Download Two distinct classes of prestalk

Development 100, 745-755 (1987)
Printed in Great Britain © The Company of Biologists Limited 1987
745
Two distinct classes of prestalk-enriched mRNA sequences in
Dictyostelium discoideum
K. A. JERMYN1, M. BERKS2, R. R. KAY2 and J. G. WILLIAMS1*
'Imperial Cancer Research Fund, Clare Hall Laboratories, Blanche Lane, South Mimms, Herts, EN6 3LD, UK
MRC Laboratory' of Molecular Biology, Hills Road, Cambridge CB2 3QH, UK
2
* To whom reprint requests should be addressed
Summary
We have isolated cDNA clones derived from three
mRNA sequences which are inducible by DIF, the
putative stalk-specific morphogen of Dictyostelium.
The three mRNA sequences are selectively expressed
in cells on the stalk cell pathway of differentiation and
we have compared them with previously characterized prestalk-enriched mRNA sequences. We find
these latter sequences are expressed without a dependence on DIF, are much less highly enriched in
prestalk over prespore cells and are expressed earlier
during development than the DIF-inducible mRNA
sequences. We propose two distinct mechanisms
whereby a mRNA may become enriched in prestalk
cells. An apparently small number of genes, represented by those we have isolated, is inducible by
DIF and accumulates only in prestalk cells. We
suggest that a second class of prestalk-enriched
mRNA sequences are induced by cAMP to accumulate in all cells during aggregation and then become
enriched in prestalk cells by selective loss from prespore cells.
Introduction
difference in the carbohydrate modification of several
lysosomal enzymes and this has been used as a
marker of prestalk cell differentiation (Oohata, 1983;
Knecht, Green, Loomis & Dimond, 1984; Loomis &
Kuspa, 1984). There are also several cDNA clones
derived from mRNA sequences enriched in prestalk
over prespore cells which have been used to determine the timing of prestalk cell differentiation and to
analyse the effect of cellular perturbations, such as
disaggregation and the addition of cAMP, upon
prestalk cell gene expression (Barklis & Lodish, 1983;
Mehdy, Ratner & Firtel, 1983).
Isolated cells of the strain V12M2, incubated at low
cell density (Town, Gross & Kay, 1976), can be
induced to form stalk cells by DIF. This is a low
molecular weight hydrophobic compound, which has
been purified and structurally identified (Kay,
Dhokia & Jermyn, 1983; H. Morris, G. W. Taylor,
M. Masento, K. Jermyn and R. Kay, unpublished
results). The mutant strain HM44 is defective in the
production of DIF (Kopachik et al. 1983) but remains
responsive to exogenous DIF. Hence it can be used to
determine whether the expression of a particular
The migrating slug of Dictyostelium discoideum contains up to 105 cells which have aggregated together in
response to pulsatile emissions of cAMP. Under
normal circumstances, cells in the anterior, prestalk,
region differentiate to form stalk cells and cells in the
posterior region differentiate to spores. The two
precursor cell types are readily purified, either by
microdissection or by Percoll gradient centrifugation,
and a number of biochemical markers which distinguish them have been identified (reviewed by
Morrissey, 1982; Williams etal. 1986). Two-dimensional gel electrophoresis indicates a continuity of gene
expression on the prespore-spore cell pathway of
differentiation (Morrissey, Devine & Loomis, 1984;
Ratner & Borth, 1983) and there are reliable enzymic, immunological, structural protein and cDNA
markers of prespore cells. There is little apparent
continuity on the stalk cell pathway of differentiation, with very few prestalk-specific proteins being
expressed in prestalk and stalk cells (Ratner & Borth,
1983; Morrissey et al. 1984). However, there is a
Key words: Dictyostelium discoideum, mRNA, DIF,
prestalk cells.
746
K. A. Jermyn, M. Berks, R. R. Kay and]. G. Williams
marker is dependent upon DIF. In the absence of
DIF, prespore cell proteins accumulate but the prestalk-specific isozyme of acid phosphatase is not produced (Kopachik et al. 1983). A similar conclusion
derives from the analysis of total cellular proteins
by two-dimensional gel electrophoresis (Kopachik,
Dhokia & Kay, 1985). We have employed this in vitro
induction system to investigate the effect of DIF upon
specific mRNA accumulation. We have used previously isolated prestalk-enriched cDNA clones and
also cDNA clones which we isolated by differential
screening for DIF-induced mRNA sequences. Analysis of the accumulation of these various mRNA
sequences - in response to DIF, during normal
development and in isolated prestalk and prespore
cells - has led us to re-evaluate the usefulness
of previously described cDNA clones as prestalk
markers and to suggest the existence of two distinct
mechanisms for establishing a differential mRNA
concentration in the two cell types.
in i h periods, depositing the cells in ice-cold KK2 to inhibit
dedifferentiation and, after pelleting, snap freezing on dry
ice. For terminally differentiated cells, freshly culminated
Ax-2 fruiting bodies were harvested and stalks were separated as described by Devine, Morrissey & Loomis (1982)
from freshly culminated Ax-2 cells.
RNA isolation
All RNA was isolated by a minor modification of the
GTME method (Maniatis, Fritsch & Sambrook, 1982)
except the spore RNA which was isolated according to
Macleod, Firtel & Papkoff (1980). Following centrifugation, RNA was routinely subjected to 1 phenol and
1 chloroform extraction. Ribosomal RNA integrity was
monitored on a 0-8% minigel (TBE pH8-3) after denaturing at 50°C for 15min.
