IHF - The Journal of Biological Chemistry

Vol. 266. No. 35, Issue of December 15, pp. 24077-24083.1991 Printed in U.S. A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1991by The American Society for Biochemistry and Molecular Biology, Inc.

A Compact Nucleoprotein Structure Is Produced by Binding of Escherichia coli Integration Host Factor (IHF) tothe Replication Origin of Plasmid R6K* (Received for publication, May 8, 1991)

Marcin FilutowiczSQ andRoss Inmann From the $Department of Bacteriology and the Vnstitute for Molecular Virology and Department of Biochemistry, University of Wisconsin, Madison, Wisconsin 53706

Understanding the role of Escherichia coli histone- in vivo replication of plasmids pSClOl (14-16) and R6K (17, like protein integration host factor (IHF) in replication 18).Each of these replicons contains a self-encoded initiator of R6K plasmid (Dellis, S., and Filutowicz, M. (1991) protein, whose binding to specific repeats is essential for J. Bacteriol. 173, 1279-1286) requires detailed anal- replication initiation (for reviewseeRefs. 19 and 20). The yses of the interaction of IHF protein with the plas- propagation of phage fl has also been reported to require IHF mid’s replication origin (y ori).We describe an electron protein (21). microscopic analysis which shows thatacompact Footprinting of the R6K y ori in the presence of IHF using structure can be formed in the presence of IHF, in neocarzinostatin (NCS) has revealed that theprotein protects which, onaverage, a 102-base pair (bp)ori segment is two segments of the ori: site 1resides within an AT-rich block, involved. IHF.y ori complexes also undergo atwo-step whereas site 2 is separated from the first by a cluster of seven conformationalchangeinan IHF concentration-dependent manner when analysed by band shift assay. iterons (17). Furthermore, the minimal R6K ori ( y ori) fails We believe that theDNA is bent atlow IHF concentra- to function in mutants altered in either the CY or /3 subunit of tions, but folded at high IHF concentrations. This idea IHF (17). The most recentstudies (18) indicate that IHF is supported by the fact that electrophoretic mobility exertsits effect by counteracting the replication-inhibitor of the IHF.y ori complexes is faster at higher concen- activity of replication protein r (22). Thus we have proposed that IHF allows the R6K ori to maintain a replication-profitrationsof IHF. Furthermore, it is shownthatthe formation of acompactnucleoproteinstructure de- cient configuration a t higher levels of K initiator protein than pends on the two regions flanking the AT-rich seg- would be possible in the absence of IHF (18). In this study we have analyzed the configuration of ori ment; the iterons to the right and the 106-bp ori domain to theleft. Finally we show thatIHF protects the DNA. IHF complexes to determine if structural alterations entire AT-richsegmentofthe ori againstnuclease could underlie the in vivo IHF-mediated effect. A novel feature cleavage. In additiontotheprotection,analtered of IHF.DNA complexes has been discovered DNA folding. cleavage pattern by DNase I , in the presence of high The formation of this structure is influenced by both the IHF levels of IHF, was observed within the iteronsbut not concentration andthe composition of the DNA fragment used. within the 106-bp domain of the ori. Implications of We demonstrate that DNA folding results from the binding the IHF-mediatedy ori folding as a possible mechanismof IHF to the entireAT-rich ori segment and that itrequires protecting the ori from replication inhibition byR6K the iteron segment and the left-most 106-bp domain of the initiator protein u are discussed. ori sequence. EXPERIMENTALPROCEDURES

Strains and Plasmids-Strain C2110 thy polAl mutant has been described (23). Plasmid pMF32 was constructed by cloning the HindIII-EglII fragment (R6Kcoordinates +I to+278) into thepUC9 vector digested with Hind111 + BamHI. The genealogy of plasmid pMF34 was also described earlier (24). Plasmid pMF239 was produced in a two-step construction originating from mutant 7134 which was isolated and sequenced byM. McEachern (25) from a bisulfiteinduced mutant bank of plasmid pMM3 (26). 7134 contains a substitution at position 21 of the first 22-bp repeat, creating a new SnaBI site (shown as S n a E P in Fig. 4). By taking advantage of this and a second SnaBI site in the same 21 position but in the seventh repeat, it was possible to delete the intervening repeat segment. The remaining R6K sequences in this deletion derivative of pMM3, flanked by EcoRI(-106) and BglII(+277) restriction sites, were cloned into the * This work wassupported by National Institutes of Health Grants EcoRI and BamHI sitesof pUC9, yielding the plasmid pMF239. From GM 40314-01 (to M. F.) and GM 14711-23 (to R. I.). The costs of our cloning approach we conclude the only difference between pMF34 publication of this article were defrayed in part by the payment of and pMF239 is a deletion of six out of seven repeats in the latter page charges. This article must therefore be hereby marked “udver- construct. tisement” in accordance with 18U.S.C. Section 1734 solelyto indicate Electron Microscopy-Preparation of DNA. IHF complexes. Carthis fact. bon films were formed across electron microscope grids by vertical § T o whom correspondence should be addressed. Tel.: 608-262- evaporation of carbon onto freshly cleaved mica discs. The films were 6849; Fax: 608-262-9865. floated onto a clean air-water interface and picked up from below on I The abbreviations used are: IHF, integration host factor; bp, base 400-mesh copper grids; it was advantageous to increase the bond pair; NCS, neocarzinostatin; EM, electron microscopy. between carbon film and grid by first dipping grids into pressure-

