 |
INTRODUCTION |
Little is known about how a set of undifferentiated cells can
break symmetry and differentiate into distinct cell types. One of the
simplest systems that exhibits this behavior is the eukaryote Dictyostelium discoideum. Dictyostelium lives as a single
cell on soil surfaces that eats bacteria and divides by fission. When the cells overgrow an area and starve, they form an aggregate of
~2 × 104 cells, which develops into a fruiting body
consisting of two cell types: an approximately 2-mm column of stalk
cells supporting a mass of spore cells.
The initial differentiation of Dictyostelium cells occurs by
a combination of asymmetric cell division and a musical chairs mechanism based on the cell cycle (1-8). At the time of starvation, each pair of sister cells in S or early G2 phase
differentiates into a prestalk cell and a null cell (a cell type that
initially expresses neither prespore nor prestalk markers). Each pair
of sister cells in the late G2 or M phase differentiates
into a prespore cell and a null cell (Dictyostelium has an
undetectable G1 phase; Ref. 9). This mechanism regulates
only initial differentiation. The eventual fate of the cell is still
plastic, and a variety of factors such as adenosine, ammonia, a
chlorinated hydrocarbon called differentiation-inducing factor, and
oxygen can change the final fate (10-17).
Using shotgun antisense (18), we identified a gene involved in the
initial differentiation of Dictyostelium cells that we called rtoA. rtoA codes for a 39.8-kDa protein,
with 11 perfect repeats of a 10-amino acid-long serine-rich acidic
sequence, a possible N-terminal transmembrane domain, and a possible
ATP/GTP-binding domain. The gene disruption mutant of rtoA
has a high prestalk:prespore ratio while maintaining a normal cell
cycle. However, prestalk and prespore cells from rtoA
disruptants originate from cells in any phase of the cell cycle at
starvation. As in wild-type cells, the sisters of the differentiated
cells are all null in the rtoA mutant. These results suggest
that RtoA is not involved in the asymmetric cell division mechanism or
cell cycle progression but is involved in a process that varies during
the cell cycle and can be monitored at starvation to select initial
cell type (19).
One parameter that varies during the cell cycle and affects cell type
choice is cytosolic pH (11, 20, 21). In synchronized populations, cells
in M and S phases have a high pH (and tend to become prestalk), whereas
cells in mid and late G2 have a low pH (and tend to become
prespore) (22). Treating cells with bases or acids alters their initial
differentiation toward becoming prestalk or prespore, respectively (22,
23). By examining the effect of a variety of proton pump inhibitors on
the plasma membrane H+-ATPase and cell fate, Gross et
al. (24) hypothesized that cytosolic pH is linked to the pH of
intracellular vesicles and that it is the pH of these vesicles that
affects cell type choice. One set of vesicles that contain these pumps
belongs to the pinocytosis/endosome/lysosome system (25-29). During
endocytosis, endosomal vesicles merge with lysosomes and acidify. The
ingested material appears to pass through nine or more compartments
before being excreted (30). This process depends upon proper vesicle fusion.
In this paper, we show that there is a cell cycle- and sister
cell-dependent expression of RtoA in the vegetative cell
population. We find that a portion of RtoA has a random coil structure
and is able to induce fusion of artificial membranes. We demonstrate that cells lacking RtoA are defective in the fusion of endosomal vesicles and correlate this with defects in endosomal vesicle and
cytosolic pH regulation. We hypothesize that the cell cycle- and sister
cell-specific expression of RtoA imposes a cell cycle- and sister
cell-specific modulation of endosomal vesicle processing and cell
physiology, which in turn results in a cell cycle- and sister
cell-specific modulation of initial cell type choice.
| |
EXPERIMENTAL PROCEDURES |
Cell Culture--
D. discoideum Ax4 wild-type and
rtoA cells (19) were grown as described previously (31),
with the exception that in the HL5 growth medium the peptone was a mix
of 7.15 g of bacterial peptone (Oxoid Limited, Basingstoke,
Hampshire, UK) and 7.15 g of BBL Thiotone E peptone (Becton
Dickinson, Cockeysville, MD)/liter, and the HL5 was supplemented, after
autoclaving, with 20 µg/liter biotin, 5 µg/liter vitamin B12, 200 µg/liter folic acid, 400 µg/liter lipoic acid, 500 µg/liter
riboflavin, and 600 µg/liter thiamine. A mixture of 0.3 g/liter
streptomycin sulfate and 0.1 g/liter ampicillin was used as an antibiotic.
Antibody Production and Western Blots--
The synthetic peptide
SSGSSNSGSESSSDSGSSSDGKTT was conjugated to keyhole limpet hemocyanin,
and this was then used to immunize a rabbit at Biosynthesis Inc.
(Lewisville, TX). Serum was collected 2 weeks after the fifth injection
and was purified using an E-Z-Sep kit (Amersham Pharmacia Biotech).
Cell fractionation and Western blots were done as described by Brock
et al. (32).
Preparation of Fragments of Recombinant RtoA--
The reverse
transcriptase polymerase chain reaction was used to generate a DNA
fragment for fusion protein production. To generate cDNA for the
serine-rich region of RtoA (S fragment) coding for amino acids
52-271, the primer 5'-GCCTCGAGTGATTCAGATCCAGAG TTG-3' was used. The
above primer as well as 5'-GCAAGCTTCTTCAATTGGCTCATCTAG-3' were used to
amplify this DNA. The purified product was ligated into the expression
vector pET23a (Novagen, Madison, WI) using the created
HindIII and XhoI sites. Fusion protein (S
fragment) was induced and purified using a nickel-agarose column
(Pierce) following the manufacturer's directions. After elution with a series of increasing imidazole concentrations, fractions containing the
fusion protein were pooled and used for assays. These fractions consisted of >85% fusion protein. A similar procedure was performed to generate a fusion protein encoding the putative nucleotide binding
domain (amino acids 295-376), except that the primers were
5'-GCCTCGAGATGGGTGACGTGGACC-3' and 5'-GCAAGCTTCTGATAGTGGATCCTCATC-3', respectively.
Overexpression of RtoA--
RtoA was expressed throughout the
cell cycle from the actin 15 promoter of a Dictyostelium
expression vector. The primers 5'-GGGGTACCGTTTGGAATTTATATAGAAAATGTGC-3'
and 5'-GCTCTAGAGATTTAATGGTGACGTGGACC-3' were used in a polymerase chain
reaction to amplify genomic DNA encoding the complete rtoA
sequence. The primers added Asp718 and XbaI sites
at the 5' and 3' ends, respectively, of the amplified region. Following
digestion with Asp718 and XbaI, the 1.6-kilobase polymerase chain reaction product was cloned in frame into the Asp718 and XbaI sites of the expression vector
pDXA-3H, and the construct was verified by DNA sequencing. The
construct was transformed into cells by electroporation, and freshly
cloned cells were used for all experiments.
Videomicroscopy--
Conditioned medium and PBM (20 mM KH2PO4, 10 µM
CaCl2, 1 mM MgCl2, pH 6.1, with
KOH) were prepared as described by Clay et al. (33).
