A versatile method for integration of genes and gene fusions into the λ attachment site of Escherichia coli

Gene, 107 (1991) 11-17 0

1991 Elsevier

GENE

Science

Publishers

B.V. All rights reserved.

11

0378-l 119/91/$03.50

06087

A versatile method for integration of genes and gene fusions into the A attachment site of Esc~erichia coli (Phage ;1; recombination;

integration

in attB; 1acZ fusions;

Southern

blot)

Tove Atlung, Anne Nielsen, Lene Juel Rasmussen, Lars J. Nellemann and Flemming Holm Department of Microbiology. The Technical University of Denmark, DK-2800 Lyngby (Denmark) Received by A.M. Campbell: 10 December Revised/Accepted: 17 May/27 July 1991 Received at publishers: 9 August 1991

1990

SUMMARY

We have developed a versatile method for integration of modified genes and gene fusions into the bacteriophage A attachment site (attB) of the Escherichia coli chromosome. The method relies on two components: (I) a DNA integration cassette, flanked by multiple restriction enzyme sites, which contains the 3, attP site and, as a selectable marker, the Tn5 uphA gene conferring kanamycin resistance (KmR); and (2) a plasmid with the I int gene transcribed from the tet promoter. A fragment carrying the gene in question is ligated to the integration cassette, resulting in a circular piece of DNA unable to replicate. The ligation product is then transformed into a strain that contains the int-carrying plasmid. Selection for KmR results in colonies with the cassette integrated into the attB site of the E. coli chromosome. This method was used for integration of several 1acZ and phoA promoter fusions. The integration products were analyzed by Southern hybridization. In addition, we found, fortuitously, that the ligated DNA circles could also integrate by homologous recombination, although usually at a much lower frequency than the Int-mediated integration into attB.

INTRODUCTION

Regulation of many E. coli genes is studied using operon and gene fusions constructed on plasmids. However, it is essential to rule out effects from changes in the plasmid Cowespondence to: Dr. T. Atlung, Department The Technical

University

Tel. (45)45933422; Abbreviations:

transcriptional

ApR-encoding

kb, kilobase

tiple cloning site;p, promoter;pf, pgk, gene

encoding

alkaline phosphatase; transposon;

Bldg. 22 1,

Lyngby (Denmark)

uphA, KmR-encoding activator

gene; bp, base pair(s);

protein;

of Microbiology,

DK-2800

Fax (45)45932809.

Ap, ampicillin;

phosphatase gration

of Denmark,

gene;

gene; uppY, acid site; blu,

art, attachment

int, gene encoding

phage

or 1000 bp; Km, kanamycin; gene encoding

phosphoglycerate R, resistance/resistant;

kinase;

pyruvate

i, inte-

MCS, mulformate

phoA, gene

lyase;

encoding

‘, sensitive/sensitivity;

Tn,

XGal, 5-bromo-4-chloro-3-indolyl-8-D-galactopyranoside;

gene truncated plasmid-carrier

at the indicated state.

side (5’ or 3’ proximal);

‘,

[ 1, denotes

copy number on the level of expression of the indicator gene when measured under different growth conditions or in different regulatory mutants. Several systems that employ in vivo or in vitro recombination onto derivatives of bacteriophage ,l have been developed to achieve integration into the chromosome (e.g., McKenney et al., 1981; Linn and Ralling, 1985). These methods, however, require the use of specialized plasmid cloning vectors. Lately we have been studying growth-phase-regulated promoters using a promoter-probe plasmid derived from pCB267 (Schneider and Beck, 1986) for which we wanted to develop a method for the incorporation of gene fusions into the chromosome. We chose to use the I att, since the method should be suitable also for fusions to essential genes. We have constructed a cassette consisting of a selectable marker and the bacteriophage I attP region. This cassette can be excised from a plasmid with a number of restriction

12 enzymes

and ligated to a fusion or a gene excised by en-

RESULTS

zymes with compatible ends. In addition, we constructed a plasmid with the phage 1 inf gene to provide the Int protein in the host cell for integration of the fusion ligated to the attP cassette. The integrated fusion or gene can easily be transferred to new strains by bacteriophage Pl-mediated transduction.

