Display of Ras on filamentous phage through cysteine replacement

Biochimie 81 (1999) 1079−1087 © 1999 Société française de biochimie et biologie moléculaire/Éditions scientifiques et médicales Elsevier SAS. All rights reserved.

Display of Ras on filamentous phage through cysteine replacement Troels Wind*, Svend Kjær**, Brian F.C. Clark University of Aarhus, Department of Molecular and Structural Biology, Gustav Wieds Vej 10C, 8000 Århus C, Denmark (Received 6 September 1999; accepted 18 October 1999) Abstract — Phage display technology has been used in a variety of contexts to understand and manipulate biomolecular interactions between proteins and other biomolecules. In this paper we describe the establishment of a phage display system for elucidation of the interactions between the GTPase Ras and its panel of effectors. It is shown how technical problems associated with phage display of a protein with unpaired cysteines, likely to be caused by the oxidizing environment of the bacterial periplasm into which the protein is directed, can be overcome by cysteine replacement based on functional and structural studies. First, the catalytic domain (residues 1–166) of mammalian H-Ras (Ras) was observed to be displayed on phage in an incorrect conformation not detectable by antibodies recognizing conformational epitopes on Ras. Although truncation of the phage coat protein used as fusion partner (g3p) resulted in minor improvements in the display, Ras was tailored for phage display by cysteine replacement. By replacing the three cysteines at positions 51, 80 and 118 of Ras with the corresponding residues in Saccharomyces cerevisiae RAS1, the resulting fusion-phage is recognized by the conformation-dependent anti-Ras antibodies. Furthermore, display of cysteine-free Ras is demonstrated by GTP-analogue dependent binding to the Ras-binding domain of the Ras-effector Raf1. These data pave the way for analysis of Ras-effector interactions using phage display technology yet demonstrate that phage display of proteins with normally reduced cysteines should be approached with caution. © 1999 Société française de biochimie et biologie moléculaire/Éditions scientifiques et médicales Elsevier SAS phage display / Ras / Raf1 / disulphide bridge / mutagenesis

1. Introduction In eukaryotic signaling the small GTPase Ras plays a pivotal role for important cellular processes in organisms ranging from yeast to man [1]. Ras is located at the inner face of the plasma membrane where it responds to extracellular stimuli and transmits these into the cytoplasm via direct interactions with a panel of effector molecules. Ras-mediated signaling is controlled by proteins responsible for switching Ras ‘on’ and ‘off’ by provoking GTP recruitment or hydrolysis, respectively [2]. This switch function is accomplished by the * Correspondence and reprints ** Present address: Karolinska Institute, Division of Molecular Neurobiology, Department of Neuroscience, Berzeliusväg 1, plan 3, 171 77 Stockholm, Sweden. Abbreviations: A, absorbance; aa, amino acid(s); Ab, antibody; cfu, colony forming units; ELISA, enzyme-linked immunosorbent assay; g3p, gene 3 product of filamentous Ff bacteriophage; GDP, guanosine 5’-diphosphate; GTP, guanosine 5’triphosphate; GppNHp, guanosine-5’-(β, γ-imido) triphosphate; GST, glutathione S-transferase; HRP, horse radish peroxidase; kbp, kilobasepair(s); kDa, kilodalton(s); MTBSG, TBS with 1% skimmed milk powder and 5% glycerol; OPD, o-phenylene diamine; OD, optical density; PCR, polymerase chain reaction; RBD, Ras-binding domain; TBS, Tris-buffered saline (50 mM Tris, pH 7.5, 100 mM NaCl).

ability of Ras to exist in two different conformations depending on the nucleotide it binds [3]. The transforming phenotype of oncogenic Ras such as Ras(G12V), which has an impaired GTPase activity and therefore predominantly exists in the GTP-bound conformation, is caused by uncontrolled signaling via multiple effectors and signaling pathways [4]. Several Ras mutants have been isolated that show preferential binding to one or few of the known effectors. Two-hybrid screenings [5] have identified Ras(T35S) and Ras(E37G) as being capable of discriminating between the kinases Raf1 and Byr2 [6], the latter of which is an effector for Ras in Schizosaccharomyces pombe [7]. Later, D38E and Y40C were included as partial loss-of-function mutations in Ras that discriminate between the known effectors Raf1, RalGDS and PI(3)kinase [8-10]. Such Ras mutants are valuable tools for delineation of Ras mediated cell transformation occurring via the plethora of Ras effectors [4]. Display on the surface of filamentous bacteriophage (phage) allows selection of (poly)peptides with desired interaction characteristics from repertoires with millions of members [11]. As the assembly of phage particles takes place in the bacterial periplasm [12], the displayed protein must be able to fold correctly under oxidative conditions where disulphide bridges are catalytically formed [13]. The interactions between hormones and their cognate

