Isolation ssDNA aptamers specific for both live and viable but nonculturable state Vibrio vulnificus using whole bacteria-SEILEX technology

Vibrio vulnificus is a ubiquitous marine bacterium that may cause rapid and deadly infection, threatening lives of people living around natural bodies of water, especially in coastal regions. However, traditional culture-based methods are time-consuming and unable to detect Viable But Non-Culturable (VBNC) V. vulnificus cells. In this work, we isolated a batch of detection aptamers specifically binding to V. vulnificus in all culture status. With traditional whole bacteria-SELEX (Systematic Evolution of Ligands by EXponential enrichment), flow cytometer analysis and imaging, we identify 18 candidates and validated two of them (V8 and V13) as applicable aptamers. Their truncated sequences also showed comparable performance. The dissociation constant (KD) value of V8 is shown to be as low as 11.22 ± 1.32 nM. Optimal aptamers V8 and V13 are also validated to be effective to detect different Vibrio vulnificus strains under different binding environments using flow cytometry. As for detection parameters, the LOD of the V8 from cytometry is 29.96 CFU mL−1, and the linear range is 102–5 × 105 CFU mL−1. This is the first case demonstrating that aptamers can detect the existence of VBNC bacteria as well as live bacteria.


Introduction
Vibrio vulnicus is a Gram-negative, halophilic, agellated, ubiquitous marine bacterium which can cause serious infection among people all over the world, especially those in coastal states and islands. As part of normal microora, it can be found in waters, oysters and other shellsh. It is one of the most prevailing marine pathogens. As the FDA (Food and Drug Administration) of the United States reported, there were 459 cases from 1992 to 2007 with a fatality rate of 51.7%. 1,2 Most cases (85%) occur in the warm water months of May to October in the Northern Hemisphere. People typically get infected through foods. V. vulnicus is responsible for 95% of seafoodrelated deaths in the US. Directly being exposed to contaminated water with a pre-existing lesion or cut can also cause infections. Patients with chronic and underlying diseases, especially liver diseases, are more likely to get infected. Infection symptoms can appear soon aer the oyster ingestion, and the onset time can be as soon as 4 h. 3 If patients can't receive antibiotic treatments in three days, the fatality rate can be 100%. 4 Thus, monitoring the presence of V. vulnicus in waters and seafood is of medical and economic importance.
Current methods for the identication and isolation of V. vulnicus from environmental or clinical samples typically rely on selective medium culture, which require further experiments such as PCR to identify the presumptive isolates. [5][6][7] Although such methods are quite accurate for the identication and validation of candidate microorganisms like V. vulnicus, however, there are still two major shortcomings for using this method in V. vulnicus clinical detection: Firstly, the culture-based methods usually take 3 or 4 days to get the nal results, which are quite time-consuming. And considering that patients infected by V. vulnicus can only survive three days aer infection, it's IMPOSSIBLE to use culture-based diagnosis method for such acute urge infection of V. vulnicus.
Secondly, under some extreme conditions (like low temperature in winter), V. vulnicus has been reported to be able to transformed into a specic state called Viable But Non-Culturable (VBNC) state, making it possible to escape traditional culture based detection. Although with adequate cells (10 8 /mL or more), 8 the VBNC V. vulnicus can be detected through PCR, however, in nature environment, such concentration of VBNC cells can't be reached, and a small number of cells (about 100 cells or less) is enough to cause serious diseases. 9 Rapid and quick direct detection of pathogens like V. vulnicus can be used as a supplement for traditional standard methods to not only provide signicant infectious information on patients quickly and effectively, like but also overcome the two shortcomings of traditional methods we mentioned above, making it possible for infected patients to get correct treatment in time.
Multiple subtypes of molecules have been applied for pathogen diagnosis as detectors. Among them, single-stranded nucleic acid can fold into unique and stable structures and makes it an ideal choice. Some single-stranded DNA (ssDNA) or RNA chains can specically bind to various targets, such as metal ions, small molecules, drugs, proteins, and even whole cells, [10][11][12][13] and they are named as aptamers. Aptamers are very suitable to be developed into diagnostic and therapeutic tools. They can be easily modied with dyes or chemical tools and can be easily immobilized on many kinds of substrates, make it more suitable for quantitative measurements. [14][15][16] Compare with antibody, aptamer is smaller, more stable and can tolerate a wide range of temperatures. Additionally, aptamer can be produced in vitro without torturing animals and can be rapidly synthesized with high purity and little batch-to-batch variation.
