Genetic Similarity Between Gaura longiflora and Its Obligately Outcrossing Derivative G. demareei

Analysis of electrophoretic variation in ten enzyme systems showed that the diploid annual plant Gaura demareei is highly similar to its progenitor G. longiflora. In both species, 12 genes were identically monomorphic and six polymorphic genes contained similar complements of alleles having similar frequencies. Since these are self-incompatible, the study demonstrated that speciation in annual plants that are obligately outcrossing can occur with minimal genetic differentiation and accords with previous results for self-pollinating species. The high degree of genetic similarity between Clarkia lingulata Lewis 8 Lewis and its progenitor C. biloba Nelson 8 Macbride (Gottlieb, 1974) and between a earlier designated as Malheurensis and its progenitor Stephanomeria exigua ssp. coronaria Gottlieb (Gottlieb, 1973a) suggests that diploid annual plant are extracted directly from the repertoire of phenotypic variation and genetic polymorphisms of their progenitors and that speciation in these organisms does not involve early reconstitution of the derivative genome. However, both C. lingulata and Malheurensis have a self-compatible breeding system that permits self-fertilization. Autogamy has been associated with accelerated formation of species in many groups of annual plants (Grant, 1963) and, therefore, it is possible that these two utilized pathways of speciation different from those generally taken by obligate outcrossers. This possibility raises the question whether plants with self-compatible breeding systems may exhibit less genetic divergence from their progenitors than obligately outcrossing species. Self-fertilization reduces the number of generations required to establish chromosomally homozygous lines in the aftermath of chromosomal rearrangement (the primary mode of reproductive isolation in annual plants) and permits rapid increases in genetic homozygosity. If adaptive genotypes are associated with the new chromosomal arrangements, it acts to perpetuate them without the inevitable recomnbination that would result from outcrossing to individuals of a progenitor population. In addition, it increases the probability of establishment following long-distance dispersal and thereby may provide a prelude to speciation. These points are discussed more fully by Stebbins (1957) and Grant (1963). Stebbins (1957) also ' We thank Peter Raven for suggesting Gaura to us and both him and Peter Hoch for collecting seeds. This research was supported by grant GB-39873 from the National Science Foundation. ' Genetics, University of California, Davis CA 95616. This content downloaded from 207.46.13.149 on Mon, 03 Oct 2016 05:44:50 UTC All use subject to http://about.jstor.org/terms 182 SYSTEMATIC BOTANY [Volume 1 showed how self-fertilization of an F1 hybrid, which was heterozygous for chromosomal or genetic factors affecting fertility, could lead to the production of new homozygous combinations and the possible formation of reproductively isolated lines. Grant (1966a, b) tested this model and demonstrated that inbred lines descendent from an experimental interspecific hybrid rapidly acquired structural homozygosity. Levin (1970) proposed that inbreeding in peripheral isolates might catalyze the formation of novel phenotypes otherwise suppressed by various regulatory and canalization devices. Grant (1956) and Stebbins (1957) also commented on the significance of self-fertilization in the formation of polyploid species. Thus, since self-fertilization appears to facilitate speciation, we tested the hypothesis that the high degree of genetic similarity between progenitor and derivative in Clarkia and Stephanomeria would not be observed when an obligate outcrosser was compared to its progenitor. Very few examples have been identified in which both progenitor and derivative are self-incompatible. The few unambiguous phylogenies of this type generally involve aneuploid changes in chromosome number, for example, in Crepis (Tobgy, 1943; Sherman, 1946), Holocarpha (Clausen, 1951), Haplopappus (Jackson, 1962), Chaenactis (Kyhos, 1965), and Calycadenia (Carr, 1975). These were either unavailable or unsuitable for our purposes, so we selected the two diploid self-incompatible annual Gaura longiflora Spach (Onagraceae) and its recent derivative G. demareei Raven G flowers of G. demareei are larger, open at sunrise, and are pollinated by bees; those of G. longiflora open at sunset and are