Gene, 39 (1985) 305-310 Elsevier GENE 1448 A rapid procedure for DNA sequencing using transposon-promoted deletions (Sequencing vector; galactose resistance; rapid plasmid extraction and sequencing) Asad Ahmed 305 in Escherichia coli Department of Genetics. University of Alberta, Edmonton, Alberta, T6G 2E9 (Canada) Tel. (403) 432-3546 (Received August 3rd, 1985) (Accepted August 29th, 1985) SUMMARY A simple procedure has been developed for sequencing long fragments of DNA. The fragment (which can be several kb in length) is cloned in pAA3.7X, and subdivided into many overlapping segments by Tn9-promoted deletions. The deletions are isolated by positive selection for galactose resistance. A rapid plasmid preparation from several hundred galactose-resistant colonies is fractionated by agarose gel electro- phoresis to pick a series of deletions terminating at approx. 200-bp intervals across the entire length of the fragment. Selected plasmids are purified by rapid alkaline extraction, and used directly for supercoil sequencing with a primer derived from IS1. Sequences of adjacent deletions contain overlaps which are used to connect individual sequences to give the complete sequence. INTRODUCTION The procedure for nucleotide sequence analysis described below is based upon two recent develop- ments. First, a DNA fragment cloned on a plasmid can be subdivided into a series of overlapping seg- ments by transposon-promoted deletions (Ahmed, 1984a). These deletions arise as a result of the for- mation of intramolecular cointegrates (Shapiro, 1979) and exhibit the unique feature of extending Abbreviations: Ap, ampicillin; bp, base pair(s); CDTA, tram-1,2- diaminocyclohexane-FV,,N,N’,N’-tetraacetic acid; Cm, chloram- phenicol; d, deletion; EtdBr, ethidium bromide; Gal. galactose; IS, insertion sequence; kb, 1000 bp; LT, see EXPERIMENTAL, section a; nt, nucleotide(s); PA, polyacrylamide; PEB, see EXPERIMENTAL, section e; R , resistant; SDS, sodium dodecyl sulfate; ‘, sensitive; Tc, tetracycline; Tn, transposon. from a fixed site at the transposon terminus to variable sites on adjoining DNA. If a transposon with a terminus X is located next to DNA of the sequence ABCDE, these deletions would generate overlapping segments of the type X.BCDE, X.CDE, X.DE and X.E. Therefore, using a primer comple- mentary to X, it is possible to determine the nu- cleotide sequences of B, C, D and E, and assemble them into the complete sequence. Second, super- coiled plasmid DNA can be used directly for dideoxy sequencing (Chen and Seeburg, 1985). These authors have demonstrated that covalently closed circular DNA, completely denatured by alkali treat- ment and adjusted to renaturing conditions, provides a more efficient template for DNA synthesis than linearized DNA. A combination of these two methods has resulted in the development of the present procedure which seems to have many 0378-l 119/85/$03.30 0 1985 Elsevier Science Publishers
Transcript
Gene, 39 (1985) 305-310
Elsevier
GENE 1448
A rapid procedure for DNA sequencing using transposon-promoted deletions
(Sequencing vector; galactose resistance; rapid plasmid extraction and sequencing)
Asad Ahmed
305
in Escherichia coli
Department of Genetics. University of Alberta, Edmonton, Alberta, T6G 2E9 (Canada) Tel. (403) 432-3546
(Received August 3rd, 1985)
(Accepted August 29th, 1985)
SUMMARY
A simple procedure has been developed for sequencing long fragments of DNA. The fragment (which can be several kb in length) is cloned in pAA3.7X, and subdivided into many overlapping segments by Tn9-promoted deletions. The deletions are isolated by positive selection for galactose resistance. A rapid plasmid preparation from several hundred galactose-resistant colonies is fractionated by agarose gel electro- phoresis to pick a series of deletions terminating at approx. 200-bp intervals across the entire length of the fragment. Selected plasmids are purified by rapid alkaline extraction, and used directly for supercoil sequencing with a primer derived from IS1. Sequences of adjacent deletions contain overlaps which are used to connect individual sequences to give the complete sequence.
