Broad-spectrum capture of clinical pathogens using engineered Fc-mannose-binding lectin enhanced by antibiotic treatment

Background: Fc-mannose-binding lectin (FcMBL), an engineered version of the blood opsonin MBL that contains the carbohydrate recognition domain (CRD) and flexible neck regions of MBL fused to the Fc portion of human IgG1, has been shown to bind various microbes and pathogen-associated molecular patterns (PAMPs). FcMBL has also been used to create an enzyme-linked lectin sorbent assay (ELLecSA) for use as a rapid (<1 h) diagnostic of bloodstream infections. Methods: Here we extended this work by using the ELLecSA to test FcMBL’s ability to bind to more than 190 different isolates from over 95 different pathogen species. Results: FcMBL bound to 85% of the isolates and 97 of the 112 (87%) different pathogen species tested, including bacteria, fungi, viral antigens and parasites. FcMBL also bound to PAMPs including, lipopolysaccharide endotoxin (LPS) and lipoteichoic acid (LTA) from Gram-negative and Gram-positive bacteria, as well as lipoarabinomannan (LAM) and phosphatidylinositol mannoside 6 (PIM 6) from Mycobacterium tuberculosis. Conclusions: The efficiency of pathogen detection and variation between binding of different strains of the same species could be improved by treating the bacteria with antibiotics, or mechanical disruption using a bead mill, prior to FcMBL capture to reveal previously concealed binding sites within the bacterial cell wall. As FcMBL can bind to pathogens and PAMPs in urine as well as blood, its broad-binding capability could be leveraged to develop a variety of clinically relevant technologies, including infectious disease diagnostics, therapeutics, and vaccines.


Introduction
Mannose-binding lectin (MBL) is a key host-defense protein associated with the lectin pathway of the innate immune system 1 , and deficiency of MBL can lead to increased susceptibility to a wide-spectrum of infectious diseases 2-4 . MBL functions as a calcium-dependent, pattern-recognition opsonin that binds a range of carbohydrate molecules associated with the surfaces or cell walls of many different types of pathogens 5 . Collectively these microbial surface carbohydrate molecules, including for example, lipopolysaccharide endotoxin (LPS) and lipoteichoic acid (LTA), are referred to as pathogen-associated molecular patterns (PAMPs) 6,7 . MBL has the intrinsic ability to distinguish foreign PAMPs from self, subsequently activating the complement system and providing protection via antibodydependent and independent mechanisms 8,9 . Due to the evolutionarily conserved recognition carbohydrate moieties of PAMPs, MBL is a broad-spectrum opsonin that can bind over 90 different species of pathogens, including Gram-negative and Gram-positive bacteria, fungi, viruses, and parasites [10][11][12][13][14] . MBL binding to these various pathogens has been demonstrated by means of flow cytometry 14,15 , radio-immunoassay 13,16 , enzyme-linked immunosorbent assay (ELISA) 13,17 , immunofluorescence and scanning electron microscopy (SEM) 18 , and Saccharomyces cerevisiae-induced MBL activation and bystander lysis of chicken erythrocytes 19 . However, many discrepancies in MBL binding have been described, depending on the method used. For example, use of flow cytometry revealed little to no MBL binding to Pseudomonas aeruginosa, while others have reported good binding of MBL to Pseudomonas aeruginosa using a hemolytic assay 15,19 .
We set out to address these conflicting results by leveraging the recent development of an engineered version of MBL that contains the carbohydrate recognition domain (CRD) and flexible neck regions of MBL fused to the Fc portion of human IgG1, which is known as FcMBL 20 . The engineered FcMBL lacks the regions of the native molecule that interact with MBL-associated serine proteases (MASPs) that activate complement and promote blood coagulation, and thus, it can be used to capture PAMPs from complex biological fluids, such as blood and urine, without activating effector functions of complement, coagulation, and phagocytosis. We have previously used FcMBL in extracorporeal therapies, such as hemofiltration, and in diagnostics to capture and detect Staphylococcus aureus from osteoarticular and synovial fluids of infected patients [20][21][22] . In the present study, we used a previously described sandwich enzyme-linked lectin sorbent assay (ELLecSA) in which both live and fragmented pathogens (PAMPs) are captured using FcMBL conjugated to magnetic beads and then detected with horseradish peroxidase (HRP)-labeled MBL 23 . This ELLecSA has demonstrated FcMBL binding to 85% (47 of 55) of pathogen species previously tested, and enabled rapid diagnosis of bloodstream infections by capturing and detecting PAMPs in whole blood from human patients 23 . Here we extend this testing to include 69 isolates from 57 more pathogen species using the ELLecSA. In total, we measure direct binding of FcMBL to over 190 pathogen isolates from over 95 different pathogen species, including bacteria, fungi, viral antigens, parasites, and bacterial cell wall molecules. As a result of this more extensive analysis, we demonstrate that FcMBL detection increases to 85% of the isolates and 87% (97 out of 112) of the pathogen species tested. Furthermore, we show that antibiotic treatment or mechanical disruption of the bacterial pathogens exposes previously concealed FcMBL binding sites on cell walls, thereby increasing the efficiency of pathogen detection and reducing variation between binding of different strains of the same species. We also show that FcMBL can detect PAMPs in urine as well as blood, making this potential diagnostic technology highly synergistic with standard of care antibiotic therapy.

