High-Speed Mass Analysis of Whole Erythrocytes by Charge

Sep 5, 2007 - audio frequency region, Peng et al.7 demonstrated that it is possible to measure the dry mass of the ..... amplifier, driven by a synthe...
1 downloads 10 Views 381KB Size
Download PDF
Anal. Chem. 2007, 79, 7401-7407

High-Speed Mass Analysis of Whole Erythrocytes by Charge-Detection Quadrupole Ion Trap Mass Spectrometry Zongxiu Nie,†,‡ Fenping Cui,†,§ Yan-Kai Tzeng,| Huan-Cheng Chang,*,†,|,⊥ Minglee Chu,# Huan-Chang Lin,⊥ Chung-Hsuan Chen,⊥ Hsin-Hung Lin,⊥ and Alice L. Yu⊥

Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 106, Taiwan, Department of Physics, Wuhan University, Wuhan 430072, China, College of Mathematics and Physics, Nanjing University of Information Science and Technology, Nanjing 210044, China, Department of Chemistry, National Taiwan Normal University, Taipei 106, Taiwan, Genomics Research Center, Academia Sinica, Taipei 115, Taiwan, and Institute of Physics, Academia Sinica, Taipei 115, Taiwan

Herein, we report an application of charge-detection quadrupole ion trap mass spectrometry to the measurement of total dry masses of mammalian and poultry erythrocytes evaporated/ionized by laser-induced acoustic desorption. The method is rapid and widely applicable. Eight different types of red blood cells (RBCs) have been successfully analyzed, including those of human, goat, cow, mouse, pig, and chicken. The measured mean masses (weights per corpuscle) range from 0.58 × 1013 Da (9.6 pg) of goat RBCs to 2.80 × 1013 Da (46.5 pg) of chicken RBCs. The total dry weights determined for human RBCs from a healthy male adult, a patient with iron-deficiency anemia, and a patient with thalassemia are 34.8, 28.8, and 20.6 pg, respectively. These weights, except that of thalassemia, are all ∼10% higher than their corresponding mean corpuscular hemoglobin values determined by a commercial automated hematology analyzer. The mass distribution profiles of the cells are all near-Gaussian, with a standard deviation of 15% for the normal human RBCs. The deviation increases significantly to 20% for RBCs with thalassemia characteristics and 27% for RBCs with iron-deficiency anemia characteristics. All the observations are in accord with their corresponding mean corpuscular volume measurements, indicating an increase in anisocytosis (variation in RBC size) in the anemic samples. Our results suggest a broad and promising application of this new technology to high-speed mass analysis of RBCs and other biological whole cells as well. Since the invention of electrospray ionization (ESI) and matrixassisted laser desorption/ionization,1-3 mass spectrometry (MS) has been well received as one of the key technologies for the * To whom correspondence should be addressed. E-mail: hcchang@ po.iams.sinica.edu.tw. † Institute of Atomic and Molecular Sciences, Academia Sinica. ‡ Wuhan University. § Nanjing University of Information Science and Technology. | National Taiwan Normal University. ⊥ Genomics Research Center, Academia Sinica. # Institute of Physics, Academia Sinica. (1) Wong, S. F.; Meng, C. K.; Fenn, J. B. J. Phys. Chem. 1988, 92, 546-550. (2) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. 10.1021/ac071207e CCC: $37.00 Published on Web 09/05/2007

