Ž .Mutation Research 374 1997 79–87

ž /A glucose-6-phosphate dehydrogenase G6PD splice siteconsensus sequence mutation associated with G6PD enzyme

deficiency

Sean Sanders a, Darrin P. Smith b, Geraldine A. Thomas a,), E. Dillwyn Williams a

a Department of Histopathology, LeÕel 5, Lab Block, Addenbrooke’s Hospital, Hills Road, Cambridge, CB2 2QQ, UKb CRC Human Cancer Genetics Research Group, LeÕel 3, Lab Block, Box 238, Addenbrooke’s Hospital, Hills Road, Cambridge, CB2 2QQ,

UK

Received 17 September 1996; accepted 10 October 1996

Abstract

Ž . Ž .A glucose-6-phosphate dehydrogenase G6PD deficient strain of mouse GPDX which was developed using theŽ .ethylating agent ethylnitrosourea ENU has been used to study clonality in epithelial tissues. While the biochemical defect

has been quantified, the genetic basis of the deficiency is unknown. The G6PD gene is composed of 13 exons. Exon 1 is nottranslated, and the ATG start site is near the 5X end of exon 2. Direct sequencing of the exonic regions of the gene from

Ž .GPDX, C3H, 101, C57BLr6 and BALBrc mice was carried out. The coding region, in which with a single exception allmutations found to cause G6PD deficiency in man are situated, showed identical sequences in three of the four strains

Ž .studied 101 coding region sequence was not examined . However, the G6PD gene in the GPDX mouse showed a singleŽ . Xbase difference from the other four strains and from the published mouse G6PD sequence BALBrc in the 5 splice site

consensus sequence at the 3X end of exon 1, part of the untranslated region. The difference was confirmed in four differentŽ .GPDX mice. This mutation was of the type A to T transversion that is known to be induced by ENU; its effect is likely to

be exerted through a defect in transcription, splicing or translation, leading to a reduction in protein levels. By Western blotwe have found a marked decrease in the G6PD protein levels in the GPDX mouse, with the C3H=GPDX heterozygoteshowing a lesser decrease. Recently, an increasing number of mutations in the untranslated regions of genes have been foundwhich have effects on protein levels. We believe that the reduced enzyme activity in the GPDX mouse is due to the mutation

X Ž .in the 5 untranslated region UTR , and that similar mutations may be relevant in other inherited conditions.

Keywords: Glucose-6-phosphate dehydrogenase; Mutation; Splice site; Ethylnitrosourea; Mouse

) Corresponding author. Tel.: q44 1223-331699; Fax: q44Ž .1223-216980; E-mail: sean-sanders@nih.gov S. Sanders and

Ž .gat1000@cam.ac.uk G.A. Thomas .

1. Introduction

Over 400 phenotypic variants of human G6PDw xdeficiency have been described 1,2 . Of these, ap-

proximately 300 have been typed by protein bio-w xchemistry using WHO guidelines 3 . The remaining

100 were found using alternative methods.

0027-5107r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved.Ž .PII S0027-5107 96 00222-9

( )S. Sanders et al.rMutation Research 374 1997 79–8780

More recently, G6PD gene defects in humanshave been examined by direct sequencing; 97 muta-tions or combinations of mutations have so far been

w xdescribed 2,4,5 of which the majority are missensemutations. Five deletion mutations have been de-

Žscribed, three 3 base pair deletions ‘G6PD Sunder-. w xland’, ‘Urayasu’ and ‘Tsukui’ 4,6 , a 24 base pair

Ž . w xdeletion ‘G6PD Nara’ 7 and a 6 base pair deletionŽ . w x‘G6PD Stoney Brook’ 5 . All these deletions aremultiples of three so that a frameshift, which isthought to be lethal, does not result. A single non-sense mutation has been found, ‘G6PD Georgia’, in

