CA1214125A - Human factor ix dna - Google Patents

Abstract

ABSTRACT
GENETIC ENGINEERING
It has been a problem to find an alternative, less time-consuming, and more reliable source of factor IX, a polypeptide which is essential to the human blood clotting process and necessary for the treatment of patients with Christmas disease.
The invention is an important step towards solving the problem by way of genetic engineering, in that it provides recombinant DNA
containing a DNA sequence occurring in the human factor IX genome.
It includes recombinant DNA comprising substantially the whole sequence of human factor IX genome inserted in a cloning vehicle and transformed into a host such as E.coli. It is conveniently characterised by a 129 or 203- nucleotide long sequence (J-J' and J'-J" in Figure 9). Other fragments of the sequence have also been cloned and the invention includes DNA molecules comprising part or all of the human factor IX DNA. The invention also includes cDNA derived from human factor IX RNA. Uses of the invention include the provision of an intermediate of value in the genetic engineering of a factor IX polypeptide precursor and thence manufacture of the factor IX polypeptide, and in making probes for use in diagnosing the presence of normal or abnormal factor IX DNA in patients with Christmas disease.

Description

~2~æ~

HUMAN FACTOR IX DNA
Back~round of the Invention 1. Field Df the invention This inventlon is in the field of genetic engineering relating to factor IX DNA.
05 2. Description of prlor art Factor IX (Christmas factor or antihaemophillc factor B) is the zymogen oE a serine protease which is required for blood coagulation via the intrinsic pathway of clotting (Jackson ~
Nemerson, Ann.Rev.Biochem. 49, 765-811, 1980). This factor is synthesised in the liver and requires vitamin K for its biosynthesis (Di Scipio & Davie, Biochem. 18, 899-904, 1979).
Human factor IX has been purified and characterised, but details of the amino acid sequence are fragmentary. It is a single-chain glycoprotein, with a molecular weight of approxi-mately 60,000 (Suomela, Eur.J.Biochem. 71, 145-154, 1976). Like other vitamin K-dependent plasma proteins~ human factor IX contains in the amino-terminal region approximately 12 gamma-carboxyglutamic acid residues (Di Scipio & Davie, Biochem. 18, 899-9049 1979).
During the clotting process, and in the presence of Ca ions, factor IX is acted upon by activated factor XI (XIa) by the cleavage of two internal peptide bonds9 releasing an activation glycopeptide of 10,000 daltons (Di Scipio et al., J.Clin.
Invest. 61, 1528~1538, 1978). The activated factor IX (IXa) is composed of two chains held together by at least one disulphide bond. Factor IXa then participates in the next step in the coagulation cascade by acting on factor X in the presence of activated factor VIII, Ca ions, and phospholipids ~Lindquist et al., J.Biol.Chem. 253, 1902-1909, 1978).
Individuals deficient in factor IX ~Christmas disease or haemophilia B) show bleeding symptoms which perslst throughout life. Bleeding may occur spontaneously or following injury. This may take place virtually anywhere. Bleeding into the joints is co~mon, and after repeated haemorrhages, may result in permanent and crippling deformities. The condition is a sex-linked disorder ~2~

affecting males. I~s frequency in the population is approxi-mately 1 in 30,000 males.
The current method oE diagnosing Christmas disease involves measurement of the titre of factor IX in plasma by a combination S Of a clotting assay and an ~nmunochemical assay. Treatment of haemorrhage ln the dlsease consists of factor IX replacement by means of lntravenous transfusion of human plasma protein concen-trates enriched in factor IX. The enrichment of plasma in factor IX is a time-consuming process.
Summar of the inven~ion Y _ . _ After considerable research and experiment, important progress has now been made towards producing artificial human factor IX by recombinant DNA technology (genetic engineering).
Thus, the cloning of DNA sequences which are substantially the same as extensive sequences occurring in the human factor IX
genome has been achieved.
The invention arises from the finding that an extensive DNA
sequence of ~he human factor IX genome can be obtained by a clever and laborious combination of chemical synthesis and artificial biosynthesis, starting from elementary nucleotide or dinucl~otide "building blocks", as will be described below.
A major feature of the invention comprises recombinant DNA
which comprises a cloning vehicle DNA sequence and a sequence forelgn thereto (i.e. foreign to the vehicle), encoding a translation produc~ convertible in vivo with the aid of vitamin K-dependent carboxylase into human factor IX.
In one specific aspect, the foreign sequence is substantially the same as a sequence occurring in the human fac~or IX genome.
A 11873 nucleotide long part of such a foreign sequence has been identified and a very large part of it has been sequenced by the Maxam-Gilbert sequencing method. A 129 nucleotide length of this sequence is more than sufficient to characterise it unamblguously as coding for a speclfic protein and a particular such length ls .... ~
~"~

æ~

- 2a -regarded herein as useful to characterise the whole sequence inserted in the cloning vehicle as one occurrlng ln the human factor IX genome. Other cloned sequences can then be veriied as belonglng to the human factor IX genome by determlning that part 05 thereof ls identical to a region of the first-mentione(l sequence, i.e. the sequences have a common identity ln an overlapping region~

A further feature of the invention therefore corr~prises recombinant DNA which comprlses a cloning vehicle or vector DNA sequerlce and a DNA sequence foreign thereto which consists of or includes substantially the following sequence of 129 nucleotides (whlch 05 should be read in rows of 30 across the page):-(5') ATGT M CATG TAACATTAAG M TGGCAGAT
GCGAGCAGTT TTGTAAAAAT AGTGCTGATA
ACAAGGTGGT TTGCTCCTGT ACTGAGGGAT
ATCGACTTGC AGAAAACCAG AAGTCCTGTG
10AACCAGCAG (3') (1) and/or its complement.
The invention includes particularly recombinant DNA whichcomprises a cloning vehicle DNA sequence and a sequence foreign to the cloning vehlcle, ~herein the foreign sequence includes substan-tially the whole of an exon sequence oE the human factor IXgenome. The 129-nucleotide sequence described above corresponds substantially ~o such an exon sequence. Another such exon sequence which independently characterises the human factor IX DNA
is the 203-nucleotide sequence substantially as follows (again reading in rows of 30 across the page):-(5~) TGCCATTTCC ATGTGGAAGA GTTTCTGTTT
CACAAACTTC TAAGCTCACC CGTGCTGAGG
CTGTTTTTCC TGATGTGGAC TATGT~AATT
CTACTGAAGC TGAAACCATT TTGGAT M CA

TCACTCGGGT TGTTGGTGGA GM GATGCCA
AACCAGGTCA ATTCCCTTGG CAG (3') and/or its complement.
The intron sequences of the human factor IX genome are excised during the transcrip~ion process by which mRNA ls made in humancells. Only exon sequences are ~ranslated into pro~ein. DNA codlng for factor IX has been yrepared frcm human mRNA. This cDNA has been partly sequenced and found to contain the same 129- and 203-nucleo-tide sequences set out above~

