Oncogene (1997) 15, 1021 ± 1033  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Mutations of critical amino acids a€ect the biological and biochemical properties of oncogenic A-Raf and Raf-1

Elizabeth Bosch, Holly Cherwinski, David Peterson and Martin McMahon

Department of Cell Signaling, DNAX Research Institute, 901 California Avenue, Palo Alto, California 94304-1104, USA

The catalytic domains of the Raf family of protein relaying signals from activated cell surface receptors via kinases (DRaf) di€er in their ability to activate MEK in MEK and MAP kinase to a variety of intracellular vitro and in vivo and in their ability to oncogenically e€ectors (Morrison, 1990; Rapp, 1991; Storm et al., transform mammalian cells. The kinase domain of B-Raf 1990). In mammals, there are three members of the Raf is more active than the equivalent portion of Raf-1 which protein kinase family: Raf-1, A-Raf and B-Raf. All three in turn is more active than A-Raf. In Raf-1 the Raf proteins are composed of a conserved amino- phosphorylation or mutation to aspartic acid of two terminal regulatory region and a carboxy-terminal key tyrosine residues upstream of the ATP binding site protein kinase domain. The regulatory domains of Raf has been demonstrated to signi®cantly potentiate proteins are composed of two conserved regions (CR1 catalytic activity. In A-Raf the analogous amino acids and CR2), with the protein kinase domain (CR3) located are also tyrosine whereas in B-Raf they are aspartic in the carboxy-terminal half of the protein. The CR1 acid. To determine if these di€erences in amino acid region encodes the Ras binding domain of the protein sequence in¯uence the relative catalytic activity of the and a `cysteine ®nger' (Dent et al., 1993; Finney et al., Raf kinase domains we constructed forms of DA-Raf, 1993; Koide et al., 1993; Moodie et al., 1993; Papin et al., DB-Raf and DRaf-1 that encode either aspartic acid 1996; Van Aelst et al., 1993; Vojtek et al., 1993; Warne et [DD], phenylalanine [FF] or tyrosine [YY] at these al., 1993; Zhang et al., 1993). Although CR1 and CR2 positions. These proteins were expressed both in contain regulatory sites of phosphorylation, it appears mammalian cells as fusions with the hormone binding that the CR2 region is the major negative regulator of domain of the estrogen receptor and as epitope-tagged protein kinase activity as a deletion encompassing this proteins in Sf9 insect cells to test their oncogenic and region in Raf-1 is constitutively activated and capable of catalytic potentials. When expressed in Rat1 or 3T3 cells promoting oncogenic transformation of several cell types in the presence of hormone all of the DRaf-1:ER and (Beck et al., 1987; Ikawa et al., 1988; Marx et al., 1988; DA-Raf:ER proteins were transforming with the excep- Stanton et al., 1989). tion of the [FF] form of DA-Raf. In general the [DD] A plethora of polypeptide factors including epider- forms of the DRaf-1:ER and DA-Raf:ER proteins were mal growth factor (EGF), platelet-derived growth the most potently oncogenic which correlated with their factor (PDGF) and nerve growth factor (NGF) elicit ability to elicit activation of the MAP kinase pathway. the activation of Ras proteins into their GTP-bound Consistent with the transformation data, the catalytic state (Avruch et al., 1994; Morrison, 1990; Rapp, activity of the [DD] forms of DA-Raf:ER and DRaf- 1991). Activated Ras recruits Raf-1 to the plasma 1:ER was about ten times greater than the cognate [FF] membrane where it is activated by a mechanism that and [YY] forms of the proteins. By contrast all of the remains poorly de®ned. It has been reported that a DB-Raf:ER proteins were highly transforming and DB- variety of protein kinases including Src family kinases Raf catalytic activity was largely una€ected by mutation (Fabian et al., 1993; Marais et al., 1995) and Protein of the aforementioned aspartic acids to either tyrosine or Kinase C (PKC) (Carroll and May, 1994; Kolch et al., phenylalanine. Similar results were obtained with 1993; Sozeri et al., 1992) as well as lipid co-factors epitope-tagged forms of DA-Raf, DB-Raf and DRaf-1 (Dent et al., 1995; Ghosh et al., 1996) can potentiate expressed in Sf9 cells. These data provide support for the Raf-1 activity. Furthermore, Raf proteins are found in model that key tyrosine residues in the protein kinase complex with members of the 14-3-3 family of proteins domains of A-Raf and Raf-1 are important in the that may both regulate activity and perhaps serve to regulation of catalytic activity. In addition they facilitate the construction of higher order signaling demonstrate that the higher intrinsic activity of B-Raf complexes within the cell (Braselmann and McCor- cannot be explained simply by the presence of aspartic mick, 1995; Freed et al., 1994; Liu et al., 1995; Papin et acids at the analogous positions. al., 1996; Xiao et al., 1995). In addition, recent evidence has suggested a role for dimerization in the Keywords: A-Raf; B-Raf; c-Raf-1; protein-serine/ regulation of Raf-1 catalytic activity (Farrar et al., threonine kinases; oncogenes; transformation 1996; Luo et al, 1996). Once activated Raf-1 can phosphorylate and activate the dual-speci®city protein kinases MEK1 and MEK2, Introduction which in turn phosphorylate and activate the Mitogen Activated/Extracellular ligand regulated (MAP/ERK) The Raf family of protein kinases play a pivotal role in protein kinases (Ahn et al., 1992; Ashworth et al., the control of cell proliferation and di€erentiation by 1992; Avruch et al., 1994; Crews et al., 1992; Dent et al., 1992; Howe and Marshall, 1993; Kyriakis et al., Correspondence: M McMahon 1992; MacDonald et al., 1993; Payne et al., 1991; Received 28 January 1997; revised 13 May 1997; accepted 13 May 1997. Samuels et al., 1993; Wu et al., 1993). Activated MAP Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1022 kinases elicit pleiotropic e€ects within the cell by therefore due simply to the presence of aspartic acid phosphorylating a number of proteins including other residues in the aforementioned positions. These protein kinases, transcription factors (Hill et al., 1993; experiments further illuminate the di€erences between Marais et al., 1993; McCarthy et al., 1997), guanine the three mammalian Raf kinases and may provide nucleotide exchange proteins (Cherniack et al., 1994; clues as to the role of these proteins in the response of

Samuels and McMahon, 1994) and phospholipase A2 cells to di€erent types of extracellular stimuli. (Lin et al., 1993; Nemeno€ et al., 1993). The inappropriate activation of this pathway is believed to be a key step in the process of Ras-mediated Results tumorigenesis in humans. The utility of estradiol-regulated forms of oncogenic Oncogenic transformation of cells by retroviruses A-Raf, B-Raf and Raf-1 (DRaf:ER) in elucidating the encoding the di€erent DRaf:ER proteins biological and biochemical changes that lead to oncogenic transformation has been established. Activa- Using standard techniques of site directed mutagenesis tion of DRaf-1:ER in NIH3T3 cells leads to the rapid we constructed forms of DA-Raf, DB-Raf and DRaf-1 activation of MEK and MAP kinase, regulation of the encoding either aspartic acid [DD], phenylalanine [FF] expression of genes such as Heparin-Binding EGF or tyrosine [YY] at the sites previously identi®ed by (HB-EGF) and ultimately to oncogenic transformation others as being involved in the regulation of Raf-1 (McCarthy et al., 1995, 1997; Pritchard et al., 1995; kinase activity (Fabian et al., 1993; Marais et al., Samuels et al., 1993). 1995). These forms of DRaf were engineered to be We have previously identi®ed biological and expressed either in replication-defective retrovirus biochemical di€erences between the di€erent DRaf:ER vectors as fusions with the hormone binding of the proteins expressed in 3T3 cells. DB-Raf:ER is the most estrogen receptor or as FLAG-tagged proteins in a potent activator of the MAP kinase pathway whereas baculovirus transfer vector as described previously and DA-Raf:ER appears to be the weakest. In vitro, as indicated in Figure 1 (Pritchard et al., 1995). epitope-tagged forms of the Raf kinases expressed in Retrovirus stocks for the infection of target cells were Sf9 cells display di€erences in catalytic activity that derived by transient transfection of Bosc23 cells (Pear re¯ect their activity in 3T3 cells. DB-Raf phosphory- et al., 1993) as described in Materials and methods. lated and activated MEK1 approximately ten times Infection of Rat1 and 3T3 cells with retroviruses more eciently than DRaf-1 and 500 times more encoding each of the DRaf:ER proteins gave rise to eciently than DA-Raf (Pritchard et al., 1995). 5105 puromycin resistant colonies that, in the absence Previous work has identi®ed two key tyrosine of hormone, displayed a ¯at, non-transformed residues (Y340, Y341) in the catalytic domain of Raf- morphology similar to that of the parental cells or 1 as being important in the regulation of protein kinase cells infected with a retrovirus encoding DRaf301:ER, a activity. Phosphorylation or mutation of these residues kinase-inactive form of DRaf-1:ER (Figure 2, 7ICI as to aspartic acid led to constitutive activation of the indicated and data not shown). This result is consistent catalytic activity of Raf-1. By contrast, mutation to with previous observations that, even in pooled phenylalanine had no e€ect on basal activity (Marais et populations, very few cells expressing DRaf-1:ER al., 1995; Fabian et al., 1993). Sequence alignment of the catalytic domains of all three Raf proteins revealed di€erences in amino acid sequence in this region that may begin to explain the di€erences observed in catalytic activity. In A-Raf the analogous amino acids are tyrosines 299 and 300 (Y299 and Y300) but in B- Raf the analogous amino acids are both aspartic acid (D492 and D493) (Ikawa et al., 1988; Marx et al., 1988). We therefore reasoned that, in a manner analogous to Raf-1, mutation of these tyrosines to aspartic acid might potentiate the activity of A-Raf whereas in B-Raf mutation of the corresponding aspartic acid residues to phenylalanine or tyrosine