INTRODUCTION
The procedure for nucleotide sequence analysis described below is based upon two recent develop- ments. First, a DNA fragment cloned on a plasmid can be subdivided into a series of overlapping seg- ments by transposon-promoted deletions (Ahmed, 1984a). These deletions arise as a result of the for- mation of intramolecular cointegrates (Shapiro, 1979) and exhibit the unique feature of extending
Abbreviations: Ap, ampicillin; bp, base pair(s); CDTA, tram- 1,2- diaminocyclohexane-FV,,N,N’,N’-tetraacetic acid; Cm, chloram-
from a fixed site at the transposon terminus to variable sites on adjoining DNA. If a transposon with a terminus X is located next to DNA of the sequence ABCDE, these deletions would generate overlapping segments of the type X.BCDE, X.CDE, X.DE and X.E. Therefore, using a primer comple- mentary to X, it is possible to determine the nu- cleotide sequences of B, C, D and E, and assemble them into the complete sequence. Second, super- coiled plasmid DNA can be used directly for dideoxy sequencing (Chen and Seeburg, 1985). These authors have demonstrated that covalently closed circular DNA, completely denatured by alkali treat- ment and adjusted to renaturing conditions, provides a more efficient template for DNA synthesis than linearized DNA. A combination of these two methods has resulted in the development of the present procedure which seems to have many
advantages over the conventional sequencing DNA (Sanger et al, 1983).
Ml3 method of 1980 ; Messing,
In the original method (Ahmed, 1984a), the dele- tions were isolated by positive selection for galac- tose-resistance on pAA3.7. This plasmid contained: transposon Tn9, a part of the gal operon which conferred galactose-sensitivity, and the ret gene which provided several sites for cloning the DNA fragment. GalR mutations consisted almost exclu- sively of overlapping deletions on the plasmid which extended from Tn9 to various points in the cloned fragment. Restriction fragments bearing these dele- tions were subcloned in M13, and sequenced using a primer derived from the Tn9 terminus. Subsequent experiments (unpublished), however, indicated that sequencing could be done directly on plasmids bear- ing these deletions by treatment with PstI + 2 exo- nuclease (or BamHI + exonuclease III) to generate single-stranded templates. Although this approach was simpler, it still required CsCl/EtdBr-purified plasmids. It became apparent that the full potential of this method would be realized only if the deletion- bearing plasmids could be extracted by a rapid, small-scale method and sequenced directly. There- fore, several rapid methods of plasmid extraction (Holmes and Quigley, 1981; Maniatis et al., 1982; Marko et al., 1982; Birnboim, 1983; Guo and Wu, 1983) were examined, and one of these was modified for DNA sequencing by the supercoil method. Appropriate changes were also made in the host, the vector, and the method of selecting GalR deletions. The resulting procedure is described below.
EXPERIMENTAL
(a) Media
LT broth contains (per liter): tryptone, 12 g; NaCl, 10 g; yeast extract, 5 g; glucose, 1 g, and thymine, 50 mg. LT agar and LT soft agar contain 12 g and 6 g agar, respectively, added to 1 liter of LT broth. MacConkey-galactose contains 40 g Mac- Conkey agar base and 10 g galactose per liter of the medium. Antibiotic concentrations used (per liter) are: Ap, 100 mg; Cm, 12.5 mg; and Tc, 12.5 mg.
