Artificial Selection in Brassica rapa

Transcription

1 Artificial Selection in Brassica rapa Biology 100 Advances in Biology in the Age of the Human Genome Project I242 Light, the Universe, and Everything Spring 2002

2

3 Table of Contents Artificial Selection in Brassica rapa Introduction... 1 Selective Breeding... 1 Artificial vs. Natural Selection... 2 Research Origins of Fast Plants... 4 Life Cycle of rapid cycling Brassica rapa... 5 Variable Traits in Brassica rapa Fast Plants and their Butterfly, Pieris rapae Our Experiment Sowing Fast Plant Seed Counting Hairs Selecting the parents of the next (second) generation Pollination of selected plants Harvesting seeds Establishing the second generation Counting the hairs in the second generation Data Analysis Inheritance Patterns Analyzing the data to describe the variation in the population Thinking about variation Writing the report General Guidelines The Conventional Format... 31

4

5 Artificial Selection in Brassica rapa Adapted and edited by Katherine Dorfman 1 Introduction Selective Breeding A trip to a supermarket, farm, pet store or garden center will offer nearly endless examples of the products of selective plant and animal breeding by humans. Over hundreds, and in some cases thousands of years, humans have altered various species of plants and animals for our own use by selecting individuals for breeding that possessed certain desirable characteristics, and continuing this selective breeding process generation after generation. In many cases the end results have been dramatic. Domestic dog varieties, from Chihuahuas to great Danes, trace their separate lineages to a common wild ancestor, the wolf (Canis lupus). Domestic fowl varieties are all derived from the wild jungle fowl (Gallus gallus), while most modern breeds of domestic cattle originate from the now-extinct giant wild ox (Bos primigenius). One example that particularly impressed Charles Darwin was that of domestic pigeon varieties (such as tumblers, fantails, carriers, pouters, and many others), which are derived from wild rock doves (Columba livia) over a period of some 5000 years. There are many similar examples among plants, including those that humans have bred for food as well as beauty. One plant group especially important to humans for food is Brassica genus of plants in the mustard family. A wide variety of familiar and highly nutritious vegetables originate from just a few species of wild Brassicas, in particular B. rapa, B. oleracea, and B. junce. Some varieties have been bred specifically for root production, others for leaves, flower buds, oil production, etc. The group is of great economic importance, and because of this, the genetic relationships among the various forms have been thoroughly studied and are now fairly well understood. It may surprise you to learn that these familiar vegetables so different in appearance have the same species as a common ancestor. Centuries of artificial selection have produced such widely divergent forms. For example, the following varieties originate from these wild species (see Figure 1): Brassica oleracea: cauliflower, broccoli, cabbage, Brussels sprouts, kohlrabi, collards, savoy cabbage, kale Brassica juncea: leaf mustard root mustard, head mustard, and many more varieties of mustard Brassica rapa: turnip, Chinese cabbage, pak choi, rapid-cycling Figure 1. Varieties (cultivar groups) of Brassica rapa.

6 2 Artificial Selection in Brassica rapa The next time you re in a grocery with a large produce section or a large farmers market, see how many of these different varieties you can find; there are numerous other varieties not listed above. That plant and animal breeders have been able to dramatically change the appearance of various lineages of organisms in a relatively short period of time is an obvious yet profound fact. It certainly did not escape the attention of Charles Darwin, who devoted the first chapter of The Origin of Species to the topic of artificial selection by humans ( Variation Under Domestication ). He used many examples of selection by humans to help support the case for his proposed mechanism resulting in evolution of natural populations natural selection. In order to gain better understanding of selection and inheritance, researchers have experimented with artificial selection involving a wide variety of traits in many different species of plants and animals. The results, obtained in a relatively short period of time, are often impressive. For example, in one experiment (started in 1896), the average oil content of corn kernels was increased from 5% to 19% in 75 generations of selective breeding. In another, average annual egg production in a flock of white leghorn chickens was doubled in only 33 years (from 126 to 250 eggs per hen). Similar examples abound, for characters useful to humans (as above) as well as more esoteric ones (such as the number of abdominal bristles on fruit flies, or the fecundity of female flour beetles). A general finding from these studies is that most variable traits in organisms respond to artificial selection (i.e., it is usually possible even easy to increase or decrease the frequency or average value of a trait in a lineage through careful selective breeding). In this lab exercise, you will attempt to accomplish the same thing. Artificial vs. Natural Selection Natural selection is a deceptively simple concept, relatively easy to understand at a basic level, but with profound implications that are intellectually challenging. The following exercise in artificial selection will serve as an introduction to natural selection, and we hope will help you to better understand this important concept. Fundamentally, artificial selection and natural selection are quite similar, but there are a few important differences. Natural selection Very briefly, natural selection occurs when certain variants within a population of organisms experience consistently greater reproductive success (i.e., leave more offspring) than do other variants. Generally, within a population of organisms (of the same species), there is variation among individuals for a great variety of obvious and not so obvious traits. Some of this variability is a result of genetic differences among individuals, while some is probably a result of different environmental influences. Here we are concerned only with variability that has a genetic basis (i.e., is inheritable passed on from parents to offspring). As a general rule, organisms produce more (often far more) offspring than can survive to reproduce themselves (in Darwin s words, there is a struggle for existence ). If individuals with certain inheritable characteristics have an advantage in survival and reproduction over individuals with other characteristics, then it follows that the next generation will differ fro the previous one perhaps slightly, perhaps considerably. Why? Because individuals with