might abrogate catalytic activity. Figure 1 Schematic representation of DRaf:ER and DRaf:FH6 In order to address these issues we have constructed constructs used in this study. Point mutations were introduced forms of the catalytic domains of Raf-1, A-Raf and B- into plasmids encoding the catalytic domains of Raf-1(DRaf-1), A-Raf (DA-Raf) and B-Raf (DB-Raf) to generate the [YY], [DD] Raf which encode either tyrosine, phenylalanine or and [FF] forms of each of the Raf protein kinases used in this aspartic acid at the aforementioned positions. We have study as described in Materials and methods. The relative expressed these proteins in mammalian cells as ER positions of the mutated amino acids and the VAVK motif of fusion proteins and analysed their biological and the ATP binding site are indicated. After DNA sequence biochemical properties. In addition, we have con- veri®cation, sequences encoding the di€erent DRaf kinase domains were subcloned into pBP3:hbER, a derivative of ®rmed the biochemical analysis by expressing epitope- pBabepuro that encodes the hormone binding domain of the tagged forms of the proteins in Sf9 cells. As predicted, HE14 allele of the human estrogen receptor. As indicated the the [DD] forms of DRaf-1 and DA-Raf display elevated expression of the DRaf:ER genes is promoted by the MLV LTR catalytic activity and enhanced oncogenic potential. By and expression of the puromycin resistance gene is promoted by contrast all of the forms of DB-Raf retain the same the SV40 early promoter. The DRaf kinase domains were also subcloned into the baculovirus transfer vector pVL1393:FH6 to elevated catalytic activity and are potently oncogenic in generate forms of the proteins epitope-tagged with the FLAG mammalian cells. The elevated activity of B-Raf is not epitope and 6 Histidine residues at the amino terminus Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1023 display signs of morphological transformation in the To con®rm the expression of each of the DRaf:ER absence of added hormone. Addition of ICI to Rat1 proteins, cell lysates were prepared from the retrovirus cells expressing the [YY], [FF] or [DD] forms of DRaf- infected Rat1 cells described above that had either been 1:ER for 48 h led to the cells displaying a rounded, untreated (0) or treated with ICI for 4, 24 or 48 h. refractile and spindle shaped morphology which is Western blots were probed with an anti-hbER characteristic of oncogenic transformation (Figure 2N, antiserum to detect the expression of the DRaf:ER

P and R, respectively). DRaf-1[DD]:ER activation elicited fusion proteins. The expression of all of the DRaf:ER the most rapid and striking alterations in cell fusion proteins was readily detectable prior to the morphology. At prolonged times after DRaf-1[DD]:ER addition of ICI and the relative levels of expression of activation we noted that cells were lost from the the di€erent DRaf:ER proteins was approximately surface of the dish presumably because of changes in equal. Addition of ICI to cells expressing the [YY], their adhesive properties or induced apoptosis (Figure [DD] and [FF] forms of DRaf-1:ER lead to an 2R and data not shown) Kau€man-Zeh et al., 1997. approximately ®vefold increase in expression after Similarly activation of all three forms of DB-Raf:ER in 24 h of treatment consistent with previous observa- Rat1 cells led to potent alterations in cell morphology tions (Figure 3a). Similar increases in expression were with the [DD] form displaying the most rapid and observed after the addition of ICI to cells expressing profound changes (Figure 2H, J, L). Activation of the the di€erent forms of DB-Raf:ER (Figure 3c). [YY] and [DD] forms of DA-Raf:ER in Rat1 cells led Prior to the addition of ICI, the basal level of to morphological transformation (Figure 2B and F, expression of all three DA-Raf:ER proteins appeared respectively) and again the [DD] form showed the most to be slightly higher compared to the level of rapid alterations in cell morphology. Interestingly, the expression of the cognate DRaf-1:ER proteins (Figure addition of ICI to Rat1 cells expressing DA-Raf[FF]:ER 3b and a, respectively). The abundance of DA- did not elicit any obvious manifestations of morpho- Raf[YY]:ER increased approximately threefold after logical transformation even after prolonged ICI 24 h of ICI treatment (Figure 3b as indicated). The treatment up to 120 h (Figure 2D). This result is [DD] and [FF] forms of DA-Raf:ER displayed similar similar to that seen when DRaf301:ER is expressed in kinetics of induction (Figure 3b). These results 3T3 and Rat1 cells (data not shown). demonstrated that the introduced point mutations did All of the transforming DRaf:ER proteins responded not have a severe e€ect on the expression of DRaf:ER to b-estradiol, 4-hydroxy-tamoxifen and ICI 182,780 as proteins. Is is also evident from this experiment that has been described previously for DRaf-1:ER (Samuels the apparent inability of DA-Raf[FF]:ER to transform et al., 1993) whereas DA-Raf[FF]:ER showed no cells is not due to a lack of expression of the fusion response to any of these hormones. We have protein. previously extensively documented that none of the estrogen receptor agonists or antagonists used in these Activation of the MAP kinase cascade in cells studies has any e€ect on the morphology of either transformed by DRaf:ER proteins parental 3T3 or Rat1 cells (Samuels et al., 1993). All of the transforming DRaf:ER proteins elicited potent Activation of DRaf-1:ER in 3T3 cells leads to the rapid, morphological transformation of Rat1 cells in the protein synthesis independent activation of MEK and absence of fetal calf serum indicating that the the p42/p44 MAP kinases (Samuels et al., 1993). We maintenance of serum growth factors plays little or have previously demonstrated that the catalytic no role in the ability of DRaf:ER proteins to give rise domains of the three di€erent Raf kinases display to alterations in cell morphology. Similar experiments di€erent biological and biochemical properties when were conducted in NIH3T3 cells with similar results expressed in cells (Pritchard et al., 1995). We wished (data not shown). therefore to assess whether the aforementioned differ- The potential of the di€erent DRaf:ER proteins to ences in amino acid sequence in the catalytic domain induce oncogenic transformation was con®rmed by could account for the di€erences in protein kinase testing the ability of Rat1 cells expressing the di€erent activity that we had previously observed. We therefore DRaf:ER proteins to form colonies in agarose. assayed the cell extracts described above for the Consistent with the morphological transformation activation of DRaf:ER proteins and for MEK and p42 data, all of the DRaf-1:ER and DB-Raf:ER proteins MAP kinase activities using the appropriate assays as elicited hormone-dependent colony formation. Rat1 described previously and in Materials and methods cells expressing either the [YY] or [DD] forms of DA- (Pritchard et al., 1995; Samuels et al., 1993). Since the Raf:ER also formed colonies in agarose in a hormone- experiments were conducted in parallel and equivalent dependent manner. However, Rat1 cells expressing exposures of the autoradiographs and Western blots are

DRaf[FF]:ER showed an extremely limited capacity to presented, it is reasonable to compare the kinase assays form colonies in agarose (data not shown). in ®gure 4, panels a-c, panels d-f and panels g-i to one another respectively. Activation of the [YY], [DD] and [FF] forms of DB- Expression of DRaf:ER proteins in Rat1 cells Raf:ER led to a robust activation of Raf kinase activity Treatment of 3T3 cells expressing DRaf:ER proteins towards GST-MEK1 (Figure 4c). The [YY] and [DD] with ICI leads to an increase in the expression of the forms of DB-Raf:ER displayed almost identical activity fusion proteins over a 24 h time course. Increased towards GST-MEK1 whereas the [FF] form displayed DRaf:ER expression is due, at least in part, to a 2 ± 3 slightly less activity under the conditions of this assay. fold increase in DRaf:ER mRNA expression as well as Equal loading was con®rmed by probing the kinase a selective stabilization of the DRaf:ER proteins within reactions with an antibody against hbER (data not the cell (Samuels et al., 1993). shown). Activation of the [DD] and [YY] forms of DB- Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1024