© 2007 American Chemical Society

high-throughput mass analysis of biological macromolecules and synthetic polymers. The subsequent development of soft ionization techniques such as laser-induced acoustic desorption (LIAD) added further versatilities to the technique,4-6 allowing intact viruses and whole cells to be evaporated to the gas phase for mass analysis.7-9 Using a quadrupole ion trap (QIT) operated in the audio frequency region, Peng et al.7 demonstrated that it is possible to measure the dry mass of the individual bioparticles inside the QIT based on a charge-differential technique.10-12 They further demonstrated recently that the speed of the mass analysis for micron-sized particles can be greatly increased by charge detection of the particles ejected axially from the QIT.9 Two quantities, the mass-to-charge ratio (m/Ze) and the absolute number of charges (Z), were measured independently. Such a detection scheme was adapted from the pioneering work of Benner and co-workers who detected and analyzed megadalton DNA,13-15 polymeric nanoparticles,13 and intact viruses16 using a (3) Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2, 151-153. (4) Lindner, B. Int. J. Mass Spectrom. Ion Processes 1991, 103, 203-218. (5) Golovlev, V. V.; Allman, S. L.; Garrett, W. R.; Chen, C. H. Appl. Phys. Lett. 1997, 71, 852-854. Golovlev, V. V.; Allman, S. L.; Garrett, W. R.; Taranenko, N. I.; Chen, C. H. Int. J. Mass Spectrom. Ion Processes 1997, 169, 69-78. (6) Perez, J.; Ramirez-Arizmendi, L. E.; Petzold, C. J.; Guler, L. P.; Nelson, E. D.; Kenttamaa, H. I. Int. J. Mass Spectrom. 2000, 198, 173-188. Shea, R. C.; Petzold, C. J.; Campbell, J. L.; Li, S.; Aaserud, D. J.; Kenttamaa, H. I. Anal. Chem. 2006, 78, 6133-6139. Shea, R. C.; Habicht, S. C.; Vaughn, W. E.; Kenttamaa, H. I. Anal. Chem. 2007, 79, 2688-2694. (7) Peng, W. P.; Yang, Y. C.; Kang, M. W.; Tzeng, Y. K.; Nie, Z. X.; Chang, H. C.; Chang, W.; Chen, C. H. Angew. Chem., Int. Ed. 2006, 45, 1423-1426. (8) Nie, Z. X.; Tzeng, Y. K.; Chang, H. C.; Chiu, C. C.; Chang, C.-Y.; Chang, C. M.; Tao, M. H. Angew. Chem., Int. Ed. 2006, 45, 8131-8134. (9) Peng, W. P.; Lin, H. C.; Lin, H. H.; Chu, M.; Yu, A. L.; Chang, H. C.; Chen, C. H. Angew. Chem., Int. Ed. 2007, 46, 3865-3869. (10) Arnold, S. J. Aerosol Sci. 1979, 10, 49-53. Philip, M. A.; Gelbard, F.; Arnold, S. J. Colloid Interface Sci. 1983, 91, 507-515. (11) Schlemmer, S.; Illemann, J.; Wellert, S.; Gerlich, D. J. Appl. Phys. 2001, 90, 5410-5418. Schlemmer, S.; Wellert, S.; Windisch, F.; Grimm, M.; Barth, S.; Gerlich, D. Appl. Phys. A: Mater. Sci. Process. 2004, 78, 629-636. (12) Peng, W. P.; Yang, Y. C.; Kang, M. W.; Lee, Y. T.; Chang, H.-C. J. Am. Chem. Soc. 2004, 126, 11766-11767. (13) Fuerstenau, S. D.; Benner, W. H. Rapid Commun. Mass Spectrom. 1995, 9, 1528-1538. (14) Schultz, J. C.; Hack, C. A.; Benner, W. H. J. Am. Soc. Mass Spectrom. 1998, 8, 305-313. (15) Schultz, J. C.; Hack, C. A.; Benner, W. H. Rapid Commun. Mass Spectrom. 1999, 13, 15-20.

