Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 2 of 9
`
`© 1994 Oxford University Press
`The detailed characterisation of a 400 kb cosmid walk
`in the BRCA 1 region: identification and localisation of
`10 genes including a dual-specifici1ty phosphatase
`
`Human Moleculnr Genetics. 1994, Vol. 3, No. 11 1927-1934
`
`Karen A.Jones •, Donald M.Biack+, Melissa A. Brown, Beatrice L.GriHiths, Hans M,Nicolal, Julie A.Chambers,
`Marfsa Bonjardlm, Chun-Fang Xu, Marie Boyd2, Robert McFarlan~. Bernhard Korn3, Annemarfe Poustka3,
`Michael A.North1, Leo Schalkwyk1 , Hans Lehrach1 and Ellen Solomon
`Somatic Cell Genetics Laboratory and 1Genoms AnalysiS Laboratory, Imperial Cancer Researcll Fund. PO Box 124, Uncoln's Inn Fields, London
`WC2A JPX. 2Beatson Institute tor Cancer Reseaich, Garscube Estate, Swrtchback Road, Beatclen, Glasgow G61 1BD, UK and 3Deutsches
`Krebslorschungszentrum, lm Neuenheimef Feld 506, 69120 Heidelberg, Germany
`
`Rece1ved September 13, 1994, R8VIsed and Accepted September 22, 1994
`
`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`contain BRCAJ' , jndicating the presence of one or more rumour
`suppressor gen1es (14-2 1). However, searches for small regions
`of deletion andl other positional clues such as translocarions in
`sporadic or familial rumours have not been successful.
`With the aim of identifying the BRCAJ gene within these
`defmed limits, several tools have been developed, including high
`density genetic maps (22,23); chromosome 17 hybrid mapping
`panels (24-261) and fluorescence in situ hybridisation (FISH)
`data, for the physical localisation of polymorphic markers and
`genes spanning the region (27). These have enabled the assembly
`of incomplete y·east an:ificia\ chromosome ('{A C) contigs for the
`purpose of isoJ.ation and analysis of the genes housed within the
`BRCA 1' region (28). Unfortunately, problems of chimaerism,
`instability and rearrangements have made Y ACs a far from ideal
`substrate for
`isolating specific genes and constructing
`comprehensive transcript maps. This work describes identification
`of a substantial hole in our Y AC coverage which bas been filled
`by cosmid end .. walking . With the conversion of our YACs into
`overlapping cosmids, a complete ordered cosmid contig spanning
`over 400 kb fr1om me gene 1A1.3B to the polymorphic marker
`D17S78 has been obtained, together with a detailed long-range
`physical map .constructed by pulsed-field gel electrophoresis
`analysis. This has enabled the isolation and precise localisation
`of l 0 genes in the region and provides further insight into the
`characteristics of the chromosomal region surrounding BRCA 1.
`
`RESULTS
`Y AC isolation and anaJysis
`In order to clone genomic DNA for detailed characterisation of
`the BRCA 1 region, large insert Y AC clones were isolated with
`probes and PCR primers corresponding to both known genes and
`anonymous markers. Several clones were identified but in the
`region flanking the RNU21ocus, problems of inter~hromosomal
`chimaerisrn, r·earrangement and
`instability prevented
`the
`generation of a complete Y AC contig (fable 1). A sizeable
`deletion was noted in one YAC, 17-5-GJO which contained onJy
`the markers DI7S859 and DI7S855 but not 1Al.3B, RNU2 and
`other cosmid markers known to reside between them. In addition
`a 400 kb Y AC , 2606 which contained me markers D 175858
`and D 17S859, was unstable and became shortened by internal
`
`We have produced a detailed physical and tran(cid:173)
`scriptional map of a 400 kb region within the narrowest
`flanking markers known to contain the hereditary breast
`and ovarian susceptibility gene, BRCA 1. The approach
`described here has avoided the problems of chlmaer(cid:173)
`lsm, Instability and rearrangements commonly ob(cid:173)
`served In yeast artificial chromosomes by converting
`the YAC clones Into ordered chromosome 17-speclflc
`cosmld contlgs and joining these contlgs by cosmld
`end-walking. A detailed long-range restriction map
`provided a framework for the cosmld contfg assembly
`and further refines existing physical mapping data. We
`have used a combined approach towards the Isolation
`of the genes housed within these cosmlds. This has
`resulted In the Isolation and precise localisation of eight
`novel genes, Including a novel G protein and an
`endogenous retrovirus related to the HERV-K family,
`and the previously described dual-specificity VHR
`phosphatase and MOX1 homeobox genes.