Aptamers are isolated by Systematic Evolution of Ligands by EXponential enrichment (SELEX) procedures. 17 There are two common strategies to isolate bacteria aptamers, using specic cell surface molecules as targets, or the whole cells as targets. For the rst strategy, however, choosing and purifying species-specic membrane molecules can be relatively difficult, and puried molecules may not be able to reserve their native structures when their microenvironments are changed, so their aptamers may not be able to bind to cells. Under such circumstance, the whole-bacteria SELEX approach can be applied and perfectly solves the problem. Whole-bacteria SELEX approach can efficiently screen aptamers with high affinity without tedious molecule purication process, and counter SELEX can ensure the high specicity of the aptamers. In recent years, whole-bacteria SELEX approach has been applied to isolating various bacteria aptamers such as Mycobacteria tuberculosis, 18 Staphylococcus aureus, 19 Candida albicans, 20 Vibrio. Alginolyticus, 21 Vibrio. Parahaemolyticus, 22 etc.
In this study, we employed whole-bacteria SELEX approaches to screen potential ssDNA sequences that can specically bind to V. vulnicus. Based on FACs (Flow Cytometry) and confocal microscopy results, we identied 2 aptamers with good performance, V8 and V13. They can bind to V. vulnicus with high binding affinity and can specically pick V. vulnicus out of other bacteria. Such optimal aptamers can detect different V. vulnicus strains, tolerate different binding environments, and detect V. vulnicus in VBNC status as well as cells in other culture phases. As for the general performance of such aptamerbased detection method, the LOD is 29.96 CFU mL À1 and its linear range is 10 2 -5 Â 10 5 CFU mL À1 . The total test time of this fast screen method is less than 1 hour. Based on our screening methods and analysis results we described above, there are three major innovations that can be summarized in this research eld.
Firstly, we developed ssDNA aptamers for V. vulnicus with the best binding affinity up to now. The K D value for Aptamer V8 is 1.22 AE 1.32 nM, while the previous studie got an aptamer with K D ¼ 26.8 AE 5.3 nM; 23 Secondly, for the rst time, we provided an optional method to detect environmental at V. vulnicus VBNC status, fullling the gaps in this research eld.
Thirdly, previous studies generally applied bacteria in one xed growth period/stage, and they didn't test whether these aptamers can bind to target bacteria in different phases. Therefore, the screened-out aptamers may not be able to identify bacteria if such bacteria entered a different stage, inducing possible false negative results. As we have mentioned above, V. vulnicus have various stages. Therefore, it's quite necessary for researcher to screen and validate the candidate aptamers using bacterium from different various stages. However, previous studies on V. vulnicus do not contain related screening and analysis. Our study lled the gap and identied the effective aptamers that can be applied to V. vulnicus at different phases.
All in all, taking advantages of whole-bacteria SELEX approaches and ow cytometry, in this study, we not only screened out two effective aptamers (V8 and V13) for further clinical and environmental detection on V. vulnicus with higher detection speed and accuracy, but also validated their ability to bind to V. vulnicus in different stages, especially in VBNC status.

Selection of aptamers against V. vulnicus
The scheme of whole-bacteria SELEX process was illustrated in Fig. 1. Thirteen rounds of selection were performed to isolate aptamers that can specically recognize V. vulnicus. During the rst seven rounds of selection, sequences bound to V. vulnicus were collected and amplied. The recovery rates increased gradually, and we decided to introduce counter-SELEX in Round 8 (see ESI Fig. S1 †). Through the subsequent six rounds of counter-selection, the unbound sequences and sequences bound to V. parahaemolyticus were discarded. The recovery rate of Round 11 increased signicantly, indicating more enriched sequences in the system can bind with V. vulnicus. 2 more rounds were added for further eliminating non-specic sequences and enrich candidate ones. PCR negative controls were set in each round to avoid possible template contamination, no detectable products yielded. When the whole SELEX process ended, the pool was puried, cloned and sequenced. 80 candidate oligonucleotides were sequenced and their homology and similarity were analyzed via Clustal X 2.0. 18 candidate aptamers for further identication (see ESI Table 1 †). This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 15997-16008 | 15999

