Clinical and Experimetual Allergy, 1992, Volume 22, pages 793-803

The allergens of dog II. Identification and partial purification of a major dander allergen A. W. FORD and D. M. KEMENY National Institute for Biological Standards and Control. Potters Bar and Department of Medicine, United Dental and Medical Schools of Guy's and St Thomas' Ho.spitals. St Thomas Street, London Summary

A dog hair and dander (DHD) extract was prepared from hair obtained from mixed breeds. By SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and immunobiotting, using sera from 32 dog-allergic subjects, a number of IgE radio-staining bands could be seen. In 78% of sera a protein of molecular weight (MW) of 21 OOOdaltons. designated Ag X, was found to bind IgE and in 34"',, it did so strongly. This allergen was isolated from DHD by size-exclusion and ion exchange chromatography. The final product was a single allergen of MW of 21 000 and an isoelectric point o\' approximately 5 2. An additional protein-staining band could still be seen of MW of 24000 daltons. Using a serum which contained IgE antibodies only to Ag X, this allergen was found only in DHD extract and dog saliva and was absent from dog serum and urine. It was the same dog allergen that we [1] reported as Ag 8 using crossed radio-immunoclcctrophoresis (CRIE) and that Blands et ai [2] and Lowenstein [3] described as Ag 13. We propose that this major dog allergen be given the title Can f I according to the new allergen nomenclature. Clinical and Experimental Allergy, Vol. 22. pp. 793-803. Submitted 30 August 1991; revised 3 February 1992; accepted 26 February 1992. Introduction Evidence has been accumulating over the past decade that exposure to dogs is a contributory factor in the development of atopic respiratory disease [4] and that a high proportion of asthmatics may have allergy to dogs as evidenced by positive skin prick tests [5]. The recent use of immunotherapy in allergies to cats and dogs [3.6-10] has necessitated the identification and isolation of the major dog allergen in addition to the major cat allergen. Once purified it could be used to monitor the potency of vaccines for immunotherapy and skin testing solutions, and be used in conjunction with appropriate antisera, to monitor the level of dog dander in the environment. Although serum albumin has been repeatedly shown to be an important allergen in hypersensitivity to a number of mammals including the dog [11 14]. many studies have pointed to a dan der-specifieallergen being of equal if not greater importance [1-2,15 16]. Our previous paper on Correspondence: Dr A. W. Ford. Standards Processing Division, National tnstilutc for Biological Standards and Control, Blanche Lane, South Minims, Potters Bar, Ilerltbrdshire EN6 3QG.

the characterization of dog hair and dander extract showed that a mixed breed hair/dander extract, when tested by CRIE with the sera of 60 subjects, had a minimum of28 antigens of which 21 acted as allergens [lj. Onedanderallergen. not found in dog serum, appeared to be the most important, binding IgE in 63% of sera and strongly in 32'Xi of them. This allergen was designated Ag 8 and it was identified as the same allergen as Ag 13 of Blands et al. [2]. This paper describes the further characterization and identification of this major hair and dander-specific allergen (designated Ag X) by immunoblotting and its partial purification by gel filtration and anion exchange chromatography. Materials and methods Dog hair I dander (DHD) extract The production of a hair/dander extract from the unwashed hair of mixed breeds of dogs has been described previously [1,17]. Briefly, allergen was extracted from batches of hair in 0.125 M ammonium bicarbonate solution at -l-4 C at a ratio of 5 g: 100 ml for 22 hr with 793

