B Lineage-Specific Interactions of an Innnunogiobulin Enhancer with Cellular Factors in vivo STOR
Anne Ephrussi, George M. Church, Susumu Tonegawa, Walter Gilbert
Science, New Series, Volume 227, Issue 4683 (Jan. 11, 1985), 134-140.
Stable URL: http://links.jstor.org/sici?sici=0036-8075%2819850111%293%3A227%3A4683%3CI34%3ABLIOAl%3E2.0.CO%3B2-I
Your use of the JSTOR archive indicates your acceptance of JSTOR' s Terms and Conditions of Use, available at http://www.jstor.org/aboutiterms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
Science is published by American Association for the Advancement of Science. Please contact the publisher for further permissions regarding the use of this work. Publisher contact information may be obtained at http://www .j stor .org/journals/aaas.html.
Science ©1985 American Association for the Advancement of Science
JSTOR and the JSTOR logo are trademarks of JSTOR, and are Registered in the U.S. Patent and Trademark Office. For more information on JSTOR [email protected].
©2002JSTOR
http://www .j stor.org/ Sun Jul 21 08:58: 162002 leaving a trace, but by measuring all of jets of hadrons that follow from the the Wand Z bosons. This exploit closed the other particles in the event, the pres quark-gluon substructure of protons and the 50-year-old quest for a clarification ence of neutrinos could be deduced by antiprotons. By the end of 1982, enough of the weak force, but the technique that imbalances of energy and momentum. data had accumulated for signs of the evolved opens a new phase of research Next, Rubbia used mUltiple detection long sought W's to be seen. The signa in elementary particle physics. Hence techniques along the paths of all particles ture was quite clear; each group of inves forth, all new colliding beam accelera so that electrons, photons, and muons tigators had a few events containing a tors and detectors wil1look back on the could be identified with high reliability. very energetic electron produced at large model pioneered by the 1984 Nobellau These particles are also harbingers of angles with respect to the colliding beam reates in physics. new physics. Such considerations were direction, but with no other visible parti LEON M. LEDERMAN essential because only about one W or Z cles to balance the electron's transverse RoY F. SCHWITTERS particle was expected for every hundred momentum. In early 1983, both groups million proton-antiproton interactions. published their findings, which were Leon M. Lederman is director of Fer As in the case of the SPS-AA com fully consistent with predicted properties mi National Accelerator Laboratory, plex, VA-l was constructed in an amaz of the W boson. VA-l had six events (6), Batavia, Illinois 60510. Roy F. Schwit ingly short time, and it began to take data whereas VA-2 presented four events (7). ters is co-manager of the Collider Detec in 1981. A second major detector known Soon thereafter, both groups obtained tor Facility at the Fermi National Accel as VA-2 was also constructed during this equally convincing evidence for the Zoo erator Laboratory and professor ofphys period. It was built by a collaboration of In this case, the production rates are ics at Harvard University. approximately 60 physicists from six Eu expected to be lower than those for the ropean institutions, under the leadership W, but the experimental signatures are References and Notes of Pierre Darriulat and Luigi DiLella, easier to interpret. By the end of 1983, 1. H. Yukawa, Proc.-Math. Soc. Jpn. 17, 48 both of CERN. The VA-2 is somewhat Rubbia's group had collected some 50 W (1935). 2. S. Coleman, Science 206, 1290 (1979). simpler than the VA-l apparatus, but it events and five ZO events, and the ex 3. G. I. Budker, At. Energ. 22, 246 (1966). is used to deal with many of the same perimenters were using the data to learn 4. S. Van der Meer, "Stochastic damping of beta tron oscillations in the ISR CERN" ISR po-n- physics questions. more about the nature of the production 31 [see also Phys. Rep. 58, 73 (1980)]. Both the VA-l and VA-2 experiments process. Given the previous successes of 5. D. B. Cline, P. Mcintyre, F. Mills, C. Rubbia, Fermilab Report TM 687 (1976). resulted in a number of important obser the unified electro-weak theory, these 6. UA-1 Collaboration, Phys. Lett. B 122, 103 (1983); ibid. 126, 398 (1983). vations during the early running. Most results were immediately accepted as 7. UA-2 Collaboration, ibid. 122,476 (1983); ibid. notable was the detection of very clean conclusive evidence for the existence of 129, 130 (1983).
