Somatic Cell and Molecular Genetics, Vol. 17, No. 2, 1991,pp. 191-200
Distribution of Genes for Gap Junction Membrane Channel Proteins on Human and Mouse Chromosomes Chih-Lin Hsieh, 1 Nalin M. K u m a r , 2 N o r t o n B. Gilula, 2 and Uta F r a n c k e ~ ~Howard Hughes Medical Institute and Department of Genetics, Stanford University Medical Center. Stanford, California 94305; and 2Department of'Molecular Biology, Research Institute of Scripps Clinic, La JoUa, California 92037 Received 20 D e c e m b e r 1990
Abstract--Gap junctions are widely distributed structures that mediate communication between cells. The channels that allow passage of small molecules between adjacent cells are made up of oligomeric proteins (connexins) that are encoded by a family of related genes. By probing somatic cell hybrid DNA on Southern filters with rat or human cDNAs or human genomic fragments, we have mapped four functioning gap junction genes, (%, f3~, 13:~and %), to different sites on human chromosomes: GJA1 (connexin43) to 6p21.1-q24.1; GJB1 (connexin32) to Xcen-q22; GJB2 (connexin26) to 13; and GJA3 (connexin46) also to 13, probably near GJB2. The GJA3 probe also hybridized to a restriction fragment that was mapped to chromosome 1. A GJAl-related pseudogene GJA1P was assigned to chromosome 5. The homologous loci in mouse were assigned to regions of known conserved syntenic groups: Gja-1 to chromosome 10; Gjb-1 to XD-F4 and Gjb-2 to 14. Of two sites of hybridization with the GJA3 probe, on mouse 14 and 5, we assume that the site on 14 corresponds" to the GJA3 locus on human 13. Based on these data, additional members of this family of related genes can be isolated and characterized, and possibte human and mouse mutations can be identified.
INTRODUCTION
Cell-to-cell communication in tissues occurs through membrane channels located in membrane specializations known as gap junctions (1-3). These plaque-like elements are characterized by two adjoining plasma membranes that are separated by a 2- to 4-nm gap. The junctions contain a polygonal arrangement of oligomers, or connexons, that form channels for the diffusion of ions and chemical mediators between cells. Structural analyses indicate that each membrane channel consists of an oligomeric association of protein, presumably a hexamer, that forms a channel space (1, 4, 5). However, not all gap junctions are structurally and function-
ally identical (reviewed in reference 3). Gap junctions are formed by a family of related proteins, and their apparent different physiological properties in different cell types may be due to differences in the constituent gap junction proteins. Thus far, five different gap junction genes have been identified 0xl, a2, a3, 13, 132) (6-13) that produce junction proteins, also called connexins, which share a basic structural motif (14) with significant sequence diversity in the cytoplasmic domains. Furthermore, some cells produce at least two different gap junction proteins simultaneously (15). Finally, gap junctions provide the mechanism for electrical synaptic transmission in excitable tissues, and communication via these junctions is thought to play
191
0740-7750/91/0300-0191506.50/0 ~ 1991 Plenum Publishing Corporation
192
a major role in embryogenesis, cellular differentiation and development (reviewed in 16). The ubiquitous distribution of gap junctions and their essential role in tissue homeostasis, cell growth, and development has made it difficult to study their potential involvement in the initiation and progression of diseases. With the recent cloning of cDNAs encoding gap junction proteins from several species, it has become possible to dissect this gene family and to determine the chromosomal localization of the different gap junction genes. By using cloned cDNA and genomic fragments, we have assigned four active gap junction genes, and one pseudogene, to different human and mouse chromosomes. This information provides a baseline for identifying novel types of gap junction genes as well as potential mutants. MATERIALS AND METHODS
Somatic Celt Hybrids and DNA from Mouse Strains. Somatic cell hybrid mapping panels of 11 Chinese hamster x human and 13 rodent x mouse hybrid clones or subclones used for primary assignments have been described before (17). Four rodent x human hybrid cell clones that retain overlapping regions of human chromosome 6 in the absence of an intact copy of this chromosome were used to regionally assign the al subunit gene. Regional assignment of the [31 subunit genes on the human and mouse X chromosomes was carried out with 14 rodent x human hybrid cell clones and two series I mouse x Chinese hamster hybrid subclones (17a). DNA from five inbred mouse strains (C57L/J, C57BL/6J, DBA/2J, AKR/J, and C3H/HeJ) was purchased from Jackson Laboratory (Bar Harbor, Maine). DNA Digestion and Probes. Genomic DNA from parental species controls and hybrid cell lines was digested with various enzymes. BamHI, BgllI, EcoRI, HincII, HindIII, PstI, XbaI, and KpnI were used to
Hsieh et al,
search for polymorphisms in inbred mouse strains. For mapping of the a~ subunit gene, a 1.6-kb rat cDNA fragment was used (18). A 1.6-kb rat (Y~3cDNA encoding a 46-kDa protein (Kumar and Gilula, in preparation) was first used to localize the gene, but this probe cross-hybridized extensively with human and mouse DNA. Therefore, a gelpurified 1.1-kb human genomic fragment was used to localize the human % subunit gene. Genes for [3 subunits were localized with a 1.5-kb human [31 cDNA (6) and a 2.2-kb rat [32 cDNA (Nishi, Kumar and Gilula, in preparation). All probes were labeled with [32P]dCTP by random hexamer priming (19). Southern Blot Hybridization. Conditions for the preparation of Southern filters, prehybridization, and hybridization were as described previously (20). Filters were washed in 2 x SSC-0.5% SDS for 15 rain at room temperature, then two times in 1 x SSC-1% SDS for 10 rain at 65°C and once in 0.5 x SSC-0.5% SDS for 10 min at 65°C before autoradiography. RESULTS
Assignment of ~1 Subunit Locus GJA1 in Human and in Mouse. When the rat a 1 cDNA probe was hybridized with EcoRIdigested DNA from a panel of 11 human x Chinese hamster cell hybrids, five fragments of 7.2, 6.6, 4.1, 3.1, and 2.1 kb were detected in human control DNA (Fig. 1A, lane 2), and three fragments of 6.0, 5.2, and 4.5 kb in Chinese hamster DNA (Fig. 1A, lane 1). The weakly hybridizing 7.2-kb human band appeared to be discordant with the other fragments but could not be scored reliably in this set of hybrids. The 2.1-kb fragment was present only in the control DNA, not in any of the hybrid cell lines, and it may represent a rare polymorphism. The human 4.1- and 3.1-kb bands were concordantly present, but only in hybrids that retained human chromosome 6. These two fragments are consistent with the presence of an EcoRI site in the
Gap Junction Genes
193 2 2 2 2
2 2 2
GJA 1
6+-+ Fig. 1. Mapping of GJA1 gene in human. (A) Southern filter of EcoRI-digested DNA from Chinese hamster control (lane 1), human control (lane 2), and Chinese hamster x human hybrid cell lines (lanes 3-8). Lanes 5 and 7 are positive for the 4.1- and 3.1-kb human bands. Only lane 8 is positive for the 6.6-kb pseudogene fragment. (B) Regional mapping of GJA1 on chromosome 6. The vertical bars represent the regions of chromosome 6 present in somatic cell hybrid lines; the GJA1 gene was assigned to region p21.1-q24.1 (bracket) based on the presence (+) and absence ( - ) of the human-specific 4.1 and 3.1 EcoRI fragments.
