Isolation and partial characterization of the human αA-crystallin gene

There are several reasons for stressing the importance of isolating ~-crystallin genes from different species. From an evolut ionary viewpoint, ~-crystall in is highly conserved (see de Jong, 1981, 1982), making small changes in pr imary s t ruc ture very interesting. Moreover, the unexpected relationship between ~-crystallin and Drosophila heat-shock proteins (Ingolia and Craig, 1982) is intr iguing and requires detai led examinat ion (see Wistow, 1985). From a developmenta l viewpoint, a~-crystallin is the first crystall in to appear in some, b u t n o t all, species (see Piat igorsky, 1981). I t is also interesting tha t the two ~-crystallin genes (~A and ~B) are differentially expressed dur ing lens cell differentiation (Delcour and Papacons tan t inou , 1974; Vermorken and Bloemendal , 1978). Another possible benefit from compara t ive studies of a-crystal l in genes is to gain a bet ter unders tanding of the molecular and biological basis for the selective appearance of the aA ins polypept ide in certain rodents (Cohen, Westerhuis, de Jong and Bloemendal , 1978); the aA ins m R N A is a product of a l ternat ive R N A splicing of the ~A-crystall in gene (King and Piat igorsky, 1983). Finally, the numerous posttranslat ional modifications associated with a-drystall in (e.g. aggregation, cleavage and racemization) dur ing lens aging and ca tarac t give special importance to the isolation of human ~-crystallin genes (see Spector, 1973; Zigler and Goosey, 1981 ; Hoenders and Bloemendal, 1983; Hard ing and Crabbe, 1984). At present, genomic clones of the aA-crystal l in gene have been isolated from the mouse (King and Piat igorsky, 1983), hamster (van den Heuval , Hendricks, Quax and Bloemendal, 1985) and chicken (Yasuda, Okazaki, Kondoh , Shimura and Okada, 1983; Thompson, Hawkins and Piat igorsky, in preparation), while a genomic clone of the aB-erystal l in gene has been obtained only from the hamster (Quax-Jeuken, Quax, van F~ens, K h a n and Bloemendal , !985). Here we have used a murine aA-crystal l in cDNA probe to isolate the h u m a n ~A-erystall in gene. Two h u m a n genomic libraries (partial Mbo I digests o f p lacenta and spleen DNA, respectively, cloned wi thou t linkers into bacter iophage A Charon 28; gift of Dr Philip Leder) were screened with pM~ACr2, a murine ~A2-crystallin cDNA (King, Shinohara and Piatigorsky, 1982). The cDNA insert was isolated electrophoret ical ly from pM~ACr2 after digestion with Ps t I and nick-translated (Maniatis, Jeffrey and Kleid, 1975) before use for hybridizat ion. Each library ~was screened (total 10 e bacter iophage plaques per l ibrary) by the colony hybridizat ion method of Benton and Davis (1977) using 1"5 x 10 e counts per filter. Hybr idizat ion (12 hr) was performed a t 68 °C in 2 x SSC (standard saline ci t rate ) and Denhard t ' s solution (1966). The washes for the first screening were a t 63 °C with 2 × SSC, while those for the second and t h i r d screenings were a t 68 °C with 2 x SSC; all washes contained 0-5 ~/o SDS. Ten positive clones were obtained after the three screenings (six from the spleen DNA and four from the p lacenta DNA). Considering t ha t t h e D N A l ibrary contains overlapping inserts, this number is consistent with the presence of very few, if no t a single, aA-crystal l in gene. R e s t r i c t i o n analysis of the 10 genomic clones indicated tha t there were five recombinant bacteriophage containing different inserts (data n o t shown) . These


Kb