RESEARCH ARTICLE independent of both the position and the orientation of the fragment (3). A smaller fragment, 307 base pairs in length (Fig. 1, base pairs 376 to 683) is capable of enhancing transcription in myeloma cells B Lineage-Specific Interactions of an from heterologous promoters and genes such as ~-globin and SV 40 T antigen Immunoglobulin Enhancer with Cellular (4). How enhancers work is not known. Factors in Vivo Although viral transcription enhancers function in most cells regardless of spe Anne Ephrussi, George M. Church cies or tissue of origin, they are most active in cells that are the virus's natural Susumu Tonegawa, Walter Gilbert hosts (6-8). The immunoglobulin heavy chain enhancer has a somewhat more restricted tissue specificity and functions most efficiently in cells of the B lineage Enhancers are DNA sequences that Analysis of deletions showed that re (3, 4). It seems likely that enhancer ac increase the level of transcription from moval of a portion of the intron of the tivity is related to and dependent on some promoters when placed within sev immunoglobulin 'Y2b gene greatly reduces specific cellular factors. In the case of an eral kilobases of them (1, 2). Mouse the level of transcription of the gene enhancer showing tissue specificity one immunoglobulin heavy chain genes con upon its transfection into myeloma cells might expect to find proteins binding to tain a tissue-specific enhancer, which is (3). Further experiments revealed that the enhancer only in cells in which the located in the intron separating the V restitution of a 990 base-pair fragment of enhancer is active. region J segment and the constant region DNA from the intron to these deletions Proteins that interact with specific se (Fig. 1) (3-5). The immunoglobulin restores high level transcription in myel quences can prevent or increase the heavy chain genes are extensively tran oma cells and that this restoration is methylation by dimethyl sulfate (DMS) scribed in myeloma cells. Cloned copies of these genes are also efficiently tran Anne Ephrussi is a graduate student and Susumu Tonegawa is professor at the Center for Cancer Research scribed when transfected into cells of the and Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139. George M. Church was a sCientist at Biogen and is now a postdoctoral fellow in the Department of Anatomy, B lineage but not in non-B cells such as L University of California at San Francisco, San Francisco 94143. Walter Gilbert is principal executive officer cells, HeLa cells, or lung cells (3, 4). at Biogen Research Corp., 14 Cambridge Center, Cambridge, Massachusetts 02142.
134 SCIENCE, VOL. 227 of specific guanine (G) residues at the N-7 position in the major groove (9). We Abstract. The mouse heavy chain immunoglobulin gene contains a tissue-specific have used DMS as a chemical probe to enhancer. The enhancer andflanking sequences were studied in vivo by carrying out observe the G residues in and around the dimethyl sulfate protection experiments on living cells, in combination with genomic heavy chain enhancer in the DNA of sequencing. Relative to reactions on naked DNA, there are changes (protections and living cells of diverse origins. The DMS enhancements) in the reactivity of guanine residues to dimethyl sulfate within the penetrates cells and methylates the ac enhancer sequence in myeloma, B, and early B cells, whereas virtually no alterations cessible G residues. Genomic sequenc appear in cells ofnon-B lineage. Most of the affected residues are infour clusters, in ing (10) can display the state of methyl sequences homologous to the octamer 5'eAGGTGGe 3' (e, cytosine; A, adenine; G. ation of the G residues in extracted guanine; T, thymine). The alterations in the pattern of G reactivity are consistent DNA. We compared the patterns of G with the tissue-specific binding of molecules to the mouse immunoglobulin heavy reactivity to DMS in several different chain enhancer. types of cells, searching for alterations in the intensity of specific G bands that might indicate that sequence-specific proteins were bound. In cells of the B While guanine residues in single position (9, 13). To determine whether lineage but not in cells of non-B lineage, stranded DNA are equally reactive to sequence-specific binding was occur the DNA of the immunoglobulin heavy DMS, in double-stranded DNA they ring, we looked for cases in which a G chain enhancer exhibited protections have distinctive reactivities. [An exam was less reactive than it was in naked and enhancements in its G pattern. The ple of this is the systematic underreac DNA (protected), equally reactive (no affected G residues were mainly in four tion of the middle G in the sequence 5' effect), or more reactive than in naked clusters, in and around sequences ho GGA 3' in double-stranded DNA (12).] DNA (enhanced). If a protein is binding, mologous to 5'CAGGTGGC 3' (C, cyto Moreover, in living cells, the double we can expect to find several such pro