coding sequence of human ~ (21). The 6.6-kb fragment, which represents a pseudogene (21), segregated with human chromosome 5 (Fig. 1A, lane 8). Since the initial signal was weak, the localization of this fragment was further confirmed with three other hybrid cell lines (not shown). The assignment of the o~1 fragments to chromosomes 5 and 6 and exclusion of all other chromosomes is summarized in Table 1. The fragments that segregated with human chromosome 6 were sublocalized to region
p21.1-q24.1 with three other hybrid cell lines (Fig. 1B). When the same rat o~I-GJ cDNA probe was hybridized with EcoRI-digested DNA from a panel of 13 mouse x rodent hybrids, a single major 7.6-kb mouse fragment (Fig. 2, lane 1), and a 7.8-kb rat fragment (Fig. 2, lane 6), were detected in addition to the 6.0-, 5.2-, and 4.5-kb hamster fragments (Fig. 2, lane 2). The presence of the 7.6-kb mouse signal was concordant with the presence of mouse chromosome 10 in the 13 hybrids
194
Hsieh et al.
Table 1. Discordancy Ratios (Percent) for Human Gap Junction Gene-Specific Restriction Fragments and Each Human Chromosome in 11-13 Chinese Hamster x Human Somatic Cell Hybrids Human chromosome Gene
1
2
3
4
5
6
7
8
9
10
GJA1 45 40 40 25 GJA1P 54 42 45 30 GJA3 55 45 40 50 GJB1 55 70 50 50 GJB2 55 45 40 50
50 0 20 80 20
0 38 45 45 45
67 60 67 56 67
60 46 27 64 27
60 42 36 60 36
36 44 18 67 56 55 30 38 45 46 31 55 62 46 45 78 64 0 44 36 40 64 56 45 78 33 55 50 45 78 64 0 44 36 40
studied; all other chromosomes were excluded by at least three discordant hybrids (Table 2). When the same probe was hybridized to a filter with BgllI-digested DNA, the major mouse fragment of 7.4 kb was also concordant with chromosome 10. Since no potymorphism was detected in inbred strains of mice with eight restriction enzymes, linkage mapping with recombinant inbred mouse strains was not possible, Assignment of ~3 Subunit Locus GJA3 in Human. When a rat a3 cDNA probe was hybridized to BamHI-digested DNA from human and mouse mapping panels, more than 20 fragments of varying intensities were
11 12
13 14
15
16 17 18 64 31 45 90 45
60 58 55 36 55
19 20 21 22 45 42 55 55 55
X
27 20 30 50 31 58 58 100 45 50 40 100 45 60 30 0 45 50 40 100
present in each lane. The most strongly hybridizing human fragment of 3.7 kb segregated with human chromosome 13 and another fragment (2.0 kb) that was also separable from hamster fragments segregated with human chromosome 1 (data not shown). When a human % genomic probe was hybridized to the same filters, a single 17.5-kb BamHI fragment was detected in human DNA and in hybrids retaining human chromosome 13 (Fig. 3, lanes 4-6, and 9) and was absent in hybrids not containing human chromosome 13 (lanes 3, 7, and 8). All other human chromosomes were ruled out by at least two discordant hybrids (Table 1). No
Fig. 2. Mapping of Gja-1 gene in mouse. Hybridization of the 32P-labeled rat cDNA probe to EcoRI-digested DNA from a panel of somatic cell hybrid cell lines. Lane 1, mouse control; lane 2, Chinese hamster; lanes 3 and 4, Chinese hamster x mouse hybrids; lane 5, rat x mouse hybrid; and lane 6, rat control. Lanes 3 and 5 are negative and lane 4 is positive for the 7.6-kb mouse signal.
Gap Junction Genes
195
Table 2. Discordancy Ratios (in Percent) for Mouse Gap Junction Gene-Specific Restriction Fragments and Each Mouse Chromosome in 11-15 Rodent x Mouse Somatic CelI Hybrids~ Mouse chromosome Gene
1
2
3
4
5
6
7
8
9
10
1I
12
13
14
15
16
17
18
19
X
Gja-t Gjb-1 Gjb-2
25 29 50
62 47 46
46 47 46
46 57 42
64 54 27
33 38 45
42 29 42
46 47 31
25 53 38
0 57 50
38 80 38
23 53 69
33 71 58
46 47 0
46 13 54
50 36 42
62 20 38
33 71 50
50 43 50
50 0 42
"Hybrids in which a particular chromosome was structurally rearranged or present in fewer than 10% of cells were excluded.
mouse signal and faint cross-hybridization with Chinese hamster DNA were detected with the human genomic probe, suggesting that it was derived from a highly diverged part of the gene. With the rat cDNA probe, one of the mouse fragments that could be distinguished from the numerous Chinese hamster fragments was concordant with mouse chromosome 5 in hybrid cell panels, and another mouse fragment was concordant with mouse chromosome 14 (data not shown). A mouse genomic probe will be required to determine which site represents the % coding gene. Assignment of ~ Subunit Locus GJB1 in Human and Mouse. When labeled human [31 cDNA was hybridized to EcoRI-digested DNA from the human mapping panel, a single 10.5-kb Chinese hamster fragment (Fig° 4A, lane 1) and a major 5.1-kb human
fragment (Fig. 4A, lane 2) were observed. The human-specific fragment was concordant with the X chromosome, and all other human chromosomes were excluded by three to nine discordant hybrids (Table 1). This result was confirmed with the rat x human hybrid cell line XXII-18A that contains the X as the only human chromosome (not shown). In the HindIII-digested DNA from an X chromosome regional mapping panel, the smallest overlapping region that was concordant with the human [3~signal is the proximal long arm, region cen-q22 (Fig. 4B). In BglII-digested DNA from the mouse mapping panel, a single 6.9-kb mouse fragment (Fig. 5, lane 1), several Chinese hamster fragments of 14.0, 8.5, 7.8, 4.9, and 3.8 kb (Fig. 5, lane 2), a major 7.3-kb and several weakly hybridizing rat bands (Fig. 5, lane 10) hybridized with the [3~ probe. The
Fig. 3. Assignment of GJA3 gene in human with a human genomic probe. Lane 2, human control; lane 1, Chinese hamster. The 11.0-kb band was very faint, but it was clearly visible after longer exposure and also in lanes 3-9, which contain Chinese hamster x human hybrid cell DNA. The 17.5-kb human fragment was observed only in hybrids retaining human chromosome 13 (lanes 4-6 and 9).
