k.) 1992 Oxford University Press
Nucleic Acids Research, Vol. 20, No. 15 4077-4081
The mechanism of trans-activation of the Escherichia coli operon thrU(tufB) by the protein FIS. A model Hans Verbeek, Lars Nilsson1 and Leendert Bosch* Department of Biochemistry, Leiden University, Gorlaeus Laboratories, PO Box 9502, 2300 RA Leiden, The Netherlands and 1Department of Cell Biology, University of Stockholm, S-106 91 Stockholm, Sweden Received March 3, 1992; Revised and Accepted July 1, 1992
ABSTRACT Transcription of the thrU(tufB) operon is trans-activated by the protein FIS which binds to the promoter upstream activator sequence (UAS). Deletions of parts of the UAS and insertions show that optimal transactivation requires occupation by FIS of the two FISbinding regions on the UAS and specific helical positioning of these regions. On the basis of these and other data, a model for the mechanism of thrU(tufB) trans-activation by FIS is proposed. This model implies that the mechanisms underlying stimulation by FIS of two totally different processes: inversion of viral DNA segments and transcription of stable RNA operons, are essentially the same. INTRODUCTION The thrU(tufl) operon of E. coli is a mixed operon, harbouring four different tRNA genes and the protein gene tuJB, encoding the polypeptide chain elongation factor EF-TuB (Fig. IA). Transcription of this operon is activated in trans by the E. coli protein FIS, which binds to a cis-acting region upstream of the promoter (designated UAS for upstream activator sequence) [ 1, 2, 3 and 4]. Similar activator sequences are also found upstream of the tyrT [5,6] and the metY [7] operons, and upstream of the P1 promoter of the rrnB operon [8]. They all specifically bind FIS [1], suggesting that this is the common trans-activator of these four operons and possibly more stable RNA operons [7]. FIS-dependent trans-activation of the rrnB operon has subsequently been demonstrated [9]. FIS (Factor for Inversion Stimulation) is a heat-stable protein of 98 amino acids, until recently only known for its role in the replication of certain bacterial viruses [10-12]. Some of these viruses alter their host range by switching the orientation of a viral DNA segment, which results in the expression of a different set of tail fiber genes [13,14]. This site-specific recombination event, mediated by a virus encoded recombinase, is stimulated by the host protein FIS which binds to a cis-acting recombinational enhancer [15,16] (Fig. 2).
*
To whom correspondence should be addressed
For some time it was not clear how E. coli benefits from FIS, a protein which enables the phage to attack bacteria. Recent studies [4] revealed that FIS plays a prominant role in accelerating bacterial growth. During a normal growth cycle early log phase cells show a large increase in FIS-dependent trans-activation of the thrU(tuJB) operon. Concomitantly, a peak of the cellular FIS concentration is seen. Cells, growing slowly in a poor medium, do not show such a rise and fall of FIS, nor a peak of operon transcription in the early log phase. After addition of nutrients to the medium these cells promptly respond with an increase both of transcriptional activity and of the FIS concentration. This increase gets higher as the medium gets more strongly enriched. Apparently, the synthesis of FIS responds to environmental signals, thereby affecting transcription of the thrU(tuJB) operon, and most likely more stable RNA operons. Furthermore, FIS allows cells to achieve very high growth rates. What is the mechanism underlying trans-activation of the thrU(tuJB) operon? The evidence obtained so far [17,3,18] and the data of the present paper show that bending of the UAS DNA double helix is essential for transcription activation. The UASs of various stable RNA operons are AT-rich [19], indicative of an unusual physical conformation, such as kinking or bending [20,21,8]. Accordingly, gel electrophoretic migration of UAS DNA fragments is retarded [1-3]. This retardation increases upon FIS binding, indicating that the nucleotide sequence-induced bending of the DNA is enhanced by FIS binding. Three FIS/DNA complexes are formed under these in vitro conditions. Cells unable to produce FIS (fis cells) do show cis-activation of thrU(tuJB) transcription. Most likely nucleotide sequence-induced bending plays a role here. FIS-directed bending (in wild-type cells) further increases transcription activation. DNase I footprinting [3,2] has delimited two FIS binding regions on the UAS of the thrU(tufB) operon (see bars in Fig. 3). Electrophoretic analysis of circularly permuted UAS DNA fragments located a bending centre around position -95 (with respect to the transcription start site), in between the two FISbinding regions [ 18]. Insertions of small DNA segments, comprising less than one or two complete helix turns at the
4078 Nucleic Acids Research, Vol. 20, No. 15 Table I. Effects of altering the helical positioning of the FIS-binding regions
A
galactokinase activity MC 1000 cells MCIOOO-fis767 cells
UU .
B
-j7,,,j-}
Fig. 1. A; The thrU(tuJB) operon of E. coli. B; The fusion operon thrU (tuiB) ':galK. Translation stop codons in three reading frames between the junction point and galK prevent translational readthrough from the cloned tuIB fragment into galK.
54 ± 2 21 ± 2 29 ± 15
Cells were grown in LB medium to an OD600 of 0.5 and galactokinase activities, expressed as activity/femtomole of plasmid, were measured as in [4]. Galactokinase activities are expressed as in Fig. 3. The insertion and the deletion alter the UAS sequence as follows: -130
wt.
recombinational enhancer
100 ± 8 12 + 3 24 + 6
no insertion one base pair deletion at -95 three base pair insertion at -95
-100
-i10
-120
-90
-80
-70
TGTTGAAAGTGTGCTAATCTGCCCTCCG.TTC3-GCTGTTTCTTCATCGTGTCGC&TAAAATG
invertible segment
{ }I-//KZJ
-lbp.
TGTTGAAAAGTGTGCTAATCTGCCCTCCGTTCGGCTccggatccggGTGTCGCATAAAATG
+3bp.