sine; A, adenine; T, thymine). stranded DNA may have proteins bound tections and enhancements over a short Preliminary considerations. We treat to it which could affect the reactivity of stretch of bases. ed whole cells briefly with DMS, to the G residues. The N-7 position of gua Control experiments in which cells partially methylate the purines (11). We nine that is methylated by DMS lies in were not treated with DMS showed that then extracted the DNA, digested it to the major groove; proteins that come the patterns observed are not due to the completion with a restriction enzyme close to this N-7 atom can prevent or existence of cell-specific endonucleases that cleaves about 100 bp from the begin increase methylation by DMS at that or depurination; controls without piperi- ning of the region being examined, and treated it with piperidine to cause cleav age at all of the methylated G residues. o 50 100 150 200 kb I I I After electrophoretic migration through a denaturing gel and electrophoretic VH 0 JH P. 8 Y3 Y1 Y2b Y2a € a' transfer to a nylon membrane, the DNA was cross-linked to the membrane by 1f1U~~11\lt1-D1-1---[}-I~~---ot-o-I-oI-DI- treatment with ultraviolet light. The membrane was probed with a short, sin 0H / J H EH SH CP.s gle-stranded, highly radioactive probe 1 2 3 4 1 2 ,3 4 made by polymerization across an M13 clone of a 100 to 150 nucleotide DNA fragment that ends at the chosen restric ~~ IIII tion site. This effectively end-labels the hybridizing DNA and displays a G-spe Xbal Pvu II Ode I Hinf I EcoRI Xba I cific sequence. Four probes were used. ~/I 7 ~ The use of the Eco 5' lower probe (see 1 382 518 566 683 992 legend to Fig. 1) and DNA cut with -1 Eco RI is equivalent to end-labeling all < )< .. hybridizable Eco RI-cleaved genomic Eco 5' lower Eco 3' lower fragments at the 5' end of the lower \ \ strand. Hybridization with the Hinf 5', <; Ode 3', Pvu 5' upper Hinf 5', Eco 3' upper Eco 3' upper probe to a Hinf I digest " gives the same result as would end Fig. 1. Scheme of the region examined and of the probes used. The immunoglobulin heavy chain proteins are encoded in several gene segments (black bars). After V-D-J rearrangement (28), a labeling the 5' end of all the upper-strand portion common to all productive rearrangements remains, namely, the intron separating J4 and Hinf I fragments at position 566 or alter elL. Within this intron are located the heavy chain enhancer (EH; white circle) and the IL switch natively, in an Eco RI digest, to end region (SH; bar). The enhancer is located on a 992-bp Xba I-Xba I fragment whose sequence is labeling the 3' end of the upper strand shown in Fig. 5. The probes (arrows) have the following nomenclature: name of restriction fragments at position 683. The four enzyme with the 5' or 3' end designation and whether it is complementary to the upper or lower strand (for example Hinf 5' upper). Arrows pointing to the left indicate that the probe is specific probes enabled us to examine the reac for upper-strand sequences, and those pointing to the right, for lower-strand sequences. For tivities to DMS of all the G residues instance, Eco 5' lower indicates that using this probe to hybridize with DNA cut with Eco RI between positions 24 and 997 on the will be equivalent to end-labeling the 5' end of lower strand genomic DNA fragments at the upper strand and, on the lower strand, Eco RI (683) site. The single-stranded DNA probes were synthesized as described (10). Hinf 1- those between 275 and 574 and 818 and Eco RI (566 to 683; Eco 5' lower, Hinf 5' and Eco 3' upper probes) and Eco RI-Xba I (683 to 992; Eco 3' lower) fragments were cloned into the M13mp8 (29) Sma I site as was the Pvu 11- 951. We first describe the upper-strand Bst NI (382-513; Dde 3' and Pvu 5' upper). The Bst NI site is 5 bp to the 5' of the Dde I site and patterns, then those of the lower strand. close enough to use the Pvu II-Bst NI fragment to probe Dde I cut DNA.
11 JANUARY 1985 135 '0 -J Fig. 2 (left). B lineage-specific protections 'E W and enhancements on the upper strand: 560- (0 -J ~ -J ro W I 347. The pattern of reactivity of G residues to -J 136 SCIENCE, VOL. 227 dine rule out the possibility that the cells (with one possible exception) are shown). Figure 4 shows results obtained patterns are due to tissue-specific DMS identical at all positions 3' to 545 (data with the Eco 5' lower probe (Fig. 1); the activated nucleases. not shown). DNA digested with Pvu II G residues at 534,533,531, and 523 are We compared two categories of cells. and probed with the Pvu 5' upper probe protected, and that at 530 is enhanced in The first consisted of immortalized cells reveals the G's between 480 and 750; a cells of the B lineage. A second cluster of ofthe B lineage that express their immu Hinf I digest probed with the Hinf 5' G's is protected in these cells, at 413, noglobulin genes, namely, two myelo upper probe reveals those between 670 408, 401, 399, 358, and 357. No other mas (J558L and MPCll), a B lymphoma