Using affinity capillary electrophoresis to study the interaction of the extracellular binding domain of erythropoietin receptor with peptides

J. Biochem. Biophys. Methods 40 (1999) 17–25

Using affinity capillary electrophoresis to study the interaction of the extracellular binding domain of erythropoietin receptor with peptides a,b ,

a,b

a,b

Gary W. Caldwell *, Patricia A. McDonnell , John A. Masucci , Dana L. Johnson a,b , Linda K. Jolliffe a,b a

The R.W. Johnson Pharmaceutical Research Institute, Welsh and McKean Roads, Drug Discovery, Spring House, PA, USA b The R.W. Johnson Pharmaceutical Research Institute, Drug Discovery, Raritan, NJ, USA Received 14 October 1998; received in revised form 9 January 1999; accepted 18 January 1999

Abstract We have shown that affinity capillary electrophoresis (ACE) can be utilized to screen peptides that bind to the extracellular binding domain of the erythropoietin receptor (EBP). The comparison of the cyclic peptides GGTYSCHFGPLTWVCKPQGG (EMP1) GGTYSCHFGPLTAVCKPQGG (EMP13), and LGRKYSCHFGPLTWVCQPAKKD (EMP37) with the linear peptides HFGPLTWV (EMP26) and FMRF as ACE buffer additives were investigated. When EMP1 and EMP37 were the buffer additives, an abrupt change in the electrophoretic mobility of EBP was observed in the electropherogram. When EMP13, EMP26, and FMRF were examined under identical ACE conditions as EMP1 and EMP37, no significant change in the electrophoretic mobility of EBP was observed. These results correlate well with previously reported IC 50 competitive binding data; that is, EMP1 and EMP37 bind to EBP while EMP13 and EMP26 bind very weakly. These observations strongly infer that peptide ? EBP dimerization were induced by EMP1, and EMP37 but not by EMP13, EMP26 or FMRF. This ACE method provides a rapid tool for the detection of small peptides or drugs that bind to EBP.  1999 Elsevier Science B.V. All rights reserved. Keywords: Erythropoietin; Capillary electrophoresis; ACE

*Corresponding author. Tel.: 1 1-215-628-5537; fax: 1 1-215-628-7064. E-mail address: [email protected] (G.W. Caldwell) 0165-022X / 99 / $ – see front matter  1999 Elsevier Science B.V. All rights reserved. PII: S0165-022X( 99 )00011-1

18

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

1. Introduction It has been demonstrated that affinity capillary electrophoresis (ACE) is a useful method for measuring binding affinities of charged ligands to proteins [1–11]. In an ACE experiment, a charged ligand is added into the electrophoresis buffer at various concentrations. A protein is injected onto the capillary where it can interact with the charged ligand during the electrophoretic analysis. The binding of a charged ligand to the protein will change the electrophoretic mobility of the protein due to a change in the protein’s charge-to-coefficient of friction ratio [1,2]. The measurement of the electrophoretic mobility of the protein in the presence and absence of ligand allows for a quantitative determination of binding constants [3–11]. The hormone erythropoietin (EPO) is produced primarily in the adult kidney, and fetal liver in response to hypoxia and circulates in the blood stream where it targets its receptor (EPOR) [12]. The EPOR belongs to the hematopoietic cytokine receptor superfamily whose members have a three domain organization [13]. The three domains are an extracellular ligand binding domain, a cytoplasmic segment and a single transmembrane portion. Receptor activation is thought to occur after homo-dimerization of two EPOR molecules with one EPO molecule [14]. This homo-dimerization process triggers a cascade of events that results in red blood cell production [12,14–17]. There has been considerable interest in understanding the interaction of EPO with EPOR and the resulting homo-dimerization process. The soluble extracellular ligand binding domain of EPOR (i.e. a 225 residue EPO binding protein, denoted EBP) has been expressed in Esherichia coli and was shown to retain similar EPO binding activity [15]. A family of disulfide-linked cyclic peptides has been identified that bind to EBP, and function in vitro and in vivo as mimetics of EPO [16,18]. From this study, the EPO mimetic peptide 1 (EMP1), which is a 20-residue cyclic peptide with the sequence GGTYSCHFGPLTWV CKPQGG, and the cyclic peptide LGRKYSCHFGPLTWVCQPAKKD (EMP37) were identified. The three-dimensional X-ray structure of ˚ resolution [17]. In the crystal an EBP-EMP1 complex was also determined at 2.8 A structure, two EMP1 molecules induced an almost perfect two-fold dimerization of two EBP molecules (i.e. [EBP ? EMP1] ? [EMP1 ? EBP]). It was also demonstrated that EMP1 mediates formation of a soluble EBP dimer complex in solution. We demonstrate that ACE can detect peptides that bind to the extracellular binding domain of erythropoietin (EBP). The ACE procedure provides a convenient method to screen for peptides or drugs that bind to EBP.