Ž .which a tyrosine codon TAC is mutated to a stopŽ . w xTAA 5 . This type of mutation, which in this caseresults in the loss of 87 C-terminal amino acids, wasthought to be lethal, but the fact that the patient, whohad no known affected relatives, was a heterozygote

w xmay explain her survival 5 .With only one exception, all mutations have been

Ž .found in the coding region exons 2 through 13 ofthe gene. The exception is ‘G6PD Vansdorf’, whichinvolves the deletion of the invariant AG dinucleo-tide at the 3X acceptor splice site at the intron 10rexon

w x11 boundary 5 . The exact effect of this mutationwas not determined, but it is thought to cause either

w xexon skipping or incorrect splicing 5 . The pheno-type of this mutation is non-spherocytic hemolytic

w xanaemia 2 . Whether the scarcity of mutations innon-coding regions is due to their rarity or the factthat these regions have not been examined is notclear from the published literature.

The GPDX strain was created in 1984 by Charlesw xand Pretsch 8 as a potential model for the human

deficiency. It was developed by treating maleŽ .101rE1=C3HrE1 F mice with ethylnitrosourea1Ž .ENU , a well-known ethylating agent, and mating

w xthese animals with untreated ‘test-stock’ females 9 .One of the offspring showed a decreased blood

w xG6PD level 10 . Stock from this animal showederythrocyte G6PD activity of 20, 60 and 15% inhemizygotes, heterozygotes and homozygotes, re-

w xspectively 10 .We have made use of this mouse strain to provide

a model in which the clonality of tissue architecturew xand tumours can be studied 11–14 . However, the

underlying genetic cause of this deficiency has notyet been elucidated.

The aim of this study was to identify the causative

mutation in the GPDX strain by sequencing, and toinvestigate the means by which this mutation pro-duces the GPDX phenotype.

2. Materials and methods

2.1. Animals

For PCR and sequencing, liver tissue was takenfrom the following animals: one male BALBrc, onemale C3H, one female C57BLr6, one male 101, 3males and one female GPDX, and a single femaleC3H=GPDX heterozygote. For Western blot analy-sis liver tissue from 2 male C3H, 3 male and 1female GPDX, and 1 female C3H=GPDX het-erozygote was used.

2.2. Extraction of DNA

A small piece of liver was homogenised in asterilised glass homogeniser in 1 ml DNA Extraction

w Ž .Buffer 10 mM Tris-HCl Sigma , pH 8.0; 100 mMŽ .ethylenediaminetetraacetic acid Sigma , pH 8.0;

Ž0.5% sodium dodecyl sulphate lauryl sulphate; SDS;. Ž .Sigma ; and 20 mgrml RNase A Sigma; 10 mgrml ,

x Žadded just before use . Proteinase K Sigma; 100 ml.of 1 mgrml solution was added, and the samples

mixed and incubated in a water bath for 3 h at 508C.DNA was then extracted using a standard protocolw x Ž15 and the final product dissolved in TE pH

.7.6–8.0; 10 mM Tris, 1 mM EDTA . DNA concen-tration and purity was determined, and samples werediluted 1 : 10 in preparation for PCR.

2.3. PCR

All primers were 20 mers with a 50% GC content,and were commercially synthesised by R&D Sys-

Ž . Xtems Abington, UK and HPLC purified. The 5primers were 5X-biotinylated to aid with later separa-

Ž .tion for sequencing see below . The other primerswere not biotinylated. Primers used are listed inTable 1.

A total reaction volume of 50 ml was used: 2–3Ž .ml DNA, 5 ml of each of the two primers 2 mM , 5

Žml 10= Promega Buffer 500 mM KCl, 15 mMMgCl , 100 mM Tris-HCl, pH 9.0, and 1.0% Triton2

( )S. Sanders et al.rMutation Research 374 1997 79–87 81

. ŽX-100 , 5 ml 2 mM dNTPs, 0.5 ml Taq Promega; 5.Urml , made up to 50 ml with water. All sample

Ž .were covered with a drop ;50 ml sterile oilŽ .Sigma . For each PCR a ‘no DNA’ control was

included. A ‘false’ hot start at 808C for ;15 s wasused, followed by 35 cycles of 958C for 30 s, 558Cfor 60 s and 728C for 2 min. A final step of 5 min at728C allowed for completion of partial polymerisa-

Table 1Primers used for sequencing of GPDX gene

a All sequences read in 5X to 3X direction.b B, 5X-biotinylated; S, used as sequencing primer.