7 Mount Holyoke College, Spring those certain favored inheritable characteristics are more likely to be parents by virtue of possessing these characteristics; the next generation will contain a disproportionate number of offspring of these naturally selected parents, and these offspring will tend to resemble their parents. As a result, over many generations the genetic structure of the population changes (in other words, the population evolves). Artificial Selection Artificial selection is essentially this same process. A population of organisms exhibits considerable variability among individuals for a trait or traits that might be of interest to humans for one reason or another (such as running speed in thoroughbred horses, petal size and shape in tulips, number of kernels in an ear of corn, egg size in chickens, etc.). By virtue of possession of certain traits, some individuals are selected by the plant or animal breeder or scientist to be parents of the next generation, while the remainder of the population not possessing those traits are excluded from breeding. To the degree that the desired trait is inheritable, the offspring will tend to resemble their parents, and hence on average the next generation will differ from the previous one i.e., will be faster, or have larger petals, or more kernels, or larger eggs, etc. Over time, substantial differences can be achieved between the original population and the artificially selected subsequent ones. As noted above, there are some important differences between artificial and natural selection. In contrast to natural selection, artificial selection: 1) favors traits that for one reason or another are preferred by humans; 2) has a goal or direction toward which the selection process is directed; 3) generally is much faster than natural selection, because the next generation can be absolutely restricted to offspring of parents that meet the desired criteria (rarely is natural selection such an all-or-none phenomenon). In artificial selection, humans are doing the selecting purposely restricting breeding to individuals with certain characteristics. In natural selection, the environment does the selecting individuals that survive and reproduce better in a given environment, for whatever reason, are naturally selected. The environment includes such influences as competitors, parasites, predators, food supply, climatic conditions, soil nutrients, and many, many others.

8 4 Artificial Selection in Brassica rapa Research Origins of Fast Plants 2 Rapid-cycling Brassica rapa, Fast Plants, were developed by Dr. Paul H. Williams, Professor of Plant Pathology at the University of Wisconsin-Madison. Fast Plants are a product of Dr. Williams research to improve the disease resistance of plants in the family Cruciferae, a large diverse group that includes mustards, radishes, cabbages, and other cole crops. In order to speed up the genetic research in the crucifers, he began breeding Brassica rapa and six related species from the family Cruciferae for shorter life cycles. The end result was a genetic line of small, prolific, rapid-cycling Brassicas. These plants are now known as Fast Plants. Dr. Williams continued to refine the rapid-cycling Brassicas to have characteristics most suitable for laboratory and classroom use. He selected seed from stock that met the following criteria: Shortest time from planting to flowering Ability to produce seeds at high planting density Petite plant type Rapid seed maturation Absence of seed dormancy Ability to grow under continuous fluorescent lighting Ability to grow well in a potting mix After about 20 years of planting, growing, and selecting, his breeding process had reduced a six-month life cycle to five weeks. Further breeding produced relative uniformity in flowering time, size, and growing conditions and yet the Fast Plants retained much variety. Over 150 genetically controlled traits have been recorded and are useful in research. These Brassicas have shortened traditional breeding programs and aided in cellular and molecular plant research. Additionally, Fast Plants have become extremely useful as a teaching tool in the classroom since all aspects of plant growth and development can be easily demonstrated. For more information on the development of Fast Plants and other rapid-cycling Brassicas see: Paul H. Williams, Curtis B. Hill, Rapid-Cycling Populations of Brassica. Science, New Series, vol 232 (4756),

9 Mount Holyoke College, Spring Life Cycle of rapid cycling Brassica rapa Under proper growing conditions, the life cycle is days from planting a seed to harvesting the next generation, as summarized in Figure 2, Figure 3, and Table 1. Figure 2. Life Cycle of rapid cycling Brassica rapa Figure 3 (overleaf): Growth of Fast Plants

10 6 Artificial Selection in Brassica rapa Insert Figure 3 (overleaf): Growth of Fast Plants Here.

11 Mount Holyoke College, Spring Table 1. Stages in the Life Cycle of Fast Plants: Concepts of Dependency 3 Stage State Condition Dependency A. seed quiescence suspended growth of independent of the parent and many (dormant embryo) embryo components of the environment B. germinating seed germination awakening of growth dependent on environment and health of the individual C. vegetative growth growth and development D. immature plant flower bud development roots, stems, leaves grow rapidly; plant is sexually immature Gametogenesis; reproductive [male (pollen) and female (egg)] cell production E. mature plant flowering mating Pollination: attracting or capturing pollen F. mature plant pollen growth gamete maturation; germination and growth of pollen tube G.mature plant H. mature parent developing fruit, plant plus embryo endosperm, and embryo I. aging parent plant plus maturing embryo J. dead parent plant plus seed dependent on environment dependent on healthy vegetative plant dependent on pollen carriers (bees and other insects) dependent on compatibility of pollen with stigma and style double fertilization union of gametes; dependent on compatibility and healthy union of sperm (n) and egg plant (n) to produce diploid zygote (2n); union of sperm (n) and fusion nucleus (2n) to produce endosperm (3n) Embryogenesis; growth and development of endosperm and embryo; growth of supporting parental tissue of the fruit (pod) interdependency among developing embryo, endosperm, developing pod and supporting mature parental plant senescence of withering of leaves of parent seed is becoming independent of the parent; plant; parent maturation of fruit; yellowing pods; seed development drying embryo; suspension of embryo growth; development of seed coat Death; Desiccation; seed quiescence drying of all plant parts; dry pods will disperse seeds seed (embryo) is independent of parent, but is dependent on the pod and the environment for dispersal