cDNA preparation
A cDNA library was prepared by the method of Gubler &
Hoffman (1983) using mRNA isolated from HM44 cells 12 h
after DIF addition. The cDNA was cloned into the Pst\ site
of the pUC8 vector, and transformed into DH5 cells. About
12000 colonies were generated, of which about 400 were
picked for secondary screening. The colonies were repliMaterials and methods
cated onto Genescreen Plus or nitrocellulose filters and
screened with 32P-labelled cDNA made with RNA isolated
Dictyostelium discoideum growth and development
from cells incubated in the absence or presence of DIF for
Strains NC4 (DdB) (obtained from P. Newell), V12M2
12 h. The filters were prehybridized for at least 2h at 37°C
(obtained from G. Gerisch) and the V12M2-derived mutant
in 50% formamide, 5xSSC, 10mM-Na2HPO4/NaH2PO4
HM44 (Kopachik et al. 1983) were all grown in association
pH6-72, lOxDenhardt's solution, 250 fig ml"1 salmon
with Klebsiella aerogenes plated on SM agar (Sussman, 1966) testes DNA (heat denatured and sonicated), 1 % SDS.
at 22°C. Ax-2 (from J. Ashworth) was grown axenically in
Hybridization was carried put in the same buffer at 37°C
suspension (Watts & Ashworth, 1970). Cells were normally
overnight. The filters were washed for 5x15 min at 37°C in
harvested and washed by centrifugation in 20 mM-KH2PO4/
prewarmed 50% formamide, 2xSSC, 1 % SDS and finally
K2HPO4, pH60 (KK2). Ax-2 cells were developed on
in
lxSSC at room temperature, before being sealed in
1-8% KK2 agar at approximately 5xlO('cellscm~2. For
polythene and exposed tofilmat —70°C using an intensifydevelopmental time courses, NC4 and V12M2 cells were
ing screen.
developed on 1-8% agar containing 40mM-Na2HPO4/
KH2PO4, 20mM-KCl and 2-5mM-MgCl2, pH6-2, at about
Northern transfer
5xl0 6 cellscm~ 2 . Amoebae of the V12M2-derived strain,
The
RNA samples (5f(g) were electrophoresed through a
HM44, were induced to form stalk cells in 14 cm Sterilin
formaldehyde
denaturing gel system and transferred onto
tissue culture dishes as described by Kopachik et al. (1985).
nitrocellulose as described by Maniatis et al. (1982) except
Cells were incubated in the presence of cAMP for 10 h ('t0'
that the filter was baked and prehybridized immediately
of the induction) and then the medium was renewed and
after transfer. Thefilterwas hybridized as described above
DIF was added to half of the plates at 3000 units ml"'.
except that 0 1 % SDS was used in all the solutions.
Scanning of the auforadiograms was performed with an
Purification of differentiated cell types
LKB Ultrascan XL laser scanning densitometer.
For migrating slugs, cells were developed as single streaks
9
1
of approximately 10 cells ml" on 2% agar containing
20mM-KCl, 20mM-NaCl and 1 mM-CaCl2 for cutting or 2 % Results
water agar for Percoll gradient separation of prestalk and
prespore cells (Ratner & Borth, 1983). These plates were
(A) Isolation and characterization of cDNA clones
kept moist and dark except for a single slit providing a
derived from three mRNA sequences which
unidirectional light source. Gradient-purified cells were
accumulate in response to DIF
separated by the method of Ratner & Borth (1983), except
The
HM44 mutant is defective in DIF production. In
that 60% gradients were normally used. Slugs were
the
absence
of DIF, only a Very low proportion
manually dissected using a microknife (a mounted, flat(<1 %) of amoebae differentiate into stalk cells
tened dissecting pin) to yield a front (10 %) or a rear (50 %)
while, in its presence, a very high proportion (>95 %)
portion. The very rearmost portion of the slug was cut away
to minimize basal disc cell contamination and to sever the
of stalk cells are formed. A cDNA clone bank was
slime trail. Usually one or other type was rapidly collected
prepared from HM44 cells incubated with cAMP for
Prestalk-enriched mRNAs in Dictyostelium
10 h and then with DIF for a further 12 h. The bank
was screened differentially with labelled cDNA prepared from the RNA used to construct the bank
and with cDNA prepared from a control, non-DIF
induced, RNA preparation. Nine cDNA clones
which showed differential hybridization were isolated. Cross-hybridization with purified inserts from
each of the nine cDNA clones and mapping with
restriction enzymes showed them to derive from three
different mRNA sequences. The longest cDNA
clones derived from each of the three mRNA sequences were used in all subsequent experiments.
These contained inserts of 0-65kb (pDd26), 2-4kb
(pDd56) and 2-3 kb (pDd63) in length. Subsequent
analysis, by Northern transfer, showed pDd26 to
contain an almost full-length insert while pDd56 and
pDd63, which derive from longer mRNA sequences,
contain portions of the coding sequences proximal to
the 3' end of their respective mRNA sequences. The
fact that multiple cDNA clones for each mRNA were
isolated (four clones derived from pDd56 mRNA,
three clones from pDd26 mRNA and two clones from
the pDd63 mRNA) suggests these to be the most
abundant mRNA sequences induced by DIF.
The mRNA hybridizing to pDd26 is 0-7 kb in
length (Fig. 1). It is not detectably expressed before
6 h of induction and maximal expression is achieved
in the terminal stages of stalk cell differentiation.
Accumulation of this mRNA is totally dependent
upon the presence of DIF.
The mRNA hybridizing to pDd56 is 3-2 kb in
length and shows a somewhat more complex pattern
of accumulation (Fig. 1). The major peak in mRNA
abundance occurs at about 6-8 h after DIF induction,
with a decline in concentration thereafter. However,
at a higher level of exposure of this Northern transfer
and in all other experiments using a number of
different mRNA preparations, the pDd56 mRNA
showed a low, but definite, level of expression within
l h of DIF addition (data not shown). Thus expression of the pDd56 mRNA is much more rapidly
induced by DIF than is the pDd26 mRNA but again
accumulation is totally dependent upon DIF.