Binding ability analysis of candidate aptamers
In order to compare binding ability of the candidate sequences, we carried out ow cytometry to analysis the incubated V. vulnicus cells. The candidate aptamers were synthesized with uorescent labels, when aptamers bind on cells, ow cytometer can detect the uorescent signals and count the uorescent cells. Mean uorescent intensity was used to screen oligonucleotides with higher binding ability. A uorescently labelled, randomized ssDNA pool was used as background of nonspecic binding. A threshold was set so that the uorescence intensity of the gated cells would be greater than those incubated with randomized ssDNA pool. 18 tested sequences were incubated with 4 Â 10 8 of V. vulnicus cells in binding buffer for 30 min, and the percentage of uorescent cells and their mean uorescence intensity are analysed and shown in Fig. 2.
Aer comparing the gated cell proportions and the mean uorescence intensity values of the gated cells, we choose V8 and V13 for further study.

Binding specicity analysis of V8 and V13
Fluorescently labelled aptamer sequences, V8 and V13, and their truncated sequences, TV8 and TV13 (sequences without primer regions) were incubated with various species of bacteria, including V. vulnicus, V. parahaemolyticus, V. alginolyticus, S. aureus, L. monocytogenes, C. albicans, and P. aeruginosa. Fluorescent intensity of different species of bacteria was tested. Fig. 3 clearly shows that when bound to V8 and V13, uorescent intensity of V. vulnicus was signicantly greater than other species. Truncated sequences, TV8 and TV13 also inherited the good specicity from V8 and V13, respectively. These results indicated that these sequences can distinguish V. vulnicus from other Vibrio bacteria as well as other species.

Fluorescence imaging of V. vulnicus-aptamer complex
In order to further test the binding ability between V. vulnicus cells and aptamer, laser scanning confocal microscopy was applied to observe the V. vulnicus-aptamer complex. V. vulnicus cells were incubated with 4 aptamer sequences respectively. Then the suspension was dropped on glass slide and thin smear was made. Through laser scanning confocal microscopy, we can see clearly that these sequences did binding to the cells efficiently (see

Binding affinity analysis of V8 and V13
In order to determine equilibrium dissociation constant of V8 and V13, and their truncated sequence TV8 and TV13, V. vul-nicus cells were incubated with different concentrations of aptamers. Fig. 5(A) shows one site saturation curves based on the ow cytometric analysis results. As Table 1 shows, K D values of four sequences reached nanomolar level, and both selected sequences and truncated sequences showed excellent binding capacity. The secondary structure of these 4 aptamers are shown in Fig. 6. V8 and TV8 have the same loop, and part of V13 and TV13 loops are the same. These results suggest that the 25-nt variable regions of the aptamer sequences are more likely to be responsible for binding to the target.

V8 and V13 bind to different concentrations of V. vulnicus
In order to further investigate binding ability of aptamers against V. vulnicus, we incubated 300 nM V8 and V13 with a series of different concentrations of V. vulnicus. According to binding affinity analysis results, 300 nM V8 and V13 can achieve binding saturation. As Fig. 5(B) shows, different concentrations of V. vulnicus showed similar uorescent signals when incubating with an excess concentration of aptamers.
V8 and V13 bind to V. vulnicus in all culture status Some articles state that aptamers have the potential to be employed to detect bacteria in VBNC status. However, none of them demonstrated this point clearly and directly. 20,24,25 We tested whether V8 and V13 can bind to V. vulnicus in VBNC status as well as other culture status. As Fig. 7 shows, V8 and V13 showed similar performance when bound to V. vulnicus in different culture phases.       Table 3 and Table 4 † for detailed data.
The linear range of V8-ow cytometer detection method The linear range of V8-ow cytometer detection method is measured according to CLSI-EP6A. 26 Good linear relationship between plate count result and V8-ow cytometer detection result at the range between 10 2 to 5 Â 10 5 CFU mL À1 is shown on Fig. 8. The regression coefficient at this range is 0.9994. Flow cytometer loses its resolution at higher concentration, probably because it cannot analyse large number of particles at fast uid speed.
Aptamer V8 bind to protein-free V. vulnicus V8 showed best binding affinity and its target molecule was explored. We treated V. vulnicus with Proteinase K and trypsin to see whether the target of V8 is protein. If the uorescence signal reduces, the target of the aptamer may be protein molecules on bacterial cell wall. However, as ESI Fig. S4 † online shows, the treatments did not cause signicant signal change. This result suggested that the target might not be membrane protein, but other cell wall composition, such as LPS, capsule polysaccharide or other kinds of molecules.