196
Hsieh et al.
il
] X
+ + +
_
GJB1
_
Fig. 4. Assignment of the GJB1 locus in human. (A) DNA from a panel of Chinese hamster x human somatic cell hybrid cell lines (lanes 3-8) and human (lane 2) and Chinese hamster (lane 1) controls was hybridized with 32p-labeled human [31-GJ cDNA probe, The 5,1-kb human fragment was clearly visible in lanes 3-6 and absent in lanes 7 and 8. (B) Regional assignment of the GJB1 gene on human X chromosome. Vertical bars represent regions of the X chromosome present in hybrid cells; the presence (+) or absence ( - ) of human 5.1-kb fragment is indicated below each vertical bar. Bracket indicates location of GJB1,
mouse signal was concordant with the mouse X chromosome. All other mouse chromosomes were excluded by at least two discordant hybrids (Table 2). With two hybrids containing complementary regions of the mouse X chromosome, the Gjb-1 gene was assigned to region D-F4 of mouse X chromo-
some distal to the T(X;16)16H translocation breakpoint (17). Assignment of [32Subunit Locus GJB2 in Human and Mouse. The rat [32 cDNA was used to probe EcoRI-digested DNA from human and mouse mapping panels. In Chinese hamster DNA, a strongly hybridiz-
Gap Junction Genes
197
Fig. 5. Assignmentof Gjb-1in mouse. Lane 1, mouse control; lane 2, Chinese hamster control; lanes 3-8, Chinese
hamster x mousehybridcell lines;lane 9, rat x mousehybridcell line;lane 10, rat control.Lanes3-7 are positivefor the mouse6.9-kbsignal;lanes 8 and 9 onlyhave the Chinese hamsteror rat fragments. ing 8.4-kb band and weakly hybridizing 21-, 18-, 6.7-, 5.3-, 4.2-, and 2.8-kb fragments were present (Fig. 6A, lane 1 and Fig. 6B, lane 2). Human fragments were 14 and 11.5 kb in size (Fig. 6A, lane 2), mouse bands were 21 and 11.5 kb (Fig. 6B, lane 1), and rat fragments were 21, 7, and 4.6 kb (Fig. 6B, lane 9). In hybrid cells, the two human fragments (Fig. 6A, lane 4) were concordant with human chromosome 13 (Table 1). The mouse 21 kb fiagment was not scored since it comigrated with Chinese hamster and rat fragments. The major ll.5-kb mouse band was present in all hybrids containing mouse chromosome 14 and absent in the hybrids not containing mouse chromosome 14. All other mouse chromosomes were ruled out by at least three discordant hybrids (Table 2). No polymorphism was found in five inbred mouse strains tested with eight enzymes. DISCUSSION
We have assigned four genes, GJA1, GJA3, GJB1, and GJB2, that encode four
different gap junction proteins (%, %, [31,and [3z) to various human (HSA) and mouse (MMU) chromosomes. In addition, a pseudogene, GJA1P, was mapped to HSA5, but no homolog of this was found in the mouse. It is clear from the Southern hybridization patterns of cDNA probes that additional related genes exist, e.g,, the GJA1 probe detected a fragment on the X chromosome not identical to GJB1 (data not shown) and the GJA3 cDNA probe detected many additional fragments of which only two could be mapped, one to HSA1 and one to MMU5. The relationships of the known coding genes mapped to the sizes of m R N A and deduced proteins are summarized in Table 3. Comparative mapping serves to define regions of conserved homology on human and mouse chromosomes. The GJA1 locus was assigned to HSA6, region p21.1-q24.1, and to MMU10. Two other genes, IFNGR1 encoding the interferon gamma receptor (binding unit), and the avian myeloblastosis viral oncogene homolog MYB, were previously assigned to HSA6, region q23-q24 and q22-q23, respectively, and to MMU10 (22,
198
Fig. 6. Chromosomal localization of the
Hsieh et al.
GJB2/Gjb-2 gene in human and mouse. (A) Southern filter of
EcoRI-digested DNA from Chinese hamster (lane 1), human (lane 2), and Chinese hamster x human hybrid cell lines (lanes 3-5) hybridized with labeled rat cDNA probe. Lane 4 is positive for both 14- and ll.5-kb human fragments and lanes 3 and 5 are negative. (B) The rat GJB2 cDNA probe was hybridized to EcoRl-digested DNA from rodent x mouse hybrid cell lines (lanes 3-7 are Chinese hamster x mouse and lane 8 is rat x mouse) and controls (lane 1, mouse; lane 2, Chinese hamster; and lane 9, rat). Hybrids in lanes 3 and 6 are positive for the ll.5-kb mouse signal, and those in lanes 4, 5, 7, and 8 are negative. 23). W e p r e d i c t t h a t t h e GJA1 locus is in t h e s a m e region, since g e n e s o n t h e s h o r t a r m a n d distal to q25 on t h e long a r m o f H S A 6 have all b e e n a s s i g n e d to M M U 1 7 , a n d g e n e s on t h e p r o x i m a l l o n g a r m o f H S A 6 h a v e b e e n m a p p e d to M M U 9 a n d M M U 4 (22, 23).
GJA1 is likely to b e i d e n t i c a l to t h e p o o r l y c h a r a c t e r i z e d gap j u n c t i o n - l i k e g e n e GJAL p r e v i o u s l y a s s i g n e d to 6 q 1 6 - q 2 2 by in situ h y b r i d i z a t i o n as r e p o r t e d in a n a b s t r a c t (24). T h e GJA1P p s e u d o g e n e o n c h r o m o s o m e 5 was a p p a r e n t l y n o t d e t e c t e d in this study.
