Cellular and molecular mechanisms that mediate basal and tumour necrosis factor-α-induced regulation of myosin light chain kinase gene activity

The patients with Crohn's disease (CD) have a ‘leaky gut’ manifested by an increase in intestinal epithelial tight junction (TJ) permeability. Tumour necrosis factor-α (TNF-α) is a proto-typical pro-inflammatory cytokine that plays a central role in intestinal inflammation of CD. An important pro-inflammatory action of TNF-α is to cause a functional opening of intestinal TJ barrier. Previous studies have shown that TNF-α increase in TJ permeability was regulated by an increase in myosin light chain kinase (MLCK) gene activity and protein expression. The major aim of this study was to elucidate the cellular and molecular mechanisms that mediate basal and TNF-α-induced increase in MLCK gene activity. By progressive 5′ deletion, minimal MLCK promoter was localized between −313 to +118 on MLCK promoter. A p53 binding site located within minimal promoter region was identified as an essential determinant for basal promoter activity. A 4 bp start site and a 5 bp downstream promoter element were required for MLCK gene activity. TNF-α-induced increase in MLCK promoter activity was mediated by NF-κB activation. There were eight κB binding sites on MLCK promoter. The NF-κB1 site at +48 to +57 mediated TNF-α-induced increase in MLCK promoter activity. The NF-κB2 site at −325 to −316 had a repressive role on promoter activity. The opposite effects on promoter activity were due to differences in the NF-κB dimer type binding to the κB sites. p50/p65 dimer preferentially binds to the NF-κB1 site and up-regulates promoter activity; while p50/p50 dimer preferentially binds to the NF-κB2 site and down-regulates promoter activity. In conclusion, we have identified the minimal MLCK promoter region, essential molecular determinants and molecular mechanisms that mediate basal and TNF-α-induced modulation of MLCK promoter activity in Caco-2 intestinal epithelial cells. These studies provide novel insight into the cellular and molecular mechanisms that regulate basal and TNF-α-induced modulation of MLCK gene activity.

Recent animal studies have shown that MLCK plays a central role in immune, stress or bacterial endotoxin-mediated increase in intestinal permeability and subsequent inflammatory response [28,29,[42][43][44]. In these studies, the increase in intestinal permeability in mice was associated with an increase in intestinal MLCK gene and protein expression [28,29,[42][43][44]; and the inhibition of MLCK activity with known pharmacologic inhibitors including   [29,42,44]. [45]. These studies suggested that the increase in MLCK protein expression and activity contributes to the intestinal permeability increase in animals and humans; and MLCK has been identified as a potential therapeutic target to induce re-tightening of intestinal TJ barrier in inflammatory conditions [28,29,[42][43][44][45].

Human studies have also shown that patients with CD have an increase in MLCK protein expression. The increase in MLCK protein expression in CD intestinal tissue directly correlated with the level of intestinal inflammation
In previous studies, we have identified and cloned a functionally active MLCK promoter region [46]. These studies have shown that TNF-␣-induced increase in Caco-2 TJ permeability was mediated by an increase in MLCK gene activity and gene transcription, MLCK protein expression and activity [46]. However, the intracellular and molecular mechanisms that mediate the basal and TNF-␣-induced increase in MLCK gene activity and subsequent increase in MLCK protein expression remain unclear. The major aim of this study was to elucidate the cellular and molecular mechanisms that regulate the basal and TNF-␣-induced increase in MLCK gene activity and the subsequent increase in MLCK protein expression, using filter-grown Caco-2 intestinal epithelial cells. In the first part of this study, we identified the minimal MLCK promoter region and the essential molecular determinants that regulate the basal MLCK promoter activity in Caco-2 monolayers. In the second part, we identified the molecular determinants and the molecular processes that mediate the TNF-␣-induced increase in MLCK promoter activity and protein expression. These studies provide novel insight into cellular and molecular mechanisms that mediate the basal and TNF-␣-induced increase in MLCK gene activity and protein expression.

Luciferase assay
After the TNF-␣ treatment (

Western blot
MLCK and NF-B p65 protein expressions were assessed by western blot as previously described [32].

Determination of Caco-2 epithelial monolayer resistance
The effect of TNF-␣ on Caco-2 monolayer epithelial electrical resistance was measured using epithelial volt-ohmmeter (World Precision Instruments, Sarasota, FL, USA) as previously described [48,49]. For resistance measurements, both the apical and basolateral sides of the epithelia were bathed in regular growth media. Electrical resistance was measured until similar values were recorded on three consecutive measurements. Each experiment was repeated at least three times in quadruplicate to ensure reproducibility.