2. Material and methods

2.1. Apparatus ACE experiments were performed with a Beckman P/ACE Model 2200 CE system (Fullerton, CA, USA). A neutral coated fused silica capillary (Beckman CA, USA), 45 cm 3 50 mm i.d., with an effective separation length of 37 cm was utilized. Neutral

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

19

coated capillaries were used to minimize the adsorption of peptides and proteins to the capillary walls. The operating temperature for the capillary chamber was set at 25618C. To reduce the adsorption of proteins / peptides to glass, all buffer reservoirs contained a plastic insert. The electrophoresis buffer consisted of 20 mM citrate solution at pH 5 7.2. Because the pH of the solution was greater than the pI of EBP (pI | 5.8) [16–18] the protein was negatively charged. Thus, the capillary electrophoresis unit was operated at 18.5 kV in the reverse polarity mode with the positive electrode at the outlet reservoir and the negative electrode at the inlet reservoir. Pressure injection was utilized with a duration of 3 s that represented about 10 nl of sample volume injected. The elutions were monitored on-capillary with the UV detector at a wavelength of 214 nm. The electropherograms were collected and analyzed with the Beckman P/ACE Windows controller software.

2.2. Reagents Sodium citrate was purchased from Aldrich (Milwaukee, WI, USA). The EBP was produced as described in the literature [16–18] and was stored at 08C in 10 mM sodium phosphate buffer, 150 mM NaCl, 1 mM EDTA at pH 5 7.2 [7]. The EBP concentration was approximately 35 mM in all experiments. More concentrated solutions of EBP were obtained using a Microcon-10 (Amicon, Inc.) at 48C at 10 000 g. Typical centrifugation times for 150 ml of 0.87 mg / ml EBP stock solution were 8 min. The peptide FMRF was purchased from PCR (Gainesville, FL, USA). The GGTYSCHFGPLTWVCKPQGG (EMP1), GGTYSCHFGPLTAVCKPQGG (EMP13), LGRKYSCHFGPLTWVC QPAKKD (EMP37) and HFGPLTWV (EMP26) peptides were synthesized by standard solid phase techniques [19]. EMP1, EMP13, and EMP37 were cyclized through a disulfide bond at the cysteine residues and amidated at the COOH-terminus [16].

2.3. Mass spectrometry The EBP sample was analyzed by electrospray ionization (ES) mass spectrometry. A 20 ml aliquot of a 0.87 mg / ml EBP solution was desalted using standard reversed-phase HPLC (Varian 9010, ODS column) techniques [20]. The desalted protein fraction contained in 50 / 50 / 0.5 water / acetonitrile / trifluoroacetic acid was infused at 6 ml / min into the electrospray ion source of a VG TRIO 2000 quadrupole mass spectrometer (Micromass, Manchester, UK). The measured monoisotopic mass of EBP was 24 714624 Da (calculated monoisotopic mass 24 713 Da). The EBP sample was also analyzed by matrix assisted laser desorption ionization (MALDI) time-of-flight (TOF) mass spectrometry. A 1 ml EBP solution (0.87 mg / ml) was combined with 1 ml of a saturated solution of a-cyano hydroxycinnamic acid. The mixture was spotted on a stainless steel target, evaporated, and irradiated at 337 nm with an N 2 laser utilizing a VG Tofspec E mass spectrometer (Micromass, Manchester, UK). The resulting ions indicated an average molecular mass of 24 8746150 Da (calculated average mass 24 728 Da).

20

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

2.4. Procedures The ACE determination was carried out by preparing a series of citrate solutions containing various concentrations of peptides (0.05–200 mM). About 10 nl of EBP (approximately 35 mM) was injected onto the capillary utilizing these different solutions and the electropherogram recorded. Affinity interaction between EBP and the peptide could be examined through the measurement of migration shifts of EBP. A simple experiment was designed to determine when significant peptide adhesion to the capillary wall occurred. The adhesion of peptides to the capillary wall will produce an EBP affinity column. In the absence of peptide, the electrophoretic mobility of EBP was measured for a 10 nl injection and this experiment served as the control. Next, a 20 mM citrate solution containing a known concentration of peptide (0.05–200 mM) was placed into the capillary and allowed to contact the capillary wall for about 20 min. The capillary was then rinsed with a fresh solution of 20 mM citrate solution. About 10 nl of EBP was injected on the capillary and the electropherogram recorded. If an equilibrium between EBP and the capillary wall is established during the electrophoretic analysis, one would expect a change in either the peak shape and / or a change in the electrophoretic mobility of EBP when compared to the control experiment. In between experiments, the capillary column was then rinsed with 0.1 N HCl for 30 s and rinsed again with 20 mM citrate solution prior to performing the next experiment.

