SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Escherichia coli MelR protein is a transcription activator that, in the presence of melibiose, activates expression of the melAB operon by binding to four sites located just upstream of the melAB promoter. MelR is encoded by the melR gene, which is expressed from a divergent transcript that starts 237 bp upstream of the melAB promoter transcript start point. In a recent study, we have identified a fifth DNA site for MelR that overlaps the melR promoter transcript start and −10 region. Here we show that MelR binding to this site can downregulate expression from the melR promoter; thus, MelR autoregulates its own expression. Optimal repression of the melR promoter is observed in the absence of melibiose and requires one of the four other DNA sites for MelR at the melAB promoter. The two MelR binding sites required for this optimal repression are separated by 177 bp. We suggest that, in the absence of melibiose, MelR forms a loop between these two sites. We argue that, in the presence of melibiose, this loop is broken as the melAB promoter is activated. However, in the presence of melibiose, the melR promoter can still be partially repressed by MelR binding to the site that overlaps the transcript start and −10 region. Parallels with the Escherichia coli araC–araBAD regulatory region are discussed.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Transcription initiation at certain bacterial promoters can be downregulated by a repressor that binds to a target that overlaps the transcript start site or the −10 element (Gralla and Collado-Vides, 1996). At some of these promoters, the binding of the repressor appears simply to block access of the RNA polymerase holoenzyme (RNAP) to the promoter. However, efficient repression sometimes also requires the binding of another repressor molecule to an auxilliary site that is distant from the transcription start (Muller-Hill, 1998). In some of these cases, the binding of repressor to the auxilliary site improves repression by increasing the local repressor concentration. In other cases, repressor molecules, bound at different sites, combine to form a repression loop, and quite complex nucleo–protein structures form that are essential for repression (e.g. see Choy and Adhya, 1996).

One well-studied example of a complex repressor is the Escherichia coli AraC protein, which functions both as an activator and as a repressor of the transcription of genes necessary for arabinose metabolism. AraC is the founder member of a large family of gene regulatory proteins, the AraC family; the activities of many members of this family are regulated by small ligands (reviewed by Gallegos et al., 1997). AraC is functional as a dimer; in the presence of arabinose, the conformation of AraC is such that the two subunits of the dimer bind to two adjacent target sites resulting in activation of the araBAD promoter. However, in the absence of arabinose, the two subunits of the AraC dimer bind to two well-separated sites, one near the araBAD promoter and the other near the adjacent araC promoter. This results in the formation of a repression loop and the repression of the araC promoter by AraC (reviewed by Schleif, 1996).

In our studies, we have focussed attention on the E. coli MelR protein, which is a member of the AraC family. MelR is essential for induction of the melAB operon that is responsible for melibiose metabolism. MelR is encoded by the melR gene that is located upstream from the melAB promoter (Webster et al., 1987). In previous work, we characterized the melR promoter; it is divergent from the melAB promoter, with the melR and melAB transcript starts separated by 237 bp (Webster et al., 1988). Expression from the melR promoter is completely dependent on the cyclic AMP receptor protein (CRP), which binds to a DNA site centered at position −41.5 upstream from the melR transcript start point (Webster et al., 1988). MelR activity is regulated by melibiose, which is required for activation of transcription initiation at the melAB promoter (Webster et al., 1989). MelR binds to 18 bp targets located at the melAB promoter (Caswell et al., 1992; Williams et al., 1994). In the accompanying paper (Belyaeva et al., 2000), we showed that, in the absence of melibiose, MelR binds to three 18 bp target sites (Site 1, Site 1′ and Site 2) located just upstream of the melAB transcription start site. Melibiose induces the occupation of a fourth site (Site 2′) which is essential for transcription initiation at the melAB promoter (Fig. 1). During this work, and in a previous study (Gostick et al., 1998), it was noted that MelR could also bind to a fifth site (Site R) that overlaps the transcription start point and −10 hexamer of the melR promoter (Fig. 1). Our in vitro studies, showed that MelR can bind to this site in both the presence and absence of melibiose, that this site accommodates just one MelR subunit, and that binding is tightened by CRP (Belyaeva et al., 2000; T. A. Belyaeva, unpublished data). In this study, we have investigated whether the binding of MelR to this site affects expression from the melR promoter. We show that the melR promoter is repressed by MelR and that this autoregulation requires MelR binding to Site R. In the absence of melibiose, greater repression occurs in the presence of MelR binding Site 2, located 177 bp upstream. We propose that, for optimal repression, MelR forms a loop between Site R and Site 2, and that this loop is broken in the presence of melibiose when the melAB promoter is activated.