The invention also includes ~ecombinant ~NA ~hich comprises a cloning vehicle sequence and a DNA sequence Eoreign to tl1e cloning vehicle, wherein the Eoreign sequence comprises a DNA sequence wllich is complementary to human faetor IX mRN~. Such a recombinan~
05 cDNA can be isolated from a lLb~ary oE recombinant cDN~ cLones derived from human liver mRNA by using an exon of the ~enomic human factor IX DNA (or part the.reof) as a probe to screen this library and thence isolating the resulting clones.
The invention also includes recombinant DNA in whlch the foreign sequence is any fragment of human factor IX DNA, particu-larly of length at leas-t S0 and preferably at least 75 nucleo-tides or base-pairs. It includes such recombinant DNA whether or not part or all of the 1~9 or 203-base-pair sequence defined above.
It includes especial]y part or all of the exon sequences of human factor IX genomic DN~. Various short lengths up to about 11 kilobases (11,000 nucleotides or base-pairs) long have been prepared by use of various restriction endonucleases. I~ethods of isolating recombinant DNA from clones are well known and some are described hereinafter. The DNA of the invention can be single or double stranded form.
The recombinant human factor IX DNA of this invention is useful as a tool of recombinant DNA technology. Thus it is useful as the first stage in the production of artificial human factor IX
and in the preparation of probes for diagnostic purposes.
In the production of artificial human factor IX i~ is contemplated that appropriate cDNA or genomic clones will be introduced into a suitable expression vector in either mammalian or bacterial systems. For mammalian studies, the gene might be too long to be conveniently retained in one clone. There~ore a suitable artificial ~minigene~ will be designed and constructed from suitable parts oE the cDNA and genomic clones. The minigene will be under the control of its own promoter or instead will be replaced by an artificial one, perhaps the mouse metallothioneine I promoter. The resultant 'minigene' will then be introduced into mammalian tissue culture cells e.g. a hepatoma cell line, and selection for clones of cells synthesising maximum amounts of biologically active factor IX will be carried out. Al~ernatively "genetic farming"
could be employed as has been demonstra-ted Eor mouse growth hormone (Palmiter _ al, Nature 300, 611-615, 1982). The minigene would_ be micro-injected into the pronucleus of fertilised eggs, followed 05 by in vivo cloning and selection for progeny produclng the largest quantity of human factor IX in blood. Alternatively, lt i8 contem-plated that the cDNA clone or selected parts of it will he linked to a suitable strong bacterial promotor, e.g. a Lac or Trp promotor or the lamdba PR or PL, and a factor IX polypeptide obtained therefrom.
The natural factor IX polypeptide is synthesised as a precursor containing both a signal and propeptide region. They are both nor~ally cleaved off in the produ~tion of the definitive length protein. Even this product is merely a precursor. It is biologically inactive and must be gamma-carboxylated at 12 specific N-terminal glutamic acid residues in the so called 'GLA' domain by the action of a specific vitamin K-dependent carboxylase. In addition, two carbohydrate molecules are added to the connecting peptide region of the molecule, but it remains unknown whether they are required for activity. The substrate for the carboxylase is unknown and could be the precursor factor IX polypeptide or alternatively the definitive length protein. Therefore vsrious relevant polypeptides both with and without the precursor domains will be "constructed" using genetic engineering methods in bacterial hosts. They will then be tested as substrates for the conversion of inactive to biologically ac~ive factor ~~in vitro by the action of partially purified preparations of the carboxylase enzyme which can be isolated from liver microsomes or other suitable sources.
For diagnostic purposes, the recombinant human genomic factor IX DNA or recombinan~ human mRNA derived factor IX DNA has a wide variety of uses. It can be cleaved by enzymes or combina-tions of two or more enzymes into shorter fragments of DNA which can be recombined into the clonin~ vehicle, producing "sub-clones".
These sub-clones can themselves be cleaved by restriction enzymes to DNA molecules suitable for preparing probes. A probe DNA (by definition) is labelled in some way, conveniently radiolabelled, ~2~2~;

and can be used to examine in detail mutations in the human DNA
whlch ordinarily would produce factor IX. Several different probes have been produced for examining several different regions of the genome where mutation was suspected ~o have occurred ln 05 patients. Failure to obtain hybridisation from such a probe indicates that the sequence of the probe differs in the patient's DNA. In particular it has been shown that Christmas disease can be detected or confirmed by such methodology. Useful probes can contain intron and/or exon regions of the genomic DNA or can contain cDNA derived from the mRNA.
The invention includes particularly probe DNA, i.e. which is labelled, and of a length suitable for the probing use envisaged.
It can be single-stranded or double-stranded over at least the human factor IX DNA probing sequences thereof and such sequences will usually have a length of at leas-t 15 nucleotides, preferably at least 19-30 nucleotides in order to have a reasonable probability of being unique They will not usually be larger than 5 kb and rarely longer than 10 kb.
The invention accordingly includes a DNA molecule, comprising part of the human factor IX DNA sequence, whether or not labelled, whether intron or exon or partly both, whether double-stranded, of the coding strand or the complementary strand there-toO It also includes human cDNA corresponding to part of all of human factor IX
mRNA. It includes particularly a solution of any DNA of the invention~ which is a form in which it is conveniently obtainable by electroelution from a gel.
The inventlon includes, of course, a host transformed with any of the recombinant DNA of the invention. The host can be a bacterium9 for example an appropriate strain of E.coli, chosen according to the nature of the cloning vehicle employed. Useful hosts may include strains of Pseudomonas, Ba _ us subtilis and Bacillus stearother~nophilus, other Bacilli, yeasts and other fungi ._ _ and mammalian (including human) cells.
One process practlsed in connection with this inventlon for preparing a host transformed with the recomblnant DNA of the invention is based on the following steps:-~ r~
~ . ~

~2~

~1~ syn~hesising an oligodeoxynucleotide having a nucleotide sequence comprising that occurring in bovine factor I~ messenger RNA coding for a~ino acids 70-75 or 348-352 of bovine factor IX, and labelling the oligodeoxynucleotide to form a probe 05 (2) preparing complementary DNA to a ~ixture of bovlne ~RNAs;
(3) inserting the complementary DNA in a cloning vector to form a mixture of recombinant bovine cDNAs;
(4) transforming a host with said mixture of recombinant bovine cDNAs to form a library of clones and multiplying said clones;
(5) probing the clones with the synthetic oligodeoxy-nucleotide probe obtained in step 1 and isolating the resultant recombinant bovine factor IX cDNA-containing clone;
(6) digesting the recombinant bov~ne factor IX cDNA from said clone with one or more enzymes to produce a bovine factor IX
cDNA molecule comprising a shorter sequence of bovine factor IX
DNA, but preferably at least 50 base-pairs long; and (7) probing a library of recombinant human genomic DNA in a transformed host with the shorter sequence bovine factor IX cDNA
molecule, to hybridise the human genomic DNA to the said recombinant bovine factor IX DNA and isolating the resultant recombinant DNA-transformed host.
Brief description of the drawings Figure 1 shows the structure of a published amino-acid sequence of bovine factor IX polypeptide, the deduced sequence of the m~NA from which it would be translated and the structures of oligonucleotides (oligo-N1 and N2) synthesised in the course of this invention;
Figures 2 and 3 show the chemical formulae of "building blocks" used to synthesise the oligonucleotides referred to in Figures 1 and 11;
Figure 4 is an elevational view, partly sectloned, showing an apparatus for synthesis~ng oligonucleotides;
Figure 5 shows the sequerlce of part of the bovine factor IX
cDNA obtained in thls invention;

Figure 6 is a map showing the organisation oE an approxi-macely 27 kb length of human factor IX genomic DNA and is divided into five portions, showing:-(a) the exon regions;
05 (b) the 11,873- nucleotide length sequenced;
(c) cDNA molecules obtained by restriction with various endonucleases, sub-cloned and subsequently used as probes;
(d) DNA molecules obtained by restriction with various endonucleases; and (e) three regions of human factor IX genomic DNA derived from three clones in lambda phage vector.
Figure 7 shows the sequence of the DNA of Figure 6(b) and in parts the encoded protein;
Figure 8 shows a restriction enzyme chart of the sequence shown in Figure 7;
Figure 9 shows part of the sequence of the human Eactor IX
cDNA and its encoded protein;
Figure 10 shows the structure of a pair of complementary oligonucleotides (oligo N3 and N4) synthesised in the course of this invention;
Figure 11 shows part of the DNA sequence of the vector pAT153/PvuII/8 of this invention, in the region where it differs from pAT153;
Figure 12 is a diagram of plasmid pHIX17 of the invention showing the origin of the 104 kb fragment used for probing and initial sequencing; and Figure 13 shows the position of the major radioactive bands on probing a "Southern blot" of normal human DNA, cut by the restriction enzymes EcoRI(E), HindIII(~ II(B) and BcII(Bc), with a sub-clone of the recombinant human factor IX DNA of this invention.
Description of preferred embodiments 1 General de6cription .

A recombinan~ DNA of the invention can be extracted by means of probes from a library of cloned human genomic D~A. This is a ~nown recombinant library and the invention does not, oE course, extend to human genomic factor IX DN~ when present in ~uch a llbrary. The probes used were of bovine factor IX cDNA (DNA
complementary to bovine mRNA), which were prepared by an elaborate 05 process lnvolving firstly the preparation of recomblna~t bovine cDNA from a bovine mRNA starting material, secondly the chemical syntheses of oligonucleotides, thirdly their use to probe the recombinant bovine cDNA, in order to extract bovine factor IX cDNA
and fourthly the preparation of suitable probes of shorter length from the rec~mbinant bovine factor IX cDNA. The first probe tried appeared to contain an irrelevant sequence and the second probe tried, not containing it, proved successful in enabling a single clone of the human genomic factor IX DNA to be isolated. This clone is designated lambda HIX-l. The steps involved are described in more detail in the sub-section "Examples" appearing hereinafter, and the second probe comprises the 247 base-pair D~A sequence of bovine factor IX cDNA indicated ln Figure 5 of the drawings. The invention therefore provides specifically a recombinant DNA which comprises a cloning vehicle sequence and a DNA sequence foreign to the cloning vehicle, which recombinant DNA hybridises to a 247 base pair sequence of bovine factor IX cDNA indicated in Figure 5 (by the arrows at each end thereof).
The cloning vehicle or vector employed in thP invention can be any of those known in the genetic engineering art (but will be cho~en to be compatible with the host). They include E.coli.
plasmids, e.g. pBR322, pAT153 and modifications thereof, plasmids with wider host ranges, e.g. RP4 plasmids specific to o~her bacterial hosts, phages, especially lambda phage, and cosmids. A
cosmid cloning vehicle contains a fragment of phage DNA including its cos (cohesive-end site) inserted in a plasmid. The resultant recombinant DNA is circular and has the capacity to accommodate very large fragments of additional foreign DNA.
Fragments of human factor IX genomic DNA can be prepared by digesting the cloned DNA with various restriction en~ymes. I~
desired, the fragments can be religated to a cloning vehicle to prepare further recombinant DNA and thereby obtain "sub-clones".

