Nuclease activity: an exploitable biomarker in bacterial infections

ABSTRACT

Introduction

In the increasingly challenging field of clinical microbiology, diagnosis is a cornerstone whose accuracy and timing are crucial for the successful management, therapy, and outcome of infectious diseases. Currently employed biomarkers of infectious diseases define the scope and limitations of diagnostic techniques. As such, expanding the biomarker catalog is crucial to address unmet needs and bring about novel diagnostic functionalities and applications.

Areas covered

This review describes the extracellular nucleases of 15 relevant bacterial pathogens and discusses the potential use of nuclease activity as a diagnostic biomarker. Articles were searched for in PubMed using the terms: ‘nuclease,’ ‘bacteria,’ ‘nuclease activity’ or ‘biomarker.’ For overview sections, original and review articles between 2000 and 2019 were searched for using the terms: ‘infections,’ ‘diagnosis,’ ‘bacterial,’ ‘burden,’ ‘challenges.’ Informative articles were selected.

Expert opinion

Using the catalytic activity of nucleases offers new possibilities compared to established biomarkers. Nucleic acid activatable reporters in combination with different transduction platforms and delivery methods can be used to detect disease-associated nuclease activity patterns in vitro and in vivo for prognostic and diagnostic applications. Even when these patterns are not obvious or of unknown etiology, screening platforms could be used to identify new disease reporters.

KEYWORDS:

• biomarker
• clinically relevant
• diagnosis
• enzyme
• clinical microbiology
• molecular
• nucleases
• nuclease activity
• bacterial pathogen

1. Introduction

Despite a continued reduction in the last decades, the global burden of infectious diseases is still immense [1–4]. Although these epidemiological studies indicate that the biggest proportion of cases, deaths, and burden occur in low-income countries, it may be misleading to assume that the impact of infectious diseases is inconsequential in more developed and higher income countries. In fact, according to data estimates in Europe, between 2009 and 2013 almost 38 million cases of infectious diseases, 1,38 million Disability-adjusted life years (DALYs) and 50,000 fatalities per year were attributable to 31 different infectious diseases including influenza, tuberculosis, HIV/AIDS, invasive pneumococcal disease (IPD), salmonellosis, and campylobacteriosis, among others [5].

In addition to the existing burden, new insights into disease mechanisms and newly arising phenomena raise serious concerns regarding the real impact of infectious diseases and our ability to battle against them. These include (1) rising evidence of causal connections between infectious and non-infectious diseases, such as different types of cancer [6–9] or chronic conditions [10]; (2) antimicrobial resistance (AMR), which menaces to difficult or impede therapeutic efforts, increase the burden, especially in clinical settings; and increase health-care costs [11–15]; and (3) threats posed by outbreaks, epidemics, and pandemics caused by re-emerging or emerging infectious agents.

Considering this grim picture, effective prevention, rapid and accurate diagnosis, and adequate therapy are the three main available tools to combat infectious diseases and reduce their burden. Diagnosis is essential; however, its scope is often overlooked. Commonly, the diagnosis of infectious disease refers to the process of identification of disease and its etiological agent in individual patients for their management in clinical settings. This is itself a complex and multifactorial process that varies widely across clinical settings and population niches, each demanding specific requirements and presenting unique challenges. However, infectious disease diagnostics also have a broader range of applications in health care beyond individual patient management, which include antimicrobial stewardship programs; control of emerging or re-emerging infectious events during outbreaks, epidemics or pandemics; population screening or epidemiological surveillance [16].

In essence, diagnosis involves the identification of disease; however, it may not be self-evident that the accuracy and timing of such identification also have profound ramifications in the subsequent management, therapy, and outcome of disease. New developments leading to improved features in the field of diagnostics, including faster pathogen identification and antimicrobial susceptibility evaluation, have considerably accelerated times to adequate treatment, reduced hospital stays and mortality, and greatly reduced health-care associated costs in clinical settings when compared with more traditional methods [17]. Nonclinical applications, such as industrial microbiological quality control, also benefit from the use of these improved diagnostic technologies [18]. As such, the importance of diagnostics warrants a continued evolution of diagnostic biomarkers, tools and methodologies .

Diagnostic biomarkers of infection of varied nature are used in clinical microbiology. For example, diagnostic modalities based on microbial culture and microscopy make use of disparate biomarkers depending on the technique, including morphology or metabolic demands, while immune or nucleic acid-based diagnostics rely on the presence or absence of antibodies/antigens or genetic determinants, respectively. Ultimately, the choice of biomarker defines the potential and features of a diagnostic technology (Table 1).

Table 1. Current and emerging diagnostic modalities in clinical microbiology. Overview of the main modalities, their advantages and disadvantages, and the biomarkers employed

Modality	Biomarker	Advantages	Disadvantages	Refs
Microscopy	Varied, including: Morphology Growth Presence or absence of dye	Rapid Sensitive and Specific*	Low throughput Technically demanding Elaborate sample preparations** Expensive**	[19,20]
Microbial culture	Varied, including: Growth characteristics Morphological characteristics Metabolic characteristics Biochemical characteristics	High sensitivity High specificity	Time-consuming Costly and cumbersome for intracellular pathogens Incompatible with certain pathogenic organisms Contamination risk	[16,19,21]
Immunoassays	Antigen-antibody interaction	High specificity Rapid Low cost High-throughput capability Compatible with automation Retrospective diagnosis	Low sensitivity Prone to interferences No antibiotic susceptibility testing Worse diagnostic accuracy than molecular tests Accuracy dependent on quality control Cost and complexity associated with antibody production	[20,22–25]
Molecular genetic techniques	Genetic determinants	Rapid* High sensitivity and specificity Low detection limits Provides genotypic antibiotic susceptibility information Multiplexing and absolute quantification capabilities High-throughput capability Compatible with automation	Overreporting (false positives) Requires specialized equipment Medium to high costs and complexity Differences between genotypic and phenotypic antibiotic resistance Need for nucleic acid isolation Prone to cross-contamination*	[133]
Mass spectrometry	Mass spectrum profile or genetic determinants	Very rapid High sensitivity and specificity Low detection limits* Low running costs Provides antibiotic susceptibility information	Need for enrichment and amplification steps High initial equipment costs	[133]
Biosensors	Varied, including: Enzymes Genetic determinants Antigens Antibodies Growth	New diagnostic possibilities and formats Novel functionalities	Slow adoption Calibration methods Quality assurance High costs Low sensitivity and high detection limits**	[18,252]

*Depends on the application.

**Depends on the technique

Nucleases are a diverse family of protein enzymes, present across all domains of life, that mediate the degradation of nucleic acids by cleaving the phosphodiester bonds that conform their backbone [26]. A great number of nucleases have been identified and characterized for many different species in the domain Bacteria, in many cases playing key roles in their associated pathogenesis. The great diversity of bacterial nucleases, their capability to degrade natural (DNA or RNA) or chemically modified nucleotides (2’-Fluoro, 2’-O-Methyl, LNA, etc.) and their varied biochemical and catalytic properties, such as substrate preference, sequence preference, cation dependency or thermostability; make them promising candidates as diagnostic biomarkers. It is known that bacterial and human nucleases can be distinguished based on their nuclease activity profile [26], and both bacterial nuclease activity [27] and nuclease patterns [28] have already been used and proposed, respectively, to identify and discriminate between bacterial species. The use of bacterial nuclease activity blueprints as novel diagnostic biomarkers is particularly promising, and novel approaches are expanding its potential for the successful identification and characterization of clinically relevant bacterial pathogens. Several methodologies to analyze nuclease activity have been reported , including Fluorescence Resonance Energy Transfer (FRET) systems, which report the oligonucleotide cleavage [29]; mass spectrometry [30] or capillary electrophoresis [31], which provide oligonucleotide fragmentation profiles. Based on some of these methodologies biosensing systems that exploit bacterial nuclease activity as a biomarker have already shown promise when complementing existing diagnostic technologies [32], addressing unmet diagnostic needs in clinical settings [33] and setting the basis for new diagnostic avenues and methodologies [29].

In this context, due to their readily accessibility in living bacteria, extracellular (membrane-associated, cell wall-anchored and secreted) nucleases represent more propitious candidates than intracellular nucleases for their use as diagnostic makers.

In this review, we briefly overview the immense repertoire of bacterial nucleases and present relevant examples from the literature of how they have been used and how could they be used as diagnostic biomarkers. Finally, we present several well-described extracellular nucleases associated with 15 clinically relevant pathogenic bacteria, whose activity blueprints could be exploited as clinical diagnostic biomarkers.

2. Nucleases in bacteria

2.1. Intracellular nucleases

A plethora of intracellular nucleases has been identified in bacteria, and most of them have been reported and characterized using Escherichia coli (E. coli) and Bacillus subtillis (B. subtillis) as model organisms. Dozens of RNases and DNases possessing different biological roles, catalytic activities, biochemical properties and regulatory mechanisms have been described. Moreover, since the advent of the genomic era, many homologues of some of these known nucleases have been identified in silico across all domains of life. Despite the fact that nucleases are typically grouped into subfamilies, families and superfamilies according to structural and sequence-related features for clarity and classification purposes [34], the real scope of their prevalence, biochemical diversity and involvement in a wide range of biological processes can be hard to grasp. In brief, roles ranging from DNA proofreading and repair, maintenance of genome stability or virulence-associated genetic recombination [35]; to RNA metabolism (e.g. maturation, turnover or regulation) or toxin-mediated growth control [36,37] are performed by a large number of nucleases displaying different modes of action (hydrolysis or phosphorolysis), types of cleavage (endonucleases, exonucleases or both) and substrate preference (DNases, RNases or both) with varying intracellular localizations (nucleus, cytosol, inner membrane or periplasm), substrate specificities (structure or sequence preferences), processivity, polarity (5´to 3´ or 3´to 5´) or co-factor requirements (Mg²⁺, Ni²⁺or Co²⁺), among others [38,39].

However, the diversity of intracellular nucleases is not limited to their number. Even between highly conserved bacterial nuclease homologues, significant operational differences may exist, as reported for RNase E homologues from different pathogenic bacterial species [40]. This is also illustrated by the differences between the Exonuclease III (Exo III) and Endonuclease IV (Endo IV) families of class II apurinic/apyrimidinic (AP) endonucleases. Exo III and Endo IV nucleases have a fundamental role in the repair of abasic DNA lesions caused by highly oxidative environments, such as the one encountered by some bacterial pathogens inside macrophages [41], and are found in organisms from bacteria to humans. However, major differences in the relative activity of Exo III and Endo IV between homologues present in bacteria, yeasts and humans have been observed. In fact, while Exonuclease III is responsible for the bulk of the AP endonuclease activity in E. coli and humans (Ape I), Endonuclease IV is the principal AP endonuclease in M. tuberculosis (End) and Saccharomyces cerevisae (Apn 1) [41,42]. Besides, when comparing in greater detail the activities of AP endonucleases between E. coli and Mycobacterium tuberculosis (M. tuberculosis), similarities are observed as expected by their homology including the ability of both nucleases to degrade ssDNA and dsDNA abasic substrates, the display of both exonuclease and endonuclease activities, the metal-ion dependency or the differential sensitivity between Exo III and Endo IV to chelating agents (EDTA) [41–43]. However, a relevant nuclease property, namely the sequence preference of the base opposite to the abasic lesion, varies between E. coli and M. tuberculosis [41,44,45]. One more example of operational differences between nuclease homologues includes Endonuclease V, a highly conserved nuclease from bacteria to humans, that varies its substrate preference (DNA to RNA) and function (DNA repair to RNA editing) between bacterial and eukaryotic organisms [46].

Furthermore, beyond nuclease homologues, differences in nuclease populations among bacterial species have been reported that suggest that different bacteria may use non-homologous nucleases to fulfill similar biological roles through disparate pathways and mechanisms pointing towards the existence of yet undescribed novel intracellular nucleases across bacterial species. This phenomenon is illustrated by nuclease populations identified in two putative model organisms, namely E. coli and B. subtillis. Despite sharing several conserved homologous nucleases, an even greater number of nucleases are unique for each species. As such, major differences in the mode of action between B. subtillis and E. coli nuclease populations exist. While 90% of the nucleases found in E. coli extracts are hydrolytic, phosphorolysis dominates the nuclease activity in B. subtillis extracts [39]. However, as reported before [26], different catalytic mechanisms are poorly correlated with biological roles, as exemplified by the fact that some of the unique non-homologous nucleases, such as RNase E (E. coli) and RNase Y in (B. subtillis) possess equivalent biological roles [47]. Similarly, several exonucleases belonging to different protein families share their preference for ssDNA and have seemingly redundant functions, as reported for E. coli [38].

