Expression, crystallization and derivatization of the complete extracellular domain of the ������c subunit of the human IL5, IL3 and GM-CSF receptors: Crystals of the ������c

Eur. J. Biochem. 268, 2905±2911 (2001) q FEBS 2001

Expression, crystallization and derivatization of the completeextracellular domain of the bc subunit of the human IL-5, IL-3and GM-CSF receptors

Sonja E. Gustin1, Alice P. Church1, Sally C. Ford1, David A. Mann1, Paul D. Carr2, David L. Ollis2 andIan G. Young1

1Division of Biochemistry and Molecular Biology, John Curtin School of Medical Research and 2Research School of Chemistry,

Australian National University, Canberra, Australia

The major signalling entity of the receptors for the

haemopoietic cytokines granulocyte-macrophage colony

stimulating factor (GM-CSF), interleukin-3 (IL-3) and

interleukin-5 (IL-5) is the shared bc receptor, which is

activated by ligand-specific a receptors. The bc subunit is a

stable homodimer whose extracellular region consists of

four fibronectin domains and appears to be a duplication of

the cytokine receptor homology module. No four domain

structure has been determined for this receptor family and

the structure of the bc subunit remains unknown. We have

expressed the extracellular domain in insect cells using the

baculovirus system, purified it to homogeneity and

determined its N-terminal sequence. N-glycosylation at

two sites was demonstrated. Crystals of the complete

domain have been obtained that are suitable for X-ray

crystallographic studies, following mutagenesis to remove

one of the N-glycosylation sites. The rhombohedral crystals

of space group R3, with unit cell dimensions 186.1 AÊ and

103.5 AÊ , diffracted to a resolution of 2.9 AÊ using syn-

chrotron radiation. Mutagenesis was also used to engineer

cysteine substitution mutants which formed isomorphous

Hg derivatives in order to solve the crystallographic phase

problem. The crystal structure will help to elucidate how

the bc receptor is activated by heterodimerization with the

respective a/ligand complexes.

Keywords: cytokine receptor; haematopoiesis; X-ray crys-

tallography; eosinophils; asthma.

Granulocyte-macrophage colony stimulating factor(GM-CSF), interleukin-3 (IL-3) and interleukin-5 (IL-5)are related cytokines involved in the regulation of haemato-poiesis and inflammation. The receptor systems for thesehaemopoietic cytokines consist of ligand-specific a sub-units and a shared common beta subunit (bc) [1,2]. The bc

subunit is believed to be the main signalling component. Itis activated by heterodimerization with the a receptors forIL-5, IL-3 or GM-CSF. The bc subunit belongs to the class Ifamily of cytokine receptors that are characterized by a200-amino-acid extracellular cytokine receptor homologymodule of two fibronectin type III domains that hasconserved sequence features [3,4]. The family includes thegrowth hormone and prolactin receptors, the erythropoietinreceptor and the IL-4 receptor a for which ligand/receptorcrystal structures are now available [5±9]. The growthhormone receptor is the paradigm for the two domainreceptors of this class and the prolactin, erythropoietin andIL-4 a receptors are of similar structure.

From sequence comparisons it is clear that the extra-cellular domains of some of the members of the class Icytokine receptor family are more complex than that of the

growth hormone receptor, although they generally includethe characteristic cytokine receptor module of two fibro-nectin type III domains as part of their structure [10]. Thestructure of this region of the gp130 receptor has beendetermined recently [11] and also resembles that of thegrowth hormone receptor. The bc subunit has some featuresthat make it quite different from the growth hormonereceptor. Its extracellular region has four fibronectin typeIII domains and it appears from sequence comparisons to bea duplication of the cytokine receptor homology module.Although structural modelling of the bc subunit has beencarried out based on the growth hormone receptor, there isno closely analogous structure available as no four-domainstructure for this family has been determined. An additionalcomplication for modelling is that the bc extracellulardomain is a stable dimer (S. E. Gustin, A. P. Church,S. C. Ford, D. A. Mann, P. D. Jeffrey, I. Walker, U. M.Wiedemann & I. G. Young, unpublished results). Veryrecently, the X-ray structure of an antibody complex ofdomain 4 of the bc subunit has been reported [12] but majorquestions about the overall structure and the mechanism ofdimerization remain unanswered.

