Molecular Characterization of GrlA, a Specific Positive Regulator of ler Expression in Enteropathogenic Escherichia coli

Abstract

Enteropathogenic Escherichia coli (EPEC) infections are characterized by the formation of attaching and effacing (A/E) lesions on the surfaces of infected epithelial cells. The genes required for the formation of A/E lesions are located within the locus of enterocyte effacement (LEE). Ler is the key regulatory factor controlling the expression of LEE genes. Expression of the ler gene is positively regulated by GrlA, which is encoded by the LEE. Here, we analyze the mechanism by which GrlA positively regulates ler expression and show that in the absence of H-NS, GrlA is no longer essential for ler activation, further confirming that GrlA acts in part as an H-NS antagonist on the ler promoter. Single-amino-acid mutants were constructed to test the functional significance of the putative helix-turn-helix (HTH) DNA binding motif found in the N-terminal half of GrlA, as well as at the C-terminal domain of the protein. Several mutations within the HTH motif, but not all, completely abolished GrlA activity, as well as specific binding to its target sequence downstream from position −54 in the ler regulatory region. Some of these mutants, albeit inactive, were still able to interact with the negative regulator GrlR, indicating that loss of activity was not a consequence of protein misfolding. Additional residues in the vicinity of the HTH domain, as well as at the end of the protein, were also shown to be important for GrlA activity as a transcriptional regulator, but not for its interaction with GrlR. In summary, GrlA consists of at least two functional domains, one involved in transcriptional activation and DNA binding and the other in heterodimerization with GrlR.

Enteropathogenic Escherichia coli (EPEC) is an important human pathogen that causes childhood diarrhea in developing countries. EPEC belongs to the attaching and effacing (A/E) family of bacterial enteropathogens, together with enterohemorrhagic E. coli (EHEC) and the mouse pathogen Citrobacter rodentium (35). The A/E pathogens intimately adhere to the membranes of epithelial cells, inducing the destruction of the brush border microvilli and the reorganization of the actin cytoskeleton into characteristic pedestal-like protrusions underneath adherent bacteria. Intimate attachment is mediated by the interaction of intimin, an outer membrane protein, with the translocated intimin receptor (Tir), a bacterial effector protein that is injected into the host cells by a specialized type III secretion system (T3SS) and integrated into the cytoplasmic membrane (19, 26).

The genes required for the generation of the A/E lesion are within a pathogenicity island (PAI) known as the locus of enterocyte effacement (LEE), which comprises 41 genes arranged in five polycistronic operons (LEE1 to LEE5) and other transcriptional units (18). The LEE1 to LEE3 operons encode proteins necessary for the assembly of the T3SS. LEE4 encodes proteins called translocators, responsible for forming a conduit and a channel at the eukaryotic plasma membrane through which effector proteins are translocated from the bacterial cytoplasm to the host cell cytosol. The LEE5 operon contains the genes coding for the proteins involved in intimate adherence, Tir and intimin. Seven effectors are encoded in the LEE (Tir, Map, EspF, EspG, EspZ, EspH, and EspB). These proteins are responsible for subverting signaling pathways within the enterocyte, resulting in cytoskeletal rearrangements, disruption of the tight junctions, and altered absorption of nutrients and ions, leading to diarrhea (14, 21, 22). Recent studies have shown that several genes encoding putative effectors are scattered along the core genome of A/E pathogens. In EHEC and EPEC, respectively, 43 and 14 functional genes encoding putative type III secreted non-LEE-encoded effectors have been identified (30, 63).

As the LEE is an essential virulence determinant of A/E pathogenesis, the expression of its genes is under the control of a complex assortment of positive and negative transcriptional regulators (47, 52, 66) and a series of posttranscriptional events (5, 25, 36, 41, 58). It has been well documented that the master positive regulator of LEE genes, as well as of genes located outside the LEE, is the LEE-encoded regulator (Ler), whose main function is to counteract the repression exerted by H-NS on these genes (1, 3, 9, 17, 23, 39, 48, 57, 62, 65, 69). It has also been reported that Ler represses its own expression in a concentration-dependent fashion (4) and that it can also act as a negative regulator for other genes outside the LEE (1, 61). The key role of Ler in LEE regulation makes its gene the main direct target of both positive and negative global regulators, such as H-NS, Hha, IHF, and QseA (9, 20, 59, 60, 69). In addition, several other proteins have been shown to regulate the expression of LEE genes by controlling ler expression or other LEE promoters directly, as in the cases of BipA, GadX, GrvA, EivF, EtrA, Fis, LexA, LrhA, RcsCDB, RegA, SdiA, YhiE, and YhiF (28, 47, 66, 71).

Despite the complexity of the network of proteins involved in LEE gene regulation, two proteins, GrlA and PerC/PchABC, have been shown to specifically derepress and enhance ler transcription, even in an E. coli K-12 laboratory strain (3, 9, 48, 53). PerC is encoded by the perABC operon in the EPEC adherence factor plasmid (pEAF), which is autoregulated by PerA (44), the product of the first per gene, which in turn also controls the expression of the bfp operon (54, 64). PerC-homologous (Pch) proteins have been shown to play a key role in LEE gene expression in EHEC by directly activating ler expression (32). In contrast to EPEC perC, these genes are found within prophages that are scattered in the EHEC chromosome (32, 53, 71).

GrlA was first identified as a positive regulator of the LEE in C. rodentium, where it was also shown to be essential in vivo (15), and was further shown to do so by controlling the expression of Ler, which in turn also controls the expression of the grlRA operon, forming a positive regulatory loop (3). The role of GrlA in the transcriptional activation of the LEE has also been demonstrated in EPEC and EHEC (29, 36, 40). In addition, GrlA regulates the expression of genes located outside the LEE in EHEC, negatively, as in the case of the flagellar master operon flhDC (31), and positively, as in the case of the enterohemolysin ehxCABD operon (55). In all cases, the specific mechanism through which GrlA activates its target genes is unknown, but it has been proposed that GrlA could counteract the negative regulation exerted by H-NS upon binding to the regulatory region (V. H. Bustamante, M. I. Villalba, V. A. García-Angulo, A. Vázquez, L. C. Martínez, R. Jiménez, and J. L. Puente, submitted for publication).

The gene encoding GrlA is found in all the LEE sequences currently reported, and homologues are present in a few other genomes, such as in Salmonella enterica, Yersinia bercovieri, and Photorhabdus luminicens, sharing approximately 37%, 29%, and 25% identity, respectively, mainly in the N-terminal half of the protein, where a helix-turn-helix (HTH) putative DNA binding motif is found (15). In addition, GrlA shares 23% identity with CaiF, a partially characterized positive transcriptional regulator of the cai and fix operons that are involved in anaerobic carnitine metabolism in E. coli and other Enterobacteriaceae (7).

grlA is cotranscribed with grlR, a gene coding for a negative regulator of LEE gene expression (3, 15, 29, 31, 33, 40). Using a yeast two-hybrid system, it was reported earlier that these two proteins interact with each other, although their functions were unknown at the time (12). This interaction has been confirmed experimentally using biochemical techniques (29, 33, 37). Based on this feature, it has been proposed that GrlR represses LEE gene expression by binding to GrlA, thus preventing its interaction with the ler promoter.

Due to its important role in A/E pathogenesis, in this work, we analyzed the mechanism by which GrlA activates ler expression in EPEC and further substantiated the notion that GrlA acts in part as an H-NS antagonist by binding specifically in the vicinity of the ler promoter. Several amino acids from the putative HTH motif of GrlA were identified as essential for ler activation and DNA binding, while some of them were also needed to interact with GrlR. Furthermore, we found that the first 100 amino acids of the protein are sufficient for GrlA-GrlR heterodimerization, but not for ler expression. In addition, charged residues at the C-terminal end of GrlA, as well as at the second predicted helix at the N terminus of the HTH motif, were also important for gene expression. In summary, we have characterized single point mutations that produced GrlA mutants showing diverse phenotypes. This study represents the first systematic study providing evidence of the functional and structural organization of GrlA and the differential roles of different amino acid residues in DNA binding/transcriptional activation and protein-protein interactions.

