Abstract
Yersinia ruckeri is a facultative intracellular enterobacterium mostly known as the causative agent of enteric redmouth disease in salmonid fish. In the present study, we applied RNA inhibition to silence twenty pre-selected genes on the genome of a fish cell line (CHSE-214) followed by a gentamicin assay to quantify the effect of silencing on the cells’ susceptibility to infection and found that silencing of 18 out of 20 genes significantly reduced the number of Y. ruckeri recovered. These findings improve our understanding of the infection process by Y. ruckeri and of the interactions between this bacterial pathogen and host cells.

Introduction, methods and results
The enterobacterium Yersinia ruckeri is a major fish pathogen worldwide and has mostly been studied as the causative agent of enteric red-mouth disease in salmonid fish [1, 2]. Several virulence factors have been described in Y. ruckeri (see reviews by Kumar et al. and Wrobel et al. [1, 3]). Y. ruckeri causes septicaemia and haemorrhages leading to high levels of mortality in infected fish [1, 2]. The bacterium also has zoonotic potential and has been associated with topical infections in humans [4]. Like several other members of the genus Yersinia, Y. ruckeri has demonstrated the ability to invade non-professional phagocytic cells [5, 6], allowing the bacterium to access restricted nutrients and protecting it from the immune system. It might help the bacterium to cross epithelial membranes as is the case for other Yersiniaceae.

Two main mechanisms of entry have been described in bacteria, including Yersiniaceae. The zipper mechanism is initiated by the binding of the bacterial adhesins to specific molecules on the cell membrane (see the review on yersinia adhesins by Chauhan et al. [2]). In Y. pseudotuberculosis and Y. enterocolitica, two autotransporter proteins invasin (Inv) and Yersinia adhesin A (YadA) interact with integrin receptors [7, 8]. Interactions of these bacterial proteins with receptors on the surface of the host’s cells lead to the recruitment of more receptors, activation of Rac1 and cytoskeletal rearrangement culminating in the uptake of the bacterium [9]. Interestingly; while Y. ruckeri does not harbour a YadA equivalent, it is known to harbour two homologs of invasin: Y. ruckeri invasin (YrInv) and Y. ruckeri invasin-like molecule (YrIlm) [3]. Furthermore, several isolates of Y. ruckeri belonging to the O1 serotype also carry a cluster of fimbrial gene homologous to the STF cluster of S. typhimurium [10, 11]. However, this cluster is absent from the ATCC 29473 type strain studied here [11].

The other main mechanism of entry studied is the trigger mechanism that relies on effector proteins secreted through the type three and type four secretion system (T3SS and T4SS). Once inside the host’s cells, these effector proteins interact with regulatory proteins in the host, in particular members of the Rho family (RhoGTPases Rac, Cdc42 or RhoG) [12], leading to a rearrangement of the cytoskeleton of the host cells and the uptake of the bacterium [12]. Intriguingly, the T3SS of Y. ruckeri actually belongs to the Ysa family, a different family from the one most studied in Yersiniaceae and is more closely related to the T3SS carried on Salmonella pathogenicity island 1 (SPI-1) of Salmonella enterica [13, 14]. Among the proteins carried on SPI-1 is the chaperon Invasion protein B (InvB); SPI-1 is known to play a role in the intracellular invasion of S. enterica [15] so the Ysa of Y. ruckeri could plausibly be involved in intracellular invasion of Y. ruckeri. However, our knowledge of Ysa T3SS, and that of Y. ruckeri in particular is still very incomplete [10] and no conclusion is currently possible.

An important feature of both invasion mechanisms is that they require active uptake of the bacterium by the cell and it is possible to prevent host cells from internalising the bacteria, for example by treating them with chemical blockers [5, 6]. Similarly, silencing of host genes has been shown to inhibit bacterial internalisation. For example, 305 host genes have been associated with the ability of Listeria monocytogenes to invade cell cultures of drosophila SL2 cells and 86 genes necessary for invasion by Mycobacterium fortuitum [16]. In human macrophages, 270 molecules have been identified whose silencing affected the bacterial load of seven different field isolates of Mycobacterium tuberculosis [17]. Finally, 252 genes have been shown to be involved in the survival of Salmonella enterica serovar Typhimurium within the human epithelial cell line MCF-7 [18]. Among these genes, the most impactful were SEC22A, Rab1B and VPS33B that were involved in vesicle trafficking as well as ATP6VOD1, an ATPase involved in vacuole acidification and the iron transporter FTHL17. Interestingly, the authors noted a significant overlap between their findings and that reported by Kumar et al. for M. tuberculosis [17] suggesting that many of these targets are well conserved even between very evolutionary distant bacteria.

In the present study, we aimed to produce siRNA to silence 20 genes commonly involved in invasion by bacterial pathogens (see Additional file 1), including surface integrins since such molecules are often targeted by the invasin molecules expressed by other Yersiniaceae [2]. Other genes were cytoskeletal molecules and genes involved in vacuolar trafficking and maturation [17, 18] as well as agents of the cytoskeletal apparatus since this plays a central part in the internalization of bacteria through both zipper and trigger mechanisms [12]. The selection also included genes and pathways shared by L. monocytogenes, M. tuberculosis as well as S. typhimurium [16,17,18] since these pathways are used by such diverse facultative intracellular pathogens, they might also be involved in the intracellular invasion of Y. ruckeri.

