JOURNAL OF BACTERIOLOGY, June 1979, p. 705-7140021-9193/79/06-0705/10$02.00/0

Vol. 138, No. 3

Organization of Structural and Regulatory Genes ThatMediate Tetracycline Resistance in Transposon TnlO

RICHARD A. JORGENSENt AND WILLIAM S. REZNIKOFF*Department ofBiochemistry, College ofAgricultural and Life Sciences, University of Wisconsin-Madison,

Madison, Wisconsin 53706

Received for publication 8 March 1979

The location of TnlO genes encoding tetracycline resistance and its regulationwas determined by analyzing the properties of recombinant plasmids carryingpartial HpaI digestion products of A::TnlO transducing phage deoxyribonucleicacid. Within a 2,700-base pair region are encoded tetracycline resistance, thestructural gene (tet) for a tetracycline-inducible polypeptide, and the regulatoryelements for the induction of both the resistance phenotype and the polypeptide.Fusion of different sequences to an HpaI site in the tet gene alters the molecularweight and stability of the polypeptide as well as the tetracycline resistancephenotype of strains producing fusion polypeptides. These results indicate theorientation of the tet gene and support the conclusion that the tet polypeptide isrequired for tetracycline resistance. A Hincd cleavage site immediately upstreamfrom the tet gene is protected by ribonucleic acid polymerase, but only in theabsence of ribonucleotide triphosphates. The possibility that tet transcription isinitiated at this site is discussed.

Tetracycline resistance specified by the R-fac-tor R100 (also called R222 and NR1) is carriedby a 9,300-base pair-long transposon termedTnlO (16, 30). The mechanism of resistance totetracycline involves an inhibition of uptake,rather than degradation or modification, of thedrug (12, 20). Uptake of tetracycline by sensitiveEscherichia coli cells involves at least two in-dependent transport systems, one active and onepassive (26); both systems are depressed in tet-racycline-resistant cells (20). In addition to up-take inhibition, resistant cells are reported topossess an internal (intracellular) mechanism forinhibition of tetracycline activity (20). Thus, acombination of decreased transport and de-creased sensitivity of intracellular componentsseems to be responsible for the resistance me-diated by R100. Further, these observations sug-gest that several gene products may be involvedin tetracycline resistance.The expression of tetracycline resistance is

regulated in that pretreatment of resistant cellswith subinhibitory levels of the drug increasesthe level of resistance (12). Levy and McMurry(19) and Levy et al. (21) have reported the R-factor-directed synthesis of a tetracycline-induc-ible membrane-associated polypeptide ("TET"protein) and suggested that this protein is re-sponsible for tetracycline resistance. That the

t Present address: Carnegie Institution of Washington, De-partment of Plant Biology, Stanford, CA 94305.

regulation of tetracycline resistance is mediated,at least in part, by a repressor has been suggestedby Yang et al. (35), who reported the partialpurification from R222-containing cells of aninhibitor ofTET protein synthesis in vitro. Thisinhibition was reversible by the addition of tet-racycline.

In this communication we described experi-ments which examine the ability ofrecombinantplasmids carrying TnlO deletions affecting tet-racycline resistance to synthesize polypeptidesin E. coli minicells. These studies demonstratethat the TET protein is indeed encoded by TnlOand is necessary for tetracycline resistance. Thegene encoding the TET protein (tet) was pre-cisely located within TnlO, and its direction oftranscription was determined. In addition, theprobable location of the tet promoter region isinferred from RNA polymerase binding and pro-tection experiments.

MATERIALS AND METHODSPhage and bacterial strains. The A::Tn1O phage

used here is A::TnlO(l), described previously (14).Strain C600 rK_ mK (from J. Davies) is thr leu lacB1- and is defective in the K-12 restriction and mod-ification system. X984 (from R. Curtiss) is a minicell-producing strain of E. coli that is met ilv his adepdx.DNA from Akan3 (6) and R-factors JR225 and NR79were kindly provided by J. Davies. Plasmid R100 (alsoknown as NR1 or R222) was obtained from R. Rowndand was introduced into C600 rK mK and X984 byconjugation.

705

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

706 JORGENSEN AND REZNIKOFF

DNA preparation and restriction enzymedigestion. Phage DNA was prepared for restrictionenzyme treatment by phenol extraction as describedby Barnes et al. (4). Plasmid DNA was prepared bythe method ofHumphreys et al. (11). Enzymatic diges-tion of DNA was performed in 0.05 ml of 60 mMMgSO4-7 mM MgSO4-10 mM Tris-hydrochloride (pH7.9)-10 mM f,-mercaptoethanol.Enzymes. HpaI, HincII, Sall, AvaI, BglI, BamHI,

XbaI, PstI, and KpnI restriction enzymes and T4 DNAligase were obtained from New England Bio-Labs.EcoRI was a gift from J. Gardner and S. Hardies.Agarose and polyacrylamide gel electropho-

retic analysis of DNA. Horizontal agarose gel elec-trophoresis was carried out essentially as described byShinnick et al. (31). DNA analysis by polyacrylamideslab gel electrophoresis was performed essentially as

described by Maniatis et al. (24), except that the gelswere made in 5% glycerol. DNA fragments were

stained for photography under UV light by immersingthe gels in a solution of 10,ug of ethidium bromide perml for 10 min, followed by rinsing and destaining inwater for 20 min.

