Article pubs.acs.org/ac
Cell-Based Galactosemia Diagnosis System Based on a Galactose Assay Using a Bioluminescent Escherichia coli Array Min-Ah Woo,†,‡ Moon Il Kim,† Daeyeon Cho,§ and Hyun Gyu Park*,† †
Department of Chemical and Biomolecular Engineering (BK21 Program), KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea ‡ Food Safety Research Group, Korea Food Research Institute, Baekhyun-dong, Bundang-gu, Seongnam-si, Gyeonggi-do 463-746, Republic of Korea § LabGenomics Co., Ltd., 1571-17 Seocho3-dong, Seocho-gu, Seoul 137-874, Republic of Korea S Supporting Information *
ABSTRACT: A new cell-based galactose assay system, which is comprised of two bioluminescent Escherichia coli strains immobilized within an agarose gel arrayed on a well plate, has been developed. For this purpose, a galT knockout strain [galT(−) cell] of E. coli was genetically constructed so that cell growth is not promoted by galactose but rather by glucose present in a sample. Another E. coli W strain (normal cell), which grows normally in the presence of either glucose or galactose, was employed. A luminescent reporter gene, which produces luminescence as cells grow, was inserted into both of the E. coli strains, so that cell growth could be monitored in a facile manner. The two strains were separately grown for 4 h on gel arrays to which test samples were individually supplied. The relative luminescence unit (RLU) values caused by cell growth were determined for each array, one of which is resulted by glucose only and the other of which is resulted by both glucose and galactose present in the sample. By employing this protocol, galactose concentrations present in the test sample are reflected in the differences between the RLU values for each array. The practical utility of the new assay system was demonstrated by its use in determining galactose levels in clinical blood spot specimens coming from newborn babies. Because it can be employed to diagnosis of galactosemia in newborn babies in a more rapid, convenient, and cost-effective manner, this cell-based solid-phase galactose assay system should become a powerful alternative to conventional methods, which require labor-intensive and timeconsuming procedures and/or complicated and expensive equipment.
G
As a result of the fact that classic galactosemia caused by a deficiency of GALT leads to an accumulation of galactose and galactose-1-phosphate in blood, diagnosis of the disease is generally achieved by either measuring the levels of galactose or galactose-1-phosphate or evaluating the activity of GALT in the blood of newborn babies.3 A number of methods for the quantitative determination of galactose levels and GALT activity in blood have been developed, the most common of which include the use of Beutler’s and microbiological tests, HPLC (high performance liquid chromatography), and enzymatic assays. The most representative of these is the Beutler’s test (fluorescent spot test), in which GALT activity is determined employing a coupled enzymatic assay involving phosphoglucomutase and glucose-6-phosphate dehydrogenase (G6PD) in association with measurement of the intensity of fluorescence arising from formed NADPH.6 Although Beutler’s testis effective for early-stage diagnosis of classic galactosemia, it possesses several disadvantages,
alactosemia, a major metabolic disorder of newborn babies that is caused by an inability to metabolize the sugar galactose properly, follows an autosomal recessive mode of inheritance conferring a deficiency in genetic enzymes, which are responsible for galactose metabolism.1 The genetic disturbance of the disease is expressed in the form of a cellular deficiency of eithergalactose-1-phosphate uridyltransferase (GALT), galactokinase (GALK), or UDP galactose 4-epimerase (GALE). These deficiencies cause three respective types of galactosemia, referred to as classic galactosemia (galactosemia type I), galactokinase deficiency (galactosemia type II), or galactose epimerase deficiency (galactosemia type III).2,3 Of these, classic galactosemia is the most common and severest form of the disease, which causes rapidly progressive symptoms in the neonate, including jaundice, cataracts, ovarian failure, hepatomegaly (an enlarged liver), renal failure, brain damage, and even neonatal death if not properly treated during early stages.4,5 Because early diagnosis of galactosemia can help prevent progression of mental and developmental disorders in infant patients, it is now mandatory in major countries including the U.S. and Korea to perform diagnostic screening of newborns. © 2013 American Chemical Society
Received: September 3, 2013 Accepted: October 21, 2013 Published: October 21, 2013 11083
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
Analytical Chemistry
Article
recombinase, was first inserted into E. coli by using electroporation with a Gene pulser system (Bio-Rad, CA). The cells were then grown in LB media containing ampicillin (Am) (50 μg/mL) at 30 °C to an OD600 of 0.4, followed by additional incubation at 37 °C for 1 h after the addition of 1 M Larabinose (Sigma-Aldrich, MO). The prepared linear cassette was then inserted by using electroporation into E. coli, carrying the ability for Red-mediated recombination. The galT genedeleted E. coli [galT(−) cell] was specifically selected in the Cm-containing (35 μg/mL) agar plate. Finally by inserting the pTAC−lucplasmid19,20 into the competent forms, bioluminescent galT(−) and normal cells were generated. Galactose Quantification Using Arrays Containing Two Bioluminescent E. coli Strains. galT(−) and normal cells were separately cultivated in LB medium