~2~

In connection with this embodiment a new cloning vehicle has been prepared. This ~s a modified pAT153 plasmid prepared by llgating a Ban~lI and UindLII double digest of pAT153 to a pair of comple-mentary double sticky-ended oligonucleotides having a DNA sequence 05 providing a Ban~lI restriction residue at one end, a ~llndIII restrlc-tlon resldue at the other end and a PvuII restriction slte ln between.
While the inYention is described herein with reference to human genomlc factor IX DNA in particular, the invention includes human factor IX cDNA (complementary to human factor IX mRNA) which contains subslantially the same sequences. A library of human cDNA
has been prepared and probed with human factor IX genomic DNA to isolate human factor IX cDNA from the library. For this purpose the probe DNA is conveniently of relatively short length and must include at least one exon sequence. The invention therefore includes a process of preparing recombinant D~A of the invention, to produce a hybrid of (a) library recombinant DNA encoding a translation product convertible in vivo with the aid of vitamin K-dependent carboxylase into human fac~or IX and (b) probe DNA, which process comprises probing a library of clones contain-ing recombinant DNA complementary to human mRNA with a probe comprising a labelled DNA comprising a sequence complementary to part or all of an exon region of the human factor IX genome, and iqolating library recombinant DNA from the probe DNA.

2. Examples A. Bacteria used E.coli K-12 strain MC 1061 (Casadaban & Cohen, J.Mol.
Blol. 138, 179-2079 1980), E.coli K-12 strain ~B 101 (Boyer &
Roulland-Dussoix~ J.Mol.Biol. 41, 459-472; 1969) and E.coli K-12 strain K803 which is a known strain used by genetic engineers~
B. Source and purification of bovine factor IX~ anti-bovlne Highly purified bovine factor IX and rabbit anci-bovine factor IX antiserum were gifts from Dr. M.P. Esnouf. Analysis of the purifled bovine factor IX on a denaturating polyacrylamide gel showed that it has a purity of greater ~han 99%. Specific anti-factor IX immunoglobulins used for immunoprecipitation experlments were purified as described by Choo et al., Bio-chem.J. 199, 527-535, 1981, by passage of the crude antiserum through a Sepharose-4B column onto which pure bovlne factor IX has been coupled 05 Bovine mXNA was obtained from calf Liver and isol~ted by the guanidine hydrochloride method (Chirgwin et al., Blo-chem. 18, 5294-5299, 1979). The mRNA preparatlon was passaged ~hrough an oligo dT-cellulose column (Caton and Rober~son, Nucl.
Acids Res. 7, 1445-1456, 1979) to isolate poly(A) ~ mRNA. Poly(A) mRNA was translated in a rabbit reticulocyte cell-free system in the presence of 35S-cysteine as described by Pelham and Jackson (Eur. J.Biochem. 67, 247-256~ 1976). At the end of the transla-tion reaction, fac~or IX polypeptide was precipitated by the addition of specific anti-factor IX immunoglobulins. The immunoprecipitation procedure was as described by Choo et al., Biochem.J. 181, 285-294, 1979. The immunoprecipitated material was washed throughly and resolved on a two-dimensional SDS-poly-acrylamide gel (Choo et al., Biochem.J. 181, 285-294, 1979), by isoelectric focussing in one dimension and elec~rophoresis in another. Some polypeptides of known molecular weight were subjected ~o this procedure, to serve as reference points. The immunoprecipitated material showed 4 pronounced spots, all in the 50,000 molecular weight region and with separated isoelectric points. These predominant spots of molecular weight abou~ 50,000 represent a single polypeptide chain plus a possible prepeptide signal sequence, a deduction compatible with published data (Katayama et al.,Proc. Natl.Acad. Sci.USA 76, 4990-4994, 1979).
~ len the gel analysi~ was repeated for the same material but immunoprecipitated in the presance of unlabelled pure bovine factor IX, the 4 spots appeared a~ reduced intensi~y, indicating that the translatlon product ls specifically competed for by pure factor IX. Thirdly, immunoprecipitation was performed uslng a control rabbit antiserum, i.e. from a rabbit which had not been i~munised with fac~or IX. None of the 4 spots appeared. These results therefore indicate ~hat the translation product was a factor IX polypeptide.

2~i The speclfic immunological/cell-free translation assay established above was used ~o monitor ~he enrichment of fac-tor IX
mRNA on sucrose gradient centrifugations. Total poly(A)~rnF~A ~as resolved by two successive separations by sucrose gradient centri~
05 fugations. When individual fractiorls from the gradlen~ were assayed by the above method, a fraction of size 20-22 Svedberg units (approx. 2.5 kilobases of RNA) region was found to be enriched (approx. ten-fold) for the bovine factor IX mRNA. This enriched fraction was used in the subsequent cloning experiments. 0 C. Synthesis o~ specific bovine factor IX deoxyoligonucleotide mixtures Starting from a knowledge of the amino acid sequence of bovine fac~or IX (Katayama e~ al. 9 Proc.Natl.Acad.Sci.
USA 76, 4990-4994, 1979), the synthesis of two mixtures of oligonucleotide probes was designed. These probes consisted of DNA sequences coding for two different regions of the protein.
The regions selected were those known to differ in sequence in the analogous serine proteases, pro~hrombin, Factor C and Factors VII
and X and were those corresponding to amino aclds 70-75 and 3~8-352 respectively. The 70-75 region was particularly favourable in ~hat the mixture of oligonucleotides synthesised, i.e. oligo N2A
and oligo N2B, contained all 16 possible sequences that might occur in a 17 nucleotide long region of the mRNA corresponding to amino acids 70-75. The oligo N2A-N2B mixture is hereinafter called "oligo N2" for brevity.
Figure 1 of the drawings shows the two selected regions of the known amino acid sequence of bovine factor IX, the corres-ponding mRNA and the oligonucleotides synthesised. Since some of the amino ~cids are coded for by more ~han one nucleotide triplet, there a}e 4 ambiguities in the mRNA sequence shown for amino acids 70-75 and ~herefore 16 possible individual sequences.
The nucleotide mix~ures oligo Nl and oligo N2 were synthe sized using the solid phase phosphotriester method of Duckworth et alO, Nucl.Acids Res. 99 1691~1706, 1981, modlfied in two ways.
Firstly, o-chlorophenyl rather than p-chlorophenyl bloeking groups were used for the phosphotriester grouping, and were incorporated in the mononucleotide and dinucleotide "building bLoc~g". Figures 2 and 3 of the drawillgs show (a) dinucLeotide and (b) mononucleotide "building blocks". DMT = 4,4' - dimethoxytrityl and B = 6-N-ben~vyl-adenin-9-yl 9 4-N-benzoylcytosin-1-yl, 2-N-isobutyrylguanin-9-yl or 05 thymin-1 yl, depending on the nucleotide seLected. Secondly, the "reactioll cell" used for the successive addition of mono- or din~lcleotide "building blocks" was miniaturised so that the cou~-ling step with the condensing agent 1-~mesi-tylene-2-sulphonyl)-3-nitro-1,2,4-triazole (MSNT) was carried out in a volume of 0.5ml pyridine containing 3.5 micromoles of polydimethylacrylamide resin, 17.5 micromoles of incoming dinucleotide (or 35 micromoles of mononucleotide) and 210 micromoles of MSNT.
Figure 4 of the drawings is an elevational view of the micro-reaction cell 1 and stopper 2 used for oligonucleotide synthesis, drawn 70% of actual si~e. The device comprises a glass-to-PTFE
tubing joint 3 at the inlet end of the stopper 2. The stopper has an internal conduit which at its lower end passes into a hollow tapered ground glass male member 5 and thence into a sintered glass outlet 6 to the stopper. The cell 1 has a ground glass female member 7 complementary to the member 5 of the stopper, leading to reaction chamber 8, the lower end of which terminates in a sintered glass outlet 9. This communicates with glass tubing lO, and a 1.2mm. "Interflow" tap 11. Further glass tubing 10, beyond the tap 11, leads to -~he outlet glass-to-PTFE
tubing joint 12. Pairs of ears 13 on the stopper and ce~l enable them to be joined -together by springs (not shown) in a liquid-tight manner.
After completion of the synthesis and deprotection, fractiona-tion was carried out by high pressure liquid chromatography ~Duckworth et al. 9 see above) and the peak tubes corresponding to the product of correct chain length were located by labelling of fractions at ~heir 5'-hydroxyl ends using [gamma-3 P3-~TP and T4 polynucleotide kinase, followed by 20% 7M urea polyacrylamide gel electrophoresis.
The position on the gel of the 17- and 14- oligonucleotides was determined by separately labelling, by the method described ' i -~2~

