The emergence of phage-resistant mutants during phage infection has beenreported in many studies 11,38–42 but the mechanisms of bacteria resistance to phagesare not yet completely understood. A previous study of our group 42 showed that the agent of furunculosis can be efficientlyinactivated by the phage AS-A (reduction of 4 Log CFU mL-1 after 8 hof treatment). However, some bacteria survived to the infection by the phagedue to the development of phage-resistance (2.24 x 10-4) 42.
The frequency of resistance was, nevertheless,limited as already reported in previous studies 13,43,44.So, with these two studies we verified that although1) a specific phage against the agent of the furunculosis can controlefficiently the bacterial growth, 2) some phage-resistant bacteria emerge aftertreatment; 3) being the resistant colonies after 4 and 5 streak-plating steps clearlydistinct from earlier streaking steps, (steps 1, 2 and 3); 4) showing asignificant modification in the expression of intracellular proteins whencompared with the phage sensitive bacteria; 4) but these modifications affectdistinct proteins after the first and the fifth streak-plating steps; 5)allowing “lysis from without” (positive spot test) after the forth streak-platingstep contrarily to that observed for bacteria from 1, 2 and 3 streak-platingsteps.It has been stated in the literature that resistance to phage can be overcomeby the phage itself because it evolves along with the host 45.
Moreover, it has been also statedthat resistance to phages entails larger costs to the bacteria 46. In fact, as observed for otherphages, colonies of AS-A phage-resistant mutants were smaller than coloniesformed by the non-phage added control 14. These results suggest that theremaining bacterial mutants (forming small size colonies) maintained theirviability in the presence of phages but their phenotypes were affected. Thesedecrease in the bacterial growth after phage exposure could be fitness cost,which can contribute to their elimination from the environment faster than theirwild-type parents. In this study, as already observed for other phages 14,15, it was also detected that phage-resistant bacteriaalso mutate after successive streak-plating steps.
Although the spot testsshowed negative results until the fourth streak-plating step, at the fourth andfifth steps, the spot test was positive, as already observed in other studies 14,15. These results were confirmed by infrared spectroscopydata of the whole cells. Infrared spectroscopy results shows that the spectraobtained from the fourth and fifth streak-plating days colonies are similar to onesfrom phage-sensitive control colonies, suggesting that these colonies are more similarto control phage-sensitive bacteria than the colonies from 1, 2 and 3streak-plating steps.
It seems that the resistant bacteria somehow “recovered”,being more similar to control bacterial populations, which are sensitive to thephage infection. The infrared peaks that contributed to these results werefound to be majorly associated with proteins. Taking this into account, wefocused the further studies on protein analysis with 1D SDS PAGE gels. Regarding the presumptively identified proteins withdifferential expression on first streak-plating phage-resistant clones, adecrease in band 8 is noticeable, being this band associated with a phagetranscriptional protein with regulation function in the transcription of phagegenes 47. This may be a response of the bacteria to the viralinfection, preventing the transcription of the viral genome. Similarly, theexpression of the protein corresponding to 9 band in first streak-platingclones decreased when compared to the control.
This protein, phage-shock Bprotein, is involved in a regulation system that responds to aggressions,habitually to phage secretins, promoting the defensive response of the bacteria48. This protein has been previously detected in theresponse of other bacteria, however, this response mechanism is not yetcompletely understood 48,49. In our case this protein is less expressed in thephage-resistant clones, which seems contradictory. Nevertheless, it was statedthat bacteria synthesise phage shock proteins after being infected with phage,that, in the case of the resistant clones, could not happen 50. Contrarily, the protein associated to band 13, tatA, increased in A. salmonicida first streak-plating clones. This protein belongs totat system(twin-arginine-translocation) which is responsible to the transport of varioussubstances at the membrane level, against the concentration gradient of thecytoplasm to the extracellular space, namely proteins, being associated to thebacterial pathogenicity in the secretion of virulence factors 51,52.
This increase suggests that these first streak-platingclones can be more virulent than control bacteria However, somestudies showed the phage-resistant clones are less pathogenic than phagesensitive bacteria 19,53. This suggestthat the increase in the expression of this protein can be associated to othermechanisms not related with pathogenicity.Regarding the proteins with differential expression onphage-resistant clones in the fifth streak-plating, that have a positivespot-test, band 16 suggests the expression of a transposase that is decreasedin these clones when comparing to control phage-sensitive bacteria. These typeof enzymes have the function to facilitate the transference of the geneticmaterial between organisms 54. The bacteria can decrease the expression of thisprotein as a defence mechanism in order to prevent the phage replication.
