Proc. Natl. Acad. Sci. USA Vol. 91, pp. 12496-12500, December 1994 Cell Biology Hereditary eosinophil peroxidase deficiency: Immunochemical and spectroscopic studies and evidence for a compound heterozygosity of the defect MAURIZIO ROMANO*, PIERLUIGI PATRIARCA*, CARLOS MELOt, FRANCIsco E. BARALLEt, AND PIETRO DRI** *Istituto di Patologia Generale, Universita di Trieste, Via A. Fleming, 22, 34127 Trieste, Italy; and tInternational Centre for Genetic Engineering and Biotechnology, Padriciano 99, I-34012, Trieste, Italy Communicated by Paul B. Beeson, July 22, 1994

ABSTRACT Hereditary eosinophil peroxidase (EPO; EC cursors derived from blood progenitor cells reveals a com- 1.11.1.7) deficiency is a rare abnormality of eosinophil gran- pound heterozygosity for the defect, with mutations consist- ulocytes characterized by decreased or absent peroxidase ac- ing of a base transition leading to an amino acid substitution tivity and decreased volume of the granule matrix. The mo- (Arg286 -- His) and an insertion in an intron-exon junction lecular basis of the defect is not known. We report here its that, by shifting the reading frame, generates a premature molecular characterization in an EPO-deficient subject and his stop codon. The family studies are compatible with an family members. The EPO-deficient eosinophils contained autosomal recessive mode of inheritance of the defect.§ EPO-related material as determined immunochemically using either monoclonal or polyclonal anti-EPO antibodies but had MATERIALS AND METHODS no spectroscopic evidence of EPO. Eosinophil precursors from Cell Isolation. Granulocytes were isolated from acid citrate the EPO-deficient subject contained normally sized EPO dextrose-anticoagulated peripheral blood (7). Cell popula- mRNA, which was reverse transcribed into the corresponding tions enriched in eosinophils were prepared by an adaptation cDNA clones encompassing the whole gene. Sequencing ofthese ofthe method ofRoberts and Gallin (8). Briefly, after dextran clones disclosed two mutations, a G -* A transition causing a sedimentation oferythrocytes, the white cell-rich plasma was nonconservative replacement of an arginine residue with a incubated with 5 ,.M fMet-Leu-Phe for 15 min at 37°C under histidine and an insertion causing a shift in the reading frame gentle agitation. The cell suspensions were then centrifuged with the appearance of a premature stop codon. The two on Lymphoprep (Nycomed, Oslo); most eosinophils local- mutations were located on different chromosomes indicating a ized with the erythrocytes at the bottom ofthe tube, whereas compound heterozygosity for the defect. Both the son and the mononuclear cells and most neutrophils remained at the cell daughter of the proband inherited the G -* A transition, and suspension/Lymphoprep interface. Eosinophil-enriched cell their eosinophils contained a peroxidase activity intermediate preparations were freed from erythrocytes by a brief hypo- between that of control subjects and the proband, suggesting tonic lysis. that the transition is a deficiency-causing mutation. Eosinophil Spectroscopy. The reduced-minus-oxidized difference precursors from the EPO-deficient subject were found to spectra of cell homogenates were obtained with a double- actively synthesize an EPO that was apparently normal in beam spectrophotometer (9). Cells [3-6 x 106 per ml in terms of cytochemical reaction for peroxidase and immunore- phosphate-buffered saline (PBS)] were sonicated and added activity with monoclonal and polyclonal anti-EPO antibodies, to both the sample and the reference cuvette. The absorption but spectroscopically abnormal. The cytochemical reaction for spectrum was then recorded after addition of a few grains of peroxidase tended to decrease or disappear in the eosinophil dithionite