`
`INTRODUCTION
`
`Breast cancer is predominantly a sporadic disease affecting about
`one in ten women in the Western world. One hereditary form
`of the disease, accounting for a few percent of all cases and
`characterised by early age of onset, has been linked to a
`susceptibility gene tenned BRCAI , on chromosome17q21 (1).
`This gene predisposes to both breast and ovarian cancer and it
`appears that almost all breast cancer families with ovarian cancer
`cases are linked to BRCAJ (2).
`Since the function of BRCAJ is unknown, efforts to clone it
`have relied upon a positional cloning strategy. Extensive
`characterisations of meiotic breakpoints in breast cancer families
`have succeeded in tuiiTOWing the region to 1.0- 1.5 Mb between
`the markers DI7S776 and D17S78 (3 -8). AdditionaJ fine
`mapping has been limited by the number of informative meioses
`available and complicated by the extensive genetic heterogeneity
`of the disease (9-l~). Both breast and ovarian rumours have
`shown extensive loss of heterozygosity in the region known to
`
`-To whom comspo1¥lence should be addressed
`+preseru address: Bes1son lnstiwc.e for Cancer Research. Glasgow, UK
`
`GeneDX 1034, pg. 1
`
`

`

`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 3 of 9
`
`1928 Human Molecular Genetics, 1994, Vol. 3, No. 1 I
`
`Table 1. Results of YAC arutlysis in the /A/.38- DI7S78 region
`
`YAC name
`
`l..ibi'IUY"
`
`Probe
`
`Size kb
`
`Feawresb
`
`106G4
`12H4
`AI67E6
`176812
`8196
`174AI2
`175GIO
`26D6
`B260EII
`172H5
`J7SH2
`109E5
`
`ICRF
`!CRF
`Wash
`CEPH17
`Wash
`CEPHl7
`CEPH1 7
`CEPH
`Wash
`CEPH 17
`CEPH17
`JCRF
`
`/AI
`!AI
`R.NU2
`R.NUl
`RNU2
`YAC 2H5e
`YAC 2H5
`PPY/pl31
`ppy
`YAC 5H2
`PPY
`PPY/p131
`
`700
`500
`200
`800
`150
`700
`nd
`250-400
`<100
`450
`350
`nd
`
`ullSillble
`chim.aeric
`chimaeric
`nd
`nd
`chimaeric
`cluma.eric and deleted
`unstable
`nd
`nd
`nd
`nd
`
`Markers
`S855
`
`/AI
`
`+
`+
`+
`
`+
`
`RNU2
`
`S858
`
`S859
`
`ppy
`
`p131
`
`+
`+
`+
`+
`
`+
`
`(+
`
`+
`+
`+)
`
`+
`+
`+
`+
`+
`
`+
`
`+
`+
`+
`
`Each YAC, indica~ in the first colwnn has been tested for each llllll'ker in the top row. Presence of a marker is indica~ as+. YAC 2606lOSt marurs DJ7S858
`a.nd D 175859 on deletion, as indicated by brackets.
`•tCRF, ICRF YAC library; Wash, WashingtOn School of Medicine library; CEPH17, CEPH chromosome 17 selected mega YAC library.
`bChimaerism was determined by in siru analysis of Y AC Alii PCR probes.
`eAlu PCR probe from YAC 172H5.
`
`deletion to 250 kb when subsequent preparations were made.
`None of the Y ACs isolated shared the RNU2 marker with any
`of the markers 0175858, 0175859, PPY or 017578 which
`indicated a gap in Y AC coverage in this region.
`
`oonlrom<ft
`..... --.,.1-mWDllDmmrr--
`\A\.38 RNU2
`
`YAC 12.H4
`
`-
`
`-
`
`-
`
`-
`
`- -....-----.-,.-,..----1~
`
`Cosmid isolation and initial contig construction
`To avoid analysis of other genomic regions present in the
`chimaeric Y AC clones, and to isolate non-deleted copies of the
`DNA that is unstable in the Y ACs, a gridded flow-sorted
`chromosome 17 cos mid library , generated at the ICRF (29) was
`screened with the available Y ACs. The resuJting pools of cosmid.s
`were arranged into conrigs on the basis of the results of PCR
`analyses and hybridisations of repeat free cosmid fragments to
`other members of !he cosmid pools (Fig. 1). These results were
`confirmed by Pstl restriction fragment size comparisons and A lu(cid:173)
`PCR fingerprinting (data not shown).
