Eur. J. Biochem. 204, 51 -56 (1992) $) FEBS 1992
Characterization of the upp gene encoding uracil phosphoribosyltransferase of Eschevichia coli K12 Paal Skytt ANDERSEN', John M. SMITH' and Bente MYGIND'
'
Enzyme Division, Institute of Biological Chemistry B, University of Copenhagen, Denmark Seattle Biomedical Research Institute, USA
(Received July 29/0ctober 13, 1991)
- EJB 91 1014
The upp gene coding for uracil phosphoribosyltransferase was subcloned on a 5-kb EcoRI restriction fragment along with the purMN operon. By a combination of complementation, deletion and minicell analyses, the upp gene was located adjacent to and divergently transcribed from the purMN operon. All three gene products could be identified in minicell extracts. The cloned upp gene shows an elevated expression upon uracil starvation. The nucleotide sequence and transcription start of the gene were determined. The sequence yields an open reading frame of 624 nucleotides encoding a protein of 22.5 kDa which is in agreement with the previously determined subunit M , of the purified enzyme. A putative 5-phosphoribosyl-a-1 -diphosphate (PRPP) binding site has been identified which is similar to the PRPP binding site of the yeast uracil phosphoribosyltransferase.
Uracil phosphoribosyltransferase (UPRTase) from Escherichia coli is a pyrimidine salvage enzyme which catalyses the synthesis of UMP from uracil and PRPP. UMP is the precursor for all pyrimidine nucleotides. The pyrimidine salvage enzymes enable the cells to utilize preformed nucleobases and nucleosides either from the growth medium or from degradation products of cellular nucleic acids (Fig. 1;[l]).The UPRTase has been purified from E. coli K12 and has been reported to be a trimer with a subunit M , of 23 500 [ 2 ] . The upp gene encoding the UPRTase has been mapped to 55 min on the E. coli linkage map [ 3 ] . The level of UPRTase has been shown to be elevated approximately tenfold upon pyrimidine starvation [2]. We have previously identified and isolated a recombinant plasmid containing the upp gene from a E. coli ColEl library [2]. In the present study we report the nucleotide sequence and the transcription initiation site of the upp gene. We have identified the gene products in minicells. In addition, we have analyzed the regulation of the upp gene in plasmid harbouring strains. Furthermore, wc report a comparative sequence analysis with the U PRTase sequence from Saccharomyces cerevisiur [4]. -.
.-
Correspondence to R. Mygind, Enzyme Division, Institute of Biological Chemistry B, University of Copenhagen, Snlvgade 83, DK1307 Copenhagen K , Denmark Abhrcviutions. UPRTase, uracil phosphoribosyltransferase; PKPP, 5-phosphoribosyl-a-I -diphosphate. Enzymes. Uracil phosphoribosyltransferase (EC 2.4.2.9); orolatc phosphoribosyltransfcrase (EC 2.4.2.1 0). Notr. The novel nucleotide sequence data for the E. cob upp genc has bcen submitted to the EMBL nucleotide sequence database and is available under accession number X57104.
CTP +pyrG
4 ndk CDP 4 crnk CMP
4 udk
UTP
4 ndk 4 w~
UDP
"'4
i
cytidine
&uridine
cytosine
ax14, uracil
de novo
synthesis
upp
-dJ
Fig. 1. Pyrimidine salvage pathway of E. coli. The enzymes are identified by their corresponding gene designations as follows: cdd, cytidine deaminase (EC 3.5.4.5); cmk, CMP kinase (EC 2.7.4.-); codA, cytosine deaminasc (EC 3.5.4.1); ndk, nucleoside-diphospho kinase (EC 2.7.4.6); pyrG, CTP synthetase (EC 6.3.4.2); pyrH, UMP kinase (EC 2.7.4.14); udk, uridine kinase (EC 2.7.2.48); udp, uridine phosphorylase (EC 2.4.2.3); upp, uracil phosphoribosyltransferase (EC 2.4.2.9).
MATERIALS AND METHODS Materials
Restriction endonucleases and DNA modifying enzymes were obtained from either Boehringer Mannheim Biochemicals (BMB) or New England Biolabs [~i-~~S]thiodATP, [y32P]ATP and [3sS]methionine were obtained from New England Nuclear, Du Pont (NEN). Bacterial strains and plasmids
The bacterial strains used in these studies are listed in 'Table 1. Constructions of BM604 and S(a6305 were carried
52
DNA sequencing
Table 1. Bacterial strains.