above, 17- and 1~~ nucleotide Long "marker" oligonucleotides ancl subjecting these to the same geL electrophoresLs.
D. ~ ion o~ libraries of cDNA sequences or bovlne ml~N~
_ __ Two diEferent approaches were used for the generatLon of 05 cloned cDNA library:-(i) MboI libr~ First strand cDN~ was synthesised usin~ the sucrose gradient-enriched poly(A) ~ bovi.ne ml~NA as template. The conditions used were as described by ~luddlestorl & Brownlee, Mucl.
Acids Res.10, 1029-1030, 1981, except that 2 micrograms of oligo N-1, 20-30 micrograms of the mRNA, 10 microcuries [alpha-3 P]
-dATP (Amersham, 3000 Ci/mmole), and 50 U of reverse transcriptase were used in a 50 microlitre reaction. Oligo N-1 hybridises to the corresponding region on the mRNA (refer to Figure 1) and thereby acts as a primer for the initiation of transcription. It was used in order to achieve a further enrichment for factor IX
mRNA. At the end of the cDNA synthesis reaction, the cDNA was e~tracted with phenol and desalted on a Sephadex-G100 column, before it was treated with alkali ~O.lM NaOH, lmM EDTA) for 30 min. at 60C to remove the mRNA st~and. Second strand DNA
synthesis was then carried out exactly as published (Huddleston &
Brownlee, Nucl.Acids Res. 10, 1029-1038, 1981).
The double-stranded DNA was next cleaved with the restriction enzyme MboI and ligated to the plasmid vector pBR322 which had been cut with BamHI and treated with calf intestinal alkaline phosphatase to minimise vector self-religation. Phosphatase treatment was carried out by incubating 5 micrograms of BamHI-cut pBR 322 plasmid with 0.5 m~crogram calf intestinal phosphatase (Boehringer; in 10mM Tris - HCl buffer, pH 8.0) in a volume of 50 microlitres at 37C for 10 minutes, see Huddleston & Brownlee supra.
The ligated DNA was used to transform E.coli strain MC 1061.
For transformation E.coli MC 1061 was grown to early exponential phase as indicated by an absorbancy of 0.2 at 600 nm and rnade "competent" by treating the pelleted bacterial cells first with one half volume, followed by reyelleting, and then with 1/50 volume ~2~ 2~

of the original ~rowth medium of 100mM CaC12 15% v/v glycerol and lOmM PIPES-NaOH, pH 6.6 at OC. Cells were lmmediately frozen in a dry ice/ethanol bath to -70C. For transformatlon7 200 micro-litre aliquots were mixed with 10 microlltres of the recomblnant 05 DNA and incubated at 0C for 10 minutes followed by 37C for 5 minutes. 200 microlitres of L-broth (bactotryptone 10g., yeas-t extract 5g , sodium chloride 10g., made up to 1 litre with deionised water) were then added and incubation continlled for a further 30 minutes at 37C. The solution was then plated on the appropriate antibiotic agar (see below). A library of about 7,000 ampicillin-resistant colonies was thus obtained. They were ampicillin-resistant because they contained the beta-lactamase gene of pBR 322. OE these, approx. 85% were found to be tetra-cycline-sensitive.
(ii) dC/d~ tailed library In the preparation of this library, _ first strand cDNA was synthesised as described for the above library except that oligo dT(12 18) was used as a primer to initiate cDNA synthesis. Following this, the cDNA was tailed with dCTP using terminal transferase and back-copied with the aid of oligo dG(12 18) primer and reverse transcriptase to give double stranded DNA, e~ac-tly according to the method of Land et al., Nucl.Acids Res. 9, 2251-2266, 1981. After a further tailing with dCTP, this material was annealed by hybridisation to a dGTP-tailed pBR322 plasmid at the PstI site. The hybrid DNA was used to transform E.coli strain MC 10610 A library of approximately 10,000 tetracycline-resistant colonies was obtained. Of these, approxi-ma~ely 80% were found to be sensitive to ampicillin, due to insertion of DNA into the ampicillin-resistant gene at the PstI
site, Eo Isolation of specific bovine factor IX clones (i) From MboI library The library of colonies, in an unordered fashion, was transferred onto 13 Whatman 541 filter papers and amplified with chloramphenicol~ to increase the number of copies of the plasmid in the colonies, as described by Gergen et al., Nucl. Acids Res., 11 2115-2136 (1979). The filters were pre-hybridised ~2~

at 65 C for 4h in 6 x NET (1 x NET = 0.15m NaCl, lmM EDTA, 15mM
Tris-HCl, pH 7.5), 5 x Denhardt'6, 0.5% NP40 non-iorllc surf~lctarlt~
and 1 microgram/ml. yeast RNA as described by Wallace et al., Nucl. Acids Res. 9, 879 894 (19~]). Hybridlsation was carried out 05 at 47 C for 20h in the same solution containlng 3 x 105cpm (().7 nanogram/m1) of labelled oli~o N-2 probe. I,abelling ~as done by phosphorylatLon of the oligonucleotides at the 5' hydroxyl end using [gamma- P] -ATP and T4 phosphokinase (Huddleston & Brownlee, Nucl.Acids Res. _ , 1029-1038, 1981). At the end of the hybrldi-sa-tio~, Eilters were washed successively at 0-4C (2h), 25C
(10 min), 37 C (10 min) and 47 C (10 min). After radioautography of the filters from this screening, one colony showed a positive signal above background. This colony was designated BIX-l clone.
(ii) From dC/dG-tailed library Screening of this library, in an ordered array fashion, using oligo N-2 probe as described above has resulted in the identifica-tion of a positive clone. This was designated BIX-2 clone~
F. Sequence characterisation of bovine factor IX cDNA clones Gharacterisation of BIX-1 clone by restriction endonuclease cleavage indicated that it contained a DNA insert of about 430 base-pairs (data omitted, for brevity). Figure S shows part of the nucleotide sequence of the coding strand, determined by the Maxam-Gilbert method, extending over 304 nucleotidefi and provides direct evidence that it has the identity of a bovine factor IX
sequence. Thus, nearly all of this 304 nucleotide sequence (correspondlng to amino acid residues 52-139) agrees with the nucleotide sequence predicted from the known bovine factor IX
amino acid sequence data (Katayama et al., Proc.Natl.Acad.Sci. 76, 4990-4994, 1979)9 Over this region, ~here are no discrepancies between BIX-L and these publlshed data for factor IXI except at nucleotides 38~40 where the amino acid coded for is Asp instead of Thr. This amino acid change was similarly observed in a second~
independent cDNA clone (BIX-2; see below). The remainder of the 304-nucleotide sequence, i.e. that shown in brackets in Figure 5, does not agree with the published bovine Eactor IX
amino acid data of Katayama.

