* Context-Collection of blood from newborns is a standard clinical procedure used for genetic screening. Typically, blood from a heel prick is absorbed onto standard collection paper and dried before analysis of metabolites, proteins, hormones, and more recently DNA.
Objective.-To evaluate strategies to purify DNA for use with automated workstations.
Design.-Two factors were used to evaluate several DNA purification protocols: residual heme contamination and amplification yield. The protocol that produced DNA with the lowest heme content and the highest amplification yield was selected. In combination with those two performance factors, the protocol with the fewest number of steps was chosen to reduce reagent use and processing time.
Setting.-Industrial research and development laboratory.
Results.-Robust amplification of DNA isolated from dried blood spots was demonstrated using both fluorescence and agarose gel-based detection methods. In addition, the samples had consistent DNA volumes and had no detectable cross-contamination. Suggested instrument settings, equipment, and supplies were included for automated processing of DNA from dried blood spots.
Conclusion.-A 4-step DNA processing protocol was developed for dried blood spots. The protocol could be performed in either a manual or automated format, making it possible to process hundreds of samples in 1 day.
(Arch Pathol Lab Med. 1999;123:1154-1160)
Collection of blood from newborns is a standard clinical procedure used for genetic screening. Typically, blood from a heel prick is absorbed onto standard collection paper for analysis of metabolites, proteins, hormones, and more recently DNA.1,2 Because blood spots are collected routinely at birth and because they are easy to collect and store, this medium is an excellent DNA source for population screening. Examples of two diseases that may be candidates for screening are hereditary hemochromatosis and insulin-dependent diabetes mellitus. DNA mutations within the HLA-H region are correlated with the incidence of hereditary hemochromatoSiS,3 and mutations within HLA DQA1 and DQB1 loci are correlated with susceptibility to insulin-dependent diabetes mellitus.4,5
Because neonatal blood specimens are dried, it is more difficult to isolate DNA. However, several methods have been developed to purify DNA in dried blood spots. Complex procedures that use methanol pretreatment, proteinase K digestion, and phenol-chloroform extraction are effective but require long processing times and toxic materials.2 However, these procedures have the advantage that the DNA is recovered in an aqueous solution suitable for multiple analyses. In contrast, the simpler protocols do not remove the DNA from the blood spot but pretreat blood spot disks with methanol2 or water6 and use the disk-- bound DNA directly for analysis. Because of the increasing number and complexity of DNA diagnostic tests being developed, there is a need for several analyses to be performed on the same DNA sample. In fact, several HLA typing protocols may require up to 50 separate polymerase chain reaction (PCR) amplification analyses to perform high-resolution DNA characterization. In these situations, DNA purification procedures in which DNA is transferred from the blood spot to an aqueous solution offer a major advantage in processing time and reliability compared with the direct method.
We describe herein the development of a rapid DNA purification protocol for use in high-throughput neonatal screening. One protocol requirement was that the DNA be eluted from the dried blood spot so that multiple analyses could be run on the same specimen. Another requirement was the use of nontoxic reagents. A third requirement was that the process be rapid and simple so that it could be performed using automated liquid handling equipment. Finally, it was required that the DNA of a purity level suitable for fluorescence-based amplification be used7 so that both the processing and analysis could be integrated into an automated, batched process.
MATERIALS AND METHODS
To create standard blood spots for optimization testing, blood from a single donor was collected in 10-mL blood collection tubes (Vacutainer, EDTA K, No. 16852; Becton Dickinson, Franklin Lakes, NJ). For this blood specimen, the white cell count was 10 X 10^sup 9^/L, and the hemoglobin concentration was 160 g/L. White cell count and hemoglobin content were determined using a counter (Coulter Counter CBC-5 Coulter Electronics, Inc, Hialeah, Fla) calibrated using hematology controls (CBC-7 Hematology Controls; R&D Systems, Minneapolis, Minn). The contents of 3 collection tubes were combined in a 50-mL polypropylene tube, and aliquots of 300 (mu)L were pipetted onto collection paper (S&S 903; Schleicher and Schuell, Keene, NH) and allowed to dry overnight. The dried spots were wrapped in plastic wrap and stored at -20 deg C. Anonymous neonatal blood spots, collected by heel prick using S&S 903 collection paper, were contributed by Neo Gen Screening (Pittsburgh, Pa). A 1/8-in hole punch was used to cut 3-mm-diameter disks from the blood spots for DNA purification. The punch was decontaminated between samples by punching clean collection paper 3 times.