12 8 Artificial Selection in Brassica rapa Days 1-3: Germination and Emergence 4 Germination is the awakening of a seed from a resting state. This resting state represents a pause in growth of the embryo. The resumption of growth, or germination, involves the harnessing of energy stored within the seed. Germination requires at least water, oxygen and a suitable temperature. The radicle (embryonic root) emerges. Seedlings emerge from the soil. Two cotyledons (seed leaves) appear and the hypocotyl (embryonic stem) extends upward. Chlorophyll and anthocyanin (green and purple pigments, respectively) can be seen. See Figure 4. Figure 4. Germination and Emergence of Brassica rapa. Days 4-12: Growth and Development As a plant germinates and matures, it undergoes the processes of growth and development. Growth arises from the addition of new cells and the increase in their size. Development is the result of cells differentiating into a diversity of tissues that makes up organs such as roots, shoots, leaves and flowers. Each of these organs has specialized functions coordinated to enable the individual plant to complete its life cycle. Days 4-9: Cotyledons enlarge. True leaves appear and develop. Flower buds appear in the growing tip of the plant. See Figure 5. Both genetics and the environment play fundamental roles in regulating growth and development, and determining the phenotype (appearance) of an organism.

13 Mount Holyoke College, Spring Figure 5. Development of true leaves Days 10-12: In Fast Plants, growth and development occur rapidly and continuously throughout the life cycle of the individual. It is most dramatic in the 10 to 12 days between seedling emergence (arrival at the soil level) and the opening of the first flowers. The stem elongates between the nodes (points of leaf attachment), and flower buds rise above the leaves, as shown in Figure 6. Leaves and flower buds continue to enlarge. Figure 6. Fast Plant, just before flowering.

14 10 Artificial Selection in Brassica rapa Days 13-17: Flowering and Reproduction Flowering is the initiation of sexual reproduction by the plant. The production of male and female gametes (sperm and eggs) within the flower is the first step in preparing for pollination and reproduction. Each part of the flower has a specific role to play in sexual reproduction. The flower dictates the mating strategy of the species. One common strategy involves employing animals, often insects, to carry pollen (containing male gametes) to the pistil (female reproductive organ). The structure of a Fast Plant flower is shown in Figure 6. Figure 6. Flower Structure 5 Pollination is the process of mating in plants whereby pollen grains developed in the anthers are transferred to the stigma. The pollen grains then germinate, forming pollen tubes that carry the sperm from the pollen to the eggs that lie within the pistil.

15 Mount Holyoke College, Spring Flowers can be cross-pollinated (from one plant to another) for 3-4 days. Pollen is viable for 4-5 days and stigmas remain receptive to pollen for 2-3 days after a flower opens. After final pollination, the remaining unopened flower buds and side shoots should be pruned off to direct the plant's energy into developing the seed. See Figure 7. Figure 7. Reproductively mature plant Days 18-40: Fertilization to Seeds Fertilization is the final event in sexual reproduction. In higher plants, two sperm from the pollen grain are involved in fertilization. One fertilizes the egg to produce the zygote and begin the new generation. The other sperm combines with the fusion nucleus to produce the special tissue (endosperm) that nourishes the developing embryo. In some plants endosperm nourishes the germinating seedling. Fertilization also stimulates the growth of the maternal tissue (seed pod or fruit) supporting the developing seed. See Figure 8. Within 24 hours of pollination, the sperm has fertilized the egg. Immediately following fertilization, the pistil and other maternal structures of the flower will grow and change functions. Within the ovules, the embryos also grow and differentiate through a series of developmental stages known collectively as embryogenesis.

16 12 Artificial Selection in Brassica rapa Figure 8. Pod development Days 18-22: Petals drop from the pollinated flowers. Pods elongate and swell. Development of the seed and young plant has begun and will continue until approximately Day 36. Days 23-38: Seeds mature and ripen. Lower leaves yellow and dry. Twenty days after final pollination (about Day 38) plants should be removed from water. Days 38-45: Plants dry down and pods turn yellow. On Day 45, pods can be removed from dried plants. Seed can be harvested. The cycle is complete.