In Figure 5, the underlined portion denotes the sequence corresponding to the ollgo N-2 probe sequence, the asterlsk denotes a nonsense codon, the brackets enclose a sequence whlch does not correspond to Katayama'.s amino acid data and the arrows 05 indicate HinEI restrictlon sites. The Katayama numbering system for amino acids is shown and this sequence is in the opposi-te orientation to the direction of trarlscrlption of ~he tetracycline-resistant gene of -the plasmid.
By similar methods, BIX-2 clone was found to have a DNA
insert of 102 nucleotides and this spans the nucleotide pos~-tions 7-108 as shown in Figure 5. The nucleotide sequences for BIX-l and BIX-2 clones over this region (nucleo~ide 7-108) were identical.
G. Isolation oF human factor IX gene (i) Initial clone - lambda HIX-l A library of cloned human genomic DNA, namely a HaeIII/AluI
lambda phage Charon 4A library prepared by Lawn et al., Cell, 15, 1157-1174, 1978, was used. 106 phage recombinants from this library were screened using the in situ plaque hybridisation procedure as described by T. Maniatis et al., Cell, 15, 687, 1978.
Pre-hybridisation and hybridisation were carried out at 42C
in 50% formamide. After hybridisation, filters were washed at room temperature with 2 x SSC (1 x SSC = 0.15mM NaCl, 15mM sodium citrate, at pH 7.2) and 0.1% SDS, then at 65C with 1 x SSC
and 0.1% SDS.
Two DNA molecules, being restrlctlon fragments from the factor IX cDNA cloned ln BIX-l~ were radiolabelled and used as probes in the hybridisation. The first fragment corresponds to nucleotide numbers -8 to 317 on the numbering system of Figure 5, and was isolated by Sau3AI digestion of BIX-l plasmid DNA. The -isolated DNA was labelled to high specific activity by incorpora-tion of [alpha - P] -dATP using a nick translation (Rigby et al., J. Mol.Biol. 113, 237-251, 1977, modifled, vide infra). Using this probe, 10 clones were isolated. These were plaque-purified and re-hybridised with a 247-nucleotide fragment from BIX-l clone.
This fragment, derived from nucleotides 3-249 can be sePn fro~

~2~

~ 18 -Figure 5. It contains only sequences in agreement with the Katayama bovine factor IX amino acid sequence and was lsolated b~
HinfI digestion of BIX-L plasmid DNA. Only a single clone gave a positive hy~ridisation signal with this 247-nucleotide probe~
05 Thls clone ~as further plaque-purified and the resulting clone was designated "lambcla HIX-l".
(ii) Subse~uent_genomic clones A sub-clone, pATIXcVII, of recombinant human factor IX cDNA
from human liver mRNA, and prepared as described in Section L
below, was linearised by digestion with HindIII and BamHI. The resulting 2 kb cDNA molecule was purified by 1% agarose gel electro-phoresis. After electroelution, about 100 ng of this cDNA was nick-translated with [alpha p] dATP (see abo~e~ and used as a hybridisation probe to screen the HaeIII/AluI lambda phage Charon 4A
human genomic DNA library for further genomic clones, using standard stringent hybridisation conditions. Two further human factor IX
genomic clones, designated lambda HIX-2 and lambda HIX-3, were thus obtained.
Ho Characterisation of human fac_or IX genomic clones (i) Restriction ma~
The initial lambda HIX-l clone was characterised by cleavage with various single and double digests with different restriction endonucleases and Southern blotting of fragments using the bovine factor IX cDNA probe (results omitted for brevity). The subse-quently isolated lambda HIX-2 and 3 clones were characterised in the same way except that the human cDNA probe, pATIXcVII (see Section L below) was used for the Southern blots. From these results it emerged that the sequences in the fac~or IX genome corresponding to lambda HIX 2 and 3 overlapped with lambda HIX-1 as shown in Figure 6(e). In Section (d) of this Figure 6 are summarised the results of the analysis using the restrlction en~ymes EcoRI (E)~ HindIII (H), B~ B), BamHI ~Ba) and PvuII (P), and this serves as a restriction enzyme map.
(ii) SequencinQ
Numerous sub-clones were isolated from a knowledge of the ~e ~ + ~ c ,-Ee~r~ enzyme map as described ln Section J~ii) below, ~he majority in a vector pAT153/PvulI/3. ~xamples o~ these sub-clones are sho~n in Figure 6(c) and a number were used and were of a convenient length fo~ sequence analysis by the Maxam-Gilbert method (Maxam & Gilbert, Proc.Natl.Acad.Sci.US~ 74, 56-564, 1980~.
05 lnitially sequencing was done on part of a 1.4 kb ~ RI
restriction fragment from the sub-clone pl-lIX-17, see below and J(i). A 403~nucleotide (base-pair) length was sequenced, of which a 129-nuc]eotide length was identified as lying ~tithin an exon region. This is the 129-nucleotide sequence used above to define the factor IX DNA.
Subsequently, a region of 11873 bases was sequenced in the central portion of the gene [see Figure 6(b)]. Figure 7 shows the sequence of one strand of the DNA. The nucleotides are arbitrarily numbered from 1 to 11873 in the 5' to 3' direction. The original 403-nucleotide sequence runs from Figure 7 nucleo-tides Nos. 4372 to 4774 and is indicated by 0-0'. The 129-nucleotide sequence lying within the 403 one, runs from Figure 7 nucleotides Nos. 4442 to ~570 and is indicated by J-J'. This corresponds exactly to the "w" exon.
In detail, the sequence of nucleotides Nos. 1-7830 contains two short exons (nucleotides 4442-4570 and 7140-7342 respectively) marked w and x in Figures 6(a) and 9, J-J' and J'-J" in Figu~re 7.
These code for amino acids 85-127, and 128-195 respectively of the amino acid sequence predicted from the human factor IX cDNA clone (Figure 9). There are no differences in amino acid sequences predicted from the genomic and cDNA clones of the invention in these two exon regions. The sequence of the gene between residues 7831-11873 is less complete, containirlg several gaps, but is still a useful characterisation oE the gene as it contains two "AluI repeat" sequences, nucleotides 7960-8155 and 9671-9938.
AluI sequences are found in many genes. The repetition is not exact but there is a typical degree of homology between them.
This further characterisation provides a useful cross check on the accuracy of the restriction enzyme map. This emerges more clearly from the re~triction enzyme chart of Figure 8.

~2~

Figure 8 1s a chart produced by a computer analysis of the sequence data of the 11873 nucleotide long sequence of Flgure 7.
Column 1 of Figure 8 gives the arbltrary nucleotlde nu~ber allotted to the nucleo~ide of Figure 7. Column 2 appor-tlons the nucleotide 05 number as a fraction of -the whole sequence. Column 3 shows the restriction enzymes wh:Lch will cut the DNA withln varlous short sequences of nucleotides shown ln Column 4, The short sequences of Column 4 begln with the nucleotide numbered in Column 1. Wlth the aid of this chart the positions of the restrictlon sltes shown in Figure 6(d) and some of the sequences shown ln Flgure 6~c) can be determined very accurately. For example sequences II-IV are produced by restriction at the following sites (denoted by the first nucleo-tide number at the 5' end of each site).

]5 III 6380 - 7378 Par-ticularly important sites are arrowed in Figure 8. Some of the relevant nucleotide numbers are shown in Figure 6(c), the number given being that of the nucleotide at the 5' end of each site.
Further sequence analysis of the sub-clones V, VI, VII
and VIII shown in Figure 6(c) indicates that the factor IX gene is divided into at least 7 exon regions separated by at least 6 introns. The positions of the exons are shown in Figure 6(a) by the solid blocks labelled t, u, v, w, x, y and z. The llZl- exon is much the longest and its 3'-end coincides with the 3'-end of the mRNA. The location of these exons relative to the cDNA sequence is discussed below (.section L) and it is clear that the "t" exon shown in Figure 6(a) is not a marker for the 5'-end of the gene, as its sequence fails to match that of the extreme 5'-end of the cDNA clone (see below). This suggests that the factor IX gene will be longer at lts 5'-end than the 27 kb region shown in Figure 6, and will contain at least one further exon.
Additionally, pHIX-17 DNA was digested with EcoRI. The digested material was resolved on 0.8% agarose gel and a 1O4 kb fragment was isolated in solution by electroelutlon. It can be stored in the usual manner. This 1.4 kb long molecule was us~d for the initial seqllencing. Only about 1.0 kb i5 lnserted DI~A, the re~aining 0.4 kb being of pBR322. A 403 nucleotlde length of 05 the lnserted DNA was sequenced and ls i~entlfied as 0--0' in ~lgure 7. The same 1.4 kb fragment was also labelled and used as a probe in Section M.
I. Construction of a vector pAT153/PvuII/8 A derivatlve of the plasmid pAT153 (Twig & Sherratt, 10Nature 283, 216-218, 1980) was prepared for subclonlng of PvuII
fragments of actor IX genomlc clones, and for ease of charac-terisation of the resultant subclones. Two partially complemen-tary synthetic deoxyoligonucleotides, oligo N3, and, oligo ~4, were synth~sised by the solid phase phosphotriester method described in Section C above. Each has "overhanging" BamHI and HindIII recognition sequences and an internal PvuII recognition sequence. Figure 10 shows the structures of oligo N3 and oligo N4.
BamHI and HindIII cleave ds DNA to leave sticky or "overhanging"
ends. For example HindIII cleaves - AAGCTT
- TTCGAA
between the adenine-carrying nucleotides of each strand leaving the sticky-ended complementary strands:-- A