To create reference white cell samples, blood collected in a citrate phosphate dextrose adenine (CPDA) bag (Baxter Healthcare Corp, Deerfield, Ill) by the Memorial Blood Centers of Minnesota, Minneapolis, was used. The blood had a white cell count of 4.4 X 10^sup 9^/L, and a hemoglobin concentration of 140 g/L. A 50-mL aliquot was separated into plasma and red cell and white cell fractions following 48 hours of gravity sedimentation at 4 deg C. The red cells and plasma were combined, and then the white cell fraction was added to make concentrations of 2, 7, and 20 X 10^sup 9^ white blood cells (WBC)/L. Using 300-(mu)L aliquots, dried blood spots were prepared using the 3 white cell concentrations as described herein.
An optimized DNA purification protocol was developed using two reagents: DNA purification solution (solution 1) and Generation DNA elution solution (solution 2) (Gentra Systems, Inc, Minneapolis, Minn). These reagents are nontoxic, aqueous solutions and were used on a laboratory bench without special ventilation. To purify DNA from blood spots, these reagents were poured into plastic reagent reservoirs (catalog No. 4870, Corning Inc, Acton, Mass, or catalog No. 37294, Beckman Coulter, Fullerton, Calif) and pipetted into 96-well plates containing 1/8-in disks. A multichannel pipette (12-Pette, catalog No. 4880, Coming) was used for manual dispensing, and a Biomek 2000 (Beckman Instruments, Inc, Fullerton, Calif) was used for automated dispensing. Three polypropylene 96-well plate configurations were provided by Marsh Biomedical Products, Inc (Rochester, NY) for DNA purification processing: PCR plate (catalog No. AB-0600), V-bottom plate (catalog No. N29074), and U-bottom plate (catalog No. N29069). After performing the washing procedure as described in the text, 150 (mu)L of solution 2 was added to the disks for DNA elution. The processing plate was sealed with an adhesive sealing sheet (catalog No. AB-0558; Marsh Biomedical) and placed into a block heater with a flat block insert prewarmed to 100 deg C (Analog Dry Bath, catalog No. 110002, and Microplate Block Module, catalog No. 110051; Boekel Scientific, Feasterville, Pa). For more even heating, an aluminum plate (catalog No. 90000-1AW; Pel Freez, Brown Deer, Wis) preheated in the block heater was placed on top of the sealed plate during the 15-minute incubation time. After heating, the adhesive sealing sheet was removed from the plate, and the eluted DNA was transferred to a clean 96-well U-bottom collection plate, leaving the disks in the processing plate to be discarded. To store the purified DNA, the plate was resealed with a clean sealing sheet and placed at -20 deg C. To remove condensation that formed after cooling and thawing, the plate was centrifuged at 1500g for 10 seconds (centrifuge model No. C412, M4 Swing Out Rotor catalog No. 11175338, and Microtitration Sealed Carrier catalog No. 11174223; jouan, Inc, Winchester, Va).
Detection of Heme Contamination
Heme, a major contaminant in DNA samples purified from blood sources, was used as a marker to evaluate the DNA purification protocols. Each DNA sample was diluted by transferring 20 to 50 (mu)L to a polystyrene microtiter plate containing a volume of solution 2 to make 200 (mu)L. Absorbence was read at 405 nm in a 96-well plate reader (EL311 Microplate Autoreader; Bio-Tek Instruments, Inc, Winooski, Vt).