(b) Sequencing vector
The vector pAA3.7X (Fig. 1) was constructed from pAA3.7 by removal of the BglII-HpaI fragment carrying the il cos site and the PstI fragment carrying the right end of Tn9. This 9.6-kb plasmid contains
Escherichia coli galKT genes, tet and amp genes from pBR322, and ISI-L and cat genes from Tn9. It confers a Gals, TcR, ApR and CmR phenotype on E:coli hosts deleted for the gal operon. The DNA fragment to be sequenced (up to 10 kb in length) is cloned at a site (BarnHI, SphI, Sal1 or XmaIII) in the tet gene, and subdivided by selecting GalR deletions. If both strands are to be sequenced, deletions are also selected from the fragment cloned in the opposite orientation. As a result of the removal of the right end of Tn9, all GalR deletions on pAA3.7X start from one fixed site located at the right terminus of IS1 -L. This is an essential requirement for direct
Fig. 1. Structure of the sequencing vector pAA3.7X. The plas-
mid contains (clockwise): a 25-bp PstI-BgJII segment carrying
ARSI from yeast; a 2.92-kb HpaI-Hind111 segment carryinggalK
and Tgenes from E. coli (zig-zag line); a 4.25-kb segment carrying
ret and amp genes from pBR322; and a 2.46-kb segment carrying
ISZ-L, cat and a truncated ISJ-R from Tn9 (double line). The
plasmid confers a Gals, TcR, ApR and CmR phenotype on E. coli hosts deleted for thegal operon. GalR deletions (inner arcs) start
from the right end of IS J -L and terminate at variable sites in, or
beyond, the galgenes. A DNA fragment cloned in the tet gene can
thus be subdivided into many overlapping segments. Restriction
sites for the enzymes BaJI, BamHI,BgJII,EcoRI, HindIII, HpaI, PstI, PvuI, SalI, SphI, TfhlllI and XmaIII are shown. Paren-
theses indicate altered restriction sites. The map scale is in kb.
307
sequencing on plasmids and for choosing a unique
sequence for the primer.
(c) Hosts
The E. coli strain A4[ = W3 110 recA trpCsup ,, strA A4(gal-chlD-pgl-att A)] has been described (Ahmed,
1984b). A restriction-defective host AA102 [recA pro
thisupE ena2 hsdR, Al(gal-chlD-pgl-att ,I.)] was con-
structed by transferring the recA 56 mutation from a
srlC : : TnlO recA donor (Csonka and Clark, 1980) to
strain 294 Al (Ahmed, 1984b). AA102 has the ad-
vantage that it can be used for cloning eukaryotic
DNA fragments directly. Moreover, it yields cleaner
plasmid preparations for sequencing, perhaps
because it is deficient in endonuclease I.
(d) Selection and fractionation of GalR deletions
Stocks of AA102 carrying pAA3.lX (with or
without the cloned fragment) are maintained on
LT + Ap plates. These Gals strains must not be
exposed to galactose-containing media at any stage
except during the selection of GalR deletions.
Colonies of plasmid-containing cells are patched
on MacConkey-galactose + Ap plates for selecting
GalR mutations. Many colonies appear in each
patch after overnight incubation at 37” C. These
colonies include GalR mutants (mainly deletions and
cointegrates) as well as Gal- segregants which
contain no plasmids but grow simply due to the
release of /Hactamase into the medium. Therefore,
mixed colonies from each patch are streaked on fresh
MacConkey-galactose + Ap plates. Deletions can
then be recognized as large pale colonies which are
ApR, CmS, GalR and TcS. The cointegrates, on the
other hand, form small reddish colonies which are
ApR, CmR, GalR and TcS. Plasmids from several
independent GalR deletions are extracted by a rapid
method (Maniatis et al., 1982) digested with
Pstl + the restriction enzyme that was used to clone
the fragment, and run on an agarose gel. A series of
deletions having end-points spaced at approx.
200-bp intervals is selected for nucleotide se-
quencing.
The random approach described above occasion-
ally results in the selection of deletions containing
large (> 200-bp) gaps between adjacent deletions.