Discussion
All in all, as we have described above, we applied modied whole-bacteria SELEX and got aptamers that can specically bind to V. vulnicus. The accuracy and efficacy of the optimal aptamers have also been validated under different circumstance and against different strains of V. vulnicus, and the aptamers showed excellent performance. Here, in this study, we divided further discussion on our project into three parts: methodological innovations, clinical signicance of new aptamers and further perspectives in this eld.
For methodological innovations, we have made three major innovations during the selection, validation and application of the optimal aptamers against V. vulnicus.
First, in this study, we emphasized on developing aptamers that can bind to different stages of V. vulnicus. In previous studies, lots of aptamers against bacteria were selected by  SELEX and many of them were developed into splendid detection method. 24,[27][28][29][30] However, most of them failed to pay attention to bacteria at different stages and relied on single phase of bacteria for aptamer selection and validation. Then, some scholar started to noticed the phase differences. Ying Zou et al. developed aptamers that can bind to different stages of E. coli O157:H7. 31 Soo HwanSuh et al. developed aptamers specically targeted bacterial cells at different growth phases. 32 Thus, we innovatively used V. vulnicus at different stages for positive SELEX, and used different stages of bacteria to validate the aptamers. We especially validated aptamers' binding ability to VBNC state. We identied aptamers with high affinity against different phases of bacteria, improving the accuracy and efficacy for aptamer selection.
Moreover, we chose V. parahaemolyticus as counter selection targets to confer the optimal aptamers potential clinical signicance. Previously, the aptamer selection for V. vulnicus has been reported by Yan, et al. 33 Their work was more focused on sh diseases and they used V. anguillarum for counter-SELEX, for further better distinguishing different water pollution pathogens for sh. While in this study, we chose V. parahaemolyticus with more clinical signicance for the counter selection, making the optimal aptamers effective enough to be applied in the differential diagnosis between V. vulnicus. and V. parahaemolyticus. For the rst time, we identied effective aptamers for V. vulnicus with potential clinical signicance.
Third, we developed an effective, fast and labour-saving method to detect V. vulnicus in VBNC status. In this work we used ow cytometer as a part of detection system. Compare to other existing proved VBNC detection methods, such as PCR, 34 qPCR, 35 PMA-LAMP (propidium monoazide-loop-mediated isothermal amplication) 36,37 PMA-based qPCR, 38 CRENAME + rtPCR 39 (rapid concentration and recovery of microbial particles, extraction of nucleic acids and molecular enrichment), the LOD of ow cytometer is slightly higher or equivalent. There is a trade-off between "easy to operate "and" extreme sensitivity". LOD of CRENAME + rtPCR can reach to 1.2 CFU/100 mL, but this approach takes 6 h to get the results. V8-ow cytometer detection is fast, with <1 h total test time and easier sample preparation.
For biological and clinical signicance of new aptamers, as we have already analysed, the optimal aptamers we screened can be further applied and modied for clinical use. Therefore, according to related publications, we compared the performance of our optimal aptamers with previous proved detection methods. In 2018, Yan, et al. 33 presented an effective detection method and identied effective aptamers which we have already mentioned above. Comparing to this study, we have two advantages: Firstly, comparing to their optimal aptamer Vapt2 with binding affinity at 26.8 AE 5.3 nM, our optimal aptamer (V8) has a better and more stable performance with binding affinity as 11.22 AE 1.30 nM. Secondly, different from previous studies using microscope to detect, which is labour-consuming and less quantitative in practical microbe detection, we applied ow cytometry to detect the performance of aptamers, which is quite fast, convenient for sample preparation and more suitable for further clinical applications.
As for future perspectives, we may still focus on the optimal aptamers we screened out in this study like aptamer V8 and V13. There are two major directions that we may contribute to in the future: Firstly, we will try to reveal the detailed biological mechanism for the aptamer to bind the target bacteria. In this study, we have already applied related experiments to make the preliminary trials in this eld. Using optimal aptamer V8 and cultured V. vulnicus strains, we incubated V8 with protein-free cells. And aer incubation uorescent did not change much, indicating that the target molecule of aptamer V8 is not protein but another cell wall component. Since the expression of some outer membrane protein may alter in different culture phases, it is reasonable that our SELEX strategy yielded aptamers that can bind to component that expresses on most V. vulnicus cells. Further studies are needed to identify the exact target molecule of V8.
Secondly, we will try to develop a reliable and fast detection kit using the aptamers for the detection of V. vulnicus at different stages. Although based on our experiments, V8cytometry method have already been validated to be effective and accurate enough for the identication of V. vulnicus at different stages, more modication should be made and a more efficient and sensitive biosensor or clinical diagnosis kit should be established. More experiments should be applied to make sure the stability and accuracy of our kit in clinics or for environment monitoring. Therefore, we may focus on the development of novel V. vulnicus detection methods, by applying technologies like aptamer-conjugated nanoparticles, quantum dots or colorimetric approaches to increase the sensitivity, handleability and throughput of the detections.