TGAAAAAGTGTGCTAATCTGCCCTCCGTTC^GGTccggatccggCATCGTGTCnCATAAAATG
FIS binding site
ED
crossover
site
Fig. 2. Invertible viral DNA segment and recombinational enhancer.
junction of the UAS and the RNA polymerase binding site (position -45), reduce transcription significantly. Partial restoration of transcriptional activity occurs when one or more full helix turns are inserted [ 18]. These data clearly demonstrate that an abnormal conformation of the UAS DNA, most likely a bent DNA double helix, is an essential requisite for transcription activation of the operon. Furthermore, that activation requires a specific spatial orientation of the DNA bend with respect to the RNA polymerase (RNAP). In the present investigation we show that optimal transactivation of the thrU(tuJB) operon not only requires occupation by FIS of both FIS binding regions on the UAS, but also specific helical positioning of these regions. These requirements are very similar to those of the site-specific recombination event, resulting in FIS-dependent inversion of viral DNA segments. Here we propose a model for the mechanism of thrU(tuJB) trans-activation taking into account various features common to both processes.
MATERIALS AND METHODS The E.coli strains MC1000 and MC100-fis767, the plasmid pDS10, harbouring the operon fusion thrU(tuiB)':galK, and its deletion derivatives, the growth medium (LB medium plus 1 % glucose) and the galactokinase activity assay, have been described previously [22,1,4]. The pDS1O derivatives harbour a BamHI linker (CCGGATCCGG) at the junctions between undeleted region and vector. For the insertion of three base pairs at position -95 (see Table I) the pDS1O derivative with a deletion up to -88 was digested with BamHI. A BglIH-BamHI fragment, carrying part of the UAS extending from -495 to -95, was then ligated in the BamHI site. For the deletion of one base pair at -95 the same procedure was carried out with a pDSIO derivative carrying a deletion up to -84 (see also legend to Table I).
RESULTS Since activation of the thrU(tuJ3) operon consists of two components: FIS-independent and FIS-dependent activation, the question may be raised whether both components have the same DNA determinants. To answer this question transcriptional activity was measured in MC1000 and MClO00-fis767 cells after deleting various portions of the UAS. To this aim expression of the thrU(tuJB) operon was studied using the plasmid-borne operon fusion thrU(tuJB)':galK (cf. Fig. 1B) measuring galactokinase activity/femtomole of plasmid, as described by Nilsson et al. [4]. Previous studies of operon expression [22] did not reveal any effect of deletions from positions -500 to -200. We, therefore, chose position -176 as deletion start point for the present experiments. As illustrated in Fig. 3 a deletion extending to 88 abolishes FIS-dependent trans-activation. FIS-independent cisactivation is also affected but to a lesser extent. DNA with such a deletion has lost the upstream FIS-binding region at 131 to 1 8 and a substantial part of the intervening sequence between the two FIS-binding regions (cf. bars in Fig. 3). DNA fragments derived from the deleted UAS form only one FIS/DNA complex [3]. We conclude that both FIS-dependent and FIS-independent transcription activation have DNA determinants upstream of position -88. A significant decrease of transcriptional activity is seen when the deletion is extended to -81. The data of Fig. 3 show that this is due to a drop of FIS-independent cis-activation. Further deletions to -56, resulting in a total elimination of the UAS, cause a further decline which can only be observed in FISproducing cells, since derepression of ribosome feedback inhibition counteracts this decline in fis cells [4]. This implies that FIS-independent cis-activation declines continuously upon successive elimination of sequences from 176 to -56, with a steep drop between -88 and -81. The results of Figure 3 show that FIS-dependent trans-
-
-
-
activation requires both FIS-binding regions. Apparently, FIS exerts its function when attached to both binding regions of the UAS. Deletion of the region -176 to 133, just upstream the FIS-binding region 131 to 118, diminishes activation in both -
-
-
Nucleic Acids Research, Vol. 20, No. 15 4079 by G-C base pairs, and the deletion removes 11 basepairs which are replaced by 10 basepairs of the BamHI linker (CCGGATCCGG). Transcriptional activity in fis cells is also reduced by the insertion and deletion. For an interpretation of these data see the Discussion.
100 -133
75
-
DISCUSSION
-176 C
50
0) c
25
CD
-150
-100
-50
endpoint of deletion
Fig. 3. The effects of deleting increasing parts of the UAS on FIS-dependent and FIS-independent activation of the thrU(tuJB) operon. Galactokinase activities of cells transformed with pDS1O derivatives harbouring the thrU(tuJB),:galK fusion operon (determined in duplicate) are expressed as percentage of the activity measured with an intact UAS. All deletions start at position -176. 0-0 MC1000 cells; A-A MC1000-fis767 cells. The two FIS-binding regions (positions -131 to -1 18 and -81 to -48) are indicated with bars.
+1
24
-131
nm
-S -48
-11 8
-81
15 nm sI:i=^8' FIS binding region
Fig. 4. Model of FIS/RNAP interaction at the thrU(tuJB) promoter and its upstream region. For details see the Discussion.
