key: cord-1010130-poptcb99
authors: Paul, Rajesh; Ostermann, Emily; Wei, Qingshan
title: Advances in point-of-care nucleic acid extraction technologies for rapid diagnosis of human and plant diseases
date: 2020-09-08
journal: Biosens Bioelectron
DOI: 10.1016/j.bios.2020.112592
sha: 4a6ee9ac07a7dfcb1f315d4bd3fd8c386cbfa186
doc_id: 1010130
cord_uid: poptcb99

Global health and food security constantly face the challenge of emerging human and plant diseases caused by bacteria, viruses, fungi, and other pathogens. Disease outbreaks such as SARS, MERS, Swine Flu, Ebola, and COVID-19 (on-going) have caused suffering, death, and economic losses worldwide. To prevent the spread of disease and protect human populations, rapid point-of-care (POC) molecular diagnosis of human and plant diseases play an increasingly crucial role. Nucleic acid-based molecular diagnosis reveals valuable information at the genomic level about the identity of the disease-causing pathogens and their pathogenesis, which help researchers, healthcare professionals, and patients to detect the presence of pathogens, track the spread of disease, and guide treatment more efficiently. A typical nucleic acid-based diagnostic test consists of three major steps: nucleic acid extraction, amplification, and amplicon detection. Among these steps, nucleic acid extraction is the first step of sample preparation, which remains one of the main challenges when converting laboratory molecular assays into POC tests. Sample preparation from human and plant specimens is a time-consuming and multi-step process, which requires well-equipped laboratories and skilled lab personnel. To perform rapid molecular diagnosis in resource-limited settings, simpler and instrument-free nucleic acid extraction techniques are required to improve the speed of field detection with minimal human intervention. This review summarizes the recent advances in POC nucleic acid extraction technologies. In particular, this review focuses on novel devices or methods that have demonstrated applicability and robustness for the isolation of high-quality nucleic acid from complex raw samples, such as human blood, saliva, sputum, nasal swabs, urine, and plant tissues. The integration of these rapid nucleic acid preparation methods with miniaturized assay and sensor technologies would pave the road for the “sample-in-result-out” diagnosis of human and plant diseases, especially in remote or resource-limited settings.

Emerging human and plant diseases are one of the major threats to global health and human civilization. Outbreaks such as the current COVID-19 pandemic upend daily life. According to Johns Hopkins University, above 25 million people are infected by the novel coronavirus, and more than 800,000 people have lost their lives worldwide when this article is written. The global economy and many other aspects of human activities have been severely impacted. Early, rapid, and accurate detection of diseases is crucial to maximizing crisis management efficiency, treatment outcomes, and economic stability. However, current practices of human and plant disease detection are mainly restricted to the centralized laboratories. Usually, patients or samples are taken to hospitals or diagnostic clinics for testing, and test results are returned within several days. Often, disease detection is delayed in developing countries or regions due to the shortage of skilled personnel and medical infrastructure. Moreover, even the healthcare systems of developed countries are facing an unprecedented challenge for laboratory-based disease detection during the ongoing COVID-19 outbreak. Therefore, the demand for portable, easy-touse, and point-of-care (POC) diagnostic tests is increasing rapidly.

POC testing of infectious human and plant diseases frees crucial time for planning, preparing, and responding to stop or limit the spread of disease in a community or an agricultural field. In POC diagnosis, patients' samples are immediately analyzed for disease screening at the sampling point. This type of testing requires a very small sample size for biomarker detection, which can be collected by patients themselves without assistance from medical personnel. After the addition of samples to the testing device, the results are displayed within a few minutes. In POC testing, various detection techniques such as nucleic acid testing (Batule et al., 2020; Leiske et al., 2015) , lateral flow assays (LFA) (Fang et al., 2014; R. H. Tang et al., 2017) , nanomaterial-J o u r n a l P r e -p r o o f based sensors (Li et al., 2020; Liu et al., 2014; Ngo et al., 2018; Padmavathy et al., 2012) , colorimetric immunosensors (Ren et al., 2017) , volatile organic compound sensors (Z. , bio-optical sensors Yoo and Lee, 2016) , and electrochemical sensors (Dutta et al., 2018; have been applied for the rapid detection of a broad range of human and plant diseases. Among these techniques, molecular assays based on nucleic acid amplification (NAA) are widely preferred. NAA-based assays examine the genomic information of pathogens or cells and thus can accurately identify microorganisms as well as their pathogenic strains, which cannot be easily achieved with other techniques. Moreover, NAA-based assays are sensitive, specific, and often can be multiplexed for the simultaneous identification of multiple pathogens (Basha et al., 2017; Stumpf et al., 2016) .

