Multiple phenotypes associated with Myc-induced transformation of chick embryo fibroblasts can be dissociated by a
basic region mutation
Multiple phenotypes associated with Myc-induced transformation of chick embryo fibroblasts can be dissociated by a basic region mutation
D. H.
Crouch*
,
R.
Gallagher
1
,
C. R.
Goding
2
,
J. C.
Neil
1
and
R.
Fulton
1,+
Beatson Institute for Cancer Research, CRC Beatson Laboratories, Garscube
Estate, Switchback Road, Bearsden,
Glasgow
G61 1BD,
UK
,
1
Molecular Oncology Laboratory, Department of Veterinary Pathology, University of
Glasgow, Garscube Estate, Switchback Road, Bearsden,
Glasgow
G61 1QH,
UK
and
2
Marie Curie Research Institute, The Chart,
Oxted
, Surrey RH8 OTL,
UK
Received April 25, 1996;
Revised and Accepted July 8, 1996
ABSTRACT
Chimaeric alleles were constructed to assay the biological functions of an N-terminal deletion and C-terminal mutations which were found in a naturally occurring mutant
of feline vMyc, T17. The mutant alleles were assayed for their ability to
transform chick embryo fibroblasts
in vitro
by a number of criteria, namely the ability to induce morphological
transformation, an accelerated growth rate and growth in soft agar. Feline cMyc could transform
the avian cells, whilst T17 vMyc could not, and the N-terminal deletion was responsible for conferring the primary
transformation defect on the mutant protein. The C-terminal mutations which consist of a point mutation adjacent to the
nuclear localisation signal and a point mutation/amino acid insertion within
the basic region (BR) could, however, dissociate the Myc-induced parameters of transformation. This effect was a specific function
of the BR mutation alone, and the mutation could be transferred into avian cMyc
with comparable biological consequences. The BR mutation did not disrupt the
sequence specific DNA binding activity of the protein
in vivo,
despite exerting a biological effect. These data suggest a novel phenotype where
the mutation may affect a subset of Myc-regulated genes through altered DNA binding specificity or protein-protein interactions.
INTRODUCTION
Deregulated expression of c-
myc
is a common feature of the neoplastic phenotype, and can occur through a number
of mechanisms including retroviral insertion, retroviral transduction and
chromosomal amplification (
1
). Myc is believed to function as a sequence specific transcription factor in
association with another protein, Max (
2
), regulating genes which are crucial to the control of cell growth,
differentiation and cell death (
3
-
6
). Both Myc and Max belong to the basic-helix-loop-helix-leucine zipper (B-HLH-LZ) class of transcription factors (
7
). Dimerisation with Max mediated through the HLH-LZ domain is a prerequisite for sequence specific DNA binding, which is itself
mediated by the basic region (BR) (
8
,
9
). After binding to specific target sequences, transactivation can occur through
multiple domains which have been mapped to the N-terminus of Myc (
10
). Recent data are consistent with a model in which the Myc-Max heterodimer is the active transforming complex (
3
,
5
).
The product of the c-
myc
protooncogene, cMyc, is transforming, although some virally transduced forms (vMyc) contain point mutations which
potentiate their activity
in vitro
(
11
,
12
). More recently, point mutations have been detected in naturally occurring tumours in Burkitt's lymphoma (
13
-
15
), suggesting that the point mutations may contribute to deregulating Myc
function
in vivo
. Point mutations may alter the gene regulatory properties of the cMyc
oncoprotein by, for example, altering sequence specific DNA binding activity (
16
) or interfering with the transactivation properties of the protein (
13
,
17
). It is worthwhile noting that the point mutations identified to date in
Burkitt's lymphoma are clustered within the transactivation domain (
15
).
A naturally occurring mutant of cMyc, T17 vMyc, was originally isolated from a
feline T-cell lymphosarcoma as a component of a virus mixture, which also contained
FeLV helper virus and a retrovirally transduced copy of the T-cell receptor (
18
,
19
). T17 vMyc was shown to be the major oncogenic component of this virus mixture,
since secondary tumours induced in neonatal cats using this virus mixture were
shown to contain clonally integrated T17 vMyc DNA, and only trace amounts of T-cell receptor (
18
). We have shown that whilst T17 vMyc may be leukaemogenic
in vivo
, it is non-transforming
in vitro
, suggesting that it represents a novel and potential tissue specific form of
the Myc protein (
20
).
