I-TRAP: A method to identify transcriptional regulator activated promoters

Background The differential expression of virulence genes is often used by microbial pathogens in adapting to the environment of their host. The differential expression of such sets of genes can be regulated by RNA polymerase sigma factors. Some sigma factors are differentially expressed, which can provide a means to identifying other differentially expressed genes such as those whose expression are controlled by the sigma factor. Methods To identify sigma factor-regulated genes, we developed a method, termed I-TRAP, for the identification of transcriptional regulator activated promoters. The I-TRAP method is based on the fact that some genes will be differentially expressed in the presence and absence of a transcriptional regulator. I-TRAP uses a DNA library in a promoter-trap vector that contains two reporter genes, one to allow the selection of active promoters in the presence of the transcriptional regulator and a second to allow screening for promoter activity in the absence of the transcriptional regulator. Results To illustrate the development and use of the I-TRAP approach, the construction of the vectors, host strains, and library necessary to identify SigmaE-regulated genes of Mycobacterium tuberculosis is described. Conclusion The I-TRAP method should be a versatile and useful method for identifying and characterizing promoter activity under a variety of conditions and in response to various regulatory proteins. In our study, we isolated 360 clones that may contain plasmids carrying SigmaE-regulated promoters genes of M. tuberculosis.

sets of genes involved in specific processes, such as genes involved in the heat-shock response or those required for sporulation in Bacillus subtilis [7,8].
Identifying a differentially expressed sigma factor can provide a means to identifying other differentially expressed genes -genes whose expression are controlled by the sigma factor. To identify sigma factor-regulated genes, we developed a method, termed I-TRAP, for the identification of transcriptional regulator activated promoters. I-TRAP takes advantage of the fact that some genes will be differentially expressed in the presence and absence of a transcription regulatory protein. To identify transcription factor-dependent promoters, a genomic DNA library is generated in a promoter-trap plasmid containing two reporter genes. The library is transformed into strain expressing the transcription factor, and recipients with plasmids containing an active promoter are selected by virtue of expression of one of the reporter genes. The plasmids are recovered and transformed into a strain that does not express the transcription factor. Promoters dependent on the expression of the regulator are differentiated from other promoters by screening for lack of reporter activity in bacteria not expressing the regulatory protein. To illustrate the development and use of the I-TRAP approach, we describe here the construction of the vectors, host strains, and library necessary to identify σ E -regulated genes of M. tuberculosis.

Bacterial strains and growth condition
Bacterial strains are listed in Table 1. Escherichia coli strains were grown in Luria Broth (LB) or on LB agar. Mycobacterium smegmatis LR222 strains were grown in Middlebrook 7H9 broth supplemented with 10% albumin-dextrose-catalase (ADC, v/v, Difco Laboratories, Detroit, MI) and 0.05% Tween 80 (v/v, Sigma, St. Louis, MO) at 37°C, with agitation on a rotating platform (100 rpm), or were grown on trypticase soy agar (TSA) plates. M. tuberculosis H37Rv TMC102 cultures were grown in Middlebrook 7H9 broth supplemented with ADC and 0.05% Tween 80 at 37°C in 250 ml nephelometer flasks on a rotating platform (60 rpm) or were grown on Middlebrook 7H10 plates with 10% oleic acid-albumin-dex-

DNA manipulations and DNA sequencing
Restriction enzyme reactions were performed as recommended by the manufacturer (Invitrogen). Wizard Plus Minipreps DNA Purification System (Promega, Madison, WI) was used to isolate plasmid DNA from E. coli bacteria and from M. smegmatis bacteria as previously described [11]. Phosphatase reactions using Calf Intestinal Alkaline Phosphatase (Invitrogen) and ligation reactions using T4 DNA Ligase (Invitrogen) were carried out according to manufacturer's directions. Sequencing reactions were carried out using the Big Dye Terminator Cycle Sequencing Ready Reaction Kit as recommended by manufacturer (Applied Biosystems). Sequence reactions were analyzed using an ABI 373 DNA sequencer and sequence data were assembled using SeqEd 675 DNA Sequence Editor, version 1.0.3 (Applied Biosystems).
Genomic DNA isolation M. smegmatis genomic DNA was purified by a modification of the glass-bead lysis method as previously described [12]. M. tuberculosis genomic DNA was isolated as previously described [13]. DNA concentration was estimated using a GeneQuant II apparatus (Amersham Biosciences).