The pDd63 cDNA clone hybridizes to a mRNA
of 5-8kb in length. This mRNA accumulates very
rapidly after DIF addition, reaching 25 to 30 % of its
maximal level by 1 h of incubation with a broad peak
of accumulation at 4-6 h after DIF addition. At early
times of incubation, the mRNA is almost totally
dependent upon DIF for its accumulation, with only a
very low level of expression detectable in overexposed autoradiograms in the absence of DIF. After
prolonged incubation without DIF, a low level of
expression becomes apparent (Fig. 1). The pDd63
gene might be expected to be responsive to a very low
level of DIF, as it is highly induced within 1 h of the
747
addition of exogenous DIF. Hence, we believe the
expression observed in the control cells results from
the low level of DIF production known to occur in the
HM44 mutant (Kopachik et al. 1983). The data
presented in Table 1 suggest that this is indeed the
case. In this experiment, DIF was added to cells at
levels which are suboptimal for stalk cell induction
and the accumulation of pDd63 mRNA was analysed.
There was a clear uncoupling of terminal stalk
cell differentiation from gene induction with, for
1
2
4
6
8
12
o +- +- +- +- +- + -
pDd63
pDd56
pDd26
Iff
—5-8 kb
•ft
— 32kb
— 0-75kb
Fig. 1. The kinetics of induction of three mRNA
sequences dependent upon DIF for their accumulation. A
DIF induction was performed with HM44 under the
standard conditions and cells were harvested at the
indicated times. The symbol 0 indicates the time at which
DIF was added (Fig. 1); the numbers indicate the number
of hours of incubation in the absence (—) or presence (+)
of DIF. Total cellular RNA was purified and 5jUg samples
were electrophoresed through an agarose gel and
transferred to nitrocellulose. The filters were hybridized
with the indicated DNA which was labelled by nick
translation. The size estimates derive from a separate
experiment where the RNA samples were
coelectrophoresed with RNA size markers.
748
K. A. Jermyn, M. Berks, R. R. Kay and J. G. Williams
Table 1. The induction ofpDd63 mRNA
accumulation by low levels of DIF
Concentration
of DIF
(units ml" 1 )
Stalk cells
induction
(%)•
12 500
1250
125
12-5
100
15
5
0-7
Fraction of
maximal pDd63
mRNA induction
(%)t
100
30
20
3
* Normalized to the fraction of amoebae (75 %) forming stalk
cells at the highest DIF concentration (12 500 units ml" 1 ).
t Measured 4 h after the addition of DIF.
example, a level of DIF sufficient to induce only 5 %
stalk cell formation giving 20 % of the control level of
pDd63 expression.
The three cDNA clones hybridize to mRNA sequences which differ in their size, in their kinetics of
accumulation and disappearance after DIF induction
and in the totality of their dependence upon exogenous DIF. However, two of the sequences encode
related proteins. A more heavily exposed autoradiogram of the hybridization with pDd56 reveals the
presence of an RNA of higher molecular weight and
reprobing of the same filter with pDd63 shows this to
be coincident with the pDd63 mRNA (data not
shown). The cross-hybridizing mRNA has precisely
the same kinetics of accumulation as the pDd63
mRNA, both in the presence and absence of DIF.
Washing at an elevated stringency abolishes the signal
obtained using pDd56 (see Fig. 4B). Reciprocal
cross-hybridization, i.e. of the pDd63 cDNA clone
with the pDd56 mRNA, is not detected. This apparent anomaly is explained by recent partial sequence
analysis of a genomic clone containing the pDd63
gene which shows the 5' proximal portion of the gene
to have a higher degree of homology to the pDd56
mRNA than the 3' proximal portion (A. Ceccarelli,
S. J. McRobbie, H. Mahbubani and J. G. Williams,
unpublished results). Since the pDd63 cDNA clone
contains only the 3' proximal half of the pDd63
mRNA it does not cross-hybridize to the pDd56
mRNA under these conditions.
(B) The kinetics of accumulation of the three mRNA
sequences during normal development is consistent
with their being induced by DIF
The three DIF-induced mRNA sequences display
quite different kinetics of accumulation after DIF
induction (Fig. 1). During normal development the
major rise in DIF levels starts at about the tipped
mound stage of development (Brookman, Town,
Jermyn & Kay, 1982). If DIF acts to induce the
accumulation of these sequences, they would be
expected to show the relative order of appearance
and kinetics of accumulation seen in the in vitro
system starting after this stage. The data presented in
Fig. 2A,B are from the strain V12M2, the parental
strain of HM44, and from strain NC4. We used both
strains because the timing of development is somewhat different, with strain V12M2 showing a more
rapid progression through the early stages of differentiation and because we wished to compare the V12
parental strain with the more commonly used NC4
isolate. When correlated with the morphological
changes occurring during development, both strains
display similar kinetics of accumulation of the three
mRNA sequences but there are clear differences.
In V12M2, the mRNA hybridizing to pDd63 appears at the tipped aggregate stage of development
(Fig. 2A). There is a rapid rise in expression to reach
a peak at the slug stage and the mRNA then
decreases in concentration to a very low level at
culmination. In NC4 cells (Fig. 2B), the pattern of
expression is similar but the mRNA persists at a high
concentration until much later into development.
(This is not the result of asynchrony in development
as almost perfect synchrony was observed for NC4.
The result probably underestimates the actual sharpness of the peak in V12M2 as it was not possible to
obtain perfect synchrony with this strain and significant numbers of migrating slugs were present in the
plates at the time culmination occurred in the majority of the aggregates.)
The pDd56 mRNA also becomes detectable at the
tipped aggregate stage in both strains. In V12M2
cells, it displays a sigmoidal accumulation curve
similar to that observed upon DIF induction, with a
very- low level initially and a sharp peak at the
'mexican hat' stage of development. The lag in NC4
cells is less pronounced, suggesting a difference in the
kinetics of induction in the two strains. Again, as with
the pDd63 mRNA, the pDd56 mRNA persists at a
higher level late into the development of NC4 with no
clear peak of accumulation being apparent. The
mRNA hybridizing to the pDd26 mRNA is not
detectable until the mexican hat stage of culmination
in either NC4 or V12M2.
Thus, the time course of accumulation of all three
mRNA sequences matches their kinetics of appearance upon DIF induction very well. The only notable
strain differences in developmental expression are
the somewhat higher level of early accumulation of
the pDd56 mRNA in NC4 cells and the more rapid
disappearance of both the pDd63 and pDd56 mRNA
sequences in V12M2 cells.
