(Received for publication, May 29, 1996, and in revised form, October 22, 1996)
From the Medical Molecular Biology Unit, Department of Molecular Pathology, University College London Medical School, The Windeyer Building, Cleveland Street, London W1P 6DB, United Kingdom
The differentiation of the ND7 neuronal cell line to a nondividing phenotype bearing numerous neurite processes is accompanied by a dramatic increase in the levels of the activating POU family transcription factor Brn-3a and a corresponding fall in the levels of the closely related inhibitory factor Brn-3b. We have previously shown that the artificial overexpression of Brn-3a in these cells can induce neurite outgrowth and the activation of genes encoding synaptic vesicle proteins in the absence of a differentiation-inducing stimulus. Here we show that overexpression of Brn-3b can reduce process outgrowth and synaptic vesicle gene expression following exposure to a stimulus which would normally induce differentiation. These inhibitory effects are abolished by altering a single amino acid in the POU homeodomain of Brn-3b to its equivalent in Brn-3a. The converse mutation in Brn-3a allows it to inhibit process outgrowth in response to a differentiation-inducing stimulus. Hence a single amino acid difference results in these closely related factors having opposite effects and allows the balance between them to regulate differentiation.
The Brn-3a, Brn-3b, and Brn-3c transcription factors (also known
as Brn-3.0, 3.2, and 3.1, respectively) (1, 2) are closely related
members of the POU family of transcription factors which are encoded by
three distinct genes (3, 4, 5, 6) (for review of POU factors see Refs. 7 and
8). Among the mammalian POU factors, the Brn-3 proteins are the most
closely related to the nematode POU factor Unc-86, which plays a
critical role in neuronal development in the nematode (9, 10)
suggesting that the Brn-3 factors may play a similar role in mammals.
In agreement with this idea the three members of the Brn-3 family are
expressed in distinct but overlapping populations of neuronal cells
during development and in the adult organism (1, 2, 3, 4, 5, 11) with Brn-3a for
example being expressed in the first differentiated neurons to appear
in the midbrain, hindbrain, and spinal cord during development (12).
Interestingly Brn-3a has been shown in co-transfection experiments to
activate several promoters including those of the neuronally expressed genes encoding pro-opiomelancortin (1), -internexin (13), and the
presynaptic vesicle protein SNAP-25 (14) as well as a thymidine kinase
(tk)1 promoter containing an added
synthetic binding site for the various forms of Brn-3 (tk-Oct) (15,
18). In contrast Brn-3b represses these promoters (13, 15, 18) and also
inhibits their activation by Brn-3a. Similarly Brn-3c has only a weak
activating effect on these promoters (13, 15, 18).
When the proliferating ND7 cell line (obtained by fusing primary dorsal root ganglion neurons with C1300 neuroblastoma cells) (17) is induced to cease dividing and differentiate to a mature neuronal-like phenotype by treatment with cyclic AMP or removal of serum from the medium (18), the levels of Brn-3a rise dramatically, and those of Brn-3b fall (3, 15), while Brn-3c levels remain unchanged.2 This differentiation process is likely to require the induction of specific genes involved in the outgrowth of neurite processes and the acquisition of a more neuronal-like phenotype. If such genes were targets for activation by Brn-3a and repression by Brn-3b, the changes in Brn-3 expression could potentially play a key role in the differentiation process by activating the expression of such genes. Hence the ND7 cell system offers a convenient model system for studying the role of the Brn-3 factors in neuronal differentiation.
Indeed we have previously used this system to show that the inhibition of Brn-3a expression using an antisense approach prevents the neurite outgrowth and elevated expression of SNAP-25, which normally occur following exposure of ND7 cells to a differentiation inducing stimulus (14). Moreover, the stable overexpression of Brn-3a in ND7 cell clones resulted in neurite outgrowth and the induction of several synaptic proteins including SNAP-25 even in the absence of a differentiation inducing stimulus (19).