to the sample cuvette. precursors of the EPO-deficient subject but not of a normal Cytochemistry. Cytochemical staining for peroxidase and subject as differentiation went on, suggesting that the Arg -+ Sudan black staining were performed as described (10, 11). His substitution causes the production of an unstable EPO that Determination ofEPO Activity. EPO activity was measured undergoes progressive degradation as the cells mature. by a method based on oxidation of homovanillic acid (12). Antibodies. Anti-EPO antiserum was obtained by immu- nizing rabbits with purified EPO. The anti-EPO monoclonal Eosinophil peroxidase (EPO; donor:hydrogen-peroxide oxi- antibodies (mAbs) H2G3 and B2G8 were prepared and char- doreductase, EC 1.11.1.7) is a highly basic heme protein acterized as described (13). Antiserum against major basic contained in the specific granules ofeosinophil granulocytes. protein (MBP) and the anti-myeloperoxidase (MPO) mAb It is a 70-kDa dimer composed of a 15-kDa light chain and a MPO3 (14) were kindly donated by G. Gleich (Mayo Clinic, 55-kDa heavy chain held together by a disulfide bond (1). Rochester, MN) and Dr. C. E. van der Schoot (Central EPO deficiency was described for the first time by Pre- Laboratory of the Netherlands Red Cross Blood Transfusion sentey in 1968 (2) and about 100 subjects with the abnormality Service, Amsterdam, The Netherlands), respectively. have been reported (3-6). Cytochemical and biochemical Immunocytochemistry. Cytospin preparations were fixed family studies have suggested an autosomal recessive pattern with 0.125% glutaraldehyde in PBS and washed with 1% oftransmission ofthe defect in some subjects, while in others human serum in PBS to neutralize excess fixative. After a definite pattern of transmission has not been established. permeabilization with methanol/acetone (1:1, vol/vol), the The molecular mechanisms underlying the defect of EPO endogenous peroxidase was inactivated by incubating the are unknown. We report here that the eosinophils from an cytospins for 60 min at 37°C with a solution containing 10 mM EPO-deficient subject lack the characteristic EPO absorption glucose, 1 unit of glucose oxidase per ml, 2 mM NaN3 (15), spectrum but contain peptides immunochemically related to EPO. Analysis of the cDNA obtained from eosinophil pre- Abbreviations: EPO, eosinophil peroxidase; mAb, monoclonal an- tibody; MBP, major basic protein; MPO, myeloperoxidase. The publication costs of this article were defrayed in part by page charge 1To whom reprint requests should be addressed. payment. This article must therefore be hereby marked "advertisement" §The sequences reported in this paper have been deposited in the in accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession nos. Z37727 and Z37728). 12496 Downloaded by guest on October 2, 2021 Cell Biology: Romano et al. Proc. Natl. Acad. Sci. USA 91 (1994) 12497 and 5 mM resorcinol (16). The cytospins were then incubated with a rabbit anti-EPO antiserum or the corresponding pre- immune serum diluted 1:1000 in PBS or with anti-EPO mAbs as ascites diluted 1:500 in PBS. Binding of antibodies was revealed by the avidin/biotin complex immunoperoxidase or alkaline phosphatase method (Vectastain ABC system; Vec- tor Laboratories) and counterstained with Mayer's hemalum. Eosinophil Precursors. These were obtained from liquid cultures of peripheral blood mononuclear cells (17). Molecular Genetic Techniques. RNA was extracted (18) from eosinophil precursors. Poly(dT)-primed cDNA was syn- thesized with a first-strand cDNA synthesis kit (Pharmacia). EPO cDNA was specifically amplified by nested PCR. In 400 450 500 550 400 450 500 550 the first round of amplification, two primers spanning the Wavelength, nm whole coding mRNA (nt -34 to 2247) were used. The second FIG. 