Polymerase chain reaction
The primers used in this study are listed in Table 2

Electroporation
Electrocompetent M. smegmatis bacteria and E. coli bacteria (Biorad) were produced and DNA was electroporated into these bacteria as previously described [14,15].

Experimental strategy
The basic I-TRAP approach ( Figure 1) uses a genomic DNA library in a plasmid vector containing a promoterless 'operon' with two reporter genes. One reporter gene, the aph gene which confers resistance to kanamycin, allows for selection of active promoters and a second reporter gene, the xylE gene which encodes catechol 2,3 dioxygenase, allows for screening for loss of promoter activity. In the first step of the I-TRAP approach, the recombinant plasmid library is transformed into bacteria expressing the transcriptional regulator. Transformants containing plasmids with active promoters are isolated by selection on kanamycin-containing media. A subset of the active promoters will be those whose expression depends on the transcriptional regulator. To differentiate between promoters dependent on the regulator and other promoters, plasmid DNA from a pool of the kanamycin-resistant transformants is transformed into bacteria that do not express the transcriptional regulator. The hygromycin Bresistant transformants are screened for XylE expression by spraying colonies with catechol. Yellow colonies arise from bacteria that contain a plasmid with a transcriptionally active promoter. White colonies arise from bacteria that contain a plasmid with a promoter that is inactive in the absence of the transcriptional regulator.
To develop the I-TRAP method for identifying M. tuberculosis promoters dependent on the expression of σ E , a strain of M. smegmatis over-expressing the M. tuberculosis σ E , a σ E mutant strain of M. smegmatis, and a M. tuberculosis promoter-trap library were created.

Construction of σ E -over-expressing strain of M. smegmatis
The σ E -over-expressing strain of M. smegmatis was made by transforming M. smegmatis LR222 bacteria with pSEint.2 ( Figure 2). The pSEint.2 plasmid contains an oriE for plasmid replication in E. coli, the mycobacteriophage L5 integrase and attP site to allow integration into the M. smegmatis chromosome, two selectable genes, ble (bleomycin or zeocin resistance) and amp (ampicillin resistance), and the M. tuberculosis RV1221 ORF cloned downstream of the M. tuberculosis hsp60 promoter. In this strain, the M. tuberculosis sigE gene is maintained in a single copy in the genome and is constitutively expressed from the hsp60 promoter. This strain is designated M. smegmatis LR222::SEint.2 ( Table 1).

Construction of sigE mutant strain of M. smegmatis
The suicide plasmid pSEKO.4 ( Figure 3) was constructed to interrupt the M. smegmatis sigE gene by gene replacement [14]. pSEKO.4 contains the M. smegmatis sigE gene interrupted with a gentamicin-resistance gene and the lacZ gene. The pSEKO.4 DNA (1 µg) was UV-irradiated in a UV Stratalinker 1800 (Stratagene) at 100 mJ cm -2 and electroporated into M. smegmatis cells. Following outgrowth at 37°C for 4 h, transformants were recovered by plating on TSA containing gentamicin and X-Gal. 24 white transformants were analyzed by PCR using primers (7997 and 7998) that bind to the sigE gene on either side of the aacC1 gene (Table 2). Of the 24 transformants, 11 had the M. smegmatis sigE gene disrupted by aacC1, 9 had only a single recombination event, and 4 only had a wildtype copy of sigE (data not shown). One transformant was chosen as the M. smegmatis sigE mutant strain and designated LR222∆ sigE ( Table 1). 20 µl total volume for 2 min at 37°C and the reaction stopped by heating at 65°C for 10 min. This generated fragments between 100 and ~6,000 bp (data not shown