( )S. Sanders et al.rMutation Research 374 1997 79–8782

tion products. Samples were removed from under oiland stored at y208C.

Products were checked by running on a 1.5%agarose gel at a constant 50 V for 45 min. Loading

Ž .buffer xylene cyanol and bromophenol blue; 0.5 mlwas added to 4 ml of sample and the entire 4.5 mlwas loaded onto the gel. Marker was 200 ng ØX 174

ŽRF DNArHaeIII digest HT Biotechnology Ltd.,Cambridge, UK; 2 ml of 1 : 2.5 dilution in bro-

.mophenol blue and xylene cyanol loading buffer .Ž .Ethidium bromide EtBr; 2 mgrml was added to the

Ž .running buffer TrisrboraterEDTA to visualise thebands under UV light.

2.4. Sequencing

PCR was performed with the 5X primers whichhad been 5X-biotinylated. This allowed the PCR prod-uct to be manipulated using Dynalw M-280 strepta-vidin beads and a Dynalw Magnetic Particle Concen-

Ž .trator MPC , following the protocols supplied withthe product.

For direct sequencing, the protocol supplied withthe Sequenasee Version 2.0 DNA Sequencing kitŽ .USB was followed. Samples were run on 40 cmsequencing gels for between 1.3 and 5.5 h at aconstant 40 W. All samples were denatured at 758C

for )3 min prior to loading 5 ml, making use of theparticle concentrator. Gels were dried under vacuumat 808C for 1 h, loaded into an X-ray cassette andexposed for between 1 and 7 days using Fuji RXX-ray film.

2.5. Extraction of protein

Ž .A small piece of liver ;0.3 g was added to achilled 1 ml glass homogeniser containing 1 ml lysis

Žbuffer 66 mm Hepes, pH 7.4;20 mM NaCl; 1%glycerol; 1% Triton X-100; 2 mM MgCl ; 6 mM2

EDTA, pH 8.0; 10 mgrml aprotinin; 2 mM phenyl-.methylsulphonyl fluoride; 1 mM benzamidine . Tis-

sue was homogenised for 1–2 min, sonicated on icetwice for approximately 10 s and spun at 48C, 13 000=g for 10 min. The supernatant was removed to afresh tube and the protein concentration determinedat 750 nm using a Bio-Rad Protein Assay kit. Thesamples were then diluted in water and 2= gel

Žloading buffer 120 mM Tris, pH 6.8; 4% SDS; 20%.glycerol; 500 mM DTT; 0.005% bromophenol blue

to yield a final protein concentration of 0.5 mgrml.

2.6. Western blot

Samples were run on a 10% polyacrylamide gelwith a 3.5% stack for 75 min at 200 V. The protein

Ž .Fig. 1. Polyacrylamide sequencing gel showing normal BALBrc , heterozygote and GPDX sequences around the mutation site. ArrowsŽ . X Xindicate the site of the mutation cf. Fig. 2 . Intronic sequences shown in lowercase, exonic in uppercase, and sequence reads 5 -3 from top

to bottom.

( )S. Sanders et al.rMutation Research 374 1997 79–87 83

Fig. 2. Schema of 5X end of the G6PD gene. The putative TATA box sequence is indicated and the 5X splice site is underlined, with thew xsplice site consensus sequence below 27 , and the sequence in all other mouse strains examined above. Intronic or non-coding sequences

are shown in lowercase, exonic in uppercase. The GPDX splice site mutation is in boldface type. The ATG start in exon 2 is indicated.