which are present in the oligo N3/N4 combinatlon.
pAT153 was digested with HindIII and BamHI and the 3393 nucleotide long linear fragment was separated from the 346 nucleo-~ide shorter fragment by 0.7% agarose gel electrophoresis, followed by electroelution of the appropriate bands visualised by ethidium bromide fluorescence under UV light. After treatment with calf intestinal phosphatase, as described in Seetion D(i), the BamHI-HindIII 3393~10ng fragment was ligated to an equimolar mixture of oligo N3 and oligo N4 which themselves had been pretreated, as a mixture, with T4 polynucleotide kinase and ATP, to phosphorylate ~ 22 -their respective 5'-terminal OH groups. After transforming competent MC 1061 cells (see above) and plating on L-broth platcs containing 20 micrograms/ml final concentration of ampicillin, 11 colonies were selected for further analysis. 1 ml plasmid prepara-tlon, see 05 Ho:Lmes and Quigley, Analytical Biochem. 114, 193-197 (1981), was isolated from the 11 colonies. The plasmid DNA was then analyse~
Eor its ability to be linearised by the restriction enzymes BamHI, H III and PvuII. Four clones were positive in this assay and one, labelled pAT153/PvuII/8, was selected for sequence analysis by the Maxam-Gilbert method across the newly constructed section of the plasmid. This part of the sequence is shown in Figure 11 along the unique restriction sites. The novel part of the plasmid sequence is underlined: the remainder is present in the parent plasmid pAT153. The vector allows blunt-end cloning (after treatment with phosphatase) into the inserted PvuII site. The cloned DNA can be excised, assuming that it lacks appropriate internal restriction sites, with BamHI/HindIII, BamHI/ClaI or BamHI/F.coRI double digests. The sites adjacent to the PvuII site are also convenient for end labelling with 3 P for characteri-zation of the ends of cloned DNA by the Maxam-Gilbert sequencing method.
J. Sub-cloning of human factor IX gene The following subcloning experiments were carried out as a first step towards sequencing of the factor IX gene, and to facili-tate the isolation of a small DNA fragment to be used as a probefor the analysis of genomic DNA from haemophilia B patients (see sections M).
(i) Sub-cloning into pBR322 plasmid An approximately 11 kilobase BglII fragment (see Figure 6) within the factor IX DNA insert in lambda HIX-l clone was inserted into the BamHI site o~ pBR322. Transformation was carried out in the E. coli strain, HB 101. The resulting "sub-clone" was desig-nated pHIX-17 (Figure 12).
(ii) Sub-cloning into pAT153/PvuI _ (a) Plasmid DNA from pHIX-17 was prepared and cleaved wi~h PvuII.
Five discrete fragments, all derived from the DNA insert of pHIX-17, 2~

were isolated. The sizes of these fragments were approxl-mately 2~3, 1.3, 1.2, 1.1 and 1.0 kilobases. These fragments were blunt-end ligated into the PvuII site of the pAT153/PvuII/8 vector and ~ransformed into ~. coli HB 101. Flve clones of recombinan-t 05 DNA which carrled the 2.3, 1.3, 1.2, 1.1 and 1.0 kb fragTnents ~7ere obtained and these were designated pATIXPvu 1, 2, 3, 4 and 5 respectively. Factor IX DNA from pATIXPvu-2 is abbreviated as IV
and pATIXPvu-5 as III in Figure 6(c).
~b) Phage DNA from the lambda HIX-l genomic clone was digested with EcoRI. Three dlfferent fragments (approximately 5, 2.3, 0.96, kb;
see Figure 6), all derived from the insert into the phage, were isolated and inserted in pAT153/PvuII/8 vector at the EcoRI site and cloned in E.coli HB 101 to form sub-clones The three result-_ _ .
ing clones for each of these fragments were designated pATIXEco-1, 2 and 4 respectively which are shown ln the restriction map of Figure 6(d). pATIXEco-1 was further digested with both EcoRI and ~_II, and the "overhanging ends" of the restriction sites filled in with deoxynucleoside triphosphates using the Klenow fragment of DNA polymerase I. After isolation of the resulting 1.1 kb fragment by agarose gel electrophoresis and electroelution, it was blunt-end ligated using T4 DNA ligase into the PvuII site of pAT153/PvuII
and allowed to transform E.coli MC 1061. The resultant sub-clone was designated pATIXBE and the factor IX DNA sequence thereof is abbreviated as II in Figure 6(c).
(c) Phage DNA from lambda HIX-2 was digested with HindIII and EcoRI giving a 1.8 kb and a 2.6 kb fragment amon~st others. These fragments were eluted separately, filled in as described in (b) above, cloned as above into the PvuII site of pAT153/PvuII/8 and allowed to transform E.coll MC 1061. The resultant clones were designated pATIXHR-l, and the facto~ IX DNA sequence thereof is abbreviated as V in Figure 6(c), and pATIXEco-6 and the factor IX
DNA sequence thereof is abbreviated as VI in Figure 6(c).
(d) Phage DNA from lambda HIX-3 was digested with F.coRI and H III and the fragments of 2.3 kb and 20 7 kb were sub-cloned exac~ly as described in (c) above. The resultant clones ~ere - 2~ -designated pAT:[XEH-1, abbreviation VII in Figure 6(c), and pATI~lE-2, abbreviation VIII in Figure 6~c).
K. Preparation of a library of cDNA clones from human liver -mRN~
Messenger RNA was extracted from a human liver and a 20-22 05 Svedberg unit enriched fraction of mRNA prepared exactly as described for bovine mRNA in Section B above, except that a 'trsnslation assay' was not used. The first steps in the con-struction of the double-stranded DNA were carried out using the 'Stanford protocol' kindly supplied from Professor P Berg1s depart-ment at Stanford University, USA. This itself is a modificationof Wickens, Buell & Schimke (J.Biol.Chem. 253, 2483-2495, 1978) and some further modifications, incorporated in the description given below were made in the present work.
For the first strand cDNA synthesis 6 micrograms of poly(A)+
~0-22S human mRNA was incubated with 5 microlitres of 10~ buffer (0.5 M Tris-chloride, pH 8.5 at room temperatur , 0.4 M KCl, 0.08M
MgC12 and 4 mM dithiothreitol), 20 microlitres of a 2.5 mM mixture of each of the four deoxynucleoside triphosphates, 0.5 microlitres of ollgo dT(12 18)~ 1 microlitre (containing 0.5 microcurie~ of [alpha- P] -dATP, 2 microlitres of reverse transcriptase (14 units per microlitre) and the volu~e made up to 50 microlltres with deionized water. After incubatlon for 1 hour at 42C, the solution was boiled for 1~ minutes and then rapidly cooled on ice.
The second strand synthesis was carried out by adding directly to the above solution 20 microlitres of 5x second strand buffer (250 mM
Hepes/KOH p~ ~.9, 250 ~M KCl, 50 mM MgC12), 4 microlitres of a 2.5 mM mixture of each of the four deoxynucleoside triphos-phates, 10 microlitres of E.coli DNA polymerase I (6 units per microlitre) and making the volume of the solution up to 100 micro-litres with deionized water. After incubation for 5 hours at 15C, Sl nuclease digestion was carried out by the addition of 400 micro-litres of S1 nuclease buffer (0.03 M sodium acetate pH 4.4, 0.25 M
NaCl, 1 mM ~nS04) and 1 microlitre of Sl nuclease (at 500 units per microlitre). After incubating for 30 minutes at 375, 10 microlitres of 0.5 M EDTA (pH ~.0) was added. Double stranded DNA
was deproteinised by shaking wi~h an equal volume of a phenol:

