The smart activatable P2&3TT probe allows accurate, fast, and highly sensitive detection of Staphylococcus aureus in clinical blood culture samples

Staphylococcus aureus bacteraemia (SAB) is associated with high mortality and morbidity rates. Yet, there is currently no adequate diagnostic test for early and rapid diagnosis of SAB. Therefore, this study was aimed at exploring the potential for clinical implementation of a nuclease-activatable fluorescent probe for early diagnosis of SAB. To this end, clinical blood culture samples from patients with bloodstream infections were incubated for 1 h with the “smart” activatable P2&3TT probe, the total assay time being less than 2 h. Cleavage of this probe by the secreted S. aureus enzyme micrococcal nuclease results in emission of a readily detectable fluorescence signal. Incubation of S. aureus-positive blood culture samples with the P2&3TT probe resulted in 50-fold higher fluorescence intensity levels than incubation with culture-negative samples. Moreover, incubation of the probe with non-S. aureus-positive blood cultures yielded essentially background fluorescence intensity levels for cultures with Gram-negative bacteria, and only ~ 3.5-fold increased fluorescence intensity levels over background for cultures with non-S. aureus Gram-positive bacteria. Importantly, the measured fluorescence intensities were dose-dependent, and a positive signal was clearly detectable for S. aureus-positive blood cultures with bacterial loads as low as ~ 7,000 colony-forming units/mL. Thus, the nuclease-activatable P2&3TT probe distinguishes clinical S. aureus-positive blood cultures from non-S. aureus-positive blood cultures and culture-negative blood, accurately, rapidly and with high sensitivity. We conclude that this probe may enhance the diagnosis of SAB.


Results
Evaluation of the P2&3TT probe in blood culture samples. To test whether the P2&3TT probe allows the direct detection of S. aureus in blood cultures of SAB patients, 17 blood cultures, identified as S. aureus-positive with conventional culture-based methods, were spiked with calcium chloride, heat-treated and centrifuged. Notably, the heat treatment was previously shown to increase the detection sensitivity of MN by several orders of magnitude through the inactivation of potential MN inhibitory antibodies in human serum 4 . The supernatants were then incubated with the probe at a concentration of 3.9 µM. Indeed, probe activation Figure 1. Evaluation of the P2&3TT probe in blood cultures. Nuclease activity assays were carried out with blood culture samples from: patients with negative-culture outcomes (n = 7, green), and patients with positive blood culture outcomes for non-S. aureus pathogens including both Gram-positive bacteria (n = 17, grey) and Gram-negative bacteria (n = 15, blue), or S. aureus (n = 17, red) (probe concentration 3.9 µM). The detected pathogens are indicated. After correction for the background based on subtraction of the fluorescence value of each blood sample without the probe, fluorescence values were normalized to the average fluorescence values of the probe signal in buffer. Variations in nuclease activity in the different blood samples are indicated by dots. Of note, a S. hominis-infected blood culture co-infected with S. saprophyticus and S. epidermidis is marked with a circle. The fluorescence intensity of this sample (1.67, SD 0.16) is higher than expected based on the T/B value for S. hominis. Possibly, this sample contained also a minor undetected fraction of S. aureus, or the detected S. hominis, S. epidermidis or S. saprophyticus secreted a nuclease that can cleave the probe. To test statistical significance, the nuclease activity measured in S. aureus-positive blood culture samples was compared with that in blood cultures testing positive for Gram-positive non-S. aureus bacteria (i), Gram-negative bacteria (ii), or negative blood culture samples (iii). **** Indicates P < 0.0001, *** indicates P = 0.0006 (Kruskal-Wallis GraphPad Prism 8.0.1). www.nature.com/scientificreports/ was observed in all 17 cases as shown in Fig. 1. In contrast, incubation of the probe with culture-negative blood (n = 7) or non-S. aureus-positive blood cultures (n = 32) did not result in probe activation and, consequently, only marginal fluorescence was detected. The differences in fluorescence signals obtained upon probe incubation with S. aureus-positive blood cultures compared to S. aureus-negative blood cultures containing other Gram-positive bacteria (p = 0.0006) or Gram-negative bacteria (p < 0.0001), or compared to culture-negative blood (p < 0.0001) were statistically significant ( Fig. 1).
To calculate T/B ratios, for each investigated blood culture sample the background fluorescence of the sample without probe was subtracted from the fluorescence measured for the sample incubated with probe. Next, the resulting background-corrected fluorescence value of the sample was normalized by dividing it by the signal of the P2&3TT probe in buffer. This showed that probe incubation in S. aureus-positive blood culture samples resulted in an average normalized fluorescence intensity of 6.08 (SD 2.03), whereas this number was 0.12 (SD 0.09) for culture-negative blood culture samples. Probe incubation in non-S. aureus-positive blood cultures yielded an average normalized fluorescence intensity of 0.17 (SD 0.08) for Gram-negative bacteria and 0.42 (SD 0.41) for non-staphylococcal Gram-positive bacteria. Probe incubation with blood cultures positive for other Staphylococcus species resulted in average normalized fluorescence intensities of 0.63 (SD 0.59) for Staphylococcus hominis, 0.67 (SD 0.34) for Staphylococcus epidermidis, and 0.19 (SD 0.04) for Staphylococcus capitis (Fig. 2).