Gap Junction Genes
199
Table 3. Localization of Genes for Gap Junction Proteins to Human and Mouse Chromosomes Locus
Gene
mRNA (kb)
Protein (kDa)
Human
Mouse
GJA1/Gja-1 GJAIP GJA3/Gja-3 GJA3L GJB1/Gjb-1 GJB2/Gjb-2
eq-GJ (Cx43)
3.6
43
6p21.1-q24.1
10
2.9
46
5 13
(14 or 5)
? 1.6 2.8
? 32 26
1 Xcen-q22 13
(?5) XD-F4 14
%-GJ (Cx46) ? 13~-GJ (Cx32) 132-GJ (Cx26)
Comparative mapping data also suggest that the mouse homolog of GJA1, Gja-1, may be near the Myb locus, which has been mapped to the proximal part of MMU10. The gene for the a 3 subunit, GJA3, was assigned to HSA13 with a human-specific genomic probe. However, the hybridization patterns observed with the rat cDNA probe indicated a second site on chromosome 1 (GJA3L). Mouse chromosome 14 was observed to segregate with one of the mouse fragments that hybridized with the rat % cDNA probe. This fragment may represent Gja-3, the mouse homolog of GJA3, because HSA13 is known to share a region of homology with MMU14 (22, 23). Other mouse GJA3 related fragments are present on MMU5. Once the mouse gene has been isolated, the map position of this locus in mouse will be clarified. Interestingly, we have found that the GJB2/Gjb-2 locus is also on HSA13 and MMU14. Three other genes, ESD (esterase D), RB1 (retinoblastoma), and HTR2 (serotonin receptor subtype 2), have been assigned to HSA13 and MMU14 (25, 26), and these three genes are close to each other in band 13q14 and by linkage analysis: cen-ESD(1.2cM)-HTR2-(2.8cM)-RBI-tel (26). Our Chinese hamster x human hybrid cell line 31-6AHAT that contains a small unidentified fragment of HSA13 is positive for both gap junction genes GJA3 and GJB2 and also for RB1 and HTR2 (26). Therefore, we postulate that GJB2/Gjb-2 is located close to these genes in both human and mouse. The GJB1 locus was assigned to the
human X chromosome, region cen-q22, and to the mouse X chromosome region D-F4. These regional assignments are consistent with the current comparative maps of human and murine X chromosomes (22, 27) and place the locus in segment 4 of the five identified conserved segments which in the mouse includes the interval between Tfm and Plp(jp) (22). Although the [31 and 132 gap junction proteins were first described in liver, and the eq gap junction gene product predominates in heart and skeletal muscle, it is now well established that individual gap junction proteins are expressed in many tissues in an overlapping pattern (13, 18, 28). Heterologous gap junctions could interact with each other and individual junctions may be made up of more than one channel type. These facts and the strong possibility that additional types may exist make it difficult to predict a phenotype for a mutation in a specific gap junction gene locus. In an attempt to obtain clues from the comparative gene map, we have noted that two mouse gene loci that can generate phenotypically identical mutations are located in regions that most likely contain GJP loci: dl (downless) on MMU10 in the region of Gja-1, and Ta (Tabby) on the X in the region of Gjb-1 (22). The underlying defects, however, involve growth abnormalities of epidermal derivatives (29). A region containing related genes could exist between MMU10 and X, as has been established for other autosomes. In summary, we present evidence that genes encoding the gap junction membrane
200
channels are dispersed over the human and mouse chromosomes. The exact number of gap junction genes remains unknown, but previously uncharacterized genes are present on chromosomes 1 and X. Analysis of these genes, together with their localization, will be of importance in understanding the evolution of the gap junction multigene family. ACKNOWLEDGMENTS This work was supported by NIH research grants GM26105 to U.F., GM37907 to N.B.G. and N.M.K., GM37904 to N.B.G., EY06884 to N.M.K., and by the Howard Hughes Medical Institute of which U.F. is an investigator and C.-L.H. was an associate. LITERATURE CITED 1, Revel, J.P., and Karnovsky, M.J. (1967). J. Cell Biol. 33:C7-C12. 2. Gilula, N.B., Reeves, O.R., and Steinbach, A. (1972). Nature 235:262-265. 3. Loewenstein, W.R. (1981). Physiol. Rev. 61:829913. 4. Makowski, L., Caspar, D.L.D., Phillips, W.C., and Goodenough, D.A. (1977). J. Cell Biol. 74:629-645. 5. Unwin, P.N.T., and Zampighi, G. (1980). Nature 283:545-549. 6. Kumar, N.M., and Gilula, N.B. (1986). J. Cell Biol. 103:767-776. 7. PauI, D.L. (1986).J. CellBiol. 103:123-134. 8. Beyer, E.C., Paul, D.L., and Goodenough, D.A. (1987).J. CellBiol. 105:2621-2629. 9. Gimlich, R.L., Kumar, N.M., and Gilula, N.B. (1988).J. CellBioL 107:1065-1073. 10. Zhang, J.T., and Nicholson, B. (1990). J. Cell BioL 109:3391-3402. 11. Ebihara, L.E., Beyer, E., Swenson, K.I., Paul, D.L., and Goodenough, D.A. (1989). Science 243:11941195. 12. Gimlich, R.L., Kumar, N.M., and Gilula, N.B. (1990). J. Cell Biol. 110:597-605.
Hsieh et al.
13. Beyer, E.C., Paul, D.L., and Goodenough, D.A. (1990). J. Membr. Biol. 116:187-194. 14. Milks, L.C., Kumar, N.M., Houghten, R., Unwin, N., and Gilula, N.B. (1988). E M B O J. 7:2967-2975. 1.5. Nicholson, B.J., and Zhang, J.T. (1988). Mol. Cell Biol. 7:207-218. 16. Guthrie, S.C., and Gilula, N.B. (1989). TINS 12:1.2-16. 17. Yang-Feng, T.L., DeGennaro, L.J., and Francke, U. (1986). Proc. Natl. Acad. Sci. U.S.A. 83:86798683. 17a. Wieacker, P., Davies, K.E., Cooke, H.J., Pearson, P.L., Williamson, R., Bhattacharya, S., Zirnmer, J., and Ropers, H.-H. (1984). Amer. J. Hum. Genet. 36:265-276. 18. Risek, B., Guthrie, S., Kumar, N., and Gilula, N.B. (1990). J. CellBiol. 110:269-282. 19. Feinberg, A., and Vogelstein, B. (1983). Anal Biochem. 132:6-13. 20. Hsieh, C.-L., Sturm, R., Herr, W., and Franeke, U. (1990). Genomics 6:666-672. 21. Fishman, G.I., Spray, D.C., and Leinwand, L.A. (1990). J. Cell BioL 111:589-598. 22. Searle, A.G., Peters, J., Lyon, M.F., Hall, J.G., Evans, E.P., Edwards, J.H., and Buckle, V.J. (1989).Ann. Hum. Genet. 53:89-140. 23. Lalley, P.A., Davisson, M.T., Graves, J.A.M., O'Brien, S.J., Womack, J.E., Roderick, T.H., Creau-Goldberg, N., Hillyard, A.L., Doolittle, D.P., and Rogers, J.A. (1989). Cytogenet. Cell Genet. 51:503-532. 24. Jayakar, P., and Rabin, M. (1989). Am. J. Hum. Genet. 45:(suppl.):A144. 25. Hsieh, C.-L., Lee, W.-H., Lee, E.Y.-H.P., Killary, A.M., Lalley, P.A., and Naylor, S.L. (1989). Somat. Cell Mol. Genet. 15:461-464. 26. Hsieh, C.-L., Bowcock, A.M., Farrer, L.A., Hebert, J.M., Huang, K.N., Cavalli-Sforza, L.L., Julius, D., and Francke, U. (1990). Somat. Cell Mol. Genet. 16:567-574. 27. Francke, U., Darras, B.T., Zander, N.F., and Kilimann, M.W. (1989). Am. J. Hum. Genet. 45:272-282. 28. Dermietzel, R., Traub, O., Hwang, T.K., Beyer, E., Bennett, M.V.L., Spray, D.C., and Willecke, K. (1989). Proc. Natl. Acad. Sci. U.S.A. 86:1014810152. 29. Lyon, M.F., and Searle, A.G. (eds.) (1989). Genetic Variants and Strains of the Laboratory Mouse, 2nd ed. (Oxford Univ. Press).