MATERIALS AND METHODS

Bacterial strains, plasmids, media, and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. Bacteria were grown in Luria-Bertani (LB) broth or agar or in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% LB. When antibiotics were necessary, they were used at the following concentrations in LB and at half the indicated concentrations in DMEM: ampicillin, 200 μg/ml; streptomycin, 100 μg/ml; kanamycin, 30 μg/ml; tetracycline, 12.5 μg/ml. For chloramphenicol acetyltransferase (CAT) activity or Western blot assays, bacterial samples were collected from shaken or static (5% CO₂ atmosphere) cultures grown at 37°C in 50 ml of DMEM supplemented with 1% LB.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain	Relevant characteristics	Reference or source
EPEC
E2348/69	WT	38
JPEP24	E2348/69 Δler::Km	Bustamante et al., submitted
JPEP35	E2348/69 carrying an in-frame deletion of grlA	A. Huerta, unpublished
JPEP36	E2348/69 Δhns::Km	A. Vázquez, unpublished
JPEP37	E2348/69 carrying mutants in ΔgrlA and Δhns::Km	This study
JPEP38	JPEP37 lacking the EAF plasmid	This study
JPEP30	E2348/69 lacking the EAF plasmid	Bustamante et al., submitted
E. coli
MC4100	F−araD139 D(argF-lac)U169 rpsL150 (Strr) relA1 flbB5301 deoC1 ptsf25 rbsR	10
JPMC1	MC4100 carrying an in-frame deletion of Δhns	3
BL21 pLys	F−ompT (lon) hsdSB(rB− mB−) gal dcm (lDE3) pLysS(Cmr)	Invitrogen
Plasmids
pKD46	Plasmid expressing the lambda Red recombinase; Apr	13
pKD4	Template plasmid containing the Km cassette	13
pler-260	pKK232-8 derivative containing a ler-cat transcriptional fusion from nucleotides −260 to +217 with respect to the transcriptional start site	Bustamante et al., submitted
pT6-HNS	pMPM-T6Ω derivative expressing WT H-NS	8
pT6-HNS/G113D	pMPM-T6Ω derivative expressing H-NSG113D	8
pMPM-T3	Low-copy-number cloning vector; Tcr	45
pTEPGrlA1	pMPM-T3 derivative containing the structural grlA gene including its ribosomal binding site (RBS) expressed under the lac promoter	Bustamante et al., submitted
pTEPGrlA1/P23A	pTEPGrlA1 derivative expressing GrlA P23A	This study
pTEPGrlA1/L24A	pTEPGrlA1 derivative expressing GrlA L24A	This study
pTEPGrlA1/Y25A	pTEPGrlA1 derivative expressing GrlA Y25A	This study
pTEPGrlA1/S29A	pTEPGrlA1 derivative expressing GrlA S29A	This study
pTEPGrlA1/W31A	pTEPGrlA1 derivative expressing GrlA W31A	This study
pTEPGrlA1/C32A	pTEPGrlA1 derivative expressing GrlA C32A	This study
pTEPGrlA1/R41A	pTEPGrlA1 derivative expressing GrlA R41A	This study
pTEPGrlA1/N42A	pTEPGrlA1 derivative expressing GrlA N42A	This study
pTEPGrlA1/I44A	pTEPGrlA1 derivative expressing GrlA I44A	This study
pTEPGrlA1/I44S	pTEPGrlA1 derivative expressing GrlA I44S	This study
pTEPGrlA1/E46A	pTEPGrlA1 derivative expressing GrlA E46A	This study
pTEPGrlA1/F48A	pTEPGrlA1 derivative expressing GrlA F48A	This study
pTEPGrlA1/F48S	pTEPGrlA1 derivative expressing GrlA F48S	This study
pTEPGrlA1/I50A	pTEPGrlA1 derivative expressing GrlA I50A	This study
pTEPGrlA1/I50S	pTEPGrlA1 derivative expressing GrlA I50S	This study
pTEPGrlA1/L52A	pTEPGrlA1 derivative expressing GrlA L52A	This study
pTEPGrlA1/R54A	pTEPGrlA1 derivative expressing GrlA R54A	This study
pTEPGrlA1/R54S	pTEPGrlA1 derivative expressing GrlA R54S	This study
pTEPGrlA1/S56A	pTEPGrlA1 derivative expressing GrlA S56A	This study
pTEPGrlA1/Y61A	pTEPGrlA1 derivative expressing GrlA Y61A	This study
pTEPGrlA1/R65A	pTEPGrlA1 derivative expressing GrlA R65A	This study
pTEPGrlA1/R132A	pTEPGrlA1 derivative expressing GrlA R132A	This study
pTEPGrlA1/R133A	pTEPGrlA1 derivative expressing GrlA R133A	This study
pTEPGrlA1/K134A	pTEPGrlA1 derivative expressing GrlA K134A	This study
pTEPGrlA1/K135A	pTEPGrlA1 derivative expressing GrlA K135A	This study
pTEPGrlA1/E136A	pTEPGrlA1 derivative expressing GrlA E136A	This study
pTEPGrlA1/Δ2-20	pTEPGrlA1 derivative expressing GrlA ΔE2-D20	This study
pTEPGrlA1/Δ21-40	pTEPGrlA1 derivative expressing GrlA ΔG21-S40	This study
pTEPGrlA1/Δ41-60	pTEPGrlA1 derivative expressing GrlA ΔR41-T60	This study
pTEPGrlA1/Δ61-80	pTEPGrlA1 derivative expressing GrlA ΔY61-N80	This study
pTEPGrlA1/Δ81-100	pTEPGrlA1 derivative expressing GrlA ΔL81-I100	This study
pTEPGrlA1/Δ101-120	pTEPGrlA1 derivative expressing GrlA ΔE101-V120	This study
pTEPGrlA1/Δ121-137	pTEPGrlA1 derivative expressing GrlA ΔG121-S137	This study
pTEPGrlA1/Δ125-137	pTEPGrlA1 derivative expressing GrlA ΔI125-S137	This study
pTEPGrlA1/Δ131-137	pTEPGrlA1 derivative expressing GrlA ΔL131-S137	This study
pGEX-4T1	Cloning vector for constructing GST fusions	Amersham Bioscience
pGST-GrlR	pGEX-4T1 derivative expressing GST-GrlR under the control of the IPTG-inducible promoter ptac	C. Lara, unpublished data
pBAD-HNS-FLAGHis6	pBADMycHis derivative expressing H-NS-FLAGHis6 under the control of the arabinose-inducible pBAD promoter	V. Bustamante, unpublished data
pMAL-c2X	Cloning vector for constructing MBP fusions	New England Biolabs
pMBP-GrlA	pMAL-c2X derivative expressing MBP-GrlA under the control of IPTG-inducible promoter ptac	This study
pMBP-GrlA/P23A	pMBP-GrlA derivative expressing MBP-GrlA P23A	This study
pMBP-GrlA/L24A	pMBP-GrlA derivative expressing MBP-GrlA L24A	This study
pMBP-GrlA/Y25A	pMBP-GrlA derivative expressing MBP-GrlA Y25A	This study
pMBP-GrlA/S29A	pMBP-GrlA derivative expressing MBP-GrlA S29A	This study
pMBP-GrlA/W31A	pMBP-GrlA derivative expressing MBP-GrlA W31A	This study
pMBP-GrlA/C32A	pMBP-GrlA derivative expressing MBP-GrlA C32A	This study
pMBP-GrlA/R41A	pMBP-GrlA derivative expressing MBP-GrlA R41A	This study
pMBP-GrlA/N42A	pMBP-GrlA derivative expressing MBP-GrlA N42A	This study
pMBP-GrlA/I44A	pMBP-GrlA derivative expressing MBP-GrlA I44A	This study
pMBP-GrlA/I44S	pMBP-GrlA derivative expressing MBP-GrlA I44S	This study
pMBP-GrlA/E46A	pMBP-GrlA derivative expressing MBP-GrlA E46A	This study
pMBP-GrlA/F48A	pMBP-GrlA derivative expressing MBP-GrlA F48A	This study
pMBP-GrlA/F48S	pMBP-GrlA derivative expressing MBP-GrlA F48S	This study
pMBP-GrlA/I50A	pMBP-GrlA derivative expressing MBP-GrlA I50A	This study
pMBP-GrlA/I50S	pMBP-GrlA derivative expressing MBP-GrlA I50S	This study
pMBP-GrlA/L52A	pMBP-GrlA derivative expressing MBP-GrlA L52A	This study
pMBP-GrlA/R54A	pMBP-GrlA derivative expressing MBP-GrlA R54 A	This study
pMBP-GrlA/R54S	pMBP-GrlA derivative expressing MBP-GrlA R54S	This study
pMBP-GrlA/S56A	pMBP-GrlA derivative expressing MBP-GrlA S56A	This study
pMBP-GrlA/Y61A	pMBP-GrlA derivative expressing MBP-GrlA Y61A	This study
pMBP-GrlA/R65A	pMBP-GrlA derivative expressing MBP-GrlA R65A	This study
pMBP-GrlA/R132A	pMBP-GrlA derivative expressing MBP-GrlA R132A	This study
pMBP-GrlA/R133A	pMBP-GrlA derivative expressing MBP-GrlA R133A	This study
pMBP-GrlA/K134A	pMBP-GrlA derivative expressing MBP-GrlA K134A	This study
pMBP-GrlA/K135A	pMBP-GrlA derivative expressing MBP-GrlA K135A	This study
pMBP-GrlA/E136A	pMBP-GrlA derivative expressing MBP-GrlA E136A	This study
pMBP-GrlA/Δ2-20	pMBP-GrlA derivative expressing MBP-GrlA ΔE2-D20	This study
pMBP-GrlA/Δ21-40	pMBP-GrlA derivative expressing MBP-GrlA ΔG21-S40	This study
pMBP-GrlA/Δ41-60	pMBP-GrlA derivative expressing MBP-GrlA ΔR41-T60	This study
pMBP-GrlA/Δ61-80	pMBP-GrlA derivative expressing MBP-GrlA ΔY61-N80	This study
pMBP-GrlA/Δ81-100	pMBP-GrlA derivative expressing MBP-GrlA ΔL81-I100	This study
pMBP-GrlA/Δ101-120	pMBP-GrlA derivative expressing MBP-GrlA ΔE101-V120	This study
pMBP-GrlA/Δ121-137	pMBP-GrlA derivative expressing MBP-GrlA ΔG121-S137	This study
pMBP-GrlA/Δ125-137	pMBP-GrlA derivative expressing MBP-GrlA ΔI125-S137	This study
pMBP-GrlA/Δ131-137	pMBP-GrlA derivative expressing MBP-GrlA ΔL131-S137	This study

DNA manipulations.

