Hum Genet (1990) 85 : 151-156

�9 Springer-Verlag 1990

Identification of sequences of chromosome 7 that are expressed in sweat gland epithelial cells

Julie Burns 1'*, GeorgMeimer 1'**, Johanna M. Rommens 1, John R. Riordan 4'5, and Manuel Buchwald 1'2's

i Research Institute, The Hospital for Sick Children, ZDepartment of Medical Genetics, 3Department of Medical Biophysics, 4Department of Biochemistry, and SDepartment of Clinical Biochemistry, University of Toronto, Toronto, Ontario, Canada

Received June 30, 1989 / Revised December 20, 1989

Summary. This paper describes an approach that can be used to identify specifically expressed coding sequences in defined regions of genomic DNA. We developed this method to identify expressed sequences from chromo- some 7 located at or near the cystic fibrosis (CF) locus. Radioactively labelled single-stranded cDNAs derived from sweat gland epithelial cells and from fibroblasts were used to screen a genomic library constructed from flow-sorted chromosomes. Differential screening of phage lifts with these two probes yielded 36 different D N A seg- ments. By using somatic cell hybrids containing different portions of chromosome 7, four of the clones were map- ped to the 7@1 region in which the CF locus is located. These four clones and two others that gave strong differ- ential epithelial signals but that were not within 7@1 were studied further. Restriction fragment length poly- morphisms (RFLPs) were identified for two of the D N A segment s within 7q31 and used for linkage analysis using a panel of CF families. One D N A segment was assigned to a location centromeric to the met locus. The other marker did not show recombination with CF but was subsequently excluded from the CF region by physical mapping. Three of the six DNA segments were found to hybridize to various RNAs using the Northern technique and therefore contain portions of genes. One of the clones showed strong differential expression when epi- thelial tissues were compared to fibroblasts and may rep- resent an epithelium-specific gene.

Introduction

Cystic fibrosis (CF) is the most common autosomal ge- netic disorder in Caucasian populations (McKusick 1988).

Present addresses: * Beatson Institute, Glasgow, UK

** Middlesex Hospital, London, UK

Offprint requests to." M. Buchwald, Department of Genetics, The Hospital for Sick Children, 555 University Avenue, Toronto, On- tario M5G 1X8, Canada

The presence of two defective alleles in CF patients leads to clinical manifestations in exocrine tissues such as lung, pancreas, and sweat glands (Boat et al. 1989), and specific defects in anion transport have been described in those tissues (Quinton 1983; Knowles et al. 1983) and in cultured cells derived from them (Frizzell et al. 1986; Welsch and Liedke 1986; Pedersen et al. 1987; Schou- macher et al. 1987; Li et al. 1988). The CF gene was lo- calized to the chromosomal region 7q31 by linkage anal- ysis to D N A markers mapped to that region and by in situ hybridization of closely linked markers (reviewed in Tsui et al. 1988; Duncan et al. 1989) and subsequently isolated by molecular cloning of the 7q31 region and identification of genes located in that region (Rommens et al. 1989; Riordan et al. 1989; Kerem et al. 1989).

An alternative strategy would be to isolate genes lo- cated in 7q31, especially those expressed in tissues impli- cated in the disease, and to determine whether they were the CF gene. We describe here the method that we u s e d for such an attempt. We screened approximately one chromosome equivalent of a flow-sorted human chromo- some 7-specific genomic library with radioactively label- led reverse-transcribed mRNA from sweat gland epithe- lial cells. Several expressed sequences located in chro- mosome region 7@1 were isolated and characterized in relation to the CF gene by both physical and genetic techniques. Included is one clone that showed a differen- tial expression in sweat gland epithelial cells when com- pared to fibroblasts. This method can be used for the iso- lation of chromosome-specific or tissue-specific genes.

Materials and methods

Cell culture

Human sweat gland epithelial cells were cultured according to our previously published methods (Collie et al. 1985; Riordan et al. 1987). Epithelial ceils derived from human nasal polyps were cul- tured employing the procedures of Yankaskas et al. (1985, 1988). Human skin fibroblasts were grown in ct-MEM medium with

152

10% foetal calf serum and passaged using published procedures (Buchwald 1984).