Ž .was blotted onto nitrocellulose Hybaid at 25 V for30 min. Filters were stained with Ponceau-S and the

Žpositions of the markers High Molecular Weight.Markers; Sigma noted. Blocking took place for 3 h

Žat room temperature in 6 ml of block solution 5%.non-fat milk powder, 3% BSA, 0.5% Tween-20 per

filter.Incubation in primary antibody was carried out

overnight at 48C and the filters washed. Two differ-ent primary antibodies were used: PM2 rabbit anti-

Žhuman G6PD antibody a gift from Dr. Philip Ma-

.son , used at 1 : 200, and RK rabbit anti-rat G6PDŽ .antibody a gift from Dr. Rolf Kletzien , used at

1 : 400. As a control for loading, a mouse mono-Ž .clonal anti-MAP kinase was used at 1 : 500 Upstate .

Ž .Secondary antibody HRP-linked incubation wasperformed at room temperature for 1 h, followed bywashing. For the PM2 and RK antibodies, a swineanti-rabbit secondary was used, while for the anti-MAP kinase antibody, a sheep anti-mouse was used.

Antibody was visualised using a chemilumines-Ž .cence system Pierce and performed according to

Fig. 3. Intronrexon boundaries sequenced. Intronic sequences in lowercase, exonic in uppercase.

( )S. Sanders et al.rMutation Research 374 1997 79–8784

the protocol. The signal using the RK antibody wasstronger, possibly due to higher specificity, so arange of exposures between 20 s and 3 min wasused.

3. Results

Through direct sequencing of PCR-amplified ge-nomic DNA, we have found a single base differenceŽ .A-T to T-A in the G6PD sequence in the GPDXstrain compared to four others: BALBrc, C57BLr6,

ŽC3H and 101 the latter two being the GPDX back-.ground strains; Fig. 1 . This mutation is in the

untranslated 3X end of exon 1, in the 5X splice siteŽ .consensus sequence Fig. 2 . No other differences

were found in the entire coding region of three of the

Ž .Fig. 4. A Western blot using PM2 rabbit anti-human G6PDantibody. Lane 1, C3H; lane 2, GPDX; lane 3, C3H=GPDX

Ž .heterozygote. B RK rabbit anti-rat G6PD antibody. Lane 1,Ž . Ž .C3H; lane 2, GPDX. C Controls for protein levels for A . Sheep

anti-mouse MAP kinase primary antibody used, giving a singleŽ . Ž .band at the 42 kDa level. Loading sequence as for A . D

Ž .Control for C . Sheep anti-mouse MAP kinase primary antibody.Ž . Ž .Same loading as in C . The upper, non-specific bands in C and

Ž .D result from the use of an anti-mouse primary antibody withmouse extract.

Ž .strains 101 strain not examined , except for a singleŽ 718silent mutation in the C57BLr6 strain C to A,

221 .coding for Gly ; data not shown . The base differ-ence was confirmed in 3 other GPDX animals, andin a C3H=GPDX heterozygote. The intronrexonborders flanking exon 6, the 3X end of exon 10, and

Ž . Xflanking exon 11 Fig. 3 , as well as part of the 3untranslated region were examined for another pur-pose, and ruled out other splice site mutations ormutations affecting mRNA stability in these regions.

Ž .Western blot analysis Fig. 4 of protein extractedfrom C3H and GPDX mice detected a specific pro-tein band at ;58 kDa, in the correct range for themouse G6PD enzyme. This same band was seen withtwo different antibodies, one raised against humanŽ . Ž .PM2 and the other against rat RK G6PD. The gelclearly shows a large decrease in the amount of

ŽG6PD protein in the GPDX mouse Fig. 4A and B,. Ž .lane 2 relative to the C3H lane 1 , when equal

amounts of total protein were used, while the C3H=

GPDX heterozygote showed an intermediate level ofŽ .protein Fig. 4A, lane 3 . This experiment was re-

peated with extracts from different animals of theŽsame strain, and identical results were seen data not

.shown . The amount of protein loaded was con-trolled for using a mouse monoclonal anti-MAP

Ž .kinase antibody Fig. 4B and D .This provides evidence that the lower G6PD ac-

w xtivity in GPDX mice shown by histochemistry 11w xand biochemistry 16 is not due to an active site

Žmutation and therefore maintained protein levels.with decreased activity , but rather to lower protein

levels. This is the expected finding for a mutationcausing a decrease in transcription, splicing or trans-lation.

