Rapid ultrasensitive diagnosis of pneumonia caused by Acinetobacter baumannii using a combination of enrichment and phage-based qPCR assay CURRENT STATUS: POSTED

Background Accurate and rapid identification of associated or nosocomial pneumonia caused by ( could improve the treatment. Methods In current study, we developed a phage-based real-time quantitative PCR (qPCR) combined with enrichment culture for rapid and specific detection of viable A. baumannii in sputum from lung infections. Through short-term plate incubation, bacteria can be enriched and the DNA polymerase reaction disturbance of sputum can be decreased greatly. This approach is based on detecting phage replication in live A. baumannii cells through Taqman qPCR. Results Through the built detection system, down to 1 CFU of A. baumannii can be detected within 6 h in spiked sputum samples without any steps of bacteria isolation and DNA extraction. The established method was then applied to detecting both A. baumannii in simulated sputum with 100% agreement with the spiked amount of the bacteria and one clinical sputum sample from an 80-year-old male lung infection patient caused by A. baumannii with perfect accuracy, demonstrating that the assay developed in this study has the merits of high rapidity, high sensitivity, good specificity and being able to detect live bacteria not dead bacteria. Conclusions assay is a potentially clinical method diagnosis of bacterial pneumonia infection caused baumannii or other bacterial infection in sputum or complicated samples through switching to types phages. 2.5 U/ µL, expressed and purified in the laboratory), 1.6 µL dNTP (2.5 mM for each, from Takara Biomedical Technology (Beijing) Co., Ltd, China), milliQ water 5.9 µL, 2 µL 10xbuffer (200 mM Tris-HCl pH 8.3, 200 mM KCl, 100 mM (NH 4 ) 2 SO 4 , 20 mM MgSO 4 and 5% NONIDET P-40 SUBSTITUTE from AMRESCO Inc., Radnor, PA, USA). Each qPCR reaction system was made of 10 µL 2xqPCR reaction mixture, 0.4 μM of the probe, 0.4 µM of the primers, 2 µL of the template and water added to total volume of 20 µL, for the detection of p53. The qPCR amplification was performed on CFX96, Bio-Rad Laboratories, Hercules, CA, USA. The procedure was set as: 95 °C for 3 min, 40 cycles of 95 °C for 5 s and 60 °C for 1 min, and the fluorescent signal was acquired at the step of 60 °C.


