R338
Dispatch
Cytoskeletal proteins: The evolution of cell division
David M. Faguy and W. Ford Doolittle
The prokaryotic cell division protein FtsZ and
eukaryotic tubulin have been shown to have very
similar structures and are most likely homologs. The
evolutionary transition from FtsZ to tubulin could
provide a window into the transition from prokaryotic
cells to eukaryotic cells.
Address: Department of Biochemistry and Canadian Institute for
Advanced Research, Dalhousie University, Halifax, Nova Scotia,
Canada B3H 4H7.
E-mail: [email protected]
Current Biology 1998, 8:R338–R341
http://biomednet.com/elecref/09609822008R0338
© Current Biology Ltd ISSN 0960-9822
A vital, and perhaps driving, force for the transition
between prokaryotic — bacterial and archaeal — and
eukaryotic cellular organisation was the development of a
cytoskeleton. It is reasonable to suppose that a cytoskeleton would allow endocytosis — and thus the eventual
capture of endosymbionts that gave rise to mitochondria
and chloroplasts — and, by eliminating the link between
chromosome replication and segregation, an unrestricted
genome size. The search for the prokaryotic origins of
eukaryotic cytoskeletal proteins has thus been a priority
for biologists interested in early evolutionary events.
Actins and tubulins are at the core of the eukaryotic
cytoskeleton, and prokaryotic homologs have been proposed for both: bacterial heat shock protein 70 (Hsp70)
and FtsA for actin [1,2], and FtsZ for tubulin [3–5]. In the
case of actin, comparisons of the three-dimensional
structure were necessary to identify possible homologs —
the resemblance had not been detected by sequence comparisons [2]. With FtsZ, an essential cell division protein
that forms a cytokinetic ring in bacteria and archaea,
homology to tubulin was first argued on the basis of
sequence and functional similarities — both are GTPases
and form proto-filaments [6]. Their cellular functions are
nevertheless distinct, and the sequence similarity is
limited and mainly in the GTP-binding region. Thus
skeptics could argue that the similarity reflects convergence rather than true homology.
Three-dimensional structures have now been reported for
both tubulin and FtsZ. Nogales et al. [7] used electron
crystallography to refine the structure of the αβ tubulin
dimer in zinc-induced crystalline sheets down to 3.7 Å,
while Löwe and Amos [8] used X-ray crystallography to
determine the structure of Methanococcus jannaschii FtsZ
crystals to 2.8 Å. The two structures are remarkably
similar. Both have amino-terminal GTPase domains with
characteristic Rossmann-fold structures. Most of the
residues involved in GDP binding (GDP from the host
expression system was crystallized with FtsZ) are conserved between FtsZ and tubulin. But even the carboxyterminal domain, where there is little sequence
conservation between FtsZ and tubulin, shows considerable structural conservation. In α and β tubulin, this
carboxy-terminal domain is thought to be involved in contacts with ancillary interacting molecules.
The similarity is not just in the overall three-dimensional
structure, but also in the assignment of secondary structure elements to regions of primary sequence (Figure 1).
The similar secondary structure elements of FtsZ and
tubulin are arrayed in a co-linear fashion along the primary
sequence, even when there is little or no underlying
sequence similarity. Taken together, the functional,
sequence, and structural similarities make an almost
unassailable case that FtsZ and tubulin are true homologs
— that is that they are derived from a common ancestral
protein by divergent evolution.
Yet there remains a problem of evolutionary rates.
Tubulin and FtsZ are conservative proteins within
eukaryotes and prokaryotes, respectively [9], and yet the
sequence divergence between them is so great that we
would probably never have suspected homology if not for
the similarity of their GTP-binding sites. Archaeal FtsZs
are very much more similar to bacterial FtsZs than to
tubulins. The phylogenetic tree of bacterial and archaeal
FtsZs, while resolving the major bacterial lineages in the
same way as do generally relied-on phylogenetic markers
such as ribosomal RNA or the translation factor EF1α,
places the archaea near Thermotoga maritima at the base of
the bacterial tree (Figure 2).
