Inactivation of the 5′-3′ exonuclease of Thermus aquaticus DNA polymerase

BB

Biochi ~mic~a et BiophysicaA~ta

ELSEVIER

Biochimica et Biophysica Acta 1264 (1995) 243-248

Inactivation of the 5'-3' exonuclease of Thermus aquaticus DNA polymerase Louise S. Merkens, Sharon K. Bryan, Robb E. Moses * Department of Molecular and Medical Genetics, Oregon Health Sciences University, Portland, OR 97201, USA Received 21 March 1995; revised 3 July 1995; accepted 10 July 1995

Abstract The gene for Thermus aquaticus (Taq) DNA polymerase enzyme (Taq Pol I) was mutagenized and sixty-two candidate clones were screened for enzyme activity. Two of the clones expressed enzymes ( *Taq-3 and *Taq-5) that showed very reduced 5'-3' exonuclease activity and normal DNA polymerase activity. These two enzymes showed heat resistance and storage stability similar to Taq Pol I and had similar effectiveness in PCR. Processivity of the polymerases was compared by measuring the extension of an end-labeled primer annealed to a single stranded DNA, as well as by a PCR method. The processivity of *Taq-3 and *Taq-5 was similar to Taq Pol I (50-80 nucleotides) and more processive than a Taq Pol I deficient in the 5'-3' exonuclease due to absence of the first 290 amino acids (Stoffel fragment). The results indicate two amino acids which are required for normal 5'-3' exonuclease activity in Taq Pol I (Arg-25 and Arg-74). Keywords: DNA polymerase; Exonuclease; Processivity; (Thermus aquaticus)

1. Introduction Heat stable DNA polymerases are useful in the polymerase chain reaction (PCR) and DNA sequencing. Thermus aquaticus DNA polymerase (Taq Pol I) is heat-resistant [1], and the gene for the enzyme has been cloned and sequenced [2]. Most DNA polymerases have two regions of activity: polymerase and 3'-5' exonuclease. They may (as in Escherichia coli DNA polymerase I) have an additional 5'-3' exonuclease. Taq Pol I has two activities: polymerase and 5'-3' exonuclease [1,2,4]. Taq Pol I has neither measurable 3'-5' exonuclease activity [3,4] nor amino acid homology in the region of other DNA polymerases having a 3'-5' exonuclease [2,5]. In contrast, Taq Pol I has 5'-3' exonuclease activity [4] and the N-terminal region (residues 1-300) shows extensive amino acid similarity with the putative 5'-3' exonuclease region of E. coli Pol I [2,6] and eight other Pol I family members and related phage exonucleases [7]. The positions of these enzymatic activities are diagrammed in Fig. 1. To investigate the behavior of an enzyme lacking the 5'-3' exonuclease activity we mutagenized the region of the Taq DNA

* Corresponding author. Fax: + 1 (503) 4946886. 0167-4781/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved

SSDI0167-4781(95)00153-0

polymerase gene encoding that activity and chose enzymes with only polymerase activity. We purified and studied two enzymes, * Taq-3 and * Taq-5, expressed by these mutagenized constructs.

2. Materials and methods

2.1. Mutagenesis of the Taq DNA polymerase gene Primers based on the published sequence of the Taq Pol I gene were used to amplify the gene from genomic DNA (Thermus aquaticus). The primers were modified to include an EcoRI restriction site in the 5'-untranslated region (5-UTR) at position - 5 8 [2] and a BgllI restriction site at the 3' end at position 2449. The pUC 18 vector was digested with EcoRI and BamHI; the Taq Pol I gene was digested with EcoRI and BgllI and then cloned into pUCI8. This recombinant plasmid (pLSM5) encoded REM-T2 enzyme. Four methods were used to mutagenize pLSM5 (Fig. 2). Three different treatments with 10 mM sodium acetate (pH 4.8) were used: 20 min at room temperature (method 1), 5 rain at 50°C (method 2), or 5 min at 70°C (method 3). After each acid treatment the DNA was neutralized with

244

L.S. Merkens et al. / Biochimica et Biophysica Acta 1264 (1995) 243-248

Taq D N A polymerase gene 5'

Y 2499 bp

I

I

]~ Taq Pol I NH 2

COOH 5'-3' "-II~ exonclel

".ql--pei~ert~--lW.- 832 aa

Fig. 1. Diagram of Taq DNA polymerase gene and activity domains of the expressed Taq Pol I based on amino acid homologies with E. coli Pol I. The 5'-3' exonuclease activity is in the N-terminal region of the enzyme (amino acids 1-300) and the polymerase activity is in the C-terminal region (amino acids 374-832).