Materials and methods
Bacterial strains and culture media V. vulnicus ATCC 27562, V. parahaemolyticus ATCC17802 and V. alginolyticus ATCC 17749 were obtained from American Type Culture Collection (ATCC), Georgetown, DC, USA. V. vulnicus MCCC 1A08743 and V. vulnicus MCCC 1H00047 were obtained from Marine Culture Collection of China. Vibrio bacteria were grown in brain heart infusion medium (Land Bridge, Beijing) with 3% NaCl at 37 C. V. vulnicus in three different culture phases early exponential phase (OD 600 ¼ 0.3), late exponential phase (OD 600 ¼ 0.7), and stationary phase (OD 600 > 1) were mixed at the ratio of 1 : 1 : 1 in number. The mixture was used for positive SELEX. V. parahaemolyticus was cultured and prepared the same as V. vulnicus for counter-SELEX.
The following bacteria were also used to identify the speci-city of the isolated aptamers: Staphylococcus aureus ATCC25923, Listeria monocytogenes ATCC19115, Candida albicans ATCC10231, and Pseudomonas aeruginosa ATCC27853. All these bacteria were cultured overnight under aerobic condition in brain heart infusion media (Land Bridge, Beijing) with 1% NaCl at 37 C and 150 rpm shaking. All these bacterial strains were obtained from ATCC.
V. vulnicus under VBNC status was obtained by refrigerating cultured cells in articial seawater (ASW) under 4 C for 7 or more days. The viability was tested on culture plates and growth should not be detected. The VBNC cells were subject to PI stain and showed negative results in ow cytometer.
Membrane protein was removed as previously described with a little modication. 40 15 min of trypsin treatment (0.25%, 37 C) and 10 min proteinase K treatment (1 mg mL À1 , 65 C) were conducted. The reaction was stopped by adding Phenylmethanesulfonyl uoride (PMSF, Beyotime, China) into the suspension.

Random DNA library and primers
The initial single strand DNA (ssDNA) library and the primers used to amplify DNA were all obtained from Sangon Biotech (Shanghai, China). This ssDNA pool consisted of a central randomized region of 25 nucleotides anked on both sides by primer regions for amplication, 5 0 -AGTATACGTATTACCTGCAGC-N25-GCAAGATCTCCGAGA-TATCG-3 0 (66-mer).
The ssDNA library contained a maximum of 10 15 different sequences, which represents high sequence diversity.
Forward primer was 5 0 -AGTATACGTATTACCTGCAGC-3 0 . Reverse primer was 5 0 -CGATATCTCGGAGATCTTGC-3 0 . A poly A-labelled reverse primer (5 0 -AAAAAAAAAAAA AAAAAAAA-Spacer 18-CGATATCTCGGAGATCTTGC-3 0 ) was used together with unmodied forward primer in PCR to get the asymmetry double strand DNA and to enable further purication of ssDNA by DNA PAGE. Unmodied forward primer and reverse primer were also used for PCR amplication aer the nal round of the selection for cloning sequencing. GoTaqHot® Start Colorless Master Mix used in PCR was purchased from Promega Corporation. A uorescent labelled random ssDNA pool was also used as negative binding control.