FIS-producing and fis cells. This region therefore is also a functional part of the UAS, but its elimination does not decrease the binding of FIS to DNA. DNA fragments derived from this partially deleted UAS still form three FIS/DNA complexes in vitro [3]. Insertion of three basepairs in position -95 (the bending centre of UAS-derived DNA fragments [18]) or a deletion of one basepair reduce trans-activation significantly (Table 1) without affecting FIS binding to DNA (not shown). These manipulations alter the spatial orientation of the two FIS-binding sites and the nucleotide sequence of the intervening region. The insertion (of three G-C base pairs) causes replacement of three A-T base pairs
An important conclusion from the present study is that proper trans-activation of the thrU(tujB) operon requires occupation by FIS of the two FIS-binding regions on the UAS (Fig.3) and correct positioning of these regions. The insertion and the deletion at position -95 (see Table I) not only alter the distance between these regions but also their spatial orientation. The helical positioning of the regions is altered by changing the number of nucleotides and the conformation of the intervening sequence. Replacement of some A-T base pairs by G-C base pairs may affect DNA bending. Apparently, the FIS-molecules bound at the two regions can only exert their trans-activating function when positioned very specifically in space. This suggests that these FIS molecules, besides interacting with DNA, also interact with other biological macromolecules during trans-activation. One possibility is that they interact with each other, bringing the two FIS-binding regions in close proximity. Whether such a FIS/FIS interaction is sterically feasible remains to be seen. An alternative is interaction with the RNA polymerase. A model for such a FIS/RNAP interaction is illustrated in Fig. 4. The RNAP structure pictured here is derived from Meisenberger et al. [23] and Heumann et al. [24] who studied the enzyme with smallangle X-ray and neutron scattering, respectively. Since the conformation of the DNA studied in the present investigation is anomalous and has a bent structure, our proposal for RNAP/DNA interaction differs from that proposed by these authors. A major characteristic of our model is that the DNA region interacting with the RNAP extends beyond position -45, the classical boundary determined by Siebenlist et al. [25]. Travers et al. [6,19,21] studying the tyrT operon, were the first to describe promoter upstream interactions with RNAP. They ascribed them to binding of more than one RNAP upstream of stable RNA promoters and suggested that RNAP binding at a secondary site might activate transcription. These early studies, performed in the absence of FIS, showed that RNAP protects the upstream region against DNaseI up to position -65, whereas RNAP protects the lacUV5 mRNA promoter to position -42. With a mutant tyrT promoter more extensive protections were found to at least -130. Our results, presented in Fig. 3, reveal a continuous decrease of FIS-independent cis-activation of the thrU(tujB) operon upon progressive deletion of the UAS starting at - 176, with a rather steep drop upon deleting from -88 to -81. Although these results may in principle be explained solely on the basis of the DNA conformation, i.e. DNA bending, RNAP/DNA interaction over large parts of the UAS may be envisioned, which is also in line with the RNAP footprints reported by Travers et al. for the tyrToperon [I. c. ]. The insertion or deletion performed in the experiments of Table I would disrupt such an interaction around and upstream of position -95. The results of the Table for fis cells, showing a substantial decrease of transcriptional activity by these manipulations, suggest that sequences upstream of -95 are involved in FIS-independent cisactivation. The effects of the manipulations in FIS-producing cells suggest that FIS-dependent trans-activation occurs through contacts
4080 Nucleic Acids Research, Vol. 20, No. 15 between RNAP and FIS molecules bound at both binding regions of the UAS. The interaction sites on the enzyme should then be identical or nearly identical. We assume that the two identical a subunits are the target sites of FIS. The topography of the RNAP as proposed by Meisenberger et al. and Heumann et al. [23,24], with a 15 nm long axis of the RNAP ellipsoid section would allow FIS molecules, separated by the intervening sequence of approximately four DNA helical turns, to interact with the ae subunits. Investigations by the group of Ishihama [26,27] showed that the a subunit plays an essential role in the communication between activator proteins and RNAP on certain promoters. Using reconstituted mutant RNAP, lacking the Cterminal region (residues 257 - 329), these authors demonstrated that the C-terminal domain of a is needed for transcription activation by activators, such as the cAMP receptor protein (CRP), and OmpR, the activator of the genes encoding the porin proteins OmpC and OmpF. The mutant RNAP did not respond to activator proteins that bind upstream of the respective promoters. Transcription by these mutant enzymes was, however, activated in the cases where activators bind to target sites that overlap the promoter -35 region. These observations indicate that the C-terminal region of a carries a contact site (site I) for some, if not all, of the transcription activators that bind (like FIS) DNA sites located upstream of the basic promoter elements. In vitro binding of increasing amounts of FIS to a DNA fragment, harbouring the entire UAS, results in the formation of up to three FIS/DNA complexes [1-3]. All complexes and the naked DNA fragment have the same bending centre (not shown), suggesting that both binding regions are occupied by FIS in all three complexes but we do not fully understand the relationship between complex formation and the occupation of the two FIS-binding regions. Our model of the RNAP/DNA interaction implies that the UAS wraps around the RNAP in the presence and absence of FIS. Although the DNA bending increases in the presence of FIS, wrapping may occur without altering the UAS bending centre in either case. In the absence of FIS the interaction may be less stable or transient. FIS/RNAP interaction may stabilize the RNAP/DNA interaction over a considerable part of the UAS, and thus may trigger conformational changes of the RNAP and of the DNA at the transcription start site, converting the closed promoter complex to an open complex. In the absence of FIS the dynamic interaction between RNAP and DNA most likely leads less frequently to effective opening of the closed promoter complex. A further reduction of this frequency occurs when an activator sequence upstream of the promoter is lacking and no DNA wrapping around the RNAP can take place. The two FIS-dependent processes: stimulation of viral DNA inversion and trans-activation of stable RNA operon transcription have various features in common. Stimulation of the inversion process occurs upon binding of FIS to two specific regions on the recombinational enhancer [15,16,28] (cf. Fig.2). These two regions are separated by a well defined intervening sequence that allows for a conformational change of enhancer DNA upon interaction with FIS. The change is manifested by DNA bending and is an essential step for recombinational enhancer activity. Insertions other than an integral number of helical turns between the two sites destroy the enhancer function, although FIS/DNA binding is not abolished [29,30]. A model of enhancer action has been proposed [29-32]. According to this model, the FISenhancer complex and the recombinase-crossover site complexes have first to interact before recombination can take place. This
putative interaction between FIS and recombinase could be necessary in order to align the two crossover sites properly in a synaptic complex. Electron microscopy of crosslinked nucleoprotein complexes, obtained with the Hin-family of recombinases, has shown that the FIS-bound recombinational enhancer is closely associated with the two Hin-bound recombination sites [33]. According to the investigators a tripartite recombination intermediate structure is formed, which they call an invertasome. Mutational analysis of FIS [34] is in line with FIS/Hin interaction. This analysis shows that there are at least two functionally distinct domains in FIS: an N-terminal region that is uniquely required for stimulating Hin-mediated inversion and a C-terminal region required for DNA binding and bending. The crystal structure of FIS [35,36] has been described as a compact ellipsoid dimer, with each subunit containing four a-helices (AD) separated by short $-turns. The C and D helices located near the C-terminus correspond to a helix-turn-helix DNA binding region. The first 24 residues were not resolved, implying that this region is disordered in the crystal. The accessibility of the N-terminal residues for FIS-Hin interaction thus remains to be determined directly from the crystal structure. The model for FIS/recombinase interaction and that of Fig.4 for FIS/RNAP interaction may explain how the FIS protein can stimulate two unrelated processes.