Construction of recombinant plasmids carry-ing TnlO HpaI fragments. Five micrograms of A::

TnlO DNA was incompletely digested with HpaI, andthe enzyme was inactivated by heating to 70°C for 10min. One microgram of pVH51 DNA was digested tocompletion with HincII, which was also inactivated byheating. The reactions were monitored by agarose gelelectrophoresis. Samples of 0.45 ug of HpaI-cleavedA::TnlO DNA and 0.1 Mug of HincII-cleaved pVH51DNA in ligase buffer (20 mM Tris [pH 7.5], 10 mMMgCl2, 1 mM dithiothreitol, 1 mM Na2EDTA, and0.05 mM ATP) were incubated in the presence of 0.4M1 of T4 DNA ligase (92 units/ml) in a 15-p1 volume at15°C overnight (29). The resulting mixture was usedto transform strain C600 rK- mK by the method ofMandel and Higa (22). Transformed cells were allowedto outgrow for 90 min in nutrient broth and were thenspread on nutrient plates containing 20ug of tetracy-cline per ml. Forty-eight clones so obtained were

screened for plasmid size by a modification of theprocedure ofBarnes (3). Twenty-five clones containingplasmids of various sizes were characterized as de-scribed in the Results section. These strains were

handled under Pl-EK1 containment conditions as pre-

scribed by the National Institutes of Health.Measurement oftetracycline resistance levels.

Tetracycline resistance conferred by plasmids in C600rK mK was measured by a modification of the pro-

cedure of Tait et al. (32). Overnight cultures were

inoculated into fresh nutrient broth with and without1 Mg of tetracycline per ml, then incubated for 45 minat 37°C and diluted. Approximately 100 to 300 cells

were plated on nutrient plates containing 0, 25, 50, 75,100, 125, 150, 175, or 200 Mg of tetracycline per ml (forexperiment 1; experiment 3 also used 225- and 250-Ag/ml plates; see Table 1) or 0, 5, 10, 20, 30, 40, 50, 65, 80,100, or 120 Mug of tetracycline per ml (experiment 2,Table 1). The number of surviving colonies was scoredafter 20 h of incubation at 370C, and the concentrationof tetracycline that gave 50% efficiency of plating (theEOP50) was determined graphically. Since the age ofagar media containing tetracycline was found to sig-

nificantly affect EOP50 values, strains that were to becompared were tested at the same time on the samebatch of plates.

Purification and labeling of minicells. Minicellswere purified essentially as described by Roozen et al.(28), except that growth of the cells was in low-phos-phate medium (18) so that minicells could be labeledwith 32PO43- (F. Schmidt, R. Jorgensen, and J. Davies,unpublished data) as well as with [nS]methionine.Minicells obtained from a 400-ml culture were washedand suspended in 1 ml of 1:4 Met Assay Medium(Difco) in M9 salts with 0.5% glucose, 2 ,ug of pyridox-ine per ml, and 40 ug of adenine per ml. Minicells werethen divided into two aliquots, one of which wasinduced with 1 or 5 yg of tetracycline per ml. After a30-min induction period, both samples were labeledwith 20 to 50 IACi of [3S]methionine (specific activity,600 to 800 Ci/mmol, Amersham/Searle) for 30 min,pelleted, and frozen. For electrophoresis, pellets weresuspended in 50 p1 of sample buffer (17) and heated at90°C for 2 to 5 min. Samples of 20 pl were loaded on15% polyacrylamide-sodium dodecyl sulfate (SDS)slab gels (17) and electrophoresed at 35 mA for 5 to 6h on a constant current power supply. Gels were driedby heating under a vacuum at 700C and were autora-diographed at room temperature using Kodak X-OmatR film. Molecular weight markers used here were[3S]methionine-labeled proteins from purified bacte-riophage A particles which were prepared for electro-phoresis by mixing with unlabeled cells and heating to900C in sample buffer.RNA polymerase "protection experiment."

Four micrograms of E. coli RNA polymerase (4 mg/ml; a gift from S. Rothstein and L. Maquat) wasincubated with 1 ,ug ofpRT61 DNA at 37°C for 30 minin buffer PBBX + 0.1 M KCI (13) to allow formationof stable polymerase-DNA complexes. Six units ofHincII was then added, and incubation was continuedfor 60 min. Two controls were treated identically ex-cept (i) one received no RNA polymerase and (ii) onewas made 400 ,uM in the four ribonucleoside triphos-phates immediately prior to addition of HincIl. AfterHincII treatment, each mixture was phenol and etherextracted and subjected to electrophoresis on 8% poly-acrylamide gels as described above.

RESULTSConstruction and characterization of

deletions of TnlO defining the tetracyclineresistance genes. In an earlier communication(14) we described the determination of a restric-tion enzyme cleavage map for TnlO (Fig. 1) as afirst step toward studying gene organization inTnlO. As illustrated by this map, TnlO containsfour HpaI cleavage sites, which divide this ele-ment into five segments. Here we report theconstruction of a set of deletions in TnlO. Thesedeletions were generated by molecular cloningof X::TnlO incomplete HpaI cleavage productsinto a plasmid vector with selection for tetracy-cline resistance (Materials and Methods). Theplasmid cloning vehicle chosen for this experi-ment was the ColEl derivative pVH51, a small

J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

VOL. 138, 1979

-ColEI- x

GENE ORGANIZATION IN TnlO 707

TnlO

HH 3H1,4

H -H-

o.-rj D,& ° vE E S H H- 3 3

I 3-s- c'.aS--- CO CD I 4Y lLo 140 00

Co CD(I {I) (I) (I)aA Y 990 1515, '375 155IssIess,5 65 L6660 #n 12iIRMI

360 117Sf 1525 2525 1600 *6951 275 1675 140[ T00940 V 1650*m 3

in II I II IIR U UH3 3 IL

a * I1775--2025- - 675-FIG. 1. Restriction enzyme cleavage map of TnlO. Numbers refer to distances in base pairs. II refers to the

HincII cleavage sites. (I) refers to those HincII cleavage sites that are also HpaI cleavage sites.

