OUP user menu

Osmoregulated periplasmic glucans are needed for competitive growth and biofilm formation by Salmonella enterica serovar Typhimurium in leafy-green vegetable wash waters and colonization in mice

Liu Liu, Shawn Tan, Won Jun, Allen Smith, Jianghong Meng, Arvind A. Bhagwat
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01462.x 13-20 First published online: 1 March 2009


Osmoregulated periplasmic glucans (OPGs) are major periplasmic constituents of Gram-negative bacteria. The role of OPGs has been postulated in symbiotic as well as pathogenic host–microorganism interactions. Here, we report the role of OPGs from Salmonella enterica serovar Typhimurium during growth and biofilm formation in leafy-green vegetable wash water. The opgGH mutant strain, which was defective in OPG biosynthesis, initiated the growth at a slower rate in wash waters obtained from spinach, lettuce and green collard and severely impaired biofilm formation. The lack of OPG synthesis did not influence biofilm formation by the opgGH mutant in low-nutrient low-osmolarity laboratory media. In coculture experiments initiated with equal proportions of cells, the opgGH mutant was outnumbered by the wild-type strain under the planktonic as well as the biofilm growth conditions. The opgGH mutant strain poorly colonized mouse organs when introduced orally along with the wild-type strain. This is the first report demonstrating the role of OPGs of Salmonella in competitive colonization of biofilms, planktonic cultures and mouse organs.

  • food safety
  • biofilms
  • glucans
  • vegetable wash waters
  • food microbiology
  • low osmotic environments


Salmonella are common agents of gastrointestinal-based diseases in humans and have been recognized as a major foodborne hazard. In humans, the organism is most commonly acquired following ingestion of contaminated food or water (D'Aoust et al., 2001). Traditionally, poultry products have been documented as a major source of contamination in many developed countries (Hald et al., 2004). However, in recent years, Salmonella infections associated with raw vegetables have occurred with increased frequency, particularly involving fresh produce (Horby et al., 2003; CDC, 2005; Singh et al., 2007). While the specific sources of contamination have not been identified, fresh produce is grown in natural habitats for Salmonella reservoirs (such as birds, amphibians, reptiles and poultry). One possible source for foodborne infections is the quality of water, either in the liquid phase used to wash the produce or in the form of ice used in shipping or storage (Bhagwat, 2006). Irrigation water quality is also of significant importance as it may be responsible for for carrying microorganisms from the field to the fork (Wachtel et al., 2002; Steele & Odumeru, 2004; Duffy et al., 2005). In several instances, foodborne illnesses have been traced to poor or unsanitary postharvest practices, specifically to nonpotable cooling water and ice (Harris et al., 2003; Steele & Odumeru, 2004). Salmonella, as a consequence of its lifestyle, endures extended periods of nutrient deprivation in natural aquatic and terrestrial environments while retaining its pathogenic potential (Foster & Spector, 1995; D'Aoust et al., 2001). A number of environmental factors, including nutrient deprivation, osmolarity and availability of oxygen, have been implicated in modulating the virulence of Salmonella, implying an empirical relationship between survival in nature and survival in the host organism (Arricau et al., 1998; Barak et al., 2005).

Relatively little is known about how Salmonella survive in nutrient-deprived, low-osmolarity environments such as irrigation and vegetable wash waters (Barrow et al., 1996; Casani et al., 2005). Growth of cells requires that the cytoplasm contains essential constituents with a total osmolarity of about 300 mosmol L−1 (Kennedy, 1996). When cells are grown in a low-osmolarity medium, swelling and rupturing of the cytoplasmic membrane is prevented by the osmolarity of the periplasmic space, which is mainly due to anionic short glucose chains, referred to in the literature as membrane-derived oligosaccharides (MDOs) or osmoregulated periplasmic glucans (OPGs) as per the new nomenclature (Miller et al., 1986; Kennedy, 1996; Bohin, 2000; Bohin & Lacroix, 2007). When cells of Escherichia coli are grown in a low-osmolarity medium (c. 30 mosmol L−1), OPGs may represent as much as 5–7% (dry weight) of the cells, and may constitute a considerable fraction of the fixed anions in the periplasm (Miller et al., 1986). OPGs are thought to play an important, but poorly understood, role in host–microorganism interactions involving specific plant and animal hosts (Bhagwat et al., 1996; Page et al., 2001; Arellano-Reynoso et al., 2005). In E. coli, the roles of OPGs in cell-to-cell signaling, chemotaxis, regulation of polysaccharide synthesis and resistance to sodium dodecyl sulfate (SDS) have been postulated (Ebel et al., 1997; Bohin, 2000; Rajagopal et al., 2003).