NAA-based human and plant disease detection involves three major steps: nucleic acid extraction/purification, amplification, and detection. For nucleic acid extraction, the first step is cell lysis, which releases nucleic acids and other intracellular molecules of interest. Several onchip cell lysis techniques such as chemical lysis Yoon et al., 2018) , mechanical lysis (J. Mahalanabis et al., 2009) , electrical lysis (Hügle et al., 2018; Kim et al., 2009; Nan et al., 2014) , ultrasonic lysis (Branch et al., 2017) , thermal lysis (Leiske et al., After nucleic acid extraction and purification, many amplification and detection strategies have been developed in the past decades. Polymerase chain reaction (PCR) and its variants such as reverse transcription-polymerase chain reaction (RT-PCR) are still considered the goldstandard method for nucleic acid detection (Petralia and Conoci, 2017) . PCR and RT-PCR assays are highly sensitive and specific. For POC applications, the PCR reagents can be lyophilized without sacrificing assay performance Czilwik et al., 2015) . Moreover, in multiplexed PCR (Cai et al., 2014; Czilwik et al., 2015) or RT-PCR assays (Chan et al., 2016a; Yaren et al., 2017; Yin et al., 2020) , simultaneous amplification of multiple pathogens' DNA or RNA is possible for high-throughput screening. Over the past few years, researchers have developed many modified versions of PCR for rapid NAA in resource-limited settings, such as continuous-flow PCR (Fu et al., 2018) , digital PCR (Hindson et al., 2011; Yin et al., 2019a) , droplet PCR (Cai et al., 2014; Chiou et al., 2013; Markey et al., 2010) , insulated isothermal PCR (Tsai et al., 2019) , and ultrafast photonic PCR (Son et al., 2016 (Son et al., , 2015 . Nevertheless, precise temperature control on a miniaturized thermal cycler is still a major challenge Park et al., 2011) . Due to this limitation, isothermal amplification methods are better suited for in-field disease detection. Representative methods include loop-mediated isothermal amplification (LAMP) Ma et al., 2019; Park et al., 2017) , nucleic acid sequence-based amplification (NASBA) (Tsaloglou et al., 2011) , strand displacement amplification (SDA) (Fang et al., 2014; Lafleur et al., 2016) , recombinase polymerase amplification (RPA) (Bender et al., 2018; Magro et al., 2017b; Rohrman and Richards-Kortum, 2012) , and helicase dependent amplification (HDA) (Linnes et al., 2014; Magro et al., 2017a; Rosenbohm et al., 2020) . Among these isothermal techniques, LAMP has been most widely researched for POC applications (Choi et al., 2016; Ma et al., 2019; Ye et al., 2018; J o u r n a l P r e -p r o o f al., 2014) . Like RT-PCR, RT-LAMP combines reverse transcription and LAMP assays in the same pot for specific RNA amplification (Estrela et al., 2019; Rodriguez et al., 2015) . In general, LAMP is more robust and inhibitor-tolerant than PCR (Damhorst et al., 2015; Kaneko et al., 2007) . In addition, its isothermal reaction condition (65 0 C) allows the use of a much simpler and lower-cost heating instrument to run the LAMP assay (Lu et al., 2016; Wang et al., 2020; . Moreover, LAMP assays can directly amplify nucleic acids from raw samples such as whole blood and swabs due to their robustness (Hoos et al., 2017) . LAMP or RT-LAMP reagents can also be lyophilized to store at room temperature up to several months (Hayashida et al., 2015; Seok et al., 2017) . Recently, several modified versions of the LAMP assay such as Tte UvrD Helicase-LAMP (New England Biolabs, USA) and UDG-LAMP (Hsieh et al., 2014) have been demonstrated to further improve the specificity and other drawbacks of the assay.

The final step for disease identification is the detection and quantification of amplicons.

In the laboratory, gel electrophoresis is usually performed to confirm the amplicons based on their molecular sizes. For POC visualization of amplicon products, lateral flow strips can be used instead (Fu et al., 2018; Huang et al., 2013; Lee et al., 2019; Rodriguez et al., 2015) . Lateral flow-based amplicon detection is sequence-specific, and a single strip can detect multiple amplicons simultaneously . Another technique commonly used for laboratorybased assay quantification is real-time detection (e.g., real-time PCR or quantitative PCR (qPCR)). For real-time detection, DNA probes such as molecular beacons, TaqMan probes, or DNA intercalating dyes are included in the amplification mixture (Borysiak et al., 2015; Loo et al., 2017; Wu et al., 2013) . However, conventional qPCR requires bulky, expensive, and sophisticated instruments. Alternatively, amplicon detection can also been achieved on cost-J o u r n a l P r e -p r o o f effective smartphone-based reader devices (Borysiak et al., 2015; Choi et al., 2016; Kaur et al., 2019; Ma et al., 2019; Wang et al., 2020; Zhu et al., 2020) . Smartphone-based nucleic acid detection platforms have been used for rapid screening of both human and plant diseases in resource-limited settings (Hernández-Neuta et al., 2019; Kong et al., 2017) . Finally, the LAMP assay can also be detected and quantified by turbidity or color change of the assay solution (Curtis et al., 2008; Estrela et al., 2019; . For colorimetric detection of the LAMP assay, pH-sensitive dyes (Kaarj et al., 2018) , metal ion indicators (Port et al., 2014; Seok et al., 2017) , or functionalized gold nanoparticles (Choi et al., 2016) have been reported as color indicators in the literature.

