• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Jun 20, 2000; 97(13): 7016–7023.
PMCID: PMC34378
Colloquium Paper

Flower color variation: A model for the experimental study of evolution

Abstract

We review the study of flower color polymorphisms in the morning glory as a model for the analysis of adaptation. The pathway involved in the determination of flower color phenotype is traced from the molecular and genetic levels to the phenotypic level. Many of the genes that determine the enzymatic components of flavonoid biosynthesis are redundant, but, despite this complexity, it is possible to associate discrete floral phenotypes with individual genes. An important finding is that almost all of the mutations that determine phenotypic differences are the result of transposon insertions. Thus, the flower color diversity seized on by early human domesticators of this plant is a consequence of the rich variety of mobile elements that reside in the morning glory genome. We then consider a long history of research aimed at uncovering the ecological fate of these various flower phenotypes in the southeastern U.S. A large body of work has shown that insect pollinators discriminate against white phenotypes when white flowers are rare in populations. Because the plant is self-compatible, pollinator bias causes an increase in self-fertilization in white maternal plants, which should lead to an increase in the frequency of white genes, according to modifier gene theory. Studies of geographical distributions indicate other, as yet undiscovered, disadvantages associated with the white phenotype. The ultimate goal of connecting ecology to molecular genetics through the medium of phenotype is yet to be attained, but this approach may represent a model for analyzing the translation between these two levels of biological organization.

Keywords: flavonoid biosynthesis, transposon, spatial autocorrelation, pollinator behavior, morning glory

The study of adaptation is a problem that intersects all disciplines of biology. The study of adaptation is also among the most difficult and challenging areas of experimental research because a complete causal analysis of adaptation involves a translation between different levels of biological organization, the ecological, the phenotypic, and the molecular level (1). In this article we review more than 20 years of research on flower color polymorphisms in the common morning glory [Ipomoea purpurea (L.) Roth] that was aimed at exploring this interface.

When the morning glory research program was initiated in the late 1970s, flower color polymorphisms appeared to be a natural starting point because (i) they represented simple discrete phenotypes that were susceptible to genetic analysis; (ii) a substantial body of work existed on the biochemistry of plant secondary metabolism; (iii) flower color was known to be important in insect pollinator behavior; and (iv) selection on reproductive performance should be among the most effective forms of selection, and, as a consequence, it should be the component of selection most likely to yield to experimental analysis. Virtually nothing was known in the late 1970s about the molecular biology of the genes that determine the anthocyanin pigments responsible for flower color, so nothing was known about the specific mutational changes associated with different flower color polymorphisms. Since that time a great deal of progress has been made in describing the molecular biology of the genes of flavonoid biosynthesis that determine flower color, but we are still some distance from a complete causal analysis that connects ecology to phenotype to genes.

We begin by discussing the natural history of the morning glory and then turn to a brief account of the genetics of flower color variation in the common morning glory. Next, we describe the flavonoid biosynthetic pathway that determines flower color, and we review pertinent work on the molecular genetics of the genes that encode enzymes within this pathway. Finally, we consider progress in the analysis of selection on flower color variation in natural and experimental populations of the common morning glory.

Natural History of I. purpurea.

The genus Ipomoea includes approximately 600 species distributed on a worldwide scale (2) that are characterized by a diversity of floral morphologies and pigmentation patterns. In addition, a wide variety of growth habits, ranging from annual species to perennial vines to longer-lived arborescent forms, are represented in the genus. The common morning glory is an annual bee-pollinated self-compatible vine with showy flowers that is a native of the highlands of central Mexico. As the name morning glory suggests, the flowers open early in the morning and are available for fertilization for a few hours, after which the flower wilts and abscises from the vine. The plant is also a common weed in the southeastern U.S., where it is found in association with field corn and soybean plantings, as well as in roadside and disturbed habitats. The common morning glory is characterized by a series of flower color polymorphisms that include white, pink, and blue (or dark blue) phenotypes. A primary pollinator in the southeastern U.S. is the bumblebee (Bombus pennsylvanicus and Bombus impatiens), but occasionally plants are visited by honey bees and some lepidopterans (3). The flower color phenotypes are thought to have been selected by pre-Columbian peoples, perhaps in association with maize culture (4). At some point, the plant was introduced into the southeastern U.S., although the routes and times of introduction remain uncertain. Early floras of the southeastern U.S. indicate the presence of I. purpurea populations by the late 1600s, providing a minimum estimate of the residence time in this geographical region.

Genetics of Flower Color Variants in I. purpurea.

At least 21 floral phenotypes are determined by five genetic loci (57). Most of these phenotypes have analogous forms in the Japanese morning glory (Ipomoea nil), and the genetics of the floral variants in both species appear to be similar, but not identical (8). A widespread polymorphism determines pink versus blue flowers (P/p locus), but the genotype at two other loci modifies the intensity of expression of the P/p locus. One modifier is an intensifier locus (I/i) that doubles the anthocyanin pigmentation in the recessive ii genotype (9). The second modifier locus is a regulatory locus that determines the patterning and degree of floral pigmentation (W/w locus). A fourth locus, the A/a locus, is epistatic to the P/p, I/i, and W/w loci in that the recessive albino phenotype yields a white floral limb independent of genotypic state at the other loci. The A/a locus is also characterized by unstable alleles (denoted a* or af) that exhibit pigmented sectors on an otherwise albino floral limb (10, 11). The pigmented sectors display the color associated with the P/p genotype. Finally, the pigmentation in the floral tube appears to be controlled separately from the outer floral limb, but the genetics of floral tube variation have not been analyzed. Fig. Fig.11 displays the flower color phenotypes determined by these genetic loci.