794

A. W. Ford and D. M. Kemenv

occasional agitation. The extract was obtained by sieving, centrifuging and filtering for sterility. It was dialysed against 0 005 M ammonium bicarbonate solution followed by distilled water, concentrated by ultrafiltration. refiltered for sterility and freeze-dried. Four batches of 500 g of dog hair were processed in this manner resulting in a total of approximately 5.9 g dry weight of freeze-dried extract containing 20% protein by amino acid analysis. Rabbit antisera A 45% saturated ammonium sulphate-precipitated fraction of antisera to DHD was produced as described previously [1], The antisera was raised in rabbits according to the immunization schedule of Harboe and Ingild [18] using three rabbits and a course of injections and bleeding over 6-12 months. A 45% ammonium sulphate precipitation was performed on the pooled sera to prepare an antibody-rich fraction of hyperimmune rabbit serum. Rabbit anti-dog serum and anti-dog IgG {H-|- L chain specific) were purchased from ICN Immuno Biologicals (High Wycombe. U.K.) and rabbit anti-dog serum albumin from Nordic Immunology (Tilburg, The Netherlands). Human sera and serum pools The 32 individual human sera, which were used to recognize the allergens, were obtained as follows: three (nos 1.2 & 12) were sera previously collected and stored at NIBSC from individuals who were allergic to dogs: the sera were positive by the radioallergosorbent test (RAST) and crossed immunoelectrophoresis (CRIE); 11 (nos 3, 4. 6-11 & 30-32) were collected from London teaching hospitals (Clinical Research Centre, St Bartholomew's & Guy's Hospital) from patients with clinical manifestations of dog allergy, positive skin prick tests (SPTs) and RASTs; one serum (no. 5) was from a patient attending the Allergy Clinic in Helsinki Hospital, Finland and 17 (nos 13-29) were kindly provided by Dr John Ohman, Boston, Massachusetts., U.S.A. aspart of a large series of sera from dog-allergic subjects, with positive skin tests and RASTs. Dog specific IgE of the sera was either determined in the clinics donating the sampling by the method of RAST® (Pharmacia Diagnostics AB, Uppsala, Sweden) following the manufacturer's instructions or was measured using discs coupled with DHD extract in modified systems of RAST or "microRAST" as described previously [1,19] where RAST class equivalents were estimated using a high titre anti-dog reference serum pool (NIBSC 79/530) in conjunction with the Pharmacia reference kits. Details of the sera are given in Table 1.

Table 1. Human sera used in immunoblotting Dog-specific No.

Source*

1 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20

N N

21 22 23 24 25 26 11 28 29 30 31 32

L L F L L L L L L N U U

u u u u u u u u u

u u u

u u u L L L

RAST SCO ret 4

4 4 4 5 3 4 5 4 4 5 2 4 2 2 4 5 2 1 \ 2 4 4 2 2 3 2 1 2 5 4 5

Other allergiesj CG C G D

CDGH CGN H CD DGH

CDGH CLA C

C C C

c c c c c c CGD

* Source: N = NIBSC, L = London teaching hospitals. F-Finland, U = U.S.A. as described in Materials and Methods. + Dog-specific RAST score equivalent determined by comparison of the binding of dilutions of positive serum pool 79/530 wilh those of Pharmacia RAST standards as defined in Materials and Methods. J Other allergies: C = Cat, G^gras.s pollen. D ^ Dermatophagoides pieronyssinus mite, H = horse, N = nuts, LA = Laboratory animals. Serum pool 1 was made from equal volumes of 10 sera and serum pool 2 from variable volumes of 13 sera from patients attending Guy's Hospital. All subjects gave strongly positive SPTs to dog skin testing solution (Bencard) and had RAST class equivalents of 2-5 to dog hair/dander extract. Included in serum pool I was the individual serum no. 30; the rest were not used indivi-

Dog dander allergen purification

dually in this particular study. In serum pool 2 the individual sera nos 7-11, 30 and 31 were included; five of the sera were common to both pools. The absorption of serum pool 1 by dog serum albumin (DSA) prior to RASTs to assess the non-albumin allergens was peribrmed using the Sigma reagent, cat. no. A-9263 (Poole, Dorset). SDS-FAGE and immunohlotting SDS-PAGE and immunoblotting were performed according to standard methods [20 23] using gradient gels of 7-5-17-5% total acrylamide and 25 /il samples. Molecular weight markers (Sigma MW-SDS-70L and MW-SDS-200) were applied to each gel. For immunoblotting approximately 500 ^\ of DHD extract was applied to the gel and electrophoreticaliy transferred to nitrocellulose. Strips of nitrocellulose were each incubated with 0 25 ml of IgE sera and traced with mouse monoclonal antibodies to human IgE myeloma protein [24], radio-labelled with '-^-iodine [25]. at 0 25-2 x lO*" c.p.m. per strip. After washing and drying, the nitrocellulose strips were placed against an X-ray film in a cassette with a single intensifier screen (Protex. Cuthbert Andew Ltd, Watford, U.K.) and retained at - 7 0 C for 3 days. The X-ray film was developed, the MWs of the bands determined and the intensity of radiostaining classified as strong, intermediate or weak.