Videomicroscopy, fixation of cells and immunofluorescence staining were
done as described by Wood et al. (19) with the following
modifications. After videotaping for approximately 18 h, the
growth medium was changed to 200 µl of 25% PBM in water, and 20 min
later cells were fixed by adding 200 µl of 1% glutaraldehyde in Z
buffer. After 3 min, the liquid was removed and replaced with 200 µl
of 3.7% formaldehyde in Z buffer; after 10 min this was removed and
replaced with 70% ethanol; the slide was then air-dried 10 min and
stored in a desiccator at 
20 °C. Alternatively, cells were fixed
with 0.6% glutaraldehyde in PBM for 5 min, then 3.7% formaldehyde in
PBM, then with a 1:2:1 mixture of 95% ethanol, water and PBM for 10 min, followed by 95% ethanol for 10 min, and then air drying as above.
The slides were thawed, incubated with 53 mM
NaBH4 (EM Sciences, Cherry Hill, NJ) for 5 min to reduce nonspecific fluorescence from the glutaraldehyde and then stained with
anti-RtoA antibody as described above.
Circular Dichroism and NMR--
For circular dichroism, 100 µl
of either the S fragment in elution buffer or elution buffer alone as a
control were concentrated and resuspended in 8 mM potassium
phosphate, pH 6.4. The volume was adjusted to 600 µl, and the
ellipticity from 190 to 300 nm was measured in a type 53 cuvette
(Starna, Atascadero, CA) with a 62ADS spectrophotometer (Aviv,
Lakewood, NJ) using a step size of 1 nm. After the spectroscopy the
samples were assayed for ability to cause an increase in the OD of a
vesicle suspension to verify that the S fragment but not the buffer
control caused vesicle fusion. Proton one-dimensional NMR was done with
a Bruker AMX 500 MHz spectrometer. Solvent suppression was achieved
using a binomial sequence.
Calcium and ATP Binding--
200 µl of protein or buffer was
mixed with 800 µl of buffer (10 mM NaPO4, pH
7.2, 50 mM KCl, 1% sucrose). This was placed into a
Spectra/Por MWCO: 3,500 dialysis membrane (Spectrum, Los Angeles, CA)
that was placed into 100 ml of the above buffer with either 50 µCi of
45Ca or [
-32P]ATP (NEN Life Science
Products) and allowed to reach equilibrium at 4 °C for 24 h.
Aliquots were removed, scintillation mixture was added, and
radioactivity was measured by scintillation counting.
Vesicle Aggregation--
Egg yolk phosphatidylcholine in
chloroform (Sigma) was evaporated under nitrogen and resuspended to 10 mg/ml in buffer (140 mM NaCl, 0.5 mM EDTA, 0.05 mM NaN3, 10 mM Tris, pH 7.2). To
make vesicles, the mixture was sonicated with an Ultrasonics W220 probe sonicator (Ultrasonics, Plainsville, NY) for 20 min on ice. The assay
was initiated when 0.5 ml of protein (0.4 mg/ml) or
PEG1 (0.5% or 1%) was mixed
with 0.25 ml of buffer and 0.25 ml of vesicles. The optical density at
360 nm was measured for 5 min after mixing.
Vesicle Imaging--
Vesicle suspensions were placed on a glass
slide, overlaid with a coverslip, and observed with unfiltered tungsten
light using a 1.4 NA 60× phase contrast objective on a Nikon Microphot
FX microscope. For negative staining, a 200 mesh copper grid with a
carbon type B support film (Ted Pella, Redding, CA) was discharged using a Zerostat 3 static discharge gun, and a drop of vesicle suspension was placed on the grid. After 1 min the suspension was
wicked off the grid, and a drop of 2% ammonium molybdate (34) was
placed on the grid for 1 min and then wicked off. The grids were
allowed to dry and then examined at 80 or 100 kV accelerating voltage
in a JEOL JEM-2010 transmission electron microscope (JEOL USA, Inc.,
Peabody, MA). Images were recorded on Kodak SO-163 electron image film.
Lipid Mixing by Fluorescence Energy Transfer--
Vesicle fusion
was measured by observing fluorescence energy transfer based on the
method of Fung and Stryer (35). Donor vesicles were prepared by mixing
egg yolk phosphatidylcholine (Sigma) with NBD-PE (Molecular Probes,
Eugene, OR) in chloroform to 100 and 20 mg/ml, respectively. The lipids
were then evaporated under nitrogen and resuspended in buffer (10 mM MOPS, pH 7.0) to 10 and 2 mg/ml, respectively. Vesicles
were formed by incubating the lipids at 50 °C for 30 min followed by
probe sonication in an ice/water bath for 3 min. Acceptor vesicles were
created in the same manner except that rhodamine DHPE (Molecular
Probes, Eugene, OR) was used instead of NBD-PE. Buffer or 20 µl of
protein (0.4 mg/ml) was incubated with a mixture of 10 µl of each
donor and acceptor vesicles for 3 h at room temperature. The
samples were then diluted to 1 ml with buffer and excited at 463 nm
(excitation maximum of NBD-PE), whereas the fluorescence was scanned
from 500 to 600 nm (the emission maxima for NBD-PE and rhodamine DHPE are 534 and 590, respectively). Transfer efficiency E was
calculated using E = 1 F/Fo (35), where F is the
donor fluorescence of the sample containing both donor and acceptor
vesicles and Fo is the fluorescence of a sample
containing only donor vesicles. Alternatively, vesicles were prepared
with egg yolk L-
-phosphatidylcholine obtained from
Avanti Polar Lipids (Alabaster, AL). One set of vesicles contained 0.8 mol % each of donor and acceptor fluorescent lipids, whereas a second
set of vesicles contained no fluorescent lipids. After mixing the two
sets of vesicles in the presence or absence of S fragment or PEG,
transfer efficiency was calculated as E = 1 F/Fo where F is the
fluorescence of mixed vesicles at 534 nm with 463 nm excitation and
Fo is the fluorescence in the presence of 0.1%
Triton X-100. A standard curve was used to correlate transfer
efficiency with the percentage of mixing.
Vesicle Contents Mixing and Leakage--
The effect of the S
fragment of RtoA on small unilamellar vesicle contents mixing was
assayed with one set of vesicles containing TbCl3 and
another set of vesicles containing dipicolinic acid following Lentz
et al. (36) omitting the resuspension of lipid in
cyclohexane, resuspending lipid in buffer to 10 rather than 15 mg/ml,
and omitting the ultracentrifugation step. The total final lipid
concentration in the reactions was 0.2 mM. Vesicle mixing
was monitored by observing the formation of the fluorescent TbCl3/dipicolinic acid complex over the course of 45 min.
The reaction was performed in a buffer containing Ca2+ and
EDTA, which destroy the fluorescent TbCl3/dipicolinic acid complex. To verify that any increase in fluorescence was due to vesicle
contents mixing, detergent was added to release the vesicle contents,
and we observed an essentially complete loss of fluorescence. The
effect of the S fragment of RtoA on small unilamellar vesicle contents
leakage was similarly assayed with vesicles containing a mixture of
TbCl3 and dipicolinic acid (36).
Light Microscopy--
To monitor pinocytosis, 200 µl of cells
at a density of 1 × 105 cells/ml in HL5 were placed
onto an 8-well coverglass bottom slide (Type 136439, Nunc, Naperville,
IL) and allowed to settle for 10 min. The medium was then gently
changed to 2 mg/ml RITC-dextran (RD) (Sigma) in HL5 and then after 10 min changed to HL5. In some experiments, cells were then examined
in situ. In other experiments, after 80 min lysosomes were
stained by adding Lysosensor Green DND-189 (Molecular Probes) to 1 µM. After 10 min, the medium was changed to either PBM or
HL5 for an additional 10 min. Cells were then fixed for 3 min in HL5
containing 1% formaldehyde, after which the medium was changed to HL5.