AND DISCUSSION

(a) Construction of the at@ cassette We have constructed two plasmids, pTAC3463 and pTAC3466 (Fig. l), both carrying a cassette consisting of the selectable marker aphi conferring Km resistance linked to a 493-bp attP fragment containing the entire region of

pTAC3575 <

phoA

Sm/T

Bst

pTAC3599

_)

>

lad?

PM

ori

bla-

X

T/Sm

>

lad

w 0

-xF-

0

pTAC3463

pTAC3466

b-

+

aphA

H3N I 27

ori

bla

B I I 28

I1

Lambda

+-

phoA ’

attP

L&3

Xm

I I 29

kb

f

attP V/xm

Int

N/V

pTAC3422 Fig. I. Construction was constructed insertion

by insertion fragment.

of the attP-aphA cassette

by digestion

of pCB267

‘tet

and the inr plasmids.

(Schneider

of one IO-bp XhoI linker (New England

appYp (arrowhead),

Biolabs,

-

At the top is shown the promoter

Beverly, MA). Plasmid

fragment

The 493-bp attPBamHI-Hind111

ofTnphoA fragment

into pUCl8

is described

and pUCl9

cloning

vehicle pTAC3575.

tilling out with E. coli DNA polymerase

pTAC3599

carries

for the uppY gene (Athmg et al., 1989) cloned into the SmaI site ofpTAC3575.

of the EcoRI-SmaI

1 kb

on’

cat

and Beck, 1986) with EcoOlO9,

(Yanisch-Perron

a 740-bp TuqI fragment

Plasmids

pTAC3463

et al., 1985), respectively,

Plasmid

I Klenow containing

and pTAC3466

in section b. Key: dark shaded

between

the XhoI and Hind111 sites in the TnphoA-derived

bar, DNA from thephoA region of the chromosome;

segment.

open bar, synthetic

pTAC3575

fragment,

and

the promoter,

were constructed

followed by insertion

was purified from a double digest of Iz DNA and first cloned into BglII + HindIII-cleaved

from which it was excised with Sal1 + Hind111 and inserted pTAC3422

I

I w

D-------j_ int tetp

of the arrP pTAC3575,

The construction DNA; blackened

of bar,

bacteriophage I DNA; horizontally striped bar, 1acZ DNA; vertically striped bar, pBR322-derived DNA; crosshatched bar, TnS-derived uphA DNA; hatched bar, pACYCl8Cderived DNA; light-shaded bar, uppYp DNA; black arrowheads indicate promoters. A, AccI; B, BamHI; Bg, BglII; Bst, BstEII; E, EcoRI; HZ, HindII; H3, HindIII; N, NdeI; P, PstI; S, SalI; MCS are given in parentheses. When two different restriction

T, TaqI; V, EcoRV; X, XhoI; Xb, XbaI; Xm, XmnI. Alternate sites were ligated in the plasmid construction, their symbols

unique restriction sites in are separated by a slash.

13 bacteriophage A needed in cis for integration (Weisberg and Landy, 1983). These cassette plasmids were designed to integrate fusions from the promoter probe vector

(c) Integration of promoter fusions

pTAC3575 (Fig. 1) and the dnaA gene, respectively. The entire phoA-MCS-lad region of pTAC3575 can only be separated from the pBR322 origin and the bla gene using the XhoI and BstEII sites flanking it (Fig. 1). To reconstitute

The purified BstEII-Sal1 fragment carrying the attP-aphA cassette from pTAC3463 was ligated to the purified BstEIIXhoI fragment ofpTAC3599 carrying thephoA gene and the appYp-1acZ fusion. The ligation mixture of DNA circles

the phoA region in the integrants we used the aphA gene from TnphoA (Manoil and Beckwith, 1985). To increase the number of restriction sites available for excision of the cassettes, they have been recloned into the multi-cloning site of pIBI25 (plasmids not shown).

unable to replicate was used to transform which carries the int plasmid, pTAC3422,

An appYp-1acZ promoter-fusion plasmid, pTAC3599 (Fig. l), was chosen for the systematic test of the system.

strain MT102 (Table I). The frequency of KmR transformants in strain TC3422 was lOO-fold higher than that in MT102, suggesting that they arose primarily by Intmediated integration into attB. More than 95% of the TC3422 transformants were ApS and had the low level of fi-galactosidase activity expected for a single copy of the