1. Reduced-minus-oxidized absorption spectra of eosino- round used primer couples designed to amplify overlapping phils from the EPO-deficient subject F.V. (Left), from a control portions ofthe 2.3-kb EPO cDNA (nt -34 to 625, 339 to 1127, subject (Center), and from an erythrocyte lysate (Right). The soni- 1050 to 1721, and 1659 to 2247). A fragment comprising nt 625 cates from 4 x 106 granulocytes per ml containing 90o eosinophils to 1608 was also amplified, to establish whether the two (Left and Center) and the lysate from about 106 erythrocytes per ml mutations detected (see Results) were located on the same (Right) were used. allele or on different alleles. Standard PCR was carried out for 35 cycles (1 min at 93°C, 1 min at 56°C, 3 min at 72°C) in minus-oxidized absorption spectra ofthe eosinophil homoge- 100-,ul reaction volumes containing 1.5 mM MgCl2, 3% nates from a control subject and from F.V. and of an (vol/vol) dimethyl sulfoxide, and 2 units of Taq DNA poly- erythrocyte lysate. The spectrum of the control subject (Fig. merase (Perkin-Elmer/Cetus). The PCR products were 1 Center) shows a broad asymmetric peak with a maximum blunt-end cloned in pUC18 Sma I/BAP (Pharmacia) and at 436 nm and a shoulder at 444 nm. The 436-nm peak is sequenced with a T7 sequencing kit (Pharmacia). contributed mainly by hemoglobin that contaminates the A 260-bp genomic DNA fragment of EPO containing exon eosinophil sonicate (Fig. 1 Right). A minor contribution to 10 and its flanking regions and a 370-bp genomic DNA this peak may be given by the low-potential cytochrome b, fragment containing exon 7 and its flanking regions were which is known to be present in eosinophils and to have a amplified from DNA extracted from circulating leukocytes; characteristic absorption peak at 428 nm (21). The shoulder for this, PCR was carried out for 35 cycles (1 min at 93°C, 1 at 444 nm is contributed by EPO, which has an absorption min at 56°C, 1 min at 72°C), 100 ,ul with 1.5 mM MgC92. The peak at 448 nm (22). The slight shift ofthe EPO peak from 448 amplified fragments were directly sequenced. nm to 444 nm is caused by hemoglobin and cytochrome b in For Southern blot analysis, the amplified 370-bp genomic the sample. In the EPO-deficient subject (Fig. 1 Left) the DNA fragment was electrophores d in a 1% agarose gel and EPO peak at 444 nm is absent. blotted to a nitrocellulose membrane (19). Blots were probed Immunochemical Studies. Immunocytochemistry was per- with oligonucleotides (5' end-labeled with 32P) at 42°C over- formed on peripheral blood eosinophils with an anti-EPO night. After hybridization the filters were washed at 47°C to antiserum and with the anti-EPO mAb B2G8. The eosinophils a final stringency of 1 x standard saline citrate and exposed of F.V. gave a positive reaction with the anti-EPO antiserum to Kodak X-Omat AR film. and with the mAb (Fig. 2). Oligonucleotides. Oligodeoxynucleotide sequences were as Genetic Studies. To characterize the nature ofthe alteration follows: -34 [number corresponds to the first base of the at the genetic level, we tried to obtain EPO cDNA. Since oligonucleotide, according to numbering of the sequence peripheral blood eosinophils do not contain detectable levels reported by Sakamaki et al. (20)], 5'-TCACTTCCCAGCTG- ofEPO mRNA (17, 23), we used eosinophil precursors grown GTGAA-3'; 2247, 5'-GTCACGAGCCTGCACTTGCT-3'; in liquid cultures from peripheral blood mononuclear cells 625, 5'-AAGCGCACAATCTGGTTGGA-3'; 339, 5'-GT- CACTGATGTGCTAACAGA-3'; 1127, 5'-GTGTCACCT- GCCAGGAAGCA-3'; 1050, 5'-TTCGACAACCTGCAC- GATGA-3'; 1721, 5'-CCAAGCATTGTACCCTGT-3'; 1659, 5'-GACCTGGCAGCTCTCAACAT-3'; 659, 5'-AGCCCT- CATGTTCATGCAGT-3'; 1608, 5'-CTAACATGGCAT- CCTGA-3'; exon 7 forward primer, 5'-CACTGTCTCCTC- TTCCAT-3'; exon 7 reverse primer, 5'-GTTTCCTGGGAA- GACACCA-3'; exon 