Given the ability of Brn-3a to induce the SNAP-25 promoter in co-transfection experiments (14), it is therefore likely that the rise in Brn-3a, which occurs during ND7 cell differentiation, directly activates the SNAP-25 promoter and is thus responsible for the rise in SNAP-25 expression, which occurs during the differentiation event. Moreover, the finding that SNAP-25 expression is essential for neurite outgrowth in several different types of neuronal cells in vitro and in vivo (20) indicates that the exocytosis of synaptic vesicles, which is required for synaptic transmission, and the constitutive exocytosis, which is required for axon outgrowth, may have several components such as SNAP-25 in common. Hence, the induction by Brn-3a of SNAP-25 gene expression and that of other synaptic proteins is likely to be responsible, at least in part, for its ability to induce neurite outgrowth.
Although these findings indicate a critical role for the rise in Brn-3a expression in inducing neuronal differentiation in ND7 cells, they also focus attention on the role of Brn-3b. Thus the levels of Brn-3b fall dramatically during ND7 cell differentiation (3, 15), while in co-transfection experiments it both represses the basal activity of a number of different promoters, which are activated by Brn-3a, and prevents their activation by Brn-3a (13, 15, 16). We have therefore investigated the response of ND7 cell clones overexpressing Brn-3b to a stimulus which would normally induce differentiation and have compared this to the response of clones overexpressing either Brn-3a or Brn-3c.
cDNA clones of wild type and mutant forms of
Brn-3a, Brn-3b, and Brn-3c, as well as the POU domains of Brn-3a and
Brn-3b, were inserted in the sense orientation under the control of the dexamethasone-inducible mouse mammary tumor virus (MMTV) promoter of
the mammalian expression vector PJ5 (21). Following co-transfection into ND7 cells (17) together with a plasmid conferring neomycin resistance (pM5G-NEO), stable transfectants were selected by the supplementation of culture medium with G418 to a final concentration of
800 µg/ml, and individual clones were isolated after 7-14 days of
selection. Putative clones were treated with dexamethasone at a final
concentration of 1 µM for 24 h to induce expression of the MMTV promoter to allow screening for clones capable of expressing the exogenous construct. RNA was isolated from cells by the
guanidinium thiocyanate method (22), treated with DNase to remove any
contaminating DNA, and subsequently used as the template for cDNA
synthesis. Resultant cDNA was amplified by PCR essentially as
described by Kawasaki (23). In initial screening experiments to confirm
that exogenous constructs were producing sense mRNAs in the cell
lines, PCR was performed using a forward primer internal to the POU
domain (Brn-3a, 5-GACTCGGACACGGACCCCGCG-3
; Brn-3b,
5
-GACGTGGATGCAGACCCGCGG-3
; Brn-3c, 5
-GATGTGGAGTCAGACCCTCGA-3
) and a
reverse primer internal to the vector sequence
(5
-AGATCTGGTACCATCGAT-3
) so as not to amplify endogenous
Brn-3mRNA, with product detected by Southern hybridization with
homologous probes. cDNA from clones capable of expressing the
exogenous construct were subsequently subjected to PCR to determine the
total level of expression of that member of the Brn-3 family by using
forward and reverse primers internal to the POU domain thus amplifying
both exogenous and endogenous mRNAs (Forward primers as above.
Reverse primers: Brn-3a, 5
-CCCTCCTCAGTAAGTGGC-3
; Brn-3b,
5
-CTAAATGCCGGCAGAGTATTTCAC-3
; Brn-3c, 5
-CAATCGTCCACAGCAGAGTATTT-3
).
Sample pairs were equalized using primers designed to amplify the
mRNA species encoding the invariant L6 ribosomal protein
(5
-ATCGCTCCTCAAACTTGACC-3
and 5
-AACTACAACCACCTCATGAA-3
). In all
cases expression was studied in the presence or absence of
dexamethasone in cells either proliferating in full serum-containing
medium or induced to differentiate by growth in serum-free medium.
In all cases cultures
differentiated in the absence of dexamethasone were analyzed in
parallel with dexamethasone-induced experimental differentiated
cultures, and both subsequently compared to lines expressing the PJ5
vector with an empty expression cassette induced to differentiate in
the presence or absence of dexamethasone. To determine neurite process
length cells were subject to immunocytochemistry using a primary
monoclonal antibody to -tubulin detected by the peroxidase/diaminobenzidene color reaction. The length of the longest
process in each of 400 cells was determined for each sample using video
capture NIH Image software (version 1.59) with the operator unaware of
the nature of the sample. Statistical analysis was performed using
Microstat software to compare the means of triplicate determinations.