795

method using 252 mg, 500 mg, I g and I g of DHD extract and pools from similar parts of the separations made giving high (HMW). medium (MMW) and low (LMW) molecular weight fractions. They were dialysed against 0 005 M ammonium bicarbonate and freeze-dried. Furification of low molecular weight allergen hy ion exchange chromatography The virtually albumin-free LMW fraction of DHD from gel filtration was further purified using Fast Protein Liquid Chromatography (FPLC) and an anion exchange column, Mono-Q (HR 5/5, Pharmacia LKB Biotechnology). The starting buffer was 20 mM ethanolamine (Sigma). pH 9 5, and the eluting buffer the same plus 1M NaCl. Both were prepared using deionized. distilled water and were filtered through a 0 22 /jm Millipore filter. The column was equilibrated with the starting buffer, a filtered sample of LMW DHD containing 3-5 mg in 500 n\ was loaded and a flow rate of 15 ml.min"' applied for 44 min. The separation included a linear gradient to 1 M NaCl over a 28 min period. Fractions with the peaks separated were collected for the first 40 min. Fractions from six separate runs were individually pooled, on a fraction to fraction basis, dialysed extensively against 0-005 M ammonium bicarbonate buffer solution and freeze-dried. They were reconstituted at a protein concentration of I mg*ml"' and examined by SDS-PAGE and RAST inhibition.

Furification of allergens Gel filtration Gel filtration was performed using a Pharmacia K.26/100 (2-6 cm X 100 cm) column packed with Ultrogel AcA 54 (Pharmacia LKB Biotechnology, Milton Keynes. U.K.) (fractionation range. 4000 to 70000 D). PBS, 0 1 5 M, pH 7-2, containing 0 02'^ w/v sodium azide was used as the running buffer. Samples of DHD extract in 5-12 ml. were filtered (0 22 ixm) and applied to the column at a flow rate of 6 ml/h '. The u.v. absorbance at 280 nm was monitored by a dual path monitor., UV-2 (Pharmacia LKB Biotechnology), and recorded on a chart recorder REC-482 (Pharmacia LKB Biotechnology). Eive millilitre fractions each were collected using a Pharmacia ERAC-3000, the absorbance at 280 nm of alternate fractions read using a Biochrom Ultraspec 4050 spectrophotometer (Pharmacia LKB Biotechnology) and plotted for each chromatographic separation. The column was calibrated using protein molecular weight calibration kit (MOL-RANGER, Pierce & Warriner, U.K.). Fractions were examined by single radial immunodiffusion (SRID) and fused rocket immunoelectrophoresis (FRIE) [26], isoelectric focusing, SDS-PAGE and RAST inhibition then combined into seven pools. Four separate gel filtration runs were performed by this

SRID and FRIE SRID and FRIE were performed according to standard methods [26], adding the antiserum to the agarose solution at a concentration of 10-20 ;d/cm', allowing diffusion for 48 hr at room temperature (SRID) or for 30 min then electrophoresing at 2 V/cm overnight (FRIE). After pressing, washing and drying, the slides were stained with Coomassie Brilliant Blue R250 solution. isoelectric focusing Isoelectric focusing was performed according to the manufacturer's instructions and used Ampholine PAGplates, pH range3 5-9 5(Pharmacia LKB Biotechnology) in conjunction with LKB 2117 Multiphor II electrophoresis unit and LKB 2103 power supply. Samples of 15 /il were applied placed approximately halfway between the anode and cathode, The gel was stained in Coomassie Brilliant Blue R-250. RAST inhibition RAST inhibition was based on well-established methods [27-29] adapted to a microplate assay system [30] and

796

A. W. Ford and D. M. Kemeny

performed as previously described for collaborative studies of allergen extracts [31. 32]. Inhibition studies were performed on the fraction pools using cither the anti-dog IgE scrum pool in its native state or after preabsorbing it with dog serum albumin (DSA. Sigma). Discs were coupled to either DSA of DHD extract at concentrations o^ 20 ;ig per disc. A 1 /5 dilution of horse serum (Gibco) in PBS was used as diluent and '-Modinelabelled monoclonal antibody to human IgE [25] used as a tracer. The number of counts bound were estimated for 30 sec. Each assay was performed in duplicate and percentage inhibition calculated. Enzyme-linked immunosorbent assay

(ELISA)

