type plants (E) than on plants carrying the cac1a-1 allele (F). Pollen was visualized on sibling plants grown side by side that were either wild type (A, C, E) or heterozygous for the cac1a-1 allele (B, D, E). Flowers were stained either with Aniline Blue (A and B) or Aniline Blue and DAPI (C–F) and viewed by fluorescence microscopy with DAPI/fluorescein isothiocyanate/Texas Red filters. Bar = 10 μm.

Figure 7.

Reduced BCCP1 abundance in CAC1A antisense plants correlates with the reduced growth phenotype. A, Phenotype of representative plants used to investigate the relationship between BCCP1 abundance and the severity of the resulting reduced growth phenotype. WT, Wild type. Bars = 1 cm. B, Abundance of BCCP1 protein in five independent wild-type plants (WT; lanes a, b, g, h, l), four independent CAC1A antisense plants expressing the M phenotype (lanes c, d, i, j), and three independent CAC1A antisense plants expressing the S phenotype (lanes e, f, k). Protein extracts were prepared from young rosette leaves of 30-d-old plants, and 50 μg of protein from each plant extract was subjected to western-blot analysis, using either BCCP1 antisera followed by ¹²⁵I-labeled Protein A or ¹²⁵I-labeled streptavidin. Numbers below each blot are signal intensities relative to the average intensity of the WT (100). C, Comparison of BCCP1 abundance in wild-type and CAC1A antisense plants analyzed with either subunit-specific antisera or streptavidin as in B. BCCP1 signal intensities from wild type (WT; n = 5), and CAC1A antisense plants expressing a M (n = 4) and S (n = 3) phenotype were averaged. Three repetitions of this experiment gave similar results. [See online article for color version of this figure.]

Figure 8.

Reduction in the abundance of BCCP1, but not BCCP2, affects fatty acid accumulation without affecting fatty acid concentration or composition. Each value in the chart represents the averaged value of (n) biological replications. Error bars indicate the sds. A, Total fatty acid concentration in seeds harvested from wild type (WT; n = 6), cac1b-1 homozygous (cac1b-1; n = 6), and cac1b-2 homozygous (cac1b-2; n = 6) mutant plants. B, Total fatty acid concentration in leaves of wild-type (WT; n = 3) plants, CAC1A antisense plants with M (n = 7) phenotype, and CAC1A antisense plants with S (n = 3) phenotype. C, Total fatty acid content (per plant) in leaves from wild-type plants (WT; n = 5), CAC1A antisense plants with M phenotype (n = 4), and CAC1A antisense plants with S phenotype (n = 3). D, Fatty acid composition of seeds harvested from homozygous plants that carried wild-type CAC1B (white bars; n = 6), and mutant cac1b-1 (gray bars; n = 6) or cac1b-2 (black bars; n = 6) alleles. E, Fatty acid composition of leaves from wild-type plants (white bars; n = 3), and of leaves from CAC1A antisense plants with M phenotype (gray bars; n = 7), or S phenotype (black bars; n = 3).

Figure 9.

Reduced accumulation of BCCP1 subunit in CAC1A antisense RNA plants does not affect the accumulation of the other subunits of htACCase. A, Protein extracts were prepared from rosette leaves of three wild-type plants (WT) and five CAC1A antisense RNA plants expressing the M or S phenotypes. Aliquots of extracts containing 50 μg of protein were fractionated by SDS-PAGE and subjected to western-blot analyses. The blots were incubated with streptavidin or antisera against each individual subunit of htACCase. B, Protein extracts prepared from excised siliques isolated at between 1 and 3 DAF from wild-type (WT) and CAC1A antisense RNA plants with a S phenotype were subjected to western-blot analysis using either streptavidin or an anti-BCCP antibody that reacts with both BCCP1 and BCCP2 (Choi et al., 1995). C, The same protein samples used in B were subjected to western-blot analysis using antisera specific for BC, α-CT, or β-CT subunit.

Figure 10.

The expression pattern of CAC1A and CAC1B genes as determined by in situ hybridization and histological staining of GUS activity expressed from promoter:GUS transgenes. A, The spatial and temporal patterns of CAC1A and CAC1B mRNAs are near identical during embryo development. The CAC1A and CAC1B mRNAs were detected by in situ hybridization procedure described in the “Materials and Methods,” using gene-specific DIG-labeled antisense RNAs as probes; control hybridizations were conducted with DIG-labeled sense RNAs. Bar = 50 μm. B, Histological staining of GUS activity expressed in different Arabidopsis organs by CAC1A- and CAC1B-GUS transgenes. These data reveal that CAC1B is only expressed in a subset of the organs and tissues, whereas CAC1A is expressed at a higher level. Organs that are shown are as follows, and these are taken from plants that are developmentally staged as defined by Boyes et al. (2001). B1: rosette of plants between principal growth stages 1.08 and 1.12 ; B2: roots of same plants as shown in B1; B3: florets of plants at principal growth stage 6.30, green horizontal arrows point to flowers at 1 DAF and red vertical arrows point to flowers at 2 DAF; B4: flowers at 2 DAF from plants at principal growth stage 6.30; B5: silique at 5 DAF; B6: silique at 8 and 9 DAF, for CAC1A:GUS and CAC1B:GUS, respectively; B7: silique at 12 DAF. Red bars = 1 mm; blue bars = 0.1 mm.

Figure 11.

BCCP1 is the major paralog that accumulates during silique development. Using recombinantly produced authentic proteins as standards and BCCP1- or BCCP-2-specific antisera the absolute amount of each paralog was determined in protein extracts prepared from siliques at the indicated stage of development. A, The indicated amounts of recombinant BCCP1 or BCCP2 proteins, and aliquots of silique extracts (μg of protein loading indicated under each lane) were subjected to electrophoresis on the same gel. Gels were then subjected to western-blot analysis with subunit-specific antisera. B, The standard curve for BCCP-1 protein. C, The standard curve for BCCP-2 protein. D, BCCP1 and BCCP2 concentrations in developing siliques. Average of three determinations of three individual harvests of siliques at each indicated developmental stage; error bars indicate sds.