Figure 1
Flower color variation in I. purpurea. Loci are described in the text. The locus that determines the phenotype shown is highlighted in bold. Dashes indicate that the phenotype is dominant and only the dominant allele is therefore indicated. In the ...

Flavonoid Biosynthetic Pathway.

To put the phenotypic variation into a biochemical context, it is useful to sketch the main outlines of the flavonoid biosynthetic pathway (Fig. (Fig.2),2), which culminates in the production of anthocyanins, the main pigments responsible for flower color. The presence or absence of these pigments affects the coloration of the floral display, which attracts pollinators. The anthocyanin pigments are therefore important to reproductive success and hence to gene transmission. In addition to pigment production, several side branches of the pathway also produce compounds that are important in plant disease defense, pollen viability, microbial interactions, and UV protection (12). The flavonoid pathway has a pleiotropic role in plants, and one must consider that a single mutation in the pathway may have multiple phenotypic effects. The pleiotropic role of the flavonoid pathway is a complication in the effort to link phenotype to molecular changes, but, in addition, many of the genes of the flavonoid pathway are now known to consist of small multigene families (Table (Table1),1), so it is essential to associate a mutation in a particular gene with a phenotype of interest. That is, one must identify the gene family member responsible for the observed phenotype.

Figure 2
Flavonoid biosynthetic pathway. Enzymes involved in anthocyanin biosynthesis and side branches leading to related flavonoid pathways are shown. PAL, phenylalanine ammonialyase; C4H, cinnamate 4-hydroxylase; 4Cl, coumarate:CoA ligase; CHS, chalcone ...
Table 1
Estimated copy number of flavonoid genes in the I. purpurea genome

Molecular Characterization of the Genes of Flavonoid Biosynthesis in Ipomoea.

The first committed step in the flavonoid biosynthetic pathway is encoded by the enzyme chalcone synthase (CHS), which catalyzes the formation of naringenin chalcone from three molecules of malonyl-CoA and one molecule of p-coumaroyl-CoA (13). Several lines of evidence suggested that the A/a locus might encode a CHS gene. First, the albino phenotype is epistatic to all other flower color variants, suggesting an early point of action in the pathway, and, second, the albino phenotype is consistent with a blockage at CHS. In an attempt to show that the A/a locus encodes a chalcone synthase enzyme, we initiated efforts to clone chalcone synthase genes from a genomic library of I. purpurea. Multiple screenings of the library by using a heterologous probe from parsley (14) were unsuccessful. Subsequent screening of the library with a tomato CHS clone (15) resulted in the cloning of four genes (CHS-A, -B, -C, and a pseudogene) initially identified as CHS on the basis of nucleotide sequence similarity to other published CHS gene sequences (16).

To provide a comparative context for Ipomoea CHS gene family evolution, one can look at the Petunia CHS gene family. Petunia is important because both Ipomoea and Petunia are in the same flowering plant order (Solanales), and the Petunia CHS gene family had been extensively characterized, with as many as 12 genes provisionally identified (17). The four I. purpurea genes appear to share a common line of descent with an unusual Petunia gene (CHS-B) that is relatively distant in nucleotide sequence from other Petunia CHS genes.

Subsequently, Fukada-Tanaka et al. (18) cloned and characterized two more CHS genes from Ipomoea (CHS-D and -E) by using differential display and AFLP-based mRNA fingerprinting (19). These genes proved to be more closely related in nucleotide sequence to the majority of CHS genes characterized in Petunia. Biochemical analysis of Ipomoea CHS genes A, B, D, and E revealed that only CHS-D and -E are capable of catalyzing the condensation reaction that results in naringenin chalcone (20). The CHS-A and -B genes appear to encode enzymes that produce bisnoryangonin but not naringenin chalcone (H. Noguchi, personal communication). All of the Ipomoea CHS genes (except the pseudogene) are expressed, but each displays differential regulatory and developmental control (21, 22). CHS-D is the most abundantly expressed transcript and is now known to be the one CHS gene solely responsible for anthocyanin production in the floral limb (21, 23). As discussed in greater detail below, the identification of CHS-D as the A/a locus followed the identification of a transposon (23) in CHS-D, which accounted for the sectoring phenotype and loss of pigment. The CHS-E gene is also expressed in the limb but at a much lower level than CHS-D (21, 22). CHS-E is now believed to be responsible for pigment production in the tube of the flower (22).

The next step in the pathway is encoded by chalcone isomerase (CHI), which catalyzes the isomerization of the naringenin chalcone product of CHS to the flavanone naringenin. In the absence of functional CHI, spontaneous isomerization may still occur in vivo because this is known to occur in vitro under physiological conditions (24). CHI appears to be a single copy gene in I. purpurea, as multiple screenings of the genomic library have yielded only a single gene (unpublished data).