DNA manipulations were performed according to standard protocols (56). Restriction enzymes were obtained from Invitrogen and used according to the manufacturer's instructions. The oligonucleotides used in this work are listed in Table 2 and were synthesized at the oligonucleotide synthesis facility of the Instituto de Biotecnología/UNAM.

TABLE 2.

Oligonucleotides used in this study

Oligonucleotide	Sequencea
ler260F	CTCCTGGggatccACTCGCT
Orf1-H3-R	GCTCTATaagcttAATGTATG
ler50BHIFW	ATCATggatccTAAATGGATTTTAAAAA
ler-50 RV	AAATCCATTTAAAATCAATG
dnaJ FW	GGCGGCGGTTTTGGCGGCGGC
dnaJ RV	AGTCTGCGGCTGTGTACCTGG
EPCiOrf11R	TACTAAGAaagcttCGTCTAACTCTCC
EPGA-XI	GCCAAATTTctcgagCCATTAATTAT
MBPCrgrlAF	ATAAAAAGAACAtctagaATGGAATCTAAA
EPAP23A-F	ACGATGGTGAAgCTCTGTATATCTTGG
EPAP23A-R	CCAAGATATACAGAGcTTCACCATCGT
EPAL24A-F	CGATGGTGAACCTgcGTATATCTTGGTT
EPAL24A-R	AACCAAGATATACgcAGGTTCACCATCG
EPAY25A-F	GGTGAACCTCTGgcTATCTTGGTTTCTC
EPAY25A-R	GAGAAACCAAGATAgcCAGAGGTTCACC
EPGAS29AF	TATATCTTGGTTgCTCTTTGGTGTAAA
EPGAS29AR	TTTACACCAAAGAGcAACCAAGATATA
EPAW31A-F	CTTGGTTTCTCTTgcGTGTAAATTGCAGG
EPAW31A-R	CCTGCAATTTACACgcAAGAGAAACCAAG
EPGAC32AF	GTTTCTCTTTGGgcTAAATTGCAGGAG
EPGAC32AR	CTCCTGCAATTTAgcCCAAAGAGAAAC
EPGAR41AF	AAATGGATTTCTgctAATGATATTGCC
EPGAR41AR	GGCAATATCATTagcAGAAATCCATTT
EPAN42A-F	TGGATTTCTCGCgcTGATATTGCCGAAG
EPAN42A-R	CTTCGGCAATATCAgcGCGAGAAATCCA
EPAI44A-F	TCTCGCAATGATgcTGCCGAAGCATTC
EPAI44A-R	GAATGCTTCGGCAgcATCATTGCGAGA
EPAE46A-F	GCAATGATATTGCCgcAGCATTCGGTATA
EPAE46A-R	TATACCGAATGCTgcGGCAATATCATTGC
EPAF48A-F	ATTGCCGAAGCAgcCGGTATAAACCTG
EPAF48A-R	CAGGTTTATACCGgcTGCTTCGGCAAT
EPAI50A-F	GAAGCATTCGGTgcAAACCTGAGGAGA
EPAI50A-R	TCTCCTCAGGTTTgcACCGAATGCTTC
EPAL52A-F	TTCGGTATAAACgcGAGGAGAGCATCA
EPAL52A-R	TGATGCTCTCCTCgcGTTTATACCGAA
EPAR54A-F	ATAAACCTGAGGgcAGCATCATTTATT
EPAR54A-R	AATAAATGATGCTgcCCTCAGGTTTAT
EPAR56S-F	CCTGAGGAGAGCAgcATTTATTATAAC
EPAR56S-R	GTTATAATAAATgcTGCTCTCCTCAGG
EPGAY61AF	TTTATTATAACTgctATATCGAGAAGA
EPGAY61AR	TCTTCTCGATATagcAGTTATAATAAA
EPGAR65AF	TATATATCGAGAgctAAAGAAAAAATT
EPGAR65AR	AATTTTTTCTTTagcTCTCGATATATA
XhxbgrlAF	CGCGGctcgagGAGGAtctagaATGGAATCTAAAA
HigrlAR	ACCCGGGaagcttCGTCTAACTCTCCTT
1-20delgrlAF	CGCGGctcgagGAGGAtctagaATGGGTGAACCTCTGT
21-40delgrlAF	GAATTACGATCGCAATGATATTGCCGAAGC
21-40delgrlAR	TATCATTGCGATCGTAATTCTTTACTGAGT
41-60delgrlAF	ATGGATTTCTTATATATCGAGAAGAAAAGA
41-60delgrlAR	TCGATATATAAGAAATCCATTTCTCCTGCA
61-80delgrlAF	TATTATAACTTTGCATTATAAGCGCCTTGA
61-80delgrlAR	TATAATGCAAAGTTATAATAAATGATGC
81-100delgrlAF	TTATGGTAATGAAAGTCCTGGATCAACCGG
81-100delgrlAR	CAGGACTTTCATTACCATAACTAACATATC
101-120delgrlAF	GGTTCCGATAGGACAGTCTAATATCTGGAA
101-120delgrlAR	TAGACTGTCCTATCGGAACCGCCTCAAGGT
121-137delgrlAR	CCCGGGaagcttCTACACAATACCATTA
EPGAΔ125R	TCATaagcttctaGATATTGAGCTGTCC
EPGAΔ131R	CCTTaagcttctaCAAGATCATTTCGTT
EPGAR132AR	CCTTaagcttctaACTCTCCTTTTTCCGCGCCAAGATCAT
EPGAR133AR	CCTTaagcttctaACTCTCCTTTTTCGCCCTCAAGAT
EPGAK134AR	CCTTaagcttctaACTCTCCTTTGCCCGCCTCAA
EPGAK135AR	CCTTaagcttctaACTCTCCGCTTTCCGCTT
EPGAE136AR	CCTTaagcttctaACTCGCCTTTTTCCG
PMPM3 FW1	GTGCCGTAAAGCACTAAATCGG
PMPM3 RV1	GCGTTATCCCCTGATTCTGTGG
Ehns H1P1	CACCCCAATATAAGTTTGAGATTACTACAATGAGCGAAGCtgtaggctggagctgcttcg
Ehns H2P2	GATTTTAAGCAAGTGCAATCTACAAAAGATTATTGCTTcatatgaatatcctccttag

^aThe sequence in lowercases indicate nucleotides that were modified with respect to the WT sequence.

Construction of an EPEC grlA and hns double mutant.

The EPEC ΔgrlA Δhns double mutant was constructed by replacing the hns gene with a kanamycin cassette from a ΔgrlA EPEC strain by the one-step method using the lambda red recombinase system (13). The PCR fragment used was amplified with the oligonucleotides Ehns H1P1 and Ehns H2P2 (Table 2) using as a template DNA from plasmid pKD4. The replacement of the hns gene was confirmed by PCR.

Construction of GrlA site-directed mutants and deletions.

Site-specific mutations in grlA were introduced by overlapping PCR as previously described (27). Briefly, pairs of complementary oligonucleotides were designed with the desired changes (Table 2). Forward and reverse mutagenic primers were combined in parallel PCRs with primers PMPM3 FW1 and PMPM3 RV1, respectively, using pTEPGrlA1 plasmid DNA as a template. The resulting PCR products were purified and mixed for a second PCR round with primers EPGA-XI and EPCiOrf11R, which allow the amplification of the entire grlA gene. Mutations at the GrlA C terminus were introduced in a single PCR using the forward primer EPGA-XI and the corresponding mutagenic reverse primers, which carry the desired changes and a restriction site (Table 2).

Sequential in-frame deletions in grlA were generated by overlapping PCR, as described above, except for the second PCR, for which primers XhxbgrlAF and HigrlAR were used. Deletions ΔE2-D20, ΔG121-S137, ΔG125-S137, and ΔG131-S137 were generated in a single PCR using the following pairs of primers, respectively: 1-20delgrlAF/HigrlAR, XhxbgrlAF/121-137delgrlAR, EPGA-XI/EPGAΔ125R, and EPGA-XI/EPGAΔ131R. The resulting PCR products were purified and digested with XhoI and HindIII and cloned into pMPM-T3 digested with the same enzymes. Each construct was verified by DNA sequencing.

Protein secretion assay.

Fifty milliliters of DMEM supplemented with 1% LB was inoculated with 500 μl of an LB overnight culture of each strain and incubated at 37°C with agitation. Samples were collected when the cultures reached an optical density (OD) of 1.0. Culture supernatants were recovered after centrifugation of 1.5 ml of each culture for 5 min at 18,000 × g. The secreted proteins were precipitated from the culture supernatant with 10% trichloroacetic acid (TCA) at 4°C overnight and concentrated by centrifugation at 20,000 × g at room temperature. The resulting precipitated proteins were analyzed by 12% SDS-PAGE.

CAT assay.