Histone-like proteins are found in all bacteria. Some, like Escherichia coli HU protein (reviewed in Ref. l),bind DNA nonspecifically (2). Others, such as E. coli integration host factor (IHF)’ (reviewed in Ref. 3), and Bacillus subtilis phage SPOl TF1protein (4)bind specifically (5);in the lattercase, binding occurs only to DNA containing hydroxymethyluracil in place of thymine (4). DNA bending could be a common mechanism underlying the diverse array of biochemical processes mediated by IHF (6-10) and HU (11). IHF and/or HU are required for, or play an accessory role in, the in vitro replication of oriC (12) and phage X (13) and

24077

24078

IHF Interactions with y ori of Plasmid R6K

sensitive adhesive (laboratory supplies No. 8091, diluted to 0.2%, w/ v, with amyl acetate) followedby drying under a heat lamp. The carbon surface was next “activated” with polylysine (27). A 7-pl drop ofpoly-L-lysine hydrobromide (1 pg/ml, degree of polymerization 3400, from Sigma) was placed on the carbon grid for 30 s and then touched to a filter paper and left to dry. The complexes for EM analysis were assembled in 30-pl aliquots containing 0.5 pg ofdigested pMF34 and 0.4 pg of IHF (for EcoRI/SalI digest) or 1.25 pg of IHF (for either PstIIScaI or SnaBI/ScaI digests). Samples were subsequently diluted in 160 pl of TBE buffer (89 mM Tris base, 89 mM boric acid, 2 mM N&-EDTA and 40 p1 of glycerol and 7 pl of the resulting dilution placed on an activated carbon film for 5 min. During this andsubsequent steps, the grid was held in self-clamping forceps. The grid was then slowly swept twice across a clean surface of a mixture of 80 ml of 40 mM ammonium acetate and 20 ml of glycerol, followed by about 15 sweeps across a similar fresh solution. A 7-pl drop of staining solution (4 volumes of 5% uranyl acetate and 1 volume of glycerol) was then applied to the carbon surface for 30 s, and thedrop was then flicked awayand a further 7pl of stain applied for 30 s. The carbon surface was again swept across two series of ammonium acetate/glycerol solutions, as described above, and blotted against filter paper. Samples were then left in a vacuum for a t least 45 min before being rotary shadowed with platinum. The pressure torr during the pump down prior to shadowing was typically 5 X torr a t 45 min. Measurements from electron after 5 min and 3 X micrographs were as described (28). Polyacrylamide Gel Electrophoresis-CsC1-purified plasmid DNA was digested with the enzymes indicated in the respective figure legends and then phenol-treated, isopropanol-precipitated, and ethanol-washed. IHF. DNA complexes were assembled using IHF dilutions in TBE buffer, always prepared fresh from the 4 mg/ml protein stock. Total reaction volumes varied from 3 to 10 pl. The composition and ratio of tracking dye mix were as described previously (24). Electrophoresis was carried out at room temperature (10 volts/cm) in 5% polyacrylamide gels prepared at least 12 h before use. Thickness of most gels was0.5 mm and never exceeded 0.75 mm. Gels were stained with ethidium bromide and photographed using short wavelength UV source. Neocarzinostatin and DNase I Protection Assays-Neocarzinostatin cleavage was carried out as described (17). The conditions for DNase I protection experiments were identical to those described earlier (24) for x “footprinting.” Under these conditions the hypersensitivity to DNase I at coordinate 99 (17) was not observed. DNA sequencing (G-specific and G + A-specific)was performed as described (24). 5’-Ends of restriction fragments were labeled with polynucleotide kinase (Bethesda Research Laboratories) and [Y-~’P] ATP according to Maniatis et al. (29). IHF Purification-IHF was purified according to the published procedure (17). As indicated earlier (17) no differences between IHF protein purified according tothe Nash and Robertson (30) and Filutowicz and Appelt (17) were noted in either the Int-mediated recombination of X phage or in IHF protection of the y ori against NCS.