Screening of chromosome 7 library for expressed sequences

A human chromosome 7 genomic DNA library in lambda Charon 21A (ATCC 57722) (described in Deaven et al. 1986) was used for screening. The average insert size in this library is about 4 kb and one genome equivalent is approximately 26,000 recombinant phage. Methods for the preparation of phage DNA have been described (Maniatis et al. 1982). After isolation, DNA inserts from phage were subcloned into the EcoRI site of pUC13. Aliquots of the phage library were plated on E. coli LE392 on LB plates. Repli- cate lifts were made to Biodyne A 1.2-pm nylon membranes (ICN Biomedicals) according to the manufacturer's protocol.

Total cellular RNA was isolated from cultured sweat gland epi- thelial and fibroblast cells derived from the same donor by the method of Chirgwin et al. (1979) and poly(A +) RNA selected after a single passage through an oligo (dT)-cellulose column (Aviv and Leder 1972). Poly(A +) RNAs from sweat gland epithelial cells and fibroblasts were used as templates for oligo (dT)-primed synthesis of 32p-labelled first strand cDNA using reverse transcriptase. Ap- proximately 2gg of poly(A § RNA was dried and then resus- pended in approximately 20gl of diethylpyrocarbonate-treated double-distilled water. After heating at 70 ~ for 5 min and chilling on ice, the reaction mixture was prepared by the addition of 5 I~1 of 10x RT buffer (500mM Tris, pH 8.3, 500mM KC1, 100mM MgC12), 2 gl of 2 mg/ml oligo d(T)12-18, 2 gl of 2 mM dGTP and of 2mM dTI'P, I gl of 0.2mM dATP and of 0.2mM dCTP, 10gl of 32p dATP and of 32p dCTP (3000 Ci/mmol), and 40 units of RAV-2 reverse transcriptase (Amersham). The reaction was carried out for 90-120 min at 37 ~ stopped by the addition of Tris HCI (10 mM) and EDTA ( lmM) and 0.1% sodium dodecyl sulphate (SDS) and the reaction products separated on a Sephadex G-50 column. Spe- cific activities > 8 • 107 cpm/pg poly(A +) RNA were routinely ob- tained.

Hybridizations were carried out in 6 x SSC (1 x SSC = 0.15 M NaC1, 0.015 M Na-citrate, pH 7.0), 10 mM NaPO4 buffer pH 6.5, 5 x Denhardt's solution (1 x Denhardt's = 0.02% Ficol1400, 0.02% polyvinylpyrrolidone, MW 360,000, 0.02% bovine serum albumin, Pentax Fraction V), 0.2% SDS, and 50% formamide at 42~ for approximately 6 h. When total human DNA was used to probe the filters, hybridizations were carried out overnight. After hybridiza- tion, the filters were washed to a final stringency of 65~ in 0.1 • SSC, 0.1% SDS and exposed to Kodak X-Omat film for 24-72 h as necessary.

To reduce nonspecific hybridization due to repeat sequences, competitive hybridization of cDNA probes with total human DNA was introduced for a second screening of the library. Immediately prior to hybridization, probes were denatured by boiling for 5 min in 0.2N NaOH with 1-3 gg human DNA sheared to sizes averaging 500 bp, neutralized with acetic acid and re-annealed for 10-20 rain at 60~ The labelled probe was then added directly to the hybrid- ization reaction. Clones hybridizing preferentially to sweat gland cDNA were selected and rescreened. When testing for the pres- ence of human repetitive DNA, total human DNA was labelled with 32p by random priming (Feinberg and Vogelstein 1983) and used for hybridization as described above.

Mapping of phage inserts within chromosome 7

Cellular DNAs isolated from a panel of human-hamster cell hy- brids containing various portions of chromosome 7 (Zengerling et al. 1987) were digested with EcoRI, subjected to electrophoresis on agarose gels, transferred to Zetabind membranes, and hybrid- ized to purified phage DNA inserts labelled with ~2p by random priming. Probes containing repetitive DNA were annealed with 100-250 gg of sheared total human DNA at 60~ for 20-25 rain to reduce nonspecific hybridization.