We recently observed that the opgGH (previously referred to as mdoGH) mutant of Salmonella enterica serovar Typhimurium was compromised in mouse virulence and required a 2-log higher oral dose to achieve a lethal dose 50% (LD50) comparable to that of the wild-type strain (Bhagwat et al., 2009). We wanted to investigate the role of OPGs in the growth and competence of Salmonella in environments critical for foodborne outbreaks such as wash waters obtained from leafy-green vegetables. Because bacteria often adhere to surfaces and form biofilm communities in their natural settings (Donlan & Costerton, 2002; Burmolle et al., 2006), we investigated the biofilm-forming ability of the opgGH mutant strain under low-nutrient low-osmolarity conditions. Lastly, we checked whether the inability to synthesize OPGs would compromise the strain's competitive colonization potential in mouse tissue as well as under biofilm and planktonic conditions.

Materials and methods

Bacterial strains and culture conditions

Salmonella enterica serovar Typhimurium wild-type strain SL1344 (Nal-r) and the opgGH null mutant SG111 (Km-r) (Bhagwat et al., 2009) were grown in Luria–Bertani (LB) medium at 37 °C in a shaking incubator at 220 r.p.m. When required, the medium was supplemented with kanamycin (50 μg mL−1), nalidixic acid (10 μg mL−1) or streptomycin (50 μg mL−1). Salmonella–Shigella agar and brilliant green agar (Difco, Franklin Lakes, NJ), Salmonella semi-selective indicator media with appropriate antibiotics, were used to isolate Salmonella from mouse tissue. The osmolarity of growth media (mosmol L−1) was measured using a Wescor vapor pressure osmometer (model 5500, Wescor Inc., Logan, UT).

The growth rates of the wild type and opgGH mutant were determined in different growth media such as LB broth and low-nutrient no-salt (LNNS) media (which are 1 : 20 diluted LB broth without NaCl) having osmolarity values of 407 ± 4 and 31 ± 3 mosmol L−1, respectively (+/− denote SD of mean). Growth was measured using a Bioscreen C automatic turbidometric analyzer (GrowthCurves USA, NJ). Starter cultures were prepared by inoculating a single colony of the appropriate strain into LB broth, followed by overnight incubation at 37 °C. This culture was diluted 1 : 10 000 into fresh media of varying osmolarities, and 250 μL per well was transferred into a 100-well honeycomb Bioscreen plate. Growth was analyzed at 37 °C with continuous shaking, and for each sample, data were collected from five replicate wells. In the initial experiments, growth was also measured by performing viable cell counts on LB agar media to ensure that the OD reflects viable cell numbers appropriately (Mellefont et al., 2005). To assess the effect of osmotic stress on growth, media were supplemented with varying amounts of salt (NaCl or KCl) or buffered with HEPES (50 mM, pH 7.1).