ACKNOWLEDGEMENTS We thank Drs. E.Vijgenboom, B.Kraal and C.W.A.Pleij for discussions, suggestions and for critically reading the manuscript and Dr. C.A.van Sluis for computer comparisons of amino acid sequences. The plasmids and deletion derivatives were generously donated by Dr. J.van Delft. The strains MC1000 and MC 1000-fis767 were kindly provided by Dr. R.C.Johnson. The investigation was supported in part by the Commission of the European Communities, Biotechnology Action Programme (BAP), Directorate-General Science, Research and Development, Brussels. L.N. was the recipient of a long-term EMBO fellowship.
REFERENCES 1. Nilsson, L., Vanet, A., Vijgenboom, E. and Bosch, L. (1990) EMBO J. 9, 727-734. 2. Bosch, L., Nilsson L., Vijgenboom E. and Verbeek H. (1990) Biochim. Biophvs. Acta 1050, 293-301. 3. Vijgenboom, E., Nilsson, L. and Bosch, L. (1988) Nucleic Acids Res. 16,
10183-10197. 4. Nilsson, L., Verbeek, H., Vijgenboom, E., van Drunen, C., Vanet, A. and Bosch, L. (1992) J. Bacteriol. 174, 921-929. 5. Lamond, A.I. and Travers, A.A. (1983) Nature, 30, 248-250. 6. Travers, A.A., Lamond, A.I., Mace, H.A.F. and Berman, M.L. (1983) Cell, 35, 265-273. 7. Verbeek, H., Nilsson, L., Baliko, G. and Bosch, L. (1990) Biochim. Biophys. Acta 1050, 302-306. 8. Gourse, R.L., de Boer, H.A. and Nomura, M. (1986) Cell, 44, 197-205. 9. Ross, W., Thompson, J.F., Newland, J.T. and Gourse, R.L. (1990) EMBO J. 9, 3733-3742. 10. Btermier, M., Lefrere, V., Koch, C., Alazard, R., and Chandler, M. (1989) Mol. Microbiol. 3, 459-468. 11. Johnson, R.C. and Simon, M.I. (1987) Trends Genet. 3, 262-267. 12. Thompson, J.F., Moitoso de Vargas, L., Koch, C., Kahmann, R. and Landy, A. (1987) Cell 50, 901-908. 13. van de Putte, P., Cramer, S. and Giphart-Gassler, M. (1980) Nature 286, 2 18-222. 14. Grundy, F.J. and Howe, M.M. (1984) Virology 134, 296-317.
Nucleic Acids Research, Vol. 20, No. 15 4081 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
28. 29. 30. 31.
32. 33. 34. 35. 36.
Johnson, R.C. and Simon, M.I. (1985) Cell 41, 781-791. Kahmann, R., Rudt, F., Koch, C. and Mertens, G. (1985) Cell 41, 771 -780. Vijgenboom, E. Ph.D. thesis, Leiden University, 1989. Verbeek, H., Nilsson, L. and Bosch, L. (1991) Biochimie 73, 713-718. Travers, A.A. (1984) Nucleic Acids Res. 12, 2605-2618. Bossi, L. and Smith, D.M. (1984) Cell 39, 643-652. Drew, G.R. and Travers, A.A. 919840 Cell 37, 491-502. van Delft, J.H.M., Marinon, B., Schmidt, D.S. and Bosch, L. (1987) Nucleic Acids Res. 15, 9511-9538. Meisenberger, O., Heumann, H. and Pilz, I. (1981) FEBSLett. 123, 22-24. Heumann, H., Lederer, H. and Baer, G. (1988) J. Mol. Bio. 201, 115-125. Siebenlist, U., Simpson, R.B. and Gilbert, W. (1980) Cell 20, 269-281. Igarashi, K. and Ishihama, A. (1991) Cell 65, 1015-1022. Igarashi, K., Hanamura, A. Makino, K., Aiba, H., Aiba, H., Mizuno, T., Nakata, A. and Ishihama, A. (1991) Proc. Natl. Acad. Sci. USA 88, 8958-8962. Huber, H.E., lida, S., Arber, W. and W. Bickle, T.A. (1985) Proc. Natl. Acad. Sci. USA 82, 3766-3780. Hubner, P. and Arber, W. (1989) EMBO J. 8, 577-585. Hubner, P., Haffter, P., lida, S. and Arber, W. (1989) J. Mol. Biol. 205, 493 -500. Bruist, M.F., Glasgow, A.C., Johnson, R.C. and Simon, M.I. (1987) Genes Dev. 1, 762-772. Kanaar, R., van de Putte and Cozzarelli, N.R. (1988) Proc. Natl. Acad. Sci. USA 85, 752-756 Heichman, K.A. and Johnson, R.C. (1990) Science 249, 511-517. Osuna, R., Finker, S.E. and Johnson, R.C. (1991) EMBO J. 10, 1595-1603. Kostrewa, D., Granzin, J., Koch C., Choe, H.-W., Raghunathan, S., Wolf, W., Labahn, J., Kahmann, R. and Saenger, W. (1991) Nature 349, 178-180. Yuan, H.S., Finkel, S.E., Feng, J.-A., Kaczor-Grzeskowiak, M., Johnson, R.C. and Dickerson, R.E. (1991) Proc. Natl. Acad. Sci. USA 88, 9558-9562.