`The 12H4 Y AC identified over 200 strong positive clones.
`Most of these clones were due to the repeated array of U2 RNA
`genes at this locus(30), as revealed by hybridisation of a duplicate
`library filter with Rll/1.12. The 12H4-positive cosmid clones which
`were positive for 1Al.3B were revealed by an additional
`hybridisation with a 1Al.3B gene probe. Cosmid B0176 was
`positive for both JA1.3B and RNU2, indicating !hat these genes
`are less than 40 kb apan, based on the average length of a cosm.id
`insen. A weakJy hybridising RNU2 positive clone, El 132 was
`found to be negative for 1Al.3B and did not overlap with cosm.id
`B0176. Hybridisati.on ofthreeEcoRI repeat-free fragments from
`E ll 32 back to the cosmid library filters isolated two further
`cosmids, Cll98 and A0413 1, both of which failed to hybridise
`to the RNU2 probe. The cosmids E1132, Cll98 and A04f3 1
`therefore represent a walk away from the RNU2 locus in the
`opposite direction to JAJ.3B (Fig. 1). The location of cosmid
`C 1198 adjacent to the RNU2 locus was confliTTled by pulsed field
`gel electrophoresis (PFGE) analysis (data not shown).
`Positive cosmid clones identified by screening withY AC 2606
`prior to deletion, were analysed by PCR for the presence of
`markers 0 175858, D 17S859, 0 17S78 and the PPY gene. Cosm.id
`AO II 83 was posilive for both DI7S78 and PPYwh.ich confinns
`the finding that these markers are within 45 kb of one another
`(31) (Fig. I). Cosmids 011 181 and G0815 1 were both positive
`for D 17578 only and represent the most distally extending
`
`I'OOI.A
`
`roots
`
`Figure 1. Ordering of cos.mid pools generated by chromosome 17 cosmJd library
`scrccrungs using YACs 12H4 nnd 2606. Cosmds CII98/A04131 and
`G05149fPJJ7131\ extend the furthest into the gap belwcen the two cosmid pools.
`
`cosmid.s of the contig. Cosmids G05149 and 807136 were
`positive for both D 17S858 and D 17S859 and were shown to link
`up to the PPY and D 17578 cosmids by hybridisation experiments.
`Cosmid DOl 179 hybridised strongly with cosmid 0 12 160,
`009100, A01 183 and itself, but more weakly with cosmids
`D 1 198 and C071 0 I. The cosmids D I I 98 and C07 1 0 I in turn
`hybridised to cosmid 807136. The overlapping cosmid contig
`of pool B suggest the markers D 175858!0 17S859 are roughly
`60 kb away from PPY!D 17578.
`
`Pulsed field gel electrophoresis analysis of the region JA1.3B
`to PPYf017S78
`Analysis of the rare-<:uner restriction fragments hybridising to
`single-<:opy probes from each of the cosm.id pools orientated the
`cosmid contigs and identified the size of the gap between them.
`A PFGE Southern ftlter containing high molecular weighr
`genomic DNA from an immonalised B cell line digested with
`a number of rare cuner restriction endonucleases in single and
`double combinations was sequentially hybridised with single-<:apy
`probes from the region (see Materials and Methods) .
`Comparisons of the autoradiographs revealed that I A/. 38,
`RNU2 and 12E/Bd, a fragment of the D 17S858f0 1 7S85~positive
`cosmid, 005149, reside on the same 750 l<b Notl fragment and
`that IA1.3B and RNU2 share the same 380 kb Mini fragment,
`but different Nrul fragments (Fig. 2). The Nrul fragment which
`contains RNU2 also hybridises to 12E/Bd, giving the order
`
`GeneDX 1034, pg. 2
`
`

`

`Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 4 of 9
`
`1, IAI. l8 PMOIIIt
`
`2. i!NU2PROOE
`
`3. llE/Bd rROBE (017S858/J\S<J locus)
`
`Human Mo•lecular Genetics. 1994. Vol. 3. No. U 1929
`
`.., , •
`
`""
`I
`
`N·r &
`I I
`
`Ea BsNr
`Ill
`.
`IAIJB
`
`EaBsNr
`
`Ill 111 111111111111111111 11
`
`MBs
`
`IWU2
`
`cao -I. N
`
`M
`I !i<IKI> I
`I
`2. N
`M
`I
`
`) . N
`
`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`e.