1080 receptors have been successfully analyzed and manipulated by display of hormone mutant repertoires on phage followed by selection versus the receptor [14-17]. Likewise, the specificity of zinc-fingers [18-20] and protease inhibitors [21, 22] has been investigated through phage display of repertoires followed by selection. Since selection from phage displayed repertoires takes place in vitro it offers the possibility of controlling the milieu. Hence, the selection can be guided by competition during the binding reaction with the bait [21, 23-26]. With this in mind, phage display may offer new routes to isolation of Ras loss-of-function mutants from a Ras-repertoire with random mutations in and around the effector region. Hence, it was decided to harness Ras for phage display. In this report we demonstrate display of the catalytic domain of Ras on the surface of filamentous phage through fusion to the minor phage coat protein g3p. Moreover, it is shown that changing the three cysteines at positions 51, 80 and 118 in Ras to the corresponding amino acids from Saccharomyces cerevisiae (S. cerevisiae) RAS1 (serine, leucine and leucine, respectively) is required for Ras to be displayed in its native conformation as judged from interactions with monoclonal antibodies recognizing conformational epitopes on Ras. Finally, it is shown that the displayed cysteine-free Ras interacts with the Ras-binding domain from Raf1 (RafRBD) in a GTPdependent manner. Thus we have prepared Ras for phage display and thereby enabled selection of Ras mutants with desired effector preferences. 2. Materials and methods 2.1. General procedures Unless indicated otherwise, all enzymes were from New England Biolabs, DNA was purified according to Hansen et al. [27], Escherichia coli (E. coli) strain TG1 (supE hsd∆5 thi ∆(lac-proAB) F’[traD36 proAB+ lacIq lacZ∆M15]) was used as host and sequencing was performed using an Applied Biosystems 373A sequencer and the ABI Prism Dye terminator cycle sequencing kit (Perkin-Elmer). Mutations were introduced with the Quickchange kit from Stratagene according to the instructions from the manufacturer, except that the final reaction product was electroporated into E. coli strain TG1. Bacteria were grown in 2 × TY broth (10 g yeast extract, 16 g casein and 5 g NaCl per liter) and, when appropriate, antibiotics were used at the following concentrations: 100 µg/mL ampicillin, 12.5 µg/mL tetracyclin or 25 µg/mL kanamycin for E. coli harbouring pHEN1 derivatives, fUSE5 derivatives or helper-phage, respectively. All oligonucleotides used for PCR, mutagenesis and sequencing are summarized in table I. GST [28] and GST-RafRBD [29] proteins were purified on glutathione agarose (Sigma) according to the instructions from the manufacturers.

Wind et al. The structures of Ras-GDP (PDB-ID 4Q21) and RasGTP (PDB-ID 5P21) were analyzed with the RasMol 2.6 software (R. Sayle, Glaxo Wellcome R&D, UK). 2.2. Cloning and mutagenesis A His6 tag was inserted into the phagemid pHEN1 [30] by inserting a oligonucleotide cassette (HHUP and HHLO) into NotI-digested pHEN1, thereby creating pHH1. The cDNA encoding residues 1 to 166 of H-Ras was amplified by PCR from pBW1699 using primers RASSFI and RASNOT2 introducing SfiI and NotI restriction sites for subcloning. Taq polymerase (LifeTechnologies) was used according to the instructions from the manufacturer and the product was extracted with phenol and precipitated with ethanol. The PCR product and pHH1 were digested with SfiI and NotI in separate reactions, purified, ligated and electroporated into TG1. The insert and flanking cloning sites in positive clones were sequenced using primers REV and M13back. A missing codon, G151 in Ras, was inserted with the Quickchange kit (Stratagene) using primers G151UP and G151LO. The integrity of the resulting phagemid, pHHras, was confirmed by sequencing with the primers REV and M13back. Ras (aa 1–166) was subcloned from pHHras to pN (a pHEN1 derivative, NJV Hansen, unpublished) via NcoI and EagI for fusion with a truncated version of g3p (aa 218–406) thereby creating pNras that was verified by sequencing with the primer REV. Phagemids based on pHHras with all combinations of Cys or Ser at position 51, Cys or Leu at position 80 and Cys or Leu at position 118 were constructed using the Quickchange kit (Stratagene) and the primers C51SUP, C51SLO, C80LUP, C80LLO, C118LUP and C118LLO. Positive clones were identified by restriction enzyme digestion since each mutation was accompanied by either removal or insertion of a restriction site (table I) and confirmed by sequencing as above. The phage fUSE5 [31] was modified by digestion with EagI and SfiI and insertion of a oligonucleotide cassette (FUHUP and FUHLO) that created SfiI, NcoI and NotI sites compatible with pHH1. The resulting multiple cloning site was confirmed by sequencing with the primers M13for and M13back. An inherent NcoI site in fUSE5 was removed with the Quickchange kit (Stratagene) and the primers FUHNCOUP and FUHNCOLO to complete the phage vector fUH1. Cysteine-free Ras was subcloned from pHHras(C51S, C80L, C118L) into fUH1 via NcoI and NotI to create fUHras(C51S, C80L, C118L), that was verified by sequencing with the primers M13for and M13back. 2.3. Production of phage particles All phage particles were produced from 100 mL cultures of E. coli strain TG1 harboring the relevant

