Grantee Research Project Results

 

 

The ability to mold structures in thermoplastics by a rapid and inexpensive hot embossing press is a cost-efficient alternative, and the flexibility of polymeric substances provides many options for sealing. We have fabricated polymeric chips in biocompatible cyclic olefin co-polymer. Later cyclic olefin co-polymer was replaced by polyester with similar properties (for our purposes) as cyclic olefin co-polymer. The cost of polyester is 25 to 30 fold less than cyclic olefin co-polymer, thus the cost per chip has decreased from approximately $2.50 per chip to 8 cents per chip. The current version of microfluidic chip has 64 reaction wells with 4 sample ports allowing simultaneous amplification of 20 genes in triplicate with 4 wells dedicated to negative controls (Figure 1). We have designed another simplified polymeric chip for the detection of 8 genes for the validation of isothermal assays and the fluorescence imaging system.

 

Figure 1.  A polyester microfulidic chip fabricated with 64-wells.

 

 

 

Waterborne bacteria and protozoa that have been commonly associated with disease outbreaks are Campylobacter jejuni, Escherichia coli O157:H7, Legionella pneumophila, Shigella flexneri, Salmonella enterica, Vibrio cholerae, Cryptosporidium parvum, and Giardia intestinalis. Some of these pathogens such as C. jejuni, E. coli O157: H7, and S. enterica have also been implicated in outbreaks of food-related illnesses. In the United States, 833 documented waterborne disease outbreaks have occurred between 1971 and 2006 resulting in 577,991 cases of illness and 106 deaths. In many cases, these deaths are due to the lack of appropriate diagnosis rather than availability of effective and economical prevention and treatment options.

 

To further simplify the assay and lower the cost of detection, we have chosen an isothermal DNA amplification reaction named loop-mediated isothermal reaction (LAMP). Features of LAMP include, moderate reaction temperature between 63-65 ^oC, excellent specificity due to the use of four to six specific primers, need for only single enzyme, and superior tolerance to substances that typically inhibit PCR. Another advantage of LAMP is the large amount of DNA generated during amplification. This permits direct visualization of amplification product based on turbidity or fluorescence, either at endpoint or in real-time. We have designed LAMP primers for the VMGs associated with major waterborne pathogens (Table 1). More information about designed LAMP primers can be found in our recent publication (Ahmad et al. 2011).

 

Table 1. Selected VMGs for major waterborne pathogens and their function

Pathoget	Gene target(s)	Description
Cryptosporidium parvum	GP60	60 kDa glycoprotein
	hsp70	70 kDa het shock protein
Glardia Intestinalis	ß-giardin	concerved protein
Leginella pneumophila	dotA	Integral cytoplasmic membrane protein
	lepB	effector protein
Vitrio cholarae	ctxA	cholera toxin
	tcpA	toxin-coregulated pilus protein
	toxR	two-component regulator
Sigella flexneri	ipaH	invasion plasmid antigen H
Campylobacter jejuni	0414	putative oxido reductase subunit
	cdtA	cytoletheal distending toxin A
Echerichia coli 0157:H7	eaeA	intimin
	stx1	Shiga toxin 1
	stx2	Shiga-toxin 2
Salmonella enterica	invA	invation protein
	phoB	phosphate regulon

 

 

Gene amplification and detection system was built by integrating light source, optical filters, thin film heater, and a CCD camera (Figure 2a).We have used SYTO dyes to monitor real-time LAMP by using a fluorescence imaging system. Due to the isothermal nature of LAMP, time-to-positivity (TTP) value was calculated by applying the signal-to-noise ratio cutoff greater than 10. With the application of SYTO dyes, we found that fluorescence LAMP (referred as RT_f-LAMP here) in microchips is faster than in a commercial real-time PCR machine (Chromo4^TM) (Figure 2b). Sensitivity, specificity, and genotyping potential of microfluidic chip-based assays were tested and compared with the performance of a commercial PCR machine.

 

Table 2. (a) Plotograph of the experimental setup consisting of an LED attached with exitation
filter (534 ± 20nm) for illumination, a thin film heater and thermocouple for temperature control,
and a monochromatic CCD camera with emission filter (572 ± 20 nm) for imaging. the inset shows
the microchip with seven V-shaped reaction wells used for RT_f-LAMP reactions. This system was
placed in the dark during imaging to avoid any ambient light. (b) Real-time flourescnence LAMP 
curves for 10⁵ DNA copies of C. parvum gp60 gene on the microchips (1 s, 3 s, and 5 s CCD exposure
time) and real-time PCR instrument (Chromo4^TM), (c) Standard curves for the C. jejuni 0414 gene 
amplification on the microchips at 5 s of CCD exposure time and real-time PCR instument. Error bars
represent the standard deviation of the mean from triplicates (Ahmat eta al. 2011)