Figure 2 Oncogenic transformation of Rat1 cells by DRaf:ER proteins. Rat1 cells were infected with replication-defective retrovirus stocks derived from the di€erent pBP3DRaf:ER plasmids described above and infected cells were selected by culture in DMEM containing puromycin. Pools of 5105 puromycin resistant colonies expressing the di€erent DRaf:ER proteins were trypsinized and established in culture. These cells were grown until they were approximately 80% con¯uent at which time they were either treated with 0.1% (v/v) ethanol (7ICI Panel A, C, E, G, I, K, M, O and Q) or with 1 mM ICI 182,780 (+ICI, Panels B, D, F, H, J, L, N, P, R) to activate the DRaf:ER proteins. Cells were cultured for a futher 48 h at which point they were photographed using a Nikon TMS micrscope at 1206 magni®cation. DA-Raf[YY]:ER (A and B), DA-Raf[FF]:ER (C and D), DA-Raf[DD}:ER (E and F), DB-Raf[YY]:ER (G and H), DB-Raf[FF]:ER (I and J), DB-Raf[DD]:ER (K and L), DRaf-1[YY]:ER (M and N), DRaf-1[FF]:ER (O and P) and DRaf- 1[DD]:ER (Q and R) Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1025 Raf:ER gave rise to similar levels of MEK and p42 potential we could detect little or no di€erence in MAP kinase activity whereas the [FF] form gave rise to catalytic activity between the [YY] and [FF] forms of a slightly lower fold activation of both of these enzymes DA-Raf:ER. Although DA-Raf[DD]:ER activated MEK (Figure 4f and i, respectively). and p42 MAP kinase very eciently (Figure 4e and h, Activation of all three forms of DRaf-1:ER also led respectively), we could detect little or no MEK activity to activation of kinase activity towards GST-MEK1 in response to activation of either the [YY] or [FF] however we detected signi®cant di€erences in catalytic forms of DA-Raf:ER. We assume that the inability to activity between the di€erent forms of the protein discriminate between the transforming [YY] and non-

(Figure 4a). DRaf-1[DD]:ER displayed a much greater transforming [FF] forms of DA-Raf:ER is due to the kinase activity towards GST-MEK1 than the [FF] and level of sensitivity of the Raf and MEK assays (Figure

[YY] forms of DRaf:1:ER. DRaf-1[YY]:ER displayed a 4e). However, consistent with the transformation higher activity towards GST-MEK1 than DRaf-1[FF]:ER assays, activation of DA-Raf[YY]:ER gave rise to in this experiment but this di€erence was not detectable p42 MAP kinase activation whereas DA- reproducible. In general, the in vitro di€erences in Raf[FF]:ER did not (Figure 4h). DRaf-1:ER catalytic activity were re¯ected in the In order to conduct a side by side comparison of the relative abilities of these kinases to activate MEK level of p42 MAP kinase activity in cells transformed by and p42 MAP kinase in the cell with the [DD] form the di€erent DRaf:ER proteins, we prepared lysates of more potent than the [FF] or [YY] forms of DRaf- Rat1 cells in which the di€erent DRaf:ER fusion 1:ER (Figure 4d and g, respectively). However, the proteins were fully activated after 48 h of treatment magnitude of MEK and p42 MAPK activation by the with 1 mM ICI. The catalytic activity of p42 MAP kinase di€erent forms of DRaf-1:ER did not exactly mirror was measured as described in Materials and methods. the di€erences in Raf catalytic activity observed in vitro The highest level of p42 MAP kinase activity was as will be discussed later. detected in lysates from cells expressing DA-Raf[DD]:ER A similar pattern of protein kinase activity was and the [YY] and [DD] forms of DB-Raf:ER and DRaf- detected with the di€erent forms of DA-Raf:ER 1:ER (Figure 5a, lanes 2, 4, 5, 7 and 8 respectively).

(Figure 4b). DA-Raf[DD]:ER was signi®cantly more Intermediate levels of p42 MAP kinase activation were active than the [YY] and [FF] forms of the protein. detected in cells expressing the [FF] forms of DB- Surprisingly, despite the di€erences in transforming Raf:ER and DRaf-1:ER and low level activation was

observed in cells expressing DA-Raf[YY]:ER. The level of p42 MAP kinase activity in cells expressing DA-

Raf[FF]:ER was similar to that seen in normal a asynchronously growing Rat1 cells (data not shown) YY DD FF consistent with the inability of this protein to elicit oncogenic transformation in these cells. These data contrast with the in vitro catalytic activities of the ∆Raf-1:ER di€erent DRaf:ER proteins. The large apparent differ- ences in DRaf:ER catalytic activity were not re¯ected by 0 4 24 48 0 4 24 48 0 4 24 48 [hrs + ICI] equivalently large di€erences in p42 MAP kinase

activity in the intact cell. For example DRaf-1[DD]:ER was apparently ten times more active than Raf-1 :ER b [YY] YY DD FF yet the di€erence in MAP kinase activation was at most twofold. These observations suggest that the ability of oncogenic Raf kinases to activate the MAP kinase ∆A-Raf:ER pathway in mammalian cells may not simply be a function of Raf catalytic activity as we have observed 0 4 24 48 0 4 24 48 0 4 24 48 [hrs + ICI] previously in other cell types (Samuels et al., 1993). Con®rmation of the level of in vivo MAP kinase activity was sought by analysing the phosphorylation of the Ets- c 2 transcription factor on threonine residue 72 which we YY DD FF have established as a good candidate target for MAP kinase in intact 3T3 cells (McCarthy et al., 1997). The level of T72 Ets-2 phosphorylation in cells transformed ∆B-Raf:ER by the di€erent DRaf:ER proteins mirrored the extent of p42 MAP kinase activation that we observed in the 0 4 24 48 0 4 24 48 0 4 24 48 [hrs + ICI] experiment described above (Ostrowski, unpublished Figure 3 Expression of DRaf:ER proteins in Rat1 cells. observations). Puromycin resistant Rat1 cells infected with retroviruses encoding the di€erent DRaf:ER proteins were either untreated or treated with 1 mM for 4, 24 or 48 h as indicated at which times the cells were harvested and extracts were prepared as described Induction of HB-EGF gene expression by DRaf:ER in Materials and methods. Approximately 60 mg of total cellular activation protein from each cell lysate was electrophoresed and analysed for DRaf:ER expression by probing Western blots with an anti-hbER We have previously characterized the HB-EGF gene as antibody. (a) Extracts of Rat1 cells expressing DRaf-1:ER [YY], an immediate-early transcriptional target of activated [DD] and [FF] proteins as indicated. (b) Extracts of Rat1 cells Ras and Raf oncogenes (McCarthy et al., 1995). We expresing DA-Raf:ER [YY], [DD] and [FF] proteins as indicated. (c) Extracts of Rat1 cells expressing DB-Raf:ER [YY], [DD] and have further demonstrated that activation of HB-EGF [FF] proteins as indicated transcription in 3T3 cells is most likely initiated by Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1026 a bc ∆Raf-1:ER ∆A-Raf:ER ∆B-Raf:ER YY DD FF YY DD FF YY DD FF

GST-Mek1

0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 [hrs + ICI]

de f YY DD FF YY DD FF YY DD FF

rp42 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 [hrs + ICI]

gh i YY DD YY YY DD FF YY DD FF

MBP

0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 0 4 24 48 [hrs + ICI] Figure 4 Activation of the MAP kinase cascade in Rat1 cells by DRaf:ER proteins. Rat1 cells expressing the di€erent DRaf:ER proteins were either untreated or treated with 1 mM ICI for 4, 24 or 48 h as indicated at which times the cells were harvested, extracts were prepared and protein kinase assays were conducted as described in Materials and methods. The activity of the [YY], [DD] and [FF] forms of DRaf-1:ER, DA-Raf:ER and DB-Raf:ER (a, b and c respectively) was measured by an immunoprecipitation kinase assay using GST-MEK1 as a substrate. Equivalent loading of the DRaf:ER porteins in each immunoprecipitate was con®rmed by probing the blots presented in a, b and c with an antisera raised against hbER (data not shown). The activation of MEK in the Rat1 cells by the [YY], [DD] and [FF] forms of DRaf-1:ER, DA-Raf:ER and DB-Raf:ER (d, e and f respectively) was measured in a whole cell lysate kinase assay using recombinant, enzymatically inactive p42 MAP kinase (rp42) as a substrate. The activation of p42 MAP kinase in the Rat1 cells by the [YY], [DD] and [FF] forms of DRaf-1:ER, DA-Raf:ER and DB-Raf:ER (g, h and i respectively) was measured by an immunoprecipitation kinase assay using Myelin Basic Protein (MBP) as a substrate. Equivalent loading of p42 MAP kinase in each immunoprecipitate was con®rmed by probing the blots presented in g, h and i with a monoclonal antibody that speci®cally recognizes p42/p44 MAP kinase (data not shown)