The function of the bc receptor has clinical relevance.GM-CSF and IL-3 are broadly acting growth regulatorswith activities on a number of lineages of myeloid cells.IL-5 is a key regulator of the eosinophil lineage and playsan obligatory role in the generation of the eosinophiliatypically seen in helminth infections and allergic diseasessuch as asthma [13,14]. IL-5 and eosinophils appear to becentral to the inflammatory mechanisms underlying asthma[14,15] and GM-CSF has been used clinically to treat

Correspondence to I. G. Young, Division of Biochemistry and

Molecular Biology, John Curtin School of Medical Research, Canberra

ACT 0200, Australia. Fax: 1 61 26125 0415, Tel.: 1 61 26125 2439,

E-mail: [email protected]

Abbreviations: GM-CSF, granulocyte-macrophage colony stimulating

factor; IL, interleukin; HPMC, hydroxypropylmethyl cellulose.

(Received 31 January 2001, accepted 20 March 2001)

neutropenia following cancer chemotherapy [16]. To gainan improved understanding of the structure of the bc

receptor and its role in signal transduction, we haveundertaken structural studies of the complete bc subunitextracellular domain. In the present work, we describe theexpression and crystallization of the human bc subunit.Mutation of an N-glycosylation site was required toimprove the crystals so as to achieve sufficient resolutionfor structure determination. Site-directed mutagenesiswas also used to introduce cysteine residues at variouspositions in the protein chain in order to generate a seriesof isomorphous heavy-atom derivatives and solve thecrystallographic phase problem.

M A T E R I A L S A N D M E T H O D S

Expression of human bc extracellular domain

A cDNA encoding the extracellular region of the human bc

receptor was isolated by PCR using Pfu DNA polymerase(Stratagene) from cDNA prepared by reverse transcriptionof HL60 eosinophils total RNA. The primers used werebased on the reported sequence of the human bc subunit[17]. The cDNA was subcloned into the EcoRI site ofpBacPAK8 (Clontech) for baculovirus expression. To createa recombinant baculovirus, the recombinant vector wastransfected into Sf9 cells together with pBacPAK6 DNA(Bsu36I digest) according to the manufacturer's instruc-tions. The virus was then propagated in Sf9 cells and usedto infect High Five cells (Invitrogen) for the production ofexpressed protein. Cells (107) were seeded into 225-cm2

flasks and grown for 72 h in Sf-900II SFM medium(Gibco). Twelve millilitres of virus in serum-containingHink's TNM-FH medium (ICN) was added and the cellsincubated with virus for 1 h, after which time the virus wasdrained off and the cells washed three times to removetraces of serum. Forty millilitres of serum-free Sf-900IISFM medium was added and incubation continued for afurther 72 h. The supernatant containing the secretedrecombinant protein was then adjusted to pH 6.5 with1 m Na2HPO4 and centrifuged at 14 600 g for 15 min.

Purification of recombinant protein

The baculovirus supernatant (650 mL) containing the bc

subunit was concentrated to 10 mL using ultrafiltrationstirred cells (Amicon 8400 with YM10 membranes),centrifuged at 14 600 g for 15 min, and filtered through a0.22-mm filter (Millipore). The concentrate was appliedto a Sephacryl S-200 (Pharmacia) gel filtration column(45 � 2 cm) equilibrated in 20 mm bis-Tris propane, 0.02%NaN3 (pH 6.5). Fractions were analysed by SDS/PAGE andthe fractions of interest pooled and concentrated byultrafiltration (Amicon 8050) to 5 mL. The bc proteinwas then chromatographed on a MonoQ HR 5/5 column(Pharmacia) at pH 6.5 using an FPLC apparatus and elutedusing a 0±350 mm NaCl gradient. Protein concentrationswere determined using a Bio-Radw protein assay kit withbovine gamma globulin as the protein standard.

Gel electrophoresis

Nonreducing SDS/PAGE was carried out using Novex 10%or Gradipore 12% polyacrylamide gels. Samples were

derivatized by boiling 5 min in SDS sample buffer and20 mL loaded and electrophoresed at 125 V for 120±150 min. The gels were stained either using Coomassieblue R-250 or silver. Molecular masses were estimatedusing Mark 12TM standards (Novex). Nondenaturing gelelectrophoresis was carried out in Tris/glycine bufferpH 8.3 using the same gels but without SDS in the sampleor electrophoresis buffers.

Capillary isoelectric focusing

Capillary IEF was carried out using a Beckman P/ACEsystem 2100. Briefly, the samples were electrophoresed at300 V´cm21 using reversed polarity (cathode at the inlet)and using catholyte (20 mm NaOH, 0.5% hydroxypropyl-methyl cellulose, HPMC) at the inlet and anolyte (40 mmphosphoric acid, 0.5% HPMC) at the outlet and a neutral-coated capillary (Beckman P/N 477445) cut to 27 cm. Thecapillary was rinsed for 10 min with 0.5% HPMC and thenfilled with ampholyte mixture (0.5% HPMC, 7.08 mL; 40%Pharmalyte pH 3-10, 0.8 mL; TEMED, N,N,N 0,N 0-tetra-methylethylenediamine, 0.12 mL) containing the sample(5 mg protein). Detection of focused proteins was bymonitoring absorbance at 280 nm.

