8. On centrosome duplication: from yeast to man
To what extent are the mechanisms of centrosome duplication
conserved across evolution? Historically, two ideas were most
prevalent to explain the apparent self-reproduction of the cen-
trosome: first, a crystallization-based mechanism, with the
local concentration of a given component acting as a nucleating
agent; second, a nucleic acid-based mechanism, whereby an
analogous principle to that governing replication of the genetic
material would hold for duplicating the centrosome. Whereas
there is currently no solid evidence in favour of the second pro-
posal, the local oligomerization of SAS-6 proteins at the site of
cartwheel assembly supports the first idea. The duplication of
the spindle pole body (SPB) in the budding yeast
Saccharomyces
cerevisiae
also relies in part on the first mechanism, with Spc42p
forming a two-dimensional crystal at the core of the satellite
that will form the new SPB [46]. However, Spc42 crystallization
does not occur at the very onset of the duplication process.
Instead, the most initial step entails duplication of the so-
called half-bridge according to a remarkably simple molecular
mechanism of mirror-image assembly (see article by John
Kilmartin) [47]. Is this principle evolutionarily conserved, per-
haps representing the core of an ancient mechanism that is
present but not yet appreciated in the context of centriole dupli-
cation? Conceivably, the two major components of the SPB
half-bridge, Cdc31p and Sfi1p, which are the only SPB com-
ponents present in
S. cerevisiae
,
Schizosaccharomyces pombe
and
vertebrates, may participate in a similar mechanism in metazo-
ans. If so, where should one look for the presence of this
mechanism in animal centrosomes. Perhaps in the connection
between the nucleus and the centrosome, which is ensured
by the half-bridge in
S. cerevisiae
; or else in the link between
the centriole and the procentriole, by analogy with the relation-
ship between the half-bridge and the SPB. An electron-dense
structure connecting the proximal end of the parental centriole
with the nascent procentriole has been observed in human cell
[48], but neither the Cd31p-related protein Centrin3 nor the
Sfi1p homologue hSfi1 seem to be enriched at that location.
If not Centrin3 or hSfi1, what else may initiate the process
of procentriolar formation in animal cells? Intriguingly, the
onset of formation is preceded by a change in the distribution
of the Plk4 kinase from being uniformly distributed around
the proximal part of the parental centriole to being con-
centrated on a single site: this transition may represent a
critical symmetry breaking event [38] (also see articles by
Kip Sluder [49], as well as by Elif Firat-Karalar and Tim
Stearns [50]). Does such Plk4 concentration occur next to a
specific triplet microtubule? Rotational asymmetries do exist
around basal bodies, as evidenced by the stereotypical distri-
bution of rootlets and other structures associate with basal
bodies in flagellates and ciliates [51]. Does this occur because
each triplet microtubule is unique in some way, or does this
stereotyped distribution reflect asymmetries inherent in the
cytoskeletal elements that are connected to the basal bodies?
Regardless of whether the local concentration of Plk4
initiates procentriole formation, it is important to note that
the current body of evidence speaks against the notion of
‘templating’ in the strict sense of the term for centriole dupli-
cation. Indeed, there is currently no evidence for the copying
of a putative mould, be it made of nucleic acids or proteins,
and which would be present in the parental centriole to then
serve as a blueprint for procentriole assembly. Therefore, the
term ‘templated formation’ that has been used often to describe
the process by which a procentriole is seeded next to a parental
centriole in proliferating cells does not seem appropriate.
Instead, we suggest using the more neutral term of ‘centriole-
guided’ procentriole formation. Whatever the term eventually
adoptedby the community, it isnecessary todistinguish the tem-
plated formation of ciliary and flagellar axonemes fromthe basal
body from such centriole-guided formation of procentrioles.
9. On the connection with the nucleus
Not only does the centrosome reproduce during the cell cycle,
but so also do cellular constituents that are associated with it.
Indeed, the centrosome is not free within the cytoplasm, but
instead is anchored to other compartments, particularly the
nucleus. This connection is essential for the migration of nuclei
in large eggs or in the fly syncytial embryo, as well as for overall
cell polarity. The flagellar apparatus is also connected to the
nucleus in most unicellular organisms, with rare exceptions
such as kinetoplastids where the basal boy is connected to the
kinetoplast [52]. Importantly, the connections that anchor centro-
somes have to be reproduced along with the centrosome itself in
order to retain them in the two daughter cells. This renders the
complete reproduction of the centrosome a topologically com-
plex process. The positioning of centrosomes at the cell centre
depends on the dynamics of microtubules that are anchored pri-
marily on the mother centriole, whereas the daughter centriole
alone cannot remain at the cell centre [53]. However, upon depo-
lymerization of cytoplasmic microtubules, the drift of themother
centriole from the central position of the cell is minimal [53], indi-
cating that other connections contribute to proper positioning.
Perhaps distal appendages play a role in anchoring the mother
centriole to another network such as the actin cytoskeleton.
Could the need to maintain connections with other
compartments explain the striking conservation of ancestral cen-
trin genes, whose products are concentrated in the distal lumen
of centrioles and accumulate early duringprocentriole formation
(figure 1
a
)? Centrin was discovered in two green algae as the
major component of calcium-dependent contractile striated fla-
gellar roots connecting the basal bodies to the nucleus, the so-
called Nucleus Basal Body Connector (NBBC) [54,55]. Calcium
treatment causes shortening of the connector fibres [55]. In
C.
reinhardtii
, this connector forms a sort of perinuclear basket,
and mutation of the centrin gene results in defective connection
between the basal bodies and the nucleus, as well as variable fla-
gella number, indicating that the NBBC is instrumental in
coordinating the segregation of basal bodies and nucleus [56].
The
C. reinhardtii
centrin gene defines a sub-family of centrin
genes (CEN2), the presence of which always correlates with that
of a basal body/axoneme motile apparatus, being for exam-
ple absent in plants, higher fungi such as yeasts, or animals
like
C. elegans
[13]. Accordingly, loss of centrin2 alters primary
ciliogenesis and promotes abnormalities related to ciliopathies
in zebrafish embryos [57]. Another centrin gene, discovered as
a Cell Division Cycle gene (CDC31) in the yeast
S. cerevisiae,
is
required for SPB duplication and cannot be complemented
by centrin2 genes, thus defining another conserved centrin
sub-family (CEN3) [58]. Present in fungi and most animals,
but absent in plants, the CEN3 subfamily apparently co-evolved
with the presence of centrosomes or SPBs, with the notable
exception of flies and worms [13]. The CEN3 subfamily could
participate in the connection between the nucleus and the
centrosome/SPB in some species, as it does in
S. cerevisiae
,
rstb.royalsocietypublishing.org
Phil. Trans. R. Soc. B
369
: 20130452
5
on October 7, 2014
rstb.royalsocietypublishing.org
Downloaded from
Book Recommendation