and 992 (Fig. 1). Over this entire 500-bp changes appear in the stretch of DNA (A20-2J), and an Abelson virus-trans region all the differentiated cells exam between 275 and 575. formed early B cell (RAW8). The second ined had identical patterns of G reactiv Outside the enhancer region, we ex was composed of non-B cells: a Friend ity to DMS in vivo. amined some additional 130 bp with the virus mouse erythroleukaemia line, The G at 708 is protected in all cells, Eco 3' lower probe on Eco RI digested (MEL) and two T lymphomas (S49 and with one exception. In one experiment, DNA's. No alterations in the G pattern BW5147). All of the B-lineage cells pro the embryonal carcinoma line PSA-l (18) appeared between 818 and 951 (data not duce some form of immunoglobulin was used and showed a pattern indistin shown). In all cases the whole cell pat RNA (14-16). Little or no immunoglob guishable from that of naked DNA at all terns were identical and were indistin ulin RNA has been detected in the non-B positions including a visible band at 708. guishable from naked DNA controls. cells (16, 17). The 708 G residue is not an intrinsically Altered G reactivity occurs in con Changes in G reactivity are B lineage weak G residue: it appears as a strong served sequences within the enchancer. specific. We examined nearly 1000 bases band in all the naked DNA's. In RA W8, On both strands, more than 30 G resi of DNA for alterations in reactivity of G A20-2J, J558L, and MPC11 it is com dues are protected or enhanced in cells residues with DMS and found 28 protec pletely obliterated, as it is in BW5147 ofthe B lineage (Fig. 5) within a region of tions and four enhancements. These al- and MEL cells. 970 bp on the top strand and 430 bp on . terations are specific to cells of the B The same strategy was used to exam the lower. These tissue-specific effects lineage and all are localized almost pre ine the lower strand in and around the are all localized within a 193-bp stretch cisely to the region defined by transfec mouse heavy chain enhancer (data not of DNA in the middle of the 992-bp tion experiments as the enhancer (3, 4). region. Most of the protections and en We first examined DNA from cells hancements occur in four clusters locat treated with DMS in vivo using the iu ed at sequences that share homology to Eco 3' upper probe (Fig. 1). The se ....i! the sequence 5'CAGGTGGC 3'. Three quence is most distinct beginning around ...Ico ...Ico of the four clusters match this sequence < 10 ...I 10 the Hinf I site and extending 5' from it Z 10 W 10 at seven out ofeight positions; the other Q .., .., (Fig. 2). Cells of the B lineage show a G ::i matches at six out of eight positions (Fig. pattern of reactivity distinct from the 6). Three of the clusters have the pattern of non-B cells and of naked 513 5'CAGGTGGC 3' homologous sequence DNA. In B cells (Fig. 2, lanes b and g), on the top strand; in the fourth cluster, it G's at positions 545, 536, 507, 494, 421, is found on the bottom strand (Fig. 5). and 411 are protected; G412 and G419 Within these sequences on the top have enhanced reactivities to DMS. In strand, the G's at positions 3, 4, and 6 some experiments G460 and G458 were are protected; the G at position 7 is protected. There is a strong enhance either protected, enhanced, or normal. ment at the 406-407 band, which is fol On the lower strand Gland G8 are lowed by a stretch of protected G resi protected. We suspect that these se dues extending to 347. There are no quences may represent recognition sites differences in relative G reactivity be for enhancer binding proteins. Given the tween DNA's from any cells and naked 530 - similarity of these sequences, the bind DNA 5' to 347. 531 - ing proteins might be identical. The stretch from 407 to 347 was exam 533 - The reason residue G7, in the octamer, ined by means of Dde I digested DNA 534 - is protected in one case, normal in the and the Dde 3' upper probe (Fig. 1). This second, and enhanced in the third and region is protected exclusively incells of fourth, may be that the positioning of a b c d the B lineage. In those cells, the protect binding molecules is slightly different in ed region begins with the strong en Fig. 4. B lineage-specific protections and each case because of variations in the hancement at 407 (Fig. 3). The G at enhancements on the lower strand. The mem brane shown in Fig. 2 (a, b, and c) was DNA sequence. An amino acid residue position 406 is protected as are residues reprobed with the Eco 5' lower probe. G in one position could protect a G from 404, 399, 387, 384, 381, 359, 356, 355, residues at positions 534, 533, 531, and 523 reacting with DMS while a slight shift in and 352. In some experiments G391 was are protected, and G530 is enhanced in myelo its position could expose the G and allow protected. G388 is a weakly reactive G ma cells and nuclei (b and d) compared with it to react and perhaps even be enhanced the non-B-cell line (c). Lane a shows dena (the middle G of a 5'GGC 3' sequence) tured plasmid DNA (Fig. 2). At the top of the because of the proximity of the protein. 