Background
Pneumonia and lung infections caused by multidrug resistant bacteria (MDR) such as Acinetobacter baumannii (A. baumannii), is the most common healthcare-acquired infections among the critical illness with significant morbidity and mortality. [1,2] Without appropriate initial antibiotic therapy, the mortality rate of community-acquired pneumonia caused by A. baumannii has been reported to be as high as 64%. [1] Accurate and rapid diagnosis of bacterial pathogens in pneumonia is pivotal for the guidance of medication and reducing mortality.
Diagnosis of the bacteria in lung infections is traditionally based on cultures and serology. [3][4][5] The methods of culture depend on culture and isolation of target bacteria and are the current ''gold 3 standard'' for the diagnoses and identifications of bacterial pathogens. While, the culture method is time-consuming, usually offering delayed results and seriously retarding the clinically choosing medicine and effective and timely treatments. [6] Serology based methods require a second convalescent-phase sample, which limits the clinical usefulness of these techniques. [6] Furthermore, the sensitivity of cultures is poor and false negative rate is usually high. Nucleic acid amplification technique are another common-used means, including conventional PCR and real-time quantitative PCR (qPCR), which usually amplify some specific genes of bacteria and are considered faster, more sensitive, and more specific than traditional cultures and serology. [7][8][9] Nevertheless, due to the high sensitivity and easy-to-contaminate of PCR of bacteria based techniques, the false positive results tend to happen. [10] Especially, the PCR of bacteria based methodologies cannot discriminate viable bacteria from dead bacteria as well as are inhibited by interference factor that existing in sputum specimen from pneumonia infection. [11,12] Bacteriophages or phages are viruses that specifically infect host bacteria. After a lytic phage binds to and infects its host, a single bacterial cell can produce hundreds of progeny phages in a few minutes.
Through facile procedures, it is easy to isolate from water or soil samples in the environment and obtain a large amount by co-culturing with the host. Bacteriophages have been developed for various assays, such as the detection of Mycobacterium tuberculosis in sputum. [13][14][15][16][17] Conventional methodologies based on phage for the recognition of bacteria are focused on counting the formed plaques by the double-layer plate culture after the infection of phage to the bacteria, where it takes two to three days to get the results. [18] In recent decades, increasing methods relying on the high affinity of the bacteriophages to the host bacteria have been developed for the detection of pathogenic bacteria. [19][20][21][22] In addition, the released components after lysis by phages can be determined to develop assays for the detection of bacteria, for example, β-galactosidase or progeny phages. [23,24] Furthermore, A SYBR green qPCR had been built for phage-based detections of bacteria, determine 1 CFU/mL within 72 h, which cannot satisfy the instant demand clinically. [25] The samples involved in these above mentioned methodologies mostly contained simple matrix such as milk, which is in favor for culturing the host bacteria and thus contributes for the detection through 4 phages, while in the complex samples such as serum or sputum, phages might not infect the host well, so it is a challenge to develop a phage-based methodology, especially in the samples of sputum.
Recently, we reported a phage-based qPCR methodology for sensitive diagnosis of bloodstream infection, where a platform was built to detect A. baumannii (down to 10 CFU in 100 µL serum) in serums within 4 h without bacteria isolation and DNA extraction. [26] However, sputum samples are more complex than serum samples and sputum disturbs the PCR reaction seriously so as to be not able to detect targets in sputum directly by qPCR.
To develop a method to detect bacterial pathogens in Pneumonia sensitively, rapidly and accurately, in this study we exploited a phage-based method combining quantitative PCR (qPCR) with a shortterm plate culture for the rapid, sensitive and noninvasive detection of viable A. baumannii in sputum samples from a patient with lung infection. As depicted in Fig. 1., after the sputum samples spread on culture plates were incubated for 2 h, the bacteria on the plates were collected for co-culturing with phage for 3 h, and then the mixture of the phage and the bacteria was subject to qPCR detection of the phage. The Ct changes between the time point 0 h and 3 indicated whether there were A. baumannii or not in sputum samples. Through the system, down to 1 CFU in 100 µL sputum samples was able to been detected by the established detection system within 6 h. Through spiking A. baumannii and several other common bacterial pathogens in sputum, the specificity of the system has been proved. Moreover, the established system has been applied to a clinical sputum sample to validate the system clinically and a satisfactory result has been obtained. The developed methodology here is potentially used to diagnose pneumonia caused by A. baumannii from sputum samples directly.

Bacterial strains and culture condition
All bacterial strains used here are listed in tab. 1. The phage p53 was isolated previously, with a host of A. baumannii strain LB8, an isolated strain from a clinical pulmonary effusion sample. All strains in this study were routinely streaked onto a Luria Bertani (LB) plate. And then the plates were incubated overnight at 37 °C. After picking a single colony on the plate into 5 mL LB broth, the bacteria were 5 grown overnight at 37 °C and 200 rpm. Then the bacteria were spread onto LB agar plates for 12 h and collected by cotton swabs. Bacteria were released by stirring cotton swabs in 1 mL LB broth. Then the bacterial suspension was centrifuged at 6000 rpm for 5 min. After discarding the supernatant, the bacterial precipitation was re-suspended in 1 mL fresh LB broth. The OD 600 of the bacteria solutions was adjusted to 0.45, which is corresponding to about 6x10 8 CFU/mL of the bacteria, which had been calculated by plate dilution as shown in Fig. 2 before use. After the determination of the concentration of the bacteria in CFU/mL, serial dilutions of the bacteria were adopted for the experiments. Unless specified, all strains listed in tab. 1 were treated by the same protocol before use.
Tab. 1. The information of the bacteria used in this study 7 lab. [26] The sequence of p53 was determined in the lab previously. The preparation of phages p53 and determination of phage titers are based on our previous publication [26]. Briefly, the purification and PFU determination of the phage were realized by double-layer plate techniques. Through put 100 µL purified phage one a double layer plate with LB8 lawn in a soft-agar layer on a regular LB agar layer and incubated overnight at 37 °C, the soft-agar layer was collected in 5 mL phage buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 , 2 mM CaCl 2 .) Tris base was bought from Sigma-Aldrich, Co, St. Louis, MO, USA. After filtered by sterile filter, the phage solution was obtained. Single clear plaques of a serial dilution of the phage solution were used for determination the concentration of a phage solution in PFU/mL. The PFU-determined phage solution was stored at 4 °C for further use.