FIG. 1. Electron micrographs of IHF-DNA complexes prepared from pMF34 cleaved with EcoRI + SalI and processed as described under “Experimental Procedures.” Inset, drawing of a bentmolecule with a blob located at thebend position. The DNA section (a)-(b)-(c) is clearly visible in the electron microscope, but DNA within the blob cannot be resolved. Point X is defined as the bend position.

RESULTS AND DISCUSSION

Bending and Folding of y ori DNA as Revealed by EMMultiple binding sites for IHF identified in the R6K y ori were described previously (17). The ihfl site contains two adjacent IHF consensus sequences in an AT-rich context which are designated a and p (formerly IHF sites I and 11, Ref. 17). The ihf2 site resides in a non-AT-rich segment, and its structurewill be discussed in more detail elsewhere? The last R6K y ori IHF sequence, ihf3, identified in this study, is located between R6K coordinates -5 to +8 (TAAAAGCTTT A M ) and has a2-bp mismatch from the consensus (underlined). EM analysis of ori-IHF complexes was carried out toidentify any unusual structurespromoted by IHF and tomap the position(s) of DNA bending. For these analyses the plasmid pMF34 was cleaved with EcoRI and SalI to produce two DNA fragments: the pUC9 vector and the entire R6K y ori from coordinates -106 to + 278. This small fragment also contains M. Filutowicz and R. Inman, manuscript in preparation.

300

350

400

450

Base pairs FIG. 2. Total lenxzth of bent and normal fragments. The hatched area represents the length distribution of 174normal fragments, whereas the nonhatched area corresponds to 127 bent or folded fragments (as shown in Fig. 1).The histograms have been scaled so that the average length of normal fragments corresponds to their known length of 386bp. The average length of the bent species (using path (a)-(b)-(X)-(b)-(c); see inset ofFig. 1) is 381 k 18 bp and represents the results of two experiments in which IHF protein was present a t 0.4 and 0.04 pg of protein/pg DNA (the two independent averages were 384 ? 17 bp and 378 & 22 bp, respectively).

12 bp of pUC9 sequences of the vector which separate the BurnHI and SalI sites.When complexes were prepared for EM, the short fragments containing the IHF binding sites often showed sharp bends and folds close to the center of each molecule (Fig. l),whereas in the absence of IHF these structures were almost always absent. The IHF-induced struc-

IHF Interactions with y ori of Plasmid R6K

24079

tures arefragile and are notfound if samples are treatedwith either cross-linking agents or ethyl alcohol as part of the drying procedure. The presence of 20% glycerol throughout the procedure, although not essential, was found to lead to more reproducible results, presumably because of the very high viscosityenvironment at thetime of drying. The proportion of bent molecules wassomewhat variable, but on average, at an IHF concentration of 0.4 pg/pg of DNA, about 20% of the fragments were bent, compared with 2% in the absence of IHF. The micrographs in Fig. 1 show that often a small blob is observed at thebend position and that this can be of variable size. The inset of this figure diagrams such a molecule, where the oual shape represents the blob. If length measurements are made on such molecules using the path (a)-(b)-(c), then molecules with large blobs have total DNA lengths that are significantly shorter than the length of unbent fragments. If a measurement path isused that includes the dashed line ( ( b ) (X)-(b))shown in the Fig. 1,inset, then total lengths are not significantly different from those of unbent fragments (Fig. 2). We take this to mean that, to a first approximation, the blob contains DNA in a folded configuration (see Fig. 1,inset). The bend position is defined by point X in this inset. Measurements on a large number of molecules shows that

scp (C)

’

Pot

0

200

400

600

FIG.4. Electron micrographs of the IHF.DNA complexes prepared from pMF34 cleaved with either SnaBIIScaI (top as panel) or PstIIScaI (bottom two panels) and processed described under “Experimental Procedures.” These three micrographs are shown a t a lower magnification than those in Fig. l.