N-Terminal sequencing

N-Terminal sequencing of proteins was performed on a PEApplied Biosystems 494 Protein Sequencer. Proteins to besequenced were absorbed overnight onto a small piece(3 � 3 mm) of PE Applied Biosystems ProBlott/poly(vinylidene difluoride) and washed free of buffer ions.The dried membrane was subjected to N-terminal Edmansequencing using the standard pulsed-liquid protocol forblotted samples.

Testing for the occurrence of the bc splice variant

Flanking primers were used to amplify cDNA covering thesite of insertion in the bc splice variant from HL60. Therecovered PCR product from HL60 cDNA was thendigested with Bsu36I. A unique site for this enzyme ispresent in the 18-bp insert of the bc splice variant.

Test for bc function in CTLL-2 cells

The cDNA encoding the full length human IL-5 a receptorwas subcloned into the EcoRI site of the expression vectorpcDEF3 [18]. The resulting expression construct wastransfected into CTLL-2 cells by electroporation (300 V,500 mF) in growth medium (RPMI with 10% fetalbovine serum, 50 mm 2-mercaptoethanol, mouse IL-2 at10 U´mL21). The transfected cells were allowed to recoverfor 24 h in growth medium and then selected with G418(400 mg´mL21) for 18 days. The G418-resistant cells weredesignated CTLL-2(hIL-5a). These cells were then trans-fected with the full length bc (HL60) cDNA which had beensubcloned into the XbaI site of the bicistronic expressionvector pEFIRES-P [19]. The transfected cells were selectedwith puromycin (2 mg´mL21) for 12 days. The puromycin-resistant cells were designated CTLL-2 (hIL-5a, hbc HL60)and tested for their ability to grow in the presence of IL-5.An analogous transfectant was also prepared using the hbc

2906 S. E. Gustin et al. (Eur. J. Biochem. 268) q FEBS 2001

(TF1) cDNA and used as a control. It was designatedCTLL-2 (hIL-5a, hbcTF1).

Thymidine incorporation assays were used to test forthe ability of CTLL-2 cells expressing the hbc(HL60)receptor to grow in response to IL-5. Cells to be used inthe test for IL-5 responsiveness were grown to about7.5 � 105 cells´mL21 in growth medium containing IL-2.They were then washed four times in factor-free basalmedium and 2 � 104 cells per well distributed in basalmedium in 96-well plates. After incubation for 44 h,0.5 mCi of tritiated thymidine was added and the incubationcontinued for a further 4 h. Cells were then harvested ontoglass fibre filters, washed three times, dried, and coveredwith 100 mL of scintillation fluid (LKB Optiscint HiSafe)and counted using a Packard Tri-Carb 1900CA.

Crystallization

Prior to crystallization, the purified protein was concen-trated to 4±7 mg´mL21 using Ultrafree-MC Filter Units(Millipore). Crystals of human bc were grown by thehanging-drop vapor diffusion method using Linbro multi-well tissue culture trays. Successful crystallization wasachieved using well buffers (1 mL) consisting of 50 mmphosphate buffer containing a range of concentrations(7±10%) of poly(ethylene glycol) 5000 monomethyl etherand covering the pH range 6.2±6.6. The drops contained4 mL of protein (4±7 mg´mL21) in 20 mm bis-Tris propane(pH 6.5), 0.1 mm EDTA and 4 mL of precipitant solutionand were equilibrated at 4 8C for about 7 days.

Cryobuffer

For freezing of crystals without loss of resolution, a bufferof the following composition was found to be satisfactory:9% (w/v) poly(ethylene glycol) 5000 monomethyl ether, 20%v/v 2-methyl-2,4-pentanediol, 50 mm NaH2PO4 /K2HPO4

buffer, pH 6.5. The crystals were initially transferred to amixture of 75% well buffer: 25% cryobuffer and allowedto soak 15 min at 4 8C. The proportion of cryobuffer insuccessive soaks was increased in 25% increments until100% cryobuffer was reached. This procedure avoidedcracking of the crystals.