In all cases the paired t test was used to compare different
cell lines either before or after steroid treatment while the Mann
Whitney test was used to compare data in a single set of cell lines
before and after steroid treatment.
cDNAs were first subject to
PCR designed to amplify the mRNA species encoding the L6 ribosomal
proteins (see above). Following subsequent Southern hybridization to
allow the equalization of cDNA input between both uninduced and
induced control and experimental samples, PCR amplification using
primers to specifically amplify the mRNA species encoding
SNAP25 (5-TGACCAGCTGGCTGATGAGTC-3
and
5
-CCCATGTCTAGGGGGCCATATGA-3
) and GAP43
(5
-GTGCTCTGGTTTCCTTAGCA-3
and 5
-GATATAGCCCTCATCCATCA-3
) was
performed. Similar amplifications were also performed, using primers
designed to specifically amplify synapsin Ia
(5
-ATGGAGACTACCGCAGTTTG-3
and 5
-CACAACTACAGGGTATGTTG-3), synaptotagmin I (5
-TCGCCCATGGCTGTGGTTGCT-3
and
5
-CTGGAAATCATAGTCCAGAG-3
) and synaptophysin I
(5
-CAGAAACAAGTACCGAGAA-3
and 5
-CCAAACACCACTGAGGTGTT-3
). All
amplifications were carried out using levels of mRNA and numbers of
cycles, which had been shown in previous control experiments to
give levels of product that were linearly related to the amount of
input RNA (data not shown). Following amplification, products were run
on agarose gels and hybridized with appropriate probes for each
amplification product.
The cell doubling rate of experiment and control cultures in serum-free medium was determined, allowing the construction of growth curves, while the S-phase proliferative fraction of each culture was determined by bromodeoxyuridine (BrdUrd) incorporation and subsequent immunocytochmical detection using a BrdUrd labeling and detection kit (Boehringer Mannheim) according to the manufacturer's instructions. The number of positive nuclei in a population of 500 cells per sample was determined for the purpose of statistical analysis.
When ND7 cells are induced to differentiate by removal of serum,
the level of Brn-3a rises and that of Brn-3b falls (Fig. 1a). To test the effect of Brn-3b on the
differentiation of ND7 cells we used cell lines prepared by stably
transfecting ND7 cell lines with cDNA clones for Brn-3a, Brn-3b or
Brn-3c under the control of the dexamethasone inducible MMTV promoter,
allowing us to regulate the level of Brn-3 in each cell line by steroid treatment. Three cell lines of each type selected on the basis of their
neomycin resistance (encoded on a co-transfected plasmid) were used to
control for clonal variation.
Each showed basal and steroid inducible expression of the appropriate transfected form of Brn-3 in cells proliferating in full-serum containing medium as assayed by PCR using one primer specific for the RNA transcript derived from a transfected plasmid and one primer specific for the appropriate form of Brn-3 (Fig. 1b). Similarly increased basal levels and steroid inducibility of each form of Brn-3 in the proliferating cells could be demonstrated by PCR using primers which would specifically amplify both the endogenous and exogenous mRNAs encoding each form of Brn-3 (Fig. 1b). Most importantly the levels of the exogenous Brn-3 in each cell line were unaffected by the induction of cell differentiation by removal of serum. Thus the fall in endogenous Brn-3b that occurs in these circumstances (see Fig. 1a) in all the cells resulted in the Brn-3b-transfected cells having a much higher level of Brn-3b than control cells following exposure to a differentiation-inducing stimulus (Fig. 1b). Moreover, the differentiated Brn-3a-transfected cells still maintained a higher total level of Brn-3a than control cells (Fig. 1b) despite the rise in endogenous Brn-3a levels that occurs upon differentiation (see Fig. 1a) (3, 15). Similarly, enhanced levels of Brn-3c were also observed in the Brn-3c transfected cells (Fig. 1b).
We therefore investigated the effect of Brn-3a, Brn-3b, or Brn-3c over
expression on the outgrowth of neurites which normally occurs following
removal of serum from the ND7 cells (18). Cells were induced to
differentiate by removal of serum and the effect of each form of Brn-3
assessed. As shown in Fig. 2 and Table I the Brn-3a- and Brn-3c-expressing cell lines exhibited enhanced neurite
length compared to control cells following the induction of
differentiation and exposure to steroid treatment to fully induce
expression of Brn-3a or Brn-3c (p < 0.05 for Brn-3a or Brn-3c cell lines compared to vector, in a paired t test).