For POC disease diagnostics, an ideal system should integrate all steps from raw sample processing to amplicon detection, and run the steps automatically with minimal human intervention. After the first demonstration of a miniaturized total chemical analysis system (μ-TAS) by Manz et al. (1990) , researchers in the last 30 years have developed numerous molecular detection systems utilizing microfluidic (Kolluri et al., 2017; Koo et al., 2017; Liu et al., 2011; Q. Liu et al., 2018; Petralia et al., 2013; Wu et al., 2014; Yan et al., 2017; Yin et al., 2019b; Zhang et al., 2016) or paper-based devices (Choi et al., 2016; Deng et al., 2017; Liu et al., 2017; Loo et al., 2017; Rodriguez et al., 2015) . Furthermore, several commercial platforms such as GeneXpert Systems (Cepheid, USA), ARIES Systems (Luminex, USA), BioFire FilmArray Torch (BioFire Diagnostics, USA), and Integrated Cycler (Focus Diagnostics, USA) have been developed for rapid molecular diagnosis of human diseases.

However, most of these detection systems do not incorporate a sample preparation step to isolate biomarkers of interest from raw sample matrices such as blood, saliva, urine, and plant tissue (Berry et al., 2014; Chan et al., 2016a; Stumpf et al., 2016) . Because of the complex nature J o u r n a l P r e -p r o o f of raw samples, many systems still depend on off-chip sample preparation (Kaur et al., 2019; Liu et al., 2017; Magro et al., 2017b; Rodriguez et al., 2015; Wang et al., 2020) or sample pretreatment steps such as plasma (Kaarj et al., 2018; Yin et al., 2020) or serum separation (Estrela et al., 2019; Tsai et al., 2019; Zhang et al., 2019) before the actual assay reactions. As a result, so far only a few number of truly integrated "sample-in-answer-out" systems have been demonstrated for practical use in real-world settings. For POC disease detection, automatic and hand-free sample preparation is a prerequisite because sample purity and contaminants directly affect the detection performance (e.g., sensitivity, accuracy, etc.) (Van Heirstraeten et al., 2014) .

In this review, we have summarized emerging POC sample preparation techniques, which demonstrate great potential for easy integration with miniaturized nucleic acid amplification and detection platforms for on-site and rapid detection of human and plant diseases from raw samples (e.g., human blood, saliva, urine, and plant tissue). This review specifically focuses on rapid extraction methods for the isolation of high-quality nucleic acid targets due to their preferred analytical performance in disease detection. The extraction techniques are discussed and grouped based on the sample matrix types, including most commonly accessible samples such as human blood, oral/nasal samples, urine, and plant tissues. For each extraction technique, we discuss their principles, advantages, disadvantages, and applications for real patient samples.

Blood is one of the most widely used body fluids for the molecular diagnosis of human diseases.

Human whole blood consists of plasma (~55% of total blood volume), buffy coat (including white blood cells plus platelets, ~1% of total blood volume), and red blood cells (~45% of total blood volume) (Alberts et al., 2002) . Undiluted plasma contains a high concentration of J o u r n a l P r e -p r o o f interfering proteins whose total concentration is typically 60-80 mg/mL, equivalent to a solution of 6-8% (w/w) BSA (Walker et al., 1990) . More than 50% of the composition of serum proteins are albumins, followed by immunoproteins (e.g., IgG, IgA, IgM, IgD), transferrin, fibrinogen, clotting factors, etc. (Walker et al., 1990) .

Nucleic acid extraction from whole blood is an essential step for DNA/RNA-based diagnosis. However, isolation of nucleic acids from whole blood is a multistep process, which is usually performed in well-equipped laboratories by skilled technicians. Standard laboratory extraction procedure involves three major steps: lysis of cell nucleus membranes with surfactants (e.g., SDS, CTAB, or Triton X-100), denaturation of proteins by proteases (e.g., proteinase K), and purification of nucleic acids (Basha et al., 2017; Kim et al., 2009) . However, the actual extraction protocols vary significantly depending on the purposes of sample preparation. For example, for genomic DNA isolation, white blood cells need to be separated from the rest of the blood components (J. . In contrast, for the detection of pathogenic nucleic acids or cell-free DNAs, either serum or pathogen-infected blood cells are separated before extraction (Zhang et al., 2019) . Removing red blood cells also helps to reduce the inference of hemoglobin, which is one of the major sources of inhibitors for downstream NAA reactions (Magro et al., 2017a) .

Several miniatured sample preparation techniques for whole blood, plasma, or serum have been reported for POC pathogen detection (Batule et al., 2020; Ganguli et al., 2017; L. Li et al., 2019) , short tandem repeat analysis (Gan et al., 2014; Lounsbury et al., 2013) , single nucleotide polymorphism detection (Lu et al., 2016) , cancer diagnosis (Zhang et al., 2010) , forensic analysis (Duarte et al., 2010) , and hereditary genetic testing (Zhuang et al., 2015) .

However, rapid sample preparation platforms for forensic analysis are beyond the scope of this J o u r n a l P r e -p r o o f review. In this section, miniaturized nucleic acid extraction systems for disease detection from human blood samples will be discussed. A summary of rapid nucleic acid extraction methods for human blood is presented in Table 1. J o u r n a l P r e -p r o o f 

Blood serum is blood plasma without the clotting factors (in presence of anticoagulants) and is often preferred over whole blood as a better testing medium. For rapid pathogenic DNA isolation from serum, several microfluidic chips have been reported in the past. Lee et al. (2006) developed a Laser-Irradiated Magnetic Bead System (LIMBS) for pathogen DNA extraction from human serum by combining laser irradiation and carboxyl terminated magnetic beads.