Images were taken with a Nikon Diaphot inverted microscope using a 60×
1.4 NA phase/fluorescence objective lens. A Photometrics Imagepoint
cooled CCD camera operating at an exposure of 16× and gain of 24 db
supplied a video image to an IBM clone computer and a direct monitor.
Nuclei were stained with 4',-6-diamidino-2-phenylindole following the
work of Wood et al. (19). Staining of contractile vacuoles
with FM4-64 was done following the methods of Heuser et al.
(37). Staining of calmodulin in fixed and permeabilized cells was done
following Zhu and Clarke (38), omitting the agarose overlay step.
Mitochondria were stained with MitoTracker Red CM-H2XRos or MitoTracker
Red CMXRos (Molecular Probes) following the manufacturer's directions.
Electron Microscopy--
EM grade 25% glutaraldehyde, osmium
tetroxide, propylene oxide, Eponate 12 embedding medium, and BEEM
capsules were obtained from Ted Pella, Inc. (Redding, CA). Cells grown
in HL5 to 2 × 106 cells/ml were pelleted in 15-ml
aliquots at 300 × g to yield 0.2 ml of packed cells.
Pellets were resuspended in 10 ml of PBM and washed two more times with
PBM. Each pellet was resuspended in 2 ml of ice-cold 2.5%
glutaraldehyde in PBM and held for 2 h on ice. After two 10-min
washes in PBM, each pellet was resuspended in 2 ml of 1% osmium
tetroxide in water and held 1 h at room temperature. After two
10-min washes in distilled water, each pellet was dehydrated in 10-min
steps in 10 ml each of 50, 70, and 95% ethanol, two 10-min steps each
in 1-1.5 ml of 100% ethanol, and two 15-min steps in propylene oxide.
Dehydration in 100% ethanol and propylene oxide and all subsequent
steps were carried out in polypropylene Eppendorf centrifuge tubes with
centrifugation at 10,000 × g. Final pellets were
resuspended in 50:50 embedding medium:propylene oxide overnight at room
temperature. After warming the samples and fresh Eponate 12 to
60 °C, cells were pelleted and resuspended in Eponate 12, formulated
for medium hardness. Resuspended cells were distributed into warmed
BEEM capsules that were centrifuged for 20 min at 700 × g. Specimens were cured overnight at 60 °C and sectioned
with glass knives on a DuPont/Sorvall MT2-B ultramicrotome. Pale to
medium gold sections were picked up on 200 mesh copper grids with
carbon type-B substrate (Pella catalog number 01811), and double
stained with uranyl acetate and Reynold's lead citrate according to
Lewis and Knight (39) with the exception that washes were done by
floating the grid on a drop of water rather than placing it in a stream
of water. Sections were examined in the electron microscope as
described above.
Pulse-Chase Labeling with Colloidal Gold--
200 µl of bovine
serum albumin/colloidal gold conjugate (5 nm diameter; Ted Pella, Inc.)
was de-salted using a Microcon-50 microconcentrator (Amicon, Beverly,
MA) and resuspended in 1 ml of PBM. 30 ml each of vegetative
rtoA and parental cells were pelleted as above, washed once
in 15 ml of PBM, and then resuspended in 15 ml PBM with 0.5 ml of
desalted gold conjugate suspension. 5 ml of each cell suspension was
removed and centrifuged 5 min after adding the colloidal gold, and the
pellets were resuspended in ice-cold 2.5% glutaraldehyde and processed
for electron microscopy as described above. The remaining suspensions
were centrifuged at the same time, pellets were resuspended in HL-5
medium, and the suspensions were incubated in shaking culture without
label. Cells were then removed and processed for electron microscopy 20 and 60 min after resuspension in HL-5.
Endocytosis and Exocytosis Assays--
Endocytosis was assayed
as described by Buczynski et al. (40). Fluid phase
exocytosis was measured using FITC-dextran (FD) (Sigma) as described
previously (41) with slight modifications. Cells grown to 2 × 106 cells/ml were harvested and resuspended in HL5
containing FD at 2 mg/ml. The cells were allowed to internalize FD for
3 h then collected and resuspended in fresh HL5. At the indicated
times, 1 ml of cells was collected by centrifugation and washed once in
5 mM glycine-NaOH (pH 8.5) containing 100 mM
sucrose. The cells were then lysed in 1 ml of the same buffer
containing 0.2% Triton X-100. The fluorescence of the lysate was
measured in an Aminco Bowman Series 2 Luminescence Spectrometer.
Excitation and emission wavelengths for FITC were 495 and 520, respectively.
Vesicle Staining--
To examine the relationship between cell
lineage and endocytic vesicles cells from mid-log phase shaking
cultures were diluted to 6 × 103 cells/ml, and 200 µl was placed in the well of an 8-well coverslip bottom chamber slide
(Type 136439, Nunc, Naperville, IL). The cells were observed and
videotaped as described above using a 4 × 0.13 NA phase
objective. After approximately 18 h, under red illumination 20 µl of 20 mg/ml fluorescein isothiocyanate, 70 kDa of dextran (Sigma)
in PBM was gently added as described by Klein and Sartre (42). After 10 min, the medium was changed to fresh HL5 with the dual channel
peristaltic pump. 60 min after adding the FITC-dextran, the cells were
observed with a 60 × 1.4 NA oil immersion phase/fluorescence
objective lens with a fluorescein filter cube. Phase and
epifluorescence images of the cells were recorded with the camera
exposure turned up to 16×. From the appearance of the FITC-dextran
fluorescence, cells were scored as having large vacuoles, multiple
small vacuoles, or no detectable ingested FITC-dextran. The cells were
then fixed and stained for RtoA as described above. To examine
contractile vacuoles, a field of cells in HL5 growing in monolayer
culture in a coverslip bottom slide was observed with a 10× objective.
After identifying dividing cells, the cells were stained with FM4-64
according to Heuser et al. (37), and the sister pair was
observed with a 60 × 1.4 NA objective, using epifluorescence with
a rhodamine filter cube.
Endocytic pH Measurements--
Following the procedures of Aubry
et al. (43), cells at 5 × 106 cells/ml
were incubated in 2 mg/ml FITC-dextran in HL5 for 10 min. They were
then collected by centrifugation at 4 °C, washed once in ice-cold
HL5, resuspended in HL5 to 5 × 107 cells/ml, and
mixed. At time points, 134-µl aliquots were removed, centrifuged at
300 × g for 20 s and resuspended in 1 ml of 50 mM MES-NaOH (pH 6.5) containing 10 mM KCl.
Cellular fluorescence was determined using emission at 520 nm and
excitation at both 450 and 495 nm. The pH of endocytic vesicles was
determined by calculating the ratio of the former value to the latter
and comparing the results to a standard curve. This curve was
constructed using cells bearing ingested probe permeabilized to ions
with 0.1% digitonin and suspended in the above buffer adjusted to
different pH values with acetic acid or NaOH.