Tris-HC1 (pH 8.0) and the fragment corresponding to the 5' end of the Taq DNA polymerase was amplified by PCR containing unbalanced dNTPs (0.75 mM dATP, and 0.15

mM each of dCTP, dGTP, and dTTP). A primer 5' to the EcoRl site (pUC18), reverse primer: 5'-caggaaacagctatgacc-3', and a primer 3' to a unique KpnI site (484), 628A: 5'-cccaaagccaggcgg-3', were used for amplification of the DNA from the first two treatments (Fig. 2). For the third treatment the reverse primer and a primer 3' to a unique BstXI site (880), 1155A: 5'-caggtccctgagggc-3' were used for the amplification of a larger fragment. In method 4, the 5' end of the Taq DNA polymerase gene was amplified in a PCR (reverse primer and primer 628A were used) for three consecutive PCR programs each with 30 cycles. The amplified products from all four treatments and the untreated pLSM5 were digested with either EcoRI and KpnI or EcoRI and BstXI, ligated and transformed into E. coli DH5 oz cells. The methods produced replacement of 542 or 938 bases for the mutagenesis.

Mutagenesis o f Taq EcoRl

org

Kpnl I

BstXl

I

I

I

I

lkb

I

I--pLSM5 2 kb

method 1:20 min pH 4.8 22 ° C method 2: S min pH 4.8 50° C method 3:5 mln pH 4.8 70 ° C method 4:3 rounds of PCR, 30 cycles each EcoRI and Kpnl or EcoRl and BstXl

I

m

BstXl !

qNg h n

or

]

mutagenized

II -100

i I 1

A

=11 I 100

BI C 200

i.I. I 300

D

i IE 400

I 500

tl F I 600

II n I 700

| 800

I 900 bp

Fig. 2. Construction of mutagenized Taq DNA polymerase genes. Plasmids with the gene were mutated by each of four different methods. The EcoRI cloning site with either a unique KpnI site (at base pair 484) or a unique BstXI (at base pair 880) was used to replace the untreated sequence with mutagenized fragments. The locations of the mutations in * Taq-3 and * Taq-5, as well as the parent plasmid (pLSM5) are shown. Positions * 241, * 319, • 559 and * 583 show the base pair positions corresponding to four conserved amino acids that are necessary for 5'-3' exonuclease activity in E. coli Pol I [6]. Stippled-box areas indicate six conserved regions in the 5'-3' exonuclease in Pol I family members and related phage exonucleases [7].

245

L.S. Merkens et al. / Biochimica et Biophysica Acta 1264 (1995) 243-248 2.2. Enzyme preparations and assay