During the laser irradiation, the photothermal effect of the magnetic beads lysed hepatitis B Later, the same group (Hwang et al., 2011) integrated on-chip PCR amplification with this rapid sample preparation technique to develop a complete "sample-in-answer-out" pathogen detection platform from whole blood. Cai et al. (2014) presented a dielectrophoretic technique in a microfluidic platform for pathogen separation from diluted whole blood, water, and other contaminated environmental samples. As shown in Figure 1e , after dielectrophoretic separation, the pathogens captured in grooves were mixed with the droplets of preloaded PCR master mix.

Then, the droplets were slipped away from the grooves to their original positions to run multiplex array PCR amplification for pathogen detection. The integrated device simultaneously detected Pseudomonas aeruginosa, Staphylococcus aureus, and Escherichia coli from blood within 3 hours. For each pathogen, the platform had a LOD of ~10 3 CFU/mL. Jin et al. (2018) utilized positively charged homobifunctional imidoesters (HI), including dimethyl pimelimidate, dimethyl adipimidate, and dimethyl suberimidate, to capture pathogens from various raw samples, such as blood plasma, swab, saliva, and urine. A premixed solution of HIs and sample was added onto the surface-modified microfluidic chip for selective binding of HI-pathogen complexes. One potential drawback is that the HI reagents may bound non-specifically to any negatively charged molecules present in the sample.

Conventional microfluidic devices depend on expensive syringe pumps for precise fluid manipulation. To eliminate the need for syringe pumps, several pump-free microfluidic devices such as hand-operated microfluidics (Byrnes et al., 2013; Li et al., 2012; , of as low as 100 copies/ml was reported. Despite their performance potential, centrifugal microfluidic devices have major drawbacks when scaling up. One of major drawbacks is high energy consumption, which limits its use in the field where the power supply is extremely limited. Furthermore, the fabrication of LabDisk devices is a complicated process, which presents a disadvantage for mass production and cost-effectiveness.

J o u r n a l P r e -p r o o f device with pre-stored extraction reagents (Figure 2d ). In a PTFF tube with a 1-mm inner diameter, lysis buffer, washing buffer, and elution buffer were stored as droplets separated by mineral oil. An external magnetic field drove magnetic beads through these droplets for sequential nucleic acid binding, purification, and releasing. This capillary tube-based system extracted genomic DNA from whole blood in 5 minutes and can be easily applied to isolate pathogenic DNA.

In addition to microfluidic devices, paper-based devices are widely used for POC diagnostics Magro et al., 2017a; , pathogen detection (J.

R. , food safety analysis Trinh et al., 2019) , and environmental monitoring (Sarwar et al., 2019; Seok et al., 2019; R. Tang et al., 2017b) . Paper-based devices are easy to fabricate and inexpensive. They do not require an external pump for fluid manipulation. In paper devices, liquids flow spontaneously due to capillary action. Thus, nucleic acid extraction, amplification, and detection can be easily integrated on a single paper device without manual sample transfer steps.

For rapid sample preparation from raw samples (e.g., whole blood and plant leaf), 

Like the FTA card, the Fusion 5 membrane has also been reported for nucleic acid extraction from whole blood. For isolating HIV-1 proviral DNA from whole blood, Jangam et al.

(2009) developed a Filtration Isolation of Nucleic Acids (FINA) method by using the Fusion 5 membrane. In the FINA method, whole blood was pipetted onto the Fusion 5 membrane to trap blood cells. NaOH was added to quickly lyse entrapped cells and wash away hemoglobin, cell debris, and other inhibitors. The FINA method has been modified to increase the limit of detection of HIV-1 from whole blood (Jangam et al., 2013; McFall et al., 2016 McFall et al., , 2015 . In the modified FINA method, whole blood lysed with Triton X was added onto the Fusion 5 membrane to capture nucleic acids. In a field trial for 61 patient samples, McFall et al. (2015) observed 100% detection sensitivity and specificity by using the modified FINA method. Tang et al. (2017a) compared the performance of the FTA card and Fusion 5 membrane for HBV detection from clinical blood samples and found that the detection limits were 10 3 copies/mL and 10 4 copies/mL for the FTA card and Fusion 5 membrane, respectively. To increase the nucleic acid extraction efficiency of the Fusion 5 membrane, Yin et al. (2020) proposed a method to modify the membrane surface with chitosan. Chitosan is a linear polysaccharide with a pKa ranging from 6.3 to 6.5 (Byrnes et al., 2015) . As a result, chitosan exhibits pH-dependent interactions with negatively charged nucleic acids ( Figure 3c ). Utilizing the chitosan-modified Fusion 5 membrane, the authors separated dengue virus (DENV) RNA from crude cell lysate.

After RNA extraction, the modified Fusion 5 membrane was directly used for PCR amplification without inhibition.