If we accept the current version of the ‘universal tree of
life’, in which the deepest divergence — about 3.5 billion
years ago — separates the bacteria from the line which
later — perhaps two-billion years ago — split to produce
archaea and the nuclear–cytoplasmic component of
eukaryotes, there are two possible solutions to the phylogenetic conundrum [9]. The first is that genes for tubulin
were derived by lateral transfer from some as yet undiscovered ‘fourth domain’, as entertained by Sogin in 1991
[10]. As there are many other components of the eukaryotic cellular ultrastructure for which no prokaryotic
homologs whatsoever have yet been identified, we might
suppose that most uniquely eukaryotic features came from
such a mysterious source. The tubulin–FtsZ similarity
Dispatch
R339
Figure 1
1
25
B1
α-tubulin
β-tubulin
Mj FtsZ
α-tubulin
β-tubulin
Mj FtsZ
α-tubulin
β-tubulin
Mj FtsZ
50
H1
75
100
B2
------------------------- --------MRECISIHVGQAGVQIG NACWELYCLEHGIQPDGQMPSDKTI GGGDDSFNTFFSETGAGKHVPRAVF
------------------------- --------MREIVHIQAGQCGNQIG AKFWEVISDEHGIDPTGSYHGDSDL QL--ERINVYYNEAAGNKYVPRAIL
MKFLKNVLEEGSKLEEFNELELSPE DKELLEYLQQTKAKITVVGCGGAGN NTITRLKME---------------- -----------------GIEGAKTV
B4
H4
H2
B3
H3
VDLEPTVIDEVRTGTYRQLFHPEQL ITGKEDAANNYARGHYTIGKEIIDL VLDRIRKLADQCTGLQGFSVFHSFG GGTGSGFTSLLMERLSVDYGKKSKL 200
VDLEPGTMDSVRSGPFGQIFRPDNF VFGQSGAGNNWAKGHYTEGAELVDS VLDVVRKESESCDCLQGFQLTHSLG GGTGSGMGTLLISKIREEYPDRIMN
AINTD-AQQLIRTKA---------D KKILIGKKLTRGLGAGGNPKIGEEA AKESAEEIKAAIQDSDMVFITCGLG GGTGTGSAPVVAEISKK--IGALTV
H8
H7
B5
H5
B6
H6
EFSIYPAPQVSTAVVEPYNSILTTH TTLEHSDCAFMVDNEAIYDICRRNL DIERPTYTNLNRLIGQIVSSITASL RFDGALNVDLTEFQTNLVPYPRAHF 300
TFSVVPSPKVSDTVVEPYNATLSVH QLVENTDETYCIDNEALYDICFRTL KLTTPTYGDLNHLVSATMSGVTTCL RFPGQLNADLRKLAVNMVPFPRLHF
AVVTLPFVMEGKVRMKNAMEGLERL KQ HTDTLVVIPNEKLFEIVPNMP LKLAFKVADEVLINAVKGLVELIT- ---KDGLINVDFADVKAVMNNGGLA
α-tubulin
β-tubulin
B7
H9
B8
H10
B9
PLATYAPVISAEKAYHEQLSVAEIT NACFEPANQMVKCDPRHGKYMACCL LYRGDVVPKDVNAAIATIKTKRTIQ FVDWCPTGFKVGINYEPPTVVPGGD 400
FMPGFAPLTSRGSQQYRALTVPELT QQMFDAKNMMAACDPRHGRYLTVAA VFRGRMSMKEVDEQMLNVQNKNSSY FVEWIPNNVKTAVCDIPPRGLKMSA
MIGIGESDS--------EKRAKEAV SMALNSPLLDVDID---GATGALIH VMGPEDLTLEEAREVVATV----SS RLDPNATIIWGATIDENLENTVRVL
α-tubulin
β-tubulin
H11
B10
H12
LAKVQRAVCMLSNTTAIAEAWARLD HKFDLMYAKRAFVHWYVGEGMEEGE FSEAREDMAALEKDYEEVGVDSV-- EGEGEEEGEEY
TFI--------GNSTAIQELFKRIS EQFTAMFRRKAFLHWYTGEGMDEME FTEAESNMNDLVSEYQQYQDATADE QGEFEEEGEED
LVITGVQSRIEFTDTGLKRKKLELT GIPKI-------------------- ------------------------- -----------
Mj FtsZ
Mj FtsZ
Current Biology
Secondary structure elements superimposed on a protein sequence
alignment for M. jannaschii FtsZ and pig α and β tubulins [7,8]. The
alignment of tubulins and FtsZ is based on alignment of similar
structural elements (α helix, red; β sheet, blue). One can clearly see
the conservation of secondary structure, even in regions of the
alignment with little or no primary sequence conservation. The
secondary structure elements are labeled as in [7].
would then reflect very ancient homology, going back to a
common ancestral gene in the last common cellular
ancestor of all four domains. The second, less romantic,
solution would be that, during the transition from a
prokaryotic to a eukaryotic cellular organization, strong
selection for new functions greatly accelerated the rate of
change of FtsZ and many other molecules — many so
greatly that their prokaryotic homologs are simply not
detectable by current methods.