(4-kilobase), amplifiable, multicopy plasmidwith a single HincII cleavage site (10). HincII(GTPy4PuAC) recognizes both HpaI(GTT4AAC) and SalI (GITCGAC) cleavagesites. The plasmid, pVH51, is not cut by eitherHpaI or Sail. Therefore, the sequence of theHincII site on the plasmid is GTC4AAC, andligation of a HpaI fragment at this site reconsti-tutes two HincII sites but only one HpaI site.To choose among plasmids for further study,

three criteria were used: (i) that all HpaI frag-ments cloned in each plasmid carry TnlO DNA(i.e., no X HpaI fragments are found in theplasmid unless that fragment also carries TnlODNA); (ii) that in each recombinant plasmidHincII cleavage sites have been reconstituted atthe junctions of TnlO and pVH51 DNA (i.e., noexonucleolytic degradation has occurred prior toligation); and (iii) that each plasmid results froman independent ligation event.Of 25 tetracycline-resistant clones examined,

8 met these criteria, established by HincII andHpaI restriction analysis of their plasmid DNAs(HpaI restriction patterns are shown in Fig. 2).The regions of TnlO carried in these eight re-combinant plasmids are illustrated in Fig. 3. Thesmallest plasmid, pRT29, carries only HpaI frag-ment 2025, indicating that this fragment alone issufficient for generating the tetracycline resist-ance phenotype.With the exception of pRT29, each plasmid

carries two or more HpaI fragments. The sim-plest way in which these plasmids could havebeen formed is by ligation of a single incompleteHpaI cleavage product to the HincII-cleavedpVH51 molecule. An alternate route is by liga-tion of two HpaI cleavage products (completelyor incompletely digested by HpaI) to the vector.The latter reaction would result in joining twofragments which do not adjoin in TnlO, or whichdo adjoin in the original sequence but in such amanner that the orientation of the fragments isreversed as compared with the original orienta-tion (i.e., the "wrong" ends of the molecules arejoined), or both. pRT4 is an obvious example ofthe former possibility (HpaI fragments 2025 and3000 are joined in this plasmid). Establishment

° c-0

::2-2 3 4 2224291<1

--> 7000

0---300

- 2025-_-- 1775

_- 675

FIG. 2. HpaI cleavage patterns ofpRT plasmidsand the indicatedphages. 1 refers topRTI, etc. Th1O-specific HpaIfragments are indicated by the arrows,and each fragment's size is given in base pairs.

of the map order and relative orientations ofneighboring fragments in each of these sevenplasmids was accomplished by double digestionexperiments (data not shown) with restrictionenzymes having only one cleavage site in thecloned region (e.g., BamHI, XbaI, EcoRI, Bgll,AvaI, and PstI; see map at the top of Fig. 3 forthe locations of these sites). Six ofthese plasmids(all except pRT4) carry the cloned HpaI frag-ments in the same map order and relative ori-entation as is found in TnlO. This suggests thatthese plasmids arose by cloning of a single in-complete HpaI cleavage product. In pRT4, HpaIfragments 2025 and 3000 are ligated with one ofthese fragments in the opposite orientation asfound in TnlO, i.e., HpaI fragment 3000 hasreversed its orientation such that its PstI cleav-age site lies within 600 base pairs of HpaI frag-ment 2025.To complete the physical mapping of the eight

-X-- Col El- ---

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

708 JORGENSEN AND REZNIKOFF

x

EcoRI

Hpa I

I 1775

Bom HI

HPOI HpOI HpoI

1 2025 I6751 3000ITII

XboI EcoRI S9IIpRT3

pRT24

pRT

pRT2

pRTI

pRT4

p RT22

Av I

tetclass

'@ tet"protein

Pst I

I I

II

I WT

I WT

WT

II WT

II f usion

IlI fusion

pRT29s I

fusion

EcoRI pRT61 PiII WT

FIG. 3. Physical maps ofrecombinant plasmids carryingportions of TnlO. At the top is a map of the TnlOportion ofa A::TnlOphage indicating appropriate cleavage sites. The lines below this map indicate the clonedportions of A::TnhO present in the indicated plasmid. pRT61, at the bottom, is an EcoRI-generated deletionderivation of a PstI TnlO fragment clone from A::ThlO. The unlabeled arrows at the end of the cloned HpaIfragments indicate the reconstituted HpaI cleavage sites. tetr (tetracycline resistance) classes are explained inthe Results section. wt refers to wild-type-size TETprotein; fusion refers to the altered TETproteins (see thetext).

recombinant plasmids, it is necessary to deter-mine the orientation of the cloned fragmentsrelative to pVH51 sequences. For pRT29 thiswas determined using the results of a KpnI-HpaIdouble digestion experiment. The results of thisexperiment (data not shown) indicated that thereconstituted HpaI cleavage site in pRT29 lieson that side of the cloned fragment nearest theKpnI cleavage site (see Fig. 4 for a map ofpRT29). Thus the nucleotide sequence of thepVH51 HincII site, reading from the EcoRI siteto the KpnI site, must be GTCJAAC. This in-formation, together with the comparisons ofHincII and HpaI restriction patterns describedearlier, allows us to orient the insertions of theother seven plasmids. In Fig. 3, that end of aninserted segment ofDNA at which the reconsti-tuted HpaI cleavage site lies is indicated by anunlabeled arrow. Interesting examples are pRT1and pRT24, in which the same cloned fragmentslie in opposite orientations relative to pVH51sequences.