Flavanone 3β-hydroxylase (F3H) follows CHI in the pathway and hydroxylates flavonones at the 3 position to form dihydroflavonols, which are required for the synthesis of anthocyanidins and flavonols. F3H consists of at least two copies in the I. purpurea genome, both of which are expressed (25). It is not known at this time whether one of the F3H genes is solely responsible for anthocyanin production in the limb (as is the case for CHS-D). Also, because they were both identified as F3H on the basis of nucleotide similarity, it is not known whether both are even capable of performing the hydroxylation step. One functional copy of F3H and one pseudogene have so far been identified in I. nil (ref. 26; S. Iida, personal communication).

Another hydroxylase, dihydroflavanol 3′ hydroxylase (F3′H), hydroxylates the 3′ position of the dihydroflavonol produced by F3H. This results in the eventual production of the red/magenta cyanidin. F3′H has been characterized in both I. purpurea and I. nil (S. Iida, personal communication). Yet another hydroxylase, dihydroflavonol 3′5′-hydroxylase (F3′5′H), hydroxylates the 3′ and 5′ position of the dihydroflavonol produced by F3H. This product ultimately leads to the production of the blue/purple delphinidin. F3′5′H has not yet been characterized in Ipomoea.

The next step in the flavonoid pathway is dihydroflavonol reductase (DFR) which reduces dihydroflavonols to leucoanthocyanidins. In I. purpurea, DFR is a small gene family consisting of at least three tandemly arranged copies (DFR-A, -B, and -C) (27). DFR-B has been identified as the gene responsible for anthocyanin production in the floral limb based on work from I. nil in which a transposon disrupts the DFR-B gene, resulting in a sectoring phenotype and loss of pigment (ref. 28; see below). The function of the two other DFR genes in I. purpurea is not known, nor is it known whether they are capable of performing the reductase reaction.

Anthocyanidin synthase (ANS) encodes a dioxygenase and appears to be single copy in I. purpurea. UDP-glucose flavonol 3-0-glucosyl transferase glycosylates anthocyanidins and flavonols on the 3 position. This gene appears to be single copy in I. purpurea. Rhamnosyl transferase adds rhamnosyl to glucose to form rutinoside. This gene is as yet uncharacterized in I. purpurea.

Most Mutant Phenotypes Appear To Be the Result of Transposon Insertions.

A wide variety of mobile elements (Table (Table2)2) have been identified in the Ipomoea genome, largely because of work from the laboratory of Shigeru Iida at the National Institute for Basic Biology in Okasaki, Japan. Some of these mobile element insertions cause phenotypic changes, including those responsible for several flower color variants (Table (Table3).3). Much of this work has concentrated on the Japanese morning glory (I. nil), where the rich history of morning glory genetics in Japan has provided an extensive research foundation. Considerable work has also been done both in the U.S. and in Japan on I. purpurea, which is a member of the same subgenus as I. nil. Both species appear to be quite closely related at the DNA sequence level. Molecular clock calculations indicate that the two species diverged roughly three million years ago based on synonymous divergence at DFR and CHS genes (unpublished data). These two closely related species provide a useful comparative framework for the analysis of genetic change over a relatively short period of evolutionary time.

Table 2
Transposons and mobile elements in I. purpurea and I. nil associated with genes of the flavonoid pathway
Table 3
Mutations linked to phenotype in I. purpurea and I. nil

A 3.9-kb Ac/Ds-like element (Tip100) has been identified in the intron of CHS-D near the 5′ junction in flaked (af) mutants of I. purpurea (23). The af flower is white with pigmented sectors on the corolla. Another CHS-D mutant phenotype (a12), which has white flowers and no pigmented sectors, has been shown to carry two copies of Tip100 in the intron in the opposite orientation (23). In another stable white phenotype (a), a rearrangement of DNA sequence between exon I and an adjacent Tip100 element has occurred (unpublished data). In yet another mutant (a*) of I. purpurea, which has a white corolla with very few pigmented sectors, an additional copy of Tip100 was found in the 5′ flanking region of CHS-D (unpublished data). The discovery that disruption of the CHS-D gene results in an unpigmented phenotype confirms that CHS-D is the only CHS gene family member that is responsible for pigment production in the floral limb (21, 22). Another Ac/Ds-like element, Tip201, has also been discovered in I. purpurea (S. Iida, personal communication). Tip201 is found inserted in exon III of F3′H, resulting in a pink phenotype rather than the wild-type blue color corresponding to the P/p locus (S. Iida, personal communication). A point mutation in F3′H in I. nil also results in a red phenotype instead of blue (S. Iida, personal communication). Interestingly, this is the only phenotypic change involving flower color that has been characterized at the molecular level in either I. nil or I. purpurea that is not attributable to the presence of a transposable element.

An En/Spm-like mobile element termed Tpn1 was characterized by Inagaki et al. (28) in the I. nil genome. Tpn1 is a 6.4-kb non-autonomous element inserted within the second intron of DFR-B in a mutant termed flecked (a-3flecked). This phenotype has predominantly white flowers with colored sectors. An unusual feature of Tpn1 is that it contains a segment of DNA consisting of four exons that encode part of an HMG-box (High Mobility Group DNA-binding proteins) (29). This finding is consistent with the observation that transposable elements can cause rearrangements of the genome (30). Not only did Inagaki et al. (28) establish that the observed sectoring phenotype was caused by the movement of this new transposable element, but the finding also established that, of the three DFR genes characterized, only one gene family member (DFR-B) is responsible for pigment production in the floral limb.