Strain
Genotype
S4114O BM605 BM604
S41537 CHSOl S4106 S41616 S46305 MC 1 000 NM522
From S41140 by P1 (H0300), [8]
DNA sequencing was performed using the dideoxy chain termination method [16] with Sequenase (United States Biochemical Corp.). Sequences were obtained from fragments cloned into either the single-stranded vectors M13mp18/19 [17], or the pBluescript phagemid vectors (Stratagene). Specific oligonucleotide primers were also used to provide a complete sequence strategy.
From BM605 by P1 (S4106)
Primer extension
Source/Reference ~~
~
HfrH thi galE A(attl-bio) deoA103 deoC nrgA IysA cytR UPP UdP HfrH thi galE A(attl-bio) deoA103 deoC argA cytR upp udp trp zch-2410 : :TnlO HfrH thi ga1.E A(artA-bio) deoAI03 deoC argA lysA cytR upp udp pyrF30 HfrR4 gulE purM48 rnetBl relAI lacZ rpsL strA thi trp lacZ rpsL strA thipyrF30 lacZ rpsL strA thi trp upp ::TnS lacZ rpsL strA thi upp:: Tn5 pyrF30 thi A(ara-leu) 7679 araD139 A (lacX) 74 galU galK rpsL supE thi A(1ac-proAB) hsd-5 [F' proAB lac19 IacZAMlS]
[71
Laboratory collection
PI
Laboratory collection R . Fast, unpublished results From S41616 by P1 (S4106)
[9]
[lo]
out by P1 mediated transduction of described by Miller [5]. The plasmid pBR322 [6] was used for cloning and gene expression. Media and growth conditions For routine purposes, the L-broth medium was used [ 5 ] . When required, ampicillin was added to a concentration of 100 pg . ml-1 and tetracycline to a concentration of 10 pg . m1-I. For determination of enzyme levels, cells were grown at 37"C in AB minimal medium [ l l ] supplemented with 0.2% glucose as the carbon source and 0.2% casamino acids. Pyrimidine starvation was accomplised by growing pyrimidine auxotrophic strains with UMP (50 pg . m1-l) as the sole pyrimidine source in a high phosphate medium. Dephosphorylation of the UMP is the limiting step and therefore the pyrimidine supply can become limiting for growth. This is seen in the growth rate; doubling times on uracil and UMP are approximately 50 and 80min, respectively. Unstarved cells were grown with uracil (20 pg . ml-') as the source of pyrimidine. The cells were harvested in the exponential phase and subsequently sonicated for analyses of enzyme levels. Enzyme assays UPRTase activity was assayed as described previously [2]. Protein determination was performed according to [12]. Gene expression in minicells Minicells from BD1854 harbouring appropriate plasmids were prepared as described [13]. Plasmid-encoded proteins were labeled with [35S]methionine and were separated on 12.5% SDSjPAGE [I41 and detected by autoradiography. Recombinant DNA techniques Restriction endonuclease digestions were performed according to the specifications of the supplier. Recombinant DNA techniques were as described by Maniatis et al. [I 51.
Cells were grown in LB medium to A436 = 0.8 and RNA was extracted by a hot phenol method [18]. Primer extension assays were carried out as follows: 1 or 10 pg RNA was hybridized to 1 pmol primer in 0.1 M KC1 by heating at 8O'C followed by slow cooling to room temperature. 4 pI of the four dNTP's each at 2.5 mM, 4 p1 5 x MLV reverse transcriptase buffer and 200 U MLV reverse transcriptase [Bethesda Research Laboratories (BRL)] were added. Reactions were incubated for 1 h at 37°C after which they were stopped by adding 20 p1 Sequenase stop solution (USB). The synthetic oligonucleotide used was 5'-CTTCGTAAGTCAGC-3' which hybridizes to nucleotides + 164151. The primer was 5'labeled using y-[32P]ATP(NEN) and T4 polynucleotide kinase (BMB). A DNA sequencing ladder obtained using the same primer, was separated by electrophoresis alongside the products of the primer extension assay.