image

Figure 1. Schematic diagram of the KK81 and KK101 fragments.The figure illustrates the KK81 fragment carrying the divergent melAB and melR promoters, and the KK101 fragment carrying just the melR promoter. Horizontal arrows indicate the transcription start sites. DNA sequences are numbered with respect to the melR transcript start as +1. Thus, the KK81 and KK101 fragments are bounded by EcoRI sites at position +76 downstream of the melR promoter, and HindIII sites at positions −272 and −74, respectively, upstream. The locations of the different DNA sites for MelR are indicated by triangles: each triangle indicates an 18 bp sequence and the position of the centre of each site is numbered (see Belyaeva et al., 2000). The filled triangles indicate Site 1 and Site 2, the grey triangles denote Site 1′ and Site R, and the open triangle denotes Site 2′ that is filled by MelR only in the presence of melibiose. The shaded boxes denote 22 bp DNA sites for CRP: one site centered at position −41.5 is responsible for the activation of the melR promoter, while the other, centered at position −155.5 is involved in activation of the melAB promoter. The complete base sequence of the KK81 fragment is given in Fig. 1B of the accompanying paper (Belyaeva et al., 2000).

Download figure to PowerPoint

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Measurement of melR promoter activity

In our previous study, we described the KK81 EcoRI–HindIII fragment carrying the starts of the divergent E. coli melR and melA genes (Belyaeva et al., 2000; Fig. 1). In order to study the melR promoter on the KK81 fragment, we constructed a new vector, pRW70, by altering the polylinker of the broad host range, low-copy number, lac expression vector, pRW50 (see Experimental procedures and Table 1). Cloning of the KK81 EcoRI–HindIII fragment into pRW70 resulted in a melR promoter::lac fusion. High levels of β-galactosidase expression were measured with this fusion in the WAM132 Δlac ΔmelR strain of E. coli, compared with very low levels of expression with the starting pRW70 vector (Fig. 2). In preliminary experiments, we found that expression of the melR promoter::lac fusion is reduced to background levels by single base changes in the melR promoter −10 hexamer or by introduction of a Δcrp allele, showing that expression is due to the melR promoter (C. L. Webster and S. Busby, unpublished data). Next, we investigated the effects of introducing melR (carried by plasmid pJW15, a pAA121 derivative) on expression of the melR promoter::lac fusion in WAM132 ΔmelR cells. Results in Fig. 2 show that β-galactosidase expression in cells carrying pJW15 is several fold lower than in cells carrying the control pAA121 plasmid. This reduction is seen in both the presence and absence of melibiose, but the reduction in expression is significantly greater in the absence of melibiose than in its presence. From our results, we conclude that MelR can downregulate expression from the melR promoter.

Table 1. Plasmids and DNA fragments used in this study.
Plasmids  
pAA121Multicopy cloning vector derived from pBR322. Encodes resistance 80 µg ml−1 ampicillin. Kelsall et al. (1985)
pJW15Derivative of pAA121 with insert carrying melR expressed from melR promoter. Williams et al. (1994)
pRW50Low copy number lac expression vector. Carries EcoRI–BamHI–HindIII polylinker Encodes resistance to 35 µg ml−1 tetracycline. Lodge et al. (1992)
pRW70Derivative of pRW50 with BamHI–HindIII–EcoRI linker.This work
DNA fragments (all EcoRI–HindIII fragments)
KK81Carries melR and melAB promoters. Bounded by EcoRI site at +76 and HindIII site at −272 (with respect to melR transcript start) Belyaeva et al. (2000)
KK81 R1–R4 Derivatives of KK81 carrying single base changes in MelR binding Site R (see Fig. 3)This work
KK81S1 Derivative of KK81 with Site R altered to: 5′ TTCGTAGGAATATCAGAA 3′ (changes underlined)This work
KK81S2 Derivative of KK81 with −10 hexamer of melAB promoter changed from CATGAT to CGTGATThis work
KK81S3Derivative of KK81 with Site 1 altered to:This work
5′ ATCGTAATAAACTCAGAT 3′ (changes underlined) 
KK81S4 Derivative of KK81 with Site 1′ altered to: 5′ ATCGTATGAAAAGCAGAG 3′ (changes underlined)This work
KK81S5 Derivative of KK81 with Site 2 altered to: 5′ ATCGTAATAAACTCAGAT 3′ (changes underlined)This work
KK101 Carries melR promoter. Bounded by EcoRI site at +76 and HindIII site at −74 (with respect to melR transcript start)This work
image