containing the antibiotics, Cm and Am and Am alone, respectively, at 37 °C for 9−12 h, under shaking conditions. After each of the cells were washed with M9 media two times, the cell solutions were mixed with 3% low melting agarose (Sigma-Aldrich, MO) in a 1:1 volume ratio to make the final cell−agarose (1.5%) solutions containing 2 × 106 cells per 100 μL. The cell−agarose mixtures (100 μL) were then independently poured into the wells of 96-well plates (Nunc, Roskilde, Denmark), followed by incubation at room temperature for 20 min to solidify the agarose gel. The assay solutions (120 μL), prepared by adding 30 μL of a test or standard sample to 90 μL of assay buffer, which is composed of M9 media including 1 nM cyanocobalamin (Sigma-Aldrich, MO), 1 mM IPTG (SigmaAldrich, MO), and 200 μM NaHCO3, were applied to the wells containing the galT(−) and normal cells. After incubation of the cell arrays at 37 °C for 4 h, luminescence intensities were measured using a luminometer (Perkin-Elmer, MA). Scanned images were obtained by using a cooled charge coupled device camera (Fujifilm, Japan) with a constant focal plane, magnification, and integration time. In order to determine the amount of galactose present in samples, standard curves were constructed from data arising from experiments using samples containing glucose solutions at concentrations of 0, 10, 20, 40, 80, 120, 160, and 200 μM in the above-mentioned assay buffer and galactose solutions at concentrations of 0, 5, 10, 20, 40, and 100 μM in the abovementioned assay buffer. The glucose and galactose solutions were applied to well arrays containing respective galT(−) and normal cell. In experiments to confirm that growth of galT(−) cells is not promoted by galactose, galactose solutions at concentrations of 0, 5, 10, 20, 40, and 100 μM were prepared in the assay buffer, which was added to the galT(−)-containing well array. The array was then incubated at 37 °C for 4 h, and luminescence was analyzed. The amounts of galactose in commercially available blood paper specimens (Bio-Rad, CA) and real infant blood paper specimens provided by Labgenomics Clinical Laboratories (Seoul, Korea) were determined. The blood paper specimens (3 mm diameter) were first punched, and the blood components of the punched papers were eluted by incubating with 3% trichloroacetic acid (TCA) (Sigma-Aldrich, MO) for 1 h at room temperature with moderate shaking. After this, 30 μL of each eluted solution was mixed with 90 μL of assay buffer to make the assay solution. The pH of each assay solution was adjusted into pH 7.3 by using 200 μM NaHCO3 (SigmaAldrich, MO). In experiments employed to construct standard curves and to quantify galactose from clinical paper specimens, 30 μL of glucose solutions (0, 10, 20, 40, 60, and 100 μM) or
including the dependence on a visual observation, fluorescence quenching by hemoglobin, and false-positive results arising from deteriorated enzymatic activities.7 Among other galactosemia screening methods that have been devised are the Paigen assay (E. coli−bacteriophage assay) and the biological inhibition test. The Paigen assay utilizes a mutant E. coli strain, which becomes resistant to bacteriophage C-21 by forming normal cell walls in the presence of galactose but is destroyed by the phage in the absence of this aldohexose.8 Thus, galactose concentrations in samples are determined by measuring the mutant E. coli growth zones around blood spots placed on a soft medium gel formed by sodium silicate. This assay procedure also possesses several limitations that are caused by the relatively long times needed to incubate E. coli (16−20 h), use of bacteriophagy culture, and requirement for a special medium and purified sodium silicate.9 The biological inhibition test for diagnosing galactosemia employs a transferase-deficient E. coli mutant whose growth is inhibited by intermediates produced during the metabolism of galactose. Unfortunately the E. coli mutant utilized in this method is generally unstable, and the zones of cell growth are difficult to measure accurately.10 These problems require that follow-up tests, which use HPLC and enzymatic assays, be performed to rule out frequent false positive results.11 Because of the experimental complexities and costs associated with HPLC analysis and enzymatic assays,12−15 a current need exists for more simple, rapid, and economical methods for measuring galactose concentrations in newborn blood samples.16 Taking into account the features that are required for a new, practical method for galactosemia diagnosis, we have designed a rapid galactose assay system that is comprised of two bioluminescent E. coli strains immobilized within agarose gels arrayed on a well plate. For this purpose, the galT knockout strain of E. coli was genetically constructed so that its growth is not promoted by galactose but rather by glucose present in a test sample. Owing to the presence of a bioluminescent reporter gene that is inserted into galT(−) cells, the magnitudes of relative luminescence units (RLU) of the grown galT(−) cells reflect glucose concentrations in the test samples. Galactose concentrations are then determined by using the difference between the RLU values associated with growthpromoted bioluminescence from galT(−) and normal E. coli cells, the latter of which have growth promoted by both galactose and glucose present in test samples. In the investigation described below, the analytical utility of this new cell-based galactose assay strategy was successfully demonstrated in a solid-phase array format by applying it to diagnose galactosemia using real infant blood paper specimens.
