Pic, an Autotransporter Protein Secreted by Different Pathogens in the Enterobacteriaceae Family, Is a Potent Mucus Secretagogue

Abstract

A hallmark of enteroaggregative Escherichia coli (EAEC) infection is a formation of biofilm, which comprises a mucus layer with immersed bacteria in the intestines of patients. While studying the mucinolytic activity of Pic in an in vivo system, rat ileal loops, we surprisingly found that EAEC induced hypersecretion of mucus, which was accompanied by an increase in the number of mucus-containing goblet cells. Interestingly, an isogenic pic mutant (EAEC Δpic) was unable to cause this mucus hypersecretion. Furthermore, purified Pic was also able to induce intestinal mucus hypersecretion, and this effect was abolished when Pic was heat denatured. Site-directed mutagenesis of the serine protease catalytic residue of Pic showed that, unlike the mucinolytic activity, secretagogue activity did not depend on this catalytic serine protease motif. Other pathogens harboring the pic gene, such as Shigella flexneri and uropathogenic E. coli (UPEC), also showed results similar to those for EAEC, and construction of isogenic pic mutants of S. flexneri and UPEC confirmed this secretagogue activity. Thus, Pic mucinase is responsible for one of the pathophysiologic features of the diarrhea mediated by EAEC and the mucoid diarrhea induced by S. flexneri.

In the last 25 years, the relevance of Escherichia coli as an etiological agent of infectious gastroenteritis has increased considerably. Through phenomena such as horizontal transfer, diarrheagenic E. coli strains have acquired mobile genetic elements, and there now are different recognized pathogenic varieties or pathotypes of this species, including enterotoxigenic E. coli (ETEC), enterohemorrhagic E. coli (EHEC), enteropathogenic E. coli (EPEC), enteroinvasive E. coli (EIEC), diffusely adherent E. coli (DAEC), and enteroaggregative E. coli (EAEC) (24). Additionally, there are E. coli pathotypes that share pathogenic mechanisms with other genetically related bacterial strains, such as EIEC and Shigella spp. (24) or EAEC and the extraintestinal E. coli strains called uropathogenic E. coli (UPEC) (1).

Kaper et al. (24) proposed a pathogenicity model for EAEC that consists of three stages. The first involves adhesion of microorganisms to the intestinal mucosa through fimbrial adhesion, such as by aggregative adherence fimbriae (AAFs). During the second stage there is intestinal mucus hypersecretion, keeping microorganisms immersed in a gel matrix favoring persistent colonization, and perhaps a malnutrition step. Finally, the third stage involves secretion of proteins with enterotoxic/cytotoxic activities that cause histopathological alterations. EAEC is able to secrete two autotransporter proteins. In addition to the Pet protein, a 104-kDa cytotoxin that cause damage to epithelial cells (35), another protein of 116 kDa has also been found in EAEC supernatants (34). This protein was named Pic, for protein involved in intestinal colonization, due to its biological activity (22). Pic, unlike Pet, is not able to increase ion secretion in the Ussing chambers or enterotoxic damage in the rat jejunal mucosa (34). The pic gene is carried in the chromosomes of EAEC, UPEC, and Shigella flexneri strains in an open reading frame of 4,116 nucleotides. In silico analysis and biological characterization revealed that the pic gene lies in the DNA coding strand, whereas two genes (set1A and set1B) encoding ShET1 protein, which is involved in intestinal toxicity, are carried on the opposite strand but completely within the pic gene (22). This toxin is identical to ShET1 toxin (55 kDa) secreted by S. flexneri 2a (15), which cause fluid accumulation in rabbit ileal loops and increase the short circuit current in Ussing chambers (14); however, the role of ShET1 in EAEC has not yet been determined. In S. flexneri, the pic gene is part of a pathogenicity island of 46,603 bp called she (2).

Henderson et al. (22) showed that Pic has diverse biological activities, such as proteolytic activity on complement system proteins, which could promote EAEC permanence in the intestine; Pic also degrades mucin from different sources (egg and bovine submaxillary gland) and crude intestinal mucus of mice. Other authors have shown that Pic does not produce cytotoxic effects on HEp-2 cells during 5 h of incubation and is unable to degrade ovine spectrin or pepsin but that is able to degrade human coagulation factor V (12). In contrast, PicU (from UPEC) degraded human spectrin as well as pepsin, coagulation factor V, and bovine submaxillary gland mucus (37). Due to these contradictory results and to the fact that pic is found in three important pathogens (EAEC, S. flexneri, and UPEC), our group sought to identify intestinal substrates for Pic. We found that Pic does not cause damage to cultured epithelial cells, cleave fodrin, or degrade host defense proteins (secretory IgA [sIgA], lactoferrin, and lysozyme). However, Pic is able to degrade intestinal and submaxilliar mucus in a dose-dependent fashion; such an effect depends on its serine protease motif (18). Interestingly, Pic is able to bind mucin, and this binding is blocked in a competitive assay using constitutive monosaccharides of mucin. Moreover, the mucinolytic activity decreased when the monosaccharide sialic acid was removed from the mucin (18). Here we show that Pic apparently has two contradictory effects: besides the mucinolytic activity due to its serine protease motif, Pic is also able to cause mucus hypersecretion and an increase in the number of mucus-producing goblet cells. These effects are independent of the serine protease motif and are shared by the Pics secreted by UPEC and S. flexneri.

