Cellular and Molecular Neurobiology, Vol. I1, No. 1, 1991

Human Acetylcholinesterase and Butyrylcholinesterase Are Encoded by Two Distinct Genes Averell Gnatt, 1 Dalia Ginzberg, 1 Judy Lieman-Hurwitz, ~ Ronit Zamir, 2 Haim Zakut, 2 and Hermona Soreq t'3 Received January I, 1990; accepted March 2, 1990 KEY WORDS: butyrylcholinesterase; acetyicholinesterase; gene mapping; cosmid; recombination; chromosome blots.

SUMMARY

1. Various hybridization approaches were employed to investigate structural and chromosomal interrelationships between the human cholinesterase genes CHE and ACHE encoding the polymorphic, closely related, and coordinately regulated enzymes having butyrylcholinesterase (BuChE) and acetylcholinesterase (ACHE) activities. 2. Homologous cosmid recombination with a 190-base pair 5' fragment from BuChEcDNA resulted in the isolation of four overlapping cosmid clones, apparently derived from a single gene with several intrans. The Cosmid CHEDNA included a 700-base pair fragment known to be expressed at the 3' end of BuChEcDNA from nervous system tumors and which has been mapped by in situ hybridization to the unique 3q26-ter position. In contrast, cosmid CHEDNA did not hybridize with full-length AChEcDNA, proving that the complete CHE gene does not include AChE-encoding sequences either in exons or in its intrans. 3. The chromosomal origin of BuChE-coding sequences was further examined by two unrelated gene mapping approaches. Filter hybridization with DNA from human/hamster hybrid cell lines revealed BuChEcDNA-hybridizing sequences only in cell lines including human chromosome 3. However, three BuChEcDNA-homologous sequences were observed at chromosomal positions 1 Department of Biological Chemistry, The Life Sciences Institute, The Hebrew University, Jerusalem 91904. 2 Department of Obstetrics and Gynecology, The Edith Wolfson Medical Center, The Sackler Faculty of Medicine, Tel Aviv University, Israel. 3 To whom correspondence should be addressed. 91 0272-4340/91/0200-0091506.50/0© 1991PlenumPublishingCorporation

92

Gnatt et al.

3q21, 3q26-ter, and 16q21 by a highly stringent in situ hybridization protocol, including washes at high temperature and low salt. 4. These findings stress the selectivity of cosmid recombination and chromosome blots, raise the possibility of individual differences in BuChEcDNAhybridizing sequences, and present an example for a family of highly similar proteins encoded by distinct, nonhomologous genes.

INTRODUCTION

The enzyme family of cholinesterases (ChEs) in humans has been the subject of intensive research for many years, with a continuous increase in the number of studies being focused on the genes encoding these enzymes (for recent relevant reviews see Rakonczay and Brimijoin, 1988; Chatonnet and Lockridge, 1989). ChEs are biochemically classified into acetylcholinesterase (ACHE; acetylcholine acetylhydrolase, EC 3.1.1.7) and butyrylcholinesterase (BuChE; acylcholine acylhydrolase, EC3.1.1.8) (Toutant and Massoulie, 1987; Taylor et al., 1987). Molecular cloning studies revealed the complete nucleotide sequences encoding human BuChE (Prody et al., 1986, 1987; McTiernan et al., 1987) and AChE (Soreq and Prody, 1989; Soreq et al., 1990), making homo sapiens the first species in which both ChE coding sequences have been cloned. This study demonstrated, quite unexpectedly, that the genomic sequences encoding AChE and BuChE in humans do not display a high sequence homology (Lapidot-Lifson et al., 1989), in spite of the considerable similarity between the protein sequences encoded by these two genes. The lack of sequence homology between AChEcDNA and BuChEcDNA proved beyond doubt that, in humans, AChE and BuChE are produced from two distinct mRNA transcripts. However, it could not answer the question whether these two transcripts are produced from two independent genes or from one, perhaps by complex posttranscriptional processing. This question became particularly intriguing in view of the findings that the CHE and the ACHE genes encoding BuChE and ACHE, respectively, are coamplified in leukemias (LapidotLifson et al., 1989) and ovarian carcinomas (Zakut et al., 1990). To determine whether the BuChE and AChE coding regions are included in a single complex gene, one would have to isolate the complete DNA sequence encompassing the regions coding for one of these enzymes and determine whether intron sequences included in this DNA contain the coding regions for the other. Here, we demonstrate cosmid recombination experiments aimed to select CHEDNA cosmid clones containing the complete BuChE coding region and their use for such analysis. Genetic linkage evidence accumulated over the years suggests that the CHE gene, encoding BuChE, resides on the long arm of chromosome 3 (Sparkes et al., 1984; Arias et al., 1985; Soreq and Zakut, 1990). Significant although considerably weaker linkage with genes on the long arm of chromosome 16 (Lovrien et al., 1978) was also observed. In situ hybridization with spread mitotic chromosomes revealed sequences complementary to BuChEcDNA on three sites,