Crude lysates of transformed E. coli were prepared from 1.5 ml of overnight liquid cultures [8,9]. The cells were collected and washed in 0.25 ml of 50 mM Tris-HC1 (pH 8.0), 1 mM EDTA, 50 mM dextrose. The cells were resuspended in 0. I ml of the above solution with 4 m g / m l lysozyme and incubated for 15 min at room temperature. Then 0.1 ml of 10 mM Tris-HC1 (pH 8.0), 50 mM KC1 was added, the mixture was incubated for 30 min at 75°C, and centrifuged at 1 4 0 0 0 X g for 10 min at 4°C. The supernatant (0.15 ml) was mixed with 0.15 ml of a 50% slurry of DEAE cellulose in 0.4 M potassium phosphate (pH 6.8). The mixture was placed on ice for 30 min, centrifuged at 14000 X g for 15 min, and the supernatant was assayed for DNA polymerase and exonuclease activity. DNA polymerase activity was assayed at 75°C for 10 min in a 75 /~1 reaction containing 25 mM Tris-HCl (pH 8.8), 4 mM MgC12, 73 ng of activated calf thymus DNA, 33 /zM each dATP, dCTP, dGTP, dTTP, 1 /~Ci of [methyl-3H]dTTP (48 /zCi/nmol), and enzyme. As needed, the enzyme was diluted first with a solution of 25 mM Tris-HCl (pH 8.8) containing 0.5% each of Tween 20 and Nonidet P-40. Reactions were terminated by addition of ice-cold 10% saturation TCA and the insoluble precipitate collected on Whatman G F / C filters. One unit of activity equals 10 nmol nucleotides incorporated into acid insoluble form in 30 min incubation at 75°C [10]. For the 5'-3' exonuclease assay a 5'-end labeled double stranded DNA was prepared as a substrate using pBluescript or pGEX, after cutting with HinclI to make blunt end fragments. The 5' ends were labeled [11]. The 5'-3' exonuclease activity was assayed at 55°C for 60 min in a 50 ~1 reaction containing 25 mM Tris-HCl (pH 8.8), 4 mM MgC12, 0.4 fmol 5' end labeled DNA substrate//xl, and 0.3 units of polymerase activity. The reaction was stopped with 0.3 ml of 10% saturation TCA on ice. After centrifugation at 14000 X g for 15 min, 0.1 ml supernatant was dried on G F / C filter disks and 32p-radioactivity measured. Activity was calculated as the total fmol of 5' label released during 60 min at 55°C. For further purification, E. coli D H 5 a with recombinant plasmids was grown in LB media with 5 mg ampicillin/100 mi. After collection, the cell pellet was washed in 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 50 mM dextrose and lysed at room temperature in the same buffer containing 4 m g / m l lysozyme (the volume was 5 times the cell weight). An equal volume (5 times cell weight) of 10 mM Tris-HC1 (pH 8.0), 50 mM KCI was added and the lysate incubated for 60 min at 75°C. The cell debris and denatured proteins were removed by centrifugation at 4°C at 10000 X g for 30 rain. The supernatant, fraction I, was combined with an equal volume of a 50% slurry of DEAE cellulose in 0.4 M potassium phosphate (pH 7.5). After centrifugation the supernatant, fraction II, was diluted 5-

fold with 0.01 M potassium phosphate (pH 7.5), applied to a phosphocellulose column (volume of 5 times cell weight) and eluted with a linear gradient 0.02 M to 4.0 M potassium phosphate (pH 7.5) of ten times column volume. Peak enzyme activity eluted at about 0.1 M potassium phosphate (pH 7.5). The pooled peak fractions containing DNA polymerase activity, fraction III, were concentrated by Amicon filter (P30) in 2-fold storage buffer (20 mM Tris-HCl (pH 8.8), 100 mM KC1, 0.1 mM EDTA, 1 mM dithiothreitol), and then Tween 20, Nonidet P-40 (0.5% each) and glycerol (50%) were added, sufficient to make the buffer normal. Protein was measured with a Sigma Protein Assay Kit P 5656 and analyzed by SDS-polyacrylamide gel (8%) electrophoresis [11] in a vertical slab unit (Hoeffer Scientific SE 600 series). Proteins were stained with Coomassie blue or silver nitrate [11]. 2.3. Gel assay o f processiuity

Conditions to evaluate processivity on a polyacrylamide gel were based on those in Innis et al. [12]. The 5' end of the primer (20-mer: ttgtaaaacgacggccagtg) was 32p endlabeled. To measure processivity, the primer (3.4 pmol primer molecules/25 /zl) was annealed to M13mpl8 single-stranded D N A template (2.9 pmol template molecules/25 /xl). After annealing, the reaction mixture was adjusted to 200 /xM for each dNTP, 0.05% each Tween 20 and Nonidet P-40, 10 mM Tris-HC1 (pH 8.0), 50 mM KCI and 2.5 mM MgC12 in a total volume of 50 /zl. The molar ratio of template/primer to enzyme was: AmpliTaq, 54; Stoffel, 66; ~ Taq-3, 16 and * Taq-5-18. The reaction (55°C) was initiated by addition of 0.2 units/tzl enzyme (fraction III for * Taq-3 and *Taq-5). The reactions were stopped at various times and the DNA denatured by heating to 95°C for 5 min and loaded onto a 6% acrylamide gel containing 8 M urea [11]. 2.4. PCR assay o f processivity

Four sets of primers that had product sizes of 41, 59, 79 and 107 base pairs were used on single-stranded M 13mp 18 DNA. Primer number 1 annealed to the template and was the same in all reactions. The opposing primers were Table 1 Location of mutations in pLSM5, pTarf3 and pTarf5 Plasmid Nucleotide Codonchange Aminoacid Aminoacid position a position change pLSM5

pTaff3 pTart3 pTarf5

- 32 814 842 2415 73 384 221

a As in [2].