Besides FTA card and Fusion 5 membrane disks, nitrocellulose and glass fiber membranes have also been reported for DNA extraction (Byrnes et al., 2015; Seok et al., 2019) . (Figure 3d ). In this device, the glass pad successfully captured Staphylococcus aureus DNA from lysates of various raw samples, such as blood, saliva, urine, water, and milk. Moreover, the purified DNA could be stored in the device up to 2 months before elution. To investigate the effect of paper pore size for chitosan modification, Byrnes et al. (2015) studied the interaction of low molecular weight (~5000 MW) chitosan with a nitrocellulose membrane with 5-10 µm pore diameter (FF80HP, GE Healthcare) and a glass fiber membrane with 10-100 µm pore diameter (Standard 17, GE Healthcare). The DNA capture capacity was lower in the modified nitrocellulose membrane than in the modified glass fiber membrane. The authors hypothesized that the integration of chitosan into smaller nitrocellulose pores reduced the amount of available chitosan for DNA binding in the nitrocellulose membrane. In addition, chitosan obstructed the convective transport of DNA in the smaller pores of the nitrocellulose membrane. As such, the authors recommended using a glass fiber membrane or other papers having large pores for chitosan modification. Batule et al. (2020) immobilized single-stranded DNA probes on a glass fiber membrane (or binding pad) for viral RNA enrichment and extraction in a paper strip device (Figure 3e ). In this device, three different probes specific to Zika, dengue, and chikungunya viruses captured viral RNAs from serum cell lysate for the early detection of mosquito-borne diseases. After hybridization, viral RNAs were eluted and manually transferred to another paper-chip device containing dry RT-LAMP reagents for amplification.

J o u r n a l P r e -p r o o f

Oral and nasal specimens (e.g., saliva, sputum, and nasal swabs) have also been extensively used for human disease diagnosis. Oral and nasal samples are non-invasive and can be easily collected at resource-limited settings. In this section, miniaturized nucleic acid extraction systems for human disease detection from saliva, sputum, throat swab and nasal swab samples are discussed.

Among oral specimens, saliva is an extracellular fluid containing >99% water. As a result, nucleic acid extraction from saliva is relatively simple and less tedious compared to other body fluids. Several miniaturized systems using saliva as a diagnostic fluid have been developed Wand et al., 2018; Zhu et al., 2020) . These integrated systems utilized SPE (Chen et al., , 2013 , paper-based extraction (Jiang et al., 2018; Seok et al., 2019; R. Tang et al., 2017a) , or pH-responsive polymer-based extraction (Zhu et al., 2020) for rapid sample preparation from saliva (also see Table 2 ).

For viral and bacterial pathogen detection from human saliva, Chen et al. (2010) integrated silica membrane-based DNA extraction, PCR amplification, and lateral flow-based amplicon detection in a microfluidic cassette ( Figure 4a ). All DNA extraction and PCR reagents were prestored in the microfluid cassette. After sample loading, the cassette was inserted into a custom-built analyzer to begin the operation (Qiu et al., 2011 ). This portable system was able to In this device, the loaded saliva sample was split into two different compartments for ELISAbased antibody detection and silica membrane-based nucleic acid extraction, respectively.

Following nucleic acid purification, on-chip RT-PCR was performed using the dry illustra™ RT- passed through the surface-modified capillary, positively charged chitosan molecules bond nucleic acids. Later, the PCR master mix (pH=8.5) was introduced into the capillary to release the nucleic acids. As demonstrated by this device, charge-switchable ploymers could be a promising strategy to selectively capture and release nucleic acids from biofluids for rapid sample preparation at the detection site.

Sputum is another oral fluid that is often used for the detection of a variety of human diseases such as tuberculosis (TB), influenza, and pneumonia (Table 2) . Unlike saliva, sputum is a highly viscous fluid (Kaur et al., 2019) . As a result, nucleic acid extraction from sputum samples involves additional pretreatment steps, including sample disinfection, liquefaction, and homogenization . In the laboratory, sample pretreatment steps are performed developed a paper-based origami device (Figure 5e ). In this device, a Fusion 5 membrane was used to capture nucleic acids from sputum lysate. For on-site applications, lysis buffer was dried on paper. This device was demonstrated to isolate E. coli DNA from spiked pig mucin (simulated sputum) with a low bacterial concentration of 33 CFU/mL. The total sample preparation time was 1.5 hours in field settings. Obviously, one of the major drawbacks of paperbased methods for viscous sputum sample preparation is the relatively long sample preparation time due to the slow diffusion rate of molecules in the paper matrix. However, given the costeffectiveness and simplicity of paper devices, their use has promising potential for POC sample preparation.

Instead of directly collecting oral fluids, swabs are often used to collect oral samples for human disease detection (Ferguson et al., 2011; Wang et al., 2019) . To detect H1N1 influenza washing, and elution steps were performed manually by loading liquid reagents into the microchip using a micropipette (Figure 5f ). After DNA extraction, the eluted DNA was split to several microchambers with pre-stored primers for LAMP amplification.

Besides oral samples, nasal specimens (e.g., nasal swab, nasopharyngeal swab, nasal wash, and nasopharyngeal aspirate) have been widely used for the molecular detection of human diseases in both laboratory and field settings (Lafleur et al., 2016; Wang et al., 2018) . Table 3 summarizes the rapid nucleic acid extraction methods for biomarker detection in nasal Figure 6a ). In the microfluidic channel, silica beads of sizes varying from 5 µm to 30 µm were packed against the etched weir by applying a vacuum for nucleic acid extraction. The total sample preparation time for this microchip was less than 10 min. However, a syringe pump was required to deliver samples and reagents to the microchip, which limits its potential POC applications. Furthermore, the packed bed of silica beads could potentially be clotted by impure samples, which limits the total sample processing capacity of the microchip (e.g., ~1 µL of nasal aspirate). (Figure 6d ). After washing the captured precipitate, the LAMP reaction mixture was directly added onto the PES membrane for paper-based LAMP amplification. This paper device detected a viral load as low as 10 6 copies/mL, which is ten-fold lower than the LOD of conventional ELISA assays. Later, the authors further modified the design to minimize the number of manual interventions necessary for pathogen detection (Rodriguez et al., 2016) . In the modified paper device, a plastic cover film was used to seal the PES membrane for LAMP amplification, which minimized cross-contamination and false-positive rates.