'and may, in some experiments, be pro-· figure, two closely spaced bands are visible in This happens in the binding of phage tected. From G347 to G24, the pattern in (a), only one of which is visible in (b, c, and lambda repressor to the operator site B cells is indistinguishable from the pat d). These closely spaced bands are part of a ORI (19), where the nearly symmetrical tern in non-B lineage cells or in naked 5'GGA 3' stretch (see Fig. 5, 511 to 513, lower strand) in which the middle G is underreacted site ORI is divided into two half-sites, DNA. in double-stranded DNA. The sequence of each of which is recognized by one sub The patterns of B-lineage and non-B this region is shown in Fig. 5. unit of the repressor dimer. Since the 11 JANUARY 1985 137 center of symmetry passes through the the same contacts can be seen in living clusters, is located in the sequence central base pair, the repressor mono cells, and (ii) that in these in vivo experi TAATTTGCAT. This sequence is ho mers do not bind with perfect symmetry; ments a further set of protections ap mologous to one of the two short se a G that is protected in one half-site, is in pears, mostly upstream of sites 3 and 4. quences found upstream of all immuno the other enhanced. The additional contacts to sites 1 and 2 globulin variable region genes and within In a companion study, where we ex could reflect the binding of the same a region, as defined by transfection ex amined the heavy chain enhancer in vitro factor to these more distant homologies periments, that is essential for correct in nuclei (20), we compared B cell to L or could reflect the binding of other transcription of those genes (21). and Friend cell nuclei. We demonstrated tissue-specific factors to the region, a Altered G reactivity correlates with tissue-specific contacts, in the enhancer binding not reflected in vitro. enhancer function. By treating whole region, to two sequences inverted with Some of the protections and enhance cells in vivo with dimethyl sulfate and respect to one another and homologous ments that we see are outside the four analyzing their DNA by genomic se to GCCAGGTGGC (sites 3 and 4). In the octamer sequences. One protection quencing, we have shown tissue-specific study presented here, we show (i) that (G545), which maps outside of these contact of factors to the mouse immuno- Xba I tctagagaggtctggtggagcctGcaaaaGtccaGctttcaaaGGaacacaGaaGtatGtGtatGGaatattaGaaGatGttGcttttactcttaaGttG ------+------+------+------+------+------+------+------+------+------+ agatctctccagaccacctcggacgttttcaggtcgaaagtttccttgtgtcttcatacacataccttat.aatcttctacaacgaaaatgagaattcaac 101 GttcctaGGaaaaataGttaaatactGtGactttaaaatGtGaGaGGGttttcaaGtactcatttttttaaatGtccaaaatttttGtcaatcaatttGa ------+-----.----+------+------+------+------+------+------+------+------+ caaggatcctttttatcaatttatgacactgaaattttacactctcccaaaagttcatgagtaaaaaaatttacaggttttaaaaacagttagttaaact 201 GGtcttGtttGtGtaGaactGacattacttaaaGtttaaccGaGGaatGGGaGtGaGGctctctcataccctatccaGaactGacttttaacaataataa ------+------+------+------+------+------~-+------+------+------+------+ ccagaacaaacacatcttgactgtaatgaatttcaaattggctccttaccctcactccgagagagtatgggataGGtcttGactGaaaattGttattatt Pst I 301 0 00 0 0 0 0 0 0 0 attaaGtttaaaatatttttaaatGaattGaGcaatGttGaGttGaGtcaaGatGGccGatcaGaaccaGaacacctGcaGcaGctGGcaGGaaGcaGGt ------+------+------+------+------+------+------+------+------+------+ taattcaaattttataaaaatttacttaactcGttacaactcaactcaGttctaccGGctaGtcttGGccttGtGGacGtcGtcGaccGtccttcGtcca 00 0 o O. O' • 0 0 catGtGGcaaGGctatttGGGGaaGGGaaaataaaaccactaGGtaaacttGtaGctGtGGtttGaaGaaGtGGttttGaaacactctGtccaGccccac ------+------+------+------+------+------+------+------+------~------+ GtacaccGttccGataaaccccttcccttttattttGGtGatccatttGaacatcGacaccaaacttcttcaccaaaactttGtGaGacaGGtcGGGGtG 000 501 o o o caaaccGaaaGtccaGGctGaGcaaaacaccacctGGGtaatttGcatttctaaaataaGttGaGGattcaGccGaaactGGaGaGGtcctcttttaact ------+------+------+------+------+------+------+------+------+------+ GtttGGctttcaGGtccGactcGttttGtGGtGGacccattaaacGtaaaGattttattcaactcctaaGtcGGctttgacctctccaggagaaaattga o .0 00 601 Eco RI tattGaGttcaaccttttaattttaGcttGaGtaGttctaGtttccccaaacttaaGtttatcGacttctaaaatGtatttaGaattcattttcaaaatt ------+------+------+------+------+------+------T------+------+------+ ataactcaagttggaaaattaaaatcgaactcatcaagatcaaaggggtttgaattcaaatagctgaagattttacataaatcttaagtaaaagttttaa 701 X aGGttatGtaaGaaattGaaGGactttaGtGtctttaatttctaatatatttaGaaaacttcttaaaattactctattattcttccctctGattattGGt ------+------+------+------+------+------+------+------+------+------+ tccaatacattctttaacttcctgaaatcacagaaattaaagattatataaatcttttgaagaattttaatgagataataagaagggagactaataacca 801 ctccattcaattattttccaatacccGaaGtctttacaGtGactttGttcatGatcttttttaGttGtttGttttGccttactattaaGactttGacatt ------+------+------+------+------+------+------+------+------+------+ gaggtaagttaataaaaGGttatGGGcttcaGaaatGtcactGaaacaaGtactaGaaaaaatcaacaaacaaaacGGaatGataattctGaaactGtaa 