Proc. Nati. Acad. Sci. USAVol. 81, pp. 3786-3790, June 1984Evolution

Eocytes: A new ribosome structure indicates a kingdom with a closerelationship to eukaryotes

(eocyta/eocytic gap/parsimony analysis/unrooted dendrogram/electron microscopy)

JAMES A. LAKE, ERIC HENDERSON, MELANIE OAKES, AND MICHAEL W. CLARK

Molecular Biology Institute and Department of Biology, University of California, Los Angeles, CA 90024

Communicated by Everett C. Olson, March 12, 1984

ABSTRACT Ribosomal large and small subunits are orga-nized in four general structural patterns. The four types arefound 'in ribosomes from eubacteria, archaebacteria, eukary-otes, and a group of sulfur-dependent bacteria (eocytes), re-spectively. All four ribosomal types share a common structuralcore, but each type also has additional independent structuralfeatures. The independent features include the eukaryoticlobes and the archaebacterial bill on the smaller subunit. Onthe larger subunit, they include the eocytic lobe, eocytic gap,and eocytic bulge and a modified central protuberance. Thesedata are most parsimoniously fit by a single unrooted evolu-tionary tree. In this tree eocytes are closely related to eukary-otes, while archaebacterla and eubacteria are closest neigh-bors. The tree is consistent with currently known molecularbiological properties and indicates that eocytes have a phyloge-netic importance equal to that of the three known kingdoms.When other properties and molecular mechanisms of these or-ganisms are better defined, we suggest that an appropriatekingdom name for this group would be the Eocyta.

~MWr

B~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r

In recent years, understanding the evolution of organismshas increasingly been based on phylogenies determined byusing fundamental molecular properties. Notably, it hasbeen proposed on the basis of oligonucleotide cataloging ofribosomal RNA sequences that archaebacteria, eubacteria,and eukaryotes represent three very primitive lines of cellu-lar descent (1). Numerous data on fundamental cellular prop-erties support this proposal (for a collection of papers, seeref. 2). In particular, ribosomes from organisms within theeubacterial, archaebacterial, and eukaryotic lineages eachhave different three-dimensional structures (3). In this paperwe report a fourth type of ribosomal structure found in thesulfur-dependent bacteria that thrive in thermal springs attemperatures above 90'C (4-6).

In the following sections, we describe these major patternsof ribosome structure in order to use the differences amongthem as indicators of deep evolutionary divergences. Thedifferences between these four types of ribosomes are ap-propriately sensitive to delineate even the evolution of lin-eages (3) and are used to derive an unrooted evolutionarytree relating taxa from each group. The four ribosome struc-tures are consistent in a most parsimonious way with only asingle unrooted tree. Eukaryotes and the sulfur-dependentorgaiss(eocytes) are adjacent in the unrooted tree, andthe archaebacteria and eubacteria are adjacent. This treeseems to be well supported by known information on otherfundamental molecular properties. Based on their uniquephylogenetic positions, it appears to us that the sulfur-de-pendent bacteria (4) consisting of the subdivisions Sulfolo-bales and Thermoproteales have a phylogenetic importanceequal to that of the other three kingdoms. If this interpreta-

DFIG. 1. Electron micrographs of the rtbosomes of representative

eubacteria, archaebacteria, eocytes, and eukaryotes. From left toright, the organisms are Synechocystis 6701, a cyanobacterium (eu-bacteria); Halobacterium cutirubrum, an extreme halophile (archae-bacteria); Thermoproteus tenax, a facultative anaerobe (eocyte);and Saccharomyces cerevisiae, a yeast (eukaryote). (x 250,000.)Small subunits are shown in row A, large subunits are shown in rowC, and diagrams of their profiles are shown below each micrograph(in rows B and D).

tion is confirmed by subsequent additional data for these or-ganisms, we propose that an appropriate kingdom name forthese organisms would be Eocyta (dawn + cell).

MATERIALS AND METHODSRibosomes and ribosomal subunits of eubacteria, archaebac-teria, and eukaryotes were prepared as described (3). In ad-dition, eocytic ribosomal subunits were resuspended in thefollowing buffer: 200 mM NH4CI/5 mM Tris-HCI, pH 7.6/10MM MgCl2. Substitution of the buffer used for eubacterialribosomes for use with archaebacterial ribosomes, eukaryot-ic ribosomes, or eocytic ribosomes produced no differencesin ribosomal profiles. Subunits in these buffers were nega-tively stained by the double-layer carbon method. Relativesizes of eukaryotic, archaebacterial, eocytic, and eubacterialsubunits were determined by electron microscopy of pair-wise mixtures of subunits from the four groups.