Urine is another common diagnostic fluid for human disease detection, which consists of ~95% water, 2% urea, various ions (chloride, sodium, potassium, etc.), and many other small molecules, which are the most valuable part for diagnosis. The analysis of urine biomarkers (e.g., proteins, glucose, electrolytes, and pathogens, etc.) can reveal the underlying health problems of a person at the early stages before symptom develops. The collection procedure of urine samples is relatively easy compared to that of other body fluids. Furthermore, urine samples are not limited by volume, and a large volume of urine samples can be enriched for highly sensitive detection of early disease markers. Table 4 summarizes the rapid nucleic acid extraction methods for biomarker detection in urine. to concentrate Brucella ovis from a large volume (<50 ml) of urine in a syringe-based sample preparation system. In this system, a Teflon syringe filter was used to trap the ADE-HI-pathogen complexes for target enrichment, cell lysis, washing, and DNA elution. The detection limit of this combined pathogen enrichment and nucleic acid extraction system was 1 CFU/ml, which is a 100-fold improvement over the commercial kit-based methods. For Chlamydia trachomatis detection in urine samples, Chan et al. (2016a) presented a medium-throughput 3D printer-based molecular diagnostic system. The authors replaced the extruder of the printer with a tip-comb attachment to conduct magnetic particle-based DNA extraction. The tip-comb attachment consisted of either 8 or 12 tips depending on the number of samples required to be J o u r n a l P r e -p r o o f simultaneously processed, and small permanent magnets were placed inside these tips for magnetic particle manipulation. This modified 3D printer automatically extracted nucleic acids from 12 samples simultaneously within 15 minutes. Following automatic nucleic acid extraction, the authors also demonstrated a water bath-based two-step PCR amplification, where the 3D printer automatically transferred PCR tubes in between two water baths with different temperatures for thermal cycling. Later, the authors integrated an isothermal RPA assay with this 3D printer-based nucleic acid extraction system for Zika virus detection from urine samples (Chan et al., 2016b) . During the RPA assay, the heated printer bed was used to directly heat the RPA tubes, which simplified the overall complexity of the system by eliminating two different temperature water baths and sample transportation between these baths. For on-site nucleic acid purification from urine samples, Pearlman et al. (2020) presented a high-gradient magnetic separation (HGMS) technique to capture DNA binding magnetic particles in the steel wool matrix (Figure 7b ). The steel wool matrix was placed into a transfer pipette, and when the mixture of sputum cell lysate and magnetic beads was pulled up through the pipette, the magnetic beads were captured on the magnetized steel wool matrix due to the dominance magnetic force exerted on beads over viscous drag. The authors demonstrated that the DNA capture efficiency of the system was ~90% from DNA-spiked urine samples, which is similar to that of Qiagen nucleic acid extraction kits. Moreover, the HGMS-based system processed a larger volume of samples compared to commercial kits. 

Paper-based devices have also been extensively utilized to isolate pathogenic nucleic However, low detection sensitivity was observed for the direct amplification assays without nucleic acid extraction (Wand et al., 2018) . Kaarj et al. (2018) presented a wax-printed paper microfluidic chip to filter Zika virus RNA from various raw samples such as urine, water, and diluted plasma (Figure 7d ). When the urine sample was loaded into the microchip, only small viral RNAs were able to flow through the microchannel at the same speed as the bulk liquid, and finally reached the opposite circular end for detection, while large molecules were stuck in the sample loading area. After microfluidic filtration, the detection area of the microchip was cut, and the RT-LAMP master mix was added onto it for a paper-based RT-LAMP amplification. By J o u r n a l P r e -p r o o f combining microfluidic filtration and paper-based DNA amplification, the authors successfully detected Zika virus at a titer as low as 1 CFU/ml in less than 15 minutes. Jiang et al. (2018) utilized cellulose chromatography paper to capture viral RNA from cell lysate. After washing the captured RNA, cellulose paper was directly used as the template for RT-LAMP amplification to detect Zika virus. The authors found no inhibitory effect of the cellulose paper on the RT-LAMP assay. To increase nucleic acid capture efficiency from the urine samples, chitosan-based surface modification of Fusion 5 membrane and glass filter paper has been reported Rosenbohm et al., 2020) . For Trichomonas vaginalis detection, Rosenbohm et al. (2020) utilized a chitosan-modified Fusion 5 membrane to capture DNA from a large sample volume (~50 mL).

To extract DNA, cell lysate was loaded into a syringe and pumped through the chitosan-modified filter using a syringe pump. After this extraction, the chitosan functionalized filter was directly used as the template for an isothermal HDA assay to detect Trichomonas vaginalis. In the HDA assay, the authors found no inhibitory effect of chitosan. For the multiplex detection of bacterial pathogens in urine, Hui et al. (2018) presented a pipette-actuated capillary array comb (PAAC) system, where a chitosan-modified glass filter paper was used for DNA extraction (Figure 7e ).