Another En/Spm-like element, Tpn2, has also been identified in I. nil.§Tpn2 is a 6.5-kb element with similarity to Tpn1. Tpn2 is inserted in the second intron of CHI in the speckled mutant. This phenotype is pale yellow with round speckles of pigment on the corolla (31). In addition to the En/Spm-like transposons, mobile element-like motifs were identified in the flanking regions of the DFR genes. These elements are similar to miniature inverted-repeat transposable elements (22, 32, 33). Mobile element-like motifs are also found in the DFR region of I. purpurea (27) and in the CHS-D region of both I. purpurea and I. nil. Three copies of one of the elements (mobile element-like motif 6) from the CHS-D region were also found in the DFR region (22).

Furthermore, three copies of a directly repeated sequence of about 193 bp are found at the 3′ end of the intron in CHS-D in some lines of I. purpurea (22). Other lines of I. purpurea contain four copies of this repeat (unpublished data). In contrast, I. nil contains only one copy of this repeat, but it is interrupted by a 529-bp insertion (22).

Another En/Spm-related element termed Tpn3 has been found in the CHS-D gene of I. nil in the r-1 mutant. This mutant bears white flowers with colored tubes. The insertion of this 5.57-kb element into CHS-D results in the accumulation of abnormal sizes of CHS-D mRNAs in the floral tissue.

A long terminal repeat retrotransposon, RTip1, has been reported associated with the ANS gene region in I. purpurea (34). The element is 12.4 kb, contains two long terminal repeat sequences of about 590 kb, and appears to be a defective Ty3 gypsy-like element. The RTip1 element resides within yet another element (MiniSip1), which is described as a minisatellite. Another minisatellite, MiniSip2, has also been described and is located within the Rtip1 element. Thus, there are three elements piggybacked on one another in the ANS 3′ flanking region. Although no phenotypic changes have been linked to these elements, DNA rearrangements in the vicinity of ANS are attributed to their presence (34).

A sine-like element (SineIp) has also been identified in the 5′ flanking region of CHS-D (unpublished data) in some lines of I. purpurea. This element is 236 bp and contains the Pol III promoters at the 5′ end of the element. It has 15 bp direct terminal repeats. No phenotypic changes have yet been associated with this element.

Another floral mutation in I. nil, not related to the flavonoid pathway, but which is caused by a transposon insertion, is the duplicated mutant. In this mutant phenotype, sexual organs are replaced by perianth organs (petals and sepals) resulting in a double floral whorl. An En/Spm-like insertion, termed Tpn-botan, was found in a C class MADS-box-like gene that is evidently responsible for this phenotype. A similar phenotype has also been described in I. purpurea (11), although the molecular basis for the mutation in I. purpurea is unknown.

The vast majority of the phenotypic variation in Ipomoea characterized to date at the molecular level appears to be caused by the insertion or deletion of transposable elements (Table (Table3).3). It is apparent that a wide variety of mobile elements exist in the Ipomoea genome, and these are evidently quite active based on the relatively modest period of evolutionary time that separates I. nil and I. purpurea. We now turn to studies of the population genetics of some flower color phenotypes in I. purpurea.

Geographic Distribution of Flower Color Polymorphisms in I. purpurea.

The geographic distribution of genetic diversity in I. purpurea appears paradoxical. Levels of flower color polymorphism are high in the southeastern U.S. whereas Mexican populations are frequently monomorphic for the blue color form (4, 35). The situation is reversed for biochemical and molecular variation. Mexican populations have levels of isozyme polymorphism that are similar to other annual plants (4, 36), but U.S. populations are depauperate in isozyme variation. Surveys of ribosomal DNA restriction fragment variation and samples of gene sequence data for the (CHS-A) locus also reveal reduced levels of variation in U.S. populations relative to Mexican populations (4, 37). We speculate that this pattern is a consequence of the introduction of horticultural forms selected for flower color diversity into the U.S. However, we do not know the source of these introductions.

There are at least two plausible introduction scenarios. One scenario posits a northward migration of the common morning glory along with maize culture over 1,000 years ago. A second scenario posits an introduction of the common morning glory into the southeastern U.S. by European settlers of this region as a horticultural plant. The genetic evidence points to a strong founder effect associated with the introduction of the common morning glory into the southeastern U.S. Maize does not appear to have experienced so extreme a founder effect, and it is not obvious why the two species would experience different population restrictions during a common migration process, so we regard the first scenario as less likely. Under either scenario, we may regard the southeastern U.S. populations as a crude series of experiments in the microevolution of flower color determining genes.

Epperson and Clegg (35) conducted geographic surveys of flower color variation within the southeastern U.S. at three different spatial scales. The smallest spatial scale (the intrapopulation scale) was analyzed via spatial autocorrelation statistics (38); second, the sub regional scale (defined as local populations that range from 0.8 to 32 km apart) was analyzed via gene frequency distances between populations (39); and third, the regional scale ranging from 80 to 560 km between populations, and including much of the southeastern U.S., was also analyzed by using genetic distance statistics. A strong result of these analyses is that there is no correlation between genetic distance and geographic distance at subregional or regional scales for either the W/w or the P/p loci. Populations separated by a few kilometers are as differentiated from one another with respect to the P/p and W/w flower color determining loci as those separated by hundreds of kilometers. Such a pattern is consistent with the hypothesis that the flower color variants were randomly introduced into multiple locations in the southeastern U.S. Analyses of spatial distributions within local populations led to two major conclusions: first, spatial autocorrelation statistics were heterogeneous between the W/w and P/p loci within the same local populations; and, second, analyses of spatial correlograms for the P/p locus revealed genetic neighborhood sizes consistent with an isolation-by-distance model of population structure (35). We begin by elaborating on the second conclusion; we then turn to the importance of the first conclusion in establishing that the W/w locus is subject to selection within local populations.