Soaks in solutions of heavy-atom compounds

The general condition used for soaking the crystals was anovernight soak at 4 8C in a 10% saturated heavy-atomsolution in cryobuffer. To avoid cracking of the crystals, thelevel of heavy-atom solution was raised from 1 to 10% infour steps with a 15-min soak at each intermediate con-centration. For the cysteine substitution mutants, an over-night soak was carried out in cryobuffer containing 5 mmdithiothreitol. Excess cryobuffer was then removed bysyringe and the crystals rinsed with cryobuffer containingincreasing levels of mercurial (methylmercurychloride ordimercuryacetate) until 10% saturation was reached. Theminimum soak was overnight but no deterioration of thecrystals occurred up to 14 days at 4 8C.

X-ray analysis

X-ray diffraction data were collected at the FIP beamline ofthe European Synchrotron Radiation Facility. The heavy

atom positions for six heavy atom derivatives preparedfrom cysteine substitution mutants were readily locatedusing Pattersons and Difference Fouriers. As there were twoprotein chains in the asymmetric unit, the resulting heavyatom positions enabled the identification not only of theposition of known residues, but also the vector betweenequivalent dimer-related positions. Data were integratedand scaled using the programs denzo and scalepack [20].Data manipulations and map calculations were carried outusing the ccp4 suite of programs [21].

Site-directed mutagenesis

Mutations were made using the QuikChangeTM site-directed mutagenesis kit (Stratagene) according to themanufacturer's instructions. The fidelity of the changes wasconfirmed by DNA sequencing.

R E S U LT S A N D D I S C U S S I O N

Expression and characterization of the human bc

receptor extracellular domain

We isolated a cDNA encoding the bc receptor from HL60eosinophils. Total RNA was used to prepare cDNA and acDNA encoding the extracellular domain of the bc subunitproduced by RT-PCR using primers based on the knownDNA sequence [17]. The DNA sequence of the PCR clonewas determined after subcloning into pBacPAK8. Thesequence agreed with the published sequence except for an18-bp insertion, in frame, which appears to have arisenfrom alternative splicing, resulting in the addition of thesequence CTCTGCTGTGCTCCTCAG to the 5 0 end ofexon 8 (Fig. 1A). This results in an insertion of six aminoacids compared to the originally described bc subunit fromTF1 cells (Fig. 1B). The 18-bp insertion carries a site forthe restriction enzyme Bsu36 I allowing the proportion ofsplice variant in the total population to be determined.Analysis of the total PCR product from HL60 obtainedusing primers flanking the site of the 18 bp insertion,showed that the splice variant which had been isolatedrepresented the predominant form of the mRNA in HL60eosinophils. The cDNA was designated hbc(HL60) andencoded residues 1±443 of the receptor. A similar analysisof receptor isoforms from TF1 cells showed only low levelsof the bc splice variant in these cells (data not shown).

To investigate whether the bc variant could elicit anormal growth signal, in response to stimulation with IL-5,a full length version of the HL60 bc variant was constructedand tested for its ability to stimulate the growth of CTLL-2cells. These cells do not normally express either the IL-5 aor bc receptors. In CTLL-2 cells transfected to express theIL-5 a receptor, the HL60 bc subunit was shown to becapable of signalling a normal growth response giving half-maximal stimulation at similar levels of IL-5 (mean of threeexperiments) to the bc subunit from TF1 cells. Based on thestructural alignment of mouse b subunits carried out byBazan [3], it is possible that the six-amino-acid insertion inthe bc variant is an addition to the B-C loop of domain 3. Atthis stage, there is no evidence that it has any particularfunctional significance.

The cDNA encoding the hbc(HL60) subunit extracellulardomain was subcloned into pBacPAK8 and expressed in

q FEBS 2001 Crystals of the bc receptor extracellular domain (Eur. J. Biochem. 268) 2907

insect cells using the baculovirus system. The expressedprotein was secreted into the medium of the insect cellcultures under the direction of its signal sequence. SDS/PAGE analysis of supernatants from cultures infected witha recombinant baculovirus expressing the bc receptorshowed the presence of glycosylated recombinant proteinsof apparent size of 54 kDa. Purification of this protein bygel filtration followed by ion-exchange chromatography onMono Q produced a preparation that was essentially purewhen examined by SDS/PAGE (Fig. 2A). The yield wasabout 9 mg per 650 mL of culture.

To verify the purity of the recombinant protein it wasexamined by capillary IEF. The protein showed the pre-sence of one major peak (pI 5.6) confirming its homo-geneity (Fig. 2B). Although SDS/PAGE analysis suggestedsome variation in the degree of glycosylation of therecombinant protein, the glycosyl chains added by theinsect cells are not usually charged and therefore nocharge heterogeneity resulting from the variable degree ofglycosylation is expected.