Distribution of Genes for Gap Junction Membrane Channel Proteins on Human and Mouse Chromosomes Chih-Lin Hsieh, 1 Nalin M. K u m a r , 2 N o r t o n B. Gilula, 2 and Uta F r a n c k e ~ ~Howard Hughes Medical Institute and Department of Genetics, Stanford University Medical Center. Stanford, California 94305; and 2Department of'Molecular Biology, Research Institute of Scripps Clinic, La JoUa, California 92037 Received 20 D e c e m b e r 1990
Abstract--Gap junctions are widely distributed structures that mediate communication between cells. The channels that allow passage of small molecules between adjacent cells are made up of oligomeric proteins (connexins) that are encoded by a family of related genes. By probing somatic cell hybrid DNA on Southern filters with rat or human cDNAs or human genomic fragments, we have mapped four functioning gap junction genes, (%, f3~, 13:~and %), to different sites on human chromosomes: GJA1 (connexin43) to 6p21.1-q24.1; GJB1 (connexin32) to Xcen-q22; GJB2 (connexin26) to 13; and GJA3 (connexin46) also to 13, probably near GJB2. The GJA3 probe also hybridized to a restriction fragment that was mapped to chromosome 1. A GJAl-related pseudogene GJA1P was assigned to chromosome 5. The homologous loci in mouse were assigned to regions of known conserved syntenic groups: Gja-1 to chromosome 10; Gjb-1 to XD-F4 and Gjb-2 to 14. Of two sites of hybridization with the GJA3 probe, on mouse 14 and 5, we assume that the site on 14 corresponds" to the GJA3 locus on human 13. Based on these data, additional members of this family of related genes can be isolated and characterized, and possibte human and mouse mutations can be identified.
INTRODUCTION
Cell-to-cell communication in tissues occurs through membrane channels located in membrane specializations known as gap junctions (1-3). These plaque-like elements are characterized by two adjoining plasma membranes that are separated by a 2- to 4-nm gap. The junctions contain a polygonal arrangement of oligomers, or connexons, that form channels for the diffusion of ions and chemical mediators between cells. Structural analyses indicate that each membrane channel consists of an oligomeric association of protein, presumably a hexamer, that forms a channel space (1, 4, 5). However, not all gap junctions are structurally and function-
ally identical (reviewed in reference 3). Gap junctions are formed by a family of related proteins, and their apparent different physiological properties in different cell types may be due to differences in the constituent gap junction proteins. Thus far, five different gap junction genes have been identified 0xl, a2, a3, 13, 132) (6-13) that produce junction proteins, also called connexins, which share a basic structural motif (14) with significant sequence diversity in the cytoplasmic domains. Furthermore, some cells produce at least two different gap junction proteins simultaneously (15). Finally, gap junctions provide the mechanism for electrical synaptic transmission in excitable tissues, and communication via these junctions is thought to play
191
0740-7750/91/0300-0191506.50/0 ~ 1991 Plenum Publishing Corporation
192
a major role in embryogenesis, cellular differentiation and development (reviewed in 16). The ubiquitous distribution of gap junctions and their essential role in tissue homeostasis, cell growth, and development has made it difficult to study their potential involvement in the initiation and progression of diseases. With the recent cloning of cDNAs encoding gap junction proteins from several species, it has become possible to dissect this gene family and to determine the chromosomal localization of the different gap junction genes. By using cloned cDNA and genomic fragments, we have assigned four active gap junction genes, and one pseudogene, to different human and mouse chromosomes. This information provides a baseline for identifying novel types of gap junction genes as well as potential mutants. MATERIALS AND METHODS
Somatic Celt Hybrids and DNA from Mouse Strains. Somatic cell hybrid mapping panels of 11 Chinese hamster x human and 13 rodent x mouse hybrid clones or subclones used for primary assignments have been described before (17). Four rodent x human hybrid cell clones that retain overlapping regions of human chromosome 6 in the absence of an intact copy of this chromosome were used to regionally assign the al subunit gene. Regional assignment of the [31 subunit genes on the human and mouse X chromosomes was carried out with 14 rodent x human hybrid cell clones and two series I mouse x Chinese hamster hybrid subclones (17a). DNA from five inbred mouse strains (C57L/J, C57BL/6J, DBA/2J, AKR/J, and C3H/HeJ) was purchased from Jackson Laboratory (Bar Harbor, Maine). DNA Digestion and Probes. Genomic DNA from parental species controls and hybrid cell lines was digested with various enzymes. BamHI, BgllI, EcoRI, HincII, HindIII, PstI, XbaI, and KpnI were used to
Hsieh et al,
search for polymorphisms in inbred mouse strains. For mapping of the a~ subunit gene, a 1.6-kb rat cDNA fragment was used (18). A 1.6-kb rat (Y~3cDNA encoding a 46-kDa protein (Kumar and Gilula, in preparation) was first used to localize the gene, but this probe cross-hybridized extensively with human and mouse DNA. Therefore, a gelpurified 1.1-kb human genomic fragment was used to localize the human % subunit gene. Genes for [3 subunits were localized with a 1.5-kb human [31 cDNA (6) and a 2.2-kb rat [32 cDNA (Nishi, Kumar and Gilula, in preparation). All probes were labeled with [32P]dCTP by random hexamer priming (19). Southern Blot Hybridization. Conditions for the preparation of Southern filters, prehybridization, and hybridization were as described previously (20). Filters were washed in 2 x SSC-0.5% SDS for 15 rain at room temperature, then two times in 1 x SSC-1% SDS for 10 rain at 65°C and once in 0.5 x SSC-0.5% SDS for 10 min at 65°C before autoradiography. RESULTS
Assignment of ~1 Subunit Locus GJA1 in Human and in Mouse. When the rat a 1 cDNA probe was hybridized with EcoRIdigested DNA from a panel of 11 human x Chinese hamster cell hybrids, five fragments of 7.2, 6.6, 4.1, 3.1, and 2.1 kb were detected in human control DNA (Fig. 1A, lane 2), and three fragments of 6.0, 5.2, and 4.5 kb in Chinese hamster DNA (Fig. 1A, lane 1). The weakly hybridizing 7.2-kb human band appeared to be discordant with the other fragments but could not be scored reliably in this set of hybrids. The 2.1-kb fragment was present only in the control DNA, not in any of the hybrid cell lines, and it may represent a rare polymorphism. The human 4.1- and 3.1-kb bands were concordantly present, but only in hybrids that retained human chromosome 6. These two fragments are consistent with the presence of an EcoRI site in the
Gap Junction Genes
193 2 2 2 2
2 2 2
GJA 1
6+-+ Fig. 1. Mapping of GJA1 gene in human. (A) Southern filter of EcoRI-digested DNA from Chinese hamster control (lane 1), human control (lane 2), and Chinese hamster x human hybrid cell lines (lanes 3-8). Lanes 5 and 7 are positive for the 4.1- and 3.1-kb human bands. Only lane 8 is positive for the 6.6-kb pseudogene fragment. (B) Regional mapping of GJA1 on chromosome 6. The vertical bars represent the regions of chromosome 6 present in somatic cell hybrid lines; the GJA1 gene was assigned to region p21.1-q24.1 (bracket) based on the presence (+) and absence ( - ) of the human-specific 4.1 and 3.1 EcoRI fragments.