Aptamer selection
The whole bacterial cell-SELEX process was performed according to previous reports with a few modications. 19,21,22 Briey, for the SELEX screening, specic concentrations of ssDNA pool dissolved in 100 mL binding buffer (pH 7.4, 0.1 mg mL À1 salmon sperm DNA, 1% bovine serum albumin (BSA), 50 mM Tris-HCl, 100 mM NaCl, 1 mM MgCl 2 , 5 mM KCl) were heated to 95 C for 10 min and cooled in an ice bath for 5 min to form the optimal structural conformation of oligonucleotides. The cell mixture containing 4 Â 10 8 V. vulnicus cells was incubated with the denatured ssDNA pool at 4 C for 2 h with end-over-end rotation, allowing the potential aptamer sequences to bind with cells. Aer the binding reaction, unbound ssDNA was removed by washing three times in 1 mL of binding buffer by centrifugation at 12 000 rpm for 3 min. The amounts of unbound ssDNA were measured and recovery rates were calculated. The cell pellets with adhesive ssDNA were resuspended in water, and PCR reagents were added directly to the cell suspension to amplify the tightly bound DNA sequences on the cells (more details can be found online in ESI Table 2 †). From 8 th selection round, the counter-SELEX was incorporated with the positive-SELEX to eliminate false-positive binding of sequences. In the counter-SELEX process, 2 Â 10 9 V. parahaemolyticus cells were initially incubated with 200 pmol of the denatured ssDNA library at 4 C. Aer centrifugation, the unbound oligonucleotides in supernatant were collected and subsequently incubated with V. vulnicus at 4 C. Then unbound ssDNA was removed by centrifugation, adhesive cells were resuspended, and ssDNA was amplied by PCR.
PCR products were further separated into ssDNA chains by 12% urea denaturing PAGE, and ssDNA chains were recovered from the gel band. The eluted ssDNA in binding buffer was collected and puried by Gel extraction kit (QIAEN, Germany). Finally, the puried ssDNA was quantied using a Qubit® 2.0 Fluorometer and used for the next selection round.

Cloning and sequencing of selected DNA
When the 13 th round ended, ssDNA sequences selected by whole-bacteria SELEX were amplied with unmodied primers through PCR. The puried PCR products were then cloned and sequenced by Shanghai Sangon Biological Science and Technology Company (Shanghai, China). These sequences were aligned and analyzed using the Clustal X 2.0 soware.