Band18, corresponds to a toxin-antitoxin protein, which is implied in themaintenance of plasmids, stress regulation and adaptation, as well as in thegrowth control and programed cellular death 55,56, also decrease in the clones of the fifthstreak-plating when compared to the control. As this protein decrease in thisstudy, it can suggest that in the fifth streak-plating clones the stress causedby the phage decrease. In fact, for the clones of the fifth streak-plating thespot test was positive, which suggest that the changes in the proteins turn thehost cells more sensitive to the phage suspension. However, the efficiency ofplating (EOP) results indicates that the fifth streak-plating clones do not replicatethe phage. Other authors 57 obtained similar results, designating this situationof “lysis from without”. The spot test lysis when the phage is not replicatedby the host (EOP is zero) has been described as a plausible mechanism whichhappens when an overload of phages simultaneously infects a bacterium leadingto lysis either from the action of phage lysins or from rapid depletion of thecells resources 58. As in the spot test the same volume of phagesuspension was used and lysis was only observed for the clones of the fourthand fifth streak-plating, the hypothesis of rapid depletion of the cellsresources does not seems plausible.
However, the lysis can be due to thepresence of lysins, explaning the positive spot test for the clones of thefourth and fifth streak-plating and the negative spot test for the clones ofthe first, second and third streak-plating. The modifications in the proteinsalong the successive streak-plating can allow the clones to recover thesensitivity to the phage lysins. This is in agreement with the infraredspectroscopy results which show that the fourth and fifth streak-plating cloneswere similar to the phage-sensitive bacteria (control), but clearly differentfrom those of the clones of the first, second and third streak-plating.
Taking into account all these results, we noticed thatthe different analysed clones present significant modifications inintracellular proteins related to phage infection, either in the first and fifth streak-plating. However, there are more proteins thatare differentially expressed in clones of the first streak-platingthan in clones of the fifth thefifth streak-plating, which is in accordance with infraredspectroscopy results. The fact that the phage-sensitive control bacteria haveinfrared spectra that are more similar to the fourth and fifth streak-plating clones may be because the cellular envelope, used by thephages to infect the bacteria, became more similar in these cases. This may berelated to the fact that the spot test turns positive again for the fourth and fifth streak-plating clones, which can be due tophenotypical similarities in the cell envelope. In order to confirm theseresults, the identification of the proteins that show differential expressionbetween the clones should be performed and future studies are being done. 4. Materials and Methods 4.1.
Bacteriaand phageThe bacteria A.salmonicida CECT 894 was used in this study. Fresh plate bacterial cultureswere maintained in solid Tryptic Soy Agar medium (TSA; Liofilchem, Italy) at 4°C. Before each assay, one isolated colony was aseptically transferred to 10 mLof Tryptic Soy Broth medium (TSB; Liofilchem, Italy) and was grown overnight at25 °C.
An aliquot of this culture (100 ?L) was aseptically transferred to 10 mLof fresh TSB medium (Liofilchem, Italy) and grown overnight at 25 °C to reachan optical density (O.D. 600) of 0.8, corresponding to about 109cells mL?1.Phage AS-A was isolated from sewage water from a liftstation of the sewage network of Aveiro, Portugal (station EEIS9 of SIMRIAMulti Sanitation System of Ria de Aveiro) using A. salmonicida as host, according to 42. The phage stocks were stored at 4 °C and 1%chloroform (final volume) (Scharlau, Spain) was added.
The phage suspensiontitre was determined by the double-layer agar method using TSA (Liofilchem,Italy) as culture medium 59. The plates were incubated at 25 °C for 12 h and thenumber of lysis plaques was counted. The results were expressed as plaqueforming units per millilitre (PFU mL-1).
4.2. Isolationof A. salmonicida phage-resistantmutantsOnly bacterial colonies that were resistant to thephage were used (bacteria that developed inside phage plates). For this,bacteria A. salmonicida and phageAS-A were plated by the double layer agar method and the plates were incubatedfor 24 h at 25 °C.
After that, several colonies that grew inside the phageplates, thus, resistant to phage infection, were visible. Three individualizedcolonies (A, B and C) were picked and used in the subsequent assays. 4.3. Detectionof bacteria sensitivity to the phage after one cycle of phage contact The phage resistant colonies obtained in the section4.
2. were used. The colonies were inoculated in TSB medium for 24 h at 25 °C.After that, the culture was used to perform a spot test and was also plated inTSA medium. This procedure was done 4 more times, making a total of 5 streakplating steps.
This procedure was made for the 3 selected colonies. 4.4.
Efficiencyof platting (EOP) The efficiency of plating was determined for bacteriathat shown positive spot tests (clear lysis area), i.e. for the bacteria fromfourth and fifth streak-plating steps, according to Pereira et al.
14 using the double-agar method 59. The EOP was calculated (average PFU on targetbacteria/ average PFU on host bacteria), three independent assays wereperformed. 4.