For clarity, it is useful to review a few elementary definitions in spatial statistics. A spatial autocorrelation measures the correlation in state of a system at two points that are separated by x distance units. For example, in the common morning glory case, the state may be the flower color determined by the P/p locus x distance units apart. The autocorrelation is often graphed as a function of distance (x), and the resulting graph is called the correlogram [y = f(x)]. For discrete characters, the convention is to transform the autocorrelation into a standard normal deviate with mean 0 and variance 1, and this transformed function is graphed as the correlogram. Let us again take the concrete example of the common morning glory P/p locus. Three graphs result from the autocorrelation of blue with blue (ybb), pink with pink (ypp), and blue with pink (ybp), each as a function of distance. When graphed as a function of distance, the data from many local I. purpurea populations in the southeastern U.S. revealed a common pattern for the P/p locus ypp or ybb > 2 and ybp < −2 initially (for the smallest values of x). This result indicates a positive autocorrelation over short spatial distances among like phenotypes (blue with blue or pink with pink) and a negative autocorrelation at short spatial scales among unlike phenotypes (blue with pink). The distance (x), where the autocorrelation first crosses the x axis (y = 0), provides an operational definition of patch size. These analyses revealed substantial patchiness for the P/p locus and yielded a minimum estimate of about 120 pink (recessive homozygous) plants per patch. The estimates of patch size are highly consistent with simulations of spatial distributions based on a pattern of pollinator flight distances that is strongly biased toward nearest neighbor moves (40, 41).

In contrast to the P/p locus pattern, the W/w locus genotypes show little or no spatial patchiness (that is, y = 0 at the smallest values of x, in populations in which y > 2 or < −2 initially for the P/p locus). Because both loci are transmitted to successive generations within populations through the same mating process, we expect homogeneous spatial patterns. Heterogeneous spatial patterns argue that isolation-by-distance is not the sole factor governing spatial patterns within local populations. Other forces must be invoked to explain the W/w locus pattern, and the most likely appears to be selection against white homozygous plants. Epperson (42) has carried out extensive simulation studies that tend to validate this conclusion, and, further, the simulations suggest an intensity of selection against white homozygotes (ww) of approximately 10%. Despite this conclusion, the frequency of white alleles varies from a low of 0% to a high of 43% across local populations with a mean value of about 10% (35). Because the various local populations can be thought of as quasi-independent experiments in which selection has operated for a number of generations, the persistence of white alleles argues strongly for countervailing selective forces favoring the retention of white phenotypes.

The spatial pattern of albino alleles, including unstable alleles (a*), is of interest because this locus also determines a white recessive phenotype that may experience selective pressures common to those experienced by the ww phenotype. Limited sampling across the southeastern U.S. indicates very low frequencies of albino alleles (<1%) in most local populations (11). In addition, unstable alleles also appear to be rare in most local populations. An exceptional local population near Athens, GA is the only population sampled with moderate frequencies of both albino (a) and unstable (a*) alleles (11).

Selection on Flower Color Phenotypes.

Because flowering plants often depend on insect pollinators for reproductive success, a natural question to investigate is whether pollinators discriminate among the various flower color phenotypes in morning glory populations. A number of experiments conducted in different years and by different investigators in both natural and experimental populations agree in revealing a bias by bumblebee pollinators against visiting white flowers when white is less than 25% of the population (3, 43, 44). In contrast, there is no evidence of pollinator discrimination among P/p (blue/pink) or I/i (dark/intense) locus phenotypes (43, 45). The under visitation of white phenotypes is correlated with an increased frequency of self-fertilization by white maternal parents based on estimates using isozyme marker loci (3, 43). However, Schoen and Clegg (45) discovered that pollinator visitation rates and outcrossing estimates did not differ between white and pigmented phenotypes when the two forms were in equal frequency. Later, Epperson and Clegg (3) and Rausher et al. (44) showed that the pollinator discrimination against whites, and the reduced outcrossing rate of white maternal plants, was frequency-dependent and disappeared when the frequency of the white phenotype approached 50%. These observations and experiments indicate that white loci act as mating system modifier loci and consequently bias their own transmission to subsequent generations.

There is substantial theoretical literature on the population genetics of modifier loci (4648). According to this literature, white genes that increase the frequency of self-fertilization are expected to increase in frequency to fixation within populations because the white gene should be transmitted differentially to progeny of white maternal plants via self-fertilization. This assumes that the selfing (white) gene is also transmitted to the outcross pool in proportion to its frequency in the population, where it can fertilize ovules of non-white maternal phenotypes (absence of pollen discounting). [“Pollen discounting” refers to the situation in which pollen transmission to other (nonmaternal) plants is reduced by a gene promoting self-fertilization, as might be the case when pollinators avoid particular floral displays.] The degree of advantage of self-fertilization genes, if any, clearly hinges on several additional factors, including (i) the extent of pollen discounting; (ii) differential maternal fertility associated with autopollination; and (c) reduced viability of progeny derived through self-fertilization because of inbreeding depression. A number of experiments have been conducted to investigate each of these additional factors.