■
EXPERIMENTAL SECTION Construction of Two Bioluminescent E. coli Strains. galT Knockout E. coli was created by using the general chromosomal gene deletion method.17 A linear cassette, serving as a replacement for the chromosomal region of galT, consisted of a selectable chloramphenicol (Cm) resistance gene and two flanking homologous sequences. The assembly and amplification of the linear cassette was achieved by performing joint PCR of the resistance gene and two homologous fragments using the primer pair consisting of 5′-TCATAATCGGCTGCCATC AC-3′ and 5′-GCGTACACATCTACAACCTC-3′ (Bioneer, Daejeon, Korea). E. coli W (ATCC1105)18 was obtained from American Type Culture Collection (ATCC, MD). A Red helper plasmid (pKD46), encoding Red 11084
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
Analytical Chemistry
Article
Figure 1. Galactose quantification on an array containing two bioluminescent E. coli. The array consists of two regions, one to quantify glucose and the other to quantify galactose. Cells A and B in 1.5% agarose were immobilized in the regions for glucose and galactose quantification, respectively. Standard (glucose or galactose) and test samples were applied to the designated wells, and the luminescent signals were measured after 4 h incubation at 37 °C. (Five test samples were applied to both cell A and B arrays.) In step (1), glucose concentration in a test sample was determined by using a standard curve which correlates cell A growth with glucose concentrations, and in step (2), the RLU value corresponding to cell B growth induced by glucose was determined by using the standard curve. The RLU value was then subtracted from the RLU value corresponding to cell B growth induced by both glucose and galactose. Finally, galactose was quantified by using the standard curve in step (3).
galactose solutions (0, 2.5, 5, 10, and 20 μM) was spotted on blank filter papers (3 mm diameter) and the papers were dried at room temperature. Glucose or galactose was then removed by elution of the papers in 3% TCA, and 30 μL of the eluted solution was used to make 120 μL of the assay solution, which was then subjected to the assay procedure described above. Detection Precisions for Galactose Quantification and Galactosemia Diagnosis. The detection precision of this method to quantify galactose was assessed by using samples
containing five different amounts of galactose (2, 8, 16, 25, and 40 μM) and clinical paper specimens including commercially available blood paper specimens (Bio-Rad, CA) (4.22, 7.50, and 10.14 μM) and real infant blood paper specimens (2.73, 3.98, 5.24, 6.60, and 9.45 μM). Galactosemia diagnosis was performed by using four different infant blood paper specimens, which contain galactose at concentrations of 2.73, 3.98, 5.24, and 6.60 μM and were supplied individually with additional amounts of galactose (5, 8, 11, and 14 μM) to create two 11085
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
Analytical Chemistry
Article
Figure 2. Correlation between cell growth of the two bioluminescent E. coli cells and concentrations of glucose or galactose. The luminescent signal produced by galT(−) cell growth was plotted with (A) glucose or (B) galactose. The luminescent signal produced by normal cell growth was also plotted with (C) glucose or (D) galactose. Below each graph are shown luminescent images produced by glucose or galactose at different concentrations.