Biofilm formation

The overnight-grown cultures were diluted 1 : 10 000 in fresh media or vegetable wash waters, placed in sterile polystyrene microplates at 100 μL per well and incubated for 24 h static at 30 °C for use in biofilm studies (O'Toole & Kolter, 1998). Eight wells were inoculated for each sample. After a 24-h incubation, microplate cultures were aspirated and washed five times with 450 μL sterile distilled water with an Elx50 plate washer (BioTek Instruments Inc., Winooski, VT). The plates were air dried, and 150 μL of Protocol crystal violet solution (0.41% w/v dye, 12.0% ethanol and 0.1% phenol in water; Fisher Scientific Company, LLC, Kalamazoo, MI) was added per well and incubated at room temperature for 45 min. The wells were then aspirated and washed five times with 450 μL sterile distilled water. After allowing the plates to air dry, 200 μL of 95% ethanol was added to each well. A multichannel pipettor was used to mix the contents of the wells and to dissolve the crystal violet dye. A600 nm was then recorded for each well using a microquant microplate spectrophotometer (BioTek Instruments Inc.). The average absorbance of eight control wells (that had contained culture medium only) was subtracted from each sample well to determine the amount of biofilm present.

In order to determine viable counts, biofilms from individual wells (before crystal violet staining) were suspended in 100 μL of saline and 100 mg glass powder (Sigma Chemical Co., St. Louis, MO). The suspended biofilms were recovered in three washes with 100 μL saline. The suspension was vortexed vigorously for 1 min, and 10-fold serial dilutions were plated on selective media to determine viable cell counts. The detection limit was 103 cells per well.

Preparation of vegetable wash water

The vegetable rinse water was prepared as described previously (Bhagwat, 2004). Briefly, fresh spinach (Spinacia oleracea), lettuce (Lactuca sativa) and collard green (Brassica oleracea) vegetables were obtained from local grocery stores. The produce (250 g) was sliced to c. 5 cm × 5 cm and washed for 30 min by gentle shaking in a plastic container (33 cm × 22 cm × 8 cm) containing 500 mL of deionized distilled water. The vegetable rinse solution was decanted from the tray and was filtered through a glass wool column placed in a 50-mL syringe. The filtered vegetable rinse-water was centrifuged for 10 min at 4000 g at room temperature. The supernatant was made bacteria-free by filtering through a 0.22-μm nylon filter and was used in biofilm experiments. The pH and osmolarity values were measured as described.

Mouse virulence studies

Five-week-old male BALB/c mice were purchased from the Small Animals Division of the National Cancer Institute (Frederick, MD). Mice were housed in an AllenTown Caging Biocontainment-isolator rack, four to five per cage, and provided with Harland-Teklad rodent chow and deionized water ad libitum. Mice were acclimated 1 week before use, and all animal protocols were approved by the Institutional Animal Care and Use Committee. Animals were fasted for c. 12 h before being inoculated with 0.2 mL of an S. enterica serovar Typhimurium suspension (in 0.9% NaCl) by an oral gavage. Bacterial strains were grown in LB medium at 37 °C without shaking for 16–18 h, suspended in saline and adjusted to the appropriate cell density before oral infection. Viable cell counts were confirmed by retrospective spread plating onto LB agar plates and incubating the plates overnight at 37 °C.

To analyze colonization of individual organs by each bacterial strain, mice were sacrificed 6 days postinfection. Individual organs (liver, spleen and intestine) were dissected, weighed and homogenized in LB medium. Cell counts were determined by spread plating appropriate dilutions onto Brilliant green agar plates (Difco) containing streptomycin (50 μg mL−1) or kanamycin (25 μg mL−1). Individual colonies were counted after an overnight incubation at 37 °C, and statistical analysis was performed using anova with post hoc analysis for multiple comparisons or the Mann–Whitney nonparametric test. A value of P<0.05 was considered significant.

Statistical analysis

For statistical analyses, sigmastat 3.0 software (Ashburn, VA) was used. Data were analyzed by the one-way anova test to determine statistical differences between the means of treatments.


OPGs are needed to achieve optimal growth rates in low-nutrient and low-osmolarity media

We examined the contributions of OPGs in growth and biofilm formation by S. enterica serovar Typhimurium strains in LNNS media (Fig. 1a). Upon 1 : 10 000-fold dilution of stationary-phase culture to a fresh LNNS medium, wild-type cells had a ‘relative OD lag time’ of 400 ± 22 min as compared with 585 ± 31 min required by the opgGH mutant (Fig. 1a, filled symbols). The delay in initiating growth was a phenomenon specific to low osmolarity of the medium, as increasing the osmolarity from 31 to 240 mosmol L−1 by addition of NaCl (0.155 M final concentration) restored normal growth in the opgGH mutant strain (Fig. 1a, open symbols).