8001200 1000

(a)

FIG. 3. Distribution of bend and fold positions determined I. a, by EM and area of protection by IHF against DNase expected error for EM data. For illustrative purposes the error bar is centered over the average bend position shown in b, but it represents the expected error for all features shown in b and c. b, bend position and extent of folding in 17 ScaIISnaBI fragments. c, bend position and extent of folding in 28 ScaIIPstI fragments. The thick burs and the open rectangles in b and c indicate the average bend and folded regions, respectively. d, map of the EcoRIISalI fragment of pMF34, showing IHF sites, AT-rich region, iterons, and restriction sites. e, bend position in 70 EcoRIISalI fragments. The thick bar indicates the most frequent bend position (as defined by X in the inset of Fig. l), whereas the open rectangle shows the actual range over which bends occurred. The distribution was highly skewed with a gradual decrease to the left and a sharp cut off to the right. folded regions in 64 EcoRI/SalI fragments. The solid rectangle represents the average ) the inset of Fig. l), whereas folded region (as defined by ( b ) - ( X ) - ( b in the open part of rectangle represents the largest observed folded areas. g, area protected by IHF against DNase I is shown as the hatched rectangle (see Fig. 6 for details).

the average bend position occurs 180 bp from one end of the 386-bp fragment. Due to the almost central position of the ihfl site within this fragment (Fig. 3d) and because of experimental error (Fig. 3a), itis not possible to definitively assign a polarity to this fragment (Fig. 3, e and f, shows the more likely polarity, in which the bend corresponds most closely to ihfl site). The distribution of bend positions is also shown in Fig. 3e, and there is a gradual decrease to theleft and a sharp cut off to theright. The extent and position of the folded area is shown in Fig. 3f, and the average amount of DNA present within the folded region contains 90-180 bp of DNA and occurs close to the ihflsite. A similar phenomenon was observedfor molecules obtained with either PstIIScaIor SnaBI/ScaI which producefragments (Fig. 4) in which the ihfl siteis now close to one end (Fig. 3, b and c). These fragments correspond to 1189 and 1139 bp, respectively. From the electron microscopic appearance of these structures and the measurements, we conclude that IHF protein molecules can fold a portion of the DNA and thereby produce a sharp bend. The observed folding does not depend on the ihf2 site, because folding is observed in its absence (Figs. 3b and Fig. 4). A Specific ori Segment Facilitates the Enhanced Electrophoretic Mobilityof IHF. DNA Complexes in the Presence of High Levels of IHF-Despite multiple IHF-binding sites the formation of IHF.ori complexes, over a wide range of protein concentration, resulted in only a one-step band shift when analyzed on agarose gels (17). Thus, we speculated that IHF binding to oricould be highlycooperative (17). Alternatively, multiple species could have been formed but were unresolved

IHF Interactions withy ori of Plasmid R6K

24080 -1 06

+I

+270

EmRl

I I I

I I I

I

IY//& I

-

II

I/////////////////I/////I/// I I/ I /////, I I I mz//5 IV

FIG.5. Conformational changes of the central segment of the R6K replication region as determined by polyacrylamide band shift assay. top Theofthe figure shows the segment of R6K replication region relevant to this study. Hatched boxes indicate positions of consensus IHF binding sites ihfl-ihf2. The genetic composition of the different restriction fragments obtained from plasmid pMF34 or pMF239 is outlined by horizontal bars, deleted segments are indicated by broken lines. Fragments designated with Roman numerals I and II were formed by cleavage of pMF34 with a combination of either EcoRI + SalI or EcoRI + SnaBI. Fragment I11 was formed by cleavage of pMF239 (see “Experimental Procedures”) with a combination of EcoRI + SalI. The Roman numerals which correspond to different restriction fragments are used to mark DNA bands resolved on polyacrylamide gels. Fragments If to IVf indicate the position of free DNA; IC toIVc, IHF-DNA complexes; and Vo, vector sequences. Multiple IHF-DNA complexes are bracketed. Band shift assays shown in A and D were carried out with 0.5 pg of the entire plasmid digest; assays shown in B and C were carried out with 0.05 pg of the purified fragments($. Samples in lanes a contain no I H F those in lanes b-e were incubated with 0.04, 0.08, 0.16, and 0.32 pg of IHF, respectively. In A, the highest concentration of IHF was 0.16 pg. All samples were processed as described under “Experimental Procedures.” Coordinates for R6K restriction sites are: EcoRI, -106; HindIII, +l;SnuBI*, +111; SnaBI, 243; BglII, 277 (43). Fragments with a SalI terminus contain 9 bp of non-R6K sequence derived from the vector.