This extends our previous conclusion (19) that Brn-3a and to a lesser extent Brn-3c can induce neurite outgrowth even in full
serum-containing medium and indicates that exogenous Brn-3a can enhance
outgrowth even in the presence of the enhanced levels of endogenous
Brn-3a produced by the differentiation-inducing stimulus.
|
In contrast however, all three Brn-3b-expressing cell lines showed decreased levels of neurite outgrowth compared to control cells following induction of differentiation even without steroid treatment (Fig. 2 and Table I, p < 0.001 in a paired t test). Following steroid treatment, these cells showed a further reduction in neurite length which was significant both in comparison to the neurite length observed in these cells in the absence of steroid treatment (p < 0.01 in a Mann Whitney test) and by comparison to the neurite length in similarly treated control cells (p < 0.01 in a paired t test) (Fig. 2b and Table I). This resulted in these processes being considerably shorter than those formed by control cells transfected with the pJ5 vector without any insert or those transfected with Brn-3a or Brn-3c. Hence the overexpression of Brn-3b can indeed interfere with neurite outgrowth following exposure of ND7 cells to a differentiation inducing stimulus. Interestingly overexpression of Brn-3a, Brn-3b, or Brn-3c had no effect on the numbers of cells bearing processes (Fig. 2c) indicating that Brn-3b acts by inhibiting the extension of the processes rather than their initial formation. In the case of Brn-3a, this effect is in contrast to its ability to enhance the number of processes formed in the absence of a differentiation inducing stimulus. This indicates that the rise in endogenous Brn-3a levels which occurs upon differentiation is sufficient for maximal numbers of processes to form so that further overexpression of Brn-3a enhances only the length of the processes which form (14, 19).
When ND7 cells are differentiated by transfer to serum-free medium,
process outgrowth is accompanied by a cessation of cell division (17,
18), although this latter effect is not observed in
Brn-3a-overexpressing cells growing in serum-containing medium (19). We
therefore investigated the effect of overexpressing each of the forms
of Brn-3 upon the growth arrest normally caused by serum removal. No
significant alteration in either the growth rate (Fig.
3a) or in the proportion of cells in S phase
as assayed by incorporation of bromodeoxyuridine (Fig. 3b)
was observed in differentiated cells overexpressing Brn-3b. Thus, like
overexpression of Brn-3a in undifferentiated cells, the overexpression
of Brn-3b in differentiated cells has a much larger effect on neurite
outgrowth than upon growth rate.
We have shown that the induction of differentiation by Brn-3a is
accompanied by the enhanced expression of four components of the
synaptic vesicle cycle SNAP-25, synapsin I, synaptophysin, and
synaptotagmin each of which plays a distinct role in different stages
of the cycle (for review, see Refs. 24 and 25). Similarly, differentiation of ND7 cells by serum removal results in enhanced expression of these proteins (Fig. 4a). We
therefore measured the expression of the mRNAs encoding each of
these proteins when the ND7 cell lines expressing each of the forms of
Brn-3 were induced to differentiate.
As shown in Fig. 4b, the Brn-3b-expressing cells showed reduced levels of SNAP-25, synaptotagmin and synaptophysin mRNAs following differentiation and steroid induction of Brn-3b expression, compared to the levels in similarly treated cells transfected with vector or Brn-3a (Fig. 4b). This was confirmed by statistical analysis of quantitative data on mRNA levels obtained by densitometric scanning of the autoradiographs (p < 0.05 comparing mRNA levels in the Brn-3b cells following differentiation and steroid induction with that in similarly treated control cells). Thus Brn-3b overexpression can indeed inhibit some of the genes normally induced during Brn-3a-induced differentiation. However, the levels of synapsin I mRNA were elevated in the differentiated Brn-3b-expressing cells compared to the control cells (Fig. 4b). This elevation in synapsin I expression is consistent with our recent co-transfection experiments which identified the synapsin I promoter as the first promoter that is inducible by all three forms of Brn-3 (26). As expected no change in GAP43 expression was noted in the Brn-3b-expressing cells (Fig. 4b), paralleling our previous findings that neither Brn-3a or Brn-3b alter the level of this protein in undifferentiated ND7 cells or regulate its promoter in co-transfections (19).
