Editor's Note: Microbiologist Carl R. Woese, a recipient of the Crafoord Prize, Leeuwenhoek Medal, and a National Medal of Science, died December 30, 2012, at the age of 84. This story was originally published in the June 1981 issue of Scientific American.

Early natural philosophers held that life on the earth is fundamentally dichotomous: all living things are either animals or plants. When microorganisms were discovered, they were di­vided in the same way. The large and motile ones were considered to be ani­mals and the ones that appeared not to move, including the bacteria, were con­sidered to be plants. As understanding of the microscopic world advanced it became apparent that a simple twofold classification would not suffice, and so additional categories were introduced: fungi, protozoa and bacteria. Ultimate­ly, however, a new simplification took hold. It seemed that life might be dichot­omous after all, but at a deeper level, namely in the structure of the living cell. All cells appeared to belong to one or the other of two groups: the eukaryotes, which are cells with a well-formed nucleus, and the prokaryotes, which do not have such a nucleus. Multicellular plants and animals are eukaryotic and so are many unicellular organisms. The only prokaryotes are the bacteria (in­cluding the cyanobacteria, which were formerly called blue-green algae).

In the past few years my colleagues and I have been led to propose a funda­mental revision of this picture. Among the bacteria we have found a group of organisms that do not seem to belong to either of the basic categories. The or­ganisms we have been studying are pro­karyotic in the sense that they do not have a nucleus, and indeed outwardly they look much like ordinary bacteria. In their biochemistry, however, and in the structure of certain large molecules, they are as different from other prokary­otes as they are from eukaryotes. Phylo­genetically they are neither prokaryotes nor eukaryotes. They make up a new "primary kingdom," with a completely different status in the history and the natural order of life.

We have named these organisms ar­chaebacteria. The name reflects an untested conjecture about their evolution­ary status. The phylogenetic evidence suggests that the archaebacteria are at least as old as the other major groups. Moreover, some of the archaebacteria have a form of metabolism that seems particularly well suited to the conditions believed to have prevailed in the early history of life on the earth. Hence it seems possible that the newest group of organisms is actually the oldest.

The evolutionary record

The earth is four and a half billion years old, and on the basis of the macro­scopic fossil record it would appear to have been inhabited for less than a sev­enth of that time: the entire evolutionary progression from the most ancient ma­rine forms to man spans only 600 mil­lion years. The fossil imprints of unicellular organisms too small to be seen with the unaided eye tell a different sto­ry. Microfossils of bacteria in particular are plentiful in sediments of all ages; they have been found in the oldest intact sedimentary rocks known, 3.5-billion­ year-old deposits in Australia. Over an enormous expanse of time, during which no higher forms existed, the bac­teria arose and radiated to form a wide variety of types inhabiting a great many ecological niches. This age of microorganisms is the most important period in evolutionary history not only because of its duration but also because of the na­ture of the evolutionary events that took place over those billions of years.

Until recently, however, almost noth­ing was known about the age of mi­croorganisms. Bacterial microfossils are not very informative structures; little can be inferred from the imprint of a small sphere or rod. The main paleon­tological indications of the nature of the early bacteria have come not from the individual microfossils but from the macroscopic structures called stromat­olites, which are thought to be fossil­ized bacterial mats: colonies of bacteria embedded with minerals. Today such structures are formed primarily by sev­eral kinds of photosynthetic bacteria, usually the cyanobacteria. Stromato­lites fossilized recently resemble the ancient ones to such an extent that it seems entirely reasonable to think the ancient structures were also made by photosyn­thetic bacteria. Therefore at least some of the ancient bacteria must have been photosynthesizers. Apart from that one fact virtually nothing could be estab­lished with certainty about the earliest microorganisms. The entire evolutionary tree of the bacteria remained ob­scure, as did the base of the tree for the higher forms of life.

In reconstructing early evolutionary events, however, biologists are not limit­ed to the geologic fossil record. The cell itself retains evidence of its past in the amino acid sequences of its proteins and in the nucleotide sequences of its nucleic acids: DNA and RNA. This living rec­ord is potentially far richer and more extensive than the fossil record, and it reaches back in time beyond the oldest fossils, to the period when the common ancestor of all life existed.

In order to read the biochemical rec­ord it was necessary to develop a tech­nology for determining (at least in part) the sequence of a gene or of the RNA or protein product encoded by a gene. For proteins this has been possible for about 25 years, but the direct sequencing of DNA and RNA has been feasible for only the past five years or so. The new technology for sequencing nucleic acids should enable biologists to uncover in relatively short order far more about the history of life than had been thought possible. It was by applying techniques of sequencing to the century-old prob­lem of the natural relations among bac­teria that my colleagues and I recog­nized the archaebacteria as a third form of life.