N-Terminal sequencing of the purified bc(HL60) proteinconfirmed its identity and revealed two sites of signalpeptide cleavage. The major sequence (about 90%) wasEETIPLQTLR with a minor sequence (about 10%) ofGAEETIPLQ. The minor site of signal peptide cleavagewas at a position two amino acids earlier in the predictedprotein sequence than the major site. Thus the major formof the mature bc (HL60) protein would consist of 419amino acids (Fig. 1B).

Analytical ultracentrifugation showed that the purified bc

subunit is a homodimer. This work will be described indetail elsewhere (S. E. Gustin, A. P. Church, S. C. Ford,D. A. Mann, P. D. Jeffrey, I. Walker, U. M. Wiedemann &I. G. Young, unpublished results).

Crystallization of the bc extracellular domain

Attempts were made to crystallize the bc receptor extra-cellular domain using commercial crystal screens (Hamp-ton Research) in sitting-drop vapour diffusion experiments.Optimization of the conditions giving the most promisingcrystals resulted in the crystallization conditions describedin Materials and methods [7±10% poly(ethylene glycol)5000 monomethyl ether, pH 6.2±6.6]. The crystals obtained

Fig. 2. Characterization of purified bc (HL60) extracellular

domain. (A) SDS/PAGE analysis, 12% gel stained with Coomassie

blue R-250. Lane 1, hbc(HL60); lane 2, hbc(TF1). (B) Capillary IEF of

hbc(HL60).

Fig. 1. Sequence of the human bc gene. (A) Sequence of relevant

region of human bc gene showing the additional 18 bases (underlined)

which are added to the 5 0 end of exon 8 in the splice variant of bc from

HL60 eosinophils Sequence is taken from GenBank 3136000. (B)

Mature sequence of the extracellular domain of the bc(HL60). The

N-terminal sequence obtained by sequencing the expressed protein is

underlined. Also indicated is the six amino acid insertion in this

variant, the WSXWS motif and the 3 potential N-glycosylation sites.

2908 S. E. Gustin et al. (Eur. J. Biochem. 268) q FEBS 2001

under these conditions were mainly plates with occasionalneedles and could not be further improved by attemptedoptimization of crystallization conditions. They diffractedto about 5 AÊ using a rotating anode source, which was notsufficient for structure determination.

Because one possible problem affecting crystal growthmay have been interference from N-glycosylation, wesystematically modified potential sites of N-glycosylationby site-directed mutagenesis. The human bc extracellularregion carries three potential sites of N-glycosylation (N34,N167, N328) that are conserved in the mouse and human bc

subunits. These three sites were individually mutated toglutamines and the mutant proteins expressed in the baculo-virus system. Analysis by SDS/PAGE showed a smallreduction in size of the mutant proteins N34Q (not shown)and N328Q (Fig. 3), numbered according to the majorspecies of mature hbc(HL60) (Fig. 1B), indicating thatthese sites are glycosylated in the parent protein. Themutation N34Q also resulted in a great reduction in yieldfrom the baculovirus system, suggesting that this glyco-sylation may serve as a signal for secretion. The mutantprotein N167Q showed no detectable change in size(Fig. 3), suggesting that this site is either not glycosylatedor carries a glycosyl chain that is too short to be detected bythis method.

The yields of N34Q were too low to be practicable butthe yields of the N328Q mutant were similar to the wild-type and therefore large scale batches of the N328Q mutant

were prepared and crystallization trials carried out. It wasfound that the quality of the crystals obtained was signifi-cantly improved by removing the site of glycosylation atN328.

The final crystals obtained were 0.15 � 0.1 mm in size,rhombohedral, of space group R3 and with unit celldimensions of 186.1 AÊ and 103.5 AÊ . A suitable buffer forcryocrystallography was developed (see Materials andmethods). Diffraction to 3.3 AÊ was obtained using arotating anode source but this was extended to 2.9 AÊ

using synchrotron radiation. The volume occupied per unitmolecular mass (Vm) was calculated [22] and found to be3.5, indicating one homodimer per asymmetric unit.

Engineering of isomorphous heavy-atom derivatives.

Molecular replacement using the growth hormone receptorstructure as a model was not successful. The formation ofisomorphous heavy-atom derivatives was therefore requiredfor determination of the detailed structure. In an extensiveset of trials we were unable to generate suitable heavy-atomderivatives by soaking the crystals in a variety of solutionscontaining representative compounds from all the majorclasses of heavy-atom compounds. We therefore soughtto engineer heavy-atom derivatives by site-directedmutagenesis.