T17 vMyc contains, in addition to a large N-terminal deletion (amino acids 49-124) which spans the transactivation domain, two other mutations. A
single point mutation (D
328
-> G) lies adjacent to the nuclear localisation signal (NLS), and a point
mutation/amino acid insertion occurs within the BR (L
362
-> FR). We set out to establish the role of the individual point mutations in
T17 vMyc, by measuring their biological activity in an
in vitro
system based on the transformation of chick embryo fibroblasts (CEF). We show
that whilst the N-terminus is essential for all measurable biological functions of Myc in
this system, the BR mutation (L
362
-> FR) can dissociate these functions by promoting anchorage independent
growth and the morphological changes associated with Myc overexpression,
without inducing an accelerated growth rate. The role of this BR mutation
within the context of Myc function is discussed.
MATERIALS AND METHODS
Cell culture and transfections
Secondary chick embryo fibroblasts (CEF) were grown at 41oC in DMEM supplemented with 10% tryptose phosphate (TP), 5% newborn calf
serum and 1% chick serum. CEF (2 * 10
5
cells) were transfected with 10 [mu]g of each SFCV construct, essentially as described previously (
22
) together with 4 [mu]g of the replication competent avian retroviral vector RCAN (
44
). Selection with G418 (1 mg/ml) was applied a few days after transfection, and
transformation assays performed on the resulting G418-resistant cells. The ability to grow in soft agar was assayed after 2 * 10
5
cells were seeded in 0.35% agar for 2 weeks. An accelerated growth rate was
measured by plating 2 * 10
5
cells in a 35 mm dish at 41oC and cumulative cell counts performed each day. The growth curves
presented have been reproduced in a number of independent experiments.
Western blot analysis
Western blot analysis of G418-resistant CEF was performed essentially as described (
22
) using a monoclonal antibody to the C-terminus of cMyc (the gift of K. Moelling, Zurich) at a concentration of
1:1000. Proteins were visualised using enhanced chemiluminescence (Amersham).
Oligonucleotide directed mutagenesis
Site directed mutagenesis was performed using the Amersham
in vitro
mutagenesis system, essentially as recommended by the manufacturers, except
that a primer concentration of 10 pmol/[mu]g was used. The mutagenic oligonucleotide (5'-CGAACGCACAACGTC
TTTCGG
GAGCGCCAGCGAAGGAAT-3') was designed to introduce the L
362
-> FR mutation into the avian cMyc background, and all mutations were
confirmed by sequencing.
Yeast strains, transformations and
[beta]
-galactosidase assays
The yeast strains, transformations and [beta]-galactosidase assays were performed essentially as described (
9
,
31
). Briefly, the C-termini of feline cMyc and T17 vMyc containing the B-HLH-LZ domains were fused to the Pho4 transactivation domain in the vector, p MA132, while Max and Max9 were cloned
into the centromeric plasmid, pRS314. Dimerisation and DNA binding
in vivo
were measured by co-transfection of the plasmids into the yeast strain Y700 essentially as
described (
31
). The resulting [beta]-galactosidase activity is representative of dimerisation and DNA
binding mediated by the C-terminus of Myc and Max.
RESULTS
Cloning and expression of feline T17 vMyc and chimaeras
A diagrammatic representation of feline cMyc, T17 vMyc and the chimaeras is
outlined in Figure
1
A. When compared with cMyc, T17 vMyc contains, in addition to the large N-terminal deletion (amino acids 49-124), a point mutation adjacent to the NLS (D
328
-> G) and a point mutation/amino acid insertion within the BR (L
362
-> FR). Chimaeras were constructed by exchanging the appropriate restriction
fragments at the
Bgl
I site, resulting in c/vMyc which contains the point mutation adjacent to the
NLS together with the BR mutation, and v/cMyc which contains only the N-terminal deletion. All mutations were verified by plasmid sequencing on
both strands.
Dissociation of Myc-mediated functions by a mutation in the basic region
Having confirmed their efficient expression in CEF, the transforming potential
of the chimaeras was determined. A number of parameters [an accelerated growth
rate (Fig.
2
A), morphological transformation (Fig.
2
B) and anchorage independent growth (Fig.
2
C)] were assayed. In agreement with others (
23
,
24
), the N-terminus was shown to be essential for all Myc functions tested (Fig.
2
A, B and C), since only cMyc and c/vMyc produced any detectable biological
activity
in vitro
. c/vMyc, however, could dissociate the Myc-mediated functions by inducing both a morphological change and growth in
agar (Fig.
2
B and C) without inducing an accelerated growth rate (Fig.
2
A). The number and size of the colonies in agar were, however, slightly reduced
with c/vMyc compared with cMyc, presumably reflecting the slower growth rate of
cells carrying this gene. c/vMyc and cMyc undergo similar rates of apoptosis
(data not shown), suggesting that the reduced growth rate of c/vMyc compared
with cMyc is not due to an increased rate of apoptosis which is associated with
Myc overexpression under appropriate culture conditions (
25
).