RESULTSRepresentative electron micrographs of ribosomal subunitsfrom eubacteria, archaebacteria, eocytes, and eukaryotesare shown in the first through fourth columns, respectively,

3786

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Janu

ary

20, 2

020

Evolution: Lake et aL

of Fig. 1. Small subunits are illustrated in row A and largesubunits are in row C. These are interpreted in diagrams be-low each figure (rows B and D). Large subunits are shown ina projection that is useful for comparative purposes, the"quasisymmetric" projection (7). Identification of this pro-jection of the eocytic and eukaryotic large subunits was de-termined by comparison with the eubacterial structure. Thesmall subunit profile that is shown is the "asymmetric pro-jection" of the small subunit. This is the same projection thathas been used previously to compare the three-dimensionalstructures of archaebacterial, eukaryotic, and eubacterialsmall subunits (3).Two fields of large subunits are shown in Fit. 2. In both

fields the predominant subunit view is the quasisymmetricprojection. Ribosomal large subunits from the eocyte Ther-moproteus tenax are shown in Fig. 2B. Subunits in the quasi-symmetric projection in this field can be recognized by thepresence of an indentation, the "eocytic gap," that givesthese ribosomes their characteristic shape. This gap sepa-rates a region of density at the bottom of the subunit, the"eocytic lobe," from a second region on the side of the sub-unit, the "eocytic bulge." For comparison, a field of largesubunits from the archaebacterium Halobacteriym cutiru-brum is shown in Fig. 2A. In both archaebacteria and eubac-teria, the lobe and bulge are absent, or nearly so.

Proc. Natl. Acad. Sci. USA 81 (1984) 3787

The eocytic gap is indicated by arrows in micrographs ofthe large subunits from five eocytic ribosomes in Fig. 3BLeft. In micrographs of representative eukaryotic large sub-units shown in Fig. 3B Right, both the lobe and the bulge arepresent, and the gap between them is filled. The eocytic lobeand bulge are present in the large ribosomal subunits of botheocytes and eukaryotes but the gap is unique to eocytic taxa.

In the small subunit, the lobes at the base of the eocyticsubunit are modified so that their structure, although pro-nounced, is intermediate between that of the archaebacterialand the eukaryotic small subunits. Both this modification inSulfolobus acidocaldarius small subunits and the lack of eu-karyotic lobes in small subunits of eubacterial and archae-bacterial subunits have been noted previously (3). For com-parison, the lobes are indicated by arrows in micrographs offive eocytic small subunits in Fig. 3A Left and in micro-graphs of representative eukaryotic small subunits in Fig. 3ARight. These features are summarized in Fig. 4.

DISCUSSIONFour Designs of Ribosomes Are Most Parsimoniously Fit by

Only One Unrooted Dendrogram. Within each of these fourtypes, three-dimensional ribosomal structure is relativelyconstant (for a discussion, see ref. 3). Hence, the variationsin structure between these four groups (outlined in Table 1)

FIG. 2. (A) An electron micrograph of a field of large ribosomal subunits from the archaebacterium Halobacterium cutirubrum. (B) Anelectron micrograph of a field of large ribosomal subunits from the eocyte Thermoproteus tenax. The eocytic gap is indicated by arrowheads.Both micrographs are at the same magnification, indicated by the 500-A bar.

Dow

nloa

ded

by g

uest

on

Janu

ary

20, 2

020

Proc. Natl. Acad. Sci USA 81 (1984)

4p,'.,

.. ... . . . ... .... ....

*: S b _

s. s&. te

,, ., _, ,. >

'^ />.,A,* . j'.t'2 t '.!. "<8

..

R~~~~~~~~~~~~~~~~

FIG. 3. Electron micrographs of small subunits (A) and large subunits (B) of eocytes (Left) and representative eukaryotes (Right). The

eocytic organisms, from left to right, are Thermoproteus tenax, Sulfolobus acidocaldarius, Desulfurococcus mucosus, Thermofilum pendens,

and Thermococcus celer. The eukaryotic organisms are, from left to right, Tetrahymena thermophila (a ciliate), Saccharomyces cerevisiae (a

yeast), and Triticum aeshivum (wheat).

provide a phylogenetic basis for relating their evolution. If,as the constancy of ribosome structure within lines and atthe resolution limit of our images suggests (3, 9), the individ-ual structural features of each ribosomal type arose onlyonce, then a parsimony analysis is appropriate. Of the threepossible unrooted dendrograms for connecting four taxa,only one dendrogram most parsimoniously fits the data inTable 1 (10). This tree, shown in Fig. 5, places the eukary-otes and eocytes on one side and the eubacteria and archae-bacteria on the other. It is not the only possible interpreta-tion of our data, but it is the simplest.Data of Others Suggest Three Possible Rootings for the