Nucleic Acids Research, Vol. 20, No. 15 4077-4081
The mechanism of trans-activation of the Escherichia coli operon thrU(tufB) by the protein FIS. A model Hans Verbeek, Lars Nilsson1 and Leendert Bosch* Department of Biochemistry, Leiden University, Gorlaeus Laboratories, PO Box 9502, 2300 RA Leiden, The Netherlands and 1Department of Cell Biology, University of Stockholm, S-106 91 Stockholm, Sweden Received March 3, 1992; Revised and Accepted July 1, 1992
ABSTRACT Transcription of the thrU(tufB) operon is trans-activated by the protein FIS which binds to the promoter upstream activator sequence (UAS). Deletions of parts of the UAS and insertions show that optimal transactivation requires occupation by FIS of the two FISbinding regions on the UAS and specific helical positioning of these regions. On the basis of these and other data, a model for the mechanism of thrU(tufB) trans-activation by FIS is proposed. This model implies that the mechanisms underlying stimulation by FIS of two totally different processes: inversion of viral DNA segments and transcription of stable RNA operons, are essentially the same. INTRODUCTION The thrU(tufl) operon of E. coli is a mixed operon, harbouring four different tRNA genes and the protein gene tuJB, encoding the polypeptide chain elongation factor EF-TuB (Fig. IA). Transcription of this operon is activated in trans by the E. coli protein FIS, which binds to a cis-acting region upstream of the promoter (designated UAS for upstream activator sequence) [ 1, 2, 3 and 4]. Similar activator sequences are also found upstream of the tyrT [5,6] and the metY [7] operons, and upstream of the P1 promoter of the rrnB operon [8]. They all specifically bind FIS [1], suggesting that this is the common trans-activator of these four operons and possibly more stable RNA operons [7]. FIS-dependent trans-activation of the rrnB operon has subsequently been demonstrated [9]. FIS (Factor for Inversion Stimulation) is a heat-stable protein of 98 amino acids, until recently only known for its role in the replication of certain bacterial viruses [10-12]. Some of these viruses alter their host range by switching the orientation of a viral DNA segment, which results in the expression of a different set of tail fiber genes [13,14]. This site-specific recombination event, mediated by a virus encoded recombinase, is stimulated by the host protein FIS which binds to a cis-acting recombinational enhancer [15,16] (Fig. 2).
*
To whom correspondence should be addressed
For some time it was not clear how E. coli benefits from FIS, a protein which enables the phage to attack bacteria. Recent studies [4] revealed that FIS plays a prominant role in accelerating bacterial growth. During a normal growth cycle early log phase cells show a large increase in FIS-dependent trans-activation of the thrU(tuJB) operon. Concomitantly, a peak of the cellular FIS concentration is seen. Cells, growing slowly in a poor medium, do not show such a rise and fall of FIS, nor a peak of operon transcription in the early log phase. After addition of nutrients to the medium these cells promptly respond with an increase both of transcriptional activity and of the FIS concentration. This increase gets higher as the medium gets more strongly enriched. Apparently, the synthesis of FIS responds to environmental signals, thereby affecting transcription of the thrU(tuJB) operon, and most likely more stable RNA operons. Furthermore, FIS allows cells to achieve very high growth rates. What is the mechanism underlying trans-activation of the thrU(tuJB) operon? The evidence obtained so far [17,3,18] and the data of the present paper show that bending of the UAS DNA double helix is essential for transcription activation. The UASs of various stable RNA operons are AT-rich [19], indicative of an unusual physical conformation, such as kinking or bending [20,21,8]. Accordingly, gel electrophoretic migration of UAS DNA fragments is retarded [1-3]. This retardation increases upon FIS binding, indicating that the nucleotide sequence-induced bending of the DNA is enhanced by FIS binding. Three FIS/DNA complexes are formed under these in vitro conditions. Cells unable to produce FIS (fis cells) do show cis-activation of thrU(tuJB) transcription. Most likely nucleotide sequence-induced bending plays a role here. FIS-directed bending (in wild-type cells) further increases transcription activation. DNase I footprinting [3,2] has delimited two FIS binding regions on the UAS of the thrU(tufB) operon (see bars in Fig. 3). Electrophoretic analysis of circularly permuted UAS DNA fragments located a bending centre around position -95 (with respect to the transcription start site), in between the two FISbinding regions [ 18]. Insertions of small DNA segments, comprising less than one or two complete helix turns at the
4078 Nucleic Acids Research, Vol. 20, No. 15 Table I. Effects of altering the helical positioning of the FIS-binding regions
A
galactokinase activity MC 1000 cells MCIOOO-fis767 cells
UU .
B
-j7,,,j-}
Fig. 1. A; The thrU(tuJB) operon of E. coli. B; The fusion operon thrU (tuiB) ':galK. Translation stop codons in three reading frames between the junction point and galK prevent translational readthrough from the cloned tuIB fragment into galK.
54 ± 2 21 ± 2 29 ± 15
Cells were grown in LB medium to an OD600 of 0.5 and galactokinase activities, expressed as activity/femtomole of plasmid, were measured as in [4]. Galactokinase activities are expressed as in Fig. 3. The insertion and the deletion alter the UAS sequence as follows: -130
wt.
recombinational enhancer
100 ± 8 12 + 3 24 + 6
no insertion one base pair deletion at -95 three base pair insertion at -95
-100
-i10
-120
-90
-80
-70
TGTTGAAAGTGTGCTAATCTGCCCTCCG.TTC3-GCTGTTTCTTCATCGTGTCGC&TAAAATG
invertible segment
{ }I-//KZJ
-lbp.
TGTTGAAAAGTGTGCTAATCTGCCCTCCGTTCGGCTccggatccggGTGTCGCATAAAATG
+3bp.