+
RESULTS Complementation of UPRTase deficiency Subcloning of the upp gene from a ColEl hybridplasmid. The upp locus has been genetically mapped to be very close to the purM gene (Mygind, unpublished results). Accordingly, it has beenshown that a hybrid ColEl plasmid, pLC1-41, carried both the purM and the upp genes [2]. To isolate the upp gene pLC1-41 further, DNA was digested with EcoRI and ligated to EcoR1-digested pBR322. From the resulting plasmids both purM+ and upp' clones were selected by complementation of purM (S41537) or upp (BM604). BM604 is unable to use uracil, but will grow on uridine as a pyrimidine source. In bothpurM' and upp+ selections the same 5-kb EcoRI fragment was obtained. The plasmids carrying the fragments in either orientation were designated pBM6 and pBM9. To localize the upp gene the plasmids were digested with different restriction enzymes and analyzed for complementation of either upp or purM mutations (Fig. 2). The PstI deletion of pBM9 resulted in plasmid pBM91 which retained the purM gene, but lost its ability to complement the upp mutation. The EcoRV deletion leading to pBM92 resulted in loss of both capabilities. However, both SalI, MluI and SspI deletions retained the upp gene but lost the purMN genes. UPRTase activity directed from the upp-hurbouring plasmids. In order to measure the activity of the plasmidencoded UPRTase protein, plasmids were transformed into ,546305 (upp ::Tn5 pyrF). This strain can be subjected to pyrimidine starvation using UMP as the sole pyrimidine source. Cell extracts of S46305, harbouring three different plasmids were assayed for UPRTase activity and compared with the level of a upp' haploid strain (Table 2). All the plasmids express upp. When grown on uracil both pBM6 and pSA4 show a marked elevation (sixfold and tenfold) compared to the chromosomal level. pSA4 shows an almost twofold-higher
53 1 kb
ET pBM6
pSA4
5
purMN I
P
*
SI I
EI1
SI
ET
s
I2 EV
pBM92
EL
Upp
PurM
+ +
+ +
+
-
-
+
I2 I2
pSA6
pBM9 1
UPP
M 2 S EV
I2 P
Fig. 2. Restriction map of the upp region and complementation of upp andpurM. Deleted derivatives of either pBM6 or pBM9 (having the 5-kb insert in the opposite oricnhtion) were produced by treatment with restriction endonucleases and subsequent religation of the deleted plasmid. Hori7ontal lines represent the inserts of each subclone and solid bars represent the position of the genes as designated. The ability of the resulting clones to complement either upp orpurM mutations were determined and their Upp and PurM phenotypes are shown in the righthand column. Restriction sites are given as follows: El, EcoRI; M, MluI; s, SspI; EV, EcoRV; P, PstI; SI, SalI. lable 2. Effect of pyrimidinestarvation on UPRTase activity. Numbers in parenthesis indicate the ratio of activity between activity in cclls grown with UMP to that of cells grown with uracil. Strain
Pyrimidine source
Specific activity of UPRTase nmol product . min-’ . mg protein-
S4106
[email protected]/pBM6 S46305/pSA4 S46305/pSA6
uracil UMP uracil UMP uracil UMP uracil UMP
5 24 ( 5 ) 32 180 (6) 50 162 (3) 17 32 (2)
level of expression than the pBM6 plasmid when grown with uracil as the pyrimidine source. The reason for this elevation could be that pSA4 unlike both pBM6 and pSA6 lacks a gene, uruA, which is involved in the transport of uracil; thus the cells may be partly starved for uracil even when grown in a medium containing uracil [18a]. pBM6 and pSA4 show a similar level of expression when grown with UMP as the pyrimidine source which is approximately sevenfold higher than the chromosomal level of UMP grown cells and also higher than the level of the unstarved cells (sixfold and threefold, respectively). One plasmid pSA6 clearly differs. The basal level is only threefold higher than the chromosomal level. Also, it does not show any elevated expression upon uracil starvation. The SspI site is positioned very close to the -35 region of the upp promoter and this suggests that the region upstream from the - 35 region is necessary for regulation. Synthesib of UPRTuse in minicells. In order to locate the transcribed region of upp and the orientation of transcription, some of the plasmids were analyzed in minicells. Results of analysis of one upp complementing and two non-complementing plasmids are shown in Fig. 3. Plasmid-encoded proteins labelled with [35S]methioninewere separated by 12.5% SDSjPAGE (Fig. 3). Plasmid pBM6 (lane 1) produced a poly-
Fig. 3. Autoradiograph of IJ5S~rnethionine-labeledpolypeptides from ininicell extracts. Lane 1, pBM6; lane 2, pBM92; lane 3, pBM91; lane 4,pBR322. The position and size (in kDa) of molecular mass markers are indicated at the right. Also the known proteins are indicated at the left. PurM, 5’-phosphoribosyl-5-aminoimidazole synthetase; Bla, I{-lactamase; PurN, 5’-phosphoribosylglycinamide transformylase; Upp, uracil phosphoribosyltransferase.