pollinated by moths (Raven 8c Gregory, 1972a). Both have the same chromosome number (n = 7). They are structurally heterozygous for several reciprocal translocations, but at meiosis the chromosomes regularly disjoin in an alternate manner so that there is no reduction in fertility. Cytogenetic studies of interspecific F1 hybrids between them indicate that they differ by at least two reciprocal translocations; however, their mean fertility (about 70%) is higher than that usually associated with this amount of chromosomal repatterning, although This content downloaded from 207.46.13.149 on Mon, 03 Oct 2016 05:44:50 UTC All use subject to http://about.jstor.org/terms 1976] GOTTLIEB & PILZ: GENETICS OF GAURA 183 it is not too dissimilar from that found in hybrids between G. longiflora and other closely related (Raven, pers. comm.). For the present study, fruits of 30-40 plants from each of three populations of G. longiflora and two populations of G. dentareei were collected in the fall of 1974 by Raven and Hoch. The populations of G. longiflora from Lonoke County (Raven 26492) and Saline County, Arkansas (Hoch 421) have strigulose pubescence in the inflorescence, typical of southern populations of the species, and the one from Jefferson County, Missouri (Raven 26494) has glandular pubescence characteristic of northern populations. The two populations of G. demareei were collected from different parts of its limited distribution: Hempstead County (Raven 26487) and Garland County, Arkansas (Hoch 420). (Representative specimens are in the Missouri Botanical Garden herbarium.) Seeds were excised from the fruits with a scalpel and germinated at ca. 15 C in moist vermiculite. Electrophoretic variation in ten enzyme systems was examined: leucine aminopeptidase (LAP-1, LAP-2), peroxidase (PER-1, PER-2), tetrazolium oxidase (TO), esterase (EST1, EST-2), glutamate oxaloacetate transaminase (GOT-1, GOT-2), malate dehydrogenase (MDH), 6-phosphogluconate dehydrogenase (6-PGD-1, 6-PGD-2), phosphoglucoisomerase (PGI-2), phosphoglucomutase (PGM-1, PGM-2), and alcohol dehydrogenase (ADH-1, ADH-2, ADH-3). The enzymes were present in crude extracts of rosette leaves (PER was extracted from roots) from fourto six-week-old plants. The procedures for extraction and electrophoresis were similar to those described by Gottlieb (1974). All systems were examined in each individual. Extracts from plants from the different populations were run side by side in several combinations to compare enzyme mobilities. With the buffer system employed, all of the enzymes migrated to the anode except PER, which migrated to the cathode. PGI was also present anodal to PGI-2 with the same mobilities in both species. However, the isozymes were not always clearly resolved and their variation is not reported here. Assays for LAP, EST, MDH, ADH, and 6-PGD were described by Gottlieb (1974); those for GOT and PGI by Gottlieb (1973b); that for PER in Gottlieb (1973a), except that the gel buffer was the standard Tris-citric/ lithium borate; TO appears as white bands on the blue background of 6-PGD; the assay for PGM was 90 ml 0.05 M Tris HCI pH 8.0, 75 mg disodium a-D-glucose-1,6-diphosphate,, 5 ml 0.00017 M dipotassium a-Dglucose-1,6-diphosphate, 5 ml 0.1 M MgCl2, 40 units G6PD, 5 mg NADP, 10 mg MTT, and 2 mg PMS. RESULTS AND DISCUSSION Genetic Variation and Similarity.-All of the individuals examined in both had enzymes with identical electrophoretic mobilities for LAP, PER, TO, GOT, MDH, PGM, and 6-PGD. These enzymes are presumed to be specified by monomorphic genes: a single gene each for MDH and TO since each individual had only a single band of enzyme activity, This content downloaded from 207.46.13.149 on Mon, 03 Oct 2016 05:44:50 UTC All use subject to http://about.jstor.org/terms 184 SYSTEMATIC BOTANY [Volume 1 TABLE 1. Observed frequencies of alleles at six polymorphic genes and observed mean proportion of the genome heterozygous per individual (of 18 genes examined) in populations of G. demareei and G. longiflora. N = number of individuals examined. G. demareei G. longiflora Gene/allele GAR HEM SAL LON JEF N= 61 N=45 N= 14 N=23 N=40 EST-1 1 .15 .01 .05 .18 * 2 .78 .97 .85 .80 3 .07 .02 .10 .01 4 .01 EST-2 1 .01 .02 .30 2 .83 .87 .71 .57 .78 3 .16 .11 .29 .13 .22 PGI-2 1 .04 .04 2 1.00 .79 .50 .96 .67 3 .21 .50 .23 4 .06 ADH-1 1 .04 2 .99 .95 .93 .95 .75 3 .01 .07 .05 .25 4 .01 ADH-2 1 .13 .05 2 .74 1.00 1.00 1.00 .90 3 .13 .05 ADH-3 1 .02 2 1.00 1.00 1.00 1.00 .92 3 .05 Heterozygous .061 .040 .072 .060 .091