Several cDNA clones have previously been described which hybridize to mRNA sequences selectively enriched in prestalk over prespore cells (Barklis
& Lodish, 1983; Mehdy et al. 1983). These were
Prestalk-enriched mRNAs in Dictyostelium
4 6 8 1 0 1 2 14 1 6 1 8 2 0 2 2 . 5
0^2
•ft
749
pDd63
.pDd63
pDd56
.pDd26
•*»#
P Dd26
B
NC4
10^
12
JL
14
16
\ •«
18
•«*
20
• •
M,
2 2 . 5
.
-D11 Prestalk
-pDd63
Fig. 2. The time course of accumulation of the three DIF-induced mRNA sequences during normal development.
Amoebae of the strain V12M2 (Panel A), or NC4 (Panel B,C) were grown and set up for development as described in
the Methods section. Development of the NC4 cells was almost perfectly synchronous with no prolonged period of slug
migration. However, it is not possible to achieve such a high degree of synchrony with the strain V12M2 and there were
significant numbers of migrating slugs in all plates harvested after the 12th hour of development. The symbols above the
time points represent the morphological stages. For NC4 these were lOh, tight aggregates; 12h, tipped aggregates; 14h,
first fingers; 16h, mexican hats; 18h, preculminants; 20h, midculminants; 22-5 h, culminants. For V12M2 these were 6h,
tight aggregates; 8h, tipped aggregates; 10h, first fingers; 12h, standing slugs (with some slugs); 14h, mexican hats
(with some slugs); 16h preculminants (with some slugs); 18h midculminants (with some slugs); 20h culminants (with
some slugs). In the experiment shown in Panel C, the filter was cut transversely, after transfer. The two portions of the
filter were separately hybridized with a pDd63 or a Dll probe in order to eliminate loading artefacts. This exposure of
the autoradiogram shows a signal for both pDd56 and pDd63 only at the first finger stage but a longer exposure of the
same filter shows these to be expression of both genes in tipped aggregates (data not shown).
750
K. A. Jermyn, M. Berks, R. R. Kay and J. G. Williams
selected by differential screening with mRNA isolated from cells fractionated on density gradients
(Tsang & Bradbury, 1981; Ratner & Borth, 1983). In
addition to correlating expression of the DIF-inducible mRNA sequences with morphological stages, we
have also determined their time of expression relative
to D l l , a prestalk gene typical of previously isolated
sequences. The pDd63 and pDd56 mRNA sequences
are first detectably expressed approximately 2h later
than the major rise in concentration of the Dll
mRNA (Fig. 2C).
(C) A class of transcripts, which become enriched in
prestalk cells, does not require DIF for their
expression
We have determined the effect of DIF on a number of
the mRNA sequences previously described as prestalk-enriched and find a variable pattern of response.
The cysteine proteinase 2 mRNA encodes a Dictyostelium protein homologous to the sulphydryl proteinase of higher eukaryotes (Pears, Mahbubani & Williams, 1985). It is totally absent from vegetative cells,
begins to accumulate late during cellular aggregation
and is prematurely expressed in response to exogenous cAMP. In previous work from this laboratory, it
has been reported to be approximately threefold
enriched in prestalk over prespore cells (Pears et al.
1985), but we have more recently obtained a somewhat lower estimate of a twofold enrichment using a
different gradient fractionation procedure (Pears &
Williams, 1987). This gene has been independently
isolated and characterized elsewhere (Mehdy et al.
1983; Datta, Gomer & Firtel, 1986) and the protein
has been shown to be specifically localized in prestalk
cells (Gomer, Datta & Firtel, 1986). Since the mRNA
is expressed at a high relative level in prespore cells it
may be that it is selectively translated or the protein
may be differentially stable in the two cell types. The
mRNA is present at t0 in the standard induction (i.e.
the time of DIF addition) and, in the absence of DIF
(Fig. 3), it displays approximately the same kineticsof disappearance as in normal development (Pears et
al. 1985). The effect of DIF is to accelerate disappearance of the mRNA from the cells, presumably as a
result of their differentiation into stalk cells.
The prestalk-enriched mRNA sequences isolated
by Barklis and Lodish have been shown to fall into
two classes termed Prestalk 1 (Pstl) and Prestalk 2
(Pst2) (Barklis & Lodish, 1983; Chisholm, Barklis &
Lodish, 1984). These two classes differ in their time
course of accumulation and in their response to
cellular disaggregation and incubation with exogenous cAMP.
The Dll (Pstl) gene encodes a hydrophilic, cysteine-rich protein of unknown function (Barklis,
Pontius & Lodish, 1985). The mRNA is expressed at
0 1
8
12
„
Time after
-DIF addition
Cysteine proteinase 2
-D11
• I
»^»
••••V
•PL1
2H6
-Ras
Fig. 3. The effect of DIF upon previously characterized
prestalk-enriched mRNA sequences. A Northern transfer
was performed with the same RNA samples used for the
experiment described in Fig. 1. The filter was hybridized
with the indicated DNA probes and washed under the
standard conditions (see Methods section) before
autoradiography. The probe for the ras mRNA was a
synthetic oligonucleotide of 23 residues in length with the
sequence 5' CGATAACTATCTTCAATTGTTGG 3'.
This is the complement of a sequence located in the 5'
proximal coding portion of the ras mRNA (Reymond et
al. 1984). The oligonucleotide was 5' end labelled using
T4 polynucleotide kinase and y-32P-ATP.
a low level during vegetative growth in axenic medium, accumulates to a 10- to 20-fold higher level
during aggregation (Barklis & Lodish, 1983; Chisholm et al. 1984) and is induced in response to
exogenous cAMP (D. Driscoll, personal communication; Oyama & Blumberg, 1986). The mRNA is
present at t0 in a DIF induction but it differs from the
cysteine proteinase 2 mRNA in that the concentration remains constant in the absence of DIF
(Fig. 3). In contrast to cysteine proteinase 2, DIF has
absolutely no effect on the concentration of the Dll
mRNA. Since the cells undergo terminal differentiation into stalk cells during this time period, these
results would suggest that the D l l mRNA is
Prestalk-enriched mRNAs in Dictyostelium
expressed in mature stalk cells and this is confirmed
below.
