| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Anatomy and Structural Biology, Albert Einstein College of Medicine, Bronx, New York 10461, USA
1Correspondence: Department of Anatomy and Structural Biology, Albert Einstein College of Medicine, Bronx, NY 10461 USA. E-mail: satir{at}aecom.yu.edu
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION REFERENCES |
|---|
and reaffirmed my decision to work as a graduate student with him at
the Rockefeller Institute for Medical Research in New York to try to
understand how the structure of the cilium was related to its mechanism
of movement.
Studying a variety of cilia, from protozoan to mammalian somatic
cilia, D. W. Fawcett, Albert Sedar, and Porter (2
, 3)
established at least two critical points: The first was that the
ciliary filaments (microtubules, as we now call them) were continuous
extensions of the basal body filaments surrounded by an extension of
the cell membrane, the ciliary membrane. The inside of the cilium was
therefore a specialized part of the cell cytoplasm. The second point
was that the 9 + 2 pattern of microtubules found in motile plant cilia
(4)
was arranged consistently and distinctively in all
motile cilia as nine doublet microtubules surrounding a central pair.
The 9 + 2 pattern of the axoneme surrounded by the ciliary membrane was
seen in cross-sections of both Paramecium and vertebrate
cilia.
With the passage of forty years, we have learned something about
the relation of structure to how cilia move. A contemporary standard
thin-section (Fig. 1
) shows the familiar outer and inner arms, spokes, and so on that
characterize the axonemal substructure. Axonemes from mussel gill
have a special conspicuous diamond-shaped bridge in one position that
defines doublets 5 and 6. This permits us to number each doublet 1–9
in the axoneme in an unequivocal manner. This sort of knowledge was
critical in the development of the sliding microtubule model of ciliary
motility, which states that motion of the cilium is produced when the
arms cause doublet microtubule N + 1 to slide tipwards with respect to
doublet N (5)
. Gibbons et al. have shown that the arms are
members of the dynein ATPase superfamily of (-) end directed molecular
motors (6)
. Given the mechanical constraints imposed by
linkages within the axoneme such as the radial spokes, this sliding is
translated into a series of bends that then propagate along the length
of the organelles to give the characteristic forms of movement.
|
To clarify the details of how sliding is converted into beating,
Michael Holwill and I have embarked on a project to model the axoneme
accurately by computer modeling techniques at a resolution approaching
4 nm (7
, 8)
. In some ways this is the obverse of the
approach of Mary Porter et al. (e.g., ref 9
) who have used
subtractive image analysis to study structural defects associated with
specific axonemal mutations. Our idea has been to use a variety of
electron microscopic images to construct a 3-dimensional axoneme,
defining each structure as specifically as possible, and then to
empower the model by providing the appropriate dynein step to see which
structures must change and to predict how they must change during
bending (10)
. To do this at high resolution we are now
using an engineering approach, finite element theory.
In our current computer image of the axoneme (8)
,
all the familiar structures are present, and the detailed match to
electron micrographs taken with a variety of procedures is excellent.
Such modeling is necessary and important, because otherwise it has
proven impossible to comprehend what happens to individual structural
elements of the axoneme when some localized dynein arms produce force.
To illustrate the complexity, Fig. 2
shows a view of our current computer model. Any physical prediction can
be reflected in the model and tested by comparison to, for example,
HVEM of quick-fixed bent cilia.
|
This image shows only a small portion of the typical 10 µm long axoneme. There are, for example, ~400 outer dynein arms (odas) along a single microtubule and a more or less similar number of inner dynein arms (idas). Although the odas are thought all to be identical in isoform composition and equivalent in structure, there are several different types of idas per spoke group repeat. The pattern of activity of this ensemble of dynein arms is crucial for an understanding of ciliary motion, and, as a first step in empowering the computer model, my laboratory has been considering just what this pattern of activity is.
Although there is normally coupling between oda and ida activity
in an axoneme (and we don’t entirely understand how this coupling
works), we have learned in the last decade that, fundamentally, the
odas and idas control different aspects of the ciliary beat. This was
first shown in the results of Brokaw and Kamiya (11)
on a
series of Chlamydomonas mutants. Brokaw and Kamiya
simultaneously recorded the ciliary stroke of the beating organelle and
the frequency of beat—which is ~60 Hz in wt cilia. It is important
to realize that this means that the pattern of arm activity that
produces the effective and recovery stroke of the cilium repeats every
16 ms—that is, the control processes that we are studying are many
orders of magnitude faster than the rate of fast axonal transport,
chromosome movement, or ameboid movement. What Brokaw and Kamiya found
was that when the odas were nonfunctional, the beat form was nearly
normal, but the beat frequency was greatly altered to 20–30 Hz. In
contrast, when the idas are mutant, beat frequency was near normal, but
beat form changed drastically. Therefore, to a first approximation the
odas regulate beat frequency while the idas regulate beat form.
Consider what happens when with no change in beat form, beat frequency
in an axoneme increases from 16.7 Hz to 25 Hz—well within the
physiological range of response of many cilia. Because there is no
change in beat form, the amplitude of the bend produced during the
effective stroke is the same at both the slower and the faster
frequency. The sliding microtubule model implies that the amount of
sliding (
lE) during the effective
stroke between adjacent doublets (N and N+1) in the axoneme is
directly proportional to the amplitude of the bend that develops. This
value, measured experimentally, is surprisingly small—0.1 µm.