Tree. Structural data alone cannot root the tree; however, incombination with the oligonucleotide catalog-derived SABdata (12) and DNA-rRNA hybridization data (13) summa-

rized in Fig. 6, the tree can be rooted. According to this anal-ysis, the three right-most branches of Fig. 5 are the mostlikely roots. In both of the trees shown in Fig. 6 A and B, thedistance between the transition organism Thermoplasma andthe archaebacteria is greater than that between any two dif-ferent archaebacterial taxa and less than that between Sulfo-lobus (an eocyte) and archaebacteria. Hence, we concludethat the branching of eubacteria is that shown in Fig. 6C. Therooted trees consistent with these constraints are the threetrees in Fig. 6 D-F.

If these arguments are correct, then certain ribosomal fea-

Central Protuberance

17/12 stalk

- Bill

Body

FIG. 4. A summary of the structural features found in the fourribosomal types. Features found in eocytes (but absent in both ar-

chaebacteria or eubacteria) are indicated by diagonal (lower left toupper right) striping. Eukaryotic features (absent in both archaebac-teria and eubacteria) are indicated by diagonal (lower right to upperleft) striping. Crosshatched features are present in both eocytes andeukaryotes and lacking in both eubacteria and archaebacteria.

tures, such as the bill, must be primitive. The three rootedtrees also imply that eocytes and eukaryotes are "older"than (Fig. 6 D and E), or at least as old as (Fig. 6F), thearchaebacteria and eubacteria. It is interesting to note thatall three rooted trees are consistent with the proposals (15,16) that the ability to splice RNA is a primitive property.Considered as one of four kingdoms, eocytes are a mono-

phyletic group. If we attempt to include eocytes in with ei-ther the archaebacteria or the eubacteria, then the largergroup defined in this way becomes paraphyletic. Consider,for example, the tree shown in Fig. 6F. If eocytes are to belumped with archaebacteria, then both the eukaryotic andeubacterial kingdoms would be derived from the archaebac-terial kingdom so that these two derived groups would havea phylogenetic status inferior to that of archaebacteria. (Sim-ilar arguments apply to the four other rooted trees.) If king-doms are to represent monophyletic groups, it is necessaryto give eocytes kingdom status. Ultimately, it is the specialrelationship of eocytes to eukaryotes, shown in the unrootedtree, that forms the basis for classifying eocytes in a separatekingdom.The Phylogeny and Molecular Properties of Eocytes Indi-

cate a New Kingdom. The close, phylogenetic relationship ofeocytes to eukaryotes, when compared with archaebacterialand eubacterial features, is supported by the similarity ofmany molecular processes in eocytes and eukaryotes. Boththe fundamental nature of these common properties and thevariety of the processes involved strongly support the pro-posed phylogeny and classification.Many molecular properties of eocytes are more like those

of eukaryotes than they are like those of archaebacteria and

Table 1. Structural features of the four ribosomal types

Large subunit Small subunit

Lobe Filled gap Bulge Bill Lobes

EubacteriaArchaebacteria - + - - +Eukaryotes + + + + + + +Eocytes + + - + + + +Thermofilum + + - + + + +Thermococcus + + - + + + +Desulfurococcus + + - + + + +Sulfolobus + + - + + + +Thermoproteus + + - + + + +Thermoplasma + - + + + +

Within eocytes, the size of the 50S lobe varies from large (+ + inThermoproteus tenax) to intermediate (+ in Thermoplasma aci-dophilus). For this reason we regard Thermoplasma as representinga transitional organism. In archaebacteria, and to a lesser extent ineubacteria (see ref. 8), a small 50S lobe is present. Details of thestructural organization of eocytes will be discussed elsewhere.

3788 Evolution: Lake et aL

Dow

nloa

ded

by g

uest

on

Janu

ary

20, 2

020

Proc. Natl. Acad. Sci. USA 81 (1984) 3789

Eubacteria

3c

Archaebacteria

"I..