Eukaryotes and prokaryotes

In order to appreciate the special status of the archaebacteria it is helpful to consider some of the defining charac­teristics of eukaryotes and prokaryotes. The eukaryotic cell is comparatively large: roughly 10 micrometers on a side. It is surrounded by a double membrane, within which a number of structures can be discerned that are themselves defined by membranes. The nucleus contains the bulk of the cell's genetic material. The rod-shaped mitochondria are the site of cellular respiration, which gen­erates the cell's main energy curren­cy, adenosine triphosphate (ATP). In plant cells the chloroplast, another rod­ shaped body, converts the energy of light into the chemical energy of ATP.

Other specialized structures such as the Golgi apparatus (for secretion) and cilia (for motility) are often present. Many eukaryotic cells are laced with a mem­brane system, the endoplasmic reticu­lum, that provides a surface on which important reactions such as the synthe­sis of proteins take place.

The prokaryotic cell is vastly differ­ent. It is typically far smaller than the eukaryotic cell: by a factor of 10 in linear measure and hence by a factor of 1,000 in volume. The prokaryotic cell too is circumscribed by a double membrane, and in addition it almost al­ways has a rigid cell wall. On the other hand, none of the internal structures characteristic of the eukaryotic cell are present; there are no mitochondria or chloroplasts and of course there is no membrane-bounded nucleus. The ge­nome—the total complement of genetic material—is limited to between 2,000 and 3,000 genes in a prokaryotic cell; the typical eukaryotic genome is larger by several orders of magnitude.

The distinction between eukaryotes and prokaryotes was initially defined in terms of subcellular structures visible with a microscope. At that level all cells appeared to be either large and com­plex, and so eukaryotic, or small and simple, and so prokaryotic. The distinc­tion between the two cell types was ulti­mately carried to the most basic biologi­cal level, the level of molecules. Here eukaryotic and prokaryotic cells have many features in common. For instance, they translate genetic information into proteins according to the same genetic code. Even where the molecular proc­esses are the same, however, the details in the two forms are different; they are either characteristically eukaryotic or characteristically prokaryotic. For ex­ ample, the amino acid sequences of various enzymes tend to be typically pro­karyotic or eukaryotic. All these differences between groups and similarities within each group made it seem certain to most biologists that the tree of life had two main stems, one stem prokary­otic and the other eukaryotic.

That conclusion was drawn too hasti­ly; the aesthetic appeal of a dichotomy was too great. Simply because there are two types of cell at the microscopic level it does not follow that there must be only two types at the molecular level.

The evolutionary relation of prokary­otes and eukaryotes is actually more complicated than the evidence cited above would indicate. Two eukaryotic organelles, the mitochondrion and the chloroplast, each have their own DNA. Moreover, the pigments in the chloro­plast (the chlorophylls and the carote­noids) are similar to those found in the cyanobacteria. Both mitochondria and chloroplasts are the size of bacteria; their apparatus for translating genetic information into proteins differs from the eukaryotic cell's own apparatus and has a number of properties in common with that of prokaryotes.

These facts and others have led to the idea that mitochondria and chloroplasts are descended from prokaryotes that be­ came trapped in a larger cell and estab­lished an endosymbiotic relation with it. The mitochondrion is thought to have been a respiring bacterium and the chlo­roplast to have been a photosynthesiz­ing relative of the cyanobacteria. This conjecture, which in its simplest form is more than a century old, was essentially proved in the case of the chloroplast by the demonstration that the nucleotide sequence of one of the kinds of RNA in the organelle, ribosomal RNA, is spe­cifically related to ribosomal-RNA se­quences in cyanobacteria. Similarly, the ribosomal RNA of the mitochondrion in plants appears to be of the bacterial type. Thus it seems that at least two lines of prokaryotic descent are represented in the eukaryotic cell.

The urkaryote

Logically the next question is: Where does the rest of the eukaryotic cell come from? What was the original host cell: the urkaryote? It is generally agreed that the bulk of the eukaryotic cell (the nucleus and the cytoplasmic structures) represents a separate line of descent. The exact nature of the ancestral cell is not clear. Most investigators have tend­ed to believe the main eukaryotic line also arose from among the ordinary bacteria. The idea is that some anaerobic bacterium deriving its energy from the fermentation of nutrients (rather than from their oxidation) at some point happened to lose its tough cell wall. Or­ganisms of this kind are known; they are the mycoplasmas. A strain of myco­plasma then evolved the capacity to en­gulf other organisms, an ability retained by many eukaryotes today. Among the many kinds of organisms such a mycoplasma might have ingested two appear to have established a stable endosym­biotic relation with their host and to have become the mitochondrion and the chloroplast. In this way the eukaryotic cell was born. (The origin of its defining characteristic, the membrane-bounded nucleus, is still far from clear.)