p42/p44 MAP kinase-mediated phosphorylation of Ets- a 2 (McCarthy et al., 1997). We wished therefore to determine the e€ects of activation of the di€erent B-Raf:ER Raf-1:ER A-Raf:ER ∆ ∆ DRaf:ER proteins on the accumulation of HB-EGF ∆ mRNA. Rat1 cells expressing the di€erent DRaf:ER proteins were either untreated (0) or treated with ICI YY DD FF YY DD FF YY DD FF for 4, 24 or 48 h. Total cellular RNA was analysed for the presence of HB-EGF and GAPDH mRNAs by a MBP simultaneous RNase protection assay as described previously and in Materials and methods (McCarthy 123456789 et al., 1995). Induction of HB-EGF mRNA expression was quantitated by taking the ratio of the HB-EGF speci®c band to the GAPDH speci®c band. b In accord with the biochemical analysis of the MAP kinase cascade described above, activation of all three Raf-1:ER A-Raf:ER forms of DB-Raf:ER led to the induction of HB-EGF B-Raf:ER ∆ ∆ mRNA (Figure 6c). The induction of HB-EGF mRNA ∆ was strongest in response to activation of the [DD] and YY DD FF YY DD FF YY DD FF [YY] forms of DB-Raf:ER. Activation of all three forms of DRaf-1:ER also led to activation of HB-EGF p42 MAPK mRNA accumulation with the [DD] form demonstrat- ing the strongest level of induction, however both the 123456789 [YY] and [FF] forms of DRaf-1:ER induced HB-EGF Figure 5 Comparison of p42 MAP kinase activation in Rat1 mRNA eciently (Figure 6a). In accord with the cells expressing the di€erent DRaf:ER proteins. The activity of oncogenic potential of the di€erent DA-Raf:ER p42 MAP kinase in extracts of Rat1 cells expressing the [YY], [DD] and [FF] forms of DA-Raf:ER (lanes 1 ± 3), DB-Raf:ER proteins the activation of the [DD] and [YY] forms (lanes 4 ± 6) and DRaf-1:ER (lanes 7 ± 9) treated with ICI for 48 h of DA-Raf:ER gave rise to HB-EGF mRNA induction was measured by an immunoprecipitation kinase assay using with the [DD] form showing the strongest level of Myelin Basic Protein (MBP) as a substrate (a). After quantitation induction. Addition of ICI to cells expressing DA- of MBP phosphorylation using a PhosphorImager the Western blot in a was reprobed with a monoclonal antibody that Raf[FF]:ER had no e€ect on HB-EGF mRNA levels recognizes p42/p44 MAP kinase to allow estimation of the consistent with the lack of biological or biochemical relative amount of p42 MAP kinase in each immunoprecipitate e€ects of this protein in Rat1 cells (Figure 6b). (b) Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1027 Comparison of relative catalytic activity of the DRaf:ER Western blot with the 4G10 anti-phosphotyrosine monoclonal antibody revealed that none of the protein kinases expressed in Rat1 Cells DRaf:ER proteins was detectably tyrosine phosphory- In order to make a direct comparison of the relative lated. catalytic activity of the di€erent DRaf:ER proteins we prepared cell extracts from Rat1 cells expressing the Comparison of relative catalytic activity of DRaf:FH di€erent DRaf:ER proteins that had been treated with 6 protein kinases expressed in baculovirus infected Sf9 1 m ICI for 48 h. Equal amounts of the di€erent M cells DRaf:ER proteins were immunoprecipitated using the anti-hbER antisera and were incubated in a kinase In order to con®rm these results we expressed all nine reaction in which the concentrations of substrate were of the DRaf proteins described above as FLAG-His6 approximately ten times higher than their published Km epitope-tagged proteins using a baculovirus transfer values (13 mM GST-MEK1 and 100 mM ATP respec- vector (pVL1393:FH6) in Sf9 cells as described tively) (Force et al., 1994). In accord with the results in previously (Pritchard et al., 1995). In order to Figure 4 all of the forms of DB-Raf:ER demonstrated measure the kinase activities of these proteins we potent catalytic activity towards GST-MEK1 (Figure immunoprecipitated approximately equal amounts of

7a, lanes 4 ± 6) indicating that the engineered mutations DRaf:FH6 with an anti-FLAG antibody (M2) coupled had no signi®cant e€ect of catalytic activity. The [YY] to agarose and measured catalytic activity using GST- and [FF] forms of DRaf-1:ER displayed catalytic MEK1 as a substrate. In good agreement with the activity that was approximately 10% of the di€erent results obtained with the DRaf:ER fusion proteins DB-Raf:ER proteins (Figure 7a, lanes 7 and 9). By expressed in Rat1 cells, we found that all of the DB- contrast, DRaf-1[DD]:ER displayed catalytic activity that Raf:FH6 displayed the highest level of kinase activity was approximately ten times greater than DRaf- towards GST-MEK1 (Figure 7b, lanes 4 ± 6 as

1[YY]:ER and DRaf-1[FF]:ER and was equivalent to that indicated). Under the conditions of this experiment of the di€erent DB-Raf:ER proteins (Figure 7a, lane 8). we could detect no signi®cant di€erence between the

Under the conditions of this experiment we detected a activities of the di€erent forms of DB-Raf:FH6.In decrease in the electrophoretic mobility of the [DD] accord with the results of the DRaf:ER kinase assays form of DRaf-1:ER that may either be due to the we observed that DRaf-1[DD]:FH6 was approximately 10 e€ects of the point mutations on SDS binding or may times more active than DRaf-1[YY]:FH6 and DRaf- re¯ect phosphorylation of the protein prior to lysis of 1[FF]:FH6 (Figure 7b, lanes 8, 7 and 9, respectively). the cells. Indeed the level of activity of DRaf-1[DD]:FH6 was

As we have previously reported, DA-Raf[YY]:ER similar to the level of activity displayed by all of the phosphorylated GST-MEK1 poorly in vitro despite DB-Raf:FH6 proteins. Finally, we found that DA- the fact that it is capable of activating the p42/p44 Raf[DD]:FH6 phosphorylated GST-MEK1 as eciently

MAP kinases and transforming mammalian cells in as DRaf-1[YY]:FH6 (Figure 7b, lanes 2 and 7, culture (Figure 7a, lane 1). Similarly, DA-Raf[FF]:ER, respectively), but we remained unable to detect any which is at best only very weakly transforming, di€erence in the ability of the [YY] and [FF] forms of displayed little or no detectable kinase activity DA-Raf to phosphorylate GST-MEK1 (Figure 7b, towards GST-MEK1 (Figure 7a, lane 3). By contrast, lanes 1 and 3, respectively).

DA-Raf[DD]:ER was able to phosphorylate GST-MEK1 Overall, the relative kinase activities of the di€erent to the same eciency as the [YY] and [FF] forms of DRaf:FH6 proteins were similar to that observed with DRaf-1:ER (Figure 7a, lane 2). Equal loading of the the di€erent DRaf:ER proteins expressed in Rat1 cells.

DRaf:ER proteins was con®rmed by probing the The [YY], [DD] and [FF] forms of DB-Raf:FH6 and

Western blot presented in Figure 7a with the anti- the [DD] form of DRaf-1:FH6 displayed the highest hbER antiserum (Figure 7c). Reprobing of this level of kinase activity. The [YY] and [FF] forms of

abc ∆Raf-1:ER ∆A-Raf:ER ∆B-Raf:ER YY DD FF YY DD FF YY DD FF

HB-EGF

042448 042448 042448 042448 042448 042448 042448 042448 042448 [hrs + ICI]

def

GAPDH

042448 042448 042448 042448 042448 042448 042448 042448 042448 [hrs + ICI]

Figure 6 Induction of HB-EGF mRNA expression in Rat1 cells by DRaf:ER proteins. Rat1 cells expressing the [YY], [DD] and [FF] forms of DRaf-1:ER, DA-Raf:ER and DB-Raf:ER (a and d, b and e, c and f respectively) were either untreated or treated with 1 mM ICI for 4, 24 or 48 h as indicated at which times the cells were harvested and total RNA was isolated using the RNeasy protocol (Qiagen). RNase protection assays to simultaneously measure HB-EGF (a, b and c) and GAPDH as an internal reference (d, e and f) of the mRNA levels were carried out as described in Materials and methods Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1028 ab 6 6 6 Raf-1:FH B-Raf:FH A-Raf:FH Raf-1:ER B-Raf:ER A-Raf:ER ∆ ∆ ∆ ∆ ∆ ∆

YY DD FF YY DD FF YY DD FF YY DD FF YY DD FF YY DD FF

GST-Mek1 GST-Mek1

123456789 123456789

cd 6 6 6 Raf-1:FH B-Raf:FH A-Raf:FH ∆ ∆ ∆ Raf-1:ER B-Raf:ER A-Raf:ER ∆ ∆ ∆ YY DD FF YY DD FF YY DD FF YY DD FF YY DD FF YY DD FF ∆Raf:FH ∆Raf:ER 6