Agarose gel electrophoresis of DNA from three recombinant bacteriophage clones: ghH~Cr77 (77), ghHaCr37 (37) and gAHaCr28 (28). Barn HI and Barn HI (B)/Pvu I double digestions, as noted. DNA size standards include A DNA digested with Hind III and ~X-I74 DNA digested with Hae III. White arrowheads indicate fragments hybridizing with the aA-crystallin eDNA probe. Electrophoresis was in 0"9 % agarose in 40 mM Tris-HCl, pH 8"2, 2 mM EDTA, 20 mM Na acetate and 18 m.~ NaCI; the gel was stained with ethidium bromide. Righl panel. Southern blot hybridization of the same gel ~th the a2P-nick-translated murine ~A-crystallin eDNA probe. clones were examined further (gAHaCr37 from the placenta DNA library, and g~HaCr28 and gAHaCr77 from the spleen DNA library). Southern blot analysis (Southern, 1975) showed that the sequences which hybridized to the ~A-crystallin eDNA are situated on a 3"5 kb Bam HI fragment of ghHaCr77 and on a 4"5 Kb Barn HI/Pvu I fragment of gAHaCr37 and gAHaCr28 (Fig. 1). Digestion of g),H~Cr37 and g~HaCr28 with Barn HI alone gave a 12 Kb band which hybridized to the nicktranslated eDNA (Fig. 1). The relatively large size of this fragment was due to the fact that only one of the end Mbo I sites of the genomic insert was converted to a Barn HI site during the cloning procedure. Thus, only approximately 3"3 Kb of each 12 Kb fragment are genomic in origin, the remainder being derived from the bacteriophage arm. Further restriction analysis revealed that gAHaCr28 and gAHaCr37 have at least 11 Kb 5' to the 3-3 Kb Barn HI/Pvu I fragment containing the ~A-crystallin sequences, while gAHczCr77 has about 12 Kb 3' to the 3-5 Kb Barn HI fragment containing the c~A-erystallin sequences (data not shown).
We next identified the aA-prystallin gene within gAH~Cr77 phage by sequencing. This phage was digested with Alu I and the resulting fragments were cloned into a Sma I-cut MI3 mp8 vector (Messing and Vieira, 1982). Positive subclones were identified by hybridization with pMaACr2 and sequenced by the dideoxy-chain termination method (Sanger, Nicklen and Coulson, 1977), as described by Biggen, Gibson and Hong (1983). The results established that the 3-5 Kb Barn HI fragment in gAHczCr77 contains the human aA-crystallin gene. The 189 nucleotides given on Flo. 2. Deduced partial structure of the human aA-crystallin gene in ghHaCr77. The protein coding sequence in exon I and for the C-terminal end of the a~A-crystallin polypeptide are shown. These sequences were derived from a single strand of DNA. The nucleotide sequence for the C-terminus of the protein is assumed to be in exon 3 by'analogy with the ~A-crystallin gene from other species (see text). The thr-ala difference between the mouse and human ~/-crystallin polypeptide at position 13 is boxed. The 25-met oligonucleotidc used in Fig. 4 is indicated. The numbers above the sequence refer to amino-acid residue number in the protein. See text for further discussion. the left side of Fig. 2 encode the first 63 amino acids of the human aA-erystallin polypeptide (de Jong, Terwindt and. Bloemendal, 1975). By analogy with the known structure of the aA-crystallin gene of the mouse (Kir.g and Piatigorsky, 1983) and hamster (van den Heuval et al., 1985), we assume that this human coding sequence is located on exon 1. Experiments to be presented elsewhere support this assumption.
Interestingly, the 3' splice site for exon 1 in the murine and human aA-crystallin gene occurs after the codon for amino acid 63 (glu) (Fig. 2). There is evidence that this exon has resulted from duplication and fusion of an ancestral sequence (Barker, Ketchan and Dayhoff, 1978;Wistow, 1985;van den Heuval et al., 1985). As noted earlier by protein sequencing (de Jong, 1982;de Jong and Goodman, 1982), there is only a single amino acid difference (thr in the human and ala in the mouse at, position 13) of the encoded aA-crystallin protein in exon 1 (boxed in Fig. 2). Position 13 is occupied by thr also in the tapir, rhinoceros, ox and rabbit, by pro in the lemur, galago and potto, and by ser in the dogfish, indicating that this position is not invariant (de Jong, 1982).