MATERIALS AND METHODS

Bacterial strains and plasmids.

The bacterial strains and plasmids used in this work are described in Table 1. All strains were cultivated under aerobic conditions at 37°C (unless otherwise specified) in Luria-Bertani (LB) broth or Luria agar for 16 to 18 h. When antibiotics were required, they were used at the following final concentrations: tetracycline, 10 μg/ml; kanamycin, 50 μg/ml; and ampicillin, 100 μg/ml.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Characteristics	Reference
EAEC 042	EAEC prototype strain, isolated in Peru	33
Shigella flexneri 2a	S. flexneri 2a wild-type strain	16
UPEC CFT073	UPEC wild-type strain, isolated from a pyelonephritis case	21
EAEC 042 Δpic	EAEC 042 pic gene mutant	20
S. flexneriΔpic::kan	S. flexneri isogenic pic gene mutant	This study
UPEC CFT073 Δpic	UPEC isogenic pic gene mutant, Cmr	21
UPEC CFT073 ΔhlyA	UPEC isogenic hlyA gene mutant, Kamr	This study
EAEC Δpic/pPic1	EAEC 042 Δpic complemented with pic gene	This study
EAEC Δpic/pPicS258I	EAEC 042 Δpic complemented with pic gene (mutation in catalytic site)	This study
EAEC 042 S258A	EAEC 042 chromosomally mutated in the serine protease motif of pic	20
E. coli HB101	E. coli K-12 derivative	4
HB101/pPic1	E. coli HB101 transformed with pPic1 (pic clone), Tetr	22
HB101/pPicS258I	E. coli HB101 transformed with pPicS258I (pic mutated in the catalytic site), Tetr	18
pKD4	Template plasmid for Kamr gene	10
pKD46	Plasmid containing phage λ Red system (Ampr), heat sensitive	10

Recombinant Pic and PicS258I proteins.

Clones E. coli HB101/pPic1 (22) and E. coli HB101/pPic258I (this study) were cultivated in LB-tetracycline for 16 h at 37°C with shaking. Supernatants were obtained by centrifugation at 7,000 ×g for 20 min at 4°C and sterilized by filtration through 0.22-μm cellulose acetate membrane filters (Corning, Cambridge, MA), concentrated 100-fold with an Ultrafree centrifugal filter device with a 100-kDa cutoff (Millipore, Bedford, MA), filter sterilized again, and stored at −20°C for up to 3 months (35).

Molecular cloning and construction of mutants.

All genetic manipulations were performed by standard methods (3). Plasmid DNA was extracted by using a plasmid Midi kit (Qiagen Inc., Chatsworth, CA). Purification of DNA fragments and extraction from agarose gel slices were performed with Geneclean (Bio 101, La Jolla, CA). Plasmid DNA was introduced into E. coli HB101 or pic mutants by transformation of competent cells (Gibco/BRL, Gaithersburg, Md.) according to the method of Hanahan (19).

To generate an isogenic pic mutant of Shigella flexneri 2a, the pic gene (accession number SFU35656) was interrupted by a gene encoding kanamycin resistance by use of the lambda Red recombinase system (10). The kanamycin resistance gene was amplified from pKD4 by PCR with primers pic-FRT-sense (5′-TTT TAC TTT TAT ATC CCT TGT AAA CAT CAT GGA GAA TCC ATA GTG TGT GTA GGC TGG AGC TGC TT) and pic-FRT-antisense (5′-GAA CAT ATA CCG GAA ATT CGC GTT TAC CGC ATT ATC CAT ATG AAT ATC CTC CTT AG). The product was treated with DpnI and introduced into Shigella flexneri 2a carrying pKD46. Colonies containing the pic::Kan interrupted gene (referred to as Shigella flexneri Δpic) were then obtained as described previously (10).

SDS-PAGE and Western blotting.

Samples were mixed with Laemmli buffer containing 5% 2-mercaptoethanol and analyzed by using polyacrylamide gel electrophoresis under denaturing conditions (SDS-PAGE) at 8, 10, or 12% (25). Gels were stained with Fairbanks stain for 1 h and destained with a methanol-water-acetic acid mix.

For Western blotting, proteins were separated by SDS-PAGE, transferred to nitrocellulose membranes, blocked for 1 h in phosphate-buffered saline (PBS)-Tween 20 (0.05%, vol/vol) and nonfat milk (5% wt/vol), and probed with primary antibodies for 2 h (46). Bound antibody was detected with horseradish peroxidase-conjugated secondary anti-rabbit antibody and enhanced chemiluminescence reagents (Amersham, Naperville, IL).

Rat ligated ileal loop assays.