2.2. Extracellular nucleases

A great number of extracellular nucleases, including both membrane-bound and secreted, with cytotoxic and non-cytotoxic roles have been identified in numerous bacterial species. In a similar fashion to intracellular nucleases, the identified extracellular nucleases display diverse biochemical and catalytic properties involving different types of cleavage, substrate and sequence preference, catalytic efficiencies, cofactor requirements, optimal pH and temperature, ionic strength or thermal stability, among others. Even between homologous nucleases displaying very high sequence identity and similarity, noticeable catalytic and biochemical differences exist that allow their distinction, as exemplified by the extracellular nuclease homologues VcEndA/Dns and VsEndA produced by Vibrio cholera (V. cholera) and Vibrio salmonicida, respectively [48,49].

2.2.1. Cytotoxic extracellular nucleases

Bacteria produce an abundance of extracellular effector molecules and toxins that make use of numerous mechanisms for their secretion, transfer, and delivery , which in many cases involve specialized secretion systems [50,51], into prokaryotic or eukaryotic targets, where they participate in bacterial warfare or act as virulence factors during host infection. Many of these effectors and toxins, as described below, are known to possess nuclease activity and rely on it to perform their function.

On the one hand, numerous plasmid and chromosomally encoded toxins involved in bacterial warfare classified in various toxin systems have been described. Among others, these include experimentally well-characterized classical toxin systems, such as bacteriocins (e.g colicins, pyocins or klebicins) [52–54] as well as a vast array of polymorphic toxin systems identified in silico across major bacterial lineages [55–57]. Interestingly, both classical bacteriocins and polymorphic toxins are known to share a number of common features. First, these toxins share different degrees of domain homology and a modular domain structure. N-terminal domains serve regulatory functions, such as trafficking, while the C-terminal domains contain the active cytotoxic domains, which can display diverse types of nuclease activity that include DNase, mRNase and tRNase. As such, these toxins tend to exert their cytotoxic functions by tampering with the flow of information inside the target cell, either by degrading nucleic acids involved in protein synthesis or by direct destruction of the genomic material. Of interest, both of these types of toxins possess co-transcribed immunity pair proteins that avoid self-intoxication and keep the toxins inactive until their delivery into the target [52–54,56].

On the other hand, many pathogenic bacteria are well known to use a wide range of effectors of different nature to hijack key pathways or induce the death of host´s cells as part of their pathogenic strategy [58]. Unsurprisingly, many of these host-targeted cytotoxic effectors and toxins found in different pathogenic bacterial species exhibit nuclease activity. Perhaps one of the best studied examples is the heterotrimeric cytolethal distending toxin (CDT), homologues of which are encoded by numerous proteobacteria. Upon translocation into the cytosol and nuclear targeting, it causes cell cycle arrest followed by gross morphological changes and cell death [59,60]. The cytotoxicity of this toxin has been attributed to the Dnase I-like enzymatic activity of the CdtB subunit that induces DNA damage [61–63]. In a similar fashion, AbOmpA, a major surface protein of Acinetobacter baumannii that has also been shown to be secreted, targets both the mitochondria and the nucleus and induces apoptotic cell death. Analogously to CdtB, nuclear localization is dependent on a nuclear localization signal (NLS), it exhibits DNase-I like activity capable of degrading chromosomal DNA and its cytotoxic effect is more prominent in macrophages compared to other cell types [64,65]. In a similar fashion, Corynebacterium diphtheriae (C. diphtheriae) toxin possesses DNase I-like activity and induces internucleosomal degradation and cytolysis [66]. In line with these observations, Lee et al. screened for and identified 49 proteins with a predicted NLS in Helicobacter pylori, and demonstrated that 26 of these proteins were indeed localized in the nucleus using a fibroblast (COS-7) in vitro cell model [67]. Further investigations by Kim et al. characterized one of these proteins (HP0425) as a cytotoxic Mn²⁺-dependent nuclease, capable of degrading supercoiled plasmid DNA and genomic DNA [68]. Additionally, Helicobacter pylori has been shown to induce contact-dependent genomic DNA damage in host´s cells through an unknown mechanism, which is known to be independent of (1) DNA synthesis, (2) the action of reactive oxygen or nitrogen species or (3) known virulence factors, such as cytotoxin associated gene pathogenicity island , vacuolating cytotoxin A or gamma-glutamyl transpeptidases [69].

2.2.2. Non-cytotoxic extracellular nucleases

Membrane-bound or secreted extracellular nucleases with non-cytotoxic roles in bacterial pathogenesis, such as nutrient scavenging, immune modulation, biofilm remodeling or horizontal gene transfer, have been described for several clinically relevant bacterial pathogens (Table 2). These include Gram-positive bacteria such as Staphyloccocus aureus, Streptococcus pneumoniae, Streptococcus pyogenes, Streptococcus agalactiae and Clostridium perfringens; as well as Gram-negative bacteria such as E.coli, Neisseria gonorrhoeae, Serratia marcescens, Helicobacter pylori, M. tuberculosis, Salmonella enterica, Campylobacter jejuni and V. cholera (Section 5). However, the aforementioned list is not exclusive and other clinically relevant bacteria not considered in this review are also known to possess non-cytotoxic extracellular nucleases.

Table 2. Biological, biochemical, and catalytic parameters of the extracellular nucleases that represent potential candidates as diagnostic biomarkers for 15 clinically relevant bacterial pathogens

Pathogen	Nucleases	Encoding	Cellular location	Function/s	Cleavage	Substrates	Cofactor	Optimal pH	Thermostability	References
S. aureus	Nuc	Chromosomal	Secreted	Biofilm regulator and immune evasion	Exo- endonuclease	RNA and ss/dsDNA	Ca2+	9–10	Yes	[135–137]
	Nuc2	Chromosomal	Membrane-bound	Biofilm regulator	Endonuclease	RNA and ss/dsDNA	Ca2+	-	Yes	[142]
	EssD	Chromosomal	Secreted	Modulate host immune response	Endonuclease	dsDNA	-	-	-	[143]
S. pneumonia	EndA	Chromosomal	Membrane-bound/secreted	Genetic transformation and immune evasion	Endonuclease	RNA and DNA	Mg2+ and Mn2+	-	-	[151–153]
S.pyogenes	SpdB/DnaseB/MF/SdaB or SpeF	Chromosomal	Secreted	Unknown	Endonuclease	RNA and ss/dsDNA	Mg2+, Ca2+, Mn2+, Sr2+, Cd2+, Cu2+, Co2+, Ni2+, Zn2+ and Fe2+	~9,5	Yes	[161–163]
	SpnA	Chromosomal	Cell-wall anchored/secreted	Immune evasion	Exo-endonuclease	RNA and ss/dsDNA	Mg2+ and Ca2+	5 to 7,5	-	[164–166]
	Spd1/MF2	Chromosomally integrated prophage-like elements	Secreted	Enhance fitness and bacterial shedding	Endonuclease	RNA and ss/dsDNA	Mg2+ and Ca2+	Only tested at 7–8	-	[168,169,171,172]
	Spd3/MF3	Chromosomally integrated prophage-like elements	Secreted	Unknown	Endonuclease	DNA	Mg2+ and Ca2+	-	-	[160,167,175]
	Spd4/MF4	Chromosomally integrated prophage-like elements	Secreted	Unkown	-	-	-	-	-	[160]
	Sda1	Chromosomally integrated prophage-like elements	Secreted	Immune evasion	Endonuclease	DNA	Mg2+ and Ca2+	Only tested at 6,8–8,8	-	[160,175,176,178]
	Sda2/SdaD/DnaseD	Chromosomally integrated prophage-like elements	Secreted	Unknown	-	DNA	-	7	-	[160,174]
	Sdn/Sdα	Chromosomally integrated prophage-like elements	Secreted	Unknown	Endonuclease	DNA	Mg2+ and Ca2+	-	-	[160,173]
S. agalactiae	NucA	Chromosomal	Secreted	Immune evasion.	Endonuclease	RNA and ss/dsDNA	Mg2+, Mn2+, Zn2+, Cu2+, and Ni2+	5,5 to 7,5	Yes	[185]
E. coli	EndA	Chromosomal	Periplasm	Unknown	Endonuclease	ss/dsDNA	Mg2+	-	-	[195,196]
N. gonorrhoeae	Nuc	Chromosomal	Membrane-bound/secreted	Biofilm regulator and immune evasion	Endonuclease	ss/dsDNA	Mg2+ and Ca2+	-	Yes	[203,204]
S. marcescens	NucA	Chromosomal	Secreted	Possible role in nutrient scavenging	Endonuclease	ss/dsRNA and ss/dsDNA	Mg2+ and Mn2+	8 to 8,5	No	[213,215–218]
H. pylori	NucT	Chromosomal	Membrane-bound/ Periplasmic	Genetic transformation and nutrient scavenging	Endonuclease	RNA and ss/dsDNA	None required	8	Yes	[230–232]
M. tuberculosis	Rv0888	Chromosomal	Membrane-bound	Possible role in pathogenesie by enhancing bacterial persistence and survival	Endonuclease	RNA and dsDNA	Mn2+, Mg2+, Ba2+, Ni2+ and Ca2+	6,5	-	[238,239]
S. enterica ser. Typhimurium	Nuc	Plasmid	Periplasmic	Unkown	Endonuclease	ss/dsDNA	Mg2+, Ca2+, Co2+, and Zn2+ (Discrepancy on cofactor dependency)	5,5 (ssDNA) and 6,8 (dsDNA)	-	[257,258]
C. jejuni	Dns	Chromosomally integrated prophage-like elements	Periplasmic (possibly secreted)	Regulation of genetic transformation	Expected Endonuclease	dsDNA	Mg2+	-	-	[261]
	CJE0566	Chromosomally integrated prophage-like elements	Periplasmic (possibly secreted)	Regulation of genetic transformation	Expected Endonuclease	DNA	-	-	-	[259]
	CJE1441	Chromosomally integrated prophage-like elements	Periplasmic (possibly secreted)	Regulation of genetic transformation	Expected Endonuclease	DNA	-	-	-	[259]
V. cholera	Dns	Chromosomal	Secreted	Regulation of genetic transformation, biofilm formation, nutrient scavenging and immune evasion	Endonuclease	RNA and ss/dsDNA	Mg2+	7,5 to 8	Yes	[48,263,268–270]
	Xds	Chromosomal	Secreted		Exonuclease	Linear DNA	Mg2+ and Ca2+	7 to 8	-	[263,267–270]
Y. eneterocolitica	YNSN	Chromosomal	Membrane-bound/Secreted	Unknown	Predicted Endonuclease	DNA (predicted RNA)	Predicted Mg2+	-	Yes	[273,274]
C. perfringens	CadA	Chromosomal	Cell-wall anchored	Unknown	Endonuclease (possibly exonuclease)	dsDNA	Mg2+ and Ca2+	-	-	[277]

For example, two extracellular secreted nucleases (EddB and EndA) have been identified in Pseudomonas aeruginosa capable of degrading extracellular genomic DNA, which is in line with previously reported unassigned extracellular DNase activity in several typed and untyped strains from clinical isolates [70,71]. EddB is a metal-ion-dependent nuclease, whose expression is increased in hyperdispersive biofilms and induced by phosphate limiting conditions in a dose-dependent manner and extracellular DNA, such as neutrophil extracellular traps (NET) DNA. Unsurprisingly, it participates in nutrient scavenging during starvation, immune evasion by degrading NETs and biofilm remodeling. Meanwhile, EndA bares sequence identity with E. coli´s Endonuclease I or V. cholera´s Dns and plays a fundamental role in biofilm remodeling [71–73]. Analogously, NTHI (non-typeable Haemophilus influenzae)-Nuc, a metal-ion-dependent nuclease homologous to micrococcal nuclease (MN) and involved in biofilm remodeling has been identified in silico in several annotated genomes of NTHI. It has been reported that NTHI-Nuc is capable of degrading both ssDNA and dsDNA, exhibits much faster kinetics than DNase I and it is inhibited by EDTA (4 mM) [74].