Distinct ChE and AChE Genesin Humans

93

designated CHEL1, CHEL2, and CHEL3, at the respective positions 3q21, 3q26, and 16q21 (Soreq et al., 1987; Zakut et al., 1989). Recently, the isolation and mapping of 3'-extended BuChEcDNA clones from brain tumor libraries demonstrated that the 3' additional fragment on these clones, which was designated Gb5, originates from the unique chromosome 3q26-ter position (Gnatt et al., 1990). We therefore employed this unique Gb5 probe to assess further the chromosomal origin of the CHEDNA cosmids, used chromosomal blots from human-hamster hybrid cells to confirm our assessment, and performed in situ hybridization under stringent washing conditions to compare the hybridization results in blots to those obtained with chromosomes. Our findings primarily support the notion of a single CHE gene, localized on chromosome 3q26-ter, which does not include AChE coding sequences, and indicate the possibility of differences between individuals in the content of BuChEcDNA hybridizing sequences.

METHODS D N A Probes and Oligonucleotides

Three cDNA probes from the CHE gene were employed. These were a 2.4-kb human BuChEcDNA encoding the complete BuChE protein (Prody et al., 1987), a genomic E c o R I fragment containing 190 base pairs from the 5' end of this sequence (Gnatt and Soreq, 1987), and a 0.7-kb 3' extension of this cDNA that is expressed in brain tumors (Gnatt et al., 1990). Oligonucleotide probes synthesized according to the BuChEcDNA sequences were as described (Prody et al., 1986; Soreq and Gnatt, 1987). To search for AChE coding regions, a 1.5-kb human AChEcDNA fragment encoding 90% of the AChE protein was employed (Soreq and Prody, 1989; Lapidot-Lifson et al., 1989). Cosmid Recombination

Human cosmid libraries, constructed from male peripheral blood DNA in the pcos2 vector, providing kanamycin resistance, and cloned in DH1 E. coli cells, were generously provided by H. Lehrach (ICRF, London), as well as the BHB3169 bacterial strain. Other bacterial strains employed were AG1, mvll90, and BHB3175. Cosmid recombination was performed according to detailed published procedures (Poustka et al., 1984) with several modifications. Briefly, the 190-bp 5'-BuChEcDNA fragment was inserted into the pucll8 ampicillinresistant plasmid (Stratagene), which does not display any sequence homology with the pcos2 cosmid vector, and transfected into BHB3175 bacteria. In order to rescue the cosmid library in the form of lambda bacteriophages, lambda 3196 rescue phages were prepared by prophage induction at 42°C from concentrated prophage-carrying BHB3169 bacteria grown at 30°C. Rescue phages at an m.o.i. (multiplicity of infection) of 50 were added to DH1 cells carrying the cosmid library and packaging was induced at 42°C. Approximately 109 cfu of the packaged phages were employed to infect 107BHB3175 cells containing the probe plasmid (carrying the 5' 190-bp

94

Gnatt et al.

BuChEcDNA fragment). Natural recombination was allowed to occur b e t w e e n the probe BuChEcDNA fragment in the plasmid and the genomic CHEDNA sequences in the cosmids. Recombinants stable to both kanamycin and ampicillin, which implies the presence of recombined pcos2 and pucll8 sequences, were selected for on dual antibiotic plates. In v i v o packaging performed on these recombinants resulted in the formation of cosmid-phage recombinants in transfected AG1 cells. These were tested by colony hybridization with [32p]. BuChEcDNA probe. Large-scale preparation of cosmid DNAs was then performed by the alkaline plasmid method (Maniatis et al., 1982).

DNA Blot Hybridizations Cosmid or genomic DNA was digested to completion by restriction endonucleases, separated by agarose gel electrophoresis, blotted under 0.5 N NaOH onto nylon Genescreen membranes (N.E.N), and UV cross-linked. Hybridization was in 35% formamide, 6× SSC, and l x Denhardt's. Where nitrocellulose filters were employed, they were denatured, neutralized, baked in vacuo for 2hr at 80°C, and hybridized in 6x SSC, 5x Denhardt's solution, and 5 mM EDTA.