Flow cytometry analysis
BD FACSCalibur ow cytometer and CellQuest pro soware (Becton, Dickinson and Company, American) were used to assess the binding performance of different sequences. Fluorescence intensity of incubated bacterial cells were measured via ow cytometry.
Candidate aptamer sequences were labelled with the uorophore (5 0 -FAM). A randomized ssDNA pool with uorescent label was used as a control for nonspecic binding in each experiment. A threshold based on uorescence intensity was set so that the uorescence intensity of the gated cells would be greater than those incubated with randomized ssDNA pool. Gated cells were counted and gated uorescence intensity was quantied. Candidate sequences were dissolved in binding buffer (pH 7.4, 50 mM Tris-HCl, 100 mM NaCl, 1 mM MgCl 2 , 5 mM KCl), preheated, and then incubated with bacterial cells. Aer incubating with 300 nM aptamer/random ssDNA pool for 30 min, the cells were washed once and resuspended in 300 mL binding buffer for immediate ow cytometric analysis.
Binding curves were created to estimate dissociation constants (K D ) values by incubating different concentrations of aptamers (0-300 nM) with a xed number of cells (10 8 CFU mL À1 ).The K D were calculated with the equation y ¼ B max x/(K D + x), using SigmaPlot 12.5 soware.
To measure whether V8 and V13 can bind to V. vulnicus in all culture status, 300 nM aptamers were incubated with V. vulnicus in early exponential phase, late exponential phase, stationary phase and VBNC status (10 8 CFU mL À1 ) at 25 C. for 30 min, respectively.
To analyse whether V8 and V13 can bind to V. vulnicus isolated from different resources, 300 nM V8 and V13 were incubated with V. vulnicus MCCC 1A08743 and MCCC 1H00047 (10 8 CFU mL À1 ) at 25 C for 30 min, respectively.
To analyse aptamer stability, bacterial cells (10 8 CFU mL À1 ) were incubated with 300 nM V8 and V13 in human serum and oyster infusion. To minimize the inuence, the sediments in liquids were removed from serum and infusion by centrifugation at 12 000 rpm for 10 min. For serum incubation, serum was diluted to 30% with binding buffer. Aer incubation, the samples were directly detected by ow cytometer.
To nd out target molecular of V8, 10 min-proteinase K-treated V. vulnicus and 15 min-trypsin-treated V. vulnicus were incubated with 300 nM V8 at 25 C for 30 min. Intact V. vulnicus was used as control. All the cells are at 10 8 CFU mL À1 .
The LOB and LOD is measured according to CLSI-17A. 41 To test the LOB of V8-ow cytometer detection method, 40 blank samples (300 nM V8 solution) were tested. Kolmogorov-Smirnov test shows that P ¼ 0.022 < 0.05, indicates the data does not conform to the normal distribution. According to ISO suggestion, the calculation formula should be LOB ¼ the measured value of the [N B (p/100) + 0.5] bit, a ¼ 5%, p ¼ 95, that is, the measured value ranked 38th. 5 different low concentrations of samples were tested to measure LOD. 300 nM aptamers were incubated with V. vulnicus at the concentrations of 25 CFU mL À1 , 50 CFU mL À1 , 100 CFU mL À1 , 150 CFU mL À1 , 300 CFU mL À1 , and the volume of each sample is 12 mL. The sample is divided into 12 equal parts and the cell number in 1 mL was measured for 12 times, then the standard deviations (SD) of these tests were calculated. The formula is LoD tent ¼ LoB + c b SD S , where SD S is the estimated deviation of the population standard deviation (SD S 2 ¼ (n 1 SD s 1 2 + n 2 SD s 2 2 + n 3 SD s 3 2 + ..+ n n SD sn 2 )/(n 1 + n 2 + n 3 ... + n n ), n n ¼ N n À 1), c b ¼ 1.645/(1-1/(4 Â f)), f is the degree of the freedom, f ¼ N S À K. K is the number of groups. f ¼ 55, The linear range is determined by reference to the CLSI-EP6A document. 26 10 6 CFU mL À1 of Vibrio vulnicus suspension was incubated with 300 nM V8 and serially diluted into 9 concentrations. Suspension of each concentration was divided into 6 parts, 3 parts were subjected to ow cytometry detection, and 3 parts were subjected to plate colony counting method. Samples of different concentrations of bacterial suspension should be randomly tested and recorded. The mean number of bacteria in each concentration of the bacterial suspension was tted to the curve, and the corresponding range of the linear segment was selected as the linear range.
Fluorescence imaging of V. vulnicus to uorescently labelled DNA aptamers Fluorescently labelled 5 0 -FAM-ssDNA aptamers (300 nM) were incubated individually with V. vulnicus cells in binding buffer for 30 min at 4 C. The concentration of V. vulnicus is 2 Â 10 9 CFU mL À1 . BSA and salmon sperm DNA were added into the binding buffer to avoid nonspecic binding. Cells were washed twice to remove the unbound aptamers by centrifugation and resuspended in 15 mL binding buffer. Then, the suspension was dropped on poly-L-lysine-coated slides. Poly-L-lysine-coated slides can avoid cell moving during scanning process. Fluorescent images of bacteria with each aptamer were observed under a uorescence microscope (Olympus, Japan) using excitation at 488 nm, with 60Â magnication.

Secondary structures prediction
Secondary structures were predicted according to M. Zuker 's method. 42 Set DNA sequence to be linear, folding temperature to be 25 C, and set Ionic conditions [Na + ] ¼ 100 mM and [Mg 2+ ] ¼ 1 mM.

DNA PAGE
To test the stability, aer incubated with diluted serum and oyster infusion for 30 min, 10 mL 300 nM V8 and V13 were subject to 10% urea denaturing PAGE electrophoresis to see if there was any degradation. The electrophoresis time was 30 min. 20bp DNA Lander (Takara) was used for ssDNA length indication.

Conclusions
Dependent on whole-bacteria SELEX technology, we screened out ssDNA sequences as potential aptamers binding to V. vulnicus. Two of them as aptamers with higher sensitivity and specicity were further conrmed to have quite good performance even in truncated forms. What's more, the performance of V8 has been conrmed to not be affected by the different strains and culture environments and V. vulnicus even in VBNC status can still be effectively detected, revealing the potentials of them to be developed as biosensors detecting the existence of V. vulnicus clinically or environmentally. In the future, we will focus on the development of more efficient and sensitive biosensors or clinical diagnosis kits.
All in all, we developed V. vulnicus detection aptamers and identied two effective aptamers (V8 and V13) with both clinical and environmental detection signicance, promoting the development of related technologies and researches in the eld of bacterial aptamers. It is the rst time for aptamer to be demonstrated its ability to detect VBNC bacteria.

Conflicts of interest
There are no conicts to declare.