Feb 24, 1988 - Cancer Instituite and Department ofPathology, Harvard Medical Sclool, Boston, Massachuisetts 021152. Received 24 ..... clonal antibody YA6-259. Samples ..... tion and applications of a highly transmissible murine retrovirus.
MOLECULAR AND CELLULAR BIOLOGY, Aug. 1988. p. 3235-3243 0270-7306/88/083235-09$02.00/0 Copyright © 1988. American Society for Microbiology
Vol. 8. No. 8
Inhibition of NIH 3T3 Cell Proliferation by a Mutant ras Protein with Preferential Affinity for GDP LARRY A. FEIG' AND GEOFFREY M. COOPER2* Department of Biochemistry, Tuifts University School of Medicinie, Bostoni, Massachutsetts 02111,1 and Danba-Farber Cancer Instituite and Department of Pathology, Harvard Medical Sclool, Boston, Massachuisetts 021152 Received 24 February 1988/Accepted 3 May 1988
Substitution of asparagine for serine at position 17 decreased the affinity of rasH p21 for GTP 20- to 40-fold without significantly affecting its affinity for GDP. Transfection of NIH 3T3 cells with a mammalian expression vector containing the Asn-17 rasH gene and a Neor gene under the control of the same promoter yielded only a small fraction of the expected number of G418-resistant colonies, indicating that expression of Asn-17 p21 inhibited cell proliferation. The inhibitory effect of Asn-17 p21 required its localization to the plasma membrane and was reversed by coexpression of an activated ras gene, indicating that the mutant p21 blocked the endogenous ras function required for NIH 3T3 cell proliferation. NIH 3T3 cells transformed by v-mos and v-raf, but not v-src, were resistant to inhibition by Asn-17 p21, indicating that the requirement for normal ras function can be bypassed by these cytoplasmic oncogenes. The Asn-17 mutant represents a novel reagent for the study of ras function by virtue of its ability to inhibit cellular ras activity in vivo. Since this phenotype is likely associated with the preferential affinity of the mutant protein for GDP, analogous mutations might also yield inhibitors of other proteins whose activities are regulated by guanine nucleotide binding. The mammalian rais gene family contains three members, , ras and rasN, which encode similar 21,000-dalton proteins referred to as p21's (for a review, see reference 1). These proteins reside on the inner face of the plasma membrane, bind GTP and GDP with an equally high affinity. and display a weak GTP hydrolysis activity (1). These properties are similar to those of G proteins that transduce signals from a wide variety of cell surface receptors to enzymes which affect the metabolism of second messengers, including adenylate cyclase, cyclic GMP phosphodiesterase, and phospholipases C and A2 and may also directly regulate ion channels (for a review, see reference 21). Based on this analogy, it has been proposed that r-as proteins also act as signal-transducing molecules. The fact that r-as genes activated by point mutations confer upon cultured cells some aspects of an oncogenic phenotype implies that their encoded proteins function in the transduction of signals that regulate cell proliferation. Other pathways are also probably affected, since p21 is expressed in nondividing cells (6. 19) and activated rcas genes can induce neuronal differentiation of PC12 cells (2, 22), concomitant with a cessation of cell
the G-protein model, decreased GTPase would extend the time that p21 remains in the activated GTP-bound state and increased rates of GDP to GTP exchange would favor the formation of the activated GTP-p21 complex. The identification of r-as-related genes in yeasts (RAS) allowed a genetic approach to the investigation of RAS function by analyzing the consequences of disrupting yeast RAS genes. Results of these studies have shown that some RAS function is required for growth (26, 46) and that yeast RAS proteins regulate adenylate cyclase (47). Consistent with the G-protein analogy, adenylate cyclase is activated by GTP but not by GDP-bound RAS protein (17). The powerful genetic approaches available in yeasts, however, are not yet applicable to mammalian cells. Efforts to elucidate the function of ras proteins in mammalian systems have instead concentrated on a comparison of the biochemistry of normal and r-as-transformed cell lines and observation of the effects on cells of microinjected rcas proteins and anti-ras antibodies. Results of these experiments have given indirect support for the mediation by *(ls of some aspects of growth factor action. For example. like growth factors, ras proteins appear to influence phospholipid turnover (3, 18. 28, 34, 49), intracellular pH (23). c-fos expression (44), and DNA synthesis (16, 43). Despite these efforts, however, the physiologic function of ras proteins remains poorly defined. In this report, we describe the isolation of a novel rcas mutant whose expression inhibits the growth of NIH 3T3 cells, apparently as a consequence of interference with endogenous rcas function. This provides an alternative approach to the analysis of r-as function by inhibition at the protein level. Since the biological activity of this mutant appears to be related to an alteration in its specificity of guanine nucleotide binding, analogous mutations might also be used to probe the function of other GTP-binding proteins.