Samples from shaken or static cultures were collected at an OD of 1 or after 6 h, respectively. CAT assays and protein quantification to determine CAT specific activities were performed as described previously (44).

Western blotting.

Whole-cell extracts were prepared by resuspending bacterial pellets from culture samples of the strains tested. Proteins were separated by 12% SDS-PAGE and transferred to a nitrocellulose membrane (Millipore) using a semidry transfer apparatus (Bio-Rad). The membranes were blocked in 5% nonfat milk and incubated with a 1:10,000 dilution of polyclonal anti-maltose binding protein (MBP) (New England Biolabs), a 1:10,000 dilution of monoclonal anti-Tir (kindly provided by B. B. Finlay), a 1:10,000 dilution of anti-EscJ (kindly provided by B. González-Pedrajo), a 1:10,000 dilution of anti-EspA (kindly provided by J. B. Kaper), and a 1:10,000 dilution of anti-DnaK (Invitrogen). The goat anti-rabbit immunoglobulin G (Pierce) and rabbit anti-mouse immunoglobulin G (Pierce) secondary antibodies were used at a 1:10,000 dilution. Positive signals were visualized using the Western Lightning Chemiluminescence Reagent Plus (Perkin-Elmer), following the manufacturer's instructions, and scientific imaging films (Kodak).

Pulldown assays.

Pulldown assays to analyze the interaction between GrlR and GrlA were carried out using as bait the GrlR protein fused to glutathione-S-transferase (GST-GrlR), whose expression is under the control of the tac promoter in plasmid pGEX-4T1 (Amersham Biosciences), and as prey the wild-type (WT) GrlA and its different mutants fused to MBP, also expressed from the tac promoter in plasmid pMAL-c2x (New England Biolabs). Expression of the GST-GrlR and MBP-GrlA proteins was induced for 4 h at 30°C with 0.5 and 0.3 mM isopropyl-β-d-thiogalactopyranoside (IPTG), respectively, in E. coli BL21 pLys grown in 50 ml of LB after reaching an OD of 0.6. Cells were collected by centrifugation at 10,000 × g at 4°C and resuspended in 4 ml of PBS (140 mM NaCl, 4.5 mM Na₂HPO₄, 1.5 mM KH₂PO₄, 2.3 mM KCl). Total cell extracts were prepared by sonication and centrifugation at 18,000 × g at 4°C. Three hundred microliters of the extract containing GST-GrlR was mixed with 50 μl of glutathione-Sepharose 4B beads (Amersham Biosciences) and incubated at 4°C for 1 h with agitation; unbound proteins were removed by washing them five times with 1 ml of cold PBS. One milliliter of the bacterial extracts containing MBP-GrlA proteins was added to the glutathione-Sepharose-GST-GrlR beads and incubated at 4°C for 2 h. Unbound proteins were removed by washing them five times with 1 ml of cold phosphate-buffered saline (PBS), and the resultant protein complexes were subjected to 12% SDS-PAGE.

Construction and purification of MBP-GrlA fusion proteins.

For constructing MBP-GrlA recombinant proteins, WT and mutant grlA genes were amplified by PCR using as templates DNA of the pTEPGrlA1-derived plasmids (Table 1) and primers MBPCrgrlAF and EPCiOrf11R. The PCR products were purified from agarose gels, digested with XbaI-HindIII, and cloned in frame at the 3′ end of the malE gene in pMAL-c2x (New England Biolabs). The correct in-frame cloning of grlA, as well as the presence of each mutation, was verified by DNA sequencing.

MBP-GrlA proteins were overexpressed in E. coli BL21 grown in 100 ml of LB with 0.2% glucose, when the culture reached an OD of 1.0, by adding 0.3 mM IPTG for 4 h at 30°C. Cells were harvested by centrifugation at 10,000 × g at 4°C and then washed once with column buffer (20 mM Tris-HCl, pH 7.4, 200 mM NaCl, 1 mM EDTA, 10 mM 2-mercaptoethanol), resuspended in 10 ml of the same buffer, and broken by sonication. Cellular debris was eliminated from the cell extracts by centrifugation. MBP-GrlA proteins were bound to amylose resin (New England Biolabs), washed with 100 ml of column buffer to remove nonspecific bound proteins, and eluted with column buffer containing 10 mM maltose.

EMSAs.

Electrophoretic mobility shift assays (EMSAs) were performed by mixing approximately 100 ng of each DNA fragment with increasing concentrations of purified WT or mutant MBP-GrlA proteins in binding buffer containing 10 mM Tris-HCl [pH 8], 50 mM KCl, 1 mM dithiothreitol (DTT), 0.5 mM EDTA, 10 μg/ml bovine serum albumin (BSA), and 5% glycerol. These reaction mixtures were incubated for 20 min at room temperature and then separated by electrophoresis in 5% polyacrylamide gels in 0.5× Tris-borate-EDTA buffer. DNA bands were stained with ethidium bromide and visualized with an Alpha-Imager UV transiluminator (Alpha Innotech Corp.).

For competitive EMSAs, the ler−50/+217 fragment was first incubated with 0.45 μM H-NS-FLAGHis₆ (H-NS-FH) for 15 min, and then increasing concentrations of MBP-GrlA were added (0, 0.17, 0.34, 0.51, 0.68, 0.85, and 1 μM) and incubated for an additional 15 min at room temperature. The complexes were visualized as described above. To detect free H-NS-FH and H-NS-FH-DNA complexes, Western blotting was performed with α-FLAG M2 antibodies (Sigma) as described above. H-NS-FH was purified from E. coli BL21 carrying plasmid pBAD-HNS-FLAGHis₆ (Table 1) as described previously (3).

Specific binding of GrlA to the ler regulatory region was confirmed by competing MBP-GrlA (0.5 μM) binding to the ³²P-5′-end-labeled ler−50/+217 probe with 10, 20, 30, and 40-fold excess of the same unlabeled PCR-derived DNA fragment as a specific competitor or of the unlabeled, nonspecific ler−260/−50 fragment for 20 min at room temperature. Samples were resolved in a 5% polyacrylamide gel at room temperature and visualized by autoradiography.

RESULTS

GrlA counteracts the repression exerted by H-NS on the ler promoter.

We previously proposed that GrlA counteracts the H-NS-mediated repression on the ler promoters of both C. rodentium and EPEC, probably by interacting with a sequence located near the promoter (3, 15; Bustamante et al., submitted). As seen in Fig. 1 A, the specific GrlA-mediated activation of the EPEC ler promoter can be recapitulated with a ler-cat transcriptional fusion in the presence of a plasmid expressing EPEC GrlA (pTEPGrlA1) (Table 1) in E. coli K-12 strain MC4100, where this promoter shows only background levels of expression in the presence of the control vector, pMPM-T3. In addition, as previously observed, when introduced into E. coli MC4100 Δhns (JPMC1), the ler promoter is derepressed, reaching an activity much higher than the background level seen in the WT strain. However, the activity of the ler promoter is further enhanced in the presence of GrlA, suggesting that GrlA may also be facilitating transcriptional activation by the RNA polymerase or counteracting additional repressors (Fig. 1A) (Bustamante et al., submitted).

FIG. 1.

LEE gene expression is GrlA independent in the absence of H-NS. (A) The expression of a ler-cat transcriptional fusion (pler-260) was analyzed in E. coli MC4100 and its isogenic Δhns mutant carrying the empty vector pMPM-T3 or its derivative pTEPGrlA1. CAT specific activity was determined from samples collected from bacterial cultures grown in 50 ml of DMEM with shaking at an optical density at 600 nm (OD₆₀₀) of 1. The results are the averages of three independent experiments done in duplicate; standard deviations are shown. (B) The expression of the ler-cat fusion was determined from samples collected from WT EPEC and ΔgrlA EPEC carrying pT6-HNS or pT6-HNS/G113D, expressing WT H-NS or its dominant negative H-NS^G113D, respectively. CAT specific activity was determined as for panel A in the absence (−) or presence (+) of 0.2% arabinose. (C) The expression levels of the ler-cat fusion in WT EPEC and its Δler, ΔgrlA, pEAF⁻, Δhns, Δhns ΔgrlA, and ΔgrlA Δhns pEAF⁻ isogenic derivatives were determined from samples collected from bacterial cultures grown in DMEM at 37°C with agitation at an OD₆₀₀ of 1 (black bars) or from static cultures at 37°C under a 5% CO₂ atmosphere after 6 h (white bars). The results are the averages of three experiments done in duplicate. (D) Total cell extracts were prepared from bacterial pellets of the same cultures grown under shaken (A) and static (S) conditions as described above and then resolved by 12% SDS-PAGE. Expression of EscJ and EspA was analyzed by Western immunoblotting using polyclonal anti-EscJ and anti-EspA antibodies. As a loading control, DnaK was also detected using a monoclonal anti-DnaK antibody.