Rausher et al. (44) found no evidence for pollen discounting in experimental populations of morning glory and instead found a nonsignificant excess of white gene fertilizations of ovules of non-white parents. In a different set of experiments, Epperson and Clegg (unpublished data) found a nonsignificant deficit of white gene fertilizations of ovules of non-white parents. Much larger experiments with greater statistical power are needed to provide a definitive answer about the role of pollen discounting, but at present there is no clear evidence that the advantage of white genes is offset by pollen discounting. Epperson and Clegg (unpublished data) have also measured the fertility of white maternal parents subjected to autopollination and find a reduction in maternal fertility that could act to moderate or eliminate the transmission advantage of white genes. Experiments to measure inbreeding depression provide evidence for a depression in fitness associated with self-fertilization, but the magnitude is not sufficient to offset the selfing advantage of white genes (49). Other studies have suggested heterogeneity among families in segregation bias favoring the transmission of either white or dark alleles in different heterozygous parents (50). Finally, overdominance in seed size among the progeny of white maternal parents has also been documented, but this evidently does not translate into a fitness advantage (51).

In summary, there is clear and convincing evidence that white phenotypes suffer some disadvantage in natural populations based on spatial autocorrelation analyses, and there is clear and convincing evidence that white genes have a transmission advantage when white maternal plants are infrequent in populations. The transmission advantage is associated with pollinator preferences and a consequent increased rate of self-fertilization among white maternal parents, but this advantage is one-sided and should diminish to zero as the frequency of white maternal types approaches 50%. Many lines of evidence argue for a compensatory disadvantage of white types to account for the global frequency in the southeastern U.S. of approximately 10%, but the precise nature of the disadvantage is unresolved. It may be that a number of separate components such as reduced maternal fertility under autopollination, inbreeding depression, and segregation bias all sum to provide the needed balance. Because small effects demand very large experiments to provide adequate statistical power, it may not prove feasible to measure each component of fitness with sufficient accuracy to identify all of the factors that comprise a balance. It is also important to note that, despite much experimental effort over two decades, no evidence has emerged for differential selection among other flower color phenotypes such as those determined by the P/p and I/i loci.

Conclusions

We began this work with the conviction that a full understanding of the mechanisms of adaptation would entail an integration of the ecological and the genetic dimensions of biology. This wisdom was by no means original but, rather, was embodied in Ledyard Stebbins' monumental contributions to plant evolutionary biology. As Stebbins emphasized, the point of contact between these two levels of biological organization is the phenotype. Whether a phenotype is adapted to a particular environment depends on a host of biotic and abiotic factors that collectively translate into survival and reproductive success. It is often difficult to identify the environmental elements that occasion phenotypic success. We began with the notion that flower color provided a simple model for the study of adaptation because one aspect of the environment seemed, a priori, to be predominant—the behavior of insect pollinators. To a large extent a substantial body of experimental work has validated this assumption, but along the way we have been forced to begin to confront the full complexity of plant reproductive biology. In a similar vein, the phenotypic bases of flower color variation appeared to be susceptible to analysis because much was already known about the biochemistry and genetics of flower color determination and because the tools for molecular analysis were rapidly appearing on the horizon. What we did not anticipate was the remarkable ecology of plant genomes in which most genes are redundant and in which mobile elements are a major player in the generation of phenotypic diversity.

One complexity that may be more apparent than real is the problem of genetic redundancy. As noted in this article, most of the genes of flavonoid biosynthesis occur in multiple copies, and sorting through this redundancy to find the particular genes responsible for individual flower color phenotypes appeared daunting at first sight. The actual findings are encouraging, in that particular gene copies of CHS and DFR are shown to be causally responsible for particular flower color phenotypes. In the case of CHS, we know that the other redundant copies are predominantly expressed at other stages in development (e.g., CHS-E in the floral tube) or they have diverged in catalytic properties (e.g., CHS-A, -B, and -C).

Another complexity is the pleiotropic nature of biosynthetic pathways. At first sight it appears that each mutation in a gene encoding a pathway component is likely to have many phenotypic effects. Is it possible to associate a main effect with one aspect of phenotype? Perhaps, if gene redundancy and specialization in expression patterns approximates a one-to-one correspondence between gene and phenotype. Whether the unraveling of phenotypic determination will be so simple remains to be seen, but there is some reason to be optimistic based on flower color analyses.

Another fascinating result of this work is the overwhelming importance of mobile elements as generators of phenotypic diversity. All except one of the flower color phenotypes analyzed so far in both I. purpurea and I. nil are the result of transposon insertions. This strongly implies that transposable elements are the major cause of mutations that yield an obvious phenotype in these plant species. Because the distribution of transposons appears to be heterogeneous across plant genomes, it seems reasonable to speculate that some species may experience higher mutation rates and therefore may be more flexible in adapting to environmental changes than other plant species. In the case of I. purpurea, a significant environmental change was the appearance of human plant domesticators who selected and propagated unusual phenotypes for esthetic and perhaps other purposes. The plant has clearly been successful, at least in the short term, by virtue of its ability to adapt to this circumstance.

