Three Salmonella serovars were used in this research. Salmonella Enteritidis ATCC BAA-1045 isolated from raw almonds, S. Agona LJH 517 from alfalfa sprouts, and S. Newport ATCC 6962 from a clinical isolate. Each strain was preserved in CryoCare Bacteria Preservers system (Scientific, Stamford, TX). The day before the experiment one bead for each serovars was removed from the ultra low freezer and transferred in 10 ml of freshly made Tryptic Soy Broth (TSB, BD Difco, Sparks, MD) and incubated for 24-h at 37 °C with 70 rpm shaking. The overnight cultures were streaked on Tryptic Soy Agar (TSA, BD Difco, Sparks, MD) to confirm purity. A single colony from each TSA plate was incubate for 24-h at the conditions described aboved. Equal amount from each bacteria suspension (1 ml) were combined to obtained a final bacterial cocktail population of 10^9 CFU/ml [17]. Each serovar solution was also enumerated (10^7 CFU/ml).

Biofilms were grown in a Center for Disease Control and Prevention (CDC) biofilm reactor (Biosurface Technologies, Bozeman, Montana) for up to 96-h using a protocol developed in our lab [17]. Briefly, during the batch phase, 1 ml of of Salmonella cocktail was added to TSB at a concentration of 3 g/l to obtain a final concentration in the reactor of 10^3 CFU/ml of bacteria. After 24 h, the continue phase started and fresh media (at a concentration of 1.5 g/l) was pumped at a rate of 7 ml/min for 72-h. The reactor was maintained at 22 ± 2 °C and the stir plate set at 120 rpm.

The following materials were selected to represent picking bags and harvest bins commonly used in the tree-fruit industry: nylon, wood, and high-density polyethylene (HDPE). All coupons had 1.27 cm diameter on both sides for biofilm growth. Nylon material was cut from used apple harvest picking bags. Sheets of basswood were purchased from a local store and cut into 1.27 cm diameter circles with a Glowforge Pro laser cutter (Glowforage, WA). HDPE coupons were purchased from Biosurface Technologies Corporation (Bozeman, MT). All coupons were scrubbed with soapy water, sonicated for five-minute, and rinsed with tap water. This process was repeated a total of three times. Clean coupons were placed under UV light for 30 min before transferred in the biofilm reactor. The fully assembled reactor was then autoclaved at 121 °C for 15-min.

Salmonella biofilms were grown as described above and at selected time intervals (2, 24 and 96 h), a coupon for each material was removed for biofilm attachment analysis. Inoculated but not rinsed coupons were considered positive controls, while un-inoculated coupons placed in TSB for 96-h served as negative controls. Coupons were gently rinsed by placing them in 50 ml of sterilized deionized water for 10 s [18]. Next, coupons were transferred in a new beaker for 10-s, followed by a third rinse to remove loosely attached cells. Afterwards, coupons were allowed to dry for 30-min prior to enumeration and then placed in 10 ml of Phosphate-Buffered Saline solutions (PBS, VWR, Solon, OH). Biofilms were detached from coupons following the ASTM method E2871-19 [19].Serial dilutions were performed in 0.1% peptone water (Bacto, Sparks, MD) and spread plated in duplicate on Tryptic Soy Agar (TSA, BD Difco, Sparks, MD). Colonies were counted and reported as log CFU/cm2 after 18–24 h of incubation at 37 °C. Two compouns for each conditions and time interval were used during each trial. Experiments were performed in triplicate.

Hydrophobicity experiments were conducted using an optical drop tensiometer (Attention Theta Flex C311, Nanoscience Instruments, Phoenix, Arizona) using Salmonella enterica serovar Typhimurium LT2 biofilms. This strain was used as a BSL 1 surrogate for S. Enteritidis ATCC BAA-1045, S. Agona LJH 517, and S. Newport ATCC 6962 since experiments were conducted in a BSL 1 laboratory. Clean not-inoculated coupons and clean inoculated coupons soaked in TSB for 96-h were used as controls for comparison purposes. The hydrophobicity of Salmonella LT2 biofilms grown on nylon, wood, and HDPE were measured at selected time intervals (2, 24, and 96-h) using sessile-drop technique [18].. All coupons were allowed to dry for 1-h prior to assay. Aliquots (10 μL) of reference liquids, deionized water and formamide, were placed on the surfaces and the contact angles were automatically captured by optical drop tensiometer, calibrated by using a steel sphere with known dimension (35 mm), over a 30-s period. Measurements were taken in quadruplicate and at two different locations on each coupon. Two compouns for each conditions and time interval were used during each trial.

Salmonella cocktail biofilms grown on nylon, wood, and HDPE were imaged under a Laser Scanning Confocal Microscopy (LSCM) (LSM-5 Pascal associated with a Zeiss Axioplan 2) at the Microscopy Facility of the Biology Department at Kansas State University. Samples were imaged at 2, 24, and 96-h. The samples were stained with SYTO 9 (ThermoFisher Scientific, Eugene, OR) and SYTOX red (ThermoFisher Scientific, Eugene, OR) to differentiate live and dead cells [18]. Images and 3D projections were developed using ImageJ (NIH, USA).

Coupons were randomly assigned treatments across experiments. The attachment assay was performed in triplicate for each time condition tested (2, 24 and 96-h) and material (nylon, wood and HDPE). Coupons inoculated with Salmonella cocktail but not rinsed were considered positive controls, while not-inoculated coupons in TSB for 96-h as negative controls. Two coupons for each treatment were used during the indipendednt trials. The hydrophoboicity assay was instead performed in triplicate. Also for this experiment two coupons for each time condition tested (2, 24 and 96-h) and material (nylon, wood and HDPE). Clean not-inoculated coupons and clean inoculated coupons soaked in TSB for 96-h were used as controls for comparison purposes. Salmonella enterica LT2 was used to form biofilm, since assays were permorned in a BSL1 laboratory. Statistical differences were determined at P < 0.05. All results were analyzed in SAS Studio (Cary, NC) using the General Linear Model (GLM) procedure with Tukey’s multiple comparison.