In order to further analyze the role of GrlA in overcoming H-NS repression at the ler promoter in the EPEC background, an EPEC ΔgrlA Δhns double mutant (JPEP37) was generated as described in Materials and Methods. To assess the role of GrlA in ler activation in the absence of H-NS, expression of the ler-cat transcriptional fusion was then analyzed in WT, Δler, ΔgrlA, pEAF-cured (pEAF⁻), Δhns, ΔgrlA Δhns, and pEAF⁻ ΔgrlA Δhns EPEC strains (Table 1) grown in static and shaken cultures in DMEM. Static cultures were used as a control, as under this growth condition ler activation in EPEC is mediated by PerC, a pEAF-encoded regulator that is coregulated with the bfp operon by PerA in a GrlA-independent manner (Bustamante et al., submitted). Consistent with our previous observations, ler was expressed under both growth conditions in WT EPEC, whereas its expression was significantly reduced under shaken and static growth conditions in the ΔgrlA (JPEP35) and pEAF⁻ (JPEP30) strains, respectively (Fig. 1C). Expression in the Δler strain (JPEP24) was partially reduced under both growth conditions. In contrast, in the Δhns mutant (JPEP36), ler expression showed a 4-fold increase with respect to the expression obtained in the WT strain independently of the growth conditions, further confirming that both regulators act as H-NS antagonists. However, in the ΔgrlA Δhns double mutant (JPEP37), as well as in the ΔgrlA Δhns pEAF⁻ triple mutant (JPEP38), ler expression showed a 2-fold to 3-fold increase in comparison to the WT strain under both growth conditions. The transcriptional activities in these strains were not as high as in the Δhns single mutant (JPEP36), consistent with the notion that GrlA enhances ler expression even in the absence of H-NS (Fig. 1C).

To monitor the expression of Ler-regulated genes within the LEE, we analyzed the protein levels of EscJ (LEE2) and EspA (LEE4) by Western blotting in the different strains grown in shaken or static cultures (Fig. 1D). Expression of EscJ and EspA was found to be similar under both growth conditions in the WT strain, while their synthesis was abolished, as expected, in the Δler mutant (JPEP24). In contrast, in the ΔgrlA (JPEP25) and pEAF⁻ (JPEP30) strains, EscJ and EspA were expressed only under static or shaken growth conditions, respectively, reflecting that ler expression is mediated by PerC in the absence of GrlA in static culture and by GrlA in the absence of pEAF, and thus of the per operon, in shaken cultures. In agreement with the transcriptional activity of the ler promoter shown in Fig. 1C, EscJ and EspA were expressed under both growth conditions in all the Δhns backgrounds, even in the absence of GrlA or both GrlA and PerC (Fig. 1D).

These observations were further supported by mimicking the hns mutation in WT EPEC by inducing the synthesis of H-NS^G113D, a mutant derivative of H-NS that carries a mutation that abolishes DNA binding without affecting oligomerization with WT H-NS then acting as a dominant negative (68). We have used this strategy successfully to inhibit the H-NS function at a defined time of the culture, preventing, at least in part, the pleiotropic and potentially deleterious effects of a chromosomal hns deletion (8). WT EPEC E2348/69 and its ΔgrlA derivative containing a ler-cat fusion and plasmids pT6-HNS and pT6-HNS^G113D were grown in DMEM with or without arabinose to induce, or not, the expression of WT H-NS and H-NS^G113D. The CAT specific activity derived from the transcriptional activity of the ler-cat fusion was determined from whole-cell extracts prepared from culture samples taken after 5 h. A slight increase in ler expression was observed in WT EPEC upon induction of H-NS^G113D, but not of WT H-NS. In contrast, ler expression was 4-fold higher upon induction of the dominant-negative version of H-NS (H-NS^G113D) than with the WT protein (Fig. 1B).

These results confirm that GrlA acts as an H-NS antagonist on the ler promoter and also support the notion that GrlA, in addition to acting as an antirepressor, further enhances ler expression, either by acting as a classical activator or by counteracting the negative regulation exerted by additional repressors.

MBP-GrlA directly interacts with the ler regulatory region and displaces H-NS.

The role of GrlA in the regulation of the LEE1 (ler) (3, 29, 36, 40), flhDC (31), and ehxCABD (55) operons, as well as the presence of a predicted HTH DNA binding domain at the N terminus of the protein, which shares homology with a known DNA binding activator, such as CaiF (3, 7), strongly suggests that it exerts its regulatory function by specifically binding to a DNA sequence motif in the ler regulatory region. However, previous attempts to demonstrate specific binding of purified GrlA to its reported regulatory sequence targets have been unsuccessful (3, 55) or not fully conclusive, as more recently it was shown that a recombinant GST-GrlA fusion protein formed protein-DNA complexes with the LEE1 regulatory region even in the absence of the HTH motif and also shifted an unrelated band (29).

We further analyzed the in vitro DNA binding activity of the MBP-GrlA fusion protein by modifying the purification procedure and binding reactions, as indicated in Materials and Methods. The MBP-GrlA fusion is fully functional, as it is still capable of complementing the EPEC ΔgrlA mutant and of interacting with GrlR (see below). EMSAs were performed by mixing increasing concentrations of amylose affinity-purified MBP-GrlA and 100 ng of PCR DNA fragments spanning the ler regulatory region between positions −260 and +217 (ler−260/+217) contained in the ler-cat fusion and, as a negative control, the EPEC structural dnaJ gene. As shown in Fig. 2 A, GrlA specifically bound to the ler fragment in a concentration-dependent manner, starting at a concentration of 0.4 μM, while no shifting was seen for the control dnaJ fragment even at the highest concentration of the protein (1 μM).

FIG. 2.

GrlA and H-NS interact with the ler regulatory region in vitro. (A and B) EMSA of PCR DNA fragments comprising the ler regulatory region (ler−260/+217) and the dnaJ coding sequence (A) or of fragments ler−50/+217 and ler−260/−50 (B). DNA fragments were mixed and incubated with increasing concentrations of purified MBP-GrlA (0, 0.28, 0.42, 0.56, 0.7, 0.84, and 1 μM). Free DNA and protein-DNA complexes were resolved by 5% polyacrylamide gel electrophoresis and stained with ethidium bromide. (C) Competitive EMSA. The ³²P-5′-end-labeled ler−50/+217 probe was mixed with 0.5 μM purified MBP-GrlA (lane 2), followed by the addition of a 10- to 40-fold excess of unlabeled specific (ler−50/+217) (lanes 3 to 6) or nonspecific (ler−260/−50) (lanes 7 to 10) competitors and resolved as indicated above. Lane 1, ³²P-5′-end- labeled ler−50/+217 probe without protein. (D) EMSA of fragments ler−50/+217 and ler−260/−50 incubated with increasing concentrations of purified H-NS-FH (0, 0.25, 0.3, 0.35, 0.4, and 0.45 μM). (E) H-NS displacement by GrlA from the ler regulatory region shown by competitive EMSA (top) and Western blotting with an α-FLAG antibody (bottom). The ler−50/+217 fragment was incubated with purified H-NS-FH (0.45 μM) (lanes 2 to 8) or MBP-GrlA (0.9 μM) (lane 9), and then increasing concentrations of MBP-GrlA (0.17, 0.34, 0.51, 0.68, 0.85, and 1 μM) were added (lanes 3 to 8). Lane 1, ler−50/+217 without protein; lane 10, free H-NS-FH (1 μM).

We previously determined that GrlA and H-NS were still able to activate and repress ler transcriptional fusions containing the ler regulatory region from positions −54 to +217 from EPEC (Bustamante et al., submitted) or −40 to −216 from C. rodentium (3), respectively, suggesting that the GrlA and H-NS recognition sequences are located in the vicinity of the ler promoter or downstream from it. Based on these results, to further delimit the GrlA binding motif in the EPEC ler regulatory region, two fragments containing the sequences between positions −260 and −50 (ler−260/−50) and −50 and +217 (ler−50/+217) were obtained by PCR and used in EMSAs, as described for the whole fragment. In agreement with the GrlA-mediated activation of a ler-cat transcriptional fusion containing the sequence between positions −50 and +217, MBP-GrlA was only able to bind to the fragment spanning these positions (Fig. 2B). Specific binding of GrlA to this region was also tested by competing binding of MBP-GrlA to the ³²P-5′-end-labeled ler−50/+217 probe with 10- to 40-fold excess of unlabeled specific and nonspecific fragments. As shown in Fig. 2C, only the unlabeled specific fragment was able to disrupt the interaction of GrlA with the labeled probe, further confirming that GrlA specifically interacts with the ler regulatory region downstream of position −50. Moreover, H-NS-His₆ shifted only the ler fragment spanning positions −50 to +217 (Fig. 1D). To further assess the role of GrlA as an H-NS antagonist in the ler regulatory region, we performed competitive EMSAs by first allowing the specific interaction of H-NS-FH with the ler−50/+217 fragment, followed by the addition of increasing concentrations of MBP-GrlA, as described in Materials and Methods. The addition of MBP-GrlA shifted the H-NS-DNA complex to a slower-migrating complex similar to that formed by MBP-GrlA alone (Fig. 2E, top) and displaced H-NS-FH from the DNA fragment, as shown by Western blotting using an anti-FLAG antibody (Fig. 2E, bottom).

The C-terminal domain of GrlA is needed for transcriptional activation, but not for the interaction with GrlR.