This view is satisfying in some re­spects, but it fails to explain the many differences between eukaryotes and pro­karyotes. In particular it does not ac­ count for the different details of molecu­lar processes or for the large differences in the amino acid sequences of function­ ally analogous proteins in the two kinds of cells. It is often taken for granted that the differences are merely a conse­quence of the many small changes in cellular design that would be necessary in passing from the simple prokaryotic condition to the more complex eukary­otic one. It is questionable that so many changes (changes in the composition of almost all enzymes, for example) can reasonably be accounted for in this way.

Essentially for this reason some biolo­gists think the line of descent that gave rise to the putative urkaryotic species may have diverged from the prokaryot­ic line at some earlier point, before the ancestor of the bacteria had itself arisen. The urkaryote could then have evolved independently to a form comparable in complexity to that of the bacteria. Such an assumption would at least provide more time for differences to emerge between prokaryotes and the urkaryote. The urkaryote, then, would represent a line of descent distinct from that of the prokaryotes, in accordance with the ba­sic phylogenetic dichotomy.

So it stood at the beginning of the 1970s. The phylogeny of the higher eukaryotes, spanning some 500 million years, was reasonably well understood except for the all-important joining of the main eukaryotic branches. There was a definite, widely accepted hypothe­sis concerning the way in which the eu­karyotic cell had evolved. Tests of the hypothesis and answers to the remain­ing questions, however, lay in the unexplored recesses of bacterial phylogeny, in the age of microorganisms.

Genetic sequencing

Bacteria constitute a world of extraor­dinary variety, far more than the micro­scope reveals. The ecological niches which they are found far exceed in vari­ety those occupied by the higher forms of life. For a century microbiologists have tried in vain to understand the nat­ural relations among bacteria and to im­pose some order on the bewildering array of forms, physiologies and ecolo­gies. Variety among bacteria is mostly variety within simplicity, and so it pro­vides little information about phyloge­netic relations. In higher organisms the eye, for example, has evolved a number of times, but the eye is complicated enough for the independently evolved examples to be readily distinguishable from one another. Such is generally not the case for the form and structure of bacteria; rods, spheres and spirals, which are the typical bacterial shapes, are easily arrived at and have evolved many times. The same principle applies to bacterial biochemistry. Although some bacterial characteristics are valid phylogenetic indicators, it is impossible to tell in advance which ones are and which are not.

The simplest way in which the cell is a record of its past is in terms of genetic sequences. Every gene that exists in a cell today is a copy of a gene that existed eons ago. It is not an exact copy because mutations have altered the original ge­netic sequence, but vestiges of the origi­nal state often persist. What makes a gene a superb record of the past is its simplicity (it is a linear array) and the fact that genetic-sequence "space" is enormous, so that over the entire span of evolution only a small fraction of the possible genetic sequences can ever be realized. Hence if two genes are similar over a stretch comprising a significant number of nucleotides, this can only mean they have an ancestor in com­mon; such genetically related molecules are said to be homologous.

A genetic sequence yields three kinds of evolutionary information. The sequence can reveal genealogical rela­tions, it can measure evolutionary time and it is a record of ancestral character­istics. To the extent that two genes for the same function in different organisms are related, the organisms are related. The extent to which two such sequences differ measures the time since the organ­isms diverged from a common ancestor. From an extensive set of related se­quences one can construct a phylogenet­ic tree in which the branch points mea­sure (approximately) the relative times of the bifurcations. Finally, comparisons sons among an extensive set of homolo­gous sequences make it possible to re­construct with some accuracy various ancestral versions of a gene.

Since the relation between a gene and its product (either a protein or one of several kinds of RNA molecule) is gen­erally a colinear one, the sequence of the product is ordinarily as useful for phy­logenetic studies as the sequence of the gene itself. Because until recently only proteins could be sequenced it was through comparisons of proteins that the first phylogenies based on molecu­lar evolution were constructed. Comparisons of the respiratory protein cyto­chrome c prove to be particularly valu­able for confirming and extending the phylogenetic tree of the higher organ­isms. On the other hand, molecules such as cytochrome c are not as effective in establishing relations among bacteria. Such proteins are not universally dis­tributed; they are not strictly constant in function and so are not entirely compa­rable, and because of the greater antiq­uity of bacterial lineages differences in sequence can be far greater among bac­teria than they are among eukaryotes. These factors make bacterial phyloge­nies deduced from protein evolution in­ complete and uncertain.

