Universal Tree of Life

    Because of the long branch attraction artifacts, we searched for molecular sequences which contained structural features, such as inserted segments. Since the insertion of segments happens much less frequently than individual nucleotide changes, they are much less sensitive to long branch artifacts, and can, therefore, be more easily interpreted.

    The molecule we chose to study was protein synthesis elongation factor EF-Tu (EF-1 in eukaryotes), (Rivera & Lake, 1992). EF-Tu is an ubiquitous protein that transports aminoacyl-tRNAs to the ribosome and participates in their selection by the ribosome. Within the GDP-binding domain of EF-Tu, the amino acid sequence, KNMITG ₉₄ , which is strictly conserved in EF-1 and EF-Tu sequences, terminates an -helix and is followed by a -strand that is terminated by GPMP ₁₁₃ at the GDP binding site. The sequence QTREH ₁₁₈ then starts a 3 ₁₀ helix. The amino acid motifs of the eukaryotic EF-1 are similar, except that the four-amino acid sequence GPMP ₁₁₃ in prokaryotes is replaced by the 11-amino acid sequence GEFEAGISKDG, and its variants, in eukaryotes, as shown below.

Eukaryotes    Human         KNMITG TSQADCAVLIVAAGV  GEFEAGISKNG             QTREH

    "         Tomato        KNMITG TSQADCAVLIIDSTT  GGFEAGISKDG             QTREH

    "         Yeast         KNMITG TSQADCAILIIAGGV  GEFEAGISKDG             QTREH



Eocytes       P.occu.       KNMITG ASQADAAILVVSARK  GEFEAGMSAEG             QTREH

   "          D.muco.       KNMITG ASQADAAILVVSARK  GEFEAGMSAEG             QTREH

   "          A.infe.       KNMITG ASQADAAIIAVSAKK  GEFEAGMSEEG             QTREH

   "          Su.sol.       KNMITG ASQADAAILVVSAKK  GEYEAGMSAEG             QTREH



Methanogens   T.celer       KNMITG ASQADAAVLVVAVTD                ---GVMP   QTKEH

& Relatives   Mc.van.       KNMITG ASQADAAVLVVNVDD                AKSGIQP   QTREH



Halobacteria  H.maris       KNMITG ASQADNAVLVVAADD                ---GVQP   QTQEH



Eubacteria    Tt.mar.       KNMITG AAQMDGAILVVAATD                ---GPMP   QTREH

    "         S.plat.       KNMITG AAQMDGAILVVSAAD                ---GPMP   QTREH

    "         Mitoch.       KNMITG AAQMDGAIIVVAATD                ---GQMP   QTREH

Since the eukaryotic 11 amino acid insert is so well conserved among eukaryotic sequences we thought that eocyte sequences might also contain the 11 amino acid insert. Using the polymerase chain reaction and DNA primers designed for use with the KNMITG and QTREH sites, we amplified, cloned, and sequenced the insert region. The eocyte amino acid sequences, translated from DNA, shared the eukaryotic motif (11 amino acids) rather than that found in methanogens, halobacteria, and eubacteria (4 amino acids). The longer 11-amino acid segment, present in eocytes and eukaryotes, shares little similarity with the shorter, four-amino acid segment found in other prokaryotes.     Based on these results, we could directly test the eocyte and archael theories for the origin of the nucleus. [Click on the trees to the left to get a higher resolution image.] The fundamental difference between these two theories is that in the eocyte tree (at the top of the figure) the eukaryotic nucleus shares a most recent common ancestor solely with the eocytes, whereas in the archaebacterial theory eocytes are no closer to eukaryotes than are methanogens or halobacteria, since they are all included within archae.     We have mapped the changes onto the trees representative of both theories. Starting from the four amino insert at the root of the tree, each solid box indicates a change from the four-amino acid segment to the 11-amino acid form. The eocyte tree is favored because it requires only a single change, whereas the archaebacterial tree requires two independent but identical changes. (The archael tree could also be explained by one appearance of the 11-amino acid form and one reappearance of the 4-amino acid form, but even so, two changes would be required.)

    Several lines of reasoning buttress the interpretation that eocytes are the closest relatives of the eukaryotes. First, the 11-amino acid segments present in eocytes and eukaryotes are very likely homologous. Eight of eleven amino acids (seven in Sulfolobus and Acidianus) are identical to the consensus eukaryotic sequence. Amino acid shuffling of the segments produced random alignments that score 6-7 standard deviations lower than those found for the eukaryotic-eocyte alignment, thereby implying homology (Waterman & Eggert, 1987). Second, the alignments are well defined. No gaps are needed to align the eukaryotic and eocytic EF-1 sequences, and no gaps are needed to align the eubacteria, methanogen, and halobacterial sequences. Third, the sequences encoding EF-1 are not likely to have been laterally transferred between organisms, since EF-1 is present in all cells and, during protein synthesis, interacts with cellular components encoded by genes dispersed throughout the bacterial genome, including aminoacyl-tRNAs, ribosomal proteins, elongation factor EF-Ts, and 16S and 18S ribosomal RNAs (Hill, Dahlberg, Garrett, Moore, Schlessinger, & Warner, 1990). These results lend strong support to the proposal that the eukaryotes and eocytes are sister taxa within the tree of life.