Factors Influencing the Ability of Listeria monocytogenes To Pass through a Membrane Filter by Active Infiltration

Abstract

Listeria monocytogenes infiltrated the reticulate structure of a membrane filter and passed through a filter with pore sizes of 0.45 μm and 0.2 μm in 6 to 24 h and 5 to 6 days, respectively. Flagellar motility and expansive pressure generated by the growing bacterial population were indicated as the driving forces of infiltration.

Listeria monocytogenes is a serious threat to humans because of food-borne infections (2, 11). The organism causes generalized infections by exhibiting a remarkable ability to disseminate to the brain, spleen, liver, and other organs (4). This invasive virulence of the organism has been investigated by focusing on the ability of L. monocytogenes to invade and multiply within host cells such as macrophages, hepatocytes, and endothelial cells (3-5). Recently it has been shown that several bacterial species (e.g., L. monocytogenes) have the distinct ability to infiltrate into the reticulate structures of membrane substrates (thickness, 150 μm), without artificial pressure, allowing the bacteria to pass through a membrane filter placed on agar media (7). In the present study, we examined the smallest pore size through which L. monocytogenes could pass, the time required for the pass-through, and other influencing factors. Then, the bacterial traits contributing to the ability to pass through membrane filters were analyzed by addressing flagellar motility and flagellum-independent infiltration forces.

Bacterial strains and examination of pass-through activity.

L. monocytogenes strain EGD was described previously (10). Strain EGDe (DH-L478), its isogenic mutants DH-L975 (Tn917 insertion in flaA [flagellin-encoding gene]), DH-L1042 (flaA in-frame deletion) and the vector pCON1 (6) were supplied by D. E. Higgins, Department of Microbiology and Molecular Genetics, Harvard Medical School. DH-L1042/pCON1FlaA is a Fla⁺ revertant obtained by electroporation of DH-L1042 with pCON1 carrying flaA DNA (PCR product by primers 5′-GGGGTACCCCCGCACAAGTAAGTAAGCCG-3′ and 5′-CGGGATCCCGTAACATTGGCTCTGTGCCCC-3′). Leifson's stain (8) was carried out to confirm flagellation of the revertant. L. monocytogenes clinical isolates CL101, CL102, CL142, and CL184, serotype reference strains F4, 1384, and 1684, and Listeria innocua 93/65 were laboratory stocks. Luria-Bertani (LB) broth and 1.5% agar medium (9) were used routinely for the growth of the above strains. Brain heart infusion (BHI) broth and agar medium (Eiken, Tokyo, Japan) and blood (sheep) agar medium (Eiken) were used in some experiments.

For examination of membrane filter pass-through activity of bacteria, logarithmic phase bacteria grown in LB broth were centrifuged, washed with sterile saline (0.15 M NaCl) and then suspended in saline (approximately 5 × 10⁸ CFU/ml). Ten microliters of the suspension was placed on the center of the autoclaved membrane filter, which was made of mixed esters of cellulose (diameter, 25 mm) and placed on LB agar medium, unless otherwise mentioned. MF-Millipore membrane filters (pore sizes, 0.45, 0.3, and 0.22 μm; thickness, 150 μm; Bedford) and ADVANTEC filters (pore size, 0.2 μm; thickness, 133 μm; and pore size, 0.1 μm; thickness, 110 μm; Tokyo) were used. After absorbance of the 10-μl droplet (horizontal diameter, <6 mm) into the filter, the plate was inverted and incubated at 30°C or 37°C. At a 6- or 12-h time interval, the filters were removed, and the remaining agar plate was incubated for 24 h and then examined for bacterial growth around the area where the original 10-μl droplet had been placed (Fig. 1). Then, the growing bacteria on the agar medium were examined for their Listeria-specific characteristics using standard techniques (1). The filter was determined to be intact by using the bubble point test as described previously (7). Pass-through times were recorded in two ways, that is, as the shortest time recorded in the four test filters (in two independent experiments with two filters) and as the absolute time needed for pass-through of bacteria in all of the nine test filters (in three independent experiments with three filters). Filters with the same lot number were used in the repeated experiments.