(b) Construction of the int plasmid The I Int protein, encoded by the int gene, is the only phage gene product required for site-specific recombination between attP and attB sites, whereas excision of a phage 2 by recombination between attL and attR requires both the Int and Xis proteins (Echols and Guameros, 1983; Weisberg and Landy, 1983). The xis gene, encoding the Xis protein, is located between bp 29078 and 28863, and the int gene between bp 28 882 and 27 8 15 (Fig. 1). We therefore isolated the XmnI-NdeI (bp 29015-27 630) fragment from phage A. This fragment contains the intact int gene and a partially deleted ‘xis gene and attP sequence; although it still contains the att core, it is unable to promote integration into attB (Weisberg and Landy, 1983). This fragment was cloned into the EcoRV site in the tet gene of pACYC184 giving rise to pTAC3422. In this plasmid the int gene is located downstream from the tet gene promoter and thus should be expressed efficiently; the tet’ and ‘xis genes are fused out of frame and the tet gene translation stops 89 bp before the ATG codon of the int gene.

TABLE

strain TC3422, and its parental

appYp-/acZ fusion. The majority of the transformants obtained in strain MT102 also had this phenotype, suggesting that they, too, had arisen by integration of the correctly ligated circle into the chromosome, for example by RecA-mediated recombination into the chromosomalphoA gene. Four and eight light blue ApS integrants from TC3422 and MT102, respectively, were chosen for further analysis. (d) Analysis of the attB integrants by phage Pl transduction P 1 lysates were grown on the four appYp-IacZ integrants in TC3422 and used to transduce the Alac Gal + strain LJ24 (a derivative of C600; Rasmussen et al., 1991) to KmR phenotype. The gal genes are located 0.4 min (17 kb) from attB on the standard genetic map of E. coli (Bachmann, 1990) and the donor strains were galK_. The cotransduction frequency of KmR and Lac + was 100% in all cases. The cotransduction frequency of KmR and Galwas

I

Transformation Strain a

test of the integration DNA a

method

using an appYp_lacZ Number

MT102

LacZ phenotype

of

transformants/ng

TC3422

fusion

pTAC3418

1.2 x IO5

Lig. mix

4x

pTAC3418 Lig. mix

1.5 x lo*

(%)”

ApR (%)”

DNA b

lo3

+

+++

>96

<2

<2

<2

1 x 106 16

12

12

17

a TC3422 is MT102[pTAC3422]. MT102 is a h.sdR K-12 derivative of MC1000 (Casadaban and Cohen, 1980). b The purified BstEII-XhoI phoA-appYp-IacZ fragment from pTAC3599 was ligated to the atrP-aphA fragment purified from a BsrEII + PvuI digest of pTAC3463 the same ligation

mixture.

(the latter two enzymes Transformants

were used to reduce

were selected

on LB medium

the size of the the vector fragment). (Miller,

The two strains

+ Sal1 + EcoRI

were transformed

1972)/50 fig Km per ml/40 pg XGal per ml.

c The LacZ phenotype was determined from the coloring of the colonies on XGal plates: - , white; + , light blue; + + + , dark blue. d Determined by replica plating onto LB + 100 pg Ap/ml. All dark blue colonies from the ligation mixture were ApR.

with

14 5-12%,

which

is within

the expected

range

taking

into

(e) Test of the integrants by Southern-blot analyses Chromosomal DNA was extracted, either from primary

account that the donors carried a 0.2 min (8.2 kb) insertion in attB. We found that the int plasmid, pTAC3422, like other pACYC 184-derived plasmids (S. Brown, personal communication), was cotransduced in approx. 10% of the transductants.

Thus

the Pl

transductions

gested that the desired integration

strongly

integrants or from integrations transferred to strain LJ24 by Pl transduction, and used for Southern blots. Fragments containing the aphA gene or thephoA gene were isolated and used as probes in Southern blots with KpnI + BamHI

sug-

event had taken place.

double digests of the DNA.

A V

V

K

lttl.