Effect ofdeletions in TnlO on tetracyclineresistance. Table 1 presents a comparison oftetracycline resistance levels conferred on E. coliC600 cells by the R-factor R100, a X::TnlO pro-phage, and multicopy recombinant plasmidscarrying various portions of TnlO. Resistancelevels are reported here as the tetracycline con-centration at which the efficiency of plating fora given strain is 50% (EOP5o) and were measuredboth before and after treatment with subinhibi-tory levels of tetracycline. Since the age of agar

FIG. 4. Physical and genetic map ofpRT29. Theheavy line indicates the 2,025-base pair HpaI frag-ment of ThlO; the rest of the plasmid is pVH51. tetrefers to the gene encoding TET protein, col immrefers to the colicin immunity gene, and ori refers tothe ColEI origin of replication.

media containing tetracycline was found to sig-nificantly affect EOP5o values, strains that wereto be compared were tested in the same experi-ment on the same batch of plates. Basal (oruninduced) EOP50 values were found to be morevariable than induced EOPw0 values (compareexperiments 1, 2, and 3 in Table 1), such thatbasal values are not reproducible within experi-mental error limits. (Measurement of the basaltetracycline resistance phenotype of a strain, byplating cells on agar containing tetracycline, pre-

J. BACTERIOL.

4,RI

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

VOL. 138, 1979

TABLE 1. Tetracycline resistance phenotypes

Expt. Stain EOP50 (jg/mI) tet'nlo. Uninduced Induced cla

1 pRT61 31 ± 10 90 ± 10 IpRT3 38 ± 10 1056 10 IpRT24 35 ± 10 102 ± 10 IpRT1 38 ± 10 90 ± 10 IpRT11 69 ± 10 190 ± 10 IIpRT4 77 ± 10 >200 IIR100 80 ± 10 178 ± 10 Wild typeA::TnlO 90 ± 10 165 ± 10 Wild type

2 pRT3 64 ± 10 113 ± 10 IpRT22 26 ± 3 39 ± 3 IIIpRT29 26 ± 3 37 ± 3 IIIC600 rK mK <2 <2 Sensitive

3 pRT3 85 ± 10 120 ± 10 IpRT2 135 ± 20 '80 ± 10 IIpRT11 190 ± 15 200 ± 15 IIpRT4 150 ± 10 180 ± 10 II

sumably leads to some induction of resistance;this induction might be expected to be variable,depending on the precise experimental condi-tions.)Three classes of tetracycline resistance phe-

notypes can be distinguished among the recom-binant plasmids in Table 1 on the basis of theinduced EOP50 levels. Class I plasmids givenEOP50 values of 90 to 105 ,ig/ml for induced cells(experiment 1). This class includes plasmidspRT3, pRT24, and pRT1, as well as pRT61, arecombinant plasmid which carries the entireTnlO transposon (the construction ofpRT61 hasbeen described elsewhere [14]). Inspection of thephysical maps of class I plasmids indicates thatsequences in TnlO to the "left" ofHpaI fragment2025 (see Fig. 3) do not contribute to tetracyclineresistance. Also, from the fact that the insertionsin pRT1 and pRT24 are identical but in oppositeorientations relative to pVH51, we conclude thatthe orientation of this insertion has no effect onthe expression of tetracycline resistance in theseplasmids. Plasmids pRT2, pRT4, and pRT11comprise class II; these plasmids confer resist-ance to much higher levels of tetracycline thando class I plasmids (experiments 1 and 3, Table1). Class II plasmids pRT2 and pRT11 differfrom class I plasmids in that they do not carryHpaI fragment 3000; suggestions for a roleplayed by fragment 3000 in tetracycline resist-ance will be discussed later. Class II plasmidpRT4 differs from class I plasmid pRT24 in thatit is missing HpaI fragment 675 and in that theorientation of HpaI fragment 3000 is reversed.Explanation of the pRT4 resistance phenotypeis complex and will be discussed below. Class IIIincludes plasmids pRT22 and pRT29, whichspecify resistance only to low concentrations of

GENE ORGANIZATION IN Tnl0 709

tetracycline (experiment 2). These plasmids dif-fer from class I plasmids in that they are missingboth HpaI fragment 3000 and HpaI fragment675, suggesting that sequences in fragment 675play some role in resistance.

Tetracycline resistance levels in Table 1 arereported both before and after a 30-min induc-tion period in the presence of 1 ,ug of tetracyclineper ml. Resistance in all the recombinant plas-mid strains described here is inducible by thistreatment, indicating that the regulatory ele-ments for the induction process lie within the2,025-base pair HpaI fragment ofTnlO. It shouldbe noted that the ratio of induced to uninducedEOP50 values may not be an exact reflection ofthe true induction ratio of those functions re-sponsible for resistance, since measurement ofthe tetracycline resistance phenotype itself pre-sumably leads to some induction.

Cells containing the R-factor R100 have muchhigher EOP50 values than pRT61 (a multicopyplasmid carrying the entire TnlO transposon) orthe other class I plasmids. This observation canbe explained in two ways: (i) R100 encodes func-tions not found in TnlO but which contribute totetracycline resistance, or (ii) tetracycline resist-ance is less well expressed when TnlO is at highcopy number. To distinguish these possibilities,the EOP50 values for cells containing a lysogenicX::TnlO phage were determined (Table 1), withthe result that a X::TnlO lysogen is, within error,resistant to the same levels of tetracycline as acell containing R100. This result favors the hy-pothesis that the low tetracycline resistance lev-els found in class I plasmids result from the highcopy number ofthese plasmids. Class II plasmidshave a resistance phenotype similar to R100;this observation suggests that a TnlO-encodedfunction that causes lower resistance levels athigh copy number maps in HpaI fragment 3000,a fragment that is missing class I plasmids. Thishypothesis will be considered further in the Dis-cussion.Effect of deletions in TnlO on the synthe-

sis of a tetracycline-inducible polypeptide.Many tetracycline-resistant R-factor strains en-code a tetracycline-inducible polypeptide (re-ferred to as TET protein) (19, 21). We haveexamined polypeptide synthesis in E. coli mini-cells containing recombinant plasmids to deter-mine whether the TET protein encoded by R100is encoded within TnlO, as well as to examinethe effects of the various TnlO deletions carriedby these recombinant plasmids on the synthesisof this protein. In this experiment, purified min-icells were labeled with [35S]methionine, and,after lysis of the minicells by heating in SDS,labeled proteins were analyzed by SDS-poly-acrylamide gel electrophoresis and autoradiog-