Statistical analysis
The values of experimental data were expressed as the mean Ϯ S.E., and analysed using two-tailed unpaired t-test ( (Fig. 1A). The sequence of MLCK promoter has no TATA box near the transcriptional initiation site. However, a CAAT box and several E boxes are located within the minimal promoter region. Also, a downstream promoter element (DPE) AGAGC which normally resides in TATA less promoter is located at ϩ36 to ϩ40 [51]. These findings suggested that the regulatory elements required for the gene transcription were present in this region (Fig. 1A).

Identification of MLCK minimal promoter region
Previously, we cloned a 2091 bp MLCK promoter region (Genbank accession number DQ 090939) into a pGL3 basic vector [46]. The sequence of the full-length (FL) 2091 bp MLCK promoter region is shown in Figure 1B. In the following studies, we determined the minimal promoter region. For this purpose, a progressively increasing 5Ј deletion constructs were generated (total of eight) and subcloned into the pGL3 basic vector (Fig. 2). Deletion constructs were then transfected into confluent Caco-2 cells and the promoter activity determined by luciferase assay (Fig. 2). As shown in Fig. 2, progressive deletion of 5Ј end extending to -314 did not result in a significant increase in promoter activity compared to the FL promoter (2091 bp). The deletion construct MLCK -313 exhibited a 2-fold increase in promoter activity compared to the FL promoter (Fig. 2). The deletion of 5Ј end extending beyond -246 resulted in a marked decrease in promoter activity. Based on these results, we identified -313 to ϩ118 as a minimal MLCK promoter region.

Regulation of basal MLCK promoter activity
As shown in Figure 2, extending the deletion to include 68 bp MLCK promoter region between -313 to -245 resulted in a marked decrease in promoter activity. These findings suggested that a reg-ulatory site within this 68 bp region could have a critical role in the maintenance of basal MLCK promoter activity. Using the Genomatix/Promoter Inspector software, a p53 transcription factor binding motif (-294 to -275) was identified within this 68 bp region (Fig. 1B). To determine the possible regulatory function of this p53 binding motif (CCCCTGCCAGGGCCTCTCCC) on basal promoter activity, the p53 binding site was mutated via site-directed mutagenesis in the construct MLCK -313 (which encodes the minimal promoter region). The mutation of p53 site (-294 to -275) resulted in a near complete inhibition of promoter activity (Fig. 3). It should be noted that the basal promoter activity of the p53 mutant MLCK promoter was similar to deletion construct MLCK -245 (which lacks the p53 binding region) (Fig. 2), indicating that p53 binding region has a critical role in the regulation of basal MLCK promoter activity. Thus, the sharp drop in promoter activity between MLCK -313 and MLCK -245 observed in Fig. 2 could be explained by the absence of p53 site. To further substantiate the role of p53 in basal promoter activity, p53 expression was knockdown by p53 siRNA transfection of Caco-2 monolayers. The p53 siRNA transfection also resulted in a marked decrease in basal MLCK promoter activity (Fig. 3B), confirming that p53 plays an important role in the regulation of basal MLCK promoter activity.
Next, the functional activity of the transcription start region was experimentally validated by selective deletion and site-directed mutagenesis. Using MLCK -313 (which encompasses the minimal promoter region) as the 'wild-type' template, a deletion construct lacking the eight base pair -6 to ϩ2 (CCACCCAC) start region was generated (Fig. 4). The deletion of 8 bp start sequence resulted in a significant decrease in luciferase activity (Fig. 4). To further narrow the start site, a four base pair mutant which replaces CCAC (-2 to ϩ2) with TTGT was generated. The 4 bp mutation also resulted in a similar decrease in luciferase activity as the 8 bp deletion construct (Fig. 4), suggesting that transcription start site (TSS) was located within this four base pair sequence. To determine the role of DPE in basal gene transcription process, a double mutation of the TSS and DPE was generated. The double mutation resulted in almost complete absence of the promoter activity (Fig. 4), confirming that these two sites were essential for the initiation of MLCK gene transcription.