Phage display of Ras

1081

Table I. Oligonucleotides used for PCR, sequencing and mutagenesis. Sequence

Commentsa

5′GGCCGCACATCATCATCATCATGG 3′ 5′GGCCCCATGATGATGATGATGATGTGC 3′

Introduces a His6 tag into pHEN1 creating pHH1

5′ATAGCGGCCCAGCCGGCCATGGCCATGACGGAATATAAGCTGGTGG 3′ 5′AAAAGGAAAAGCGGCCGCATGCTGCCGAATCTCACGTACTAGTGTG 3′

PCR primers for cloning of Ras (aa 1–166 via SfiI and NotI

G151UP G151LO

5′GCCAAGACCCGGCAGGGTGTAGAGGACGCGTTCTACACACTAGTACG 3′ 5′CGTACTAGTGTGTAGAACGCGTCCTCTACACCCTGCCGGGTCTTGGC 3′

QC primers for introduction of Gly 151 into Ras. Introduces AflIII site

FUHUP FUHLO

5′GGCCCAGCCGGCCATGGCCGCGGCCGCTGCCG 3′ 5′CAGCGGCCGCGGCCATGGCCGGCTG 3′

Makes fUSE5 compatible with pHEN1 and its derivatives

FUHNCOUP FUHNCOLO

5′ATTACGTTACTCGATGCGATGGGGATTGGCCTT 3′ 5′AAGGCCAATCCCCATCGCATCGAGTAACGTAAT 3′

QC primers for removal of a NcoI site in fUSE5

C51SUP C51SLO

5′CATTGATGGGGAGACATCTCTTCTAGACATCTTAG 3′ 5′CTAAGATGTCTAGAAGAGATGTCTCCCCATCAATG 3′

QC primers for C51S mutation in Ras. Removes AflIII site

C80LUP C80LLO

5′GGGGAGGGCTTCCTACTAGTATTTGCCATCAACAACACC 3′ 5′GGTGTTGTTGATGGCAAATACTAGTAGGAAGCCCTCCCC 3′

QC primers for C80L mutation in Ras. Introduces SpeI site

C118LUP C118LLO

5′GGTGGGCAAGCTTGACCTGGCCGCTCGC 3′ 5′GCGAGCGGCCAGGTCAAGCTTGTTGCCCACC 3′

QC primers for C118L mutation in Ras. Introduces HindIII site

REV M13back M13for

5′AAACAGCTATGACCATG 3′ 5′TGAATTTTCTGTATGAGGTTTTG3′ 5′TTTCGACACAATTTATCAGG 3′

Primers for sequencing

Oligo name HHUP HHLO RASSFI RASNOT2

a

QC primers are the complementart primer paid used with the Quickchange kit (Stratagene) to induce mutations.

phagemid or phage by propagation at 25 °C for 24 h. For the phagemids, VCSM13 (Stratagene) or KM13 [32] was used as helper-phage. Phage was concentrated from the culture supernatants by precipitation with NaCl and PEG6000 (final concentrations of 500 mM and 4% (w/v) respectively) at 4 °C for 1 h followed by centrifugation (15 min, 10 000 g). The phage was resuspended in 30 mL TBS (50 mM Tris, 100 mM NaCl, pH 7.5), remaining cell debris removed by centrifugation (15 min, 30 000 g) and the phage precipitated as above. Finally, the phage was resuspended in 1 mL TBS supplemented with 10% glycerol and stored at –20 °C. The titre was determined as colony forming units (cfu) by allowing serially diluted phage to infect log-phase E. coli strain TG1 and plating on 2 × TY agar-plates supplemented with the appropriate antibiotic.

wells were blocked with 1% skimmed milk powder and 5% glycerol in TBS (MTBSG) and 1010 cfu phage per well was incubated in MTBSG for 1 h. Glycerol was included to decrease non-specific phage adsorption [33]. In competition experiments 20 µg/mL soluble Ras(G12V) [34] was added along with the phage. After washing with TBS supplemented with 0.1% (v/v) Tween20, HRP-conjugated goat-anti-phage serum (Pharmacia) was diluted 2000-fold in MTBSG and added for detection of bound phage. After a final wash as above the reactions were for most purposes developed with OPD/H2O2 (KemEnTech), quenched with 1 vol 1 M H2SO4 and absorbance measured at 490 nm using an ELISA reader (Biorad). Alternatively, wells were added 100 µL chemiluminescence solution (Pierce) and emission detected with a light-sensitive film for 30 s [35].

2.4. Capture of phage particles with monoclonal antiRas antibodies

2.5. Western blotting

Wells of an ELISA plate (Nunc Maxisorp) were coated overnight at 4 °C with monoclonal anti-Ras antibody F111-85, Y13-259 or Y13-238 (Oncogene Research Products, 5 µg/mL) in 100 µL 50 mM NaHCO3, pH 9.6. The monoclonal anti c-Fos antibody 6-2H-2F (Santa Cruz Biotechnology) was included as negative control. Coated

Phage proteins from 5 × 1011 cfu were separated by SDS-PAGE (8% acrylamide) and transferred to nitrocellulose membranes (Hybond C, Amersham) by semi-dry electroblotting. After blocking with MTBSG, Ras mutants fused to g3p were detected with 1 µg/mL of the monoclonal anti-Ras antibody F111-85 (Oncogene Research Products) in MTBSG followed by HRP-conjugated rabbit-