‘real biochemistry’ [11–13]. These sequences, when analyzed phylogenetically, divide into two distinct branches.
Of the four completed and available archaeal genome
sequences — M. jannaschii [14], Methanobacterium thermoautotrophicum [15], Archaeoglobus fulgidus [16], and Pyrococcus horikoshii — three have two ftsZ genes. Among
bacteria, only Rhizobium meliloti has two ftsZ genes, and
these are recently diverged paralogs (they can even be
considered as developmentally regulated isotypes [17]).
This second scheme would likely require that archaeal
FtsZs and tubulins share a common ancestor more like the
former than the latter in sequence and function. We know
almost nothing about archaeal cell division mechanisms.
The presence of FtsZ and minD — a protein involved in
division site selection — proteins in archaea might argue
that their cell division is more bacterial than eukaryotic in
character, which would be consistent with available
functional analyses. In Halobacterium salinarum, overexpression of FtsZ has been found to impair the normal
process of cell division [11], and in H. volcanii the FtsZ
protein has been localized to a ring that coincides with the
division constriction site [12].
In archaea with two ftsZ sequences, none branch together,
as they would be expected to do if they were recently
diverged paralogs. As shown in Figure 2, these archaeal
genes diverged well before any of the organisms themselves diverged. That it is possible to have two ftsZ genes
independently diverging in the same organism when only
one is needed — M. thermoautotrophicum, for example, has
only paralog 1 — makes a scheme for the evolution of the
tubulins from one of two archaeal ftsZ genes appealing,
although there is no direct evidence for this. The divergence of the two archaeal paralogs is consistent with their
having (at least partially) different functions in the cell.
Conserved gene order among distantly related organisms
is often a clue to function, which means that close examination of the genes that lie close the ftsZ paralogs could
be informative.
Nothing we know is inconsistent with there having been a
progressive, but very rapid, conversion of a single
archaeal-type FtsZ into a tubulin, but it might be easier
still to imagine such conversion occurring in a duplicate
ftsZ gene, released from its customary functional constraints. So it is interesting that the possession of two ftsZ
genes seems to have been ancestral in the archaea, at least
the euryarchaeotes. Eleven FtsZ sequences are available
from archaea, seven from genome projects and four from
In bacteria, ftsZ is almost always found in an operon with
ftsA and genes for other cell-division proteins. Examination of the genes around the ftsZ homologs in archaea
reveals that there is considerable conservation near those
labeled paralog 1 in Figure 2. This paralog is present in
all genomes sequenced and, in all but P. horikoshii,
R340
Current Biology, Vol 8 No 10
Figure 2
M.jannaschii 1
M.thermoautotrophicum Paralog 1
57/58
P.horikoshii 1
100/100
P.woesii
35/*
100/99
Archaea
H.volcanii
100/100
A.fulgidus 1
T.acidophilum
75/76
P.horikoshii 2
Paralog 2
40/27
M.jannaschii 2
100/89
H.salinarum
96/100
A.fulgidus 2
T.maritima
Thermotogales
Cc
100/100
α Proteobacteria
Bb
63/*
Wsp
45/*
β Proteobacteria
Ng
Av
100/100
91/73
Pa
52/*
γ Proteobacteria
Pp
100/100
Hi
97/53
Ec
100/100
Bsp
38/36
Ef
96/69
Sa
Low GC Gram +ve
100/*
Bs
73/25
Mt
92/100
Sf
100/100
High GC Gram +ve
Sc
100/100
59/20
Csp
100/100
32/*
Csp
At plastid
100/52
Pp plastid
Cyanobacteria
69/*
Ssp.