`
`Eo
`I
`
`(MXEol
`I I
`
`-300Kb
`
`Sa
`I
`
`(MXEal
`I I
`
`N
`
`Nr
`
`N
`
`Nr
`
`Nr
`
`s. r;., N T
`s.
`I ~I I
`12E/ Bd
`
`I
`
`Bs e. N
`Bs
`I
`1 I
`1 .
`12E/Bd
`
`Nr
`M
`I
`
`1+2-+J..N
`M
`I
`
`Ho
`I
`
`NrBs
`I I
`
`Ea Bs Nr
`
`M S.
`
`•" 1 11111111111111 1111 11111
`
`11\J.JB
`
`RNUl
`
`Figure 2. Construction of a physical map from PFGE data using probes RNU2, IAJ.38 and 12EIB, a fragment of the D 17S85810 178859-positive cosmid, 005149
`(see Materials and Methods for probe details). Arrowed bands indicate the restriction fragments which hybridise to more than one probe. Restriction maps were
`constructed using the information provided by each hybridisation experiment. Superimposing the maps (I + 2+3) gives a rough estimation of the distance between
`12EfB and RNU2 as 300 kb. N. Norl ; M, Mlul; 13s, BssHU. Ea, Eogl: Nr. Nnd; (M) and (Ea). restriction sites showing paniaJ methylation ..
`
`1Ai.3B - RNU2-12E!Bd . The probes PPYand pl31 hybridised
`to a smaller 550 kb Notl fragment, which indicates that cosmid
`005149 is the most proximally extending cosmid of pool B (data
`not shown). Therefore, combining the cosmid contig analysis and
`the pulsed-field electrophoresis analysis, the order of the markers
`is cen-1AI.38-RNU2 - DI7S858/9-PPY-Dl7S78-tel.
`The detailed restriction map integrating all of the restrictiOJ
`fragment sizes indicated that the gap between the most distal!)
`extending cosm.id of pool A, cosm.id Cll98 and the mos
`proximally extending cosmid of pool B, G05149 wa:
`approximately 300 kb in the B cell line tested (Fig. 2).
`
`Generation of a complete cosmid walk between RNU2 anc
`D17S858
`In order to complete the genomic coverage of this region, tht
`cosmids extending the furthest into the gap were used to initiatt
`cosmid walks towards one another. The walking strateg)
`employed the dual opposed SP6 and T7 promoters flanking the
`cloning site in the cosmid vector, Lawrist 4 (D.Nizetic,
`unpublished dara) to create probes from each end of the cosm.id.
`Neighbouring cosmids were isolated when each end-specific
`probe was hybridised in tum to the cosmid library, as shown
`(Fig. 3). The walks proceeded proximally from cosmid 005149
`and distally from cosm.id C 1198 until they overlapped at cosmids
`C02179 and GI24 (Fig. 4B). The distance covered in the cosm.id
`walk correlates roughly with that suggested by the pulsed field
`electrophoresis analysis above. The ordering of the cosm.ids
`throughout the region JA/.38 to DI7S78 was confirmed by FISH
`analysis (32). This gave the order as CIJ98-G05149- D 11181
`and C 1198-0 124- 005149 and confirms that the cosmids are
`contiguous (Fig. 5 and data not shown). In addition, FISH
`
`SP6
`I77A
`~
`809121
`~r---------1~
`
`T1
`
`1!1132
`
`AO<IIll
`
`COSI2J
`
`COSI73
`
`809117 (SIP6) probe
`
`809127 ml Probe
`
`rlglll'\! 3. Hybridisax.ion of nboprobes from each end ofcosmid B09127toduplicate
`filters containing gridded cosmids from the chromosome 17 cosmid library. The
`17 riboprobe extemds the furthest into the gap as it has hybridi.sed with previously
`unidentified cosmids, COSI23 (5) and COSI73 (4), whilst the SP6 riboprobe
`hybridi.sed to previous cosmids in the walk, E1132 (2) and A04131 (3). Both
`riboprobes hybrid,ised to the ccsmid 809127 (I) from where they originated,
`serving as an inte,maJ positive con1rol.
`
`analysis on chromatin released from nuclei (33) confinned the
`overlap of the two cosmid walks at cosmids 0124 and C02179
`(data not shown).