BCECF Staining to Measure Cytosolic pH--
200 µl of 1 × 104 cells/ml in HL5 added to the well of an 8-well
slide. The settled cells were then followed by time lapse videomicroscopy using a Nikon Diaphot inverted microscope with a
10 × 0.5 NA phase/fluorescence objective lens. A Photometrics Imagepoint cooled CCD camera operating at an exposure of 4× and gain
of 21 db supplied a video image to a Panasonic 6740 time lapse video
recorder operating in 480-h mode. After approximately 16 h, a dual
channel peristaltic pump operating at 75 µl/min was used to gently
change the media over the course of 10 min to 20 M (20 mM MES, pH 6.1, with KOH). For a dye stock, 50 µg of
BCECF AM (Molecular Probes) was dissolved in 50 µl of anhydrous
methyl sulfoxide (Aldrich, Milwaukee, WI). 1 µl of the BCECF AM stock was added to 100 µl of 20 M, and 40 µl of this mixture
was gently added to the well with a Hamilton syringe. After 20 min, the
medium was changed to 20 M as described above. The cells
were then observed by epifluorescence using a Chroma Technology
(Brattleboro, VT) 71001A BCECF filter cube set. Continued videotaping
of cells after this type of pH measurement verified that at this light
intensity we were not changing any observable aspect of cell morphology or movement (both translocation and pseudopod extension).
The fluorescence images were transferred from the CCD camera to a
computer and analysis performed using NIH Image 1.61 (National Institutes of Health). Each data set consisted of three images; phase,
green filter, and purple filter. Using the phase image as a reference,
individual cells were identified on each filter image, measured, and
labeled. A background measurement was also taken adjacent to each cell
counted. The cell brightness value was obtained by subtracting each
cells background value from its density value. A color value was then
obtained by dividing the green brightness of each cell by its purple
brightness. Cell pH values were obtained by correlating the color value
to a calibration curve generated by adding dye stock to 20 M that had been adjusted to different pH values, and small
droplets of this mixture were imaged as described above.
| |
RESULTS |
RtoA Is Associated with Both Small Organelles and the
Cytosol--
Rabbit polyclonal antibodies were generated against a
synthetic peptide corresponding to one and a half copies of the
11-amino acid-long serine-rich repeated motif of RtoA. Western blot
analysis of cells starved for varying lengths of time showed that the
anti-RtoA antibodies stain a 40-kDa band that is present in vegetative
and developing wild-type cells. No staining was seen in proteins from rtoA cells or with preimmune sera in either cell type (data
not shown). This suggests that the rtoA cells contain very
little of the serine-rich repeat region of RtoA protein (the portion of
the protein the antibody recognizes).
To measure the average concentration of RtoA in vegetative cells, we
expressed a recombinant fragment of RtoA consisting of 6 histidines
preceded by the entire serine-rich motif (the S fragment). A Western
blot containing known amounts of the S fragment, and protein from 1, 2.5, and 5 × 106 cells was stained with anti-RtoA. We
found that the staining intensity of the 40-kDa RtoA in 5 × 106 cells was roughly equivalent to the staining intensity
of 1.8 µg of the 27-kDa S fragment. Correcting for the difference in molecular masses, we found that there is an average of 0.59 ± 0.03 pg of RtoA/cell, or roughly 8.9 × 106 molecules
of RtoA/cell.
To determine the subcellular localization of RtoA, vegetative cells
were lysed and fractionated by centrifugation. The fractions were
electrophoresed on a SDS-polyacrylamide gel, and Western blots were
stained with anti-RtoA antibodies. After a low speed spin to remove
nuclei, RtoA was present in the supernatant (Fig. 1). This supernatant was then
recentrifuged to pellet large organelles, and after this centrifugation
RtoA was still present in the supernatant. This second supernatant was
then fractionated by centrifugation at 6 × 106
g-min, and RtoA was present in both the supernatant and
pellet (Fig. 1). When the same fractions were stained with antibodies against the cytosolic protein SmlA (32), only the 6 × 106 g-min supernatant showed staining (data not
shown). These experiments suggest that RtoA resides mostly in the
cytosol with a percentage of the protein associated with small
organelles or vesicles.
View larger version (58K):
[in this window]
[in a new window]
|
Fig. 1.
Distribution of RtoA in subcellular
fractions. Vegetative DH1 cells were broken open with a Dounce
homogenizer and fractionated by centrifugation. A low speed spin was
used to pellet nuclei and unbroken cells. The supernatant from this
spin was centrifuged at a higher speed to pellet organelles. The
supernatant from the second spin was centrifuged again to pellet
microsomes. Samples of pellets (P) and supernatants
(S) from the three spins were electrophoresed on an
SDS-polyacrylamide gel. A Western blot of the gel was then stained with
anti-RtoA antibodies.
|
|
RtoA Levels Show That There Is a Heterogeneity in Vegetative
Cells--
Immunofluorescence with anti-RtoA antibodies showed
staining throughout vegetative and developing Ax4 cells, and very
little staining of rtoA cells (Fig.
2). Preimmune sera stained Ax4 and rtoA cells with an intensity identical to that of
rtoA cells stained with immune sera (data not shown).
Interestingly, the Ax4 cells showed a heterogeneity of staining
intensities. This was observed for both axenically grown cells and
cells grown on bacteria, and these results were obtained with a variety
of fixation techniques (data not shown). When slugs and fruiting bodies
were stained for RtoA, the prestalk regions had no detectable RtoA,
whereas there was a high level in the prespore regions (data not
shown).
View larger version (120K):
[in this window]
[in a new window]
|
Fig. 2.
Distribution of RtoA protein in vegetative
cells. Vegetative cells growing in shaking culture were allowed to
adhere to a coverslip and were then fixed and stained for RtoA by
indirect immunofluorescence. A, Ax4 cells; B,
rtoA cells. C and D are the
corresponding phase images. No staining was seen with preimmune
antibodies. The bar in D is 20 µm.
WT, wild type.
|
|
There Are Low Levels of RtoA in S and Early G2 Cells
and Higher Levels in One Sister of Each Pair in Late G2
Cells--
To determine whether the RtoA staining heterogeneity
correlates with any observable property of the cells, vegetative
wild-type cells were videotaped for approximately two cell generations
and then stained for RtoA. When the time difference between the time of
fixation and the time of the most recent cytokinesis was plotted for
these cells, an obvious grouping was observed. Using the observation that cytokinesis occurs 20 min after the beginning of S phase (9), a
plot of the cell cycle phase of the strongly RtoA-positive cells showed
that these were cells in the late three-quarters of the cell cycle
(Fig. 3). When the sisters of the
strongly RtoA-positive cells were followed, with two exceptions all
showed a very low level of RtoA staining. Cells in S and early
G2 had an intermediate level of RtoA staining, with only
slight differences between the observed levels in the sister pairs
(data not shown). A Northern blot of RNAs from synchronized cells
indicated that cells in mid to late G2 had higher levels of
rtoA mRNA than did cells in S and early G2
(data not shown).
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 3.