ras
division. G proteins characterized to date have a common structural design. They exist as oligomers made up of ox, ,B, and y subunits (21). In the inactive form. the a subunit is bound to GDP as well as to the a and -y subunits (21). An excited membrane receptor activates the G protein by enhancing the rate of exchange of bound GDP for GTP. Go then dissociates from 13 and y and alters the activity of its target enzyme or ion channel. The ot chain then hydrolyzes its bound GTP to GDP, thereby terminating activation (21). Although no 1 and -y subunits have been found to be associated with p21, p21 regulation by guanine nucleotides appears to be similar to that of Got. For example, the transforming potential of ras genes is activated by single amino acid substitutions that decrease GTPase activity (11, 20, 29, 31. 45) or increase rates of GDP to GTP exchange (13, 27, 41, 50). According to
MATERIALS AND METHODS Mutant isolation. The Asn-17 raSH gene was isolated by randomly mutagenizing a cellular raSH bacterial expression
* Corresponding author. 3235
3236
MOL . CELL . B IOL .
FEIG AND COOPER
vector (pXCR) and screening for GTP-binding mutants with a bacterial colony GTP-binding assay as described previously (14, 15). Briefly, 30 ,ug of pXCR was treated with 100 mM hydroxylamine, which introduces C to T substitutions, for 90 min at 70°C in a reaction volume of 150 p.1. After residual hydroxylamine was removed, the bacterial strain PR13Q was transformed with a portion of the mutagenized DNA and bacterial colonies were selected on kanamycinand ampicillin-containing plates. Plates were replicated, and one copy of the filters was used to assay binding of 10' M [a-32P]GTP to lysed bacterial colonies. The entire ras-coding sequence of the mutant was determined by the dideoxy method (37). DNA clones. To recombine the Asn-17 mutation with other rasH mutants, DNA sequences distal to a unique PvuII restriction site at codon 24 were exchanged. For example, Asn-17 Thr-59 Ser-186 rasH was generated by deleting codons 25 to 189 of pXCR Asn-17 as an -700-base-pair PvuII-BamHI fragment and replacing it with a comparable fragment of pBW1225, a v-rasH (Arg-12 Thr-59) clone containing a cysteine to serine substitution at codon 186 (51). For Asn-17 Thr-59 and Asn-17 Leu-61 rasH, the same approach was used, except that -700-base-pair PvuIIBamHI fragments were derived from bacterial expression constructs bearing these mutations. To assay their biological activities, the coding sequences of ras mutants were isolated from pXCR as -750-base-pair BglII-BamHI fragments and subcloned into the BamHI site of the mammalian expression vector pZIPneoSV(X) (5), such that expression of the rasH and neor genes were both under the control of the viral long terminal repeat. In cotransfection experiments, a 6.6-kilobase (kb) BamHI fragment of a genomic clone of normal or Leu-61 rasH in pBR322 was used (11). pHT25, obtained from G. Vande Woude, contained an 8.0-kb EcoRI-HindIII fragment of v-mos proviral DNA in pBR322 (4). This construct included the viral long terminal repeat as well as sufficient v-mos sequence for biological activity. pF4, containing the entire v-raf proviral genome (36), was obtained from the American Type Culture Collection (Rockville, Md.). pSrcll (38) contained the v-src gene of Rous sarcoma virus under the control of the viral long terminal repeat. Interactions with guanine nucleotides. p21's were expressed in and purified from bacteria as described previously (15). Briefly, bacteria were induced to synthesize p21 by exposure to 5 mM isopropyl-p-D-thiogalactopyranoside for 1 to 2 h. The cells were lysed, and p21 was extracted from the insoluble fraction with 3.5 M guanidine hydrochloride. Binding affinities were determined by incubating 5 to 50 ng of p21 with various concentrations of [a-32P]GTP or [ax-32P]GDP for 60 min at 30°C in 50 p.1 of binding buffer (20 mM Tris hydrochloride [pH 7.21, 0.35 M guanidine hydrochloride, 150 mM NaCl, 1 mM dithiothreitol, 1 mM MgCl2, 40 p.g bovine serum albumin per ml). Nucleotides bound to p21 were trapped by filtering the reaction mixture through nitrocellulose and quantitated by scintillation counting. Nucleotide exchange rates were determined by allowing p21 to bind [ot-32P]GTP or [CL-32P]GDP, as described above. Samples were then diluted into 2 ml of binding buffer containing >1,000-fold excess unlabeled GTP or GDP. The rate at which cold nucleotide replaced labeled nucleotide bound to p21 at either 32 or 0°C was quantitated by filtration
threitol-20 p.g of bovine serum albumin per ml at 37°C for 15 to 120 min. The release of 32p; from GTP was determined by chromatography of 2-,ul samples on polyethyleneimine plates as described previously (11). Transfection assays. NIH 3T3 cells were transfected with 10 to 1,000 ng of cloned DNAs, together with 20 p.g of carrier NIH 3T3 DNA, as described previously (9). Foci of transformed cells were counted 10 to 14 days after transfection. Alternatively, transfected cells were subcultured (1:10 split)
through nitrocellulose. GTPase activity was assayed by incubating 5 p.g of p21 with 5 x 10-9 M [_y-32P]GTP (36 Ci/mmol) in 50 p.1 of 20 mM Tris hydrochloride (pH 7.4)-0.2 mM MgCl2-5 mM dithio-
asparagine.