Figure 1

Growth of Salmonella enterica serovar Typhimurium wild-type and opgGH mutant strains in (a) LNNS medium, (b) wash waters obtained from spinach, (c) green collard and (d) lettuce. Growth was measured every 15 min by measuring A600 nm using a BioScreen growth chamber either at low osmolarity (•, ▪; 25–31 mosmol L−1) or after addition of NaCl to 0.155 M (final concentration) (○, □; 240–255 mosmol L−1). Wild type (•, ○) and opgGH mutant strain (▪, □).

We then compared the growth potential of wild-type and opgGH mutant strains in vegetable wash waters obtained from leafy-green vegetables such as spinach, collard green and lettuce (Fig. 1b–d). Similar to what was observed with LNNS growth medium, lack of OPG synthesis in the opgGH mutant severely affected its ability to initiate growth in vegetable wash waters that had osmolarity values of 25–29 mosmol L−1 (Table 1). The observed delay to initiate growth by the opgGH mutant does not appear to be related to its ability to utilize certain nutrients in vegetable wash waters. By the time strains reached the stationary growth phase, all three leafy-green wash waters supported c. 11 doublings of the wild type and opgGH mutant, and the cell densities measured as viable cell counts after 24 h of growth were not significantly different (P>0.05). Moreover, adjusting the osmolarity of vegetable wash waters by adding NaCl (0.155 M final concentration) to 240 mosmol L−1 restored the normal growth of the opgGH mutant in all three wash waters (Fig. 1b–d, open symbols).

View this table:
Table 1

Ability of various wash waters to support the growth of Salmonella with and without addition of external osmoticum

Growth mediumAddition of external osmoticum (NaCl, 0.15 M final concentration)pHOsmolarity (mosmol L−1)Viable count after 24 h of growth in pure culture (CFU mL−1)
Wild typeopgGH
LNNS6.8 ± 0.131 ± 35.48 ± .03 × 1085.85 ± 1.1 × 108
LNNS+6.8 ± 0.1240 ± 115.28 ± 0.6 × 1085.3 ± 0.6 × 108
Spinach wash water6.0 ± 0.129 ± 45.85 ± 0.6 × 1086.6 ± 0.7 × 108
Spinach wash water+6.0 ± 0.1244 ± 65.48 ± .03 × 1087.22 ± 1.1 × 108
Green collard wash water6.6 ± 0.128 ± 59.9 ± 2.0 × 1085.3 ± 3.0 × 108
Green collard wash water+6.6 ± 0.1255 ± 71.3 ± 0.5 × 1081.1 ± 0.27 × 108
Lettuce wash water5.7 ± 0.126 ± 61.9 ± .07 × 1080.9 ± .04 × 108
Lettuce wash water+5.7 ± 0.1250 ± 41.5 ± 0.1 × 1081.4 ± .2 × 108
  • * Osmolarity was adjusted by addition of 5 M NaCl to 0.15 M final concentration.

  • Viable cell counts measured immediately after inoculation were 7.3 ± 1.91 × 104 (wild type) and 7.29 ± 1.98 × 104 (opgGH mutant).

OPGs and biofilm formation

Attachment of Salmonella to food-processing surfaces and subsequent development of biofilms may have significant economic and public health consequences (Donlan & Costerton, 2002; Burmolle et al., 2006; Agle, 2007). It is suggested that Salmonella biofims adapt structurally to changes in the medium osmolarity and nutrients (Lapidot et al., 2006; Mangalappalli-Illathu et al., 2008). We examined biofilm formation by wild-type and opgGH mutant strains in LNNS medium (Fig. 2). In general, the opgGH mutant strain appeared to form less biofilm than the wild type in LNNS medium, but the differences were not statistically significant (P>0.05). However, the opgGH mutant formed significantly reduced or no biofilms in leafy-green wash waters (Fig. 2). Biofilm formation by the opgGH strain in wash waters obtained from green collard and lettuce was below the detection limit (A600 nm after crystal violet staining <0.05). In spite of the fact that the opgGH mutant was able to support a number of cell divisions and was able to grow in vegetable wash waters, it formed significantly lower quantities of biofilm in spinach wash waters. In general, wild-type Salmonella cells formed reduced quantities of biofilm in vegetable wash waters in comparison with LNNS medium.

