Flower Breeding and Genetics: Issues, Challenges and Opportunities for the 21st Century (2007) pp. 321-323
edited by Neil O. Anderson

Chapter 11. Petunia. Petunia x hybrida
Robert J. Griesbach

5.1.2 Flavonoid Genetic Engineering

One of the first practical examples of plant genetic engineering involved the development of a novel flower color in petunia through the engineering of dihydroflavonol reductase (DFR). The Petunia DFR has an extremely low substrate specificity for dihydrokaempferol; therefore, pelargonidin is rarely found (Huitts, et. al., 1994). Mutants with a defective Htl gene would be expected to produce unpigmented flowers (Figure 11-1). Two leaky hr hfl- mfl- mutants (RL01 and W80) and one leaky ht- hfl- hf2- mutant (Skr4 x Sw63) were used as parent plants in transformation. Mutants RL01 and W80 both produced one tenth the normal amount of anthocyanin that was composed of 28% delphinidin, 63% cyanidin and 9% pelargonidin (Griesbach, 1993; Johnson, et al., 1999). Mutant Skr4 x Sw63 produced one tenth the normal amount of anthocyanin that was composed of 24% pelargonidin, 7% peonidin, 4% petunidin and 65% malvidin (Tanaka, et al., 1995).

A Cymbidium DFR gene was introduced into mutant W80 (Johnson et al., 1999). The Cymbidium enzyme only recognizes dihydroquercetin as a substrate. Therefore, only cyanidin and peonidin are produced (Tatsuzawa et al., 1996). When the Cymbidium gene was expressed in W80, there was no change in flower color.

A Zea mays DFR gene (AI) was introduced into mutant RLO1 (Meyer, et al., 1987). The Zea enzyme recognizes both dihydroquercetin and dihydrokaempferol as a substrate, but has a much stronger affinity for dihydroquercetin. Pelargonidin is only produced in the absence of dihydroquercetin. When AI was expressed in RL01, there was a ten-fold increase in the total amount of anthocyanin, as well as an increase in the relative amount of pelargonidin from 9% to 55% (Griesbach, 1993).

A Gerbera DFR gene (Gdfr) was introduced into RL01 (Elomaa et al., 1995). The Gerbera enzyme recognizes both dihydroquercetin and dihydrokaempferol as a substrate, but has a much stronger affinity for dihydrokaempferol. Pelargonidin is produced irrespective of what precursor is present (Asen, 1984). When the Gdfr was expressed in RL01, the total amount anthocyanin and the relative amount of pelargonidin did not differ significantly from RL01 plants expressing A1.

A Rosa DFR gene was introduced into Skr4 x Sw63 (Tanaka et al., 1995). The Rosa enzyme recognizes both dihydroquercetin and dihydrokaempferol, but has a stronger affinity for dihydroquercetin. A small amount of pelargonidin is always produced (Asen 1982). When the Rosa gene was expressed Skr4 x Sw63, there was a ten-fold increase in the total amount of anthocyanin, as well as, an increase in the relative amount of pelargonidin from 24% to 97%.

In order to identify the DNA sequence leading to substrate specificity, chimeric DFR genes were constructed and introduced into P. x hybrida 'W80' (Johnson, et al., 1999). It was determined that the substrate binding region was between amino acids 132 and 158 with amino acid 134 critical in substrate specificity. A switch from asparagine to leucine at position 134 caused a change in substrate preference from dihydroquercetin to dihydrokaempferol.

When grown in the greenhouse, RL01 plants expressing A1 produced solid orange flowers; while when grown outdoors, they produced weakly pigmented flowers with reduced levels of A1 expression (Meyer, et al., 1992). Reduced expression (gene silencing) was due to DNA methylation within either the 35S promoter in the A1 construct or in A1 itself (Meyer and Heidmann, 1994).

Unlike plants expressing A1, plants expressing Gdfr very rarely produced weakly pigmented flowers with reduced levels of transgene expression. In those rare plants with reduced expression, only the 35S promoter was methylated. Gdfr itself was never methylated. It was suggested that the increase in the stability of Gdfr expression over A1 expression was due to the lower GC content (39%) of Gdfr than AI (60%). Dicotyledonous plants have a lower GC content than monocotyledonous plants. Transgene silencing can be induced by differences in the GC content between the foreign and host DNA (Fagard and Vaucheret, 2000).

There are two mechanisms for gene silencing (Fagard and Vaucheret, 2000). In transcriptionally gene silencing (TGS), there is a decrease in mRNA synthesis because of promoter methylation. As described above, the A1 gene was silenced through TGS. In post-transcriptionally gene silencing (PTGS), there is decrease in the mRNA because of sequence specific degradation. The exact mechanism(s) for gene silence are still unknown; however, both TGS and PTGS involve the production of small double stranded RNA (dsRNA) molecules (Sijen et al., 2001).

The introduction of a transgene can also lead to PTGS of its endogenous homologous gene. The silenced phenotype is dependent upon the orientation of the transgene. Petunia x hybrida plants expressing a chalcone synthase (Chs) transgene in the same orientation as the endogenous gene (co-suppression) produced flowers with reduced pigmentation at the junctions between adjacent petals (Napoli, et al., 1990). While P. x hybrida plants expressing a chalcone synthase (Chs) transgene in opposite orientation as the endogenous gene (anti-sense suppression), produced a different phenotype with white petal margins and reduced pigmentation throughout the flower (van der KroI, et al., 1988). Through breeding it was possible to combine both phenotypes in a single plant (Que, et al., 1998). It was suggested that sense and anti-sense suppression occur in different cellular compartments or at different times in development.

Genetic engineering has been used to study the expression of flavonoid regulatory genes. In genetic complementation studies, the Zea mays regulatory genes Lc (encodes a bHLH protein) and C1 (encodes a MYB protein) together were able to up-regulate the An6 promoter and increase anthocyanin production in a white-flowered an2- mutant (Quattrocchio, et al., 1993). Lc alone had no effect. It was suggested that structural gene activation requires the interaction of a bHLH protein with a MYB protein (Bradley, et al., 1998). In the complementation study, the Lc encoded bHLH protein by itself was not able to up-regulate the An6 promoter because of the absence of a MYB protein due to the an2- mutation.

Regulatory genes are highly conserved (Solano, et al., 1997). All animal c-MYB-domain proteins bind to a single specific DNA sequence (AAAC(G/C)GTTA) called MBSI, while the Petunia encoded MYB-domain protein binds to both the MBSI sequence and another sequence (AGTTAGTTA) called MBSII. A single amino acid change from glutamine to leucine at position 71 in the animal protein switched its single DNA binding specificity to dual binding.

Figure 11-1. The flavonoid biosynthetic pathway, with the identified corresponding genes, operating in Petunia x hybrida.