into medium containing G418 (400 ,ug/ml) 3 days after exposure to DNA. Drug-resistant colonies appearing 14 to 21 days later were either isolated for further analysis or quantitated after they were stained with crystal violet. Immunoprecipitation. Cells were labeled by incubation with 200 p.Ci of [35S]methionine (500 Cilmmol) per ml in methionine-free medium for 18 h. ras proteins were then immunoprecipitated with anti-p21 monoclonal antibody YA6-259, as described previously (10). Proteins were analyzed by electrophoresis in 7.5 to 15% sodium dodecyl sulfate (SDS)-polyacrylamide gels and by autoradiography. Northern blot analysis. Cytoplasmic RNA was isolated by suspending 107 trypsinized cells in 0.25 ml of ice-cold 20 mM Tris hydrochloride (pH 7.2)-i mM disodium EDTA-100 mM NaCl. RNAsin (5 ,ul) was added, cells were lysed by the addition of 24 p.1 of 10% Nonidet P-40, and nuclei were pelleted by centrifugation for 1 min in a microfuge. The supernatant was added to an equal volume of 200 mM Tris hydrochloride (pH 8.0)-0.35 M NaCI-20 mM disodium EDTA-1% SDS and then extracted twice with phenolchloroform-isoamyl alcohol (25:24:1) and once with chloroform. RNA was precipitated with ethanol. For blot hybridization, 15 p.g of RNA was electrophoresed in 1% agarose-formaldehyde gels, transferred to nitrocellulose filters, and probed with a nick-translated, 750-base-pair BglIIBamHI fragment of pXCR. Hybridization was performed in 50% formamide-Sx SSC (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-5 x Denhardt solution-50 mM NaPi (pH 6.8)-0.2% SDS-100 p.g of single-stranded sonicated salmon sperm DNA per ml at 42°C for 18 h. Filters were washed in 2x SSC plus 0.2% SDS at 68°C and autoradiographed. Transcript sizes were estimated by comparison with the migration of rRNAs. RESULTS Isolation of a rasH mutant with preferential affinity for GDP. In an effort to test the hypothesis that p21 function is dependent on the ability to bind GTP in vivo, we were interested in isolating and characterizing a ras mutant with preferential affinity for GDP. Such a mutant was identified by screening a series of rasH GTP-binding mutants for one that retained its affinity for GDP. The GTP-binding mutants were isolated by random mutagenesis of a normal cellular rasH bacterial expression vector with hydroxylamine and screening by a GTP-binding assay performed directly on individual bacterial colonies expressing ras p21 (15). One mutant with the desired characteristics was identified, and its ras-coding sequence was determined. Two A to T substitutions were found, one at position 48 and the other at position 50. The first was in the third position of the Lys-16 codon and thus did not alter the encoded protein. However, the second substitution changed amino acid 17 from sefine to The mutant protein was purified from bacteria, and its equilibrium binding properties were compared with those of normal cellular rasH p21 (Fig. 1A). For Asn-17 p21, the Kd
VOL. 8. 1988
A
~.
INHIBITORY
Normal p21
c
-;o
100 80
0 0
o9
O.
60
0 0
40 0
z
20 10-
B
MUTANT
3237
17 Asn p21 GDP
c
GDP
100
.........GTP 7;0 10 c
-
80
0
/
.TP
60
-
40
z
20
I.I
I
. I. .,i .
10-"
1o-8
[GTP]
ris
or
i0o9
7o-,
10-8
10-6
[GTP] or [GDP] (M)
[GDP] (M)
Normal p21
c
100
CL .
80
0
0.5.rcm X
60
._
E
40
0
20 i07
or
100
OX
80
C,
60 e
_ tvg40 0
GDP GTP
M
20
..... ......
........