Figure 2

Biofilm formation by Salmonella enterica serovar Typhimurium wild-type and opgGH mutant strains. Growth was measured by determining A600 nm. Biofilm was measured after staining with crystal violet. The ratio of A600 nm after crystal violet staining to initial growth turbidity (A600 nm) is plotted for the wild-type (black bars) and the opgGH strain (gray bars) after incubation in LNNS and vegetable wash waters. *A600 nm of crystal violet <0.001.

OPGs and competitive growth in free-living and biofilm states

We examined how the initial delay in resuming growth of the mdoGH mutant might affect the strains' ability to compete in biofilm and free-living settings when coinoculated in equal proportions with the wild-type strain (Fig. 3). Biofilms were initiated by inoculating equal numbers of wild-type and opgGH mutant cells in LNNS medium of varying osmolarities. After 24 h of static growth in the medium of osmolarity <40 mosmol L−1, the opgGH mutant was outnumbered by the wild-type strain by c. 10-fold in the planktonic as well as in the biofilm state [competitive index (CI)≤0.1]. Upon addition of NaCl to increase the medium osmolarity, although the total biofilm formation was negatively affected (data not shown), the competitiveness of the opgGH mutant was restored, and CI values closer to 1.0 were obtained at osmolarity >100 mosmol L−1 (Figs 3 and 4, gray bars). Likewise, in vegetable wash waters obtained from spinach and green collard (Fig. 4), the CI for the opgGH mutant was <0.01 in biofilms, indicating that the OPG-lacking cells were outnumbered by the wild-type strain in biofilms. Viable count data from biofilms obtained from vegetable wash waters indicated that the biofilms were mainly composed of the wild-type strain (<1.0% of the total viable cells carried the opgGH mutation). It may be noted that in pure cultures, very low or no biofilm formation by the opgGH mutant strain was observed in vegetable wash waters (Fig. 2) (<103 cells mL−1 were recovered from individual wells, data not shown).

Figure 3

Competitive growth of Salmonella enterica serovar Typhimurium wild-type and opgGH mutant strain in planktonic and biofilm states in LNNS medium at varying osmolarities. Wild-type and opgGH mutant strains were inoculated in equal proportions in LNNS media with varying osmolarities. Viable cell counts were determined from biofilms (○) and planktonic cells (•) on selective media containing antibiotics. CI is the ratio of viable cell counts for the opgGH mutant to that of the wild-type cells at the time of harvest.

Figure 4

CI during biofilm formation by Salmonella enterica serovar Typhimurium wild-type and opgGH mutant strains in vegetable wash waters. Wild-type and opgGH mutant strains were mixed in equal proportions and inoculated in LNNS or vegetable wash waters with (gray bars) or without (black bars) addition of NaCl to 0.15 M (final concentration). Viable cell counts were determined from biofilms by plating on selective media containing antibiotics. CI is the ratio of viable cells of the opgGH mutant to that of the wild type recovered 24 h postinoculation from biofilms. *Viable cell counts of the opgGH mutant strain were below the detection limit (<102 cells mL−1).

Role of OPGs in mouse organ colonization

Recently, we observed that lack of OPG synthesis rendered Salmonella less virulent, and c. 2-log more cells were needed to achieve an LD50 comparable to that of wild-type cells (Bhagwat et al., 2009). We inoculated mice with equal proportions of wild-type and opgGH mutant strains and examined their colonization in the intestine, spleen and liver (Fig. 5). Although there was variation among individual mice, the average CI for the intestine was 0.0001, while the values for the spleen and the liver were 0.08 and 0.09, respectively. The data indicated that the Salmonella strain lacking OPGs was unable to colonize the intestinal track in competition with the wild-type strain. On the other hand, during the later stages of infection involving spleen and liver colonization, the opgGH mutant was outnumbered by the wild-type strain only by c. 10-fold.