Ribosomal RNA

There are other gene products that can serve as indicators of bacterial rela­tions. All self-replicating entities neces­sarily have systems for maintaining and propagating genetic information and for translating it into the chains of amino acids that constitute proteins. Most of the large molecules engaged in these processes must trace their origin back to the very early stages in the evolution of the cell; they certainly emerged be­fore cells became complex enough to be called prokaryotes. Therefore one would expect these molecules to have the requisite properties of a phylogenet­ic marker.

The most reasonable first choices among such molecules are the RNA molecules that are complexed with pro­teins to form the ribosomes. It is on the ribosomes that genetic information is translated into proteins. The ribosomal RNA is easy to isolate in workable quantities because a typical bacterial cell has from 10,000 to 20,000 ribo­somes. Moreover, ribosomal-RNA mol­ecules seem to have remained constant in function over great evolutionary dis­tances. This is important because func­tional changes in a molecule bring with them additional changes in sequence that make it difficult or even impossi­ble to compare one molecular sequence with another and thereby deduce phy­logenetic relations. Still another advantage of the ribosomal RNAs is that at least some portions of their sequences change slowly enough for the common ancestral sequence not to be totally obliterated. In other words, the sequen­ces make it possible to detect the deep­est phylogenetic relations.

There are three kinds of ribosomal­ RNA molecules. In bacteria the "large" ribosomal RNA is the 23S RNA (S stands for Svedberg unit, a measure of the rate of sedimentation in an ultracen­trifuge and hence an indirect measure of molecular size); it is approximate­ly 2,900 nucleotides long. The "small" one, designated 16S ribosomal RNA, is about 1,540 nucleotides long. A very small one (5S) has only 120 nucleotides. The sizes are similar in eukaryotic cells: 18S, 25-28S and 5S. One might think that ease of characterization would make the small 5S RNA the most suit­able one for phylogenetic studies. Actu­ally it is not as accurate an indicator of phylogenetic relations as the larger ribo­somal RNAs, chiefly for statistical rea­sons. (The 5S RNA sometimes exhibits anomalous large differences in sequence from one species to another.) The 16S ribosomal RNA is the molecule of choice, because the 23S molecule is almost twice as large and more than twice as difficult to characterize.

RNA dictionaries

At the University of Illinois in 1969 I decided to explore bacterial relations by comparing the sequences of the 16S ri­bosomal RNAs in different species. It was not yet feasible (as it is now) to determine the nucleotide sequence of the entire molecule. The technology did ex­ist, however, for sequencing short seg­ments of the molecule. The enzymes called ribonucleases yield short frag­ments of RNA by cutting an RNA strand at specific sites. Each nucleotide of RNA is composed of a sugar called ribose, a phosphate group and one of four nitrogenous bases: adenine (A), uracil (U), guanine (G) or cytosine (C). The enzyme T-sub-1 ribonuclease cuts an RNA strand at a particular bond on one side of each nucleotide that incorpo­rates a guanine base. The T-sub-1 enzyme therefore digests an RNA "text" into short "words," called oligonucleotides. Each oligonucleotide includes, and ends with, a single G as in AACUCG or UC­CUAUCG.

The oligonucleotides made in this way were short enough to be sequenced by the available techniques. The small­est words are of little value because they recur many times in each molecule. By the time the word length reaches six nucleotides, however, a particular se­quence is unlikely to appear more than once in a 16S RNA molecule. (Given the constant terminal G, there are 3^5, or 243, possible six-letter sequences of this kind, and a typical 16S RNA molecule has roughly 25 such words.) When 16S RNAs from different organisms include the same six-letter sequence, it almost always reflects a true homology. By con­fining attention to words six letters long or longer one can generate a "diction­ary" characteristic of a given organism, which can readily be compared with other such dictionaries to determine genealogical relations.

A simple way to analyze the data is in terms of an association coefficient S-sub-AB, which is defined as twice the number of nucleotides in the words common to both of the dictionaries A and B di­vided by the number of nucleotides in all words in the two dictionaries. S-sub-AB ranges from 1 when dictionaries A and B are identical to less than .1 when they are unrelated. (The coefficient is usually greater than zero even for unrelated se­quences because of chance correspon­dences.) By compiling the S-sub-AB values for a number of organisms in a matrix one can discern a pattern of relatedness or unrelatedness among organisms. More­ over, it is possible by straightforward statistical methods to construct from a set of SAB values for a group of or­ganisms a dendrogram, or tree, show­ing the relations among members of the group.

To date the ribosomal RNAs of al­most 200 species of bacteria and eu­karyotes have been characterized. Most of the bacteria form a coherent but very large (which is to say ancient) group. They are the eubacteria, or true bacte­ria, and as would be expected they are quite distinct from the eukaryotes. The relations among the various genera (rep­resented by the branchings of the tree) determined through ribosomal-RNA analysis are at variance with many of the established prejudices about bacteri­al relations. What is important at this point is that the eubacteria are divided into a number of major branches and hat several of the branches include pho­tosynthetic bacteria. This finding sug­gests all eubacteria stem from a com­ mon photosynthetic ancestor.