The PL1 mRNA is a Pst2 transcript, encoding a
protein of unknown function. It is reported to be
enriched in prestalk over prespore cells (Barklis &
Lodish, 1983) but we do not find this to be the case
(see below). In contrast to the D l l mRNA sequence,
the PL1 mRNA is expressed only after the loose
aggregate stage of development and the mRNA is not
inducible with cAMP (Oyama & Blumberg, 1986).
The PL1 mRNA is not expressed at t0 in a DIF
induction (Fig. 3), a result consistent with its late
stage of expression during normal development and
its nonresponsiveness to cAMP. However, after 4-6 h
of incubation in the presence of DIF-1, the PL1
mRNA is expressed. This observation suggests that
the PL1 mRNA may play a role in the terminal stages
of stalk cell differentiation and its presence in mature
stalk cells (see below) is consistent with this hypothesis.
The 2H6 cDNA clone hybridizes to a mRNA of
unknown coding potential, which is enriched in
prestalk cells (Mehdy et al. 1983) and which seems to
show the properties of a Pstl mRNA. This mRNA is
expressed at a finite level at t0 in the DIF induction
and increases in abundance at 4-6 h after DIF addition (Fig. 3). This is the time of overt stalk cell
differentiation and we find that this mRNA sequence
is expressed at culmination in stalk cells; although in
contrast to PL1, which shows similar kinetics of
induction, the 2H6 sequence is also expressed in
mature spore cells.
The Dictyostelium ras gene is expressed during
vegetative growth; the mRNA declines in abundance
during early development and reaccumulates late
during aggregation when it is reported to be enriched
in prestalk over prespore cells (Mehdy et al. 1983;
Reymond, Gomer, Medhy & Firtel, 1984). The ras
mRNA is expressed in HM44 cells at t0 in the
induction and DIF has no effect on accumulation of
the mRNA (Fig. 3).
Thus, the cysteine proteinase 2, D l l , ras and 2H6
mRNA sequences, which are enriched in prestalk
over prespore cells, share one important common
feature with respect to DIF induction. They are all
expressed at high levels in HM44 cells incubated in
submerged conditions in the presence of cAMP for
10h (i.e. t0 in the standard induction conditions).
Hence, in this in vitro system at least, they do not
require DIF for their expression.
(D) The DIF-induced mRNA sequences are more
highly enriched in the prestalk-stalk cell pathway of
differentiation than previously described sequences
We have determined the relative degree of enrichment of these different mRNA sequences in purified
751
Table 2. The relative degree of enrichment of prestalk
and prespore mRNA sequences in purified cells
mRNA sequence
pDd63
pDd56
pDd26
Dll
PL1
ras
2H6
cysteine
proteinase 2*
D19t
Approximate
prestalk/prespore
ratio
Approximate
stalk/spore
ratio
S10:l
3=10:1
Not expressed
4:1
1:1
3:1
5:1
1-5:1
50:1
50:1
50:1
50:1
50:1
6:1
1:1
1:1-5
s=l:10
N.D.
*Data derived from Pears & Williams (1987).
tD19 is a prespore 2 clone (Barklis & Lodish, 1983) which
was used to monitor the purity of the prestalk fraction.
prestalk and prespore cells. Migrating slugs were
fractionated by manual microdissection or, after
disaggregation, by Percoll gradient centrifugation
(Ratner & Borth, 1983). Microdissection yields an
anterior, prestalk fraction, which should in principle
be totally devoid of prespore cells. The posterior
portion of slugs is predominantly composed of prespore cells but about 10-15 % of the cells display the
morphological and biochemical properties of prestalk
cells and hence are termed 'anterior-like cells' (Sternfeld & David, 1981, 1982; Devine & Loomis, 1985).
Based upon staining with a prespore-specific antiserum, the prespore cells purified by density centrifugation were estimated to be between 5 and 10 %
contaminated with prestalk cells. The degree of
enrichment observed for the various mRNA sequences was very similar using RNA isolated from
cells purified by either of the two methods.
The Dll, the PL1 and the ras mRNA sequences all
show a lower prestalk enrichment than previously
reported. The most extreme example is PL1, which
we find to be present at approximately equal levels in
prespore and prestalk cells (Fig. 4A, Table 2). We do
not understand the reason for this contradiction but it
appears unlikely to be due to a strain difference, since
M. Oyama and D. Blumberg (personal communication) have obtained similar results to ours using
slugs of the strain NC4. Hence, it may reflect some
difference in conditions of development or gradient
fractionation. The cysteine proteinase 2, D l l , 2H6
and ras mRNA sequences are somewhat prestalkenriched. We have previously estimated the cysteine
proteinase 2 mRNA to be approximately threefold to
fourfold enriched using gradient fractions prepared
by the method of Tsang & Bradbury (1981), but we
obtain a slightly lower estimate of a twofold enrichment using cells fractionated by the gradient pro-
752
K. A. Jennvn. M. Berks, R. R. Kav and]. G. Williams
A Dissection Gradient
Pst PspPstPspSt Sp
I
I I I
I I
_pDd63
— D11
-PL1
§!•••
— Ras
pDd26
D19
B
PstPspStSp
Low
w
stringency 49
High
*
stringency w
•
IB
pDd63
pDd56
pDd63
V
pDd56
Fig. 4. A comparison of specific mRNA abundance in
prestalk and prespore cells. Prestalk. prespore. stalk and
spore cells were purified as described in the Methods
section and a 5,ug sample of the RNA from each cell
population was analysed by Northern transfer using the
indicated probes (Fig. 4A). The D19 cDNA clone
hybridizes to a prespore-enriched mRNA sequence
(Barklis & Lodish. 1983) and is included as a measure of
the purity of the prestalk fraction. The filter hybridized
with pDd56 (Fig. 4B) was washed firstly at low (3xSSC.