3. Results and discussion Under our ACE conditions, two EBP peaks (denoted EBP(1) and EBP(2)) were resolved as shown in the electropherogram in Fig. 1. These two peaks likely represent multiple isoelectric forms of EBP and / or possibly multiple EBP molecules with different masses. Because of the resolution afforded by CE, it is not unusual that different isoforms of a protein possess different mobilities [21]. When the EBP sample was allowed to remain in water at room temperature over several days, it was observed that the ratio of these two peaks changed with time (data not shown). The later eluting peak decreased in area while the earlier eluting peak increased indicating that the EBP charge state was changing. If one or more of the five asparagine or the three glutamine side-chain amide groups hydrolyzed, then multiple isoforms of EBP with different masses would result [22]. Considering the above observation, the two peaks observed in the ACE experiment are probably due to either the asparagine and / or glutamine groups being hydrolyzed to acids in solution. The EBP sample was analyzed by ES and MALDI-TOF mass spectrometry. A single species was observed that had a monoisotopic mass of 24 714624 Da, which compared well with the theoretical value of EBP (24 713 Da). A mixture of proteins with a mass difference between 1 and 8 Da would not be detected utilizing our low resolution mass spectrometers. In Fig. 1, we demonstrated that the ACE assay can be used to screen peptides that bind to EBP. As the concentration of EMP1 (GGTYSCHFGPLTWVCKPQGG) is increased, from 0.5 to 40 mM, the EBP peak slightly broadens but no significant migration shift of EBP is observed. At 50 mM of EMP1, the EBP peaks are significantly

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

21

Fig. 1. Affinity capillary electrophoresis of a mixture of isoforms of extracellular ligand binding domain of erythropoietin receptor (denoted EBP) as a function of EMP1 (GGTYSCHFGPLTWVCKPQGG) buffer concentration. The electrophoresis buffer used was 20 mM citrate solution at pH 7.2. The total length of the capillary was 45 cm (37 cm from injection to detection) being operated at 18.5 kV. The inverted peak is due to citrate anions not being present in the injection plug but present in the buffer and therefore, is used as the standard reference marker. It is also noted that the electrophoretic mobility of the inverted peak did not change significantly with increases in the concentration of EMP1. This observation is consistent with the EOF being constant during these ACE experiments and that minimal adsorption of EMP1 onto the capillary wall occurred.

broader and an abrupt change in the electrophoretic mobility of EBP(1) and EBP(2) has occurred (DmP,L 5 2 3.1 3 10 25 cm 2 / Vs). This migration is due to the complexation of EBP to EMP1 resulting in a change in electrophoretic mobility (DmP,L 5 mP,L 2 mP ) of EBP(1) and EBP(2), where mP is the electrophoretic mobility of EBP(1) and (2) in the absence of EMP1 and mP,L is the electrophoretic mobility of EMP1 at 50 mM. Additionally when LGRKYSCHFGPLTWVCQPAKKD (EMP37) was used in the buffer, similar results were observed. However, the abrupt change in the mobility of EBP was observed at approximately 20 mM. Since EMP1 and EMP37 are known to have binding affinities (i.e. IC 50 5 mM, Table 1) with EBP [17,18], these changes in DmP,L and peak shapes are interpreted to occur due to the binding of EBP to these peptides. As a control, the peptides GGTYSCHFGPLTAVCKPQGG (EMP13) and HFGPLTWV (EMP26) were examined under identical ACE conditions as EMP1 and EMP37. The EMP13 is structurally different from EMP1 by only a single amino acid change (i.e. an A substituted for a W) and EMP26 represents the peptide residues between the disulfide bond of EMP1. Fig. 2 presents a set of experimental elec-

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

22

Table 1 Comparison of ACE results with competitive binding IC 50 data Sequence No.a