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007 7401

charge-detection time-of-flight mass spectrometer. Hundreds of cells or synthetic microparticles can be analyzed in an hour. However, to take on this detection scheme, the analyte particles have to carry thousands of charges in order to defeat the high electronic backgrounds (typically 500 electron charges). Both ESI and LIAD are useful ion sources for this application, and it has been proven that highly charged particles can be readily generated by LIAD when silicon wafers are employed as the acoustic desorption substrate without the need of external ionization.7-9 Red blood cells (RBCs) are the most common cells in blood. They are readily accessible and well suited for the total dry mass measurement with charge-detection LIAD-QIT-MS. Human RBCs are biconcave disks with a diameter of 8.5 µm and greatest and least thicknesses of 2.4 µm and 1.0 µm, respectively.17 Hemoglobin is the most abundant component (>90%) in the RBC.18 An index commonly used to characterize the RBCs is MCH (mean corpuscular hemoglobin), which is a measure of the average weight of hemoglobin per corpuscle.19 The MCH index provides a useful reference for baseline screening and the differential diagnosis of various disease states. The measurement of MCH is also of interest in various pathologic conditions and in the study of the drug-induced destruction of RBCs. Various methods for the quantitative characterization of hemoglobin have been developed in the past few decades.19 Presently, one of the most used methods to measure MCH levels is the photometric detection of cyanmethemoglobin, which is a stable compound derived from hemoglobin. The advantage of using this method is that it is simple, rapid, and does not require sophisticated equipment. However, the chemicals used in the assay are toxic, and in addition, the method can only provide a mean value of the hemoglobin content per corpuscle. Alternative detection methods are therefore needed to obtain the RBC mass distribution, which sometimes is more informative than the corresponding mean mass value in hematological diagnosis. Measurement of the total dry mass (and the dry mass distribution) of human erythrocytes was a subject of active research in the 1950s using interference microscopy.20-24 In that method, RBCs were inspected individually under a microscope to determine the optical path differences (φ) between water and the cells. The dry mass (m) of each cell was then determined from the relation m ) φA/100R, where A is the area projected by the cell and R ≈ 0.0018 is the refractive increment of the substance other than water in the cell.25 Uncertainties in this mass measurement due to the variation of R with the cell constituents were in the range of (10%. To speed up the mass measurement process, particularly for the dry mass distribution, Mitchison et al.21 showed (16) Fuerstenau, S. D.; Benner, W. H.; Thomas, J. J.; Brugidou, C.; Bothner, B.; Siuzdak, G. Angew. Chem., Int. Ed. 2001, 40, 542-544. (17) Weinstein, R. S. In The Red Blood Cell, 2nd ed.; Surgenor, D. M., Ed.; Academic: New York, 1975; Chapter 5. (18) Pennell, R. B. In The Red Blood Cell, 2nd ed.; Surgenor, D. M., Ed.; Academic: New York, 1975; Chapter 3. (19) Mosby’s Manual of Diagnostic and Laboratory Tests; Pagana, K. D., Pagana, T. J., Eds.; Mosby: St. Louis, MO, 1998. (20) Largerlof, B.; Thorell, B.; Akerman, L. Exp. Cell Res. 1956, 10, 752-754. (21) Mitchison, J. M.; Passano, L. M.; Smith, F. H. Q. J. Microsc. Sci. 1956, 97, 287-302. (22) Ponder, E. Nature 1959, 183, 1330-1331. (23) Gamble, C. N.; Glick, D. J. Biophys. Biochem. Cytol. 1960, 8, 53-60. (24) Thaer, A. J. Microsc. 1969, 89, 237-250. (25) Davies, H. G.; Wilkins, M. H. F.; Chayen, J.; La Cour, L. F. Q. J. Microsc. Sci. 1954, 95, 271-304.

7402

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

that an approximately linear relationship between the mass and the integrated intensity over the image of the isolated cell can be achieved if the refractive index difference between the medium and the RBCs is properly adjusted with albumin. Utilizing this technique, Thaer24 measured the integrated image intensities of the individual cells in an albumin solution and determined fairly rapidly (∼100 cells within 60 min) their relative dry masses. A mass distribution of approximately (20% in width was observed by the author for both reticulocytes and mature human RBCs. This article describes a mass-spectrometric approach toward high-speed measurements for the total dry masses and dry mass distributions of human and animal RBCs. Masses of the individual cells were determined by two independent measurements for m/Ze using a frequency-scan QIT and for Z using an image-current detection plate. Compared with quantitative light microscopy, charge-detection LIAD-QIT-MS is a more direct approach with a potential for broader biological applications. The result so obtained is completely independent of factors such as the shape, size, density, refractive index, and absorption characteristics of the analyzed cells. The method is rapid, the spectrometer is compact, and, more importantly, it can be operated by automation. More than 600 RBCs can be analyzed in 1 h, and both the mean mass and the mass distribution width of the cells can be obtained simultaneously. Moreover, the identification of anemic human RBCs can be readily achieved with this whole-cell mass spectrometer. EXPERIMENT Polystyrene size standards (SRM 1692) with a nominal diameter of 3 µm were obtained from the U.S. National Institute of Standards and Technology (NIST). The sample, a colloidal suspension, contained 0.25% (w/v) monodisperse polystyrene spheres and 50 ppm sodium azides as biocides. The certified number-averaged diameter of the spheres is 2.982 µm, with a standard deviation (SD) of 0.016 µm. To remove the sodium azide contaminants from the suspension, the as-received polystyrene microspheres were thoroughly washed with deionized water several times and recovered by centrifugation. They were resuspended in filtered (0.2 µm pore size filter) distilled water at a concentration of ∼1 × 1010 particles/mL before use. Anticoagulated human blood samples (typically 2 mL) were taken from a healthy male adult and patients with iron-deficiency anemia and thalassemia. Animal blood samples including those of indigenous goat, cow, pig, and chicken were obtained from the Department of Veterinary Medicine and the Department of Animal Science and Technology of National Taiwan University. The blood sample of mice (the BALB/c strain) was obtained from the Genomics Research Center, Academia Sinica. RBCs of all the samples were separated by centrifugation at 390g and rinsed 3 times in phosphate buffer saline (PBS) at room temperature. Prior to the mass measurement, they were fixed by mixing 10 mL of 0.05% (v/v) glutaraldehyde in PBS with 1 mL of the centrifugepacked RBCs for 1 h.26,27 The fixed RBCs were then thoroughly washed with deionized water, separated by centrifugation, and resuspended in filtered distilled water at a concentration of ∼1 × 107 cells/mL. The corresponding red-cell indices (such as MCH, (26) Morel, F. M. M.; Baker, R. F.; Wayland, H. J. Cell Biol. 1971, 48, 91-100. (27) The mass increase due to the cell fixation with 0.05% glutaraldehyde for 1 h was estimated to be less than 1% (ref 26).