Bst

B

f--b-+ phoA

appb’p

t K

v

v

V

B

B

vv

aphA uttR

Iad

Int

VK

B

lttB

1 kb

K

B

B

V

V

B

Red K

B

K

V

-4

phoA phoAp

Fig. 2. Integration of the uppYp-1acZ fusion into the afrB site (A) or the phoA gene (B). At the bottom of each map are shown the ligated circles of the cassette and the phoA-uppYp-IucZ fragments aligned with the uttB site or the phoA gene in the chromosome. Restriction map of the urrB region (A) of the E. coli chromosome

was deduced

from the maps of Redfield

and Campbell

(1987) and Kohara

and Ross, 1977). Restriction map ofthephoA region (B) is based on the maps of Kohara maps produced by the correct integrations of the circle. Keys as in Fig. 1.

et al. (1987), and from the sequence

et al. (1987). The upper portions

show the structure

of aMI (Landy and restriction

15 The at@ site is located

on a 1.2-kb KpnI-BamHI

frag-

ment on the E. coli chromosome (Fig. 2A). If one copy of the promoter fusion has been correctly integrated into attB the aphA probe should hybridize exclusively to a 2.2-kb BamHI fragment spanning attR (Fig. 2A). ThephoA probe should hybridize to a 3.5kb KpnI-BamHI fragment spanning attL. According to the restriction enzyme map of the E. coli chromosome (Kohara et al., 1987) the phoA probe should in addition hybridize to a 15.5kb BamHIKpnI fragment carrying the normal chromosomal phoA gene. The uphA and phoA probes both hybridized to fragments of the expected size in DNA prepared from the four appYp-lacZ attB integrants which had been Pl transduced into strain LJ24 (Fig. 3A and B, lanes 3). In Southern blots using DNA from primary

integrants

plasmid, the uphA-hybridizing

still containing

the int

fragment was always a single

r

1234

1234

7.8 kb

sm

*

5.7 kb

5.7 kb+

1.2 kb-j

Fig. 3. Southern-blot

analysis

1ucZ fusion. Chromosomal gel electrophoresis

(0.7 % agarose

as described

the

nonradioactive

Boehringer-Mannheim, as described

DNA

by the suppliers.

(a BarnHI-BsrEII

DNA were separated

fragment

in attB; 4, uppYp-lad

and

from pTAC3575). fragment

(DuPont). Hybridienzyme fragments detection

and detection

(Panel A) Hybridization

the aphA probe (BarnHI-Hind111 A and B): 1, M, markers;

labelling

and hybridization

(Atlung

(Sigma) in TBE buffer], and

the DNA was transferred to GeneScreen membranes zation probes were prepared from purified restriction using

uppYp-

of the integrated

was prepared

ofKpn1 + BumHI-digested

et al., 1989). Fragments by agarose

of the structure DNA

kit

was carried

from out

to thephoA probe

(Panel B) Hybridization from pTAC3463).

2, strain LJ24; 3, uppYp-IucZ fusion integrated fusion integrated

in phoA.

to

Lanes (for

band,

whereas

attL appeared

the phoA-hybridizing as a doublet

band

fragment

spanning

(data not shown).

This

band was a doublet with all the different promoters tested (see below) and also when the DNA was digested with EcoRV + BamHI instead of KpnI + BamHI. The extra band could arise from the presence in a small part of the population of a copy of the ligated attP circle integrated into the int plasmid, which contains an att core sequence. Upon integration into pTAC3422, the phoA-hybridizing band would be 3.7 kb, and the aphA-hybridizing band 1.9 kb. Since we have never observed a doublet band hybridizing to the aphA probe we think that this is an unlikely explanation. Alternatively, the extra phoA-hybridizing band might be due to a double strand break present near attL in the DNA preparations from the Int protein over-producing strains. The two probes hybridized to fragments of completely different sizes in the DNA from the eight uppYp-ZacZ integrants into MT102, indicating that the integration was not in attB (Fig. 3A and B, lanes 4). The sizes of the fragments hybridizing to the aphA (15 kb) and phoA probes (15 and 5.1 kb) were the same in all eight isolates and are those expected to arise from a single crossover event integrating the ligated circle of the two purified fragments into thephoA gene (Fig. 2B). To demonstrate the correct ligation at the XhoI and Sal1 sites, EcoRV + BamHI double digests were probed with a large fragment from the 1acZ gene. The expected fragment of 2.6 kb spanning the junction between ZacZ and the pUC multiple cloning site was seen in all integrants (data not shown). (f) Integration of other promoter fusions Using the procedure described above we have also integrated six differentpfl’-laci! fusions and we have integrated five different pgk’ -phoA fusions using a phoA 8 derivative of strain C600 carrying pTAC3422. All transformants analysed for thepgk’-phoA fusions had integrated properly into attB. Two of 17 pjl’-facZ integrants analysed, however, gave hybridization patterns strongly suggesting th& the primary integration into attB had been followed by a RecA mediated recombination between the two phoA genes leading to inversion of the lo-min region of the chromosome between phoA and attB. All the simple integrants into attB could be separated from pTAC3422 by phage P l-mediated transduction. Six integrants of a pj?‘-1acZ fusion into MT102 were also analysed. Three of these were integrated in phoA and three in pfl by homologous recombination (the pfl’ fusion carried a 2-kb promoter fragment). (g) Integration of an in vitro manipulated dnaA gene We have also used the system for integration of an in vitro created fusion between the E. coli and the S. typhimurium