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

710 JORGENSEN AND REZNIKOFF

raphy (see Materials and Methods). Minicellscontaining R100 or pRT3 synthesize a 36,000-dalton (36K) tetracycline-inducible polypeptide(see Fig. 5). In other experiments (data notshown), we have found that recombinant plas-mids pRT24, pRTl, pRT11, and pRT61 alsodirect the synthesis of the same inducible poly-peptide. Thus, the R100 TET protein of Levyand co-workers (19, 21) is indeed encoded byTnlO. (It should be noted here that the TETprotein encoded by these recombinant plasnidshas not always been observed in each SDS-polyacrylamide gel analysis. It may be that theinability to detect the TET protein is due to thefailure to solubilize it from its site in the minicellmembrane.)Plasmids pRT22 and pRT29 direct the syn-

V>>: pPT3 pRT4t i. ,4_.~~~~

thesis of a larger (37K) tetracycline-induciblepolypeptide than does R100; this 37K poly-peptide is probably unstable, being accompaniedin the gel by many other lower-molecular-weightpolypeptides whose synthesis is also induced bytetracycline (Fig. 5). pRT4, on the other hand,directs the synthesis ofa stable 34K polypeptide.From the known physical structures of theseplasmids (Fig. 3), it is possible to explain theoccurrence of these different-sized polypeptidesby concluding that a HpaI site lies within thestructural gene encoding the wild-type 36K poly-peptide. This gene (tet) must then lie mostlywithin HpaI fragment 2025 and cross one of theHpaI sites defining this fragment. pRT22 andpRT29, which encode inducible polypeptides ofthe same size, both possess the same fusion of

RIOO pPRT22 |pPT29 ' m-E-- +1--- + - +

sv.. *m

AK ,.

4MWnW . s8,,: .. ; .:.. :. : . : .: :*:.: .: ..... :. ,' ... : .:: .:_ -_tt:t w :S:.'.:'Ss:. :: ..

}. ... . .jjlw 8z{8.. '_sl_ J_.R

_E_'_|-F _-:

FIG. 5. Autoradiogram of 15% polyacrylamide-SDS gel displaying plasmid-encoded polypeptides labeledwith [MS]methionine in minicells. Exposure tirne was 155 h.-and + indicate whether or not miniceUs wereinduced with I jg of tetracycline per ml at the time of labeling.

J. BACTERIOL.

VP" mm:qw,:i:: :..:-

i.%.":%X: :..,:::. .4 1,.

,:. W .:

.1

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

GENE ORGANIZATION IN TnlO 711

fragment 2025 to pVH51 sequences, indicatingthat the HpaI site in the tet gene is the one thatlies between HpaI fragments 2025 and 675 (seeFig. 3). In pRT4, the end of fragment 2025defined by this site is joined to A DNA sequencesin fragment 3000 (Fig. 3), and so pRT4 encodesa different fusion polypeptide than do pRT22and pRT29. These plasmids also encode regu-latory functions for the induction of both resist-ance and the tet gene product; thus it wouldseem that the promoter region and the amino-terminus of the tet gene must be intact in plas-mids pRT4, pRT22, and pRT29. If this is thecase, the HpaI site in the tet gene lies in thecarboxy-terminus of the TET protein, and wecan conclude that the direction of tet gene tran-scription is left to right, as the TnlO map isdrawn in Fig. 3.Yang et al. (35) have reported the synthesis of

a 15K tetracycline-inducible polypeptide inR100-containing minicells. We have also ob-served this polypeptide in R100-containing min-icells, but not in minicells containing any of therecombinant plasmids described here. The pos-sibility thus exists that this protein is encodedby regions of R100 outside TnlO. We also ob-served several other tetracycline-inducible pol-ypeptides, synthesized by both R100 and pRTseries recombinant plasmid-containing mini-cells. These appear in the 15K-25K region of thegel but are present in relatively small amounts.These polypeptides may represent fragments ofthe 36K protein derived from mild proteolysis orpremature translation termination.RNA polymerase protection of an HincH

cleavage site in Tn10: possible location ofthe tet gene promoter. We have concludedabove that the start point of transcription of thetet gene lies in HpaI fragment 2025 because arecombinant plasmid (pRT29) carrying only thisfragment shows essentially normal regulation oftetracycline resistance and tet gene product syn-thesis. RNA polymerase protection of HincIIcleavage sites has been observed in several E.coli promoters, e.g., the trp operon promoter (5,13) and the bacteriophage A PL and PR pro-moters (1, 2, 23, 25, 34). As shown by the "pro-tection experiment" presented in Fig. 6, priorincubation of TnlO DNA with E. coli RNApolymerase inhibits HincII cleavage of theHincII site within HpaI fragment 2025, i.e.,HincII/HpaI fragment 2025 is incompletely di-gested to fragments 1275 and 695. Of the nineHincII sites within TnlO, only this site is pro-tected by RNA polymerase. Addition of the fourribonucleoside triphosphates to the RNA po-lymerase binding mixture prior to HincII treat-ment allows complete cleavage at this HincII

RNP- RNP XTP's

2025- 2 75

*- 695

FIG. 6. RNA polymerase "protection experiment."Columns are labeled as follows: -, no RNA polym-erase added before HincII; RNP, RNA polymeraseboundprior to HincII treatment; RNP + XTP's, RNApolymerase and ribonucleoside triphosphate addedprior to HincII. 695 and 1275 are HincII fragmentsfrom TnIO. 2025 is the fragment resulting from RNApolymerase protection of the HincII cleavage sitebetween fragments 1275 and 695.

site, indicating that, in vitro at least, this RNApolymerase binding site can also act as a site fortranscription initiation. Although the directionof transcription from this site is not yet known,the obvious suggestion is that this site corre-sponds to the tet promoter.