qPCR assay
The primers and probes were designed for specifically detecting p53, as described previously. [26] The sequences of the primers and the probe were: forward, 5'-CGGATGTGGCAATATTAC-3'; reverse, After incubation at 37 °C for 8 h, the bacteria on LB agar plates were collected by cotton swabs in 1 mL LB broth. Through serial dilution of the bacteria, different concentrations of LB8 (10 0 , 10 1 , 10 2 , 10 3 , 10 4 , 10 5 , and 10 6 CFU/mL) in LB broth were obtained according to the description above. 500 µL LB8 in LB broth were mixed with 500 µL of p53 at the concentrations of 10 2 , 10 3 , 10 4 , 10 5 , 10 6 PFU/mL, respectively. 50 µL of the mixture of LB8 and p53 was immediately taken out and stored at -20 °C, which was set as 0 min. Next, the mixture of LB8 and p53 were incubated at 37 °C with shaking at 200 rpm. Then 50 µL of the mixture was taken out every 50 min. All the samples were stored at -20 °C immediately after taken out for the qPCR detection of p53 after sampling was completed.

Optimization of the plate culture time for the detection of A. baumannii
The culture time of LB8 spiked sputum samples on LB agar plates was optimized for the detection platform. Before spiking the bacteria, the pure sputum samples were investigated by culture and plaque assays for the absence of either LB8 or p53. The bacteria of serial 10-fold dilutions of LB8, was mixed with the pure sputum to the final concentrations of LB8 with of 10 0 , 10 1 , 10 2 , 10 3 , 10 4 , 10 5 and 10 6 CFU/mL. Then 100 µL of LB8 spiked sputum was transferred onto LB agar plates containing three glass beads and spread by moving the glass beads on the plates. The plates were incubated at 37 °C.
And one plate was taken out at the time points of 0 h, 2 h, 4 h, 6 h, 8 h and 10 h, respectively.
Substances including bacteria on the plates were collected by cotton swabs and released by stirring cotton swabs in 1 mL LB broth. Then the bacterial suspension was centrifuged at 6000 rpm for 5 min, discard the supernatant and bacterial precipitation was re-suspended in 1 mL fresh LB broth.
Afterward, 500 µL bacteria in LB broth were mixed with 500 µL 10 3 to 10 4 PFU/mL p53 in phage buffer. The other 500 µL bacteria were for qPCR detection of gltA gene of A. baumannii. After the mixture was cultured for 180 min, 50 µL of the mixture was taken out for qPCR detection of p53. The experiments have been run for three times.

Results
Before exhibiting the results in this study, several results and conclusions should be shared from our previous publication. [26] 1. As determined by regular PCR and the agarose gel analysis in the publication, the set of the primers/probe was specific to p53 without any interference from different strains of A. baumannii as well as other four common bacteria (E. coli, S. aureus, P. aeruginosa, S. pyogene). 2. The burst time for p53 was determined as 50 min by qPCR. As well, the phage genome was released well through initial heating of 95 °C for 3 min from the amplification efficiency. 3. The detection limit of p53 in phage buffer was 10 2 PFU/mL, so for the following optimizations of the p53 concentration for the detection of A. baumannii, the concentration of phage p53 started from 10 2 PFU/mL. However, sputum has more profound inhibition to qPCR than serum, which is the matrix in the previous study. [26] We could not get any Ct values for any concentration of p53 in sputum by direct qPCR. Therefore, we changed the strategy to decrease the interference of sputum through short culture on LB agar plates. Therefore, here we optimized the phage concentration by the new strategy.
Optimization of phage concentration and incubation time of p53 with LB8 for the detection of A.
baumannii Through optimization of phage concentration, the sensitivity and the concentration range of the detection for the host bacterial pathogen can be improved. The bacteria were scraped from culture plates, where the bacteria were different from the ones cultured in broth. The optimization of the concentration of p53 was started from 10 2 PFU/mL since the lowest concentration of 10 2 PFU/mL can be detected. [26] Serial 10-fold from 10 2 to 10 6 PFU/mL of the concentration of p53 (totally 5 concentrations of 10 2 , 10 3 , 10 4 , 10 5 , 10 6 ) were adopted. And at each concentration of p53, the host concentrations were variated from 10 0 to 10 6 CFU/mL, including all possible CFUs of the bacteria in sputum samples. The sample were collected every 50 min for the detection of qPCR until 250 min and Ct values were recorded. The plots of Ct values vs. the culture time of the mixture of p53 and LB8 in LB broth were illustrated in Fig. 3. When the concentration of p53 was higher than 10 4 PFU/mL, an increasing detection range of the bacteria concentration and sensitivity of the bacteria can be obtained as shown in Fig. 3C-E, while at low concentration of p53 (10 2 PFU/mL) the sensitivity became low, as indicated in Fig. 3A. Therefore, when the concentration of phage p53 was 10 3 to 10 4 PFU/mL as revealed in Fig. 3B and 3C, the sensitivity was the highest and the detected range of the bacteria is the broadest, where 100 CFU/mL of A. baumannii was detected at the time point of 100 min. Therefore 10 3 to 10 4 PFU/mL of phage p53 was selected for the following detection of bacteria in sputum samples. In addition, longer incubation time will lead to higher sensitivity. For example, at the culture time of 200 min, the detection limit arrived at 10 CFU/mL with obvious signal (Ct changes) as shown in Fig. 3B Fig. 4A) 10 1 CFU/ mL A. baumannii was able to be detected easily, where ΔCt is much higher than the cut-off value 0.5. At the culture time of more than 4 h, 10 1 CFU/mL of the host bacteria can be also detected, but less ΔCt values were obtained for the detection of different concentrations of the bacteria than those at 2 h or 4 h. When running the experiments, we found that at long culture time, the things on the plate became sticky, probably resulted from other bacteria or easy-to-grow organisms, which might inhibit the infection of phage to the bacteria. This might be the reason why at long culture time, less difference of phage numbers arose. Furthermore, less time is better for rapid detection. Therefore, 2 h was used for the following detection of bacteria in sputum samples.
At the same time, the established method was compare with the traditional qPCR method for amplifying gltA gene of A. baumannii. The bacteria were scraped from the plates applied with sputum samples and subjected to DNA extraction by kit, followed by qPCR for the bacteria. Here a commercial kit for extracting genome DNA of A. baumannii, because the substances scraped from the culture plates still interfere directly detecting the specific gene of A. baumannii by qPCR. The corresponding results are revealed in Fig. 4B. At the culture time of 2 h, the detection limit is about 10 4 CFU/mL. For detecting gltA gene of A. baumannii, the limit of detection could reach 10 1 CFU/mL but culture time was about 6 h, which takes total about 9 h (6 h for plate culture and 3 h for DNA extraction and qPCR detection) for the whole detection from the sputum to final results.