10 forward primer, 5'-AGAGGCT- GGTCCAATCTGT-3'; exon 10 reverse primer, 5'-AGCCT- GGCCCGAGCAAAGCT-3'. The oligonucleotides used for Southern blots were the following: normal probe, 5'-TTTCTTCC(jCTCGGCAC-3'; mutant probe, 5'-TTTCTTCCACTCGGCAC-3'. RESULTS The propositus (F.V.) was a blood donor at the blood bank of Trieste Hospital. The EPO defect was identified by auto- mated flow cytochemistry and confirmed by (i) lack of FIG. 2. Immunochemical detection of EPO in eosinophils from a cytochemical and biochemical activity and absence of Sudan normal subject (a and b) and the EPO-deficient subject (c and d). (a black staining in his eosinophils (ref. 6, subject 2; ref. 12) and and c) Anti-EPO antiserum (ABC alkaline phosphatase). (b and d) (ii) presence of the characteristic decrease in the volume of Anti-EPO mAb B2G8 (ABC peroxidase). (Insets) Eosinophils from the specific granule matrix (ref. 6, subject 2). EPO-deficient subject reacted with a rabbit preimmune serum (a) and Spectrophotometric Studies. Fig. 1 shows the reduced- with a nonspecific ascites (b). (x500.) Downloaded by guest on October 2, 2021 12498 Cell Biology: Romano et al. Proc. Natl. Acad. Sci. USA 91 (1994)

A exon 9 exon 10 153 1/511 1561/521 TAT GAA G gtgacc...caccag GG C ATC GAC CCC ATC CTC CGG GGC CTC ATG GCC ACC CCT tyr glu g ly gly ile asp pro ile leu arg gly leu met ala thr pro

1591/531 GCC AAG CTG AAC CGT CAG GAT GCC ATG TTA GTG GAT GAG ala lys leu asn arg gln asp ala met leu val asp glu

B exon 9 exon 10 FIG. 3. Nucleotide and deduced amino 153 1/511 1561/521 acid sequence of an EPO genomic DNA TAT GAA G gtgacc ... caccag GG GG CAT CGA CCC CAT CCT CCG GGG CCT CAT GGC CAC CCC tyr glu g ly gly his arg pro his pro pro gly pro his gly his pro fragment from a normal subject (A) and the EPO-deficient subject (B). The region con- 1591/531 taining the G insertion is underlined. Upper- TGC CAA GCT GAA CCG TCA GGA TGC CAT GTT AGT GGA TGA case, exon 9 and exon 10; lowercase, intron cys gln ala glu pro ser gly cys his val ser gly OPA between exon 9 and exon 10. (17). By using EPO-specific primers for PCR, four partially genomic DNA clones ofexon 7 fragments from both F.V. and overlapping fragments of cDNA encompassing the whole his offspring (data not shown). gene were obtained, cloned, and sequenced. The possibility that the G -- A substitution was a popu- In comparison with the sequence ofthe coding region ofthe lation polymorphism was tested by examining its occurrence EPO gene reported by Sakamaki et al. (20), two different point in the genomic DNA from 50 unrelated subjects belonging to mutations were found: a G -) A transition at position 857 that the same ethnic group. Exon 7, which contains the mutation, predicts the nonconservative substitution of histidine for ar- was amplified by PCR and then hybridized with specific ginine at codon 286 and an insertion of a G within the G tract probes. The substitution was shown to be unique to F.V. and 1537-1541 at the intron-exon 10 junction. This causes a shift his offspring (data not shown), indicating that the variation is in the reading frame with the generation of a stop codon after unlikely to be a population polymorphism. codon 538 (Fig. 3), leading to a truncated EPO-precursor that Characterization of EPO in Eosinophil Precursors. At day lacks 177 aa and has a completely subverted sequence of the 13, in liquid cultures of peripheral blood mononuclear cells 24-aa carboxyl-terminal tail downstream of the insertion. incubated with interleukin 5 and cell line 5637-conditioned To establish whether the two mutations were located on medium, eosinophil promyelocytes and myelocytes were the same allele or on different alleles, a cDNA fragment clearly evident, as described before (17), and the precursors corresponding to a region