Dao-Pin et al. [23] found that useful heavy-atomderivatives could be obtained by introducing cysteineresidues at positions close to the surface of the proteinbut which were not fully exposed to solvent. These authors,and several other groups who have used this approach[24±27], have favoured Ser!Cys mutations because 50%of serine residues have a surface accessibility of at least30% and serine is nearly isostructural with cysteine.However, successful mutation of other residues likely tohave a surface position has also been reported [28,29].

Thirteen cysteine substitution mutants were designedusing the above principles and nine mutants (S54, Q116,S157, S249, K204, R267, N328, S368, and E409) on thebasis of a comparison with the growth hormone receptorstructure [5]. The growth hormone receptor extracellulardomain has two fibronectin domains and the assumptionwas made that the bc receptor structure would resemble aduplication of the growth hormone receptor structure. Theprimary criterion was the predicted accessibility of theintroduced cysteines. Of the 22 mutants, only eight crystal-lized under the conditions established for the wild-typeprotein and gave usable-sized crystals (Table 1). Soakingof the crystals with mercurials (methylmercury chloride

Fig. 3. Site-directed mutagenesis of potential N-glycosylation sites

in hbc(HL60). SDS/PAGE analysis (Coomassie-blue staining): lane 1,

N167Q; lane 2, N328Q; lane 3, wild-type.

Table 1. Properties of cysteine substitution mutants. To obtain Hg derivatives, crystals were soaked in dimercuryacetate as described in Materials

and methods.

Type of crystals Mutations Hg derivative

Wild-type crystals S54C, S157C, S166C, K204C, S368C, T418C 1

P142C, R389C ±

Small crystals R197C, R267C, E348C

Needles/plates Q116C, D120C, S181C, N328C, S387C

No crystals S17C, S78C, R93C, S249C, T401C, E409C

q FEBS 2001 Crystals of the bc receptor extracellular domain (Eur. J. Biochem. 268) 2909

or dimercuryacetate) established that six (S54C, S157C,S166C, K204C, S368C, T418C) of the crystallizablemutants formed isomorphous heavy-atom derivatives. Theother mutants proved unsatisfactory for a variety of reasons:four mutants gave low yields, four others did not crystallizeand the remainder gave very small crystals or needles andplates (Table 1). Of the six reactive cysteine mutants,S166C gave the best derivative (Table 2). Dimercuryacetatewas generally used for derivative formation as it gave astronger signal due to its two Hg atoms. It also provedpossible to crystallize a triple mutant (S54C, S166C,S368C) which gave a further improvement in phasingpower (Table 2). All but T418C were useful for phasing.MIRAS phases were calculated using mlphare. Theoverall figure of merit was 0.682.

Although there is only a limited sample of mutants, ourexperience would suggest that cysteine substitution ofserines is the most successful approach, as four heavy-atomderivatives were produced from a total of nine suchmutants. Our data also suggest that serines predicted tobe in loops and which are conserved throughout a proteinfamily and with adjacent charged residues should beincluded when designing mutants as these gave a highyield of heavy-atom derivatives. In the present work,conditions of purification were used which aimed tominimize the oxidation of cysteine residues. Buffers wereat pH 6.5 as the oxidation of cysteine residues is greatlyreduced at this pH [30]. EDTA (0.1 mm) was included in allbuffers except those used for chromatography on Mono Q(Pharmacia) as an added precaution against oxidation ofcysteines catalyzed by trace heavy metal contamination.The cysteines were subsequently reduced by treatment withdithiothreitol prior to derivative formation. Using theseapproaches, useful isomorphous derivatives were success-fully prepared from 75% of the proteins which crystallizednormally under the conditions established for the parentprotein.

The bc subunit is one of the more complex members ofthe class I cytokine receptor family with an extracellularregion consisting of four fibronectin domains. It clearlyconsists of two cytokine receptor homology modules. This,coupled with its dimerization in the absence of ligand,suggests the structure may be quite different from thestructures of the other members of this receptor family thathave been determined to date. The present work will allowthe structure of the bc subunit and the nature of itsdimerization to be solved. This should greatly assist ourunderstanding of this interesting receptor.

A C K N O W L E D G E M E N T S

We thank D. Shaw for N-terminal protein sequencing, T. Gill for

assistance with the initial purification, S. Palmer for help with the b

splice variant and D. Woltring for assistance with mutagenesis. We are

grateful for the help and friendly advice of Michel Roth whilst

collecting data at the FIP beamline at the ESRF and the assistance of

Denis Verger in data collection. The Australian Synchrotron Radiation

Project is thanked for travel support.