PCR Amplification Using Fluorescence Detection
Purified DNA samples were evaluated in an amplification assay. All amplification reactions were set up in a horizontal flow hood. DNA samples were amplified using a quantitative Gene Amp PCR System 9700 (PE Applied Biosystems, Foster City, Calif). The amplification target, 390 base pairs in size, was located within the human HLA-H locus.3 DNA was amplified in a reaction volume of 25 (mu)L by adding 2.5 (mu)L of DNA to 22.5 (mu)L of a master mix. This mix consisted of reagents provided by the manufacturer (PE Applied Biosystems): 1X Universal Master Mix, 1X Internal Positive Control Mix, 1X Internal Positive Control DNA, and 200 nM custom Taqman fluorescent-labeled probe (5'-FAM-tgc ctc ctt tgg tga agg tga cac atc-TAMRA-3'). In addition, forward and reverse primers (5'-tgg caa ggg taa aca gat cc- 3', 5'-ctc agg cac tcc tct caa cc- 3 [Feder et al3], synthesized by Research Genetics, Huntsville, Ala) were included in each reaction at 1 (mu)M each. Six "no amplification control" wells (containing 2.5 (mu)L of TaqMan Exogenous Internal Positive Control Block) and six "no template control" wells (containing 2.5 (mu)L of solution 2) were run concurrently with each sample set according to the manufacturer's recommendation. Samples were amplified in the presence of a fluorescent probe so that the amplification yield could be quantitated using a DNA sequence detector (PE Applied Biosystems ABI Prism 7200 Sequence Detector). Both amplification and probe detection were performed in a polypropylene plate (PCR Plate AB-0600; Marsh Biomedical) sealed with adhesive sealing sheet (AB-0558; Marsh Biomedical) using a rubber roller (catalog No. 40101-1004, Dick Blick, Inc, Galesburg, Ill) to exclude air pockets. The amplification conditions were as follows: 50 deg C for 2 minutes, 95 deg C for 10 minutes; 40 cycles of 94 deg C for 30 seconds, 58 deg C for 30 seconds, 72 deg C for 30 seconds; and 72 deg C for 6 minutes, 4 deg C hold. Determination of amplification yield was computed as an Rn value using the Sequence Detector Version 1.6 software supplied by the manufacturer.
PCR Amplification Using Agarose Gel Detection
Purified DNA samples were also evaluated using standard PCR amplification followed by analysis using agarose gel electrophoresis. Samples were amplified in a 25-(mu)L volume using primers specific for the HLA-H locus as cited herein. The reaction mixture contained 1 X Taq polymerase buffer, 0.05 U/(mu)L Taq polymermase, 1.5 mM MgCl^sub 2^, and 0.2 mM each dNTP (Promega, Madison, Wis), and forward and reverse primers at 1 (mu)M (Research Genetics). The amplification conditions were as follows: 94 deg C for 5 minutes; 35 cycles of 94 deg C for 30 seconds, 58 deg C for 30 seconds, 72 deg C for 30 seconds; and 72 deg C for 7 minutes, 4 deg C hold. Following amplification, 10 (mu)L of each sample was loaded into a 2% gel and electrophoresed at 80 V for 45 minutes with buffer recirculation. Both gel and running buffer contained ethidium bromide at 0.125 (mu)g/mL to allow visualization on a UV transilluminator (312 nm transilluminator, catalog No. PBT1819; Fisher Scientific, Pittsburgh, Pa).
Estimation of Recovered Sample Volume
To determine the consistency of sample recovery by the Biomek 2000 workstation, the final DNA eluates were transferred directly from the processing plate to a 96-well quartz plate on the workstation deck. The sample volumes were calculated using the PathCheck algorithm supplied with the 96-well UV plate reader (SpectraMax Plus UV Plate Reader, Softmax Pro Version 2.2.1 Software; Molecular Devices, Inc, Sunnyvale, Calif).
Statistical analyses were performed using Microsoft Excel (Microsoft Corporation, Seattle, Wash) and StatView version 4.5 (Abacus Concepts, Berkeley, Calif). The coefficient of variation, expressed as a percentage, was defined as mean/SD X 100.