due to thesmall extent of retardation of the complexes in the agarose matrix. Therefore, all subsequent band shift assays were carried out in polyacrylamide gels which considerably enhance the difference in migration of free DNA versus DNAprotein complexes (31). We carried out a series of gel retardation assays with a set of DNA fragments designated I, 11,111, and IV which represent the minimal R6K y ori and the y ori with either ihf2, six of the seven iterons, or both ihf2 and ihf3 deleted, respectively (Fig. 5). The pattern of retardation with fragment I clearly differs from the pattern observed with fragments I11 or IV. As shown inFig. 5, fragmentI forms two species when analyzed individually (B) or when mixed in equimolar amounts with fragment I11 (A). This mixed experiment was employed to allow different DNA fragments to compete for IHF from the same pool. The unique feature of fragment I is that the species formed at high IHF concentrations migrated more quickly than thatformed at low IHF concentration. The formation of the structure responsible for enhanced electrophoretic mobility of fragment I at high IHF concentration is abolished by the deletion of six repeats (A and C). We reasoned that enhanced mobility could result from increased compactness resulting from interaction between IHF promoters bound to ihfl and ihf2. Such interactionwould require DNA looping, which would be diminished if less DNA was available for such a loop (as was observed with fragment 111). We tested this model by asking whether a fast migrating band would appear if ihf2 were removed. Such a fragment, designated 11, was generated by EcoRI and SnaBIcleavage of

plasmid pMF34. Fig. 5 0 indicates that a distinct fast-migrating band is gradually generated as more IHF is added to fragment 11. Thus we conclude that ihf2 does not contribute to theformation of a more compact structure. Recall that the ihf2 site was dispensable for the DNA folding as determined by EM (Fig. 4). In the subsequent experiments we determined if ihf3 and/ or a 106-bp left-most domain of the y ori is/are required for the formation of the compact structure. As shown in Fig. 5E, in which the HindIIIISnaBI fragment (IV) was used, only a single species was observed over a wide range of IHF concentration. From these experiments we conclude that the formation of the compact structure requires two segments flanking the AT-rich block of the R6K ori, but notihf2. These two segments most likely fold around more than one IHF protomer. Interestingly, another HU-like protein, TF1, mentioned earlier (4), forms nested complexes around binding sites in hydroxymethyluracil-containing DNA. The resulting complexes can be resolved on native polyacrylamide gels (31); withincertain DNA contexts, complexes with more TF1 protein bound migrate faster than those containing less? At High IHF Concentrations the Entire AT-rich y ori Segment Is Protected against Nuclease Cleavage-Thefolded DNA segment utilizes a surprisingly large part of the y ori DNA, and its formation does not require ihf2 site, as determined from EM analysis and from the gel mobility shift assays. We reasoned that itis rather unlikely that binding of a single IHF protomer could be responsible for such a dramatic

’P.Geiduschek and G . Schneider, personal communication.