The PAAC system consisted of six capillaries. Each capillary was embedded with a functionalized glass filter paper to capture DNA from the cell lysate. Moreover, target-specific LAMP primers were preloaded onto these chitosan-modified filter papers for multiplexed amplification. J o u r n a l P r e -p r o o f

Every year, plant diseases cause approximately 220 billion dollars of crop losses worldwide (Sarkozi, 2019) . According to the Food and Agriculture Organization (FAO) of the United Nations, more than 30% of global crop production is lost annually due to plant pathogens (e.g., fungi, bacteria, and viruses) and pests (Sarkozi, 2019) . For plant disease detection, molecular methods via nucleic acid amplification play a vital role in modern agriculture for protecting crops and increasing agricultural yield to fulfill the ever-growing demand for food supply (Lau and Botella, 2017; Martinelli et al., 2015) . However, most nucleic acid-based methods for plant pathogen identification and genotyping are laboratory-based. At present, field samples need to be transported to a laboratory for nucleic acid extraction, amplification, and detection (Lau and Botella, 2017; Ristaino et al., 2020) . Sample preparation from plant tissues is a complicated multistep process due to the rigid polysaccharide cell walls of plant cells. In the laboratory, the CTAB-based extraction method is widely used to extract nucleic acid from plant samples (Demeke and Jenkins, 2010) . A typical CTAB extraction protocol involves mechanical grinding, chemical cell lysis, temperature-assisted incubation, and organic phase-based nucleic acid extraction (Mahuku, 2004; Murray and Thompson, 1980) . Following DNA extraction, PCR or LAMP amplification is performed for pathogen detection. As a result, conventional methods are time-consuming, and they require expensive laboratory instruments and skilled technicians (Khiyami et al., 2014) . On the other hand, plant diseases spread rapidly, and some diseases such as late blight can destroy entire field crops in a few days if not treated (Gugino et al., 2013) .

Therefore, rapid and accurate on-site molecular detection platforms are essential for plant disease management (Donoso and Valenzuela, 2018; Nezhad, 2014) .

In recent years, several portable PCR systems (Julich et al., 2011; Koo et al., 2013; Radhakrishnan et al., 2019) and isothermal methods such as LAMP (Hodgetts et al., 2011; Ristaino et al., 2020) , RPA (Cha et al., 2020; Lau et al., 2016; Mekuria et al., 2014) and HDA (Lau and Botella, 2017; Schwenkbier et al., 2015) have been reported for facilitating field diagnosis of plant diseases. However, a lack of simple, instrument-free nucleic acid extraction methods remains one of the major challenges when converting laboratory-based molecular diagnostics to sample-to-answer field tests for plant diseases (Ali et al., 2017; Nezhad, 2014) .

In recent years, several rapid nucleic acid extraction methods have been reported for plant sample preparation (see Table 5 ). To simplify nucleic acid extraction from complex plant samples, Edwards et al. (1991) developed a sodium dodecyl sulfate (SDS)-based method to isolate PCR amplifiable genomic DNA in 15 minutes without using any hazardous chemicals such as chloroform or phenol. However, the protocol involved high-speed centrifugation for sample preparation. Thus, like the CTAB method, SDS-based DNA extraction is also confined to a laboratory setting. Nevertheless, an instrument-free, NaOH-based rapid DNA extraction method has been developed (Wang et al., 1993) . In this NaOH-based nucleic acid extraction method, a small piece of sample (e.g., leaf, plantlet, callus) was homogenized with 0.5 N NaOH solution. Then, 5 µL of the homogenized solution was quickly transferred to a new tube containing 495 µL Tris buffer (pH=8) for neutralization. While this NaOH-based method enables rapid sample preparation from plant samples, the neutralization step significantly dilutes the sample (e.g., 100x dilution), which reduces DNA concentration and thus detection sensitivity.

Moreover, this NaOH method works best for young tissues, not aged or dry plant samples. Chomczynski and Rymaszewski (2006) developed an alkaline polyethylene glycol (PEG)-based J o u r n a l P r e -p r o o f one-step nucleic acid extraction method. Unlike the NaOH method, this PEG-based method did not involve a neutralization step. Cha et al. (2020) utilized both NaOH and PEG and developed a DNA extraction buffer to extract nucleic acids from pinewood chips. Tomlinson et al. (2010) combined chemical cell lysis and a lateral flow strip-based nucleic acid purification for rapid sample preparation. For chemical-free sample preparation, Hieno et al. (2019) tested a waterbased extraction method to detect Phytophthora nicotianae from infected tomato, eggplant, and cucumber. In this water-based extraction method, fine slices of infected vegetables were vortexed with DI water. Then, the supernatant was used as a template for LAMP amplification.

However, the non-specific amplification rate of the water-based extraction solution was higher than the conventional extraction buffer (Hieno et al., 2019) . (Tangkanchanapas et al., 2018) . The sample extraction bag (Agdia) can also perform rapid nucleic acid extraction in the field ( Figure   8a ) (McCoy et al., 2020) . For in-field applications, the plant sample is placed inside of a plastic mesh bag containing the Agdia general extraction buffer 2. Then, the outside of the bag is rubbed with a blunt object such as a pen (Mekuria et al., 2014) . After sample maceration, the crude extract can be directly used for RPA amplification to detect Phytophthora species, which infect various food crops such as potato, tomato, and soybean, etc.

