Toni A. Voelker

Palmitoyl-acyl carrier protein (ACP) thioesterase and the evolutionary origin of plant acyl-ACP thioesterases.

Acyl-acyl carrier protein (ACP) thioesterases play an essential role in chain termination during de novo fatty acid synthesis and in the channeling of carbon flux between the two lipid biosynthesis pathways in plants. We have discovered that there are two distinct but related thioesterase gene classes in higher plants, termed FatA and FatB, whose evolutionary divergence appears to be ancient. FatA encodes the already described 18:1-ACP thioesterase. In contrast, FatB representatives encode thioesterases preferring acyl-ACPs having saturated acyl groups. We unexpectedly obtained a 16:0-ACP thioesterase cDNA from Cuphea hookeriana seed, which accumulate predominantly 8:0 and 10:0. The 16:0 thioesterase transcripts were found in non-seed tissues, and expression in transgenic Brassica napus led to the production of a 16:0-rich oil. We present evidence that this type of FatB gene is ancient and ubiquitous in plants and that specialized plant medium-chain thioesterases have evolved independently from such enzymes several times during angiosperm evolution. Also, the ubiquitous 18:1-ACP thioesterase appears to be a derivative of a 16:0 thioesterase.

Expression of a maize sucrose phosphate synthase in tomato alters leaf carbohydrate partitioning.

We isolated a complementary DNA sequence for the enzyme sucrose phosphate synthase (SPS) from maize utilizing a limited amino acid sequence. The 3509-bp cDNA encodes a 1068-amino acid polypeptide. The identity of the cDNA was confirmed by the ability of the cloned sequence to direct sucrose phosphate synthesis in Escherichia coli. Because no plant-specific factors were necessary for enzymatic activity, we can conclude that SPS enzyme activity is conferred by a single gene product. Sequence comparisons showed that SPS is distantly related to the enzyme sucrose synthase. When expressed from a ribulose bisphosphate carboxylase small subunit promoter in transgenic tomatoes, total SPS activity was boosted up to sixfold in leaves and appeared to be physiologically uncoupled from the tomato regulation mechanism. The elevated SPS activity caused a reduction of starch and increase of sucrose in the tomato leaves. This result clearly demonstrates that SPS is involved in the regulation of carbon partitioning in the leaves.

In vitro mutated phytohemagglutinin genes expressed in tobacco seeds: role of glycans in protein targeting and stability.

Phytohemagglutinin is a glycoprotein that accumulates in the protein storage vacuoles of bean seeds. The mature glycoprotein has a high-mannose and a complex glycan. We describe here the use of site-directed mutagenesis and expression of the mutated genes in transgenic tobacco to study the role of glycans in intracellular targeting. The reading frame for phytohemagglutinin-L was mutated so that either one or both of the glycosylation signals were disrupted to specifically prevent the attachment of asparagine-linked glycans. Expression of these genes with the beta-phaseolin promoter in the seeds of transgenic tobacco plants showed that phytohemagglutinin-L with only one glycan or without glycans was correctly targeted to the protein storage vacuoles of the seeds. Furthermore, the absence of either the complex glycan or the high-mannose glycan did not alter the processing of the other glycan. On the basis of these results, we propose that the targeting signal of this vacuolar protein is contained in its polypeptide domain and not in its glycans.

Broad-range and binary-range acyl-acyl-carrier protein thioesterases suggest an alternative mechanism for medium-chain production in seeds.

In the current model of medium-chain (C8–14) fatty acid biosynthesis in seeds, specialized FatB acyl-acyl-carrier-protein (ACP) thioesterases are responsible for the production of medium chains. We have isolated and characterized FatB cDNAs from the maturing seeds of elm (Ulmus americana) and nutmeg (Myristica fragrans), which accumulate predominantly caprate (10:0)- and myristate (14:0)-containing oils, respectively. In neither species were we able to find cDNAs encoding enzymes specialized for these chain lengths. Nutmeg FatB hydrolyses C14–18 substrates in vitro and expression in Brassica napus seeds leads to an oil enriched in C14–18 saturates. Elm FatB1 displays a binary specificity: one activity is centered on 10:0-ACP, and a second is centered on palmitate (16:0)-ACP. After expression in B. napus seeds the oil is enriched in C10–18 saturates, predominantly 16:0, 14:0, and 10:0. The composition of free fatty acids produced by elm FatB1 in Escherichia coli shifts from C14–16 to mostly C8–10 by increasing the rate of chain termination by this enzyme. These results suggest the existence of an alternative mechanism used in the evolution of medium-chain production, a model of which is presented.

On Defining T-DNA.