Unassigned extracellular nuclease activity blueprints have also been reported in numerous anaerobic bacteria. These include Gram-positive anaerobes, including pathogenic peptostreptococci and Clostridium spp, such as Peptostreptococcus anaerobius or the aforementioned C. perfringens respectively; and Gram-negative anaerobes, including pathogenic fusobacteria or bacteroides, such as Fusobacterium necrophorum or Bacteroides fragilis, respectively [75]. Similar unassigned activity has been reported in members of the corynebacteria, including Corynebacterium ulcerans and C. diphtheriae, the activity of the latter species being independent of the activity of its toxin [76]. Membrane-associated and secreted extracellular nuclease activity of unassigned origin has also been described for numerous pathogenic species and strains of mycoplasma [28], some of which have been shown to induce immortalization and malignant transformation of different human cells in vitro [77,78] and have been associated with the development of malignancies in humans [79–82].

In fact, recent studies have identified membrane-associated extracellular nucleases in Mycoplasma meleagridis (Mm19) [83] and Mycoplasma hyopneumoniae (mhp379) [84], common animal pathogens, as well as in species isolated from humans, such as Mycoplasma pneumoniae (M. pneumoniae) (Mpn133) [85], Mycoplasma genitalium (MG-168) [86] or Mycoplasma penetrans (P40) [87,88]. Interestingly, mhp379, MG-168 and Mpn133 are Ca2+-dependent nuclease homologues with broad substrate specificity belonging to the MN cluster of orthologous proteins (COG1525) that are encoded upstream of genes encoding for homologous ABC transport systems, which is in line with the previously proposed nutrient scavenging roles of mycoplasma nucleases [28]. This hypothesis is further supported by the ability of other mycoplasmas to use undegraded DNA and RNA as a nutrient source [89] and the observed reduction of the cytotoxic effects of M. pneumoniae in the presence of adenine supplement [90]. However, their ability to induce internucleosomal DNA degradation [86,91,92] and observations in in vitro models of human-derived cell of membrane binding, internalization, reduction of viability and induction of apoptosis [85,88], also suggest a role of mycoplasmas´ extracellular nucleases as pathogenic determinants.

3. Nuclease activity as a diagnostic biomarker

As early as the 1950s, the production of extracellular deoxyribonuclease activity was proposed as a useful phenotypic trait to aid in the biochemical identification, characterization, and discrimination of bacterial species in clinical microbiology. As such, different assays, referred to collectively as DNase test, have been developed over the last decades for the detection of nuclease activity [93–95] and are still in use today [96]. The use of the DNase test has been proposed as a means to discriminate Serratia spp from other species of the Enterobacteriaceae family based on the presence or absence of extracellular nuclease activity [97–99]. Similarly, the DNase test has been postulated as a cost-effective and simple screening tool for C. diphteria, for its ability to discriminate between diphterial and non-diphterial corynebacteria with high sensitivity (100%) and specificity (93,9%) [76]. Ultimately, the DNase test has most commonly been used to specifically identify Staphylococcus aureus (S. aureus). However, to overcome the specificity issues posed by the production of thermolabile extracellular nucleases by micrococci and coagulase-negative Staphylococci, such as Staphylococcus epidermidis [100], a derivative of the DNase test, known as the thermonuclease (TNase) test, that exploits the thermostability of MN was developed. TNase is not only more specific than the DNase test, but its accuracy has been shown to match that of the tube coagulase test (TCT) for the identification of S. aureus in food and clinical isolates of Gram-positive cocci [27,101–103]. Moreover, it represents a simple, rapid (~ 2,5 hours) and inexpensive methodology for the detection at very low concentrations (5–10 ng/g – approx. 10–3 units/g) directly from food samples of MN, which is a good indicator of S. aureus contamination, even when viable bacteria are no longer available after food processing [104,105]. Unsurprisingly, the TNase test has and continues to be proposed as a very accurate, rapid (1 -4 hours), simple and inexpensive alternative to molecular methods for the identification of S. aureus bacteremia from positive blood cultures that exhibit Gram-positive cocci on a direct Gram stain [96,106–108]. However, the production of thermostable nucleases by typical bacteremia suspects like Enterococcus faecalis and coagulase-negative staphylococci, amid with much lower prevalence and different thermostability profiles [109], may occasionally lead to the misidentification of the causative agent. It is also worth mentioning that nuc, the gene coding for S. aureus putative nuclease (MN), has been employed as a targeted genetic determinant in different diagnostic NAATs for its identification in food and clinical isolates, such as positive blood cultures [110–115]. However, concerns have been raised about the risks of using nuc as a sole genetic determinant, due to the appearance of strains with marked gene sequence heterogeneity that may lead to false negatives. Importantly, all the identified genetic alterations involved same sense mutations [116], and therefore, do not affect the applicability and the accuracy of the DNase/TNase test.

The scope of the approaches that take advantage of the catalytic activity of MN as a biomarker for the identification of S. aureus has been expanded of late. Some recently reported biosensing platforms utilize nucleic acid-based recognition elements in combination with fluorescence-based transduction mechanisms to detect the enzymatic activity of MN both qualitatively and quantitatively with high selectivity and very low reported limits of detection, ranging from 2,9x10−3 to 2,7x10−5 unit/ml (approx. 0,1 ng/ml) in a reproducible manner. However, in most cases their performance in complex matrices remains to be validated, which limits their applicability to the detection of S. aureus´s nuclease solely from pure cultures [117–119]. Similar approaches have also been adopted for other nucleases, such as DNase I, while adding new features like real-time monitoring of enzymatic activity [120]. One of the most explored and flexible biosensing platforms to detect the catalytic activity of bacterial nucleases is based on the use of short, chemically modified self-quenched fluorescent oligonucleotide probes that can be easily tuned to serve as their specific substrates, allowing bacterial identification in vitro in complex clinical specimens, such as urine or plasma, as well as in vivo. Specific examples of the use of nuclease activity as biomarkers for the detection and identification of both Gram positive [29] and Gram negative [32,33] bacteria have been reported. Hernandez et al. could image a S. aureus infection in vivo by taking advantage of the catalytic properties of its secreted nuclease (MN) in a murine model of pyomyositis [29]. Based on an analogous but optimized platform; a cost-effective, ultrasensitive and highly specific assay for the rapid (3 hours) identification of S. aureus bacteremia directly from non-enriched blood plasma has also been reported [121]. Flenker et al. also exploited this platform to develop a rapid (3 hours), cost-effective and simple assay, based on the detection of the dominant activity of E. coli´s chromosomally encoded endonuclease I, for the diagnosis of urinary tract infection (UTI) caused by this bacterium directly from urine specimens with minimal processing and reportedly equivalent or higher sensitivity (S) and negative predictive values (NPV) for both E. coli UTI (S: 97%; NPV: 98%) and general UTI (S: 95,3% NPV: 87,5%) than rapid dipstick-based urinalysis. Furthermore, it holds promise to provide accurate pathogen identification and quantification while being pliable to automation and multiplexing [33]. Yet again, Machado et al. used the same biosensing platform to rapidly (8 hours) and accurately detect Salmonella enterica ser. Typhimurium from pure cultures and cultured homogenates of fattening-pigs mesenteric lymph nodes by exploiting the nuclease blueprint associated with the bacterium. By using a screening approach to identify the oligonucleotide probes that are preferentially catalyzed by Salmonella enterica ser. Typhimurium´s nuclease blueprint, pathogen identification with matching accuracy levels to the gold standard methods is attained, without the need to be versed in the etiology of the enzymatic activity [32]. In addition to pathogen identification, the same biosensing platform has been employed for antimicrobial susceptibility testing based on two seemingly opposing strategies: growth-associated nuclease production and lysis associated nuclease release. However, both have proven more rapid (3–6 hours) and equally as accurate as classical phenotypic methods (e.g. broth microdilution method) to obtain quantitative antimicrobial susceptibility information from bacterial suspensions derived from purified colonies for both Gram-positive and Gram-negative bacterial species, respectively [122,123].

Recently, a variation of the fluorescent detection platform has also been used for the detection and identification of bacteria based on the catalytic activity (cleavage of oligonucleotide probe) of bacterial nucleases. However, the transduction mechanisms in these variations were based on magnetic resonance imaging [124] and lateral flow assays [125].

4. Significance, current diagnostics, and nuclease candidates as potential diagnostic biomarkers of 15 clinically relevant bacterial pathogens

4.1. Staphylococcus aureus (S. aureus)

4.1.1. Significance and current diagnostic methods

S. aureus is one of the most common and burdensome human pathogens, being responsible for a multitude of clinical infections including (1) bacteremia; (2) skin and soft tissue conditions, such as cutaneous abscesses or impetigo; (3) osteoarticular infections, such as osteomyelitis or septic arthritis; (4) pulmonary infections, such as ventilator-associated pneumonia; and to a lesser extend (5) toxic shock syndrome (TSS), (6) meningitis and, in rare cases, (7) UTI [126]. The appearance of antimicrobial resistant strains to penicillin, which were already observed in 1942; methicillin, quinolone and vancomycin complicates the clinical picture and entails a huge economical and medical burden. In fact, annual deaths attributable to nosocomial methicillin-resistant S. aureus (MRSA) infections have surpassed HIV/AIDS in the United States [127].

The diagnosis of S. aureus usually involves time-consuming, culture-based sample enrichment and colony isolation, followed by identification and characterization methods. Gram stains and/or colony morphology evaluations can provide presumptive pathogen identification [128]. For definitive identification different methods are available. These include different biochemical assays, which range from rapid, accurate, and inexpensive tests that are well adapted to low resource settings, such as the coagulase test [129], to automated panel-based commercial platforms that offer phenotypic identification and susceptibility information [130]. Immunological assays, such as ELISA and latex agglutination, identify S. aureus based on the detection of characteristic virulence factors, such as secreted toxins, and are especially important to identify contaminated food in the absence of bacteria [131,132]. Numerous commercial platforms based on molecular diagnostic methodologies, such as NAATs or mass spectrometry (MS), are also available which provide identification and antimicrobial susceptibility testing [133,134]. These platforms exhibit different requirements relating to specimen type, processing steps and sample purity, while some offer rapid results with minimal sample processing, others require extensive manual processing for sample preparation or culturing steps to obtain pure colonies.

4.1.2. Nuclease candidates

S. aureus produces Nuc, also known as micrococcal nuclease (MN) and Nuc2. On the one hand, MN is a well characterized secreted exo-endonuclease that can degrade both single-stranded (ss) and double-stranded (ds) RNA and DNA, with a higher preference toward ssDNA. Its main cofactor is Ca2+, though it retains DNase activity in the presence of Sr2+. It can operate in a wide range of pH values, with an optimal Ca2+ -dependent activity at basic pH levels (9 to 10), it is remarkably thermostable and its catalytic activity varies depending on pH levels, Ca2+ concentration and substrate nature. Additionally, MN activity displays both nucleotide preference towards adenine and thymine and sequence preference [135–137]. Apart from being a key regulator in biofilm formation [138], there is strong evidence, both from in vitro and in vivo models that S. aureus uses MN as a virulence factor to escape neutrophil action by degrading their extracellular traps [139]. Furthermore, MN degradation of NETs promotes a chain of events that leads to the generation of toxic nucleotide derivates that induce macrophage apoptosis [140]. This realization has even led to the pursue of therapeutic approaches that target MN, like inhibiting its catalytic activity [141]. On the other hand, Nuc2 is an extracellular facing surface attached nuclease with similar biochemical characteristics to MN, but with a considerably diminished level of activity. Importantly, Nuc2 has been reported to be expressed and functional during infection using a murine in vivo model [142]. Additionally, S. aureus uses a specialized secretion system to release a DNase effector (EssD) into the extracellular environment that has been shown to contribute to its pathogenic strategy by inducing a nuclease dependent proinflammatory response in a bloodstream infection mouse model [143].

4.2. Streptococcus pneumoniae (S. pneumoniae)

4.2.1. Significance and current diagnostic methods

S. pneumoniae is an extracellular bacterium whose main habitat is the human upper respiratory tract (UPT) [128]. Colonization of the UPT is more common among children (20–50%) and can be asymptomatic and self-resolved, leading to temporary serotype-specific immunity, or it can evolve into highly burdensome invasive disease in the form of pneumonia, bacteremia, meningitis, middle-ear infection, mastoiditis, or sinusitis. In fact, just the lower respiratory infections attributable to S. pneumoniae were estimated to cause the death of over a million people globally in 2016 [144]. Airborne or contact-dependent transmission occur upon close contact with colonized or infected individuals or fomites, and the rate of transmission is enhanced by inflammatory events, such as viral co-infection. Co-infection also increases the risk of invasive disease and mortality [145,146]. Since the introduction of the serotype-specific pneumococcal vaccines, the burden of disease has significantly decreased. However, the rising incidence of non-vaccine serotypes and the appearance of multidrug resistance in a typically penicillin-susceptible species represent new challenges [147].