FcMBL binding to bacteria
We first set out to determine the range of pathogens that FcMBL can capture by screening multiple species of bacteria, fungi, viral antigens, and parasites using the ELLecSA detection technology. In the FcMBL ELLecSA, pathogen materials in experimental samples are captured with FcMBL immobilized on superparamagnetic beads (1 µm diameter), magnetically separated, washed, detected with human MBL linked to horseradish peroxidase (HRP), magnetically separated again, washed, and then tetramethylbenzidine (TMB) substrate is added to measure the amount of pathogen material bound ( Figure 1A). Results were quantified by interpolating against an internal standard curve generated using yeast mannan in buffer [50 mM Tris-HCl, 150 mM NaCl, 0.05% Tween-20, 5 mM CaCl 2 , pH 7.4 (TBST 5 mM CaCl 2 )] ( Figure 1B). In addition, we show that while the curve sensitivity is reduced when mannan is spiked into whole human blood, the limit of detection remained similar (1 ng/ml) as it does in buffer ( Figure 1B).
Initially our focus was on screening bacteria and as such we compiled a comprehensive list of clinically relevant bacterial pathogens (Table 1) 23 . When we screened 82 different species of bacteria to compare FcMBL binding to live versus fragmented cells, we found that FcMBL detected 59 out of 82 live microbes (72%) and that more could be detected (70 out of 82; 85%) after they were fragmented by treating with antibiotics, or mechanically disrupted using zirconia/silica beads in a mixer mill (Table 1) 23 . The antibiotics we used in this study were clinical grade cefepime, ceftriaxone, meropenem, amikacin, gentamicin, and vancomycin, to provide enough coverage to target this diverse range of bacteria. We dosed each bacterial class with a single appropriate antibiotic dose (≤1 mg/ml) to obtain acute fragmentation within 4 hours.
To determine if inducing bacterial fragmentation via antibiotic treatment or bead milling would reduce variation in FcMBL binding between different strains of the same species, we screened 134 isolates from 21 of the 88 Gram-positive and Gram-negative bacterial species, including strains tested by Cartwright et al. (Figure 2) 23 . As before, FcMBL bound a greater proportion of the pathogens when fragmented (113/134 = 84%) than when live and intact (77/134 = 57%). For some bacterial species such as Enterobacter cloacae, Escherichia coli, Klebsiella oxytoca, and Klebsiella pneumoniae we found that mechanical disruption Live and fragmented bacteria are captured by FcMBL, which has been coated on superparamagnetic beads via an N-terminal aminooxy-biotin on the Fc to allow oriented attachment to streptavidin (FcMBL Bead). FcMBL beads with captured bacteria are then magnetically separated and detected using recombinant human MBL linked to horseradish peroxidase (Human MBL-HRP). Tetramethylbenzidine (TMB) substrate is added to quantify captured bacteria and results are read at OD 450nm. (B) Mannan standard curve showing FcMBL binding to mannan in buffer and spiked into whole human blood (<24 h from draw), measured at optical density 450 nm. The mannan standard curve in buffer serves as an internal assay control, and is used to determine FcMBL binding to pathogen material in units of PAMPs/ml (where 1 ng/ml of mannan bound by FcMBL is equivalent to 1 PAMP unit. PAMP units are then multiplied by the dilution factor of the sample volume to give PAMPs/ml).
or antibiotic-induced fragmentation greatly increased FcMBL binding, whereas other bacteria like Pseudomonas aeruginosa, Yersinia pseudotuberculosis, and MRSA bound equally well when live and intact ( Figure 2). With the exception of Proteus mirabilis and Enterococcus faecalis, which FcMBL did not bind at all, the capture of fragmented bacteria was equal to or greater than that of live bacteria (Figure 2).
These findings are consistent with past studies that showed the efficiency of MBL binding to live bacteria differs between isolates from the same bacterial genus and species, possibly due to differences in encapsulation 15,16 . To illustrate that the heterogeneity of MBL binding to live isolates of the same species can be reduced by using antibiotics to disrupt previously cryptic binding sites, we used two different isolates of Escherichia coli, and two different isolates of Streptococcus pneumoniae. E. coli 41949 and S. pneumoniae 3 exhibited equivalent FcMBL binding whether they were live or fragmented with antibiotics (1 mg/ml cefepime or ceftriaxone, respectively, for 4 hours), whereas fragmented forms of E. coli RS218 and S. pneumoniae 19A isolates bound much more effectively to FcMBL than living forms ( Figure 3A-D). This difference was further supported visually using scanning electron microscopy (SEM) in which magnetic FcMBL beads could be seen to bind both live and fragmented versions of E. coli 41949 and S. pneumoniae 3, but with E. coli RS218 and S. pneumoniae 19A, the FcMBL beads only bound to fragmented material ( Figure 3E-H). FcMBL binding increases upon fragmentation, which is correlated with LPS release measured using a limulus amebocyte lysate (LAL) assay: equal amounts of LPS were detected for E. coli 41949 whether live or fragmented, whereas LPS levels were higher in antibiotic treated E. coli RS218 ( Figure 3I, J). These results suggest that antibiotic treatment results in exposure of previously cryptic PAMPs in the cell wall, including toxins such as LPS and LTA, which leads to greatly increased binding of FcMBL.
In these studies, we found that 12 bacterial species, including multiple species of Enterococcus and Proteus, failed to bind to  FcMBL even when treated for 4 hours with combinations of antibiotics (500 µg/ml vancomycin and 500 µg/ml amikacin for Gram-positive isolates or 500 µg/ml cefepime and 500 µg/ml amikacin for Gram-negative isolates) ( Table 1) 23 . Importantly however, FcMBL was able to detect 84% (172 out of 204) of the bacterial isolates (Figure 4), which includes 9 of the 10 pathogens responsible for most healthcare-associated infections in acute care hospitals in the U.S., with Enterococcus species being the one exception 24 .