A number of questions remain to be resolved before this work can be seen as complete. First, the gene that is clearly subject to selection in nature is not a structural gene determining an enzyme in the anthocyanin pathway but, instead, a regulatory gene that determines the floral distribution of pigmentation (W/w locus). To date, this gene has not been characterized at the molecular level, and we do not know the nature of the mutational changes that cause the white phenotype. We do know a lot about the A/a locus, which also determines a white (albino) phenotype, and we can probably assume that this phenotype is also discriminated against by insect pollinators, but the albino phenotype is rare in populations. So major gaps exist in our present knowledge, but we have every reason to expect these gaps to fill in over time. Interestingly, we do know the molecular bases for the P/p locus phenotypes, but so far there is no evidence of selection in nature at this locus, although in a broader sense it is clear that man has selected this polymorphism for propagation.

A second limitation is that most population work has focused on the southeastern U.S., where the plant is introduced, rather than on native populations in the central highlands of Mexico. The southeastern U.S. is not the geographical region in which the species evolved, and it is not possible to make inferences about the longer term ecological circumstances that shaped the evolution of this species based on investigations in other geographical and ecological settings. Despite this reservation, the southeastern U.S. populations represent a crude series of experiments that trace to introductions within the last several hundred to one thousand years, and the ability to place temporal limits on the history of these populations is a strength of the morning glory program. Common patterns among populations are indicative of common initial conditions or common selective forces. Thus, the lack of spatial autocorrelation for the W/w locus phenotypes over populations is strongly indicative of a disadvantage that is independent of location. Similarly, the case for an isolation-by-distance model of population structure for the P/p locus is strongly supported by the consistency of this result over local populations.

A third consideration is the problem of statistical power. Because the structure of statistical hypothesis testing is deliberately biased against accepting the alternative hypothesis (power is usually much less than the size of the critical region, α, under the null hypothesis), very large experiments must be carried out to detect moderately large effects. Put another way, magnitudes of selection that can be quite effective from an evolutionary standpoint cannot be detected in experiments of reasonable size. This may explain the failure to detect pollen discounting and the failure to detect significant inbreeding depression in experimental populations of I. purpurea. Limited statistical power can in part be overcome by experiments that run over many generations in which the effects are cumulative. This is the strength of the geographical and population studies in the southeastern U.S., which represents the cumulative outcome of many generations rather than the result of single generation experiments.

Despite the limitations noted above, the study of model evolutionary systems is just as important as the study of other model systems. Model systems should reveal the important questions and provide a basis for inventing novel approaches that can then be extended to more refractory systems. As one example, Stebbins pioneered the analysis of plant development as an essential prerequisite to understanding the determination of phenotype and hence to the understanding of adaptation. This approach is just as important today as it was more than 40 years ago, when Stebbins began his classic work with barley awn development as his model system. Today we have a rich suite of molecular tools that assist in unraveling the determination of phenotype and in penetrating the complexities associated with the coordinated action of many genes. The determination of floral color development is an area of special promise because the translation between genes and phenotype is tractable. Similarly, the translation between environment and phenotype is more transparent for flower color than in most other cases, and there is still much to be learned from this research strategy.

Acknowledgments

We thank Dr. Shigeru Iida of the National Institute for Basic Biology in Okazaki, Japan for his critical review and suggestions and for sharing his unpublished data. We also thank Dr. Hiroshi Noguchi from the School of Pharmaceutical Sciences, University of Shizuoka, Shizuoka, Japan for sharing his unpublished data on the substrate specificity testing of the I. purpurea CHS genes. This work was supported in part by a grant from the Alfred P. Sloan Foundation.

Abbreviations

CHS
chalcone synthase
CHI
chalcone isomerase
F3H
flavanone 3β-hydroxylase
DFR
dihydroflavonol reductase
ANS
anthocyanidin synthase

Footnotes

This paper was presented at the National Academy of Sciences colloquium “Variation and Evolution in Plants and Microorganisms: Toward a New Synthesis 50 Years After Stebbins,” held January 27–29, 2000, at the Arnold and Mabel Beckman Center in Irvine, CA.

Morita, Y., Hoshino, A., Tanaka, Y. Kusumi, T., Saito, N. & Iida, S. (1999) Plant Cell Physiol. 40, Suppl., 124 (abstr.).

§Hoshino, A. & Iida, S. (1997) Genes Genet. Syst. 72, 422 (abstr.).

Hoshino, A. & Iida, S. (1999) Plant Cell Physiol. 40, Suppl., 26 (abstr.).

Nitasaka, E. (1997) Genes Genet. Syst. 72, 421 (abstr.).