1082 anti-mouse serum (DAKO) diluted 2000-fold in MTBSG. After each incubation with Ab the blot was washed with TBS supplemented with 0.1% (v/v) Tween-20. The blot was developed with the chemiluminescence system from Pierce. 2.6. Capture of phage from fUHras(C51S, C80L, C118L) with GST-RafRBD immobilized on plastic Wells of an ELISA plate (Nunc, Maxisorp) were coated overnight at 4 °C with 500 ng GST-RafRBD or GST (expressed from pGEX-2T-RafRBD [29] and pGEX1 [28], respectively) in 100 µL 50 mM NaHCO3, pH 9.6 and blocked with MTBSG. Phage produced from fUHras(C51S, C80L, C118L) was loaded with 10 µM GDP or GppNHp and Ras(G12V) was loaded with 10 µM GTP in TBS supplemented with 1 mM EDTA at 30 °C for 30 min. Nucleotide exchange was quenched with 10 mM (final concentration) MgCl2 and the phage (1010 cfu in 100 µL) was added to the coated wells with or without 0.5 µM GTP-loaded Ras(G12V). After 1 h of incubation, bound phage was detected as described above with OPD/H2O2. 2.7. Capture of phage from pHHras(C51S, C80L, C118L) with GST-RafRBD in solution Phage produced from pHHras(C51S, C80L, C118L) with KM13 as helper-phage were loaded with GppNHp or GDP as described above and 5 × 109 cfu allowed to equilibrate overnight at 4 °C with 1 nM GST-RafRBD in 200 µL TBS supplemented with 5% glycerol and 1% BSA. For competition, 0.5 µM GTP-loaded Ras(G12V) was included and in a parallel experiment GST-RafRBD was replaced with 1 nM GST (negative control). The reactions were transferred to wells of a 96-well ELISA plate (Nunc, Maxisorp) coated with polyclonal goat-antiGST serum (Pharmacia, diluted 1:1000 in 50 mM NaHCO3, pH 9.6) and blocked with MTBSG. After 1 h of incubation the wells were washed 20 times with TBS supplemented with 0.1% Tween and 10% glycerol over a period of 30 min. Bound phage were eluted with 1 mg/mL trypsin (Sigma) in 50 mM Tris, 1 mM CaCl2, pH 7.5 for 1 h at 37 °C. The eluted phage was allowed to infect E. coli TG1 and plated on 2 × TY agar-plates supplemented with 100 µg/mL ampicillin and 1% (w/v) glucose for titre-determination as cfu. 3. Results and discussion 3.1. Phage display of wild type Ras The cDNA coding for the 166 N-terminal residues of H-Ras (referred to as Ras in the following) was cloned for phage display in the phagemid pHH1, a modified version

Wind et al. of pHEN1 [30]. Ras was thereby fused to the N-terminus of the minor phage coat protein g3p and the resulting phagemid denominated pHHras (figure 1A). This region of Ras was chosen as it comprises both the catalytic GTPase domain and the so-called effector region (residues 32–40) known to be the common docking site for Ras effectors [36]. Phage particles were produced from E. coli by rescue of the phagemid with VCSM13 helper-phage and subsequently tested for the display of Ras. To this end, monoclonal anti-Ras antibodies were used for capture of phage followed by detection with antibodies to proteins of the phage coat in a sandwich ELISA. When the monoclonal anti-Ras antibody Y13-238 was used as bait, display could be detected only with a sensitive assay that combines ELISA and chemiluminescence [35], whereas capture with the F111-85 antibody was readily detectable with the less sensitive OPD/H2O2 system (data not shown). As opposed to the F111-85 antibody, which recognizes a linear epitope within the first 21 residues of Ras, Y13-238 is believed to recognize a conformational epitope on Ras, hence these results suggested that Ras was either mostly displayed in a misfolded conformation or that the display format resulted in steric hindrance of the Y13-238 antibody-Ras interaction. To elucidate the latter possibility, the phagemid pNras was constructed in order to improve the display of Ras (figure 1B). While pHHras resulted in fusion of Ras to full-length g3p, pNras has Ras fused to truncated g3p comprising only a glycine rich linker (aa 219–256) plus the C-terminal domain (aa 257–406) of g3p responsible for anchorage in the phage coat. This approach has previously been shown to improve the display of human growth hormone on phage [37]. Although this truncation of g3p gave a moderate improvement of display, the interaction with monoclonal anti-Ras antibodies towards conformational epitopes was detectable only with the chemiluminescence system (data not shown). Taken together these observations indicated that Ras was displayed on phage in a corrupt conformation and prompted us to consider the possibility of disruptive disulphide bridge formation involving the normally reduced cysteines in Ras. 3.2. Phage display of cysteine-free Ras The structures of mammalian Ras in complex with GDP [38] or the non-hydrolysable GTP analogue GppNHp [39] reveal that the three cysteines at positions 51, 80 and 118 play no role in nucleotide binding. Furthermore, they are not conserved in the two homologues from S. cerevisiae (RAS1 and RAS2) suggesting that they can be mutated to the corresponding residues of yeast RAS (serine, leucine and leucine, respectively) without compromising the activity of Ras. In support of this notion, mammalian H-Ras and S. cerevisiae RAS1 and RAS2 show conserved biochemical properties [40] and the 153

Phage display of Ras

1083

Figure 1. Constructs for phage display of Ras. A. pHHras was created by inserting a 6 × His tag into pHEN1 and cloning of the catalytic domain of H-Ras (residues 1–166) via SfiI and NotI thereby enabling monovalent display of Ras. The pelB leader sequence ensures translocation to the periplasmic space where it is proteolytically removed; the amber codon TAG allows expression of soluble Ras in non-suppresser strains of E. coli and the c-myc tag (EQKLISEEDLN) allows immuno-detection with the monoclonal antibody 9E10. B. pNras was the result of cloning the Ras-insert from pHHras into another pHEN1 derivative via NcoI and EagI resulting in monovalent display by fusion to truncated g3p. C. For polyvalent display of cysteine-free Ras, the fUSE5 phage was modified to be compatible with pHEN1 and its derivatives and Ras was cloned via NcoI and NotI. The leader sequence in the resulting phage comprises the wild type g3p leader and the six C-terminal residues from the pelB leader; whether the latter were proteolytically processed prior to phage assembly has not been determined.