123456789 123456789

Figure 7 Comparison of the relative catalytic activities of the di€erent DRaf protein kinases. The relative activity of the [YY], [DD] and [FF] forms of DA-Raf:ER (lanes 1 ± 3), DB-Raf:ER (lanes 4 ± 6) and DRaf-1:ER (lanes 7 ± 9) in extracts of Rat1 cells treated with ICI for 48 h was measured by an immunoprecipitation kinase assay using GST-MEK1 as a substrate (a). After quantitation of GST-MEK1 phosphorylation using a PhosphorImager the Western blot in a was reprobed with a polyclonal anti-ER antibody that recognizes all of the DRaf:ER proteins to allow estimation of the relative amount of DRaf:ER proteins in each immunoprecipitate (c). DNA sequences encoding all of the DRaf protein kinases were subcloned into pVL1393:FH6 such that they were tagged at the carboxy-terminus with the FLAG epitope followed by six consecutive histidine residues. The catalytic activity of the [YY], [DD] and [FF] forms of DA-Raf:FH6 (Lanes 1 ± 3), DB-Raf:FH6 (lanes 4 ± 6) and DRaf-1:FH6 (lanes 7 ± 9) was measured by an immunoprecipitation kinase assay using GST-MEK1 as a substrate (b) After quantitation of GST-MEK1 phosphorylation using a PhosphorImager the Western blot in b was reprobed with the M2 monoclonal antibody that recognizes the FLAG epitope to allow estimation of the relative amount of DRaf:FH6 proteins in each immunoprecipitate (d).

DRaf-1:FH6 and the [DD] form of DA-Raf:FH6 contrasts with the observation that changing the displayed similar activity that was about tenfold less tyrosine residues found at the analogous positions of

than the DB-Raf:FH6 proteins. We could detect little or DA-Raf and DRaf-1 increased catalytic activity no activity with the [YY] and [FF] forms of DA- approximately tenfold. In order to determine if the

Raf:FH6. changes in amino acid sequence of DB-Raf had an Equal loading of the Western blot in panel b was e€ect of the kinetics of protein kinase activity we con®rmed by probing with the M2 anti-FLAG measured GST-MEK1 phosphorylation by the di€erent

monoclonal antibody. This also revealed that many forms of DB-Raf:FH6 as a function of time under

of the DRaf:FH6 proteins migrated as doublets conditions where the substrates were in excess. consisting of a major band of faster mobility and a Consistent with the data in Figure 7 we found that minor band of reduced mobility (Figure 7d). These the initial rate of GST-MEK1 phosphorylation by the

data suggested that a fraction of the DRaf:FH6 proteins di€erent forms of DB-Raf:FH6 was almost identical were phosphorylated when expressed in Sf9 cells. (data not shown). Indeed when this Western blot was reprobed with a monoclonal anti-phosphotyrosine antibody (4G10) we

observed that a sub-population of the DB-Raf[YY]:FH6 protein, but none of the others, was tyrosine Discussion phosphorylated (data not shown). In view of the lack of tyrosine phosphorylation of the di€erent DRaf:ER The mechanisms of regulation of the Raf family of proteins expressed in Rat1 cells and the elevated protein kinases have been the subject of considerable activity of the non-tyrosine phosphorylated DB- interest. Current models hold that the activation of

Raf[FF]:FH6 protein, it is unlikely that the tyrosine Raf-1 in response to growth factors such as EGF is phosphorylation is having a signi®cant in¯uence on the initiated by the Ras-mediated recruitment of Raf-1 to interpretation of the relative kinase activities of the the plasma membrane where it is activated by a di€erent DB-Raf kinases expressed in Sf9 cells. mechanism that most likely involves tyrosine phos- The results obtained in Rat1 and Sf9 cells expressing phorylation by members of the Src family of protein

the di€erent DB-Raf:ER and DB-Raf:FH6 proteins tyrosine kinases (Fabian et al., 1993; Marais et al., suggested that DB-Raf catalytic function is largely 1995). una€ected by mutation of the negatively charged Support for a regulatory role of tyrosine phosphor- aspartic acid residues found at positions 492 and 493 ylation has come from evidence that Raf-1 is tyrosine to either tyrosine or phenylalanine. This result phosphorylated in response to ligand stimulation Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1029 particularly in cells of hematopoietic origin (Carroll et Despite the fact that DA-Raf interacts with MEK1 al., 1991; Turner et al., 1991), in cells transformed by in the two-hybrid system in S. Cerevisiae we have been oncogenic Ras (Jelinek et al., 1996) and in cells unable to show that DA-Raf[YY] is able to phosphor- exposed to ionizing radiation (Kasid et al., 1996). ylate and activate MEK1 in vitro (Bosch and Furthermore, experiments in both mammalian and Sf9 McMahon, unpublished observations; Pritchard et al., cells have implicated two adjacent tyrosine residues in 1995). Others have however reported that immunopre- the catalytic domain of Raf-1 (Y340 and Y341) as cipitates of full length p68A-Raf from primary cardiac being important for the regulation of Raf-1 kinase myocytes stimulated with endothelin 1 (ET1) or TPA, activity. Indeed it has been shown that the conversion or from Hela cells stimulated with EGF were capable of either one of these two tyrosine residues to aspartic of phosphorylating MEK in vitro (Bogoyevitch et al., acid is sucient to activate the oncogenic potential of 1995; Wu et al., 1996). The results presented here may full length p74Raf-1 (Fabian et al., 1993). It is assumed therefore help to resolve this apparent discrepancy that the substitution of negatively charged aspartic acid since the [DD] form of DA-Raf eciently phosphory- residues into these sites mimics, at least in part, the lated MEK in vitro. Since the extent of tyrosine e€ects of tyrosine phosphorylation of the protein at phosphorylation of p68A-Raf was not assessed in the these sites. Similarities of sequence suggest that A-Raf aforementioned studies it is possible that the e€ects of may be regulated in a manner analogous to Raf-1 the [DD] mutation reveal that the catalytic activity of however, the catalytic domain of B-Raf encodes A-Raf has a strong requirement for negative charges in aspartic acid residues in the analogous positions these sites. It is also possible that phosphorylation of suggesting that the regulation of B-Raf may diverge A-Raf on other residues may also increase its kinase from that of the other family members. activity toward MEK1. We have previously demonstrated di€erences in We had anticipated that changing the analogous activity between the di€erent Raf kinases when aspartic acid residues in DB-Raf would reduce the expressed in mammalian or insect cells. Catalytic catalytic activity of the protein to that seen with DRaf- activity of the di€erent Rafs correlated with the 1[YY]. By contrast all of the forms of DB-Raf:ER/DB- ability of the proteins to both activate MEK and Raf:FH6 were highly active and not signi®cantly altered oncogenically transform NIH3T3 cells. Interestingly by mutation of these amino acids. These data indicate the mitogenic properties of the di€erent Raf kinases that the simple model that suggests that the increased appeared to show an inverse correlation with catalytic catalytic activity of B-Raf is a consequence of a minor activity as DA-Raf:ER elicited a stronger mitogenic di€erence in the sequence of the catalytic domain is response in quiescent 3T3 cells than DB-Raf:ER and largely incorrect. The fact that the catalytic activity of DRaf-1:ER (Pritchard et al., 1995). It appears that 3T3 DB-Raf was independent of negative charge in these cell proliferation is promoted by a low level of Raf residues, suggest that the higher intrinsic catalytic activity whereas high level Raf activity elicits cell cycle activity of B-Raf depends on other characteristics of arrest (Woods et al., manuscript submitted). the molecule. Experiments are under way to address Given the di€erences in sequence between the Raf this possibility. catalytic domains and the potential importance of these Although none of the DRaf:ER proteins expressed in di€erences in determining catalytic activity we have mammalian cells was detectably tyrosine phosphory- analysed the e€ects of expressing mutant forms DB- lated we found that a sub-population of the DB-