• Figure 3 compares the nucleotide st~quence of the coding region of e~n 1 of the human and routine aA-crystallin gene. There are only 12 nucleotide differences between these sequences (6"3 ~/o change) despite tile 70-80 million years of divergence between mouse and man (de Jong, 1982). All but one of these occur at the third position of the codons and do not result in an amino acid change. The ala -~ thr change noted above is due to a G -~ A change at position one of, eodon 13 (boxed in Fig. 3). The 3' end of the coding region of the ~A-crystallin gene was found in another Alu I fragment derived from gAH~Cr77 (Fig. 2, right side). Again, by analogy with the  The 25-mer oligonucleotide probe indicated in Fig. 2 was synthesized by OCS Labs, Denton, Texas. It was labeled to a specific activity of about 10 ~ cpmflg -~ DNA by hybridization with a 15-mer complement followed by extension with Klenow fragment of DNA polymerase using four a-a2P-dNTPs. Hybridization procedures are given by Miyada, Klofelt, Reyes, McLaughlin-Tayior and Wallace (1985). This probe was hybridized with a dried denatured, neutralized and rebydrated 0-9 % agarose gel using 2"0 x 106 cpm ml -l. Hybridization was carried out at 68 °C for 3 hr followed by two 15-rain washes at 25 °C, one 2-hr wash at 25 °C and one 90-sec wash at 68 °C. All washes were in 6 x SSC (0.9 M NaCI, 90 m,,~ sodium citrate, pH 7.2). Autoradiography was with Kodak XAR-5 film for 15 hr at --70 °C with an intensifying screen. known structure of the hamster aA-crystallin gene (van den Heuval et al., 1985) we assume that the sequences encoding the C-terminu's of the protein are located on exon 3. Tile numbering of these exons depends on the assumption that the insert exon present in the mouse (King and Piatigorsky, 1983) and hamster (called exon 2 in this species, van den Heuval et al., 1985) is absent in the human.
Finally, we performed a Southern blot on Barn HI-digested DNA from tissue samples of five individuals in order to provide further evidence that there is only a single aA-crystallin gene in humans. Hybridization was performed with a radioactively labeled 25-mer oligonucleotide probe encoding amino acids 10 to 17 of the aA-crystallin polypeptide (indicated in Fig. 2) on the dried agarose gel. The 25-mer probe has minimum and maximum mismatches with the human gB-crystallin gene of 32 % and 64 %, respectively, as calculated from the known sequence of the human ~B protein (Kramps, de Man and de Jong, 1977) and the ambiguity of the genetic code. The oligomeric probe hybridized exclusively to a 3.5 Kb Barn HI DNA fragment from each individual (Fig. 4). In another test, this 3"5 Kb Barn HI fragment co-migrated with the 3"5 Kb insert in gAHgCr77 which contains the ~A-crystallin gene. This result is consistent with the interpretation that there is a single gA-crystallin gene in the human and that this gene is present in gAH~Cr77. Additional experiments using pMgACr2 (the cDNA probe) revealed the presence of this hybridizing 3"5 Kb Barn HI DNA fragment in 16 individuals. Thus, it appears as if the human ~A-crystallin gene is not highly polymorphic.
In summary, we have isolated the gA-crystallin gene in humans by using a nearly full-length cDNA probe from mice. There appears to be only one gA-crystallin gene in man as in mouse Piatigorsky, 1983), hamster (van den Heuval et al., 1985) and chicken (Thompson et al., in preparation). A recent investigation using somatic cell hybrids indicates that this gene is located on human chromosome 21 (Quax-Jeuken et al., 1985). Our data indicate that the size and structure of the human ~A-crystallin gene are generally similar to that in the other organisms examined. The first intron of the human gene is situated after codon 63 as it is in the mouse (King and Piatigorsky, 1983), hamster (van den Heuval et al., 1985) and chicken (Yasuda et al., 1983;Thompson et al., in preparation). We presume that the human gA-crystallin gene lacks an insert exon (King and Piatigorsky, 1983) and that codons 64-104 are contained on a separate exon as in the other organisms, but this remains to be established. The present isolation of the human gA-crystallin gene extends our ability to conduct comparative studies on this highly conserved lens protein, which should deepen our understanding of its selective expression (Chepelinsky, King, Zelenka and Piatigorsky, 1985;Okazaki, Yasuda, Kondoh and Okada, 1985;Overbeek, Chepelinsky, Khillan, Piatigorsky and Westphal, 1985) and alternative RNA splicing (King and Piatigorsky, 1983). The availability of the human gA-crystallin gene should also facilitate studies concerning the possible involvement of~A-crystallin in cataract.