N-terminal amino acids of mammalian H-Ras can be replaced functionally by the corresponding region from RAS1 [41]. In addition, none of these cysteines are involved in intramolecular hydrogen bonds via their thiol-group: C118 is surface-exposed whereas C51 and C80 are facing α-helix 5 and 3, respectively, both with the nearest neighbor to the thiol-group more than 3.5 Å away. To assess whether the replacement of C80 with the more bulky leucine could alter the conformation of Ras, the surroundings of C80 in the protein-structure of Ras were evaluated [38, 39]. First, the side chain of C80 is located in a predominantly hydrophobic environment composed of the side chains from V9, F78, Y96, I100 and M111, all identified as being closer than 5 Å to the thiol-group of C80. Second, the van der Waals surfaces of these hydrophobic side chains constitute a cavity around the thiolgroup of C80 sufficiently large to accommodate a leucine side chain. Third, the residues making up this cavity are conserved between mammalian Ras and S. cerevisiae RAS1, except M111 being a valine in RAS1, suggesting a similar environment for C80 in mammalian Ras and the

corresponding L87 in yeast RAS1. Taken together, these observations, which are valid for both the GDP- and GTP-bound conformation of Ras, advocate that the C80L mutation in mammalian Ras does not jeopardize the integrity of the protein. For the two other cysteines in mammalian Ras, their mutation to the corresponding residues from RAS1 should not impose any steric problems as C51 is replaced with the homologous amino acid serine and as C118, replaced with leucine, is surfaceexposed. Consequently the three cysteines were mutated to the corresponding residues in S. cerevisiae RAS1, i.e., the mutations C51S, C80L and C118L were introduced into pHHras in the seven possible combinations. As shown in figure 2, two or three cysteine replacements dramatically improved the quality of display as evaluated by capture ELISA with the monoclonal antibodies Y13-259 and Y13-238, both recognizing conformational epitopes on Ras. Of the three cysteines, C80 appears to be the most disruptive for display since its sole replacement results in a detectable improvement in the ELISA. In accordance with this, the double mutation that

1084

Figure 2. Antibody-capture of Ras-bearing phage. Phage from pHH1 (negative) and pHHras with all possible combinations of Cys or Ser at position 51; Cys or Leu at position 80 and Cys or Leu at position 118 were captured with 500 ng of the monoclonal anti-Ras antibodies Y13-259 (upper panel) or Y13-238 (lower panel) coated in wells of an ELISA plate followed by detection with HRP-conjugated anti-phage antibody and OPD/H2O2. Amino acids replacing wildtype cysteines are underlined. All experiments were performed with (shaded bars) or without (black bars) soluble Ras(G12V) as competitior. The average and standard deviation of three experiments are shown.

leaves C80 intact (C51S, C118L) shows less improvement of display than the other two (figure 2). What makes C80 more prone to jeopardize the correct folding of Ras when displayed on phage remains elusive, but it may exist in a sequence context making it more reactive in terms of disulphide bridge formation with cysteines from g3p or other periplasmic proteins. To ensure that these observations were indeed the consequences of better folding of Ras, and not the result of a greater proportion of the phage particles actually carrying a Ras-g3p fusion, the phage coat proteins were analyzed by Western blotting. As seen from figures 2 and 3, there is no correlation between ELISA signal and amount of full-length Ras-g3p in the fusion-phage preparations. The improvements observed in the ELISA upon cysteine replacement can therefore not be ascribed to a quantitative increase in display level. Hence, removal of cysteines causes a qualitative improvement of Ras display.

Wind et al.

Figure 3. Western blot of phage proteins. Proteins from 5 × 1011 cfu phage, produced from pHH1 (negative control, lane 1), pHHras (lane 2) or the four pHHras derivatives with two or three cysteines removed (lanes 3–6, deviations from wild type underlined), were separated by SDS-PAGE and probed with the monoclonal antibody F111-85, recognizing an epitope between residues 1 and 21 in Ras. The full-length Ras-g3p fusion is indicated with an arrow and its migration (≈ 90 kDa) is in accordance with the apparent molecular mass of g3p (≈ 63 kDa) plus the weight of Ras (aa 1–166) and the 6 × His- and c-myc tags (≈ 22 kDa).

3.3. Stability of the Ras-g3p fusion protein In addition to full-length Ras-g3p with an apparent molecular mass of 90 kDa, the Western blot reveals a faster migrating dominant band (figure 3). This extra band is recognized by the F111-85 antibody, which is directed towards the N-terminal portion of Ras. Consequently, the extra band results from a C-terminal truncation of the Ras-g3p fusion. Furthermore, the apparent size of the truncated protein, 70–75 kDa, indicates that a fragment corresponding to 15–20 kDa is missing. A similar phenomenon has been observed with phagemid-based display of staphylokinase (Sak) where a Western blot of Sakphage proteins with an anti-Sak antibody revealed a faster migrating band in addition to the expected full-length Sak-g3p fusion product [42]. As in the present study, the apparent size of the truncated Sak-g3p fusion was 15–20 kDa lower than that of the full-length protein. Furthermore, the presence of a truncated g3p-fusion with