FIG. 1.

Membrane filter pass-through activity of L. monocytogenes. (A) A membrane filter (pore size, 0.45 μm) placed on LB agar medium was point-inoculated with strain EGD and incubated at 30°C for 24 h. (B) Bacterial growth on the agar medium after removal of the membrane filter. Arrows indicate the growing bacterial population on the filter and the agar medium under the filter. Ring-shaped characteristic bacterial growth corresponding to the circular margin of the inoculated area on the filter is visible on the agar surface.

The influence of filter pore size and incubation temperature on the ability of Listeria strains to pass through a filter.

Eight L. monocytogenes strains with swimming activity at 30°C and not at 37°C and one L. innocua strain with swimming activity at 30°C and 37°C were examined for the shortest time required to pass through a membrane filter with a pore size ranging from 0.2 to 0.45 μm on LB medium at 30°C and 37°C (Table 1). All strains were able to pass through a membrane filter with a pore size of 0.2 μm. Generally, there were no remarkable differences among the examined strains. More time was required for the strains to pass through a membrane filter with a smaller pore sizes (e.g., with the EGD strain at 30°C, it took 18 h with a pore size of 0.45 μm and 120 h with a pore size of 0.2 μm). Pass-through of a membrane filter with a pore size of 0.1 μm was examined with L. monocytogenes EGD. No pass-through was seen, even after a 21-day incubation at 30°C (data not shown). It was noteworthy that the pass-through time for a 0.45-μm-pore-size filter was shorter at 30°C than at 37°C (Table 1); however, this was not the case with 0.22- or 0.2-μm-pore-size filters. No differences in growth rates of EGD between 30°C- and 37°C-LB broth cultures were recognized (data not shown). Since L. monocytogenes is known to be active even at low temperatures, the filter pass-through activity of an EGD strain at 15°C was examined. The shortest time required to pass through at 15°C was 48 h with a 0.45-μm-pore-size filter, 168 h with a 0.3-μm filter, and 456 h with a 0.22-μm filter.

TABLE 1.

Membrane filter pass-through activity of Listeria spp. at 30°C and 37°C

Strain	Timea for pass-through of filter with pore size (μm):
	0.45		0.3		0.22		0.2b
	30°C	37°C	30°C	37°C	30°C	37°C	30°C	37°C
L. monocytogenes
EGD	18	48	54	96	132	120	120	120
CL101	30	42	48	66	132	120	132	108
CL102	24	30	48	60	120	108	120	96
CL142	30	54	114	108	144	144	144	132
CL184	24	48	72	72	132	108	132	108
F4	42	66	72	72	144	132	144	132
1384	30	48	72	90	120	108	120	108
1684	24	48	72	96	132	132	132	120
L. innocua
93/65	18	24	30	54	96	96	120	96

^aShortest pass-through time (h) in four filters examined.

^bThe 0.2-μm-pore-size filter is an ADVANTEC filter; others are MF-Millipore filters.

Micrograph of L. monocytogenes infiltrating into the reticulate structure of a membrane filter.

For scanning electron microscopy, membrane specimens were fixed and vacuum dried as described previously (7). Critical point drying was not performed because of the incompatibility of the used filter to ethanol and isoamyl acetate. The membrane filter was mechanically torn to see the inside structures. An S-5000N microscope (Hitachi) was used. Infiltration of L. monocytogenes EGD into the reticulate structure of the membrane filter with a 0.2-μm pore size is evident in Fig. 2A, 2B, and 2C. No remarkable changes in the size of bacterial cells during pass-through were recognized. It is noteworthy that void spaces in the polymer nets of the membrane substrate are larger than the diameter of the rod-shaped cell body (Fig. 2), especially in the 0.45-μm-pore-size filter (Fig. 2D). In other words, the pore size described by the manufacturers is not a morphologically determined size, rather an estimated value from the pressure data in mercury porosimetry and calculation based on Washburn's equation (12).