`
`GeneDX 1034, pg. 3
`
`

`

`Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 5 of 9
`
`1930 Human Molecular Genetics, 1994, Vol. 3, No. 11
`
`lioLill -
`
`. •
`
`• 0
`
`0
`
`I
`I
`
`'
`
`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`' I
`
`i
`L
`
`..
`
`-
`
`Figure 4. Detailed cosmid clone and uanscript map of the region between RNU2 and DI7S78. (A) Localisation of tO genes (blue markers) on the cosmid contig
`is indicated by dashed venical lines. The localisation of exon-trapped products (filled squares) and sublibrary eDNA contigs (filled circles) is also given. Open squares
`indicates the position of EcoRI fragments in the cosmid contig which conll!in puU!cive CpG islands. (U) Cosmid walk between RNU2 and Dl8S78. Cosmids in red
`constiWte the minimal overlapping set use4 in the creation of eDNA sublibraries and exon-trapping. llile orientarion of the SP6 and T7 ends of these cosmids within
`the walk is shown. (C) Zoo blot results from conserved EcoRI fragments in the region which have resulted in the isolation of genes.
`
`Isolation and localisation of genes in the cosmid contig
`A minimal set of overlapping cosmids spanning the region
`RNU2-Dl7S858/Dl7S859 (see Fig. 4B) was used as the
`substrate for the isolation of eDNA clones using a number of
`techniques. They include direct eDNA selection (34,35),
`surveying for evolutionary conservation (36), exon trapping (37)
`and localisation of CpG islands (38,39). Clusters of eDNA clones
`were localised throughout the cosmid contig and Northern
`analyses confirmed which of them corresponded to separate gene
`loci (Fig. 6). The genes isolated include the dual specificity VHR
`phosphatase (40), the newly identified MOXJ gene (41), a novel A.
`GTP-binding protein (G protein) HAL64 (Black et al., in
`preparation), a locus for an endogenous retrovirus and six novel
`genes with no database sequence homologies to date.
`
`Direct eDNA selection. Overlapping eDNA contigs were
`identified at the gene loci indicated on the transcript map (Fig.
`4A). Members of the eDNA contigs were used as probes to isolate
`fuU length eDNA clones from random plated eDNA libraries and
`sequencing of these clones has identified homologies to other
`known genes. Nearly all the genes isolated in this region were
`represented in the subtracted eDNA libraries with the exception B.
`of the BCCB gene, discovered by a search for evolutionarily
`conserved sequences and four exon-trapped products.
`
`Evolurio!Ulry conservarion. Expression-independent methods were
`also employed to identify genomic fragments corresponding to
`cDNAs within the BRCAJ region. All of the EcoRI fragments
`from the minimal set of overlapping cosmids were hybridised
`
`c
`
`Figure S. FISH analysis of the cosmid contig spanning the region from RNU2
`to DI7S78. (A) Cosmid Cll98 (red), Dl I 181 (green) and G05159 (red) give
`the order red-red-green. (B) Cosmid Cll98 {red), 011181 (red) and G05 149
`(green) give the order red-green-red. (C) Cosmid Cl 198 {red), Gl24 (green)
`and G05149 (gre.:n) give the order red-green- green. These results confi.rm that
`the walk is contiguious and indicate that the cosmid order given by the walk is
`correct.
`
`GeneDX 1034, pg. 4
`
`

`

`Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 6 of 9
`
`Human Molecular Genetics, 1994, Vol. 3, No. 11 1931
`
`~ 6 1
`
`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`ET6B
`l4~(.Jli
`
`HERV-K
`2145()78
`
`BCCS
`9 10 II ll ll I~ IS Ill
`
`9 10 J I 12 11 I~ J~ lb
`
`9 10 II 12 13 14 I~ 16
`
`Acun
`
`-
`
`HSDPH
`
`BCCLI
`
`Acun
`
`rtgUr-e 6. Multiple tissue Northern analyses of the 10 gene loci identified tn the region between RNU2 and 017S78. Northern filu:rs (Clomech) contain hean (I),
`bra.in {2), placenta (3), lung (4), liver (5). skeletal muscle (6), kidney (7), pancreas (8), spleen (9), thymus (10). prostate (II), testis (12), ovary (13), small intestine
`(14), coloo (15) and peripheral blood leukocyte (16).