To evaluate Pic activity on intestinal mucin in vivo, the rat ligated ileal loop model was used (9, 42, 48). Sprague-Dawley rats (n = 4) of either sex and between 70 and 100 g in weight were used. Before the assay, rats were starved for 16 h. All the animal experiments were done after approval from the CINVESTAV-IPN Animal Ethics Committee (CICUAL).

To perform laparotomy and expose the small intestines, animals were anesthetized with Xylazine (6.5 mg/kg) and Ketamine (34.5 mg/kg); once the intestines were exposed, three ileal loops of 3 cm with 1 cm between them were ligated. Such ileal loops were inoculated with 200 μl of solutions of 1.5 × 10⁸ CFU/ml of the indicated strains (Table 1). In other experiments, ileal loops were also inoculated with 100 μg of Pic, PicS258I, or denatured Pic or Pet proteins. As a negative control, basal mucus secretion, PBS, or E. coli HB101 was used. After inoculation, intestines were returned to the abdominal cavity and the incision was sutured. Inoculated rats were kept alive for 12 h and then sacrificed by cervical dislocation. Ileal loops were dissected, and the intestinal contents were collected by gentle pressure using cold sterile deionized water. Glycoprotein concentrations were quantified by a colorimetric method using periodic acid-Schiff staining (PAS).

For histopathological analyses, intestinal tissues fixed in 4% formaldehyde were embedded in paraffin to obtain sagittal sections (5-μm thick). Tissues were deparaffinized and rehydrated in 3% acetic acid for 2 min. Staining was performed by submerging tissue sections for 30 min in a 1% Alcian blue solution in 3% acetic acid, the dye excess was removed with distillated water, nuclear fast red was applied for 1 to 5 min, and that dye excess was also removed with distillated water. Finally, tissues were washed twice with 100% methanol and once with xylene, mounted in Gelvatol (Sigma), and observed under a Leica DM4500 B microscope (39).

Glycoprotein determination by the PAS method.

To determine glycoprotein concentrations in samples from rat intestinal mucus, a colorimetric method was used (29). Schiff reagent was prepared by dissolving 1 g basic fucsin in 100 ml of water (at 100°C). When this solution reached 50°C, 20 ml of 1 N HCl and 300 mg of activated charcoal were added, mixed for 5, min and then filtered. Before use, sodium metabisulfite (0.1 g for 6 ml of reagent) was added and incubated until the reagent became uncolored. For the glycoprotein standard curve, hog gastric mucins were used and adjusted to 100, 50, 25, 12.5, 6.2, 3.1, 1.5, and 0.7 mg/ml in water. For determination of glycoprotein concentrations in experimental samples, 5 μl of intestinal contents in 155 μl of PBS was used. To each enzyme-linked immunosorbent assay (ELISA) plate well containing the samples, 20 μl of a solution of acetic acid and periodic acid (1 ml of 7% acetic acid and 1 ml of 50% periodic acid) was added and incubated for 2 h at 37°C. After incubation, 20 μl of Schiff reagent (Sigma-Aldrich) was added, and color was allowed to develop for 30 min at room temperature. Absorbance was read at 595 nm in an ELISA microplate reader and compared with the standard curve values to obtain the mucin concentration in the sample.

Kinetics of mucus secretion induced by bacteria harboring the pic gene.

The ileal loop assay was performed as described above using EAEC 042 and UPEC CFT073. However, for extraction of intestinal contents and tissue, rats were sacrificed at the 1, 2, 4, 6, 8, 10, 12, 14, and 16 h postinfection. The glycoprotein concentration was determined by using the PAS method as described above.

RESULTS

Secretion of Pic by EAEC, UPEC, S. flexneri, and their pic isogenic mutants.

Bacterial strains were grown overnight, and the supernatants were analyzed for Pic by 9% SDS-PAGE and Western blotting. Wild-type EAEC 042, UPEC CFT073, and S. flexneri 2a showed a protein band of high molecular mass (109 kDa), which corresponded to the Pic protein detected with anti-Pic antibodies (Fig. 1A). However, their isogenic pic mutants (EAEC Δpic, UPEC Δpic, and S. flexneri Δpic) were unable to secrete Pic. To demonstrate that genetic manipulation in constructing the mutants did not interfere with Pic expression and secretion, a mutant with a mutation in a different gene, i.e., hylA from UPEC, was constructed. The UPEC ΔhlyA mutant (hemolysin negative) was able to produce and secrete Pic (Fig. 1A). These data indicate that independently of the different genetic backgrounds of these three pathogens, all of them are able to efficiently secrete the Pic protein. Since the pic gene has been cloned in a plasmid, pPic1 (22), and we have constructed a mutant with a site-directed mutation in the serine protease motif of Pic, pPicS258I (18), both plasmids were used to complement the isogenic mutant EAEC Δpic. Both complemented strains (EAEC Δpic/pPic and EAEC Δpic/pPicS258I) recovered Pic expression and secretion (Fig. 1B). Furthermore, anti-Pic antibodies specifically recognized Pic (109 kDa) in supernatants from EAEC 042 strain and from the complemented strains but not in supernatants from EAEC Δpic (Fig. 1C). However, Pet was detected in all the supernatants by using anti-Pet antibodies. Thus, in the complemented strains we have the advantage that one strain secretes Pic which is inactive at the catalytic site (PicS258I) and the other secretes native Pic (Fig. 1C), allowing us to determine the role of the serine protease motif.