17 Mount Holyoke College, Spring Variable Traits in Brassica rapa One can observe several fairly obvious variable traits in a large population of mature Fast Plants, such as: arrangement of flowers, total number of flowers, total number of leaves, arrangement of leaves, length of the lower leaf, surface area of the lower leaf, plant height, total number of seeds per plant, total mass of plant, stem length between first and second leaves, etc. In contrast, there are other traits that don t appear to vary at all, such as: number of petals per flower (4); number of sepals (green petal-like structures just below the petals) per flower (4); petal color (bright yellow); and number of cotyledons or seed-leaves (2). When you first have plants to observe, look carefully at them and note the variability you can see. Remember, all your plants are almost exactly the same age, so differences you see (such as height, leaf size, etc.) are not due to differences in age. Figure 9 shows a typical Fast Plant at about two weeks old, in which the cotyledons are distinguished from the true leaves. Figure 9. Fast Plant, about 14 days old. One variable trait that you might not have noted in your list is hairiness of leaves, petioles (leaf stalks) and stems, but if you look more closely (especially with a hand lens and desk lamp) you should see these hairs on some plants. Technically, plant hairs are called trichomes, and they have been shown to have a specific function. What do you suppose this function might be? Fast Plants and their Butterfly, Pieris rapae 6 Pieris rapae, the ubiquitous cabbage white butterfly, is found across North America in many natural places and in gardens with cabbage, broccoli, and other crucifers. It hatches as a little caterpillar from a tiny egg laid on the surface of a leaf by the mother. It goes through 5 such larval stages in about 18 days, finally pupating and metamorphosing into an adult, which emerges from the pupal case about 8 days later. The adult can live up to three weeks, making the entire life cycle less than 2 months long. Read: Jon Agren, Doublas W. Schemske, The cost of defense against herbivores: An experimental study of trichome production in Brassica rapa. American Naturalist, Vol 141 (2): , and Jon Agren, Doublas W. Schemske, Evolution of trichome number in a naturalized population of Brassica rapa, American Naturalist, Vol 143 (1):1-13 for more information on the relationship between Pieris and Brassica. Figure 10 (overleaf). Life cycle of Brassica rapa and Pieris rapae.

18 14 Artificial Selection in Brassica rapa Paste part 1 of Figure 10 (overleaf). Life cycle of Brassica rapa and Pieris rapae. Here.

19 Mount Holyoke College, Spring Paste part 2 of Figure 10 (overleaf). Life cycle of Brassica rapa and Pieris rapae. Here.

20 16 Artificial Selection in Brassica rapa Sowing Fast Plant Seed Our Experiment Label your pots, following the example of the first one. Put a MiraCloth wick in each pot. Make sure the pointed end comes out through the hole in the bottom of the pot, and the top extends about 1 mm above the top of the pot Fill each pot halfway with moist soil mix. Put a fertilizer pellet on the soil in each pot. Add more soil, up to the bottom of the pot lip. Position the wicks so they extend out of the soil by about 1 mm. Use the butt end of a dissecting needle to make a little depression in the soil for the seed. Touch a toothpick to a glue stick, then pick up one or a few seeds with the sticky end. Carefully place one seed in the little depression, flicking it off the toothpick with the dissecting needle. Sprinkle vermiculite over the seed to cover it. Put water into each pot with a plastic transfer pipet, until it drips out of the bottom. Decide on a watering schedule, to ensure that the soil is kept moist. Take the pots upstairs to the environmental chamber. Conscientiously follow the watering schedule over the next week, taking care not to tip over any pots.

21 Mount Holyoke College, Spring Counting Hairs As you have probably guessed, the variable trait that you are going to attempt to alter in this lineage is hairiness. In this wild type population, some plants should be noticeably hairy, many are slightly hairy, while others are apparently hairless. But such an observation is too general, and needs further quantification. To accomplish this, you could count all hairs on all parts of each plant, but this would be a time-consuming task, and an unnecessary one. It turns out that hairiness of one part of a plant (such as the leaf surface) is strongly correlated with hairiness on other parts (such as the leaf margin). In other words, a plant s hairiness in general can be quantified by assessing hairiness of a specific structure. The structure we will use for this hairiness index is the petiole, or leafstalk, where trichomes are large, conspicuous and rather easily counted, and the structure is relatively small with a defined starting and ending point. For consistency, we will use the petiole of the first (lowermost) true leaf, and we will define the limits of the petiole as follows: from its junction with the main stem (usually marked by a small bulge or ridge, often differently colored)to its junction with the lowermost leaf vein (see Figure 11). Figure 11. The limits of the petiole An important note: the two lowermost squarish, two-lobed and rather thick leaves are actually cotyledons ( seed leaves ), not true leaves. The first true leaf is just above these two cotyledons. What you need to do now is to count the total number of trichomes on the petiole of the first true leaf of each plant in your sample. Your data will then be combined with data from the rest of the section. Use the hand lens and desk lamp; the trichomes are most conspicuous if strongly illuminated against a dark background, such as the black table top. If present, the trichomes will generally be concentrated on the lower side of the petiole, but some could occur on the sides or top as well. Be sure to rotate the plant so that you can see all aspects of its petiole. Count a second time for verification, then record your data in your data sheet.