4. Discussion

We have identified a mutation in the 5X untrans-lated sequence of the G6PD gene from a glucose-6-phosphate dehydrogenase deficient mouse strainŽ .GPDX by direct sequencing. This transversion mu-

Ž .tation A-T to T-A was in the penultimate base ofŽ .exon 1 Fig. 3 . It is compatible with the type of

mutation induced by ENU, the alkylating agent usedto create the GPDX strain by intraperitoneal injec-

w xtion of males 8 . We have not formally excluded the

( )S. Sanders et al.rMutation Research 374 1997 79–87 85

unlikely possibility that mutation arose by chance inthe test-stock strain. However, ENU has been shownto induce A to T transversions in mouse germline

w xcells 17–20 , and it is highly probable that it in-duced the mutation observed.

Western blot analysis showed severely loweredŽlevels of G6PD in the GPDX strain Fig. 4A and C,

.lane 2 and intermediate levels in the C3H=GPDXŽ .heterozygote Fig. 4A, lane 3 . This result was con-

firmed using two different antibodies and multipleanimals, and suggested that the GPDX phenotypewas due to a reduced level of G6PD enzyme ratherthan a decrease in protein activity caused by amissense mutation in the coding sequence. An alter-native but unlikely explanation was a missense muta-tion, insertion, or deletion causing epitope loss orreduced antigenicity.

We excluded a coding region mutation by demon-strating exact identity of the entire G6PD codingsequence from four GPDX mice with that of 3 otherstrains, and with the published mouse coding se-quence; intron sequences were not determined, ex-cept in the regions shown in Fig. 3. One exception ofa single silent mutation in the C57BLr6 strain in thetriplet coding for Gly 221 was found. This excludesthe possibility of the GPDX mutation causing achange in the protein sequence. The observed pheno-type must therefore have been due to lower proteinlevels.

Comparison of the 5X UTR in rat and mouseŽshows strong sequence conservation 40 out or 41

. Xbases in the 3 region of exon 1, suggesting that thisregion is important in rodents. The AG dinucleotideat the 3X end of the exon is also conserved in humansw x21 .

Splice site mutations are increasingly being iden-wtified as responsible for a variety of disorders 4,22–

x26 . The particular base change identified here isnovel, but there is good supporting evidence for itsinvolvement in the regulation of transcription.

A 9 base consensus sequence for the 5X splice sitewas identified around the essentially invariant GTdinucleotide at the start of an intron, by examining

w x139 coding nuclear and viral genes 27 . In theC < A Ž <.consensus sequence r AG GT r AGT the barA G

represents the splice site and bases to the rightŽ .intronic are numbered from q1, while those to the

Ž . w xleft exonic are numbered from y1 27 .

There is much and varied evidence for the impor-tance of the 5X splice site consensus sequencew x w x24,28–33 , some of which is contradictory 34,35 .The majority of findings thus far indicate that the 9base pair consensus sequence plays a critical part inthe splicing process. However, its relative impor-tance depends on the particular RNA undergoingsplicing, as well as the environment. In addition, the5X splice site has been shown to be an absoluterequirement for formation of the splicing complex or

w xspliceosome 36 .Work done on the relative lifetimes of spliceo-

some complexes bound to normal and mutant splicesites supports a model in which the usage of acertain splice site is influenced by the dissociation

w xrate of the complex 30,37 . So, a mutated donorŽsplice site could influence the dissociation or associ-

.ation rate of the spliceosome components, leading toa relatively lower splicing rate.