TGAAAAAGTGTGCTAATCTGCCCTCCGTTC^GGTccggatccggCATCGTGTCnCATAAAATG
FIS binding site
ED
crossover
site
Fig. 2. Invertible viral DNA segment and recombinational enhancer.
junction of the UAS and the RNA polymerase binding site (position -45), reduce transcription significantly. Partial restoration of transcriptional activity occurs when one or more full helix turns are inserted [ 18]. These data clearly demonstrate that an abnormal conformation of the UAS DNA, most likely a bent DNA double helix, is an essential requisite for transcription activation of the operon. Furthermore, that activation requires a specific spatial orientation of the DNA bend with respect to the RNA polymerase (RNAP). In the present investigation we show that optimal transactivation of the thrU(tuJB) operon not only requires occupation by FIS of both FIS binding regions on the UAS, but also specific helical positioning of these regions. These requirements are very similar to those of the site-specific recombination event, resulting in FIS-dependent inversion of viral DNA segments. Here we propose a model for the mechanism of thrU(tuJB) trans-activation taking into account various features common to both processes.
MATERIALS AND METHODS The E.coli strains MC1000 and MC100-fis767, the plasmid pDS10, harbouring the operon fusion thrU(tuiB)':galK, and its deletion derivatives, the growth medium (LB medium plus 1 % glucose) and the galactokinase activity assay, have been described previously [22,1,4]. The pDS1O derivatives harbour a BamHI linker (CCGGATCCGG) at the junctions between undeleted region and vector. For the insertion of three base pairs at position -95 (see Table I) the pDS1O derivative with a deletion up to -88 was digested with BamHI. A BglIH-BamHI fragment, carrying part of the UAS extending from -495 to -95, was then ligated in the BamHI site. For the deletion of one base pair at -95 the same procedure was carried out with a pDSIO derivative carrying a deletion up to -84 (see also legend to Table I).
RESULTS Since activation of the thrU(tuJ3) operon consists of two components: FIS-independent and FIS-dependent activation, the question may be raised whether both components have the same DNA determinants. To answer this question transcriptional activity was measured in MC1000 and MClO00-fis767 cells after deleting various portions of the UAS. To this aim expression of the thrU(tuJB) operon was studied using the plasmid-borne operon fusion thrU(tuJB)':galK (cf. Fig. 1B) measuring galactokinase activity/femtomole of plasmid, as described by Nilsson et al. [4]. Previous studies of operon expression [22] did not reveal any effect of deletions from positions -500 to -200. We, therefore, chose position -176 as deletion start point for the present experiments. As illustrated in Fig. 3 a deletion extending to 88 abolishes FIS-dependent trans-activation. FIS-independent cisactivation is also affected but to a lesser extent. DNA with such a deletion has lost the upstream FIS-binding region at 131 to 1 8 and a substantial part of the intervening sequence between the two FIS-binding regions (cf. bars in Fig. 3). DNA fragments derived from the deleted UAS form only one FIS/DNA complex [3]. We conclude that both FIS-dependent and FIS-independent transcription activation have DNA determinants upstream of position -88. A significant decrease of transcriptional activity is seen when the deletion is extended to -81. The data of Fig. 3 show that this is due to a drop of FIS-independent cis-activation. Further deletions to -56, resulting in a total elimination of the UAS, cause a further decline which can only be observed in FISproducing cells, since derepression of ribosome feedback inhibition counteracts this decline in fis cells [4]. This implies that FIS-independent cis-activation declines continuously upon successive elimination of sequences from 176 to -56, with a steep drop between -88 and -81. The results of Figure 3 show that FIS-dependent trans-
-
-
-
activation requires both FIS-binding regions. Apparently, FIS exerts its function when attached to both binding regions of the UAS. Deletion of the region -176 to 133, just upstream the FIS-binding region 131 to 118, diminishes activation in both -
-
-
Nucleic Acids Research, Vol. 20, No. 15 4079 by G-C base pairs, and the deletion removes 11 basepairs which are replaced by 10 basepairs of the BamHI linker (CCGGATCCGG). Transcriptional activity in fis cells is also reduced by the insertion and deletion. For an interpretation of these data see the Discussion.
100 -133
75
-
DISCUSSION
-176 C
50
0) c
25
CD
-150
-100
-50
endpoint of deletion
Fig. 3. The effects of deleting increasing parts of the UAS on FIS-dependent and FIS-independent activation of the thrU(tuJB) operon. Galactokinase activities of cells transformed with pDS1O derivatives harbouring the thrU(tuJB),:galK fusion operon (determined in duplicate) are expressed as percentage of the activity measured with an intact UAS. All deletions start at position -176. 0-0 MC1000 cells; A-A MC1000-fis767 cells. The two FIS-binding regions (positions -131 to -1 18 and -81 to -48) are indicated with bars.
+1
24
-131
nm
-S -48
-11 8
-81
15 nm sI:i=^8' FIS binding region
Fig. 4. Model of FIS/RNAP interaction at the thrU(tuJB) promoter and its upstream region. For details see the Discussion.