 

To establish standard curves for quantitative analysis, real-time fluorescence LAMP assays targeting 0414 gene were performed on microchips with 5 s CCD exposure time and on a real-time PCR instrument with 10-fold serial dilutions of C. jejuni DNA ranging from 10⁵ to 1 copy. The standard curves were established by plotting the Tt values versus log of the number of genomic DNA copy used in RT_f-LAMP assays (Figure 2c). Both the microRT_f-LAMP and RT_f-LAMP assay on commercial real-time PCR instrument were sensitive to a single DNA copy. The correlation coefficients of the log linear regression plots between Tt values and DNA copy numbers for microRT_f-LAMP and real-time PCR instrument were the same, 0.99. However, microRT_f-LAMP assay was approximately twice faster than the RT_f-LAMP assay on the real-time PCR instrument. The result indicates that microRT_f-LAMP enables the reproducible and rapid quantification of DNA.

 

Figure 3. Real-time flourensence LAMP curves of 10⁵ DNA copies of 6 wterborne pathogent
(2 virulent genes for each) on the microchips with 5 s of CCD exposure time (filled circles
and squares) and real-time PCR instument (Chromeo4^TM) (open circles and squares). Error

bars represent the standard deviation of the mean from triplicates (Ahmad et al., 2011)

 

CCD-based imaging system was validated for 12 virulent genes of 6 waterborne pathogens using SYTO-82 dye and real-time fluorescence LAMP on microfluidic chip (Figure 3). Major waterborne pathogenic bacteria including, S. enterica (invA and phoB gene), C. jejuni (0414 and cdtA gene), L. pneumophila (dotA and lepB gene), E. coli O157:H7 (stx2 and eae gene), and V. cholerae (toxR and ctxA gene), and a protozoan, C. parvum (hsp70 and gp60 gene) were selected from the list of Centers for Disease Control and Prevention. MicroRT_f-LAMP assays for 12 virulent genes of waterborne pathogens (10⁵ gene copies) were performed at 5 s of CCD exposure time and compared with the real-time fluorescence LAMP on a real-time PCR instrument (Fig. 4). A reduction in Tt values ranging from 2.7 min to 9.8 min was achieved for microRT_f-LAMP assays in comparison to real-time PCR instrument.

 

 

Specificity of real-time LAMP assays was tested from blind samples containing targeted genomic DNA mixed with background DNA from environmental samples (data not shown). Detection sensitivity was in the range of 10-100 genomic DNA copies for these VMGs. For a total of 60 real-time LAMP reactions, only 4 false positive calls were observed. These false positives were only observed for less than 10 targeted genomic DNA copies. However, signal increased later in the reaction (over 20 min) indicating primer-dimer formation. In fact, these LAMP primers were insensitive to less than 10 genomic DNA copies as tested on real-time PCR machine (Chromo4^TM).

 

LAMP primers were designed to discriminate between alleles of a given gene (targeting single nucleotide polymorphisms) to genotype assemblages of Giardia. Primer sets (targeting the beta-giardin gene) to discriminate between assemblages of Giardia were designed with a mismatch on the 2^nd, 5^th, and 8^th primer on the the 5’ end of the FIP, and the BIP primer has a mismatch on the 2^nd and 5^th base. An experiment with Giardia intestinalis (Portland-1 strain, assemblage A) mixed with primers targeting assemblage A and primers targeting assemblage B, only showed amplification with the assemblage A primer set (Figure 2).

 

Figure 4. Testing Giardia intestinalis primers for assemblage discriminatino. 
The Portland strain (assemplage A) only amplifies with primers designed to 
target assemblage A. 

 

 

In 2008, we reported the validation of 20 waterborne by using high throughput on-chip PCR system (Stedtfeld et al., 2008). We also evaluated selected data for virulence and marker genes for their melting characteristics in order to provide higher level of confidence in the identities (Pozhitkov et al., 2008). We have also, demonstrated the correlation between the prevalence of antibiotic resistance genes and potential indicator markers (2008 EPA report). Additionally, we have evaluated the efficacy of multiple displacement amplification technique for amplifying small amounts of genomic DNA or whole genomes (2008 EPA report). Two manuscripts about the i) correlation of potential indicators with antibiotic resistance genes, and ii) the quantitative evaluation of bias during whole genome amplification is in preparation.