Raf, DA-Raf and DRaf-1 encoding either aspartic acid, Raf[YY]:FH6 protein expressed in Sf9 cells was tyrosine tyrosine or phenylalanine at the aforementioned phosphorylated. Since neither the [DD] nor the [FF] positions. We found that mutation of these tyrosine forms of DB-Raf:FH6 were tyrosine phosphorylated it residues to aspartic acid signi®cantly increased the seems likely that the site(s) of phosphorylation are activity of DRaf-1 and DA-Raf. These data support those that we have generated by mutation. By contrast previous work that demonstrated that negative charges the [YY] forms of DRaf-1 or DA-Raf were not at these positions in Raf-1 increased catalytic activity detectably tyrosine phosphorylated indicating a speci- and lends support for a role for tyrosine phosphoryla- ®city for the DB-Raf[YY]:FH6 protein. These results tion in the regulation of Raf-1 and A-Raf kinase suggest that DB-Raf[YY]:FH6 is being phosphorylated by activity. In contrast, DB-Raf catalytic activity was not a protein tyrosine kinase that is intrinsic to Sf9 cells. changed signi®cantly by mutating the analogous To date only one Raf homologue has been identi®ed in aspartic acid residues to either tyrosine or phenylala- the simple eukaryotes D. melanogaster and C. elegans. nine. Consequently, these aspartic acids residues do not These Raf homologues have been described to be most fully account for the higher catalytic activity exhibited like mammalian B-Raf and consistent with this by DB-Raf. hypothesis both genes encode aspartic acid residues in An interesting result to emerge from this study was the sites analogous to those under study in this report the apparent inability of the [FF] form of DA-Raf:ER (Han et al., 1993; Mark et al., 1987). The presence in to transform Rat1 and NIH3T3 cells whereas the [FF] Sf9 cells of a protein tyrosine kinase that speci®cally forms of DRaf-1:ER and DB-Raf:ER were ecient phosphorylates mutated mammalian DB-Raf[YY] may oncogenes. These data indicate that the presence of either be a fortuitous observation or an indication that hydroxyl groups or a negative charge at these positions the Spodoptera frugiperda Raf homologue may encode in DA-Raf is essential for the protein to be active in the tyrosine(s) in the aforementioned sites or may be an cells currently under analysis. Interestingly the DA- indicator that there are additional Raf homologues yet

Raf[FF]:ER protein may not be completely biologically to be discovered in insect cells. inactive as it is able to convert FDC-P1 cells to an IL-3 All of the transforming DRaf:ER proteins were able independent phenotype at a low frequency (Hoyle et to activate the expression of the previously identi®ed al., unpublished observations). immediate-early Raf target gene, HB-EGF (McCarthy Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1030 et al., 1995). In accord with the transformation data, Materials and methods the [DD] forms of DRaf:ER activated the expression of the gene to the greatest extent but the fold induction of Construction of Raf mutants HB-EGF did not necessarily match with the fold Site directed mutagenesis was used to mutate tyrosine di€erences in catalytic activity between the di€erent residues 299 and 300 in the catalytic domain of mouse A- DRaf:ER proteins. For example, despite the di€erences Raf to either aspartic acid or phenylalanine and aspartic in catalytic acitvity between the [YY] and [DD] forms acid residues 492 and 493 in mouse B-Raf to either tyrosine of DRaf-1:ER the kinetics and extent of HB-EGF or phenylalanine (Deng & Nickolo€, 1992). The primers activation were similar. These data suggest that there is used to construct these mutations were as follows: A-Raf a rate-limiting step in the activation of gene transcrip- (Y299D,Y300D): 5'-CCGGGACTCAGGCGATGACTGG- tion that occurs downstream of the activation of the GAGGTGC-3', A-Raf (Y299F,Y300F): 5'-CCGGGACTC- AGGCTTTTTCTGGGAGGTGC - 3; B - Raf (D49 - 2Y, DRaf:ER proteins. This idea is supported by the D493Y); 5'-GGTAGAAGAGATTAAGTTATTACTGGG- observation that DRaf-1[DD]:ER is approximately ten AGATTCCTGATG-3', B-Raf (Y492F,Y493F): 5'-GGTA- times more active than DRaf-1[YY]:ER but their relative GAAGAGATTCAAGTTTTTTCTGGGAGATTCCTGAT- abilities to activate p42 MAP kinases are only 2 ± 3- GG-3'. fold di€erent. Indeed we have previously demonstrated To improve the frequency of recovering mutants we that other factors such as the activity of cellular simultaneously replaced an XmnI site in pGEM7:DA-Raf, phosphoprotein phosphatases can play a role in and AlwNI site pGEM7:DB-Raf with the unique restriction determining the outcome of signaling through the sites EcoRV and SpeI, respectively using the following selection Raf pathway such that activation of DRaf-1:ER in primers: A-Raf trans-oligo XmnI/EcoRV, 5'-GCTCAT- Rat1a cells leads to MEK activation in the absence of CATTGGATATCGTTCTTCGGG-3'; B-Raf trans-oligo AlwN I/SpeI, 5'-GCAGCCACTAGTAACAGGATT-3';B- detectable MAP kinase activation (Samuels et al., Raf switch-oligo SpeI/AlwN I 5'-GCAGCCACTGGTAA- 1993). It is therefore possible that such molecules are CAGGATT-3' (Clontech). The selection trans-oligo XmnI/ having an in¯uence on MAP kinase activation under EcoRV together with the appropriate mutagenic A-Raf primers

the circumstances described here. It is also worthy of were used to create pGEM7:DA-Raf[DD] and pGEM7:DA-

note that we are analysing the activity of constitutively Raf[FF]. The selection trans-oligo AlwNI/SpeI together with the activated forms of the Raf kinases whereas the Raf appropriate B-Raf mutagenic primer was used to create

proteins are normally activated transiently in response pGEM7:DB-Raf[YY]. Having isolated pGEM7:DB-Raf[YY],it to the engagement of plasma membrane receptors. It is was used as a template for the selection switch-oligo SpeI/ possible that the relevant di€erences in Raf catalytic AlwnI and the appropriate B-Raf mutagenic primer to create pGEM7:DB-Raf . activity may be more signi®cant in determining the [FF] Mutants of human Raf-1 [Y340D,Y341D] and extent of activation of signaling pathways contolled by [Y340F,Y34F] were very kindly provided by Dr Deborah Raf kinases in response to physiological concentrations Morrison (NCI, Frederick, Maryland). P¯MI ± BgIII frag- of exogenous growth factors. ments of Raf-1 that included the mutated codons were At present it is not clear what the consequences of subcloned into pBS:DRaf-1, encoding the kinase domain of tyrosine phosphorylation (or aspartic acid substitution) human Raf-1 as reported previously, to generate pBS:DRaf-

have on the di€erent parameters of catalytic activity. In 1[DD] and pBS:DRaf-1[FF] (Samuels et al., 1993). XhoI ± EcoRI addition to increasing the catalytic activity of the isolated fragments encoding the di€erent forms of the kinase domains kinase domain it is possible that tyrosine phosphoryla- of Raf-1, A-Raf and B-Raf we subcloned into pBP3:hbER, a tion of Raf-1 may play a role in relieving the inhibitory replication-defective murine retrovirus vector derived from e€ects of the CR2 region of the protein (not contained pBabePuro (Morgenstern and Land, 1990) that encodes the hormone binding domain of the HE14 allele of the human within the portion of the three Rafs under investigation estrogen receptor (hbER) (Figure 1). This created in-frame here). Given the importance of tyrosine phosphorylation fusions between sequences encoding the [YY], [DD] and [FF] in mediating protein-protein interactions it is also forms of DRaf-1, DA-Raf and DB-Raf and hbER as has been possible that such modi®cations could target Raf for described previously (Pritchard et al., 1995; Samuels et al., interactions with other cellular proteins such as has been 1993). We thus generated retrovirus vectors (pBP3DRaf:ER)

reported with Lck and Fyn (Clark et al., 1994; Iwashima encoding DRaf-1[YY]:ER, DRaf-1[DD]:ER, DRaf-1[FF]:ER, DA- et al., 1994). In the context of the isolated Raf kinase Raf[YY]:ER, DA-Raf[DD]:ER, DA-Raf[FF]:ER, DB-Raf[YY]:ER, DB-Raf[DD]:ER and DB-Raf[FF]:ER as depicted in Figure 1. domains, tyrosine phosphorylation could a€ect the Km

and/or the Vmax of the enzyme. We are presently using

standard enzyme kinetic measurements and the BIAcore Expression of epitope-tagged DRaf:FH6 kinases in Sf9 cells to measure these e€ects. The di€erent DRaf mutants were subcloned into a In conclusion, we propose that these data support baculovirus transfer vector as described previously the hypothesis that suggests that the di€erent Raf (Pritchard et al., 1995). NcoI ± EcoRI fragments encoding kinases may be di€erentially coupled to upstream the kinase domains of the di€erent forms of Raf-1, A-Raf

signaling pathways. Consistent with a variety of in and B-Raf were cloned into pVL1393:FH6 to generate

