Proc. Natl. Acad. Sci. USA Vol. 83, pp. 24-27, January 1986 Biochemistry Hemoglobin Long Island is caused by a single mutation (adenine to cytosine) resulting in a failure to cleave amino-terminal methionine (globin/initiation codon/methionine cleavage/protein processing) JOSEF T. PRCHAL*t, DANIEL P. CASHMANt, AND Y. W. KANt tHoward Hughes Institute and Department of Medicine, University of California, San Francisco, CA 94143; and *Hemoglobin and Red Cell Laboratory and Division of Hematology, University of Alabama at Birmingham, Birmingham, AL 35294 Communicated by Helen M. Ranney, August 26, 1985 ABSTRACT Hemoglobin Long Island has two separate trout cells (11). It is removed from the nascent polypeptide amino acid abnormalities of /3-globin structure: an extension of chain by posttranslational modification in most nonsecreted the NH2 terminus by a methionine residue and a histidine-to- prokaryotic and eukaryotic proteins (12). Exceptions to this proline substitution at the normal second position. The NH2- rule are found in various prokaryotic and eukaryotic proteins terminal methionine residue, the translation product of an (13) and, also perhaps relevant to this report, in some AUG initiation codon, is present only transiently in nascent ruminant p-globins such as bovine f3- and A-related y-globin proteins. Because of the general biological implications of this (14, 15) as well as in other ruminants-i.e., sheep and goat abnormality, we investigated the nature of the genetic defect of (16). this mutant. We determined the sequence of the relevant portion of the f8-globin mRNA by means of dideoxynucleotide MATERIALS AND METHODS chain termination ofthe complementary DNA (cDNA) in which an oligonucleotide complementary to codons 10-17 was used as Subjects and Reticulocyte RNA Preparation. The mutant a primer for reverse transcriptase. A histidine-to-proline hemoglobin was discovered because ofits interference with the substitution was confirmed in the mutant mRNA by identifying detection of glycosylated hemoglobin in the propositus who is an adenine-to-cytosine transversion in the second codon. How- a diabetic, and the same mutant was also subsequently found in ever, we were unable to find any other abnormality at either the his nondiabetic mother. No apparent deleterious effect was AUG initiation codon or in the 56 bases upstream from the associated with the Hb Long Island in either the propositus or adenine-to-cytosine transversion (encompassing most of the 5' his mother. Heparinized blood from the propositus and his untranslated region of the mutant (3-globin mRNA). Thus, it mother (both heterozygous for Hb Long Island) and from a appears that this single lesion probably interferes with the control subject undergoing an exchange transfusion for pro- poorly understood methionine-cleaving mechanism that mod- longed sickle-cell crisis served as the source of globin mRNA. ulates most of prokaryotic and eukaryotic proteins. Erythrocytes were selectively lysed by exposure to 0.1444 M NH4Cl/0.001 M NH4HCO3 solution, and crude RNA was We have recently described a hemoglobin mutant, hemoglo- precipitated by acid titration. RNA was further purified by bin Long Island (1), that is characterized by two separate phenol/chloroform, 50:50 (vol/vol) extraction and ethanol pre- abnormalities in the primary amino acid structure of the cipitation (17). Neither ofthe two subjects with Hb Long Island P3-globin chain: (i) an extension of the NH2 terminus by a had evidence of shortened erythrocyte survival; thus, only 30 methionine residue and (ii) the substitution of a proline for a tkg (mother) and 34 tkg (propositus) of total reticulocyte RNA histidine at the second amino acid position. A similar mutant were obtained from each 100 ml peripheral blood. has also been described in France as hemoglobin Marseille Oligonucleotide Primer Synthesis. The deoxyoligonucleo- (2). Although more than 400 hemoglobin mutants have been tide primer was 24 nucleotides long and was complementary described (3), this P-globin is ofinterest for two reasons: first, to P-globin codons 10-17. It was constructed on a DNA the extension of the NH2 terminus and, second, of more synthesizer (Applied Biosystems, model 380A, Foster City, general biological significance, the fact that it is a mutant CA) utilizing the phosporamidite (18, 19) method with con- nonsecreted vertebrate protein in which the NH2-terminal trolled-pore-glass solid-phase support and was further puri- methionine, coded by the AUG initiation codon, is preserved fied on an HPLC (IBM, model LC 9533, San Jose, CA) in the protein structure. equipped with a C4 reverse-phase column. Its composition To establish the molecular lesion responsible for this was as follows: dCACCTTGCCCCACAGGGCAGTAAC (5' mutation, we have directly sequenced mRNA from the to 3' orientation). normal and mutant P-globin