including both mutations (nt 659- from F.V. were positive for peroxidase cytochemistry, al- 1608) was amplified by PCR, cloned, and sequenced. Each though with a degree of reactivity less pronounced than that clone was found to contain only one of the two mutations, of the precursors from a normal subject (see Fig. 7 a and d). indicating a compound heterozygosity for the mutant alleles. Day 13 cultures from F.V. contained cells that strongly Family Studies. The wife, the son, and the daughter ofF.V. reacted with the anti-EPO antiserum (Fig. 5a) and with the were available for studies. The proband's parents were not anti-EPO mAbs B2G8 and H2G3 (Fig. 5 d and e). alive, and his two brothers didn't cooperate with the study. Cells with morphological characteristics comparable to Table 1 shows the results of the biochemical assay of EPO those of the cells that reacted with the anti-EPO antibodies activity in F.V. and his family members. The peroxidase also reacted with an anti-MBP antiserum (Fig. Sb). Since activity of the wife was normal, while that of the son and the MBP is known to be exclusively present in eosinophils, this daughter was 33-40% that of the control subjects, a finding suggests that the peroxidase-positive cells found in these compatible with an autosomal recessive mode of inheritance cultures actually belong to the eosinophilic lineage. This was of the defect. confirmed by counting the cytochemically peroxidase- To identify which mutation was inherited by the children of positive cells and the cells reacting with the anti-EPO anti- F.V., we directly sequenced exon 10 DNA fragments con- serum, the anti-MBP antiserum, and an anti-MPO mAb that the G insertion and blot studies recognizes both the MPO precursor and the mature MPO taining performed Southern (14). The number of cells was equivalent on exon 7 DNA fragments containing the G -- A transition. peroxidase-positive The sequence of F.V. exon 10 was found to contain the to the number of anti-EPO- and anti-MBP-reactive cells insertion, whereas it was normal in both offspring (data not (Table 2). The number of cells reactive with the anti-MPO shown). Exon 7 DNA was shown to hybridize with both the mAb was negligible. These data clearly rule out the possi- normal and the mutant indi- bility that the peroxidase-positive cells in these cultures oligonucleotide probes (Fig. 4), to other as cating that both offspring inherited the allele with the G -3 A might belong peroxidase-containing cells such transition. This mutation was also detected by sequencing F M S D Table 1. EPO activity of eosinophils from the EPO-deficient subject and his family members 4i 4m 40 Arg286 Source of Fluorescence, granulocytes % of control F M S D Control subjects 100

EPO-deficient subject <0.5 S -.rsa.ops :: His286 Wife 120 Son 40 Daughter 33 FIG. 4. Allele-specific hybridization ofArg/His EPO sequence in Granulocyte suspensions containing 7-20% eosinophils were di- the family members of the EPO-deficient subject. A DNA fragment luted in the assay buffer and the fluorescence generated by homo- containing exon 7 was amplified from genomic DNA, run on a 1% vanillic acid oxidation was measured after 120 min of incubation at agarose gel, transferred to nitrocellulose membrane, and hybridized 37°C (12). The activity of control subjects was 125 ± 18 (mean ± to oligonucleotide probes specific for G at 857 (Arg286) or A at 857 SEM, n = 3) fluorescence units per 106 eosinophils. (His286). Lanes: F, father (F.V.); M, mother; S, son; D, daughter. Downloaded by guest on October 2, 2021 Cefl Biology: Romano et al. Proc. Natl. Acad. Sci. USA 91 (1994) 12499 _._ I OD = 0.005