FIG. 5. The unrooted dendrogram relating the steps in the evolu-tion of taxa from the four lineages. This is the most parsimonioustree (11) relating the four groups. Characters listed as "+ +" in Table1 are assumed to represent ordered transitions (11) from + + to + to-. The ribosomal subunits shown at the nodes correspond to theribosomes present in the most parsimonious interpretation.

eubacteria. The DNA-dependent RNA polymerases of eo-

cytes are composed of protein subunits with molecularweights and immunological properties that resemble those ofeukaryotic polymerase A(I) more closely than the patternsfound in archaebacteria and eubacteria (14). Introns similarto those found in eukaryotic tRNA genes have been found in

A Woese el ox B Zillig et a/ARCHAE- Thermo- Sulfo- ARCHAE- MTermo-

plosmo lobus BACTERIA plosmo EOCYTES

SAB

.24 SF

.2 r40_.4.16-{ 35-

CInterpretation Using Unrooted Dendrogram

EU- ARCHAE-

BACTERIA BACTERIA EOCYTES

Conclusion Consistent With:

D E

FIG. 6. The three rooted trees that are consistent with data ob-tained from oligonucleotide catalogs and DNArRNA hybridizationdata. A SAB of 1.0 indicates perfect oligonucleotide correspondencebetween 16S rRNAs of taxa A and B, and a SAB of 0.0 indicates no

common oligonucleotides. Similarly, SFS of 1.0 and 0.0 represent100% and 0% hybridization, respectively, between pairs of taxa. (A)Summary of the oligonucleotide catalog data using SAB analysis(12). Sulfolobus is an eocyte. (B) Summary of the DNA-rRNA cross-

hybridization data using fractional stability, SF, analysis (13). Thesulfur-dependent bacteria are listed as eocytes. (C) The unique inter-pretation for the placement of eubacteria that is consistent with theunrooted dendrogram and with the data in A and B. (D-F) The threealternative trees that are consistent with the interpretation in C.

tRNA genes in the eocyte Sulfolobus (17), whereas none hasyet been found in archaebacteria or eubacteria. The second-ary structures of 55 ribosomal RNA from Sulfolobus, andfrom Thermoplasma to a lesser degree, while related to theeukaryotic 55 pattern, differ more from the archaebacterialand the eubacterial palterns than does even eukaryotic 5SrRNA (18). The ii'.. tator tRNAs of archaebacteria show acloseness to euba% aria (for example, the sequence of theinitiator tRNA of Halococcus resembles those of eubacteriamore than those of eukaryotes), while the sequence of Sulfo-lobus initiator tRNA is closer to those of eukaryotes, andthat of Thermoplasma is intermediate (19). Significantamounts of long polyadenylylated sequences are found inSulfolobus RNA that are similar to those in eukaryoticmRNAs, whereas only much lower amounts (1/30th) arefound in eubacteria (20).The modes of cell division in eocytes also mimic those

found in eukaryotes. Eocytes do not divide equally by sep-tum formation as the archaebacteria and eubacteria do, rath-er they divide unequally by budding, constriction, branch-ing, and other mechanisms (4). While archaebacteria andeocytes both utilize glycerolipids with ether linkages, signifi-cant differences are found between the kingdoms. The te-traether membrane lipids of eocytes (Sulfolobus and Ther-moplasma to a lesser extent) contain cyclopentanol C40-bi-phytanol chains. In addition, their neutral lipids also containbranched alkyl benzenes (21). The membrane lipids of ar-chaebacteria and eubacteria contain neither. The lipids ofeocytes contain a high percentage of tetraethers, in contrastwith archaebacteria, which contain relatively less or none(21). Thus, although the molecular mechanisms of the eocyteare only starting to be explored, it appears that eocytes are agroup with unique cellular properties appropriate for thephylogeny and kingdom classification proposed in this pa-per.We think that the reasons for considering eocytes to be a

separate kingdom are as compelling as those for the choiceof the other three. The evolutionary tree is supported by dataon a broad range of basic molecular properties. In this treeeocytes occupy a special position with respect to the eukary-otes and, hence, as a group are likely to be extremely usefulin future investigations into the origin of the eukaryotic cell.Although only a small number of eocytes have been isolated,both anaerobic and aerobic taxa are known, indicating phy-logenetic diversity. Furthermore, many eocytes are capableof a novel anaerobic, purely chemolithoautotrophic metabo-lism utilizing H2, C02, and elemental sulfur as a terminalelectron acceptor (6). Hence, they prominently occupyniches appropriate for modern day descendents of primitiveorganisms. We think their crucial evolutionary position andunique molecular biological properties, when more fully ex-plored, will support the designation of independent kingdomstatus.