The Agrobacterium-mediated DNA transfer system is the most widely used method for inserting genes into plant genomes; it is utilized for both basic and applied purposes. Previous studies have indicated that Agrobacterium T-DNA is present in the genomes of transformed host plants as single units or in multiple, tandemly arrayed copies (De Block et al., 1984; Spielmann and Simpson, 1986; Jorgensen et al., 1987; Deroles and Gardner, 1988a) and have emphasized that truncated T-DNA regions can frequently be observed (Holsterset al., 1983; Derolesand Gardner, 1988b). For those studies in which host plant DNA was analyzed using hybridization methods, probes specifically designed to detect DNA sequences located beyond the T-DNA border repeats in the original bacterial plasmid were usually not employed. One exceptional study described a single example of a T-DNA region that contained Ti plasmid sequences that extended beyond the usual border repeats in transformed plant tumor tissue (Ooms et al., 1982). However, the generally accepted view has been that DNA from beyond the T-DNA border direct repeats is not transferred into the host plant (Zambryski, 1992). In the course of our development of transgenic crop plants, we have conducted a thorough analysis of the DNA inserted into our transformed plants via the Agrobacterium-mediated transfer process. Using the DNA blot hybridization technique (Southern, 1975) under standard reaction conditions (see, for example, Sanders et al., 1992), we have now examined transferred DNA in several hundred plants (representing several crop species) independently transformed using an Agrobacterium-mediated binary vector system (McBride and Summerfelt, 1990). DNA from beyond the classically defined T-DNA region has been integrated into the genomes of approximately 20 to 30% of these plants. For example, in one cocultivation experiment, 22 of 96 independently generated plants transformed with the same gene construct contained DNA sequences from beyond the border repeats. Our examination of offspring from plants transformed using another Agrobacterium-mediated binary vector system (An et al., 1985) revealed a similar frequency (20%) of “beyond the border” DNA transfer. That stable transfer of DNA from beyond the T-DNA borders can occur in transformed plants may not come as a surprise to some plant scientists. “Read-through” of border sequences during T-strand formation in the bacterium has been reported (Stachel et al., 1987; Veluthambi et al., 1988), for example. Also, DNA sequences from far beyond the T-DNA region have been detected in petunia protoplasts during a 24-hr period after infection with Agrobacterium (Virts and Gelvin, 1985). Based on these observations in transient systems, one might predict that sequences outside the classically defined T-DNA would be available for stable incorporation into a plant genome. However, we find the high frequency at which T-DNA extensions occur in transgenic plants to be quite unexpected. We bring these results to the attention of our colleagues for two reasons. The first is of a practical nature. Although T-DNA binary vectors are completely disarmed of the virulence and oncogenic genes normally present on Ti plasmids, assays for sequences beyond the border repeats should, and can easily, be performed when characterizing potential products. Also, plant scientists involved in projects such as cloning genes via T-DNA tags should be aware of the potential for considerably larger “tags” than had been heretofore expected, a factor that might complicate the identification of adjacent plant DNA. Second, and perhaps more importantly, we hope that this information will contribute to the design of ongoing investigations into the molecular mechanism of T-DNA transfer.

Medium-chain fatty acid biosynthesis and utilization in Brassica napus plants expressing lauroyl-acyl carrier protein thioesterase

We have examined production of mediumchain fatty acids by Brassica napus L. plants transformed with a California bay (Umbellularia californica) medium-chain acyl-acyl carrier protein (ACP) thioesterase (UcFatB1) cDNA under the control of the constitutive cauliflower mosaic virus 35S promoter. These plants were found to accumulate medium-chain fatty acids in seeds but not in leaves or roots. Assay of thioesterase activity in extracts of leaves indicated that lauroyl-ACP thioesterase activity is comparable to oleoyl-ACP thioesterase (EC 3.1.2.14) activity in transformant leaves. Furthermore, leaf lauroyl-ACP thioesterase activity was in excess of that which produced a significant increase in the amount of laurate (12:0) in seed. Studies in which isolated chloroplasts were 14C-labelled were used to evaluate whether medium-chain fatty acids were produced in transformed leaves. Up to 34% of the fatty acids synthesized in vitro by isolated chloroplasts were 12:0. These results demonstrate that the normally seed-localized lauroyl-ACP thioesterase can be expressed in active form in leaves, imported into chloroplasts and can access acyl-ACP intermediates of leaf de-novo fatty acid synthesis. The most likely explanation for the lack of accumulation of 12:0 in transformed leaves is its rapid degradation by β-oxidation. In support of this hypothesis, isocitrate lyase (EC 4.1.3.1) activity was found to be significantly increased in plants transformed with 35S-UcFatB1.

Correct glycosylation, Golgi-processing, and targeting to protein bodies of the vacuolar protein phytohemagglutinin in transgenic tobacco.