FIG. 1.

Pet secretion induced by bacterial pathogens harboring the pic gene. Proteins were precipitated from supernatants with 100% trichloroacetic acid and then resolved by 10% SDS-PAGE (B) and transferred to nitrocellulose membranes for immunoblotting (A and C). Membranes were incubated with a primary antibody against Pic from EAEC 042 (dilution, 1:500) (A) or against Pet from EAEC 042 (dilution 1:500) (C) and an alkaline phosphatase-conjugated goat anti-rabbit IgG secondary antibody (1:3,000). The reaction was developed using the one-step nitroblue tetrazolium (NBT)/5-bromo-4-chloro-3-indolylphosphate (BCIP) substrate (Pierce).Wt, wild type.

Pic from EAEC causes an increase in intestinal mucus secretion in ileal loops.

To analyze the biological role of Pic in intestinal mucus components in an in vivo model, we used the rat ileal loop model, which has been helpful in characterizing histopathological lesions caused by EAEC 042 (48). Bacterial strains were grown and adjusted to 1.5 × 10⁸ CFU/ml to be inoculated into the lumina of intestinal loops of Sprague-Dawley rats (n = 4). After 12 h of incubation, the rats were sacrificed and the loops dissected to collect the intestinal mucus in 1 ml of sterile water. Glycoprotein concentrations secreted into the intestinal lumen were determined by use of the PAS colorimetric method, and the intestinal tissues were fixed for histopathological analyses.

With the prototype EAEC 042 strain, an increase of the intestinal mucus secretion into the ileal loops was observed; the glycoprotein concentration was 11.34 mg/ml, representing a significant increase of 2-fold compared to that found in the ileal loops inoculated with the laboratory strain E. coli K-12 HB101 (4.84 mg/ml) (P < 0.05) (Fig. 2). E. coli HB101 was unable to increase intestinal mucus secretion, and the values were similar to those basally detected in ileal loops injected with the vehicle, PBS. Surprisingly, when the ileal loops were inoculated with the isogenic pic mutant (EAEC 042 Δpic), the amount of glycoproteins secreted to the intestinal lumen decreased significantly, to 3.79 mg/ml, compared to that induced by the wild-type strain (P < 0.05). On the other hand, in ileal loops inoculated with the complemented strains EAEC Δpic/pPic1 and EAEC Δpic/pPicS258I, secreted glycoprotein concentrations were 8.59 mg/ml and 7.89 mg/ml, respectively. Such concentrations were higher than that obtained with the EAEC Δpic mutant, which was 3.79 mg/ml (P < 0.05) (Fig. 2). The effect of inoculation of the strain containing the pic gene cloned in E. coli HB101 (HB101/pPic1) was similar to that of the mutant bacteria complemented with pic; the glycoprotein concentration was 7.78 mg/ml, which was significantly higher than that obtained by inoculating the parental HB101 (P < 0.05) (Fig. 2). Since the complemented cells recovered the induction of mucus secretion in the ileal loops but never to the levels reached by the wild type and since complementation with the pic gene mutated in the serine protease motif still caused intestinal mucus hypersecretion, we decided to use a strain which was mutated in the active site constructed directly in the chromosome of EAEC 042 (20) to avoid complementation in trans. Inoculation of the EAEC S258A mutant in the intestinal ileal loops induced an increase in glycoprotein secretion to 11 mg/ml, a value not significantly different from that obtained with the wild-type strain (Fig. 2). These results indicate that the serine protease motif is not related to the mucin secretagogue activity induced by Pic and that complementation in trans (by using a plasmid) does not restore the secretagogue activity of Pic to 100%.

FIG. 2.

Pic-secreting EAEC bacteria induce glycoprotein hypersecretion to the intestinal lumen. Rat ileal loops were inoculated with 1.5 × 10⁸ CFU/ml of bacteria that are unable (HB101 and EAEC Δpic) or able (EAEC, EAEC Δpic/pPic1, EAEC Δpic/pPicS258I, EAEC S258I, and HB101/pPic1) to secrete Pic. After 12 h, intestinal mucus (detected as glycoprotein in mg/ml) was quantified colorimetrically by using the PAS method and analyzed statistically with the Mann-Whitney test. *, P < 0.05 compared with HB101; **, P < 0.05 compared with EAEC 042; ***, P < 0.05 compared with EAEC Δpic. Error bars indicate standard deviations.

Pics from UPEC and S. flexneri also cause intestinal mucus hypersecretion.