peptide with a M , of 23 000. This is in fine agreement with the published M , of 23 500 [2]. It also coded for polypeptides with an M , of 39000 and 27000. These M , values are in agreement
54 ACGCGTTACTTGCGGTAGAT
-274
purklt -250 -183 -116 -49
19
74
125
AAAATTCGGCGCAATTCTAACAGGGAAAGCAAACGTTTGCGAGACTGCTTTACACAACCTTTTTGCA CGTCTTTTCCCCAGGCGCGCGGCGAAAGAAGACTTGTGCCAGGG~AAAGGTTAGTTTTCGGATGG~ TAATCTTCTTTCATAACCATCTGAATATAAAATAACTTTATCTCAAACCG~TATCATTTTGACTAAA -35 -1 0 1 GTCAACGAAAAGAATATTGCCGCCTTG~G~GGAGGTATAATCCGTC~ATTTTTTTTGTGGCTGC SD CCCTCAAAGGAGAAAGAGT ATG AAG ATC GTG GAA GTC AAA CAC CCA CTC GTC AAA M K I V E V K H P L V K
CAC AAG CTG GGA CTG ATG CGT GAG CAA GAT ATC AGC ACC AAG CGC TTT CGC H K L G L M R E Q D I S T K R F R GAA CTC GCT TCC E L A S
GAA
E
GTG GGT AGC CTG CTG ACT TAC GAA GCG ACC GCC GAC V G S L L T Y E A T A D
176
CTC GAA ACG GAA AAA GTA ACT ATC GAA GGC TGG AAC GGC CCG GTA GAA ATC L E T E K V T I E G W N G P V E I
227
GAC CAG ATC AAA GGT AAG AAA ATT ACC GTT GTG CCA ATT CTG CGT GCG GGT D Q I K G K K I T V V P I L R A G
210
CTT GGT ATG ATG GAC GGT GTG CTG GAA AAC GTT CCG AGC GCG CGC ATC AGC L G M M D G V L E N V P S A R I S
32 9
GTT
V
GTC GGT ATG TAC CGT AAT GAA GAA ACG CTG GAG CCG GTA CCG TAC TTC V G M Y R N E E T L E P V P Y F
380
CAG A A A CTG GTT TCT AAC ATC GAT GAG CGT ATG GCG CTG ATC GTT GAC CCA Q K L V S N I D E R M A L I V D P
431
ATG CTG GCA ACC GGT GGT TCC GTT ATC GCG ACC ATC GAC CTG CTG AAA AAA M L A T G G S V I A T I D L L K K
482
GCG GGC TGC AGC AGC ATC AAA GTT CTG GTG CTG GTA GCT GCG CCA GAA GGT A G C S S I K V L V L V A A P E G
533
ATC GCT GCG CTG GAA AAA GCG CAC CCG GAC GTC GAA CTG TAT ACC GCA TCG I A A L E K A H P D V E L Y T A S
584
ATT GAT
I
D
CAG GGA CTG AAC GAG CAC GGA TAC ATT ATT CCG GGC CTC GGC GAT Q G L N E H G Y I I P G L G D
635
T GCC GGT GAC AAA ATC TTT GGT ACG AAA T A A A G A A T ~ T A A T T A G C C G A C T T T A A G D K I F G T K
693
AGAGTCGGCTTTTTTTTGAGTAAAGCGCCTATAC
Fig. 4. Nucleotide sequence and derived amino acid sequence of the upp gene. The promoter region, transcriptional initiation point, the ShineDalgarno (SD) and the terminator regon (T) are underlined. The purM transcriptional start and orientation is indicated by a bold letter and an arrow.
with those for 5’-phosphoribosyl-5-aminoimidazolesynthetase (PurM) [19] and 5’-phosphoribosylglycineamide transformylase (PurN) [20]. The two non-complementing plasmids lack the 23-kDa polypeptide. pBM92 (lane 2) produced a protein pattern very similar to that of the parent vector pBR322 (lane 4) whereas the pBM91 (lane 3) also produced a polypeptide with an approximate M, of 19 in addition to the products of purM andpurN. We suggest that this polypeptide is a truncated form of U PRTase. These results suggest that the upp gene is very close to and is transcribed divergently from the the purMN operon.