Figure 1. Whole erythrocyte mass spectrometer consisting of a LIAD ion source, a QIT mass analyzer, and a charge-sensitive detector.

mean corpuscular volume (MCV), and red-cell distribution width (RDW)) were measured independently using an automated hematology analyzer (Sysmex, XT1800i) before the cell fixation. Figure 1 shows a three-dimensional drawing of the experimental setup. It consisted of a QIT mass analyzer, a LIAD ion source, and a charge-monitoring detector with a design identical to that described previously.9 To prepare the sample for LIAD, an aliquot (10 µL) of the aforementioned purified particle (or cell) suspension was deposited on a 0.5 mm thick Si wafer without any matrix. After being air-dried, the sample-loaded Si wafer was positioned near one of the four holes (3.1 mm in diameter) drilled orthogonally on the ring electrode (10 mm in radius) of the QIT. A frequency-doubled Nd:YAG laser (wavelength of 532 nm, pulse width of 7 ns, and pulse energy of 30 mJ) was shined on the backside of the Si wafer. Particles, desorbed acoustically, entered the trap through the hole and were captured by the ion trap operating at a frequency of 450 Hz and a voltage amplitude of 1510 V in the presence of 40 mTorr of helium as buffer gas. A He-Ne laser (632 nm) illuminated the charged particles in the trap center for visual monitoring with a charge-coupled device camera. The frequency scan of the QIT was conducted in a monopolar mode,28 where the ring electrode was driven by a power amplifier and the endcap electrodes were electrically grounded. The amplifier, driven by a synthesized function generator, provided ac voltages with the frequency variable from 10 to 700 Hz at a constant amplitude.29 No significant distortion of the sinusoidal waveform was detected at the peak-to-peak amplitude of less than 3500 V. Scanning the ac frequency from 450 to 100 Hz at a constant voltage amplitude (zero-to-peak) of 1510 V ejected the trapped particles from the QIT through the holes (3.1 mm in diameter each) on the endcap electrodes in a mass-selective axial instability mode.28 A home-built charge-sensitive detection plate (10 mm in diameter), located ∼10 mm away from the exit of the endcap electrode, measured the absolute number of the charges carried by the ejected particles. The detector comprised a charge (28) March, R. E.; Hughes, R. J. Quadrupole Storage Mass Spectrometry; Wiley: New York, 1989; Chapter 2. (29) Ting, J. EDN 2001, 46, 136-138.

Figure 2. Electrical signals of a typical mass spectrum scan. Each negative pulse represents the event when a single positively charged red blood cell struck the charge collection plate. The background rms noise was measured to be 500 e. The height of each peak gives a direct measure of the number of the charges carried by the detected particle.

integrator circuit built next to the solder-coated conducting plate. The integrator used a low-noise junction field-effect transistor as the input stage and an operational amplifier for further amplification and basic low-pass noise filtering. Elimination of the ac pickup from the trapping field was accomplished by placing an electrically grounded mesh (85% transmission) in front of the detector. The typical background in the charge detection was 500 electron charges (Figure 2). The time of the scan to cover the entire frequency range was 5 s. RESULTS AND DISCUSSION When a particle is ejected axially from the QIT operating in the monopolar frequency-scan mode,30,31 its mass-to-charge ratio can be evaluated according to the equation28

m/Ze ) 4V/qejectr02Ωeject2

(1)