vitro studies, B-Raf activation may only require pVLDRaf-1[YY]:FH6,pVLDRaf-1[DD]:FH6,pVLDRaf-1[FF]: binding to Ras.GTP (Moodie et al., 1994; Tamamori FH6,pVLDA-Raf[YY]:FH6,pVLDA-Raf[DD]:FH6,pVLDA- et al., 1995). Although there is no evidence that Ras- Raf[FF]:FH6,pVLDB-Raf[YY]:FH6,pVLDB-Raf[DD]:FH6 and pVLDA-Raf :FH respectively. This cloning vector is a mediated activation of Raf-1 has an absolute require- [FF] 6 ment for tyrosine phosphorylation the results reported modi®ed form of pVL1393 into which sequences encoding the FLAG epitope (DYKDDDK, Kodak/IBI) followed by here are consistent with a role for tyrosine phosphor- six consecutive histidine residues had been inserted. The ylation being able to strongly potentiate Raf-1 catalytic insertion of the DRaf kinase domains into this vector leads activity. Finally activation of A-Raf may have strong to the expression of epitope-tagged forms of the proteins in dependency on both Ras activation and tyrosine Sf9 cells (Figure 1). Baculovirus stocks encoding the [YY], phosphorylation. [DD] and [FF] forms of the DRaf kinase domains were Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1031 producedandusedtoinfectSf9cellsbystandard incubated with the appropriate secondary antibody coupled techniques. to horseradish peroxidase (hrp). The enhanced chemilumi- nescence detection system (Amersham) was utilized to visualize the antigen-antibody complexes. Cell culture, retrovirus production, and infection All cells were cultured in Phenol Red free Dulbecco's modi®ed Eagle's medium supplemented with 10% fetal Assay for Raf kinase activity bovine serum (FCS), penicillin and streptomycin at 378Cin Immune-complexes of DRaf:ER or DRaf:FH6 were pre- a humidi®ed atmosphere containing 6% (v/v) CO2. Cells pared by incubating cell lysates with either the anti-hbER were photographed with a Nikon TMS photomiscroscope. and protein A-Sepharose 4B (Pharmacia) or M2-agarose ICI 182,780 (hereafter referred to as ICI), an estrogen anti-FLAG antisera (Kodak/IBI). Immune-complexes were receptor antagonist, was provided by Dr Alan Wakeling collected, washed in lysis bu€er and the activities of

(Zeneca Pharmaceuticals, Maccles®eld, UK) and was DRaf:ER and DRaf:FH6 were assessed by incubation for prepared as a 1 mM stock in ethanol and stored at 30 min at 308C in a reaction mix containing 25 mM HEPES

7208C. Virus stocks encoding the DRaf:ER alleles were (pH 7.4), 10 mM MgC12,1mM MnC12,1mM DTT, 50 obtained by Lipofectamine (Gibco-BRL) mediated trans- mM ±100mM ATP and 10 mCi of [g-32P]ATP (Amersham) fection of the pBP3DRaf:ER vectors into Bosc23 cells (Pear with 3 ± 15 mg(2.6±13mM) of puri®ed recombinant et al., 1993). Brie¯y 12.5 mg of supercoiled plasmid DNA enzymatically inactive GST-MEK1[K97R] as a substrate. was mixed with 90 ml of Lipofectamine in a total volume of The reactions were analyzed by polyacrylamide gel 2 ml of OPTI-MEM-I (Gibco-BRL) medium. The DNA/ electrophoresis and Western blotting. Western blots were Lipofectamine mix was incubated for 6 h with 56106 ®rst quantitated using a Molecular Dynamics Phosphor- Bosc23 cells that had been plated onto a 10 cm dish 12 h Imager to measure GST-MEK1[K97R] phosphorylation previously. After this incubation 10 ml of OPTI-MEM-I and subsequently probed with either the anti-hbER or the media containing 20% FCS was added and the incubation anti-FLAG antisera as appropriate to quantitate the continued for a further 19 h. Virus stocks were harvested amounts of the DRaf kinases in each immunoprecipitate. over the following 24 h into 5 ml of Phenol Red free DMEM containing 10% (vv) FCS and then used to infect Rat1 and 3T3 cells. Forty-eight hours following infection, Assay for MEK activity the cells were trypsinized and placed in medium containing Aliquots from unfractionated cell lysates containing 60 mg 2 mg/ml puromycin (Sigma) to select for virus-infected of total protein were incubated for 30 min at 308Cina cells. Standard virus stocks gave rise to 5106 puromycin reaction mix containing 40 mM HEPES pH 7.8, 10 mM resistant colonies per ml of virus. Following selection in MgCl 1mM DTT and 40 mlATPand2mg of bacterially puromycin, cells were pooled, expanded and tested for the 2 expressed rp42, which is an enzymatically inactive form of expression of DRaf:ER proteins by Western blotting. p42 MAP kinase. The reaction mixtures were electrophor- Under these conditions 595% of the puromycin resistant esed through a 10% polyacrylamide gel, which was then Rat1 or 3T3 colonies express the DRaf:ER fusion proteins Western blotted and exposed to X-ray ®lm. Tyrosine (Cherwinski and McMahon, unpublished observations). phosphorylation of rp42 was quantitated by probing the Western blot with 4G10-hrp, a horseradish peroxidase- Assessment of DRaf:ER-induced oncogenic transformation conjugated anti-phosphotyrosine monoclonal antibody. To test for DRaf:ER induced morphological transforma- tion, the pooled populations described above were plated Assay for p42 kinase activity and grown in DMEM containing 10% FCS until they were approximately 80% con¯uent. At this time the cells were Sixty mg of cell extract was incubated for 2 h at 48C with 5 treated with either 0.1% (v/v) ethanol as a solvent control ml of an anti-p42 MAP kinase polyclonal antiserum conjugated to agarose beads. Immune-complexes were or with either 1 mM b-estradiol, 4-Hydroxy-Tamoxifen or ICI 182,780, and the cells were monitored over the collected and washed twice with lysis bu€er and once following 72 h. All of these compounds have been shown with a bu€er containing 25 mM Tris (pH 7.5), 137 mM MgC1 and 10% (v/v) glycerol. Immune-complexes were to elicit the biochemical and biological activation of DRaf- 2 1:ER (Samuels et al., 1993). The ability of the di€erent incubated for 30 min at 308C in a reaction mix containing 40 mM HEPES pH 7.4 40 mM MgC1 ,200mM ATP, and DRaf:ER proteins to elicit morphological transformation 2 32 was also assessed in Rat1 cells that were grown to 80% 5 mCi of [g- P]ATP with 20 mg of Myelin Basic Protein con¯uency in DMEM containing 10% FCS and then (MBP) as a substrate. The reactions were analysed and switched to DMEM containing 1% (w/v) BSA with no quantitated on SDS polyacrylamide gels as described FCS prior to the addition of hormones with identical above. Western blots were ®rst quantitated using a results. Molecular Dynamics PhohphorImager to measure MBP phosphorylation and subsequently probed with anti-p42 MAP kinase to quantitate the amounts of the MAP kinases Preparation of cells extracts and analysis by Western blot in each immunoprecipitate. Triton X-100-soluble cell lysates were prepared as described previously (Samuels et al., 1993). Cells were lysed in Gold Ribonuclease protection assays lysis bu€er (GLB), and protein concentrations were measured using the BCA protein assay kit (Pierce). Rat1 cells expressing the di€erent DRaf:ER proteins grown Aliquots of cell lysates were electrophoresed through to 80% con¯uency on 15 cm cell culture plates were either polyacrylamide gels and Western blotted onto Immobilon untreated or treated with 1 mMICIfor4,24or48hat P polyvinylidene di¯uoride membranes (Millipore). Western which time total cellular RNA was isolated using the blots were probed with 1:2500 dilutions of either an anti- RNeasy kit (Qiagen) according to the manufacturer's hbER polyclonal antiserum (HC-20) (Santa Cruz Biotech- instructions and as described previously (McCarthy et al., nology), an anti-FLAG monoclonal antibody (M2) (Kodak/ 1995). Simultaneous quantitation of the level of expression IBI), or an anti-phosphotyrosine monoclonal antibody of HB-EGF and GAPDH mRNAs was conducted using a (4G10, UBI) for at least 1 h. The ®lters were washed in Ribonuclease protection assay as described previsouly Tris-bu€ered saline containing 0.5% (v/v) Nonidet P-40 and (McCarthy and Bicknell, 1992; McCarthy et al., 1995). Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1032 Acknowledgements oligonucleotide synthesis, Toshio Kitamura for advice on We thank Debbie Morrison, Richard Marais and Chris retrovirus production and Gary Burgett for graphics Marshall for the provision of reagents and the commu- support. Finally we would like to thank the members of nication of results prior to publication and Jay Morgen- the McMahon and Lees' laboratories for helpful discus- stern for the provision of the pBabepuro retrovirus vector. sions, advice and for reviewing the manuscript. The DNAX In addition we are grateful to our colleagues at DNAX for Research Institute is supported by Schering Plough their advice and assistance; Mary Weiss and Satish Menon Corporation. for puri®cation of GST-MEK1, Debbie Liggett for