`
`to zoo blots containing digested genomic DNA from a variety
`of species, in a systematic survey for evolutionarily conserved
`sequences. Six fragments, Cll98f, Al028h, Al028c, B07165c,
`B07165d and G ll5la appeared to be clearly conserved (Fig. 4C),
`while the remaining fragments were too repetitive to detect single
`bands by Southern hybridisation. Sets of overlapping eDNA
`clones were identified when each of the conserved fragments were
`hybridised to the relevant sublibrary (Fig. 4A). These eDNA
`clones were sequenced and used as probes on large insert eDNA
`libraries to isolate full length eDNA clones.
`Fragment C ll98f identified the env and pol genes of the
`HERV-K endogenous retrovirus locus, whilst AI028h identified
`gene BCCIO, a ubiquitously-expressed novel gene with no
`sequence database homologies to date. Fragments c and d of
`cosmid 807165 both identified another novel gene, BCC7 which
`is expressed strongly in pancreas and skeletal muscle with
`message sizes of 1.8 and 4.8 kb (Fig. 6). Fragment G 115Ia did
`not hybridise to any clones in the sublibrary suggesting it was
`not well represented in the pool of eDNA libraries used in the
`direct selection process. However, the convincing Northern result
`with this fragment gives strong indication that it contains a gene,
`
`termed BCC8, with high tissue specificity (Fig. 6). A eDNA clone
`isolated from a retina eDNA library with fragment G 1151 a has
`shown no sequence database homologies to date. Fragment
`AI028c did not hybridise with any of the mRNA samples on
`both types of Northern filters tested, suggesting that either it
`contains a gene of high tissue specificity not represented by these
`Northern fLlters, or it is a fragment of genomic DNA with no
`function. It has therefore been omitted from the transcript map,
`but the possibility remains that this is a locus of an additional
`gene.
`
`Exon-trapping. Exon-trapping was performed with the same sets
`of cosmids used to create the above-mentioned subtracted eDNA
`libraries. A total of 14 products were obtained, 10 of which
`hybridised to clones in the subtracted eDNA libraries. The loci,
`HERV-K LTR, ET6B, 8Eg2, HAL64, BCC10, VHR phosphatase
`and MOXJ were all identified by exon-trapping. Four products
`failed to isolate any cDNAs from the subtracted libraries
`suggesting that they represent genes not selected in the
`sublibraries. Cloning, sequencing and hybridisation
`to
`
`GeneDX 1034, pg. 5
`
`

`

`Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 7 of 9
`
`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`The identification of a complex integration site for the human
`endogenous retrovirus HERV -K at one end of the transcript map
`may explain the cloning difficulties experienced in this region;
`a phenomenon also observed at the G6PD locus of chromosome
`Xq28 which also contains a HERV-K integration site (43). Many
`of the eDNA clones isolated from this locus from random plated
`eDNA libraries and the gridded sublibraries failed to localise to
`chromosome 17. These clones may represent homologous
`retroviral inserts from elsewhere in the genome. Sequence
`anaJysis of the BRCAJ region-specific HERV-K eDNA clones,
`and of subcloned genomic DNA from cosmid C 1198, suggest
`that this locus contains pol and env genes and a long tenninal
`repeat (LTR). Completion of the sequencing at this locus will
`indicate whether this HERV-K is transcribed and is capable of
`encoding proteins. RelrOvi.rus-related sequences exhibit a number
`of features giving
`them a potential for
`involvement
`in
`carcinogenesis (reviewed in ref. 44). Proviral integration can
`cause activation of adjacent cellular protooncogenes by promoter
`insertion, as observed by induction of lymphomas by the viral
`promoter-induced activation of the cellular gene, c-myc (45). In
`addition, stimulation of HERV-K expression in the breast
`carcinoma cell line, T47D by female steroid hormones has been
`observed. It has been suggested that this enhanced expression
`may be involved in the aetiology of cenain human breast cancers,
`similar to the MMTV-induction of tumours in mice (46) . Hence,
`the HERV-K locus and its neighbouring genes inside the BRCAI
`region merit further investigation.
`Another interesting gene in this region is a new ADP(cid:173)
`ribosylation fuctor, HAL64, related to the hlgh.ly conserved family
`of GTP-binding proteins with homology to both the ras oncogene
`superfamily and the heterotrimeric G protein a subunits (47) .
`HAL64 appears most closely related to a processed chicken
`pseudogene, CPSJ which has been implicated as an oncogene
`due to its similarity to the c-H-ras oncogene and possible
`involvement in early embryonic development (48) . This gene may
`represent a new oncogene and further studies are required to see
`whether it is involved in tumour development.