Northern blot analysis

Samples (10 gg) of poly(A +) RNA isolated from epithelial and fi- broblast cells after one round of chromatography in oligo (dT)-cel- lulose were denatured with formaldehyde, subjected to electro- phoresis on 1% agarose, transferred to Biotrans nylon membrane, and hybridized for 24h with 32p-labelled probe, competed, if necessary, with total human DNA. The hybridization mix was as described above with the addition of 20 pg/ml tRNA and 50 pg/ml polyA. Posthybridization washings were carried out under varying levels of stringency.

Identification of restriction fragment length polymorphisms (RFLPs) and linkage analysis

Genomic DNA from individuals in families with two or more CF children (Tsui et al. 1985) and from ten random individuals was digested with at least 25 restriction enzymes using protocols de- scribed by the manufacturers, subjected to electrophoresis on agarose, transferred to Zetabind membrane, and hybridized with 32p-labelled probe and competed, if necessary, with total human DNA. The linkage data was analyzed as described previously (Tsui et al. 1985) using the LIPED program (Ott 1974).

Results

Screening of the chromosome 7 library for expressed sequences

O u r init ial ob jec t ive was to ob ta in c lones to genomic D N A tha t c o d e d for sequences expres sed in sweat g land ep i the l ia l cells and tha t were loca l ized to 7@1 since these wou ld be poss ib le cand ida t e C F genes. In add i t ion , we were in te res ted in ob ta in ing genomic c lones coding for sequences tha t were specif ical ly exp res sed in sweat g land cells, even if they d id no t res ide in 7q31, because these might l ead to the iden t i f i ca t ion of r egu la to ry se- quences specific for sweat g land cells.

O u r bas ic s t ra tegy was to screen a genomic ch romo- some 7 l ibrary with s ing le -s t randed c D N A ob ta ined f rom e i the r sweat g land ep i the l ia l ceils or f ib rob las t s and to select c lones tha t hybr id ized m o r e s t rongly to the ep i the- l ial c D N A . W e r ea soned that some of these clones would inc lude genes that were p re fe ren t i a l ly exp res sed in these cells and o the r s tha t wou ld be expres sed in bo th cell types . To min imize d i f fe ren t ia l signals ar is ing f rom the gene t ic var iab i l i ty o b s e r v e d in h u m a n popu la t i ons , we used ep i the l ia l and f ib rob las t cells de r ived f rom the same ind iv idua l as the source of the m R N A used to p r e p a r e the screening p robes .

O n the first r o u n d of screening, one g e n o m e equiva- lent of the c h r o m o s o m e 7 l ib ra ry (26,000 phage p l a t e d at a dens i ty o f 2,600 p laques p e r 15 cm p la te ) was hybr id - ized to p robes p r e p a r e d f rom 32p-labelled, f i r s t -s t rand c D N A s . Clones that hybr id ized d i f ferent ia l ly with the two p robes were se lec ted . Resc reen ing o f the p i cked phage with l abe l led , s ing le-s t rand c D N A s and to ta l hu- m a n D N A ind ica ted that the ma jo r i t y of se lec ted c lones con ta ined repe t i t ive D N A . H o w e v e r , seven c lones were i so la ted tha t hybr id ized with one or bo th c D N A p robes and showed only weak hybr id i za t ion to to ta l h u m a n D N A . One c lone, 5 -21 , was found to be wi thin 7@1 when hybr id ized against a pane l of h u m a n h a m s t e r cell

153

Fig. 1A, B. Screening for differentially expressed clones. Autoradiograms of filter lifts of recombinant phage from a chromosome 7 library were plated at a density of 700 phage/plate and probed with radiolabelled single-stranded cDNA that had been competed with total human DNA prior to hybridization. A Part of a filter disk probed with sweat gland cDNA; B the replica probed with fibroblast cDNA. The arrow points to a differentially labelled clone

Fig. 3. Recombination event between DNA segment 89 and cystic fibrosis (CF). Autoradiogram of a Southern blot of genomic DNA from three families with two children affected with CF probed with DNA segment 89. Family i is noninformative, family 2 shows con- cordance of inheritance of the b allele from the father, while family 41 shows recombination of the b allele from the mother

Fig. 2. Mapping of probes to 7q31. Portions of autoradiograms of Southern blots of genomic DNA from several human-hamster hy- brid cell lines containing various parts of chromosome 7 as indi- cated above each lane. The DNA segments used are identified on the left of the autoradiograms and their location in relation to q31 is given to the right

lines containing various portions of chromosome 7 (see below).