1-phosphate.21 Therefore, growth of galT(−) cells does not take place in the presence of galactose alone. However, growth of these cells is still promoted by glucose in a concentration dependent manner. In contrast, normal E. coli cells grow in the presence of both glucose and galactose in a concentrationdependent manner. In order to visualize and quantitate cell growth, both the galT(−) and normal cells were designed to contain a luminescence-generating plasmid pTAC-luc, which is constructed by inserting the luciferase gene (luc) derived from Photinuspyralis (firefly) into pET−pTAC containing an IPTGinducible promoter.19,20 As a result, the luminescent galT(−) and normal E. coli strains each produce a luminescence signal, whose intensity reflects the extent of growth promoted by galactose and glucose in the case of normal cells and by glucose in the case of galT(−) cells. Thus, growth data arising from analysis of the luminescence of the cells can readily be manipulated to give the amounts of galactose in blood samples. The array system (Figure 1) used for the assay consists of separate regions for glucose and galactose quantification that contain wells which are loaded with galT(−) (cell A) and normal (cell B) cell in 1.5% (w/v) agarose, respectively. For the construction of standard curves, glucose solutions of different
normal and two galactosemia samples. The actual galactose concentrations of the blood paper specimens were determined by using an enzymatic galactose assay kit (Bio-Rad, CA) followed by analysis using a chemistry analyzer (Hitachi 7180, Hitachi, Japan), according to the manufacturer’s instructions and protocols. The precision and reproducibility of the assay for galactose quantification and galactosemia diagnosis were determined using the recovery rate [recovery (%) = measured value/expected value × 100] and coefficient of variation [CV (%) = SD/mean × 100], resulting from the assay data.
■
RESULTS AND DISCUSSION Galactose Quantification Using Arrays Containing Two Bioluminescent E. coli Strains. The design of the new cell-based galactosemia diagnosis system is based on a strategy in which two genetically engineered E. coli strains are employed to determine the amount of galactose in a test sample. One, a galT knockout strain of E. coli, was constructed by using chromosomal gene deletion of the galT gene of E. coli, which is involved in the biosynthesis of UDP−galactose from galactose-1-phosphate. This gene deletion disrupts normal metabolism of galactose and leads to accumulation of galactose11086
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
Analytical Chemistry
Article
Figure 3. Evaluation of analytical capability of the assay for galactose quantification of artificial samples. Galactose quantification in M9 media containing 2, 8, 16, 25, and 40 μM of galactose (A) without and (B) with additional 60 μM glucose. Above each graph are given a luminescent image produced by galactose at different concentrations. All samples were analyzed with three replicates in a single run, and the results are averaged.
Figure 4. Evaluation of analytical capability for galactose quantification from blood paper specimens. Galactose quantification of (A) commercial blood or (B) real infant blood samples. All samples were analyzed with three replicates in a single run, and the results are averaged.
concentrations of glucose or galactose was examined. Inspection of plots of luminescence signals produced in association with cell growth versus glucose or galactose concentrations show that galT(−) cell growth is linearly dependent on glucose but not on galactose concentrations (Figure 2, A and B). This observation confirms the expectation that deletion of the galT gene prevents galactose promoted growth of E. coli. In contrast, normal E. coli cells grow in a manner that is linearly dependent on the concentrations of both glucose and galactose (Figure 2, panels C and D). The data obtained from these experiments were used to construct the standard curves employed in the galactose quantification procedure. To evaluate precision and reproducibility of the standard curves, we analyzed the average linearity and standard deviation arising from three measurements on a single set of experiments. As a result, the precision was quite high, yielding CVs less than 5.9%, which demonstrate that the standard curves are sufficiently accurate for use in the galactose assay (Figure S1 of the Supporting Information). To verify the analytical capability of the new galactose assay method, artificial test samples were prepared in M9 media containing galactose at concentrations of 2, 8, 16, 25, and 40 μM, with or without added glucose at a fixed concentration of 60 μM. First, the glucose concentration of test samples containing glucose was determined using the RLU value arising from galT(−) cell growth induced by this sugar. Next, galactose concentrations were determined by employing the subtractive
