translational modifications on the binding of microtubule-
associated proteins or motors on neuronal microtubules [18],
one possibility is that such modifications promote the associ-
ation of PCM proteins with centriolar microtubules, which in
turn could contribute to centriole stability. Such reciprocal
stabilization of the two compartments would increase over
time, resulting in the accumulation of post-translational modi-
fications, which are thus candidate biomarkers of centriole age.
Another question in the realm of centriolar microtubules that
deserves further investigation concerns the rare
d
- and
1
-tubu-
lin, which are required, respectively, for the addition of C or B
plus C microtubules [19–22]. What exactly do these tubulin
variants bring to centrioles in those species that have them
and how can they be dispensable in other species?
6. On centriole architecture
The structural complexity of the centriole is simply amazing,
from the intricate cartwheel present within the procentriole at
the onset of the assembly process to the elaborate appendages
of the mother centriole added at the end of it (see article by
Mark Winey and Eileen O’Toole) [23]. A central question in
the field has been to unravel the mechanisms governing the
near universal ninefold symmetry of this intricate biological
edifice. The discovery that proteins of the SAS-6 family can
self-assemble into ring-like structures from which emanate
nine spokes that resemble a slice of the cartwheel provides an
organizing principle at the root of such symmetry [24,25]
(figure 1
b
; see article by Masafumi Hirono [26]). Because the
nine spokes in these structures correspond to the coiled-coil
domain of SAS-6 proteins, this model suggests that the
length of this domain may determine the diameter of the cen-
triole, although whether this is the case remains to be tested.
An interesting question pertains to how length of the
centriole is controlled. The procentriole reaches a length approxi-
mately 80%of that of the parental centriole by the time of mitosis,
with the remaining elongation taking place in the subsequent cell
cycle [27,28]. How is centriole length precisely regulated? In
human cells and in
Drosophila
, POC1 proteins are important for
such regulation, since their depletion causes shorter centrioles
while their overexpression results in overly long centrioles
[29,30]. In human cells, overexpression of CPAP also leads to
overly long centrioles, as does depletion of CP110, a protein
which normally caps the very distal end of centrioles [31–33].
Although uncovering these components is an important step
forward, the underlying mechanisms remain sketchy.
Is centriole length set by the extent of microtubule polymer-
ization? What components, if any, regulate the polymerization
of centriolar microtubules is not known. Given that the centriole
is approximately 450 nm long in human cells, and that the
polymerization rate of interphasic cytoplasmic microtubules is
approximately 10
m
m min
2
1
[16], it should take only a few
seconds for microtubules exhibiting regular dynamics to reach
this length. And yet procentriolar elongation takes hours in pro-
liferating somatic cells [27]. Perhaps some of the centriolar
proteins that bind tubulin dimers or microtubules, such as
CPAP or Cep135 in human cells [34,35], can modulate polymer-
izationdynamics in amanner that differs substantially fromthat
of conventional microtubule-associatedproteins (MAPs). Alter-
natively, perhaps negative regulation of conventional MAPs,
which would lead them to function less efficiently than usual,
takes place during centriole assembly. Clearly, the analysis of
centriolar microtubules and their regulation is an exciting fron-
tier of research. Centriole elongation also involves the assembly
of an intra-luminal structure in the distal end (figure 1). Except
for the presence of Centrin proteins and the Centrin-binding
protein POC5, which is required for the assembly of the distal
part of centrioles [36], little is known about the molecular com-
position of this intra-luminal structure or about its interaction
with the centriole wall. This interaction could, however, partici-
pate inmaintaining the cohesion of the centriole and controlling
its diameter [37].
Increased understanding of questions related to centriole
architecture offers the exciting prospect of obtaining molecu-
lar handles to tinker with the underlying principles so as to
probe their biological significance: what would be the conse-
quences of having a procentriole that has the wrong diameter
or is too short? A thorough understanding of the system
might even enable one to engineer centrioles at will. Synthetic
centrioles are on the horizon.
7. On pericentriolar material organization
How centrioles can help organize the PCM is not yet clear, but
perhaps the exceptional stability of centriolar microtubules pro-
vides a favourable milieu for recruiting
g
-tubulin-containing
nucleation complexes. Recent results obtained by super-
resolution microscopy indicate that the PCM is organized in a
stereotyped manner around the centriole [38,39], in line with
the notion that the latter organizes the former. However, the
influence may be bidirectional here as well. Perhaps the organ-
ized network of PCM proteins surrounding the proximal part
of the parental centriole is key in ensuring that the procentriole
forms orthogonal to it. There is a land of discovery ahead regard-
ing the mechanisms mediating interaction between centrioles
and PCM, as exemplified by work with the
Drosophila
PCM
protein Cnn [40]: what are the association kinetics of proteins
that bridge the external wall of centrioles with the innermost
part of the PCM and have properties that allow them to trans-
form the order inherent to centrioles into ordered assembly of
the surrounding PCM? More generally, what physico-chemical
properties explain why the PCM excludes ribosomes, for
example, and allow the concentration of many specific proteins
and activities? More generally, how is the boundary of the
PCM controlled (see article by Tony Hyman and collaborators)
[41]? During the cleavage divisions of early embryos, part of
the answer is cell size, as was evident from the days of Boveri
and established quantitatively more recently in
C. elegans
[42].
How does the PCM interact with the two sets of append-
ages associated with the mother centriole? Insights into this
question could come from analysing how appendages
attach to the centriole during mitotic exit. Analysis by elec-
tron microscopy established that centrosome organization is
modified during mitosis, with the PCM forming a perfect
halo around each parental centriole and the appendages tran-
siently disappearing from the mother centriole before
reforming on both parental centrioles [43–45]. This is also
the moment when the former daughter centriole reaches its
full length and thus completes the centriole maturation pro-
cess. Investigations of centrosome remodelling during
mitosis promises to yield interesting insights about the com-
pletion of centriole biogenesis, which may be coupled to the
disengagement step and the priming of centriole duplication
that take place at this moment.
rstb.royalsocietypublishing.org
Phil. Trans. R. Soc. B
369
: 20130452
4
on October 7, 2014
rstb.royalsocietypublishing.org
Downloaded from
Book Recommendation