To circumvent the effect of the repetitive sequences on the differential signals, the second round of screening was per formed with single-stranded c D N A probes that had been preincubated with total human D N A before hybridization to the filters. In this round, approximately 7,000 recombinant phage were screened at a density of 700 per 15-cm plate. After rescreening, 29 clones were isolated and mapped , and three clones, 85B, 89, and 101, were localized within 7q31 (see below). Two other clones not located within 7q31 (clones 105 and 117) but that showed strong differential signals between epithelial and fibroblast c D N A were also characterized further.

An example of a clone demonstrat ing a differential sig- nal on this round of screening is illustrated in Fig. 1.

Chromosomal localization o f clones

Clones that produced differential signals after rescreen- ing were localized to 7q31 by Southern hybridization against a panel of DNAs f rom human-rodent somatic cell hybrids containing varying port ions of human chro- mosome 7 (Zengerling et al. 1987). When necessary, hy- bridization was carried out with probes annealed to sheared and sonicated total human D N A to reduce the signals resulting f rom the repetitive sequences present in some of the clones.

Examples of four D N A segments mapped with the hybrid cell panel are shown in Fig. 2. D N A segments 85B, 89, and 101 showed a signal when hybridized to D N A from cell lines carrying the complete 7 (lane B), partial 7s missing the q22 region (lane D), a 7 containing the distal port ion of the long arm (lane G), and total human D N A . On the other hand, these clones gave no signal when hybridized to total Chinese hamster ovary (CHO) , lane A) or mouse (lane E) D N A or when hy- bridized to a cell line carrying an interstitial deletion in 7@1 (lane F). In contrast, D N A segment 105 did show a

154

Fig. 4. Restriction fragment length polymorphism (RFLP) iden- tified for DNA segment 5-21. Autoradiogram of genomic DNA from three families with two or more affected individuals with CF probed with DNA segment 5-21. DNAs were digested with the enzyme ScaI. Family 40 is noninformative for this RFLP while the other two families show concordant segregation for the a allele from the father

Fig. 5. Identification of expressed sequences by Northern blot analysis. Autoradiogram of blots prepared from poly(A § RNA isolated from fibroblasts (odd lanes') and sweat gland epithelial cells (even lanes). The location of the ribosomal RNA bands is in- dicated. Lanes 1 and 2 were probed with DNA segment 105, lanes 3 and 4 with DNA segment 5-21, lanes 5 and 6 with DNA segment 85B, and lanes 7 and 8 with DNA segment 117. The arrows point to bands of 2.5 kb (lane 2, clone 105), 6 kb (lanes 3, 4, clone 5-21) and 3kb (lanes 7, 8, clone 117). Final washes were performed at 65 ~ 0.2 x SSC, 0.1% SDS (clone 105); 63 ~ 0.2 x SSC, 0.1% SDS (clone 5-21); 65 ~ 0.1 • SSC, 0.1% SDS (clone 85B), and 60 ~ 0.1 • SSC, 0.1% SDS (clone 117)

signal when hybridized to lane F. Thus, this clone cannot be located within 7q31. In similar mapping experiments we determined that D N A segment 5-21 was located within 7q31 and segment 117 was not (data not shown).

Detect ion o f R F L P s and linkage analysis

To determine the relationship of the D N A segments f rom within 7q31 to the CF gene, linkage analysis with a panel of CF families (Tsui et al. 1985) was performed with D N A segments 89 and 5-21. D N A segment 89 is a 1.2-kb single-copy sequence and an RFLP with MspI was identified. Linkage analysis reveals a recombinat ion

event between CF and 89 in one family (Fig. 3). This places 89 further from CF (closer to the centromere) than both the met oncogene and pA37 (D7Sl15, Rom- mens et al. 1988), which do not show recombination with CF in this family.