Figure 2. Inhibition of the melR promoter by MelR.The figure shows β-galactosidase expression (in Miller units) in WAM132 ΔmelR Δlac E. coli cells containing a melR promoter::lac fusion carried by pRW70 with the KK81 fragment insert. Cells carried either pJW15 that encodes MelR, or the control pAA121 plasmid with no melR insert. As a control, the figure shows results with pRW70 vector carrying no promoter insert. Assays were performed in media either with or without melibiose as indicated. The results derive from over 10 independent determinations.

Download figure to PowerPoint

MelR-dependent repression of the melR promoter depends on Site R

WAM132 Δlac ΔmelR cells carrying pRW70 containing the KK81 fragment are phenotypically Lac plus on MacConkey lactose indicator plates (i.e. they give deep-red colonies). However, cells that also contain pJW15, encoding melR, score clearly as Lac minus, due to the MelR-dependent repression of the melR promoter (i.e. they give pale pink–white colonies). To pinpoint the DNA sequences in the KK81 fragment responsible for this repression, we used error-prone PCR to synthesise a library of mutant KK81 EcoRI–HindIII fragments. These fragments were cloned into pRW70 and recombinants were transformed into WAM132 cells carrying pJW15. The resulting transformants were screened for their Lac phenotype; from four different experiments, we screened 500 colonies and selected four mutant pRW70 derivatives carrying the KK81 fragment that gave a Lac-plus phenotype (denoted KK81 R1–R4). We reasoned that MelR-dependent repression of the melR promoter must be less efficient in these mutants.

The entire base sequence of the KK81 insert in each mutant was determined. This analysis showed that, in each of the four cases, the KK81 fragment carries a single base change in the Site R MelR-binding site (Fig. 3). The activity of the melR promoter carried by each mutant KK81 fragment was then measured in vivo in the presence and absence of MelR. Results in Fig. 3 show that, in each case, the single base substitution in the Site R MelR-binding site results in a significant reduction in MelR-dependent repression of the melR promoter in both the absence and presence of melibiose. From this, we conclude that the MelR-binding Site R plays an essential role in MelR-dependent repression of the melR promoter. To confirm this, we used site directed mutagenesis to create a further KK81 derivative (KK81S1), that carried a four base substitution, previously shown to substantially reduce MelR binding to Site 1 and Site 2 at the melAB promoter (Williams et al., 1994). Again, this change in the base sequence of Site R resulted in a substantial reduction in MelR-dependent repression of the melR promoter (Fig. 3). Electromobility shift assays showed that, while MelR can bind to Site R in the starting KK81 fragment, MelR is unable to bind to Site R in the KK81S1 mutant (J. T. Wade, unpublished data).

image

Figure 3. Base sequences of different MelR binding sequences.The figure lists the 18 bp MelR-binding sequence found both at Site 1 and Site 2, the 18 bp MelR-binding Site R, and the Site R sequence found in four KK81 mutant derivatives (KK81 R1–R4) selected after mutagenesis by error-prone PCR. The bottom line of the figure shows the Site R sequence in the KK81S1 derivative, that carries four base changes created by site-directed mutagenesis. Bases that accord with the consensus Site 1 and Site 2 are shaded. The figure also shows the measured amount of MelR-dependent repression of the melR promoter on the KK81 fragment carrying the different Site R substitutions. This repression was determined by measurement of β-galactosidase expression in WAM132 ΔmelR Δlac cells containing the melR promoter::lac fusion carried by pRW70 with the different KK81 fragment inserts. Repression is expressed as the ratio of the measured β-galactosidase activity in cells carrying pJW15, that encodes MelR, to the measured β-galactosidase activity in cells carrying the control pAA121 plasmid, with no melR insert. This ratio was determined in media either with or without melibiose as indicated. The ratios are based on four independent determinations, with a standard error of ±20%.