Sequence

ACE b

IC 50 (mM)c

EMP1 EMP37 EMP13 EMP26

GGTYSCHFGPLTWVCKPQGG LGRKYSCHFGPLTWVCQPAKKD GGTYSCHFGPLTAVCKPQGG HFGPLTWV FMRF

Yes Yes No No No

5 5 . 500 . 500 ND d

a

Sequence numbers same as those in Ref. [18]. Yes denotes that the electrophoretic mobility of EBP significantly changed as the CE buffer concentration of the corresponding peptide was increased from 0.5 to 50 mM. No implies that no change in the electrophoretic mobility of EBP was observed over a same concentration range. c The IC 50 is defined as the concentration of peptide which reduced the binding of [ 125 I]EPO to the EBP which was covalently attached to Sulfolink agarose beads by 50% [18]. d Not determined. b

Fig. 2. Affinity capillary electrophoresis of a mixture of isoforms of extracellular ligand binding domain of erythropoietin receptor (denoted EBP) as a function of HFGPLTWV (EMP26) concentration. The electrophoresis buffer used was 20 mM citrate solution at pH 7.2. The total length of the capillary was 45 cm (37 cm from injection to detection) being operated at 18.5 kV. The inverted peak is due to citrate anions not being present in the injection plug but present in the buffer and therefore, is used as the standard reference marker. It is also noted that the electrophoretic mobility of the inverted peak did not change significantly with increases in the concentration of EMP26. This observation is consistent with the EOF being constant during these ACE experiments and that minimal adsorption of EMP26 onto the capillary wall occurred.

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

23

tropherograms showing that no significant peak broadening, and no abrupt change in DmP,L , is observed for EMP26. Similar results were obtained for EMP13 (data not shown). The peptides EMP13 and EMP26 both bind very weakly to EBP (IC 50 . 500 mM; Table 1) [18]. As an additional control, the peptide FMRF was tested in the ACE assay (data not shown). With FMRF as the ligand, no significant peak broadening, and no significant change in DmP,L , was observed up to a final FMRF concentration of 200 mM. It was surprising to observe this abrupt change in the electrophoretic mobility of EBP in the presence of EMP1 and EMP37. It is more typical to observe a smooth increase or decrease in the electrophoretic mobility of a protein as the ligand concentration in the running buffer is increased [1–3]. If the binding of a charged ligand to a protein (i.e. stoichiometry of 1:1) does not significantly change the electrophoretic mobility of the protein–ligand complex by either changing the charge or the size of the complex or both, the electrophoretic mobilities of the free and complexed proteins will be experimentally indistinguishable. We propose, without definitive evidence, that the abrupt change in the electrophoretic mobility of EBP may suggest that the peptide is either aggregating at the higher concentrations first and then interacting with EBP or that EBP is consecutively binding to the peptides – that is, the stoichiometry of the EBP ? peptide complex is greater than 1:1 (i.e. 1:2 or 1:3 etc.). While the exact mechanism is not known, it is clear, from our data, that peptides that bind to EBP can be distinguished from peptides that do not. An assumption made in ACE experiments is that the protein is only interacting with the peptide in the running buffer and not with peptide adsorbed to the capillary wall [1–3]. The neutral coated capillaries used in these experiments are designed to reduce the adsorption of molecules onto the capillary wall during ACE analysis since adhesion of peptides may potentially interfere with the interpretation of the experiment. The electroosmotic flow (EOF) of the coated capillary is estimated at less than 5% of the EOF determined for a similar uncoated capillary [23]. Therefore, the coated capillary used in these experiments retain a small percentage of the silanol groups normally found in uncoated fused silica. Peptides may adhere to the capillary walls at these active silanol sites. Since EBP has bioaffinity for these peptides, peptides adsorbed to the capillary wall will reduce or even eliminate the EBP signal and increase the EBP migration times. The adhesion of peptides to the capillary wall was examined for EMP1 and EMP37 as outlined in the Section 2. When the capillary was exposed to low peptide concentrations (e.g. 0.5–50 mM) and rinsed with citrate solution, the peak shape and electrophoretic mobility of EBP was not significantly affected – thus, suggesting a minimum amount of EMP1 or EMP37 was adsorbed to the capillary wall. When the capillary was exposed to higher peptide concentration (ca. 100 mM of EMP1 or EMP37) and rinsed, the elution of EBP was not observed in the electropherogram. The baseline at these high peptide concentrations was highly distorted and may have obscured the EBP signal or the EBP may have been completely bound-up in the capillary. The capillary could be recovered in each case by a 0.1 N HCl rinse followed by a citrate solution rinse. It is also noted in Figs. 1 and 2 that the electrophoretic mobility of the inverted peak in the electropherograms, which is the citrate anion, did not change significantly with increases in the concentration of EMP1 or EMP26. We are using this peak as our

24

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

standard reference marker. This observation is consistent with the EOF being constant during these experiments. As a final check on our data, we repeated the experiments utilizing a 20 mM phosphate buffer at pH 7.2 with O-phospho-DL-tyrosine as the standard reference marker. These results were totally consistent with the results reported here.