4 DISCUSSIONTnlO, which originates from the R-factor

R100, carries all the structural genes necessaryto confer the same level oftetracycline resistanceon E. coli cells as R100. This was demonstratedby the observation (Table 1) that a lysogen of aX::TnlO transducing phage is resistant to thesame levels of tetracycline as is an R100-contain-ing cell. Further, tetracycline resistance in bothA::TnlO- and R100-containing cells is inducibleby subinhibitory concentrations of tetracycline,indicating that regulatory elements required forthe induction process are also encoded by TnlO.

VOL. 138, 1979

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

712 JORGENSEN AND REZNIKOFF

Levy and co-workers (19, 21) have reported theR-factor-directed synthesis of a tetracycline-in-ducible polypeptide (TET protein) in E. coliminicells. Minicell analysis of a series of recom-binant plamids carrying various portions ofTnlO (Fig. 5) demonstrates that theTET proteinis, as might be expected, encoded by TnlO, asare its regulatory elements.The recombinant plasmids described in this

paper provide a set ofTnlO deletions with whichit is possible to study the role played by differentregions of TnlO in tetracycline resstance andthe regulation of its expression. By inspection ofthe physical structures of these plasmids, thestructural gene (tet) encoding the TET proteincan be shown to lie within the 2,700-base pair-long region of TnlO c-rried by plasnid pRT11(see Fig. 3). In addition, an Hp,a[ cleavage sitewithin this region lies in the tet gene; this wasdemonstrated by the observatioon that fusion ofdifferent sequences to this site results in altera-tions in the molecular weight and itability of thetet gene product (Fig. 5). In all cases these al-tered tet gene products are inducible. Thus, theregulatory elements are not affocted in the fu-sion strains and lie wholly within HpaI fragment2025, as do the promoter and amino-terminus ofthe tet gene. The HpaI site in tet then lies in thecarboxy-terminus of the gene, and the directionof tet transcription is from left to right as theTnlO physical map is drawn in Fig. 1 and 3.The concliWon that the gene for tetracycline

resistance resides largely within HpaI fragment2025 and more precisely within HincH fragment1275 is consistent with structural studies of theTnlO transposon carried out on plasmid R6-5 bySharp et al. (30). R6-5 is a tet plasmid in whichan insertion is located within TnlO approxi-mately 1,300 base pairs from one inverted repeat.Tetracycline-resistant derivatives of R6-5 ariseby loss of this insertion. This insertion musteither be in HindI-HincdI fragment 685 on the"left side" of the transposon, which our datashow not to be required for tetracycline resist-ance, or in HincH fragment 1275 (see restrictionmap in Fig. 1).

I

In Fig. 7 we present schematically a simplemodel for the location and organization of thetet gene and its regulatory elements. The loca-tion of the tet gene and direction of its transcrip-tion are shown as stated above. The location ofthe promoter is inferred from the RNA polym-erase "protection experiment" (Fig. 6). This ex-periment shows that an Hincd cleavage site inHpaI fragment 2025 lies in a region of in vitrotranscription initiation. Recent experiments (L.Wray, R. Jorgensen, and W. Reznikoff, unpub-lished data) have shown that removal of Hindcfragment 695 and fusion of other Hincd frag-ments to HincII fragment 1275 affects the levelof tetracycline resistance but not the size of thetet gene product. This result strongly suggeststhat the HincII cleavage site in HpaI fragment2025 lies in the tet gene promoter, with the resultthat fusions at this site lead to altered promoteractivity.Mapping the tet promoter at this site is also

consistent with the location of the start point oftet translation, which can be approximately de-termined from the size of the tet gene product.The carboxy-terminus of this protein has beenshown here to lie inHpaI fragment 675, probablyvery near (within 100 base pairs of) the HpaIsite, which divides fragments 2025 and 675 (as-su1ming that only a small portion of the carboxy-terminus of this protein can be replaced with anew amino acid sequence while maintaining theprotein's resistance function). The apparent mo-lecular weight of the tet gene product is 36,000(36K) in Tris-glycine-buffered SDS-polyacryl-amide gels (this paper and ref. 34) and 50,000(50K) in a phosphate-buffered SDS-polyacryl-amide gel (19). If the mass of one amino acid is120 daltons, a 36K protein contains 300 aminoacids and requires a 900-base pair-long codingsequence, while a 50K protein contains 420amino acids and requires a 1,260-base pair-longcoding sequence. With 100 base pairs as an es-timate of the length of tet extending into HpaIfragment 675, it follows that the amino terminusof the tet gene lies roughly 800 to 1,160 basepairs to the left of HpaI fragment 675 within

T'et' protein

I II II

tetinactive

repressorFIG. 7. Model for organization of tet gene and its regulatory elements. RNP is RNA polymerase, R is

repressor. I indicates HpaI cleavage sites in ThlO. Unlabeled vertical bars are HincII cleavage sites in TnhO.

eLb, -16.I lI I

J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

GENE ORGANIZATION IN TnIO 713

HpaI fiagment 2025. Although this is a widerange on the scale of the size of promoters, it iscertainly consistent with a promoter location atthe HincH site 1,275 base pairs left of HpaIfragment 675. Also constent is the finding thatHincII rapgents 695 and 1275 display no RNApolymerase binding activity (15; R. Jorgensen,Ph.D. Thesis, University ofWisconsin, Madison,1978); thus the "protected" Hinc]I site is theonly known in vitro candidate for the tet pro-moter in HpaI fragment 2025.