Phage display of Ras approximately 20 kDa missing from the C-terminus was observed in preparations of fusion-phage displaying an alkaline phosphatase monomer in a phage-based system [43]. Such truncated versions of the g3p-fusion proteins, however, are not a general trait as only the fulllength protein was present in preparations of fusion-phage displaying hirudin [44]. Considering the structure of g3p, which contains three domains (N1, N2 and CT) separated by glycine-rich linkers [45], it appears that the C-terminal domain of g3p (CT; aa 257–406, ≈ 16 kDa) is missing from the truncated variants mentioned above. As this domain is responsible for anchorage of g3p in the phage coat, its absence prohibits direct incorporation of the truncated protein into the phage particle. Thus, the presence of Ras-N1-N2 protein in the phage preparations can be explained either as the product of a proteolytic event occurring after purification or as co-purification of the truncated protein along with the phage particles. Intriguingly, Western blots of lysates from bacteria harboring pHHras revealed the presence of Ras-N1-N2 prior to infection with helperphage and phage assembly (data not shown). A similar observation was reported for bacteria harboring a phagemid encoding a recombinant single-chain antibody (scFv) fused to g3p: in addition to the full-length scFv-g3p protein these bacteria contained two truncated versions of which one was missing ≈ 15 kDa from the C-terminal end [46]. Taken together, these data suggest that fusion of foreign proteins to the N-terminus of g3p can result in appearance of a truncated version lacking the CT domain, e.g., Ras-N1-N2, that can co-purify with phage particles. Furthermore, as N1 and N2 are engaged in extensive interactions in normal g3p [45], it may be speculated that the Ras-N1-N2 protein adsorbs to phage particles through intermolecular N1-N2 association with g3p from the phage. The influence of the truncated g3p-fusion protein on eventual selections from a Ras-repertoire can be expected to be minimal as such products are likely to be present in many of the previously reported phage display systems that nevertheless have been used with success. 3.4. Capture of cysteine-free Ras-phage with the Rasbinding domain from Raf1 To ensure that the removal of cysteines and the display on phage did not compromise Ras function we wanted to establish whether phage displaying cysteine-free Ras could interact with the Ras-binding domain (RBD) from Raf1 expressed as a GST-fusion protein [29]. Phage were produced from E. coli harboring pHHras(C51S, C80L, C118L) by rescue of the phagemid with VCSM13 and loaded with the non-hydrolysable GTP analogue GppNHp. These phage were subsequently allowed to interact with GST-tagged RafRBD immobilized on a plastic surface. However, we failed to detect any captured phage

1085 with the standard OPD/H2O2 ELISA detection system (data not shown). Consequently, we chose to change the display format from phagemid-based to phage-based in order to impose a quantitative improvement of Ras display. The phage vector fUSE5 [31] was made compatible with pHEN1 and its derivatives, thereby creating fUH1, and Ras(C51S, C80L, C118L) was cloned for display (figure 1C). In theory, this results in as many as five copies of Ras per phage particle, but again some proteolytic instability of the Ras-g3p fusion product must be expected [43]. Phage were produced from the resulting fUHras(C51S, C80L, C118L) and loaded with either GDP or GppNHp prior to capture on coated GST-RafRBD. As shown in figure 4A, phage displaying cysteine-free Ras loaded with GppNHp were readily captured by GSTRafRBD but not by GST alone, and the specificity of the interaction was confirmed by competition with GTP loaded Ras(G12V). This result demonstrates that Ras(C51S, C80L, C118L) is displayed in its native conformation when it is fused to g3p. That capture of cysteine-free Ras by immobilized GST-RafRBD could only be detected using the phagebased display as opposed to the phagemid-based display suggests a poor display level of Ras in the latter system. Hence, a sensitive micro-panning assay was used to confirm functional phagemid-based display of Ras. Phage were rescued from E. coli harboring pHHras(C51S, C80L, C118L) with the modified helper-phage KM13 [32]. This helper-phage is characterized by a trypsin cleavage-site introduced into g3p rendering phage carrying no g3pfusions from the phagemid non-infectious upon trypsin treatment. Consequently, as the majority of the phage particles carries only g3p from the helper-phage (> 95%, data not shown), treatment with trypsin dramatically decreases the infectivity of non-specific background phage and thereby increases the signal-to-noise ratio when infectivity is used as a measure of phage-presence. Furthermore, as the c-myc-tag in pHH1 is susceptible to trypsin-cleavage, treatment with trypsin will simultaneously elute specific binders (see figure 1A). Importantly, when using KM13 as helper-phage, the infectivity of the trypsin-treated phage relies on the g3p portion of the fusion protein encoded by the phagemid, hence the pNras construct with Ras fused to a truncated version of g3p (figure 1B) could not be used in this context. Phage produced from pHHras(C51S, C80L, C118L) with KM13 as helper-phage were loaded with either GDP or GppNHp and allowed to bind GST-RafRBD in solution. Subsequently, any formed complexes were captured through polyclonal anti-GST serum and the presence of phage determined by treatment with trypsin and titration in E. coli. As shown in figure 4B, this assay revealed a nucleotide-dependent interaction between phage from pHHras(C51S, C80L, C118L) and GST-RafRBD that could be competed with soluble Ras(G12V)-GTP thereby

1086

Wind et al. affinity. In contrast, selections from phage-based repertoires suffers from avidity effects when fusion-phage can bind the bait via multiple displayed molecules. Furthermore, the use of trypsin-elution in combination with the KM13 helper-phage will increase the ratio between specific and non-specific phage eluted from each panning cycle and hence decreases the required number of iterations. In conclusion, functional display of cysteine-free Ras on filamentous phage has been demonstrated. Hence, a system has been designed that may be used for generation of novel Ras loss-of-function mutants with defined effector preferences. Furthermore, our results indicate that unpaired cysteines can constitute a significant problem in phage display approaches but can be dealt with through heuristic replacements. Acknowledgments We thank Greg Winter and Hennie Hoogenboom for the phagemid pHEN1 and the E. coli strain TG1; Peter Kristensen for the KM13 helper-phage; Berthe Willumsen for the vector pBW1699; Alfred Wittinghofer for recombinant Ras(G12V) and helpful comments; George Smith for the fUSE5 phage and Christian Herrmannn for the GST-RafRBD expression vector. This work was supported by The Danish Natural Science Research Council.