`Another biologically plausible candidate gene, the dual(cid:173)
`specificity VHl-related VHR phosphatase was also found to
`reside within this transcript map. A member of the protein(cid:173)
`tyrosine-phosphatases (PTPases). this gene has potential roles in
`cell signalling, cell growth and proliferation and oncogenic
`transfonnarion (40). A possible role in the reversal of the effects
`of protein tyrosine Jcinases has implicated the PTPases as
`candidate rumour suppressor genes (49). However, investigations
`of other PTPases have fruled to find any involvement in lung
`and renal cancen; (50,51 ). We have searched samples of genomic
`DNA and eDNA in our breast and ovarian cancer families and
`in breast and ovarian tumours for oncogenic mutations, but have
`found no alterations to date. However, it may be interesting to
`see whether any loss of function of this gene within breast and
`ovarian tumours has any effect on tumour progression ,
`The remaining genes in thjs transcript map require further study
`into possible roles in breast and ovarian cancer. Since they all
`reside within a region commonly deleted in breast and ovarian
`rumours, the implications for tumourigenesis and development
`must be investigated.
`
`MATERIALS AND METHODS
`Ubrarles
`Y A C clones were isolllled from libl'!l.ries genenued 81 the Imperial Canrx:r Research
`Fund (lCRf), Co;ure d'Erude du Polymorph.ismc Humain (CEPH) and Washington
`
`1932 Human Molecular Genetics, 1994, Vol. 3, No. 11
`
`conventional eDNA libraries is now in progress to further
`characterise these products.
`
`ldentificaJion of CpG islands. In addition, the minimal cosmid
`set was screened by restriction enzyme analysis for the presence
`of CpG islands (38). A putative CG island-containing fragment
`was defined as a fragment which was cut by both the rare-cutters
`Sacll and BssHD. All such fragments appear to reside in close
`proximity to genes isolated in the region (Fig. 4A). Further
`characterisations of these fragments is underway to confmn they
`are the 5' regions of these genes.
`
`DISCUSSION
`
`In this paper, we describe a detailed physical and transcriptional
`analysis of a 400 kb region residing within the narrowest limits
`known to contain BRCAJ. This work has identified and located
`ejght novel genes, including a novel G protein, and an endogenous
`retrovirus related to the HERV-K fanilly, and the previously
`described dual-specificity VHR phosphatase (40) and MOXI
`homeobox genes (41 ). All reside within a chromosomal region
`shown to be frequently deleted in sporadic and familial breast
`and ovarian tumours and may therefore play a role in breast and
`ovarian tumourigenesis and progression.
`The long-range restriction map described here combined with
`analyses of YAC cosrnid pools gives the order of markers in the
`BRCAI region as cen- JAJ .3B- RNU2- D17S858/D17S859 -
`PPY-D17S78-tel. This further ref111es the physical mapping
`data provided by multicolour fluorescence in situ hybrisation with
`A/u-PCR-amplified Y AC clone DNA (27). A rough analysis of
`the restriction fragment data provided by PFGE Southern blot
`hybridisation with single copy probes in this region suggested
`that the D 175858/D 175859 locus was approximately 300 kb distal
`to RNU2 in the B ceU line tested. Although restriction site
`polymorphisms and differences in methylation could exist
`between the B cell line used in the PFGE analysis here, and the
`cell lines used to construct the cosmid and Y AC libraries, the
`distance covered by the cosmjd contig correlated roughly with
`estimates provided by the restriction map. The cosmid contig was
`confirmed as contiguous by FISH analysis which also agreed the
`order of the cosmids withln the walk, providing further evidence
`to support the physical mapping data.
`The assembly of an ordered cosmid contig has enabled the
`precise localisation of gene loci in the region. A combination
`of different gene isolation approaches was employed to make the
`transcript map as complete as possible. The method of direct
`eDNA selection (34,35) proved
`the most efficient and
`straightforward of the eDNA isolation techniques and has
`provided eDNA clones at every gene locus on our r:ranscript map
`with the exception of BCC8 wh.ich appears to be a gene with
`hlgh tissue specificity (Fig. 6). The use of a pool of eDNA
`libraries has probably increased the likelihood of isolating most
`of the genes in this region. In addition, limiting the complexity
`of the input genomic DNA to a maximum of five cosmids may
`have also increased the efficiency of the hybrid selection. We
`have combined thjs technique with expression-independent gene
`isolation methods of cross species homology searches, exon(cid:173)
`trapping and CpG island hunting. Whilst each of these methods
`have their lirni1ations (discussed in ref. 42) , we believe that the
`combination of all of them will result in a complete transcript
`map of the region.