Gene Mapping Experiments Chromosomal blots were donated by Bios corporation (New Haven, CT) and included DNA from 27 different hybrid human/hamster cell lines carrying hamster chromosomes as well as one or more human chromosomes or fragments thereof, all restricted with H i n d l I I . Hybridization was performed as recommended in the Bios Timeframe System using the Large Bios Hybridization Cassette (Model TF-2500). Experiments were performed at 65°C in an incubator with 2.4-kb BuChEcDNA at 10ng/ml. Probe DNA was multiprime labeled (Feinberg and Vogelstein, 1983), yielding a specific activity of 6 x 108cmp//~g. Hybridization was for 60-90 min. Washes were at 65°C in 0. l x SSC. In situ hybridization to spread mitotic chromosomes was performed essentially as described elsewhere (Soreq et al., 1987; Zakut et al., 1989) except that an additional wash was performed for 2 x 15 min at 60°C in 0.1 x SSC. Following the hybridization, chromosomes were Giemsa stained and covered by a thin Optilux (0.75%, Falcon) layer prior to emulsion autoradiagraphy, to prevent interaction between the photographic emulsion and the Giemsa stain.

RESULTS The presence of both BuChE and AChE coding sequences in the cosmid library was first verified by DNA blot hybridization. For this purpose, the cosmid library was amplified and total DNA extracted from it was digested to completion with E c o R I and electrophoretically separated in parallel to human genomic DNA. Blot hybridization with BuChEcDNA and AChEcDNA probes confirmed that both sequences are represented in this library (not shown). Homologous recombination screening, using a 5' 190-bp long fragment from BuChEcDNA

Distinct ChE and AChE Genes in Humans

95

cloned in puc118, resulted in the isolation of four different cosmid pCosCHEDNA clones designated C1-4. Figure la schematically presents the recombination experiment. DNA extracted from each of the pCosCHEDNA clones was tested for its ability to hybridize with BuChEcDNA and AChEcDNA. All four clones were clearly positive with the first probe but negative with the latter, demonstrating that they all included BuChE coding sequences but not AChE coding sequences (Fig. lb). Enzymatic restriction, agarose gel electrophoresis, and ethidium bromide staining of these pCosCHEDNA preparations revealed that the four isolated cosmid DNA fregments w e r e atl approximately 50kb in size and displayed highly similar although not identical restriction patterns with four different enzymes. Blot hybridization with BuChEcDNA again demonstrated similar restriction patterns, suggesting that all four cosmid clones were derived from the same gene and encompassed overlapping DNA fragments (Fig. 2). All four cosmid clones originally had to include the 5' 190-bp sequence to recombine with the puc118 plasmids which contributed their ampicillin resistance. Moreover, they all hybridized with the 0.7-kb 3' extension of BuChEcDNA, designated Gb5, that we recently found in cDNA libraries from nervous system tumors (Gnatt et al., 1990; not shown). This demonstrated that all four cosmids included the entire sequence encoding BuChE plus adjacent sequences. In addition, the cosmid DNAs appeared to contain several intron sequences. For example, the enzyme HincII, having a unique restriction site in BuChEcDNA (Prody et al., 1987; McTiernan et al., 1987), restricted the various cosmid DNAs into at least three BuChEcDNA hybridizing fragments (Fig. 2). This suggested the existence of more than one intron in these DNA fragments, in complete agreement with observations of others (Chatonnet and Lockridge, 1989). It should be mentioned in this respect that the 190-bp fragment that was used to recombine with the CHEDNA cosmids had to be actively integrated into the isolated cloned sequences. Therefore, the organization of these DNA fragments should not necessarily be identical to that of the native gene and these blots were hence not used to construct detailed restriction maps for the CHE gene. Since the Gb5 3'-extended fragment was mapped to the unique 3q26-ter position by in situ hybridization, these findings also indicated that all of the isolated pCosCHEDNA clones were derived from a single BuChE encoding gene, localized on the long arm of chromosome 3. The assignment of the CHE1 gene to a unique position on chromosome 3 was further confirmed by DNA hybridization employing chromosome blots (Bios. Inc.). These included DNA from different human/hamster hybrid cell lines carrying one or more human chromosomes (D'Eustachio and Ruddle, 1983). Only DNA from cell lines including human chromosome 3 contained the informative 2.4-kb H i n d I I I restriction fragment hybridizing with BuChEcDNA (Prody et al., 1989; McGuire et al., 1989). Furthermore, the pCosCHEDNA fragments were identical to those observed in parallel lanes loaded with DNA from human lymphocytes. In contrast, hamster genomic DNA or hamster/human cell hybrids carrying human chromosomes other than No. 3 displayed no positive bands when hybridized with [3zP]BuChEcDNA. Figure 3 presents representative

96

Gnatt et al. 1

a

2

3

AMP PUO118 1 " ~ ~ AM

Kn~

b

32P-AChEcDNA 32P-BuChEcDNA eDNA BuChE

C1

C2

C3

C4

DNA eDNA

Herring

AChE

PCosChEDNA

Fig. 1. Recombinant screening and isolation of CHE cosmid clones. (a) A 190-bp fragment from the 5' domain of BuChEcDNA (Gnatt and Soreq, 1987) was inserted into the ampicillin (AMP)resistant vector puc 118 (PCHE) and transfected into the rec + E. Coli strain BHB3169, which permits natural recombination. PCHEcarrying bacteria were in turn infected by the genomic kanamycinresistant cosmid phage library and recombination between PCHE and CHE cosmid clones (CosCHE) took place. Dual antibioticresistant recombinant clones (PCosCHE) were purified for analysis. (b) One microgram of DNA from four such clones (PCosChEDNA clones C1-CA) and 60 ng of either BuCHE or AChEcDNA were blotted onto nitrocellulose filters with herring DNA used as a negative control. Probes employed were labeled with 32p as detailed under Methods. Note that all cosmid clone DNAs hybridized with BuChEcDNA but not with AChEcDNA.