FIS-producing and fis cells. This region therefore is also a functional part of the UAS, but its elimination does not decrease the binding of FIS to DNA. DNA fragments derived from this partially deleted UAS still form three FIS/DNA complexes in vitro [3]. Insertion of three basepairs in position -95 (the bending centre of UAS-derived DNA fragments [18]) or a deletion of one basepair reduce trans-activation significantly (Table 1) without affecting FIS binding to DNA (not shown). These manipulations alter the spatial orientation of the two FIS-binding sites and the nucleotide sequence of the intervening region. The insertion (of three G-C base pairs) causes replacement of three A-T base pairs
An important conclusion from the present study is that proper trans-activation of the thrU(tujB) operon requires occupation by FIS of the two FIS-binding regions on the UAS (Fig.3) and correct positioning of these regions. The insertion and the deletion at position -95 (see Table I) not only alter the distance between these regions but also their spatial orientation. The helical positioning of the regions is altered by changing the number of nucleotides and the conformation of the intervening sequence. Replacement of some A-T base pairs by G-C base pairs may affect DNA bending. Apparently, the FIS-molecules bound at the two regions can only exert their trans-activating function when positioned very specifically in space. This suggests that these FIS molecules, besides interacting with DNA, also interact with other biological macromolecules during trans-activation. One possibility is that they interact with each other, bringing the two FIS-binding regions in close proximity. Whether such a FIS/FIS interaction is sterically feasible remains to be seen. An alternative is interaction with the RNA polymerase. A model for such a FIS/RNAP interaction is illustrated in Fig. 4. The RNAP structure pictured here is derived from Meisenberger et al. [23] and Heumann et al. [24] who studied the enzyme with smallangle X-ray and neutron scattering, respectively. Since the conformation of the DNA studied in the present investigation is anomalous and has a bent structure, our proposal for RNAP/DNA interaction differs from that proposed by these authors. A major characteristic of our model is that the DNA region interacting with the RNAP extends beyond position -45, the classical boundary determined by Siebenlist et al. [25]. Travers et al. [6,19,21] studying the tyrT operon, were the first to describe promoter upstream interactions with RNAP. They ascribed them to binding of more than one RNAP upstream of stable RNA promoters and suggested that RNAP binding at a secondary site might activate transcription. These early studies, performed in the absence of FIS, showed that RNAP protects the upstream region against DNaseI up to position -65, whereas RNAP protects the lacUV5 mRNA promoter to position -42. With a mutant tyrT promoter more extensive protections were found to at least -130. Our results, presented in Fig. 3, reveal a continuous decrease of FIS-independent cis-activation of the thrU(tujB) operon upon progressive deletion of the UAS starting at - 176, with a rather steep drop upon deleting from -88 to -81. Although these results may in principle be explained solely on the basis of the DNA conformation, i.e. DNA bending, RNAP/DNA interaction over large parts of the UAS may be envisioned, which is also in line with the RNAP footprints reported by Travers et al. for the tyrToperon [I. c. ]. The insertion or deletion performed in the experiments of Table I would disrupt such an interaction around and upstream of position -95. The results of the Table for fis cells, showing a substantial decrease of transcriptional activity by these manipulations, suggest that sequences upstream of -95 are involved in FIS-independent cisactivation. The effects of the manipulations in FIS-producing cells suggest that FIS-dependent trans-activation occurs through contacts
4080 Nucleic Acids Research, Vol. 20, No. 15 between RNAP and FIS molecules bound at both binding regions of the UAS. The interaction sites on the enzyme should then be identical or nearly identical. We assume that the two identical a subunits are the target sites of FIS. The topography of the RNAP as proposed by Meisenberger et al. and Heumann et al. [23,24], with a 15 nm long axis of the RNAP ellipsoid section would allow FIS molecules, separated by the intervening sequence of approximately four DNA helical turns, to interact with the ae subunits. Investigations by the group of Ishihama [26,27] showed that the a subunit plays an essential role in the communication between activator proteins and RNAP on certain promoters. Using reconstituted mutant RNAP, lacking the Cterminal region (residues 257 - 329), these authors demonstrated that the C-terminal domain of a is needed for transcription activation by activators, such as the cAMP receptor protein (CRP), and OmpR, the activator of the genes encoding the porin proteins OmpC and OmpF. The mutant RNAP did not respond to activator proteins that bind upstream of the respective promoters. Transcription by these mutant enzymes was, however, activated in the cases where activators bind to target sites that overlap the promoter -35 region. These observations indicate that the C-terminal region of a carries a contact site (site I) for some, if not all, of the transcription activators that bind (like FIS) DNA sites located upstream of the basic promoter elements. In vitro binding of increasing amounts of FIS to a DNA fragment, harbouring the entire UAS, results in the formation of up to three FIS/DNA complexes [1-3]. All complexes and the naked DNA fragment have the same bending centre (not shown), suggesting that both binding regions are occupied by FIS in all three complexes but we do not fully understand the relationship between complex formation and the occupation of the two FIS-binding regions. Our model of the RNAP/DNA interaction implies that the UAS wraps around the RNAP in the presence and absence of FIS. Although the DNA bending increases in the presence of FIS, wrapping may occur without altering the UAS bending centre in either case. In the absence of FIS the interaction may be less stable or transient. FIS/RNAP interaction may stabilize the RNAP/DNA interaction over a considerable part of the UAS, and thus may trigger conformational changes of the RNAP and of the DNA at the transcription start site, converting the closed promoter complex to an open complex. In the absence of FIS the dynamic interaction between RNAP and DNA most likely leads less frequently to effective opening of the closed promoter complex. A further reduction of this frequency occurs when an activator sequence upstream of the promoter is lacking and no DNA wrapping around the RNAP can take place. The two FIS-dependent processes: stimulation of viral DNA inversion and trans-activation of stable RNA operon transcription have various features in common. Stimulation of the inversion process occurs upon binding of FIS to two specific regions on the recombinational enhancer [15,16,28] (cf. Fig.2). These two regions are separated by a well defined intervening sequence that allows for a conformational change of enhancer DNA upon interaction with FIS. The change is manifested by DNA bending and is an essential step for recombinational enhancer activity. Insertions other than an integral number of helical turns between the two sites destroy the enhancer function, although FIS/DNA binding is not abolished [29,30]. A model of enhancer action has been proposed [29-32]. According to this model, the FISenhancer complex and the recombinase-crossover site complexes have first to interact before recombination can take place. This
putative interaction between FIS and recombinase could be necessary in order to align the two crossover sites properly in a synaptic complex. Electron microscopy of crosslinked nucleoprotein complexes, obtained with the Hin-family of recombinases, has shown that the FIS-bound recombinational enhancer is closely associated with the two Hin-bound recombination sites [33]. According to the investigators a tripartite recombination intermediate structure is formed, which they call an invertasome. Mutational analysis of FIS [34] is in line with FIS/Hin interaction. This analysis shows that there are at least two functionally distinct domains in FIS: an N-terminal region that is uniquely required for stimulating Hin-mediated inversion and a C-terminal region required for DNA binding and bending. The crystal structure of FIS [35,36] has been described as a compact ellipsoid dimer, with each subunit containing four a-helices (AD) separated by short $-turns. The C and D helices located near the C-terminus correspond to a helix-turn-helix DNA binding region. The first 24 residues were not resolved, implying that this region is disordered in the crystal. The accessibility of the N-terminal residues for FIS-Hin interaction thus remains to be determined directly from the crystal structure. The model for FIS/recombinase interaction and that of Fig.4 for FIS/RNAP interaction may explain how the FIS protein can stimulate two unrelated processes.