Plastids
100/*
Asp
100/*
Asp
H.pylori
21/43
M.pulmonis
10
Current Biology
74/89
Phylogenetic tree of 26 bacterial and 11 archaeal FtsZ homologs.
While the archaea form a well-supported clade, the sequences clearly
divide into two divergent paralogs. The bacterial FtsZ tree is similar to
that derived from rRNA sequences, with the exception of Helicobacter
pylori and Mycoplasma pulmonis. As cell division is closely tied to cellwall growth in bacteria, M. pulmonis, lacking a cell wall, would be
expected to be divergent. This tree was inferred from amino-acid
translations by distance methods. Bootstrap proportions are shown
(from left to right) for distance and protein parsimony; an asterisk
indicates that the node is not present in the bootstrap consensus tree.
All phylogenetic methods were performed using the PHYLIP package.
Most bacterial species are indicated by the initials of the genus and
species. GenPept accession numbers are available on request. The
T. acidophilum and P. horikoshii sequences are unpublished;
P. horikoshii sequences were obtained through the Pedant genome
analysis web site (http://pedant.mips.biochem.mpg.de/
frishman/pedant.html).
immediately downstream of it lies a gene encoding a
homolog of the SecE/Sec61Y protein that is involved in
protein translocation in both bacteria and eukaryotes
[18]. Farther downstream lie genes for ribosomal proteins. The SecE/Sec61Y and ribosomal protein gene
order is also conserved — but without ftsZ — in many
bacteria. This is significant, as conservation of gene order
across large evolutionary distances is rare [19]. Around
paralog 2, found in all sequenced archeal genomes
except that of M. thermoautotrophicum and also isolated
from Thermoplasma acidophilum and H. salinarum, no conserved gene order is apparent. Interestingly, in P.
horikoshii paralog 2 does have a minD homolog next to it.
All of the archaeal FtsZ sequences to date are from
euryarchaeotes, one of two kingdoms within the archaea.
Without further biochemical and physiological studies to
define the roles of these FtsZ homologs, we are limited in
our conclusions. If, as Figure 2 indicates, ftsZ gene duplication goes as deep as we can so far see in the archaea,
then the duplication might have been present at the base
of the archaea and in the first eukaryotic nuclear genome.
Tubulin-like proteins and functions might thus have
arisen in cells which also retained FtsZ performing its traditional roles in cell division. It will be very interesting to
see if both FtsZ-like and tubulin-like sequences are
present in crenarchaeotes, such as Pyrobaculum aerophilum,
whose genome is completely sequenced but not yet
released, or Sulfolobus solfataricus, whose genome should
be completed within the year.
The elucidation of the structures of tubulin and FtsZ are
not only exciting to evolutionary biologists. This structural
information will help cell biologists understand tubulin
function and dynamics and microbiologists understand
cell division in prokaryotes. It is gratifying to note that
biology has not specialised and diversified so far that
major discoveries in one branch are not still important contributors to the science as a whole.
References:
1. Flaherty KM, McKay DB, Kabsch W, Holmes KC: Similarity of the
three-dimensional structures of actin and the ATPase fragment of
a 70-kDa heat shock cognate protein. Proc Natl Acad Sci USA
1991, 88:5041-5045.
2. Bork P, Sander C, Valencia A: An ATPase domain common to
prokaryotic cell cycle proteins, sugar kinases, actin, and hsp70
heat shock proteins. Proc Natl Acad Sci USA 1992, 89:7290-7294.
3. Mukherjee A, Dai K, Lutkenhaus J: Escherichia coli cell division
protein FtsZ is a guanine nucleotide binding protein. Proc Natl
Acad Sci USA 1993, 90:1053-1057.
4. de Boer P, Crossley R, Rothfield L: The essential bacterial celldivision protein FtsZ is a GTPase. Nature 1992, 359:254-256.
5. RayChaudhuri D, Park JT: Escherichia coli cell-division gene ftsZ
encodes a novel GTP-binding protein. Nature 1992, 359:251-254.
6. Erickson HP, Taylor DW, Taylor KA, Bramhill D: Bacterial cell
division protein FtsZ assembles into protofilament sheets and
minirings, structural homologs of tubulin polymers. Proc Natl
Acad Sci USA 1996, 93:519-523.
7. Nogales E, Wolf SG, Downing KH: Structure of the ab tubulin
dimer by electron crystallography. Nature 1998, 391:199-203.