16 dnaA genes. In this case the attP-aphA cassette was excised from pTAC3466 with Eco0109 + BamHI and ligated to an EcoO109-BarnHI fragment containing the dnaA promoter region and intact structural gene. Upon transformation into

adapted for use with any gene or gene fusion from any plasmid, simply by recloning the attP-aphA cassette into another suitable MCS, from which it can then be excised

a dnaA(ts) strain carrying pTAC3422, we found that all KmRApS transformants had become thermoresistant, indicating that they had received a functional dnaA gene. When integration was carried out in a dnaA(ts) strain containing a transducing phage ARBl, carrying a dnaA’-

with the desired restriction enzymes. Another advantage over the systems based on recombination onto a complete phage 1, is that our system has a high probability for singlecopy integrations, and if double integration occurs one copy will be lost by subsequent Int-mediated recombination between the attP and attR or attL sites.

ZacZ fusion, in attB (TC1926; Andrup et al., 1988) we found that the ilRB1 phage was lost very rapidly upon continued selection for KmR. This suggests that the Int-

We are presently recloning the int gene into the tet gene of a rep(ts) plasmid to allow rapid curing of the plasmid after transformation of the ligated attP cassette circles.

mediated recombination between the attP site, recreated by integration of the cassette, and the attL or attR site (Weisberg and Landy, 1983), is too efficient to allow maintenance of the double ‘lysogen’. A few stable LacZ + , KmR thermoresistant clones were obtained after repeated reisolation. These strains still carried the int plasmid, pTAC3422, and Southern-blot analysis (data not shown) showed that they had arisen from DNA rearrangements.

When we tested the system we found that it was possible to get integrants by RecA-mediated homologous recombination of the ligated circles with a frequency that will allow purposeful exploitation of this process to obtain insertions, for example, of lacZ fusions situated at the normal chromosome location analogously to those widely used in Bacillus subtilis studies.

(II) Conclusions We have developed a versatile method for integration of in vitro genetically manipulated genes into the E. coli chromosome in the bacteriophage I attachment site, which relies on introduction of a ligated circle containing attP into a cell expressing the Int protein, as previously described by Koob and Szybalski (1990). We have successfully integrated several promoter fusions to IacZ or phoA and an in vitro manipulated dnaA gene. The integrated fusions can easily be transferred to other strains by Pl-mediated transduction. A priori it could be expected that the ligated attP circle would integrate also into the int plasmid since this contains an intact att core sequence from attP. Such an attP-derived att core (AattP) was found to be very effective in the attP x AattP Int-mediated reaction (Podhajska et al., 1985). Our data, however, indicate that the end product is nearly always integration into the chromosomal attB. We have presently used the method for integration of more than 25 different promoter fusions and analysed in all close to 200 integrants and never found one where the fusion was not in the chromosome. The evidence is: (I) the integrants always show a uniform light blue coloring of the colonies indicative of a stable single copy lacZ or phoA fusion; (2) the integrated fusion could in all cases easily be separated from the int plasmid by Pl-mediated transduction, and (3) the bands found in the hybridizations are different in size from those that would arise from integration into the int plasmid. It is possible that some elements remaining in our partially deleted attP sequence interfered with its function as attB. Our system has the great advantage that it can be easily

ACKNOWLEDGEMENTS

We are grateful to Bjarne Albrechtsen for critical reading of the manuscript. This work was supported by grants from the Danish Technical Science Research Council, the Christian Hansen Laboratories and from the Danish Center of Microbiology.