400 450 500 550 600 400 450 500 550 600 Wavelength, nm FIG. 6. Reduced-minus-oxidized absorption spectra of eosino- phil precursors from a normal subject (Left) and the EPO-deficient subject (Right) at day 21 of culture. The homogenates from 20 x 106 S.. cells per ml containing about 15% eosinophil precursors were used. FIG. 5. Immunocytochemical detection of EPO and MBP in the deficient subject have, in addition, no spectroscopic evidence precursors from the EPO-deficient subject. (a) Anti-EPO antiserum. of the EPO heme group but do contain material antigenically (b) Anti-MBP antiserum. (d) Anti-EPO mAb H2G3. (e) Anti-EPO related to EPO as judged by immunochemical analysis. This mAb B2G8. (c and I) Cell preparations reacted with a rabbit latter finding is strengthened by the observation that the preimmune serum and a nonspecific ascites, respectively. eosinophil precursors ofthe EPO-deficient subject contained neutrophils and monocytes. The reduced-minus-oxidized ab- normally sized EPO mRNA and actively synthesized the protein as indicated by their sorption spectra of eosinophil precursor homogenates from strong cytochemical reaction for peroxidase (Fig. 7) and immunoreactivity with an anti-EPO both F.V. and a normal subject (Fig. 6) show the 430-nm peak antiserum and two anti-EPO mAbs (Fig. 5). The synthesis of of cytochrome b, but the characteristic peak of EPO at 448 EPO was confirmed by pulse labeling of the precursors with nm, which is present in the control subject, is shifted to the [35S]methionine for 75 min followed by a 16-hr chase. The right in F.V., with a maximum absorption at 520 nm. This is band pattern observed after precipitation with the anti-EPO at variance with the mature eosinophils of F.V. which do not antiserum was similar in a normal subject and the EPO- exhibit any absorption peak in this region (see Fig. 1). Precursors from the normal subject (Fig. 7 a-c) were homogeneously positive for peroxidase activity, with a high degree of staining throughout the observation time (days 13, 21, and 27). The precursors from the EPO-deficient subject (Fig. 7 d-J), apart from being less positive than their normal counterparts, displayed some heterogeneity in cell staining. This heterogeneity increased from day 13 to day 21 and was particularly evident at day 27 of culture, when cells almost totally negative for peroxidase were observed. DISCUSSION This study reports a detailed description of the molecular basis of EPO deficiency. Peripheral blood EPO-deficient eosinophils are characterized by the absence ofEPO catalytic activity. We show here that the eosinophils of an EPO- Table 2. Cytochemically peroxidase-positive precursors from the

EPO-deficient subject react with antibodies against the eosinophil :..,...... '.,".'..', granular proteins EPO and MBP -.,.7.' No. of positive cells Test per cytospin Peroxidase cytochemistry 314 Anti-EPO antiserum 304 Anti-MBP antiserum 280 Anti-MPO mAb 16 Cytospin preparations from day 13 cultures of precursors were FIG. 7. Cytochemical staining for peroxidase activity in eosino- reacted for peroxidase cytochemistry and for immunocytochemistry phil precursors in liquid cultures from a normal subject (a-c) and the with the indicated antibodies. The positive cells in each condition EPO-deficient subject (d-f), at day 13 (a and d), day 21 (b and e), and were counted microscopically. day 27 (c and f) of culture. Downloaded by guest on October 2, 2021 12500 Cell Biology: Romano et al. Proc. Natl. Acad. Sci. USA 91 (1994) deficient subject (data not shown). A previous immunoelec- resolution the region containing Phe'46 resides in a zone that tron microscopic study from our laboratory (6) based on the covers the a-helical core involved in heme binding. It is likely so-called postembedding technique failed to detect EPO- that a new His147 in the immediate heme environment may immunoreactive material in the eosinophils of the same alter the formation of the coordination bonds and