901 Xba I ctGGtcaaaacGGcttcacaaatctttttcaaGaccactttctGaGtattcattttaGGaGaaatattttttttttaaatGaatGcaattatctaGa ------+------+------+------+------+------+------+------+------+------997 GaccaGttttGccGaaGtGtttaGaaaaaGttctGGtGaaaGactcataaGtaaaatcctctttataaaaaaaaaatttacttacgttaatagatct Fig. 5. Summary of the protections and enhancements. The complete sequence of the Xba I-Xba I fragment which contains the mouse immunoglobulin heavy chain enhancer is shown. Residues whose reactivities to DMS have been examined are indicated by capital letters: G residues 14 through 996 on the top strand and on the lower strand, G residues 275 through 574 and 818 through 951. A fragment spanning bases 376 to 683 has approximately full enhancer activity (4). B lineage-specific protections (0) and enhancements (el are indicated above upper strand and below lower strand residues. Sequences homologous to those found upstream of all immunoglobulin variable region gene segments (21) are located at 451 to 465 and 539 to 548. The protection at 545 is within one of these sequences. A cross at 708 marks the G which is protected in all differentiated cells examined. 138 SCIENCE, VOL. 227 globulin heavy chain enhancer. Almost cells make no detectable immunoglob pIe consequence of transcription since all of the protections and' enhancements ulin mRNA (16). the amount of transcription in the cell we found are specific to cells of the B In most cells the enhancer exists in lines varies. A more likely interpretation lineage, which are committed to express two copies which, in the case of B is that these patterns are a consequence ing their immunoglobulin genes. All of lineage cells, are often located in differ of binding of proteins that cause the the alterations, on both strands, are lo ent regions of the genome. In J558L, for enhancer activation. calized within 200 bp that coincide al instance, one copy of the enhancer is There are other examples of activation most exactly with a fragment defined as appropriately located within the large of eukaryotic genes by binding proteins. sufficient for almost full enhancer activi intron of a productively rearranged IX In yeast, elements called upstream acti ty (3, 4). gene. The other copy has been translo vator sequences (U AS) are regulated by Banerji et al. (4) transfected into cells cated from the immunoglobulin heavy trans-acting positive regulatory proteins the enhancer linked to the SV 40 T anti chain locus to the c-myc locus on chro and have some properties similar to gen gene, and assayed T antigen-posi mosome 15 (22). Since some of the pro those of enhancers (23). The glucocorti tive cells. When they deleted sequences tections in these cells are complete, it is coid receptor, which can stimulate the 3' to 496, thereby eliminating site 4, likely that the reactivities of these G's in transcription of mouse mammary tumor activity was reduced. When they deleted both copies of the immunoglobulin gene virus (24, 25), binds in vitro (26) to a sequences 5' to 400, which include sites are being affected by the same factors in fragment of DNA called the glucocorti 1 and 2, the activity fell noticeably, while the same way. coid response element (GRE) (27). The deleting site 1 alone had no effect. Gillies In double-stranded DNA, as opposed GRE, reminiscent of enhancers, has et at. (3), assaying gpt transformants, to single-stranded DNA, the sequence been shown to confer hormone inducibil found that the fragment spanning 387 to 5'GGA 3' consistently yields an under ity to heterologous promoters in a dis 520 (lacking site 4) had half the activity, reacted middle G in the reaction with tance and orientation-independent fash while Banerji et al. (4) showed that the DMS (12). There are five 5'GGA 3' se ion. segment from 456 to 627 (lacking 1, 2, quences distributed within the enhancer. These experiments demonstrate tis and 3) was one-half to one-third as active The reactivities of the middle G's in all sue-specific contacts, most likely made as the full region, in stimulating ~-globin 5'GGA 3' sequences in vivo are charac by a binding protein, that correlate with synthesis. These results suggest that teristically weak, like those of double enhancer function. In general, the tissue these sites can independently provide stranded naked DNA. Thus the immuno specific action of the enhancer could be some enhancer activity. globulin enhancer is for the most part due to factors positively acting on it in B These distinct patterns, including double-stranded. cells or to negatively acting factors. in complete protections, are all in cells that These patterns are probably not a sim- non-B cells. Our findings support the make immunoglobulin messenger RNA first model. (mRNA). There is no striking difference in pattern between immortalized cells References and Notes believed to represent three distinct 1. J. Banelji, S. Rusconi, w. Schaffner, Cell 27, stages of B cell development in spite of 299 (1981). 