FcMBL binding to bacterial cell wall components
We further explored FcMBL's ability to bind bacterial cell wall components because when bead mill or antibiotics were used to disrupt the membranes of Gram-negative isolates (n = 117), there was a significant boost in FcMBL detection efficiency with fragmented cells (89%) versus live intact cells (58%) (Figure 4), which is likely due to exposure of LPS that is present in high concentrations in their cell wall 25 . Similarly, FcMBL also detected a greater percentage (77%) of fragmented Gram-positive isolates (n = 78), versus live intact cells (65%) (Figure 4). Thus, to better understand some of the major targets that FcMBL binds when bacteria are fragmented, we extended our analysis using purified samples of LPS and LTA 25,26 .
Using the ELLecSA, we screened LPS purified from Gramnegative bacteria (Serratia marcescens, Klebsiella pneumoniae, and Salmonella enterica serovar enteritidis), as well as LTA from Gram-positive bacteria (Enterococcus hirae, Staphylococcus aureus, and Streptococcus pyogenes) in TBST 5 mM CaCl 2 , as well as in more clinically relevant human whole blood samples. We found that FcMBL was able to detect LPS from all 3 Gramnegative species, with S. marcescens being the best (1 ng/ml limit of detection in buffer and 3.9 ng/ml in blood) ( Figure 5A-C).
FcMBL also bound to E. hirae LTA very well (15.6 ng/ml limit of detect in buffer and 62.5 ng/ml in blood) ( Figure 5D), which is consistent with past findings 27 . Also consistent with past findings, we found that FcMBL binds S. aureus LTA through the carbohydrate recognition domain ( Figure 5E and Supplementary Figure 1) 28,29 . Notably, however, FcMBL also bound S. pyogenes LTA ( Figure 5F), which is in contrast to the past finding that MBL binds well to E. hirae LTA but poorly to LTA from S. pyogenes due to lack of glycosyl substituents 27,30 .
We next tested FcMBL's ability to bind lipoarabinomannan (LAM) and its biosynthetic precursors, phosphatidylinositol mannoside 1 & 2 and 6 (PIM 1,2 and PIM 6 ) from Mycobacterium tuberculosis (TB) strain H37Rv 31,32 . LAM released from metabolically replicating or degrading TB bacteria has been detected in both blood and urine 33,34 . Thus, we assessed the ability of FcMBL to capture and detect LAM, as well as PIM 1,2 and PIM 6 , spiked into both of these complex biological fluids as well as buffer. Our initial screen in buffer confirmed that FcMBL can detect LAM and PIM 6 at levels down to 4 ng/ml, but it did not detect PIM 1,2 ( Figure 6A-C). FcMBL also bound to LAM in both blood and urine but its binding sensitivity was reduced as it could only detect 15.6 ng/ml and 250 ng/ml, respectively. FcMBL binding to PIM 6 exhibited a similar sensitivity in buffer, but it could only detect 62.5 ng/ml and 15.6 ng/ml in blood and urine, respectively.    FcMBL binding to fungi, parasites, viral antigens, and bacterial cell wall antigens In addition to screening multiple bacteria, we also tested FcMBL's ability to bind to 12 different species of fungi, 10 viral antigens, 2 species of parasites, and 6 purified bacterial cell wall antigens (Table 1) 23 . In contrast to studies with bacteria, FcMBL was found to bind 100% of live fungal cells from all 12 species and 13 isolates tested (Figure 4). Of the two parasites tested in this preliminary analysis, only Trichomonas vaginalis was bound by FcMBL, whereas 80% of the viral antigens screened and 92% of the purified bacterial cell wall antigens were detected (Figure 4 and Figure 7). The handful of pathogen material FcMBL did not detect included the E1 protein from Chikungunya virus, the NS1 protein from Tick-borne encephalitis virus, Plasmodium falciparum, and PIM 1,2 from TB. In total, the overall FcMBL binding profile respectively detected 85% (194 out of 229) of isolates, and 87% (97 out of 112) of the different pathogen species tested.
Raw data for the present study are available on OSF 29 .