Optimization of a DNA Purification Protocol
To begin optimization of a simplified method for DNA isolation from dried blood spots, a series of wash protocols were evaluated. Nine protocols were designed to test different combinations of two reagents, DNA purification solution (solution 1) and DNA elution solution (solution 2). For all optimization testing, blood spots from a single donor were used to reduce variation. After placing a 3-- mm. disk into each well of a 96-well plate, the basic purification method was as follows. (1) Add 150 (mu)L of the designated solution, incubate for 15 minutes at room temperature, remove solution, and discard. (2) Repeat as specified. (3) Add 150 (mu)L of solution 2, seal plate with adhesive film, and incubate the plate at 100 deg C for 15 minutes to elute DNA. (4) Remove solution 2 and transfer eluted DNA to a clean 96-well plate. To compare the efficiency of each purification protocol, heme contamination and amplification yield were measured. The average and SD (n = 4) for each protocol are shown in Figure 1. The criteria used to select the optimum protocol were number of steps, heme concentration, and amplification yield. The protocol that gave the lowest heme concentration and highest amplification yield, yet required the fewest steps to perform, was protocol 5.
This protocol was performed as follows. (1) Add 150 (mu)L of solution 1, incubate for 15 minutes at room temperature, and discard solution 1. (2) Repeat step 1. (3) Add 150 (mu)L of solution 2, incubate for 15 minutes at room temperature, and discard solution 2. (4) Add 150 (mu)L of solution 2, seal plate with adhesive film, and incubate the plate at 100 deg C for 15 minutes to elute DNA. (5) Remove solution 2 and transfer eluted DNA to a clean 96-well plate.
To further optimize the method, the effects of plate configuration and mixing were examined. Using protocol 5, 3 plate types were evaluated: PCR, U-bottom, and V-bottom. All plates tested were made of polypropylene to withstand the heating required for DNA elution. In addition, we found that polypropylene was less prone to generating static electricity when handling the dry disks compared with other plastics. The effect of adding one mix step was tested to determine whether purification efficiency would be increased. A mix step was defined as aspirating and dispensing the full volume of solution just before removing it. The results, given in Figure 2, show the mean heme concentration for each treatment. The amplification yields (not shown) were not significantly different. Within each plate type pair, the coefficients of variation were lower in the samples with an incorporated mix step, protocols 11, 13, and 15. The protocol that gave the lowest heme contamination (0.044 +/- 0.002) and lowest variation (4%) was protocol 15, and this protocol was selected for automated processing.
The effect of wash volume was also tested because of the varying volume requirements for probes and pipettes used with automated liquid handling systems. For example, the aerosol-resistant pipette tips, recommended for use with the Biomek 2000, have a maximum capacity of 130 (mu)L. In this experiment, the range of wash volumes tested was 200 to 100 (mu)L, but the elution volume remained constant at 150 (mu)L for each sample. An elution volume of 150 (mu)L was chosen to allow for more than 50 separate amplification reactions to be analyzed from each DNA sample; this assumes a DNA template volume of 2.5 (mu)L. Results given in Figure 3 showed an increase in the variation in heme concentration and a significant reduction in amplification yield at the 100-(mu)L volume. The acceptable wash volume range was found to be 200 to 120 (mu)L, and the recommended volume was 150 (mu)L, which was used for the remainder of this study.
Evaluation of DNA Purification Protocol
To validate the purification protocol, 107 neonatal blood spots were tested for successful amplification (Figure 4). Disks of 3 mm diameter were placed into the wells of a U-bottom polypropylene plate and purified according to the optimized protocol (protocol 15). Reference samples I through 4 were used to determine the amplification yield range and were run concurrently with the neonatal DNA samples. The calculated quantity and concentration of DNA expected from each of the reference samples were 36 ng of DNA eluted in 150 (mu)L, or 0.24 ng / (mu)L, DNA for the 2 X 10^sup 9^ WBC/L, 126 ng of DNA eluted in 150 (mu)L or 0.84 ng/(mu)L DNA for the 7 x 10^sup 9^ WBC/L, and 360 ng of DNA eluted in 150 (mu)L or 2.4 ng / (mu)L DNA for the 20 x 10^sup 9^ WBC/L. These calculations assume 3 (mu)L of blood per disk, 6 pg of DNA per white blood cell, and 100% DNA yield; nucleated blood cells found in neonatal specimens were not included in the calculation. DNA concentration determination by standard UV absorbence was found to be unreliable probably due the presence of UV-absorbing materials copurifying with the DNA. Instead, successful PCR amplification was used to verify the presence of DNA.