IHF Interactions with y ori of Plasmid R6K effect. Thus, we considered the possibility that IHF binds to more then one site in theAT-rich segment at a suitable ratio of protein to DNA. The failure to detect these interactions in the earlier studies (17) could have been due to inadequate electrophoretic conditions for the detection of a large area of protection. NCS-induced cleavage products, correspondingto the left-most part of the AT-rich segment,would have run off such an acrylamide gel. This scenario is supported by the fact that the ori can adopt a more compact structure only if the R6K sequence between coordinates -106 to +1 is provided. In addition, the R6K segment between coordinates -106 to +1, which is required for these IHF-DNA interactions, was not present in previous analyses (17). Totestthese two possibilities we have examined IHF footprints with both a truncated ori fragment, which was electrophoresed for considerably shorter time, and a complete ori segment. IHF binding patterns for fragments labeled at the 5'-end of either the HindIII site (coordinate +1)or EcoRI site (coordinate -106) were determined by NCS and DNase I footprinting, and the results obtained are displayed in Fig. 6. At a low IHF to DNA ratio (Fig. 6A) an IHF footprint between coordinates +61 and +88 can beseen. It encompasses the sequences designated ihf 1a and p. However, an additional 2-fold increase in the IHF concentration resulted in a footprint extending to the left of coordinate +61. Although the extent of protection is weaker in comparison with ihfl and the boundaries of the IHFfootprint inthis case are somewhat less defined, it appears that a specific IHF interaction to the left of the ihfldoes occur since several NCS cleavage products at the bottom of the gel, beyond the area protected by IHF, remain equally pronounced in all samples analyzed. These data suggest that IHF is able to interact with a segment of the ori DNA considerably larger than previously claimed. Figs. 6, B-E, represent footprinting analyses carried out with the extended ori cleaved with either NCS or DNase I, respectively. In these cases there is no ambiguity as towhether or not IHF interacts with the sequences within the AT-rich segment extending beyond ihfl. Decreased cleavage frequencies by NCS and DNase I between coordinates -5 to +90 can be clearly seen. Therefore, the areastrongly protected by IHF is at least 95 bp long and approximately three times the size reported earlier. It is unlikely that IHFbinding to ihf3, which contains two mismatches from the consensus IHF recognition site (TAAAAGCTTTA_AA,with bases differing from that of the consensus sequence (33) underlined) plays any role in the extended "footprint" because only 5 bp are protected beyond the HindIII site. Thus, nuclease protection data reported here demonstrated the ability of IHF protein to interact with the AT-rich domain of the y ori. As with EM and the band shift assays, the presence of ihf2 on a fragment does not appear to make any difference in the extent of area protected by IHF (compare Fig. 6C (+ihf2)with Fig. 6D (-ihf2)). Detailed comparison of the DNase I digestion pattern of DNA with and without IHF shows a region between +90 and the endof a fragment (+278) where the DNase I cutting patternsdiffer considerably (Fig. 7). Thusit is clear thatthe DNaseI cleavage pattern is altered not only within the area of the ATrich domain but also within the interons themselves (Fig. 7). Importantly, a comparison of the DNase I cleavage pattern of the fragment lacking the R6K sequences -106 to +1 shows identical cutting patterns within the iterons withand without IHF protein (Fig. 6F). It can be concluded therefore, that although protection of the AT-rich domain of the ori can occur without the 106-bp fragment, this fragment, or a t least part of it, is required to produce an altered cutting pattern within the iterons in samples containing IHF.

24081

Clearly in situ footprinting experiments will be required to determine the limits of DNase I protectionand cuttingin the two electrophoretically distinct nucleoprotein species observed in polyacrylamide gels. Nevertheless, it is likely that complexes analyzed by footprinting and gel retardation are similar despite the fact that theywere formed under different buffer conditions. If this is true, then the failure to observe the fastmigrating band with the fragment containingdeleted iterons (Fig. 5C) can be explained by a direct role of the iterons in DNA folding. Furthermore, it is reasonable to assume that thelarge area of protection as well as ori folding, require binding of more then one IHF protomer per DNA molecule since both features depend on IHF concentration. Finally, there is a good correlation between EM and nuclease protection data regarding the domain of the ori which is able to interactwith IHF (compare diagrams e, f , and g in Fig. 3). Stoichiometry and chemical protection studies willnowbe required to better understand the structure of IHF complexes with different domains of the y ori DNA. Comparison of these structures to the tentative model available for IHF complexed with the attBsites of phage X (34) will then be possible. What is the functional meaning of our in vitro observations? To answer this question one has to take into account A

D

C

B 0

E

F

0

?,::: z 0 : q o :

30:

...-..

I'?

+EO+

.89 I

+40+

+67 I il

+

:" I

*54 I

-*.. -40+

-"-

-

=&a

I-m-

&

mr

f=.S=..

a-

85

a:

-::

add--

""

.to4

L

I- 4, 1 . 8" --- -._ I-

-;t

-mz

.

a-

44-10)

:

-EO*

a-

FIG.6. IHF protection of the y ori fragments, generated from plasmid pMF34, analyzedby NCS and DNase I sensitivity. Two ng of ""P-labeled HindIII/SalI fragment were used in the NCS protection assay shown inA. Five ng of the '"P-labeled EcoRI/ Sal1 fragment were used inNCS ( R ) andDNase I ( C and D ) protection assays. Five ng of '*P-labeled EcoRI/SnaBI fragment were used in DNase I protection assay shown in E. Five ng of "2P-labeled HindIIIISalI fragment were used in DNaseI protection assay shown in F.Brackets indicate segmentsof DNA protected by IHF. Protection assays with NCS and DNase I were carried out in 100 and 20 pl of the appropriate buffers, respectively (see "Materials and Methods"). Inset (bottom left) shows sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of 4 pg of IHF protein electrophoresed on 15%polyacrylamide and stained withCoomassie Blue.The two bands correspond to a (top) and p (bottom) subunits of IHF protomer (42).