The fish cell line Chinook Salmon Embryo (CHSE-214) was used since we wanted to investigate an epithelial cell line derived from salmonids. In previous experiments using multiple cell lines we showed that it was well suited for invasion by Y. ruckeri. PCR and sequencing were applied to confirm that the CHSE214 cells were indeed derived from Chinook salmon (Oncorhynchus tshawytscha). Similarly; the study focussed on the type strain ATCC 29473, a type strain belonging to biotype 1, since these experiments showed that it had a high potential for invasiveness [5].

siRNA transfection
Sequences for the siRNA were designed using the Silencer Select siRNA design algorithm. For the genes for which no O. tshawytscha sequences were available, sequences from other members of the Oncorhynchus genomes were used. Upon reception, the siRNA were resuspended to 20 µM and stored at −20 °C until use, according to the manufacturer’s instructions.

CHSE-214 cells were grown in 24 well plates at 20 °C supplied with 1250 µL Minimum Essential Medium (MEM-glutaMAX™, Gibco, Thermo-Fisher Scientific, Waltham, USA) containing 2% FBS, where they reached about 80% confluence 1 day after seeding. On the day of the assay, 8 µL of the siRNA solution was added to 400 µL of Opti-MEM medium (Gibco). Seventy-two microliters of RNAiMax lipofectamine reagent (Sigma Aldrich, St Louis, USA) was diluted in 1200 µL of Opti-MEM medium. The two solutions were mixed and incubated 5 min at room temperature. Afterwards, the culture medium on top of two adjacent rows of the 24 wells was replaced with 50 µL of the solution and, after 20 min, 450 µL of fresh MEM-Glutamax medium was added.

Ambion®Silencer® Negative Control #1 (Thermo-Fischer Scientific) was applied as a negative control. The cells were incubated at 20 °C and since the temperature commonly used for transfection was lower than for other cell lines, incubation time was extended to 3 days.

Bacterial invasion assay
The gentamicin assay was performed as previously described [5]. Briefly, Y. ruckeri ATCC 29473 was cultivated overnight in brain heart infusion (BHI, Oxoid, Thermo-Fisher Scientific). Optical density of the culture was assessed by spectrophotometry (Eppendorf Biophotometer, Hamburg, Germany) and adjusted to an optical density of .5 at 600 nm. Bacteria were pelleted by centrifugation at 3220 g and 20 °C (in an Eppendorf Centrifuge 5810 R) and resuspended in 10 times the original volume of MEM-Glutamax. Then, the culture medium in the leftmost row was replaced by this bacterial solution and the bacteria were left to interact with the cells for 5 h at room temperature. In the other row, the medium was replaced by fresh MEM-glutamax without bacteria. After that time, the medium was removed, the cells were washed three times with phosphate buffered saline (PBS) and a fresh volume of MEM-glutamax, supplemented with the antibiotic gentamicin (Sigma-Aldrich) at a concentration of 100 μg/mL was added to the wells. The antibiotic was left to act for 4 h. Afterwards the cells were washed twice with PBS before replacing the medium supplemented with 1% Triton-X (Sigma-Aldrich). After 10 min of exposure to the detergent, the cells were triturated with a micropipette and serially diluted from 10−1 to 10−4 before being plated onto Brain heart infusion agar (BHIA, Oxoid). Each of the three wells represented a biological replicate and each dilution was plated in technical quadruplicates, meaning that 48 plates were inoculated for each siRNA. The agar plates were incubated at 22 °C until clear colonies were visible and counted (generally after 48 h) and the average CFU per mL value was calculated.

Controls
For each siRNA, the cells numbered 1D, 3D and 5D were left un-inoculated in each 24 well plate then lysed and plated to act as a negative control and detect any contamination of the reagents. Similarly, the cells in the wells 2D, 4D and 6D were also left un-exposed to the bacteria and a trypan blue assay was performed to rule-out a toxic effect of the siRNA silencing: The cells were washed three times in PBS, then .2% Trypan Blue (Gibco, .4% diluted 1:1 in PBS) was added. After 1 min, fixation was performed with 4% formalin for 10 min. The cells were then rinsed until any trace of blue dye had disappeared. The plates were kept at 4 °C until quantification of the cells using an inverted microscope (Leica DM IRB, Wetzlar, Germany). One hundred cells were counted and the number of blue stained cells among them was recorded. The procedure was repeated 4 more times for each culture unit to result in 5 percentage values for each siRNA. These were then compared to the survival of the control cells without siRNA to confirm that the silencing procedure did not have a toxic effect. In no instances did these numbers differ significantly from that of the control.

Preparation of the cDNA samples
Finally, the cells in the rightmost row for each treatment were lysed in buffer RLT (Qiagen, Hilden, Germany). The cell suspension was homogenised using QIAshredder columns (Qiagen) and centrifugation at 145 000 RPM for two minutes at room temperature using MiniSpin tabletop centrifuge (Eppendorf). Afterwards, the RNA were extracted using the Rneasy mini kit (Qiagen). RNA were immediately quantified using a Nanodrop machine (Thermo-Fisher) and cDNA were immediately synthesised using Iscript kits (Bio-Rad) in a C1000 Touch thermocycler (Bio-Rad, Hercules, USA). cDNA were stored at 4 °C until use.