Figure 5

Competitive colonization of mouse organs after coinoculation with Salmonella enterica serovar Typhimurium wild-type and opgGH mutant strain 6 days postinfection. CI is the ratio between the opgGH mutant strain and the wild type in the individual organs divided by the ratio of the two strains in the inoculum. The distribution of CI from individual animals is shown in a box and whisker format. Boxes range from the 25th to the 75th percentile and are intersected by the median line. Whiskers extending below and above the box range from the 10th to the 90th percentile, respectively.


Water reuse is increasingly regarded as a necessity. Currently, reconditioned water is used for initial washing of vegetables, fluming of unprepared products and scalding water of meat and poultry (Rajkowski et al., 1996; Rajkowski & Baldwin, 2003; Casani et al., 2005). Agricultural and farm wash waters are likely to have low-nutrient low-osmolarity conditions. However, little is known about the microbial factors that are critical to growth, survival and biofilm formation by food-borne pathogens such as Salmonella under such conditions (Rychlik & Barrow, 2005). We have demonstrated here that the ability to synthesize OPGs is beneficial for maintaining the optimal growth potential. The longer time needed by Salmonella under low-osmotic conditions could be due to the instability of the membrane structures created due to the lack of OPGs. As the name indicates, OPGs are osmoregulated and their synthesis and accumulation is inversely proportional to the osmotic strength of the environment (Bohin, 2000). Thus, after increasing medium osmolarity from 31 to 240 mosmol L−1, bacterial growth was relatively independent of OPG synthesis (Fig. 1 and Table 1). Salmonella strain lacking OPGs was a poor competitor in various mouse organs as well as in low-osmolarity low-nutrient media in planktonic and biofilm states (Figs 35).

OPGs were suggested to play a critical role in conferring resistance to SDS in E. coli (Rajagopal et al., 2003). It was postulated that anionic detergents such as SDS would be repelled by the OPGs. In E. coli, lack of OPGs also resulted in increased colonic acid production and reduced motility (Kennedy, 1996; Ebel et al., 1997). It was further demonstrated that secondary mutations in the rcs gene family (namely, rcsB, rcsD and rcsF) restored motility in opgGH mutant strains of E. coli (Giris et al., 2007). The authors suggested that the rcs pathway could be hyperactivated in opgGH mutants because rcsB and rcsD act as repressors of flagella synthesis (Majdalani & Gottesman, 2005). Based on what is observed in E. coli, the Rcs pathway might be overactivated under low-osmolarity conditions in the opgGH mutant strain of Salmonella. Hyperactivation of the Rcs pathway in Salmonella has been shown to attenuate its virulence response (Mouslim et al., 2004; Garcia-Calderon et al., 2005). Thus, OPGs may be required to maintain the low level of stress response regulators such as Rcs B, RcsD and RcsF during mouse colonization.

Comparatively little is known about how Salmonella strains grow under low-nutrient, low-osmolarity conditions. In-depth growth analysis of enteric human pathogens under such challenging conditions may provide insights into strategies to reduce human health risk associated with farm waters. We have presented evidence to indicate that the ability to synthesize OPGs is of a significant advantage to bacteria in order to grow, form biofilms and compete in environments such as leafy-green wash waters as well as mouse colonization. It is possible that the presence of OPGs may confer structural stability, which is required for efficient nutrient uptake. Further studies are required to determine whether the opgGH mutant is impaired in nutrient uptake, especially under low-osmolarity conditions.


This study was supported in part by the China Scholarship Council, Ministry of Education, China (L.L.) and the overseas Industrial Attachment Program of the School of Life Sciences and Chemical Technology, Ngee Ann Polytechnic, Singapore (S.T.).


  • Editor: Stephen Smith


View Abstract