Distribution of the 30% of cells with the
highest RtoA brightness. Wild-type Ax4 cells were plated at low
density in growth medium in submerged culture and videotaped. After at
least 12 h the cells were fixed for immunofluorescence labeling
with antibodies RtoA. The videotape was reviewed, and the cell cycle
phase of each strongly stained cell was determined and mapped on a
diagram of the cell cycle.
|
|
To further elucidate the function of RtoA, we examined the effect of
overexpressing it. A recombinant gene was made using the actin 15 promoter to drive expression of the rtoA cDNA. Previous studies have shown that this promoter causes expression in all vegetative cells. Vector-alone transformants formed normal fruiting bodies (data not shown), whereas RtoA overexpressor transformants had
somewhat misshapen fruiting bodies (data not shown). Immunofluorescence showed that freshly isolated clones of these transformants had some
cells with much more RtoA than control cells (data not shown). Interestingly, as with wild-type and vector-alone transformants, the
RtoA overexpressor transformant cells showed a marked heterogeneity in
anti-RtoA staining. This suggested that there may be some activity that
either regulates RtoA expression at a post-transcriptional level or
degrades RtoA in some cells.
The Serine-rich Region of RtoA Is a Random Coil--
We were
puzzled by the possible function of RtoA. Dr. David Worcester of the
University of Missouri at Columbia suggested that the serine-rich
repeat sequence of RtoA would have a random coil structure and thus
behave like PEG, a known membrane fusiogen. To test the first part of
this hypothesis, we measured the molar ellipticity of the S fragment.
As shown in Fig. 4, the S fragment has a
spectrum characteristic of a polypeptide with a very high percentage of
random coil (44). Addition of 6 M urea only slightly affected the spectrum above 210 nm; below this the urea absorbed strongly, and we were unable to obtain a spectrum (data not shown). One-dimensional proton NMR of the S fragment showed peaks between 0 and
7.3 ppm, which are characteristic of protons in amino acid side chains.
There were no detectable peaks between 7.4 and 11 ppm, which would
characterize slowly exchanging amide protons of the polypeptide
backbone (Ref. 45 and data not shown). The absence of these peaks and
the circular dichroism spectrum suggest that most of the S fragment is
randomly coiled rather than structured.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4.
Circular dichroism spectrum of the 220-amino
acid-long serine-rich region of RtoA. This spectrum is essentially
indistinguishable from the spectra of known random coil proteins.
|
|
At physiological pH, the serine-rich region of RtoA (S fragment) would
have a slight negative charge. Because Ca2+ is involved in
many forms of vesicle fusion, and Ca2+ levels fluctuate
during the cell cycle (46), one possibility is that RtoA binds
Ca2+ directly. In addition, the carboxy third of RtoA
contains a potential nucleotide binding site (19). We expressed
portions of the RtoA protein in bacteria and measured binding of
45Ca2+ and [3H]ATP by equilibrium
dialysis. We found that for both portions of RtoA the
KD for Ca2+ was greater than 2 × 10
5 M1 and the KD for
ATP was greater than 2 × 104
M1, thus indicating that these portions of
RtoA do not bind these ligands with high affinity.
The Serine-rich Region of RtoA Catalyzes Vesicle Fusion in
Vitro--
Because the S fragment of RtoA appeared to have a random
coil structure like poly(ethylene glycol), we then examined whether this region of RtoA could cause vesicle-vesicle interactions. We first
monitored its ability to catalyze the aggregation of artificial
membranes. Vesicles were made from egg phosphatidylcholine and exposed
to the S fragment, PEG as a positive control and bovine serum albumin
(BSA) as a negative control. Aggregation was measured by an increase in
optical density at 360 nm. We found that addition of the S fragment
caused a 16-fold increase in the rate of aggregation over that of BSA
(Fig. 5A). In comparison, 0.5 and 1% PEG caused 7- and 56-fold increases in the rate of vesicle
aggregation, respectively. Optical (Fig. 5, B-D) and
electron (Fig. 5, E-G) microscopy of the fusion reactions
showed that both S fragment and PEG caused an increase in the observed
size of the vesicles.
View larger version (84K):
[in this window]
[in a new window]
|
Fig. 5.
The effect of S fragment on synthetic lipid
vesicles. A, the effect of the recombinant S fragment
of RtoA on the light scattering of synthetic lipid vesicles. Synthetic
vesicles were mixed with S fragment, BSA, or PEG, and the change in
absorbance at 360 nm was measured over the course of 1 min. An increase
in particle size in the suspension causes an increase in light
scattering. The bars are S.E.; the errors for BSA and 0.5%
PEG were smaller than the plot symbol. B-D, light
microscopy of vesicles. Synthetic phospholipid vesicles were treated
with buffer (B), recombinant S fragment of RtoA
(C), or 1% PEG (D). The bar in
D is 20 µm. E-G, electron microscopy of
vesicles. The same preparations of vesicles in B-D were
examined by negative stain electron microscopy. D, vesicles
treated with buffer; E, vesicles treated with recombinant S
fragment of RtoA; F, vesicles treated with 1% PEG. The
bar in G is 200 nm.
|
|
Because the increase in optical density observed in the previous
experiment could have been due to either vesicle fusion or vesicle
aggregation without fusion, we used fluorescence energy transfer to
directly measure membrane mixing. Two types of vesicles were prepared,
donor and acceptor. Donor vesicles were made from egg
phosphatidylcholine with 2% NBD-PE, and acceptor vesicles were made
from egg phosphatidylcholine with 2% rhodamine DHPE. When the two
types of vesicles fuse, the energy from the donor phospholipids
(NBD-PE) is transferred to the acceptor phospholipids (rhodamine DHPE).
Fusion of the two vesicles was measured as an increase in energy
transfer from the donor vesicles to the acceptor vesicles. We found
that buffer alone allowed 0.8% energy transfer (Fig.
6A). Addition of S fragment
increased the energy transfer to 11.8%, whereas the presence of 1%
PEG caused 8.4% energy transfer. As an alternative assay for membrane
fusion, we prepared vesicles as before except that one set was labeled
with a mixture of 0.8% donor and 0.8% acceptor probes, and the other
set of vesicles was unlabeled. Fusion was then measured as a dilution
of the probes, as indicated by a decrease in energy transfer. As shown
in Fig. 6B, the S fragment of RtoA induces a dilution of the
mixed probes. We observed similar results with higher concentrations of
the probe lipids and using unlabeled lipids from two different sources (data not shown). Thus by two different criteria, mixing of separate probe lipids and dilution of mixed probe lipids, our data suggest that
the S fragment of RtoA causes phospholipids of different vesicles to
mix with each other.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 6.
The effect of the recombinant S fragment of
RtoA on fluorescence energy transfer in lipid vesicles.
A, a suspension of donor vesicles containing one fluorescent
lipid was mixed with a suspension of acceptor vesicles containing a
second fluorescent lipid. This was incubated in the presence of buffer,
S fragment, or PEG for 3 h. The mixture was diluted with buffer,
and the fluorescence energy transfer efficiency from the first lipid to
the second lipid was determined. B, a suspension of vesicles
containing both donor and acceptor lipids was mixed with unlabeled
vesicles in the presence of buffer, S fragment, or PEG as above, and
fluorescence energy transfer was similarly measured, with a decrease in
transfer efficiency indicating that the fluorescent lipids had been
diluted because of membrane mixing. The bars are S.E.
|
|
To determine whether the S fragment-induced mixing of membrane lipids
is accompanied by a mixing of vesicle contents, we loaded TbCl3 into the aqueous interior of one set of vesicles and
dipicolinic acid into the interior of another set of vesicles. The
addition of TbCl3 to dipicolinic acid causes the formation
of a fluorescent complex, which can then be monitored with a
fluorescence spectrophotometer (47). As shown in Table
I, S fragment causes the aqueous contents of vesicles to mix. To determine whether this mixing was accompanied by
vesicle contents leakage, we examine the effect of S fragment on
vesicles loaded with a TbCl3/dipicolinic acid mixture. As
with the vesicle mixing assay, the buffer outside the vesicles
contained Ca2+ and EDTA, which destroy the
TbCl3/dipicolinic acid complex and abolish the
fluorescence. We found that S fragment caused no significant leakage of
vesicle contents (Table I). The above results suggest that the
serine-rich region of RtoA is sufficient to induce fusion of
phospholipid vesicles.