50% formamide. 37°C) and then at high stringency
(0-6XSSC. 50% formamide. 37°C).
cedure of Ratner & Borth (1983) and by slug-cutting
(Pears & Williams, 1987). We find the ras mRNA
sequence to be only approximately threefold enriched in prestalk over prespore cells (Fig. 4A.
Table 2). The D l l mRNA sequence is fourfold to
fivefold enriched in prestalk over prespore cells but
this is again a lower enrichment than previous results
suggest (Barklis & Lodish, 1983). A proportion of the
apparent expression of these mRNA sequences in
prespore cells could result from contamination with
prestalk cells. There must, however, be a finite level
of expression of these genes in prespore cells because
we can show, using these same fractions, that the
pDd63 mRNA sequence is much more highly enriched in prestalk over prespore cells (Fig. 4A.
Table 2). In a large series of experiments, using many
different preparations of purified cells, we have
established that the low level of apparent prespore
expression we observe for the pDd63 mRNA can be
accounted for entirely by contamination with prestalk
or anterior-like cells (Williams et al. 1987).
The pDd56 mRNA is expressed in slugs but at
submaximal levels. We have not determined the
extent of enrichment to the same degree of precision
but, using gradient-purified fractions, we find the
pDd56 mRNA is enriched in prestalk over prespore
cells. In the experiment shown in Fig. 4B. we simultaneously assayed the relative degree of enrichment
of both the pDd56 and pDd63 by hybridizing with the
pDd56 probe at low stringency. As discussed above,
the two sequences share sufficient sequence homology to cross-hybridize under these conditons. This
result clearly shows the pDd56 mRNA is as highly
prestalk enriched as the pDd63 mRNA. The pDd26
mRNA sequence is not expressed at a detectable
level in the slug. This sequence is detectable in stalk
cells isolated from fruits in the terminal stages of
culmination, but it is not detectable in purified spore
cells (Fig. 4A). The pDd63 and pDd56 mRNA sequences are also expressed in stalk, but not in spore
cells (Fig. 4A.B). The D l l and PL1 mRNA sequences are both highly enriched in stalk over spore
cells, but the 2H6 and ras mRNA sequences are
expressed at a significant level in both cell types
(Fig. 4A) as is the cysteine proteinase 2 mRNA
(Pears & Williams, 1987).
Discussion
During normal development on agar. the mutant
strain HM44 becomes arrested after aggregation at a
stage prior to slug formation (Kopachik et al. 1983).
Cells within the aggregate have been shown to
express morphological and biochemical markers
characteristic of prespore cells (Kopachik et al. 1983.
Prestalk-enriched mRNAs in Dictyostelium
1985). The prestalk-specific isozyme of acid phosphatase does not accumulate in HM44 cells in the absence
of DIF but is produced in the presence of DIF
(Kopachik et al. 1983). This result provides strong
support for a major role of DIF in directing prestalk
cell differentiation. In apparent contrast to this result,
we observe a high level of expression of several
prestalk-enriched mRNA sequences in HM44 cells
developing in the absence of DIF. We believe the
explanation for this apparent paradox lies in the two
different mechanisms whereby a product may come
to be enriched within a specific cell type; differential
accumulation or differential loss.
The most compelling evidence for selective loss of a
gene product in prespore cells derives from several
monoclonal antibodies which stain a major fraction of
the cell population early during aggregation but
which stain only prestalk cells in the migrating slug
(Tasaka, Noce & Takeuchi, 1983; Krefft, Voet, Gregg
& Williams, 1985). The D l l , cysteine proteinase 2,
2H6 and ras genes are expressed prior to tip formation and they are rapidly induced by cAMP which
may, therefore, act as the initial trigger for their
accumulation during normal development (Barklis &
Lodish, 1983; Mehdy et al. 1983; Pears et al. 1985).
Only the most extreme cell-sorting models for pattern
formation in Dictyostelium involve cellular differentiation concomittant with the initiation of cAMP
signalling. Hence, it seems reasonable to assume that
transcripts of these genes are present at equal concentrations in the precursors of both prestalk and prespore cells. It is then only necessary to assume a
differential loss in prespore cells, due to an altered
rate of transcription or post-transcriptional processing, to account for the different mRNA concentrations observed in the two cell types. Since cAMP is
present in the in vitro assay with HM44, we would
expect these genes to be expressed and, with the
above interpretation of their prestalk-enrichment,
their nonresponsiveness to DIF would be explained.
There is, of course, no direct evidence that these
genes are expressed in all cells during aggregation. It
may be that there is a subset of aggregating cells,
fated to become stalk cells, which expresses the genes
at a higher level than presumptive prespore cells.
However, such a model requires that the genes later
be turned on in 'definitive' prespore cells, where we
observe their expression, or alternatively that the
subset of expressing cells sediments with prespore
cells on a gradient. Given the existence of other
'prestalk' markers initially expressed in most or all
cells (Tasaka et al. 1983; Krefft et al. 1985), we feel it
reasonable to suggest differential loss from prespore
cells for these mRNA sequences. Support for this
suggestion derives from our identification of pDd63
and pDd56, two mRNA sequences which are more
753
highly enriched in prestalk over prespore cells than
these other sequences and which differ radically in
their mode of induction and in their time of appearance during development.
The level of expression in prespore fractions of
pDd63 and pDd56 is low enough to be accounted for
by prestalk cell contamination and thus they are the
first potentially presta\k-speciftc mRNA sequences to
be identified. As such, their time of appearance
during normal development can be used as a definitive marker for the appearance of prestalk cells. The
pDd63 and pDd56 mRNAs are first detectably
expressed at the tipped aggregate stage of development, several hours later than the cysteine proteinase
2 (Pears et al. 1985) Dll, ras and 2H6 mRNA
sequences (Fig. 4C). Again, this is consistent with our
suggestion that these latter mRNA sequences accumulate in all cells, becoming selectively enriched in
prestalk cells. The early appearance of such sequences has led to the suggestion that prestalk cell
differentiation occurs very early during development.