2. P. Moreau et al., Nucleic Acids Res. 9, 6047 the fact that these cells make different (1981). amounts of immunoglobulin mRNA: my 3. S. D. Gillies, S. L. Morrison, V. T. Oi, S. 2 718 Tonegawa, Cell 33, 717 (1983). eloma cells express more immunoglob 4. J. Banelji, L. Olson, W. Schaffner, ibid., p. 729. ulin mRNA both from their native genes 5. M. S. Neuberger, EMBO J. 2, 1373 (1983). and from transfected copies than do pre 6. J. de Villiers and W. Schaffner, Nucleic Acids Res. 9, 6251 (1981). B cells. However, experiments measur 3 7/8 7. J. de Villiers, L. Olson, C. Tyndall, W. ing the intrinsic strength of the enhancer Schaffner, ibid. 10, 7965 (1982). 8. L. A. Laimins, G. Khoury, C. Gorman, B. in pre-B, B, and myeloma cells have not Howard, P. Gruss, Proc. Natl. Acad. Sci. U.S.A. 79, 6453 (1982). been reported. The enhancer may be 9. W. Gilbert, A. Maxam, A. D. Mirzabekov, in equally active in all cells of the B lineage 4 718 Control ofRibosome Synthesis, The Alfred Ben zon Symposium IX, N. O. Kjelgaard and O. regardless of their developmental stage; Maaloe, Eds. (Munksgaard, Copenhagen, 1976), this pattern may reflect general tissue pp. 139-148. 10. G. M. Church and W. Gilbert, Proc. Natl.Acad. specific events that make the enhancer Sci. U.S.A. 81, 1991 (1984). active and the gene available for tran 00 00 11. A. M. Maxam and W. Gilbert, Methods Enzy CAGGTGGC mol. 65, 499 (1980). saiption; subsequent changes might GTCCACCG 12. L. Johnsrud, Proc. Natl. Acad. Sci. U.S.A. 75, o then regulate the precise level of immu 5314 (1978). 34324443 13. R. T. Ogata and W. Gilbert, J. Mol. BioI. 132, noglobulin mRNA in cells at different 709 (1979). stages of development. 14. S. D. Gillies and S; Tonegawa, unpublished Fig. 6. Alignment of clustered protections and results. We do not see this pattern of contact 15. U. Schibler, K. B. Marcu, R. P. Perry, Cel/15, enhancements: the octamer CAGGTGGC. 1495 (1978). in T cells or Friend cells. In fact, the T The tissue-specific protections and enhance 16. D. J. Kemp, A. w. Harris, S. Cory, J. M. cell line S49 makes no detectable immu ments shown in Fig. 5 are located in clusters Adams, Proc. Natl. Acad. Sci. U.S.A. 77, 2876 within the immunoglobulin heavy chain en (1980). noglobulin mRNA (16) and BW5147 is 17. M. C. Zuniga, P. D'Eustachio, N. H. Ruddle, reported as making little (three mole hancer. The four clusters have sequence ho ibid. 79, 3015 (1982). mologies and are similar to the octamer 18. G. R. Martin, L. M. Wiley, I. Damjanov, Dev. cules per cell) (17) or no (16) detectable 5'CAGGTGGC 3'. When these G's occur in Bioi. 61, 230 (1977). 19. A. Johnson, B. J. Meyer, M. Ptashne, Proc. immunoglobulin RNA. Thus the immu the sequence, Gl, G3, G4, and G6 are protect Natl. Acad. Sci. U.S.A. 75, 1783 (1978). noglobulin enhancers in these T cells do ed, and G7 can be either protected, normal, or 20. G. M. Church, A. Ephrussi, W. Gilbert, S. not seem to be very active. Although the enhanced. Three clusters (350 to 357, 382 to Tonegawa, Nature (London), in press. 389, 401 to 408) are found in one orientation in 21. F. G. Falkner and H. G. Zachau, ibid. 310, 71 cell, line used in these experiments was (1984); T. G. Parslow, D. L. Blair, W. J. Mur the enhancer, the other (536 to 529) occurs in phy, D. K. Granner, Proc. Natl. Acad. Sci. not tested, Friend virus erythroleukemia the opposite orientation. U.S .A. 81, 2650 (1984). 11 JANUARY 1985 139 22. J. M. Adams, S. Gerondakis, E. Webb, L. M. 31. M. Potter, in Methods in Cancer Research, H. Spring Laboratory, Cold Spring Harbor, N.Y., Corcoran, S. Cory, Proc. Nat!. Acad. Sci. Busch, Ed. (Academic Press, New York, 1967). 1982). U.S.A. 80, 1982 (1983). 32. J. Kappler, J. White, D. Wegmann, E. Mustain, 39. A. Ephrussi and G. Church, unpublished obser 23. L. Guarente, Cell 36, 799 (1984). P. Marrack, Proc. Natl. Acad. Sci. U.S.A. 79, vations. 24. H. A. Young, T. Y. Shih, E. M. Scolnik, W. P. 3604 (1982). 40. We thank Anne Weber for the MEL line, Ther Parks, J. Virol. 21, 139 (1977). 33. P. Ralph, I. Nakoinz, W. C. Raschke, Biochem. eza Imanishi-Kari for MPCll, and Gail Martin 25. G. M. Ringold, K. R. Yamamoto, J. M. Bishop, Biophys. Res. Commun. 61, 1268 (1974). for PSA-l; and Roger Brent, Joseph Heilig, Ann H. E. Varmus, Proc. Nat!. Acad. Sci. U.S.A., 34. V. Volloch and D. Housman, J. Cell Bioi. 93, Hochschild, Nancy Hopkins, Mary Weiss, and 74, 2879 (1977). 390 (1982). Kai Zinn for their comments on the manuscript. 26. F. Payvar et al., Cell 35, 381 (1983). 35. R. Hyman and V. Stallings, J. Natl. Cancer Supported by American Cancer Society grant 27. V. L. Chandler, B. A. Maler, K. R. Yamamoto, Inst. 52, 429 (1974). AI-17879 (to S.T.), by grants CA14041 (a core ibid. 33, 489 (1983). 36. K. Horibata and A. W. Harris, Exp. Cell Res. grant to S. Luria) and GM-09541-22 (to W.G.) 28. S. Tonegawa, Nature (London) 302, 575 (1983). 60, 61 (1970). from the National Institutes of Health, a grant 29. J. Messing, Methods Enzymol. 101,20 (1983). 