Discussion
MBL has been reported to bind to over 90 different pathogen species as well as PAMPs released from these microbes based on studies in which binding was assessed by means of flow cytometry, ELISA, radio-immunoassay, immunofluorescence and SEM, or hemolytic assays 12-19 ; however, different results have been obtained with different methods. Here we explored the broad-spectrum binding capabilities of an engineered form of MBL, known as FcMBL, using a previously described magnetic ELLecSA detection assay to quantify binding of MBL to over 190 isolates from over 95 different pathogen species, which include bacteria, fungi, viral antigens, parasites, and bacterial cell wall antigens 23 . FcMBL was previously shown to bind PAMPs released from 47 of 55 (85%) pathogen species tested, including 38 species of bacteria and 9 species of fungi 23 . The FcMBL ELLecSA also was able to detect infectious PAMPs in whole blood of sepsis patients, regardless of antibiotic therapy (blood culture positive or negative) with a detection sensitivity and specificity of 85% and 89%, respectively 23 . In the present study, we utilized the ELLecSA to extend this testing to include 69 isolates from 57 more pathogen species. In total, our results confirm that FcMBL binds to 85% of the isolates and 97 of the 112 species tested, which corresponds to an increased detection sensitivity of 87%.
MBL binding to different clinical bacterial isolates of the same species has previously produced conflicting results 15,16 . These same studies also described that most Gram-negative isolates (encapsulated strains) bound little or no MBL. We have reported similar results as we previously found that FcMBL only bound 38% of live clinical E. coli isolates tested; however, upon fragmentation and release of PAMPs, FcMBL detection of these same isolates increased to 92% 23 . Broader examination of Gramnegative bacteria in the present study revealed a similar pattern: FcMBL only detected 68/117 (58%) of live isolates, but when these same microbes were treated with antibiotics or bead mill, the detection sensitivity significantly increased to 89% (104/117 isolates). Apparently, by treating bacteria with antibiotics or bead mill we were able to disrupt the encapsulated cell wall, exposing and presenting previously hidden PAMPs, thereby increasing binding and reducing variability between isolates within the same bacterial species. However, even with cell wall disruption, FcMBL did not bind 12 bacterial species, including multiple isolates of E. faecalis and P. mirabilis. These microbes likely lack the complex polysaccharide antigens which FcMBL and MBL bind. Alternatively, the binding sites might still be present, but if they are, they remain inaccessible due to the unique structure of their cell wall (e.g. carbohydrate conformation, sugar density or composition). Alternatively, the antibiotics we used might not be optimal for disrupting the cell walls of these bacteria.
To emphasize the ability of FcMBL to be used to detect the presence of a systemic pathogenic infection even when blood cultures are negative, we tested its ability to bind LPS and LTA that are major PAMP-associated toxins released by multiple species of bacteria. FcMBL was able to detect both LPS and LTA from all 6 bacterial species tested in both buffer and blood, although detection sensitivity was consistently higher in buffer.