A presumptive diagnosis of pneumococcal disease is typically done based on clinical symptoms and unspecific infectious markers, which in some cases leads to false positives and antibiotic overuse [148]. Phenotypic assays used for identification, such as bile solubility or optochin susceptibility testing, are based on time-consuming microbial culture, whose sensitivity is negatively affected by prior empirical antibiotic treatment. Furthermore, the presence of similar characteristics in other Streptococcus spp and S. pneumoniae strains reduce the specificity of these tests. Molecular methods have been proposed based on PCR, RT-qPCR and mass spectrometry (MALDI-TOF). However, due to genetic similarity between S. pneumoniae and the viridians group of streptococci, PCR-based methods lack specificity when used on upper respiratory tract or sputum specimens. Direct assessment of whole blood using PCR-based methods exhibits high specificity, but poor sensitivity. Although rarely used in clinical settings due to the marginal sensitivity, it has proven more rapid and sensitive than blood culture in pediatric patients with pneumonia. However, it is slower and less accurate than the urinary antigen testing in adult patients with bacteremia or pneumonia [148,149], a culture-free method based on the detection of S. pneumoniae´s capsular polysaccharide in urine. The precise identification and serotyping of S. pneumoniae are also important to reduce antimicrobial resistance and develop effective vaccines. Some of the methods used for species serotyping include multilocus sequence typing, sequential multiplex PCR and whole-genome sequencing [145,148,150].

4.2.2. Nuclease candidates

EndA is the major extracellular nuclease of S. pneumoniae, which is expressed at different levels by multiple S. pneumoniae strains and plays a key role as a virulence factor. EndA is a metal-ion-dependent endonuclease capable of degrading both DNA and RNA, which similarly to other bacterial nucleases, like Serratia marcescens´s nuclease, belongs to the family of nucleases containing a ββα-metal finger motif and possesses an N-terminal signal sequence and an active catalytic site characterized by the presence of a DRGH motif. Despite these common features, the sequence and structural similarities are minimal, and EndA presents differential structural features among its family members, like the presence of a ‘finger-loop’ interrupting the α-helix D, whose function is unknown. Both Mg2+ and Mn2+ can act as cofactors, while other divalent metal-ions, like Ca2+ and Zn2+, fail to support maximal nuclease activity. Mutagenesis analysis has identified several critical residues for EndA activity, including His160, Glu205 or Arg127/128, which are fundamental for nucleic acid binding and catalysis [151]. EndA is an extracellular nuclease that can be membrane-associated or secreted, in line with its signal peptide, and its production varies in vitro according to the growth rate. It plays a fundamental role in genetic transformation by degrading extracellular DNA and therefore facilitating its uptake. However, EndA activity is not dependent on the competence status. While competence-dependent activity is rapid but weak and transient, competence-independent activity has been shown to be involved in immune evasion during infection by degrading NETs. In this manner, EndA acts as a virulence factor that increases invasiveness and leads to reduced survival in murine models [152,153].

4.3. Streptococcus pyogenes (S. pyogenes)

4.3.1. Significance and current diagnostic methods

S. pyogenes, also known as group A streptococcus (GAS), colonizes epithelial surfaces of the skin, throat, vagina and rectum. Transmission is either airborne, foodborne or through direct contact with infected or colonized individuals. It is responsible for a wide range of human diseases that range from uncomplicated infections, such as pharyngitis or impetigo, to life-threatening invasive diseases, including cellulitis, bacteremia, streptococcal toxic shock syndrome, pneumonia, necrotizing fasciitis, meningitis, septic arthritis or osteomyelitis [128,154]. Additionally, it also triggers severe post-infectious immune sequelae, including acute glomerulonephritis or acute rheumatic fever, which can develop into rheumatic heart disease (RHD). It is therefore not surprising that the global burden of GAS disease is substantial. For reference, global death estimates of invasive GAS and RHD amount to 163,000 and 233,000 deaths per year, respectively, while the global prevalence of pyoderma in children and the global annual incidence of pharyngitis are 111 million and over 600 million, respectively [155]. A vaccine is not yet available despite continued efforts [154].

Rapid and accurate identification that allows prompt and adequate treatment is key to reducing immune sequelae and mortality associated with invasive disease. Like S. pneumonia, the identification of GAS relies heavily on microbial culture, which involves lengthy incubation times (24 to 48 hours), is highly dependent on specimen collection and incubation conditions; is prone to interferences from contaminating flora and does not allow discrimination between colonizing and invading organisms. Traditional phenotypic identification tests using different traits such as morphology, Gram staining, beta-hemolysis, Lancefield group (serologic test), bacitracin susceptibility or presence of pyrrolidonyl aminopeptidase (PYR test) need to be used in combination to avoid misidentification [156].

Fortunately, existing non-culture-based rapid and specific antigen test for the detection of S. pyogenes from throat swabs help to guide treatment and reduce antimicrobial misuse. However, culture-based confirmation in case of a negative result is necessary due to low sensitivities. Moreover, false positives may occur in colonized or previously infected patients. Rapid molecular methods based on probe hybridization and PCR techniques offer similar accuracy to culture-based diagnosis while providing same-day results. However, they cannot differentiate viable from non-viable or colonizing from invading organisms and some of them are not true POCT. In general, rapid tests are also highly dependent on sampling quality and disease severity and cannot provide antimicrobial susceptibility information [157–159]. At the same time, commercial automated platforms using a suit of physiologic tests, mass spectrometry or PCR offer accurate identification of S. pyogenes from other specimen types, such as positive blood cultures or culture isolated organisms. In the case of immune sequelae, diagnosis relies on monitoring, through neutralization assays, the antibody response toward streptolysin O and DNase B, the latter being a more specific marker of S. pyogenes infection [156].

4.3.2. Nuclease candidates

S. pyogenes is a prolific producer of extracellular nucleases. To date, 8 nucleases have been identified [160]. On the one hand, SpdB and SpnA are chromosomally encoded nucleases. SpdB, which is also known as DnaseB, MF, SdaB or SpeF, is a secreted nuclease and is monocistronically encoded by the mf gene. Catalytically, SpdB is a metal-ion-dependent, thermostable endonuclease capable of degrading both ssDNA and dsDNA as well as RNA. The type of cleavage is analogous to the one displayed by DNase I, rendering oligonucleotide products with 5´- phosphorylated terminus. Activity is supported by a wide range of cofactors, including Mg2+, Ca2+, Mn2+, Sr2+, Cd2+, Cu2+, Co2+, Ni2+, Zn2+ and Fe2+. It optimally operates at a very narrow pH level range of around 9,5, and its activity is both inhibited by EDTA and ionic salts, including NaCl and KCl. The thermal stability is biphasic, as the activity decreases at 60°C, but is restored at 80°C, and it persists after prolonged thermal stress at 100°C (up to 1 hour). SpdB is specific to GAS, being present across hundreds of clinical isolates of S. pyogenes. Interestingly, antibodies against SpdB have been detected in patients suffering invasive GAS infections. However, despite antibodies raised against SpdB neutralize nuclease activity, no difference in virulence was detected in an in vivo murine model of sepsis between infection with a wild type and a nuclease deficient S. pyogenes strain [161–163]. SpnA is the only cell wall-anchored nuclease of S.pyogenes, despite it has also been reported to be secreted, that has been detected in a multitude of S. pyogenes strains belonging to different M types and displaying varying clinical features. Structurally, SpnA is a 910 amino acid (aa) nuclease possessing both an N-terminal signal peptide and a C-terminal cell wall-anchoring domain containing a LPXTG motif. Additionally, it contains three N-terminal OB-folds of which at least two of them are required for its nuclease activity. In silico analysis and directed mutagenesis analysis have identified critical amino acids involved in catalysis, like Asp810/D810 and cofactor binding, like Glu592/E592. In a similar fashion to SpdB, SpnA is a metal-ion nuclease capable of degrading RNA and both ssDNA and dsDNA, including both plasmid and chromosomal forms, like NETs. However, SpnA can operate in a wider range of pH (5 to 7,5) and its activity is dependent on the presence of both Mg2+ and Ca2+ cofactors. The latter has also been shown to increase its structural integrity, enhancing the resistance to protease degradation. Activity neutralizing antibodies against SpnA can also be found in patients´ convalescent plasma at higher frequencies compared to healthy donors, potentially serving as serological markers for acute rheumatic fever [164]. However, opposite to SpdB, SpnA has been reported as a key virulence factor, which promotes bacterial survival and helps to evade extracellular responses of the innate immune system, like NETs, during infection. Yet, Chalmers et al. recently reported that the nuclease activity of SpnA, despite necessary, is not the only factor contributing to the the SpnA associated virulence of S. pyogenes [164–166].

On the other hand, Spd1/MF2, Spd3/MF3 [167], Spd4/MF4, Sda1, Sda2/SdaD/DnaseD and Sdn/Sdα are homologous secreted extracellular nucleases of chimeric nature encoded by prophage-like elements integrated into the bacterial chromosome, of which the most abundant among S. pyogenes strains are Spd1 and Spd3 [160,167]. Spd1 is a divalent metal-ion-dependent nuclease, belonging to the family of nucleases containing a ββα-metal finger motif, capable of degrading ssDNA, dsDNA and RNA in the presence of Mg2+ and Ca2+, that shares more than 50% sequence identity with Spd3 and structural homology with EndA, with which it even shares the characteristic ‘finger loop’ interruption feature, apart from the conserved DRGH motif [168]. Spd1 is co-transcribed with SpeC, which codes for a known superantigen, and their expression is associated with lysogenic prophage induction and it is stimulated by a yet unknown factor derived from pharyngeal cells. Indeed, it has been suggested that Spd1 activity could enhance the fitness of both the bacteria and the phage [169]. However, the expression of prophage encoded nucleases is not necessarily accompanied by phage induction and can be regulated by chromosomally encoded regulators, such as the Rgg transcriptional regulator, as it is the case for Spd3 [170]. Functionally, Spd1 has been connected to enhanced bacterial shedding upon infection, but not to increase invasiveness, despite the discovery of a significant association between an unusual upsurge in invasive disease and a S. pyogenes emm3 lineage characterized by the gain of the prophage element coding for Spd1 and SpeC [171,172]. The other nucleases are only present in specific emm/M types. For example, Spd4 has only been identified in S. pyogenes strains of emm3 and emm5 types [160]. Sdn (Sdα), a divalent metal-ion-dependent DNase sharing 97% sequence identity to the putative Dnase of S.equisimilis (SdC), is slightly more prolific being present across strains belonging to 11 different emm/M types and isolates associated with different clinical manifestations [160,173]. The exact contribution, if any, of most of these prophage-encoded nucleases to S. pyogenes pathogenesis remains unclear. Different studies have shed some light on the topic, but further research is due. For example, the presence of Sda2-specific antibodies in patient´s sera indicates its production during infection, suggesting a role in virulence [174]. Furthermore, the nuclease activity of Sda1, a potent ββα-metal, ion-dependent DNase that shares homology with SdaD, Sdn, EndA and NucA among others, is one of three nucleases (SpdB and Spd3) produced by the globally distributed, highly virulent M1T1 S. pyogenes clone [175,176], and it has been reported to contribute to its immune evasion by promoting escape from NET and eluding TLR9-dependent production of proinflammatory cytokines and macrophage-mediated killing [177,178].