R E F E R E N C E S

1. Tavernier, J., Devos, R., Cornelis, S., Tuypens, T., Van der Heyden,

J., Fiers, W. & Plaetinck, G. (1991) A human high affinity

interleukin-5 receptor (IL5R) is composed of an IL5-specific alpha

chain and a beta chain shared with the receptor for GM-CSF. Cell

66, 1175±1184.

2. Takaki, S., Murata, Y., Kitamura, T., Miyajima, A., Tominaga, A.

& Takatsu, K. (1993) Reconstitution of the functional receptors for

murine and human interleukin 5. J. Exp. Med. 177, 1523±1529.

3. Bazan, J.F. (1990) Structural design and molecular evolution of a

cytokine receptor superfamily. Proc. Natl Acad. Sci. USA 87,

6934±6938.

4. Sato, N. & Miyajima, A. (1994) Multimeric cytokine receptors:

common versus specific functions. Curr. Opin. Cell Biol. 6,

174±179.

5. De Vos, A.M., Ultsch, M. & Kossiakoff, A.A. (1992) Human

growth hormone and extracellular domain of its receptor: crystal

structure of the complex. Science 255, 306±312.

6. Somers, W., Ultsch, M., De Vos, A.M. & Kossiakoff, A.A. (1994)

The X-ray structure of a growth hormone-prolactin receptor

complex. Nature 372, 478±481.

7. Hage, T., Sebald, W. & Reinemer, P. (1999) Crystal structure of the

interleukin-4/receptor alpha chain complex reveals a mosaic

binding interface. Cell 97, 271±281.

8. Livnah, O., Stura, E.A., Johnson, D.L., Middleton, S.A., Mulcahy,

L.S., Wrighton, N.C., Dower, W.J., Jolliffe, L.K. & Wilson, I.A.

(1996) Functional mimicry of a protein hormone by a peptide

agonist: the EPO receptor complex at 2.8 AÊ . Science 273,

464±471.

9. Syed, R.S., Reid, S.W., Li, C., Cheetham, J.C., Aoki, K.H., Liu, B.,

Zhan, H., Osslund, T.D., Chirino, A.J., Zhang, J., Finer-Moore, J.,

Elliott, S., Sitney, K., Katz, B.A., Matthews, D.J., Wendoloski, J.J.,

Egrie, J. & Stroud, R.M. (1998) Efficiency of signalling through

cytokine receptors depends critically on receptor orientation.

Nature 395, 511±516.

10. Mott, H.R. & Campbell, I.D. (1995) Four-helix bundle growth

factors and their receptors: protein±protein interactions. Curr.

Opin. Struct. Biol. 5, 114±121.

11. Bravo, J., Staunton, D., Heath, J.K. & Jones, E.Y. (1998) Crystal

structure of a cytokine-binding region of gp130. EMBO J. 17,

1665±1674.

12. Rossjohn, J., McKinstry, W.J., Woodcock, J.M., McClure, B.J.,

Hercus, T.R., Parker, M.W., Lopez, A. & Bagley, C.J. (2000)

Structure of the activation domain of the GM-CSF/IL-3/IL-5

receptor common beta-chain bound to an antagonist. Blood 95,

2491±2498.

13. Kopf, M., Brombacher, F., Hodgkin, P.D., Ramsay, A.J.,

Milbourne, E.A., Dai, W.J., Ovington, K.S., Kohler, G., Young,

I.G. & Matthaei, K.I. (1996) IL-5-deficient mice have a

developmental defect in CD51 B-1 cells and lack eosinophilia

but have normal antibody and cytotoxic T cell responses. Immunity

4, 15±24.

14. Foster, P.S., Hogan, S.P., Ramsay, A.J., Matthaei, K.I. & Young,

I.G. (1996) Interleukin 5 deficiency abolishes eosinophilia,

airways hyperreactivity, and lung damage in a mouse asthma

model. J. Exp. Med. 183, 195±201.

Table 2. Isomorphous heavy atom derivatives. Phasing power �S|Fh|/SõFPHobs|-|FPHcalcõ, where F is the structure factor. Rcullis �SõFPH| ^ |FPõ-|FHõ/SõFPH| ^ |FPõ for centric reflections.

Mutant

Phasing

power Rcullis

S166C 1�.44 0�.76

S368C 1�.06 0�.83

S157C 0�.94 0�.85

K204C 0�.92 0�.87

S54C 0�.82 0�.88

S54C, S368C, S166C 1�.87 0�.67

2910 S. E. Gustin et al. (Eur. J. Biochem. 268) q FEBS 2001

15. Mould, A.W., Matthaei, K.I., Young, I.G. & Foster, P.S. (1997)

Relationship between interleukin-5 and eotaxin in regulating

blood and tissue eosinophilia in mice. J. Clin. Invest. 99,

1064±1071.