Transcriptional regulators, like most proteins, can be modular and have more than one functional domain. According to the predicted secondary structure shown in Fig. 3 B, the N-terminal half of GrlA contains four alpha helices. Helices III and IV correspond to a putative HTH DNA binding motif, whose presence correlates with the transcriptional activation function and in vitro DNA binding property of GrlA (Fig. 1 and 2). As mentioned above, GrlR interacts with GrlA, forming heterodimers (12, 29, 33), and it has been proposed that this interaction inhibits the GrlA-mediated activation of the ler promoter, thus also repressing LEE gene expression (3, 15). However, the role of this interaction in LEE regulation is not fully understood. In all, these observations suggest that this protein contains at least two functional domains. To investigate the potential modular nature of GrlA, we performed a systematic deletion analysis of the protein by generating consecutive deletions of 20 amino acids along the entire protein (Fig. 4 A), as described in Materials and Methods. All deletions were unable to induce the expression of a ler-cat transcriptional fusion in E. coli MC4100 (Fig. 4B) or to complement the EPEC ΔgrlA strain (JPEP25) (data not shown), suggesting that each deletion imposed a structural constraint or modification that altered the entire folding of the protein or that several parts of the protein are essential for transcriptional activation.

FIG. 3.

GrlA homologues and secondary-structure prediction. (A) The amino acid sequence multialignment of GrlA from EPEC E2348/69 (AAC 38375.1), SGH from S. enterica serovar Typhimurium LT2 (NP_945169), Yber from Yersinia bercobieri ATCC 43970 (ZP_04627362), Plu from Photorhabdus luminicens subsp. Laumondii TTO1 (NP_927633), and CaiF from E. coli K-12 substrain MG1655 (NP_414576), was done using the ClustalW sequence alignment program from the European Bioinformatics Institute (EMBL-EBI). Identical amino acids are boxed, and similar amino acids are shaded in gray. The dots over the alignment indicate the GrlA amino acid residues mutated in this work. (B) Schematic representation of the predicted secondary structure of GrlA using the PSIPRED server (6). The arrows indicate beta strands, and the rectangles represent alpha helices. Amino acids replaced by alanine are underlined. Helices III and IV corresponding to the putative GrlA HTH DNA binding domain are shown in gray. The symbols above the amino acid sequence indicate alanine substitutions that did not affect GrlA function (white circles), affected ler expression and DNA binding (black squares) or abolished the interaction with GrlR (arrowheads).

FIG. 4.

The complete GrlA protein is required for ler expression, but not for GrlA-GrlR interactions. (A) Schematic representation of the predicted secondary structure of WT GrlA and of the GrlA 20-amino-acid sequential deletions. (B) The effects of the deletions on the activation of the ler-cat transcriptional fusion were determined in E. coli MC4100 transformed with pMPM-T3, pTEPGrlA1, and its derivatives expressing the GrlA deletion variants. The results are the average of three experiments done in duplicate. Standard deviations are shown. (C) Pulldown assay to analyze the interaction between the GrlA deletion mutants fused to MBP and GrlR fused to GST. Protein samples were resolved by 12% SDS-PAGE. The bands corresponding to MBP-GrlA, GST-GrlR, and GST are indicated. (Bottom) Western blot of whole-cell extracts of strains expressing the different MBP-GrlA variants used in the pulldown assay.

To gain insights into the sequence motifs of GrlA required to interact with GrlR, all GrlA deletion mutants were MBP tagged, as described in Materials and Methods. Then, pulldown assays were performed using as prey WT MBP-GrlA and the MBP-GrlA deletion mutants expressed from pMAL-c2X derivatives (Table 1) and, as bait, GrlR fused in frame to GST-GrlR expressed from an IPTG-inducible promoter from plasmid pGEX-4T1. Full-length GST-GrlR was bound to glutathione beads and used to pull down MBP-tagged GrlA proteins from whole-cell extracts of E. coli BL21 pLys expressing them. MBP-GrlA_WT was pulled down by GST-GrlR, but not by GST alone, confirming the interaction between the two proteins (Fig. 4C).

Interestingly, only mutants GrlA_Δ101-120 and GrlA_Δ121-137, albeit inactive, were pulled down by GST-GrlR, indicating that the last 37 amino acids of the protein, while important for transcriptional activation, are not needed for the correct folding of the GrlR-interacting motif (Fig. 4C, top). Moreover, the elimination of all 37 amino acids produced a deletion mutant (GrlA_Δ101-137) with the same phenotype seen for GrlA_Δ101-120 and GrlA_Δ121-137, confirming that the C terminus is dispensable for protein-protein interactions, but not for transcriptional activation (data not shown). Further deletion of the C terminus up to amino acid 80 (GrlA_Δ81-137) generated a mutant that was no longer able to interact with GrlR (data not shown). All variants were expressed at levels similar to those seen for the WT protein, as shown by Western blotting of whole-cell extracts of E. coli BL21 pLys expressing the MBP-GrlA deletion mutants (Fig. 4C, bottom).

The predicted HTH motif of GrlA contains amino acid residues that are essential for GrlA transcriptional activity and DNA binding.

The results described above indicated that the entire protein was needed for the function of GrlA as a transcriptional regulator and also suggested that protein folding or stability was not being compromised for all deletion mutants, as the deletion of the last 37 amino acids did not affect the interaction with GrlR. GrlA is a 137-amino-acid protein that is highly conserved in strains belonging to the A/E family of bacterial pathogens. Uncharacterized homologues of GrlA are present in S. enterica, Y. bercovieri, and Photorabdus sp. (Fig. 3A). A partially characterized homologue, CaiF, is found in members of the family Enterobacteriaceae. CaiF is a DNA binding protein that positively regulates the cai and fix operons (7). The multialignment analysis of these proteins revealed several conserved amino acid residues, mainly in the N-terminal half of the protein (Fig. 3A). To investigate in more detail the roles of helices II, III, and IV, and thus of the HTH motif, as well as to determine the contributions of the conserved amino acid residues to GrlA function (Fig. 3A), alanine substitutions were generated at residues P23, L24, Y25, S29, W31, and C32 (in helix II) and R41, N42, I44, E46, F48, I50, L52, R54, S56, Y61, and R65 (in the HTH motif, helices III and IV) by site-directed mutagenesis, as described in Materials and Methods. The effects of these changes on GrlA function were assessed by determining the capacities of the GrlA mutants to activate the expression of a ler-cat transcriptional fusion in E. coli MC4100 in comparison to WT GrlA (Fig. 5 A and Table 3). Alanine substitutions at residues Y25, W31, I44, F48, I50, R54, Y61, and R65 produced nonfunctional GrlA mutants, as they were unable to activate the expression of the ler-cat transcriptional fusion. In contrast, changes at residues P23, L24, S29, C32, N42, E46, L52, and S56 produced mutants that still activated the ler-cat fusion to different extents, but without abolishing the function of GrlA. These results indicated that the highly conserved residues I44, F48, I50, R54, and Y61 in the HTH motif (helices III and IV) play critical roles in transcriptional activation and suggested that these residues may also be important for the putative regulatory function of GrlA homologues present in other bacteria (Fig. 3A). Interestingly, alanine substitutions at residue Y25 of the highly conserved PLY motif and residue W31 in helix II also produced inactive mutants, indicating that this helix also has an important function in transcriptional activity, likely having a role in the correct folding of the HTH motif or in DNA recognition or binding.

FIG. 5.

Different amino acid residues in the HTH motif are important for GrlA function. Site-directed mutants were generated at different positions near the putative HTH motif. (A) Functional analysis of GrlA site-directed mutants in E. coli MC4100 carrying a ler-cat transcriptional fusion. Plasmids pMPM-T3 (empty vector) and pTEPGrlA1 and its derivatives expressing one of the single-amino-acid mutants pTEPGrlA1/P23A, L24A, Y25A, W31A, N42A, I44A, E46A, F48A, I50A, L52A, R54A, and S56A were transformed into E. coli MC4100 containing the ler-cat transcriptional fusion. Bacterial cultures were grown in DMEM at 37°C with shaking, and samples were collected at an OD₆₀₀ of 1 to determine CAT specific activity. Standard deviations are shown. (B) Complementation analysis of EPEC ΔgrlA. Plasmids expressing WT MBP-GrlA and MBP-GrlA single-amino-acid mutants were introduced into EPEC ΔgrlA. The resulting strains were grown as indicated above. The secreted proteins were concentrated from culture supernatants by TCA precipitation and resolved by 12% SDS-PAGE. (C) Western blot analysis of whole-cell extracts of the strains shown in panel B, using anti-EscJ, anti-Tir, anti-MBP, and anti-DnaK, which was used as a loading control. (D) EMSAs were performed to analyze the capacities of purified MBP-GrlA mutants (MBP-GrlA/Y25A, MBP-GrlA/W31A, MBP-GrlA/I44A, MBP-GrlA/F48A, MBP-GrlA/I50A, and MBP-GrlA/R54A) to bind to the ler−50/+217 fragment at a concentration of 1.2 μM. The DNA-protein complexes were resolved in 5% polyacrylamide gels and stained with ethidium bromide.

TABLE 3.