hence EPO-deficient subject described in the present paper. The account for changes in both the spectroscopic characteristics discrepancy may be explained by the treatments employed in and the stability of the mutated EPO. This possibility is the postembedding procedure to etch the ultrathin sections, supported by the results ofthe spectrophotometric studies on which may have seriously altered the antigenic properties of the eosinophil precursors. In these cells the typical EPO the residual EPO-related immunoreactive material, as has absorption peak at 448 nm was replaced by a broad and been described for other proteins (24). asymmetric peak with a maximum at 520 nm. This suggests The molecular biology studies showed changes in the an alteration of the normal amino acid environment around amino acid sequence of EPO in the deficient subject. Anal- the heme within the newly synthesized EPO protein. ysis of the EPO cDNA and genomic DNA showed that the mutations in the We thank Dr. P. Spessotto and Dr. R. Menegazzi for help in culture proband is a compound heterozygote for two of precursors and assay of EPO, respectively. This work was coding region of the EPO gene: an insertion in a sequence supported by grants from Ministero Universita Ricerca Scientifica e adjacent to a splice site and a transition. Neither mutation Tecnologica (40% and 60%) and by Associazione Italiana per la was detected in another previously described EPO-deficient Ricerca sul Cancro and Consiglio Nazionale delle Ricerche (Target subject (ref. 6, subject 5; ref. 12, subject Z.F.), suggesting Project on Biotechnology and Bioinstrumentation). that the genetic basis for EPO deficiency is heterogeneous. 1. Olsen, R. & Little, C. (1983) Biochem. J. 209, 781-787. The insertion puts exon 10 out of frame and leads to a 2. Presentey, B. (1968) Am. J. Clin. Pathol. 49, 887-890. premature stop codon. This results in a truncated protein 3. Presentey, B. (1984) Acta Haematol. 71, 334-340. lacking 177 aa and containing an abnormal stretch of 24 aa in 4. Presentey, B. (1970) Acta Haematol. 44, 345-354. its carboxyl terminus. As mentioned above we could not 5. Spry, C. J. F. (1988) Eosinophils (Oxford Univ. Press, Oxford), detect the truncated EPO by pulse-chase experiments. It is pp. 259-262. likely that the prematurely terminated EPO undergoes rapid 6. Zabucchi, G., Soranzo, M. R., Menegazzi, R., Vecchio, M., degradation, as it has been shown for other truncated pro- Knowles, A., Piccinini, C., Spessotto, P. & Patriarca, P. (1992) teins (25-27). This process is likely to be speeded up by the Blood 80, 2903-2910. alteration in the sequence of the carboxyl-terminal 24 aa. 7. Zabucchi, G., Menegazzi, R., Cramer, R., Nardon, E. & P. Immunology 69, 580-587. -- to Patriarca, (1990) The transition (CGC CAC) would lead the synthesis 8. Roberts, R. L. & Gallin, J. I. (1985) Blood 65, 433-440. of a protein with a nonconservative amino acid substitution 9. Dri, P., Cramer, R., Soranzo, M. R., Comin, A., Miotti, V. & (Arg286 - His). That the G -- A substitution was unique to Patriarca, P. (1982) Blood 60, 323-327. the subject studied and was not found in the DNA samples of 10. Kaplow, L. S. (1965) Blood 26, 215-219. 50 unrelated random controls suggests that the variation is 11. Bennet, J. M. & Reed, C. E. (1979) in The Leukemia Cell, eds. unlikely to be a population polymorphism. Evidence that the Rubin, A. D. & Waxman, S. (CRC, Boca Raton, FL), pp. 7-24. Arg -> His substitution is a deficiency-causing mutation 12. Menegazzi, R., Zabucchi, G., Zuccato, P., Cramer, R., Picci- comes from the study of the proband family. In fact, both his nini, C. & Patriarca, P. (1991) J. Immunol. Methods 137, 55-63. son and daughter inherited the transition but not the insertion 13. Romano, M. (1990) Degree thesis (Univ. of Trieste, Trieste, Italy). and displayed a level of peroxidase activity which was 14. van der Schoot, C. E., Daams, G. M., Pinkster, J., Vet, R. & intermediate between the normal level and that of the father. von dem Borne, A. E. G. K. (1990) Br. J. Haematol. 