97

Distinct ChE and AChE Genes in Humans 32

BKSHB

P-CHE

BBH KBKSHBBBH H KSS H HKSS

Cosmid t

K K S H

Cosmid 2

BKSHBBBHKKBKSHBBBHKK HKSSSH

Cosmid 3

HKSSSH Cosmid 4

Fig. 2. Blot hybridization of human cosmid CHEDNAs. DNA from cosmid clones was digested to completion by one or more of the restriction enzymes SacI (S), HincII (H), BamHI (B), and KpnI (K) and electrophosesed in a 0.8% TBE agarose gel (TBE = 0.0089M Tris-borate, 0.089M boric acid, and 0.002M EDTA). [32P]BuChEcDNA probe was employed in hybridization. Size markers were lambda HindIII- and PhiX HaeIII-digested fragments; the migration positions of some are illustrated. Highly similar restriction patterns can be observed in all of the cosmid DNA preparations in both single and double enzymatic restrictions. Note that KH double restrictions reveal a closer restriction pattern between Cosmid t and Cosmid 3 as opposed to Cosmid 2 and Cosmid 4.

examples for these hybridization results and Table I summarizes the somatic-cell hybrid mapping panel, clearly demonstrating full concordance with the chromosome 3 assignment in all of the five cell lines examined which carded this chromosome. In contrast, all other chromosomes, including chromosome 16, appeared not to carry BuChEcDNA positive sequences. Exclusion of chromosome 16 was shown by two independently derived discordant hybrids, and one hybrid cell line with chromosome 3 as the only human chromosome was clearly positive for the human specific fragment. Cosmid recombination is an efficient and highly selective process, which takes place only with fully complementary sequences (Poustka et al., 1984). The

98

Gnatt et al.

Er,. "1" ¢o

:Cell line ,i¢

'I-

Fig. 3. Human/hamster chromosomal blot hybridization. DNA from human/hamster somatic-cell hybrids containing one or

more human chromosomeon a backgroundof hamster chromosomes was enzymaticallyrestricted with HindIII. Blots were prepared by Bios. Corp. and hybridizationwas performed with [32p]-BuChEcDNA. For chromosomalcontents see Table I. A 2.3-kb HindlII fragmentcharacteristicof the human CHE gene appeared only in lanes loaded with DNA from cell hybrids containinghumanchromosome3 (423, 860) or with total human DNA, and not in lanes loadedwith hamsterDNA or with DNA from hybridcell lines that do not carry the human chromosome 3 (867,854). chromosome blots were hybridized under denaturating conditions and washed at high temperature and low salt, ensuring the stringency of the hybridization reaction. In contrast, in situ hybridization with spread chromosomes is limited in that it is performed in the presence of chromatin and washing conditions are generally rather mild (Rabin et al., 1985). To reexamine the validity of our previous in situ hybridization with BuChEcDNA, which demonstrated positive labeling of chromosome 16 in addition to 3, we therefore performed these mapping experiments under highly stringent washing conditions. Significant labeling again appeared on chromosomes 3 and 16 (Fig. 4), reconfirming the presence of chromosome 16 BuChEcDNA-like sequences at least in some individuals.

DISCUSSION The findings presented in this report imply that distinct genes encode for BuChE and AChE in man. Although they encode for highly similar proteins (>50%) (Soreq and Prody, 1989), these two genes do not display sequence homology at the level of DNA, suggesting that they diverged very early in evolution. While the CHE gene remained rather primordial in its A, T-rich base composition, the ACHE gene continued to evolve and gradually acquired the G, C-rich composition characteristic of certain genes (Holmquist, 1987). It is not yet clear whether the CHE and the ACHE genes are colocalized on the same chromosomal position; ACHE gene mapping experiments by in situ hybridization, pulse field gel electrophoresis, and RFLP analyses will be required to resolve this issue and determine whether the CHE and ACHE genes coamplify

Distinct ChE and AChE

G e n e s in H u m a n s

I

I

)

Human acetylcholinesterase and butyrylcholinesterase are encoded by two distinct genes.

1. Various hybridization approaches were employed to investigate structural and chromosomal interrelationships between the human cholinesterase genes ...
1MB Sizes 0 Downloads 0 Views