Another complicating factor to take into accountis the position of the base change. Being at the endof the first exon, the machinery with which it inter-acts might well be different from that used at the

w xother downstream splice sites. Ohno et al. 38showed that the 5X cap structure found in all mRNAsappears to play a role in the splicing of the first

w xintron 39 .The y2 position of the 5X splice site is evidently

an important site. The evidence indicates that theconsensus A at this position provides the most effi-

w xcient splicing 24,29 . The presence of a T results inw xlower levels of splicing 29 . In addition, an A

residue is found significantly more often in donorsites examined from both mouse and human genes

w xthan a T 27,40 .It is highly unlikely that intronic mutations in the

invariant dinucleotides resulting in exon skipping orintron inclusion are present, as no difference inmRNA transcript size was seen in the GPDX strainafter RT-PCR. Five further intron-exon boundarieswere sequenced for another purpose, and all showed

Ž .complete identity in all strains Fig. 3 . An additionalmutation in an intron affecting mRNA stability re-mains a theoretical possibility, but the mutation wehave identified is of the type induced by ENU, isrestricted to the GPDX strain, and is in a highlyconserved splice site region where a mutation wouldbe expected to influence splicing efficiency.

( )S. Sanders et al.rMutation Research 374 1997 79–8786

In summary, we have identified a single base atthe 3X end of exon 1 present in all GPDX miceexamined which differs from the sequence found infour other mouse strains; it is the only such differ-ence. The entire coding region from three of the four

Ž .strains 101 not examined , as well as partial intronsequences and part of the 3X UTR, was examined.We believe that this mutation causes a decrease inthe amount of enzyme through a splicing defect. Wehave demonstrated, using Western blot analysis, thatthere is a decrease in G6PD protein in mice carryingthe GPDX mutation.

Of the 400 known phenotypes of human G6PDdeficiency, only 97 have been accounted for on a

w xgenetic level 4 . The demonstration of a splice sitemutation associated with low protein levels empha-sises the importance of examining the untranslatedregions of the gene in patients with G6PD deficiencyfor possible causative mutations. Missense mutationsin highly conserved regions can usually be desig-nated causative with a good amount of certainty.However, the interpretation of mutations in less wellconserved regions should be carried out with morecaution, and both coding and non-coding regionsexamined to avoid misinterpretation of simple poly-morphisms as causative. Splice site mutations of thetype found here may well be important in otherheritable diseases in which the underlying geneticdefect has not yet been found.

Acknowledgements

This work has been supported by grants from theŽ . Ž .Wellcome Trust SS and the BBSRC GAT . We

are very grateful to both Dr. Philip Mason and Dr.Rolf Kletzien for the kind gifts of G6PD antibodies.We would also like to acknowledge Mr. Chris Bur-ton for photography.

References

w x Ž .1 Beutler, E. 1990 The genetics of glucose-6-phosphate dehy-Ž .drogenase deficiency. Semin. Hematol., 27 2 , 137–164.

w x Ž . Ž .2 Beutler, E. 1994 G6PD deficiency. Blood, 84 11 , 3613–3636.

w x Ž .3 Betke, K. 1967 Standardization of procedures for study of

glucose-6-phosphate dehydrogenase. WHO Technical ReportSeries, 366, 2–33.

w x Ž .4 Beutler, E., T. Vulliamy and L. Luzzatto 1996 Hematologi-cally important mutations: Glucose-6-phosphate dehydroge-

Ž .nase. Blood Cells Mol. Dis., 22 4 , 49–56.w x5 Xu, W., B. Westwood, C.S. Bartsocas, J.J. Malcorra-Azpiazu,

Ž .K. Indrak and E. Beutler 1995 Glucose-6-phosphate dehy-´drogenase mutations and haplotypes in various ethnic groups.

Ž .Blood, 85 1 , 257–263.w x6 MacDonald, D., M. Town, P. Mason, T. Vulliamy and L.