The basic region mutation is sufficient to dissociate Myc-induced parameters of transformation
Immunocytochemical analysis showed that each of the mutant proteins localised to
the nucleus, suggesting that the mutation (D
328
-> G) adjacent to the NLS did not interfere with this intrinsic property of
the Myc protein (
26
; D. Crouch, data not shown). Comparisons of the sequences of the BR from a
number of related proteins show an extremely high degree of conservation within
this region (Table
1
). We reasoned that if the BR mutation (L
362
-> FR) alone was sufficient to dissociate Myc function, the mutation should
be transferable into an avian cMyc background with similar consequences. This
construct would also test the possibility that the feline Myc background
contributes to the dissociated phenotype in avian cells. Site directed
mutagenesis was used to introduce this mutation into avian cMyc, which was
subsequently cloned into SFCV sa
-
(
27
). The resulting construct, avian c/vMyc, was assayed for transforming potential
in CEFs. The results, summarised in Table
2
, show that the BR mutation alone in an avian cMyc background is sufficient to
dissociate the Myc-induced parameters of transformation, since the growth promoting functions
of Myc can be dissociated from the morphological changes and anchorage
independent growth. These data confirm that this effect is not limited to the
feline background, and that the specific BR mutation has general and
reproducible biological effects.
Amino acid sequence homology between the basic regions of myc family members
(20,32) and Max (29)
Amino acids in one-letter code are numbered according to the full-size feline cMyc. Numbering 1-15 identifies the amino acids when plotted on a helical wheel
(see Fig. 3A). The conserved amino acids (Lys-3, Arg-4, His-7, Asn-8, Glu-11, Arg-12, Arg-14 and Arg-15) are boxed and, in Max, have
been shown to make contact with either DNA bases or the phosphate backbone
(29).
A mutation (L
362
-> FR) within the feline T17 vMyc basic region (20) transferred into the
avian cMyc background results in dissociation of the multiple phenotypes
associated with Myc-induced transformation in CEF
Morphological
Growth in
Accelerated
transformation
agar
growth rate
Avian cMyc
+
+
+
Avian c/vMyc
+
+
-
Vector
-
-
-
E-box mediated DNA binding is unaffected by the basic region mutation
in vivo
The BR is essential for mediating sequence specific DNA binding through an E-box motif (
8
,
9
), and it has been proposed that homologous amino acids of Myc and Max basic
regions make equivalent contacts in their respective protein-DNA complexes (
28
). Given the stringent requirement for conserved amino acids within the BR, we
have projected the amino acids onto a helical wheel to compare the basic
regions of cMyc, T17 vMyc and Max (Fig.
3
A). To take into account the displacement which occurs due to the extra amino
acid insertion in the T17 vMyc basic region, the helical wheels have been
plotted to align Leu-10
in cMyc and Arg-10 in T17 vMyc with the equivalent Leu residue in the Max basic region.
Conserved amino acids which participate in DNA binding (boxed residues in Table
1
; circled residues in Fig.
3
A) lie on one face of the [alpha]-helix, and in the Max-DNA crystal structure make contact with either DNA bases
(His-7, Arg-15 and two contact point for Glu-11) or the phosphate backbone (Lys-3, Arg-14, Arg-4, Asn-8 and Arg-12). Within Max and Myc (Fig.
3
A) and other CACGTG-binding proteins (
9
), six residues are most highly conserved (Lys-3, Arg-14, His-7, Glu-11, Arg-4 and Arg-15). We had initially assumed that the amino
acid insertion in the T17 vMyc BR would result in disruption of the [alpha]-helical structure. It was, therefore, a surprise when the helical
wheel analysis of the T17 vMyc BR showed that of the key residues along the
face of the [alpha]-helix, only three amino acid changes resulted from the amino acid
insertion: Lys-3, His-7 and Asn-8 become Arg-3, Asn-7 and Val-8, respectively. In the Max-DNA crystal structure (
29
), Lys-3 and Asn-8 are proposed to make contacts with the phosphate backbone, whilst
in both Myc and Max, His-7 has been shown to contact a DNA base (
28
,
29
). It should be noted that Glu-11 which is crucial for DNA binding (
30
), is retained in the equivalent position in the T17 vMyc BR, and the
hydrophobic Phe-9 in T17 vMyc is located at an equivalent position to the hydrophobic Val-9 in cMyc.