known concentrations are applied to both cell A and cell B regions (usually the first rows of cell A and cell B regions), while galactose solutions of different known concentrations are applied only to the cell B region (usually the second row). Samples being tested are usually applied to the last rows of both the cell A and cell B regions. In response to added glucose and galactose, growth of the two cells takes place on the array and luminescent signals are generated. The measured luminescence intensities from regions (1), (2), and (3) are used to construct three standard curves that correlate the relative luminescence unit (RLU) values obtained from wells containing each type of cells with the concentrations of glucose or galactose. (1) The first step in determining the amount of galactose in a test sample is to determine its glucose concentration based on a comparison of the RLU values obtained from analysis of the glucose quantification region for cell A growth with those obtained by using the standard curve. (2) The RLU value associated with cell B growth solely promoted by glucose is next determined using the standard curve. (3) Finally, the amount of galactose present in the test sample is obtained by subtraction of the RLU value corresponding to glucose promoted growth from the RLU value arising from analysis of the galactose quantification region, which corresponds to cell B growth induced by both glucose and galactose. Analytical Capability for Galactose Quantification. The existence of correlations between luminescence associated with the growth of normal and galT(−) E. coli cells and 11087
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
Analytical Chemistry
Article
Table 1. Galactosemia Diagnosis Using Real Infant Blood Samples actual galactose concentration (μM)
added (μM)
expected (μM)
measureda (μM)
SDb (μM)
CVc (%)
recoveryd (%)
diagnosise
2.73 3.98 5.24 6.60
5 8 11 14
7.73 11.98 16.24 20.60
7.03 10.82 17.95 22.08
0.72 0.94 1.81 2.37
10.24 8.69 10.08 10.73
90.9 90.3 110.5 107.2
normal normal galactosemia galactosemia
Mean of three measurements. bStandard deviation. cCoefficient of variation. dMeasured value/expected value × 100. eThe cutoff value for galactosemia is 14.8 μM. a
10.73%. These values demonstrate the assay method is sufficiently accurate to be applied for galactosemia diagnosis in real clinical applications.
methodology described in Figure 1. The data obtained by assaying the test samples in this way are displayed as bar graphs in Figure 3. The results show that the new method correctly quantifies the amount of galactose present in test samples containing galactose exclusively and ones comprised of mixtures of galactose and glucose. The reproducibility of the assay is high such that the CVs for almost all measurements are less than 10% (Tables S-1A and S-1B of the Supporting Information). The precision of the assay for galactose quantification was assessed using blood paper specimens. First, by employing filter papers with varying amounts of absorbed glucose or galactose, three standard curves were constructed in order to correlate cell growth of the two bioluminescent E. coli strains with the amounts of absorbed glucose or galactose (Figure S-2 of the Supporting Information). With the employment of the assay technique along with the constructed standard curves, the amounts of galactose (4.22, 7.50, and 10.14 μM) absorbed on the blood paper samples were determined correctly with recovery percentages of 90.3−110.9% and CVs in the range of 6.62−7.99% (Figure 4A and Table S-2A of the Supporting Information). The method was also utilized to determine various amounts of galactose (2.73, 3.98, 5.24, 6.60, and 9.45 μM) present in real infant blood paper samples with high precision (Figure 4B and Table S-2B of the Supporting Information). Galactose Assay for Galactosemia Diagnosis. Galactosemia is an inherited autosomal-recessive disorder associated with galactose metabolism. Because it can lead to severe clinical consequences and life-threatening situations,5 this disease needs to be diagnosed at a very early stage of newborn development and treatment needs to be started as soon after birth as possible. Consequently, a significant need exists for simple and rapid methods that diagnose galactosemia, which in its most common form is characterized by the presence of high galactose concentrations in blood. To address this issue, we explored the diagnostic capability of the new cell-based galactose assay method by performing measurements on test samples comprised of four real infant blood paper samples (Table 1). The original galactose concentrations (2.73, 3.98, 5.24, and 6.60 μM) of the samples were determined by using the conventional enzymatic method. Different amounts of galactose (5, 8, 11, or 14 μM) were added to one set of the paper samples to create sets containing two normal and two galactosemia-type paper samples. The cutoff value for galactosemia in our assay is 14.8 μM, which is the diluted concentration from the original cutoff value of 444 μM22−24 with a volume ratio of 1:30. The assay results show that the new method reliably determined galactose concentrations in the samples and, as a result, successfully diagnosed galactosemia in all tested blood paper samples. Importantly, the precision and reproducibility of the assay were quite high, yielding recovery rates in the range of 90.3−110.5% and CVs less than
■
CONCLUSIONS A new cell-based assay method, employing rapidly growing luminescent E. coli strains, was developed to determine galactose concentrations and to diagnose galactosemia using newborn baby blood samples. The assay, performed using a solid-phase array format, displays excellent linearity for galactose quantification and high precision and reproducibility. In addition, the method is both rapid and convenient, it does not require any pre- and post-treatment, and it gives results within a 4 h period. On the basis of the results described above and the fact that the new procedure overcomes the laborious and time- and cost-consuming drawbacks of conventional methods, we strongly believe that the new galactose assay will serve as a viable alternative for high-throughput screening for galactosemia in newborn babies.