References

Ahn NG, Robbins DJ, Haycock JW, Seger R, Cobb MH and Koide H, Satoh T, Nakafuku M and Kaziro Y. (1993). Proc. Krebs EG. (1992). J. Neurochem., 59, 147 ± 156. Natl, Acad. Sci. USA, 90, 8683 ± 8686. Ashworth A, Nakielny S, Cohen P and Marshall C. (1992). KolchW,HeideckerG,KochsG,HummelR,VahidiH, Oncogene, 7, 2555 ± 2556. Mischak H, Finkenzellar G, Marme D and Rapp UR. Avruch J, Zhang XF and Kyriakis JM. (1994). Trends (1993). Nature, 364, 249 ± 252. Biochem. Sci., 19, 279 ± 283. Kyriakis JM, App H, Ziang X-f., Banerjee P, Brautigan DL, Beck TW, Huleihel M, Gunnell M, Bonner TI and Rapp UR. Rapp UR and Avruch J. (1992) Nature, 358, 417 ± 421. (1987). Nucleic Acids Res., 15, 595 ± 609. Lin LL, Wartmann M, Lin AY, Knopf JL, Seth A and Davis Bogoyevitch MA, Marshall CJ and Sugden PH. (1995). J. RJ. (1993). Cell, 72, 269 ± 278. Biol. Chem., 270, 26303 ± 26310. Liu D, Bienkowska J, Petosa C, Collier RJ, Fu H, and Braselmann S and McCormick F. (1995). EMBO J., 14, Liddington R. (1995). Nature, 376, 191 ± 194. 4839 ± 4848. Luo Z, Tzivion G, Belshaw PJ, Vavvas D, Marshall M and Carroll MP and May WS. (1994). J. Biol. Chem., 269, 1249 ± Avruch J. (1996). Nature, 383, 181 ± 185. 1256. MacDonald SG, Crews CM, Wu L, Driller J, Clark R, Carroll MP, Spivak JL, McMahon M, Weich N Rapp UR Erikson RL and McCormick F. (1993). Mol. Cell. Biol., and May WS. (1991). J. Biol. Chem., 266, 14964 ± 14969. 13, 6615 ± 6620. Cherniack AD, Klarlund JK and Czech MP. (1994). J. Biol. Marais R, Wynne J and Treisman R. (1993). Cell, 73, 381 ± Chem., 269, 4717 ± 4720. 394. Clark MR, Johnson SA and Cambier JC. (1994). EMBO J., Marais R, Light Y, Paterson H and Marshall CJ. (1995). 13, 1911 ± 1919. EMBO J., 14, 101 ± 110. Crews CM, Asessandrini A and Erikson RL. (1992). Science, Mark GE, Mac IR, Digan ME, Ambrosio L and Perrimon N. 258, 478 ± 480. (1987). Mol. Cell. Biol., 7, 2134 ± 2140. Deng WP and Nickolo€ JA. (1992). Anal. Biochem., 200, Marx M, Eychene A, Laugier D, Bechade C, Crisanti P, 81 ± 88. Dezelee P, Pessac B and Calothy G. (1988). EMBO J., 7, Dent P, Haser W, Haystead TAJ, Vincent LA, Roberts TM 3369 ± 3373. and Sturgill TW. (1992). Science, 257, 1404 ± 1407. McCarthy SA and Bicknell RB. (1992). J. Biol. Chem., 267, Dent P, Reardon DB, Morrison DK and Sturgill TW. (1995). 21617 ± 21622. Mol. Cell. Biol., 15, 4125 ± 4135. McCarthy SA, Chen D, Yang B, Cherwinski H, Hauser CA, Dent P, Wu J, Romero G, Vincent LA, Castle D and Sturgill Chen X, Klagsbrun ML, Ostrowski MC and McMahon TW. (1993). Mol. Biol. Cell., 4, 483 ± 493. M. (1997). Mol. Cell. Biol. 17, 2401 ± 2412. Fabian JR, Daar IO and Morrison DK. (1993). Mol. Cell. McCarthy SA, Samuels ML, Pritchard CA, Abraham JA and Biol., 13, 7170 ± 7179. McMahon M. (1995). Genes and Development, 9, 1953 ± Farrar MA, Alberola-Ila J and Perlmutter RM. (1996). 1964. Nature, 383, 178 ± 181. Moodie SA, Paris MJ, Kolch W and Wolfman A. (1994). Finney RE, Robbins SM and Bishop JM. (1993). Curr. Biol., Mol. Cell. Biol., 14, 7153 ± 7162. 3, 805 ± 812. Moodie SA, Willumsen BM, Weber MJ and Wolfman A. Force T, Bonventre JV, Heidecker G, Rapp U, Avruch J and (1993). Science, 260 1658 ± 1661. Kyriakis M. (1994). Proc. Natl. Acad. Sci. USA, 91, 1270 ± Morgenstern JP and Land H. (1990). Nucleic Acids Res., 18, 1274. 3587 ± 3596. Freed E, Symons M, Macdonald SG, McCormick F and Morrison DK. (1990). Cancer Cell, 2, 377 ± 382. Ruggieri R. (1994). Science, 256, 1713 ± 1716. Nemeno€ RA, Winitz S, Qian NX, Van Putten V, Johnson GhoshS,StrumJC,SciorraVA,DanielLandBellRM. GL and Heasley LE. (1993). J. Biol. Chem., 268, 1960 ± (1996). J. Biol. Chem., 271, 8472 ± 8480. 1964. Han M, Golden A, Han Y and Sternberg PW. (1993). Nature, Papin CA, Denouel A, Calothy G and Eychene A. (1996). 363, 133 ± 139. Oncogene, 12, 2213 ± 2221. Hill CS, Marais R, John S, Wynne J, Dalton S and Treisman Payne DM, Rossomando AJ, Martino P, Erickson AK, Her R. (1993). Cell, 73, 395 ± 406. JH, Shabanowitz J, Hunt DF, Weber MJ and Sturgill TW. Howe LR and Marshall CJ. (1993). J. Biol. Chem., 268, (1991). EMBO J., 10 885 ± 892. 20717 ± 20720. Pear WS, Nolan GP, Scott ML and Baltimore D. (1993). Ikawa S, Fukui M, Ueyama Y, Tamaoki N, Yamamoto T Proc.Natl.Acad.Sci.USA,90, 8392 ± 8396. and Toyoshima K. (1988). Mol. Cell. Biol., 8, 2651 ± 2654. Pritchard CA, Samuels ML, Bosch E and McMahon M. Iwashima M, Irving BA, van Oers NS, Chan AC and Weiss (1995). Mol. Cell. Biol., 15, 6430 ± 6442. A. (1994). Science, 263, 1136 ± 1139. Rapp UR. (1991). Oncogene, 6, 495 ± 500. Jelinek T, Dent P, Sturgill TW and Weber MJ. (1996). Mol. Samuels ML and McMahon M. (1994). Mol. Cell. Biol., 14, Cell. Biol., 16, 1027 ± 1034. 7855 ± 7866. Kasid U, Guy S, Dent P, Ray S, Whiteside TL and Sturgill Samuels ML, Weber MJ, Bishop JM and McMahon M. TW. (1996). Nature, 382, 813 ± 816. (1993). Mol. Cell. Biol., 13, 6241 ± 6252. Kau€man-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert, Sozeri O, Vollmer K, Liyanage M, Frith D, Kour G, Mark C, Co€er P, Downward J and Evan, G. (1997). Nature, G. and Stabel S. (1992). Oncogene, 7, 2259 ± 2262. 385, 544 ± 548. Key residues modulate A-Raf and Raf-1 activity E Bosch et al 1033 Stanton VJ, Nichols DW, Laudano AP and Cooper GM. Wu X, Noh SJ, Zhou G, Dixon JE and Guan KL. (1996). J. (1989). Mol. Cell. Biol., 9, 639 ± 647. Biol. Chem. 271, 3265 ± 3271. Storn SM, Brennscheidt U, Sithanandam G and Rapp UR. Xiao B, Smerdon SJ, Jones DH, Dodson GG, Soneji Y, (1990). Crit. Rev. Oncog., 2, 1±8. Aitken A and Gamblin SJ. (1995). Nature, 376, 188 ± 191. Turner B, Rapp U, App H, Greene M, Dobashi K and Reed Yamamori B, Kuroda S, Shimizu K, Fukui K, Ohtsuka T J. (1991). Proc.Natl.Acad.Sci.USA,88, 1227 ± 1231. and Takai Y. (1995). J. Biol. Chem., 270, 11723 ± 11726. Van Aelst L, Barr M, Marcus S, Polverino A and Wigler M. Yao B, Zhang Y, Delikat S, Mathias S, Basu S and Kolesnick (1993). Proc. Natl. Acad. Sci. USA, 90, 6213 ± 6217. R. (1995). Nature, 378, 307 ± 310. Vojtek AB, Hollenberg SM and Cooper JA. (1993). Cell, 74, Zhang XF, Settleman J, Kyriakis JM, Takeuchi SE, Elledge 205 ± 214. SJ, Marshall MS, Bruder JT, Rapp UR and Avruch J. Warne PH, Viciana PR and Downward J. (1993). Nature, (1993). Nature, 364, 308 ± 313. 364, 352 ± 355. Wu J, Harrison JK, Dent P, Lynch KR, Weber MJ and Sturgill TW. (1993). Mol. Cell. Biol., 13, 4539 ± 4548.