Figure 4. Capture of Ras bearing phage with GST-RafRBD. A. The Ras binding domain from Raf as a GST-fusion (GSTRafRBD) was coated in wells of an ELISA plate. Phage produced from the phage fUHras(C51S, C80L, C118L) was loaded with either GDP or the GTP-analogue GppNHp and 1010 cfu added to the wells in 100 µL buffer. GppNHp-loaded phage were in parallel experiments competed with 0.5 µM GTP-loaded Ras(G12V) or added to wells coated with GST. The average and standard deviation of three experiments are shown. B. Phage produced from the phagemid pHHras(C51S, C80L, C118L) with KM13 helper-phage was loaded with either GDP or GppNHp and 5 × 109 cfu allowed to equilibrate with 1 nM GST-RafRBD in 200 µL buffer. GppNHp-loaded phage were in parallel experiments competed with 0.5 µM Ras(G12V)-GTP or incubated with 1 nM GST. Formed complexes were subsequently captured on immobilized goat-anti-GST serum and phage eluted with trypsin. Output titres were normalized to that from incubation with GST. The average and standard deviations of three experiments are shown.

demonstrating functional phagemid-based display of cysteine-free Ras. When compared to the phage-based system, phagemidbased display of Ras(C51S, C80L, C118L) holds the promise of being superior as this system through its low display level (< 1 Ras molecule per phage, [47]) allows selection to be performed based on one-to-one binding

References [1] Lowy D.R., Willumsen B.M., Function and regulation of ras, Annu. Rev. Biochem. 62 (1993) 851–891. [2] Boguski M.S., McCormick F., Proteins regulating Ras and its relatives, Nature 366 (1993) 643–654. [3] Wittinghofer A., Pai E.F., The structure of Ras protein: a model for a universal molecular switch, Trends Biochem. Sci. 16 (1991) 382–387. [4] Joneson T., Bar-Sagi D., Ras effectors and their role in mitogenesis and oncogenesis, J. Mol. Med. 75 (1997) 587–593. [5] Chien C.T., Bartel P.L., Sternglanz R., Fields S., The two-hybrid system: a method to identify and clone genes for proteins that interact with a protein of interest, Proc. Natl. Acad. Sci. USA 88 (1991) 9578–9582. [6] White M.A., Nicolette C., Minden A., Polverino A., Van Aelst L., Karin M., Wigler M.H., Multiple Ras functions can contribute to mammalian cell transformation, Cell 80 (1995) 533–541. [7] Masuda T., Kariya K., Shinkai M., Okada T., Kataoka T., Protein kinase Byr2 is a target of Ras1 in the fission yeast Schizosaccharomyces pombe, J. Biol. Chem. 270 (1995) 1979–1982. [8] Khosravi-Far R., White M.A., Westwick J.K., Solski P.A., Chrzanowska-Wodnicka M., Van Aelst L., Wigler M.H., Der C.J., Oncogenic Ras activation of Raf/mitogen-activated protein kinaseindependent pathways is sufficient to cause tumorigenic transformation, Mol. Cell Biol. 16 (1996) 3923–3933. [9] Joneson T., White M.A., Wigler M.H., Bar-Sagi D., Stimulation of membrane ruffling and MAP kinase activation by distinct effectors of RAS, Science 271 (1996) 810–812. [10] Rodriguez-Viciana P., Warne P.H., Khwaja A., Marte B.M., Pappin D., Das P., Waterfield M.D., Ridley A., Downward J., Role of phosphoinositide 3-OH kinase in cell transformation and control of the actin cytoskeleton by Ras, Cell 89 (1997) 457–467.

Phage display of Ras [11] Smith G.P., Scott J.K., Libraries of peptides and proteins displayed on filamentous phage, Methods Enzymol. 217 (1993) 228–257. [12] Marvin D.A., Filamentous phage structure, infection and assembly, Curr. Opin. Struct. Biol. 8 (1998) 150–158. [13] Missiakas D., Raina S., Protein folding in the bacterial periplasm, J. Bacteriol. 179 (1997) 2465–2471. [14] Lowman H.B., Wells J.A., Affinity maturation of human growth hormone by monovalent phage display, J. Mol. Biol. 234 (1993) 564–578. [15] Toniatti C., Cabibbo A., Sporena E., Salvati A.L., Cerretani M., Serafini S., Lahm A., Cortese R., Ciliberto G., Engineering human interleukin-6 to obtain variants with strongly enhanced bioactivity, EMBO J. 15 (1996) 2726–2737. [16] Ballinger M.D., Jones J.T., Lofgren J.A., Fairbrother W.J., Akita R.W., Sliwkowski M.X., Wells J.A., Selection of heregulin variants having higher affinity for the ErbB3 receptor by monovalent phage display, J. Biol. Chem. 273 (1998) 11675–11684. [17] Saggio I., Gloaguen I., Poiana G., Laufer R., CNTF variants with increased biological potency and receptor selectivity define a functional site of receptor interaction, EMBO J. 14 (1995) 3045–3054. [18] Choo Y., Klug A., Toward a code for the interactions of zinc fingers with DNA: selection of randomized fingers displayed on phage, Proc. Natl. Acad. Sci. USA 91 (1994) 11163–11167. [19] Jamieson A.C., Kim S.H., Wells J.A., In vitro selection of zinc fingers with altered DNA-binding specificity, Biochemistry 33 (1994) 5689–5695. [20] Rebar E.J., Pabo C.O., Zinc finger phage: affinity selection of fingers with new DNA-binding specificities, Science 263 (1994) 671–673. [21] Dennis M.S., Lazarus R.A., Kunitz domain inhibitors of tissue factor-factor VIIa. II. Potent and specific inhibitors by competitive phage selection, J. Biol. Chem. 269 (1994) 22137–22144. [22] Wang C.I., Yang Q., Craik C.S., Isolation of a high affinity inhibitor of urokinase-type plasminogen activator by phage display of ecotin, J. Biol. Chem. 270 (1995) 12250–12256. [23] Cunningham B.C., Lowe D.G., Li B., Bennett B.D., Wells J.A., Production of an atrial natriuretic peptide variant that is specific for type A receptor, EMBO J. 13 (1994) 2508–2515. [24] Ames R.S., Tornetta M.A., Jones C.S., Tsui P., Isolation of neutralizing anti-C5a monoclonal antibodies from a filamentous phage monovalent Fab display library, J. Immunol. 152 (1994) 4572–4581. [25] Parsons H.L., Earnshaw J.C., Wilton J., Johnson K.S., Schueler P.A., Mahoney W., McCafferty J., Directing phage selections towards specific epitopes, Protein Eng. 9 (1996) 1043–1049. [26] Stausbøl-Grøn B., Wind T., Kjær S., Kahns L., Hansen N.J., Kristensen P., Clark B.F., A model phage display subtraction method with potential for analysis of differential gene expression, FEBS Lett. 391 (1996) 71–75. [27] Hansen N.J., Kristensen P., Lykke J., Mortensen K.K., Clark B.F., A fast, economical and efficient method for DNA purification by use of a homemade bead column, Biochem. Mol. Biol. Int. 35 (1995) 461–5. [28] Smith D.B., Johnson K.S., Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase, Gene 67 (1988) 31–40. [29] Herrmann C., Martin G.A., Wittinghofer A., Quantitative analysis of the complex between p21ras and the Ras-binding domain of the human Raf-1 protein kinase, J. Biol. Chem. 270 (1995) 2901–2905.