ACKNOWLEDGEMENTS We thank Drs. E.Vijgenboom, B.Kraal and C.W.A.Pleij for discussions, suggestions and for critically reading the manuscript and Dr. C.A.van Sluis for computer comparisons of amino acid sequences. The plasmids and deletion derivatives were generously donated by Dr. J.van Delft. The strains MC1000 and MC 1000-fis767 were kindly provided by Dr. R.C.Johnson. The investigation was supported in part by the Commission of the European Communities, Biotechnology Action Programme (BAP), Directorate-General Science, Research and Development, Brussels. L.N. was the recipient of a long-term EMBO fellowship.
REFERENCES 1. Nilsson, L., Vanet, A., Vijgenboom, E. and Bosch, L. (1990) EMBO J. 9, 727-734. 2. Bosch, L., Nilsson L., Vijgenboom E. and Verbeek H. (1990) Biochim. Biophvs. Acta 1050, 293-301. 3. Vijgenboom, E., Nilsson, L. and Bosch, L. (1988) Nucleic Acids Res. 16,
10183-10197. 4. Nilsson, L., Verbeek, H., Vijgenboom, E., van Drunen, C., Vanet, A. and Bosch, L. (1992) J. Bacteriol. 174, 921-929. 5. Lamond, A.I. and Travers, A.A. (1983) Nature, 30, 248-250. 6. Travers, A.A., Lamond, A.I., Mace, H.A.F. and Berman, M.L. (1983) Cell, 35, 265-273. 7. Verbeek, H., Nilsson, L., Baliko, G. and Bosch, L. (1990) Biochim. Biophys. Acta 1050, 302-306. 8. Gourse, R.L., de Boer, H.A. and Nomura, M. (1986) Cell, 44, 197-205. 9. Ross, W., Thompson, J.F., Newland, J.T. and Gourse, R.L. (1990) EMBO J. 9, 3733-3742. 10. Btermier, M., Lefrere, V., Koch, C., Alazard, R., and Chandler, M. (1989) Mol. Microbiol. 3, 459-468. 11. Johnson, R.C. and Simon, M.I. (1987) Trends Genet. 3, 262-267. 12. Thompson, J.F., Moitoso de Vargas, L., Koch, C., Kahmann, R. and Landy, A. (1987) Cell 50, 901-908. 13. van de Putte, P., Cramer, S. and Giphart-Gassler, M. (1980) Nature 286, 2 18-222. 14. Grundy, F.J. and Howe, M.M. (1984) Virology 134, 296-317.
Nucleic Acids Research, Vol. 20, No. 15 4081 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.
28. 29. 30. 31.
32. 33. 34. 35. 36.
Johnson, R.C. and Simon, M.I. (1985) Cell 41, 781-791. Kahmann, R., Rudt, F., Koch, C. and Mertens, G. (1985) Cell 41, 771 -780. Vijgenboom, E. Ph.D. thesis, Leiden University, 1989. Verbeek, H., Nilsson, L. and Bosch, L. (1991) Biochimie 73, 713-718. Travers, A.A. (1984) Nucleic Acids Res. 12, 2605-2618. Bossi, L. and Smith, D.M. (1984) Cell 39, 643-652. Drew, G.R. and Travers, A.A. 919840 Cell 37, 491-502. van Delft, J.H.M., Marinon, B., Schmidt, D.S. and Bosch, L. (1987) Nucleic Acids Res. 15, 9511-9538. Meisenberger, O., Heumann, H. and Pilz, I. (1981) FEBSLett. 123, 22-24. Heumann, H., Lederer, H. and Baer, G. (1988) J. Mol. Bio. 201, 115-125. Siebenlist, U., Simpson, R.B. and Gilbert, W. (1980) Cell 20, 269-281. Igarashi, K. and Ishihama, A. (1991) Cell 65, 1015-1022. Igarashi, K., Hanamura, A. Makino, K., Aiba, H., Aiba, H., Mizuno, T., Nakata, A. and Ishihama, A. (1991) Proc. Natl. Acad. Sci. USA 88, 8958-8962. Huber, H.E., lida, S., Arber, W. and W. Bickle, T.A. (1985) Proc. Natl. Acad. Sci. USA 82, 3766-3780. Hubner, P. and Arber, W. (1989) EMBO J. 8, 577-585. Hubner, P., Haffter, P., lida, S. and Arber, W. (1989) J. Mol. Biol. 205, 493 -500. Bruist, M.F., Glasgow, A.C., Johnson, R.C. and Simon, M.I. (1987) Genes Dev. 1, 762-772. Kanaar, R., van de Putte and Cozzarelli, N.R. (1988) Proc. Natl. Acad. Sci. USA 85, 752-756 Heichman, K.A. and Johnson, R.C. (1990) Science 249, 511-517. Osuna, R., Finker, S.E. and Johnson, R.C. (1991) EMBO J. 10, 1595-1603. Kostrewa, D., Granzin, J., Koch C., Choe, H.-W., Raghunathan, S., Wolf, W., Labahn, J., Kahmann, R. and Saenger, W. (1991) Nature 349, 178-180. Yuan, H.S., Finkel, S.E., Feng, J.-A., Kaczor-Grzeskowiak, M., Johnson, R.C. and Dickerson, R.E. (1991) Proc. Natl. Acad. Sci. USA 88, 9558-9562.