Nucleotide sequence of upp The finding that the PstI deletion in pBM91 removed the end of the upp gene resulting in the production of a 19-kDa polypeptide positioned the start of the coding region about
0.6 kb upstream from the PstI site. The nucleotide sequence of this region has previously been determined in connection with the sequencing of thepurM gene [19]. We have sequenced the entire upp region using either universal or specific primers following subcloning of appropriate fragments into the pBluescript phagemid vectors or the MI 3mpl8/19 phages. The sequence was determined on both strands for the entire region. The nucleotide sequence (Fig. 4) reveals an open reading frame of 624 nucleotides starting with an AUG codon at position + 37. Proximal to this is a possible Shine-Dalgarno sequence with a spacing of seven nucleotides. This start codon is the most likely start codon based on the Shine-Dalgarno sequence. The open reading frame corresponds to a polypeptide with a molecular mass of 22.5 kDa which is in agreement with the published subunit molecular mass [2]. The open reading frame terminates at position 661 with a UAA stop codon. 17 bp downstream from the stop codon there is a region of dyad symmetry that resembles a rho-independent termina-
55 tor [21]. The strength of this possible hairpin structure is -56.7 kJ/mol. However, we believe that this sequence is an internal terminator of a polycistronic operon [Izla].
DISCUSSION
In the present study we have identified the upp gene whose product UPRTase is a central enzyme in the pyrimidine salvage pathway. The regulation of expression of the genes in Mapping of the transcription start point the pyrimidine de novo pathway has been extensively studied. Expression of all genes encoding pyrimidine metabolic enIn order to determine the upp transcriptional start point, zymes is increased under conditions of pyrimidine depletion. we performed a primer-extension analysis. RNA was isolated However, no common pyrimidine regulatory system is responfrom a strain harbouring the upp complementing plasmid sible for this coordinated response. Transcriptional attenupBM6. The 5' end of the mRNA was localized 37 bp upstream ation control of pyrB [22] and pyrE [23], and translational from the presumed translational initiation codon (Fig. 5). control of pyrC [18] and pyrD, [24] have been analyzed in These results were confirmed in S1 nuclease protection studies detail. At present, the mechanisms by which pyrimidines regu(data not shown). Six base pairs upstream from the transcrip- late the expression of the carAB and pyrF genes remain untion start a consensus - 10 region is found. A region resem- known. The regulatory mechanisms controlling expression of bling the - 35 consensus sequence in a proper distance is also the genes, codA (encoding cytosine deaminase), udk (encoding identified. uridine kinase) and upp, are even less clear. At present only the regulation of codA has been investigated. For the expression of this gene, purine repression [25] and nitrogen control are observed along with pyrimidine repression [26]. The expression of the udk and upp genes increase upon pyrimidine starvation [l]. Neither of these genes respond markedly to the availability of nitrogen or purines as is the case for the codA gene [l, 261. In the present study we have observed the same fold of pyrimidine regulation of upp gene expression when the upp gene appears on high copy number pbdsmids as in haploid cells which suggests that the regulatory mechanism may not include a titratable element. The entire nucleotide sequence and the transcription start site have been determined. The sequence between the transcriptional start site and the coding sequence does not reveal any regions similar to the attenuator sequences of pyrB or pyrE. However, it is notable that a stretch of eight thymine residues follows immediately after the transcription initiation point. Therefore, the RNA polymerase requires many UTP molecules in the first elongation cycles after transcription initiation. We do not know whether the polymerization of these residues is involved in the regulation of gene expression. We find in one of our plasmid constructions (pSA6) that upp expression is not increased upon uracil starvation. The plasmid has been deleted of the region upstream from the -35 region but the -35 region is left intact. Therefore, it Fig, 5. Primer extension mapping of the 5' terminus of the upp mRNA. appears that the upstream region is necessary for pyrimidine Total RNA was isolated as described in Materials and Methods. regulation. The nucleotide sequence in that region does not Lancs G, A, T and C contain a sequcnce ladder of the upp promoter reveal any obvious regions of dyad symmetry, and it is not region. Lane 1. primcr extension with 1 kg RNA; lane 2, primer cxtension with 10 pg RNA. at present clear what kind of regulation is involved in the
E.c.