4.4. Streptococcus agalactiae (S. agalactiae)

4.4.1. Significance and current diagnostic methods

S. agalactiae, also known as group B streptococcus (GBS), is a pathobiont that typically colonizes the gastrointestinal and genitourinary tracts and may cause invasive disease in the form of soft tissue and skin infections, pneumonia, urinary tract infections, bacteremia, endocarditis, osteomyelitis, septic arthritis or meningitis. Increasing incidence among non-pregnant adults has been reported, with immunocompromised and elderly populations being the most susceptible. However, GBS is best known for its role in fetal, neonatal, and maternal invasive disease. In fact, 57,000 stillbirths, 319,000 cases of neonatal disease, 33,000 cases of maternal disease and up to 3,5 million preterm births were estimated to be attributable to GBS in 2015 [179]. Colonization, which can be transient, intermittent, or persistent is a pre-requisite for invasive GBS disease and 18% of women are estimated to be asymptomatically colonized, although the rates and serotype prevalence vary geographically. Intravenous intrapartum antibiotic prophylaxis in colonized women has dramatically reduced the incidence of early onset neonatal disease, but late onset neonatal disease remains refractive to it. Interestingly, reduced odds of maternal colonization by GBS during pregnancy has been associated with higher antibody titers against immunogenic surface proteins [180]. As such, the ongoing development of high coverage vaccines holds promise to protect against maternal, fetal and both types of neonatal disease [179,181,182].

Diagnosis of GBS also relies heavily on culture-dependent methodologies impeding rapid diagnosis, which compromises the adequate management of severe invasive infections and limits preventive screening efforts in pre-partum pregnant women. In the latter case, collection of vaginal and rectal swabs is typically followed by broth enrichment and colony isolation in agar and subsequent identification. Traditional phenotypic methods are based on the detection of beta-hemolysis, granadaene pigment or specific enzyme systems using specialized agar media; the detection of CAMP factor (CAMP test) or the detection of sodium hippurate hydrolysis. However, the existence of non-hemolytic, non-pigmented isolates and the existence of CAMP positive GAS, compromise their specificity. Immunoassays based on the determination of the Lancefield B group are highly specific. Rapid antigen detection tests have been developed but lack clinical impact due to their low sensitivity. Aiming to increase diagnostic speed while maintaining high accuracy, different molecular methods based on MALDI-TOF MS and PCR have been developed. In fact, PCR-based POCT exist, amid associated with excessive costs, that in combination with intrapartum prophylaxis could help further reduce the incidence of GBS disease in neonates [183,184].

4.4.2. Nuclease candidates

In silico analysis of S. agalactiae genome reveals the presence of 7 genes encoding secreted DNases [185]. One of these genes, gbs0661, codes for NucA, the major extracellular secreted nuclease of S. agalactiae. NucA sequence bares a high degree of identity with the nucleases from S. pneumoniae (EndA), S. pyogenes (Sda1) and Serratia marcescens (NucA). Similarly to other extracellular nucleases, it possesses a signal peptide sequence. In its mature, secreted form, NucA is a 25 kDa metal-ion-dependent nuclease capable of degrading ssDNA and dsDNA PCR products, close and open forms of plasmid DNA, chromosomal DNA, and RNA. The nuclease activity can be activated by different divalent cations, such as Mg2+, Mn2+, Zn2+, Ni2+ or Cu2+, and it is both thermostable and pH-stable, with activity observed in a wide range of temperatures (4°C and 95°C) and pH levels (5–8). NucA belongs to the family of nucleases containing a ββα-metal finger motif. It also possesses a H-N-N motif, whose histidine residue is fundamental for the nuclease activity, which incidentally is inhibited by the presence of chelating agents. Interestingly, extracellular nuclease activity is enhanced in the absence of glucose and during respiration permissive growth. It has also been shown using both in vitro and in vivo models that NucA protects S. agalactiae from NETs, prevents immune clearance during the early stages of infection and enhances its virulence, playing a role in persistence and dissemination in later stages of infection [185].

4.5. Escherichia coli (E. coli)

4.5.1. Significance and current diagnostic methods

Pathogenic E. coli can be subdivided into intestinal pathogenic E. coli (IPEC) and extra-intestinal pathogenic E. coli (ExPEC) [186]. IPEC are classified into different pathotypes according to phenotypic and genotypic traits and include: enteropathogenic E. coli (EPEC), shiga toxin-producing E. coli (STEC), enteroaggregative E. coli (EAEC), enterotoxigenic E. coli (ETEC), shigella/Enteroinvasive E. coli (EIEC), diffusely adherent E. coli (DAEC) and adherent invasive E. coli (AIED). IPEC are the causative agent of enteric infections and are prone to cause epidemics or pandemics. Generally, they manifest as self-limiting diarrheas with different levels of severity that may develop into complications such as persistent diarrhea (e.g. EPEC), hemolytic-uremic syndrome (e.g. STEC/shigella), growth abnormalities in children (e.g. ETEC), intestinal perforations (e.g. STEC) or increase risk for inflammatory intestinal diseases (e.g. EAEC, DAEC and AIEC), among others [186,187].

Clinical guidelines for diarrheagenic diseases recommend the use of routine culturing techniques or NAATs for pathogen identification, or the use of exploratory panel-based multiplex RT-qPCR assays when enteric fever or bacteremia is suspected [188]. Biochemical identification is still commonly used, though sometimes it may lack specificity (e.g. Shigella vs EIEC), differential sensitivity (e.g. O157 STEC vs SFO157:NM STEC) or the ability to distinguish pathogenic from commensal bacteria (e.g. ETEC). Rapid probe hybridization or PCR-based pathogen identification for pathotype-specific identification is becoming common. However, the genomic variability of E. coli may affect the accuracy, due to the lack of genetic determinants (AIEC), loss of determinants in vivo (e.g. STEC) or in vitro (e.g. ETEC), between-strain cross reactivity or inability to distinguish pathogenic from non-pathogenic strains (e.g. EAEC/DAEC). Immunological assays based on EIA and LFA are used for the rapid detection of toxins in ETEC or STEC, such as stx1 and stx2. In general, these rapid techniques tend to require specimen enrichment to reach acceptable sensitivity levels. Subspecies typing is important in the control of outbreaks and the gold standard technique is pulse-field gel electrophoresis (PFGE), though approaches based on sequencing, immunoassays, NAAT or MALDI-TOF, exist both for clinical and food samples [187,189].

ExPEC pathotypes include uropathogenic (UPEC), meningitis associated (MNEC) and septicemia associated (SEPEC) E. coli. Strains belonging to these groups are responsible for UTI, sepsis and meningitis, skin infections, myositis, osteomyelitis, surgical site infections and hospital-acquired pneumonias, among others. In fact, E. coli is one of the most prevalent causes of both complicated and uncomplicated UTI, neonatal sepsis and meningitis. E. coli can rapidly modify their pathogenic phenotype, even during infections, and incorporate new features, such as antibiotic resistance plasmids, which make the diagnosis and treatment challenging [190–192].

Pathogen identification and characterization in ExPEC infections, such as UTI or bacteremias, rely heavily on culture-based enrichment that ranges from 24 h in urine cultures to ≥ 5 days for blood cultures. For UTI, cumbersome and insensitive microscopic techniques as well as rapid, insensitive biochemical urinalysis tests with low positive predictive values and unable to provide pathogen identification exist. After enrichment, urine culture results are interpreted based on culture-isolated microorganisms and colony counts. Identification and AST from positive blood cultures or isolated colonies can be performed within hours using commercially available platforms based on rapid molecular methods, such as microarrays, probe hybridization assays, automated panel-based RT-qPCR systems or MALDI-TOF [134,193,194].

4.5.2. Nuclease candidates

Endonuclease I is a 26,7 kDa, 235 aa long, chromosomally encoded, metal-ion-dependent (Mg2+), periplasmic nuclease expressed by E. coli under the control of a weak promoter. It can cleave both ssDNA and dsDNA and displays sequence specificity [195]. It has a 22 aa signal peptide, in line with its periplasmic localization, and bares sequence similarity with the endonuclease from V. cholera (see below) and Aeromonas hydrophila, which also belong to the EndA/NucM nuclease family [196]. Interestingly, Endonuclease I is competitively inhibited by different RNA species, such as tRNA, which upon complexing with the nuclease and in high salt concentrations shifts its endonucleolytic activity profile, from a double-strand break activity to a single-strand nicking activity [197].

4.6. Neisseria gonorrhoeae (N. gonorrhoeae)

4.6.1. Significance and current diagnostic methods

N. gonorrhoeae causes infections in the mucosal epithelia of the urogenital tract, rectum, pharynx and conjunctiva. The most common manifestations are urethritis in men and cervicitis and urethritis in women. N. gonorrhoeae infections increase the risk for sexually transmitted infections, like HIV, and increase the risk of suffering serious complications in women, such as chronic pelvic pain, ectopic pregnancy, and infertility. Unfortunately, 66% of men and 50% of infected women are asymptomatic at any given time which facilitates unwilling horizontal and vertical transmission [198–200]. Data estimates from 2016, indicate that the global prevalence varied between 0.9% for women and 0.7% for men aged 15 to 49, and it was higher in lower income countries. Meanwhile, the incidence was estimated at 86.9 million, a noticeable increase over the estimated 78 million in an identical study from 2012 [201,202]. Despite its high incidence and prevalence, there is no available vaccine and the appearance of fit multi-resistant strains have literally dried out treatment options [198].

Currently, microscopic examination of Gram-stained smears is the method of choice for low resource settings due to their low cost. However, their application is limited to urogenital specimens, due to low sensitivities when using oral and rectal specimens. Even then, their sensitivity varies depending on the type of specimens and is significantly reduced in asymptomatic patients, which is likely due to low bacterial loads. NAATs have become the new gold standard and have become a mainstay in high-income countries, with numerous commercially available options. NAATs offer rapid turnaround times and consistently high sensitivities and specificities across specimen types. However, detection of non-viable organisms after treatment, high costs and inability to determine antibiotic susceptibility are the main drawbacks. Culture is necessary for AST and is frequently used in combination with biochemical tests, immunological assays, NAATs or mass spectrometry to offer definitive pathogen identification. However, its accuracy depends on the time of collection after exposure, transport conditions or specimen type. Low cost, rapid, sensitive and specific POCT, which are capable of accurately diagnosing asymptomatic patients are desired for screening efforts. LFA or optical-based POCT exist but suffer from very poor sensitivity. NAAT-based POCT have been developed and continue to be improved [199].

4.6.2. Nuclease candidates

N. gonorrhoeae produces a divalent metal-ion-dependent thermonuclease, denoted Nuc, that bares 25% identity and 40% similarity with S. aureus´s MN. The enzyme is coded by the nuc gene, which is contained within a 7 gene operon. Nuc is capable of degrading ssDNA, supercoiled plasmid DNA and eukaryotic and prokaryotic genomic DNA; and its activity is inhibited by chelators (EDTA 4 mM). Interestingly, the nuclease shows significantly diminished activity towards methylated DNA, which suggest a regulatory role of methylation in its catalytic action. This is in line with observations that Nuc, similarly to MN, is involved in the remodeling and degradation of N. gonorrhoeae biofilms, which are densely packed with highly methylated DNA when compared to biofilms from other bacteria such as Lactobacillus spp. Nuclease activity has been detected both in association with the bacterial cell (e.g. bacterial lysates) and as part of the secreted proteome (e.g. conditioned media), the latter being in line with the identification of a predicted 34 aa secretion signal peptide in its open reading frame. Furthermore, among the many strategies used by N. gonorrhoeae to evade the innate and adaptative immune action [198], Nuc is used as a key virulent factor that increases bacterial survival upon colonization by degrading NETs [203,204].

4.7. Serratia marcescens (S. marcescens)

4.7.1. Significance and current diagnostic methods

S. marcescens causes a wide range of infections including urinary, respiratory, bloodstream and ocular infections [205]. S. marcescens thrives in acute care center environments (e.g. hospitals), helped by its ability to survive in disinfectants, such as chlorhexidine, for long periods and to easily contaminate sterilized plastic surfaces in the presence of water [206,207]. Numerous nosocomial outbreaks have been described in the literature, such as in neonatal, pediatric or neurosurgical wards [208,209], which in conjunction with antimicrobial resistant strains can render mortality of up to 40% [210]. In fact, according to the 2014 annual epidemiological report from the European Centre for Disease Prevention and Control, Serratia spp were the 10th most frequent cause of intensive care unit-acquired pneumonia and bloodstream infections, with similar levels to Enterobacter spp and Acinetobacter spp, respectively [211]. Additionally, high levels of S. marcescens infections with a community onset have also been described [212].

Clinical diagnosis of Serratia spp relies heavily on phenotypic assays. Serratia spp members are easily grown in culture from clinical specimens. Biochemical characterization is used to distinguish them from other species of the Enterobacteriacea family and can also be used for sub-species identification. Different panel-based automated commercial systems capable of identification and AST exist. Molecular methodologies, such as PFGE or sequencing, can be used for typing, which is important for monitoring nosocomial outbreaks [205,210].