22 18 Artificial Selection in Brassica rapa Reliability of hair counts The number of hairs we count in this experiment may vary from one observer to another (inter-observer reliability), or between counts performed by the same person (intra-observer reliability). So we can never be entirely certain about the number of hairs on a given plant. Take some time to think about why the count includes this inherent variability. These kinds of uncertainties are often called error, but that is not meant to imply mistakes on the part of the experimenter (or the equipment manufacturer). Making multiple measurements is a way of acknowledging the uncertainty in each one, and reducing the risk of letting a single odd measurement carry too much weight. We then represent our result as an average of the many repeated measurements we took. But this glosses over the variation among measurements. We don t have time to count each plant a dozen times in this lab period, so we will compromise as follows: To improve the accuracy of our counting, the hairs on each plant will be counted by two students, and the counters will collaborate (or enlist the aid of a third counter) to come to a fair estimate of the actual number of hairs. Make a data table for yourself, in which you record the plant number and number of hairs. After you have adjudicated the number of hairs with your counting partners, enter your data in the official data sheet. Enter your data in the spread sheet on the computer. Compiling the data When all groups have finished assessing trichome number of individual plants, your lab instructor will coordinate the effort of combining the separate data into a single data table. After tabulating the data from all sections, your instructor will make this summary available to you. Selecting the parents of the next (second) generation Only a small fraction (about 10%) of this population of plants will be selected to be parents of the next generation. These won t be randomly chosen, but instead will be the hairiest 10% of plants in the population, and the 10% least hairy. Sort the data table by number of hairs. Find the top 10% in hairiness. Print out or write down their ID numbers. If more than 10% of the population has no hairs, then use a randomizing technique to choose the 10% to be parents. Your instructor may have created a column in the spread sheet with random numbers in it; pick the 10% of the population that is bald and has the highest numbers in this column. (You may also pick numbers out of a hat or toss dice or coins to pick the winners in this selection process.) Indicate on the data sheet which plants have been chosen as breeders for which group. Separate the chosen plants into hairy parents and bald parents, and dump the contents of the remaining pots into the compost pile. From now on the two sets of parents must be kept strictly separate.

23 Mount Holyoke College, Spring Pollination of selected plants The plants you selected last week to be parents of the second generation should be in heavy flower. Today, you will assist these plants in the process of sexual reproduction. Brassica rapa plants need assistance because in nature they are totally dependent on certain insects for transferring sperm-bearing pollen from the male part of the flowers of one plant to the female part of the flowers of another plant. In these plants, the most conspicuous floral structures are the yellow petals. These petals enclose the male sexual structures (stamens, terminating in the pollen-producing anthers) and female structures (pistil, with the pollen-receiving stigma, style and egg-producing ovary), as shown in Figure 6 and Figure 7. Although each flower has both male and female parts, sperm from one plant are incapable of successfully fertilizing eggs of the same plant. This self-incompatibility ensures out-crossing (mating between different individuals). The insect pollinators don t act as pollen couriers just to be nice; the plants lure and reward them with nutrients (nectar and edible pollen), and the insects inadvertently pick up the sticky pollen on various body parts and then carry it with them to the next flower (Figure 12). Honeybees are common pollinators of Brassicas in the wild. We will use artificial bees (little pollinator wands) to help us pollinate these plants. Figure 12. Honeybee (the usual pollinator of Brassica) Designate sides of the room for hairy and bald plants. Do not mix plants or pollinating wands! No student should pollinate more than one type! Each student group should obtain one or two of the selected mature plants. The crosspollination activity will be accomplished over the course of the week, so that even early- and late-bloomers can mate with each other. The object is to transfer pollen from each plant to every other plant. This can be done in the following way. Using a single pollinating wand, one group will lightly rub and twirl the bristly end for several seconds onto the anthers and stigma of each open flower on their plant(s). Then pass the wand, now loaded with yellow pollen, to the next group, who will repeat the process. After the bee stick has made one complete round (all open flowers on all plants have been visited by the wand), make a second round in the same manner. When finished, insert the wand into the soil (pointed end down, bee end up) of one of the pollinated plants don t return the used wand to its original container, where other students might confuse it with fresh, unused ones. Finally, return all plants to the watering tray. Fertilization will result in the development of the ovules (each containing an embryo) into mature seeds, which will be contained in the fruit or seed pod (the elongated ovary). The

24 20 Artificial Selection in Brassica rapa length of time from fertilization to mature seed is about 3-4 weeks. We don t need to do anything more with these plants until it is time to harvest the seeds except to remove them from water in 3 weeks, which hastens the seed-maturation process. Next week, you should easily be able to see the elongating fruits. In two weeks, they will have become even longer (2-5 cm) and the individual seeds inside (perhaps 5-20 per pod) will have become visible. In three weeks, the maturing process is nearly complete. You will plant these seeds (generation 2) in about 4 weeks. Harvesting seeds The Fast Plants that you selected from the initial large population and pollinated should have mature seeds ready to be harvested. Each student group should take one or two of the plants to your table. Remove several pods, then break open the pods into a small plastic dish, releasing the seeds. You should be able to retrieve 5-20 or so seeds from each pod. The plant from which you removed the pod is the mother of each of those seeds; the father could have been one or more of any of the other plants in the selected group. Similarly, your plant could have been the father of seeds from any of the other plants in the selected group. Our mass cross-pollination technique ensured that all individuals in the selected group had an equal opportunity to be parents. Establishing the second generation The goal is to plant and grow generation 2 seeds in order to assess the average hairiness of this new generation (the offspring of the selected generation 1 parents). Each student will be responsible for planting a small number of seeds. (This time we will plant 2 seeds per pot, since germination rates tend to be lower in classroom-harvested seeds than seeds we get from the supplier.) Your lab instructor will give you further directions. After planting, your section s flat will be placed under the lights in the lab room, and maintained by the staff. By next week s lab, most of the seeds will have germinated. In two weeks they will be ready for the final analysis of the experiment. Counting the hairs in the second generation Follow the counting procedure from the first generation, keeping separate records for the offspring of the hairy parents and the offspring of the bald parents.