MKIVEVKHPLVKHKLGLMREQDISTKRFRELASEVGSLLTYEATADLETEKVTIEGWNGPVEIDQIK-GK ::I::::: I . . .- : l l : I : : : I : : I : : I : : : : : : ::
69
S.C.
NVYLLPQTNQLLGLYTIIRNKNTTRPDFIFYSDRI:[RLLVEEGLNHLPVQKQIVET-DTN~NFEGVSFMG
111
G--XXXXDD-X-TGGT :Ill1 I Ill: KITVVPILRAGLGMMDGVLENVPSARISWGMYRNEETLEPVPYFQKLVSNIDERMALIVDPMLATGGSV 139 / I I : I : / I I :I :I: : :l:ll: : : I : I I I I :::I/ ::l:ll ::::II/IIIIII: KICGVSIVRAGESMEQGLRDCCRSVRIGKILIQRDEETALPLKFYEKLPEDISER~FLLDPMLATGGSA 181
IATIDLLKKAGCSSIKV--LVLVAAPEGIAALEKAHPDVELYTASIDQGLNEHGYIIPGLGDAGDKIFGTK I ::::I I I :: : : I I : : I l l : : I I:I : l:::l:ll:l: l::lllll 11: : IMATEVLIKRGVKPERIYFLNLICSKEGIEKYHAAFPEVRIVTGALDRGLDENKYLVPGLGDFGDRYYCV
208 251
Fig. 6. Scqucnce alignment of the UPRTases of E. cofi (E. c.) and 5'. ceirrvisiuc (S. c.). Computer-aided scquence alignment was made using programs descrihcd by Devereux et al. [27] as implemcnted on the Eh"BL sequence data base. Thc consensus PRPP binding site has been aligned with the sequences [2X]. Vertical lines (1) indicate identical amino acids. Colons (1) indicate similar amino acids. X denotes isoleucine, leucinc or valin residucs.
56 pyrimidine-mediated regulation. Whether the region contains secondary promoters or a recognition site for a trans-acting factor is presently under investigation. The nucleotide sequence of the gene encoding URPTdse from S. cerevisiae has recently been published [4]. This gene encodes a 28.7-kDa protein. We have compared the two deduced amino acid sequences (Fig. 6).The two UPRTases share many similarities in discrete areas despite only 32% overall identities. There are two regions of extensive similarity between the two UPRTases. One is the C-terminal region which may correspond to a uracil binding site. The second region covers amino acids 122- 137 in the E. colienzyme. This region has been aligned with the consensus sequence of PRPP binding sites based on 22 sequences of PRPP binding sites ([28]; HoveJensen, B. & Nilsson, D., University of Copenhagen, personal communication) (Fig. 6 ) . Both the E. coli and the S. cerevisiae UPRTase resemble this consensus sequence. It is interesting that the second aspartate residue of the consensus sequence which is conserved in all other PRPP-utilizing enzymes is replaced by a proline residue in both UPRTases. The region in which the two enzymes differ the most is the N-terminal region. The S. cerevisiae UPRTase has an Nterminal extension of approximately 40 amino acids which is not present in the E. coli enzyme. The enzyme orotate phosphoribosyltransferase encoded by the p y r E gene of the pyrimidine de novo pathway catalyses a reaction similar to that of the UPRTase. The nucleotide sequence of the p y r E gene has been published [29]. Despite the fact that the pyrimidine substrates of the two enzymes differ only by the presence of a carboxyl group in orotate, their amino acid sequences show very little similarity apart from the sequence around the putative PRPP binding site. We believe that UPRTase is involved in the uptake and metabolism of uracil and preliminary experiments show that upp is the first gene in a polycistronic operon including a region which is necessary for uptake of uracil at very low concentrations PO]. We wish to thank Mogcns Kilstrup and Bjarne Hove-Jensen for careful reading of the manuscript and Robert Fast for providing the upp : : Tn5 strain. This work was supported by the Danish Center of Microbiology (B. M.) and NIH grant A1 20068 (J. M. S.).