Phenotypes shown by the GrlA mutants

Mutation	Phenotype	Interaction with GrIR	Location
P23A	Active	Yes
L24A	Active	No	Helix II
Y25A	Inactive	Yes	Helix II
S29A	Active	Yes	Helix II
W31A	Inactive	No	Helix II
C32A	Active	Yes	Helix II
R41A	Active	Yes	Helix III
N42A	Active	Yes	Helix III
I44A	Inactive	No	Helix III
E46A	Active	Yes	Helix III
F48A	Inactive	No	Helix III
I50A	Inactive	No	HTH loop
L52A	Active	Yes	Helix IV
R54A	Inactive	Yes	Helix IV
S56A	Active	Yes	Helix IV
Y61A	Inactive	Yes	Helix IV
R65A	Inactive	Yes	Helix IV
Δ125-137	Inactive	Yes	C terminal
Δ131-137	Inactive	Yes	C terminal
R132A	Active	Yes	C terminal
R133A	Inactive	Yes	C terminal
K134A	Inactive	Yes	C terminal
K135A	Active	Yes	C terminal
E136A	Active	Yes	C terminal

To further evaluate the effects of these mutations on GrlA activity, MBP fusions of all mutants were constructed by subcloning the grlA mutant genes from pMPM-T3 derivatives into pMAL-c2x (Table 1), as described in Materials and Methods. Plasmids expressing the MBP-GrlA variants were then introduced into EPEC ΔgrlA to assess their capacities to complement the mutation by analyzing type III secretion (T3S) protein profiles by SDS-PAGE of proteins recovered from culture supernatants (Fig. 5B) and expression of LEE-encoded proteins by Western blotting (Fig. 5C) of whole-cell extracts prepared from culture samples collected from shaken cultures in DMEM. In both cases, the results were fully consistent with the phenotypes observed for the different mutants in E. coli K-12 carrying the ler-cat fusion (Fig. 5A and Table 3), as inactive mutants were not capable of reestablishing T3S, as determined by the presence of protein bands corresponding to EspA, EspB, and EspD (Fig. 5B) or expression of LEE-encoded proteins, such as EscJ or Tir (Fig. 5C, top two gels).

Proper expression of all GrlA mutants was evaluated by detecting the MBP portion of each fusion using an anti-MBP antibody. As seen in Fig. 5C (third gel from the top), all mutants were expressed at similar levels in comparison to the MBP fusion of WT GrlA, which complemented ler-cat expression in E. coli K-12 and T3S and the expression of LEE-encoded proteins in EPEC ΔgrlA (Fig. 5).

Helices II, III, and IV seem to comprise a typical trihelical bundle or HTH domain that includes the HTH motif (helices III and IV), a structural feature commonly found in DNA binding transcription factors (2). We found that mutations in conserved residues located in the three helices abolished the capacity of GrlA to activate ler transcription. The third helix of the HTH domain is usually involved in the interaction with the DNA and thus is known as the recognition sequence. However, the first and second helices also play important roles in DNA recognition and specificity (2). Considering that the majority of the conserved residues in the HTH domain were essential for GrlA function, we explored the possibility that these residues were also required for DNA binding by performing EMSAs with the purified MBP-GrlA mutants. An excess of 1.2 μM purified MBP-GrlA_WT or MBP fusions to GrlA mutants Y25A, W31A, I44A, F48A, I50A, and R54A was mixed with a PCR fragment spanning the ler regulatory region from positions −50 to +217. In agreement with the predicted function of the HTH domain and the phenotypes described above (Fig. 5A to C), all mutants tested by EMSA were unable to form a protein/DNA complex with the regulatory region of ler, in contrast to WT GrlA, showing that helices II, III, and IV, as well as the linker region between helices III and IV, participate in DNA recognition and thus in transcriptional activation (Fig. 5D).

Amino acid residues in helices II and III are involved in GrlA-GrlR protein-protein interactions.

To determine the roles of single amino acid mutations in the interaction of GrlA with GrlR, MBP-tagged GrlA mutants were used in pulldown assays with GST-GrlR, as described above. The active GrlA variants P32A, S29A, C32A, R41A, N42A, E46A, L52A, and S56A were still able to interact with GrlR, without any evident change in affinity, indicating that these residues do not play essential roles in either function (Fig. 6 A and Table 3). In contrast, GST-GrlR did not pull down mutants L24A, W31A, I44A, F48A, and I50A; however, while most of these mutants were also not able to activate ler expression, MBP-GrlA_L24A was still active (Fig. 5 and Table 3). Interestingly, inactive GrlA mutants, such as Y25A, R54A, Y61A, and R65A, were pulled down by GST-GrlR, indicating that despite losing the DNA binding/transcriptional activation properties, the GrlR-interacting motif was still functional. These results, together with the observation that these proteins are expressed at similar levels, as tested by Western blotting of the same whole-cell extracts used for the pulldown assay using an anti-MBP antibody (Fig. 6B), also supported the notion that these proteins are being expressed and folded correctly.

FIG. 6.

Amino acid positions in the HTH motif are also essential for GrlA-GrlR interactions. (A) A pulldown assay was performed to analyze the interactions of GrlA single-amino-acid mutants with GrlR. Whole-cell extracts of strains expressing the different MBP-GrlA proteins were mixed with GST-GrlR bound to glutathione-Sepharose beads. The bound proteins were resolved by 12% SDS-PAGE. Protein bands corresponding to GST, GST-GrlR, and MBP-GrlA proteins are indicated. (B) Western blot analysis of whole-cell extracts of the strains expressing the MBP-GrlA fusion variants with anti-MBP antibodies.

In all, these data suggested that the DNA binding/activation and heterodimerization motifs overlap at the N terminus of GrlA but do not involve the exact same set of amino acids. In addition, although further analysis is required, they also suggested that helices II and III (Fig. 3B) are part of the motif that interacts with GrlR, which is not the case for the putative recognition helix at the HTH motif (helix IV), as all 5 mutants in this region, including inactive mutants R54A, S56A, and Y61A, maintained its interaction with GrlR (Fig. 5A and Table 3).

The GrlA C terminus contains residues involved in ler activation.

In addition to the functional motifs identified in the N-terminal half of the protein, the deletion analysis shown in Fig. 4 also revealed that the last 17 amino acids of GrlA were required for ler expression, but not for interaction with GrlR. To further characterize this putative functional motif, two additional deletions, GrlA_Δ125-137 and GrlA_Δ131-137, were constructed and tested for the capacity to activate the ler-cat fusion and to complement the EPEC ΔgrlA mutant. Both mutants were not active in both assays, indicating that the C-terminal tail of GrlA, characterized by the presence of 5 out of 6 charged amino acid residues (RRKKE), plays an important role in DNA binding and activation (Table 3). To determine if all of the residues of this motif were important for GrlA activity, alanine substitutions were generated at residues R132, R133, K134, K135, and E136 and tested. We found that changes at positions R132, K135, and E136 did not affect GrlA function, whereas mutations in R133 and K134 produced protein mutants defective in ler activation (Fig. 7 A), but not in their interaction with GrlR, as expected (Fig. 7B).

FIG. 7.

Amino acid residues at the C terminus of GrlA are important for ler expression. (A) The functionality of the GrlA deletion mutant lacking amino acids 131 to 137 and of point mutants at residues R132, R133, K134, K135, and E136 was analyzed by testing their capacities to activate the expression of the ler-cat fusion in E. coli MC4100, as described for Fig. 5A. Standard deviations are shown. (B) The interaction of MBP-GrlA proteins containing the mutations described above with GrlR was tested by performing a pulldown assay as described in the legend to Fig. 6.

DISCUSSION

The genome of a bacterial pathogen is characterized by the presence of large DNA segments that have been acquired by horizontal gene transfer (HGT) events and that are not present in closely related bacteria. In many cases these genomic islands encode proteins with virulence properties and thus are called PAIs (24). It has been proposed that to prevent the potential deleterious effects caused by the uncontrolled expression of incoming new genetic information, the global regulator H-NS has acted as a guardian silencer during the evolution of enteric bacterial pathogens by binding to AT-rich DNA tracks commonly found in PAIs, thus repressing the expression of this information until regulatory mechanisms evolved to control its spatiotemporal expression in an appropriate manner (16, 42, 50, 51). As with other PAIs, the core LEE has a low G+C content (an average of 38.4%), which contrasts with the 50% G+C content of the entire genome, a characteristic of DNA regions acquired horizontally (46, 49). Accordingly, it has been widely documented that the expression of LEE genes is repressed by H-NS (3, 9, 17, 23, 48, 57, 62, 69). As summarized in the introduction, the expression of LEE genes is regulated by a complex network of negative and positive regulators, most of them modulating the expression of the master regulator of the island, Ler, which in turn counteracts H-NS repression at all the promoters in the LEE, except for the LEE1 (ler) promoter. Despite this complexity, GrlA, a second LEE-encoded regulator, has been shown to play a key role in the regulation of ler (and thus in the expression of LEE genes) and of other virulence-related genes outside the LEE, as well as in virulence in vitro and in vivo. In addition, it has previously been shown that Ler and GrlA regulate each other, establishing a positive regulatory loop that ensures the appropriate level of Ler production to activate LEE gene expression (3). Despite its importance, the mechanism whereby GrlA activates ler expression is not well understood.