If we are correct in our interpretation of the mechanism whereby these mRNA sequences become enriched, then this conclusion is rendered invalid.
Similarly, they cannot safely be used to assess the
response of prestalk cells to inducing agents such as
cAMP or DIF, or to determine the effect of other
potential control signals such as cell contact (Chisholm etal. 1984).
Of course, we cannot exclude the existence of
prestalk-specific markers expressed earlier in development than pDd56 and pDd63. There are cells at
the tight aggregate stage of development which do
not express prespore-specific antigens (Takeuchi,
Hayashi & Tasaka, 1977; Krefft et al. 1984) but there
is no direct proof from fate mapping that these are the
precursors of prestalk cells. There is also no indirect
evidence from protein markers, because analysis of
proteins by two-dimensional gel electrophoresis has
shown that, while there is continuity in the genes
expressed on the prespore-spore cell pathway of
differentiation, this is not the case on the prestalkstalk pathway (Ratner & Borth, 1983; Morrissey etal.
1984). This has led to the suggestion that most
prestalk-enriched proteins are involved in processes
such as cAMP signal transduction and other processes
peculiar to the slug stage (Morrissey, 1982). There is
now good evidence that the ras mRNA sequence,
which we confirm to be marginally enriched in prestalk over prespore cells, encodes a protein which is
involved in signal relay (Reymond et al. 1986). The
pDd63 and pDd56 mRNA sequences appear, therefore, to be exceptional in that they are expressed in
both prestalk and stalk cells. The pDd56 and pDd63
genes encode closely related, cysteine-rich proteins
composed of a tandem array of a simple peptide
754
K. A. Jermyn, M. Berks, R. R. Kay and J. G. Williams
repeat (A. Ceccarelli, S. J. McRobbie, H. Mahbubani & J. G. Williams; unpublished results). These
are characteristic features of extracellular structural
proteins of lower eukaryotes and we have preliminary
evidence that these two proteins form part of the stalk
wall of Dictyostelium (S. J. McRobbie & J. G.
Williams, unpublished results).
The pDd56 and pDd63 mRNA sequences are
rapidly induced in response to DIF but they differ in
their precise kinetics of accumulation and in their
level of expression in the absence of DIF. The pDd56
mRNA displays a sigmoidal accumulation curve and
is totally dependent upon DIF for its accumulation.
The pDd63 mRNA accumulates rapidly with linear
kinetics and there is a low level of expression in the
absence of DIF which we believe to be the result of
slight leakiness in the DIF-deficient mutant. The
difference in response to DIF may not reflect a
fundamental difference in mechanism of induction.
Both mRNA sequences increase in concentration
within 1 h of DIF addition and both sequences display
more similar patterns of accumulation during normal
development of the strain NC4. The lag observed for
pDd56, in the in vitro system with HM44, may,
therefore, reflect its derivation from the strain
V12M2.
The mechanism whereby those cells responding to
DIF come to constitute the prestalk fraction is as yet
unknown. It may involve a gradient of extracellular
DIF within the slug or it may be a result of differentiation in situ followed by cell sorting. The pDd63 and
pDd56 cDNA clones provide markers with which to
distinguish these possibilities.
prespore and prestalk cells of Dictyostelium. Proc. natn.
Acad. Sci. U.S.A. 79, 7361-7365.
DEVINE, K. & LOOMIS, W. F. (1985). Molecular
characterization of anterior like cells. Devi Biol. 107,
364-372.
GOMER, R. H., DATTA, S. & FIRTEL, R. A. (1986).
Cellular and sub-cellular distribution of a cAMPregulated prestalk protein and prespore protein in
Dictyostelium discoideum: A study on the ontogeny of
prestalk and prespore cells. J. Cell Biol. 103,
1999-2015.
GUBLER, U. & HOFFMAN, B. J. (1983). A simple and very
efficient method for generating cDNA libraries. Gene
25, 263-269.
KAY, R. R., DHOKIA, B. & JERMYN, K. A. (1983).
Purification of stalk cell-inducing morphogens from
Dictyostelium discoideum. Eur. J. Biochem. 136, 51-56.
KNECHT, D., GREEN, E., LOOMIS, W. F. & DIMOND, R.
(1984). Developmental changes in the modification of
lysosomal enzymes in Dictyostelium discoideum. Devi
Biol. 107, 490-502.
KOPACHIK, W., DHOKIA, B. & KAY, R. R. (1985).
Selective induction of stalk cell-specific proteins in
Dictyostelium. Differentiation 28, 209-216.
KOPACHIK, W., OOHATA, A., DHOKIA, B . , BROOKMAN, J.
J. & KAY, R. R. (1983). Dictyostelium mutants lacking
DIF, a putative morphogen. Cell 33, 397-403.
KREFFT, M., VOET, L., GREGG, J. H . , MAIRHOFER, H. &
WILLIAMS, K. L. (1984). Evidence that positional
information is used to establish the prestalk-prespore
pattern in Dictyostelium discoideum aggregates. EM BO
J. 3, 201-206.
KREFFT, M., VOET, L., GREGG, J. H . & WILLIAMS, K. L.
(1985). Use of a monoclonal antibody recognizing a cell
surface determinant to distinguish prestalk and
prespore cells of Dictyostelium discoideum slugs. J.
Embryol. exp. Morph. 88, 15-24.
LOOMIS, W. F. & KUSPA, A. (1984). Biochemical and
References
BARKLIS, E. & LODISH, H. F. (1983). Regulation of
Dictyostelium discoideum mRNAs specific for prespore
or prestalk cells. Cell 32, 1139-1148.
BARKLIS, E., PONTIUS, E. & LODISH, H. F. (1985).
Structure of the Dictyostelium discoideum prestalk D l l
gene and protein. Molec. cell. Biol. 5, 1473-1479.