C to G TIT to ATI' CTC to CCC GAG to GAA CGC to TGC AAG to AAA CGC to CAC

272 281 805 25 128 74

5' UTR Phe to Ile Leu to Pro Glu (no change) Arg to Cys Lys (no change) Arg to His

246

L.S. Merkens et al. / Biochimica et Biophysica Acta 1264 (1995) 243-248

spaced to define the different product sizes and could only anneal to the product of primer 1. For each 25 /xl PCR reaction, template (120 fmol template molecules//xl) and primers (1 pmol primer molecules//zl) were used with a standard buffer: 67 mM Tris-HC1 (pH 8.8), 16 mM (NH4)2504, 10 mM 2-mercaptoethanol, 6.7/xM EDTA, 2 mM MgC12. and 150 /xM each dNTP and 0.01 or 0.03 units//zl enzyme (fraction III). The PCR was twelve cycles: 20 s at 94°C and 30 s at 48°C. The products were analyzed on a 12% non-denaturing polyacrylamide gel and stained with ethidium bromide.

AmpliTaq

Stoffel

*Taq-3

*Taq-5

- 102

,72

3. Results and discussion 3.1. Cloning the Taq DNA polymerase gene

26

We isolated the gene for Taq Pol I by using PCR and cloned the product into pUC18. Sequencing showed four single base differences compared to the published sequence (Table 1). These were: a change from C to G at position - 3 2 (non-coding), T to A at position 14 (Phe to lie), T to C at position 842 (Leu to Pro) and G to A at position 2415 (no change). Presumably these changes occurred in the PCR recovering the gene. These differences appeared to have no effect on function. The specific activity of purified enzyme was 70 000 units/mg protein with the same apparent molecular mass (94 kDa) as AmpliTaq (Perkin Elmer) or Taq DNA polymerase (Promega) analyzed by PAGE in SDS. DNA polymerase activity was

5'-3' Exonuclease Activity of Taq Pol I

~n

!i~i ~I

:

i~

H

T 11 Stoffel

I R E M T2 I *Taq-3

I *Taq-5

Fig. 3.5'-3' exonuclease activity. Exonuclease was measured for AmpliTaq, Stoffel, REM-T2, ~ Taq-3, and * Taq-5 as described in Section 2. 0.3 units of each DNA polymerase was used for each assay (n = 3). For • Taq-3, " Taq-5 and REM-T2, fraction 3 was used. Error bars represent the standard error of the mean. Activity was calculated as the total fmol of 5' label released during 60 min incubation at 55°C,

Fig. 4. Comparison of products synthesized from a 5'-end labeled 20-mer and single stranded DNA template by AmpliTaq, Stoffel, * Taq-3 and • Taq-5 (0.02 polymerase units//xl of reaction for all enzymes). Reactions were sampled at 0, 15, 30, 45, and 60 s (lane 1-5).

affected by Mg 2+ concentration; activity at 4 mM was higher than activity at 2 mM. Neither N-ethylmaleimide nor 2-mercaptoethanol had an effect on enzyme activity or amount of product formed under PCR conditions. 3.2. * Taq-3 and * Taq-5 enzymes lack exonuclease

Because the 5'-3' exonuclease activity is located in the amino terminal end of the enzyme, this region was targeted for mutagenesis (Fig. 2). In E. coli DNA polymerase I, removal of the amino terminal portion causes loss of 5'-3' exonuclease activity [13]. Of four mutations in E. coli DNA polymerase I which inactivate the 5' exonuclease [6], two of the conserved amino acids in Taq Pol I are 5' to the KpnI restriction site and all four are 5' to the BstXI restriction site (Fig. 2). Also five of the conserved regions [7] are 5' to the KpnI restriction site and all six are 5' to the BstXI restriction site (Fig. 2). After pLSM5 was mutagenized, the fragment 5' to either KpnI or BstXI sites was substituted for the corresponding unmutagenized fragment (Fig. 2). DNA polymerase activity and 5'-3' exonuclease activity were measured in cell lysates from 62 colonies with recombinant plasmids. Two of the clones (pTarf3 and pTarf5) expressed enzymes (*Taq-3 and * Taq-5) that showed very reduced 5'-3' exonuclease activity (Fig. 3) and normal DNA polymerase activity. The Stoffel fragment (Perkin Elmer), which begins at amino acid 290 and has no 5'-3' exonuclease activity [14], was also tested. There was no significant difference in the 5'-3' exonuclease activity of these three enzymes (indicated by the error bars). The plasmid DNA in pTarf3 had been mutagenized by method 4 and the DNA in pTarf5 had been mutagenized