Evaluation of P2&3TT probe sensitivity for S. aureus detection in blood.
Considering the strong performance of the probe in blood cultures, we next evaluated its potential in detecting S. aureus directly in patient blood samples. An S. aureus-positive blood culture was step-wise diluted in healthy donor blood to simulate S. aureus-positive, uncultured blood. We spiked the sample with calcium chloride, heated and incubated the heat-treated sample supernatants with 3.9 µM probe. CFU counting was performed in parallel to determine the input bacterial loads. As expected, the highest fluorescence intensity was observed with undiluted S. aureuspositive blood culture, but with as much as 10 4 -fold dilution corresponding to ~ 7000 CFU/mL, a fluorescence signal that was clearly distinguishable from the control signal of the corresponding S. epidermidis-positive blood culture dilution was obtained (Fig. 3).
Lastly, to determine whether S. aureus-positive blood could be identified with lower concentrations of P2&3TT probe, we incubated two SAB blood culture samples with step-wise reduced P2&3TT probe concentrations. This resulted in step-wise reduction of the fluorescence signal, which was still clearly detectable at probe concentrations of ~ 0.5 µM (Fig. 4A). In parallel, we investigated whether a substantially lowered S. aureus load would be detectable with the P2&3TT probe at different concentrations. S. aureus-positive cultured blood was 10 4 -fold diluted with blood from healthy volunteers and incubated with different concentrations of probe. For the 10 4 -fold diluted S. aureus-positive blood sample, the probe used at the standard concentration of 3.9 µM still gave a clear fluorescence signal that could be distinguished from that of the S. epidermidis control. However, this signal was strongly enhanced by using higher probe concentrations, which was tested for a maximum final probe concentration of 39.1 µM (Fig. 4B). From these sensitivity studies which, together, involved three different

Discussion
Considering the high risk of BSIs for patients, a point-of-care test that provides reliable diagnoses within minutes would be ideal. Unfortunately, current diagnostic methods take much longer, ranging between one to seven days, or more. This delayed diagnosis results in poorer outcome for patients. The empirical broad-spectrum antimicrobial therapy initiated when BSI is suspected can exhibit side-effects and is prone to elicit antibiotic resistance 19 . In worst case, empirical antimicrobial therapy does not cover the causative micro-organism, potentially leading to death of the patient. These concerns are particularly relevant for SAB, because of the high S. aureus virulence 18,20 .
Since the previously developed P2&3TT probe may offer significantly faster SAB detection 18 , this validation study aimed at assessing its sensitivity and specificity using clinical blood culture samples.
Here we demonstrate recovery of active MN from S. aureus-positive blood cultures. We specifically detect this enzyme with the P2&3TT probe, even after 10 4 -fold dilution of cultures prior to probe incubation (corresponding to ~ 7,000 CFU/mL). Accordingly, we conclude that the detection of SAB with the P2&3TT probe is at least 10 4 -fold more sensitive than the sensitivity achieved by diagnostic culturing based on the detection of CO 2 as implemented in the BD BACTEC blood culture system (BACTEC Fluorescent Series User's Manual) 16 . Of note, this will generally apply to SAB since the P2&3TT probe, under presently applied assay conditions, is unlikely to detect BSIs caused by organisms other than S. aureus. Further, the nuclease assay takes less than 2 h and is inexpensive. Methods currently in common use require plating on agar media and incubating until colonies are formed, taking a minimum of 12-14 hours 1 .
Our data support the notion that the P2&3TT probe assay can provide the needed specificity for rapid S. aureus identification. However, we cannot exclude the possibility that rare infections (e.g. caused by uncommon coagulase-negative staphylococci such as S. argenteus, S. hyicus, S. intermedius, and S. schweitzeri) may yield false positive results since they were not encountered in our proof-of-concept study [21][22][23] . Based on its time-to-result, P2&3TT probe-based detection of S. aureus provides the means to expedite identification of S. aureus bactaeremia by at least several hours versus current methods in common use. As a standalone assay that is specific for a single bloodstream pathogen, this assay will not replace current methods which detect pathogens regardless of their species. Rather, we envision its use as a complementary tool to current methods (e.g. mass spectrometryand PCR-based identification of culture isolates) that provides more rapid identification of the most impactful bacterial bloodstream pathogen.
Lastly, in the present experimental set-up, we have sampled blood culture bottles which, for ethical reasons, was done once culture-positive signals were obtained, or when blood samples tested negative after 4-7 days of culturing. Since CO 2 production during blood culturing is detected with fluorescence tracers in some blood culture devices, it is well conceivable that next-generation MN-activatable probes can be directly implemented www.nature.com/scientificreports/ in diagnostic blood culture systems. We conclude that MN-activatable probes, like the P2&3TT probe can enable rapid and precise diagnosis of SAB.