25 Mount Holyoke College, Spring Data Analysis Inheritance Patterns Is hairiness genetic? Allele dominance and frequency An allele is dominant if it takes only one to express a trait. Dominant alleles are generally those that do something: specify a working version of an enzyme, an antigen, a hormone, a transcription factor, a receptor, etc. Recessive alleles, on the other hand, are generally those that represent a loss-of-function. That is, they do not specify a working version of the protein in question. It seems more likely that alleles that result in production of trichomes are dominant. Allele frequency depends not on dominance, but on current prevalence and usefulness. Alleles that reduce the reproductive capability of their bearers tend to become less common, while those that enhance reproductive capability tend to become more common. In a large population with random mating and no slection, allele frequencies do not change. Allele frequencies change when there is selection for one allele over the other. When you read The cost of defense against herbivores, by Agren and Schemske, you should be thinking about how herbivores act as agents of selection on their food plants. If herbivorous insects prefer to eat smooth stemmed plants, then the smooth stemmed plants will, on average, produce fewer offspring. Therefore, if hairiness is a heritable trait, the frequency of alleles for hairiness should increase in the population, and the frequency of alleles for smooth stems should decline *. If there are no predatory insects to maim or kill the smooth-stemmed plants, then the smoothstemmed plants may actually have an advantage they can spend their energy on making babies instead of making trichomes. In this case, the allele for smooth stems should increase in frequency in the population. For neither scenario (selection against smooth plants by herbivorous insects or selection for smooth plants in the absence of herbivores) does it matter which allele is dominant. Whichever allele is most useful under the current conditions will become more common in the population. We have acted as very severe agents of selection in our population of plants. We have chosen two subsets of the original heterogeneous population to be parents, and have killed all the rest! We reduced the reproductive capability of the non-chosen plants all the way to zero! In one subset, only smooth-stemmed plants were allowed to breed. In the other subset, only those in the hairiest 10% or so of the population were allowed to breed. * Of course, the plants also act as agents of selection on the insects. If all the plants are hairy, the individual insects that cannot abide trichomes do not breed as much (because they don t get enough to eat) as those that find a way to eat hairy plants. Allele frequencies will change in the insects, producing a population more willing to eat hairy plants. This will likely result in further selection by the insects against the relatively less hairy plants, and an arms race of sorts will ensue.

26 22 Artificial Selection in Brassica rapa If hairiness is an inherited trait, then we should be able to create two populations by this technique, one in which the hairy allele(s) is(are) common, and one in which the smooth allele(s) is(are) common. No matter which allele is dominant, we can change its frequency in the population. This should further reinforce the idea that either the dominant or the recessive allele can be more common in the population, depending on the circumstances. Realized heritability Selection can act on any phenotypic variation, but can only cause evolutionary change if the variation is genetic. Population biologists often use an index called realized heritability, h 2, to quantify the degree to which a trait in a population can be pushed by selection. To calculate this index, we find the response to selection by subtracting the average of the second generation from that of the entire first generation. We also find the selection differential by subtracting the average of the selected parents from that of the entire first generation. Realized heritability is the response divided by the differential. realized heritability = response to selection selectiondifferential st nd avg1 gen avg2 gen realized heritability = st avg1 gen avg selected parents A low h 2 (for example, less than 0.01) occurs when the offspring of the selected parents differ little from the original population, in spite of a big difference between the population as a whole and the selected parents. A high h 2 (for example, greater than 0.6) occurs when the offspring of the selected parents differ from the original population almost as much as the selected parents do. Is hairiness controlled by a single gene? Complete dominance One of the first questions we might ask of our results is whether there is a single on/off type gene for hairiness. To answer such a question, we need to predict what the results of our experiment might be if that were the case. So here is a hypothetical inheritance pattern for hairiness in Fast Plants: Suppose there is a single gene which determines whether or not trichomes are produced. Suppose that there are exactly two alleles of this gene, H, which directs the production of hairs, and h, which does not. Since every plant gets one of these alleles from its mother and one from its father, each has a pair of alleles, and there are three possible pairs of alleles, or genotypes: HH, Hh, and hh. Since, in our hypothetical situation, H allows hairs to be made, any plant with at least one H will have trichomes, that is, plants with either the genotype HH or Hh will produce hairs and plants with the genotype hh will not. The appearance of the plant with respect to the genes in question is called its phenotype.