Promoter selection and scoring
To recover plasmids containing active promoters, the plasmid library DNA was electroporated into M. smegmatis moter fragment. When the initial pool of kanamycinresistant colonies sprayed with catechol, the colonies were light to dark yellow indicating a wide range of XylE activity.
To differentiate between σ E -dependent and -independent promoters, approximately 40,000 kanamycin-resistant transformants were pooled, and plasmid DNA was isolated and electroporated into M. smegmatis LR222::∆ sigE bacteria, in which the wild type copy of the sigE gene had been disrupted with the gentamicin-resistance gene. Transformants were plated on hygromycin B and assayed for catechol 2,3 dioxygenase activity by spraying colonies with 0.5 M catechol. About 20,000 colonies were sprayed with catechol, and a range of color intensity from white to bright yellow was observed. A total of 360 white and light yellow colonies were recovered. Overall, about 1.8% of hygromycin B-resistant colonies were white or light yellow.
To confirm the phenotypes of plasmids recovered in this two-step procedure, plasmids were recovered from 18 white colonies and individually transformed into M. smegmatis LR222::SEint.2 (SigE-expressing) and M. smegmatis LR222::∆ sigE bacteria. All 18 plasmids generated only white colonies when transformed into the σ E -mutant strain. However, although all 18 plasmids generated yellow colonies when transformed into σ E -expressing bacteria, some plasmids produced a mixture of white and yellow colonies. These white colonies were consistently white on retesting and their plasmids usually contained deletions that removed the cloned M. tuberculosis sequence. Replating of the yellow colonies produced a mixture of white and yellow colonies, suggesting that the expression of the xylE or aph gene in the SigE-expressing strain may be detrimental to the cell and may lead to plasmid instability.
A preliminary analysis of the sequences of the inserts in the 18 clones identified 6 inserts that had matches with the SigE consensus sequence [6]. Interestingly, none of these genes were identified as being SigE-induced genes in microarray studies [6], although the expression of one of the genes, Rv3223c (sigH), is thought to be SigE-regulated [5].