1087 [30] Hoogenboom H.R., Griffiths A.D., Johnson K.S., Chiswell D.J., Hudson P., Winter G., Multi-subunit proteins on the surface of filamentous phage: methodologies for displaying antibody (Fab) heavy and light chains, Nucleic Acids Res. 19 (1991) 4133–4137. [31] Scott J.K., Smith G.P., Searching for peptide ligands with an epitope library, Science 249 (1990) 386–390. [32] Kristensen P., Winter G., Proteolytic selection for protein folding using filamentous bacteriophages, Fold. Des. 3 (1998) 321–328. [33] Kjær S., Stausbøl-Grøn B., Wind T., Ravn P., Jensen K.H., Kahns L., Clark B.F., Glycerol diversifies phage repertoire selections and lowers non-specific phage absorption, FEBS Lett. 431 (1998) 448–452. [34] John J., Frech M., Wittinghofer A., Biochemical properties of Ha-ras encoded p21 mutants and mechanism of the autophosphorylation reaction, J. Biol. Chem. 263 (1988) 11792–11799. [35] Schmid J.A., Billich A., Simple method for high sensitivity chemiluminescence ELISA using conventional laboratory equipment, Biotechniques 22 (1997) 278–280. [36] Marshall M.S., The effector interactions of p21ras, Trends Biochem. Sci. 18 (1993) 250–254. [37] Lowman H.B., Bass S.H., Simpson N., Wells J.A., Selecting high-affinity binding proteins by monovalent phage display, Biochemistry 30 (1991) 10832–10838. [38] Tong L.A., de Vos A.M., Milburn M.V., Kim S.H., Crystal structures at 2 A resolution of the catalytic domains of normal ras protein and an oncogenic mutant complexed with GDP, J. Mol. Biol. 217 (1991) 503–516. [39] Pai E.F., Krengel U., Petsko G.A., Goody R.S., Kabsch W., Wittinghofer A., Refined crystal structure of the triphosphate conformation of H-ras p21 at 1.35 A resolution: implications for the mechanism of GTP hydrolysis, EMBO J. 9 (1990) 2351–2359. [40] Temeles G.L., Gibbs J.B., D’Alonzo J.S., Sigal I.S., Scolnick E.M., Yeast and mammalian ras proteins have conserved biochemical properties, Nature 313 (1985) 700–703. [41] DeFeo-Jones D., Tatchell K., Robinson L.C., Sigal I.S., Vass W.C., Lowy D.R., Scolnick E.M., Mammalian and yeast ras gene products: biological function in their heterologous systems, Science 228 (1985) 179–184. [42] Jespers L., Jenne S., Lasters I., Collen D., Epitope mapping by negative selection of randomized antigen libraries displayed on filamentous phage, J. Mol. Biol. 269 (1997) 704–718. [43] McCafferty J., Jackson R.H., Chiswell D.J., Phage-enzymes: expression and affinity chromatography of functional alkaline phosphatase on the surface of bacteriophage, Protein Eng. 4 (1991) 955–961. [44] Wirsching F., Opitz T., Dietrich R., Schwienhorst A., Display of functional thrombin inhibitor hirudin on the surface of phage M13, Gene 204 (1997) 77–84. [45] Lubkowski J., Hennecke F., Pluckthun A., Wlodawer A., The structural basis of phage display elucidated by the crystal structure of the N-terminal domains of g3p, Nat. Struct. Biol. 5 (1998) 140–147. [46] Breitling F., Dubel S., Seehaus T., Klewinghaus I., Little M., A surface expression vector for antibody screening, Gene 104 (1991) 1 47–53. [47] Lowman H.B., Bacteriophage display and discovery of peptide leads for drug development, Annu. Rev. Biophys. Biomol. Struct. 26 (1997) 401–424.