37. T. G. Easton, J. E. Valinsky, E. Reich, Cell 13, from Biogen N.V. (W.G.), and NIH training 30. V. T. Oi, S. L. Morrison, L. A. Herzenberg, P. 475 (1978). grant 2T 32 CA 09255 (A. E.). Berg, Proc. Natl. Acad. Sci. U.S.A. 80, 825 38. T. Maniatis, E. F. Fritsch, J. Sambrook, Molec (1983). ular Cloning; a Laboratory Manual (Cold 1 November 1984; accepted 5 December 1984 clones covering most of the LDL recep RESEARCH ARTICLE tor gene, which spans at least 60 kilo bases (kb) of DNA (10). The amino acid sequence of the LDL receptor was de duced from the nucleotide sequence of Mutation in LDL Receptor: its cDNA (9). This sequence, coupled with protein chemistry studies (8), re Alu-Alu Recombination Deletes Exons vealed that the normal receptor consists of at least five structural and functional Encoding Transmembrane and domains. Listed in order from the amino terminus, the domains are (i) a cysteine Cytoplasmic Domains rich region of 322 amino acids that con tains the binding site for LDL; (ii) a Mark A. Lehrman, Wolfgang J. Schneider, Thomas C. Sudhof region of ~350 amino acids that is ho mologous to the precursor for epidermal Michael S. Brown, Joseph L. Goldstein, David W. Russell growth factor; (iii) a serine- and threo nine-rich region of 48 amino acids that is the site for O-linked glycosylation; (iv) a stretch of 22 hydrophobic amino acids Approximately one person in every receptor was purified, and antibodies that spans the plasma membrane; and (v) 500 inherits one mutant gene for the low were generated. These advances led to a carboxyl terminal region of 50 amino density lipoprotein (LDL) receptor, a the demonstration that some mutant al acids that projects into the cytoplasm cell surface protein that carries the cho leles produce reduced amounts of recep (9). The mature messenger RNA lesterol transport protein LDL from tor protein, whereas others encode re (mRNA) for the receptor is unusual in plasma into cells by receptor-mediated ceptor proteins with a demonstrably al that its 3' untranslated region contains endocytosis in coated pits (1). When the tered structure (4). several copies of a repetitive DNA se receptors are defective, LDL cannot be Among the most informative muta quence of the Alu family (9). removed normally from plasma, it accu tions have been those in which the LDL We now report the use of our cDNA mulates to high levels, and premature receptor reaches the cell surface but and genomic clones to characterize a atherosclerosis ensues. Mutations in the does not cluster in coated pits, a defect mutation in the structural gene for the gene for the receptor are important clini that prevents the receptors from carrying LDL receptor in a family with FH. The cally because a single copy produces LDL into cells (5-7). Study of these index case is a young man (B.H.), here heterozygous familial hypercholesterol mutations demonstrated the essential after designated FH 274, who has all of emia (FH), a common cause of heart role of coated pits in receptor-mediated the clinical features of homozygous FH attacks in middle-aged people (1). About endocytosis (2, 5). As a result of kinetic (7). Previous functional studies revealed one in I million persons inherits two and genetic complementation studies, that cultured fibroblasts from FH 274 copies of a mutant LDL receptor gene. we proposed that mutations that block bound about one-third of the normal In such individuals, called FH homozy internalization reside in the gene for the amount of 125I-Iabeled LDL. However, gotes, plasma levels of LDL-cholesterol LDL receptor, and not in some other the receptors in FH 274 did not cluster in are even higher than in heterozygotes, gene necessary for receptors to be incor coated pits and hence did not transport and heart attacks occur before 20 years porated into coated pits (5, 7). their bound LDL into the cell (7). Thus, of age (1). Recently it has become possible to FH 274 was categorized functionally as Mutations in the LDL receptor are study the molecular genetics of these an "internalization-defective" mutation. important subjects of study because they mutations as a result of the cloning of a Studies of fibroblasts from the relatives provide insights into the general mecha full-length complementary DNA (cDNA) of FH 274 revealed that he had inherited nism of receptor-mediated endocytosis for the human LDL receptor (8, 9) and two different mutant alleles; the allele (2). These mutations were originally de the subsequent isolation of genomic encoding the internalization-defective fined through functional assays of LDL receptor activity in cultured fibroblasts Mark A. Lehrman and Thomas C. Siidhof are postdoctoral fellows, Wolfgang J. Schneider and David W. Russell are assistant professors, and Michael S. Brown and Joseph L. Goldstein are professors in the and freshly isolated lymphocytes from Department of Molecular Genetics, University of Texas Health Science Center at Dallas, Southwestern FH patients (3). Subsequently, the LDL Medical School, Dallas 75235. 140 SCIENCE, VOL. 227