74, The changes induced by the Arg -> His substitution seem 173-178. to lead, in the early precursors, to the synthesis of an EPO 15. Ohno, Y. & Gallin, J. I. (1985) J. Biol. Chem. 260, 8438-8446. protein that is spectroscopically abnormal but enzymatically 16. Cramer, R., Bisiacchi, B., de Nicola, G. & Patriarca, P. (1980) and immunochemically similar to the normal protein. That Adv. Exp. Med. Biol. 121A, 91-99. the EPO can no longer be detected either cytochemically or 17. Romano, M., Melo, C., Baralle, F. & Dri, P. (1992) J. Immunol. spectroscopically in mature peripheral blood eosinophils Methods 154, 265-267. suggests that the protein may be unstable and undergo 18. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162, 156-159. degradation. This possibility is supported by the observation 19. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular of cells with a decreased or absent cytochemical reaction for Cloning: A Laboratory Manual (Cold Spring Harbor Lab. peroxidase in older cultures of eosinophil precursors as Press, Plainview, NY), p. 382. opposed to the uniformly and more intensely stained cells 20. Sakamaki, K., Tomonaga, M., Tsukui, K. & Nagata, S. (1989) present in younger cultures. On this basis it can be concluded J. Biol. Chem. 264, 16828-16836. that the observed amino acid substitution is responsible, in 21. Segal, A. W., Garcia, R., Goldstone, A. H., Cross, A. R. & this subject, for the protein alterations leading to the absence Jones, 0. T. G. (1981) Biochem. J. 196, 363-367. of both the spectroscopic properties and the enzymatic 22. Wever, R., Hamers, M. N., Weening, R. S. & Roos, D. (1980) activity ofEPO. A known example ofa genetic defect leading Eur. J. Biochem. 108, 491-495. 23. Gruart, V., Truong, M. J., Plumas, J., Zandecki, M., Kusnierz, to synthesis ofan unstable protein is provided by the glucose- J. P., Prin, L., Vinatier, D., Capron, A. & Capron, M. (1992) 6-phosphate dehydrogenase deficiency where a normal Blood 79, 2592-2597. amount of enzymatically active protein is present in bone 24. Horisberger, M. (1992) Int. Rev. Cytol. 136, 227-287. marrow cells and reticulocytes but mature erythrocytes con- 25. Fei, Y. J., Stoming, T. A., Kutlar, A., Huisman, T. H. J. & tain a decreased number ofthe molecules, accounting for the Stamatoyannopoulos, G. (1989) Blood 73, 1075-1077. decreased enzymatic activity (28). 26. Thein, S. L., Hesketh, C., Taylor, P., Temperley, I. J., The Arg286- His substitution described in this paper is not Hutchinson, R. M., Old, J. M., Wood, W. G., Clegg, J. B. & included in the putative heme-binding region proposed on the Weatherall (1990) Proc. Natl. Acad. Sci. USA 87, 3924-3928. basis ofhomology with other mammalian peroxidases (20, 29) 27. Bolscher, B. G. J. M., de Boer, M., de Klein, A., Weening, R. S. & Roos, D. (1991) Blood 77, 2482-2487. but is adjacent to Phe285. This amino acid, which corresponds 28. Beutler, E. (1978) The Metabolic Basis ofInherited Disease, to Phe146 according to the amino acid numbering of Zeng and eds. Stanbury, J. B., Wyngaarden, J. B. & Fredrickson, D. S. Fenna (30), has been shown by those authors to be univer- (McGraw-Hill, New York), 4th Ed. pp. 1430-1451. sally conserved among mammalian peroxidases and to belong 29. Kimura, S. & Saito, M. I. (1988) Proteins Struct. Funct. Genet. to a group of residues lying in close proximity to the heme 3, 113-120. group. In fact in the proposed x-ray crystal structure at 3-A 30. Zeng, J. & Fenna, R. E. (1992) J. Mol. Biol. 226, 185-207. Downloaded by guest on October 2, 2021