where m is the mass, Z is the charge number, e is the elementary charge, V is the amplitude (zero-to-peak) of the trapping ac field, qeject is the point of ejection (0.908 for the ideal case), r0 is the radius of the ring electrode, and Ωeject is the frequency at the point of the particle ejection. For a micron-sized particle ejected from the QIT in the presence of 1 mTorr He, our previous calibration of the ion trap operating in the voltage scan mode indicated that there is a delayed ejection due to the buffer gas damping effect.32 The point of the particle ejection qeject increased significantly from 0.908 to 0.949. This delayed ejection, however, was not fully characterized in the present case, which gives rise to an uncertainty in evaluating m/Ze using eq 1. Another uncertainty associated with the mass determination lies in Z. In the Z measurement, the image charge was integrated over a small capacitor (1 pF) in the charge-detection circuit that converted the collected charge to a voltage. However, the capacitance of the capacitor was marked (30) Schlunegger, U. P.; Stoeckli, M.; Caprioli, R. M. Rapid Commun. Mass Spectrom. 1999, 13, 1792-1796. (31) Cai, Y.; Peng, W.-P.; Chang, H.-C. Anal. Chem. 2003, 75, 1805-1811. (32) Cai, Y.; Peng, W. P.; Kuo, S. J.; Lee, Y. T.; Chang, H. C. Anal. Chem. 2002, 74, 232-238. Cai, Y.; Peng, W.-P.; Kuo, S.-J.; Chang, H.-C. Int. J. Mass Spectrom. 2002, 214, 63-73.

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

7403

Table 1. Comparison of the Measured Total Dry Weights with the Corresponding Red-Cell Indices of Mammalian and Poultry Erythrocytesa sources human (normal) human (iron deficiency)b human (thalassemia)b goat cow mouse pig chicken

MCW RDWm MCV RDW MCH MCH/ (pg) (%) (fL) (%) (pg) MCW 34.8

15

88.9

12

31.0

0.89

28.8

27

80.4

17

26.5

0.92

20.6

20

66.5

18

21.5

1.04

9.6 15.5 18.4 20.7 46.5

16 20 11 17 13

34.4 40.4 46.2 60.4 130.8

30 27 20 22 9

14.8 16.2 19.3 33.5

0.95 0.88 0.93 0.72

a Key to abbreviations: MCW, mean corpuscular weight; MCH, mean corpuscular hemoglobin; MCV, mean corpuscular volume; RDW, red-cell distribution width; RDWm, red-cell distribution width determined by the total dry mass measurement. b Averages of four independent measurements with an uncertainty of (1% in MCW.

Figure 3. Mass histograms of (a) polystyrene size standards (SRM 1692), (b) normal human erythrocytes, (c) cow erythrocytes, and (d) goat erythrocytes. The asterisk in d indicates dimer formation. All the samples were analyzed under the same experimental conditions. The typical number of the cells loaded on the Si wafer to obtain the histograms in panels b-d is 2 × 105. Insets: Photos of the corresponding polystyrene spheres and glutaraldehyde-fixed cells. The scale bar is 10 µm in width.

to have a (25% tolerance, which introduces a fixed error in the mass measurement. We therefore calibrated the spectrometer (both accuracy and resolution) against NIST polystyrene size standards, which have been suggested to serve as a mass standard in particle MS.33,34 The polystyrene microspheres used in this work have a certified mean diameter of 2.982 µm, corresponding to a molar (33) Peng, W. P.; Yang, Y. C.; Lin, C. W.; Chang, H. C. Anal. Chem. 2005, 77, 7084-7089. (34) Trevitt, A. J.; Wearne, P. J.; Bieske, E. J. Int. J. Mass Spectrom. 2007, 262, 241-246.