View this table:
[in this window]
[in a new window]
|
Table I
The effect of the S fragment of RtoA on the mixing of vesicle contents
and the leakage of vesicle contents
The S fragment concentration was 0.17 mg/ml and caused a roughly linear
increase of vesicle contents mixing with time. Values are means of
three experiments ± standard deviations.
|
|
Disruption of rtoA Inhibits Fusion and Excretion of Endocytic
Vesicles--
To determine whether the absence of RtoA affects the
appearance or distribution of intracellular organelles and to possibly identify the unfused vesicles, parental DH1 and rtoA cells
were stained with a variety of markers. Staining of the contractile vacuole in live cells with the dye FM4-64 (37) revealed no observable differences between DH1 and rtoA cells. This was true under
a variety of conditions (such as in growth medium or in starvation buffer). Similarly, when the contractile vacuoles in fixed and permeabilized cells were stained with anti-calmodulin (38), there was
no detectable difference. Staining of nuclei, mitochondria, and
lysosomes also revealed no detectable alteration in these organelles in
rtoA cells (data not shown). Thus, RtoA is not involved in
regulating the morphology of the contractile vacuole, nucleus, mitochondria, or lysosome.
To determine whether the absence of RtoA has any effect on
intracellular structure, thin sections of fixed cells were examined by
electron microscopy. The only consistent difference observed between
the parental and the rtoA cells was that the parental cells
had small vesicles that have apparently fused, whereas the rtoA cells had many vesicles that appeared to be pressed
against each other (Fig. 7, A
and B). In these cases the intervening membranes were
intact, suggesting that membrane fusion had not occurred. Similar
results were observed under a variety of fixation techniques. Thus,
RtoA appears to be required for the fusion of a subset of vesicles.
View larger version (146K):
[in this window]
[in a new window]
|
Fig. 7.
Ultrastructure of cells with a disrupted
rtoA gene. Vegetative cells growing in shaking
culture were fixed with glutaraldehyde and osmium, embedded in resin,
sectioned on a ultramicrotome, and then stained for electron
microscopy. The sections were then examined by electron microscopy. The
gross appearance of the cells was very similar; the only observable
difference was that one can see what appears to be vesicles fusing in
the wild-type (WT) cells (A), and similar
vesicles pushed up against one another but not fusing in the
rtoA cells (B). The bar in
B is 500 nm. Wild-type (C and D) and
rtoA (E) cells were allowed to ingest colloidal
gold-BSA for 5 min. The cells were harvested by centrifugation and
resuspended in growth medium. 60 min later the cells were fixed and
sectioned. N designates nuclei; arrows indicate
those vesicles that contain colloidal gold (as verified by examination
at high magnification). Heavy arrows in D
indicate two closely opposed vesicles. The bars in
C and E are 2 µm.
|
|
One set of vesicles that is known to fuse with each other in
Dictyostelium are endocytic/exocytic pathway vesicles. To
determine whether these are the vesicles that are defective in fusing
in the rtoA cells, we briefly fed cells colloidal gold. The
cells ingested the gold particles, and 20 min after the pulse we
observed gold particles in small vesicles in both rtoA and
parental cells (data not shown). 60 min after ingestion, many of the
parental cells had accumulated the gold in a small number of relatively large vesicles (Fig. 7, C and D). In the
rtoA cells, the gold particles were in numerous smaller
vesicles, which in some cases were closely opposed to each other
(heavy arrows, Fig. 7E). These data suggest that
cells lacking RtoA have a decreased rate of fusion of
endocytic/exocytic vesicles into larger vesicles.
To alternatively visualize the endosomal system we allowed both DH1 and
rtoA cells to internalize RD. 5 min after ingesting RD, both
cell lines showed a punctate localization of RD (Fig. 8, B and D)
representing multiple endocytic vesicles. 55 min later, the DH1 cells
showed a heterogeneity of RD distributions (Fig. 8F). Some
cells had no detectable RD, suggesting that the RD had been secreted.
Other cells had a punctate distribution, indicative of RD in the
endosomal system. The final group of cells had localized most of the RD
into a large vesicle, suggesting deposition of the RD into the lysosome
or post-lysosome. Similar results were obtained using Ax4 cells and FD.
At the same time, almost all of the rtoA cells showed a
punctate distribution (Fig. 8H), which indicates that the RD
is stuck in the endosomal system. To determine whether there is a
correlation between cell cycle phase and the appearance of the
endosomal vesicles in the vegetative wild-type cells, cells were
videotaped to obtain lineages, and the cells were then fed a pulse of
FD. 1 h after the pulse, the appearance of ingested FD was
examined at high magnification, and the cells were then fixed and
stained for RtoA. In a polar plot of the data (Fig.
9), r (the distance from the
center) represents the approximate amount of RtoA staining intensity,
and 
represents cell cycle phase. During S and early G2
phase, cells have an intermediate amount or RtoA staining. For the
cells we observed in this sector, one sister had multiple small
vesicles, whereas the other sister had either a large vesicle or no
detectable vesicles. For sister cells in mid and late G2,
one sister had a low level of RtoA, and its sister had a high level of
RtoA. Almost all of the sister cells had a difference in the appearance
of their endocytic vesicles, and all of the sisters that had a low
level of RtoA had either no detectable vesicles or vesicles with a
punctate appearance. The above data suggest that the mid and late
G2 wild-type cells with a low level of RtoA, as well as the
rtoA cells themselves, tend not to have large endocytic
vesicles.
View larger version (94K):
[in this window]
[in a new window]
|
Fig. 8.
Endocytic vesicles in wild-type and
rtoA cells. Wild-type (WT;
A, B, E, and F) and
rtoA (C, D, G, and
H) cells were pulsed for 10 min with FITC dextran and chased
in HL5. After five (A-D) or 60 (E-H) minutes of
chase, fields of cells were imaged. Phase images are A,
C, E, and G. The corresponding
fluorescence images are shown in B, D,
F, and H. The bar in H is
10 µm.
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 9.
The appearance of vesicles in sister
cells. Cells were grown at low density in monolayer culture and
videotaped to obtain lineages. Cells were then given a pulse of FD and
chased in growth medium for 1 h, and the appearance of vesicles
was scored as in Fig. 8. The cells were then fixed and stained for
RtoA. The data show the phase of the cell cycle that pairs of sister
cells happened to be in, with the radial distance from the center
indicating the approximate RtoA staining intensity, and the symbol
showing the appearance of the ingested FD. Circles indicate
the presence of a large FD-containing vesicle, squares
indicate the presence of numerous small FD-containing vesicles,
and × indicates that no FD could be detected in the cell.
|
|
We then investigated the possible source of the difference between the
two sister cells. We videotaped cells to obtain lineages and then
stained for a variety of markers. We observed that although some
sisters get slightly unequal amounts of glycogen, most sister pairs get
roughly equal amounts of glycogen. We observed that sister cells have
no consistent difference in the amount or appearance of their
contractile vacuoles. Staining of nuclei, mitochondria, and lysosomes
also revealed no detectable cell-cell heterogeneity except for the
occasional mitotic figure seen staining for nuclear DNA (data not shown).