L.S. Merkens et aL / Biochimica et Biophysica Acta 1264 (1995) 243-248

by method 3. Nucleotide sequence was determined and single amino acid changes compared to pLSM5 were identified at amino acid residue 25 and 74 (Table 1 and Fig. 2). The mutation in arginine-25 is located on the 3' edge of conserved region A [7]. Six of six bacterial DNA polymerases have an arginine at this position. The mutation in arginine 74 is within conserved region C. This arginine residue is invariant in the bacterial DNA polymerases and four bacteriophage 5'-3' exonucleases. Plasmid pTarf3 had an additional base change at position 128 which did not change the amino acid encoded. These data are consistent with functionally conserved regions and identify two additional amino acids necessary for 5'-3' exonuclease activity. Among the remaining clones, 15 showed l o w DNA polymerase activity (between 2 and 10% of REM-T2), and several others had reduced 5'-3' exonuclease activity (up to 25%, normalized on a polymerase basis). Each of the mutagenesis methods produced clones with altered enzyme activity (data not shown). One clone showed nearly normal 5'-3' exonuclease activity with no detectable DNA polymerase activity. After incubation at 88-90°C for 10 min, the remaining percent of activities of the purified enzymes was similar to AmpliTaq (Table 2). These enzymes function effectively under PCR conditions (including RT-PCR), and they synthesize products as long as 4.8 kb in our trials (data not shown). In storage buffer at - 2 0 ° C the enzymes lose only 20-30% activity over 14 months.

247

Table 2 Heat stability of DNA polymerases Polymerase

% Activity after 10 rain at 88-90°C a

AmpliTaq * Taq-3 * Taq-5

86.7+ 12.0 b 92.8 + 6.6 94.3 + 6.9

Incubation in PCR buffer, polymerase activity by standard assay. b % Activity relative to activity before incubation: mean +_S.E., n = 3.

3.3. Processivil3, Processivity of * Taq-3 and * Taq-5 was estimated by two methods. The first method was a gel analysis of the extension products of a 5'-end labeled primer. The second was a novel PCR method. Fig. 4 shows the results of the gel analysis of the extension products formed under conditions with excess primer/template concentration and limiting enzyme concentration. The length of the product formed (minus the primer) should approximate the processivity, because the probability of multiple initiations on a given template is minimized. * Taq-3 and * Taq-5 have an estimated processivity of 52 nucleotides, close to AmpliTaq (82 nucleotides) but much higher than Stoffel fragment (Perkin Elmer) at 6 nucleotides. These are similar to reported values: AmpliTaq, 50-60 nucleotides [15] and Stoffel fragment l 0-fold lower than AmpliTaq [ 14]. We developed another strategy to measure processivity

PCR Processivity Assay 3'

5t

....

~

-,4----- primer 1

rst synthesis round

primer I primer 2 ~

ssDNA template

or

- - ~

primer 1

- - ~

primer 1

- - ~

primer I

-

primer 2 - - - ~ ~. . . .

~ ~:---~l----

primer I

primer 2 - - - - ~ . . . .

~ ~-'---'~--

primer 1

primer 2 - - - - - ~ . . . .

~

primer I

~ ....

~

Fig. 5. PCR processivityassay.

....

~

primer I

248

L.S. Merkens et al. / Biochimica et Biophysica Acta 1264 (1995) 243-248

1234

1234

-,91107 -~m 79 -'91 59 -qm 41 *Taq.3 1234

*Taq-5 1234

-~ "~ "~ "q

107 79 59 41

product formed would serve as template for a subsequent round of synthesis. Synthesis of new templates would lead to amplification of the full length product with continued cycles. In the case where the processivity is shorter than the gap, the product will be of variable length and would not serve as a template for primer 2 for synthesis in succeeding cycles (right hand path, Fig. 5). Each denaturation cycle disassociates the complex and re-establishes conditions of excess primer for annealing. To visualize small amounts of these short products with ethidium bromide the reactions were run on 12% polyacrylamide gels. The results for AmpliTaq, * Taq-3 and *Taq-5 (Fig. 6) show that the processivity of these enzymes is at least 87 nucleotides (product minus the primer), whereas the Stof/el fragment (at 3-times the enzyme concentration) has a processivity of much less, in agreement with other methods. Results from both processivity methods indicate that the full length peptide of Taq Pol I is necessary for optimal processivity and that the 5'-3' exonuclease activity is not required for normal processivity.