Download figure to PowerPoint

Optimal MelR-dependent repression of the melR promoter depends on upstream sequences

Transcription initiation of the melR promoter is dependent on the binding of CRP to a 22 bp DNA site centered at position −41.5, but sequences further upstream are not essential (Webster et al., 1988). Thus, we constructed a shorter derivative of the KK81 EcoRI–HindIII fragment carrying the upstream HindIII site at position −74 (KK101; Fig. 1). This fragment was cloned into pRW70 and melR promoter activity was measured in the absence and presence of pJW15, encoding MelR. The results shown in Fig. 4 indicate that, with the shorter KK101 fragment, repression of the melR promoter by MelR differs from the repression found with the longer KK81 fragment. In the absence of melibiose, MelR-dependent repression of the melR promoter is more efficient with the longer KK81 fragment. However, in the presence of melibiose, the degree of repression by MelR is similar with the long KK81 and the short KK101 fragments.

image

Figure 4. Repression of the melR promoter on short- or long-cloned fragments.The figure shows the degree of MelR-dependent repression of the melR promoter carried on the long KK81 fragment or the short KK101 fragment. This repression was determined by measurement of β-galactosidase expression in WAM132 ΔmelR Δlac cells containing the melR promoter::lac fusion carried by pRW70 with either the KK81 or the KK101 fragment insert. Repression is expressed as the ratio of the β-galactosidase level in cells carrying pJW15, that encodes MelR, to the β-galactosidase level in cells carrying the control pAA121 plasmid, with no melR insert. This ratio was determined in media either with or without melibiose as indicated. The results are the average of six independent determinations. WAM132 ΔmelR Δlac cells carrying pRW70 with either the KK81 or the KK101 fragment insert contain, respectively, 500 units or 380 units of β-galactosidase, in the absence of melR.

Download figure to PowerPoint

In addition to the melR promoter, the long KK81 fragment also carries the divergent melAB promoter (Fig. 1). We reasoned that some part of the melAB promoter might be responsible for the high level of MelR-dependent repression of the melR promoter seen with the KK81 fragment. To investigate this, we constructed a series of KK81 derivatives carrying different mutations in the melAB promoter (Table 1) and measured the consequences of these mutations on MelR-dependent repression of the melR promoter in either the presence of absence of melibiose (Fig. 5). First, in KK81S2, the melAB−10 hexamer was altered from CATGAT to CGTGAT, a change that completely suppresses all melAB promoter activity (S. Busby, unpublished data): this change has little or no effect on repression of the melR promoter by MelR. Next, we made 4 bp changes at MelR binding Site 1, Site 1′ and Site 2 (KK81S3, S4 and S5: the four base changes were the same as those introduced into Site R to give the KK81S1 derivative: see Table 1). Results in Fig. 5 show that MelR-dependent repression of the melR promoter is little altered by the changes in Site 1 (KK81S3) or Site 1′ (KK81S4). However, with the changes in Site 2 (KK81S5), MelR-dependent repression in the absence of melibiose is greatly reduced. In contrast, the changes in Site 2 have very little effect on MelR-dependent repression in the presence of melibiose. Thus, when MelR-binding Site 2 is altered in the KK81 fragment, the pattern of MelR-dependent repression of the melR promoter is similar to that seen with the shorter KK101 fragment (Fig. 4). We conclude that MelR-binding Site 2 is the key sequence element upstream of the melR promoter that is needed for efficient repression of the melR promoter in the absence of melibiose.

image

Figure 5. Repression of the melR promoter carried on different KK81 fragments.The figure shows the degree of MelR-dependent repression of the melR promoter carried on the starting KK81 fragment and the mutant derivatives, S2–S5. This repression was determined by measurement of β-galactosidase expression in WAM132 ΔmelR Δlac cells containing the melR promoter::lac fusion carried by pRW70 containing the starting KK81 fragment or the S2–S5 derivatives. Repression is expressed as the ratio of the β-galactosidase level in cells carrying pJW15, that encodes MelR, to the β-galactosidase level in cells carrying the control pAA121 plasmid. This ratio was determined in media either with or without melibiose as indicated. The results are the average of three independent determinations. WAM132 ΔmelR Δlac cells carrying pRW70 with the KK81, KK81S2, KK81S3, KK 81S4 or KK81S5 inserts, contain, respectively, 620 units, 600 units, 750 units, 500 units or 570 units of β-galactosidase, in the absence of melR.