coding sequence of human ~ (21). The 6.6-kb fragment, which represents a pseudogene (21), segregated with human chromosome 5 (Fig. 1A, lane 8). Since the initial signal was weak, the localization of this fragment was further confirmed with three other hybrid cell lines (not shown). The assignment of the o~1 fragments to chromosomes 5 and 6 and exclusion of all other chromosomes is summarized in Table 1. The fragments that segregated with human chromosome 6 were sublocalized to region
p21.1-q24.1 with three other hybrid cell lines (Fig. 1B). When the same rat o~I-GJ cDNA probe was hybridized with EcoRI-digested DNA from a panel of 13 mouse x rodent hybrids, a single major 7.6-kb mouse fragment (Fig. 2, lane 1), and a 7.8-kb rat fragment (Fig. 2, lane 6), were detected in addition to the 6.0-, 5.2-, and 4.5-kb hamster fragments (Fig. 2, lane 2). The presence of the 7.6-kb mouse signal was concordant with the presence of mouse chromosome 10 in the 13 hybrids
194
Hsieh et al.
Table 1. Discordancy Ratios (Percent) for Human Gap Junction Gene-Specific Restriction Fragments and Each Human Chromosome in 11-13 Chinese Hamster x Human Somatic Cell Hybrids Human chromosome Gene
1
2
3
4
5
6
7
8
9
10
GJA1 45 40 40 25 GJA1P 54 42 45 30 GJA3 55 45 40 50 GJB1 55 70 50 50 GJB2 55 45 40 50
50 0 20 80 20
0 38 45 45 45
67 60 67 56 67
60 46 27 64 27
60 42 36 60 36
36 44 18 67 56 55 30 38 45 46 31 55 62 46 45 78 64 0 44 36 40 64 56 45 78 33 55 50 45 78 64 0 44 36 40
studied; all other chromosomes were excluded by at least three discordant hybrids (Table 2). When the same probe was hybridized to a filter with BgllI-digested DNA, the major mouse fragment of 7.4 kb was also concordant with chromosome 10. Since no potymorphism was detected in inbred strains of mice with eight restriction enzymes, linkage mapping with recombinant inbred mouse strains was not possible, Assignment of ~3 Subunit Locus GJA3 in Human. When a rat a3 cDNA probe was hybridized to BamHI-digested DNA from human and mouse mapping panels, more than 20 fragments of varying intensities were
11 12
13 14
15
16 17 18 64 31 45 90 45
60 58 55 36 55
19 20 21 22 45 42 55 55 55
X
27 20 30 50 31 58 58 100 45 50 40 100 45 60 30 0 45 50 40 100
present in each lane. The most strongly hybridizing human fragment of 3.7 kb segregated with human chromosome 13 and another fragment (2.0 kb) that was also separable from hamster fragments segregated with human chromosome 1 (data not shown). When a human % genomic probe was hybridized to the same filters, a single 17.5-kb BamHI fragment was detected in human DNA and in hybrids retaining human chromosome 13 (Fig. 3, lanes 4-6, and 9) and was absent in hybrids not containing human chromosome 13 (lanes 3, 7, and 8). All other human chromosomes were ruled out by at least two discordant hybrids (Table 1). No
Fig. 2. Mapping of Gja-1 gene in mouse. Hybridization of the 32P-labeled rat cDNA probe to EcoRI-digested DNA from a panel of somatic cell hybrid cell lines. Lane 1, mouse control; lane 2, Chinese hamster; lanes 3 and 4, Chinese hamster x mouse hybrids; lane 5, rat x mouse hybrid; and lane 6, rat control. Lanes 3 and 5 are negative and lane 4 is positive for the 7.6-kb mouse signal.
Gap Junction Genes
195
Table 2. Discordancy Ratios (in Percent) for Mouse Gap Junction Gene-Specific Restriction Fragments and Each Mouse Chromosome in 11-15 Rodent x Mouse Somatic CelI Hybrids~ Mouse chromosome Gene
1
2
3
4
5
6
7
8
9
10
1I
12
13
14
15
16
17
18
19
X
Gja-t Gjb-1 Gjb-2
25 29 50
62 47 46
46 47 46
46 57 42
64 54 27
33 38 45
42 29 42
46 47 31
25 53 38
0 57 50
38 80 38
23 53 69
33 71 58
46 47 0
46 13 54
50 36 42
62 20 38
33 71 50
50 43 50
50 0 42
"Hybrids in which a particular chromosome was structurally rearranged or present in fewer than 10% of cells were excluded.
mouse signal and faint cross-hybridization with Chinese hamster DNA were detected with the human genomic probe, suggesting that it was derived from a highly diverged part of the gene. With the rat cDNA probe, one of the mouse fragments that could be distinguished from the numerous Chinese hamster fragments was concordant with mouse chromosome 5 in hybrid cell panels, and another mouse fragment was concordant with mouse chromosome 14 (data not shown). A mouse genomic probe will be required to determine which site represents the % coding gene. Assignment of ~ Subunit Locus GJB1 in Human and Mouse. When labeled human [31 cDNA was hybridized to EcoRI-digested DNA from the human mapping panel, a single 10.5-kb Chinese hamster fragment (Fig° 4A, lane 1) and a major 5.1-kb human
fragment (Fig. 4A, lane 2) were observed. The human-specific fragment was concordant with the X chromosome, and all other human chromosomes were excluded by three to nine discordant hybrids (Table 1). This result was confirmed with the rat x human hybrid cell line XXII-18A that contains the X as the only human chromosome (not shown). In the HindIII-digested DNA from an X chromosome regional mapping panel, the smallest overlapping region that was concordant with the human [3~signal is the proximal long arm, region cen-q22 (Fig. 4B). In BglII-digested DNA from the mouse mapping panel, a single 6.9-kb mouse fragment (Fig. 5, lane 1), several Chinese hamster fragments of 14.0, 8.5, 7.8, 4.9, and 3.8 kb (Fig. 5, lane 2), a major 7.3-kb and several weakly hybridizing rat bands (Fig. 5, lane 10) hybridized with the [3~ probe. The
Fig. 3. Assignment of GJA3 gene in human with a human genomic probe. Lane 2, human control; lane 1, Chinese hamster. The 11.0-kb band was very faint, but it was clearly visible after longer exposure and also in lanes 3-9, which contain Chinese hamster x human hybrid cell DNA. The 17.5-kb human fragment was observed only in hybrids retaining human chromosome 13 (lanes 4-6 and 9).