Assay performance with simulated clinical sputum samples
LB8 in Logarithmic-phase was spiked to the sputum to the final bacteria concentrations of 10 0 CFU/ mL to 10 6 CFU/ mL. The p53 concentration of 10 3 to 10 4 PFU/mL was used as optimized previously. consistent results with the traditional culture method. By the developed method, down to 10 1 CFU/mL of A. baumannii can be detected well in 6 h with a high sensitivity. Here 100 µL sputum was taken for the detection, which means 1 CFU can be probed by the current methodology. +++, ++ and + mean the plaque of the phage on the tested bacteria is very clear, clear and faint of the plaques, respectively. ---means no lysis plaque.

The performance of the method in diagnosing an 80-year-old male lung infection patient
To evaluate the performance of this method, the developed method as well as a regular culture method has been used for detecting A. baumannii in a real clinical sputum sample from a patient. The regular culture method was realized through isolating the bacteria from the sputum sample and identifying the bacteria through 16S rDNA PCR and sequencing. As shown in Fig. 5., the results showed that the method could effectively detect bacteria in a clinical sample, the result of which was consistent with the traditional methods (the data of the traditional methods were not shown here). Compared with the method based on the conventional microbiological culture, the specificity of the developed method was 100%. The detection limit of this method was 1 CFU in 100 µL sputum.

Discussion
Otherwise, the method can be applied for antibiotics-resistant detection through directly culturing the sputum samples with antibiotics for 1 or 2 hours, [28,29] and then the amount of the bacteria will be different for drug-sensitive and drug-resistant bacteria. Otherwise, there cases of pulmonary infection based on phage therapy are increasing [30,31] and this method provides ideas for phage screening and monitoring in phage therapy.
However, there are still some questions to resolve for the developed method for clinical application. A common problem based on phage for detecting host bacteria is the host spectra of the phages. Up to now, there have been no phages discovered infecting all strains of a species, which might be solved by engineered phage. Therefore, the developed method in this study can only detect p53-sensitive A.
baumannii strains, not all A. baumannii. Circumventing this problem, we can combine a phage cocktail and multiplex-qPCR to detect more bacterial strains and species in sputum samples. Owing to the resource limitation, the real clinical patient specimens were small and we hope to evaluate the clinical application prospect of this method more accurately through more actual patient samples in future.

Conclusion
Here, a phage-based qPCR methodology combined with enrichment culture through a short-term The correlation between bacteria concentration logarithm of A. baumannii in CFU/mL and  The proposed system was applied to detect A. baumannii in a sputum specimen from a clinical patient.