In addition, we explored whether FcMBL binds to the antigenic PAMPs, LAM, PIM 1,2 , and PIM 6 from M. tuberculosis H37Rv because these are active virulence factors associated with TB pathogenesis, and hence, they are critically important targets for point-of-care diagnostic and vaccine applications 35-38 . We found that FcMBL can detect LAM and PIM 6 , but not PIM 1,2 , in buffer, urine, and blood; this difference in binding is likely due to the fact that PIM 1,2 has 4 fewer branched mannose residues than PIM 6 39 .
Preliminary viral antigen detection in the ELLecSA was encouraging, demonstrating FcMBL binding to 8 of 10 (80%) species. We screened these viral antigens at a range of 0.1 µg/ml to 10 µg/ml that is within or below the range at which viral proteins are known to induce an immune response in vaccines 40-42 .
While the FcMBL ELLecSA cannot distinguish between different types of infections, we have previously shown that it can be used to rapidly (<1 h) detect the presence of blood infections in whole blood samples from patients suspected of sepsis regardless of whether or not they have positive blood cultures 23 . Use of the FcMBL ELLecSA in conjunction with other tools, such as C-reactive protein (CRP) and Procalcitonin (PCT), could help inform and assist the physician in deciding if there is an infection, whether hospitalization is critical, or whether antibiotics should be administered when a patient first enters a care center.
In addition, we have found that we can combine the FcMBL capture of PAMPs with additional molecular diagnostic tools, such as PCR, to distinguish different pathogen types based on molecular composition of the samples 22 .
In summary, FcMBL's ability to both bind to numerous types of infectious pathogens and capture many of the cell wall PAMPs released by these microbes when treated with antibiotics in complex biological fluids, further demonstrates the potential value of using FcMBL capture for rapid detection of bloodstream infections, even when blood cultures are negative. As a result of extending our FcMBL-based ELLecSA studies to a broader range of different pathogens, we determined a higher (87%) binding efficiency than that observed in preliminary studies. We also now better understand how FcMBL interacts with different bacterial strains of the same species when mechanically disrupted or fragmented by antibiotics. To our knowledge, this is the broadest range and largest number of pathogens and PAMPs that have been shown can be detected by a single blood opsonin or lectin. The ability of FcMBL to detect cell wall fragments also synergizes well with standard of care antibiotic therapy, and it's broad-range pathogen capture and detection can be leveraged to develop a wide range of infectious disease diagnostics, therapeutics, and vaccines. Preparation of fungi, viral antigens, parasites, and bacterial cell wall antigens Fungi species were primarily subcultured in RPMI 10 mM glucose; however other media, such as potato dextrose broth (Teknova, USA), were used to facilitate growth. In these cases, the fungal cells were pelleted at 3,000 × g for 5 minutes at 22°C (Eppendorf 5424, USA), washed 3x in 50 mM Tris-HCl, 150 mM NaCl, 0.05% Tween-20, 5 mM CaCl 2 , pH 7.4 (TBST 5 mM CaCl 2 ) (Boston BioProducts, USA) to remove residual growth media, and then resuspended in TBST 5 mM CaCl 2 . Testing by FcMBL ELLecSA was performed on titers of live fungi at ≤10 7 CFU/ml. Purified or inactivated viral antigens, parasites, and bacterial antigens were resuspended or diluted in TBST 5 mM CaCl 2 for testing directly by FcMBL ELLecSA. Trichomonas vaginalis was screened at 0.46 µg/mL and Plasmodium falciparum Circumsporozoite protein was screened at 50 µg/mL. Concentrations of viral antigens tested are indicated in the legend of Figure 7.