NF-B regulation of MLCK promoter activity
Previously, we reported that TNF-␣-induced increase in MLCK gene transcription was mediated by nuclear transcription factor NF-B [32]. In the following studies, we expanded on our previous findings to delineate the specific molecular determinants on MLCK promoter that mediated the NF-B action. Eight putative NF-B binding sites were identified on MLCK promoter using the Genomatix/Promoter Inspector software. The location of each of the eight NF-B binding sites is shown in Figure 1B. Note that seven B sites (NF-B2 to 8 sites) were located outside the minimal promoter region and one B site (NF-B1 site) was located within the minimal region. To delineate the specific B sites that mediated the TNF-␣-induced upregulation of MLCK promoter activity, the TNF-␣ effect on promoter activity of deletion constructs having progressively larger 5Ј end deletion was determined. As shown in Fig. 5 Fig. 5B and C), indicating that NF-B2 to 8 sites were not necessary for the TNF-␣-induced up-regulation of MLCK promoter activity. These findings suggested that the NF-B1 site (located within the minimal region) may be the regulatory site that mediated the TNF-␣-induced up-regulation of MLCK promoter activity. To test this possibility, NF-B1 motif was mutated in MLCK -313. In the wild-type MLCK-313, TNF-␣ caused an increase in MLCK promoter activity (Fig. 6A). The mutation of the NF-B1 site completely prevented the TNF-␣-induced increase in promoter activity (Fig. 6A). In separate studies, the effect of NF-B1 mutation on FL MLCK promoter activity was also determined. The mutation of the NF-B1 site in the FL MLCK also prevented the TNF-␣-induced increase in promoter activity (Fig. 6B). These results indicated that the NF-B1 site was required for the TNF-␣-induced increase in promoter activity.

, TNF-␣ caused a significant increase in MLCK promoter activity in the FL MLCK promoter construct compared to the control (P Ͻ 0.05). TNF-␣ also caused a similar proportional increase in MLCK promoter activity in the deletion constructs lacking the NF-B2 to 8 sites (
Next, the binding of TNF-␣ activated NF-B to the NF-B1 binding site on MLCK promoter region was determined using an ELISA-based DNA-binding assay. In these studies, 50 bp doublestranded DNA probe encoding the NF-B1 site (30 bp up-and 10 bp downstream of NF-B1 site) on MLCK promoter region was synthesized and used as a probe to assess the binding of NF-B to the NF-B1 motif. The nuclear extracts from control and TNF-␣ treated Caco-2 monolayers were used to assess the binding of NF-B to the DNA probe. TNF-␣ treatment resulted in a significant increase in NF-B binding to the NF-B1 binding site on the DNA probe (Fig. 6C). The mutation of NF-B1 binding site on DNA probe prevented the TNF-␣ increase in NF-B binding to the DNA probe (Fig. 6C). These findings confirmed that TNF-␣ causes an increase in NF-B binding to the NF-B1 binding motif, which presumably leads to the activation of MLCK promoter.

p53 binding site does not play a role in TNF-␣induced increase in MLCK promoter activity
Above studies indicated that the p53 site (-294 to -275) plays a crucial role in basal MLCK promoter activity. The possible involvement of this site in mediating the TNF-␣ modulation of MLCK promoter activity was also examined by site-directed mutagenesis of p53 binding motif. The mutation of p53 binding site did not inhibit the TNF-␣-induced increase in MLCK promoter activity (Fig. 7).

Molecular mechanisms of TNF-␣ regulation of MLCK promoter activity
To delineate the molecular mechanisms involved in NF-B regulation of MLCK promoter activity, the role of NF-B1 and 2 sites in the regulation of MLCK promoter activity was examined. As shown in Fig. 2 (Fig. 2), suggesting that the NF-B2 site may have a repressive effect on promoter activity. To investigate the possible interactive relationship between NF-B1 and 2 site, the effect of NF-B1 site mutation in the presence of NF-B2 site was determined using MLCK -509 (which contains the NF-B1 and 2 sites) (Fig. 8). The mutation of NF-B1 site resulted in a marked decrease in MLCK -509 promoter activity (Fig. 8A)     corresponding to the NF-B1 region and 5Ј-TGCAGGAAG-GCAGCTCCCATGGCCT-3Ј corresponding to the NF-B2 region). As shown in Fig. 9A, TNF-␣ (10 ng/ml) treatment resulted in a marked increase in NF-B p50/p65 heterodimer binding to 32 Plabelled NF-B1 DNA probe. There was only a small increase in p50/p50 dimer binding. In contrast, p50/p50 dimer was the predominant dimer-type binding to the NF-B2 site in response to TNF-␣ treatment (Fig. 9A). These findings indicated that p50/p65 was the predominant dimer-type binding to the NF-B1 site while p50/p50 was the predominant dimer-type binding to the NF-B2 site following TNF-␣ stimulation of Caco-2 cells. The specificity of NF-B dimer binding to the 32 P-labelled probes was confirmed by addition of excess (100-fold higher concentration) 'cold' or nonradioactive labelled probe. The excess of non-radiolabelled probes inhibited the NF-B dimer binding to the 32 P-labelled probes (Fig. 9A). To further validate the above EMSA results, ELISA-based antibody labelling studies were also carried out. In these studies, two 50 bp DNA probes were synthesized, encoding either the NF-B1 or 2 site plus 30 bp up-and 10 bp downstream of the B motif. TNF-␣ treatment resulted in a similar proportional increase in both p65 and p50 binding to the NF-B1 site (Fig. 9B). In contrast, TNF-␣ caused a significant increase in p50 binding but did not cause an increase in p65 binding to the NF-B2 site (Fig. 9C). Together, these results indicated that in response to TNF-␣ treatment, p50/p65 dimer was the predominant dimer-type binding to the NF-B1 site, while p50/p50 dimer was the major dimer-type binding to the NF-B2 site. These results suggested that p50/p65 dimer binding to the NF-B1 site up-regulates the promoter activity, while p50/p50 binding to the NF-B2 site downregulates the promoter activity [52][53][54][55].