IHF Interactions with y ori of Plasmid R6K

24082 A

39 and 40 and R. Durland and M. Filutowicz, unpublished observations discussed in Ref. 41). Thus we propose that the domain can be required only if T is presentat inhibitory levels and that theability of IHF tofold the R6K y ori could bethe principal mechanism responsible for the ability of IHF to counteract the replication inhibitor activity of A protein observed in vivo. Finally, it is important to mention that therole of the ihf2 binding site inan R6K biologyis notunderstood at thepresent time. Nevertheless, we have recently learned that mutants of this site, which completely eliminate IHF binding in vitro, have undetectable effect on the ability of a y ori replicon to function in IHF-proficient bacteria producing a wide range of ?r protein levels?

A

0

Acknowledgments-We thank Rick Gourse and Bill Reznikoff for critical comments on the manuscript. We also thank Maria Schnos and David Inman for assistance in electron microscopy. The help of Stephanie Dellis, Frank Wu, and Dona York in preparation of the manuscript is greatly appreciated. FIG. 7. Densitometric analysisof p a r t of the autoradiogram showing the iteron segment (bottom).”P-Labeled EcoRI/SnaBI fragments werecleavedby DNase I either in the absence (A) or present ( B ) of IHF. The arrows indicate positions of a decreased frequency of DNase I cleavage in samples containing IHF.

the known in vivo role of IHF in the replication of plasmid R6K. We have recently proposed a model in which IHF counteracts the replication inhibitor activity of A protein (18). The model implies that IHF and A are antagonistic in some way. The relevant inhibited complex might be one in which both A and IHF are bound or one in which only A is bound. Obviously a complex containing IHF would beexpected to be inert, since A is required for initiationin addition to its inhibitory activity (35). The molecular mechanism underlying the inhibition of replication by A is not known. Nevertheless studies involving R6K (36) and RK2 (37), another iteroncontaining plasmid, suggest that replication control might involve trans interactions between sets of iterons brought about by a respective initiator protein (“handcuffing,” Ref. 36). IHF can, at sufficiently high levels, fold largesegment of y ori DNA which encompases some of the iterons in addition to the AT-rich domain of the ori. A might be able to bind to the resulting structure in a manner which would allow replication initiation to occur, but would prohibit, due to a sterical hindrance, the formation of the “handcuffed molecules.’’ Future experiments will address whether IHF alters the c k and trans interactions known to occur between sets of iterons in vitro (Refs. 36 and 38). In this investigation we show that the sequence to the left of theHindIII site, referred to as the 106-bp domain, is required for IHF-mediated DNA folding. Most recent results indicate that the domain is dispensable for R6K replication in uiuo under the conditions which permit plasmid replication in theabsence of IHF.4 These two conditions include reduced intracellular concentration of A protein or normal levels of mutated variants of A, which allow initiation of DNA replication to occur at much higher frequency (18). The insight gained from the results reported in this paper also shed light on some puzzlingobservations regarding the sequence requirements for a minimal R6K replicon. It had been originally demonstrated by Kolter and Helinski (39) that the -106 to +1 fragment could not be deleted without loss of y ori activity. This conclusion has been challenged by several observations indicating the dispensability, under certain circumstances, of the 106-bp R6K sequence to theleft of the HindIII site (Refs.

‘F. Wu, I. Goldberg and M. Filutowicz, manuscript

in preparation.