D N A segment 5-21 is a 6.0-kb fragment containing repetitive DNA. No subclone was found to be entirely free of repeat sequences and a mixture of two AccI frag- ments of about 3 kb and 2 kb was used as a probe after competi t ion with total human DNA. D N A segment 5 - 21 detected RFLPs with SphI , ScaI, and SacI (of 35 en- zymes tested) (Fig. 4) and linkage analysis revealed no recombinants between CF and 5-21 in 43 meioses that could be scored. To further localize 5-21, mapping ex- periments were performed with 5-21 using a human- hamster hybrid deleted for parts of 7q31 (S. Naylor, per- sonal communication) and these showed that 5-21 is fur- ther from the CF gene than met.

Expression studies

The six D N A segments described above were analyzed further to identify the transcripts to which they hybrid- ize. Northern blots prepared with sweat gland and fibro- blast poly(A +) R N A were probed with inserts from plas- mid subclones of the original phage isolates. In the case of segments 101, 105, 117, and 5-21 the hybridization was performed with probes that had been annealed to total human D N A to reduce the signal f rom repetitive sequences. Transcripts were detected by three of the six D N A segments tested. Figure 5 illustrates results ob- tained when probing the fibroblast (left lane) and epithe- lial (right lane) poly(A +) R N A with D N A segments 105, 5-21, 85B, and 117. D N A segment 5-21 detected a faint band of 6 kb in both sweat glands and fibroblasts (lanes 3 and 4), D N A segment 105 a strong band of 2.5 kb that is present in sweat glands but not fibroblasts (lanes 2 and 1, respectively), D N A segment 117 a weak signal in both sweat glands and fibroblasts (lanes 7 and 8), whereas D N A segment 85B did not detect any transcripts under the conditions used in this expriment (lanes 5 and 6) or even when washed under reduced stringency (not shown). In experiments not illustrated we found that D N A segments 89 and 101 failed to detect transcripts. Thus, one of the probes that map within 7q31 (5-21) and two that map elsewhere on chromosome 7 (105 and 117) contain coding sequences expressed in sweat glands that can be detected at this level of sensitivity. D N A segment 105 contains a portion of a gene that shows a markedly increased expression in sweat gland epithelial cells rela- tive to fibroblasts.

Discussion

The objective of this study was to isolate genomic clones located on chromosome 7q31 that contained coding re- gions, since these could be considered, in the first in- stance, candidate CF genes. Because the CF defect was thought to be expressed in sweat gland cells and not in skin fibroblasts, we employed a differential screening protocol with radioactive c D N A made from m R N A iso-

155

lated from sweat gland cells and fibroblasts. Differential screening methods have been widely used to isolate cDNAs that have tissue-specific expression (e.g., St. John and Davis 1979; Sargent and Dawid 1983; Garrison et al. 1985) but chromosome specificity was not attempted. On the other hand, Neve et al. (1986) reported the isola- tion of chromosome-21-encoded cDNAs by identifying human-specific clones in a cDNA library constructed from a hybrid cell line containing only human chromo- some 21 and also by using single-copy fragments mapped to chromosome 21 to screen a human foetal brain cDNA library. They were able to confirm that three of the posi- tive clones contained genes by hybridization to RNA iso- lated from foetal tissues. Similarly, Wiles et al. (1988) identified an abundantly expressed, human X-chromo- some-specific cDNA by differential screening of a cDNA library constructed from a hybrid cell carrying an X :6 translocation as its only human component . Most of the cDNA clones giving positive signals in the screening had repetitive sequences that interfered in the subsequent analysis.

We screened a genomic library made from flow- sorted chromosomes from a somatic cell hybrid contain- ing chromosome 7 as its only human component . In our first at tempt we found a majority of the genomic clones that showed a differential signal contained repeated se- quences. Therefore the signals reflected the presence of the repeated D N A rather than coding sequences. We tried to overcome this problem by subsequently screen- ing (and rescreening) with probes prehybridized to total human D N A (e.g., Ardeshir et al. 1983). This procedure yielded 4 out of 36 D N A segments that gave signals with the radioactively labelled cDNA probe derived from sweat gland poly(A) + RNA and that also mapped to 7q31, a region that spans about one-sixth of chromo- some 7 and contains the CF gene. However , through the combined use of linkage analysis and mapping with so- matic cell hybrids, all four were excluded as candidate genes.