, deletion of six upstream NF-B sites (NF-B3 to 8) between -1973 and -509 did not affect the basal MLCK promoter activity. However, extending the deletion to include NF-B2 site between -325 and -316 resulted in a 2-fold increase in basal promoter activity
Based on the above results, we hypothesized that the TNF-␣induced increase in MLCK promoter activity and subsequent increase in MLCK gene and protein level were regulated by NF-B p50/p65 dimer binding to the NF-B1 site. To validate the role of NF-B p50/p65 dimer on MLCK promoter activation, NF-B p65 expression was silenced in Caco-2 monolayers by NF-B p65 siRNA transfection. As shown in Figure 10A, p65 siRNA transfection resulted in a near-complete depletion of NF-B p65 in Caco-2 monolayers. NF-B p65 silencing resulted in a complete inhibition of TNF-␣-induced increase in MLCK promoter activity in MLCK -313 (Fig. 10B), indicating that NF-B p65 was required for the NF-B1 site up-regulation of promoter activity. In contrast, NF-B p65 siRNA did not affect the TNF-␣-induced decrease in promoter activity in MLCK -509 Mu (which contains functionally active NF-B2 site, but a mutated NF-B1 site) (Fig. 10B), confirming that p50/p65 was not involved in the NF-B2 site down-regulation of promoter activity. The NF-B p65 silencing also inhibited the TNF-␣-induced increase in MLCK protein expression, indicating that p50/p65 mediated the TNF-␣ increase in MLCK protein expression (Fig. 10C). Moreover, NF-B p65 silencing also prevented the TNF-␣-induced drop in Caco-2 transepithelial resistance (Fig. 10D). Together, these results indicated that NF-B p50/p65 binding to the NF-B1 site causes an up-regulation of MLCK gene activity, which leads to an increase in MLCK protein expression and functional opening of the Caco-2 TJ barrier. [41,46]. These studies indicated that TNF-␣ causes a rapid activation of NF-B resulting in a cytoplasmic-to-nuclear translocation of NF-B. The activated NF-B binds to its binding site on the MLCK promoter and causes a sequential increase in MLCK gene activity, MLCK protein expression and MLCK activity, and opening of TJ barrier [41,46]. Although previous studies have shown that MLCK gene activation was a key intracellular process mediating the TNF-␣-induced increase in MLCK protein expression/activity [46], the molecular mechanisms and the intracellular determinants that regulate MLCK gene activity remain unclear. In the present study, we extended on our previous observations and investigated the cellular and molecular mechanisms that mediate the basal and TNF-␣-induced regulation of MLCK gene activity. Our studies have for the first time localized the minimal MLCK promoter region to -313 to ϩ118 and identified p53 as a key transcription factor regulating the basal promoter activity. Our studies also delineated the TSS and the DPE that were essential for the initiation of MLCK gene transcription. Additionally, using computerized software we identified eight B sites on the MLCK promoter region and delineated the specific B site responsible for mediating the TNF-␣-induced up-regulation of MLCK promoter activity. Using combination of molecular approaches, we were also able to provide mechanistic insight into the molecular interactions that mediate the TNF-␣ regulation of MLCK promoter activity.

