ARTHUR C. GIBSON, MEMBG Director
Many of our readers are probably unaware of a recent revolution - or revelation - in the plant sciences. For more than 200 years, all botanists have utilized classifications in which the flowering plants, i.e., angiosperms (260,000 species), were defined as consisting of two fundamental branches, the dicotyledons and the monocotyledons. This has been a universal lesson around the globe. Both amateurs and professionals use shorthand to abbreviate these as dicots and monocots. Counting species, there are at least three times more dicots than monocots (60,000 species). With the assistance of evidence from DNA sequences, now scientists with confidence place monocots as a branch within the dicotyledons, destroying one old model that depicted two branches at the base. Don't throw your old botany books away yet, but certainly expect there to be major revisions in the upcoming generation of textbooks.
Every basic textbook discusses how to distinguish dicots from monocots using easily observed characters. Monocots have leaves with parallel principal veins, whereas prominent netted or reticulate venation typifies dicots. Stems on monocots have vascular bundles evenly distributed in a cross section, whereas dicots tend to have the vascular bundles arranged in a ring or cylinder. Monocots bear flowers with parts produced in multiples (whorls) of threes, whereas dicots generally have floral parts in multiples of fours or fives. As the name implies, the embryo of a monocot has only one cotyledon, whereas a dicotyledonous embryo almost universally has two cotyledons, rarely more. Whether the plant is sterile, in flower, or in fruit, nobody needs to be an expert to determine with reasonable certainty to which class an angiosperm belongs. If you were handed a bag of unknown flowering plants and asked sort them into two piles, using just the characters provided you probably would get an A. Of course, beware of the teacher who tries to trick you with the exceptions!
In the late 17th century, it was John Ray (1627-1705), dubbed the father of British botany, who published the first classification in which he recognized monocotyledons and dicotyledons as natural groups. The next major leap in knowledge came in 1789 when Antoine Laurent de Jussieu published a natural system of classification in which he recognized "monocotyledons" as one of three major groups of plants, and subdivided monocots into major lineages that we now recognize as families, e.g., grasses, palms, amaryllids, orchids, rushes, sedges, bromeliads, aroids, and irises. Whereas a score of the groups named by de Jussieu are still recognized as families, now the number of monocot families exceeds 100, and that may increase as little known forms are analyzed with greater sophistication.
The revolution was initiated during the early 1990s, when numerous labs were doing nucleotide sequencing of the rbcL gene from plant chloroplasts and pooled results for one grand computer analysis. Techniques for isolating and sequencing nucleic acids were becoming rapid and relatively inexpensive. For this segment of DNA, which was 1428 base pairs in length, the computer could be used to align the nucleotide sequences from species of more than 100 plant families, and then to calculate the mutation steps required to best explain the time sequence of changes in that gene's structure. The resultant DNA tree for angiosperms should reflect the overall pattern of branching of the families. The goal was to produce a phylogeny of plants, called a topology or cladogram, that shows when each branch of plants diverged.
One gene sequence cannot give perfect resolution of all branching events during evolution. Indeed, a scientific reconstruction of the branching events would be really simple if each time a lineage began to diverge there would be a mutation in the gene. That permanent alteration in the DNA could then be used to tag the starting point of each family. Unfortunately, sometimes divergence has occurred without changes in the gene; sometimes too many mutations occurred to clearly resolve the sequence of divergence; never is the information stored in one gene sequence sufficient to resolve patterns of divergence to anyone's level of total satisfaction. Hence, molecular plant biologists have worked toward an analysis utilizing sequences of several different DNA genes.
The most useful multigene analysis so far has been one published three years ago by 16 authors based on nucleotide sequences of three genes, rbcL and atpB from chloroplasts and 18s for ribosomal DNA sequences from the nucleus (Soltis et al. 2000. Botanical Journal of the Linnean Society 133:381-461). By analyzing the patterns of evolution for all three genes at the same time, the computer could produce an evolutionary tree having a very high degree of support. From this analysis, one conclusion was inescapable: the monocots evolved from dicotyledonous stock, not from the base of the tree, and monocots originated after several groups of dicotyledons had already evolved.
The "consensus" topology estimating the known early evolutionary branches of angiosperms is presented in the above diagram, which is a simplification from the published version in Soltis et al. (2000). What this shows, reading left to right, is that the oldest recognized branch is represented by Amborella, today found only on New Caledonia. The next distinctive branch includes the fully aquatic, widespread water lilies, the Nymphaeaceae (not including Nelumbo, the sacred lotus). The third divergence includes woody plants in the order Illiciales and Austrobaileya, and then next appears to be a fourth lineage, named the eumagnoliids, which includes the subgroup order Chloranthales and five other branches, one of which became monocots. These may not be the only early branches, but they are the ones discovered so far and fully justified with molecular data from the species examined. The origin of monocots is nested within the dicotyledons. In professional jargon, the scientist would say that there is "well-supported congruence of the tree." All gene sequence data clearly show that monocots evolved within dicots and not the reverse. So appears to end that age-old debate.
Whether using nucleotide sequences of DNA molecules or robust analyses of more traditional structural and bio-chemical features, sophisticated computer analyses tell us that we can be essentially 100% certain that ALL monocots arose as a single evolutionary lineage, i.e., monocots are "monophyletic." Theoretically this means that all 2800 monocot genera on earth today can trace their origins back to a single ancestral population. Thus probably ends another age-old debate, whether monocots evolved more than one time. The well-defined group of herbs (they lacked the ability to form wood) were evolving in the Early Cretaceous, perhaps 120 million years ago or more, although the oldest dependable monocot fossil has a date of only 100 million years.
A closer view of the origin of monocots has revealed that the oldest branch found so far is sweet flag, Acorus, which was formerly classified as an aroid (Araceae). Acorus is a rhizomatous perennial growing knee high around pond margins, common in the Midwest where I grew up. Its leaves are folded along a midvein and have no petiole. Nobody ever suspected that at this moment in time sweet flag would represent the deepest branch of monocots, i.e., with the nucleotide sequence closest to that of the ancestral monocot. Of course that may change as other Araceae and primitive-looking monocots are sequenced and analyzed. Acorus calamus L. has small bisexual flowers with six greenish yellow "petals", six stamens, and a superior ovary, whereas typical aroids have separate male and female flowers arranged in a spike (the spadix) subtended or wrapped by a leaflike bract (the spathe)..
During the 20th century, many evolutionary plant biologists speculated on what are the closest living relatives to monocots. Focus of attention was on dicots that posses any monocotyledonous features. These are, for example, certain dicot families that form flower parts in multiples of threes, many of which are in the eumagnoliids. Some dicots have scattered vascular stem bundles. The standard type of pollen grains in monocots is monosulcate, a feature that rarely appears among dicots, and the characteristic type of protein crystals in monocot phloem tissue most resembles crystalline inclusions in the Aristolochiaceae. It appears now that scientists who suggested Piperales and their similarities to Araceae were in the right neighborhood, if Acoraceae and Araceae are very near the base of monocot evolution.
Certainly this is not the final word, but DNA is becoming invaluable in resolving issues that are centuries old.