DISCUSSION
Analysis of mutations of cMyc, either naturally occurring or experimentally
designed, has yielded useful information about this important oncoprotein, providing strong links between both the biological
function of the protein and its molecular mechanism of action as a transcription factor (
5
,
32
). In this study, we have investigated the role of the N- and C-terminal mutations contained within a novel mutant of vMyc, T17, to
assess the relative contribution of each to its biological activity
in vitro,
and to relate this to its function as a transcription factor.
Compared with cMyc, the T17 vMyc mutant has certain defects: firstly, at the N-terminus, it contains a large deletion spanning the transactivation
domain, and secondly, at the C-terminus, it contains a point mutation adjacent to the NLS and a point
mutation/amino acid insertion within the basic region (BR), the domain
responsible for mediating sequence specific DNA binding (
8
,
9
). We have previously shown that whilst this mutant is oncogenic
in vivo
, it is non-transforming in an
in vitro
transformation system (
20
). We now show that the individual T17 vMyc mutations have a differential effect
on the biological activity of the protein. In agreement with other workers (
23
,
24
), we show that an intact N-terminus of Myc is essential for all the transformation parameters assayed
in this system, and the mutation adjacent to the NLS does not affect nuclear
localisation (D. Crouch, data not shown). However, by inducing the
morphological changes and anchorage independent growth without inducing an
accelerated growth rate, the BR mutant (c/vMyc) can dissociate the multiple
phenotypes associated with Myc-induced CEF transformation, without affecting the ability of the mutant
protein to bind to a consensus CACGTG motif
in vivo
or
in vitro
(
20
). The normal growth rate of this mutant cannot be explained by an increased
rate of apoptosis, since c/vMyc induces apoptosis at an equivalent rate to cMyc
(D. Crouch, data not shown). The ability of this mutation to dissociate transforming activity
in vitro
suggests that it is biologically relevant. Dissociation of Myc function has previously been described in avian myogenic cells (
27
), where a mutant of cMyc, cMyc[Delta]7, is capable of transforming myoblasts, but is unable to block myogenic
differentiation. In that study, however, the dissociation may be attributable
to selective binding of the mutant protein to Max, since the particular mutant
was defective in the LZ domain. We have shown, however, that T17 vMyc can
dimerise with Max and Max9 to the same degree as wild type cMyc.
T17 vMyc may transform cells in a T-cell specific manner, and the mutations may provide clues to the cell-type-specific regulatory domains of the protein required for this.
Cell-type-specific functions of Myc have previously been reported, and mutants have
been identified that can transform macrophages but not fibroblasts (
33
,
34
). The functional consequences of these mutations are, however, not known. Of
the three transactivation domains mapped to the N-terminus of Myc using transient Gal4 fusion experiments (
10
), only one is retained in T17 vMyc, which may conceivably perform a specific
function in T-cells. Two highly conserved elements, Myc Box I (MBI, amino acids 1-104) and Myc Box II (MBII, amino acids 104-143) are found in the N-terminus of Myc (
23
), and MBII has been shown to play a more major role in transformation than MBI,
since it is more sensitive to mutation (
23
). A minimal boundary of MBII which is essential for
myc
-
ras
cotransformation has been defined between amino acids 129-145 (
35
). Recent studies have attempted to dissect the gene regulatory properties of
the N-terminus and relate these to the transforming properties of the proteins, resulting in
controversy regarding the precise mechanism of transcriptional action (
36
,
37
). Li
et al.
(
36
) suggest that the ability of MBII to repress transcription through a non-E-box sequence is more relevant to transformation than the E-box-mediated transcriptional activation functions mediated by MBI. A more recent report agrees that MBII
plays a more crucial role in transformation than MBI, but shows that whilst
both domains can mediate E-box-specific transactivation on a bona fide target sequence of Myc, [alpha]-prothymosin, they are differentially affected by the position of the E-box (
37
). Whilst the precise gene regulatory nature of the N-terminus is a matter of debate (
36
,
37
), we have shown that in this cell system, T17 vMyc is clearly transformation-defective even though the N-terminal deletion (amino acids 49-124) removes MBI, whilst MBII (amino acids 129-145) is intact. T17 vMyc may also act in a MBII-dependent dominant negative fashion, by
titrating out the cellular factor which interacts with MBII (
35
). Within this context, it will be interesting to test if T17 vMyc acts as an
inhibitor of cMyc in transformation assays.