`
`GeneDX 1034, pg. 6
`
`

`

`Downloaded from
`
`http://hmg.oxfordjournals.org/
`
` at Stanford Medical Center on August 22, 2013
`
`Case 2:13-cv-00640-RJS Document 120 Filed 09/05/13 Page 8 of 9
`
`Human Moleculnr Genetics, 1994, Vol. 3, No. I 1 1933
`
`School of Mc:&ine (52-54). Scr=Ung was carried oot by hybridisatlon 10 gridded
`Y AC ftlters and PCR analysis of Y AC pools (55) . Alu PCR prtXlUcts (56) and
`YAC iruens (57) from lhe positive YAC clones were: used to :;crcen duplicate
`filters of lhe chromosome 17 cosmid libr.uy ooruaio.ing 2(]736 clones per membrane,
`representing a 4-fold genomic coverage (29) .
`
`purifted using Qiagen columns (Dil\gen) and used to tranSfect COS-7 cells by
`electroporation. The cells were grown for 48 h and cytoplasmic total RNA was
`prepared . First-wand c 0 N A symhcosis and PCR amp I ification were carr'lcd out
`as dcscn'bed (3 7). Secondary PCR and cloning of the ~on-trapped prodll(;IS was
`carried out as described (66).
`
`DNA preparallon
`Gt:oomic DNA for PFGE analysis was ~ from culrured lymphoblastoid
`ceU lines as described (58). C05TT1id DNA was prepared by alkaline lysis (59).
`
`Restriction digestion of agarose blocks
`Each block was digestc:d Ill 190 I' I of sterile distilled water and 15 - 20 units of
`enzyme (Northumbria biolabs) with 25 p.J of 10x retriction enzyme buffer
`(North umbria baolabs). Digests were carried out for 4 hat 37•c (BssHD dagest5
`at so•q,
`
`Southern transftf' and b.ybridlsatl!los
`All agarose gels, including pulsed-field gels, were transferred to Hybond-N+
`by lllkali transfer wx1 hybridised at65•c ao:ortling to lhe rnanufllCIUf"e'r (Amct>ham
`LIFE SCIENCE).
`
`DNA labelling
`All probes were labc:Ucd by the ra00om priming method using la-12P)dCfP (60)
`and ndiooctively-labelled probe was .separnted from Wlincorporatcd [aJ2PidCTP
`on a scphadcx G50 column. Probes were denatured by boiling and preanncaled
`a1 65°C for 90 min with a 5000-fold eJtccss (w/w) of sonicated human placental
`DNA prior to hybrid.isation to compete out repecltive elemcnlS.
`
`Single copy probes for Southern analysis
`Probes used include a 3.7 kb insen from a clone containing the gene IA/.38
`(61 ). an RNU2 gene probe(a gift from A . We~r. Yale Univors.uy), a PCR product
`from the coding region of the gene for panci'Clltic polypeptide (PPY) (62), the
`probe pl31 from the locus DI7S78 (63) and 12E/Bd, a fragment of the
`017S858/D17S859-positive rosmid, 005149. Between each hybridisation, the
`blot was stripped and checked by automdiogmphy for any residual signal.
`
`Pulsed f'idd get electrophoresis
`All gels were made up in 0.5 XTris-borutc- BDTA elot,'IIOphoresis buffer with
`1% agarose. Gels were run using parameters obtained from the Biorad CHEF
`DRD CHEFMAPPER programme in order 1.0 scpara!C fragme:nlS below 1.2 Mb.
`All gels were run at a circulating temperature of I4°C using the Biorad CHEFIJ
`syst.em.
`
`Cosmld walking
`Cosmid DNA (I /lg) was digested to completion using &al. An in 1ritro
`transcriJX.ion was then perfonned using the riboprobe n core system kit (Promega)
`according to manufacturer's instructions. The radioaaively labelled riboprobe
`was then hybridised dlteetly with the eosmid library fillers.
`
`fluorescence in situ hybridlsation
`Interphase nuclei were prepared and FISH was carried oot as descnbed previously
`(32). Antibody detection and signal visua.lisation were as described (64).
`
`Direct eDNA hybrid sclection
`A min1mal set of founeen cosmid.s, which together span the cosmid walk, was
`divided into three sets of four or five cosmid.s for the construction of eDNA
`sublibrwies by hybrid selection (34,35,43) using a pool of