To determine whether other bacteria harboring pic were able to induce intestinal mucus secretion, we used the same rat ileal loop model inoculated with 1.5 × 10⁸ CFU/ml per loop of S. flexneri 2a, S. flexneri Δpic, UPEC, and UPEC Δpic (both mutants were constructed by chromosomal insertion). Ileal loops inoculated with S. flexneri 2a or UPEC induced glycoprotein secretion of 18 mg/ml and 12 mg/ml, respectively (Fig. 3); both strains induced a significantly higher secretion than the laboratory HB101 strain (4.84 mg/ml) (P < 0.05). In contrast, their isogenic mutants did not show the increase in mucus secretion detected with the wild-type strains, showing glycoprotein concentrations of only 5 mg/ml and 3.5 mg/ml in ileal loops inoculated with UPEC Δpic and S. flexneri Δpic, respectively (Fig. 3). The mucin secretagogue activities of these mutants were compared with that of an isogenic UPEC hylA mutant to control for interference by genetic manipulation during the construction of the mutants. This mutant caused an increase in glycoprotein secretion into the ileal loops of 12 mg/ml, which was similar to that caused by the wild type (Fig. 3).

FIG. 3.

Pic-secreting UPEC and S. flexneri also cause mucus hypersecretion to the intestinal lumen. Rat ileal loops were inoculated with the indicated bacterial strains (1.5 × 10⁸ CFU/ml per loop). After 12 h, intestinal mucus was collected and the glycoprotein concentration was quantified by the colorimetric PAS method. Four rats were used for each bacterial strain, and data are expressed as averages ± standard deviations (SD). Statistical analyses were performed by using Student's t test. *, P < 0.05 compared with HB101; **, P < 0.05 compared with Wt S. flexneri; ***, P < 0.05 compared with Wt CFT073.

Pic increases intestinal mucus secretion and goblet cell cavitation.

To confirm the intestinal mucus secretion and determine the rapid cavitation (compound exocytosis) of goblet cells induced by Pic, intestinal sections from the ileal loops were stained with Alcian blue/nuclear fast red and analyzed by light microscopy. With this staining, it was possible to observe secreted mucus and mucus inside the goblet cells, which were stained in blue, while intestinal epithelium was stained in pink. A normal aspect of the intestinal tissue was observed when the ileal loops were inoculated with E. coli HB101; intestinal villi had their classical digitiform projections and their lamina propria composed of smooth muscle cells, and leukocytes were scarce. Goblet cells were observed distributed among enterocytes of the villus (Fig. 4A). However, when the ileal loops were inoculated with the wild-type EAEC 042, we observed intestinal villus hypertrophia, with abundant leukocytes infiltrated in the lamina propria, as well as a remarkable increase of goblet cells exuding abundant mucus to the intestinal lumen by pinching (cavitation) of the cells (Fig. 4B). When ileal loops were inoculated with the mutant EAEC 042 Δpic, we also observed villus hypertrophia and infiltrated leukocytes in the lamina propria, but there was only scarce mucus exudation to the intestinal lumen. Additionally, there was no increase in mucus-secreting goblet cells as was previously seen in tissue inoculated with the wild-type strain (Fig. 4C). As previously shown, this mutant does not secrete Pic in vitro but secretes Pet (see Fig. 1), which has enterotoxic and cytotoxic activities (13, 35), suggesting that the morphological alterations in the ileal loops were due to Pet effects. This lack of mucus hypersecretion in the pic mutant was reversed by complementation with the pic gene (EAEC 042 Δpic/pPic) (Fig. 4D). In the same way, ileal sections from intestinal loops inoculated with EAEC S258A, which is mutated in the serine protease motif directly in the chromosome, showed mucus hypersecretion (Fig. 4E) comparable to that observed when the wild-type EAEC 042 was inoculated (Fig. 4B). Again, this result suggests that the Pic serine protease motif is not involved in the secretagogue activity or the increase of mucus-secreting goblet cells induced by Pic.

FIG. 4.

Staining of mucus in intestinal sections from ileal loops inoculated with bacteria harboring the pic gene and their isogenic mutants. Ileal loops were inoculated with the indicated bacterial strains (1.5 × 10⁸ CFU/ml per loop). After 12 h, intestinal segments were removed and embedded in paraffin to obtain sections that were stained with Alcian blue and nuclear fast red. Images correspond to representative sections observed under a light microscope with a 40× objective.

On the other hand, in ileal sections from intestinal loops inoculated with UPEC CFT073 (Fig. 4F), we also observed an increase in mucus secretion to the intestinal lumen, accompanied by an increase in the number of mucus-containing goblet cells in the intestinal villus as well as cells exocytosing mucin granules, in comparison to intestinal sections from ileal loops inoculated with HB101 (Fig. 4A). However, when analyzing rat ileal sections from intestinal loops inoculated with the isogenic pic mutant, UPEC Δpic (Fig. 4G), we did not observe any increase in mucus secretion or in the number of mucus-containing goblet cells, and they were similar to ileal sections from loops inoculated with E. coli HB101 (Fig. 4A). On the other hand, in rat ileal sections from intestinal loops inoculated with S. flexneri 2a, we observed severe intestinal epithelial damage, a huge amount of mucus secreted in the intestinal lumen, goblet cells exocytosing mucin granules, and an increase in the number of mucus-containing goblet cells in the intestinal villi (Fig. 4H). In contrast, in intestinal sections from ileal loops inoculated with the mutant S. flexneri Δpic, villi were observed upright, with a moderate amount of goblet cells (Fig. 4I), similar to the case for the negative control (Fig. 4A).