Hence as well as reducing the length of neurite processes, which form
upon differentiation, Brn-3b can also inhibit some but not all of the
changes in gene expression that occur during this process. In previous
experiments, using chimeric proteins containing different regions of
Brn-3a or Brn-3b, we have shown that the POU domain plays a key role in
the different effects of these factors on target promoters in
co-transfection assays (15, 16). Thus, although the POU domain of both
factors can bind to DNA (5, 27), only the isolated POU domain of Brn-3a
can activate the tk-Oct promoter (13), whereas the POU domain of Brn-3b
cannot do so. Hence as well as acting as the DNA binding domain, the POU domain of Brn-3a can also act as a distinct activation domain. We
therefore prepared ND7 cell lines overexpressing the isolated POU
domain of Brn-3a or Brn-3b and tested their effect on neurite outgrowth. As shown in Fig. 5 the differentiated cells
expressing the Brn-3a POU domain showed a similar enhancement of
neurite length compared to control cells as did those expressing
full-length Brn-3a (p < 0.05 in a paired t
test comparing Brn-3a POU domain cells with controls). Moreover, the
differentiated cells expressing the POU domain of Brn-3b showed a
similar steroid inducible reduction in neurite length as occurred in
cells expressing full-length Brn-3b (p < 0.01 in a
paired t test comparing Brn-3b POU domain cells with
controls). As expected no effects on the number of processes formed
were observed in the cells overexpressing the POU domain paralleling
the lack of effect of the full-length factors in differentiated cells
(Fig. 6).
Interestingly, as well as affecting neurite length in a similar fashion
to the full-length protein, the POU domains also either enhanced
(Brn-3a) or reduced (Brn-3b) the expression of SNAP-25 in a manner
similar to that observed in differentiated cells overexpressing the
full-length protein (Fig. 7). Hence the Brn-3b POU
domain appears to be able to inhibit process outgrowth and SNAP-25
expression, whereas the Brn-3a POU domain has the opposite effect.
This indicates that one or more of the seven amino acid differences between the POU domains of Brn-3a and Brn-3b (3, 16) plays a critical role in this effect. Six of these differences are in the relatively nonconserved linker region joining the POU-specific and POU homeodomains, which together make up the POU domain (7, 8). We have therefore focused our attention on the single difference in the POU homeodomain at position 22 and have shown that this plays a critical role in the different effects of Brn-3a and Brn-3b on a target promoter in co-transfection assays. Thus alteration of the isoleucine at this position in Brn-3b to the valine found in Brn-3a allows the mutant Brn-3b to activate the tk-Oct promoter (28).
We therefore tested the effect of overexpressing mutant forms of full-length Brn-3a or Brn-3b in which the amino acid at this position has been mutated to its equivalent in the other factor. In these experiments, mutant Brn-3b with valine instead of isoleucine at this position failed to inhibit neurite outgrowth (Fig. 5) in the differentiated cells compared to the levels observed in control differentiated cells. Hence the mutation of a single amino acid in Brn-3b results in its losing the ability to inhibit neurite outgrowth (p < 0.05 for Brn-3b valine compared to wild type Brn-3b in a paired t test). Similarly, the mutant Brn-3b also failed to inhibit SNAP-25 expression in the differentiated cells (Fig. 7).
Moreover, cells expressing Brn-3a with isoleucine instead of valine at position 22 showed a much smaller enhancement of SNAP-25 expression than did cells expressing intact Brn-3a (Fig. 7). More dramatically, these cells not only failed to show the enhancement of process formation characteristic of Brn-3a but actually showed a steroid-inducible reduction in process formation compared to control differentiated cells which was almost as large as that observed in cells expressing Brn-3b (Fig. 5). Thus the mutant Brn-3a isoleucine shows a dramatically different effect on neurite outgrowth compared to wild type Brn-3a (p < 0.01 for Brn-3a isoleucine compared to Brn-3a in a paired t test). Hence alterations at a single amino acid can dramatically affect the ability of Brn-3a or Brn-3b to regulate ND7 cell differentiation.