In this study, ELISA was used to investigate the presence of the major cat allergen, Fel d I, in the dog hair and dander extracts and fractions. The method used employed two monoclonal antibodies (MoAbs), one as ihe "capture" antibody and the second as the biotinylated tracer, and were used according to the method kindly supplied by Dr Martin Chapman, University of Virginia, Charlottesville, Virginia. U.S.A. The Fel d I standard used was from an assay kit kindly supplied by D r H . Baer, Office of Biologies, FDA, Bcthesda, Maryland. U.S.A. and the MoAbs from Dr Chapman. The absorbance was read at 405 nm in an ELISA microplate reader (ICN Biomedicals Ltd. High Wycombe, U.K.). The amount of Fel d I present in the extracts was found by interpolation ofthe readings on the standard curve and by parallel line assay. Results Identification of importance of the LMW allergens hy immunoblotting To assess the importance of the DHD allergens to dog allergic subjects, a reference DHD extract was separated by SDS-PAGE and the IgE binding of two positive serum pools and of 32 individual sera from dog allergic people was examined by immunoblotting. Eleven distinct IgEbinding bands, numbered I-XI. were detected by the positive serum pools ranging in molecular weight from 14000-68000 D whilst no bands were detected by the negative pools. Each individual serum bound to one or more bands with varying intensity and at various frequencies (Fig. I). Some ofthe sera {nos 5. 8. 10, 17,26.30,32) bound to virtually all of the separated components whereas a few {9. 23, 3!) bound solely to Ag X. at 21 000 D. The evaluation of the binding of the 32 sera is summarized in a histogram (Fig. 2). Ag X was bound by 25/32 sera (78%) and strongly by 11/32 (34%) of them. Serum albumin, which ran as Agsl and II, was bound by

50 59% of the sera (Agl, 16/32; Agll 19/32) and strongly by 13-19% of them. The light chain of IgG, AglX, was bound by 15/32 (47%) of sera but strongly by only 6/32 (19%) of them. Separation of DHD into HMW, fractions hy gel filtration

MMW

and

LMW

The separation of whole DHD extract on AcA 54 resulted in a chromatogram showing two major peaks of u.v. absorbance (Fig. 3), The first peak eluted with the MW markers of BSA (67 000) and ovalbumin (45 000) and the second with a MW of less than that of cytochrome C (12 500). Alternate fractions were taken directly from the separation without prior dialysis or concentration and examined by FRIE, Fractions 36 to 70 were precipitated by anti-DHD whereas the remainder of the fractions, 72116, were not (Fig. 4a). A large number ofthe precipitated fractions could also be precipitated by anti-dog serum, as shown by the incorporation of an intermediate gel containing anti-dog serum and only some of the later fractions, 52 66. contained hair and dander-specific material with little evidence of serum proteins (Fig. 4b). Further confirmation ofthe presence of serum proteins in the earlier fractions was obtained by SRID of the fractions in agarose gels containing rabbit anlisera to whole dog serum, dog serum albumin (DSA) and dog IgG. By this method dog serum proteins and DSA were found in fractions 37 50 and dog IgG was strongly precipitated in fractions 37-46 and weakly precipitated in fractions 47 53 (results not shown). Seven separate pools of the gel filtration fractions, representing the main peaks and portions of the trough were made (Fig. 3). These were dialysed against 0 005 M ammonium bicarbonate solution and freeze-dried. The total dry weight recovered from these pools was 73-3 mg which was 29% ofthe initial dry weight. The allergenic activity of each of these pools was examined by RAST inhibition to test their abiUty to prevent binding of a dog-specifk IgE serum pool to sohd phase DHD antigen. Two separate experiments were performed; one in which serum pool I was used in its native state and the second where this same pool was absorbed with an excess of DSA to remove all the antiDSA IgE activity. All fraction pools (1 to 7) inhibited binding of the unabsorbed serum with fractions 2 4 being the most inhibitory (Fig. 3. solid symbols). However, after DSA-absorption the inhibition of pools 1, 2 and 3 had dropped somewhat whereas those of 4 and 5 remained as high as before absorption (Fig. 3. open symbols). With DSA on the discs, fractions 1, 2 and 3 were strongly inhibitory with the native serum pool but

Dog dander allergen purification

797

I 2

Fig. 1. Immunoblols ofthe DHD reference extract by 32 individual sera plus positive and negative control sera. Conditions as in Materials and Methods. The sera arc numbered on the figure and the positions of allergens I XI marked.