J o u r n a l P r e -p r o o f

Besides commercial sample extraction bags and nucleic acid extraction kits, paper devices made of FTA card and cellulose paper are extensively used for on-site rapid plant sample preparation, including cell lysis, nucleic acid extraction, and storage (Grund et al., 2010; Ndunguru et al., 2005) . For FTA card-based sample preparation, infected plant samples are pressed onto the FTA card ( Figure 8b) . Alternatively, the homogenized sample with Tris-EDTA buffer can be added to the FTA card (Grund et al., 2010) . 

Every commercial product or existing extraction method mentioned above requires 

In addition to human blood, oral/nasal, urine, and plant tissue samples, POC nucleic acid extraction devices or methods have also been developed for various other raw sample types.

Given the broadness of sample types, this section will only briefly summarize representative sample matrices such as stool samples and vaginal/cervical swabs to highlight the recent progress on POC nucleic acid extraction.

Stool samples are non-invasive and can be used to simultaneously test for multiple enteropathogenic bacteria and diseases. Although the composition of feces is variable, human feces has a median moisture content of 75%, with bacterial biomass comprising 25-54% of the organic solids (Rose et al., 2015) . The complex composition of feces poses a major challenge for direct testing. chemically lysed stool samples. After washing the captured DNA, the filter disc was cut and transferred to the LAMP mix for rapid rotavirus A detection. The total sample-to-answer detection time of the system was 30 minutes, and the LOD was 1 x 10 3 copies of rotavirus A/mL.

Vaginal and cervical swabs are typically used to diagnose sexually transmitted infections.

Vaginal swabs obtain samples of vaginal discharge, which is composed of cervical mucus, shed cells, and bacteria. Gulliksen et al. (2012) designed an automated lab-on-a-chip platform for human papillomavirus (HPV) detection from cervical specimens, consisting of two separate chips for sample preparation and nucleic acid amplification. The sample preparation chip was 

Over the past few years, significant progress has been made to simplify and speed up the nucleic acid isolation process from various raw human samples (e.g., blood, saliva, urine, sputum, stool, etc.) and difficult-to-lyse plant tissues. This review highlights a few representative systems such as microfluidic chips, paper-based devices, microneedle patches, and extraction bags to showcase the current frontiers of various POC nucleic acid extraction platforms. The J o u r n a l P r e -p r o o f microfluidic platform is a very cost-effective solution for sample preparation because of the reduced sample and reagents size. However, most of the sample preparation microchips presented in the literature do not have on-chip reagent storage capability and depend on external syringe pumps for fluid manipulation, which limits their applicability to resource-limited settings. To overcome this limitation, several pump-free microfluidic methods such as centrifugal microfluidics, digital microfluidics, hand-operated microfluidic, and capillary-based microfluidic devices have emerged. Beside these, paper-based devices have become popular for rapid sample preparation. Paper devices are inexpensive and easy-to-fabricate. Sample preparation reagents can be easily dried and stored on the device for long-term use or transportation. Moreover, paper-based devices facilitate nucleic acid extraction by providing additional purification functions. For example, paper membranes can filter blood cells from whole blood and thus eliminate the need for centrifugation. Furthermore, paper disks with entrapped nucleic acids can be directly added to the amplification reactions without the need for nucleic acid elution, due to the inert chemical properties of the cellulose matrix. To further increase the nucleic acid extraction efficiency of the paper device, chitosan-based charge switching extraction methods can be integrated to selectively bind target DNA/RNA from complex biofluids for rapid sample preparation at the detection site.

There are also several challenges remaining to be solved for the development of future POC diagnostic devices, including 1) further speeding nucleic acid extraction and purification within minutes, 2) better POC methods for the extraction of fragile RNA molecules, and 3) robust extraction strategies that can be potentially applied to several different sample matrices (e.g., blood, saliva, and nasal swab). The last point is especially relevant for accurate diagnosis of newly emerging infectious diseases such as COVID-19, when the optimal diagnostic media J o u r n a l P r e -p r o o f has not been identified at the beginning of outbreaks or one specific sample matrix is subject to significant variation of false positive or false negative rates.

Some of the sample preparation methodologies reviewed here can in principle be applied to isolate protein or other biomarkers from body fluids or plants, in addition to nucleic acids.

Many of these extraction platforms or methods have already been integrated with portable nucleic acid amplification and detection systems, such as smartphone devices, to generate fully integrated sample-to-answer detection platforms that have great potential for rapid screening of human diseases in resource-limited settings. In addition to NAA, many rapid nucleic acid extraction methodologies may also pave the road for POC DNA sequencing analysis in the future, which is expected to generate an even more profound impact on POC diagnostics and precision medicine.

Compared to human disease detection, in-field molecular diagnosis of plant pathogens is still lagging slightly behind. However, several promising recent progressions, such as rapid DNA isolation from infected plant leaves via a MN patch 

Lab a Chip -Miniaturisation

Proc. Natl. Acad. Sci

New standards to curb the global spread of plant pests and diseases

Clinical Methods: The History, Physical, and Laboratory Examinations

The authors gratefully acknowledge the funding support from the NCSU Chancellor's Faculty 

The authors declare no conflict of interest.

The authors declare no conflict of interest.