Antibodies
Anti-protein A (catalog number ab19483) was purchased from Abcam (Cambridge, USA).

Biological reagents
Fresh whole human blood (sodium heparin) was purchased from Research Blood Components, LLC. (Boston, USA), and normal single donor human urine (IR100007) was purchased from Innovative Research Inc. (Novi, USA).

FcMBL ELLecSA
The key metric used to quantify direct FcMBL binding to pathogen-associated molecular patterns (PAMPs) from bacteria, fungi, viral antigens, parasites, and bacterial cell wall antigens is a 96 well ELLecSA, which has been previously published 23 . The assay uses FcMBL coated superparamagnetic beads (1 µm MyOne Dynabead [Thermo Fisher Scientific, USA]) where FcMBL, biotinylated at the N termini of the Fc protein using an N-terminal amino-oxy reaction, is coupled to streptavidin beads in an oriented array ( Figure 1A). Each sample was screened using 100 µl or 200 µl of test sample added to 900 µl or 800 µl of assay solution respectively, which contains 5 µg of the FcMBL beads at 5 mg/ml and 10 mM glucose in TBST 5 mM CaCl 2 to total 1 ml (50 µl heparin is added if testing blood

4.
alternative to the time consuming and often false negative blood culture. It can also be a fantastic tool, helpful in the screening and simple identification of MBL ligands (for example -after SDS-PAGE-WB and immunostaining with PAMP-specific antibodies).
Minor points Mannosylated LAM H37Rv strain should be described as "ManLAM" Mycobaterium tuberculosis Page 4: probably should be "including multiple isolates" instead of "including multiple species" On the page 4, referring to Table 1 -the authors stated that they tested FcMBL binding to 12 different species of fungi. However , in Table 1 only 3 isolates belonging to 3 species are listed and in Figure 4 -results of FcMBL interaction with 13 fungi isolates is presented. It is confusing. On page 3 the Authors stated that pathogens were treated with antibiotic within 4 hours, however according to the method section -"bacteria were treated with antibiotics for ≥4 hours". It should be clarified.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
28 January 2019 Reviewer Report https://doi.org/10.5256/f1000research.19079.r43622 © 2019 Eisen D. This is an open access peer review report distributed under the terms of the Creative Commons , which permits unrestricted use, distribution, and reproduction in any medium, provided the original Attribution Licence work is properly cited.

Damon Eisen
James Cook University, Townsville, Australia The authors, who are the inventors of the FcMBL capture assay, have undertaken rigorous experimental work to produce a clinically oriented paper that extends the range and numbers of microbes tested in prior publications. The current publication describes the detection of bacteria, fungi, parasites and viral antigens by the enzyme-linked lectin sorbent assay (ELLecSA). Microbiologically relevant quantities of these microbes were tested and high rates of detection by ELLecSA are reported. As before, in vitro fragmentation of bacteria using antibiotics and physical fragmentation increased the numbers of bacteria detected.
The mechanisms of this increased sensitivity of detection were explored by the quantification of FcMBL-binding to the major virulence determinants of gram-negative and gram-positive bacteria. Greater sensitivity of binding to LPS was shown compared with LTA. This result provides some explanation for FcMBL's failure to detect E. faecalis.
The current publication does not include any further investigation of patients with the sepsis syndrome relying on the author's prior work to imply that a broader range of bacterial causes of sepsis will be detected. This is a reasonable assumption.
Further work using the same techniques could incorporate assessment of FcMBL's detection of in vitro rickettsia, leptospires and treponemes. These bacteria all cause diseases that are predominantly diagnosed by serology. Expedited diagnosis of blood stream or central nervous system infection would be clinically beneficial and direct appropriate antimicrobial treatment in patients with appropriate epidemiological characteristics. This could be particularly relevant to use of FcMBL technology as part of biodefence.
As the authors emphasise, the use of FcMBL can only indicate the presence of a microbial pathogen in the bloodstream due to the non-specific characteristics of binding to PAMPs. A positive ELLecSA may provide clinicians confidence that a patient presenting with the sepsis-syndrome has an infectious cause. Further testing in such patients will give additional information on the validity of the test and may provide additional encouragement for its incorporation in a microbiological testing algorithm.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound?

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
Reviewer Expertise: Infectious Diseases physician Clinical scientist with extensive history of MBL biology research.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com