196
Hsieh et al.
il
] X
+ + +
_
GJB1
_
Fig. 4. Assignment of the GJB1 locus in human. (A) DNA from a panel of Chinese hamster x human somatic cell hybrid cell lines (lanes 3-8) and human (lane 2) and Chinese hamster (lane 1) controls was hybridized with 32p-labeled human [31-GJ cDNA probe, The 5,1-kb human fragment was clearly visible in lanes 3-6 and absent in lanes 7 and 8. (B) Regional assignment of the GJB1 gene on human X chromosome. Vertical bars represent regions of the X chromosome present in hybrid cells; the presence (+) or absence ( - ) of human 5.1-kb fragment is indicated below each vertical bar. Bracket indicates location of GJB1,
mouse signal was concordant with the mouse X chromosome. All other mouse chromosomes were excluded by at least two discordant hybrids (Table 2). With two hybrids containing complementary regions of the mouse X chromosome, the Gjb-1 gene was assigned to region D-F4 of mouse X chromo-
some distal to the T(X;16)16H translocation breakpoint (17). Assignment of [32Subunit Locus GJB2 in Human and Mouse. The rat [32 cDNA was used to probe EcoRI-digested DNA from human and mouse mapping panels. In Chinese hamster DNA, a strongly hybridiz-
Gap Junction Genes
197
Fig. 5. Assignmentof Gjb-1in mouse. Lane 1, mouse control; lane 2, Chinese hamster control; lanes 3-8, Chinese
hamster x mousehybridcell lines;lane 9, rat x mousehybridcell line;lane 10, rat control.Lanes3-7 are positivefor the mouse6.9-kbsignal;lanes 8 and 9 onlyhave the Chinese hamsteror rat fragments. ing 8.4-kb band and weakly hybridizing 21-, 18-, 6.7-, 5.3-, 4.2-, and 2.8-kb fragments were present (Fig. 6A, lane 1 and Fig. 6B, lane 2). Human fragments were 14 and 11.5 kb in size (Fig. 6A, lane 2), mouse bands were 21 and 11.5 kb (Fig. 6B, lane 1), and rat fragments were 21, 7, and 4.6 kb (Fig. 6B, lane 9). In hybrid cells, the two human fragments (Fig. 6A, lane 4) were concordant with human chromosome 13 (Table 1). The mouse 21 kb fiagment was not scored since it comigrated with Chinese hamster and rat fragments. The major ll.5-kb mouse band was present in all hybrids containing mouse chromosome 14 and absent in the hybrids not containing mouse chromosome 14. All other mouse chromosomes were ruled out by at least three discordant hybrids (Table 2). No polymorphism was found in five inbred mouse strains tested with eight enzymes. DISCUSSION
We have assigned four genes, GJA1, GJA3, GJB1, and GJB2, that encode four
different gap junction proteins (%, %, [31,and [3z) to various human (HSA) and mouse (MMU) chromosomes. In addition, a pseudogene, GJA1P, was mapped to HSA5, but no homolog of this was found in the mouse. It is clear from the Southern hybridization patterns of cDNA probes that additional related genes exist, e.g,, the GJA1 probe detected a fragment on the X chromosome not identical to GJB1 (data not shown) and the GJA3 cDNA probe detected many additional fragments of which only two could be mapped, one to HSA1 and one to MMU5. The relationships of the known coding genes mapped to the sizes of m R N A and deduced proteins are summarized in Table 3. Comparative mapping serves to define regions of conserved homology on human and mouse chromosomes. The GJA1 locus was assigned to HSA6, region p21.1-q24.1, and to MMU10. Two other genes, IFNGR1 encoding the interferon gamma receptor (binding unit), and the avian myeloblastosis viral oncogene homolog MYB, were previously assigned to HSA6, region q23-q24 and q22-q23, respectively, and to MMU10 (22,
198
Fig. 6. Chromosomal localization of the
Hsieh et al.
GJB2/Gjb-2 gene in human and mouse. (A) Southern filter of
EcoRI-digested DNA from Chinese hamster (lane 1), human (lane 2), and Chinese hamster x human hybrid cell lines (lanes 3-5) hybridized with labeled rat cDNA probe. Lane 4 is positive for both 14- and ll.5-kb human fragments and lanes 3 and 5 are negative. (B) The rat GJB2 cDNA probe was hybridized to EcoRl-digested DNA from rodent x mouse hybrid cell lines (lanes 3-7 are Chinese hamster x mouse and lane 8 is rat x mouse) and controls (lane 1, mouse; lane 2, Chinese hamster; and lane 9, rat). Hybrids in lanes 3 and 6 are positive for the ll.5-kb mouse signal, and those in lanes 4, 5, 7, and 8 are negative. 23). W e p r e d i c t t h a t t h e GJA1 locus is in t h e s a m e region, since g e n e s o n t h e s h o r t a r m a n d distal to q25 on t h e long a r m o f H S A 6 have all b e e n a s s i g n e d to M M U 1 7 , a n d g e n e s on t h e p r o x i m a l l o n g a r m o f H S A 6 h a v e b e e n m a p p e d to M M U 9 a n d M M U 4 (22, 23).
GJA1 is likely to b e i d e n t i c a l to t h e p o o r l y c h a r a c t e r i z e d gap j u n c t i o n - l i k e g e n e GJAL p r e v i o u s l y a s s i g n e d to 6 q 1 6 - q 2 2 by in situ h y b r i d i z a t i o n as r e p o r t e d in a n a b s t r a c t (24). T h e GJA1P p s e u d o g e n e o n c h r o m o s o m e 5 was a p p a r e n t l y n o t d e t e c t e d in this study.