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Address: Department of Chemical and Biomolecular Engineering, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea. E-mail:
[email protected]. Tel: +82-42350-3932. Fax: +82-42-350-3910. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This investigation was supported by grants from Basic Science and Public Welfare & Safety research program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (Grant 20090080602) and (Grant 2012M3A2A1051683), the Technology Development Program (Grant 609002-5 and Grant 110065-3) for Agriculture and Forestry of Ministry for Food, Agriculture, Forestry and Fisheries, Republic of Korea.
■
REFERENCES
(1) Isselbacher, K. J.; Anderson, E. P.; Kurahashi, K.; Kalckar, H. M. Science 1956, 13 (123), 635−636. (2) Holden, H. M.; Rayment, I.; Thoden, J. B. J. Biol. Chem. 2003, 278, 43885−43888. (3) Segal, S. Mol. Genet. Metab. 2004, 81, 253−254. (4) Bosch, A. M. J. Inherited Metab. Dis. 2006, 29 (4), 516−525. (5) Leslie, N. D. Annu. Rev. Nutr. 2003, 23, 59−80. (6) Beutler, E.; Baluda, M. J. Lab. Clin. Med. 1966, 68, 137−141.
11088
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
Analytical Chemistry
Article
(7) Akie, F.; Yoshiyuki, O.; Tomiko, M.; Gen, I.; Toshiaki, O. Clin. Chem. 2000, 46 (6), 806−810. (8) Paigen, K.; Pacholec, F.; Levy, H. L. J. Lab. Clin. Med. 1982, 99 (6), 895−907. (9) Yoshida, A.; Tadokoro, Y.; Shima, K.; Ishii, S. Bull. Jpn. Soc. Screening Metab. Disord. Relat. Dis. 1979, 4, 121−123 (in Japanese). (10) David, C. J.; Daniel, V.; Adam, O.; Robert, G. Clin. Biochem. 1987, 20, 353−357. (11) Mohammad, R. S.; Sadeq, V. J. Appl. Sci. 2008, 8 (17), 3026− 3031. (12) Diepenbrock, F.; Heckler, R.; Schickling, H.; Engelhard, T.; Bock, D.; Sander, J. Clin. Biochem. 1992, 25 (1), 37−39. (13) Hongl, S. P.; Yoon, H. R.; Kim, M. K. Chromatographia 2001, 54, 83−86. (14) Jeong, J. S.; Yoon, H. R.; Hong, S. P. J. Chromatogr., A 2007, 1140, 157−162. (15) Ko, D. H.; Jun, S. H.; Park, H. D.; Song, S. H.; Park, K. U.; Kim, J. Q.; Song, Y. H.; Song, J. Clin. Chem. 2010, 56 (5), 764−771. (16) Kim, M. I.; Shim, J.; Li, T.; Woo, M. A.; Cho, D.; Lee, J.; Park, H. G. Analyst 2012, 137 (5), 1137−1143. (17) Datsenko, K. A.; Wanner, B. L. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 6640−6645. (18) Burkholder, P. R. Science 1951, 114, 459−460. (19) Kim, M. I.; Yu, B. J.; Woo, M. A.; Cho, D.; Dordick, J. S.; Cho, J. H.; Choi, B. O.; Park, H. G. Anal. Chem. 2010, 82, 4072−4077. (20) Woo, M. A.; Kim, M. I.; Yu, B. J.; Cho, D.; Kim, N. J.; Cho, J. H.; Choi, B. O.; Chang, H. N.; Park, H. G. Anal. Chem. 2011, 83 (8), 3089−3095. (21) Kalckar, H. M.; Kurahashi, K.; Jordan, E. Proc. Natl. Acad. Sci. U.S.A. 1959, 45 (12), 1776−1786. (22) Jensen, U. G.; Brandt, N. J.; Christensen, E.; Skovby, F.; Nørgaard-Pedersen, B.; Simonsen, H. Clin. Chem. 2001, 47 (8), 1364− 1372. (23) Park, I. S.; Cho, H. J.; Lee, D. H.; Song, J. H. Korean J. Pediatr. 2003, 46 (5), 440−446. (24) Jeong, J. S.; Kwon, H. J.; Yoon, H. R.; Lee, Y. M.; Choi, T. Y.; Hong, S. P. Anal. Biochem. 2008, 376 (2), 200−205.
11089
dx.doi.org/10.1021/ac4027912 | Anal. Chem. 2013, 85, 11083−11089