mRNA of two subjects cDNA Synthesis. The following conditions were established heterozygous for Hb Long Island. The knowledge of the to be optimal for full-length transcription of p-globin cDNA. exact nature of the mutation associated with Hb Long Island 360 pmol of oligonucleotide was mixed with 5 ,ug of total is of interest since the protein abnormality could be caused reticulocyte RNA, heated for 3 min at 98°C, rapidly cooled, by: (i) a general abnormality ofthe poorly defined methionine suspended in annealing buffer [0.4 M NaCl, Pipes (40 mM, cleavage pathway, (ii) two separate mutations, or (iii) a single pH 6.4), 80% (vol/vol) formamide] for 3 hr at 300C, and RNA mutation that interferes with the cleavage of the initial NH2 was ethanol precipitated. Each reaction mixture consisted of terminus. This NH2-terminal methionine has been found to RNA (0.2 ,ug in 2-,u aliquots) suspended in a solution of be transiently present in nascent bacterial polypeptides (4), RNasin (0.2 unit/ml; Promega Biotec, Madison, WI), acti- human tumor cells in ascites fluid (5), globins (6-10), and nomycin D, and reverse transcriptase buffer (see below). This RNA solution was incubated in a total volume of 5 ,ul containing 800 units of avian myeloblastosis virus reverse The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 24 Biochemistry: Prchal et al. Proc. Natl. Acad. Sci. USA 83 (1986) 25 transcriptase per ml, Tris HCl (100 mM, pH 8.3), 10 mM propositus heterozygous for Hb Long Island, one with CAC MgCl2, 50 mM KCl, 10 mM dithiothreitol, actinomycin D (40 coding for histidine in the second B-globin codon position and jug/ml), and various amounts of deoxynucleotides with the other with a CCC codon corresponding to proline. Also in approximately 1% of the dCTP labeled with 32p. this gel the adenine-to-thymine substitution at codon 6 resulting ,t-Globin mRNA Sequencing. Nucleotide sequences of the in a glutamine-to-valine substitution (sickle-cell mutation) is desired 13-globin anticodon region of interest in the mutant seen clearly in the control reticulocyte RNA obtained from a and control sickle-cell samples were determined by the patient with sickle-cell disease (Fig. 1). No other abnormality dideoxy method (20). The following concentrations were was seen in this analysis. Results ofthe 12% and 8% acrylamide found to be optimal for the sequencing reactions: gels confirmed the adenine-to-cytosine transversion. Separate deoxynucleotides (dCTP, 1.1 ,uM; dATP, dGTP, and dTTP, sequencing of RNA from the propositus' mother revealed the 26 AuM) and dideoxynucleotides (ddCTP, 0.4 ,uM; ddATP, 40 identical abnormality to that of the propositus. ,M; ddGTP, 8 ,uM; ddTTP, 20 ,uM). After 60 min at 37°C, each reaction was stopped by adding 4 ,ul of a solution containing 0.03% xylene cyanol FF, 0.03% bromphenol blue, DISCUSSION EDTA (10 mM, pH 8.3), 90%o (vol/vol) formamide and heated To our knowledge this is the first described naturally occur- at 65°C for 15 min. These mixtures were then rapidly cooled ring protein mutant characterized by NH2-terminal methio- and applied in 3-,ul aliquots to acrylamide gels containing 7 M nine extension. The analysis of the mRNA sequence of the urea. To increase the resolution of the various mRNA /3-globin mutant, Hb Long Island, confirmed the histidine- regions, 8%, 12%, and 20% acrylamide gels were used and to-proline substitution (1) resulting from an adenine-to- analyzed by autoradiography. cytosine transversion. The RNA sequence immediately ad- jacent to the AUG initiation codon was normal, as was the RESULTS more upstream region of the untranslated portion of 83-globin mRNA (21). Although the mutant protein consisted of two An essentially full-length cDNA transcript (better than 95% separate amino acid abnormalities (1), no other nucleic acid complete) was obtained at concentrations of deoxynucleotides mutation was found. greater than 50 ,uM. This transcript of the translated and 5' We have determined the nature of the mutation leading to untranslated regions of 8-globin mRNA consisted of 107 bases two separate amino acid abnormalities in the P-globin mutant (data not shown). Even when concentrations of deoxynucleo- of Hb Long Island by direct, rapid sequencing ofthe mRNA. tides optimal for sequencing were used, most of the in vitro This method bypasses the customary, time-consuming DNA synthesized cDNA was completely transcribed. Sequence anal- cloning and sequencing methods. RNA sequencing of homo- ysis of the reticulocyte mRNA derived from the carriers of Hb geneous viral RNA by the dideoxynucleotide termination Long Island showed that both the CAC and CCC sequences method has been accomplished (23). Eukaryotic mRNA were present at codon number two (Fig. 1). This confirmed the [poly(A)-rich RNA] has also been sequenced by this method histidine-to-proline substitution (1) in this position (Fig.