Of the six clones that were analyzed for expression using Northern blots, three recognized transcripts pre- sent in normal sweat gland epithelial cells. D N A seg- ment 105 appeared not to be expressed in fibroblasts and may be epithelium - or even sweat gland specific. The D N A segment 5-21 was of interest as a candidate CF gene since in initial experiments no recombination with CF was detected. This clone detects a 6-kb transcript that appeared to be expressed at similar, low levels in both sweat glands and fibroblasts. While the CF gene is expressed in epithelial tissues such as the sweat gland, the level of its expression in tissues where no phenotype is observed in patients has not been extensively charac- terized (Riordan et al. 1989). Since the defect in chloride ion transport seen in epithelial cells (Schoumacher et al. 1987; Li et al. 1988) has now also been observed in fibro- blasts (Bear 1988) and lymphoblasts (Chen et al. 1989), expression of 5-21 in fibroblasts does not, a priori, elim- inate it from consideration as a candidate gene. How- ever, further mapping experiments showed 5-21 not to be in the right part of 7q31 and therefore excluded it from being a candidate gene. The other three segments

presumably contain genes that are expressed at a level that is too low to be detected by Northern blot hybridiza- tion. At this point, however, it also cannot be ruled out that they were detected because of cross-reacting repeti- tive sequences.

The confirmation by Northern blot hybridization that the isolated genomic clones contained coding regions was also complicated by the presence of repetitive se- quences. Our results suggest that many expressed se- quences will contain repetitive elements and that there- fore an effective protocol to block the repetitive sequences is necessary. Even though we were able to detect trans- cripts with D N A segments 117 and 5-21 by competing out the repetitive sequences with total human DNA, the protocol that we used was clearly not suitable to gener- ate a large, or even representative, number of clones from a single chromosome.

Different protocols have been designed to compete out repeated sequences. The most thorough investiga- tion was reported by Sealey et al. (1985) who showed that the most effective blockage of repeated sequences in genomic regions, such as those in cosmid clones, was achieved by preannealing the labelled probes with soni- cated total D N A to a Cot value of 100. This requires the use of lm g D N A in 100 ~1 for 100 min. However , as re- ported by Scaly et al. (1985), the success of the protocol depends on the probe used and is therefore most likely related to the extent and type of repeated sequence pre- sent in the probe. For example, it has been shown by Kariya et al. (1987) that Alu repeats contain long stretches of A-rich sequences. These are likely to be labelled by methods, such as ours, that use oligo (dT) as primers for the reverse transcriptase.

In comparison to the protocol of Sealey et al. (1985), we used less D N A for shorter times and therefore did not achieve as high Cot values. Sealey et al. (1985) have suggested that under standard conditions competit ion continues throughout the hybridization reaction and that higher effective Cot values are obtained. On the other hand, if the Cot values are too high, signal intensity is lost. Our own results (G. Melmer, unpublished results) show that the use of less D N A (100~g) for longer times (5 h to overnight) can be effective in blocking repeated sequences. However , these results as well as those re- ported by Sealey et al. (1985) were obtained with geno- mic clones as probes rather than with cDNA probes gen- erated by reverse transcription of mRNA.

With the improvements in the conditions for the pre- hybridization of total human D N A to the cDNA probe suggested above, the method described in this paper should be of use not only for the study of disorders with tissue-expression but also for the general identification of chromosome-specific cDNAs.

Acknowledgements. We thank Dr. Lap-Chee Tsui for helpful dis- cussions; members of his laboratory for assistance with the identifi- cation of RFLPs, mapping of clones, and linkage analysis; and Gail Collie and Tim Jensen for help in the culture of epithelial cells. This research was supported by grants from the Canadian Cystic Fibrosis Foundation (CFF), Cystic Fibrosis Foundation (USA), Sellers Fund, National Institutes of Health (USA) and the North Dakota Cystic Fibrosis Association. J.B. was a Fellow of

156

the CCFF and J. M. R. a Fellow of the Medical Research Council (Canada). G.M. was supported in part by the Deutsche For- schungsgemeinschaff. We also thank an anonymous reviewer for comments that have improved this manuscript.