7404 Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

mass of 0.881 × 1013 Da (density 1.055 g/cm3).35 The SD of the size distribution is 0.016 µm, which translates to a coefficient of variance (CV, defined as the ratio of SD to the mean) of 0.54% in size or 1.6% in mass. Figure 3a shows a mass histogram of the polystyrene size standards measured by LIAD-QIT-MS. The observed mass distribution is characterized by a SD ) 0.14 × 1013 Da or CV ) 16%. Since the mass variation of the polystyrene microspheres is only (1.6%, the observed (16% mass distribution width must be a combined result of the low mass resolution of the QIT operated in the frequency scan mode and the moderate signal-to-noise ratios (typically 5) of the image currents recorded by the charge detector.36 Seven different types of mammalian RBCs have been examined in this study, including those of human (a healthy male adult, a patient with iron-deficiency anemia, and a patient with thalassemia), goat, cow, pig, and mouse. These RBCs bear the common feature that they do not contain a cell nucleus and thus contain no DNA. Shown in parts b-d of Figure 3 are the typical mass histograms of RBCs from normal human, cow, and goat, respectively. They all exhibit a fairly symmetrical mass distribution profile with respect to the mean mass, although doublets are often found for the goat RBCs. The measured mean mass for the human RBCs was 2.10 × 1013 Da, obtained by fitting the mass histogram with a Gaussian profile (blue curve in Figure 3b). This molar mass corresponds to a mean corpuscular weight (MCW) of 34.8 pg. Compared with this weight, the MCH value (31.0 pg) determined by the automated hematology analyzer is ∼10% smaller. The ratio, i.e., MCH/MCW ∼ 0.9, similarly holds for RBCs of other mammals such as pig, mouse, and cow (Table 1). Since MCH only quantifies the total weight of hemoglobin (∼270 million molecules in each human erythrocyte) in the RBC, this ∼10% mass increment in MCW is most likely contributed by membranes, membrane-associated proteins, enzymes, and other intracellular components.18 The MCH values reported herein were determined by using a commercial instrument, which was optimized for the analysis (35) Kahler, H.; Lloyd, B. J., Jr. Science 1951, 114, 34-35. (36) See ref 9 for the charge histogram.

of RBCs with MCV in the range of 30-250 fL (10-15 liters). Shown in Table 1 are the MCV values of the mammalian RBCs examined in this work, and they all fall in this volume range. However, it is known that goat RBCs have the smallest diameter (3.2 µm) of the domestic animals’ RBCs and are most variable in shape.37 The goat RBCs presently examined have a MCV value of 34.4 fL, which is very close to the lower detection limit of the instrument. As a result, we could not obtain a reliable MCH value for this sample. The mass-spectrometric measurement, in contrast, has no such limitation. The measured value of MCW ) 9.6 pg is in good agreement with the literature values of MCH ) 8.2-9.5 pg as determined by Tibbo et al.38 using conventional methods to measure the RBC count and the hemoglobin concentration. The ability to analyze RBCs with the total dry weight ranging from 9 pg of goat to 50 pg of chicken (vide infra) demonstrates the generality of this method for hematological diagnosis. This MS-based approach is also well suited for evaluation of various types of anemias, since the disease is characterized by a decrease in RBC mass.39 Figure 4 compares the results of the mass measurement between normal and anemic RBCs. For the RBCs from a healthy male adult, the measured mean mass is 2.10 × 1013 Da, with a SD of 0.46 × 1013 Da in the mass histogram (Figure 4a). It is noted that this observed mass distribution profile is a result of the convolution of the source and the instrumental functions. Assuming that both functions are Gaussian, the SD of the observed mass distribution (also a Gaussian) follows the relation40

SD2 ) SDs2 + SDi2

(2)

where SDs and SDi are the standard deviations of the source and the instrumental functions, respectively. To deduce the information of SDs (or CV) in eq 2 for RBCs, we assume that the QIT mass spectrometer has the same resolution (i.e., CV ) 16% in Figure 3a) throughout the mass analysis range covered by this work. At mass in the range of 2 × 1013 Da, we have SDi ) 0.34 × 1013 Da. With this SDi and SD ) 0.46 × 1013 Da, an estimate of CV ) 15% (or RDWm in Table 1) was obtained for the intrinsic mass distribution width of the normal RBCs. Notably, the width agrees satisfactorily with the measurement of ∼20% by Thaer24 using interference microscopy. Figure 4b shows the mass histogram of RBCs from a patient with iron-deficiency anemia. Iron deficiency is the most common cause of anemia.39 The disease may be caused by diets low in iron, blood loss, and body changes (such as during periods of rapid growth and pregnancy), etc. For healthy adults, the MCH

(37) Thrall, M. A.; Baker, D.; Campbell, T.; DeNicola, D.; Fettman, M.; Lassen, D.; Rebar, A.; Weiser, G. Veterinary Hematology and Clinical Chemistry; Lippincott Williams & Wilkins: Philadelphia, PA, 2004; p 70. (38) Tibbo, M.; Jibril, Y.; Woldemeskel, M.; Dawo, F.; Aragaw, K.; Rege, J. E. O. Int. J. Appl. Res. Vet. Med. 2004, 2, 297-309. (39) Conrad, M. E. In Clinical Methods: The History, Physical and Laboratory Examinations, 3rd ed.; Walker, H. K., Hall, W. D., Hurst, J. W., Eds.; Butterworth: Boston, MA, 1990; Chapter 147. (40) Arfken, G. Mathematical Methods for Physicists, 3rd ed.; Academic: Orlando, FL, 1985; pp 810-814. See also http://mathworld.wolfram.com/Convolution.html.