106
[GTP]
rc
10-i
[GDP]
(M)
10-"
10-
10-6
[GTP]
or
[GDP] (M)
FIG. 1. Affinity of Asn-17 riasH p21 for GTP and GDP. (A) A total of 10 to 50 ng of normal (O. *) or Asn-17 (0. *) p21 was incubated for 1 h at 30°C with the indicated concentrations of either [o-32P]GTP (U. 0) or [Qx-32P]GDP (EO. 0). Samples were filtered through nitrocellulose, and nucleotide that bound to p21 was quantitated by scintillation counting. Data are presented as the percentage of maximal binding (which was similar for both normal and Asn-17 p21's) and represent the average of at least duplicate experiments. (B) A total of 50 ng of normal (EO, *) or Asn-17 (0. 0) p21 was incubated for 1 h at 320C with 5 x 10-7 M [a-32P]GDP (3,000 Ci/mmol) in the presence of the indicated concentrations of unlabeled GDP (O. 0) or GTP (-. 0). Bound [32P]GDP was assayed by membrane filtration. Data are expressed as the percentage of [32P]GDP bound in the presence of unlabeled nucleotide compared with that in the absence of unlabeled nucleotide.
for GTP binding was found to be 3 x 10' M which represented an -40-fold decrease in affinity compared with that of normal rasH p21 (Kd = 8 10-9 M). In contrast, Asn-17 p21 had a Kd for r32P]GDP binding of 5 x 10-9 M, a value comparable to that observed for the normal rcas protein. These results were confirmed by comparing the affinities of the proteins for GTP and GDP by a competition assay (Fig. 1B). [32P]GDP (5 x 10-7 M) was incubated with Asn-17 p21 or normal p21 in the presence of various amounts of unlabeled GDP or GTP, and the ability of these nucleotides to compete with [32P]GDP for binding to p21 was compared. Unlike normal rasH p21, for which these two nucleotides competed equally, GDP competed -20-fold better than GTP for binding to Asn-17 p21. Thus, in both assays Asn-17 displayed a 20- to 40-fold higher affinity for GDP than it did for GTP. This difference might be an underestimate because of the possibility of GDP contamination of GTP preparax
tions.
Previous analysis of GTP-binding mutants has shown that decreased affinities for nucleotides are associated with enhanced rates of nucleotide exchange (13. 15). We therefore analyzed this property by allowing p21 to reach equilibrium binding with [32P]GTP or [32P]GDP and quantifying the rate at which excess unlabeled nucleotide displaced radioactively labeled nucleotide bound to p21. Although a Asn-17 substitution had no effect on the affinity of p21 for GDP, the rate of
GDP exchange at 32°C was increased at least 30-fold (>1.0/ min for the mutant compared with 0.03/min for normal p21) (Fig. 2A). Similar results were obtained for GTP exchange (Fig. 2A). Because these exchange rates were too rapid to quantitate, the experiments were repeated at 0°C (Fig. 2B). Under these conditions, normal p21 did not show detectable nucleotide exchange. However, for Asn-17 p21, GTP exchange (>0.5/min) was shown to be at least 30 times faster than GDP exchange (0.03/min), a difference in exchange rates comparable to the difference in the affinities of the proteins of these two nucleotides. This difference in GTP and GDP exchange rates indicates that most of the GTP binding assayed was, in fact, due to GTP rather than to contaminating GDP. The GTPase activity of Asn-17 p21 appeared to be reduced compared with that of the normal protein (data not shown). However, the significance of this measurement is not clear, since small amounts of GDP that formed early in the course of the reaction acted as a potent end product inhibitor. Asn-17 rasH p21 inhibits cell proliferation. The biological activity of Asn-17 raSH was assayed by subcloning it into the mammalian expression vector pZIPneoSV(X) (5) and quantitating its transforming activity by transfection of NIH 3T3 cells. Unlike normal raSH, which displayed weak but significant focus-forming activity in this efficient expression vector (10 to 20 foci per ,ug of DNA), Asn-17 raSH had no
3238
~ GTP
MOL. CELL. BIOL.
FEIG AND COOPER
A r100
50 C -0 I
C
CL0 0o0 a.-
10
4
*_ a
, E 5
at
(
4 GDP
B
320°C
s,
10
20
100
0.
50 -0
Ic m0
o
cL
'r
10
O to
E
5
a.t
30 40 50 Time (Min)
60
5
10
15
20
25
30
(Min) FIG. 2. Rate of exchange of guanine nucleotides bound to Asn-17 p21. A total of 50 ng of normal (O, *) or Asn-17 (0, 0) p21 was incubated at 32°C in 50 ,ul of binding buffer containing 10-6 M [a-32PIGTP (H, 0) or [a-32P]GDP (O, 0) for 1 h. A portion of the sample was then diluted into 2 ml of binding buffer containing 5 x 10-4 M unlabeled GTP or GDP and incubated at either 32°C (A) or 0°C (B). The percentage of [32P]GTP or [32P]GDP remaining bound to p21 was determined as a function of time.
measurable transforming activity (