4. Conclusion In summary, we have developed a method using ACE to screen peptides for their interaction with EBP. The ACE protocol takes less than 20 min per sample and is easily automated. The results were consistent with previously reported IC 50 data obtained from radioactive competition assays [18]. That is, peptides with an IC 50 in the low micromolar range (i.e. 5 mM or less) produced changes in the electrophoretic mobility of EBP while peptides with IC 50 . 500 mM did not show this effect (Table 1). The observations of an abrupt change in the electrophoretic mobility of EBP over a small change in the peptide concentration for EMP1 and EMP37 suggests that more than one peptide is binding to EBP. This model system also serves to demonstrate the feasibility of ACE being used in applications to screen peptides or drugs that bind and may induce homo-dimerization of other proteins.

Acknowledgements The authors thank Mr William Jones for performing the mass spectrometry studies of EBP and helpful discussions with Dr William Murray. We also thank Mr Kenway Hoey for preparing peptides for this project.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

Colton IJ, Carbeck JD, Rao J, Whitesides GM. Electrophoresis 1998;19:367–82. Chu Y-H, Cheng CC. Cell Mol Life Sci 1998;54:663–83. Gomez FA, Avila LZ, Chu Y-H, Whitesides GM. Anal Chem 1994;66:1785–91. Heegard NHH, Hansen BE, Svejgaard A, Fugger LH. J Chromatogr 1997;781:91–7. Gomez FA, Mirkovich JN, Dominguez VM, Liu KW, Macias DM. J Chromatogr 1996;727:291–300. Kwak E-S, Gomez FA. Chromatographia 1996;43:659–62. Shimura K, Kasai K. Anal Biochem 1995;227:186–94. Kuhn R, Frei R, Christen M. Anal Biochem 1994;218:131–5. Liu J, Volk KJ, Lee MS, Pucci M, Handwerger S. Anal Chem 1994;66:2412–6. Handwerger S, Pucci M, Volk KJ, Liu J, Lee MS. J Bacteriol 1994;176:260–4. Chu Y-H, Avila LZ, Gao J, Whitesides GM. Acc Chem Res 1995;28:461–8. Krantz SB. Blood 1991;77:419–34. Nicola NA, editor. Guidebook to cytokines and their receptors, New York: Oxford University Press, 1994. [14] Philo JS, Aoki KH, Arakawa T, Narhi LO, Wen J. Biochemistry 1996;35:1681–91.

G.W. Caldwell et al. / J. Biochem. Biophys. Methods 40 (1999) 17 – 25

25

[15] Johnson DL, Middleton SA, McMahon F, Barbone FP, Kroon D, Tsao E, Lee WH, Mulcahy LS, Jolliffe LK. Protein Express Purif 1996;7:104–13. [16] Wrighton NC, Farrell FX, Chang R, Kashyap AK, Barbone FP, Mulcahy LS, Johnson DL, Barrett RW, Jolliffe LK, Dower WJ. Science 1996;273:458–63. [17] Livnah O, Stura EA, Johnson DL, Middleton SA, Mulcahy LS, Wrighton NC, Dower WJ, Jolliffe LK, Wilson IA. Science 1996;273:464–71. [18] Johnson DL, Farrell FX, Barbone FP, McMahon FJ, Tullai J, Hoey K, Livnah O, Wrighton NC, Middleton SA, Loughney DA, Stura EA, Dower WJ, Mulcahy LS, Wilson IA, Jolliffe LK. Biochemistry 1998;37:3699–710. [19] Atherton E, Sheppard RC, editors. Solid phase peptide synthesis: a practical approach, Oxford, UK: IRL Press, 1989, p. 203. [20] Mant CT, Hodges RS, editors. High performance liquid chromatography of peptides and proteins, Boca Raton, FL: CRC Press, 1991. [21] Okun VM. Electrophoresis 1998;19:427–32. [22] White DH, Shultz J, Hurst R. BioPharm 1995;8(8):40–7. [23] Beckman Instruments, eCAP TM Neutral Capillary Methods Development Kit / Proteins, Fullerton, CA, 1994.