Here, we have further shown that GrlA acts as a positive regulator of ler expression by binding directly to its regulatory region, thus counteracting the H-NS repression exerted on its promoter. However, our data also confirmed that GrlA further enhances ler transcription in the absence of H-NS, although it remains to be investigated whether it does so by counteracting other repressors or by facilitating the formation of productive RNA polymerase-promoter complexes. The secondary structure of GrlA predicts a putative HTH DNA binding domain in the N-terminal half of the protein between residues 41 and 65, which was shown to be functional by site-directed mutagenesis. Mutations in this motif completely abolished GrlA-mediated activation of the ler promoter and DNA binding. Although the exact sequence motif bound by GrlA has not been identified, the interaction with the ler regulatory region seems to be specific, as GrlA did not bind other DNA fragments under the conditions used in this work. Ler (9) and GrlA share the capacity to counteract H-NS repression by binding to sequences within the LEE in different promoter regions and seem to do it specifically based on the observation that these proteins do not activate each other's target promoters in a non-EPEC background, such as E. coli K-12 (3). Moreover, expression of Tir and EspB in a C. rodentium Δler ΔgrlA double mutant can be restored only with a plasmid expressing Ler from a constitutive promoter, while a plasmid expressing GrlA restores only ler expression (15).

As in the case of GrlA, several virulence-regulatory proteins activate virulence gene expression by counteracting the repression exerted by H-NS. Among others, this role has been reported for AraC-like proteins, such as CfaD from enterotoxigenic E. coli (34), HilD from S. enterica serovar Typhimurium (8), and ToxT from Vibrio cholerae (72), and also for VirB, an unusual transcription factor from Shigella flexneri (67); for RovA, a member of the SlyA/Hor family of transcriptional regulators from Yersinia sp. (11); Ler, an H-NS paralogue from A/E pathogens (3, 9, 23, 69); and SsrB, a response regulator from S. enterica (70).

The ability of GrlA to activate ler expression correlates with the prediction of a putative HTH DNA binding motif in the N-terminal domain of the protein, suggesting that GrlA must bind to the ler promoter region to activate its expression. Previous attempts to demonstrate the specific interaction of purified MBP-GrlA with the C. rodentium ler promoter (3) or of a 6XHis-GrlA fusion to the EHEC ehxCABD and flhDC promoters (55) were unsuccessful. Binding of GST-GrlA to the EHEC ler promoter region was recently shown; intriguingly, binding was still observed even for a deletion mutant lacking the HTH motif (29). In this work, we demonstrated that a functional MBP-GrlA fusion specifically binds to a DNA fragment spanning the ler promoter region from positions −50 to +217, but not to the ler fragment spanning positions −260 to −50 or to an unrelated control fragment. This result is in agreement with experiments demonstrating that GrlA still enhances the expression of a ler-cat transcriptional fusion containing the region between positions −50 and +217, which is also repressed by H-NS (3; Bustamante et al., submitted). In addition, we showed that GrlA binding displaces H-NS from its binding site, also located between positions −50 and +217. Thus, GrlA could alleviate H-NS repression by competing for overlapping binding sites or by generating architectural changes in this region that prevent H-NS binding or the formation of a nucleoprotein repressor complex.

A variety of structural motifs are involved in DNA binding. One of the most frequently found in bacterial transcription factors is the HTH motif. The basic structure of the HTH motif consists of a bundle of three helices, where the second and third helices, connected by a tight turn, are involved in DNA recognition (2). The predicted secondary structure of GrlA shows the presence of four alpha helices (helices I to IV) in the N-terminal half of the protein and an additional alpha helix (helix V) toward the C terminus. Helices II, III, and IV constitute the 3-helical HTH domain, and helices III and IV are the putative core helices (Fig. 3B). The relevance of the HTH domain in GrlA-mediated activation was analyzed by site-directed mutagenesis. Several amino acids were found to be critical for GrlA activity (Y25, W31, I44, F48, I50, and R54). GrlA mutants carrying alanine substitutions at these residues were incapable of activating the expression of a ler-cat transcriptional fusion in E. coli K-12 and unable to complement an EPEC ΔgrlA mutant. Mutants carrying serine substitutions for residues I44, F48, I50, and R54 showed the same phenotype as the corresponding alanine mutants (data not shown), confirming the critical role that these amino acid residues play in the functionality of GrlA. Consistent with the expected role of the HTH domain, inactive mutants were also defective in DNA binding, confirming their relevance in GrlA function; however, it remains to be determined if these residues establish direct contacts with the DNA backbone or play a structural role.

Amino acids Y25 and W31, located in helix II, are hydrophobic, suggesting that they may participate in the correct folding of the HTH domain by interacting with other amino acids. The second HTH helix (helix III in GrlA) has been associated with the correct positioning of the recognition helix (helix IV) at the DNA binding site. In addition, the replacement of hydrophobic amino acids I44 and F48 within the second HTH helix (helix III) produced inactive GrlA mutants. Taken together, these results suggest that hydrophobic amino acids in the first and second HTH helices (helices II and III) probably form a hydrophobic pocket that is important to maintain the compact fold of the HTH motif. However, the possibility that these amino acids participate directly in DNA binding—as shown for the corresponding helices 1 and 2 from the winged HTH motif of the LysR family of transcriptional regulators, which were implicated in DNA binding, as its mutation produced a protein that lost the interaction with DNA (73)—cannot be ruled out at this point. The turn connecting the second and third helices is also important for maintaining the correct folding of the HTH motif, and it does not support changes or distortions in its sequence (2). According to this notion, a mutation in residue I50, which is located in the middle of the turn, abolished GrlA activity. Furthermore, in agreement with the putative role of helix IV in the interaction with DNA, the mutation at residue R54 produced a GrlA variant that was deficient in activating ler and in DNA binding. It is also consistent with the observation indicating that positively charged amino acids are the main residues involved in the interaction with DNA phosphate groups (43). In contrast, several residues located in the HTH were not essential for GrlA function as a transcriptional regulator. The fact that not all alanine substitutions in this motif caused significant defects in GrlA function suggests that the integrity of the protein is, in general, maintained with this mutagenesis strategy and that changes generating a defective phenotype reflect only local effects and the importance of those amino acids in establishing contacts with the DNA target sequence or the stability and spatial orientation of the HTH fold in GrlA. In support of this, helical-wheel projections of helices III and IV of the putative GrlA HTH motif showed that residues that are essential for GrlA activity are clustered on one side of each of the helices (data not shown). In particular, at the putative recognition helix (helix IV), hydrophilic amino acid residues, which, if mutated, produced inactive GrlA proteins, are clustered on one side, probably favoring sequence-specific contacts with the DNA.

At this point, pulldown experiments and two-hybrid assays have indicated that GrlA does not interact with itself (data not shown). These observations suggest that GrlA acts on the DNA as a monomer, which is consistent with the apparent absence of direct or inverted repeats in the sequence of the fragment bound by GrlA. However, further work is required to completely rule out GrlA dimerization, as CaiF, the only GrlA homologue partially characterized, has been shown to be a dimer in solution and to bind to a palindromic sequence (7). Only two nonfunctional single-amino-acid mutants have been described for CaiF. An N-terminal A27V mutation seems to dramatically affect its structure and function, while an I62N mutant, located in the central part of the protein, showed a defect in DNA binding (7). Both CaiF A27 and I62N are not conserved with GrlA or other GrlA homologues (Fig. 3A).

The interaction between GrlA and GrlR has been documented previously (29, 33), an observation that has led to the proposal that GrlR acts as a repressor of LEE gene expression by heterodimerizing with GrlA (3, 31, 33, 40). However, we have recently shown that this interaction rather serves to prevent repression by GrlR, which inhibits LEE gene transcription specifically and independently of its interaction with GrlA (37). Thus, GrlA has a dual and key role in the transcriptional positive regulation of the LEE, as, in addition to acting as a DNA binding protein in the ler regulatory region to antagonize H-NS repression and to further promote ler transcription, it also acts as an antirepressor by inactivating GrlR through direct protein-protein interactions. In view of the relevance of this interaction, in this work, we also identified several GrlA mutants that were incapable of interacting with GrlR.

Mutations in residues L24, W31, I44, F48, and I50 produced GrlA versions that no longer interacted with GrlR. Interestingly, while most of these mutants also lost the capacity to activate ler expression, GrlA L24A was still functional. Although a full alanine-scanning analysis would be required to determine all positions important for the interaction with GrlR, the locations of these mutations suggest that the region spanning helices II and III forms a functional domain for protein-protein (GrlA-GrlR) interactions that partially overlaps with the DNA binding domain. This region is not as rich in basic residues as the C terminus, which was suggested to be the region potentially involved in the interaction with GrlR, for which a negatively charged EDED motif was shown to be important for GrlA-GrlR heterodimerization (33). In this regard, several basic residues, including the RRKK motif at the GrlA C terminus, had no role in the interaction with GrlR.

Of note, the 3 residues of the highly conserved PLY motif at helix II seem to play different roles in the functionality of the protein. While mutant P23A did not show any defect in ler activation, DNA binding, or the interaction with GrlR, L24A was active but lost the interaction, and in contrast, Y25A, although inactive, was still able to interact with GrlR. In time, it will be interesting to investigate if this motif also has a role in protein-protein interactions or DNA binding for other GrlA-like proteins.

This work furthers our understanding of the functional organization of GrlA and provides insights for future studies into the molecular structure-function relationship of this novel family of regulatory proteins and, in particular, into the importance of the GrlR-GrlA duo in the regulation of the LEE and the evolution of A/E pathogens.