Franklin and Cook (8) and Yang et al. (35)have postulated that tetracycline resistance isnegatively regulated. Our results, regardless ofthe mode of regulation, map the regulatory ele-ments that control tetracycline resistance andTET protein synthesis in HpaI fragment 2025.The model in Fig. 7 suggests a more preciselocation (in HincH fragment 695) for the codingregion for the "repressor" of Yang et al. (35).This is consistent with the observation by Reeve(27) that a tet' insertion mutation ofR6 (plasmidR6-5) carries an active TET-repressor gene, al-though more complicated models such as theexistence of overlapping tet and repressor genesare not ruled out.Levy et al. (21) have suggested that the TET

protein is responsible for tetracycline resistance,based on observations that the amount of thisprotein synthesized in minicells seems to becorrelated with the level of tetracycline resist-ance conferred by the particular plasmid encod-ing it and that tetracycline-sensitive mutants ofR222 do not synthesize TET protein. Our obser-vations concerning the effect of TnlO deletionson tetracycline resistance and on the molecularweight and stability of the TET protein areentirely consistent with this view. We have iso-lated other fusions at the HpaI site in the tetgene through attempts to use pRT29 as an HpaIcloning vehicle for bacteriophage T7 HpaI frag-ments (while maintaining selection for tetracy-cline resistance) and for various HpaI fragmentswith certain selectable characters (e.g., amino-glycoside resistance) (N. Panayotatos, L. Wray,R. Jorgensen, and W. Reznikoff, unpublisheddata). These fusions also have altered tetracy-cline resistance levels, supporting a conclusionthat the TET protein is both necessary andsufficient for tetracycline resistance.Taylor et al. (33) have described a copy num-

ber mutant of NR1 called ROR12 with a mark-edly lower level of tetracycline resistance andsuggest that this phenotype is a direct result ofthe higher copy number of ROR12. Recombi-nant plasmid pRT61, a derivative of the multi-copy plasmid ColEl containing the entire TnlOtransposon, is resistant to only about one-half

the level of tetracycline as is a single-copy A::TnlO lysogen or R100 (Table 1). pRT61 appearsto be resitant to higher levels of tetracyclinethan has been reported for ROR12; however,these two plamsids have not been tested in thesame host strain. Taylor et al. (33) have furthersuggested that the multicopy effect on resistancein ROR12 is due to overproduction of the tetgene repressor by multicopy plasmids. Anotherpossibility is that some TnlO-encoded functioninteracts directly with the tet gene product. Ev-idence for the existence of such a TnlO-encodedfunction is provided by the observed differencesbetween class I and class II plasmids, such aspRT1 and pRT11. Although both of these plas-mids encode a wild-type tet gene product, pRT1is markedly less resistant to tetracycline than ispRT11. The only difference between these plas-mids is that HpaI fragment 3000 is not presentin pRT11. This suggests that fragment 3000encodes a function which contributes to thelowered resistance in multicopy plasmids.

ACKNOWLEDGMENTEWe thank S. Hardies, L Wray, N. Panayotatos, and P.

Kiberstis for helpful discuions and comments on the manu-script, as well as for permission to mention unpublished re-sults, and to the editor and reviewers of the Journal ofBacteriology for helpful comments on the manuscript.We are grateful for support from the National Institutes of

Health (Public Health Service grant GM-19670) to W.S.R.W.S.R. was also supported in part by Public Health ServiceCareer Development Award (GM-30970) from the NationalInstitutes ofHealth and is a recipient of the Harry and EvelynSteenbock Career Development Award. This work was alsosupported in part by Public Health Service training grantGM-00236-BCH from the National Institute of General Med-ical Sciences.

LITERATURE CITED1. Allet, B., R. J. Roberts, R. F. Gesteland, and R.

Solem. 1974. Class of promoter sites for Escherichiacoli DNA-dependent RNA polymerase. Nature (Lon-don) 249:217-221.

2. Allet, B., and R. Solem. 1974. Separation and analysis ofpromoter sites in bacteriophage lambda DNA by spe-cific endonucleases. J. Mol. Biol. 85:475-484.

3. Barnes, W. 1977. Plasmid detection and sizing in singlecolony lysates. Science 195:393-394.

4. Barnes, W. M., W. S. Reznikoff, F. R. Blattner, R. C.Dickson, and J. Abelson. 1975. The isolation of RNAhomologous to the genetic control elements of the lac-tose operon. J. Biol. Chem. 250:8184-8192.

5. Bennett, G., ML Schweingruber, K. Brown, C.Squires, and C. Yanofsky. 1976. Nucleotide sequenceof region preceding thp mRNA initiation site and itsrole in promoter and operator function. Proc. Natl.Acad. Sci. U.S.A. 73:2351-2355.

6. Berg, D., J. Davies, B. Allet, and J.-D. Rochaix. 1975.Transposition of R-factor genes to bacteriophage A.Proc. Natl. Acad. Sci. U.S.A. 72:3628-3632.

7. Cameron, J., R. Philippsen, and R. Davis. 1977. Anal-ysis of chromosomal integration and deletions of yeastplasmids. Nucleic Acids Res. 4:1429-1448.

8. Franklin, T. J., and J. M. Cook. 1971. R-Factor with a

VOL. 138, 1979

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.

714 JORGENSEN AND REZNIKOFF

mutation in the tetracycline resistance marker. Nature(London) 229:273-274.

9. Hardies, S., R. Patient, R. Klein, F. Ho, W. Reznikoff,and R. Wells. 1979. Construction and mapping ofrecombinant plasmids used for the preparation of DNAfragments conaining the E. coli lactose operator andpromoter. J. Biol. Chem., in press.