4.7.2. Nuclease candidates

A trademark of S. marcescens, and of the Serratia spp in general, is the production of an extracellular nuclease, known as Sm nuclease or NucA, which is used for different biotechnological applications (e.g. DNA and RNA decontamination) and has even been proposed for therapeutic approaches [213,214]. Unsurprisingly, it has been thoroughly characterized [215–218] and comprehensively reviewed elsewhere [213]. In brief, Sm is a divalent metal-ion-dependent homodimeric nuclease that specifically cleaves the 3´end of a phosphodiester bond and is active in a wide range of pH levels (6 to 10) and temperatures (35 to 44°C). It degrades both ss/dsDNA and ss/dsRNA substrates at similar rates. As long as the substrates are longer than 5-mer, it has a high catalytic efficiency which is approximately 4 and 34 times higher than that of MN and DNase I, respectively. Moreover, it displays sequence preference towards d(G).d(C)-tracts versus d(A).d(T)-tracts in dsDNA substrates that diminishes in ssDNA substrates. The two isoforms of Sm nuclease (Sm1 and Sm2) are encoded by the nucA gene, whose transcription is regulated by the growth phase and by environmental cues through the bacterial SOS system. Both nucA and nucC, which codes for nucA´s transcriptional regulator NucC, are regulated by the growth phase. At the same time, both genes are regulated by the SOS system through the repression of LexA. Thus, nuclease production is regulated by the environment and directly proportional to growth. Both isoforms are structurally almost identical with Sm1 differing only in the length of their N-terminal tail by 3 aas in favor of Sm2. Despite this high degree of similarity small differences in sequence preference have been observed. Moreover, the Sm2 isoform is produced during exponential growth and immediately secreted extracellularly, while the Sm1 isoform is produced during logarithmic growth and stays longer in the periplasm before being secreted. Independently of the isoform, the oxidizing environment of the periplasm is needed to stabilize and activate the nuclease.

4.8. Helicobacter pylori (H. pylori)

4.8.1. Significance and current diagnostics methods

H. pylori has co-evolved with humans and it has developed strategies to cause chronic infections of the gastric mucosa and survive in such a harsh environment [219]. It is estimated that around 50% of the human population is positive for H. pylori, with levels as high as 80% in developing countries. Infection with H. pylori is characterized by a rarely symptomatic gastritis, that if symptomatic typically manifests as dyspepsia. Depending on several factors including genetic susceptibility, immune response, level of acid production and type of strain; infections may develop into a chronic gastritis, ulcers, and if not eradicated, significant risk of developing progressive tissue atrophy and gastric cancer [220]. In fact, H. pylori has been labelled as a type I carcinogen and it was estimated to be responsible for 89% of all non-cardia gastric cancers worldwide in 2012 [9]. Despite combinatorial antibiotic therapies for eradication are available, resistance is becoming worrying, with a recent systematic review reporting estimated primary and secondary resistance rates higher than 15% for most antibiotics, and a worrying upwards trend [221]. Additionally, recent associations between H. pylori and impaired cognitive function in adults [222], as well as a suggested role of the infection in the pathophysiology of Alzheimer’s disease and patients’ outcomes have been reported [223–225], though contradictory reports exist [226].

Several diagnostic methodologies are available, and their implementation depends on the clinical presentation and health-care setting. Serology and endoscopy, which are followed by biopsy culture and rapid urease test or histological evaluation, are used in acute care centers. Endoscopy has great sensitivity and specificity, and it is able to provide ancillary information on disease status. Narrow band imaging or confocal laser endomicroscopy facilitate targeted biopsies and lesion grading. The most common and accurate non-invasive methods are based on detecting the bacteria´s urease activity (urea breath test) or specific antigens in stool specimens (immunoassays). Serology is sometimes used, but it is not recommended due to poor sensitivity derived from variable rates of seroconversion and inaccuracy in confirming eradication. Molecular tests allow pathogen detection from cultures, biopsies or stool specimens and display higher rates of positivity than culture. Real-time PCR targeting 23S ribosomal RNA is a commonly employed method and allows pathogen detection from biopsies with low bacterial loads, though detection of non-viable organisms (false positives) is always a concern. The use of isothermal amplification and digital PCR has been explored and offer simpler and quantitative assays, respectively. Molecular tests are also becoming a suitable alternative to traditional phenotypic AST due to their high sensitivity and rapid turnaround times (hours). These include PCR-based and in-situ hybridization methods for the identification of mutations associated with resistance directly from biopsy specimens. However, in some cases, they suffer from lack of standardization [220,227,228].

4.8.2. Nuclease candidates

Nuclease activity has previously been reported for 50 different strains of H. pylori [229]; however, O´Rourke et al. was the first to identify, purify and characterize NucT, the main H. pylori nuclease, as shown by a 200-fold reduction in the activity of bacterial extracts from insertional mutants. NucT belongs to the phospholipase D nuclease sub-family and it is a 17.75 kDa membrane-associated thermostable endonuclease that shares high structural homology with its family counterparts Nuc (Salmonella enterica serovar Thyphimurium) and Zucchini endonucleases. NucT is encoded by the monocistronically transcribed hp0323 gene, which also codes for a 23 aa periplasmic peptide signal sequence. NucT shows optimal catalytic activity at 80°C and pH 8, though it can work at body temperature (37°C), albeit with a > 20-fold reduction in activity. It is not inhibited by chelating agents (10 mM EDTA) and it is enhanced by reducing agents (DTT or β-mercaptoethanol), inhibited by oxidizing agents (glutathione) and neither dependent nor inhibited by divalent cations at 1 mM (data reported but not shown). Structurally it consists of a homodimer that conforms a deep and narrow positively charged groove that binds with high affinity and catalyzes the degradation of RNA, ssDNA, dsDNA and different forms (supercoiled, circular or linear) of plasmid DNA, showing a clear preference toward DNA over RNA and for ssDNA over dsDNA substrates. Catalytic activity is entirely dependent on the H124 aa and possesses a positively charged loop that is not shared by its nuclease sub-family counterparts. This loop represents a variable region among the identified NucT homologues in the Helicobacteraceae family, suggesting differential substrate affinities among them. NucT has been shown to be involved in the natural state of competence for transformation, with 10 and 100-fold reductions in plasmid and chromosomal DNA transformation efficiency, respectively, in hp0323 mutants. Moreover, NucT is also fundamental for bacterial nutrition and growth thanks to the nuclease’s ability to scavenge and process environmental DNA into purines [230–232].

4.9. Mycobacterium tuberculosis (M. tuberculosis)

4.9.1. Significance and Current diagnostic methods

M. tuberculosis is the causative agent of tuberculosis (TB), an airborne, not highly infectious, treatable disease that primarily affects the lungs, though lymphatic or hematogenous spread of M. tuberculosis can cause infections in other tissues and organs (extrapulmonary TB), such as bones or kidneys [233]. Despite a decreasing trend, the global burden of TB is mighty, especially in infants and immunocompromised patients. In fact, it became the 6th and 4th cause of death in 2013 among young adults aged 15 to 19 and 20 to 24, respectively, and it was responsible for 1,3 million deaths in 2016 [1,3]. Moreover, the increasing prevalence of antibiotic resistance, especially in India and China, limits the therapeutic options and, in some cases, it renders the disease incurable [234,235]. TB is a capricious disease as it can disproportionately manifest as an innate or adaptative immune-system-cleared asymptomatic infection; as a non-transmissible, latent infection; or as an asymptomatic, mildly symptomatic or symptomatic transmissible active infection that may develop into a life-threatening condition if untreated. Preventive strategies are currently limited to the unreliable BCG vaccine [236].

On the one hand, the tuberculin skin tests (TST) and the interferon-γ release assay (IGRA) are used to identify TB during the latent phase of the disease. These assays can identify current infection or previous exposure by measuring cell-mediated immune responses to tuberculin or region-of-difference 1 encoded antigens. They are widely used, especially in low resource settings, due to their few requirements in terms of equipment, expertise and costs. Limitations include the inability to distinguish between past and present infection or between latent and active disease, low predictive value for progression, low levels of reproducibility that deems them unfit for screening purposes, and reduced sensitivity in immunocompromised patients. Moreover, in the case of TST, specificity is compromised in patients having received the Bacillus-Calmette-Guerin vaccine and in patients having infections caused by non-tuberculous mycobacteria [237].

On the other hand, diagnostic tools for the identification of active TB include imaging modalities such as X-ray or PET-CT, microscopic examination of sputum smears, culturing methods, LFA for antigen detection in urine and isothermal or PCR-based NAATs that allow sensitive and specific diagnosis and resistance determination from sputum specimens. Due to the very low sensitivity and specificity of microscopic examinations, the high costs and low specificity of imaging modalities, the lack of sensitivity of immunoassays and the prolonged waiting times associated with culturing methods, qPCR-based NAATs are currently recommended as the first-line diagnostic test, despite high costs and complexity, the inability to monitor treatment efficacy and the reliance on genotypic resistance [236].

4.9.2. Nuclease candidates

Recently, Rv0888 has been described as the first extracellular nuclease produced by M. tuberculosis. Rv0888 had been previously described as an outer membrane-associated extracellular sphingomyelinase, whose expression is highly upregulated in the presence of sphingomyelin, a lipid found in eukaryotic cells; with roles in M. tuberculosis nutrition and intracellular persistence in macrophages [238]. Dang et al. further characterized Rv0888 as a divalent metal-ion-dependent nuclease encoded by rv0888, with a predicted 31 aa signal sequence and whose D438 residue is fundamental for its catalytic activity. Despite its ability to degrade both circular, linear and chromosomal DNA, it shows higher substrate preference towards RNA. Rv0888 operates optimally at 41°C and pH of 6,5 in the presence of Mn2+ and Ca2+, though its relative activity remains higher than 50% when operating at different ranges of temperature (39°C to 45°C) and pH levels (6,0 to 8,0) and with the presence of different ions (Mg2+ + Mn2+; Ca2++ Mg2+; Mn+2 or Ca+2). A possible key role of this nuclease in the pathogenesis of M. tuberculosis has been suggested. Using a murine model of lung infection, Dang et al. showed that rv0888-expressing Mycobacterium smegmatis significantly increased persistence of infection and induced noticeable pathological changes in the lung tissue compared to controls [239]. However, these results oppose the observations by Speer et al. of an unaltered persistence of both wild type M. tuberculosis and a mutant lacking rv0888 in an in vivo murine model of lung infection. These differences are most likely due to variations in the model, such as the mouse strains (C57BL/6 and BALB/c), the bacterial species used (M. tuberculosis versus M. smegmatis) and the levels of Rv0888 expression (wild-type expression versus overexpression), respectively [238,239].

4.10. Foodborne bacterial pathogens

4.10.1. Significance and current diagnostic methods

Salmonella enterica (S. enterica), Campylobacter jejuni (C. jejuni), V. cholera, Yersinia enterocolitica (Y. enterocolitica) and Clostridium perfringens (C. perfringens) are some of the most frequent bacterial agents causing enteric disease. S. enterica generally causes self-limited gastroenteritis and in some cases self-promoted phagocytosis by macrophages leads to systemic infections. The emergence of antibiotic resistance in common serotypes (serovar Typhimirium and Enteriditis) not only complicates treatment but leads to more severe presentations [240–242]. C. jejuni can invade the lower intestinal tract causing enterocolitis and very rarely sepsis. However, it increases the risk for inflammatory bowel disease, and it may lead to extraintestinal manifestations, including immune disorders, such as Guillain-Barré syndrome or reactive arthritis [128,243,244]. V. cholera´s serotypes (O1 and O139) are the causative agents of cholera, an enterotoxin-mediated profuse watery diarrhea that can be life-threatening if left untreated [128,245]. Y. enterocolitica typically causes enterocolitis by invading the proximal colon, from where it can invade the lymph nodes or spread to other organs (blood, lungs or heart) [246]. C. perfringens causes enterotoxin-associated, self-limiting food poisoning. It is also known to cause toxin-α-mediated myonecrosis (gas gangrene) in deep wounds [247]. Despite typically causing mildly symptomatic self-limiting disease, these are some of the most burdensome, deadly and widespread waterborne and foodborne pathogens both in low- and high-income countries due to their epidemiological status [5,248,249].