chloroform (1:1) mix~ure, followed by ether extraction of the aqueous phase and precipi~ation of ds DNA by addition of 2 volumes of ethanol. After 16 hours a-t -20 C, ds DNA was recovered by centrifugation. DNA polymerase I "fill in" of S1 ends was carried 05 out by a further incubation of the sample dissolved in 25 micro-litres of 50 mM tris~chloride, pH 7.5, 10 mM MgC12, 5 mM
dithiothreitol and containing 0.02 mM dNTP and 6 units of DNA
polymerase I. After incubating for 10 minutes at room tempera-ture 9 5 microlitres of RDl'A (0.1 M at pH 7.4) and 3 microlitres of 5% sodium dodecyl sulphate (SDS) were added.
The following part of the protocol differs from the 'Stanford protocol'. The sample was fractionated on a "mini"-Sephacryl S400 column run in a disposable 1 ml pipette in 0.2 M NaCl, 10 mM
Tris-chloride, pH 7O5 and 1 mM EDTA. The first 70% of the "break-through" peak of radioactiYity was pooled (0.4 ml) and deproteinisedby shaking with an equal volume of n-butanol:chloroform (1:4). To the aqueous phase was added 1 microgram of yeast RNA (BD~) as carrier followed by 2 volumes of ethanol. After 16 hours at -20C
double stranded DNA was recovered by centrifugation for blunt-end ligation into calf intestinal phosphatase-treated, PvuII-cu~
pAT153/PvuII/8, using T4 DNA ligase (see I and J(ii) above).
After performing a trial experiment, it was found that when the bulk of the sample was incubated with 200 nanograms of vector DNA
in a suitable buffer (1 mM ATP, 50 mM Tris-chloride, pH 7.4, 10 mM
25 MgC12 and 12 mM dithiothreitol) and using 10 microlitres of T4 DNA
ligase in a total volume of 0.2 ml, then on subsequent transforma-tion of competent E.coli MC 1061 cells a total of 58,000 ampicillin-resistant colonies were obtained. Up to 20% of these were estimated to derive from "background" non-recombinants derived by religation of the vector itself. This 20-22S cDNA library was amplified by growing the E.coli for a further 6 hours at 37 C. 1 ml aliquots of this amplified library ~ere stored at -20C in L broth contain-ing 15% glycerol, before screening for factor IX cDNA clones.
L. Isolation and sequence analysls of human factor IX cDNA clones 6000 colonies of the amplified 20-22S human cDNA library were plated on each of ten 15 cm agar plates and after growing overnight ~2~
- 26 ~
were blotted into Whatman 541 filter paper. After preparing filters for hybridisation as described in section E(i) above~
the lmmobilised colonies were probed wlth a 1.1 kb molecule o~
[alpha-32P] -nick translated human factor IX genomlc DNA isolated 05 from the pATIXBE subclone (Section J, above). This llnea~ 1.1 kb section of factor IX genomic cDNA was isolated from pATIXBE by cleavage with the restrictiol1 enzymes BamHI and HindIII, followed by separation of the 1.1 kb section from the vector by 1.5~ agarose gel electrophoresis. After electroelution, nick-translation was carrled out as before and the material used in a hybr1disation reaction for 16 hours at 65 C in 3x SSC, 10x Denhardts solu-tion, 0.1% SDS and 50 micrograms/ml sonicated denatured E.coli DNA
and lO0 micrograms/ml of sonicated denatured herring sperm DNA.
After hydridisation filters werP washed at 65 C successively 15 in 3x SSC, 0.1% SDS (2 changes, half an hour each) and 2x SSC, 0.1%
~DS (2 changes, half an hour each). After radioautography, 7 clones were selected as positive~ but on dilution followed by re-screening by hydridisation as above, only 5 proved to be positive. Plasmid DNA was isolated from each of these 5 clones and one, designated pATIXcVII, was selected for sequence analysis as it appeared to be the longest of the 5 clones as ~udged by its electrophoretic mobility on 1% agarose gel electrophoresis. A second shorter clone, designated pATIXcVII was also selected for partial sequence analysis.
Sequencing was carried out by the Maxam-Gilbert method and a 2778 nucleotide long section of sequence is shown in Figure 9.
Nucleotides 115 2002 were derived by sequencing clone pATIXcVII.
~The actual extent of this clone is greater as it extends in a 5' direction to nucleotide 170 The sequence between 17 and 111 is inverted with respect to the remainder of the sequence presumably due to a cloning artefact.) Nucleotid s 1-130 were derived from clone pATIXcVI which extends fro~ nucleotides 1-1548 of Figure 9.
The sequence from Nos. 2002-2778 was derived by isolating 4 additional clones designated pATIX108.1, pATIX108.2~ pATIX108.3 and pATIXDB. The first 3 were derived from a mini-library (designa-ted GGB108) of cDNA clones constructed exactly as ~æ~

described in sectlon K above except tha~ sucrose density gradient centrifugation was used instead of chromatography on '1Sephacryl"
S~400 to fractionate the double-stranded DNA according to si~e. A
fractlon of m.w. from 1 kb-5 kb was selected and an amplified 05 ].ibrary of 10,~00 independent clones containing approximately 20%
background non-recombinan-t clones was obtained. Clone pATIXDB
derived from another cDN~ library (designated DBl) constructed as described ln section K except that total poly A~ human liver mRNA
was used as the starting material and sucrose density gradient centrifugation was used to fractionate the DNA according to size as in the construction of the mini-library GGB108. The complexity of this library was 95,000 with an estimated background of non-recomibinants of 50%. Clones pATIX108.1 and pATIX10~.2 were selected from a group of 30 hybridization-positive clones isolated by Grunstein-Hogness screening of the mini library GGB108 using a P-nick ~ranslated probe derived from a Sau3~I restriction en~yme fragment, itself derlved from nucleotides 1796-2002 of clone pATIXcVII. From pATIX108.1 the sequence of nucleo-tides 2009-2756 was determined (Figure 9). Following this the sequence of a part of pATIX108.2, specifically nucleotides 1950-2086, provided the overlap with p~TIXcVII. The remaining 28 hybridization posltive clones were screened by carrying out a triple enzymatic digestion with the restriction en~ymes EcoRI, Ba~HI and HindIII
and screening the product of the digest for an EcoRI restriction fragment extending in the 3' dlrection from the cut at postion 2480.
By this approach, clone pATIX108.3 was selected and sequenced from nucleotides 2642-2778. This clone was followed by three A nucleo-tides, which sequence was conEirmed as a vestigial marker for the poly A tail, by the subsequent isolation of clone pATIXDB by a similar method. pATIXDB was sequenced from Nos. 2760-2778 and ended in 42 A nucleo~ides, thus marking the 3' end of the mRNA.
Figure 9 shows that the predicted amino acid sequence codes codes for a protein of 456 amino acids, but included in this are 41 residues of precursor amino acid sequence preceding the N-terminal tyrosine residue (Y) of the deflnitive length factor IX protein.
The precursor section of the protein shows a basic amino acid ~2~

domain (amino acids -1 to -4~ as well as the more usual hydrophobic signal peptide domain (amino acids -21 to ~36).
The definitive factor IX protein consists of 415 amino acids with 12 potential gamma-carbo~yglutamlc acid residues at amino 05 acids 7, 89 15, 17, 20, 21, 26, 27, 30, 33, 36 and 40. Two potential carbohydrate attachment sites occur at amino acid residues 157 and 167. The activation peptide encompasses residues 146-180, which are cut out in the activation of Factor IX
(see Background of Invention) by the peptide cleavage of an R-A
and R-V bond. This leaves a light chain spanning residues 1-145 and a heavy chain spanning residues 181-415.
The exact location of the boundaries between exons (see Section H, above) and how they are joined in the mRNA is marked in Figure 90 The exons are marked t, u, v, w, x, y, z. It can be seen tha~ there is a rough agreement between the exon domains and the protein regions. For example, the exon for the signal peptide is distinct from that of the ~LA region. Also that of the activa-tion peptide is separated from the serine protease domain.
The 3' non-coding region of the mRNA is ext~nsive, consisting of 1390 residues (including the UAAUGA double terminator 1389-1394 but excluding the poly A tail).
The factor IX cDNA is cleavable by the restriction en~yme HaeIII to give a fragment from nucleotides 133-1440 i.e. a 1307 nucleotide long region of DNA entirely encompassing the definitive factor IX proteir sequence. The nucleotide sequence recognised by ~-~ HaeIII is GGCC. This fragment should be a suitable starting material J for the expression of factor IX protein from suitable promoters in or bacterial, yeast ~ mammalian cells. Another suitable fragment could be deri~ed using the unique S~uI site at residue 41 (corres-ponding to an early part of the hydrophobic slgnal peptide region) and linking lt to a suitable promoter. The nucleotide sequence recognised by StuI is AGGCCT
M. Southern Analysis of normal and patient Christmas disease DNA
(i) Normal The standard (Southern) blotting procedure, Southern, J.Mol.
Biol. 98, 503-5179 1975) was used. In a typical experiment, 10-20 ~2~