Figure 4. Mass histograms of human erythrocytes from (a) a healthy male adult, (b) a patient with iron-deficiency anemia, and (c) a patient with thalassemia. Insets: Photos of the corresponding glutaraldehydefixed cells. The scale bar is 10 µm in width.

value ranges from 27 to 31 pg,41 and it falls below this range for patients with iron-deficiency syndrome. As shown in Table 1, the anemic RBC has an MCH value of 26.5 pg, which is ∼15% smaller than the normal value and ∼10% lower than the total dry weight of MCW ) 28.8 pg measured for this cell. Additionally, the intrinsic mass distribution of this sample is very broad, RDWm ) 27%, which is nearly twice as large as that of the normal RBCs. Such a distinct increase in the variation of the RBC mass serves as a useful indicator for the evaluation of this anemia. Thalassemia is a type of hereditary anemia that results in the production of an abnormal ratio of hemoglobin subunits (i.e., R and β chains).42 Inspection of the RBCs from the patient with this anemia under an optical microscope indicated that they are thinner than the normal cells but are more or less uniform in diameter (inset in Figure 4c for the glutaraldehyde-fixed cells). Quantitative measurement by the LIAD-QIT-MS revealed that these cells have a mean mass of 1.24 × 1013 Da, accompanied with a SD ) 0.31 × 1013 Da (Figure 4c). To our surprise, this measured MCW value (20.6 pg) is slightly smaller than its MCH counterpart (21.5 pg), (41) Sarma, P. R. In Clinical Methods: The History, Physical and Laboratory Examinations, 3rd ed.; Walker, H. K., Hall, W. D., Hurst, J. W., Eds.; Butterworth: Boston, MA, 1990; Chapter 152. (42) Bank, A.; Rifkind, R. A.; Marks, P. A. In The Red Blood Cell, 2nd ed.; Surgenor, D. M., Ed.; Academic: New York, 1975; Chapter 22.

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

7405

Figure 6. Mass histogram of 1:2 mixture of normal human and cow erythrocytes. Inset: Photo of the corresponding glutaraldehyde-fixed cells. The scale bar is 10 µm in width.

Figure 5. Mass histograms of (a) 2:1 and (b) 1:1 mixtures of human erythrocytes from a healthy male adult and a patient with thalassemia.

contrary to the result of MCW > MCH for normal human and other mammalian RBCs (Table 1). A possible explanation for this discrepancy is that the MCH value reported here was determined by using a colorimetric method, which probes only the heme groups of hemoglobin.19 Therefore, it is an indirect measure for the total dry weight of hemoglobin in each cell. The present observation of MCH ≈ MCW implies that there is a significant decrease (up to 10%) in the amount of either R or β chains of the hemoglobin molecules in this patient’s RBCs, and this decrease was not properly reflected by the measured MCH. Compared with the normal cells, these anemic RBCs not only have a smaller total dry weight but also exhibit a wider distribution. The intrinsic mass distribution width of the cells is RDWm ) 20%. Our measured mass distribution width can be compared closely with RDW, a red-cell index defined as the coefficient of variation of the RBC size.19 This index is useful in hematological diagnosis and also a good indicator of the degree of anisocytosis. In most commercial instruments, the value of this index is determined based on the Coulter principle, which states that when a particle traverses an orifice concurrent with an electrical current, the change in impedance is proportional to the volume of the particle. The method is rapid, with an inaccuracy of ∼3% in the total volume measurement.19 For human RBCs, the normal RDW value is 13 ( 1.5%, which can increase to 15% for anemic RBCs with β-thalassemia and to 20% for RBCs with iron-deficiency anemia.41 Our measurements of RDWm ) 15, 20, and 27% (Table 1) for the intrinsic mass distribution widths of the normal RBCs and the two anemic types of RBCs correlate well with the variation of these RDW values. The good correlation supports our previous assessment7 of using the mass-spectrometric method to differentiate the type of anemias. An interesting application of LIAD-QIT-MS to the hematological study is to be able to distinguish normal from abnormal RBCs in a mixture. The feasibility of such an experiment is demonstrated in Figure 5, where the mass histograms of mixed normal and anemic (with thalassemia) RBCs at two different molar ratios are 7406 Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