Traditional methodologies both in the clinic and in industry have typically relied on amplification techniques based on culturing methods. These consist of pre-enrichment and enrichment steps using specialized culturing media to isolate the pathogens in vitro, which are then followed by biochemical, morphological and serological characterization for identification. These traditional methodologies may suffer from low sensitivity depending on the pathogen and specimen matrix and are laborious and time-consuming, usually requiring days for preliminary results and up to a week for confirmation. Despite their drawbacks, these methods are still considered the gold standard due their low cost and high selectivity, among other variables. In fact, clinical diagnostic guidelines for enteric pathogens (S. enterica, Campylobacter spp or Yersinia spp) [188], as well as different international standard methodologies for the detection of foodborne pathogens, such as ISO 10273:2003 (Y. enterocolitica) [250] or ISO 6579:2002/Amd 1:2007 (Salmonella spp) [251], are heavily based on culture-based methods.

The development of rapid molecular and immune-based methodologies, and their commercial implementation, have advanced the field. qPCR-based NAATs or DNA microarrays are now capable of rapid and, in some cases, multiplexed detection of foodborne pathogens with low detection limits from pure cultures; identification of multiple serotypes; or evaluation of antibiotic susceptibility. Similarly, immune assays, such as EIA and LFA, despite their lower sensitivity, offer low cost, rapid and specific detection of pathogen-associated toxins in different food matrices. Furthermore, a small number of biosensor devices based on optical and electrochemical transduction techniques have passed agency validation and are now being commercialized for the detection of Salmonella spp or Campylobacter spp, promising very low detection limits without sample pre-treatment or enrichment. However, the inconvenient truth is that most of these rapid techniques still rely on culture-dependent enrichment and isolation ranging from 6 h to 48 h depending on the specimen matrix, pathogen or technique. Furthermore, other drawbacks include high costs, complexity and high rate of false positives for NAATs, or high cross reactivity and elevated false-negative rates for immune-based assays. New approaches based on the flexibility of biosensing techniques, including one using the activity of nucleases as a biomarker [32], have been explored and are reviewed elsewhere [18,252–255]. However, a reduced level of industry adoption of biosensing techniques due to issues with quality assurance, stability and calibration is a concern [18,132,256].

4.10.2. Candidate nucleases

4.10.2.1. Salmonella enterica serovar Typhimurium (S. enterica ser. Typhimurium)

. Salmonellas´s Nuc nuclease is a member of the phospholipase D (PLD) family of proteins, and despite a low sequence identity, it bears a predicted strong structural homology with NucT nuclease from H. pylori [232]. Nuc, however, is one of the few proteins of the PLD family to only possess a single copy of the invariant motif (HxK(x)4D(x)6GSxN) that characterizes the family. Nuc is encoded by the drug resistance pKM101 plasmid, which has roles in mutagenesis and survival, as well as by mutant plasmid derivatives (pGW12, pGW21 and pGW46) present in different strains of S. enterica ser. Typhimurium. Like NucT, Nuc possesses a 23 aa signal sequence [257,258]. Characterization of its activity first by Lackey et al. [257] and then by Zhao et al. [258], describe Nuc as a periplasmic associated endonuclease that can cleave both ssDNA and dsDNA to render 3’-OH/5’-P products, equally to Serratia nuclease. Interestingly, it exhibits different optimal pH conditions for catalysis depending on the nature of the substrate (ssDNA or dsDNA). However, some discrepancies exist between the reported catalytic activity. Lackey et al. observed a total dependence on divalent cations, including Mg2+, Ca2+, Zn2+ and Co2+, with optimal conditions requiring 10 mM of Mg2+, which is consistent with the identified roles of Mg2+, Ca2+ and Mn2+ as essential cofactors of S. enterica ser. Typhimurium´s global nuclease activity by Machado et al. [32]. However, Zhao et al. claimed (not showing the data) that divalent cations have no influence in the catalytic process and added that activity is not inhibited by the presence of chelating agents (1 mM EDTA), paralleling its homologue NucT. Discrepancies aside, Lackey et al. reported a slight sequence preference towards adenine nucleobases, which matches the purine salvage function attributed to NucT [231]. In this line, Machado et al. also observed a sequence preference of S. enterica ser. Typhimurium´s global extracellular nuclease blueprint for purine nucleobases, especially adenine bases, as well as a lack of inhibition by chelating agents, like EGTA and NTA (5 mM), but not for 5 mM EDTA. Overall, these observations suggest that the nuclease blueprint of S. enterica ser. Typhimurium may be dominated by the activity of the Nuc endonuclease and that NucT and Nuc share very similar catalytic properties, as already reported [232].

4.10.2.2. Campylobacter jejuni (C. jejuni)

Lior et al. reported abundant DNase activity among Campylobacter spp, including C.jejuni and C.coli. In particular, 137 out of 272 different strains of C. jejuni demonstrated DNase activity [229], as assessed by using toluidine blue DNA agar [94]. This is harmonious with the fundamental role of DNases in the natural competence for transformation of C. jejuni strains. Gaasbeek et al. showed that DNase activity is inversely correlated with transformation efficiency. Three extracellular nucleases encoded by the prophage integrated elements CJIE1 (dns), CJIE2 and CJIE4 (CJE0566 and CJE1441) were found to be responsible for this activity in different C. jejuni strains. Dns displays sequence similarity to other nucleases such as EndA (E. coli) or Dns/VcEndA (V. cholera) and it is also predicted to possess a typical endonuclease I domain. CJE0566 and CJE1441 nucleases contain a NUC superfamily domain and a DRGH motif and show structural similarity to the ββα-metal nuclease NucA from Anabaena spp, which is itself structurally homologous to NucA from S. marcescens [259,260]. All three nucleases contain a signal sequence, and their activity has been measured in enriched periplasmic fractions (Dns) and live bacteria treated with polymyxin B (CJE0566 and CJE1441), suggesting a periplasmic or secreted localization [259,261].

4.10.2.3. Vibrio cholera (V. cholera)

. V. cholera possesses two well-known nucleases (Dns and Xds) initially reported by Newland et al. [262] and Focareta et al. [263,264]. Since then both proteins have been crystallized and comprehensively characterized. In brief, Dns (also known as VcEndA) is a 24,7 kDa extracellular metal-ion-dependent endonuclease belonging to the Endonuclease I superfamily and encoded by the endA gene. Its monomeric structure is stabilized by NaCl [49] and exhibits a Mg2+ binding site (Glu79) with pH-dependent affinity within the characteristic ββα-metal finger motif and next to the catalytic residue (Gly 80) [265]. Consequently, in low pH (< 6), the nuclease activity is absent due to the lack of Mg2+ binding [266]. Dns shows good thermostability, preserving up to 50% of its activity after incubation at 70°C for 30 minutes. It shows very poor activity toward RNA substrates, especially at higher NaCl concentrations. However, it is capable of degrading dsDNA in linear and circular forms and ssDNA. It exhibits optimal activity at 50°C, pH level of 7,5 to 8 and in the presence of NaCl (175 mM) [48]. Meanwhile, Xds is a 94,3 kDa extracellular metal-ion-dependent strict exonuclease that possesses a 28 aa signal sequence and three characteristic domains (LTD, OB and EEP), with the OB domain being indispensable for nuclease activity. It shows optimal activity at low temperatures (<25°C), low NaCl concentrations (0 to 100 mM), neutral pH levels (7 to 8) and in the presence of Mg2+ (10 mM) and Ca2+ (20 mM), the latter being a requirement for its activity. Activity is unaffected in reducing conditions (1,4 dithiotreitol (DTT)). Being a strict exonuclease, it can only degrade linear forms of DNA, and it has been shown to preferentially cleave AT-rich over GC-rich dsDNA substrates [267]. These two extracellular nucleases act cooperatively to regulate and control natural transformation and biofilm formation, the latter having important roles in the colonization efficiency. They are also involved in immune evasion through NET degradation, and phosphate sourcing by degradation of extracellular DNA. In all these functions Dns has a more prominent role than Xds, which usually presents a more residual, but nevertheless complementary role. Expression of both nucleases, though at different levels, is induced by phosphate limiting conditions and by extracellular DNA, including DNA from NETs. Interestingly, their transcription is governed by different regulatory mechanisms. While Dns has been shown to be under the transcriptional control of the quorum sensing regulator HapR, which not only represses the transcription of Dns but also acts as a negative regulator of the exopolysaccharide (VPS) synthesis required for biofilm formation; Xds is independent of HapR. This may explain the differences in temporal expression patterns during biofilm formation and it is in line with the cell density-dependent Dns expression that controls transformability [268–270].

4.10.2.4. Yersinia enterocolitica (Y. enterocolitica)

Nakajima et al. observed the degradation of PCR products in crude-boiled bacterial extracts of both pathogenic and non-pathogenic clinical and environmental isolates of Y. enterocolitica, but not in extracts of Y. pseudotuberculosis isolates. This observations suggested the presence of a thermoresistant DNase activity associated to Y. enterocolitica that could persist at 4°C and was inhibited by EDTA and proteinase K [271]. Nakajima´s reported DNase activity could have been attributed to the action of a thermonuclease identified and characterized by Shi´s group in Y. enterocolitica subspecies palearctica, a common foodborne pathogen [272]. Bioinformatic analysis showed that this 283 aa nuclease, referred to as YNSN, is a homologue of the nuclease produced by Y. enterocolitica subps. enterocolitica 8081 and it contains two functional domains with high homology to the NUC superfamily and Endonuclease_NS family, respectively. It has been reported to possess activity towards both DNA and RNA substrates, it has a predicted binding site for a divalent cofactor and it can operate in a wide range of temperature (reported activity at 37°C and 55°C) and resist thermal shock (30 min at 80°C). Experimentally, it displays equivalent levels of DNase activity to Serratia marcescens nuclease, but dissimilar to its homologous nuclease in Y. enterocolitica subspecies enterocolitica 8081. Its activity and thermostability depend on three key residues (Glu202, Ile203 and Asp264). The only predicted transmembrane domain of the protein coincides with its signal peptide, for which a cleavage site has also been predicted, which suggest a membrane-associated or secreted extracellular localization [273–275].

4.10.2.5. Clostridium perfringens (C. perfringens)

DNase activity has been reported in human clinical isolates [75] and in a high percentage of C. perfringens isolates collected from mammals and birds. Interestingly, DNase activity was most prevalent among animals presenting disease manifestations, suggesting a role of DNase as a virulence factor [276]. Okumura et al. identified a 193 kDa cell wall-anchored endonuclease (CadA) encoded by CPE1368 gene (cadA) under the transcriptional control of the VirR/VisS-VR-RNA system, which is a system known to control the expression of several virulence-related and toxin genes [277]. CadA possesses a characteristic cell wall anchoring motif characteristic of several streptococcal nucleases, such as SpnA from S. pyogenes, SsnA from Streptococcus suis or SWAN from Streptococcus sanguinis [160,277,278]. CadA shows metal-ion dependency (tested only in the presence or absence of Mg2+ and Ca2+) when degrading both plasmid and chromosomal DNA, however, it preserves nicking activity for plasmid DNA in their absence [277]. Besides, the reported existence of a CadA-unrelated, cell surface associated nuclease activity and the presence of detectable activity in culture supernatants are consistent with the existence of other extracellular nucleases [277].

5. Conclusions

To combat existing and arising challenges in infectious diseases advances in the field of diagnostics are key. The nature of biomarkers and the existing technology define the scope and the limitations of currently employed diagnostic tools. As such, the addition of new members to the catalog of available biomarkers that allow the development of diagnostic tools with the ability to complement, substitute or add new functionalities to the existing crop are likely to be welcomed by physicians and clinical microbiologists.

Nucleases are a diverse group of enzymatic proteins present across all domains of life that degrade nucleic acids. Due to their abundance, biochemical and catalytic diversity and involvement in fundamental biological roles, including pathogenesis; their activity shows promise as a novel diagnostic biomarker in bacterial infections. In fact, using different approaches, nuclease activity has already been shown to be a useful diagnostic biomarker for a limited number of pathogens that are bound to be further expanded.

6. Expert opinion

Due to nucleases´ vast diversity and their important roles in an ample range of bacterial pathogens, their catalytic activity could become a multipurpose diagnostic biomarker in clinical microbiology allowing to address some of the existing unmet needs for different diagnostic applications, including prolonged times to diagnosis, diminished accuracy or excessive costs and complexity. Additionally, its implementation into clinical practice has the potential to drive the development of entirely new modalities, such as non-invasive in vivo bacterial identification and visualization.