Acknowledgements

AmpliTaq

Stoffei

Fig. 6. PCR products formed under conditions to determine processivity. For each lane the primer set used defined a different product size, lane 1: 41 bp, lane 2:59 bp, lane 3:79 bp and lane 4:107 bp. Primer sequences were: 1. 5'-ttgtaaaacgacggccagtg-3'g, 41. 5'-cctgcaggcatgcaagc-3', 59. gggatcctc tagagtcg-3', 79. 5'-gaattcgagctcggtac-3', 107. 5'-caggaaacagctatgacc-3'. Gap sizes were respectively: 21,39, 59, and 87. The concentration of * Taq-3, * Taq-5 (fraction HI) and AmpliTaq was 0.01 unit//xl reaction, and the concentration of Stoffel was 0.03 units//zl.

based on PCR (Fig. 5). A single stranded DNA, Ml3mpl8, was used for the template. Sets of primers were used on M13mpl8; each set defined a different gap size. Although primer 1 of each set (the same in all cases) could anneal to the vector, primer 2 could only anneal to an extended product (Fig. 5). If the newly synthesized DNA did not extend far enough to include the sequence to which the opposing primer annealed, then the subsequent cycle could not form a full length product. The probability of multiple initiations on a given template was minimized by using excess template (120 fmol//zl) and primer (I pmol//xl), limiting enzyme concentration (0.01-0.03 units//xl), and limiting the PCR to 12 cycles. Under these conditions, the number of products synthesized in each cycle is determined by the number of active enzyme molecules. Because short products require more rounds of synthesis to be visualized than do longer products, we first determined the lowest concentration of enzyme required to visualize the smallest product (gap size 21 or 24 and product size 42 base pairs) for a low number of cycles. This enzyme concentration was used with all sets of primers. If the processivity is greater than the number of nucleotides required (left hand path, Fig. 5) to fill the gap, then the

This work was supported by USPHS GM 19122 and a grant from the MRF.

References [1] Chien A., Edgar, D.B. and Trela, J.M. (1976) J. Bacteriol. 127, 1550-1557. [2] Lawyer, F.C., Stoffel, S., Saiki, R.K., Myambo, K., Drummond, R. and Gel/and, D.H. (1989) J. Biol. Chem. 264, 6427-6437. [3] Tindall, K.R. and Kunkel, T.A (1988) Biochemistry 27, 6008-6013. [4] Longley, M.J., Bennett, S.E. and Mosbaugh, D.W. (1990) Nucleic Acids Res. 18, 7317-7322. [5] Beruad, A., Blanco, L., Lazaro, J.M., Martin, G. and Salas, M. (1989) Cell 59, 219-228. [6] Joyce, C.M., Fujii, D.M. Laks, H.S., Hughes, C.M. and Grindley, N.D.F. (1985) J. Mol. Biol. 186, 283-293. [7] Gutman, P.D, Minton, K.W. (1993) Nucleic Acids Res. 21, 44064407. [8] Niwa, O., Bryan, S.K. and Moses, R.E. (1979) Proc. Natl. Acad. Sci. USA 76, 5572-5576. [9] Engelke, D.R., Krikos, A., Bruck, M.E. and Ginsburg, D. (1990) Anal. Biochem. 191,396-400. [10] Moses, R.E. and Richardson, C.C. (1970) Proc. Natl. Acad. Sci. USA 67, 674-681. [11] Sambrook, J., Maniatis, T. and Fritsch, E.F. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor. [12] Innis, M.A., Myambo, K.B., Gel/and, D.H. and Brow, M.A.D (1988) Proc. Natl. Acad. Sci, USA 85, 9436-9440. [13] Jacobsen, H., Klenow, H. and Overgaard-Hansen, K. (1974) Eur. J. Biochem. 45, 623-627. [14] Lawyer, F.C., Stoffel, S., Saiki, R.K., Chang, S-Y., Landre, P.A., Abramson, R.D. and Gel/and, D.H. (1993) PCR Methods and Applications Vol. 2, pp. 275-287. [15] Abramson, R.D., Stoffel, S. and Gel/and, D.H. (1990) FASEB J. 4, A2293.