Since there are multiple members of the B-HLH-LZ family which recognise the same CACGTG motif (
38
), there must be differential regulation of these proteins, by, for example,
interaction with different partners or subtle differences in DNA binding specificities, which ensure
complete specificity of function by directing the complexes to specific subsets
of target sequences. In isolation, the T17 vMyc BR mutation had a profound
effect on the
in vitro
transformation properties of the protein, whilst, at the molecular level, not
appearing to affect dimerisation and DNA binding
in vitro
or
in vivo
. This biological effect presumably reflects subtle alterations in DNA binding
specificity or affinity, and three key residues proposed to be involved in DNA
binding, and most notably His-7 (
28
,
29
) are altered in T17 vMyc (Fig.
3
A). A recent report has identified the position and number of E-box elements within a bona fide target sequence ([alpha]-prothymosin) as being critical determinants of specificity for Myc in
transcriptional activation (
37
). Assessing the transcriptional activity of T17 vMyc on such a target sequence
may help to define how the BR mutation specifically affects its gene regulatory
properties. A point mutation at an equivalent position to the T17 vMyc BR
mutation in N-Myc can broaden the specificity of DNA binding (
16
), whilst methylation-sensitive sequence-specific DNA binding has been demonstrated for the cMyc BR, but not
two other members of the HLH family (
39
). Different members of the bHLH family have distinct preferences for DNA
sequences (
30
,
40
), and specific sequences flanking the core E-box motif strongly affect binding by Myc-Max heterodimers, but not Max-Max homodimers (
41
,
42
). In addition, the BR can affect the transactivation potential of a protein.
For MyoD, BR mutants have been identified which affect the ability to
transactivate without affecting sequence-specific DNA binding, perhaps by interfering with the correct alignment of
the proteins (
43
). Whatever the subtle changes associated with the T17 vMyc BR mutation, it may result in T17 vMyc being directed to only a subset of target genes,
which, in our system, results in the dissociation of the phenotypes associated
with Myc transformation.
The demonstration here that the N- and C-terminal mutations in T17 vMyc have distinct and independent effects
raises interesting questions regarding the order in which they arose
in vivo
, and their relative importance for the activity of the mutant protein. The C-terminal mutation may be silent in the presence of the severe
transformation defect conferred by the N-terminal deletion, and it arose because the selective pressure to maintain
gene regulatory activity was lost in the tumour. Alternatively, the BR mutation
may have a modulatory effect on the oncogenic effects of the mutant gene
in vivo
. This hypothesis may be tested by analysing the
in vivo
oncogenic potential of the chimaeric genes. However, whatever its origin, the C-terminal BR mutation clearly has an independent effect on the activity of
cMyc
in vitro
, and will be a valuable tool for analysing the pleiotropic functions of this
critical effector of oncogenic transformation.
ACKNOWLEDGEMENTS
We wish to thank David Brighty for helpful discussions, Jen Blake and Wei Li for
reading the manuscript. This work was supported by the Cancer Research Campaign
of Great Britain.
16 Feldmann, T., Alex, R., Suckow, J., Dildrop, R., Kisters-Woike, B. and Muller-Hill, B. (1993) Nucleic Acids Res. 21, 5050-5058.MEDLINE Abstract
17 Gupta, S., Seth, A. and Davis, R.J. (1993) Proc. Natl. Acad. Sci. USA90, 3216-3220.MEDLINE Abstract
18 Terry, A., Fulton, R., Stewart, M., Onions, D. and Neil, J.C. (1992) J. Virol. 66, 3538-3549.MEDLINE Abstract
19 Fulton, R., Forrest, D., McFarlane, R., Onions, D. and Neil, J.C. (1987) Nature326, 190-194.MEDLINE Abstract
20 Fulton, R., Gallagher, R., Crouch,D. and Neil, J.C. (1996) J. Virol. 70, 1154-1162.MEDLINE Abstract
21 Fuerstenberg, S., Beug, H., Introna, M., Khazaie, K., Munoz, A., Ness, S., Nordstrom, K., Sap, J., Stanley, I., Zenke, M and Vennstrom, B. (1990) J. Virol.64, 5891-5902.MEDLINE Abstract
22 Crouch, D.H., Lang, C. and Gillespie, D.A.F. (1990) Oncogene 5, 683-689.
23 Stone, J., Lange, T., Ramsay, G., Jakobovits, E., Bishop, J.M., Varmus, H.E. and Lee, W. (1987) Mol. Cell. Biol.7, 1697-1709.MEDLINE Abstract
*
To whom correspondence should be addressed at present address: University of
Dundee, Biomedical Research Centre, Ninewells Hospital and Medical School,
Dundee DD1 9SY, UK
+
Present address: Department of Biological Sciences, Caledonian University, City
Campus, Cowcaddens, Glasgow G4 OBA, UK