REFERENCES 1 . Neuhard, J. (1983) in Metaholism ofnucleotides, nucleosides and
nucleohases in microorganisms (Munch-Petersen, A,, ed.) pp. 97 - 148, Academic Press, London. 2. Rasmussen, U . B., Mygind. B. & Nygaard, P. (1986) Biochern. B i o p h ~ .A~ .C ~ 881, U 268 - 275. 3. Pierard. A,, Glansdorff. N. & Yashphe, J. (1972) Mol. Gen. Genet. 118,235-245.
4. Kern, L., de Montigny, J., Jund, R. & Lacroute, F. (1990) Gene 88, 149- 157. 5. Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor, Cold Spring Harbor Press, New York. 6. Bolivar, R., Rodriguez, R. L., Greene, P. J., Belach, M., Heyneker, H. L. & Boyer, H. W. (1977) Gene 2,95- 113. 7. Munch-Petersen, A., Mygind, B., Nicolaisen, A. & Pihl, N. J. (1979) J . Biol. Chem. 254, 3730-3737. 8. Hove-Jensen, B. (1983) J . Bacteriol. 154, 177-184. 9. Casadaban, M. J. & Cohen, S . M. (1980) J. Mol. Biol. 138, 179207. 10. Mead, D. A,, Skorupa, E. S . & Kemper, B. (1985) Nucleic Acids Res. 13, 1103-1110. 11. Clark, D. J. & Maalse, 0.(1967) J. Mol. B i d . 23,99- 112. 12. Lowry, 0. H., Rosebrough, N. F., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193, 265-275. 13. Jensen, K. F., Larsen, J. N., Schack, L. & Sivertsen, A . (1984) Eur. J. Biochem. 140,343 - 352. 14. Laemmli, U. K. (1970) Nature 227, 680-685. 15. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular cloning: a laboratory manual, Cold Spring Harbor, Cold Spring Harbor Press, New York. 16. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Nut1 Acud. Sci. USA 74, 5463 - 5467. 17. Yanisch-Perron, C., Vieira, J. & Messing, J. (1985) Gene33,103119. 18. Ssrensen, K . I. & Neuhard, J. (1991) Mol. Gen. Genet. 225,249256. 18a. Andersen, P.-S. & Mygind, B. (1990) Abstr. Annu. Meet. Am. SOC.Microbiol. K75. 19. Smith, J. M. & Daum 111, H. A. (1986) J . B i d . Chem. 261, 10632-10636. 20. Smith, J. M. & Daum 111, H. A. (1987) J . B i d . Chem. 262, 10565- 10569. 21. Rosenberg, M. & Court, D. (1979) Annu. Rev. Genet. 13, 319353. 22. Turnbough, C. L., Hicks, K. L. & Donahue, J. P. (1983) Proc. Nut1 Acad. Sci. USA 80, 368 - 372. 23. Poulsen, P., Bonekamp, F. & Jensen, K. F. (1984) EMBO J . 3, 1783-1790. 24. Frick, M. M., Neuhard, J . & Kelln, R. A. (1990) Eur. J . Biochern. 194,573 - 578. 25. Kilstrup, M., Meng, L. M., Neuhard, J. & Nygaard, P. (1989) J . Bacteriol. 171,2124 - 2127. 26. Andersen, L., Kilstrup, M. & Neuhard, J. (1989) Arch. Microhiol. 152, 115-118. 27. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acidy Res. 12, 387-395. 28. Hove-Jensen, B., Harlow, K . W., King, C. J. & Switzer, R. L. (1986) J. Biol. Chem. 261, 6765-6711. 29. Poulsen, P., Jensen, K. F., Valentin-Hansen, P., Carlsson, P. & Lundberg, L. G. (1983) Eur. .I. Biochem. 135, 223229. 30. Munch-Petersen, A. & Mygind, B. (1983) in Metuholisrn o j nucleotides, nucleosides and nucleobases in microorganisms (Munch-Petersen, A,, ed.) pp. 259 - 305, Academic Press, London.