BROOKMAN, J. J., TOWN, C. D . , JERMYN, K. & KAY, R.
R. (1982). Development regulation of a stalk cell
differentiation inducing factor in Dictyostelium
discoideum. Devi Biol. 91, 191-196.
CHISHOLM, R. L., BARKLIS, E. & LODISH, H. F. (1984).
Mechanism of sequential induction of cell-type specific
mRNAs in Dictyostelium discodeum. Nature, Lond. 310,
67-69.
DATTA, S., GOMER, R. H. & FIRTEL, R. A. (1986). Spatial
and temporal regulation of a foreign gene by a
prestalk-specific promoter in transformed Dictyostelium
discoideum. Molec. cell. Biol. 6, 811-820.
DEVINE, K. M., MORRISSEY, J. H. & LOOMIS, W. F.
(1982). Differential synthesis of spore coat proteins in
genetic analysis of prestalk specific acid phosphatase in
Dictyostelium. Devi Biol. 102, 498-503.
MACLEOD, C , FIRTEL, R. A. & PAPKOFF, J. (1980).
Regulation of actin gene expression during spore
germination in Dictyostelium discoideum. Devi Biol. 76,
263-274.
MEHDY, M. G., RATNER, D. & FIRTEL, R. A. (1983).
Induction and modulation of cell-type specific gene
expression in Dictyostelium discoideum. Cell 32,
763-771.
MANIATIS, T., FRITSCH, E. F. & SAMBROOK, J. (1982).
Molecular Cloning: A Laboratory Manual. Cold Spring
Harbor, New York: Cold Spring Harbor Laboratory.
MORRISSEY, J. H. (1982). In The Development of
Dictyostelium discoideum (ed. W. F. Loomis), pp.
411-451. New York: Academic Press.
MORRISSEY, J. H., DEVINE, K. M. & LOOMIS, W. F.
(1984). The timing of cell-type specific differentiation in
Dictyostelium discoideum. Devi Biol. 103, 414-424.
OOHATA, A. (1983). A prestalk cell specific acid
phosphatase in Dictyostelium discoideum. J. Embryol.
exp. Morph. 74, 311-319.
Prestalk-enriched mRNAs in Dictyostelium
OYAMA, M. & BLUMBERG, D. D. (1986). Cyclic AMP and
N H 3 / N H 4 + both regulate cell-type specific mRNA
accumulation in the cellular slime mould Dictyostelium
discoideum. Devi Biol. 117, 557-566.
PEARS, C. J., MAHBUBANI, H. & WILLIAMS, J. G. (1985).
Characterization of two highly diverged but
developmentally co-regulated cysteine proteinase genes
in Dictyostelium discoideum. Nucl. Acids Res. 13,
8853-8866.
PEARS, C. J. & WILLIAMS, J. G. (1987a). Identification of
a DNA sequence element required for efficient
expression of a developmentally regulated and cAMPinducible gene of Dictyostelium discoideum. EMBO J. 6,
195-200.
RATNER, D. & BORTH, W. (1983). Comparison of
differentiating Dictyostelium discoideum cell types
separated by an improved method of density gradient
centrifugation. Expl Cell Res. 143, 1-73.
REYMOND, C. D., GOMER, R. H., MEHDY, M. C. &
FIRTEL, R. A. (1984). Developmental regulation of
Dictyostelium gene containing a protein homologous to
mammalian ras protein. Cell 30, 141-148.
REYMOND, C. D., GOMER, R. K., NELLEN, W., THEIBERT,
755
SUSSMANN, M. (1966). Biochemical and genetic methods
in the study of cellular slime mould development.
Meth. Cell Physio!. 2, 397-410.
TAKEUCHI, I., HAYASHI, M. & TASAKA, M. (1977). Cell
differentiation and pattern formation in Dictyostelium.
In Developments and Differentiation in the Cellular Slime
Moulds (ed. P. Cappuccinelli & J. Ashworth), pp.
1-16. Amsterdam: Elsevier/North Holland Biomedical
Press.
TASAKA, M., NOCE, T. & TAKEUCHI, I. (1983). Prestalk
and prespore differentiation in Dictyostelium as
detected by cell type-specific monoclonal antibodies.
Proc. natn. Acad. Sci. U.S.A. 80, 5340-5344.
TOWN, C. D., GROSS, J. D. & KAY, R. R. (1976). Cell
differentiation without morphogenesis in Dictyostelium
discoideum. Nature, Lond. 262, 717-719.
TSANG, A. & BRADBURY, J. M. (1981). Separation and
properties of prestalk and prespore cells of
Dictyostelium discoideum. Expl Cell Res. 32, 433-441.
WATTS, D. J. & ASHWORTH, J. M. (1970). Growth of
myxamoebae of the cellular slime mould Dictyostelium
discoideum in axenic culture. Biochem. J. 119, 171-174.
WILLIAMS, J. G., CECCARELLI, A., MCROBBIE, S. J.,
A., DEVROETES, P. & FIRTEL, R. A. (1986). Phenotypic
MAHBUBANI, H., BERKS, M. M., KAY, R. R. & JERMYN,
changes induced by a mutated ras gene during the
development of Dictyostelium transformants. Nature,
Lond. 323, 340-343.
STERNFELD, J. & DAVID, C. N. (1981). Cell sorting during
pattern formation in Dictyostelium. Differentiation 20,
10-21.
K. A. (1987). Direct induction of Dictyostelium prestalk
induction by DIF provides evidence that DIF is a
morphogen. Cell 49, 185-192.
STERNFELD, J. & DAVID, C. N. (1982). Fate and
regulation of anterior-like cells in Dictyostelium slugs.
Devi Biol. 93,51-61.
WILLIAMS, J. G., PEARS, C. J., JERMYN, K. A., DRISCOLL,
D. M., MAHBUBANI, H. & KAY, R. R. (1986). The
control of gene expression during cellular
differentiation of Dictyostelium discoideum. In
Regulation of Gene Expression (ed. I. Booth & C.
Higgins). Cambridge University Press.