The amplification results from the reference and neonatal samples are shown in Figure 4. Amplification yields given as mean Rn +/- SD (n = 6) were as follows: (1) "no template control," 5.49 +/- 0.19; (2) 2 million white blood cells per milliliter, 21.28 +/- 4.33; (3) 7 million white blood cells per milliliter, 28.49 +/- 1.75; and (4) 20 million white blood cells per milliliter, 33.66 +/- 2.91. The amplification results for the neonatal test samples ranged from a low Rn value of 14.10 (sample 105) to a high of 38.40 (sample 21). For this group of specimens, all of the amplification yields were easily distinguished from the "no template control" mean, with the lowest value (sample 10, Rn 14.1) exceeding it by about 2.5-fold. The robustness of the optimized purification protocol was demonstrated by the 107 of 107 positive amplifications.
Processing Disks in an Automated Liquid Handling System
To transfer the optimized purification protocol to an automated liquid handling system, several factors needed to be examined. The most difficult problem to overcome was prevention of disk loss by sticking to the pipette tips. With the manual method, it was easy to verify visually that the disks were not lost during the purification procedure. However, with automated processing it was important to program the system to reliably perform the wash steps without losing samples. To prevent disk loss, the pipette position and aspiration volume were found to be important factors. We found that the 10% aspirate height worked well for the standard Biomek disposable pipette tips. However, pipette tips from other manufacturers had different optimum heights. Another step that helped prevent disk removal by the pipette tips was to increase the aspiration volume from 150 to 175 (mu)L. Because the preceding step was a 150-(mu)L liquid dispense, a large air gap of approximately 25 (mu)L formed. This air gap broke the partial vacuum that formed occasionally between the disk and the pipette tip. In addition, the large air gap allowed liquids to be transferred safely over the surface of the deck, reducing any chance of cross-contamination. In addition, it was found that a slow aspiration rate helped to reduce disk loss during processing. Finally, the tip-touch function was used for each aspiration step, and the blow-- out function was used at each dispensing step. The settings recommended for use with the Biomek 2000 are summarized in Table 1. They were used to process successfully hundreds of disks without disk loss.
Additional testing was performed to validate the performance of the automated liquid handling system. Because diagnostic testing laboratories may require many separate analyses of a single DNA sample, it is important that the instrument be able to recover a consistent final DNA volume. To test this, 96 disks of 3 mm diameter were punched from a reference dried blood spot and placed into the wells of a 96-well polypropylene U-bottom plate. The samples were purified according to the procedure described in Table 1, and then the eluted DNA samples were transferred directly to a 96-well quartz plate to determine sample volume (see "Materials and Methods" for details). The results, shown in Figure 5, indicate consistent recovery across the plate. The average volume recovered was 145.6 +/- 3.7 (mu)L (+/-SD, n = 96), with a maximum volume of 155.2 and a minimum of 136.1 (mu)L. Even the sample with the lowest recovered volume (136 (mu)L) had enough DNA to perform more than 50 reactions using 2.5 (mu)L.
Another step necessary in using an automated processing system is to ensure that there is no cross-contamination among samples. This is a concern when samples are arrayed in close proximity, as they are in 96- or 384-well formats. Another consideration is the cost of using disposable tips when screening thousands of samples. To test for cross-contamination risk, an experiment was set up in which 48 disks were punched from a reference blood spot and placed into a 96-well plate alternating with 48 blank disks. The plate was processed in the Biomek 2000 automated workstation according to the protocol described in Table 1. To test the most conservative use of tips, a single rack of standard disposable tips (nonaerosol resistant) was used for the procedure so that the same tip was used for each sample throughout the protocol.