Ž .Luzzatto 1991 Deficiency in red blood cells. Nature, 350,115.

w x Ž .7 Hirono, A., H. Fujii, M. Shima and S. Miwa 1993 G6PDNara: A new class 1 glucose-6-phosphate dehydrogenase

Ž .variant with an eight amino acid deletion. Blood, 82 11 ,3250–3252.

w x Ž .8 Charles, D.J. and W. Pretsch 1984 A mouse mutant defi-cient in erythrocyte glucose-6-phosphate dehydrogenase afterparental ethylnitrosourea treatment. Genetics, 107, s19.

w x Ž .9 Russell, W.L. 1951 X-ray-induced mutations in mice. ColdSpring Harbor Symp. Quant. Biol., 16, 327–335.

w x Ž .10 Charles, D.J. and W. Pretsch 1987 Linear dose-responserelationship of erythrocyte enzyme-activity mutations in off-spring of ethylnitrosourea-treated mice. Mutation Res., 176,81–91.

w x Ž .11 Thomas, G.A., D. Williams and E.D. Williams 1988 Thedemonstration of tissue clonality by X-linked enzyme histo-chemistry. J. Pathol., 155, 101–108.

w x12 Griffiths, D.F.R., S.J. Davies, D. Williams, G.T. WilliamsŽ .and E.D. Williams 1988 Demonstration of somatic muta-

tion and colonic crypt clonality by X-linked enzyme histo-Ž .chemistry. Nature, 333 6172 , 461–463.

w x Ž .13 Thomas, G.A., D. Williams and E.D. Williams 1989 Theclonal origin of thyroid nodules and adenomas. Am. J.

Ž .Pathol., 134 1 , 141–147.w x14 Seddon, S.V., D.M. Walker, D. Williams and E.D. Williams

Ž .1992 The clonal organization of the squamous epitheliumof the tongue. Cell Prolif., 25, 115–124.

w x Ž .15 Sambrook, J., E.F. Fritsch and T. Maniatis 1989 Molecularcloning: A Laboratory Manual, 2nd Edn., Cold Spring Har-bor, New York: Cold Spring Harbor Laboratory Press.

w x Ž .16 Pretsch, W., D.J. Charles and S. Merkle 1988 X-linkedglucose-6-phosphate dehydrogenase deficiency in Mus mus-

Ž .culus. Biochem. Genet., 26 1r2 , 89–103.w x17 Douglas, G.R., J. Jiao, J.D. Gingerich, J.A. Gossen and L.M.

Ž .Soper 1995 Temporal and molecular characteristics of mu-tations induced by ethylnitrosourea in germ cells isolatedfrom seminiferous tubules and in spermatozoa of lacZ trans-

Ž .genic mice. Proc. Natl. Acad. Sci. USA, 92 16 , 7485–7489.w x Ž .18 Zingg, B.C., W. Pretsch and H.W. Mohrenweiser 1995

Molecular analysis of four ENU induced triosephosphateisomerase null mutants in Mus musculus. Mutation Res.,

Ž .328 2 , 163–173.w x Ž .19 Provost, G.S. and J.M. Short 1994 Characterization of

mutations induced by ethylnitrosourea in seminiferous tubulegerm cells of transgenic B6C3F1 mice. Proc. Natl. Acad. Sci.

Ž .USA, 91 14 , 6564–6568.

( )S. Sanders et al.rMutation Research 374 1997 79–87 87

w x Ž .20 Shibuya, T. and K. Morimoto 1993 A review of the geno-toxicity of 1-ethyl-1-nitrosourea. Mutation Res., 297, 3–38.

w x21 Chen, E.Y., A. Cheng, A. Lee, W.-J. Kuang, L. Hillier, P.Ž .Green, D. Schlessinger, A. Ciccodicola and M. D’Urso 1991

Sequence of human glucose-6-phosphate dehydrogenasecloned in plasmids and a yeast artificial chromosome. Ge-nomics, 10, 792–800.

w x22 Treisman, R., N.J. Proudfoot, M. Shander and T. ManiatisŽ .1982 A single-base change at a splice site in a b8-thalas-semic gene causes abnormal RNA splicing. Cell, 29, 903–911.

w x Ž .23 Treisman, R., S.H. Orkin and T. Maniatis 1983 Specifictranscription and RNA splicing defects in five cloned b-thalassaemia genes. Nature, 302, 591–596.