REFERENCES Andrup, L., Athmg, T., Ogasawara, N.,Yoshikawa, H. and Hansen, F.G.: Interaction of the BaciNus subtilis DnaA-like protein with the Escherichiu coli DnaA

protein.

J. Bacterial.

Atlung, T., Nielsen, A. and Hansen,

170 (1988) 1333-1338.

F.G.: Isolation,

of appY, a regulatory

nucleotide

sequence

dependent

gene expression

characterization

and

gene for growth-phase-

in Escherichiu coli. J. Bacterial.

I71 (1989)

1683-1691. Bachmann,

map of Escherichia coli K-12, Edition

B.J.: Linkage

crobiol. Casadaban,

8. Mi-

Rev. 54 (1990) 130-197. M.J. and Cohen,

S.N.: Analysis

of gene control

signals

by

DNA fusion and cloning in Escherichiu cofi. J. Mol. Biol. 138 (1980) 179-207. Chang,

A.C.Y. and Cohen,

amplifiable cryptic Echols,

multicopy

miniplasmid.

S.N.: Construction

and characterization

DNA cloning vehicles derived J. Bacterial.

H. and Guameros,

134 (1978) 1141-I 156.

G.: Control

Hendrix,

R.W., Roberts,

Lambda

II. Cold Spring

of integration

and excision.

J.W., Stahl, F.W. and Weisberg, Harbor

of

from the Pl5A

Laboratory,

In:

R.A. (Eds.),

Cold Spring

Harbor,

NY, 1983, pp. 75-92. Kohara,

Y., Akiyama,

E. coli chromosome: and sorting

K. and Isono, K.: The physical application

of a large genomic

of a new strategy library.

map of the whole for rapid analysis

Cell 50 (1987) 495-508.

Koob, M. and Szybalski, W.: Cleaving yeast and Escherichia cofigenomes at a single site. Science 250 (1990) 271-273. Landy, A; and Ross, W.: Viral integration and excision: lambda aft sites. Science 197 (1977) 1147-I 161.

structure

of the

17 Linn, T. and Ralling, G.: A versatile multiple- and single-copy vector system for the in vitro construction oftranscriptional fusions to IacZ. Plasmid 14 (1985) 134-142. Manoil, C. and Beckwith, J.: TnphoA: a transposon probe for protein export signals. Proc. Natl. Acad. Sci. USA 82 (1985) 8129-8133. McKenney, K., Shimatake, H., Court, D., Schmeissner, U., Brady, C. and Rosenberg, M.: A system to study promoter and terminator signals recognized by Escherichiu coli RNA polymerase. In: Chirikijan, J.G. and Papas, T.S. (Eds.), Gene Ampii~cation and Analysis, 2. Analysis ofNucleic Acids by Enzymatic Methods. Elsevier, Amsterdam, 1981, pp. 383-415. Miller, J.H.: Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1972. Podhajska, A.J., Hasan, N. and Szybalski, W.: Control of cloned gene expression by promoter inversion in vivo: construction of the heatpulse-activated atf-null-p-art-N module. Gene 40 (1985) 163-168.

Rasmussen, L.J., Mdler, P.L. and Atlung, T.: Carbon metabolism regulates the expression of the $I (pyruvate formate-lyase) gene in E~c~e~c~a CC&.J. Bacterial. (1991) in press. Redfield, R.J. and Campbell, A.M.: Structure of cryptic d prophages. J. Mol. Biol. 198 (1987) 393-404. Schneider, K. and Beck, C.F.: Promoter-probe vectors for the analysis of divergently arranged promoters. Gene 42 (1986) 37-48. Weisberg, R.A. and Landy, A.: Site-specitic recombination in phage lambda. In: Hendrix, R.W., Roberts, J.W., Stahl, F.W. and Weisberg, R.A. (Eds.), Lambda II. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1983, pp. 21 l-250. Yanisch-Perron, C., Vieira, J. and Messing, J.: Improved Ml3 phage cloning vectors and host strains: nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene 33 (1985) 103-119.