Download figure to PowerPoint

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Transcription initiation at the melR promoter is downregulated by MelR, which binds to an operator site that overlaps the transcription start point and the −10 region. MelR can occupy this site in both the presence and absence of melibiose, and this appears to provide the cell with a fairly simple autoregulatory mechanism to ensure that MelR protein is never overexpressed. However, the situation is not so simple, and, in the absence of melibiose, efficient repression requires DNA sequences upstream of the melR promoter; in particular, MelR-binding Site 2 is required. In the accompanying paper (Belyaeva et al., 2000), we showed that Site 2 is occupied by MelR in both the absence and presence of melibiose whereas the neighbouring Site 2′ is only occupied by MelR in the presence of melibiose. We concluded that the occupation of Site 2′ is essential for the activity of the melAB promoter and that this explains why this activity is completely dependent on melibiose. Together with the present results, this suggests a simple model (illustrated in Fig. 6) to explain how MelR-dependent repression of the melR promoter is modulated by the elements at the upstream melAB promoter. In the absence of melibiose, we propose that a MelR dimer simultaneously occupies Site R and Site 2, creating a repression loop. The presence of melibiose triggers a conformational change in MelR such that the dimer now occupies Site 2 and Site 2′ and activation of the melAB promoter ensues. Thus, in the presence of melibiose, repression of the melR promoter is solely due to the occupation of Site R; repression is no longer aided by the upstream sequences, and, consequently, is less efficient. According to this model, there are striking similarities in the regulation of the E. coli melR–melAB and araC–araBAD genes; in particular both MelR and AraC can form repression loops in the absence of their cognate inducer ligand. However, the detailed arrangement and orientation of the cognate sites differs between the two systems. Additionally, at the araC promoter, separate and distinct DNA sites for AraC, Sites O1 and O2, are required for autoregulation in the absence and presence of arabinose respectively (Schleif, 1996). We assume that bacteria have evolved many different ways of deploying transcription factors that can be interconverted from repressor to activator by small ligands. We anticipate that the detailed study of other members of the AraC family will reveal yet more interesting subtleties in the way that regulatory regions can be organized.

image

Figure 6. Model for MelR-dependent repression of the melR promoter.In the absence of melibiose, MelR subunits occupy Site R and Site 2 forming a repression loop. The presence of melibiose induces a conformation change in MelR such that it occupies Site 2 and the neighbouring Site 2′, and the repression loop is broken. MelR bound at Site 2′ is able to interact with RNAP and promotes transcription initiation at the melAB promoter. In the presence of melibiose, repression is solely due to the occupation of Site R and is no longer dependent on the upstream sequences.

Download figure to PowerPoint

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Strains, plasmids and promoters

In this work we used the WAM132 Δlac ΔmelR strain of E. coli (Belyaeva et al., 2000). The plasmids and DNA fragments carrying the E. coli mel operon regulatory region used in this study are listed in Table 1. To make pRW70, two complementary synthetic oligos D17920 (5′ AATTGGATCCCAAGCTTGCGGAATTCGC 3′) and D17921 (5′ AGCTGCGAATTCCGCAAGCTTGGGATCC 3′) were annealed and cloned between the EcoRI and HindIII sites of pRW50. The consequence of this is to change the EcoRI–BamHI–HindIII polylinker of pRW50 into a BamHI–HindIII–EcoRI polylinker.

The starting point of this work was the EcoRI–HindIII KK81 fragment described by Belyaeva et al. (2000), illustrated in Fig. 1. By convention, in this paper, sequences are numbered with respect to the transcription start point of the melR promoter with upstream and downstream locations denoted by ‘–’ and ‘+’ prefixes respectively. PCR was used to construct the shorter KK101 fragment and five KK81 derivatives (S1–S5; Table 1). KK81S1 contains four base changes that disrupt MelR binding Site R, KK81S2 carries a single base substitution in the melAB−10 hexamer, and KK81S3, S4 and S5 each carry four base pair changes that disrupt MelR binding Site 1, Site 1′ and Site 2 respectively. Error-prone PCR (Zhou et al., 1991) was used to construct a library of mutant KK81 derivatives, from which we selected mutants where MelR-dependent repression of the melR promoter was reduced (KK81 R1, R2, R3 and R4; Fig. 3). For simplicity, this experiment was performed in the absence of melibiose.