Methods
Nuclease-activatable P2&3TT probe. The structure of the nuclease-activatable P2&3TT probe was previously described 4,18 . In short, it consists of an oligonucleotide comprising 2′-O-methyl modified uridines flanking a pair of unmodified deoxythymidines, which is coupled to fluorescein amidite on the 5′-end and the ZEN www.nature.com/scientificreports/ and Iowa Black RQ quenchers on the 3′-end. For the present study, it was synthesized and purified by Integrated DNA Technologies, Inc. (IDT; Coralville, IA). The lyophilized probe was dissolved in 10 mM Tris-HCl pH 8.0, 1 mM EDTA to a final concentration of 391.6 μM and stored at -80 °C. For the assay, 1 µL of probe was diluted in 9 µL of 10 mM Tris-HCl pH 9.0, 10 mM CaCl 2 to yield a working stock of 39.16 µM 18 .

Blood culture samples. Clinical blood cultures generated in BD Bactec bottles (Becton, Dickinson and
Company, USA) were collected from the department of Medical Microbiology (University Medical Center Groningen [UMCG]) after one to two weeks of culturing and storage. All blood culture samples were incubated until marked positive by a BD BACTEC FX Blood Culture System, or until unloaded as culture-negative after 4-7 days. Each blood sample was sub-cultured in parallel according to standard procedures to verify the microbiological diagnosis by plating and MALDI-TOF mass spectrometry.
Blood samples from healthy volunteers. Blood from healthy volunteers was collected in BD Vacutainer Heparin tubes to perform sensitivity assays on the day of the experiment. These blood samples were also used to prepare dilutions of infected-blood cultures to assess the sensitivity of the P2&3TT probe.
Determination of bacterial titres. To determine CFUs, serially diluted blood cultures were plated on blood agar media immediately prior to carrying out the nuclease assays, and the respective plates were incubated overnight at 37 °C.
Detection of micrococcal nuclease activity in blood. Blood  Statistical analyses. In each experiment, unless stated otherwise, mean fluorescence levels of background controls (e.g., blood without the P2&3TT probe) were subtracted from the signal of the measured condition and divided by the signal of the P2&3TT probe in buffer to calculate a target-to-background (T/B) signal. The blood was divided into four equal volumes, 1 mL each, of which two were incubated with the probe and two were incubated in parallel without probe. All fluorescence measurements were made in triplicate. All p-values were calculated using a one-way ANOVA in GraphPad Prism 8.1.0. P-values < 0.05 were considered significant.
Ethical approval. Permission for this study was obtained via the Medical Ethical Review Board Committee of the UMCG (permission number METc2017/098). The study was performed with adherence to the guidelines of the Declaration of Helsinki and local regulations, and all patient samples were treated pseudo-anonymously. Of note, during the present study period, patients admitted to the UMCG complied with hospital guidelines in an opt-out research consent procedure. Consequently, individual written consent was not required for inclusion of their diagnostic waste materials in our study, as consent was given through this opt-out practise unless stated otherwise in the medical file. Blood donations from healthy volunteers were collected with approval of the medical ethics committee of the UMCG (approval no. METc 2016/621). All blood donations were obtained after written informed consent from all volunteers, in accordance to the Helsinki Guidelines and local regulations, and all the samples were anonymized.