16. Hohaus, S., Martin, H., Wassmann, B., Egerer, G., Haus, U.,

Farber, L., Burger, K.J., Goldschmidt, H., Hoelzer, D. & Haas, R.

(1998) Recombinant human granulocyte and granulocyte-macro-

phage colony-stimulating factor (G-CSF and GM-CSF) adminis-

tered following cytotoxic chemotherapy have a similar ability to

mobilize peripheral blood stem cells. Bone Marrow Transplant.

22, 625±630.

17. Hayashida, K., Itoh, N., Kitamura, T., Schreurs, J., Yonehara, S.,

Yahara, I., Arai, K. & Miyajima, A. (1990) Molecular cloning of a

second subunit of the receptor for human granulocyte-macrophage

colony-stimulating factor (GM-CSF): reconstitution of a high-

affinity GM-CSF receptor. Proc. Natl Acad. Sci. USA 87,

9655±9659.

18. Goldman, L.A., Cutrone, E.C., Kotenko, S.V., Krause, C.D. &

Langer, J.A. (1996) Modifications of vectors pEF-BOS, pcDNA1

and pcDNA3 result in improved convenience and expression.

Biotechniques 21, 1013±1015.

19. Hobbs, S., Jitrapakdee, S. & Wallace, J.C. (1998) Development of

a bicistronic vector driven by the human polypeptide chain

elongation factor 1alpha promoter for creation of stable mammal-

ian cell lines that express very high levels of recombinant proteins.

Biochem. Biophys. Res. Commun. 252, 368±372.

20. Otwinowski, Z. & Minor, W. (1997) Processing of X-ray

diffraction data collected in oscillation mode. Methods Enzymol.

276, 307±326.

21. Collaborative Computational Project, Number, 4. (1994) The

CCP4 Suite: Programs for Protein Crystallography. Acta Crystal-

logr. D50, 760±763.

22. Matthews, B.W. (1968) Solvent content of protein crystals. J. Mol.

Biol. 33, 491±497.

23. Dao-Pin, S., Alber, T., Bell, J.A., Weaver, L.H. & Matthews, B.W.

(1987) Use of site-directed mutagenesis to obtain isomorphous

heavy-atom derivatives for protein crystallography: cysteine-

containing mutants of phage T4 lysozyme. Protein Eng. 1,

115±123.

24. Tucker, A.D., Baty, D., Parker, M.W., Pattus, F., Lazdunski, C. &

Tsernoglou, D. (1989) Crystallographic phases through genetic

engineering: experiences with colicin A. Protein Eng. 2, 399±405.

25. Stock, A.M., Mottonen, J.M., Stock, J.B. & Schutt, C.E. (1989)

Three-dimensional structure of CheY, the response regulator of

bacterial chemotaxis. Nature 337, 745±749.

26. Klein, C., Vogel, W., Bender, H. & Schulz, G.E. (1990) Engin-

eering a heavy atom derivative for the X-ray structure analysis of

cyclodextrin glycosyltransferase. Protein Eng. 4, 65±67.

27. Martinez, C., de Geus, P., Stanssens, P., Lauwereys, M. &

Cambillau, C. (1993) Engineering cysteine mutants to obtain

crystallographic phases with a cutinase from Fusarium solani pisi.

Protein Eng. 6, 157±165.

28. Hatfull, G.F., Sanderson, M.R., Freemont, P.S., Raccuia, P.R.,

Grindley, N.D. & Steitz, T.A. (1989) Preparation of heavy-atom

derivatives using site-directed mutagenesis. Introduction of

cysteine residues into gamma delta resolvase. J. Mol. Biol. 208,

661±667.

29. Nagai, K., Oubridge, C., Jessen, T.H., Li, J. & Evans, P.R. (1990)

Crystal structure of the RNA-binding domain of the U1 small

nuclear ribonucleoprotein A. Nature 348, 515±520.

30. Van der Laan, J.M., Swarte, M.B., Groendijk, H., Hol, W.G. &

Drenth, J. (1989) The influence of purification and protein

heterogeneity on the crystallization of p-hydroxybenzoate

hydroxylase. Eur. J. Biochem. 179, 715±724.

q FEBS 2001 Crystals of the bc receptor extracellular domain (Eur. J. Biochem. 268) 2911