As mentioned before, any diagnostic modality or assay is highly constraint by the characteristics of the biomarkers they interrogate (Table 1). The biomarkers at the core of all diagnostic immunoassays and molecular genetic methods are antibody–antigen interactions and genetic determinants, respectively. As such, these two biomarkers define and set the limitations and disadvantages of the assays based on them, which include low sensitivity and proneness to interference in the case of immunoassays; or elevated false positives rates, inaccurate reporting of antimicrobial susceptibility and high complexity in the case of molecular genetic assays. For example, the disadvantages associated with the latter assays are due to the fact that: i) Genetic determinants can be detected even if pathogens are no longer viable leading to overreporting of positive results. ii) Genetic determinants of resistance are not a guarantee of phenotypic resistance, as resistance may depend on unknown or novel mechanisms or on gene expression levels [279,280], leading to misreporting. iii) Most molecular genetic methods depend on purification/isolation and amplification of the targeted determinants to provide adequate levels of sensitivity and specificity. These steps increase the risk for cross-contamination of specimens and usually involve many reagents and specialized equipment, such as thermocyclers, which need to be optimized to perform in different applications and matrices, increasing complexity and unavoidably impacting costs.

In this respect, the use of nuclease activity as a biomarker has several advantages over established diagnostic biomarkers in clinical microbiology, including genetic determinants and antibody–antigen interactions. Compared to genetic determinants or antibody–antigen interactions, nuclease activity represents a phenotypic trait, avoiding issues related to genotypic-phenotypic correlation. Moreover, its dynamic nature (enzymatic activity) opposes the static nature of nucleic acid hybridization or antibody–antigen interactions. Therefore, independently of the transduction technology, a single nuclease can interact with more than one reporter, acting as an intrinsic signal amplifier. This is not the case for the aforementioned biomarker counterparts, as a reporter can only interact with a single specific target, relying on additional amplification methods to enhance the signal. Additionally, both genetic determinants and antibody–antigen interactions are at the mercy of the genetic variability (e.g. mutations or recombination), which can lead to diagnostic resistance, requiring constant validation of the targeted biomarkers. Nuclease activity is also vulnerable to diagnostic resistance, however, given some of the fundamental roles of nucleases in the biology and pathogenesis of bacteria, it is likely that their function, and therefore their catalytic activities, are preserved despite the occurrence of genetic variations. This hypothesis is reinforced by the existence of numerous nuclease homologues whose most conserved sites are those relating to their catalytically active sites and cofactor-binding sites. Further support comes from the observation of converging functionalities and catalytic activities of non-homologous bacterial nucleases, as mentioned before (section 3).

Diagnostic assays using nuclease activity as a biomarker have already been developed, and in some cases clinically implemented, for the identification and characterization of pathogens, like S. aureus. With this in mind, and due to their straightforward accessibility, different extracellular nucleases have been proposed in this review as candidate diagnostic markers for the identification and characterization of numerous clinically relevant bacterial pathogens (Table 2). It is worth noticing that in some bacterial pathogens, such as S. aureus or S. marcescens, there exist a dominant extracellular nuclease activity of known origin and characteristics, which is readily detectable both in vitro and in vivo. However, in other cases. several nucleases with similar activity coexist, as is the case in S. pyogenes, C. perfringens or C. jejuni. Meanwhile, in other bacterial pathogens the origin of the nuclease activity detected is unknown. Furthermore, in bacteria whose dominant nuclease activity is produced by non-secreted nucleases, as it is the case for the periplasmic endonuclease I from E. coli, in vivo diagnostic applications are limited. Moreover, for in vitro diagnostic applications extra processing steps, such as bacterial lysis or cellular fractionation, may be necessary to access their activity, potentially affecting the accuracy and increasing the complexity and time to results. For these reasons, in order to utilize the full potential of nuclease activity as a diagnostic biomarker in multi-nuclease environments, a stepwise targeting approach limited to known and characterized nucleases is flawed. Consequently, a screening strategy that can target the unique nuclease activity blueprint of pathogens, independently of individual nucleases, and allows the selection and simultaneous optimization of sensitive and specific diagnostic reporters is necessary. These circumstances led Balian et al. [281] to develop a robust and easy to implement screening platform that takes advantage of the modularity of nucleic acids to iteratively screen oligonucleotide libraries for the selection of substrates that serve as highly specific/selective reporters of characteristic nuclease activity blueprints, even when their etiology is unknown. In this context, the composition of the libraries will therefore be crucial. Libraries containing substrates of different natures that are capable to question the catalytic properties, such as substrate and sequence preference, can be evaluated in a standardized way for all pathogens and iteratively re-designed and tested in screening rounds to enhance aspects, such as specificity and sensitivity levels. Obviously, previous knowledge about the catalytic properties of characteristic nucleases or the existence of homologues of known nucleases in targeted bacterial pathogens is bound to guide the design of libraries, but it is not indispensable. Of note, the use of nucleic acid libraries is simple, cost-effective and well-established compared to libraries of other natures (e.g. peptides and phages). Examples of these types of libraries are reported elsewhere [32].

Importantly, these screening platforms like the one described by Balian et al. can be implemented both in vitro, ex vivo or in vivo [281]. When implemented in vitro, once the sample matrix is defined (e.g. blood serum, saliva, sputum), questioning different catalytic parameters affecting the enzymatic activity derived from different relevant pathogens towards selected oligonucleotides using a screening approach is possible. The parameters that can be tested for include temperature, pH level and the presence or absence as well as the type, combination and concentration of cofactors, chelators, ionic compounds, oxidizing and reducing agents during the enzymatic process. As such, extra screening rounds can question an additional set of variables affecting nuclease activity to further identify the ideal conditions to maximize specificity and sensitivity for already selected reporters. It is plausible to imagine that analogous screening platforms could set the basis for the development of diagnostic biosensors based on activatable reporters, which by benefiting from the favorable features associated with the use of nuclease activity as a biomarker have the potential to enable accurate, single or multiplex identification and characterization of pathogens, even in polymicrobial infections, directly from complex clinical specimens.

Some of this potential has already been demonstrated for some applications ranging from rapid, easy to implement, specific and very sensitive assays for the identification of S. aureus bacteremia directly from positive blood cultures or directly from blood specimens; to quantitative phenotypic antimicrobial susceptibility tests for both Gram-positive and Gram-negative bacteria that are as accurate, but significantly faster than classical gold standard methods, as described in section 4. Importantly, the simplicity and rapidity offered by any of these methods already promise to be of clinical relevance by reducing the time to diagnosis (days to hours), which is an important parameter that impacts patient management, outcomes and associated health care costs. However, it is just a matter of time that novel and improved diagnostic approaches using nuclease activity as a biomarker are developed for different clinical and industrial applications given the abundance and diversity of nucleases in bacterial pathogens, the availability of suitable screening platforms and the advances in biosensing approaches. In fact, even now, it is not difficult to conceive that some of these already reported diagnostic approaches could be easily adapted for analogous purposes in different pathogens just by switching the specificity of the reporter so that it specifically targets their characteristic nuclease activity blueprints. For example, the screening and biosensing approach employed by Machado et al. to detect S. enterica ser. Typhimurium rapidly and accurately from pig-derived samples for food safety monitory could be adapted for the detection of other foodborne pathogens that also possess extracellular nuclease activities, such as C. jejuni or Y. enterocolitica. It is also conceivable that the approach used by Hernandez et al. for the specific identification and imaging of S. aureus infections in vivo could be adapted for other pathogens. In the case of H. pylori or M. tuberculosis this approach could be used to screen, diagnose, and monitor infections, while addressing issues related to current approaches, such as invasive diagnostic procedures, difficult specimen collection and low diagnostic sensitivity and specificity. However, to adopt such an approach for the aforementioned pathogens, novel or optimized delivery methods, as well as further advances in transduction technology, are due. For example, in the case of pulmonary infections, delivery methods could take advantage of the concepts already explored in pulmonary gene therapy [282–286], while in the case of gastric infections, biodegradable liquid gel-like matrices could be used as a cost-effective, clinically applicable vehicle to deliver reporters to the site of infection (e.g. stomach mucosa), while protecting them from unspecific activation during the process. Additionally, using activatable reporters employing magnetic resonance tuning systems [287,288], rather than the near-infrared fluorescence energy transfer system employed by Hernandez et al., would allow deep tissue detection and imaging using clinically available magnetic resonance imaging systems.

The use of nuclease activity as a biomarker is not without drawbacks. Perhaps the biggest limitation is the need for the existence of a characteristic nuclease activity associated to a specific bacterial pathogen. In some cases, different bacterial species may present or express the same type of nuclease which may hinder diagnostic specificity. For example, both commensal and pathogenic Neisseria spp, including Neisseria meningitidis and N. gonorrhea possess the complete nuc gene in their genomes [204]. However, if consistent differences in the pattern of expression of these nucleases exist between species, these would still translate into measurable and useful differences in nuclease activity. Strain typing may also be out of reach for assays based on nuclease activity, as strain-associated differences in nuclease populations seem to be absent in some bacterial species, such as Mycoplasma pulmonis [28].

It is also worth noting that nuclease production may be subjected to fluctuations associated to intraspecies variations [289] or growth conditions (e.g. oxygen tension or pH levels) [137]. Furthermore, in vivo there exists the possibility that nuclease production, and therefore activity, could vary during the course of infection depending on the functional role of the nuclease and the associated regulatory mechanisms. Ultimately, these variations could affect diagnostic sensitivity and specificity. However, given the participation of nucleases in virulence-associated activities, such as biofilm remodeling, immune evasion or host cytotoxicity, the quantification and evaluation of differences in nuclease activity could open a possibility to obtain information related to the stage of infection or level of virulence of the causative pathogen. This type of information could help guide prognosis and treatment.

However, it is likely that this type of evaluations would only be feasible in combination with techniques capable of very low limits of detection and a large dynamic range. Low limits of detection are also desirable in vitro for performing diagnostic assessments directly from complex clinical specimens, avoiding time-consuming enrichment and purification steps. To lower the detection limits in vitro, antibody-mediated nuclease enrichment has already proven successful [121], however, it requires previous knowledge of the nuclease activity being used as a characteristic marker. Another option involves the specific inhibition of ubiquitous eukaryotic nucleases [290], such as DNase I, to reduce possible sources of noise, increasing specificity and reducing the limit of detection.

Overall, we envision that the use of nuclease activity as a diagnostic biomarker aided by new technological advances could become a cornerstone in the development of evolutionary and revolutionary diagnostic assays and methodologies able to complement and add new features to the available arsenal of diagnostic tools in clinical microbiology, addressing unmet needs and opening new avenues for screening, diagnosis and prognosis. It is worth mentioning that nucleases are also present in non-bacterial infectious agents, such as pathogenic yeasts [291–293], protozoans and viruses; where they also play a role in their virulence [294]. In fact, some of these nucleases act as virulence factors by contributing to immune evasion in pathogenic plasmodial parasites [295] or coronaviruses [296,297], including the infamous SARS-Cov-2 [298,299]. Consequently, the use of nuclease activity as a diagnostic biomarker for non-bacterial infections also represents an attractive option that remains, however, out of the scope of this review.

Article highlights

• Timely and accurate diagnosis is fundamental for the successful management, therapy and outcome of infectious diseases.

• The nature of the biomarkers used defines the advantages and limitations of currently diagnostic techniques in clinical microbiology.

• Extracellular and intracellular nucleases are widespread in bacteria and play crucial biological roles, including pathogenesis.

• Readily accessible extracellular nucleases have been described in both Gram-positive and Gram-negative bacterial pathogens of humans.

• Due to their ubiquity, biochemical and functional diversity and catalytic nature, nucleases postulate themselves as promising biomarkers of disease.

• Diagnostic test using nuclease activity as a biomarker have been employed in the past to identify and discriminate between bacterial species.

• Novel technologies and approaches are expanding the scope nuclease activity-based diagnostic tests and show promise to become successful alternatives to gold standard techniques and to address unmet needs in clinical microbiology.

Declaration of Interest

FJ Hernandez is inventor in several patents that describe the use of nucleases as biomarkers. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

The authors would like to acknowledge Dr. Baris Ata Borsa and Anna Sophie Fröhlich for the careful revision of the manuscript.

Additional information

Funding

This work was supported by The Knut and Alice Wallenberg Foundation and The Swedish Research Council -Grant 2021-05641.