References

Ardeshir F, Giulotto E, Zieg J, Brison O, Liao WSL, Stark GR (1983) Structure of amplified D N A in different Syrian hamster cell lines resistant to N-(phophonacetyl)-L-aspartate. Mol Cell Biol 3 : 2076-2088

Aviv H, Leder P (1972) Purification of biologically active globin messenger R N A by chromatography on oligothymidylic acid- cellulose. Proc Natl Acad Sei USA 69 : 1408-1412

Bear C (1988) Phosphorylation-activated chloride channels in hu- man skin fibroblasts. FEBS Lett 237 : 145-149

Boat TJ, Welsh M, Beaudet AL (1989) Cystic fibrosis. In: Scriver CR, Beaudet AL, Sly WS, Valle D (eds) The metabolic basis of inherited disease, 6th edn. McGraw-Hill, New York, pp 2649-2680

Buchwald M (1984) Use of cultured human cells for biochemical analysis. Clin Biochem 17 : 143-150

Chen JH, Schulman H, Gardner P (1989) A cAMP-regulated chloride channel in lymphocytes that is affected in cystic fibro- sis. Science 243 : 657-660

Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ (1979) Isolation of biologically active ribonucleic acid from sources enriched in ribonucleases. Biochemistry 18:5294-5299

Collie G, Buchwald M, Harper P, Riordan JR (1985) Culture of sweat gland epithelial cells from normal individuals and pa- tients with cystic fibrosis. In Vitro 21 : 597-602

Deaven LL, Van Dilla MA, Bartholdi MF, Carrano AV, Cram LS, Fuscoe JC, Gray JW, Hildebrand CE, Moyzis RK, Perl- man S (1986) Construction of human chromosome-specific D N A libraries from flow-sorted chromosomes. Cold Spring Harb Symp Quant Biol 51 : 159-167

Duncan A, Buchwald M, Tsui L-C (1989) In situ hybridization of two cloned chromosome 7 sequences tightly linked to the cystic fibrosis locus. Cytogenet Cell Genet 49:309-310

Feinberg AP, Vogelstein B (1983) A technique for radiolabelling D N A restriction endonuclease fragments to high specific activ- ity. Anal Biochem 132:6-13

Frizzell RA, Rechkemmer G, Shoemaker RL (1986) Altered regu- lation of airway epithelial cell chloride channels in cystic fibro- sis. Science 233 : 558-560

Garrison JC, Hardeman E, Wade R, Kedes L, Gunning P (1985) Isolation of full-length cDNAs encoding abundant adult hu- man skeletal muscle mRNAs. Gene 38 : 177-188

Kariya Y, Kato K, Hayashizaki Y, Himeno S, Tarui S, Matsubara K (1987) Revision of consensus sequence of human Alu re- peats: a review. Gene 53 : 1-10

Kerem B-s, Rommens JM, Buchanan JA, Markiewicz D, Cox TK, Chakravarti A, Buchwald M, Tsui L-C (1989) Identification of the cystic fibrosis gene: genetic analysis. Science 245:1073- 1080

Knowles MR, Stutts MJ, Spock A, Fisher N, Gatzy JT, Boucher RC (1983) Abnormal ion permeation through cystic fibrosis re- spiratory epithelium. Science 221 : 1067-1070

Li M, McCann JD, Liedke CM, Nairn AC, Greengard P, Welsch MJ (1988) Cyclic AMP-dependent protein kinase opens chlo- ride channels in normal but not cystic fibrosis airway epithe- lium. Nature 331 : 358-360

Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

McKusick VA (1988) Mendelian inheritance in man, 8th edn. Johns Hopkins University Press, Baltimore London

Neve RL, Stewart GD, Newcomb P, Van Keuren ML, Patterson D, Drabkin HA, Kurnit DM (1986) Human chromosome 21- encoded cDNA clones. Gene 49 : 361-369

Ott J (1974) Estimation of the recombination fraction in human pedigrees: efficient computation of the likelihood for human linkage studies. Am J Hum Genet 26 : 588-597

Pedersen PS, Larsen EH, Hainau B, Brandt NJ (1987) Transepi- thelial ion transport in sweat duct cell cultures derived from normals and patients with cystic fibrosis. Med Sci Res 15: 1009-1016