- 29 ~
micrograms of human genomic DNA (prepared from uncultured blood cells or cultured lymphocy~lc cells) were digested with one of a number of restriction endonucleases and loaded onto a single gel slo-t. Following electrophoresis on 0.8% agarose gel and transfer 05 onto nitrocellulose it was hybridised with a probe of 3 P- labelled probe II or of 1.4 kb EcoRI fragment (see Sectlon H). Label1ing of the probe was carried out by nick translation using the method of Rigby et al., supra~ modified as follows. About 100 nanograms _ ~ 3~
of the probe was mixed with 40 microcuries of [alpha 'P] dATP
(activity about 3,000 Curies/mMole, obtained from Amersham International PLC~ in 0.05M Tris-HCl, pH 7.5, O.OlM MgC12, 0.001M
dithiothreitol and dCTP, dGTP, dTTP each at a final concentration of 20 micromolar in a volume of 29 microlitres. To this was added 1 microlitre of "solution X" made up of a mixture of 6 units of DNA polymerase I ~E.coli)~ 0.6 nanograms of pancreatic DNase I
(Worthington), 1 microgram of cr~stalline BSA in 10 microlitres of 50~ v/v glycerol containin~ 0.05M Trls-HCl, pH 7~5s 0 OlM MgC12 and O.OOlM dithiothreitol. The mixture was incubated for 2 hours at 15C, after which high molecular weight DNA was purified by chromatography on G-100 "Sephadex"~ Figure 13 shows the major bands obtained with DNA from normal individuals probed with either probe II ~Figure 6) or labelled 1.4 kb EcoRI fragment. With earh of the 4 en~ymes used, EcoRI, HindIII~ II and BclI, a single major band of about 4~8, 5.2, 11 and 1.7 kb was obtained.
The fact that these restriction fragments had tha same length as -those observed in the restriction map of clone lambda ~IIX-l confirmed that the conditions of Southern blotting were precise enough to detect the factor IX gene in total DNA preparations.
This provides the basis for analysis of DNA from the blood of patients with Christmas disease.
(ii) Christmas patients_wi~h~_~ene deletions The value of the probes of the invention for the assay of alterations of genes of some patients suffering from Christmas disease has been demonstrated as follows. Two patients with severe Christmas disease, who also developed antibodies to factor IX, were selec~ed for study. The DNA from 50 ml of blood ~4~$

was digested separately with EcoRI, HindIII, ~_II and BclI and Southern blots prepared for probing with 32P~nick translated probe II (Figure 6). No specific bands wPre observed with either patient under conditions where a control diges-t gave the pa-ttern 05 shown ln Figure 13. Similarly no bands were observed in the patients when probe I, III or IV (Figure 6) was 6ubstituted for probe II. In order to control for possible mischance of some experimental artefac~ giving the observed 'negatlve' signal, a factor IX gene probe (this time pATIXcVII the cDNA probe) was mixed with an irrelevant autosomal gene probe which was specific for the human A1 apolipoprotein (Shoulders and Baralle, Nucl.Acids Res. 10, 4873-4882, 1982). This experiment showed that patient 1 had the normal Al apolipoprotein gene, characterised by a 12 kb band ln an EcoRI digest, and confirmed that he lacked the 5.5 kb band observed with pATIXcVII and characteristic of the normal factor IX gene. It was concluded that both patlents have a sequence of at least 18 kb deleted from their factor IX gene. Two other patients, designa~ed patients 3 and 4, who had also developed antibodies to factor IX gave bands in the normal or abnormal positions on Southern blots with some factor IX gene probes of the invention, but not with others. This suggested that these patients had less extensive deletions of the gene, possibly about 9 kb in length.
These results suggest that diagnosis of haemophiliacs and the heterozygous (carrier) females would be possible in families and this is now under examination. The altered pattern seen in the patient's DNA, whether absence of a band or the presence of a band in an abnormal position, serves as a "disease marker", which can be used to assess for its presence or absence in a suspected carrier. This same test can be applied to antenatal diagnosis, if DNA from foetal cells are available from an amniocentesis. "Genetic diagnosis" should considerably improve existing methods of antena~al diag~osis based on the assay of foe~al factor IX protein levels, with the added advantage that the test can be carried out earlier in pregnancyO Ge~etic methods using natural polymorphisms within the factor IX gene as allelic markers should also ~ake 100% carrier ~f ~1' d~,~ec~,`on ~e~s~ a reality, thereby improving the existing somewhat unsatis-factory methods where probability values are offered to patients.
The following names used in the specifiation are believed to be Registered Trade Marks: "SEPtIACROSE", "SEPIIADEX", "WtIA'rMAN"
05 and "SEPHACRYL".

... ` ~.

Claims (22)

1. Recombinant DNA which comprises a cloning vehicle DNA sequence and a DNA sequence foreign to the cloning vehicle, said foreign sequence encoding a translation product convertible in vivo with the aid of vitamin K-dependent carboxylase into human factor IX.

2. Recombinant DNA according to Claim 1, wherein the foreign sequence comprises substantially the following 129-nucleotide sequence (read in rows of 30 across the page):- and/or its complement.

3. Recombinant DNA according to Claim 1, the foreign sequence comprises substantially the following 203-nucleotide sequence (read in rows of 30 across the page):- and/or its complement.

4. Recombinant DNA according to Claim 1 which comprises a cloning vehicle DNA sequence and a sequence foreign to the cloning vehicle, the foreign sequence being substantially the same as 2 sequence occurring in the human factor IX genome.

5. Recombinant DNA according to Claim 4 wherein the human factor IX
sequence has a length of at least 50 nucleotides.

6. Recombinant DNA according to Claim 4 wherein the length of the human factor IX sequence is from 75 to 11,000 nucleotides.

7. Recombinant DNA according to Claim 1 which comprises a cloning vehicle sequence and a DNA sequence foreign to the cloning vehicle, wherein the foreign sequence includes substantially the whole of an exon sequence of the human factor IX genome.

8. Recombinant DNA which comprises a cloning vehicle sequence and a DNA sequence foreign to the cloning vehicle, wherein the foreign sequence comprises a DNA sequence which is complementary to the human factor IX mRNA.

9. Recombinant DNA according to Claim 4, wherein the cloning vehicle is a modified pAT153 plasmid prepared by ligating a BamHI
and HindIII double digest of pAT153 to a pair of complementary double sticky-ended oligonucleotides having a DNA sequence providing a BamHI restriction residue at one end, a HindIlI
restriction residue at the other end and a PvuII restriction site in between.

10. Recombinant DNA according to Claim 9 wherein the pair of complementary oligonucleotides are of formula:-5' GATCCAGCTCA 3' .......
.......
3' GTCGACTTCGA 5'

11. A host transformed with at least one molecule per cell of recombinant DNA claimed in Claim 1.

12. A host according to Claim 11 in the form of E.coli.

13. A host according to Claim 11 in the form of mammalian tissue cells.

14. A process of preparing recombinant DNA comprising a cloning vehicle DNA sequence and a DNA sequence foreign to the cloning vehicle said foreign sequence encoding a translation product convertible in vivo with the aid of vitamin K-dependent carboxylase into human factor IX, which process comprises probing a library of clones containing recombinant DNA complementary to human mRNA with a probe comprising a labelled DNA comprising a sequence complementary to part or all of an exon region of the human factor IX genome to produce a hybrid of (a) library recombinant DNA encoding a said translation product and (b) probe DNA and isolating the library recombinant DNA from the probe DNA.

15. An artificial DNA molecule encoding a translation product convertible in vivo with the aid of vitamin K-dependent carboxylase into human factor IX.

16. An artificial DNA molecule comprising an at least 15 nucleotide long sequence of part or all of substantially the 129-nucleotide sequence set forth in Claim 2 and/or its complement.

17. An artificial DNA molecule comprising an at least 15 nucleotide long sequence of part or all of substantially the 203-nucleotide sequence set forth in Claim 3 and/or its complement.

18. An artificial DNA molecule according to Claim 15 comprising a sequence of length at least 15 nucleotides substantially the same as a sequence complementary to part or all of that occurring in human factor IX mRNA.

19. An artificial DNA molecule according to Claim 15 comprising a sequence substantially the same as a sequence of length at least 15 nucleotides occurring in the human factor IX genome.

20. An artificial DNA molecule according to Claim 19 comprising substantially only exon sequences.

21. A labelled diagnostic probe comprising a DNA molecule having a single-stranded or double-stranded probe sequence of from 15 to 10,000 nucleotides long of DNA sequence defined in Claim 15.

22. A probe according to Claim 21 having a probe sequence from 20 to 5,9000 nucleotides long.