w x Ž .24 Nakai, K. and H. Sakamoto 1994 Construction of a noveldatabase containing aberrant splicing mutations of mam-malian genes. Gene, 141, 171–177.

w x Ž .25 Baronciani, L., P. Bianchi and A. Zanella 1996 Hematolog-ically important mutations: Red cell pyruvate kinase. Blood

Ž .Cells Mol. Dis., 22 7 , 85–89.w x Ž .26 Cross, A.R., J.T. Curnette, J. Rae and P.G. Heyworth 1996

Hematologically important mutations: X-linked chronic gran-Ž .ulomatous disease. Blood Cells Mol. Dis., 22 8 , 90–95.

w x Ž .27 Mount, S.M. 1982 A catalogue of splice junction se-Ž .quences. Nucleic Acids Res., 10 2 , 459–472.

w x Ž .28 Montell, C. and A.J. Berk 1984 Elimination of mRNAsplicing by a point mutation outside the conserved GU at 5X

Ž .splice sites. Nucleic Acids Res., 12 9 , 3821–3827.w x29 Aebi, M., H. Hornig, R.A. Padgett, J. Reiser and C. Weiss-

Ž .mann 1986 Sequence requirements for splicing of highereukaryotic nuclear pre-mRNA. Cell, 47, 555–565.

w x Ž .30 Eperon, L.P., J.P. Estibeiro and I.C. Eperon 1986 The roleof nucleotide sequences in splice site selection in eukaryoticpre-messenger RNA. Nature, 324, 280–282.

w x Ž .31 Reed, R. and T. Maniatis 1986 A role for exon sequences

and splice-site proximity in splice-site selection. Cell, 46,681–690.

w x32 Mertes, G., M. Ludwig, B. Finkelnburg, M. Krawczak, R.Ž .Schwaab, H.-H. Brackmann and K. Olek 1994 A G -to-Tq3

donor splice site mutation leads to skipping of exon 50 inVon Willebrand factor mRNA. Genomics, 24, 190–191.

w x Ž .33 Cote, J., J. Beaudoin, R. Tacke and B. Chabot 1995 The U1ˆ ´small nuclear ribonucleoproteinr5X splice site interactionaffects U2AF65 binding to the downstream 3X splice site. J.

Ž .Biol. Chem., 270 8 , 4031–4036.w x Ž .34 Wieringa, B., F. Meyer, J. Reiser and C. Weissmann 1983

Unusual splice sites revealed by mutagenic inactivation of anauthentic splice site of the rabbit b-globin gene. Nature, 301,38–43.

w x Ž .35 Parent, A., S. Zeitlin and A. Efstratiadis 1987 Minimalexon sequence requirements for efficient in vitro splicing of

Ž .mono-intronic nuclear pre-mRNA. J. Biol. Chem., 262 23 ,11284–11291.

w x Ž .36 Frendewey, D. and W. Keller 1985 Stepwise assembly of apre-mRNA splicing complex requires U-snRNPs and specificintron sequences. Cell, 42, 355–367.

w x Ž .37 Green, M.R. 1991 Biochemical mechanisms of constitutiveand regulated pre-mRNA splicing. Annu. Rev. Cell Biol., 7,559–599.

w x Ž .38 Ohno, M., H. Sakamoto and Y. Shumira 1987 Preferentialexcision of the 5X proximal intron from mRNA precursorswith two introns as mediated by the cap structure. Proc. Natl.Acad. Sci. USA, 84, 5187–5191.

w x39 Padgett, R.A., P.J. Grabowski, M.M. Konarska, S. Seiler andŽ .P.A. Sharp 1986 Splicing of messenger RNA precursors.

Ann. Rev. Biochem., 55, 1119–1150.w x Ž .40 Ohshima, Y. and Y. Gotoh 1987 Signals for the selection of

a splice site in pre-mRNA: Computer analysis of splicejunction sequences and like sequences. J. Mol. Biol., 195,247–259.