Previous studies from our laboratory have shown that the TNF-␣induced increase in intestinal epithelial TJ permeability was regulated in part by NF-B-induced increase in MLCK gene activity and MLCK protein expression
The NF-B family of transcription factors is known to regulate a wide range of biological activities, including inflammatory responses, cellular proliferation and apoptosis [56]. The NF-B family consists of p50, p65, RelB, cRel and p52 subunits. These five NF-B subunits form various combinations of dimers that mediate NF-B dependent transcriptional regulation [57]. The p50/p65 and p50/p50 dimers are the most common dimer types in cells [58][59][60]. It has been shown that p50 and p65 subunits are ubiquitously expressed in various cell types [58][59][60]. The p50/p65 is the predominant dimer type (Ͼ90%) and p50/p50 is a minor dimer type present in cells [61][62][63]. The basal expression of other family members varies greatly and their presence depends on the specific cell type.
In this study, we examined the regulatory actions of two B sites; within (NF-B1: ϩ48 to ϩ57) and external (NF-B2: -325 to -316) to the minimal promoter region (Fig. 1B). Interestingly, our data indicated that the two B sites had opposite effects on promoter activity (Fig. 8). To provide molecular explanation for the opposite effects, we examined the possibility that different NF-B dimer types may be binding to the B sites. The NF-B subunitbinding studies indicated that p50/p65 preferentially binds to the NF-B1 site, while p50/p50 preferentially binds to the NF-B2 site (Fig. 9). Based on these findings, we hypothesized that p50/p65 binding to the NF-B1 site induces an activation of promoter activity, while p50/p50 binding to the NF-B2 site causes a repression of the promoter activity. In support of such possibility, previous studies have shown that p65 subunit contains a potent transactivation domain that leads to the activation of gene promoter region [64,65]. In contrast, p50 lacks the transactivation domain and functions only in DNA binding [64]. The p50/p50 dimers have been shown to act as a negative regulator of NF-B activity by competing with p50/p65 for the NF-B response elements [65]. The p50/p50 homodimers preferentially bind to the GC rich binding sites in comparison to p50/p65 dimers [66]. Consistent with this, our data showed that p50/p50 dimers have higher affinity to   the NF-B2 site (GGCAGCTCCC) (which has higher GC content) than the p50/p65 dimers. Correspondingly, p50/p65 dimers had higher affinity to the NF-B1 site (GGAGCTTCCC) (which has lower GC content) than the p50/p50 dimers. The mutation and siRNA knock-down studies validated the hypothesis that p50/p65 binding to the NF-B1 site causes an increase in MLCK promoter activity, while p50/p50 binding to the NF-B2 site causes a decrease in MLCK promoter activity.
The net result of TNF-␣ stimulation of Caco-2 cells is to activate MLCK promoter activity, which in turn leads to an increase in MLCK protein expression and increase in TJ permeability in Caco-2 cells (Fig. 10) [46]. Our data suggested an interesting possibility that during quiescent (unstimulated) conditions, the NF-B2 site may have a repressive action on MLCK promoter activity (which leads to an enhancement of TJ barrier function). In constrast, during TNF-␣ stimulation, activated p50/p65 dimers bind to the NF-B1 site, causing an increase in MLCK promoter activity. Although NF-B1 and NF-B2 sites have opposite effects on MLCK promoter activity, it appears that the TNF-␣-induced stimulatory effect of NF-B1 site is greater than the repressive effect of NF-B2 site, resulting in a net stimulation of MLCK promoter activity.
Although our studies indicated that TNF-␣-induced activation of NF-B leads to an increase in MLCK promoter activity, an interesting question remained as to whether NF-B activation by other cytokines can also cause an increase in MLCK gene activity. To address this question, we examined the effect of another potent NF-B activating cytokine IL-1␤ on MLCK promoter activity. IL-1␤ treatment resulted in a rapid activation of NF-B and an increase in MLCK promoter activity (unpublished data), suggesting that NF-B activation by other pro-inflammatory mediators can also induce activation of MLCK gene.
In conclusion, our studies have for the first time identified several molecular determinants that play an essential role in the regulation of basal and TNF-␣-stimulated MLCK promoter activity. Our data indicated that MLCK minimal promoter region was located between -313 and ϩ118 on MLCK promoter region. The p53 binding site (-294 to -275) located within the minimal promoter region was an essential determinant for basal promoter activity. The TNF-␣-induced increase in MLCK promoter activity was mediated by NF-B1 site (ϩ48 to ϩ57), and deletion of this site resulted in a complete loss of response to TNF-␣ treatment. The NF-B1 and NF-B2 (-325 to -316) sites have opposite regulatory action on MLCK promoter activity. A likely molecular explanation for the differential effect relates to the differences in the NF-B dimer type that preferentially binds to the B sites. The p50/p65 dimer binding to the NF-B1 site up-regulates the MLCK promoter activity, while p50/p50 binding to the NF-B2 site down-regulates the promoter activity. These findings provide novel insight into the molecular mechanisms that regulate basal and TNF-␣-induced modulation of Caco-2 MLCK gene activity and the subsequent opening of the intestinal epithelial TJ barrier.