The discovery of archaebacteria

As the screening of bacteria contin­ued a surprise emerged. In collaboration with Ralph S. Wolfe I looked at the ribo­somal RNA of the methanogenic bacte­ria. These unusual organisms live only in oxygen-free environments and gener­ate methane (CH4) by the reduction of carbon dioxide (C02), We discovered that methanogens do not fall within the phylogenetic group defined by the other bacteria. Indeed, they appear to represent an evolutionary branching that far antedated the common ancestor of all true bacteria. Not only were the meth­anogens separate but also the group they formed seemed to be about as deep phy­logenetically—as ancient—as the group defined by the eubacteria.

There can be no doubt that the meth­anogens and their relatives are bacteria. They are the size of bacteria, they have no nuclear membrane, they have a low DNA content and so on. Surely, then, one would have expected them to be re­lated more closely to other bacteria than to eukaryotes. Our analysis showed they are not. Methanogens are related as closely to eukaryotes as to eubacteria.

How could this be? There were sup­posed to be only two primary lines of descent, the eukaryotic and the prokary­otic. Here was a new group of organ­ isms: the methanogens and their rela­tives, which together have come to be called archaebacteria. They were obvi­ously like other bacteria in their superfi­cial characteristics, and so they had been assumed to be in the prokaryotic line of descent. It is not striking differences in morphological characteristics, however, that distinguish the prokaryote phylo­genetically from the eukaryotic cell; it is the subtler and more ancient differences in molecular sequences and in details of function at the molecular level that dis­tinguish them. Hence there is no reason two prokaryotic lines of descent cannot be just as distinct from each other as either one is from the eukaryotic line.

This idea was too novel to be easily accepted, and initially some biologists rejected out of hand the notion of a "third form of life." How could some­ thing that looked like a bacterium not be a bacterium and indeed not be related to bacteria? In time the simplicity of our argument and the accumulation of evi­dence prevailed. Although a few biolo­gists still dispute our interpretation, the idea that archaebacteria represent a sep­arate grouping at the highest level is be­ coming generally accepted.

The supposed great antiquity of the archaebacteria remains an unproved prejudice, but it is a plausible one. The methanogenic phenotype seems to cov­er a phylogenetic span as great as or greater than the span covered by any other comparable bacterial phenotype. This implies that the methanogens are as old as or older than any other bacterial group. Moreover, methanogenic metab­olism (the reduction of carbon dioxide to methane) is ideally suited to the kind of atmosphere thought to have existed on the primitive earth: one that was rich in carbon dioxide and included some hydrogen but virtually no oxygen. The name archaebacteria implies that these organisms were the dominant ones in the primeval biosphere. When condi­tions changed, the methanogens' need for an anaerobic environment confined them to a limited range of relatively in­ accessible niches.

The measurements that revealed the existence of the archaebacteria (differ­ences in R N A sequences) were genetic ones and were purely quantitative. They revealed nothing about the quality of the differences—the phenotypic differ­ences—between the archaebacteria and the true bacteria. If our interpretation of the archaebacteria as a primary king­dom separate from that of the true bac­teria is correct, then on detailed inspec­tion the archaebacteria should prove to be as different from true bacteria in their molecular phenotype as either group is from eukaryotic cells.

Archaebacterial forms

The archaebacteria are indeed unusu­al organisms. The group is now known to include three very different kinds of bacteria: methanogens, extreme halo­philes and thermoacidophiles.

The dominant form (in the sense that it constitutes a deep phylogenetic group­ing) is the methanogen. Bacteria that give off methane have been known for some time. Alessandro Volta discovered in 1776 that "combustible air" is gener­ated in bogs, streams and lakes whose sediments are rich in decaying vegeta­tion, but the fact that a microorganism is responsible for generating "marsh gas" became known only much later. Methanogens are widely distributed in na­ture, but they are not commonly en­ countered because they are killed by oxygen and do not exist in the open.

In ancient times methanogens could have existed almost anywhere. Today they live only where oxygen has been excluded and where hydrogen and car­ bon dioxide are available. This general­ly means living in close association with other bacteria, such as the clostridia, that metabolize decaying organic mat­ter and give off hydrogen as a waste product. Methanogens are found in stagnant water and in sewage-treatment plants (in amounts that have made it commercially feasible to manufacture methane). They are also found in the rumen of cattle and other ruminants and in the intestinal tract of animals in gen­eral. Methanogens can be isolated from the ocean bottom and from hot springs. In spite of their intolerance of oxygen they are obviously globally distributed.