3

10

12

13

14 15 16 4-

I 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Fig. 2. Histogram to show IgE binding of 32 sera to the allergens of the DHD referenee extraet by SDS-PAGE immunoblotting. Intensity ol" radioslaining: • . strong: M, intermediate; D, weak.

no fractions inhibited after DSA absorption ofthe serum (not shown). SDS-PAGE performed on samples of each of the seven pools from the gel filtration separation showed that considerable separation on a molecular weight basis had

100

30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0

50

I

n Allergen no.

taken place with very little material present in fraction pool 1, material of MW up to 68 000 in fractions 2 and 3 and of less than 29 000. 24 000 and 14 000 in fractions 4, 5 and 6 respectively. Virtually no material was seen in fraction 7. Both fractions 2 and 3 contained material of

798

A. W. Ford and D. M. Kemenv

70

HMW MMW LMW

50

30

10

80

120

Fraction number

-

Fig. 3. Gel filtration on Ultrogel AcA 54 of DHD extract, run no. I. Sample = 252 mg of DHD in 6 ml of running buffer; running bulTer = PBS, 0 15 M pH 7 2 + 002% sodium azide; flow rate = 6 ml/hr; fractions = 50 min (5 ml). The elution position of the molecular weight markers. BSA 67000, ovalbumin 45000, chymotrypsinogcn 25000 and cytochrome C 12 500 are marked on thefigures.The pools (1-7) are indicated along the top of the figure, and their ultimate destination in the HMW, MMW and LMW pool from separate runs are represented by shaded areas under the curve. Percentage RAST inhibition is shown lor each of the pools. Discs were coated with unfractionated DHD extract. Equal volumes of serum pool 1 and fractions 1-7 were incubated lor I hr at 37 C before a DHD disc was added. • ^percentage inhibition using unabsorbed scrum and 0 = percentage inhibition using DSA-absorbed serum.

(o)

anti-OHO

I I 1 I I I I I 1 I 3G 40 44 48 52 56 SO 64 68 72

I I I I I I I 1 I I 78 82 86 90 94 98 106 114 102 IIO

b)

I I I I I 36 40 44 48 52 56 60 64 66 72

I I I I I 78 8? 86 90 94 98

106 114 102 IIO

Fig. 4. Fused rocket immunoeleetrophoresis (FRIE) of Ultrogel AcA 54 fractions of DHD extract. Rabbit anti-DHD was incorporated into the lower (cathodic) portion of the gel, an intermediate gel contained (a) no antiserum or (b) antidog serum and the remainder (anodie) portion also contained anti-DHD. Alternate fractions from 36-116 were applied to the wells, allowed to diffuse for 30 min and electrophorescd at 2 V/ cm overnight. The gel was stained wilh Coomassie Brilliant Blue.

Dog dander allergen purification

799

IOO 0-4

0-3

50

0-2

Fig. 5. Anion exchange FPLC of LMW Iraction of DHD extract. Mono-Q HR 5/5 column. Solid line, absorbance at 280 nm; dashed line, percentage concentration of buffer B (— buffer A + 1 M NaCl). BufTer A = 20mM cthanolaminc solution. pH 9 5. SampIe-5 mg LMW DHD.

0-1

approximately 68 000, the MW of serum albumitis, but there was a difTerence in the MW distribution of the lower MW components. A gel where higher sample loading was used showed a band in the albumin region in all seven fractions but with fraction pools 5 and 6 to be virtually free of albumin. Pool 7, which was the entire second chromatographic peak, contained very little Coomassiestaining material (gel not shown here). Inter-run pools of HMW. a MMW and a LMW material were made by combining four gel filtration separations. The HMW pool comprised the descending arm of the first chromatographic peak (fraction pool 3 in the first run shown here. Fig. 3) and, although 45000 D and less by column calibration appeared to contain a range of MW from 68 000 to < 14000 D by SDS-PAGE. It included the major part of the serum proteins (as shown FRIE and by SRID) and exhibited activity by RAST inhibition before and after DSA-absorption. The MMW pool contained the latter part of the descending portion of the first peak together with the first part of the trough (fraction pool 4, Fig. 3). Itcontained very little albumin as seen in SDS-PAGE and was a mixture of hair and danderspecific proteins and some serum proteins. The LMW pool was the remainder of the trough between the two peaks on the chromatogram (fraction pool 5, Fig. 3). It appeared to contain extremely little albumin on SDSPAGE and was comprised almost entirely of danderspecific proteins with a minimum of serum protein contamination as seen on FRIE and SRID. It also had strong inhibitory action in RAST even after absorption of ihe serum pool with DSA. Examination of the HMW., MMW, and LMW pools by