Discussion
A key feature of the I-TRAP method is that it allows two degrees of promoter analysis, i.e., with and without a transcriptional regulator. The first step allows selection of all active promoters under one condition, and the second step sorts promoters by activity dependent on a second condition. In the study reported here, we used this method to identify potential σ E -induced genes. The basic I-TRAP approach might also be used to identify promoters repressed by a particular protein. The transformation steps are reversed so that the promoters are selected in the mutant background first (i.e., no repressor) and then screened for loss of activity in the presence of the repressor protein.
The I-TRAP method might also be used to characterize promoter activity in two environmental conditions in a manner similar to the IVET approach [16]. This is useful because, in some cases, the factor necessary for transcriptional activation during a particular environmental condition is unknown. In this case, after exposing bacteria containing the recombinant library to an environmental condition (e.g., a heat shock or acid shock or growth in medium A), antibiotics could be added to the medium to kill any bacteria that had not expressed the antibiotic resistance reporter gene during the exposure. For mycobacteria, a two-hour treatment with kanamycin would be sufficient to kill bacteria that had not expressed the kanamycin-resistance gene during the stress. The resistant bacteria would then be grown in a second medium or in the absence of a stress, and colonies scored for lack of promoter activity in the second condition by expression of the xylE reporter gene. White colonies would contain a plasmid with a promoter that is active only during the first condition. Using I-TRAP for this purpose may provide a means to identify promoters dependent on any of several regulatory proteins that may be active during the stress and to obtain a global overview of gene expression during the stress.
In the I-TRAP method, the recipient strain can be the native bacterium or a surrogate host. We used M. smegmatis as the host because it grows much more rapidly than does M. tuberculosis (generation time of 3 hrs vs. 24 hrs), does not require BSL-3 facilities, and is more easily manipulated genetically than M. tuberculosis. One possible advantage of using a surrogate host is that studies could be focused on one particular regulator which might avoid complications due to other regulatory proteins in the cell that may be able to recognize promoters in the library. Of course, the surrogate host must not have a regulatory protein that recognizes the same promoters as the regulator being studied. In our studies, the M. smegmatis σ E protein has 92% homology with the M. tuberculosis σ E protein [17]. While this suggests that the M. tuberculosis σ E protein should be able to interact properly with the M. smegmatis RNA polymerase, it also suggests that the M. smegmatis σ E homolog might be able to recognize M. tuberculosis σ Edependent promoters. Because of this, we constructed an M. smegmatis host strain that lacked any functional σ E protein.
An assumption of the I-TRAP approach is that the sequences cloned in the plasmids will be recognized by transcriptional regulatory proteins in the same manner as the intact sequences in the genome. The cloning process might generate false-positive or false-negative results because of the cloned sequences being recognized out of the context of the surrounding genomic sequences.
Another assumption of the I-TRAP approach is that overexpression of a transcriptional regulator will be sufficient to induce gene expression from its regulated promoters. This suggests that proteins, such as PhoP, that require activation to promote transcription may not be approachable with this method unless the over-expressed regulatory protein can also be activated. Also, certain regulatory proteins may not be suitable for study in a surrogate host bacterium [18]. For proteins that require the presence of additional factors, such as σ N , or interaction with host proteins to be active, such as σ E with RNA polymerase, use of the native bacterium or closely related surrogate host may be required to ensure the presence of the necessary interacting proteins [19]. Some of these additional factors, however, may only be expressed or active under a particular growth condition and therefore may also need to be conditionally expressed along with the transcription factor. Promoters that require both a positive transcription regulator and the absence of a repressor protein may be missed because expressing the positive transcription factor alone may not be sufficient to induce transcription at such promoters. Another limitation of the I-TRAP method is that some transcription factors may not be stably maintained in the bacterium or may be lethal to the bacterium when over-expressed. To avoid this problem, an inducible promoter could be used to promote transcription of the gene encoding the transcription factor.
Another possible complication of this approach is that expressing a foreign protein may turn on the stress response of the host bacterium. Transcription factors mediating the stress response to the expressed foreign protein may be able to recognize promoters within the library and promote transcription of the reporter genes, thus producing the false positives in the two-step screen.
Because the first step of the I-TRAP method involves selection for active promoters, one could easily begin with a library containing recombinants representing all possible promoters in a bacterial genome. Unfortunately, the recombinant DNA library used in this study contained only ~65% of the genome in the correct orientation upstream of the reporter genes, which will limit the number of different promoters we can recover using this library. For the screening step, we used the xylE gene because catechol 2,3 dioxygenase activity is easily measured by spraying colonies with catechol and the amount of color produced is a rough measure of promoter activity [10]. In our screen, white, pale yellow, and bright yellow colonies were observed, indicating a wide range of promoter activity. Pale yellow colonies may represent weak σ E -independent promoters or promoters that are recognized by another sigma factor in addition to σ E , providing a low level of constitutive expression. These possibilities could be distinguished by transformation of plasmids from individual clones into the σ E -over-expressing strain and directly comparing of XylE activity in the presence and absence of σ E . Unfortunately, it appears that expression of XylE in M. smegmatis may be detrimental and may lead to plasmid instability and loss of the inserted DNA. This raises the possibility that some white colonies may arise in the second step of I-TRAP by loss of a promoter fragment from a plasmid recovered in the first step. Nonetheless, the I-TRAP method should greatly enrich for promoters with the desired activities. As is generally true for genetic screens, it is essential to study any identified promoters in their native state to understand the regulation of their expression.

Conclusions
The I-TRAP method is a versatile and useful method for characterizing promoter activity under a variety of conditions and in response to various regulatory proteins. In our study, we isolated 360 clones that may contain plas-