8. Löwe J, Amos LA: Crystal structure of the bacterial cell-division
protein FtsZ. Nature 1998, 391:203-206.
9. Doolittle RF: The origins and evolution of eukaryotic proteins.
Philos Trans R Soc Lond B Biol Sci 1995, 349:235-240.
10. Sogin ML: Early evolution and the origin of eukaryotes. Curr Opin
Genet Dev 1991, 1:457-463.
11. Margolin W, Wang R, Kumar M: Isolation of an ftsZ homolog from
the archaebacterium Halobacterium salinarium: implications for
the evolution of FtsZ and tubulin. J Bacteriol 1996, 178:13201327.
12. Wang X, Lutkenhaus J: FtsZ ring: the eubacterial division
apparatus conserved in archaebacteria. Mol Microbiol 1996,
21:313-319.
13. Baumann P, Jackson SP: An archaebacterial homologue of the
essential eubacterial cell division protein FtsZ. Proc Natl Acad Sci
USA 1996, 93:6726-6730.
14. Bult CJ, White O, Olsen GJ, Zhou L, Fleischmann RD, Sutton GG,
Blake JA, FitzGerald LM, Clayton RA, Gocayne JD, et al.: Complete
genome sequence of the methanogenic archaeon,
Methanococcus jannaschii. Science 1996, 273:1058-1073.
Dispatch
15. Smith DR, Doucette-Stamm LA, Deloughery C, Lee H, Dubois J,
Aldredge T, Bashirzadeh R, Blakely D, Cook R, Gilbert K, et al.:
Complete genome sequence of Methanobacterium
thermoautotrophicum DH: functional analysis and comparative
genomics. J Bacteriol 1997, 179:7135-7155.
16. Klenk HP, Clayton RA, Tomb JF, White O, Nelson KE, Ketchum KA,
Dodson RJ, Gwinn M, Hickey EK, Peterson JD, et al.: The complete
genome sequence of the hyperthermophilic, sulphate-reducing
archaeon Archaeoglobus fulgidus. Nature 1997, 390:364-370.
17. Margolin W, Long SR: Rhizobium meliloti contains a novel second
homolog of the cell division gene ftsZ. J Bacteriol 1994,
176:2033-2043.
18. Hartmann E, Sommer T, Prehn S, Gorlich D, Jentsch S, Rapoport TA:
Evolutionary conservation of components of the protein
translocation complex. Nature 1994, 367:654-657.
19. Mushegian AR, Koonin EV: Gene order is not conserved in bacterial
evolution. Trends Genet 1996, 12:289-290.
R341
If you found this dispatch interesting, you might also want
to read the April 1998 issue of
Current Opinion in
Structural Biology
which includes the following reviews, edited
by Murray Stewart and John E Johnson, on
Macromolecular assemblages:
Spherical viruses
John J Rux and Roger M Burnett
Filamentous phage structure, infection and assembly
DA Marvin
Chaperonins
Kerstin Braig
F-actin-binding proteins
Amy McGough
Intermediate filament assembly: fibrillogenesis is
driven by decisive dimer–dimer interactions
Harald Herrman and Ueli Aebi
The functional significance of multimerization in
ion channels
M Louise Tierney and Michael HB Stowell
The role of assembly in insulin’s biosynthesis
Gut Dodson and Don Steiner
GTPase-activating proteins and their complexes
Steven J Gamblin and Stephen J Smerdon
The use of solid physical models for the study of
macromolecular assembly
Michael J Bailey, Klaus Schulten and John E Johnson
the same issue also includes the following
reviews, edited by Julia M Goodfellow and
Ronald M Levy, on Theory and simulation:
Electrostatic effects in macromolecules: fundamental
concepts and practical modeling
Arieh Warshel and Arno Papazyan
Protein hydration density: theory, simulations
and crystallography
B Montgomery Pettitt, Vladimir A Makarov
and B Kim Andrews
Simulations of protein folding and unfolding
Charles L Brooks III
Simulations of the molecular dynamics of
nucleic acids
Pascal Auffinger and Eric Westhof
Models and simulations of ion channels and related
membrane proteins
Mark SP Sansom
The full text of Current Opinion in Structural Biology is in
the BioMedNet library at
http://BioMedNet.com/cbiology/stb