Our previous studies (14, 19) have shown that the overexpression of Brn-3a can enhance both the number and length of neurite processes, which form in ND7 cells under conditions that do not normally promote differentiation, and can also activate the expression of several genes encoding proteins involved in the synaptic vesicle cycle. The differentiation process in ND7 cells, however, normally involves both a rise in the level of Brn-3a and a fall in the level of Brn-3b (3, 15). In the experiments reported here we have investigated the role of Brn-3b in this process and have shown that, under conditions which normally promote differentiation, overexpression of Brn-3b can reduce the length of the neurites that form and reduce the expression of some but not all of the genes that are activated during Brn-3a-induced differentiation.
These findings indicate therefore that regulation of ND7 cell differentiation involves a balance between the differentiation-inducing factor Brn-3a and the inhibitory factor Brn-3b. As Brn-3a and Brn-3b are expressed in distinct but overlapping patterns in the developing and adult nervous system (1, 2, 3, 5, 11, 12), it is possible that they play a similar role in regulating neurite outgrowth during development and in response to nerve injury in the adult.
Although further studies will be required to confirm this possibility,
our findings clearly establish Brn-3b as the first transcription factor
to have an inhibitory effect on neurite outgrowth. Our recent
identification of the gene encoding the neuronal nicotinic acetylcholine receptor 2 subunit as the first example of a
gene whose promoter is activated by Brn-3b and not Brn-3a (29, 30) raises the possibility that Brn-3b might inhibit neurite outgrowth by
activating the expression of other genes whose products play a directly
inhibitory role in outgrowth.
However, it is more likely that the inhibitory effect of Brn-3b is brought about by its ability to repress the activity of specific genes. Thus in co-transfection experiments we have previously shown that Brn-3b can repress the basal activity of several different promoters, which are inducible by Brn-3a, and can also interfere with their activation by Brn-3a (13, 15, 16). Moreover, in the experiments reported here the inhibition of neurite outgrowth was accompanied by a reduction in SNAP-25, synaptophysin, and synaptotagmin expression in the Brn-3b-overexpressing cells.
The ability of Brn-3b to prevent the rise in SNAP-25 expression which normally occurs during ND7 cell differentiation is of particular interest since this presynaptic vesicle protein (31) has itself been shown to be necessary for process formation by different neuronal cell types both in vitro and in vivo (20). It is likely therefore that the ability to inhibit the expression of SNAP-25 and potentially of other genes whose protein products are involved in the process of neurite outgrowth underlies the inhibitory effect of Brn-3b on process formation.
This association of the inhibitory effect of Brn-3b on specific gene expression and on neurite outgrowth is further supported by the critical role of the amino acid at position 22 in the homeodomain in this process. Thus alteration of the isoleucine at this position to the valine in Brn-3b prevents the inhibitory effect of Brn-3b not only on a co-transfected promoter (28) but also, as shown here, on endogenous SNAP-25 expression and on neurite outgrowth. Moreover, the reciprocal substitution in Brn-3a not only abolishes its ability to enhance neurite outgrowth but actually allows it to inhibit neurite outgrowth during differentiation.
We previously suggested (28) that Brn-3b may inhibit Brn-3a-induced genes simply by binding to the DNA and being unable to activate transcription. It would therefore passively inhibit gene expression by preventing Brn-3a from binding to the DNA. In this model position 22 in the homeodomain which is known to have no effect on DNA binding (28) would be involved in allowing Brn-3a to act as an activator perhaps by allowing Brn-3a to recruit a second activating factor (for further discussion, see Dawson et al. (28)), while Brn-3b could not do so. However, the finding that mutation of this position in Brn-3a allows it to actually inhibit neurite outgrowth suggests that Brn-3b (and the mutant Brn-3a) may actually act as an active repressor by interacting with wild type Brn-3a or the basal transcriptional complex (for discussion of the mechanisms of transcriptional repression see Refs. 32 and 33). In this case isoleucine at this position would allow this inhibitory interaction to occur, whereas valine would not do so.
Whatever the precise mechanism of this effect, however, it is already clear that Brn-3b is the first example of a transcription factor which can inhibit neurite outgrowth. The further analysis of this factor and its interaction with Brn-3a should greatly enhance our understanding of the mechanisms that regulate neurite outgrowth in development and may potentially have therapeutic implications for the treatment of human spinal injury when nerve regeneration is very poor.
We thank Tarik Moroy for the gift of Brn-3 cDNA clones.