In addition, intestinal sections inoculated with the mutant with a mutation in an unrelated gene, hlyA, showed no difference compared with sections from ileal loops inoculated with the wild-type UPEC CTF073 (data no shown). Together, the data indicate that Pic is a factor responsible for mucus secretion to the intestinal lumen induced by three pathogens harboring the pic gene, EAEC, UPEC, and S. flexneri.

Purified Pic protein induces mucus hypersecretion by goblet cells.

To avoid the influence of other bacterial components on the mucus hypersecretion induced by Pic, we decided to use purified Pic protein obtained from the supernatant from a culture of HB101/pPic1. Ileal loops inoculated with 100 μg of purified Pic protein showed a glycoprotein concentration of 12.23 mg/ml, which represented a significant 2-fold increase compared to that reached by inoculation with PBS (5.5 mg/ml) (P < 0.05) (Fig. 5A). As expected, when Pic protein was denatured by heating (boiling for 10 min) before its inoculation, mucus hypersecretion was not observed; i.e., the glycoprotein concentration was 6.36 mg/ml, significantly lower than that after inoculation of Pic (P < 0.05) (Fig. 5A).

FIG. 5.

Purified Pic induces mucus hypersecretion to intestinal lumen. (A) Glycoprotein hypersecretion. Ileal loops were inoculated with purified Pic, PicS258I, and Pet proteins, PBS was used as vehicle. Glycoprotein quantification and analyses were performed as for Fig. 4. *, P < 0.05 compared with PBS; **, P < 0.05 compared with Pic. (B to D) Mucus hypersecretion detected in intestinal sections. Intestinal segments were removed and embedded in paraffin to obtain ileal sections that were stained with Alcian blue and nuclear fast red. Images are representative sections observed under a light microscope with a 40× objective. Rat ileal loops were inoculated with PBS (B), 100 μg of purified Pic protein (C), or 100 μg of heat-denatured (Δ) purified Pic protein (D).

To directly determine the role of the active site of Pic, we used the purified Pic mutant with a mutation in the serine protease motif (PicS258I). When PicS258I was inoculated in the ileal loops, we observed intestinal mucus hypersecretion, since the glycoprotein concentration was 11.22 mg/ml, significantly higher than that after inoculation with PBS (P < 0.05) (Fig. 5A). Considering that Pet represents an important virulence factor in EAEC, we decided to inoculate Pet protein into the ileal loops under the same conditions as for Pic. Ileal loops inoculated with Pet showed a glycoprotein concentration of 6. 82 mg/ml, which was similar to that obtained by using PBS (5.5 mg/ml) and significantly lower than that after inoculation of Pic (P < 0.05) (Fig. 5A). These data show and confirm that the Pic autotransporter protein secreted by EAEC is an intestinal mucin secretagogue and that its activity is independent of its serine protease motif.

The effect of inoculating purified proteins into the intestinal loops was also evaluated through histopathological analyses. When ileal loops were inoculated with PBS, the villi were observed upright, keeping their digitiform projections, with scarce presence of leukocytes in the lamina propria (Fig. 5B). When purified Pic protein was inoculated in the ileal loops, we observed no epithelium destruction but slight villus hypertrophia with few infiltrated leukocytes and, particularly, an increase in the mucus-containing goblet cells, which were detected from the Lieberkühn crypts to the distal parts of the villus, as well as mucus exudation to the intestinal lumen (Fig. 5C). As expected, when the ileal loops were inoculated with denatured Pic, mucus hypersecretion was abolished, and the intestinal sections were similar to those of negative controls (Fig. 5D).

Pics secreted by EAEC, UPEC, and S. flexneri increase the number of mucus-containing goblet cells in the intestinal villi.

To quantify the increase in mucus-containing goblet cells, we microscopically counted mucus-producing goblet cells in 10 intestinal villi by using computing software (Imag.Pro Plus) and selecting goblet cells stained in blue. When the ileal loops were inoculated with the EAEC 042 strain, an average of 30 mucus-producing goblet cells per villus was counted (Fig. 6), which was significantly higher than the 19 goblet cells per villus in ileal loops inoculated with the laboratory strain HB101 (P < 0.05). When the ileal loops were inoculated with the mutant 042 Δpic, 23 mucus-producing goblet cells were quantified, an amount significantly smaller than that quantified in ileal loops inoculated with the wild-type EAEC 042 (P < 0.05).

FIG. 6.

Bacteria harboring the pic gene, but not their isogenic mutants, increase the number of mucus-producing goblet cells in intestinal villi. Ileal loops were inoculated with the indicated bacterial strains (1.5 × 10⁸ CFU/ml per loop). After 12 h, intestinal segments were removed and embedded in paraffin to obtain sections that were stained with Alcian blue and nuclear fast red. The graph represents the number of blue goblet cells counted in 10 villi of the intestinal sections, expressed as average ± SD. Statistical analyses were performed by using Student's t test. *, P < 0.05 compared with HB101; **, P < 0.05 compared with EAEC; ***, P < 0.05 compared with Wt UPEC; ****, P < 0.05 compared with Wt S. flexneri.