 

 

The inability to differentiate between viable and non-viable organisms with PCR based methods necessitates a complementary monitoring device. The unique characteristics exhibited with Dye-doped nanoparticles (DDNPs) such as small size, strong fluorescent signals, photostability and large surface areas provide a sensitive and quantitative means of detecting viable organisms. We demonstrated that dye-doped nanoparticle based method could detect growth of cells much earlier than many conventional techniques including cell and DNA-based real time PCR, absorbance and plate methods. The NP/antibody-based assay itself takes about 20 min and its sensitivity is comparable to antibody-based methods (~1 cell per ml of water). Using this assay, the growth of E. coli cells could be observed after 2 hr. A manuscript entitled, “Quantification and detection of microorganisms in viable-but-non-culturable state by nanoparticle-based bioassay” is in preparation for submission.

 

 

 

Previous studies on the membrane-based concentration of viruses involved crossflow systems operated at very low transmembrane pressures and, therefore, at very low permeate fluxes. At higher pressures, pressure instabilities result in a highly variable permeate fluxes and difficult to reproduce pathogen recoveries. We have re-configured the concentrator to avoid instabilities and to be suitable for higher pressure/flux operation. The appropriate components (e.g. high pressure peristaltic pump, etc.) have been identified and incorporated into the bench-scale concentrator. Results on filtration with calf serum blocked layer at various pressures and crossflow rates show that we were able to maintain average recoveries observed in literature while increasing permeate flux (Figure 5) We showed that by increasing both the crossflow rate and the transmembrane pressure we could greatly accelerate the filtration without significantly decreasing the average recovery of P22 from deionized water. Filtration of surface water (Lake Lansing, Haslett, MI) seeded with P22 bacteriophage was also performed in the same conditions and produced similar results. Using a transmembrane pressure of 60 psi we were able to reach a filtration rate of 170 L/min/m². This value is much higher than 1.4 L/min/m² permeate flux that we measured in experiments with 17 psi of the transmembrane pressure; lower pressures of the same order of magnitude (15 psi or less) have been used in previous studies (Morales-Morales et al., 2003; Hill et al., 2007).

 

Figure 5. Rate of concentration (top) and % recovery of virus surrogate (P22 bacteriophage) from 
deionized and surface water

 

 

In order to improve the reproducibility of pathogen recovery, we have developed a new approach to membrane blocking by using anti-adhesive membrane coatings based on multilayer polyelectrolyte films. In an effort to minimize irreversible deposition of pathogens on the membrane operated at desirable high permeation rates, we focused on the development of a concentrator that is operated at higher crossflow pressures and considered membrane coatings as sacrificial films. Specifically, we employed alternate deposition of polyanion and polycation layers on the ultrafiltration membrane support to produce polyelectrolyte multilayer (PEMs) films; these films dissolve at high pH to recover any virus that does adsorb to the film surface. This strategy is vital for recovering highly adsorbing viruses, particularly in water matrices with a high fouling potential. The approach enables fast (due to high water flux through ultrathin coatings), efficient (due to high virus rejection), and most importantly, reproducible (due to recovery of any viruses attached to the membrane coating) concentration of target viruses from high-volume water samples to the low volumes needed for quantitative, rapid detection by qPCR (Table 2).

 

Table 2. Relative benefits of rationally designed, removable polyelectrolyte multilayer (PEM) 

coatings with respect to the currently used methods of virus preconcentration. 
¹ Note that although ultrafiltration membranes are used in this method, after pre-blocking with
calf serum, the permeate flux decreases to levals typical for nanofiltration membranes. 

 

A bench-scale crossflow system - a prototype for the smaller scale, portable filtration unit - has been constructed and employed for testing the proposed method under a range of hydrodynamic conditions using feed water samples seeded with virus surrogates. Bacteriophage P22 was used as a model virus. One liter of P22 feed suspension (10⁷ PFU/ml) was concentrated to 250 ml sample under conditions of different transmembrane pressures and different crossflow rates. Calf-serum blocked membrane and membrane coated with polyelectrolyte multilayer (PEM) films were comparatively evaluated in these experiments.

 

After each concentration test, the membrane was subjected to elution procedure to recover viruses that adhered to the membrane surface. P22 concentration in feed, concentrate, permeate and eluate was measured by qPCR to evaluate the number of viruses lost to the permeate and to the membrane. We observed that for higher values of the crossflow flux to the initial permeate flux ratio (J_x/J_ip), P22 removal was higher for PEM coated membrane and was very low for calf serum coated membrane. Also, pre-elution recovery by membranes coated with PEM was higher than for calf serum blocked membranes, and could be improved by increasing J_x/J_ip for the former but not the latter. Our results show that coating a membrane with a PEM film, a 1-hour process, results in removal and recovery values that are at least as high as those obtained with calf serum coated membranes, and significantly better removal and pre-elution recovery for higher J_x/J_ip. In addition the high recovery observed for PEM coated membrane at high J_x/J_ip values shows potential for the development of a membrane surface that would eliminate the need for an elution step, thus reducing both the amount of time required for filtration and the volume of the final sample.

 

Figure 6. Pre-elution recovery of P22 from deionized water