Figure 7. Mass histogram of chicken erythrocytes. Inset: Photo of the glutaraldehyde-fixed cells. The scale bar is 10 µm in width.

presented. As seen, two peaks can be well identified in the histogram of the 2:1 (normal/anemic) mixture (Figure 5a), even though the total dry weights of these two samples differ by only ∼40%. While the two peaks may no longer be discernible in the mass histogram of the 1:1 mixture (Figure 5b), the asymmetric and deformed shape of the profile clearly indicates that the analyte is composed of more than one type of RBCs. The LIAD-QIT-MS technique can also be applied to distinguish the RBCs of different species.43 For the analyte containing both normal human RBCs and cow RBCs, which differ by a factor of 2 in their total dry masses, a bimodal profile can be more readily found in the mass histogram of the 1:2 mixture (Figure 6). Deconvolution of the histogram with two Gaussian profiles properly reproduces their respective mean masses of 2.0 × 1013 Da for human RBCs and 1.0 × 1013 Da for cow RBCs. Chicken RBCs are the only poultry erythrocytes examined in this work. Unlike that of the mammalian species, the chicken RBC has a nucleus, which appears as an ellipsoidal core in the cell center (inset in Figure 7). For this sample, the mass-spectrometric measurement revealed a peak centering at 2.80 × 1013 Da with a SD ) 0.58 × 1013 Da in the histogram (Figure 7). This measured mass corresponds to a MCW of 46.5 pg, which, interestingly, is ∼30% larger than its MCH value of 33.5 pg. Assuming that the hemoglobin molecules in this cell contribute only 90% of the total cell weight excluding that of the nucleus, we estimate that the nucleus itself has a total dry weight of 9 pg (or 5.6 × 1012 Da in (43) Just, W. W.; Leonv, J. O.; Werner, G. Anal. Biochem. 1975, 67, 590-601.

mass). It is known that the chicken karyotype is made up of 38 autosomes and one pair of sex chromosomes.44 Additionally, the entire chicken genome comprises 1.06 billion base pairs. Taking into account the genome mass of 7 × 1011 Da, our measurement of the nucleus mass suggests that more than 10% of the total dry weight of the chicken nucleus is given by DNA. To the best of our knowledge, this is the first time that the percentage composition (w/w) of DNA in a cell nucleus was determined by MS.45 CONCLUSION We have developed a high-speed spectrometric method capable of measuring the total dry masses and dry mass distributions of red blood cells in the gas phase. More than 600 cells can be analyzed individually in 1 h, and both the mean mass and the mass distribution width of the cells can be obtained simultaneously. The method has been successfully applied to analyzing (44) International Chicken Genome Sequencing Consortium. Nature 2004, 432, 695-716. (45) The determination of the total dry weight also allows us to estimate the content of water in each RBC based on its MCV value (Table 1). Assuming that hemoglobin has an average density of 1.41 g/cm3 (Fischer, H.; Polikarpov, I.; Craievich, A. F. Protein Sci. 2004, 13, 2825-2828), we estimate that the normal human RBC is composed of 65% of water, as opposite to 72% of water for the thalassemia RBC, in isotonic buffered saline.

the erythrocytes of mammals (such as human, goat, cow, mouse, and pig) and poultry (such as chicken). For human erythrocytes, the results we obtained are consistent and also closely correlated with the red-cell indices (such as MCH, MCV, and RDW) commonly provided by automated hematology analyzers equipped in most clinical laboratories. Additionally, our measurements reveal the mass differences not coming from hemoglobin only, which is a piece of valuable information that cannot be obtained by other current methods. Therefore, the method presented in this work is a good complementary technology to other available methodologies. ACKNOWLEDGMENT We thank Prof. Ching-Ho Wang (Department of Veterinary Medicine, NTU), Prof. Pey-Hwa Wang (Department of Animal Science and Technology, NTU), and M. D. Ching-Ling Ho (TriService General Hospital) for providing the blood samples. This research was supported by grants from Academia Sinica and the National Science Council (Grant No. NSC 92-3112-B-001-012-Y) of Taiwan, Republic of China. Received for review June 7, 2007. Accepted July 24, 2007. AC071207E

Analytical Chemistry, Vol. 79, No. 19, October 1, 2007

7407