This is because Tn9-promoted deletions exhibit
slight end-point specificity, and tend to terminate at
certain sequences preferentially. A simple solution to
this problem is to isolate and fractionate a large
number (several hundred) GalR deletions so that
deletions terminating in hot as well as cold spots are
recovered with equal efficiency. About 500 colonies
of plasmid-containing cells are patched individually
on MacConkey-galactose + Ap plates (50 patches/
plate) and incubated at 37°C. The GalR colonies,
which arise after l-2 days, are inoculated en masse
into ten cultures (3 ml each) of LT broth + Ap, and
grown for rapid plasmid extraction (Maniatis et al.,
1982). The mixed GalR plasmid preparations are
fractionated on a 1% low-melting-point agarose gel
(for example, a 13-slot 13 cm x 11.5 cm x 0.3 cm gel
run at 3 5 V for 20 h at 4 o C). The parent plasmid and
GalRA41 (a deletion which extends up to nt 642 of
pBR322) can be applied as markers in a slot near
the margin to delineate the region of interest on the
gel. DNA is extracted from consecutive 2-mm wide
gel slices (Maniatis et al., 1982) from the marked
region and used to transform strain AA102 on
LT + Ap plates. Recovery of transformants is great-
ly improved by plating transformed cells in LT soft
agar. ApR transformants arising from consecutive
fractions contain plasmids harboring deletions of
progressively increasing sizes. Plasmids are extract-
ed from a few colonies purified from each plate, and
their sizes are determined by restriction analysis. In
this manner, deletions spaced at approx. 200-bp
intervals across the entire length of the cloned frag-
ment are selected. It should be emphasized that
different ApR colonies, even on the same transfor-
mation plate, contain small variations in plasmid
size. Therefore, it is advisable to save the transfor-
mation plates at 4°C until a complete set of uni-
formly spaced deletions has been assembled. Using
this approach it is possible to isolate deletions of any
size, terminating in virtually any region of the cloned
fragment.
(e) Rapid plasmid preparation for nucleotide sequencing
A rapid method of plasmid isolation (Mark0 et al.,
1982) which involves two alkaline extractions, gives
excellent results. (A single alkaline extraction is
usually not sufficient to remove all traces of chro-
mosomal DNA.) A single colony of AA102 harbor-
308
ing the deletion-containing plasmid is inoculated into
6.5 ml of LT broth containing 100 pg Ap/ml. The
culture is grown with vigorous agitation for 20 h at
37 o C. Cells are collected in a 1.5~ml eppendorf tube
by four successive 20-s centrifugations in an eppen-
dorf model 5414 microcentrifuge. 100 ~1 of a fresh
1 mg/ml lysozyme solution prepared in PEB I
(50 mM glucose, 10 mM CDTA, and 25 mM
Tris * HCl, pH 8.0) is added to the cell pellet and
mixed by vigorous vortexing. The suspension is kept
at 0°C for 7 min. 200 ~1 PEB II (0.2 M NaOH,
1% SDS) is added, mixed gently, and the tube is
kept at 0°C for 10 min. 150 ~1 of PEB III (3 M
sodium acetate, pH 6.0) is added, and mixed by
vortexing. After 15 min at 0” C, the tube is cen-
trifuged for 5 min at 4°C. The supernatant (ap-
prox. 450 ,ul) is transferred to another tube, and
900 ~1 of cold ethanol is added. The tube is kept at
-70°C for 10 min, and then centrifuged for 10 min
at 4°C. The supernatant is discarded, and the pellet
is dried under vacuum. The pellet is dissolved in
20 ~1 PEB I, and then 40 ~1 of PEB II is added.
After keeping for 5 min at room temperature,
30 ilPEB III is added and the tube is allowed to
stand for 5 more min. The tube is centrifuged for
4 min, and the supernatant (90 ~1) is transferred to
another tube. 180 ~1 of cold ethanol is added and
mixed. The tube is kept at -70°C for 10 min, and
centrifuged for 10 min at 4°C. The supernatant is
discarded, and the pellet is dried. The pellet is dis-
solved in 54 ~1 of Tris-CDTA (50 mM Tris + HCl,
1.5 mM CDTA, pH 8.0). 27 ~1 of a LiCl solution
(15 M LiCl, 50 mM Tris + HCl, pH 8.0) is added and
the tube is kept at 0” C for 7 min. It is then cen-
trifuged for 5 min at 4’ C and, using a micropipette
with a siliconized tip, the supernatant (80 ~1) is
transferred to another tube. 160 ~1 cold ethanol is
added, and the tube is kept at -70” C for 10 min.