We used a heterologous system (transgenic Nicotiana tabacum L.) to investigate the processing, assembly and targeting of phytohemagglutinin (PHA), the lectin of the common bean, Phaseolus vulgaris L. In the bean, this glycoprotein accumulates in the protein bodies of the storage parenchyma cells in the cotyledons, and each polypeptide has a high-mannose glycan attached to Asn12 and a complex glycan on Asn60. The gene for PHA-L, dlec2, with 1200 basepairs (bp) 5′ upstream and 1600 bp 3′ downstream from the coding sequence was introduced into tobacco using Agrobacterium-mediated transformation (T. Voelker et al., 1987, EMBO J. 6, 3571–3577). Examination of thin sections of tobacco seeds by immunocytochemistry with antibodies against PHA showed that PHA-L accumulated in the amorphous matrix of the protein bodies in the embryo and endosperm. This localization was confirmed using a non-aqueous method to isolate the protein bodies from mature tobacco seeds. The biochemical analysis of tobacco PHA indicated that the signal peptide had been correctly removed, and that the polypeptides formed 6.4 S oligomers; tobacco PHA had a high-mannose glycan at Asn12 and a complex glycan at Asn60. The presence of the complex glycan shows that transport to the protein bodies was mediated by the Golgi complex. At seed maturity, a substantial portion of the PHA-L remained associated with the endoplasmic reticulum and the Golgi complex, as indicated by fractionation experiments using aqueous media and the presence of two high-mannose glycans on some of the polypeptides. Taken together, these data show that insertion of the nascent PHA into the endoplasmic reticulum, signal peptide processing, glycosylation, assembly into oligomers, glycan modification in the Golgi, and targeting of the protein occur faithfully in this heterologous system, although transport may not be as efficient as in bean cotyledons.

Differences in expression between two seed lectin alleles obtained from normal and lectin-deficient beans are maintained in transgenic tobacco

Using Agrobacterium‐mediated transformation, two genes for phytohemagglutinin‐L (PHA‐L), the lectin seed protein of the common bean Phaseolus vulgaris, were stably integrated into the tobacco genome. The two alleles for PHA‐L, dlec2 and pdlec2, were obtained from a normal cultivar (Greensleeves) and a lectin‐deficient cultivar (Pinto) respectively. In the bean embryos, the expression of dlec2 is 30 times greater than the expression of pdlec2. In the dlec2‐transformed tobacco, PHA‐L accumulated specifically in the seeds at the same stages as the tobacco seed storage proteins and was degraded after germination. PHA‐L was found in the embryo, and at a 5–7 times lower concentration in the endosperm tissue of the mature tobacco seeds. No PHA could be detected in other parts of the plants. We conclude that the signals for temporal and spatial regulation of the dlec2 gene are present in the DNA fragment used for transformation. Transformation with the second PHA‐L allele pdlec2 from the cultivar Pinto caused the accumulation of about 50 times less PHA‐L in tobacco seeds when compared to dlec2. We conclude from analyzing the 5′ sequences of dlec2 and Pdlec2 that the low expression phenotype of the Pdlec2 allele could be due to the absence or mutation of a cis‐acting element carried by the dlec2 fragment.

Cotyledon nuclear proteins bind to DNA fragments harboring regulatory elements of phytohemagglutinin genes.

The effects of deleting DNA sequences upstream from the phytohemagglutinin-L gene of Phaseolus vulgaris have been examined with respect to the level of gene product produced in the seeds of transgenic tobacco. Our studies indicate that several upstream regions quantitatively modulate expression. Between -1000 and -675, a negative regulatory element reduces expression approximately threefold relative to shorter deletion mutants that do not contain this region. Positive regulatory elements lie between -550 and -125 and, compared with constructs containing only 125 base pairs of upstream sequences (-125), the presence of these two regions can be correlated with a 25-fold and a 200-fold enhancement of phytohemagglutinin-L levels. These experiments were complemented by gel retardation assays, which demonstrated that two of the three regions bind cotyledon nuclear proteins from mid-mature seeds. One of the binding sites maps near a DNA sequence that is highly homologous to protein binding domains located upstream from the soybean seed lectin and Kunitz trypsin inhibitor genes. Competition experiments demonstrated that the upstream regions of a bean beta-phaseolin gene, the soybean seed lectin gene, and an oligonucleotide from the upstream region of the trypsin inhibitor gene can compete differentially for factor binding. We suggest that these legume genes may be regulated in part by evolutionarily conserved protein/DNA interactions.

Expression analysis of a pseudogene in transgenic tobacco: a frameshift mutation prevents mRNA accumulation.

Seeds of the Pinto cultivar of the common bean, Phaseolus vulgaris, are deficient in phytohemagglutinin (PHA), a lectin normally composed of two different polypeptides (PHA-E and PHA-L). In Pinto seeds, there is no PHA-E and only small amounts of PHA-L. The gene coding for the Pinto PHA-E, Pdlec1, is a pseudogene as a result of a single base pair deletion in codon 11, causing a frameshift and premature termination of translation. This mutation explains the absence of the PHA-E polypeptide but not the several-hundredfold reduction of the cytoplasmic Pdlec1 mRNA in developing seeds when compared with a normal PHA-E gene. To find the cause for this reduction in mRNA levels, we swapped gene fragments of Pdlec1 with the homologous parts of a normal PHA gene from the cultivar Greensleeves and introduced these fusions into tobacco. Analysis of the transgenic seeds showed that the Pdlec1 promoter is fully functional. We also repaired the Pdlec1 coding frame in vitro and inserted the repaired and unrepaired versions into a PHA gene expression cassette. In transgenic tobacco, both constructs showed Pdlec1 transcript accumulation in the second half of seed maturation. The single-base frame repair boosted the peak transcript levels by a factor of 40 and resulted in the synthesis of PHA-E at normal levels. We propose that the premature translational stop caused by the frameshift leads to a faster breakdown of the Pdlec1 mRNA, thereby preventing this transcript from accumulating to high levels.