10. Hershfield, V., H. Boyer, L Chow, and D. Helindki.1976. Characterization of a mini-ColEl plasnid. J. Bac-teriol. 126:447453.

11. Humphreys, G. O., G. A. Wilishaw, and E. S. Ander-son. 1975. A simple method for the preparation of largequantities of pure plasmid DNA. Biochim. Biophys.Acta 383:457-463.

12. Izakl, K, K. Kiuchi, and K. Arima. 1966. Specificityand mechanism of tetracycline resistance in a multipledrug-resistant strain of Escherichia col. J. Bacteriol.91:628-633.

13. Jones, B. B., and W. S. Reznikoff. 1977. Tryptophan-transducing bacteriophages: in vitro studies with re-striction endonucleases HindII+III and Escherichiacoli ribonucleic acid polymerase. J. Bacteriol. 132:270-281.

14. Jorgensen, R., D. Berg, B. Allet, and W. Reznikoff.1979. A restriction endonuclease cleavage map for thetransposable tetracycline resistant detenninant TnIO.J. Bacteriol. 137:681-685.

15. Jorgensen, R., D. Berg, and W. S. Reznikoff. 1978.Genetic organization in the transposable tetracyclineresistance determinant TnlO, p. 181-183. In D. Schles-singer (ed.), Microbiology-1978. American Society forMicrobiology, Washington, D.C.

16. Kleckner, N., R. Chan, B.-K. Tye, and D. Botstein.1975. Mutagenesis by insertion of a drug-resistanceelement carrying an inverted repetition. J. Mol. Biol.97:561-575.

17. Laemmli, U. 1970. Cleavage of structural proteins duringthe assembly of the head of bacteriophage T4. Nature(London) 227:680-68.

18. Landy, A., J. Abelson, M. Goodman, and J. Smith.1967. Specific hybridization of tyrosine transfer ribo-nucleic acids with DNA from a transducing bacterio-phage 060 carrying the amber suppressor gene SuIII. J.Mol. Biol. 29:457471.

19. Levy, S., and L McMurry. 1974. Detection of an induc-ible membrane protein aociated with R-factor-me-diated tetracycline resistance. Biochem. Biophys. Res.Commun. 56:1060-1068.

20. Levy, S., and L McMurry. 1978. Probing the expressionof plasmid-mediated tetracycline resistance in Esche-richia coli, p. 177-180. In D. Schlessinger (ed.), Micro-biology-1978. American Society for Microbiology,Washington, D.C.

21. Levy, S.,L McMurry, P. Onigman, and R. Saunders.1978. Plasmid mediated tetracycline resistance in Esch-erichia coli, p. 181-206. In J. Drews and G. Hogenauer(ed.), R-factors: their properties and posible control.Springer-Verlag, New York.

22. Mandel, M, and A. Higa. 1970. Calcium-dependent bac-teriophage DNA infection. J. Mol. Biol. 43:159-162.

23. Maniatis, T., A. Jeffrey, and D. Kleid. 1975. Nucleotidesequence of the rightward operator of phage lambda.Proc. Natl. Acad. Sci. U.S.A. 73:1184-1186.

24. Maniatis, T., A. Jeffrey, and H. Van deSande. 1975.Chain length detenrination ofsmall double- and single-stranded DNA molecules by polyacrylamide gel electro-phoresis. Biochemistry 14:3787-3794.

25. Maniatls, T., M. Ptashne, B. G. Barrell, and J. Do-nelson. 1974. Sequence of a represor binding site inthe DNA of bacteriophage lambda. Nature (London)250:394-397.

26. MeMurry, L*, and S. Levy. 1978. Two transport systemsfor tetracycline in sensitive Escherichia coli: criticalrole for an initial rapid uptake system insensitive toenergy inhibitors. Antimicrob. Agents Chemother. 14:201-209.

27. Reeve, E. C. R. 1978. R plasmid R6-5 retains an activetet protein repressor gene. Plasmid 1:681-583.

28. Roosen, D., R. Fenwick, and R. Curti. 1971. Synthe-sis of ribonucleic acid and protein in plasmid-containingminicelLs of Escherichia coli K-12. J. Bacteriol. 107:21-33.

29. Sgaramella, V. 1972. Enzymatic oligomerization of bac-teriophage P22 DNA and of linear aniian virus 40 DNA.Proc. Natl. Acad. Sci. U.S.A. 69:3389-3393.

30. Sharp, P., S. Cohen, and N. Davidson. 1973. Electronmicrocope heteroduplex studies of sequence relationsamong plammids of Escherichia coli. II. Structure ofdrug resistance (R) factors and F factors. J. Mol. Biol.75:236-255.

31. Shinnlck, F. M., E. Lund, 0. Smithies, and F. R.Blattner. 1975. Hybridization of labeled RNA to DNAin agarose gels. Nucleic Acids Res. 2:1911-1929.

32. Tait, R., R. Rodriquez, and H. Boyer. 1977. Alteredtetracycline resistance in pSC101 recombinant plas-mids. Mol. Gen. Genet. 151:327-331.

33. Taylor, D., J. Greenberg, and R. Rownd. 1977. Gen-eration of miniplasmids from copy number mutants ofthe R plasmid NR1. J. Bacteriol. 132:986-995.

34. Walz, A., and V. Pirotta. 1975. Sequence of the P,,promoter of phage lambda. Nature (London) 254:118-121.

35. Yang, H., G. Zubay, and S. Levy. 1976. Synthesis of anR plasmid protein asociated with tetracycline resist-ance is negatively regulated. Proc. Nadl. Acad. Sci.U.S.A. 73:1509-1512.

J. BACTERIOL.

Dow

nloa

ded

from

http

s://j

ourn

als.

asm

.org

/jour

nal/j

b on

03

Febr

uary

202

2 by

14.

47.1

57.4

9.