A drastic difference between control and rtoA cells became
apparent when the cells were osmotically stressed. Cells were allowed to ingest FD and were then harvested and resuspended in a low salt
buffer (PBM). After 10 min in buffer, the DH1 cells showed a rapid
decrease in the levels of RD in internal vesicles, leaving little or no
staining, whereas in the rtoA cells the levels of RD remain
high (data not shown). The poor exocytosis of endocytic vesicle
contents when cells were osmotically stressed suggested the possibility
that rtoA cells might also have a defect in exocytosis under
normal conditions. To examine the rate of efflux in DH1 and
rtoA cells, we measured the release of FD from exponentially growing cells in shaking culture. Cells were loaded with FD for 3 h, washed, and released into fresh growth medium. At the indicated times, samples were taken and processed for fluorometry. After 90 min,
approximately 30% of the internalized FD was retained in DH1 cells,
whereas rtoA cells retained 60% of the RD (Fig. 10A). This indicated a block
in exocytosis. A similar assay was performed to determine the rate of
FD uptake. Both DH1 and rtoA cells internalized FD at the
same rate (data not shown). The above data suggest that rtoA
cells have normal endocytosis but poor exocytosis.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 10.
A, exocytosis rates in wild-type and
rtoA cells. Cells were allowed to internalize FITC dextran
for 3 h, washed, and resuspended in fresh HL5. At the indicated
times, cells were collected by centrifugation, washed once, and lysed
in wash buffer containing Triton X-100. The percentage of FITC dextran
retained within the cells was calculated by dividing the intracellular
fluorescence at the indicated time by the fluorescence measurement at 0 min. The bars are S.E. B, endosomal pH in
wild-type and rtoA cells. Cells were fed FITC-dextran for 10 min, collected, and resuspended in fresh HL5. At the indicated times,
aliquots of cells were removed, centrifuged, and resuspended in assay
buffer, and fluorescence was measured using excitation at 450 and 490 nm. The pH values were obtained from a standard curve of the FITC
dextran fluorescence ratio as a function of pH. The bars are
S.E. A t test on the data showed that the differences
between parental and rtoA vesicle pH levels are significant
to p < 0.05 (5 min); <0.2 (10 min); < 0.025 15 min); < 0.05 (20 min); < 0.1 (25-75 min); and < 0.15 (95 min).
|
|
To discover other possible defects in the endosomal system, we examined
the pH evolution of endosomes in both DH1 and rtoA cells. As
vesicles progress from endosomes to lysosomes to post-lysosomes, they
undergo distinctive pH changes. During the first 20 min, the pH drops
from that of the medium to 5.0 and then rises over the next 40 min to
plateau at a value of 6.2 (43). We observed this progression in the DH1
parental cells (Fig. 10B). In the rtoA cells,
vesicles appeared to be acidified much more quickly and acidified to a
lower pH (Fig. 10B). The neutralization of these vesicles
followed a similar time course to the DH1 cells and the wild-type cells
studied by Aubry et al. (43) but still remained more acidic
at all time points examined (Fig. 10B). Therefore, rtoA cells are not only defective in exocytosis but are also
unable to properly regulate the pH of their endocytic vesicles.
Cytosolic pH Is Altered in rtoA Cells and Differs in the Two Sister
Cells from a Cell Division--
Gross et al. (24) predicted
that cytosolic pH and vesicular pH should be linked. To determine
whether rtoA cells have abnormal cytosolic pH levels
possibly because of having endocytic vesicles with abnormal pH levels,
we used BCECF AM to measure cytosolic pH levels by ratiometric
fluorescence imaging. As described previously, wild-type cells have a
distribution of cytosolic pH levels (Refs. 11, 20, 21, and 23 and Fig.
11). When rtoA cells were
similarly stained, they had a similar range of pH levels, but the mean
was skewed toward a higher pH (Fig. 11).
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 11.
Distribution of cytosolic pH levels.
Vegetative wild-type (WT) and rtoA cells were
allowed to grow in submerged monolayer culture. BCECF AM was added, and
after gentle washing the cells were observed by epifluorescence at two
wavelengths. After generating a standard curve under the same
conditions, the ratio of the two fluorescence intensities was used to
determine the cytosolic pH. The data were then binned and plotted.
Open bars are wild-type cells, and shaded bars
are rtoA cells. The figure shows data from a representative
experiment.
|
|
Both RtoA levels and the initial choice of cell type vary during the
cell cycle (Fig. 3 and Ref. 48). On closer examination, we found that
both RtoA and cell type choice are different in the two sister cells
from each division. Previous work indicated that there is a cell
cycle-dependent variation of cytosolic pH in growing
Dictyostelium cells and that cytosolic pH may affect cell
differentiation (22). We thus examined whether cytosolic pH is also
different in the two sister cells from a division. Fields of cells
growing in monolayer culture were videotaped, and then, while
continuously videotaping, the cells were stained for cytosolic pH with
BCECF AM. As shown in Fig. 12, there is
a heterogeneity of cytosolic pH levels. We found that the distribution of cytosolic pH levels changes depending on the cell density (Fig. 12).
For cells starved at densities corresponding to the density used for
previous experiments (48), the cells with the highest 10% of pH levels
were distributed unevenly in the cell cycle (Fig. 12). The high pH
cells were predominantly clustered in late G2, M, S, and
early G2 with only a few high pH cells present in mid G2. The sisters of these cells all had low cytosolic pH
levels. This distribution is different from that of the cells, which
would become prestalk when starved. However, in cells lacking RtoA, the
high pH cells are distributed evenly throughout the cell cycle (Fig.
12). This suggests that cells lacking RtoA are unable to link cell
cycle position with pH regulation.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 12.
Distribution of cells with high cytosolic pH
levels. Wild-type (WT) Ax4 cells (A) or
rtoA cells (B) were plated at low density in
submerged culture and videotaped. 18 h later BCECF-AM was added,
and ratiometric fluorescence imaging was used to determine the
cytosolic pH of individual cells. The videotape was reviewed, and the
cell cycle phase of each high pH cell was determined and mapped on a
diagram of the cell cycle. In the wild-type, high pH cells originate
generally from cells in late G2, S, or early G2
phase. In the rtoA mutant, high pH cells originate randomly
from cells in any phase of the cell cycle.
|
|
| |
DISCUSSION |
RtoA Mediates Vesicle Fusion--
We were puzzled by the presence
of the 10 tandem repeats of the 11-amino acid-long sequence SSGSSNSGSES
in the middle of the RtoA protein (19). Using the algorithms of Chou
and Fasman (49), the predicted structure of this sequence is a random
coil. In agreement with this prediction, the circular dichroism
spectrum (Fig. 4) and one-dimensional NMR indicate that the serine-rich repeat portion of RtoA is a random coil. The absence of RtoA from the
nuclear or vesicular fractions and the diffuse staining seen with
anti-RtoA antibodies by immunofluorescence suggest that RtoA is a
cytosolic protein. This would then suggest that the potential signal
sequence seen in the predicted amino acid sequence of RtoA (19) is nonfunctional.