27 Mount Holyoke College, Spring According to this scheme, we cannot tell the genotype of a hairy plant by looking at it, but all the hairless plants have the genotype hh. So when two hairless plants are mated, they each put one h into their respective gametes, and their offspring would all be hh, as well. On the other hand, some of the hairy plants might have the genotype HH and some, Hh, so there are three possible matings: HH x HH, HH x Hh, Hh x Hh. It is possible in this third mating for an offspring to be produced with the genotype hh, making it bald. It is left as an exercise for the reader to figure out the probabilities of the various genotypes in the offspring of these matings. We can examine our data to find out whether all hairless parents had hairless offspring, and some hairy parents had hairless offspring and thus answer the question of the presence of a single on/off gene. We can also examine our data for the converse: is there a single on/off gene whose dominant allele prevents trichome production? Again, the reader is left to construct the relevant diagrams and use the data to answer this question. Incomplete dominance Suppose the hairiness gene we ve been discussing isn t a simple on/off switch, but a modulator of some sort. Suppose there is a T gene (for trichome), which influences the number of trichomes, rather than the mere presence or absence of them. Suppose that there are still two alleles, T, and t, but that T is incompletely dominant to t. This means that the heterozygote, Tt, would have a phenotype (a number of hairs) intermediate between that of the two homozygotes, TT and tt. Plants with the genotype TT would have many hairs, those with the genotype Tt would have some hairs, and those with the genotype tt would have very few hairs. Each T can be thought of as contributing one phenotypic unit, which in this case would be some number of hairs. This is illustrated in Table 2. Table 2. Genotypic and phenotypic categories for a singlegene, incomplete dominance model of trichome number inheritance. # of phenotypic units Intermediately phenotype Hairiest hairy smoothest Genotype TT Tt tt Matings between similar homozygous parents (TT x TT, or tt x tt) should produce only offspring just like the parents. Matings in which at least one of the parents is heterozygous (Tt) would produce offspring with greater variation. Calculating the results of such matings is left as an exercise for the reader.

28 24 Artificial Selection in Brassica rapa Cross Results: T t T t Genotype: Phenotype: Proportion (fraction) of offspring of a Tt x Tt cross: We can examine our data for evidence that a single gene with incomplete dominance controls trichome number in our plants. Are there offspring of two very hairy parents that have as few hairs as the least hairy parents? Are there offspring of two bald parents that have as many hairs as one of the hairy parents? Is hairiness controlled by multiple genes? Hairiness in Fast Plants most likely differs from these simple traits in that there are not just two or three phenotypes, but a whole range (from zero to many ). Explaining the inheritance of such traits involves a more complicated model. Traits such as body size or skin pigmentation in humans, or milk production in cattle, for which there exists a continuum of phenotypes between two extreme values, are known to geneticists as quantitative (or continuous) traits. Traits such as petiole trichome number in Fast Plants, number of bristles on fruit fly abdomens, or number of eggs laid by house fly females, are in a related category called meristic traits. These often exhibit wide variability among individuals, but the variability is incremental, not continuous (e.g., an individual might have 5 or 6 hairs, but not 6.3 hairs). Despite the difference between these two kinds of traits, geneticists feel their mechanism of inheritance is similar. Such traits are very important most traits in most organisms are quantitative or meristic, and relatively few are simple two-phenotype traits such as exhibited by Mendel s peas (red/white flowers) or the Macintosh rabbits (straight/floppy ears). Quantitative and meristic traits are also often called polygenic traits, because the model that best explains their inheritance involves multiple gene loci, with two or more alleles at each locus. Two loci. Let s imagine a situation in which there two genes contribute to the hairiness of the petiole in Fast Plants. As in the previous example, each locus has two alleles ( A and a for one locus; B and b for the other), and each uppercase allele contributes one phenotypic unit. Thus, an individual with genotype AABB would exhibit a phenotype of 4 units; individuals with genotype AABb or AaBB would each have a phenotype of 3 units; etc., as shown in Table 3.

29 Mount Holyoke College, Spring Table 3. Genotypic and phenotypic categories for a 2-gene model of trichome number inheritance. Number of phenotypic units Very many Many Some A few Almost no phenotypes hairs hairs hairs hairs hairs AAbb AABb Aabb Possible genotypes AABB AaBb aabb AaBB aabb aabb As before, when both parents have only the dominant alleles (AABB x AABB) or only the recessive alleles (aabb x aabb), the offspring will have the same genotypes as their parents. Suppose, however, that our hairiest parents were in category 3, that is, with genotypes AABb and AaBB. There are three possible matings in this population: AABb x AABb, AABb x AaBB, and AaBB x AaBB. What are the possible genotypes and phenotypes in the offspring of these matings? What are the highest and lowest phenotypic categories that might result from matings between parents with category 3 phenotype? We will have to consider whether our data are consistent with a 2-gene system such as this. Three loci. If there were 3 genes, A, B, and C, there would be 7 phenotypic categories. The reader should fill in Table 4 like that for two loci, showing how many different genotypes would there would be for each category. Table 4. Genotypic and phenotypic categories for a 3-gene model of trichome number inheritance. # of phenotypic units phenotype Hairiest smoothest Possible genotype(s) The more genes involved in adding to the number of trichomes, the more phenotypic categories there are, and the more closely the theoretical distribution can match the real distribution, which after all, had 24 different numbers of trichomes on their petioles. Another factor is how many trichomes on average are added by each dominant allele. Biological things being the way they are, we should expect a rather loose correlation between the number of dominant alleles and the number of trichomes, as opposed to a precise one-to-one correspondence.