Measurement of promoter activities in vivo

DNA fragments containing the melR promoter were cloned into pRW70 to generate pmelR::lac fusions. The levels of β-galactosidase in cells carrying these recombinants were measured by the Miller (1972) method; cells were grown in media either with or without melibiose exactly as in our previous work (Webster et al., 1987). Assays were performed in the WAM132 Δlac ΔmelR host strain carrying either plasmid pJW15 encoding melR or the control ‘empty’ vector plasmid pAA121. Note that, in pJW15, melR is expressed from the melR promoter, but that the upstream limit of melR promoter sequence carried by pJW15 is at position −59 upstream of the melR promoter transcript start (Webster et al., 1988; Williams et al., 1994). Thus, the melR promoter is not efficiently repressed by MelR and MelR is overexpressed.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

This work was generously supported by the UK BBSRC with a project grant G09595 to S.J.W.B. and E.I.H. and a PhD studentship award to J.T.W. We are grateful to C. Webster for constructing pRW70.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Belyaeva, T., Wade, J., Webster, C., Howard, V., Thomas, M., Hyde, E., et al. (2000) Transcription activation at the Escherichia coli melAB promoter: the role of MelR and the cyclic AMP receptor protein. Mol Microbiol 36 : 211–222.
  • Caswell, R., Webster, C., Busby, S. (1992) Studies on the binding of the E coli MelR transcription activator protein to operator sequences at the melAB promoter. Biochem J 287: 501508.
  • Choy, H. & Adhya, S. (1996) Negative control. In Escherichia Coli and Salmonella, Vol 1. Neidhardt, F. (ed. ). Washington: American Society for Microbiology Press, pp. 12871299.
  • Gallegos, M.-T., Schleif, R., Bairoch, A., Hofmann, K., Ramos, J.-L. (1997) AraC/XylS family of transcriptional regulators. Microbiol Mol Biol Rev 61: 393410.
  • Gostick, D., Green, J., Irvine, A., Gasson, M., Guest, J. (1998) A novel regulatory switch mediated by the FNR-like protein of Lactobacillus casei. Microbiology 144: 705717.
  • Gralla, J. & Collado-Vides, J. (1996) Organization and function of transcription regulatory elements. In Escherichia coli and Salmonella, Vol 1. Neidhardt, F. (ed. ). Washington: American Society for Microbiology Press, pp. 13321245.
  • Kelsall, A., Evans, C., Busby, S. (1985) A plasmid vector that allows fusion of the E. coli galactokinase gene to the translation startpoint of other genes. FEBS Lett 180: 155159.
  • Lodge, J., Fear, J., Busby, S., Gunasekaran, P., Kamini, N.-R. (1992) Broad host-range plasmids carrying the E. coli lactose and galactose operons. FEMS Lett 95: 271276.
  • Miller, J. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
  • Muller-Hill, B. (1998) Some repressors of bacterial transcription. Curr Opinion Microbiol 1: 145151.
  • Schleif, R. (1996) Two positively regulated systems, ara and mal. In Escherichia coli and Salmonella, Vol 1. Neidhardt, F. (ed. ). Washington: American Society for Microbiology Press, pp. 13001309.
  • Webster, C., Gardner, L., Busby, S. (1989) The E. coli melR gene encodes a DNA binding protein with affinity for specific sequences located in the melibiose-operon regulatory region. Gene 83: 207213.
  • Webster, C., Gaston, K., Busby, S. (1988) Transcription from the E. coli melR promoter is dependent on the cyclic AMP receptor protein. Gene 68: 297305.
  • Webster, C., Kempsell, K., Booth, I., Busby, S. (1987) Organisation of the regulatory region of the Escherichia coli melibiose operon. Gene 59: 253263.
  • Williams, J., Michán, C., Webster, C., Busby, S. (1994) Interactions between the E coli MelR transcription activator protein and operator sequences at the melAB promoter. Biochem J 300: 757763.
  • Zhou, Y., Zhang, X., Ebright, R. (1991) Random mutagenesis of gene-sized molecules by use of PCR with Taq DNA polymerase. Nucleic Acids Res 19: 6052.