PEG, another macromolecule with a random coil structure, causes bilayer
phospholipid membranes to fuse. However, the mechanism by which
proteins or PEG induce membrane fusion is poorly understood. One
hypothesis is that PEG is a very flexible hydrophilic polymer and can
thus interact with a large number of water molecules. This then causes
a local dehydration, which in turn leads to vesicle fusion (50, 51). In
experiments where PEG was placed on one side of a dialysis membrane and
vesicles were placed on the other, the vesicles fused. This supports
the idea that a local dehydration mediates vesicle fusion and shows
that PEG can cause fusion without being in physical contact with the
vesicles (34). The random coil structure of the serine-rich region of
RtoA could allow it to exhibit the same sort of flexibility that PEG
does. In addition, our observation that the S fragment of RtoA can
directly cause vesicles to fuse suggests that RtoA could cause vesicle
fusion by the same mechanism as PEG. A naive prediction of this
observation is that other proteins with large random coil domains may
also induce vesicle fusion. It is also important to note that the
in vitro fusion experiments were performed with only a
portion of RtoA and not the full-length protein. It is therefore
possible that the fusion activity detected may not exist in the whole
protein. However, because the S fragment represents 70% of the
protein, and cells lacking RtoA have defects in vesicle fusion, it
seems reasonable to conclude that the full-length RtoA protein mediates vesicle fusion.
In mammalian cells, vesicle fusion is mediated by membrane proteins
called SNAREs (52). Although these proteins are required for membrane
fusion, their association precedes membrane fusion. The factors that
are actually responsible for fusion of the lipid bilayer are still
unidentified. If a similar complex is formed in
Dictyostelium, it is possible that RtoA represents one of
the factors required for lipid bilayer fusion. Although there appears to be a considerable amount of the 40-kDa RtoA protein in cells, there
is no protein of this size associated with endocytic vesicles (53). In
addition, much of RtoA is in the high speed supernatant of fractionated
cells, suggesting that RtoA is a predominantly cytosolic protein. Only
a portion of the RtoA is associated with small organelles. This may be
because RtoA only associates with endocytic vesicles transiently, being
recruited into a protein complex specifically at the time of membrane
fusion, or like the effect of PEG across a dialysis membrane (34), RtoA
does not need to bind to vesicles to mediate their fusion.
There Is a Complex Heterogeneity in the Vegetative Cell
Population--
The expression pattern of RtoA suggests that at the
moment of cell division, the two daughter cells have roughly equal
amounts of RtoA. Approximately 1.5 h after cytokinesis the level
of RtoA increases in one sister and then decreases in the other sister. Near M phase, the levels appear to equilibrate; the level drops in the
high RtoA sister and increases in the low RtoA sister. This complex
pattern of protein production suggests that the expression of RtoA is
regulated by a cell cycle-dependent mechanism, as well as
by an unknown mechanism involving asymmetric cell division. When we
expressed the rtoA cDNA under control of an actin
promoter (which when used to drive expression of most proteins causes
all cells to have an increased level of the protein), RtoA levels varied enormously from cell to cell, suggesting there is a
post-transcriptional mechanism that modulates RtoA
levels.2
We and others had previously observed that there is a correlation
between the phase of the cell cycle that a cell happened to be in at
the time of starvation and its subsequent initial choice of cell type.
Other workers, having observed that a variety of factors that are
generated by cells many hours after starvation can regulate the
expression of cell type-specific genes, surmised that these factors,
rather than being involved in adjusting ratios of cell types, actually
were responsible for the initial generation of a heterogeneity. The
observation that there is a pre-existing heterogeneity in the
vegetative population, as evidenced by the heterogeneity of RtoA
staining intensities, which correlates with initial differentiation
into prespore cells, shows that the initial choice has already been
made by the vegetative cells.
RtoA and Endosomal Function--
Disruption of RtoA causes an
alteration in vesicle trafficking, specifically in the endosomes.
Although proficient for endocytosis, cells lacking RtoA are defective
in post-lysosomal exocytosis. It has been proposed by others (54) that
post-lysosomes are not stable compartments but instead arise from the
fusion of smaller lysosomes. One possibility is that RtoA mediates this
fusion step, so that in the absence of RtoA, post-lysosomes are
improperly formed, leading to defective exocytosis. Our data indicate
that in addition to RtoA, there is another mechanism that mediates exocytosis. In the rtoA cells, there is still some amount of
exocytosis (Fig. 10A). We also observed that wild-type cells
all ingest FITC-dextran (Fig. 8) and that during the mid and late parts
of the cell cycle, cells with both a low level of RtoA as well as cells
with a high level of RtoA are able to excrete all of the detectable
FITC-dextran (Fig. 9).
We observed that when a cell divides, the two sisters have different
cytosolic pH levels. It is unclear why Dictyostelium evolved
a mechanism that makes the two sister cells from a cell division have
different physiological properties. One possibility is that this
creates a sort of biodiversity, with the low pH sister able to
withstand stresses that the high pH sister cannot and vice
versa. An intriguing finding is that Dictyostelium
cells with a higher cytosolic pH are able to withstand hypertonic
stress better than cells with a low cytosolic pH (55). This process for
increasing diversity may have been co-opted by the developmental program as a mechanism for allowing the differentiation of cells into
stalks and spores.
RtoA, pH, and Cell Type Choice--
It has long been believed that
pH may participate in controlling differentiation in
Dictyostelium (56). Acidification favors development of
prestalk cells, whereas alkalization promotes prespore differentiation.
However, Gross et al. (24) predicted this regulation is not
based on the pH of the cytoplasm, but on the pH of specific vesicles.
They hypothesized that prestalk cells would tend to have some class of
intracellular vesicles with a lower internal pH compared with the same
vesicles in prespore cells. We previously found that rtoA
cells have a higher percentage of prestalk cells (19), and in this
report we show that rtoA cells also tend to have vesicles
with a lower internal pH. This correlates beautifully with the Gross
et al. (24) prediction. We observed a very well defined
sector of the cell cycle that gave rise to prestalk cells (S and early
G2), and this corresponds very nicely to the sector in
which both sister cells have an intermediate level of RtoA (Ref. 19 and
Figs. 3 and 9). Similarly, the region of the cell cycle that gives rise
to prespore cells closely corresponds to the sector where one sister
has a high level of RtoA and the other sister has a low level of RtoA.
However, neither of these sectors correspond well to, or are the same
size as, the region of the cell cycle where cells have a high cytosolic
pH (Fig. 12). Thus, although the absence of RtoA randomizes where
eventual prestalk, eventual prespore, and high pH cells are in the cell
cycle, it does not appear that cytosolic pH can be an accurate
predictor for eventual cell fate. This could be explained in two ways.
One possibility is that RtoA has several activities and its function in
vesicle fusion and pH regulation is separate from its role in initial
cell fate determination. Another possibility is that there is another
level of control, such as cytosolic Ca2+ levels (46), that
aids in determining initial cell fate.
§¶,
,
TOP
ABSTRACT
INTRODUCTION
Present address: Dept. of Biological Sciences, Walla Walla
College, College Place, WA 99324.