Gap Junction Genes
199
Table 3. Localization of Genes for Gap Junction Proteins to Human and Mouse Chromosomes Locus
Gene
mRNA (kb)
Protein (kDa)
Human
Mouse
GJA1/Gja-1 GJAIP GJA3/Gja-3 GJA3L GJB1/Gjb-1 GJB2/Gjb-2
eq-GJ (Cx43)
3.6
43
6p21.1-q24.1
10
2.9
46
5 13
(14 or 5)
? 1.6 2.8
? 32 26
1 Xcen-q22 13
(?5) XD-F4 14
%-GJ (Cx46) ? 13~-GJ (Cx32) 132-GJ (Cx26)
Comparative mapping data also suggest that the mouse homolog of GJA1, Gja-1, may be near the Myb locus, which has been mapped to the proximal part of MMU10. The gene for the a 3 subunit, GJA3, was assigned to HSA13 with a human-specific genomic probe. However, the hybridization patterns observed with the rat cDNA probe indicated a second site on chromosome 1 (GJA3L). Mouse chromosome 14 was observed to segregate with one of the mouse fragments that hybridized with the rat % cDNA probe. This fragment may represent Gja-3, the mouse homolog of GJA3, because HSA13 is known to share a region of homology with MMU14 (22, 23). Other mouse GJA3 related fragments are present on MMU5. Once the mouse gene has been isolated, the map position of this locus in mouse will be clarified. Interestingly, we have found that the GJB2/Gjb-2 locus is also on HSA13 and MMU14. Three other genes, ESD (esterase D), RB1 (retinoblastoma), and HTR2 (serotonin receptor subtype 2), have been assigned to HSA13 and MMU14 (25, 26), and these three genes are close to each other in band 13q14 and by linkage analysis: cen-ESD(1.2cM)-HTR2-(2.8cM)-RBI-tel (26). Our Chinese hamster x human hybrid cell line 31-6AHAT that contains a small unidentified fragment of HSA13 is positive for both gap junction genes GJA3 and GJB2 and also for RB1 and HTR2 (26). Therefore, we postulate that GJB2/Gjb-2 is located close to these genes in both human and mouse. The GJB1 locus was assigned to the
human X chromosome, region cen-q22, and to the mouse X chromosome region D-F4. These regional assignments are consistent with the current comparative maps of human and murine X chromosomes (22, 27) and place the locus in segment 4 of the five identified conserved segments which in the mouse includes the interval between Tfm and Plp(jp) (22). Although the [31 and 132 gap junction proteins were first described in liver, and the eq gap junction gene product predominates in heart and skeletal muscle, it is now well established that individual gap junction proteins are expressed in many tissues in an overlapping pattern (13, 18, 28). Heterologous gap junctions could interact with each other and individual junctions may be made up of more than one channel type. These facts and the strong possibility that additional types may exist make it difficult to predict a phenotype for a mutation in a specific gap junction gene locus. In an attempt to obtain clues from the comparative gene map, we have noted that two mouse gene loci that can generate phenotypically identical mutations are located in regions that most likely contain GJP loci: dl (downless) on MMU10 in the region of Gja-1, and Ta (Tabby) on the X in the region of Gjb-1 (22). The underlying defects, however, involve growth abnormalities of epidermal derivatives (29). A region containing related genes could exist between MMU10 and X, as has been established for other autosomes. In summary, we present evidence that genes encoding the gap junction membrane
200
channels are dispersed over the human and mouse chromosomes. The exact number of gap junction genes remains unknown, but previously uncharacterized genes are present on chromosomes 1 and X. Analysis of these genes, together with their localization, will be of importance in understanding the evolution of the gap junction multigene family. ACKNOWLEDGMENTS This work was supported by NIH research grants GM26105 to U.F., GM37907 to N.B.G. and N.M.K., GM37904 to N.B.G., EY06884 to N.M.K., and by the Howard Hughes Medical Institute of which U.F. is an investigator and C.-L.H. was an associate. LITERATURE CITED 1, Revel, J.P., and Karnovsky, M.J. (1967). J. Cell Biol. 33:C7-C12. 2. Gilula, N.B., Reeves, O.R., and Steinbach, A. (1972). Nature 235:262-265. 3. Loewenstein, W.R. (1981). Physiol. Rev. 61:829913. 4. Makowski, L., Caspar, D.L.D., Phillips, W.C., and Goodenough, D.A. (1977). J. Cell Biol. 74:629-645. 5. Unwin, P.N.T., and Zampighi, G. (1980). Nature 283:545-549. 6. Kumar, N.M., and Gilula, N.B. (1986). J. Cell Biol. 103:767-776. 7. PauI, D.L. (1986).J. CellBiol. 103:123-134. 8. Beyer, E.C., Paul, D.L., and Goodenough, D.A. (1987).J. CellBiol. 105:2621-2629. 9. Gimlich, R.L., Kumar, N.M., and Gilula, N.B. (1988).J. CellBioL 107:1065-1073. 10. Zhang, J.T., and Nicholson, B. (1990). J. Cell BioL 109:3391-3402. 11. Ebihara, L.E., Beyer, E., Swenson, K.I., Paul, D.L., and Goodenough, D.A. (1989). Science 243:11941195. 12. Gimlich, R.L., Kumar, N.M., and Gilula, N.B. (1990). J. Cell Biol. 110:597-605.
Hsieh et al.
13. Beyer, E.C., Paul, D.L., and Goodenough, D.A. (1990). J. Membr. Biol. 116:187-194. 14. Milks, L.C., Kumar, N.M., Houghten, R., Unwin, N., and Gilula, N.B. (1988). E M B O J. 7:2967-2975. 1.5. Nicholson, B.J., and Zhang, J.T. (1988). Mol. Cell Biol. 7:207-218. 16. Guthrie, S.C., and Gilula, N.B. (1989). TINS 12:1.2-16. 17. Yang-Feng, T.L., DeGennaro, L.J., and Francke, U. (1986). Proc. Natl. Acad. Sci. U.S.A. 83:86798683. 17a. Wieacker, P., Davies, K.E., Cooke, H.J., Pearson, P.L., Williamson, R., Bhattacharya, S., Zirnmer, J., and Ropers, H.-H. (1984). Amer. J. Hum. Genet. 36:265-276. 18. Risek, B., Guthrie, S., Kumar, N., and Gilula, N.B. (1990). J. CellBiol. 110:269-282. 19. Feinberg, A., and Vogelstein, B. (1983). Anal Biochem. 132:6-13. 20. Hsieh, C.-L., Sturm, R., Herr, W., and Franeke, U. (1990). Genomics 6:666-672. 21. Fishman, G.I., Spray, D.C., and Leinwand, L.A. (1990). J. Cell BioL 111:589-598. 22. Searle, A.G., Peters, J., Lyon, M.F., Hall, J.G., Evans, E.P., Edwards, J.H., and Buckle, V.J. (1989).Ann. Hum. Genet. 53:89-140. 23. Lalley, P.A., Davisson, M.T., Graves, J.A.M., O'Brien, S.J., Womack, J.E., Roderick, T.H., Creau-Goldberg, N., Hillyard, A.L., Doolittle, D.P., and Rogers, J.A. (1989). Cytogenet. Cell Genet. 51:503-532. 24. Jayakar, P., and Rabin, M. (1989). Am. J. Hum. Genet. 45:(suppl.):A144. 25. Hsieh, C.-L., Lee, W.-H., Lee, E.Y.-H.P., Killary, A.M., Lalley, P.A., and Naylor, S.L. (1989). Somat. Cell Mol. Genet. 15:461-464. 26. Hsieh, C.-L., Bowcock, A.M., Farrer, L.A., Hebert, J.M., Huang, K.N., Cavalli-Sforza, L.L., Julius, D., and Francke, U. (1990). Somat. Cell Mol. Genet. 16:567-574. 27. Francke, U., Darras, B.T., Zander, N.F., and Kilimann, M.W. (1989). Am. J. Hum. Genet. 45:272-282. 28. Dermietzel, R., Traub, O., Hwang, T.K., Beyer, E., Bennett, M.V.L., Spray, D.C., and Willecke, K. (1989). Proc. Natl. Acad. Sci. U.S.A. 86:1014810152. 29. Lyon, M.F., and Searle, A.G. (eds.) (1989). Genetic Variants and Strains of the Laboratory Mouse, 2nd ed. (Oxford Univ. Press).