FIG. 2.

Scanning electron micrographs of membrane filters (A through C, pore size, 0.2 μm; D, pore size, 0.45 μm). (A) The upper membrane surface inoculated with L. monocytogenes EGD. (B) Inside of the torn membrane with infiltrating EGD bacterial cells. (C) The exit side of the membrane filter permitting pass-through of EGD bacterial cells. (D) Surface of uninoculated membrane filter. Arrows indicate bacteria in the process of passing through. Each bar, 1 μm.

Effect of medium and inoculum conditions.

The time required for L. monocytogenes to pass through a filter was shorter when the filter was placed on a blood agar medium. An EGD strain grown in BHI broth was point inoculated on 0.45-μm-pore-size membrane filters on three kinds of agar media and incubated at 30°C. Absolute pass-through times with all nine filters examined in each experiment were 6 h on blood agar medium, 18 h on BHI agar medium, and 24 h on LB agar medium. Differences in growth phase (log or stationary) of the inoculated bacteria gave no remarkable differences in the pass-through times of L. monocytogenes (data not shown). However, inoculum prepared with the bacterial culture itself without centrifugation and suspension into saline gave faster pass-through times with a 0.45-μm filter at 30°C (12 h absolute pass-through time by inoculation of bacteria in culture broth, 24 h by bacteria suspended in saline). It was noted by microscopic examination that the motility of saline-suspended L. monocytogenes was remarkably reduced (data not shown).

Bacterial driving forces in pass-through of the membrane filter.

L. monocytogenes is known to be motile and flagellated at 30°C but not at 37°C. Since the slower filter pass-through of L. monocytogenes at 37°C (Table 1) seemed to reflect the downregulated flagellation and motility of L. monocytogenes at 37°C, we compared the ability of L. monocytogenes EGDe strains with and without flagella to pass through a membrane filter. The results indicated the distinct role of the flagella in pass-through and the permissive pore size for effective flagellar movement. As shown in Table 2, flaA gene mutants showed slower absolute pass-through times with a 0.45-μm filter and a revertant obtaining the plasmid-borne flaA gene passed through at the same rate as wild-type EGDe. However, as shown in Table 1, which presents the temperature-independent pass-through for 0.22- and 0.2-μm pore sizes, the flagellum-dependent faster pass-through of flagellated wild-type and revertant strains was not seen with a 0.22-μm filter (Table 2). A pore size of 0.45 μm in the membrane filter seems to give enough space for flagellar rotation as seen in Fig. 2D, but the pore size may be too small for flagellar rotation in a 0.2-μm filter (Fig. 2C). For slow pass-through of the membrane filter with smaller pore sizes, expansion of the cellular population mass by bacterial multiplication seems to be the main driving force. Similarly, slow pass-through behavior of L. monocytogenes in an agar gel (reticulate structure of agarose/agaropectin polymer net) has been observed previously (10) as three-dimensional fractal growth of the bacterial population infiltrating into hard agar medium in seven weeks. The infiltration ability of microorganisms into reticulate structures seems to be an important factor for invasion into animal and plant tissues and in contamination of food materials.

TABLE 2.

Role of L. monocytogenes flagella in the pass-through of membrane filters

L. monocytogenes straina	Timeb for pass-through of filter with pore size (μm):
	0.45	0.22
DH-L478 EGDe wild type	24	144
DH-L975 EGDe flaA:Tn	84	168
DH-L1042 EGDe ΔflaA	84	168
DH-L1042/pCON1	84	168
DH-L1042/pCON1FlaA	24	156

^aA flagellin-encoding flaA gene was mutated by transposon insertion (DH-L975) or in-frame deletion (DH-L1042). pCON1FlaA is a vector that carries the flaA gene.

^bAbsolute time (h) for pass-through of all of the nine tested filters incubated at 30°C.