After a 1 0-min centrifugation at 4’ C, the supernatant
is poured off gently and the DNA pellet is dried. The
DNA is dissolved in 100 ~1 of 0.5 M NaCl in TE
buffer (10 mM Tris . HCl, 1 mM EDTA, pH 7.2)
and purified further by NACS chromatography. A
NACS PREPAC cartridge (Bethesda Research
Laboratories, Gaithersburg, MD, catalog
No. 1526 NP) is hydrated with 2 M NaCl in TE, and
equilibrated with 0.5 M NaCl in TE. The DNA
sample is loaded on the column and washed with
4 ml of the same buffer. Plasmid DNA is then eluted
three times with 100 ~1 of 0.7 M NaCl in TE. Eluted
DNA is precipitated with 2 ~01s. of ethanol, washed
with 80% ethanol, and dried. The yield is approx.
2-3 pg of plasmid DNA.
The boiling method (Holmes and Quigley, 1981)
has also been modified for rapid plasmid preparation
for DNA sequencing. It has the advantage of being
faster and the yield of plasmid DNA is equally good.
This modified procedure is included in the detailed
protocol (see section g, below).
(f ) Nucleotide sequencing
The supercoil sequencing method of Chen and
Seeburg (1985) is used for direct sequencing of plas-
mid DNA. The primer used is a 16-nt oligodeoxy-
ribonucleotide of the sequence:
G-C-C-A-C-T-G-G-A-G-C-A-C-C-T-C
which corresponds to residues 60 to 45 of ISI
(Ohtsubo and Ohtsubo, 1978). Hence, the primer
terminus is located 44 nt before the junction created
by the deletion. This -44 ISI primer is available from
New England Biolabs, Beverly, MA (catalog No.
1216). The -4 ISI primer described previously for
sequencing on M 13 (Ahmed, 1984a) cannot be used
for sequencing on plasmids as it is complementary to
both terminal inverted repeats of ISI.
(g) Results and conclusions
This method has been used to determine the
nucleotide sequences of several GalR deletions ter-
minating at different sites in the gal and tet regions
of pAA3.7X. In each case, the deletion fused the ISI
terminal sequence
T-T-G-G-C-A-G-C-A-T-C-A-C-C
to a new sequence derived from the gal or the tet
region. A sequence gel obtained by this procedure is
shown in Fig. 2. The following conclusions were
drawn from these studies.
(1) Sequences determined by the Ml3 method
and the present method were identical. The length of
the sequence that could be read from each gel in a
single run varied from 175 to 200 nt.
309
A C G T (2) Each GalR deletion fused the right terminus of ISI-L to other sequences without causing any detectable aberration at the junction. An example of this fusion by deletion is shown in Fig. 3(a). GalR dll5 fuses the ISI terminus (boxed sequence) to nt 97 of pBR322 (Sutcliffe, 1979), and the se- quence reads normally beyond this point. This result shows that Tn9-promoted deletions can be used confidently for subdividing DNA fragments for se- quence analysis.
(3) Perfect overlaps were found between the se- quences of adjacent deletions. Some examples of such overlaps are presented in Fig. 3(b-d). This result shows that long DNA sequences can be read continuously from one end by a systematic use of uniformly spaced deletions.
The present method of sequencing on plasmids bearing specific deletions offers many advantages over the random, shotgun approaches (reviewed by Deininger, 1983). It requires significantly less cloning and sequencing work, does not depend on the pres- ence of specific restriction sites within the fragment, avoids complex in vitro manipulations, allows pro- gressive determination of long DNA sequences, and generates sequence data in an ordered fashion. Per- haps the most important feature of the method is that it is fast, simple and economical.
A detailed protocol for DNA sequencing using this method is available from the author.
ACKNOWLEDGEMENTS
I thank Doreen Harcus for technical assistance and David Landry for supplying the primer. This
work was supported by a grant from the Natural Sciences and Engineering Research Council of Canada.
Fig. 2. A sequence gel from GalRd6.9 obtained by the supercoil
sequencing method (Chen and Seeburg, 1985). The plasmid
DNA preparation (from section e) was denatured with alkali,
neutralized, quickly annealed to the -44 IS1 primer, and
sequenced by the dideoxy method. After the reaction, samples
were run in an 8 y0 PA gel (0.4 mm x 33 cm x 40 cm) containing
8 M urea for 2.2 h at 60 W. Arrow indicates the deletion which
fused the IS1 -L terminus (below) to a new sequence from thegul