Quinton PM (1983) Chloride impermeability in cystic fibrosis. Na- ture 301:421-422

Riordan JR, Burns JE, Tsui L-C, Reddy MM, Quinton PM, Buch- wald M (1987) Utilization of cultured epithelial cells from the sweat gland in studies of the CF defect. In: Riordan JR, Buch- wald M (eds) Genetics and epithelial cell dysfunction in cystic fibrosis. Liss, New York, pp 59-71

Riordan JR, Rommens JM, Kerem B-s, Alon N, Rozmahel R, Grzelczak Z, Zielenski J, Lok S, Plavsic N, Chou J-L, Drumm ML, Iannuzzi MC, Collins FS, Tsui L-C (1989) Identification of the cystic fibrosis gene: cloning and characterization of the complementary DNA. Science 245 : 1066-1073

Rommens JM, Zengerling S, Burns J, Melmer G, Kerem B-s, Plavsic N, Zsiga M, Kennedy D, Markiewicz D, Rozmahel R, Riordan JR, Buchwald M, Tsui L-C (1988) Identification and regional localization of D N A markers on chromosome 7 for the cloning of the cystic fibrosis gene. Am J Hum Genet 43: 645 -663

Rommens JM, Iannuzzi MC, Kerem B-s, Drumm ML, Melmer G, Dean M, Rozmahel R, Cole JL, Kennedy D, Hidaka N, Zsiga M, Buchwald M, Riordan JR, Tsui L-C, Collins FS (1989) Identification of the cystic fibrosis gene: chromosome walking and jumping. Science 245 : 1059-1065

Sargent TD, David IB (1983) Differential gene expression in the gastrula ofXenopus laevis. Science 222:135-139

Schoumacher RA, Shoemaker RL, Halm DR, Tallant EA, Wal- lace RW, Frizzell R A (1987) Phosphorylation fails to activate chloride channels from cystic fibrosis airway cells. Nature 330 : 752-754

Sealey PG, Whittaker PA, Southern EM (1985) Removal of re- peated sequences from hybridisation probes. Nucleic Acids Res 13 : 1905-1922

St John TP, Davis RW (1979) Isolation of galactose-inducible DNA sequences from Saccharomyces cerevisiae by differential plaque filter hybridization. Cell 16 : 443-452

Tsui L-C, Buchwald M, Barker D, Braman JC, Knowlton R, Schumm JW, Eiberg H, Mohr J, Kennedy D, Plavsic N, Zsiga M, Markiewicz D, Akots G, Brown V, Helms C, Gravius T, Parker C, Rediker K, Donis-Keller H (1985) Cystic fibrosis locus defined by a genetically linked polymorphic D N A mark- er. Science 230 : 1054-1057

Tsui L-C, Rommens JM, Burns J, Zengerling S, Riordan JR, Car- lock LR, Grzeschik K-H, Buchwald M (1988) Progress to- wards cloning the cystic fibrosis gene. Philos Trans R Soc Lond [Biol] 319 : 263-273

Welsch M J, Liedke CM (1986) Chloride and potassium channels in cystic fibrosis airway epithelium. Nature 322 : 467-470

Wiles MV, Alexander CM, Goodfellow PN (1988) Isolation of an abundantly expressed sequence from the human X chromo- some by differential screening. Somat Cell Mol Genet 14:31- 39

Yankaskas JR, Cotton CU, Knowles MR, Gatzy JT, Boucher RC (1985) Culture of nasal epithelial cells on collagen matrix sup- ports. Am Rev Respir Dis 132 : 1281-1287

Yankaskas JR, Stutts MJ, Cotton CU, Knowles MR, Gatzy JT, Boucher RC (1988) Primary culture of cystic fibrosis airway epithelial cells: preservation of disease-specific ion transport abnormalities. In: Mastella G, Quinton PM (eds) Cellular and molecular basis of cystic fibrosis. San Francisco Press, San Francisco, pp 369-382

Zengerling S, Tsui L-C, Grzeschik K-H, Olek K, Riordan JR, Buchwald M (1987) Mapping of DNA markers linked to the cystic fibrosis locus on the long arm of chromosome 7. Am J Hum Genet 40 : 228-236