Remember, however, that the time taken for the effective stroke
(tE) is also small—about one-quarter of the
total beat period. At 16.7 Hz, tE = 15 ms, while
at 25 Hz, the beat period is only 40 ms and tE =
10 ms. The critical parameter that changes then is the velocity of
microtubule sliding
(lE/tE)
produced by oda activity. The conclusion from this calculation is that
beat frequency increases because the sliding velocity produced by oda
activity increases (12)
. The question then becomes how the
sliding velocity produced by an oda is controlled—or, more
specifically, in what way is the mechanochemistry of an oda changed so
that the oda can operate more quickly? To answer this question, we have
turned to a specific oda—22S dynein of the ciliate
Paramecium. In P., isolated 22S dynein is a
bouquet structure with three (DHC) heads 
ß
and several
intermediate and light chains. Inner arm dynein by contrast breaks down
into six or so single-headed molecules that run at 14S.
Hennesey et al. (13)
have shown that beat frequency in
P., as in many other cilia, increases when there is an
increase in cAMP around the axoneme, which is easily monitored as an
increase in forward swimming speed in detergent permeabilized cells
(14
, 15)
. After treatment with cAMP in the 10–100 µM
range, the detergent treated population reactivated with ATP and
monitored by automated computer tracking, swims faster than the
control. Average swimming speed rises from ~175 µm/s (about half
body length displacement per s) to ~350 µm/s. Some cells swim much
faster.
When P. are treated with cAMP in the presence of
nonreactivating concentrations of -S ATP and then the S-ATP and
cAMP are sufficiently diluted before the cells are reactivated, the
swimming rate still goes up, suggesting that cAMP works by
phosphorylating (thiophosphorylating) some axonemal protein. However,
when we treat with cAMP in the presence of 10-4M
Ca2+ (pCa4), swimming speed—and consequently
beat frequency—are unaltered versus the control group. The protein
that is phosphorylated in the detergent-treated axoneme in the presence
of cAMP and ATP, but not in the presence of cAMP, ATP, and
Ca2+, is p29 (15
, 16)
. Labeled p29
associates with the 22S odas but not the 14S idas. Barkalow et al.
(16)
showed that this association favored a specific DHC
() and was stoichiometric, suggesting that p29 is an oda
cAMP-regulated, Ca2+-sensitive dynein light
chain. To summarize, in P., faster beat frequency is
reflected by faster swimming. This is correlated with the
phosphorylation of p29, which functions as an oda regulatory LC. The
hypothesis proposed is that this is a causal relationship because
phosphorylation of p29 increases microtubule sliding velocity within
the axoneme.
This proposal has been tested directly using in vitro
microtubule translocation assays. The 22S dynein to be tested was as
used as a substratum on a glass slide. Microtubules polymerized from
purified tubulin were added to the chamber and ATP was added;
microtubule movement was recorded and velocity calculated. The results
have been summarized in Hamasaki et al. (17
, 18)
. In each
case, 22S dynein with phosphorylated p29 produced faster sliding that
matched the original speedup of swimming speed. Therefore, we conclude
that beat frequency in P. increases when the p29 associated
with odas is phosphorylated, because p29 phosphorylation changes the
Vmax of the dynein mechanochemical
cycle to increase the rate of microtubule sliding within the axoneme
(17
, 18)
. Win Sale et al. (19)
have shown
that cAMP-dependent phosphorylation of a 138 kDa protein associated
with I1 of the inner arm also controls sliding velocity produced by the
idas in Chlamydomonas. However, here dephosphorylation
produces faster sliding. Perhaps this difference is a manifestation of
the mechanisms of oda and ida coupling.
Recall that the displacement between doublets during an effective
stroke (lE) is only ~100 nm; if
an outer dynein arm step is 8 nm, only ~12 steps out of 400 odas
would be needed to produce this displacement. An oda step of 16 nm,
which is geometrically feasible, would require even fewer steps. The
force produced by one oda is ~6 pN, sufficient to move a 10 µm long
microtubule weakly held against a substratum, and the timing of a step
is also reasonable. It follows that the odas that are active in
producing sliding during a single beat are probably only a small
percentage of the total odas on a doublet, which suggests that a
stochastic process operates to produce sliding and bending during the
effective stroke. In this model, active odas are random in timing and
position along the microtubule. Whether this type of force generation
is consistent with localized bending during the effective stroke and
how phosphorylation of an active arm could speed up a stochastic
process are questions that we are presently trying to address using
modeling techniques. Three-dimensional structural modeling has led to
these interesting current biophysical questions about mechanisms of
ciliary beat. Keith Porter knew the beginnings of this story; he would
have appreciated these developments.
|
REFERENCES |
|---|
|
TOPINTRODUCTIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. A. Zariwala, M. W. Leigh, F. Ceppa, M. P. Kennedy, P. G. Noone, J. L. Carson, M. J. Hazucha, A. Lori, J. Horvath, H. Olbrich, et al. Mutations of DNAI1 in Primary Ciliary Dyskinesia: Evidence of Founder Effect in a Common Mutation Am. J. Respir. Crit. Care Med., October 15, 2006; 174(8): 858 - 866. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