The other Pic-secreting pathogens (UPEC CFT073 and Shigella flexneri 2a), as well as the strain with a chromosomal mutation in the serine protease motif (EAEC S258A), caused an increase in the number of mucus-producing goblet cells of 1.6-fold (P < 0.05) compared with the E. coli HB101 strain (Fig. 6). However, when S. flexneri Δpic was inoculated, the number of mucus-producing goblet cells per villus decreased to the levels seen with HB101 (Fig. 6), which were significantly lower than those in cells inoculated with the wild-type S. flexneri (P < 0.05). These results were similar to those obtained when the isogenic pic mutant of UPEC CFT073 was inoculated, when a significant decrease of the mucus-producing cells compared with that after inoculation with the wild-type strain UPEC CFT073 was observed (P < 0.05) (Fig. 6). These data are consistent with the thesis that infection with Pic-secreting pathogens produces an increase in mucus-producing goblet cells, mucus cavitation, and mucus hypersecretion, events that are induced by Pic and are independent of its serine protease motif.

Kinetics of mucin secretion induced by Pic-secreting pathogens in ileal loops.

Once the rapid secretion of mucin induced by Pic-secreting bacteria was established, it was interesting to determine how mucin secretion occurs during the infection with these pathogens. For this, we performed a kinetic analysis of Pic incubation in ileal loops by inoculation of bacterial strains harboring pic (1.5 × 10⁸ CFU/ml in each loop). After 1, 2, 4, 6, 8, 10, 12, 14, and 16 h, mucus was collected and glycoproteins colorimetrically quantified by the periodic acid-Schiff stain (PAS) method. Glycoprotein secretion induced by UPEC and EAEC inoculation significantly increased with the incubation time (P < 0.05) in comparison to the kinetics of mucin secretion in ileal loops inoculated with PBS (PBS and HB101 induce similar amount of mucus secretion), which was kept constant during all times. At 4 h postinoculation (Fig. 7), we started observing a significant difference between ileal loops inoculated with pathogens and controls. However, the increment at this time was 2-fold higher in ileal loops inoculated with UPEC than in those inoculated with EAEC. At 6 h, both strains reached the same level of mucin secretion into the intestinal lumen, and both increased at 8 and 10 h; finally, EAEC caused a greater increase in mucus secretion into the ileal loops at 12 and 14 h than UPEC, but both strains reached similar concentrations at 16 h (Fig. 7). Like with UPEC, similar kinetics of mucin secretion occurred when S. flexneri was inoculated in the ileal loops, whereas isogenic pic mutants maintained glycoprotein secretion like that in the ileal loops inoculated with PBS (data not shown).

FIG. 7.

Kinetics of mucus secretion into ileal loops by Pic-secreting bacteria. Ileal loops were inoculated with the indicated bacterial strains (1.5 × 10⁸ CFU/ml per loop). After the times established for the kinetics analysis (1 to 16 h), intestinal mucus was collected and the glycoprotein concentration was quantified by the colorimetric PAS method for each time period. Data are expressed as average ± SD. Statistical analyses were performed by using Student's t test.

DISCUSSION

The pathogenesis of EAEC infection is based on two main histopathological characteristics: (i) formation of a thick layer of mucus on the intestinal mucosa (47) and (ii) damage to the intestinal epithelium (23, 48). One of the virulence factors of this bacterium is Pic, which has mucinolytic activity on bovine submaxillary mucin gland (BSM) and hog gastric mucin (HGM), and its proteolytic activity depends on its serine protease motif (18). In this work, we show that Pic protein is secreted by three important pathogens harboring the pic gene (EAEC, UPEC, and S. flexneri) and once secreted is able to induce mucin hypersecretion to the intestinal lumina of rat ileal loops inoculated with these bacteria. Histopathological studies showed an increase in mucus secretion to the intestinal lumen, the presence of goblet cells exocytosing mucin granules, and an increase in the number of mucus-producing goblet cells in the intestinal villi. These effects induced by Pic do not depend on its serine protease motif. Thus, Pic mucinase is responsible for one of the histopathological features of EAEC pathogenesis: mucus hypersecretion.

When analyzing culture supernatants from pathogenic strains harboring the pic gene (EAEC, UPEC, and S. flexneri), we found the Pic protein of 109 kDa in the supernatants from prototype strains of these pathogens, while in the pic mutants constructed for the three pathogens, Pic secretion was abolished. These data indicate that pic gene detected in the three pathogens (2, 22, 37) is active and that, based on their secretion features, the protein is autosecreted to the extracellular medium and thereby to the intestinal lumen during bacterial infection.