REFERENCES 1. Drlica, K., and Rouviere-Yaniv, J. (1987)Microbwl.Reu. 51, 301-319 2. Rouviere-Yaniv, J., and Gros, F. (1975)Proc. Natl. Acad Sci. U. S. A. 72,3428-3432 3. Friedman, D. I. (1988)Cell 55,545-554 4. Johnson, G.G., and Geiduschek, E. P.(1977) Biochemistry 16, 1473-1485 5. Craig, N. L., and Nash, H. A. (1984)Cell 39, 707-716 6. Prentki, M., Chandler, M., and Galas, D. J. (1987)EMBO J. 6, 2419-2487 7. Robertson, C.A., and Nash, H. A. (1988)J. Biol. Chem. 263, 3554-3557 8. Thompson, J. F., and Landy, A. (1988)Nucleic Acids Res. 16, 9681-9705 9. Kosturko, L. D., Daub, E., and Murialdo, H. (1989)Nucleic Acids Res. 17,317-334 10. Kur, J., Hasan, N., and Szybalski, W. (1989)Gene (Amst.) 81, 1-15 11. Hodges-Garcia, Y.,Hagerman, P. J., and Pettijohn, D. E. (1989) J. Biol. Chem. 264,14621-14623 12. Bramhill, D., and Kornberg, A. (1988)Cell 52,743-755 13. Mensa-Wilmot, K., Carroll, K., and McMacken, R. (1989)EMBO J. 8,2393-2402 14. Gamas, P., Burger, A.C., Churchward, G., Caro, L., Galas, D., and Chandler, M. (1986)Mol. & Gen. Genet. 204,85-89 15. Stenzel, T. T., Patel, P., and Bastia, D. (1987)Cell 49, 709-717 16. Biek, D. P., and Cohen, S. N. (1989)J. Bacteriol. 171, 20562065 17. Filutowicz, M., and Appelt, K. (1988)NucleicAcidsRes. 16, 3829-3843 18. Dellis, S., and Filutowicz, M. (1991)J. Bacteriol. 173,1279-1286 19. Filutowicz, M., McEachern, M. J., Mukhopadhyay, P., Greener, A., Yang, S. L., and Helinski, D. R. (1987)J. Cell Sci. Suppl. 7,15-31 20. Bramhill, D., and Kornberg, A. (1988)Cell 54,915-918 21. Greenstein, D., Zinder, N. D., and Horiuchi, K. (1988)Proc. Natl. Acad. Sei. U. S. A. 85,6262-6266 22. Filutowicz, M., McEachern, M. J., and Helinski, D.R. (1986) Proc. Natl. Acad. Sei. U. S. A. 83,9645-9649 23. Kolter, R., and Helinski, D. R. (1978)J. Mol. Biol. 124,425-441 24. Filutowicz, M., Uhlenhopp, E., and Helinski, D. R. (1986)J. Mol. Biol. 187,225-239 25. McEachern, M. J. (1987)Dissertation, University of California, San Diego, La Jolla, CA 26. McEachern, M. J., Filutowicz, M., and Helinski, D.R. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 1480-1484 27. Williams, R.C. (1977)Proc. Natl. Acad. Sei. U. S. A. 74, 23112315 28. Littlewood, R. K., and Inman, R.B. (1982)Nucleic Acids Res. 10,1691-1706 29. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982)Molecular

S. Dellis, T. Schatz and M. Filutowicz, unpublished observation.

IHF Interactions with y ori of Plasmid R6K 30. 31. 32. 33. 34. 35. 36.

C1oning:ALaboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Nash, H. A., and Robertson, C. A. (1981) J. Biol. Chem. 2 5 6 , 9246-9253 Wu, H.-W., and Crothers, D. (1984) Nature 308,509-513 Greene, J. R., Morrissey, L. M., Foster, L. M., and Geiduschek, E. P. (1986) J. Eiol. Chem. 4 6 1 , 12820-12827 Leong, J. M., Nunes-Duby, S., Lesser, C.F., Youderian, P., Susskind, M., and Landy, A. (1985) J. Biol. Chem. 260,44684477 Yang, C. C., and Nash, H.A. (1989) Cell 67,869-880 Inuzuka, M., and Helinski, D. R. (1978) Proc. Natl. Acad. Sci. U. S. A. 76,5381-5385 McEachern, M. J., Bott, M. A., Tooker, P. A., and Helinski, D. R. (1989) Proc. Natl. Acad. Sci. U.S. A. 8 6 , 7942-7946

24083

37. Lewis-Kittell, B., and Helinski, D. R. (1991) Proc. Nutl. Acad. Sei. U. S. A. 88,1389-1393 38. Mukherjee, S., Erickson, H., and Bastia, D. (1988) Cell 6 2 , 375383 39. Kolter, R., and Helinski, D. R. (1982) J. Mol. Biol. 161,45-56 40. Patel, I., and Bastia, D. (1986) Cell 4 7 , 785-792 41. McEachern, M. J., Filutowicz, M., Yang, S., Greener, A., Mukhopadhyay, P., and Helinski, D. R. (1987) in Banbury Conference on Evolution and Environmental Spread of Antibiotic Resistance Genes (Levy, S., and Novick, R., eds) Harbor Cold Spring Laboratory, Cold Spring Harbor, NY 42. Nash, H. A., Robertson, C. A., Flamm, E., Weisberg, R. A., and Miller, H. I. (1987) J. Bacteriol. 169,4124-4127 43. Stalker, D. M., Kolter, R., and Helinski, D. R. (1979) Proc. Natl. Acad. Sci. U. S. A. 76,1150-1154