Note Added in Proof. The unrooted evolutionary tree described inthis paper differs from the alternative advanced by Woese and Fox(ref. 1; also see ref.12). In their proposal, the earliest divisions areinto the three lines consisting of the archaebacteria, the eubacteria,and the eukaryotes. In their scheme, Sulfolobus (and the eocytes)subsequently branch from the archaebacterial lineage. We have re-ferred to the unrooted tree corresponding to this sequence of events(shown below) as the "archaebacterial tree." (For the rooted versionof this tree see ref. 12.) Also shown is the tree proposed in this pa-per, the "eocytic tree." When more extensive data become availablefor taxa from all four groups, we anticipate that they will permitfurther testing of these two alternative theories. Because evolution-ary distances between taxa normally are interpreted by assumingconstant rates of evolution (an assumption not likely to apply whengroups as diverse as eubacteria, archaebacteria, eocytes, and eu-karyotes are considered), we emphasize that it is the topology ofthetree that must be tested and not the distances.

Evolution: Lake et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

20, 2

020

Proc. NatL. Acad. Sci. USA 81 (1984)

Eocytes Eubacteria

Archaebacteria Eukaryotes

ARCHAEBACTERIAL TREE(from reference 12)

Eubacteria Eocytes

Archoebacteria Eukaryotes

EOCYTIC TREE(this paper)

We thank J. Washizaki for expert electron microscopy and photo-graph assistance and R. Dickerson, D. Eisenberg, W. Fitch, W. Zil-lig, A. Matheson, E. Smith, and A. Klug for discussions. We thankJ. William Schopf for suggesting the name eocytes. We thank W.Zillig for providing cells from eocytes, A. Matheson for providingcells and ribosomal subunits from archaebacteria and eocytes, andC. Brunk for providing cells from Tetrahymena. This work was sup-ported by research grants from the National Science Foundation(PCM 76-14710 to J.A.L.) and the National Institute of GeneralMedical Science (GM 24034 to J.A.L.).

1. Woese, C. R. & Fox, G. E. (1977) Proc. Nati. Acad. Sci. USA

74, 5088-5090.

2. Kandler, O., ed. (1982) in Archaebacteria (Fischer, Stuttgart).3. Lake, J. A., Henderson, E., Clark, M. W. & Matheson, A. T.

(1982) Proc. NatI. Acad. Sci. USA 79, 5948-5952.4. Zillig, W., Schnabel, R. & Stetter, K. O., Curr. Top. Microbi-

ol., in press.5. Zillig, W., Stetter, K., Schafer, W., Janekovic, D., Wunderl,

S., Holz, L. & Palm, P. (1981) Zentralbl. Bakteriol. Mikrobiol.Hyg. Ser. C 2, 205-207.

6. Fischer, F., Zillig, W., Stetter, 0. & Schreiber, G. (1983) Na-ture (London) 301, 511-513.

7. Lake, J. A. (1981) Sci. Am. 245, 84-97.8. Marquis, D. M., Fahnestock, S. R., Henderson, E., Clark,

M., Schwinge, S., Clark, M. & Lake, J. A. (1981) J. Mol. Biol.150, 121-132.

9. Henderson, E., Pierson, B. K. & Lake, J. A. (1983) J. Bacteri-ol. 155, 900-902.

10. Fitch, W. M. (1971) Syst. Zool. 20, 406-416.11. Fitch, W. M. (1977) Am. Nat. 111, 223-257.12. Woese, C. R. (1981) Sci. Am. 244, 98-122.13. Tu, J., Prangishvilli, D., Huber, H., Wildgruber, G., Zillig, W.

& Stetter, K. 0. (1982) J. Mol. Evol. 18, 109-114.14. Zillig, W., Stetter, K., Schnabel, R., Madon, J. & Gierl, A.

(1982) in Archaebacteria, ed. Kandler, 0. (Fischer, Stuttgart),pp. 218-227.

15. Darnell, J. (1978) Science 202, 1257-1260.16. Doolittle, W. F. (1978) Nature (London) 272, 581-582.17. Woese, C., Kaine, B. & Gupta, R. (1983) Proc. NatI. Acad.

Sci. USA 80, 3309-3314.18. Fox, G. E., Luchrsen, K. R. & Woese, C. R. (1982) in Ar-

chaebacteria, ed. Kandler, 0. (Fischer, Stuttgart), pp. 330-345.

19. Kuchino, Y., Hiara, M., Yabusaki, Y. & Nishimura, S. (1982)Nature (London) 298, 684-685.

20. Ohba, M. & Oshima, T. (1982) in Archaebacteria, ed. Kandler,0. (Fischer, Stuttgart), p. 353.

21. Langworthy, J. A., Tornabene, T. G. & Hoker, G. (1982) inArchaebacteria, ed. Kandler, 0. (Fischer, Stuttgart), pp. 228-244.

3790 Evolution: Lake et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

20, 2

020