The extreme halophiles are bacteria that require high concentrations of salt in order to survive; some of them grow readily in saturated brine. They can give a red color to salt evaporation ponds and can discolor and spoil salted fish. The extreme halophiles grow in salty habitats along the ocean borders and in inland waters such as the Great Salt Lake and the Dead Sea. Although the extreme halophiles have been studied by microbiologists for a long time, they have recently become particularly inter­esting for two reasons. They maintain large gradients in the concentration of certain ions across their cell membrane and exploit the gradients to move a variety of substances into and out of the cell. In addition the extreme halophiles have a comparatively simple photosynthetic mechanism based not on chlorophyll but on a membrane-bound pigment, bacterial rhodopsin, that is remarkably like one of the visual pigments.

The thermoacidophiles

The third known type of archaebac­terium is the thermoacidophile, and the members of this group too are nota­ ble for their habitat. Sulfolobus, one of the two genera of thermoacidophiles, is found in hot sulfur springs. Its various species generally grow at temperatures near 80 degrees Celsius (176 degrees Fahrenheit); growth at temperatures above 90 degrees has been observed for some varieties. Moreover, the springs in which Sulfolobus flourishes are extreme­ly acidic; pH values lower than 2 are common (a pH of 7 is neutral). Thermoplasma, the other genus of thermoacid­ophile, has so far been found only in smoldering piles of coal tailings. It is a mycoplasma: it has no cell wall but merely the limiting cell membrane.

Although archaebacterial thermoacidophiles can grow only in an acidic environment, the internal milieu of the cell has a quite moderate pH, near neutrality; this requires that a sizable pH gradient be maintained across the cell membrane. As in the extreme halophiles the gradient may play a role in pumping other molecules into and out of the cell. It is interesting that when the temperature is reduced and as a consequence Sulfolobus stops metabolizing, the cell's internal pH can no longer be maintained near neutrality and the cell dies.

For some time it had been recognized that various organisms now classified as archaebacteria are individually somewhat peculiar. In each instance the idiosyncrasy has been seen as just that: an adaptation to some peculiar niche or biochemical quirk. The ribosomal-RNA phylogenetic measurement, however, showed at least some of the idiosyncrasies might instead be general characteristics of a new group of organisms. Thus informed, investigators in many countries have undertaken to find the general properties that link archaebacteria to one another and to see how those properties either distinguish the archaebacteria from the other two major forms or relate them specifically to one or the other of those forms.

One generalization about bacteria has been that they have a cell wall incor­porating the sugar derivative muramic acid, which is the basis of a complex polymer called a peptidoglycan. One extreme halophile and the thermoacid­ophile been that they have a cell wall incor­porating the sugar derivative muramic acid, which is the basis of a complex polymer called a peptidoglycan. One ex­treme halophile and the thermoacidophile Sulfolobus were known to be exceptions to this generalization; they were considered to have an idiosyncratic wall structure. Otto Kandler of the University of Munich, collaborating with Wolfe, made a systematic study of cell-wall structure in other known ar­chaebacteria. All of them turned out to be atypical. The archaebacteria have a variety of wall types, but none of them is of the muramic-acid-based peptidogly­ can type.

Lipids and RNAs

It was also known that the cell mem­brane of the extreme halophiles and of the thermoacidophiles is composed of unusual lipids. The lipids of both eu­karyotes and eubacteria consist in the main of two straight-chain fatty acids bound at one end to a glycerol molecule through an ester linkage (-CO-O-). The lipids of the extreme halophiles and the thermoacidophiles are also composed of a glycerol group linked to two long hydrocarbon chains, but the connection between the glycerol and the chains is an ether (-O-) link rather than an ester link. Moreover, the hydrocarbon chains are not straight but branched: every fourth carbon atom in the chain carries a meth­yl group (CHa). The basic archaebacterial lipid, in other words, is a diether composed of glycerol and two mole­cules of an alcohol, phytanol. When a number of methanogens were examined for lipid composition, our expectation was confirmed: their lipids turned out to be typically archaebacterial branched­ chain glycerol ethers.

In the course of the ribosomal-RNA studies another unexpected archaebac­terial property emerged, one that was to provide the first clue to the significance of the differences between archaebacte­ria and true bacteria. Central to the process of translation is the transfer­ RNA molecule. It recognizes a three­ base "codon" in messenger RNA speci­fying a particular amino acid, and it delivers that amino acid to be incorpo­rated into the protein chain. A number of the nucleotides in a transfer-RNA molecule are modified, that is, their structure is altered chemically after they have been incorporated into the mole­cule; most often a methyl group is add­ed to the nucleotide at some position on either the base or the sugar. Biologists had come to believe one particular mod­ification was characteristic of a certain position in almost all transfer-RNA molecules in almost all organisms: at that position the base uracil has been methylated to form thymine (which is normally present only in DNA, not in RNA). It turns out that all transfer RNA's of all archaebacteria lack this thymine unit; instead the uracil has been modified in one of two other ways to yield a pseudouridine or an as yet un­ identified nucleotide.