30 26 Artificial Selection in Brassica rapa What happens in the second generation depends on the nature of the inheritance of trichome number in fast plants. What if, for example, the hairiest individuals in our first generation did not happen to have all possible trichome-generating alleles? (I.e., What if our hairiest individuals were in fact heterozygous for one or more trichome genes?) How would that affect the possible range of trichome number in their offspring? What if there is more than one genotype that results in a smooth petiole, so that some of the smooth individuals were heterozygous? What might we see in their offspring? Analyzing the data to describe the variation in the population In addition to the variation in counting, of course, there is variation between plants. Our goal in this semester long experiment will be to see how much of this variation can be accounted for by genetics. We will compare the hairiness of the original population with the next generation, after selection. We need measures not only of hairiness, but also of variation in hairiness. The mean (average) Calculating the average number of hairs per plant is one way to indicate hairiness. Suppose the 6 plants planted by one student had the following number of hairs (arranged in descending order): 15, 13, 10, 8, 5, 0. The average number of hairs is 8.5, calculated as follows: average = = 51 6 = 8.5 The Excel file has been set up to calculate this automatically as the data are entered, since adding up hundreds of numbers is tedious. You can use the function wizard (f x ) to calculate the average: pick a cell to contain the average, click on f x, choose AVERAGE, highlight the column containing the number of hairs, and click OK. Frequency distribution One way to look at the hairiness in the population is to construct a histogram, or bar graph, to illustrate the variation. To do this, we need to lump some of our data together in categories. A convenient starting point would be to count how many plants have between 0 and 4 hairs, how many between 5 and 9, how many between 10 and 14, etc. This effectively creates hairiness categories. The student at the computer highlights just those cells whose entries range from 0 through 4, and counts them. If she keeps her finger down on the mouse, Excel will tell her, in the upper left corner of the spread sheet, how many rows she has highlighted. If she lets her finger up, she has to count them by hand. A student at the board constructs a table showing the frequency distribution. Table 5 is such a table for our 6 plant data given in the section on averages, above.

31 Mount Holyoke College, Spring Table 5. Frequency distribution of number of hairs on the first petiole of 6 week-old plants number of hairs number of plants Each student should construct a histogram illustrating the frequency distribution of hairs on all the plants. The histogram illustrating the data in Table 5 is given in Figure 13. Figure 14 illustrates the distribution of trichome number in the first generation of Fast Plants grown in spring 2001 for an experiment similar to ours. The height of each bar represents the number of plants whose number of hairs fell into the category given below the bar. Talk with your instructor about learning how to construct such a histogram with Excel. number of plants #of first generation plants unselected selected smooth selected hairy 0 0 to 4 5 to 9 10 to to 19 number of trichomes per petiole Figure 13. Distribution of number of hairs per petiole 0 0 to 4 5 to 9 10 to to19 number of trichomes per petiole 20 to 24 Figure 14. Frequency distribution of trichome number in first generation Fast Plants, Spring The standard deviation One common method of expressing the variation within a population is to calculate one of several conventional indices of uncertainty. An especially common such index is the standard deviation. It is a statistical measure of the precision in a series of repetitive measurements. It is calculated this way: Where sd sd is standard deviation, n is the number of data, x i is each individual measurement, x is the mean of all measurements, and Σ means the sum of = ( x n x 1 i ) 2

32 28 Artificial Selection in Brassica rapa The standard deviation is something like the average deviation from the mean. We subtract each measurement from the mean of the measurements. Since some of these are positive and some are negative, and will cancel each other out when we add them together, we square each deviation to make it positive. Then we add all these squares together, and divide by the number of measurements (actually, we fudge it here, using n-1 instead of n, because this gives a better result sorry) to get a sort of average. Finally, we correct for having squared the deviations by taking the square root of the whole thing. The standard deviation has become popular because of an interesting feature. For a large sample, two thirds of the values can be found within one standard deviation of the mean, 95% within two standard deviations, and 99% within three. In the example above, the standard deviation would be 5.5, calculated as follows: sd = (15 8.5) 2 + (13 8.5) 2 + (10 8.5) ( ) 2 + (5 8.5) 2 + (0 8.5) 2 = ( 0.5 ) 5 + ( ) + ( 8.5) 2 = = = 29.9 = 5.47 A greater spread among my values would give a higher standard deviation; a smaller spread, a lower one. Table 6 illustrates this: the data set from the example above is listed between two other data sets, each with the same average and different standard deviations. Table 6. Three data sets with the same average and different standard deviations. raw data average standard deviation 12, 11, 9, 8, 7, , 13, 10, 8, 5, , 17, 11, 3, 1, It is a good thing we have computers to do the calculations for us, but for the answer to have meaning, we need to know how it is done. The standard deviation is often used in histograms, to illustrate the spread of values in the data, as well as to help compare one group of data with another. Figure 15 llustrates the use of means (the bars) and standard deviations (the lines) to compare the number of trichomes in the various groups of Fast Plants from a previous semester s experiment. The heights of the bars tell us that the two sets of offspring were different from the first generation of plants, and the standard deviation marks tell us that there was more variation among the offspring than among the selected parents.