10

20

30

40

Elution volume (ml)

isoelectric focusing indicated that the HMW material had bands in the albumin region (pi =4-5) whereas the MMW pool and, to a greater extent, the LMW pools exhibited bands ranging from pi 4-5-5-5 (not shown). Separation by anion exchange FPLC The absorbance at 280 nm of the LMW pool after separation by FPLC on a Mono-Q anion exchange column is given in Fig. 5. The major portion of the material eluted after the passage of 6-12 ml of buffer as the concentration of sodium chloride rose from 0 1 - 0 4 M. Fractions were collected and combined on a fraction to fraction basis from six identical runs. Fractions 2-22 were examined individually by SDS-PAGE stained with Coomassie Brilliant Blue. Staining was evident for bands of MW between < 14000-29000 D. Progressive loss of one Coomassie-staining band and replacement with another occurred between fractions2-i7 on SDS-PAGE but there were no bands visible from fraction 18 onwards (not shown). The appearance by SDS-PAGE of the extract before and after chromatography is shown in Fig. 6; Figure 6(a) shows the appearance of the extract before and after gel filtration to produce HMW. MMW and LMW pools and Fig. 6(b) shows the appearance of the LMW fraction after anion exchange chromatography. RAST inhibition activity was most pronounced in fractions 6-12 following ion exchange chromatography ranging from 50-70'^. inhibition and in these fractions stained bands of between 20000-29000 D were present in SDS-PAGE (Fig. 6b). This activity, however, did not appear to correlate with the intensity of staining of a

800

{o )

A. W. Ford and D. M. Kemenv

t

2

3

4

Iss

5

(b)

.-

I

2

3

4

5

6

7

205 M6 97-4 66

66

-- 45 - 36 — 29 24

-

45 36 29 24

20-1

20-1 -^ 14-2

LMW fr6

fr7 frS

(r9 frIO MW MW

DHO HMW MMW LMW MW MW « 10 (c)

I

2

66

45 36 29 24

20-1 DHD

LMW

fr 4 - 6

fr7-l2

— . 14-2

MW

prominent band at 24000 D but rather with a 21 000 D band seen more intensely after immunoblotting with serum pool 2 (Fig. 6c). Interestingly, although serum albumin appeared virtually absent in the LMW pool when examined by Coomassie staining of SDS-PAGE and was only seen faintly by blotting, concentration was somewhat enhanced in the later fractions by ion exchange chromatography. Fractions 4-6 bound IgE very strongly at 21000 with no albumin detectable and gave strong RAST inhibition with the DSA-absorbed serum pool.

Presence of AgX in other dog materials Using marker sera nos 9 & 31. which bound only to AgX in immunoblotting of SDS-PAGE, this Ag X was shown also to be present in dog saliva but to be absent from both dog urine and serum (not shown). These same two human

Fig. 6. SDS-PAGE and immunoblotting of dog hair and dander (DHD) extract at various stages of purification of Ag X. (a) SDS-PAGE on 7-5-17 5% acrylamide gel stained with Coomassie Brilliant Blue R-250. Lanes 1-4 results of get filtration on Uilrogel AcA 54: lane I. unfractionated DHD; lane 2, HMW; lane 3. MMW; lane 4, LMW; lane 5, MW markers, (b) SDS-PAGE on 7'5-17'5% acrylamide gel stained with Coomassie Brilliant Blue R-250. Lane 1. LMW DHD; lanes 2 6. fractions 6-10 from anion exchange on Mono-Q HR 5/5; lane 7. MW markers. (c) Immunoblotting of SDS-PAGE. The gel was blotted and incubated successively with serum pool 2 al 0 25 ml per track and radiolabelled antihuman IgE then autoradiographed for 3 days at —70 C. Lane I, unfractionated DHD extract; lane 2. LMW DHD from gel filtration; lane 3. pooled fractions 46 from anion exchange FPLC. The position of the MW markers is shown on the RHS.

sera had bound to only a single precipitin arc when studied by CRIE [1] and thus Ag X by SDS-PAGE was identified as the same allergen as Ag 8 in CRIE. Non-identity of Ag X and Fel d I The MMW and LMW fractions, both seen on SDSPAGE to contain AgX. were tested for the presence of Fc/ (/1. Whereas the reference cat extract contained 3 units per milligram dry weight, the whole DHD extract contained