References

1. Clegg M T. Limits to Knowledge in Evolutionary Genetics. 2000. , ed. M. T. Clegg (Plenum, New York), in press.
2. Austin D F, Huaman Z. Taxon. 1996;45:3–38.
3. Epperson B K, Clegg M T. Evolution (Lawrence, Kans) 1987;41:1302–1311.
4. Glover D E, Durbin M L, Huttley G, Clegg M T. Plant Species Biol. 1996;11:41–50.
5. Ennos R A, Clegg M T. J Hered. 1983;74:247–250.
6. Epperson B K, Clegg M T. J Hered. 1988;79:64–68.
7. Barker E E. Bull Cornell Univ Agric Exp Station. 1917;392:121–154.
8. Imai Y. J Coll Agric Imp Univ Tokyo. 1927;9:199–222.
9. Schoen D J, Giannasi D E, Ennos R A, Clegg M T. J Hered. 1984;75:113–116.
10. Epperson B K, Clegg M T. J Hered. 1987;78:346–352.
11. Epperson B K, Clegg M T. J Hered. 1992;83:405–409.
12. Koes R E, Quattrocchio F, Mol J N M. BioEssays. 1994;16:123–132.
13. Kreuzaler F, Hahlbrock K. Arch Biochem Biophys. 1975;169:84–90. [PubMed]
14. Reinhold U, Krüger M, Kreuzaler F, Hahlbrock K. EMBO J. 1983;2:1801–1805. [PMC free article] [PubMed]
15. O'Neill S D, Tong Y, Sporlein B, Forkmann G, Yoder J. Mol Gen Genet. 1990;224:279–288. [PubMed]
16. Durbin M L, Learn G H, Huttley G A, Clegg M T. Proc Natl Acad Sci USA. 1995;92:3338–3342. [PMC free article] [PubMed]
17. Koes R E, Spelt C E, Mol J N M, Gerats A G M. Plant Mol Biol. 1987;10:159–169. [PubMed]
18. Fukada-Tanaka S, Hoshino A, Hisatomi Y, Habu Y, Hasebe M, Iida S. Plant Cell Physiol. 1997;38:754–758. [PubMed]
19. Habu Y, Fukada-Tanaka S, Hisatomi Y, Iida S. Biochem Biophys Res Commun. 1997;234:516–521. [PubMed]
20. Shiokawa, K., Inagaki, Y., Morita, H., Hsu, T., Iida, S. & Noguchi, H. (2000) Plant Biotechnol., in press.
21. Durbin M L, McCaig B, Clegg M T. Plant Mol Biol. 2000;42:79–92. [PubMed]
22. Johzuka-Hisatomi Y, Hoshino A, Mori T, Habu Y, Iida S. Genes Genet Syst. 1999;74:141–147. [PubMed]
23. Habu Y, Hisatomi Y, Iida S. Plant J. 1998;16:371–376. [PubMed]
24. Miles C O, Main L. J. Chem. Soc. Perkin Trans. 2. 1985. 1639–1642.
25. McCaig B. Ph.D. dissertation. Riverside: Univ. of California; 1998.
26. Hoshino A, Abe Y, Saito N, Inagaki Y, Iida S. Plant Cell Physiol. 1997;38:970–974. [PubMed]
27. Inagaki Y, Johzuka-Hisatomi Y, Mori T, Takahashi S, Hayakawa Y, Peyachoknagul S, Ozeki Y, Iida S. Gene. 1999;226:181–188. [PubMed]
28. Inagaki Y, Hisatomi Y, Suzuki T, Kasahara K, Iida S. Plant Cell. 1994;6:375–383. [PMC free article] [PubMed]
29. Takahashi S, Inagaki Y, Satoh H, Hoshino A, Iida S. Mol Gen Genet. 1999;261:447–451. [PubMed]
30. Lönnig W-E, Saedler H. Gene. 1997;205:245–253. [PubMed]
31. Abe Y, Hoshino A, Iida S. Genes Genet Syst. 1997;72:57–62.
32. Wessler S R, Bureau T E, White S E. Curr Opin Gen Dev. 1995;5:814–821. [PubMed]
33. Hisatomi Y, Yoneda Y, Kasahara K, Inagaki Y, Iida S. Theor Appl Genet. 1997;95:509–515.
34. Hisatomi Y, Hanada K, Iida S. Theor Appl Genet. 1997;95:1049–1056.
35. Epperson B K, Clegg M T. Am Nat. 1986;128:840–858.
36. Hamrick J L, Godt M J. In: Plant Population Genetics, Breeding, and Genetic Resources. Brown A D H, Clegg M T, Kahler A L, Weir B S, editors. Sunderland, MA: Sinauer; 1989. pp. 43–63.
37. Huttley G A, Durbin M L, Glover D E, Clegg M T. Mol Ecol. 1997;6:549–558.
38. Sokal R R, Oden N L. J Linn Soc. 1978a;10:199–228.
39. Nei M. Am Nat. 1972;106:283–292.
40. Turner M E, Stephens J C, Anderson W W. Proc Natl Acad Sci USA. 1982;79:203–207. [PMC free article] [PubMed]
41. Sokal R R, Wartenberg D E. Genetics. 1983;105:219–237. [PMC free article] [PubMed]
42. Epperson B K. Genetics. 1990;124:757–771. [PMC free article] [PubMed]
43. Brown B A, Clegg M T. Evolution (Lawrence, Kans) 1984;38:796–803.
44. Rausher M C, Augustine D, Vanderkooi A. Evolution (Lawrence, Kans) 1993;47:1688–1695.
45. Schoen D J, Clegg M T. Evolution (Lawrence, Kans) 1985;29:1242–1249.
46. Fisher R. Ann Eugen. 1941;11:53–63.
47. Karlin S, McGregor J. Theor Pop Biol. 1974;5:59–103. [PubMed]
48. Holsinger K E. In: Evolutionary Biology. Hecht M K, MacIntyre R J, Clegg M T, editors. New York: Plenum; 1996. pp. 107–149.
49. Chang S M, Rausher M D. Evolution (Lawrence, Kans) 1999;53:1366–1376.
50. Fry J D, Rausher M D. Evolution (Lawrence, Kans) 1997;51:66–78.
51. Mojonnier L E, Rausher M D. Evolution (Lawrence, Kans) 1997;51:608–614.

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...