All three pathogens harboring the pic gene, but not the isogenic pic mutants, induced mucus hypersecretion. These data indicate that the Pic protein is an intestinal mucus secretagogue because it stimulates the rapid mucus secretion pathways, which result in a massive fusion of mucin granules with the apical membrane of goblet cells, leading to total cavitation of these cells (45). This mucus hypersecretion can occur in response to a variety of stimuli (32), including intrinsic stimuli such as interleukin-1 (IL-1), IL-4, IL-6, IL-9, prostaglandin E2, acetylcholine, histamine, and neurotensin and external stimuli such as cholera toxin, listeriolysin O from Listeria monocytogenes, elastase from Pseudomonas aeruginosa, lipopolysaccharide (LPS) from Bordetella pertussis, and an unknown factor from Entamoeba histolytica (5, 7, 8, 11, 38). In this work, we describe a new intestinal mucus secretagogue of bacterial origin: Pic from EAEC, UPEC, and S. flexneri.

Rapid mucin secretion (which was also evidenced by the kinetics of glycoprotein secretion induced by Pic) could be quantified by the PAS colorimetric method, which detects neutral mucins (40, 41). Moreover, Pic was also able to induce an increase in acidic mucins (sulfomucins and sialomucins) that was qualitatively detected by Alcian blue in histological samples, as well as the quantitative increase of the number of goblet cells containing sulfo-sialomucins. Using Alcian blue and PAS, the intestinal epithelium showed an increase of a mixture of acidic and neutral mucins induced by Pic protein, as well as by Pic-producing bacterial strains. Interestingly, MUC2 is a prominent intestinal mucin, suggesting that Pic may be induced by MUC2 hypersecretion, since a correlation between Alcian blue-PAS and immunohistochemical expression of MUC2 in Barrett's esophagus has been reported (27).

Rapid mucin secretion represents a defense mechanism against various chemical, physical, and biological agents, providing a barrier between the intestinal lumen and epithelium (45). The defensive nature of mucins lies in their capacity to entrap microbes and their subsequent removal through peristaltic movement (11). Due to these facts, microorganisms have developed proteases to pass through this mucus layer and interact with the intestinal epithelium. One of these mucinases is Pic; our evidence indicates a double functionality on mucus during infection by pathogens harboring Pic. The protein is secreted from the bacteria and recognized by goblet cells to induce mucus hypersecretion. However, Pic also has a mucinolytic activity (18), which favors intestinal colonization, because it could help the bacterium to go across the mucus layer lining the intestinal epithelium (18, 20). Thus, the data from this work describe, for the first time, a bacterial protein with mucinase and mucus secretagogue activities. Availability for degrading mucus confers various advantages to intestinal microorganisms for colonization, as suggested recently for Pic (20), since mucin and its oligosaccharides represent a direct source of carbohydrates and peptides. This source can positively favor colonization by providing a nutrient source for bacterial growth. Thus, mucus-colonizing bacteria, especially EAEC, are able to avoid rapid expulsion via the hydrokinetic properties of the intestine and would have a growth advantage. In fact, the host mucus could be the more suitable bacterial niche (11).

Remarkably, unlike the case for other pathogen secretagogues, the Pic active site (serine protease motif) is not needed for the secretagogue activity, leaving a possibility for the therapeutic use of PicS258I for pathologies where deficient intestinal mucus secretion exists. A precise definition of the regulatory networks that interface with goblet cells by using Pic also may have broad biomedical applications, because mucus alterations appear to characterize most diseases of mucosal tissues. Additionally, the lack of a role for the Pic serine protease might suggest that Pic is not using the PAR₂ pathway for intestinal glandular exocrine secretion, since it is activated by serine proteases (28). Thus, here we demonstrate that the presence of a large amount of mucus in the intestinal mucosa induced by EAEC (24), as well as the mucus present in basilar dysentery induced by shigellosis (36), can be due to Pic induction. Interestingly, we describe for the first time an increase in the mucus-producing goblet cells as a new feature in both infections.

Goblet cells are present in the respiratory, urogenital, and gastrointestinal tracts. There is evidence that these cells originated from differentiation, and not from proliferation, from precursor epithelial cells (6, 31, 43). According to this evidence, it is probable that the increase in the number of mucus-producing goblet cells during infection by Pic-secreting bacteria may be due to a differentiation process or simply an induction of mucus production in preexisting goblet cells. Therefore, more studies must be conducted to clarify this process. However, this is not the first study in which an increase in goblet cells in a pathological process was seen, since infection with the Nippostrongylus brasiliensis parasite also increases goblet cell rates in intestinal villi of rats inoculated with this protozoan (30).

It is known that UPEC first colonizes the intestine and then by anal-urethral cross-contamination infects and colonizes the urinary tract, producing an extraintestinal infection (17). Furthermore, UPEC infection is characterized by the presence of high concentrations of glycoproteins in the urine; among these glycoproteins are GP51, which is the major component of the mucin layer in the bladder, and Tamm-Horsfall protein, also known as uromucoid or uromodulin, which is produced by renal tubular cells of the Henle's loop. There are various types of mucin in the urinary tract, i.e., MUC1, -3, and -4, which are bound to the membranes of uroepithelial cells (26, 44). There are no reports that higher expression of these mucins occurs during UPEC infection, but from our data, we can hypothesize that the Pic protein might be involved in this process.