If one compares both ribosomal RNA and transfer RNA in eukaryotes, eubac­teria and archaebacteria, one finds a general pattern, of which the replace­ment of thymine in archaebacterial transfer RNA's is only one example. The same regions in the RNA's tend to be modified in all three primary lines of descent, but the nature of the modifi­cation tends to vary from one kingdom to another. The differences are of two kinds. Either the modification of a given base is different in each of the kingdoms, or a given base is modified in one kingdom and in another kingdom the modi­fication is made to an adjacent base. These modes of variation suggest that the modifications have evolved sepa­rately in each major line of descent.

Several other molecular distinctions between the archaebacteria and the oth­er groups are known (for example, in the subunit structure of the enzyme RNA polymerase), but the list is not long. The reason is not that additional differences do not exist; it is rather that the world of archaebacteria remains virtually un­explored. The study of archaebacterial genetics is in a primitive state; few mu­tants have even been isolated for genet­ic study. Nothing whatever is known about the control of gene expression in archaebacteria. The basic molecular biology of archaebacteria is not under­ stood. And yet to the extent that the ar­chaebacteria have been characterized they have been found to differ significantly from both of the other major groups.

A new perspective

The discovery of a new primary king­dom of organisms is a major finding in its own right (comparable to going into the backyard and seeing an organism that is neither a plant nor an animal), but the real importance of the discovery lies in what it may reveal about the early history of life. When there were only two known primary lines of descent, one could not readily interpret the differ­ences between the two. The recognition of three lines of descent equidistant from one another gives a much better perspective for judging which properties are ancestral and which have evolved recently. With the discovery of the ar­chaebacteria two central evolutionary problems therefore become approach­able: the nature of the common ances­tor of all life and the evolution of the eukaryotic cell.

At what stage in the evolution of the cell did the fundamental division into the primary kingdoms take place? What was the nature of the universal ancestor? The assumption has been that the uni­versal ancestor was a prokaryote, the simplest of today's living forms. Long ago, however, there must have been still simpler forms of the cell. Although nothing is known about such forms, one can make an educated guess as to cer­tain of their general properties.

Consider the following argument. The translation mechanism is complex, comprising on the order of 100 large molecular components. It is also high­ly accurate in its functioning. In mak­ing proteins of the size common today (chains of from 100 to 500 amino acids), obtaining a flawless product 90 percent of the time or more requires an error rate of no more than a few parts in 10,000. Moreover, this accuracy must be attained by a mechanism of molecu­lar dimensions. For such a mechanism to have evolved in a single step is clearly impossible. The primitive version of the mechanism must have been far simpler, smaller and less accurate.

Imprecision in translation would have required the synthesis of proteins that were smaller, and therefore less specif­ic in their action, than proteins are to­ day. (Otherwise the probability of error in making a protein strand would have been too great.) Among the smaller and less specific proteins would have been the enzymes required to process genetic information. If those enzymes were less precise than today's versions are, the cell's mutation rate must necessarily have been higher and the size of its genome correspondingly smaller. The translation process is the link between genotype and phenotype, between infor­mation and its expression; as the process evolved to become more precise, the cell itself necessarily passed through a corresponding series of evolutionary refinements. It evolved from an entity having simple properties, imprecise and general functions and a rather small complement of genes to an entity that functioned with many highly specific enzymes and a complex, precise genetic apparatus.  To emphasize the primitive genetic and translational mechanisms of the earlier, simpler cells, I call them progenotes.

The progenote as ancestor

The discovery of the archaebacteria provides the perspective needed to approach the question of whether the universal ancestor was a prokaryote or a progenote. Although the question is far from settled, the initial indications are that the universal ancestor was indeed a progenote. First of all consider that true bacteria and archaebacteria have proba­bly existed for at least 3. 5 billion years. The time needed for the evolution of the first true bacteria or archaebacteria, then, had to be less than a billion years, and perhaps much less. Still, the kinds of evolutionary changes that have arisen within each of the bacterial kingdoms over the later interval of three billion years or more are minor compared with the differences that separate archaebac­teria from true bacteria, such as the dif­ferences in lipids, in transfer-RNA and ribosomal-RNA sequences and modifi­cation patterns and in enzyme-subunit structure.