A sensitive PCR amplification assay with gel detection was used to evaluate cross-contamination risk. Using this sensitive DNA assay, as little as a single cell equivalent (6 pg) can be detected in a negative control sample. The assay, which uses primers specific to the HLA-H locus,3 was found to be 3 orders of magnitude more sensitive than the fluorescence-based assay.8 A group of 18 samples was selected for testing from the 96-well plate; 9 of the samples were processed from blood spots and the other 9 from blank disks. A volume of 2.5 (mu)L from each sample was tested for the presence of DNA by PCR amplification. After amplifying the samples, 10 (mu)L of the 25-(mu)L reaction volume was loaded into a 2% agarose gel as shown in Figure 6. Lanes containing amplified product from the blood spot disks gave the expected 390-base pair band within the HLA-H locus. In addition, the large quantity of amplification product indicated that the purified DNA template did not contain appreciable PCR inhibitors. Close examination of the gel image in Figure 6 shows that samples from the blank disks showed no amplified DNA, indicating the absence of detectable contamination from neighboring wells.
In summary, we developed an optimized DNA purification protocol for isolating DNA from dried blood spots in a 96-well plate format. Two factors were used to evaluate several purification protocols: residual heme contamination and amplification yield. The protocol that produced DNA with the lowest heme content and the highest amplification yield was selected. In combination with those two performance factors, the protocol with the fewest number of steps was chosen to reduce reagent use and processing time. A 4-step processing protocol was selected that could be performed either in a manual or automated format, making it possible to process hundreds of samples in 1 day. Robust amplification of DNA isolated from dried blood spots was demonstrated using both fluorescence-- and agarose gel-based detection systems. In addition, the automated protocol was found to have no detectable sample cross-contamination. Suggested instrument settings, equipment and supplies were included for automated processing of DNA from dried blood spots, summarized in Tables 1 and 2.
This work was funded in part by National Institutes of Health, Bethesda, Md, Small Business Innovation Research contract N43-- DK-6-222. We thank N. Morken, BS, P. Olson, and D. Tkach for excellent tedinical support.
1. McCabe ERB, Huang SZ, Seltzer WK, Law ML. DNA microextraction from dried blood spots on filter paper blotters: potential applications to newborn screening. Hum Genet. 1987;75:213-216.
2. McCabe ERB. Utility of PCR for DNA analysis from dried blood spots on filter paper blotters. PCR Meth Appl. 1991;1:99-106.
3. Feder JN, Gnirke A, Thomas W, et al. A novel MHC Class ]-like gene is mutated in patients with hereditary haemochromatosis. NatGenet. 1996;13:399408.
4. Morel P, Dorman J, Todd J, McDevitt H, Trucco M. Aspartic acid at position 57 of the HLA-DQ beta chain protects against type I diabetes: a family study. Proc Nail Acad Sci U S A@ 1988;85:8111-8115.
5. She JX. Susceptibility to type I diabetes: HLA-DQ and -DR revisited. Hum Immunol. 1996;17:323-332.
6. Makowski GS, Davis EL, Aslanzadeh J, Hopfer SM. Enhanced direct amplification of Guthrie card DNA following selective elution of PCR inhibitors. Nucleic Acids Res. 1995;23:3788-3789.
7. Faas SI, Mennon R, Braun E, Rudert WA, Trucco M. Sequence-specific priming and exonuclease released fluorescence detection of HLA DQBI alleles. Tissue Antigens. 1996;48:97-112.
8. Heath EM, 0' Brien DIP, Holmes DR, Morken N, Detert ME Evaluation of rapid purification methods for cross-contamination risk using HIV-1 provirus and HLA-H markers [poster GT 11 ]. Assoc Molec Pathol. 1998;4:34.
Accepted for publication July 19, 1999.
From Gentra Systems, Inc, Minneapolis, Minn Qrs Heath and O'Brien), and Neo Gen Screening, Inc, Pittsburgh, Pa (Mr Banas and Drs Naylor and Dobrowolski).
Presented at the Eighth Annual William Beaumont Hospital DNA Technology Symposium, DNA Technology in the Clinical Laboratory, Royal Oak, Mich, March 25-27, 1999.
Reprints: Ellen M. Heath, PhD, Director of Research, Gentra Systems, Inc, 13355 loth Ave N, Suite 120, Minneapolis, MN 55441
Copyright College of American Pathologists Dec 1999
Provided by ProQuest Information and Learning Company. All rights Reserved