5. PhageadsorptionThe determination of phage adsorption was performedaccording to Pereira et al. 14. Briefly, ten microliters of phage suspension ofabout 106 PFU/mL were added to 10 mL of A. salmonicida culture of about 109 CFU/mLcorresponding to an optical density (600 nm) of 0.8 60 and incubated at 25 °C.
Aliquots of this culture werecollected after 0, 5, 10, 15, 20, 25, 30, 40, 50, 60 and 70 minutes ofincubation and chloroform was added to a final concentration of 1%. The mixturewas centrifuged at 12000g for 5 minutes, after that the supernatants werefiltered using 0.2 µL membranes (Millipore, Bedford, USA). The filtratescontaining unadsorbed phages were then diluted and titrated.
The plates werethen incubated at 25 °C and observed after 8h for plaques formation. The valueswere calculated as decrease of phage titre in supernatant (percentage) comparedwith time zero. Three independent assays were performed. 4.6. InfraredspectroscopyIn order to access the spectral differences ofsensitive A.
salmonicida colonies andphage resistant mutant colonies, mid-infrared spectroscopy was used, as it waspreviously described 14,25. They were used the A. salmonicida phage resistant colonies A, B and C (from section 43).To analyse the whole cells, colonies A, B and C were analysed along the5 days of streaking (section 4.
3), as well as control sensitive colonies Ct1and CT5 (after 1 and 5 streak plating steps). The colonies were collected witha loop and placed in the crystal of a horizontal single reflection ATRaccessory. The colonies were gently air dried and the spectra were acquired.
Spectra were done in a MIR (Bruker ALPHA FTIR spectrometer, Germany)with a resolution of 4 cm-1 and 32 scans, in the infrared region(4000 to 600 cm-1). At least 5 replicate spectra were performed foreach colony. Mid-infrared spectra were obtained in OPUS format (OPUS 6.5,Bruker, Germany) and transferred via JCAMP.DX format for use in a housedeveloped data analysis software (CATS build 97). The spectra were SNV(standard normal deviate) corrected previous to multivariate analysis.Principal component analysis (PCA) was done in order to find the major sourcesof variability in the spectra and to detect groups. 4.
7. Intracellularproteins extraction and quantificationThe proteins extracts were obtained from the growthuntil the late exponential phase of the strains (OD 0.9 at 550 nm) in LuriaBertani Broth (Merck, Germany). The cells were separated from the supernatantby centrifugation at 8000xg for 10 min at 4 °C. The protein extractions weremade in three independent experiments per each strain and the proteinquantification was performed in triplicate.
The cell pellets were washed three times in 10mMphosphate buffered saline pH 7.4. After that they were resuspended in 1 mL oflysis buffer solution 7 M urea, 2 M thiourea, 4% cholamidopropyldimethylammonio-1-propanesulfonate (CHAPS), 30 mM Tris base, pH 8.5. Crudecell-free extracts were obtained by sonication in ice to minimize proteindamage, during 2min, using a 30% duty cycle, 2 s pulses with interveningperiods of 3 s. The intracellular protein solution was incubated with 1 mg.
mL-1of Dnase I (GE Healthcare, Sweden) and 10mM of protease inhibitor mix (GEHealthcare, Sweden) during 1 h at 15 °C . The final solution wascollected by centrifugation at 20000xg for 40min at 4 °C and then,the protein concentration was measured using the 2-D Quant Kit (GE Healthcare, Sweden), following themanufacturer’s instructions. The procedure was performed in triplicate. 4.
8. Proteinseparation by 1-D electrophoresisProteinswere separated by 12.5 % SDS-PAGE 61, in aMini-PROTEAN 3 Cell (Bio-Rad, USA),for 50 min at 150 V. Proteins were visualized by colloidal Coomassie BrilliantBlueG-250 (CBB) staining 62. Gel imageswere acquired using the Gel DocTM XR+ (Bio-Rad, USA).
The comparative analysisof the acquired images was performed in Image Lab v3.0 software (Biorad, USA)and based in the optical density measurement of each band. The result wasexpressed in band percentage, resulting from the value of the optical densityof a given band in the total of the bands per lane x 100.
The comparison of thedifferential expression of the intracellular proteins of the different tested A. salmonicida clones in the differentanalysis times was made through a two-way ANOVA, using GraphPad Prism softwarev7 (USA). The differences were considered statistically significant when p <0.05. 4.9. Presumptive identification of theproteins in differentially expressed bandsThemolecular weight of the bands that were differentially expressed betweencontrol and A.
salmonicida clones onday 1 and between control and day 5, recurring to databases UniProtKB (www.uniprot.org) andNCBI (www.ncbi.nlm.
nih.gov/pubmed) allowed us to presumptively identify theproteins and their function, based on the deposited genome of Aeromonas salmonicida A449.