It would seem that the nature of evo­lution some four billion years ago was very different from what it was later. This implies that the organisms under­ going the evolution were also very dif­ferent. The possibility that the universal ancestor was in the process of develop­ing the cell wall is suggested by the fact that archaebacterial cell walls are as un­like true bacterial walls as eukaryotic walls are. Perhaps the universal ances­tor was still developing or refining bio­chemical pathways as well; lipids are synthesized differently in the two bac­terial kingdoms, and many coenzymes are different. If archaebacteria should be found to differ from true bacteria in their mechanisms for controlling gene expression (a possibility that has yet to be investigated), the implication would be that their common ancestor may have had only rudimentary mechanisms of genetic control.

The key question is whether the uni­versal ancestor was still developing the genotype-phenotype link when it gave rise to its descendant lines. Two obser­vations suggest it may have been. RNA polymerase is the enzyme that tran­scribes the gene into its messenger-RNA complement (which is then translated into protein). The subunit structure of the RNA polymerase is quite constant among the true bacteria, whereas the archaebacterial polymerase structure is different. Could this mean the RNA­ polymerase function was still being re­ fined at the time the two bacterial lines separated?

The second observation concerns the modified nucleotides in transfer and ri­bosomal RNAs. As I mentioned above, the patterns of modification are almost invariant within any primary kingdom, but they tend to differ between king­doms. Although the function of modi­fied nucleotides in transfer and riboso­mal RNAs is not understood, it is rea­sonable to assume most of them serve to "fine tune" translation: to make it more precise. If that is so, it would appear that many of the modifications have evolved independently in each of the primary kingdoms. The independence of these modifications implies in turn that the universal ancestor did not have today's highly specialized transfer-RNA and ri­bosomal-RNA molecules but made do with more rudimentary translation ma­chinery. At this stage one can say only that the facts are consistent with, and in­ deed suggestive of, the universal ances­tor's having had rudimentary transcrip­tion and translation mechanisms, and so having been a progenote.

The origin of the urkaryote

In terms of their ribosomal-RNA cat­alogues the archaebacteria, eubacteria and eukaryotes appear to be equidis­tant from one another genealogically; no specific relation between any two of the three has been detected. Nevertheless, in terms of the amino acid sequence of one protein, the ribosomal A protein, the ar­chaebacteria clearly seem to be relatives of the eukaryotes. Therefore it may be that archaebacteria as well as true bacte­ria participated in forming the eukaryotic cell. Perhaps it is to the archaebac­teria one should look for the origin of the unexplained stage of the eukaryotic cell: the urkaryote that played host to the endosymbiont ancestors of the mito­chondrion and the chloroplast. (Rather than searching for the hypothetical host, however, one should instead question whether there was such an entity. This is not a time to shape new discoveries in accordance with old prejudices.)

As I have indicated, the differences between the eukaryotic cell and the other major cell types at the molecular level are more extensive and pervasive than any of the differences visible with a microscope. The eukaryotic nucleus appears to contain at least three kinds of genes: those of eubacterial origin (presumed to have been appropriated from the genomes of the eukaryote's organelles), those of archaebacterial origin (for example the gene for the ribosomal A protein), and those of an unidentified third origin (exemplified by the cytoplasmic ribosomal RNA). To what extent is the eukaryotic nucleus genetically a chimera: an entity composed of parts assembled from disparate sources? At what stage (or stages) of evolution did the presumed assembly take place? And what was the nature of the organisms that supplied the various genes and structures?

Biologists have tended to look at the eukaryotic cell as having been formed by the association of fully evolved prokaryotic cells; their association is assumed to have created a "higher" type of cell, the eukaryotic. (The term prokaryote—"before the nucleus"—carries just this implication).) The question that an profitably be asked now is whether the evolutionary agents that gave the eukaryotic cell its basic molecular character really were of this nature. The eu­karyotic cell appears to be a chimera at a very basic level. Even the eukaryotic ribosome seems to be chimeric, its com­ponent RNAs having come from a source other than that of at least one of its proteins. If this is a correct interpreta­tion of the data (and future investiga­tions must settle the point), the eukary­otic cell may be a different kind of entity than it is now taken to be. It may have been chimeric even before it reached a stage of complexity comparable to that of today's prokaryotes; it may have been chimeric as it emerged from the proge­note condition. Rather than being an advanced, "higher" form, the eukaryotic cell may represent a throwback to the evolutionary dynamics of its long-gone ancestor, the progenote.

Perhaps the most exciting thing about the recent discoveries in molecular phy­logeny is that they show how much in­formation about the very early stages of evolution is locked into the cell itself. It is no longer necessary to rely solely on speculation to account for the origins of life. It has become customary to think of the last decades of this century as a time in biology when "genetic engineering" will make possible exciting develop­ments in medicine and industry. It must also be recognized that biology is now on the threshold of a quieter revolution, one in which man will come to under­ stand the roots of all life and thereby gain a deeper understanding of the evolutionary process.